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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
109c18cc-6532-4c8e-9b1d-7e2b12ea0203 | nnformer-interleaved-transformer-for | 2109.03201 | null | https://arxiv.org/abs/2109.03201v6 | https://arxiv.org/pdf/2109.03201v6.pdf | nnFormer: Interleaved Transformer for Volumetric Segmentation | Transformer, the model of choice for natural language processing, has drawn scant attention from the medical imaging community. Given the ability to exploit long-term dependencies, transformers are promising to help atypical convolutional neural networks to overcome their inherent shortcomings of spatial inductive bias. However, most of recently proposed transformer-based segmentation approaches simply treated transformers as assisted modules to help encode global context into convolutional representations. To address this issue, we introduce nnFormer, a 3D transformer for volumetric medical image segmentation. nnFormer not only exploits the combination of interleaved convolution and self-attention operations, but also introduces local and global volume-based self-attention mechanism to learn volume representations. Moreover, nnFormer proposes to use skip attention to replace the traditional concatenation/summation operations in skip connections in U-Net like architecture. Experiments show that nnFormer significantly outperforms previous transformer-based counterparts by large margins on three public datasets. Compared to nnUNet, nnFormer produces significantly lower HD95 and comparable DSC results. Furthermore, we show that nnFormer and nnUNet are highly complementary to each other in model ensembling. | ['Yizhou Yu', 'Liansheng Wang', 'Lequan Yu', 'Yinghao Zhang', 'Jiansen Guo', 'Hong-Yu Zhou'] | 2021-09-07 | null | null | null | null | ['volumetric-medical-image-segmentation'] | ['medical'] | [ 3.15313905e-01 5.60570061e-01 6.59946306e-03 -6.36377990e-01
-7.65633583e-01 -4.03539211e-01 5.89228570e-01 2.11162016e-01
-4.15526658e-01 6.29691839e-01 3.01842183e-01 -5.56189895e-01
1.90740407e-01 -8.54978800e-01 -8.69521499e-01 -5.21132529e-01
-1.17744997e-01 3.04398924e-01 4.75321025e-01 -1.12098731e-01
-4.07356918e-02 4.13420111e-01 -8.44095349e-01 5.25825202e-01
8.98624599e-01 1.34228516e+00 2.80279696e-01 3.59427184e-01
-4.39690441e-01 1.04308140e+00 -2.19153047e-01 -5.40630102e-01
7.65258148e-02 -3.03559929e-01 -1.12131274e+00 -3.42853665e-01
5.47285795e-01 -4.39723611e-01 -3.89878273e-01 7.48397887e-01
6.33326471e-01 -9.11542252e-02 5.98653436e-01 -7.23433673e-01
-9.44302678e-01 9.63451564e-01 -5.90763450e-01 8.30650508e-01
-1.26646042e-01 3.24249685e-01 9.19694185e-01 -6.89326108e-01
4.35308248e-01 1.10027432e+00 1.01290774e+00 4.64174986e-01
-1.27451670e+00 -4.80691344e-01 1.74265593e-01 -5.61266281e-02
-1.15945506e+00 -2.33415663e-01 6.94094718e-01 -2.53795564e-01
1.29232383e+00 3.50700140e-01 7.75255799e-01 1.08568478e+00
4.06088978e-01 1.01951110e+00 1.02725291e+00 -1.27695441e-01
6.09581023e-02 -4.50748242e-02 1.99876770e-01 7.79811502e-01
-6.45584241e-02 -2.01465800e-01 -7.87726194e-02 -1.17732771e-02
1.22077608e+00 -3.83076072e-02 -3.94392192e-01 -2.98972577e-01
-1.34606302e+00 8.61939490e-01 1.27062738e+00 5.67208111e-01
-4.43159848e-01 5.38453817e-01 6.55428231e-01 -7.40210861e-02
6.10464513e-01 4.67480481e-01 -3.86472046e-01 1.51147872e-01
-9.85629678e-01 -1.09777510e-01 9.94514376e-02 7.46541142e-01
2.96610773e-01 1.19301312e-01 -7.30514705e-01 7.48860657e-01
8.34322870e-02 2.03365371e-01 7.22972631e-01 -5.19391179e-01
3.82991225e-01 6.29377007e-01 -5.73950469e-01 -6.15231037e-01
-5.78611135e-01 -9.54171658e-01 -1.08097696e+00 -1.37380943e-01
2.24485919e-01 1.19123660e-01 -1.34578228e+00 1.67749882e+00
7.19860569e-02 2.63138443e-01 -2.86021352e-01 9.07330275e-01
1.06392550e+00 1.83693066e-01 3.70750874e-01 6.79328889e-02
1.26308966e+00 -1.05759823e+00 -6.28294408e-01 -2.08190367e-01
7.25795209e-01 -4.55686003e-01 1.16808701e+00 5.67244105e-02
-1.30161655e+00 -3.62274349e-01 -9.57805634e-01 -4.42645878e-01
-3.92643541e-01 -1.24354661e-01 1.00539351e+00 6.92007065e-01
-1.48395324e+00 6.64562047e-01 -9.84472394e-01 -1.52978048e-01
1.05821514e+00 4.72491026e-01 -9.39562768e-02 1.04306340e-01
-1.15454519e+00 9.36523736e-01 2.48175398e-01 3.10543954e-01
-7.74623215e-01 -1.09253311e+00 -9.32905376e-01 1.22868091e-01
2.39925571e-02 -8.27041566e-01 1.33786118e+00 -9.32449877e-01
-1.25256538e+00 8.13173056e-01 -2.62387879e-02 -8.94205272e-01
4.94865894e-01 -1.04860589e-01 -3.68574597e-02 1.62972897e-01
-2.84611024e-02 1.04266095e+00 5.56743979e-01 -9.60856378e-01
-1.02358848e-01 -3.35219741e-01 1.01115368e-01 -1.34317044e-04
-2.80544341e-01 -3.38371217e-01 -5.30357122e-01 -8.99890482e-01
2.11380422e-01 -4.60816920e-01 -5.25009751e-01 9.81909633e-02
-4.95102495e-01 -2.68407345e-01 7.39650667e-01 -5.53890467e-01
1.08858168e+00 -1.89296591e+00 -5.24463318e-02 1.50141895e-01
6.63017452e-01 1.20803364e-01 -1.13407172e-01 -2.76397407e-01
-2.68772304e-01 1.79579705e-01 -6.47920847e-01 -3.31954390e-01
-3.33414257e-01 2.72700310e-01 -1.03818767e-01 5.13555408e-01
4.58925158e-01 1.51425624e+00 -8.98376763e-01 -6.98053956e-01
3.44288886e-01 7.44879603e-01 -7.52022803e-01 -1.16864547e-01
-1.33201718e-01 4.56778198e-01 -4.59852844e-01 6.34582698e-01
8.00873518e-01 -4.20374244e-01 5.38184121e-02 -5.11333704e-01
1.00785196e-01 4.16258603e-01 -2.75832355e-01 2.07583117e+00
-6.37793005e-01 4.14559752e-01 -4.70951460e-02 -1.26082003e+00
6.04384661e-01 3.69300872e-01 5.84058106e-01 -1.11570585e+00
4.63186145e-01 2.00258285e-01 1.64187420e-02 -2.71497279e-01
2.92038530e-01 -3.44589263e-01 7.97541365e-02 1.33176908e-01
4.09979165e-01 2.99305338e-02 -1.48540303e-01 3.69200379e-01
1.21531272e+00 1.89452782e-01 1.31856486e-01 -5.46824932e-01
3.95681530e-01 -3.17474216e-01 3.90115082e-01 6.86038315e-01
-2.62667686e-01 9.43305254e-01 5.39080679e-01 -4.70822871e-01
-8.03732455e-01 -1.34928048e+00 -3.54073286e-01 8.22987199e-01
-1.35475948e-01 -2.27723867e-01 -8.03628862e-01 -9.76824582e-01
-1.71195462e-01 5.19405305e-01 -1.12119710e+00 -7.37841353e-02
-6.97189510e-01 -9.04183209e-01 7.43202865e-01 9.78275478e-01
5.90278566e-01 -1.01547742e+00 -8.73423576e-01 3.73303711e-01
-2.42720008e-01 -1.03568995e+00 -6.39902115e-01 4.82351720e-01
-1.36143458e+00 -9.02660012e-01 -1.19823420e+00 -5.59091568e-01
6.81285381e-01 -1.04076013e-01 1.46241558e+00 1.98239744e-01
-3.34810019e-01 2.10165754e-01 -2.64253825e-01 -2.05015302e-01
3.14386524e-02 4.56654638e-01 -6.24046564e-01 -2.67333955e-01
3.73250246e-02 -7.02434063e-01 -8.62756073e-01 1.31086418e-02
-1.04622161e+00 2.46235535e-01 7.52523184e-01 9.79578018e-01
5.52847207e-01 -5.75322568e-01 5.55645227e-01 -1.09981775e+00
5.36381006e-01 -4.54253435e-01 -1.34561419e-01 3.14574242e-01
-4.27166492e-01 2.53028810e-01 3.55163276e-01 -2.11544737e-01
-9.80812609e-01 -1.15662999e-01 -5.71583390e-01 -5.85058928e-01
9.90193561e-02 4.75686938e-01 1.47491813e-01 -6.99822605e-02
4.43349719e-01 2.97051847e-01 5.92396520e-02 -4.72584009e-01
3.92771989e-01 1.45855293e-01 5.89808941e-01 -4.26974267e-01
2.49417722e-01 4.55532461e-01 -2.16420025e-01 -4.49028075e-01
-6.77417099e-01 -1.67673901e-01 -6.93337023e-01 -1.07717603e-01
1.21386611e+00 -7.88846135e-01 -5.00736773e-01 2.71862835e-01
-1.19659173e+00 -5.65470278e-01 -5.41641474e-01 1.78518757e-01
-4.17856663e-01 1.23762734e-01 -7.69335687e-01 -5.04051208e-01
-7.57899761e-01 -1.46924436e+00 1.16693664e+00 4.33737189e-02
-2.06833392e-01 -1.10889351e+00 -2.72251993e-01 2.29355752e-01
9.05448794e-01 3.90660077e-01 1.05475950e+00 -5.81655383e-01
-5.08776546e-01 1.47435218e-01 -6.34177327e-01 1.83945358e-01
5.48087768e-02 -5.98614395e-01 -1.16465878e+00 -1.85738519e-01
-8.10718015e-02 -2.92952865e-01 1.16998780e+00 7.81374156e-01
1.64736676e+00 -1.02171220e-01 -3.30509454e-01 9.59601581e-01
1.45130062e+00 8.36540982e-02 8.57505143e-01 2.15997979e-01
7.85790741e-01 1.57557875e-01 -1.29687622e-01 1.03853136e-01
4.55356359e-01 4.49170977e-01 6.15139484e-01 -6.85987175e-01
-3.94842952e-01 4.01102416e-02 -7.20548183e-02 7.49386072e-01
-2.00533625e-02 -1.23952292e-02 -9.42207575e-01 6.16092741e-01
-1.51199949e+00 -6.56573653e-01 -1.84796616e-01 1.84602213e+00
7.51283109e-01 3.38645190e-01 -1.71155900e-01 -7.31015801e-02
1.94184616e-01 1.72277153e-01 -5.55934846e-01 -5.85122764e-01
-1.44567341e-01 7.72821903e-01 7.45555460e-01 2.28546426e-01
-1.22088325e+00 8.01969230e-01 6.35743523e+00 9.76356626e-01
-1.37798131e+00 6.34099901e-01 1.22012126e+00 -1.69307172e-01
-6.92986250e-01 -4.00407583e-01 -1.09985143e-01 1.90014869e-01
7.84781039e-01 4.27350551e-01 -2.32367571e-02 5.21867454e-01
-3.26158777e-02 -2.20379028e-02 -1.06682932e+00 7.44419992e-01
-9.61878225e-02 -1.63665831e+00 1.61720097e-01 -1.11861795e-03
5.56845069e-01 3.49433601e-01 2.11123675e-01 4.30065423e-01
1.78265452e-01 -1.46986079e+00 7.88934648e-01 3.90365452e-01
9.88258123e-01 -6.44843876e-01 9.25867438e-01 -1.19242385e-01
-1.33615005e+00 1.04421295e-01 -2.38429815e-01 2.30168253e-01
1.80269495e-01 7.20275342e-01 -8.16203058e-01 6.23480260e-01
8.46684635e-01 6.26542449e-01 -6.74124062e-01 1.01527739e+00
1.18486345e-01 7.06054628e-01 -3.38346571e-01 3.39021921e-01
6.62199438e-01 1.55806452e-01 9.80576053e-02 1.35420334e+00
2.01680660e-01 1.98028028e-01 -4.28492501e-02 1.07729065e+00
-2.59498775e-01 1.93435088e-01 -4.15752351e-01 1.61241129e-01
-8.85066763e-02 1.14886892e+00 -1.08390319e+00 -5.35070956e-01
-2.75358588e-01 1.08816683e+00 4.17474777e-01 2.50292838e-01
-1.12474275e+00 -1.29706234e-01 3.70152086e-01 3.16168189e-01
4.16536599e-01 -5.28246984e-02 -8.25035691e-01 -7.79742062e-01
-5.14712930e-02 -4.28926319e-01 3.43807757e-01 -7.03926265e-01
-9.96160328e-01 1.08215141e+00 -1.71272293e-01 -8.84823203e-01
2.27392167e-01 -4.29418772e-01 -7.55549371e-01 9.66167331e-01
-1.62993789e+00 -1.40682888e+00 -2.07964942e-01 6.01835489e-01
6.35831952e-01 2.16522038e-01 8.50227952e-01 5.11214137e-01
-3.87477160e-01 7.94946611e-01 -1.89077631e-01 2.75938690e-01
4.45929557e-01 -1.31605518e+00 5.08877397e-01 6.50516272e-01
-6.98433518e-02 6.70678973e-01 2.74238080e-01 -6.15614235e-01
-1.03550768e+00 -1.25226283e+00 5.46636164e-01 -3.65169764e-01
4.78982627e-01 -1.88943818e-01 -9.99558806e-01 6.65101469e-01
5.27948380e-01 5.42302251e-01 4.42322344e-01 -2.96659749e-02
-3.47397476e-01 -1.30503532e-03 -1.38241076e+00 2.92261690e-01
1.09169698e+00 -3.57657939e-01 -4.82864529e-01 1.06243879e-01
9.73779142e-01 -6.44051373e-01 -9.43816841e-01 5.55688739e-01
3.57900828e-01 -9.63895380e-01 1.17169535e+00 -1.84784874e-01
6.50215864e-01 1.41814381e-01 -1.85819939e-02 -1.09994280e+00
-3.59789848e-01 -1.66009694e-01 1.07234456e-01 1.09626675e+00
5.54312110e-01 -6.86661541e-01 7.63141334e-01 4.30477500e-01
-5.82417965e-01 -1.35121512e+00 -1.27195203e+00 -4.14062232e-01
5.07275879e-01 -5.07100821e-01 7.09524035e-01 9.07735586e-01
-2.24249721e-01 1.61431849e-01 -1.24244086e-01 -1.76275745e-01
6.05450809e-01 -1.79368109e-01 -5.77617660e-02 -9.01992321e-01
-1.39111415e-01 -7.08973110e-01 -2.51965344e-01 -1.08662486e+00
-1.96360424e-01 -1.41375983e+00 -1.90076567e-02 -1.77352345e+00
2.68742174e-01 -7.51004755e-01 -6.85823739e-01 6.89497471e-01
7.95170572e-03 6.46463573e-01 5.26129603e-02 -1.47396311e-01
-3.33343863e-01 6.52243197e-01 1.56996834e+00 -4.37955290e-01
-4.96313116e-03 -3.21567118e-01 -7.34768689e-01 4.76756096e-01
7.52749979e-01 -1.69270337e-01 -5.93540728e-01 -8.17089975e-01
-1.85817018e-01 2.55358648e-02 5.11887550e-01 -1.06537986e+00
2.28467863e-02 2.89084971e-01 7.17405438e-01 -5.58180749e-01
1.16812140e-01 -7.56248355e-01 -1.23334989e-01 4.15108860e-01
-3.74600589e-01 2.49579266e-01 4.69639748e-01 2.77318865e-01
-1.87731162e-01 1.82276830e-01 7.36545384e-01 -4.00436550e-01
-5.38858593e-01 5.99263251e-01 -3.51277351e-01 4.67717871e-02
7.63283432e-01 -2.59081513e-01 -3.61009762e-02 7.67829120e-02
-8.19865108e-01 2.45804682e-01 -1.43116117e-02 2.80373067e-01
7.74233460e-01 -1.17381763e+00 -5.08538723e-01 2.14798808e-01
-1.76746219e-01 2.92097896e-01 5.55722833e-01 1.41685057e+00
-5.58797300e-01 6.26177430e-01 -1.77610189e-01 -8.22794557e-01
-8.18159342e-01 3.84123594e-01 6.14476681e-01 -6.59326077e-01
-1.01380730e+00 1.27326322e+00 5.84967852e-01 -4.07308251e-01
1.89170972e-01 -9.32928205e-01 -5.09470776e-02 -8.86204392e-02
3.39376479e-01 -2.35071525e-01 5.10619760e-01 -4.19316322e-01
-5.41163504e-01 2.16525063e-01 -2.26918802e-01 1.43197313e-01
1.49789190e+00 2.30383780e-02 -8.18841159e-02 1.42737910e-01
1.30820394e+00 -5.27279139e-01 -1.10767019e+00 -3.33187848e-01
-8.01751763e-02 -1.05844714e-01 5.36530375e-01 -1.09603846e+00
-1.74881840e+00 1.15224361e+00 9.00915205e-01 -1.24344490e-01
1.29849207e+00 1.20446876e-01 1.03052521e+00 -1.26077414e-01
2.59194136e-01 -6.55975878e-01 -3.40298153e-02 4.50714380e-01
9.82797503e-01 -1.12484312e+00 -2.64787346e-01 -3.44067812e-01
-4.70389694e-01 9.68792260e-01 8.21692824e-01 -5.85054234e-02
6.73170388e-01 5.03124475e-01 -1.13831889e-02 -4.40220118e-01
-4.71593291e-01 -1.30306527e-01 4.81096089e-01 5.49423277e-01
9.72181439e-01 -5.64379729e-02 -1.05468206e-01 6.72372818e-01
-1.07189305e-01 1.27900720e-01 1.26909867e-01 8.19016516e-01
-1.58458427e-01 -9.04242992e-01 -2.17534512e-01 8.25734556e-01
-5.61356843e-01 -4.75319058e-01 6.88028485e-02 6.65203691e-01
2.76659638e-01 2.04665005e-01 3.03926229e-01 -1.56296402e-01
2.53879488e-01 1.68790221e-02 7.19978094e-01 -3.47749472e-01
-1.12912154e+00 1.92853004e-01 -2.06467956e-01 -6.80400789e-01
-2.88610637e-01 -3.42494279e-01 -1.44766152e+00 -4.34140861e-02
-1.52956501e-01 -1.14171751e-01 2.91855305e-01 1.05682111e+00
1.32013470e-01 1.34372270e+00 6.98228404e-02 -8.94934416e-01
-2.62457460e-01 -1.08659804e+00 -3.39135706e-01 2.54316032e-01
3.43713582e-01 -7.15818226e-01 1.45175532e-01 -2.88334042e-01] | [14.57932186126709, -2.5322625637054443] |
1aa562c8-6cd3-4306-adc2-b13b9ec39967 | non-stationary-dynamic-pricing-via-actor | 2208.09372 | null | https://arxiv.org/abs/2208.09372v3 | https://arxiv.org/pdf/2208.09372v3.pdf | Non-Stationary Dynamic Pricing Via Actor-Critic Information-Directed Pricing | This paper presents a novel non-stationary dynamic pricing algorithm design, where pricing agents face incomplete demand information and market environment shifts. The agents run price experiments to learn about each product's demand curve and the profit-maximizing price, while being aware of market environment shifts to avoid high opportunity costs from offering sub-optimal prices. The proposed ACIDP extends information-directed sampling (IDS) algorithms from statistical machine learning to include microeconomic choice theory, with a novel pricing strategy auditing procedure to escape sub-optimal pricing after market environment shift. The proposed ACIDP outperforms competing bandit algorithms including Upper Confidence Bound (UCB) and Thompson sampling (TS) in a series of market environment shifts. | ['Henghsiu Tsai', 'Chi-Hua Wang', 'Po-Yi Liu'] | 2022-08-19 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [ 2.67789923e-02 2.80568004e-01 -9.75531518e-01 -1.71792299e-01
-8.33639145e-01 -7.01457322e-01 2.35396683e-01 -1.86153099e-01
-6.73436940e-01 1.28315210e+00 -3.35784793e-01 -6.98271155e-01
-6.17251337e-01 -8.29948604e-01 -6.82854891e-01 -6.78746462e-01
-4.19166416e-01 1.13727796e+00 -1.19389892e-01 3.54136527e-01
5.73347032e-01 1.80356815e-01 -1.26329136e+00 -2.94311494e-01
1.14761114e+00 1.35203481e+00 2.96260059e-01 5.24103701e-01
-1.45708246e-03 2.25519285e-01 -1.53381079e-01 -5.47797859e-01
1.04032886e+00 -1.61163867e-01 -1.42969295e-01 -1.33585781e-01
-5.88652968e-01 -9.72467959e-01 3.84863555e-01 1.42830491e+00
1.40066966e-01 1.53711677e-01 6.66881502e-01 -1.43214071e+00
-4.98402953e-01 9.78008926e-01 -1.03868413e+00 3.46400678e-01
6.97942749e-02 3.67347062e-01 1.31736231e+00 -2.42689624e-02
3.08894604e-01 1.33449340e+00 2.52617747e-01 4.14517373e-01
-1.66814387e+00 -8.01130712e-01 3.77334803e-01 1.16108790e-01
-8.66660237e-01 1.66532144e-01 7.18326271e-01 -2.27436759e-02
8.20449173e-01 4.57003713e-01 1.13750422e+00 1.04159117e+00
4.64971662e-01 9.92300153e-01 1.72840035e+00 -5.32543600e-01
9.91354883e-01 2.34892368e-01 -3.25100064e-01 -2.65827656e-01
9.76848722e-01 7.73037612e-01 -2.75379010e-02 -7.35624254e-01
7.10683763e-01 1.10951200e-01 1.91483825e-01 -6.11911118e-01
-6.26456797e-01 1.11696303e+00 -2.01126233e-01 -5.34414828e-01
-1.25712287e+00 -7.63358846e-02 2.70646304e-01 8.34670544e-01
4.41036582e-01 2.16094613e-01 -9.80223536e-01 -4.34787720e-01
-6.52750969e-01 7.57225931e-01 1.21845603e+00 9.74026740e-01
4.40494150e-01 1.54352084e-01 -1.57035381e-01 3.72916162e-01
5.46933353e-01 9.53002930e-01 4.74538445e-01 -1.20217502e+00
4.13606852e-01 -1.99671939e-01 1.13033390e+00 -4.82977688e-01
-8.16912577e-02 -4.60268229e-01 -7.78584555e-02 2.54443407e-01
2.73155034e-01 -6.87747717e-01 -5.42274773e-01 1.47155273e+00
3.84547174e-01 8.52684304e-03 8.01385287e-03 6.35512471e-01
-7.23908901e-01 5.49162745e-01 -3.04810470e-03 -1.00095832e+00
1.25768149e+00 -5.55130422e-01 -7.54232824e-01 1.06359169e-01
2.64361620e-01 -4.96915460e-01 4.88031060e-01 1.01921892e+00
-1.19547117e+00 1.28321007e-01 -6.92403853e-01 1.05521393e+00
-1.30403386e-02 -7.47526109e-01 8.84758711e-01 8.59503210e-01
-5.38261771e-01 6.79292977e-01 -4.70369756e-01 1.10796973e-01
4.59396094e-01 3.56014818e-01 8.21510494e-01 1.64349765e-01
-1.13548207e+00 7.92745650e-01 5.31548440e-01 3.33175361e-02
-9.01274025e-01 -7.28563905e-01 -4.22887802e-01 2.34373406e-01
1.19893932e+00 -4.18551028e-01 1.97920752e+00 -1.16571414e+00
-1.71220124e+00 1.39137313e-01 3.35678130e-01 -1.36157286e+00
9.04503763e-01 -2.84243450e-02 -4.22000140e-01 5.36770709e-02
1.64711148e-01 2.53398001e-01 7.01826811e-01 -1.23228788e+00
-1.34338081e+00 -4.13019478e-01 -2.73023754e-01 5.40525496e-01
5.45426726e-01 -2.06932098e-01 5.93743503e-01 -5.03108144e-01
-1.35047242e-01 -8.82984757e-01 -6.96738660e-01 -8.72741163e-01
-3.55579615e-01 -6.95690885e-02 2.65817225e-01 -1.51958883e-01
9.32484388e-01 -1.58626103e+00 -5.21483660e-01 5.76652884e-01
-4.60910112e-01 -4.39905733e-01 1.51038289e-01 6.11811519e-01
1.86934128e-01 2.43294761e-01 1.51667476e-01 2.10815191e-01
7.64756739e-01 5.37176430e-01 -4.43080485e-01 3.65942419e-01
-4.35136288e-01 5.50578594e-01 -5.19729078e-01 2.55385190e-02
-2.58295843e-03 -8.31877589e-01 -6.67880118e-01 -2.25422055e-01
-7.96504080e-01 -8.88513774e-02 -8.32519412e-01 8.16741645e-01
1.10319352e+00 2.82323688e-01 3.43449742e-01 6.32831872e-01
-3.94313097e-01 6.28509521e-02 -1.49154675e+00 7.28995085e-01
-2.10038722e-01 -1.65904298e-01 4.43875313e-01 -1.24305201e+00
4.76619422e-01 1.36248007e-01 5.56482077e-01 -8.75462174e-01
1.95135966e-01 2.50668019e-01 2.38026232e-01 -1.87232271e-01
2.64184743e-01 -5.54296792e-01 -2.06401065e-01 1.02085197e+00
-5.34317255e-01 2.48032603e-02 -1.25938997e-01 -2.68502414e-01
5.01506746e-01 1.86085582e-01 5.90281188e-01 -7.48351157e-01
-1.21293753e-01 1.50868058e-01 1.01481402e+00 1.36278212e+00
-4.39339817e-01 -4.92417306e-01 7.11634338e-01 -3.84169966e-01
-1.00256538e+00 -1.01225209e+00 -2.18858466e-01 1.07108665e+00
1.12260059e-01 6.82734132e-01 -7.29688942e-01 -4.55628574e-01
9.23432350e-01 1.40611517e+00 -6.96848631e-01 3.69519562e-01
7.35739991e-02 -9.61309791e-01 -4.74975467e-01 1.15110138e-02
6.48831844e-01 -8.75592411e-01 -9.58161473e-01 5.35582602e-01
3.88000727e-01 -5.76710165e-01 -4.11257058e-01 2.48241857e-01
-6.98314667e-01 -9.86108005e-01 -4.48662639e-01 1.32938221e-01
3.91289622e-01 1.47181347e-01 6.83753550e-01 -6.81845784e-01
3.76120880e-02 5.43314934e-01 -3.07313129e-02 -1.36353743e+00
-1.90581575e-01 -6.56943768e-02 1.38323113e-01 2.12021604e-01
8.01984012e-01 -2.26074144e-01 -1.05903482e+00 3.28677624e-01
-5.95310748e-01 -2.37240031e-01 5.43968499e-01 9.18112814e-01
7.50074625e-01 6.86218083e-01 8.69853139e-01 -1.07531583e+00
1.03462815e+00 -6.68590784e-01 -1.53436577e+00 1.97290197e-01
-1.40928054e+00 9.65712126e-03 3.76777440e-01 -5.91996312e-01
-1.52730656e+00 -3.19208950e-01 6.47901058e-01 -1.39205739e-01
-1.05654404e-01 6.01503730e-01 -5.21538258e-02 4.73435640e-01
8.39815214e-02 1.87685564e-01 3.10535967e-01 -6.05830729e-01
1.44493476e-01 5.48363626e-01 1.14396572e-01 -8.66543055e-01
8.42556596e-01 3.50578457e-01 -8.22241977e-03 -9.19618234e-02
-6.03653848e-01 -6.98582530e-02 -1.96133941e-01 -1.08949821e-02
1.81472480e-01 -7.92751670e-01 -1.28467321e+00 1.17635064e-01
-4.48857844e-01 -2.65232384e-01 -5.28267980e-01 9.37361598e-01
-1.28803766e+00 1.35768810e-02 -1.94433987e-01 -1.75445497e+00
-1.19012952e-01 -7.94046283e-01 2.71080524e-01 3.52511942e-01
-1.19273804e-01 -8.24102104e-01 2.07532838e-01 3.69346023e-01
2.25996390e-01 3.80915582e-01 7.28535235e-01 -9.63258147e-01
-9.34392929e-01 -2.58233190e-01 3.47709596e-01 -9.52001959e-02
3.80956344e-02 -1.07411295e-01 -5.25849342e-01 -6.46980762e-01
3.55198801e-01 -5.32573126e-02 3.37213337e-01 1.19353271e+00
6.43291295e-01 -1.06933463e+00 -5.66264391e-01 2.28735954e-01
1.67597520e+00 1.24058735e+00 1.76324472e-01 1.10422039e+00
-5.33821642e-01 6.59167528e-01 1.26805604e+00 1.13096797e+00
2.31602073e-01 4.73082513e-01 6.67734861e-01 4.13852543e-01
1.32872045e+00 -4.28099602e-01 1.37916967e-01 -3.44628364e-01
1.33309171e-01 1.38970062e-01 -2.89227843e-01 5.01183748e-01
-1.96306968e+00 -1.32420456e+00 3.87553424e-01 2.61763382e+00
7.01932669e-01 6.51868224e-01 8.65019441e-01 -2.78080493e-01
8.38135183e-01 -2.76592761e-01 -1.34001684e+00 -1.19248366e+00
-1.57319475e-02 -3.82826366e-02 1.34019637e+00 4.69137192e-01
-6.20184839e-01 4.79893148e-01 6.71721363e+00 8.06243479e-01
-4.97431546e-01 -1.36446785e-02 9.79321659e-01 -3.25279504e-01
-9.10600483e-01 2.77931124e-01 -8.99634719e-01 6.94832325e-01
1.27066338e+00 -8.85198951e-01 8.94569755e-01 1.14030993e+00
6.90933824e-01 -6.64005697e-01 -8.80367458e-01 7.00587392e-01
-7.22593725e-01 -1.31110907e+00 -4.24943954e-01 7.42050827e-01
7.96666563e-01 -5.90587966e-02 2.83357054e-01 5.25926232e-01
1.02914953e+00 -5.90685070e-01 9.10636544e-01 4.13121432e-01
1.38278365e-01 -1.28336215e+00 7.96243668e-01 6.87746108e-01
-8.61206710e-01 -7.98213780e-01 -3.19674909e-01 -3.66746634e-01
1.12905815e-01 1.91767618e-01 -8.16482425e-01 5.12130618e-01
5.32853663e-01 -1.19765490e-01 3.68675977e-01 1.13618827e+00
2.83711642e-01 3.53877723e-01 -3.28661025e-01 -6.09016120e-01
7.26190150e-01 -1.07090199e+00 5.11520207e-01 5.43843448e-01
4.62940723e-01 8.96263644e-02 2.89199561e-01 1.00760150e+00
2.62263060e-01 3.71475555e-02 -6.41326666e-01 7.09872544e-02
8.56896162e-01 6.06538236e-01 -8.50012898e-01 -1.55762970e-01
-4.10232693e-01 5.10841966e-01 -6.27670825e-01 5.13049722e-01
-5.94070554e-01 -1.52267352e-01 7.54255950e-01 -5.99433072e-02
7.01068044e-01 2.46589154e-01 -4.45598036e-01 -5.43173671e-01
-1.47315711e-01 -6.72963798e-01 8.69829237e-01 -1.86886519e-01
-1.17299187e+00 -5.61991930e-01 3.98813009e-01 -1.17442620e+00
-5.56243062e-01 -4.37143028e-01 -6.35803163e-01 9.19129074e-01
-1.81679428e+00 -2.25010619e-01 1.02225006e+00 1.50781423e-01
7.17746913e-01 -1.65617973e-01 4.28959399e-01 -6.45700336e-01
-4.04003918e-01 1.95791572e-01 9.83332098e-01 -3.34507585e-01
-1.17307268e-01 -1.67313743e+00 -3.75263132e-02 3.38664472e-01
-4.64487016e-01 3.83858263e-01 9.25527155e-01 -9.73084569e-01
-1.72186828e+00 -6.09674335e-01 3.25076208e-02 2.10313395e-01
9.34537649e-01 -7.99899101e-02 -4.88296837e-01 4.82538432e-01
5.75346529e-01 -6.26036644e-01 8.08844090e-01 -5.36100604e-02
-3.66584072e-03 -5.67982793e-01 -1.80268550e+00 7.34008789e-01
5.80563605e-01 2.14791998e-01 -6.06799722e-01 1.08996548e-01
6.61414802e-01 -2.17736326e-02 -7.90239930e-01 1.90779969e-01
9.23627079e-01 -7.01192200e-01 5.42786658e-01 -6.29375458e-01
-6.65434301e-01 3.08321804e-01 -1.11731537e-01 -1.51113713e+00
-3.38804692e-01 -1.76508284e+00 1.34457454e-01 8.19976807e-01
6.31351471e-01 -1.39491820e+00 7.27353692e-01 1.02044213e+00
5.22493839e-01 -4.01785016e-01 -1.65892613e+00 -1.02480042e+00
3.04586709e-01 -2.71714866e-01 9.84784186e-01 6.83754921e-01
4.99099828e-02 -3.23150277e-01 -3.36245000e-01 1.36353001e-01
1.42884779e+00 7.48617828e-01 5.15719771e-01 -1.15135658e+00
-4.92578387e-01 -7.19856203e-01 4.98652846e-01 -7.56888330e-01
1.03452187e-02 -1.31509051e-01 -1.59576833e-01 -8.53516281e-01
2.22866133e-01 -2.15553567e-01 -8.46060872e-01 -1.29887134e-01
5.50352156e-01 -7.70007312e-01 1.40604243e-01 1.23997927e-01
-3.90919954e-01 3.62319797e-01 1.31029415e+00 -5.47838807e-02
-5.16066194e-01 5.74284256e-01 -1.02159059e+00 5.06468296e-01
9.84956264e-01 -4.80517179e-01 -7.09758997e-01 4.74401385e-01
2.96331525e-01 6.23961449e-01 -1.58109099e-01 -1.95920225e-02
-3.32477912e-02 -1.12166524e+00 2.41071954e-01 -1.04007947e+00
-2.86521316e-01 -1.15495598e+00 6.73069298e-01 1.12428975e+00
-4.16297734e-01 1.69652611e-01 5.63847041e-03 1.21240759e+00
3.33531052e-01 -5.64310253e-01 6.41486287e-01 -6.15797520e-01
-4.28865999e-01 2.84426123e-01 -6.76491082e-01 -1.16592109e-01
1.36379409e+00 -5.51125526e-01 -2.71613210e-01 -5.06501734e-01
-7.09663272e-01 7.77821660e-01 3.96474242e-01 1.81051940e-01
2.08137020e-01 -1.12040651e+00 -3.97912592e-01 2.39084801e-03
-3.19181144e-01 -1.69981450e-01 3.56428117e-01 5.49737394e-01
-1.82367593e-01 7.15182066e-01 -1.14081807e-01 -2.50887424e-01
-5.16062498e-01 8.04469883e-01 4.57537562e-01 -3.09344769e-01
-4.31900471e-01 3.31850797e-01 -1.20812736e-01 -1.23397699e-02
1.94454134e-01 -3.76860678e-01 2.62291074e-01 5.53633630e-01
4.45038438e-01 5.39206028e-01 -4.03747708e-01 1.39352113e-01
2.16256291e-01 1.09672949e-01 -3.00908893e-01 -4.93169159e-01
1.32775605e+00 -5.76093733e-01 2.83211768e-01 4.43486899e-01
3.76282752e-01 -3.70169431e-01 -1.79951930e+00 -3.47501665e-01
8.31493258e-01 -9.80343103e-01 1.68070734e-01 -1.04043686e+00
-6.10616744e-01 -1.11322150e-01 6.07402802e-01 1.08273137e+00
9.07206118e-01 -6.51531816e-01 3.58735204e-01 2.24385887e-01
8.92470300e-01 -1.98985517e+00 -7.18015075e-01 -3.32565129e-01
5.05187273e-01 -1.04388666e+00 -5.97090758e-02 2.34653533e-01
-8.39551151e-01 6.12681568e-01 1.49876177e-01 -2.65149891e-01
9.06855643e-01 1.54882267e-01 -3.19045007e-01 2.98622310e-01
-1.35467899e+00 -5.02860323e-02 -3.93676221e-01 6.92747772e-01
-5.40904999e-01 6.94919109e-01 -6.18165791e-01 8.00958633e-01
-1.30529836e-01 1.36511654e-01 7.65175700e-01 1.02910721e+00
-6.89674079e-01 -8.73608232e-01 -6.31181240e-01 7.32251346e-01
-7.44462371e-01 8.27904344e-02 2.06928179e-02 1.19695985e+00
-2.88072258e-01 5.58869183e-01 4.23675805e-01 4.75705832e-01
4.14965957e-01 1.91769153e-01 3.42028379e-01 -2.73462057e-01
-1.91820357e-02 1.02500629e+00 1.55666411e-01 -3.45109046e-01
-4.66260433e-01 -1.26986980e+00 -7.67931223e-01 -3.63478929e-01
-5.26976347e-01 6.09393597e-01 4.97373492e-01 7.87474275e-01
1.08134307e-01 -2.73817182e-01 1.44435501e+00 -5.00568807e-01
-1.63811111e+00 -6.95419312e-01 -1.41173923e+00 -7.97040947e-03
3.15564364e-01 -9.87723291e-01 -6.64675832e-01 -6.83078289e-01] | [4.49793815612793, 3.2678756713867188] |
deee8450-4c17-482a-9488-ce81547a75a1 | pixelrnn-in-pixel-recurrent-neural-networks | 2304.05440 | null | https://arxiv.org/abs/2304.05440v1 | https://arxiv.org/pdf/2304.05440v1.pdf | PixelRNN: In-pixel Recurrent Neural Networks for End-to-end-optimized Perception with Neural Sensors | Conventional image sensors digitize high-resolution images at fast frame rates, producing a large amount of data that needs to be transmitted off the sensor for further processing. This is challenging for perception systems operating on edge devices, because communication is power inefficient and induces latency. Fueled by innovations in stacked image sensor fabrication, emerging sensor-processors offer programmability and minimal processing capabilities directly on the sensor. We exploit these capabilities by developing an efficient recurrent neural network architecture, PixelRNN, that encodes spatio-temporal features on the sensor using purely binary operations. PixelRNN reduces the amount of data to be transmitted off the sensor by a factor of 64x compared to conventional systems while offering competitive accuracy for hand gesture recognition and lip reading tasks. We experimentally validate PixelRNN using a prototype implementation on the SCAMP-5 sensor-processor platform. | ['Gordon Wetzstein', 'Piotr Dudek', 'Laurie Bose', 'Haley M. So'] | 2023-04-11 | null | null | null | null | ['hand-gesture-recognition', 'hand-gesture-recognition-1', 'gesture-recognition'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.10614264e+00 -1.45481199e-01 -2.33428031e-01 -2.12850496e-01
-5.95161021e-01 -1.59309328e-01 2.45235965e-01 1.48704322e-02
-8.12239528e-01 9.83731449e-02 9.20988247e-02 -2.42700204e-01
3.35264862e-01 -6.23950064e-01 -5.68543017e-01 -5.57823360e-01
1.17138498e-01 -1.60622969e-01 5.86051106e-01 2.17397630e-01
1.92578465e-01 3.85792524e-01 -2.09108829e+00 5.65663517e-01
-1.44496918e-01 1.43387675e+00 4.77783054e-01 1.19677114e+00
1.63987219e-01 5.28512359e-01 -6.52386248e-01 1.75679833e-01
1.43633425e-01 -1.60703555e-01 -1.43272713e-01 -2.83114910e-01
4.22609180e-01 -7.79006302e-01 -7.20377564e-01 9.94388580e-01
6.84722900e-01 -3.50924432e-01 2.00230345e-01 -9.60562050e-01
-2.49617532e-01 5.60913742e-01 -4.71836597e-01 3.64496112e-01
3.10066551e-01 2.27684200e-01 5.68205357e-01 -6.02958679e-01
5.36364079e-01 9.74980712e-01 7.88928688e-01 7.07083881e-01
-9.90371585e-01 -4.99050975e-01 -1.54480174e-01 1.59438655e-01
-1.15187204e+00 -9.66272831e-01 5.64173043e-01 1.72977120e-01
1.81292260e+00 2.31961802e-01 8.81183803e-01 1.48255384e+00
5.08133471e-01 4.85756159e-01 9.69034612e-01 -2.03744873e-01
4.82914567e-01 -4.74750370e-01 4.67294008e-01 4.31474626e-01
3.39768052e-01 3.13526690e-02 -1.24677384e+00 3.88899595e-01
1.08855712e+00 2.71432191e-01 -3.18154812e-01 3.91038895e-01
-9.62298870e-01 9.83411223e-02 2.19825834e-01 1.28817558e-01
-6.65147483e-01 8.31020176e-01 5.22786856e-01 2.04349354e-01
-1.69396549e-02 -2.26075485e-01 -1.95443913e-01 -8.65616322e-01
-8.98532569e-01 -3.05179149e-01 6.07375801e-01 7.70130754e-01
2.93591749e-02 2.92365253e-01 1.35221824e-01 3.30434889e-01
3.71432006e-01 7.89848030e-01 8.19559574e-01 -8.56594265e-01
5.25449336e-01 5.65337956e-01 -1.33779183e-01 -6.38262510e-01
-5.77864647e-01 3.84464413e-01 -9.01868999e-01 5.83131671e-01
2.90143520e-01 -2.25852996e-01 -1.26490915e+00 1.31241500e+00
-1.62868902e-01 -1.70621499e-02 1.79983482e-01 1.09172785e+00
7.19180942e-01 8.49094570e-01 1.07847519e-01 -3.15775461e-02
1.65767038e+00 -3.80655468e-01 -6.77165627e-01 -3.18169862e-01
1.76680461e-01 -5.16506255e-01 1.05211639e+00 5.18733084e-01
-1.15358579e+00 -6.51253343e-01 -1.57503080e+00 -4.12284940e-01
-3.91344160e-01 2.78355069e-02 6.57554388e-01 7.87430525e-01
-1.39911866e+00 5.57995260e-01 -1.49438095e+00 -2.10982174e-01
5.11902928e-01 9.46388841e-01 -4.61219698e-02 2.74703711e-01
-5.55239141e-01 3.16134244e-01 2.44945481e-01 8.15038085e-02
-2.79254973e-01 -7.42388725e-01 -7.02752650e-01 3.20873410e-01
-1.16151519e-01 -4.57418412e-01 1.42065847e+00 -7.06934094e-01
-2.14180350e+00 4.72974122e-01 -3.94730777e-01 -8.86416435e-01
1.51684985e-01 -5.64722978e-02 -5.50585210e-01 3.66884410e-01
-5.01104057e-01 9.15102959e-01 1.34335530e+00 -4.38404143e-01
-7.02544093e-01 -5.27442634e-01 -4.26528066e-01 5.41852117e-02
-5.24130642e-01 1.24258362e-02 -4.59744871e-01 -2.77058512e-01
3.69661272e-01 -8.46627653e-01 -3.64797115e-02 7.20686987e-02
-9.33542568e-03 -1.95043936e-01 1.56540728e+00 -3.33831280e-01
9.74517822e-01 -2.22436666e+00 -4.42174435e-01 1.52034596e-01
2.11396009e-01 5.62640309e-01 -6.33036047e-02 -8.54750350e-03
1.79638237e-01 -2.44373888e-01 1.91831551e-02 -3.31437707e-01
-3.00636888e-01 2.00489566e-01 -4.31939572e-01 1.59056842e-01
1.74276046e-02 1.11586785e+00 -4.11349595e-01 -5.76265454e-02
4.01846200e-01 9.87236440e-01 -2.21726775e-01 -5.63786067e-02
-2.20961526e-01 -1.73773661e-01 -2.15256512e-01 9.69462276e-01
2.28187025e-01 -4.78804290e-01 3.25936794e-01 -3.72835547e-01
-3.57287377e-01 6.83110178e-01 -1.05350757e+00 1.90879989e+00
-3.76922756e-01 1.09238613e+00 3.52181077e-01 -5.44568658e-01
6.01718128e-01 5.04732728e-01 4.69308078e-01 -1.07277071e+00
6.34586692e-01 1.18112415e-01 -2.48573571e-01 -4.93229061e-01
7.24490702e-01 2.04986528e-01 -1.89139277e-01 5.67933619e-01
-2.91084051e-01 2.67905910e-02 -7.64647573e-02 -3.05065602e-01
1.46952796e+00 -1.19866364e-01 -2.10385561e-01 -1.67000089e-02
-2.64595270e-01 -1.00251935e-01 2.39971086e-01 8.54225755e-01
-2.78363436e-01 4.13077891e-01 5.39146513e-02 -4.49423522e-01
-1.17338157e+00 -1.07933438e+00 -2.20940158e-01 9.62137282e-01
1.04066737e-01 -5.57095408e-01 -8.99216533e-01 7.70274028e-02
-2.13752836e-01 -7.05062672e-02 3.82646769e-02 1.40156418e-01
-6.82831824e-01 -5.94318569e-01 7.71569788e-01 9.26897943e-01
7.05239117e-01 -9.86941576e-01 -2.16395664e+00 5.49118519e-01
4.94674593e-01 -1.37342179e+00 -1.81917861e-01 4.78974819e-01
-8.64791691e-01 -6.64513767e-01 -4.47990686e-01 -5.97567320e-01
4.52326626e-01 2.58858472e-01 6.67169034e-01 -3.66993994e-01
-7.13690937e-01 4.96468097e-01 -6.59555942e-02 -7.02011645e-01
6.08840175e-02 6.31459504e-02 3.33756655e-01 -3.35406512e-01
6.82179511e-01 -6.99855804e-01 -8.72294426e-01 -3.21278572e-01
-8.06860030e-01 2.11805090e-01 6.37471378e-01 6.04235590e-01
6.94326937e-01 -6.34813905e-02 3.41991335e-01 -4.27879184e-01
5.92346728e-01 2.85878360e-01 -9.91145313e-01 -1.99309304e-01
-5.89117587e-01 1.55731454e-01 5.58254242e-01 -6.07466519e-01
-9.34771538e-01 4.97683555e-01 -2.06903294e-01 -2.95193553e-01
9.23697799e-02 5.46477400e-02 9.34053361e-02 -1.07655816e-01
3.65659684e-01 -2.17115991e-02 3.21403593e-01 -1.19367428e-01
3.64154615e-02 1.08139420e+00 9.84445035e-01 3.05739373e-01
-9.49379206e-02 5.79605043e-01 4.80172634e-02 -1.25981379e+00
1.53664768e-01 -1.05863571e-01 -1.96358398e-01 -1.13665000e-01
8.83571267e-01 -1.04794073e+00 -1.37657130e+00 7.85327077e-01
-1.13885188e+00 -5.73653877e-01 -3.80652338e-01 4.36712056e-01
-2.19894856e-01 -9.63764042e-02 -9.65170026e-01 -9.10317838e-01
-8.53650153e-01 -1.10888469e+00 1.30305159e+00 6.61059618e-01
-5.40051043e-01 -3.72027695e-01 -4.20487702e-01 8.98735151e-02
6.22860849e-01 -1.66860744e-01 4.71102118e-01 1.62108973e-01
-9.69014704e-01 -2.59826362e-01 -4.49705780e-01 -1.00793704e-01
-6.93668723e-02 -1.20368041e-01 -1.25309336e+00 -2.16001809e-01
1.55622378e-01 -2.47687206e-01 1.00757670e+00 8.10302734e-01
1.35034037e+00 -3.98886278e-02 -6.07116759e-01 5.42423069e-01
1.43685746e+00 4.63455737e-01 6.42698526e-01 -3.79018217e-01
6.53061748e-01 -8.32265615e-02 -9.34528410e-02 3.84020865e-01
1.88573062e-01 6.67486072e-01 9.29664671e-02 5.53938225e-02
-4.66947109e-01 -2.54399031e-01 6.42770648e-01 8.68783057e-01
4.06564660e-02 -4.37805980e-01 -8.08790505e-01 3.16271842e-01
-1.64009035e+00 -8.09402764e-01 8.47932473e-02 1.97342944e+00
5.67141116e-01 3.58682156e-01 -1.63403809e-01 5.02848625e-01
4.64104146e-01 2.60538757e-01 -9.76335526e-01 -8.65869641e-01
2.57053040e-03 8.32033932e-01 9.90564466e-01 3.91302168e-01
-7.54971266e-01 9.49103534e-01 7.46477509e+00 3.79201442e-01
-1.86631310e+00 1.21092377e-02 4.35949922e-01 -7.19639003e-01
1.91366926e-01 -5.21043539e-01 -1.03504109e+00 5.54877400e-01
1.68490803e+00 4.44615394e-01 4.69169378e-01 7.56810129e-01
1.34734660e-01 -3.99421453e-01 -1.11211467e+00 1.67342758e+00
2.34221220e-02 -1.48558950e+00 -2.24316210e-01 1.43370852e-01
1.57770976e-01 3.03503364e-01 3.62670302e-01 -3.53839785e-01
-2.82328911e-02 -1.16918874e+00 3.79060090e-01 3.11525136e-01
1.36847484e+00 -6.29596055e-01 2.87777930e-01 -4.81549092e-02
-1.18442261e+00 -2.38131285e-01 -2.70098388e-01 -6.22570038e-01
4.18060660e-01 5.96269369e-01 -9.29448128e-01 -5.10221124e-01
1.16750228e+00 4.72100288e-01 -9.74632204e-02 7.64213055e-02
8.97326022e-02 4.68279779e-01 -8.50999832e-01 -3.68268400e-01
1.35541214e-02 1.12833939e-01 3.62179875e-01 1.09879911e+00
4.56456393e-01 4.92251605e-01 -4.12669480e-01 4.83463436e-01
-4.18480873e-01 -7.40147710e-01 -7.23275840e-01 -7.13245869e-02
6.65354133e-01 1.02251387e+00 -8.82240534e-01 -4.78858858e-01
-4.12717909e-01 1.35831964e+00 -6.82736039e-02 1.36678725e-01
-3.88960063e-01 -6.74735725e-01 8.31239402e-01 5.76738566e-02
3.71429414e-01 -6.30009055e-01 -7.23075986e-01 -6.89772308e-01
3.65206331e-01 -6.26577854e-01 1.76815148e-02 -8.03653896e-01
-4.88087386e-01 5.26734591e-01 -6.12307489e-01 -7.74200618e-01
-6.21408999e-01 -1.06200552e+00 -2.06303775e-01 4.93606389e-01
-1.11848426e+00 -7.80130446e-01 -4.33768451e-01 4.82423604e-01
5.13445675e-01 -5.03084660e-02 1.07898307e+00 1.07445620e-01
-6.56669855e-01 6.75459087e-01 -2.38077372e-01 1.65529415e-01
8.28539208e-02 -6.79595113e-01 9.06907320e-01 9.22270298e-01
2.30256662e-01 3.91603768e-01 4.18543607e-01 -5.21898568e-01
-2.22525311e+00 -7.27603018e-01 6.08286321e-01 -2.93810040e-01
4.26124930e-01 -5.63323677e-01 -6.97177529e-01 4.47556883e-01
2.97597110e-01 1.06409669e-01 4.66434628e-01 -2.33037144e-01
-4.25015062e-01 -2.41516069e-01 -1.06815338e+00 7.89807916e-01
1.32300568e+00 -9.75718677e-01 -2.15113804e-01 -1.92129463e-01
5.83509445e-01 -5.14282286e-01 -8.25972497e-01 3.42386156e-01
1.02671242e+00 -8.82621825e-01 7.77125359e-01 1.80034891e-01
4.21008207e-02 -1.46917745e-01 -1.09889105e-01 -7.37518668e-01
1.85533851e-01 -9.60733294e-01 -3.94531518e-01 7.88405716e-01
3.47692311e-01 -6.30940080e-01 1.29412031e+00 8.51248085e-01
3.88887571e-03 -7.06180274e-01 -1.47455609e+00 -3.85994911e-01
-7.83153653e-01 -7.72028565e-01 4.82863039e-01 2.56243683e-02
5.68769991e-01 4.77945805e-01 1.00650517e-02 3.82069238e-02
5.38535237e-01 -8.89792442e-02 2.43643612e-01 -8.12701523e-01
-1.63024768e-01 -5.27281582e-01 -9.38433588e-01 -1.46679282e+00
-2.32224435e-01 -2.80829042e-01 1.16506949e-01 -1.02531040e+00
-2.43918508e-01 -2.22643942e-01 -7.40395412e-02 5.97367287e-01
3.62813890e-01 7.45277941e-01 1.47697985e-01 1.38067409e-01
-3.80528480e-01 -9.93498415e-02 5.90537548e-01 -4.91397493e-02
-5.34792066e-01 -4.03800815e-01 -2.13948995e-01 7.19671369e-01
7.07192421e-01 -1.02145739e-01 -4.57776576e-01 -7.93622077e-01
2.07702145e-01 1.65154949e-01 4.57996488e-01 -1.56664336e+00
7.59393394e-01 5.24608672e-01 7.06137240e-01 -6.57346725e-01
9.25205231e-01 -9.18269396e-01 3.63210589e-02 6.07138991e-01
-2.30143979e-01 3.25881004e-01 4.19065297e-01 3.48858505e-01
-5.57805970e-02 2.29878470e-01 5.15023947e-01 2.16432557e-01
-1.00675201e+00 8.04499015e-02 -7.01925755e-01 -5.14345109e-01
7.50702143e-01 -6.06805265e-01 -4.24667984e-01 -2.28875905e-01
-3.93119454e-01 -3.73256177e-01 3.53216738e-01 4.15559947e-01
9.94711995e-01 -1.02497208e+00 -9.92672816e-02 8.99898350e-01
-2.83532083e-01 -1.20727997e-02 2.02837586e-01 4.06206250e-01
-6.05292141e-01 7.73713529e-01 -4.17363286e-01 -9.48699594e-01
-1.58792603e+00 1.84378669e-01 2.25225598e-01 3.10284257e-01
-1.22391033e+00 9.42501605e-01 -6.62241578e-01 7.02714920e-01
4.37051296e-01 -9.26507711e-01 2.22360954e-01 -7.24673644e-02
9.41020787e-01 3.38002264e-01 1.26953200e-01 -1.69338182e-01
-2.58714229e-01 7.54907429e-01 -2.01553002e-01 -6.19833529e-01
1.01880181e+00 8.65821466e-02 3.14380676e-01 4.41712141e-01
1.23383796e+00 -5.63430667e-01 -1.77012467e+00 9.80908796e-02
-3.12263761e-02 -9.44883153e-02 5.38542151e-01 -5.14356196e-01
-9.18176234e-01 8.48588049e-01 1.24870694e+00 9.01590288e-02
1.38764834e+00 -1.47170663e-01 1.31847131e+00 4.80029672e-01
5.13442874e-01 -1.58773530e+00 -3.05595621e-02 3.21786374e-01
3.78135324e-01 -9.92769659e-01 -2.68591493e-01 -2.71642387e-01
-2.15710357e-01 1.09994078e+00 2.17329100e-01 -1.65602833e-01
6.98922217e-01 1.31127119e+00 1.92655846e-01 -2.48855501e-01
-7.58593202e-01 1.46279288e-02 -2.06412405e-01 7.52426147e-01
2.96735615e-01 2.99715102e-01 1.26083896e-01 3.65435660e-01
-2.85826504e-01 4.38747793e-01 3.82259965e-01 1.20411348e+00
-3.41044813e-01 -9.46068048e-01 -1.75779283e-01 6.21121109e-01
-3.31034780e-01 -1.50175646e-01 2.00182456e-03 2.36703858e-01
-1.49557412e-01 1.06446195e+00 8.60730231e-01 -6.31364107e-01
1.20539024e-01 1.58195406e-01 4.96302485e-01 -2.04224721e-01
-5.52303612e-01 3.09645082e-03 -7.38333762e-02 -9.65162456e-01
-1.80949450e-01 -6.11719429e-01 -1.72863805e+00 -3.35472047e-01
2.43165851e-01 -6.83086693e-01 1.30252433e+00 7.68212318e-01
1.04416597e+00 7.94999540e-01 5.35953641e-02 -1.05742061e+00
-3.81196320e-01 -7.46039391e-01 -2.28095442e-01 -1.54821903e-01
5.19728243e-01 -3.72421406e-02 1.32902816e-01 1.89971879e-01] | [8.298930168151855, 2.4460160732269287] |
4d2e6912-aeb0-48aa-b4d0-da73a98fba37 | contcommrtd-a-distributed-content-based | 2301.12984 | null | https://arxiv.org/abs/2301.12984v1 | https://arxiv.org/pdf/2301.12984v1.pdf | ContCommRTD: A Distributed Content-based Misinformation-aware Community Detection System for Real-Time Disaster Reporting | Real-time social media data can provide useful information on evolving hazards. Alongside traditional methods of disaster detection, the integration of social media data can considerably enhance disaster management. In this paper, we investigate the problem of detecting geolocation-content communities on Twitter and propose a novel distributed system that provides in near real-time information on hazard-related events and their evolution. We show that content-based community analysis leads to better and faster dissemination of reports on hazards. Our distributed disaster reporting system analyzes the social relationship among worldwide geolocated tweets, and applies topic modeling to group tweets by topics. Considering for each tweet the following information: user, timestamp, geolocation, retweets, and replies, we create a publisher-subscriber distribution model for topics. We use content similarity and the proximity of nodes to create a new model for geolocation-content based communities. Users can subscribe to different topics in specific geographical areas or worldwide and receive real-time reports regarding these topics. As misinformation can lead to increase damage if propagated in hazards related tweets, we propose a new deep learning model to detect fake news. The misinformed tweets are then removed from display. We also show empirically the scalability capabilities of the proposed system. | ['Adrian Paschke', 'Ciprian-Octavian Truică', 'Elena-Simona Apostol'] | 2023-01-30 | null | null | null | null | ['community-detection'] | ['graphs'] | [-5.50235569e-01 1.30732954e-01 2.22898081e-01 -1.26861632e-01
-6.17707610e-01 -3.24336916e-01 8.71335924e-01 1.46751916e+00
-7.86950648e-01 8.04294527e-01 7.88725495e-01 -2.80920975e-02
8.36284906e-02 -1.70902765e+00 -5.22262990e-01 -4.76232380e-01
-9.31341112e-01 5.23152709e-01 7.28957474e-01 -8.11098933e-01
2.76867688e-01 4.28827345e-01 -1.13277388e+00 3.42717558e-01
4.07655925e-01 5.72733819e-01 1.42276540e-01 3.92427236e-01
-4.07460362e-01 7.47693479e-01 -1.28242886e+00 1.94151938e-01
-2.26343796e-01 3.65194753e-02 -7.20251441e-01 -3.66453499e-01
-2.16648027e-01 -5.40271282e-01 -5.29774606e-01 6.76689625e-01
7.33045399e-01 6.24883510e-02 5.64996243e-01 -1.50308502e+00
-3.23259503e-01 1.09841013e+00 -8.90816927e-01 7.27678895e-01
4.88081127e-01 -5.69113314e-01 2.95244753e-01 -6.28545165e-01
6.64968908e-01 1.21633947e+00 1.02203119e+00 -3.13414931e-01
-7.43834138e-01 -8.56158316e-01 4.05495197e-01 1.54759020e-01
-1.72912085e+00 -2.21148040e-02 6.19074762e-01 -7.59513319e-01
7.88515449e-01 2.68758059e-01 2.10436091e-01 8.91542912e-01
4.31149542e-01 1.62735283e-01 5.51398933e-01 -1.14194304e-01
4.57630932e-01 5.38135767e-02 -9.03366506e-03 1.89909428e-01
3.55905414e-01 -2.36812890e-01 -8.85776103e-01 -7.32513368e-01
-4.97657135e-02 4.98573899e-01 -3.29332769e-01 6.29778028e-01
-1.26600480e+00 1.22644281e+00 1.04871047e+00 5.62081277e-01
-7.59024620e-01 8.08911398e-03 6.17294490e-01 4.19189453e-01
1.42003036e+00 4.41037118e-02 5.12694754e-02 4.85880703e-01
-1.04672205e+00 3.53516310e-01 8.07057321e-01 5.36530793e-01
9.67008293e-01 -1.30341634e-01 -3.81725021e-02 5.22131562e-01
1.07507013e-01 6.72486961e-01 4.63259220e-01 -1.92368001e-01
5.00916183e-01 5.61227202e-01 4.63523418e-01 -2.21274185e+00
-1.09917367e+00 -2.84847111e-01 -1.17687500e+00 -2.87547916e-01
-2.66945250e-02 -6.44544303e-01 -3.04265171e-01 1.46656907e+00
5.60394883e-01 3.65059793e-01 -3.01175922e-01 7.55834520e-01
6.60372674e-01 1.01899111e+00 3.39875430e-01 -2.40902290e-01
1.00700569e+00 -1.03304327e-01 -8.38178754e-01 1.35020480e-01
6.91036463e-01 -5.11755228e-01 3.34130637e-02 6.87342584e-02
-5.73914289e-01 -3.60140875e-02 -3.88631821e-01 3.43740880e-01
-1.07186306e+00 -7.55771935e-01 1.10447600e-01 2.10454196e-01
-1.29435682e+00 5.21022677e-01 -5.81505358e-01 -5.89218497e-01
2.28939593e-01 8.07118043e-02 -1.56867459e-01 3.42330366e-01
-1.75783849e+00 6.61293745e-01 4.24558669e-01 -2.83031166e-01
-7.71816611e-01 -5.40395319e-01 -5.00195622e-01 -1.40149146e-01
3.81825976e-02 -2.90943891e-01 8.99096310e-01 -2.30144247e-01
-3.27800155e-01 5.14470696e-01 2.96619479e-02 -6.13611102e-01
4.40302402e-01 1.35987729e-01 -7.10159779e-01 2.61986494e-01
7.63074279e-01 2.82532424e-01 4.53113288e-01 -1.17830563e+00
-1.15870464e+00 -2.83865273e-01 -2.06230864e-01 -3.19528401e-01
-8.78793120e-01 3.99180800e-01 2.62000918e-01 -8.47672701e-01
1.38585046e-02 -5.37750483e-01 -1.97708771e-01 -2.95853585e-01
-6.80366814e-01 -2.74629116e-01 1.04791987e+00 -7.24361122e-01
1.59286773e+00 -1.96935463e+00 -4.40289706e-01 2.86853313e-01
6.01421773e-01 -2.80315459e-01 2.14285523e-01 1.38165307e+00
3.08786213e-01 3.42466742e-01 -2.09438264e-01 -5.26974976e-01
-3.07414889e-01 -1.20531671e-01 -7.92305350e-01 8.47612977e-01
-1.90440387e-01 2.05104321e-01 -1.21015465e+00 -2.79729992e-01
-2.78850585e-01 4.76337731e-01 -2.84124851e-01 -3.33436430e-02
1.29727349e-02 3.53553951e-01 -5.71292460e-01 9.71865356e-02
9.56125081e-01 -2.04678595e-01 -1.46437958e-01 2.98008680e-01
-5.61235547e-01 2.73994118e-01 -9.60957110e-01 7.05950320e-01
-2.64252871e-01 1.03872764e+00 -3.57488953e-02 -9.06609774e-01
1.22107613e+00 4.72257912e-01 7.52846241e-01 -5.24643362e-01
1.56859197e-02 2.67519034e-03 -8.84094238e-01 -3.09597850e-01
1.05493569e+00 -2.87057534e-02 -4.34303373e-01 1.11386216e+00
-5.12308657e-01 3.67670566e-01 -8.88650939e-02 5.43462753e-01
1.22409928e+00 -1.21836901e+00 1.12581901e-01 -1.66489989e-01
-1.00603215e-02 1.89466104e-01 3.95631082e-02 9.39139903e-01
1.09087722e-02 5.49858034e-01 3.36363822e-01 -6.03408575e-01
-8.27964902e-01 -8.16854298e-01 7.45499060e-02 1.40525222e+00
1.66818723e-01 -2.26190075e-01 -2.68646359e-01 -3.51643652e-01
3.01460624e-01 7.35698044e-01 -7.80635118e-01 9.98821035e-02
-4.91116226e-01 -1.35980821e+00 6.37858987e-01 4.41181473e-04
4.47593898e-01 -1.12446153e+00 -3.25422645e-01 6.45500481e-01
-8.43151689e-01 -7.29648411e-01 -1.51886865e-01 -1.62548140e-01
-4.19308662e-01 -7.75720477e-01 -8.50573778e-01 -6.43831909e-01
4.48540211e-01 8.12410891e-01 6.84946656e-01 1.16149858e-01
2.02282622e-01 3.29694718e-01 -5.51277876e-01 -7.06301153e-01
-3.84933054e-01 2.38611288e-02 2.86856174e-01 3.06166887e-01
1.21675834e-01 -6.78292334e-01 -7.03702867e-01 2.31189325e-01
-1.11797488e+00 -3.86355758e-01 -4.59178835e-01 1.17041416e-01
-1.57590300e-01 5.05030215e-01 1.19338119e+00 -7.88414478e-01
1.01042426e+00 -1.72125125e+00 -2.70978987e-01 -3.77898782e-01
-3.77633363e-01 -5.03916681e-01 2.69463718e-01 -3.70790690e-01
-8.42905700e-01 -4.56695884e-01 2.87647069e-01 5.00915408e-01
-2.44166628e-01 1.37435353e+00 7.05719352e-01 1.67743400e-01
1.06374168e+00 1.18248895e-01 -2.88080573e-01 -4.73309308e-01
1.93394795e-01 1.11888611e+00 4.76685792e-01 7.85132274e-02
8.69311273e-01 1.29892850e+00 -6.03914261e-01 -1.16250134e+00
-4.16169286e-01 -1.05197585e+00 -1.66327864e-01 -3.04779470e-01
5.08748651e-01 -1.19513667e+00 -5.36385357e-01 9.62795317e-01
-1.75238812e+00 -5.48136272e-02 8.17021206e-02 3.53563517e-01
1.71483129e-01 4.22030926e-01 -8.72804821e-01 -8.88987601e-01
-4.07935202e-01 -8.13478380e-02 7.77800918e-01 -7.11724386e-02
-3.14863533e-01 -1.39411855e+00 5.94479263e-01 -3.40742201e-01
1.02649677e+00 7.98053622e-01 3.61543864e-01 -1.10705781e+00
-2.11662203e-01 -6.13418818e-01 -5.02374530e-01 -4.83138710e-01
3.46801430e-01 -5.17610371e-01 -9.73403156e-01 -4.59949195e-01
-9.97429341e-02 2.62073755e-01 9.56021905e-01 1.16555043e-01
6.10895276e-01 -9.63506818e-01 -9.08177674e-01 1.31119505e-01
1.08962464e+00 -2.54010707e-01 4.06318694e-01 9.38854158e-01
5.80862880e-01 1.01696706e+00 4.61346984e-01 1.31540775e+00
9.35848355e-01 3.27487648e-01 6.89045966e-01 -4.07213062e-01
3.17596406e-01 -2.16918945e-01 3.15139890e-01 6.40270829e-01
1.65215477e-01 -1.06688190e+00 -1.34720993e+00 1.10817444e+00
-1.62443841e+00 -1.38761532e+00 -6.09120905e-01 2.07077670e+00
7.80980945e-01 -1.19628705e-01 4.79851604e-01 4.66521643e-02
1.16375780e+00 4.99980658e-01 7.43377283e-02 2.48389140e-01
-3.75892729e-01 -2.93342650e-01 9.16297019e-01 8.61280620e-01
-1.18443966e+00 6.09134316e-01 5.60865021e+00 4.05013829e-01
-1.36177433e+00 5.68556368e-01 4.06306893e-01 1.16895407e-01
-5.60568869e-01 -3.58714253e-01 -7.19735980e-01 6.20531142e-01
9.47799087e-01 -6.31917059e-01 -3.00130486e-01 5.12996316e-01
9.07230318e-01 -3.86193544e-01 -2.03487381e-01 3.09706807e-01
1.54242679e-01 -1.65139365e+00 8.47123191e-03 2.71761924e-01
7.31798232e-01 6.43956244e-01 -2.05652252e-01 -2.89921165e-01
3.47301155e-01 -3.44989806e-01 7.55804241e-01 6.95310295e-01
3.04937959e-01 -8.41503978e-01 9.55613256e-01 7.31419802e-01
-1.14848304e+00 -1.26369223e-01 -4.24952000e-01 -1.87196612e-01
6.49006605e-01 1.04792666e+00 -1.26477623e+00 3.54780287e-01
8.06556106e-01 8.87674510e-01 -4.05463159e-01 1.45648146e+00
1.29472658e-01 7.23053336e-01 -7.37196922e-01 -4.27537970e-02
2.44243130e-01 3.75443906e-01 7.94284761e-01 1.69057167e+00
6.01896107e-01 2.24330034e-02 3.82591277e-01 5.08742392e-01
5.94212972e-02 1.65177658e-02 -9.12079573e-01 3.81561309e-01
8.19116950e-01 7.15631902e-01 -9.04840231e-01 -4.76868749e-01
-3.46613303e-03 6.98148489e-01 1.27657935e-01 4.41229790e-01
-4.95389968e-01 -5.85750282e-01 2.74672657e-01 8.32630217e-01
1.45002073e-02 -4.49936658e-01 1.12961620e-01 -6.71192467e-01
-3.86685878e-01 -2.74503648e-01 6.47682130e-01 -4.70260948e-01
-1.50234818e+00 7.09267676e-01 3.33650500e-01 -1.21382737e+00
-7.76270628e-02 3.53222340e-01 -1.17732143e+00 6.44221425e-01
-1.59468734e+00 -8.81955802e-01 -3.58226091e-01 8.05405319e-01
6.98039904e-02 -8.28587338e-02 6.74238801e-01 5.35823166e-01
-1.42989099e-01 2.16855317e-01 2.62748897e-01 4.26042169e-01
7.94690847e-01 -8.51711631e-01 8.47099781e-01 5.63323557e-01
-3.27309161e-01 3.26893777e-01 8.07711661e-01 -1.25661516e+00
-3.81497681e-01 -1.61215389e+00 1.77750409e+00 -1.75374433e-01
1.12148869e+00 -1.85235828e-01 -1.13507533e+00 4.59460586e-01
1.22611418e-01 -4.44800258e-01 7.48598754e-01 1.12401530e-01
-2.10824117e-01 7.11266920e-02 -1.09235370e+00 2.12482646e-01
5.16291738e-01 -6.57462656e-01 -3.75758618e-01 1.14172649e+00
9.56369698e-01 1.55827671e-01 -5.65857053e-01 -3.31648648e-01
-1.83825776e-01 -8.50855350e-01 8.03485990e-01 -3.79960060e-01
1.03785079e-02 -2.81062841e-01 1.83759540e-01 -1.69681215e+00
-5.16525805e-01 -7.07367539e-01 4.10348058e-01 1.26662242e+00
3.08454603e-01 -1.06577837e+00 4.38005060e-01 2.11367324e-01
-1.28265647e-02 3.73620600e-01 -1.10098493e+00 -5.07067919e-01
8.55902582e-02 -4.90853578e-01 6.87456250e-01 1.32286322e+00
5.31885803e-01 -9.89211649e-02 -5.57730675e-01 7.61796832e-01
5.22102118e-01 -1.23142347e-01 5.75071514e-01 -1.50062835e+00
4.65296417e-01 -1.74148232e-01 -3.56588721e-01 -3.05155426e-01
-3.72033060e-01 -6.04105055e-01 -1.02396950e-01 -1.51597917e+00
-1.51533529e-01 -6.85120821e-01 -2.29085431e-01 5.01079500e-01
2.15587333e-01 2.76572376e-01 -2.22254500e-01 5.36934078e-01
-6.24722421e-01 3.93015027e-01 3.22614193e-01 -3.64697158e-01
-4.53888535e-01 1.86533615e-01 -4.06206906e-01 6.81347013e-01
9.48728859e-01 -1.24833751e+00 2.33361095e-01 -3.87475133e-01
9.11224484e-01 5.46686389e-02 6.40669703e-01 -1.02997530e+00
7.59109616e-01 -8.78930744e-03 -1.13801315e-01 -9.76703882e-01
-1.98526345e-02 -5.04058063e-01 1.15870439e-01 5.07882357e-01
-2.73963213e-01 2.18388006e-01 1.59304485e-01 1.05656183e+00
-3.78191084e-01 3.58893126e-01 1.62366465e-01 6.44631609e-02
-4.10261869e-01 3.99750054e-01 -1.38366008e+00 -4.08346355e-02
7.85052299e-01 1.49765313e-01 -8.04268241e-01 -1.03698337e+00
-8.14491034e-01 3.26250225e-01 1.88103631e-01 4.16447043e-01
5.91116726e-01 -1.13777852e+00 -1.19839633e+00 -2.50702798e-01
2.32467398e-01 -2.82774448e-01 4.03509080e-01 7.22785592e-01
-4.45678174e-01 1.33672431e-01 -1.21796727e-02 -2.07761437e-01
-7.82584608e-01 5.30329823e-01 -2.03863867e-02 1.44777521e-01
-4.72695678e-01 7.64091372e-01 -8.18144605e-02 -2.13343918e-01
2.27666304e-01 -1.74144775e-01 -6.48361385e-01 9.77540374e-01
1.01635206e+00 7.75531650e-01 5.83074205e-02 -8.47217202e-01
-3.86861354e-01 5.67914657e-02 4.15039584e-02 -2.19201446e-01
1.54434943e+00 -5.71308076e-01 -3.10470104e-01 5.87139904e-01
1.19806302e+00 1.64290830e-01 -5.60352266e-01 -4.93960440e-01
3.19100022e-01 -2.66446918e-01 3.08874667e-01 -4.23843294e-01
-9.17102098e-01 5.25890350e-01 4.47627336e-01 1.05825651e+00
7.95766771e-01 2.01618239e-01 1.13401031e+00 3.51026356e-01
4.41329896e-01 -9.95443165e-01 1.24005051e-02 6.10370755e-01
1.06881559e+00 -1.15695810e+00 -1.03586791e-02 -1.51165023e-01
-3.17439049e-01 9.41501558e-01 -1.08133420e-01 -2.27510817e-02
1.31894898e+00 1.35217682e-01 1.69823050e-01 -5.84778190e-01
-3.78401071e-01 -1.96074739e-01 -1.90071985e-01 8.30541492e-01
1.41374379e-01 1.69695884e-01 -2.27926284e-01 3.05861622e-01
-2.24967539e-01 -3.08116406e-01 9.77318287e-01 9.32969511e-01
-1.30368972e+00 -4.70972955e-01 -8.67566228e-01 1.74007758e-01
-4.24489051e-01 5.70461014e-03 -3.34824502e-01 4.41503286e-01
5.63681545e-03 1.38895190e+00 4.56336021e-01 -5.75502813e-01
4.91681844e-02 -4.48796511e-01 -7.95127451e-01 -5.97643316e-01
-8.21364582e-01 -2.52856374e-01 1.18217498e-01 -2.50497073e-01
-8.35719034e-02 -5.89686394e-01 -1.29513609e+00 -8.34782064e-01
-3.06331009e-01 5.02627432e-01 7.22329259e-01 8.47556531e-01
7.18799233e-01 6.39440715e-02 1.06878662e+00 -9.46019351e-01
5.01286536e-02 -1.13198912e+00 -7.16043293e-01 4.25821751e-01
9.12554920e-01 -5.00677168e-01 -6.81091487e-01 -1.46987155e-01] | [8.454626083374023, 9.755800247192383] |
e92e5068-3d00-468a-b854-4cacb000eb7d | learning-rate-free-bayesian-inference-in | 2305.14943 | null | https://arxiv.org/abs/2305.14943v1 | https://arxiv.org/pdf/2305.14943v1.pdf | Learning Rate Free Bayesian Inference in Constrained Domains | We introduce a suite of new particle-based algorithms for sampling on constrained domains which are entirely learning rate free. Our approach leverages coin betting ideas from convex optimisation, and the viewpoint of constrained sampling as a mirrored optimisation problem on the space of probability measures. Based on this viewpoint, we also introduce a unifying framework for several existing constrained sampling algorithms, including mirrored Langevin dynamics and mirrored Stein variational gradient descent. We demonstrate the performance of our algorithms on a range of numerical examples, including sampling from targets on the simplex, sampling with fairness constraints, and constrained sampling problems in post-selection inference. Our results indicate that our algorithms achieve competitive performance with existing constrained sampling methods, without the need to tune any hyperparameters. | ['Christopher Nemeth', 'Lester Mackey', 'Louis Sharrock'] | 2023-05-24 | null | null | null | null | ['bayesian-inference'] | ['methodology'] | [ 2.32895434e-01 -2.85414439e-02 -5.71149409e-01 -2.82290518e-01
-1.03633988e+00 -4.54144746e-01 7.61396229e-01 -4.94303793e-01
-6.75393283e-01 1.47723353e+00 1.40216798e-01 -3.17600578e-01
-3.28120917e-01 -7.01387346e-01 -7.46818721e-01 -6.60580516e-01
-7.95597658e-02 1.07404375e+00 -1.19774707e-01 1.71823457e-01
5.48249066e-01 1.91200554e-01 -1.23840892e+00 -3.66401404e-01
8.69756162e-01 7.41008043e-01 -9.35674757e-02 7.29590058e-01
1.81681320e-01 3.08713675e-01 -2.77510583e-01 -6.43627167e-01
2.04672292e-01 -3.91349822e-01 -6.42124474e-01 -3.34749073e-02
2.25461662e-01 -1.77660018e-01 8.84681195e-02 1.22773075e+00
5.09633660e-01 5.95992863e-01 8.16464663e-01 -1.02325773e+00
-4.95545387e-01 5.48886776e-01 -2.14276448e-01 -4.27500233e-02
3.15238714e-01 1.62661001e-01 1.32120240e+00 -9.69528079e-01
6.61515713e-01 1.55559266e+00 8.03083777e-01 6.01191998e-01
-1.61619592e+00 -4.17477280e-01 1.41759828e-01 -2.50737906e-01
-1.15835249e+00 -4.53569233e-01 3.78615290e-01 -4.13077742e-01
4.62852985e-01 6.16240382e-01 8.48767519e-01 1.14417136e+00
-2.00131401e-01 8.86151731e-01 1.37960994e+00 -5.46230674e-01
9.59711730e-01 3.21532607e-01 -2.99155712e-02 8.11549485e-01
4.16901439e-01 6.60696805e-01 -6.80529356e-01 -9.47984695e-01
8.76687407e-01 -1.95394367e-01 -1.91345900e-01 -8.39154482e-01
-1.20686793e+00 1.44293272e+00 -7.83575922e-02 -5.67213655e-01
-1.48626909e-01 6.25829399e-01 1.31526008e-01 -1.06354862e-01
8.58217359e-01 3.59272093e-01 -2.89144218e-01 -1.61087766e-01
-1.44208181e+00 8.22458804e-01 1.38230228e+00 1.01110923e+00
5.86777925e-01 -2.09461197e-01 -7.22051561e-01 4.00998175e-01
8.63567889e-01 8.53268623e-01 -1.94525942e-01 -1.31883764e+00
2.91391164e-01 -1.36775896e-01 9.91229475e-01 -2.89999574e-01
-5.58895059e-02 -2.79983938e-01 -5.82059443e-01 2.83023834e-01
6.18687868e-01 -3.95440936e-01 -6.59641504e-01 1.73921871e+00
6.64616883e-01 2.39706039e-01 -5.21420419e-01 9.31923032e-01
-2.04398632e-01 3.35218549e-01 -2.25566760e-01 -4.89711761e-01
1.12228572e+00 -8.09299111e-01 -5.18667281e-01 2.06396535e-01
6.68330118e-02 -5.71104109e-01 1.16247475e+00 3.78375560e-01
-1.45566547e+00 1.39459535e-01 -9.23970342e-01 1.28732279e-01
3.10232192e-02 -1.20951414e-01 8.47204685e-01 1.13077438e+00
-9.12152886e-01 1.11285460e+00 -9.31927681e-01 -9.09660757e-02
7.98339665e-01 1.89261526e-01 5.80723882e-01 1.86895698e-01
-8.85498226e-01 8.63593936e-01 -8.62924531e-02 -1.47569269e-01
-9.33394372e-01 -9.42442000e-01 -3.47986490e-01 -6.48978949e-02
4.51675415e-01 -1.10933781e+00 1.47630823e+00 -1.02350846e-01
-2.17719269e+00 5.96574187e-01 -4.74863946e-01 -6.02851868e-01
1.24845850e+00 -8.82469788e-02 3.99470627e-01 -2.61964619e-01
7.72238448e-02 1.63214207e-01 9.37984824e-01 -8.97482932e-01
-4.23935026e-01 -1.56207234e-01 -6.00985289e-02 3.42457265e-01
9.98627767e-02 -6.03032671e-02 -1.35436684e-01 -3.38071525e-01
-2.20294848e-01 -1.17914009e+00 -6.14218354e-01 2.76395708e-01
-6.70468509e-01 -3.52698058e-01 7.74789751e-02 1.34637803e-01
9.99921083e-01 -1.40439594e+00 3.91470224e-01 5.21101177e-01
1.58088431e-02 -1.45967945e-01 2.67912954e-01 2.76786029e-01
6.02756202e-01 4.02082205e-01 -4.65075493e-01 -7.65478730e-01
8.62279654e-01 4.50778492e-02 -4.43539679e-01 7.63281047e-01
-2.50132293e-01 1.14727831e+00 -1.14872134e+00 -4.91906226e-01
7.23170638e-02 6.93783164e-02 -8.63383293e-01 -1.95836611e-02
-6.42438948e-01 3.08191776e-01 -5.26806056e-01 6.59359694e-01
6.12259090e-01 -4.07498568e-01 2.71761209e-01 2.06103668e-01
-2.38835718e-02 4.32973891e-01 -1.52972126e+00 1.58065474e+00
-2.40394026e-01 1.66843116e-01 3.69478822e-01 -7.98249125e-01
4.69610304e-01 -5.90620283e-03 8.69094953e-02 1.60137549e-01
-2.86766496e-02 4.18461323e-01 -5.60914218e-01 -1.64089240e-02
6.09636009e-01 -7.20599532e-01 -2.80281585e-02 9.11606431e-01
-1.89304307e-01 -6.41291261e-01 2.31403351e-01 -6.62044287e-02
6.13201380e-01 4.52292234e-01 1.84354588e-01 -7.33181477e-01
1.82615459e-01 6.26278371e-02 2.88590878e-01 1.42730188e+00
-1.34505644e-01 4.70716447e-01 2.78400123e-01 -2.29626998e-01
-1.21111083e+00 -1.32857955e+00 -5.50751150e-01 1.22925460e+00
-2.77297758e-02 -1.17811941e-01 -5.67891181e-01 -4.05880719e-01
5.21841764e-01 6.72626019e-01 -6.76356435e-01 4.18501586e-01
-2.78129667e-01 -1.21432102e+00 4.58222389e-01 1.39837995e-01
1.56115577e-01 -6.98325813e-01 -4.82150257e-01 2.34577686e-01
1.06943414e-01 -4.86424357e-01 -7.90670514e-01 1.41013190e-01
-1.04188824e+00 -1.09415126e+00 -8.71680677e-01 -3.20990264e-01
4.36438799e-01 -1.77246779e-01 1.30381715e+00 -2.81828314e-01
-2.60077477e-01 4.80164289e-01 3.97359312e-01 -6.34300649e-01
-2.31464356e-01 3.37601602e-02 2.90942639e-01 -1.91922322e-01
2.32925743e-01 -4.30051535e-01 -6.09764636e-01 2.57361352e-01
-4.01633263e-01 -3.18289936e-01 6.03097901e-02 1.20051348e+00
7.13185012e-01 -4.05412585e-01 1.54560506e-01 -9.52761233e-01
1.09834814e+00 -5.45808315e-01 -1.07990921e+00 1.40207469e-01
-7.45060623e-01 3.20482582e-01 1.28963217e-01 -6.01672232e-01
-9.70926404e-01 -3.41256410e-01 2.79400557e-01 -3.36894959e-01
5.26406765e-01 3.55881244e-01 2.23235369e-01 -2.76464075e-01
6.27938032e-01 1.66235730e-01 3.20671856e-01 -3.27575177e-01
6.89863682e-01 5.20533025e-01 1.35444373e-01 -1.25205863e+00
7.62910962e-01 1.06661952e+00 2.73196757e-01 -4.78688568e-01
-9.83785987e-01 -8.21144581e-02 -3.11346591e-01 1.22449115e-01
5.26168644e-01 -6.69045329e-01 -1.20283735e+00 1.51044384e-01
-9.01127636e-01 -6.12042725e-01 -9.06955659e-01 6.46511972e-01
-1.05591404e+00 4.02790934e-01 -4.43016976e-01 -1.38581276e+00
-4.11004186e-01 -1.04747438e+00 1.01188195e+00 1.11663312e-01
-4.48731691e-01 -1.33083892e+00 5.79640806e-01 2.40053743e-01
5.98057985e-01 6.31960705e-02 5.83005548e-01 -3.34483594e-01
-1.03450990e+00 -1.54006600e-01 2.26168465e-02 -3.19480263e-02
-3.78251791e-01 -1.21480368e-01 -6.53849781e-01 -5.07091701e-01
1.68515161e-01 -5.89375794e-01 9.54583943e-01 7.46636391e-01
7.88699985e-01 -2.92706311e-01 -5.45811594e-01 9.20748770e-01
1.35748005e+00 -3.86789173e-01 2.97971517e-01 1.58588529e-01
2.78082937e-01 1.95776984e-01 4.21402067e-01 9.21800733e-01
2.43146256e-01 5.48443079e-01 1.98731259e-01 4.92195725e-01
3.41789365e-01 -3.96531910e-01 9.44595560e-02 3.23336750e-01
-4.82569188e-02 -5.04246801e-02 -2.56696105e-01 4.95862007e-01
-2.10366225e+00 -1.26444280e+00 1.52453721e-01 2.55594802e+00
1.26324403e+00 1.03929237e-01 3.40634257e-01 -4.24279958e-01
8.12757075e-01 8.04182589e-02 -1.14516962e+00 -3.26977253e-01
2.00650707e-01 7.47187912e-01 7.66805410e-01 9.91280615e-01
-1.00067210e+00 7.05358803e-01 8.29499149e+00 1.13988078e+00
-2.74382174e-01 3.92816097e-01 3.89981896e-01 -6.62349522e-01
-6.72705710e-01 1.87606305e-01 -1.14182723e+00 7.01324403e-01
6.70901716e-01 -4.84883130e-01 9.22067165e-01 7.29151487e-01
1.92635104e-01 -2.46055454e-01 -1.28595066e+00 7.07838476e-01
-3.46982837e-01 -1.82870603e+00 -2.66538888e-01 3.60856891e-01
1.00965643e+00 2.02927068e-01 2.17371628e-01 2.37524673e-01
1.15741873e+00 -1.24516904e+00 7.12129533e-01 6.69631124e-01
8.10545027e-01 -5.87515771e-01 2.46806189e-01 3.01882565e-01
-6.04460537e-01 3.67187798e-01 -5.84574878e-01 -4.03051078e-01
5.13612509e-01 9.35708642e-01 -5.34641623e-01 1.03935907e-02
1.80994168e-01 3.64996940e-01 5.33380032e-01 1.23393393e+00
-2.78380096e-01 6.85128510e-01 -9.23128068e-01 -8.28441799e-01
9.80803370e-02 -7.61031747e-01 9.63750005e-01 1.12671041e+00
1.46563843e-01 -1.45500004e-01 1.31977692e-01 1.59068096e+00
-1.36025622e-01 -1.78271383e-01 -3.25175256e-01 2.95856565e-01
8.66619885e-01 1.03404868e+00 -3.96472812e-01 -3.91301095e-01
4.37134728e-02 5.67250907e-01 3.50653470e-01 4.26278681e-01
-9.54027295e-01 -1.21106148e-01 9.63283300e-01 -2.37156823e-01
5.94707727e-01 -2.47985125e-01 -3.56680065e-01 -1.37178838e+00
-4.49932069e-02 -6.13734126e-01 1.46316737e-01 -2.96407580e-01
-1.56132269e+00 -2.70473540e-01 1.56886369e-01 -7.54237056e-01
-2.29785129e-01 -4.98371899e-01 -7.21692026e-01 1.14202297e+00
-1.51017749e+00 -4.40511972e-01 3.39407355e-01 5.30638933e-01
1.56598717e-01 3.58729735e-02 7.65455127e-01 -2.20206276e-01
-2.86467522e-01 5.18288553e-01 6.92313254e-01 -4.41895694e-01
1.98035151e-01 -1.62411070e+00 5.41517437e-01 4.51511830e-01
1.02094389e-01 7.97515750e-01 7.10586667e-01 -5.75345874e-01
-1.60090005e+00 -8.25711071e-01 5.07389307e-01 -7.99167991e-01
8.30516458e-01 -4.96736050e-01 -2.75370866e-01 5.41547835e-01
-1.12806700e-01 -1.54204257e-02 6.63147867e-01 4.65525150e-01
4.93854936e-03 4.03089255e-01 -1.47261369e+00 7.11722136e-01
1.23892391e+00 -4.75456327e-01 -3.94606531e-01 9.73235786e-01
4.20394719e-01 -7.46583462e-01 -8.93320739e-01 -2.36459114e-02
7.15080500e-01 -7.48074293e-01 1.24277604e+00 -7.74951458e-01
1.14230141e-01 -6.99425191e-02 -2.65323460e-01 -1.29386806e+00
-4.03317809e-01 -1.29054403e+00 -3.76412094e-01 6.44206762e-01
3.23030353e-01 -7.89519370e-01 1.25110006e+00 7.87250042e-01
3.61725032e-01 -9.56644773e-01 -1.25377810e+00 -8.55065942e-01
4.50562596e-01 -4.07049447e-01 6.20440900e-01 5.51129460e-01
7.27253631e-02 1.46949843e-01 -5.21439135e-01 -1.87369362e-01
1.49320161e+00 3.97053123e-01 6.88689590e-01 -1.33842337e+00
-6.79686129e-01 -4.74924684e-01 3.22955489e-01 -1.30160069e+00
2.35679060e-01 -1.11591589e+00 2.22352520e-02 -1.12759006e+00
4.53938305e-01 -6.94445312e-01 8.53823647e-02 -9.95981693e-02
-2.51397073e-01 2.21537784e-01 1.90224394e-01 4.29318696e-01
-6.75654888e-01 8.33253920e-01 1.17700851e+00 -1.30950715e-02
-1.49226591e-01 4.59692448e-01 -5.40333271e-01 7.28220105e-01
5.82608521e-01 -5.28055429e-01 -2.86896914e-01 -1.26964366e-02
4.28979278e-01 -2.93780962e-04 3.78408968e-01 -6.36670828e-01
2.69815117e-01 -6.58521831e-01 1.80216417e-01 -5.23188472e-01
4.89631653e-01 -1.05702169e-01 6.57927394e-02 6.19148076e-01
-4.48271066e-01 -5.01425922e-01 -1.95158377e-01 8.74922514e-01
5.45254648e-01 -5.72141111e-01 9.20195937e-01 -4.85479385e-01
3.82094495e-02 3.20623815e-01 -2.98454940e-01 8.44733059e-01
5.98711967e-01 2.45170314e-02 -1.06861293e-01 -2.91235447e-01
-7.22899497e-01 4.17587489e-01 4.39752907e-01 -3.84779513e-01
3.85667056e-01 -1.31236422e+00 -8.28369319e-01 -4.91201878e-02
-4.65553015e-01 6.30532131e-02 -4.42883402e-01 7.44371235e-01
-2.45038599e-01 4.55012977e-01 4.59226787e-01 -5.96821606e-01
-7.13005900e-01 2.55693346e-01 4.46610510e-01 -4.57394660e-01
-2.30699897e-01 1.24021947e+00 -3.72428298e-01 -9.14216280e-01
3.18602949e-01 -2.94690758e-01 6.31793797e-01 1.64533015e-02
3.52964133e-01 8.27798426e-01 -4.92333263e-01 2.91925311e-01
-3.13019007e-01 4.32910323e-01 1.27699435e-01 -6.94487631e-01
1.05552781e+00 -2.23314822e-01 -7.34525025e-02 5.07754087e-01
9.14906025e-01 1.53900981e-01 -1.51676440e+00 -3.36231023e-01
-1.01504669e-01 -5.01135230e-01 -4.17115428e-02 -7.08462000e-01
-4.04057175e-01 3.76880467e-01 1.65211305e-01 8.64088237e-02
9.20831263e-02 -2.55923271e-01 5.70558846e-01 5.07946551e-01
7.42421985e-01 -1.20268786e+00 -1.28188625e-01 4.14248645e-01
4.46368665e-01 -1.14202917e+00 6.86134040e-01 -1.50562907e-02
-3.52318734e-01 8.20709944e-01 -1.71829924e-01 -6.21137977e-01
8.65676701e-01 1.43211886e-01 -5.12332082e-01 -9.08918679e-02
-8.01531672e-01 -1.83198437e-01 1.29648939e-01 5.38510680e-01
1.17564887e-01 2.73263544e-01 -5.43264449e-01 2.29611620e-01
-3.30974519e-01 4.09373045e-01 5.10883212e-01 8.10213149e-01
-4.98114020e-01 -1.12800765e+00 -3.66805822e-01 6.52157724e-01
-4.04516101e-01 -2.54714251e-01 7.02593997e-02 6.78830564e-01
-4.00058478e-01 7.48246074e-01 -8.73240978e-02 3.28518987e-01
-6.02018610e-02 2.07216129e-01 7.84825742e-01 -6.10222280e-01
-2.03004450e-01 -1.61115959e-01 3.17142531e-02 -5.40881753e-01
-5.67536831e-01 -1.06763113e+00 -6.41722918e-01 -6.37524843e-01
-7.14935482e-01 5.03005564e-01 6.86554492e-01 9.34731066e-01
2.44098738e-01 -7.86942169e-02 5.18882453e-01 -1.26715124e+00
-1.71159422e+00 -7.54970610e-01 -9.41238582e-01 9.33758020e-02
2.69364178e-01 -8.32211971e-01 -7.29336917e-01 -5.66615641e-01] | [6.793126106262207, 4.052186965942383] |
ac1afe61-d738-4d71-bfea-74f3d162d44b | combining-recurrent-convolutional-and | 2110.13985 | null | https://arxiv.org/abs/2110.13985v1 | https://arxiv.org/pdf/2110.13985v1.pdf | Combining Recurrent, Convolutional, and Continuous-time Models with Linear State-Space Layers | Recurrent neural networks (RNNs), temporal convolutions, and neural differential equations (NDEs) are popular families of deep learning models for time-series data, each with unique strengths and tradeoffs in modeling power and computational efficiency. We introduce a simple sequence model inspired by control systems that generalizes these approaches while addressing their shortcomings. The Linear State-Space Layer (LSSL) maps a sequence $u \mapsto y$ by simply simulating a linear continuous-time state-space representation $\dot{x} = Ax + Bu, y = Cx + Du$. Theoretically, we show that LSSL models are closely related to the three aforementioned families of models and inherit their strengths. For example, they generalize convolutions to continuous-time, explain common RNN heuristics, and share features of NDEs such as time-scale adaptation. We then incorporate and generalize recent theory on continuous-time memorization to introduce a trainable subset of structured matrices $A$ that endow LSSLs with long-range memory. Empirically, stacking LSSL layers into a simple deep neural network obtains state-of-the-art results across time series benchmarks for long dependencies in sequential image classification, real-world healthcare regression tasks, and speech. On a difficult speech classification task with length-16000 sequences, LSSL outperforms prior approaches by 24 accuracy points, and even outperforms baselines that use hand-crafted features on 100x shorter sequences. | ['Christopher Ré', 'Atri Rudra', 'Tri Dao', 'Khaled Saab', 'Karan Goel', 'Isys Johnson', 'Albert Gu'] | 2021-10-26 | combining-recurrent-convolutional-and-1 | http://proceedings.neurips.cc/paper/2021/hash/05546b0e38ab9175cd905eebcc6ebb76-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/05546b0e38ab9175cd905eebcc6ebb76-Paper.pdf | neurips-2021-12 | ['sequential-image-classification'] | ['computer-vision'] | [ 2.60353208e-01 -3.89547199e-01 -2.61033595e-01 -2.56067425e-01
-3.28732044e-01 -3.80906552e-01 6.82340026e-01 -1.51662037e-01
-6.22079492e-01 6.41236544e-01 3.52501608e-02 -8.84536922e-01
-4.05537814e-01 -5.67086577e-01 -1.01356578e+00 -7.52119005e-01
-6.30942941e-01 2.21527386e-02 -3.36285025e-01 -4.76461738e-01
1.19823754e-01 5.25295138e-01 -1.15298653e+00 1.12300731e-01
5.61427832e-01 1.32737577e+00 2.18004286e-01 7.62426734e-01
-5.59101701e-02 1.19777298e+00 -3.26211959e-01 1.23068862e-01
4.36707318e-01 -3.56231689e-01 -8.63934219e-01 -2.56138682e-01
8.17436725e-02 -2.47530147e-01 -9.19508696e-01 5.92417836e-01
4.15374666e-01 4.32675391e-01 4.93349016e-01 -1.05586481e+00
-9.91403341e-01 5.49157917e-01 -3.49773556e-01 6.21554494e-01
-1.06793553e-01 5.69209516e-01 8.45941842e-01 -4.83830035e-01
5.29484451e-01 1.03324592e+00 1.17181611e+00 7.29994059e-01
-1.44378805e+00 -7.79297054e-01 5.05384386e-01 1.10219702e-01
-1.04207945e+00 -3.15679580e-01 5.22737384e-01 -3.51920605e-01
1.82199705e+00 1.95236832e-01 6.69243634e-01 1.44673634e+00
7.63582170e-01 9.00200784e-01 1.01193106e+00 -5.60940616e-02
2.94923365e-01 -5.56073070e-01 3.08997869e-01 8.42536926e-01
-3.20049584e-01 3.69241476e-01 -4.70548213e-01 8.23504233e-04
9.97668445e-01 5.32148778e-01 -2.68869340e-01 -1.56800747e-01
-1.35206592e+00 7.90992022e-01 5.18372178e-01 4.43644106e-01
-5.14941514e-01 7.12755501e-01 6.73103094e-01 7.62806714e-01
3.42282504e-01 7.05665708e-01 -7.57326186e-01 -2.79925853e-01
-8.19348335e-01 1.62935063e-01 6.60327852e-01 8.31595182e-01
4.62857306e-01 7.12826669e-01 -8.22656527e-02 7.13161588e-01
-2.90485173e-01 5.84778368e-01 1.09565246e+00 -1.07375717e+00
2.19997957e-01 1.90168962e-01 -1.63260028e-01 -7.14785635e-01
-7.60495245e-01 -8.18630636e-01 -1.45213127e+00 -2.33201310e-01
1.01426467e-02 -3.06975573e-01 -1.23639929e+00 2.19635391e+00
-4.16743129e-01 6.10776067e-01 1.11307904e-01 4.97260153e-01
2.50270635e-01 1.09295881e+00 -7.83764720e-02 -4.61013496e-01
8.08520317e-01 -9.12780941e-01 -6.43524706e-01 -2.83691853e-01
7.41549194e-01 -2.39310443e-01 9.08503950e-01 2.53812462e-01
-1.26749289e+00 -6.13298118e-01 -9.41793680e-01 -9.28197522e-03
-4.71521914e-01 -7.51779079e-02 8.59434605e-01 1.74330220e-01
-1.59478199e+00 1.24836099e+00 -1.35091639e+00 -2.97889024e-01
1.97201878e-01 5.83600223e-01 7.24549545e-03 2.95266867e-01
-1.40968871e+00 9.73021626e-01 -1.54553233e-02 3.00481915e-01
-1.02306414e+00 -1.16996098e+00 -8.98576975e-01 2.35311668e-02
1.88968360e-01 -7.85701394e-01 1.65524554e+00 -7.63823986e-01
-1.65948176e+00 4.25436467e-01 -1.25151008e-01 -1.40199697e+00
2.28123784e-01 -1.24189861e-01 -6.05923951e-01 -1.86844379e-01
-1.41277075e-01 6.80888593e-01 9.27888930e-01 -3.27869952e-01
-3.42301428e-01 8.66755843e-02 -1.69501901e-01 -2.07347870e-01
-3.20234388e-01 -2.67491043e-01 -4.77690957e-02 -9.21843767e-01
6.97527900e-02 -1.13156259e+00 -6.81746066e-01 -1.86383620e-01
-2.35022515e-01 -2.13923603e-01 7.51087844e-01 -5.99154413e-01
1.59938490e+00 -2.05058265e+00 2.16783062e-01 7.73473382e-02
3.27340007e-01 3.66498798e-01 -4.77265537e-01 6.58685684e-01
-6.44888341e-01 1.27965197e-01 -3.09266180e-01 -3.60945940e-01
1.23788364e-01 3.71528178e-01 -6.63279831e-01 3.78110319e-01
3.29802334e-01 1.31452024e+00 -7.59170353e-01 1.42021090e-01
1.73493117e-01 4.33637708e-01 -3.25229138e-01 -1.65063530e-01
-3.26687366e-01 1.25597760e-01 -2.11711839e-01 1.35616034e-01
1.39755100e-01 -7.22807765e-01 9.12225097e-02 -7.33605474e-02
-2.17930943e-01 5.88267982e-01 -6.11427128e-01 1.83730590e+00
-6.13754272e-01 8.86843503e-01 -3.24841976e-01 -1.48809302e+00
5.80800951e-01 3.24347079e-01 8.73652458e-01 -1.06377590e+00
1.22706629e-01 2.55474091e-01 1.67567015e-01 -4.17168170e-01
2.34625116e-01 -2.13074371e-01 -1.18660234e-01 7.66385317e-01
1.67295292e-01 -4.71162684e-02 1.18015006e-01 1.00525148e-01
1.51165795e+00 -9.57630128e-02 1.67259991e-01 -1.45163476e-01
1.49418414e-01 -3.35988402e-01 3.37937295e-01 1.02739561e+00
-1.10597618e-01 2.57815093e-01 3.49684983e-01 -7.20900595e-01
-1.30737102e+00 -1.06549299e+00 6.19906597e-02 1.12313771e+00
-5.01425505e-01 -3.00663173e-01 -3.78280908e-01 -1.01568684e-01
-1.40945213e-02 5.99571228e-01 -9.77805138e-01 -4.08401340e-01
-8.95622551e-01 -7.95637965e-01 6.74231112e-01 8.53604853e-01
4.46239263e-01 -1.24011886e+00 -8.46786559e-01 6.62025034e-01
-3.89756039e-02 -7.93793321e-01 -7.65626132e-01 7.30225205e-01
-9.73697066e-01 -5.87779880e-01 -8.17499995e-01 -7.90477216e-01
2.29115844e-01 6.62427098e-02 1.15218532e+00 -1.47027746e-01
-3.22587907e-01 4.72138375e-01 1.19556829e-01 -2.43419006e-01
-2.39987284e-01 2.65757173e-01 5.28180063e-01 -1.32699490e-01
2.57233474e-02 -9.22852933e-01 -6.94080949e-01 -9.34831798e-03
-9.68293369e-01 9.98981819e-02 6.19797707e-01 1.07586944e+00
5.18683493e-01 -2.13204250e-01 6.24241233e-01 -4.33115661e-01
9.27678943e-01 -4.28593367e-01 -4.37553793e-01 1.42026037e-01
-6.88408732e-01 5.26854157e-01 7.64097154e-01 -8.73492718e-01
-4.50852424e-01 -3.88579294e-02 -2.05141649e-01 -9.66958642e-01
3.70163202e-01 7.80575752e-01 7.38526940e-01 9.42106470e-02
5.40467739e-01 9.05821919e-01 5.09595394e-01 -2.55808651e-01
3.50174457e-01 8.68246928e-02 5.71657121e-01 -5.27897120e-01
3.62068504e-01 3.69022787e-01 -1.92717854e-02 -7.53321111e-01
-6.93514049e-01 -9.45268050e-02 -2.31656224e-01 2.08984882e-01
6.31294131e-01 -8.13288391e-01 -1.04557323e+00 6.99843407e-01
-1.01110923e+00 -9.30030942e-01 -6.33097112e-01 4.23603207e-01
-9.63174284e-01 -1.46550713e-02 -1.28564179e+00 -7.46519625e-01
-5.34592867e-01 -8.54969144e-01 6.90685332e-01 -8.07403326e-02
-2.14034334e-01 -1.04785073e+00 3.68822590e-02 -7.30565786e-01
8.49804401e-01 1.88797563e-01 1.11475742e+00 -4.67372268e-01
-3.23469162e-01 2.26118956e-02 8.44691619e-02 6.60851657e-01
-2.31497437e-02 -2.62078911e-01 -5.06636679e-01 -4.16792810e-01
3.81466001e-01 -2.20641002e-01 1.01990843e+00 9.12414372e-01
1.45783937e+00 -5.30653417e-01 -4.10741806e-01 7.03587234e-01
1.22261143e+00 7.43793011e-01 4.60110575e-01 1.12077385e-01
3.43442172e-01 5.96653819e-02 -1.11486405e-01 3.55193198e-01
2.94392288e-01 2.25577950e-01 2.26641148e-01 -1.15459614e-01
3.34788620e-01 -7.82428086e-02 6.37419701e-01 1.17957103e+00
-5.63470758e-02 -1.42639831e-01 -9.18644607e-01 4.78323221e-01
-1.92636549e+00 -1.24187219e+00 3.46347421e-01 1.80355489e+00
8.11990380e-01 3.59781533e-01 1.15359187e-01 2.60188133e-02
3.51811796e-01 4.51235384e-01 -1.17663777e+00 -4.83932227e-01
-2.09705248e-01 6.70927286e-01 6.27421141e-01 2.54694134e-01
-1.09346616e+00 6.92247391e-01 6.97524834e+00 6.85675383e-01
-1.72312367e+00 9.28885266e-02 8.86871219e-01 -5.78630209e-01
-4.52833287e-02 -6.70593500e-01 -5.55124402e-01 3.45400989e-01
1.70146966e+00 -2.52099484e-01 8.25614512e-01 4.74693835e-01
3.02789807e-01 5.32330155e-01 -1.42744231e+00 1.23110425e+00
-1.51930511e-01 -1.82193375e+00 -1.15252763e-01 2.66268682e-02
7.98206210e-01 6.15805805e-01 7.06272125e-01 7.84352779e-01
6.12688184e-01 -1.37938571e+00 6.29103839e-01 6.18251562e-01
7.67606258e-01 -4.47409719e-01 2.34844446e-01 3.44430894e-01
-1.18015480e+00 -5.32970309e-01 -5.37062809e-02 -3.93417865e-01
2.25573391e-01 1.77713722e-01 -4.10167933e-01 1.45968810e-01
7.65998065e-01 1.11859405e+00 -1.51359409e-01 5.76944530e-01
4.64096636e-01 6.59805954e-01 -5.16291678e-01 -3.39156806e-01
9.10588562e-01 -1.02458354e-02 2.06653744e-01 1.15288377e+00
3.96984607e-01 3.29108566e-01 -8.26736093e-02 8.26517105e-01
3.79001983e-02 -3.73440951e-01 -6.74748838e-01 -5.36550224e-01
2.17676356e-01 6.18471742e-01 -3.15139085e-01 -4.57522690e-01
-3.62242311e-01 8.92333031e-01 4.00245011e-01 6.74405634e-01
-1.00973701e+00 -3.17263901e-01 1.04262555e+00 -5.80344386e-02
5.74745297e-01 -6.35913908e-01 -6.20880015e-02 -1.09980822e+00
-7.55336583e-02 -9.99786377e-01 3.10857981e-01 -7.81688511e-01
-1.38117218e+00 8.95849526e-01 -3.22235703e-01 -1.36275578e+00
-8.32346380e-01 -7.05751061e-01 -2.84100890e-01 8.87268484e-01
-1.25367641e+00 -5.77692866e-01 3.88511628e-01 7.83426583e-01
7.27254808e-01 6.91786921e-03 8.46201658e-01 1.52930826e-01
-7.07096636e-01 5.16211808e-01 4.46753711e-01 2.03606039e-01
1.13461174e-01 -9.69453752e-01 1.00017893e+00 7.25774646e-01
3.73978689e-02 1.11579108e+00 6.48663521e-01 -4.09433037e-01
-1.65068662e+00 -1.20594573e+00 6.41970932e-01 -2.34828472e-01
1.09831071e+00 -3.69986415e-01 -9.54703450e-01 1.10103095e+00
3.13635468e-01 4.39622924e-02 2.79558301e-01 4.13161330e-02
-5.10251045e-01 -8.74367952e-02 -6.64842427e-01 7.23824203e-01
1.33362079e+00 -8.63937318e-01 -4.00684744e-01 3.47057432e-01
9.61737394e-01 -6.39115691e-01 -8.00845921e-01 3.60815555e-01
7.12194145e-01 -6.42763317e-01 1.15520561e+00 -1.16703153e+00
6.37012005e-01 1.88966647e-01 -8.72971639e-02 -1.57515430e+00
-3.96546751e-01 -1.05140483e+00 -4.05142814e-01 2.95630783e-01
4.81059432e-01 -8.34368169e-01 5.31745791e-01 4.59339559e-01
-4.34747100e-01 -1.22448719e+00 -9.10826504e-01 -1.11974847e+00
4.62831289e-01 -6.71242714e-01 5.58420181e-01 9.79179382e-01
2.94072777e-02 2.74532825e-01 -5.35018504e-01 -6.82353675e-02
8.87116417e-03 8.59642029e-02 1.25660777e-01 -8.82980525e-01
-3.54847521e-01 -9.36430871e-01 -9.74567384e-02 -1.51159072e+00
2.24614754e-01 -7.67092943e-01 2.14996785e-02 -1.21281123e+00
-2.13174328e-01 -3.94015193e-01 -8.09660673e-01 5.97478449e-01
1.58350393e-01 -2.01053068e-01 2.11150914e-01 6.45506084e-02
-4.78517324e-01 4.65974361e-01 9.52789843e-01 -2.26149067e-01
-2.37353176e-01 -7.33348951e-02 -5.35231054e-01 4.18825179e-01
6.90186203e-01 -1.52638584e-01 -5.81926227e-01 -6.91967368e-01
8.57809186e-02 5.44892013e-01 3.31542134e-01 -1.05763555e+00
5.17908931e-01 -2.14936495e-01 2.07930878e-01 -6.18461430e-01
5.75442135e-01 -5.78933775e-01 2.30488852e-01 9.71273839e-01
-7.26235271e-01 7.62312055e-01 3.79095227e-01 6.45146489e-01
3.20088081e-02 4.30085510e-01 5.61416805e-01 -2.50982136e-01
-7.18852878e-01 7.13540673e-01 -6.74942195e-01 -1.77292228e-02
6.32027090e-01 1.06817484e-02 -2.59363532e-01 -6.89157128e-01
-9.37554419e-01 2.26331025e-01 -2.87904054e-01 6.25823557e-01
4.40754116e-01 -1.28938615e+00 -4.15924489e-01 3.44411433e-01
-3.34282964e-01 -3.42893034e-01 4.90936995e-01 1.01561892e+00
-2.01361910e-01 8.56296778e-01 -2.07986802e-01 -7.85795510e-01
-5.18900156e-01 8.05092096e-01 8.22595179e-01 -5.06113529e-01
-7.21825123e-01 8.73720050e-01 1.59849554e-01 -3.75586718e-01
3.41219932e-01 -1.11538923e+00 3.11724961e-01 -1.39522791e-01
4.58355963e-01 9.21760723e-02 2.32223775e-02 -1.54394194e-01
-2.39840046e-01 4.52536732e-01 -2.95394599e-01 -8.85462910e-02
1.73318315e+00 -1.63905751e-02 1.51060864e-01 8.56909513e-01
1.50626075e+00 -8.89943480e-01 -1.49759471e+00 -6.14533067e-01
-1.22048147e-01 4.48665857e-01 -2.08981484e-01 -6.59915626e-01
-1.15439177e+00 1.02023387e+00 6.46963716e-01 3.83039355e-01
1.24617505e+00 -3.22480142e-01 1.07641768e+00 7.28285491e-01
2.34294206e-01 -1.00312448e+00 3.00011903e-01 1.05530989e+00
8.27140033e-01 -9.23772693e-01 -4.41883415e-01 3.77595365e-01
-4.11078334e-01 1.06078494e+00 3.80365163e-01 -3.97785664e-01
1.03274882e+00 5.10945678e-01 -1.71375602e-01 -1.06637880e-01
-1.36813843e+00 8.00543725e-02 -2.76605971e-03 2.67475992e-01
3.61566573e-01 -5.19732293e-03 -6.93390891e-02 5.93957603e-01
-3.64005506e-01 2.04284310e-01 2.92666107e-01 1.05114782e+00
-7.19895735e-02 -6.31532550e-01 1.85629562e-01 7.73051977e-01
-2.69382477e-01 -4.02898520e-01 2.80075014e-01 8.44470620e-01
-3.64357352e-01 4.64396745e-01 5.51517725e-01 -6.12260640e-01
1.57500833e-01 2.31953248e-01 3.47698152e-01 -2.89001316e-01
-7.63817012e-01 -1.62032340e-02 -2.85156727e-01 -8.08857262e-01
-2.71939605e-01 -8.73432755e-01 -1.28493488e+00 -6.31503403e-01
8.06197897e-02 -1.26344681e-01 5.48782349e-01 1.09839463e+00
6.37280643e-01 8.76416862e-01 5.69706202e-01 -8.28221798e-01
-9.54530060e-01 -8.91854405e-01 -4.15901989e-01 1.71209961e-01
9.90925193e-01 -2.74907023e-01 -2.05025271e-01 -1.13146212e-02] | [7.432005405426025, 3.3980491161346436] |
cd901cc0-9670-4e41-a35f-4e4cb9333c61 | self-supervised-learning-across-domains | 2007.12368 | null | https://arxiv.org/abs/2007.12368v2 | https://arxiv.org/pdf/2007.12368v2.pdf | Self-Supervised Learning Across Domains | Human adaptability relies crucially on learning and merging knowledge from both supervised and unsupervised tasks: the parents point out few important concepts, but then the children fill in the gaps on their own. This is particularly effective, because supervised learning can never be exhaustive and thus learning autonomously allows to discover invariances and regularities that help to generalize. In this paper we propose to apply a similar approach to the problem of object recognition across domains: our model learns the semantic labels in a supervised fashion, and broadens its understanding of the data by learning from self-supervised signals on the same images. This secondary task helps the network to focus on object shapes, learning concepts like spatial orientation and part correlation, while acting as a regularizer for the classification task over multiple visual domains. Extensive experiments confirm our intuition and show that our multi-task method combining supervised and self-supervised knowledge shows competitive results with respect to more complex domain generalization and adaptation solutions. It also proves its potential in the novel and challenging predictive and partial domain adaptation scenarios. | ["Antonio D'Innocente", 'Yujun Liao', 'Tatiana Tommasi', 'Barbara Caputo', 'Silvia Bucci', 'Fabio Maria Carlucci'] | 2020-07-24 | null | null | null | null | ['partial-domain-adaptation'] | ['methodology'] | [ 3.05795521e-01 7.17455521e-02 -4.43214595e-01 -6.47850275e-01
7.13793263e-02 -6.15949333e-01 5.22538245e-01 4.15168047e-01
-4.32214528e-01 8.21115613e-01 9.17502306e-03 2.96294242e-01
-4.99833494e-01 -8.18702161e-01 -5.51711798e-01 -8.58158469e-01
-1.37909368e-01 8.87134135e-01 5.84622264e-01 -3.95602167e-01
8.52210373e-02 5.32653332e-01 -1.78914547e+00 2.97402471e-01
8.71718645e-01 8.91322136e-01 3.95333380e-01 3.45173746e-01
-2.50712991e-01 8.34348798e-01 -3.00445050e-01 -1.78732231e-01
2.37129286e-01 -2.56275773e-01 -1.02638376e+00 3.06352913e-01
3.97206515e-01 1.96837261e-01 1.56245679e-01 8.66841137e-01
1.77793071e-01 3.08374703e-01 8.60523999e-01 -1.13795424e+00
-7.52317131e-01 4.58794534e-01 -3.95417333e-01 1.72750086e-01
2.56724745e-01 -1.42512158e-01 9.33937073e-01 -6.31810009e-01
5.96186280e-01 1.06876850e+00 7.81991541e-01 7.39819705e-01
-1.49488425e+00 -3.23246419e-01 4.24714684e-01 3.27130288e-01
-1.14226389e+00 -2.58356124e-01 9.59721446e-01 -5.48067868e-01
7.01265037e-01 -1.51030719e-01 6.00084960e-01 1.11350703e+00
-1.93983436e-01 8.44797969e-01 1.27987289e+00 -6.64136767e-01
5.04276752e-01 5.30379891e-01 1.07886933e-01 4.91546929e-01
1.25241697e-01 1.98645771e-01 -6.45559132e-01 3.06222200e-01
5.30891359e-01 1.53837800e-01 3.55701777e-03 -9.82471228e-01
-9.89450872e-01 6.60550892e-01 6.65166557e-01 6.79948688e-01
-3.08900803e-01 -3.62254202e-01 3.85927588e-01 6.10916674e-01
4.22588766e-01 7.01884747e-01 -8.28696966e-01 1.35974020e-01
-8.98798168e-01 3.23348902e-02 7.60865390e-01 8.73686135e-01
1.14014733e+00 -1.06947750e-01 1.23231873e-01 1.10504127e+00
-2.86737923e-03 4.29385513e-01 7.58862853e-01 -7.59506524e-01
-8.22669081e-03 9.33018506e-01 -2.06862628e-01 -8.18858743e-01
-7.40900934e-01 -7.35238075e-01 -8.35140646e-01 4.87312168e-01
4.60616589e-01 4.94072847e-02 -9.93602276e-01 1.97392714e+00
2.03481987e-01 -1.64244384e-01 1.84209235e-02 6.87485337e-01
5.84161639e-01 2.06497043e-01 2.87843466e-01 -1.76631257e-01
9.33470666e-01 -8.62673104e-01 -2.97266632e-01 -5.03161073e-01
3.94247711e-01 -3.16285551e-01 9.26567972e-01 4.41927701e-01
-8.24840069e-01 -1.02068543e+00 -9.42441404e-01 1.76052094e-01
-8.32485676e-01 9.67768729e-02 6.19407237e-01 4.58175480e-01
-1.01402450e+00 5.97112477e-01 -6.38592124e-01 -8.44060779e-01
6.57714307e-01 5.66334248e-01 -4.15769845e-01 -2.81608738e-02
-8.19558501e-01 1.03711128e+00 8.72307062e-01 -3.76649946e-01
-6.86946213e-01 -5.63643813e-01 -7.66867638e-01 -4.49786987e-03
4.16202128e-01 -6.12504780e-01 8.99044275e-01 -1.53934348e+00
-1.35230708e+00 1.14541709e+00 -1.80529505e-02 -5.78601837e-01
2.44525537e-01 -8.69316608e-02 -1.90564811e-01 5.87976128e-02
1.31545335e-01 9.51543629e-01 9.85346913e-01 -1.40658736e+00
-7.00407684e-01 -6.92921042e-01 -4.18842062e-02 5.65779842e-02
-7.62145877e-01 -5.14401972e-01 -1.98106542e-02 -7.38162875e-01
3.67582083e-01 -8.91300142e-01 -2.22563922e-01 -1.26438543e-01
-3.93318385e-02 -3.22209358e-01 8.95237446e-01 -2.97224641e-01
8.52594256e-01 -2.13210654e+00 4.81442422e-01 4.23046380e-01
1.19009867e-01 2.52611071e-01 -1.19547896e-01 2.09847093e-01
-2.24002063e-01 -3.10797900e-01 -4.45630401e-01 -1.03645861e-01
-1.46653771e-01 6.01466954e-01 -2.59264648e-01 1.56442702e-01
2.87720442e-01 1.07298636e+00 -9.59979236e-01 -4.14213926e-01
7.95808062e-02 -6.85178638e-02 -4.29916024e-01 1.90466836e-01
-4.50770438e-01 8.13430130e-01 -4.80140239e-01 5.34765542e-01
4.62223232e-01 -3.40100050e-01 3.20713669e-01 1.04974583e-01
1.75842255e-01 -8.89191311e-03 -1.10447884e+00 1.94693232e+00
-5.21015525e-01 4.40746427e-01 -1.99585199e-01 -1.83216488e+00
1.30230713e+00 -1.63231399e-02 6.58601046e-01 -8.24378431e-01
-2.66634692e-02 1.62479922e-01 -8.95158127e-02 -6.47449076e-01
8.82469863e-02 -2.33789086e-01 -6.37737960e-02 5.31623840e-01
7.03918040e-01 -1.77007139e-01 2.63442725e-01 -4.63684872e-02
6.71772718e-01 4.71784115e-01 6.02884710e-01 -4.10007805e-01
7.24600911e-01 4.84097488e-02 4.03310597e-01 7.78507352e-01
-2.59705279e-02 1.93845958e-01 1.67001784e-01 -6.72185302e-01
-9.16403294e-01 -1.24715269e+00 -2.22813934e-01 1.69075620e+00
1.23870738e-01 -7.57491738e-02 -5.21996856e-01 -1.04858100e+00
1.23792149e-01 3.95781845e-01 -9.12190497e-01 -4.12610978e-01
-4.00483102e-01 -5.04175663e-01 1.35441110e-01 8.89422834e-01
5.29412866e-01 -1.24644899e+00 -7.04930723e-01 1.51651531e-01
2.61839092e-01 -1.04532456e+00 2.30125055e-01 7.45813012e-01
-1.07327342e+00 -1.01941979e+00 -6.85447812e-01 -1.06674623e+00
8.33358467e-01 -1.67830233e-02 1.15995443e+00 -5.98121025e-02
-3.20687562e-01 6.91413522e-01 -5.34361005e-01 -5.88978410e-01
-4.84657645e-01 3.12450230e-01 2.57481903e-01 1.25733659e-01
6.09164536e-01 -1.18430793e+00 -6.45561516e-02 4.37664062e-01
-8.54175210e-01 -2.38866672e-01 6.84284568e-01 9.92944837e-01
4.19289380e-01 1.20466493e-01 9.51475263e-01 -9.61313903e-01
2.85718888e-01 -3.26889336e-01 -4.46048558e-01 4.60786074e-01
-7.71448374e-01 3.38349491e-01 8.46119106e-01 -5.77209830e-01
-1.13805354e+00 5.91522694e-01 1.76896676e-01 -1.49778992e-01
-7.14524329e-01 1.47951365e-01 -1.98303327e-01 -2.60441184e-01
1.03354287e+00 4.15010035e-01 9.09027178e-03 -6.49937451e-01
5.33758342e-01 3.77307564e-01 6.48279190e-01 -7.75596857e-01
9.40530121e-01 6.45893872e-01 1.69217855e-01 -7.81381249e-01
-1.04980803e+00 -6.10944033e-01 -1.53674066e+00 -1.10718317e-01
6.94039226e-01 -6.54547513e-01 -6.05718732e-01 3.23278993e-01
-8.50696802e-01 -3.98588151e-01 -8.95863891e-01 3.22339028e-01
-8.01239252e-01 2.52583027e-01 -1.71909221e-02 -5.79000175e-01
2.34445557e-01 -6.28985882e-01 7.29911745e-01 3.99930775e-01
-2.02301010e-01 -1.41837645e+00 2.33577624e-01 4.89232829e-03
3.61267030e-01 2.62732133e-02 9.57639813e-01 -1.16996837e+00
-2.37702042e-01 5.15452065e-02 -9.47820619e-02 6.28184736e-01
3.27691346e-01 -5.34883499e-01 -1.10891819e+00 -2.67984420e-01
-5.97789465e-03 -7.10252345e-01 1.17484820e+00 2.63735354e-01
1.16783488e+00 1.81730743e-03 -3.80154669e-01 5.51838636e-01
1.25509810e+00 7.98295427e-04 1.83293819e-01 3.87252450e-01
5.81184447e-01 1.02930152e+00 4.99867767e-01 3.20570856e-01
2.92060226e-01 7.31065035e-01 2.93034881e-01 -1.59505919e-01
-1.60110250e-01 -1.89503834e-01 9.96066928e-02 5.08017004e-01
-2.50032008e-01 3.64023596e-01 -1.05573273e+00 5.10184944e-01
-1.91318381e+00 -9.07356918e-01 1.64401680e-01 2.20043898e+00
9.56632614e-01 1.42411903e-01 4.21295464e-01 2.42154390e-01
5.34853101e-01 -5.95753267e-02 -6.81880534e-01 -2.51410872e-01
-3.48033339e-01 4.07500714e-01 2.64047980e-01 1.95567340e-01
-1.33749712e+00 1.08568466e+00 6.60977602e+00 6.66711211e-01
-1.09563649e+00 6.27917796e-02 4.59631830e-01 2.81530440e-01
3.27308103e-02 -6.10215254e-02 -5.98394096e-01 1.35032162e-01
4.31447893e-01 -1.37071218e-02 4.67092425e-01 9.92526829e-01
-2.27815837e-01 -2.21879721e-01 -1.50377381e+00 6.97724938e-01
2.35542476e-01 -1.06014836e+00 1.66931041e-02 -1.26414120e-01
9.62153614e-01 -1.57018766e-01 6.02775328e-02 3.40606034e-01
3.77281994e-01 -9.57185745e-01 5.47670841e-01 5.90483904e-01
4.81952399e-01 -4.01692361e-01 5.11585474e-01 7.24644482e-01
-9.87027884e-01 -5.67265630e-01 -4.18022454e-01 -2.53603101e-01
-2.71083683e-01 2.67765522e-01 -9.07349348e-01 4.92109537e-01
8.06034207e-01 1.09262991e+00 -9.40765798e-01 9.17777956e-01
-4.44219202e-01 3.65047574e-01 -2.73197711e-01 9.26299691e-02
4.83680554e-02 -1.03997365e-02 5.11955261e-01 1.25288951e+00
-1.11199200e-01 -1.21217690e-01 4.73843396e-01 7.67792761e-01
1.51892602e-01 1.66698441e-01 -7.23038435e-01 2.76481658e-01
1.44811764e-01 9.97402370e-01 -9.03837800e-01 -3.74245137e-01
-3.66539448e-01 1.03677607e+00 5.92549503e-01 4.24927175e-01
-3.38468909e-01 -4.66467477e-02 1.77316338e-01 -2.85546891e-02
5.35707891e-01 -2.25104973e-01 -5.29043198e-01 -1.08069515e+00
5.39878989e-03 -6.97080374e-01 6.42514944e-01 -5.74207544e-01
-1.44425619e+00 3.80940974e-01 8.95995945e-02 -1.32600236e+00
-4.18388009e-01 -9.47772026e-01 -4.31324631e-01 4.29275066e-01
-1.72870886e+00 -1.31333899e+00 -4.21610624e-01 8.98304760e-01
6.41235650e-01 -5.12535453e-01 9.75117981e-01 -7.23850206e-02
3.94420438e-02 5.64175904e-01 1.83621839e-01 2.09369808e-01
8.09818447e-01 -1.39230084e+00 -5.71498685e-02 6.06453180e-01
6.08636260e-01 4.45485532e-01 4.94330615e-01 -4.49543864e-01
-8.76299381e-01 -8.65882337e-01 5.99807024e-01 -4.66660440e-01
6.05048776e-01 -3.29603672e-01 -1.12023365e+00 4.80598509e-01
-8.44371915e-02 5.97124361e-02 6.41325414e-01 4.69666630e-01
-6.36871338e-01 -4.05708104e-01 -9.97810245e-01 1.67044327e-01
1.14815056e+00 -3.48337561e-01 -8.57258916e-01 4.65837568e-01
3.88141930e-01 3.04285996e-02 -6.70680344e-01 4.37549949e-01
4.36744481e-01 -1.14109206e+00 1.01643908e+00 -1.02938366e+00
2.96519488e-01 -1.05776832e-01 -1.17533565e-01 -1.51605284e+00
-4.96873498e-01 -2.52105385e-01 7.44471401e-02 1.18017435e+00
4.33460534e-01 -6.48797929e-01 7.69188166e-01 1.62129421e-02
8.37200284e-02 -4.90871459e-01 -7.54572570e-01 -1.12744105e+00
2.04698324e-01 -3.25538844e-01 3.98108244e-01 1.12838197e+00
-2.27715215e-03 4.12460595e-01 -1.15165971e-01 6.15097545e-02
6.20179713e-01 4.27023441e-01 7.00186849e-01 -1.81205893e+00
-3.62681180e-01 -4.82289433e-01 -7.02452719e-01 -9.47306812e-01
4.36675221e-01 -1.14431834e+00 -7.58339837e-02 -1.13805544e+00
2.90308833e-01 -5.70476651e-01 -3.91065240e-01 8.20758462e-01
9.04328823e-02 2.06095457e-01 1.12653702e-01 2.90546387e-01
-8.54823828e-01 4.33865458e-01 1.09066057e+00 -2.68837780e-01
-3.83228868e-01 2.86886096e-01 -6.45315826e-01 9.35383558e-01
7.45453477e-01 -3.86569083e-01 -5.13020694e-01 -3.08183193e-01
2.92998906e-02 -5.55953860e-01 6.22106016e-01 -1.19313216e+00
4.59366590e-01 -2.82132655e-01 7.39541829e-01 -7.65212700e-02
1.26253292e-01 -1.14624941e+00 -5.28009355e-01 3.14313799e-01
-5.16073108e-01 -3.77127737e-01 1.56339139e-01 6.00841165e-01
-2.47040316e-01 -2.34572932e-01 1.01256216e+00 -3.53629798e-01
-1.23355520e+00 1.54718250e-01 -4.81523760e-02 1.66244581e-01
1.06277776e+00 -5.76481700e-01 -7.25872740e-02 -2.36225635e-01
-1.41512144e+00 2.89478928e-01 5.33957720e-01 6.10539496e-01
4.76411909e-01 -1.13212967e+00 -5.90162337e-01 5.40253103e-01
4.63871866e-01 -1.21764183e-01 -1.91449828e-03 6.99181437e-01
9.51338485e-02 2.96106160e-01 -6.98074996e-01 -1.10574424e+00
-1.07966745e+00 8.78122151e-01 2.67861426e-01 -8.43212008e-02
-4.19633001e-01 8.28281760e-01 3.75648618e-01 -5.23599386e-01
3.74828398e-01 -1.97415818e-02 -5.59728384e-01 2.13764772e-01
2.97011554e-01 2.69553542e-01 3.96869667e-02 -6.48323059e-01
-3.63388121e-01 9.44055200e-01 -2.77435575e-02 1.22839540e-01
1.63404500e+00 -2.33096138e-01 -1.53974146e-01 5.93239963e-01
1.06274402e+00 -1.86790541e-01 -1.33785868e+00 -7.47855723e-01
5.08542001e-01 -2.63728470e-01 -4.39090431e-01 -1.07037282e+00
-7.73450673e-01 9.42007780e-01 5.42593718e-01 3.92149925e-01
1.54907894e+00 5.42027175e-01 1.27105266e-01 6.79836392e-01
4.02677447e-01 -1.40752554e+00 4.82671678e-01 6.68952107e-01
7.21758425e-01 -1.33498836e+00 1.07524157e-01 -3.11437577e-01
-6.48411155e-01 1.39628100e+00 7.41986632e-01 -1.60334826e-01
5.59998214e-01 7.78798237e-02 -1.50533430e-02 -1.23278871e-01
-5.76930285e-01 -5.74659050e-01 4.79582042e-01 1.08943260e+00
2.33202830e-01 -7.08438531e-02 9.41503514e-03 4.58232522e-01
-1.72457606e-01 -9.29312482e-02 7.18907118e-02 8.24272394e-01
-7.82855272e-01 -1.38579655e+00 -2.39607617e-01 1.41813338e-01
1.95090957e-02 2.79486567e-01 -5.31371355e-01 7.34770238e-01
6.77234113e-01 6.60286605e-01 6.44564861e-03 -3.20120811e-01
2.22632200e-01 3.25986624e-01 7.22293019e-01 -8.39057565e-01
-3.11986804e-01 -1.96905315e-01 -3.66397798e-01 -4.18240666e-01
-8.48044097e-01 -7.08437443e-01 -1.18793869e+00 4.76807117e-01
1.39546208e-02 1.62443951e-01 5.33876300e-01 1.23984993e+00
9.17541906e-02 4.82940853e-01 8.77254665e-01 -8.10541987e-01
-5.66005468e-01 -8.45243275e-01 -4.56929922e-01 7.07748413e-01
4.42431360e-01 -9.23900127e-01 -8.94483551e-02 5.62852681e-01] | [9.773608207702637, 2.646498441696167] |
2cd911e0-720a-4cf7-b680-f9de38accebc | revisiting-prototypical-network-for-cross | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhou_Revisiting_Prototypical_Network_for_Cross_Domain_Few-Shot_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhou_Revisiting_Prototypical_Network_for_Cross_Domain_Few-Shot_Learning_CVPR_2023_paper.pdf | Revisiting Prototypical Network for Cross Domain Few-Shot Learning | Prototypical Network is a popular few-shot solver that aims at establishing a feature metric generalizable to novel few-shot classification (FSC) tasks using deep neural networks. However, its performance drops dramatically when generalizing to the FSC tasks in new domains. In this study, we revisit this problem and argue that the devil lies in the simplicity bias pitfall in neural networks. In specific, the network tends to focus on some biased shortcut features (e.g., color, shape, etc.) that are exclusively sufficient to distinguish very few classes in the meta-training tasks within a pre-defined domain, but fail to generalize across domains as some desirable semantic features. To mitigate this problem, we propose a Local-global Distillation Prototypical Network (LDP-net). Different from the standard Prototypical Network, we establish a two-branch network to classify the query image and its random local crops, respectively. Then, knowledge distillation is conducted among these two branches to enforce their class affiliation consistency. The rationale behind is that since such global-local semantic relationship is expected to hold regardless of data domains, the local-global distillation is beneficial to exploit some cross-domain transferable semantic features for feature metric establishment. Moreover, such local-global semantic consistency is further enforced among different images of the same class to reduce the intra-class semantic variation of the resultant feature. In addition, we propose to update the local branch as Exponential Moving Average (EMA) over training episodes, which makes it possible to better distill cross-episode knowledge and further enhance the generalization performance. Experiments on eight cross-domain FSC benchmarks empirically clarify our argument and show the state-of-the-art results of LDP-net. Code is available in https://github.com/NWPUZhoufei/LDP-Net | ['Yanning Zhang', 'Wei Wei', 'Lei Zhang', 'Peng Wang', 'Fei Zhou'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['cross-domain-few-shot', 'cross-domain-few-shot-learning'] | ['computer-vision', 'computer-vision'] | [ 3.25655013e-01 -1.17966942e-01 -2.78744489e-01 -5.68182588e-01
-5.30043423e-01 -4.16685820e-01 5.35836220e-01 5.80360368e-03
-4.81157720e-01 8.68221998e-01 -1.06241882e-01 1.76693708e-01
-5.57381868e-01 -1.09485459e+00 -7.33409822e-01 -7.32613802e-01
1.59823343e-01 2.30006114e-01 4.41585779e-01 -4.44147378e-01
1.20238304e-01 2.99612880e-01 -1.72778094e+00 2.55669296e-01
1.06289542e+00 1.20231998e+00 2.79762447e-01 -1.00614525e-01
-3.49627405e-01 5.60757160e-01 -6.04811072e-01 -3.36629629e-01
2.70075709e-01 -6.59949481e-01 -9.30194736e-01 -2.77201626e-02
2.87871540e-01 -7.99653977e-02 3.48006412e-02 1.36693645e+00
3.87012333e-01 5.80532134e-01 6.20575607e-01 -1.41101015e+00
-8.64036679e-01 4.92830843e-01 -4.36514884e-01 8.92034993e-02
1.36658520e-01 3.29674333e-01 1.12843573e+00 -6.87914610e-01
8.33309054e-01 1.00742400e+00 4.60799307e-01 7.27087617e-01
-1.05068195e+00 -8.36463809e-01 4.07672256e-01 4.68873322e-01
-1.37643421e+00 -1.01275973e-01 1.00831676e+00 -1.42755941e-01
7.03317881e-01 1.94314003e-01 5.95931649e-01 1.44640529e+00
-1.09617583e-01 7.43458688e-01 1.07567024e+00 -2.99767494e-01
5.62721789e-01 2.15308413e-01 3.06449980e-01 5.50987124e-01
3.26890051e-01 1.04373477e-01 -4.53352809e-01 1.32627442e-01
4.18539107e-01 1.25494391e-01 -4.60149735e-01 -6.75060630e-01
-8.97274852e-01 1.05175006e+00 6.13469005e-01 6.50708318e-01
-7.04184473e-02 -1.28810167e-01 6.43829465e-01 4.05222863e-01
4.70721066e-01 7.72094071e-01 -5.04769981e-01 -1.47074118e-01
-7.50964582e-01 1.52950570e-01 5.57705164e-01 9.73461151e-01
1.12599742e+00 -7.43773580e-02 -2.06269577e-01 1.14436626e+00
-2.90640742e-01 1.05419777e-01 7.18937993e-01 -8.98369074e-01
3.19291621e-01 7.28770554e-01 -3.24561179e-01 -1.03383660e+00
-2.77574956e-01 -5.61708212e-01 -8.45969200e-01 1.36389667e-02
2.14351825e-02 -1.49018794e-01 -8.61925542e-01 1.93254244e+00
1.98476136e-01 2.63700157e-01 7.68482732e-03 9.46166217e-01
8.38286459e-01 4.28161740e-01 3.97044830e-02 -1.07353568e-01
1.11730754e+00 -8.98987949e-01 -4.02949601e-01 -2.04642639e-01
6.74072266e-01 -1.77163392e-01 1.17624891e+00 1.55089855e-01
-4.69783962e-01 -7.56374300e-01 -1.29409015e+00 7.64736533e-02
-9.42591608e-01 -3.01325142e-01 6.46160483e-01 4.79271978e-01
-6.95515454e-01 8.87123346e-01 -4.52095479e-01 -5.58208048e-01
6.05998993e-01 7.55102634e-02 -2.21490949e-01 -2.23583683e-01
-1.65191996e+00 8.52544606e-01 9.73358214e-01 -3.05012554e-01
-6.61069989e-01 -8.88251424e-01 -9.20434654e-01 2.24120408e-01
7.34410107e-01 -4.84682888e-01 8.66361976e-01 -1.34651136e+00
-1.45898330e+00 7.88222849e-01 1.41488969e-01 -3.90637845e-01
4.00017232e-01 8.50716159e-02 -6.79146469e-01 3.86423133e-02
3.36147249e-01 7.88993478e-01 7.44044363e-01 -1.21995640e+00
-8.60384703e-01 -1.52880937e-01 3.47130507e-01 6.38800189e-02
-6.27879679e-01 -2.51672804e-01 -3.59527498e-01 -6.36457264e-01
-6.34797066e-02 -7.82795608e-01 -2.56200582e-02 6.23226017e-02
-2.49383971e-01 -3.93947393e-01 8.62869382e-01 -7.35628977e-02
1.22255492e+00 -2.33289385e+00 -4.34039310e-02 1.34503469e-01
-2.24963203e-03 3.34969759e-01 -3.99630010e-01 1.26372784e-01
-2.61703014e-01 -1.37413181e-02 -5.02458215e-01 -1.07180931e-01
-1.58267599e-02 4.62530851e-01 -2.17716858e-01 2.93524921e-01
4.58639055e-01 8.85442615e-01 -1.06039262e+00 -3.53217393e-01
1.65459603e-01 1.21540852e-01 -2.98837006e-01 -2.34899908e-01
-1.94705158e-01 1.46801993e-01 -5.34254491e-01 6.96052790e-01
8.20170164e-01 -3.58709961e-01 6.93749711e-02 -9.00065899e-02
1.06391266e-01 -2.80277189e-02 -1.21400547e+00 1.94450569e+00
-4.35541660e-01 3.73385340e-01 -4.81986046e-01 -1.63742280e+00
1.14585102e+00 -4.15604077e-02 5.32775223e-01 -8.02082956e-01
2.80660957e-01 3.01188350e-01 4.40496616e-02 -2.65716106e-01
3.38036746e-01 -2.61656791e-01 -1.73777297e-01 1.29294083e-01
3.55713278e-01 2.16004308e-02 3.61746699e-01 1.73382033e-02
8.95865560e-01 7.29063600e-02 4.66196865e-01 -4.48681116e-01
5.54793477e-01 1.04215950e-01 9.75094020e-01 6.97770000e-01
-5.67914844e-01 7.62394845e-01 4.16215986e-01 -4.71241444e-01
-7.19731510e-01 -1.03551066e+00 -2.19160214e-01 1.23610294e+00
6.10849142e-01 -2.93422848e-01 -4.96455848e-01 -9.58437026e-01
-1.34377964e-02 9.07039344e-01 -8.92544448e-01 -6.00281477e-01
-2.54845440e-01 -7.52313316e-01 4.90138531e-01 5.92305064e-01
8.52959871e-01 -1.15275788e+00 -7.69381404e-01 2.90064961e-01
-6.14073016e-02 -9.40724134e-01 -1.98000997e-01 3.63122433e-01
-5.90754092e-01 -1.18676031e+00 -8.12579274e-01 -7.67343462e-01
4.91099119e-01 3.78999770e-01 9.97203887e-01 -2.61830371e-02
-2.57009506e-01 1.31555200e-01 -8.00767064e-01 -1.78464711e-01
-1.06903985e-02 1.89642861e-01 -9.06631425e-02 7.45714083e-02
7.02410579e-01 -6.25117481e-01 -4.89201039e-01 3.01081866e-01
-9.85550463e-01 9.21071395e-02 2.42624149e-01 1.05378962e+00
4.51656014e-01 2.28944898e-01 7.48591602e-01 -9.26143229e-01
5.03793478e-01 -5.85606754e-01 -4.15682346e-01 5.39459825e-01
-6.03705525e-01 -5.62793389e-02 8.69564950e-01 -6.18716121e-01
-1.13312268e+00 -1.44710183e-01 9.21399295e-02 -5.02048016e-01
-2.86981702e-01 4.42319572e-01 -3.54374051e-01 3.30000706e-02
7.07796156e-01 3.83512855e-01 -9.41015854e-02 -2.54180759e-01
3.28236699e-01 3.51733893e-01 2.25623399e-01 -8.63357604e-01
7.33890951e-01 4.49273080e-01 -2.17310190e-01 -5.17844975e-01
-1.03027296e+00 -4.88652110e-01 -4.51739669e-01 -8.02093893e-02
8.33426535e-01 -6.21241629e-01 -3.60454232e-01 3.75378191e-01
-9.94800508e-01 -2.13486835e-01 -4.89268064e-01 3.06424320e-01
-5.71856081e-01 1.60931513e-01 -3.17529321e-01 -3.68722290e-01
-2.65817404e-01 -1.02012849e+00 7.80803740e-01 4.56818014e-01
-1.02967352e-01 -8.97792220e-01 -7.13275447e-02 -2.35071555e-02
4.56197292e-01 3.88818145e-01 9.26141024e-01 -8.69207025e-01
-1.79635242e-01 1.55766338e-01 -4.43385243e-01 5.64446211e-01
4.27803189e-01 -1.45278424e-01 -9.56482947e-01 -3.14455956e-01
1.81298144e-02 -5.04122019e-01 1.06159210e+00 2.11963207e-01
1.47669291e+00 6.39619380e-02 -2.39646181e-01 8.47486079e-01
1.70170808e+00 4.06641513e-01 4.69635397e-01 6.40809238e-01
3.74387920e-01 5.10826826e-01 7.77216017e-01 4.37622696e-01
1.01998098e-01 4.94636685e-01 2.29087338e-01 1.05629213e-01
-9.89782214e-02 5.24842087e-03 2.31424093e-01 4.62048680e-01
-1.34169355e-01 -6.67943656e-02 -6.95656955e-01 5.73872149e-01
-1.90833342e+00 -9.32611108e-01 4.27477449e-01 1.93225825e+00
8.87163281e-01 1.76933840e-01 -4.47513573e-02 -1.41201034e-01
9.75218952e-01 3.48370492e-01 -7.66995847e-01 -2.97703147e-01
-3.48446697e-01 4.33933586e-01 4.19557571e-01 -5.24987318e-02
-1.07162976e+00 1.01342511e+00 4.35999584e+00 1.33941615e+00
-1.32898045e+00 2.36360818e-01 6.23249471e-01 -1.10212259e-01
-1.47515953e-01 -4.41084690e-02 -7.56461382e-01 5.55678785e-01
4.23024505e-01 -4.17431384e-01 4.58126366e-01 1.01839161e+00
-2.44190425e-01 -3.02667990e-02 -1.05502892e+00 9.02918637e-01
1.15037449e-01 -1.37266886e+00 5.57852760e-02 -1.69453621e-01
9.51072216e-01 -2.40687542e-02 5.26882336e-02 7.97462940e-01
3.13464701e-01 -7.78600872e-01 5.43934762e-01 2.15194270e-01
9.21572685e-01 -8.69623244e-01 6.54853821e-01 2.55799055e-01
-1.18381655e+00 -2.08934367e-01 -6.69232965e-01 8.54401663e-02
-8.90486687e-02 4.87335920e-01 -4.19910908e-01 9.38229501e-01
9.26342845e-01 9.30853963e-01 -6.91075623e-01 8.05410922e-01
-1.90818340e-01 1.94797650e-01 -6.06725439e-02 -2.22548209e-02
5.60900211e-01 -1.28784284e-01 4.06385660e-01 1.01949930e+00
4.33302015e-01 1.23258671e-02 1.29040629e-01 1.04908407e+00
5.75087732e-03 2.13393848e-02 -5.54620564e-01 1.28996268e-01
4.99245733e-01 1.06415629e+00 -8.21373582e-01 -6.27216041e-01
-5.40919960e-01 1.05780566e+00 4.91037369e-01 4.60989028e-01
-1.14433837e+00 -7.03475118e-01 7.66027987e-01 -1.91134751e-01
4.54778165e-01 2.94354975e-01 -3.15365195e-01 -1.25291264e+00
1.75099507e-01 -8.17773461e-01 5.82151592e-01 -3.65824491e-01
-1.68615568e+00 5.54816008e-01 1.19906828e-01 -1.51546550e+00
6.22267015e-02 -6.15673304e-01 -7.49333680e-01 6.46907985e-01
-1.82617772e+00 -9.41253960e-01 -4.62256223e-01 8.05501163e-01
7.40325689e-01 -2.24119931e-01 7.77464092e-01 1.76145643e-01
-5.99038839e-01 6.92223847e-01 2.13823542e-01 2.17681453e-01
8.79101753e-01 -1.03202987e+00 1.25375837e-01 7.16637015e-01
-1.07717186e-01 7.65658617e-01 4.70547467e-01 -4.98422712e-01
-7.94192791e-01 -1.35369039e+00 4.28305566e-01 -2.89143156e-03
6.74210131e-01 -1.77465275e-01 -1.19692397e+00 3.66035074e-01
6.15109829e-03 3.08455586e-01 5.75940788e-01 2.34154254e-01
-5.76027334e-01 -3.09245139e-01 -1.15311146e+00 3.81720662e-01
1.25109577e+00 -4.00735795e-01 -7.36464798e-01 1.86982930e-01
7.66497552e-01 9.29818489e-03 -7.36584067e-01 6.47773504e-01
3.68075848e-01 -1.05691683e+00 8.02485585e-01 -6.37418866e-01
6.76609218e-01 -1.90412819e-01 -2.99916983e-01 -1.48456597e+00
-4.17451382e-01 -8.97704661e-02 5.30483350e-02 1.31325424e+00
3.46632928e-01 -9.07949209e-01 7.09532142e-01 4.17062908e-01
-1.89005807e-01 -7.60577917e-01 -9.78948057e-01 -1.50524902e+00
2.97459275e-01 -3.64524275e-01 7.86341846e-01 1.43257892e+00
-3.74076553e-02 2.04852477e-01 -1.78427964e-01 -1.15370899e-01
4.09033328e-01 3.95104080e-01 4.63302404e-01 -1.42863774e+00
-3.19179416e-01 -5.86991608e-01 -4.16195273e-01 -6.89971864e-01
4.49529529e-01 -9.83754575e-01 1.05608784e-01 -1.18282354e+00
3.77331257e-01 -5.14978409e-01 -7.82601297e-01 6.42935574e-01
-2.30030343e-01 3.41559984e-02 4.04515475e-01 -9.37200561e-02
-7.60010242e-01 8.29299033e-01 1.31076181e+00 -3.78059000e-01
-2.43463129e-01 -1.78192034e-01 -8.11074913e-01 4.73329216e-01
9.86280799e-01 -5.19085646e-01 -5.65879107e-01 -1.85406908e-01
-1.06830373e-01 -4.04726028e-01 4.48706746e-01 -1.15512776e+00
2.26267472e-01 -4.56060380e-01 3.49816918e-01 -1.22271381e-01
2.81727314e-01 -7.67906070e-01 -1.04587160e-01 4.53132361e-01
-3.21372330e-01 -4.78282094e-01 2.84078345e-02 4.64152724e-01
-3.77455652e-01 -4.82003927e-01 1.02348733e+00 -2.61195153e-01
-1.32617688e+00 3.61556053e-01 1.72235951e-01 3.44292730e-01
1.17801332e+00 -4.96325821e-01 -5.72695553e-01 -1.88243333e-02
-4.90494043e-01 2.40372762e-01 5.35592198e-01 7.25389361e-01
5.65499067e-01 -1.44338298e+00 -4.55905288e-01 9.53657776e-02
5.83559871e-01 5.19610755e-02 3.77523035e-01 6.32496178e-01
-1.08268857e-01 3.11219960e-01 -4.76272613e-01 -4.21316296e-01
-9.32277322e-01 7.00685322e-01 3.54967177e-01 -1.82761490e-01
-6.63066089e-01 8.99023652e-01 4.28006470e-01 -4.61884141e-01
7.08370358e-02 -1.98378548e-01 -8.57540369e-02 1.17425725e-01
4.38947022e-01 2.75545210e-01 -2.22116951e-02 -2.78149337e-01
-4.94874388e-01 4.94208455e-01 -1.41207382e-01 3.17191064e-01
1.43652880e+00 5.34965321e-02 5.65502346e-02 4.45399076e-01
1.51127493e+00 -6.22027397e-01 -1.29941499e+00 -4.48707044e-01
6.20394461e-02 -3.83989781e-01 -1.52699515e-01 -8.16312313e-01
-1.21678066e+00 6.64747596e-01 6.77281916e-01 4.57006395e-02
1.35839736e+00 -4.58352454e-02 8.08386385e-01 4.48443592e-01
4.53026891e-01 -1.55035102e+00 7.72861391e-02 6.55272007e-01
6.90742612e-01 -1.39771962e+00 -7.65133649e-02 -2.43809611e-01
-7.69025624e-01 1.16321039e+00 9.98042941e-01 -2.46067092e-01
4.96695280e-01 -2.41556436e-01 -2.96039522e-01 -2.49016300e-01
-5.12543380e-01 -4.32794064e-01 2.47005627e-01 6.75443411e-01
1.46781191e-01 8.74322131e-02 -4.06547755e-01 7.73692608e-01
-5.41428253e-02 5.74970692e-02 2.47556269e-01 9.57745433e-01
-5.12701213e-01 -8.97222221e-01 2.06539817e-02 3.59015018e-01
-1.94175094e-02 -8.38100389e-02 -1.60246626e-01 1.01245964e+00
6.73201978e-01 7.82641649e-01 1.10999078e-01 -3.21341276e-01
3.13987345e-01 4.80717942e-02 3.08646709e-01 -7.41472006e-01
-3.31570327e-01 -3.70964587e-01 -1.69500262e-01 -5.54219186e-01
-4.31754887e-01 -3.90840441e-01 -1.23345912e+00 -1.50495559e-01
-2.87690103e-01 2.13659909e-02 3.14754128e-01 1.08177853e+00
2.88046300e-01 6.78798854e-01 6.36678934e-01 -4.01320755e-01
-5.27920723e-01 -7.92055190e-01 -6.89820111e-01 7.39199221e-01
1.01768866e-01 -1.10327435e+00 -4.25303370e-01 -1.60018981e-01] | [9.963224411010742, 2.9332547187805176] |
6b8de838-7371-4c32-99ae-4c458827dbf6 | deep-learning-provides-rapid-screen-for | 2301.05938 | null | https://arxiv.org/abs/2301.05938v1 | https://arxiv.org/pdf/2301.05938v1.pdf | Deep Learning Provides Rapid Screen for Breast Cancer Metastasis with Sentinel Lymph Nodes | Deep learning has been shown to be useful to detect breast cancer metastases by analyzing whole slide images of sentinel lymph nodes. However, it requires extensive scanning and analysis of all the lymph nodes slides for each case. Our deep learning study focuses on breast cancer screening with only a small set of image patches from any sentinel lymph node, positive or negative for metastasis, to detect changes in tumor environment and not in the tumor itself. We design a convolutional neural network in the Python language to build a diagnostic model for this purpose. The excellent results from this preliminary study provided a proof of concept for incorporating automated metastatic screen into the digital pathology workflow to augment the pathologists' productivity. Our approach is unique since it provides a very rapid screen rather than an exhaustive search for tumor in all fields of all sentinel lymph nodes. | ['Andy N. D. Nguyen', 'Hongxia Sun', 'Amer Wahed', 'Karan Saluja', 'Kevin Chiu', 'Jianmin Ding', 'Songlin Zhang', 'Xiaohong Iris Wang', 'Kareem Allam'] | 2023-01-14 | null | null | null | null | ['whole-slide-images'] | ['computer-vision'] | [ 2.78760284e-01 1.73413932e-01 -2.43029729e-01 -1.52859464e-01
-1.06259823e+00 -7.28261888e-01 -2.83054076e-03 6.24562621e-01
-5.62724352e-01 2.60803312e-01 4.27922700e-03 -1.01431084e+00
1.98496968e-01 -9.09388959e-01 -4.81481761e-01 -1.01755011e+00
-2.24084370e-02 5.91325462e-01 5.48956156e-01 -6.40183315e-03
-1.66737005e-01 1.09000993e+00 -7.61390626e-01 6.29812479e-01
1.57426864e-01 6.43732429e-01 2.89385468e-01 1.27845514e+00
-2.23226130e-01 9.44915414e-01 -2.62538254e-01 -1.94919780e-01
2.05811188e-01 -3.02683771e-01 -6.77178085e-01 -2.60706961e-01
5.69145560e-01 -6.49553776e-01 -1.41130805e-01 9.29516196e-01
8.53330612e-01 -8.20557892e-01 2.37694517e-01 -5.87045729e-01
1.51230931e-01 3.42605531e-01 -6.85536802e-01 5.52422583e-01
-7.32144043e-02 5.15433192e-01 6.50792599e-01 -4.65798169e-01
9.98918295e-01 6.53272927e-01 1.18196559e+00 6.30116105e-01
-1.15348423e+00 -8.35547149e-01 -7.13910103e-01 -1.12097234e-01
-9.51595366e-01 -1.82114840e-01 1.38915405e-01 -5.33578098e-01
8.12372863e-01 4.10928130e-01 1.19223428e+00 7.41337836e-01
9.59954321e-01 5.29979825e-01 8.19625020e-01 -3.95921975e-01
9.05459374e-02 1.80114985e-01 1.67392984e-01 9.39095497e-01
3.67082715e-01 -1.03939436e-01 -5.86137883e-02 -2.16528013e-01
7.05056489e-01 5.09554327e-01 2.58611858e-01 -1.69047207e-01
-1.03927445e+00 7.36037433e-01 6.44887924e-01 4.95762527e-01
-2.49287903e-01 3.93438071e-01 8.17580819e-01 1.88488454e-01
1.73944980e-01 1.87930882e-01 -1.24809787e-01 1.27916157e-01
-8.95031750e-01 -2.65105158e-01 5.39732993e-01 4.59576882e-02
4.93052721e-01 -7.74968624e-01 -2.63391227e-01 2.87617266e-01
1.89519912e-01 5.29659867e-01 5.83985627e-01 -5.77468991e-01
-3.23674381e-01 1.09575522e+00 -1.19749963e-01 -8.95858884e-01
-1.04835212e+00 -3.76342714e-01 -9.32182491e-01 4.44100559e-01
5.02789915e-01 -2.13824868e-01 -1.25287890e+00 1.04579544e+00
5.58829725e-01 -2.63124347e-01 -3.17787796e-01 5.60472250e-01
9.79654491e-01 1.81279749e-01 3.74072194e-01 2.91250736e-01
1.60097587e+00 -6.00109637e-01 -5.26695311e-01 -5.50894104e-02
1.44306874e+00 -6.72460794e-01 8.18547010e-01 9.99161378e-02
-9.22273636e-01 -1.39355808e-01 -7.86655188e-01 -9.17135552e-02
-6.39786482e-01 3.46699744e-01 9.76309240e-01 7.98296809e-01
-1.52186871e+00 3.55929434e-02 -1.38397276e+00 -9.64209855e-01
8.94203484e-01 6.91681266e-01 -6.43806994e-01 -2.42985301e-02
-6.58432364e-01 7.51682639e-01 4.78786118e-02 1.46380737e-01
-9.57625568e-01 -9.79008615e-01 -5.56278110e-01 5.88320568e-03
-1.25463799e-01 -3.71784776e-01 1.30881023e+00 -1.08534729e+00
-9.69688892e-01 1.47561800e+00 -3.37379843e-01 -2.62866229e-01
5.59729576e-01 6.70449138e-01 -7.86401704e-02 4.05814081e-01
1.11943431e-01 7.77523816e-01 -2.24949643e-02 -5.06399751e-01
-8.85484457e-01 -4.61242199e-01 -3.88625890e-01 -1.91225633e-01
-3.41781586e-01 3.88469584e-02 -6.00717306e-01 -3.55686732e-02
-1.86883122e-01 -8.27996135e-01 -7.40417600e-01 5.74865639e-01
-2.33909398e-01 2.27809906e-01 6.59884870e-01 -6.79719448e-01
8.16416800e-01 -2.21791840e+00 -5.91257095e-01 4.73290563e-01
5.24472952e-01 3.99855942e-01 -8.73280019e-02 1.13714315e-01
-4.47592475e-02 6.08346045e-01 5.62507987e-01 1.08445272e-01
-5.11026442e-01 -1.14789151e-01 4.38294202e-01 7.57986426e-01
1.82914168e-01 1.23718131e+00 -7.87065208e-01 -8.55292022e-01
1.19860165e-01 4.13180113e-01 -3.63091022e-01 -1.01571895e-01
7.30121508e-02 2.39796251e-01 -3.96080166e-01 9.84551013e-01
7.70791769e-01 -6.40330732e-01 5.17094493e-01 8.14411789e-02
1.95311189e-01 -2.08699971e-01 -6.58510208e-01 1.34450841e+00
-1.22968182e-01 8.86599898e-01 5.31395793e-01 -3.69169176e-01
4.77626145e-01 2.68470019e-01 7.15346813e-01 -7.30960131e-01
2.71180898e-01 3.32932085e-01 3.05455536e-01 -7.13432789e-01
-3.54336724e-02 -3.22873294e-01 3.06728750e-01 5.98863482e-01
-2.44536579e-01 1.95621103e-01 2.82395691e-01 1.84339926e-01
1.91083872e+00 -7.84856975e-01 4.72363472e-01 -3.04369599e-01
3.35524797e-01 7.06535220e-01 3.45553845e-01 8.39787066e-01
-5.55489421e-01 4.23796713e-01 9.20658469e-01 -7.34492898e-01
-9.68925476e-01 -8.41406345e-01 -3.67422312e-01 8.18351209e-01
-1.47527203e-01 2.03437939e-01 -3.74069452e-01 -8.23913753e-01
1.64266467e-01 -3.48321408e-01 -1.13797414e+00 2.32130989e-01
-4.79766577e-01 -1.04021454e+00 7.67916739e-01 5.85173726e-01
2.36479014e-01 -8.51661205e-01 -7.99286067e-01 2.04933256e-01
3.30027908e-01 -4.61452901e-01 9.98307690e-02 6.03606105e-01
-7.61036038e-01 -1.62195385e+00 -7.47092843e-01 -1.16959405e+00
1.20929158e+00 7.68623576e-02 8.83130670e-01 4.65312719e-01
-1.15030551e+00 4.10799161e-02 -1.05193935e-01 -7.84459174e-01
-9.56847847e-01 1.51605889e-01 -6.31944180e-01 -2.61202365e-01
7.96284139e-01 5.34821041e-02 -7.64389277e-01 9.12950635e-02
-9.26602185e-01 7.20726103e-02 1.11424112e+00 8.89222503e-01
8.24491441e-01 7.36502782e-02 2.83622537e-02 -1.17577589e+00
2.92339146e-01 -5.32376826e-01 -5.51486850e-01 9.62203220e-02
-2.53242571e-02 -5.56394279e-01 4.28012908e-01 -2.47423008e-01
-6.61104858e-01 2.88925827e-01 -4.03063595e-01 2.36632034e-01
-1.78003967e-01 5.45574546e-01 4.29309785e-01 -4.88336623e-01
8.71588349e-01 -1.43460870e-01 7.05302179e-01 1.61753558e-02
-6.51160896e-01 4.82092261e-01 3.20511609e-01 5.01722991e-01
3.16997826e-01 1.03322077e+00 4.69332576e-01 -6.80530310e-01
-3.37374449e-01 -9.18700159e-01 -4.68406290e-01 -3.74653101e-01
8.03869367e-01 -7.03998327e-01 -9.52635944e-01 3.23632210e-01
-6.77170992e-01 -7.60391772e-01 -1.95594758e-01 2.93911308e-01
2.28433698e-01 -2.90278066e-02 -1.09137034e+00 -1.72553614e-01
-6.56173885e-01 -1.18702161e+00 1.32785416e+00 3.24631304e-01
-3.55997711e-01 -1.26907766e+00 5.16497970e-01 -1.16264589e-01
7.79243231e-01 5.67926109e-01 8.81225049e-01 -8.50230813e-01
-4.07292813e-01 -1.16769123e+00 -3.86391342e-01 -2.31702879e-01
2.04821885e-01 5.43703973e-01 -9.03141737e-01 -3.62982810e-01
-5.49494743e-01 -2.49341100e-01 8.56411397e-01 5.81708908e-01
1.14960670e+00 2.45780423e-01 -1.19766653e+00 7.29783177e-01
1.61645937e+00 1.26364216e-01 5.83498240e-01 5.29780626e-01
2.77651012e-01 2.28495657e-01 3.16868216e-01 1.26336232e-01
-2.55700886e-01 -3.24627720e-02 5.24882317e-01 -1.04922855e+00
-2.11291224e-01 1.81012258e-01 7.34958798e-02 -2.46050358e-01
4.22347754e-01 -3.11293662e-01 -1.52008963e+00 7.07191169e-01
-1.13193178e+00 -5.90916634e-01 -4.62157965e-01 1.59117508e+00
7.95033395e-01 -1.33672550e-01 -1.23896025e-01 -1.51805788e-01
7.08680093e-01 -4.10064936e-01 -4.90784198e-01 -2.97937900e-01
8.19121003e-02 3.89929563e-01 6.69951200e-01 3.76832575e-01
-1.01453650e+00 5.20482183e-01 7.85399199e+00 5.50309241e-01
-1.67235720e+00 4.16742153e-02 1.12982535e+00 -3.24690819e-01
-1.32816806e-02 -3.73991758e-01 -8.12311649e-01 1.47407576e-01
8.00783694e-01 1.69599921e-01 -4.18337792e-01 6.21946990e-01
3.13392460e-01 -6.13485992e-01 -1.09396935e+00 5.05610347e-01
-4.56212640e-01 -1.97785795e+00 -1.88649148e-01 4.04203951e-01
4.18059468e-01 4.17814553e-01 -5.37746698e-02 -6.86201602e-02
4.97295380e-01 -1.11491847e+00 -3.46368343e-01 4.49943066e-01
1.01619351e+00 -3.94566059e-01 1.59852660e+00 5.96279204e-02
-5.77975333e-01 2.34070676e-03 -5.45155555e-02 3.15059394e-01
-4.44100380e-01 6.44169509e-01 -1.89650774e+00 -2.54938662e-01
7.86492527e-01 3.87880266e-01 -1.16496658e+00 8.92343581e-01
4.32217687e-01 7.26051331e-01 -4.17862773e-01 -2.85257727e-01
2.47929230e-01 8.22123408e-01 1.98715832e-02 1.57645464e+00
4.01694119e-01 -3.75930220e-01 -1.48090392e-01 2.56692290e-01
1.81238994e-01 2.53565431e-01 -3.54368031e-01 -1.76708087e-01
1.79052159e-01 1.81496966e+00 -1.32087719e+00 -1.68014109e-01
-3.82383078e-01 5.49652755e-01 1.25834662e-02 8.26072842e-02
-5.39238632e-01 -2.82470495e-01 9.26980525e-02 4.76232797e-01
9.98505279e-02 4.91743684e-01 -4.56534922e-01 -3.00372958e-01
-3.30412537e-01 -6.88247263e-01 6.78196490e-01 -4.46803331e-01
-9.61627364e-01 8.82061943e-02 -6.96928501e-01 -9.53520954e-01
-6.75676689e-02 -1.06423175e+00 -9.79533911e-01 7.87072718e-01
-1.43678832e+00 -1.25896251e+00 -7.88491368e-01 4.64601696e-01
-1.17472835e-01 5.13487086e-02 1.06219041e+00 2.18838565e-02
-5.35261393e-01 6.60154939e-01 2.55238354e-01 5.21201313e-01
8.63359451e-01 -1.36355102e+00 -2.02757359e-01 4.76626366e-01
-1.03591943e+00 5.13698637e-01 3.25758278e-01 -6.49544060e-01
-1.69572866e+00 -1.12097597e+00 8.11326504e-01 -5.82427144e-01
8.43542397e-01 -1.76692873e-01 -4.10875052e-01 8.57250750e-01
2.24711433e-01 3.81548822e-01 1.35919869e+00 -1.02440067e-01
3.08636665e-01 -3.82367909e-01 -1.40947175e+00 6.45403385e-01
1.49644479e-01 -8.19149464e-02 4.26698297e-01 5.50522327e-01
1.11917004e-01 -6.97828531e-01 -8.59592199e-01 3.88696551e-01
7.31719971e-01 -9.41913009e-01 6.30285680e-01 -1.86769173e-01
5.20762622e-01 -1.41648501e-01 4.08973217e-01 -6.96645439e-01
-4.39410359e-01 -7.87220597e-02 5.43569207e-01 7.60417342e-01
6.60905361e-01 -5.62497914e-01 1.43397903e+00 3.74600679e-01
-1.99823617e-03 -7.73344040e-01 -8.53405416e-01 5.73694855e-02
1.30204737e-01 -7.73260668e-02 2.39917383e-01 4.71262515e-01
-1.67475417e-02 -3.34466994e-01 6.74403727e-01 1.16962619e-01
2.02588543e-01 -2.61677295e-01 8.82194638e-01 -9.38435018e-01
-2.26584181e-01 -6.44569516e-01 -8.53020012e-01 2.90827155e-02
-3.90784264e-01 -1.16656983e+00 -9.83219668e-02 -1.54950440e+00
8.36612821e-01 -4.09812093e-01 -4.31749791e-01 8.03184509e-01
-1.45476177e-01 6.78449571e-01 -4.60719168e-01 8.88234824e-02
-5.67302346e-01 -7.76116610e-01 1.46752441e+00 -4.50291663e-01
-6.74935207e-02 -1.23242944e-01 -9.92693305e-01 6.40024185e-01
7.11853862e-01 -5.05562603e-01 2.79818386e-01 -3.24192226e-01
5.06422520e-01 -1.98132209e-02 5.76422811e-01 -9.76470113e-01
4.68429834e-01 5.28669134e-02 1.14530981e+00 -8.86180520e-01
-1.03903338e-01 -9.56656873e-01 2.89891690e-01 1.12165213e+00
-4.18121845e-01 -3.50362062e-01 6.60841107e-01 3.23086917e-01
-1.39603585e-01 9.97997671e-02 8.98953021e-01 -4.88892019e-01
-5.81803203e-01 1.46908507e-01 -1.00574934e+00 -7.41788626e-01
1.33117259e+00 -5.49786150e-01 -3.59311849e-01 5.13803996e-02
-8.38660181e-01 4.72383678e-01 7.99516857e-01 -4.77870882e-01
1.79028049e-01 -9.22159851e-01 -8.18955779e-01 2.84100682e-01
2.33165428e-01 2.10242853e-01 4.75408852e-01 1.33415031e+00
-1.43197370e+00 3.52670312e-01 -3.15868288e-01 -9.93097782e-01
-1.73589170e+00 3.21783543e-01 9.06808376e-01 -9.32669342e-01
-6.19747460e-01 1.18890989e+00 9.33771394e-03 -4.70687121e-01
3.65244091e-01 -3.25291544e-01 -1.40333295e-01 6.67125881e-02
8.37839603e-01 3.96417081e-02 3.87606800e-01 1.56138033e-01
-5.33470154e-01 1.63295448e-01 -4.76961672e-01 2.92774200e-01
1.14379418e+00 2.89523154e-01 -5.03323853e-01 1.68410584e-01
1.31816351e+00 3.47333923e-02 -8.27079475e-01 1.36297911e-01
-3.61247033e-01 -1.30411997e-01 2.96137810e-01 -1.03960657e+00
-1.06745887e+00 7.42049038e-01 1.06800997e+00 2.24136472e-01
1.35522556e+00 1.30516440e-01 4.41730678e-01 4.43044126e-01
-6.03834242e-02 -6.78549647e-01 -1.48332268e-01 1.26925319e-01
2.83898711e-01 -1.61058331e+00 1.58898443e-01 -1.90773100e-01
-1.05663255e-01 1.57972419e+00 4.90867227e-01 -3.09808515e-02
7.03515351e-01 9.25745368e-01 6.03299379e-01 -5.97041667e-01
-7.82433689e-01 1.41223103e-01 -4.72164661e-01 4.88120079e-01
9.27794576e-01 1.08056717e-01 -2.94662286e-02 1.89249590e-01
8.85645375e-02 3.79794836e-01 5.37814140e-01 1.25890303e+00
-5.23382246e-01 -7.99203217e-01 -3.85276228e-01 7.73670614e-01
-8.67044330e-01 1.40294328e-01 -9.34082806e-01 1.20803857e+00
-2.26614662e-02 4.12274122e-01 3.52288187e-01 -1.00122549e-01
-5.69433868e-02 -1.71602786e-01 6.56737909e-02 -6.22078657e-01
-1.00293589e+00 2.66573280e-01 -2.32054628e-02 -5.91849744e-01
-1.06200993e-01 -5.52023590e-01 -1.22854280e+00 -4.34689403e-01
-5.05085476e-02 -1.68375984e-01 7.56013989e-01 5.03640652e-01
3.00175786e-01 8.44648182e-01 4.88356024e-01 -6.19846582e-01
5.16607501e-02 -9.22446728e-01 -8.56307447e-01 -1.48638725e-01
6.60814941e-01 8.28759149e-02 -3.01090807e-01 -4.69379174e-03] | [15.132198333740234, -3.106757164001465] |
7af6a8b3-b2b9-4564-b548-b5585fbee7df | smaclite-a-lightweight-environment-for-multi | 2305.05566 | null | https://arxiv.org/abs/2305.05566v1 | https://arxiv.org/pdf/2305.05566v1.pdf | SMAClite: A Lightweight Environment for Multi-Agent Reinforcement Learning | There is a lack of standard benchmarks for Multi-Agent Reinforcement Learning (MARL) algorithms. The Starcraft Multi-Agent Challenge (SMAC) has been widely used in MARL research, but is built on top of a heavy, closed-source computer game, StarCraft II. Thus, SMAC is computationally expensive and requires knowledge and the use of proprietary tools specific to the game for any meaningful alteration or contribution to the environment. We introduce SMAClite -- a challenge based on SMAC that is both decoupled from Starcraft II and open-source, along with a framework which makes it possible to create new content for SMAClite without any special knowledge. We conduct experiments to show that SMAClite is equivalent to SMAC, by training MARL algorithms on SMAClite and reproducing SMAC results. We then show that SMAClite outperforms SMAC in both runtime speed and memory. | ['Stefano V. Albrecht', 'Filippos Christianos', 'Adam Michalski'] | 2023-05-09 | null | null | null | null | ['multi-agent-reinforcement-learning', 'starcraft-ii', 'smac-1', 'starcraft', 'smac'] | ['methodology', 'playing-games', 'playing-games', 'playing-games', 'playing-games'] | [-6.43100977e-01 -2.06422716e-01 -2.02984229e-01 3.59988242e-01
-6.51829243e-01 -7.54693329e-01 8.09691131e-01 4.06389415e-01
-6.73852265e-01 1.05546594e+00 -5.49594648e-02 -4.33632344e-01
-1.80222392e-01 -8.81501734e-01 -7.36499727e-01 -4.97853249e-01
-3.74861002e-01 9.22265947e-01 6.73978806e-01 -9.19179499e-01
3.18050869e-02 2.12606072e-01 -1.58781266e+00 5.62112220e-02
5.81761837e-01 3.15682620e-01 1.97699636e-01 1.05814493e+00
2.60248512e-01 1.31185675e+00 -6.90972686e-01 -2.31196240e-01
4.49483007e-01 -6.80650830e-01 -9.87792492e-01 -3.76988620e-01
-1.24855526e-01 -1.82101890e-01 9.94282868e-03 6.20557606e-01
5.94324887e-01 3.39884788e-01 1.88504547e-01 -1.82305408e+00
6.16037212e-02 6.00776672e-01 -4.47719783e-01 -1.24497369e-01
4.76700157e-01 6.29171073e-01 1.00145721e+00 -2.31064528e-01
8.14452350e-01 1.17912400e+00 7.59555876e-01 6.48005128e-01
-9.75876868e-01 -3.61539006e-01 -1.82465702e-01 1.97999194e-01
-9.06518698e-01 -2.36304492e-01 6.23025000e-01 -1.48929745e-01
1.22294354e+00 1.89122573e-01 8.38365078e-01 1.22951162e+00
2.05658287e-01 1.03585100e+00 1.34148061e+00 -5.13223290e-01
8.83393466e-01 -2.72775829e-01 -2.81888038e-01 1.08896959e+00
1.68237045e-01 4.82761621e-01 -4.80935127e-01 -3.26372713e-01
9.51704860e-01 -6.36511326e-01 1.80255905e-01 -9.13116813e-01
-1.32387197e+00 1.00365162e+00 6.80914149e-02 1.84592396e-01
-4.10811007e-01 6.69990182e-01 7.20432281e-01 5.97048700e-01
-7.64982626e-02 7.21904695e-01 -3.98931086e-01 -8.94146144e-01
-5.64502418e-01 9.91325021e-01 1.10987949e+00 5.21421134e-01
7.24075973e-01 1.55663148e-01 2.86753953e-01 6.16376936e-01
3.18720639e-01 3.17280293e-01 7.35333741e-01 -1.51434457e+00
3.67442518e-02 3.07495415e-01 3.76897633e-01 -7.10025430e-01
-6.70113087e-01 -3.19585316e-02 -1.32867128e-01 1.00280225e+00
3.36398780e-01 -5.87229371e-01 -2.03202397e-01 1.60198522e+00
4.62722570e-01 4.89485443e-01 5.73754489e-01 9.54311550e-01
7.42894113e-01 5.56196153e-01 -1.06018834e-01 1.29403616e-03
1.02921772e+00 -1.39754879e+00 -2.67817944e-01 -3.83835912e-01
8.71412218e-01 -4.61582929e-01 1.07301652e+00 4.02406275e-01
-1.23035932e+00 -6.26647398e-02 -1.34781766e+00 4.81504053e-01
-3.67636770e-01 -5.37653029e-01 1.06126666e+00 6.73465371e-01
-1.15889692e+00 6.24312043e-01 -1.20944667e+00 -2.72985220e-01
-1.35194674e-01 2.55095184e-01 -3.00699979e-01 1.71134129e-01
-1.05362356e+00 1.24397051e+00 6.62232697e-01 -6.93104565e-01
-1.25801694e+00 -7.91239619e-01 -1.11159396e+00 -2.72764146e-01
7.08306432e-01 -6.39282584e-01 1.91852534e+00 -8.60619247e-01
-2.18685579e+00 4.79512751e-01 5.26708543e-01 -7.12863326e-01
7.04459965e-01 -1.22037277e-01 -2.48655573e-01 7.72667602e-02
2.34923899e-01 5.14586270e-01 5.59387624e-01 -1.21509433e+00
-9.26722229e-01 2.37941682e-01 5.71975052e-01 3.21801573e-01
3.21896315e-01 -2.50588767e-02 -1.19744807e-01 -4.28521603e-01
-9.30723608e-01 -9.35649037e-01 -5.01971185e-01 -4.12588447e-01
3.15527350e-01 -2.82317370e-01 8.61047506e-01 -1.47329062e-01
7.19382882e-01 -1.71450543e+00 1.68918386e-01 -1.01349622e-01
2.79382885e-01 4.05101448e-01 -7.85839438e-01 9.55572307e-01
3.98441494e-01 -1.62601084e-01 4.53744978e-02 -3.12557369e-02
4.09208745e-01 6.47355616e-01 2.37251535e-01 4.52990413e-01
-1.03688031e-01 9.08112884e-01 -1.56582272e+00 -4.06323612e-01
4.90380168e-01 -8.45550969e-02 -8.90788674e-01 1.12648241e-01
-8.19435775e-01 2.51924932e-01 -2.74479896e-01 4.46740776e-01
2.89792925e-01 -4.36964352e-03 3.37000996e-01 5.09407938e-01
-2.69309461e-01 1.12523742e-01 -1.35207760e+00 2.14115596e+00
-7.20781565e-01 2.27093533e-01 2.21372977e-01 -8.55868518e-01
6.09250128e-01 2.88470566e-01 6.15809321e-01 -8.17883372e-01
5.64156733e-02 3.66796076e-01 2.99808457e-02 -4.28142458e-01
3.89225572e-01 1.69517682e-03 -2.93278813e-01 9.21434164e-01
3.78120989e-01 -4.00860161e-01 8.77539754e-01 1.46034464e-01
1.47320092e+00 5.08303165e-01 5.16012073e-01 -2.48868614e-01
3.90716314e-01 4.77607638e-01 7.24096715e-01 1.12218165e+00
-3.14210385e-01 -1.93453893e-01 4.37228113e-01 -7.16779113e-01
-9.15349782e-01 -1.22414315e+00 6.53148055e-01 1.28464901e+00
2.14583442e-01 -8.05964828e-01 -7.59379566e-01 -8.15601766e-01
9.11700428e-02 7.40973413e-01 -4.81165111e-01 1.27605751e-01
-4.63912249e-01 -5.86099625e-01 7.58242726e-01 7.11523890e-02
8.31255138e-01 -1.63513017e+00 -1.26458645e+00 5.99913776e-01
1.47760347e-01 -9.40125525e-01 -2.28253201e-01 -3.96433752e-03
-2.96121091e-01 -1.48691177e+00 -2.94203877e-01 -6.84414089e-01
-9.17138830e-02 8.22349191e-02 1.53667951e+00 2.32703537e-01
-3.16172808e-01 8.39141607e-01 -7.86465526e-01 -4.63736534e-01
-9.02281582e-01 2.03619391e-01 -1.49950907e-02 -5.69516778e-01
-8.01142380e-02 -7.78531373e-01 -3.19319099e-01 9.02479738e-02
-7.12568641e-01 3.67011279e-01 3.73229951e-01 8.81476939e-01
-4.94185947e-02 1.93264753e-01 6.93889797e-01 -6.44372463e-01
8.55757117e-01 -3.67778540e-01 -1.03517616e+00 -5.56207933e-02
-3.46729487e-01 1.50887236e-01 6.10711813e-01 -1.40166566e-01
-6.29903853e-01 -1.14961155e-01 -2.93835551e-01 -1.00275569e-01
-1.90191165e-01 6.94397151e-01 1.65893540e-01 -3.19524676e-01
7.20523179e-01 1.62977189e-01 5.92767060e-01 -1.76460177e-01
4.47782934e-01 2.43373886e-01 5.56835175e-01 -9.98009086e-01
6.17982090e-01 1.11036201e-03 -5.45831621e-02 -5.66757023e-01
-4.76680666e-01 -9.25756916e-02 -4.88426797e-02 -3.00740778e-01
4.81818914e-01 -7.78570652e-01 -1.25767720e+00 5.80723941e-01
-8.41819584e-01 -1.26334286e+00 -3.75042945e-01 3.16067696e-01
-1.06678486e+00 8.85470957e-02 -7.47869551e-01 -4.67598498e-01
-2.32858822e-01 -1.13992059e+00 5.04042447e-01 2.29160979e-01
-2.04150826e-01 -1.33858061e+00 9.36013579e-01 2.28821009e-01
6.22076094e-01 4.81637299e-01 6.96561992e-01 -6.75628066e-01
-2.06421733e-01 -8.89982097e-03 3.32707793e-01 6.60855472e-02
7.10002035e-02 -3.76368724e-02 -4.87741679e-01 -6.97869360e-01
-4.97607410e-01 -9.64316189e-01 2.27001101e-01 9.41414610e-02
3.62170249e-01 -3.58993858e-01 1.31450176e-01 3.03532213e-01
1.65348053e+00 2.40125120e-01 4.27747607e-01 1.25887764e+00
1.96893498e-01 1.02676786e-01 6.32792950e-01 8.97855759e-01
6.92354679e-01 6.66012704e-01 8.27216089e-01 -1.03036769e-01
1.86888799e-01 -1.61503226e-01 7.02018082e-01 7.67376304e-01
-3.04483533e-01 6.07677475e-02 -9.51054692e-01 4.84887153e-01
-2.45156789e+00 -1.08450317e+00 2.22254172e-01 1.94231248e+00
1.03917408e+00 3.54609229e-02 5.74895144e-01 -9.82036069e-02
2.00745776e-01 2.94534385e-01 -4.98205513e-01 -7.66143739e-01
3.05169933e-02 6.20236814e-01 3.87905627e-01 7.86379516e-01
-1.04618466e+00 1.28279805e+00 6.83025837e+00 8.83312285e-01
-8.51579070e-01 2.37111583e-01 -1.84710175e-01 -1.26907706e-01
1.10883480e-02 1.82466626e-01 -3.11559528e-01 1.81905076e-01
9.58841264e-01 -6.53773606e-01 9.86654222e-01 1.03667533e+00
3.44594330e-01 -2.51041204e-01 -8.50709140e-01 6.56892776e-01
-1.99964806e-01 -1.76928914e+00 -2.38970712e-01 -2.11376980e-01
6.85635328e-01 4.50961411e-01 -3.36752594e-01 9.31254029e-01
1.55389631e+00 -8.39158654e-01 6.79205298e-01 -1.60689741e-01
2.46066675e-01 -1.12053537e+00 6.62400067e-01 5.25234759e-01
-1.00784814e+00 6.73829615e-02 -2.45060503e-01 -5.61633408e-01
-1.50723845e-01 -1.00919507e-01 -7.68012583e-01 6.64344251e-01
5.63800633e-01 6.57491207e-01 -5.77001929e-01 1.14824891e+00
-4.53543454e-01 5.10656953e-01 -3.98004539e-02 -2.95469195e-01
7.70580649e-01 -1.83984607e-01 4.77995664e-01 9.23720837e-01
-3.07363626e-02 -2.42734522e-01 7.67950118e-01 4.87275332e-01
2.70054966e-01 3.85830365e-02 -5.49380660e-01 -6.42021000e-02
4.68432009e-01 1.45114863e+00 -3.78554642e-01 -1.95470482e-01
-5.49302578e-01 9.41443741e-01 4.85088378e-01 2.03221366e-02
-1.00042248e+00 -3.88019651e-01 1.10186505e+00 -1.56569809e-01
2.57856667e-01 -3.42604816e-01 3.43777001e-01 -9.90960836e-01
-4.78780091e-01 -1.55886757e+00 4.67628777e-01 -7.36189723e-01
-1.04465961e+00 4.60715830e-01 -3.19758058e-01 -1.01511145e+00
-8.61662805e-01 -4.27857816e-01 -7.83041596e-01 3.81142408e-01
-1.49240506e+00 -1.10278797e+00 -2.09419161e-01 7.96115637e-01
4.80256706e-01 -6.73530221e-01 1.26575351e+00 -9.39405933e-02
-6.57065868e-01 3.22752506e-01 2.81057835e-01 -9.54768732e-02
5.23711503e-01 -1.63307750e+00 4.81214523e-01 5.89104831e-01
-1.14578910e-01 -5.14693484e-02 9.53369737e-01 -4.19283748e-01
-1.66391826e+00 -9.57583547e-01 1.90461740e-01 -2.49334872e-01
1.21186399e+00 -2.37222567e-01 -4.04243469e-01 6.70324266e-01
7.64761209e-01 1.12987228e-01 5.21175742e-01 1.64954349e-01
-2.43541777e-01 1.09010801e-01 -1.00758719e+00 9.15924489e-01
6.79253519e-01 -1.76336452e-01 -5.20628572e-01 2.85749614e-01
5.56413054e-01 -6.19371474e-01 -9.37250316e-01 1.31969184e-01
3.45287979e-01 -1.25212538e+00 7.96873093e-01 -6.58075809e-01
4.71355855e-01 -4.74983871e-01 8.39622878e-03 -1.90254200e+00
-2.83517867e-01 -9.48616445e-01 -3.90878655e-02 8.52706909e-01
2.11046502e-01 -6.67743742e-01 9.94379461e-01 7.88680539e-02
-2.01601967e-01 -3.64575565e-01 -8.55249465e-01 -1.11739290e+00
3.31309438e-01 -4.15689081e-01 6.33785486e-01 8.70953023e-01
4.64251220e-01 3.13315719e-01 -4.59616274e-01 9.62343216e-02
6.51410222e-01 2.34516680e-01 1.32338786e+00 -8.88869643e-01
-9.00405526e-01 -4.96458143e-01 -1.96326301e-01 -4.15643364e-01
2.93415189e-01 -8.55239630e-01 1.64803252e-01 -1.49706984e+00
-5.40998019e-02 -5.23627877e-01 -1.26200616e-01 8.11748326e-01
7.38270953e-02 2.84645874e-02 5.25774598e-01 5.15968502e-02
-1.23220527e+00 7.23802149e-01 1.16205180e+00 -8.34678635e-02
-4.30737615e-01 -2.70071328e-01 -4.41621333e-01 8.35633397e-01
1.06631017e+00 -2.94761062e-01 -4.68597829e-01 -1.90238535e-01
3.68697613e-01 3.03626895e-01 4.14136052e-01 -1.43215013e+00
9.35535654e-02 -5.29642045e-01 -2.99414426e-01 -2.48603169e-02
1.01970069e-01 -4.70481306e-01 1.13186404e-01 8.47420394e-01
-1.43800735e-01 4.71792728e-01 5.14801860e-01 9.08036530e-02
-1.51989996e-01 -5.33863127e-01 7.58077204e-01 -4.68091637e-01
-1.11217165e+00 -6.38343617e-02 -1.03599381e+00 4.16406840e-01
1.41900551e+00 3.24507833e-01 -6.40359402e-01 -4.65213895e-01
-4.30309504e-01 5.67879200e-01 7.37010181e-01 3.38546097e-01
3.18064988e-01 -1.27333951e+00 -8.20861220e-01 1.32234156e-01
1.72353029e-01 -2.70191491e-01 -7.07740560e-02 3.97798330e-01
-1.14029312e+00 6.60747737e-02 -6.71871483e-01 -6.69576898e-02
-1.13441288e+00 6.34565115e-01 5.47536671e-01 -8.04310977e-01
-8.01294386e-01 2.13514328e-01 -4.54298019e-01 -8.54903698e-01
2.04014760e-02 1.14043467e-01 1.09799549e-01 -3.97461921e-01
5.05004466e-01 2.71898270e-01 -1.05358362e-02 -3.65463197e-02
-2.90178955e-01 -3.08477227e-03 4.47409451e-02 -5.84287822e-01
1.60813177e+00 3.73612106e-01 8.51333141e-03 2.15947151e-01
3.92737508e-01 -4.24135439e-02 -1.30210018e+00 -4.26251180e-02
1.77252173e-01 -3.39987464e-02 3.12260129e-02 -9.79028285e-01
-9.49036956e-01 9.59896818e-02 3.67593348e-01 -6.31517842e-02
7.93844819e-01 -3.57464015e-01 7.36350834e-01 7.86240935e-01
9.83694017e-01 -1.24760997e+00 4.07043368e-01 8.62031162e-01
9.06022966e-01 -1.09570575e+00 -5.32553233e-02 9.30843800e-02
-8.05348277e-01 8.81381810e-01 8.14364910e-01 -4.80963737e-01
3.68787944e-01 6.16296113e-01 2.57214457e-01 -2.05220759e-01
-1.27501512e+00 -4.18743134e-01 -6.36913717e-01 9.73459661e-01
-3.86315808e-02 1.09100214e-03 -2.92736888e-01 5.06342113e-01
-3.61942261e-01 2.37094522e-01 9.06880438e-01 1.56558156e+00
-3.92509073e-01 -1.54658544e+00 -3.60212117e-01 -3.95903848e-02
-5.56721911e-02 1.77756116e-01 -2.02903703e-01 1.16067672e+00
4.51681316e-02 8.64653885e-01 -1.03556372e-01 -2.07445845e-01
1.85927063e-01 -3.31773371e-01 7.95362532e-01 -3.83772850e-01
-9.03029323e-01 -2.27596775e-01 4.15587336e-01 -8.13748896e-01
-4.62457627e-01 -4.50572193e-01 -1.33378530e+00 -7.26626337e-01
2.50159681e-01 5.57912052e-01 6.14922047e-01 7.87009537e-01
3.44390512e-01 3.78025889e-01 4.57375735e-01 -1.01789093e+00
-3.67307752e-01 -5.39067686e-01 -7.14209080e-01 1.19082429e-01
1.47420213e-01 -7.99980581e-01 2.14660689e-02 -4.53967124e-01] | [3.8120429515838623, 1.7160718441009521] |
a0789f57-c15f-424a-84f6-3a74bc271d60 | semantic-hierarchical-priors-for-intrinsic | 1902.03830 | null | https://arxiv.org/abs/1902.03830v2 | https://arxiv.org/pdf/1902.03830v2.pdf | Semantic Hierarchical Priors for Intrinsic Image Decomposition | Intrinsic Image Decomposition (IID) is a challenging and interesting computer vision problem with various applications in several fields. We present novel semantic priors and an integrated approach for single image IID that involves analyzing image at three hierarchical context levels. Local context priors capture scene properties at each pixel within a small neighbourhood. Mid-level context priors encode object level semantics. Global context priors establish correspondences at the scene level. Our semantic priors are designed on both fixed and flexible regions, using selective search method and Convolutional Neural Network features. Our IID method is an iterative multistage optimization scheme and consists of two complementary formulations: $L_2$ smoothing for shading and $L_1$ sparsity for reflectance. Experiments and analysis of our method indicate the utility of our semantic priors and structured hierarchical analysis in an IID framework. We compare our method with other contemporary IID solutions and show results with lesser artifacts. Finally, we highlight that proper choice and encoding of prior knowledge can produce competitive results even when compared to end-to-end deep learning IID methods, signifying the importance of such priors. We believe that the insights and techniques presented in this paper would be useful in the future IID research. | ['P. J. Narayanan', 'Saurabh Saini'] | 2019-02-11 | null | null | null | null | ['intrinsic-image-decomposition'] | ['computer-vision'] | [ 6.51780725e-01 1.68943942e-01 9.59933698e-02 -5.11976063e-01
-3.81895900e-01 -2.05915362e-01 5.38649619e-01 -1.16493486e-01
-1.02988444e-01 4.17018682e-01 5.02991259e-01 2.13532388e-01
-2.62731403e-01 -6.17662430e-01 -5.39295197e-01 -8.14150631e-01
1.24665964e-02 2.39531342e-02 5.28136969e-01 -7.19298497e-02
4.32277173e-01 6.31255388e-01 -1.77969921e+00 4.32300061e-01
6.00220144e-01 1.14412236e+00 8.37005496e-01 4.24945951e-01
-9.37398449e-02 7.33948529e-01 -2.67226070e-01 -6.03698986e-03
7.23957062e-01 -2.57133693e-01 -7.70036161e-01 4.44394410e-01
8.55174482e-01 -1.47102490e-01 -1.48582757e-01 1.10917819e+00
5.43316424e-01 2.21569672e-01 5.06105185e-01 -7.54587948e-01
-6.12069905e-01 -5.60748987e-02 -8.50402415e-01 2.03608334e-01
2.89300025e-01 1.06498502e-01 1.15123045e+00 -1.18423414e+00
6.70502245e-01 1.28271556e+00 8.13861251e-01 2.38127425e-01
-1.46911955e+00 -6.03480404e-03 6.27690256e-01 6.82059228e-02
-1.10935330e+00 -3.56103033e-01 1.21605074e+00 -5.37353754e-01
1.15679431e+00 1.26678318e-01 4.32866871e-01 7.14231193e-01
1.22279182e-01 9.48842645e-01 1.35739589e+00 -5.98746002e-01
2.58953929e-01 -8.91360417e-02 2.62924939e-01 7.84056067e-01
1.42852142e-01 2.15397537e-01 -7.16112316e-01 4.92864754e-03
1.02433383e+00 -2.45381609e-01 -3.46806914e-01 -6.20672166e-01
-1.05497122e+00 6.88973010e-01 3.87262881e-01 -2.48658732e-02
-5.16134381e-01 3.12431842e-01 1.50790274e-01 -2.21246611e-02
6.74321532e-01 7.77224973e-02 -6.37381196e-01 4.03730571e-01
-9.64364588e-01 2.42950264e-02 5.71272492e-01 8.81390631e-01
9.11369860e-01 4.38457541e-02 -2.52858132e-01 1.17922521e+00
4.93406981e-01 4.30291951e-01 -5.04030026e-02 -1.37311101e+00
-1.00759808e-02 4.02167052e-01 1.66559532e-01 -1.32404828e+00
-4.79296625e-01 -5.41145623e-01 -7.44746029e-01 6.25951409e-01
1.49122998e-01 3.14817995e-01 -9.82922256e-01 1.66594768e+00
4.41139102e-01 2.40142107e-01 -9.68614295e-02 1.09524000e+00
1.12384486e+00 4.70756024e-01 -4.21810783e-02 -1.01387195e-01
1.48708296e+00 -1.08591032e+00 -5.29990792e-01 -4.55270022e-01
-1.50429443e-01 -1.06267881e+00 1.16892374e+00 5.66826761e-01
-1.32200623e+00 -8.61755669e-01 -7.76422739e-01 -4.55137581e-01
-6.07118495e-02 3.67570132e-01 7.89000213e-01 5.46400070e-01
-1.42464709e+00 3.90779167e-01 -5.81474185e-01 -4.60081875e-01
4.07933682e-01 3.70099247e-01 -7.30621740e-02 -1.38049930e-01
-8.02974761e-01 6.44122422e-01 6.02489822e-02 2.66739279e-01
-9.37595904e-01 -5.92113793e-01 -9.34064031e-01 -1.56912386e-01
2.15717793e-01 -8.69396448e-01 8.28497529e-01 -1.03356540e+00
-1.50503993e+00 1.37591791e+00 -3.87433529e-01 -3.87579650e-01
1.23514034e-01 -3.18268895e-01 1.91092744e-01 4.02369618e-01
1.17240317e-01 8.08367074e-01 1.14481735e+00 -1.71854067e+00
-5.75864911e-01 -2.64289081e-01 1.25965655e-01 4.68447745e-01
1.14618316e-01 -3.07461582e-02 -7.77146637e-01 -8.96016240e-01
5.30691385e-01 -6.31783664e-01 -3.07582080e-01 3.16477299e-01
-2.18449563e-01 7.50998482e-02 8.16816688e-01 -8.13141048e-01
8.98680270e-01 -2.12661886e+00 2.80608684e-01 6.16511442e-02
1.90014452e-01 -4.55499776e-02 -1.14928201e-01 1.72501326e-01
2.17934642e-02 -2.71460652e-01 -6.40048862e-01 -4.82411474e-01
1.19124986e-01 2.75690705e-01 -2.66457617e-01 5.17882466e-01
1.20011799e-01 5.22508204e-01 -8.02551508e-01 -3.68759841e-01
6.89158738e-01 7.31344104e-01 -7.31656373e-01 1.87541172e-01
-4.15727049e-01 4.45400685e-01 -3.78775328e-01 6.30930305e-01
9.60146606e-01 -1.95589676e-01 3.83128524e-02 -7.94759393e-01
-2.29676291e-01 1.61834568e-01 -1.36046052e+00 1.92159247e+00
-4.81688559e-01 7.31190026e-01 5.82274020e-01 -1.26666772e+00
1.04686940e+00 -5.17590716e-02 5.88907957e-01 -6.77430034e-01
-1.16408154e-01 5.31986468e-02 -4.84750181e-01 -5.17684519e-01
4.94531482e-01 -7.25585371e-02 2.22776324e-01 2.15852372e-02
2.42162752e-03 -5.71575284e-01 -4.74168099e-02 -4.87745441e-02
6.86312377e-01 6.15788281e-01 2.59065449e-01 -7.98415959e-01
7.37424731e-01 -1.33920878e-01 7.38118291e-01 8.29451740e-01
-3.14768702e-01 1.08411586e+00 2.72812754e-01 -6.54077709e-01
-8.35371733e-01 -1.02400219e+00 -3.71718675e-01 9.90095139e-01
5.22839725e-01 -8.27216431e-02 -5.40555835e-01 -3.21484953e-01
-2.18833864e-01 5.37908137e-01 -6.54451549e-01 3.41222644e-01
-5.88265777e-01 -8.25581312e-01 -2.57466063e-02 4.47776049e-01
9.63761866e-01 -9.01749194e-01 -8.12351108e-01 1.71255603e-01
-8.58573020e-02 -1.27137637e+00 -3.61585885e-01 3.69440883e-01
-1.02109945e+00 -8.80479038e-01 -6.53899014e-01 -1.04956830e+00
6.84379756e-01 6.06602073e-01 1.51849782e+00 2.99589839e-02
-5.59148908e-01 9.20699120e-01 -1.86184749e-01 -4.64257859e-02
9.91404131e-02 -4.51861501e-01 -2.33888283e-01 9.18146893e-02
2.31597841e-01 -7.21434772e-01 -1.05019510e+00 3.83884698e-01
-7.97873795e-01 3.15502465e-01 4.28441346e-01 8.52754235e-01
9.61818635e-01 -1.04861660e-02 5.07423691e-02 -7.21068680e-01
2.14994907e-01 -2.10001126e-01 -6.76431894e-01 1.10718206e-01
-3.25962096e-01 6.78008348e-02 2.62993366e-01 2.27390587e-01
-1.68889439e+00 3.07826221e-01 -1.34988323e-01 -3.78382295e-01
-3.73451293e-01 1.54632866e-01 -3.63699764e-01 -2.29915053e-01
6.01094484e-01 2.99394518e-01 -1.60863623e-01 -7.69843400e-01
4.59906459e-01 1.91395253e-01 5.06418169e-01 -9.88861203e-01
4.40410167e-01 1.00891685e+00 2.11657539e-01 -1.18286693e+00
-1.11650646e+00 -5.63149631e-01 -7.23357499e-01 -2.16208950e-01
1.08776402e+00 -1.02956510e+00 -4.55458522e-01 5.27619958e-01
-1.03907442e+00 -5.80237567e-01 -4.85181421e-01 2.53729701e-01
-7.40418136e-01 6.76182628e-01 -4.57333416e-01 -5.52930474e-01
-1.05842657e-01 -1.31382811e+00 1.40273201e+00 1.81325302e-01
2.38194801e-02 -1.27883387e+00 5.04616424e-02 5.78582764e-01
2.81282425e-01 3.04268867e-01 7.76081383e-01 1.06544271e-01
-8.91217947e-01 4.92195576e-01 -6.44291818e-01 4.67507571e-01
1.50176972e-01 -1.46861568e-01 -1.27354693e+00 -1.87738195e-01
3.20004791e-01 -2.07828462e-01 1.10505867e+00 1.00662291e+00
1.35656130e+00 1.49951242e-02 -2.08981782e-01 9.43547666e-01
1.93826783e+00 -3.62258181e-02 5.47487020e-01 3.52749616e-01
5.98641098e-01 8.50731671e-01 5.26650369e-01 5.05356073e-01
2.47313470e-01 6.35683477e-01 6.73642695e-01 -3.85104328e-01
-8.33848774e-01 8.81897733e-02 1.33182481e-01 4.14374977e-01
-1.25670925e-01 -1.35832548e-01 -7.50770509e-01 7.45882690e-01
-1.78035915e+00 -8.20122421e-01 -2.49328777e-01 2.02445054e+00
6.77703917e-01 -4.82440181e-03 -1.90957516e-01 -1.75082430e-01
6.77869737e-01 3.92636091e-01 -5.65137029e-01 -3.97171885e-01
-4.81525719e-01 3.29187721e-01 3.98584932e-01 8.92835855e-01
-1.10264051e+00 9.12585318e-01 6.82624292e+00 8.14453125e-01
-8.92937005e-01 5.35025783e-02 8.01170051e-01 9.28794146e-02
-5.45347571e-01 9.35839713e-02 -7.25570083e-01 2.00178683e-01
5.04692318e-03 6.26476645e-01 2.37493783e-01 6.86068714e-01
3.35882843e-01 -5.12499571e-01 -8.59470367e-01 1.08936727e+00
1.83904469e-01 -1.41729152e+00 -9.32333022e-02 -1.43405870e-01
1.00894868e+00 1.58878505e-01 7.04297125e-02 -2.93802857e-01
3.53149712e-01 -8.58495653e-01 8.05917382e-01 6.09032035e-01
5.01685381e-01 -4.19662982e-01 2.60536730e-01 1.65673497e-03
-1.41420257e+00 -2.72320881e-02 -5.43310881e-01 -1.69835836e-01
5.47864325e-02 9.53727305e-01 -2.54913270e-01 4.91630524e-01
9.76527870e-01 9.62226748e-01 -3.64637583e-01 8.28343332e-01
-2.89136052e-01 4.88150328e-01 -5.19360483e-01 3.48879367e-01
2.44571522e-01 -5.13221443e-01 7.52943277e-01 1.40785217e+00
-6.58846647e-02 3.07040632e-01 3.13163608e-01 9.27735507e-01
2.08427444e-01 -1.32147983e-01 -5.26970923e-01 6.79409802e-01
9.76811871e-02 1.05863178e+00 -1.00348616e+00 -2.03696206e-01
-6.45151794e-01 1.15661991e+00 1.32169336e-01 5.57604849e-01
-4.85291153e-01 -9.97206196e-02 7.70645559e-01 -1.46647299e-05
5.72300553e-01 -4.46463972e-01 -7.86793530e-01 -1.20962036e+00
2.71747913e-03 -4.68795031e-01 4.45290387e-01 -9.08724546e-01
-1.35506642e+00 3.93147737e-01 1.60793453e-01 -1.08830106e+00
4.22485173e-01 -8.96989048e-01 -4.93929595e-01 7.36475050e-01
-1.99150538e+00 -1.26482999e+00 -5.30655146e-01 6.24216974e-01
9.39387739e-01 8.69663581e-02 5.25125265e-01 6.18634559e-02
-3.42400163e-01 1.17605045e-01 1.88287169e-01 -1.53994083e-01
3.85435253e-01 -1.25388539e+00 1.51496619e-01 1.02881098e+00
3.51780504e-02 4.94481146e-01 7.23762393e-01 -5.11252642e-01
-1.24031830e+00 -8.35253954e-01 5.68039298e-01 -2.09982559e-01
1.94094002e-01 -2.27348343e-01 -7.29672313e-01 4.11074370e-01
5.14686406e-01 1.67683348e-01 5.06377459e-01 1.44907488e-02
-8.85205269e-02 -1.96240753e-01 -1.28977835e+00 5.56973875e-01
1.30936480e+00 -4.16691393e-01 -6.88753486e-01 3.37129205e-01
6.79566443e-01 -2.80075967e-01 -6.87667608e-01 5.26616752e-01
3.58781040e-01 -1.55004835e+00 1.53113687e+00 1.24335736e-01
4.37178463e-01 -5.71691155e-01 -6.82962596e-01 -7.35904574e-01
-5.00103772e-01 -5.12359142e-01 3.34122665e-02 1.00117886e+00
-1.32955968e-01 -4.23262537e-01 7.77576327e-01 6.81397200e-01
-3.63029957e-01 -6.31632924e-01 -7.15709984e-01 -5.66313565e-01
-2.13344440e-01 -6.36528373e-01 1.11660942e-01 7.51601279e-01
-6.37355983e-01 1.10184550e-01 -3.65364432e-01 4.33340967e-01
1.29320216e+00 5.78612506e-01 5.03580213e-01 -1.05762053e+00
-3.02218825e-01 -4.65506166e-01 -2.14294821e-01 -1.51459396e+00
7.72773847e-02 -7.39197791e-01 8.45014900e-02 -1.61877894e+00
1.25424564e-01 -5.26863635e-01 -2.53506660e-01 2.38293886e-01
-7.30002448e-02 5.52833140e-01 -2.55744439e-02 1.92460090e-01
-3.03648382e-01 6.36439383e-01 1.26996231e+00 1.74123999e-02
-2.20637605e-01 -3.47649753e-01 -5.24002969e-01 1.11509788e+00
5.83333433e-01 -1.09918714e-01 -6.35448396e-01 -7.59196341e-01
1.11443721e-01 -3.14876229e-01 6.36766076e-01 -6.97180688e-01
1.57623067e-01 -3.31831872e-01 4.46851343e-01 -7.71872640e-01
6.25546277e-01 -8.99033666e-01 -9.41124931e-02 2.52398372e-01
-1.43520460e-01 -5.73201001e-01 2.64077008e-01 7.02200294e-01
-3.00761998e-01 -3.30309600e-01 1.14854097e+00 -4.97935116e-01
-1.28000867e+00 2.00570792e-01 -3.30664694e-01 1.09542988e-01
6.86601341e-01 -7.06270576e-01 3.31908196e-01 -1.20989680e-01
-9.11643028e-01 1.61539968e-02 5.85582197e-01 6.23979904e-02
8.73570025e-01 -1.00630724e+00 -6.00228012e-01 2.87515134e-01
-1.24480359e-01 2.44624808e-01 2.52669990e-01 6.33449733e-01
-6.67294800e-01 2.68932551e-01 -2.06638813e-01 -1.06077135e+00
-1.15909398e+00 3.73070151e-01 3.69169921e-01 1.11278314e-02
-8.81326556e-01 1.22579014e+00 9.16360199e-01 -3.41556877e-01
4.54567790e-01 -4.54813004e-01 -4.00912240e-02 -3.64445709e-02
1.94846317e-01 2.27540061e-01 -1.59903139e-01 -5.45163929e-01
-4.28460628e-01 1.17778647e+00 1.83842137e-01 3.14040668e-02
1.60813200e+00 -6.61965609e-01 -3.19359481e-01 1.15829065e-01
9.57477748e-01 -1.26553020e-02 -1.85150599e+00 -4.29405272e-01
-1.47552118e-01 -8.32411170e-01 5.27580321e-01 -7.79842377e-01
-1.21272933e+00 9.00626123e-01 6.90847754e-01 8.30422807e-03
1.75421274e+00 -5.22673130e-02 5.23620665e-01 1.44398853e-01
2.85983205e-01 -1.17776906e+00 2.95919895e-01 4.33813542e-01
1.01482844e+00 -1.43617928e+00 3.07260424e-01 -6.73110604e-01
-4.93629903e-01 1.10220504e+00 5.03690124e-01 -4.08672422e-01
9.30896461e-01 2.37915292e-01 -1.06678814e-01 -4.56053704e-01
-4.68379021e-01 -4.93590176e-01 4.44365382e-01 7.66430140e-01
4.21965718e-01 -4.31704253e-01 -2.28587389e-01 2.39376888e-01
3.60691100e-01 -3.19854409e-01 1.72484875e-01 7.69755304e-01
-6.78022385e-01 -1.01966083e+00 -5.39284110e-01 -1.31868301e-02
-2.43630946e-01 -2.82600731e-01 -1.53962865e-01 5.96287906e-01
3.46188933e-01 8.01533937e-01 -9.14742798e-02 1.49785042e-01
1.28024831e-01 -1.72282472e-01 7.48364270e-01 -6.61452889e-01
-4.23247129e-01 4.38190520e-01 -1.60219073e-02 -7.55324960e-01
-1.02490234e+00 -7.24894941e-01 -1.13710248e+00 2.62511581e-01
-3.98989394e-02 -3.60010982e-01 5.95146298e-01 6.72628105e-01
4.05067474e-01 2.75938600e-01 5.80573499e-01 -1.16579497e+00
6.43278062e-02 -5.04412413e-01 -4.15259928e-01 4.01867568e-01
5.04899561e-01 -7.74131894e-01 -2.86198586e-01 5.84969163e-01] | [9.588708877563477, -2.7665090560913086] |
7c01f305-6d2c-471a-9f51-4c2167312b60 | continuous-adaptation-of-multi-camera-person | 1607.00417 | null | http://arxiv.org/abs/1607.00417v1 | http://arxiv.org/pdf/1607.00417v1.pdf | Continuous Adaptation of Multi-Camera Person Identification Models through Sparse Non-redundant Representative Selection | The problem of image-base person identification/recognition is to provide an
identity to the image of an individual based on learned models that describe
his/her appearance. Most traditional person identification systems rely on
learning a static model on tediously labeled training data. Though labeling
manually is an indispensable part of a supervised framework, for a large scale
identification system labeling huge amount of data is a significant overhead.
For large multi-sensor data as typically encountered in camera networks,
labeling a lot of samples does not always mean more information, as redundant
images are labeled several times. In this work, we propose a convex
optimization based iterative framework that progressively and judiciously
chooses a sparse but informative set of samples for labeling, with minimal
overlap with previously labeled images. We also use a structure preserving
sparse reconstruction based classifier to reduce the training burden typically
seen in discriminative classifiers. The two stage approach leads to a novel
framework for online update of the classifiers involving only the incorporation
of new labeled data rather than any expensive training phase. We demonstrate
the effectiveness of our approach on multi-camera person re-identification
datasets, to demonstrate the feasibility of learning online classification
models in multi-camera big data applications. Using three benchmark datasets,
we validate our approach and demonstrate that our framework achieves superior
performance with significantly less amount of manual labeling. | ['Amit K. Roy-Chowdhury', 'Abir Das', 'Rameswar Panda'] | 2016-07-01 | null | null | null | null | ['person-identification'] | ['computer-vision'] | [ 3.82238835e-01 -4.07158256e-01 -1.93879813e-01 -7.52570868e-01
-6.59973502e-01 -6.09891236e-01 2.52714366e-01 2.49290988e-01
-6.18306339e-01 6.43081546e-01 -8.42135325e-02 1.92344338e-01
6.61744103e-02 -5.27884483e-01 -6.09268606e-01 -6.32556379e-01
3.13633859e-01 7.78520644e-01 -1.59799412e-01 2.03667700e-01
2.30836943e-02 4.80116248e-01 -1.67556798e+00 4.74680886e-02
6.07886374e-01 9.82073605e-01 1.06822707e-01 4.82460797e-01
5.35538979e-02 5.40641785e-01 -4.38289940e-01 -6.18451297e-01
7.18667388e-01 -1.77422449e-01 -7.29784191e-01 7.99509287e-01
8.72917295e-01 -4.20276284e-01 -6.92067519e-02 1.22897601e+00
5.13974965e-01 1.88346222e-01 3.78747582e-01 -1.13604772e+00
-4.45802122e-01 3.12211096e-01 -7.65391290e-01 7.05840588e-02
5.26863217e-01 -1.01355962e-01 7.28072524e-01 -7.95555592e-01
3.39167774e-01 1.06357789e+00 9.99352932e-01 5.89463770e-01
-1.27588058e+00 -6.52281165e-01 3.49679142e-01 2.63435066e-01
-1.72077727e+00 -6.76138997e-01 7.73282468e-01 -3.96103233e-01
3.32364440e-01 3.35571170e-01 7.64834940e-01 7.66428888e-01
-5.54716945e-01 6.18001342e-01 1.27133763e+00 -4.79026198e-01
1.67543471e-01 6.84891880e-01 5.41142523e-01 7.00635314e-01
4.59867746e-01 -7.52387866e-02 -6.23283088e-01 -2.41225898e-01
4.83298302e-01 5.61944485e-01 -1.02152452e-01 -3.77172410e-01
-8.77401650e-01 5.47967792e-01 1.05211936e-01 9.52953324e-02
-4.10411090e-01 4.97304201e-02 2.06158668e-01 1.96905494e-01
2.65939683e-01 -3.79603244e-02 -2.77757138e-01 -1.07759885e-01
-1.17141092e+00 2.35817172e-02 6.99114740e-01 9.03657854e-01
1.21852911e+00 -2.48158023e-01 9.69448388e-02 1.13094640e+00
1.91863418e-01 6.43539369e-01 5.12243211e-01 -7.05647588e-01
2.36983836e-01 9.13483143e-01 3.02326709e-01 -1.09965158e+00
-1.84947178e-01 -4.89022017e-01 -9.59025025e-01 -2.05148652e-01
4.00415093e-01 -1.42011285e-01 -7.17220485e-01 1.52977085e+00
4.98246819e-01 5.42140663e-01 -9.99626964e-02 9.26281631e-01
3.00664216e-01 3.00254256e-01 9.60755162e-03 -2.48238370e-01
1.37459254e+00 -8.50875735e-01 -5.19245863e-01 -4.39983398e-01
4.48613703e-01 -5.44890046e-01 7.45491087e-01 3.23480636e-01
-6.05836928e-01 -6.29334211e-01 -8.14335167e-01 1.86537728e-01
-3.59901637e-01 5.03087580e-01 7.56098807e-01 1.10625196e+00
-8.69258463e-01 5.36727846e-01 -4.79482353e-01 -6.27450466e-01
4.74956214e-01 7.39261687e-01 -6.80570841e-01 -4.02079284e-01
-5.91287673e-01 5.28329253e-01 2.90736496e-01 3.23648512e-01
-7.04817295e-01 -5.49797773e-01 -7.78072000e-01 -1.24590158e-01
3.32948864e-01 -4.76723254e-01 8.97112787e-01 -1.32054818e+00
-1.24426341e+00 9.88779962e-01 -5.73382914e-01 -4.69966084e-01
4.69163597e-01 -2.65534401e-01 -4.09036130e-01 4.03235890e-02
1.42302871e-01 6.48779809e-01 1.12509573e+00 -1.48956680e+00
-7.80077517e-01 -6.35425687e-01 7.13284239e-02 2.26544201e-01
-7.86576331e-01 3.33119072e-02 -6.08728945e-01 -5.03624916e-01
9.85382870e-02 -1.19987631e+00 -2.98356771e-01 -8.34968388e-02
-2.17459336e-01 -3.31463548e-03 8.81801963e-01 -8.77478600e-01
8.69614005e-01 -1.99634635e+00 -8.19421257e-04 4.42275733e-01
1.11814693e-01 1.66827008e-01 1.18560933e-01 1.19206823e-01
6.29562512e-02 -2.09265664e-01 -2.92605937e-01 -9.30948436e-01
-1.67397156e-01 1.06728829e-01 -3.22136618e-02 6.20122850e-01
-2.38967136e-01 6.50141299e-01 -7.59102702e-01 -5.99235594e-01
3.85674268e-01 5.32374680e-01 -4.21620637e-01 2.21511483e-01
3.03058088e-01 5.65484941e-01 -1.01129755e-01 1.01438808e+00
6.80714607e-01 -4.73887622e-01 2.38486752e-01 -4.18405205e-01
9.73926336e-02 -5.99594474e-01 -1.75146532e+00 1.70192409e+00
-3.75211507e-01 2.70693630e-01 1.60702810e-01 -1.28559053e+00
7.71874189e-01 2.83281118e-01 7.31244028e-01 -1.97120309e-01
1.26919329e-01 7.19070286e-02 -4.47477192e-01 -2.19009981e-01
4.84166712e-01 5.25594950e-02 -3.41563746e-02 6.17303014e-01
3.48583236e-02 5.56130946e-01 1.56380609e-01 4.11576852e-02
7.43622482e-01 -1.63637787e-01 2.94212401e-01 1.10397977e-03
7.73686707e-01 4.39152382e-02 7.90389836e-01 8.69950891e-01
-2.70421654e-01 4.71541673e-01 -4.17661846e-01 -7.00786769e-01
-1.10185945e+00 -4.47843283e-01 -7.88986906e-02 9.91721511e-01
3.36254299e-01 -3.90526772e-01 -7.68517673e-01 -6.71003759e-01
4.67871800e-02 -5.64202340e-03 -4.91739303e-01 1.34584904e-01
-4.57372993e-01 -1.03306723e+00 2.78817117e-01 3.59698057e-01
7.77157247e-01 -5.50020456e-01 -3.85973722e-01 4.77497168e-02
-2.48553809e-02 -1.29262924e+00 -8.65420520e-01 -1.70459837e-01
-6.24517024e-01 -1.06223691e+00 -7.08388448e-01 -9.54509079e-01
1.26319766e+00 6.92277730e-01 7.08335221e-01 2.58782774e-01
-4.97122884e-01 8.42199981e-01 -2.40011066e-01 -5.61045259e-02
5.79353198e-02 -1.48593411e-01 4.65100825e-01 8.58025014e-01
5.48602164e-01 -5.04623294e-01 -6.47162199e-01 4.33069617e-01
-5.34501910e-01 1.03480062e-02 5.53903580e-01 1.00340652e+00
7.07235098e-01 2.70828545e-01 3.97696614e-01 -1.02138102e+00
2.92667031e-01 -2.33555928e-01 -6.93820715e-01 5.27689099e-01
-6.88216150e-01 -1.19063050e-01 6.91645682e-01 -6.71176314e-01
-9.07554150e-01 8.68625104e-01 1.96001217e-01 -4.89726603e-01
-1.91256061e-01 2.52202243e-01 -1.53156072e-01 -5.92992783e-01
2.37785459e-01 5.06455660e-01 -1.42480105e-01 -6.98487163e-01
2.24786878e-01 9.27419126e-01 6.57007217e-01 -6.92687988e-01
1.00438774e+00 7.37565875e-01 -1.15345657e-01 -8.67125511e-01
-8.26808691e-01 -9.40710485e-01 -1.04744756e+00 -3.40283722e-01
5.24136662e-01 -1.20504117e+00 -1.00612891e+00 4.50099051e-01
-8.31372976e-01 1.37009233e-01 -3.01789820e-01 3.11654240e-01
-2.51179766e-02 8.20522964e-01 -3.12501580e-01 -9.81022358e-01
-3.95774603e-01 -1.04191124e+00 1.18887401e+00 4.31031674e-01
-5.99813722e-02 -9.16388512e-01 -2.81787720e-02 8.27062249e-01
2.25028604e-01 -3.94032039e-02 1.50099859e-01 -6.42923355e-01
-6.22270465e-01 -8.12352419e-01 -2.18312576e-01 3.34200561e-01
3.99237305e-01 -5.73577642e-01 -1.16570079e+00 -7.71913826e-01
-1.96470171e-01 -4.09650534e-01 6.38093591e-01 6.88334256e-02
1.26852369e+00 -2.81673372e-01 -4.30348545e-01 7.26440310e-01
1.58698118e+00 3.80690508e-02 5.76223284e-02 1.02575459e-01
1.14426386e+00 5.15036404e-01 4.45256382e-01 7.60624349e-01
6.03442907e-01 8.31826866e-01 -7.12852329e-02 7.13004591e-03
7.66492113e-02 -3.41542244e-01 2.49242932e-01 7.37680733e-01
-9.72276106e-02 1.78591713e-01 -7.71371484e-01 4.99184430e-01
-1.95182312e+00 -1.00853097e+00 3.33147794e-01 2.56196618e+00
7.53299057e-01 -3.49250019e-01 3.46760958e-01 2.46831775e-01
1.02081788e+00 -2.26742938e-01 -6.24147475e-01 3.04875314e-01
-8.38329643e-02 3.76400584e-03 8.21939230e-01 5.76304495e-01
-1.31936908e+00 8.28097880e-01 5.86614132e+00 5.92386544e-01
-9.94566441e-01 2.14154616e-01 6.62204504e-01 -6.23853467e-02
2.11660445e-01 9.48760062e-02 -1.27974594e+00 6.42760158e-01
8.64308894e-01 3.85382064e-02 6.39365315e-01 9.14216280e-01
1.43647939e-01 -1.63566023e-01 -1.34052503e+00 1.81686699e+00
3.65019709e-01 -1.23179317e+00 7.80465407e-03 1.73331290e-01
7.18278885e-01 -4.85400319e-01 -1.18969850e-01 -1.25028059e-01
2.45164320e-01 -9.27348077e-01 6.12988174e-01 3.90315920e-01
6.67948902e-01 -6.14386916e-01 8.02412391e-01 3.94696385e-01
-1.43539822e+00 -3.24134409e-01 -4.64780897e-01 5.20603545e-02
6.63700700e-02 5.12082875e-01 -7.01318204e-01 3.78094882e-01
7.76204348e-01 9.87459600e-01 -8.20042789e-01 1.07007337e+00
2.38749236e-01 4.09545630e-01 -4.35603887e-01 2.51920432e-01
-2.05711782e-01 -3.24969411e-01 1.79975152e-01 1.02945113e+00
1.50395945e-01 1.74993664e-01 7.75959849e-01 3.63535821e-01
-1.47082016e-01 1.78656265e-01 -4.55623537e-01 1.81980222e-01
7.25973129e-01 1.41030014e+00 -6.16935015e-01 -4.36817229e-01
-6.49533272e-01 1.34336948e+00 3.33948761e-01 9.31680948e-02
-5.92817545e-01 2.40416184e-01 3.45467597e-01 4.43200059e-02
1.07097924e-01 -1.59242153e-01 -8.05464983e-02 -1.38881791e+00
2.02089638e-01 -1.02273965e+00 5.69011033e-01 -1.66146591e-01
-1.36140227e+00 4.37880367e-01 -1.96290046e-01 -1.30907369e+00
-8.66559520e-02 -3.34246427e-01 -1.57811239e-01 7.33002424e-01
-1.33630252e+00 -1.51348078e+00 -7.52747536e-01 9.03251171e-01
5.13767183e-01 -3.96890849e-01 9.40903902e-01 8.08635652e-01
-9.18132365e-01 8.72368932e-01 1.48922235e-01 4.26076204e-01
7.28212774e-01 -1.15333021e+00 -1.48671359e-01 1.01247931e+00
3.50903034e-01 7.24405348e-01 5.31069577e-01 -6.25790656e-01
-1.70184636e+00 -1.13101590e+00 6.43777668e-01 -2.93330640e-01
1.87817365e-01 -3.35182875e-01 -6.23551369e-01 8.30432475e-01
-1.19402736e-01 2.41763011e-01 1.07284808e+00 2.32596964e-01
-2.02384248e-01 -6.20709598e-01 -1.30304241e+00 1.90686747e-01
9.19150114e-01 -6.17181420e-01 -1.90493345e-01 5.92795610e-01
8.28238279e-02 -5.34693040e-02 -9.30484354e-01 6.39313608e-02
6.35005295e-01 -6.59517944e-01 1.17091107e+00 -2.32232779e-01
-4.26524341e-01 -6.04889035e-01 -2.44980067e-01 -9.38944399e-01
-2.20627487e-01 -5.18785298e-01 3.02047580e-02 1.47370958e+00
-1.36838034e-02 -5.33480585e-01 1.28621757e+00 1.06532121e+00
3.44671667e-01 -1.69604987e-01 -8.68358612e-01 -7.75446355e-01
-6.64989293e-01 -1.64229870e-01 5.55413723e-01 1.03604829e+00
-1.98225498e-01 2.19915375e-01 -1.01117754e+00 5.30827761e-01
1.29416227e+00 2.25529447e-01 9.68419731e-01 -1.42974365e+00
-4.68067110e-01 2.30868652e-01 -6.07591152e-01 -9.69114542e-01
2.25067750e-01 -8.26361179e-01 -2.11946607e-01 -1.03570724e+00
6.91169739e-01 -7.70588577e-01 -3.03531766e-01 6.68782413e-01
-2.10615799e-01 5.55916011e-01 2.13857010e-01 5.57838202e-01
-7.98194408e-01 1.68101728e-01 5.32294989e-01 -5.01413643e-01
-1.78746000e-01 5.13018258e-02 -7.40736902e-01 6.08226836e-01
4.55532879e-01 -4.11053330e-01 -4.46782053e-01 -4.50226307e-01
-2.50955284e-01 -2.63801575e-01 5.33896208e-01 -1.21433604e+00
5.70206881e-01 5.10153547e-03 6.56220675e-01 -4.42276984e-01
6.87436044e-01 -1.25894499e+00 4.60783273e-01 3.44840020e-01
-9.32994559e-02 1.13354633e-02 -1.18748881e-01 6.84745550e-01
-2.49272332e-01 -2.85082012e-01 9.56673324e-01 -3.39089602e-01
-1.10386503e+00 5.97592056e-01 1.17117606e-01 -4.00658488e-01
1.12127924e+00 -5.08758187e-01 2.02893853e-01 -2.64405668e-01
-8.22831869e-01 1.84534788e-01 7.25040197e-01 2.80908138e-01
5.58007419e-01 -1.32722390e+00 -4.94300604e-01 2.99488991e-01
1.67847738e-01 -2.01156169e-01 3.04059774e-01 5.18617153e-01
-2.71529049e-01 3.64130795e-01 -8.67512301e-02 -7.56569743e-01
-1.71146011e+00 5.69840789e-01 3.43440115e-01 -1.13915898e-01
-5.42310953e-01 7.94114947e-01 -1.90707922e-01 -3.54687512e-01
3.39904726e-01 2.76512861e-01 -1.48980662e-01 9.17491317e-02
7.09504485e-01 2.89766461e-01 -1.13082804e-01 -9.41090286e-01
-4.69598264e-01 8.08085322e-01 -2.79194057e-01 5.53283766e-02
1.32705903e+00 -3.82585257e-01 -1.58436552e-01 3.65086049e-01
1.07435608e+00 -9.78294015e-03 -1.28725171e+00 -4.85214651e-01
1.11909099e-01 -8.72332036e-01 -3.09875011e-02 -4.85610604e-01
-1.14285159e+00 5.46747029e-01 1.07591248e+00 -1.82745099e-01
1.11878526e+00 -1.92334652e-01 8.32538724e-01 5.99349380e-01
7.58409977e-01 -1.36668098e+00 -4.80593443e-02 -7.44811594e-02
3.95892113e-01 -1.75159061e+00 2.17295095e-01 -4.60342109e-01
-5.75382411e-01 8.06062579e-01 5.27803838e-01 6.34260625e-02
7.56829381e-01 -7.06300279e-03 -8.43769237e-02 8.73954296e-02
-2.31627092e-01 -1.84119314e-01 5.23053668e-02 5.69089353e-01
-3.14138122e-02 9.74546075e-02 7.20140785e-02 4.48376119e-01
-4.77052992e-03 1.82337426e-02 2.21793830e-01 9.95414615e-01
-3.12120944e-01 -1.45838034e+00 -7.15262175e-01 4.39716548e-01
-3.11407924e-01 5.58594149e-03 -2.95596659e-01 2.98515201e-01
4.31396544e-01 9.74303067e-01 -3.60223167e-02 -4.40884560e-01
4.64245416e-02 1.46083653e-01 5.02423942e-01 -6.63205802e-01
-5.12034118e-01 -7.44491890e-02 -9.01250318e-02 -2.61877835e-01
-8.58345032e-01 -9.70774889e-01 -7.59261906e-01 -4.77724165e-01
-2.22297654e-01 1.24780864e-01 5.88231564e-01 9.94587600e-01
3.23727012e-01 -6.87378794e-02 8.45740736e-01 -8.61132085e-01
-4.83323127e-01 -7.84611404e-01 -6.61241174e-01 7.27461159e-01
2.62049645e-01 -6.52328193e-01 -5.87451831e-02 6.34637177e-01] | [14.739654541015625, 1.0266761779785156] |
36a820ba-7f03-40b5-a0ae-dfe3ec1f8f2a | learning-ergodic-averages-in-chaotic-systems | 2001.04027 | null | https://arxiv.org/abs/2001.04027v2 | https://arxiv.org/pdf/2001.04027v2.pdf | Learning ergodic averages in chaotic systems | We propose a physics-informed machine learning method to predict the time average of a chaotic attractor. The method is based on the hybrid echo state network (hESN). We assume that the system is ergodic, so the time average is equal to the ergodic average. Compared to conventional echo state networks (ESN) (purely data-driven), the hESN uses additional information from an incomplete, or imperfect, physical model. We evaluate the performance of the hESN and compare it to that of an ESN. This approach is demonstrated on a chaotic time-delayed thermoacoustic system, where the inclusion of a physical model significantly improves the accuracy of the prediction, reducing the relative error from 48% to 7%. This improvement is obtained at the low extra cost of solving two ordinary differential equations. This framework shows the potential of using machine learning techniques combined with prior physical knowledge to improve the prediction of time-averaged quantities in chaotic systems. | ['Francisco Huhn', 'Luca Magri'] | 2020-01-09 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [ 1.09683469e-01 5.20121939e-02 4.68681455e-01 1.46105438e-01
-5.14431775e-01 -3.54650170e-01 9.34706509e-01 9.57970843e-02
-5.95397234e-01 7.68956900e-01 -1.27777785e-01 -4.23001975e-01
-2.04969302e-01 -6.89635277e-01 -2.91194081e-01 -1.24840367e+00
-4.21604127e-01 5.01797736e-01 2.85032660e-01 -3.85494113e-01
5.09931326e-01 7.32388794e-01 -1.37816036e+00 -5.02126515e-01
6.96856558e-01 9.76252854e-01 6.59089014e-02 1.09033525e+00
6.60760403e-02 7.53916264e-01 -5.43126822e-01 2.85237730e-01
2.13410228e-01 -5.75353742e-01 -4.73233908e-01 -5.91187179e-01
-1.08840764e-01 -1.81870148e-01 -8.07545900e-01 8.17252457e-01
5.20993531e-01 3.61557215e-01 8.35765541e-01 -8.81069362e-01
-1.12468516e-02 5.18808901e-01 8.52202550e-02 2.85521895e-01
-8.39708149e-02 2.23055825e-01 6.39030755e-01 -4.43225265e-01
2.32948408e-01 7.85515726e-01 1.11115026e+00 4.14415985e-01
-1.45196915e+00 -4.21979725e-01 -6.87518299e-01 1.53910160e-01
-1.23503888e+00 -4.72890824e-01 9.33694005e-01 -3.26754093e-01
9.87051368e-01 1.27927303e-01 8.18809748e-01 5.47651768e-01
6.74302757e-01 1.78265184e-01 1.64882481e+00 -7.98219383e-01
6.11550152e-01 8.15960839e-02 1.95059001e-01 6.59609199e-01
2.92060256e-01 8.14361513e-01 -3.31156909e-01 -3.56499791e-01
5.19568145e-01 -5.69221258e-01 -1.43946350e-01 -1.84209287e-01
-1.03901660e+00 6.43943727e-01 1.62226394e-01 6.08433843e-01
-4.18354064e-01 3.36308211e-01 1.38717428e-01 3.73296678e-01
4.94764537e-01 8.77791047e-01 -3.37058365e-01 -3.63056153e-01
-1.32633257e+00 3.02294213e-02 1.60530412e+00 1.72563493e-01
7.87376642e-01 4.10518438e-01 3.74701291e-01 2.18455449e-01
1.31881252e-01 1.20858777e+00 5.45520008e-01 -1.23098540e+00
-2.90393293e-01 5.59863783e-02 4.78617102e-01 -8.41427922e-01
-5.52672565e-01 -2.86686271e-01 -8.94752920e-01 3.46356362e-01
4.85446572e-01 -4.64706391e-01 -8.34425509e-01 1.37124360e+00
-2.23158579e-02 7.25246847e-01 5.21374047e-01 5.53240001e-01
5.17829001e-01 1.05693638e+00 -3.06669623e-01 -4.47180688e-01
5.92860937e-01 -4.16656852e-01 -8.81626308e-01 1.73095688e-01
6.77249014e-01 -6.15862310e-01 3.47683102e-01 3.25038761e-01
-8.62758934e-01 -3.60468328e-01 -1.20438206e+00 5.18268704e-01
-5.11492908e-01 -2.37746805e-01 3.07737619e-01 7.88340986e-01
-1.24391079e+00 1.38453448e+00 -1.15041411e+00 -4.23902869e-02
-5.56199968e-01 6.12179875e-01 -5.51692173e-02 5.33990860e-01
-1.36834264e+00 1.21080589e+00 4.84326988e-01 1.45806968e-01
-6.63563848e-01 -6.98534369e-01 -7.59592235e-01 1.14188969e-01
-1.10772714e-01 -1.80321768e-01 1.28397548e+00 -4.04433250e-01
-2.08656526e+00 1.27811804e-01 -1.50488645e-01 -6.63559079e-01
4.75166798e-01 1.85397476e-01 -5.30899048e-01 4.27983731e-01
-3.97748083e-01 1.77087128e-01 6.40773475e-01 -1.24044108e+00
-3.03663760e-02 4.66586679e-01 -3.08585167e-01 -2.55242229e-01
-1.04680270e-01 -6.10154271e-01 2.91155279e-01 -1.31973922e-01
1.33881718e-01 -1.37409925e+00 -3.53625536e-01 -3.98092985e-01
-1.48045138e-01 -3.02708149e-02 1.05210435e+00 -6.54133201e-01
9.68093097e-01 -1.82670605e+00 -9.02131200e-03 4.68461365e-01
1.06232248e-01 5.52704096e-01 3.40313241e-02 9.63827670e-01
7.65150785e-02 -3.18709761e-02 -6.01038814e-01 -2.83473015e-01
-1.12866752e-01 3.64101529e-01 -4.27926570e-01 4.58623290e-01
5.12901619e-02 7.19595671e-01 -8.82572055e-01 -3.26961845e-01
4.85722721e-01 5.02020597e-01 -5.22480309e-02 2.05548659e-01
2.10535508e-02 7.21906066e-01 -2.10117161e-01 -1.15053125e-01
6.20663702e-01 -2.33693287e-01 2.18933448e-02 7.19827786e-02
-5.60552180e-01 2.58828193e-01 -1.10628796e+00 1.08107996e+00
-7.31334031e-01 9.50973511e-01 -1.15374818e-01 -1.16132081e+00
1.28504026e+00 5.08471370e-01 7.72106111e-01 -6.61296785e-01
9.20329317e-02 5.96204102e-01 3.62384796e-01 -5.38596630e-01
5.27660429e-01 -5.61872721e-01 -9.02066156e-02 9.01761413e-01
1.60554618e-01 -5.96501827e-01 -1.01238571e-01 2.74939448e-01
1.10045862e+00 -1.48149366e-02 3.11361849e-01 -8.58372271e-01
6.66424930e-01 4.47892472e-02 1.97722226e-01 9.74399924e-01
-1.32715806e-01 -6.67614415e-02 5.01057327e-01 -4.51098114e-01
-1.39917672e+00 -6.71386957e-01 -4.15758878e-01 1.70082211e-01
2.97558874e-01 -2.95926124e-01 -7.14446127e-01 3.24741155e-02
-2.93370876e-02 7.89077759e-01 -6.68708503e-01 -3.67690086e-01
-8.89687479e-01 -8.46450806e-01 6.24523699e-01 7.24405572e-02
5.35414279e-01 -1.17518651e+00 -5.88092387e-01 5.01190484e-01
-8.01091269e-02 -1.09325886e+00 1.78669110e-01 3.10346395e-01
-9.63718235e-01 -6.95161760e-01 -6.29407883e-01 -1.59645736e-01
1.93675399e-01 -2.55216360e-01 7.75838971e-01 1.63340583e-01
-1.41761735e-01 6.56055391e-01 -2.14661285e-01 -3.58642846e-01
-1.23049140e+00 8.74373764e-02 5.40205181e-01 -1.16462007e-01
1.48768127e-01 -7.23766685e-01 -5.03581583e-01 3.40800248e-02
-7.34046996e-01 -2.56460965e-01 4.45838600e-01 8.99418890e-01
6.35013953e-02 2.51422197e-01 5.12663066e-01 -2.50159442e-01
4.06588197e-01 -3.70693862e-01 -9.49175894e-01 1.31130116e-02
-9.02758658e-01 6.21880651e-01 8.84241045e-01 -6.77966475e-01
-9.41078484e-01 -3.42678912e-02 7.12574050e-02 -2.53430724e-01
-4.44272086e-02 5.27072668e-01 6.49580121e-01 -6.89891160e-01
4.12967771e-01 6.75641537e-01 3.62048626e-01 -3.80340129e-01
-6.08000308e-02 5.54463446e-01 4.92810875e-01 -4.60855335e-01
8.09136093e-01 3.62127334e-01 8.22530389e-01 -1.31341910e+00
-4.08144325e-01 -3.11509013e-01 -9.73487854e-01 -4.05043781e-01
3.50871503e-01 -4.32781488e-01 -9.99687970e-01 8.27206433e-01
-8.89329791e-01 -6.79318666e-01 -4.28139925e-01 9.53224003e-01
-6.81457281e-01 5.04388869e-01 -7.57728398e-01 -1.22234750e+00
-2.67389625e-01 -7.39008665e-01 6.25933290e-01 1.66762605e-01
9.09912586e-02 -1.57710695e+00 7.70559192e-01 -4.14394975e-01
8.74356568e-01 2.94795722e-01 6.91914141e-01 -6.86859429e-01
-3.51098597e-01 -4.17971402e-01 1.62024930e-01 2.94841319e-01
-1.34812742e-01 1.84458569e-01 -1.09859669e+00 -3.96221250e-01
2.79637665e-01 -1.24082886e-01 9.64472711e-01 4.84402388e-01
2.69103467e-01 -2.14109138e-01 -3.22902054e-01 1.65370673e-01
1.56757939e+00 1.59766674e-01 3.23192835e-01 -3.37435380e-02
1.59425378e-01 5.20542443e-01 9.98467878e-02 4.55040067e-01
1.07427631e-02 3.90803516e-01 1.87576577e-01 1.24053918e-01
7.57797509e-02 7.88670406e-02 4.94627774e-01 1.47681367e+00
-2.09346890e-01 -8.66418034e-02 -1.19519079e+00 4.37903792e-01
-1.50221848e+00 -1.01730525e+00 -3.71626735e-01 2.19116330e+00
5.39046586e-01 -5.24034761e-02 -2.15680942e-01 2.99030006e-01
6.35803103e-01 1.37597993e-01 -5.06176293e-01 -6.32464707e-01
6.05445355e-02 3.43453497e-01 4.76488054e-01 8.50590587e-01
-8.94798279e-01 6.13986433e-01 7.57008505e+00 5.70839763e-01
-1.55379832e+00 -9.30917561e-02 1.14221565e-01 3.46617609e-01
1.45542100e-01 1.72540799e-01 -6.56059146e-01 4.57326949e-01
1.88815594e+00 -3.87861282e-01 5.49650073e-01 2.21708238e-01
3.91996920e-01 -4.07340974e-01 -6.08488619e-01 4.63129997e-01
-2.13842615e-01 -1.19626367e+00 -4.64443177e-01 2.76245713e-01
6.26106024e-01 2.69696563e-01 -1.27012670e-01 2.96151459e-01
1.45996943e-01 -7.04467833e-01 3.23718995e-01 1.27852297e+00
7.09602416e-01 -6.42686546e-01 1.09692311e+00 6.23319447e-01
-1.13669503e+00 5.02883308e-02 -2.92715400e-01 -3.63291413e-01
3.60072851e-01 6.69401407e-01 -7.30231822e-01 3.98063272e-01
4.31398988e-01 5.28271616e-01 -2.99532473e-01 1.12294841e+00
2.68155813e-01 1.01697087e+00 -8.88193846e-01 -5.18643737e-01
4.61119831e-01 -2.33723164e-01 9.90658104e-01 1.17652237e+00
6.31426513e-01 4.85709339e-01 -1.68564081e-01 7.91504860e-01
4.51720625e-01 -2.46651262e-01 -7.69819379e-01 -9.82387215e-02
3.01718712e-01 1.26680303e+00 -7.32216954e-01 -5.38001955e-01
7.83012509e-02 4.69576657e-01 -2.58936316e-01 1.89355984e-01
-4.70293462e-01 -7.00201750e-01 1.12708695e-02 -1.85626447e-01
4.36671406e-01 -5.02985656e-01 5.28672114e-02 -9.53415573e-01
-3.32540065e-01 -1.92851394e-01 -2.45498732e-01 -7.09885955e-01
-1.10642743e+00 7.46529996e-01 2.27821231e-01 -1.31085312e+00
-5.97294748e-01 -6.83535278e-01 -8.91731501e-01 9.73493218e-01
-1.43236434e+00 -5.32234728e-01 -8.73784944e-02 2.22398400e-01
-2.40990624e-01 -1.99515775e-01 1.08158576e+00 -2.20004693e-01
-2.99815357e-01 1.05878688e-01 8.25738370e-01 1.06488252e-02
2.82641768e-01 -1.46552122e+00 3.24998260e-01 5.86957157e-01
-2.26008400e-01 4.18017656e-01 1.27425170e+00 -6.20326400e-01
-1.21167529e+00 -6.60737634e-01 9.87533569e-01 -2.75210232e-01
1.04135525e+00 -2.56979346e-01 -1.18051326e+00 2.83330772e-02
3.59579861e-01 2.09700078e-01 4.08730000e-01 -2.91626096e-01
7.95994475e-02 -1.00876344e-02 -1.15861511e+00 1.26214996e-01
1.56509623e-01 -5.95642328e-01 -6.46566451e-01 1.10615000e-01
3.52643639e-01 -8.59805495e-02 -1.07507634e+00 4.07082796e-01
7.17589259e-01 -9.04435873e-01 5.77255189e-01 -1.89440593e-01
1.09653957e-01 -1.30225480e-01 4.81093451e-02 -1.47964048e+00
-2.45322272e-01 -9.33276713e-01 -3.56024176e-01 7.63085961e-01
3.01655859e-01 -1.01305175e+00 5.88604510e-01 5.76262653e-01
1.03117399e-01 -5.05170286e-01 -1.21521866e+00 -1.17099214e+00
5.28379679e-01 -2.42287040e-01 2.85310209e-01 8.85477960e-01
2.65646636e-01 7.15798093e-03 -4.82496798e-01 3.93152505e-01
7.04155505e-01 1.18213750e-01 3.64929050e-01 -1.55339634e+00
-2.58794993e-01 -2.62751013e-01 -4.75500435e-01 -7.03845620e-01
2.69633234e-01 -5.59011221e-01 3.49443495e-01 -9.44865823e-01
-1.04384214e-01 -6.84722245e-01 -4.76240069e-01 7.72152171e-02
2.26587281e-01 2.58370519e-01 1.67019099e-01 4.62322950e-01
-1.44778386e-01 6.51061535e-01 9.06581819e-01 2.63384134e-01
-2.65060991e-01 2.47138336e-01 2.34184131e-01 5.28374374e-01
1.01923132e+00 -7.51283586e-01 4.80274186e-02 4.20942366e-01
-4.18495424e-02 4.74705637e-01 5.86347640e-01 -1.51347399e+00
6.47300184e-01 1.85970262e-01 5.11298850e-02 -5.78890800e-01
6.26748025e-01 -6.84988737e-01 1.60454884e-01 1.10493457e+00
-2.60023624e-01 -6.47425950e-02 4.61501062e-01 7.14907706e-01
-1.91713810e-01 -5.65303624e-01 9.54190195e-01 -4.51968051e-02
-4.07493711e-01 -1.32786602e-01 -8.24867785e-01 -4.10449058e-01
6.52428329e-01 5.37654385e-02 -2.33648434e-01 -7.39445925e-01
-6.88139498e-01 -3.10744792e-01 2.12780699e-01 -2.40754858e-01
1.39154613e-01 -7.82241404e-01 -5.52887678e-01 3.11603457e-01
-3.24213356e-01 -6.73419774e-01 2.20262349e-01 1.15478158e+00
-7.05303013e-01 8.51433098e-01 -6.29239157e-02 -6.87770903e-01
-8.67288947e-01 3.38624746e-01 8.92315924e-01 -4.51989740e-01
-7.09643066e-01 2.16822132e-01 -4.74753290e-01 -6.25579953e-01
-3.31496060e-01 -1.91151083e-01 -1.63327247e-01 -2.98069090e-01
2.84312069e-01 6.18139863e-01 -3.52363549e-02 -7.83505797e-01
-2.12010238e-02 7.73967564e-01 3.68311197e-01 -6.18690252e-01
1.30190110e+00 -1.85779575e-02 -3.55044007e-01 9.02465880e-01
1.17590284e+00 1.06213028e-02 -1.29175508e+00 -3.97670388e-01
9.41574499e-02 2.67521471e-01 4.47133273e-01 -7.69063771e-01
-6.89104319e-01 1.17072499e+00 7.19670594e-01 8.18981767e-01
8.07645619e-01 -3.30028147e-01 8.48069012e-01 8.67911756e-01
3.03701222e-01 -8.67555976e-01 -1.80813119e-01 8.57511282e-01
4.79343534e-01 -1.11211514e+00 1.43894240e-01 1.82558075e-01
-3.09358150e-01 1.57694352e+00 -2.91078035e-02 -5.83841562e-01
1.24644792e+00 5.33928692e-01 1.12681977e-01 4.37553599e-02
-8.28601956e-01 1.14006914e-01 1.56549066e-01 2.49338567e-01
-2.32188031e-02 4.84183915e-02 -9.71609801e-02 -7.14305267e-02
-3.44477922e-01 -1.75498098e-01 8.52342486e-01 8.10085714e-01
-5.48628867e-01 -1.01278603e+00 -3.13410372e-01 2.86520839e-01
-3.31292093e-01 -2.07902640e-01 -8.99202079e-02 1.00382936e+00
-3.04728866e-01 9.49083030e-01 1.41635492e-01 -3.63547474e-01
-1.03214286e-01 3.04655254e-01 2.36001700e-01 5.90115488e-02
-3.45285088e-01 -1.02758664e-03 -1.45276576e-01 -2.07765177e-01
-6.68537259e-01 -8.08394730e-01 -1.15902150e+00 -6.05466664e-01
-4.96624827e-01 6.88796043e-01 8.49996686e-01 1.23250329e+00
2.04264954e-01 2.89080203e-01 9.40117598e-01 -1.19854045e+00
-6.13721073e-01 -1.10774076e+00 -8.46326172e-01 -3.00306052e-01
9.36906815e-01 -5.28639436e-01 -1.08269882e+00 -1.58474147e-01] | [6.57594633102417, 3.4478542804718018] |
031c4e65-c0f1-48c7-85bc-4819fc73710c | multilingual-event-extraction-from-historical | 2305.10928 | null | https://arxiv.org/abs/2305.10928v1 | https://arxiv.org/pdf/2305.10928v1.pdf | Multilingual Event Extraction from Historical Newspaper Adverts | NLP methods can aid historians in analyzing textual materials in greater volumes than manually feasible. Developing such methods poses substantial challenges though. First, acquiring large, annotated historical datasets is difficult, as only domain experts can reliably label them. Second, most available off-the-shelf NLP models are trained on modern language texts, rendering them significantly less effective when applied to historical corpora. This is particularly problematic for less well studied tasks, and for languages other than English. This paper addresses these challenges while focusing on the under-explored task of event extraction from a novel domain of historical texts. We introduce a new multilingual dataset in English, French, and Dutch composed of newspaper ads from the early modern colonial period reporting on enslaved people who liberated themselves from enslavement. We find that: 1) even with scarce annotated data, it is possible to achieve surprisingly good results by formulating the problem as an extractive QA task and leveraging existing datasets and models for modern languages; and 2) cross-lingual low-resource learning for historical languages is highly challenging, and machine translation of the historical datasets to the considered target languages is, in practice, often the best-performing solution. | ['Isabelle Augenstein', 'Natalia da Silva Perez', 'Nadav Borenstein'] | 2023-05-18 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [ 0.03319117 0.01425107 -0.35121462 -0.00620659 -1.4682889 -1.2385551
1.1500462 0.40698755 -0.9672983 1.2559034 0.4773311 -0.44050166
0.13023174 -0.7242053 -0.75044745 -0.51761407 -0.07582855 0.9433783
-0.05734904 -0.4263887 0.24925889 0.46469927 -1.0918185 0.18703364
0.81891644 0.5343131 0.22792651 0.46571162 -0.2564778 0.764713
-0.6791158 -1.0285804 0.17250244 -0.3618982 -1.2223322 -0.29160842
0.4618975 0.13934976 -0.24820447 0.7395949 0.33419943 -0.16015768
0.7277435 -0.51471543 -0.5001143 1.1166557 -0.2587641 0.54761606
0.44857463 -0.29294446 1.450923 -0.835266 1.1182721 0.9814669
0.56309956 0.20911694 -1.2600263 -0.27077115 -0.15141739 0.42552492
-1.384472 -0.52678 0.75184274 -0.5983407 1.0584787 0.16374646
0.51689744 1.4581251 0.17030933 0.7725006 1.212166 -0.53852135
-0.06425631 0.21320851 -0.25903544 0.35936156 -0.03435702 -0.11488654
-0.7478058 -0.19698997 0.3560779 -0.58776957 -0.2409802 0.34271222
-1.6165832 0.96289384 -0.07931311 0.8146688 -0.2982323 -0.13750772
0.7179559 0.47133514 0.8336915 0.6101613 -0.6817245 -0.6052302
-1.3910933 0.47575805 0.9463393 0.7669501 0.52624047 -0.18096054
0.48750514 0.8520029 -0.2160032 0.60339135 0.3603582 -0.32936606
0.9951291 0.22221781 0.05133403 -0.9993364 -0.45092693 -0.2852237
-0.5595289 -0.35006762 0.87031865 -0.14753975 -0.5407271 1.5604173
0.24049479 -0.39061338 0.24840833 0.8216589 0.56649214 1.05885
0.18562917 -0.58394843 1.6427841 -0.5168562 -0.47509143 -0.60936046
0.33242399 -1.053329 1.0923612 0.56905 -1.0758144 -0.1537192
-0.92802674 -0.32612386 -0.50433975 0.16883947 0.55501765 0.24328077
-0.5275383 0.7527802 -0.93445295 -0.6649562 0.17332606 0.0787567
-0.4432578 0.1008155 -1.3824857 1.3535831 0.63858664 -0.02740818
-0.8226478 -0.6684393 -0.88500375 -0.19375576 0.53704274 0.02392029
1.3491738 -0.97364837 -1.1133559 1.2537295 0.10948566 -0.49101177
0.76060116 -0.429653 -0.4581262 0.07635985 0.24051434 -0.03202031
0.69656646 -0.9561456 -0.7286032 -0.37484515 -0.10854956 -0.06077532
-0.5140011 0.4696953 -0.23036191 -0.9915291 -0.22640139 -0.8299938
-0.0820444 -0.56013155 -0.24775757 -0.3077118 0.4069548 -1.4273491
1.3817734 -1.7373136 0.5290742 -0.04982392 -0.19910467 0.12593812
0.21203567 0.96169496 0.4287825 0.03085745 -0.59902304 -0.22009325
0.12193765 0.42313132 -0.63410705 0.55697227 0.2818664 0.7357715
-1.0583968 -0.894755 -0.13663675 0.31199878 -0.1732601 -0.06069969
-0.34444088 0.56055063 -0.3382696 0.72296894 0.09929032 0.04115038
0.46409613 0.05439113 -0.54429233 0.6846276 -0.75156003 1.9588727
-0.7397864 1.0520205 -0.143041 -0.9440112 0.6719266 0.5025028
0.47049966 -0.643704 0.19282152 0.6586794 -0.22287108 -0.6260994
0.890558 -0.6258297 -0.79591715 0.38326663 0.366514 -0.2880649
0.67078215 0.03772146 1.1299503 0.37842774 0.6424832 -0.37039763
0.46628657 0.6295461 0.6412305 0.35660174 0.22987626 0.42389393
0.38478577 -0.6696324 -1.5111805 -1.0346434 -0.49446285 1.0273677
-0.4049792 -0.60387754 -0.4323428 -0.7482883 -0.32819778 0.6883784
-0.34670988 0.55094975 -1.16351 -1.0829234 0.7760271 0.04710348
0.04439946 -1.1784205 -0.6332279 0.6905722 -0.55907464 -1.3769904
-0.07685893 -0.0069965 -0.4149781 -0.92649543 -0.799067 -0.78354275
0.20924953 -0.5584645 1.3288018 -0.31335557 -0.24859989 -0.06186268
-0.51862043 -0.4509191 -0.6394656 0.69725204 -0.07717492 -0.16925642
0.33479393 -0.6124244 -0.04052265 -0.20194204 -0.8576296 -0.1703294
0.53118724 0.72685724 0.21072592 -0.04297104 0.76767945 -1.023828
0.2609408 -0.6838177 -0.60199463 0.3241613 -0.24904366 -0.05863792
1.1406113 -0.44743386 -1.2753516 -0.02477561 -0.3870594 0.48470625
0.09321737 0.9227212 0.0701001 0.44625306 0.918158 0.19182816
-0.6148378 -0.6892203 0.49531147 0.81681865 0.90257025 -0.85586536
1.0849006 0.46161684 -0.23601566 -1.1597059 -1.0420667 -0.2630896
-0.977517 -0.23264733 0.8700245 -0.8111415 -0.10860669 0.16381195
-1.2203186 -0.2852639 -0.16901678 0.46312574 -0.47785592 0.36711305
-0.91220045 -0.5255904 -0.44670388 -0.7791056 0.96396995 -0.02747845
-0.39920613 -1.1841444 0.61708957 0.29904088 0.0228822 0.6957468
1.2563938 -0.7011297 -0.2206624 -0.0693321 0.15751086 0.04500913
0.04861033 0.01588719 -0.744807 -0.2795425 -0.1026556 -0.56869936
0.66322684 -0.2111819 0.18349573 -0.5104938 -0.15550151 0.17150514
1.5657161 -0.05690526 0.3970333 0.7266377 0.46904823 0.87795776
0.58696616 0.43852642 0.4740969 0.579883 -0.33582833 0.17930515
0.08544751 -0.18647097 0.62291574 1.1743733 -0.2979139 -0.19878359
-1.3702923 1.2170596 -1.7504413 -1.0202866 -0.14669387 2.0142438
1.4939288 0.18398774 0.14670296 0.11209855 0.2297958 0.31477463
0.02682635 -0.1888086 -0.37618053 0.6142338 0.40485913 0.3839669
-1.2199259 1.1634524 5.554428 1.0223866 -1.0256317 0.4976581
0.15358612 -0.33651307 -0.2395117 0.34537736 -0.82279605 0.5222631
1.2640425 -0.2640772 0.4038509 0.6096738 0.07590564 -0.02728579
-1.249565 0.8908082 0.27877262 -1.4097283 -0.2895504 0.10190921
0.61188865 0.24837759 -0.4043631 0.33619225 0.1998057 -0.89145166
1.1675799 0.2761812 0.65113413 -0.94312686 0.52673084 0.45216143
-1.0147073 0.11466748 -0.3759138 0.12293138 0.789144 0.5786328
-0.6480654 0.707178 0.49254128 0.70108587 -0.68389374 0.56863254
-0.5578259 0.8378008 -0.43935534 0.05978092 0.5151716 -0.24676135
0.7492864 1.5045449 0.46544424 -0.04739557 0.11440248 0.54601425
-0.17708871 0.5846014 -0.62152404 -0.38647282 0.23361096 1.3350099
-0.7302271 -0.22946967 -0.51907665 0.87516236 0.77297163 -0.01161381
-0.7405852 -0.4249988 0.10614394 0.281261 -0.00972771 -0.583658
-0.02241531 -1.4705039 0.21019025 -1.0294752 0.71663964 -0.12201273
-1.530345 0.87720263 0.02388071 -0.8853242 -0.5601646 -0.53832877
-0.1390006 0.55749774 -1.4630861 -1.5902001 0.36808246 0.38146037
0.7705696 -0.28907436 0.71317106 0.732944 -0.3125619 0.17789684
0.22652555 0.30915883 0.88540274 -1.2178947 0.37851143 1.0850769
0.7006311 0.22213337 0.880427 -0.68823266 -1.5138441 -0.8837938
1.5843776 -0.64442813 1.3379301 -0.7372573 -0.9886518 0.84883904
0.52026963 -0.641996 0.6296311 0.28010425 -0.15052833 0.2606759
-0.62952596 0.6906559 0.802736 -0.7423484 -1.0586138 0.8726633
0.21444778 -0.30667582 -1.2123818 -0.03612181 0.39411157 -0.23398182
0.63545984 -0.45721197 0.6992945 -0.11762588 -0.0563556 -1.2877667
0.24629457 -1.0024978 0.1846413 1.8155395 0.8476161 -0.2835894
0.5247399 0.26042607 0.018346 -0.16706619 -1.3052868 -0.75456554
0.47210476 -0.53482 0.20585698 1.2612787 0.20911223 0.76763064
-0.59288174 -0.03022577 0.52263355 0.08436168 0.55843765 -1.2206697
-0.19728912 -0.2280508 -0.19980854 -0.48696214 0.5024454 -1.0575637
0.18153556 -1.306207 -0.00863864 -0.515134 0.44199777 0.4866011
0.08588703 0.39758635 0.10427684 0.5060801 -0.02424785 0.3516108
0.7875068 -0.03031106 -0.1308685 -0.2226805 -0.3334488 0.8027599
0.6836338 -0.82895917 0.04956278 -0.552845 0.7703866 0.17994599
0.15994911 -0.71043617 0.22532882 -0.29888144 0.07833625 -0.6108577
0.15964015 -0.6699701 0.16244869 0.13364069 0.03244666 0.22034717
0.1438348 0.2572657 -0.43097115 -0.53557473 0.650658 -0.46295428
-0.6687199 0.0305011 -0.6643707 0.51432925 0.86347187 0.373018
-0.36102948 -0.00864311 -0.56467444 -0.08574692 0.4753414 0.2915047
0.09275923 -1.2374821 -1.1679273 -0.27400592 0.06850742 -0.07542156
-0.03684386 0.6961445 -0.89357376 0.40753928 -0.07272465 -0.2288663
-0.9052717 0.5190937 -0.22449546 -0.65498185 -0.6926585 0.46254316
-0.28134444 -0.3816281 -0.21765482 0.11792726 -0.27842033 0.67049
0.39340147 0.23509057 0.25683042 -0.94551945 -0.29319412 0.4209588
-0.16627644 -0.64539856 1.749571 -0.1470898 -0.3164043 0.5969454
1.0459098 0.5595607 -0.7701673 -0.4400032 0.65537167 -0.18481624
-0.24267034 -0.6425284 -0.53803915 0.70654124 -0.2230627 0.2016965
0.98015463 0.2640897 1.0927572 0.35222 0.5618166 -1.4895319
-0.37034804 0.64419556 0.88578945 -1.1365278 0.32234368 0.03462364
-0.6796978 1.2190523 0.05273923 -0.08235435 0.31372002 0.10084762
-0.02351783 -0.34905183 -0.5826346 -0.07167966 0.4054731 0.11470478
0.4728306 -0.08570196 -0.8414833 0.531184 -0.80686516 -0.47388354
0.6035474 1.0987735 -0.20827886 -1.4540265 -0.3874459 0.18805201
-1.1214554 -0.29359454 -0.5030074 1.170883 0.14510444 0.7038952
-0.08218177 0.06720918 0.21186619 0.16994247 0.5766991 -0.577603
-0.9098022 0.14627084 0.59658295 -0.05091995 -0.65749466 -1.0677475
-0.9159943 -0.46737608 0.02780366 0.47215968 0.70563495 1.2285621
-0.26812407 -0.0136738 0.317347 -0.8671175 -0.2736947 -0.91068983
-0.46432105 0.20269401 0.17050686 -0.2953601 -0.34677067 0.5081497 ] | [10.595136642456055, 10.059042930603027] |
b25bcacc-848a-41cc-8e2d-51a7b274b059 | coherent-reconstruction-of-multiple-humans-1 | 2006.08586 | null | https://arxiv.org/abs/2006.08586v1 | https://arxiv.org/pdf/2006.08586v1.pdf | Coherent Reconstruction of Multiple Humans from a Single Image | In this work, we address the problem of multi-person 3D pose estimation from a single image. A typical regression approach in the top-down setting of this problem would first detect all humans and then reconstruct each one of them independently. However, this type of prediction suffers from incoherent results, e.g., interpenetration and inconsistent depth ordering between the people in the scene. Our goal is to train a single network that learns to avoid these problems and generate a coherent 3D reconstruction of all the humans in the scene. To this end, a key design choice is the incorporation of the SMPL parametric body model in our top-down framework, which enables the use of two novel losses. First, a distance field-based collision loss penalizes interpenetration among the reconstructed people. Second, a depth ordering-aware loss reasons about occlusions and promotes a depth ordering of people that leads to a rendering which is consistent with the annotated instance segmentation. This provides depth supervision signals to the network, even if the image has no explicit 3D annotations. The experiments show that our approach outperforms previous methods on standard 3D pose benchmarks, while our proposed losses enable more coherent reconstruction in natural images. The project website with videos, results, and code can be found at: https://jiangwenpl.github.io/multiperson | ['Georgios Pavlakos', 'Wen Jiang', 'Xiaowei Zhou', 'Kostas Daniilidis', 'Nikos Kolotouros'] | 2020-06-15 | coherent-reconstruction-of-multiple-humans | http://openaccess.thecvf.com/content_CVPR_2020/html/Jiang_Coherent_Reconstruction_of_Multiple_Humans_From_a_Single_Image_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Jiang_Coherent_Reconstruction_of_Multiple_Humans_From_a_Single_Image_CVPR_2020_paper.pdf | cvpr-2020-6 | ['3d-depth-estimation', '3d-human-reconstruction'] | ['computer-vision', 'computer-vision'] | [-9.84511804e-03 2.86125660e-01 1.30775183e-01 -4.13827002e-01
-5.48774660e-01 -2.19212547e-01 3.72499496e-01 -2.41580922e-02
-4.78405029e-01 5.33886909e-01 1.48724839e-01 3.24265152e-01
1.40945911e-01 -6.65050685e-01 -8.51435959e-01 -7.00491369e-01
1.99192554e-01 1.15550709e+00 3.91236633e-01 -1.32252216e-01
-1.82263568e-01 3.59883249e-01 -1.43750215e+00 2.53439378e-02
8.52412462e-01 8.74714136e-01 2.14121625e-01 5.44784427e-01
2.70869404e-01 2.16552407e-01 -4.42822307e-01 -6.34191036e-01
6.07894897e-01 -3.15098464e-01 -6.78836584e-01 4.77694303e-01
8.15607250e-01 -5.74564219e-01 -3.36798251e-01 9.75201726e-01
6.66635990e-01 2.89493024e-01 5.65243781e-01 -1.20281792e+00
2.45864689e-02 2.99848706e-01 -8.95199060e-01 -3.06789160e-01
5.95453084e-01 2.48893231e-01 8.96806002e-01 -6.63859248e-01
7.22031236e-01 1.54433644e+00 6.88247383e-01 6.19619191e-01
-1.30616105e+00 -4.02933538e-01 4.02616322e-01 -1.26499295e-01
-1.35269046e+00 -2.16975108e-01 7.79384017e-01 -4.95093554e-01
5.16448855e-01 5.31919822e-02 1.02257562e+00 1.28588688e+00
-6.24543289e-03 8.64170909e-01 8.73137057e-01 -2.79516399e-01
9.17666685e-03 -6.86412305e-02 2.08101496e-02 7.79382408e-01
4.28068399e-01 -7.90283829e-03 -6.30710185e-01 6.06653653e-02
9.52189744e-01 5.07490002e-02 -2.03311086e-01 -8.39418769e-01
-1.00399458e+00 5.73117614e-01 5.82597673e-01 -7.88550526e-02
-3.11106771e-01 6.41673803e-02 2.08423778e-01 -1.40030026e-01
4.42119688e-01 9.28296521e-03 -3.15106511e-01 1.36939332e-01
-8.44677389e-01 9.01622951e-01 5.90425253e-01 8.40212643e-01
7.19067454e-01 -3.15291762e-01 -1.82196379e-01 6.35534346e-01
4.95164841e-01 4.37465101e-01 -1.01470321e-01 -1.33979917e+00
6.76144540e-01 5.91238260e-01 2.95511901e-01 -1.05127358e+00
-4.98354942e-01 -5.26899040e-01 -7.57993460e-01 3.37913603e-01
8.28815699e-01 -1.07254155e-01 -9.62626934e-01 1.82274640e+00
8.17857087e-01 6.58471510e-02 -4.19537872e-01 1.49845374e+00
6.39458060e-01 4.16484833e-01 -7.77132511e-02 1.41750038e-01
1.48180223e+00 -1.05142140e+00 -4.52630758e-01 -6.52490139e-01
2.10974663e-01 -4.88445222e-01 8.09717238e-01 4.74400759e-01
-1.42060769e+00 -6.13394201e-01 -8.94776404e-01 -2.11882025e-01
2.30051726e-01 4.06856127e-02 4.61703092e-01 3.37708473e-01
-8.18341374e-01 5.96027493e-01 -9.28886473e-01 -4.24446940e-01
4.47757691e-01 3.04261416e-01 -3.87885034e-01 -2.66807258e-01
-8.69733691e-01 7.16746628e-01 1.75066441e-01 3.90881121e-01
-7.26844728e-01 -4.72558051e-01 -9.99588430e-01 -1.86604336e-01
7.61765778e-01 -1.20940602e+00 1.11673498e+00 -8.97752047e-01
-1.13160062e+00 1.05906117e+00 -1.77720904e-01 -3.66160959e-01
1.31654501e+00 -7.01781094e-01 3.62333357e-01 2.84821749e-01
3.90706241e-01 8.63250196e-01 6.69112802e-01 -1.59633029e+00
-5.62986016e-01 -7.24184036e-01 8.50427970e-02 5.17655253e-01
2.61691034e-01 -2.72983819e-01 -1.05412841e+00 -4.25876319e-01
1.94292262e-01 -1.05022311e+00 -4.63315487e-01 3.57206762e-01
-8.14578712e-01 -1.38030306e-01 2.37895012e-01 -7.33928502e-01
6.45968020e-01 -1.89915884e+00 6.71691775e-01 2.36718774e-01
4.32999223e-01 -1.67846352e-01 1.61467925e-01 1.25938758e-01
2.55001873e-01 -2.38002598e-01 -4.20819521e-01 -1.09912169e+00
5.43508492e-02 1.30831853e-01 7.30416849e-02 8.36366653e-01
3.18296179e-02 7.07598448e-01 -7.38949120e-01 -6.33437335e-01
3.49716306e-01 8.03112686e-01 -7.73621917e-01 3.13329130e-01
-2.91771799e-01 9.57217038e-01 -4.86118376e-01 4.39924151e-01
6.92198277e-01 -3.39373291e-01 1.42151967e-01 -2.49960467e-01
7.57340789e-02 1.27605885e-01 -1.43655658e+00 2.02692127e+00
-8.87155980e-02 4.90435846e-02 3.10513198e-01 -7.42534637e-01
4.53301847e-01 1.30346239e-01 7.07983851e-01 -4.68193144e-01
2.40250766e-01 1.42755240e-01 -1.99020669e-01 -2.46386170e-01
2.14353472e-01 -6.12214655e-02 -1.25166342e-01 1.01253524e-01
-1.32123217e-01 4.71656350e-03 1.90734059e-01 2.34848917e-01
7.40612566e-01 5.27723968e-01 -7.66419694e-02 6.00886755e-02
1.97583094e-01 -1.37040943e-01 9.42231894e-01 6.83394194e-01
-1.62339553e-01 9.75926638e-01 4.79198605e-01 -5.55652499e-01
-1.08073747e+00 -1.18295538e+00 1.93631777e-03 5.48048198e-01
6.01239204e-01 -4.49948847e-01 -7.79296517e-01 -6.02724195e-01
-4.78649931e-03 3.19023609e-01 -4.57754731e-01 1.45759508e-01
-8.50233495e-01 -4.08935219e-01 2.86003739e-01 3.59556764e-01
4.53357160e-01 -8.42670560e-01 -8.19033980e-01 5.13151623e-02
-7.16353893e-01 -1.32746363e+00 -4.85033512e-01 -7.41858184e-02
-7.98769891e-01 -1.12467635e+00 -8.31398606e-01 -5.81098437e-01
8.64021480e-01 1.26401410e-02 1.11668515e+00 2.02201083e-01
-3.62809241e-01 1.77827954e-01 -8.76646787e-02 -1.80833775e-03
1.31789781e-02 -8.25842619e-02 1.64003491e-01 -2.74052192e-02
3.94489728e-02 -5.26457906e-01 -9.79490638e-01 3.15290481e-01
-3.44748169e-01 2.89395481e-01 4.03326541e-01 6.15639687e-01
8.17822576e-01 8.55805129e-02 -1.18457057e-01 -7.47219443e-01
-5.00874035e-02 -9.84118804e-02 -6.01272941e-01 -3.48019823e-02
-9.50165763e-02 -1.34135649e-01 2.82757968e-01 -3.02283496e-01
-9.57314789e-01 3.74805182e-01 -2.94097453e-01 -5.17144084e-01
-3.32152694e-01 -1.18887581e-01 -5.26628733e-01 2.23851010e-01
4.02225465e-01 3.69697027e-02 -1.91137437e-02 -6.04133606e-01
1.17738172e-01 8.01124275e-02 6.03863955e-01 -7.71155655e-01
8.26993406e-01 6.32082283e-01 7.57026002e-02 -7.43487477e-01
-9.39357042e-01 -3.68957043e-01 -7.76186287e-01 -4.39616978e-01
1.07388532e+00 -1.09694493e+00 -8.10983241e-01 6.40761912e-01
-1.30673361e+00 -4.55606699e-01 -2.40333200e-01 4.81191248e-01
-6.74235582e-01 4.21753228e-01 -7.38749027e-01 -7.86458790e-01
-6.90760911e-02 -1.32458496e+00 1.42894769e+00 1.66234881e-01
-3.37518752e-01 -6.31668687e-01 -5.79364970e-02 8.30589473e-01
-2.20143184e-01 5.74651480e-01 4.07202065e-01 -1.65928960e-01
-6.97871983e-01 -1.21163055e-01 4.97729406e-02 1.00875452e-01
-2.26363957e-01 -2.64080435e-01 -7.83888102e-01 -2.98332632e-01
-8.23157430e-02 -2.52919257e-01 8.99326980e-01 7.60959923e-01
9.62007403e-01 -2.91137487e-01 -5.11689007e-01 7.58686960e-01
1.15894771e+00 -4.10489172e-01 3.90002936e-01 1.57130480e-01
1.05475533e+00 1.10087025e+00 6.27935708e-01 5.41201770e-01
7.52281845e-01 1.03564548e+00 5.96535504e-01 -1.82101771e-01
-1.91533625e-01 -5.08453906e-01 1.64472401e-01 2.25974053e-01
-2.74796546e-01 -4.52979356e-01 -7.35342741e-01 2.50070035e-01
-1.97297847e+00 -7.43093014e-01 -2.53356606e-01 2.33924961e+00
5.70080221e-01 3.60856682e-01 5.14747500e-01 -8.52676928e-02
7.23011255e-01 5.26515320e-02 -6.16856635e-01 2.86237031e-01
1.05013967e-01 -1.90477982e-01 3.68964225e-01 7.15328097e-01
-1.15181935e+00 8.28646004e-01 4.75754642e+00 4.70999867e-01
-6.95866585e-01 1.52056004e-04 6.97908580e-01 -3.74588162e-01
-8.87911469e-02 -1.17051184e-01 -1.07190025e+00 4.45545554e-01
2.97330201e-01 3.95094037e-01 1.15958162e-01 5.43104231e-01
4.03046370e-01 -2.06959561e-01 -1.26387966e+00 1.04488814e+00
6.49061278e-02 -8.74766529e-01 -7.58070350e-02 3.35866570e-01
4.64546859e-01 -2.36031935e-01 -2.02176601e-01 -6.15100116e-02
1.82303220e-01 -8.47456098e-01 1.23482490e+00 5.88313222e-01
4.51717794e-01 -7.83705056e-01 6.05257213e-01 5.68865061e-01
-1.19478226e+00 1.00145333e-01 -1.04709432e-01 -1.20587960e-01
6.56557620e-01 8.26786220e-01 -3.40032429e-01 6.01096094e-01
8.75271320e-01 5.97942472e-01 -2.86602050e-01 1.05381501e+00
-3.81098598e-01 -1.05163334e-02 -5.11379540e-01 5.15565753e-01
-8.47127587e-02 -3.41385841e-01 7.60595381e-01 9.52239990e-01
1.83105826e-01 7.36862049e-02 6.78475022e-01 9.04988587e-01
2.29337495e-02 -1.33272842e-01 -2.71323562e-01 5.56723654e-01
2.20088229e-01 1.05519998e+00 -8.33757877e-01 -1.81987748e-01
-2.25263700e-01 1.29631507e+00 4.41933274e-01 3.09704691e-01
-9.46410716e-01 2.72133142e-01 6.99471533e-01 7.36287892e-01
1.90745234e-01 -2.29416326e-01 -2.71994472e-01 -1.30903113e+00
4.80834961e-01 -7.66869903e-01 2.90298045e-01 -5.26747823e-01
-1.32125521e+00 4.63559389e-01 1.64902776e-01 -1.10591638e+00
-1.77152932e-01 -3.33576560e-01 -3.15830976e-01 7.84160495e-01
-1.23512042e+00 -1.22631550e+00 -4.02120531e-01 4.69254136e-01
5.23647368e-01 3.95883560e-01 1.82978883e-01 5.16710460e-01
-6.12923086e-01 4.50878024e-01 -7.70479560e-01 1.82662457e-01
7.86328912e-01 -1.32818198e+00 2.85149872e-01 8.79614055e-01
-4.94725490e-03 4.38474596e-01 8.86062026e-01 -8.86570096e-01
-1.11603117e+00 -9.49863732e-01 9.01886821e-01 -4.82626617e-01
1.68760009e-02 -6.18265569e-01 -7.28914142e-01 7.21657515e-01
-2.03647316e-01 4.98071164e-02 1.88918814e-01 -2.96413936e-02
-1.58853903e-01 -3.12137250e-02 -1.12175202e+00 7.22132921e-01
1.36388779e+00 6.17624335e-02 -3.74757588e-01 3.36193472e-01
5.94028652e-01 -8.53712142e-01 -5.90427220e-01 2.99602002e-01
7.18570888e-01 -1.21874928e+00 1.31286240e+00 -1.91429302e-01
5.48258603e-01 -4.63342160e-01 2.53933519e-02 -9.59674537e-01
-1.64668933e-02 -4.00383204e-01 -1.55257702e-01 9.94088888e-01
8.94551054e-02 -3.84471804e-01 1.04831457e+00 8.33920717e-01
-2.27653701e-02 -8.70451748e-01 -9.05222297e-01 -5.66057324e-01
3.54256108e-03 -3.08534741e-01 1.96356907e-01 4.70496774e-01
-5.33596516e-01 3.92427355e-01 -6.86549664e-01 3.74155283e-01
1.18638492e+00 4.75100838e-02 1.04102123e+00 -1.32667565e+00
-6.61601245e-01 -3.16098154e-01 -2.72280395e-01 -1.62273061e+00
8.31026360e-02 -6.14217281e-01 2.84514815e-01 -1.66922891e+00
2.74992138e-01 -5.10064483e-01 3.99480045e-01 3.31885040e-01
-1.71765208e-01 2.54198045e-01 3.86461675e-01 1.78552613e-01
-7.24767268e-01 6.36143923e-01 1.53083181e+00 3.61216031e-02
-3.36126029e-01 2.63606220e-01 -5.79364896e-01 1.01536047e+00
5.60152948e-01 -5.49302399e-01 -2.14922085e-01 -5.55137813e-01
1.92404985e-01 2.64536470e-01 9.63680327e-01 -9.93688285e-01
2.17846826e-01 7.11142272e-02 7.01759398e-01 -7.41551161e-01
8.39000165e-01 -8.50904524e-01 3.08442503e-01 5.95843494e-01
-7.13463128e-02 -1.77291974e-01 -1.45555630e-01 5.35855889e-01
1.49109602e-01 9.31494385e-02 8.72686207e-01 -4.86657172e-01
-3.44199330e-01 5.72903156e-01 1.70498013e-01 2.02076137e-01
7.92782068e-01 -3.37238193e-01 1.15187466e-01 -3.29563767e-01
-8.41365516e-01 5.60778558e-01 7.68425643e-01 2.95502633e-01
5.63478351e-01 -1.13246775e+00 -8.80809963e-01 9.57247242e-02
-1.32793978e-01 7.78788745e-01 3.84811908e-01 9.22556937e-01
-5.15694022e-01 2.56557222e-02 2.62339599e-03 -9.03164506e-01
-1.23948967e+00 5.44704683e-02 6.39537334e-01 -4.00135040e-01
-1.11785734e+00 8.47561181e-01 6.21216714e-01 -5.46075463e-01
4.37195867e-01 -1.18799217e-01 5.14212325e-02 -1.17654391e-01
3.18386883e-01 3.92958045e-01 -2.09522083e-01 -9.34699476e-01
-4.44130003e-01 9.07325983e-01 3.06029990e-02 -2.16069296e-01
1.34700263e+00 -4.32005525e-01 7.52045438e-02 3.18299264e-01
9.09747183e-01 5.36554009e-02 -1.79188788e+00 -9.52749699e-02
-4.78525281e-01 -6.44146085e-01 -3.66235912e-01 -6.16415024e-01
-1.13580358e+00 6.86115921e-01 3.95249128e-01 -2.62612194e-01
8.96086216e-01 3.32988501e-01 1.00324833e+00 5.08513078e-02
6.03101373e-01 -9.56575572e-01 1.48527548e-01 2.89508104e-01
7.54307449e-01 -1.24018860e+00 7.14826658e-02 -6.54128790e-01
-6.05502605e-01 7.43569136e-01 9.48939860e-01 -2.70812124e-01
2.36183584e-01 7.32900426e-02 -1.61949188e-01 -1.98448911e-01
-3.93479288e-01 -1.29513279e-01 2.48710155e-01 5.44312656e-01
2.44438842e-01 5.35257086e-02 -9.99794900e-02 5.58727920e-01
-4.14985061e-01 -2.71519840e-01 2.91789711e-01 6.17973864e-01
-2.18548283e-01 -1.09147906e+00 -6.06935501e-01 1.85542881e-01
-4.00642604e-01 3.83274108e-01 -3.06614846e-01 6.95830643e-01
4.57597882e-01 7.37605155e-01 7.63950720e-02 -3.72698642e-02
6.78025126e-01 -2.73891300e-01 7.25819767e-01 -6.28401101e-01
-4.09670979e-01 4.33807373e-01 1.83880199e-02 -9.03354347e-01
-4.94446248e-01 -8.59589517e-01 -1.37630248e+00 -3.63712221e-01
-3.23634297e-02 -9.64773670e-02 3.23274106e-01 9.40254688e-01
1.27411246e-01 4.37157810e-01 2.42577627e-01 -1.37483954e+00
-2.49649689e-01 -5.99342346e-01 -6.00758135e-01 6.42221332e-01
4.40213650e-01 -9.20698345e-01 -4.06459004e-01 -2.55346373e-02] | [7.108514308929443, -1.0335441827774048] |
41403fe0-aff7-41a3-a115-9a551b8c6486 | feature-compatible-progressive-learning-for | 2304.10305 | null | https://arxiv.org/abs/2304.10305v2 | https://arxiv.org/pdf/2304.10305v2.pdf | Feature-compatible Progressive Learning for Video Copy Detection | Video Copy Detection (VCD) has been developed to identify instances of unauthorized or duplicated video content. This paper presents our second place solutions to the Meta AI Video Similarity Challenge (VSC22), CVPR 2023. In order to compete in this challenge, we propose Feature-Compatible Progressive Learning (FCPL) for VCD. FCPL trains various models that produce mutually-compatible features, meaning that the features derived from multiple distinct models can be directly compared with one another. We find this mutual compatibility enables feature ensemble. By implementing progressive learning and utilizing labeled ground truth pairs, we effectively gradually enhance performance. Experimental results demonstrate the superiority of the proposed FCPL over other competitors. Our code is available at https://github.com/WangWenhao0716/VSC-DescriptorTrack-Submission and https://github.com/WangWenhao0716/VSC-MatchingTrack-Submission. | ['Yi Yang', 'Yifan Sun', 'Wenhao Wang'] | 2023-04-20 | null | null | null | null | ['video-similarity'] | ['computer-vision'] | [-3.86039056e-02 -7.14849055e-01 -2.76578486e-01 -6.62725493e-02
-1.26584756e+00 -4.22907829e-01 6.56682491e-01 -8.43580142e-02
-1.30449906e-01 3.92460734e-01 3.74882609e-01 3.89523320e-02
-1.98021770e-01 -2.18568608e-01 -7.58696973e-01 -2.67966062e-01
-5.30217588e-01 1.69745490e-01 4.09033298e-01 -1.06144466e-01
5.17738760e-01 1.68483302e-01 -1.73031747e+00 8.07920218e-01
5.98677516e-01 9.84612525e-01 5.65437853e-01 6.57100260e-01
8.83639827e-02 8.25162590e-01 -2.48262092e-01 -3.64432663e-01
6.90920532e-01 -4.18380171e-01 -7.88383067e-01 -7.80704916e-02
5.55193186e-01 -3.41476768e-01 -6.80756986e-01 1.17902827e+00
4.99637812e-01 2.92223934e-02 5.86099386e-01 -1.88790047e+00
-6.30917728e-01 5.15190721e-01 -7.83890784e-01 5.21259844e-01
8.47794354e-01 9.80114564e-02 1.04417098e+00 -1.34374988e+00
8.09132040e-01 1.14496315e+00 6.65288031e-01 5.36799192e-01
-6.75020695e-01 -9.65531826e-01 9.38986614e-02 8.78565788e-01
-1.78540421e+00 -4.69114214e-01 6.80130720e-01 -3.16243798e-01
8.02272081e-01 4.26406860e-01 4.86676455e-01 1.15770578e+00
-1.12270461e-02 1.40678656e+00 8.09070766e-01 -4.04573917e-01
-5.06142490e-02 1.80411208e-02 7.07027549e-03 7.69078076e-01
5.10898717e-02 -2.70744544e-02 -6.47550762e-01 -2.51412481e-01
5.35393476e-01 8.26414749e-02 -6.16843641e-01 -4.21459913e-01
-1.16263628e+00 5.84482610e-01 2.28590831e-01 4.78106171e-01
-2.10334063e-01 3.54860015e-02 5.21817267e-01 7.60940850e-01
9.07386318e-02 5.20468473e-01 -6.51544034e-02 -3.58811766e-01
-9.89252567e-01 3.61611754e-01 5.13053715e-01 1.56160843e+00
6.03268087e-01 -3.11671495e-01 -3.56278390e-01 8.18525612e-01
2.20900357e-01 2.86191434e-01 6.51960433e-01 -1.26714432e+00
5.57194829e-01 3.09280813e-01 -3.23581211e-02 -1.17461932e+00
3.68245691e-02 -2.11759880e-01 -6.64176464e-01 -2.23173663e-01
-1.97812900e-01 2.26261005e-01 -6.04448259e-01 1.51162064e+00
2.91818231e-02 9.55023587e-01 -2.66969562e-01 1.00511253e+00
9.62549686e-01 7.20682621e-01 -1.72807023e-01 -4.62850444e-02
9.97337639e-01 -1.22823751e+00 -2.90100008e-01 2.54073054e-01
6.52492225e-01 -1.06453788e+00 7.97553539e-01 2.26380035e-01
-1.03236175e+00 -7.29341209e-01 -1.08634698e+00 1.75244257e-01
-1.10216044e-01 -1.03504144e-01 3.19937408e-01 1.89089969e-01
-1.20786798e+00 7.54893959e-01 -5.85782290e-01 -3.76433045e-01
5.32908142e-01 1.39923111e-01 -4.90906954e-01 -3.41050148e-01
-1.28180456e+00 6.30944788e-01 3.95859331e-01 -2.50309497e-01
-8.64623249e-01 -8.48525941e-01 -5.75636327e-01 -8.02172199e-02
5.33727646e-01 -3.40309352e-01 1.19547188e+00 -7.87565529e-01
-9.14319932e-01 8.78056824e-01 -1.46658853e-01 -4.67492193e-01
8.05589676e-01 -4.40669656e-01 -6.46537364e-01 4.42020535e-01
3.29492003e-01 7.30601788e-01 8.21215808e-01 -1.33012486e+00
-9.46708977e-01 7.02461079e-02 -1.39218226e-01 1.79686382e-01
-3.16247970e-01 2.73389697e-01 -1.24964786e+00 -8.22211742e-01
-7.36977458e-02 -1.01294243e+00 6.93718642e-02 9.01966691e-02
-3.32123399e-01 -4.10533935e-01 1.01164031e+00 -5.06559372e-01
1.27342224e+00 -2.31749654e+00 7.52082735e-04 3.98690611e-01
2.94456869e-01 5.83852887e-01 -7.05247581e-01 5.12527287e-01
1.63897555e-02 1.61260217e-01 3.72327827e-02 -4.65851068e-01
6.51217857e-03 -2.35510632e-01 -2.06997946e-01 3.20455551e-01
3.18570495e-01 8.97442818e-01 -9.14483130e-01 -6.06773555e-01
1.10844262e-01 3.00069720e-01 -4.43500370e-01 3.19163769e-01
6.99740499e-02 -4.10212576e-02 -4.60880131e-01 8.85258496e-01
9.19866562e-01 -4.43244070e-01 1.68698505e-01 -3.16842169e-01
2.02630505e-01 -7.38339499e-02 -1.33115053e+00 1.77098465e+00
8.10777396e-02 6.70886457e-01 -3.34987670e-01 -9.33061779e-01
7.50265360e-01 2.08851263e-01 6.54356599e-01 -7.97925532e-01
1.23198114e-01 2.96312571e-01 -3.36679429e-01 -3.62632841e-01
7.30963409e-01 5.39413929e-01 1.00109078e-01 3.92966896e-01
3.24679911e-01 2.43027523e-01 3.00129205e-01 6.70671940e-01
1.15578806e+00 2.61189610e-01 2.64856905e-01 -1.81063816e-01
8.34484279e-01 -6.43687919e-02 6.19272649e-01 9.58396256e-01
-5.67631900e-01 8.61567914e-01 1.69121191e-01 -1.91395879e-01
-1.00278604e+00 -1.09273076e+00 6.37031272e-02 7.49306679e-01
6.03710353e-01 -8.53443682e-01 -4.00470614e-01 -6.47310317e-01
2.56158203e-01 4.69467968e-01 -4.73669857e-01 -2.79137373e-01
-5.80439270e-01 1.21266685e-01 5.83590329e-01 3.73996288e-01
6.02746665e-01 -7.94031680e-01 -2.02671811e-01 -3.50397266e-02
-4.20414090e-01 -1.09083509e+00 -9.01761532e-01 -3.26154441e-01
-4.16239113e-01 -1.25082886e+00 -8.93695593e-01 -1.03940797e+00
2.79316574e-01 1.00569904e+00 1.00501585e+00 5.35843432e-01
-3.96622896e-01 7.80810356e-01 -6.44718647e-01 -1.09473847e-01
-3.76316875e-01 3.24457628e-03 2.56951600e-01 -5.90951033e-02
6.42806351e-01 -3.66122872e-01 -5.16264319e-01 5.30787408e-01
-6.55904710e-01 -1.72223356e-02 3.01764578e-01 7.03903019e-01
6.80096865e-01 -1.70954511e-01 3.67193937e-01 -5.50608456e-01
6.22349203e-01 -7.59721041e-01 -4.83238816e-01 5.97999275e-01
-4.37886626e-01 -2.57734418e-01 3.21324944e-01 -3.72140318e-01
-7.31087923e-01 -2.53160715e-01 8.53200331e-02 -1.28285682e+00
-3.46107148e-02 4.28697854e-01 3.14041264e-02 -1.72046944e-01
3.37644964e-01 5.49922764e-01 3.75578105e-02 -4.49279636e-01
2.39099488e-01 7.37252951e-01 6.82333648e-01 -5.52635491e-01
1.12583673e+00 7.06266090e-02 -3.63273859e-01 -5.91768324e-01
-3.08236152e-01 -8.77708495e-01 -4.30215567e-01 -5.29307485e-01
4.29122299e-01 -1.15587461e+00 -4.20958430e-01 5.85644245e-01
-8.51280868e-01 -6.45085052e-02 7.31937215e-02 6.43396795e-01
-5.59987843e-01 7.26139545e-01 -6.13976061e-01 -4.32079792e-01
-2.85115391e-01 -1.01675892e+00 8.15493405e-01 3.09408784e-01
-2.59000450e-01 -7.31891215e-01 5.96985519e-02 4.37040120e-01
3.22303921e-01 -8.93130079e-02 4.93220001e-01 -9.23232079e-01
-7.22270668e-01 -3.21447551e-01 -2.34098628e-01 2.87562937e-01
8.95562395e-02 1.70071781e-01 -6.79503739e-01 -7.28855610e-01
-3.73526573e-01 -5.02831340e-01 9.18323338e-01 7.54303187e-02
1.21390975e+00 2.26419549e-02 -4.25535649e-01 5.56678891e-01
1.38176250e+00 1.52738050e-01 7.36690223e-01 6.14429951e-01
5.69643199e-01 3.99524301e-01 8.76498938e-01 5.36580026e-01
4.91183996e-01 8.08117449e-01 1.05243802e-01 2.83912778e-01
-3.83827835e-01 -4.01460558e-01 4.69155192e-01 9.27913845e-01
-2.28900969e-01 -3.69841933e-01 -8.75616968e-01 5.86955011e-01
-1.82660234e+00 -1.21675694e+00 -1.41796574e-01 2.05681920e+00
5.57292342e-01 9.74544510e-02 1.31065741e-01 -1.18469834e-01
1.02292097e+00 1.33428290e-01 -3.30580205e-01 -5.21235913e-03
-2.77320355e-01 -2.00934373e-02 2.32435182e-01 1.87654495e-01
-1.34453964e+00 8.63580525e-01 4.65266848e+00 1.37710178e+00
-9.25624073e-01 1.60576507e-01 3.49708796e-01 -2.76864976e-01
-2.48170435e-01 -1.13084160e-01 -6.11726820e-01 7.83723116e-01
6.25013053e-01 -6.36174500e-01 3.43124032e-01 7.98296452e-01
-1.73466519e-01 1.49099469e-01 -1.09787035e+00 1.35946047e+00
3.74810934e-01 -1.69625914e+00 2.78336316e-01 -1.06579341e-01
8.35402668e-01 2.86892891e-01 1.46403223e-01 5.34153759e-01
4.86321747e-02 -5.64914465e-01 6.55550301e-01 4.84051555e-01
6.14601135e-01 -5.59676230e-01 7.38831699e-01 -8.04123506e-02
-1.57018960e+00 -1.15257822e-01 -4.35964942e-01 5.66441298e-01
1.97974011e-01 1.06940903e-01 -6.29805028e-01 9.53345001e-01
9.60227549e-01 1.14473498e+00 -8.45096648e-01 1.57328224e+00
6.63552135e-02 4.54828650e-01 -1.31205559e-01 1.81285739e-01
2.23375618e-01 1.67448118e-01 8.69919240e-01 1.29902184e+00
6.34290457e-01 -7.24921748e-02 3.37334186e-01 4.49132562e-01
-3.35902452e-01 2.78753757e-01 -4.72051769e-01 1.34522974e-01
9.22057450e-01 9.67732608e-01 -3.75283480e-01 -4.53667760e-01
-5.95562518e-01 1.09023666e+00 1.89884543e-01 7.81451762e-02
-1.00662875e+00 -3.26856464e-01 6.93983078e-01 -5.09640314e-02
5.41094005e-01 5.61614819e-02 4.02161270e-01 -1.42684495e+00
2.59670556e-01 -1.00404871e+00 5.87513506e-01 -7.61460602e-01
-1.66625500e+00 6.81990445e-01 -8.20614249e-02 -2.07369328e+00
-4.70950939e-02 -3.34446996e-01 -6.87213480e-01 4.89636093e-01
-1.42575777e+00 -1.00005770e+00 -3.25553209e-01 9.69724655e-01
7.05043018e-01 -6.57504857e-01 5.70396721e-01 5.47829151e-01
-2.68886685e-01 1.10003674e+00 2.75092453e-01 1.49060056e-01
1.04352963e+00 -8.21946800e-01 5.11593044e-01 9.11192119e-01
2.58024931e-01 4.34480816e-01 3.17016244e-01 -6.92102969e-01
-1.49750292e+00 -1.26170433e+00 6.53131545e-01 -2.12486461e-01
7.24006653e-01 -3.44929993e-02 -1.22315919e+00 3.37532192e-01
2.19713807e-01 -2.48139035e-02 8.46570790e-01 -3.22639108e-01
-7.38823593e-01 -8.19762945e-02 -9.78703678e-01 5.60816646e-01
1.45868719e+00 -7.97749221e-01 -5.97350478e-01 2.59204715e-01
6.80839539e-01 -1.58541903e-01 -9.33901727e-01 3.73482198e-01
5.92872918e-01 -9.51185226e-01 8.45695913e-01 -6.57636702e-01
4.68104959e-01 -4.87679482e-01 -5.85194409e-01 -8.11203957e-01
-6.38183713e-01 -6.41198516e-01 -1.93572745e-01 1.44002104e+00
2.90126532e-01 -4.08033848e-01 5.82331121e-01 3.63524824e-01
-2.19133854e-01 -7.00794399e-01 -8.34909141e-01 -1.36450243e+00
-4.75315340e-02 -5.20339549e-01 5.28380811e-01 1.19801319e+00
5.42290919e-02 -3.28790605e-01 -6.79558098e-01 1.39049605e-01
7.77719796e-01 2.14339659e-01 7.80133724e-01 -8.49343300e-01
-4.68464941e-01 -5.78770757e-01 -8.09620082e-01 -1.07397842e+00
3.93618755e-02 -1.13410485e+00 -1.57101214e-01 -1.18906415e+00
6.31132364e-01 -2.95566291e-01 -6.95408642e-01 3.28148514e-01
-2.95364320e-01 3.93029660e-01 7.48288810e-01 7.19271481e-01
-1.17568219e+00 4.93690491e-01 8.23878765e-01 -2.23330513e-01
3.88861597e-02 -1.48140594e-01 -5.71695447e-01 3.40058565e-01
1.11109018e+00 -4.76417303e-01 -3.01691413e-01 -3.49853188e-01
-3.76822323e-01 -1.62836567e-01 3.14799517e-01 -1.07367837e+00
6.18735492e-01 1.55371606e-01 3.66322726e-01 -7.67301559e-01
1.81764290e-01 -5.99158347e-01 3.42790365e-01 5.62966168e-01
-4.38010544e-01 8.36384296e-02 1.72626093e-01 6.07474625e-01
-4.72702503e-01 -3.19801480e-01 4.22466874e-01 -7.05123842e-02
-1.24058795e+00 4.89641756e-01 -2.89319366e-01 1.67921111e-01
1.14871490e+00 -3.04461211e-01 -5.10644257e-01 -4.94483083e-01
-3.52181107e-01 7.76974440e-01 5.72991908e-01 1.05186605e+00
1.00495601e+00 -1.66028774e+00 -1.04669654e+00 -7.42527144e-03
4.85391378e-01 -7.03993320e-01 4.88964319e-01 8.24737549e-01
-4.08434302e-01 4.53294039e-01 -3.44804406e-01 -5.93568683e-01
-1.74794376e+00 6.86449528e-01 2.70769279e-02 -1.00850977e-01
-4.22224134e-01 1.06337118e+00 -2.24415720e-01 -2.28192016e-01
2.99113601e-01 4.03816849e-01 -6.50806576e-02 -1.51493371e-01
6.45435631e-01 5.16958416e-01 -1.27612025e-01 -7.16829777e-01
-7.20501184e-01 5.96484840e-01 -5.72169840e-01 2.93689400e-01
1.29066896e+00 -1.70348525e-01 4.89830188e-02 -1.35410288e-02
1.63841140e+00 -1.70426741e-01 -1.02554703e+00 -6.26147449e-01
2.89484560e-01 -1.00420392e+00 -1.60023689e-01 -4.97660637e-01
-1.15030873e+00 3.93315941e-01 9.51202691e-01 -2.24070549e-01
1.19844961e+00 1.14792861e-01 7.45388687e-01 4.68035907e-01
6.88343704e-01 -1.04930866e+00 2.67524153e-01 2.99926758e-01
1.01317751e+00 -1.29920292e+00 1.03212260e-01 -3.56663227e-01
-8.99238765e-01 1.08864462e+00 7.41909504e-01 -2.78271794e-01
6.42373741e-01 -3.24410088e-02 -6.19436465e-02 -4.16867547e-02
-9.11415279e-01 -3.66374701e-02 4.97999758e-01 5.51095188e-01
1.31476402e-01 -8.83354545e-02 -2.73589224e-01 4.15318489e-01
1.09008111e-01 8.99987593e-02 3.45802695e-01 1.14694738e+00
-1.70293555e-01 -1.30477536e+00 -1.43945709e-01 5.99353909e-01
-1.30328864e-01 -2.34687909e-01 -3.65997791e-01 7.01008022e-01
5.57565391e-02 8.50844860e-01 8.08684677e-02 -9.74655807e-01
1.54601097e-01 -1.08740538e-01 4.50439364e-01 -2.77090520e-01
-3.47755462e-01 2.04694107e-01 5.36711626e-02 -8.76529038e-01
-5.25251627e-01 -9.53745902e-01 -9.95948017e-01 -4.30201203e-01
-2.28811488e-01 3.88184816e-01 1.49698272e-01 4.35387492e-01
7.46516466e-01 3.87073830e-02 1.06129456e+00 -7.23135769e-01
-5.09495676e-01 -7.52891064e-01 -3.99696350e-01 4.55840230e-01
1.08555876e-01 -7.50407696e-01 -3.36884260e-01 2.08313242e-02] | [10.264575958251953, 0.736552357673645] |
89dac179-e3bf-4b0f-a653-cda50302593b | ranking-distance-calibration-for-cross-domain | 2112.00260 | null | https://arxiv.org/abs/2112.00260v2 | https://arxiv.org/pdf/2112.00260v2.pdf | Ranking Distance Calibration for Cross-Domain Few-Shot Learning | Recent progress in few-shot learning promotes a more realistic cross-domain setting, where the source and target datasets are from different domains. Due to the domain gap and disjoint label spaces between source and target datasets, their shared knowledge is extremely limited. This encourages us to explore more information in the target domain rather than to overly elaborate training strategies on the source domain as in many existing methods. Hence, we start from a generic representation pre-trained by a cross-entropy loss and a conventional distance-based classifier, along with an image retrieval view, to employ a re-ranking process for calibrating a target distance matrix by discovering the reciprocal k-nearest neighbours within the task. Assuming the pre-trained representation is biased towards the source, we construct a non-linear subspace to minimise task-irrelevant features therewithin while keep more transferrable discriminative information by a hyperbolic tangent transformation. The calibrated distance in this target-aware non-linear subspace is complementary to that in the pre-trained representation. To impose such distance calibration information onto the pre-trained representation, a Kullback-Leibler divergence loss is employed to gradually guide the model towards the calibrated distance-based distribution. Extensive evaluations on eight target domains show that this target ranking calibration process can improve conventional distance-based classifiers in few-shot learning. | ['Chengjie Wang', 'Yanwei Fu', 'Shaogang Gong', 'Pan Li'] | 2021-12-01 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Li_Ranking_Distance_Calibration_for_Cross-Domain_Few-Shot_Learning_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Ranking_Distance_Calibration_for_Cross-Domain_Few-Shot_Learning_CVPR_2022_paper.pdf | cvpr-2022-1 | ['cross-domain-few-shot', 'cross-domain-few-shot-learning'] | ['computer-vision', 'computer-vision'] | [ 4.15448844e-01 2.06863657e-02 -3.55218351e-01 -4.29893821e-01
-1.16271698e+00 -4.92643058e-01 8.49640071e-01 5.07751144e-02
-4.83153105e-01 5.66904843e-01 2.94694334e-01 2.97045916e-01
-5.22220373e-01 -6.99828625e-01 -3.99990082e-01 -9.45706964e-01
3.18933055e-02 6.52899861e-01 4.34592128e-01 -2.11492449e-01
3.13065082e-01 1.87425718e-01 -1.51964915e+00 2.53192157e-01
6.79509044e-01 9.14795160e-01 2.46974796e-01 2.51236320e-01
4.08692518e-03 4.87216711e-01 -4.94505972e-01 -3.48397523e-01
6.53655767e-01 -6.31218851e-01 -8.11920881e-01 3.05668414e-01
3.65825593e-01 6.00898080e-02 -2.62827039e-01 1.25105608e+00
4.02874887e-01 5.54685652e-01 1.16692960e+00 -1.21259141e+00
-6.91570401e-01 1.31745651e-01 -5.38183510e-01 7.82965720e-02
3.19474220e-01 -1.64780125e-01 1.02100551e+00 -1.05773556e+00
9.74101067e-01 1.04892933e+00 6.79060340e-01 4.50483501e-01
-1.52867723e+00 -4.84838843e-01 -4.65823822e-02 3.92706603e-01
-1.62050629e+00 -4.48946059e-01 1.01976216e+00 -5.00499368e-01
5.93029976e-01 1.74471643e-02 3.82775187e-01 1.24836397e+00
8.91321823e-02 3.83834332e-01 1.06464767e+00 -7.45031118e-01
6.19784415e-01 7.78136551e-01 -5.03679626e-02 2.12036580e-01
-8.70293304e-02 2.06014916e-01 -4.12157863e-01 -2.73298264e-01
4.03339505e-01 1.27529614e-02 -3.56594980e-01 -1.70667958e+00
-9.37253654e-01 1.26027215e+00 5.89475453e-01 4.88157958e-01
-2.15604201e-01 -4.33662653e-01 4.97855335e-01 4.23445284e-01
5.12793899e-01 5.55324793e-01 -2.68895447e-01 8.97641629e-02
-1.00072801e+00 5.76909296e-02 9.04663801e-01 9.91305292e-01
1.07439792e+00 -3.74581337e-01 -6.69954717e-02 1.20171976e+00
1.62396669e-01 1.44482464e-01 7.08148301e-01 -8.33487928e-01
4.73513156e-01 4.97983575e-01 -7.82933831e-02 -7.96120524e-01
4.34282832e-02 -4.19732898e-01 -6.06760800e-01 4.00463045e-01
5.81474900e-01 1.98874876e-01 -7.48847604e-01 1.77781820e+00
5.28292835e-01 1.46421641e-01 2.70841420e-01 9.64583218e-01
6.08709008e-02 4.68673229e-01 5.81913590e-02 -2.43813202e-01
1.14675283e+00 -6.96772814e-01 -2.22333267e-01 -3.29061568e-01
8.80272388e-01 -6.13709450e-01 1.16692472e+00 2.03902841e-01
-4.79244411e-01 -5.29749572e-01 -1.36088824e+00 8.49406794e-02
-8.18405867e-01 -2.36498505e-01 2.77514067e-02 4.48360026e-01
-6.54325068e-01 7.44229496e-01 -1.44198373e-01 -6.40632391e-01
4.06697989e-01 5.53552322e-02 -6.18512154e-01 -5.29617250e-01
-1.58305168e+00 1.20835960e+00 7.40578830e-01 -7.14228570e-01
-6.90403938e-01 -8.57556820e-01 -9.83662486e-01 -1.74450621e-01
4.80928659e-01 -4.70734805e-01 8.71263683e-01 -1.09847915e+00
-1.36461806e+00 1.06904256e+00 3.33319277e-01 -4.04531986e-01
5.75490296e-01 -1.35728503e-02 -3.80085468e-01 8.71619284e-02
4.62195605e-01 4.97429579e-01 1.18851399e+00 -1.36733031e+00
-5.06061316e-01 -4.76118386e-01 -1.30736321e-01 6.44429922e-01
-4.25513744e-01 -1.85777619e-01 -3.79673809e-01 -6.60198808e-01
6.81007206e-02 -1.00502551e+00 9.39786620e-03 1.06731094e-01
3.39030363e-02 -3.42855081e-02 9.30147529e-01 -2.58658886e-01
1.03147328e+00 -2.36362386e+00 2.78569520e-01 2.43748724e-01
1.57946814e-02 3.02322775e-01 -3.52892220e-01 4.15080875e-01
-3.81671786e-01 -5.53946376e-01 -5.05846083e-01 -4.31946032e-02
-1.13149859e-01 2.29354545e-01 -4.48633999e-01 8.47428977e-01
2.06534386e-01 5.04953325e-01 -1.25405693e+00 -5.33134043e-01
5.11056542e-01 5.06773055e-01 -3.43286425e-01 1.48841918e-01
8.51117373e-02 2.02550083e-01 -2.93634802e-01 3.15971375e-01
7.48696923e-01 -1.02639116e-01 6.64655790e-02 -6.14470355e-02
2.70085722e-01 -9.07163881e-03 -1.20719719e+00 2.00322556e+00
-5.69398940e-01 4.82194722e-01 -3.03877532e-01 -1.48111486e+00
1.14694262e+00 3.06284204e-02 6.35972917e-01 -6.40761197e-01
8.99662729e-03 1.50147438e-01 -1.39410108e-01 -2.52931565e-01
3.59018380e-03 -4.75293249e-01 -2.22745195e-01 2.57804871e-01
4.73100781e-01 -2.74492204e-01 -5.50591722e-02 8.58371332e-02
9.74207461e-01 4.08404261e-01 6.59198046e-01 -4.37790543e-01
6.08291507e-01 9.76802260e-02 4.10467029e-01 5.53230047e-01
-4.85444725e-01 8.04251969e-01 1.45932823e-01 -2.84093112e-01
-1.21030068e+00 -1.27355552e+00 -5.17909884e-01 1.40637004e+00
2.50621378e-01 -3.05271506e-01 -6.35341227e-01 -1.07545221e+00
8.91924948e-02 1.00862551e+00 -9.50643957e-01 -7.54813254e-01
-2.87501365e-01 -3.22163194e-01 1.42604023e-01 1.96153402e-01
3.65592480e-01 -6.24546647e-01 -5.68890750e-01 1.86297044e-01
-3.09949610e-02 -6.67087078e-01 -4.28060055e-01 7.20646858e-01
-7.98009813e-01 -1.00001860e+00 -1.13612783e+00 -7.98343182e-01
6.07804179e-01 3.51403683e-01 9.17178452e-01 -8.59596550e-01
-3.13534081e-01 4.35886860e-01 -3.98167282e-01 -2.06022620e-01
-2.70086437e-01 -1.95415318e-01 6.21715188e-02 6.91862181e-02
1.06591201e+00 -5.01523793e-01 -3.14417064e-01 3.74959946e-01
-7.17710316e-01 -4.27778125e-01 5.71323633e-01 1.15452993e+00
4.28821981e-01 4.31146733e-02 5.04743993e-01 -7.81259120e-01
4.53899264e-01 -7.74183691e-01 -4.01774228e-01 3.48958552e-01
-7.34598160e-01 1.81255400e-01 6.00559473e-01 -8.96777391e-01
-9.72143471e-01 3.60984296e-01 5.68958640e-01 -9.13618267e-01
-1.62003279e-01 1.70648664e-01 -3.29110950e-01 7.29911253e-02
1.27258289e+00 3.04571629e-01 5.91581278e-02 -2.45554641e-01
5.99606574e-01 8.26072514e-01 3.44989330e-01 -3.33460540e-01
9.20439839e-01 3.57689530e-01 -1.23541877e-01 -8.46479177e-01
-1.06618381e+00 -9.43281233e-01 -1.12012315e+00 7.64973788e-03
6.53867781e-01 -8.49080503e-01 1.35892212e-01 1.94984786e-02
-8.49750996e-01 -1.48233045e-02 -8.00155759e-01 6.11399472e-01
-8.59682262e-01 3.13056141e-01 1.58523783e-01 -7.18557477e-01
1.74143493e-01 -8.39998007e-01 1.11163247e+00 -1.07808039e-01
-5.00289977e-01 -1.27789629e+00 6.22565985e-01 1.62434772e-01
2.54664809e-01 -1.88964512e-02 9.09509838e-01 -1.22131789e+00
-3.66095863e-02 -4.76544052e-01 -2.44056001e-01 3.95626247e-01
2.71490514e-01 -6.91273451e-01 -1.14771712e+00 -3.73647898e-01
4.18361425e-01 -5.81110775e-01 9.00034249e-01 1.84971899e-01
7.46739686e-01 -2.90720239e-02 -4.05392885e-01 5.65238297e-01
1.49235392e+00 1.82514250e-01 5.03212094e-01 5.64769566e-01
3.80231172e-01 9.23195243e-01 9.51111555e-01 3.62453282e-01
1.90874953e-02 7.76065350e-01 9.69414562e-02 3.17861080e-01
-4.35112305e-02 -3.09552968e-01 3.19932401e-01 3.49607378e-01
4.20124650e-01 2.18336180e-01 -9.19842064e-01 6.50896192e-01
-1.79416680e+00 -1.12128711e+00 6.40987873e-01 2.73182654e+00
8.94737899e-01 1.15352310e-01 1.43689960e-01 1.62670225e-01
9.21305180e-01 2.00051799e-01 -7.79974878e-01 -1.34495154e-01
4.09433097e-02 -3.02826744e-02 4.88822132e-01 3.33577305e-01
-1.45522797e+00 8.46970439e-01 5.60032606e+00 1.11322212e+00
-1.10274458e+00 6.51457012e-02 2.87442505e-01 -1.82111204e-01
-9.33545157e-02 1.01244219e-01 -6.49542630e-01 4.37258661e-01
9.36971486e-01 -5.47511816e-01 3.58001143e-01 1.23381269e+00
-3.42094660e-01 -9.70820487e-02 -1.36156988e+00 1.09655571e+00
3.72820199e-01 -9.83368278e-01 -6.98373001e-03 1.72067851e-01
6.10385239e-01 -5.56730218e-02 9.98880565e-02 7.30729759e-01
3.66678715e-01 -7.96175838e-01 3.72316927e-01 3.61119360e-01
8.83081079e-01 -7.26829827e-01 5.30550063e-01 4.47529197e-01
-1.05813277e+00 -2.16976315e-01 -9.37137783e-01 2.27045402e-01
-2.33565748e-01 4.29333895e-01 -9.61456776e-01 3.93707126e-01
5.54475605e-01 9.05148506e-01 -4.35918391e-01 8.94427538e-01
2.11652264e-01 -5.45101799e-02 -4.78073247e-02 2.39762872e-01
2.37735569e-01 -5.33858538e-01 6.74193382e-01 1.05850947e+00
3.00862402e-01 -1.20142184e-01 1.81006923e-01 6.20899558e-01
8.38607699e-02 1.92737207e-01 -1.21769905e+00 2.60692507e-01
5.22385836e-01 1.08363879e+00 -4.77943689e-01 -4.50855672e-01
-5.73701262e-01 1.28464687e+00 4.64906991e-01 3.98721635e-01
-7.36549973e-01 -6.59259558e-01 7.32607543e-01 1.68437809e-01
4.74807322e-01 1.87972531e-01 -9.76238102e-02 -1.14264083e+00
-1.14029553e-02 -6.48051500e-01 6.34545863e-01 -5.21013558e-01
-1.66052997e+00 4.97048855e-01 3.11729729e-01 -1.86232245e+00
-4.55407828e-01 -4.12948310e-01 -4.43139553e-01 1.04720521e+00
-1.40763664e+00 -9.84190524e-01 -2.83393301e-02 8.56214702e-01
6.40612483e-01 -4.29095358e-01 1.08669412e+00 6.99508116e-02
-5.68246543e-02 5.08566260e-01 6.39482558e-01 2.10073460e-02
1.28397191e+00 -1.21593010e+00 -1.05514251e-01 4.60460335e-01
8.77633467e-02 6.00589335e-01 7.85127938e-01 -5.60912132e-01
-1.07873833e+00 -1.17304862e+00 6.10406339e-01 -5.93680978e-01
8.70074987e-01 -4.57251459e-01 -1.23219073e+00 4.29996371e-01
1.38951233e-02 4.44156796e-01 8.56501162e-01 8.18954110e-02
-8.71173859e-01 -1.43530652e-01 -1.28327560e+00 4.02161628e-01
9.06836450e-01 -8.77708793e-01 -1.11387837e+00 2.86287308e-01
3.83490413e-01 1.31164610e-01 -8.01766038e-01 2.50077516e-01
4.19914931e-01 -8.44775736e-01 1.10963571e+00 -6.95559919e-01
1.46670207e-01 -9.41344202e-02 -5.01571894e-01 -1.67492485e+00
-4.72703993e-01 -1.95210636e-01 8.08484256e-02 1.17223597e+00
7.51191378e-02 -4.08270419e-01 9.02698278e-01 5.41481316e-01
2.04911038e-01 -5.46945632e-01 -1.24713337e+00 -1.27340603e+00
4.43645746e-01 -2.56760925e-01 6.79765046e-02 1.18129599e+00
3.58027428e-01 6.64665222e-01 -2.42757618e-01 -1.59095794e-01
9.78387535e-01 8.84045660e-02 6.29990637e-01 -1.57139218e+00
-1.95873469e-01 -1.87539056e-01 -5.72092712e-01 -7.40386128e-01
2.83431768e-01 -1.05194294e+00 2.33906046e-01 -1.05245674e+00
3.26210529e-01 -4.11342651e-01 -5.37285686e-01 1.80502102e-01
8.58243629e-02 2.30271351e-02 1.54865995e-01 4.32049513e-01
-7.28671253e-01 8.39284718e-01 9.01340008e-01 -2.21546471e-01
-2.93755293e-01 -1.48119584e-01 -6.34990275e-01 7.01534152e-01
5.34110904e-01 -7.22523391e-01 -8.01625013e-01 1.39613211e-01
-3.81436437e-01 -1.42652571e-01 1.65752232e-01 -1.10574460e+00
2.21473470e-01 -2.19132677e-01 5.86463511e-01 -6.38697147e-02
7.04848230e-01 -1.02483284e+00 -1.11315288e-01 2.83674538e-01
-6.11915946e-01 -7.75018990e-01 -2.11652935e-01 7.81800330e-01
-3.42704147e-01 -3.87410611e-01 1.26668012e+00 -4.76960242e-02
-8.60423505e-01 1.37923330e-01 -3.30265164e-02 3.72517735e-01
1.59680474e+00 -7.59847701e-01 1.51858581e-02 -2.50437379e-01
-7.90944219e-01 -1.10947505e-01 7.04793990e-01 6.37894511e-01
6.02246642e-01 -1.64461792e+00 -4.65565950e-01 3.04687202e-01
8.77608895e-01 -4.02655602e-01 2.06993863e-01 5.42487264e-01
2.26041123e-01 3.39437008e-01 -4.62918013e-01 -7.14392424e-01
-1.00456047e+00 1.06793320e+00 2.71429747e-01 -2.11027145e-01
-6.12718403e-01 7.38899767e-01 5.37198186e-01 -6.62263036e-01
2.62005329e-01 3.01482469e-01 -2.43525475e-01 5.85938096e-01
5.86032629e-01 4.14690435e-01 -5.42687885e-02 -8.19221854e-01
-5.85721791e-01 7.48763800e-01 -1.83015347e-01 -2.27226108e-01
1.23789144e+00 -2.89045453e-01 4.96150225e-01 6.65111005e-01
1.85565341e+00 -2.96223730e-01 -1.46737659e+00 -7.61569262e-01
4.15871888e-01 -7.97898889e-01 6.39967769e-02 -5.47383130e-01
-5.32534301e-01 8.41524839e-01 8.04640114e-01 2.82554291e-02
1.03843510e+00 2.44688272e-01 1.94978848e-01 2.64508843e-01
4.30145949e-01 -1.51561928e+00 2.86479235e-01 6.40738070e-01
1.03426933e+00 -1.48156691e+00 -1.48141505e-02 -2.26183578e-01
-1.00550294e+00 1.08493948e+00 5.36907673e-01 -4.07178402e-01
8.37276101e-01 -1.81717902e-01 9.79073271e-02 -3.02160326e-02
-5.56956530e-01 -3.72360796e-01 4.61668193e-01 1.14543402e+00
1.67570055e-01 -2.21492171e-01 1.38652563e-01 2.71792561e-01
1.03275843e-01 -6.59353435e-02 1.90209329e-01 7.90349722e-01
-4.87258434e-01 -1.04642248e+00 -4.66505527e-01 3.88783038e-01
5.91472238e-02 5.53475730e-02 -3.49885702e-01 8.12989652e-01
6.87637255e-02 7.53172398e-01 1.21187583e-01 -3.52982044e-01
3.17528993e-01 3.80075455e-01 3.13406348e-01 -8.96971524e-01
4.84485887e-02 9.22388807e-02 -1.64292499e-01 -5.73540568e-01
-1.60610452e-01 -6.42948389e-01 -8.18342566e-01 8.62780884e-02
-3.69418412e-01 2.09656745e-01 5.53840399e-01 8.72094154e-01
2.79958129e-01 3.99346240e-02 7.75839746e-01 -9.11701977e-01
-1.15355527e+00 -8.73611391e-01 -7.91650891e-01 8.51064384e-01
1.74593925e-01 -1.14111435e+00 -6.17789626e-01 -7.97529817e-02] | [10.01702880859375, 2.8964693546295166] |
76dc5399-b629-4c6e-91fa-553cf8bce504 | complementary-bi-directional-feature | 2207.02437 | null | https://arxiv.org/abs/2207.02437v1 | https://arxiv.org/pdf/2207.02437v1.pdf | Complementary Bi-directional Feature Compression for Indoor 360° Semantic Segmentation with Self-distillation | Recently, horizontal representation-based panoramic semantic segmentation approaches outperform projection-based solutions, because the distortions can be effectively removed by compressing the spherical data in the vertical direction. However, these methods ignore the distortion distribution prior and are limited to unbalanced receptive fields, e.g., the receptive fields are sufficient in the vertical direction and insufficient in the horizontal direction. Differently, a vertical representation compressed in another direction can offer implicit distortion prior and enlarge horizontal receptive fields. In this paper, we combine the two different representations and propose a novel 360{\deg} semantic segmentation solution from a complementary perspective. Our network comprises three modules: a feature extraction module, a bi-directional compression module, and an ensemble decoding module. First, we extract multi-scale features from a panorama. Then, a bi-directional compression module is designed to compress features into two complementary low-dimensional representations, which provide content perception and distortion prior. Furthermore, to facilitate the fusion of bi-directional features, we design a unique self distillation strategy in the ensemble decoding module to enhance the interaction of different features and further improve the performance. Experimental results show that our approach outperforms the state-of-the-art solutions with at least 10\% improvement on quantitative evaluations while displaying the best performance on visual appearance. | ['Yao Zhao', 'Zhijie Shen', 'Kang Liao', 'Lang Nie', 'Chunyu Lin', 'Zishuo Zheng'] | 2022-07-06 | null | null | null | null | ['feature-compression'] | ['computer-vision'] | [ 6.61911488e-01 1.30067766e-01 -1.03949390e-01 -2.83419788e-01
-4.54683751e-01 -2.51139045e-01 3.82578254e-01 -2.24755079e-01
-3.60336870e-01 2.67003775e-01 3.69469285e-01 8.10785592e-02
-5.27626835e-02 -1.03240407e+00 -6.77489996e-01 -1.02664173e+00
4.48321313e-01 3.66559066e-02 6.35034323e-01 -2.31641144e-01
3.70774418e-01 2.75503725e-01 -1.73100197e+00 3.54789764e-01
1.21055603e+00 1.23896050e+00 6.05249286e-01 2.14586750e-01
-1.69960141e-01 2.90055573e-01 -1.75049230e-01 -1.28755346e-01
3.58114034e-01 -4.26334947e-01 -4.87799078e-01 2.63804883e-01
2.91893512e-01 -4.04230714e-01 -5.24546027e-01 1.22384751e+00
4.83003736e-01 -7.74436072e-02 4.16873842e-01 -6.77602112e-01
-5.07087350e-01 4.36693251e-01 -8.73294950e-01 -1.22794241e-01
1.71435550e-01 -1.16690665e-01 7.42509127e-01 -7.58631229e-01
6.48561060e-01 1.07177079e+00 4.23373133e-01 4.50097144e-01
-1.03169703e+00 -6.47062719e-01 2.05203876e-01 2.44758397e-01
-1.19947374e+00 -3.01639438e-01 1.07016027e+00 -1.19842753e-01
6.35184109e-01 4.83972758e-01 7.15939760e-01 6.04322970e-01
7.59237483e-02 9.52105165e-01 1.20357466e+00 -3.62621158e-01
1.56429350e-01 -8.57044458e-02 -1.83360875e-01 5.59732318e-01
5.49645908e-02 -7.76625052e-02 -3.24152172e-01 1.55911803e-01
8.55720878e-01 2.69819796e-01 -5.56019127e-01 -5.00214517e-01
-1.11227953e+00 7.69271553e-01 7.07362354e-01 4.08744961e-01
-4.12524462e-01 -1.14757158e-01 1.14303745e-01 -1.23947993e-01
4.03623641e-01 2.86873668e-01 -1.98068693e-01 1.83828413e-01
-1.20097220e+00 1.96177304e-01 3.56094778e-01 9.49033797e-01
7.76936412e-01 -1.59920141e-01 -2.15094373e-01 1.11928642e+00
2.68887430e-01 4.17281181e-01 8.67477834e-01 -8.59754324e-01
6.96473658e-01 6.43954635e-01 -2.79053181e-01 -1.13980901e+00
-3.29885930e-01 -3.42156112e-01 -9.49250340e-01 -8.54070708e-02
5.52032851e-02 1.67521805e-01 -1.15231752e+00 1.52945089e+00
2.03801081e-01 -3.69617604e-02 1.61494757e-03 1.02824926e+00
7.81358659e-01 9.07866240e-01 -3.51385146e-01 -4.30103630e-01
1.41932309e+00 -1.23587298e+00 -7.79585242e-01 -1.45142004e-01
1.72340274e-01 -9.84338164e-01 7.86984086e-01 5.17976701e-01
-1.32713020e+00 -4.34629411e-01 -1.34683120e+00 -1.79925889e-01
-2.24995136e-01 2.36470670e-01 4.74951118e-01 4.82515663e-01
-8.12250137e-01 4.13987130e-01 -7.49791443e-01 -1.05275795e-01
4.29024696e-01 2.75859684e-01 -9.19446424e-02 -2.99764723e-01
-1.01614583e+00 4.50937510e-01 4.58162993e-01 -5.18474467e-02
-3.51690263e-01 -3.06697696e-01 -9.30189669e-01 2.10840166e-01
3.01366717e-01 -4.81383979e-01 8.71110559e-01 -6.68739617e-01
-1.42552185e+00 5.15302002e-01 -2.49238789e-01 -1.75402075e-01
4.88338947e-01 -1.51582211e-02 -3.81546021e-01 5.10716617e-01
-1.23103661e-02 9.37759697e-01 8.41959238e-01 -1.31916809e+00
-7.95234919e-01 -5.57351828e-01 -1.82524145e-01 7.21781671e-01
-7.45087504e-01 -2.90570527e-01 -1.05490863e+00 -9.37420964e-01
9.90871012e-01 -7.29606748e-01 -3.12005818e-01 -1.35740772e-01
-3.82458359e-01 2.30162591e-01 1.03894615e+00 -8.45610321e-01
1.47800112e+00 -2.25984216e+00 3.56073290e-01 2.92994767e-01
2.15334564e-01 1.84658840e-01 7.28252232e-02 3.94743159e-02
1.18530907e-01 -8.69009495e-02 -6.75379574e-01 -2.84450531e-01
-2.79713154e-01 1.95624188e-01 -4.42011476e-01 3.38581979e-01
-7.59610832e-02 6.10104680e-01 -5.38115442e-01 -6.23374820e-01
3.40607136e-01 5.03961325e-01 -8.31546307e-01 -7.56912399e-03
1.03982039e-01 3.55882406e-01 -5.10793209e-01 6.31491899e-01
1.25090265e+00 6.78046942e-02 1.31170198e-01 -2.31232151e-01
-3.78647953e-01 1.32578626e-01 -1.21076500e+00 1.94497299e+00
-3.55382323e-01 2.48877987e-01 1.01104982e-01 -1.02130973e+00
9.88853574e-01 6.82513565e-02 5.60390472e-01 -9.36060488e-01
1.51761696e-01 3.78177434e-01 -1.22646086e-01 -2.51448005e-01
7.29504168e-01 2.59290379e-03 -5.72916418e-02 2.87502766e-01
-6.44619316e-02 -5.12514651e-01 7.76619241e-02 -2.39815526e-02
5.12713969e-01 -3.09214965e-02 3.66576649e-02 -7.10292011e-02
7.37631381e-01 -4.32806671e-01 8.88578713e-01 1.65233389e-01
1.25807021e-02 1.18326831e+00 2.11885050e-01 -1.98139161e-01
-9.58531201e-01 -8.31800342e-01 -4.15296704e-01 7.05244005e-01
9.19531226e-01 -5.14026225e-01 -1.06679463e+00 -4.82535243e-01
-3.68597597e-01 5.06573915e-01 -4.92411256e-01 -2.19491377e-01
-8.08632493e-01 -9.53450084e-01 2.62537360e-01 6.54530644e-01
1.02739501e+00 -8.64673316e-01 -8.06930721e-01 8.93828124e-02
-4.93989885e-01 -1.01528001e+00 -4.91428941e-01 -6.06804192e-02
-1.10519516e+00 -6.41460538e-01 -1.05495250e+00 -8.92956555e-01
7.89113343e-01 5.85994720e-01 4.06236529e-01 -1.99099764e-01
-1.04635082e-01 -8.33090767e-02 -4.08566475e-01 -2.17894763e-01
7.21228942e-02 3.76813114e-02 -2.96431452e-01 -4.03374322e-02
2.05549181e-01 -8.38549256e-01 -1.05653071e+00 5.95501900e-01
-1.23600280e+00 5.38616240e-01 9.14565802e-01 7.82659113e-01
8.91319811e-01 2.64150023e-01 1.15804687e-01 -5.53125203e-01
1.94902405e-01 -2.95293719e-01 -3.88845205e-01 1.14421889e-01
-6.33587658e-01 -6.03082478e-02 6.46643639e-01 -2.43031546e-01
-1.23847961e+00 1.26440704e-01 -2.65699863e-01 -3.74387324e-01
-3.66960838e-02 1.45841718e-01 -5.70384145e-01 -5.91755472e-02
2.95541435e-01 8.07438314e-01 -2.13102147e-01 -5.85887134e-01
4.45185870e-01 7.27794945e-01 5.56680679e-01 -2.88041472e-01
6.41514659e-01 7.72561252e-01 -2.55217582e-01 -8.01486731e-01
-3.42154801e-01 -5.47077596e-01 -5.41364849e-01 -4.18177843e-02
9.26836550e-01 -7.75529087e-01 -1.02595173e-01 7.08871186e-01
-9.08009708e-01 2.48422042e-01 -4.06877697e-01 5.53220153e-01
-7.49923050e-01 6.27835810e-01 -5.58364689e-01 -3.46889168e-01
-3.58880252e-01 -1.39716542e+00 1.09628057e+00 6.30801737e-01
2.24019006e-01 -3.28356683e-01 -1.54979497e-01 4.92476732e-01
2.76836187e-01 -1.08188726e-01 8.49902809e-01 -2.31313080e-01
-6.74423277e-01 -1.97159350e-02 -4.73813206e-01 3.77080947e-01
-9.45185944e-02 -3.57880920e-01 -9.83820379e-01 -2.77261972e-01
2.92706788e-01 -1.31715268e-01 1.34359944e+00 5.23260951e-01
1.73549294e+00 -2.86161125e-01 -4.30421144e-01 1.13862395e+00
1.28312767e+00 3.10416669e-01 1.01463544e+00 3.63495737e-01
8.23598742e-01 7.63203144e-01 5.53583264e-01 2.20700741e-01
3.82443428e-01 7.68546999e-01 4.67870712e-01 -1.17638126e-01
-2.87579447e-01 -3.30973774e-01 1.55124679e-01 1.05201793e+00
-2.90086865e-01 -1.30010977e-01 -5.00293911e-01 3.89150769e-01
-1.64037287e+00 -7.89830148e-01 1.27883106e-01 2.25644851e+00
8.45873356e-01 1.58062428e-02 -2.34505504e-01 3.87314081e-01
7.32318342e-01 5.12936652e-01 -4.86232311e-01 -3.58614475e-01
-3.40030074e-01 5.53340241e-02 6.49797201e-01 1.95720568e-01
-1.15599620e+00 8.57422352e-01 5.70849228e+00 1.31985235e+00
-1.30722654e+00 3.81962582e-02 5.59017658e-01 -1.52263818e-02
-7.38144577e-01 -5.65807056e-03 -6.21403039e-01 6.29228473e-01
1.60019666e-01 2.02104017e-01 2.65075356e-01 7.38281012e-01
-3.10356349e-01 -1.77355215e-01 -5.35797000e-01 1.07884622e+00
3.08122456e-01 -1.26053715e+00 1.42151773e-01 3.00760746e-01
8.78210247e-01 -1.51373893e-01 2.50578344e-01 -2.31081080e-02
-3.17879170e-01 -8.50247681e-01 7.49100804e-01 5.39447784e-01
1.06763518e+00 -9.08137321e-01 5.56467533e-01 4.41443115e-01
-1.26511216e+00 -2.33400509e-01 -4.76128638e-01 2.42808789e-01
3.18645775e-01 6.41324937e-01 -1.24849267e-01 9.33434129e-01
7.76746750e-01 6.76642716e-01 -3.68166029e-01 1.03394103e+00
-1.40041962e-01 3.00086319e-01 -3.56105864e-01 3.83094698e-01
2.62410522e-01 -4.58392739e-01 6.17978990e-01 1.06649613e+00
5.07396817e-01 3.50723505e-01 -8.78243968e-02 7.62993813e-01
6.10149875e-02 2.08970591e-01 -3.92150760e-01 2.88997501e-01
3.12343448e-01 1.26751530e+00 -9.04453278e-01 -2.98598796e-01
-4.98782247e-01 1.24877536e+00 -1.13668181e-02 3.81594658e-01
-6.81212664e-01 -5.87474108e-01 2.22149968e-01 8.11037719e-02
6.19509041e-01 -7.29450723e-03 -6.72155857e-01 -1.32780778e+00
2.67638862e-01 -6.01073503e-01 2.48999655e-01 -5.15306532e-01
-6.73269689e-01 6.28588378e-01 -6.43155351e-02 -1.53114593e+00
-1.39298409e-01 -2.94913054e-01 -5.88312447e-01 8.59584391e-01
-1.52861774e+00 -1.17641664e+00 -4.01057243e-01 5.10582089e-01
7.12141871e-01 3.96568365e-02 5.84664226e-01 3.67443383e-01
-4.91093129e-01 5.72078168e-01 2.80228883e-01 -1.35192677e-01
5.02895892e-01 -9.18346763e-01 -9.67468414e-03 1.00650609e+00
-5.85310161e-02 4.02626425e-01 3.40101182e-01 -5.69623828e-01
-1.09398186e+00 -9.20396447e-01 6.24962330e-01 2.62848020e-01
-6.62272722e-02 -2.71301508e-01 -9.30895209e-01 2.24268839e-01
-9.12894979e-02 -5.50804548e-02 5.97385347e-01 -4.22037333e-01
-2.36880571e-01 -1.38850853e-01 -1.13477755e+00 5.48265696e-01
1.00844705e+00 -3.05808246e-01 -7.97719479e-01 -1.01714402e-01
6.24031067e-01 -4.82784063e-01 -5.95566213e-01 8.05373073e-01
6.84926867e-01 -1.24163604e+00 1.04239154e+00 2.46454343e-01
8.23393881e-01 -4.17806208e-01 -4.34596390e-01 -1.05627871e+00
-3.02306235e-01 -3.17627430e-01 -1.30048975e-01 1.08121431e+00
3.90330963e-02 -6.20382428e-01 6.64498389e-01 5.45156784e-02
-4.05502558e-01 -1.21463740e+00 -8.77811134e-01 -3.90247375e-01
1.25922756e-02 -3.28325570e-01 7.27165639e-01 7.65815496e-01
-6.65009394e-03 1.08032906e-03 -3.09645504e-01 -5.56225814e-02
4.48429137e-01 5.92067719e-01 3.35951924e-01 -9.06072795e-01
-2.43451729e-01 -7.33889580e-01 -4.28426504e-01 -1.61862123e+00
-2.96055555e-01 -7.33471155e-01 1.42879158e-01 -1.52741492e+00
4.23284203e-01 -5.46092689e-01 -2.52450377e-01 8.29166770e-02
-1.47219449e-01 4.47242856e-01 2.55292207e-01 5.93538284e-01
-1.97849780e-01 9.31425869e-01 1.62970686e+00 1.94988735e-02
-2.38120049e-01 -4.19154763e-02 -7.73691595e-01 1.09535801e+00
6.21888280e-01 -1.34350464e-01 -5.27319670e-01 -5.93289614e-01
-1.56927988e-01 -2.46724024e-01 6.86126575e-02 -1.19977272e+00
2.20035627e-01 3.64862271e-02 6.27117693e-01 -9.52882349e-01
5.54401636e-01 -7.38560975e-01 -3.49402189e-01 3.24456990e-01
-5.25999665e-02 -4.30695325e-01 5.71445674e-02 6.39388561e-01
-4.96806532e-01 -6.43340051e-02 7.60452688e-01 1.30239585e-02
-7.20220983e-01 3.52143645e-01 -9.93310586e-02 -3.35524470e-01
1.16291630e+00 -6.38215840e-01 -2.85703272e-01 -1.31356493e-01
-4.05282378e-01 2.70607114e-01 7.54925370e-01 3.82743806e-01
9.12761986e-01 -1.22493732e+00 -4.61615711e-01 6.37463987e-01
1.43208634e-02 4.49605376e-01 6.47935927e-01 9.39972401e-01
-6.60401762e-01 4.72752362e-01 -5.57311654e-01 -6.77772284e-01
-1.03063440e+00 4.04822379e-01 9.43977444e-04 -1.50448367e-01
-7.36155033e-01 9.51727688e-01 8.46486568e-01 -3.72318119e-01
1.29836276e-01 -2.93978333e-01 -6.49347186e-01 1.75175279e-01
5.45838952e-01 4.29895639e-01 2.16706432e-02 -9.55245435e-01
-1.68655708e-01 1.08145988e+00 -2.39627212e-01 -3.33122879e-01
1.30231130e+00 -3.52645904e-01 -1.18182346e-01 1.22643873e-01
1.10191131e+00 1.21702895e-01 -1.44255328e+00 -2.65514135e-01
-6.16101384e-01 -7.15312719e-01 2.55349189e-01 -4.88045514e-01
-1.35853434e+00 1.12359703e+00 6.93000376e-01 1.39104566e-02
1.71004367e+00 3.19778025e-02 1.27373528e+00 -2.47688949e-01
1.59522727e-01 -1.31614685e+00 -3.95993292e-02 1.62971184e-01
8.79851580e-01 -8.30833793e-01 1.46113738e-01 -9.55043137e-01
-7.23586857e-01 1.12014747e+00 6.80287540e-01 -4.80588563e-02
5.08303523e-01 9.80488956e-02 -2.60176569e-01 -1.16542831e-01
-1.96209833e-01 -5.78861423e-02 4.96600330e-01 3.37861270e-01
1.35138810e-01 1.12401523e-01 -7.17376173e-01 7.79102504e-01
-3.69800746e-01 -2.83234328e-01 1.90049186e-01 7.51227260e-01
-5.45445859e-01 -9.34960127e-01 -4.49611723e-01 3.82531494e-01
-2.89287359e-01 -1.05989560e-01 2.94941161e-02 4.49228823e-01
5.90705514e-01 6.96000695e-01 3.08311790e-01 -6.23054326e-01
3.04230005e-01 -2.76201993e-01 4.04518962e-01 -3.20513457e-01
-7.91913867e-02 6.23340547e-01 -3.20388764e-01 -7.50516295e-01
-3.23182762e-01 -6.46718264e-01 -1.35889685e+00 3.49401525e-04
-4.14743662e-01 -5.31536676e-02 8.07804525e-01 7.72000611e-01
2.49394715e-01 6.40460908e-01 7.87520826e-01 -1.01963830e+00
-3.06279510e-01 -8.82815301e-01 -5.42007983e-01 2.37677485e-01
-7.30909780e-03 -6.27342761e-01 -2.17534378e-01 -2.62093358e-03] | [10.942795753479004, -1.6325080394744873] |
64e54f6e-6783-41b2-aafe-4aa57b34f459 | wganvo-monocular-visual-odometry-based-on | 2007.13704 | null | https://arxiv.org/abs/2007.13704v1 | https://arxiv.org/pdf/2007.13704v1.pdf | WGANVO: Monocular Visual Odometry based on Generative Adversarial Networks | In this work we present WGANVO, a Deep Learning based monocular Visual Odometry method. In particular, a neural network is trained to regress a pose estimate from an image pair. The training is performed using a semi-supervised approach. Unlike geometry based monocular methods, the proposed method can recover the absolute scale of the scene without neither prior knowledge nor extra information. The evaluation of the system is carried out on the well-known KITTI dataset where it is shown to work in real time and the accuracy obtained is encouraging to continue the development of Deep Learning based methods. | ['Taihú Pire', 'Javier Cremona', 'Lucas Uzal'] | 2020-07-27 | null | null | null | null | ['monocular-visual-odometry'] | ['robots'] | [-3.42919618e-01 2.36655086e-01 -1.10791415e-01 -5.03641367e-01
-5.38989343e-03 -3.51856470e-01 7.93819904e-01 -5.37141144e-01
-5.12092650e-01 7.78863430e-01 -1.42498568e-01 1.86571702e-02
2.93768764e-01 -5.01501620e-01 -7.47025847e-01 -4.02716488e-01
1.96354032e-01 9.94416833e-01 1.27433375e-01 -1.14316337e-01
2.47071967e-01 9.55030739e-01 -1.28981018e+00 -5.43246508e-01
2.76704371e-01 9.49089587e-01 2.79126883e-01 7.37184405e-01
5.10542810e-01 9.62734044e-01 -3.16764504e-01 9.83052179e-02
7.60984898e-01 -1.31229147e-01 -6.64850771e-01 2.22976029e-01
1.10620093e+00 -5.74889123e-01 -8.33010852e-01 8.39951873e-01
6.58162653e-01 -2.15851516e-01 6.66676939e-01 -8.79299283e-01
1.88731737e-02 -3.62510026e-01 -2.78238177e-01 -6.37157336e-02
4.72052276e-01 1.74037635e-01 7.02723026e-01 -1.11419499e+00
1.17712736e+00 1.00290716e+00 8.09604764e-01 7.07968324e-02
-1.04666400e+00 -2.03537241e-01 -5.13358176e-01 2.05993131e-01
-1.53153241e+00 -3.08861583e-01 9.19020057e-01 -6.74507618e-01
1.15093863e+00 -4.54703599e-01 7.68864870e-01 4.75469798e-01
4.28759903e-01 4.91084665e-01 1.30055141e+00 -6.05583966e-01
1.76255792e-01 5.69408434e-03 -2.93158889e-01 9.86020982e-01
5.48061609e-01 6.69370592e-01 -6.59727454e-01 3.51407349e-01
1.16334021e+00 -7.92581514e-02 -7.14390352e-02 -1.32976198e+00
-1.03726864e+00 8.27747643e-01 1.09852862e+00 7.21976534e-02
-3.00742567e-01 4.83301193e-01 1.34794101e-01 3.96033913e-01
4.88122910e-01 3.95555377e-01 -2.88150847e-01 3.63049246e-02
-1.04438829e+00 1.05788745e-01 6.74081683e-01 9.19298053e-01
1.08320701e+00 3.94221544e-01 7.35325456e-01 3.79982233e-01
5.50451636e-01 6.55989587e-01 3.80337298e-01 -1.08325171e+00
2.25164667e-01 5.59939086e-01 2.02224791e-01 -9.68786180e-01
-8.76418948e-01 -2.91111887e-01 -4.02353644e-01 9.56579685e-01
2.97133118e-01 -5.02940238e-01 -1.11109531e+00 1.13623798e+00
2.59640723e-01 -5.23514003e-02 -1.10932954e-01 1.22021723e+00
8.87499928e-01 1.42880931e-01 -6.87360883e-01 3.01429749e-01
6.66969299e-01 -7.87580252e-01 -4.84484732e-01 -6.46813452e-01
5.59129357e-01 -9.28394437e-01 2.75781304e-01 4.31452811e-01
-8.36870372e-01 -6.25174046e-01 -1.62185752e+00 -3.43784273e-01
-3.20930421e-01 6.08675480e-01 5.58851957e-01 4.55431998e-01
-1.17547822e+00 6.08396888e-01 -9.54697311e-01 -6.60731196e-01
-1.48108611e-02 6.49875581e-01 -6.64593697e-01 -8.73239785e-02
-7.74009049e-01 1.32458925e+00 6.53930008e-01 2.36057773e-01
-7.58822560e-01 -1.54968515e-01 -1.28461623e+00 -5.19510090e-01
2.47766115e-02 -8.60705853e-01 1.25067914e+00 -7.76629567e-01
-1.97518826e+00 1.37069559e+00 -1.67235926e-01 -7.46544063e-01
7.68707275e-01 -4.01104182e-01 4.39366437e-02 3.04885149e-01
-1.34537622e-01 9.93333399e-01 8.36991608e-01 -1.21438396e+00
-4.47320074e-01 -4.21453506e-01 1.66118369e-01 5.03616989e-01
5.90532362e-01 -4.23707813e-01 -2.90933371e-01 -2.79753171e-02
8.71872485e-01 -1.29085100e+00 -1.28798515e-01 2.09585160e-01
-2.54132748e-01 2.94541150e-01 6.69189990e-01 -5.24159253e-01
1.98708907e-01 -1.67490745e+00 2.39025071e-01 -1.29072070e-02
2.86969244e-01 2.66357005e-01 3.98251384e-01 3.69407386e-01
4.81073968e-02 -8.10713530e-01 2.31959686e-01 -3.84965479e-01
-2.21811369e-01 3.35909516e-01 -3.67900245e-02 1.15869105e+00
-1.34870768e-01 9.74474788e-01 -6.73103988e-01 -1.37184948e-01
8.85496676e-01 3.84546369e-01 -4.36368167e-01 3.55572850e-01
-1.11327119e-01 6.16928339e-01 2.05800042e-01 4.86497611e-01
8.67520213e-01 -1.05509005e-01 1.84744388e-01 -2.95717537e-01
-2.57865191e-01 6.21026635e-01 -1.24618685e+00 2.00503755e+00
-5.49797714e-01 1.32474172e+00 -4.30073999e-02 -9.66510892e-01
1.20159113e+00 2.12814495e-01 2.33062223e-01 -8.19658756e-01
4.10441071e-01 4.87472415e-01 -1.02886640e-01 -2.51617700e-01
5.41359544e-01 -2.85292119e-01 2.49028787e-01 3.92116606e-02
4.10447389e-01 -8.10479224e-01 -4.15682495e-02 -1.45429626e-01
6.91506088e-01 7.84477234e-01 7.18852937e-01 -2.63262153e-01
5.58879137e-01 4.68547553e-01 1.45530537e-01 2.77197301e-01
-2.45640144e-01 7.55418301e-01 -1.19895302e-01 -8.90734017e-01
-1.35488653e+00 -9.34105575e-01 -2.47377932e-01 9.83048826e-02
4.78069901e-01 1.23619452e-01 -3.35854739e-01 -3.61410737e-01
2.71848232e-01 1.06078424e-01 -3.93307000e-01 1.50704652e-01
-6.51435912e-01 -1.82880238e-01 2.73669779e-01 6.40785336e-01
6.57768011e-01 -7.99537897e-01 -8.97512853e-01 1.41322255e-01
2.29389489e-01 -1.38107502e+00 -3.45407352e-02 2.23780036e-01
-1.06164527e+00 -1.23852551e+00 -6.82133198e-01 -9.32657778e-01
5.36617219e-01 3.67163718e-01 9.54928935e-01 -4.10431415e-01
-2.62463450e-01 2.72166342e-01 2.11910427e-01 -3.89208049e-01
-4.54009604e-03 1.23935759e-01 2.93413192e-01 -5.44138432e-01
5.66527605e-01 -6.47425532e-01 -5.50686717e-01 3.35916281e-01
-2.46168390e-01 2.55041923e-02 5.50208807e-01 7.75034487e-01
3.33982140e-01 -7.53571033e-01 -9.92483050e-02 -4.65474010e-01
-1.77648768e-01 -2.30465792e-02 -1.21606612e+00 -4.36131179e-01
-6.19557858e-01 1.78373784e-01 1.97452605e-01 2.42014206e-03
-8.87326479e-01 7.18280733e-01 -1.53188789e-02 -4.39298332e-01
-3.51779431e-01 5.68339646e-01 2.58066654e-01 -7.14153111e-01
9.77367997e-01 -2.52603218e-02 1.95449725e-01 -4.54972953e-01
2.49957129e-01 4.70862955e-01 7.60060728e-01 3.69887166e-02
1.13765442e+00 1.02934325e+00 5.65849602e-01 -1.01229656e+00
-7.04066396e-01 -7.57781863e-01 -1.47111046e+00 -1.76743463e-01
7.27156281e-01 -1.50415587e+00 -6.63467705e-01 5.66431999e-01
-1.18838477e+00 -4.13484126e-01 3.00917495e-03 1.10311890e+00
-9.17205811e-01 2.87882924e-01 -4.81174409e-01 -5.24972677e-01
-1.44785091e-01 -8.64482343e-01 1.14355016e+00 2.20412746e-01
-1.55824974e-01 -1.27285278e+00 7.12426901e-01 1.95274711e-01
-3.35098244e-02 3.32488745e-01 -2.82850955e-02 -1.41677365e-01
-1.00233757e+00 -4.57120985e-01 -3.53182375e-01 1.43986076e-01
9.98523384e-02 -3.41563582e-01 -1.09147739e+00 -5.05819798e-01
-8.33774544e-03 -6.68266177e-01 5.84103107e-01 4.13377881e-01
9.06381905e-02 3.43383342e-01 -1.61168635e-01 9.75286782e-01
1.80663347e+00 5.12173846e-02 5.19547760e-01 5.94852924e-01
9.99225140e-01 2.91372329e-01 5.46317577e-01 1.06545158e-01
4.01115954e-01 1.05664742e+00 7.36877859e-01 -2.66963094e-01
-2.40515068e-01 -3.65941852e-01 1.17111094e-01 6.17278814e-01
-1.50961637e-01 4.42682892e-01 -1.09119916e+00 4.91282701e-01
-1.65147948e+00 -6.02743804e-01 -1.54524252e-01 2.12344050e+00
3.94709319e-01 1.23097420e-01 -1.51773855e-01 8.47279504e-02
1.36329010e-01 2.62419611e-01 -3.84931147e-01 -4.33204383e-01
-9.39192846e-02 1.60743490e-01 9.65899348e-01 9.60202873e-01
-1.20679533e+00 1.20763206e+00 7.05561447e+00 -9.48410332e-02
-1.70251381e+00 -1.30523592e-01 -1.50014058e-01 8.75607207e-02
4.22792941e-01 1.39104396e-01 -8.19166541e-01 -2.36951932e-01
5.92613518e-01 -2.94285119e-02 3.52730691e-01 1.13964355e+00
9.44811106e-02 -5.42083979e-01 -1.01387882e+00 1.20879221e+00
4.23729807e-01 -1.39791715e+00 -5.30417919e-01 1.90117598e-01
9.69864309e-01 7.75764287e-01 -3.42264324e-01 -7.84642845e-02
4.41926539e-01 -9.13615882e-01 6.82963252e-01 3.32457572e-01
7.58967757e-01 -5.14103234e-01 8.94380450e-01 4.79683220e-01
-1.16013837e+00 2.59228379e-01 -6.61303103e-01 -7.71043599e-01
6.61533400e-02 4.42141056e-01 -1.43160462e+00 7.83334136e-01
4.37225282e-01 1.14259994e+00 -5.99254847e-01 1.41639507e+00
-6.36268198e-01 1.03969201e-01 -4.88685131e-01 1.13228485e-01
2.51886725e-01 -2.53193408e-01 4.79972422e-01 7.65177786e-01
2.09539637e-01 -3.13601851e-01 3.06818306e-01 4.86708820e-01
2.66531967e-02 -1.99685991e-01 -1.14411557e+00 3.46712142e-01
-1.96841806e-01 1.02462435e+00 -3.47471088e-01 -2.32525483e-01
-3.42891902e-01 8.76951456e-01 4.37043071e-01 1.81834415e-01
-3.07737350e-01 -2.31331259e-01 4.93176341e-01 1.35926008e-01
5.72277546e-01 -8.02044451e-01 -5.95030844e-01 -1.34876406e+00
-2.47116257e-02 -2.80617386e-01 -1.58307612e-01 -1.24785948e+00
-5.95613301e-01 4.32038337e-01 -5.08444346e-02 -1.54213560e+00
-8.50819349e-01 -1.11744869e+00 -4.42743778e-01 1.09370911e+00
-1.75094056e+00 -9.57460940e-01 -6.87565029e-01 3.11543584e-01
2.41747737e-01 -2.64958054e-01 4.88054305e-01 2.21263781e-01
1.45536855e-01 7.61404708e-02 1.68806568e-01 1.71345115e-01
7.65038311e-01 -1.32470810e+00 5.05183518e-01 8.62136900e-01
4.22617376e-01 3.16745102e-01 1.14186919e+00 -5.92663527e-01
-1.42987919e+00 -7.60348022e-01 9.78019297e-01 -5.44723690e-01
4.50586408e-01 -2.47586474e-01 -4.72721159e-01 1.15193212e+00
2.03490525e-01 3.36717814e-01 -2.15229049e-01 -2.21898854e-01
-1.49149150e-01 -1.45445004e-01 -1.04481411e+00 2.09583998e-01
8.95178139e-01 -7.53484368e-01 -8.01388860e-01 2.49733582e-01
2.51899570e-01 -1.18437660e+00 -3.98368090e-01 4.32629049e-01
6.31959021e-01 -1.06878519e+00 9.49313819e-01 -8.74133222e-03
1.89912871e-01 -5.10572970e-01 -7.53904879e-02 -1.44001102e+00
-1.59872491e-02 -3.98167640e-01 -1.59281686e-01 3.47198248e-01
4.51308377e-02 -6.58570766e-01 1.38840425e+00 -7.04757273e-02
2.71320771e-02 -1.96832761e-01 -1.15969181e+00 -8.86446953e-01
-1.53246194e-01 -1.12603478e-01 1.08043337e-02 6.49821281e-01
-1.18092373e-01 4.45473969e-01 -6.32740319e-01 3.59225601e-01
7.95433044e-01 1.25533313e-01 1.34936965e+00 -1.44086158e+00
-2.03128666e-01 2.08837956e-01 -1.19910145e+00 -1.31459200e+00
6.79385811e-02 -6.00576878e-01 1.34310350e-01 -1.42771959e+00
-3.54802221e-01 1.11245424e-01 3.70881736e-01 8.04968476e-02
3.97567421e-01 6.12957597e-01 -2.11089142e-02 3.41767251e-01
-2.72697657e-01 5.59143364e-01 1.13751018e+00 1.06269397e-01
-1.51226580e-01 4.02897745e-02 2.54043698e-01 9.97102022e-01
7.01478839e-01 -3.59565675e-01 -3.33506435e-01 -3.47293645e-01
2.57299423e-01 1.08412225e-02 5.46643019e-01 -1.38805199e+00
2.98764944e-01 1.10670932e-01 6.75182939e-01 -1.04566681e+00
6.17600024e-01 -9.07339811e-01 6.41825795e-02 8.99197161e-01
3.00247878e-01 -2.48789228e-02 1.83415443e-01 3.27299029e-01
-3.15286666e-01 -1.23429887e-01 8.76177013e-01 -1.20427445e-01
-9.79252934e-01 2.32186243e-01 -3.74776870e-02 -1.30864799e-01
7.96251714e-01 -3.90329003e-01 -1.56357631e-01 -6.02270484e-01
-7.05783427e-01 -1.28602818e-01 9.42380250e-01 4.60526347e-02
5.32916129e-01 -1.43660438e+00 -2.84584135e-01 4.00929719e-01
4.59242940e-01 6.76141307e-02 -2.30464578e-01 6.74809456e-01
-1.46152127e+00 1.03322208e+00 -5.85720837e-01 -1.02096057e+00
-1.12406659e+00 4.25165802e-01 8.76707196e-01 8.72091949e-03
-7.23707259e-01 3.66695076e-01 -7.69659579e-02 -8.80823195e-01
8.51076841e-02 -3.39783937e-01 -1.85090080e-01 -3.42266351e-01
1.14436448e-01 3.33113968e-01 2.92723745e-01 -1.00990760e+00
-3.11042249e-01 8.40618134e-01 3.60107630e-01 -4.63999212e-01
1.29081619e+00 -2.61098385e-01 1.34763986e-01 4.54792857e-01
1.27765381e+00 -2.45645344e-01 -1.68080652e+00 -6.15497530e-01
4.41814847e-02 -5.44070959e-01 3.22812259e-01 -4.63760585e-01
-8.81638050e-01 1.16302097e+00 8.42381120e-01 -4.09955829e-01
5.73743284e-01 -2.91598380e-01 4.00808632e-01 8.28148901e-01
7.01176643e-01 -9.97687101e-01 1.96145438e-02 8.87906849e-01
8.57446551e-01 -1.68664217e+00 5.74511468e-01 -3.01799655e-01
-2.08872199e-01 1.32401979e+00 6.01053119e-01 -8.26159537e-01
7.21890211e-01 8.58708546e-02 6.05845690e-01 -3.90937418e-01
-2.48031154e-01 -3.41250330e-01 5.71766913e-01 6.97587311e-01
4.99107957e-01 -9.30039957e-02 -2.14883909e-01 -8.19954395e-01
-3.60414773e-01 2.58582771e-01 6.82181060e-01 1.23404706e+00
-5.64929545e-01 -1.02321506e+00 -3.79673898e-01 -2.13119343e-01
7.63223469e-02 1.75121859e-01 -5.72548091e-01 1.11867559e+00
-1.22811787e-01 4.47017759e-01 5.28360531e-02 -1.45975217e-01
2.74393976e-01 6.15171604e-02 9.24508631e-01 -6.91272616e-01
-2.33653575e-01 -6.74989745e-02 1.38555318e-01 -6.44301891e-01
-5.30193746e-01 -5.22488654e-01 -1.12529349e+00 -1.95312113e-01
-1.87326577e-02 -5.17408788e-01 1.06196201e+00 1.18081677e+00
5.82202263e-02 1.97406486e-02 6.43823922e-01 -1.55442500e+00
-3.63533169e-01 -8.71826410e-01 -5.49345732e-01 1.44235551e-01
5.92764020e-01 -8.67745042e-01 -2.24592596e-01 -1.19277455e-01] | [7.987481117248535, -2.2041714191436768] |
9d9a9df3-65a2-4e9f-9a22-cdcae8c3a835 | deep-face-quality-assessment | 1811.04346 | null | http://arxiv.org/abs/1811.04346v1 | http://arxiv.org/pdf/1811.04346v1.pdf | Deep Face Quality Assessment | Face image quality is an important factor in facial recognition systems as
its verification and recognition accuracy is highly dependent on the quality of
image presented. Rejecting low quality images can significantly increase the
accuracy of any facial recognition system. In this project, a simple approach
is presented to train a deep convolutional neural network to perform end-to-end
face image quality assessment. The work is done in 2 stages : First, generation
of quality score label and secondly, training a deep convolutional neural
network in a supervised manner to predict quality score between 0 and 1. The
generation of quality labels is done by comparing the face image with a
template of best quality images and then evaluating the normalized score based
on the similarity. | ['Vishal Agarwal'] | 2018-11-11 | null | null | null | null | ['face-image-quality', 'face-image-quality-assessment'] | ['computer-vision', 'computer-vision'] | [ 4.06022608e-01 2.09462214e-02 2.88231879e-01 -9.71339405e-01
-3.20019215e-01 -1.92406952e-01 3.18333000e-01 -2.15443015e-01
-4.14859653e-01 3.47583473e-01 -9.60790887e-02 -3.88494916e-02
-1.82559386e-01 -9.85322893e-01 -2.85221964e-01 -5.68318963e-01
1.58568308e-01 3.52981299e-01 -1.74649894e-01 -3.26455310e-02
3.46784800e-01 1.11680913e+00 -1.86639583e+00 5.46932578e-01
3.75036389e-01 1.41947734e+00 -4.77897763e-01 8.44489813e-01
-1.31844908e-01 5.95346689e-01 -9.24725652e-01 -5.09308159e-01
4.80854362e-01 -5.79500914e-01 -8.62320065e-01 2.98060566e-01
6.83575630e-01 -4.84491080e-01 -1.04082145e-01 1.11274827e+00
6.75228417e-01 -3.41612473e-02 6.08980119e-01 -1.36679554e+00
-2.67778546e-01 -6.61766157e-02 -6.26285300e-02 1.93902701e-01
3.97674084e-01 2.30371639e-01 4.82141227e-01 -1.13068473e+00
3.67367536e-01 1.32551050e+00 4.20313418e-01 6.86204553e-01
-1.05644870e+00 -7.36729443e-01 -7.47562647e-01 3.06675941e-01
-1.29446852e+00 -8.56761158e-01 7.02507317e-01 -4.26761478e-01
8.96969855e-01 -1.43283248e-01 5.83887339e-01 4.35270369e-01
4.64567207e-02 2.64106840e-01 1.10587239e+00 -5.35152555e-01
3.11088890e-01 -2.42942516e-02 -2.11051807e-01 8.07269633e-01
-9.11779255e-02 4.80359435e-01 -4.81080949e-01 1.30084932e-01
4.67864662e-01 -2.60032088e-01 1.11174069e-01 -5.31848483e-02
-4.27373737e-01 6.90471828e-01 6.13732755e-01 5.48593402e-01
-5.42003334e-01 8.05708617e-02 3.51265699e-01 6.37004077e-01
2.55188704e-01 1.69254109e-01 -1.02210827e-01 9.15666446e-02
-1.48780060e+00 -5.57887889e-02 5.43090582e-01 8.04118738e-02
7.14340687e-01 1.92491964e-01 -3.39343727e-01 8.69854391e-01
5.37188649e-01 5.45545816e-01 3.85247797e-01 -1.05207276e+00
-1.68682203e-01 8.20852995e-01 -9.66972783e-02 -1.02307463e+00
-3.72437358e-01 -1.55092865e-01 -6.66010737e-01 1.26823032e+00
4.69552189e-01 -6.57237098e-02 -1.23420227e+00 1.40537548e+00
2.56292582e-01 -2.57124990e-01 7.48340711e-02 7.97826827e-01
1.07682133e+00 5.50099373e-01 5.65841161e-02 -1.35379583e-01
1.03988838e+00 -4.17449951e-01 -7.55537927e-01 2.52313495e-01
8.66564959e-02 -1.08975506e+00 7.28557587e-01 5.91834068e-01
-1.17100942e+00 -1.05879891e+00 -1.17731774e+00 1.84171230e-01
-4.07265157e-01 3.94335568e-01 5.44942580e-02 1.22587562e+00
-1.51705861e+00 8.00911427e-01 -3.92757416e-01 -8.26844648e-02
7.62765050e-01 8.19852531e-01 -8.88712525e-01 -4.50340696e-02
-9.30125237e-01 1.02441359e+00 4.44263488e-01 1.87283516e-01
-1.03537726e+00 -3.33236933e-01 -6.59947634e-01 1.94541395e-01
-3.69587690e-01 -1.74309358e-01 1.00107288e+00 -1.65341711e+00
-1.55019665e+00 1.16676950e+00 -9.63596925e-02 -1.37555450e-01
4.96066988e-01 2.54622430e-01 -6.38261616e-01 2.12100312e-01
-1.23756722e-01 8.37562799e-01 1.11051440e+00 -1.05180359e+00
-7.16790974e-01 -5.33013344e-01 -1.85746163e-01 -2.96361074e-02
-2.31929943e-01 4.62295979e-01 -2.67925292e-01 -1.14178210e-01
3.67305055e-02 -5.00956595e-01 3.25776756e-01 3.81675422e-01
3.53875816e-01 -4.41551417e-01 8.04090083e-01 -7.76743114e-01
7.75160015e-01 -2.08638477e+00 -3.94983828e-01 5.90564430e-01
-8.62389505e-02 7.51232922e-01 -4.25985992e-01 -1.30059466e-01
-4.00082856e-01 -6.58505708e-02 7.87714794e-02 1.17387087e-03
-9.23616290e-02 -1.32792071e-01 3.53785455e-01 2.90490240e-01
8.46060395e-01 5.28233171e-01 -5.99904060e-01 -3.66666526e-01
3.13246936e-01 8.02160203e-01 -3.99697632e-01 5.13600588e-01
1.58935145e-01 1.29803896e-01 1.99718609e-01 6.35329783e-01
9.69390869e-01 1.76374555e-01 -5.14718853e-02 -5.55139542e-01
1.17254995e-01 -4.54853401e-02 -1.29135323e+00 1.07835197e+00
-3.53239477e-01 8.08416009e-01 1.60508696e-02 -7.48310626e-01
1.21845829e+00 5.63531339e-01 3.97072881e-01 -1.08994043e+00
4.80066448e-01 4.82955128e-02 1.61775798e-01 -4.91176397e-01
2.23121613e-01 -3.60548615e-01 4.88525510e-01 5.45049548e-01
5.01491010e-01 5.09923808e-02 3.04435790e-01 -2.32950702e-01
8.72949541e-01 -3.96240242e-02 6.62356466e-02 -1.87694207e-02
8.44655395e-01 -5.32590568e-01 3.40003729e-01 1.69675529e-01
-6.18143916e-01 5.83410740e-01 5.56964874e-01 -7.65051484e-01
-1.24046350e+00 -8.72805655e-01 -2.37490788e-01 7.61670232e-01
-3.41413945e-01 7.64150918e-02 -1.06255627e+00 -7.04346538e-01
-1.26609206e-01 1.06911950e-01 -8.30380559e-01 -4.03537363e-01
-3.97183239e-01 -4.34805453e-01 6.11966491e-01 2.73316026e-01
5.79066396e-01 -1.54889739e+00 -6.35743558e-01 1.44815876e-03
1.40798509e-01 -6.25454724e-01 -6.47783056e-02 4.27479893e-02
-6.20540559e-01 -1.09233701e+00 -5.22501051e-01 -6.86948478e-01
7.98531651e-01 -9.47101265e-02 1.03018892e+00 5.86220741e-01
-5.09035051e-01 -9.37563106e-02 -1.20371342e-01 -3.64049852e-01
-7.39422858e-01 -4.31925774e-01 -7.42254406e-02 3.40717107e-01
5.05825698e-01 -1.48746565e-01 -7.22972572e-01 3.23048353e-01
-1.01501095e+00 -4.38060105e-01 7.35732377e-01 7.85135567e-01
3.16225857e-01 3.33621502e-01 4.11110640e-01 -4.34133083e-01
5.88469982e-01 1.40703827e-01 -6.49851799e-01 3.24661940e-01
-7.26612747e-01 9.23102796e-02 3.58340681e-01 -1.71974152e-02
-8.81512105e-01 3.10263723e-01 -7.18716502e-01 -2.00756013e-01
-4.23447520e-01 1.86457187e-01 -4.26232815e-01 -5.11097252e-01
7.55669832e-01 -8.32317621e-02 3.71304929e-01 -4.51953746e-02
3.36560607e-02 7.72983313e-01 4.72823888e-01 8.66496563e-02
5.63149333e-01 1.60017475e-01 2.14143887e-01 -6.79851830e-01
-4.19000775e-01 -2.19817013e-01 -7.91616142e-01 -9.54016685e-01
7.05636680e-01 -4.42931890e-01 -9.08968985e-01 6.49173200e-01
-1.01645792e+00 -8.29787627e-02 -5.83662428e-02 2.39769697e-01
-2.15743974e-01 1.36924461e-01 -3.66046369e-01 -7.56922364e-01
-6.36351466e-01 -1.22446716e+00 8.47724497e-01 4.33230639e-01
-3.67242135e-02 -5.41236341e-01 -5.53853512e-02 4.20020074e-01
6.80921257e-01 1.52161554e-01 6.59573495e-01 -3.50496203e-01
-1.31644949e-01 -6.21615887e-01 -5.26566327e-01 8.13340902e-01
2.37609535e-01 3.21376592e-01 -1.26902533e+00 -2.90642232e-01
-1.79203302e-01 -5.18597722e-01 5.08419573e-01 3.08626652e-01
8.28767180e-01 -1.24459513e-01 3.10804307e-01 4.97420788e-01
1.48662281e+00 2.68673956e-01 1.07745934e+00 8.53219032e-02
9.67805684e-02 6.32638872e-01 4.64347303e-01 1.80468500e-01
-3.49357486e-01 6.61097288e-01 5.67794800e-01 -2.73356855e-01
-4.38886315e-01 1.31084442e-01 3.12078923e-01 6.38389913e-03
-2.32938770e-02 5.54988533e-02 -7.75581598e-01 2.33850375e-01
-1.13948905e+00 -1.26001012e+00 1.73053116e-01 2.14876175e+00
5.63115597e-01 1.38206407e-01 2.39843309e-01 6.69816256e-01
7.30041623e-01 -3.46057683e-01 -2.31102094e-01 -7.07146764e-01
7.20662251e-02 6.78769112e-01 -1.34496912e-02 4.76952314e-01
-1.00868356e+00 8.23789001e-01 7.08116817e+00 5.00883281e-01
-1.54342973e+00 -4.88178395e-02 8.62966478e-01 -1.13299519e-01
2.89655954e-01 -5.12532234e-01 -5.55555701e-01 4.07364100e-01
1.07417715e+00 8.63798559e-02 3.33929330e-01 7.01237023e-01
3.01577061e-01 -2.18298972e-01 -8.92470241e-01 1.10441935e+00
3.49491894e-01 -9.42639053e-01 2.13507682e-01 -5.88346198e-02
6.04703069e-01 -2.80418783e-01 1.30571127e-01 -2.48267688e-02
1.00666098e-01 -1.47332156e+00 6.68608189e-01 7.95806348e-01
8.65835488e-01 -1.08980775e+00 1.13339365e+00 -1.60658434e-01
-8.11592042e-01 2.71112844e-03 -5.34449220e-01 9.89441574e-02
-2.15129182e-01 5.17398715e-01 -9.32135642e-01 1.29111692e-01
7.93433905e-01 -4.24883924e-02 -6.16484523e-01 1.08607674e+00
-5.37843630e-02 4.15111661e-01 -1.00290328e-01 2.01160327e-01
-2.50089802e-02 1.43386805e-02 -5.47202080e-02 1.12725008e+00
4.06875849e-01 8.94702673e-02 -2.32142508e-01 6.49519622e-01
-3.01704615e-01 1.82543948e-01 -3.76810551e-01 1.14077687e-01
7.90624246e-02 1.43980598e+00 -8.15379083e-01 -2.76061535e-01
-3.00958842e-01 1.01336038e+00 8.98059011e-02 -1.65059447e-01
-3.00660282e-01 -5.33341348e-01 3.99076611e-01 3.18914875e-02
1.32573843e-01 2.79075652e-01 -1.95036799e-01 -3.80860806e-01
4.39186841e-02 -9.61748600e-01 2.97405601e-01 -7.19749331e-01
-8.75320852e-01 1.06480086e+00 -6.52624071e-01 -9.79527950e-01
-3.55282545e-01 -8.40356529e-01 -6.25471592e-01 1.20106184e+00
-1.31274998e+00 -9.50345039e-01 -7.42321312e-01 6.59110129e-01
2.24620491e-01 -4.56644803e-01 9.06710804e-01 6.02900684e-01
-2.49454275e-01 8.75179827e-01 -2.75443077e-01 4.65616286e-01
5.71811199e-01 -9.54639733e-01 6.70011938e-02 8.67081702e-01
-9.22691524e-02 1.19296238e-01 5.77587008e-01 -3.97333562e-01
-9.25360560e-01 -9.49229956e-01 1.02035654e+00 -4.86197285e-02
-4.12592441e-02 2.09738433e-01 -8.24697196e-01 -2.38047138e-01
2.54291922e-01 2.56945461e-01 8.19529533e-01 -3.24760586e-01
-3.04009467e-01 -6.11814797e-01 -1.72983050e+00 4.34699766e-02
4.88799065e-01 -5.38429141e-01 -2.50579268e-01 -2.27294583e-02
-2.17512235e-01 1.06155939e-01 -9.55610454e-01 5.35972059e-01
8.77563775e-01 -1.23949957e+00 6.81268215e-01 -3.89329910e-01
3.56046051e-01 -4.59509701e-01 4.95471209e-02 -1.13189411e+00
-3.68437916e-01 6.32473230e-02 4.85778093e-01 1.09768438e+00
4.97624218e-01 6.95015350e-03 1.12390113e+00 6.17146194e-01
3.51828545e-01 -3.53681266e-01 -9.49846327e-01 -4.22237664e-01
-2.81011403e-01 -1.75713167e-01 7.21461117e-01 5.41528404e-01
-4.18269306e-01 1.29114151e-01 -2.36517116e-01 1.53255373e-01
6.83412910e-01 -2.86114454e-01 5.80834746e-01 -1.47725046e+00
3.41778129e-01 -7.29962289e-01 -1.03798652e+00 2.67729647e-02
-1.42871616e-02 -7.71045148e-01 1.73155352e-01 -1.40243471e+00
1.23825289e-01 -8.89905468e-02 -4.93174404e-01 5.10508835e-01
1.37870267e-01 9.89447594e-01 4.99631353e-02 -1.28327325e-01
-4.51414764e-01 3.67848456e-01 9.79742527e-01 -1.98668584e-01
1.02308787e-01 1.31976515e-01 -1.64701328e-01 3.44744295e-01
8.82730305e-01 -3.49936545e-01 -1.14748761e-01 -2.39136562e-01
9.31451246e-02 7.68893436e-02 1.54385135e-01 -1.59128451e+00
9.51866433e-02 2.72120796e-02 1.24785089e+00 -4.57503080e-01
2.20373631e-01 -9.38473046e-01 2.49510080e-01 6.99765801e-01
-4.78950828e-01 2.46906560e-02 2.53275454e-01 -1.92228869e-01
-4.84586447e-01 -5.30819535e-01 1.42676497e+00 -8.19517225e-02
-6.58039033e-01 3.64482105e-01 -2.39726365e-01 -7.01656342e-01
8.44213903e-01 -5.65323770e-01 2.01287508e-01 -2.52991885e-01
-1.01290703e+00 -3.19961339e-01 3.12816918e-01 3.40881944e-01
9.91839588e-01 -1.51347029e+00 -8.86974216e-01 6.57263517e-01
9.93033573e-02 -9.10630882e-01 1.32755652e-01 1.97358385e-01
-7.52111077e-01 9.63437334e-02 -9.35100257e-01 -4.28165704e-01
-1.83684719e+00 2.65617102e-01 7.83635616e-01 2.35309880e-02
9.69327241e-02 8.08932960e-01 -3.85608345e-01 -2.20034808e-01
4.16711599e-01 2.36326382e-01 -6.77435815e-01 3.07682659e-02
1.10014188e+00 2.82582104e-01 6.96603596e-01 -1.10738683e+00
-2.85671383e-01 6.21097207e-01 9.86043662e-02 -1.83617279e-01
1.32780409e+00 2.73500890e-01 -3.11002433e-01 -2.24817783e-01
1.43742919e+00 -4.76393670e-01 -1.01160324e+00 -2.02813316e-02
4.48451713e-02 -6.63240731e-01 4.65438068e-01 -1.17063951e+00
-1.51684153e+00 9.34613466e-01 1.59218454e+00 -6.08148314e-02
1.50908554e+00 -4.00033414e-01 2.29303688e-01 2.09487855e-01
2.68355403e-02 -1.27078331e+00 4.03765857e-01 2.79420406e-01
1.00256741e+00 -1.42659259e+00 -1.94187075e-01 7.95273557e-02
-2.73816615e-01 1.49394536e+00 6.29038036e-01 7.87685663e-02
7.86987901e-01 1.64888680e-01 4.72919464e-01 -4.10606384e-01
-4.38892692e-01 -2.87070185e-01 6.80953443e-01 6.49695575e-01
7.90241539e-01 -1.36221960e-01 -2.46901318e-01 1.20715372e-01
-1.45451874e-01 5.42605817e-01 1.47271514e-01 7.16018140e-01
-7.58560479e-01 -1.22277498e+00 -5.23390830e-01 5.58349967e-01
-7.64474690e-01 2.28125751e-01 -5.50054252e-01 1.22196145e-01
4.63112861e-01 1.19089818e+00 1.16642095e-01 -4.34657454e-01
4.50154811e-01 2.12022632e-01 6.00003242e-01 -3.46781969e-01
-7.79837668e-01 -1.77480891e-01 -2.38050133e-01 -6.40253246e-01
-4.21901852e-01 -3.65486085e-01 -1.00043380e+00 -5.81766725e-01
-1.47661969e-01 -3.64798144e-03 1.13072848e+00 8.95851374e-01
6.17736951e-02 3.16598445e-01 9.01852071e-01 -7.83624113e-01
-2.62758046e-01 -9.87118185e-01 -2.64054537e-01 8.30870807e-01
2.90783346e-01 -5.55938065e-01 -9.27928463e-02 8.30107033e-02] | [13.067625045776367, 0.8382455110549927] |
10cd535a-5285-47cf-96a5-2f52ffad7d15 | domain-adversarial-training-for-accented | 1806.02786 | null | http://arxiv.org/abs/1806.02786v1 | http://arxiv.org/pdf/1806.02786v1.pdf | Domain Adversarial Training for Accented Speech Recognition | In this paper, we propose a domain adversarial training (DAT) algorithm to
alleviate the accented speech recognition problem. In order to reduce the
mismatch between labeled source domain data ("standard" accent) and unlabeled
target domain data (with heavy accents), we augment the learning objective for
a Kaldi TDNN network with a domain adversarial training (DAT) objective to
encourage the model to learn accent-invariant features. In experiments with
three Mandarin accents, we show that DAT yields up to 7.45% relative character
error rate reduction when we do not have transcriptions of the accented speech,
compared with the baseline trained on standard accent data only. We also find a
benefit from DAT when used in combination with training from automatic
transcriptions on the accented data. Furthermore, we find that DAT is superior
to multi-task learning for accented speech recognition. | ['Mei-Yuh Hwang', 'Ching-Feng Yeh', 'Mari Ostendorf', 'Sining Sun', 'Lei Xie'] | 2018-06-07 | null | null | null | null | ['accented-speech-recognition'] | ['speech'] | [ 4.42621201e-01 5.19314051e-01 2.00137243e-01 -7.25937426e-01
-1.14580655e+00 -9.53662217e-01 5.57296991e-01 -4.81033921e-01
-6.30941033e-01 9.40488875e-01 5.75522482e-01 -5.76716185e-01
3.50835681e-01 -3.07609856e-01 -7.09792256e-01 -6.94171250e-01
3.83360147e-01 7.91998744e-01 -2.44817644e-01 -3.90417546e-01
-4.43753600e-01 2.99871624e-01 -7.39944220e-01 2.05221489e-01
9.64575648e-01 5.73981106e-01 5.72330952e-02 6.04104757e-01
-4.87167351e-02 7.24016130e-01 -1.19829810e+00 -4.40410435e-01
4.29234415e-01 -3.34110230e-01 -1.04549944e+00 2.06161991e-01
6.87160492e-01 -1.29881278e-01 -3.72169673e-01 9.85989451e-01
7.21898615e-01 2.26808131e-01 4.82024640e-01 -8.84889305e-01
-8.54950786e-01 9.72815812e-01 -3.20685595e-01 1.91027835e-01
-3.95490155e-02 2.65232679e-02 8.29865456e-01 -8.19296718e-01
5.87006986e-01 1.34458101e+00 4.02197897e-01 8.34008992e-01
-1.39710569e+00 -7.30149090e-01 1.90879956e-01 -1.30260974e-01
-1.26830149e+00 -8.33142161e-01 1.02699268e+00 -8.91645253e-03
9.23817277e-01 1.95899382e-01 -6.78777397e-02 1.51830065e+00
-2.41561338e-01 8.16688180e-01 1.27659965e+00 -6.56880558e-01
3.20467025e-01 4.25486788e-02 -9.28676780e-03 1.59612343e-01
-4.12964374e-01 3.81926894e-01 -3.39418292e-01 -7.06263771e-03
4.26587880e-01 -6.39113247e-01 -1.42701920e-02 -5.93352281e-02
-1.11841416e+00 8.48262310e-01 1.07373476e-01 2.70194501e-01
-4.42186832e-01 -3.57757688e-01 4.95778173e-01 7.13253617e-01
8.01383853e-01 5.80750704e-01 -9.31510687e-01 -2.99884174e-02
-7.08498120e-01 1.72538310e-01 7.74214864e-01 8.98921788e-01
6.37693584e-01 8.42014074e-01 -1.28022791e-03 1.47297359e+00
6.39496222e-02 9.54027593e-01 5.79633892e-01 -8.20515573e-01
6.65176809e-01 2.18960792e-02 -6.56745136e-02 -1.32104784e-01
-1.94862112e-01 -3.58582318e-01 -7.26146102e-01 2.87318826e-01
7.85541058e-01 -9.69780922e-01 -1.48745060e+00 2.02054262e+00
9.14670825e-02 8.74108896e-02 7.40234911e-01 7.70073175e-01
4.40653205e-01 9.12626207e-01 3.65685374e-01 -1.31964743e-01
1.16720641e+00 -9.25215125e-01 -8.76922429e-01 -7.02893972e-01
5.67664146e-01 -1.01694286e+00 1.14145768e+00 2.12914407e-01
-9.75283086e-01 -7.57694721e-01 -9.34933484e-01 1.59949452e-01
-5.79995215e-01 -1.83060572e-01 3.37351412e-02 9.38048959e-01
-1.17957461e+00 -1.05235986e-01 -5.79711735e-01 -4.07436639e-02
5.31148650e-02 5.05933642e-01 -3.73274237e-01 -9.91907064e-03
-1.63553643e+00 1.16646075e+00 4.78857189e-01 -3.16427350e-01
-8.49545300e-01 -7.82913744e-01 -1.03010726e+00 -1.35847852e-01
1.65317789e-01 -1.41258642e-01 1.46454799e+00 -1.57092595e+00
-1.59984636e+00 7.84731030e-01 2.73389113e-03 -7.27733433e-01
1.82752252e-01 -2.99747348e-01 -8.19170713e-01 -2.52059400e-01
-1.37740493e-01 8.10107589e-01 9.34934318e-01 -1.18229878e+00
-5.58121562e-01 -2.35401690e-01 -1.48011729e-01 3.67174298e-01
-3.07764262e-01 3.91460985e-01 -5.62756397e-02 -1.15319109e+00
-3.67007077e-01 -1.18882918e+00 -2.45080739e-01 -9.05392766e-01
-5.19913077e-01 -1.39910892e-01 1.22582483e+00 -1.07031953e+00
7.47088790e-01 -2.35502386e+00 2.38835081e-01 1.61697030e-01
-2.83279032e-01 8.72710645e-01 -4.19007629e-01 -3.12445015e-02
-4.79012460e-01 6.29414320e-02 -3.57078135e-01 -3.88191998e-01
3.38961147e-02 6.33525908e-01 -4.28689510e-01 7.59732351e-02
5.83126724e-01 6.27302349e-01 -6.29377425e-01 -1.28950283e-01
1.87334180e-01 4.95329022e-01 -3.14222932e-01 5.53282022e-01
-3.85809362e-01 5.85026383e-01 -5.94586730e-02 3.95432621e-01
8.97737801e-01 4.51796383e-01 4.63703305e-01 1.51841834e-01
6.04609549e-02 5.89693904e-01 -1.05512667e+00 1.50743032e+00
-7.19127893e-01 4.95628625e-01 3.75124276e-01 -8.98388922e-01
1.12627828e+00 6.79586470e-01 1.49306487e-02 -7.78729498e-01
-2.33691856e-01 1.46919876e-01 3.50200087e-01 1.65525064e-01
3.79085600e-01 -5.17394483e-01 -4.07832444e-01 3.86863738e-01
3.70741904e-01 -1.66868985e-01 -2.51100600e-01 -1.79436922e-01
1.10993707e+00 -7.71005228e-02 -6.89395051e-03 -2.92981744e-01
4.78213817e-01 -6.45688623e-02 7.78272748e-01 4.74886239e-01
-2.59865075e-01 5.76246321e-01 2.90840805e-01 -4.02798325e-01
-1.37679291e+00 -1.23983014e+00 -1.22153200e-01 1.53170931e+00
-4.67062503e-01 5.44687510e-02 -1.05560708e+00 -1.11597979e+00
-1.84279665e-01 1.12420452e+00 -3.07826936e-01 -4.98091727e-02
-1.06435359e+00 -7.56592631e-01 1.05082774e+00 6.76565826e-01
4.73358631e-01 -1.20582747e+00 2.79238582e-01 2.07678944e-01
-1.12762704e-01 -1.26789451e+00 -8.34674776e-01 8.17353487e-01
-3.16433966e-01 -2.39884868e-01 -8.97524536e-01 -1.19759297e+00
4.57170546e-01 -4.37050045e-01 1.06800461e+00 -6.13163173e-01
5.37037969e-01 4.38193679e-02 -3.88513535e-01 -6.10070527e-01
-1.18722725e+00 4.92359102e-01 2.78802037e-01 -5.54716438e-02
4.62376356e-01 -4.35486227e-01 1.92592219e-01 3.58169347e-01
-8.86913300e-01 -1.85371161e-01 6.45564854e-01 1.12914157e+00
3.95266861e-01 -1.97091132e-01 1.11029041e+00 -1.09675109e+00
4.96572137e-01 -3.42492729e-01 -5.26943624e-01 3.55948508e-02
-4.62375224e-01 2.75450766e-01 8.35780799e-01 -4.86022770e-01
-1.56882417e+00 3.26192588e-01 -7.20091164e-01 -5.42589128e-01
-8.06467593e-01 3.44628572e-01 -6.24655128e-01 2.08911791e-01
6.79905713e-01 1.82886586e-01 1.48975953e-01 -5.64292669e-01
5.03127694e-01 9.98247564e-01 6.77483737e-01 -5.29979289e-01
8.42152357e-01 -3.54765832e-01 -5.82669437e-01 -9.60166574e-01
-8.83429825e-01 -1.66302398e-01 -7.37451315e-01 2.55382776e-01
9.24342334e-01 -1.03877521e+00 1.17278010e-01 7.08871663e-01
-1.28265154e+00 -7.73291886e-01 -2.74320155e-01 3.34626883e-01
-4.65855777e-01 7.16922730e-02 -6.52214408e-01 -6.44086182e-01
-2.77721971e-01 -1.02923393e+00 7.13903427e-01 -6.84206784e-02
-2.35659450e-01 -1.37017787e+00 1.77004293e-01 5.25465727e-01
5.27577817e-01 -1.74904112e-02 9.03545201e-01 -1.47028708e+00
-8.80941078e-02 2.34481379e-01 1.77617386e-01 1.03780484e+00
4.70662504e-01 -4.23257738e-01 -1.34023046e+00 -2.39523008e-01
3.27724330e-02 -5.05006075e-01 5.79377413e-01 3.32843214e-01
6.89393163e-01 -4.42096949e-01 1.90151364e-01 5.04052103e-01
9.46096778e-01 6.29033506e-01 6.46537423e-01 3.19635391e-01
7.65971243e-01 5.41008055e-01 5.46345532e-01 -8.62916261e-02
1.64181903e-01 7.96127856e-01 -2.06959844e-02 -6.37790680e-01
-4.52400416e-01 -1.27675429e-01 6.40440702e-01 9.17479932e-01
3.51353228e-01 -5.28764784e-01 -1.19141960e+00 7.28712857e-01
-1.34035707e+00 -7.71018744e-01 2.56693691e-01 1.97482967e+00
1.28397727e+00 2.81494766e-01 2.88588405e-01 -5.86620085e-02
1.08335054e+00 3.25959116e-01 -4.21569228e-01 -8.67400169e-01
-5.39666831e-01 6.77815855e-01 6.88816667e-01 9.21508014e-01
-1.58091867e+00 1.34497523e+00 6.87063837e+00 7.45112240e-01
-1.15645468e+00 1.69726565e-01 6.95298374e-01 2.06435770e-01
-1.94084466e-01 -2.69932747e-01 -7.95369565e-01 5.07255554e-01
1.69039059e+00 -5.57185002e-02 3.92668813e-01 9.09249902e-01
1.98944807e-02 5.44949770e-01 -8.81717086e-01 3.95626187e-01
-4.89626788e-02 -9.85636473e-01 -1.93160921e-02 1.08218104e-01
8.12336087e-01 3.02417338e-01 2.40792856e-01 5.64104378e-01
9.16161895e-01 -1.01394224e+00 5.68870485e-01 -1.78503573e-01
8.30145717e-01 -1.16755545e+00 9.96542573e-01 1.97928250e-01
-6.01345479e-01 3.37691277e-01 -1.66208088e-01 1.38771445e-01
2.34867826e-01 2.98666686e-01 -1.46353030e+00 4.35685813e-01
1.93446904e-01 1.34055436e-01 -2.65520096e-01 2.92958170e-01
-2.31317788e-01 1.16086638e+00 -2.67670661e-01 2.69401133e-01
4.06377524e-01 -1.08694427e-01 6.03844583e-01 1.50447094e+00
-1.85405031e-01 -2.06634358e-01 2.08788440e-01 5.34114659e-01
-3.09845388e-01 -5.52969240e-02 -7.81110883e-01 -1.12858582e-02
6.11612320e-01 8.08571339e-01 -1.38076663e-01 -3.46698195e-01
-5.28365970e-01 1.24591172e+00 3.44544917e-01 4.92965579e-01
-6.01001263e-01 -5.03529966e-01 1.00566447e+00 -2.46282652e-01
4.88469929e-01 -2.54789382e-01 -4.32448238e-01 -7.04047024e-01
-2.30788797e-01 -1.39962220e+00 2.02184841e-01 -5.29050767e-01
-1.39433002e+00 8.88639152e-01 -2.67815739e-01 -7.22169816e-01
-7.71523416e-01 -5.48851550e-01 -5.46655476e-01 1.50764954e+00
-1.42796040e+00 -1.33150816e+00 5.20926178e-01 6.37132883e-01
8.73889923e-01 -4.96664733e-01 1.19316030e+00 4.40497279e-01
-4.70435113e-01 9.18087780e-01 2.74349511e-01 7.97808707e-01
1.02990556e+00 -1.78332961e+00 1.00580657e+00 9.95725930e-01
1.78516164e-01 3.98171693e-01 7.59940028e-01 -5.05782843e-01
-9.32679951e-01 -1.40016246e+00 1.01292324e+00 -5.73245108e-01
5.48313141e-01 -4.99167323e-01 -1.09997666e+00 1.09413946e+00
7.67317533e-01 -2.23610908e-01 9.64949608e-01 3.59747320e-01
-3.18210870e-01 -3.98236811e-02 -1.17170048e+00 4.30801064e-01
5.62911153e-01 -6.44027948e-01 -1.17193496e+00 1.24376222e-01
1.14957762e+00 -5.79216123e-01 -9.61714923e-01 1.90028831e-01
4.63583171e-02 -3.18227470e-01 8.80809665e-01 -8.48853171e-01
7.05397949e-02 -1.65826321e-01 -2.53693104e-01 -2.07845092e+00
-1.79116488e-01 -6.37842596e-01 3.73067170e-01 1.68538630e+00
8.93106282e-01 -5.43282866e-01 6.51457250e-01 5.25077939e-01
-6.29413307e-01 9.83859673e-02 -1.06589019e+00 -1.00289798e+00
7.16417253e-01 -1.90694153e-01 5.73162496e-01 1.26184130e+00
-2.40406349e-01 7.11480141e-01 -4.80638474e-01 4.01115328e-01
3.64407897e-01 -4.28645730e-01 5.59842408e-01 -8.73664975e-01
-1.57562733e-01 7.40629509e-02 -1.82243899e-01 -8.36598516e-01
8.22779834e-01 -8.17848086e-01 2.73543805e-01 -8.70490670e-01
-4.29748803e-01 -5.08076668e-01 -4.43293691e-01 8.34166944e-01
-2.62385637e-01 8.88938084e-02 1.82574913e-01 -2.93119162e-01
-1.60506815e-01 3.54795992e-01 9.55358744e-01 -2.37515107e-01
-2.48780444e-01 1.21076636e-01 -7.03025103e-01 6.50236189e-01
1.05765247e+00 -4.54410404e-01 -3.07962269e-01 -5.52965820e-01
-6.26072645e-01 4.03115302e-02 -1.10663146e-01 -8.34438145e-01
-1.64898578e-02 -1.50023296e-01 3.52208734e-01 -2.94534445e-01
3.72021645e-01 -5.73454678e-01 -1.35800213e-01 3.92594114e-02
-5.53351283e-01 -9.85697284e-02 6.50366843e-01 1.37542844e-01
-4.23548877e-01 -6.42828792e-02 1.17352831e+00 1.61226153e-01
-7.37738073e-01 -9.95239243e-02 -6.33619010e-01 4.41755831e-01
5.81727266e-01 2.49144226e-01 -3.66374224e-01 -3.43855500e-01
-1.07139754e+00 8.95962119e-02 3.41526896e-01 5.91824234e-01
1.83250070e-01 -1.39874470e+00 -1.02889407e+00 5.31871617e-01
-1.84382111e-01 -6.11156709e-02 9.20311064e-02 1.35893926e-01
-8.69637802e-02 4.61719036e-01 -2.14582652e-01 -2.73230344e-01
-1.25691032e+00 5.80504417e-01 4.44927782e-01 -2.18076319e-01
-1.58657953e-01 8.54422212e-01 2.67920285e-01 -1.08719397e+00
3.23761910e-01 7.53149623e-04 1.77236304e-01 -1.79513693e-01
2.91606784e-01 4.97847609e-02 3.48076731e-01 -7.19204545e-01
-4.09129471e-01 -9.36837420e-02 -5.22220552e-01 -5.49617231e-01
1.17472470e+00 -8.53042454e-02 3.99151117e-01 1.97202846e-01
9.73942399e-01 4.31221008e-01 -1.35897577e+00 -3.57457876e-01
2.11368114e-01 3.17698419e-02 1.95200920e-01 -1.33794224e+00
-9.06117618e-01 6.10435426e-01 5.85592449e-01 3.02447796e-01
1.09346426e+00 9.56170708e-02 7.21971154e-01 3.32437783e-01
-1.22238055e-01 -1.30456567e+00 -3.09480071e-01 9.79437470e-01
8.00047100e-01 -1.15125585e+00 -8.58974695e-01 -2.19955191e-01
-1.17738998e+00 6.98324442e-01 7.22966194e-01 4.48061302e-02
5.53794563e-01 6.99005187e-01 8.49731028e-01 3.32468271e-01
-6.11117065e-01 -2.10575804e-01 4.82295491e-02 1.16642737e+00
5.36660671e-01 2.49326289e-01 1.52547956e-01 4.24377114e-01
-5.07930934e-01 -5.69692791e-01 4.25868332e-01 6.77331746e-01
-2.91408420e-01 -1.69028664e+00 -5.75176835e-01 7.79185221e-02
-7.23918796e-01 -4.83307987e-01 -7.56956756e-01 6.92873716e-01
2.93434579e-02 1.04514670e+00 3.47860664e-01 -3.96581650e-01
4.98251647e-01 6.88393772e-01 1.77275091e-02 -7.43293703e-01
-7.08116055e-01 2.31419697e-01 5.54834545e-01 -8.53443705e-03
-2.23855093e-01 -8.43406618e-01 -1.11066616e+00 -1.36330009e-01
-1.17450222e-01 3.04532468e-01 5.71372449e-01 1.07100356e+00
2.69488662e-01 6.08149946e-01 8.39598596e-01 -3.96639735e-01
-7.10760295e-01 -1.25247800e+00 -6.90429926e-01 4.02907401e-01
5.75514495e-01 -3.32652867e-01 -3.73083800e-01 4.50613350e-01] | [14.354082107543945, 6.769536972045898] |
7f223160-c9f9-4351-b092-995ba59f9e7a | the-syn-series-corpora-of-written-czech | null | null | https://aclanthology.org/L14-1267 | https://aclanthology.org/L14-1267.pdf | The SYN-series corpora of written Czech | The paper overviews the SYN series of synchronic corpora of written Czech compiled within the framework of the Czech National Corpus project. It describes their design and processing with a focus on the annotation, i.e. lemmatization and morphological tagging. The paper also introduces SYN2013PUB, a new 935-million newspaper corpus of Czech published in 2013 as the most recent addition to the SYN series before planned revision of its architecture. SYN2013PUB can be seen as a completion of the series in terms of titles and publication dates of major Czech newspapers that are now covered by complete volumes in comparable proportions. All SYN-series corpora can be characterized as traditional, with emphasis on cleared copyright issues, well-defined composition, reliable metadata and high-quality data processing; their overall size currently exceeds 2.2 billion running words. | ["Hana Skoumalov{\\'a}", "Pavel Proch{\\'a}zka", 'Michal K{\\v{r}}en', "Milena Hn{\\'a}tkov{\\'a}"] | 2014-05-01 | null | null | null | lrec-2014-5 | ['morphological-tagging'] | ['natural-language-processing'] | [-1.98894903e-01 3.66068602e-01 -3.18153113e-01 1.94214821e-01
-8.50123107e-01 -1.14724708e+00 1.14375365e+00 7.71627545e-01
-1.06643355e+00 7.39618897e-01 7.15220273e-01 -3.04328501e-01
-1.30447736e-02 -3.70638311e-01 -3.37054640e-01 -2.32191414e-01
4.11967844e-01 7.25087464e-01 1.88698739e-01 -3.68709028e-01
3.23898852e-01 3.31520557e-01 -8.73673499e-01 1.73487216e-01
5.92009306e-01 3.55980515e-01 5.58615446e-01 3.42725366e-01
-5.64621508e-01 3.22860569e-01 -9.35219765e-01 -1.01650381e+00
4.45724279e-02 -1.90006182e-01 -1.12018919e+00 -2.94247717e-01
1.72710977e-04 2.84201205e-01 -1.00345150e-01 9.70458448e-01
2.73582757e-01 -4.50736463e-01 3.64317775e-01 -6.59698486e-01
-2.52445430e-01 1.36341465e+00 -4.07258749e-01 4.09766912e-01
4.69093382e-01 -2.78040498e-01 1.22972608e+00 -7.72385299e-01
1.41826856e+00 7.48338223e-01 7.01324940e-01 1.79564565e-01
-9.17252660e-01 -5.49095571e-01 -2.24153444e-01 -2.04875916e-01
-1.16950011e+00 -6.16668522e-01 4.31369931e-01 -6.67730272e-01
1.26890326e+00 1.62328467e-01 9.33836102e-01 9.31171298e-01
4.33497310e-01 4.42576379e-01 1.14209259e+00 -8.01131845e-01
-2.72288416e-02 2.01142490e-01 -1.91106030e-03 -5.93441799e-02
9.27466035e-01 -5.15951633e-01 -3.23122054e-01 -1.25237972e-01
5.39131582e-01 -5.18711269e-01 4.32866567e-04 1.21762091e-02
-1.68896484e+00 4.73568648e-01 -5.84505796e-01 1.11568749e+00
-4.04422879e-01 -2.54732400e-01 1.11568749e+00 3.10620546e-01
4.78304207e-01 4.91770357e-01 -7.50309229e-01 -4.80126321e-01
-1.19777668e+00 3.58588457e-01 9.08863962e-01 1.49658918e+00
1.10765651e-01 5.32402620e-02 2.74254262e-01 8.70845437e-01
3.22187781e-01 8.30866098e-01 5.27274072e-01 -7.45286822e-01
8.92318010e-01 3.48195225e-01 -8.25549513e-02 -5.66573799e-01
-4.70066756e-01 -1.68128893e-01 -4.99079049e-01 -4.68215346e-01
4.27625924e-01 -2.27383181e-01 -5.17708659e-01 1.29505634e+00
9.80630666e-02 -7.58334935e-01 3.54103476e-01 3.18080969e-02
1.32058024e+00 7.24534988e-01 2.79876232e-01 -8.28260362e-01
1.77444768e+00 -4.92771477e-01 -1.27507544e+00 3.21579427e-02
3.31057698e-01 -1.58346415e+00 5.73226511e-01 4.31994379e-01
-1.59538603e+00 -2.07987316e-02 -1.01625824e+00 -1.09698527e-01
-7.02627778e-01 7.29839057e-02 4.50290382e-01 7.07108378e-01
-9.63080943e-01 3.18395525e-01 -7.33314097e-01 -6.69859231e-01
9.02012736e-02 -2.94215977e-02 -5.69614470e-01 6.11080229e-01
-1.14280939e+00 1.16528344e+00 6.78036273e-01 -2.19612390e-01
4.64410223e-02 -5.87147534e-01 -9.07884061e-01 -2.84694076e-01
3.82738948e-01 -9.15455073e-02 1.45765710e+00 -5.80387175e-01
-1.37048864e+00 1.67493141e+00 2.11985573e-01 -4.89334792e-01
4.29634511e-01 -1.73283741e-01 -9.41456437e-01 3.73731732e-01
3.99961203e-01 7.22089112e-02 3.94022375e-01 -7.60855258e-01
-9.29370582e-01 -1.47932306e-01 -4.51510996e-01 -4.23223116e-02
-1.17372081e-01 9.81424034e-01 -7.24547029e-01 -1.19555724e+00
3.41389165e-03 -6.16284907e-01 -1.42391458e-01 -9.75716472e-01
-2.05216542e-01 -1.53078754e-02 1.39178351e-01 -1.28236091e+00
1.69022703e+00 -1.78997111e+00 -1.11916944e-01 -2.53616758e-02
3.00502211e-01 2.86497772e-01 3.53328407e-01 1.15695012e+00
-5.40484153e-02 4.69025195e-01 -1.67936802e-01 -1.94889992e-01
1.60765514e-01 3.38473469e-01 -3.92545968e-01 8.77126098e-01
-9.33116227e-02 9.50796008e-01 -9.37011719e-01 -6.08300626e-01
1.13063417e-02 2.90332437e-02 1.95586249e-01 -5.37691355e-01
2.16828093e-01 5.42585887e-02 -1.84746787e-01 5.55799127e-01
6.42839849e-01 2.93107152e-01 7.13040113e-01 1.63055986e-01
-8.92795563e-01 8.81264687e-01 -1.11058152e+00 1.64123392e+00
-3.75008076e-01 7.01122642e-01 2.85066694e-01 -5.35603285e-01
7.73495674e-01 7.91861832e-01 5.82187295e-01 -5.79494476e-01
3.22873324e-01 6.02509141e-01 -2.00035810e-01 -5.83563969e-02
1.36358190e+00 -4.18488495e-02 -9.16845083e-01 2.05451831e-01
5.19951940e-01 -4.50017273e-01 7.87722528e-01 4.56001222e-01
6.64232135e-01 2.43360758e-01 1.14339912e+00 -6.49144292e-01
5.68295538e-01 2.60358363e-01 7.57702529e-01 1.49799883e-01
1.05771050e-02 4.22728062e-01 6.28191411e-01 -3.75639461e-02
-1.86052704e+00 -6.22867048e-01 -6.34270251e-01 3.61399680e-01
-3.09516221e-01 -9.64592755e-01 -8.76651108e-01 -4.64804977e-01
-3.90221000e-01 4.89109188e-01 -4.01113778e-01 7.54510581e-01
-9.87984538e-01 -4.58291739e-01 9.12453353e-01 5.91906048e-02
1.90446213e-01 -1.55316973e+00 -7.36776114e-01 3.29309464e-01
-3.50029439e-01 -1.35165405e+00 -2.64582425e-01 1.70997724e-01
-5.18564105e-01 -1.10085523e+00 -8.28496099e-01 -1.05853784e+00
5.56650162e-02 -2.12713450e-01 1.21929407e+00 -2.86056131e-01
8.41920525e-02 2.07585841e-01 -4.63818341e-01 -6.29986346e-01
-9.62623358e-01 4.23759997e-01 -3.15478444e-02 -8.01540136e-01
2.45958403e-01 -3.67594093e-01 2.69102097e-01 -4.86372828e-01
-9.52264190e-01 -2.19140545e-01 5.27677119e-01 6.12773836e-01
6.10788047e-01 -3.13937128e-01 6.28362119e-01 -1.42668617e+00
4.96134281e-01 -3.68200302e-01 -8.66661310e-01 2.90276170e-01
-8.70401978e-01 -3.14904988e-01 4.43813890e-01 3.13453004e-02
-1.17472446e+00 -1.69919148e-01 -3.27042788e-01 4.53117311e-01
-1.76765934e-01 7.90048897e-01 -1.80216596e-01 4.83241558e-01
1.10947624e-01 3.70862372e-02 -3.23189706e-01 -8.38875830e-01
5.10354877e-01 7.64639437e-01 1.04862189e+00 -4.41190630e-01
9.06301260e-01 2.32734561e-01 -3.45612019e-01 -9.13830459e-01
-1.96465820e-01 -9.12352145e-01 -1.06335104e+00 -1.80620492e-01
7.55203366e-01 -9.80144441e-01 -3.74433726e-01 6.31353915e-01
-1.33099222e+00 -1.47551864e-01 -6.62198126e-01 4.38336581e-01
-2.98877507e-01 4.78639632e-01 -8.22658479e-01 -6.20135784e-01
-6.71473026e-01 -5.58041155e-01 7.12123990e-01 -4.18099724e-02
-6.05766475e-01 -1.24935603e+00 4.96120125e-01 -4.77850251e-02
-1.07958809e-01 4.56854194e-01 7.07976937e-01 -9.18874621e-01
4.14605349e-01 -2.15268418e-01 1.94735736e-01 5.15549779e-02
-1.30619988e-01 4.41638738e-01 -5.63985169e-01 -2.37409603e-02
-4.11820970e-03 2.09095567e-01 5.26841640e-01 2.79165179e-01
-2.24651068e-01 -5.00017941e-01 -1.03404842e-01 4.13344018e-02
1.66084945e+00 2.82303154e-01 7.82774627e-01 9.92571831e-01
1.83439508e-01 7.30433464e-01 6.83504462e-01 4.88027006e-01
3.91587943e-01 5.26710868e-01 -1.93720266e-01 3.21753919e-01
-3.79954994e-01 -1.23118013e-01 5.73376536e-01 1.86269605e+00
-2.00867150e-02 5.29600754e-02 -1.26263177e+00 1.14663565e+00
-1.56872022e+00 -1.23282194e+00 -7.01550901e-01 2.17577696e+00
1.16873896e+00 4.79001820e-01 2.63806909e-01 2.39322692e-01
8.38619232e-01 4.25544322e-01 4.91311878e-01 -5.83579600e-01
-6.18920267e-01 6.46189511e-01 9.55614984e-01 4.43982959e-01
-9.99040842e-01 1.12537217e+00 7.24657345e+00 9.72406983e-01
-6.49398923e-01 3.79642785e-01 -6.44117668e-02 -5.20506799e-02
-3.99648458e-01 2.61631846e-01 -1.02768469e+00 4.17654723e-01
1.37252235e+00 -7.33345389e-01 -1.07490243e-02 4.81016248e-01
2.21648350e-01 -2.30366901e-01 -2.96657711e-01 6.72098994e-01
-2.48230472e-02 -1.84202445e+00 -2.91868627e-01 4.11854118e-01
8.08347106e-01 3.26818168e-01 -2.86544979e-01 1.75281361e-01
4.34334159e-01 -5.51514506e-01 1.59995663e+00 3.25788140e-01
1.20395291e+00 -8.38807046e-01 1.17155397e+00 -6.49205744e-02
-1.25557697e+00 3.53738487e-01 -2.61180133e-01 2.20361296e-02
6.48251116e-01 6.58853531e-01 -2.57441640e-01 1.10632968e+00
6.34799302e-01 8.92610967e-01 -4.59987342e-01 9.42228436e-01
-3.48713666e-01 8.50649297e-01 -2.07301930e-01 -5.32593466e-02
2.90440410e-01 -4.62349474e-01 8.11861455e-01 1.74947953e+00
1.25625595e-01 4.30517942e-02 -2.79697813e-02 1.40814483e-01
-1.18013933e-01 8.07514966e-01 -3.61737221e-01 -3.57383966e-01
8.21914792e-01 1.38527131e+00 -1.23865366e+00 -5.05398691e-01
-4.99785483e-01 5.44774234e-01 7.54023716e-03 -1.38119563e-01
-4.43660766e-01 -6.58251524e-01 1.70006275e-01 8.47489759e-02
4.59731787e-01 -3.09164613e-01 -5.71350753e-01 -8.77898812e-01
1.42722160e-01 -7.60721266e-01 5.12371778e-01 -2.20102385e-01
-8.86701047e-01 9.79556501e-01 1.89754352e-01 -1.04566789e+00
-1.10410072e-01 -2.94501722e-01 1.22483084e-02 8.41407359e-01
-1.04302955e+00 -1.22252762e+00 4.88288730e-01 2.23106429e-01
3.90530705e-01 -3.49569827e-01 9.90624845e-01 3.70199203e-01
-4.14154142e-01 3.28525603e-01 5.93162119e-01 2.94196844e-01
9.27706003e-01 -1.46448028e+00 8.16666663e-01 8.33163559e-01
2.34626934e-01 5.90885520e-01 8.63956451e-01 -8.99955988e-01
-7.77597010e-01 -7.29679644e-01 2.17819262e+00 -4.81041998e-01
1.37699234e+00 -5.38938999e-01 -4.04605299e-01 6.67544186e-01
1.00441575e+00 -6.22750163e-01 3.97191286e-01 -1.65818527e-01
-9.80712175e-02 6.80749342e-02 -9.99862552e-01 4.14666623e-01
7.09485292e-01 -4.20058966e-01 -1.11805713e+00 3.64183784e-01
6.96475744e-01 -6.38661027e-01 -1.28327703e+00 -2.00757861e-01
5.18776238e-01 -2.77958035e-01 3.52233201e-01 -8.22966397e-02
3.48638773e-01 -2.61116289e-02 3.02630544e-01 -7.58410096e-01
-2.14732423e-01 -1.40244901e+00 5.60519814e-01 1.89826417e+00
5.71006298e-01 -6.64239526e-01 1.97682738e-01 -1.19836934e-01
-3.34813803e-01 1.74697563e-02 -1.14537013e+00 -8.36925864e-01
4.26986367e-01 -5.34483016e-01 4.12194610e-01 1.15499604e+00
7.13879943e-01 3.93540829e-01 -9.69727114e-02 -6.32321835e-01
2.68514991e-01 -1.13595210e-01 4.65345293e-01 -1.43461728e+00
2.09729254e-01 -8.11063170e-01 -2.08330736e-01 -3.32029223e-01
-7.39377365e-02 -1.02882755e+00 -1.59502015e-01 -1.60021520e+00
7.70642236e-02 -1.86487854e-01 3.12454641e-01 3.71731788e-01
3.96276087e-01 3.05078149e-01 2.11101100e-01 6.81244910e-01
-3.89729023e-01 2.57017855e-02 7.30690122e-01 1.47997022e-01
-3.07254225e-01 -1.00222573e-01 -5.00783920e-01 6.03903770e-01
7.77646780e-01 -6.14839435e-01 1.49745375e-01 4.04323898e-02
7.22116411e-01 -4.35374565e-02 -3.95009011e-01 -6.55970812e-01
4.69959863e-02 -1.97292939e-01 -1.47621647e-01 -9.77095366e-01
-1.47088230e-01 -6.87908828e-01 5.23213506e-01 4.50611711e-01
-1.97600275e-01 7.61709630e-01 5.67265987e-01 7.13546351e-02
-3.78626794e-01 -5.57355106e-01 5.89641333e-01 -3.71020079e-01
-6.80747807e-01 -6.25436753e-02 -9.84735608e-01 4.96820182e-01
1.03554595e+00 -1.71696797e-01 -2.48310566e-01 -1.39729112e-01
-6.90371990e-01 -1.93581492e-01 6.25558555e-01 1.97301269e-01
-9.01833735e-03 -1.03460598e+00 -9.84948516e-01 -5.25274992e-01
1.31755605e-01 -3.19635153e-01 -1.11420378e-01 8.52250516e-01
-9.92960095e-01 8.63541186e-01 -4.36511099e-01 1.66547805e-01
-1.31556547e+00 4.86441761e-01 -4.20668751e-01 -9.92498100e-01
-1.10546303e+00 -1.91438049e-02 -5.88198483e-01 -3.81604671e-01
-2.09315307e-02 -2.93959767e-01 -8.60537648e-01 5.98668516e-01
4.27834451e-01 1.74086258e-01 2.41304204e-01 -1.30813718e+00
-4.65813667e-01 2.38783389e-01 7.40667582e-02 -8.74476135e-01
1.64571452e+00 -4.99138594e-01 -6.44671679e-01 8.44668627e-01
6.87860012e-01 1.02504659e+00 -3.74949515e-01 -2.33568698e-01
7.22541749e-01 3.40631396e-01 -1.95611238e-01 -8.45122218e-01
-7.62160718e-01 2.65070528e-01 -2.16654986e-01 1.48559719e-01
6.79895401e-01 1.04503013e-01 8.45398962e-01 -1.33733466e-01
1.03818066e-01 -1.57549834e+00 -8.20843875e-01 9.46546376e-01
9.34224486e-01 -3.96524996e-01 5.01429379e-01 -4.87406075e-01
-6.96373522e-01 1.22814465e+00 -3.59906316e-01 3.58230323e-02
7.93272674e-01 6.27687454e-01 1.46870017e-01 -3.08637768e-01
-4.00595039e-01 -5.40146120e-02 1.30556062e-01 5.91650844e-01
5.67028582e-01 9.26083699e-02 -1.53171372e+00 1.04820561e+00
-7.19442070e-01 -5.06047249e-01 8.54623258e-01 1.02823424e+00
-9.95448604e-02 -1.57550085e+00 -3.28783810e-01 9.38455462e-02
-1.14866066e+00 -6.40146554e-01 -4.61554796e-01 1.40318429e+00
5.07820062e-02 6.64929390e-01 1.08756907e-01 1.76557922e-03
3.83784860e-01 1.00328319e-01 4.87664938e-01 -7.19224870e-01
-1.41628289e+00 5.36419392e-01 6.48950934e-01 1.47279790e-02
-6.47532701e-01 -1.40027058e+00 -1.23938167e+00 -6.29605293e-01
-1.54120773e-01 4.92549926e-01 8.20612967e-01 9.00706351e-01
-2.74079353e-01 2.67843604e-01 2.02395003e-02 -6.30593300e-01
-1.67057112e-01 -1.22915435e+00 -9.62545693e-01 -1.22754544e-01
-1.14685565e-01 2.08559097e-03 -9.75688398e-02 4.07108188e-01] | [10.33865737915039, 10.212453842163086] |
4b116dd2-3bb5-471b-8f83-6ee9757e70da | conditional-graphical-lasso-for-multi-label | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Li_Conditional_Graphical_Lasso_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Li_Conditional_Graphical_Lasso_CVPR_2016_paper.pdf | Conditional Graphical Lasso for Multi-Label Image Classification | Multi-label image classification aims to predict multiple labels for a single image which contains diverse content. By utilizing label correlations, various techniques have been developed to improve classification performance. However, current existing methods either neglect image features when exploiting label correlations or lack the ability to learn image-dependent conditional label structures. In this paper, we develop conditional graphical Lasso (CGL) to handle these challenges. CGL provides a unified Bayesian framework for structure and parameter learning conditioned on image features. We formulate the multi-label prediction as CGL inference problem, which is solved by a mean field variational approach. Meanwhile, CGL learning is efficient due to a tailored proximal gradient procedure by applying the maximum a posterior (MAP) methodology. CGL performs competitively for multi-label image classification on benchmark datasets MULAN scene, PASCAL VOC 2007 and PASCAL VOC 2012, compared with the state-of-the-art multi-label classification algorithms. | ['DaCheng Tao', 'Qiang Li', 'Wei Bian', 'Maoying Qiao'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['multi-label-image-classification'] | ['computer-vision'] | [ 0.5712614 -0.44724947 -0.44571272 -0.7419278 -1.2119291 -0.38855907
0.5944479 0.18325472 -0.37646657 0.6854227 -0.2613044 0.16681673
-0.10913214 -0.27591825 -0.6295596 -1.0325112 0.6209314 0.44438055
-0.07112507 0.67382216 0.41581172 0.07776556 -1.6026481 0.42669755
0.67567414 1.1070812 0.37075403 0.49582472 -0.02367615 1.0584062
-0.04853946 -0.28775933 -0.1609157 -0.36158392 -0.85325986 0.41985032
0.87295824 -0.05305179 0.365864 1.3009177 0.2597811 -0.07610396
1.254535 -1.2880998 -0.33114856 0.13266014 -0.9814363 -0.28045064
0.13229868 -0.36902314 1.3212717 -0.99884903 0.63624877 1.3770674
0.70106834 0.4341197 -1.668989 -0.59211075 0.22277798 0.18357804
-1.471309 -0.1639138 0.7340117 -0.75425947 0.4842086 0.17883044
0.12656812 1.0801821 0.2780784 1.1258295 1.6660134 -0.48371768
-0.00995883 0.23397382 0.36403346 1.0567173 -0.20161153 -0.15213458
-0.6156899 -0.46011144 0.21139167 0.06182204 -0.01539286 -0.58107877
-1.0684793 1.0892335 -0.06460539 -0.11071575 -0.28567505 0.3924164
0.45253068 -0.18823484 0.6940144 -0.06446687 -0.41407046 0.3975347
-0.9818018 0.2536561 0.6029924 0.85566294 1.0656564 -0.38928315
-0.38666508 1.2011235 0.9719026 0.5255952 0.12350313 -1.1908962
0.06786467 0.26131383 -0.16792594 -0.84966975 -0.50239724 -0.50242877
-0.78402454 0.15824333 0.2996112 0.13302669 -0.93141246 1.6846721
0.48429197 0.3229574 -0.10075384 0.61898404 0.80409247 0.67510384
0.5298962 -0.520794 1.328641 -1.3005861 -0.8406464 -0.2243817
0.92742556 -0.83749324 0.77482414 0.5095452 -0.56236494 -0.5001799
-0.8329935 0.04617365 -0.10621372 0.43750408 0.6211002 0.676335
-0.6081326 0.38264224 -0.55202824 -0.11123095 0.6389076 0.31878445
-0.3444009 -0.47052673 -0.6309815 0.693057 0.59076136 -0.06118448
-1.2084054 -0.6221004 -0.89682704 -0.30903694 0.470443 -0.5687386
1.0513903 -0.84804815 -1.4340494 1.3687731 -0.27725807 -0.14157039
0.37515697 -0.18742009 -0.05639686 0.26775196 0.42538154 0.87521404
1.0754092 -1.656763 -0.8201271 -0.2509723 -0.2451877 0.12831797
-0.12839726 0.20777577 -0.26836607 -0.45748332 0.28948176 -1.2179004
-0.19920847 -0.10869657 -0.55207264 -0.58064264 0.8347487 -0.30103374
0.9265113 -2.107486 0.34417668 0.08456351 0.21060154 -0.08719976
-0.08044633 0.18407378 0.1415821 0.03189585 -0.22093615 -0.8909075
0.06459239 0.33551773 -0.07229318 0.9523381 0.00648637 0.6898744
-0.79018635 -1.2723211 0.29120666 0.36672685 -0.601252 0.17468454
-0.38819686 0.8612007 -0.5059396 0.8600639 0.74833685 -0.8480559
0.30419827 -0.3493894 -0.1074393 -0.34462264 -1.0903053 1.9418437
-0.32725605 0.12158984 0.17479399 -1.0372117 0.6384875 0.33890387
0.69880515 -0.1257826 0.21044146 0.27790734 -0.789887 -0.42336804
0.04644172 -0.34326342 -0.0295778 0.41445553 0.31658718 -0.06024833
0.2767007 0.1459195 0.6520492 0.42054954 0.32928047 -0.31209415
0.66356456 -0.26277167 0.8101823 0.9906024 -0.26558018 0.55392015
0.3753469 -0.22158132 -0.5936153 -0.97509736 -0.6461396 1.4300542
0.12420004 -0.4968614 -0.4970999 -1.0003136 -0.11763012 0.5244568
-0.48237345 0.25581282 -0.22648197 -1.147575 0.34590897 0.1365805
0.59893364 -0.732657 -0.16548097 0.17914017 -0.3517425 -1.2097287
-0.36179596 0.42313313 -0.69522786 -1.0014354 -0.6206499 -0.9642718
0.59456074 0.11710739 1.0334675 -0.20473658 -0.4049628 0.48603314
-0.1644301 -0.0172815 -0.29823515 0.02947275 -0.13225655 0.4330544
0.28800586 -0.31510007 -0.3690683 0.45962465 -0.76701313 0.39219838
0.6711149 1.2472374 1.2070355 -0.09050843 0.4745519 -1.315172
0.16974223 -0.471269 -0.6619192 0.6351729 -0.94174224 0.085942
0.25027585 -0.36730158 -1.2183703 0.67400885 0.05560965 -0.48339164
-0.45119026 0.5671579 -0.06499384 -0.3245887 0.28999105 0.18275718
-0.26946303 -0.5789457 0.4322193 0.6955411 0.21160032 -0.76877403
0.17561826 0.55250615 0.49114218 -0.5135176 -1.757402 -0.93064463
-0.83118016 -0.42719474 1.3204772 -1.1213623 -0.74987787 0.6735519
-0.970224 -0.05861717 0.4100593 0.5980128 -0.78280175 0.50831395
-0.85098344 -0.6660359 -0.04389412 -1.4388391 1.3580507 0.1242292
0.14477085 -1.1663406 0.18109563 0.94233626 -0.06512 0.27437264
0.8760364 -0.37064677 -0.5809129 -0.16580407 -0.57005274 0.52273035
-0.24981318 -0.23103984 -1.2718909 -0.46835208 0.02018526 -0.9352021
1.0912426 0.5910521 1.2676072 0.05420265 -0.56020766 0.59714377
1.742538 -0.33816135 -0.01372834 0.06047681 1.0595505 0.6366404
0.8464125 0.4965748 0.4607495 0.73710394 0.5241941 0.12242807
-0.02418733 -0.01568083 0.05526867 0.8538963 0.31883746 -0.14796655
-0.8281447 0.22092189 -2.1050766 -0.5509347 -0.49537542 1.7073023
0.81361115 -0.12413222 -0.20621215 -0.19796875 0.7657261 0.28972322
-0.56523496 0.12956491 -0.22219618 -0.18945375 0.66905284 0.33488154
-1.6443399 0.8756367 6.475919 1.3483074 -0.67287916 0.5177853
0.6482927 0.17662823 0.03105079 0.17799293 -1.2724137 0.29191256
0.7738001 0.56987816 0.08558095 0.8754737 -0.21055639 -0.37323818
-1.0458847 1.2343704 0.36051363 -1.0293007 -0.12480586 0.24975772
1.1129136 -0.0235448 0.22341031 0.21172883 0.296529 -0.9495084
0.6077414 0.5553444 0.777574 -0.5053678 0.55733025 0.4086661
-1.0850785 -0.1427811 -0.29985842 0.22997452 0.3495832 0.78646666
-0.40088496 0.49593818 0.4924611 1.1439447 -0.68968296 0.83668965
-0.2874336 0.7418335 -0.09573208 0.21034655 0.34514868 -0.4429142
0.21117482 1.1518512 -0.09859516 -0.42361277 0.7720196 0.64433664
-0.04789995 0.44274992 -0.27572733 0.06858014 -0.02072561 1.5255227
-0.82420284 -0.26327512 -0.7981357 1.012555 0.587887 0.2969438
-0.92355454 0.25427037 0.07963462 -0.44558242 0.17531659 -0.05348314
-0.23510171 -1.1924194 -0.37331912 -0.7050121 0.62391555 -0.68141544
-1.4866529 0.07659473 0.1028879 -1.0150772 -0.00734498 -0.8038622
0.03165817 0.6030051 -1.7657615 -1.7644455 -0.09822611 0.48706457
0.49684745 -0.06079479 1.0073252 0.5757808 -0.681392 0.3185121
0.5332546 -0.29965347 0.9464188 -1.3080118 -0.70768255 0.42230615
0.293706 0.14786032 0.4988971 -0.49604958 -0.91824037 -1.1837486
0.7315398 -0.2908037 0.48658872 -0.1840247 -0.61850464 0.63383764
0.1820463 0.35539582 1.2223406 0.22387008 -0.5742765 0.1011049
-0.97415084 0.14620249 0.62873435 -0.6781339 -0.09004715 0.8299122
0.3437606 -0.13099028 -1.1523902 0.57344675 0.625904 -0.81616765
0.937085 -0.3520414 0.42713928 -0.41179407 -0.57142127 -1.0278643
-0.43516752 -0.02584223 -0.03279377 1.4013612 0.2818485 -0.40225613
0.6000993 0.27025616 0.0367065 -0.79488736 -0.79528874 -0.44826633
0.06377798 -0.3409704 -0.0545858 0.889142 -0.5329541 0.48789492
-0.9182312 0.13502148 1.3567501 0.4393705 0.4133968 -1.4211915
-0.43915492 -0.19918606 -0.17194556 -1.0547614 0.8629391 -1.184007
0.2066997 -1.317956 0.95441365 -0.50465465 -0.5298812 0.46787512
-0.28404275 0.62014186 0.17263395 0.49523675 -1.2977593 0.50326556
1.2153076 -0.2905102 0.47360238 0.03719473 -0.2505165 0.85894614
0.6627598 -1.0254102 -0.2317521 -0.13669442 0.32468718 0.04712647
0.35419592 -0.8741153 0.17014644 -0.32376334 0.2049179 -0.8838088
0.47978672 -0.65232086 0.19935568 0.21547836 -0.6174934 -0.552268
-0.26456556 0.9992838 -0.28972444 -0.69254535 1.1722811 -0.16401277
-0.6839163 0.3788066 -0.49233255 -0.02409464 1.0419273 0.2700595
-0.07863791 0.06937819 -0.7795061 0.3986686 0.2403454 0.10547262
0.4069518 -1.4285383 -0.7127841 0.00748322 0.41811985 -0.0982646
0.37900117 1.030118 -0.2613061 0.3755547 0.11041878 -1.101997
-1.6226845 0.501659 0.1866414 -0.6072682 -0.24026911 0.93643785
0.5225024 -0.67472106 0.18467869 0.32672834 -0.32736263 0.30492967
0.2856002 0.24949726 -0.3360777 -0.9290496 -0.20760755 1.0152099
-0.33022815 0.01854166 1.0249783 -0.5016777 -0.357082 0.8771131
1.5748398 -0.45953342 -1.3262677 -0.33777586 0.21592447 -0.4063471
0.42257333 -0.6159811 -0.872919 0.83600754 0.8104444 -0.13108145
0.84022653 0.08301762 0.36064774 0.2252399 0.33155194 -1.1592865
0.34054816 0.36545733 0.37988314 -1.8163373 0.2131179 -0.6400215
-0.53746814 0.9274141 0.3669644 0.22253463 1.032926 -0.06823897
0.13676414 -0.33144405 -0.666002 -0.16705401 0.32046112 0.24862595
0.43642738 0.12320177 -0.46335593 0.13078943 0.63259166 -0.08715794
0.04918268 0.87213874 -0.4176586 -1.3121005 -0.3109499 0.5995662
-0.92781204 -0.04398885 0.21692987 0.23241267 0.21792197 1.093035
-0.44351947 -0.01851441 -0.42789394 0.3579789 0.6219358 -0.7296321
-0.26001412 0.45862877 -0.09413169 -0.49082795 -1.0995146 -0.96724564
-0.9463786 0.25781572 -0.7758038 -0.01435907 0.89750695 1.110777
0.00902707 0.36237878 0.53324383 -0.81232214 -0.53218055 -1.0015397
-0.98047227 0.5208028 0.07995777 -0.92982095 -0.48353353 0.2662529 ] | [9.507450103759766, 4.1321563720703125] |
fd910602-ab01-4597-9b2e-390ec55b5c2e | memory-efficient-episodic-control | 1911.09560 | null | https://arxiv.org/abs/1911.09560v1 | https://arxiv.org/pdf/1911.09560v1.pdf | Memory-Efficient Episodic Control Reinforcement Learning with Dynamic Online k-means | Recently, neuro-inspired episodic control (EC) methods have been developed to overcome the data-inefficiency of standard deep reinforcement learning approaches. Using non-/semi-parametric models to estimate the value function, they learn rapidly, retrieving cached values from similar past states. In realistic scenarios, with limited resources and noisy data, maintaining meaningful representations in memory is essential to speed up the learning and avoid catastrophic forgetting. Unfortunately, EC methods have a large space and time complexity. We investigate different solutions to these problems based on prioritising and ranking stored states, as well as online clustering techniques. We also propose a new dynamic online k-means algorithm that is both computationally-efficient and yields significantly better performance at smaller memory sizes; we validate this approach on classic reinforcement learning environments and Atari games. | ['Anil Anthony Bharath', 'Pierre Richemond', 'Marta Sarrico', 'Kai Arulkumaran', 'Andrea Agostinelli'] | 2019-11-21 | null | null | null | null | ['online-clustering'] | ['computer-vision'] | [-3.69731963e-01 -3.08324665e-01 -1.12322234e-01 1.44341774e-03
-5.19011855e-01 -4.36593503e-01 4.92348373e-01 2.87583202e-01
-8.43848348e-01 1.27492440e+00 1.09881256e-02 1.55564649e-02
-7.59939849e-01 -9.44247007e-01 -6.98037088e-01 -8.48955154e-01
-4.28216726e-01 8.02381158e-01 4.91401255e-01 -2.71994084e-01
4.73924667e-01 4.68685657e-01 -2.07066417e+00 -1.12919606e-01
8.93601596e-01 9.12460208e-01 4.09701794e-01 5.55364549e-01
-1.01345770e-01 9.01580036e-01 -6.43108249e-01 1.69960652e-02
1.03564769e-01 -6.38094842e-01 -8.46022248e-01 -1.11089520e-01
-2.30838567e-01 -3.46660346e-01 -6.18964672e-01 7.64158010e-01
5.19711494e-01 8.13431919e-01 3.82684678e-01 -9.43588257e-01
-3.62438917e-01 5.69983959e-01 -1.28956929e-01 6.17928147e-01
-1.58326179e-02 8.95574912e-02 4.54718262e-01 -4.15588379e-01
4.59234834e-01 8.57151926e-01 6.83640778e-01 7.33064055e-01
-1.24923050e+00 -4.20978814e-01 1.95055649e-01 7.13572919e-01
-1.38569939e+00 -5.41040123e-01 4.11252260e-01 2.25298777e-01
1.26313829e+00 7.99707994e-02 1.18275118e+00 9.66272235e-01
1.91972986e-01 9.05080795e-01 1.26152539e+00 -3.86959165e-01
1.08718812e+00 -1.40339419e-01 -3.18868011e-01 5.78825295e-01
1.39863536e-01 5.10645390e-01 -7.58309186e-01 -2.68465400e-01
7.51994193e-01 6.67803064e-02 1.38958087e-02 -6.49562299e-01
-7.41805017e-01 8.06854248e-01 1.18791118e-01 4.64694083e-01
-6.42367899e-01 5.47429502e-01 2.98459768e-01 6.34549916e-01
5.28548732e-02 5.67145109e-01 -4.53221977e-01 -7.64785707e-01
-1.17595482e+00 3.31682593e-01 8.57169449e-01 6.51188970e-01
6.44652069e-01 3.87962103e-01 -3.73148695e-02 7.59791255e-01
-1.72424480e-01 3.83954167e-01 1.00656009e+00 -1.24419105e+00
1.13280704e-02 1.31727159e-01 4.70925421e-01 -7.78702319e-01
-4.07337040e-01 -3.94778907e-01 -6.13499939e-01 2.26904884e-01
3.08100909e-01 -5.27665988e-02 -1.02057791e+00 1.89163363e+00
3.15132082e-01 3.49262744e-01 7.27402419e-02 8.02947998e-01
-2.06533685e-01 6.69520795e-01 1.48473829e-01 -5.33135414e-01
6.21414065e-01 -5.88326991e-01 -7.91258097e-01 -8.23657289e-02
6.18592739e-01 4.83815446e-02 9.58897889e-01 6.50563061e-01
-1.43382788e+00 -3.11636984e-01 -1.03244936e+00 4.28546101e-01
-5.25558591e-01 -4.08703595e-01 6.29922092e-01 7.05213308e-01
-1.40240073e+00 1.21485782e+00 -1.30895710e+00 -3.65225554e-01
3.48301619e-01 5.77089012e-01 1.78575724e-01 2.54454408e-02
-1.32052565e+00 1.09740961e+00 7.89825618e-01 -1.80405244e-01
-1.28643525e+00 -4.60185289e-01 -3.51376295e-01 2.85990685e-01
7.00454175e-01 -4.39122945e-01 1.22819126e+00 -8.52872789e-01
-1.86007428e+00 2.37159923e-01 8.47959891e-02 -9.70882952e-01
3.16316336e-01 -3.25793415e-01 -3.06356788e-01 1.02505051e-01
-5.64732850e-01 4.75878298e-01 8.02435577e-01 -9.27781284e-01
-5.17898917e-01 -3.78617525e-01 -8.30835178e-02 4.39128160e-01
-6.18941844e-01 -4.35710758e-01 -1.91157013e-01 -4.20693010e-01
-4.69166599e-02 -8.97857189e-01 -4.80930746e-01 -1.55226082e-01
4.41664338e-01 -9.50308889e-02 5.99034369e-01 -4.94225204e-01
1.35079730e+00 -1.94973505e+00 3.46282870e-01 4.22019064e-01
-1.59697041e-01 5.46336591e-01 -4.96289060e-02 6.20245993e-01
3.94961923e-01 -3.11437875e-01 -1.69260502e-01 1.00315519e-01
6.21781871e-02 7.61722803e-01 -2.51753539e-01 2.25910798e-01
-2.36921847e-01 9.78684247e-01 -9.91075695e-01 -4.11669701e-01
1.68313041e-01 2.23017141e-01 -5.47998250e-01 1.61782578e-01
-4.48131144e-01 5.97719625e-02 -3.37894619e-01 2.64845759e-01
1.97877929e-01 -2.09996313e-01 7.44779706e-01 3.61089200e-01
-5.27656339e-02 1.26092657e-01 -1.38163626e+00 1.82565641e+00
-5.96519768e-01 2.20061809e-01 -6.57653138e-02 -1.41932738e+00
8.48992646e-01 1.42032877e-01 5.93352914e-01 -1.30511665e+00
2.36434609e-01 1.32957429e-01 -2.77098864e-01 -2.54266679e-01
5.86009204e-01 1.36947960e-01 1.87311456e-01 6.03357375e-01
2.37525940e-01 1.61030330e-02 2.64163136e-01 2.53506061e-02
1.30293012e+00 2.44138360e-01 2.44042128e-01 -4.41484720e-01
2.33538732e-01 1.44981995e-01 4.99171793e-01 1.17520225e+00
-1.95172176e-01 1.95060164e-01 2.32485026e-01 -5.64154446e-01
-9.79505301e-01 -1.16059673e+00 3.74477535e-01 1.16968584e+00
1.39133409e-01 -4.55516398e-01 -7.59628832e-01 -2.66802460e-01
8.37165639e-02 7.94808447e-01 -5.22944391e-01 -7.23145068e-01
-6.65241659e-01 -8.32297444e-01 3.41781765e-01 6.48724258e-01
5.28237700e-01 -1.22677171e+00 -1.20151806e+00 7.10889280e-01
1.21666975e-01 -4.63835448e-01 8.53964686e-03 7.11726904e-01
-1.29159939e+00 -8.70658219e-01 -5.63208163e-01 -6.27786100e-01
3.45689684e-01 5.10962456e-02 1.20974588e+00 2.78877497e-01
-2.96079665e-01 6.12731099e-01 -2.83133566e-01 4.28877119e-03
-1.92202240e-01 -8.83249007e-03 3.39931428e-01 -3.35940033e-01
2.33112082e-01 -7.07397997e-01 -7.23212063e-01 8.54242370e-02
-9.58240867e-01 -2.88423568e-01 5.84693611e-01 1.21037340e+00
7.99218178e-01 5.51169991e-01 7.70556152e-01 -5.38050056e-01
8.38605523e-01 -5.46094656e-01 -8.05404007e-01 3.90348136e-01
-1.09740722e+00 4.02522773e-01 8.81348670e-01 -6.12699032e-01
-1.15651846e+00 -6.18805364e-02 1.89362206e-02 -7.05259979e-01
-2.13186350e-02 5.67123771e-01 2.49639317e-01 1.40600577e-01
6.06399059e-01 8.67106140e-01 1.40936151e-01 -3.98296833e-01
2.74338990e-01 2.93131888e-01 4.24936175e-01 -8.37533057e-01
3.33788842e-01 2.69797564e-01 -1.95151433e-01 -5.96594155e-01
-6.12251043e-01 -1.68729857e-01 -3.34740937e-01 -2.04627782e-01
3.82076502e-01 -6.87196076e-01 -7.09073961e-01 4.37602252e-01
-5.46914756e-01 -8.23401630e-01 -6.40242100e-01 3.49195480e-01
-1.11637616e+00 1.18217386e-01 -7.82204688e-01 -1.08140314e+00
-1.81996956e-01 -6.16416633e-01 3.60006779e-01 3.81858826e-01
-7.43600950e-02 -8.26661944e-01 4.71612811e-01 -1.38695657e-01
7.38320649e-01 -5.71701750e-02 9.63888824e-01 -4.73046571e-01
-3.98869544e-01 3.53580326e-01 2.89387017e-01 -1.14486344e-01
-2.04345331e-01 -5.04550815e-01 -5.82825303e-01 -6.75691485e-01
-7.98736047e-03 -6.43088818e-01 8.13518405e-01 3.35193127e-01
1.49929452e+00 -4.45339680e-01 -1.87319651e-01 2.27125123e-01
1.59910965e+00 5.72933197e-01 6.42617941e-01 5.45247793e-01
-2.81468742e-02 4.27124828e-01 7.55842745e-01 9.12253082e-01
2.77277052e-01 5.13565302e-01 3.62571359e-01 6.06618345e-01
4.37861010e-02 -3.06251377e-01 3.66705656e-01 9.87810850e-01
-8.49303529e-02 -1.96493492e-01 -7.50582755e-01 8.99149477e-01
-2.24843931e+00 -1.25687802e+00 6.48823380e-01 2.46989298e+00
1.18587780e+00 1.47492468e-01 2.89894909e-01 2.93572307e-01
4.20142323e-01 -9.27466527e-02 -1.13133502e+00 -4.42610502e-01
-8.00289959e-02 6.53822839e-01 2.90160626e-01 6.40047714e-02
-7.68584847e-01 1.08090007e+00 6.97735071e+00 9.26174760e-01
-9.77220535e-01 1.99308261e-01 4.96980309e-01 -5.88731825e-01
-1.45208631e-02 -2.98858225e-01 -2.94616967e-01 6.62868261e-01
1.67462122e+00 -2.25993454e-01 1.14773500e+00 9.35732365e-01
-2.21823016e-03 -5.27148366e-01 -7.02664793e-01 1.18444860e+00
-1.65451050e-01 -1.50233400e+00 -1.91931471e-01 -1.19989514e-01
7.72716761e-01 5.50693125e-02 1.36254862e-01 5.11553347e-01
7.19498158e-01 -9.06308234e-01 6.42064393e-01 7.20758319e-01
5.13814926e-01 -1.34710836e+00 3.85287166e-01 3.71380657e-01
-9.19339240e-01 -7.20981300e-01 -6.86077654e-01 -1.06712133e-01
-1.03066765e-01 2.42229998e-01 -9.66565311e-02 2.04497650e-01
1.02984464e+00 3.51939410e-01 -4.06664759e-01 1.19675040e+00
1.54440269e-01 4.94149566e-01 -5.07203937e-01 -4.48688924e-01
3.88891846e-01 -1.94193900e-01 1.51156887e-01 6.15943372e-01
4.98558670e-01 3.12123030e-01 -6.90048561e-02 6.58064425e-01
2.97576845e-01 -1.79095328e-01 -4.70997095e-01 -1.14323050e-01
8.18869114e-01 7.59918332e-01 -1.01866508e+00 -3.57061654e-01
4.31891419e-02 9.44426179e-01 7.42640913e-01 2.18163475e-01
-8.89202535e-01 -3.90413791e-01 5.50501525e-01 -5.18134795e-02
8.11506510e-01 -7.08754778e-01 1.87941357e-01 -9.71777618e-01
-3.39563549e-01 -9.68329608e-01 5.50620377e-01 -4.87211734e-01
-8.24834585e-01 3.83312762e-01 4.57594171e-02 -8.71364653e-01
-6.54462159e-01 -2.49460384e-01 -3.11868280e-01 -5.89854270e-02
-1.39938736e+00 -1.84841573e-01 -3.35271135e-02 1.03075731e+00
4.40805376e-01 -2.32673451e-01 9.82540071e-01 1.62213787e-01
-4.94466901e-01 5.10463178e-01 9.11804378e-01 -4.85504240e-01
2.44809091e-01 -1.29268396e+00 -5.69152832e-02 5.03423095e-01
2.96193153e-01 5.04227698e-01 7.81701565e-01 -4.81973946e-01
-1.91440868e+00 -8.36477935e-01 3.54652971e-01 1.94833595e-02
5.57754517e-01 -2.50332147e-01 -1.09114134e+00 1.88094333e-01
1.24188997e-01 -2.06428975e-01 5.57624161e-01 1.20467588e-01
2.25265697e-02 -1.42147422e-01 -1.12100661e+00 6.35658920e-01
1.19892776e+00 -4.50272441e-01 -6.85450494e-01 1.61952570e-01
1.67537048e-01 -3.19024503e-01 -8.26169431e-01 -1.79953158e-01
3.11116785e-01 -1.13338482e+00 9.22084868e-01 -6.28976166e-01
-1.14509255e-01 -9.74986553e-02 1.20854583e-02 -1.54638672e+00
-4.21840340e-01 -6.43906474e-01 -9.05460000e-01 7.54522800e-01
-3.84462848e-02 -4.52109069e-01 1.06021857e+00 6.54680789e-01
1.01093106e-01 -8.90712321e-01 -1.14167750e+00 -1.22463965e+00
5.37896268e-02 -1.38524041e-01 6.74961150e-01 7.13089824e-01
7.32409954e-02 -2.19087005e-01 -4.59445894e-01 -2.23348692e-01
5.26194274e-01 3.95418435e-01 2.47541964e-01 -1.05642045e+00
-4.71355826e-01 -2.47050866e-01 -9.89615470e-02 -7.86188483e-01
1.50025636e-01 -4.47082788e-01 1.21815622e-01 -1.22285604e+00
-5.17724194e-02 -5.76228678e-01 -8.25326979e-01 6.44007742e-01
2.38721251e-01 1.33211529e-02 1.50356606e-01 2.19641671e-01
-1.09150445e+00 1.03281307e+00 7.36796796e-01 1.05577938e-01
-4.27013397e-01 -3.03550720e-01 -2.48169884e-01 3.01329851e-01
1.10221601e+00 -5.93928516e-01 -8.58940005e-01 -3.93865198e-01
2.77341068e-01 3.10762614e-01 1.58930391e-01 -1.61014056e+00
6.09831333e-01 -2.46123910e-01 5.89889884e-01 -5.26439786e-01
5.10079563e-01 -8.12893927e-01 3.35371524e-01 9.11568820e-01
-4.87774402e-01 4.56860781e-01 2.38390058e-01 9.74048078e-01
-4.25641574e-02 -3.49637002e-01 8.34674478e-01 -3.97444159e-01
-1.00905788e+00 1.88431874e-01 -8.69326532e-01 1.26618400e-01
1.22944605e+00 -3.14018488e-01 -7.12759867e-02 -5.44348478e-01
-1.01678801e+00 3.72223049e-01 4.10732925e-01 3.33337486e-01
7.51866758e-01 -1.13984084e+00 -2.42567975e-02 4.86390442e-02
-3.73126477e-01 -4.37928289e-01 4.12978649e-01 6.67151570e-01
-4.21916485e-01 2.87440628e-01 -6.96423352e-01 -1.77113101e-01
-8.04062068e-01 8.52505326e-01 4.21507746e-01 -3.08262765e-01
-6.05634570e-01 3.85437429e-01 -6.18563354e-01 2.80563366e-02
5.04487157e-01 7.43539482e-02 -2.08778009e-01 1.67950407e-01
5.96837461e-01 6.90456152e-01 2.31713325e-01 1.09658487e-01
-1.68650374e-01 6.50237724e-02 -1.96616650e-01 -2.96887994e-01
1.58707762e+00 -2.84256220e-01 1.80134743e-01 4.08923298e-01
6.44838691e-01 -7.53632128e-01 -1.40298045e+00 -2.78040797e-01
4.38929707e-01 -3.40181857e-01 3.12656730e-01 -8.29265058e-01
-1.17472613e+00 5.21172404e-01 9.00961816e-01 1.19150832e-01
1.22104931e+00 -3.77837718e-01 9.24948633e-01 9.95383263e-01
8.80216599e-01 -2.00638986e+00 3.65079850e-01 7.09570408e-01
4.58345681e-01 -8.12747359e-01 -4.11826931e-02 6.47131681e-01
-6.07319236e-01 1.04348683e+00 7.65719175e-01 -2.88264751e-01
5.99833548e-01 2.09170029e-01 -2.77610928e-01 -4.27512601e-02
-1.14348316e+00 -3.76740366e-01 -4.66402084e-01 7.11728990e-01
-2.19562113e-01 -3.07645023e-01 -3.92497957e-01 4.14871275e-01
-8.10800679e-03 1.30226716e-01 5.19667745e-01 1.42203665e+00
-6.54978454e-01 -1.05059564e+00 -9.34279487e-02 5.15372276e-01
-2.10084096e-01 1.65373087e-01 -3.39735448e-02 6.67229950e-01
-3.67282301e-01 7.85646200e-01 2.58280516e-01 -2.74396718e-01
2.71325465e-02 2.57990837e-01 7.02364028e-01 -1.00471474e-01
-7.07538784e-01 -1.15223706e-01 -1.17884271e-01 -9.96473312e-01
-2.55623698e-01 -6.87519312e-01 -1.46166217e+00 -4.57717896e-01
-1.00271039e-01 4.24822927e-01 5.39665103e-01 7.53189266e-01
8.05914164e-01 2.30789274e-01 4.92531776e-01 -6.56266212e-01
-1.04657018e+00 -4.54594553e-01 -7.85657823e-01 2.64172316e-01
-1.06749780e-01 -1.00528240e+00 -5.88358194e-02 -3.43618900e-01] | [4.078391075134277, 1.8987270593643188] |
0d052a32-2707-4aa2-819d-14e541983fd6 | can-label-noise-transition-matrix-help-to | null | null | https://openreview.net/forum?id=c0AD3ll9Wyv | https://openreview.net/pdf?id=c0AD3ll9Wyv | Can Label-Noise Transition Matrix Help to Improve Sample Selection and Label Correction? | Existing methods for learning with noisy labels can be generally divided into two categories: (1) sample selection and label correction based on the memorization effect of neural networks; (2) loss correction with the transition matrix. So far, the two categories of methods have been studied independently because they are designed according to different philosophies, i.e., the memorization effect is a property of the neural networks independent of label noise while the transition matrix is exploited to model the distribution of label noise. In this paper, we take a first step in unifying these two paradigms by showing that modelling the distribution of label noise with the transition matrix can also help sample selection and label correction, which leads to better robustness against different types of noise. More specifically, we first train a network with the loss corrected by the transition matrix and then use the confidence of the estimated clean class posterior from the network to select and re-label instances. Our proposed method demonstrates strong robustness on multiple benchmark datasets under various types of noise. | ['Masashi Sugiyama', 'Gang Niu', 'Bo Han', 'Mingming Gong', 'Alan Blair', 'Tongliang Liu', 'Xuefeng Li', 'Yu Yao'] | 2021-09-29 | null | null | null | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 5.03779471e-01 2.73098927e-02 -1.67286262e-01 -5.95664322e-01
-7.28997409e-01 -3.71924609e-01 5.90146005e-01 3.90373528e-01
-5.82142353e-01 8.62570345e-01 6.56744912e-02 4.77372073e-02
-1.27310187e-01 -7.80088961e-01 -7.49240577e-01 -1.18944895e+00
3.67210627e-01 1.82916984e-01 1.75203666e-01 2.14305803e-01
3.84088755e-01 2.39920348e-01 -1.82202625e+00 1.14068404e-01
9.58252132e-01 1.15757203e+00 -9.56031978e-02 1.77971169e-01
-2.65538335e-01 7.83444703e-01 -5.92461228e-01 -1.70497507e-01
2.19813928e-01 -5.94122469e-01 -7.09801674e-01 -1.67808294e-01
1.93401679e-01 1.72550663e-01 -3.18804309e-02 1.60940313e+00
7.32448518e-01 3.54650468e-01 8.85807097e-01 -9.10677731e-01
-5.98716974e-01 8.64826560e-01 -4.31858450e-01 -1.39550820e-01
7.16584176e-02 2.71869870e-03 7.97330081e-01 -8.30450296e-01
5.41921198e-01 1.09923208e+00 9.08616006e-01 8.04812789e-01
-1.41034985e+00 -8.46913338e-01 2.28852957e-01 3.45489115e-01
-1.34791565e+00 -4.06420767e-01 8.38681221e-01 -4.85393524e-01
2.77306527e-01 6.80508390e-02 3.75381172e-01 1.12728429e+00
1.79661766e-01 7.18874812e-01 1.47959268e+00 -6.30891860e-01
5.24019361e-01 6.28710568e-01 4.66775060e-01 5.04618347e-01
1.83746114e-01 1.56473234e-01 -5.80429018e-01 -1.28725410e-01
1.76732376e-01 -1.87842295e-01 -3.75138670e-01 -3.77894819e-01
-8.30765545e-01 7.99016118e-01 3.99455935e-01 3.11804414e-01
-5.72642796e-02 4.64329906e-02 5.53410769e-01 2.78594732e-01
6.84171617e-01 3.85727674e-01 -2.62122065e-01 3.96155030e-01
-1.06934869e+00 -9.11867770e-04 6.33825541e-01 6.56071544e-01
8.57441783e-01 1.79620087e-02 -4.17789191e-01 1.06704390e+00
4.90191698e-01 3.69544238e-01 7.13172615e-01 -6.21134162e-01
2.41352469e-01 2.11933926e-01 1.03759147e-01 -7.88613439e-01
-5.60399830e-01 -7.64652193e-01 -1.11031544e+00 3.08845937e-01
5.80834687e-01 -1.95176035e-01 -1.01156723e+00 2.20208740e+00
2.50294775e-01 2.86547422e-01 1.27388593e-02 4.71810699e-01
7.96654880e-01 4.90166485e-01 1.15190275e-01 -4.35432583e-01
1.08577216e+00 -8.65898430e-01 -1.03892028e+00 -1.11757018e-01
7.63024151e-01 -5.27621865e-01 8.69430065e-01 5.82393587e-01
-9.31722760e-01 -6.31676614e-01 -1.16066802e+00 2.00638235e-01
-4.49671477e-01 1.18316352e-01 1.44812331e-01 7.36496806e-01
-1.00345981e+00 9.91960287e-01 -4.70210791e-01 -3.19556117e-01
4.08387423e-01 2.23211586e-01 -1.64308324e-01 -9.08677280e-03
-1.30723798e+00 8.88284564e-01 4.74456817e-01 6.38989508e-02
-6.98229015e-01 -5.10875344e-01 -6.24691904e-01 9.26248431e-02
2.44787529e-01 -5.18467724e-01 1.11266208e+00 -1.06005406e+00
-1.59981465e+00 7.73900330e-01 -1.37330815e-01 -4.81635183e-01
6.45064771e-01 -2.21435100e-01 -1.31753474e-01 -2.43787691e-01
9.95706469e-02 4.08695191e-01 9.79369640e-01 -1.55247247e+00
-7.86260188e-01 -3.15147161e-01 -4.33004797e-01 5.56294136e-02
-3.49484384e-01 -3.63347888e-01 8.48043896e-03 -7.11917222e-01
2.42949441e-01 -7.64212489e-01 8.50866511e-02 -2.06726819e-01
-5.42719662e-01 -3.85623604e-01 5.23038030e-01 -4.53480661e-01
1.24716842e+00 -2.31890988e+00 -9.02761221e-02 3.71406227e-01
3.59834135e-02 4.20683533e-01 -9.87454578e-02 2.40533531e-01
-3.56423914e-01 2.17427343e-01 -2.68828511e-01 -5.61232984e-01
5.96306361e-02 -6.95079565e-02 -4.30609614e-01 4.75398898e-01
2.61008181e-02 5.65766573e-01 -9.88712251e-01 -2.77144015e-01
-1.10450305e-01 4.15364504e-01 -1.33998737e-01 1.61253422e-01
-2.69970478e-04 6.56233490e-01 -2.03962997e-01 2.34709889e-01
8.16452324e-01 4.00698790e-03 2.06443340e-01 -3.53900909e-01
1.12776324e-01 3.49224478e-01 -1.36984432e+00 1.30853367e+00
-2.29767844e-01 2.16043249e-01 2.84881406e-02 -1.07670259e+00
1.00244248e+00 3.23696107e-01 3.04876089e-01 -5.71032643e-01
3.59852374e-01 3.99557978e-01 -1.93741500e-01 -3.57440203e-01
1.96689337e-01 -4.61891860e-01 1.20758675e-01 7.50074148e-01
2.34672010e-01 3.15708630e-02 1.42416909e-01 -5.63636869e-02
8.42393339e-01 1.18817449e-01 1.91128537e-01 -3.06646377e-01
4.76562619e-01 -5.37584066e-01 7.77698576e-01 1.17824948e+00
-2.85631806e-01 4.76168752e-01 3.34842145e-01 -4.14090663e-01
-7.74828613e-01 -1.01338387e+00 -3.63778442e-01 1.04976988e+00
2.06047282e-01 -1.08411506e-01 -8.40496302e-01 -1.00856864e+00
-2.53124833e-01 8.16459060e-01 -8.68935227e-01 -6.51596665e-01
-2.40574971e-01 -1.04326355e+00 4.93743062e-01 4.31042165e-01
5.93377113e-01 -1.08941078e+00 -2.07548067e-01 6.20316602e-02
-3.10934782e-01 -4.64735270e-01 -2.64396399e-01 7.78270960e-01
-7.94868290e-01 -9.88212049e-01 -4.66240019e-01 -7.64153719e-01
7.12921321e-01 6.29008338e-02 9.89232659e-01 1.68206140e-01
2.50562638e-01 1.26810446e-01 -4.36359704e-01 -5.26262522e-01
-4.98715192e-01 1.46419168e-01 2.23221496e-01 2.17351049e-01
3.13467413e-01 -6.78100228e-01 -2.04446629e-01 1.13970198e-01
-1.09609485e+00 -1.00831665e-01 4.23214257e-01 1.11814868e+00
6.62526906e-01 3.12088162e-01 8.04501414e-01 -1.28653491e+00
6.69105232e-01 -5.10044694e-01 -3.72250885e-01 3.55637491e-01
-8.90579581e-01 3.59987348e-01 5.72039425e-01 -5.29999435e-01
-1.15445924e+00 2.09669754e-01 -3.36780161e-01 1.54068530e-01
-2.26072833e-01 3.63932729e-01 -2.16371194e-01 -8.35194215e-02
6.33213222e-01 3.30217451e-01 -2.24500239e-01 -5.99651277e-01
3.15577388e-01 5.53110600e-01 3.05721879e-01 -4.96072948e-01
7.12514818e-01 4.27412003e-01 1.57135472e-01 -3.12470973e-01
-1.33078182e+00 -2.01912045e-01 -5.76570153e-01 -2.25577399e-01
6.94683611e-01 -5.54474235e-01 -4.83315289e-01 8.16399693e-01
-8.93557668e-01 -3.13907295e-01 -6.08549535e-01 5.18627405e-01
-4.75426257e-01 4.41355169e-01 -7.26683915e-01 -8.85422647e-01
-3.66593421e-01 -1.14050901e+00 6.11175716e-01 3.93671602e-01
1.03407525e-01 -9.45552051e-01 7.29726478e-02 -1.53280064e-01
4.71588969e-01 1.67779729e-01 1.15527570e+00 -9.21188533e-01
-8.01896304e-02 -3.18502754e-01 -2.46484265e-01 7.57343948e-01
1.31878987e-01 -2.10222036e-01 -1.31860077e+00 -3.38157058e-01
4.23531860e-01 -4.50870842e-01 1.27475333e+00 4.57294464e-01
1.10487354e+00 -1.12883776e-01 -2.99438238e-01 4.50351954e-01
1.38038504e+00 1.41817883e-01 5.84579229e-01 3.33245069e-01
5.64978957e-01 6.57276511e-01 4.39637125e-01 3.10946912e-01
1.26083940e-02 5.64192295e-01 4.05953050e-01 1.63138397e-02
-1.29443750e-01 -2.39110827e-01 1.43148258e-01 1.02710307e+00
2.89591283e-01 -2.30116099e-01 -5.92724979e-01 3.01368982e-01
-1.91892469e+00 -6.96380258e-01 -3.97335887e-02 2.51202464e+00
1.15995777e+00 3.47715557e-01 -4.09270599e-02 3.30842435e-01
9.08927619e-01 3.12240445e-03 -6.25430167e-01 -6.67460263e-02
-1.08839318e-01 7.30400383e-02 4.48723316e-01 5.60035348e-01
-1.12494087e+00 5.42288423e-01 6.37199497e+00 1.21159601e+00
-1.09543324e+00 2.11806208e-01 7.78387129e-01 1.02025665e-01
-2.23123133e-01 -6.28663599e-02 -8.77494216e-01 7.13628650e-01
6.47512317e-01 2.34458491e-01 1.98527202e-01 6.90081894e-01
-9.05928314e-02 -3.06493908e-01 -1.00590742e+00 7.48434186e-01
6.51823506e-02 -8.79947901e-01 -2.91786045e-02 -3.14660907e-01
8.71757507e-01 -2.21343488e-01 1.99508056e-01 3.86945158e-01
3.90894413e-01 -9.04222667e-01 8.79418314e-01 9.01483595e-01
5.42677283e-01 -8.26854885e-01 1.17660093e+00 5.07309198e-01
-8.10222387e-01 -1.20408133e-01 -3.99960607e-01 4.95400429e-02
9.83385593e-02 1.21560669e+00 -3.53456020e-01 5.22914052e-01
5.96986711e-01 6.17050946e-01 -6.70529962e-01 1.26694417e+00
-5.50585985e-01 8.12903166e-01 -1.86616153e-01 1.76312059e-01
-1.74556538e-01 -2.52877325e-01 4.40593302e-01 1.01333272e+00
3.24226230e-01 -5.76989114e-01 5.37147895e-02 7.30167389e-01
-2.15430275e-01 9.29243639e-02 -2.35919878e-01 3.89945894e-01
6.64036870e-01 1.15271699e+00 -9.56878483e-01 -4.01282132e-01
4.11635973e-02 6.52121007e-01 5.41975558e-01 3.36912185e-01
-6.69291317e-01 -5.54640770e-01 1.44988269e-01 -6.20191768e-02
1.55112877e-01 3.31088930e-01 -5.66807270e-01 -1.11740720e+00
-7.73999617e-02 -6.72209918e-01 1.64824128e-01 -5.00311911e-01
-1.55325389e+00 6.54037476e-01 -1.46507904e-01 -1.07089472e+00
1.13007605e-01 -4.11126494e-01 -4.95459437e-01 7.72224784e-01
-1.60782731e+00 -6.93147779e-01 -2.64865786e-01 1.70256674e-01
2.16815442e-01 1.47084385e-01 8.89845908e-01 3.68122637e-01
-6.65132880e-01 7.86873460e-01 6.08266354e-01 1.50071848e-02
1.14808428e+00 -1.44566441e+00 -1.29612535e-01 7.01070905e-01
4.65908088e-02 6.26991451e-01 8.13879669e-01 -6.29167497e-01
-5.12927651e-01 -1.21145093e+00 9.13815618e-01 -2.58595437e-01
2.95630574e-01 -5.20113528e-01 -1.12674785e+00 4.66778994e-01
6.02101162e-02 -7.03759864e-02 6.53498113e-01 7.66262561e-02
-5.48632026e-01 -3.76878738e-01 -1.25418282e+00 2.93938249e-01
7.53167391e-01 -3.65518391e-01 -3.72572988e-01 4.11153406e-01
5.04806101e-01 -1.53880984e-01 -5.81198812e-01 5.79302132e-01
4.68266517e-01 -1.10037732e+00 6.53266132e-01 -2.86217362e-01
1.68774977e-01 -3.44907671e-01 2.05846746e-02 -1.58542681e+00
-4.41347510e-01 -1.10368706e-01 1.12725146e-01 1.66014063e+00
3.13423812e-01 -8.16669703e-01 6.01593077e-01 3.66046071e-01
8.67145211e-02 -8.29301655e-01 -8.99206340e-01 -6.28517747e-01
1.14170149e-01 -3.42687368e-01 4.61174905e-01 9.48134780e-01
-2.66602248e-01 3.08136821e-01 -6.41366184e-01 -2.60818601e-01
6.14461184e-01 -2.85090774e-01 2.77722120e-01 -1.49568141e+00
-1.92149580e-01 -3.66762191e-01 -1.00235343e-01 -8.91275227e-01
2.55258828e-01 -9.66110826e-01 5.51575840e-01 -1.36708343e+00
3.01566899e-01 -6.30791843e-01 -6.55914903e-01 5.25816023e-01
-3.88069481e-01 4.75165099e-01 7.91632310e-02 3.41948271e-01
-6.81482077e-01 6.28053129e-01 8.02939773e-01 -1.09293714e-01
-9.17877853e-02 2.46852368e-01 -7.61219084e-01 9.01002407e-01
8.97553980e-01 -1.01555741e+00 -3.97498041e-01 -2.05738902e-01
4.96753693e-01 -5.37843585e-01 5.25687239e-04 -1.25146627e+00
1.75591186e-01 1.13891341e-01 2.90676624e-01 -2.84525096e-01
1.44192830e-01 -7.42381573e-01 -2.18653325e-02 4.76004362e-01
-8.63316178e-01 -4.29678679e-01 -6.16802461e-02 8.55534494e-01
-1.45090476e-01 -8.11954558e-01 1.18919766e+00 -1.78064764e-01
-2.22250670e-01 -1.50680974e-01 -6.02094591e-01 5.12402281e-02
6.85188353e-01 -4.11072336e-02 -3.66915524e-01 -4.46233064e-01
-8.45047295e-01 -5.17082587e-02 2.47635111e-01 1.07706279e-01
2.61462092e-01 -1.49546480e+00 -5.63759804e-01 2.57120550e-01
-4.19948623e-02 -1.45878464e-01 2.93494701e-01 9.34158862e-01
9.17579457e-02 5.89622408e-02 -1.21310716e-02 -4.91300315e-01
-1.13354301e+00 6.23104036e-01 3.66890728e-01 -6.43627524e-01
-2.29859918e-01 1.07433259e+00 2.23603055e-01 -6.06719792e-01
6.31938279e-01 -5.08970357e-02 -5.09959936e-01 2.86466897e-01
6.28557265e-01 3.11062396e-01 1.56650156e-01 -6.24940395e-01
-1.97958067e-01 7.49278665e-01 -1.67408779e-01 1.88787669e-01
1.18761599e+00 -2.79919446e-01 -2.18050569e-01 8.78218234e-01
1.08932042e+00 -7.17321634e-02 -1.11534989e+00 -7.13538229e-01
1.82635397e-01 -2.31368646e-01 -1.34018892e-02 -9.04860258e-01
-1.00888121e+00 9.13015902e-01 9.31340933e-01 4.73423719e-01
1.26727057e+00 -3.28932732e-01 5.68411648e-01 1.93233535e-01
3.00306797e-01 -1.34805071e+00 8.32581222e-02 5.83910644e-01
4.13169593e-01 -1.16058612e+00 -7.21753165e-02 -3.77794892e-01
-4.00769562e-01 9.94226933e-01 3.86890382e-01 -1.68793142e-01
8.92924726e-01 1.88724753e-02 1.89976782e-01 1.13486722e-01
-4.30201203e-01 -9.12322029e-02 3.00170422e-01 7.17049539e-01
3.90526563e-01 -1.19750537e-01 -6.26515329e-01 8.32762480e-01
1.69554055e-01 5.80137856e-02 3.96057278e-01 7.78754592e-01
-6.53538227e-01 -1.24645627e+00 -4.30793196e-01 6.02384984e-01
-5.11363387e-01 3.90174985e-03 -2.05495149e-01 2.35043615e-01
5.93676388e-01 9.92635548e-01 -4.19366121e-01 -4.97693151e-01
2.49571711e-01 5.31774044e-01 2.29640067e-01 -7.55165040e-01
-6.63980961e-01 -2.05380723e-01 -2.70238549e-01 -2.76839972e-01
-5.99797368e-01 -4.90309775e-01 -8.19133520e-01 6.43429309e-02
-8.29945862e-01 4.50596899e-01 6.80792630e-01 1.16086543e+00
2.18016490e-01 6.19052052e-01 4.64927226e-01 -8.27468276e-01
-9.18020666e-01 -1.21150637e+00 -9.03867543e-01 6.33014619e-01
2.20635146e-01 -8.88015270e-01 -7.40531743e-01 -4.51068729e-02] | [9.31523609161377, 3.885422945022583] |
98109f88-ef06-4f5d-849c-4eb7ad342c1d | opi-at-semeval-2023-task-1-image-text | 2304.07127 | null | https://arxiv.org/abs/2304.07127v1 | https://arxiv.org/pdf/2304.07127v1.pdf | OPI at SemEval 2023 Task 1: Image-Text Embeddings and Multimodal Information Retrieval for Visual Word Sense Disambiguation | The goal of visual word sense disambiguation is to find the image that best matches the provided description of the word's meaning. It is a challenging problem, requiring approaches that combine language and image understanding. In this paper, we present our submission to SemEval 2023 visual word sense disambiguation shared task. The proposed system integrates multimodal embeddings, learning to rank methods, and knowledge-based approaches. We build a classifier based on the CLIP model, whose results are enriched with additional information retrieved from Wikipedia and lexical databases. Our solution was ranked third in the multilingual task and won in the Persian track, one of the three language subtasks. | ['Sławomir Dadas'] | 2023-04-14 | null | null | null | null | ['word-sense-disambiguation'] | ['natural-language-processing'] | [-3.01459078e-02 -1.65152803e-01 -4.38971013e-01 -9.56162438e-02
-8.19441855e-01 -7.05637276e-01 8.90294790e-01 6.82329953e-01
-1.02287674e+00 6.89882576e-01 5.79440355e-01 6.58241585e-02
1.16008662e-01 -2.84251750e-01 -2.12966219e-01 -3.74057651e-01
2.70820767e-01 5.59827745e-01 2.24399075e-01 -4.21389908e-01
4.90617275e-01 3.17657560e-01 -1.57428205e+00 7.02601016e-01
7.17492104e-01 9.14282560e-01 6.54368341e-01 3.83023888e-01
-6.95864975e-01 7.30668187e-01 -3.82504463e-01 -4.54703063e-01
-1.49778634e-01 3.22028389e-03 -1.06268609e+00 -2.47821435e-01
9.74458396e-01 5.33548355e-01 4.41405065e-02 1.17305911e+00
7.41339803e-01 1.91850945e-01 6.89094067e-01 -1.36455262e+00
-1.06417727e+00 4.88949537e-01 -4.80884552e-01 3.49693358e-01
5.76203525e-01 -2.18758225e-01 1.27659905e+00 -1.34628332e+00
1.34595668e+00 1.49800181e+00 2.04994351e-01 5.41281700e-01
-1.11179686e+00 -4.87065881e-01 2.28263706e-01 8.05327237e-01
-1.66622651e+00 -2.28209272e-01 4.98067528e-01 -8.35051715e-01
1.34217465e+00 -1.34716168e-01 6.95242643e-01 1.18714750e+00
-1.44303977e-01 1.02506733e+00 1.30510485e+00 -7.73622811e-01
6.19395971e-02 3.21388811e-01 3.85010064e-01 7.37518191e-01
2.26677746e-01 -1.50287926e-01 -8.71122539e-01 3.41916680e-02
1.87628120e-01 -2.95222521e-01 -2.54105121e-01 -5.71715772e-01
-1.33030713e+00 7.27138638e-01 5.54139793e-01 7.44859159e-01
-2.30861142e-01 6.20126091e-02 5.31697989e-01 1.33221582e-01
2.89917737e-01 8.45776439e-01 -2.80960649e-01 2.58021355e-01
-9.26029980e-01 2.76826382e-01 5.00815213e-01 7.59112298e-01
9.29804921e-01 -2.60186970e-01 -5.14921844e-01 1.08808613e+00
3.80980104e-01 6.16745472e-01 6.97275519e-01 -5.78654230e-01
2.22596273e-01 7.95682251e-01 7.47320009e-03 -1.16332984e+00
-4.27934706e-01 5.39096519e-02 -4.41710800e-01 3.36880803e-01
1.33263558e-01 1.75247580e-01 -1.14326859e+00 1.53991020e+00
1.16734765e-01 2.16510962e-03 2.62943655e-01 1.01228762e+00
1.42541027e+00 4.28469658e-01 7.86032975e-01 1.43177342e-02
1.90291297e+00 -7.83231378e-01 -7.94773102e-01 -5.54336667e-01
3.11906844e-01 -1.15204978e+00 9.56287384e-01 1.62743509e-01
-5.18374562e-01 -5.79635501e-01 -1.28661335e+00 -3.54278862e-01
-1.00094628e+00 2.14939922e-01 2.09417075e-01 2.64655083e-01
-1.16614974e+00 -4.95479032e-02 -1.82761177e-01 -1.01253140e+00
1.81125849e-01 -1.61768720e-01 -8.46759021e-01 -2.17102632e-01
-1.34083128e+00 1.45886278e+00 8.93188477e-01 -4.29603338e-01
-6.07682168e-01 -4.54778105e-01 -1.17746139e+00 -3.46981883e-01
4.56050634e-01 -7.65131474e-01 7.64792264e-01 -9.92234647e-01
-6.66376114e-01 1.73780620e+00 -2.92836457e-01 -5.45100391e-01
1.04829475e-01 -2.78784215e-01 -6.01799548e-01 2.39561960e-01
5.40936768e-01 1.21384227e+00 7.05629706e-01 -1.51146090e+00
-8.31504285e-01 -6.04528785e-01 -1.20210601e-02 3.95260692e-01
-1.23808712e-01 1.05736196e-01 -8.79219830e-01 -6.01989508e-01
-1.68160632e-01 -7.65913129e-01 -7.11956099e-02 -1.36520267e-01
-2.93061495e-01 -6.39468908e-01 7.16146588e-01 -7.37986863e-01
1.04719007e+00 -2.16499591e+00 2.58956581e-01 1.69166446e-01
3.90923083e-01 1.40654504e-01 -3.76473606e-01 4.59132075e-01
-1.34811863e-01 2.19809830e-01 1.80167228e-01 -3.34338546e-01
1.30776405e-01 2.13292882e-01 -1.45186946e-01 1.15819097e-01
1.51946187e-01 8.46608996e-01 -1.07854903e+00 -1.01687825e+00
2.74476141e-01 4.03771818e-01 5.80622591e-02 -3.95305753e-02
-2.67529011e-01 1.06569983e-01 -3.03080350e-01 6.47543550e-01
4.08752084e-01 -1.14061557e-01 4.13531899e-01 -7.96518385e-01
-2.98855424e-01 -1.64444476e-01 -1.20599663e+00 2.01616311e+00
-5.24550915e-01 9.84546661e-01 -2.52999336e-01 -8.27935696e-01
8.52284431e-01 1.87289104e-01 2.23440751e-01 -1.13351095e+00
-3.86689380e-02 4.83118705e-02 -5.39649189e-01 -8.15306783e-01
8.48657489e-01 -1.52023315e-01 -3.10381532e-01 -1.12514151e-02
5.43874979e-01 -5.89185655e-02 5.25862038e-01 4.80741709e-01
6.16134167e-01 1.66118592e-01 9.45771277e-01 -3.55873317e-01
5.47810972e-01 3.72225374e-01 4.01094586e-01 6.18655324e-01
-3.28631997e-01 4.24426764e-01 3.03912848e-01 -5.17477512e-01
-9.80097234e-01 -1.13074994e+00 -3.82741354e-02 1.20298600e+00
4.19169039e-01 -7.66552627e-01 -4.01261151e-01 -5.60568452e-01
4.71104160e-02 7.07364261e-01 -7.34940708e-01 2.00319916e-01
-2.08721548e-01 -2.70339519e-01 4.22911465e-01 3.47441018e-01
3.64944845e-01 -1.22177279e+00 -5.65926015e-01 -4.97452728e-02
-2.98014939e-01 -1.31862330e+00 -3.82273287e-01 -8.49867836e-02
-9.57167521e-02 -1.43429673e+00 -7.91194141e-01 -1.18169188e+00
1.85630336e-01 2.01748550e-01 1.43998182e+00 -1.07872756e-02
-8.68468344e-01 8.27691436e-01 -5.07680237e-01 -4.88744169e-01
-6.77022263e-02 5.31424470e-02 -9.28566754e-02 -1.13229111e-01
7.25210547e-01 1.87971517e-01 -3.87925953e-01 -1.52460828e-01
-8.06943655e-01 -1.25637114e-01 5.29229701e-01 8.32283080e-01
8.83633792e-01 -5.13992667e-01 2.40571395e-01 -5.73603690e-01
6.60629153e-01 -2.48105153e-01 -3.56007367e-01 7.63348341e-01
-7.46061623e-01 4.17350054e-01 -3.33944738e-01 -1.84832498e-01
-1.01815212e+00 5.56335688e-01 1.27631605e-01 -3.88363898e-01
-2.67814249e-01 5.41022897e-01 -3.06583017e-01 -1.80895887e-02
6.85491800e-01 2.65282542e-02 -3.70561332e-01 -4.55655962e-01
1.05496991e+00 4.74904209e-01 6.92334414e-01 -5.26020348e-01
2.13212997e-01 3.03164244e-01 -2.59599864e-01 -1.17628050e+00
-1.04872620e+00 -1.00033832e+00 -8.07959735e-01 -4.88755941e-01
1.49385393e+00 -1.10811424e+00 -3.49826157e-01 -6.53482899e-02
-1.46073258e+00 3.08875859e-01 -1.08535118e-01 4.38007087e-01
-3.73781353e-01 3.68560493e-01 2.35852405e-01 -6.21361017e-01
-4.45710719e-01 -9.55852866e-01 1.08283532e+00 2.60639966e-01
-4.11545783e-01 -8.98480535e-01 4.96773690e-01 4.59382147e-01
1.15932964e-01 2.52162904e-01 8.95972133e-01 -8.33708107e-01
-2.71301329e-01 6.52759671e-02 -6.10473871e-01 1.18094526e-01
-1.80012241e-01 -1.14426136e-01 -9.58858848e-01 -9.67060104e-02
-6.91984594e-01 -5.67306578e-01 1.37650776e+00 2.28587866e-01
5.04234254e-01 1.37978166e-01 -6.41946018e-01 1.35308355e-01
1.81643832e+00 1.24860704e-01 4.62259799e-01 5.95778823e-01
7.99938917e-01 6.82407141e-01 6.35464549e-01 2.07359657e-01
5.47838151e-01 1.00728405e+00 2.67555207e-01 -5.84567226e-02
-5.34467101e-01 -2.36776397e-01 1.22387446e-01 4.14792687e-01
1.04647599e-01 -1.67967841e-01 -1.41273367e+00 9.57524836e-01
-1.90860832e+00 -9.83161509e-01 -8.68060142e-02 1.94034290e+00
6.71187222e-01 -2.03338683e-01 4.65094857e-02 -4.23167735e-01
7.77456403e-01 2.66843915e-01 -7.49841109e-02 -3.74702990e-01
-5.10099769e-01 2.47182637e-01 3.76926810e-01 6.10517919e-01
-1.47382927e+00 1.49973452e+00 6.48644972e+00 9.50713575e-01
-8.29421401e-01 4.34100121e-01 1.27146924e-02 1.84444830e-01
-2.87048399e-01 -9.99406055e-02 -7.39631355e-01 -3.30527872e-02
3.39732081e-01 -3.37706655e-01 5.06751351e-02 6.52521789e-01
-2.50497103e-01 -3.42766166e-01 -9.15730774e-01 1.42999709e+00
8.78549337e-01 -1.35389316e+00 4.38020915e-01 -4.26717013e-01
4.46946621e-01 2.02033460e-01 -7.72087350e-02 2.57006079e-01
3.66269082e-01 -1.16249537e+00 7.26812959e-01 8.09502840e-01
8.31914961e-01 -7.31419027e-01 6.92528546e-01 -2.30223686e-01
-1.50000048e+00 2.01165900e-01 -2.80423909e-01 3.83112162e-01
6.06022887e-02 2.99506396e-01 -9.37772393e-01 5.18673718e-01
8.73573124e-01 9.81031239e-01 -9.36499119e-01 1.09534526e+00
-4.96454686e-01 -5.55402003e-02 2.09334329e-01 -1.82545796e-01
1.31347224e-01 2.33236164e-01 7.30581462e-01 1.65233612e+00
2.13206351e-01 -9.39231664e-02 7.25110412e-01 6.06933951e-01
-7.69571215e-02 7.29240537e-01 -8.76128674e-01 -1.24366418e-01
3.46459985e-01 1.43286753e+00 -6.70835197e-01 -5.31734049e-01
-4.06863153e-01 1.08661556e+00 5.75028837e-01 2.50174552e-01
-1.98075548e-01 -2.82918602e-01 6.81667030e-01 -3.54255229e-01
1.17980927e-01 -1.83134273e-01 -9.00816098e-02 -1.10643125e+00
-2.48958439e-01 -6.40011430e-01 8.23547304e-01 -1.25357294e+00
-1.48971629e+00 8.38828981e-01 8.80001485e-02 -1.10501671e+00
-1.44275323e-01 -9.55210865e-01 -4.88398932e-02 8.07240903e-01
-1.51953852e+00 -1.44360960e+00 -2.75820404e-01 6.98021710e-01
5.06215036e-01 -4.57098305e-01 1.07159877e+00 3.94753397e-01
-1.78036317e-01 3.61725837e-01 -2.88858503e-01 2.01780081e-01
1.21871781e+00 -1.46487582e+00 5.46840776e-04 5.93173504e-01
7.46085346e-01 3.03173721e-01 7.95455039e-01 -7.25106299e-01
-9.71267581e-01 -8.07014167e-01 1.41445100e+00 -4.99323398e-01
9.15242732e-01 -2.71692500e-02 -6.45308137e-01 3.50159049e-01
8.37432802e-01 -2.25408420e-01 8.31649065e-01 1.37907147e-01
-9.00730312e-01 1.87804848e-01 -8.15289378e-01 4.29307967e-01
7.17071474e-01 -7.64337778e-01 -1.20858347e+00 2.16576040e-01
4.75381672e-01 -1.66561723e-01 -6.36018574e-01 -1.72454081e-02
5.12151480e-01 -2.30351657e-01 1.24149942e+00 -7.87632406e-01
3.39959562e-01 -5.33808291e-01 -6.29558563e-01 -1.47886050e+00
-1.71929181e-01 6.58341199e-02 5.34244835e-01 1.22620022e+00
5.72548509e-01 6.73577189e-02 2.94391304e-01 -9.06102452e-03
2.07182601e-01 7.28361961e-03 -1.13377678e+00 -6.30593777e-01
-1.32842986e-02 -4.64917392e-01 1.00913674e-01 1.26878822e+00
1.73860759e-01 8.71188045e-01 -3.69463593e-01 7.05916062e-02
5.66826880e-01 1.83594719e-01 3.27294558e-01 -1.34740412e+00
2.35557213e-01 -5.75296521e-01 -8.44526112e-01 -2.97400683e-01
5.57389677e-01 -1.34585190e+00 2.03969236e-03 -1.99724841e+00
6.11799896e-01 2.06342250e-01 -5.90519249e-01 8.33322048e-01
-2.83808261e-01 4.06726003e-01 7.06890225e-01 1.49298785e-03
-1.13165402e+00 1.81693301e-01 8.68539631e-01 -7.78491914e-01
1.39707670e-01 -9.64228332e-01 -5.74591339e-01 7.49671876e-01
5.90915501e-01 -3.36390167e-01 -1.00383602e-01 -4.61034238e-01
4.94537294e-01 -2.70925522e-01 5.20765364e-01 -8.81457567e-01
2.43413284e-01 -2.43206948e-01 3.65518332e-01 -6.53134704e-01
3.63790214e-01 -7.18868732e-01 -2.42693529e-01 2.77994692e-01
-4.58704263e-01 4.48400557e-01 3.53352755e-01 4.87565011e-01
-4.86619592e-01 -1.13085903e-01 7.01948762e-01 -2.01450333e-01
-1.88967371e+00 7.09131034e-03 -3.60908359e-01 4.20625716e-01
1.07861996e+00 1.27958477e-01 -3.80645275e-01 -2.39926577e-02
-1.24010217e+00 4.71368432e-01 3.21462512e-01 1.04898894e+00
7.92095423e-01 -1.68938565e+00 -8.12828004e-01 -4.04459774e-01
1.04501069e+00 -9.26471114e-01 3.29371333e-01 3.94426107e-01
-3.47895205e-01 4.03803527e-01 -5.61456442e-01 -4.99387592e-01
-1.69157660e+00 8.59566152e-01 2.02091530e-01 -1.58088684e-01
-3.59914929e-01 6.74945951e-01 -2.62129586e-02 -2.88217515e-01
1.50780469e-01 2.32779264e-01 -1.07674837e+00 8.63988221e-01
8.90978217e-01 1.25209689e-01 -7.82289505e-02 -1.17213869e+00
-8.20663929e-01 8.48243594e-01 1.84768409e-01 -3.93839449e-01
8.94736707e-01 -2.63625860e-01 -2.95979172e-01 6.22683048e-01
1.20697093e+00 -1.38164490e-01 -3.23443443e-01 -5.69321632e-01
6.01408899e-01 -2.34430164e-01 1.06849559e-01 -1.21506155e+00
-8.66046667e-01 9.66862381e-01 1.38175154e+00 -8.89032334e-02
8.29516292e-01 4.96824741e-01 3.36505592e-01 5.44849038e-01
2.59425908e-01 -1.37220716e+00 6.17898144e-02 7.06997752e-01
1.15574658e+00 -1.57254350e+00 3.73925455e-02 -2.04176009e-01
-1.02070427e+00 1.08274305e+00 6.32594168e-01 1.40344098e-01
5.00598490e-01 -7.52217993e-02 4.97294366e-01 -4.32542026e-01
-6.74967706e-01 -1.04507172e+00 9.96643960e-01 8.85216296e-01
5.24079025e-01 2.89909273e-01 -7.45448887e-01 6.21024072e-01
1.71750173e-01 -1.83006838e-01 2.71284487e-03 6.39753044e-01
-7.10834682e-01 -1.27893889e+00 -3.25242460e-01 -3.32913585e-02
-2.31034651e-01 -4.60163623e-01 -8.23778987e-01 8.63615930e-01
5.15609443e-01 8.51157486e-01 4.57276478e-02 -2.44048044e-01
4.08802688e-01 3.78229976e-01 4.40752864e-01 -6.43833756e-01
-2.90266901e-01 -3.58085148e-02 1.94068089e-01 -7.54211903e-01
-8.32592010e-01 -5.56923866e-01 -1.28208649e+00 3.15531284e-01
3.49089146e-01 -9.42186415e-02 6.16015077e-01 1.07628942e+00
2.32307494e-01 3.57730389e-01 -1.80941790e-01 -6.90590978e-01
2.86433697e-01 -6.77927554e-01 -4.27970707e-01 6.94664538e-01
7.30283633e-02 -8.94718051e-01 2.28021264e-01 2.92862684e-01] | [10.79671859741211, 1.5213834047317505] |
e8fa6e5c-c508-4342-91cc-55afc446d8ce | generalizing-to-unseen-domains-with | 2207.04913 | null | https://arxiv.org/abs/2207.04913v1 | https://arxiv.org/pdf/2207.04913v1.pdf | Generalizing to Unseen Domains with Wasserstein Distributional Robustness under Limited Source Knowledge | Domain generalization aims at learning a universal model that performs well on unseen target domains, incorporating knowledge from multiple source domains. In this research, we consider the scenario where different domain shifts occur among conditional distributions of different classes across domains. When labeled samples in the source domains are limited, existing approaches are not sufficiently robust. To address this problem, we propose a novel domain generalization framework called Wasserstein Distributionally Robust Domain Generalization (WDRDG), inspired by the concept of distributionally robust optimization. We encourage robustness over conditional distributions within class-specific Wasserstein uncertainty sets and optimize the worst-case performance of a classifier over these uncertainty sets. We further develop a test-time adaptation module leveraging optimal transport to quantify the relationship between the unseen target domain and source domains to make adaptive inference for target data. Experiments on the Rotated MNIST, PACS and the VLCS datasets demonstrate that our method could effectively balance the robustness and discriminability in challenging generalization scenarios. | ['Yang Li', 'Shao-Lun Huang', 'Yao Xie', 'Liyan Xie', 'Jingge Wang'] | 2022-07-11 | null | null | null | null | ['rotated-mnist'] | ['computer-vision'] | [ 2.31027469e-01 -1.69599354e-01 -1.87684044e-01 -8.26990247e-01
-1.08849061e+00 -8.78516078e-01 5.15679717e-01 3.19557078e-02
-3.45686793e-01 1.09828520e+00 -9.66348127e-02 -4.66040000e-02
-5.90946376e-01 -8.11228812e-01 -5.82985640e-01 -8.41822028e-01
1.43302873e-01 6.24399781e-01 3.40117931e-01 2.30184887e-02
8.42861682e-02 6.22164249e-01 -1.20596254e+00 4.55582701e-02
1.33856392e+00 1.03952360e+00 1.48999751e-01 2.16642827e-01
1.30981021e-02 -2.64345314e-02 -8.07197988e-01 -7.53476769e-02
4.46605712e-01 -9.65302363e-02 -5.90019882e-01 6.29037321e-02
3.54064673e-01 -1.90179884e-01 -1.54481411e-01 1.26878262e+00
5.06916404e-01 5.91591775e-01 1.17003548e+00 -1.46260893e+00
-1.07796299e+00 4.30685729e-01 -4.16828066e-01 5.02966762e-01
7.10816756e-02 2.24960931e-02 7.03577816e-01 -6.55963778e-01
4.85218078e-01 1.43342054e+00 3.05740595e-01 7.17391968e-01
-1.40560758e+00 -8.83681774e-01 5.37062168e-01 1.20575324e-01
-1.37285340e+00 4.67197672e-02 7.21637845e-01 -5.22741437e-01
3.52918565e-01 -1.25641063e-01 -3.03697497e-01 1.70298243e+00
1.18244916e-01 6.72017753e-01 1.20430827e+00 -1.96041107e-01
6.70366585e-01 4.34003860e-01 1.82667524e-01 -1.02357782e-01
4.01068300e-01 2.30659544e-01 -4.07492220e-01 -2.38178238e-01
5.06512344e-01 -5.84807880e-02 -2.65952349e-01 -7.27216840e-01
-1.13633919e+00 7.71901727e-01 4.00842726e-01 1.82253480e-01
-2.81099193e-02 -4.69497353e-01 2.33895019e-01 3.82633179e-01
7.57599354e-01 3.93423378e-01 -6.81919873e-01 2.30374753e-01
-6.22542381e-01 3.91525358e-01 5.74478924e-01 1.30249250e+00
6.69569373e-01 -5.38494103e-02 -4.30638969e-01 1.02718771e+00
3.25667083e-01 8.40462387e-01 6.67898357e-01 -6.60520434e-01
6.12317264e-01 4.71466064e-01 2.52220005e-01 -5.52307665e-01
-3.22616309e-01 -5.12505472e-01 -5.22991300e-01 8.83339718e-02
6.91946030e-01 -2.95677960e-01 -1.08521056e+00 1.98572528e+00
6.40828788e-01 2.83311665e-01 3.78578782e-01 7.60709465e-01
4.29134995e-01 3.53042334e-01 3.23132843e-01 1.28207117e-01
8.63135755e-01 -3.52854103e-01 -3.46424699e-01 -3.52114797e-01
4.75404650e-01 -3.91747594e-01 1.21150005e+00 2.95532018e-01
-3.55492115e-01 -5.73400438e-01 -1.20611179e+00 2.59039760e-01
-5.88467956e-01 -2.17199266e-01 -8.62686615e-03 7.13665843e-01
-4.73074526e-01 6.78552210e-01 -6.28335297e-01 -4.52526957e-01
5.85879564e-01 8.11884552e-02 -2.52523482e-01 -3.74033451e-01
-1.25996292e+00 7.80990839e-01 9.89154398e-01 -3.66558969e-01
-1.23634577e+00 -9.13592756e-01 -8.63886595e-01 -1.57535553e-01
3.84459823e-01 -3.52133572e-01 1.10233569e+00 -7.36393869e-01
-1.30938983e+00 7.26992428e-01 1.22649103e-01 -5.44754326e-01
6.07937157e-01 -1.50106072e-01 -8.06882560e-01 -1.11858159e-01
2.45825902e-01 4.58192647e-01 1.04990304e+00 -1.27406728e+00
-7.04807222e-01 -6.24539196e-01 -2.56926924e-01 1.30794525e-01
-3.86717647e-01 -4.52221662e-01 1.88905895e-01 -8.47026408e-01
1.40251413e-01 -8.79319906e-01 -5.31547032e-02 -2.30418965e-01
-3.32835406e-01 -4.93405133e-01 1.02935874e+00 -3.16050708e-01
8.97210836e-01 -2.44523597e+00 9.76066366e-02 4.02030081e-01
2.07496178e-03 1.46825805e-01 -2.86040813e-01 -9.09968764e-02
4.47581476e-03 3.51887872e-03 -4.21229362e-01 2.06343353e-01
3.37403059e-01 4.85063076e-01 -7.00257063e-01 4.96326566e-01
4.70287830e-01 3.80583823e-01 -1.02308464e+00 -2.26518497e-01
7.85471797e-02 2.68207908e-01 -3.49419355e-01 2.50520736e-01
-3.75806034e-01 7.67972648e-01 -8.36895943e-01 6.52000487e-01
1.13507342e+00 -1.33247763e-01 1.49298474e-01 1.47033453e-01
5.07352412e-01 -1.02774769e-01 -1.38025856e+00 1.42772245e+00
-2.89772570e-01 2.56999850e-01 -2.12902933e-01 -1.18802059e+00
1.30234432e+00 -1.72310188e-01 2.26999179e-01 -4.38017786e-01
-7.72447418e-03 4.42938745e-01 -6.61800504e-02 -2.52090067e-01
2.78486431e-01 -4.98782098e-01 -3.44492525e-01 1.47436276e-01
3.48915726e-01 -2.04699874e-01 -6.94003925e-02 -6.64298907e-02
7.77174234e-01 2.87638634e-01 2.85849065e-01 -6.70583904e-01
3.80292445e-01 -1.71093345e-01 7.65596390e-01 8.29635918e-01
-6.63105667e-01 6.41938686e-01 3.61799031e-01 -1.74297497e-03
-8.62477243e-01 -1.85961127e+00 -8.86160016e-01 1.40495825e+00
4.53177392e-01 2.74561971e-01 -6.47871912e-01 -1.12762702e+00
4.42426145e-01 1.13663983e+00 -5.76661706e-01 -6.03100896e-01
-1.83409616e-01 -7.83220530e-01 4.59350199e-01 7.90287793e-01
4.89457369e-01 -5.73602736e-01 -2.11213008e-01 8.49168748e-02
1.06152426e-03 -1.17579162e+00 -3.14093918e-01 3.93252879e-01
-9.10245538e-01 -1.05007887e+00 -9.24523294e-01 -5.43534517e-01
4.24927533e-01 -5.46379872e-02 9.22230661e-01 -9.16076481e-01
-1.95613150e-02 5.20997882e-01 -3.90834808e-01 -5.47202647e-01
-3.70421827e-01 1.77607592e-02 6.19895577e-01 1.50436470e-02
8.17856908e-01 -4.09483761e-01 -2.77320147e-01 6.83293164e-01
-1.16283917e+00 -8.84298086e-01 2.86852688e-01 8.07650983e-01
7.12579548e-01 2.45203838e-01 1.11772680e+00 -7.64160454e-01
8.14679325e-01 -9.70218003e-01 -6.62918508e-01 5.38633943e-01
-6.22605205e-01 3.75552475e-01 6.81638598e-01 -8.92982483e-01
-1.45701301e+00 -2.92175412e-01 4.67826217e-01 -3.35843027e-01
-4.96416539e-01 1.98101088e-01 -6.35280192e-01 2.11447999e-01
1.15493250e+00 5.68114500e-03 -3.32204789e-01 -4.37076122e-01
4.05880183e-01 8.51262212e-01 5.32666564e-01 -1.23403454e+00
9.42267895e-01 5.32553434e-01 -1.86764181e-01 -7.80382276e-01
-1.01865470e+00 -4.68084753e-01 -9.41246450e-01 1.45870432e-01
7.16462731e-01 -8.70186031e-01 1.81664638e-02 2.10653439e-01
-6.92780793e-01 -3.64647388e-01 -5.36973178e-01 6.81087554e-01
-5.87363124e-01 4.67973441e-01 1.43053308e-01 -6.69960082e-01
1.59460202e-01 -1.15719426e+00 1.10934591e+00 3.86742711e-01
-1.65781096e-01 -1.41704249e+00 3.86914797e-02 -3.23114358e-02
2.31495649e-01 4.34162349e-01 9.98806059e-01 -1.35332561e+00
-2.62062937e-01 -7.82077014e-02 -3.54547709e-01 7.97613978e-01
1.92359880e-01 -3.71310055e-01 -1.09827375e+00 -5.04462779e-01
-4.75338250e-02 -3.84944707e-01 8.16421688e-01 3.55263293e-01
1.40075994e+00 3.78460661e-02 -5.41329861e-01 6.76973403e-01
1.23958516e+00 2.26458445e-01 3.13651145e-01 4.73980308e-01
3.10995817e-01 6.69022083e-01 9.11885381e-01 5.66237390e-01
9.60511491e-02 3.50111574e-01 2.02033758e-01 4.92598951e-01
1.34638146e-01 -3.17360878e-01 3.75473410e-01 -1.57826804e-02
4.72624540e-01 -4.07202601e-01 -1.06162417e+00 7.48365581e-01
-1.58311200e+00 -6.60911560e-01 3.14647555e-01 2.36549115e+00
7.86001086e-01 1.91619813e-01 2.86478363e-02 -2.49863610e-01
9.49372828e-01 -1.22064531e-01 -1.21663249e+00 4.76126932e-03
-2.41016269e-01 1.81163520e-01 5.79478443e-01 4.06498164e-01
-1.24537051e+00 7.75730252e-01 6.34172153e+00 1.00885630e+00
-8.15897942e-01 -3.52778099e-02 4.14050162e-01 -5.07013202e-02
-4.20986384e-01 -2.40810648e-01 -1.02191448e+00 5.57800591e-01
8.85111094e-01 -4.91779685e-01 1.33565739e-01 1.19878793e+00
-1.95387647e-01 1.94679067e-01 -1.43299484e+00 7.64832318e-01
-1.49810597e-01 -5.93203127e-01 8.35091621e-02 3.84696648e-02
8.11351001e-01 1.57179281e-01 5.62424362e-01 7.12458789e-01
7.90273905e-01 -9.78204310e-01 3.98984134e-01 4.05434132e-01
7.81467557e-01 -6.90696180e-01 4.84152138e-01 4.74397182e-01
-7.18302488e-01 -3.19497585e-01 -8.17718804e-01 4.63845789e-01
-1.52626693e-01 7.40777075e-01 -1.04300308e+00 5.44822693e-01
8.73267829e-01 7.60579467e-01 -4.70539004e-01 8.90882015e-01
-1.81909561e-01 4.15700823e-01 -3.70626390e-01 2.48061150e-01
9.72908810e-02 -1.79806873e-01 7.79232979e-01 1.10797763e+00
5.13648331e-01 -6.08971454e-02 2.66167164e-01 1.02948487e+00
-1.14031151e-01 -2.02657819e-01 -6.86622560e-01 1.60524637e-01
8.12025309e-01 8.91600788e-01 -4.78103787e-01 -2.23798767e-01
-2.52040643e-02 8.36631715e-01 3.15345168e-01 8.44384909e-01
-9.70100641e-01 -4.13586646e-01 9.26669121e-01 -1.85617983e-01
1.99960455e-01 -1.61555529e-01 -3.15012008e-01 -1.35728276e+00
1.72035515e-01 -7.23673344e-01 8.87301505e-01 -4.49223071e-01
-2.14380002e+00 4.27209198e-01 5.48602641e-01 -1.46796179e+00
-1.62624102e-02 -1.05421555e+00 -3.48117471e-01 1.04056489e+00
-1.81288803e+00 -8.60185444e-01 -1.08776033e-01 8.78181994e-01
4.07280684e-01 -4.17151392e-01 6.20546818e-01 7.04520522e-03
-4.52873141e-01 8.57759356e-01 8.25437784e-01 -5.42544387e-02
1.24921286e+00 -1.45743525e+00 -6.91548362e-02 8.02303016e-01
-2.07061931e-01 6.15515649e-01 6.94073498e-01 -6.31742835e-01
-6.50799811e-01 -1.43294895e+00 1.46092519e-01 -7.09282398e-01
8.03006113e-01 -3.31908077e-01 -1.21231973e+00 8.45130742e-01
-2.09008843e-01 1.73668474e-01 9.04673636e-01 1.91844031e-01
-8.22710633e-01 -2.40328759e-01 -1.73204374e+00 3.56998563e-01
9.66429651e-01 -4.69923317e-01 -9.14062321e-01 5.45066833e-01
6.79032147e-01 -3.72993201e-01 -1.15593600e+00 5.55310369e-01
1.39972806e-01 -5.48255503e-01 1.07663357e+00 -1.04190254e+00
2.21468005e-02 -2.69728541e-01 -6.64743483e-01 -1.89238524e+00
-2.35376418e-01 -1.47841558e-01 5.38660176e-02 1.34497571e+00
3.77786964e-01 -9.18022871e-01 5.18865943e-01 7.18551159e-01
7.65940920e-02 -1.21178344e-01 -1.24102569e+00 -1.39394319e+00
8.58556390e-01 -5.73497474e-01 9.36245620e-01 1.02241611e+00
-7.33049810e-02 1.26897646e-02 1.57048196e-01 6.50102019e-01
9.57823038e-01 6.96982518e-02 5.10431647e-01 -1.55062079e+00
-1.04410090e-01 -2.69206971e-01 -4.27794039e-01 -9.41576481e-01
5.84746778e-01 -1.03022277e+00 7.12420046e-02 -1.11771429e+00
-2.24362575e-02 -5.88602066e-01 -6.26477957e-01 2.05895379e-01
-2.49013916e-01 -3.98658037e-01 -6.08557835e-02 1.74279362e-01
-6.36308789e-01 8.08708966e-01 1.21688020e+00 -2.92631209e-01
-1.73449650e-01 1.96065120e-02 -8.78462791e-01 6.43288672e-01
8.68127704e-01 -5.19477427e-01 -8.02839160e-01 -4.13340390e-01
-3.96446764e-01 -4.99284178e-01 2.95717090e-01 -1.02271950e+00
-8.34906697e-02 -5.98745823e-01 6.68632448e-01 -3.75870526e-01
3.74170169e-02 -1.01799464e+00 -4.42285925e-01 -5.03001846e-02
-5.07271707e-01 -4.70404327e-01 3.54456633e-01 1.13918614e+00
-1.29539758e-01 -8.55587721e-02 1.34720421e+00 2.71729887e-01
-1.01401663e+00 4.11784977e-01 4.44350801e-02 6.68647587e-01
1.50064623e+00 -1.77809507e-01 -5.10651708e-01 3.89860682e-02
-9.01923597e-01 4.89764184e-01 2.97827601e-01 6.44396663e-01
6.05196834e-01 -1.48841751e+00 -8.36857021e-01 2.90743649e-01
6.64188743e-01 2.53026128e-01 3.02464545e-01 2.56696999e-01
4.66158725e-02 2.57027805e-01 -1.82293251e-01 -9.61811423e-01
-6.18113816e-01 7.78839588e-01 5.42597055e-01 -6.10440746e-02
-2.78994441e-01 9.18622375e-01 5.40252626e-01 -9.84926820e-01
1.34538904e-01 -5.44207752e-01 1.47693455e-02 -3.25736776e-02
5.04439890e-01 5.24796724e-01 -1.92647595e-02 -3.73758018e-01
-4.20734137e-01 3.21795434e-01 -1.48785681e-01 1.74294431e-02
1.08807266e+00 -3.07133585e-01 4.53427494e-01 4.11200345e-01
1.25235987e+00 -3.35500687e-01 -1.79200017e+00 -6.19404197e-01
4.82118517e-01 -6.17918670e-01 -2.27161109e-01 -9.93583143e-01
-5.72292626e-01 8.45037460e-01 8.89004409e-01 2.20528156e-01
1.07352138e+00 1.72101483e-01 4.82403338e-01 2.80228615e-01
2.43791088e-01 -1.48282588e+00 1.47841200e-01 5.90956390e-01
7.75896192e-01 -1.47459400e+00 -2.46731296e-01 -1.22687317e-01
-9.16246295e-01 1.09703910e+00 9.22576368e-01 -1.82057828e-01
7.76154637e-01 2.96801999e-02 4.31610793e-02 1.48689628e-01
-3.16202432e-01 -1.06287427e-01 5.92616498e-01 1.33573461e+00
6.92608282e-02 3.23105067e-01 2.06737816e-01 8.05325508e-01
-7.72974193e-02 -1.81944460e-01 2.79063642e-01 6.68849587e-01
-6.06344640e-01 -1.04013157e+00 -6.19344771e-01 2.81599551e-01
-2.58433908e-01 2.15061307e-01 -2.11035073e-01 8.61646175e-01
4.79748510e-02 1.00080860e+00 -4.86611500e-02 -2.70917743e-01
5.45413077e-01 3.35653871e-01 3.46675217e-01 -7.79903471e-01
2.72028357e-01 -2.38823846e-01 -4.01433110e-01 -3.25224310e-01
-2.65169710e-01 -6.64739370e-01 -1.21316266e+00 2.93680485e-02
-1.21754855e-01 1.08578773e-02 3.20216358e-01 9.70252454e-01
3.43960404e-01 3.41716886e-01 5.86137474e-01 -4.20856029e-01
-1.48478520e+00 -9.23226595e-01 -1.11338842e+00 6.62382066e-01
5.19704700e-01 -1.12890422e+00 -5.77884912e-01 -1.42521292e-01] | [10.339449882507324, 3.1254024505615234] |
5eb15c14-8310-4f47-a5ee-d0de42d0b796 | pnen-pyramid-non-local-enhanced-networks | 2008.09742 | null | https://arxiv.org/abs/2008.09742v1 | https://arxiv.org/pdf/2008.09742v1.pdf | PNEN: Pyramid Non-Local Enhanced Networks | Existing neural networks proposed for low-level image processing tasks are usually implemented by stacking convolution layers with limited kernel size. Every convolution layer merely involves in context information from a small local neighborhood. More contextual features can be explored as more convolution layers are adopted. However it is difficult and costly to take full advantage of long-range dependencies. We propose a novel non-local module, Pyramid Non-local Block, to build up connection between every pixel and all remain pixels. The proposed module is capable of efficiently exploiting pairwise dependencies between different scales of low-level structures. The target is fulfilled through first learning a query feature map with full resolution and a pyramid of reference feature maps with downscaled resolutions. Then correlations with multi-scale reference features are exploited for enhancing pixel-level feature representation. The calculation procedure is economical considering memory consumption and computational cost. Based on the proposed module, we devise a Pyramid Non-local Enhanced Networks for edge-preserving image smoothing which achieves state-of-the-art performance in imitating three classical image smoothing algorithms. Additionally, the pyramid non-local block can be directly incorporated into convolution neural networks for other image restoration tasks. We integrate it into two existing methods for image denoising and single image super-resolution, achieving consistently improved performance. | ['Kai-Kuang Ma', 'Feida Zhu', 'Chaowei Fang'] | 2020-08-22 | null | null | null | null | ['image-smoothing'] | ['computer-vision'] | [ 4.78183866e-01 -1.91071421e-01 2.35510275e-01 -5.74921787e-01
-7.14470983e-01 1.00989491e-01 4.61073995e-01 4.85243201e-02
-8.70554745e-01 6.57765865e-01 1.53524280e-01 1.66440353e-01
-2.45414793e-01 -1.11190140e+00 -8.08803082e-01 -9.28384423e-01
-1.06490679e-01 -6.09611869e-01 6.69461846e-01 -3.29972863e-01
2.72401810e-01 7.72533178e-01 -1.76905107e+00 5.98645866e-01
1.09492970e+00 1.00097537e+00 6.67162418e-01 5.86686432e-01
-1.96024343e-01 5.85788190e-01 -1.60995007e-01 -1.86454520e-01
2.58750618e-01 -2.06169546e-01 -4.91351962e-01 1.22451141e-01
7.29993403e-01 -2.79846311e-01 -2.39144921e-01 1.45994663e+00
5.74810863e-01 3.04027677e-01 3.05516422e-01 -4.46544290e-01
-8.94116044e-01 4.51038778e-01 -7.61987567e-01 4.42548931e-01
-3.29793453e-01 -1.73307694e-02 6.59285069e-01 -1.06710267e+00
4.12015468e-01 1.11170936e+00 1.02055168e+00 2.96141982e-01
-1.34709144e+00 -2.15242356e-01 1.92405194e-01 4.43028092e-01
-1.24126840e+00 -3.03519845e-01 8.50988030e-01 6.65810704e-02
9.12071228e-01 1.30471021e-01 3.56538951e-01 5.20645082e-01
3.96917522e-01 3.36646676e-01 1.39969993e+00 -3.80978405e-01
9.32873189e-02 2.45785415e-02 1.98914230e-01 7.42848635e-01
1.13740228e-01 1.26989931e-01 -5.68808973e-01 3.26340467e-01
1.22265482e+00 1.62277758e-01 -3.37264299e-01 -5.33985980e-02
-1.10981822e+00 5.02060473e-01 1.02058077e+00 7.57866025e-01
-8.02544475e-01 5.81189506e-02 3.15046608e-01 2.76011467e-01
6.03833258e-01 -3.23233046e-02 -5.70709705e-01 5.49153328e-01
-1.23893166e+00 -8.96923095e-02 2.55083412e-01 4.21370834e-01
1.15714204e+00 1.96334571e-01 -2.87899315e-01 9.27287519e-01
-1.21846728e-01 -1.17427027e-02 4.78351235e-01 -8.89184892e-01
1.51252210e-01 6.41425908e-01 5.43744713e-02 -1.09897494e+00
-5.80123723e-01 -4.54384625e-01 -1.46901655e+00 8.71336937e-01
2.86439121e-01 1.51242226e-01 -1.01220477e+00 1.55364001e+00
3.29090953e-01 4.68121052e-01 1.87633574e-01 9.33480799e-01
1.00817013e+00 5.68097830e-01 1.87106624e-01 -2.93250978e-01
1.64697838e+00 -1.08737493e+00 -5.60031354e-01 -1.31109983e-01
3.02330375e-01 -9.08017457e-01 1.01224256e+00 1.92268878e-01
-1.29316545e+00 -1.13082600e+00 -1.06215680e+00 -5.18231332e-01
-6.11557543e-01 4.68308747e-01 5.47685921e-01 3.82924646e-01
-1.50459433e+00 1.20227468e+00 -8.09766412e-01 -1.83558092e-01
4.82153893e-01 4.00774777e-01 -6.25388861e-01 -7.16095697e-03
-1.10462797e+00 8.57764482e-01 4.05056328e-01 5.96372008e-01
-1.32437602e-01 -7.79316068e-01 -9.45760250e-01 2.83780754e-01
4.84581068e-02 -5.98652184e-01 6.18997514e-01 -1.07756150e+00
-1.34911227e+00 6.96305811e-01 -3.63756508e-01 -5.58694601e-01
3.79638344e-01 -1.87021494e-01 -5.12542546e-01 2.35438585e-01
-1.82960659e-01 5.31909525e-01 1.26697755e+00 -1.07188010e+00
-6.42898381e-01 -3.22389454e-01 -4.63571623e-02 2.32894197e-01
-5.35821378e-01 6.25761449e-02 -3.50066125e-01 -9.62163210e-01
2.50738323e-01 -1.41577229e-01 -4.42387998e-01 2.17588127e-01
-5.20364717e-02 -9.44840163e-03 7.49067783e-01 -1.11414170e+00
1.15142262e+00 -2.02932334e+00 -1.69162545e-02 4.30924334e-02
1.38780713e-01 5.66945016e-01 -4.57059294e-01 1.01556152e-01
-2.33461231e-01 -2.74145752e-01 -3.89336050e-01 -3.71752441e-01
-3.29465091e-01 3.78224179e-02 -6.16827514e-03 3.74006033e-01
5.53982913e-01 7.83388972e-01 -5.01716971e-01 -5.36376715e-01
6.02129936e-01 1.14430892e+00 -5.11537910e-01 -3.19012739e-02
2.70732164e-01 4.39091355e-01 -1.96905971e-01 3.40639353e-01
1.24695134e+00 -7.57708549e-02 -2.59190321e-01 -9.45505261e-01
-6.35575891e-01 -1.36116803e-01 -1.46126533e+00 1.67202342e+00
-7.29200959e-01 4.70229745e-01 5.15299022e-01 -1.19405115e+00
1.03523290e+00 4.36479747e-02 5.72428107e-02 -6.43835545e-01
3.60850431e-02 1.09763183e-01 -2.11892277e-01 -3.25334072e-01
6.33606732e-01 4.07856591e-02 5.14057875e-01 -1.89350694e-01
2.06174469e-03 7.26309121e-02 1.84499294e-01 -1.29972979e-01
9.36168909e-01 1.30214497e-01 4.30678457e-01 -5.12985945e-01
1.14097404e+00 -4.82195318e-01 4.76643085e-01 8.02984595e-01
-2.10346892e-01 7.32047379e-01 1.09247655e-01 -6.23852491e-01
-1.16604817e+00 -9.56240892e-01 -2.64421225e-01 1.13176012e+00
1.68830931e-01 -1.31080747e-01 -9.52702761e-01 -3.67517829e-01
-2.08781213e-01 2.83646494e-01 -8.43545675e-01 9.89450216e-02
-8.30592155e-01 -7.22911239e-01 6.34392202e-02 6.95754707e-01
1.20746493e+00 -1.26422548e+00 -4.17276502e-01 3.49111587e-01
2.97064543e-01 -1.03757572e+00 -5.45761168e-01 3.31983775e-01
-1.11366963e+00 -7.70702004e-01 -8.63013148e-01 -1.02619803e+00
7.96418667e-01 4.27427888e-01 8.20478141e-01 2.34062687e-01
-4.26376224e-01 -5.14412001e-02 4.02168408e-02 2.29077443e-01
-1.53853282e-01 -1.99014083e-01 -2.05981210e-01 2.91650176e-01
1.51456773e-01 -9.42908525e-01 -9.23323274e-01 1.12061366e-01
-1.02842641e+00 1.35122269e-01 9.37468767e-01 1.18880749e+00
7.90519118e-01 4.45585310e-01 5.44587374e-01 -6.77736163e-01
5.27073026e-01 1.36723459e-01 -6.21544480e-01 2.58288175e-01
-2.52699435e-01 1.89577460e-01 9.06502247e-01 -2.74604559e-01
-1.56726503e+00 1.04368106e-01 -4.41385746e-01 -1.18575856e-01
-3.58267128e-01 2.87459433e-01 -4.36423421e-02 -5.38647354e-01
5.96149564e-01 3.87781590e-01 -7.33797327e-02 -7.23626494e-01
4.86190528e-01 3.39841247e-01 8.57943416e-01 -2.29913354e-01
9.07597423e-01 7.17845976e-01 2.38464445e-01 -1.09984446e+00
-8.05998445e-01 -3.17229569e-01 -1.03498614e+00 1.67955887e-02
1.03273797e+00 -9.36663628e-01 -6.70293331e-01 8.36821318e-01
-1.07881021e+00 -2.12464482e-01 -3.39258641e-01 2.70637691e-01
-2.88711637e-01 5.45125246e-01 -8.84087682e-01 -4.46537554e-01
-5.53380847e-01 -1.11721301e+00 9.46688890e-01 6.38244927e-01
4.70993102e-01 -1.06966436e+00 -3.92739624e-01 -9.91922021e-02
9.83301461e-01 4.20900695e-02 6.16659760e-01 -5.16899377e-02
-6.62526250e-01 -3.00513431e-02 -9.15548325e-01 8.67630839e-01
2.65264094e-01 -2.99985588e-01 -9.24909294e-01 -3.40585351e-01
1.57024533e-01 -1.67449620e-02 1.34062624e+00 7.46034682e-01
1.41304684e+00 -1.21977724e-01 4.96459529e-02 8.86976123e-01
1.75335693e+00 -4.25215662e-01 9.07543778e-01 3.80574822e-01
6.79446161e-01 5.34847021e-01 2.79402435e-01 1.50183603e-01
3.88720259e-02 4.65336949e-01 2.56104082e-01 -6.55111372e-01
-3.92522871e-01 2.23793879e-01 1.22620285e-01 7.23259866e-01
-4.89519656e-01 4.90491271e-01 -3.35098684e-01 5.11406600e-01
-1.69864726e+00 -1.06128478e+00 -2.40417093e-01 2.13621330e+00
8.72065187e-01 1.27002701e-01 -3.51305664e-01 1.28689766e-01
8.61416996e-01 4.22147095e-01 -3.69102895e-01 -4.33315516e-01
-4.72653985e-01 6.63814545e-01 6.19955122e-01 6.77839220e-01
-1.47798347e+00 8.86078060e-01 4.95525599e+00 1.14227176e+00
-1.01580453e+00 9.87125039e-02 6.83520555e-01 3.34570110e-01
7.60838240e-02 -2.75292993e-01 -7.23024011e-01 1.51592791e-01
5.42330921e-01 5.34531586e-02 4.00879532e-01 6.62503362e-01
4.29177880e-01 -2.79253513e-01 -6.39052689e-01 8.68167818e-01
-1.19694829e-01 -1.48102236e+00 7.00965971e-02 -3.56587708e-01
9.18447793e-01 -7.10302219e-02 7.49945045e-02 6.08914644e-02
-8.20856169e-02 -9.85197902e-01 2.58112133e-01 7.61307180e-01
6.12456977e-01 -9.08378541e-01 8.10831308e-01 1.37914747e-01
-1.63498652e+00 -1.18116692e-01 -7.17522502e-01 7.24005513e-04
1.14594951e-01 7.32784390e-01 2.92455386e-02 7.30686426e-01
1.04063988e+00 7.79867411e-01 -6.09806716e-01 9.51580048e-01
-1.66922733e-01 1.81854069e-01 -1.63226843e-01 4.34967667e-01
3.72930110e-01 -4.31077421e-01 2.93262661e-01 1.42696404e+00
4.48613733e-01 1.57162696e-01 -7.85153657e-02 7.49433160e-01
-6.13014139e-02 1.13179967e-01 -2.24759668e-01 5.62455356e-01
3.89918648e-02 1.71836162e+00 -8.49510610e-01 -4.40846860e-01
-7.56464720e-01 1.36434174e+00 4.79647100e-01 5.01958072e-01
-4.67252374e-01 -6.58015132e-01 7.80594528e-01 -1.05674580e-01
7.47555792e-01 -2.14146018e-01 -5.08248091e-01 -1.01986873e+00
6.35115877e-02 -5.43723345e-01 2.17709959e-01 -7.24793971e-01
-1.35675752e+00 8.82288933e-01 -4.89297926e-01 -1.00100517e+00
3.16850156e-01 -7.11712837e-01 -7.79768169e-01 1.45655441e+00
-2.08789158e+00 -1.29467976e+00 -6.21045768e-01 9.12191272e-01
5.41684151e-01 1.26151979e-01 6.62579238e-01 4.18332398e-01
-4.17052388e-01 4.02779102e-01 3.16967934e-01 -1.00748740e-01
7.87856996e-01 -1.32518709e+00 3.89219344e-01 1.18261576e+00
-2.65518308e-01 6.69738114e-01 5.04505634e-01 -5.58737278e-01
-8.56707096e-01 -1.22753072e+00 9.32225049e-01 2.44396225e-01
6.84675992e-01 -2.52398103e-02 -1.51112199e+00 3.41006935e-01
2.94981599e-01 5.08138120e-01 -6.02165982e-02 -2.81511605e-01
-1.94741413e-01 -3.85602355e-01 -1.23678553e+00 6.30585372e-01
8.80357862e-01 -5.92367113e-01 -5.09490192e-01 -1.08892992e-02
6.43124342e-01 -2.74954528e-01 -9.45656657e-01 4.90123510e-01
2.39389986e-01 -1.26310325e+00 1.42718852e+00 -2.43666872e-01
5.15284300e-01 -5.22522688e-01 2.30611116e-02 -1.09788680e+00
-5.87991655e-01 -3.56269598e-01 2.54189610e-01 1.14843774e+00
7.56115615e-02 -6.63170338e-01 6.96287692e-01 3.57966870e-01
-2.61171758e-01 -7.06154704e-01 -1.00254846e+00 -5.32178402e-01
-6.97493851e-02 -1.63936242e-01 2.54644275e-01 7.49775767e-01
-5.82458198e-01 -1.93882976e-02 -5.45861602e-01 6.74333811e-01
1.08494139e+00 7.70834088e-02 3.44728708e-01 -1.08739996e+00
-1.05508707e-01 -5.23825407e-01 -4.37354147e-01 -9.48854446e-01
-6.21928126e-02 -7.38540471e-01 -1.68841779e-02 -1.64112592e+00
1.39543086e-01 -3.36067230e-02 -6.29056871e-01 3.18710893e-01
-4.48495388e-01 7.95436859e-01 -3.22369039e-02 -5.71922623e-02
-2.00449988e-01 4.92281973e-01 1.43663597e+00 9.00389999e-02
-3.19690347e-01 7.53905717e-03 -4.77789819e-01 1.04617798e+00
8.36011648e-01 1.04613066e-01 -4.29039747e-02 -3.94738793e-01
-2.84468353e-01 -2.49134362e-01 6.79053307e-01 -1.26922679e+00
5.28043389e-01 9.33966562e-02 7.08350420e-01 -6.22516751e-01
3.32467616e-01 -7.96015620e-01 -6.32859319e-02 3.07376027e-01
-4.00591761e-01 -1.54998958e-01 3.03770334e-01 4.96583790e-01
-4.37332511e-01 -2.82969534e-01 1.28628659e+00 -1.83181286e-01
-1.19442797e+00 1.95847750e-01 -1.71199098e-01 -5.23910642e-01
7.15095758e-01 -3.85621667e-01 -2.35549599e-01 7.37810805e-02
-1.07441282e+00 -2.59504080e-01 3.29305083e-01 1.24641009e-01
6.96928382e-01 -1.22438979e+00 -6.75476372e-01 3.93064827e-01
-3.24462771e-01 -2.44610325e-01 9.50165689e-01 1.00370073e+00
-4.85326797e-01 4.79659587e-02 -5.88470459e-01 -3.86603177e-01
-1.48770857e+00 5.84228396e-01 4.51326907e-01 -3.56871307e-01
-9.11700368e-01 8.81047368e-01 3.55265647e-01 -1.93623364e-01
2.76868157e-02 -4.90481466e-01 -4.24188673e-01 -5.90724275e-02
1.00460994e+00 3.66683871e-01 1.61711663e-01 -7.15395868e-01
-1.33531958e-01 9.56160069e-01 -2.15966582e-01 3.82072538e-01
1.71800745e+00 -3.74854386e-01 -5.82525492e-01 -5.22722751e-02
1.22292662e+00 -7.42371753e-02 -1.59959674e+00 -7.16522932e-01
-1.17547169e-01 -4.78166193e-01 5.45125186e-01 -5.51377892e-01
-1.28070998e+00 9.76848245e-01 9.38792467e-01 7.56074935e-02
1.71904433e+00 -3.44518840e-01 6.58071399e-01 2.88283378e-01
1.09677315e-01 -1.09370065e+00 -1.46107972e-02 3.30520272e-01
1.04149413e+00 -1.40013123e+00 1.17551245e-01 -5.11865795e-01
-2.06509188e-01 1.45649290e+00 4.87993985e-01 -4.16436523e-01
7.25848615e-01 3.51977527e-01 -3.71631980e-02 -6.55101426e-03
-2.76040077e-01 -5.20206749e-01 5.70176363e-01 5.33548474e-01
4.76156354e-01 -1.34775460e-01 -4.81885970e-01 3.95166367e-01
3.23260188e-01 2.14439668e-02 3.46672297e-01 6.43725038e-01
-7.70188928e-01 -1.06132722e+00 -3.28735650e-01 3.46039891e-01
-4.84238625e-01 -5.11863232e-01 4.02797073e-01 6.38526917e-01
3.29036951e-01 5.96500397e-01 8.46650973e-02 -9.53212753e-02
4.04078633e-01 -2.06751376e-01 3.19370151e-01 -8.07332993e-02
-7.28043556e-01 9.01404321e-02 -2.08498061e-01 -8.40362430e-01
-7.40742445e-01 -3.71616662e-01 -1.04183519e+00 -2.33448878e-01
-1.15701035e-01 -2.22497642e-01 5.42748213e-01 6.36454165e-01
2.61799097e-01 8.01366568e-01 4.02141005e-01 -1.46867347e+00
-4.55011606e-01 -9.25329566e-01 -7.72824168e-01 3.60778868e-01
5.40134072e-01 -3.38829517e-01 -3.24185371e-01 3.34667772e-01] | [10.942111015319824, -1.8173872232437134] |
5b4806eb-37b7-4513-98f2-13fd3d23ea74 | discovering-the-real-association-multimodal | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zang_Discovering_the_Real_Association_Multimodal_Causal_Reasoning_in_Video_Question_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zang_Discovering_the_Real_Association_Multimodal_Causal_Reasoning_in_Video_Question_CVPR_2023_paper.pdf | Discovering the Real Association: Multimodal Causal Reasoning in Video Question Answering | Video Question Answering (VideoQA) is challenging as it requires capturing accurate correlations between modalities from redundant information. Recent methods focus on the explicit challenges of the task, e.g. multimodal feature extraction, video-text alignment and fusion. Their frameworks reason the answer relying on statistical evidence causes, which ignores potential bias in the multimodal data. In our work, we investigate relational structure from a causal representation perspective on multimodal data and propose a novel inference framework. For visual data, question-irrelevant objects may establish simple matching associations with the answer. For textual data, the model prefers the local phrase semantics which may deviate from the global semantics in long sentences. Therefore, to enhance the generalization of the model, we discover the real association by explicitly capturing visual features that are causally related to the question semantics and weakening the impact of local language semantics on question answering. The experimental results on two large causal VideoQA datasets verify that our proposed framework 1) improves the accuracy of the existing VideoQA backbone, 2) demonstrates robustness on complex scenes and questions. | ['Wei Liang', 'Mingtao Pei', 'Hanqing Wang', 'Chuanqi Zang'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['video-question-answering'] | ['computer-vision'] | [ 2.17532605e-01 -1.37956157e-01 -2.19461784e-01 -4.97728527e-01
-5.60997486e-01 -4.89762366e-01 8.00019026e-01 1.74387380e-01
-2.16449827e-01 4.69882041e-01 9.01470542e-01 1.61601994e-02
-3.65689546e-01 -4.50957417e-01 -7.69952595e-01 -6.07131660e-01
3.16688687e-01 3.20847780e-02 4.98726100e-01 -2.13555157e-01
3.09324384e-01 -2.20375553e-01 -1.68378842e+00 8.18257034e-01
6.57390118e-01 1.09498692e+00 1.68686464e-01 4.79580283e-01
-3.79568070e-01 1.50487649e+00 -3.63093287e-01 -5.31544924e-01
-8.87010172e-02 -5.62905788e-01 -7.82715440e-01 1.17063522e-01
8.62512410e-01 -5.06936610e-01 -4.81172532e-01 1.08614683e+00
1.97028771e-01 -8.22058097e-02 4.75945204e-01 -1.57415295e+00
-6.18909299e-01 4.88886297e-01 -5.19537747e-01 2.17172951e-01
8.91124487e-01 8.65881369e-02 1.33303821e+00 -8.32064092e-01
6.20479465e-01 1.62355614e+00 3.84784281e-01 2.64142424e-01
-7.78671026e-01 -4.10505623e-01 5.99695146e-01 7.99998879e-01
-1.35346127e+00 -5.15350521e-01 1.06568122e+00 -4.08981025e-01
3.64538878e-01 5.47448575e-01 2.22361118e-01 1.21033621e+00
-6.33136109e-02 9.81655955e-01 1.00833416e+00 -1.97136983e-01
2.03308519e-02 -1.73332822e-02 2.83024192e-01 6.93952084e-01
-1.88428145e-02 -3.44627321e-01 -9.57685769e-01 -2.50288904e-01
3.64481688e-01 7.27002369e-03 -4.25541431e-01 -4.68808472e-01
-1.61203921e+00 6.33372903e-01 2.50762016e-01 2.20661953e-01
-3.54258269e-01 1.75393865e-01 2.81785727e-01 1.99129194e-01
-4.72178012e-02 2.71325037e-02 -4.18065697e-01 -5.85938767e-02
-4.99031425e-01 3.97367984e-01 5.47171354e-01 1.06865907e+00
6.94183707e-01 -4.14912611e-01 -4.22098875e-01 5.21567345e-01
7.14751899e-01 6.21852040e-01 1.62053093e-01 -1.10216916e+00
7.64892161e-01 1.01697588e+00 7.59316236e-02 -1.64120638e+00
-2.20972165e-01 1.15606427e-01 -5.50844789e-01 -5.83842576e-01
4.80087250e-01 1.30572110e-01 -5.01156926e-01 1.83611834e+00
4.91715372e-01 8.18848163e-02 -5.66633679e-02 1.20310378e+00
1.34040368e+00 4.82998192e-01 2.01912969e-01 -4.66399997e-01
1.57943261e+00 -8.67348135e-01 -1.11367333e+00 -6.38287812e-02
3.54618102e-01 -7.90645361e-01 1.17747116e+00 4.62656543e-02
-7.85075068e-01 -5.44679046e-01 -7.60011852e-01 -3.28838110e-01
-2.45508015e-01 2.30347458e-02 4.12803411e-01 2.40987524e-01
-5.91864347e-01 -1.20371372e-01 -3.45225185e-01 -4.76870328e-01
1.21478729e-01 1.99992985e-01 -4.81446922e-01 -4.40368235e-01
-1.36970198e+00 5.96342862e-01 4.65597093e-01 2.35157266e-01
-7.63109267e-01 -3.83132070e-01 -6.46749377e-01 -1.95984468e-01
9.64360654e-01 -8.99640262e-01 1.01806152e+00 -1.20774150e+00
-1.08915246e+00 4.14739430e-01 -5.38754940e-01 7.31037110e-02
2.81912506e-01 -4.02284086e-01 -3.87664497e-01 6.81077659e-01
1.19812086e-01 4.51037169e-01 1.02041900e+00 -1.55768192e+00
-5.90436816e-01 -5.81298113e-01 4.13766474e-01 3.99674654e-01
-3.28315735e-01 8.52364451e-02 -7.29014277e-01 -4.24104065e-01
2.03391925e-01 -6.44102573e-01 2.52345055e-01 -1.63100794e-01
-2.28383765e-01 -3.24947417e-01 9.96359944e-01 -7.23424077e-01
1.29782224e+00 -2.03711295e+00 4.34323967e-01 1.33679494e-01
3.48492652e-01 -2.91758001e-01 -2.00448781e-01 5.45002043e-01
1.02620959e-01 3.25139239e-02 -3.86531316e-02 7.07179382e-02
1.80599913e-02 5.12996554e-01 -5.54094911e-01 3.32425445e-01
2.91903913e-01 1.05127788e+00 -9.41575468e-01 -1.05664146e+00
-1.07919529e-01 1.49492845e-01 -4.83476877e-01 4.12865490e-01
-4.19823349e-01 3.72377247e-01 -7.95576692e-01 8.17297101e-01
5.34532070e-01 -4.18444365e-01 3.01646262e-01 -9.31704104e-01
3.72707471e-02 -2.60699913e-03 -1.17813683e+00 1.88740051e+00
7.63691515e-02 3.43337178e-01 6.80545121e-02 -9.75201428e-01
5.71618140e-01 3.71102244e-01 5.59172273e-01 -6.65208757e-01
-6.96998369e-03 -1.58080593e-01 -1.25787660e-01 -1.33042884e+00
4.36273992e-01 4.26521003e-02 -2.52228770e-02 1.53442696e-01
1.88518003e-01 2.78676629e-01 1.03811704e-01 6.18773818e-01
8.93317580e-01 3.97522807e-01 1.95480093e-01 -4.31319512e-02
8.15282583e-01 9.63796750e-02 5.27613759e-01 7.44810402e-01
-2.69447118e-01 5.18366694e-01 7.81630695e-01 -1.69361442e-01
-5.75513542e-01 -1.02947628e+00 1.92895785e-01 1.21441853e+00
7.61887431e-01 -8.38607490e-01 -4.13092345e-01 -9.20318127e-01
-2.03362554e-01 4.94132459e-01 -7.29398608e-01 -5.58805279e-02
-4.71471280e-01 -5.93220055e-01 4.64284360e-01 3.34349871e-01
5.37166059e-01 -5.97640753e-01 -3.04994881e-01 -1.78128794e-01
-9.80056345e-01 -1.43882036e+00 -3.85069370e-01 -5.56544423e-01
-6.26219273e-01 -1.38456988e+00 -2.57747889e-01 -3.72205615e-01
5.52151203e-01 5.55503607e-01 1.05701053e+00 2.50050396e-01
7.09966198e-02 1.10170114e+00 -5.98234296e-01 -1.17492296e-01
-3.28477807e-02 -3.15072030e-01 -5.19342422e-02 5.12621284e-01
6.12479091e-01 -1.98442444e-01 -6.41380370e-01 5.35257936e-01
-9.94759619e-01 7.70261139e-02 4.89464819e-01 6.76316142e-01
4.23416436e-01 -1.35982111e-01 5.48560083e-01 -4.08829927e-01
5.02076149e-01 -7.20733225e-01 -8.35375339e-02 6.96751475e-01
-2.67438948e-01 1.86807692e-01 2.44375765e-01 -3.90420049e-01
-1.47722435e+00 -8.85360613e-02 4.59663957e-01 -5.14328957e-01
-1.90832779e-01 5.56722403e-01 -5.18890202e-01 2.25491717e-01
2.47706518e-01 1.62608311e-01 -8.03164393e-02 -3.13348919e-01
5.68465769e-01 2.44034022e-01 4.57771331e-01 -9.40723300e-01
7.89307237e-01 7.69677520e-01 1.55115634e-01 -6.66267753e-01
-8.85307431e-01 -7.24352002e-01 -6.76474214e-01 -6.45741165e-01
1.11706972e+00 -9.78754461e-01 -9.91356194e-01 -6.73421621e-02
-1.37709963e+00 5.66439152e-01 1.88424021e-01 5.41217327e-01
-3.58247548e-01 8.91836286e-01 -3.28844190e-01 -8.48060250e-01
1.80848867e-01 -9.57690239e-01 1.19965708e+00 6.33165315e-02
-1.67652965e-01 -6.39913380e-01 -5.18613160e-02 9.58510935e-01
-3.95528087e-03 3.93434912e-02 1.00782263e+00 -4.94888008e-01
-9.21194136e-01 6.33348003e-02 -5.12678206e-01 -4.15501855e-02
1.23698659e-01 2.37636596e-01 -9.49661374e-01 2.19063088e-01
8.78636539e-02 -4.37298924e-01 9.14148986e-01 1.30617283e-02
9.00570393e-01 -4.22649801e-01 -1.12406470e-01 -6.48998022e-02
1.13342917e+00 -5.79280928e-02 4.49227273e-01 5.74231334e-02
1.02879179e+00 1.21147323e+00 7.88085341e-01 2.47921452e-01
8.95289600e-01 5.69852352e-01 7.75157273e-01 2.05838516e-01
-3.43870819e-02 -5.68838656e-01 4.78095382e-01 1.12547088e+00
-2.83187419e-01 -1.18522562e-01 -8.37593675e-01 4.78085279e-01
-2.50567675e+00 -1.13188648e+00 -6.40757024e-01 1.78150594e+00
7.14589596e-01 -3.96247864e-01 -9.70537513e-02 -1.43656865e-01
5.71062565e-01 2.24978462e-01 -1.78868890e-01 2.17133686e-01
-4.57349211e-01 -5.93466640e-01 5.48744202e-03 3.97310287e-01
-9.18347418e-01 6.90078437e-01 5.55692530e+00 7.25640833e-01
-4.83123213e-01 1.05555095e-01 1.89370036e-01 -7.24119991e-02
-6.95325196e-01 2.08227828e-01 -5.96861362e-01 2.22371250e-01
3.75523448e-01 1.61543727e-01 2.02508956e-01 2.21998408e-01
3.54284525e-01 -2.76136458e-01 -1.18159425e+00 1.00617552e+00
5.38316965e-01 -1.04774261e+00 5.90930343e-01 -1.58013523e-01
4.29302722e-01 -5.05922854e-01 -2.24077329e-01 1.21430516e-01
-1.55987576e-01 -8.41569424e-01 9.04977322e-01 1.07100928e+00
7.97634721e-02 -4.50141788e-01 7.87813008e-01 3.72608989e-01
-1.17696428e+00 -1.59960315e-01 -1.69787645e-01 -4.01257612e-02
2.45644599e-01 2.73371220e-01 -4.54498917e-01 9.25731361e-01
8.90602767e-01 8.10505271e-01 -1.01792884e+00 5.15720904e-01
-1.38137892e-01 5.03187001e-01 -1.68162689e-01 2.01304443e-03
6.85102269e-02 -6.41196594e-02 6.04274631e-01 9.94943142e-01
1.59119423e-02 4.60642606e-01 1.25054732e-01 7.32561648e-01
7.73619860e-02 4.41096038e-01 -5.42862058e-01 -1.28270388e-01
1.45491481e-01 1.00447130e+00 -3.35541129e-01 -3.04517806e-01
-8.31677318e-01 7.16764033e-01 1.03908844e-01 6.32320464e-01
-1.06045043e+00 2.05733284e-01 2.86636651e-01 -2.33319968e-01
1.84127480e-01 -2.24348739e-01 -9.21116397e-02 -1.51590049e+00
3.56293261e-01 -1.00905728e+00 7.19123244e-01 -1.09576023e+00
-1.42092383e+00 2.34569952e-01 2.42438585e-01 -1.18729877e+00
3.70096639e-02 -4.90230560e-01 -1.58853531e-01 3.69159043e-01
-1.41582763e+00 -1.54837704e+00 -5.39627373e-01 1.16272318e+00
3.88968617e-01 5.45435213e-02 3.88716996e-01 3.35390627e-01
-4.67176259e-01 1.26465753e-01 -5.23996830e-01 -3.37366350e-02
1.11906099e+00 -7.48676658e-01 -8.18590760e-01 8.29710841e-01
2.72068322e-01 7.62376904e-01 8.58325720e-01 -6.79182529e-01
-1.93539596e+00 -6.13400757e-01 1.02807093e+00 -7.70148575e-01
9.37684238e-01 -8.60378668e-02 -9.80189919e-01 4.47553486e-01
4.94867980e-01 -3.41148317e-01 7.61163414e-01 1.32153541e-01
-7.29763091e-01 -1.91429570e-01 -6.79081917e-01 7.08365023e-01
1.06606829e+00 -7.96622932e-01 -9.97050226e-01 1.46857783e-01
7.94063091e-01 9.07722488e-02 -7.75096953e-01 6.49953246e-01
5.68459332e-01 -9.63056624e-01 9.11166668e-01 -8.91707778e-01
8.10226858e-01 -7.75743186e-01 -6.86843038e-01 -5.09234190e-01
-1.22504115e-01 -1.94295391e-01 -4.18863535e-01 1.47457588e+00
1.77557096e-01 1.59488097e-01 3.06333154e-01 7.15224206e-01
3.28087002e-01 -2.54538447e-01 -9.88487005e-01 -1.96661577e-01
-5.34740329e-01 -5.58368862e-01 4.72710222e-01 1.19049191e+00
6.54433146e-02 5.94954014e-01 -6.40341282e-01 5.86562812e-01
6.70316756e-01 3.21319193e-01 7.42654204e-01 -9.20925796e-01
-1.82681605e-01 -2.36159340e-01 -3.55885983e-01 -1.12571633e+00
2.14210153e-01 -4.28616077e-01 -3.35527696e-02 -1.45941138e+00
5.50960302e-01 2.77496368e-01 -2.46812820e-01 3.16288918e-01
-5.21364748e-01 -6.81663603e-02 3.30421627e-01 3.30296993e-01
-1.16400647e+00 7.72804677e-01 1.30165184e+00 -3.56037140e-01
2.95381397e-01 -6.27794206e-01 -5.80923796e-01 9.41168606e-01
4.42548424e-01 -2.17952192e-01 -6.49054170e-01 -6.06413841e-01
7.02270925e-01 1.74710959e-01 8.00250232e-01 -3.76184642e-01
4.39321011e-01 -5.58490634e-01 1.25386521e-01 -5.91977417e-01
2.67812997e-01 -1.19128752e+00 -1.45475179e-01 -4.75601889e-02
-3.92063856e-01 1.56681463e-01 -1.27159670e-01 1.05173242e+00
-5.96341670e-01 -1.01634532e-01 -2.99479924e-02 -3.60226706e-02
-9.77198243e-01 7.08157346e-02 -1.51004806e-01 1.22108772e-01
8.24565589e-01 -1.21949963e-01 -5.50316393e-01 -6.39569163e-01
-4.60208029e-01 5.24178028e-01 1.48419276e-01 8.53447080e-01
7.65174031e-01 -1.60670352e+00 -6.23912334e-01 -3.69173676e-01
5.88001013e-01 -3.09186608e-01 4.57782447e-01 1.20085394e+00
6.68225139e-02 4.67390895e-01 1.55100212e-01 -7.97058165e-01
-1.48566771e+00 7.73218930e-01 5.70559651e-02 2.82095760e-01
-8.99605751e-02 5.60038447e-01 6.06887937e-01 -3.79330851e-02
2.13731498e-01 -2.40459695e-01 -6.60780549e-01 3.75795394e-01
4.52744246e-01 2.88718134e-01 -4.12025660e-01 -1.08426476e+00
-5.73520720e-01 6.48983836e-01 1.39895812e-01 -9.52602997e-02
7.93640554e-01 -7.05056608e-01 -4.42896217e-01 7.02175200e-01
9.90593255e-01 1.14285953e-01 -1.05705237e+00 -4.63821292e-01
1.56736374e-01 -6.15408063e-01 -5.67811765e-02 -6.20774329e-01
-9.10890102e-01 8.14738452e-01 3.02128315e-01 1.12652719e-01
1.09455884e+00 3.33964407e-01 3.96365792e-01 3.88212740e-01
1.25929369e-02 -1.00531960e+00 4.10043329e-01 3.64038527e-01
1.06279027e+00 -1.49526238e+00 8.83055478e-02 -5.49772203e-01
-8.37339282e-01 1.32410586e+00 7.41243839e-01 4.33262527e-01
4.37510937e-01 -2.86310285e-01 3.16773690e-02 -5.74051261e-01
-1.07868099e+00 -2.77410984e-01 6.47792876e-01 1.87301636e-01
3.42517167e-01 -1.40428185e-01 -4.59306389e-01 7.83338726e-01
3.37687075e-01 -1.08399838e-01 2.52252579e-01 7.40117610e-01
-2.65041113e-01 -7.98403740e-01 -5.80157042e-01 2.30539925e-02
-3.14053297e-01 -1.71299040e-01 -7.27677226e-01 6.71769202e-01
3.57705683e-01 1.35281742e+00 -1.18016347e-01 -3.24600369e-01
3.23453575e-01 1.46261156e-01 5.23030937e-01 6.48960024e-02
-2.38128632e-01 2.37567544e-01 2.82966077e-01 -8.03692341e-01
-1.23819685e+00 -7.28033125e-01 -1.09476626e+00 -4.17414010e-02
-2.28451967e-01 -1.56773478e-02 3.87140036e-01 1.28193438e+00
3.31524462e-01 4.09958601e-01 3.65179926e-01 -5.94566166e-02
-2.70963013e-01 -6.74673140e-01 -1.39929593e-01 7.78190315e-01
3.73568147e-01 -7.00874567e-01 -4.27785009e-01 5.14941454e-01] | [10.374366760253906, 1.1180083751678467] |
146ea12e-68e8-400f-ba99-b3e053fa9535 | a-review-of-driver-gaze-estimation-and | 2307.01470 | null | https://arxiv.org/abs/2307.01470v1 | https://arxiv.org/pdf/2307.01470v1.pdf | A Review of Driver Gaze Estimation and Application in Gaze Behavior Understanding | Driver gaze plays an important role in different gaze-based applications such as driver attentiveness detection, visual distraction detection, gaze behavior understanding, and building driver assistance system. The main objective of this study is to perform a comprehensive summary of driver gaze fundamentals, methods to estimate driver gaze, and it's applications in real world driving scenarios. We first discuss the fundamentals related to driver gaze, involving head-mounted and remote setup based gaze estimation and the terminologies used for each of these data collection methods. Next, we list out the existing benchmark driver gaze datasets, highlighting the collection methodology and the equipment used for such data collection. This is followed by a discussion of the algorithms used for driver gaze estimation, which primarily involves traditional machine learning and deep learning based techniques. The estimated driver gaze is then used for understanding gaze behavior while maneuvering through intersections, on-ramps, off-ramps, lane changing, and determining the effect of roadside advertising structures. Finally, we have discussed the limitations in the existing literature, challenges, and the future scope in driver gaze estimation and gaze-based applications. | ['Pranamesh Chakraborty', 'Pavan Kumar Sharma'] | 2023-07-04 | null | null | null | null | ['gaze-estimation'] | ['computer-vision'] | [-1.81455016e-01 -2.30512731e-02 -4.12275225e-01 -6.91398501e-01
-1.30707189e-01 -2.01185688e-01 -1.15545660e-01 -3.39215249e-01
-3.83680671e-01 4.30967838e-01 -5.53699434e-02 -7.23751783e-01
-2.78160185e-01 9.03426334e-02 -1.63531274e-01 -8.72411311e-01
3.04872364e-01 -4.44976896e-01 1.46737084e-01 -6.06085420e-01
1.12411654e+00 2.31616870e-01 -2.68500233e+00 -3.93145263e-01
6.47152066e-01 9.74790156e-01 1.09901607e-01 1.09528863e+00
5.52216778e-03 9.01383281e-01 -8.44304740e-01 -4.08600867e-02
-2.81850457e-01 -3.01340163e-01 -4.10127372e-01 -2.77380466e-01
9.13396358e-01 -4.54712123e-01 5.69399334e-02 6.59388661e-01
7.65651762e-01 2.64080316e-01 4.66578096e-01 -2.09930825e+00
-5.97560048e-01 -4.51838076e-01 -9.62269962e-01 1.04433858e+00
5.88056862e-01 4.45596546e-01 2.78261930e-01 -4.76483524e-01
-7.99338240e-03 9.94687319e-01 4.29938108e-01 8.69756401e-01
-3.38160604e-01 -1.20889425e+00 3.09169471e-01 8.48957658e-01
-1.04635084e+00 -7.10844100e-01 8.87017131e-01 -7.67601609e-01
7.12185562e-01 3.97691756e-01 4.24052030e-01 7.18248546e-01
5.64105570e-01 8.42588246e-01 1.42455757e+00 -6.61490977e-01
-3.87190849e-01 5.40815592e-01 1.10197544e+00 5.43908656e-01
1.47530496e-01 5.51865637e-01 -1.02185810e+00 4.45081562e-01
-1.76927373e-01 -1.66731730e-01 -2.99922079e-01 1.38762742e-01
-6.14156544e-01 7.68390715e-01 5.81707537e-01 -3.44090790e-01
-2.58559585e-01 -1.36070788e-01 1.69788688e-01 2.06097335e-01
3.75821650e-01 -6.29155487e-02 -1.09802127e-01 -3.68604094e-01
-5.64858079e-01 3.71357948e-01 3.45130384e-01 1.36294281e+00
8.07169735e-01 -7.61369765e-02 -2.95768589e-01 5.52209198e-01
6.41448677e-01 9.00478244e-01 2.72624761e-01 -6.82704210e-01
2.31101662e-01 3.77889425e-01 2.10113943e-01 -6.82726204e-01
-6.24984443e-01 3.20063055e-01 1.90277755e-01 9.21284437e-01
2.66708553e-01 -4.89047021e-01 -7.67786264e-01 1.36391246e+00
3.81042868e-01 1.10273376e-01 -2.91743159e-01 1.13835073e+00
1.28249657e+00 -1.44229345e-02 3.77707809e-01 -3.49564075e-01
1.93957484e+00 -9.18674111e-01 -1.55165255e+00 -4.03682292e-01
8.01266849e-01 -8.67232800e-01 1.19769537e+00 1.28676474e-01
-9.65432048e-01 -7.63523817e-01 -9.36750114e-01 -6.71968520e-01
-6.63338721e-01 -1.14407167e-01 3.43230933e-01 1.48896325e+00
-1.07292235e+00 -6.57378078e-01 -3.56542379e-01 -3.92589808e-01
4.30157602e-01 5.71093500e-01 -3.50787006e-02 1.48244709e-01
-8.43238473e-01 1.30855536e+00 -3.18159699e-01 3.94728363e-01
-3.25300604e-01 -8.70081663e-01 -1.02655435e+00 -7.88882002e-02
2.93499917e-01 -4.81673717e-01 1.75483894e+00 -6.08441472e-01
-1.44069719e+00 1.15230811e+00 -1.48018229e+00 -1.67526618e-01
-4.29032259e-02 -3.46720517e-01 -9.49330807e-01 -5.26747942e-01
-1.34275228e-01 9.39388692e-01 9.67481613e-01 -1.00985563e+00
-1.59658706e+00 -5.39442062e-01 2.15126589e-01 1.58777103e-01
4.43298280e-01 5.71810126e-01 5.96175976e-02 5.95181227e-01
-6.12053931e-01 -1.12994087e+00 6.38423622e-01 -2.64605790e-01
-3.90459359e-01 -8.16213369e-01 1.49972117e+00 -3.03979039e-01
1.54625094e+00 -2.20743108e+00 -6.35547101e-01 -5.23119047e-02
9.55127597e-01 3.36238265e-01 3.15539181e-01 3.83173972e-02
-3.19801927e-01 -1.72212139e-01 3.85430425e-01 -3.03587884e-01
5.30108213e-02 -1.44669116e-01 -4.16544564e-02 4.20401007e-01
1.70824945e-01 1.14066160e+00 -6.48293495e-01 -4.25087929e-01
5.17986059e-01 3.97520602e-01 3.32966223e-02 2.99469978e-01
7.09815085e-01 5.75913906e-01 -3.79085205e-02 6.36929572e-01
8.65460038e-01 4.13897753e-01 -8.47169220e-01 2.29471717e-02
-1.07728076e+00 4.68718380e-01 -4.18363541e-01 6.32503450e-01
-3.84811938e-01 1.83127129e+00 -1.97221562e-01 -3.40125710e-02
7.83437908e-01 -1.21494666e-01 -1.93951860e-01 -1.13494122e+00
7.27550805e-01 -1.38840541e-01 3.94732267e-01 -1.20592582e+00
7.97120988e-01 3.40458512e-01 2.42269754e-01 7.46807456e-01
-4.26192343e-01 6.42476678e-01 1.35991797e-01 -3.11292857e-01
3.25439483e-01 -4.17396128e-01 3.07932287e-01 -3.29669029e-01
9.37086225e-01 -1.43851399e-01 -9.38445479e-02 1.36983529e-01
-1.27453995e+00 1.16093732e-01 6.27014697e-01 -4.28066432e-01
-1.38159037e-01 -4.31199640e-01 -1.38304397e-01 1.82779229e+00
5.30787647e-01 1.52134836e-01 -1.27556133e+00 -5.56311131e-01
-6.54898062e-02 1.19694901e+00 -1.16922629e+00 -3.52118134e-01
-4.79161024e-01 -2.16854274e-01 1.09237410e-01 4.03953075e-01
3.72776687e-01 -1.18201363e+00 -1.20857775e+00 -8.79264474e-01
-2.06867903e-01 -5.85932136e-01 -7.40873873e-01 -3.16964924e-01
-1.86475396e-01 -1.40359282e+00 -3.50479156e-01 -8.22572112e-01
4.93490130e-01 1.25431502e+00 9.14397120e-01 3.85304451e-01
-1.02564096e-01 4.77903068e-01 1.25065163e-01 -1.63896573e+00
-1.82035528e-02 3.45333338e-01 4.38865162e-02 -2.48812558e-03
1.79583311e+00 4.64726478e-01 -9.46106553e-01 6.05301857e-01
-3.26152034e-02 -5.42962849e-01 3.53638858e-01 3.03267896e-01
-1.07918195e-01 -4.06614393e-01 5.50426483e-01 -8.76480639e-01
1.14546955e+00 -4.60923761e-01 -7.21771777e-01 1.01760574e-01
-1.04564118e+00 -3.66557240e-01 -7.63969004e-01 -2.36194152e-02
-1.21212935e+00 -6.29969299e-01 -5.69329131e-03 -2.19267771e-01
-7.43296266e-01 -9.53306407e-02 1.21798664e-01 -4.29255068e-01
1.04216301e+00 -2.70319283e-01 6.13408804e-01 -8.69716257e-02
-2.02953015e-02 1.33882344e+00 1.24670029e-01 5.08550227e-01
2.66172051e-01 3.26147616e-01 -1.14534222e-01 -8.54020774e-01
-9.49841678e-01 -9.56072569e-01 -7.84050524e-01 -8.68611515e-01
9.07927990e-01 -6.01474106e-01 -2.02160597e+00 7.23073304e-01
-7.58411646e-01 -8.53200108e-02 2.46473581e-01 5.85740805e-01
-4.03715402e-01 -5.04804432e-01 6.29709288e-02 -1.40473568e+00
-3.20282608e-01 -1.51222932e+00 1.19866920e+00 8.17864180e-01
-3.14045936e-01 -8.76777589e-01 1.18756317e-01 8.74557018e-01
6.38742924e-01 -1.36400789e-01 5.09558439e-01 -4.05002274e-02
-4.22513634e-01 -2.33589545e-01 -3.77243519e-01 -4.54522520e-02
3.15995216e-01 1.90910265e-01 -1.70171976e+00 -1.26169235e-01
-3.81645709e-02 1.19790860e-01 3.20209026e-01 1.03368223e+00
7.93934584e-01 3.35735083e-01 -6.96390510e-01 3.67715538e-01
7.03892887e-01 4.51657921e-01 9.00126755e-01 4.82167155e-01
7.71670461e-01 1.33366966e+00 1.04721963e+00 -1.71922326e-01
9.12714183e-01 2.15957850e-01 5.35929620e-01 -5.95907941e-02
-2.84346670e-01 2.02194244e-01 2.21527770e-01 2.57451802e-01
-2.59722233e-01 -3.80810022e-01 -1.01723897e+00 4.32305485e-01
-1.26623929e+00 -1.07961464e+00 -9.43889320e-01 2.13859224e+00
1.16175778e-01 -1.02528237e-01 8.10623765e-01 2.46833801e-01
1.01903033e+00 -2.44720414e-01 -9.41099465e-01 -8.96592677e-01
5.18745303e-01 -2.00344101e-01 7.78757989e-01 6.94972575e-01
-8.45656753e-01 7.49287605e-01 7.44228315e+00 -1.05317213e-01
-1.53960037e+00 3.09764147e-01 2.66051352e-01 -5.38128018e-01
3.14378232e-01 -5.03576458e-01 -1.47609842e+00 7.50222862e-01
1.56060433e+00 -1.61910877e-01 1.47869453e-01 7.66054332e-01
6.08265936e-01 -8.43622804e-01 -8.32472444e-01 1.28103542e+00
5.18556476e-01 -7.38478601e-01 -8.81181896e-01 4.53073531e-01
-5.53549603e-02 2.44873866e-01 6.50224149e-01 5.08375168e-01
-6.63667262e-01 -9.27262247e-01 5.35305798e-01 6.15313709e-01
9.73793089e-01 -8.92933726e-01 8.20267022e-01 1.38515323e-01
-6.71261787e-01 -3.86716545e-01 6.31233454e-02 -3.81505400e-01
2.88282692e-01 -4.20114428e-01 -8.69793177e-01 -2.76733428e-01
8.15492570e-01 6.10843837e-01 -8.55676472e-01 1.06870902e+00
-1.26089320e-01 4.36098188e-01 1.52199104e-01 -3.61122996e-01
2.08539162e-02 2.49687918e-02 3.73724729e-01 8.45080912e-01
-1.00529768e-01 -4.20174040e-02 -7.99176514e-01 7.18858600e-01
3.52741838e-01 -5.35004973e-01 -6.08000994e-01 6.68211639e-01
4.27356005e-01 1.08801842e+00 1.46761509e-02 1.87944189e-01
-6.01357281e-01 -5.85177056e-02 -1.01695299e-01 6.73079073e-01
-9.26545501e-01 -8.19679439e-01 1.44406378e+00 7.10667312e-01
-2.08735958e-01 4.26593162e-02 -4.48795676e-01 -5.31208664e-02
-7.45725259e-02 -3.78418684e-01 1.79096982e-01 -1.24707556e+00
-6.57787383e-01 4.47476923e-01 3.48418504e-01 -1.36863065e+00
-3.25902402e-01 -8.45836580e-01 -6.32417917e-01 1.61844969e+00
-2.41170621e+00 -8.82218480e-01 -1.15757024e+00 7.94291675e-01
7.21904457e-01 -3.43554646e-01 3.09772253e-01 3.52080613e-01
-1.07908905e+00 1.05991864e+00 -5.41068435e-01 -2.56697237e-01
1.09879160e+00 -1.08902681e+00 3.40165168e-01 5.80875695e-01
-9.38476443e-01 7.34615743e-01 8.11796784e-01 -1.16906539e-01
-1.18707383e+00 -4.50669736e-01 1.04154241e+00 -1.02366447e+00
1.78969935e-01 -1.96028337e-01 -9.41342056e-01 9.38533247e-01
9.11693335e-01 -3.64398241e-01 1.03423154e+00 5.01798630e-01
5.15052304e-02 -2.78885424e-01 -1.02511168e+00 7.33982265e-01
7.42499828e-01 -4.73426402e-01 -4.73885059e-01 7.17135370e-02
1.85407490e-01 -7.74595261e-01 -1.27336755e-01 -2.76826769e-01
8.91189039e-01 -1.15953457e+00 5.35938382e-01 -6.83163047e-01
-2.67056495e-01 -3.23538363e-01 7.07775056e-01 -9.92588639e-01
-4.60563868e-01 -8.18200350e-01 -1.81198195e-01 7.52242744e-01
1.35385051e-01 -1.30227542e+00 6.10927820e-01 1.14776468e+00
-3.16840738e-01 -3.96271288e-01 -3.32413256e-01 -4.02058661e-02
-2.93147445e-01 -4.05822217e-01 8.64723027e-01 4.62759107e-01
1.01371109e-01 5.60455322e-01 -1.30939528e-01 3.40388149e-01
5.31251013e-01 -4.55385625e-01 1.38523710e+00 -1.28863490e+00
1.06173670e+00 -7.75133967e-01 -8.03885400e-01 -1.10089207e+00
2.31196687e-01 1.59917429e-01 2.65375912e-01 -9.88351762e-01
-3.63309205e-01 -1.67807430e-01 -4.11955923e-01 1.32042887e-02
-6.77797914e-01 6.40074015e-02 -1.19453676e-01 2.98406810e-01
-1.51318118e-01 -1.12919234e-01 1.27692306e+00 3.26848269e-01
-3.66930574e-01 5.88799894e-01 -1.16166842e+00 3.49527061e-01
7.96159625e-01 -3.61919552e-01 -7.59239078e-01 -4.64223295e-01
6.45001531e-02 -4.03163284e-01 3.67390096e-01 -6.87223434e-01
7.78523743e-01 -3.41292918e-01 -1.76476300e-01 -1.09501171e+00
3.60237986e-01 -6.42295957e-01 -5.83193243e-01 4.46570851e-02
-3.15483391e-01 3.94370139e-01 6.28169775e-01 4.24952656e-01
-1.40282169e-01 -3.05570513e-01 8.67916048e-01 5.77997267e-01
-9.93140399e-01 8.71446058e-02 -5.72435796e-01 -1.61784649e-01
1.53764367e+00 -9.72392857e-01 -9.18763041e-01 -4.12927419e-01
-3.28820080e-01 5.09739757e-01 1.54801965e-01 1.02648890e+00
3.78642172e-01 -8.07153642e-01 -3.72595787e-01 8.60490203e-01
6.25404000e-01 -3.45257998e-01 4.92079556e-01 1.55096161e+00
-2.83264786e-01 9.97695804e-01 -5.78978121e-01 -1.16078150e+00
-1.78326929e+00 6.86743915e-01 4.57187802e-01 9.45141375e-01
1.09127667e-02 8.29860508e-01 6.04974747e-01 2.97942162e-01
1.39951602e-01 -3.43899071e-01 -1.00679803e+00 6.58459887e-02
1.19970131e+00 9.56081212e-01 4.33383346e-01 -1.42648125e+00
-4.40257728e-01 7.90788829e-01 -3.04887563e-01 3.75483543e-01
2.65963554e-01 -1.15876782e+00 2.24830598e-01 7.77654946e-01
1.04652071e+00 -1.50350571e-01 -9.58992064e-01 2.17136487e-01
-2.22993523e-01 -7.02599943e-01 4.71754789e-01 -6.14004254e-01
-8.86684775e-01 1.35623288e+00 1.29141414e+00 4.91229028e-01
1.11016703e+00 -4.85381745e-02 7.65286982e-01 1.02416240e-01
-1.16567716e-01 -8.57856870e-01 -6.85712039e-01 3.27584177e-01
6.03785753e-01 -1.96213365e+00 -3.53646398e-01 -4.34753209e-01
-9.56437230e-01 6.36209965e-01 8.69527578e-01 3.49226207e-01
1.17586601e+00 -1.28331050e-01 8.76067281e-01 -7.34500647e-01
-6.93289697e-01 -7.96208024e-01 8.07276011e-01 1.20365584e+00
5.24792671e-01 -9.61112827e-02 1.71676844e-01 -2.57108957e-01
-6.30180180e-01 9.89884958e-02 4.99694347e-01 1.11044526e+00
-5.96454382e-01 -5.39391279e-01 -7.07802415e-01 6.13994181e-01
-2.31071651e-01 1.90600976e-01 -3.49183559e-01 1.04299831e+00
1.05729349e-01 1.54182243e+00 4.64011669e-01 -3.42111200e-01
8.50642025e-01 3.45211267e-01 1.25946939e-01 -3.67080957e-01
-4.15658593e-01 -4.36378568e-01 -1.77229002e-01 -4.97158110e-01
-6.93815231e-01 -8.52177382e-01 -5.44678271e-01 -9.82423246e-01
-7.23183692e-01 -3.65738332e-01 9.80821013e-01 1.01087952e+00
7.08960593e-01 7.98391640e-01 4.56629753e-01 -8.56580496e-01
3.32179844e-01 -1.17400110e+00 -4.43274826e-01 -7.01584741e-02
1.26189756e+00 -1.52292395e+00 -5.48117697e-01 2.63806731e-01] | [14.075566291809082, 0.11102213710546494] |
d909c40d-ac27-4b76-a2ff-df2975eae3c7 | exploring-sentence-community-for-document | null | null | https://aclanthology.org/2021.findings-emnlp.32 | https://aclanthology.org/2021.findings-emnlp.32.pdf | Exploring Sentence Community for Document-Level Event Extraction | Document-level event extraction is critical to various natural language processing tasks for providing structured information. Existing approaches by sequential modeling neglect the complex logic structures for long texts. In this paper, we leverage the entity interactions and sentence interactions within long documents and transform each document into an undirected unweighted graph by exploiting the relationship between sentences. We introduce the Sentence Community to represent each event as a subgraph. Furthermore, our framework SCDEE maintains the ability to extract multiple events by sentence community detection using graph attention networks and alleviate the role overlapping issue by predicting arguments in terms of roles. Experiments demonstrate that our framework achieves competitive results over state-of-the-art methods on the large-scale document-level event extraction dataset. | ['Weijia Jia', 'Yusheng Huang'] | null | null | null | null | findings-emnlp-2021-11 | ['document-level-event-extraction'] | ['natural-language-processing'] | [ 3.71403813e-01 6.91415310e-01 -5.74904025e-01 -5.53057075e-01
-4.95696604e-01 -6.21388495e-01 8.72835696e-01 1.07423151e+00
-3.52689564e-01 9.14168239e-01 9.32952225e-01 -4.27438676e-01
-1.43814251e-01 -1.21749532e+00 -7.82136798e-01 -7.87591189e-02
-4.74093229e-01 3.35257202e-01 4.67629761e-01 -9.18368548e-02
-5.91934659e-02 2.46919230e-01 -7.87309051e-01 7.12976754e-01
7.07743943e-01 4.31979030e-01 -2.96755463e-01 5.52995801e-01
-5.13501465e-01 2.01185727e+00 -6.24794722e-01 -7.41540015e-01
-3.90992671e-01 -4.35564786e-01 -1.03294408e+00 8.78263712e-02
-8.94197002e-02 -2.09460437e-01 -7.66851604e-01 9.11384583e-01
9.20755565e-02 -3.64058129e-02 4.36596274e-01 -1.45281410e+00
-5.12464404e-01 1.70520246e+00 -8.16386163e-01 5.63141346e-01
6.42151117e-01 -4.80573505e-01 1.87202346e+00 -1.00287199e+00
9.25939143e-01 1.42385840e+00 3.94108713e-01 -1.33174360e-02
-1.15875256e+00 -4.57041949e-01 7.74629712e-01 3.99590909e-01
-8.22948694e-01 -1.22863665e-01 1.16378832e+00 -2.27728307e-01
1.67062807e+00 1.94466412e-01 6.96662247e-01 1.09111428e+00
3.36771607e-01 1.08408439e+00 3.67658675e-01 -3.77431333e-01
-5.38288616e-02 -4.79694486e-01 1.07403123e+00 8.18026900e-01
7.83484340e-01 -8.16948354e-01 -6.85763538e-01 -4.35135096e-01
1.91931784e-01 6.50303364e-02 -1.34199500e-01 3.24435160e-02
-1.30870998e+00 9.50906336e-01 2.83934861e-01 3.95774424e-01
-4.82830375e-01 4.21408594e-01 6.83446109e-01 1.45544633e-01
7.27220118e-01 3.21208298e-01 -5.72447419e-01 3.35742772e-01
-5.18030047e-01 2.26416990e-01 1.17090595e+00 9.92389560e-01
1.61065504e-01 -3.24854314e-01 -4.52323377e-01 3.83957505e-01
4.80445534e-01 -1.59646556e-01 2.82081757e-02 -5.83672643e-01
9.11999166e-01 1.13507330e+00 -1.32624045e-01 -1.40711200e+00
-4.95883942e-01 -3.78996581e-01 -7.96546459e-01 -7.09489703e-01
1.64212450e-01 -2.38235235e-01 -3.51691186e-01 1.68022573e+00
3.93031538e-01 1.06695265e-01 4.96067479e-02 2.76099086e-01
1.03791022e+00 6.81635857e-01 3.31865400e-01 -5.93331695e-01
1.47460556e+00 -1.06013107e+00 -1.14399004e+00 -5.13516545e-01
4.87370700e-01 -1.62715539e-01 4.76080894e-01 1.16642946e-02
-1.13622355e+00 9.12809074e-02 -1.09979022e+00 -2.48700082e-01
-2.04738721e-01 -1.59320697e-01 8.96852195e-01 -1.52631059e-01
-6.21456683e-01 2.03850031e-01 -8.52048874e-01 -8.32437947e-02
5.53768933e-01 2.11476967e-01 -2.39969999e-01 1.06872052e-01
-1.80295670e+00 5.73111951e-01 7.37837553e-01 -7.61884078e-02
-4.92509782e-01 -5.21082401e-01 -1.42684269e+00 4.48070526e-01
9.30019677e-01 -6.93087697e-01 1.21652937e+00 -3.03980500e-01
-7.27818727e-01 8.05855691e-01 -4.84358668e-01 -8.78181398e-01
8.23125616e-02 -2.01261669e-01 -4.85855311e-01 4.29288477e-01
3.67118448e-01 -4.76108640e-02 3.43079567e-01 -9.19251978e-01
-5.21202207e-01 -2.48806477e-01 5.60315132e-01 1.33386984e-01
-5.50482094e-01 6.70489430e-01 -2.85719782e-01 -6.54204726e-01
1.68715134e-01 -3.90402943e-01 -3.98420542e-01 -4.67458040e-01
-9.50246811e-01 -8.11162531e-01 6.95969045e-01 -5.02047777e-01
1.71608675e+00 -1.67334986e+00 2.17858970e-01 4.72955219e-02
8.84589672e-01 -5.26857376e-01 1.32588208e-01 8.24641943e-01
-2.69939542e-01 4.53044772e-01 -2.04448014e-01 -3.61028016e-01
1.29634395e-01 2.27040216e-01 -5.50835609e-01 3.36747795e-01
5.10561824e-01 1.06025076e+00 -1.22855508e+00 -1.10351861e+00
-2.93282300e-01 5.54242469e-02 -3.83193552e-01 -1.56413168e-02
-5.22476614e-01 -1.45367116e-01 -7.15830564e-01 5.30040503e-01
2.68658906e-01 -8.01445425e-01 8.38473856e-01 -2.09641412e-01
2.51280874e-01 9.03187752e-01 -1.04018295e+00 1.36536682e+00
-1.34025188e-02 4.44380194e-01 8.69008452e-02 -1.04644644e+00
3.63955081e-01 4.38475460e-01 6.12670243e-01 -1.08989906e-02
5.64830713e-02 -2.78255671e-01 1.07081145e-01 -3.28665376e-01
3.67462575e-01 5.47740310e-02 -5.69014311e-01 7.98798263e-01
1.56804353e-01 3.06798160e-01 9.28078115e-01 1.28864634e+00
1.43090141e+00 -3.35968405e-01 8.21007133e-01 -1.57886982e-01
5.55279672e-01 7.64483726e-03 7.56844819e-01 8.14503908e-01
7.34188408e-02 1.68518409e-01 1.32048178e+00 -2.61859298e-01
-6.16534531e-01 -9.15173650e-01 2.48413771e-01 1.08369839e+00
-6.79566562e-02 -1.10601473e+00 -3.22067916e-01 -1.24555540e+00
-1.25185112e-02 7.00292349e-01 -6.14306450e-01 1.27197877e-01
-8.76651049e-01 -1.01074278e+00 5.98224401e-01 8.32543612e-01
1.93820506e-01 -1.09649324e+00 -1.06359787e-01 6.83612585e-01
-5.54251790e-01 -1.57065821e+00 -4.85860407e-01 1.71792611e-01
-5.09483099e-01 -1.55226839e+00 2.74731249e-01 -9.10850644e-01
6.82686388e-01 2.06282847e-02 1.57529593e+00 1.30355775e-01
-2.24528946e-02 1.88940421e-01 -3.25412810e-01 -5.71678042e-01
-2.30497763e-01 -1.42761424e-01 -2.17880368e-01 4.59670611e-02
4.22745436e-01 -7.36300647e-01 -3.02200973e-01 -2.91029036e-01
-9.24783409e-01 8.84225219e-02 2.21654177e-01 7.38244176e-01
4.50653195e-01 4.66900706e-01 6.78820014e-01 -1.39658201e+00
8.94985616e-01 -7.56755054e-01 -2.94531941e-01 4.79797572e-01
-3.94482553e-01 1.18953519e-01 5.94581723e-01 -2.49546796e-01
-1.26873136e+00 -2.51436085e-01 2.96762019e-01 3.85193378e-01
1.84010848e-01 1.19574547e+00 -2.46170759e-01 7.67139733e-01
1.72794998e-01 -4.08828594e-02 -5.82603872e-01 5.49728721e-02
4.79119629e-01 2.15215251e-01 2.35280111e-01 -6.57166779e-01
6.47754550e-01 6.22622490e-01 1.15116285e-02 -5.31521678e-01
-1.50823998e+00 -4.46931571e-01 -5.42674482e-01 -1.38258263e-01
9.71316218e-01 -1.01328290e+00 -8.41631234e-01 1.44926012e-01
-1.76424217e+00 1.52209222e-01 -1.72097385e-01 3.01467746e-01
2.01194152e-01 4.94008124e-01 -1.34200060e+00 -9.01077509e-01
-4.59504128e-01 -3.93691331e-01 8.99565578e-01 -3.09540890e-02
-4.46799427e-01 -1.10246217e+00 1.41300023e-01 1.84048727e-01
-2.75710702e-01 5.16866565e-01 1.17593765e+00 -9.83447731e-01
-5.35705924e-01 -2.90858269e-01 -2.78406084e-01 -3.67542773e-01
7.20315725e-02 5.67997480e-03 -3.82724762e-01 1.02013774e-01
-6.55274391e-02 -2.57627308e-01 1.05982316e+00 2.23843336e-01
1.12595391e+00 -6.04666710e-01 -6.25175953e-01 -1.35211766e-01
1.04603398e+00 -4.47591096e-02 2.11794063e-01 1.01114064e-01
9.59521532e-01 7.05240965e-01 2.55611300e-01 4.77039218e-01
7.43502259e-01 1.53993862e-02 2.80446529e-01 -2.13066973e-02
1.65829752e-02 -5.00905037e-01 3.55073184e-01 8.53087664e-01
2.59764612e-01 -6.99771047e-01 -9.07083929e-01 6.46031916e-01
-2.33933425e+00 -1.43707740e+00 -7.80699313e-01 1.31104088e+00
1.09524393e+00 6.52481794e-01 -1.32809028e-01 2.05939516e-01
8.83484244e-01 5.97574651e-01 -2.33314827e-01 -1.39876172e-01
-3.92958820e-01 2.26586554e-02 2.40215123e-01 5.49236834e-01
-1.24531746e+00 8.10581863e-01 6.01460695e+00 3.80093813e-01
-1.17808379e-01 1.14291057e-01 4.91090357e-01 -1.13631234e-01
-6.03380084e-01 2.18982160e-01 -8.86082888e-01 2.05222502e-01
8.47758830e-01 -6.98189020e-01 -1.32437602e-01 5.65484285e-01
2.64941156e-01 1.97058190e-02 -1.18246317e+00 3.01548511e-01
2.38206282e-01 -1.65681267e+00 1.94651484e-01 -1.46251112e-01
5.85182369e-01 -3.41361202e-02 -8.23328853e-01 2.00966939e-01
8.92303884e-01 -7.23895073e-01 7.59031892e-01 1.67627648e-01
2.49589071e-01 -6.51490927e-01 6.38225853e-01 3.98734868e-01
-1.78802276e+00 -1.77338660e-01 4.84593771e-02 -3.97879779e-01
5.93694031e-01 1.04106569e+00 -8.75898123e-01 9.06354666e-01
3.75667930e-01 1.18654525e+00 -5.50542951e-01 4.00007844e-01
-8.32494318e-01 9.79378819e-01 -1.64982826e-01 -3.79939854e-01
1.56399444e-01 -2.56937891e-02 7.66609907e-01 1.39334655e+00
-3.66142452e-01 3.53369027e-01 4.82392937e-01 8.57327998e-01
-7.40581751e-01 -1.25409424e-01 -8.23369801e-01 -5.41474342e-01
4.41284567e-01 1.19752049e+00 -1.03594649e+00 -6.40215039e-01
-6.88359261e-01 6.94129288e-01 5.86292028e-01 3.11155200e-01
-9.36673284e-01 -4.94807661e-01 1.14875026e-01 8.98971129e-03
3.50877531e-02 -3.19916397e-01 -1.89403594e-01 -1.46557260e+00
1.96190178e-01 -5.32014012e-01 1.19106042e+00 -6.91021979e-01
-1.48554027e+00 4.82928157e-01 1.80291981e-01 -6.42059982e-01
-1.90407693e-01 -3.54738414e-01 -9.28130031e-01 3.16286087e-01
-1.21573365e+00 -1.34658980e+00 2.23276824e-01 6.43419504e-01
7.42609501e-01 2.94723421e-01 6.00918055e-01 1.69224009e-01
-6.74064696e-01 4.99043614e-02 -6.22080564e-01 7.73233175e-01
3.84943724e-01 -1.55696702e+00 4.97942716e-01 1.20044637e+00
4.06036794e-01 8.84149790e-01 5.05688846e-01 -1.13849890e+00
-1.40194786e+00 -1.15189624e+00 1.82052040e+00 -5.29034853e-01
1.38262224e+00 -6.68480456e-01 -7.58149207e-01 1.23433411e+00
5.79784513e-01 3.33952159e-02 8.40976179e-01 4.66167867e-01
-4.82725263e-01 2.27904007e-01 -7.01784253e-01 6.98214412e-01
1.37597930e+00 -7.00341284e-01 -1.26768661e+00 6.57298923e-01
1.37482262e+00 -1.03553928e-01 -6.83365703e-01 3.48183900e-01
6.13722876e-02 -2.89792597e-01 9.61288869e-01 -1.18618309e+00
8.59957159e-01 -2.49851018e-01 1.88094079e-01 -8.30445290e-01
-3.77542853e-01 -6.65578067e-01 -9.13726389e-01 1.47191048e+00
9.52193141e-01 -4.60954815e-01 5.26108503e-01 6.22668922e-01
7.97532424e-02 -6.31457686e-01 -4.28833187e-01 -2.77073383e-01
-3.67853165e-01 -5.65780342e-01 4.08917308e-01 1.16041505e+00
6.72985613e-01 1.19364381e+00 -1.76498786e-01 3.78783405e-01
9.02165711e-01 6.99683428e-01 3.35938297e-02 -1.41721547e+00
-5.19751310e-01 -1.90467641e-01 1.14431009e-01 -5.85388660e-01
8.08805168e-01 -1.20423174e+00 -2.55655438e-01 -2.11689878e+00
9.11813974e-01 4.41672504e-01 -3.96314412e-01 5.51318884e-01
-4.46089000e-01 -3.33433747e-01 -1.00145236e-01 -1.20676622e-01
-1.30484056e+00 4.91251737e-01 1.03014457e+00 -4.86162543e-01
-1.17851071e-01 -2.54197717e-01 -9.56570745e-01 1.12262046e+00
6.09436035e-01 -9.02074575e-01 -4.72995013e-01 -3.26398313e-01
8.57246935e-01 2.93377042e-01 1.60308182e-01 -2.19128251e-01
5.70965648e-01 -1.98731661e-01 2.33927280e-01 -7.76593626e-01
-1.69246495e-01 -5.31310678e-01 -2.55948305e-01 3.62910211e-01
-8.58462095e-01 1.22799248e-01 -7.97273219e-02 1.14077377e+00
-4.91882473e-01 -3.02544013e-02 -5.13115264e-02 -2.26221144e-01
-3.83971840e-01 2.75499195e-01 -4.22792077e-01 3.81219566e-01
9.33029473e-01 4.56634402e-01 -5.74692070e-01 -4.68348384e-01
-8.98717582e-01 6.94664359e-01 -1.91514894e-01 3.23319912e-01
7.14633703e-01 -1.19156516e+00 -9.46855307e-01 -3.42571139e-01
2.56661940e-02 2.26720601e-01 -5.48099838e-02 8.30769658e-01
-8.70012268e-02 2.46515721e-01 5.82255304e-01 1.32305156e-02
-1.55836427e+00 4.83732969e-01 -1.53146088e-01 -1.11408567e+00
-6.25262022e-01 8.54478359e-01 1.63754046e-01 -1.84532642e-01
1.23946585e-01 -4.70109105e-01 -6.27618909e-01 3.36776614e-01
5.90881109e-01 -6.23720884e-02 -1.51961565e-01 -2.69386321e-01
-6.92536771e-01 -2.10959017e-01 -2.52503753e-01 -1.25244036e-01
1.57206607e+00 1.55797764e-03 -9.13548708e-01 3.05513442e-01
8.61859620e-01 3.95689964e-01 -7.29675174e-01 -2.20298111e-01
5.95573545e-01 1.55786842e-01 -5.02262563e-02 -5.85015178e-01
-5.60016334e-01 3.68000180e-01 -9.03901219e-01 6.40618980e-01
7.82289326e-01 5.35971105e-01 7.70321727e-01 6.52794421e-01
1.61366016e-01 -8.33251774e-01 2.20541984e-01 7.69053459e-01
1.05532324e+00 -9.79229629e-01 3.09238225e-01 -1.02541876e+00
-5.90112448e-01 9.94273484e-01 5.16053915e-01 -5.54463454e-02
6.92283332e-01 6.49054825e-01 -7.08518744e-01 -6.08347893e-01
-1.05367577e+00 -1.23009972e-01 2.83535749e-01 -1.36653811e-01
6.98338807e-01 2.53713522e-02 -7.41989374e-01 1.23894835e+00
2.97980905e-01 -3.08401734e-01 7.12470651e-01 1.15532482e+00
-2.92123199e-01 -1.17091334e+00 1.80846378e-02 6.37411714e-01
-9.75400507e-01 -5.04869163e-01 -9.38327372e-01 6.26562953e-01
-1.93109766e-01 1.27762139e+00 -2.89329365e-02 -7.78895766e-02
2.76910454e-01 1.95244014e-01 3.53486776e-01 -9.66980636e-01
-7.83459306e-01 5.03699481e-02 9.77370262e-01 -4.32370424e-01
-5.52754760e-01 -6.25033557e-01 -1.86420989e+00 -3.79933566e-02
-2.33703613e-01 3.75855178e-01 -2.75439061e-02 1.04320228e+00
3.48721623e-01 9.99289811e-01 3.55702788e-01 -1.42750949e-01
-2.84981370e-01 -8.19915652e-01 -3.97227168e-01 5.54019511e-01
1.27818227e-01 -3.61869782e-01 -3.55583817e-01 1.38308421e-01] | [9.06202507019043, 9.082265853881836] |
cec12f63-4886-403b-8c8d-53a5dc580344 | learning-based-dequantization-for-image | 1803.01532 | null | http://arxiv.org/abs/1803.01532v2 | http://arxiv.org/pdf/1803.01532v2.pdf | Learning-Based Dequantization For Image Restoration Against Extremely Poor Illumination | All existing image enhancement methods, such as HDR tone mapping, cannot
recover A/D quantization losses due to insufficient or excessive lighting,
(underflow and overflow problems). The loss of image details due to A/D
quantization is complete and it cannot be recovered by traditional image
processing methods, but the modern data-driven machine learning approach offers
a much needed cure to the problem. In this work we propose a novel approach to
restore and enhance images acquired in low and uneven lighting. First, the ill
illumination is algorithmically compensated by emulating the effects of
artificial supplementary lighting. Then a DCNN trained using only synthetic
data recovers the missing detail caused by quantization. | ['Xiao Shu', 'Chang Liu', 'Xiaolin Wu'] | 2018-03-05 | null | null | null | null | ['tone-mapping'] | ['computer-vision'] | [ 8.67443085e-01 -2.13715270e-01 2.51443863e-01 -1.04067922e-01
-5.32124937e-01 -1.35612741e-01 3.15775156e-01 -2.14339226e-01
-1.20221950e-01 1.04622591e+00 2.73794532e-01 1.21955136e-02
1.48075446e-01 -8.54393899e-01 -5.85402012e-01 -8.92304897e-01
4.22600001e-01 -2.99203992e-01 2.37281825e-02 -5.15302122e-01
4.15087789e-01 3.45871866e-01 -2.04851151e+00 1.81080401e-01
1.11422992e+00 8.22739899e-01 3.07346404e-01 8.64825070e-01
-3.35592598e-01 9.69738245e-01 -6.70323670e-01 -9.29779336e-02
4.97355342e-01 -7.89099693e-01 -5.17710388e-01 4.21112865e-01
5.72964549e-01 -9.37033474e-01 -4.23354387e-01 1.37115788e+00
7.11376786e-01 -1.05384558e-01 3.42412680e-01 -1.02532554e+00
-9.03169394e-01 -2.12775201e-01 -8.35554719e-01 -6.86174259e-02
4.69806492e-01 1.49453700e-01 2.95040399e-01 -9.45944846e-01
5.03503263e-01 9.95815039e-01 5.76266110e-01 7.18680859e-01
-1.40221286e+00 -3.63048792e-01 -6.13579750e-01 1.96406782e-01
-1.22197211e+00 -5.77987671e-01 1.09342003e+00 2.48410869e-02
5.81008255e-01 4.62503791e-01 5.13806164e-01 4.71168071e-01
2.70109493e-02 4.05129492e-01 1.84937263e+00 -8.69175434e-01
1.78517312e-01 1.98699266e-01 -4.37904269e-01 3.65784824e-01
-3.51793878e-02 4.36616659e-01 -5.49009860e-01 2.29892164e-01
8.98845911e-01 2.28032283e-02 -5.40112019e-01 -4.18601232e-03
-8.12635541e-01 2.57151484e-01 3.53525609e-01 -3.18653099e-02
-4.32554007e-01 2.04061437e-02 9.54146981e-02 5.62578738e-01
5.41873932e-01 3.99994493e-01 -4.03170854e-01 1.61801800e-01
-1.29024875e+00 -2.95379087e-02 3.04640085e-01 7.80733883e-01
1.20298290e+00 5.10441422e-01 5.86443096e-02 8.79154682e-01
9.52324867e-02 6.36958480e-01 4.42108661e-01 -1.37063313e+00
7.64994174e-02 2.89466411e-01 3.02095801e-01 -8.96804929e-01
-7.77630252e-04 7.50970170e-02 -1.18546176e+00 9.27964032e-01
1.81064412e-01 4.58090799e-03 -1.09273839e+00 1.22497249e+00
2.17110440e-01 1.88773852e-02 2.14095056e-01 9.26277399e-01
6.85582995e-01 8.73710275e-01 -1.54109076e-01 -5.92753708e-01
1.01521635e+00 -7.33498752e-01 -1.52173090e+00 -5.51301762e-02
-2.44704690e-02 -1.17786920e+00 1.23497701e+00 9.25558209e-01
-1.45229948e+00 -7.14983821e-01 -1.35898936e+00 -5.44763565e-01
-3.28283995e-01 -8.25150609e-02 2.51629710e-01 6.82209492e-01
-1.32974720e+00 7.27322638e-01 -2.08440974e-01 1.27451802e-02
3.69863957e-01 6.97553232e-02 -2.38540486e-01 -5.04326344e-01
-1.17462349e+00 1.05599535e+00 2.37514153e-01 2.66703665e-01
-5.66215634e-01 -8.83224964e-01 -7.70611584e-01 -1.04140565e-01
1.06137291e-01 -2.93550998e-01 7.66971111e-01 -1.19079053e+00
-1.64322352e+00 8.43881667e-01 -3.05299789e-01 -1.95724145e-01
5.37485421e-01 -2.02014759e-01 -4.96381193e-01 3.53499323e-01
-2.71668553e-01 5.34069180e-01 1.42604351e+00 -1.67157996e+00
-3.01335216e-01 -3.19062859e-01 -3.90569329e-01 2.36378387e-01
-3.15855056e-01 -9.57653150e-02 -4.97181267e-02 -7.24072099e-01
1.17618695e-01 -3.30816269e-01 -1.26420498e-01 4.66788501e-01
-2.70949602e-01 7.80650437e-01 1.17273831e+00 -1.26703393e+00
1.05957901e+00 -2.07345772e+00 -2.95226336e-01 -2.69629415e-02
1.09474629e-01 4.45458382e-01 -7.25975335e-02 4.44411248e-01
-1.93977579e-01 -9.20143500e-02 -7.46086955e-01 -4.18526828e-01
-2.81038553e-01 3.32618833e-01 -5.35058200e-01 4.79250461e-01
2.41414160e-01 5.68168402e-01 -6.57320619e-01 -4.61231887e-01
7.52436399e-01 1.04592431e+00 -1.02840401e-01 2.56175280e-01
6.10221643e-03 3.73892337e-01 3.97802949e-01 7.99473286e-01
1.27343917e+00 2.77586222e-01 -1.57054737e-01 -4.41313177e-01
-3.69349808e-01 -1.59376279e-01 -1.21126080e+00 1.44902515e+00
-3.87390375e-01 9.15601611e-01 1.70751795e-01 -7.60287464e-01
1.17119086e+00 3.27676892e-01 2.74738550e-01 -1.32228923e+00
-3.67413536e-02 3.97195578e-01 -5.79977930e-01 -7.06749022e-01
7.92405307e-01 -4.79048967e-01 5.25662601e-01 3.59854728e-01
-1.82003260e-01 -3.89578730e-01 -1.04993835e-01 -7.87478779e-03
8.21948349e-01 2.66433269e-01 2.26567775e-01 -1.11304812e-01
5.20322084e-01 -2.30753683e-02 4.43609357e-01 3.48478764e-01
-2.59784341e-01 1.19452703e+00 7.88692012e-02 -4.98058438e-01
-1.75162232e+00 -7.75171280e-01 -1.82278991e-01 6.57030284e-01
3.94367546e-01 2.32783198e-01 -8.76429439e-01 5.93185723e-02
-3.44712824e-01 6.30979896e-01 -3.92675847e-01 -2.93974489e-01
-5.83130419e-01 -1.07193410e+00 2.29078427e-01 9.47839245e-02
1.07207727e+00 -1.25929427e+00 -5.32698750e-01 2.56941199e-01
-4.66044933e-01 -1.03155851e+00 -2.50806510e-01 4.96666133e-02
-9.67082083e-01 -8.76211941e-01 -8.86207461e-01 -5.94679654e-01
8.02877963e-01 5.70575595e-01 1.00430691e+00 4.22349304e-01
-7.89407253e-01 -2.24951096e-03 -2.76650041e-01 -2.09290385e-01
-5.37729621e-01 -7.54006684e-01 -1.76804185e-01 9.01887268e-02
8.43036100e-02 -6.56735003e-01 -8.55477750e-01 1.18567720e-01
-1.25627589e+00 1.80584311e-01 5.71561158e-01 8.16714346e-01
8.25075686e-01 6.92920148e-01 2.84573913e-01 -5.86077750e-01
4.95139390e-01 1.67284399e-01 -3.88976455e-01 1.66948736e-02
-7.87232041e-01 1.00577297e-02 6.95981562e-01 -3.34146440e-01
-1.56927896e+00 5.35950102e-02 -1.30308151e-01 -4.25234973e-01
-2.55031586e-01 -2.19314650e-01 -3.47825319e-01 -4.02214438e-01
6.92324519e-01 6.26964569e-01 -2.57895943e-02 -5.76819062e-01
3.43088448e-01 6.23827934e-01 8.79443288e-01 6.67314231e-02
1.00490463e+00 8.90280247e-01 2.57374018e-01 -9.71722543e-01
-5.39325237e-01 -7.84970447e-02 -4.26795512e-01 -4.71187592e-01
7.80000687e-01 -8.35387349e-01 -2.32579410e-01 7.58844495e-01
-9.72651422e-01 -4.67691600e-01 -7.42666602e-01 2.02875897e-01
-6.71688318e-01 5.21312296e-01 -5.75415432e-01 -8.80546033e-01
-3.67239326e-01 -7.98871160e-01 6.66448236e-01 4.36675519e-01
3.57170224e-01 -7.07805455e-01 4.59626019e-02 2.26169944e-01
8.25479031e-01 3.61912102e-01 8.40693057e-01 5.89266300e-01
-5.45151949e-01 7.93695152e-02 -4.55580950e-01 8.75363410e-01
4.78145987e-01 1.30996317e-01 -1.21066189e+00 -1.23473294e-01
3.74621391e-01 -2.30695054e-01 8.31535220e-01 5.56071520e-01
1.17231524e+00 -3.46396536e-01 3.70210409e-01 8.64839435e-01
1.98609543e+00 1.90990139e-02 1.48057556e+00 4.61251438e-01
3.59494984e-01 4.95381355e-01 6.25331521e-01 3.44272077e-01
4.03753184e-02 3.19614202e-01 7.03078389e-01 -8.86124730e-01
-9.10981596e-01 -1.87931255e-01 1.69816121e-01 5.06329834e-01
-6.34734482e-02 -3.25821303e-02 -4.05628383e-01 4.56518263e-01
-1.13156164e+00 -9.16540742e-01 -5.25392473e-01 2.19105864e+00
1.19519901e+00 -1.27694741e-01 -3.80088776e-01 7.89038479e-01
9.92659867e-01 4.34523337e-02 -5.34927905e-01 -6.49842978e-01
-7.15688705e-01 2.82940865e-01 7.12640524e-01 7.59974003e-01
-8.38757157e-01 6.14684463e-01 7.29768562e+00 7.21745789e-01
-1.14054656e+00 -1.28651828e-01 8.08350146e-01 9.69964787e-02
-3.22898328e-01 -5.54379113e-02 -2.16333792e-01 4.82246011e-01
5.25941789e-01 2.30278537e-01 8.37318063e-01 4.53965008e-01
4.11061287e-01 -4.07281280e-01 -3.82896811e-01 1.05443919e+00
3.47054750e-01 -9.59722519e-01 -3.78172882e-02 2.26313993e-02
1.00014400e+00 -5.51622033e-01 2.76229441e-01 -2.73170710e-01
-9.68729611e-03 -7.47839093e-01 3.90491426e-01 7.25468516e-01
1.19344711e+00 -6.96837842e-01 5.24885178e-01 1.94497019e-01
-7.31298685e-01 -6.48584962e-03 -5.66212833e-01 -2.24812433e-01
2.96430528e-01 8.80129337e-01 -2.91377515e-01 3.32544327e-01
8.42542350e-01 4.69445348e-01 -5.33194005e-01 9.12799299e-01
-3.91032755e-01 3.41309369e-01 -5.56609631e-02 7.43736982e-01
-3.38630915e-01 -3.79972637e-01 4.38556671e-01 9.76837039e-01
3.87617350e-01 2.46681094e-01 -6.68712080e-01 8.27347279e-01
-1.33520201e-01 -2.14497879e-01 -5.40616870e-01 3.30731988e-01
1.49762109e-01 1.28364766e+00 -4.91514772e-01 -3.50562781e-01
-3.57465744e-01 1.52504456e+00 -3.37556958e-01 6.75028741e-01
-5.54722488e-01 -6.82868600e-01 2.56726503e-01 3.61607879e-01
6.01750687e-02 7.55042434e-02 -4.62988138e-01 -9.63071942e-01
2.06017807e-01 -7.08553016e-01 -2.07883999e-01 -1.47087550e+00
-1.13850510e+00 3.74874055e-01 -5.09152234e-01 -1.33592808e+00
1.23221114e-01 -3.14579785e-01 -5.68457007e-01 1.00457180e+00
-2.27630734e+00 -8.15153003e-01 -6.42846048e-01 7.72790194e-01
4.90984738e-01 1.77483350e-01 6.73269749e-01 6.80423379e-01
-9.21320915e-02 1.46709248e-01 1.89908117e-01 -3.06723148e-01
1.05160451e+00 -1.37524080e+00 -7.32741058e-02 1.04056549e+00
-5.96963406e-01 -5.70929460e-02 1.05437708e+00 -4.54621702e-01
-1.34407854e+00 -1.03970623e+00 8.80573273e-01 2.38889784e-01
1.11312084e-01 2.80507412e-02 -1.31978428e+00 2.54284531e-01
5.22821605e-01 1.76048338e-01 3.95041317e-01 -8.45238209e-01
-2.34121367e-01 -3.92119139e-01 -1.72620988e+00 3.59575093e-01
4.99011308e-01 -5.70138454e-01 -3.34254831e-01 1.07423306e-01
7.25510538e-01 -4.62231040e-02 -7.26355314e-01 2.83827394e-01
3.39128435e-01 -1.28747916e+00 1.12860870e+00 3.75875644e-03
8.52700174e-01 -5.61771393e-01 -2.67244250e-01 -1.29701293e+00
5.04891090e-02 -7.63814211e-01 -7.04767108e-02 1.35280967e+00
-5.54705225e-02 -4.22251582e-01 5.45060813e-01 5.15364349e-01
-8.16680938e-02 -3.80654186e-02 -7.07033813e-01 -3.32684189e-01
-2.41981670e-01 7.85079412e-03 6.51566684e-01 1.23108411e+00
-3.94912601e-01 -2.75887609e-01 -9.37084556e-01 -5.08268178e-02
1.16937184e+00 -1.88403577e-01 4.55690056e-01 -8.02313864e-01
8.93171355e-02 -6.64010346e-02 -6.98365420e-02 -6.82005346e-01
-5.62188804e-01 -1.62071839e-01 3.10802907e-01 -1.59601009e+00
-3.25853936e-02 -3.23082626e-01 -7.83413202e-02 2.33297572e-01
-3.06261748e-01 9.75953937e-01 1.97587758e-02 3.40640783e-01
-1.10565126e-01 6.75842762e-01 1.65327811e+00 5.57747111e-02
-1.57313764e-01 -5.38726687e-01 -6.27276480e-01 7.26471782e-01
9.25493002e-01 -2.68926948e-01 -2.70465076e-01 -5.80851376e-01
4.55013037e-01 1.35113165e-01 6.13128960e-01 -9.74259973e-01
1.03696875e-01 -2.44059172e-02 8.08664501e-01 -5.93555093e-01
3.23355943e-01 -9.19160843e-01 1.17389500e-01 2.87583619e-01
-2.36236602e-01 -2.78644502e-01 6.62779510e-02 3.06472242e-01
-5.09707570e-01 -7.97929317e-02 1.22521925e+00 -2.95705646e-01
-5.40069103e-01 -1.74644198e-02 -3.35952848e-01 -3.73640448e-01
7.30252981e-01 -5.83916247e-01 -4.61126983e-01 -4.31397498e-01
-6.53065860e-01 -3.82993966e-01 7.21733510e-01 -6.43979087e-02
9.62952137e-01 -1.39496124e+00 -7.63359070e-01 5.16530871e-01
-4.19869214e-01 -1.25503451e-01 6.08912051e-01 6.28451765e-01
-8.86244476e-01 -3.18584293e-01 -5.25067389e-01 -1.49223104e-01
-1.21291399e+00 7.63538599e-01 3.79012614e-01 6.90047070e-02
-8.37936163e-01 3.82818222e-01 -1.35390639e-01 -1.76760256e-01
1.84841640e-02 1.70856267e-01 -2.43328083e-02 -1.34086609e-01
9.61749375e-01 7.07115829e-01 2.58623213e-01 -5.31730771e-01
2.18904376e-01 7.48499691e-01 3.15716624e-01 -8.89289975e-02
1.32708645e+00 -7.19972908e-01 -3.47963125e-01 1.24520555e-01
1.16697490e+00 -2.13340431e-01 -1.44970775e+00 -2.00584784e-01
-7.18335509e-01 -8.76901865e-01 5.08489311e-01 -8.65704417e-01
-1.21209145e+00 1.19494760e+00 1.04036999e+00 4.76893365e-01
2.03592324e+00 -6.09849870e-01 9.75096524e-01 -1.87327608e-01
-1.34474158e-01 -1.42302346e+00 8.27987045e-02 -1.22618645e-01
7.45209634e-01 -1.25064445e+00 2.48156324e-01 -3.10172647e-01
-4.91018862e-01 1.16210532e+00 2.52753407e-01 -2.47289523e-01
4.82193440e-01 5.30432642e-01 1.70632496e-01 8.92469361e-02
-4.25842404e-01 -1.95855200e-01 -2.49197617e-01 9.47261572e-01
-1.86922401e-03 -2.93342024e-01 -2.68512070e-01 -2.08121672e-01
5.98637108e-03 3.17921400e-01 9.91617084e-01 9.66751575e-01
-7.99142897e-01 -7.88805962e-01 -8.88824105e-01 1.58775225e-01
-4.49992329e-01 -2.19356656e-01 -5.91777638e-02 5.75629175e-01
3.79898071e-01 9.20064867e-01 -2.61929375e-03 -1.83044806e-01
3.55396628e-01 -1.83824763e-01 5.40539920e-01 1.89440727e-01
-1.63994268e-01 3.10437381e-01 -5.10332763e-01 -5.65789580e-01
-7.77779877e-01 -3.15281779e-01 -1.11182332e+00 -5.34912646e-01
-2.97028303e-01 -5.16729772e-01 9.83402133e-01 5.13884068e-01
-3.46593447e-02 6.83307648e-01 9.46418405e-01 -8.07010412e-01
-1.94829732e-01 -7.40340412e-01 -9.18552399e-01 6.28086209e-01
8.99102628e-01 -2.65537679e-01 -7.92627692e-01 6.82069600e-01] | [10.920714378356934, -2.284127950668335] |
70f4060c-cad8-46cb-a413-4f78ba1b2fe7 | margin-optimal-classification-trees | 2210.10567 | null | https://arxiv.org/abs/2210.10567v4 | https://arxiv.org/pdf/2210.10567v4.pdf | Margin Optimal Classification Trees | In recent years there has been growing attention to interpretable machine learning models which can give explanatory insights on their behavior. Thanks to their interpretability, decision trees have been intensively studied for classification tasks, and due to the remarkable advances in mixed-integer programming (MIP), various approaches have been proposed to formulate the problem of training an Optimal Classification Tree (OCT) as a MIP model. We present a novel mixed-integer quadratic formulation for the OCT problem, which exploits the generalization capabilities of Support Vector Machines for binary classification. Our model, denoted as Margin Optimal Classification Tree (MARGOT), encompasses the use of maximum margin multivariate hyperplanes nested in a binary tree structure. To enhance the interpretability of our approach, we analyse two alternative versions of MARGOT, which include feature selection constraints inducing local sparsity of the hyperplanes. First, MARGOT has been tested on non-linearly separable synthetic datasets in 2-dimensional feature space to provide a graphical representation of the maximum margin approach. Finally, the proposed models have been tested on benchmark datasets from the UCI repository. The MARGOT formulation turns out to be easier to solve than other OCT approaches, and the generated tree better generalizes on new observations. The two interpretable versions are effective in selecting the most relevant features and maintaining good prediction quality. | ['Laura Palagi', 'Marta Monaci', 'Giorgio Grani', "Federico D'Onofrio"] | 2022-10-19 | null | null | null | null | ['interpretable-machine-learning'] | ['methodology'] | [ 4.95484799e-01 4.97319490e-01 -7.51927972e-01 -5.77628076e-01
-3.52942288e-01 -1.67472586e-01 3.15512925e-01 2.91872293e-01
-4.08236831e-02 1.01735961e+00 -1.45685300e-01 -6.12686574e-01
-9.06823397e-01 -4.40811515e-01 -4.00309533e-01 -7.94227540e-01
-3.17138225e-01 3.98849696e-01 -4.00839776e-01 1.55268162e-01
3.76966029e-01 4.78614867e-01 -1.49906981e+00 3.39457542e-01
1.47930396e+00 1.21379364e+00 2.13674046e-02 2.81264722e-01
1.38447896e-01 3.73366326e-01 -2.44654506e-01 -3.74203831e-01
3.32894504e-01 -7.30714947e-02 -9.79765058e-01 5.90789735e-01
1.36286244e-01 2.20521823e-01 1.41293600e-01 6.61362112e-01
-1.56127959e-01 4.20162007e-02 8.85676920e-01 -1.75220048e+00
-7.77694762e-01 4.11940783e-01 -4.99215126e-01 -4.99643050e-02
1.14658445e-01 -1.92521989e-01 1.36025083e+00 -8.08408618e-01
3.49750847e-01 1.27027535e+00 3.92615885e-01 1.32092729e-01
-1.64487123e+00 -2.03139380e-01 1.63096994e-01 4.99275059e-01
-1.13933766e+00 4.18738499e-02 7.94038475e-01 -5.88712394e-01
4.98380125e-01 7.12848425e-01 5.90889931e-01 7.54975736e-01
4.88695204e-01 1.04515827e+00 1.40845549e+00 -6.61339641e-01
2.81060010e-01 5.25595129e-01 7.25522935e-01 6.35341227e-01
5.40358305e-01 2.05375571e-02 -3.08076352e-01 -2.17074916e-01
3.00946385e-01 3.05302106e-02 -4.13661987e-01 -7.69116282e-01
-1.05103922e+00 1.14666021e+00 6.74733877e-01 1.18460163e-01
-3.25663060e-01 -4.17499870e-01 2.28270620e-01 1.41202047e-01
7.04674125e-01 7.66529441e-01 -5.44537544e-01 4.03767794e-01
-5.73674202e-01 3.38840574e-01 7.78105855e-01 5.77125609e-01
6.71348512e-01 -1.56879559e-01 -1.88288227e-01 8.46908987e-01
4.24899995e-01 1.10413678e-01 5.40578663e-01 -6.10819399e-01
8.90883565e-01 1.02869821e+00 -6.66642096e-03 -1.11965823e+00
-6.13581121e-01 -8.71713161e-01 -1.03476179e+00 2.77072012e-01
3.52293581e-01 5.73689230e-02 -7.37536311e-01 1.25535131e+00
3.00288409e-01 -2.06665605e-01 2.94416368e-01 8.73096347e-01
3.53521556e-01 7.21248090e-01 -8.06446597e-02 -2.78299004e-01
1.12418485e+00 -1.09717798e+00 -5.58842897e-01 -2.55144179e-01
9.38431799e-01 -2.19927117e-01 8.08891535e-01 6.52800560e-01
-4.30962026e-01 -4.12316769e-01 -1.27912033e+00 2.30729461e-01
-3.29719394e-01 3.71509194e-01 9.11054432e-01 7.28497684e-01
-6.93976521e-01 7.62124956e-01 -5.81219912e-01 8.06245655e-02
4.76999253e-01 6.59299910e-01 -5.80121160e-01 -1.14844091e-01
-9.01209235e-01 8.41727436e-01 9.15703714e-01 5.49533427e-01
8.55105221e-02 -2.36803457e-01 -1.07458127e+00 1.25915632e-01
4.35351282e-01 -5.39309561e-01 7.29454696e-01 -1.17408526e+00
-1.22724259e+00 5.97224891e-01 -4.12989825e-01 -5.42868555e-01
3.81622672e-01 -1.73050538e-02 -5.66390045e-02 -1.74372271e-01
1.86250389e-01 4.38522547e-01 8.55000257e-01 -1.13423121e+00
-4.78348434e-01 -7.10299432e-01 1.99960433e-02 7.69625604e-03
-4.68323380e-01 -2.95666277e-01 2.81525582e-01 -7.15730011e-01
4.59606946e-01 -1.17275953e+00 -4.55713570e-01 -2.04525992e-01
-7.68520772e-01 -2.63547808e-01 9.55537677e-01 -6.62000895e-01
1.52370155e+00 -1.95985925e+00 6.01443470e-01 3.80056083e-01
1.26542106e-01 1.32605061e-01 4.48982939e-02 9.24460888e-02
-5.43854713e-01 3.85778010e-01 -5.62452674e-01 -3.23988825e-01
-3.69763710e-02 5.60988188e-01 -1.93144143e-01 3.02779168e-01
5.64111471e-01 6.62946641e-01 -4.76970792e-01 -4.03159589e-01
3.05651259e-02 -6.86597452e-02 -5.50821602e-01 2.34533735e-02
-1.43610924e-01 4.96197283e-01 -6.92453086e-01 6.81660414e-01
7.65764415e-01 -2.90571630e-01 2.07370862e-01 2.34542251e-01
-7.26185665e-02 -1.54782474e-01 -1.37731647e+00 9.18336689e-01
-4.35009956e-01 6.65454149e-01 -2.80896306e-01 -1.57654095e+00
1.14188731e+00 1.16514593e-01 3.99610311e-01 -3.97894025e-01
-1.89836308e-01 2.14173764e-01 6.55476525e-02 -5.04466176e-01
2.37438515e-01 -8.25475752e-02 -2.96112634e-02 3.36857215e-02
-3.05209547e-01 8.94678235e-02 2.48202682e-01 -4.72745806e-01
5.61293900e-01 -1.13198040e-02 7.75737822e-01 -4.85544086e-01
9.29120541e-01 1.18703783e-01 6.29054248e-01 4.60994154e-01
2.40152270e-01 3.72203201e-01 8.08015406e-01 -8.34954023e-01
-7.31065094e-01 -6.78631902e-01 -7.08369136e-01 6.08791590e-01
1.28680663e-02 -2.62419224e-01 -5.23098171e-01 -8.12040687e-01
1.45474195e-01 8.75036359e-01 -6.65812373e-01 2.03712331e-03
-4.95892882e-01 -1.18908441e+00 2.94625219e-02 4.65294719e-01
6.63600743e-01 -7.11215913e-01 -5.64975917e-01 1.02485858e-01
-1.17376275e-01 -9.71526384e-01 6.97533786e-02 6.48404121e-01
-1.34434652e+00 -1.09739685e+00 -4.45955068e-01 -5.35725653e-01
7.86039233e-01 1.51110947e-01 7.14577019e-01 4.04398702e-02
-3.26156378e-01 -1.10114351e-01 -5.04918635e-01 -4.23366278e-01
-2.43774563e-01 3.72567475e-01 -4.65536378e-02 5.32491148e-01
-1.09471992e-01 -1.82752654e-01 -1.30041629e-01 6.05601549e-01
-1.03207338e+00 4.76409763e-01 8.01630974e-01 1.40010881e+00
4.94602472e-01 2.80431211e-01 6.30640745e-01 -9.75662172e-01
5.11924684e-01 -6.48164690e-01 -6.45111203e-01 3.28652322e-01
-9.99235988e-01 7.83733487e-01 8.80785346e-01 -1.98795155e-01
-1.01538014e+00 1.81129098e-01 2.58032620e-01 -6.34185830e-03
-1.00528531e-01 8.15204382e-01 -3.41103494e-01 -1.65683821e-01
5.03116012e-01 6.13230690e-02 -1.27779439e-01 -4.11199927e-01
7.38514736e-02 9.34972882e-01 -1.62086159e-01 -5.89294791e-01
7.51895547e-01 1.30916283e-01 5.85048556e-01 -8.46955895e-01
-1.07892799e+00 -3.91857862e-01 -1.07364190e+00 1.12055831e-01
7.64721274e-01 -2.18908146e-01 -4.23046321e-01 2.10437983e-01
-1.03100681e+00 3.23054790e-02 -1.61789097e-02 4.65646088e-01
-6.22825503e-01 3.43430370e-01 1.07785258e-02 -9.33456063e-01
-1.32281050e-01 -1.28269851e+00 1.09957016e+00 1.53850436e-01
-2.86812901e-01 -1.27946889e+00 -1.82016194e-01 6.90438807e-01
-1.40004471e-01 6.63255632e-01 1.52247822e+00 -7.39655733e-01
-6.84708774e-01 -2.65662819e-01 -2.19072312e-01 3.99823308e-01
6.12349473e-02 -3.02748680e-02 -8.61072600e-01 -1.82630941e-01
9.58712250e-02 -1.32306730e-02 7.61744142e-01 4.29240108e-01
1.36297214e+00 -9.09223616e-01 -5.63043475e-01 6.70622945e-01
1.40259027e+00 3.73158276e-01 3.30070108e-01 7.92630970e-01
3.68698955e-01 8.48998725e-01 9.81313884e-01 4.09325480e-01
1.76016912e-01 1.00359499e+00 6.54115438e-01 -2.22376794e-01
7.38096237e-01 1.32253021e-01 7.46189803e-02 2.58838654e-01
-2.66773820e-01 -2.14132801e-01 -7.96693981e-01 1.28183484e-01
-2.12348080e+00 -6.76168084e-01 -4.49057162e-01 2.30886459e+00
3.56688887e-01 1.01078898e-01 1.42748235e-02 7.45997727e-01
5.49825251e-01 1.53031871e-01 -4.43152428e-01 -9.49543417e-01
-1.63441196e-01 1.50522724e-01 3.02774876e-01 4.79306757e-01
-1.42431474e+00 3.41528118e-01 5.56617498e+00 6.80850685e-01
-1.10725403e+00 -4.26263750e-01 1.07473695e+00 3.89869541e-01
-1.28106683e-01 1.61581889e-01 -9.12688553e-01 3.11242878e-01
6.73071325e-01 -2.99286187e-01 1.30258113e-01 1.11379099e+00
3.96417201e-01 -2.12027729e-01 -1.39858246e+00 7.81844139e-01
-9.63304341e-02 -1.09553277e+00 8.85640010e-02 3.84277880e-01
6.43175304e-01 -7.77444899e-01 3.34132761e-01 2.85777837e-01
-2.85160035e-01 -1.33337724e+00 5.54426908e-01 2.53858536e-01
4.01031971e-01 -6.51921511e-01 9.82377231e-01 7.09001422e-01
-8.95274878e-01 -5.82457602e-01 -3.34741116e-01 -3.52739006e-01
-2.54284859e-01 7.35968173e-01 -1.16522884e+00 1.00014687e+00
3.49150389e-01 7.45730400e-01 -9.54829752e-01 1.16565645e+00
-1.11461535e-01 4.30827260e-01 -2.38774300e-01 -1.77830949e-01
3.85548204e-01 -4.77106452e-01 6.24657035e-01 8.66034508e-01
7.73700103e-02 -1.66203141e-01 1.69869497e-01 8.18372965e-01
4.24516946e-01 4.75502908e-01 -5.59355140e-01 1.68165684e-01
-1.69769958e-01 1.06278718e+00 -5.96668780e-01 8.62153023e-02
-1.92955896e-01 7.44712830e-01 4.52958435e-01 2.51162112e-01
-7.63484299e-01 -9.38354060e-02 3.29574138e-01 -5.02736047e-02
7.68037811e-02 -1.70478582e-01 -8.37274015e-01 -1.19429219e+00
4.70585346e-01 -7.47932613e-01 5.21018863e-01 -5.02350807e-01
-9.24218953e-01 8.08359087e-01 3.50352317e-01 -1.31259871e+00
-3.43633085e-01 -8.76294374e-01 -3.56432438e-01 7.38178134e-01
-1.37132251e+00 -1.04074419e+00 -1.59605816e-01 7.11117089e-02
6.75583303e-01 1.84505537e-01 8.39076221e-01 -8.60763788e-02
-9.90052998e-01 4.41654533e-01 4.21045691e-01 -3.46393019e-01
1.75069403e-02 -1.40383077e+00 -1.70456141e-01 6.38157308e-01
1.13720141e-01 3.54139715e-01 6.86262369e-01 -3.02091181e-01
-1.15393758e+00 -1.20139122e+00 9.56019878e-01 -1.70086488e-01
4.61854666e-01 -1.99486181e-01 -8.97245824e-01 8.05423141e-01
-2.17805177e-01 -2.20311806e-01 8.19222152e-01 2.89132327e-01
4.88401204e-02 -1.96149975e-01 -1.11704290e+00 5.22325933e-01
6.11018479e-01 3.67955089e-01 -6.79578900e-01 5.10575414e-01
3.43099147e-01 -1.21842735e-01 -8.53788018e-01 6.38256907e-01
4.29807663e-01 -9.53015864e-01 7.82422900e-01 -8.75236332e-01
7.05577374e-01 -1.42692417e-01 -3.51296534e-04 -1.47732770e+00
-3.99764240e-01 -2.32094079e-01 9.72357243e-02 1.00195503e+00
6.19620800e-01 -9.88151014e-01 8.79473388e-01 8.38165641e-01
8.06333944e-02 -1.52580225e+00 -1.15337920e+00 -9.62007701e-01
-1.60910949e-01 -2.27769256e-01 3.53094578e-01 7.28043973e-01
1.87107116e-01 3.76927108e-01 -4.23748910e-01 1.09665953e-01
6.38918102e-01 4.92228776e-01 7.20955193e-01 -1.60850859e+00
-4.85323727e-01 -3.89396220e-01 -7.06712306e-01 -9.46283937e-01
3.44854891e-01 -1.13231814e+00 -3.62659514e-01 -1.17845917e+00
1.09938234e-01 -8.33039403e-01 7.89671987e-02 5.99803805e-01
-3.68073732e-01 -1.43404737e-01 3.27445656e-01 2.00892165e-01
-1.73434466e-01 8.70723784e-01 1.09379601e+00 -2.77269542e-01
-4.56959039e-01 7.72593021e-01 -7.05721319e-01 8.33459496e-01
7.42860198e-01 -4.36422884e-01 -3.06524575e-01 -8.84058103e-02
-8.51546042e-03 3.53903711e-01 3.04032385e-01 -8.21205616e-01
-2.58456647e-01 -4.07682389e-01 3.82855296e-01 -4.50540334e-01
3.29730719e-01 -1.03091812e+00 2.87386358e-01 5.48236430e-01
-4.25734371e-01 -1.03427403e-01 9.84879211e-02 4.99636412e-01
-5.20462453e-01 -6.57507479e-01 6.49100542e-01 3.17264646e-01
-5.70063114e-01 8.31294060e-02 -2.52970606e-01 -3.22827965e-01
1.41119397e+00 -6.11649454e-01 1.19088300e-01 -1.84640028e-02
-9.47948337e-01 3.89856696e-01 3.89512219e-02 3.62693876e-01
5.59116781e-01 -1.11351144e+00 -6.21858895e-01 2.11822510e-01
2.27734789e-01 1.04237422e-01 -6.73646778e-02 8.30287218e-01
-4.83702213e-01 1.09003782e+00 -1.40210077e-01 -8.88765156e-01
-1.41234541e+00 6.26682460e-01 1.86256409e-01 -7.94972181e-01
-4.36733544e-01 3.68106097e-01 2.93141454e-01 -4.74402219e-01
-2.08997093e-02 -6.03284240e-01 -5.44934988e-01 -2.03478467e-02
8.37194920e-02 5.78979731e-01 6.35850728e-02 -4.78413969e-01
-2.04326317e-01 5.18594205e-01 8.18326548e-02 2.17763290e-01
1.44321227e+00 7.49789700e-02 -9.43445936e-02 4.04135197e-01
1.12775552e+00 -2.92508870e-01 -8.93555999e-01 -3.33120041e-02
5.17263532e-01 -6.92845941e-01 -1.92395583e-01 -4.94997978e-01
-7.86324561e-01 6.97024763e-01 1.53364077e-01 5.42707205e-01
1.32908428e+00 -2.83577800e-01 2.53758281e-01 4.26787794e-01
3.00324142e-01 -5.97105086e-01 -3.11371475e-01 2.90683538e-01
1.17112219e+00 -1.38185287e+00 1.86607569e-01 -1.05612075e+00
-6.81827605e-01 1.64588189e+00 4.84391779e-01 9.01477635e-02
4.80035663e-01 -1.80854470e-01 -3.93158317e-01 2.39727274e-01
-6.15039051e-01 8.60079899e-02 6.19976163e-01 4.43651736e-01
1.86902210e-01 2.97429472e-01 -6.05453193e-01 7.11436927e-01
-3.51385295e-01 -4.17296253e-02 3.53780329e-01 5.84321022e-01
-2.50546157e-01 -1.19590938e+00 -6.89274490e-01 5.89843214e-01
-2.36252785e-01 6.74902946e-02 -4.24099207e-01 1.00756741e+00
8.20034444e-02 9.57186639e-01 -3.76440108e-01 -2.48330504e-01
2.65002489e-01 9.85127315e-02 1.51147336e-01 -6.35710120e-01
-2.69462615e-01 -4.52874124e-01 1.05962835e-01 -2.20872685e-01
-2.61932492e-01 -6.50796354e-01 -8.81038129e-01 1.63846210e-01
-6.85383439e-01 3.79397571e-01 6.41351163e-01 1.31350255e+00
2.03478679e-01 4.24201697e-01 9.09471929e-01 -8.36880505e-01
-7.64564693e-01 -7.23689258e-01 -6.49784863e-01 2.65891254e-01
2.00036794e-01 -8.85622382e-01 -3.95778030e-01 -1.22892223e-01] | [8.463187217712402, 4.176764011383057] |
1d775824-42df-4550-9ec6-328c401d4c46 | aerial-scene-parsing-from-tile-level-scene | 2201.01953 | null | https://arxiv.org/abs/2201.01953v2 | https://arxiv.org/pdf/2201.01953v2.pdf | Aerial Scene Parsing: From Tile-level Scene Classification to Pixel-wise Semantic Labeling | Given an aerial image, aerial scene parsing (ASP) targets to interpret the semantic structure of the image content, e.g., by assigning a semantic label to every pixel of the image. With the popularization of data-driven methods, the past decades have witnessed promising progress on ASP by approaching the problem with the schemes of tile-level scene classification or segmentation-based image analysis, when using high-resolution aerial images. However, the former scheme often produces results with tile-wise boundaries, while the latter one needs to handle the complex modeling process from pixels to semantics, which often requires large-scale and well-annotated image samples with pixel-wise semantic labels. In this paper, we address these issues in ASP, with perspectives from tile-level scene classification to pixel-wise semantic labeling. Specifically, we first revisit aerial image interpretation by a literature review. We then present a large-scale scene classification dataset that contains one million aerial images termed Million-AID. With the presented dataset, we also report benchmarking experiments using classical convolutional neural networks (CNNs). Finally, we perform ASP by unifying the tile-level scene classification and object-based image analysis to achieve pixel-wise semantic labeling. Intensive experiments show that Million-AID is a challenging yet useful dataset, which can serve as a benchmark for evaluating newly developed algorithms. When transferring knowledge from Million-AID, fine-tuning CNN models pretrained on Million-AID perform consistently better than those pretrained ImageNet for aerial scene classification. Moreover, our designed hierarchical multi-task learning method achieves the state-of-the-art pixel-wise classification on the challenging GID, bridging the tile-level scene classification toward pixel-wise semantic labeling for aerial image interpretation. | ['Deren Li', 'Gong Cheng', 'Liangpei Zhang', 'Gui-Song Xia', 'Yang Long'] | 2022-01-06 | null | null | null | null | ['scene-parsing'] | ['computer-vision'] | [ 9.12599981e-01 4.76647764e-02 -1.94767505e-01 -4.90351856e-01
-7.98015237e-01 -7.74476826e-01 2.80486673e-01 4.15500313e-01
-2.31675625e-01 5.00281930e-01 -2.05025211e-01 -3.77242774e-01
-1.63196936e-01 -1.26686943e+00 -1.10353684e+00 -6.51401877e-01
4.86830249e-02 3.69480222e-01 3.87413830e-01 -2.68184334e-01
-9.68141854e-02 3.60088319e-01 -1.98979759e+00 8.44579399e-01
7.42195308e-01 1.47376919e+00 4.55007762e-01 7.66232073e-01
-3.72947484e-01 8.07598710e-01 -5.43865740e-01 -7.94259086e-02
3.98604035e-01 -2.69615889e-01 -1.21250951e+00 5.47489226e-01
1.02350628e+00 -1.28809080e-01 4.83817071e-01 1.18447697e+00
-1.31403352e-03 -2.24895418e-01 2.79881924e-01 -1.28402603e+00
-3.92993271e-01 5.11851788e-01 -4.20566201e-01 -2.12647170e-01
-9.82713327e-02 1.28209993e-01 1.23734605e+00 -5.30587614e-01
4.47197288e-01 9.86020684e-01 9.20083165e-01 1.61781862e-01
-1.16556573e+00 -1.48374721e-01 6.18837416e-01 -1.27811730e-01
-1.18790305e+00 1.53760657e-01 7.13478088e-01 -6.04873538e-01
7.51171052e-01 5.79337895e-01 6.85469151e-01 6.08476341e-01
-1.43256709e-01 8.55818689e-01 1.22219110e+00 -2.48730361e-01
2.67380267e-01 -2.53714323e-01 2.98629314e-01 8.25225234e-01
1.70718133e-01 -3.69388312e-01 -1.30492657e-01 1.87343657e-01
6.04258657e-01 -1.57297794e-02 -2.09880754e-01 -4.04357135e-01
-1.19592607e+00 5.47408819e-01 8.58395159e-01 2.43092850e-01
-4.30235773e-01 2.51112640e-01 5.13136923e-01 -4.84121870e-03
7.58898318e-01 4.81263757e-01 -9.26225543e-01 5.28974295e-01
-1.06981337e+00 1.69641688e-01 4.70361859e-01 1.04475784e+00
9.46267366e-01 7.68073648e-02 -1.56722292e-01 7.84068346e-01
1.68046772e-01 5.08710980e-01 8.75424370e-02 -7.46963620e-01
5.96546113e-01 1.11668646e+00 -1.75877810e-01 -8.64848197e-01
-5.29884458e-01 -5.37503779e-01 -1.00559270e+00 3.10341030e-01
4.26105976e-01 -4.47748266e-02 -1.19399238e+00 1.43317509e+00
2.60445774e-01 2.18374897e-02 3.11866015e-01 9.29500520e-01
1.15258586e+00 8.04896712e-01 4.93690223e-01 2.87500173e-01
1.75119627e+00 -1.21930158e+00 -2.23715931e-01 -5.73523939e-01
8.50289762e-01 -4.03817117e-01 1.34010196e+00 3.76320899e-01
-7.44881868e-01 -7.71232426e-01 -1.06419158e+00 5.80376759e-02
-8.50818157e-01 6.92114055e-01 6.31628871e-01 5.53083181e-01
-1.06123948e+00 3.31416637e-01 -4.66151208e-01 -5.67718625e-01
8.44230831e-01 3.58060509e-01 -3.60077351e-01 -3.06491414e-03
-9.63132143e-01 3.45410526e-01 8.60229492e-01 2.22952828e-01
-1.13532257e+00 -8.75127614e-01 -9.25624788e-01 2.01735813e-02
5.29784024e-01 -6.96344376e-01 1.09999800e+00 -1.34870648e+00
-1.03636909e+00 1.35725105e+00 8.06390569e-02 -5.45637071e-01
2.57987648e-01 -1.05164886e-01 -1.05848834e-01 1.23485535e-01
4.19658273e-01 9.73770618e-01 7.36046970e-01 -1.59648359e+00
-1.14220488e+00 -2.91904479e-01 6.23942971e-01 2.29729116e-01
-2.88524896e-01 -3.71617645e-01 -2.65545428e-01 -5.54200947e-01
2.31364474e-01 -6.14281952e-01 -3.36485237e-01 1.38148651e-01
-6.85489714e-01 6.00632280e-02 1.09230924e+00 -3.96965295e-01
8.71354342e-01 -2.08074355e+00 1.39615625e-01 -1.42706126e-01
8.69209170e-02 2.74327755e-01 -2.95633107e-01 1.36185423e-01
-3.68596077e-01 3.61944318e-01 -8.57123554e-01 -3.36333156e-01
-2.65258402e-01 2.45598376e-01 -5.17180383e-01 1.66934460e-01
4.54043299e-01 9.77644563e-01 -7.38305390e-01 -5.08974195e-01
5.22718906e-01 -2.72545200e-02 -4.73172724e-01 2.78371930e-01
-8.30653608e-01 4.72431272e-01 -6.10795081e-01 1.03497577e+00
7.15878725e-01 -3.63677889e-01 6.79171607e-02 -6.63709402e-01
-3.09020877e-01 -3.12214941e-01 -8.31755280e-01 1.88516152e+00
-6.56147599e-01 5.13692737e-01 2.20755428e-01 -1.35148346e+00
8.66234720e-01 -4.79998700e-02 5.65826952e-01 -4.94047076e-01
1.47993609e-01 1.06054336e-01 -5.06099522e-01 -3.30271393e-01
6.48297429e-01 2.52643805e-02 -3.88366342e-01 4.25920449e-02
1.07462034e-01 -6.25967145e-01 2.09088787e-01 -2.36449078e-01
6.93932533e-01 4.14648265e-01 2.97317624e-01 -4.71554071e-01
7.29520440e-01 1.04621196e+00 2.43340820e-01 5.04507065e-01
1.72092944e-01 7.23181844e-01 4.11101341e-01 -1.05408168e+00
-1.03659952e+00 -6.83077753e-01 -2.65760154e-01 1.14818108e+00
3.66502166e-01 -2.91294008e-01 -1.21519244e+00 -7.30126500e-01
-1.47688761e-01 4.29952353e-01 -6.21139765e-01 2.22610861e-01
-3.82087171e-01 -1.25586200e+00 6.80862606e-01 4.70215023e-01
1.22378647e+00 -1.04416990e+00 -9.13617611e-01 1.12150930e-01
-2.27365881e-01 -1.51736438e+00 2.19500527e-01 5.62694192e-01
-8.55146885e-01 -1.26020944e+00 -5.06457269e-01 -8.19496572e-01
7.76705027e-01 3.76586050e-01 1.38877106e+00 2.38378555e-01
-5.81109762e-01 3.19012046e-01 -5.88695586e-01 -3.07166249e-01
-1.72086090e-01 3.80322069e-01 -5.16962469e-01 1.22708179e-01
-1.37005657e-01 -2.18452320e-01 -4.17774737e-01 3.59490395e-01
-1.45159996e+00 4.68471885e-01 5.81139445e-01 6.36537313e-01
1.02291536e+00 3.31009656e-01 -9.48724300e-02 -1.04605544e+00
2.43483130e-02 -1.85620204e-01 -8.27886224e-01 5.67501485e-01
-6.69932365e-03 -1.77310571e-01 9.58707750e-01 1.74142629e-01
-9.01543558e-01 5.82981110e-01 -2.98943728e-01 -1.16601072e-01
-9.56607819e-01 5.54710031e-01 -3.07042718e-01 -1.34662226e-01
6.06569886e-01 2.76126742e-01 -5.81103683e-01 -2.18081072e-01
4.40060318e-01 4.57568973e-01 6.07808948e-01 -7.16403961e-01
5.98739207e-01 6.69737279e-01 2.46855453e-01 -1.09294069e+00
-1.53090072e+00 -4.56809580e-01 -1.17497563e+00 -2.28307143e-01
1.46449947e+00 -1.21381509e+00 -4.02343452e-01 6.70341730e-01
-1.22977233e+00 -9.42381024e-01 -3.86685848e-01 -1.72746465e-01
-6.94854796e-01 6.59184828e-02 -1.21294178e-01 -4.49407637e-01
-1.82413116e-01 -1.36267984e+00 1.89705181e+00 -8.00556913e-02
2.25657150e-02 -1.06830633e+00 -2.82327741e-01 6.09364808e-01
7.65210614e-02 6.99850440e-01 1.20823121e+00 -2.86670715e-01
-7.96931565e-01 2.43476275e-02 -5.76297164e-01 3.80955338e-01
1.37482375e-01 -4.12063152e-02 -1.19507885e+00 1.24134898e-01
-3.93848509e-01 -5.73125184e-01 9.27215397e-01 6.05461776e-01
1.71799684e+00 3.59481946e-02 -1.01824373e-01 1.04242754e+00
1.91504955e+00 -4.38846275e-02 5.05517542e-01 6.52223647e-01
1.01854336e+00 8.10937703e-01 9.08344746e-01 3.15693378e-01
3.93620342e-01 6.48746073e-01 1.10021448e+00 -7.37050533e-01
-2.41246209e-01 -1.74996495e-01 -8.57033953e-02 4.23777439e-02
-7.77450874e-02 -6.81075215e-01 -1.07376289e+00 4.36265469e-01
-1.89253020e+00 -4.17123079e-01 -7.44316459e-01 1.79450345e+00
3.91758084e-01 -1.45538762e-01 -7.23053841e-03 2.13692755e-01
5.70161223e-01 5.41945219e-01 -4.27810103e-01 -1.11190408e-01
-5.20808756e-01 1.09116197e-01 9.90145326e-01 2.45623231e-01
-1.85046566e+00 1.51977360e+00 5.72017527e+00 8.65154088e-01
-1.11976123e+00 7.21170679e-02 8.62583220e-01 4.77541775e-01
1.46787897e-01 -1.70967430e-01 -9.60119247e-01 5.04418463e-02
5.14777005e-01 4.63895977e-01 2.88375933e-02 1.16190267e+00
8.04683343e-02 -2.54175901e-01 -8.36964905e-01 7.88269222e-01
-6.67056441e-02 -1.45495868e+00 4.28573906e-01 -1.37624294e-01
8.70580733e-01 3.04666553e-02 -7.64850676e-02 -1.32934988e-01
3.42632502e-01 -9.33362663e-01 1.04556441e+00 3.71139534e-02
9.95067894e-01 -2.74497718e-01 8.75346839e-01 2.13485658e-01
-1.83961236e+00 -2.29242384e-01 -4.31390852e-01 7.79105797e-02
4.09996621e-02 6.06078207e-01 -3.01674157e-01 8.91593516e-01
9.18560088e-01 1.06184280e+00 -7.90079892e-01 6.81323409e-01
-3.04934740e-01 5.12695611e-01 -1.42021120e-01 3.97064000e-01
7.79705822e-01 -3.20750684e-01 1.77244283e-02 1.13710964e+00
3.10938150e-01 1.99228480e-01 5.74708939e-01 9.07316446e-01
-9.35801119e-02 1.27655212e-02 -7.52043009e-01 -8.62151198e-03
-2.06414506e-01 1.58093536e+00 -1.24547172e+00 -5.03968775e-01
-2.09323093e-01 1.17907536e+00 -2.32980829e-02 2.00361729e-01
-8.92125726e-01 1.08545255e-02 7.51268446e-01 7.76776252e-03
2.01341867e-01 -1.64112464e-01 -5.72197258e-01 -8.80135655e-01
-1.50380814e-02 -6.25063121e-01 3.02744091e-01 -9.51312184e-01
-8.87624919e-01 9.07525182e-01 1.48405164e-01 -1.37047744e+00
3.28989923e-01 -1.07815444e+00 -4.70880091e-01 5.32757282e-01
-1.82850122e+00 -1.81358254e+00 -1.02968144e+00 4.42891538e-01
1.05659890e+00 1.37579009e-01 9.79052901e-01 2.18367875e-01
-6.03101432e-01 -1.13991022e-01 -3.51300061e-01 1.80008724e-01
1.36507064e-01 -1.37048495e+00 4.01793927e-01 8.59969079e-01
1.18155137e-01 -2.59389430e-01 2.49269530e-01 -4.41408157e-01
-1.11667335e+00 -1.93205070e+00 3.72603595e-01 -3.85189980e-01
5.78272164e-01 -3.76531839e-01 -6.41088486e-01 5.87970614e-01
-1.44875005e-01 2.90154815e-01 4.03987110e-01 -2.39485651e-01
-7.98861682e-02 -2.53138572e-01 -9.66947317e-01 3.00068110e-01
1.28641033e+00 -3.46667528e-01 -2.29889676e-01 7.46940136e-01
9.13342535e-01 -3.91412228e-01 -8.27114224e-01 8.03305805e-01
1.85160428e-01 -1.03868043e+00 1.23314726e+00 -5.00941694e-01
7.91272998e-01 -6.70166433e-01 -5.41347504e-01 -1.22080266e+00
-1.72745332e-01 3.14015985e-01 8.50773394e-01 1.05960417e+00
4.81364131e-01 -2.50466019e-01 8.34604383e-01 3.80141921e-02
-7.29110777e-01 -5.14874518e-01 -4.58390355e-01 -7.14013100e-01
-1.05836369e-01 -6.88148975e-01 7.39868701e-01 7.79660940e-01
-8.83879125e-01 -5.04949428e-02 -2.30902910e-01 6.44108951e-01
5.84298551e-01 5.24608910e-01 8.58574212e-01 -1.28208423e+00
4.08484861e-02 -4.66724426e-01 -3.44200134e-01 -9.70637619e-01
2.98785031e-01 -9.20294583e-01 3.31385732e-01 -1.79703879e+00
-6.43321127e-02 -5.05586028e-01 1.92166939e-01 7.92885125e-01
-2.00066343e-02 6.86447561e-01 9.18061063e-02 -3.44408154e-02
-7.27181137e-01 3.75562549e-01 1.09191668e+00 -6.15577817e-01
3.59196425e-03 -1.84451565e-01 -3.89079064e-01 9.35221493e-01
8.12392652e-01 -3.44142675e-01 -4.13998574e-01 -8.50265563e-01
1.75026238e-01 -3.71994585e-01 8.98130536e-01 -1.23385477e+00
2.22769938e-02 -1.16101511e-01 1.78936541e-01 -6.76771760e-01
3.12623382e-01 -1.14172292e+00 1.85099006e-01 4.59406078e-01
-3.25576514e-01 -1.40061408e-01 4.92234975e-01 2.28564173e-01
-6.20010316e-01 -1.64504007e-01 8.11589301e-01 -5.07145882e-01
-1.44482684e+00 4.42295402e-01 -3.28969896e-01 -7.31564090e-02
1.19202149e+00 -3.06124955e-01 -2.44705439e-01 2.52738595e-01
-6.40582919e-01 2.12117508e-01 4.22698975e-01 2.67498016e-01
4.39039379e-01 -9.22593057e-01 -5.12531757e-01 3.07843983e-01
6.11139655e-01 6.78028464e-01 3.37568879e-01 5.88245571e-01
-1.01474941e+00 5.94723225e-01 -3.08826685e-01 -1.07039368e+00
-1.48496914e+00 2.60125577e-01 3.92182171e-01 -3.28555077e-01
-4.07110453e-01 8.68952930e-01 6.55851781e-01 -7.73668349e-01
-7.40119815e-02 -9.71734166e-01 -2.29962334e-01 1.61223918e-01
1.39484540e-01 -1.36314392e-01 2.70534873e-01 -7.28150189e-01
-2.00973943e-01 1.25137556e+00 7.91385829e-01 4.46298182e-01
1.47854412e+00 8.68694186e-02 -4.21815574e-01 3.33617806e-01
9.41205919e-01 -6.62346423e-01 -1.16460824e+00 1.65593182e-03
7.27409795e-02 -3.25228930e-01 2.45931908e-01 -8.23080540e-01
-1.20027339e+00 1.01160717e+00 5.19798279e-01 4.02582258e-01
1.48431933e+00 4.84914705e-02 5.14126718e-01 4.67288792e-01
6.80524111e-01 -7.94478297e-01 -6.98683113e-02 4.82943147e-01
7.51638412e-01 -1.58872712e+00 1.98462997e-02 -1.02339613e+00
-6.01874709e-01 1.24003065e+00 5.96048474e-01 -2.60095596e-02
5.10669589e-01 1.80455789e-01 1.16439946e-02 -5.57684422e-01
-3.10413778e-01 -7.48025239e-01 4.59721327e-01 5.17495751e-01
2.47453034e-01 3.40451062e-01 4.34363872e-01 4.05091882e-01
-1.23627380e-01 -1.11206919e-01 7.44294524e-02 7.02920020e-01
-5.96463025e-01 -1.12284017e+00 -4.65548933e-01 2.84430385e-01
-1.48271829e-01 -3.32071662e-01 -6.71104550e-01 9.36829150e-01
6.27302229e-01 7.80152678e-01 1.55224279e-01 -3.72326910e-01
5.70706189e-01 -9.78564397e-02 1.88349247e-01 -9.37107265e-01
-6.82564795e-01 -2.64265269e-01 2.13237390e-01 -6.66796625e-01
-8.95536184e-01 -3.52565140e-01 -9.93108928e-01 1.39264196e-01
3.24353622e-03 -1.27305597e-01 8.12085271e-01 1.05129719e+00
2.29807924e-02 8.79917383e-01 4.23906982e-01 -9.32878792e-01
2.73708910e-01 -5.78455091e-01 -5.20346999e-01 2.89296478e-01
2.43048087e-01 -3.57382059e-01 -1.37884647e-01 4.54030424e-01] | [9.642585754394531, 0.3259478211402893] |
bb13bdd5-0979-4167-a4b5-0ddd0aaa1f11 | zero-shot-long-form-voice-cloning-with | 2201.10375 | null | https://arxiv.org/abs/2201.10375v2 | https://arxiv.org/pdf/2201.10375v2.pdf | Zero-Shot Long-Form Voice Cloning with Dynamic Convolution Attention | With recent advancements in voice cloning, the performance of speech synthesis for a target speaker has been rendered similar to the human level. However, autoregressive voice cloning systems still suffer from text alignment failures, resulting in an inability to synthesize long sentences. In this work, we propose a variant of attention-based text-to-speech system that can reproduce a target voice from a few seconds of reference speech and generalize to very long utterances as well. The proposed system is based on three independently trained components: a speaker encoder, synthesizer and universal vocoder. Generalization to long utterances is realized using an energy-based attention mechanism known as Dynamic Convolution Attention, in combination with a set of modifications proposed for the synthesizer based on Tacotron 2. Moreover, effective zero-shot speaker adaptation is achieved by conditioning both the synthesizer and vocoder on a speaker encoder that has been pretrained on a large corpus of diverse data. We compare several implementations of voice cloning systems in terms of speech naturalness, speaker similarity, alignment consistency and ability to synthesize long utterances, and conclude that the proposed model can produce intelligible synthetic speech for extremely long utterances, while preserving a high extent of naturalness and similarity for short texts. | ['Ivan Ozhiganov', 'Artem Gorodetskii'] | 2022-01-25 | null | null | null | null | ['voice-cloning'] | ['speech'] | [ 2.38956839e-01 3.79954934e-01 2.72726119e-01 -4.06203032e-01
-9.34854209e-01 -3.77291203e-01 6.26473188e-01 -4.87585723e-01
-5.00655621e-02 5.73261201e-01 5.05225003e-01 -3.04368258e-01
4.04234976e-01 -3.02409530e-01 -6.25135303e-01 -6.55445576e-01
4.91705358e-01 4.11131114e-01 7.62111023e-02 -2.19131127e-01
-1.27898365e-01 6.42174840e-01 -1.70783293e+00 2.54485551e-02
9.06115353e-01 7.88587332e-01 5.87561011e-01 1.09674919e+00
-1.31466612e-01 6.02105260e-01 -9.74689603e-01 -2.06953228e-01
-1.90365925e-01 -8.46460938e-01 -7.86709726e-01 2.67736435e-01
3.48408401e-01 -2.71140993e-01 -3.96960765e-01 8.40963483e-01
1.05075085e+00 4.93169874e-01 5.85693359e-01 -8.68063033e-01
-8.69332552e-01 8.94015014e-01 2.80395836e-01 -1.90134980e-02
3.91280860e-01 2.15607405e-01 8.85275900e-01 -1.01793039e+00
3.67611378e-01 1.46561408e+00 3.76326770e-01 1.05500734e+00
-1.15444708e+00 -4.20152366e-01 -1.33160248e-01 -6.10293932e-02
-1.30213916e+00 -1.22973740e+00 7.76434302e-01 -1.26610398e-01
1.49569666e+00 6.53149784e-01 3.37578237e-01 1.37943769e+00
3.24839205e-01 4.73777920e-01 4.66886759e-01 -6.08415604e-01
3.38741481e-01 2.62130052e-01 -2.81146258e-01 3.68153363e-01
-5.64404190e-01 2.00834826e-01 -4.36001152e-01 8.77461806e-02
4.94016558e-01 -4.85410661e-01 -4.85051185e-01 1.98287129e-01
-1.10125637e+00 7.21836090e-01 2.21966952e-01 5.48077464e-01
-5.48559964e-01 3.80477086e-02 5.20389020e-01 4.33110118e-01
5.39018095e-01 4.49305445e-01 -7.18261823e-02 -1.67712465e-01
-1.12374485e+00 1.01671079e-02 9.30092454e-01 1.21794164e+00
1.54096007e-01 1.02921152e+00 -3.07357162e-01 1.14379013e+00
9.97297466e-02 7.03720987e-01 1.19777131e+00 -6.27172112e-01
2.99823165e-01 -1.71372116e-01 -7.88752362e-02 -5.32835782e-01
-2.70219713e-01 -6.78768039e-01 -7.83992350e-01 1.47628307e-01
-3.30374122e-01 -3.67542386e-01 -9.31462824e-01 1.87352705e+00
2.28530541e-01 6.23982102e-02 5.11439145e-01 8.34912658e-01
9.88052785e-01 1.09774125e+00 -2.57418603e-01 -2.39207610e-01
1.02172422e+00 -1.43720400e+00 -1.16106510e+00 -2.48378813e-01
2.25173250e-01 -9.61637914e-01 1.34444427e+00 5.87425530e-02
-1.32683718e+00 -1.02683759e+00 -1.12339079e+00 -1.83103815e-01
-9.92928222e-02 9.30497497e-02 -2.02296019e-01 8.14195871e-01
-1.21648681e+00 5.55283666e-01 -5.67445219e-01 -4.67196465e-01
-2.52478480e-01 3.62052888e-01 -1.51097789e-01 5.05150139e-01
-1.06056440e+00 1.01707411e+00 3.58468235e-01 -1.00182310e-01
-1.00083053e+00 -5.21659791e-01 -8.78657997e-01 4.90290940e-01
3.42094265e-02 -8.27491581e-01 1.62111521e+00 -1.10014892e+00
-2.48815060e+00 3.38662773e-01 -2.34635904e-01 -6.21308744e-01
1.85191676e-01 -1.79903433e-01 -9.32338774e-01 -7.52649037e-03
-1.36091307e-01 5.81905544e-01 1.19849682e+00 -1.00491571e+00
-3.29852909e-01 -1.98869947e-02 -4.55918789e-01 3.47713500e-01
-1.74502477e-01 1.72016114e-01 -8.24034140e-02 -7.61660457e-01
-1.80454582e-01 -8.85701418e-01 6.91003129e-02 -5.66227019e-01
-5.22513330e-01 -2.20322534e-02 9.50466275e-01 -9.43392217e-01
1.11860132e+00 -2.11599994e+00 4.83630419e-01 -3.65105331e-01
-4.10429657e-01 4.38302338e-01 -3.49081784e-01 6.12640917e-01
-1.64260358e-01 -3.34496461e-02 -2.93984294e-01 -7.29677439e-01
4.32642847e-02 1.08642220e-01 -7.04673409e-01 2.62531638e-01
2.98533887e-01 8.10019612e-01 -5.75645745e-01 -1.80295303e-01
3.12848181e-01 6.64008081e-01 -5.82686305e-01 7.02022374e-01
-2.92020530e-01 5.27650476e-01 4.50202860e-02 2.71278471e-01
2.13883787e-01 3.65519434e-01 -1.33841872e-01 1.40782848e-01
-3.25690001e-01 6.73384070e-01 -8.21908832e-01 1.72414851e+00
-9.59761500e-01 6.94603860e-01 2.83850968e-01 -7.03461766e-01
1.10698807e+00 1.06886160e+00 -3.54096591e-02 -3.93546373e-01
2.36741289e-01 3.45404297e-01 2.05138043e-01 -4.76696640e-01
6.84601784e-01 -4.38724965e-01 3.85283642e-02 3.92912269e-01
4.73249584e-01 -8.52876186e-01 -2.66867071e-01 -1.48659453e-01
6.98976338e-01 2.08172891e-02 2.05811143e-01 -1.51623815e-01
7.54353940e-01 -4.73556399e-01 1.98994964e-01 4.75423574e-01
1.34900734e-01 8.54431272e-01 -1.31874859e-01 1.17763050e-01
-1.41075790e+00 -9.82014060e-01 3.36978063e-02 1.17568970e+00
-1.94282338e-01 -1.80531934e-01 -1.15261352e+00 -1.22115403e-01
-3.07791889e-01 1.43182468e+00 -1.06104046e-01 -4.57496911e-01
-6.99531257e-01 -1.98783092e-02 8.04339230e-01 3.45546931e-01
1.99385464e-01 -1.41078389e+00 -2.84905821e-01 6.05983257e-01
-2.07040772e-01 -1.16750789e+00 -9.50546443e-01 1.87468708e-01
-6.64586067e-01 -1.71555310e-01 -9.12950516e-01 -1.00293183e+00
1.53214127e-01 2.17134431e-02 8.74200225e-01 -2.27580011e-01
8.96032378e-02 1.67741388e-01 -2.70038694e-01 -3.87544304e-01
-1.24965596e+00 2.76392609e-01 5.15840232e-01 1.50219202e-01
-2.38325372e-01 -6.72337890e-01 -5.51133938e-02 2.61431277e-01
-7.87494779e-01 1.08104728e-01 4.37955528e-01 1.12727380e+00
2.27314919e-01 -2.09381551e-01 1.19550657e+00 -3.05865079e-01
8.31386566e-01 -2.41122648e-01 -3.98690283e-01 1.47370040e-01
-4.11720067e-01 2.11689398e-01 1.17810571e+00 -5.73283553e-01
-1.33213782e+00 2.13752743e-02 -7.38993227e-01 -5.68787217e-01
-1.56296313e-01 6.39779419e-02 -5.38178623e-01 2.60321259e-01
6.33873105e-01 6.87712133e-01 2.01841295e-01 -6.08569086e-01
6.58252597e-01 1.39431918e+00 8.86935890e-01 -2.75578618e-01
6.69555545e-01 -2.64849246e-01 -6.25665784e-01 -1.31733644e+00
-3.04600865e-01 -2.16793418e-01 -5.14409006e-01 -2.52082311e-02
7.51646817e-01 -8.06893289e-01 -5.38596392e-01 5.86447001e-01
-1.28251386e+00 -2.70252109e-01 -5.02580225e-01 7.26256847e-01
-8.61180604e-01 2.59154260e-01 -5.24496377e-01 -8.61328900e-01
-7.90309429e-01 -1.35939193e+00 1.20153809e+00 6.45372123e-02
-3.34768504e-01 -8.08535755e-01 8.81671980e-02 2.26868317e-01
9.12187219e-01 -3.33604187e-01 8.46262038e-01 -9.04996753e-01
-3.25422525e-01 2.85191531e-03 4.27698314e-01 8.10684502e-01
4.78849471e-01 3.35901156e-02 -1.33569288e+00 -3.71941328e-01
4.46692735e-01 -2.63846695e-01 5.72784543e-01 3.25760782e-01
6.83165908e-01 -5.83240867e-01 8.31880048e-02 5.93823850e-01
7.87016153e-01 6.00734890e-01 5.61536670e-01 -3.98568064e-01
4.86337155e-01 5.37269950e-01 8.55965316e-02 8.62182453e-02
-1.35953024e-01 9.01959896e-01 2.46068344e-01 -5.22809103e-03
-5.50863206e-01 -1.74994245e-01 6.62534297e-01 1.58706558e+00
2.28157863e-01 -6.15652263e-01 -4.92146283e-01 5.56963861e-01
-1.32341683e+00 -1.09145761e+00 3.60482723e-01 2.16593194e+00
8.88179898e-01 -4.92186025e-02 4.22666110e-02 1.12881497e-01
9.57104623e-01 3.88313502e-01 -5.53234637e-01 -1.05896270e+00
-4.78208177e-02 3.12077671e-01 -7.98279494e-02 9.43975389e-01
-6.43866956e-01 1.24271941e+00 6.71367979e+00 8.64896178e-01
-1.67176557e+00 1.72196940e-01 2.58739620e-01 -2.01880008e-01
-3.57486844e-01 -3.17314923e-01 -6.65885448e-01 3.58837187e-01
1.67367744e+00 -5.69413245e-01 7.11113632e-01 9.52031076e-01
3.27511489e-01 6.10170722e-01 -1.24007368e+00 8.16243887e-01
3.20572644e-01 -1.18353891e+00 1.75991908e-01 -3.17836553e-01
6.23925745e-01 -4.31294553e-02 7.27400631e-02 4.56115246e-01
-1.26245603e-01 -1.19637227e+00 1.05590069e+00 3.14876765e-01
1.06771266e+00 -7.91062593e-01 4.79581147e-01 4.79315817e-01
-1.08137310e+00 9.06285569e-02 -3.25511068e-01 1.17146745e-01
3.14237535e-01 1.39291510e-01 -1.37793911e+00 5.21525919e-01
2.46249828e-02 1.31360963e-01 5.08050993e-02 5.85389912e-01
-7.35567138e-02 6.99678004e-01 -1.01561569e-01 -2.79736608e-01
2.10923582e-01 6.33833259e-02 8.33327830e-01 1.18900275e+00
7.26655722e-01 -4.39262167e-02 -1.78962588e-01 8.62946510e-01
-2.56407857e-01 3.41369003e-01 -5.71746349e-01 -3.77811342e-01
6.04873240e-01 9.02865291e-01 -3.44924442e-02 -4.05326962e-01
-2.32741639e-01 1.40008211e+00 8.12207907e-02 4.77684736e-01
-8.96604955e-01 -6.70741260e-01 6.02220774e-01 -2.75668651e-01
3.01736325e-01 -1.37380540e-01 2.33862177e-02 -1.05935729e+00
-1.34409405e-02 -1.10148168e+00 -4.49431390e-01 -9.54037011e-01
-8.17546606e-01 1.37391055e+00 -5.26577592e-01 -1.00984406e+00
-8.79139781e-01 -1.29278213e-01 -8.85939360e-01 1.30745375e+00
-1.02023268e+00 -9.58412588e-01 1.66555811e-02 3.56072903e-01
1.26264524e+00 -6.82269454e-01 1.20551753e+00 1.64659396e-01
-5.91874123e-01 6.56912982e-01 2.46774733e-01 -3.17073405e-01
6.07845604e-01 -1.11591446e+00 1.00684023e+00 7.64879346e-01
1.83212966e-01 3.12625170e-01 9.43241596e-01 -4.15723145e-01
-1.26115406e+00 -1.21673703e+00 1.08509219e+00 1.32664919e-01
3.60151380e-01 -5.34690142e-01 -1.07548952e+00 6.48021042e-01
7.26209641e-01 -3.67972374e-01 5.13634443e-01 -2.79066384e-01
-2.36859061e-02 -1.27651945e-01 -9.68157947e-01 7.30267227e-01
8.27291787e-01 -8.75936091e-01 -9.16405022e-01 3.19332272e-01
1.26641715e+00 -3.72297138e-01 -7.22387433e-01 6.02718107e-02
4.03452665e-01 -7.51878381e-01 7.05337048e-01 -5.18052399e-01
1.42145053e-01 -5.77959744e-03 -3.15900356e-01 -1.71011925e+00
-3.49690229e-01 -1.09505832e+00 1.04017466e-01 1.50483906e+00
4.91149455e-01 -5.68973064e-01 1.52869806e-01 2.65362740e-01
-8.30974221e-01 -4.04810935e-01 -1.22637570e+00 -1.05833423e+00
1.25657871e-01 -2.45434150e-01 8.49085212e-01 4.76548672e-01
-1.01542454e-02 8.80385816e-01 -7.16022968e-01 3.72008473e-01
1.10914171e-01 3.11217606e-02 7.07390010e-01 -7.16453969e-01
-4.89045292e-01 -3.92669708e-01 -7.31293559e-02 -1.12852848e+00
4.79622215e-01 -9.73093152e-01 5.06569266e-01 -1.24922371e+00
-3.09623688e-01 7.33185792e-03 1.90193459e-01 2.70206612e-02
-7.28649870e-02 -2.13845730e-01 1.59440935e-01 7.88640324e-03
1.32973164e-01 1.24045622e+00 1.07469940e+00 -1.43364981e-01
-4.36851263e-01 2.03271508e-01 -2.50668049e-01 4.79031444e-01
6.91786170e-01 -3.51931870e-01 -5.44409096e-01 -3.48276049e-01
-8.38838696e-01 5.72235882e-01 -3.40977572e-02 -1.29734361e+00
2.47199863e-01 3.15384381e-02 -7.96081722e-02 -4.72183466e-01
7.53665209e-01 -6.25517488e-01 5.13855278e-01 4.48022008e-01
-6.27826214e-01 -2.67148931e-02 2.39249036e-01 2.04864442e-01
-4.27042007e-01 -4.10074800e-01 9.92551744e-01 1.49042800e-01
-2.31182113e-01 -2.68490184e-02 -5.63025296e-01 -1.62860051e-01
6.66340232e-01 8.22794996e-03 -6.41349629e-02 -6.76739275e-01
-8.33279371e-01 -2.93083727e-01 2.77823210e-01 6.80039704e-01
5.88828743e-01 -1.48638022e+00 -9.80433583e-01 5.49965441e-01
-2.31022879e-01 -3.83021921e-01 3.39927584e-01 2.86195874e-01
-3.56590867e-01 8.22827578e-01 -1.22936234e-01 -4.72430229e-01
-1.23134613e+00 7.32131302e-01 5.77863038e-01 2.89865047e-01
-7.62960792e-01 8.04371119e-01 1.54802725e-01 -5.25458694e-01
4.95913833e-01 -4.54530418e-01 1.51873350e-01 -2.39382997e-01
5.11280239e-01 1.69739768e-01 1.71527907e-01 -9.29062724e-01
-2.77047604e-01 3.69705230e-01 1.60876602e-01 -5.09201109e-01
1.08791018e+00 -3.25016588e-01 3.14138561e-01 5.63859701e-01
1.19255328e+00 3.26447517e-01 -9.59118307e-01 -1.94268018e-01
-3.10333997e-01 -1.16938829e-01 8.23124498e-02 -6.45930588e-01
-8.34871054e-01 9.92760420e-01 3.89481366e-01 3.48162025e-01
1.03746331e+00 -8.40100497e-02 1.14799738e+00 3.68394703e-01
5.00249453e-02 -9.76596236e-01 7.15258494e-02 6.51092410e-01
1.55002046e+00 -9.79467273e-01 -6.44968331e-01 -1.32484317e-01
-7.60590494e-01 1.16672802e+00 2.83031225e-01 8.55429620e-02
3.26399148e-01 3.87084335e-01 1.83822379e-01 3.80181640e-01
-1.00434625e+00 6.96865544e-02 3.73174995e-01 5.28854907e-01
5.24043381e-01 3.10782343e-02 -1.15707070e-02 6.15244925e-01
-7.47339487e-01 -3.41651976e-01 5.63226044e-01 2.93361485e-01
-7.30562627e-01 -9.00088370e-01 -4.61547107e-01 -1.66913748e-01
-4.86426890e-01 -3.37456852e-01 -3.13987195e-01 3.57969612e-01
-3.72911990e-01 1.20074892e+00 1.46167770e-01 -4.51001734e-01
3.85015488e-01 5.53480148e-01 1.76702023e-01 -8.04448962e-01
-8.43302369e-01 3.53269815e-01 4.05232698e-01 -2.37626672e-01
1.12783276e-02 -2.97803432e-01 -1.22853386e+00 -1.17904663e-01
-5.85373223e-01 3.37428421e-01 9.92924154e-01 8.37869048e-01
4.79910344e-01 1.00153565e+00 9.48628068e-01 -9.45874512e-01
-1.13218188e+00 -1.34752095e+00 -6.79291010e-01 1.22024082e-01
6.48164451e-01 -1.68489978e-01 -4.64212149e-01 1.17793508e-01] | [14.903270721435547, 6.5775628089904785] |
9a64cd79-ecd3-4f84-b96e-a46baf77d83a | scalable-variable-selection-for-two-view | 2307.01558 | null | https://arxiv.org/abs/2307.01558v1 | https://arxiv.org/pdf/2307.01558v1.pdf | Scalable variable selection for two-view learning tasks with projection operators | In this paper we propose a novel variable selection method for two-view settings, or for vector-valued supervised learning problems. Our framework is able to handle extremely large scale selection tasks, where number of data samples could be even millions. In a nutshell, our method performs variable selection by iteratively selecting variables that are highly correlated with the output variables, but which are not correlated with the previously chosen variables. To measure the correlation, our method uses the concept of projection operators and their algebra. With the projection operators the relationship, correlation, between sets of input and output variables can also be expressed by kernel functions, thus nonlinear correlation models can be exploited as well. We experimentally validate our approach, showing on both synthetic and real data its scalability and the relevance of the selected features. Keywords: Supervised variable selection, vector-valued learning, projection-valued measure, reproducing kernel Hilbert space | ['Juho Rousu', 'Tat Hong Duong Le', 'Riikka Huusari', 'Sandor Szedmak'] | 2023-07-04 | null | null | null | null | ['variable-selection'] | ['methodology'] | [ 2.58893460e-01 -9.85521898e-02 -2.87845671e-01 -4.44648951e-01
-4.96644616e-01 -5.78020036e-01 4.95420635e-01 2.02757880e-01
-3.17349672e-01 1.04915118e+00 -8.97892192e-02 1.84652552e-01
-4.97227132e-01 -1.00916147e+00 -7.52675310e-02 -9.22623813e-01
-4.02935654e-01 7.11952090e-01 -2.32811317e-01 -2.17137024e-01
3.04361254e-01 5.02212107e-01 -1.77519166e+00 -1.43545315e-01
4.39779252e-01 8.71886671e-01 -1.13017328e-01 6.33066535e-01
-7.75738582e-02 2.21552745e-01 -4.24936414e-01 -2.22436458e-01
4.49277878e-01 -5.15748084e-01 -5.95432162e-01 9.13620647e-03
2.47435838e-01 4.48769689e-01 4.02953744e-01 1.03487146e+00
4.68436539e-01 2.65579581e-01 8.47450197e-01 -1.42235947e+00
-1.79703787e-01 5.25496900e-01 -5.72104275e-01 -1.54435486e-01
3.62641931e-01 -1.78490162e-01 1.39710045e+00 -1.22837567e+00
8.78059685e-01 1.22367001e+00 5.67828238e-01 1.81420386e-01
-1.80839145e+00 -4.80205983e-01 -2.46041700e-01 1.78785115e-01
-1.19888818e+00 -5.71509413e-02 7.50587523e-01 -6.45392358e-01
5.75972736e-01 5.78405619e-01 8.46378565e-01 9.51778114e-01
1.67071670e-01 5.08046925e-01 1.36589921e+00 -5.56630254e-01
4.99511123e-01 8.34964931e-01 3.64672095e-01 7.19423354e-01
8.82884860e-02 1.62349746e-01 -7.33059704e-01 -6.15688026e-01
5.99551857e-01 4.71552601e-03 -3.13354999e-01 -1.16005647e+00
-1.52621961e+00 1.24644685e+00 -2.53878146e-01 1.80497870e-01
-4.05458182e-01 -3.53982598e-01 5.95363796e-01 8.23852777e-01
3.15046281e-01 9.32853580e-01 -7.76330829e-01 -9.55979619e-03
-6.35292470e-01 6.49560019e-02 9.28360224e-01 9.60918367e-01
1.00831521e+00 -2.81376809e-01 -3.02822918e-01 7.27491498e-01
-6.91368133e-02 6.37175560e-01 5.51404953e-01 -6.18586004e-01
2.65853524e-01 6.41395271e-01 -1.40957311e-01 -7.33208418e-01
-7.55991817e-01 -2.02124879e-01 -1.04307294e+00 5.36072552e-01
2.54199922e-01 -1.81520298e-01 -4.26540405e-01 1.52193606e+00
1.70161605e-01 -1.73314631e-01 1.13379247e-01 6.49392009e-01
2.81040549e-01 1.71202615e-01 -3.55529040e-01 -8.20355415e-01
1.14421535e+00 -5.79667449e-01 -5.44639409e-01 5.65387070e-01
5.32454014e-01 -4.88076717e-01 1.05824494e+00 6.42146111e-01
-8.22860122e-01 -3.38970095e-01 -9.93504047e-01 6.06589258e-01
-3.33661079e-01 1.75846428e-01 8.68969262e-01 6.64333165e-01
-8.88689280e-01 7.77497113e-01 -4.93107349e-01 -3.07728291e-01
2.62605935e-01 6.20143473e-01 -8.08783054e-01 4.11795914e-01
-1.15654194e+00 6.32158458e-01 2.75097489e-01 -3.70169699e-01
-2.95847088e-01 -5.74796319e-01 -6.65682316e-01 2.11892813e-01
1.73035234e-01 -6.71886265e-01 5.93137860e-01 -8.89402390e-01
-1.50375724e+00 7.68482804e-01 -6.25432730e-02 -3.16321641e-01
4.72757429e-01 8.15688223e-02 -4.63500470e-01 4.38677799e-03
6.41006157e-02 -1.27308652e-01 1.18547392e+00 -8.93657863e-01
-4.73609954e-01 -7.21760869e-01 -3.81190658e-01 1.59918927e-02
-6.17560387e-01 -4.93699685e-02 4.46473286e-02 -4.66059804e-01
9.41258371e-02 -1.06884837e+00 -5.34121394e-01 -7.94641748e-02
-3.50352198e-01 -2.94453837e-02 8.37671340e-01 -5.07473797e-02
1.07361042e+00 -2.22078013e+00 6.53182328e-01 9.45573688e-01
3.07513922e-01 -6.81811497e-02 1.18864784e-02 5.33862591e-01
-3.57173383e-01 1.58869505e-01 -4.20538187e-01 5.34529723e-02
-3.24771941e-01 -1.68155864e-01 -1.73314944e-01 7.35592782e-01
3.25079739e-01 5.02016485e-01 -8.33091259e-01 -3.81213546e-01
2.87814111e-01 3.05883110e-01 -2.48088449e-01 1.98031619e-01
2.15549186e-01 3.91180009e-01 -5.77451050e-01 4.85897273e-01
4.48319942e-01 -2.93021083e-01 2.60609716e-01 1.37809739e-01
3.57726887e-02 -3.71955752e-01 -1.63298857e+00 1.21500742e+00
-5.87873280e-01 7.71622300e-01 -1.69711173e-01 -1.03130817e+00
1.42817569e+00 3.44903678e-01 7.94492364e-01 -3.44215661e-01
-9.70265344e-02 1.53349265e-01 -2.39047945e-01 -4.59852397e-01
1.65797055e-01 -4.14886266e-01 2.88931318e-02 4.69632983e-01
2.82686710e-01 -1.27073899e-01 1.41422018e-01 -4.42565121e-02
9.12303686e-01 -2.02406377e-01 1.12011051e+00 -4.49352115e-01
1.06358910e+00 -1.80766761e-01 3.29907387e-01 4.53545183e-01
1.32154733e-01 5.96532643e-01 1.04561090e+00 -2.68651962e-01
-9.34729695e-01 -1.06311858e+00 -4.90257263e-01 8.66305053e-01
-3.30026634e-02 -4.48167890e-01 -2.06671700e-01 -8.11587214e-01
3.05768371e-01 5.07026196e-01 -8.95882487e-01 -9.99043137e-02
-2.63781697e-01 -5.78455806e-01 1.33509442e-01 3.70263904e-01
-6.21140748e-02 -1.01009905e+00 -6.03944182e-01 6.62532374e-02
5.00772357e-01 -4.99042094e-01 1.00197464e-01 6.88478470e-01
-1.09445226e+00 -1.04638028e+00 -4.83852178e-01 -5.43152392e-01
5.50438583e-01 -1.29538085e-02 1.20551348e+00 -2.46124119e-01
-2.50822097e-01 2.63990641e-01 -3.19047302e-01 -3.62084240e-01
-6.47702068e-02 -1.71759650e-01 3.24072182e-01 1.84448183e-01
3.52322042e-01 -6.18049741e-01 -1.18610285e-01 3.72813046e-01
-6.37410641e-01 -4.31537032e-01 6.69932365e-01 1.70266366e+00
8.94526839e-01 1.02311343e-01 4.40314353e-01 -1.40510809e+00
7.65829802e-01 -3.45989794e-01 -7.51449227e-01 3.33114147e-01
-1.07581949e+00 5.94168365e-01 9.52122569e-01 -4.87414420e-01
-5.97246706e-01 2.44600624e-01 6.28356516e-01 -3.20711762e-01
7.33973905e-02 6.14491582e-01 -2.57120371e-01 -9.47505310e-02
8.56469452e-01 2.15414818e-02 1.98390052e-01 -3.62568051e-01
3.72808069e-01 3.80225152e-01 1.05995173e-02 -2.68840373e-01
9.12766218e-01 5.94027579e-01 6.20707512e-01 -8.55675280e-01
-8.62289295e-02 -7.26414323e-01 -1.13279545e+00 3.67548279e-02
5.30957282e-01 -5.47028720e-01 -7.06065178e-01 -1.36464601e-02
-6.07628584e-01 2.95343548e-01 -6.35912120e-01 9.03926969e-01
-9.31200624e-01 -6.31584898e-02 2.62223394e-03 -7.69982874e-01
-2.35943079e-01 -1.03543365e+00 9.71956074e-01 -4.39904770e-03
-3.61398131e-01 -1.20930350e+00 2.73715019e-01 -2.76551574e-01
2.99552500e-01 3.79241645e-01 1.14848006e+00 -1.05426729e+00
-1.06827430e-01 -4.24400359e-01 2.47531742e-01 2.51617223e-01
6.18766136e-02 9.52237993e-02 -7.82850564e-01 -4.00959939e-01
8.54841620e-02 -2.29858458e-01 8.18002582e-01 2.94834435e-01
1.00293231e+00 1.09240308e-01 -2.45064259e-01 6.76040590e-01
1.50693703e+00 -1.21585995e-01 3.99739593e-01 3.49650741e-01
3.78511935e-01 7.48827040e-01 9.40736413e-01 9.09600973e-01
-4.59970415e-01 7.31031418e-01 3.25382091e-02 -1.81836575e-01
5.61142504e-01 4.97311056e-02 1.43818781e-01 5.91652513e-01
-2.21204773e-01 1.20418049e-01 -7.07854509e-01 1.65562391e-01
-1.75955272e+00 -1.16894054e+00 -4.25970823e-01 2.76784015e+00
5.09093225e-01 -5.60726738e-03 2.93063462e-01 3.67015570e-01
6.39473975e-01 -9.75772925e-03 -5.26585519e-01 -4.78549421e-01
-3.16013932e-01 5.82884192e-01 7.50344336e-01 4.25721586e-01
-1.18751335e+00 5.31687140e-01 6.53428745e+00 4.04436797e-01
-1.23436856e+00 -1.76155716e-01 3.55111510e-01 -1.21030271e-01
-4.44318831e-01 3.08435094e-02 -4.90320921e-01 1.05991460e-01
6.89793944e-01 -5.29679894e-01 3.09189916e-01 9.97731209e-01
1.25022039e-01 1.83206543e-01 -1.27835560e+00 1.07754660e+00
4.86974604e-02 -9.85468268e-01 -2.67279744e-01 1.71298578e-01
5.52502036e-01 -4.60323691e-01 3.12082499e-01 2.75934756e-01
-2.56050937e-02 -1.34527910e+00 9.67263654e-02 5.44585824e-01
7.90189385e-01 -9.22402084e-01 8.40204597e-01 -2.36464990e-03
-1.09969628e+00 -3.64315920e-02 -3.29210013e-01 1.28090596e-02
-1.49425492e-01 7.90092051e-01 -9.46334898e-01 5.48480392e-01
3.52779180e-01 9.60250080e-01 -7.34915793e-01 8.18581283e-01
-4.07933928e-02 5.14511228e-01 -8.68576020e-02 -2.85571635e-01
-2.30912253e-01 -7.26356030e-01 6.46116436e-01 1.03772366e+00
2.46056810e-01 -1.91582978e-01 2.86670886e-02 5.84761918e-01
3.30972970e-01 7.43066549e-01 -9.20236349e-01 1.34639576e-01
2.49028996e-01 1.41073751e+00 -6.01430357e-01 -3.74128193e-01
-5.35557032e-01 9.73560572e-01 2.12652609e-01 4.71703053e-01
-4.08315331e-01 -5.92251003e-01 7.99103320e-01 -2.22226441e-01
3.66345644e-01 2.83085424e-02 -6.43301725e-01 -1.34804535e+00
1.55878603e-01 -9.14437175e-01 2.88247049e-01 -1.12955242e-01
-1.25658417e+00 4.50160831e-01 -1.09251603e-01 -1.60452569e+00
-4.49481755e-01 -7.55324543e-01 -3.47769827e-01 1.09455073e+00
-9.62204397e-01 -5.57038963e-01 1.05782591e-01 8.79622281e-01
2.03582585e-01 -7.31416941e-01 1.20222151e+00 -1.16497934e-01
-1.74484506e-01 6.47468507e-01 5.13818145e-01 -2.58134305e-01
7.44797587e-01 -1.69307351e+00 -4.47045192e-02 1.13356411e-01
3.78244489e-01 7.40601599e-01 7.84746051e-01 -3.12087059e-01
-1.49773467e+00 -7.25086868e-01 6.94969118e-01 -3.49890769e-01
6.59238458e-01 -3.28520834e-01 -8.20122838e-01 2.71368742e-01
-2.42278188e-01 1.52715787e-01 1.16082656e+00 6.98228478e-01
-3.89216661e-01 -1.00945167e-01 -1.20698655e+00 2.65022963e-01
7.09324181e-01 -4.81050402e-01 -1.91095859e-01 5.02276182e-01
2.84876585e-01 1.77000880e-01 -1.25428009e+00 3.36005002e-01
8.73855412e-01 -1.14402759e+00 1.03988409e+00 -9.64359760e-01
1.40327871e-01 -1.76826105e-01 -1.37879640e-01 -1.60765529e+00
-4.16375190e-01 -4.33481365e-01 -5.08532673e-02 1.02344954e+00
7.60099590e-01 -7.91482687e-01 7.72916615e-01 3.96967053e-01
7.63792634e-01 -8.13028097e-01 -8.27723622e-01 -8.78138125e-01
-1.84681609e-01 -1.31583750e-01 4.95422721e-01 1.09412718e+00
3.37569386e-01 7.84669876e-01 -6.54164791e-01 -1.50238484e-01
5.83597898e-01 3.94192964e-01 7.84127116e-01 -1.73579073e+00
-7.50158906e-01 -4.85667348e-01 -1.21244943e+00 -7.63669461e-02
1.58829644e-01 -8.47296238e-01 -3.90492767e-01 -9.27546084e-01
1.96615741e-01 -4.79123205e-01 -5.15479982e-01 1.05622396e-01
-2.76722550e-01 -2.18994126e-01 1.60033926e-01 1.84070140e-01
-1.76132694e-01 4.78333712e-01 1.01790917e+00 1.21621722e-02
-5.49780786e-01 5.32570124e-01 -2.78961867e-01 5.08458197e-01
6.76147759e-01 -3.74526978e-01 -4.73823339e-01 4.90140140e-01
1.75829634e-01 4.52288598e-01 -1.65129714e-02 -8.28942239e-01
-1.81289569e-01 -4.41637933e-01 5.32482266e-01 -3.00035775e-01
3.21359664e-01 -1.06971419e+00 2.98834294e-01 4.36442018e-01
-7.02684760e-01 2.11859241e-01 -4.20897305e-01 6.64142013e-01
-4.22935367e-01 -2.58403629e-01 7.80494988e-01 -3.76051068e-02
-5.90916932e-01 1.98352098e-01 -1.81357563e-01 -1.00438349e-01
1.25893509e+00 -5.34845777e-02 1.95805132e-01 -3.89456540e-01
-7.99235821e-01 1.99910566e-01 6.66398168e-01 2.07517162e-01
8.01135063e-01 -1.40660155e+00 -7.95006990e-01 5.86527407e-01
5.41653275e-01 -5.63700557e-01 -5.02967358e-01 8.58652413e-01
-1.22672603e-01 3.00813049e-01 -3.30498636e-01 -5.87671578e-01
-1.68593168e+00 8.86538208e-01 -2.10763644e-02 -3.60157639e-01
-5.58455467e-01 6.86405480e-01 -1.10560820e-01 -5.62698305e-01
1.14838764e-01 -3.59379351e-02 -6.61531150e-01 6.48547530e-01
5.68809271e-01 4.05802310e-01 2.18986627e-02 -4.90774214e-01
-3.45541567e-01 6.02146447e-01 2.31272429e-01 -4.38257843e-01
1.45634353e+00 1.48944139e-01 -1.62170514e-01 1.00161707e+00
1.73238480e+00 2.02019259e-01 -6.50743961e-01 -3.20973396e-01
3.29418302e-01 -6.93198204e-01 -3.20834070e-01 -3.33651364e-01
-1.01314020e+00 8.21846902e-01 8.89062762e-01 2.60259807e-01
1.20202935e+00 -3.11494060e-02 3.69019993e-02 6.72959387e-01
5.48736274e-01 -1.09602952e+00 -1.16193384e-01 3.50577712e-01
7.83767104e-01 -1.45414603e+00 2.28231251e-01 -5.12025833e-01
-1.05824637e+00 1.56237519e+00 2.40301654e-01 -2.09872052e-01
7.77118206e-01 1.96672961e-01 1.59995537e-02 -2.41805404e-01
-8.61893415e-01 -2.95480013e-01 1.30449310e-01 7.11667061e-01
6.43313646e-01 5.16157866e-01 -6.89968526e-01 1.17725283e-01
-3.09729248e-01 -3.94571036e-01 5.69283366e-01 5.81146240e-01
-2.21258774e-01 -1.39204693e+00 -4.63586748e-01 9.08554316e-01
-3.45387422e-02 -1.08366877e-01 -5.23010433e-01 7.66009212e-01
-3.47173452e-01 6.14071369e-01 -1.68349266e-01 -6.81838810e-01
4.76079822e-01 1.86396882e-01 1.45716265e-01 -6.32035732e-01
-6.74975634e-01 -1.37442946e-01 -6.38870969e-02 -5.22456110e-01
-3.36843103e-01 -1.01818573e+00 -1.11161935e+00 1.84397295e-01
-5.45745790e-01 2.30542734e-01 7.34559834e-01 5.54801881e-01
5.86526506e-02 1.67350590e-01 1.21722567e+00 -4.22094047e-01
-1.03127015e+00 -7.43314564e-01 -1.17750943e+00 7.60127187e-01
3.14820766e-01 -9.51746047e-01 -3.83797944e-01 -9.00621787e-02] | [7.914144992828369, 4.37347412109375] |
aebbed39-723e-462d-a2e8-88c51b7dfcd0 | implementing-a-portable-clinical-nlp-system | 1811.06179 | null | http://arxiv.org/abs/1811.06179v1 | http://arxiv.org/pdf/1811.06179v1.pdf | Implementing a Portable Clinical NLP System with a Common Data Model - a Lisp Perspective | This paper presents a Lisp architecture for a portable NLP system, termed
LAPNLP, for processing clinical notes. LAPNLP integrates multiple standard,
customized and in-house developed NLP tools. Our system facilitates portability
across different institutions and data systems by incorporating an enriched
Common Data Model (CDM) to standardize necessary data elements. It utilizes
UMLS to perform domain adaptation when integrating generic domain NLP tools. It
also features stand-off annotations that are specified by positional reference
to the original document. We built an interval tree based search engine to
efficiently query and retrieve the stand-off annotations by specifying
positional requirements. We also developed a utility to convert an inline
annotation format to stand-off annotations to enable the reuse of clinical text
datasets with inline annotations. We experimented with our system on several
NLP facilitated tasks including computational phenotyping for lymphoma patients
and semantic relation extraction for clinical notes. These experiments
showcased the broader applicability and utility of LAPNLP. | ['Yuan Luo', 'Peter Szolovits'] | 2018-11-15 | null | null | null | null | ['computational-phenotyping'] | ['medical'] | [ 1.85610741e-01 5.02415955e-01 -4.40649152e-01 -6.96966767e-01
-1.07156289e+00 -6.48658276e-01 3.55347358e-02 8.82071972e-01
-2.42575228e-01 1.17097127e+00 4.29915696e-01 -5.66881120e-01
-7.16870964e-01 -6.29288733e-01 -1.66791864e-02 -2.37458527e-01
1.55711830e-01 1.14824378e+00 4.16261584e-01 2.76179105e-01
-1.47634313e-01 7.20664859e-01 -1.06293786e+00 8.69906962e-01
8.56348276e-01 7.25685537e-01 2.96851605e-01 5.21966398e-01
-7.15080142e-01 8.25353920e-01 -6.25419915e-01 -8.54462460e-02
-5.96624166e-02 -3.37887675e-01 -1.11491966e+00 -3.65912735e-01
-3.40287745e-01 2.64663786e-01 3.46723348e-01 5.62378764e-01
6.62127137e-01 -3.01373541e-01 3.46152842e-01 -1.26351535e+00
-3.68531138e-01 6.16390765e-01 1.50311051e-03 -6.41348660e-02
1.03363049e+00 -2.12098494e-01 7.85849631e-01 -3.78182828e-01
1.41800094e+00 7.35426664e-01 9.18502986e-01 4.67121452e-01
-9.62657094e-01 -5.80271363e-01 -5.54230273e-01 1.77548593e-03
-1.58798480e+00 -2.26111814e-01 4.86644246e-02 -4.29656982e-01
1.52283370e+00 7.25072861e-01 6.11873507e-01 5.36141753e-01
3.01732868e-01 3.29047889e-01 8.11695755e-01 -6.90792799e-01
3.11812848e-01 2.95123488e-01 3.94734979e-01 8.41715395e-01
2.38421306e-01 -3.95183474e-01 -4.91046667e-01 -7.82061815e-01
5.25213838e-01 -1.86445430e-01 -1.27158567e-01 -1.40639067e-01
-1.05480230e+00 4.99319553e-01 -2.72719204e-01 5.20280540e-01
-3.75331432e-01 -2.69657046e-01 7.53500402e-01 -8.59274063e-03
1.44407406e-01 4.23815876e-01 -1.02341509e+00 -2.65317746e-02
-9.52178061e-01 -5.44048622e-02 1.30824757e+00 1.55428517e+00
2.91078836e-01 -8.95160913e-01 -4.21502024e-01 9.87389863e-01
2.73112953e-01 -2.24248856e-01 7.87657678e-01 -9.92483556e-01
-4.73361835e-03 9.99315381e-01 2.78019935e-01 -3.53816211e-01
-1.08339334e+00 -3.79279628e-02 -2.52190113e-01 -3.60054046e-01
3.26927781e-01 -1.44128263e-01 -8.19718063e-01 1.37100899e+00
4.31783110e-01 -3.84926945e-01 3.97191793e-01 3.31262499e-01
1.02989542e+00 2.08693743e-01 6.41994953e-01 -5.52402198e-01
1.90103102e+00 -5.23211658e-01 -1.12541151e+00 4.07764107e-01
1.37022924e+00 -8.98267448e-01 7.08014369e-01 2.58151472e-01
-8.70196223e-01 1.25894338e-01 -6.52691901e-01 -3.92543972e-01
-8.92279029e-01 9.89744291e-02 6.52015805e-01 6.15060329e-01
-1.02011263e+00 3.28480363e-01 -9.43309426e-01 -1.22503960e+00
3.58134925e-01 4.21432108e-01 -5.79301775e-01 5.81217222e-02
-1.18098032e+00 1.02173841e+00 9.57900167e-01 -5.54509699e-01
-1.07105263e-01 -1.23396480e+00 -7.19592214e-01 -6.05124198e-02
3.22931409e-01 -1.10700810e+00 1.33436441e+00 -1.75620735e-01
-1.23139203e+00 1.17940140e+00 -1.61295086e-01 -3.95077527e-01
2.23696753e-01 3.11121643e-01 -8.24987888e-01 3.80996436e-01
3.76338154e-01 5.59477568e-01 -3.36968422e-01 -4.56278563e-01
-7.96783507e-01 -2.40053818e-01 -4.24590081e-01 -6.07528687e-02
1.25711694e-01 5.77684760e-01 -5.79735816e-01 -6.01303935e-01
-7.13459253e-02 -4.67727333e-01 -2.08362028e-01 1.18055448e-01
-2.81078398e-01 -4.09434468e-01 5.44137895e-01 -7.71161139e-01
1.56965625e+00 -1.70870483e+00 -5.49959242e-01 3.32605988e-01
-2.48749405e-02 -1.07146474e-02 2.83058584e-01 9.72086132e-01
-3.18749607e-01 2.02197254e-01 -7.73752630e-02 2.67023206e-01
1.47698000e-01 6.43070936e-01 1.29389375e-01 -4.23318855e-02
-1.27618834e-02 8.84606063e-01 -8.01301241e-01 -1.16569340e+00
-9.98679549e-03 4.31846887e-01 -4.46649283e-01 1.12499502e-02
-5.37253559e-01 2.52207160e-01 -5.42878211e-01 9.66340005e-01
4.08942223e-01 -5.29008031e-01 7.04145968e-01 -8.73584971e-02
-1.81796104e-01 3.67586583e-01 -8.92997921e-01 2.19373631e+00
-2.59849370e-01 1.69511214e-02 -2.84720049e-03 -3.50680918e-01
9.67136681e-01 8.59523535e-01 8.79991055e-01 -1.50127262e-01
-1.28118947e-01 2.84243703e-01 -3.99358213e-01 -1.09126318e+00
2.49478281e-01 -2.07297325e-01 -1.37700930e-01 4.77957577e-01
1.84223920e-01 -2.49898117e-02 4.57413644e-01 8.48364457e-02
1.47533000e+00 3.88271868e-01 1.36311424e+00 -4.08246309e-01
7.09420860e-01 5.68129957e-01 7.40487754e-01 4.79454219e-01
5.19056953e-02 2.13383898e-01 6.69895470e-01 -5.16275585e-01
-6.99889004e-01 -8.14915419e-01 -7.75931239e-01 9.16538835e-01
-6.85698569e-01 -9.74462509e-01 -3.48952949e-01 -8.64486873e-01
-2.17493519e-01 7.81205833e-01 -3.11360776e-01 5.96442819e-01
-1.49680376e-01 -8.19903791e-01 9.53572154e-01 6.43359125e-01
2.74609663e-02 -1.05471730e+00 -8.78074050e-01 5.78013182e-01
-1.66141838e-01 -1.14162004e+00 -2.84365237e-01 4.17709231e-01
-6.04644299e-01 -1.41656399e+00 -3.65143955e-01 -8.63536179e-01
6.76054537e-01 -8.46752703e-01 1.09141219e+00 -2.46014684e-01
-7.11668432e-01 4.41941530e-01 -4.74644065e-01 -7.39727795e-01
-5.01092374e-01 1.25152066e-01 -3.65589589e-01 -8.76342535e-01
8.61194730e-01 -3.32173288e-01 -1.57826409e-01 2.86198378e-01
-1.21856511e+00 3.20540480e-02 3.11668813e-01 5.95294654e-01
9.57764268e-01 -1.60301208e-01 7.82838047e-01 -1.17724431e+00
7.89875686e-01 -4.10804778e-01 -5.56376815e-01 7.07084835e-01
-7.94336259e-01 2.13029593e-01 3.16086143e-01 1.35036379e-01
-1.08331060e+00 3.96590650e-01 -5.28581440e-01 2.98363894e-01
-5.19336879e-01 8.68733704e-01 -2.88389087e-01 3.00708145e-01
6.04266584e-01 -7.36795589e-02 -5.53823833e-04 -5.86394072e-01
1.72057807e-01 8.88433635e-01 4.57150042e-01 -5.63271999e-01
-1.43751621e-01 2.22417369e-01 1.82661593e-01 -4.42634135e-01
-5.36742806e-01 -6.19535744e-01 -7.75005102e-01 4.32278693e-01
9.64751601e-01 -5.84258914e-01 -6.37691379e-01 -3.09244692e-01
-1.04435110e+00 -8.85037258e-02 -7.28541374e-01 4.85588193e-01
-6.73912227e-01 3.41108173e-01 -5.71117163e-01 -2.23923445e-01
-5.80232978e-01 -7.97970474e-01 8.59600723e-01 -3.74194644e-02
-9.76547301e-01 -1.08767176e+00 4.38967049e-01 3.96945663e-02
2.40257126e-03 3.01740795e-01 1.45438278e+00 -1.52080357e+00
-4.77508418e-02 -3.48722786e-01 -3.01279753e-01 -3.24299753e-01
3.29752088e-01 -2.65908446e-02 -6.10576510e-01 2.89689690e-01
-3.52973193e-01 1.07623369e-01 5.62312379e-02 1.52044132e-01
1.18082058e+00 -3.61503333e-01 -1.03292918e+00 6.22070312e-01
1.40389848e+00 5.93058407e-01 5.34967840e-01 3.70595813e-01
-3.40902247e-02 7.89591014e-01 7.49343872e-01 4.68633324e-01
2.90726095e-01 5.91425598e-01 -2.93432176e-01 5.88615537e-02
2.86110938e-02 -1.76686764e-01 -3.98365259e-01 4.46781248e-01
1.73011139e-01 -2.66150147e-01 -1.40764070e+00 4.51384664e-01
-1.82828093e+00 -3.40093672e-01 -4.72069494e-02 1.79776990e+00
1.32175350e+00 -2.96483994e-01 -8.92058387e-02 -1.53986007e-01
3.94356012e-01 -6.43648386e-01 -2.29280233e-01 -5.91682792e-01
9.56467390e-02 6.64063573e-01 3.58496457e-01 4.47428018e-01
-5.45290887e-01 7.04330385e-01 7.04419231e+00 7.20138550e-01
-5.21556020e-01 2.69192636e-01 1.51831722e-02 2.94948742e-02
-9.46956873e-02 -9.25748423e-02 -1.00645983e+00 2.70142257e-01
1.20176232e+00 -5.07593870e-01 -3.65360916e-01 7.54728794e-01
3.61545593e-01 -8.11791867e-02 -1.36016691e+00 6.76070035e-01
-2.01814830e-01 -1.71079469e+00 -7.20354617e-02 -5.31618632e-02
1.36852279e-01 -1.06777914e-01 -7.45311677e-01 -4.62033935e-02
4.84441638e-01 -8.28553915e-01 -1.88982952e-02 6.48356855e-01
1.21415007e+00 -3.78622562e-01 9.95293915e-01 -3.87199558e-02
-1.03752887e+00 2.51623571e-01 -1.24006890e-01 3.21021616e-01
2.21578196e-01 6.95667624e-01 -1.83632016e+00 1.09386384e+00
5.39247930e-01 4.59712356e-01 -3.48754466e-01 1.15076315e+00
-2.70675272e-01 -1.61698982e-02 -4.78076607e-01 1.16865359e-01
-1.98287591e-01 9.44020301e-02 2.91666180e-01 1.74003637e+00
4.86224383e-01 3.39485854e-01 4.96317670e-02 4.22309607e-01
1.65484458e-01 6.24183655e-01 -3.71309459e-01 -1.44150615e-01
6.44793808e-01 1.28090644e+00 -8.97348523e-01 -5.09112656e-01
-3.02189767e-01 5.12789726e-01 -2.00518593e-01 1.40726864e-01
-7.69787848e-01 -8.27350974e-01 4.43002135e-01 2.20861733e-01
1.81185022e-01 2.72344112e-01 -1.48187026e-01 -7.19807267e-01
-2.78180957e-01 -7.22426355e-01 1.31998670e+00 -1.18370020e+00
-1.23365963e+00 6.85340405e-01 2.99483180e-01 -9.95052278e-01
-4.48765457e-01 -6.23278558e-01 -9.33527648e-02 9.48250771e-01
-1.01711297e+00 -1.26355195e+00 -1.80184364e-01 6.28965318e-01
5.85719682e-02 1.78155392e-01 1.61564124e+00 3.80329192e-01
-4.59982425e-01 3.83517325e-01 -1.88785329e-01 9.36081931e-02
1.11880422e+00 -1.08731151e+00 -3.94163370e-01 1.89313427e-01
-3.87424588e-01 7.29661644e-01 3.80530804e-01 -9.83899951e-01
-7.79760242e-01 -1.16549718e+00 1.54056680e+00 -5.09334564e-01
6.97282732e-01 -7.20970184e-02 -8.64482343e-01 1.06315362e+00
2.11748987e-01 -1.18690498e-01 1.78502846e+00 -9.85258296e-02
-2.44754091e-01 -8.32657702e-03 -1.72582662e+00 7.51295611e-02
8.95460844e-01 -3.56476903e-01 -1.06656814e+00 7.67729402e-01
8.29675913e-01 -4.21123862e-01 -1.56852138e+00 3.81530493e-01
6.47322595e-01 -4.41607654e-01 7.07205415e-01 -7.69069850e-01
-2.38350872e-02 -5.05737126e-01 3.55192162e-02 -7.70992935e-01
-8.33747089e-02 -7.03812003e-01 1.04269236e-01 1.25534964e+00
8.07322621e-01 -8.77487004e-01 6.56420171e-01 9.51585710e-01
-2.57285953e-01 -7.21671343e-01 -8.83456409e-01 -5.39957404e-01
-3.98892581e-01 -5.11669934e-01 9.98163402e-01 1.20723891e+00
8.88162851e-01 4.92707863e-02 4.28101957e-01 2.91621804e-01
4.67419326e-02 3.76069658e-02 1.07207172e-01 -1.34847736e+00
-2.14712754e-01 7.97132999e-02 -3.91786903e-01 -2.83102930e-01
-1.05977652e-03 -1.32612813e+00 -4.00920779e-01 -2.04952097e+00
1.39404088e-01 -6.37671471e-01 -2.30513945e-01 1.28229582e+00
4.82086062e-01 -4.65017632e-02 -3.50155354e-01 1.21246479e-01
-4.89745557e-01 -5.22123218e-01 5.96377909e-01 2.95532227e-01
-5.42805552e-01 -3.57696295e-01 -7.84808278e-01 7.35968351e-01
8.16480935e-01 -1.00103688e+00 -3.65536273e-01 -1.32036477e-01
5.12834966e-01 2.81208754e-01 -8.74501392e-02 -8.21991026e-01
6.09745622e-01 -2.73362279e-01 4.15621907e-01 -1.06721044e+00
-7.31186867e-02 -1.08317232e+00 8.40252936e-01 5.71444511e-01
-4.49219435e-01 -4.75670174e-02 4.72956181e-01 1.44673511e-01
-8.85449722e-02 -5.06175458e-01 4.82516319e-01 -3.44875276e-01
-4.71750677e-01 -4.93821464e-02 -5.59994578e-01 -5.66414110e-02
1.35426569e+00 -2.74412483e-01 -4.36003953e-01 1.98517218e-01
-1.31446695e+00 5.34656048e-01 5.04125655e-01 -1.30764525e-02
3.17989051e-01 -9.46880996e-01 -3.97010833e-01 1.37525246e-01
4.14323658e-01 -5.37872091e-02 7.80226514e-02 9.85329270e-01
-1.01610267e+00 1.16115785e+00 -3.20827186e-01 -4.37795162e-01
-1.47933066e+00 8.34483147e-01 1.72608897e-01 -7.53266513e-01
-6.12859070e-01 5.85817158e-01 -2.69380450e-01 -6.41178012e-01
1.63960725e-01 -7.83236206e-01 -1.17897406e-01 2.21664175e-01
6.64022267e-01 2.18333289e-01 5.25462210e-01 7.66365305e-02
-8.34318697e-01 2.98880726e-01 5.88331856e-02 -1.01551093e-01
1.32780099e+00 1.07107358e-03 -5.32588363e-01 2.74200767e-01
8.23988557e-01 2.43261501e-01 -2.20182940e-01 -4.08511423e-02
6.54955268e-01 1.44067779e-01 -4.54972029e-01 -1.60957313e+00
-4.00541812e-01 5.46603650e-02 1.43782690e-01 -2.14500222e-02
1.32484138e+00 2.77567178e-01 3.57095093e-01 5.25872767e-01
3.49213779e-01 -9.24020767e-01 -8.86909127e-01 2.45580614e-01
6.73917830e-01 -5.06020844e-01 2.39254072e-01 -7.65611768e-01
-5.19987345e-01 1.33733177e+00 4.00501564e-02 7.55425930e-01
7.09126055e-01 8.41128826e-01 2.28834927e-01 -4.77237880e-01
-1.09174895e+00 -3.19579393e-02 7.19880220e-03 9.66169775e-01
7.35262275e-01 -1.60514831e-03 -1.02601647e+00 1.05452645e+00
-1.29850015e-01 1.00401688e+00 3.08456868e-01 1.23202908e+00
-1.66297078e-01 -1.75193644e+00 -4.71590817e-01 5.21320343e-01
-8.31610382e-01 -2.10923910e-01 -4.72401828e-01 1.03770828e+00
5.87176740e-01 6.06352329e-01 4.94130626e-02 2.55500197e-01
2.37766877e-01 8.54020178e-01 4.26290095e-01 -1.04200327e+00
-7.84449637e-01 1.74048007e-01 5.87513924e-01 -5.34814775e-01
-5.90237260e-01 -5.75345695e-01 -1.81070328e+00 4.44659472e-01
4.74205948e-02 6.01516843e-01 6.06682241e-01 6.74761415e-01
7.58742511e-01 7.13167131e-01 -3.79131377e-01 4.14213330e-01
5.27643673e-02 -8.26400161e-01 -5.55835962e-01 -7.09004775e-02
-2.11218283e-01 -1.95714831e-01 2.88027376e-01 4.68115419e-01] | [8.530489921569824, 8.672755241394043] |
3d63e5dd-9ed8-492d-aad0-423e4a4f00fa | il-mcam-an-interactive-learning-and-multi | 2206.03368 | null | https://arxiv.org/abs/2206.03368v1 | https://arxiv.org/pdf/2206.03368v1.pdf | IL-MCAM: An interactive learning and multi-channel attention mechanism-based weakly supervised colorectal histopathology image classification approach | In recent years, colorectal cancer has become one of the most significant diseases that endanger human health. Deep learning methods are increasingly important for the classification of colorectal histopathology images. However, existing approaches focus more on end-to-end automatic classification using computers rather than human-computer interaction. In this paper, we propose an IL-MCAM framework. It is based on attention mechanisms and interactive learning. The proposed IL-MCAM framework includes two stages: automatic learning (AL) and interactivity learning (IL). In the AL stage, a multi-channel attention mechanism model containing three different attention mechanism channels and convolutional neural networks is used to extract multi-channel features for classification. In the IL stage, the proposed IL-MCAM framework continuously adds misclassified images to the training set in an interactive approach, which improves the classification ability of the MCAM model. We carried out a comparison experiment on our dataset and an extended experiment on the HE-NCT-CRC-100K dataset to verify the performance of the proposed IL-MCAM framework, achieving classification accuracies of 98.98% and 99.77%, respectively. In addition, we conducted an ablation experiment and an interchangeability experiment to verify the ability and interchangeability of the three channels. The experimental results show that the proposed IL-MCAM framework has excellent performance in the colorectal histopathological image classification tasks. | ['Marcin Grzegorzek', 'Xinyu Huang', 'Hongzan Sun', 'Changhao Sun', 'Wanli Liu', 'Yixin Li', 'Weiming Hu', 'Md Mamunur Rahaman', 'Xiaoyan Li', 'Chen Li', 'HaoYuan Chen'] | 2022-06-07 | null | null | null | null | ['histopathological-image-classification'] | ['medical'] | [ 8.85097757e-02 1.16922997e-01 -2.53456272e-02 -2.76336998e-01
-6.91573560e-01 -1.41938493e-01 3.52922857e-01 3.66563201e-01
-8.65931630e-01 3.70230079e-01 -9.93117839e-02 -6.02132559e-01
-3.25443409e-02 -5.22428751e-01 -5.44213951e-01 -8.82168770e-01
-3.60876471e-02 1.20242164e-01 2.32998028e-01 8.44790190e-02
2.08666742e-01 2.02226475e-01 -1.09718072e+00 6.84027493e-01
1.13004768e+00 9.49673355e-01 2.99421430e-01 7.92747438e-01
-1.47223637e-01 8.49541306e-01 -3.58172238e-01 -2.65681416e-01
-7.12889880e-02 -3.15440387e-01 -8.21299434e-01 -7.89936110e-02
-1.67924970e-01 -1.93140417e-01 -5.42367287e-02 1.03870571e+00
5.55508077e-01 -2.46032059e-01 5.03198206e-01 -9.87029791e-01
-4.58604813e-01 4.20952857e-01 -7.20315039e-01 2.95432091e-01
1.83122344e-02 2.83988476e-01 4.58766520e-01 -7.33344615e-01
3.30958694e-01 9.63728130e-01 6.74069345e-01 3.91623199e-01
-9.19735134e-01 -8.05114090e-01 1.13050334e-01 3.30319405e-01
-1.38703763e+00 -1.36047110e-01 6.06862187e-01 -5.03490865e-01
6.79642558e-01 4.41908628e-01 7.01790869e-01 5.99747658e-01
7.38237381e-01 9.51387465e-01 1.07753253e+00 -5.10905504e-01
-6.32131565e-03 4.38491434e-01 2.41745427e-01 6.77965879e-01
2.21036509e-01 -7.33043114e-03 1.11923449e-01 -3.24728154e-03
8.40192795e-01 1.91834554e-01 -3.98891628e-01 -5.03500253e-02
-1.25277758e+00 8.37435842e-01 1.03465068e+00 4.60738271e-01
-7.69751295e-02 -3.88054438e-02 6.40015781e-01 9.00338367e-02
2.73675114e-01 3.76232326e-01 -8.44694823e-02 3.35455805e-01
-5.50162315e-01 -3.79059106e-01 4.06719387e-01 5.79046607e-01
1.40906870e-01 -5.44477701e-01 -3.51408660e-01 6.69344366e-01
4.90306884e-01 1.72504500e-01 1.03238571e+00 -5.63735254e-02
1.33302122e-01 9.02911425e-01 -1.27378508e-01 -7.77997375e-01
-7.02211499e-01 -6.58481061e-01 -1.29023969e+00 2.29158744e-01
2.10452378e-01 -1.22122936e-01 -9.83900785e-01 1.40092599e+00
1.63214788e-01 1.58604518e-01 1.73121870e-01 9.16285574e-01
1.19653189e+00 4.05901521e-01 4.65101391e-01 -1.14407372e-02
1.30779576e+00 -1.26358235e+00 -8.04540694e-01 2.11035550e-01
9.83438313e-01 -5.97593963e-01 1.05674005e+00 4.16296124e-02
-7.69232213e-01 -6.08134866e-01 -1.07099855e+00 -4.33630832e-02
-2.76819259e-01 6.49757326e-01 6.85810983e-01 3.84176701e-01
-8.75944197e-01 8.10648128e-02 -1.02299535e+00 -2.60612160e-01
5.20940423e-01 6.02355421e-01 -4.55720305e-01 1.19547263e-01
-1.15635741e+00 5.63836873e-01 4.17936921e-01 4.20077264e-01
-7.65747368e-01 -4.05444145e-01 -6.93527222e-01 1.92228273e-01
-4.71466444e-02 -4.76972342e-01 1.29449153e+00 -1.46627879e+00
-1.45302296e+00 9.06147778e-01 1.72261074e-01 -2.86318213e-01
6.19052231e-01 -2.56122574e-02 -3.81648153e-01 9.27046612e-02
-5.61817065e-02 6.77572012e-01 3.19746554e-01 -9.68714952e-01
-6.48456573e-01 -3.01653713e-01 -2.19708178e-02 3.03841382e-01
-3.99403960e-01 -3.36290359e-01 -6.06637299e-01 -4.71767604e-01
-1.14565678e-01 -9.22715425e-01 -3.03795278e-01 1.62290648e-01
-4.27353293e-01 -7.21027851e-02 6.21541321e-01 -6.02730155e-01
1.13114369e+00 -2.39682722e+00 -1.61120906e-01 2.86566019e-01
2.27575019e-01 5.63162029e-01 -8.07247236e-02 -1.02765970e-01
-1.90703988e-01 2.06778288e-01 -6.34809583e-02 -9.13541541e-02
-4.76219386e-01 -2.59129554e-01 4.37618494e-01 5.52145779e-01
2.11339623e-01 1.02662373e+00 -8.53874803e-01 -7.90336549e-01
2.75517315e-01 4.00309533e-01 -5.66773772e-01 3.47993582e-01
1.82022706e-01 6.79460168e-01 -2.92644322e-01 5.28949857e-01
6.62427545e-01 -5.01351416e-01 2.70402819e-01 -3.14080536e-01
-7.68259317e-02 -4.67328787e-01 -6.97666883e-01 1.62204921e+00
-6.39183819e-01 6.80286884e-01 -2.58287285e-02 -8.27975392e-01
7.36562669e-01 3.84251773e-01 1.76334083e-01 -8.82597923e-01
4.65440810e-01 2.33623534e-01 5.29503822e-01 -8.63622069e-01
2.51714233e-02 1.44721851e-01 8.50719959e-02 -4.92547117e-02
-1.27807055e-02 5.38789034e-01 -1.28288686e-01 3.02515645e-02
8.66863251e-01 -4.31400001e-01 5.62353313e-01 -5.84530950e-01
1.03590536e+00 7.17962831e-02 4.82369006e-01 6.21137023e-01
-2.62347460e-01 2.55608112e-01 4.07060981e-01 -5.43256104e-01
-6.15608931e-01 -6.77881598e-01 -3.52698833e-01 5.35072684e-01
5.01405478e-01 2.84996554e-02 -7.06382155e-01 -1.08810771e+00
-2.78724670e-01 3.59487504e-01 -8.80714655e-01 -2.36888364e-01
-1.81875870e-01 -1.09103811e+00 3.62351567e-01 5.38823485e-01
9.45638657e-01 -1.17878795e+00 -6.62631691e-01 8.60598013e-02
-1.32252276e-01 -6.03648126e-01 -3.89476806e-01 1.07929066e-01
-6.43198013e-01 -1.26090848e+00 -6.94949925e-01 -1.17572117e+00
1.15511250e+00 1.59574255e-01 7.40964413e-01 6.03524625e-01
-5.43720007e-01 -1.77145883e-01 -3.59497100e-01 -6.49244428e-01
-4.03684467e-01 2.53374398e-01 -5.92310369e-01 3.20134670e-01
2.35823736e-01 -8.29709694e-02 -9.42389607e-01 3.51637095e-01
-8.54446292e-01 5.06638646e-01 1.28323543e+00 1.11424768e+00
6.16349161e-01 -6.03246177e-03 5.27391255e-01 -1.29425073e+00
3.88332993e-01 -4.67211813e-01 -5.07277131e-01 2.08792821e-01
-3.22123736e-01 -1.02526240e-01 6.91891134e-01 -4.86646414e-01
-1.04338145e+00 2.84704685e-01 -4.12514925e-01 -6.76224828e-02
-3.43064368e-02 8.10438216e-01 -3.04182410e-01 -2.35701859e-01
5.49657047e-01 1.12430699e-01 1.50759444e-01 -1.94541603e-01
-1.45442069e-01 1.05636990e+00 4.35456753e-01 1.15917012e-01
2.99396694e-01 2.75002480e-01 -3.16495895e-01 -6.84997201e-01
-5.03394365e-01 -5.21650791e-01 -3.07733446e-01 -4.05776799e-01
1.02093625e+00 -9.95741904e-01 -8.98622870e-01 9.14364934e-01
-9.56747591e-01 -4.25452441e-01 2.05328047e-01 6.54763341e-01
-9.89569575e-02 1.51595354e-01 -8.61321211e-01 -5.31902075e-01
-6.91405952e-01 -1.60349154e+00 8.97232413e-01 5.52397609e-01
-1.12686427e-02 -1.06926990e+00 -1.99938998e-01 2.93840170e-01
5.63724875e-01 2.84776360e-01 1.02758145e+00 -8.29903066e-01
-2.74722189e-01 -4.95656282e-01 -4.81872439e-01 -1.23713993e-01
2.72782389e-02 -1.32761940e-01 -9.89324868e-01 -4.70325649e-01
-1.76779002e-01 -2.79264152e-01 8.24894071e-01 3.23817343e-01
1.60681915e+00 -1.16956763e-01 -7.77617514e-01 6.10725582e-01
1.61778486e+00 6.26122534e-01 9.93144512e-01 3.97613108e-01
4.43199098e-01 1.07918568e-01 5.10054469e-01 -4.38702554e-02
1.74221456e-01 4.22084510e-01 5.71080804e-01 -8.89823020e-01
2.90646516e-02 2.42128801e-02 -1.96614675e-02 7.38835812e-01
7.03631118e-02 -3.95482302e-01 -1.08060312e+00 4.49718177e-01
-1.84113598e+00 -6.44115269e-01 -2.93317527e-01 2.22385025e+00
7.49997735e-01 2.29873389e-01 -3.89039248e-01 1.50117323e-01
7.89341211e-01 -4.86662179e-01 -5.23154438e-01 -2.36373931e-01
3.85051727e-01 -6.72840476e-02 2.92270184e-01 3.88079911e-01
-1.39515889e+00 2.51609027e-01 5.21559572e+00 8.54837537e-01
-1.41657233e+00 1.00804910e-01 1.02770841e+00 2.71803379e-01
9.45380256e-02 -4.51231092e-01 -5.15638947e-01 6.61124349e-01
5.53899944e-01 6.45995811e-02 -2.32798740e-01 7.62557149e-01
1.05066724e-01 -2.93227136e-01 -1.04559529e+00 9.99020159e-01
-7.61460289e-02 -1.24480569e+00 1.18838027e-01 2.37597153e-02
5.26003420e-01 -3.17232907e-01 6.61677867e-02 4.28298265e-01
-1.39237717e-01 -1.02746081e+00 1.57524362e-01 7.41832435e-01
9.48121190e-01 -8.43400717e-01 1.58516419e+00 5.16952276e-01
-1.04977381e+00 -1.35422468e-01 1.97174456e-02 1.41062155e-01
-4.03496832e-01 4.88317847e-01 -8.66457403e-01 6.59899712e-01
5.93338430e-01 4.82122838e-01 -8.63827169e-01 1.36341190e+00
-1.28738031e-01 3.84102643e-01 6.01195730e-02 -2.60228723e-01
2.37069651e-01 2.45155066e-01 7.00868666e-02 1.41367066e+00
-2.40650191e-03 1.99577976e-02 1.54410541e-01 5.04952490e-01
-1.37352675e-01 3.86524290e-01 -7.49836937e-02 2.28206381e-01
1.80404663e-01 1.61821699e+00 -9.93934453e-01 -3.68543327e-01
-4.25096333e-01 9.54675674e-01 3.83484721e-01 3.84925939e-02
-1.16922462e+00 -7.97800303e-01 -3.79506089e-02 9.85353291e-02
-1.41360566e-01 2.77560145e-01 -1.26507074e-01 -9.70948458e-01
-3.53313953e-01 -7.81903625e-01 4.47206408e-01 -4.88480061e-01
-1.16678095e+00 7.33559787e-01 -4.81151819e-01 -1.30933809e+00
3.44356507e-01 -4.39743698e-01 -9.63651121e-01 8.10535491e-01
-1.41388071e+00 -1.11570370e+00 -8.96231353e-01 4.78281587e-01
5.63584387e-01 2.05037724e-02 8.82397771e-01 3.80166590e-01
-7.64048576e-01 1.08628857e+00 2.07125112e-01 4.74894911e-01
6.12818241e-01 -1.20135701e+00 -2.16247022e-01 5.21443129e-01
-4.39346969e-01 4.08305824e-01 -4.34734002e-02 -4.09715712e-01
-9.63049114e-01 -1.50382972e+00 4.50491160e-01 1.25624478e-01
3.69018465e-02 -3.01552683e-01 -8.19698572e-01 5.79458773e-01
2.27217481e-01 1.55056164e-01 8.87618184e-01 -2.42252484e-01
3.76779646e-01 1.82778277e-02 -1.18422699e+00 5.08162439e-01
3.53436202e-01 -4.07420732e-02 -1.05416372e-01 2.92687327e-01
5.24339795e-01 -6.02231383e-01 -7.21729755e-01 6.48053229e-01
6.28188670e-01 -6.98379159e-01 5.90564132e-01 -2.74257481e-01
5.52611291e-01 -4.74472255e-01 2.75683373e-01 -1.19204044e+00
-5.01757860e-01 1.14625461e-01 3.47395718e-01 9.98978853e-01
5.69392920e-01 -5.96604049e-01 5.60419858e-01 3.05336535e-01
-1.66057721e-01 -1.15409732e+00 -6.69020295e-01 -3.50903362e-01
2.74112988e-02 6.57974184e-02 2.52517760e-01 9.35296595e-01
4.96655852e-02 3.85569930e-01 3.93093154e-02 2.00503752e-01
2.93073326e-01 9.91235375e-02 5.34294009e-01 -9.13595378e-01
-2.94316202e-01 -4.16246057e-01 -6.79002523e-01 -6.89052701e-01
-1.79080293e-01 -1.15291238e+00 -9.52353422e-03 -1.42066133e+00
7.42672026e-01 -4.92963791e-01 -5.69037497e-01 4.59402770e-01
-5.19298613e-01 3.28470081e-01 -1.15837149e-01 1.61330566e-01
-7.75710821e-01 3.82194698e-01 1.30549502e+00 -3.68565381e-01
-2.97504038e-01 -3.62994149e-02 -4.85983759e-01 6.55444920e-01
7.85404980e-01 -1.18813574e-01 -2.58400738e-01 -3.83046061e-01
-2.97108442e-01 7.20435381e-02 3.91116142e-01 -1.23397803e+00
3.82837325e-01 1.81189612e-01 9.51483846e-01 -3.49846095e-01
-2.02542707e-01 -7.97148168e-01 2.51235247e-01 1.08490121e+00
-4.68780518e-01 -3.75982881e-01 4.78817075e-01 4.78329659e-01
-3.54611844e-01 -2.70098969e-02 1.02931368e+00 -1.23869509e-01
-6.47747219e-01 2.78830647e-01 -5.55384219e-01 -4.34169799e-01
1.36719429e+00 1.83075443e-02 -2.58978575e-01 -2.81391684e-02
-9.06663418e-01 5.67726493e-01 2.45008588e-01 2.87385911e-01
5.39305985e-01 -1.29318810e+00 -6.43180490e-01 3.72101575e-01
2.92288601e-01 1.89489961e-01 6.53110623e-01 1.27785611e+00
-7.94503868e-01 3.55513543e-01 -1.40826046e-01 -7.35849857e-01
-1.42175114e+00 6.43439233e-01 6.65608168e-01 -7.57121921e-01
-2.84817606e-01 7.63420343e-01 5.58875680e-01 -3.47322851e-01
2.19239414e-01 -1.89603791e-01 -4.84620363e-01 -3.58805060e-01
5.48799574e-01 5.69687523e-02 1.98679581e-01 -1.95760727e-01
-3.18363786e-01 3.21930796e-01 -4.75540340e-01 2.71049947e-01
1.00579762e+00 7.08930716e-02 4.27837521e-02 1.83638096e-01
1.19605541e+00 -2.18635127e-01 -7.88443625e-01 9.59325582e-02
-1.50821328e-01 -1.50653630e-01 2.14357495e-01 -1.17978811e+00
-1.27566457e+00 1.02694309e+00 1.17814982e+00 -1.43091157e-01
1.45280755e+00 -3.13446254e-01 5.64203978e-01 3.99584249e-02
3.98674607e-02 -5.70846438e-01 3.37508023e-02 1.60019040e-01
8.38741958e-01 -1.40371847e+00 -2.02416599e-01 -4.54549670e-01
-5.33561945e-01 1.11161065e+00 1.08635306e+00 4.67072167e-02
6.82419777e-01 2.28710264e-01 2.78682470e-01 -1.90659791e-01
-5.03700256e-01 7.39195198e-02 8.91451463e-02 1.48283288e-01
7.35803246e-01 6.94003180e-02 -6.20677471e-01 1.02862740e+00
2.89853603e-01 2.38862231e-01 3.99128526e-01 9.43110824e-01
-1.89943299e-01 -6.86015487e-01 -2.08578587e-01 6.76785827e-01
-5.69180548e-01 -3.23005840e-02 -2.31017962e-01 9.44720149e-01
3.14193845e-01 7.93796241e-01 1.64987683e-01 -4.72185493e-01
1.63853422e-01 -9.84795839e-02 1.73106894e-01 -3.95468444e-01
-6.85430467e-01 1.26631916e-01 -2.52958506e-01 -3.03549677e-01
-2.17505395e-01 -2.35765830e-01 -1.32698607e+00 6.74992874e-02
-7.26166308e-01 3.12909305e-01 4.76981848e-01 5.91233313e-01
3.59500587e-01 9.62670207e-01 7.55216420e-01 -5.72099388e-01
-1.59290403e-01 -1.11196268e+00 -2.89659500e-01 4.22323823e-01
2.26709619e-01 -5.01631260e-01 -1.92711309e-01 1.93528365e-02] | [14.962089538574219, -2.806582450866699] |
dfe1a34e-8c0a-4939-bc14-daa1e8c0111d | an-interpretable-classifier-for-high | 2002.07613 | null | https://arxiv.org/abs/2002.07613v1 | https://arxiv.org/pdf/2002.07613v1.pdf | An interpretable classifier for high-resolution breast cancer screening images utilizing weakly supervised localization | Medical images differ from natural images in significantly higher resolutions and smaller regions of interest. Because of these differences, neural network architectures that work well for natural images might not be applicable to medical image analysis. In this work, we extend the globally-aware multiple instance classifier, a framework we proposed to address these unique properties of medical images. This model first uses a low-capacity, yet memory-efficient, network on the whole image to identify the most informative regions. It then applies another higher-capacity network to collect details from chosen regions. Finally, it employs a fusion module that aggregates global and local information to make a final prediction. While existing methods often require lesion segmentation during training, our model is trained with only image-level labels and can generate pixel-level saliency maps indicating possible malignant findings. We apply the model to screening mammography interpretation: predicting the presence or absence of benign and malignant lesions. On the NYU Breast Cancer Screening Dataset, consisting of more than one million images, our model achieves an AUC of 0.93 in classifying breasts with malignant findings, outperforming ResNet-34 and Faster R-CNN. Compared to ResNet-34, our model is 4.1x faster for inference while using 78.4% less GPU memory. Furthermore, we demonstrate, in a reader study, that our model surpasses radiologist-level AUC by a margin of 0.11. The proposed model is available online: https://github.com/nyukat/GMIC. | ['Nan Wu', 'Laura Heacock', 'Kyunghyun Cho', 'Krzysztof J. Geras', 'Kangning Liu', 'Jungkyu Park', 'Sudarshini Tyagi', 'Linda Moy', 'S. Gene Kim', 'Yiqiu Shen', 'Jason Phang'] | 2020-02-13 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 6.84409201e-01 4.98741537e-01 -5.78655005e-01 -4.36285734e-01
-1.19033182e+00 -6.12373352e-02 1.31974250e-01 5.62372029e-01
-4.49866563e-01 4.54160273e-01 1.80207044e-01 -6.71545625e-01
3.96931581e-02 -9.39781904e-01 -9.84628856e-01 -4.75923747e-01
-5.50194643e-02 2.15581074e-01 3.91737193e-01 1.19335949e-02
-2.55792439e-02 4.14598405e-01 -1.29082084e+00 7.51836419e-01
7.29923308e-01 1.04252875e+00 4.27056164e-01 9.43627119e-01
2.81103075e-01 1.06028938e+00 -2.84802675e-01 -1.47854224e-01
3.03206407e-02 -4.16735768e-01 -8.31896305e-01 -9.62086692e-02
5.36522090e-01 -5.60681820e-01 -2.80237526e-01 1.02045000e+00
6.46640241e-01 -1.76436245e-01 3.63674104e-01 -3.96798909e-01
-5.27689219e-01 6.30808115e-01 -7.76651740e-01 6.35590196e-01
-1.52845848e-02 3.64586487e-02 7.57298768e-01 -5.97953677e-01
6.52523518e-01 8.94270062e-01 8.01407099e-01 6.05457246e-01
-9.78622258e-01 -5.62880754e-01 -1.34338379e-01 5.64084239e-02
-1.24694169e+00 -3.82156074e-01 3.59097034e-01 4.60258275e-02
8.72606158e-01 5.82169771e-01 5.57988226e-01 9.23361242e-01
7.52680421e-01 9.02558088e-01 9.67679501e-01 -4.61362958e-01
6.75855670e-03 1.36917382e-02 2.23576292e-01 1.01655865e+00
3.45274508e-01 -1.72019482e-01 -1.33213043e-01 -2.19706506e-01
9.89881754e-01 2.61946350e-01 -1.07975788e-01 1.18387369e-02
-1.32128131e+00 7.82382429e-01 9.77428496e-01 4.27117944e-01
-5.55156827e-01 2.41644382e-01 3.18707079e-01 -2.27471933e-01
5.34049273e-01 5.10499060e-01 -1.84908971e-01 3.80685538e-01
-1.12310791e+00 -1.65717334e-01 4.46864992e-01 5.30819952e-01
5.13187289e-01 -3.61625731e-01 -3.72370392e-01 8.19500625e-01
-6.61268383e-02 3.66543204e-01 7.49786973e-01 -8.81540060e-01
-2.84995660e-02 6.49053514e-01 -2.12652877e-01 -1.19930875e+00
-8.67488623e-01 -7.87783206e-01 -1.13393319e+00 -4.69016768e-02
2.93246090e-01 -8.36497732e-03 -1.24516618e+00 1.43542242e+00
1.83215037e-01 2.90054828e-01 -3.89538445e-02 9.75924432e-01
1.12138629e+00 2.39376411e-01 3.17087591e-01 1.67273447e-01
1.71556151e+00 -1.09365439e+00 -4.65869695e-01 -4.17908370e-01
8.81696403e-01 -4.17122096e-01 7.75455892e-01 9.80709717e-02
-1.40305853e+00 -4.79792267e-01 -1.04113054e+00 -1.31914690e-01
-1.88042507e-01 4.55832839e-01 7.01583147e-01 5.07508039e-01
-1.39034557e+00 3.96624565e-01 -1.12665009e+00 -4.85828996e-01
7.21411765e-01 3.35785717e-01 -8.38633850e-02 -1.89678967e-01
-1.00786865e+00 1.06567657e+00 3.75410438e-01 1.07928611e-01
-8.64583492e-01 -8.59387279e-01 -9.38109338e-01 1.48275457e-02
3.71174455e-01 -7.64763713e-01 1.47386563e+00 -1.15844226e+00
-9.54057455e-01 8.91535580e-01 -4.16401565e-01 -8.32858622e-01
4.55319881e-01 1.30746856e-01 -3.60515624e-01 6.20481253e-01
2.84110069e-01 1.05066693e+00 4.22093332e-01 -8.96465540e-01
-7.74198592e-01 -1.79306194e-01 8.61300156e-03 2.46312976e-01
-9.56113636e-02 -4.89584208e-02 -7.81874299e-01 -5.63780665e-01
2.13045746e-01 -8.93713534e-01 -8.52313757e-01 1.17229328e-01
-4.85480875e-01 3.49788696e-01 4.00083482e-01 -8.50526273e-01
1.00251043e+00 -2.01937509e+00 -3.65988880e-01 3.35579962e-01
4.65669304e-01 1.18131541e-01 -3.01696174e-02 -5.11536241e-01
-7.10555762e-02 1.77063793e-01 -1.90416694e-01 2.44841892e-02
-6.89298868e-01 -4.79061007e-02 3.35098088e-01 4.54096198e-01
4.52526540e-01 1.32734179e+00 -9.45612907e-01 -7.40795434e-01
3.72096330e-01 4.57065433e-01 -4.88960057e-01 -3.77982944e-01
-1.23005770e-02 3.53447795e-01 -4.66441661e-01 8.79644930e-01
5.94992280e-01 -1.01525259e+00 3.81644040e-01 -3.28396738e-01
3.00231189e-01 4.90365326e-02 -7.05413222e-01 1.74055576e+00
-4.80691105e-01 6.38661921e-01 -2.63989368e-03 -8.87934685e-01
6.14983916e-01 2.18736425e-01 6.35958374e-01 -9.61357713e-01
1.98407650e-01 2.72215545e-01 2.31495306e-01 -3.29394281e-01
4.68965113e-01 2.60349691e-01 8.78543854e-02 2.45353058e-01
-2.68258840e-01 2.50570357e-01 1.03067450e-01 1.37618780e-01
1.52556980e+00 -2.19256461e-01 6.22169375e-01 -3.51804674e-01
2.89104372e-01 2.79719323e-01 3.96089137e-01 1.04007578e+00
-2.71216303e-01 8.49493623e-01 3.40020776e-01 -6.10083163e-01
-9.63000894e-01 -9.38699901e-01 -4.27462786e-01 8.71246636e-01
3.47363025e-01 -4.92893755e-02 -5.58166385e-01 -5.63905060e-01
-1.23808920e-01 6.26194239e-01 -9.66873765e-01 -5.33106849e-02
-7.41361678e-01 -1.11191845e+00 5.55914462e-01 6.51874840e-01
5.47085345e-01 -1.09498525e+00 -1.02549303e+00 2.90083438e-01
-3.79725873e-01 -1.02369547e+00 -1.31373614e-01 1.04064606e-01
-9.65901136e-01 -1.14380455e+00 -1.06273866e+00 -8.47938836e-01
1.09073985e+00 3.38324696e-01 1.30983174e+00 2.99654394e-01
-8.10925841e-01 1.08340450e-01 -1.71591595e-01 -5.03736496e-01
-6.29308403e-01 2.44407907e-01 -6.74171031e-01 -2.90368974e-01
3.50973189e-01 2.19272524e-02 -1.04707015e+00 2.21017510e-01
-9.07435954e-01 6.17261887e-01 1.14677703e+00 9.08039212e-01
8.97119284e-01 -2.43684590e-01 4.14277375e-01 -9.65826452e-01
2.24627554e-01 -6.87936008e-01 -8.90737623e-02 3.50386530e-01
-3.01880628e-01 -2.26651967e-01 3.16545248e-01 -3.08318436e-01
-9.94793952e-01 2.43588880e-01 1.53469201e-02 -1.25286698e-01
-2.28147283e-01 4.74630177e-01 7.35292912e-01 -6.98842257e-02
9.85193610e-01 1.97454810e-01 2.15208307e-01 -1.15227312e-01
-2.52471548e-02 4.55496460e-01 6.02853119e-01 -1.54136680e-03
1.33189037e-01 6.88008964e-01 8.84193480e-02 -6.23908401e-01
-1.05269873e+00 -4.74033147e-01 -3.30729187e-01 -1.95973754e-01
9.91424859e-01 -1.01880264e+00 -3.99622589e-01 1.14295788e-01
-8.43646348e-01 -3.66923451e-01 -2.60578543e-01 5.30811965e-01
-2.69885272e-01 6.17300086e-02 -9.42012668e-01 -3.21160316e-01
-5.61233461e-01 -1.32117939e+00 1.26063168e+00 3.97067994e-01
-2.34657928e-01 -8.15797091e-01 -5.12994826e-01 3.71034056e-01
5.33568025e-01 3.82339150e-01 6.45970821e-01 -5.45367360e-01
-6.02308393e-01 -2.55706131e-01 -6.34932399e-01 -6.22070394e-03
2.83618689e-01 -1.51347741e-01 -8.66622508e-01 -2.05460161e-01
-2.44742796e-01 -2.20901713e-01 1.23706973e+00 1.14723361e+00
1.80514586e+00 5.01345545e-02 -7.21840501e-01 6.78906024e-01
1.33357346e+00 1.67303637e-01 5.62794089e-01 4.70500708e-01
4.90811527e-01 3.83193791e-01 5.88459969e-01 2.09746376e-01
3.96587193e-01 2.80701220e-01 4.82019931e-01 -8.51100683e-01
-4.19998974e-01 1.51298111e-02 -1.47513568e-01 3.82660717e-01
4.09665657e-03 -1.42927974e-01 -1.14027572e+00 5.98452926e-01
-1.70786119e+00 -6.44097745e-01 -4.86787111e-02 1.80916095e+00
8.93583894e-01 1.03784882e-01 -1.87282518e-01 -2.92279989e-01
8.13461065e-01 1.38607612e-02 -8.10774744e-01 -2.18625501e-01
-7.68132359e-02 3.29636812e-01 1.03635383e+00 2.90152609e-01
-1.33442891e+00 6.13175929e-01 6.57551765e+00 9.22412038e-01
-1.20815301e+00 2.92349905e-01 1.49454820e+00 -1.43131629e-01
-1.51181132e-01 -4.67148513e-01 -6.10041559e-01 2.38626212e-01
1.10263681e+00 1.34829953e-01 -6.79000989e-02 8.04473460e-01
1.31666824e-01 -4.61844712e-01 -8.19937229e-01 6.01995766e-01
1.84213638e-01 -1.83670676e+00 -6.04231805e-02 -6.22553527e-02
6.86159968e-01 2.80577630e-01 1.49840623e-01 -6.58367993e-03
2.51090020e-01 -1.31369519e+00 3.47414821e-01 4.97939110e-01
9.13184822e-01 -5.74017644e-01 9.37340260e-01 2.35855505e-01
-9.15474057e-01 1.82774253e-02 -3.69587928e-01 3.58654946e-01
-2.48874992e-01 6.88529849e-01 -1.32444835e+00 4.88194317e-01
8.95161033e-01 6.50251389e-01 -1.05067050e+00 1.01879013e+00
2.06320271e-01 5.63705862e-01 -3.72529507e-01 -5.05085848e-02
3.23061466e-01 5.06571949e-01 1.08961381e-01 1.45728874e+00
3.88409019e-01 2.31946073e-02 7.09645310e-03 6.31466746e-01
6.05969504e-02 1.66760132e-01 -3.53515327e-01 5.07916629e-01
1.81557864e-01 1.46417296e+00 -1.18032873e+00 -6.35483563e-01
-4.11292732e-01 8.56472433e-01 1.49044797e-01 1.18070699e-01
-1.01675558e+00 -1.41402230e-01 -1.08343087e-01 3.70304100e-02
1.14325203e-01 3.52620363e-01 -5.49861968e-01 -8.37081134e-01
-1.52322099e-01 -8.41887355e-01 4.31713939e-01 -7.57390678e-01
-8.45668256e-01 6.27478600e-01 -1.41359359e-01 -1.20735765e+00
-2.92147934e-01 -5.76708078e-01 -3.38342726e-01 7.21001387e-01
-1.71097457e+00 -1.17149723e+00 -6.23829484e-01 1.70972437e-01
5.86960673e-01 9.62983742e-02 8.85964811e-01 1.47385895e-01
-5.64486980e-01 5.48582315e-01 2.66894735e-02 2.75673121e-01
5.69605470e-01 -1.16957533e+00 4.52256233e-01 6.36035025e-01
-3.07185531e-01 3.80504906e-01 3.11959565e-01 -7.70258904e-01
-9.91291761e-01 -1.40605450e+00 6.47032499e-01 -2.38527268e-01
4.37627435e-01 1.21806011e-01 -7.50680387e-01 7.17027068e-01
6.43279180e-02 3.60624045e-01 6.52675807e-01 -2.86860973e-01
8.94483477e-02 5.65433390e-02 -1.33715963e+00 6.86428905e-01
8.63633215e-01 -2.89288871e-02 -1.04236998e-01 4.80512738e-01
5.75163424e-01 -8.86498988e-01 -9.49249566e-01 6.99618578e-01
4.92803156e-01 -8.69453132e-01 1.15439975e+00 -1.86937258e-01
7.91744769e-01 6.68161660e-02 3.85133713e-03 -9.23605561e-01
-4.72099245e-01 2.67890282e-03 1.45849556e-01 3.18978757e-01
7.84121931e-01 -6.60292089e-01 8.68357658e-01 4.61609185e-01
-2.02167764e-01 -1.16235971e+00 -7.28229284e-01 -1.71593502e-01
-5.85447364e-02 -3.71401250e-01 4.35496390e-01 7.76544034e-01
-3.56446207e-01 -2.83574313e-01 -2.77124136e-03 2.13629588e-01
5.81739426e-01 1.19571105e-01 1.11748509e-01 -7.62953937e-01
-3.62237811e-01 -6.02750719e-01 -3.94621223e-01 -7.34061897e-01
-2.09995225e-01 -1.08086360e+00 1.89607516e-02 -1.67778373e+00
6.46666706e-01 -5.55728078e-01 -6.93537295e-01 7.17682064e-01
-3.63547146e-01 7.14719772e-01 -3.37392278e-02 1.80474862e-01
-6.26351416e-01 -2.64865816e-01 1.58508921e+00 -3.14797640e-01
9.20582097e-03 -4.11998034e-02 -9.49782550e-01 8.01819861e-01
1.07757151e+00 -2.70191789e-01 -1.24434158e-01 -3.73853981e-01
-1.29463986e-01 2.99257755e-01 6.40776694e-01 -1.07961774e+00
1.76978767e-01 -6.82779029e-02 1.01092494e+00 -6.77716494e-01
2.59283155e-01 -4.59514767e-01 3.47711146e-02 8.91171634e-01
-7.03359008e-01 -2.63497949e-01 4.07754660e-01 4.02782530e-01
-1.03257351e-01 -1.96653917e-01 8.91375959e-01 -5.25567293e-01
-7.42105186e-01 1.64896905e-01 -3.56141269e-01 -4.10441399e-01
1.13880634e+00 -3.34883630e-01 -4.52006012e-01 -1.40642509e-01
-7.66073167e-01 1.39543191e-01 3.05625409e-01 2.07084686e-01
6.58847868e-01 -9.93655324e-01 -8.42846036e-01 1.37815669e-01
-1.77916884e-03 2.56335318e-01 5.63798785e-01 1.10848105e+00
-8.42518508e-01 4.66395289e-01 -9.04711708e-02 -9.35394466e-01
-1.26467979e+00 1.44400045e-01 5.86172521e-01 -5.33401370e-01
-7.68354356e-01 7.83310115e-01 4.10566747e-01 -2.50735283e-01
-3.59289870e-02 -7.03467011e-01 -2.76506603e-01 -4.07832772e-01
6.54371977e-01 1.33832961e-01 2.44034559e-01 -4.69559103e-01
-3.69597048e-01 3.19394559e-01 -3.60662609e-01 1.95524186e-01
1.13635945e+00 1.39366418e-01 -2.41082744e-03 8.20793286e-02
1.10926199e+00 -2.92751819e-01 -8.82174015e-01 -2.90560246e-01
-6.59750104e-02 -2.45963126e-01 4.30498391e-01 -1.00444651e+00
-1.12990403e+00 4.44888920e-01 9.01071906e-01 6.13401197e-02
1.45087576e+00 1.71200722e-01 7.35067904e-01 1.78923473e-01
4.92280014e-02 -7.42988467e-01 1.08472342e-02 5.05614057e-02
5.56843221e-01 -1.57362854e+00 1.35676116e-01 -4.07343864e-01
-6.52038574e-01 1.08415926e+00 7.27434635e-01 -2.30850711e-01
5.49436748e-01 3.81200284e-01 6.03041537e-02 -2.42352962e-01
-6.78088605e-01 -2.23003048e-02 4.92124081e-01 3.36000949e-01
6.18042171e-01 3.25756907e-01 -1.45102054e-01 2.76342511e-01
-8.51108730e-02 1.80009052e-01 4.32727814e-01 1.02068698e+00
-5.32603562e-01 -5.55302978e-01 -3.17185134e-01 1.17984879e+00
-8.60735893e-01 -1.44251823e-01 7.56809711e-02 7.51427412e-01
1.19441776e-02 7.68825650e-01 4.42182660e-01 -1.02234073e-01
5.62970601e-02 -3.73802274e-01 3.04789960e-01 -6.73164666e-01
-6.42855644e-01 1.85129240e-01 1.47409335e-01 -7.37379968e-01
-5.20438731e-01 -5.69621742e-01 -1.38133168e+00 -6.93198740e-02
-2.12097988e-01 -3.43815088e-01 6.76610470e-01 6.02721393e-01
4.28430378e-01 1.09552932e+00 3.71112376e-01 -8.09587777e-01
-2.66433328e-01 -8.53231370e-01 -3.22108477e-01 1.12520307e-01
4.35385555e-01 -2.06987128e-01 1.62394196e-02 -9.91064012e-02] | [15.072592735290527, -2.5372424125671387] |
78637963-195e-46d9-a832-399452ad271e | individualized-and-global-feature | 2211.04409 | null | https://arxiv.org/abs/2211.04409v1 | https://arxiv.org/pdf/2211.04409v1.pdf | Individualized and Global Feature Attributions for Gradient Boosted Trees in the Presence of $\ell_2$ Regularization | While $\ell_2$ regularization is widely used in training gradient boosted trees, popular individualized feature attribution methods for trees such as Saabas and TreeSHAP overlook the training procedure. We propose Prediction Decomposition Attribution (PreDecomp), a novel individualized feature attribution for gradient boosted trees when they are trained with $\ell_2$ regularization. Theoretical analysis shows that the inner product between PreDecomp and labels on in-sample data is essentially the total gain of a tree, and that it can faithfully recover additive models in the population case when features are independent. Inspired by the connection between PreDecomp and total gain, we also propose TreeInner, a family of debiased global feature attributions defined in terms of the inner product between any individualized feature attribution and labels on out-sample data for each tree. Numerical experiments on a simulated dataset and a genomic ChIP dataset show that TreeInner has state-of-the-art feature selection performance. Code reproducing experiments is available at https://github.com/nalzok/TreeInner . | ['Qingyao Sun'] | 2022-11-08 | null | null | null | null | ['additive-models'] | ['methodology'] | [ 3.35483879e-01 3.57209533e-01 -5.80611944e-01 -7.77620912e-01
-7.08416402e-01 -1.99838310e-01 2.42549628e-01 1.26843795e-01
1.45353064e-01 7.60190487e-01 -7.44689554e-02 -1.92593887e-01
-3.33884865e-01 -7.88736582e-01 -7.76388526e-01 -9.45815802e-01
-1.62237510e-01 4.46044713e-01 -1.86917171e-01 -1.01644158e-01
-6.41719028e-02 8.71061906e-02 -1.62963998e+00 -1.74374461e-01
8.49081635e-01 1.06369007e+00 -3.30776036e-01 5.87158203e-01
1.37063324e-01 5.81816554e-01 -2.66480505e-01 -7.09056497e-01
1.78883538e-01 -3.46922219e-01 -5.60603321e-01 -1.72729772e-02
5.79425037e-01 1.28993616e-01 -7.49300271e-02 9.74780500e-01
5.80849588e-01 -3.29065830e-01 8.24270368e-01 -1.54149330e+00
-6.66492462e-01 1.10219789e+00 -7.48136222e-01 -4.33698408e-02
-2.93063939e-01 -1.42659515e-01 1.35962689e+00 -8.92945051e-01
3.92326951e-01 9.93678391e-01 1.10853672e+00 3.83811980e-01
-1.77524602e+00 -8.25156152e-01 1.81365788e-01 1.76677667e-02
-1.37816381e+00 -2.89809793e-01 6.91793859e-01 -6.62915289e-01
2.60136783e-01 6.20891690e-01 3.78695905e-01 6.72492921e-01
4.82660711e-01 7.68976510e-01 8.99076462e-01 -4.75501835e-01
3.62675190e-01 2.61976659e-01 8.72926533e-01 1.03296030e+00
2.54225403e-01 2.59947181e-01 -8.53130043e-01 -5.25532842e-01
3.34481508e-01 2.63544917e-02 -1.28655493e-01 -5.69246531e-01
-7.59673715e-01 1.03466403e+00 3.97028685e-01 2.50170920e-02
-1.20629504e-01 4.35147077e-01 3.12687218e-01 3.01089346e-01
8.18170428e-01 -1.14364460e-01 -6.42484784e-01 2.69526422e-01
-7.56050885e-01 7.50190811e-03 5.24964511e-01 8.76554370e-01
1.15029716e+00 2.13223156e-02 -3.48925143e-01 9.17213917e-01
1.31181240e-01 3.50538939e-01 4.07934815e-01 -8.27492654e-01
-1.95335910e-01 8.01139176e-01 -2.43287176e-01 -4.82811600e-01
-6.41254246e-01 -1.08713174e+00 -1.13913202e+00 1.43333331e-01
3.81498724e-01 7.28440331e-03 -8.57345998e-01 2.14299679e+00
5.46557009e-01 -4.01784293e-02 -4.05440331e-01 5.03346264e-01
4.91450161e-01 1.80808660e-02 5.68160564e-02 -2.60711521e-01
1.36233950e+00 -6.99744165e-01 -3.91986519e-01 -2.99811333e-01
9.67918217e-01 -3.21274847e-01 1.12127495e+00 2.53007501e-01
-5.28426051e-01 -3.35285991e-01 -9.74211037e-01 7.93634728e-02
-1.47249386e-01 5.61464906e-01 1.10842955e+00 1.18658006e+00
-9.20557857e-01 8.67818534e-01 -8.62562299e-01 -2.58638561e-01
7.90189505e-01 5.24513900e-01 -2.93946534e-01 3.74187306e-02
-8.36101592e-01 4.40520048e-01 -1.33713469e-01 -2.51829457e-02
-8.57489109e-01 -1.08322084e+00 -4.16102588e-01 7.87730049e-03
1.75031796e-01 -8.41867089e-01 1.00606120e+00 -8.49297345e-01
-1.29781580e+00 8.92607152e-01 -2.34926462e-01 -4.08159852e-01
4.38918799e-01 -1.63672075e-01 1.19026579e-01 -7.27512300e-01
3.37656051e-01 4.33178961e-01 1.12821507e+00 -7.29064643e-01
-5.55340588e-01 -8.72059286e-01 -4.73155111e-01 -1.51561156e-01
-5.14037430e-01 -3.98343086e-01 6.81459755e-02 -5.14273047e-01
3.79704684e-01 -1.03299558e+00 -3.72525126e-01 1.65460184e-01
-7.52036572e-01 -1.80903092e-01 5.88217258e-01 -5.21528602e-01
1.27105534e+00 -2.03888583e+00 2.01312631e-01 2.61509418e-01
4.67290342e-01 -3.40591162e-01 -2.26930459e-03 1.40422165e-01
-2.14577481e-01 2.80369818e-01 -4.74889964e-01 -3.58463645e-01
9.11346003e-02 1.41298369e-01 -2.28537902e-01 7.53650248e-01
-1.14378996e-01 5.75493634e-01 -6.52059197e-01 -2.96257257e-01
-1.83800325e-01 3.36447656e-01 -5.49902320e-01 -1.03211097e-01
-8.06336775e-02 2.64423966e-01 -5.14111578e-01 8.50275099e-01
7.96639800e-01 -3.96121860e-01 2.36829534e-01 1.18505016e-01
1.17673323e-01 3.36092454e-03 -7.46570885e-01 1.21964061e+00
-1.74215585e-01 2.97335625e-01 -2.44668759e-02 -1.06499350e+00
1.19585013e+00 -8.51360634e-02 3.31792206e-01 -6.52686413e-03
1.85942456e-01 2.38742843e-01 -9.97469276e-02 3.93222034e-01
2.89102256e-01 -4.41913515e-01 -3.33718777e-01 3.62511486e-01
2.31700882e-01 -3.03527489e-02 -3.37276198e-02 2.43427783e-01
1.31048560e+00 3.46595980e-02 5.44871449e-01 -5.88298500e-01
4.07674640e-01 -3.48188043e-01 8.58571768e-01 9.26305830e-01
-2.29216948e-01 3.73940796e-01 7.50183463e-01 -3.47219706e-01
-9.00918126e-01 -1.05702174e+00 -5.76295972e-01 1.70733416e+00
-3.83009553e-01 -4.57490891e-01 -6.27861261e-01 -7.05512226e-01
3.88174027e-01 8.86964500e-01 -1.03598034e+00 -4.31596994e-01
-2.53893901e-03 -1.30960965e+00 5.00615537e-01 4.79774952e-01
2.50181824e-01 -3.78780007e-01 -3.85404617e-01 2.96602342e-02
9.90124792e-02 -3.24562401e-01 -4.47424382e-01 9.20424521e-01
-1.03633475e+00 -8.17320406e-01 -3.63537163e-01 -3.90280217e-01
6.41089320e-01 -7.41833225e-02 1.04087722e+00 9.72122103e-02
-6.36999965e-01 2.04856187e-01 -3.57195824e-01 -4.44001883e-01
-3.21684837e-01 3.64762455e-01 1.29244536e-01 1.66632786e-01
2.70410001e-01 -8.94014776e-01 -2.32346401e-01 4.79994059e-01
-3.18219393e-01 1.10613266e-02 4.23707783e-01 1.41813982e+00
7.78555691e-01 -2.67827660e-01 6.64133012e-01 -1.06236136e+00
1.75267011e-01 -4.82908070e-01 -6.22059703e-01 3.32489431e-01
-9.49433625e-01 3.10000628e-01 3.98521036e-01 -2.21508294e-01
-8.46433878e-01 3.46983969e-01 5.78215756e-02 -2.50126272e-01
3.16574126e-01 3.94222528e-01 -1.77276835e-01 -1.35890201e-01
9.12803769e-01 6.46381751e-02 -9.55396891e-03 -6.04894936e-01
5.56269228e-01 4.99342799e-01 1.56194136e-01 -5.26381493e-01
5.53820431e-01 3.96622747e-01 4.06763196e-01 -5.74668288e-01
-9.48012769e-01 9.11281724e-03 -6.33965075e-01 -1.37302324e-01
3.14663082e-01 -6.76221728e-01 -9.42293465e-01 4.91625935e-01
-3.82712692e-01 -2.76558727e-01 -6.59687698e-01 2.20301658e-01
-7.65167654e-01 9.61632431e-02 -1.61920905e-01 -9.80604053e-01
-5.62791467e-01 -7.93872416e-01 8.41937125e-01 1.82918478e-02
-1.84613124e-01 -8.33525777e-01 1.78596620e-02 2.64160484e-01
3.92183870e-01 2.26957172e-01 1.27584827e+00 -7.31818199e-01
-2.91971385e-01 -6.16672397e-01 1.47753716e-01 3.10699821e-01
9.86232460e-02 6.47361949e-02 -1.06871617e+00 -2.08556712e-01
-6.23609032e-03 -4.04240221e-01 1.23279226e+00 7.31223583e-01
1.04978955e+00 -1.55846134e-01 -6.30849361e-01 7.30525970e-01
1.10837281e+00 -2.44203895e-01 3.27170670e-01 -5.52555323e-02
6.46649480e-01 4.18699771e-01 5.19781232e-01 8.63112211e-01
1.87668473e-01 7.05928326e-01 4.56959635e-01 1.93240687e-01
-6.32760003e-02 -3.94514084e-01 4.00526226e-01 5.48140347e-01
7.01874867e-02 1.72743145e-02 -7.11147964e-01 3.55193049e-01
-1.90530634e+00 -5.80312192e-01 -3.86968225e-01 2.64587379e+00
7.48388350e-01 3.22932154e-02 -1.61830150e-03 2.09496289e-01
9.21783388e-01 -1.37822926e-01 -9.88683283e-01 -3.37031037e-01
-2.71184921e-01 2.12282985e-01 8.08074653e-01 5.56276858e-01
-1.22451639e+00 6.79393351e-01 5.99927950e+00 1.23646164e+00
-8.85866642e-01 4.31196034e-01 1.19302797e+00 -2.68332679e-02
-2.87991881e-01 2.43996993e-01 -1.03784621e+00 3.68253946e-01
7.85787344e-01 -5.18271029e-01 1.74196288e-01 1.21257079e+00
-3.56688678e-01 -9.86481532e-02 -1.35836053e+00 5.17164767e-01
-3.26782763e-01 -1.01105237e+00 -3.78957182e-01 1.94296077e-01
5.31192183e-01 -3.12129594e-02 4.02541280e-01 3.98217708e-01
6.99564457e-01 -8.88856530e-01 6.37334704e-01 3.65842581e-01
7.92102277e-01 -5.22697210e-01 6.47732556e-01 3.54953498e-01
-1.04826617e+00 -1.26557454e-01 -4.93223190e-01 -2.13368665e-02
-2.23817974e-01 1.32116389e+00 -1.00712299e+00 6.11025512e-01
5.96151829e-01 7.10026264e-01 -7.94108927e-01 6.61150038e-01
-2.43591487e-01 1.06528187e+00 -6.63072705e-01 9.75701660e-02
-4.95368719e-01 -6.32089972e-02 5.80864906e-01 7.52369225e-01
4.05034244e-01 -3.19992453e-01 -1.71914902e-02 7.51875579e-01
-1.61179423e-01 1.92148432e-01 -2.69455314e-01 1.19500838e-01
4.48499769e-01 1.35001910e+00 -5.85582316e-01 -1.59374222e-01
1.73799768e-01 7.30051100e-01 5.38479269e-01 -1.02152325e-01
-7.33994782e-01 -6.35504052e-02 7.46567845e-01 1.68554112e-02
4.20598954e-01 5.70784569e-01 -6.95331037e-01 -9.43328738e-01
-1.74982190e-01 -4.87095237e-01 6.33312166e-01 -4.08565849e-01
-1.59444094e+00 3.00040871e-01 -2.67908424e-01 -1.17435944e+00
9.10601765e-02 -4.63413179e-01 -4.70845193e-01 6.85341120e-01
-8.52642059e-01 -1.33447123e+00 -2.41978541e-01 3.01425666e-01
1.11837439e-01 -3.18958238e-02 1.02158439e+00 -1.58319194e-02
-9.06483352e-01 1.23014629e+00 5.26742339e-01 -4.55529958e-01
4.74960059e-01 -1.09486401e+00 2.12655067e-01 4.08559412e-01
-9.14003626e-02 4.40706223e-01 7.89721191e-01 -6.45462155e-01
-1.33679116e+00 -1.14234185e+00 6.59274578e-01 -3.01259160e-01
5.31967282e-01 -4.05980289e-01 -7.36150861e-01 1.02326405e+00
-3.18831593e-01 2.19905317e-01 9.41394389e-01 7.18400776e-01
-6.16487324e-01 -3.96322995e-01 -1.37399781e+00 1.36944100e-01
1.26244807e+00 -1.21519171e-01 1.26847312e-01 5.88740289e-01
5.67635477e-01 -9.83484015e-02 -1.13612998e+00 5.39372265e-01
6.80990815e-01 -1.05173039e+00 7.86162913e-01 -4.81615454e-01
2.42262572e-01 -9.88706946e-02 -5.32673120e-01 -1.23655474e+00
-7.00228572e-01 -5.06127536e-01 1.39525771e-01 1.52056301e+00
5.67700028e-01 -9.87239480e-01 1.11992562e+00 4.50595170e-01
-1.86208189e-01 -9.41686094e-01 -1.35654604e+00 -8.17740262e-01
2.73722947e-01 -3.42290670e-01 4.68373001e-01 8.24984968e-01
1.34117290e-01 5.81164896e-01 -6.33018017e-01 -2.09982589e-01
1.10245645e+00 1.50108859e-02 7.40980983e-01 -1.67283475e+00
-4.68118072e-01 -3.02819937e-01 -6.28796279e-01 -4.90143448e-01
9.53711867e-02 -1.33472013e+00 -6.97953999e-02 -8.46713006e-01
5.90965927e-01 -9.09132123e-01 -2.87111610e-01 9.82042968e-01
-3.15656126e-01 1.23599447e-01 -4.69120666e-02 1.86184734e-01
-6.90136328e-02 7.83270359e-01 6.92335308e-01 -6.70438260e-02
-1.56077549e-01 2.74890751e-01 -1.03932321e+00 8.08100462e-01
8.52500737e-01 -9.44495142e-01 -1.05727471e-01 2.31032506e-01
6.84066936e-02 7.42747914e-04 2.58580118e-01 -8.28898251e-01
-4.57121879e-02 1.02514818e-01 3.57017785e-01 -6.52218223e-01
4.74270135e-01 -4.48714405e-01 5.82784951e-01 7.36126900e-01
-5.09678006e-01 -4.91059303e-01 -3.39260809e-02 6.42229080e-01
2.98144996e-01 -2.65734553e-01 7.63030052e-01 2.22083017e-01
-1.47883251e-01 1.03712693e-01 -2.86653459e-01 -1.43987283e-01
9.04596865e-01 6.70790672e-02 -4.17494416e-01 -3.94111753e-01
-8.23569715e-01 2.23594353e-01 4.17973489e-01 -1.86404631e-01
2.58220345e-01 -1.20214009e+00 -9.83650923e-01 2.95181900e-01
1.74164578e-01 -4.73994493e-01 4.11579281e-01 1.05294335e+00
1.80229902e-01 7.39762783e-02 -2.81176325e-02 -7.10875809e-01
-1.48899961e+00 4.23424661e-01 3.24902505e-01 -4.12157744e-01
-2.54612505e-01 1.46575618e+00 5.44680595e-01 -7.49887407e-01
2.76417695e-02 5.40788546e-02 3.26887369e-01 7.86203071e-02
2.73054004e-01 5.76661229e-01 4.10581864e-02 -4.28611308e-01
-6.13940537e-01 5.73851347e-01 -2.78762460e-01 1.47211611e-01
1.17941093e+00 1.20375969e-01 -3.75889391e-01 5.89337349e-01
1.05100465e+00 -9.10436735e-02 -1.00533998e+00 -3.03277701e-01
-1.63133144e-01 -4.55101281e-01 9.03241783e-02 -8.07393253e-01
-1.05722797e+00 7.80096054e-01 9.20916259e-01 1.27916746e-02
1.21250355e+00 1.89550698e-01 1.02500422e-02 1.95454776e-01
4.50294882e-01 -7.09314346e-01 -2.42525697e-01 1.97290897e-01
8.21684003e-01 -9.48302925e-01 3.83653671e-01 -7.59762824e-01
-6.12308323e-01 7.59150624e-01 4.18624938e-01 -3.16424146e-02
8.89240563e-01 2.51567304e-01 -3.43243599e-01 1.19154036e-01
-1.01720726e+00 -3.35140526e-02 -1.93769336e-01 5.08393884e-01
6.82568431e-01 5.66182613e-01 -5.10547519e-01 1.19413972e+00
-4.82696593e-01 5.90649843e-02 4.29606438e-01 7.90198803e-01
-5.40674210e-01 -1.22370589e+00 -3.47297043e-01 1.02036381e+00
-3.69368732e-01 -4.15284187e-02 -4.36228544e-01 4.30225283e-01
2.58527249e-01 6.94763660e-01 -7.23327771e-02 -9.13529336e-01
1.02960683e-01 4.61828738e-01 3.11344415e-01 -3.08122337e-01
-7.40708947e-01 -9.17162597e-02 1.95899576e-01 -1.89975232e-01
-1.97764058e-02 -7.93758631e-01 -9.84801829e-01 -5.01769304e-01
-7.37492383e-01 1.98724940e-01 6.84327841e-01 4.71592814e-01
5.09028792e-01 5.00557005e-01 7.99881041e-01 -7.26180375e-01
-8.95127654e-01 -1.11259854e+00 -1.03607333e+00 4.08852026e-02
7.17329159e-02 -8.36972654e-01 -6.29993379e-01 -1.80939324e-02] | [8.136127471923828, 4.692221641540527] |
c4ee9e64-9414-4291-8b1f-bd5e70f33790 | geowine-geolocation-based-wiki-image-news-and | 2104.14994 | null | https://arxiv.org/abs/2104.14994v2 | https://arxiv.org/pdf/2104.14994v2.pdf | GeoWINE: Geolocation based Wiki, Image,News and Event Retrieval | In the context of social media, geolocation inference on news or events has become a very important task. In this paper, we present the GeoWINE (Geolocation-based Wiki-Image-News-Event retrieval) demonstrator, an effective modular system for multimodal retrieval which expects only a single image as input. The GeoWINE system consists of five modules in order to retrieve related information from various sources. The first module is a state-of-the-art model for geolocation estimation of images. The second module performs a geospatial-based query for entity retrieval using the Wikidata knowledge graph. The third module exploits four different image embedding representations, which are used to retrieve most similar entities compared to the input image. The embeddings are derived from the tasks of geolocation estimation, place recognition, ImageNet-based image classification, and their combination. The last two modules perform news and event retrieval from EventRegistry and the Open Event Knowledge Graph (OEKG). GeoWINE provides an intuitive interface for end-users and is insightful for experts for reconfiguration to individual setups. The GeoWINE achieves promising results in entity label prediction for images on Google Landmarks dataset. The demonstrator is publicly available at http://cleopatra.ijs.si/geowine/. | ['Ralph Ewerth', 'Jens Lehmann', 'Sherzod Hakimov', 'Eric Müller-Budack', 'Endri Kacupaj', 'Golsa Tahmasebzadeh'] | 2021-04-30 | null | null | null | null | ['photo-geolocation-estimation'] | ['computer-vision'] | [-5.56122303e-01 1.03622086e-01 -1.88049972e-01 -2.29892448e-01
-9.83268142e-01 -5.31590819e-01 9.98513460e-01 6.35009646e-01
-8.02261591e-01 4.41184729e-01 4.19224828e-01 1.07591301e-01
-4.11013216e-01 -1.02833617e+00 -6.04453146e-01 -4.93741453e-01
-3.47442508e-01 5.60448945e-01 4.45790708e-01 -2.64340669e-01
3.04627717e-01 4.17856634e-01 -1.77076209e+00 2.16447264e-01
2.28197023e-01 1.08979082e+00 3.52688551e-01 8.02921116e-01
-2.81590432e-01 5.81717789e-01 -3.36436749e-01 -3.49973947e-01
-8.01899377e-03 1.61157027e-01 -7.78235316e-01 -6.07107520e-01
5.06530762e-01 6.86896890e-02 -6.01590633e-01 8.33480656e-01
8.69771540e-01 1.95668817e-01 8.10150146e-01 -1.64074004e+00
-8.45725596e-01 5.74783206e-01 -3.05140227e-01 4.35224563e-01
5.32527268e-01 -4.22309816e-01 1.06252956e+00 -1.36075509e+00
1.06786466e+00 8.99255812e-01 4.83408391e-01 -1.46807939e-01
-5.05580604e-01 -4.14694816e-01 -1.29142702e-01 6.95832133e-01
-2.12815213e+00 -3.92136067e-01 3.44691843e-01 -4.98742670e-01
9.71353233e-01 3.43672663e-01 5.16580284e-01 6.45841002e-01
5.68665266e-02 7.17274249e-01 4.81062949e-01 -5.46833277e-01
7.18551734e-03 5.42195082e-01 4.44283858e-02 8.41365516e-01
1.62262265e-02 -2.01504022e-01 -9.49606657e-01 -1.17997192e-01
4.89960849e-01 5.24337701e-02 -4.10104096e-01 -5.17631829e-01
-1.57424819e+00 6.89890265e-01 1.02748346e+00 4.29618865e-01
-4.98574257e-01 2.83720642e-01 4.06070441e-01 1.76404238e-01
5.82619548e-01 3.24357957e-01 -3.94093357e-02 5.19530587e-02
-7.90721655e-01 7.94210881e-02 8.82168770e-01 1.09157038e+00
9.96858954e-01 -4.24802303e-01 -1.79376602e-02 8.57637167e-01
5.59693038e-01 5.57021976e-01 6.59214616e-01 -2.93367654e-01
5.40221393e-01 5.89867175e-01 2.32375190e-01 -1.63539898e+00
-4.54534620e-01 -1.08678862e-01 -3.60051453e-01 -2.86464602e-01
6.84988797e-02 1.19676702e-01 -7.87167490e-01 1.29653811e+00
4.80291873e-01 4.06726360e-01 2.51031145e-02 9.18785930e-01
1.53226101e+00 8.51008415e-01 4.11054462e-01 6.11371756e-01
1.61839569e+00 -1.00132632e+00 -4.88710970e-01 -1.95210159e-01
7.27436125e-01 -7.80070126e-01 5.14303386e-01 -4.34714049e-01
-6.93525612e-01 -3.32806677e-01 -8.62958848e-01 -1.33738741e-01
-1.65383220e+00 5.34262896e-01 6.28504992e-01 2.63155788e-01
-1.46930659e+00 8.19425583e-02 -3.22959661e-01 -1.01127958e+00
-3.09186932e-02 3.03331047e-01 -9.59490418e-01 -7.72741139e-02
-1.50196826e+00 1.04219770e+00 7.89684236e-01 -2.78480258e-02
-4.81945693e-01 -6.77772582e-01 -1.37677574e+00 1.21926442e-01
1.02751620e-01 -6.91400588e-01 6.83707237e-01 -4.25066084e-01
-9.01681960e-01 1.38434362e+00 8.78933668e-02 -3.66438806e-01
3.44956398e-01 1.91741690e-01 -8.43788922e-01 4.80266273e-01
5.88558555e-01 1.00167572e+00 6.41658664e-01 -1.01769865e+00
-7.95286417e-01 -2.50999779e-01 -9.00998339e-02 3.91507417e-01
-3.17130357e-01 8.14254060e-02 -1.11042070e+00 -4.83798712e-01
-4.76778112e-02 -8.34354401e-01 2.01439783e-01 -3.40755619e-02
-3.29494447e-01 -1.93761840e-01 4.81326967e-01 -6.70423746e-01
1.36180520e+00 -2.38859463e+00 -6.78885961e-03 3.84997934e-01
1.41789779e-01 -1.44859180e-01 -1.35066003e-01 1.01347113e+00
-2.57614762e-01 4.86124717e-02 3.69920760e-01 -3.39596897e-01
2.92535752e-01 -1.82970658e-01 -9.34205130e-02 4.84623313e-01
-2.21404225e-01 1.03607082e+00 -1.17288780e+00 -7.85057724e-01
5.13419986e-01 5.56935906e-01 2.58291978e-02 2.69954186e-02
3.12192798e-01 -4.94457744e-02 -3.82804006e-01 7.98002899e-01
5.39719701e-01 -2.80826479e-01 -4.80435602e-02 -4.67292726e-01
-3.90739888e-01 -2.28686333e-01 -1.33393550e+00 1.83127868e+00
-6.33432209e-01 8.17228079e-01 -3.43924969e-01 -6.95902348e-01
6.14228964e-01 3.28075618e-01 4.70387876e-01 -6.77037120e-01
3.86132561e-02 2.16441423e-01 -1.04802823e+00 -6.47005141e-01
1.19300079e+00 5.39868355e-01 -5.80407023e-01 3.47140193e-01
5.66911280e-01 2.77708799e-01 3.70292485e-01 5.87458313e-01
8.68064344e-01 -1.02635119e-02 4.88320708e-01 -7.26170912e-02
4.73593891e-01 1.92686945e-01 -5.49352504e-02 8.30091417e-01
1.52405119e-02 7.30663002e-01 3.85997705e-02 -4.26947117e-01
-8.32403123e-01 -1.10823977e+00 -9.30332914e-02 1.22747564e+00
4.81701434e-01 -7.80167699e-01 -4.14184779e-01 -5.46914101e-01
1.43473476e-01 4.90781367e-01 -8.07032049e-01 -2.01583258e-03
4.85527504e-04 -6.75342143e-01 5.29624104e-01 2.85908401e-01
5.70326388e-01 -9.65499520e-01 -3.39764744e-01 -8.38987902e-02
-3.14637601e-01 -9.06899691e-01 -5.52179158e-01 -1.97516575e-01
3.60759906e-02 -1.13714552e+00 -1.07319927e+00 -9.36920881e-01
7.24036992e-01 4.32279617e-01 1.17952228e+00 -1.61164682e-02
-6.29934132e-01 1.40638304e+00 -5.27140379e-01 -3.99137884e-01
3.00493360e-01 4.31959361e-01 -1.45777449e-01 3.07281435e-01
4.74677235e-01 -1.64537221e-01 -8.31117094e-01 3.53605598e-01
-1.04823852e+00 -1.29148915e-01 3.19395065e-01 5.22692621e-01
6.09327555e-01 -1.65520802e-01 4.62306947e-01 -6.01426840e-01
3.41221064e-01 -1.25798440e+00 -6.34370983e-01 6.01353943e-01
-2.66667843e-01 -1.85193688e-01 -1.09526210e-01 -2.05185506e-02
-8.23857665e-01 2.12338313e-01 1.28238499e-01 -2.57645130e-01
-2.63936698e-01 9.52614367e-01 9.97409746e-02 -3.15021992e-01
6.11120045e-01 2.09339708e-01 -4.95719910e-01 -3.42300743e-01
6.29699230e-01 8.65444303e-01 5.79517543e-01 -1.20426111e-01
6.43025279e-01 5.15132427e-01 -2.10588232e-01 -1.08051121e+00
-4.34351832e-01 -1.22336054e+00 -5.74807763e-01 -6.34746075e-01
1.02279341e+00 -1.28777146e+00 -5.26971519e-01 3.12089026e-01
-9.07235146e-01 1.62960086e-02 -6.51640967e-02 6.25819027e-01
-3.14337879e-01 -3.38223600e-03 -1.19007729e-01 -4.83929664e-01
-1.07628889e-01 -6.41817689e-01 1.30069959e+00 5.85154891e-01
1.25876084e-01 -1.33831310e+00 1.99530423e-01 9.44198743e-02
4.21025246e-01 1.64574370e-01 3.53455186e-01 -8.20591688e-01
-8.33451986e-01 -6.68519974e-01 -4.68631029e-01 -3.61741722e-01
-2.63770223e-01 -2.33355358e-01 -9.97840464e-01 -8.19972716e-03
-9.87979472e-01 -1.25421077e-01 9.47557032e-01 8.18289295e-02
7.12195039e-01 3.10596749e-02 -6.77358329e-01 6.02935553e-01
1.83513355e+00 -2.20320150e-01 7.28851557e-01 7.96999812e-01
6.85490131e-01 4.44441795e-01 8.04631412e-01 4.30204690e-01
8.44419658e-01 8.06801260e-01 4.92838830e-01 -1.40351862e-01
-6.10934943e-02 -4.02967125e-01 6.72373921e-02 5.84421813e-01
5.48322201e-02 -6.40893459e-01 -1.18520367e+00 1.14301550e+00
-2.19023156e+00 -1.07006764e+00 -3.83979790e-02 2.18403268e+00
1.08110994e-01 -5.92627645e-01 -1.75972492e-01 -4.62073058e-01
8.91211927e-01 4.64914054e-01 6.13426194e-02 1.27939448e-01
-1.11106634e-01 -2.59048223e-01 9.39101040e-01 3.58573645e-01
-1.39641833e+00 1.15037000e+00 5.38221121e+00 9.93998230e-01
-1.01154053e+00 4.17557329e-01 5.10965548e-02 2.19741702e-01
-1.71694860e-01 -2.03826595e-02 -1.04378796e+00 4.70829993e-01
9.38943803e-01 -4.27804381e-01 5.74921630e-02 1.01011634e+00
-2.48278245e-01 -5.36806941e-01 -6.94813311e-01 1.32801557e+00
5.07125795e-01 -1.76157606e+00 -4.38637659e-02 -1.43802866e-01
4.40155685e-01 4.76920754e-01 1.23556368e-02 3.94625336e-01
4.85795401e-02 -6.84523046e-01 7.99638212e-01 9.57233787e-01
8.81345332e-01 -6.31211460e-01 8.59641612e-01 -1.34661021e-02
-1.60891819e+00 1.25826150e-01 -3.59121919e-01 6.68915570e-01
4.33631092e-01 3.56251031e-01 -1.12964523e+00 9.35677111e-01
9.85387385e-01 8.08767378e-01 -1.08164537e+00 1.61916602e+00
-1.45391956e-01 -2.30419300e-02 -5.20349324e-01 7.59662762e-02
2.95688391e-01 6.20185994e-02 5.58231175e-01 1.58473849e+00
6.40921175e-01 -2.54081994e-01 5.74327633e-02 3.43319595e-01
-2.12196574e-01 3.93696398e-01 -1.06727409e+00 2.12675668e-02
6.04146659e-01 1.53238976e+00 -9.14161861e-01 -3.24209273e-01
-3.04843068e-01 1.23892784e+00 3.83937240e-01 4.61277723e-01
-8.03050756e-01 -8.70202839e-01 3.09647053e-01 4.63171564e-02
3.88197303e-01 -7.11245611e-02 6.18281364e-01 -1.20265174e+00
-2.59991914e-01 3.89967114e-02 9.47829902e-01 -1.38883984e+00
-1.04385304e+00 6.63434863e-01 3.02337259e-01 -1.28122926e+00
-2.15456769e-01 -5.19110620e-01 -4.21770453e-01 6.32332027e-01
-1.71481347e+00 -1.46047020e+00 -7.24362969e-01 8.08485329e-01
1.10817403e-01 -2.89917797e-01 1.04343748e+00 9.32299912e-01
-4.62735087e-01 4.01448280e-01 3.43520403e-01 3.14155638e-01
1.05431700e+00 -1.33849311e+00 1.60747260e-01 5.09355247e-01
5.39975286e-01 4.28479463e-01 2.56008834e-01 -5.45972228e-01
-1.24137962e+00 -1.26091158e+00 1.43731081e+00 -5.62442482e-01
8.20005894e-01 -1.63307920e-01 -3.65634710e-01 8.62254620e-01
3.22197974e-01 3.15925866e-01 8.48916650e-01 -1.76911995e-01
-2.26846293e-01 -1.21141575e-01 -8.88485193e-01 5.65934241e-01
6.71744585e-01 -8.71239841e-01 -4.01531577e-01 6.15574479e-01
2.80083865e-01 -4.90368187e-01 -1.10186219e+00 -1.59790099e-01
3.28743786e-01 -6.09068930e-01 1.41931415e+00 -1.79955155e-01
-9.15456414e-02 -6.49146736e-01 -2.85571575e-01 -1.35108447e+00
-1.56644866e-01 -1.14129350e-01 3.24656904e-01 1.33250320e+00
6.28940046e-01 -6.75311208e-01 3.02664757e-01 4.29806173e-01
-2.04388276e-02 -2.25489810e-02 -9.09337759e-01 -5.55529773e-01
-7.40582824e-01 -3.64729375e-01 5.03879905e-01 1.15741360e+00
1.65193170e-01 6.01089820e-02 -3.60699087e-01 6.85977936e-01
4.38718170e-01 -3.01481038e-02 7.75791883e-01 -9.96188700e-01
1.55133158e-01 -7.22633749e-02 -1.10611224e+00 -5.06935537e-01
1.87877774e-01 -1.24425721e+00 -1.63692340e-01 -1.94341838e+00
-3.32120545e-02 -4.42538798e-01 -5.05930126e-01 4.85863745e-01
1.85534894e-01 7.81570077e-01 1.72292128e-01 3.20677489e-01
-1.24736595e+00 2.91492879e-01 3.56709808e-01 -3.20666969e-01
9.59844291e-02 -2.99086839e-01 -2.64041364e-01 4.54220027e-01
6.97458327e-01 -6.98936164e-01 -2.58596808e-01 -3.88248652e-01
7.93121815e-01 -6.03897534e-02 7.94106185e-01 -1.06148958e+00
8.45059514e-01 3.29580635e-01 3.68754357e-01 -7.62504637e-01
4.85325396e-01 -8.46099854e-01 5.29604971e-01 -1.52967677e-01
-8.60619843e-02 2.79098868e-01 1.67725265e-01 7.85792232e-01
-6.90494537e-01 -5.06436586e-01 1.51454031e-01 -2.84847230e-01
-1.61623251e+00 4.30615634e-01 -3.75210524e-01 -1.19674906e-01
1.22510171e+00 -6.93110675e-02 -3.29482555e-01 -5.50413966e-01
-9.99437094e-01 4.05164123e-01 3.30747068e-01 5.46840489e-01
5.41199744e-01 -1.62973058e+00 -4.14622605e-01 -1.56739846e-01
9.51116383e-01 -4.83756393e-01 5.72958946e-01 8.19418669e-01
-7.89638281e-01 5.32668710e-01 -6.29317388e-02 -4.59870547e-01
-1.23098648e+00 3.79360139e-01 1.90890431e-01 -1.01715252e-01
-4.46679890e-01 7.94178843e-01 -1.26548320e-01 -5.33789992e-01
2.12317675e-01 3.46963078e-01 -6.01188183e-01 7.71509111e-01
7.38569438e-01 2.93956876e-01 2.76660293e-01 -1.05877459e+00
-5.77990174e-01 6.73114717e-01 1.71997830e-01 -3.10438365e-01
1.44758797e+00 -4.21540886e-01 -1.40239030e-01 4.33587492e-01
1.46163821e+00 -4.00205217e-02 -4.31035310e-01 -2.07968161e-01
2.66842991e-01 -5.08975804e-01 3.04871768e-01 -6.59259856e-01
-9.67325032e-01 4.81243342e-01 8.26611221e-01 2.38692611e-01
8.06796312e-01 3.12971979e-01 4.15531307e-01 4.11528260e-01
5.99142253e-01 -1.26434398e+00 -2.65576065e-01 4.21541214e-01
9.71396208e-01 -1.38448966e+00 3.24268974e-02 -1.47597492e-01
-7.73747206e-01 1.01091576e+00 1.71674162e-01 1.36459664e-01
1.15935254e+00 -4.05113548e-01 1.08437635e-01 -7.41758347e-01
-2.98155040e-01 -6.66549683e-01 5.41094661e-01 6.34727061e-01
3.88300568e-02 -1.81809738e-02 -1.00472160e-01 2.02618957e-01
4.89604659e-02 -2.23116532e-01 1.86754897e-01 9.29819345e-01
-2.83028454e-01 -7.92539716e-01 -4.79006499e-01 1.15053348e-01
-2.72412837e-01 -2.90853083e-01 -2.43994057e-01 9.15030479e-01
7.94263706e-02 6.64752305e-01 1.40972272e-01 -2.16286972e-01
2.70284891e-01 1.43877387e-01 -3.66202369e-02 -5.09496212e-01
-6.13122940e-01 -3.58935505e-01 3.14245999e-01 -7.19570935e-01
-4.34336692e-01 -4.17579383e-01 -9.30549145e-01 -2.26576239e-01
-2.96836585e-01 4.05886233e-01 1.25630879e+00 3.54170948e-01
8.49379361e-01 2.38042161e-01 2.48577565e-01 -1.27396560e+00
4.61749852e-01 -5.80087841e-01 -7.76981831e-01 3.10291529e-01
-1.32521447e-02 -7.60958433e-01 -3.19489777e-01 5.04594818e-02] | [7.6894354820251465, -1.8071354627609253] |
2e7d57cf-b7cd-40c3-afe1-c49ce6391f7b | stratified-graphical-models-context-specific | 1309.6415 | null | http://arxiv.org/abs/1309.6415v2 | http://arxiv.org/pdf/1309.6415v2.pdf | Stratified Graphical Models - Context-Specific Independence in Graphical Models | Theory of graphical models has matured over more than three decades to
provide the backbone for several classes of models that are used in a myriad of
applications such as genetic mapping of diseases, credit risk evaluation,
reliability and computer security, etc. Despite of their generic applicability
and wide adoptance, the constraints imposed by undirected graphical models and
Bayesian networks have also been recognized to be unnecessarily stringent under
certain circumstances. This observation has led to the proposal of several
generalizations that aim at more relaxed constraints by which the models can
impose local or context-specific dependence structures. Here we consider an
additional class of such models, termed as stratified graphical models. We
develop a method for Bayesian learning of these models by deriving an
analytical expression for the marginal likelihood of data under a specific
subclass of decomposable stratified models. A non-reversible Markov chain Monte
Carlo approach is further used to identify models that are highly supported by
the posterior distribution over the model space. Our method is illustrated and
compared with ordinary graphical models through application to several real and
synthetic datasets. | ['Timo Koski', 'Jukka Corander', 'Henrik Nyman', 'Johan Pensar'] | 2013-09-25 | null | null | null | null | ['computer-security'] | ['miscellaneous'] | [ 5.61899900e-01 3.55471522e-02 -2.78419614e-01 -6.90319896e-01
-2.30063781e-01 -3.88544977e-01 8.70280683e-01 1.63678974e-01
-2.81228900e-01 9.58509326e-01 -8.88633355e-02 -6.19039953e-01
-8.07612836e-01 -9.26643968e-01 -3.23814243e-01 -8.23494375e-01
-4.11972493e-01 6.98899865e-01 4.20252055e-01 1.93628967e-01
4.15527791e-01 7.11496413e-01 -1.53655159e+00 -1.62966892e-01
8.79695356e-01 3.39108825e-01 -3.60271186e-02 4.94160652e-01
-4.45153527e-02 2.86244869e-01 -3.34517539e-01 -4.39638942e-01
-2.54274696e-01 -5.49585700e-01 -5.78847051e-01 1.68452710e-01
-3.74921352e-01 -6.30159229e-02 -2.50391085e-02 1.09732234e+00
2.38455072e-01 -1.22196689e-01 9.21846032e-01 -1.16363251e+00
-2.01548830e-01 5.81200719e-01 -5.99127352e-01 8.73227939e-02
2.87329882e-01 -3.85304451e-01 9.29275930e-01 -6.28586829e-01
5.54898560e-01 1.43776345e+00 4.95774925e-01 5.93126416e-02
-1.63889992e+00 -3.72068048e-01 1.43478081e-01 1.72102243e-01
-1.56293869e+00 -1.11506790e-01 6.59351885e-01 -3.35610896e-01
4.44707841e-01 3.20974052e-01 5.35166502e-01 8.91927183e-01
3.32425922e-01 4.48438317e-01 1.50740635e+00 -6.97355568e-01
4.06016082e-01 6.48589805e-02 4.04515088e-01 7.38033175e-01
6.71409011e-01 1.45013258e-01 -4.11195129e-01 -8.12599182e-01
6.08551860e-01 -4.46218662e-02 -3.22717205e-02 -7.00903237e-01
-7.10376441e-01 1.15679526e+00 -1.53918773e-01 2.25552335e-01
-5.12351990e-02 -1.20732948e-01 4.25878502e-02 -1.29363596e-01
3.26309681e-01 -1.33979633e-01 -1.28787920e-01 1.36299357e-01
-8.36724162e-01 3.24610382e-01 9.51028526e-01 6.15568101e-01
6.44931793e-01 -7.02511072e-02 1.01769723e-01 6.22972608e-01
7.98235238e-01 4.54303741e-01 -5.58958054e-02 -4.98737991e-01
1.20731279e-01 4.55841333e-01 3.01499903e-01 -8.37083936e-01
-4.30970609e-01 -4.06700253e-01 -9.31221128e-01 -8.05086941e-02
6.09472513e-01 9.99055803e-02 -8.57753038e-01 1.83757126e+00
3.31968606e-01 2.14051045e-02 -2.29465708e-01 2.94770062e-01
8.61498192e-02 7.03757942e-01 6.46192878e-02 -5.35364211e-01
1.03290915e+00 -3.92242670e-02 -7.66427100e-01 2.55329311e-01
4.34050262e-01 -5.14286041e-01 6.73075974e-01 6.01872146e-01
-8.53879511e-01 -9.80770402e-03 -7.60937035e-01 6.87421203e-01
-1.77492723e-01 -3.20084512e-01 7.41252780e-01 1.10523784e+00
-8.94672692e-01 5.64324915e-01 -1.01091826e+00 -4.81892645e-01
1.71794981e-01 4.68763351e-01 2.06064787e-02 -7.48294741e-02
-1.19638467e+00 8.59367013e-01 2.72677928e-01 4.14312690e-01
-6.96686387e-01 -1.05927713e-01 -5.16406119e-01 6.61943853e-02
3.09119761e-01 -5.26247919e-01 7.79949069e-01 -6.86700761e-01
-1.28404474e+00 7.26699114e-01 -1.92311317e-01 -3.63307267e-01
6.03320956e-01 2.27478296e-01 -5.53427756e-01 -1.94725767e-02
-2.20860928e-01 -3.16597670e-02 7.00900018e-01 -1.12428308e+00
-2.24632815e-01 -5.24033427e-01 -1.37243539e-01 -7.42015149e-03
-4.66015674e-02 3.77294570e-01 -3.24184239e-01 -5.56973696e-01
3.96557063e-01 -1.08919263e+00 -5.04490316e-01 -3.02232891e-01
-8.47508073e-01 -5.69868125e-02 4.91083652e-01 -3.98544371e-01
1.37934697e+00 -1.81992269e+00 2.89051026e-01 9.35425341e-01
-2.40379304e-01 -4.35047299e-02 3.02568018e-01 8.42098355e-01
-6.49846196e-02 2.98353314e-01 -7.68960118e-01 -3.54351569e-03
-2.32518502e-02 3.61931324e-01 -2.64168710e-01 7.66327202e-01
1.86734572e-01 3.30378085e-01 -5.69314361e-01 -5.17524421e-01
1.56598181e-01 6.02853537e-01 -4.11245853e-01 -5.24028065e-03
-6.55797198e-02 3.78076136e-01 -6.92544520e-01 4.04192865e-01
6.34538054e-01 -4.17437822e-01 8.79078925e-01 4.54520434e-01
-2.26597458e-01 1.60050057e-02 -1.51736975e+00 1.05726993e+00
3.01775485e-01 1.40158832e-01 -1.75875917e-01 -1.15221870e+00
9.32537496e-01 2.34206334e-01 2.41660148e-01 -9.59974825e-02
1.08727373e-01 4.62834314e-02 1.26230225e-01 -3.57924521e-01
1.96806222e-01 -3.99301082e-01 -2.58902997e-01 5.15543461e-01
-2.73669451e-01 1.59518886e-02 1.05474159e-01 1.35529876e-01
7.78229594e-01 2.20820725e-01 6.28908455e-01 -3.97298604e-01
4.87074345e-01 -2.33256117e-01 6.45823836e-01 9.46198821e-01
7.29116648e-02 2.61110038e-01 7.71616697e-01 -2.02242613e-01
-8.52599502e-01 -1.22993255e+00 -6.39811397e-01 5.15325487e-01
5.07118329e-02 -1.45352110e-01 -6.47488058e-01 -3.10400069e-01
-4.49164547e-02 5.81621587e-01 -5.80008745e-01 -1.22498676e-01
-3.47786576e-01 -1.58870935e+00 5.72263241e-01 1.93014234e-01
1.34426594e-01 -6.01496041e-01 -4.27558750e-01 1.77461952e-01
2.02131778e-01 -5.95919371e-01 3.34524155e-01 2.21113607e-01
-1.23601317e+00 -1.07580709e+00 -6.56048536e-01 -1.96413621e-01
7.42942214e-01 -1.40613895e-02 7.27778137e-01 1.05953828e-01
-1.96200624e-01 3.80204052e-01 -2.08012551e-01 -2.39547506e-01
-7.18986511e-01 -6.66166991e-02 2.51309127e-01 2.52701759e-01
2.21276253e-01 -5.16664565e-01 -1.33822665e-01 4.63287860e-01
-1.25053704e+00 -9.33635682e-02 5.23269117e-01 7.66054153e-01
4.89640176e-01 1.75474510e-01 6.02734506e-01 -1.08094621e+00
4.50261056e-01 -6.53613150e-01 -1.07926893e+00 5.82262874e-01
-7.90100515e-01 2.81730920e-01 3.05267572e-02 -1.39895290e-01
-1.40847576e+00 7.18251467e-02 6.78172559e-02 3.42585713e-01
-2.61747330e-01 1.04196787e+00 -4.29645419e-01 -8.57270285e-02
2.38166079e-01 1.68548554e-01 -4.81361598e-02 -5.70387840e-01
1.96831152e-02 5.69672346e-01 1.28870696e-01 -4.75479633e-01
6.60329521e-01 4.80249256e-01 6.29639983e-01 -1.05276394e+00
-5.62373996e-01 -5.61282709e-02 -6.60048246e-01 -2.80564278e-01
7.28150606e-01 -3.58159959e-01 -4.94043052e-01 5.86527467e-01
-9.53234911e-01 1.69392563e-02 4.00632322e-01 6.59608901e-01
-5.98417878e-01 8.04907560e-01 -3.62331241e-01 -1.30332708e+00
4.94228214e-01 -1.08969784e+00 3.58802855e-01 1.14137046e-01
-1.51322931e-01 -1.32802069e+00 4.10560668e-01 7.43144229e-02
-5.94541011e-03 2.85859287e-01 1.45098281e+00 -5.97827792e-01
-6.06069803e-01 -4.75941062e-01 2.87355520e-02 -1.94671401e-03
2.53568422e-02 5.59547961e-01 -7.75340497e-01 -2.24239379e-01
-2.66100258e-01 7.86366761e-02 7.09997296e-01 5.58481991e-01
9.43462670e-01 9.25792605e-02 -6.51398420e-01 2.87511557e-01
1.43079388e+00 4.07085299e-01 7.05613017e-01 -1.22348972e-01
1.65619478e-01 7.81924188e-01 4.65128243e-01 5.26186049e-01
5.30968346e-02 8.36114764e-01 4.39656466e-01 1.87815979e-01
5.43448448e-01 -2.28509009e-01 1.40261859e-01 5.75449288e-01
-2.70288169e-01 -4.58408862e-01 -1.01057482e+00 2.65253067e-01
-1.90812671e+00 -1.01528347e+00 -5.49191475e-01 2.68332672e+00
6.80375516e-01 3.14962029e-01 3.29035133e-01 2.64840007e-01
1.13934529e+00 -1.10052660e-01 -4.49736744e-01 -2.99877197e-01
-2.50796765e-01 1.73690692e-01 4.58568633e-01 6.31318927e-01
-7.98153758e-01 3.97035003e-01 7.65433311e+00 6.50804818e-01
-5.83247542e-01 -2.27192506e-01 6.21491790e-01 4.70829964e-01
-4.29225236e-01 3.95651340e-01 -9.76177752e-01 4.80228573e-01
1.01346767e+00 -1.00386895e-01 -1.05336094e-02 3.13771695e-01
3.56831998e-01 -5.48648357e-01 -7.88680017e-01 3.78968179e-01
-2.26053685e-01 -1.04719865e+00 1.41814694e-01 4.87805814e-01
7.44757712e-01 -4.20206070e-01 4.80847955e-02 -2.53622830e-01
5.38348436e-01 -8.58770967e-01 4.98892814e-01 7.55996704e-01
5.18101752e-01 -8.76351714e-01 5.36622763e-01 4.44531441e-01
-6.78041577e-01 -2.35155108e-03 -1.75331920e-01 -2.44500816e-01
4.69811231e-01 9.00727570e-01 -8.55803728e-01 8.19983423e-01
2.30571330e-01 3.21107358e-01 -3.56929034e-01 1.36027086e+00
-2.45086581e-01 8.27942252e-01 -3.97663325e-01 -1.25627443e-01
5.43100201e-02 -3.40779364e-01 5.68460524e-01 8.74025643e-01
3.37525487e-01 -8.79465267e-02 -1.76033556e-01 7.24693537e-01
4.98527199e-01 -5.44346757e-02 -6.18154109e-01 -9.83808488e-02
5.69644928e-01 7.59393752e-01 -1.23270631e+00 -1.12020716e-01
-3.74915481e-01 4.20877010e-01 4.71736193e-02 4.51924384e-01
-7.46044457e-01 -2.53970951e-01 1.24234676e-01 1.72122732e-01
2.19088405e-01 -3.87629241e-01 -4.81594354e-02 -1.10587323e+00
-4.00372446e-01 -5.44273734e-01 6.36150956e-01 -4.38709855e-01
-1.14398277e+00 2.78991163e-01 6.48221731e-01 -9.47092593e-01
-4.93458599e-01 -5.97225845e-01 -4.92030352e-01 9.33073759e-01
-1.37636971e+00 -7.27861166e-01 2.07097784e-01 4.76714611e-01
-1.26362175e-01 3.30030769e-02 9.31265891e-01 -1.04883453e-02
-7.02924728e-01 5.28400578e-02 4.56832945e-01 -2.41571695e-01
4.21047509e-01 -1.03693807e+00 2.47975923e-02 9.12791848e-01
-2.64263581e-02 6.93814039e-01 1.07929623e+00 -9.47055697e-01
-1.09142184e+00 -6.86364233e-01 1.16219628e+00 -1.71701297e-01
6.90520823e-01 -4.25731212e-01 -8.33368242e-01 6.05932534e-01
-2.85079032e-01 -3.57104331e-01 8.23738515e-01 2.50528395e-01
-2.49698788e-01 6.21723896e-03 -1.00983059e+00 5.25036156e-01
6.74387395e-01 -1.89804733e-01 -5.06374955e-01 1.84153050e-01
6.40380681e-02 2.86634147e-01 -6.12311721e-01 4.41433847e-01
6.40036583e-01 -1.25108695e+00 7.67520368e-01 -7.57405937e-01
-6.24065474e-02 -4.15672094e-01 -1.63696736e-01 -8.48130345e-01
-2.31148392e-01 -6.98983431e-01 1.43008366e-01 1.03211761e+00
3.85854214e-01 -7.74496078e-01 6.35499120e-01 5.45883119e-01
2.02924356e-01 -5.13298988e-01 -1.15912259e+00 -7.02636480e-01
-1.20478854e-01 -2.66259670e-01 4.36477602e-01 8.20478976e-01
-3.57623957e-02 -6.46229535e-02 -4.47548807e-01 2.28609905e-01
9.06427383e-01 1.73554674e-01 4.35319424e-01 -1.72200453e+00
-4.81867790e-01 -2.24166140e-01 -5.32416999e-01 -6.85928047e-01
1.95678413e-01 -5.77419698e-01 -2.66177952e-02 -1.25906956e+00
4.86270458e-01 -6.81980014e-01 -2.26905376e-01 7.02582970e-02
-3.64523493e-02 1.39840662e-01 -2.46357292e-01 2.29659155e-01
-2.12291494e-01 2.49534681e-01 6.85062647e-01 3.36026818e-01
-3.15569416e-02 5.44961512e-01 -4.56898034e-01 9.08258200e-01
6.53084993e-01 -6.59575880e-01 -4.01100844e-01 5.18850014e-02
6.97180748e-01 3.07067901e-01 4.73859847e-01 -5.72025657e-01
-1.71494454e-01 -4.13360536e-01 1.56721562e-01 -7.01340139e-01
2.62956113e-01 -5.89735031e-01 8.47095788e-01 6.99514270e-01
-3.69421452e-01 -9.50817913e-02 -2.28041247e-01 9.36216772e-01
-8.20015669e-02 -5.80601156e-01 7.96602786e-01 4.12515029e-02
-2.06034526e-01 3.43976095e-02 -7.39334702e-01 -4.71937865e-01
1.09888434e+00 -3.08822691e-01 -6.05714449e-04 -5.73082209e-01
-1.10365868e+00 -1.32335216e-01 4.62291837e-01 -1.16273753e-01
4.08600658e-01 -9.34428096e-01 -5.07404685e-01 1.88385710e-01
-6.47689998e-02 -4.63805735e-01 3.65318745e-01 9.33537185e-01
-3.70753974e-01 5.56040943e-01 -5.67722172e-02 -6.60258889e-01
-1.34805977e+00 4.87028241e-01 1.53653473e-01 -5.32158732e-01
-2.93475807e-01 2.97972262e-01 6.16410002e-02 -8.53829682e-02
1.30102097e-03 -8.45434293e-02 -1.36247993e-01 -5.06858751e-02
3.05334598e-01 5.68487108e-01 -1.86039850e-01 -6.22136652e-01
-3.89464080e-01 3.78873318e-01 7.46696442e-02 -3.21141690e-01
1.31496429e+00 -2.29953110e-01 -3.02359790e-01 7.72685409e-01
5.56301653e-01 -4.74645086e-02 -1.12225878e+00 -1.43210873e-01
5.24299920e-01 -2.34736145e-01 -2.43925229e-01 -6.88415408e-01
-4.52127010e-01 8.45733106e-01 2.89869189e-01 7.40789771e-01
8.92654538e-01 -2.50551514e-02 -2.38221228e-01 2.79751241e-01
4.52943057e-01 -6.96777225e-01 -4.00112808e-01 3.20222735e-01
4.56335008e-01 -7.17094123e-01 3.18103135e-01 -6.14450693e-01
-2.07415208e-01 1.07687795e+00 -2.80413240e-01 3.09830322e-03
9.09386873e-01 2.23046228e-01 -4.97435927e-01 -6.09974435e-04
-6.24770463e-01 2.01458801e-02 1.25519127e-01 6.48431778e-01
4.06222701e-01 1.77145779e-01 -6.04156613e-01 3.83428872e-01
3.47395718e-01 9.70695540e-02 7.24531591e-01 1.05089104e+00
-5.33315539e-01 -1.42434132e+00 -5.69882512e-01 4.80835885e-01
-5.57470202e-01 1.14530578e-01 -3.99475455e-01 8.93480301e-01
-3.64951074e-01 8.72555494e-01 -6.06020726e-02 2.18314081e-01
-1.24548212e-01 3.48230720e-01 6.27582967e-01 -4.31847721e-01
-6.23034425e-02 2.60889739e-01 2.14811832e-01 -2.19072495e-02
-8.28080118e-01 -1.07930398e+00 -7.92102695e-01 -2.29402512e-01
-5.87599993e-01 4.28326160e-01 6.10798061e-01 1.09067547e+00
4.30683084e-02 9.98247787e-02 5.93935549e-01 -6.44276619e-01
-6.51453316e-01 -8.24134946e-01 -1.04516494e+00 -1.90319359e-01
-4.98314649e-02 -9.48725641e-01 -4.14935172e-01 -1.48951299e-02] | [7.1421332359313965, 4.692811489105225] |
4872d780-93ef-43d0-bef2-3bafaad6b485 | the-sound-of-silence-efficiency-of-first | 2210.02746 | null | https://arxiv.org/abs/2210.02746v1 | https://arxiv.org/pdf/2210.02746v1.pdf | The Sound of Silence: Efficiency of First Digit Features in Synthetic Audio Detection | The recent integration of generative neural strategies and audio processing techniques have fostered the widespread of synthetic speech synthesis or transformation algorithms. This capability proves to be harmful in many legal and informative processes (news, biometric authentication, audio evidence in courts, etc.). Thus, the development of efficient detection algorithms is both crucial and challenging due to the heterogeneity of forgery techniques. This work investigates the discriminative role of silenced parts in synthetic speech detection and shows how first digit statistics extracted from MFCC coefficients can efficiently enable a robust detection. The proposed procedure is computationally-lightweight and effective on many different algorithms since it does not rely on large neural detection architecture and obtains an accuracy above 90\% in most of the classes of the ASVSpoof dataset. | ['Simone Milani', 'Federica Latora', 'Daniele Mari'] | 2022-10-06 | null | null | null | null | ['synthetic-speech-detection'] | ['audio'] | [ 5.47040641e-01 -1.15748927e-01 2.29609028e-01 9.44839343e-02
-9.64879930e-01 -6.55611753e-01 8.83265436e-01 1.34633273e-01
-4.06715125e-01 8.58588219e-01 -5.37294038e-02 -3.94064635e-01
-1.70119151e-01 -4.60108370e-01 -1.14059031e-01 -8.91683400e-01
1.66030765e-01 6.75060004e-02 3.60919178e-01 -2.11945385e-01
3.55448186e-01 8.31182241e-01 -1.82614923e+00 3.58861208e-01
6.66873813e-01 1.02396464e+00 1.24481551e-01 8.52787435e-01
4.98215389e-03 5.15443623e-01 -1.18985200e+00 -6.08451009e-01
1.04213944e-02 -7.54395485e-01 -2.55986452e-01 -9.69166011e-02
1.44796297e-01 -4.75097187e-02 -1.21862426e-01 1.13587379e+00
9.61410046e-01 -4.97840345e-02 7.44794667e-01 -7.48083532e-01
-2.28224039e-01 6.97164357e-01 -1.54199421e-01 6.70847178e-01
5.18223941e-01 3.05517968e-02 6.38336837e-01 -1.02916181e+00
3.59910309e-01 8.77928257e-01 5.67532361e-01 2.04370663e-01
-1.06496441e+00 -6.62246585e-01 -6.38830602e-01 3.14687312e-01
-1.39719784e+00 -9.17285919e-01 9.06381369e-01 -2.21615210e-01
9.43078756e-01 5.44288397e-01 5.25197983e-01 1.56322181e+00
2.49672346e-02 6.28790677e-01 1.14484596e+00 -7.14644909e-01
1.93454519e-01 2.79145896e-01 -1.12826973e-01 4.87518817e-01
4.23149079e-01 3.55908334e-01 -5.45166373e-01 -1.15101583e-01
4.21143442e-01 -5.55334628e-01 -2.04761386e-01 2.39666671e-01
-7.44370997e-01 7.14265049e-01 -4.26832378e-01 8.96579862e-01
-2.79241204e-01 -5.54389507e-03 5.08614123e-01 4.19712633e-01
1.64876446e-01 3.18882465e-01 8.19562823e-02 -4.84155864e-01
-1.28810990e+00 1.90317720e-01 7.36363888e-01 5.20661891e-01
-3.00968215e-02 7.17873573e-01 1.09653130e-01 9.02668536e-01
1.35611355e-01 4.60241765e-01 6.73489630e-01 -6.39969468e-01
3.47231746e-01 3.14059675e-01 -7.42807314e-02 -1.12365484e+00
-2.53871083e-01 -7.50455439e-01 -7.76809096e-01 1.55975580e-01
3.83801520e-01 -6.16474785e-02 -8.24357092e-01 1.34068191e+00
2.09191024e-01 -8.38954151e-02 3.72135229e-02 3.87015998e-01
5.23787260e-01 5.78321636e-01 -2.19229490e-01 -2.54104465e-01
1.15991235e+00 -1.07570954e-01 -9.21753824e-01 -1.46266386e-01
2.66014934e-01 -1.15375388e+00 6.54437840e-01 7.25288272e-01
-1.05199909e+00 -6.18857205e-01 -1.31164610e+00 3.66406858e-01
-4.93785948e-01 4.72494364e-01 3.24766159e-01 1.54914713e+00
-8.70860696e-01 4.17155653e-01 -5.57024598e-01 -1.00426646e-02
1.72539294e-01 3.14794302e-01 -2.20818073e-01 2.03878999e-01
-1.13666272e+00 8.38051736e-01 5.55038452e-01 2.21011862e-01
-6.44715786e-01 -1.21222109e-01 -6.08534276e-01 1.14736572e-01
2.05942929e-01 -1.19558841e-01 9.26288724e-01 -6.77820981e-01
-1.49748039e+00 7.98392296e-01 2.22890645e-01 -8.41865122e-01
7.93096840e-01 -5.24830297e-02 -8.88844848e-01 3.40837985e-01
-2.04679146e-01 -3.94484363e-02 1.48571908e+00 -7.08342969e-01
-5.47657430e-01 -2.19054118e-01 -5.63134015e-01 -2.53918558e-01
-1.14324532e-01 2.73358822e-01 1.22177996e-01 -1.06675351e+00
1.32771239e-01 -6.75151408e-01 1.93854362e-01 -6.71020269e-01
-4.35915202e-01 1.56789109e-01 9.50859547e-01 -8.71064246e-01
1.25577295e+00 -2.32320571e+00 -2.28218451e-01 3.50770354e-01
-2.63514131e-01 8.00957978e-01 2.61939347e-01 6.42788708e-01
-9.61653739e-02 -2.62996666e-02 -2.53489107e-01 -2.27507845e-01
1.06732575e-02 2.46978179e-02 -5.81355214e-01 5.67704439e-01
1.02927186e-01 4.56430703e-01 -5.18635213e-01 -4.79477733e-01
4.88592386e-01 7.18411088e-01 -2.98309237e-01 -6.37680218e-02
3.25715572e-01 1.80931747e-01 -1.83982298e-01 8.37871969e-01
2.77503043e-01 4.01970059e-01 1.73565090e-01 7.36190304e-02
-1.82119817e-01 3.15810591e-01 -1.27201080e+00 1.24054646e+00
-2.80244470e-01 7.49869466e-01 5.22966757e-02 -1.10412097e+00
1.09720814e+00 7.47554898e-01 1.38049453e-01 -6.62437916e-01
5.60928762e-01 6.04601085e-01 2.23842144e-01 -3.81229162e-01
4.94200438e-01 -1.46133497e-01 -1.28059173e-02 1.09227695e-01
3.38233024e-01 -2.88740307e-01 1.56343400e-01 -1.79707006e-01
9.57931638e-01 -1.17011912e-01 4.22734082e-01 3.51927206e-02
9.01164651e-01 -5.33183515e-01 2.83472329e-01 7.83643782e-01
-2.74697274e-01 4.91403550e-01 2.61760145e-01 -4.98329885e-02
-8.79131377e-01 -9.80667710e-01 -8.26093853e-02 6.32261336e-01
-4.11509722e-01 -8.56597126e-02 -9.12707806e-01 -2.53625125e-01
-3.81790578e-01 8.19018304e-01 -2.40758628e-01 -3.39226186e-01
-5.58129370e-01 -7.27986097e-01 1.12337279e+00 1.19930454e-01
3.45070779e-01 -1.18448973e+00 -7.10427880e-01 5.14599442e-01
-1.46089867e-02 -1.35916221e+00 6.22601993e-02 2.35980153e-01
-7.49694526e-01 -9.26299155e-01 -6.62210345e-01 -5.76436162e-01
1.16142437e-01 -1.27629563e-01 5.48481762e-01 -7.42658749e-02
-5.86240351e-01 5.74136563e-02 -5.85934997e-01 -4.97142375e-01
-9.84876394e-01 1.72244608e-01 1.32042384e-02 1.72948793e-01
1.83927819e-01 -6.37444675e-01 -1.54593408e-01 1.25355095e-01
-8.20670128e-01 -4.96404886e-01 7.05543578e-01 7.68654823e-01
8.85203928e-02 1.38344228e-01 8.99476945e-01 -6.42604113e-01
1.11916220e+00 -8.32963660e-02 -7.34570503e-01 1.45462155e-01
-5.05760849e-01 -1.76011741e-01 5.91895878e-01 -3.39693338e-01
-1.08537948e+00 -6.23764470e-02 -6.45627141e-01 -1.76459968e-01
-3.88892978e-01 9.95602906e-02 -1.76504344e-01 -2.51660883e-01
7.57871091e-01 6.62419021e-01 -2.78537840e-01 -4.35619235e-01
1.17650643e-01 8.17451298e-01 7.83265293e-01 -2.37105370e-01
8.84630442e-01 1.95021957e-01 1.41094932e-02 -1.53051388e+00
-1.09311059e-01 -3.33585113e-01 -2.85700858e-01 -2.61384547e-01
6.36209667e-01 -3.74120593e-01 -4.83595341e-01 6.32302165e-01
-1.16635251e+00 4.22872126e-01 -1.16493620e-01 5.77894092e-01
-3.64964664e-01 8.34141433e-01 -6.26991093e-01 -1.49611127e+00
-4.81365949e-01 -1.23260128e+00 7.29954243e-01 1.00237899e-01
-2.74449468e-01 -4.96715456e-01 -1.25621241e-02 4.65980768e-01
3.43890458e-01 1.87591597e-01 6.72273219e-01 -9.08205330e-01
-3.90880406e-01 -7.18491495e-01 1.93601653e-01 7.03810155e-01
2.02948973e-02 2.32371658e-01 -1.31364167e+00 -2.82111857e-03
2.61783957e-01 -8.95335823e-02 6.92691863e-01 2.29764014e-01
7.06530511e-01 -1.11728139e-01 1.00438394e-01 2.21700072e-01
1.07838023e+00 8.24577510e-01 7.76028037e-01 -7.25081284e-03
6.86546490e-02 4.28816885e-01 3.09108913e-01 4.65515137e-01
-5.09300053e-01 6.26738787e-01 2.08753005e-01 1.68989629e-01
-1.67513505e-01 4.11076508e-02 5.30555069e-01 1.01328504e+00
-3.03031772e-01 -3.87055695e-01 -6.76747859e-01 4.75503772e-01
-1.24363875e+00 -1.37621307e+00 -7.85507336e-02 2.19236374e+00
4.65718359e-01 4.88127917e-01 4.79363427e-02 1.12206972e+00
7.71515071e-01 1.45474389e-01 -1.83317177e-02 -7.21368968e-01
-4.58266705e-01 7.47320831e-01 3.72134894e-01 3.02787334e-01
-1.07604337e+00 7.52853572e-01 6.57366276e+00 1.22456956e+00
-9.60895240e-01 1.82013124e-01 3.67976576e-01 9.91917122e-03
1.60255641e-01 -6.18421912e-01 -6.39708340e-01 6.92467093e-01
1.18457615e+00 1.90064386e-01 1.10044040e-01 5.80602825e-01
2.46979624e-01 -3.16315114e-01 -6.05661631e-01 1.11317670e+00
2.60078132e-01 -1.12766707e+00 -2.60595810e-02 5.10097854e-02
3.67042124e-01 -3.14985007e-01 3.99242878e-01 4.71696854e-02
-2.68556654e-01 -9.63561535e-01 8.69514287e-01 1.10806912e-01
6.65845871e-01 -1.10423934e+00 8.08910251e-01 3.70931208e-01
-8.75049949e-01 -1.11337520e-01 -6.62119687e-02 2.88319647e-01
2.06489712e-01 4.37415838e-01 -1.24861979e+00 5.39285302e-01
3.27700317e-01 -1.22281685e-01 -3.18007916e-01 1.15180540e+00
-3.06746751e-01 8.58734965e-01 -4.00900453e-01 -9.97565910e-02
2.06973016e-01 -1.77282542e-01 7.30004191e-01 1.44004393e+00
5.97797394e-01 -2.80361444e-01 -4.36441690e-01 4.91485506e-01
1.99855819e-01 1.57449350e-01 -6.95756853e-01 -2.68353939e-01
4.38322574e-01 9.47532177e-01 -9.80760098e-01 -6.71746507e-02
-2.41978779e-01 8.65517735e-01 -3.87364179e-01 1.30778238e-01
-8.64920199e-01 -5.64832747e-01 3.76515299e-01 1.67289689e-01
3.70231539e-01 -3.40032369e-01 -9.93463472e-02 -7.14453399e-01
-1.18278071e-01 -1.09908652e+00 1.88052267e-01 -3.09762686e-01
-6.51642323e-01 8.10058057e-01 -1.56151518e-01 -1.07780099e+00
-7.29210138e-01 -5.77439725e-01 -4.80088800e-01 7.24974334e-01
-1.12187994e+00 -9.32644248e-01 1.79675803e-01 5.55182457e-01
7.05654442e-01 -6.13832474e-01 8.90327692e-01 7.17021286e-01
-5.24911106e-01 6.61914706e-01 7.89738074e-02 1.44824952e-01
2.56738931e-01 -8.26720059e-01 3.21884513e-01 1.28283393e+00
5.04859567e-01 4.11544234e-01 9.59423423e-01 -4.78117734e-01
-9.28209305e-01 -3.90898138e-01 9.63254213e-01 -6.02044025e-03
4.91081327e-01 -3.41714859e-01 -7.54270732e-01 -6.53445572e-02
1.67014822e-01 -4.46413755e-01 7.09304333e-01 -3.89563590e-01
-8.70883912e-02 -2.38677803e-02 -1.10775423e+00 4.44182694e-01
6.64673567e-01 -9.01530981e-01 -9.55395401e-01 3.70433070e-02
9.45998952e-02 -7.52120912e-02 -3.05330694e-01 2.31712818e-01
5.18025160e-01 -1.45421755e+00 1.12850451e+00 -2.16014653e-01
-2.04473883e-01 -1.28549308e-01 -1.85774639e-01 -7.77081728e-01
2.18462855e-01 -9.10079479e-01 -1.53763294e-01 1.35026646e+00
5.64536810e-01 -6.17799699e-01 7.61310816e-01 2.55008154e-02
-6.77048638e-02 -3.40341479e-01 -1.37631762e+00 -7.91638315e-01
-6.00427687e-01 -7.09543228e-01 1.13908641e-01 6.86732113e-01
-2.25877777e-01 3.77643816e-02 -5.34458280e-01 2.25531193e-03
5.64426363e-01 -6.13360293e-02 4.01654303e-01 -1.11791670e+00
-3.97770137e-01 -5.17382264e-01 -7.57625759e-01 -3.13480049e-01
-9.42365900e-02 -5.33368826e-01 -2.89003134e-01 -8.74735832e-01
-6.32390738e-01 -5.10847569e-02 -2.35658959e-01 5.71039729e-02
1.94346160e-01 4.14802343e-01 7.53992572e-02 -1.39609739e-01
3.34829390e-02 3.29533100e-01 5.86763501e-01 -8.30003321e-02
9.38038528e-03 4.14673209e-01 -1.36465624e-01 7.83496082e-01
8.58192623e-01 -6.83684170e-01 -4.46207553e-01 3.29017192e-01
2.26567850e-01 1.70572892e-01 2.24665508e-01 -1.48167980e+00
1.62704468e-01 4.37910378e-01 4.00981814e-01 -7.12490380e-01
6.64480627e-01 -7.29592502e-01 3.26065123e-01 7.31644869e-01
1.72643550e-02 7.76628479e-02 2.75354609e-02 5.01518011e-01
-4.33337420e-01 -5.63350737e-01 9.16281581e-01 -1.32464662e-01
-4.94292855e-01 -3.70216221e-01 -7.07446992e-01 -1.39876530e-01
8.82638872e-01 -7.08586216e-01 1.37985973e-02 -4.42973614e-01
-7.04166770e-01 -7.15780854e-01 1.70821901e-02 2.54376024e-01
6.68174565e-01 -7.43869424e-01 -5.91196656e-01 3.44445139e-01
-2.89224684e-01 -6.45159781e-01 2.31384322e-01 6.91952348e-01
-6.66613519e-01 4.81062889e-01 -2.93897808e-01 -2.56511450e-01
-1.53986776e+00 4.14916396e-01 2.08019257e-01 -8.67408365e-02
-3.46746117e-01 8.13260794e-01 -3.77669036e-01 3.48300695e-01
2.32408687e-01 -2.35672250e-01 -2.76035935e-01 5.08427680e-01
5.22080064e-01 7.29977667e-01 4.99338508e-01 -7.82515347e-01
-3.16549212e-01 2.33672768e-01 1.87235042e-01 -4.50235128e-01
1.10506880e+00 4.49694470e-02 3.23747694e-01 3.12240303e-01
9.55808938e-01 4.57087040e-01 -4.91191983e-01 2.01422796e-01
3.08289140e-01 -3.42666805e-01 5.80470748e-02 -7.29798377e-01
-7.12439358e-01 9.09178495e-01 6.70122564e-01 5.75740278e-01
1.16147780e+00 -3.89158517e-01 6.01106346e-01 2.75254250e-01
4.05560613e-01 -1.36255026e+00 -1.39700428e-01 1.63078681e-01
8.03291678e-01 -7.69851863e-01 -5.21951132e-02 -3.04815829e-01
-3.47998112e-01 1.15660489e+00 -1.46889016e-01 -1.91371534e-02
3.96281302e-01 5.13913214e-01 9.11649391e-02 2.94166114e-02
-2.10188299e-01 -2.74050891e-01 2.01473936e-01 6.31176591e-01
2.86664695e-01 -1.12936057e-01 -6.01836741e-01 3.49079490e-01
-3.13409954e-01 -2.86640286e-01 4.27718014e-01 9.01805162e-01
-5.07204473e-01 -1.27494216e+00 -6.97746217e-01 3.31228286e-01
-1.03875470e+00 -9.46933851e-02 -6.84147000e-01 8.28639150e-01
2.27069169e-01 1.09909201e+00 -4.63488340e-01 -3.32391232e-01
1.89402670e-01 3.98536056e-01 4.45564121e-01 -1.98592827e-01
-8.52747858e-01 4.31066602e-01 2.63947099e-01 -3.30437750e-01
-3.99842769e-01 -7.76821494e-01 -8.19628000e-01 -4.11781222e-02
-5.79867542e-01 1.96424350e-01 9.85176623e-01 8.88231695e-01
-3.46505046e-02 4.98034865e-01 4.32013899e-01 -6.88964725e-01
-7.02393711e-01 -1.02891409e+00 -7.33194709e-01 5.68967797e-02
3.11879784e-01 -6.56746984e-01 -4.44686949e-01 -9.09203589e-02] | [14.809823989868164, 5.776573181152344] |
8ef8bee9-3066-4a72-99f6-b16eae03503f | self-supervised-augmentation-consistency-for | 2105.00097 | null | https://arxiv.org/abs/2105.00097v1 | https://arxiv.org/pdf/2105.00097v1.pdf | Self-supervised Augmentation Consistency for Adapting Semantic Segmentation | We propose an approach to domain adaptation for semantic segmentation that is both practical and highly accurate. In contrast to previous work, we abandon the use of computationally involved adversarial objectives, network ensembles and style transfer. Instead, we employ standard data augmentation techniques $-$ photometric noise, flipping and scaling $-$ and ensure consistency of the semantic predictions across these image transformations. We develop this principle in a lightweight self-supervised framework trained on co-evolving pseudo labels without the need for cumbersome extra training rounds. Simple in training from a practitioner's standpoint, our approach is remarkably effective. We achieve significant improvements of the state-of-the-art segmentation accuracy after adaptation, consistent both across different choices of the backbone architecture and adaptation scenarios. | ['Stefan Roth', 'Nikita Araslanov'] | 2021-04-30 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Araslanov_Self-Supervised_Augmentation_Consistency_for_Adapting_Semantic_Segmentation_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Araslanov_Self-Supervised_Augmentation_Consistency_for_Adapting_Semantic_Segmentation_CVPR_2021_paper.pdf | cvpr-2021-1 | ['synthetic-to-real-translation'] | ['computer-vision'] | [ 6.42476439e-01 3.75615507e-01 1.22814618e-01 -5.90917289e-01
-1.07454324e+00 -8.61945510e-01 6.21771872e-01 -4.14863318e-01
-6.59067154e-01 8.02877426e-01 -2.25909233e-01 -2.97320366e-01
1.09399192e-01 -5.13116896e-01 -8.83951068e-01 -5.74016392e-01
3.70940924e-01 5.50615907e-01 2.34143257e-01 -3.56461257e-01
1.52686551e-01 3.84879172e-01 -1.13098562e+00 -1.52136207e-01
1.03745711e+00 9.53010380e-01 -1.20331600e-01 7.59364367e-01
-1.36733606e-01 5.45123935e-01 -3.94073099e-01 -8.58102024e-01
5.48007488e-01 -6.67632639e-01 -1.17899525e+00 3.59624505e-01
5.95894039e-01 -7.86958635e-02 -2.22277194e-01 1.08649230e+00
4.58606869e-01 2.56234884e-01 6.42725587e-01 -1.06168771e+00
-8.62360120e-01 2.54857868e-01 -5.08255124e-01 -7.92460050e-03
-8.75073299e-02 3.89587015e-01 7.65953958e-01 -5.99466622e-01
9.13669169e-01 9.53764141e-01 1.11576033e+00 1.09081721e+00
-1.71710229e+00 -5.16338825e-01 3.15650165e-01 -1.36941239e-01
-1.14097190e+00 -5.58116794e-01 8.47624719e-01 -3.58231902e-01
6.43490732e-01 4.86743972e-02 4.61359978e-01 1.22518408e+00
-2.55072296e-01 5.71615279e-01 1.27641916e+00 -6.22219503e-01
3.06228310e-01 1.98944643e-01 -1.33255363e-01 7.28931844e-01
-8.56693387e-02 1.35100204e-02 -2.02482849e-01 1.53717175e-01
8.25972497e-01 -3.82019043e-01 5.58986664e-02 -7.28188574e-01
-8.86911094e-01 7.94455051e-01 4.84751791e-01 1.51520133e-01
-3.26168276e-02 3.84313524e-01 3.88373673e-01 3.69354546e-01
7.33134747e-01 5.77045918e-01 -7.37363458e-01 -7.11151063e-02
-9.79842126e-01 1.89587340e-01 5.64710140e-01 1.08922637e+00
7.74766088e-01 2.65605479e-01 1.47737220e-01 8.85549247e-01
7.99056515e-02 2.21113160e-01 4.02726054e-01 -1.56435049e+00
1.84024930e-01 2.76111364e-01 9.49897766e-02 -4.08970118e-01
-1.64892972e-01 -4.24805194e-01 -5.72834015e-01 5.13287246e-01
6.29472136e-01 -3.37854952e-01 -1.37767649e+00 1.97427213e+00
3.80446464e-01 2.85551816e-01 9.62080210e-02 4.18716967e-01
3.00087690e-01 1.33970484e-01 4.34033543e-01 2.31920049e-01
1.12391615e+00 -1.24509585e+00 -4.07222837e-01 -5.00430584e-01
4.84950006e-01 -8.04606259e-01 1.35051620e+00 1.88844681e-01
-1.35663259e+00 -5.69699228e-01 -1.05735373e+00 -1.86938614e-01
-4.50877577e-01 -1.67731792e-01 7.77225256e-01 8.33761632e-01
-1.28714836e+00 9.65137184e-01 -1.01780140e+00 -4.52091515e-01
8.52446198e-01 5.31552732e-01 -3.55531782e-01 -1.14998536e-03
-7.64940977e-01 1.07246959e+00 4.38237846e-01 -1.85547099e-01
-6.98222518e-01 -8.73472929e-01 -8.05852592e-01 -4.18808430e-01
2.87396342e-01 -1.05628788e+00 1.54401541e+00 -1.64017069e+00
-1.88297391e+00 1.32114887e+00 -1.18877232e-01 -5.74610651e-01
8.82750988e-01 -2.01906815e-01 -1.76611468e-01 3.76041862e-03
1.13771539e-02 1.10511100e+00 9.74857628e-01 -1.37260973e+00
-4.59116191e-01 -1.04676753e-01 -4.95838225e-02 4.00436401e-01
-2.14531317e-01 2.50191037e-02 -5.50104439e-01 -8.13595951e-01
-2.11583227e-02 -1.18983042e+00 -5.60273409e-01 3.45397770e-01
-2.08318993e-01 1.33792162e-01 8.58614087e-01 -6.89488769e-01
5.82884848e-01 -2.15838766e+00 2.09017217e-01 1.29450604e-01
-3.02765425e-02 4.29322243e-01 -1.47204399e-01 -9.24599692e-02
-1.19598441e-01 2.15877831e-01 -9.72518504e-01 -6.94236159e-01
1.16539881e-01 4.34174478e-01 -2.48558998e-01 3.16260725e-01
5.26457250e-01 1.13884974e+00 -8.53052557e-01 -4.79507744e-01
3.20986450e-01 5.00009954e-01 -7.15586424e-01 3.49656552e-01
-3.92402589e-01 9.65327919e-01 -3.64259005e-01 4.02285457e-01
5.54435372e-01 -3.02894980e-01 1.24062467e-02 1.06433615e-01
2.05862671e-01 1.98631793e-01 -8.92726123e-01 2.42735362e+00
-4.80402380e-01 4.25082445e-01 1.65723577e-01 -1.22951317e+00
7.60078549e-01 1.31247640e-01 3.13120753e-01 -4.85738963e-01
2.91727502e-02 4.44332510e-01 -1.69025466e-01 -2.64038563e-01
2.23148853e-01 -4.29898620e-01 -5.63210547e-02 3.58034074e-01
4.68632758e-01 -5.38439512e-01 1.09142782e-02 -6.67172894e-02
9.24917340e-01 9.11498010e-01 4.83498760e-02 -3.48067641e-01
3.42800796e-01 1.91723093e-01 5.34984112e-01 4.75620717e-01
-3.46707553e-01 9.74937439e-01 2.68485844e-01 -4.07816589e-01
-1.36518955e+00 -1.18479121e+00 -1.63326323e-01 1.27624750e+00
2.87030060e-02 1.81191891e-01 -1.28321993e+00 -1.05434549e+00
-7.94963092e-02 7.22499490e-01 -7.37257719e-01 -1.67528406e-01
-7.48626888e-01 -8.21240783e-01 8.29163909e-01 7.08497822e-01
6.90083385e-01 -1.14550304e+00 -2.58970231e-01 1.24159284e-01
-4.17106859e-02 -1.30636239e+00 -4.92805690e-01 2.38317207e-01
-1.07299280e+00 -8.20288062e-01 -8.30889106e-01 -9.66609836e-01
8.64087164e-01 -2.45923966e-01 1.39354527e+00 -8.88349414e-02
-3.16908777e-01 4.15985048e-01 -8.15851986e-02 -2.90422738e-01
-5.96739352e-01 3.34951520e-01 -2.77705193e-01 -1.86459541e-01
1.43310115e-01 -8.47326398e-01 -8.10844183e-01 6.32388666e-02
-7.97216177e-01 -6.14505447e-02 4.96955365e-01 9.07088637e-01
5.93601704e-01 -4.10400480e-01 5.15437484e-01 -1.25376785e+00
3.38590115e-01 -1.85927823e-01 -4.59785253e-01 1.67080984e-01
-8.70860875e-01 1.59689084e-01 7.19834149e-01 -2.93977231e-01
-1.26825130e+00 2.17275649e-01 -4.18422490e-01 -5.00365257e-01
-5.25919437e-01 -1.88034907e-01 2.53907982e-02 -5.71409941e-01
8.91601264e-01 9.47336573e-03 5.21462671e-02 -3.92041564e-01
7.91018307e-01 3.65505874e-01 8.54897738e-01 -7.67175794e-01
1.02524209e+00 6.44776583e-01 -8.63292441e-02 -5.09483814e-01
-7.70742416e-01 -7.47670904e-02 -9.59760547e-01 1.32399574e-01
1.02602232e+00 -7.77713954e-01 -2.88573623e-01 5.25014102e-01
-9.61247087e-01 -6.82622910e-01 -7.24049926e-01 -4.03832868e-02
-9.37149048e-01 4.45357531e-01 -6.37712061e-01 -4.32229400e-01
-3.75355035e-01 -1.11416042e+00 9.59218562e-01 1.64167538e-01
-3.20994467e-01 -1.37605929e+00 6.35354444e-02 5.12277901e-01
5.30413330e-01 5.34482062e-01 7.05670476e-01 -5.85319102e-01
-4.64734375e-01 -3.76773183e-03 -2.64924288e-01 6.10629022e-01
1.60469919e-01 -7.29636624e-02 -1.13933921e+00 -2.53642559e-01
-1.99064597e-01 -5.79767644e-01 7.31671929e-01 3.27047944e-01
1.38659954e+00 -4.90481369e-02 -1.40100226e-01 1.06853461e+00
1.34204805e+00 5.85571826e-02 6.39849544e-01 4.82284486e-01
9.11065936e-01 6.72723711e-01 1.84966609e-01 -1.54582798e-01
4.30537224e-01 5.80711782e-01 1.36976331e-01 -5.00631511e-01
-4.31862533e-01 -2.66588956e-01 8.85316730e-02 3.52608442e-01
-1.91395536e-01 2.18400806e-02 -8.80288184e-01 6.98439419e-01
-1.73979270e+00 -6.32259846e-01 3.85905206e-01 2.07092261e+00
1.07393992e+00 1.23040877e-01 1.15186229e-01 -1.42979518e-01
5.89459658e-01 8.50866958e-02 -8.64705324e-01 -5.32096982e-01
-3.78292166e-02 8.31393421e-01 7.68541753e-01 5.46945632e-01
-1.19527555e+00 1.42129743e+00 7.45200062e+00 7.33767211e-01
-9.94559824e-01 3.21272701e-01 9.07855392e-01 2.89986469e-03
-3.90232176e-01 1.75622895e-01 -3.68122190e-01 4.04388368e-01
8.49862933e-01 2.84367770e-01 6.97164953e-01 7.82071948e-01
-2.92851388e-01 2.01365590e-01 -1.05897820e+00 6.47590935e-01
5.67634590e-02 -1.20168161e+00 -1.80965424e-01 -3.00318897e-01
9.69880164e-01 6.16387688e-02 2.84182191e-01 1.88042760e-01
9.06854033e-01 -1.03343844e+00 6.44740582e-01 2.28007615e-01
9.55109835e-01 -5.86379826e-01 2.15648204e-01 -2.30837911e-02
-6.91525936e-01 1.33786872e-01 -3.02112736e-02 1.69344023e-01
2.27940708e-01 1.89277530e-01 -4.50057596e-01 3.06331635e-01
7.47615457e-01 5.19388616e-01 -4.88678187e-01 4.45432067e-01
-3.53784978e-01 6.16930008e-01 -2.96379179e-01 5.39047122e-01
2.95250177e-01 -2.82212913e-01 2.94079006e-01 1.22762525e+00
1.10451601e-01 -1.07328214e-01 7.00088516e-02 8.23798954e-01
-3.26727837e-01 -1.05222650e-01 -6.18438482e-01 1.82789043e-01
2.80952692e-01 1.12075043e+00 -8.64064693e-01 -4.77478623e-01
-2.97763050e-01 1.43966687e+00 5.37919402e-01 5.27390122e-01
-7.44049191e-01 -3.55847716e-01 7.48978019e-01 -1.19155981e-01
5.82763493e-01 -2.14259043e-01 -7.79765904e-01 -1.17005336e+00
1.99005231e-01 -7.90877700e-01 2.61862904e-01 -6.08141840e-01
-1.43154883e+00 4.70378786e-01 -1.25948682e-01 -9.12467062e-01
-3.89253378e-01 -6.38898432e-01 -5.80838978e-01 8.94830585e-01
-1.67252767e+00 -1.41597271e+00 -1.17576897e-01 6.07891381e-01
5.19186735e-01 2.66626514e-02 9.60745633e-01 1.39216647e-01
-5.97213030e-01 9.17697191e-01 1.19687334e-01 1.18132688e-01
8.31841946e-01 -1.48214698e+00 9.97492969e-01 1.02895558e+00
1.32705122e-01 4.17826951e-01 7.11993039e-01 -3.39154959e-01
-7.95539916e-01 -1.05350411e+00 5.59445143e-01 -7.43275106e-01
7.16985285e-01 -4.20769185e-01 -8.23275864e-01 1.02256477e+00
4.68040109e-01 2.36875087e-01 6.33821905e-01 5.95130846e-02
-4.65946496e-01 3.30006294e-02 -1.39750564e+00 7.54311740e-01
1.39252961e+00 -5.68609238e-01 -4.75260288e-01 3.07258993e-01
7.49691129e-01 -6.10965252e-01 -9.52252448e-01 4.17925119e-01
3.95459384e-01 -8.01140785e-01 1.17564738e+00 -8.19634259e-01
4.00433093e-01 -1.28989026e-01 -7.78879598e-02 -1.34374332e+00
-2.73087144e-01 -8.66707265e-01 2.10970327e-01 1.24801755e+00
5.01097023e-01 -7.82129467e-01 1.05046320e+00 1.02715695e+00
-3.63684297e-01 -5.10920107e-01 -9.67987239e-01 -6.08357251e-01
5.18462658e-01 -2.95185804e-01 4.17118460e-01 1.12771761e+00
-4.29883242e-01 1.19888783e-01 -1.95676297e-01 -2.86194962e-02
7.29203880e-01 -1.04886346e-01 7.90389240e-01 -1.04026389e+00
-4.53688055e-01 -6.13534033e-01 -2.44957328e-01 -9.02762949e-01
3.97959948e-01 -8.51320267e-01 1.45580515e-01 -1.21228373e+00
-9.63677764e-02 -6.52802646e-01 -3.50919306e-01 6.41925514e-01
-2.86783159e-01 7.61778057e-01 1.60628140e-01 1.75200582e-01
-5.41212559e-01 4.21414226e-01 1.19838428e+00 4.48266938e-02
-8.86526257e-02 -1.96056768e-01 -9.63541150e-01 9.28307772e-01
8.92177582e-01 -4.14776891e-01 -4.83418912e-01 -7.92050660e-01
-2.16097578e-01 -4.69814509e-01 4.43647832e-01 -8.59786570e-01
-1.22932673e-01 -3.97670567e-02 3.55584919e-01 3.13814074e-01
3.12929749e-01 -7.17020512e-01 9.23567414e-02 2.79180825e-01
-3.93729717e-01 -5.50724193e-03 4.85429555e-01 3.69792670e-01
2.49040760e-02 -2.35438973e-01 1.12314510e+00 -2.63575047e-01
-8.04132104e-01 3.43506634e-01 1.03977345e-01 3.56133044e-01
1.00361991e+00 -3.66561204e-01 -7.10077286e-02 -4.33118604e-02
-9.18293595e-01 1.91603564e-02 9.34479535e-01 2.15657383e-01
7.53383934e-02 -1.15310967e+00 -4.92249042e-01 1.68148801e-01
1.63662974e-02 3.11596155e-01 2.16549575e-01 5.09898663e-01
-7.12769389e-01 -4.86257672e-02 -2.81893641e-01 -5.32515168e-01
-7.97127306e-01 4.10708487e-01 5.56375027e-01 -2.88745731e-01
-5.92925251e-01 1.06658208e+00 1.02173932e-01 -8.21887255e-01
2.46729776e-02 7.58313909e-02 3.06632280e-01 -3.37431103e-01
-7.73580819e-02 1.56489938e-01 4.32576844e-03 -4.33602780e-01
-1.81833625e-01 8.93733561e-01 -1.60322115e-01 -3.98241907e-01
1.19494009e+00 -1.58480838e-01 2.13277519e-01 2.58316427e-01
1.02865505e+00 -3.18116903e-01 -2.01832223e+00 -2.15696692e-01
-1.62874773e-01 -2.85770416e-01 -1.17575623e-01 -9.23948586e-01
-1.12599087e+00 8.04704666e-01 6.56267583e-01 -6.92495927e-02
1.23469102e+00 -6.61586225e-02 9.59807217e-01 1.90428078e-01
1.85517356e-01 -1.30484569e+00 -1.27932101e-01 2.96933889e-01
5.03777444e-01 -1.37114453e+00 -1.28803059e-01 -3.86302710e-01
-6.56888783e-01 6.31803036e-01 5.55219829e-01 -5.11866510e-01
4.60368454e-01 1.70355111e-01 6.00574791e-01 8.67122859e-02
-3.10450852e-01 -2.34085754e-01 2.56935984e-01 9.46909726e-01
4.66777354e-01 -1.44864589e-01 -1.70335263e-01 7.29383081e-02
-2.43916944e-01 -3.95194963e-02 2.52256785e-02 9.36885059e-01
9.44927055e-03 -1.57080376e+00 -4.45025899e-02 1.47853047e-01
-6.38007700e-01 -7.90392607e-02 -3.19789678e-01 8.87654662e-01
2.01470688e-01 5.29473245e-01 1.04347236e-01 -2.01193854e-01
3.63682151e-01 4.84041929e-01 7.06208110e-01 -5.04256368e-01
-5.18612504e-01 -7.01249689e-02 -4.33945358e-02 -4.55615491e-01
-6.95237219e-01 -8.10351193e-01 -1.17871404e+00 -2.31524587e-01
-7.19926953e-02 -2.18201190e-01 6.43332243e-01 1.08897197e+00
5.06547451e-01 5.90147138e-01 5.22647917e-01 -9.25500572e-01
-5.57702839e-01 -7.84946263e-01 -1.06936015e-01 9.15324986e-01
1.75863713e-01 -5.19978106e-01 -1.43932790e-01 5.04062355e-01] | [9.906726837158203, 1.0942490100860596] |
00bd5404-0517-4558-8713-7a1eafc69216 | train-smarter-not-harder-learning-deep | 2211.15717 | null | https://arxiv.org/abs/2211.15717v3 | https://arxiv.org/pdf/2211.15717v3.pdf | Learning deep abdominal CT registration through adaptive loss weighting and synthetic data generation | Purpose: This study aims to explore training strategies to improve convolutional neural network-based image-to-image deformable registration for abdominal imaging. Methods: Different training strategies, loss functions, and transfer learning schemes were considered. Furthermore, an augmentation layer which generates artificial training image pairs on-the-fly was proposed, in addition to a loss layer that enables dynamic loss weighting. Results: Guiding registration using segmentations in the training step proved beneficial for deep-learning-based image registration. Finetuning the pretrained model from the brain MRI dataset to the abdominal CT dataset further improved performance on the latter application, removing the need for a large dataset to yield satisfactory performance. Dynamic loss weighting also marginally improved performance, all without impacting inference runtime. Conclusion: Using simple concepts, we improved the performance of a commonly used deep image registration architecture, VoxelMorph. In future work, our framework, DDMR, should be validated on different datasets to further assess its value. | ['Frank Lindseth', 'Ole-Jakob Elle', 'Thomas Langø', 'Shanmugapriya Survarachakan', 'David Bouget', 'Egidijus Pelanis', 'André Pedersen', 'Javier Pérez de Frutos'] | 2022-11-28 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 4.61143851e-01 4.29454952e-01 -5.22336289e-02 -9.01973069e-01
-8.45144629e-01 -2.65799224e-01 4.72754180e-01 2.57421374e-01
-9.87453997e-01 5.53938150e-01 1.25340139e-02 -2.60026157e-01
-8.34562257e-02 -8.38354051e-01 -7.93983638e-01 -5.58430672e-01
-3.84718060e-01 6.65620685e-01 3.89801770e-01 -1.41489774e-01
-1.14816017e-01 5.67594230e-01 -9.17750895e-01 4.13264811e-01
7.49718130e-01 9.60475981e-01 1.73321068e-01 4.28813130e-01
4.25424129e-02 5.90292037e-01 -4.09549683e-01 -3.91925842e-01
5.40480971e-01 -4.27853465e-01 -1.01581979e+00 -3.98292929e-01
7.07322836e-01 -4.35228914e-01 -1.41904727e-01 7.48230755e-01
9.80509579e-01 3.32056761e-01 3.84675920e-01 -6.82212949e-01
-3.23168367e-01 8.34889889e-01 -3.41629177e-01 3.04504514e-01
1.50533810e-01 2.50737518e-01 3.24509978e-01 -3.51155430e-01
6.42695904e-01 7.93389082e-01 1.26422632e+00 6.91166043e-01
-1.31919146e+00 -7.01227248e-01 -4.05368239e-01 -1.82864338e-01
-1.11694825e+00 -1.30048275e-01 6.02164805e-01 -3.48683625e-01
9.63818967e-01 3.54365230e-01 5.70876420e-01 8.93281996e-01
3.39893758e-01 3.53399664e-01 1.18760467e+00 -2.80078799e-01
-5.86113706e-02 -9.60881561e-02 -2.14342419e-02 9.09320712e-01
3.75663102e-01 4.00961071e-01 6.43085763e-02 -2.39318907e-01
8.95559013e-01 -3.90308231e-01 -1.91667035e-01 -5.20366669e-01
-1.05743933e+00 8.63116324e-01 8.93281698e-01 6.93095267e-01
-4.49459046e-01 2.26575807e-01 7.77440488e-01 2.25134194e-01
5.50674021e-01 7.17393935e-01 -4.51601118e-01 4.69715744e-02
-1.28150153e+00 2.15421058e-02 6.97698832e-01 5.33652246e-01
6.00353897e-01 -8.17707330e-02 -3.33240688e-01 7.02182949e-01
2.04361200e-01 -6.63430020e-02 7.61323571e-01 -9.72249806e-01
3.86359572e-01 5.60468316e-01 -5.17156720e-01 -8.61720324e-01
-1.04440641e+00 -5.18656194e-01 -8.99258912e-01 2.48753771e-01
5.96677721e-01 -3.36893171e-01 -1.07269955e+00 1.62181127e+00
3.51601779e-01 3.66372645e-01 -2.00557500e-01 1.11181211e+00
1.03385568e+00 -1.64889723e-01 3.51661325e-01 3.28022361e-01
1.31027186e+00 -8.51648390e-01 -3.51558357e-01 2.10331157e-01
1.02203929e+00 -7.83783734e-01 1.06563795e+00 1.41411647e-01
-1.43870020e+00 -4.66057658e-01 -9.69466150e-01 -5.85901514e-02
-1.44988805e-01 1.74616456e-01 9.14370000e-01 8.93957675e-01
-1.36914444e+00 1.00465381e+00 -1.29146230e+00 -3.58912617e-01
6.38472676e-01 1.03028429e+00 -4.27943379e-01 2.89792679e-02
-1.17266035e+00 1.08593798e+00 5.51918626e-01 3.40038654e-03
-6.63515270e-01 -1.15743089e+00 -1.08049381e+00 -1.60791799e-01
-4.40830961e-02 -9.67839897e-01 9.81960058e-01 -1.13937402e+00
-1.59321165e+00 1.29697537e+00 3.76564264e-01 -8.02774310e-01
8.29698145e-01 6.13116585e-02 -1.30579203e-01 1.59936801e-01
-8.11263099e-02 1.04120362e+00 4.23455030e-01 -9.94174898e-01
1.02560923e-01 -2.12683484e-01 2.85354648e-02 -1.30772719e-03
-9.83784795e-02 3.46005894e-02 -2.74492025e-01 -8.56055677e-01
-1.18145049e-01 -1.07423508e+00 -5.51123023e-01 8.12740717e-03
7.80957416e-02 1.66146070e-01 2.97168136e-01 -9.12596405e-01
6.96414292e-01 -1.85336292e+00 -2.36620307e-01 4.66728926e-01
2.03558244e-02 3.98654729e-01 -4.21016604e-01 -2.25797936e-01
-3.75049621e-01 -8.42036307e-02 -6.12766325e-01 -3.00719172e-01
-4.04945284e-01 2.03777179e-01 4.79578346e-01 7.27796137e-01
1.29551411e-01 1.26600301e+00 -8.16197455e-01 -6.22053564e-01
3.79969656e-01 7.78416157e-01 -9.18156385e-01 2.41033390e-01
1.92206278e-01 1.01689506e+00 2.07933374e-02 1.81966111e-01
8.21704268e-01 -1.49286479e-01 4.96654026e-02 -4.75247532e-01
1.64209202e-01 2.10250661e-01 -8.17328811e-01 2.38746095e+00
-7.84312725e-01 3.35450053e-01 1.79354981e-01 -1.13357306e+00
8.76947939e-01 3.19185108e-01 1.03344285e+00 -1.02692115e+00
4.49004740e-01 2.20434338e-01 3.86584938e-01 -5.43354034e-01
3.66174012e-01 -1.93667173e-01 1.96920544e-01 3.48457307e-01
2.99141884e-01 -1.35337859e-01 4.59147133e-02 -2.14871287e-01
9.89634335e-01 4.82594579e-01 -1.09592870e-01 -6.05836630e-01
6.74774647e-01 1.89801082e-01 2.13784203e-01 6.62622213e-01
-1.06145233e-01 9.63038981e-01 8.80097449e-02 -6.95206404e-01
-8.15371335e-01 -8.01702261e-01 -4.28864658e-01 1.01390707e+00
-7.07665980e-02 6.93191588e-02 -1.20732224e+00 -8.04662764e-01
-2.99782395e-01 4.39690471e-01 -8.11176240e-01 -2.01649949e-01
-1.02269495e+00 -1.16210437e+00 1.06275249e+00 5.60381830e-01
4.77055490e-01 -1.19352198e+00 -7.73638129e-01 3.13787162e-01
-1.21888362e-01 -9.56490993e-01 -5.45403659e-01 3.53484720e-01
-1.36766064e+00 -1.14939666e+00 -9.96693432e-01 -7.98698366e-01
9.58460033e-01 -4.86120731e-01 1.29501128e+00 2.51711637e-01
-6.02688551e-01 5.09699643e-01 2.92668995e-02 -8.16409960e-02
-5.87819219e-01 4.10588026e-01 -2.29199842e-01 -4.60627168e-01
-4.76555154e-02 -4.97632235e-01 -9.40505147e-01 2.19359875e-01
-1.02146637e+00 -2.58042932e-01 5.38666487e-01 8.48059952e-01
5.23134112e-01 -4.43829745e-01 1.29016697e-01 -1.00537312e+00
5.11506379e-01 -2.38259673e-01 -4.76087511e-01 6.64173141e-02
-6.45430028e-01 2.20387772e-01 4.15970325e-01 -4.83402342e-01
-9.41990614e-01 9.30017680e-02 -5.41216552e-01 -2.54428327e-01
-1.66074201e-01 4.94903594e-01 3.82716984e-01 -6.42822802e-01
8.81452382e-01 -5.78074530e-02 4.47042525e-01 -4.44889396e-01
1.54171035e-01 1.25722848e-02 5.78870177e-01 -8.05521965e-01
3.50200862e-01 4.73561555e-01 2.90837914e-01 -3.70929360e-01
-4.19170350e-01 -3.52812946e-01 -8.62833679e-01 9.38421115e-02
1.01755321e+00 -8.02764177e-01 -5.44739187e-01 3.28084886e-01
-1.01647484e+00 -9.21026468e-01 -5.51816285e-01 8.04120183e-01
-5.00942826e-01 2.68212467e-01 -7.70218909e-01 1.15863308e-01
-8.92782927e-01 -1.42073953e+00 9.45170224e-01 -4.38461080e-03
-3.49768221e-01 -1.48314214e+00 3.27711850e-01 2.83394426e-01
9.56920624e-01 5.68296909e-01 5.66947520e-01 -9.84165132e-01
-1.51581511e-01 -7.87698403e-02 -2.57648528e-01 4.20476288e-01
2.74033874e-01 -4.69212115e-01 -7.98899651e-01 -5.19611299e-01
2.03065239e-02 -3.45448464e-01 7.50144362e-01 6.07723713e-01
1.51028931e+00 -1.10020056e-01 -2.38168370e-02 1.03970599e+00
1.35357606e+00 -1.12240024e-01 8.79176617e-01 5.89346170e-01
6.98503017e-01 3.57274681e-01 2.12231234e-01 8.26822296e-02
3.33862811e-01 7.12808907e-01 3.50714982e-01 -7.89271712e-01
-6.08321965e-01 3.44795257e-01 -6.02803677e-02 5.81516445e-01
-3.79193515e-01 3.98875087e-01 -1.05873942e+00 2.63402373e-01
-1.23732960e+00 -5.57860136e-01 -1.09654844e-01 2.14692497e+00
1.05030668e+00 -5.87577820e-02 3.60082500e-02 -3.27306360e-01
3.79878104e-01 -1.95681170e-01 -2.04458505e-01 -4.04862851e-01
2.17265517e-01 8.36193442e-01 7.63188779e-01 4.56095099e-01
-1.20534563e+00 6.30243838e-01 6.15358019e+00 5.88609755e-01
-1.71863496e+00 4.43909019e-01 7.17378497e-01 1.02608569e-01
-1.01503521e-01 -2.97750443e-01 -3.26457828e-01 2.95055687e-01
1.03127408e+00 2.59396017e-01 3.44806463e-01 5.79589665e-01
1.27300933e-01 -3.59835587e-02 -1.00236440e+00 6.20040953e-01
1.95589796e-01 -1.23402190e+00 -1.49861857e-01 -9.14045274e-02
5.71806192e-01 4.93109316e-01 -1.42083168e-02 2.64169633e-01
1.74924880e-01 -1.12702668e+00 2.07175851e-01 4.60685819e-01
8.83108735e-01 -6.27521217e-01 1.06808782e+00 -1.34805351e-01
-8.64585221e-01 4.54692543e-01 -1.14913628e-01 2.46631756e-01
-3.73483263e-02 4.30219978e-01 -1.23569024e+00 7.88754582e-01
6.81069553e-01 4.15894091e-01 -8.81697178e-01 1.11047077e+00
2.06096277e-01 5.54816008e-01 -3.59245569e-01 5.49212039e-01
3.67743939e-01 -1.00445397e-01 2.72729874e-01 1.53365099e+00
1.80506900e-01 -6.36321530e-02 2.05666408e-01 8.05488169e-01
-1.02281250e-01 1.36709914e-01 -3.77843976e-01 6.46361470e-01
-1.95453838e-01 1.60016119e+00 -9.54462469e-01 -1.33776203e-01
-2.77970314e-01 9.22362745e-01 1.13148116e-01 -4.69918512e-02
-1.08826077e+00 -1.64061114e-01 3.06697816e-01 4.97971445e-01
4.59755547e-02 -4.29839827e-03 -4.57591116e-01 -8.41398239e-01
-2.78893381e-01 -7.00511932e-01 5.90698302e-01 -2.49802187e-01
-1.16689253e+00 8.15891147e-01 1.31175667e-01 -1.04693055e+00
-2.01179922e-01 -3.67726892e-01 -6.43039644e-01 8.81350636e-01
-1.65798318e+00 -1.24139154e+00 -5.10012984e-01 7.34359503e-01
2.43949443e-02 6.32163286e-02 9.81060624e-01 7.31639504e-01
-1.93290785e-01 1.02450597e+00 -2.77059227e-01 4.80064273e-01
8.91282320e-01 -1.18749464e+00 3.03019285e-01 4.47900772e-01
-1.56643391e-01 7.12659299e-01 2.51097888e-01 -5.06064296e-01
-8.16260874e-01 -1.29796588e+00 4.10492718e-01 -2.10732803e-01
4.29559916e-01 5.10610342e-02 -9.71354783e-01 6.03210032e-01
3.50559026e-01 4.47802722e-01 7.47010946e-01 -1.32546455e-01
1.05947256e-01 -3.96022908e-02 -1.66034102e+00 3.18422765e-01
7.86150634e-01 -1.38457477e-01 -4.58988130e-01 1.20032318e-01
5.19684553e-01 -1.12709439e+00 -1.50438893e+00 7.73618937e-01
4.79624718e-01 -6.06612444e-01 1.24971581e+00 -4.48746949e-01
3.45193982e-01 6.23153113e-02 2.29055420e-01 -1.17966282e+00
-2.16546223e-01 -3.93072397e-01 1.18616991e-01 9.78298545e-01
3.34155351e-01 -7.58468151e-01 9.39355314e-01 9.38493609e-01
-3.34136099e-01 -7.66232908e-01 -1.16912282e+00 -6.74474418e-01
4.15971130e-01 -2.99155086e-01 6.34626985e-01 1.08822417e+00
-4.04233009e-01 -2.51338363e-01 5.80254123e-02 1.30049922e-02
5.30435920e-01 -3.24155778e-01 5.04400253e-01 -1.00357902e+00
-2.30208337e-01 -4.78890985e-01 -4.76972878e-01 -3.66941750e-01
5.72947003e-02 -1.43224692e+00 -1.01364054e-01 -1.31785500e+00
8.09589401e-02 -9.57535803e-01 -5.49166560e-01 8.13547373e-01
-1.57117784e-01 7.14961529e-01 1.86633691e-01 -4.95802462e-02
-3.45344305e-01 1.56261742e-01 1.51773334e+00 -1.50284618e-01
-2.64359117e-01 -1.40207401e-02 -3.32516253e-01 6.26662195e-01
9.34679687e-01 -7.31910527e-01 -2.12373555e-01 -5.76432765e-01
-2.61913657e-01 -8.73956010e-02 5.31000197e-01 -1.19266820e+00
1.62943393e-01 4.41716969e-01 5.65410972e-01 -8.06565806e-02
2.47999951e-02 -9.91368771e-01 2.23169565e-01 9.29184198e-01
-4.27112848e-01 9.93169770e-02 6.49577856e-01 -2.06402451e-01
2.90961433e-02 -3.31946462e-01 9.72102106e-01 -1.67142153e-01
-2.42205128e-01 3.83702725e-01 1.75690223e-02 1.34224892e-01
9.04812872e-01 -1.99117690e-01 1.66426718e-01 5.17587215e-02
-9.66056287e-01 -3.42239141e-02 3.68963689e-01 3.27564955e-01
3.30407411e-01 -1.12202621e+00 -8.05928051e-01 2.73749918e-01
-3.15406084e-01 2.06039265e-01 1.65070534e-01 1.36815274e+00
-9.49695706e-01 3.19042861e-01 -4.80007321e-01 -8.31587076e-01
-1.21973395e+00 3.44001919e-01 8.50719750e-01 -7.81310320e-01
-7.37893224e-01 7.63763905e-01 -1.77913249e-01 -1.08794999e+00
1.81361511e-01 -7.53288865e-01 -2.16460750e-01 -1.80218041e-01
2.17090115e-01 3.32648903e-01 6.40488863e-01 -5.36134005e-01
-4.63897854e-01 4.16210741e-01 -7.16098920e-02 8.53596106e-02
1.45292246e+00 1.17274240e-01 -2.36722738e-01 -1.91706717e-01
1.36709189e+00 -2.62046993e-01 -1.02728653e+00 -1.28115162e-01
7.57559240e-02 -3.51557918e-02 2.34393135e-01 -1.11747599e+00
-1.64200282e+00 5.71143270e-01 1.13472974e+00 -3.29022914e-01
1.22141063e+00 -3.00876975e-01 9.36948717e-01 3.90358828e-02
3.54588628e-01 -7.16151953e-01 -1.27263069e-01 3.62430036e-01
7.14012980e-01 -1.41866732e+00 1.63864970e-01 -1.20320179e-01
-5.66593051e-01 1.15574813e+00 5.79534531e-01 -4.12835062e-01
6.18791699e-01 6.17790401e-01 1.67302638e-01 -3.43734473e-01
1.15987159e-01 8.72635022e-02 6.56516671e-01 5.49093068e-01
8.90354872e-01 -1.49180830e-01 -5.84790051e-01 4.69665438e-01
-1.97807267e-01 4.10499841e-01 6.61099404e-02 7.38221526e-01
3.36324841e-01 -1.44463003e+00 -3.07997435e-01 5.49308598e-01
-8.70399058e-01 -2.85939366e-01 5.49674593e-02 1.00418890e+00
3.90315235e-01 3.60328913e-01 1.92008361e-01 6.56951815e-02
3.48286599e-01 -2.14677975e-01 8.32841277e-01 -5.63846409e-01
-1.50681305e+00 -2.16169223e-01 -6.89995363e-02 -8.47415090e-01
-6.66020453e-01 -5.39349139e-01 -1.55186093e+00 -1.28593028e-01
-3.09379250e-01 -2.84777768e-02 8.05056274e-01 8.16895306e-01
3.57304394e-01 9.19666290e-01 2.94619590e-01 -1.06137121e+00
-4.42180157e-01 -1.02929831e+00 -6.81153908e-02 5.23569286e-01
4.75218100e-03 -3.45808983e-01 1.42958969e-01 7.57154226e-02] | [14.25632381439209, -2.4935178756713867] |
e80fe562-75d6-4c1a-ba4a-71d743f39f82 | combining-multiscale-features-for | 1606.04985 | null | http://arxiv.org/abs/1606.04985v1 | http://arxiv.org/pdf/1606.04985v1.pdf | Combining multiscale features for classification of hyperspectral images: a sequence based kernel approach | Nowadays, hyperspectral image classification widely copes with spatial
information to improve accuracy. One of the most popular way to integrate such
information is to extract hierarchical features from a multiscale segmentation.
In the classification context, the extracted features are commonly concatenated
into a long vector (also called stacked vector), on which is applied a
conventional vector-based machine learning technique (e.g. SVM with Gaussian
kernel). In this paper, we rather propose to use a sequence structured kernel:
the spectrum kernel. We show that the conventional stacked vector-based kernel
is actually a special case of this kernel. Experiments conducted on various
publicly available hyperspectral datasets illustrate the improvement of the
proposed kernel w.r.t. conventional ones using the same hierarchical spatial
features. | ['Sébastien Lefèvre', 'Yanwei Cui', 'Laetitia Chapel'] | 2016-06-15 | null | null | null | null | ['classification-of-hyperspectral-images'] | ['computer-vision'] | [ 7.06710398e-01 -4.16448683e-01 1.16862031e-02 -3.06362450e-01
-5.57156503e-01 -6.66335940e-01 3.98976803e-01 5.31145692e-01
-5.62180996e-01 8.28111172e-01 -2.40719274e-01 -2.11950839e-01
-6.31627619e-01 -8.21730077e-01 -3.23571652e-01 -1.16159725e+00
5.20987622e-02 -2.87847906e-01 2.33410165e-01 -4.47130948e-02
4.00430471e-01 7.18237400e-01 -1.77331913e+00 -1.09723322e-01
1.58587289e+00 1.19514692e+00 3.27724576e-01 3.80339116e-01
-3.80705118e-01 3.74220401e-01 -3.42828602e-01 -1.40114635e-01
1.26319468e-01 -3.09241623e-01 -6.06971741e-01 7.05557704e-01
1.22317486e-01 1.41925082e-01 1.61663935e-01 1.34324610e+00
1.16341040e-02 3.73319805e-01 8.95711720e-01 -8.46035004e-01
-2.80898571e-01 4.05938596e-01 -7.76665092e-01 -3.21518369e-02
-6.66894615e-02 -4.28331196e-01 8.63858044e-01 -9.84448075e-01
1.91220269e-01 5.90468466e-01 6.33762896e-01 -4.10683244e-01
-1.38489997e+00 -1.11917965e-01 -8.96140262e-02 4.34577137e-01
-1.46129930e+00 1.27128148e-02 9.31667686e-01 -7.27641702e-01
3.79072845e-01 5.25955021e-01 7.55308211e-01 5.97173989e-01
-5.26317209e-02 5.43926537e-01 1.66455042e+00 -4.64222193e-01
3.20939422e-01 2.85409600e-01 7.81152070e-01 5.12511492e-01
4.09499913e-01 -3.11005563e-01 3.03384271e-02 -4.35840040e-01
4.12526220e-01 1.60406068e-01 -5.55180550e-01 -6.87093019e-01
-9.66647327e-01 9.64019358e-01 5.27636290e-01 6.83777273e-01
-3.82616550e-01 -5.74700773e-01 2.74627388e-01 -8.35730582e-02
7.26790547e-01 1.93351194e-01 -3.54563385e-01 3.55646074e-01
-1.19072533e+00 2.08095685e-02 5.85281491e-01 2.44870424e-01
1.30451202e+00 1.29970178e-01 1.90653414e-01 1.08037102e+00
5.99882379e-02 2.49495700e-01 5.65814376e-01 -3.41458946e-01
9.70342830e-02 8.62053990e-01 1.23226419e-01 -8.77098858e-01
-5.99465370e-01 -6.41525984e-01 -9.54980910e-01 3.10797453e-01
4.05249834e-01 -1.03645995e-01 -6.74989522e-01 1.10974991e+00
2.86956221e-01 3.95341486e-01 2.07543358e-01 7.09758520e-01
4.72333908e-01 8.11745703e-01 4.40379530e-02 -2.83264726e-01
1.09801614e+00 -7.61836588e-01 -5.10352314e-01 1.77949265e-01
6.23069346e-01 -6.17175698e-01 7.62074947e-01 6.72900617e-01
-4.08482790e-01 -4.49255735e-01 -1.22445381e+00 4.90961730e-01
-7.97706008e-01 6.12734973e-01 6.45270050e-01 7.81309426e-01
-7.90724695e-01 8.32388759e-01 -4.49256331e-01 -4.97124314e-01
1.41320983e-02 3.23648810e-01 -6.96293592e-01 2.57153332e-01
-8.05888772e-01 7.09245265e-01 9.93042707e-01 7.69103095e-02
1.21585086e-01 -2.93047071e-01 -7.75130630e-01 1.15694568e-01
3.88311923e-01 -4.07186896e-02 8.43607605e-01 -1.03709280e+00
-1.59710002e+00 6.12739503e-01 -7.10745826e-02 -3.55177850e-01
7.11489618e-02 -3.21004204e-02 -2.27278873e-01 3.97615045e-01
-1.69241726e-01 -1.12876650e-02 1.27613425e+00 -1.37631750e+00
-5.78431904e-01 -6.51531398e-01 -8.10051188e-02 1.55066745e-02
-9.23882306e-01 -1.87367275e-01 2.66031772e-01 -5.33610046e-01
4.40496475e-01 -9.53919172e-01 -2.79870927e-01 -4.34451878e-01
-6.52546406e-01 -2.92322129e-01 9.91538167e-01 -7.50843287e-01
8.83316934e-01 -2.02501512e+00 4.53449368e-01 4.21724260e-01
2.86251344e-02 2.86702573e-01 1.52122185e-01 6.92834496e-01
-3.86887550e-01 -1.21157452e-01 -8.61939907e-01 -1.30438833e-02
-3.20883751e-01 -1.06246568e-01 -1.74500883e-01 7.52450705e-01
1.89580813e-01 3.73132974e-01 -6.95927560e-01 -4.01521921e-01
4.74856049e-01 4.47609305e-01 -1.61444783e-01 1.30130380e-01
1.06993495e-02 6.19317353e-01 -5.03753662e-01 3.36489797e-01
1.01609814e+00 -2.55492628e-01 -5.90349035e-03 -6.33745611e-01
-6.15910888e-01 -6.40777647e-01 -1.28996408e+00 1.50797200e+00
-3.26153845e-01 2.33917013e-01 -8.82405136e-03 -1.67160308e+00
1.04565024e+00 1.32338628e-01 6.97802901e-01 1.76150411e-01
1.18733957e-01 3.52148324e-01 -2.85356253e-01 -4.77249861e-01
5.46645701e-01 -1.25843927e-01 1.78003222e-01 -1.09974593e-01
2.20541194e-01 -6.41174838e-02 6.72444180e-02 -3.44101906e-01
6.15914822e-01 2.49169201e-01 9.75122392e-01 -4.89710480e-01
1.06658769e+00 1.58899099e-01 1.01723857e-01 4.29251879e-01
1.52576804e-01 4.17651355e-01 3.20121199e-01 -6.66507557e-02
-9.74239528e-01 -7.26981997e-01 -5.71184993e-01 8.85772824e-01
-4.56938073e-02 -2.00621262e-01 -7.45960474e-01 -6.42357349e-01
3.09904460e-02 3.03088456e-01 -4.85030413e-01 1.09183483e-01
-1.08948536e-01 -1.09279144e+00 4.27204370e-01 2.97141105e-01
7.41936207e-01 -4.86408412e-01 -7.23171651e-01 2.90239334e-01
1.29090115e-01 -1.02008486e+00 2.43119195e-01 5.05256414e-01
-1.06981874e+00 -8.51594210e-01 -9.86905336e-01 -4.91124034e-01
2.57335812e-01 5.69925725e-01 3.85706902e-01 -2.69182265e-01
-4.34228867e-01 1.85517445e-01 -9.30121005e-01 -9.07970816e-02
-6.36908710e-02 1.66525260e-01 -1.79923564e-01 7.38315642e-01
-1.36623651e-01 -6.69247806e-01 -4.14701253e-01 -1.17202103e-01
-1.29365444e+00 4.51250151e-02 7.02536047e-01 9.33571100e-01
3.78597260e-01 3.25737000e-01 5.09386718e-01 -7.13850439e-01
2.85072416e-01 -4.74289536e-01 -8.34692359e-01 3.59667182e-01
-3.91539931e-01 1.11916423e-01 8.77350032e-01 -2.51915395e-01
-1.10553145e+00 3.66609991e-01 -4.18422893e-02 -8.68592262e-02
-5.73440015e-01 9.54744279e-01 -1.56927124e-01 -6.58295095e-01
3.82878274e-01 6.82717860e-01 -1.74795032e-01 -6.81611240e-01
4.72620785e-01 9.24399853e-01 2.99311191e-01 -4.14450496e-01
9.20735955e-01 4.40057725e-01 2.05910489e-01 -1.51426661e+00
-5.82870305e-01 -7.48938024e-01 -8.21497381e-01 -2.19908938e-01
9.78341818e-01 -4.33795393e-01 -5.47811568e-01 5.33770919e-01
-9.03791308e-01 9.02236626e-02 -1.10222794e-01 7.76749551e-01
-5.02932608e-01 8.61819565e-01 -3.91769260e-01 -1.19022334e+00
-7.66935721e-02 -1.03061557e+00 1.03360379e+00 4.77263957e-01
4.82365578e-01 -9.86181498e-01 6.59128502e-02 2.42111355e-01
2.40041524e-01 3.66041005e-01 1.04103577e+00 -8.15867066e-01
-6.39388934e-02 -1.70518085e-01 -6.00281596e-01 6.75252557e-01
2.80719638e-01 2.12754607e-01 -9.43172336e-01 -1.29381478e-01
3.11612725e-01 -1.04011625e-01 1.12385476e+00 2.79889226e-01
1.38858414e+00 6.22317269e-02 -1.92807361e-01 8.02285910e-01
1.93888485e+00 2.01545954e-01 4.78838533e-01 3.50846797e-01
6.98867381e-01 7.30334640e-01 5.44128478e-01 6.31121576e-01
-7.46236369e-02 5.19497097e-01 3.88149500e-01 -1.73255771e-01
7.90314019e-01 3.74745637e-01 6.27948297e-03 7.36985087e-01
-3.33749563e-01 2.27249935e-02 -7.30136335e-01 6.41108900e-02
-1.90472269e+00 -8.41973007e-01 -2.71357089e-01 2.25382495e+00
6.41152382e-01 -2.69047707e-01 6.03635535e-02 6.98381722e-01
8.53814125e-01 3.26519758e-01 -2.37817138e-01 -1.17609575e-02
-4.99380767e-01 4.83129382e-01 6.75724864e-01 4.01484400e-01
-1.61065102e+00 6.80824220e-01 4.97890711e+00 1.08291566e+00
-1.36854136e+00 6.26524240e-02 2.51911581e-01 6.40190780e-01
1.59067050e-01 6.83481172e-02 -4.24285293e-01 1.95923522e-01
5.93517900e-01 -3.18336248e-01 2.18634114e-01 8.04977417e-01
7.52116963e-02 -4.05912608e-01 -4.27211195e-01 1.11138725e+00
1.41979530e-01 -1.08649993e+00 -1.08441925e-02 2.22854063e-01
5.89183509e-01 -4.69950706e-01 8.16514865e-02 3.49172838e-02
-4.19951648e-01 -8.34427357e-01 3.74912024e-01 5.56326985e-01
5.07419288e-01 -7.38259971e-01 7.40409672e-01 3.29151720e-01
-1.38735282e+00 -1.39807656e-01 -3.55581611e-01 7.23107830e-02
-3.46126929e-02 9.38376307e-01 -3.96757632e-01 1.27111948e+00
2.60134906e-01 9.49495137e-01 -6.72527075e-01 1.49456763e+00
1.89168498e-01 6.27961695e-01 -2.11867169e-01 9.63672623e-02
6.47598743e-01 -1.01330769e+00 6.77500963e-01 1.25167465e+00
7.36963630e-01 -3.07102837e-02 2.72308052e-01 4.85679120e-01
3.38659942e-01 8.14007461e-01 -7.48224437e-01 -8.25651959e-02
-1.74934819e-01 1.81881154e+00 -1.07879210e+00 -4.21027929e-01
-6.15109980e-01 1.29106939e+00 1.54755920e-01 2.65465707e-01
-6.47028565e-01 -4.77925628e-01 3.23127806e-01 -3.88655156e-01
5.11575580e-01 -3.45859617e-01 -1.17273770e-01 -1.30062497e+00
-1.84614241e-01 -4.35747713e-01 2.04909578e-01 -7.37643600e-01
-1.44726014e+00 7.09302902e-01 1.91087976e-01 -1.49994969e+00
1.62381709e-01 -9.40222323e-01 -4.14990693e-01 8.16166580e-01
-1.49571919e+00 -1.16440511e+00 -6.10464513e-01 7.96248078e-01
3.14746320e-01 1.49022847e-01 9.32056248e-01 -1.69518068e-02
-7.33350515e-01 -1.59084499e-01 5.70933044e-01 -1.95229411e-01
5.78303754e-01 -1.58060575e+00 -5.78394890e-01 6.60099745e-01
7.40626529e-02 4.27449197e-01 6.37367427e-01 -5.23471653e-01
-1.16813815e+00 -1.14239991e+00 2.87549227e-01 3.62348437e-01
9.70074594e-01 1.28921375e-01 -1.12516999e+00 3.53341371e-01
3.58863801e-01 -2.76891533e-02 1.12126410e+00 -2.00307578e-01
-2.92950362e-01 -2.68112272e-01 -1.04167378e+00 3.24165672e-01
4.35763270e-01 -3.80635351e-01 -5.55945933e-01 3.91412944e-01
4.30778444e-01 1.75736904e-01 -1.12222266e+00 6.14335835e-01
4.16485518e-01 -1.27279341e+00 8.58746946e-01 -3.63349497e-01
-3.92540805e-02 -4.13139582e-01 -3.58822793e-01 -1.57317662e+00
-2.59551495e-01 -9.25556570e-02 3.44502807e-01 9.10546303e-01
1.68768540e-01 -1.04278266e+00 5.15227437e-01 -1.32654980e-01
-4.19150777e-02 -7.69779205e-01 -8.70709836e-01 -9.39673424e-01
-9.69460607e-02 -1.50378823e-01 3.88489366e-01 1.10411990e+00
1.00151993e-01 1.20240442e-01 -2.12459773e-01 4.32581425e-01
9.72073615e-01 4.89170641e-01 2.48533875e-01 -1.72397411e+00
-2.64120877e-01 -4.84313369e-01 -7.71400750e-01 -4.74188954e-01
4.32345301e-01 -9.03944433e-01 -8.09724405e-02 -1.43177485e+00
1.86985433e-01 -3.78232121e-01 -3.85190368e-01 1.16873458e-01
-3.31147283e-01 4.68092501e-01 1.68133274e-01 -5.46346745e-03
-2.39845160e-02 6.47303581e-01 7.38193452e-01 -1.10050160e-02
-4.08300728e-01 1.00265235e-01 -3.04330230e-01 8.74086320e-01
1.11662507e+00 -1.08790062e-01 -2.16126949e-01 1.53838307e-01
-2.67440856e-01 5.10654822e-02 5.14938474e-01 -1.33267045e+00
2.55267411e-01 -1.70731097e-01 3.53273302e-01 -7.83947766e-01
5.61274052e-01 -9.21777964e-01 1.35315895e-01 1.83435977e-01
2.80043688e-02 -4.52934265e-01 7.97978044e-02 5.10785997e-01
-3.57690513e-01 -5.84931731e-01 7.62411654e-01 4.55104150e-02
-8.33140969e-01 1.13679998e-01 -4.76693660e-01 -7.03300655e-01
1.18717313e+00 -3.33863378e-01 -8.03004652e-02 5.64364567e-02
-1.00697637e+00 -4.28053707e-01 3.22360903e-01 -2.28302702e-01
5.32330036e-01 -9.49001670e-01 -5.86502254e-01 -4.14910801e-02
3.28973919e-01 -6.15041673e-01 3.40921819e-01 1.22405338e+00
-4.37830299e-01 4.45207715e-01 -2.81551719e-01 -7.83533871e-01
-1.33618450e+00 4.97809201e-01 3.06744963e-01 -2.30283171e-01
-5.34275472e-01 5.64152718e-01 4.83000167e-02 -8.05334747e-02
-3.18284452e-01 -3.38095695e-01 -6.64873362e-01 5.71230352e-01
4.71203029e-01 4.18978721e-01 1.82676196e-01 -8.67822528e-01
-1.98873654e-01 9.55255389e-01 5.43578625e-01 -2.63986558e-01
1.34883189e+00 8.79352540e-02 -5.75927258e-01 7.24379480e-01
1.28467655e+00 1.84331328e-01 -9.24899876e-01 -2.62789100e-01
1.79390937e-01 -3.68506283e-01 2.66886920e-01 -2.32723460e-01
-7.34858572e-01 7.55348384e-01 6.52424097e-01 8.03566635e-01
1.42005026e+00 -2.54711092e-01 2.94710487e-01 4.53144103e-01
4.93269295e-01 -5.54867268e-01 -4.67579931e-01 2.33414829e-01
6.83679223e-01 -1.09550202e+00 -2.33932994e-02 -9.04557168e-01
-4.60275769e-01 1.75232136e+00 2.16311127e-01 -2.24377409e-01
8.63202155e-01 -4.13884828e-03 -2.24627793e-01 2.52434254e-01
8.76105204e-02 -9.43540633e-01 2.80386567e-01 3.82445067e-01
2.91598111e-01 2.85620719e-01 -6.43080950e-01 5.61210811e-01
8.40541646e-02 -5.92544749e-02 3.93093854e-01 8.69166791e-01
-7.78305352e-01 -1.16773963e+00 -6.77384257e-01 7.00346470e-01
-2.16411039e-01 -4.06488664e-02 -1.69291466e-01 3.49913239e-01
2.71040618e-01 1.10991383e+00 -4.55322683e-01 -4.99041259e-01
2.05103114e-01 4.02511418e-01 5.55558145e-01 -5.29620707e-01
-4.25954491e-01 1.79229409e-01 -2.98711419e-01 -2.60768652e-01
-8.58773351e-01 -5.29638946e-01 -8.90820861e-01 1.41594872e-01
-5.85122585e-01 4.05480027e-01 9.02807593e-01 1.06087816e+00
-3.24108005e-01 2.61097640e-01 9.46253419e-01 -1.25832057e+00
-6.68970883e-01 -9.29843128e-01 -1.29023039e+00 1.79301530e-01
3.92593741e-01 -8.04922223e-01 -5.86279988e-01 4.13904451e-02] | [9.962021827697754, -1.8925178050994873] |
0d4485bc-2668-4b71-889c-ada7fe50a9c2 | an-empirical-study-of-topic-transition-in | 2111.14188 | null | https://arxiv.org/abs/2111.14188v3 | https://arxiv.org/pdf/2111.14188v3.pdf | An Empirical Study of Topic Transition in Dialogue | Transitioning between topics is a natural component of human-human dialog. Although topic transition has been studied in dialogue for decades, only a handful of corpora based studies have been performed to investigate the subtleties of topic transitions. Thus, this study annotates 215 conversations from the switchboard corpus and investigates how variables such as length, number of topic transitions, topic transitions share by participants and turns/topic are related. This work presents an empirical study on topic transition in switchboard corpus followed by modelling topic transition with a precision of 83% for in-domain(id) test set and 82% on 10 out-of-domain}(ood) test set. It is envisioned that this work will help in emulating human-human like topic transition in open-domain dialog systems. | ['Vincent Wade', 'Benjamin R. Cowan', 'Christian Saam', 'Emer Gilmartin', 'Brendan Spillane', 'Mayank Soni'] | 2021-11-28 | null | https://aclanthology.org/2022.codi-1.12 | https://aclanthology.org/2022.codi-1.12.pdf | coling-codi-crac-2022-10 | ['open-domain-dialog'] | ['natural-language-processing'] | [-7.76322857e-02 8.47717464e-01 -1.63922027e-01 -6.09834909e-01
-4.82332498e-01 -9.21527624e-01 1.26155961e+00 3.94284278e-01
-8.01358670e-02 1.01904547e+00 5.53729475e-01 -4.62976635e-01
7.33143687e-02 -6.22824550e-01 9.71719399e-02 -1.35041818e-01
-2.47127280e-01 1.32034719e+00 6.72801077e-01 -4.94173408e-01
2.14138761e-01 -1.68057606e-01 -9.47365165e-01 3.82528365e-01
8.33602309e-01 4.73337680e-01 2.10250407e-01 8.65028739e-01
-7.52356231e-01 5.93806744e-01 -1.18790388e+00 -3.14952582e-01
1.47805493e-02 -6.15822732e-01 -1.67993200e+00 3.44459474e-01
2.83021778e-01 -1.15482286e-01 -1.73836365e-01 5.31398714e-01
3.34229887e-01 4.93169904e-01 8.57068956e-01 -1.58038139e+00
-5.27345650e-02 9.23923075e-01 -5.27279712e-02 5.47679126e-01
9.53584194e-01 1.40848711e-01 1.03571582e+00 -2.02502429e-01
1.11803865e+00 1.64783096e+00 4.63699490e-01 4.49922979e-01
-1.18397486e+00 -6.10994160e-01 1.31867781e-01 -3.46167654e-01
-6.75567508e-01 -1.87689096e-01 5.81570268e-01 -6.71330750e-01
1.06069171e+00 2.01887667e-01 5.71845353e-01 1.24184597e+00
2.07998291e-01 6.62229538e-01 1.50493574e+00 -4.10531461e-01
3.63501534e-02 1.00337422e+00 8.01991045e-01 4.02771942e-02
-3.51846337e-01 -3.71134698e-01 -6.40490592e-01 -6.30519331e-01
4.54122037e-01 -5.05537450e-01 -2.03034863e-01 1.68450639e-01
-1.13290799e+00 1.20783401e+00 -1.63280413e-01 6.36768162e-01
-2.43856534e-01 -7.28846848e-01 5.13885617e-01 7.74110079e-01
8.22067380e-01 6.97418809e-01 -6.06263041e-01 -7.91433215e-01
-7.83677161e-01 8.36477518e-01 1.93348765e+00 1.13066626e+00
5.77576101e-01 -5.74904859e-01 -3.94578457e-01 9.62372005e-01
1.53592736e-01 1.89390272e-01 5.92492878e-01 -7.02412724e-01
6.43584371e-01 7.26869941e-01 3.01663607e-01 -6.81765139e-01
-5.81186235e-01 5.32153308e-01 -2.81693727e-01 -2.35252991e-01
9.36820745e-01 -7.05981374e-01 -3.67130190e-01 1.46795511e+00
5.07246196e-01 -5.19613445e-01 1.69734359e-01 5.17304897e-01
1.13338554e+00 9.07028079e-01 2.63985097e-01 -3.11780006e-01
1.91630983e+00 -6.53742135e-01 -1.08049655e+00 -6.53395876e-02
4.96871144e-01 -1.09742343e+00 1.15772247e+00 2.39059597e-01
-1.11155975e+00 -4.34237450e-01 -4.36143309e-01 -1.35735512e-01
-4.34346825e-01 -4.54591453e-01 5.98740518e-01 8.52417648e-01
-9.47650492e-01 1.99432701e-01 -4.99066025e-01 -1.19177759e+00
-2.88237721e-01 3.71040016e-01 -1.60677791e-01 5.75299025e-01
-1.59392893e+00 1.17190373e+00 2.45594859e-01 -6.95190668e-01
-6.76601708e-01 -7.50763059e-01 -5.27231097e-01 -1.56766221e-01
4.78388816e-01 -3.69025558e-01 1.85579777e+00 -4.50047940e-01
-1.90391397e+00 1.02561557e+00 -1.84409633e-01 -6.54518366e-01
4.78629559e-01 -2.93162495e-01 -3.35490018e-01 7.72015601e-02
8.77764523e-02 9.54139769e-01 3.81581753e-01 -1.03938639e+00
-9.06272054e-01 -1.38876736e-01 3.76534373e-01 5.38043082e-01
-6.35736957e-02 3.36831510e-01 1.30011723e-03 -2.84103304e-01
-2.07535222e-01 -1.19312513e+00 1.88462988e-01 -5.79709947e-01
-6.40003383e-01 -9.79855359e-01 1.16773832e+00 -6.33580089e-01
1.18128097e+00 -1.74115956e+00 -1.79005668e-01 -2.51547426e-01
2.16620281e-01 -3.95920500e-02 4.45485890e-01 1.12825525e+00
2.93108970e-01 2.07303062e-01 -5.74670769e-02 -3.91007185e-01
3.24278086e-01 -1.68299332e-01 -7.53689587e-01 1.23677686e-01
-8.37083459e-02 4.03716564e-01 -7.74654508e-01 -6.15467846e-01
1.59769192e-01 2.59598792e-02 -4.13013339e-01 6.36344314e-01
-6.00703001e-01 5.62213898e-01 -5.23670077e-01 1.44519553e-01
3.66582453e-01 2.93469224e-02 2.04406127e-01 4.00200516e-01
-3.15462112e-01 1.01744747e+00 -7.68968821e-01 1.33031261e+00
-3.59876633e-01 1.20950675e+00 9.97357219e-02 -2.76062191e-01
1.17159152e+00 8.77531052e-01 1.25454634e-01 -1.90991849e-01
1.89244777e-01 -3.56590450e-01 3.81003350e-01 -3.80867749e-01
1.12831450e+00 -3.59460682e-01 -5.85790157e-01 9.19368744e-01
1.12140678e-01 -7.06763685e-01 2.24907890e-01 4.54840899e-01
7.69511461e-01 -3.94864082e-01 5.55783272e-01 -7.03933239e-01
2.89571881e-01 4.68933970e-01 1.36846051e-01 6.50436759e-01
-6.30757570e-01 2.61330366e-01 9.97550428e-01 -2.38181457e-01
-8.67852092e-01 -9.70216572e-01 -3.17248553e-01 1.32809865e+00
-2.28779539e-02 -3.13582510e-01 -9.76863027e-01 -5.72517514e-01
-2.76376724e-01 9.61476743e-01 -4.31700617e-01 2.96173960e-01
-4.75559264e-01 -3.56818676e-01 6.55723095e-01 -8.47216100e-02
6.99868560e-01 -1.25233579e+00 -5.12603343e-01 2.97031879e-01
-5.24775326e-01 -1.22021461e+00 -3.60461682e-01 7.59352371e-03
-8.79786491e-01 -9.37702417e-01 -5.38367689e-01 -7.16182888e-01
-1.08907856e-02 -1.21897168e-03 1.28324938e+00 -4.24926847e-01
1.22326374e-01 3.38820666e-01 -4.49638695e-01 -5.91087222e-01
-8.46128464e-01 4.92549062e-01 -1.21973656e-01 -6.96945310e-01
6.70841217e-01 -3.06524485e-01 -3.24973762e-01 6.16903007e-01
-5.67210972e-01 -5.89908473e-02 -9.05127302e-02 6.32722020e-01
-5.89631617e-01 -2.36111879e-01 5.98429739e-01 -1.38845336e+00
1.42835057e+00 -7.14243829e-01 -3.98359030e-01 -2.36593261e-02
-2.98291564e-01 -2.33662292e-01 2.92758167e-01 -5.76586723e-01
-1.55946147e+00 -5.15490234e-01 1.15734607e-01 4.65733588e-01
-7.63099551e-01 1.54579774e-01 1.21062878e-03 5.11112273e-01
6.09945953e-01 -9.96512771e-02 3.35414186e-02 -2.89060175e-01
1.57106712e-01 1.06578374e+00 4.64667790e-02 -6.13316357e-01
3.98465753e-01 2.59947836e-01 -7.38831460e-01 -1.50750828e+00
-5.11887729e-01 -8.08406591e-01 -4.60249245e-01 -3.23151559e-01
9.61868763e-01 -9.28323865e-01 -8.64564717e-01 3.32222909e-01
-1.30811226e+00 -8.73574317e-01 5.85781187e-02 3.53628546e-01
-4.32664961e-01 1.26868576e-01 -7.00249195e-01 -9.70244706e-01
-2.93169141e-01 -1.03429556e+00 9.46341217e-01 4.84247208e-01
-1.33742678e+00 -1.43404663e+00 4.87475812e-01 3.81763577e-01
3.01699042e-01 -1.48903593e-01 7.53197551e-01 -1.49608421e+00
-2.52582610e-01 9.18064564e-02 1.69045970e-01 -3.77760947e-01
3.02171588e-01 -1.73823237e-01 -1.06775951e+00 -1.05844438e-01
1.40470430e-01 -3.76542002e-01 8.60905647e-02 2.61482060e-01
1.32735353e-02 -3.29794794e-01 -5.90439856e-01 -4.12313372e-01
5.67061782e-01 4.75443810e-01 4.30064559e-01 2.47845143e-01
-2.09129050e-01 1.12604094e+00 9.85272408e-01 5.45156658e-01
6.78881645e-01 8.14721882e-01 -4.52325672e-01 3.01060736e-01
5.25567010e-02 4.37640734e-02 3.34873021e-01 5.81725299e-01
3.82725745e-01 -6.79020643e-01 -1.16970944e+00 7.00148284e-01
-1.39612591e+00 -8.28918636e-01 -2.21834689e-01 1.82465959e+00
1.17807806e+00 3.01770955e-01 6.36578798e-01 -2.67879337e-01
8.44970226e-01 3.15622956e-01 7.59348571e-02 -7.99887121e-01
2.75628865e-01 -4.26132325e-03 -9.65234712e-02 9.64461982e-01
-1.00877130e+00 1.38677299e+00 6.90408325e+00 4.24965948e-01
-8.41217995e-01 1.81961671e-01 7.00145185e-01 3.19462836e-01
-3.97350751e-02 2.83772528e-01 -1.27819574e+00 5.40345252e-01
1.08129025e+00 -4.00447816e-01 -1.72602654e-01 7.96531260e-01
3.59875947e-01 -6.19862318e-01 -1.21231055e+00 2.17749462e-01
-1.48194775e-01 -7.50981092e-01 -3.25628519e-01 1.93957284e-01
6.45911098e-01 -3.25606257e-01 -2.81180441e-01 6.92133367e-01
8.19779277e-01 -8.06463480e-01 1.86529994e-01 1.03987418e-01
2.52121955e-01 -2.75980264e-01 7.25610077e-01 5.39978623e-01
-8.15670013e-01 2.98247248e-01 -1.94085360e-01 -2.75177509e-01
4.50741231e-01 3.93133402e-01 -1.88642967e+00 1.76947072e-01
5.56984961e-01 -5.85994832e-02 -4.55439948e-02 7.80032694e-01
9.61875767e-02 1.04587281e+00 -4.24344182e-01 -3.58293295e-01
3.56962442e-01 -2.73267657e-01 8.81937087e-01 1.34107852e+00
-2.09930271e-01 4.30651784e-01 4.95022178e-01 6.31820798e-01
1.73735186e-01 6.80644810e-02 -7.15929985e-01 -1.85596853e-01
7.38860190e-01 1.06109643e+00 -1.07468677e+00 -5.09441316e-01
-9.33260098e-02 7.15875924e-01 -5.21404371e-02 1.30574971e-01
-6.27991557e-01 -2.80777663e-01 7.65329421e-01 1.91531226e-01
-1.38266847e-01 -2.04869673e-01 -2.49275610e-01 -4.64806795e-01
-4.14100558e-01 -9.23502028e-01 3.60225707e-01 -3.08120459e-01
-1.22833800e+00 9.73135829e-01 6.33821189e-01 -8.32172453e-01
-6.18345857e-01 1.03658466e-02 -1.08480954e+00 9.35088336e-01
-8.05565596e-01 -7.86615431e-01 -3.70047092e-01 1.63572297e-01
1.25602841e+00 -1.61939502e-01 8.89333427e-01 -1.86152086e-01
-2.13662997e-01 3.47599179e-01 -2.30022877e-01 5.94828464e-02
1.21100652e+00 -1.36595547e+00 6.62608147e-01 -2.25740541e-02
-3.40278804e-01 8.28891814e-01 1.35128987e+00 -9.45675969e-01
-5.99911213e-01 -5.81198275e-01 1.34387290e+00 -6.40929461e-01
7.50305533e-01 -5.54818571e-01 -9.05451059e-01 7.11282253e-01
1.13581550e+00 -9.77381945e-01 8.95567060e-01 5.86432695e-01
1.53216720e-01 5.41391015e-01 -1.02917194e+00 5.25665879e-01
5.25698900e-01 -5.95548093e-01 -1.35443294e+00 5.53841412e-01
8.77851963e-01 -6.08374953e-01 -1.07426560e+00 -1.18622564e-01
2.94536293e-01 -8.09496582e-01 6.15139484e-01 -5.24056733e-01
2.97969967e-01 3.27977657e-01 3.01944226e-01 -1.38776398e+00
3.24705005e-01 -1.14281821e+00 2.56435096e-01 1.71644711e+00
5.34535587e-01 -6.40718400e-01 8.34062874e-01 8.63385379e-01
-1.88232794e-01 -3.02697092e-01 -8.09325516e-01 -4.93081123e-01
3.82980049e-01 1.49353687e-02 2.83715159e-01 1.05342281e+00
8.79305422e-01 6.83227301e-01 -1.93498001e-01 -2.65789747e-01
-7.65414238e-02 7.64263421e-02 1.12842703e+00 -1.49408031e+00
4.95944470e-02 -2.40231633e-01 -7.36555457e-02 -1.19407272e+00
1.65041581e-01 -2.66296864e-01 2.47824535e-01 -1.37113559e+00
6.53736144e-02 -6.25823796e-01 8.73586357e-01 6.07811585e-02
-2.35808849e-01 -4.69090730e-01 2.29590356e-01 2.34045520e-01
-4.41591740e-01 4.28703547e-01 1.16189897e+00 1.91512838e-01
-6.79869771e-01 3.49124283e-01 -3.87481034e-01 6.66634321e-01
7.87845016e-01 -4.84448075e-01 -4.92083997e-01 3.67953628e-01
-3.75196099e-01 5.39069772e-01 -1.71651527e-01 -8.29722166e-01
3.22399646e-01 -1.32984474e-01 -3.54571551e-01 -8.77956808e-01
5.62193036e-01 -4.95275497e-01 -1.58527061e-01 2.59433866e-01
-7.64914632e-01 2.47907941e-03 2.74310231e-01 4.54885066e-01
-3.22526872e-01 -2.81880498e-01 4.27004129e-01 -2.37794861e-01
-4.68261808e-01 -1.43660009e-01 -9.54090178e-01 5.71375251e-01
1.14743626e+00 -2.68624902e-01 -3.95891756e-01 -7.55437434e-01
-7.31796324e-01 3.55621666e-01 3.75232399e-01 5.36290228e-01
-9.57095176e-02 -5.44379294e-01 -4.25554365e-01 -3.40549529e-01
-1.34358361e-01 -4.53110375e-02 1.82450220e-01 5.31384170e-01
-2.89258540e-01 1.04412913e+00 -2.07974598e-01 -6.96937442e-01
-1.52141750e+00 -1.99302763e-01 -1.23408794e-01 -6.07185066e-01
-6.23414278e-01 8.81910503e-01 4.25366789e-01 -6.88031673e-01
4.27712381e-01 -3.26833397e-01 -2.81178027e-01 5.23251474e-01
2.58320600e-01 3.71253878e-01 -4.07944113e-01 -4.42693770e-01
-4.37568426e-02 -1.03006259e-01 -5.24594963e-01 -6.43958032e-01
6.48754954e-01 -4.14439052e-01 1.13354035e-01 1.06600559e+00
9.07155454e-01 -1.71938300e-01 -1.00123644e+00 -2.65367806e-01
3.79122645e-01 -1.95654303e-01 -5.08690059e-01 -7.72700131e-01
8.17991719e-02 7.28325844e-01 2.63386309e-01 9.46236908e-01
3.35549831e-01 3.76120150e-01 8.89264882e-01 4.51636225e-01
1.15135007e-01 -1.19756317e+00 2.23695666e-01 1.01643336e+00
9.00698900e-01 -1.44244039e+00 -1.31300062e-01 -8.07623804e-01
-1.19746757e+00 8.61553013e-01 8.25560749e-01 8.84570852e-02
6.75588548e-01 1.37365133e-01 2.40992025e-01 -5.48605382e-01
-1.05204332e+00 1.28388450e-01 -3.20488843e-03 5.26917160e-01
9.98943448e-01 2.43479729e-01 -5.20738006e-01 4.28304732e-01
-8.89740169e-01 -3.28712285e-01 8.07605922e-01 1.01054084e+00
-7.23749340e-01 -1.11335456e+00 -5.05111635e-01 3.50651175e-01
-4.98018622e-01 -1.44126285e-02 -1.05783558e+00 1.30698025e+00
-4.31747139e-01 1.33046806e+00 4.51847702e-01 -9.72614884e-02
1.14616871e-01 7.00313747e-01 -2.99908277e-02 -1.01801443e+00
-1.33639145e+00 -7.87942261e-02 8.25927556e-01 4.22358960e-02
-2.07688749e-01 -9.30469751e-01 -1.07356024e+00 -4.56593782e-01
-5.64034343e-01 8.46492589e-01 4.45767939e-01 1.03784740e+00
3.20198350e-02 3.62602234e-01 3.64599705e-01 -3.07177275e-01
-4.11918223e-01 -1.90297842e+00 -5.49705446e-01 4.43675488e-01
-5.53589165e-02 -6.82478905e-01 -3.57644349e-01 1.43013239e-01] | [12.918129920959473, 8.029666900634766] |
342939f0-f74d-47d3-a40e-44c95d53d309 | griprank-bridging-the-gap-between-retrieval | 2305.18144 | null | https://arxiv.org/abs/2305.18144v1 | https://arxiv.org/pdf/2305.18144v1.pdf | GripRank: Bridging the Gap between Retrieval and Generation via the Generative Knowledge Improved Passage Ranking | Retrieval-enhanced text generation, which aims to leverage passages retrieved from a large passage corpus for delivering a proper answer given the input query, has shown remarkable progress on knowledge-intensive language tasks such as open-domain question answering and knowledge-enhanced dialogue generation. However, the retrieved passages are not ideal for guiding answer generation because of the discrepancy between retrieval and generation, i.e., the candidate passages are all treated equally during the retrieval procedure without considering their potential to generate the proper answers. This discrepancy makes a passage retriever deliver a sub-optimal collection of candidate passages to generate answers. In this paper, we propose the GeneRative Knowledge Improved Passage Ranking (GripRank) approach, addressing the above challenge by distilling knowledge from a generative passage estimator (GPE) to a passage ranker, where the GPE is a generative language model used to measure how likely the candidate passages can generate the proper answer. We realize the distillation procedure by teaching the passage ranker learning to rank the passages ordered by the GPE. Furthermore, we improve the distillation quality by devising a curriculum knowledge distillation mechanism, which allows the knowledge provided by the GPE can be progressively distilled to the ranker through an easy-to-hard curriculum, enabling the passage ranker to correctly recognize the provenance of the answer from many plausible candidates. We conduct extensive experiments on four datasets across three knowledge-intensive language tasks. Experimental results show advantages over the state-of-the-art methods for both passage ranking and answer generation on the KILT benchmark. | ['Zhoujun Li', 'Zhao Yan', 'Xinnian Liang', 'Jian Yang', 'Jiaheng Liu', 'Hongcheng Guo', 'Jiaqi Bai'] | 2023-05-29 | null | null | null | null | ['dialogue-generation', 'passage-ranking', 'open-domain-question-answering', 'answer-generation', 'dialogue-generation'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'speech'] | [ 1.68448627e-01 3.72523218e-01 1.17937941e-02 -2.80327126e-02
-1.60423136e+00 -7.90785551e-01 8.55853617e-01 2.48554006e-01
-3.77668589e-01 1.11952412e+00 5.89615226e-01 -3.45289558e-01
-2.20180556e-01 -1.22026467e+00 -8.32449436e-01 -3.36328894e-01
8.35288912e-02 1.07747018e+00 3.91497612e-01 -6.22925997e-01
4.01494384e-01 -3.10715020e-01 -1.56743312e+00 5.92817605e-01
1.50260687e+00 6.25175893e-01 3.78945380e-01 1.02176523e+00
-5.07828355e-01 8.88356209e-01 -8.54294419e-01 -5.70164680e-01
-5.09564839e-02 -9.54852045e-01 -1.36307037e+00 -6.29489362e-01
2.57182360e-01 -5.40362895e-01 -1.74960002e-01 5.71782887e-01
6.35090411e-01 4.55103695e-01 7.94311464e-01 -8.27076733e-01
-9.25839543e-01 1.02382505e+00 1.95478454e-01 2.98537970e-01
1.05928481e+00 -4.46740910e-02 1.46871829e+00 -8.45773399e-01
7.80598998e-01 1.39237428e+00 1.04448915e-01 7.63593674e-01
-9.58042145e-01 -1.97359815e-01 -3.97344604e-02 2.74972647e-01
-1.02751696e+00 -1.07880905e-01 5.03793895e-01 -1.98490575e-01
9.04149055e-01 4.40517753e-01 5.58972359e-01 8.96788180e-01
-1.70084432e-01 1.18487680e+00 7.09210813e-01 -5.50322294e-01
2.77951390e-01 2.68783748e-01 -3.97530161e-02 4.98719633e-01
-9.31633785e-02 2.32493188e-02 -6.70985520e-01 -4.72994566e-01
4.11410242e-01 -4.38542902e-01 -5.41966438e-01 -1.93069931e-02
-1.13708043e+00 1.11310017e+00 3.99331272e-01 3.35612223e-02
-3.54939044e-01 -8.08290616e-02 1.80526525e-01 4.58758742e-01
2.34006599e-01 1.13066936e+00 -5.03845632e-01 -2.11107880e-01
-8.96696627e-01 9.18921232e-01 1.30917966e+00 8.42784703e-01
6.78993762e-01 -4.64480758e-01 -1.15728056e+00 8.37027311e-01
3.42427760e-01 4.19805050e-01 4.42200780e-01 -9.65084791e-01
7.46920049e-01 6.76058650e-01 3.41856837e-01 -5.90734124e-01
3.72564793e-01 -3.27485710e-01 -3.31712991e-01 -3.98301750e-01
4.90695000e-01 -3.16378832e-01 -6.65345669e-01 1.76547217e+00
4.56212372e-01 2.11754814e-02 4.68554944e-01 8.61713588e-01
1.13320267e+00 1.14961183e+00 2.51389910e-02 1.42524680e-02
1.41921878e+00 -1.07873940e+00 -3.47926497e-01 -1.59334779e-01
6.39468968e-01 -7.52292752e-01 1.30731177e+00 1.18272744e-01
-1.18090236e+00 -5.07064104e-01 -7.83680856e-01 -2.29373977e-01
-8.06062669e-02 -2.37351358e-01 1.94876239e-01 1.22637101e-01
-8.86074483e-01 4.66371149e-01 -1.89751104e-01 -9.48739499e-02
7.20310956e-02 -6.66286573e-02 1.90839738e-01 -3.00774843e-01
-1.81854486e+00 7.83549488e-01 5.86117566e-01 -2.71424085e-01
-1.11603856e+00 -9.62068975e-01 -7.21489668e-01 2.32318178e-01
4.33827788e-01 -1.32477355e+00 1.65312243e+00 -5.23738623e-01
-1.65317297e+00 4.39744055e-01 -2.18909413e-01 -4.36626554e-01
5.03173649e-01 -3.76855433e-01 3.71965207e-02 4.90982831e-01
2.65966028e-01 8.43394101e-01 6.38374984e-01 -1.18173468e+00
-8.86933625e-01 3.04038167e-01 5.41629195e-01 6.65817142e-01
-9.58258193e-03 -2.72635162e-01 -3.61028135e-01 -3.16575259e-01
-2.31622979e-01 -6.70819759e-01 -1.91208586e-01 -6.42113566e-01
-5.13400733e-01 -7.43864000e-01 2.84146935e-01 -6.57098174e-01
1.35327029e+00 -1.44653094e+00 1.17628708e-01 1.23241715e-01
1.18792273e-01 3.20011944e-01 -4.23288196e-01 7.77835846e-01
4.67952847e-01 2.03595951e-01 -5.68152703e-02 -2.43437216e-02
3.15342426e-01 2.13256732e-01 -8.95405769e-01 -5.28664052e-01
2.76560605e-01 1.19105768e+00 -1.52331853e+00 -5.65387070e-01
-2.89400458e-01 2.07932994e-01 -7.68265188e-01 7.59120464e-01
-9.04425859e-01 5.76857746e-01 -9.70502436e-01 2.15207383e-01
9.04086530e-02 -2.83758312e-01 -2.48156846e-01 3.82275373e-01
3.43444496e-01 9.29556489e-01 -8.72200668e-01 1.60553432e+00
-6.43349707e-01 2.22605571e-01 -5.39563656e-01 -4.55373645e-01
9.82063830e-01 4.33303207e-01 -7.86598474e-02 -8.19127679e-01
-3.06777865e-01 3.17613751e-01 -5.81288375e-02 -6.32500052e-01
1.01499331e+00 -1.89418897e-01 -2.88300931e-01 8.15099061e-01
1.05420783e-01 -5.92152655e-01 6.96380436e-01 8.72343421e-01
1.11879992e+00 3.32310438e-01 -4.74087223e-02 1.06670402e-01
7.19476581e-01 2.11601019e-01 2.98778027e-01 1.26829672e+00
2.55215585e-01 7.11862504e-01 4.91747320e-01 -2.56407671e-02
-8.64841998e-01 -1.24506199e+00 3.91260654e-01 1.16284943e+00
-3.85091268e-02 -5.76967776e-01 -7.94368923e-01 -8.05330515e-01
-1.30830437e-01 1.05468404e+00 -3.80975217e-01 -3.33761007e-01
-6.87316000e-01 -2.84095168e-01 5.66768289e-01 4.13225800e-01
3.69791359e-01 -1.37214065e+00 -4.52859610e-01 4.41537738e-01
-9.13605034e-01 -7.66934752e-01 -5.11676788e-01 -2.13951796e-01
-7.04317987e-01 -8.89645815e-01 -8.26782882e-01 -7.43223310e-01
6.74033403e-01 -1.28875664e-02 1.73543453e+00 3.35596323e-01
4.45601605e-02 5.15070021e-01 -6.26639366e-01 -2.37333894e-01
-8.96500409e-01 3.09823722e-01 -4.30248082e-01 -4.18589175e-01
2.15661675e-01 -2.98211396e-01 -8.23897660e-01 1.16876811e-01
-9.73115206e-01 1.40710473e-01 4.08431381e-01 9.25025225e-01
6.25566900e-01 -2.00504020e-01 1.02999985e+00 -7.94016063e-01
1.30890167e+00 -5.90820909e-01 -3.87364417e-01 5.69694102e-01
-4.37571347e-01 4.96944457e-01 8.56917143e-01 -3.82079780e-01
-1.36540854e+00 -7.12562144e-01 -2.64297962e-01 1.29271507e-01
1.61960885e-01 7.82732069e-01 4.33292203e-02 5.03369689e-01
7.78234124e-01 4.89850074e-01 -4.42320138e-01 -2.04889402e-01
7.87047446e-01 4.40937221e-01 4.38279033e-01 -1.17958224e+00
8.39399517e-01 -1.12231247e-01 -4.79662061e-01 -3.60838383e-01
-1.18381953e+00 -5.34013569e-01 -9.80337113e-02 -8.77555087e-02
5.81606328e-01 -7.68154562e-01 -4.51949984e-01 -5.83721790e-03
-1.33911586e+00 -2.59447247e-01 -5.91863573e-01 2.77986079e-01
-6.90460324e-01 1.56537175e-01 -6.26321793e-01 -7.28937626e-01
-7.63770044e-01 -8.66695940e-01 1.02785802e+00 5.58543801e-01
-5.41652799e-01 -1.04674780e+00 5.45586407e-01 6.16568863e-01
4.36823070e-01 -4.00048904e-02 1.40580893e+00 -9.43275392e-01
-9.64714766e-01 -2.26390228e-01 7.26094693e-02 1.82183370e-01
-2.57783771e-01 -2.58067697e-01 -8.50366890e-01 -6.85314834e-02
-1.80149704e-01 -7.76186764e-01 9.81912315e-01 -1.14354663e-01
7.16560483e-01 -6.17583752e-01 -1.14267290e-01 5.38668968e-02
1.09492731e+00 -3.06606591e-02 7.63681829e-01 3.06894720e-01
3.17983717e-01 7.07623065e-01 6.39945209e-01 2.21788004e-01
6.75468504e-01 3.99894953e-01 2.03620307e-02 3.12399089e-01
-2.30149940e-01 -9.36776400e-01 3.36210757e-01 9.31627214e-01
1.47996873e-01 -5.54099262e-01 -7.30938077e-01 8.08805108e-01
-1.81162763e+00 -1.04241443e+00 7.25689158e-02 2.31620431e+00
1.43780756e+00 -1.39127141e-02 -1.18911132e-01 -1.98169097e-01
3.47312808e-01 1.57942064e-02 -3.05989832e-01 -3.24876875e-01
9.47244763e-02 4.88904715e-01 -3.58800650e-01 7.39835918e-01
-5.52167356e-01 9.47117507e-01 5.66271448e+00 9.75851655e-01
-4.57736075e-01 -1.58017963e-01 4.26416099e-01 7.90026113e-02
-8.43546093e-01 3.48952472e-01 -1.01381981e+00 2.41920277e-01
9.51716542e-01 -8.23663771e-01 3.21492881e-01 5.04530311e-01
1.11628585e-02 -2.88794905e-01 -1.27674150e+00 2.57229209e-01
1.04358599e-01 -1.23849523e+00 6.71759665e-01 -3.67333174e-01
9.56541240e-01 -2.91168511e-01 -6.52602464e-02 8.89408350e-01
5.96137285e-01 -9.18239117e-01 5.74396789e-01 7.85307527e-01
2.34285846e-01 -7.54939795e-01 6.19386613e-01 6.79661870e-01
-9.18178022e-01 2.94455420e-02 -4.28888410e-01 5.58288395e-03
3.05504084e-01 4.77631688e-01 -1.43046439e+00 8.18918407e-01
2.88085133e-01 9.42650139e-02 -4.25566226e-01 1.16393518e+00
-8.45085263e-01 8.03893149e-01 -1.56166166e-01 -4.82039034e-01
3.13033462e-01 -4.36527133e-02 7.32536197e-01 9.11930263e-01
4.09783810e-01 3.03809345e-01 2.74245799e-01 1.13404191e+00
-2.95407683e-01 1.84965953e-01 -3.03695619e-01 -7.89949968e-02
5.99256873e-01 1.04078376e+00 -3.01414818e-01 -5.14002979e-01
-6.89754076e-03 8.20821762e-01 3.60034138e-01 5.03194451e-01
-4.34639126e-01 -5.69287241e-01 2.50609577e-01 -8.18679929e-02
2.62612909e-01 2.30448052e-01 2.42304415e-01 -1.05869293e+00
1.40961513e-01 -1.06672573e+00 6.00518942e-01 -8.92594278e-01
-1.35310328e+00 6.20331645e-01 1.22599490e-01 -9.98893976e-01
-9.17606354e-01 -1.10190152e-03 -9.22424197e-01 1.26549470e+00
-1.94794679e+00 -7.12306976e-01 4.87149172e-02 3.39973986e-01
6.40231967e-01 9.65268686e-02 9.83676374e-01 -8.27976130e-03
-5.27015142e-02 4.69806165e-01 -1.24902807e-01 9.69066694e-02
6.69448078e-01 -1.50664270e+00 3.75953346e-01 7.50482023e-01
1.95038036e-01 8.26467633e-01 7.53546476e-01 -6.97435617e-01
-1.21082103e+00 -1.00488675e+00 1.39295542e+00 -7.44431198e-01
5.47532082e-01 -2.14939043e-01 -1.10324776e+00 1.65800959e-01
3.04391205e-01 -5.58929324e-01 7.35793591e-01 9.79253948e-02
-2.67626345e-01 1.88304737e-01 -8.28828454e-01 7.07273126e-01
6.58739209e-01 -5.72762668e-01 -1.18619156e+00 4.15488124e-01
1.06138253e+00 -4.92729336e-01 -6.29070878e-01 2.03382134e-01
2.77724057e-01 -6.08468533e-01 9.14288580e-01 -6.98293984e-01
9.30082083e-01 -3.87159228e-01 1.02234691e-01 -1.48692334e+00
2.90283095e-02 -8.01045477e-01 -5.21072209e-01 1.41174769e+00
8.34631860e-01 -3.17184746e-01 6.83797002e-01 5.99182546e-01
-1.53276339e-01 -9.66000795e-01 -7.17492998e-01 -5.85506737e-01
3.93613935e-01 -8.78854468e-02 8.33172858e-01 3.49110633e-01
6.57987222e-02 7.79357374e-01 -1.04003459e-01 -1.42901361e-01
2.26460308e-01 4.64854538e-01 8.23271334e-01 -1.07322836e+00
-6.12182796e-01 -4.00423765e-01 3.33815813e-01 -1.56455052e+00
-2.05370598e-02 -1.07352102e+00 5.12029052e-01 -2.08040619e+00
2.54654109e-01 -4.35851276e-01 -3.82342115e-02 1.52297214e-01
-8.87804925e-01 -3.59312773e-01 2.15328723e-01 1.35502979e-01
-9.77697313e-01 8.10867310e-01 1.75222325e+00 -1.40094191e-01
-3.41880739e-01 1.80463910e-01 -9.98443604e-01 3.95410925e-01
5.18248320e-01 -5.83224177e-01 -8.48163366e-01 -5.15736341e-01
6.90200686e-01 5.17000914e-01 3.17194641e-01 -6.39062583e-01
4.18146849e-01 3.63764465e-02 1.19802050e-01 -5.03153205e-01
3.85899469e-02 -4.78286929e-02 -3.61681730e-01 1.95323616e-01
-9.00512695e-01 -9.63521227e-02 -2.60726959e-02 5.86497188e-01
-3.33972722e-01 -5.02301097e-01 1.83467552e-01 -3.27581853e-01
-4.77052808e-01 1.67601556e-01 -2.23904878e-01 8.40225875e-01
4.05396640e-01 -2.19699442e-02 -5.40352464e-01 -7.43413150e-01
-3.89109284e-01 7.68271506e-01 -2.92775035e-02 5.20265341e-01
8.45339179e-01 -1.24911427e+00 -1.21629262e+00 -1.60876021e-01
2.84913033e-01 3.22965294e-01 1.78352118e-01 3.35086405e-01
-2.79662460e-01 6.84322238e-01 4.01580632e-01 -2.69496143e-01
-7.75797725e-01 1.28772631e-01 2.55463064e-01 -1.04628253e+00
-3.90443355e-01 1.06759882e+00 8.08844715e-02 -7.47281730e-01
1.61121175e-01 -2.25533679e-01 -3.81351292e-01 5.51851913e-02
7.75558352e-01 2.67260522e-01 2.21756063e-02 -3.60396411e-03
2.04792336e-01 2.27189332e-01 -2.56070971e-01 -4.49425668e-01
9.16214883e-01 -1.05476052e-01 -9.41376612e-02 9.06859711e-02
8.98353696e-01 -2.27639396e-02 -1.04724884e+00 -4.46429014e-01
1.19737862e-02 -2.68679053e-01 -3.25920969e-01 -1.24464607e+00
-3.43949139e-01 6.49789989e-01 -1.60154164e-01 3.03843796e-01
8.47164571e-01 1.64254427e-01 1.26668751e+00 7.94993281e-01
4.14241999e-01 -8.86018991e-01 4.85176235e-01 1.03727257e+00
9.17234004e-01 -9.68016326e-01 -3.29437464e-01 -1.60167694e-01
-6.50819242e-01 9.33325529e-01 7.77611554e-01 1.71185434e-01
3.06432527e-02 -3.33052874e-01 1.13929771e-01 -1.80071428e-01
-1.11860478e+00 -1.94697455e-01 5.85464776e-01 4.35511053e-01
4.30511177e-01 -4.72383425e-02 -4.46068436e-01 5.67202628e-01
-6.19101286e-01 -4.91352566e-02 4.21379864e-01 8.89209151e-01
-7.65453935e-01 -1.32906032e+00 -2.69899011e-01 4.75993931e-01
-2.50811815e-01 -4.24248278e-01 -5.22293985e-01 3.37592095e-01
-1.91614315e-01 1.30139434e+00 -2.19964877e-01 9.44194477e-03
3.19350004e-01 3.48736942e-01 4.23534364e-01 -1.00389993e+00
-8.28689694e-01 -3.64597887e-01 2.75842845e-01 -1.62337095e-01
-5.37788495e-03 -3.18852991e-01 -1.39171362e+00 1.47357592e-02
-4.52429563e-01 1.07350707e+00 2.51652360e-01 1.14904547e+00
3.69344831e-01 5.98930836e-01 5.89444280e-01 -1.53739288e-01
-8.87682080e-01 -8.60402465e-01 -1.14021413e-01 3.99821758e-01
2.06126362e-01 -2.15084001e-01 -2.60591894e-01 -7.06988871e-02] | [11.457934379577637, 8.069330215454102] |
caf9e2f8-42a8-4c98-a856-852f76e699da | deepfakes-a-new-threat-to-face-recognition | 1812.08685 | null | http://arxiv.org/abs/1812.08685v1 | http://arxiv.org/pdf/1812.08685v1.pdf | DeepFakes: a New Threat to Face Recognition? Assessment and Detection | It is becoming increasingly easy to automatically replace a face of one
person in a video with the face of another person by using a pre-trained
generative adversarial network (GAN). Recent public scandals, e.g., the faces
of celebrities being swapped onto pornographic videos, call for automated ways
to detect these Deepfake videos. To help developing such methods, in this
paper, we present the first publicly available set of Deepfake videos generated
from videos of VidTIMIT database. We used open source software based on GANs to
create the Deepfakes, and we emphasize that training and blending parameters
can significantly impact the quality of the resulted videos. To demonstrate
this impact, we generated videos with low and high visual quality (320 videos
each) using differently tuned parameter sets. We showed that the state of the
art face recognition systems based on VGG and Facenet neural networks are
vulnerable to Deepfake videos, with 85.62% and 95.00% false acceptance rates
respectively, which means methods for detecting Deepfake videos are necessary.
By considering several baseline approaches, we found that audio-visual approach
based on lip-sync inconsistency detection was not able to distinguish Deepfake
videos. The best performing method, which is based on visual quality metrics
and is often used in presentation attack detection domain, resulted in 8.97%
equal error rate on high quality Deepfakes. Our experiments demonstrate that
GAN-generated Deepfake videos are challenging for both face recognition systems
and existing detection methods, and the further development of face swapping
technology will make it even more so. | ['Pavel Korshunov', 'Sebastien Marcel'] | 2018-12-20 | null | null | null | null | ['lip-sync-1'] | ['computer-vision'] | [ 4.96580526e-02 4.63906042e-02 1.87825620e-01 -1.39361560e-01
-5.87808490e-01 -8.75295401e-01 5.79552412e-01 -7.05828786e-01
-1.28925040e-01 8.60209465e-01 1.43071413e-01 -2.70253327e-02
3.52993935e-01 -7.37094164e-01 -7.98278093e-01 -8.21001768e-01
-5.64008169e-02 -4.44121249e-02 2.16971301e-02 2.67277323e-02
1.56175494e-01 6.70697868e-01 -1.73427749e+00 5.85739911e-01
3.87611657e-01 1.03995717e+00 -3.37137580e-01 9.90916371e-01
7.88298398e-02 4.61033225e-01 -1.13360095e+00 -1.02005827e+00
5.37481248e-01 -6.71726942e-01 -5.59866130e-01 -1.56153440e-01
9.97187078e-01 -7.33036339e-01 -4.75861430e-01 1.04718709e+00
8.75803709e-01 -3.82874608e-01 6.73857987e-01 -1.81467569e+00
-5.36410928e-01 3.28284293e-01 -4.71783906e-01 2.81420171e-01
5.18093467e-01 4.55759287e-01 1.51334003e-01 -7.89537966e-01
5.19517124e-01 1.45037782e+00 9.05729711e-01 1.00573909e+00
-9.89208579e-01 -1.42672300e+00 -6.81826174e-01 2.58913904e-01
-1.44199038e+00 -8.52075040e-01 6.67289019e-01 -5.48050046e-01
8.00407767e-01 2.75011986e-01 6.59159720e-01 1.54599440e+00
2.21093521e-01 1.37881458e-01 1.09888911e+00 -3.84408325e-01
-1.22062594e-01 1.75778568e-01 -5.00962079e-01 6.83526039e-01
3.20578098e-01 3.62988681e-01 -4.35965180e-01 -2.04269364e-01
7.25086451e-01 -3.68075967e-01 -3.24219853e-01 9.91151854e-02
-6.81729138e-01 1.06640792e+00 -7.72593468e-02 3.23537976e-01
-2.62378216e-01 2.60063499e-01 6.18996024e-01 4.79803890e-01
-5.33715822e-02 1.23424329e-01 1.25553742e-01 -2.42756352e-01
-1.27110779e+00 -7.17704892e-02 6.25003994e-01 5.76692700e-01
4.45463419e-01 4.11759168e-01 -2.37461463e-01 6.71325147e-01
3.56581211e-01 5.79907477e-01 6.03192985e-01 -8.49710047e-01
1.05005987e-01 1.68369591e-01 -1.31180361e-01 -1.22233307e+00
1.35940582e-01 1.55030593e-01 -5.79983056e-01 5.39512098e-01
4.96081948e-01 -4.07309920e-01 -1.15145266e+00 1.58094454e+00
2.28870437e-01 4.99899477e-01 1.18625417e-01 7.00312376e-01
1.09682822e+00 5.70419669e-01 2.81244721e-02 -1.64689943e-01
1.31908572e+00 -6.16126060e-01 -8.18212926e-01 3.54812026e-01
2.57224709e-01 -1.27540362e+00 8.92574191e-01 6.30003095e-01
-9.49726224e-01 -6.59331083e-01 -1.17752051e+00 4.73770320e-01
-2.85992891e-01 1.92417547e-01 2.87210822e-01 1.79911017e+00
-1.38828838e+00 3.41391772e-01 -2.08414286e-01 -4.97510225e-01
7.34995782e-01 5.32232285e-01 -7.85201311e-01 7.90326372e-02
-1.10101449e+00 6.66290462e-01 1.60238773e-01 -1.76084246e-02
-1.47770905e+00 -4.26387221e-01 -5.44613779e-01 -9.62949917e-02
1.84348762e-01 -1.89408422e-01 8.98429215e-01 -1.49472260e+00
-1.52427649e+00 1.20333624e+00 2.64304280e-01 -5.05825043e-01
6.62442207e-01 -1.52227774e-01 -9.86133873e-01 4.28161979e-01
-2.05812111e-01 8.31531644e-01 1.59583855e+00 -1.02799094e+00
-2.22434387e-01 -1.04171686e-01 -1.41246706e-01 -4.42783087e-01
-2.98971832e-01 4.30641443e-01 3.22903916e-02 -6.91941679e-01
-7.69828081e-01 -7.95543969e-01 7.90276587e-01 1.36206830e-02
-3.97350788e-01 -8.10110047e-02 1.38207233e+00 -1.00545394e+00
8.71185482e-01 -2.18766665e+00 -5.40494502e-01 1.82285547e-01
4.65700515e-02 9.56609786e-01 -3.56907487e-01 4.07294482e-01
-3.57630879e-01 4.43984807e-01 2.85227835e-01 2.33176388e-02
-1.85561553e-01 -4.67726476e-02 -3.53796512e-01 6.18953824e-01
1.82160035e-01 7.23081827e-01 -4.62749302e-01 -2.87603199e-01
2.26265669e-01 9.66798067e-01 -6.29648030e-01 3.42102975e-01
3.55728954e-01 1.76614910e-01 2.19536543e-01 8.18886817e-01
1.04996264e+00 4.65650409e-01 -1.46218210e-01 -5.03031731e-01
3.16130817e-01 -2.03486726e-01 -9.74666417e-01 1.05048680e+00
-2.03729630e-01 1.02702630e+00 1.18832447e-01 -8.42260599e-01
9.52262700e-01 6.02868915e-01 1.85856164e-01 -5.18772840e-01
4.17084575e-01 1.52560219e-01 1.14988722e-01 -6.68141782e-01
2.43809193e-01 -3.30999717e-02 3.18750471e-01 1.57695621e-01
5.15797257e-01 3.99290249e-02 -4.17004786e-02 1.01864256e-01
9.82050240e-01 -7.12039843e-02 -9.93734300e-02 -1.38961077e-01
6.20699048e-01 -6.10060692e-01 2.76106000e-01 6.72005534e-01
-4.83260840e-01 6.60371065e-01 5.19962966e-01 -2.87941605e-01
-1.06656075e+00 -9.94805098e-01 -2.82465778e-02 6.39887094e-01
-3.47374856e-01 -3.49634588e-01 -1.31727958e+00 -8.87884200e-01
-5.84039576e-02 2.85751909e-01 -4.90223408e-01 -3.13986331e-01
-3.71325612e-01 -3.87836218e-01 1.41398966e+00 2.11584046e-01
7.45142698e-01 -1.19022572e+00 -4.06389743e-01 -1.34581821e-02
7.24591687e-02 -1.06764925e+00 -5.53993642e-01 -5.29399037e-01
-2.68649131e-01 -1.40692663e+00 -1.00726938e+00 -7.36722529e-01
4.96152222e-01 -7.96974078e-02 8.92480433e-01 2.26739049e-01
-6.77731097e-01 6.09669924e-01 -3.77726763e-01 -2.70635039e-01
-9.00992036e-01 -5.40233254e-01 1.79218382e-01 3.00805598e-01
4.63835865e-01 -3.87300044e-01 -6.68722451e-01 4.80781227e-01
-9.30112123e-01 -3.51538926e-01 3.36053163e-01 7.74182081e-01
-8.73031691e-02 4.37753126e-02 8.51217330e-01 -5.50049484e-01
6.86044812e-01 -2.60377616e-01 -4.82676059e-01 1.71593636e-01
-3.04383010e-01 -2.68505722e-01 2.84139544e-01 -7.03789592e-01
-7.47379959e-01 7.74711510e-03 -4.64373648e-01 -6.62459373e-01
-3.70707035e-01 -2.28421420e-01 -2.87216038e-01 -5.74008822e-01
7.18450248e-01 1.92592889e-01 2.78639644e-01 -9.30952430e-02
1.06543250e-01 9.71721530e-01 6.92781925e-01 -7.39997625e-02
7.33476102e-01 2.37591580e-01 -1.52557313e-01 -9.57417548e-01
2.83082604e-01 -2.95399339e-03 2.95585692e-02 -7.13703990e-01
8.45475078e-01 -8.78573179e-01 -9.59939718e-01 1.15056384e+00
-1.17303526e+00 -1.20505929e-01 2.12659091e-01 3.17323208e-01
-2.86423534e-01 5.98985970e-01 -5.20652413e-01 -8.33854556e-01
-4.18766588e-01 -1.24925828e+00 9.69107091e-01 2.44571969e-01
-6.14865348e-02 -6.14784300e-01 5.82601726e-02 5.03115296e-01
6.96384668e-01 4.45415765e-01 5.49081087e-01 -4.96947587e-01
-1.84798792e-01 -2.23029345e-01 -1.53277114e-01 8.24140251e-01
3.07184368e-01 4.60459977e-01 -1.42681789e+00 -4.32449996e-01
1.26544442e-02 -4.44663256e-01 7.53450692e-01 2.48601303e-01
1.10993564e+00 -7.02350140e-01 -1.20238341e-01 5.25775075e-01
1.33412111e+00 4.40989792e-01 1.40632331e+00 -1.04196034e-01
5.57763994e-01 4.55415487e-01 1.87611282e-01 1.94690734e-01
-3.54048461e-01 7.86768079e-01 5.93932986e-01 2.43946537e-02
-6.17359519e-01 -3.26200962e-01 7.74430692e-01 4.32060622e-02
-5.37052862e-02 -4.62097257e-01 -5.85599959e-01 2.53528953e-01
-1.00185347e+00 -1.35493839e+00 1.22194126e-01 2.27420902e+00
6.52209520e-01 -1.57152042e-01 4.62129444e-01 3.59546363e-01
1.15714920e+00 -1.71044275e-01 -1.04836106e-01 -6.88787341e-01
-8.96412209e-02 5.83858311e-01 5.40670753e-01 3.38364035e-01
-1.10683692e+00 8.81125093e-01 6.52791262e+00 9.54355776e-01
-1.64486969e+00 3.57694417e-01 7.28265047e-01 -9.47491229e-02
8.81532282e-02 -5.58735907e-01 -8.65456581e-01 9.27533031e-01
1.13438141e+00 7.90581629e-02 3.95443052e-01 7.86213100e-01
4.55932543e-02 3.58219333e-02 -8.48094642e-01 1.56640565e+00
6.20424390e-01 -1.27366841e+00 1.04974769e-01 1.75779000e-01
4.93943810e-01 -3.94483835e-01 1.19235739e-01 2.06204861e-01
1.13255896e-01 -1.59209383e+00 5.19113064e-01 2.10190371e-01
1.22652531e+00 -1.02762020e+00 8.85792136e-01 -3.28031063e-01
-9.39544499e-01 3.38760167e-02 -2.24591628e-01 3.49357456e-01
-1.97916422e-02 2.71900266e-01 -1.02755320e+00 -1.77591350e-02
7.65180767e-01 1.59080282e-01 -5.75639129e-01 1.08023834e+00
3.25028896e-02 7.20465124e-01 -3.15504670e-01 2.67282814e-01
-1.87772453e-01 3.66130918e-01 5.64586103e-01 1.24187028e+00
6.98747218e-01 -4.25571203e-01 -5.59275627e-01 6.63135409e-01
-3.03642452e-01 -6.20778948e-02 -9.15030479e-01 -2.57181019e-01
3.71221602e-01 1.33496201e+00 -5.29849589e-01 -1.22012801e-01
-3.09269935e-01 1.19736838e+00 -3.08474362e-01 1.24562867e-01
-1.18169594e+00 -3.19962233e-01 8.21878016e-01 3.45767885e-01
4.01950002e-01 2.43290693e-01 5.10089755e-01 -9.15776253e-01
-4.60977107e-02 -1.23936963e+00 2.71858841e-01 -8.04520965e-01
-1.19394279e+00 9.25244033e-01 -1.61197305e-01 -1.07876265e+00
-4.37539965e-01 -7.08622694e-01 -7.16415048e-01 6.89716220e-01
-1.04912174e+00 -1.05857825e+00 -5.85542560e-01 9.37601388e-01
4.08148617e-01 -6.70609653e-01 8.17098498e-01 5.91901004e-01
-3.10301900e-01 1.21294379e+00 -2.90443838e-01 3.90316486e-01
9.03624535e-01 -5.88249385e-01 2.43853033e-01 9.83303785e-01
1.60571054e-01 1.63839981e-01 5.64328551e-01 -5.96775234e-01
-1.34665620e+00 -8.89587760e-01 4.45753157e-01 -2.12899983e-01
1.22503504e-01 -4.76907670e-01 -7.15820611e-01 2.84658879e-01
5.27626693e-01 -1.08947856e-02 6.77976370e-01 -7.25830376e-01
-4.69801098e-01 -1.82908908e-01 -1.90450096e+00 2.18001649e-01
7.22773671e-01 -6.88147306e-01 -3.41014713e-01 3.71396206e-02
1.04100369e-01 -3.91321518e-02 -5.57486653e-01 4.53239918e-01
8.01445723e-01 -1.32705235e+00 8.86513114e-01 -3.82650137e-01
1.23252831e-01 -2.13450506e-01 5.67517169e-02 -1.12485862e+00
2.24471807e-01 -8.58852684e-01 -3.34034786e-02 1.69506645e+00
-1.68246720e-02 -7.47757494e-01 8.95144105e-01 2.85300761e-01
2.75849670e-01 -9.89891738e-02 -1.13466418e+00 -8.09319079e-01
-2.22997427e-01 -5.98419718e-02 5.60659647e-01 9.11105812e-01
-4.11928922e-01 -2.22375825e-01 -8.57818246e-01 7.00489804e-02
3.87232244e-01 -7.25743830e-01 1.05086350e+00 -8.67025435e-01
-1.28619269e-01 -3.25111419e-01 -1.09966016e+00 -2.52179623e-01
1.01628140e-01 -4.02018011e-01 -3.28335375e-01 -7.43587375e-01
6.56489953e-02 -6.61105216e-02 7.56842569e-02 5.83126903e-01
3.20782810e-01 1.07710302e+00 1.31202966e-01 -1.71515182e-01
2.21583620e-01 1.04221836e-01 7.13855028e-01 -1.61694139e-01
1.74178496e-01 -3.90123457e-01 -3.75855088e-01 4.93279755e-01
8.33111763e-01 -4.77623343e-01 -3.54300976e-01 1.40042305e-01
-4.23815437e-02 -1.32769927e-01 7.42910922e-01 -1.37502706e+00
-2.08743259e-01 1.94247052e-01 7.02616155e-01 -7.05143735e-02
4.02003765e-01 -7.49438524e-01 5.89780748e-01 6.52627349e-01
5.37930168e-02 4.13618945e-02 5.11639655e-01 1.56113684e-01
-2.87669867e-01 -3.22164714e-01 1.01050377e+00 4.15776893e-02
-6.65594220e-01 8.97835121e-02 -4.58766192e-01 -1.33193195e-01
1.30425572e+00 -5.18581748e-01 -5.49751759e-01 -7.11109519e-01
-5.77317476e-01 -5.82384408e-01 4.38486755e-01 6.40374720e-01
7.34288752e-01 -1.42266774e+00 -7.50870407e-01 7.33896673e-01
-2.70767152e-01 -8.98094177e-01 3.70871007e-01 4.89290565e-01
-8.20400715e-01 9.59286615e-02 -9.02828932e-01 -4.01362032e-01
-1.96328139e+00 4.90909934e-01 5.34614265e-01 4.85366821e-01
-2.45244637e-01 8.92020345e-01 1.02215625e-01 1.96552992e-01
2.46606082e-01 7.30179325e-02 -2.56141573e-01 7.91813135e-02
8.16352844e-01 4.01527405e-01 2.77943134e-01 -9.24476624e-01
-5.10168254e-01 3.91366154e-01 2.39755437e-02 1.16072251e-02
9.12311733e-01 2.34368086e-01 1.89185813e-01 -4.09575015e-01
1.31972790e+00 4.65454280e-01 -1.08990824e+00 6.86657131e-01
-6.60960793e-01 -8.50973248e-01 1.70302298e-02 -8.43671083e-01
-1.55995262e+00 7.32266426e-01 1.30314553e+00 3.17270756e-01
1.25687766e+00 -1.57370329e-01 7.72212505e-01 -2.00898007e-01
2.91849017e-01 -7.30722368e-01 3.50066125e-01 -8.72792974e-02
9.50587511e-01 -1.01875567e+00 -5.42889178e-01 -4.74886000e-01
-4.54584450e-01 1.18150330e+00 6.47243202e-01 -1.54477820e-01
3.78904402e-01 4.51795489e-01 1.37713134e-01 1.63462907e-02
-3.34505618e-01 2.79248118e-01 1.15106396e-01 1.02940869e+00
2.59399027e-01 -8.40415582e-02 -1.42134532e-01 3.39234620e-01
-3.05924773e-01 2.48796642e-01 6.57635808e-01 4.96940464e-01
-4.89453003e-02 -1.27279222e+00 -6.83212817e-01 4.33437526e-01
-1.05453634e+00 2.96239499e-02 -5.69651783e-01 7.62484372e-01
3.83025855e-01 1.04271162e+00 4.40778024e-02 -6.58118367e-01
-1.58286482e-01 1.99907631e-01 6.28934264e-01 -1.05443723e-01
-6.82655454e-01 -1.86077446e-01 1.63257644e-01 -6.51855171e-01
-5.22656918e-01 -3.42393517e-01 -6.25530362e-01 -7.83985257e-01
-2.73354590e-01 2.45517381e-02 6.94193423e-01 5.87378085e-01
4.68454719e-01 -2.10140850e-02 5.01586914e-01 -8.12856436e-01
-2.27847815e-01 -9.92265105e-01 -4.73520219e-01 5.21592975e-01
3.71115953e-01 -8.17893624e-01 -6.40159488e-01 2.69106358e-01] | [12.692086219787598, 1.0992428064346313] |
af2a176b-2d58-40e2-ba52-d502bf372d35 | optiforest-optimal-isolation-forest-for | 2306.12703 | null | https://arxiv.org/abs/2306.12703v2 | https://arxiv.org/pdf/2306.12703v2.pdf | OptIForest: Optimal Isolation Forest for Anomaly Detection | Anomaly detection plays an increasingly important role in various fields for critical tasks such as intrusion detection in cybersecurity, financial risk detection, and human health monitoring. A variety of anomaly detection methods have been proposed, and a category based on the isolation forest mechanism stands out due to its simplicity, effectiveness, and efficiency, e.g., iForest is often employed as a state-of-the-art detector for real deployment. While the majority of isolation forests use the binary structure, a framework LSHiForest has demonstrated that the multi-fork isolation tree structure can lead to better detection performance. However, there is no theoretical work answering the fundamentally and practically important question on the optimal tree structure for an isolation forest with respect to the branching factor. In this paper, we establish a theory on isolation efficiency to answer the question and determine the optimal branching factor for an isolation tree. Based on the theoretical underpinning, we design a practical optimal isolation forest OptIForest incorporating clustering based learning to hash which enables more information to be learned from data for better isolation quality. The rationale of our approach relies on a better bias-variance trade-off achieved by bias reduction in OptIForest. Extensive experiments on a series of benchmarking datasets for comparative and ablation studies demonstrate that our approach can efficiently and robustly achieve better detection performance in general than the state-of-the-arts including the deep learning based methods. | ['Xiaolong Xu', 'Amin Beheshti', 'Mark Dras', 'Wanchun Dou', 'Lianyong Qi', 'Hongsheng Hu', 'Xuyun Zhang', 'Haolong Xiang'] | 2023-06-22 | null | null | null | null | ['anomaly-detection', 'intrusion-detection', 'benchmarking', 'benchmarking'] | ['methodology', 'miscellaneous', 'miscellaneous', 'robots'] | [ 4.93808985e-02 -2.75039345e-01 -4.29878503e-01 -1.07600123e-01
-1.87768310e-01 -3.63745153e-01 5.00856280e-01 2.75807530e-01
-3.66686016e-01 2.16549367e-01 -2.99270004e-01 -7.66044974e-01
-3.57300431e-01 -8.18494439e-01 -2.86197782e-01 -8.97523701e-01
-4.73307848e-01 2.13124275e-01 4.27787662e-01 -3.56148295e-02
4.41712379e-01 7.25378752e-01 -1.15515745e+00 -5.57654537e-02
8.93327653e-01 1.37815571e+00 -4.24543768e-01 3.41915220e-01
-2.10385486e-01 3.97447050e-01 -5.72359204e-01 -4.20012236e-01
6.92333519e-01 -2.69823104e-01 -5.82655489e-01 -1.55382216e-01
1.72568023e-01 -4.63584781e-01 -2.77280301e-01 1.04319859e+00
2.51567215e-01 -1.05883718e-01 6.62336648e-01 -1.57146549e+00
-1.68308064e-01 5.45897782e-01 -1.07100713e+00 7.40714788e-01
-8.88545290e-02 2.69127578e-01 1.17000151e+00 -5.67974448e-01
4.87888278e-03 1.12596047e+00 6.12480521e-01 3.19450915e-01
-1.08647895e+00 -1.02077007e+00 4.30667698e-01 4.40054715e-01
-1.31788254e+00 -1.26710739e-02 6.64549649e-01 -1.12815000e-01
7.94613898e-01 2.82802314e-01 5.42541623e-01 1.11856949e+00
4.87734914e-01 8.29839766e-01 9.89784837e-01 -3.68412197e-01
3.57268393e-01 -1.23333335e-01 2.31743440e-01 7.12374806e-01
7.39889503e-01 4.41390187e-01 -3.43727142e-01 -3.74469250e-01
7.43447065e-01 4.09423441e-01 -8.07869360e-02 -3.84499133e-01
-7.90492892e-01 1.10595942e+00 7.00810432e-01 2.82138109e-01
-2.55850106e-01 5.84329246e-03 6.03980660e-01 4.44934875e-01
3.84834409e-01 5.46001673e-01 -3.46802145e-01 1.27685353e-01
-1.04993677e+00 1.39240339e-01 7.47540474e-01 4.28692758e-01
3.74611050e-01 2.15817973e-01 -9.03448388e-02 4.25166696e-01
3.70046765e-01 3.11638087e-01 1.78600699e-01 -5.63036501e-01
2.66945183e-01 7.65140593e-01 -2.42107838e-01 -1.03264785e+00
-4.87971336e-01 -8.84101570e-01 -1.24085808e+00 2.47295454e-01
3.95491391e-01 2.74824677e-03 -8.48183393e-01 1.42130816e+00
4.51470226e-01 3.44568104e-01 -2.89056420e-01 8.40543985e-01
7.19550774e-02 3.37225735e-01 -6.75188452e-02 -1.92410111e-01
1.25120711e+00 -6.36396587e-01 -3.55704188e-01 -3.32565397e-01
5.43740332e-01 -4.24245775e-01 6.63902581e-01 5.92281044e-01
-1.03350550e-01 -1.21286727e-01 -1.21840596e+00 6.98422432e-01
-3.48713696e-01 -5.79998195e-01 8.07594597e-01 1.04738724e+00
-5.80117404e-01 4.30580586e-01 -9.96149659e-01 -5.28793335e-01
5.42136669e-01 1.41600460e-01 -9.02628079e-02 1.03645220e-01
-1.13014042e+00 7.77933061e-01 6.07405365e-01 -1.52016655e-02
-8.44417274e-01 -3.85095298e-01 -4.36166495e-01 6.74454048e-02
7.76559114e-01 -6.01106286e-01 1.11246943e+00 -6.75716937e-01
-1.06338012e+00 5.39020360e-01 1.96780920e-01 -1.14353883e+00
4.77257192e-01 -4.18080628e-01 -4.36791450e-01 1.70646414e-01
1.60294119e-02 1.69256270e-01 1.16924596e+00 -9.45877492e-01
-9.81408596e-01 -5.85608661e-01 7.03917351e-03 -1.24227695e-01
-5.64501286e-01 1.29055791e-02 2.42445748e-02 -7.00749874e-01
2.12837279e-01 -7.41052568e-01 -4.91855741e-01 -7.60687068e-02
-6.19786859e-01 -1.52355701e-01 1.14100492e+00 -3.65942270e-01
1.67593169e+00 -2.00719190e+00 -3.15791726e-01 5.60754657e-01
5.42013824e-01 4.91476089e-01 1.16215698e-01 3.66260141e-01
-8.87123719e-02 1.76048815e-01 -4.31248844e-01 2.48631939e-01
-1.69438794e-01 1.06537610e-01 -6.06687367e-01 6.70278549e-01
2.31135234e-01 4.84779030e-01 -8.25706303e-01 -3.46908092e-01
1.50957912e-01 3.51030491e-02 -6.59502149e-01 2.85136968e-01
1.59061119e-01 3.36126268e-01 -6.38152301e-01 8.97476614e-01
5.31838357e-01 -1.45752952e-01 -1.01194583e-01 1.55431464e-01
-4.06526141e-02 1.61997959e-01 -1.18815804e+00 1.03533340e+00
-7.59758279e-02 3.27505350e-01 2.82321364e-01 -1.37961280e+00
1.04791439e+00 1.03362948e-02 5.37596703e-01 -6.34543538e-01
7.96616152e-02 3.62329632e-01 3.57356966e-01 -1.47263214e-01
2.63887525e-01 -2.17781991e-01 -1.17472239e-01 4.60430026e-01
-3.54663610e-01 4.46266592e-01 -3.81091312e-02 2.06442356e-01
1.54980457e+00 -3.71973813e-01 6.10247612e-01 -2.26216078e-01
4.72674757e-01 -1.92302495e-01 5.89359939e-01 1.05777979e+00
-6.30404770e-01 1.45357087e-01 5.06694794e-01 -7.27378190e-01
-6.84510171e-01 -1.16281962e+00 -3.70575845e-01 9.30775046e-01
1.45455793e-01 -3.33058119e-01 -6.63472354e-01 -1.02473974e+00
1.43170223e-01 6.11853182e-01 -5.37663639e-01 -3.76016766e-01
-5.60738683e-01 -1.13811183e+00 7.01513350e-01 4.28914577e-01
8.34801733e-01 -9.14968431e-01 -1.07079005e+00 1.05575122e-01
-2.49596834e-02 -1.02031505e+00 -2.44572595e-01 5.16825855e-01
-1.03595114e+00 -1.16698110e+00 -8.90244097e-02 -2.07452938e-01
3.46116424e-01 4.52003270e-01 8.84386420e-01 2.07578331e-01
-3.46198678e-01 1.75057590e-01 -5.25895715e-01 -5.70797682e-01
-2.99461454e-01 3.67215097e-01 3.47984850e-01 1.73362717e-01
5.62328041e-01 -7.11086571e-01 -7.24394023e-01 3.69701624e-01
-1.09040320e+00 -4.87007946e-01 1.02945197e+00 8.16329360e-01
2.39654541e-01 3.44121248e-01 6.17208004e-01 -8.42971265e-01
5.89460790e-01 -6.56409204e-01 -5.32504857e-01 -4.10314426e-02
-1.13813424e+00 2.19358608e-01 6.21667385e-01 -1.38626903e-01
-4.51138377e-01 -3.96291554e-01 -6.57007694e-02 -6.33170605e-01
-2.61263013e-01 3.10847253e-01 2.05299072e-02 5.34060188e-02
7.24215209e-01 2.08283335e-01 9.72784460e-02 -3.20256203e-01
1.02699362e-01 5.60299754e-01 3.59045267e-01 -5.10802925e-01
1.04272377e+00 5.22671044e-01 2.57568717e-01 -7.36835897e-01
-8.76929641e-01 -7.09066510e-01 -5.30046463e-01 -9.08258557e-03
3.90194207e-01 -4.27930772e-01 -6.67698562e-01 4.56356436e-01
-8.14392745e-01 1.53330341e-01 -5.76068759e-02 1.19267209e-02
-8.54049250e-02 6.52235627e-01 -4.95190054e-01 -1.20654607e+00
-7.69180715e-01 -7.80989766e-01 8.00116599e-01 3.45267504e-02
-1.52000546e-01 -7.34817266e-01 -3.52531597e-02 -2.84987893e-02
5.84469199e-01 5.06515503e-01 1.00819898e+00 -1.05923402e+00
-6.20723903e-01 -4.53581810e-01 -4.04723525e-01 3.77027005e-01
8.33349079e-02 -1.55804977e-02 -7.90999115e-01 -5.38651288e-01
5.47414310e-02 2.04232246e-01 1.06202495e+00 2.59177089e-01
1.25314689e+00 -6.10701740e-01 -4.83162999e-01 6.32896245e-01
1.31706548e+00 2.72247851e-01 5.69256961e-01 8.04744244e-01
6.94543183e-01 5.09992301e-01 6.77304149e-01 5.61395228e-01
4.04950604e-02 4.46883678e-01 9.30662572e-01 8.12603459e-02
4.12305683e-01 -6.83893040e-02 4.93022352e-01 4.80008215e-01
2.12544635e-01 -7.82059431e-02 -1.11791503e+00 4.15161759e-01
-1.80256665e+00 -1.12547183e+00 1.13756470e-01 2.49089646e+00
3.80614370e-01 6.96826756e-01 5.24559200e-01 5.66898584e-01
6.85172141e-01 3.13039780e-01 -6.52054191e-01 -4.62287098e-01
1.70461401e-01 8.57724473e-02 5.96766412e-01 7.99642038e-03
-1.41215634e+00 7.04959393e-01 6.16630268e+00 9.64120090e-01
-1.32211506e+00 -1.86454669e-01 8.06215703e-01 -1.53479939e-02
2.46676624e-01 8.57816339e-02 -9.28377151e-01 4.82500315e-01
9.73829150e-01 9.39847156e-03 2.15235859e-01 9.58199322e-01
1.14222765e-02 -4.16626558e-02 -9.38059688e-01 7.55143166e-01
-3.01670730e-01 -9.01381373e-01 1.25798315e-01 4.70956862e-01
1.95905343e-01 4.13289741e-02 8.70715156e-02 4.59387332e-01
1.08999856e-01 -8.88073623e-01 4.69480604e-01 1.56337733e-03
4.55056578e-01 -9.64728832e-01 8.57727587e-01 5.36754668e-01
-1.16594195e+00 -5.28308809e-01 -1.02935098e-01 -1.36874586e-01
-1.65443301e-01 1.08683121e+00 -9.72646356e-01 7.47685254e-01
9.37470436e-01 4.98063356e-01 -5.92289269e-01 1.17300057e+00
-9.39606801e-02 9.44603562e-01 -5.38398206e-01 -6.33127466e-02
4.76076126e-01 -1.87028468e-01 9.30450022e-01 1.24594426e+00
2.12708369e-01 -3.12717676e-01 4.47303742e-01 6.26296163e-01
1.83214501e-01 2.09116265e-02 -6.64161503e-01 -5.90243551e-04
6.90474153e-01 1.27210557e+00 -9.67126906e-01 1.05191715e-01
-2.53290087e-01 5.03112495e-01 1.40210077e-01 6.72757402e-02
-8.69303346e-01 -3.36042404e-01 7.28797913e-01 4.84662354e-01
4.94961709e-01 -1.71874195e-01 -5.91271341e-01 -9.63703513e-01
6.90648258e-02 -1.09130979e+00 8.31720233e-01 2.71344781e-01
-1.63252842e+00 7.06166327e-01 1.19407050e-01 -1.41526127e+00
-1.26945794e-01 -7.84460545e-01 -8.89936566e-01 3.63411278e-01
-1.38553309e+00 -8.47865641e-01 -1.73222646e-01 5.18058300e-01
2.60599315e-01 -2.67672867e-01 6.66234076e-01 7.81383365e-02
-1.04975128e+00 6.99092329e-01 -4.35453244e-02 4.12704945e-01
5.55689752e-01 -1.29098058e+00 4.67802465e-01 1.34034276e+00
2.08138332e-01 6.04295313e-01 7.32193351e-01 -6.13678634e-01
-1.17257440e+00 -1.06611860e+00 2.31723487e-02 -3.88543010e-01
8.29261839e-01 -1.44557148e-01 -9.75125372e-01 3.67946059e-01
-4.43329178e-02 6.54691905e-02 6.34163380e-01 4.42085683e-01
-7.95549810e-01 -4.01104003e-01 -1.18787086e+00 5.79614282e-01
9.80200708e-01 -1.30040675e-01 -4.42153603e-01 -1.40993237e-01
4.64830130e-01 -1.08533753e-02 -5.84478617e-01 8.20705414e-01
4.57798928e-01 -1.25029135e+00 8.71230781e-01 -4.57668394e-01
1.21771097e-01 -3.01270336e-01 -1.69582054e-01 -1.06486630e+00
-4.40052092e-01 -5.53282082e-01 -6.69633806e-01 9.86954987e-01
1.71351880e-01 -1.13035381e+00 7.80902028e-01 5.11430688e-02
1.41828865e-01 -1.21799970e+00 -1.23411810e+00 -1.09889376e+00
1.32167161e-01 -4.04201657e-01 4.20456648e-01 8.61437201e-01
-2.95425951e-01 4.64742929e-01 -2.31347084e-01 3.38390559e-01
8.34803998e-01 2.94505030e-01 8.09767306e-01 -1.65761757e+00
-2.74829239e-01 -7.08903551e-01 -6.78877175e-01 -8.75665545e-01
-2.22806916e-01 -6.66900635e-01 -2.08023190e-01 -1.09347701e+00
1.58188000e-01 -5.22143960e-01 -6.79535985e-01 5.14948428e-01
-3.81334543e-01 -1.45722449e-01 9.30080637e-02 4.63338733e-01
-7.81084299e-01 4.42384720e-01 6.72768354e-01 -6.22671954e-02
-9.52945799e-02 1.85874969e-01 -8.53787839e-01 6.71488404e-01
9.36837256e-01 -6.64872348e-01 -1.09212838e-01 9.38439146e-02
-1.32420421e-01 -2.78316468e-01 2.88812846e-01 -1.14154625e+00
1.00595251e-01 -1.59007341e-01 2.85891771e-01 -4.53763515e-01
-1.35559142e-01 -8.60486448e-01 -3.94163638e-01 8.48651767e-01
1.10459730e-01 3.95007521e-01 1.69941008e-01 9.97456670e-01
-1.48700967e-01 -2.00260867e-04 9.95338500e-01 1.61057904e-01
-6.48611307e-01 5.13391376e-01 -2.86161691e-01 1.23924337e-01
1.10695302e+00 -3.03356558e-01 -2.64749855e-01 -3.75529855e-01
-3.69247757e-02 2.48350158e-01 2.29235157e-01 5.52828729e-01
5.83839834e-01 -1.12784994e+00 -7.39610553e-01 4.10025269e-01
1.20052241e-01 -9.25507918e-02 -2.01178372e-01 1.05063820e+00
-4.15618211e-01 4.37366366e-01 -2.02786267e-01 -9.79624093e-01
-1.03782225e+00 6.60269916e-01 3.30956191e-01 -9.16424215e-01
-6.66182637e-01 6.26788378e-01 1.71786860e-01 -2.33191267e-01
4.11104083e-01 -9.45436805e-02 1.80850908e-01 -1.58467278e-01
6.04118526e-01 4.47779208e-01 1.90129459e-01 -2.79330581e-01
-6.08136058e-01 1.23788230e-01 -5.23078144e-01 1.83267027e-01
1.02966022e+00 1.76512599e-01 -6.45444319e-02 1.50429145e-01
7.49085128e-01 -2.40706488e-01 -1.06156552e+00 -3.45855594e-01
5.48991144e-01 -6.26286149e-01 1.22444525e-01 -5.79396605e-01
-1.19639361e+00 8.13869834e-01 7.82815814e-01 7.22850144e-01
1.39210212e+00 -3.57368231e-01 8.37696195e-01 4.78731006e-01
6.95467234e-01 -8.13235223e-01 2.84365773e-01 5.14099777e-01
5.81780910e-01 -1.17724097e+00 5.49630933e-02 -3.38748783e-01
-3.67957532e-01 9.73151565e-01 8.14062238e-01 -2.33215064e-01
7.74216652e-01 3.53425711e-01 -1.66136354e-01 -2.27842346e-01
-6.65477276e-01 -8.62780064e-02 -1.31068360e-02 5.13405561e-01
1.76452443e-01 1.70579687e-01 -7.33946189e-02 4.20056671e-01
-1.49124749e-02 -5.58399856e-01 1.46195799e-01 8.79041255e-01
-7.38198459e-01 -1.06639254e+00 -6.12406254e-01 8.77089202e-01
-7.19694495e-01 3.76753882e-03 -4.18938071e-01 8.12998712e-01
-3.73818278e-02 1.01954210e+00 -2.92121582e-02 -5.47242999e-01
3.34472269e-01 1.02995746e-01 1.61157012e-01 -3.66245002e-01
-6.52120531e-01 -1.22926034e-01 -2.40906194e-01 -6.55804932e-01
-9.86589268e-02 -5.14848769e-01 -9.64955151e-01 -5.84271729e-01
-4.98888999e-01 2.36719653e-01 3.42962772e-01 1.02163470e+00
3.64061087e-01 5.30943036e-01 8.55519772e-01 -4.88710910e-01
-9.18501437e-01 -9.55944061e-01 -5.62101543e-01 1.38636976e-01
4.39806938e-01 -7.88276672e-01 -6.17070794e-01 -7.80475736e-01] | [7.558152675628662, 2.627147674560547] |
cf58f7d2-c04f-4c01-8075-1211d76052b3 | unsupervised-video-summarization-with-a | 2105.11131 | null | https://arxiv.org/abs/2105.11131v1 | https://arxiv.org/pdf/2105.11131v1.pdf | Unsupervised Video Summarization with a Convolutional Attentive Adversarial Network | With the explosive growth of video data, video summarization, which attempts to seek the minimum subset of frames while still conveying the main story, has become one of the hottest topics. Nowadays, substantial achievements have been made by supervised learning techniques, especially after the emergence of deep learning. However, it is extremely expensive and difficult to collect human annotation for large-scale video datasets. To address this problem, we propose a convolutional attentive adversarial network (CAAN), whose key idea is to build a deep summarizer in an unsupervised way. Upon the generative adversarial network, our overall framework consists of a generator and a discriminator. The former predicts importance scores for all frames of a video while the latter tries to distinguish the score-weighted frame features from original frame features. Specifically, the generator employs a fully convolutional sequence network to extract global representation of a video, and an attention-based network to output normalized importance scores. To learn the parameters, our objective function is composed of three loss functions, which can guide the frame-level importance score prediction collaboratively. To validate this proposed method, we have conducted extensive experiments on two public benchmarks SumMe and TVSum. The results show the superiority of our proposed method against other state-of-the-art unsupervised approaches. Our method even outperforms some published supervised approaches. | ['Yanning Zhang', 'Shizhou Zhang', 'Shucheng Li', 'Yanbing Lv', 'Guoqiang Liang'] | 2021-05-24 | null | null | null | null | ['unsupervised-video-summarization'] | ['computer-vision'] | [ 3.61724764e-01 -1.77945912e-01 -1.13100477e-01 -2.49063194e-01
-9.60814536e-01 -1.35434479e-01 6.17166221e-01 -1.94218953e-03
-3.59017164e-01 7.73841977e-01 5.85946500e-01 2.10919932e-01
2.43516028e-01 -7.70955503e-01 -7.58855999e-01 -8.10611248e-01
1.57125086e-01 8.76021236e-02 2.81542957e-01 -8.64829868e-02
3.66710275e-01 -4.85344231e-02 -1.34578645e+00 2.97476351e-01
1.03732598e+00 1.19414902e+00 2.73504287e-01 4.85384345e-01
5.28866192e-03 1.42736924e+00 -8.15156817e-01 -4.19978172e-01
4.25239578e-02 -9.16978180e-01 -6.08622372e-01 1.52641207e-01
2.42361173e-01 -4.48822945e-01 -8.01811874e-01 1.25166309e+00
6.02858245e-01 3.87467593e-01 5.86265862e-01 -1.07645500e+00
-7.43972480e-01 6.73748732e-01 -4.64524329e-01 3.99753779e-01
2.13838905e-01 3.83140802e-01 1.19040847e+00 -6.36250913e-01
4.31382537e-01 9.44607615e-01 3.44523042e-01 4.88076508e-01
-8.45480084e-01 -5.30217886e-01 7.30732381e-02 5.58773220e-01
-1.21470785e+00 -4.26499724e-01 1.14261627e+00 -1.99443892e-01
4.60735619e-01 1.91636324e-01 6.64392948e-01 1.19418466e+00
7.98882693e-02 1.03661752e+00 6.60001874e-01 7.58559778e-02
2.72205681e-01 -2.40876094e-01 -3.48011017e-01 6.51183724e-01
-1.00179620e-01 -3.13529700e-01 -4.82869685e-01 1.03634752e-01
7.14612842e-01 2.07038492e-01 -4.80509907e-01 -1.34734541e-01
-1.16328382e+00 6.69398904e-01 5.15723765e-01 3.11522096e-01
-5.89013278e-01 2.23698109e-01 5.68305969e-01 2.14944884e-01
7.00018764e-01 2.21503779e-01 7.23710954e-02 -2.49351636e-01
-1.14406967e+00 3.13669324e-01 4.73575294e-01 6.11648321e-01
5.22147179e-01 1.24268793e-01 -5.91757655e-01 7.81370878e-01
6.90149590e-02 2.04881638e-01 7.24532545e-01 -8.92829895e-01
6.90556288e-01 6.79866195e-01 -9.06679407e-03 -1.26210690e+00
6.95429370e-02 -5.79703093e-01 -1.19727886e+00 -1.13842383e-01
6.63537532e-02 -2.70089716e-01 -6.59880459e-01 1.74577999e+00
4.27279212e-02 5.19139051e-01 8.21016058e-02 1.14118564e+00
8.97430599e-01 1.12840736e+00 1.22468963e-01 -4.53225017e-01
9.88060415e-01 -1.30579948e+00 -6.23760879e-01 1.07490399e-03
1.73477769e-01 -6.14660621e-01 8.52802396e-01 2.15470448e-01
-1.24479258e+00 -7.67317832e-01 -1.19043291e+00 -4.75590266e-02
1.72886491e-01 9.38908011e-02 3.51984262e-01 8.56093541e-02
-9.06306446e-01 6.97875500e-01 -6.74139440e-01 -5.85611798e-02
7.27491140e-01 1.68857858e-01 -2.41246998e-01 -2.04382464e-02
-1.19568729e+00 4.15421546e-01 6.13552272e-01 5.96070401e-02
-1.18311536e+00 -3.27164501e-01 -7.07897067e-01 4.48458910e-01
5.64177930e-01 -7.14016438e-01 1.07150376e+00 -1.41355979e+00
-1.55033422e+00 5.50075352e-01 9.86622721e-02 -6.47944689e-01
5.95982850e-01 -2.48413205e-01 -2.78568178e-01 4.03783053e-01
1.58383250e-01 4.71423715e-01 9.17717874e-01 -9.91615355e-01
-7.74460673e-01 4.82873693e-02 1.31235301e-01 3.17019582e-01
-6.63788378e-01 -4.74957414e-02 -6.25393569e-01 -1.06024754e+00
-3.24066430e-01 -4.43692625e-01 -1.44637108e-01 -3.21167052e-01
-3.78837943e-01 -3.65994602e-01 6.44683301e-01 -8.48325789e-01
1.46438575e+00 -2.06899381e+00 5.33617020e-01 -2.78979719e-01
3.66954088e-01 4.63931262e-01 -1.61041871e-01 2.65921026e-01
1.64425559e-02 -8.38049427e-02 -4.65041667e-01 -3.90360057e-01
-1.24442615e-01 -7.23708719e-02 -2.79010057e-01 1.90450490e-01
3.53998333e-01 8.89631391e-01 -1.17428493e+00 -6.03069961e-01
8.87480378e-02 3.71246099e-01 -7.03134775e-01 6.62774324e-01
-4.24962819e-01 5.50445557e-01 -6.21423900e-01 3.94763350e-01
3.60496104e-01 -3.70342284e-01 -1.05624394e-02 -2.07698211e-01
-3.94075662e-02 1.68012857e-01 -8.86233091e-01 1.85376883e+00
-2.06553370e-01 6.02395475e-01 -2.97137201e-01 -1.47729647e+00
8.91941965e-01 3.45098525e-01 6.13307118e-01 -5.50494432e-01
3.39929700e-01 1.60479903e-01 -1.24064013e-01 -6.22510076e-01
4.60503250e-01 2.63153147e-02 -1.47638589e-01 3.07674825e-01
2.23035604e-01 1.43751130e-01 2.58775622e-01 2.93742716e-01
1.28764415e+00 4.21693809e-02 2.54480481e-01 8.36735517e-02
9.18211997e-01 -2.02683404e-01 7.48291671e-01 3.82364452e-01
-1.29122496e-01 8.45322013e-01 7.73613572e-01 -5.05409420e-01
-1.04047763e+00 -7.47428954e-01 5.74012935e-01 9.30825174e-01
4.22322124e-01 -4.23878700e-01 -9.47939038e-01 -9.75833118e-01
-4.39046115e-01 4.51822340e-01 -5.58110893e-01 -3.99524182e-01
-6.67495251e-01 -5.68448544e-01 3.43967855e-01 4.13253665e-01
9.34874356e-01 -1.26459217e+00 -4.45789605e-01 2.71308511e-01
-5.58340669e-01 -9.88033712e-01 -7.81994700e-01 -3.27394545e-01
-6.06251180e-01 -9.23335791e-01 -1.07696605e+00 -8.13268721e-01
5.69113016e-01 2.12551087e-01 1.02735913e+00 3.08696032e-01
1.56524494e-01 2.14504879e-02 -5.78111887e-01 -1.44014999e-01
-3.34878385e-01 3.27169716e-01 -2.95621336e-01 5.11333048e-01
6.64940998e-02 -7.74977565e-01 -9.53879774e-01 1.42178476e-01
-1.19981933e+00 2.73946643e-01 7.70375073e-01 7.91244328e-01
6.95122004e-01 7.80984387e-02 7.73359954e-01 -8.19593787e-01
6.37915134e-01 -6.68916106e-01 -3.05274129e-01 1.38416409e-01
-1.45121306e-01 7.52828494e-02 1.06004143e+00 -2.27243021e-01
-9.54986095e-01 -2.15267420e-01 -3.27912956e-01 -5.85997403e-01
4.21825051e-02 5.94341815e-01 -4.44277853e-01 4.41095084e-01
2.91076332e-01 6.08218908e-01 -2.00102925e-01 -3.12015742e-01
2.40505785e-01 5.59887826e-01 9.25488174e-01 -3.45450521e-01
7.57815063e-01 3.09810102e-01 -2.68270224e-01 -6.01878464e-01
-1.01072800e+00 -2.25586832e-01 -2.14621723e-01 -3.61481100e-01
9.20459390e-01 -9.13154900e-01 -3.98107678e-01 5.02112687e-01
-1.22127688e+00 -1.45145670e-01 -3.58942091e-01 2.60367006e-01
-6.35842144e-01 4.21425760e-01 -3.72929931e-01 -4.81038690e-01
-5.96231520e-01 -1.22101498e+00 8.85712743e-01 4.71174061e-01
2.19896324e-02 -7.81404614e-01 1.62689403e-01 5.10344207e-01
4.66594994e-01 5.12796700e-01 6.34568095e-01 -7.24600315e-01
-7.14186192e-01 -1.94098562e-01 -1.99802652e-01 7.05241919e-01
1.89792871e-01 -1.44194841e-01 -6.55398011e-01 -1.55014381e-01
1.42842233e-01 -5.40167689e-01 1.15618777e+00 2.17084333e-01
1.54120803e+00 -6.02942407e-01 1.35624139e-02 5.84705710e-01
1.28623748e+00 1.58631533e-01 8.85094404e-01 2.51426131e-01
7.72006929e-01 1.57678545e-01 5.20802140e-01 5.18255293e-01
3.70347530e-01 5.62250733e-01 6.78873599e-01 1.02007352e-01
-1.08990885e-01 -3.91635507e-01 4.64436978e-01 1.10764205e+00
-5.10270655e-01 -4.88331109e-01 -4.70468283e-01 4.14703310e-01
-2.06146932e+00 -1.35458481e+00 3.77250552e-01 2.05218577e+00
9.13233876e-01 3.10280621e-01 1.32427350e-01 2.40336880e-02
7.55429268e-01 7.61449695e-01 -6.23459518e-01 1.04946330e-01
-7.82033876e-02 1.41920030e-01 6.17802292e-02 7.54452497e-02
-1.33561313e+00 7.36265481e-01 4.62566710e+00 1.06655395e+00
-1.11792886e+00 1.76086202e-01 9.66218531e-01 -2.46182188e-01
-2.33538091e-01 -1.48330599e-01 -2.00102285e-01 1.00968409e+00
7.42389381e-01 -3.06828260e-01 3.06099325e-01 7.73192525e-01
3.55178773e-01 2.14621484e-01 -8.45854580e-01 9.83796179e-01
3.65906954e-01 -1.57079399e+00 3.24065477e-01 -3.42751265e-01
9.98423696e-01 -9.76398438e-02 6.15785867e-02 4.64382112e-01
1.95365748e-03 -9.15808380e-01 7.37504780e-01 7.74265409e-01
7.00458527e-01 -1.01903796e+00 9.93745387e-01 4.45191801e-01
-1.20950067e+00 1.08782649e-02 -4.10828143e-01 -5.43458126e-02
1.31513923e-01 5.12372136e-01 -1.51098534e-01 7.26448655e-01
4.49575961e-01 1.03052866e+00 -3.41553837e-01 1.10054684e+00
-4.53658223e-01 8.02995682e-01 2.35099733e-01 -1.85999479e-02
3.88605237e-01 -1.98750958e-01 6.30960107e-01 1.05165088e+00
3.32079560e-01 2.44622976e-01 2.56366581e-01 6.15727067e-01
-7.48964608e-01 1.90913349e-01 -3.00294846e-01 -5.23462631e-02
2.90687293e-01 1.27932155e+00 -6.34216189e-01 -5.48535168e-01
-3.73560429e-01 1.27121449e+00 4.02601540e-01 9.42466259e-02
-1.08235049e+00 -5.63655019e-01 3.78794909e-01 -5.03449179e-02
3.30704480e-01 7.22508878e-02 4.34972569e-02 -1.43445623e+00
1.19242467e-01 -1.10550487e+00 2.74219453e-01 -6.47564769e-01
-1.09263515e+00 6.50222301e-01 -2.94534683e-01 -1.52304685e+00
-1.89023077e-01 -3.48051190e-02 -1.12092865e+00 6.80193424e-01
-1.41340125e+00 -8.23287964e-01 -5.57920039e-01 5.12316763e-01
1.05482996e+00 -4.39863235e-01 3.63404930e-01 4.38824743e-01
-8.09669793e-01 5.00531554e-01 2.43414357e-01 3.90521377e-01
6.62405849e-01 -1.06193745e+00 2.66546577e-01 1.08893371e+00
6.35177046e-02 2.70727694e-01 6.55589938e-01 -5.37123024e-01
-1.15046978e+00 -1.34718525e+00 6.36961401e-01 3.99675369e-02
5.84788203e-01 2.36574560e-02 -8.92667592e-01 3.57168138e-01
5.29521167e-01 1.13573954e-01 4.50581580e-01 -5.98816276e-01
-2.62966920e-02 -2.83671498e-01 -7.53454387e-01 4.95169193e-01
1.03038788e+00 -1.97022721e-01 -6.82792127e-01 2.95461029e-01
8.83481205e-01 -3.70626450e-01 -5.92139721e-01 3.77364188e-01
3.62085313e-01 -1.12668324e+00 7.68366098e-01 -4.53330636e-01
1.30152726e+00 -2.74808913e-01 3.00523620e-02 -1.45514727e+00
-2.91709572e-01 -6.97286725e-01 -3.55297178e-01 1.43784130e+00
9.96908639e-03 -2.75926352e-01 8.14817190e-01 1.32438079e-01
-3.79256546e-01 -1.01970279e+00 -7.54610658e-01 -4.45707738e-01
-3.14810395e-01 -7.16927871e-02 5.24952650e-01 6.85239196e-01
-3.08238745e-01 6.32787764e-01 -7.11343825e-01 -8.33129287e-02
5.80792367e-01 1.54301003e-01 8.17760587e-01 -9.22226191e-01
-4.40793872e-01 -6.36830747e-01 -5.11559844e-01 -1.22408223e+00
2.15594783e-01 -8.28202188e-01 5.11483923e-02 -1.64627969e+00
5.24597645e-01 -9.47011709e-02 -5.79529166e-01 2.64253497e-01
-6.50014460e-01 2.83716202e-01 2.09966108e-01 3.88178468e-01
-1.14434075e+00 1.01947391e+00 1.29338241e+00 -4.85674024e-01
-6.21745586e-02 3.75279710e-02 -8.39714229e-01 8.46168637e-01
7.65388846e-01 -4.57803637e-01 -5.30011296e-01 -6.56482041e-01
7.35602453e-02 2.03874886e-01 3.13916802e-01 -1.33498335e+00
2.79398054e-01 -1.46051452e-01 3.56856614e-01 -5.18114865e-01
9.16562229e-02 -5.91215670e-01 -2.64683571e-02 3.10955644e-01
-5.34745097e-01 -7.61809349e-02 -2.72113740e-01 6.86435282e-01
-6.19559944e-01 -1.68932706e-01 7.49211133e-01 -1.27773687e-01
-6.93721414e-01 6.60803795e-01 -4.08183411e-02 3.13887596e-01
9.33698416e-01 1.66627631e-01 -2.31780633e-01 -4.46830571e-01
-3.45648915e-01 3.20823282e-01 3.77128214e-01 3.73508692e-01
7.20409274e-01 -1.47378063e+00 -1.05487061e+00 -1.11083888e-01
-6.27274960e-02 1.96698427e-01 4.77708817e-01 7.04427302e-01
-6.53385818e-01 7.00015202e-02 -2.25573361e-01 -2.85268486e-01
-9.91345465e-01 4.69619691e-01 2.13482101e-02 -5.73155701e-01
-5.63224435e-01 7.44765460e-01 2.16864258e-01 2.59750307e-01
2.65072197e-01 -3.75101314e-04 -4.85349596e-01 6.53700531e-02
7.60524035e-01 4.19665456e-01 -1.88610062e-01 -6.87776685e-01
9.24187005e-02 1.25123739e-01 -8.48210528e-02 2.02280343e-01
1.58688676e+00 -3.59608829e-02 -1.38875008e-01 1.79548591e-01
1.34254086e+00 -1.36534989e-01 -1.48189938e+00 -3.38818580e-01
-3.41787428e-01 -3.37101907e-01 6.77908286e-02 -3.84357184e-01
-1.37858212e+00 8.76337171e-01 2.12878510e-01 2.54604489e-01
1.46328366e+00 -1.11820176e-01 1.19633281e+00 1.26346543e-01
1.34910166e-01 -9.51648414e-01 4.29129303e-01 3.52065474e-01
8.65164876e-01 -1.17563319e+00 -9.89146307e-02 -1.26645956e-02
-7.09960699e-01 9.90995944e-01 5.80989838e-01 -4.16608751e-01
1.43930078e-01 -1.65794849e-01 -3.23905617e-01 3.15987952e-02
-6.89571500e-01 -4.20716517e-02 3.13016534e-01 1.88568309e-01
3.51204723e-01 -1.12980999e-01 -4.56086010e-01 9.81849611e-01
-1.66859105e-02 1.61086157e-01 2.57190973e-01 5.61293423e-01
-5.99921048e-01 -9.20055389e-01 -1.34354010e-01 6.04664564e-01
-7.19806910e-01 1.65849971e-03 -9.61640701e-02 3.88108999e-01
1.57615036e-01 7.72164702e-01 6.68920251e-03 -5.38729012e-01
3.62174809e-01 -1.51918203e-01 1.10477999e-01 -5.48221290e-01
-5.34238756e-01 -6.13934845e-02 -3.21465820e-01 -4.58169401e-01
-6.67645037e-01 -5.58749974e-01 -1.10150909e+00 -2.18497485e-01
-3.91931310e-02 3.88302743e-01 1.31212905e-01 9.48017836e-01
2.00680569e-01 9.48931217e-01 1.09829557e+00 -9.74866629e-01
-5.21180987e-01 -9.32150185e-01 -2.68832862e-01 6.57746613e-01
2.32023448e-01 -5.06088555e-01 -1.64846942e-01 3.69149327e-01] | [10.39399528503418, 0.44246768951416016] |
598b7d7d-cd0f-4b1f-9524-e62a5ec1d68e | a-wireless-vision-dataset-for-privacy | 2205.11962 | null | https://arxiv.org/abs/2205.11962v1 | https://arxiv.org/pdf/2205.11962v1.pdf | A Wireless-Vision Dataset for Privacy Preserving Human Activity Recognition | Human Activity Recognition (HAR) has recently received remarkable attention in numerous applications such as assisted living and remote monitoring. Existing solutions based on sensors and vision technologies have obtained achievements but still suffering from considerable limitations in the environmental requirement. Wireless signals like WiFi-based sensing have emerged as a new paradigm since it is convenient and not restricted in the environment. In this paper, a new WiFi-based and video-based neural network (WiNN) is proposed to improve the robustness of activity recognition where the synchronized video serves as the supplement for the wireless data. Moreover, a wireless-vision benchmark (WiVi) is collected for 9 class actions recognition in three different visual conditions, including the scenes without occlusion, with partial occlusion, and with full occlusion. Both machine learning methods - support vector machine (SVM) as well as deep learning methods are used for the accuracy verification of the data set. Our results show that WiVi data set satisfies the primary demand and all three branches in the proposed pipeline keep more than $80\%$ of activity recognition accuracy over multiple action segmentation from 1s to 3s. In particular, WiNN is the most robust method in terms of all the actions on three action segmentation compared to the others. | ['Yuanwei Liu', 'Zhiyuan Shi', 'Yanling Hao'] | 2022-05-24 | null | null | null | null | ['action-segmentation'] | ['computer-vision'] | [ 6.42713249e-01 -3.97252381e-01 -3.22864294e-01 -1.96718901e-01
-4.05854076e-01 -1.75518375e-02 2.62506187e-01 -4.12203759e-01
-5.41764677e-01 8.70871842e-01 4.19363454e-02 3.50625068e-02
-3.83223057e-01 -6.96123064e-01 -4.34381425e-01 -9.78028297e-01
8.45151767e-02 -1.89214036e-01 4.32879627e-01 1.02778770e-01
3.65246236e-02 3.17635894e-01 -1.75143754e+00 1.03426874e-01
8.66881013e-01 1.53192985e+00 1.61902219e-01 4.65456337e-01
-1.03181764e-01 9.35033262e-01 -6.77512586e-01 7.81957656e-02
2.52155513e-01 -3.19693714e-01 -3.72773200e-01 2.13019941e-02
1.11559458e-01 -4.16157782e-01 -5.93650281e-01 6.69733107e-01
5.99205554e-01 6.73201215e-03 2.51572043e-01 -1.56985164e+00
-4.60293964e-02 3.19587052e-01 -7.65723228e-01 3.25087756e-01
6.31726980e-01 1.60247460e-01 2.21360177e-01 -4.79143590e-01
3.62545885e-02 6.66864932e-01 7.35294104e-01 3.75769585e-01
-4.30571765e-01 -7.65735149e-01 -5.97267486e-02 8.70621443e-01
-1.42388499e+00 -4.92499620e-01 7.33048499e-01 -1.94342136e-01
9.36445653e-01 3.28280061e-01 9.40600336e-01 1.40876901e+00
1.65494025e-01 1.02881038e+00 9.34514523e-01 -3.30359370e-01
9.68315303e-02 -2.74479777e-01 -4.11341665e-03 4.48219121e-01
3.33025187e-01 -1.23010531e-01 -7.05830276e-01 1.32301047e-01
6.02577031e-01 4.41592634e-01 -3.78999829e-01 -1.09203532e-01
-1.58778262e+00 1.95326388e-01 2.23878384e-01 4.83024478e-01
-5.00791788e-01 2.48610616e-01 2.11732417e-01 -5.56258149e-02
4.63998392e-02 -7.51862600e-02 -4.99406308e-01 -5.84160089e-01
-9.56147611e-01 -2.08954841e-01 6.09665811e-01 7.86031425e-01
2.99946994e-01 4.60885733e-01 -3.03531945e-01 8.22188973e-01
4.45588171e-01 8.13392818e-01 6.53415442e-01 -9.51421678e-01
5.03372133e-01 5.50192654e-01 2.80718207e-01 -1.04415917e+00
-5.90598881e-01 -3.77811730e-01 -1.03454304e+00 -2.90560001e-03
4.79061842e-01 -2.74486750e-01 -6.52471006e-01 1.38363159e+00
2.84999877e-01 7.67278790e-01 -6.83534965e-02 6.94220483e-01
7.82397091e-01 6.96612060e-01 1.73593014e-01 -5.58342874e-01
1.17357683e+00 -1.03761172e+00 -9.81047153e-01 -2.28520364e-01
3.15879256e-01 -3.68763804e-01 7.72828162e-01 7.90561855e-01
-6.63546085e-01 -7.65411258e-01 -1.46811962e+00 4.94608998e-01
-2.44205475e-01 2.66767234e-01 7.09307551e-01 1.04694998e+00
-6.31380379e-01 3.32287490e-01 -1.11858773e+00 -7.12712526e-01
7.83150911e-01 4.70116198e-01 -3.82414222e-01 -1.37955710e-01
-1.01414526e+00 4.88532037e-01 1.82833448e-01 3.74030381e-01
-7.58575559e-01 -1.71621457e-01 -5.31316280e-01 1.44905252e-02
3.49637032e-01 -2.34882101e-01 1.01917005e+00 -1.05914187e+00
-1.47290409e+00 3.62622947e-01 5.78956772e-03 -5.43958724e-01
5.61295569e-01 -3.05283517e-01 -9.49237108e-01 1.36141032e-01
-8.09252933e-02 1.13416009e-01 6.45990610e-01 -5.34372449e-01
-8.46562803e-01 -4.21321690e-01 -1.59781754e-01 2.34828738e-04
-5.18606246e-01 -6.34110272e-02 -4.24887419e-01 -4.07554418e-01
3.08887273e-01 -7.50753284e-01 1.45754710e-01 2.20300823e-01
-2.86963433e-01 -1.83622807e-01 1.17775702e+00 -6.43134892e-01
1.30729866e+00 -2.03685689e+00 -4.81875807e-01 1.31872162e-01
-1.50698423e-01 5.17924845e-01 1.91334948e-01 1.00875050e-01
2.93088943e-01 -3.10340017e-01 -2.07228780e-01 4.30603027e-02
-2.51021504e-01 3.54340136e-01 3.48928690e-01 6.25977397e-01
-1.65734991e-01 6.36208832e-01 -7.30829895e-01 -5.65862834e-01
4.84784752e-01 4.35658574e-01 6.88029407e-03 4.57754061e-02
3.05501431e-01 6.59420669e-01 -5.21845996e-01 9.43463027e-01
6.96519077e-01 -1.84457049e-01 7.28630461e-03 -3.02013218e-01
-1.08919345e-01 -3.13014209e-01 -1.63343978e+00 1.95697689e+00
-3.06947410e-01 4.32994008e-01 4.84467931e-02 -1.33179295e+00
8.13786209e-01 4.34112668e-01 1.17694390e+00 -1.02046394e+00
1.83368251e-01 5.36870174e-02 -9.29792747e-02 -1.01635647e+00
1.51822332e-03 5.45340776e-01 -1.76546238e-02 1.94446653e-01
9.03202966e-03 7.11404800e-01 1.56333908e-01 -1.79429367e-01
1.40597701e+00 4.25751507e-01 3.75574857e-01 7.12080747e-02
7.81240940e-01 -2.44350269e-01 9.90568817e-01 7.49440730e-01
-6.75844073e-01 2.62013525e-01 -2.60904610e-01 -2.45197847e-01
-4.10620838e-01 -1.08349562e+00 -8.41128007e-02 9.15804386e-01
5.70429385e-01 -4.33013327e-02 -6.84080303e-01 -6.63065732e-01
-2.82987744e-01 2.75779992e-01 -2.39041001e-01 2.21552458e-02
-4.12222475e-01 -8.00306618e-01 9.24122810e-01 7.91343927e-01
1.45353162e+00 -1.10387313e+00 -8.80638361e-01 2.27075174e-01
-4.61474806e-01 -1.22827017e+00 1.37965754e-03 3.21692452e-02
-6.85955286e-01 -1.38519609e+00 -6.77660704e-01 -6.09185278e-01
2.62808025e-01 5.53985596e-01 4.39652354e-01 -1.83326676e-02
-3.29278141e-01 5.22671938e-01 -3.65928859e-01 -5.68673015e-01
2.23325193e-01 -7.16301203e-02 1.63078710e-01 4.27479714e-01
6.70713365e-01 -7.89809704e-01 -8.36960733e-01 6.34241760e-01
-5.23934066e-01 -2.14100286e-01 6.31309152e-01 4.06570941e-01
4.27857131e-01 1.44411445e-01 5.49875915e-01 -3.91123623e-01
4.41226400e-02 -6.11868680e-01 -3.65201592e-01 3.60706329e-01
-4.92858112e-01 -4.16665614e-01 3.85946989e-01 -4.11514908e-01
-1.23980498e+00 2.92452812e-01 -1.99353680e-01 -8.66461173e-02
-4.86057699e-01 2.80823380e-01 -6.64961278e-01 -8.47412050e-02
7.93701589e-01 3.33849669e-01 -2.01762408e-01 -5.00709236e-01
-1.90479800e-01 1.09680641e+00 6.44129097e-01 -1.62324324e-01
5.82476974e-01 6.39424145e-01 -6.35450780e-02 -1.21208334e+00
-5.28026938e-01 -5.91474652e-01 -4.62867975e-01 -6.58677995e-01
1.03362679e+00 -1.07127845e+00 -8.41420352e-01 1.13286150e+00
-8.40952814e-01 -7.26583302e-02 2.33522151e-02 8.65382075e-01
-2.78088570e-01 6.41215146e-01 -3.16349536e-01 -1.06232035e+00
-4.78194267e-01 -9.56629276e-01 7.25614488e-01 6.84814215e-01
3.82325165e-02 -3.97759795e-01 -2.39072576e-01 7.94330359e-01
6.45077109e-01 4.48796988e-01 1.53529257e-01 -3.21901828e-01
-5.36294937e-01 -3.01915914e-01 -3.77662629e-02 3.18266988e-01
3.01532835e-01 -7.50554577e-02 -1.15774131e+00 -9.63165518e-03
-1.12100523e-02 -1.33620918e-01 5.40048838e-01 5.17489552e-01
1.42022896e+00 -6.72795027e-02 -5.67514241e-01 6.94175363e-01
1.30621707e+00 7.77822793e-01 1.17291951e+00 4.40324128e-01
6.54355288e-01 5.31956293e-02 7.91111529e-01 6.44365132e-01
4.08663340e-02 6.73336089e-01 6.00981295e-01 -1.57783404e-02
7.14059696e-02 -1.71355233e-02 4.88118559e-01 4.35660809e-01
-4.10796225e-01 -3.27038407e-01 -7.68771768e-01 2.88333803e-01
-2.07954335e+00 -1.39143682e+00 -4.17933613e-01 2.27292347e+00
2.40720823e-01 2.15590775e-01 2.47769669e-01 8.38197351e-01
6.04770184e-01 2.72532076e-01 -6.85587347e-01 2.07730800e-01
-5.65855317e-02 2.91247547e-01 6.73789680e-01 -1.99537262e-01
-1.41290998e+00 3.79847765e-01 5.53121805e+00 7.73319483e-01
-1.17018294e+00 2.55946636e-01 1.15646847e-01 -1.45446897e-01
4.21850860e-01 -4.57829177e-01 -5.13722241e-01 8.64589572e-01
8.45193088e-01 4.11479503e-01 2.83416778e-01 1.10721457e+00
6.47734582e-01 -6.57105744e-01 -5.65491617e-01 1.48520911e+00
3.53086740e-02 -1.07845247e+00 -4.43828434e-01 -1.41364172e-01
3.57452720e-01 1.32641494e-01 -3.58423561e-01 2.49420390e-01
-2.40526170e-01 -7.82284915e-01 5.43073714e-01 7.63347566e-01
8.50496590e-01 -6.06929600e-01 8.19951117e-01 4.70101327e-01
-1.47134793e+00 -3.52913946e-01 7.05848634e-02 -2.00379044e-01
1.80002034e-01 4.33265626e-01 -6.01681918e-02 5.99579334e-01
1.12624538e+00 9.32610869e-01 -4.34510976e-01 1.22278333e+00
-9.87302065e-02 8.69930029e-01 -4.61125642e-01 -9.27729979e-02
-1.79443568e-01 -8.88344646e-02 2.06331417e-01 1.05049670e+00
5.74041843e-01 1.32066458e-01 3.03512216e-01 1.67491794e-01
2.19362900e-01 -1.20325491e-01 -5.16256094e-01 4.41369936e-02
3.37311625e-01 1.16920781e+00 -7.31369674e-01 -6.00498170e-02
-6.95712507e-01 9.84788656e-01 -3.77956867e-01 2.80306578e-01
-1.18245447e+00 -5.68716347e-01 5.94565809e-01 1.48566842e-01
1.33180156e-01 -4.47630197e-01 -1.33789569e-01 -1.15644455e+00
7.43368715e-02 -5.27743101e-01 4.11480397e-01 -8.51983905e-01
-6.45075619e-01 8.78376141e-02 -1.75480351e-01 -1.50310361e+00
1.76042691e-01 -5.44860244e-01 -6.69033408e-01 3.86053681e-01
-1.34188819e+00 -1.05263448e+00 -1.14021075e+00 1.08856523e+00
7.16235876e-01 -2.14650691e-01 6.36915863e-01 8.19220424e-01
-9.73824501e-01 6.23599410e-01 1.56232744e-01 3.71460021e-01
4.82211083e-01 -6.20019317e-01 -3.33559185e-01 1.19807005e+00
2.00604409e-01 1.65439576e-01 3.22039127e-01 -3.57834905e-01
-1.46667039e+00 -8.95712376e-01 4.86804068e-01 -9.29447189e-02
2.27038682e-01 -2.93837115e-02 -4.84092981e-01 3.90113622e-01
4.00431864e-02 2.65565127e-01 7.84882069e-01 -3.55186760e-01
1.52698934e-01 -7.29401171e-01 -1.35426331e+00 1.46655515e-01
1.41448689e+00 -1.79013729e-01 -1.90873399e-01 1.81918919e-01
7.60089308e-02 -1.65744811e-01 -7.32874453e-01 4.88292426e-01
9.65855479e-01 -1.03639901e+00 8.70001316e-01 -2.54781187e-01
-4.93522435e-02 -7.27611423e-01 -4.64991003e-01 -6.13954425e-01
-2.47873619e-01 -4.70323741e-01 -4.43582058e-01 1.13374043e+00
4.58890386e-02 -4.98675525e-01 1.04068041e+00 3.57155532e-01
-1.12862378e-01 -3.93823326e-01 -1.10727179e+00 -8.17035854e-01
-1.04422855e+00 -8.64382625e-01 4.35828090e-01 6.85026705e-01
3.73874828e-02 6.07938394e-02 -6.38792336e-01 1.00066334e-01
5.38755476e-01 -3.72717142e-01 8.08706582e-01 -1.33920145e+00
-2.30818614e-01 7.30587691e-02 -7.84885705e-01 -1.14338517e+00
-3.29735368e-01 -3.50005060e-01 -2.87377425e-02 -1.86204600e+00
-1.21567994e-01 -2.36124709e-01 -7.67003357e-01 6.12914681e-01
3.55147123e-01 3.20440352e-01 -6.46616593e-02 1.54732555e-01
-8.25101793e-01 4.75447834e-01 9.17160392e-01 -5.56201696e-01
-1.16170667e-01 4.54838872e-01 -4.06271189e-01 7.65230119e-01
8.47009301e-01 -2.21252859e-01 -5.28007686e-01 -4.97183353e-01
-7.27575198e-02 9.25304443e-02 3.07582825e-01 -1.88385367e+00
3.53975236e-01 -2.77049303e-01 6.83160365e-01 -5.35656571e-01
4.01541114e-01 -1.23310363e+00 3.01888168e-01 6.55701935e-01
8.03544968e-02 -3.69885087e-01 -1.07561409e-01 8.37346733e-01
-3.75047624e-02 2.16748208e-01 6.98517144e-01 2.00102068e-02
-1.18418860e+00 3.74387294e-01 -6.12664759e-01 -3.12585652e-01
1.24602830e+00 -8.22446227e-01 -3.20627123e-01 -3.21153045e-01
-6.85541809e-01 1.59473106e-01 -1.07506096e-01 4.73359257e-01
5.09952366e-01 -1.37188101e+00 -3.04931283e-01 2.68090218e-01
3.23231459e-01 -3.02812755e-01 5.50478816e-01 1.29661775e+00
-4.05734569e-01 2.42580175e-01 -5.05359828e-01 -8.08472037e-01
-1.30590785e+00 1.68410912e-01 4.11952436e-01 6.38931394e-02
-3.02393466e-01 4.89540964e-01 -5.57130873e-01 1.30435035e-01
6.53521061e-01 -3.36121172e-01 -5.32292068e-01 -4.30009514e-02
7.01315105e-01 8.82414758e-01 2.01627146e-02 -6.56435847e-01
-6.28493786e-01 5.48461378e-01 4.09553468e-01 2.25991160e-01
1.10672688e+00 -1.51969731e-01 4.45694327e-01 2.63131976e-01
7.35675216e-01 -2.65767545e-01 -1.39191854e+00 -1.26737356e-01
-5.66055393e-03 -4.45665479e-01 5.65151162e-02 -8.61989260e-01
-1.18355906e+00 8.13718975e-01 1.33223736e+00 1.05872661e-01
1.28082442e+00 -5.22091866e-01 9.29542422e-01 4.95042652e-01
8.68907809e-01 -1.49938869e+00 -1.26375407e-02 1.38083056e-01
3.41131717e-01 -1.26826811e+00 -4.40089591e-02 -1.50862068e-01
-3.73326629e-01 9.85571861e-01 6.46476090e-01 1.40510455e-01
7.89507627e-01 1.85396537e-01 3.61862816e-02 1.38964146e-01
-4.67670783e-02 -3.33935052e-01 -2.13085726e-01 1.08895206e+00
1.31989077e-01 -1.81787908e-01 -4.49457794e-01 6.48678958e-01
3.57400775e-01 5.39398313e-01 2.05657482e-01 1.09424245e+00
-5.21182537e-01 -6.20752275e-01 -4.05542463e-01 6.81262732e-01
-5.34586370e-01 5.72602272e-01 1.47243172e-01 5.86006463e-01
7.24445879e-01 1.57524002e+00 -2.47218058e-01 -6.29278421e-01
3.99783343e-01 2.16662404e-04 4.50319558e-01 -1.50800049e-02
-2.15904355e-01 -1.12774141e-01 1.58389807e-01 -1.00989699e+00
-1.07304776e+00 -6.81967437e-01 -1.26634383e+00 -1.33067057e-01
-2.64273882e-01 -7.53613636e-02 7.60939538e-01 1.29647386e+00
3.12801868e-01 5.88998199e-01 5.84746599e-01 -7.47105300e-01
-1.88124944e-02 -9.31286454e-01 -8.35843265e-01 2.34315261e-01
-5.88268898e-02 -8.21511149e-01 -1.94272220e-01 1.06468402e-01] | [7.181371688842773, 0.6580095887184143] |
a31357c8-130d-4099-aafb-28efed428f10 | particle-filtering-for-plca-model-with | 1703.09772 | null | http://arxiv.org/abs/1703.09772v1 | http://arxiv.org/pdf/1703.09772v1.pdf | Particle Filtering for PLCA model with Application to Music Transcription | Automatic Music Transcription (AMT) consists in automatically estimating the
notes in an audio recording, through three attributes: onset time, duration and
pitch. Probabilistic Latent Component Analysis (PLCA) has become very popular
for this task. PLCA is a spectrogram factorization method, able to model a
magnitude spectrogram as a linear combination of spectral vectors from a
dictionary. Such methods use the Expectation-Maximization (EM) algorithm to
estimate the parameters of the acoustic model. This algorithm presents
well-known inherent defaults (local convergence, initialization dependency),
making EM-based systems limited in their applications to AMT, particularly in
regards to the mathematical form and number of priors. To overcome such limits,
we propose in this paper to employ a different estimation framework based on
Particle Filtering (PF), which consists in sampling the posterior distribution
over larger parameter ranges. This framework proves to be more robust in
parameter estimation, more flexible and unifying in the integration of prior
knowledge in the system. Note-level transcription accuracies of 61.8 $\%$ and
59.5 $\%$ were achieved on evaluation sound datasets of two different
instrument repertoires, including the classical piano (from MAPS dataset) and
the marovany zither, and direct comparisons to previous PLCA-based approaches
are provided. Steps for further development are also outlined. | ['D. Cazau', 'W. Yuancheng', 'O. Adam', 'G. Revillon'] | 2017-03-28 | null | null | null | null | ['music-transcription'] | ['music'] | [ 2.02784374e-01 -2.87525117e-01 2.71022201e-01 1.12700164e-01
-9.79412615e-01 -7.46580005e-01 5.53459585e-01 1.30448356e-01
-4.61721957e-01 6.65029943e-01 2.88869619e-01 7.24224374e-02
-6.18767440e-01 -4.57735270e-01 -1.90447971e-01 -9.54888284e-01
-8.03195536e-02 5.49378872e-01 1.25714503e-02 -7.50233904e-02
2.33356833e-01 4.13382888e-01 -1.87792587e+00 -4.81814332e-03
6.33145273e-01 8.08606327e-01 4.40990120e-01 7.81138182e-01
-5.69967143e-02 3.65239203e-01 -9.02039051e-01 -2.07436562e-01
-9.90387276e-02 -4.68732506e-01 -4.15053457e-01 3.11910510e-02
-2.84540541e-02 2.19134331e-01 3.57825398e-01 7.68852353e-01
6.50440037e-01 3.12095404e-01 9.02605176e-01 -7.81076908e-01
3.23254541e-02 7.20523179e-01 -7.75289088e-02 3.71067747e-02
6.01219594e-01 -1.24874532e-01 9.84623849e-01 -7.85053194e-01
3.77597123e-01 9.91469443e-01 1.04002476e+00 1.01476558e-01
-1.44489491e+00 -5.66572726e-01 -2.84950316e-01 2.50426710e-01
-1.66787469e+00 -5.23915350e-01 8.81044269e-01 -6.71106398e-01
8.76207411e-01 5.07141829e-01 6.87902808e-01 1.07352066e+00
1.66945662e-02 5.05026877e-01 1.31054711e+00 -8.19991291e-01
4.72246945e-01 1.34343088e-01 -9.91961956e-02 6.28014654e-02
-2.31870040e-01 -6.30807951e-02 -7.10853696e-01 -4.94614989e-01
6.52889550e-01 -5.00532925e-01 -3.00833702e-01 -1.55696467e-01
-1.14547634e+00 7.65745044e-01 -4.10906076e-01 6.25768661e-01
-7.54711092e-01 8.68516788e-02 3.42330456e-01 4.76310588e-02
3.40573519e-01 5.45211434e-01 -1.87196940e-01 -7.54283369e-01
-1.31243813e+00 5.07173955e-01 9.37876344e-01 4.47741121e-01
3.88048410e-01 3.41171533e-01 -1.76014230e-01 9.68570232e-01
4.96242970e-01 5.99385440e-01 6.36311233e-01 -8.96540165e-01
1.67209849e-01 1.78674787e-01 2.18534738e-01 -1.02095771e+00
-3.72748762e-01 -5.44015408e-01 -6.08465731e-01 1.58753514e-01
4.65793371e-01 -1.17365025e-01 -5.30780911e-01 1.70857286e+00
2.63362169e-01 3.06124419e-01 -4.25648689e-02 7.44307339e-01
1.90196455e-01 7.41251886e-01 -6.16915226e-02 -6.55964851e-01
1.33389127e+00 -3.31785440e-01 -1.00751388e+00 2.97914743e-01
1.68727115e-02 -1.35515797e+00 9.38618839e-01 1.10840917e+00
-1.09009647e+00 -1.01651275e+00 -9.56694007e-01 5.21885693e-01
-7.77328163e-02 5.15705645e-01 3.97673875e-01 1.10831523e+00
-7.30288148e-01 7.52438784e-01 -8.06741834e-01 -3.12383031e-03
-2.04054445e-01 4.56734836e-01 -5.65666705e-02 5.50274372e-01
-9.85672891e-01 5.88584363e-01 4.32029575e-01 2.60394663e-01
-5.74115992e-01 -6.14050627e-01 -4.98340845e-01 1.29211172e-01
8.00122246e-02 -4.29171413e-01 1.17839587e+00 -7.63877094e-01
-2.17536020e+00 3.78683686e-01 -1.32872358e-01 -5.57747424e-01
4.60010588e-01 -3.25640738e-01 -5.22916973e-01 1.49736881e-01
-1.61959171e-01 1.27713680e-01 1.34694469e+00 -1.06486714e+00
-3.93881321e-01 -9.28308368e-02 -4.04523313e-01 -1.46640306e-02
-1.09088659e-01 2.83317920e-02 -2.72910684e-01 -1.00725353e+00
2.09828049e-01 -1.19407535e+00 8.48628506e-02 -7.12106049e-01
-2.24939734e-01 -1.87153518e-01 4.60559130e-01 -8.83867800e-01
1.63769054e+00 -2.30465603e+00 3.92982423e-01 3.72445703e-01
-2.60569453e-01 3.62612605e-01 3.38619798e-01 7.58892298e-01
-5.78335673e-02 -3.85413975e-01 -2.48948991e-01 -6.33329570e-01
3.00100267e-01 1.62894711e-01 -4.72110361e-01 4.28702921e-01
-2.59849578e-01 3.07113200e-01 -5.99869311e-01 -3.83766592e-01
5.20178854e-01 7.94237375e-01 -3.90249819e-01 1.48890883e-01
-1.80859342e-01 5.77444732e-01 -3.16253118e-02 2.56041080e-01
5.17887473e-01 4.32114065e-01 2.96078444e-01 -2.21885994e-01
-4.22711045e-01 2.80577987e-01 -1.88924587e+00 1.85704827e+00
-6.65150523e-01 3.85266036e-01 2.47554168e-01 -6.71450615e-01
1.17975843e+00 7.68033862e-01 5.70487440e-01 1.20309740e-02
2.24713489e-01 4.07776475e-01 2.08345607e-01 -5.53273022e-01
4.79454368e-01 -3.15466613e-01 1.01007886e-01 3.44318599e-01
3.70605081e-01 -3.47992331e-01 3.88264716e-01 -3.15823436e-01
5.65388858e-01 3.40704143e-01 3.18660051e-01 -4.13126051e-01
7.21105993e-01 -4.04994637e-01 3.62353593e-01 5.02512097e-01
1.40631855e-01 7.09986985e-01 1.98078424e-01 1.09489202e-01
-8.25700343e-01 -1.00668335e+00 -2.37877309e-01 7.34038949e-01
-5.61912060e-01 -7.27016866e-01 -9.00777996e-01 1.21844951e-02
-2.40854129e-01 7.25543320e-01 -3.20020616e-01 1.42364174e-01
-5.08748353e-01 -8.48937988e-01 5.25251269e-01 1.35122865e-01
-3.08622830e-02 -1.09238088e+00 -5.72732687e-01 7.16950834e-01
-3.67266655e-01 -8.37078452e-01 -1.03259496e-01 2.91447520e-01
-8.07297468e-01 -5.90761721e-01 -7.14091003e-01 -2.18296543e-01
-2.26596426e-02 -3.53278846e-01 8.15898478e-01 -4.38772023e-01
-2.41619781e-01 6.14729762e-01 -6.23783052e-01 -6.04670107e-01
-6.33050919e-01 4.11737040e-02 3.78062069e-01 4.66322780e-01
1.07771665e-01 -1.09882617e+00 -3.86069924e-01 2.31483042e-01
-8.16838682e-01 -2.15687960e-01 4.27358449e-01 5.66131830e-01
5.83640277e-01 2.60916919e-01 4.84094292e-01 -6.50349557e-01
8.47608447e-01 -1.29366443e-01 -5.34075439e-01 -1.18422076e-01
-4.75919276e-01 -1.74333751e-01 6.08564436e-01 -6.07158124e-01
-1.07513976e+00 2.72269994e-01 -6.64404035e-01 -5.00649035e-01
-3.99859786e-01 3.68414551e-01 -2.96854585e-01 9.76993665e-02
7.32123733e-01 3.42124552e-01 -1.94200158e-01 -8.57249737e-01
1.52860865e-01 7.07191110e-01 6.62713051e-01 -7.05935001e-01
6.44818008e-01 2.74405658e-01 -4.57775109e-02 -1.12372029e+00
-4.16291058e-01 -7.15711832e-01 -7.23203182e-01 -4.13851917e-01
6.98856235e-01 -6.91970825e-01 -7.82947719e-01 4.38892752e-01
-9.41414118e-01 1.40981870e-02 -4.98119175e-01 1.10532737e+00
-8.04390490e-01 4.39114422e-01 -4.18401688e-01 -1.25238323e+00
-2.67175198e-01 -1.07184958e+00 9.40971613e-01 3.66552323e-02
-6.90636039e-01 -8.26282144e-01 4.56348985e-01 3.50742728e-01
3.08173954e-01 1.91784754e-01 7.27871895e-01 -3.06710094e-01
-8.41546208e-02 -2.05475301e-01 5.37430406e-01 5.96680522e-01
1.27090335e-01 1.32901117e-01 -1.34566164e+00 -2.64786005e-01
2.90231228e-01 1.50842994e-01 5.73613584e-01 5.30412912e-01
7.75361240e-01 3.21622789e-02 4.31210771e-02 3.30270529e-01
1.38925910e+00 2.62966931e-01 5.96749663e-01 5.95687330e-02
3.43936533e-01 4.56606328e-01 7.08821774e-01 7.86730468e-01
-1.88516676e-01 1.03713703e+00 1.33105770e-01 3.67363811e-01
-2.04444721e-01 -2.17628449e-01 4.55480129e-01 1.23919272e+00
-5.10722101e-01 -1.72344282e-01 -7.16480017e-01 4.27473009e-01
-1.64531267e+00 -8.55903149e-01 -3.03304106e-01 2.36922073e+00
6.76727712e-01 2.58967340e-01 4.56548393e-01 9.22781289e-01
4.97821480e-01 -5.38114309e-02 3.29227857e-02 -4.60263312e-01
-5.98137379e-02 6.27427697e-01 2.61079222e-01 5.05392432e-01
-1.01893723e+00 5.37777066e-01 5.85954857e+00 1.31490839e+00
-9.80570853e-01 2.44742811e-01 -1.17550530e-01 8.04241002e-02
-1.15642138e-01 4.66227392e-03 -7.86300182e-01 6.61415637e-01
1.23845422e+00 2.74238348e-01 4.41802531e-01 4.69156116e-01
5.16276777e-01 -2.87390828e-01 -6.70815289e-01 1.22919106e+00
1.14207365e-01 -1.11694455e+00 -1.58510819e-01 -4.03333269e-02
4.29199934e-01 -3.37602496e-01 1.65404066e-01 1.71985120e-01
-4.39222574e-01 -7.59532809e-01 1.01628089e+00 7.37773657e-01
5.38596451e-01 -9.35625613e-01 7.15404749e-01 4.77078497e-01
-1.11008811e+00 -1.74554586e-02 -1.07601710e-01 -2.37972260e-01
3.50452662e-01 7.72162914e-01 -1.01636600e+00 7.18412399e-01
6.52227223e-01 1.95616171e-01 -2.34729961e-01 1.14211905e+00
-2.78645754e-01 1.13866246e+00 -5.69348812e-01 8.58898684e-02
-1.37443095e-01 -4.73059028e-01 9.67492521e-01 1.46657658e+00
5.42157590e-01 -2.40882531e-01 -1.04383610e-01 8.47068548e-01
5.17296731e-01 4.73422945e-01 4.07458749e-03 -8.41035470e-02
4.57530409e-01 1.31020045e+00 -1.04386473e+00 -4.37386259e-02
9.19728950e-02 7.48440385e-01 -2.60423541e-01 1.22335799e-01
-7.21211135e-01 -1.79982901e-01 3.80100191e-01 5.19189537e-02
5.60902476e-01 -4.85020578e-01 3.54454364e-03 -7.71875858e-01
-1.19460486e-01 -9.72130299e-01 1.35749578e-01 -6.19993269e-01
-1.03336418e+00 6.80117726e-01 1.84027433e-01 -1.44778621e+00
-6.94535553e-01 -5.45656562e-01 -3.38839918e-01 1.10152996e+00
-9.30395126e-01 -1.05283237e+00 2.70154238e-01 4.93031681e-01
4.92158502e-01 -2.01353118e-01 1.28403580e+00 4.63336140e-01
-3.02712500e-01 2.52804756e-01 2.48060450e-01 -3.51311743e-01
6.10647738e-01 -1.40842450e+00 9.07622930e-03 7.14437604e-01
7.64470398e-01 6.21962070e-01 1.20554173e+00 -3.14382046e-01
-1.17542899e+00 -4.55165356e-01 8.77438486e-01 -3.70051652e-01
6.17946148e-01 -4.89782661e-01 -7.85694242e-01 1.95436835e-01
1.87954992e-01 -6.12464488e-01 1.02688885e+00 3.08272421e-01
2.73564626e-02 5.83486538e-03 -7.85807610e-01 1.95352197e-01
3.52730036e-01 -6.50520146e-01 -6.98589683e-01 -1.13084644e-01
1.95217177e-01 -3.00930738e-01 -1.17135262e+00 2.66875386e-01
6.54607058e-01 -1.19828546e+00 9.82354343e-01 2.17664152e-01
-3.44719529e-01 -6.53022170e-01 -1.84191257e-01 -1.24453104e+00
-2.52375424e-01 -1.14495230e+00 -1.64698094e-01 1.65597951e+00
2.14215040e-01 -4.18211251e-01 4.98260379e-01 5.31460438e-03
-1.00508548e-01 -3.81987870e-01 -1.08259308e+00 -7.46497929e-01
-3.70379895e-01 -9.76971030e-01 3.17983896e-01 7.03969836e-01
-8.53157118e-02 4.09602404e-01 -5.91503799e-01 1.75772592e-01
5.11222899e-01 2.09236279e-01 6.95304871e-01 -1.40059698e+00
-8.40981722e-01 -4.08906519e-01 -3.45965654e-01 -6.21708274e-01
2.39683986e-02 -5.04464865e-01 3.79105136e-02 -9.58924890e-01
-3.34839910e-01 -3.44668955e-01 -3.64220917e-01 1.94881279e-02
1.21898621e-01 3.72572720e-01 3.70571822e-01 1.63787112e-01
-1.24284521e-01 4.18324202e-01 7.71749437e-01 1.91534638e-01
-4.37370777e-01 6.10885262e-01 -7.52453208e-02 8.92042279e-01
7.41726398e-01 -6.01038516e-01 -4.04903203e-01 1.28850684e-01
3.01678240e-01 1.92499131e-01 1.32799998e-01 -1.38851619e+00
1.15759835e-01 2.71343619e-01 2.44455785e-01 -7.10092545e-01
8.02229047e-01 -7.73739159e-01 7.59718180e-01 3.06428134e-01
-7.87358135e-02 -5.87174892e-02 3.90352964e-01 4.78544354e-01
-4.04386461e-01 -6.07506216e-01 6.13890827e-01 1.06409118e-02
-4.51662779e-01 -2.65118569e-01 -7.59186149e-01 -3.84411097e-01
5.33358097e-01 -2.49709487e-01 5.10794520e-01 -3.89015824e-01
-1.20042908e+00 -6.05538487e-01 1.91115662e-02 1.65975213e-01
2.52313942e-01 -1.04997349e+00 -7.00759530e-01 1.79751456e-01
-9.93846729e-02 -4.35930818e-01 7.38596857e-01 1.03814447e+00
-3.78060102e-01 4.38674986e-01 -2.26272359e-01 -6.20703042e-01
-1.39573896e+00 4.61470336e-01 1.24622285e-01 -2.50486732e-01
-3.53474915e-01 8.24767232e-01 -3.41177404e-01 -2.22613677e-01
2.74908453e-01 -3.92694861e-01 -5.14854908e-01 4.44727302e-01
4.03281063e-01 5.42099416e-01 2.69178569e-01 -8.53662908e-01
-1.77897304e-01 6.53212070e-01 5.68676889e-01 -5.98563612e-01
1.17200708e+00 -2.45821565e-01 -8.40776116e-02 9.17146266e-01
8.85661542e-01 6.93082452e-01 -9.63336408e-01 4.56405953e-02
1.33961558e-01 -2.48406574e-01 2.29053527e-01 -8.09362710e-01
-3.21780175e-01 8.11123252e-01 8.03613365e-01 4.40422475e-01
1.22898531e+00 -3.87024462e-01 4.79603887e-01 -4.29203995e-02
3.49163353e-01 -1.15091932e+00 -3.59865189e-01 2.13359609e-01
8.12207401e-01 -5.67346334e-01 3.24889533e-02 -3.92314821e-01
-3.67685676e-01 1.25955915e+00 -1.94720134e-01 -1.86412692e-01
7.07702219e-01 3.66903126e-01 8.97736698e-02 1.08780257e-01
-4.81248081e-01 -3.01135123e-01 4.45651889e-01 4.73789781e-01
5.20449162e-01 2.63099998e-01 -5.05284250e-01 8.67470562e-01
-6.21500254e-01 -5.46818078e-02 2.55058050e-01 6.37520730e-01
-1.51074246e-01 -1.49110365e+00 -9.07251000e-01 -4.94349515e-03
-7.35221267e-01 -6.16376810e-02 -9.88560393e-02 6.04964197e-01
5.70181370e-01 1.08320284e+00 -1.86693460e-01 -3.50157171e-01
3.58542770e-01 3.48732382e-01 7.60155737e-01 -4.60691839e-01
-7.96650589e-01 8.06110561e-01 2.57598683e-02 -2.17205882e-01
-6.90030396e-01 -1.15867925e+00 -8.70419681e-01 1.42806187e-01
-4.32495058e-01 6.22398555e-01 1.07289433e+00 8.56625617e-01
1.17161172e-02 6.84145331e-01 3.67354512e-01 -1.12442374e+00
-4.72879797e-01 -1.24667096e+00 -9.07653809e-01 1.68844357e-01
1.70178652e-01 -6.26012802e-01 -3.72955948e-01 2.98373759e-01] | [15.722349166870117, 5.452995300292969] |
54ce921e-2c56-47e1-bcbd-83b92058467b | decomposed-human-motion-prior-for-video-pose | 2305.18743 | null | https://arxiv.org/abs/2305.18743v2 | https://arxiv.org/pdf/2305.18743v2.pdf | Decomposed Human Motion Prior for Video Pose Estimation via Adversarial Training | Estimating human pose from video is a task that receives considerable attention due to its applicability in numerous 3D fields. The complexity of prior knowledge of human body movements poses a challenge to neural network models in the task of regressing keypoints. In this paper, we address this problem by incorporating motion prior in an adversarial way. Different from previous methods, we propose to decompose holistic motion prior to joint motion prior, making it easier for neural networks to learn from prior knowledge thereby boosting the performance on the task. We also utilize a novel regularization loss to balance accuracy and smoothness introduced by motion prior. Our method achieves 9\% lower PA-MPJPE and 29\% lower acceleration error than previous methods tested on 3DPW. The estimator proves its robustness by achieving impressive performance on in-the-wild dataset. | ['Kai Zhang', 'Weixi Gu', 'Zhaoyu Zheng', 'Zhengdi Yu', 'Xiang Zhou', 'Wenshuo Chen'] | 2023-05-30 | null | null | null | null | ['pose-estimation'] | ['computer-vision'] | [ 1.05544589e-01 1.35086421e-02 -2.70464242e-01 -1.25978798e-01
-5.20695269e-01 -2.14848325e-01 4.56701785e-01 -4.45909798e-01
-9.38568115e-01 6.48706853e-01 3.11304480e-01 1.40043646e-01
2.51309633e-01 -2.84835726e-01 -1.03667307e+00 -4.60814923e-01
-3.30050960e-02 3.67232151e-02 2.93753982e-01 -2.95389682e-01
-6.15162030e-02 4.40898746e-01 -1.13634312e+00 -1.56785697e-01
3.64954352e-01 8.52234423e-01 7.06306770e-02 7.14266121e-01
7.23619401e-01 8.25457275e-01 -5.32456815e-01 -3.71845901e-01
6.25908017e-01 -2.74880379e-01 -5.64266443e-01 4.05883081e-02
9.20663714e-01 -5.71308851e-01 -6.90983176e-01 8.90213668e-01
5.92717409e-01 3.54827017e-01 6.84441686e-01 -1.18024290e+00
-1.91877484e-01 -1.25150189e-01 -7.81784773e-01 3.86337101e-01
4.99455392e-01 2.85779744e-01 8.42920840e-01 -7.24398911e-01
7.60137200e-01 1.11717904e+00 1.11249018e+00 7.01997995e-01
-1.25632417e+00 -4.97079372e-01 1.88402310e-01 1.84158608e-01
-1.21950710e+00 -4.46343333e-01 8.12727392e-01 -5.16262889e-01
8.71585429e-01 -7.28156790e-02 6.76075876e-01 1.37194288e+00
3.78102779e-01 9.52701271e-01 6.53775275e-01 -2.76071072e-01
-1.15331840e-02 -4.17582005e-01 -2.86743224e-01 7.17823625e-01
2.48496667e-01 9.22746956e-02 -6.43918753e-01 2.47708991e-01
1.06734920e+00 -2.21319832e-02 -3.91525537e-01 -7.20041335e-01
-1.26248169e+00 6.41633570e-01 5.35324633e-01 -1.21732458e-01
-3.84484649e-01 5.58058858e-01 2.46663317e-01 -1.87124405e-02
4.93783563e-01 1.51272461e-01 -4.43511695e-01 -1.48206845e-01
-1.00563276e+00 6.69333458e-01 6.67393923e-01 8.09989333e-01
2.39790618e-01 2.86551207e-01 -2.05399897e-02 6.11318886e-01
3.71284217e-01 6.48620427e-01 3.13570052e-01 -1.11316228e+00
7.44705915e-01 1.76702097e-01 2.94534236e-01 -1.33001709e+00
-5.48989654e-01 -4.05201942e-01 -9.19808209e-01 4.38408822e-01
7.61437416e-01 -2.00758725e-01 -1.03928518e+00 2.02330232e+00
4.52532023e-01 1.09484173e-01 -3.89182955e-01 1.17056119e+00
5.10811925e-01 4.13987905e-01 1.09744005e-01 1.71434686e-01
1.18340039e+00 -1.07311416e+00 -5.31119406e-01 -5.96944988e-01
3.20229858e-01 -5.03103197e-01 8.06101084e-01 5.31296372e-01
-1.08680153e+00 -8.01480591e-01 -1.09486330e+00 -1.14521869e-01
1.38554513e-01 1.49393052e-01 4.26969856e-01 4.54020917e-01
-7.92106569e-01 8.40646744e-01 -1.31689894e+00 -3.54107171e-01
3.45392078e-01 6.24362111e-01 -6.20784879e-01 1.85169473e-01
-9.84823823e-01 1.04761493e+00 2.73882359e-01 2.80497134e-01
-7.48344362e-01 -4.98304516e-01 -1.13723159e+00 -1.79706395e-01
4.74548340e-01 -1.04849315e+00 1.26987791e+00 -7.01990724e-01
-1.81845295e+00 6.20045781e-01 -7.49037415e-02 -6.95643604e-01
9.98718262e-01 -9.20185089e-01 1.50175989e-02 1.90235376e-01
-6.12454209e-03 9.17019904e-01 1.19347370e+00 -7.81828880e-01
-4.86037016e-01 -2.76772887e-01 6.07819296e-02 2.65276551e-01
-2.32835025e-01 -4.03962851e-01 -7.67299831e-01 -1.04823685e+00
5.39158098e-02 -1.33860385e+00 -3.90138209e-01 2.57432699e-01
-1.56606942e-01 3.02736321e-03 6.28497183e-01 -9.41596448e-01
9.16292608e-01 -1.81539845e+00 4.67488706e-01 -5.46758026e-02
7.72586912e-02 4.16377574e-01 4.61753011e-02 7.49975815e-02
1.08292043e-01 -2.97694087e-01 -1.70355469e-01 -4.38691676e-01
2.27478370e-02 2.48636648e-01 -1.61887750e-01 8.39364409e-01
4.46252584e-01 9.07694578e-01 -7.15477407e-01 -3.31621438e-01
2.58046597e-01 6.88805282e-01 -9.81130898e-01 2.15764061e-01
-5.05104400e-02 7.15287805e-01 -2.76142985e-01 4.29009497e-01
4.31495160e-01 -1.59522399e-01 -7.02009574e-02 -2.48621300e-01
3.05869609e-01 -8.63023996e-02 -1.30666912e+00 2.11921525e+00
-2.90144384e-01 6.56173229e-01 1.24781139e-01 -1.10124254e+00
5.68320155e-01 3.40622425e-01 5.49581468e-01 -4.58003104e-01
2.95169413e-01 -1.66941360e-01 -6.71195053e-03 -5.07064760e-01
5.54755628e-01 -1.71593517e-01 -1.73483729e-01 2.37567793e-03
1.32180944e-01 5.20899743e-02 5.09180240e-02 -1.91650897e-01
1.08184946e+00 9.14684594e-01 2.81483293e-01 -1.54394865e-01
4.72964466e-01 -1.09916985e-01 7.85518408e-01 5.78733623e-01
-4.45317924e-01 8.04762602e-01 6.30842596e-02 -6.30033016e-01
-9.57302749e-01 -1.10244846e+00 1.55435696e-01 9.27654028e-01
1.00077696e-01 -2.36320496e-01 -6.71585500e-01 -8.10539842e-01
1.35812834e-01 5.99879287e-02 -6.58549249e-01 -1.45919666e-01
-1.07849145e+00 -5.20311415e-01 5.90180933e-01 8.95047665e-01
5.96091092e-01 -6.88087344e-01 -9.45760667e-01 2.58383274e-01
-3.64697993e-01 -1.37036467e+00 -6.55506730e-01 -9.19907987e-02
-9.44163740e-01 -9.62924004e-01 -1.15100527e+00 -5.83243251e-01
5.90402246e-01 -5.07511012e-02 1.05242240e+00 -3.51717502e-01
-4.06371504e-01 6.45426929e-01 -1.60920352e-01 -3.93141598e-01
-8.00967216e-02 1.23343751e-01 3.90339792e-01 -1.33622944e-01
1.22012608e-01 -5.18156111e-01 -8.33592832e-01 3.90359670e-01
-7.31213391e-01 -8.07140321e-02 5.04330039e-01 8.80620956e-01
5.16017914e-01 -2.49390289e-01 1.58619866e-01 -3.91585022e-01
2.64366418e-01 -2.20865116e-01 -5.18941820e-01 -2.20455408e-01
-2.32347101e-01 1.38468847e-01 4.84534711e-01 -6.66716039e-01
-8.35239768e-01 5.11817276e-01 -3.21473032e-01 -6.69404030e-01
-1.91358492e-01 1.24659151e-01 -8.66015777e-02 -2.30773181e-01
6.48847580e-01 -9.61213186e-02 2.17367560e-01 -5.04955292e-01
3.58319044e-01 -6.96288645e-02 9.34138358e-01 -4.92444783e-01
1.00149512e+00 6.32633388e-01 3.27044338e-01 -8.08332980e-01
-8.44108284e-01 -4.12527174e-01 -9.27636981e-01 -2.19797030e-01
1.18221033e+00 -1.16437364e+00 -7.73497224e-01 3.80773246e-01
-1.12515628e+00 -2.18426704e-01 -9.71675068e-02 7.96417236e-01
-7.56474435e-01 5.94109893e-01 -6.76310062e-01 -5.29825568e-01
-3.18607271e-01 -1.08105445e+00 8.93070817e-01 2.73549054e-02
-5.75499475e-01 -9.16092932e-01 2.51379400e-01 3.38047296e-01
1.53316081e-01 7.34641433e-01 9.49617177e-02 -2.12005526e-01
-4.31511790e-01 -3.67085785e-01 1.37967750e-01 5.65674365e-01
2.72120740e-02 -4.91691917e-01 -8.25621068e-01 -6.13609195e-01
9.39248428e-02 -3.29988062e-01 9.52475071e-01 6.04620934e-01
7.48968661e-01 -2.85705894e-01 -1.89650148e-01 7.99446881e-01
1.15631521e+00 -2.67086118e-01 4.35024232e-01 5.36043227e-01
9.31473315e-01 4.91545230e-01 5.11068583e-01 3.89513642e-01
2.60960996e-01 9.62758958e-01 3.29454005e-01 -7.63144270e-02
-1.71463713e-01 -2.73767382e-01 5.61039805e-01 5.40111542e-01
-5.69455683e-01 -7.37113133e-02 -9.27756011e-01 4.57823932e-01
-1.92810261e+00 -8.88815522e-01 9.37362462e-02 2.02491641e+00
6.80851817e-01 3.71940076e-01 4.00680721e-01 9.33005065e-02
4.72314745e-01 2.61733860e-01 -4.63427275e-01 4.52127457e-02
2.62500942e-01 1.93929568e-01 6.01422191e-01 4.00337547e-01
-1.49869454e+00 7.72310376e-01 6.33663750e+00 3.92644256e-01
-1.07213521e+00 -4.19150479e-02 2.97057629e-01 -4.41313863e-01
4.33219761e-01 -2.35226884e-01 -9.50151742e-01 3.55145216e-01
6.46162271e-01 2.58204609e-01 1.31282620e-02 8.50978255e-01
2.61911303e-01 -1.21578112e-01 -1.31613183e+00 9.74250197e-01
2.70860255e-01 -9.14099336e-01 -5.74628301e-02 2.65506860e-02
6.26871824e-01 9.32664275e-02 6.87836930e-02 2.58866727e-01
1.68961346e-01 -1.02806032e+00 8.42861295e-01 5.64260304e-01
5.02675831e-01 -7.60644257e-01 5.42719007e-01 5.59298515e-01
-1.14936614e+00 2.44419813e-01 -2.12425873e-01 -3.98440361e-01
3.85803670e-01 2.29880691e-01 -8.27665269e-01 4.66889620e-01
7.72587001e-01 7.34342158e-01 -4.61766094e-01 9.57245350e-01
-4.26534057e-01 5.27658284e-01 -4.33979183e-01 2.27964416e-01
1.93084739e-02 6.46185726e-02 7.39304900e-01 1.22999310e+00
2.05958232e-01 -1.50076017e-01 3.88995409e-01 4.74043518e-01
-7.93508589e-02 -1.00822724e-01 -6.68971181e-01 3.52146775e-01
-1.32452682e-01 8.47138405e-01 -5.78597546e-01 -6.19421750e-02
-3.42639983e-01 1.28320825e+00 3.31807226e-01 2.88366616e-01
-9.66001451e-01 -2.50048220e-01 6.27786160e-01 2.63937891e-01
6.22150064e-01 -6.75316572e-01 2.43535042e-02 -1.29508495e+00
4.13403839e-01 -7.94310510e-01 2.84711063e-01 -4.68530357e-01
-1.07094526e+00 2.49232396e-01 1.35423675e-01 -1.44070530e+00
-6.28704548e-01 -8.35629642e-01 -3.37682754e-01 7.00798094e-01
-1.32364082e+00 -1.23497939e+00 -2.32405782e-01 6.64628088e-01
5.18460989e-01 1.29594430e-01 4.43792164e-01 3.34423989e-01
-5.24018526e-01 6.90428495e-01 -2.56135315e-01 3.73900503e-01
8.79396021e-01 -1.12042940e+00 6.37385428e-01 1.14103842e+00
2.66470015e-01 6.26576364e-01 9.81412947e-01 -6.16171539e-01
-1.48602843e+00 -9.36247647e-01 5.58307052e-01 -9.11607563e-01
3.98264229e-01 -3.25212836e-01 -7.63066173e-01 7.39386380e-01
5.51586151e-02 2.64600277e-01 4.10097301e-01 -1.14995785e-01
-3.85645807e-01 -2.02683844e-02 -9.76367772e-01 6.92120433e-01
1.15412128e+00 -2.73803651e-01 -8.06448638e-01 -5.67737222e-03
5.10704100e-01 -7.79994547e-01 -8.40627313e-01 6.10073090e-01
7.98467577e-01 -6.68238759e-01 1.27446067e+00 -5.69998980e-01
5.51458955e-01 -3.66421878e-01 -3.46623585e-02 -1.14708364e+00
-3.60302955e-01 -6.17492139e-01 -5.34954488e-01 7.83086777e-01
5.41004427e-02 -1.82799384e-01 1.31249094e+00 6.36631072e-01
2.32876405e-01 -5.81315637e-01 -9.70975399e-01 -8.79963100e-01
1.70768782e-01 -4.66148973e-01 -1.35585010e-01 7.11393595e-01
-2.74113268e-01 2.91055083e-01 -9.81557488e-01 2.27894455e-01
7.85583317e-01 -3.37508798e-01 1.28898513e+00 -1.06556761e+00
-6.53269470e-01 -2.94941336e-01 -6.20615363e-01 -1.45492518e+00
1.71556443e-01 -7.23062396e-01 2.09245354e-01 -1.21160436e+00
-1.67976350e-01 3.11475158e-01 -1.39768586e-01 4.40348029e-01
-2.61487663e-01 6.31448865e-01 4.35379952e-01 2.02006757e-01
-5.30020714e-01 5.97477198e-01 1.23182487e+00 -8.72962177e-02
-6.37293532e-02 6.04902543e-02 -1.29609421e-01 1.04779434e+00
6.59587860e-01 -5.26576400e-01 -3.04573804e-01 -5.92289150e-01
8.52349475e-02 1.76549535e-02 7.10245788e-01 -1.47277772e+00
2.69154042e-01 4.43337932e-02 8.17003369e-01 -5.82095146e-01
6.41812205e-01 -7.31987715e-01 4.83881347e-02 6.67798042e-01
-1.33723661e-01 2.58125961e-01 4.30879354e-01 8.45329225e-01
-1.39391661e-01 9.89332497e-02 7.47242510e-01 -1.97579861e-01
-9.28800344e-01 4.43314701e-01 -1.71464950e-01 3.12658995e-01
8.03913653e-01 -3.32434535e-01 3.56222987e-01 -3.64171863e-01
-8.32174301e-01 3.88639718e-02 4.75003093e-01 5.48435450e-01
5.57303786e-01 -1.32777297e+00 -6.65644467e-01 1.54445648e-01
-5.57795800e-02 9.45510417e-02 1.28080562e-01 8.93779397e-01
-6.77970767e-01 2.76763558e-01 -3.73962879e-01 -8.07559967e-01
-1.27445614e+00 3.39233637e-01 2.75396854e-01 -2.68104613e-01
-8.86450768e-01 8.37125480e-01 5.11428416e-02 -2.63715863e-01
4.97319460e-01 -2.59937972e-01 -6.22423328e-02 -2.34294429e-01
3.26225430e-01 4.50858146e-01 -1.35978162e-01 -7.85331786e-01
-3.75515401e-01 7.22352028e-01 -1.68576073e-02 -1.63016275e-01
1.31526780e+00 6.26074970e-02 4.07090247e-01 1.74587965e-01
1.28387654e+00 3.49792838e-02 -1.89277303e+00 -2.30170071e-01
2.83892378e-02 -4.80949134e-01 -9.50318351e-02 -4.68890190e-01
-9.36128259e-01 9.00334656e-01 6.92886114e-01 -3.91744614e-01
7.88050890e-01 -3.27589869e-01 9.04593349e-01 5.77291846e-01
2.50418603e-01 -1.20470393e+00 2.49695584e-01 5.54165483e-01
8.41277719e-01 -1.34619617e+00 4.03333455e-01 -1.40138850e-01
-4.39763397e-01 9.67484176e-01 6.53650165e-01 -6.05675459e-01
5.63859880e-01 1.10632926e-01 1.23741962e-01 1.19682841e-01
-3.37829113e-01 -4.57813889e-02 7.33887732e-01 5.31702161e-01
4.28633451e-01 -3.81615222e-01 -2.07120612e-01 3.35172027e-01
-1.56436652e-01 2.73030818e-01 1.59286916e-01 1.13315320e+00
-1.30013019e-01 -1.00913715e+00 -3.74823064e-01 -1.88484658e-02
-9.07865942e-01 1.72174811e-01 -6.14177696e-02 1.10017610e+00
4.53450792e-02 5.86732924e-01 -1.94708362e-01 -4.17044997e-01
4.93082821e-01 -2.81489901e-02 6.40622675e-01 -3.08829695e-01
-4.19047505e-01 1.50781557e-01 6.91314787e-02 -8.03182542e-01
-6.71582699e-01 -6.48536801e-01 -9.70729589e-01 -5.09586670e-02
-1.55371696e-01 -2.57927984e-01 5.24912715e-01 9.00157392e-01
2.29950041e-01 5.68462968e-01 1.27974302e-01 -1.43398952e+00
-7.44616508e-01 -8.82119119e-01 -2.15327919e-01 7.78732836e-01
5.58080018e-01 -8.98034871e-01 -1.91769212e-01 3.29856426e-01] | [7.1696391105651855, -0.7052644491195679] |
ba021f9e-ff6c-4a55-a393-5db6210fee07 | recommendation-system-based-upper-confidence | 1909.04190 | null | https://arxiv.org/abs/1909.04190v1 | https://arxiv.org/pdf/1909.04190v1.pdf | Recommendation System-based Upper Confidence Bound for Online Advertising | In this paper, the method UCB-RS, which resorts to recommendation system (RS) for enhancing the upper-confidence bound algorithm UCB, is presented. The proposed method is used for dealing with non-stationary and large-state spaces multi-armed bandit problems. The proposed method has been targeted to the problem of the product recommendation in the online advertising. Through extensive testing with RecoGym, an OpenAI Gym-based reinforcement learning environment for the product recommendation in online advertising, the proposed method outperforms the widespread reinforcement learning schemes such as $\epsilon$-Greedy, Upper Confidence (UCB1) and Exponential Weights for Exploration and Exploitation (EXP3). | ['Elena Simona Lohan', 'Flavian vasile', 'Nhan Nguyen-Thanh', 'Kinda Khawam', 'Dana Marinca', 'Steven Martin', 'Dominique Quadri', 'David Rohde'] | 2019-09-09 | null | null | null | null | ['product-recommendation'] | ['miscellaneous'] | [-3.30611646e-01 2.35478237e-01 -9.84373689e-01 -8.17114636e-02
-9.84237552e-01 -3.64921749e-01 2.66236663e-01 -2.76534140e-01
-4.56825048e-01 1.32584107e+00 3.55327711e-03 -1.02924299e+00
-1.01305747e+00 -6.22236729e-01 -7.31467903e-01 -7.31895447e-01
-2.90562898e-01 6.27853513e-01 -1.40009085e-02 -3.86081785e-01
4.86111969e-01 2.97378749e-01 -1.48540747e+00 -9.69762430e-02
7.03323364e-01 1.51000988e+00 7.09105581e-02 5.75598538e-01
1.90242026e-02 3.66306931e-01 -2.00530767e-01 -5.68876147e-01
5.10002673e-01 -1.56416997e-01 -4.61064667e-01 -8.83245468e-03
-1.96107551e-01 -3.28319103e-01 -5.32615967e-02 1.06102407e+00
1.73551708e-01 6.26481831e-01 6.85527265e-01 -9.94424045e-01
-4.68588650e-01 9.32256520e-01 -1.02171969e+00 7.27266133e-01
1.73356105e-02 -8.73292163e-02 1.09891307e+00 5.45221083e-02
1.84955463e-01 1.46503067e+00 2.77199447e-01 3.83500189e-01
-1.19998276e+00 -6.40894353e-01 5.18438756e-01 2.96818733e-01
-7.67944753e-01 3.05696297e-03 3.84105861e-01 6.93423599e-02
8.19691300e-01 4.53365743e-01 1.02419090e+00 1.08989251e+00
3.40047240e-01 8.48191738e-01 1.65586495e+00 -5.38572252e-01
6.81432366e-01 3.62658709e-01 5.65129638e-01 1.56780496e-01
5.78667402e-01 1.30025768e+00 -5.91204762e-02 -5.20163536e-01
7.48056591e-01 -2.40682751e-01 5.23133934e-01 -2.06959158e-01
-3.55051994e-01 1.61692536e+00 2.23658293e-01 -5.46542965e-02
-9.45813119e-01 4.15361881e-01 2.42166352e-02 4.74275857e-01
6.22870386e-01 4.21765208e-01 -5.14729381e-01 -2.64048845e-01
-5.77415764e-01 4.28918451e-01 7.10805535e-01 1.00229108e+00
3.67329270e-01 3.87275815e-01 -3.05206567e-01 6.45970047e-01
8.89115155e-01 6.09104872e-01 8.01193178e-01 -8.06140244e-01
9.51460972e-02 -4.85602356e-02 8.90781105e-01 -4.77303952e-01
-3.39173883e-01 -7.46887088e-01 -2.49367297e-01 1.64648339e-01
1.40312925e-01 -6.54772103e-01 -9.16973770e-01 1.39628899e+00
6.82009339e-01 2.34912470e-01 3.90495300e-01 6.71263158e-01
7.18611404e-02 9.52408135e-01 2.56999493e-01 -9.76013422e-01
1.24363399e+00 -1.09601700e+00 -1.03465497e+00 -2.10321754e-01
1.36465177e-01 -6.81400359e-01 6.46302819e-01 9.93174672e-01
-1.20387101e+00 -6.03497326e-01 -9.84837770e-01 9.98243630e-01
-3.05590153e-01 -1.65900052e-01 1.07559490e+00 1.27299011e+00
-8.04407954e-01 5.86831748e-01 -5.82169235e-01 -1.81166410e-01
1.94905072e-01 4.59247828e-01 5.82517028e-01 6.63820058e-02
-1.34947705e+00 9.53003347e-01 7.82218754e-01 1.45161659e-01
-8.70983243e-01 -1.76439509e-01 -4.58110124e-01 -4.63451399e-03
7.04474270e-01 -1.14695854e-01 1.54604542e+00 -9.92394626e-01
-2.13985705e+00 -1.66947469e-01 7.63727844e-01 -1.00209916e+00
2.57930309e-01 -3.05217505e-01 -9.58618581e-01 -2.10925594e-01
-2.89409190e-01 2.67679214e-01 1.09224641e+00 -1.01358402e+00
-1.13047802e+00 -5.50629437e-01 6.19038269e-02 1.53588966e-01
8.25664699e-02 4.84408997e-02 2.98680842e-01 -6.50589406e-01
-1.10747725e-01 -1.05375266e+00 -8.47923934e-01 -1.17793918e+00
1.58056334e-01 -4.25952882e-01 3.73931557e-01 -5.23462415e-01
1.34125054e+00 -1.99283028e+00 -2.10367545e-01 7.46891856e-01
-1.02194881e+00 1.75744072e-01 -8.46278146e-02 3.86822850e-01
3.69486511e-02 -1.79222167e-01 5.88062286e-01 4.03176963e-01
2.57897675e-01 5.86221993e-01 -4.36548263e-01 3.90655637e-01
-3.46048832e-01 6.15440786e-01 -6.99724793e-01 -4.61858541e-01
-6.29231781e-02 -3.82200837e-01 -6.32469475e-01 4.38809097e-02
-7.80488789e-01 4.31543328e-02 -9.67044175e-01 5.29452920e-01
6.24821067e-01 -1.79398209e-02 3.13108176e-01 1.03945091e-01
-3.79022896e-01 -2.34537929e-01 -1.73610091e+00 1.30747807e+00
-5.06836593e-01 -4.25047845e-01 -1.20893484e-02 -1.28113902e+00
1.08885813e+00 2.80761272e-01 4.33505327e-01 -8.15929413e-01
3.60161722e-01 2.20415279e-01 2.09567860e-01 -2.83025891e-01
5.02309263e-01 -2.09457725e-01 -1.42753273e-02 6.47404611e-01
1.38341859e-01 1.45169497e-01 4.69702661e-01 -1.90792289e-02
5.62580109e-01 5.02411604e-01 6.74960613e-01 -6.74394488e-01
3.63560975e-01 4.09478396e-02 4.29400861e-01 1.17415130e+00
-3.73723716e-01 -6.71019137e-01 3.69378388e-01 -2.46205866e-01
-7.93520868e-01 -7.86944985e-01 -3.67219478e-01 1.51052594e+00
2.04276949e-01 1.78791121e-01 -3.35405976e-01 -7.94253647e-01
4.38947797e-01 1.27133989e+00 -6.49902880e-01 -2.05755085e-01
-8.17282945e-02 -1.03840733e+00 -2.40904838e-01 3.12373608e-01
3.33126485e-01 -1.05100310e+00 -4.38951433e-01 8.24802995e-01
3.73484254e-01 -4.65601116e-01 -2.18520448e-01 5.47201693e-01
-1.20519900e+00 -6.35190487e-01 -5.30314386e-01 -3.75174075e-01
2.46462673e-02 -1.30422696e-01 7.92075038e-01 -6.14513278e-01
1.32552534e-01 6.39546037e-01 -7.18547404e-01 -6.55509055e-01
-2.97523946e-01 -1.44655153e-01 -1.93170700e-02 1.45815119e-01
5.51029325e-01 -3.05908471e-01 -7.43083239e-01 6.22665763e-01
-5.49492657e-01 -5.57726562e-01 9.37942743e-01 1.28157294e+00
6.35971487e-01 1.55644953e-01 1.22691119e+00 -7.72782147e-01
9.27809060e-01 -5.44719636e-01 -1.29794741e+00 2.69078940e-01
-1.26810920e+00 3.56444210e-01 2.18666077e-01 -9.55278635e-01
-1.34834468e+00 -6.39161691e-02 -1.19976357e-01 1.78877562e-01
1.05201252e-01 6.86959863e-01 4.96069521e-01 6.54287543e-03
8.54143083e-01 7.65192062e-02 1.89831242e-01 -4.46612656e-01
6.07305169e-01 9.72499430e-01 -2.00199082e-01 -7.59475470e-01
6.65934235e-02 -9.30219740e-02 -7.13207424e-02 -1.99434221e-01
-1.13716733e+00 -5.57847917e-01 2.57526308e-01 -1.59936517e-01
3.89597207e-01 -7.92164803e-01 -1.00301492e+00 -2.10446194e-01
-3.50336343e-01 -3.52630466e-01 -5.22914052e-01 1.01060545e+00
-1.28347194e+00 1.60498723e-01 -3.87596816e-01 -1.60454798e+00
-4.10900682e-01 -9.35230851e-01 3.17168206e-01 5.33427536e-01
4.34526116e-01 -6.77558601e-01 2.46743977e-01 3.94770771e-01
2.75963843e-01 -1.75236702e-01 6.69829190e-01 -8.15241098e-01
-1.11861780e-01 -3.00639644e-02 1.98904455e-01 4.09892470e-01
-3.48196328e-01 -2.68804848e-01 -2.35847369e-01 -5.69254696e-01
-1.90447241e-01 -7.05983222e-01 4.15910363e-01 9.25251961e-01
9.67610240e-01 -4.55297530e-01 -2.25733474e-01 3.33026461e-02
1.51899838e+00 9.79179323e-01 6.06254458e-01 8.45979393e-01
-6.69730306e-01 1.68538094e-02 1.74461663e+00 1.06489909e+00
-4.02622491e-01 4.78679150e-01 8.55189860e-01 2.79831320e-01
6.27104759e-01 -1.39721990e-01 3.87739182e-01 -2.43551116e-02
-2.29610503e-01 -2.64392108e-01 4.71998751e-02 2.81443357e-01
-2.26459551e+00 -9.01922584e-01 3.46129656e-01 2.16745710e+00
6.76070273e-01 4.87717748e-01 6.42803192e-01 -1.10469319e-01
8.39920223e-01 -2.46826395e-01 -9.57194865e-01 -1.14788866e+00
2.33751640e-01 5.00402510e-01 1.01851010e+00 4.86842692e-01
-9.51319933e-01 8.84834230e-01 6.70286322e+00 1.26069117e+00
-4.54780787e-01 3.74295712e-01 6.80547476e-01 1.56853318e-01
-6.30381703e-02 1.83935289e-03 -1.19552052e+00 4.91663933e-01
1.43728697e+00 -9.71588343e-02 8.70727301e-01 1.21759963e+00
2.01026171e-01 -4.46695328e-01 -2.40285948e-01 6.48277342e-01
-4.57133681e-01 -9.59536910e-01 -1.63676634e-01 3.03655446e-01
9.08469796e-01 -2.37057790e-01 4.63813841e-01 1.03652513e+00
1.17461014e+00 -4.48745728e-01 5.17114639e-01 2.71427423e-01
3.41760337e-01 -1.34318662e+00 7.13617444e-01 3.88624787e-01
-4.08040971e-01 -1.07303798e+00 -7.54457414e-01 1.01247191e-01
2.90527552e-01 2.88687386e-02 -5.47480822e-01 6.00637436e-01
5.08796394e-01 1.78418472e-01 1.13824613e-01 1.07895434e+00
4.11007144e-02 7.85137475e-01 -3.35901856e-01 -9.14937556e-01
1.13470924e+00 -7.31372833e-01 2.46296391e-01 6.35754704e-01
7.04079509e-01 3.56365561e-01 1.70567140e-01 1.80210158e-01
7.86274135e-01 5.09811997e-01 -2.31964532e-02 -2.31877267e-01
4.98162396e-02 1.20555973e+00 -4.82374251e-01 -3.18360567e-01
-2.48091087e-01 2.03033388e-01 1.47780091e-01 3.70827854e-01
-1.25140560e+00 7.77294785e-02 1.55957207e-01 -5.00451364e-02
1.01442850e+00 1.37137443e-01 3.08785349e-01 -3.08313072e-01
-8.34027529e-01 -9.98518705e-01 1.00155187e+00 -5.12217283e-01
-1.20894861e+00 5.32040000e-01 3.79913121e-01 -1.04101431e+00
-4.40230668e-01 -9.55040276e-01 -4.16345261e-02 5.31884789e-01
-1.45659494e+00 -8.40198874e-01 5.46170473e-01 8.36514890e-01
4.96820211e-01 -6.95110023e-01 6.02860272e-01 1.59054413e-01
-4.91006792e-01 4.80324715e-01 1.00357080e+00 -1.06412363e+00
3.16071570e-01 -1.19308817e+00 -4.02389407e-01 2.64730901e-01
-3.74770388e-02 3.16937059e-01 1.06204915e+00 -8.42570066e-01
-1.44451213e+00 -7.41965652e-01 -2.98572749e-01 4.93100047e-01
1.13993156e+00 2.07952052e-01 1.29108146e-01 7.36796558e-01
3.65155011e-01 -5.17625473e-02 6.09516978e-01 6.01211548e-01
2.04930045e-02 -4.84082222e-01 -1.32745206e+00 4.22132492e-01
5.20266533e-01 6.61589384e-01 -4.20452207e-01 6.70497715e-01
6.78378344e-01 -2.13999897e-01 -1.16453636e+00 1.68328494e-01
6.24827087e-01 -5.31739533e-01 9.33844149e-01 -1.33472943e+00
-2.95515120e-01 3.77371401e-01 -2.81146675e-01 -1.72901881e+00
-7.92254329e-01 -1.04412127e+00 -5.88724434e-01 6.13845289e-01
7.79939711e-01 -7.88127303e-01 8.07008684e-01 1.68929353e-01
-4.57481928e-02 -6.10426188e-01 -1.38166094e+00 -1.08035278e+00
-1.49720132e-01 -7.21042901e-02 5.77446461e-01 1.19101562e-01
1.02489077e-01 1.53397545e-01 -8.28318596e-01 5.75244837e-02
6.52567565e-01 3.14127147e-01 3.57236475e-01 -9.75750864e-01
-1.04315376e+00 -1.02367766e-01 1.28093213e-01 -1.15666270e+00
-1.66641563e-01 -1.72242805e-01 -4.51819561e-02 -9.02874351e-01
-6.66854829e-02 -7.17912078e-01 -9.63189542e-01 -4.23178114e-02
1.97838068e-01 -7.40854740e-02 -9.22420546e-02 -3.13781589e-01
-7.33661175e-01 4.63240921e-01 1.53561354e+00 -4.60180603e-02
-3.68899524e-01 9.43071604e-01 -8.89045715e-01 3.64524662e-01
8.76164734e-01 -4.40569609e-01 -5.35585999e-01 5.89912891e-01
3.81602407e-01 6.67114258e-01 -2.29775995e-01 -5.88257611e-01
-3.01391214e-01 -6.42336011e-01 -3.30032110e-02 -1.02291262e+00
3.23313236e-01 -1.11230087e+00 3.80899638e-01 9.73367810e-01
-4.11906391e-01 8.48501250e-02 2.89817721e-01 1.25604141e+00
1.66619003e-01 -9.72769678e-01 8.09121430e-01 -2.73317158e-01
-6.05554581e-01 2.27650195e-01 -4.43670273e-01 -3.28723013e-01
1.22969306e+00 -2.65779961e-02 -2.29562044e-01 -5.65832675e-01
-1.29972160e+00 1.77402765e-01 -6.88517928e-01 4.86079268e-02
1.49081290e-01 -1.42274249e+00 -4.24783260e-01 -1.92154914e-01
-1.11660704e-01 -1.01635587e+00 3.31127137e-01 7.38616824e-01
-1.18605904e-02 6.25109792e-01 -3.47290963e-01 -3.17642353e-02
-8.18427265e-01 1.10968542e+00 2.02098951e-01 -9.63950515e-01
-4.45806012e-02 5.03795147e-01 -5.14580488e-01 6.47039935e-02
3.12380731e-01 9.03505832e-04 -5.98049283e-01 2.35811938e-02
3.58059108e-01 5.37962198e-01 -8.55880603e-02 1.21827852e-02
3.52638781e-01 -1.16021045e-01 -5.46211600e-01 -5.26498616e-01
1.35906136e+00 -3.24912816e-01 4.31833625e-01 9.18127671e-02
2.78247237e-01 -4.07093972e-01 -1.30190980e+00 -7.14781061e-02
-1.46018220e-02 -5.53416371e-01 6.43528521e-01 -1.22470415e+00
-7.82977402e-01 4.79150414e-02 1.33716214e+00 7.87119865e-01
7.51085401e-01 -3.44151765e-01 3.98116797e-01 4.96564239e-01
5.03478289e-01 -1.79105175e+00 1.52881116e-01 9.28996801e-02
8.04313958e-01 -1.29483712e+00 2.53091633e-01 1.37530997e-01
-7.71010399e-01 8.92676294e-01 7.82874942e-01 -4.15742874e-01
1.04159963e+00 -4.08933237e-02 -3.54970038e-01 1.01108164e-01
-8.06352198e-01 -4.19210643e-01 1.50110601e-02 4.59149420e-01
4.72034953e-05 2.72285640e-01 -1.34223390e+00 8.07116687e-01
1.01537310e-01 1.73142835e-01 3.29580396e-01 9.53069925e-01
-7.67593622e-01 -1.16994715e+00 -6.06417716e-01 7.66290247e-01
-5.46951294e-01 8.76874328e-02 4.27799672e-01 1.05878937e+00
-1.68584079e-01 1.23013222e+00 7.42036663e-03 4.02108468e-02
2.54092813e-01 -3.75987262e-01 7.08447099e-01 4.06774553e-03
-5.50125420e-01 8.60893190e-01 4.97928262e-01 -4.54501539e-01
-5.04594266e-01 -8.52435827e-01 -4.70311224e-01 9.12572965e-02
-1.07087743e+00 9.44013238e-01 9.71787453e-01 5.75148284e-01
2.82514960e-01 4.18553561e-01 1.03237534e+00 -3.10894310e-01
-1.79206145e+00 -1.36170137e+00 -1.35789871e+00 -8.50120708e-02
-2.37030104e-01 -1.23611474e+00 -2.85973370e-01 -7.99039721e-01] | [4.527740955352783, 3.2103469371795654] |
eb9fc1e6-de2e-4b28-a915-afd2630a7840 | empirical-risk-minimization-and-stochastic | 1806.10701 | null | http://arxiv.org/abs/1806.10701v2 | http://arxiv.org/pdf/1806.10701v2.pdf | Empirical Risk Minimization and Stochastic Gradient Descent for Relational Data | Empirical risk minimization is the main tool for prediction problems, but its
extension to relational data remains unsolved. We solve this problem using
recent ideas from graph sampling theory to (i) define an empirical risk for
relational data and (ii) obtain stochastic gradients for this empirical risk
that are automatically unbiased. This is achieved by considering the method by
which data is sampled from a graph as an explicit component of model design. By
integrating fast implementations of graph sampling schemes with standard
automatic differentiation tools, we provide an efficient turnkey solver for the
risk minimization problem. We establish basic theoretical properties of the
procedure. Finally, we demonstrate relational ERM with application to two
non-standard problems: one-stage training for semi-supervised node
classification, and learning embedding vectors for vertex attributes.
Experiments confirm that the turnkey inference procedure is effective in
practice, and that the sampling scheme used for model specification has a
strong effect on model performance. Code is available at
https://github.com/wooden-spoon/relational-ERM. | ['Peter Orbanz', 'Wenda Zhou', 'Morgane Austern', 'Victor Veitch', 'David M. Blei'] | 2018-06-27 | null | null | null | null | ['graph-sampling'] | ['graphs'] | [ 1.00585982e-01 5.69005847e-01 -5.20259023e-01 -4.08153027e-01
-9.44566309e-01 -2.68334746e-01 4.50653583e-01 1.18828893e-01
-1.10146195e-01 6.73255503e-01 -2.07070276e-01 -5.57463944e-01
-3.87538105e-01 -9.13472831e-01 -6.45406604e-01 -6.39551163e-01
-2.45551452e-01 7.51058936e-01 -1.54957697e-01 1.10439815e-01
1.96315780e-01 6.59516454e-01 -1.06525743e+00 -2.58702576e-01
8.04576159e-01 9.74319577e-01 -1.56615764e-01 8.33830476e-01
2.06468761e-01 7.46097863e-01 -1.00937048e-02 -8.11112523e-01
2.95899719e-01 -5.20146370e-01 -1.01515508e+00 1.79143339e-01
2.05352962e-01 -1.91608816e-01 -1.91013739e-01 1.14671433e+00
2.90090740e-01 4.43020314e-02 1.11714256e+00 -1.46394026e+00
-1.82863235e-01 8.37269843e-01 -5.28603613e-01 -5.21106906e-02
-4.27585375e-03 -2.31475651e-01 1.58241630e+00 -9.29603696e-01
5.47657847e-01 1.06530309e+00 1.03128004e+00 3.35656226e-01
-1.98906827e+00 -3.93589497e-01 -1.20573960e-01 -2.21037388e-01
-1.68619037e+00 -3.51379693e-01 8.22111785e-01 -8.10840905e-01
5.75423360e-01 3.98788810e-01 5.92652857e-01 9.75812256e-01
5.49366288e-02 8.85251880e-01 7.90241182e-01 -3.75911206e-01
3.63832027e-01 5.24816394e-01 4.65463340e-01 8.69130254e-01
4.18900311e-01 2.89303690e-01 -1.04648642e-01 -7.12583244e-01
6.05287910e-01 -1.20755821e-01 -7.82935098e-02 -9.01690841e-01
-5.21664262e-01 1.31713486e+00 1.89013928e-01 -1.54756412e-01
-2.93673068e-01 4.09141839e-01 2.78120756e-01 3.54588956e-01
8.39469492e-01 2.41180927e-01 -4.08333242e-01 2.68213630e-01
-9.70391512e-01 3.99581283e-01 1.27148259e+00 8.83349478e-01
8.24063420e-01 1.60851758e-02 8.46158639e-02 8.74024928e-01
5.42178452e-01 3.32502395e-01 -1.76732376e-01 -1.19046199e+00
2.76355386e-01 2.57095605e-01 -1.23989582e-01 -1.05360699e+00
-3.30078095e-01 -5.41145802e-01 -8.65658343e-01 -1.17241675e-02
7.16840625e-01 -2.67769843e-01 -3.86382937e-01 1.72751570e+00
5.71223557e-01 1.24254366e-02 -2.49822214e-01 3.81601810e-01
2.90514499e-01 5.11293769e-01 -5.62563129e-02 -2.94782668e-01
9.93204236e-01 -5.70613503e-01 -4.27406520e-01 -8.62440211e-04
1.09223378e+00 -3.01058173e-01 8.77626657e-01 4.19452459e-01
-1.05947924e+00 -4.63447496e-02 -6.95900679e-01 -4.38958667e-02
-1.01832792e-01 2.20366612e-01 9.08402979e-01 9.36625123e-01
-1.08183777e+00 1.05394769e+00 -8.42578769e-01 -1.16301686e-01
6.56707764e-01 4.56836492e-01 -2.56524414e-01 2.60019928e-01
-1.08102572e+00 6.70812726e-01 1.62639603e-01 2.01259598e-01
-8.50876868e-01 -8.47593904e-01 -9.34915006e-01 7.99992532e-02
5.95243633e-01 -6.14227057e-01 1.23617673e+00 -8.63425612e-01
-1.34801865e+00 8.53526413e-01 -4.33011679e-03 -5.72041571e-01
7.52436399e-01 5.45457602e-02 2.66594887e-01 -1.93639826e-02
-5.41865453e-02 1.73284151e-02 8.69827509e-01 -1.14986122e+00
-1.26538172e-01 -3.94465238e-01 -1.82015046e-01 -3.46553847e-02
-2.26800814e-01 -5.96582368e-02 -3.10761750e-01 -5.41822076e-01
-4.59345989e-02 -9.45165753e-01 -5.91137588e-01 -8.87671486e-03
-7.64024734e-01 -2.38338366e-01 7.09244907e-02 -7.59401917e-01
1.26598740e+00 -1.88066614e+00 2.26843029e-01 8.72955263e-01
3.71050626e-01 -1.40714824e-01 1.84878498e-01 6.73000336e-01
-1.52152032e-01 3.78808916e-01 -6.01942718e-01 -5.62173486e-01
1.93873331e-01 5.00662532e-03 -1.38439909e-01 6.49382830e-01
1.18945487e-01 8.16245675e-01 -7.30069816e-01 -5.44623733e-01
2.05060970e-02 3.98754716e-01 -7.93339431e-01 1.42601103e-01
-1.28284693e-01 -1.80040181e-01 -5.29002547e-01 3.13090593e-01
6.00590467e-01 -7.09580660e-01 5.04584253e-01 -6.17373586e-02
4.45772320e-01 3.98506641e-01 -1.40564322e+00 1.09860206e+00
-2.86830395e-01 3.87651294e-01 2.41356671e-01 -1.10353768e+00
6.91369116e-01 7.94538558e-02 4.50287014e-01 3.48785520e-01
2.66799271e-01 -2.67387144e-02 -3.39593202e-01 -2.22951043e-02
2.31905490e-01 -2.21511617e-01 1.26134783e-01 5.69729090e-01
7.28286207e-02 -2.25837961e-01 1.31726831e-01 5.34581780e-01
1.00965416e+00 -3.03071160e-02 3.77964199e-01 -6.10048473e-01
2.19392061e-01 2.90879840e-03 3.97163838e-01 8.46538186e-01
1.34310707e-01 4.18356121e-01 1.23757672e+00 -1.06274322e-01
-8.85067940e-01 -1.00311208e+00 -4.45382476e-01 7.10952640e-01
-4.57284421e-01 -5.79488814e-01 -8.74640107e-01 -9.39757288e-01
2.68269986e-01 8.35462987e-01 -8.59480500e-01 -1.26065418e-01
-7.88224190e-02 -9.47907984e-01 2.77673006e-01 3.85450661e-01
-2.00684845e-01 -5.65132737e-01 4.89911996e-02 -7.78786466e-02
1.48212612e-01 -6.81801260e-01 -5.76985359e-01 3.55156541e-01
-1.11158288e+00 -1.24221659e+00 -4.56156880e-01 -5.26476145e-01
8.19534600e-01 -1.31426200e-01 1.26781142e+00 2.24533334e-01
-2.99741149e-01 4.69692916e-01 4.92036268e-02 -1.35480374e-01
-5.63460767e-01 3.99872422e-01 -1.00066505e-01 1.43507630e-01
1.53560877e-01 -4.89220768e-01 -2.60081559e-01 2.34593645e-01
-6.13633752e-01 1.71139855e-02 2.89746881e-01 9.86909986e-01
6.40245855e-01 -1.63266063e-02 3.57135713e-01 -1.47721398e+00
7.28953302e-01 -6.49956107e-01 -1.15599036e+00 1.75648928e-01
-1.09265459e+00 2.51920372e-01 3.42597663e-01 -2.07012400e-01
-6.13249481e-01 1.83636919e-01 -1.30406365e-01 -3.22427809e-01
4.18667585e-01 6.93413675e-01 -3.65859736e-03 -2.69369017e-02
5.98988950e-01 -2.51503915e-01 4.86697525e-01 -4.14669842e-01
3.07361573e-01 4.15204793e-01 -1.81285828e-01 -7.04575837e-01
9.05055344e-01 2.40362987e-01 3.42923641e-01 -9.49175239e-01
-8.15957069e-01 -1.11038454e-01 -5.60120285e-01 -7.83400536e-02
6.74898863e-01 -7.50716865e-01 -7.86381960e-01 1.80861145e-01
-7.88024962e-01 -8.47937524e-01 -4.87280220e-01 4.34005469e-01
-6.62869751e-01 2.24003419e-01 -7.77043104e-01 -1.19959033e+00
-2.01797664e-01 -9.50636864e-01 9.25892174e-01 -3.32080781e-01
-1.94059193e-01 -1.47496331e+00 1.65830582e-01 2.09687650e-01
-4.47957553e-02 4.45142649e-02 8.61694753e-01 -7.47140586e-01
-5.86228311e-01 -4.57125872e-01 -1.48944452e-01 6.16786003e-01
-2.40206629e-01 2.93398440e-01 -8.47831309e-01 -2.51474530e-01
-1.88475084e-02 -3.19890290e-01 8.03140938e-01 7.77654052e-01
1.22757614e+00 -3.48003030e-01 -4.18953538e-01 7.46947229e-01
1.48335886e+00 -5.06585062e-01 4.22872066e-01 -2.82573581e-01
8.24124634e-01 8.41484487e-01 4.59764749e-01 4.49801385e-01
4.62216377e-01 6.47666097e-01 1.10751830e-01 9.85083729e-02
3.48172814e-01 -5.32014549e-01 3.55110973e-01 7.38987923e-01
1.78606287e-01 2.28845537e-01 -1.01436222e+00 4.65343833e-01
-1.84077621e+00 -1.02547455e+00 -3.67020816e-01 2.41575456e+00
9.97599900e-01 5.12198731e-02 4.50854748e-01 6.80364668e-02
5.93655646e-01 2.93841194e-02 -3.73330474e-01 -3.36482227e-01
2.77405679e-01 9.33974236e-02 8.54073584e-01 1.06057823e+00
-9.67709184e-01 7.79970288e-01 6.83476305e+00 1.01688290e+00
-5.48498690e-01 7.72030652e-02 1.10898268e+00 1.10119544e-01
-4.87683624e-01 2.32930034e-01 -1.00864553e+00 1.68369070e-01
1.11540294e+00 -3.24986905e-01 6.75545812e-01 1.17435956e+00
1.33385986e-01 2.80996058e-02 -1.26675081e+00 6.05815053e-01
-2.60807008e-01 -1.26513267e+00 -4.42654014e-01 5.31202376e-01
6.50874794e-01 -2.23735999e-02 -4.55317125e-02 2.29215592e-01
7.71893144e-01 -1.11108005e+00 2.80022621e-01 4.38035816e-01
6.38180673e-01 -8.84992003e-01 4.84083325e-01 2.70841926e-01
-9.61666346e-01 2.04983711e-01 -4.87771481e-01 1.55029356e-01
-5.64929210e-02 1.11581314e+00 -8.99899364e-01 5.13049006e-01
3.31926554e-01 7.91612625e-01 -5.15883684e-01 5.98686993e-01
-1.44919530e-01 9.27631080e-01 -5.46819925e-01 -4.84036282e-02
-2.59633273e-01 -5.91393173e-01 4.58034396e-01 9.87380326e-01
1.10105857e-01 -2.16854274e-01 -6.51851967e-02 1.14011896e+00
-1.13241233e-01 2.05153063e-01 -9.42123413e-01 -1.32447377e-01
3.93274754e-01 1.29658127e+00 -6.32394373e-01 -1.04115173e-01
-3.05439293e-01 6.04492247e-01 6.68970406e-01 4.23318803e-01
-6.74317420e-01 -3.19463938e-01 6.07915819e-01 2.37228259e-01
4.36839521e-01 1.50822857e-02 -3.81486595e-01 -1.20065832e+00
-3.21619734e-02 -1.00227153e+00 4.99281138e-01 -4.19492453e-01
-1.43981373e+00 2.17375860e-01 2.04796657e-01 -7.00920105e-01
-4.53982323e-01 -7.70016074e-01 -3.57855916e-01 9.60146070e-01
-1.04928374e+00 -6.90153360e-01 1.40436411e-01 2.73436278e-01
-8.94532055e-02 1.69108689e-01 6.79392517e-01 1.01333708e-01
-9.71378922e-01 7.20273972e-01 2.13515222e-01 2.18892843e-01
5.95918819e-02 -1.53208685e+00 4.57232624e-01 7.27461934e-01
2.71776140e-01 6.78506792e-01 7.03700662e-01 -7.19786108e-01
-1.47621012e+00 -1.01151586e+00 9.35093880e-01 -6.44153178e-01
1.05822921e+00 -6.47189438e-01 -8.40480268e-01 9.95467067e-01
-2.70949870e-01 1.08346596e-01 7.30566144e-01 5.02170801e-01
-1.74528211e-01 -2.24151120e-01 -1.12479460e+00 6.12818897e-01
9.78529572e-01 -5.72376609e-01 5.15471771e-02 6.71631455e-01
4.50388014e-01 -9.73331183e-02 -1.20531213e+00 2.59233892e-01
4.47762221e-01 -8.46002460e-01 9.04309392e-01 -8.06275427e-01
3.70094270e-01 1.74505025e-01 -2.51442015e-01 -1.15129423e+00
-1.01036347e-01 -1.03540373e+00 -3.03569406e-01 1.14952016e+00
8.39505851e-01 -1.05000436e+00 1.02635860e+00 9.09692347e-01
4.36831832e-01 -1.04791903e+00 -5.24929404e-01 -7.30822504e-01
2.67477572e-01 -5.41266024e-01 2.30667368e-01 8.52785110e-01
9.36726667e-03 4.53043878e-01 -4.33955222e-01 1.68006733e-01
1.13441527e+00 -1.24390036e-01 9.59510803e-01 -1.49091351e+00
-6.86122000e-01 -4.79386181e-01 -3.46629918e-01 -9.07799482e-01
4.82025087e-01 -1.29649282e+00 -1.60380766e-01 -1.26138163e+00
2.54181027e-01 -7.08027899e-01 6.49336427e-02 1.40383646e-01
-1.98626414e-01 -1.65198103e-01 -1.96544424e-01 -1.67819019e-02
-1.35294765e-01 5.74145138e-01 6.65515780e-01 2.88452655e-01
-9.94309187e-02 4.64136869e-01 -7.97785223e-01 5.88700294e-01
8.83950651e-01 -7.00832963e-01 -6.22201800e-01 1.35597333e-01
4.64151800e-01 2.72268623e-01 4.74651277e-01 -5.06601572e-01
-5.01519889e-02 -1.78472981e-01 1.80452168e-02 -3.71758819e-01
1.68265924e-01 -7.84192920e-01 4.11099434e-01 5.19069731e-01
-7.44448304e-01 -3.44479047e-02 -1.31514102e-01 6.71471834e-01
2.05931365e-01 -7.88734019e-01 8.21929634e-01 2.13838264e-01
2.04432428e-01 5.46991646e-01 -2.15316772e-01 4.52595711e-01
8.94371271e-01 2.09641322e-01 1.10068761e-01 -5.25595248e-01
-8.94446015e-01 2.54273117e-01 5.27452528e-01 -4.25813258e-01
3.80435795e-01 -1.35733879e+00 -7.03832567e-01 1.61786899e-01
-1.84064180e-01 -1.23247258e-01 -1.44289628e-01 1.14857411e+00
-5.09103894e-01 -3.06665315e-03 6.33054197e-01 -5.35030484e-01
-1.25729930e+00 4.48284686e-01 5.59503794e-01 -5.11881769e-01
-4.68344688e-01 8.86856675e-01 5.27756400e-02 -6.36996329e-01
2.63369143e-01 -8.78810585e-02 2.06298679e-01 1.78501859e-01
2.75381446e-01 4.61568952e-01 -1.11445114e-01 -2.06970438e-01
-2.48352349e-01 3.38117748e-01 -8.68087560e-02 -1.95557028e-01
1.35381317e+00 -1.66937746e-02 -2.55921006e-01 6.96271539e-01
1.39687717e+00 -3.65758911e-02 -1.16538060e+00 -2.60223806e-01
2.12420911e-01 -3.77549767e-01 1.33139580e-01 -1.33077726e-01
-1.14164722e+00 7.66335249e-01 8.17279443e-02 6.94066644e-01
7.11393714e-01 2.11818535e-02 1.03439778e-01 3.81357551e-01
3.10838580e-01 -1.08189023e+00 -3.49042863e-01 1.72814474e-01
7.63406456e-01 -1.20820260e+00 4.15461898e-01 -7.28860617e-01
-6.32210672e-01 8.32805634e-01 2.90809404e-02 -3.47317338e-01
1.18340278e+00 2.69988388e-01 -3.21833730e-01 -2.84532964e-01
-9.15833831e-01 -6.54507801e-02 3.62296104e-01 4.09109324e-01
4.52173442e-01 1.55455589e-01 -1.21198922e-01 4.74323332e-01
-1.70886666e-01 -1.91146269e-01 4.48659897e-01 3.79197866e-01
-1.16648808e-01 -1.16312683e+00 1.41739562e-01 1.02522671e+00
-4.76639688e-01 -2.56571800e-01 -4.07821953e-01 9.33828712e-01
-7.95545936e-01 7.01561093e-01 -2.18259320e-01 -3.52907211e-01
4.99070389e-03 2.79146969e-01 4.89313126e-01 -6.84380174e-01
-2.87908137e-01 -4.05979939e-02 4.89163607e-01 -5.51626444e-01
-6.37983978e-02 -1.08635271e+00 -8.20152164e-01 -5.48264802e-01
-6.87164605e-01 4.17405069e-01 5.54552317e-01 5.69429219e-01
2.10692957e-01 2.11092070e-01 8.12317789e-01 -6.90066099e-01
-1.25231278e+00 -7.92893410e-01 -8.74325097e-01 8.62408429e-02
1.72878280e-01 -5.08467257e-01 -9.22511041e-01 -2.41957322e-01] | [7.570588111877441, 4.508666038513184] |
687a01a4-5349-408f-858f-0ee4f10fd48b | pct-point-cloud-transformer | 2012.09688 | null | https://arxiv.org/abs/2012.09688v4 | https://arxiv.org/pdf/2012.09688v4.pdf | PCT: Point cloud transformer | The irregular domain and lack of ordering make it challenging to design deep neural networks for point cloud processing. This paper presents a novel framework named Point Cloud Transformer(PCT) for point cloud learning. PCT is based on Transformer, which achieves huge success in natural language processing and displays great potential in image processing. It is inherently permutation invariant for processing a sequence of points, making it well-suited for point cloud learning. To better capture local context within the point cloud, we enhance input embedding with the support of farthest point sampling and nearest neighbor search. Extensive experiments demonstrate that the PCT achieves the state-of-the-art performance on shape classification, part segmentation and normal estimation tasks. | ['Shi-Min Hu', 'Ralph R. Martin', 'Tai-Jiang Mu', 'Zheng-Ning Liu', 'Jun-Xiong Cai', 'Meng-Hao Guo'] | 2020-12-17 | null | null | null | null | ['3d-part-segmentation'] | ['computer-vision'] | [ 6.56065345e-03 -3.96076739e-01 -1.83039501e-01 -3.46483916e-01
-5.71179688e-01 -6.46972775e-01 5.93388021e-01 2.56183773e-01
-1.91390306e-01 -7.14161340e-03 -3.28148991e-01 -5.22646010e-01
-2.07550317e-01 -1.11768901e+00 -1.13530290e+00 -5.17323494e-01
-3.08541596e-01 7.06003964e-01 1.56666517e-01 -1.50162159e-02
3.99373621e-01 1.25238657e+00 -1.35376894e+00 4.09723252e-01
5.06339431e-01 1.08600974e+00 4.67501372e-01 3.06841820e-01
-7.02030420e-01 3.78466882e-02 -3.85100931e-01 -4.76594061e-01
4.44347471e-01 4.27677989e-01 -6.05824649e-01 3.41111533e-02
7.66997218e-01 -1.37292489e-01 -3.55237603e-01 9.66740012e-01
3.79930466e-01 -9.63013768e-02 7.29548812e-01 -1.30969477e+00
-9.42542434e-01 2.40024209e-01 -6.50377989e-01 1.68187395e-01
4.38574739e-02 4.78766300e-02 1.15382266e+00 -1.41658819e+00
4.04195070e-01 1.24502385e+00 9.60550666e-01 1.98606357e-01
-9.39328611e-01 -5.48469245e-01 7.03864172e-02 2.45617107e-01
-1.47622919e+00 1.02220811e-01 9.72967923e-01 -3.32538813e-01
1.14904308e+00 2.56148160e-01 1.00600040e+00 7.04884827e-01
2.23578095e-01 1.01179338e+00 5.32889485e-01 -1.57010227e-01
3.03039998e-01 -4.73373950e-01 -1.06491484e-01 6.18108928e-01
-2.41519487e-03 4.91975546e-02 -3.68036062e-01 -3.15462530e-01
1.29301858e+00 4.08259869e-01 1.09269172e-01 -7.92715192e-01
-1.52509594e+00 7.60007381e-01 1.09392297e+00 1.43351391e-01
-4.09973413e-01 4.85352874e-01 2.50764847e-01 1.85539737e-01
5.19172847e-01 3.66895616e-01 -5.15298247e-01 5.07420674e-02
-8.67472053e-01 4.38297749e-01 3.27476770e-01 1.29433441e+00
6.92742288e-01 4.04399075e-02 -5.54387420e-02 8.43292475e-01
3.69647145e-01 7.27048278e-01 4.52286117e-02 -8.94339919e-01
5.53140521e-01 8.37978005e-01 -2.44300842e-01 -1.01456630e+00
-2.29307845e-01 -3.97692919e-01 -9.82256114e-01 4.41486448e-01
-1.25397667e-01 4.71446037e-01 -1.11811924e+00 1.04544294e+00
2.25998461e-01 6.16642475e-01 -1.48692340e-01 7.54501700e-01
9.09619153e-01 8.03010583e-01 -3.29872593e-02 4.34328556e-01
1.28815913e+00 -5.84596574e-01 -1.72197029e-01 -1.38239220e-01
1.01542741e-01 -6.38533175e-01 1.15513396e+00 1.92774862e-01
-1.00287879e+00 -7.18449712e-01 -1.08635485e+00 -2.86110431e-01
-3.81160229e-01 1.47257775e-01 8.93346965e-01 2.45886773e-01
-1.00195062e+00 8.15341055e-01 -1.23231626e+00 -3.07511594e-02
1.00784743e+00 5.61184466e-01 -3.20652038e-01 -2.19452783e-01
-5.62481701e-01 3.46713126e-01 8.16671997e-02 6.81363121e-02
-3.57950658e-01 -1.08958650e+00 -9.52332556e-01 2.16742203e-01
-4.29278053e-02 -9.88498390e-01 1.10527802e+00 -3.03399146e-01
-1.26687980e+00 9.39414144e-01 -4.05926049e-01 -4.08718407e-01
3.40571284e-01 -1.46809816e-01 -6.06970079e-02 1.94891930e-01
3.64668369e-01 1.01846337e+00 1.15927756e+00 -1.17199361e+00
-5.21804452e-01 -6.94886565e-01 -3.87008637e-01 1.64924189e-01
-1.13212978e-02 -1.75932348e-01 -8.17651629e-01 -6.11039281e-01
6.62385464e-01 -9.34013724e-01 -2.75058985e-01 4.26245540e-01
-2.40692407e-01 -8.26742351e-01 1.17657089e+00 -1.60062298e-01
4.57822353e-01 -2.37340069e+00 -2.26565108e-01 4.17031139e-01
3.17545980e-01 2.61381090e-01 -2.65270084e-01 5.55099666e-01
-2.28348881e-01 2.54309297e-01 -4.22369778e-01 -3.61155450e-01
6.53536096e-02 3.78188074e-01 -5.69673002e-01 5.11996508e-01
6.40969753e-01 1.36120129e+00 -7.39525020e-01 -2.58724689e-01
7.17215478e-01 6.31894767e-01 -5.83135486e-01 -1.20075062e-01
-2.56249428e-01 1.64202929e-01 -5.11471331e-01 9.17958677e-01
1.07351720e+00 -3.62735569e-01 -6.24838054e-01 -2.37754926e-01
-2.24928558e-03 4.04045619e-02 -9.65135336e-01 1.78247273e+00
-2.78886259e-01 6.59490764e-01 -2.50552267e-01 -7.01353967e-01
1.02982056e+00 2.11310070e-02 7.81280100e-01 -4.02913332e-01
-1.48613051e-01 5.82450069e-02 -1.71983466e-01 -2.08166197e-01
3.82450461e-01 7.04662427e-02 9.47109312e-02 1.07633090e-02
-7.60650113e-02 -7.68096268e-01 -2.67931223e-01 -1.33829832e-01
8.76162291e-01 -1.58102766e-01 9.82458666e-02 -1.51105374e-01
3.21182877e-01 6.39253631e-02 3.72398287e-01 4.91903067e-01
1.35334311e-02 8.87221396e-01 1.05601631e-01 -7.01507568e-01
-1.12537277e+00 -1.74913979e+00 -3.56754541e-01 6.48765266e-01
2.62271941e-01 -3.28903854e-01 -3.03525627e-01 -4.35376674e-01
4.63023335e-01 4.56661075e-01 -2.58641154e-01 8.43374878e-02
-7.64806271e-01 -2.24389732e-01 4.01978254e-01 9.24517512e-01
6.80344462e-01 -1.16587222e+00 -2.60971427e-01 2.52942014e-02
9.70557481e-02 -1.23199487e+00 -3.70443553e-01 -2.77587883e-02
-1.27152562e+00 -9.38978136e-01 -6.79660916e-01 -1.12140727e+00
7.59947419e-01 5.47055364e-01 1.29459167e+00 -2.95577280e-04
-4.91403267e-02 4.58359867e-01 -2.65429109e-01 -7.45803356e-01
1.96751788e-01 1.23486318e-01 -1.33671775e-01 -3.08130890e-01
7.01188385e-01 -6.41042650e-01 -4.67387736e-01 3.77561562e-02
-8.82553458e-01 -3.70686918e-01 6.91080749e-01 8.25136781e-01
1.15052605e+00 1.57042816e-01 -1.14249820e-02 -3.67118657e-01
6.14799976e-01 -1.69424668e-01 -6.69195652e-01 -8.41623694e-02
-1.20096371e-01 -2.44237944e-01 4.52419966e-01 -2.63726525e-02
-2.69643068e-01 2.41748452e-01 -5.10811865e-01 -1.00348258e+00
-2.67321646e-01 3.45223457e-01 -1.48983940e-01 -3.52848351e-01
2.87214100e-01 3.62330258e-01 6.09185621e-02 -4.78082776e-01
3.64319623e-01 3.87158602e-01 7.46294320e-01 -4.75879580e-01
1.04932213e+00 8.62198293e-01 2.06433594e-01 -1.27317286e+00
-2.91711807e-01 -7.10370839e-01 -1.06401014e+00 1.46791443e-01
8.82844090e-01 -8.72848690e-01 -8.21634889e-01 3.41692895e-01
-1.49433196e+00 -8.31244327e-03 -3.97494525e-01 1.49744734e-01
-6.23283863e-01 4.53154117e-01 -3.88781726e-01 -3.77677679e-01
-6.45739973e-01 -1.08848917e+00 1.68512309e+00 -9.64367688e-02
1.73324525e-01 -8.54340374e-01 -2.23623514e-01 -4.29658331e-02
-1.07656606e-02 1.93465099e-01 1.12810302e+00 -3.01473260e-01
-1.12281263e+00 -2.69764692e-01 -3.30088317e-01 4.22594786e-01
8.07769969e-02 1.32303536e-01 -7.19395757e-01 -3.22367460e-01
-4.92562167e-02 2.19696298e-01 7.90504634e-01 4.86597508e-01
1.93436372e+00 9.73436013e-02 -5.64085364e-01 1.09828281e+00
1.58984160e+00 -1.96260810e-01 7.53374219e-01 3.64697576e-01
1.00225961e+00 1.46462426e-01 4.07187849e-01 2.40679681e-01
4.59235132e-01 5.82141638e-01 7.17096150e-01 -1.23580538e-01
3.83786857e-03 -4.84378576e-01 -8.81757364e-02 8.17523956e-01
9.35136005e-02 2.09680405e-02 -1.33238363e+00 6.32525027e-01
-1.55997467e+00 -8.00206602e-01 -2.12564379e-01 1.81923473e+00
2.50570238e-01 9.88392383e-02 -2.96564698e-01 9.59501192e-02
4.72102642e-01 1.75207019e-01 -5.05743325e-01 -2.89894074e-01
-1.42538711e-01 6.49089277e-01 6.89755440e-01 1.79675475e-01
-1.25609589e+00 1.18368471e+00 6.65474129e+00 8.84087324e-01
-1.18941581e+00 -1.91068694e-01 2.78954327e-01 2.62846649e-01
-3.20882410e-01 -4.11792338e-01 -6.10192299e-01 3.54638040e-01
2.51719922e-01 1.06402397e-01 9.34692696e-02 1.07554889e+00
1.28864750e-01 4.23938572e-01 -1.13763309e+00 1.42545187e+00
-2.21490651e-01 -1.79970694e+00 4.40951109e-01 8.43549073e-02
6.28023922e-01 7.53490746e-01 3.50148529e-01 1.74735636e-01
-1.91898085e-02 -1.13225532e+00 6.10610723e-01 2.50587910e-01
8.82252574e-01 -9.69725490e-01 6.36303186e-01 4.55584347e-01
-1.32552731e+00 5.82292266e-02 -9.07745600e-01 8.73750523e-02
6.13417514e-02 5.87678850e-01 -1.01809824e+00 4.84946996e-01
9.89525676e-01 1.08580732e+00 -5.02934217e-01 1.31183577e+00
-6.82864115e-02 4.05992508e-01 -6.22787833e-01 1.30751386e-01
4.65309411e-01 -3.96991819e-01 8.01698446e-01 1.04102814e+00
4.80408877e-01 6.64976835e-02 2.10816339e-01 1.06799316e+00
-1.74764752e-01 -5.50849512e-02 -1.02974558e+00 1.05276488e-01
7.06068695e-01 9.24951792e-01 -8.99865210e-01 -1.60439223e-01
-3.09027761e-01 8.89896691e-01 3.37407649e-01 1.94430575e-01
-4.33574528e-01 -2.90001482e-01 1.17409384e+00 8.74674320e-02
8.15703928e-01 -8.82014632e-01 -9.13900375e-01 -8.35513651e-01
2.68513232e-01 -4.57274884e-01 -2.24789716e-02 -7.39795268e-01
-1.69272029e+00 5.52211404e-01 4.03739810e-02 -1.68498099e+00
1.43513866e-02 -1.05436885e+00 -9.34913456e-01 7.66292751e-01
-1.65171802e+00 -1.31929445e+00 -1.83029905e-01 6.74282253e-01
5.54032207e-01 -2.44203970e-01 6.47295058e-01 2.15926856e-01
3.44236419e-02 4.30963904e-01 2.45909825e-01 3.53238106e-01
2.11791415e-02 -1.35417795e+00 1.29570961e+00 6.47510111e-01
5.87061405e-01 8.19389582e-01 1.82383180e-01 -6.77804828e-01
-1.73293996e+00 -1.35253167e+00 6.44341767e-01 -6.12660944e-01
2.59863734e-01 -5.69156349e-01 -1.03655529e+00 5.24861395e-01
-1.74751058e-01 2.18374863e-01 4.59944099e-01 4.78010550e-02
-3.17907453e-01 -1.58898458e-01 -1.03395069e+00 7.14536071e-01
1.16722095e+00 -6.58062398e-01 -8.49579215e-01 5.29635429e-01
8.89696777e-01 -6.48394942e-01 -8.73242795e-01 7.83597708e-01
6.11693636e-02 -7.25852668e-01 1.58597839e+00 -2.28381276e-01
3.46050531e-01 -3.69936347e-01 -1.39067456e-01 -1.29345417e+00
-6.80184066e-01 -2.17220798e-01 -2.48321146e-02 9.06374514e-01
1.23081960e-01 -4.55392331e-01 1.35856748e+00 1.78121746e-01
-4.37270612e-01 -1.03308499e+00 -1.21806479e+00 -9.27119493e-01
4.15856719e-01 -9.10523355e-01 1.14363194e+00 8.83417666e-01
-5.80348611e-01 4.04672921e-02 4.00111675e-01 6.46108389e-01
6.80680454e-01 4.51106697e-01 7.94728875e-01 -1.35112417e+00
2.74705142e-01 -6.20799303e-01 -9.00504947e-01 -1.54123759e+00
3.16784739e-01 -1.30358207e+00 -1.44413516e-01 -1.83847487e+00
-4.91755158e-01 -6.81991220e-01 -9.13080852e-03 3.55190963e-01
6.06897734e-02 4.42397565e-01 2.95515418e-01 3.94641727e-01
-1.96564659e-01 9.06957150e-01 1.29500771e+00 -5.50692320e-01
-7.86716789e-02 4.17627037e-01 -3.41940820e-01 6.43994629e-01
7.37699389e-01 -2.72213519e-01 -2.98257947e-01 -1.05679965e+00
7.87090287e-02 -4.17825431e-01 5.81071258e-01 -1.13757849e+00
3.42076540e-01 1.16364487e-01 7.16312945e-01 -1.47285819e+00
5.80714941e-01 -1.14677620e+00 -1.26965612e-01 1.61990792e-01
7.04753399e-02 5.12855589e-01 4.83483523e-01 7.10793555e-01
-3.53955537e-01 -1.79405287e-02 4.09372985e-01 -2.24150807e-01
-1.04374635e+00 9.27879632e-01 -1.26126017e-02 -3.78719330e-01
8.33039641e-01 -6.35516405e-01 1.63083032e-01 -1.23535637e-02
-4.61644083e-01 3.53623301e-01 4.81355309e-01 6.95163727e-01
1.29871511e+00 -1.64946032e+00 -6.97931528e-01 5.92566073e-01
3.22106510e-01 6.05051696e-01 -7.72309080e-02 3.05658102e-01
-8.99089336e-01 5.42943001e-01 -1.52065665e-01 -1.36315846e+00
-1.08470976e+00 4.68392670e-01 1.44290492e-01 2.97784448e-01
-1.21069598e+00 9.72309351e-01 8.43636319e-02 -8.97062719e-01
3.15629207e-02 -8.54712963e-01 5.83666191e-03 -4.95816022e-01
2.32999846e-01 2.25117266e-01 3.91163290e-01 -5.68344474e-01
-4.06100303e-01 1.02529740e+00 1.02693170e-01 1.70731440e-01
1.39830184e+00 2.68149108e-01 -4.76274014e-01 4.03821766e-01
1.41750467e+00 -2.08399579e-01 -1.09179521e+00 -3.23646963e-01
2.58124378e-02 -7.72713304e-01 1.39388323e-01 -2.82624424e-01
-9.48256135e-01 1.04093134e+00 4.41284060e-01 7.30383918e-02
7.19782770e-01 1.70886278e-01 9.42755818e-01 6.22864664e-01
6.06554925e-01 -6.76280200e-01 -6.04433566e-02 9.48316932e-01
1.07259345e+00 -1.23786211e+00 -3.36496793e-02 -7.36559331e-01
-3.53115648e-01 1.28065610e+00 4.18294549e-01 -6.06733263e-01
1.06479907e+00 1.95982322e-01 3.87404971e-02 -5.83353698e-01
-4.28991914e-01 -9.69252586e-02 6.09840274e-01 9.26246345e-01
7.53303478e-03 2.85069048e-01 2.78832316e-01 1.28100559e-01
-7.17943311e-01 -2.33000398e-01 -1.08654099e-02 8.41664612e-01
-3.10892135e-01 -1.04703689e+00 -4.01136398e-01 7.35978723e-01
-5.84571287e-02 -4.19035815e-02 -4.37088668e-01 7.27759421e-01
1.07012011e-01 3.90234858e-01 5.59245288e-01 -2.90057391e-01
4.47941065e-01 -1.66267902e-01 4.57190275e-01 -7.01862335e-01
-3.63380253e-01 -1.75799996e-01 -6.47013187e-01 -6.74000502e-01
-1.88221410e-01 -7.53407836e-01 -1.40144289e+00 -3.10235709e-01
-2.15036273e-02 -1.10663883e-01 9.01572824e-01 7.35596120e-01
6.63095176e-01 2.70227283e-01 5.19106269e-01 -1.24300289e+00
-4.31607097e-01 -6.31708324e-01 -4.43548560e-01 2.22076803e-01
5.04201353e-01 -5.91854930e-01 6.66270927e-02 -3.00095141e-01] | [7.941089153289795, -3.619476079940796] |
48eebc34-eaff-44ac-b4d7-b4f26857d11d | learning-off-road-terrain-traversability-with | 2305.18896 | null | https://arxiv.org/abs/2305.18896v1 | https://arxiv.org/pdf/2305.18896v1.pdf | Learning Off-Road Terrain Traversability with Self-Supervisions Only | Estimating the traversability of terrain should be reliable and accurate in diverse conditions for autonomous driving in off-road environments. However, learning-based approaches often yield unreliable results when confronted with unfamiliar contexts, and it is challenging to obtain manual annotations frequently for new circumstances. In this paper, we introduce a method for learning traversability from images that utilizes only self-supervision and no manual labels, enabling it to easily learn traversability in new circumstances. To this end, we first generate self-supervised traversability labels from past driving trajectories by labeling regions traversed by the vehicle as highly traversable. Using the self-supervised labels, we then train a neural network that identifies terrains that are safe to traverse from an image using a one-class classification algorithm. Additionally, we supplement the limitations of self-supervised labels by incorporating methods of self-supervised learning of visual representations. To conduct a comprehensive evaluation, we collect data in a variety of driving environments and perceptual conditions and show that our method produces reliable estimations in various environments. In addition, the experimental results validate that our method outperforms other self-supervised traversability estimation methods and achieves comparable performances with supervised learning methods trained on manually labeled data. | ['Inwook Shim', 'Sungdae Sim', 'Junwon Seo'] | 2023-05-30 | null | null | null | null | ['one-class-classification'] | ['miscellaneous'] | [ 2.25446001e-01 2.11743005e-02 -4.66184139e-01 -9.95865583e-01
-7.03008294e-01 -6.27010405e-01 4.72881675e-01 2.29175851e-01
-4.11286175e-01 8.37084711e-01 -1.47098631e-01 -5.42768776e-01
-1.05564278e-02 -1.05372036e+00 -8.48356366e-01 -4.11889404e-01
-3.85918498e-01 3.36303413e-01 5.76859772e-01 -3.22171092e-01
2.90777832e-01 5.08379579e-01 -2.25445294e+00 1.03878453e-01
1.26650190e+00 8.88725102e-01 2.47601807e-01 6.60212398e-01
8.77623111e-02 9.63088810e-01 -3.94580394e-01 -2.03790903e-01
3.39826256e-01 -1.70291271e-02 -8.79866004e-01 1.20276652e-01
8.75514746e-01 -4.40247953e-01 -4.66992736e-01 8.70981574e-01
-1.99908957e-01 3.52084339e-01 1.01324332e+00 -1.46859431e+00
-2.54235417e-01 1.96953565e-01 -2.91173548e-01 4.62103516e-01
2.84601837e-01 2.94859588e-01 8.23434174e-01 -5.75725675e-01
5.05685270e-01 9.83143568e-01 7.18720138e-01 1.19159207e-01
-9.63402748e-01 -8.34252536e-01 5.35929620e-01 7.34645903e-01
-1.44058585e+00 -5.65561056e-01 7.69569516e-01 -6.57596588e-01
8.82298708e-01 6.68289140e-02 6.45736992e-01 9.42039490e-01
3.64232600e-01 8.84847641e-01 1.17178881e+00 -1.60319917e-02
3.29031080e-01 3.11333746e-01 1.71905162e-03 9.21483636e-01
2.01056227e-01 6.41016006e-01 -3.27454060e-01 4.24781680e-01
4.03827220e-01 -2.80876786e-01 3.39361995e-01 -6.21881783e-01
-1.19856620e+00 6.86941862e-01 6.72926962e-01 -2.54289687e-01
-3.84934098e-02 1.42323866e-01 4.94782329e-01 3.06596071e-01
5.82695723e-01 3.07214200e-01 -5.06017134e-02 -3.33054096e-01
-1.01645041e+00 2.20823601e-01 5.28142989e-01 1.30589497e+00
1.22885489e+00 2.66457766e-01 1.45219386e-01 5.22362053e-01
-1.50655408e-03 7.97624588e-01 1.30737841e-01 -9.65994179e-01
5.01596987e-01 3.40800971e-01 3.06710988e-01 -1.09796476e+00
-5.45269132e-01 -3.29563648e-01 -4.61033762e-01 5.31119645e-01
3.00935775e-01 -1.01082528e-03 -1.11119878e+00 1.39501929e+00
2.40583614e-01 1.07792787e-01 2.81190664e-01 7.31734335e-01
6.20551169e-01 6.89419448e-01 2.44428694e-01 1.54392257e-01
8.16784680e-01 -1.34674144e+00 -6.30629778e-01 -5.54980874e-01
8.31630886e-01 -2.98260093e-01 1.08415437e+00 2.99619049e-01
-5.08141875e-01 -1.08985722e+00 -1.25761616e+00 1.00459933e-01
-7.46037424e-01 1.53954983e-01 6.82684958e-01 5.76040566e-01
-9.57951844e-01 5.32573998e-01 -8.14421356e-01 -3.84660810e-01
4.73569810e-01 -5.70110679e-02 -2.56507128e-01 3.40175838e-03
-1.22361600e+00 1.25034642e+00 5.39205015e-01 -1.36984289e-02
-1.42552793e+00 -2.09280029e-01 -1.41767013e+00 -3.83460552e-01
4.05670166e-01 -4.97321673e-02 1.31818724e+00 -7.56598115e-01
-1.21647143e+00 7.59560466e-01 -3.66778642e-01 -6.19061768e-01
6.54881775e-01 -1.94927350e-01 -7.11966932e-01 7.30542839e-02
6.34890497e-01 9.50808108e-01 8.96413565e-01 -1.57513094e+00
-1.28678107e+00 4.90671061e-02 2.65157968e-01 5.08130431e-01
-7.88933486e-02 -7.33177960e-01 -4.15075757e-02 -1.50055900e-01
1.43909439e-01 -8.98838222e-01 -3.25848103e-01 3.05191614e-02
-3.23242366e-01 -2.25560188e-01 1.03907037e+00 -2.87682116e-01
9.25278962e-01 -2.19073987e+00 -3.99423659e-01 3.05907130e-01
1.51986331e-01 6.54151887e-02 1.05400616e-02 2.17150062e-01
2.60804027e-01 -1.29522830e-01 -2.76590556e-01 -4.32100073e-02
-1.09616436e-01 4.34075058e-01 -4.05428320e-01 2.95117438e-01
2.60789841e-01 7.42649198e-01 -1.51994693e+00 -7.99625099e-01
6.00460172e-01 -1.31050065e-01 -1.94867581e-01 2.94302553e-01
-7.62666808e-03 7.50503242e-01 -4.91426855e-01 7.26029575e-01
6.17388308e-01 1.36177734e-01 -1.52566776e-01 -1.31436452e-01
-4.84107137e-01 2.66208142e-01 -9.01025891e-01 1.45827639e+00
-7.28319287e-01 1.08163285e+00 -6.25003695e-01 -1.11612284e+00
1.27896070e+00 -2.32639581e-01 1.94055527e-01 -9.12405431e-01
-1.27984256e-01 1.88771933e-01 -3.58453184e-01 -1.02623475e+00
8.00060391e-01 6.75702244e-02 -4.07927513e-01 9.69098955e-02
-1.39941975e-01 -5.68418503e-01 4.95674759e-01 -8.02252069e-03
7.95306623e-01 5.53601563e-01 1.75639600e-01 -1.11901261e-01
3.23188663e-01 6.20221436e-01 4.80053067e-01 8.30438495e-01
-6.81396127e-01 7.69843087e-02 9.43754390e-02 -9.19917047e-01
-1.01858389e+00 -1.32310486e+00 -4.40752149e-01 1.08033800e+00
8.68014574e-01 -2.38636881e-01 -5.47232211e-01 -8.63348842e-01
8.68649483e-02 8.49103153e-01 -5.44498444e-01 -2.77107745e-01
-4.19182658e-01 -2.48182297e-01 5.65571070e-01 8.78722489e-01
8.73048246e-01 -9.49475050e-01 -9.50276613e-01 1.74880505e-01
-3.68148744e-01 -1.39920616e+00 -2.29975414e-02 3.45030010e-01
-6.61763608e-01 -1.00510919e+00 -1.91336542e-01 -1.04716814e+00
8.02911699e-01 5.70592821e-01 8.69624138e-01 -1.48308247e-01
-1.99454263e-01 1.24511644e-01 -2.96928048e-01 -4.29112703e-01
-3.44893456e-01 3.00342888e-01 2.13687032e-01 -2.10605949e-01
4.61237729e-01 -2.94047683e-01 -4.31941122e-01 7.62586474e-01
-2.51445979e-01 3.23864251e-01 5.11494577e-01 5.24491668e-01
6.63181305e-01 6.13070786e-01 6.84852600e-01 -8.21440756e-01
2.23522216e-01 -6.49652719e-01 -6.56424522e-01 -3.18363197e-02
-6.00994945e-01 -2.08332259e-02 7.31719971e-01 -3.31235826e-01
-1.25722456e+00 3.77790838e-01 -1.15760133e-01 -3.10710957e-03
-6.78616643e-01 3.18430364e-01 2.28396952e-02 -1.84584334e-01
1.06495929e+00 7.67765567e-02 -2.40888789e-01 7.94111192e-02
5.65707266e-01 8.64929855e-01 7.08251715e-01 -5.09154737e-01
1.18951356e+00 7.01884210e-01 -1.57764956e-01 -7.87141383e-01
-1.31724143e+00 -5.99097073e-01 -1.02943075e+00 -7.88833380e-01
7.41259396e-01 -1.04122460e+00 -3.76262337e-01 4.10703927e-01
-4.22398567e-01 -7.86151648e-01 -3.22868079e-01 6.03597879e-01
-8.20517600e-01 2.95139879e-01 -1.18679397e-01 -8.73060107e-01
3.75652760e-01 -1.12703216e+00 9.69034851e-01 3.73650610e-01
-2.65875697e-01 -1.09511054e+00 -2.64584459e-02 4.22195256e-01
3.73536378e-01 3.47719461e-01 5.04654229e-01 -1.38133988e-01
-5.81976831e-01 -2.08945885e-01 -3.50415409e-01 1.49543107e-01
3.69380444e-01 -1.16724141e-01 -1.15158284e+00 -1.25418589e-01
-6.27793908e-01 -6.33535385e-01 9.99989092e-01 1.08918533e-01
1.31918406e+00 -9.20076221e-02 -6.44904971e-01 6.14427984e-01
1.13187230e+00 3.01816612e-01 4.75501806e-01 7.02871203e-01
7.84773111e-01 9.57412839e-01 1.32401097e+00 5.55585586e-02
1.03066957e+00 3.29477996e-01 5.55069029e-01 -4.01771426e-01
9.64180455e-02 -6.46995902e-01 2.28651732e-01 4.11893159e-01
-3.09438054e-02 -1.13125309e-01 -1.09157944e+00 8.76325250e-01
-1.89619040e+00 -9.54229891e-01 -1.78350687e-01 2.12730455e+00
5.79301953e-01 6.77673757e-01 2.36066535e-01 2.89244652e-01
4.40871745e-01 2.09726572e-01 -9.65117693e-01 -5.11881888e-01
2.01812685e-01 -3.62032264e-01 9.35136378e-01 5.90803444e-01
-1.59762394e+00 1.37022269e+00 7.50262308e+00 4.77714837e-01
-9.49686289e-01 -8.28067958e-02 5.09084404e-01 1.94906801e-01
-1.31075412e-01 1.29743842e-02 -7.65049875e-01 1.81339517e-01
9.86388385e-01 -1.71533264e-02 1.96301550e-01 1.14684474e+00
3.09873909e-01 -4.10978645e-01 -1.08987486e+00 6.71886265e-01
7.99691305e-02 -8.72393548e-01 -1.93637893e-01 -2.80178487e-01
8.54707837e-01 6.02659397e-02 1.13225833e-01 5.94381988e-01
5.58249414e-01 -1.02638686e+00 9.33263421e-01 4.10045296e-01
1.00126517e+00 -7.88217783e-01 6.67802989e-01 4.84806955e-01
-1.56094742e+00 -2.49579132e-01 -4.25328135e-01 -3.96979451e-01
3.03864986e-01 3.92695785e-01 -9.15652871e-01 3.74637991e-01
9.38130736e-01 1.18757808e+00 -7.38612235e-01 1.06446123e+00
-5.10162055e-01 6.15237415e-01 -1.21222988e-01 -5.76656424e-02
5.43888271e-01 -3.87166888e-02 3.59095514e-01 1.28247452e+00
9.03423280e-02 -2.54384995e-01 4.86699790e-01 6.50599480e-01
4.26151633e-01 -2.36968901e-02 -1.24531126e+00 3.49170625e-01
5.16647518e-01 1.12291062e+00 -6.89297736e-01 -6.00024641e-01
-2.48319492e-01 7.48588383e-01 5.89610517e-01 6.01303637e-01
-1.00091183e+00 -5.97850144e-01 4.93828952e-01 2.52995342e-01
-1.89256277e-02 -5.35011947e-01 -2.30578229e-01 -8.15250337e-01
2.69678775e-02 -3.09854031e-01 1.59290478e-01 -8.20213735e-01
-9.99359071e-01 8.01090658e-01 4.77247924e-01 -1.79953241e+00
-2.13685125e-01 -4.92974639e-01 -4.96050060e-01 3.14828515e-01
-2.15313983e+00 -1.02812445e+00 -8.57002616e-01 2.47254178e-01
7.72861540e-01 -1.07121497e-01 5.36828995e-01 1.07740410e-01
-3.19956124e-01 5.22854328e-01 1.14092547e-02 -9.49791744e-02
6.89875901e-01 -1.34151042e+00 7.25176275e-01 7.36849844e-01
-2.70258114e-02 8.24407861e-02 6.02378726e-01 -7.71824777e-01
-7.91286826e-01 -1.57743073e+00 6.96979284e-01 -4.79754269e-01
6.10706687e-01 -3.52811962e-01 -8.92559588e-01 8.20486903e-01
-1.50312319e-01 3.76722999e-02 4.14539069e-01 2.34817695e-02
-2.43936032e-01 -3.24687183e-01 -1.04567945e+00 6.54401541e-01
1.45200789e+00 -4.85657185e-01 -5.25884211e-01 4.76172656e-01
3.96260887e-01 -5.99635899e-01 -6.06186569e-01 5.93423188e-01
5.28282702e-01 -8.57443213e-01 8.19907963e-01 -3.18019092e-01
1.99630097e-01 -5.32483935e-01 4.22569551e-02 -1.42094827e+00
-3.46827239e-01 -1.75266549e-01 1.59674317e-01 7.14692175e-01
6.57673240e-01 -6.80411696e-01 9.02012289e-01 2.73092002e-01
-4.34687316e-01 -4.64818686e-01 -8.04136276e-01 -1.06711686e+00
9.64460336e-03 -5.71202934e-01 6.41467154e-01 7.15152442e-01
1.99154794e-01 3.61005217e-02 -4.55212921e-01 3.37569088e-01
8.84251356e-01 9.64535996e-02 1.00099742e+00 -1.19024169e+00
5.44141531e-01 -3.23908962e-02 -9.02843058e-01 -1.30055261e+00
5.25805593e-01 -8.89822125e-01 7.54172027e-01 -1.68420744e+00
-5.88054843e-02 -8.68494272e-01 -7.48290122e-02 6.66344523e-01
-9.32767391e-02 3.83774906e-01 -4.80699420e-01 2.99383968e-01
-9.39335644e-01 5.55666983e-01 1.30238616e+00 -4.82697695e-01
-2.15713292e-01 1.62665676e-02 -3.34170431e-01 9.33534801e-01
1.14642906e+00 -3.24360549e-01 -8.66174817e-01 -4.37758386e-01
1.01259902e-01 -3.13885242e-01 2.30666116e-01 -1.43788707e+00
1.78922489e-01 -5.10111272e-01 3.49841595e-01 -8.79809678e-01
1.78929359e-01 -7.83792436e-01 -2.86983758e-01 2.78292686e-01
-3.68564159e-01 -1.60884317e-02 4.05019283e-01 7.48278558e-01
-1.92782372e-01 -1.64873093e-01 5.59247613e-01 1.03212427e-02
-1.62354016e+00 2.10290700e-01 -1.08709979e+00 1.17163971e-01
1.31552768e+00 -6.63890898e-01 -1.14086121e-01 -5.09017408e-01
-6.54817939e-01 7.60894775e-01 5.00676274e-01 7.18520045e-01
8.11960757e-01 -1.32264256e+00 -5.30111730e-01 4.47081894e-01
8.17926049e-01 1.03379823e-01 1.55105993e-01 3.09557438e-01
-6.40961528e-01 1.94240317e-01 -4.08118874e-01 -9.83595669e-01
-7.97234714e-01 5.01470268e-01 3.07119757e-01 6.58205822e-02
-8.51335347e-01 3.53291363e-01 5.81424870e-02 -7.75894523e-01
2.74676144e-01 -4.62527961e-01 -5.79541028e-01 -1.63617358e-01
2.68643856e-01 3.95950645e-01 -2.30594240e-02 -1.03128052e+00
-1.68927476e-01 7.48099089e-01 1.56032443e-02 -1.59347132e-02
7.13007033e-01 -4.04177487e-01 6.56145871e-01 8.75667691e-01
9.99276459e-01 -4.17468369e-01 -1.82470572e+00 -1.93042368e-01
-8.47564563e-02 -6.69491589e-01 9.26847756e-02 -5.11025250e-01
-7.21194327e-01 8.35629702e-01 7.28551090e-01 2.59601790e-02
7.47623086e-01 -2.02753007e-01 8.42462659e-01 7.44273722e-01
7.87406683e-01 -1.31340384e+00 2.14245498e-01 4.98102456e-01
5.08956909e-01 -1.74974310e+00 -2.41093904e-01 -5.78171551e-01
-8.89728129e-01 9.36299026e-01 9.75104928e-01 1.45007268e-01
6.37747824e-01 2.26929709e-01 5.41678190e-01 -2.17102662e-01
-4.10962313e-01 -4.89334673e-01 3.51676047e-01 1.00446880e+00
-1.19807698e-01 2.67373592e-01 3.31998497e-01 -2.76939571e-01
-6.92932010e-01 -1.45530075e-01 5.18652022e-01 1.13126016e+00
-8.51486087e-01 -6.12190604e-01 -1.52735099e-01 6.56962752e-01
3.83489192e-01 2.75146395e-01 9.47504863e-02 7.45678186e-01
3.24767560e-01 1.30833066e+00 2.94643313e-01 -6.77079618e-01
4.70542610e-01 -1.89792946e-01 1.81050271e-01 -6.63628817e-01
1.20925047e-01 -6.31798565e-01 4.41233784e-01 -5.46951175e-01
-5.06628931e-01 -6.86365664e-01 -1.32878280e+00 -1.46945447e-01
-3.12331706e-01 9.25229564e-02 6.13868594e-01 9.86815989e-01
-2.69329213e-02 6.26219332e-01 9.13952529e-01 -1.02251995e+00
-2.01569021e-01 -5.90452969e-01 -4.43343937e-01 5.21267533e-01
5.22972286e-01 -1.16874862e+00 -4.99377638e-01 3.21843594e-01] | [8.170353889465332, -1.775512933731079] |
c97df95f-3785-4161-8094-f5083d98e03e | smart-a-situation-model-for-algebra-story | 2012.14011 | null | https://arxiv.org/abs/2012.14011v1 | https://arxiv.org/pdf/2012.14011v1.pdf | SMART: A Situation Model for Algebra Story Problems via Attributed Grammar | Solving algebra story problems remains a challenging task in artificial intelligence, which requires a detailed understanding of real-world situations and a strong mathematical reasoning capability. Previous neural solvers of math word problems directly translate problem texts into equations, lacking an explicit interpretation of the situations, and often fail to handle more sophisticated situations. To address such limits of neural solvers, we introduce the concept of a \emph{situation model}, which originates from psychology studies to represent the mental states of humans in problem-solving, and propose \emph{SMART}, which adopts attributed grammar as the representation of situation models for algebra story problems. Specifically, we first train an information extraction module to extract nodes, attributes, and relations from problem texts and then generate a parse graph based on a pre-defined attributed grammar. An iterative learning strategy is also proposed to improve the performance of SMART further. To rigorously study this task, we carefully curate a new dataset named \emph{ASP6.6k}. Experimental results on ASP6.6k show that the proposed model outperforms all previous neural solvers by a large margin while preserving much better interpretability. To test these models' generalization capability, we also design an out-of-distribution (OOD) evaluation, in which problems are more complex than those in the training set. Our model exceeds state-of-the-art models by 17\% in the OOD evaluation, demonstrating its superior generalization ability. | ['Song-Chun Zhu', 'Siyuan Huang', 'Daniel Ciao', 'Ran Gong', 'Qing Li', 'Yining Hong'] | 2020-12-27 | null | null | null | null | ['mathematical-reasoning'] | ['natural-language-processing'] | [ 2.21741334e-01 2.33376533e-01 9.56188887e-02 -5.17459452e-01
-2.47474387e-01 -4.97524261e-01 3.27117980e-01 1.40326470e-01
-2.93799281e-01 4.59411830e-01 7.16852471e-02 -4.97405797e-01
-4.90003645e-01 -1.30837452e+00 -6.58898711e-01 -2.28505805e-01
1.91917002e-01 6.70379698e-01 -5.91481216e-02 -3.91844571e-01
2.49395713e-01 6.85547590e-02 -1.23250830e+00 5.58570027e-01
1.45258617e+00 8.37535739e-01 3.01784903e-01 1.01485781e-01
-4.23658788e-01 1.08644831e+00 -7.88217604e-01 -7.43981719e-01
7.11036101e-02 -4.01229233e-01 -1.04721427e+00 -1.15351744e-01
3.44198793e-01 -3.31780761e-01 -4.57599938e-01 1.19710386e+00
1.63615018e-01 2.16995955e-01 5.48454285e-01 -1.16512156e+00
-9.31447148e-01 1.12967479e+00 -1.27053052e-01 1.22928657e-01
6.06534243e-01 1.98194355e-01 1.09287310e+00 -5.53725839e-01
3.92643273e-01 1.12525570e+00 5.65409124e-01 5.97212553e-01
-1.06478810e+00 -6.94102347e-01 3.06497961e-01 7.07046986e-01
-1.37711000e+00 4.88865227e-02 9.32899952e-01 -1.02894939e-01
1.05252385e+00 3.34782809e-01 7.24294722e-01 1.01450312e+00
-6.78328425e-02 9.14120138e-01 1.03612339e+00 -3.34488004e-01
3.23007882e-01 1.86796840e-02 6.51077807e-01 7.31742144e-01
3.23598415e-01 -2.09400326e-01 -5.67498982e-01 1.76850393e-01
8.00045311e-01 -4.46697891e-01 -4.15237427e-01 -4.40652631e-02
-1.16992176e+00 7.91387737e-01 4.25072521e-01 2.76717752e-01
-2.34531790e-01 2.02694889e-02 1.53547287e-01 2.23417401e-01
4.67302799e-02 9.98539567e-01 -3.93927038e-01 -1.03216954e-01
-6.70421004e-01 7.10109711e-01 9.92612958e-01 1.02816951e+00
2.92539179e-01 9.10424590e-02 -1.38028324e-01 6.52444303e-01
9.55440924e-02 3.73000294e-01 5.45697927e-01 -1.11352026e+00
8.59367728e-01 1.06213164e+00 -2.44559243e-01 -1.11733115e+00
-4.40868974e-01 -6.87936127e-01 -9.10233855e-01 -3.88431042e-01
7.26291776e-01 -5.09732254e-02 -6.44307673e-01 1.85318279e+00
1.07961245e-01 3.08893383e-01 4.58862424e-01 9.14522052e-01
1.24574447e+00 8.76747966e-01 -2.27286071e-02 -1.44117579e-01
1.27390838e+00 -7.86231816e-01 -7.95420349e-01 -4.66374069e-01
8.32225144e-01 -1.15009569e-01 1.24537516e+00 7.11208224e-01
-1.20692670e+00 -3.42277110e-01 -1.04325771e+00 -2.03231856e-01
-3.31860662e-01 8.03442001e-02 1.03121746e+00 4.95891422e-01
-6.89451993e-01 5.30502737e-01 -3.95973951e-01 3.87962237e-02
4.67356473e-01 3.12915683e-01 -1.71976373e-01 -3.29401523e-01
-1.55085337e+00 1.02420473e+00 8.50898921e-01 1.44744217e-01
-6.43029988e-01 -9.67307866e-01 -1.24866056e+00 5.08419156e-01
7.29590178e-01 -6.22299910e-01 1.24098730e+00 -6.66879714e-01
-1.36723602e+00 7.61016548e-01 4.28724103e-03 -4.44770515e-01
4.52139154e-02 -1.81899413e-01 -2.72255659e-01 1.02339774e-01
-1.91366151e-01 3.11886966e-01 3.96448612e-01 -1.04309106e+00
-3.20264071e-01 -3.03886533e-01 7.61798739e-01 2.20314786e-01
-4.05664831e-01 -1.25751466e-01 -4.60094303e-01 -7.61073172e-01
3.95890296e-01 -5.50589263e-01 -1.87724739e-01 -3.70415121e-01
-2.50505716e-01 -6.36949301e-01 1.30914852e-01 -6.41660392e-01
1.47749710e+00 -1.93213546e+00 4.74599034e-01 2.71343708e-01
5.32818794e-01 3.23255539e-01 -1.91605166e-01 8.75811949e-02
-2.98651338e-01 7.99597129e-02 -3.26511890e-01 4.12268238e-03
3.68952006e-01 3.23760718e-01 -3.86704981e-01 -1.03100367e-01
3.29658866e-01 9.52828884e-01 -7.72775054e-01 -4.81604546e-01
9.20139700e-02 -7.63564259e-02 -8.92032623e-01 2.73054034e-01
-6.59490526e-01 -1.51573196e-01 -4.79511082e-01 3.37415218e-01
5.79360664e-01 -4.98701990e-01 3.65457237e-01 -5.84014431e-02
5.29558718e-01 4.85057354e-01 -1.37488508e+00 1.82959914e+00
-3.92189056e-01 3.59524101e-01 -3.54684711e-01 -1.32225084e+00
9.96562898e-01 1.52950481e-01 2.71403673e-03 -8.04713070e-01
2.43020773e-01 -1.33793354e-01 4.13498282e-01 -7.42600262e-01
2.93817014e-01 -1.48334861e-01 -2.70342380e-01 3.57767075e-01
4.32721488e-02 -3.53760779e-01 5.47785044e-01 4.89721060e-01
1.14172065e+00 3.63095962e-02 1.29276291e-01 -2.45044857e-01
6.67249024e-01 1.19361944e-01 7.96991169e-01 6.70755029e-01
9.63999182e-02 3.92239302e-01 9.54666913e-01 -5.80693662e-01
-6.64890289e-01 -9.51811135e-01 1.50235603e-02 7.16581166e-01
3.12260121e-01 -8.91038120e-01 -9.79954779e-01 -6.28779769e-01
-2.21123859e-01 1.38647377e+00 -4.21194702e-01 -3.22252572e-01
-7.71417677e-01 -9.41109002e-01 6.71490550e-01 6.17683887e-01
8.12458694e-01 -1.19971645e+00 -4.22660261e-01 1.49567053e-01
-5.36346734e-01 -1.41362405e+00 3.58562529e-01 7.49311736e-03
-6.54789925e-01 -1.14784241e+00 -4.60803211e-02 -7.89369226e-01
6.82789326e-01 -2.36779258e-01 1.32531714e+00 3.08559686e-01
-1.49727445e-02 2.96645388e-02 -3.40585262e-01 -4.65030551e-01
-3.40597361e-01 9.23299715e-02 -8.40683505e-02 -4.42848921e-01
6.82234406e-01 -7.51218259e-01 -4.88829315e-02 3.46709378e-02
-9.53455329e-01 4.90870386e-01 4.90781903e-01 5.98673582e-01
1.77166283e-01 6.31651938e-01 4.49403584e-01 -9.00766551e-01
9.30379391e-01 -4.28597063e-01 -6.71138346e-01 4.45410401e-01
-4.28785920e-01 1.78050458e-01 1.00783443e+00 -3.89488339e-01
-1.16146398e+00 -2.75410950e-01 5.58873080e-02 -1.61924258e-01
-3.38383198e-01 1.00676620e+00 -5.47881424e-01 3.62515181e-01
6.88833356e-01 3.79656881e-01 -2.19732299e-01 -2.80041099e-01
2.84831256e-01 3.32021058e-01 6.00371063e-01 -1.31860173e+00
8.90262783e-01 -1.72877192e-01 -9.05884951e-02 -5.76088190e-01
-1.23155594e+00 8.79680216e-02 -3.01210880e-01 -2.69916821e-02
6.26486719e-01 -6.83937907e-01 -9.77549016e-01 5.31257093e-01
-1.34985113e+00 -4.85976756e-01 -1.55994564e-01 4.94341582e-01
-4.35660034e-01 2.77979136e-01 -8.43061447e-01 -5.04476607e-01
-1.39639765e-01 -9.89625692e-01 6.01541698e-01 3.19011450e-01
-5.18817067e-01 -8.65332425e-01 -3.56030613e-01 7.61864007e-01
2.28926480e-01 7.11570829e-02 1.68566573e+00 -8.94079268e-01
-5.67479432e-01 -1.54924672e-02 -3.42115134e-01 2.99133629e-01
-3.01385909e-01 -3.88356954e-01 -7.72824764e-01 2.39900902e-01
4.36817288e-01 -4.85715091e-01 4.99493450e-01 6.88077062e-02
1.71403229e+00 -2.66787201e-01 4.13559005e-02 6.37447834e-01
1.19099867e+00 8.24823007e-02 9.96918023e-01 3.30567092e-01
5.25630713e-01 6.63398623e-01 3.28712612e-01 3.32768410e-01
6.42308176e-01 2.66670078e-01 3.66228908e-01 2.93990821e-01
1.89893752e-01 -1.85003728e-01 1.44181147e-01 9.21063721e-01
-1.44601256e-01 -3.21134210e-01 -1.21791768e+00 3.33814770e-01
-1.71737254e+00 -7.85569370e-01 -2.16406673e-01 1.73541808e+00
1.28186142e+00 4.21909362e-01 -3.73929560e-01 5.21424532e-01
3.76131445e-01 -1.06574073e-02 -3.87860954e-01 -2.59513348e-01
-4.74928413e-03 5.74731648e-01 -1.87633514e-01 4.15416420e-01
-8.65911663e-01 1.18979073e+00 5.80644321e+00 1.10813975e+00
-8.13010633e-01 -4.36588883e-01 4.09908712e-01 -4.51588035e-02
-3.78289104e-01 -1.63357645e-01 -6.66923463e-01 3.37154090e-01
7.34866023e-01 -2.91152835e-01 6.55554235e-01 8.00225914e-01
-7.14879781e-02 -1.40343457e-02 -1.48377442e+00 1.01119041e+00
2.79713303e-01 -1.28463924e+00 4.20989454e-01 -2.48573095e-01
3.30524415e-01 -7.78552651e-01 -1.59261048e-01 8.31951797e-01
2.72986174e-01 -1.48064446e+00 4.79196787e-01 4.86535162e-01
3.04043472e-01 -7.09243000e-01 7.88409173e-01 7.96602130e-01
-8.23472142e-01 -3.75373736e-02 -3.87296051e-01 -7.28716969e-01
-1.18676923e-01 4.29776132e-01 -5.86599469e-01 6.90309644e-01
4.78306234e-01 6.82140648e-01 -7.83534408e-01 7.47875154e-01
-7.03599155e-01 6.64459646e-01 -3.87200683e-01 -2.89006472e-01
1.73910469e-01 -2.36622497e-01 2.71040201e-01 8.38071644e-01
1.33318678e-01 7.83190012e-01 -1.16299056e-02 1.44279253e+00
-2.55049691e-02 -6.87576756e-02 -3.66232723e-01 -1.89387023e-01
3.98662120e-01 1.14344645e+00 -4.98778284e-01 -4.20256525e-01
-2.35894695e-01 6.14024699e-01 6.45934522e-01 4.38436359e-01
-1.15019810e+00 -4.52567607e-01 3.61029327e-01 -6.71852753e-03
2.13531591e-02 -2.86023766e-01 -5.85787117e-01 -1.36281681e+00
2.94010937e-01 -1.32579243e+00 3.14100862e-01 -9.91843045e-01
-1.26701629e+00 4.26166654e-01 4.50771302e-01 -6.92862809e-01
-1.77746817e-01 -9.88470614e-01 -6.21124983e-01 7.76779592e-01
-1.17379344e+00 -8.48748744e-01 -5.78370750e-01 4.76050049e-01
5.32179654e-01 -2.00254679e-01 8.62896502e-01 1.07419811e-01
-8.87943447e-01 5.73944569e-01 -5.72193503e-01 4.96723324e-01
1.51715562e-01 -1.31553185e+00 2.15351969e-01 8.68952870e-01
1.23313427e-01 8.85486722e-01 7.27354825e-01 -4.50226367e-01
-1.63612819e+00 -8.97699893e-01 8.48186791e-01 -5.05949557e-01
6.20652974e-01 -3.13432515e-01 -1.18680632e+00 8.22879434e-01
1.57422319e-01 -4.08523142e-01 6.90353215e-01 2.49444827e-01
-4.10077125e-01 1.63095258e-02 -9.93863940e-01 7.86285818e-01
1.25174379e+00 -1.72526136e-01 -1.46316075e+00 4.08220351e-01
8.48707676e-01 -7.16702700e-01 -8.56073201e-01 5.62640786e-01
7.70313814e-02 -7.99457371e-01 9.50841486e-01 -8.82037103e-01
1.10638678e+00 -1.22985996e-01 -1.26214296e-01 -1.29777360e+00
-3.41413707e-01 -3.08319002e-01 -2.85262495e-01 1.17206967e+00
4.79826719e-01 -3.94208372e-01 9.11675572e-01 1.06555855e+00
-2.12070480e-01 -9.88015592e-01 -6.12805307e-01 -6.91837847e-01
3.36381465e-01 -8.90998304e-01 8.29943895e-01 1.25273299e+00
4.93434876e-01 5.71696222e-01 1.65570140e-01 2.83273727e-01
5.87696671e-01 3.01668465e-01 5.79770625e-01 -1.17560136e+00
-4.24824238e-01 -6.13059998e-01 -1.40210867e-01 -1.07527840e+00
5.63022256e-01 -1.13869870e+00 -1.32242784e-01 -1.68217230e+00
1.59954473e-01 -4.29476291e-01 -1.12307511e-01 6.36863530e-01
-3.46220046e-01 -2.56824702e-01 1.66629046e-01 -2.60588855e-01
-6.81793332e-01 6.30094945e-01 1.35431075e+00 -1.58554941e-01
1.24776661e-01 -3.39261264e-01 -8.31714869e-01 1.05355716e+00
7.62163341e-01 -1.55411661e-01 -6.26321852e-01 -6.96203768e-01
6.01416171e-01 1.36222750e-01 3.54616165e-01 -1.07432806e+00
2.75023758e-01 -4.79373604e-01 3.35892528e-01 -3.27039868e-01
3.40510994e-01 -7.04344630e-01 -1.90966681e-01 2.85485357e-01
-4.77396458e-01 9.91126522e-02 2.58640110e-01 1.09238729e-01
-1.98411331e-01 -5.79349279e-01 3.31655979e-01 -2.39547327e-01
-8.40809584e-01 -2.82601546e-02 -3.14042032e-01 4.96694803e-01
1.01739550e+00 3.73288803e-02 -4.40326720e-01 -3.34472775e-01
-6.19823754e-01 5.10395885e-01 4.41771466e-03 2.23807916e-01
7.82201231e-01 -1.24023640e+00 -7.30418503e-01 2.41297156e-01
-6.94341213e-02 6.54912293e-01 2.09072560e-01 4.19009238e-01
-8.45943511e-01 2.94507146e-01 -5.78753985e-02 -2.96502411e-01
-9.46099281e-01 5.42589664e-01 4.35862809e-01 -4.74802822e-01
-6.44744396e-01 9.01547015e-01 3.31138730e-01 -5.97042859e-01
4.78648841e-01 -7.83852220e-01 -4.41934228e-01 -3.24207187e-01
5.68437755e-01 9.67429429e-02 -8.84430707e-02 -2.66561471e-02
-2.20001280e-01 3.16197127e-01 -1.42921031e-01 2.10742518e-01
1.36392963e+00 5.63206553e-01 -3.53575617e-01 3.73367108e-02
4.89963204e-01 -3.91926408e-01 -5.37628949e-01 -3.80940318e-01
9.05284584e-02 -3.05106759e-01 -5.87990694e-02 -9.80510354e-01
-1.00730884e+00 8.18837285e-01 -2.02672884e-01 1.59490228e-01
1.10602558e+00 -1.41202644e-01 7.78446078e-01 8.67720962e-01
2.13567659e-01 -9.90704596e-01 2.20682085e-01 9.35560405e-01
1.01913977e+00 -9.24998581e-01 -3.28695402e-02 -9.41072583e-01
-5.08889198e-01 1.16171575e+00 1.25809228e+00 -1.72280133e-01
1.62440211e-01 3.33722889e-01 -1.86978668e-01 -3.31983566e-01
-8.01262856e-01 1.46022052e-01 2.68292904e-01 3.39750916e-01
3.39258909e-01 -1.22840196e-01 -2.89148837e-01 1.42767394e+00
-8.10358524e-01 1.08217351e-01 5.02332449e-01 7.03450263e-01
-1.75370708e-01 -8.51056814e-01 -3.24191391e-01 4.15731192e-01
-2.20872417e-01 -3.97993505e-01 -4.86388475e-01 9.13457632e-01
2.51960129e-01 7.95200467e-01 3.12765166e-02 -3.02968591e-01
4.21082288e-01 2.17260882e-01 8.34844589e-01 -6.42712533e-01
-6.04194820e-01 -5.80851316e-01 3.23206782e-01 -5.65115333e-01
-9.18986723e-02 -3.44498396e-01 -1.50580239e+00 -5.65860569e-01
-1.73800915e-01 1.78387597e-01 2.33074278e-01 1.36264896e+00
-4.98853102e-02 8.84753883e-01 -1.74448211e-02 -2.39399448e-01
-8.29663694e-01 -8.93590391e-01 -3.36716831e-01 6.49224818e-01
-3.55631262e-01 -4.13943172e-01 -3.56521398e-01 -1.30653247e-01] | [9.649306297302246, 7.485450267791748] |
706897f0-d759-4c93-bfe7-c1cfd4f1bb2b | solving-a-new-3d-bin-packing-problem-with | 1708.05930 | null | http://arxiv.org/abs/1708.05930v1 | http://arxiv.org/pdf/1708.05930v1.pdf | Solving a New 3D Bin Packing Problem with Deep Reinforcement Learning Method | In this paper, a new type of 3D bin packing problem (BPP) is proposed, in
which a number of cuboid-shaped items must be put into a bin one by one
orthogonally. The objective is to find a way to place these items that can
minimize the surface area of the bin. This problem is based on the fact that
there is no fixed-sized bin in many real business scenarios and the cost of a
bin is proportional to its surface area. Our research shows that this problem
is NP-hard. Based on previous research on 3D BPP, the surface area is
determined by the sequence, spatial locations and orientations of items. Among
these factors, the sequence of items plays a key role in minimizing the surface
area. Inspired by recent achievements of deep reinforcement learning (DRL)
techniques, especially Pointer Network, on combinatorial optimization problems
such as TSP, a DRL-based method is applied to optimize the sequence of items to
be packed into the bin. Numerical results show that the method proposed in this
paper achieve about 5% improvement than heuristic method. | ['Xiaodong Zhang', 'Yinghui Xu', 'Haoyuan Hu', 'Xiaowei Yan', 'Longfei Wang'] | 2017-08-20 | null | null | null | null | ['3d-bin-packing'] | ['miscellaneous'] | [-4.74089056e-01 -2.53157198e-01 -4.80360001e-01 -5.21834679e-02
3.16926360e-01 -4.79768902e-01 -4.34146374e-01 2.90172964e-01
-3.05661559e-01 1.01631832e+00 -1.09147735e-01 -5.32780051e-01
-5.77093899e-01 -1.32334149e+00 -8.00115526e-01 -7.17150569e-01
-4.42755789e-01 9.96022880e-01 2.11410195e-01 -6.05407476e-01
7.38181293e-01 8.97824526e-01 -1.15070748e+00 6.50744662e-02
8.97890270e-01 1.29699719e+00 6.19447827e-01 2.24852160e-01
-5.69051921e-01 7.24317729e-02 -7.88787842e-01 -2.68161744e-01
6.04821324e-01 -2.59978563e-01 -8.90081286e-01 2.59112716e-01
-3.42858732e-01 -1.57961145e-01 -2.19600752e-01 9.76621449e-01
2.62116998e-01 2.37701729e-01 6.51961386e-01 -1.45661712e+00
-7.14337230e-01 7.60867476e-01 -8.45444918e-01 4.46436554e-01
2.15125486e-01 -1.33420140e-01 1.08796930e+00 -1.93735957e-01
7.97221363e-02 1.06225646e+00 3.75082850e-01 1.31318405e-01
-1.00607908e+00 -4.36844289e-01 2.44672522e-01 5.43257654e-01
-1.15270150e+00 3.47237051e-01 6.18736625e-01 -6.19620048e-02
1.00961816e+00 2.82971025e-01 1.26443577e+00 1.86402082e-01
6.46557033e-01 5.84500670e-01 6.97168350e-01 -6.29956663e-01
3.93474758e-01 -2.12695256e-01 1.25904962e-01 3.78035039e-01
9.57773685e-01 2.36890335e-02 -1.22752592e-01 3.22838247e-01
8.09016466e-01 2.00132668e-01 1.77745670e-02 -8.28358710e-01
-6.61254168e-01 1.17433035e+00 8.35073113e-01 2.57292211e-01
-3.68889391e-01 1.00767270e-01 1.22805245e-01 -1.36692658e-01
-1.06570028e-01 7.70909250e-01 -2.92452216e-01 -1.80126429e-02
-6.25144362e-01 6.62866592e-01 1.11081910e+00 1.29526937e+00
5.25495470e-01 -2.19469503e-01 -1.42225966e-01 5.63216865e-01
3.94758910e-01 4.52629119e-01 1.61843970e-01 -4.59121436e-01
9.27529097e-01 6.40775919e-01 3.66964757e-01 -1.38382804e+00
-6.37010872e-01 -2.41155744e-01 -7.60637462e-01 -1.26075119e-01
3.69301200e-01 -8.57447237e-02 -9.57858264e-01 1.09427881e+00
2.77952164e-01 -3.19852889e-01 -2.43829742e-01 1.17779148e+00
3.86523217e-01 1.02196217e+00 -3.90179634e-01 -2.55135417e-01
1.17714846e+00 -1.10729539e+00 -7.38601923e-01 -1.81594610e-01
1.01808660e-01 -6.20303094e-01 3.72198939e-01 3.87288660e-01
-1.41500592e+00 -3.43211472e-01 -1.29435194e+00 4.42768902e-01
-7.06326425e-01 -6.14445508e-01 6.75319433e-01 7.52163291e-01
-6.00890756e-01 6.94368243e-01 -5.36856711e-01 -5.30052036e-02
3.43662441e-01 7.16369569e-01 5.03500775e-02 -3.36089253e-01
-1.18157017e+00 1.12424874e+00 8.16215038e-01 5.18597066e-01
-4.67438310e-01 -1.31089002e-01 -7.68315136e-01 4.06754434e-01
6.24898612e-01 -2.85274714e-01 1.02620065e+00 -2.42668495e-01
-1.34590077e+00 2.75474757e-01 2.25562565e-02 -5.50360322e-01
9.30893943e-02 6.32875636e-02 -1.08264290e-01 2.90404446e-02
1.37196958e-01 4.18153763e-01 3.29631001e-01 -1.10306931e+00
-1.06863117e+00 -3.80080760e-01 3.51224899e-01 4.28191930e-01
-1.52532458e-01 -4.42406237e-01 -6.09898448e-01 -3.87265794e-02
2.19851196e-01 -6.51244462e-01 -6.04317844e-01 -6.81586564e-01
-3.89965236e-01 -4.35591668e-01 2.74009436e-01 -3.09646547e-01
1.27878702e+00 -1.72612965e+00 4.74913567e-01 6.13300323e-01
7.85826221e-02 1.78836703e-01 7.37140700e-02 5.94822407e-01
1.01837263e-01 1.83168143e-01 2.38283902e-01 3.57971996e-01
3.80953014e-01 5.37026525e-01 6.14742264e-02 3.62887532e-01
-2.44739410e-02 7.49882221e-01 -8.51261020e-01 -4.26364541e-01
2.53923297e-01 -3.51309299e-01 -7.77283430e-01 1.47852927e-01
-2.28736043e-01 -2.87924975e-01 -4.58960623e-01 7.51498342e-01
1.26873040e+00 2.87517700e-02 1.77458495e-01 -1.40259802e-01
-3.73945415e-01 3.71734351e-01 -1.34149969e+00 1.18380296e+00
-1.24720549e-02 7.92562217e-02 -4.28383201e-02 -1.59097600e+00
1.10359740e+00 -2.32026055e-01 5.39429188e-01 -1.10028315e+00
2.38807082e-01 2.95988530e-01 3.71419907e-01 -5.51559210e-01
7.36614645e-01 -1.60064861e-01 -2.24070847e-01 1.20186217e-01
-3.68488848e-01 -2.67327160e-01 9.29030716e-01 -2.48448864e-01
7.12493122e-01 -2.33869299e-01 2.44685262e-01 -2.36088112e-01
2.07351610e-01 3.39085191e-01 8.21522355e-01 7.59800255e-01
-1.48286551e-01 1.31873325e-01 7.71163523e-01 -6.06686473e-01
-1.11113942e+00 -8.23936939e-01 -2.44735509e-01 5.23961186e-01
1.03188038e+00 1.35712668e-01 -5.38226902e-01 -2.96723604e-01
5.38699567e-01 6.34573579e-01 -4.87265170e-01 -2.24804915e-02
-7.02565372e-01 -6.38212502e-01 -3.32671165e-01 5.00153899e-01
4.61892128e-01 -1.12936747e+00 -6.90486431e-01 5.92679262e-01
-8.09259489e-02 -6.72119558e-01 -3.71575594e-01 6.69506729e-01
-7.92166233e-01 -1.10461867e+00 -5.12517095e-01 -1.03332269e+00
6.84225678e-01 7.09021866e-01 1.02327752e+00 7.30575547e-02
-2.43134782e-01 -3.86718482e-01 -8.01677465e-01 -5.08605838e-01
2.85483867e-01 1.34103969e-01 -7.02025136e-03 -6.16483808e-01
5.36123276e-01 -2.31178284e-01 -6.76575005e-01 7.76767731e-01
-7.45495677e-01 5.80761349e-03 5.94491243e-01 7.10231960e-01
6.96678638e-01 8.78024399e-01 3.79823506e-01 -2.90323019e-01
6.62741005e-01 -6.10006392e-01 -1.00360060e+00 1.67828351e-01
-3.73288572e-01 -1.04562521e-01 8.08501244e-01 -9.09295604e-02
-2.09672228e-01 -4.19746280e-01 2.60072481e-02 -1.76042214e-01
1.75521404e-01 5.99651396e-01 -2.98642159e-01 -1.47020081e-02
-2.65551955e-01 7.87386298e-02 -5.44891227e-05 -2.49190688e-01
-2.33411491e-02 4.54257607e-01 5.23355091e-03 -5.71229517e-01
5.39757013e-01 2.95228194e-02 3.91734987e-01 -2.80207276e-01
-1.00229859e+00 -4.16211635e-01 -6.28183067e-01 -2.56842878e-02
7.30046034e-01 -1.10907279e-01 -1.36128664e+00 4.98250686e-02
-9.86840487e-01 -2.90672444e-02 -6.01270534e-02 3.72585982e-01
-6.63849115e-01 1.78277731e-01 -3.74843448e-01 -7.19897270e-01
-1.77131787e-01 -1.23078120e+00 2.82717526e-01 6.98593616e-01
1.58727884e-01 -7.02644825e-01 -1.06398046e-01 3.69329274e-01
3.32227618e-01 4.98656891e-02 1.22101438e+00 -4.87187326e-01
-6.73632920e-01 -2.27144971e-01 -2.94891417e-01 -7.09071755e-02
2.12973520e-01 -3.04325819e-01 4.03022319e-01 -1.84929714e-01
-1.65580675e-01 1.71544880e-01 3.51171464e-01 8.20029914e-01
1.36071968e+00 -3.67583424e-01 -5.69607019e-01 5.85799038e-01
1.80976939e+00 1.07758403e+00 6.91172719e-01 9.12701905e-01
4.65314120e-01 3.89874816e-01 1.01597261e+00 4.91448820e-01
3.52337360e-01 6.34551764e-01 1.17414641e+00 8.56333673e-02
5.09281039e-01 -1.24271847e-01 -3.40405971e-01 5.44139385e-01
-6.11950345e-02 -8.93600106e-01 -7.61769414e-01 5.12150586e-01
-1.90652061e+00 -1.19834936e+00 -8.07128623e-02 2.21676397e+00
5.73696315e-01 6.18630886e-01 3.00579667e-01 2.80087322e-01
1.04400015e+00 -2.68156469e-01 -5.10944188e-01 -1.04114604e+00
2.08074421e-01 1.80923015e-01 1.07289267e+00 2.88302630e-01
-9.29828405e-01 6.68676972e-01 6.30353022e+00 8.41099441e-01
-8.51154745e-01 -5.30268252e-01 4.23828512e-01 -2.01818302e-01
6.81325942e-02 -2.91042447e-01 -1.37205434e+00 1.02604854e+00
5.08268774e-01 -2.43527398e-01 7.79433012e-01 6.61187351e-01
2.28671446e-01 -5.02913892e-01 -1.09589803e+00 7.22829700e-01
-2.67627612e-02 -1.60273862e+00 -8.28931108e-02 3.92642081e-01
6.18379235e-01 -5.63462615e-01 -1.54739432e-03 2.85293907e-01
2.81860292e-01 -1.18485558e+00 5.55541515e-01 2.92694986e-01
9.78386402e-02 -1.46024024e+00 1.16725802e+00 3.60332251e-01
-1.09179819e+00 -4.40922767e-01 -1.05880463e+00 -2.24858463e-01
4.33684707e-01 6.15917206e-01 -8.29744935e-01 6.10297382e-01
1.01264179e+00 3.02665472e-01 2.50342429e-01 1.94624794e+00
-1.60626978e-01 4.55116816e-02 -5.58871448e-01 -8.45671475e-01
7.83975959e-01 -7.05558479e-01 2.28240654e-01 6.65168941e-01
4.53395545e-01 2.93331623e-01 3.85467678e-01 9.24328089e-01
2.92555410e-02 -5.39900959e-02 -4.65220749e-01 -8.75943247e-03
7.82316327e-01 1.04522538e+00 -1.00387716e+00 2.05797940e-01
-9.90365148e-02 3.93569589e-01 3.78597021e-01 -3.81934904e-02
-1.02347159e+00 -8.38103592e-01 3.71376276e-01 2.25794584e-01
1.13144898e+00 -3.23923260e-01 -3.76903683e-01 -2.54468381e-01
-9.63376090e-02 -5.66720963e-01 2.21816525e-01 -5.24536490e-01
-1.39573300e+00 1.63655177e-01 8.23364854e-02 -1.04806268e+00
3.43912840e-01 -9.42455053e-01 -6.74837172e-01 9.27004993e-01
-1.56682682e+00 -2.05281407e-01 -3.19104493e-02 3.41667026e-01
4.67344046e-01 -4.84068096e-02 3.66414666e-01 1.99042618e-01
-8.40208352e-01 3.78035635e-01 4.02671069e-01 1.75435897e-02
-1.05538808e-01 -1.32965648e+00 -8.59253574e-03 4.33288217e-01
-3.53760242e-01 2.95961410e-01 7.16604829e-01 -5.96619785e-01
-1.78215206e+00 -5.75343013e-01 6.95314467e-01 2.66496688e-01
4.99790907e-01 -2.19716847e-01 -3.28856379e-01 3.65294039e-01
2.29765669e-01 -3.82665098e-01 6.34839952e-01 2.15663999e-01
4.64683712e-01 -3.57443273e-01 -1.22748685e+00 4.94804621e-01
9.27018702e-01 8.57710481e-01 -7.24584997e-01 5.26793480e-01
8.09503138e-01 -7.84947634e-01 -5.99465013e-01 2.46393055e-01
2.35083178e-01 -9.53797519e-01 9.22504663e-01 -7.36821830e-01
6.85693741e-01 -2.14950800e-01 -2.85832435e-01 -1.64133430e+00
-9.53150451e-01 -2.29615971e-01 -1.73041284e-01 7.23060668e-01
3.54579031e-01 -4.56833154e-01 1.10840642e+00 4.10033971e-01
-3.77067387e-01 -1.37450528e+00 -9.46792960e-01 -9.33047056e-01
4.94127981e-02 1.92004636e-01 1.12759805e+00 3.21853489e-01
2.07671359e-01 8.55939165e-02 -1.54122472e-01 2.09941879e-01
4.93296832e-01 6.93706989e-01 4.01764274e-01 -1.05018127e+00
-1.37381658e-01 -6.12943470e-01 -3.80395651e-01 -1.33796668e+00
-2.94462234e-01 -6.41304910e-01 3.02961260e-01 -2.12876153e+00
5.03201522e-02 -1.05084050e+00 -4.50536281e-01 3.68263572e-01
2.42885575e-01 -1.65432394e-01 2.51207650e-01 -1.72060058e-01
-6.46837771e-01 3.52624476e-01 1.68261778e+00 -7.02669621e-02
-3.67433786e-01 2.77105212e-01 -7.26100862e-01 3.64576131e-01
9.50981796e-01 -4.09966052e-01 -4.55845594e-02 -5.27737558e-01
4.82214600e-01 3.72934312e-01 -4.33925927e-01 -6.41649365e-01
2.62098491e-01 -7.30759621e-01 6.46564066e-01 -1.38321829e+00
3.27664584e-01 -1.32738495e+00 -2.70110577e-01 7.51626551e-01
2.26109415e-01 3.93610835e-01 3.72784674e-01 3.86893064e-01
-8.79371352e-03 -9.43888724e-01 5.11331379e-01 -2.97662050e-01
-5.43593585e-01 4.30217594e-01 -3.42786282e-01 -1.36170372e-01
1.55273771e+00 -7.72772789e-01 -3.66321504e-01 7.31584150e-03
-4.54448760e-01 9.26951349e-01 -9.95226391e-03 2.28787810e-02
5.53005874e-01 -1.47158766e+00 -2.76741236e-01 1.49646223e-01
-5.34032702e-01 4.45665330e-01 1.91082716e-01 5.55950105e-01
-1.18975854e+00 1.00473654e+00 -7.46230543e-01 -4.44117725e-01
-8.33186686e-01 8.97750854e-01 2.73453593e-01 -7.19478011e-01
-3.54338914e-01 1.11218333e+00 -3.96826684e-01 -1.20218180e-01
3.74853551e-01 -1.49770454e-01 -2.75651395e-01 9.08423439e-02
2.47341588e-01 4.80337054e-01 6.95187300e-02 4.28865515e-02
-3.23409051e-01 5.27855992e-01 -4.17334139e-01 3.40893328e-01
1.84786272e+00 8.25741217e-02 -3.80404711e-01 -1.10427998e-01
8.22125256e-01 -2.13318080e-01 -7.65917838e-01 2.29391083e-01
-2.54736006e-01 -8.71275127e-01 -1.75606906e-01 -6.33528531e-01
-1.20185065e+00 6.60347283e-01 1.46647796e-01 9.18125987e-01
9.57064688e-01 -2.37636864e-01 9.83548224e-01 2.67610461e-01
6.87873602e-01 -1.43349588e+00 1.85696959e-01 8.69185030e-01
6.81143165e-01 -7.69482493e-01 2.07758307e-01 -3.75440717e-01
-2.62700677e-01 1.25001395e+00 9.76602793e-01 -5.46883285e-01
4.94597495e-01 1.67448178e-01 -8.25172067e-01 -1.54739693e-01
-3.61385882e-01 -1.40085965e-01 -5.13503440e-02 6.08098567e-01
1.99557971e-02 1.83327049e-01 -6.12378478e-01 4.83469814e-01
-5.59458315e-01 -3.94397974e-01 4.36395645e-01 1.10005736e+00
-1.00968492e+00 -1.09146070e+00 -7.89410591e-01 5.56252122e-01
-1.04632460e-01 1.58902168e-01 1.35345191e-01 9.07092929e-01
4.33066994e-01 9.92199719e-01 6.62308455e-01 -1.01626009e-01
4.38816220e-01 -4.21756625e-01 8.63326132e-01 -6.26424730e-01
-1.94993064e-01 -1.34222075e-01 -2.02168539e-01 -2.84161001e-01
-1.46645606e-01 -3.41276556e-01 -1.40417778e+00 -7.42385924e-01
-6.06451869e-01 6.38598740e-01 8.01870644e-01 9.24012184e-01
-9.29753669e-03 7.05497861e-01 8.60835969e-01 -9.94122207e-01
-4.94089901e-01 -4.93773788e-01 -1.05421698e+00 -5.13337553e-02
-1.02425635e-01 -1.12832403e+00 -1.68451533e-01 -8.14713717e-01] | [5.020750522613525, 2.7419192790985107] |
7fa247af-12bf-4d05-acaa-2bfa772e73d9 | leverage-interactive-affinity-for-affordance | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Luo_Leverage_Interactive_Affinity_for_Affordance_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Luo_Leverage_Interactive_Affinity_for_Affordance_Learning_CVPR_2023_paper.pdf | Leverage Interactive Affinity for Affordance Learning | Perceiving potential "action possibilities" (i.e., affordance) regions of images and learning interactive functionalities of objects from human demonstration is a challenging task due to the diversity of human-object interactions. Prevailing affordance learning algorithms often adopt the label assignment paradigm and presume that there is a unique relationship between functional region and affordance label, yielding poor performance when adapting to unseen environments with large appearance variations. In this paper, we propose to leverage interactive affinity for affordance learning, i.e., extracting interactive affinity from human-object interaction and transferring it to non-interactive objects. Interactive affinity, which represents the contacts between different parts of the human body and local regions of the target object, can provide inherent cues of interconnectivity between humans and objects, thereby reducing the ambiguity of the perceived action possibilities. Specifically, we propose a pose-aided interactive affinity learning framework that exploits human pose to guide the network to learn the interactive affinity from human-object interactions. Particularly, a keypoint heuristic perception (KHP) scheme is devised to exploit the keypoint association of human pose to alleviate the uncertainties due to interaction diversities and contact occlusions. Besides, a contact-driven affordance learning (CAL) dataset is constructed by collecting and labeling over 5,000 images. Experimental results demonstrate that our method outperforms the representative models regarding objective metrics and visual quality. Code and dataset: github.com/lhc1224/PIAL-Net. | ['DaCheng Tao', 'Yang Cao', 'Jing Zhang', 'Wei Zhai', 'Hongchen Luo'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['human-object-interaction-detection'] | ['computer-vision'] | [-4.12958376e-02 -7.24583631e-03 -2.83626139e-01 -3.87120157e-01
-2.43182674e-01 -4.86653298e-01 2.83979326e-01 -1.45189971e-01
-2.05027401e-01 4.50679719e-01 2.74374217e-01 1.81633905e-01
-2.85816789e-01 -2.38627642e-01 -8.71429443e-01 -4.99104023e-01
-1.97221801e-01 4.67531770e-01 4.91123497e-01 -2.91671872e-01
9.18163881e-02 7.47680426e-01 -1.79912376e+00 3.25154781e-01
9.36530650e-01 1.06500506e+00 5.63077271e-01 3.39646518e-01
2.27813363e-01 4.34628010e-01 -4.05420214e-01 5.86617328e-02
3.79175335e-01 -1.80987403e-01 -6.83097124e-01 2.69448966e-01
5.74108720e-01 -5.99919319e-01 -4.70994473e-01 8.56809139e-01
4.52152193e-01 4.84097242e-01 7.63607800e-01 -1.66960919e+00
-7.83607066e-01 2.55001068e-01 -3.98794174e-01 2.62993366e-01
7.39234507e-01 5.02506137e-01 1.16102135e+00 -1.06158698e+00
7.78533936e-01 1.61456609e+00 9.31160673e-02 6.51292562e-01
-1.22624779e+00 -5.02993286e-01 5.92573822e-01 5.53693354e-01
-1.47770691e+00 -1.91669181e-01 9.61336493e-01 -6.46192729e-01
7.34518945e-01 4.87140447e-01 1.03355646e+00 1.23482656e+00
5.22428975e-02 1.25461674e+00 9.65143979e-01 -2.55525410e-01
-3.53307575e-02 9.16756515e-04 -1.09641291e-01 8.60046208e-01
-6.23616017e-02 3.86929959e-01 -5.61659813e-01 -6.02090172e-02
1.20722389e+00 3.42254817e-01 -5.20213783e-01 -1.05632079e+00
-1.57714641e+00 3.28142256e-01 1.01863027e+00 -8.18746164e-02
-2.14558572e-01 -6.13796785e-02 7.41455629e-02 1.14264257e-01
-2.52262168e-02 4.47230667e-01 -5.77953637e-01 2.80723810e-01
-1.20234387e-02 4.34918702e-01 5.44684112e-01 1.38279092e+00
8.26050401e-01 -4.28880095e-01 -3.66100311e-01 6.89291358e-01
4.86529231e-01 3.03202391e-01 1.49725094e-01 -9.77843642e-01
3.39576453e-01 8.49155664e-01 2.33163729e-01 -9.91942763e-01
-5.26539803e-01 8.57093707e-02 -4.04542655e-01 3.70370537e-01
4.08191532e-01 2.77084261e-01 -9.25024927e-01 1.64187026e+00
8.40913475e-01 -1.50107831e-01 -2.18548700e-01 1.53519332e+00
9.26977932e-01 2.56502420e-01 2.91070372e-01 1.39896914e-01
1.39209783e+00 -1.20924509e+00 -4.31281179e-01 -1.86114162e-01
3.32293600e-01 -5.54708004e-01 1.61401045e+00 8.30058903e-02
-5.92988014e-01 -7.75726438e-01 -9.04046118e-01 -2.78987944e-01
-4.48328525e-01 2.72433311e-01 9.25832093e-01 -7.81128705e-02
-3.42743158e-01 4.08205688e-01 -7.15361118e-01 -3.07233781e-01
3.69647563e-01 3.98686707e-01 -5.37914455e-01 -5.76363318e-02
-9.87803698e-01 7.39740312e-01 5.07197320e-01 2.33038470e-01
-1.13181531e+00 -4.59408671e-01 -9.82105255e-01 -1.43789098e-01
9.72405195e-01 -6.98485911e-01 8.16695631e-01 -9.17712629e-01
-1.23098075e+00 5.29147685e-01 1.88265547e-01 2.74543405e-01
7.31365025e-01 -4.45861816e-01 -2.20142961e-01 1.24564007e-01
4.99024168e-02 1.20192015e+00 8.70967329e-01 -1.60271561e+00
-5.98558068e-01 -4.48711455e-01 4.74343598e-01 5.63909888e-01
-4.51154932e-02 -4.04876739e-01 -4.97755408e-01 -7.08852708e-01
5.64098775e-01 -1.11535025e+00 -6.43846095e-02 6.89692736e-01
-5.83337009e-01 -7.05338955e-01 8.41121793e-01 -4.28981185e-01
7.69596994e-01 -2.16540289e+00 6.08424962e-01 1.61073953e-02
2.79402852e-01 -1.66855738e-01 -1.73856556e-01 1.62716180e-01
2.03172937e-02 -2.82179922e-01 5.33863232e-02 9.67701301e-02
2.78724611e-01 4.73437756e-01 -5.37896380e-02 4.81892139e-01
2.38003030e-01 9.90110993e-01 -1.04403210e+00 -7.66658604e-01
5.02788365e-01 5.26762366e-01 -5.33860564e-01 6.44767523e-01
-4.60058302e-01 8.84291351e-01 -6.88919544e-01 1.12803137e+00
2.97549367e-01 -8.96666646e-02 -3.92230786e-02 -6.14704907e-01
1.30638823e-01 -1.79379046e-01 -1.29430079e+00 2.01460338e+00
-1.59502625e-01 2.99982578e-01 -2.14037046e-01 -4.08324331e-01
7.23498583e-01 8.20844918e-02 3.66902024e-01 -4.42007005e-01
2.36172542e-01 7.89745077e-02 2.70481974e-01 -9.57737803e-01
1.19215861e-01 4.08038467e-01 -1.21594872e-02 9.07390341e-02
1.94257975e-01 -6.85629100e-02 -2.33861376e-02 -1.29400015e-01
6.09054446e-01 7.26791084e-01 3.57563794e-01 -2.56495506e-01
3.43698055e-01 -2.07529977e-01 4.70506102e-01 3.50842088e-01
-5.64746678e-01 4.45689321e-01 2.42740780e-01 -7.00974703e-01
-9.70320821e-01 -1.49398553e+00 -1.97807968e-01 1.42361522e+00
8.27358067e-01 -2.19758358e-02 -5.85969746e-01 -7.67663598e-01
1.82278812e-01 3.99807274e-01 -7.98868418e-01 -1.44150063e-01
-5.14089227e-01 -1.23933405e-01 -1.72880486e-01 6.46656752e-01
3.81344855e-01 -1.43976116e+00 -9.62143481e-01 -2.16136694e-01
-3.11215103e-01 -9.34976101e-01 -7.37291157e-01 -3.41623127e-02
-5.26554167e-01 -1.23854685e+00 -5.13423741e-01 -7.52970099e-01
8.67980838e-01 2.76618332e-01 9.39346611e-01 1.47432953e-01
-6.90367460e-01 4.91387546e-01 -3.10731113e-01 -3.63382809e-02
9.02791172e-02 -7.23304003e-02 2.94594854e-01 3.19705047e-02
2.57842362e-01 -4.65341091e-01 -9.76975083e-01 6.84544921e-01
-7.11820066e-01 2.40397915e-01 8.73266578e-01 8.35466981e-01
8.06690395e-01 -2.61750937e-01 2.62220472e-01 -2.83243001e-01
2.88360149e-01 -3.60011160e-01 -2.85269618e-01 4.55621243e-01
-2.17270046e-01 -8.19075331e-02 1.41091987e-01 -9.77382660e-01
-9.65964019e-01 4.33177680e-01 3.29465389e-01 -7.15113461e-01
-5.51228106e-01 -6.66162968e-02 -6.85944557e-01 -1.07597753e-01
7.17886925e-01 -1.80780560e-01 -2.66977102e-01 -3.90476167e-01
6.33148372e-01 4.37542379e-01 5.52379727e-01 -9.11403239e-01
5.78424215e-01 4.42164183e-01 -1.55639812e-01 -5.80800891e-01
-8.20543289e-01 -5.78610420e-01 -1.19520199e+00 -6.37086749e-01
1.07438850e+00 -8.01707566e-01 -9.01541233e-01 5.80610782e-02
-1.12876904e+00 -2.72653073e-01 -2.70608395e-01 6.50647163e-01
-8.24104130e-01 1.32268742e-01 -3.42789054e-01 -7.05666184e-01
1.66294590e-01 -1.29473007e+00 1.28381991e+00 3.56155813e-01
-3.83111864e-01 -5.88898957e-01 -4.98224109e-01 3.37998927e-01
-2.46263996e-01 6.41800702e-01 8.02818716e-01 -4.30349439e-01
-9.58855271e-01 -5.96137643e-02 -4.22839671e-01 3.49554121e-02
2.88099289e-01 -2.54399896e-01 -7.52238035e-01 -2.20732301e-01
-3.33030611e-01 -6.78915977e-01 4.10512924e-01 1.58359423e-01
1.18973422e+00 -3.02190542e-01 -5.13520837e-01 3.88395935e-01
9.81458485e-01 6.29499648e-03 2.76026577e-01 8.59451741e-02
1.23104692e+00 1.11082542e+00 1.02754939e+00 4.47732449e-01
4.28992659e-01 7.50774980e-01 5.73535800e-01 -9.71502587e-02
-2.03161389e-01 -4.53318149e-01 8.81701782e-02 3.49328935e-01
-1.74656376e-01 -3.39162024e-03 -7.00855434e-01 2.85125464e-01
-1.97404516e+00 -6.08389378e-01 2.84687281e-02 1.99778152e+00
7.69863427e-01 1.83681294e-01 1.95061684e-01 -3.27313215e-01
5.90001881e-01 -9.30670053e-02 -1.12947690e+00 3.07959855e-01
5.89397661e-02 -4.84914720e-01 2.90147942e-02 3.30747604e-01
-1.03508329e+00 9.28305864e-01 5.10384321e+00 5.54843485e-01
-7.57584274e-01 4.81184572e-02 2.77314305e-01 -6.57041511e-03
-1.72738880e-01 -1.84488371e-01 -5.75117171e-01 1.98868468e-01
-1.87419564e-01 3.35832566e-01 5.44970930e-01 1.00432551e+00
1.43413870e-02 -8.40207785e-02 -1.52569711e+00 9.37666416e-01
1.19229212e-01 -5.73858738e-01 1.90309286e-01 -7.67831057e-02
5.48685133e-01 -3.24101150e-01 9.48825106e-03 1.94337219e-01
8.05160925e-02 -8.26111257e-01 8.54152501e-01 8.58749568e-01
6.59981787e-01 -4.37854558e-01 3.42198282e-01 4.62718546e-01
-1.40581346e+00 -3.69160742e-01 -2.42637917e-01 -1.44884195e-02
-2.38076746e-02 -1.53563395e-01 -5.67797840e-01 3.00355971e-01
1.01214206e+00 4.79811370e-01 -6.25670969e-01 9.19145942e-01
-2.94703960e-01 -2.99857408e-01 -3.26060057e-01 -1.26791522e-01
-6.25872705e-03 -7.06755966e-02 6.39138401e-01 7.31851399e-01
-5.09272926e-02 2.31976748e-01 6.59084916e-01 1.11022270e+00
3.18216681e-01 8.74542966e-02 -4.68851179e-01 1.53249696e-01
5.19027233e-01 1.24381006e+00 -7.52179861e-01 -2.04213470e-01
-2.13856727e-01 1.18432546e+00 5.93332231e-01 6.16078198e-01
-9.29212213e-01 -8.15492272e-02 6.17465436e-01 1.69868901e-01
1.81468517e-01 -3.14775944e-01 2.01542616e-01 -1.00508094e+00
4.06347394e-01 -7.18805671e-01 3.68684411e-01 -9.83798444e-01
-1.33015990e+00 5.66367984e-01 3.99129778e-01 -1.42524159e+00
1.19826496e-01 -7.14233458e-01 -5.23547865e-02 7.41262019e-01
-1.05138087e+00 -1.48568773e+00 -7.25112975e-01 6.53572798e-01
8.53191435e-01 5.23801818e-02 4.86584842e-01 -3.38176219e-03
-1.56306341e-01 4.59993213e-01 -6.25595510e-01 -6.86976612e-02
5.37159264e-01 -1.26911509e+00 2.83106836e-03 2.86221206e-01
2.04894975e-01 6.64013743e-01 6.08024895e-01 -5.68160713e-01
-1.43056726e+00 -8.66539538e-01 3.55120242e-01 -8.91777337e-01
4.26029265e-01 -6.18702531e-01 -8.61766636e-01 5.58687806e-01
-2.78948814e-01 4.25771028e-01 4.47440118e-01 -5.11763468e-02
-4.97513741e-01 1.07041650e-01 -9.62943852e-01 1.08490348e+00
1.76431191e+00 -6.66053236e-01 -7.67995000e-01 4.56212938e-01
9.62559879e-01 -7.23243654e-01 -9.31607127e-01 6.49857581e-01
9.93138075e-01 -8.02845597e-01 1.20436013e+00 -6.21887803e-01
1.03913315e-01 -7.18890667e-01 -3.98303777e-01 -1.01369679e+00
-4.14568067e-01 -2.32896388e-01 -3.79031867e-01 8.39343429e-01
3.10489647e-02 -2.60804296e-01 4.88432348e-01 5.95394611e-01
-2.09583014e-01 -9.16205943e-01 -9.32001472e-01 -6.93959594e-01
-4.15303707e-01 -8.08299184e-02 6.67996764e-01 8.82318437e-01
-1.81538314e-02 3.50578696e-01 -2.35361144e-01 3.93383950e-01
6.24435961e-01 2.11952627e-01 9.22131836e-01 -1.21892536e+00
-3.92213315e-01 -3.71543318e-01 -4.64966744e-01 -1.36519992e+00
3.33747357e-01 -7.08685338e-01 3.79447758e-01 -1.30491602e+00
9.57289636e-02 -6.66860640e-01 -3.06527764e-01 5.29399574e-01
-3.01786691e-01 2.12937832e-01 1.88428715e-01 5.23682356e-01
-7.69332409e-01 7.64791667e-01 2.10201859e+00 -2.79693037e-01
-4.18214321e-01 -1.34278715e-01 -7.92007446e-02 7.05642998e-01
4.76706207e-01 -1.59934014e-01 -6.45834446e-01 -2.84466624e-01
-8.48391354e-02 -3.23709697e-02 9.69578743e-01 -1.07873452e+00
3.18162404e-02 -4.18181539e-01 5.45620084e-01 -4.93906587e-01
4.86331642e-01 -1.07240188e+00 1.44146428e-01 3.73103619e-01
-4.37790781e-01 -6.78025633e-02 -9.38679725e-02 9.45834935e-01
1.87089682e-01 1.17818005e-01 3.64855617e-01 -2.57690340e-01
-1.16075730e+00 6.53329253e-01 1.90661758e-01 -1.33147389e-01
1.17355764e+00 -2.88240671e-01 -4.61090505e-02 1.21615514e-01
-1.09979141e+00 4.07183260e-01 5.69485068e-01 9.91124868e-01
8.24194133e-01 -1.47830570e+00 -3.28401089e-01 2.84076959e-01
6.86005235e-01 2.41622567e-01 5.48049092e-01 7.21334636e-01
-1.73667133e-01 2.45112590e-02 -5.16131222e-01 -1.00759435e+00
-1.16348386e+00 8.16774428e-01 4.01906580e-01 5.32052040e-01
-7.04647601e-01 9.19698358e-01 6.57527208e-01 -6.53379858e-01
5.33283114e-01 -5.05787730e-01 -2.28789687e-01 -1.88361138e-01
1.66631937e-01 3.53548646e-01 -5.39857507e-01 -9.16007161e-01
-1.89172775e-01 6.11012220e-01 4.67316471e-02 2.14806437e-01
8.05105567e-01 -3.32581043e-01 2.20888760e-02 7.16122210e-01
1.03810537e+00 -5.31122684e-01 -1.96886635e+00 -2.79692978e-01
-1.53738379e-01 -7.37522066e-01 -4.17629182e-01 -7.95052111e-01
-6.91797435e-01 8.41924131e-01 9.77622390e-01 -8.15673992e-02
9.06831503e-01 4.73943770e-01 4.29085732e-01 5.98092139e-01
5.49051523e-01 -1.15758312e+00 6.30423784e-01 -1.44718969e-02
1.33763695e+00 -1.61018741e+00 2.43291259e-02 -6.46468282e-01
-6.38860881e-01 8.94798458e-01 1.27610970e+00 1.21975407e-01
5.70350766e-01 -3.05533707e-01 2.28937380e-02 -4.25604165e-01
-3.79463226e-01 -2.75640577e-01 7.67421722e-01 8.07693481e-01
1.36752099e-01 3.37947577e-01 -1.63848456e-02 3.96061808e-01
-3.34224589e-02 -4.34355855e-01 -1.75103143e-01 8.78486753e-01
-4.04533178e-01 -6.66319013e-01 -1.68664470e-01 3.52906227e-01
2.81701326e-01 2.14361563e-01 -4.13776755e-01 8.94128561e-01
5.90635717e-01 6.45250142e-01 3.33627090e-02 -3.79659891e-01
6.52413189e-01 -1.76103368e-01 7.34931707e-01 -5.58311462e-01
-2.01166347e-01 1.77366659e-01 -2.95241475e-01 -8.89734745e-01
-5.48229933e-01 -6.48259461e-01 -1.49830663e+00 9.73648429e-02
-4.31549668e-01 -2.26981834e-01 2.52514720e-01 9.53278601e-01
2.45621756e-01 6.06128812e-01 3.57929140e-01 -1.33253670e+00
-4.28210378e-01 -8.80727172e-01 -5.22702992e-01 6.27125502e-01
4.04020578e-01 -1.43258584e+00 -2.08081543e-01 7.75571018e-02] | [5.166680812835693, -0.0973045825958252] |
85498c03-0366-4482-b28b-b5d898b9c314 | architect-regularize-and-replay-arr-a | 2301.02464 | null | https://arxiv.org/abs/2301.02464v1 | https://arxiv.org/pdf/2301.02464v1.pdf | Architect, Regularize and Replay (ARR): a Flexible Hybrid Approach for Continual Learning | In recent years we have witnessed a renewed interest in machine learning methodologies, especially for deep representation learning, that could overcome basic i.i.d. assumptions and tackle non-stationary environments subject to various distributional shifts or sample selection biases. Within this context, several computational approaches based on architectural priors, regularizers and replay policies have been proposed with different degrees of success depending on the specific scenario in which they were developed and assessed. However, designing comprehensive hybrid solutions that can flexibly and generally be applied with tunable efficiency-effectiveness trade-offs still seems a distant goal. In this paper, we propose "Architect, Regularize and Replay" (ARR), an hybrid generalization of the renowned AR1 algorithm and its variants, that can achieve state-of-the-art results in classic scenarios (e.g. class-incremental learning) but also generalize to arbitrary data streams generated from real-world datasets such as CIFAR-100, CORe50 and ImageNet-1000. | ['Davide Maltoni', 'Gabriele Graffieti', 'Lorenzo Pellegrini', 'Vincenzo Lomonaco'] | 2023-01-06 | null | null | null | null | ['class-incremental-learning'] | ['computer-vision'] | [ 1.75945401e-01 -3.75383973e-01 -1.01948194e-01 -5.78658104e-01
-6.52706683e-01 -5.44020951e-01 9.87645566e-01 2.29347140e-01
-6.84868634e-01 7.71715343e-01 3.88507321e-02 -3.39583158e-01
-5.07653534e-01 -7.09958017e-01 -5.17544091e-01 -8.55974376e-01
-2.46712476e-01 7.19094217e-01 1.56182364e-01 -2.41747558e-01
3.14281076e-01 7.31489718e-01 -2.01429033e+00 1.77512735e-01
5.40431917e-01 9.87294793e-01 1.06114559e-01 6.54372871e-01
-2.00363427e-01 7.23793805e-01 -7.40064204e-01 -1.14920847e-01
2.27404431e-01 -9.68739092e-02 -5.05752146e-01 -1.71926841e-01
1.71913520e-01 7.71444365e-02 -2.28822470e-01 8.28409255e-01
9.21758413e-01 2.19588041e-01 7.27588117e-01 -1.21487677e+00
-5.36758184e-01 8.32164109e-01 -6.86637282e-01 6.66107178e-01
-1.68258362e-02 1.68693900e-01 6.22145712e-01 -5.89064360e-01
1.97036222e-01 1.33430755e+00 5.02072334e-01 5.63054800e-01
-1.48904598e+00 -7.43647456e-01 4.62352663e-01 4.62035388e-01
-1.21470189e+00 -4.72369552e-01 6.35225773e-01 -3.92332405e-01
7.61753917e-01 1.85614884e-01 2.09662169e-01 1.78880274e+00
-2.04026580e-01 9.37474847e-01 1.05898452e+00 -4.16318595e-01
5.93240499e-01 2.48627305e-01 1.96826175e-01 -3.17899161e-03
3.44965994e-01 1.07630165e-02 -3.77780467e-01 -1.85511649e-01
3.44058752e-01 -6.81987181e-02 -3.38499933e-01 -3.74967873e-01
-1.06195462e+00 9.87007618e-01 6.77644238e-02 3.31845909e-01
-4.59224850e-01 2.02698916e-01 8.00797999e-01 4.03744191e-01
5.05661547e-01 1.86746210e-01 -6.04003429e-01 -1.97780490e-01
-9.61889148e-01 5.49015701e-01 7.70163298e-01 8.74541342e-01
6.01208210e-01 4.63144481e-01 -3.64831060e-01 8.43078434e-01
1.61822915e-01 3.10181528e-01 9.34710205e-01 -5.10919631e-01
3.35911244e-01 2.27050439e-01 9.03230160e-02 -5.71836352e-01
-6.52601480e-01 -9.76621866e-01 -1.10358942e+00 4.38774936e-02
3.34650815e-01 -1.74060971e-01 -9.22012269e-01 1.84008574e+00
2.69271880e-01 6.04565203e-01 2.31881484e-01 7.21936405e-01
6.92200780e-01 6.02451086e-01 3.25248033e-01 -9.95399058e-02
1.28346765e+00 -4.57217991e-01 -4.36040550e-01 -6.17117658e-02
3.33968729e-01 -6.45056009e-01 1.22535217e+00 7.39103556e-01
-7.05846906e-01 -7.54336238e-01 -1.10711241e+00 5.08558095e-01
-4.59152669e-01 4.41322997e-02 5.11649907e-01 9.03154850e-01
-9.69497859e-01 5.37700891e-01 -8.40435624e-01 -3.50016475e-01
4.17948157e-01 2.35540316e-01 8.80412832e-02 4.10750974e-03
-1.16105354e+00 6.91075325e-01 4.51130658e-01 -1.42664211e-02
-1.24054778e+00 -8.54332566e-01 -2.64675081e-01 4.68392260e-02
4.21439111e-01 -6.62551284e-01 1.40827167e+00 -1.09101653e+00
-1.80284703e+00 8.34235191e-01 2.64381021e-01 -1.08075619e+00
7.99563706e-01 -4.98554379e-01 -5.47565043e-01 -1.69157669e-01
-3.23518604e-01 4.40235823e-01 8.92972529e-01 -1.18369079e+00
-6.34561896e-01 -3.50256592e-01 5.06025814e-02 -2.63536423e-02
-2.86284059e-01 6.91564754e-02 5.07197492e-02 -8.06119144e-01
-3.23078990e-01 -8.95963192e-01 -4.09377754e-01 -3.54251385e-01
-2.02544168e-01 -3.72058898e-01 7.06632376e-01 1.63765967e-01
1.09179199e+00 -2.16581011e+00 9.49781463e-02 1.64529532e-01
-3.17808241e-01 6.89225733e-01 -2.39402115e-01 3.99320602e-01
-1.76197574e-01 3.24112992e-03 -9.19555202e-02 -3.75376433e-01
2.27268133e-02 3.52899879e-01 -6.43042326e-01 7.53264129e-01
-1.10469274e-02 3.62820715e-01 -8.42409015e-01 -3.58933769e-02
3.06031615e-01 6.51833534e-01 -5.35389543e-01 2.80671835e-01
-4.14012909e-01 4.39112127e-01 -3.14161837e-01 2.90547967e-01
6.98628724e-01 -2.30510637e-01 1.07819801e-02 6.18573539e-02
-8.86805654e-02 4.77079535e-03 -1.37957346e+00 1.61678743e+00
-5.79245627e-01 5.93576193e-01 -4.18712199e-02 -1.53907228e+00
1.22630942e+00 2.45361194e-01 4.97032255e-01 -6.63610280e-01
1.89868480e-01 1.72070265e-01 -5.37233464e-02 -4.93167043e-01
5.26653230e-01 -1.08700611e-01 1.82643369e-01 1.79806054e-01
1.26841605e-01 3.29178751e-01 1.41718224e-01 -2.98760027e-01
9.55985129e-01 2.35339090e-01 3.88175219e-01 -4.94056404e-01
6.59520030e-01 -2.70082027e-01 3.85535628e-01 1.11279118e+00
-6.90099373e-02 8.37009490e-01 3.37915689e-01 -6.74618602e-01
-7.70875096e-01 -1.07866621e+00 -3.46030444e-01 1.44305956e+00
-9.53267142e-02 4.61569540e-02 -5.31966209e-01 -3.79299998e-01
-2.79695615e-02 9.90427792e-01 -5.87663710e-01 -3.48918766e-01
-6.84525192e-01 -1.22955215e+00 6.50045335e-01 3.43683213e-01
4.48112726e-01 -1.12767410e+00 -1.30384195e+00 5.80856383e-01
8.51937160e-02 -1.07840502e+00 2.93949425e-01 5.66560090e-01
-8.01794291e-01 -7.38328099e-01 -7.20081568e-01 -2.29299501e-01
6.47873506e-02 1.83355257e-01 1.39159799e+00 -4.02126014e-01
-2.30434224e-01 3.66179436e-01 -7.28360951e-01 -7.28945732e-01
-2.41365552e-01 4.24288720e-01 1.03034839e-01 4.78423923e-01
3.48314166e-01 -8.51494968e-01 -7.59426951e-01 2.15676844e-01
-1.38658237e+00 -4.30107206e-01 4.57519770e-01 6.61830246e-01
5.54384410e-01 -1.82291642e-01 1.07700920e+00 -9.84695077e-01
6.75848126e-01 -1.07604098e+00 -6.40765786e-01 2.53953516e-01
-7.42871106e-01 2.67767191e-01 8.93107414e-01 -8.19462121e-01
-1.20051706e+00 -2.83287138e-01 -2.99355447e-01 -5.63144505e-01
-5.66820979e-01 4.63995129e-01 -1.18499964e-01 4.68691349e-01
1.04505706e+00 4.08881128e-01 -2.14783862e-01 -5.89161456e-01
4.42472905e-01 7.50720024e-01 5.62141597e-01 -9.87935007e-01
4.13201541e-01 5.61770439e-01 -1.38212070e-01 -8.45638037e-01
-7.46477604e-01 -3.28959972e-01 -1.86006039e-01 -4.80138920e-02
4.32999372e-01 -1.01838899e+00 -4.04760182e-01 5.08239269e-01
-9.12010133e-01 -4.39143598e-01 -6.00320637e-01 5.97416222e-01
-7.29987204e-01 1.21499829e-01 -1.66483633e-02 -8.34266305e-01
-4.13012773e-01 -1.11468172e+00 9.54482257e-01 4.49188769e-01
-4.67607193e-02 -7.69183934e-01 3.06332856e-01 -2.31604591e-01
9.02074158e-01 3.02838266e-01 7.26467133e-01 -1.09426332e+00
-3.50172073e-01 4.33246009e-02 -1.19889192e-01 4.54929322e-01
-2.76176110e-02 3.10832206e-02 -1.29590321e+00 -5.01063049e-01
-8.40630829e-02 -2.79117137e-01 8.93061817e-01 5.36137760e-01
1.48763919e+00 -3.43675226e-01 -7.56874010e-02 8.25526953e-01
1.54814601e+00 2.24701867e-01 6.30865037e-01 5.95540762e-01
1.26497298e-02 4.32963490e-01 3.34483206e-01 1.02082276e+00
1.26242295e-01 6.51504517e-01 6.56062126e-01 2.78218776e-01
1.04778312e-01 6.83557689e-02 3.03507656e-01 2.46528432e-01
-4.96627297e-03 -6.12327099e-01 -7.64687777e-01 5.77267051e-01
-1.83997893e+00 -1.05868423e+00 1.94016770e-01 2.34904218e+00
6.41489387e-01 2.85427600e-01 3.20369273e-01 3.36220086e-01
5.70325017e-01 3.89310986e-01 -6.86275125e-01 -2.90506899e-01
-1.57938674e-01 3.80110115e-01 5.81328988e-01 -1.62475362e-01
-1.04445517e+00 6.79541469e-01 5.66396952e+00 9.76815343e-01
-1.44783914e+00 1.65181741e-01 7.76370823e-01 -2.26081237e-01
-2.71586835e-01 -3.69238675e-01 -8.37600529e-01 5.07002950e-01
1.42498875e+00 -2.52219081e-01 4.01645273e-01 9.49975908e-01
1.29886121e-02 1.43529549e-01 -1.07868266e+00 1.17595398e+00
-7.22596869e-02 -1.23087823e+00 6.36259913e-02 -2.26921082e-01
4.57203656e-01 4.66733485e-01 3.49351287e-01 6.33071840e-01
4.29837108e-01 -9.63258207e-01 8.92417252e-01 6.01729512e-01
5.56637704e-01 -8.23177338e-01 7.22607315e-01 4.32020903e-01
-9.19150114e-01 -2.41696268e-01 -5.11268616e-01 1.37673110e-01
6.98385481e-03 6.90592051e-01 -5.59876323e-01 6.98029459e-01
9.04952526e-01 4.49683607e-01 -4.81836468e-01 1.29109991e+00
2.46686026e-01 9.36097980e-01 -5.28072000e-01 -1.28208622e-01
3.46999109e-01 9.11708698e-02 5.10549486e-01 1.46950746e+00
4.67667967e-01 -3.25001299e-01 6.07832633e-02 5.33388793e-01
7.98133239e-02 2.32243910e-01 -4.85885084e-01 3.84483784e-01
6.54281139e-01 1.13385236e+00 -7.44859159e-01 -3.95395160e-01
-3.68833691e-01 2.91680455e-01 1.76642433e-01 5.80885470e-01
-1.05519974e+00 -1.01096042e-01 8.17216218e-01 2.17050746e-01
4.87681955e-01 -1.18531369e-01 2.62213927e-02 -1.24505246e+00
-1.89826906e-01 -1.08414698e+00 6.31532192e-01 -2.19871894e-01
-1.51255512e+00 9.83150542e-01 5.92099726e-01 -1.30344141e+00
-3.01795542e-01 -6.07217789e-01 -5.30130744e-01 3.91143024e-01
-1.93708575e+00 -6.32373571e-01 -3.94606352e-01 8.21681619e-01
6.69437230e-01 -5.54367006e-01 7.58586407e-01 4.83407706e-01
-5.14208078e-01 6.65052950e-01 3.71571302e-01 -2.83570379e-01
5.68553090e-01 -1.01702547e+00 7.98008516e-02 5.76735616e-01
3.72527391e-01 4.31541771e-01 1.08104372e+00 1.49573177e-01
-1.14680326e+00 -1.23935258e+00 3.65380719e-02 2.97765876e-03
5.24584889e-01 -2.61291981e-01 -1.03963101e+00 5.48697650e-01
2.38580704e-01 2.42826164e-01 5.30036092e-01 6.95019439e-02
-6.18343592e-01 -6.15099490e-01 -1.09104216e+00 4.47292715e-01
9.42538381e-01 -8.21684301e-02 -3.61731648e-01 2.36951903e-01
4.60318506e-01 -3.19236159e-01 -5.12042940e-01 5.79523265e-01
3.02849203e-01 -1.23481798e+00 1.20048690e+00 -7.09632099e-01
-2.12148309e-01 -1.61628008e-01 -5.37361979e-01 -1.25283384e+00
-1.54453725e-01 -7.34486282e-01 -1.75772727e-01 1.32602346e+00
1.56599402e-01 -7.77404189e-01 6.30595088e-01 1.64302841e-01
-1.02998376e-01 -6.61961734e-01 -1.22220516e+00 -8.60768795e-01
1.87424242e-01 -6.47438049e-01 9.14315343e-01 6.05934441e-01
-7.17292488e-01 1.55838057e-01 -4.94186997e-01 1.19489552e-02
6.44624591e-01 1.41207241e-02 9.16215777e-01 -1.46969545e+00
-3.36124808e-01 -6.84131861e-01 -3.43027681e-01 -1.03278458e+00
9.30759460e-02 -7.60450184e-01 -1.35671154e-01 -9.83105659e-01
-2.49275446e-01 -6.41429245e-01 -7.46869087e-01 2.10115477e-01
1.02851577e-01 -1.81088254e-01 7.55830631e-02 1.62282050e-01
-5.51105440e-01 7.24328279e-01 4.38328087e-01 -5.78536876e-02
-2.83166498e-01 3.48001301e-01 -8.32662702e-01 5.25082767e-01
1.14839220e+00 -5.45857608e-01 -8.08303058e-01 -4.59491163e-01
1.94750279e-01 -1.40790746e-01 2.53276199e-01 -1.36509883e+00
1.16491862e-01 -1.80453569e-01 1.78319469e-01 -4.49722141e-01
5.33543527e-02 -7.92429209e-01 2.71613449e-01 1.84203967e-01
-7.85343468e-01 3.55777144e-01 1.47320867e-01 8.93010259e-01
-1.77412570e-01 -2.56419778e-01 1.09950781e+00 -2.07147166e-01
-6.40883327e-01 4.23970133e-01 -4.09727633e-01 3.74708951e-01
1.16439700e+00 7.76203126e-02 -3.93481821e-01 -2.16828212e-01
-4.68189955e-01 -9.51596797e-02 -1.38123453e-01 6.36754811e-01
4.65665221e-01 -1.03636229e+00 -1.10615480e+00 1.20241463e-01
4.22007106e-02 1.19140469e-01 3.72094959e-01 6.41241312e-01
-2.29831234e-01 2.87591547e-01 -1.57888189e-01 -8.35140765e-01
-8.24216008e-01 6.63494945e-01 3.14602882e-01 -4.64856029e-01
-6.53186142e-01 6.17041767e-01 -3.89513560e-02 -3.29375386e-01
6.11677289e-01 -2.92108655e-01 -3.87258261e-01 3.58820736e-01
5.75678587e-01 3.58569026e-01 3.10928911e-01 -3.02481323e-01
-4.01883900e-01 2.63698727e-01 -7.65193701e-02 8.06196779e-02
1.60489881e+00 1.20863775e-02 4.87951398e-01 5.92074156e-01
1.05532956e+00 -4.82252449e-01 -1.37546885e+00 -5.49842060e-01
2.04181001e-01 -2.96894968e-01 -2.57211104e-02 -6.25099003e-01
-1.26090646e+00 9.25471187e-01 9.80567932e-01 4.72606093e-01
1.34115970e+00 -1.32280082e-01 3.19318891e-01 3.58450502e-01
5.32993317e-01 -9.46594536e-01 8.18351284e-02 3.90504956e-01
7.89347351e-01 -1.00968361e+00 -1.38188198e-01 5.51130950e-01
-5.63573599e-01 1.20247614e+00 5.34844339e-01 -3.39145184e-01
8.72591197e-01 3.17989737e-01 2.61839684e-02 2.25833550e-01
-9.52231288e-01 -2.74687350e-01 -9.06997547e-02 7.74707258e-01
3.79575163e-01 -9.19410288e-02 -2.08948210e-01 5.32436490e-01
1.08376879e-03 1.17683731e-01 4.56748724e-01 7.68347144e-01
-3.75120789e-01 -1.02521217e+00 -2.90894181e-01 3.23259264e-01
-5.86382449e-01 1.38879105e-01 3.44689161e-01 9.73776400e-01
2.18178425e-02 8.07923436e-01 7.72669092e-02 -2.10131511e-01
4.60335821e-01 -1.69395935e-02 3.01884443e-01 -3.32198024e-01
-6.67021632e-01 -1.08985417e-01 -8.25619996e-02 -3.72557372e-01
-7.38272488e-01 -7.36379743e-01 -9.69718039e-01 -1.78576320e-01
-4.74762544e-03 9.98542607e-02 6.93007529e-01 7.65909612e-01
5.02439976e-01 6.36352718e-01 5.60630679e-01 -9.86667097e-01
-1.16392350e+00 -1.04791355e+00 -4.75077987e-01 3.49846870e-01
4.35093850e-01 -8.17556739e-01 -3.99122864e-01 -1.90256000e-01] | [9.407938003540039, 3.159766435623169] |
32401cdd-fa4a-4aa3-ac48-d60fff7817b2 | hdformer-high-order-directed-transformer-for | 2302.01825 | null | https://arxiv.org/abs/2302.01825v2 | https://arxiv.org/pdf/2302.01825v2.pdf | HDFormer: High-order Directed Transformer for 3D Human Pose Estimation | Human pose estimation is a challenging task due to its structured data sequence nature. Existing methods primarily focus on pair-wise interaction of body joints, which is insufficient for scenarios involving overlapping joints and rapidly changing poses. To overcome these issues, we introduce a novel approach, the High-order Directed Transformer (HDFormer), which leverages high-order bone and joint relationships for improved pose estimation. Specifically, HDFormer incorporates both self-attention and high-order attention to formulate a multi-order attention module. This module facilitates first-order "joint$\leftrightarrow$joint", second-order "bone$\leftrightarrow$joint", and high-order "hyperbone$\leftrightarrow$joint" interactions, effectively addressing issues in complex and occlusion-heavy situations. In addition, modern CNN techniques are integrated into the transformer-based architecture, balancing the trade-off between performance and efficiency. HDFormer significantly outperforms state-of-the-art (SOTA) models on Human3.6M and MPI-INF-3DHP datasets, requiring only 1/10 of the parameters and significantly lower computational costs. Moreover, HDFormer demonstrates broad real-world applicability, enabling real-time, accurate 3D pose estimation. The source code is in https://github.com/hyer/HDFormer | ['Xuansong Xie', 'Yifeng Geng', 'Bin Luo', 'Hanbing Liu', 'Zhi-Qi Cheng', 'Wei Liu', 'Wangmeng Xiang', 'Jun-Yan He', 'Hanyuan Chen'] | 2023-02-03 | null | null | null | null | ['3d-pose-estimation', '3d-human-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-5.07151783e-01 4.16492932e-02 -1.37930572e-01 -1.69845298e-01
-7.75534868e-01 -5.73282801e-02 1.41290084e-01 -2.67958373e-01
-4.77149546e-01 5.41438460e-01 3.88935328e-01 1.01709992e-01
-3.62041332e-02 -6.27801538e-01 -8.51893306e-01 -4.27145392e-01
-1.92085519e-01 8.66680741e-01 2.51611561e-01 -4.34610784e-01
-2.23095313e-01 3.01967651e-01 -1.30799019e+00 -3.58157754e-02
6.55019879e-01 8.57414067e-01 2.02679738e-01 5.32746375e-01
2.95772016e-01 2.73920238e-01 -3.46231967e-01 -4.31905806e-01
2.11029202e-01 -4.51236032e-02 -7.13148773e-01 -2.24634305e-01
4.68255848e-01 -6.00450099e-01 -3.94806921e-01 5.98383665e-01
8.66779983e-01 2.00682282e-01 3.31059903e-01 -1.08738256e+00
-1.16166309e-01 1.57901138e-01 -8.45358789e-01 1.27553672e-01
5.52144647e-01 4.03447330e-01 1.04559469e+00 -1.02132189e+00
5.20495117e-01 1.35060704e+00 7.57215142e-01 4.66149777e-01
-1.09786975e+00 -8.02827775e-01 2.98182011e-01 1.24168023e-01
-1.43318713e+00 -1.59914091e-01 8.14344764e-01 -4.30112004e-01
1.18040562e+00 1.87298939e-01 1.01827455e+00 1.11108327e+00
5.70832312e-01 9.57210779e-01 6.41392171e-01 2.11252202e-03
-1.58128500e-01 -7.14545071e-01 -7.97367319e-02 7.66254663e-01
-3.37248622e-03 1.85198307e-01 -8.06662619e-01 1.09179996e-01
1.08465743e+00 -1.12792023e-01 -1.74994394e-01 -4.13700372e-01
-1.43472564e+00 6.60508573e-01 7.84260511e-01 -2.25655474e-02
-4.37096149e-01 5.99539876e-01 4.17654753e-01 -2.29308754e-01
4.79803681e-01 1.96507573e-01 -6.16380334e-01 -3.34359229e-01
-7.88088977e-01 6.76053047e-01 4.16022778e-01 9.43840981e-01
4.43107367e-01 -9.86287072e-02 1.65656283e-02 8.09145629e-01
6.27399802e-01 5.28287947e-01 2.07798004e-01 -9.35438335e-01
6.30666852e-01 5.32315910e-01 -1.05625149e-02 -1.04273009e+00
-9.50647175e-01 -9.08595085e-01 -9.49000597e-01 4.60410155e-02
2.73367524e-01 8.86675864e-02 -9.78560448e-01 1.94702864e+00
7.89417863e-01 -2.57886767e-01 -5.53228319e-01 1.35899520e+00
7.98355997e-01 4.59197462e-01 1.33445457e-01 2.62287825e-01
1.64820242e+00 -1.17528415e+00 -4.02806938e-01 -5.05060077e-01
2.71167874e-01 -7.09642112e-01 1.15842080e+00 2.96226948e-01
-1.21367884e+00 -7.71640599e-01 -9.15168881e-01 -4.77555484e-01
6.62354380e-03 1.44126251e-01 6.31403446e-01 8.90181810e-02
-7.38775551e-01 4.16044205e-01 -1.06691027e+00 -1.96842432e-01
1.79785192e-01 6.24214590e-01 -5.54943979e-01 6.23620525e-02
-1.28120148e+00 8.31718922e-01 7.58905429e-04 4.74093467e-01
-7.09772825e-01 -7.94998348e-01 -1.01873589e+00 -1.33658618e-01
5.48314273e-01 -1.09060943e+00 1.32104445e+00 -2.37820372e-01
-1.50315917e+00 7.40883112e-01 -1.83872171e-02 -1.67676389e-01
7.59588003e-01 -1.06446993e+00 5.93105331e-02 8.61894786e-02
1.29205838e-01 8.79969358e-01 6.09940410e-01 -9.56568360e-01
-1.88039795e-01 -7.06348300e-01 5.13096601e-02 6.06395841e-01
-6.58942312e-02 -2.94702470e-01 -1.02855933e+00 -9.05676842e-01
2.79503882e-01 -1.18154383e+00 -3.04704636e-01 1.97967231e-01
-5.50257325e-01 -1.71229199e-01 5.25640130e-01 -8.92668664e-01
1.19459426e+00 -1.79854643e+00 6.85289681e-01 1.70124229e-02
4.12532330e-01 1.44806489e-01 4.82826727e-03 4.68258053e-01
1.12187386e-01 -2.93139368e-01 7.61085330e-03 -5.14190137e-01
1.06090344e-01 2.08552435e-01 2.93480605e-01 4.64284509e-01
5.61865866e-02 1.04349959e+00 -7.02152073e-01 -5.91714442e-01
4.28857893e-01 7.81268001e-01 -9.15783942e-01 2.75366932e-01
-1.35880768e-01 6.99100673e-01 -4.60753024e-01 7.97771454e-01
5.08468568e-01 -3.12097639e-01 1.08660251e-01 -6.67513132e-01
-4.34022322e-02 1.95948347e-01 -1.08246577e+00 2.02628207e+00
-4.32549924e-01 2.97219604e-01 2.24950612e-01 -6.33811355e-01
7.08922565e-01 1.76878095e-01 8.44961643e-01 -7.47194469e-01
4.43849862e-01 1.23368777e-01 6.09968975e-02 -3.42659682e-01
3.42950225e-01 5.62316040e-03 -1.75152496e-01 9.46786255e-02
1.44407144e-02 -3.84786837e-02 1.15483716e-01 1.48614654e-02
9.54147279e-01 6.14317536e-01 3.09515923e-01 -3.26283187e-01
2.67830759e-01 -3.80087048e-01 6.58414960e-01 1.58048809e-01
-1.93168610e-01 8.44703615e-01 2.46051639e-01 -5.32846451e-01
-8.40584934e-01 -1.28130746e+00 9.57033783e-02 1.26250708e+00
1.72962695e-01 -6.08412921e-01 -7.92153060e-01 -2.39571527e-01
1.79944396e-01 9.24087092e-02 -6.65909410e-01 -9.86369848e-02
-9.44981813e-01 -7.17057467e-01 3.71678263e-01 9.84631360e-01
5.42376399e-01 -9.46497381e-01 -1.05414283e+00 4.65969801e-01
-5.23166955e-01 -9.40263093e-01 -6.71273172e-01 7.51913711e-02
-7.83825397e-01 -1.04292870e+00 -9.53507662e-01 -4.25510019e-01
5.00367880e-01 -1.66587800e-01 1.42972696e+00 1.24623552e-01
-5.78257501e-01 8.99857134e-02 -2.26299435e-01 -1.04841098e-01
4.40979332e-01 2.63520598e-01 1.18297450e-01 -3.76731813e-01
-1.34680122e-02 -6.15552306e-01 -1.03549933e+00 5.57642460e-01
-4.67364669e-01 2.52442807e-01 8.11460376e-01 9.45654750e-01
6.50273681e-01 -3.12917978e-01 2.48725310e-01 -3.15667152e-01
1.84848443e-01 -2.19149172e-01 -2.66778588e-01 -9.33490023e-02
-4.21177074e-02 -7.63349384e-02 3.19895178e-01 -4.78850424e-01
-8.00629675e-01 2.05041826e-01 -7.08604634e-01 -4.75903481e-01
7.03273714e-02 5.08246541e-01 -3.14722210e-01 1.90128595e-01
3.06120574e-01 8.88160150e-03 7.05049112e-02 -7.48113930e-01
2.59383142e-01 1.17761344e-01 8.14313829e-01 -8.00876439e-01
5.89619815e-01 4.38336611e-01 -3.73078138e-02 -7.07840025e-01
-6.81136787e-01 -3.68642300e-01 -8.11547339e-01 -4.64141130e-01
9.27363813e-01 -1.03220344e+00 -1.03205276e+00 7.25897551e-01
-1.01798272e+00 -4.72522080e-01 4.08069454e-02 5.72308183e-01
-6.21250570e-01 2.88432389e-01 -8.62568796e-01 -6.26606703e-01
-7.21881568e-01 -1.31789160e+00 1.53814769e+00 1.73622053e-02
-6.37145758e-01 -5.61238229e-01 -3.84754166e-02 6.85628653e-01
1.66712612e-01 3.47708255e-01 7.27477610e-01 -1.06774934e-01
-4.87968206e-01 -2.01549068e-01 -1.50180191e-01 -1.97025966e-02
-2.60925412e-01 -1.96810082e-01 -4.04783100e-01 -5.37137985e-01
-5.06613016e-01 -3.86273175e-01 4.14690882e-01 5.08152306e-01
9.02145863e-01 2.32302193e-02 -2.89641619e-01 6.92453861e-01
9.87133920e-01 -1.82255447e-01 6.68537915e-01 3.71458709e-01
9.71251428e-01 5.89087367e-01 6.98413432e-01 5.93249500e-01
7.94652462e-01 1.15894711e+00 5.97212553e-01 -3.17260176e-01
-2.79020935e-01 -2.91597903e-01 5.19386716e-02 9.42700744e-01
-3.15918475e-01 -1.16420507e-01 -1.10551572e+00 3.86293232e-01
-1.73125505e+00 -4.75418985e-01 -8.85024294e-02 1.95629311e+00
7.58589983e-01 4.16482717e-01 3.33606571e-01 5.70012210e-03
3.63068610e-01 1.81483731e-01 -6.71737671e-01 5.78504875e-02
4.11836147e-01 3.03899616e-01 1.88241497e-01 3.28462511e-01
-1.04124117e+00 7.79464006e-01 5.49307871e+00 7.33819664e-01
-1.02409959e+00 6.76296875e-02 3.46108466e-01 -4.01051730e-01
1.73861254e-02 -3.03713709e-01 -8.05999875e-01 5.58893263e-01
3.69415134e-01 3.25251997e-01 1.89043973e-02 7.55704343e-01
1.21808887e-01 -2.76239455e-01 -1.07667208e+00 9.04004931e-01
-1.32198706e-01 -8.32796752e-01 -1.51621461e-01 2.59222656e-01
2.32173234e-01 9.31769758e-02 -6.02693744e-02 3.49484146e-01
1.15043446e-01 -7.85067439e-01 9.51040924e-01 3.64126712e-01
7.78770626e-01 -7.59338796e-01 7.02936888e-01 3.90834630e-01
-1.73588634e+00 1.18406929e-01 1.10723086e-01 -6.12281263e-02
5.02382815e-01 6.03676736e-01 -2.77941197e-01 7.63743758e-01
1.14526927e+00 5.36754727e-01 -3.50588232e-01 8.27430129e-01
-2.13200718e-01 3.03540323e-02 -6.80303514e-01 1.90295786e-01
8.66069645e-03 8.12752992e-02 4.96437669e-01 9.47286069e-01
2.29385048e-01 1.50750160e-01 3.72346789e-01 4.97650355e-01
1.17599808e-01 -1.54183596e-01 -7.73069784e-02 4.11418855e-01
2.75639564e-01 1.01124489e+00 -6.61393285e-01 -1.24793053e-01
-2.22057909e-01 9.28184390e-01 4.52187896e-01 6.48326576e-02
-1.25228524e+00 -8.00869092e-02 8.58407676e-01 4.47457790e-01
1.28825650e-01 -5.43265522e-01 -2.23137870e-01 -1.11106873e+00
3.34907562e-01 -9.76644158e-01 4.50479716e-01 -7.61788964e-01
-9.97988939e-01 6.36799157e-01 2.09594622e-01 -1.23601639e+00
-2.72118211e-01 -4.67461824e-01 -2.69764185e-01 5.61841011e-01
-9.30826128e-01 -1.39933383e+00 -5.26205838e-01 5.16850352e-01
7.69241989e-01 4.01791662e-01 6.73304915e-01 6.24692738e-01
-6.60433114e-01 6.68786943e-01 -4.32251334e-01 9.87034589e-02
5.80021679e-01 -1.00550270e+00 5.51737726e-01 5.19877791e-01
-1.65342629e-01 7.60731041e-01 7.02669263e-01 -7.93846965e-01
-1.44478273e+00 -6.98289633e-01 7.76603281e-01 -3.46087307e-01
3.46887052e-01 -4.64306682e-01 -6.56881452e-01 5.83776832e-01
-1.23016134e-01 1.16481088e-01 4.87826496e-01 2.68073410e-01
-2.86343217e-01 -7.10707977e-02 -8.24411452e-01 7.82966018e-01
1.35928953e+00 -1.70377925e-01 -5.55082023e-01 2.41186157e-01
5.95709801e-01 -9.78226185e-01 -1.07227123e+00 8.63807797e-01
1.10963702e+00 -9.18740749e-01 1.36826634e+00 -2.05588818e-01
5.89523613e-01 -1.80604160e-01 -8.57661292e-02 -9.77173984e-01
-4.42049414e-01 -4.32175517e-01 -4.43810403e-01 7.44167268e-01
8.64123777e-02 -3.85775208e-01 1.05210340e+00 4.14671838e-01
-2.26922169e-01 -1.37000930e+00 -1.12692964e+00 -5.84207475e-01
-2.41592787e-02 -4.34853077e-01 3.70161623e-01 4.98018414e-01
-2.05253854e-01 2.42066935e-01 -7.83122540e-01 6.46603778e-02
7.39923000e-01 8.45131651e-02 1.04086041e+00 -9.40665424e-01
-5.30427337e-01 -3.10007930e-01 -4.07125503e-01 -1.48543501e+00
-1.62342802e-01 -3.01458746e-01 2.62051523e-01 -1.51634192e+00
7.98248872e-02 -3.16589862e-01 -1.26211718e-02 5.46413124e-01
-1.88821971e-01 5.13809621e-01 2.43286267e-01 1.07932016e-01
-5.05688608e-01 8.15717280e-01 1.45143032e+00 4.72871289e-02
5.76238614e-03 -8.71946216e-02 -1.73847303e-01 8.57870519e-01
6.18673325e-01 -2.25215182e-01 -2.67020643e-01 -7.45820105e-01
3.08559556e-02 1.59193873e-01 6.24729395e-01 -1.30624115e+00
1.64157391e-01 1.22873997e-02 6.14483714e-01 -1.10381651e+00
8.06315482e-01 -6.42817199e-01 4.59181458e-01 7.81560957e-01
-7.52544999e-02 4.14231658e-01 1.19747117e-01 3.90470147e-01
-1.44991621e-01 4.26126152e-01 7.81447172e-01 -2.70728230e-01
-6.37641668e-01 5.53575933e-01 -1.02943264e-01 1.07478641e-01
9.16036665e-01 -1.58295706e-01 1.99811347e-02 -2.52721936e-01
-9.27525938e-01 4.41440701e-01 2.63830602e-01 5.92586696e-01
5.78981340e-01 -1.36009681e+00 -5.53336620e-01 2.20708311e-01
1.25596644e-02 4.93276268e-01 6.57230616e-01 1.17053711e+00
-8.24683607e-01 3.71581852e-01 -3.08551490e-01 -7.73680389e-01
-1.21110857e+00 1.55805051e-01 2.70066708e-01 -5.08040369e-01
-8.27277303e-01 1.20604038e+00 3.96030068e-01 -4.83923495e-01
2.41267785e-01 -1.18910216e-01 2.54908085e-01 -5.52358888e-02
1.06193170e-01 4.60001886e-01 1.53082997e-01 -8.25655580e-01
-7.01092660e-01 8.81434202e-01 -1.51097346e-02 -1.29304258e-02
1.21331775e+00 -9.31192189e-02 -1.16396332e-02 6.57783896e-02
1.20676553e+00 -1.64017946e-01 -1.45625448e+00 -1.05718814e-01
-3.77327383e-01 -4.99561220e-01 -1.84489265e-01 -7.40184307e-01
-1.25759006e+00 9.36444402e-01 3.76331329e-01 -5.29421210e-01
1.15579700e+00 9.37126428e-02 1.13866949e+00 2.97864955e-02
7.05503583e-01 -1.16450441e+00 4.99565214e-01 4.78656292e-01
1.25224316e+00 -9.70306277e-01 4.09418225e-01 -6.59143388e-01
-3.82458210e-01 6.93680525e-01 1.08803439e+00 -1.63705185e-01
6.35491192e-01 3.63893270e-01 3.42468880e-02 -3.76794159e-01
-6.46508574e-01 -1.39675543e-01 6.27115011e-01 3.46343100e-01
6.82708740e-01 7.74908587e-02 -4.58809108e-01 6.33442521e-01
-4.38185662e-01 -1.51382834e-01 -3.63000780e-01 1.11055875e+00
-1.68125913e-01 -1.05537689e+00 -4.23695892e-01 3.83640766e-01
-4.15077090e-01 6.65349066e-02 -1.65219456e-01 1.12110281e+00
2.34104812e-01 5.25221109e-01 7.10785985e-02 -6.26817524e-01
4.88802254e-01 -2.25040868e-01 5.47686636e-01 -4.23797488e-01
-6.88890517e-01 6.20640397e-01 1.36612579e-01 -1.04811335e+00
-2.08663836e-01 -5.89455128e-01 -1.39936757e+00 -3.90714556e-01
-2.02329919e-01 -1.97849095e-01 4.48088199e-01 8.02472591e-01
4.87377971e-01 7.43243575e-01 4.10654545e-02 -1.48376572e+00
-4.17490780e-01 -9.79942501e-01 -2.34107301e-01 4.54868525e-01
1.38737038e-01 -1.28399014e+00 1.14829808e-01 -1.89039081e-01] | [7.137869358062744, -0.6967793107032776] |
1ef79cee-fbe8-4005-bcbd-c8ebf4d1bbb4 | referring-transformer-a-one-step-approach-to | 2106.03089 | null | https://arxiv.org/abs/2106.03089v2 | https://arxiv.org/pdf/2106.03089v2.pdf | Referring Transformer: A One-step Approach to Multi-task Visual Grounding | As an important step towards visual reasoning, visual grounding (e.g., phrase localization, referring expression comprehension/segmentation) has been widely explored Previous approaches to referring expression comprehension (REC) or segmentation (RES) either suffer from limited performance, due to a two-stage setup, or require the designing of complex task-specific one-stage architectures. In this paper, we propose a simple one-stage multi-task framework for visual grounding tasks. Specifically, we leverage a transformer architecture, where two modalities are fused in a visual-lingual encoder. In the decoder, the model learns to generate contextualized lingual queries which are then decoded and used to directly regress the bounding box and produce a segmentation mask for the corresponding referred regions. With this simple but highly contextualized model, we outperform state-of-the-arts methods by a large margin on both REC and RES tasks. We also show that a simple pre-training schedule (on an external dataset) further improves the performance. Extensive experiments and ablations illustrate that our model benefits greatly from contextualized information and multi-task training. | ['Leonid Sigal', 'Muchen Li'] | 2021-06-06 | null | http://proceedings.neurips.cc/paper/2021/hash/a376802c0811f1b9088828288eb0d3f0-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/a376802c0811f1b9088828288eb0d3f0-Paper.pdf | neurips-2021-12 | ['referring-expression-segmentation'] | ['computer-vision'] | [ 4.44219381e-01 3.32718730e-01 -2.54959404e-01 -4.84522909e-01
-1.53690052e+00 -7.72095263e-01 7.78961599e-01 4.96941507e-02
-3.81127328e-01 4.14481878e-01 4.07224268e-01 -4.73325878e-01
6.11188829e-01 -3.86762768e-01 -1.06596851e+00 -4.48169827e-01
5.10994375e-01 4.63525593e-01 1.56144276e-01 -1.49649113e-01
5.60140610e-02 4.27183025e-02 -1.28716373e+00 5.60458839e-01
8.69111240e-01 9.55123842e-01 2.56467700e-01 5.70400715e-01
-3.73713374e-01 7.70632505e-01 -4.79452461e-01 -6.57602668e-01
-1.05086781e-01 -6.22389853e-01 -9.79082346e-01 1.62506402e-01
4.73396331e-01 -3.14944267e-01 1.05497912e-01 9.29485261e-01
3.60140592e-01 -6.40655756e-02 6.00572228e-01 -1.17102075e+00
-8.36370051e-01 6.38886154e-01 -8.34968925e-01 -1.48855865e-01
5.19625962e-01 1.39905155e-01 1.26864910e+00 -1.11397398e+00
6.16992235e-01 1.37568212e+00 4.13140059e-01 5.79901755e-01
-1.51121116e+00 -4.18558121e-01 6.00865006e-01 4.69870120e-02
-1.34941077e+00 -6.07225001e-01 7.63476908e-01 -3.21768403e-01
8.99771214e-01 4.47128639e-02 5.01537800e-01 1.23489320e+00
-2.79054493e-01 1.19711721e+00 1.25041473e+00 -4.54278290e-01
8.49139765e-02 3.95677425e-02 -1.80043474e-01 1.07342255e+00
-1.79628059e-01 -2.53525585e-01 -7.00613260e-01 2.90146440e-01
6.76051974e-01 -1.91520035e-01 -3.08711052e-01 -4.58543420e-01
-1.18810225e+00 7.43974805e-01 6.97447240e-01 1.62630141e-01
-2.15308532e-01 4.29404140e-01 3.98571670e-01 1.07688941e-01
6.02333724e-01 1.80215657e-01 -1.49150074e-01 2.09137425e-02
-1.17822206e+00 1.36300683e-01 5.51042438e-01 1.13090193e+00
9.54675972e-01 -1.28111929e-01 -5.64172745e-01 8.03686142e-01
3.59135270e-01 4.57633883e-01 7.01688882e-03 -8.36916387e-01
8.31368148e-01 5.51716506e-01 -8.87331292e-02 -5.28281689e-01
-4.04200554e-01 -3.41887534e-01 -4.55612183e-01 1.15710422e-01
5.46004653e-01 -9.20234248e-02 -1.32238281e+00 1.96386111e+00
1.83833569e-01 -2.09701224e-03 2.40366794e-02 1.07426095e+00
8.04906130e-01 5.31883955e-01 4.91575718e-01 2.11140916e-01
1.79353988e+00 -1.40616310e+00 -6.36534691e-01 -6.87010527e-01
7.10559070e-01 -6.59819782e-01 1.56849515e+00 6.30528033e-02
-1.27945185e+00 -2.81346917e-01 -9.20450211e-01 -8.04067910e-01
-2.98426211e-01 2.57352680e-01 5.42920828e-01 2.83261329e-01
-1.15761065e+00 -7.65700191e-02 -9.36659098e-01 -3.76155525e-01
7.08254099e-01 -1.17155559e-01 -2.83617258e-01 2.40508723e-03
-8.26335311e-01 9.48683143e-01 3.67702469e-02 1.67898089e-01
-1.15682960e+00 -6.42971575e-01 -1.16669881e+00 2.09569007e-01
6.17409110e-01 -1.03871346e+00 1.50123215e+00 -1.04680943e+00
-1.41845882e+00 1.28025138e+00 -5.24950504e-01 -3.95250797e-01
6.21310353e-01 -4.56187278e-01 1.41837940e-01 4.47526664e-01
3.34332824e-01 1.00995338e+00 8.83798957e-01 -1.50348115e+00
-4.58321750e-01 -3.23198676e-01 2.78784901e-01 2.99073637e-01
9.28658321e-02 1.33011356e-01 -1.09594238e+00 -6.33885324e-01
9.90338102e-02 -8.49754453e-01 -1.29736215e-01 1.66994780e-01
-7.18605816e-01 -8.02583396e-02 5.39435327e-01 -8.38585734e-01
1.06942976e+00 -2.16113806e+00 4.80053842e-01 -7.05066994e-02
2.21052989e-01 -1.67751223e-01 -1.31242022e-01 2.33902171e-01
5.74479774e-02 3.25344443e-01 -4.12959188e-01 -9.49367225e-01
1.89872608e-01 3.93152609e-02 -4.85508651e-01 2.87652254e-01
4.07941461e-01 1.46806180e+00 -6.58765197e-01 -6.88930452e-01
1.07883438e-01 4.95941848e-01 -5.35731494e-01 3.84260386e-01
-7.39733100e-01 6.75956786e-01 -3.96255583e-01 9.04457748e-01
3.58022809e-01 -5.60581267e-01 7.62511194e-02 -3.45144540e-01
1.09658211e-01 3.09166282e-01 -5.28569281e-01 2.33018494e+00
-9.06548440e-01 7.06074536e-01 2.30791762e-01 -9.88828659e-01
5.21075547e-01 2.85786718e-01 -6.65900186e-02 -1.09395170e+00
8.98885205e-02 1.14978693e-01 -4.50561851e-01 -4.83321309e-01
3.25663716e-01 -1.24631688e-01 -4.37344283e-01 6.49149597e-01
9.20304060e-02 -1.85844660e-01 -7.50372978e-03 4.00454640e-01
7.38478065e-01 7.86124289e-01 1.39091253e-01 1.32939368e-01
2.38568455e-01 8.47178176e-02 2.95476615e-01 6.56904161e-01
-6.14925809e-02 9.11263764e-01 9.12450373e-01 -5.67191318e-02
-1.11689913e+00 -9.95901942e-01 2.54896373e-01 1.46966910e+00
2.10373536e-01 -5.09284616e-01 -9.05450284e-01 -7.34846473e-01
-1.71942204e-01 9.03628588e-01 -6.28806174e-01 1.71376001e-02
-7.19524920e-01 -3.04659009e-01 5.51604986e-01 8.33922744e-01
5.22067964e-01 -9.67954218e-01 -7.60229886e-01 -4.54288721e-02
-2.79428452e-01 -1.65069342e+00 -4.07476813e-01 2.53130347e-01
-6.34813249e-01 -7.36322165e-01 -8.91897500e-01 -8.16083014e-01
7.72452533e-01 2.59455085e-01 1.33912599e+00 8.14949870e-02
2.71415716e-04 3.63815576e-01 -3.19453865e-01 -2.48973474e-01
-1.73679307e-01 1.78181320e-01 -7.72877693e-01 -1.39454473e-02
9.18641463e-02 -4.49537814e-01 -7.57420778e-01 1.67469352e-01
-6.99224532e-01 6.07443571e-01 8.50967765e-01 8.14423025e-01
7.90361464e-01 -9.34727192e-01 3.39020818e-01 -1.10942245e+00
5.70003748e-01 -2.59843737e-01 -6.63486063e-01 6.70603812e-01
-4.33315098e-01 3.47862035e-01 4.27834809e-01 -1.41643956e-01
-1.15939486e+00 1.01862215e-01 -1.49143711e-01 -6.39413655e-01
-8.23766440e-02 5.33366859e-01 -2.16250896e-01 2.41884619e-01
3.81626099e-01 1.02829888e-01 -2.60476738e-01 -5.14507711e-01
1.04894066e+00 3.36108625e-01 7.43556857e-01 -8.15141618e-01
6.96485639e-01 5.04560113e-01 -1.47469178e-01 -2.56219804e-01
-1.21074677e+00 -1.60871729e-01 -6.74850047e-01 -4.06203978e-02
1.40917754e+00 -1.30745125e+00 -6.52291059e-01 1.72498599e-02
-1.28310597e+00 -7.50264108e-01 -1.93819515e-02 -3.53348367e-02
-6.91422820e-01 1.23090275e-01 -5.19583404e-01 -7.11257339e-01
-4.35456067e-01 -1.28141737e+00 1.76505363e+00 9.93590429e-02
-1.85632765e-01 -9.51862752e-01 -4.19272989e-01 6.02496445e-01
4.07705337e-01 2.82886833e-01 1.19786263e+00 -4.31143939e-01
-1.03353941e+00 2.13481441e-01 -6.05731070e-01 -6.91427812e-02
-3.13010693e-01 -2.40807220e-01 -1.23466241e+00 -6.54496178e-02
-3.80415291e-01 -8.95324409e-01 1.21937120e+00 1.50281742e-01
1.29421139e+00 -5.98103739e-02 -4.79141504e-01 9.99397635e-01
1.23600543e+00 -1.58445179e-01 3.59134883e-01 2.36331686e-01
1.03251576e+00 5.89203179e-01 6.46527648e-01 5.30918054e-02
8.47232819e-01 7.42957711e-01 5.23538172e-01 -6.59771740e-01
-3.88596803e-01 -6.08733177e-01 2.79304385e-01 4.09107804e-01
8.66408721e-02 -2.64129847e-01 -1.08710706e+00 6.89257920e-01
-1.98304176e+00 -7.59146452e-01 2.75248945e-01 1.72973549e+00
9.60744619e-01 5.52517883e-02 1.26441926e-01 -3.07298422e-01
3.77197206e-01 3.71068686e-01 -5.94480097e-01 -4.29117054e-01
-1.60405725e-01 2.79573649e-01 3.21324408e-01 6.70152545e-01
-1.05712497e+00 1.55576968e+00 6.05789185e+00 6.42699838e-01
-1.20914412e+00 3.79902750e-01 8.17593396e-01 -1.06336713e-01
-7.29063988e-01 2.41009012e-01 -6.42969728e-01 6.07960261e-02
6.44172072e-01 7.99283683e-02 3.93593758e-01 6.63063407e-01
6.28521070e-02 -2.85730511e-01 -1.29143929e+00 1.01156127e+00
3.11831385e-01 -1.22353137e+00 8.79323855e-02 -1.52302042e-01
5.31222105e-01 -3.32219377e-02 2.09906578e-01 3.98366958e-01
2.56436944e-01 -1.13790095e+00 1.17854226e+00 2.31831625e-01
1.23908365e+00 -5.32181263e-01 1.79344833e-01 4.09324728e-02
-1.22037697e+00 9.85537395e-02 3.30822766e-02 2.59676367e-01
6.47883713e-01 1.84918150e-01 -6.13830566e-01 5.38958073e-01
5.64102173e-01 5.93339145e-01 -6.76492810e-01 5.91011882e-01
-8.16513956e-01 6.65446937e-01 -2.52924293e-01 8.86497274e-02
4.37161118e-01 -4.66479510e-02 2.13517249e-01 1.27962267e+00
-3.60028706e-02 -1.63005799e-01 1.73038945e-01 1.28519988e+00
-2.99566031e-01 4.17188108e-02 -4.24164563e-01 -6.80954941e-03
2.74578720e-01 1.39212513e+00 -6.58979654e-01 -3.55495930e-01
-8.68220210e-01 1.31771851e+00 7.38189578e-01 8.22007537e-01
-1.03815544e+00 -2.14580804e-01 3.24457526e-01 -4.31532636e-02
5.77743649e-01 -1.84812471e-01 -5.88889301e-01 -1.38973939e+00
1.92798153e-01 -7.86892354e-01 4.48543847e-01 -1.02241671e+00
-8.39501381e-01 6.39213979e-01 -3.78529690e-02 -8.54840517e-01
-4.92372423e-01 -4.63943750e-01 -5.70338666e-01 1.01550364e+00
-1.92627454e+00 -1.85829151e+00 -4.41039264e-01 6.68942213e-01
6.53184056e-01 2.14981988e-01 8.45274389e-01 1.72821209e-01
-6.77028835e-01 6.98440731e-01 -5.10156989e-01 2.44805157e-01
7.22000182e-01 -1.39888918e+00 5.75648725e-01 1.03500307e+00
4.46177036e-01 5.80242574e-01 5.54950356e-01 -4.46894586e-01
-1.30070961e+00 -9.62925076e-01 7.28067219e-01 -4.66822982e-01
6.38294399e-01 -1.00594211e+00 -8.29108477e-01 1.13868904e+00
5.46702623e-01 -2.46983697e-03 6.03728712e-01 3.80499214e-01
-6.88715398e-01 1.82505250e-01 -6.71283662e-01 9.04923558e-01
1.10400367e+00 -1.07099998e+00 -6.56201363e-01 1.80536836e-01
9.18152153e-01 -6.15192711e-01 -4.73217487e-01 1.05733261e-01
4.98725235e-01 -9.03381884e-01 9.05326128e-01 -5.71451962e-01
6.65376365e-01 -2.73880035e-01 -9.35787857e-02 -1.10201669e+00
1.17798038e-01 -7.49656498e-01 4.02281852e-03 1.30535996e+00
7.67732263e-01 -2.00334445e-01 6.32940412e-01 6.49571419e-01
-3.13407063e-01 -9.05038059e-01 -7.68378913e-01 -2.37834021e-01
4.91539352e-02 -4.98303980e-01 3.06973934e-01 6.41441882e-01
3.75255495e-02 8.73764634e-01 -3.07051569e-01 6.60728812e-02
4.20722216e-01 5.83111882e-01 8.49597275e-01 -5.95670164e-01
-2.00210541e-01 -4.12937880e-01 1.96459085e-01 -1.61478078e+00
3.88771355e-01 -9.74369824e-01 1.60676911e-01 -1.84074330e+00
2.81394780e-01 -2.58754253e-01 -9.30261761e-02 8.31394494e-01
-4.66234595e-01 3.01571041e-01 3.99075449e-01 2.83250809e-01
-9.55966115e-01 5.37701428e-01 1.30526519e+00 -7.85705596e-02
-1.12180464e-01 -3.79568607e-01 -9.61640775e-01 6.17778480e-01
4.20115560e-01 -2.73085475e-01 -4.59527999e-01 -9.50752199e-01
3.10820162e-01 1.61652118e-01 6.62892342e-01 -4.42394465e-01
2.88558632e-01 1.13805585e-01 2.99786985e-01 -6.56358600e-01
5.70682526e-01 -6.22412264e-01 -4.07352895e-01 2.70062536e-02
-4.85965401e-01 2.25824416e-01 1.90943897e-01 4.92226869e-01
-3.89980406e-01 8.91026482e-02 5.17985582e-01 -1.15608782e-01
-6.99202597e-01 6.28317818e-02 -3.32031474e-02 2.88340896e-01
9.50259805e-01 -2.72809826e-02 -5.02273083e-01 -5.77842534e-01
-7.46759057e-01 5.01902521e-01 5.61813056e-01 2.95125812e-01
4.98783976e-01 -1.10153306e+00 -5.06684840e-01 -8.54079500e-02
2.94827402e-01 3.71777475e-01 6.01716749e-02 1.03102624e+00
-4.41207200e-01 3.81704509e-01 -4.60903655e-04 -9.05802906e-01
-1.01986527e+00 6.24164760e-01 3.32041502e-01 -2.22402468e-01
-6.77474141e-01 9.92072642e-01 5.90881348e-01 1.77879888e-03
2.97002673e-01 -4.00628924e-01 6.95601925e-02 1.94895923e-01
2.84938931e-01 -3.04941684e-01 -1.68513924e-01 -6.14809692e-01
-3.54281604e-01 8.21161866e-01 -2.29585469e-01 -5.12219548e-01
1.09144616e+00 -3.37029517e-01 1.57442600e-01 3.70233417e-01
1.13852143e+00 3.19089331e-02 -1.55980170e+00 -3.51318032e-01
1.46768928e-01 -2.19020084e-01 -1.68292541e-02 -9.60515022e-01
-9.88982856e-01 1.23957324e+00 1.17115222e-01 -1.33147106e-01
1.30649018e+00 5.65855443e-01 6.63779855e-01 1.63412780e-01
2.10689798e-01 -8.79327476e-01 2.53485739e-01 2.72034347e-01
1.00100911e+00 -1.30829942e+00 -1.12709686e-01 -5.74926734e-01
-1.06303763e+00 8.73898983e-01 6.67119563e-01 1.23739220e-01
1.43320784e-01 3.07280332e-01 4.31809068e-01 -2.59943724e-01
-8.57819378e-01 -6.55854881e-01 3.45449895e-01 4.05562937e-01
5.46908796e-01 -2.03668132e-01 7.99139887e-02 5.79675972e-01
-8.53323713e-02 -2.98663676e-01 3.22717614e-03 8.57702732e-01
-2.20285341e-01 -9.83785868e-01 -9.34285894e-02 -2.03656573e-02
-4.96573418e-01 -3.83140385e-01 -3.92114490e-01 8.38945985e-01
-2.93299586e-01 7.61777401e-01 1.58498213e-01 5.24686091e-03
4.14606482e-01 2.88252145e-01 5.23889959e-01 -6.64890289e-01
-6.21535361e-01 4.16967422e-01 2.46494785e-01 -9.66122508e-01
-4.36286151e-01 -5.89294732e-01 -1.38393950e+00 1.87455684e-01
-1.24778822e-01 -1.97916135e-01 6.75276220e-01 1.05211556e+00
5.18223524e-01 6.22364223e-01 8.73890147e-02 -7.21217334e-01
-1.95263326e-01 -7.91521430e-01 1.26115326e-02 6.02938294e-01
4.49583203e-01 -5.63951373e-01 -2.50889547e-02 2.11524725e-01] | [10.660099983215332, 1.5343003273010254] |
2f4fbb3e-f829-4ec8-b7b1-8d8a12a8eb3e | nips-conversational-intelligence-challenge | null | null | https://aclanthology.org/C18-1312 | https://aclanthology.org/C18-1312.pdf | NIPS Conversational Intelligence Challenge 2017 Winner System: Skill-based Conversational Agent with Supervised Dialog Manager | We present bot{\#}1337: a dialog system developed for the 1st NIPS Conversational Intelligence Challenge 2017 (ConvAI). The aim of the competition was to implement a bot capable of conversing with humans based on a given passage of text. To enable conversation, we implemented a set of skills for our bot, including chit-chat, topic detection, text summarization, question answering and question generation. The system has been trained in a supervised setting using a dialogue manager to select an appropriate skill for generating a response. The latter allows a developer to focus on the skill implementation rather than the finite state machine based dialog manager. The proposed system bot{\#}1337 won the competition with an average dialogue quality score of 2.78 out of 5 given by human evaluators. Source code and trained models for the bot{\#}1337 are available on GitHub. | ['Yurii Kuratov', 'Idris Yusupov'] | 2018-08-01 | nips-conversational-intelligence-challenge-1 | https://aclanthology.org/C18-1312 | https://aclanthology.org/C18-1312.pdf | coling-2018-8 | ['goal-oriented-dialog', 'short-text-conversation', 'goal-oriented-dialogue-systems'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 1.64362371e-01 8.44066322e-01 5.77183425e-01 -3.83364022e-01
-7.12674677e-01 -6.50964200e-01 7.88466990e-01 2.29074107e-03
-4.99617785e-01 8.07392061e-01 4.44702774e-01 -1.43628493e-01
1.99307457e-01 -3.51740807e-01 -7.77567551e-02 -3.20678502e-01
2.65499085e-01 1.12948143e+00 4.04028654e-01 -6.46494985e-01
4.72314000e-01 -2.02880487e-01 -1.03934360e+00 7.02293336e-01
9.30955708e-01 5.20036042e-01 5.83925903e-01 1.39258170e+00
-2.19330534e-01 1.41235137e+00 -1.36332107e+00 -4.42434669e-01
-1.85416907e-01 -9.50169325e-01 -1.67413604e+00 -9.37445238e-02
6.94541484e-02 -4.95869130e-01 -6.84341639e-02 7.50824451e-01
6.28556848e-01 3.37132037e-01 6.85772240e-01 -1.42889738e+00
-8.77361298e-02 9.90317166e-01 3.15721691e-01 1.65294811e-01
7.93891013e-01 6.56800747e-01 1.02361286e+00 -4.71284837e-01
6.74528003e-01 1.48464739e+00 2.27345929e-01 1.19806921e+00
-9.25080061e-01 -2.72864014e-01 -2.68217802e-01 1.60514325e-01
-5.44102192e-01 -5.30763030e-01 5.26743531e-01 -5.59541464e-01
1.29376757e+00 3.01962316e-01 3.45396727e-01 1.36000538e+00
-2.57057883e-02 8.72680485e-01 1.07859290e+00 -3.24694306e-01
4.15267527e-01 4.81392324e-01 3.27443957e-01 4.76445526e-01
-4.80823636e-01 -5.61606705e-01 -6.10592246e-01 -3.86143923e-01
4.24373090e-01 -7.53208876e-01 -5.56254908e-02 4.90890265e-01
-1.14303517e+00 1.10658169e+00 6.74589202e-02 2.83363789e-01
-4.12354827e-01 -1.15809001e-01 6.49360478e-01 4.74740237e-01
3.30991983e-01 1.02465081e+00 -3.68036747e-01 -7.62705505e-01
-2.29561061e-01 4.76553887e-01 1.68403757e+00 1.02709317e+00
3.04030418e-01 -2.26629257e-01 -9.39383984e-01 1.01791060e+00
1.04096800e-01 3.21694493e-01 3.27638239e-01 -1.57006323e+00
6.50789976e-01 7.97123909e-01 4.10030931e-01 -5.33318460e-01
-5.58641672e-01 5.58118105e-01 -6.18674397e-01 1.46719575e-01
8.44975770e-01 -9.17905390e-01 -3.80660772e-01 1.50585508e+00
2.09978446e-01 -3.99001122e-01 4.26583558e-01 5.39642930e-01
1.53141952e+00 7.38640189e-01 1.55762374e-01 -7.96643570e-02
1.69089234e+00 -1.37219453e+00 -7.56243885e-01 -1.88349411e-01
6.54708266e-01 -8.52241755e-01 1.12920475e+00 2.10821867e-01
-1.13498092e+00 -2.70522594e-01 -5.44919968e-01 1.30054876e-02
-6.84006140e-02 4.61406559e-02 2.42179576e-02 3.12509567e-01
-1.30907822e+00 3.62371951e-01 -3.54199290e-01 -8.36312056e-01
-1.37629360e-01 3.16605270e-01 -4.69114445e-02 4.61216092e-01
-1.33030570e+00 1.24768615e+00 1.54131800e-01 -2.46463194e-01
-7.74889410e-01 -2.20163852e-01 -6.16079748e-01 3.03597394e-02
3.85811538e-01 -6.68563843e-01 2.06388426e+00 -6.40567482e-01
-2.25765705e+00 8.94336820e-01 1.55935422e-01 -7.10957289e-01
6.66161358e-01 -1.83321267e-01 1.40392318e-01 4.87321168e-01
1.96060464e-01 7.05270886e-01 4.70044822e-01 -7.35677242e-01
-7.00621247e-01 2.00061217e-01 6.21011555e-01 4.64149058e-01
-5.08607328e-02 4.44480568e-01 -4.15233113e-02 4.79758671e-03
-6.89484954e-01 -9.45423424e-01 -2.06580088e-01 -8.62725019e-01
-7.83700943e-01 -8.96364093e-01 7.84095049e-01 -8.30956340e-01
8.28602672e-01 -1.47662258e+00 1.09543495e-01 -5.15086651e-01
1.74020618e-01 5.35150647e-01 -1.61799073e-01 1.00962830e+00
4.62925255e-01 -3.74857523e-02 -1.44918948e-01 -5.42879522e-01
9.13889110e-02 -1.73503220e-01 -9.44183171e-02 -1.76832214e-01
2.78259188e-01 7.21741557e-01 -1.00557363e+00 -5.42598665e-01
2.90708601e-01 2.08292734e-02 -5.89734435e-01 9.43412542e-01
-7.96466529e-01 7.35020459e-01 -6.15652502e-01 -4.20373194e-02
2.04159915e-01 -2.80369312e-01 6.49093017e-02 4.30687279e-01
-1.84277683e-01 8.98839951e-01 -5.30251384e-01 1.51653564e+00
-4.63472098e-01 6.95356488e-01 4.40193683e-01 -5.24620950e-01
8.17434192e-01 8.43780339e-01 2.06376359e-01 -1.38067618e-01
3.76386732e-01 -2.33777732e-01 2.83679128e-01 -7.91875184e-01
7.19296038e-01 3.92249972e-01 -4.04748827e-01 1.00875378e+00
3.14715624e-01 -5.16535580e-01 4.47239548e-01 6.76727235e-01
1.59830761e+00 -2.41548389e-01 2.01833218e-01 -2.14016184e-01
7.44118810e-01 4.72828358e-01 7.27234781e-02 1.02865374e+00
-4.61362749e-01 3.06913882e-01 9.00679708e-01 -1.57437205e-01
-8.50660563e-01 -4.52780783e-01 5.27419329e-01 1.48142099e+00
-2.49630302e-01 -4.69999582e-01 -1.42702985e+00 -8.20664287e-01
-6.09324872e-01 9.78101194e-01 -3.91943485e-01 4.99718785e-02
-7.17910945e-01 -6.82707876e-02 7.90011346e-01 -1.60755768e-01
9.58789945e-01 -2.21721578e+00 -9.66003597e-01 3.49965006e-01
-8.20707917e-01 -1.29972374e+00 -6.51816189e-01 -7.00812042e-03
-2.73548633e-01 -9.47767198e-01 -5.25075376e-01 -7.76033700e-01
2.02355415e-01 1.12443380e-02 1.16235387e+00 2.04704404e-01
3.04591432e-02 5.45217872e-01 -6.78611457e-01 -5.24667382e-01
-1.24065697e+00 5.75254202e-01 -2.85320401e-01 -3.65016580e-01
2.26678848e-01 -3.67881358e-01 -6.15170121e-01 4.68884796e-01
-4.15302098e-01 4.19926047e-01 3.12024027e-01 8.29388499e-01
-6.08484268e-01 -5.07946730e-01 8.31489325e-01 -1.10011971e+00
1.45494521e+00 -4.02077883e-01 -3.02937984e-01 1.89708650e-01
-1.33405954e-01 -9.68485996e-02 7.09701717e-01 -3.55976433e-01
-1.30689633e+00 5.49049117e-02 -3.88520032e-01 4.06080753e-01
-4.68504220e-01 8.01542699e-02 8.50664899e-02 4.08433110e-01
8.47821772e-01 1.65687785e-01 1.45966830e-02 -4.56046879e-01
3.26649159e-01 1.17818809e+00 5.01576304e-01 -5.75820863e-01
3.47886145e-01 -2.59436548e-01 -6.87184572e-01 -8.93773079e-01
-6.89451218e-01 -5.36560893e-01 -4.63969827e-01 -3.91489297e-01
1.28098130e+00 -6.04301393e-01 -1.47899771e+00 5.66311359e-01
-1.82211959e+00 -9.44158256e-01 -8.21809396e-02 -9.60483328e-02
-6.58366919e-01 3.37800407e-03 -8.49515378e-01 -1.12928462e+00
-1.04723644e+00 -9.54644620e-01 7.93759584e-01 6.21110022e-01
-9.64362204e-01 -8.22588265e-01 2.25097120e-01 1.00833631e+00
5.57148933e-01 -2.35821363e-02 7.05240905e-01 -1.34512579e+00
-1.59967110e-01 -1.74016893e-01 2.15159804e-02 3.25692415e-01
8.49446505e-02 -1.14639036e-01 -1.08354282e+00 3.07792891e-02
1.30607948e-01 -8.84593487e-01 3.21179390e-01 5.45792561e-03
3.49356413e-01 -7.23534048e-01 2.41270401e-02 -4.51719135e-01
3.94003481e-01 4.21492368e-01 3.24942589e-01 1.40681446e-01
2.99557269e-01 1.03560042e+00 4.83385980e-01 4.99251455e-01
7.40141094e-01 7.01338232e-01 2.96252906e-01 4.02711779e-01
5.95034510e-02 -9.91051346e-02 5.00838161e-01 6.52935088e-01
1.40795141e-01 -4.32006657e-01 -1.04186499e+00 3.95929903e-01
-2.05737400e+00 -9.88448918e-01 -2.23734841e-01 1.66717923e+00
1.10592723e+00 8.74624029e-02 7.08627820e-01 -3.26441556e-01
8.46039176e-01 5.59587069e-02 -3.27637196e-01 -8.82908642e-01
3.31581652e-01 1.22834705e-01 -1.88887402e-01 7.83892214e-01
-8.43794107e-01 1.32310867e+00 5.62180185e+00 4.74781662e-01
-7.13759005e-01 1.24697715e-01 4.58722889e-01 2.28198513e-01
3.28058809e-01 6.05389550e-02 -8.32613707e-01 4.31826919e-01
1.19877064e+00 -4.76441741e-01 4.21069771e-01 8.12653720e-01
4.62808460e-01 -5.05807638e-01 -9.83948886e-01 2.74956375e-01
-5.95536120e-02 -1.28103054e+00 -3.76587272e-01 -1.45358428e-01
4.27878499e-01 -4.23573479e-02 -4.12647992e-01 6.77499712e-01
1.01838553e+00 -1.07560849e+00 3.48032564e-01 3.29336196e-01
2.93041766e-01 -4.00221944e-01 7.70370722e-01 1.04857540e+00
-5.92063367e-01 1.22741200e-02 -1.72968820e-01 -3.05553347e-01
2.33605757e-01 -7.58346021e-02 -1.85218728e+00 2.17096552e-01
5.20934820e-01 1.48737459e-02 -3.43268335e-01 5.95642865e-01
-4.60416377e-01 8.09066415e-01 -1.69308349e-01 -7.15443015e-01
7.11162090e-02 -3.80718708e-02 8.64063978e-01 1.41738594e+00
-2.17005521e-01 5.24726927e-01 3.31956685e-01 8.40403736e-01
-1.70344546e-01 1.13562182e-01 -3.74459714e-01 -2.13299125e-01
7.29777157e-01 1.64287817e+00 -5.27301252e-01 -4.55867320e-01
1.13659561e-01 1.07334042e+00 3.44729066e-01 4.70414683e-02
-4.24559891e-01 -5.40639460e-01 2.62638986e-01 -2.44651750e-01
-7.00291172e-02 9.89642739e-03 -2.12878734e-01 -7.19263732e-01
-2.89348096e-01 -1.21688139e+00 4.32215899e-01 -8.47029984e-01
-1.18922067e+00 1.25077784e+00 1.68731645e-01 -6.71190917e-01
-9.18677568e-01 -2.53413856e-01 -1.27028251e+00 1.04164135e+00
-5.40932536e-01 -9.12642300e-01 -4.20331568e-01 4.85448867e-01
1.12206066e+00 -3.86437476e-01 1.17516053e+00 -3.87741834e-01
-5.68154514e-01 3.79738301e-01 -5.73785126e-01 3.37870270e-01
7.91433871e-01 -1.54437792e+00 6.97064161e-01 2.49804959e-01
-3.40822190e-01 4.71530408e-01 1.17658353e+00 -6.49439931e-01
-9.50729012e-01 -8.23994815e-01 1.21721101e+00 -7.96903849e-01
6.62223220e-01 -5.72168648e-01 -6.64641082e-01 4.80216414e-01
1.12461126e+00 -9.09107327e-01 3.78754854e-01 -1.59017131e-01
9.89703909e-02 4.69756424e-01 -1.48768568e+00 6.43411279e-01
5.44413149e-01 -3.20221543e-01 -6.83083057e-01 7.54165292e-01
9.34307635e-01 -5.69519401e-01 -6.04103148e-01 -9.92057547e-02
5.21109030e-02 -8.98067176e-01 4.81431246e-01 -5.94666481e-01
7.74539411e-01 1.65269896e-01 3.45284671e-01 -1.34983146e+00
2.00296223e-01 -1.33729315e+00 1.25485793e-01 1.45981669e+00
4.88132298e-01 -4.92738813e-01 6.64252102e-01 9.50270772e-01
-8.78874063e-02 -4.11476254e-01 -7.93597817e-01 -2.95470685e-01
1.89056173e-01 9.15528759e-02 1.50541902e-01 4.57889974e-01
7.06213832e-01 1.20122504e+00 -5.06219208e-01 -2.68537074e-01
2.32729182e-01 -1.96404204e-01 1.16507697e+00 -1.12907457e+00
-4.01135027e-01 -4.16343778e-01 2.96930283e-01 -1.09557354e+00
1.16125114e-01 -5.13073266e-01 6.06920242e-01 -1.83964419e+00
1.59113072e-02 5.10324873e-02 5.41085243e-01 5.74519753e-01
-2.64805347e-01 -1.62891760e-01 4.06223655e-01 1.88137800e-01
-9.01273489e-01 4.66763407e-01 1.19609499e+00 -1.39266904e-02
-4.86117542e-01 5.99098563e-01 -7.86398292e-01 6.26348853e-01
1.04243791e+00 -3.94026488e-01 -3.62916857e-01 8.52218887e-04
-2.00978145e-01 4.88042653e-01 2.19476804e-01 -7.09565520e-01
6.23357713e-01 -8.14531371e-02 -3.53291720e-01 -3.88024777e-01
5.33186615e-01 -2.37618789e-01 -3.68007213e-01 4.70814943e-01
-8.82449627e-01 -7.33792111e-02 8.74921530e-02 1.31351024e-01
1.19080536e-01 -5.99519372e-01 6.95139050e-01 -3.53681862e-01
-1.77712634e-01 -2.81328976e-01 -1.11928761e+00 3.50996763e-01
8.43287230e-01 2.87941724e-01 -7.68680036e-01 -1.21604562e+00
-5.33082962e-01 7.01671779e-01 -3.84459272e-02 4.09802258e-01
2.23810047e-01 -6.92590296e-01 -9.54876900e-01 -5.29883862e-01
-6.41797036e-02 -5.16588241e-02 1.80816367e-01 6.36332750e-01
-5.39724052e-01 5.50238550e-01 -1.83229163e-01 -3.26282591e-01
-1.56172490e+00 -1.56502917e-01 2.96290725e-01 -7.47910857e-01
-5.45023918e-01 1.03950310e+00 -9.72457826e-02 -6.97305083e-01
4.14650291e-01 8.50000605e-02 -7.20244884e-01 -5.02446331e-02
6.85067594e-01 4.82533395e-01 -1.45807147e-01 -3.60503823e-01
-2.32989788e-01 -3.79914165e-01 -2.86098838e-01 -7.40349889e-01
1.12283516e+00 -1.17697172e-01 -1.60420120e-01 3.92892569e-01
6.80182099e-01 -4.53820258e-01 -1.05063534e+00 -2.61561126e-01
2.66909599e-01 8.21841508e-02 -7.17708111e-01 -1.27893639e+00
-1.72092810e-01 6.73475444e-01 2.53768195e-03 8.31237376e-01
5.47592282e-01 1.32917956e-01 8.79417658e-01 9.30677354e-01
1.73941076e-01 -1.18610632e+00 5.42700231e-01 1.34307194e+00
1.35625029e+00 -1.32969058e+00 -4.59957004e-01 -3.37226450e-01
-1.29349625e+00 9.75574791e-01 1.10384047e+00 -1.44167185e-01
2.02872857e-01 6.40640706e-02 4.40978348e-01 -3.80996376e-01
-1.34299374e+00 -9.08020884e-03 -7.43818879e-02 6.24939799e-01
6.36963964e-01 -6.93744570e-02 -3.07309836e-01 6.23752832e-01
-6.91432178e-01 -3.30240637e-01 8.44928682e-01 8.30590606e-01
-5.98769128e-01 -9.92231727e-01 -1.47436425e-01 2.49179348e-01
-2.70325124e-01 1.49187481e-03 -1.25089276e+00 4.45522010e-01
-5.16376615e-01 1.82004130e+00 -2.31781095e-01 -5.49319446e-01
5.75383902e-01 3.68687123e-01 4.15087081e-02 -1.08073390e+00
-1.46930623e+00 -3.48224223e-01 9.12810147e-01 -2.40628108e-01
-1.26723900e-01 -5.69127262e-01 -1.29764485e+00 -3.24855000e-01
-5.55472001e-02 6.08751416e-01 4.66590554e-01 1.22632301e+00
2.88294464e-01 3.37853104e-01 7.65280128e-01 -8.33262444e-01
-8.39270473e-01 -1.92979002e+00 -4.70681712e-02 3.15498054e-01
4.06601131e-02 -1.05419300e-01 -3.34271491e-01 1.05952404e-01] | [12.843008995056152, 8.025388717651367] |
6082d7fa-91bb-4c94-917a-349fcae8e0d5 | an-overview-of-open-ended-evolution-editorial | 1909.04430 | null | https://arxiv.org/abs/1909.04430v1 | https://arxiv.org/pdf/1909.04430v1.pdf | An Overview of Open-Ended Evolution: Editorial Introduction to the Open-Ended Evolution II Special Issue | Nature's spectacular inventiveness, reflected in the enormous diversity of form and function displayed by the biosphere, is a feature of life that distinguishes living most strongly from nonliving. It is, therefore, not surprising that this aspect of life should become a central focus of artificial life. We have known since Darwin that the diversity is produced dynamically, through the process of evolution; this has led life's creative productivity to be called Open-Ended Evolution (OEE) in the field. This article introduces the second of two special issues on current research in OEE and provides an overview of the contents of both special issues. Most of the work was presented at a workshop on open-ended evolution that was held as a part of the 2018 Conference on Artificial Life in Tokyo, and much of it had antecedents in two previous workshops on open-ended evolution at artificial life conferences in Cancun and York. We present a simplified categorization of OEE and summarize progress in the field as represented by the articles in this special issue. | ['Takashi Ikegami', 'Steen Rasmussen', 'Norman Packard', 'Alastair Channon', 'Tim Taylor', 'Mark A. Bedau', 'Kenneth O. Stanley'] | 2019-09-10 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 1.02886453e-01 2.38765806e-01 3.37740660e-01 -6.66249618e-02
6.64199591e-01 -7.71295190e-01 8.16523373e-01 3.24408114e-02
-3.32580179e-01 1.05404103e+00 3.01783681e-01 -9.31730717e-02
-1.63057923e-01 -6.04906499e-01 -5.55207074e-01 -6.19271040e-01
-2.75986016e-01 2.59973288e-01 4.61265706e-02 -8.70281279e-01
3.83463383e-01 3.93470109e-01 -1.96915054e+00 -3.42249542e-01
1.03760779e+00 1.48327753e-01 2.51045197e-01 9.31847632e-01
-8.40277895e-02 3.77282947e-01 -8.99760306e-01 -4.44879234e-01
2.87729800e-02 -1.15605855e+00 -7.15641260e-01 -1.05604097e-01
-4.03289080e-01 3.22955459e-01 -1.58033352e-02 5.44410408e-01
5.97666383e-01 -5.12537286e-02 5.01836896e-01 -8.93685400e-01
-7.49885201e-01 5.62243044e-01 -3.68792154e-02 2.23177090e-01
2.81676710e-01 1.68797925e-01 8.29246283e-01 -5.28953493e-01
8.77337158e-01 9.29730535e-01 8.52177203e-01 6.80369377e-01
-9.63734567e-01 3.54187526e-02 4.36346829e-02 -1.24116495e-01
-9.98657525e-01 -4.71840292e-01 1.60534233e-01 -6.20634019e-01
1.09733498e+00 5.58352947e-01 1.71564579e+00 7.39981413e-01
7.95562029e-01 4.89321381e-01 1.06335509e+00 -4.72571522e-01
4.40283626e-01 1.85122609e-01 -4.13449168e-01 4.08759564e-01
7.42933154e-01 2.60083467e-01 -5.34047663e-01 -1.31752357e-01
5.43472230e-01 -4.77854073e-01 -5.00294507e-01 7.20974132e-02
-1.23612440e+00 5.04881859e-01 2.27849126e-01 7.16574013e-01
-1.84875265e-01 1.83718175e-01 1.04496166e-01 3.79555136e-01
3.27747971e-01 9.37048495e-01 -4.97627437e-01 -6.32448614e-01
-3.28871012e-01 1.73722394e-02 1.29939628e+00 5.61905622e-01
3.36889863e-01 1.08337604e-01 6.00052297e-01 9.34206188e-01
2.77645260e-01 3.03468436e-01 6.19798839e-01 -7.19111681e-01
-7.21582294e-01 5.89329183e-01 -4.62937318e-02 -5.48238933e-01
-4.71855313e-01 -6.25688255e-01 -4.00013238e-01 3.01735908e-01
1.30520180e-01 -3.69106233e-01 -5.34855843e-01 1.78189266e+00
5.55330098e-01 -2.16929302e-01 1.02740079e-01 6.60264611e-01
7.25371003e-01 7.86256433e-01 -2.99120724e-01 -3.87768805e-01
1.17822015e+00 -7.30637252e-01 -6.60415232e-01 1.02496505e-01
3.39845270e-01 -5.82851887e-01 5.17054915e-01 2.37579107e-01
-8.86803091e-01 3.06782941e-03 -1.38017523e+00 3.94634068e-01
-6.56164527e-01 -8.13150048e-01 8.35467100e-01 1.04586136e+00
-1.21620941e+00 7.91000485e-01 -6.72765613e-01 -1.44462073e+00
-1.14533104e-01 2.72979448e-03 -9.03563201e-02 6.06987417e-01
-1.14312482e+00 1.30466652e+00 2.05362409e-01 -5.70799038e-02
-4.41443354e-01 -5.32572985e-01 -2.41147190e-01 -5.05131245e-01
2.21034810e-01 -1.00594032e+00 9.21807349e-01 -9.41787004e-01
-1.69155538e+00 1.29468298e+00 2.40363464e-01 -2.46516660e-01
4.18326229e-01 9.15483013e-02 -4.96738076e-01 -2.17794105e-01
-1.82955608e-01 4.53704953e-01 5.20893037e-02 -1.24293101e+00
-6.10232711e-01 -4.00537044e-01 -3.67461815e-02 1.11968964e-01
-2.16925994e-01 8.84514824e-02 -7.63535127e-02 -6.72029495e-01
-9.48741212e-02 -9.81612444e-01 -1.89512804e-01 2.06604838e-01
6.64312392e-02 -3.39665830e-01 3.26416552e-01 -4.23232585e-01
1.09999561e+00 -2.07618213e+00 4.78487879e-01 -5.08550763e-01
2.79943436e-01 3.40509601e-02 3.15194279e-01 1.12024117e+00
2.69372582e-01 3.87234420e-01 -7.63820231e-01 1.49911776e-01
2.07749978e-02 5.10387897e-01 4.48520444e-02 5.57648659e-01
-1.70440927e-01 8.77894819e-01 -1.25386262e+00 -1.32800043e-01
-2.14633346e-03 4.81763214e-01 -6.68447539e-02 -5.86808398e-02
-2.48400375e-01 5.40284157e-01 6.84896260e-02 6.94767416e-01
2.44781464e-01 -1.70872182e-01 5.35472855e-02 5.27457654e-01
-8.45260203e-01 2.89319474e-02 -3.84481639e-01 1.69010580e+00
-1.43996701e-02 8.81724060e-01 9.33887511e-02 -5.25217175e-01
9.27851677e-01 3.74195367e-01 5.89210808e-01 -2.40194991e-01
4.46685068e-02 6.45653129e-01 3.77991885e-01 -4.93817925e-01
4.14487690e-01 -5.02117872e-01 6.81293681e-02 6.43114507e-01
-2.60924064e-02 -8.14411700e-01 3.09636593e-01 -7.65620098e-02
1.01539385e+00 8.52489889e-01 6.58374667e-01 -8.61276448e-01
3.69861275e-01 4.51569296e-02 7.50538051e-01 5.09677291e-01
-6.20827913e-01 3.79817337e-01 6.08750060e-03 -7.59180546e-01
-1.18050492e+00 -8.93945396e-01 -6.15269840e-01 6.42135620e-01
2.32475445e-01 -6.67744100e-01 -7.91805565e-01 2.15080734e-02
-1.70000970e-01 8.12116504e-01 -1.08463860e+00 -3.73459071e-01
-5.33958375e-01 -1.05538332e+00 6.04012370e-01 -1.17057174e-01
3.55581075e-01 -1.34438968e+00 -1.32000589e+00 2.55037874e-01
2.67558396e-02 -4.05692667e-01 9.73574072e-02 3.73207361e-01
-7.17859328e-01 -6.24853313e-01 -9.33170199e-01 -5.64400315e-01
1.53345197e-01 -1.17701301e-02 1.04666841e+00 3.42912704e-01
-7.27737486e-01 3.33536774e-01 -6.31557345e-01 -8.82958174e-01
-6.78230524e-01 3.79143693e-02 1.95680961e-01 -3.91117036e-01
8.95136520e-02 -9.47584748e-01 -4.89165157e-01 2.52602249e-01
-9.75703716e-01 5.49347289e-02 2.60447741e-01 6.33850992e-01
1.56655341e-01 -2.07835272e-01 8.85208786e-01 -5.44288933e-01
3.29282314e-01 -5.83681464e-01 -2.53871620e-01 5.67716837e-01
-5.17562151e-01 -7.08593130e-02 4.28478539e-01 -1.78473741e-02
-9.64438140e-01 -4.32298928e-01 -1.89802364e-01 7.43225455e-01
1.53046623e-01 5.26698768e-01 9.34118927e-02 -1.39057353e-01
5.85297525e-01 3.35958034e-01 -1.85086019e-02 -2.87494510e-01
1.71033695e-01 8.78399491e-01 3.69779974e-01 -4.76401895e-01
5.88722348e-01 5.44984519e-01 -1.43231107e-02 -1.56222427e+00
-2.80172437e-01 1.66986808e-01 -4.57988530e-01 -6.86046898e-01
8.31993043e-01 -3.65407407e-01 -4.15666848e-01 8.36514294e-01
-8.25694621e-01 -3.84214491e-01 -8.85155380e-01 4.15152699e-01
-5.27108431e-01 9.61714238e-02 -1.85898885e-01 -1.00013888e+00
-2.89271742e-01 -5.26989102e-01 5.93668699e-01 6.71323478e-01
-5.08478999e-01 -1.36809719e+00 8.66769791e-01 3.45864445e-02
7.85202920e-01 8.28560591e-01 7.40713775e-01 -1.72405675e-01
-1.47611022e-01 3.99944261e-02 6.30815864e-01 -4.22090851e-02
4.18991089e-01 5.93388736e-01 -8.15688789e-01 -6.96328953e-02
2.63352066e-01 -1.29911765e-01 6.96779728e-01 8.13583732e-02
5.34902923e-02 2.06651956e-01 -4.53559071e-01 7.72399902e-01
1.42114377e+00 5.38905799e-01 5.94827771e-01 8.77029955e-01
-3.45074125e-02 1.09587967e+00 3.75155240e-01 6.26939595e-01
2.00183213e-01 3.76357913e-01 3.45171183e-01 1.79063559e-01
-1.10572144e-01 4.68938462e-02 3.09703201e-01 1.19385684e+00
-3.25388551e-01 -5.61192989e-01 -8.97839129e-01 5.81561983e-01
-1.62463558e+00 -9.22428548e-01 -2.17559978e-01 2.17249250e+00
7.09846556e-01 -2.58320153e-01 3.30928683e-01 -2.93999344e-01
8.12790990e-01 -6.14914037e-02 -8.62264872e-01 -8.04533124e-01
-8.32371950e-01 9.17246714e-02 7.76787922e-02 3.69238675e-01
-4.96248871e-01 7.83278525e-01 7.99906635e+00 3.20494696e-02
-1.14674139e+00 -1.55749202e-01 1.87032714e-01 -1.16464272e-01
-6.90708697e-01 2.31355608e-01 -3.95334035e-01 4.95362788e-01
1.06109858e+00 -7.81669378e-01 4.48945463e-01 3.40046048e-01
8.90328512e-02 -3.44045013e-01 -7.29344964e-01 3.92587900e-01
2.27028638e-01 -1.02298188e+00 -3.47789377e-01 2.30330706e-01
8.05703521e-01 2.89724767e-01 -4.05349553e-01 3.03585548e-02
3.21344912e-01 -1.05478036e+00 8.33320320e-01 5.32879651e-01
4.97883320e-01 -5.97456276e-01 2.81093806e-01 3.25156868e-01
-8.44444335e-01 -1.29592657e-01 -4.36523259e-01 -4.71237212e-01
7.03147233e-01 6.93154693e-01 -3.33376080e-01 4.93928462e-01
9.36979592e-01 7.53043175e-01 -4.13070172e-01 1.47628403e+00
-2.67287880e-01 4.49214280e-01 -4.76297617e-01 -6.18076563e-01
-1.73078358e-01 -4.92637694e-01 1.11644828e+00 9.86202836e-01
3.60007852e-01 2.80772299e-01 -6.43040836e-01 8.15482855e-01
4.22119439e-01 4.42093723e-02 -4.80654240e-01 -7.96189964e-01
3.70415270e-01 1.04093444e+00 -9.26807642e-01 -2.13543132e-01
-8.84436220e-02 9.55403924e-01 -1.47764057e-01 -1.44935369e-01
-7.45171309e-01 -5.52953899e-01 7.28537083e-01 -5.51344492e-02
6.21703230e-02 -1.95204392e-01 -3.69854450e-01 -1.02629888e+00
-2.91732937e-01 -6.53324842e-01 1.10308342e-01 -6.89076900e-01
-1.07817876e+00 7.91869342e-01 -2.06438988e-01 -7.50469327e-01
-2.09273845e-02 -2.81650424e-01 -4.87449616e-01 5.13219059e-01
-6.36622131e-01 -8.10380042e-01 -3.66226196e-01 -3.67700338e-01
3.23692977e-01 2.45742872e-02 9.33254361e-01 -2.07418382e-01
-6.52008057e-01 2.42058218e-01 6.29084408e-01 -7.81412601e-01
5.86119592e-01 -1.17370081e+00 8.32209289e-01 5.65513730e-01
-2.15044126e-01 6.53794944e-01 1.11433983e+00 -8.57589066e-01
-1.46189487e+00 -3.82981449e-01 1.04380596e+00 -3.88836741e-01
5.73461831e-01 -3.89992982e-01 -4.00690198e-01 5.60007811e-01
7.65583873e-01 -7.59289563e-01 9.23314393e-01 -3.33177477e-01
2.21157014e-01 3.14042658e-01 -1.22092378e+00 8.14872086e-01
1.44461906e+00 1.16898119e-01 -7.78475285e-01 2.05888987e-01
1.09323108e+00 6.80376589e-02 -9.58444476e-01 3.86315227e-01
1.17310584e+00 -1.11786366e+00 5.27238369e-01 -3.22400928e-01
3.48890007e-01 -5.28060794e-01 -5.27236499e-02 -1.29898024e+00
-5.12080014e-01 -9.70261514e-01 2.82053411e-01 1.34735692e+00
2.85778493e-01 -1.26471555e+00 2.00860977e-01 1.81203708e-01
-4.95200753e-01 -8.48495901e-01 -8.79290998e-01 -8.95181179e-01
3.50935936e-01 4.70630765e-01 5.92870712e-01 1.00043237e+00
2.39041463e-01 2.83668250e-01 -8.46890733e-02 -6.13017738e-01
6.66273236e-01 -2.22834498e-02 8.21859300e-01 -1.50222075e+00
-5.69553256e-01 -6.57083571e-01 -7.37947583e-01 -4.14328426e-01
-5.39932013e-01 -7.20967412e-01 1.75822914e-01 -1.84109318e+00
1.01077572e-01 1.01961405e-03 2.62183305e-02 -6.59379736e-02
-4.87308763e-02 1.78525686e-01 -1.39333569e-02 4.60124969e-01
-9.90664288e-02 7.06792951e-01 1.15699708e+00 4.54003990e-01
-2.91137755e-01 -2.78460979e-01 -1.01576543e+00 3.99232388e-01
7.98507810e-01 -2.20361710e-01 -1.53109565e-01 -2.31733382e-01
6.68464720e-01 -2.69757271e-01 -2.18687296e-01 -1.16839218e+00
2.85874092e-04 -3.89261186e-01 1.15223601e-01 -1.81865305e-01
4.07550126e-01 -3.97223294e-01 1.02188849e+00 1.00658095e+00
1.11120299e-01 -4.33142148e-02 2.41808355e-01 2.93530107e-01
2.77258366e-01 -3.57500643e-01 9.48948443e-01 -4.05953825e-01
-5.98719656e-01 -1.93812102e-01 -1.00024140e+00 -7.93325007e-02
1.41553545e+00 -7.05032229e-01 -5.82070470e-01 -1.16037048e-01
-6.64650738e-01 2.11098596e-01 1.35723484e+00 3.12828898e-01
2.54885107e-01 -7.68927217e-01 -9.60590065e-01 -2.59158671e-01
1.81371167e-01 -4.81520057e-01 1.33260474e-01 9.41132784e-01
-1.19343317e+00 2.62746364e-01 -5.08463442e-01 -5.03858507e-01
-8.76598775e-01 4.98254299e-01 5.86178005e-01 3.51251513e-01
-7.32783318e-01 1.12608850e+00 2.82757908e-01 -6.61283612e-01
-3.10269445e-01 5.14859855e-01 -1.08673356e-01 -2.48418286e-01
6.13802433e-01 5.37661254e-01 -4.80896115e-01 -6.90728426e-01
-6.02094531e-01 7.48699963e-01 3.93639207e-01 -4.76868898e-01
1.82061493e+00 -3.54062408e-01 -1.03442419e+00 1.23552132e+00
6.97732449e-01 -2.27577034e-02 -9.02329087e-01 3.95479828e-01
-8.88755471e-02 -2.92985290e-01 -5.05641878e-01 -1.19037116e+00
-2.89532751e-01 5.08974791e-01 2.83512831e-01 2.89748549e-01
1.07586074e+00 5.77479489e-02 6.27342165e-01 2.43138358e-01
5.92713833e-01 -1.07362914e+00 -2.26260334e-01 5.28385401e-01
1.18121457e+00 -3.70407254e-01 1.46616265e-01 -3.92319597e-02
-2.15423167e-01 1.12113631e+00 5.23648262e-01 5.18354066e-02
7.30750799e-01 3.95387918e-01 -8.60582292e-03 -1.80052102e-01
-1.09584069e+00 -5.09526618e-02 -3.97585839e-01 6.90682232e-01
8.61639977e-01 -1.01239629e-01 -1.11175525e+00 1.55220792e-01
-5.37861764e-01 -1.98472645e-02 7.45986044e-01 1.49312770e+00
-1.14847684e+00 -1.22218871e+00 -3.80874842e-01 3.56929563e-02
-1.37024060e-01 1.08993568e-01 -1.26891947e+00 5.39502263e-01
3.58276010e-01 9.42477703e-01 -1.37912080e-01 -4.57645237e-01
-5.60419448e-02 -5.14066704e-02 8.21331203e-01 -3.03094000e-01
-1.03855586e+00 -3.89855057e-01 1.52066529e-01 -1.95783705e-01
-4.02227640e-01 -7.76286364e-01 -1.25440407e+00 -6.74002826e-01
-3.28845292e-01 7.91044891e-01 1.23031378e+00 7.72362292e-01
2.64758438e-01 5.17869234e-01 3.61794025e-01 -6.67815745e-01
-1.90558121e-01 -8.72942269e-01 -7.85012603e-01 -7.41169453e-02
8.12687501e-02 -4.43540096e-01 -6.69346094e-01 3.09776254e-02] | [5.570057392120361, 4.201204299926758] |
8bf9cb97-1113-4ee4-8c95-cd137a313b3e | data-mining-textual-responses-to-uncover | 1703.08544 | null | http://arxiv.org/abs/1703.08544v2 | http://arxiv.org/pdf/1703.08544v2.pdf | Data-Mining Textual Responses to Uncover Misconception Patterns | An important, yet largely unstudied, problem in student data analysis is to
detect misconceptions from students' responses to open-response questions.
Misconception detection enables instructors to deliver more targeted feedback
on the misconceptions exhibited by many students in their class, thus improving
the quality of instruction. In this paper, we propose a new natural language
processing-based framework to detect the common misconceptions among students'
textual responses to short-answer questions. We propose a probabilistic model
for students' textual responses involving misconceptions and experimentally
validate it on a real-world student-response dataset. Experimental results show
that our proposed framework excels at classifying whether a response exhibits
one or more misconceptions. More importantly, it can also automatically detect
the common misconceptions exhibited across responses from multiple students to
multiple questions; this property is especially important at large scale, since
instructors will no longer need to manually specify all possible misconceptions
that students might exhibit. | ['Andrew S. Lan', 'Joshua J. Michalenko', 'Richard G. Baraniuk'] | 2017-03-24 | null | null | null | null | ['misconceptions'] | ['miscellaneous'] | [ 7.95198604e-02 -8.83188546e-02 -2.53141411e-02 -4.83538508e-01
-6.73220754e-01 -8.27294230e-01 1.39828131e-01 1.02613580e+00
6.73396736e-02 4.58962947e-01 6.91404119e-02 -1.06294513e+00
-4.04215485e-01 -8.12591970e-01 -6.96803987e-01 -1.32390574e-01
6.55519009e-01 -1.41741931e-01 5.02253115e-01 -3.20546180e-01
1.02192545e+00 3.51303309e-01 -1.88553321e+00 5.49319685e-01
1.30259621e+00 3.42853457e-01 6.26359284e-02 8.10257912e-01
-3.65526527e-01 1.63195682e+00 -8.06443810e-01 -5.28103173e-01
-6.06698990e-01 -4.00388032e-01 -1.24176908e+00 -6.36107549e-02
9.40707922e-01 4.23484445e-02 6.48823380e-02 1.56989145e+00
-1.31824046e-01 5.93274176e-01 4.91702348e-01 -1.13076377e+00
-8.17193329e-01 6.61546290e-01 -4.18086797e-01 2.40223631e-01
9.38756287e-01 -3.83212268e-01 6.35695517e-01 -6.34186804e-01
1.19313434e-01 1.33621049e+00 4.43393081e-01 5.63207984e-01
-1.01117957e+00 -6.07425213e-01 -3.52134183e-02 4.49172407e-01
-9.95613813e-01 3.56102176e-02 7.36553371e-01 -7.86746264e-01
-1.41045935e-02 7.80739665e-01 8.20966884e-02 7.71611035e-01
9.97897983e-01 7.37830162e-01 1.24379909e+00 -5.76384366e-01
2.22730011e-01 9.50244129e-01 9.46006715e-01 6.48783028e-01
4.49945211e-01 -3.68684798e-01 -3.90146583e-01 -4.22806293e-01
1.42586589e-01 3.17240208e-01 -6.99272573e-01 -1.64610460e-01
-4.40639317e-01 9.31077302e-01 1.85813699e-02 3.63008261e-01
5.25315106e-02 -5.30645549e-01 2.28969336e-01 6.93269610e-01
-4.26857769e-02 7.92033374e-01 -6.47901952e-01 -4.54529375e-01
-5.35624981e-01 3.93164247e-01 1.17695928e+00 8.94037187e-01
7.84905136e-01 -5.39348602e-01 7.76220784e-02 9.93858159e-01
4.22342628e-01 1.99955598e-01 8.85340691e-01 -8.15811694e-01
-3.91512811e-02 1.10563731e+00 1.87292278e-01 -1.67036366e+00
2.31351275e-02 -1.88138336e-01 -3.20442080e-01 -2.54925102e-01
3.30387503e-01 -8.81314650e-02 -2.14735687e-01 1.26712036e+00
2.24071965e-01 -4.30517755e-02 6.02091849e-02 5.76169729e-01
1.23242569e+00 4.75302488e-01 3.80850658e-02 5.93033480e-03
1.76017439e+00 -6.32897079e-01 -8.68929207e-01 -1.39186457e-01
1.18158937e+00 -1.46419680e+00 1.09597373e+00 7.28515327e-01
-6.94519460e-01 -6.83040142e-01 -5.87239981e-01 1.26242042e-02
-1.69729516e-01 1.09495841e-01 2.69087791e-01 8.40660870e-01
-4.40474808e-01 5.70243120e-01 -1.39628828e-01 2.67268787e-03
-1.33420244e-01 -5.53114899e-02 -1.22259915e-01 -6.26003444e-01
-1.24792647e+00 8.38024437e-01 1.17493808e-01 -2.41364375e-01
-4.17153776e-01 -1.04198492e+00 -1.06967962e+00 3.80683690e-01
6.31319284e-01 -6.88169524e-02 1.70073009e+00 -6.56517088e-01
-1.47285771e+00 6.90129876e-01 -3.16449136e-01 1.83257610e-01
2.43429944e-01 -3.19534659e-01 -5.92356622e-01 -8.34584460e-02
9.65269879e-02 -8.58179256e-02 3.32704723e-01 -9.75278378e-01
-3.92620474e-01 -2.70678222e-01 -5.04453443e-02 -3.57173800e-01
-5.08611500e-01 1.14738517e-01 3.41136962e-01 -3.49541605e-01
4.79237407e-01 -9.05282795e-01 -1.01839729e-01 -4.40318376e-01
-1.49724334e-01 -8.61404717e-01 8.26720655e-01 -8.78385678e-02
1.37388468e+00 -2.01325750e+00 -7.47946560e-01 2.10942447e-01
4.08674449e-01 4.95867997e-01 9.65082124e-02 6.61684811e-01
-3.19279432e-01 2.51137465e-01 2.90323555e-01 5.79507649e-01
7.07624927e-02 -1.35574918e-02 -7.25521624e-01 1.99233040e-01
6.53218664e-03 3.75430733e-01 -1.42484105e+00 -4.93448555e-01
2.61487186e-01 -1.18073754e-01 -4.20129031e-01 6.11779094e-01
-4.53999452e-02 -4.21097651e-02 -8.37207615e-01 2.36655638e-01
9.83213544e-01 -1.66395426e-01 -3.37024406e-02 5.95500171e-01
-2.69356012e-01 9.30865169e-01 -9.53281283e-01 1.11802566e+00
-5.30312419e-01 8.79187882e-01 -3.17540824e-01 -8.22154939e-01
1.29450595e+00 3.85291159e-01 2.69284472e-02 -1.88089654e-01
3.23147446e-01 1.78533241e-01 6.74094483e-02 -1.01894534e+00
7.79740870e-01 -4.12683547e-01 -1.72380388e-01 7.60782838e-01
4.39084461e-03 -4.37411994e-01 1.97905689e-01 5.33883870e-01
1.11106431e+00 -5.63386559e-01 2.31898651e-01 -6.78235710e-01
9.43820536e-01 3.32648866e-02 4.70483720e-01 9.79837358e-01
-5.65693021e-01 2.32432321e-01 7.03498363e-01 -4.69510496e-01
-1.28980838e-02 -7.70260572e-01 -1.30602866e-01 1.12176704e+00
3.44352983e-02 -5.82208574e-01 -4.96597171e-01 -8.01917851e-01
-2.07495987e-01 1.14329159e+00 -4.76897240e-01 -5.08173287e-01
1.03480369e-01 1.70157775e-01 2.37893194e-01 8.72063041e-02
2.04722226e-01 -7.55547583e-01 -6.05596185e-01 1.34050950e-01
-4.73598719e-01 -9.72995937e-01 -4.90363032e-01 -1.10493563e-02
-8.00234139e-01 -1.19876587e+00 -1.76610395e-01 -9.47128952e-01
9.23093021e-01 8.54080081e-01 1.14311564e+00 5.27314067e-01
1.26831727e-02 5.23277402e-01 -5.18989444e-01 -5.48678577e-01
-9.49673295e-01 -4.28903759e-01 -8.51933733e-02 -2.60386199e-01
1.20280993e+00 -2.33926654e-01 -2.46110976e-01 4.85872597e-01
-9.82269347e-01 -2.95183569e-01 3.73505563e-01 3.02278340e-01
-1.34661794e-01 3.50058645e-01 5.28590143e-01 -1.28322184e+00
1.18427622e+00 -5.33404291e-01 -5.45949280e-01 4.53180850e-01
-5.44448733e-01 1.71648353e-01 8.05551946e-01 -6.33873761e-01
-1.09646690e+00 -4.17625636e-01 -2.11916000e-01 -1.49116695e-01
-9.19661224e-01 7.03991771e-01 2.12113053e-01 -3.25411379e-01
9.70035732e-01 3.10604751e-01 -1.91766098e-01 -2.56118715e-01
-1.59608081e-01 8.77516866e-01 3.58724147e-01 -9.72214580e-01
5.07947505e-01 -3.04873347e-01 -1.90010965e-01 -9.36380327e-01
-1.22362185e+00 -9.03877437e-01 -3.16597015e-01 -4.12689000e-01
2.30581999e-01 -6.92798018e-01 -1.32140374e+00 3.84955877e-03
-1.32902861e+00 2.89276093e-01 3.86292547e-01 6.13191605e-01
-1.37593105e-01 7.21980095e-01 -6.85106874e-01 -7.79689133e-01
3.29479605e-01 -1.18357539e+00 2.83005964e-02 7.41506279e-01
-8.66391420e-01 -1.19546282e+00 4.28559542e-01 9.58869755e-01
1.69948205e-01 -2.13297993e-01 1.14692163e+00 -1.00902832e+00
-9.49241593e-03 -4.30168599e-01 2.87459586e-02 3.19367170e-01
2.10665390e-01 3.00539374e-01 -8.28303516e-01 1.90815665e-02
5.64686656e-01 -6.91365898e-01 4.76383299e-01 -1.95018560e-01
1.35590553e+00 -5.85177958e-01 8.62370953e-02 -4.64365900e-01
1.06439984e+00 -9.22393948e-02 4.20568258e-01 6.35947138e-02
3.91193867e-01 1.18291926e+00 7.59519696e-01 1.74020156e-01
4.28544581e-01 -5.96313551e-02 1.83649421e-01 7.76852548e-01
4.77044940e-01 -3.46015722e-01 5.97886741e-01 1.14392996e+00
7.60397732e-01 -8.71172473e-02 -9.34847951e-01 7.17322648e-01
-1.52346098e+00 -8.67421448e-01 -1.09740806e+00 2.23421574e+00
1.12291968e+00 -3.08491588e-02 -5.52411854e-01 1.29891574e-01
7.31989324e-01 -2.19530463e-01 6.43813461e-02 -1.34546983e+00
8.12183738e-01 1.29977331e-01 -1.66727200e-01 5.19697666e-01
-5.73160291e-01 6.28667414e-01 5.06317043e+00 7.12086678e-01
-8.61179411e-01 -4.07772303e-01 2.65691131e-01 6.83546126e-01
-8.02703440e-01 4.58014868e-02 -7.50446439e-01 1.73438057e-01
1.13811827e+00 -4.17988390e-01 -3.40228677e-01 1.14965534e+00
3.24498057e-01 -4.09454703e-01 -1.12128365e+00 3.54235828e-01
1.55622318e-01 -9.30171788e-01 7.40771517e-02 -2.51639724e-01
9.34504867e-01 -8.13323140e-01 -8.65894258e-02 4.88516897e-01
6.11163676e-01 -7.71079183e-01 9.69801247e-02 1.37929514e-01
-3.01243931e-01 -9.45105255e-01 7.89715588e-01 8.42094600e-01
-5.37544608e-01 8.32272694e-02 -7.70701468e-01 -5.86888731e-01
-8.81259441e-01 6.79716110e-01 -1.27415442e+00 2.05911309e-01
4.09036696e-01 3.80589336e-01 -7.79798210e-01 1.10845470e+00
-6.98940992e-01 8.47220600e-01 3.30210775e-01 -6.29042327e-01
1.41775951e-01 5.89911863e-02 5.79758659e-02 9.91833627e-01
2.66149998e-01 5.10130286e-01 1.28648147e-01 1.18827271e+00
-6.18649162e-02 3.71040910e-01 -6.99696779e-01 -1.43502027e-01
5.71146965e-01 1.17902291e+00 -3.49285364e-01 -2.30974436e-01
-2.37038061e-01 5.07022619e-01 1.90418676e-01 1.20742947e-01
-4.15452272e-01 -7.24534512e-01 4.19141144e-01 4.92143407e-02
-2.09893316e-01 3.29272538e-01 -1.42414227e-01 -1.22339571e+00
-2.47522548e-01 -1.13710093e+00 3.18586409e-01 -7.30331659e-01
-1.41777849e+00 1.95477143e-01 -4.05650079e-01 -1.16723680e+00
8.35340396e-02 -7.64528811e-01 -1.13622236e+00 8.57263029e-01
-1.44873488e+00 -9.95123163e-02 -3.12805235e-01 4.10409719e-01
7.17082858e-01 4.13202584e-01 9.01972771e-01 7.01417774e-02
-6.23042941e-01 4.72744673e-01 -3.93921047e-01 1.67019129e-01
1.32690489e+00 -1.21249080e+00 -3.19580287e-01 8.34916651e-01
-5.09538986e-02 1.38794065e+00 1.13329852e+00 -4.82749820e-01
-1.31521797e+00 -8.71473730e-01 1.46889710e+00 -5.37923932e-01
7.74613857e-01 3.86169136e-01 -1.64631796e+00 5.91885984e-01
3.49994719e-01 -3.02776903e-01 1.66596293e+00 2.59581149e-01
-4.30478752e-01 4.44701642e-01 -7.68518388e-01 6.68694913e-01
-1.70478165e-01 -5.12549579e-01 -1.40376389e+00 4.03142601e-01
5.08667111e-01 -5.66776454e-01 -8.43073547e-01 -3.18305232e-02
3.97172511e-01 -1.05404663e+00 3.42705965e-01 -1.05460012e+00
1.09359598e+00 -6.11437410e-02 1.81867242e-01 -1.30292892e+00
-1.89484715e-01 -6.63012981e-01 3.73160899e-01 1.26996219e+00
2.37460628e-01 -3.70574355e-01 6.29371524e-01 1.34384942e+00
-2.05869019e-01 -5.05438209e-01 -2.95530319e-01 -4.57986295e-01
3.84166509e-01 -5.13178885e-01 2.26952374e-01 1.58479941e+00
7.63208747e-01 3.27217430e-01 -2.06067488e-02 3.95461142e-01
5.18712044e-01 6.75735176e-01 8.71539891e-01 -1.32807410e+00
9.32743549e-02 -4.52232033e-01 -1.95929945e-01 -1.20710194e+00
3.57888490e-01 -5.29489934e-01 4.50843394e-01 -8.14037621e-01
3.59345675e-01 -1.46040916e-01 -3.00542206e-01 1.99737251e-01
-7.51754642e-01 -3.61250609e-01 -2.14626163e-01 -2.92093307e-01
-6.74877584e-01 3.39356512e-01 1.33605003e+00 1.18977651e-01
-9.48557630e-02 4.99056846e-01 -1.22224498e+00 1.26872277e+00
6.71131134e-01 -9.13484752e-01 -3.23741347e-01 -1.83063969e-01
3.97705823e-01 1.72490001e-01 2.64285833e-01 -9.00709569e-01
6.27942622e-01 -8.89456093e-01 -7.04161897e-02 -4.74949569e-01
-5.47541022e-01 -7.91126609e-01 -7.52606809e-01 4.63108838e-01
-7.17097878e-01 6.75331578e-02 5.08751035e-01 4.38085169e-01
-3.94479990e-01 -1.08085144e+00 7.01142251e-01 -1.58816114e-01
-4.49611455e-01 -3.98865640e-01 -1.09182632e+00 4.81185280e-02
8.62930954e-01 2.06224784e-01 -4.93432015e-01 -4.18203205e-01
-1.80310339e-01 3.95745188e-01 1.48375615e-01 6.54309213e-01
1.04408348e+00 -1.03528738e+00 -6.53429866e-01 1.84406593e-01
4.24051285e-01 -2.46232525e-01 5.07607460e-01 5.35085320e-01
-4.91748393e-01 6.48206592e-01 -1.01433523e-01 -4.57371771e-01
-1.67904115e+00 4.65992510e-01 -4.23069783e-02 -2.42860019e-01
5.71338944e-02 1.04017019e+00 -3.60658653e-02 -8.88153374e-01
4.36486676e-02 -7.46735871e-01 -4.64292884e-01 -2.45325863e-01
1.00677168e+00 4.87934142e-01 3.11274856e-01 -2.12404355e-01
-1.90731555e-01 2.43462533e-01 -5.48828125e-01 5.07868528e-01
7.15214074e-01 -7.80200884e-02 -2.90601969e-01 7.45599985e-01
1.06582808e+00 5.03339887e-01 -2.40499914e-01 -3.29298735e-01
1.83523670e-01 -7.82697260e-01 3.24314535e-02 -7.96882212e-01
-3.05124730e-01 1.28694594e+00 -8.21688548e-02 5.30694246e-01
6.60874665e-01 -3.81388247e-01 6.65075660e-01 7.87044764e-01
5.03721423e-02 -5.78455269e-01 2.36949340e-01 8.19303513e-01
5.65479100e-01 -1.44429994e+00 -2.93037359e-04 -7.31814742e-01
-3.72661769e-01 1.59576344e+00 1.06901860e+00 5.88292256e-02
4.06531215e-01 -2.73798794e-01 3.89468342e-01 -2.87863761e-01
-1.25460815e+00 6.29734516e-01 5.38726449e-01 -1.10882141e-01
1.16517806e+00 1.66446060e-01 -4.97780889e-01 6.72003269e-01
-3.35262984e-01 -4.85094786e-02 1.49999762e+00 1.15653241e+00
-1.03460503e+00 -1.11948907e+00 -8.92827332e-01 2.58149147e-01
-6.40449703e-01 -9.17369649e-02 -8.36046755e-01 4.36275676e-02
-2.58612871e-01 1.68797016e+00 -3.33771914e-01 -4.94462132e-01
2.87760824e-01 2.58389562e-01 2.64705300e-01 -1.14318347e+00
-7.40824401e-01 -5.68382204e-01 -4.56826895e-01 -1.31694481e-01
-2.13770326e-02 -3.23802173e-01 -1.13975179e+00 -8.07599783e-01
-8.56544733e-01 9.77162421e-01 3.75791162e-01 1.23537624e+00
3.32906038e-01 5.08587182e-01 7.36038446e-01 4.18360442e-01
-1.05954897e+00 -9.51842725e-01 -2.05208942e-01 4.37885344e-01
5.08039951e-01 -3.55685413e-01 -4.47492242e-01 -6.24849461e-02] | [10.161417007446289, 7.432184219360352] |
0b3b6241-ddfd-4ffa-8531-30d50bf24f47 | cross-lingual-transfer-can-worsen-bias-in | 2305.12709 | null | https://arxiv.org/abs/2305.12709v1 | https://arxiv.org/pdf/2305.12709v1.pdf | Cross-lingual Transfer Can Worsen Bias in Sentiment Analysis | Sentiment analysis (SA) systems are widely deployed in many of the world's languages, and there is well-documented evidence of demographic bias in these systems. In languages beyond English, scarcer training data is often supplemented with transfer learning using pre-trained models, including multilingual models trained on other languages. In some cases, even supervision data comes from other languages. Does cross-lingual transfer also import new biases? To answer this question, we use counterfactual evaluation to test whether gender or racial biases are imported when using cross-lingual transfer, compared to a monolingual transfer setting. Across five languages, we find that systems using cross-lingual transfer usually become more biased than their monolingual counterparts. We also find racial biases to be much more prevalent than gender biases. To spur further research on this topic, we release the sentiment models we used for this study, and the intermediate checkpoints throughout training, yielding 1,525 distinct models; we also release our evaluation code. | ['Adam Lopez', 'Björn Ross', 'Seraphina Goldfarb-Tarrant'] | 2023-05-22 | null | null | null | null | ['cross-lingual-transfer'] | ['natural-language-processing'] | [-3.38240862e-01 1.48708057e-02 -7.04327881e-01 -8.51501465e-01
-8.28190506e-01 -7.48995364e-01 8.78791988e-01 3.13902497e-01
-7.52306521e-01 1.15414548e+00 6.12484217e-01 -6.91571832e-01
5.81146955e-01 -6.89658582e-01 -7.30511308e-01 -2.35660255e-01
2.94980347e-01 5.33768296e-01 -3.99315864e-01 -6.76136196e-01
1.14693351e-01 -1.10392369e-01 -1.12293911e+00 1.44558147e-01
1.16715837e+00 1.39959708e-01 -3.76458794e-01 1.75177619e-01
-1.25596285e-01 6.39768004e-01 -5.27726948e-01 -1.03416634e+00
2.64920890e-01 -2.63731241e-01 -8.62598538e-01 -2.25890771e-01
8.21695924e-01 2.58487668e-02 2.73327619e-01 1.02364731e+00
4.73994672e-01 -1.72938734e-01 7.77441025e-01 -1.17327690e+00
-9.44597781e-01 8.59659314e-01 -9.97620165e-01 1.11736961e-01
2.64320642e-01 -9.42839403e-03 1.11447012e+00 -9.45556939e-01
8.14719200e-01 1.62254298e+00 9.49372530e-01 4.43522811e-01
-1.44238949e+00 -1.12655783e+00 3.04803550e-01 -3.22270900e-01
-1.08677578e+00 -5.77423453e-01 5.88440716e-01 -6.26018763e-01
6.07865334e-01 -1.35558084e-01 5.71488857e-01 1.34418285e+00
3.00197929e-01 5.08329630e-01 2.00170970e+00 -5.58507621e-01
-2.84928292e-01 9.76017177e-01 1.12246729e-01 4.11844343e-01
6.08317971e-01 -1.84035987e-01 -8.65655065e-01 -2.35842824e-01
1.98993877e-01 -2.24396244e-01 1.94882408e-01 -1.85940787e-01
-1.05363226e+00 1.11787784e+00 2.15150923e-01 3.99535090e-01
1.05450265e-01 -2.70313919e-01 6.31045580e-01 7.08196223e-01
1.12684989e+00 5.89097202e-01 -8.14088643e-01 -2.53943801e-01
-8.84327710e-01 1.49940133e-01 8.82828116e-01 7.84253240e-01
1.03314865e+00 -5.62852249e-02 3.53623807e-01 1.14909315e+00
1.04170807e-01 7.49071300e-01 5.06821454e-01 -6.77958131e-01
5.57449639e-01 5.41516125e-01 -4.62444611e-02 -6.52469873e-01
-2.05892846e-01 -2.90487379e-01 -4.47429746e-01 -8.25544670e-02
6.20975494e-01 -6.03581667e-01 -6.90188289e-01 2.09781647e+00
8.75597224e-02 -5.50934970e-01 1.53595299e-01 5.01020610e-01
2.86519527e-01 3.86942953e-01 4.42942023e-01 4.87420484e-02
1.45912802e+00 -6.67579114e-01 -4.36703295e-01 -6.82491422e-01
1.03947842e+00 -1.07473934e+00 1.37646699e+00 1.27264351e-01
-7.12729871e-01 -3.31680000e-01 -8.93075645e-01 -2.62818843e-01
-7.70592868e-01 2.63941810e-02 9.28843439e-01 1.20988226e+00
-1.08475184e+00 2.10925683e-01 -5.66194713e-01 -9.55458999e-01
3.53035122e-01 1.21544957e-01 -5.00250995e-01 -2.52545118e-01
-1.31268609e+00 1.18904805e+00 -1.20564237e-01 -3.09613258e-01
-4.56001729e-01 -7.89769828e-01 -1.03636611e+00 -5.46365023e-01
-3.14085409e-02 -3.48610729e-01 1.09925318e+00 -1.65523148e+00
-1.02535450e+00 1.49201584e+00 -3.97961199e-01 4.72045429e-02
4.30514812e-01 -3.37627977e-01 -4.69849855e-01 -5.44967830e-01
8.39928150e-01 6.66399360e-01 4.18665439e-01 -1.25816190e+00
-6.99871123e-01 -6.17769420e-01 1.31369293e-01 3.42865467e-01
-6.90323174e-01 6.00141764e-01 7.07197785e-02 -6.78839207e-01
-5.10566533e-01 -1.17336702e+00 3.84603105e-02 -5.67197502e-01
-6.80716559e-02 -4.07590479e-01 3.28153491e-01 -9.65434134e-01
9.14558113e-01 -1.97008789e+00 -2.53059566e-01 1.24338113e-01
-1.31211251e-01 -9.07391831e-02 2.40020473e-02 4.64749545e-01
-1.42311305e-01 4.95907992e-01 -9.48271677e-02 -6.13355041e-01
-6.59395084e-02 1.03223331e-01 -3.79897982e-01 4.46433306e-01
2.99909055e-01 4.49405998e-01 -7.50575304e-01 -6.11960351e-01
-2.68605649e-01 2.61080265e-01 -7.78830349e-01 -3.76713634e-01
3.00462961e-01 4.64855313e-01 -4.22419235e-02 8.35227907e-01
5.81230342e-01 7.94417858e-02 4.72586691e-01 1.93753704e-01
-3.79667759e-01 5.90129137e-01 -3.64415377e-01 1.46820354e+00
-7.19918609e-01 8.31347764e-01 1.64324254e-01 -7.00381815e-01
8.87777507e-01 4.86257672e-02 1.54405264e-02 -6.43281221e-01
1.31357998e-01 5.89929104e-01 2.91069806e-01 -6.91971034e-02
7.82199383e-01 -6.69489205e-01 -4.86836880e-01 6.86103880e-01
4.95887399e-02 -3.43093276e-01 2.67143011e-01 2.32546777e-01
3.90432149e-01 9.62043703e-02 6.21484146e-02 -7.44453907e-01
3.09271485e-01 2.75966644e-01 7.89437771e-01 4.30620551e-01
-4.20238465e-01 2.26137146e-01 6.88537776e-01 -3.83876115e-01
-1.09674180e+00 -7.82852829e-01 -5.73002636e-01 1.60543072e+00
-3.05052280e-01 -3.88759792e-01 -5.85276961e-01 -8.67667496e-01
2.65586317e-01 7.05749035e-01 -8.79700780e-01 -2.02597920e-02
-3.41367394e-01 -1.16089439e+00 6.57718182e-01 3.56446981e-01
2.95509994e-01 -7.25241482e-01 6.26801923e-02 -3.30210775e-01
-1.48193017e-01 -9.71204579e-01 -3.80114496e-01 1.47278309e-01
-8.59894276e-01 -7.34907866e-01 -7.48059094e-01 -9.13147569e-01
8.01137626e-01 -1.45993456e-01 1.53477240e+00 2.00130846e-02
1.82421535e-01 2.23155901e-01 -2.55072385e-01 -9.72487211e-01
-4.95145023e-01 7.64518499e-01 3.75863791e-01 -3.02514672e-01
9.87579942e-01 -2.31405556e-01 -2.61116415e-01 1.64884135e-01
-4.76590812e-01 -2.32639998e-01 4.75780159e-01 6.41806901e-01
-9.02272388e-02 -4.29811686e-01 9.06417549e-01 -1.46004975e+00
6.48966491e-01 -7.63576984e-01 -1.86115786e-01 1.29875541e-01
-8.04499030e-01 -3.75951409e-01 5.16153336e-01 -2.22660437e-01
-1.19007301e+00 -5.82654655e-01 1.23845741e-01 2.66261190e-01
-1.53151393e-01 7.88975358e-01 1.12299040e-01 2.49001339e-01
8.22514176e-01 -4.82990116e-01 -7.81796128e-03 -3.05516750e-01
4.80408631e-02 9.50221956e-01 -7.95351788e-02 -1.16805124e+00
8.34487975e-01 4.54531968e-01 -7.89799631e-01 -7.82312214e-01
-1.08335304e+00 -9.39816460e-02 -6.47301078e-01 -5.27247190e-02
8.27388406e-01 -1.48750687e+00 -1.62898540e-01 4.14515883e-01
-6.10968351e-01 -7.44740427e-01 2.96616033e-02 7.56343126e-01
2.68023703e-02 -2.79034525e-01 -7.23383367e-01 -6.37055039e-01
-1.55078724e-01 -1.05154312e+00 8.85537863e-01 2.33534783e-01
-7.01587796e-01 -1.47142041e+00 4.94993240e-01 5.29919088e-01
3.70955199e-01 -7.28127584e-02 1.02695453e+00 -7.69096613e-01
2.04718530e-01 -7.18515366e-02 -1.33791760e-01 4.00525838e-01
3.94932210e-01 2.30341390e-01 -1.01667631e+00 -6.99978828e-01
-2.94974059e-01 -9.06103313e-01 7.02894330e-01 2.19430000e-01
4.55250591e-01 1.69391826e-01 -2.32682779e-01 3.96287143e-01
1.26375282e+00 -1.17036082e-01 1.91746458e-01 6.61473811e-01
6.35411322e-01 1.17992747e+00 6.50879800e-01 -4.66659553e-02
1.12875009e+00 1.57105789e-01 -5.39349973e-01 -3.93166453e-01
2.97802687e-01 -5.13796747e-01 8.89416635e-01 1.09970069e+00
-1.08385578e-01 3.23079020e-01 -1.30606687e+00 8.86904895e-01
-1.32509923e+00 -7.11691380e-01 1.21996356e-02 2.23986173e+00
1.20859897e+00 2.49528378e-01 2.58866519e-01 -3.80059689e-01
7.00753093e-01 2.46145681e-01 -4.51772153e-01 -8.19802105e-01
-3.68145585e-01 2.61761963e-01 7.18144178e-01 5.93986869e-01
-1.00078797e+00 1.34854543e+00 7.35409832e+00 4.77698743e-01
-1.35331857e+00 9.74928141e-02 1.21307075e+00 -1.23317339e-01
-5.45643330e-01 1.80321217e-01 -8.40207696e-01 2.75568008e-01
9.75530207e-01 -4.55062270e-01 1.54675990e-01 6.95505440e-01
-4.99066245e-03 -2.72800118e-01 -1.05859947e+00 5.34348488e-01
3.41458470e-01 -5.25671303e-01 -2.58511811e-01 1.87653080e-01
1.12933135e+00 3.30979675e-01 4.18666810e-01 9.21501994e-01
1.01820874e+00 -1.08764064e+00 7.84556866e-01 -2.70573556e-01
1.13499868e+00 -9.14764643e-01 7.77719915e-01 6.63410872e-02
-5.37615538e-01 1.49456352e-01 -4.21589375e-01 -5.01644552e-01
-1.47757912e-02 5.41473389e-01 -7.37783313e-01 3.26480955e-01
1.05161035e+00 9.50890541e-01 -1.03444016e+00 1.19558983e-01
-2.74343938e-01 9.14565146e-01 -1.15668833e-01 1.13523178e-01
7.78212696e-02 -4.23869640e-01 2.93374676e-02 1.16481376e+00
1.80377066e-01 -4.98193979e-01 -1.02797402e-02 3.64908516e-01
-3.71065259e-01 5.66029489e-01 -1.06627047e+00 -1.86504528e-01
3.88793409e-01 1.34342825e+00 -6.44404888e-01 -4.75773513e-01
-9.46051717e-01 6.46956444e-01 4.90449667e-01 3.72375816e-01
-4.85734284e-01 -2.38679737e-01 8.14302206e-01 1.16358839e-01
-2.69157261e-01 -9.89812464e-02 -4.69936818e-01 -1.51916742e+00
-2.02599555e-01 -1.35006082e+00 4.28116530e-01 -5.07945061e-01
-1.75060177e+00 2.92407155e-01 -1.69627458e-01 -6.34585261e-01
-1.52826026e-01 -6.71671510e-01 -3.72500241e-01 1.17603576e+00
-1.56295836e+00 -1.38636136e+00 2.19692096e-01 5.83419323e-01
2.69511223e-01 -4.14802253e-01 8.22244525e-01 5.27920425e-01
-6.28601551e-01 8.87592554e-01 -1.13721870e-01 4.75260854e-01
1.94248247e+00 -1.33932495e+00 3.15550148e-01 6.14187658e-01
-1.40171230e-01 1.15676498e+00 6.28675938e-01 -8.71654689e-01
-9.08890128e-01 -8.27451646e-01 1.28057647e+00 -1.00904214e+00
9.48133767e-01 -7.35309243e-01 -6.51149511e-01 1.47065091e+00
7.28580356e-01 -5.48132181e-01 1.23633504e+00 1.19880843e+00
-6.67423666e-01 -1.59934700e-01 -9.61790621e-01 7.00857460e-01
7.20743179e-01 -8.94611537e-01 -6.45918608e-01 7.57446960e-02
5.41360855e-01 -1.81462333e-01 -9.59166169e-01 2.08944112e-01
6.59277201e-01 -8.48371804e-01 4.95162576e-01 -7.73959458e-01
9.73592162e-01 1.34423658e-01 -1.33830011e-01 -1.68486738e+00
7.80769251e-03 -2.83730209e-01 1.02862680e+00 1.64015019e+00
9.81616080e-01 -1.08313537e+00 6.51557088e-01 7.13854849e-01
2.02806992e-03 -3.56725633e-01 -5.57401359e-01 -4.73061144e-01
1.16493952e+00 -2.89446801e-01 5.07517874e-01 1.86515760e+00
5.45924157e-02 6.49506092e-01 -1.97979167e-01 -5.62254824e-02
4.44358408e-01 9.41391941e-03 1.13664436e+00 -1.04875183e+00
1.13035053e-01 -3.35836291e-01 1.02598190e-01 -2.57987738e-01
7.16755509e-01 -1.19869149e+00 -1.69097349e-01 -1.17896032e+00
4.48365420e-01 -8.38378727e-01 -1.02862246e-01 5.79637647e-01
-4.25117493e-01 4.06801403e-01 4.49932627e-02 3.10064834e-02
-2.31001824e-01 2.97212929e-01 9.89910305e-01 6.84902295e-02
-1.12098806e-01 -3.52989286e-01 -1.38776219e+00 9.55138028e-01
7.83771038e-01 -5.11269927e-01 -2.04118654e-01 -6.88503742e-01
6.92536294e-01 -5.49806535e-01 -1.55224681e-01 -4.54542667e-01
-2.46820569e-01 -2.09433183e-01 4.87224728e-01 -1.12986386e-01
6.26783967e-02 -5.46149433e-01 -3.74954134e-01 2.17779294e-01
-2.91280627e-01 5.30586660e-01 3.99585962e-01 -1.89564094e-01
-4.52797502e-01 7.82875493e-02 5.91841280e-01 1.97493727e-03
-1.53225303e-01 6.43918291e-02 -6.39605403e-01 6.21223986e-01
7.81438649e-01 1.62995189e-01 -5.66175938e-01 -4.62560326e-01
-2.22854897e-01 3.37478340e-01 1.17180490e+00 6.11112356e-01
-3.13075483e-01 -1.27639222e+00 -1.02055252e+00 1.73485205e-01
3.61040056e-01 -3.45682830e-01 -1.69196963e-01 1.04334283e+00
-3.51435244e-01 1.86061859e-01 -2.39950880e-01 -3.13582152e-01
-1.05462873e+00 1.29561335e-01 2.22497866e-01 -2.13763833e-01
1.94565371e-01 7.35909045e-01 2.70275414e-01 -1.31927264e+00
-4.65133578e-01 4.84696627e-02 -8.81638825e-02 6.12922549e-01
1.25770271e-01 1.80159271e-01 -1.92946672e-01 -8.04151535e-01
-4.73771811e-01 4.81545776e-01 -3.33048075e-01 -5.23530662e-01
1.38029134e+00 -6.43435791e-02 -4.58458781e-01 9.65665996e-01
1.08712661e+00 8.66877913e-01 -6.02970243e-01 -1.99531287e-01
-4.94981110e-02 -4.21469241e-01 -3.58271450e-01 -7.66853690e-01
-8.81668866e-01 7.64552116e-01 2.27553189e-01 5.99467643e-02
6.46312296e-01 -3.49989557e-03 2.77068228e-01 -1.05373301e-02
5.57441413e-01 -1.44193935e+00 -3.25630188e-01 7.34054387e-01
6.08483553e-01 -1.61450684e+00 2.33518690e-01 -1.28450379e-01
-8.91178548e-01 3.96290839e-01 1.00501621e+00 -4.45962548e-02
7.34777868e-01 2.07676500e-01 7.60429382e-01 -1.48828765e-02
-8.20237100e-01 9.00414512e-02 -2.28286043e-01 4.58265185e-01
1.38107538e+00 2.65701562e-01 -4.65426445e-01 5.18478870e-01
-9.80093956e-01 -2.77016968e-01 5.74784517e-01 8.36639702e-01
2.62905627e-01 -1.53388429e+00 -5.69093466e-01 5.86626589e-01
-9.14751709e-01 -4.20609266e-01 -6.89575136e-01 1.05273843e+00
2.79520422e-01 9.26014543e-01 4.49230611e-01 -2.17032477e-01
1.17086127e-01 3.96857172e-01 4.02395785e-01 -7.81570792e-01
-8.41858208e-01 -1.07717983e-01 4.39838409e-01 2.03813724e-02
-6.44729793e-01 -1.18547821e+00 -7.26025701e-01 -7.33720243e-01
-9.55124348e-02 2.46854931e-01 6.30068481e-01 7.10038364e-01
1.60256296e-01 9.34759825e-02 5.26021719e-01 -5.10775030e-01
-1.50952786e-01 -1.23500085e+00 -5.43634951e-01 5.21651208e-01
1.56361356e-01 -4.36611056e-01 -6.29027724e-01 3.25879008e-02] | [9.864645957946777, 10.163237571716309] |
2f995554-e308-452e-a1d0-a998d08e7af9 | saint-improved-neural-networks-for-tabular | 2106.01342 | null | https://arxiv.org/abs/2106.01342v1 | https://arxiv.org/pdf/2106.01342v1.pdf | SAINT: Improved Neural Networks for Tabular Data via Row Attention and Contrastive Pre-Training | Tabular data underpins numerous high-impact applications of machine learning from fraud detection to genomics and healthcare. Classical approaches to solving tabular problems, such as gradient boosting and random forests, are widely used by practitioners. However, recent deep learning methods have achieved a degree of performance competitive with popular techniques. We devise a hybrid deep learning approach to solving tabular data problems. Our method, SAINT, performs attention over both rows and columns, and it includes an enhanced embedding method. We also study a new contrastive self-supervised pre-training method for use when labels are scarce. SAINT consistently improves performance over previous deep learning methods, and it even outperforms gradient boosting methods, including XGBoost, CatBoost, and LightGBM, on average over a variety of benchmark tasks. | ['Tom Goldstein', 'C. Bayan Bruss', 'Avi Schwarzschild', 'Micah Goldblum', 'Gowthami Somepalli'] | 2021-06-02 | saint-improved-neural-networks-for-tabular-1 | https://openreview.net/forum?id=nL2lDlsrZU | https://openreview.net/pdf?id=nL2lDlsrZU | null | ['insurance-prediction'] | ['miscellaneous'] | [-2.03025728e-01 -6.46900535e-02 -8.52999210e-01 -5.08437276e-01
-8.89134586e-01 -2.19107121e-01 6.24120414e-01 4.54724401e-01
-2.09384918e-01 1.20258832e+00 3.23434651e-01 -5.70100486e-01
-2.11340904e-01 -1.15756249e+00 -7.53191948e-01 -7.48242497e-01
7.85626546e-02 9.11589503e-01 -4.66939062e-01 -3.57227951e-01
2.60328114e-01 1.03570677e-01 -1.25076509e+00 6.34246349e-01
7.81711936e-01 1.44824243e+00 -6.79125309e-01 2.51836330e-01
-5.43686569e-01 1.20950198e+00 -3.68211061e-01 -1.00653303e+00
1.85931087e-01 -2.23175004e-01 -9.77404714e-01 -4.29078102e-01
5.84962368e-01 -2.32584745e-01 -6.17224276e-02 5.95606327e-01
6.15568221e-01 -2.74828821e-01 7.41297722e-01 -1.35761142e+00
-6.46545112e-01 6.15021467e-01 -9.16464984e-01 -4.95834388e-02
1.94547653e-01 -1.61588475e-01 1.54529130e+00 -7.81390071e-01
6.58355415e-01 1.19181037e+00 1.27909827e+00 4.79374230e-01
-1.32364583e+00 -7.14437306e-01 8.42146277e-02 6.76441967e-01
-7.06031024e-01 -1.82450697e-01 8.27084363e-01 -4.62131321e-01
8.59432042e-01 2.43259043e-01 7.93452144e-01 1.36211991e+00
4.54309374e-01 1.12832999e+00 1.36023962e+00 -4.37468946e-01
4.09035504e-01 -1.66066028e-02 2.93275565e-01 8.81011188e-01
4.52758789e-01 7.08300471e-02 -8.73759747e-01 -6.57489359e-01
1.96499214e-01 2.47406095e-01 1.98439717e-01 -7.84275353e-01
-9.97708201e-01 1.34886837e+00 4.36205715e-01 -9.73433480e-02
-4.27513123e-01 3.96026611e-01 7.85356879e-01 4.67165768e-01
8.90182614e-01 5.10215819e-01 -7.65979469e-01 -5.48690259e-02
-8.30198288e-01 5.73850691e-01 7.52937138e-01 5.16052186e-01
7.18271732e-01 -2.44352221e-01 -2.50356972e-01 1.11193395e+00
2.84801535e-02 1.91941947e-01 4.73608434e-01 -7.91083694e-01
5.26172876e-01 8.03299546e-01 1.75333649e-01 -9.14695442e-01
-7.59239435e-01 -5.78406215e-01 -1.14285815e+00 3.52321640e-02
3.99309933e-01 2.19727568e-02 -1.04507339e+00 1.24014604e+00
6.10296667e-01 -5.20510316e-01 -2.90340006e-01 6.90889716e-01
8.91383350e-01 4.47857380e-01 9.45214257e-02 -1.81816313e-02
1.19380820e+00 -1.15122366e+00 -7.62386620e-01 -4.08616930e-01
8.03075969e-01 -3.86936009e-01 8.63578320e-01 9.20021415e-01
-9.08172965e-01 -1.22984514e-01 -9.00644600e-01 -4.79570776e-01
-7.29521573e-01 -2.93275237e-01 1.21876526e+00 8.89235318e-01
-8.73013794e-01 7.36604691e-01 -3.76783639e-01 -5.16981706e-02
1.00974488e+00 2.46915981e-01 -4.79253709e-01 -3.97472948e-01
-1.30412650e+00 8.76443207e-01 5.68187516e-03 1.04560569e-01
-6.22909665e-01 -5.41612327e-01 -7.61680007e-01 -2.44165435e-02
2.86159635e-01 -1.09428263e+00 1.28503549e+00 -5.82283735e-01
-1.25584757e+00 1.10691714e+00 -8.99118781e-02 -8.08900952e-01
8.20524871e-01 -4.28221375e-01 3.98703143e-02 -1.79574728e-01
1.83238089e-01 5.29221117e-01 7.68080235e-01 -9.66678023e-01
-6.53345942e-01 -7.59471714e-01 -4.75678928e-02 9.95790213e-02
-5.92606425e-01 -3.55061948e-01 -4.25563902e-02 -3.42116177e-01
2.13706642e-01 -6.21956468e-01 -4.64111239e-01 -3.70716974e-02
-5.82314253e-01 -4.76314545e-01 3.21288377e-01 -7.09101200e-01
9.54505801e-01 -1.54558623e+00 1.34223397e-03 6.01072088e-02
6.07931018e-01 -1.50307059e-01 2.13108346e-01 5.61171830e-01
-1.19563892e-01 1.68646351e-01 -4.88502048e-02 -3.55685502e-01
1.63451105e-01 5.80202341e-02 -2.45434016e-01 5.54370522e-01
-5.41230962e-02 1.13157058e+00 -8.64649951e-01 -5.26358604e-01
7.55143538e-02 1.27011925e-01 -6.45804524e-01 -8.81004333e-02
-4.47935402e-01 -1.18465640e-01 -1.29629642e-01 1.02687776e+00
6.83493674e-01 -5.77732325e-01 4.74185556e-01 1.41317442e-01
1.99907988e-01 7.92906165e-01 -6.28579915e-01 1.47423244e+00
-4.34109122e-01 4.66244698e-01 -2.10026830e-01 -1.42076576e+00
9.43531215e-01 2.74757738e-03 7.39792526e-01 -1.07071590e+00
-1.15676522e-01 2.27993697e-01 -3.41426700e-01 -1.96779296e-01
3.38687778e-01 -2.35995606e-01 -4.95159365e-02 3.26906860e-01
-4.62195650e-02 8.03997144e-02 8.88544768e-02 7.78264478e-02
1.14718914e+00 1.12585329e-01 6.13778710e-01 -3.40945095e-01
6.66629747e-02 5.39497793e-01 6.53829098e-01 1.05761945e+00
-5.54541126e-02 4.28850710e-01 9.57963586e-01 -1.39498997e+00
-1.03303313e+00 -7.08051085e-01 -3.64569575e-01 1.48359525e+00
-1.37522787e-01 -5.72652459e-01 -4.11476046e-01 -1.10797858e+00
6.63127244e-01 3.85684907e-01 -1.05657983e+00 5.32731041e-02
-3.53253812e-01 -1.52315903e+00 7.56659731e-02 6.38598084e-01
4.90834862e-01 -1.02811313e+00 -2.84176737e-01 5.03415585e-01
-1.51875347e-01 -5.55670619e-01 2.58246422e-01 9.67359066e-01
-1.21104264e+00 -1.13986480e+00 -4.70759571e-01 -4.93987620e-01
1.51238635e-01 -4.91428077e-02 1.72581887e+00 7.11633870e-03
-3.26719344e-01 -3.31756771e-01 -3.24078262e-01 -5.48039615e-01
9.87335294e-03 4.55617607e-01 -2.40303189e-01 -3.44189942e-01
6.80581212e-01 -1.95906788e-01 -8.75385284e-01 1.86098859e-01
-3.61125410e-01 1.05047405e-01 4.90688890e-01 1.67488790e+00
5.41273355e-01 -2.59527683e-01 6.36067808e-01 -1.68224907e+00
6.81292713e-01 -6.97883248e-01 -4.07480538e-01 4.06363457e-02
-1.21579707e+00 1.43766567e-01 5.99947333e-01 1.82167321e-01
-5.56499362e-01 -2.72639692e-01 -4.25531536e-01 -1.95728429e-02
1.36695877e-01 8.82277191e-01 8.60662237e-02 2.89484501e-01
9.48100686e-01 -9.98445302e-02 -7.12649226e-02 -5.55905461e-01
3.90323728e-01 6.14455104e-01 5.88572696e-02 -4.63806629e-01
4.62971061e-01 7.24044502e-01 8.93607065e-02 -1.16041061e-02
-1.04374528e+00 -4.10328567e-01 -2.81070471e-01 6.15244620e-02
4.68659073e-01 -8.85724545e-01 -7.26769924e-01 5.40446639e-01
-8.36717427e-01 -3.12819004e-01 -4.57562469e-02 -6.83807116e-03
-5.37904263e-01 4.39158566e-02 -6.81169868e-01 -6.95175707e-01
-7.19178081e-01 -7.08374381e-01 1.19806767e+00 -3.27705801e-01
-2.37943500e-01 -1.03250098e+00 3.00014406e-01 6.69943988e-01
2.86719173e-01 4.97347534e-01 1.28611159e+00 -6.74877763e-01
-4.09812361e-01 -2.39839941e-01 -3.14803988e-01 5.76959029e-02
-3.22067440e-02 -3.99413794e-01 -1.11266923e+00 -2.95869023e-01
-4.18291867e-01 -7.87005901e-01 1.58190620e+00 5.61895847e-01
1.59150159e+00 -4.67188299e-01 -6.52939439e-01 8.01437259e-01
1.34026384e+00 1.35421202e-01 6.00005805e-01 1.05876875e+00
7.34722316e-01 6.49751961e-01 5.95090270e-01 4.05248940e-01
6.43164039e-01 4.27212089e-01 7.10810184e-01 -5.05309105e-01
4.23278511e-01 -2.54768670e-01 -8.72973949e-02 4.57994521e-01
1.43351525e-01 -1.39834553e-01 -1.26031578e+00 6.19767249e-01
-2.29913163e+00 -9.15882468e-01 -1.57337934e-01 1.99132228e+00
8.57770741e-01 3.99059355e-01 3.91483366e-01 5.37715793e-01
1.86378479e-01 2.46424288e-01 -8.15177619e-01 -7.54629493e-01
-6.38844147e-02 4.04796064e-01 6.84518456e-01 8.13589897e-03
-1.39110148e+00 7.27160335e-01 7.24924564e+00 7.62074828e-01
-8.26853275e-01 2.04127908e-01 1.34243548e+00 -2.34604329e-01
-4.02914524e-01 -3.49258482e-01 -7.36376286e-01 3.99774313e-01
7.36325562e-01 1.88527867e-01 3.22336882e-01 1.30501103e+00
-2.31243417e-01 -5.52698374e-02 -1.27093136e+00 1.10740876e+00
-8.99244994e-02 -1.80308652e+00 -1.80194959e-01 1.48742378e-01
8.58744621e-01 1.38727769e-01 3.20247769e-01 6.03155911e-01
8.75190735e-01 -1.36342537e+00 4.76374805e-01 2.93690115e-02
8.36413264e-01 -8.40515137e-01 1.01684213e+00 7.78874680e-02
-4.40507650e-01 -4.58680063e-01 -4.24266338e-01 -3.75222027e-01
-5.51804900e-01 1.19746411e+00 -7.65186012e-01 6.29006267e-01
1.27481866e+00 8.17469001e-01 -4.17821318e-01 9.36166942e-01
-1.03605032e-01 6.78378522e-01 -2.54417770e-02 -3.54238421e-01
3.19961935e-01 5.62960319e-02 -1.91704571e-01 1.05443263e+00
-4.35460582e-02 -5.99928975e-01 -1.43434480e-01 4.18630689e-01
-4.67794597e-01 2.82618999e-01 -6.71197891e-01 1.57016620e-01
1.77389786e-01 1.01261568e+00 -3.98080647e-01 -3.76837164e-01
-5.12440205e-01 5.05414248e-01 6.81369960e-01 3.54762264e-02
-4.89601672e-01 -2.40421265e-01 7.47795880e-01 1.13092661e-01
3.56790841e-01 3.10296148e-01 -8.71723771e-01 -1.09438682e+00
4.41222563e-02 -1.31013644e+00 8.69713783e-01 -5.12399971e-01
-1.69405496e+00 3.36973906e-01 -3.21168005e-01 -8.43697369e-01
-4.98063952e-01 -9.94644463e-01 -1.92930877e-01 6.83069050e-01
-1.83893871e+00 -9.37149167e-01 -2.20822960e-01 4.30412292e-01
5.01764901e-02 -4.47661310e-01 8.74224544e-01 2.53302246e-01
-5.28716564e-01 6.43699586e-01 8.40663731e-01 6.31761029e-02
6.89940453e-01 -1.57591796e+00 6.67891383e-01 1.97825238e-01
1.87117487e-01 5.13700962e-01 6.62509620e-01 -4.36034203e-01
-1.38864398e+00 -1.07993436e+00 9.80262160e-01 -4.65853930e-01
5.03825605e-01 -7.03654289e-01 -5.02247572e-01 6.02576971e-01
2.62715638e-01 -4.14187759e-02 8.11217904e-01 1.14291286e+00
-7.72201657e-01 -6.02691650e-01 -1.13970542e+00 3.27345043e-01
1.00201058e+00 -3.97781372e-01 -3.06570441e-01 7.72192597e-01
1.76680818e-01 -3.61805975e-01 -5.59787869e-01 5.44986129e-01
9.56544399e-01 -1.16324651e+00 9.56750214e-01 -1.10328770e+00
1.21278644e+00 3.23828071e-01 2.76859999e-02 -1.63330865e+00
-5.58219254e-01 -3.85467440e-01 -4.63945389e-01 5.51336348e-01
3.28388125e-01 -6.11987591e-01 1.42582357e+00 2.37015054e-01
2.41746157e-01 -1.21052396e+00 -8.77338588e-01 -6.47379518e-01
5.62149644e-01 -9.27432328e-02 8.85826886e-01 1.31437302e+00
5.86240776e-02 3.49952072e-01 -6.36894822e-01 -5.90822816e-01
7.55619049e-01 6.83211207e-01 1.00630510e+00 -1.48173773e+00
-1.47965148e-01 -5.61425626e-01 -2.90875912e-01 -7.37455964e-01
-2.88901683e-02 -9.70176518e-01 -2.96221465e-01 -2.01113605e+00
4.74625647e-01 -5.74443817e-01 -8.17616284e-01 7.42457986e-01
-3.27240050e-01 5.23225486e-01 -3.30508560e-01 5.22042029e-02
-5.95170379e-01 3.84553224e-01 9.12159741e-01 -6.81569576e-01
1.33599475e-01 -1.50774032e-01 -1.14697015e+00 4.06528771e-01
9.07009304e-01 -8.13193142e-01 1.65589958e-01 -4.57803726e-01
8.78604889e-01 6.46567419e-02 1.65174574e-01 -6.17758214e-01
-1.30422011e-01 -1.85105547e-01 8.14508438e-01 -7.17288792e-01
1.31730527e-01 -5.09239078e-01 -9.88467187e-02 5.08213997e-01
-3.67802173e-01 2.64449865e-01 1.26197487e-01 4.77236360e-01
-3.39216709e-01 -9.23623913e-04 5.35584152e-01 -2.37452462e-01
-3.27932119e-01 3.40776622e-01 -6.08722083e-02 1.12465667e-02
5.97081840e-01 -2.41112821e-02 -7.78937936e-01 -4.73434150e-01
-4.51266378e-01 5.04769087e-01 1.98054373e-01 1.61625087e-01
4.05438483e-01 -1.41992402e+00 -8.36150050e-01 -3.56549695e-02
3.48006129e-01 -1.37027770e-01 -1.07142948e-01 8.06182027e-01
-6.51890755e-01 5.11956394e-01 -4.60269839e-01 -4.20102894e-01
-9.38856125e-01 6.50688589e-01 2.31490061e-01 -7.97155797e-01
-4.72384840e-01 1.07457709e+00 6.11711070e-02 -7.52442718e-01
2.61003345e-01 -3.44472006e-02 -9.16116610e-02 4.60002333e-01
1.86558977e-01 4.69023615e-01 6.39039397e-01 1.85750365e-01
-4.51392472e-01 3.09431963e-02 -5.76348424e-01 2.63916373e-01
1.68327367e+00 4.49931324e-01 -2.68813014e-01 4.43052590e-01
1.12643087e+00 -2.16697812e-01 -6.74764991e-01 -1.51651561e-01
1.82969600e-01 -4.50105280e-01 2.85814375e-01 -1.24963403e+00
-9.66206551e-01 1.02623200e+00 4.05657113e-01 4.47739989e-01
9.99299228e-01 -3.54814142e-01 7.85904527e-01 6.59901619e-01
4.14274693e-01 -1.12519002e+00 -1.02286898e-01 2.99099028e-01
5.18545270e-01 -1.66386151e+00 4.46794122e-01 1.90380104e-02
-3.72323573e-01 9.84824538e-01 4.76436198e-01 -5.73263839e-02
4.28603321e-01 2.76612103e-01 2.37813473e-01 -3.14919859e-01
-1.24090672e+00 -3.31131294e-02 6.58203289e-02 5.00552475e-01
6.31197155e-01 1.59000278e-01 -5.71262717e-01 3.53727251e-01
-3.48885804e-01 2.09755152e-01 6.99249953e-02 7.31728554e-01
-1.52732879e-01 -1.15333009e+00 -4.60616291e-01 1.30198324e+00
-7.19023347e-01 -3.57924521e-01 -7.19683170e-01 8.11738193e-01
5.61512727e-03 7.00548172e-01 -3.36820446e-02 -3.85224164e-01
-8.22112188e-02 2.91802108e-01 3.60507727e-01 -1.75950930e-01
-8.00109684e-01 -2.75192827e-01 3.83759201e-01 -6.65303349e-01
-8.84407684e-02 -7.65731931e-01 -4.03238624e-01 -6.82651043e-01
-2.34046459e-01 2.00112104e-01 6.54364049e-01 6.68557644e-01
1.79486006e-01 3.50358874e-01 6.10084653e-01 -3.66141498e-01
-6.50098443e-01 -8.54229152e-01 -4.78428453e-01 4.66978878e-01
5.52188516e-01 -9.67367947e-01 -1.46161675e-01 -4.81107652e-01] | [8.593734741210938, 4.160357475280762] |
0ff94cce-af1d-42d1-863f-7866ca891ba7 | seq2path-generating-sentiment-tuples-as-paths | null | null | https://aclanthology.org/2022.findings-acl.174 | https://aclanthology.org/2022.findings-acl.174.pdf | Seq2Path: Generating Sentiment Tuples as Paths of a Tree | Aspect-based sentiment analysis (ABSA) tasks aim to extract sentiment tuples from a sentence. Recent generative methods such as Seq2Seq models have achieved good performance by formulating the output as a sequence of sentiment tuples. However, the orders between the sentiment tuples do not naturally exist and the generation of the current tuple should not condition on the previous ones. In this paper, we propose Seq2Path to generate sentiment tuples as paths of a tree. A tree can represent “1-to-n” relations (e.g., an aspect term may correspond to multiple opinion terms) and the paths of a tree are independent and do not have orders. For training, we treat each path as an independent target, and we calculate the average loss of the ordinary Seq2Seq model over paths. For inference, we apply beam search with constrained decoding. By introducing an additional discriminative token and applying a data augmentation technique, valid paths can be automatically selected. We conduct experiments on five tasks including AOPE, ASTE, TASD, UABSA, ACOS. We evaluate our method on four common benchmark datasets including Laptop14, Rest14, Rest15, Rest16. Our proposed method achieves state-of-the-art results in almost all cases. | ['Longjun Cai', 'Xiaoying Zhu', 'Jingchao Yang', 'Yi Shen', 'Yue Mao'] | null | null | null | null | findings-acl-2022-5 | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 3.02897483e-01 1.53356031e-01 -3.05665415e-02 -8.26744914e-01
-8.37325633e-01 -6.59359217e-01 4.46539134e-01 9.16581899e-02
-9.48776379e-02 9.39681232e-01 3.87586355e-01 -4.01127517e-01
2.77245790e-01 -1.08186865e+00 -9.39842045e-01 -6.74603820e-01
2.21487314e-01 6.77215695e-01 5.62976189e-02 -4.27101254e-01
4.89322431e-02 -1.80534974e-01 -1.25816178e+00 6.81181133e-01
7.67305017e-01 1.04699993e+00 -1.03956632e-01 6.47283316e-01
-5.05175889e-01 6.89362049e-01 -1.01914775e+00 -1.08809459e+00
4.84988801e-02 -8.72573614e-01 -8.44380081e-01 -7.29566962e-02
-1.12549439e-01 6.54652044e-02 2.70041436e-01 9.72326100e-01
4.82617408e-01 -1.93809532e-02 5.58503687e-01 -1.17198765e+00
-6.26476586e-01 1.13417208e+00 -5.79436660e-01 -2.40870699e-01
3.23979706e-01 -2.08412021e-01 1.43685496e+00 -7.76099741e-01
5.06842494e-01 1.31485069e+00 3.09484333e-01 5.44853151e-01
-9.29279029e-01 -5.68545043e-01 5.00990927e-01 -1.16143733e-01
-9.98840988e-01 -2.51547933e-01 5.59602380e-01 7.51705617e-02
9.16587532e-01 5.26134014e-01 7.41237521e-01 1.12938380e+00
2.88241208e-01 1.12663412e+00 9.75042582e-01 -2.12512612e-01
4.57639515e-01 2.06806406e-01 1.64434478e-01 4.20817941e-01
1.30052760e-01 -3.82381231e-01 -9.17973578e-01 -1.94491014e-01
8.24967027e-02 -2.50515729e-01 1.15062609e-01 1.81658641e-01
-1.01341140e+00 7.88385391e-01 2.89719939e-01 -1.35470405e-01
-3.23047101e-01 7.06497207e-03 4.13829416e-01 2.14103803e-01
3.68450403e-01 3.49386275e-01 -9.78885353e-01 -2.35107005e-01
-5.65203846e-01 5.12038052e-01 9.72246110e-01 1.28065395e+00
8.34284723e-01 -2.63612747e-01 -5.88246405e-01 7.60711014e-01
3.26819152e-01 6.45515025e-01 5.51986992e-01 -2.95422852e-01
6.38717353e-01 5.49082577e-01 2.70299278e-02 -7.86220968e-01
-1.66204214e-01 -3.31890255e-01 -8.58928621e-01 -4.19550300e-01
-2.98322458e-02 -7.40377367e-01 -1.23231637e+00 1.62424982e+00
5.06334007e-01 3.29721496e-02 3.55460376e-01 7.07455754e-01
7.85747826e-01 9.19568062e-01 -2.83224450e-04 -2.42022410e-01
1.46661556e+00 -1.15149200e+00 -6.15205884e-01 -5.11520326e-01
6.71363592e-01 -7.77568519e-01 9.26868975e-01 6.02949679e-01
-9.03900504e-01 -2.60874242e-01 -8.49397182e-01 1.70480922e-01
-2.81324655e-01 1.30421463e-02 9.06759202e-01 6.35407627e-01
-6.83186948e-01 3.35501820e-01 -7.56005168e-01 1.32363796e-01
2.72811174e-01 2.97942758e-01 -4.60094661e-02 -3.66813205e-02
-1.34743655e+00 2.62201577e-01 4.07202393e-01 3.73486519e-01
-6.34728670e-01 -5.69972575e-01 -9.27861333e-01 -7.98647180e-02
4.89224017e-01 -9.64311898e-01 1.24941027e+00 -9.34772551e-01
-1.40918231e+00 5.01209140e-01 -8.10097754e-01 -2.79066980e-01
1.73731327e-01 -4.18534636e-01 -5.62017202e-01 -3.30493152e-01
3.32208246e-01 4.76811528e-01 5.04720926e-01 -1.17339468e+00
-8.96654725e-01 -3.72229904e-01 3.33092660e-01 3.09289604e-01
-1.81750700e-01 1.60241544e-01 -5.96741676e-01 -4.77241129e-01
3.67678627e-02 -1.07930243e+00 -5.22640228e-01 -8.02542090e-01
-1.10730767e+00 -2.57327318e-01 3.72178972e-01 -2.54319727e-01
1.29321957e+00 -1.95430100e+00 5.13851456e-02 2.07961529e-01
-3.17010909e-01 1.00130588e-01 -6.22966401e-02 3.66257191e-01
1.12815678e-01 3.88991177e-01 -4.25443470e-01 -5.10013521e-01
-1.37712145e-02 3.22557807e-01 -5.25051415e-01 -2.75003850e-01
3.64663303e-01 9.85966086e-01 -8.53856802e-01 -5.05159020e-01
-5.26726007e-01 2.58468926e-01 -7.08089292e-01 3.75092477e-01
-7.99695909e-01 2.76813090e-01 -7.41001844e-01 6.11476660e-01
7.03228414e-01 -1.97285280e-01 2.49403864e-01 3.42015959e-02
9.84483287e-02 7.58269727e-01 -1.07159030e+00 1.60033727e+00
-7.21239984e-01 1.31789684e-01 -5.45116603e-01 -5.58812499e-01
1.05418491e+00 1.69906065e-01 2.54695322e-02 -6.00605011e-01
9.07673687e-02 8.83153155e-02 2.57321093e-02 -3.91171038e-01
7.12152839e-01 -3.53663385e-01 -5.22961318e-01 5.33239603e-01
-8.51323456e-03 -5.43552861e-02 5.50403357e-01 2.74039835e-01
8.90915334e-01 1.19872533e-01 6.89481944e-02 9.39656794e-02
5.40278018e-01 3.73284682e-03 8.13607395e-01 6.98364377e-01
5.47964394e-01 7.70115018e-01 1.21995997e+00 -8.16684663e-02
-7.34975696e-01 -8.44897211e-01 2.07921155e-02 9.58943546e-01
1.03421934e-01 -8.46783817e-01 -5.32547891e-01 -1.14317238e+00
-3.60888839e-01 1.02333272e+00 -7.16962993e-01 -9.77313891e-02
-4.87426698e-01 -1.24576128e+00 2.56508321e-01 5.09834051e-01
2.92258918e-01 -1.39415967e+00 -2.43107036e-01 2.12809667e-01
-3.29545319e-01 -1.18567169e+00 -4.52748746e-01 2.52141267e-01
-6.32003188e-01 -8.03994954e-01 -3.01619649e-01 -7.49308884e-01
8.65910649e-01 -1.65778011e-01 1.40569317e+00 -2.64904559e-01
2.47039273e-01 -4.60815966e-01 -7.46950448e-01 -6.01760983e-01
-1.52225241e-01 3.99110913e-01 -3.29685450e-01 3.19935650e-01
4.76784796e-01 -3.75899106e-01 -6.28235996e-01 1.32079199e-01
-8.73664379e-01 3.27856243e-02 7.14637160e-01 8.79101574e-01
1.03009701e+00 2.03038275e-01 5.76722503e-01 -1.62191617e+00
8.13331544e-01 -6.66822910e-01 -2.46049553e-01 1.88351229e-01
-5.24806499e-01 1.46614194e-01 1.03161383e+00 -1.16844065e-01
-1.20540750e+00 -4.17230204e-02 -4.13081557e-01 2.83853356e-02
-2.02081621e-01 9.08023357e-01 -5.50190866e-01 8.71389925e-01
2.92135805e-01 2.91386038e-01 -5.55211246e-01 -2.11658761e-01
3.19656014e-01 6.07787728e-01 9.85056832e-02 -6.36384368e-01
5.50762236e-01 2.93005735e-01 -2.35175062e-02 -5.00166714e-01
-1.45327318e+00 -1.31103784e-01 -3.14507276e-01 1.59468949e-01
6.24286056e-01 -8.33076060e-01 -5.34301758e-01 4.73563641e-01
-1.26632440e+00 -1.60268620e-01 -1.95898026e-01 1.56425044e-01
-1.13479555e-01 -1.29159287e-01 -3.82305980e-01 -9.36397195e-01
-8.43779325e-01 -1.09440088e+00 1.20898914e+00 4.43069935e-01
-4.50995207e-01 -7.31593490e-01 9.13810283e-02 1.50965571e-01
7.95665309e-02 8.12051892e-02 9.99810398e-01 -8.38531256e-01
-5.32917023e-01 -1.97342962e-01 3.10524315e-01 4.12023693e-01
1.09043755e-01 8.84138197e-02 -8.11164439e-01 1.69151336e-01
-1.55685693e-01 -3.29658061e-01 9.53940749e-01 7.94577077e-02
1.06268501e+00 -6.29452229e-01 -2.72098184e-01 5.54718196e-01
1.22037542e+00 4.14714754e-01 7.04998791e-01 1.85628220e-01
7.98097968e-01 6.19810820e-01 8.13719451e-01 3.88322562e-01
5.64179063e-01 2.89989948e-01 3.63075346e-01 3.17801744e-01
3.12647074e-01 -5.25286973e-01 6.38459027e-01 9.78588641e-01
3.90674919e-01 -7.20653236e-01 -6.09219670e-01 6.40206516e-01
-1.77119911e+00 -5.98748446e-01 -4.22763556e-01 2.15151739e+00
1.02232170e+00 5.50442934e-01 -8.55702311e-02 8.91024843e-02
3.23772907e-01 2.88032979e-01 -6.04137719e-01 -6.89322948e-01
-2.55106539e-01 4.52889353e-01 8.06867629e-02 4.14228171e-01
-9.56463993e-01 1.18727076e+00 5.25360680e+00 6.16355419e-01
-9.92374182e-01 -1.04531527e-01 1.08037019e+00 -3.23770076e-01
-9.29352403e-01 1.78740680e-01 -1.17871916e+00 5.96090376e-01
7.53485858e-01 -2.00810522e-01 1.32600456e-01 7.26558089e-01
-9.39533189e-02 -1.62709996e-01 -1.00047016e+00 3.42563212e-01
1.78709686e-01 -8.66947055e-01 4.04447585e-01 -4.54928875e-01
9.94868755e-01 -2.74366796e-01 1.66135833e-01 4.60773408e-01
3.81073445e-01 -1.05483520e+00 6.20635927e-01 3.27451617e-01
5.24072826e-01 -1.17249238e+00 9.84130740e-01 2.09365219e-01
-9.99540567e-01 1.69111535e-01 -3.68249625e-01 2.13337705e-01
4.19578671e-01 1.04611146e+00 -7.62291789e-01 8.74426544e-01
7.21394479e-01 6.35034800e-01 -3.57334226e-01 6.56361520e-01
-7.91698635e-01 8.73149872e-01 -1.41395569e-01 -6.37317717e-01
2.59214878e-01 -2.96601474e-01 4.18621182e-01 1.14523399e+00
3.43546659e-01 9.18316692e-02 -1.58030331e-01 7.18198538e-01
-1.88660488e-01 2.99657941e-01 -2.83792794e-01 -5.06163538e-02
3.31350476e-01 1.38747907e+00 -6.51783049e-01 -3.34427118e-01
-3.54676545e-01 9.90150154e-01 8.52749124e-02 3.36532235e-01
-7.38108397e-01 -6.37864232e-01 8.22295785e-01 -2.88521528e-01
6.18335962e-01 2.01756015e-01 -5.24117291e-01 -1.08240795e+00
3.08669031e-01 -1.09667242e+00 2.95377761e-01 -7.84642279e-01
-1.19051516e+00 9.38398421e-01 -3.47900331e-01 -9.82096970e-01
-4.02106911e-01 -3.31061393e-01 -8.96690726e-01 9.81520832e-01
-1.31260526e+00 -1.05346370e+00 5.00966869e-02 3.25025678e-01
5.16865790e-01 -2.23442875e-02 9.42519724e-01 6.75240904e-02
-7.09294081e-01 7.80662239e-01 -3.28743547e-01 1.98662087e-01
6.03440225e-01 -1.33591855e+00 8.02445054e-01 9.45784390e-01
1.51067272e-01 8.22577894e-01 7.27665305e-01 -8.18331718e-01
-1.43829489e+00 -1.26228559e+00 1.24337351e+00 -4.12013501e-01
5.47387600e-01 -6.59513354e-01 -6.20798945e-01 8.35509658e-01
3.00110161e-01 -4.48795885e-01 8.48790050e-01 5.28149545e-01
-2.22255543e-01 -2.62148023e-01 -6.14440560e-01 8.26274991e-01
9.90807295e-01 -1.09884866e-01 -2.24938035e-01 2.84943551e-01
1.19966412e+00 -5.12788415e-01 -4.38994408e-01 4.72926825e-01
4.71669883e-01 -9.46652710e-01 4.01929677e-01 -9.14146662e-01
1.16432691e+00 -2.44737387e-01 -7.76812360e-02 -1.67592847e+00
1.09861545e-01 -6.44247115e-01 -2.88813174e-01 1.59002423e+00
1.23337400e+00 -4.56153333e-01 9.86326516e-01 4.05627549e-01
-1.56245589e-01 -1.26772022e+00 -5.42122781e-01 -4.64916825e-01
-1.88533217e-02 -5.47097862e-01 1.28150034e+00 5.51775753e-01
1.37809096e-02 9.42624688e-01 -4.22653586e-01 2.19276384e-01
3.28829527e-01 5.60942769e-01 8.16232502e-01 -7.26138115e-01
-6.22727215e-01 -2.18616098e-01 6.47541694e-03 -1.24130833e+00
1.11928374e-01 -7.87896514e-01 2.31087327e-01 -1.66712666e+00
2.62595117e-01 -7.25061595e-01 -2.51953036e-01 5.34194291e-01
-6.48881614e-01 -7.69447461e-02 7.66309304e-03 -3.26682478e-01
-6.49068773e-01 7.85864353e-01 1.42939937e+00 -1.41643837e-01
-1.70846149e-01 3.91870737e-01 -1.24302208e+00 4.94108140e-01
8.99107099e-01 -7.22886741e-01 -5.86448252e-01 -5.49473345e-01
1.14876056e+00 2.64943749e-01 -2.48582035e-01 -4.72745478e-01
2.13509709e-01 -3.84063348e-02 2.62910128e-01 -6.87401056e-01
2.53784657e-01 -5.03271103e-01 -4.26145867e-02 1.21655680e-01
-4.21400338e-01 7.34443292e-02 3.07474844e-02 4.47229683e-01
-4.84358966e-01 -2.61357695e-01 9.41666365e-02 -4.61728498e-02
-3.30236524e-01 4.39863980e-01 -1.06961273e-01 2.24845842e-01
7.45202303e-01 4.08904910e-01 -3.05236638e-01 -4.06905472e-01
-5.19534945e-01 5.24069011e-01 3.34321931e-02 6.67410433e-01
6.41050696e-01 -1.23724318e+00 -7.42601812e-01 2.32043490e-01
1.16054356e-01 7.08850682e-01 1.67324588e-01 4.94491488e-01
-5.63854985e-02 3.31283838e-01 3.34106147e-01 -2.73262531e-01
-1.10209799e+00 2.84404665e-01 1.60320684e-01 -5.06657958e-01
-6.00792766e-02 1.23888803e+00 3.17631036e-01 -6.95424020e-01
-1.96070001e-01 -1.39614180e-01 -3.82903993e-01 2.11496070e-01
4.41174001e-01 -1.18851006e-01 1.61972463e-01 -3.56897116e-01
-3.33980858e-01 3.11172783e-01 -3.76363009e-01 -2.91665167e-01
1.12624121e+00 1.69714004e-01 -4.20964628e-01 6.45700395e-01
1.17532361e+00 3.30963761e-01 -6.85072184e-01 -1.11510616e-03
-2.32505128e-01 -2.48648331e-01 -5.11992455e-01 -9.63648140e-01
-1.22230351e+00 6.53592706e-01 -2.70200402e-01 3.19527835e-01
1.20338368e+00 1.58936709e-01 1.13671088e+00 1.56871974e-01
3.34355444e-01 -7.79244483e-01 -7.22469464e-02 8.65781307e-01
7.02496827e-01 -1.09472680e+00 -2.65016854e-01 -5.12190878e-01
-1.13876021e+00 7.40943611e-01 7.29323208e-01 5.73611669e-02
5.83898246e-01 4.40744907e-01 2.14302748e-01 -1.29887432e-01
-1.21566856e+00 -1.17198236e-01 9.09738019e-02 2.02585265e-01
5.67603350e-01 2.63467222e-01 -4.90862101e-01 1.06752765e+00
-9.19077694e-01 -1.71356276e-01 4.86785263e-01 7.41943300e-01
-4.57851999e-02 -1.37615871e+00 -7.88399130e-02 7.30492830e-01
-8.07595551e-01 -5.38800240e-01 -4.37257707e-01 2.96351314e-01
6.94841295e-02 1.12979519e+00 1.01015143e-01 -6.56172454e-01
6.02948964e-01 9.48780403e-02 1.23996250e-01 -7.59026289e-01
-6.76154494e-01 4.76226434e-02 5.38459480e-01 -2.65702695e-01
-1.45564690e-01 -9.29078698e-01 -1.41063142e+00 -1.25226006e-01
-2.72068411e-01 4.82369751e-01 5.79182446e-01 9.38193500e-01
3.92463028e-01 8.55029643e-01 9.67233241e-01 6.02015257e-02
-2.27620646e-01 -9.34807062e-01 -4.30955440e-01 2.02997267e-01
1.42300010e-01 -9.35020298e-03 -1.03873670e-01 -4.65567270e-03] | [11.546422004699707, 6.730895519256592] |
778868bf-56b9-406c-ad3b-ec4b1976a018 | rb-ccr-radial-based-combined-cleaning-and | 2105.04009 | null | https://arxiv.org/abs/2105.04009v1 | https://arxiv.org/pdf/2105.04009v1.pdf | RB-CCR: Radial-Based Combined Cleaning and Resampling algorithm for imbalanced data classification | Real-world classification domains, such as medicine, health and safety, and finance, often exhibit imbalanced class priors and have asynchronous misclassification costs. In such cases, the classification model must achieve a high recall without significantly impacting precision. Resampling the training data is the standard approach to improving classification performance on imbalanced binary data. However, the state-of-the-art methods ignore the local joint distribution of the data or correct it as a post-processing step. This can causes sub-optimal shifts in the training distribution, particularly when the target data distribution is complex. In this paper, we propose Radial-Based Combined Cleaning and Resampling (RB-CCR). RB-CCR utilizes the concept of class potential to refine the energy-based resampling approach of CCR. In particular, RB-CCR exploits the class potential to accurately locate sub-regions of the data-space for synthetic oversampling. The category sub-region for oversampling can be specified as an input parameter to meet domain-specific needs or be automatically selected via cross-validation. Our $5\times2$ cross-validated results on 57 benchmark binary datasets with 9 classifiers show that RB-CCR achieves a better precision-recall trade-off than CCR and generally out-performs the state-of-the-art resampling methods in terms of AUC and G-mean. | ['Michał Woźniak', 'Colin Bellinger', 'Michał Koziarski'] | 2021-05-09 | null | null | null | null | ['classification'] | ['methodology'] | [ 1.80880576e-01 -1.49386570e-01 -6.09786451e-01 -5.71842492e-01
-9.85668898e-01 -3.37184876e-01 3.20947707e-01 6.68169439e-01
-2.88781703e-01 1.14274764e+00 -1.71584114e-01 -4.55712348e-01
-2.80916393e-01 -1.05542660e+00 -7.06254184e-01 -7.10096061e-01
2.14397073e-01 6.23879552e-01 1.38674006e-01 1.39471799e-01
3.20444405e-01 4.63894844e-01 -1.64666867e+00 5.11877120e-01
1.40268397e+00 9.00828481e-01 -1.94785580e-01 2.23687381e-01
-9.68199596e-02 3.07577819e-01 -7.39202976e-01 -2.04795033e-01
1.86114579e-01 -4.79805976e-01 -5.27751863e-01 -2.64851004e-01
1.30208999e-01 8.26714039e-02 2.84862697e-01 1.06543195e+00
4.74853009e-01 -2.72780284e-02 8.32685828e-01 -1.06659627e+00
-2.95848697e-01 4.19913650e-01 -7.90916085e-01 1.61566064e-01
-3.33399288e-02 1.17266446e-01 5.87295651e-01 -6.86597288e-01
4.05362219e-01 9.38282847e-01 8.96261871e-01 3.48298460e-01
-1.47275066e+00 -8.71674597e-01 3.81313972e-02 8.78194347e-02
-1.51622367e+00 -1.74736619e-01 6.70101464e-01 -5.17952561e-01
8.18246305e-01 4.82745081e-01 6.04416668e-01 5.39554596e-01
4.87474144e-01 3.54841322e-01 1.12020910e+00 -4.19064164e-01
6.72457576e-01 2.29310825e-01 2.60465592e-01 2.56129503e-01
8.25711429e-01 1.54866382e-01 -3.30336750e-01 -5.18083632e-01
2.40691260e-01 2.17229292e-01 -1.81722507e-01 -3.05053443e-01
-7.74691522e-01 7.66245663e-01 3.79124403e-01 9.81432199e-02
-4.94188994e-01 -1.59012705e-01 3.71449798e-01 -5.00067696e-02
8.59799862e-01 6.59003854e-01 -6.27800763e-01 1.37224542e-02
-1.17283905e+00 4.56394076e-01 5.18172503e-01 4.82235402e-01
5.95199764e-01 -3.08072686e-01 -3.86466324e-01 1.04795742e+00
-2.64132451e-02 4.15898621e-01 4.71021235e-01 -4.24451977e-01
5.24523377e-01 8.17404032e-01 1.66395247e-01 -6.10400617e-01
-4.30786699e-01 -7.65993595e-01 -9.40302670e-01 1.25778064e-01
4.15439725e-01 -1.01962313e-02 -1.16698539e+00 1.47799575e+00
7.20152259e-01 -1.05368160e-01 -4.44899015e-02 6.82542622e-01
3.90960574e-01 3.72137934e-01 3.19962293e-01 -3.34659517e-01
1.37072885e+00 -4.07432020e-01 -5.75088441e-01 -2.05888167e-01
4.98117983e-01 -4.77334112e-01 9.57321286e-01 4.38906312e-01
-9.08888757e-01 -4.35734451e-01 -1.20064008e+00 2.31778577e-01
-3.50072980e-01 1.43040463e-01 4.77783203e-01 8.22264433e-01
-1.71958193e-01 5.93825579e-01 -8.69744480e-01 -2.49127634e-02
7.84935951e-01 2.54123449e-01 -1.04024850e-01 -2.72445947e-01
-1.12416613e+00 8.36646020e-01 2.89911926e-01 -2.04062536e-01
-4.37202513e-01 -1.42427969e+00 -7.18000233e-01 5.74593320e-02
1.92989528e-01 -6.90143168e-01 1.01064646e+00 -7.38560855e-01
-1.04254150e+00 7.48499751e-01 -2.45498329e-01 -6.02552354e-01
8.19821537e-01 -2.26819031e-02 -3.28492045e-01 -2.97469478e-02
2.79480796e-02 3.05606127e-01 6.84694350e-01 -1.21254671e+00
-6.37212276e-01 -5.83104968e-01 -2.70038575e-01 5.01557477e-02
3.83529812e-02 -3.91442537e-01 3.23892273e-02 -7.51817346e-01
2.17626885e-01 -6.88282311e-01 -3.20521235e-01 -3.81006271e-01
-3.52289140e-01 1.07059315e-01 6.54549420e-01 -4.67707098e-01
1.34274089e+00 -2.05385470e+00 -3.62083107e-01 3.02326769e-01
4.55233408e-03 3.17799956e-01 4.50170398e-01 1.87110797e-01
-3.56666207e-01 3.13924015e-01 -5.36105514e-01 -9.09416098e-03
-3.47117007e-01 -1.38206810e-01 -1.35237902e-01 7.33645916e-01
3.76564771e-01 4.55111116e-01 -9.40720260e-01 -2.34856695e-01
3.75523955e-01 5.45523703e-01 -6.78718150e-01 1.20307850e-02
-2.53310263e-01 2.82502264e-01 -2.58927435e-01 6.42212987e-01
1.05635583e+00 -3.29013497e-01 2.90814012e-01 -3.31907302e-01
2.99145907e-01 3.69168460e-01 -1.27197337e+00 1.08676720e+00
-5.01676142e-01 1.20156728e-01 -2.97993153e-01 -1.02758634e+00
9.60931778e-01 -9.98003706e-02 6.51267886e-01 -7.07602143e-01
-5.85229993e-02 3.10593396e-01 1.03253527e-02 -2.51169831e-01
4.21593219e-01 -4.59707826e-01 -1.43610248e-02 3.23100626e-01
-4.23870504e-01 -2.21035779e-01 2.50760559e-02 -2.70833552e-01
9.53831553e-01 -1.14433905e-02 6.22268200e-01 -5.70252895e-01
2.90117592e-01 2.38835320e-01 8.67256224e-01 5.79996884e-01
1.64326951e-02 8.84454727e-01 5.27943194e-01 -3.42787027e-01
-9.62653756e-01 -8.68073702e-01 -7.66213119e-01 5.22020221e-01
1.49328306e-01 -6.94335997e-02 -8.41203630e-01 -8.43611240e-01
3.55416238e-01 1.10532117e+00 -6.41113400e-01 -4.74663109e-01
-3.46882075e-01 -1.55242538e+00 3.35614085e-01 4.21588451e-01
4.43017811e-01 -6.68161035e-01 -8.10146809e-01 2.23201543e-01
-1.69352084e-01 -7.77228236e-01 -2.06574157e-01 4.88031089e-01
-9.54284549e-01 -1.35029101e+00 -3.12048376e-01 -2.11940289e-01
1.03844881e+00 -5.20752668e-02 1.16950786e+00 1.07942596e-01
-4.67316002e-01 -3.66334558e-01 -3.00025940e-01 -6.82536125e-01
-4.38492656e-01 2.23021656e-01 -8.90624151e-02 -1.70988634e-01
5.66599846e-01 -2.69661725e-01 -8.69342089e-01 4.70619857e-01
-1.00864065e+00 -6.85112476e-02 2.92882651e-01 1.04789138e+00
9.74919975e-01 4.29429591e-01 1.02106702e+00 -1.20414531e+00
4.73429978e-01 -6.40789032e-01 -6.10659957e-01 2.32509285e-01
-9.32942569e-01 8.48302469e-02 6.62795663e-01 -4.56386387e-01
-8.85114372e-01 4.85655526e-03 -6.61708415e-02 -2.38327280e-01
3.50988284e-02 4.35509056e-01 -3.20084065e-01 3.74994248e-01
9.08062339e-01 -2.68582165e-01 -4.53452393e-02 -4.55230027e-01
-1.81441069e-01 8.50407779e-01 1.88086033e-01 -5.88300645e-01
3.38288426e-01 5.10150254e-01 1.37367114e-01 -4.95514303e-01
-7.52033412e-01 -3.50592285e-01 -3.32637191e-01 6.75161332e-02
5.42209327e-01 -8.44870985e-01 -4.92034405e-01 5.56676626e-01
-6.22662961e-01 -3.68714869e-01 -6.56127036e-01 4.11701828e-01
-2.78766543e-01 -7.55180418e-02 -8.76052678e-02 -9.00118530e-01
-4.62245643e-01 -1.21747398e+00 1.13358164e+00 3.35662335e-01
-2.91538835e-01 -7.22565711e-01 -1.49808824e-01 3.29738379e-01
3.58894080e-01 4.72189754e-01 1.12889159e+00 -7.93774486e-01
-1.54977322e-01 -4.42074478e-01 -8.47142637e-02 3.40478569e-01
3.90653104e-01 5.47264554e-02 -1.13815928e+00 -2.68936127e-01
-8.80933553e-02 6.12393655e-02 9.11187351e-01 7.51446724e-01
1.75444806e+00 -5.76647045e-03 -6.18233860e-01 5.03045619e-01
1.47208953e+00 2.37214819e-01 7.19330072e-01 2.10147157e-01
2.45088607e-01 4.42581445e-01 1.09114230e+00 5.89657068e-01
2.91641206e-01 6.38944387e-01 3.62352222e-01 -1.75988629e-01
-1.02406442e-02 -2.72582442e-01 -9.60495025e-02 -2.46510040e-02
2.91825444e-01 -1.72721386e-01 -1.00040781e+00 6.08557940e-01
-1.61734748e+00 -7.85831392e-01 -4.54969145e-02 2.72210860e+00
1.20168948e+00 3.56684417e-01 1.28669664e-01 5.96913338e-01
9.10543501e-01 -1.93961114e-01 -7.76300967e-01 -2.34838188e-01
1.98306739e-01 2.74506927e-01 8.18992496e-01 4.79977548e-01
-9.63801742e-01 3.53056192e-01 6.00192070e+00 9.38486755e-01
-1.17390978e+00 2.63005476e-02 1.20466328e+00 -2.50948161e-01
-3.12406898e-01 -1.78727746e-01 -1.01479125e+00 7.55280674e-01
9.31542873e-01 -1.41341180e-01 1.54072642e-01 7.05338538e-01
2.03740761e-01 -5.17082572e-01 -1.07507753e+00 7.58029461e-01
-2.07683757e-01 -1.26862633e+00 -2.83082962e-01 1.25408113e-01
6.89376950e-01 -2.53523678e-01 -1.06844909e-01 3.02794814e-01
2.27483854e-01 -1.10674632e+00 5.97133994e-01 4.72083271e-01
1.13882291e+00 -9.50169444e-01 9.91223454e-01 3.70198727e-01
-8.43292177e-01 -1.50652215e-01 -2.43068591e-01 1.96831137e-01
-1.38382446e-02 1.62035668e+00 -1.01517940e+00 6.83575809e-01
7.63974190e-01 3.35859090e-01 -4.81247276e-01 1.11244559e+00
2.25318268e-01 6.46653354e-01 -4.43993121e-01 5.20796068e-02
-4.90949780e-01 -3.86952460e-02 2.78573126e-01 1.07997990e+00
2.68728316e-01 3.36205848e-02 -3.44617255e-02 7.33418226e-01
-1.40842110e-01 -2.88165035e-03 -1.39781013e-01 2.60577321e-01
8.35327029e-01 7.68631220e-01 -8.48360121e-01 -4.83027428e-01
1.66401014e-01 1.62522629e-01 1.95455179e-02 1.67245924e-01
-1.02084887e+00 -4.85869914e-01 7.72088230e-01 5.64162254e-01
2.22048908e-01 3.72246772e-01 -8.79470944e-01 -9.61897671e-01
1.42656446e-01 -1.03170335e+00 4.80780721e-01 -4.06165123e-01
-1.30931580e+00 3.16999525e-01 2.89680094e-01 -1.32611680e+00
-4.75616641e-02 -2.26506472e-01 -1.70430884e-01 8.88996303e-01
-1.55334401e+00 -7.00448215e-01 -5.57984471e-01 4.78549339e-02
4.20069665e-01 1.20310828e-01 6.38451397e-01 3.87870967e-01
-5.13013721e-01 7.16848552e-01 1.60799757e-01 -2.13617712e-01
7.76992083e-01 -1.10992765e+00 2.20644325e-01 6.07790947e-01
-5.11518955e-01 3.64408135e-01 9.18779135e-01 -9.22773838e-01
-9.72108364e-01 -1.34903264e+00 5.74034274e-01 -2.62980908e-01
8.92751515e-02 -1.97757050e-01 -1.07478535e+00 2.47603923e-01
-4.43197012e-01 3.65740955e-01 6.73859417e-01 7.28849024e-02
-2.33738631e-01 -5.89418411e-01 -1.95189416e+00 2.06602797e-01
8.17109704e-01 -1.61005497e-01 -3.88659000e-01 3.20166022e-01
4.08493012e-01 -6.84949338e-01 -9.70918536e-01 9.20638621e-01
5.05781889e-01 -9.55528498e-01 1.05281842e+00 -5.93169272e-01
4.06268626e-01 -4.24234897e-01 -1.97987512e-01 -1.51149213e+00
-4.84154522e-02 -6.42612651e-02 3.65112945e-02 1.21807766e+00
4.15532142e-01 -8.52414846e-01 7.20419168e-01 5.37485301e-01
1.25242665e-01 -1.11800838e+00 -9.48182940e-01 -5.92468858e-01
1.44755527e-01 -4.19160426e-01 1.14921188e+00 9.13630724e-01
-1.37421757e-01 -1.62715137e-01 1.25134379e-01 1.04501553e-01
6.46287203e-01 1.36155754e-01 5.82522452e-01 -1.24262536e+00
-1.62191331e-01 -3.67783427e-01 -2.40548015e-01 -3.33313376e-01
-1.01371959e-01 -6.99848354e-01 -4.26864438e-02 -1.43532515e+00
2.12612316e-01 -1.04716194e+00 -3.11041355e-01 5.16590416e-01
-4.99235302e-01 2.24364102e-01 -2.27534622e-01 -5.78588732e-02
-8.09645206e-02 3.53597939e-01 9.40775692e-01 -2.16371626e-01
-4.45833445e-01 1.58273280e-01 -8.78958881e-01 4.88114744e-01
9.65037227e-01 -7.96587050e-01 -3.96788538e-01 1.31308213e-01
1.62914574e-01 -6.13043420e-02 2.18760803e-01 -9.21313941e-01
-2.20908359e-01 -4.72111136e-01 7.07396448e-01 -7.03662992e-01
-7.42034465e-02 -8.20756733e-01 4.58487868e-01 7.01260626e-01
-3.06242675e-01 -4.74719368e-02 2.88947999e-01 4.98648107e-01
-8.87377933e-02 -2.20275745e-01 1.24115527e+00 1.58579826e-01
1.51332945e-01 -5.35658048e-03 5.76910097e-03 3.00926715e-01
1.15497708e+00 -1.24948889e-01 -5.01112103e-01 1.52768806e-01
-3.02551359e-01 1.22095026e-01 6.21520579e-01 2.68699914e-01
3.58626604e-01 -1.11556542e+00 -6.85575306e-01 4.35323179e-01
2.44684443e-01 4.86381412e-01 3.36526304e-01 7.55632818e-01
-4.95434850e-01 9.22557861e-02 5.80615848e-02 -7.90196359e-01
-1.19574153e+00 4.18921649e-01 6.86555505e-01 -4.82224226e-01
-4.40783799e-01 4.75056291e-01 -1.03591353e-01 -4.08057988e-01
-4.99888416e-03 -6.35004044e-01 6.76201135e-02 2.01029867e-01
5.50250769e-01 5.21021307e-01 8.22887897e-01 -1.39468368e-02
-5.86984634e-01 2.78919280e-01 2.72119343e-02 2.58762270e-01
1.33239627e+00 2.73177326e-01 -1.74395239e-03 3.60328406e-01
8.91639352e-01 7.93925971e-02 -1.07331932e+00 1.68310687e-01
-2.06162125e-01 -7.00515389e-01 9.92607176e-02 -1.03396559e+00
-1.06585801e+00 5.18230855e-01 6.13849282e-01 1.86287373e-01
1.28387785e+00 -3.02603126e-01 4.36591774e-01 -2.68460304e-01
3.71518075e-01 -1.21656156e+00 -3.38105530e-01 4.42131907e-02
5.68545461e-01 -1.20224643e+00 4.79880691e-01 -4.28293049e-01
-5.82629442e-01 7.00378954e-01 6.14777744e-01 -3.55774499e-02
8.80032837e-01 3.58766913e-01 -1.77706644e-01 1.36868004e-02
-7.36021817e-01 3.14035118e-01 2.69625604e-01 3.94087285e-01
3.48214686e-01 3.13962251e-01 -5.81704319e-01 7.40677118e-01
-2.80930310e-01 2.64857471e-01 3.02132219e-01 9.69879925e-01
-2.08585680e-01 -9.66656446e-01 -7.09508836e-01 1.01385832e+00
-6.79037273e-01 7.38819018e-02 7.20386729e-02 6.44542873e-01
3.61747086e-01 8.42234433e-01 2.74120659e-01 -2.90286213e-01
4.71627235e-01 2.11930454e-01 4.63012159e-01 -4.96836334e-01
-7.04250515e-01 -1.81313083e-01 7.65949860e-02 -2.50044882e-01
-2.38477841e-01 -8.25008214e-01 -1.34422874e+00 -3.17350060e-01
-5.47089875e-01 2.38059506e-01 8.59932363e-01 8.78187180e-01
5.67351162e-01 8.41490626e-01 5.39693296e-01 -5.71198881e-01
-7.04655349e-01 -9.44459200e-01 -3.86004478e-01 4.11879480e-01
5.39521813e-01 -9.30053711e-01 -4.78148371e-01 -1.40577838e-01] | [8.76247501373291, 4.2169508934021] |
03dc263f-de01-4f24-8819-957db3b8d620 | a-sliced-wasserstein-distance-based-approach | 2302.01459 | null | https://arxiv.org/abs/2302.01459v1 | https://arxiv.org/pdf/2302.01459v1.pdf | A sliced-Wasserstein distance-based approach for out-of-class-distribution detection | There exist growing interests in intelligent systems for numerous medical imaging, image processing, and computer vision applications, such as face recognition, medical diagnosis, character recognition, and self-driving cars, among others. These applications usually require solving complex classification problems involving complex images with unknown data generative processes. In addition to recent successes of the current classification approaches relying on feature engineering and deep learning, several shortcomings of them, such as the lack of robustness, generalizability, and interpretability, have also been observed. These methods often require extensive training data, are computationally expensive, and are vulnerable to out-of-distribution samples, e.g., adversarial attacks. Recently, an accurate, data-efficient, computationally efficient, and robust transport-based classification approach has been proposed, which describes a generative model-based problem formulation and closed-form solution for a specific category of classification problems. However, all these approaches lack mechanisms to detect test samples outside the class distributions used during training. In real-world settings, where the collected training samples are unable to exhaust or cover all classes, the traditional classification schemes are unable to handle the unseen classes effectively, which is especially an important issue for safety-critical systems, such as self-driving and medical imaging diagnosis. In this work, we propose a method for detecting out-of-class distributions based on the distribution of sliced-Wasserstein distance from the Radon Cumulative Distribution Transform (R-CDT) subspace. We tested our method on the MNIST and two medical image datasets and reported better accuracy than the state-of-the-art methods without an out-of-class distribution detection procedure. | ['Gustavo K Rohde', 'Yan Zhuang', 'Abu Hasnat Mohammad Rubaiyat', 'Mohammad Shifat E Rabbi'] | 2023-02-02 | null | null | null | null | ['medical-diagnosis', 'feature-engineering'] | ['medical', 'methodology'] | [ 3.87982011e-01 -2.20518529e-01 4.29538451e-02 -2.85589010e-01
-6.37827814e-01 -3.56766939e-01 3.91868979e-01 4.05065566e-02
-2.26794958e-01 7.22571850e-01 -5.98930359e-01 -3.96103233e-01
-3.15615624e-01 -8.69643569e-01 -5.06654859e-01 -1.10207129e+00
1.74710333e-01 5.27665317e-01 2.26395577e-01 9.67952162e-02
7.18502626e-02 7.29577065e-01 -1.53200221e+00 5.90685606e-02
1.00932074e+00 1.13898790e+00 -8.30150321e-02 3.51735204e-01
9.18702036e-02 3.60080004e-01 -5.71732879e-01 -1.14356443e-01
7.06732646e-02 -4.75733250e-01 -3.16671103e-01 4.42781508e-01
8.41938853e-02 -2.76841164e-01 -4.34827864e-01 1.15502310e+00
5.43434739e-01 8.32621939e-03 8.66836250e-01 -1.32689798e+00
-7.74043202e-01 -2.10497454e-01 -5.36448479e-01 1.68410450e-01
-8.40695277e-02 2.99115658e-01 2.15259269e-01 -6.90048695e-01
3.03055257e-01 8.72915566e-01 5.09559691e-01 7.40973830e-01
-1.11638963e+00 -8.51931930e-01 -2.76080668e-01 1.61880270e-01
-1.28473473e+00 -1.99183837e-01 8.09283853e-01 -5.78359604e-01
2.94316411e-01 3.50391179e-01 3.56899083e-01 1.10527539e+00
4.54102665e-01 6.29369318e-01 1.21221685e+00 -1.85409710e-01
4.17677879e-01 3.53524715e-01 -9.70494747e-02 7.48040378e-01
3.21239620e-01 2.51087099e-01 4.99335863e-02 -2.48684332e-01
7.70189464e-01 3.31131339e-01 -2.40849406e-01 -3.85519087e-01
-8.98037076e-01 8.82228196e-01 2.89396048e-01 3.23013067e-01
-2.94230014e-01 -1.69349074e-01 1.98163882e-01 -1.24926008e-01
5.96918344e-01 7.92616457e-02 -1.90470725e-01 2.44760618e-01
-7.36957192e-01 -8.25160518e-02 4.95433092e-01 7.05170453e-01
6.25470340e-01 1.43648446e-01 -5.42469434e-02 6.41700804e-01
1.46618471e-01 7.68442154e-01 6.14211738e-01 -4.18515861e-01
2.16053292e-01 3.87356013e-01 -2.10614711e-01 -1.07882035e+00
-3.57768804e-01 -7.54806638e-01 -1.28852570e+00 2.40815401e-01
3.99214238e-01 -4.32228930e-02 -9.64755416e-01 1.47310841e+00
6.51726782e-01 3.28363448e-01 2.02250078e-01 7.87822723e-01
6.01822019e-01 4.56784040e-01 -6.77998289e-02 -1.79198086e-01
1.09646142e+00 -5.19182861e-01 -4.87172544e-01 -1.08688354e-01
3.85691315e-01 -5.97777426e-01 9.01153862e-01 4.47142214e-01
-6.13073230e-01 -6.30591333e-01 -9.79739606e-01 4.44841951e-01
-2.52368718e-01 4.45613302e-02 5.58382213e-01 9.39496815e-01
-5.01355112e-01 5.82399845e-01 -8.79993498e-01 -1.90103710e-01
8.78660262e-01 2.87098020e-01 -3.39349836e-01 -3.57810438e-01
-9.20672238e-01 5.99076033e-01 5.75837940e-02 1.86606735e-01
-1.10850585e+00 -6.22388601e-01 -8.41428816e-01 -1.89603567e-01
2.56084532e-01 -5.78988492e-01 7.56673157e-01 -8.86357486e-01
-1.31338000e+00 6.80640996e-01 3.04216668e-02 -1.70086235e-01
6.21559322e-01 1.38244420e-01 -6.22422457e-01 1.46388680e-01
5.05075902e-02 2.36761257e-01 1.10214353e+00 -1.14228344e+00
-3.41629207e-01 -7.06563771e-01 -2.94283807e-01 -1.64558157e-01
-3.50114167e-01 -2.60061443e-01 3.14078107e-02 -4.90430325e-01
2.97070831e-01 -9.39553380e-01 -2.97329932e-01 1.06921464e-01
-6.04279280e-01 5.04737534e-02 1.19773078e+00 -4.20547068e-01
7.63452590e-01 -2.31501889e+00 -2.66471326e-01 2.54060715e-01
1.84105486e-01 3.75059903e-01 6.67738616e-02 1.56030450e-02
-1.90212391e-02 1.69581443e-01 -5.36084771e-01 -2.94284131e-02
-3.29104125e-01 2.41044492e-01 -3.29237968e-01 8.67183924e-01
4.09804553e-01 6.80838227e-01 -9.34951186e-01 -3.31895530e-01
4.14305538e-01 6.02869391e-01 -2.23442331e-01 7.95913190e-02
-1.69774331e-02 8.99454772e-01 -6.92500174e-01 6.87410355e-01
7.66988873e-01 -1.59559593e-01 -2.48739779e-01 -9.35500190e-02
3.33344251e-01 -4.05681252e-01 -9.46683109e-01 1.14089549e+00
-3.84627253e-01 3.57986093e-01 -1.44753769e-01 -1.41003585e+00
8.86773169e-01 2.12525770e-01 6.82318032e-01 -6.36185110e-01
1.87813580e-01 3.94224435e-01 1.87876493e-01 -7.53486454e-01
4.88369018e-02 -4.78102952e-01 3.12040113e-02 2.90156662e-01
-1.29896700e-01 -1.08749963e-01 -9.69655961e-02 -2.17762783e-01
1.02440333e+00 -1.13861039e-01 1.09565303e-01 -2.12689310e-01
6.28149569e-01 -1.86724395e-01 4.37845558e-01 5.29985011e-01
-2.98009545e-01 8.35665882e-01 1.83655500e-01 -3.83114010e-01
-7.91358471e-01 -1.05073929e+00 -5.34847677e-01 2.49668822e-01
3.25132251e-01 3.57154489e-01 -8.23547840e-01 -8.77131104e-01
-6.37660697e-02 6.15771413e-01 -4.11555260e-01 -4.95354444e-01
-3.21448892e-01 -1.12415206e+00 6.05050325e-01 4.06660050e-01
5.81088245e-01 -8.74587715e-01 -5.61773896e-01 2.11282313e-01
-4.61128913e-02 -1.30040884e+00 -2.08678186e-01 -1.33410692e-01
-9.15051222e-01 -1.31932104e+00 -6.85126185e-01 -5.55127501e-01
1.06764543e+00 1.41786365e-02 5.96945703e-01 1.00114278e-01
-7.79540718e-01 3.11797440e-01 -6.76926076e-02 -5.64745128e-01
-6.44550860e-01 -1.47967234e-01 3.80988568e-02 4.26027596e-01
1.94517687e-01 -3.08140159e-01 -7.15876758e-01 6.76613092e-01
-1.17351711e+00 -3.28396440e-01 6.10006452e-01 1.08364570e+00
7.07308531e-01 5.48220932e-01 9.57688868e-01 -9.41155314e-01
3.96154642e-01 -6.18631661e-01 -4.46297735e-01 9.80085209e-02
-5.62746227e-01 -8.23596865e-02 7.83663630e-01 -6.46302640e-01
-8.13964486e-01 -3.83778661e-02 -3.52370560e-01 -5.30491769e-01
-4.76509750e-01 2.37827718e-01 -3.38462681e-01 -1.45390168e-01
7.26795852e-01 4.76577342e-01 3.97513181e-01 2.05242690e-02
-1.32148713e-01 7.40331590e-01 4.75450754e-01 -4.29387718e-01
7.93388665e-01 8.35104465e-01 3.46010447e-01 -1.07994986e+00
-7.51749873e-01 -4.05104727e-01 -3.53970915e-01 -2.57576376e-01
8.30686092e-01 -4.56938952e-01 -3.68759573e-01 8.82979453e-01
-9.24556136e-01 -7.08867386e-02 -3.37859601e-01 5.93461156e-01
-4.66690063e-01 6.17174029e-01 -3.13464642e-01 -8.98228943e-01
-3.28480035e-01 -1.40638006e+00 1.03901863e+00 3.82561177e-01
1.62807271e-01 -1.15251291e+00 -3.74179095e-01 5.25588214e-01
6.24448955e-01 5.64520121e-01 1.13217425e+00 -6.27190888e-01
-4.96010780e-01 -5.29080689e-01 6.28394261e-02 6.80057168e-01
4.14512336e-01 -1.20594382e-01 -1.07803786e+00 -3.99764776e-01
2.86265016e-01 -2.00723067e-01 5.93808949e-01 4.08864051e-01
1.58776104e+00 -1.66545138e-01 -4.93661612e-01 5.73677897e-01
1.14901674e+00 4.22174871e-01 8.16080570e-01 -6.59730434e-02
4.53865021e-01 5.67795694e-01 5.31323016e-01 2.54813135e-01
-2.22885720e-02 3.24903429e-01 4.99884307e-01 -2.92463779e-01
-1.81443226e-02 6.73710480e-02 2.02175975e-01 3.79044771e-01
2.29030892e-01 -3.82338285e-01 -6.20169222e-01 4.31446284e-01
-1.51907170e+00 -9.77426887e-01 3.40338144e-03 2.43650126e+00
5.54302037e-01 1.39323950e-01 -1.62417203e-01 4.09033507e-01
9.22156334e-01 -2.93765336e-01 -9.33323741e-01 8.60584900e-02
2.81496383e-02 2.76054144e-01 3.69575024e-01 5.83487423e-03
-1.19218230e+00 4.22193468e-01 5.60980415e+00 8.84573519e-01
-1.50102913e+00 2.20066056e-01 1.07608294e+00 3.30906063e-01
-1.70574009e-01 -2.78303415e-01 -6.06188774e-01 6.42861247e-01
5.59243500e-01 2.40895137e-01 1.04542658e-01 7.56668568e-01
5.81815913e-02 -2.32450098e-01 -1.01763391e+00 1.12309599e+00
2.39077404e-01 -9.88646567e-01 -2.56432921e-01 2.78132021e-01
5.97602427e-01 -1.82932153e-01 3.22615534e-01 1.79904297e-01
-2.71772861e-01 -1.07010877e+00 4.60326970e-01 2.66288906e-01
9.39151227e-01 -6.61482275e-01 7.81186461e-01 6.81774497e-01
-5.82107246e-01 -7.67519847e-02 -3.97741228e-01 3.34039629e-01
9.56334770e-02 1.08044302e+00 -6.58363223e-01 5.07442713e-01
5.10503113e-01 3.70944917e-01 -3.11262190e-01 1.01080501e+00
-1.88783705e-01 6.78825200e-01 -3.57474208e-01 -2.31061950e-02
1.43439658e-02 -1.91395342e-01 5.21546125e-01 7.44752049e-01
5.90890586e-01 8.21286365e-02 1.94642991e-01 9.39374030e-01
6.77306950e-02 -1.65479779e-01 -6.08802259e-01 -1.26793580e-02
4.87471335e-02 1.37154627e+00 -1.00708890e+00 -7.71369636e-02
-1.46053031e-01 7.61611700e-01 -2.73169309e-01 3.57489824e-01
-9.64555979e-01 -2.78850764e-01 4.42333877e-01 4.75018948e-01
8.52594599e-02 -2.25847423e-01 -2.21048117e-01 -1.04005873e+00
1.11298181e-01 -6.99216902e-01 2.89102674e-01 -3.96612763e-01
-1.48486710e+00 6.36565864e-01 1.04299620e-01 -1.26891756e+00
-2.53300518e-01 -6.63286448e-01 -6.49686635e-01 6.28224134e-01
-1.45908844e+00 -1.05714428e+00 -4.42017019e-01 6.90654874e-01
4.07989562e-01 -3.43639821e-01 7.28845358e-01 5.87969780e-01
-5.61341524e-01 4.97920960e-01 2.85521150e-01 2.47106507e-01
4.22038943e-01 -7.55495965e-01 -9.47982818e-02 8.31696749e-01
-4.32739500e-03 2.48945817e-01 5.37700295e-01 -5.27607024e-01
-1.27462959e+00 -1.32959020e+00 1.94416359e-01 -1.58332646e-01
4.61708963e-01 -3.34078044e-01 -9.36160326e-01 2.52623856e-01
-2.83079416e-01 6.45572722e-01 6.39080167e-01 -3.72933626e-01
1.20197916e-02 -3.25114161e-01 -1.62638676e+00 2.46081099e-01
6.30303502e-01 -3.04907352e-01 -2.11134553e-01 4.52383280e-01
4.09975536e-02 -4.58548427e-01 -5.85285068e-01 5.81867158e-01
4.10817832e-01 -8.37285399e-01 8.96515012e-01 -4.10525084e-01
1.90250069e-01 -3.24087977e-01 -2.42852252e-02 -1.28413713e+00
5.20621650e-02 -3.25708449e-01 6.34218529e-02 1.09700048e+00
2.23941058e-01 -1.03173876e+00 7.61956215e-01 4.47353065e-01
-1.32417068e-01 -9.61207092e-01 -1.14058435e+00 -8.56843054e-01
6.41565099e-02 -4.82663006e-01 4.73618835e-01 7.95530021e-01
-5.23736238e-01 7.35631287e-02 -1.16051257e-01 4.21166599e-01
8.38639379e-01 3.77186079e-04 5.75239122e-01 -1.22408879e+00
-2.81723350e-01 -1.65620834e-01 -8.12842369e-01 -5.67552269e-01
2.00474322e-01 -8.55090976e-01 1.46695018e-01 -1.51872230e+00
1.09715022e-01 -8.74214888e-01 -2.37436548e-01 3.01957220e-01
-9.36790183e-02 2.70773530e-01 -2.45110720e-01 2.47058168e-01
-1.55137464e-01 4.50156718e-01 1.35309589e+00 -4.44531679e-01
1.65070787e-01 3.80845815e-01 -5.29068708e-01 7.20326304e-01
8.57983828e-01 -6.50818765e-01 -6.78596914e-01 -1.33875042e-01
-2.96605349e-01 -1.89855937e-02 7.19575703e-01 -1.26436484e+00
9.29084495e-02 -2.16808796e-01 5.49462318e-01 -5.10495842e-01
2.81709194e-01 -8.65304828e-01 2.82638580e-01 6.73992693e-01
1.06632829e-01 -2.62697935e-01 3.50423083e-02 5.79126835e-01
-2.05799922e-01 -2.87955880e-01 1.13407385e+00 -2.94952672e-02
-4.34721261e-01 6.23351872e-01 -2.94596672e-01 -6.31265119e-02
1.52152336e+00 -3.05173606e-01 -4.30860281e-01 -2.51943052e-01
-5.31659901e-01 -1.77461877e-01 1.69164091e-01 3.22933793e-01
8.60161126e-01 -1.10703552e+00 -7.65748620e-01 6.64907813e-01
1.29045144e-01 4.40671682e-01 5.51605582e-01 9.68790531e-01
-4.26835388e-01 8.27773511e-02 1.87801160e-02 -9.76117671e-01
-8.89436901e-01 6.76273763e-01 4.78971869e-01 -1.24086991e-01
-5.81282854e-01 3.55487764e-01 5.44026077e-01 -4.28940564e-01
-1.22919120e-01 -3.50347340e-01 1.04749456e-01 -3.13471973e-01
4.17191207e-01 3.45789939e-01 3.15664709e-01 -5.98195493e-01
-3.29596937e-01 5.28935492e-01 1.16777308e-01 2.35957563e-01
1.07393277e+00 8.42246860e-02 1.77949220e-01 1.56250313e-01
1.10860872e+00 -2.63425320e-01 -1.11029637e+00 -3.95153314e-02
-2.55235106e-01 -4.53343302e-01 5.16809672e-02 -5.71189404e-01
-1.39172089e+00 1.22671604e+00 8.13516498e-01 4.12987232e-01
1.07595634e+00 6.47479743e-02 9.27196980e-01 3.28964293e-02
5.12628257e-01 -9.50966120e-01 2.35952869e-01 1.01122446e-01
6.36654198e-01 -1.22296727e+00 -9.21896622e-02 -5.09909630e-01
-4.86968160e-01 1.07666349e+00 5.66730857e-01 4.44864742e-02
8.69723558e-01 3.20559055e-01 5.96533753e-02 -1.08717382e-01
-2.67649084e-01 7.12435693e-02 2.96962172e-01 7.98515916e-01
5.72512634e-02 -6.48802333e-03 4.92797084e-02 3.72222900e-01
1.45510510e-01 7.81892464e-02 4.85195220e-01 8.92787457e-01
-2.00893208e-01 -7.75647163e-01 -5.42671084e-01 6.36129737e-01
-6.52603209e-01 2.20265388e-01 2.18205631e-01 7.18158007e-01
2.44944185e-01 8.92390907e-01 9.26651135e-02 -2.29051694e-01
2.24554569e-01 -1.12415947e-01 4.96969372e-01 -4.93433774e-01
2.53493246e-02 -7.51649216e-02 -4.55903649e-01 -1.28788009e-01
-4.41146314e-01 -6.07819080e-01 -1.30016255e+00 7.90513121e-03
-6.42888606e-01 -5.33204004e-02 8.24865639e-01 1.11008096e+00
2.99540013e-01 5.23402214e-01 9.19128776e-01 -5.37378669e-01
-8.85367334e-01 -7.05319881e-01 -6.76332116e-01 4.69732881e-01
3.39330733e-01 -7.79483736e-01 -4.14944857e-01 -1.29786521e-01] | [7.585346698760986, 1.9625024795532227] |
d86128cb-77ba-4bff-8245-afd7b289fb4d | thompson-sampling-on-symmetric-stable-bandits | 1907.03821 | null | https://arxiv.org/abs/1907.03821v2 | https://arxiv.org/pdf/1907.03821v2.pdf | Thompson Sampling on Symmetric $α$-Stable Bandits | Thompson Sampling provides an efficient technique to introduce prior knowledge in the multi-armed bandit problem, along with providing remarkable empirical performance. In this paper, we revisit the Thompson Sampling algorithm under rewards drawn from symmetric $\alpha$-stable distributions, which are a class of heavy-tailed probability distributions utilized in finance and economics, in problems such as modeling stock prices and human behavior. We present an efficient framework for posterior inference, which leads to two algorithms for Thompson Sampling in this setting. We prove finite-time regret bounds for both algorithms, and demonstrate through a series of experiments the stronger performance of Thompson Sampling in this setting. With our results, we provide an exposition of symmetric $\alpha$-stable distributions in sequential decision-making, and enable sequential Bayesian inference in applications from diverse fields in finance and complex systems that operate on heavy-tailed features. | ['Alex Pentland', 'Abhimanyu Dubey'] | 2019-07-08 | null | null | null | null | ['sequential-bayesian-inference'] | ['time-series'] | [-8.67899358e-02 -2.01367915e-01 -7.62727022e-01 -4.32620078e-01
-1.02777565e+00 -5.10371149e-01 2.82722771e-01 -2.05973148e-01
-3.38946611e-01 1.27241480e+00 7.12929741e-02 -6.39369428e-01
-6.04075849e-01 -8.08962405e-01 -8.67166281e-01 -7.58246243e-01
-6.30808175e-02 8.08353364e-01 -2.69261807e-01 2.54588425e-01
4.14846599e-01 2.37563521e-01 -1.32980978e+00 -2.20137928e-02
3.77138048e-01 1.50847018e+00 3.86171453e-02 7.40231097e-01
-4.08774167e-02 8.68917704e-01 -2.52066880e-01 -8.91479671e-01
4.67207223e-01 -2.13817403e-01 -6.22256994e-01 -6.68974668e-02
-1.35625303e-01 -6.68558180e-01 -5.00759259e-02 1.18271148e+00
2.61700839e-01 3.90473634e-01 8.71907651e-01 -1.10575056e+00
-4.87109363e-01 1.13887942e+00 -1.03576005e+00 6.38636827e-01
1.84455171e-01 -1.88590348e-01 1.49959254e+00 -4.15411592e-01
1.74737200e-01 1.37386644e+00 5.69931090e-01 2.44036630e-01
-1.32088733e+00 -6.38836443e-01 4.62104917e-01 1.53771654e-01
-6.90513194e-01 -2.63569415e-01 6.47108257e-01 -4.79285061e-01
3.87004495e-01 4.23208624e-01 7.66460776e-01 1.31807435e+00
1.31182894e-01 1.52430832e+00 1.30280316e+00 -4.27883834e-01
6.99234009e-01 -5.72756352e-03 3.47275615e-01 2.45498791e-01
6.52270138e-01 3.54849339e-01 -7.56279945e-01 -7.54558921e-01
9.62062120e-01 3.79348576e-01 3.26991756e-03 -2.98437148e-01
-8.73952627e-01 1.32513797e+00 -2.27340892e-01 -3.15793753e-01
-7.18513429e-01 5.45274019e-01 1.07000761e-01 2.79249489e-01
7.41254389e-01 -2.30906904e-01 -5.60525537e-01 -3.79224211e-01
-1.06129122e+00 6.88155353e-01 1.01075649e+00 1.16663241e+00
4.91747648e-01 -4.06474695e-02 -8.90664518e-01 5.76866686e-01
4.72921968e-01 1.02803278e+00 1.71337262e-01 -1.10873520e+00
5.69017470e-01 -5.24621069e-01 1.07801366e+00 -4.53053385e-01
-5.91306351e-02 -7.11567640e-01 -6.21347606e-01 5.73332049e-02
7.58522093e-01 -4.47098911e-01 -5.47307730e-01 1.70264924e+00
3.87498856e-01 -3.39137092e-02 -3.61955673e-01 6.92969918e-01
-1.84738591e-01 3.55249286e-01 -1.85077667e-01 -7.22261667e-01
1.35688221e+00 -7.02928364e-01 -6.58183515e-01 -5.82825281e-02
1.05094820e-01 -6.98751330e-01 7.81585634e-01 7.33170271e-01
-1.39658022e+00 -2.05957871e-02 -4.34983343e-01 3.64953399e-01
2.66914427e-01 -2.78178006e-01 1.10898292e+00 9.79232848e-01
-5.50151825e-01 9.91511405e-01 -8.55446815e-01 1.41435012e-01
8.01993608e-01 -1.93410963e-01 5.81211865e-01 -4.23704553e-03
-8.21528852e-01 5.43665648e-01 -9.16330703e-03 -4.50121686e-02
-9.88008082e-01 -7.08595812e-01 -2.63782382e-01 2.40936086e-01
6.36495590e-01 -8.79280984e-01 1.98593831e+00 -7.34537125e-01
-1.71210861e+00 4.68604773e-01 -3.73090148e-01 -8.55226874e-01
1.19010746e+00 -4.37977344e-01 6.91709667e-02 -6.82159960e-02
3.28631222e-01 -3.37934047e-01 1.16818726e+00 -6.09284639e-01
-9.45156872e-01 -4.07309324e-01 1.35998605e-02 -8.51864442e-02
7.91601390e-02 -3.47170024e-03 8.66600648e-02 -8.67223024e-01
-1.44513503e-01 -7.41642773e-01 -5.29227614e-01 -3.27764332e-01
-5.15705407e-01 -2.34831586e-01 -1.51814327e-01 -2.53240764e-01
1.12488699e+00 -1.57718205e+00 -2.10916340e-01 5.08852065e-01
-4.31714617e-02 -4.53830987e-01 3.16115320e-01 3.75266850e-01
2.11836770e-01 4.15331274e-02 -4.44027781e-03 -4.02692586e-01
6.42540991e-01 1.10382035e-01 -7.93253005e-01 6.50994241e-01
-4.67734218e-01 1.02162313e+00 -9.05499876e-01 -1.64419413e-01
-1.25155076e-01 -4.81006980e-01 -4.66830254e-01 6.31876290e-02
-4.90855038e-01 1.41533554e-01 -9.60832953e-01 7.63011992e-01
6.98880613e-01 -6.47601187e-01 7.46660903e-02 3.95379752e-01
1.23009317e-01 7.98609015e-03 -1.30016351e+00 1.19094157e+00
-3.43951643e-01 4.60925847e-02 1.80138022e-01 -1.22604489e+00
5.28255999e-01 1.00489184e-01 4.15941983e-01 -3.23617011e-01
5.37286162e-01 1.97939619e-01 -4.51401800e-01 -2.25651979e-01
4.17398095e-01 -7.72736609e-01 -1.67071521e-01 1.19108307e+00
-3.08237225e-01 -8.01360682e-02 3.43487769e-01 2.48329621e-03
7.26766646e-01 -2.28878688e-02 3.83346379e-01 -5.09200037e-01
-2.77244180e-01 -3.27580065e-01 4.57263023e-01 1.70433235e+00
-2.41850793e-01 1.63049042e-01 5.36149740e-01 -4.45922166e-01
-9.24669802e-01 -1.20439601e+00 -2.62964994e-01 1.38865006e+00
-2.40375414e-01 8.73980522e-02 -3.20205659e-01 -5.14798939e-01
6.54728949e-01 8.56660247e-01 -9.64169443e-01 3.53327394e-01
1.38969168e-01 -1.23341954e+00 -7.26656476e-03 5.79862833e-01
2.27403939e-01 -5.77718556e-01 -7.96400607e-01 4.75523502e-01
-1.06872534e-02 -6.48532331e-01 -4.11151886e-01 3.10515434e-01
-9.76735651e-01 -9.30381775e-01 -1.21172726e+00 8.89645293e-02
1.32804573e-01 1.96799934e-01 1.31261396e+00 -5.26438653e-01
-6.91459477e-02 6.44294918e-01 -1.50658146e-01 -9.67911363e-01
6.48432225e-02 -8.65767002e-02 1.07253544e-01 2.34945774e-01
3.93542677e-01 -5.84289551e-01 -7.80379891e-01 2.20692351e-01
-6.11352503e-01 -2.27983996e-01 4.28024292e-01 9.24478948e-01
6.77494466e-01 -2.36704215e-01 4.91785496e-01 -1.04910886e+00
7.33892977e-01 -8.05101991e-01 -1.19688785e+00 3.39585125e-01
-4.60364133e-01 2.13045225e-01 2.54573256e-01 -4.67565805e-01
-1.05271542e+00 -3.16722602e-01 1.20347306e-01 -3.88458073e-01
3.39752555e-01 7.15698957e-01 3.31827432e-01 3.41662049e-01
3.80715251e-01 2.44441167e-01 -4.70993035e-02 -6.91503763e-01
3.19011480e-01 5.52571297e-01 2.14265600e-01 -1.26563251e+00
4.54268008e-01 7.85941184e-01 1.89321712e-01 -1.96428120e-01
-1.75377643e+00 -2.68538207e-01 1.15765788e-01 9.10548717e-02
2.88236380e-01 -8.35004628e-01 -1.20923722e+00 3.58361632e-01
-6.84730768e-01 -4.59916949e-01 -5.66678286e-01 7.28623152e-01
-1.31864059e+00 1.56401321e-01 -5.80039620e-01 -1.80641794e+00
-2.87494779e-01 -6.26112998e-01 8.57178271e-01 3.19649249e-01
6.33975491e-02 -9.98084009e-01 2.02105820e-01 3.64886642e-01
4.58399504e-01 9.14542284e-03 6.09409451e-01 -5.71300089e-01
-7.02107370e-01 -1.14648290e-01 -3.27980444e-02 7.01537058e-02
-1.88011974e-01 -2.29923546e-01 -7.38240421e-01 -3.92800599e-01
2.17385516e-01 -4.27160621e-01 1.07381701e+00 1.09859228e+00
1.48006094e+00 -3.46674234e-01 -3.20705682e-01 3.95325452e-01
1.06621349e+00 3.56527679e-02 1.38555139e-01 4.30124104e-01
-1.73940241e-01 3.65087509e-01 7.45213568e-01 1.39128864e+00
1.82996050e-01 1.38505414e-01 3.78442675e-01 5.29495955e-01
7.90297747e-01 -3.21045101e-01 1.95741765e-02 1.51318982e-01
-2.03744322e-01 -4.53198403e-02 -3.26704234e-01 6.11084580e-01
-2.16299534e+00 -1.32013059e+00 2.20108852e-01 2.41993785e+00
1.01616335e+00 2.85619825e-01 6.23835981e-01 -1.50632769e-01
8.40064764e-01 -2.81099398e-02 -1.09808719e+00 -1.16952136e-01
-1.30756080e-01 5.24503708e-01 9.29963529e-01 4.29165572e-01
-1.05757189e+00 5.21142304e-01 7.71779823e+00 1.24397874e+00
-5.07587075e-01 2.74347037e-01 1.10714376e+00 -7.43764937e-01
-4.42259520e-01 -1.87741414e-01 -1.09370601e+00 7.52245367e-01
9.10889089e-01 -6.78241193e-01 5.13262033e-01 1.17477870e+00
2.67856717e-01 -3.26177120e-01 -1.18573868e+00 1.07765460e+00
-5.65274477e-01 -1.48066151e+00 -3.08087349e-01 1.80682883e-01
1.08067715e+00 1.87364705e-02 3.64224434e-01 2.48379841e-01
1.37857175e+00 -9.14379120e-01 1.03583038e+00 6.26292169e-01
5.10043085e-01 -1.01959050e+00 5.51138282e-01 5.03294408e-01
-6.16197705e-01 -3.10955226e-01 -6.36241198e-01 -4.43466187e-01
3.98598790e-01 1.15264463e+00 -1.54283345e-01 4.22362536e-01
8.62493098e-01 5.59417367e-01 3.85881513e-01 1.19044232e+00
-1.72600180e-01 8.18022907e-01 -6.58932507e-01 -6.23128772e-01
1.95472777e-01 -3.63877296e-01 3.99404019e-01 8.88378143e-01
5.39021254e-01 4.04960327e-02 5.53145073e-02 9.22727764e-01
-1.11626342e-01 -5.19374497e-02 -2.45562419e-01 1.15042493e-01
5.04369676e-01 7.24731803e-01 -6.95112050e-01 -6.12216473e-01
-2.70048857e-01 4.74294275e-01 2.79416084e-01 4.16116506e-01
-1.03890133e+00 -3.03058457e-02 8.29502702e-01 -1.02785937e-01
9.69546258e-01 -5.20285405e-03 -4.39464658e-01 -1.40599835e+00
1.06552904e-02 -5.41530013e-01 6.91004038e-01 -3.60707372e-01
-1.80693007e+00 -1.10605478e-01 4.06213194e-01 -8.06506872e-01
-4.18719172e-01 -6.29805624e-01 -4.17314261e-01 8.65559220e-01
-1.60002065e+00 -5.38059711e-01 5.63718140e-01 7.68763006e-01
3.63763034e-01 -5.41560873e-02 3.29573721e-01 -2.17204913e-01
-4.57215220e-01 5.40426314e-01 1.06424332e+00 -2.15736911e-01
1.72163382e-01 -1.37248671e+00 2.22537845e-01 6.13738596e-01
2.17409357e-01 6.03365123e-01 1.07819247e+00 -5.98390579e-01
-1.45171857e+00 -7.10315466e-01 1.01946704e-01 -3.40928555e-01
1.13130796e+00 -2.15321053e-02 -1.15812354e-01 9.63480949e-01
-9.61227864e-02 -1.21021397e-01 9.27360713e-01 4.11090463e-01
-2.66401589e-01 -3.08972429e-02 -1.32384038e+00 3.83129984e-01
9.54850435e-01 -6.56631663e-02 -5.19185066e-01 6.24806643e-01
2.79640883e-01 -3.15688610e-01 -7.74693549e-01 -1.07709214e-01
9.85881746e-01 -1.00690925e+00 8.83297980e-01 -9.78242636e-01
3.09869170e-01 3.10896844e-01 -4.95471478e-01 -1.27319145e+00
-4.89941329e-01 -1.22739136e+00 -5.18530726e-01 7.25676477e-01
3.60397965e-01 -9.13977683e-01 1.07153714e+00 6.83378279e-01
4.89870250e-01 -7.25469589e-01 -1.27903128e+00 -8.73297095e-01
4.70275521e-01 -7.79151499e-01 7.08063185e-01 4.32373166e-01
-1.00276344e-01 -1.09823875e-01 -8.18563223e-01 -2.37769350e-01
1.30869150e+00 8.20642412e-01 6.51900232e-01 -1.33262622e+00
-9.33208704e-01 -6.07001066e-01 3.90859962e-01 -1.67955470e+00
4.52111512e-02 -3.72918040e-01 -5.30489795e-02 -1.04657793e+00
7.26027548e-01 -4.86884117e-01 -5.31570971e-01 4.03806940e-02
-2.06846058e-01 9.88573581e-02 2.12599915e-02 2.09511697e-01
-7.35231221e-01 6.66855276e-01 1.16201329e+00 -7.55491257e-02
2.95675129e-01 8.04307401e-01 -1.09401524e+00 6.74623728e-01
5.98263681e-01 -5.19693077e-01 -2.90411413e-01 -1.57913104e-01
6.29210949e-01 4.33352202e-01 1.70046791e-01 -2.21442476e-01
-6.52411059e-02 -8.84266794e-01 4.86369252e-01 -9.42436159e-01
2.97298998e-01 -5.22971570e-01 2.34478284e-02 5.60642958e-01
-3.73113334e-01 -1.95394948e-01 -1.52898893e-01 1.04380667e+00
3.19694012e-01 -4.49519932e-01 7.99698889e-01 -5.59140384e-01
9.13535282e-02 5.25367975e-01 -4.64793980e-01 2.83338189e-01
9.16884243e-01 1.62542224e-01 -1.26548097e-01 -9.24810231e-01
-9.04010177e-01 4.11860973e-01 -4.31876965e-02 -2.98831820e-01
1.78224862e-01 -1.28268313e+00 -7.62001395e-01 -1.09660216e-01
-3.96691054e-01 -1.79077193e-01 3.95113856e-01 9.56223428e-01
1.41960531e-02 4.09880280e-01 1.35657310e-01 -4.12556648e-01
-6.47016764e-01 5.34256935e-01 -3.31538096e-02 -6.61641657e-01
-1.46295533e-01 1.08046019e+00 4.10092734e-02 2.45666564e-01
2.69500613e-01 -5.55030406e-01 2.44322687e-01 1.84178531e-01
9.22599077e-01 6.93089902e-01 -2.45353699e-01 2.68343747e-01
-3.26607600e-02 9.00119245e-02 -3.02487284e-01 -5.58874011e-01
1.50461471e+00 -4.49579924e-01 1.48575455e-01 7.31603980e-01
6.10140860e-01 2.74265781e-02 -1.52914810e+00 -5.12162626e-01
7.44411275e-02 -8.93106461e-01 3.06110494e-02 -6.07040823e-01
-8.89492393e-01 5.47942162e-01 7.39528164e-02 6.99349403e-01
6.01720273e-01 1.14616632e-01 6.39367163e-01 4.32456702e-01
8.90292764e-01 -1.24376094e+00 -2.12869257e-01 3.79052341e-01
5.62824488e-01 -1.18701410e+00 1.95370629e-01 1.30331472e-01
-5.24606884e-01 9.39789534e-01 -3.13069075e-01 -3.35865766e-01
1.01560581e+00 2.35644475e-01 -4.48436648e-01 1.55854464e-01
-9.39510763e-01 -2.07510382e-01 -1.19344898e-01 3.53605390e-01
7.45042041e-02 4.78726983e-01 -3.47951949e-01 1.13889134e+00
-6.05472326e-01 2.98262954e-01 6.13791227e-01 1.08383775e+00
-6.36263132e-01 -9.29965079e-01 -6.11458778e-01 9.94767785e-01
-1.10918689e+00 -1.69369370e-01 2.41181850e-01 4.15239364e-01
-7.66549289e-01 9.78423059e-01 6.40664697e-02 2.84318537e-01
-1.15714267e-01 -4.65727225e-02 7.75341809e-01 -2.89624423e-01
-5.61459810e-02 3.81841868e-01 1.50608383e-02 -3.72582257e-01
-4.46287751e-01 -1.35503864e+00 -2.50067979e-01 -8.41543555e-01
-3.14101815e-01 4.76022542e-01 4.64932442e-01 1.01005447e+00
5.27238846e-02 5.87467700e-02 8.20004880e-01 -8.68956566e-01
-1.60216177e+00 -1.12233400e+00 -1.38360751e+00 1.44529626e-01
4.07065034e-01 -9.68402267e-01 -5.21669567e-01 -3.28053117e-01] | [4.557809352874756, 3.288393974304199] |
3dc753bf-1892-4c0c-a64f-87f8d0b9eab6 | towards-learning-to-detect-and-predict | 1910.03973 | null | https://arxiv.org/abs/1910.03973v1 | https://arxiv.org/pdf/1910.03973v1.pdf | Towards Learning to Detect and Predict Contact Events on Vision-based Tactile Sensors | In essence, successful grasp boils down to correct responses to multiple contact events between fingertips and objects. In most scenarios, tactile sensing is adequate to distinguish contact events. Due to the nature of high dimensionality of tactile information, classifying spatiotemporal tactile signals using conventional model-based methods is difficult. In this work, we propose to predict and classify tactile signal using deep learning methods, seeking to enhance the adaptability of the robotic grasp system to external event changes that may lead to grasping failure. We develop a deep learning framework and collect 6650 tactile image sequences with a vision-based tactile sensor, and the neural network is integrated into a contact-event-based robotic grasping system. In grasping experiments, we achieved 52% increase in terms of object lifting success rate with contact detection, significantly higher robustness under unexpected loads with slip prediction compared with open-loop grasps, demonstrating that integration of the proposed framework into robotic grasping system substantially improves picking success rate and capability to withstand external disturbances. | ['Michael Yu Wang', 'Zicheng Kan', 'Yazhan Zhang', 'Weihao Yuan'] | 2019-10-09 | null | null | null | null | ['contact-detection'] | ['robots'] | [ 3.94170195e-01 -5.40175438e-01 1.54044449e-01 -1.40857831e-01
-3.30878705e-01 -6.67715311e-01 7.09487824e-03 -9.26685613e-03
-3.22198838e-01 4.87788200e-01 -2.29303554e-01 1.09329395e-01
-4.57588732e-01 -8.35754395e-01 -8.70173454e-01 -7.60912180e-01
-3.31061304e-01 6.73183352e-02 2.84343481e-01 -1.12290755e-01
6.25685930e-01 6.58430338e-01 -1.92870080e+00 5.97114444e-01
6.06920898e-01 1.70337379e+00 1.07958531e+00 4.97000962e-01
-3.75481579e-03 6.75697625e-02 -2.81339616e-01 2.46071458e-01
2.59913921e-01 3.94309759e-01 -3.79613012e-01 -4.64714438e-01
-4.22483012e-02 -8.51962566e-01 -3.43675554e-01 5.93631268e-01
7.77049601e-01 -3.10953826e-01 8.40338826e-01 -8.49253476e-01
-7.99591362e-01 4.09248471e-01 -3.32312047e-01 -3.46866786e-01
4.14954513e-01 1.87622026e-01 5.91427982e-01 -1.01191783e+00
4.58732486e-01 1.12970018e+00 5.28386772e-01 6.57072306e-01
-1.04546463e+00 -2.66075879e-01 -1.90504804e-01 4.20217067e-01
-4.50321555e-01 -6.06315695e-02 9.95626509e-01 -4.60173875e-01
8.88739228e-01 6.83065653e-02 4.52412993e-01 1.78796017e+00
6.57475770e-01 9.03507054e-01 1.06512678e+00 -1.83382049e-01
3.77334833e-01 -5.59621334e-01 -9.51669067e-02 1.94491223e-01
3.82715940e-01 2.64314294e-01 -4.96054977e-01 8.50456282e-02
1.15101826e+00 3.33116114e-01 -8.29897970e-02 -3.35683018e-01
-1.37699568e+00 7.17639998e-02 7.52366126e-01 2.74039030e-01
-9.04225051e-01 2.47487113e-01 6.29857123e-01 3.53427112e-01
-2.32040256e-01 5.74049830e-01 -5.93159556e-01 -3.18635851e-01
-8.72864872e-02 2.07208112e-01 9.24986184e-01 7.90307403e-01
2.49422356e-01 1.84581671e-02 -3.26867223e-01 9.13864911e-01
8.75890553e-02 9.39644217e-01 3.76816154e-01 -9.88353550e-01
3.19650650e-01 5.64397693e-01 6.48142934e-01 -1.13369823e+00
-4.95480746e-01 2.58170545e-01 -7.27862179e-01 5.06583214e-01
3.91132861e-01 -2.05876917e-01 -9.41518843e-01 1.36646760e+00
-8.00396577e-02 -5.14852405e-01 -4.90316749e-02 1.23789334e+00
3.83710265e-01 2.11351350e-01 1.72792450e-01 -5.10371896e-03
8.93845081e-01 -8.96253884e-02 -7.68499970e-01 -1.37572825e-01
-2.12905228e-01 -7.15318918e-01 1.22409654e+00 6.32764757e-01
-8.40080023e-01 -8.01289558e-01 -1.00994611e+00 3.18482161e-01
-3.62604856e-01 6.58402264e-01 7.72423625e-01 -1.51695952e-01
-3.94168705e-01 1.00575185e+00 -1.15174377e+00 -4.23072666e-01
2.78729886e-01 4.52584535e-01 -2.68334627e-01 6.51186854e-02
-9.74160373e-01 9.64388072e-01 3.17761093e-01 4.81328607e-01
-8.08244467e-01 -4.63543981e-01 -2.45269522e-01 2.13501714e-02
1.01524152e-01 -2.93974392e-02 1.03839636e+00 -1.28929585e-01
-1.68674541e+00 4.84136909e-01 6.58384636e-02 -4.45874929e-02
6.55734897e-01 -7.22246170e-01 1.64453641e-01 2.86690891e-01
-4.65086550e-02 3.42793047e-01 9.62281823e-01 -1.55894244e+00
-1.51589215e-01 -5.60304880e-01 -2.54777938e-01 -1.94684535e-01
-4.69974697e-01 -4.76710618e-01 3.53251219e-01 -4.54956472e-01
5.76956153e-01 -8.42046082e-01 2.78804988e-01 3.55193913e-01
-2.80820936e-01 -3.67548555e-01 1.19818807e+00 -4.79869694e-01
4.55343217e-01 -2.05436206e+00 2.07369372e-01 1.01331286e-01
-2.77072430e-01 4.03672040e-01 -2.93784648e-01 1.00124741e+00
2.95607895e-01 -3.49366546e-01 7.81333447e-02 1.96595013e-01
1.92259967e-01 3.56023043e-01 -8.20272148e-01 4.75937575e-02
4.91781145e-01 1.08084631e+00 -8.81730497e-01 9.70489345e-03
5.07130921e-01 1.36656687e-01 -2.39447621e-03 4.74907458e-01
-3.04239124e-01 2.68141359e-01 -8.50214958e-01 1.23696280e+00
7.12857127e-01 2.82129318e-01 2.09276333e-01 -5.74900329e-01
-2.63826221e-01 -1.53660566e-01 -8.32808673e-01 1.72286248e+00
-5.97596169e-01 2.69028276e-01 3.92009765e-01 -1.07544470e+00
1.45520711e+00 2.49801099e-01 7.23673165e-01 -9.30087984e-01
3.24666828e-01 6.90227628e-01 -1.62781149e-01 -1.15180004e+00
7.89313540e-02 3.20806354e-01 4.66846488e-02 1.80892870e-01
1.95863657e-02 -2.38546312e-01 -2.94131458e-01 -5.49051285e-01
1.06255496e+00 6.08008862e-01 -4.69503224e-01 -2.92155862e-01
-7.47313574e-02 -2.38260224e-01 1.04400307e-01 1.00060952e+00
-1.63779035e-01 3.72654080e-01 1.25814021e-01 -4.55359340e-01
-1.00578773e+00 -1.62614512e+00 -3.03445011e-01 1.05854988e+00
6.73716545e-01 6.86940849e-01 -3.59701902e-01 7.06700161e-02
8.48983467e-01 1.33134469e-01 -5.55375814e-01 -2.70477504e-01
-6.84360087e-01 -4.20422256e-01 3.08565140e-01 9.99435604e-01
5.43781102e-01 -1.69072902e+00 -1.25527942e+00 7.19319642e-01
-7.54197389e-02 -9.52395201e-01 4.90610152e-01 5.02875984e-01
-8.15113902e-01 -1.17022061e+00 -8.57189000e-01 -1.06913114e+00
3.42269808e-01 1.90978602e-01 2.32097298e-01 -2.65101641e-01
-1.01408482e+00 5.16209126e-01 -6.51054442e-01 -5.90098858e-01
-7.53706992e-02 -1.44396216e-01 4.03759509e-01 -3.52990955e-01
3.27469558e-01 -5.94108999e-01 -9.18391585e-01 3.95338684e-01
-7.71776736e-01 -3.57279450e-01 9.74877715e-01 1.06275117e+00
2.50807166e-01 -3.70531589e-01 1.20882249e+00 3.88384044e-01
1.01058793e+00 -2.16940522e-01 -1.00954160e-01 4.00914133e-01
-1.44677714e-01 -3.66673142e-01 5.30429900e-01 -8.25542390e-01
-1.04816413e+00 3.21849249e-02 5.94634525e-02 -3.03102851e-01
-2.76865929e-01 5.34977317e-01 2.67406493e-01 -2.40283236e-01
6.35808945e-01 2.59521604e-01 3.99295002e-01 -4.59021062e-01
-1.02083370e-01 1.09380746e+00 6.53283954e-01 -1.09753191e+00
1.20327272e-01 2.26554066e-01 1.03164650e-01 -7.83836722e-01
-2.17012584e-01 -1.22818321e-01 -8.25531960e-01 -7.36906290e-01
5.94527066e-01 -4.88858610e-01 -1.63165712e+00 1.16220629e+00
-1.32159281e+00 -5.93370914e-01 1.62675112e-01 4.00264502e-01
-8.71851921e-01 3.43695402e-01 -9.90283251e-01 -1.13067186e+00
-6.36899054e-01 -9.02396619e-01 1.27654564e+00 2.09135205e-01
-3.46090905e-02 -2.59428263e-01 -3.89866143e-01 -2.13003427e-01
8.86798561e-01 6.16808057e-01 8.98133576e-01 -9.94013548e-02
-2.30599821e-01 -6.36750877e-01 -4.00349081e-01 3.75860810e-01
7.38483906e-01 -9.93665457e-02 -9.19806361e-01 -5.43355763e-01
5.40293977e-02 -9.25689697e-01 7.67542899e-01 3.69363487e-01
1.39331257e+00 2.40126867e-02 -5.07999897e-01 -2.56761760e-01
1.62351477e+00 5.86343229e-01 6.61550045e-01 -5.48728965e-02
2.83942759e-01 7.34729648e-01 9.01696503e-01 6.91489935e-01
-2.67518252e-01 4.62497830e-01 7.97209322e-01 2.21487701e-01
1.59614354e-01 -1.14040738e-02 8.83735046e-02 5.11085510e-01
-4.99198437e-01 -1.02451950e-01 -7.43340254e-01 6.18978441e-01
-1.81593513e+00 -8.23672354e-01 6.73593879e-02 2.03521705e+00
7.95281768e-01 2.00978726e-01 -4.03826118e-01 2.45816633e-01
5.82896590e-01 -3.62285614e-01 -1.12004137e+00 -2.13682413e-01
2.35600714e-02 3.25080395e-01 2.91858405e-01 -4.44497727e-02
-8.76491070e-01 6.84379280e-01 6.03505516e+00 3.03142905e-01
-1.74934244e+00 -5.89678526e-01 -1.64137155e-01 3.71661074e-02
2.50354201e-01 -7.54643619e-01 -2.63299868e-02 5.28353810e-01
-1.17446529e-02 2.53333747e-01 6.09022439e-01 9.02074814e-01
1.62254602e-01 -1.81520328e-01 -1.25691330e+00 8.01375091e-01
-4.31176513e-01 -9.41669166e-01 -9.06161070e-02 -4.05527681e-01
1.81785002e-01 -1.61371566e-02 1.18154986e-02 7.26413308e-03
-2.11601496e-01 -7.92913318e-01 5.63500583e-01 9.56596375e-01
8.78652096e-01 -1.43299341e-01 7.02144802e-01 1.73511162e-01
-1.02049673e+00 -6.04196966e-01 -5.43845415e-01 -3.97200912e-01
1.97897628e-01 3.33759487e-01 -6.94247067e-01 5.26761830e-01
1.10069072e+00 4.75985140e-01 2.36533776e-01 8.29677701e-01
1.39960602e-01 2.45139420e-01 -3.55053365e-01 -5.74824989e-01
-4.53976393e-02 2.52655536e-01 5.15428722e-01 9.99364555e-01
5.50847888e-01 -4.98860106e-02 1.69340417e-01 9.51442778e-01
4.45831776e-01 -5.09010375e-01 -7.84248590e-01 -1.65184498e-01
8.09847653e-01 1.04712009e+00 -4.52341229e-01 2.90788542e-02
1.51694849e-01 1.02664816e+00 1.04346871e-01 6.10976815e-01
-3.93375039e-01 -8.74373138e-01 4.40742612e-01 -2.09927306e-01
3.09364110e-01 -7.26161659e-01 -6.67031407e-01 -7.78229833e-01
7.54349291e-01 -2.38034487e-01 -5.56957126e-01 -1.01200402e+00
-1.78775811e+00 2.58524641e-02 -2.53306568e-01 -1.13750064e+00
1.35207713e-01 -1.47241628e+00 -5.54150760e-01 8.67286146e-01
-1.15706408e+00 -9.44623232e-01 -7.62495995e-01 3.49723518e-01
5.57614267e-01 1.31921485e-01 1.08159196e+00 -3.62598225e-02
8.70636478e-02 1.76018000e-01 1.96948007e-01 -5.01298681e-02
8.12634110e-01 -1.10445285e+00 -5.31553999e-02 1.18399926e-01
-9.16519642e-01 8.38438272e-01 5.63956857e-01 -8.81046474e-01
-2.30916667e+00 -7.05790043e-01 -2.26826378e-04 5.78207746e-02
7.56462336e-01 -3.77935082e-01 -9.57506418e-01 1.30922630e-01
-1.94307864e-01 -2.91888088e-01 2.72588562e-02 -3.93268555e-01
-7.19014555e-03 -2.73063689e-01 -1.24470520e+00 5.21359324e-01
1.06366980e+00 -5.17402172e-01 -8.04411888e-01 1.39862806e-01
1.97626263e-01 -4.02398914e-01 -1.24776447e+00 1.17229819e+00
1.40756440e+00 -4.44851100e-01 1.04705489e+00 -5.16583264e-01
6.78227127e-01 6.48981556e-02 -2.93641359e-01 -1.02649748e+00
-3.32694769e-01 -1.89565048e-01 -9.07277167e-02 8.57323527e-01
-1.90645665e-01 -6.60944104e-01 5.15968263e-01 5.20286679e-01
-2.97647744e-01 -9.44757044e-01 -1.12987995e+00 -7.89013088e-01
9.31053162e-02 2.86522880e-02 1.43393457e-01 4.70555365e-01
4.23451036e-01 -6.38869762e-01 -1.91783324e-01 1.68508396e-01
6.75207138e-01 3.81241143e-01 2.67253309e-01 -1.43025994e+00
1.29951864e-01 -3.36067349e-01 -3.24191570e-01 -9.18148518e-01
-9.39348191e-02 -3.76530588e-01 6.73990786e-01 -1.81533408e+00
-7.36861750e-02 -6.58374071e-01 -5.33428013e-01 7.58163750e-01
-2.41597686e-02 -6.57787025e-02 4.65241931e-02 2.48701304e-01
1.33516537e-02 5.81009388e-01 1.54692066e+00 -3.13451171e-01
-1.29938245e-01 -9.25295278e-02 1.97340399e-01 1.51664808e-01
8.61043513e-01 2.68602073e-01 -2.57704761e-02 -7.17700243e-01
-1.66939512e-01 3.97246242e-01 8.40607822e-01 -1.11366892e+00
1.77480564e-01 -2.35077754e-01 7.16789484e-01 -5.00318885e-01
3.96419346e-01 -9.56491888e-01 -1.17463283e-01 8.85302901e-01
-4.40033078e-01 -5.18188655e-01 4.48981404e-01 6.60384178e-01
-6.32209256e-02 -1.05178514e-02 2.81350911e-01 -6.73940331e-02
-8.24252367e-01 -8.70928541e-02 -4.69351172e-01 -8.15072238e-01
1.07245970e+00 -2.94970989e-01 -4.50086385e-01 2.35500291e-01
-9.24294531e-01 1.00475863e-01 2.16828361e-01 8.25973570e-01
9.36010778e-01 -1.31239676e+00 -1.57921672e-01 2.13112876e-01
-5.04676551e-02 -1.71847060e-01 5.08518219e-01 4.87061709e-01
-3.61659646e-01 2.65733302e-01 -9.13369060e-01 -8.45607042e-01
-7.01929569e-01 2.55464911e-01 -7.46431947e-02 5.77508807e-01
-5.38737178e-01 5.60995758e-01 -4.99032289e-01 -1.93619773e-01
5.75984359e-01 -5.11281073e-01 1.08922005e-03 -1.62165523e-01
1.03445826e-02 4.23609495e-01 2.07578376e-01 4.47100818e-01
-4.13679481e-01 7.95070171e-01 1.64935231e-01 2.87558764e-01
1.45008445e+00 1.41983181e-01 9.15563852e-03 5.66150069e-01
8.58345091e-01 -6.92229450e-01 -1.77177644e+00 1.44201323e-01
-6.45169169e-02 -3.27843487e-01 -4.34960246e-01 -1.25834453e+00
-5.65987468e-01 1.03728926e+00 9.41069961e-01 4.22923356e-01
9.54903603e-01 -1.65922567e-01 9.63412821e-01 9.93879080e-01
1.06706142e+00 -1.40217996e+00 6.13876283e-01 5.28077006e-01
1.56213379e+00 -1.44129407e+00 -5.47401130e-01 -1.74327314e-01
-3.41311604e-01 1.67024291e+00 7.87878692e-01 -6.20850265e-01
4.57286477e-01 5.20887136e-01 1.68260597e-02 -3.04902121e-02
-4.75540638e-01 2.81201422e-01 -6.47771508e-02 6.88356817e-01
-2.32594702e-02 3.00861359e-01 -2.67887056e-01 4.95771736e-01
4.10227239e-01 5.66421449e-01 -1.07902894e-03 1.56637692e+00
-7.83692241e-01 -7.14137197e-01 -1.02122627e-01 6.77075088e-01
-1.13587022e-01 4.42642212e-01 -2.01588660e-01 4.88614827e-01
-6.59307167e-02 8.50813746e-01 5.19444905e-02 -6.98633671e-01
6.13818765e-01 6.26807241e-03 8.55168223e-01 -9.46802869e-02
-1.03732824e-01 -1.76747039e-01 -2.68740863e-01 -8.67642045e-01
-2.48343065e-01 -4.67683822e-01 -1.28029120e+00 2.74832666e-01
-2.08072275e-01 -4.71894413e-01 1.12365150e+00 7.97889173e-01
6.29349172e-01 5.34969866e-01 6.82394624e-01 -1.60232747e+00
-1.13587201e+00 -1.29916310e+00 -6.12299383e-01 4.89359856e-01
2.22506955e-01 -1.28884697e+00 -1.95943877e-01 -3.75103615e-02] | [5.846062660217285, -0.8230483531951904] |
e26d07eb-faf7-4642-88cc-429ef42cd228 | deep-logismos-deep-learning-graph-based-3d | 1801.08599 | null | http://arxiv.org/abs/1801.08599v1 | http://arxiv.org/pdf/1801.08599v1.pdf | Deep LOGISMOS: Deep Learning Graph-based 3D Segmentation of Pancreatic Tumors on CT scans | This paper reports Deep LOGISMOS approach to 3D tumor segmentation by
incorporating boundary information derived from deep contextual learning to
LOGISMOS - layered optimal graph image segmentation of multiple objects and
surfaces. Accurate and reliable tumor segmentation is essential to tumor growth
analysis and treatment selection. A fully convolutional network (FCN), UNet, is
first trained using three adjacent 2D patches centered at the tumor, providing
contextual UNet segmentation and probability map for each 2D patch. The UNet
segmentation is then refined by Gaussian Mixture Model (GMM) and morphological
operations. The refined UNet segmentation is used to provide the initial shape
boundary to build a segmentation graph. The cost for each node of the graph is
determined by the UNet probability maps. Finally, a max-flow algorithm is
employed to find the globally optimal solution thus obtaining the final
segmentation. For evaluation, we applied the method to pancreatic tumor
segmentation on a dataset of 51 CT scans, among which 30 scans were used for
training and 21 for testing. With Deep LOGISMOS, DICE Similarity Coefficient
(DSC) and Relative Volume Difference (RVD) reached 83.2+-7.8% and 18.6+-17.4%
respectively, both are significantly improved (p<0.05) compared with contextual
UNet and/or LOGISMOS alone. | ['Jianhua Yao', 'Zhihui Guo', 'Milan Sonka', 'Le Lu', 'Ronald M. Summers', 'Ling Zhang', 'Mohammadhadi Bagheri'] | 2018-01-25 | null | null | null | null | ['unet-segmentation'] | ['computer-vision'] | [ 2.71587372e-01 3.49194556e-01 -3.11649799e-01 -6.86763227e-02
-8.10263216e-01 -2.33541161e-01 1.14168987e-01 5.99517405e-01
-3.22375625e-01 5.21588504e-01 -1.64571386e-02 -3.05352867e-01
-6.89475983e-02 -7.35016942e-01 -4.61774647e-01 -1.04074705e+00
-3.40913445e-01 6.72953546e-01 2.71850497e-01 1.16675436e-01
2.64339834e-01 8.03730905e-01 -4.71106380e-01 4.37379591e-02
1.30552912e+00 1.14202225e+00 4.97051865e-01 7.10962832e-01
-4.73893464e-01 2.00748622e-01 -2.01955318e-01 6.97699711e-02
2.61430264e-01 -4.41498250e-01 -7.72868752e-01 4.56473500e-01
1.88087538e-01 1.96738751e-04 5.02299629e-02 1.25724101e+00
3.28859508e-01 -3.02703176e-02 1.08161712e+00 -7.90183187e-01
-2.45563149e-01 5.92173994e-01 -8.76735866e-01 1.49989873e-01
-2.64386773e-01 1.03463612e-01 5.75418472e-01 -7.09733009e-01
5.77406883e-01 7.49783993e-01 7.82719672e-01 4.24754709e-01
-1.15960622e+00 -3.69533509e-01 -2.06390947e-01 -5.01325786e-01
-1.38812673e+00 2.49632090e-01 7.14348674e-01 -4.30015773e-01
7.49154031e-01 3.08063835e-01 1.20278716e+00 3.65644723e-01
1.00097728e+00 7.23112702e-01 9.31426764e-01 -3.95982772e-01
1.28703326e-01 -1.16423354e-01 2.20217541e-01 9.91737545e-01
3.46699029e-01 -3.38799417e-01 3.36410850e-01 8.93261284e-02
1.21104014e+00 1.35993566e-02 -1.16161019e-01 -4.30673122e-01
-1.04124522e+00 8.69667232e-01 9.76213753e-01 4.85789806e-01
-6.58314168e-01 4.40776832e-02 3.62157196e-01 -1.40276626e-01
5.24346888e-01 1.60156399e-01 -1.02743074e-01 5.22184193e-01
-7.82447517e-01 -1.08930312e-01 7.12308943e-01 6.08545125e-01
4.58219439e-01 -1.42930999e-01 -1.59318492e-01 6.51863217e-01
5.29607713e-01 2.11468011e-01 6.66152656e-01 -7.96060383e-01
-3.07068806e-02 1.15562403e+00 -2.72926688e-01 -8.32268298e-01
-8.82627308e-01 -7.58850694e-01 -1.23732722e+00 1.02868609e-01
6.46930635e-01 -5.43507785e-02 -1.51394808e+00 1.28103530e+00
5.61809599e-01 2.14935020e-01 2.49614902e-02 9.27663743e-01
1.05854750e+00 3.24509174e-01 2.93681502e-01 -4.71827090e-01
1.46435905e+00 -9.43813562e-01 -5.30301869e-01 -4.72969785e-02
1.17844975e+00 -6.24317646e-01 7.18962371e-01 2.21117232e-02
-1.03008378e+00 3.39808688e-02 -7.94614911e-01 2.67916739e-01
-5.68460301e-02 8.13418105e-02 5.56071997e-01 4.97423917e-01
-1.29886425e+00 3.00354570e-01 -1.20327783e+00 -5.11581361e-01
8.54396820e-01 6.91082895e-01 -3.29350859e-01 1.45247102e-01
-7.40542471e-01 7.17310429e-01 4.21882093e-01 5.32197207e-02
-1.04454505e+00 -8.72655928e-01 -8.42396557e-01 -1.45566657e-01
1.19417652e-01 -8.38975728e-01 9.17825818e-01 -8.66221607e-01
-1.42252851e+00 1.14879537e+00 5.62502407e-02 -5.45784831e-01
6.73915565e-01 3.88096422e-01 1.93424687e-01 4.04261976e-01
2.19787091e-01 7.52035260e-01 4.38549638e-01 -1.12028122e+00
-5.17609000e-01 -6.68807566e-01 -6.23903632e-01 7.94819832e-01
8.37500319e-02 -3.41098189e-01 -4.60561216e-01 -5.24942935e-01
7.96210647e-01 -9.67635155e-01 -7.61038423e-01 7.60486722e-02
-7.72702754e-01 -1.63353875e-01 8.46600294e-01 -9.51693952e-01
8.85255754e-01 -1.78289437e+00 2.86939502e-01 6.35024548e-01
5.22906542e-01 -4.68011200e-02 2.55780131e-01 -3.52328092e-01
-1.50487041e-02 1.94635451e-01 -6.98161721e-01 -3.69662121e-02
-5.42402089e-01 1.28401771e-01 7.68162012e-01 8.08555841e-01
-2.40931734e-01 1.06571543e+00 -6.79673254e-01 -9.99146283e-01
3.60055953e-01 3.12350303e-01 -4.52413976e-01 3.79461050e-02
-2.65654445e-01 6.96958125e-01 -6.52520418e-01 5.73041141e-01
7.78947592e-01 -2.48524636e-01 1.75583854e-01 -1.94292739e-01
1.77071467e-02 -4.25600737e-01 -8.24303389e-01 1.83726907e+00
-1.13235451e-01 3.41847926e-01 3.05577785e-01 -1.01914859e+00
9.76359665e-01 2.58979768e-01 1.33431339e+00 -4.02885824e-01
6.17837191e-01 4.59816009e-01 7.30800852e-02 -6.00828171e-01
-1.98222511e-02 -2.49078348e-01 1.60551537e-02 2.34104499e-01
-3.09650809e-01 -5.13219237e-01 3.82862315e-02 1.61937192e-01
8.39561820e-01 -2.54542857e-01 3.98960263e-01 -7.50778258e-01
6.54541612e-01 4.28065240e-01 4.35497910e-01 3.21691453e-01
-3.35759968e-01 5.79476595e-01 7.32380629e-01 -4.46959764e-01
-7.14315295e-01 -1.03624940e+00 -4.60040510e-01 2.44574532e-01
4.38105762e-01 2.18703315e-01 -1.09228551e+00 -1.00974655e+00
-9.45115462e-02 4.93597955e-01 -7.56386220e-01 1.98585875e-02
-7.25262582e-01 -1.09965777e+00 4.09475751e-02 3.55834126e-01
5.42831004e-01 -9.12871540e-01 -6.78490877e-01 4.11410481e-01
-1.21063955e-01 -8.15176249e-01 -5.90783656e-01 2.69057781e-01
-1.39334130e+00 -1.19319487e+00 -9.09068763e-01 -1.11478138e+00
1.25353885e+00 2.53213849e-02 8.11454475e-01 2.52201676e-01
-6.27906442e-01 2.58194301e-02 -2.68271826e-02 -1.66056558e-01
-7.04599500e-01 4.78065945e-02 -4.42522466e-01 -1.37524545e-01
4.61474434e-02 -3.68977726e-01 -7.43372917e-01 2.64108241e-01
-6.94115877e-01 3.91800731e-01 8.02327216e-01 8.53268206e-01
1.11085689e+00 1.29727438e-01 2.25769043e-01 -8.30984473e-01
4.45804685e-01 -3.46600652e-01 -5.48423290e-01 4.84481677e-02
-4.46981877e-01 -3.71436149e-01 2.20947981e-01 -3.25198084e-01
-7.15959191e-01 1.99247256e-01 -9.55215748e-03 -3.86101663e-01
-1.44934252e-01 6.11260951e-01 2.01666564e-01 -2.63810277e-01
6.28841758e-01 1.21655203e-01 4.96638179e-01 1.12796230e-02
2.87063234e-03 3.63755971e-01 2.80962408e-01 -2.70801987e-02
2.34295547e-01 4.80216771e-01 3.58265132e-01 -7.39141047e-01
-4.46344137e-01 -2.41895035e-01 -8.56060922e-01 -3.46231699e-01
1.39676166e+00 -3.31165582e-01 -4.41261888e-01 5.78074157e-01
-7.05302179e-01 -6.55038059e-01 -2.95327425e-01 7.40359604e-01
-5.20837665e-01 9.09239650e-02 -9.19746935e-01 -3.54768127e-01
-8.68224442e-01 -1.56373596e+00 8.28746915e-01 5.58441877e-01
-7.23416507e-02 -1.33856273e+00 -3.36866289e-01 2.49528959e-01
3.64376873e-01 7.85814166e-01 1.07287943e+00 -7.46381044e-01
-5.70249200e-01 -4.21577424e-01 -2.93026954e-01 2.25241594e-02
3.15161258e-01 -7.53857270e-02 -4.12252396e-01 -2.58046269e-01
5.74974855e-03 2.20606327e-01 7.08097756e-01 1.38163650e+00
1.08768547e+00 -8.60070437e-03 -9.19573009e-01 8.30843925e-01
1.64265287e+00 5.13539135e-01 2.61487126e-01 2.14035034e-01
8.95814121e-01 1.94913015e-01 2.50532061e-01 2.35612348e-01
2.40667015e-01 3.20265263e-01 7.55718708e-01 -3.88018042e-01
-4.46608543e-01 1.22168064e-01 -2.55074024e-01 7.70123482e-01
1.48674637e-01 -2.09449217e-01 -1.20915127e+00 4.53131795e-01
-1.29336786e+00 -1.86450496e-01 -4.40906227e-01 1.83148551e+00
6.01588130e-01 2.14722082e-01 -1.63260534e-01 -1.95058256e-01
9.28570390e-01 -2.50346899e-01 -7.31325090e-01 -1.75345048e-01
1.78299457e-01 -2.30630720e-03 5.69082558e-01 7.11767495e-01
-1.08710492e+00 8.77272606e-01 5.17312098e+00 9.06503916e-01
-1.03869236e+00 -8.77325088e-02 1.09397435e+00 2.78059781e-01
-1.66597560e-01 -1.80461586e-01 -4.35761005e-01 1.35744587e-01
4.27549243e-01 -4.27776836e-02 -4.19723429e-02 4.11852151e-01
3.22626650e-01 -6.90518260e-01 -5.60214341e-01 8.76529872e-01
-1.69648036e-01 -1.31877637e+00 3.31987254e-02 3.69203329e-01
8.92120540e-01 3.01153194e-02 -1.94081292e-01 -6.05118908e-02
3.01876038e-01 -1.11674118e+00 2.02720389e-01 4.65827852e-01
7.73912013e-01 -7.77764738e-01 8.94770741e-01 2.97622800e-01
-1.22588229e+00 1.90769821e-01 -8.47298130e-02 7.71497488e-01
-8.34279228e-03 5.48485398e-01 -1.36020827e+00 6.48880064e-01
3.17623734e-01 6.53956413e-01 -3.82869601e-01 1.25652122e+00
8.35889876e-02 3.57723802e-01 -6.11933827e-01 -1.02928340e-01
4.62393373e-01 -4.14000005e-01 8.82003963e-01 9.29609120e-01
3.45469236e-01 1.14926800e-01 3.29900116e-01 7.43571937e-01
-7.96992481e-02 5.43057263e-01 -8.90096352e-02 3.02345991e-01
2.27596059e-01 1.55533624e+00 -1.70565605e+00 -2.73164213e-01
5.35538532e-02 7.49969363e-01 7.55679384e-02 1.98606655e-01
-7.82205939e-01 -5.62921539e-02 2.46635675e-02 1.39818385e-01
-1.72450483e-01 4.76310551e-02 -6.42218947e-01 -6.54269576e-01
-4.53318477e-01 -3.60663444e-01 7.93412447e-01 -4.27412510e-01
-1.03136861e+00 5.43839037e-01 -5.26903607e-02 -9.50686812e-01
2.29833037e-01 -4.48089421e-01 -8.51751864e-01 8.70858610e-01
-1.15758109e+00 -9.79126573e-01 -7.64872372e-01 2.75070846e-01
6.04053974e-01 1.13099702e-01 6.12676799e-01 -1.96162596e-01
-6.62742853e-01 3.65794986e-01 -2.15510413e-01 2.78030008e-01
1.79439858e-01 -1.41325223e+00 -3.22787642e-01 5.44047534e-01
-4.39409912e-01 9.30999778e-03 3.09720933e-01 -1.06788135e+00
-1.26146460e+00 -1.32051492e+00 3.23871255e-01 -6.79240152e-02
3.60077918e-01 8.31546038e-02 -8.74106944e-01 5.76105952e-01
1.17717505e-01 1.28868163e-01 5.75667322e-01 -8.24970722e-01
7.14121401e-01 2.01110438e-01 -1.66989172e+00 6.76219940e-01
6.27704918e-01 3.95753533e-01 -8.89768228e-02 4.16941524e-01
7.27330506e-01 -1.05469704e+00 -1.31312633e+00 6.77917957e-01
2.59930849e-01 -7.25537717e-01 7.79688358e-01 -8.36312473e-02
1.42403498e-01 -1.93521589e-01 1.22970611e-01 -1.27430272e+00
-3.03874016e-01 -2.90322721e-01 4.42559183e-01 4.69506472e-01
6.21726871e-01 -6.06679201e-01 1.19851959e+00 6.18117929e-01
-8.60073984e-01 -1.21594870e+00 -9.49580431e-01 -2.25361511e-01
4.51672047e-01 -3.60884219e-01 3.28486562e-01 8.63646805e-01
-3.04212779e-01 -1.41641110e-01 5.79344273e-01 2.56237149e-01
9.28123593e-01 -8.21063761e-03 1.99504480e-01 -1.08143592e+00
4.07713234e-01 -9.08328056e-01 -3.92511070e-01 -6.28072560e-01
-1.35288790e-01 -1.51105249e+00 -1.95867896e-01 -2.23613334e+00
2.50027418e-01 -5.65495908e-01 -1.88683435e-01 3.61776948e-01
-1.53353602e-01 1.22133330e-01 -1.83001474e-01 3.01151305e-01
-7.55739212e-02 2.49435052e-01 2.11606741e+00 -1.28285065e-01
-7.29865909e-01 5.48561886e-02 -5.19918799e-01 6.85590863e-01
1.00003576e+00 -3.84042054e-01 -2.73138106e-01 -8.32840279e-02
-5.51945031e-01 4.52446610e-01 1.91323340e-01 -9.27475631e-01
3.05460185e-01 -1.12899311e-01 7.94849634e-01 -8.04217160e-01
-9.87216979e-02 -8.61710012e-01 1.72684148e-01 9.29511964e-01
-2.00088084e-01 -4.06701237e-01 2.64203191e-01 3.75814497e-01
1.02470934e-01 -2.20069095e-01 1.18262851e+00 -4.50558305e-01
-3.18313986e-01 8.18248630e-01 -4.14357126e-01 -1.43293351e-01
1.65624118e+00 -6.30443990e-01 3.07180852e-01 5.42723872e-02
-1.25231111e+00 3.91834259e-01 2.15662464e-01 -2.87937552e-01
8.14135015e-01 -1.24734628e+00 -9.33455229e-01 3.05067599e-01
-3.20861101e-01 8.09562266e-01 4.86868888e-01 1.52438033e+00
-1.12767911e+00 1.82779729e-01 -1.67456698e-02 -1.13225961e+00
-1.08351147e+00 1.72060311e-01 1.07445014e+00 -5.78110516e-01
-1.02444470e+00 1.03710747e+00 3.34244579e-01 -2.21801624e-01
9.64725912e-02 -6.56233966e-01 -3.88183624e-01 -6.56494871e-02
-9.46981907e-02 1.39232278e-01 9.68861673e-03 -6.61649108e-01
-3.70556176e-01 7.60866761e-01 -1.12096734e-01 7.91749079e-03
1.09207237e+00 -9.07543972e-02 -2.87642241e-01 -1.97897106e-01
1.28953016e+00 -4.69550163e-01 -1.21559048e+00 -4.07838412e-02
2.63119582e-02 -3.96068208e-02 4.17629987e-01 -7.82363534e-01
-1.55051947e+00 6.21409416e-01 9.02067363e-01 2.27905124e-01
1.18067396e+00 1.85043961e-01 8.61012995e-01 -4.09522414e-01
-7.62574151e-02 -6.02060676e-01 1.07772157e-01 2.59991169e-01
7.96607971e-01 -1.05237901e+00 -1.12137876e-01 -5.94115913e-01
-6.00831568e-01 1.28042209e+00 7.96854436e-01 -5.08223116e-01
8.26916575e-01 4.23411191e-01 1.40623689e-01 -4.75819170e-01
-2.32045516e-01 1.59703135e-01 3.85097861e-01 3.07085156e-01
3.10956866e-01 3.80890578e-01 -3.27526689e-01 4.58272308e-01
-9.99027491e-02 -1.61748752e-01 3.16512764e-01 7.85775125e-01
-6.79867566e-01 -5.31656682e-01 -3.26946318e-01 8.01905513e-01
-4.17135090e-01 1.15966015e-01 -2.51446553e-02 9.87408161e-01
-1.37334868e-01 4.42742169e-01 1.91993773e-01 5.36843725e-02
1.33143291e-01 -1.85652062e-01 4.30321783e-01 -5.60486257e-01
-6.30704582e-01 8.53438377e-01 -2.20203266e-01 -3.81128818e-01
-6.43800944e-02 -6.14156961e-01 -1.95691419e+00 1.80029914e-01
-3.44446570e-01 1.95492566e-01 8.41398418e-01 7.12544143e-01
-1.39854282e-01 7.86680400e-01 6.19166136e-01 -7.00664937e-01
-8.10389742e-02 -9.48523283e-01 -6.02873027e-01 2.09756464e-01
1.09335810e-01 -5.41418791e-01 -3.50973874e-01 -1.51312232e-01] | [14.558305740356445, -2.5583221912384033] |
d852ddab-8456-4467-b824-9eca5a0cb85c | igformer-interaction-graph-transformer-for | 2207.12100 | null | https://arxiv.org/abs/2207.12100v1 | https://arxiv.org/pdf/2207.12100v1.pdf | IGFormer: Interaction Graph Transformer for Skeleton-based Human Interaction Recognition | Human interaction recognition is very important in many applications. One crucial cue in recognizing an interaction is the interactive body parts. In this work, we propose a novel Interaction Graph Transformer (IGFormer) network for skeleton-based interaction recognition via modeling the interactive body parts as graphs. More specifically, the proposed IGFormer constructs interaction graphs according to the semantic and distance correlations between the interactive body parts, and enhances the representation of each person by aggregating the information of the interactive body parts based on the learned graphs. Furthermore, we propose a Semantic Partition Module to transform each human skeleton sequence into a Body-Part-Time sequence to better capture the spatial and temporal information of the skeleton sequence for learning the graphs. Extensive experiments on three benchmark datasets demonstrate that our model outperforms the state-of-the-art with a significant margin. | ['Jun Liu', 'James Bailey', 'Hossein Rahmani', 'Qiuhong Ke', 'Yunsheng Pang'] | 2022-07-25 | null | null | null | null | ['human-interaction-recognition'] | ['computer-vision'] | [ 1.72333941e-01 2.11884543e-01 -3.77892137e-01 -3.49152952e-01
2.50722587e-01 -1.00278236e-01 3.50218773e-01 -2.00965270e-01
1.08112954e-01 2.28838578e-01 5.58763087e-01 3.42714727e-01
-3.22655857e-01 -7.65493631e-01 -4.28580552e-01 -4.18860555e-01
-2.23678544e-01 6.23804510e-01 5.09303927e-01 -3.03697228e-01
-1.13394402e-01 2.11540818e-01 -1.40558434e+00 1.96353152e-01
8.30521524e-01 9.97472346e-01 -1.34994432e-01 3.10594559e-01
-7.16204867e-02 6.70346379e-01 -3.67150784e-01 -4.33615416e-01
9.94360894e-02 -7.95718431e-01 -8.12876999e-01 4.63400006e-01
9.15477574e-02 -3.03591732e-02 -6.55308783e-01 8.41453433e-01
2.49976262e-01 2.88074702e-01 6.51695490e-01 -1.43055904e+00
-2.87048161e-01 6.87074244e-01 -8.23453426e-01 -4.57147546e-02
9.88210142e-01 8.44379067e-02 1.01762879e+00 -5.44919789e-01
6.57060981e-01 1.59747386e+00 5.30291677e-01 3.15567583e-01
-7.50002444e-01 -7.45471358e-01 4.44640011e-01 4.92274761e-01
-1.22652972e+00 1.25732645e-01 1.03095257e+00 -5.30455649e-01
5.90599239e-01 7.73484260e-02 1.27562773e+00 8.15362453e-01
2.63660491e-01 1.24231946e+00 3.91093910e-01 -2.15002716e-01
-2.59001851e-01 -6.17755532e-01 4.08050179e-01 1.17660785e+00
1.29009724e-01 -3.05390209e-01 -7.03970373e-01 -5.58867939e-02
9.28269982e-01 4.62799430e-01 -2.57415861e-01 -6.91483498e-01
-1.15417147e+00 4.78044599e-01 5.83542526e-01 2.53296494e-01
-2.79028237e-01 1.03604645e-01 6.56944573e-01 -1.20929606e-01
2.05746412e-01 -2.91520774e-01 -1.29703447e-01 6.41807243e-02
-4.41125840e-01 1.30415127e-01 8.75599384e-01 8.92493904e-01
4.73126620e-01 -2.65795767e-01 -2.77706593e-01 8.79784405e-01
6.99750900e-01 3.47702861e-01 3.84553224e-01 -6.80116296e-01
5.69001734e-01 1.26933694e+00 -4.58427161e-01 -1.39115977e+00
-4.42991972e-01 -2.82932758e-01 -1.05564547e+00 -1.93851441e-01
2.73665220e-01 2.30655059e-01 -8.24565589e-01 1.54642653e+00
5.69610715e-01 3.03272575e-01 -2.85133421e-01 1.02244484e+00
9.64136004e-01 1.89813882e-01 1.36161402e-01 2.64640171e-02
1.60804081e+00 -1.14792979e+00 -7.02951968e-01 -2.85988986e-01
4.36706185e-01 -1.78406179e-01 7.47413218e-01 2.24392548e-01
-7.32533157e-01 -8.75340879e-01 -1.03997386e+00 1.60916775e-01
-2.60542538e-02 2.74824113e-01 7.62820005e-01 2.42748633e-01
-3.63097817e-01 6.36382222e-01 -9.23954606e-01 -5.30065596e-01
2.70010114e-01 4.76313055e-01 -6.40255690e-01 1.09345485e-02
-1.24271667e+00 4.58190382e-01 4.51048195e-01 2.79141188e-01
-4.00532156e-01 -2.09813446e-01 -1.32942259e+00 -4.15115543e-02
4.81596589e-01 -1.00702345e+00 7.35210121e-01 -9.43197191e-01
-1.31726217e+00 6.83777392e-01 -8.24160874e-02 -3.15110266e-01
5.97873688e-01 -3.46538097e-01 -4.12463993e-01 2.62565494e-01
-6.53063804e-02 3.61632973e-01 6.90607607e-01 -1.05302250e+00
-6.04376137e-01 -7.65943944e-01 -5.45233153e-02 4.87922221e-01
-1.84897915e-01 -1.39554068e-01 -1.07799065e+00 -6.33831322e-01
2.89148301e-01 -9.95584071e-01 -9.08521414e-02 5.00963069e-02
-5.39912641e-01 -3.90276432e-01 9.39448178e-01 -9.89456832e-01
1.32455766e+00 -2.13155007e+00 6.15565777e-01 4.69651014e-01
5.11701822e-01 1.29476383e-01 -8.46255571e-03 3.37590694e-01
-1.91110715e-01 -3.81602913e-01 -2.28315338e-01 -2.72178471e-01
1.72021091e-02 3.31173331e-01 2.37664178e-01 4.18336421e-01
-4.18639600e-01 1.10582578e+00 -1.04261672e+00 -8.83760750e-01
4.52704847e-01 5.62933743e-01 -3.47285122e-01 3.17778021e-01
1.67329028e-01 6.50318563e-01 -5.68472147e-01 5.12849927e-01
3.17242593e-01 -2.42639378e-01 5.18435657e-01 -4.00947154e-01
6.55335307e-01 -1.37626797e-01 -1.12291563e+00 1.90546203e+00
2.06864830e-02 1.06223010e-01 -1.63348839e-01 -1.01646304e+00
1.03239965e+00 1.08422764e-01 8.27278197e-01 -5.23945808e-01
2.24982947e-01 -2.02190518e-01 3.40876244e-02 -5.12931943e-01
-5.15780598e-02 1.57397032e-01 -8.19375068e-02 9.66235027e-02
1.88500807e-01 4.14078861e-01 2.28502885e-01 2.96924353e-01
1.16183925e+00 3.56388867e-01 6.22404754e-01 9.71937105e-02
8.70491624e-01 -5.36830902e-01 6.40260816e-01 9.67226624e-02
-2.12972626e-01 3.88956964e-01 3.76239389e-01 -4.97897625e-01
-3.09383690e-01 -1.19229877e+00 5.78729570e-01 8.45659435e-01
5.81962168e-01 -7.55397975e-01 -8.24323475e-01 -1.15439320e+00
8.30591619e-02 1.69595614e-01 -7.97196150e-01 -5.58634400e-01
-5.69318652e-01 -1.31793052e-01 3.83437574e-01 1.01892543e+00
8.49694192e-01 -1.14553034e+00 -3.47637981e-01 3.02889813e-02
-5.00759542e-01 -9.71589684e-01 -8.56070518e-01 -4.38565046e-01
-8.39563608e-01 -1.39329612e+00 -6.41388118e-01 -9.50574577e-01
8.02534819e-01 1.88798800e-01 8.89214456e-01 4.23793137e-01
-3.40504050e-01 4.31065917e-01 -4.16160852e-01 5.35676181e-02
1.35011092e-01 -1.11112423e-01 -7.28302449e-02 3.91793936e-01
2.31789589e-01 -7.55952954e-01 -7.46477425e-01 6.18021190e-01
-6.12423658e-01 3.95953685e-01 5.60004592e-01 7.86096632e-01
5.36475360e-01 2.53370553e-01 -4.49484251e-02 -1.04077327e+00
3.72116715e-01 -2.94202805e-01 9.41501483e-02 3.82907927e-01
-2.50784874e-01 4.00923640e-02 5.46828151e-01 -5.19641459e-01
-1.08120608e+00 4.75077450e-01 8.98387879e-02 -6.53211653e-01
-9.43654031e-02 5.40615737e-01 -6.39646709e-01 6.45961165e-02
2.67182142e-01 3.31499696e-01 9.78182256e-02 -3.98380429e-01
3.03970009e-01 2.88224727e-01 8.09293211e-01 -6.17852151e-01
6.81285679e-01 5.15206456e-01 1.26347095e-01 -6.19795740e-01
-7.51510561e-01 -7.81967998e-01 -1.04354572e+00 -6.86818182e-01
1.04301739e+00 -6.08559430e-01 -9.99420285e-01 5.82056224e-01
-9.83560443e-01 -1.16192661e-01 -2.33324274e-01 4.41662312e-01
-6.34456873e-01 9.08077478e-01 -6.54111087e-01 -7.06363857e-01
-2.61865139e-01 -8.63334417e-01 1.23779655e+00 3.68073821e-01
-4.89448607e-01 -9.76639509e-01 1.21128388e-01 6.69952512e-01
-4.83098656e-01 5.11283100e-01 6.49980783e-01 -5.19155264e-01
-2.71884292e-01 -3.58733892e-01 -2.15941057e-01 4.23853137e-02
3.07931006e-01 -5.91632947e-02 -3.10125291e-01 2.87889317e-02
-3.03704679e-01 2.53835954e-02 8.24323833e-01 3.19639236e-01
1.09685588e+00 -1.65291786e-01 -8.09900343e-01 4.63441581e-01
7.60749459e-01 2.42294729e-01 7.29419291e-01 -2.16857731e-01
1.37349737e+00 8.09033394e-01 8.26499701e-01 6.04702592e-01
6.10710382e-01 8.35640371e-01 2.77310163e-01 -1.89225003e-01
-3.18778247e-01 -6.66715205e-01 3.48202050e-01 9.47674334e-01
-5.36352813e-01 -1.60534859e-01 -8.73206794e-01 2.80945659e-01
-2.38172483e+00 -9.37748373e-01 -3.18545789e-01 1.99139190e+00
3.87667000e-01 2.59222746e-01 4.31095511e-01 1.99834585e-01
9.46656823e-01 1.27478406e-01 -5.16557992e-01 2.60392785e-01
2.85973132e-01 8.55647922e-02 9.36110467e-02 8.54909569e-02
-1.04997623e+00 8.61315727e-01 5.76252365e+00 7.17928767e-01
-5.33160090e-01 -2.39823207e-01 2.36823857e-01 5.18853605e-01
1.41988039e-01 -7.39710480e-02 -4.97171819e-01 5.30336142e-01
2.16344386e-01 -1.42929152e-01 2.90815949e-01 8.33677530e-01
3.34177800e-02 7.04991668e-02 -1.18992043e+00 1.24982691e+00
3.55778992e-01 -5.86474180e-01 3.18969876e-01 3.61112729e-02
2.00204775e-01 -7.28828728e-01 -3.47956300e-01 2.30374888e-01
3.87535244e-01 -7.72917032e-01 5.16450703e-01 9.85649407e-01
4.65081960e-01 -6.85543120e-01 8.53003860e-01 3.30654472e-01
-2.06597662e+00 5.37566356e-02 1.46155193e-01 -1.69863075e-01
1.94132909e-01 3.77540171e-01 -4.41679627e-01 1.06829715e+00
6.05436265e-01 1.26927269e+00 -5.60070753e-01 9.18587506e-01
-5.38719416e-01 4.61542696e-01 -1.70697123e-01 2.22939059e-01
-2.82654822e-01 -5.38852453e-01 3.11639696e-01 1.00187767e+00
5.06453253e-02 3.49487305e-01 6.73708916e-01 5.45523882e-01
9.39919800e-02 2.19884455e-01 -5.73481917e-01 -5.15176170e-02
1.57500029e-01 1.14165735e+00 -8.69889796e-01 -3.42873633e-01
-4.67585117e-01 1.34005678e+00 3.89062852e-01 1.64855093e-01
-1.00397897e+00 -4.79920477e-01 4.15624470e-01 1.76334113e-01
1.06215291e-01 -2.75179833e-01 -3.85143384e-02 -1.13732815e+00
1.56039074e-01 -6.97709322e-01 9.51887012e-01 -6.73171341e-01
-1.27517128e+00 2.52627313e-01 7.93093592e-02 -1.23870993e+00
-1.98430404e-01 -3.94324630e-01 -5.80210149e-01 5.40267289e-01
-5.03875673e-01 -1.51635587e+00 -9.27386999e-01 8.75627279e-01
5.73560357e-01 6.81912452e-02 4.22000319e-01 1.33571252e-01
-5.29884994e-01 5.17667472e-01 -6.75035000e-01 7.23268688e-01
4.20753062e-01 -1.06702232e+00 4.81416613e-01 6.10969841e-01
2.60603845e-01 6.94058180e-01 4.21146005e-01 -1.09346473e+00
-1.26755691e+00 -9.44474816e-01 5.38795054e-01 -1.98420614e-01
3.72308880e-01 -2.89959282e-01 -7.79795408e-01 8.12524498e-01
-1.16453990e-01 6.72519347e-03 6.37193024e-01 3.02778155e-01
-4.95163471e-01 -8.71010777e-03 -8.37951958e-01 6.12597942e-01
1.87647212e+00 -5.14463603e-01 -8.93800318e-01 1.18367136e-01
4.51493323e-01 -4.82870221e-01 -8.44471991e-01 7.24788487e-01
9.82690871e-01 -8.80187273e-01 1.19404542e+00 -3.46776307e-01
4.15399253e-01 -4.24720258e-01 2.09495276e-01 -1.18806779e+00
-5.89486599e-01 -1.82976872e-01 -4.39603120e-01 1.07185483e+00
-1.71856567e-01 -4.66939867e-01 9.42370474e-01 3.12648237e-01
3.24445665e-02 -7.05121040e-01 -5.03959060e-01 -7.50012159e-01
-8.70776176e-01 -2.72781879e-01 4.45867032e-01 8.58302474e-01
3.70713741e-01 5.82839072e-01 -4.56385225e-01 -1.10714547e-02
6.34175301e-01 1.55426368e-01 1.04480016e+00 -1.47053862e+00
-6.33356750e-01 -2.84365654e-01 -1.21012664e+00 -1.29865074e+00
3.05525094e-01 -9.02926683e-01 3.32806110e-02 -1.80276155e+00
6.13060594e-01 1.88865483e-01 -2.69495845e-01 6.22766912e-01
-3.91483247e-01 1.14014998e-01 2.44931161e-01 1.57276273e-01
-9.73915756e-01 7.08862305e-01 1.46634519e+00 -4.54751849e-01
-1.96673051e-01 9.70750675e-02 -3.02789450e-01 9.80367124e-01
4.34797913e-01 -4.62514861e-03 -6.66216493e-01 8.42353031e-02
-3.30438405e-01 3.17161292e-01 3.96280080e-01 -1.38023603e+00
3.30244541e-01 -9.96866748e-02 5.14108717e-01 -7.23350585e-01
3.43285054e-01 -8.80937040e-01 6.19727492e-01 5.77380836e-01
-2.88551956e-01 -3.93956453e-01 -1.78267553e-01 9.55445349e-01
-2.86712497e-01 3.27511162e-01 4.48897183e-01 -1.20130554e-01
-7.01142728e-01 5.50660193e-01 2.64562387e-02 -6.45431131e-02
1.15842330e+00 -4.50483948e-01 8.56647417e-02 -5.01284838e-01
-1.12418377e+00 3.96240771e-01 2.66041785e-01 6.17448807e-01
7.65398026e-01 -1.52454507e+00 -2.42611498e-01 9.06740427e-02
3.60313505e-01 -1.43518001e-01 4.30110127e-01 8.29612315e-01
-2.61785537e-01 7.89317638e-02 -4.09242004e-01 -6.19735956e-01
-1.93993533e+00 4.62555766e-01 2.84228623e-01 -5.05757928e-01
-8.85270000e-01 6.65165424e-01 7.89019346e-01 -4.35557753e-01
3.25977325e-01 -2.39041626e-01 -5.11179447e-01 -1.67888030e-01
4.01722580e-01 4.56221014e-01 -3.95387173e-01 -1.04613292e+00
-5.09653628e-01 8.28486621e-01 1.06562063e-01 1.22094207e-01
9.20603514e-01 5.25864922e-02 -1.71323195e-01 4.42569017e-01
1.05356908e+00 -1.81230783e-01 -1.10896325e+00 -2.61213362e-01
-6.38994351e-02 -5.22786319e-01 -5.05282760e-01 -5.26261270e-01
-1.42350161e+00 7.76625037e-01 3.50934446e-01 -2.10641325e-01
1.24150920e+00 2.43152007e-01 1.00219691e+00 9.52800810e-02
6.05658531e-01 -8.45886409e-01 4.10336256e-01 4.10025716e-01
8.86679292e-01 -7.37785339e-01 7.36251622e-02 -1.10939658e+00
-7.72796810e-01 9.94619012e-01 8.05284321e-01 -1.26439855e-01
8.03150475e-01 -6.98386952e-02 -2.30416164e-01 -6.10762358e-01
-9.61989984e-02 -3.48453820e-01 8.23355734e-01 7.25671709e-01
3.95729572e-01 3.60277951e-01 -5.67001998e-01 9.62766886e-01
-3.47372144e-02 4.29611057e-02 -3.94823164e-01 7.71965921e-01
-2.87814826e-01 -1.21278477e+00 -2.20549807e-01 5.34217179e-01
5.06942645e-02 3.74205261e-01 -9.04936016e-01 6.19759738e-01
3.72957736e-01 7.49812961e-01 -1.62148774e-01 -1.00458205e+00
5.47010481e-01 6.32812902e-02 5.30350804e-01 -6.99315965e-01
-5.26440918e-01 2.01024320e-02 -5.89076132e-02 -1.00689459e+00
-4.28955942e-01 -7.54405320e-01 -1.78795493e+00 -9.70002487e-02
-1.63080484e-01 1.80014238e-01 9.30750445e-02 1.10550773e+00
2.79770762e-01 7.40219116e-01 3.81536782e-01 -8.65095794e-01
4.80595715e-02 -8.58396411e-01 -8.63991678e-01 9.63141322e-01
-3.00533861e-01 -1.16612184e+00 -5.42649068e-02 2.28004143e-01] | [8.07697868347168, 0.47686266899108887] |
c2386f3f-46d6-4434-89ea-0fd5073fd854 | realsmilenet-a-deep-end-to-end-network-for | 2010.03203 | null | https://arxiv.org/abs/2010.03203v1 | https://arxiv.org/pdf/2010.03203v1.pdf | RealSmileNet: A Deep End-To-End Network for Spontaneous and Posed Smile Recognition | Smiles play a vital role in the understanding of social interactions within different communities, and reveal the physical state of mind of people in both real and deceptive ways. Several methods have been proposed to recognize spontaneous and posed smiles. All follow a feature-engineering based pipeline requiring costly pre-processing steps such as manual annotation of face landmarks, tracking, segmentation of smile phases, and hand-crafted features. The resulting computation is expensive, and strongly dependent on pre-processing steps. We investigate an end-to-end deep learning model to address these problems, the first end-to-end model for spontaneous and posed smile recognition. Our fully automated model is fast and learns the feature extraction processes by training a series of convolution and ConvLSTM layer from scratch. Our experiments on four datasets demonstrate the robustness and generalization of the proposed model by achieving state-of-the-art performances. | ['Shafin Rahman', 'Tom Gedeon', 'Md Zakir Hossain', 'Yan Yang'] | 2020-10-07 | null | null | null | null | ['smile-recognition'] | ['computer-vision'] | [ 1.75358076e-02 1.03283748e-02 4.06581044e-01 -8.53102326e-01
-2.57018626e-01 -2.22052038e-01 8.80623639e-01 -3.33224952e-01
-2.95979291e-01 4.20994252e-01 1.82427224e-02 1.33566573e-01
2.19452053e-01 -4.85378802e-01 -3.29976380e-01 -4.94738489e-01
-1.80945784e-01 4.82867360e-01 -2.36289665e-01 -1.27369221e-02
6.34684622e-01 7.66092539e-01 -1.35816455e+00 3.76286536e-01
5.24690866e-01 1.20958340e+00 -5.28301597e-01 7.99848974e-01
-1.37422821e-02 9.12243366e-01 -6.23558223e-01 -7.74321377e-01
1.52655810e-01 -3.82105649e-01 -7.37206817e-01 2.76739031e-01
6.08482897e-01 -5.21270156e-01 -3.41780663e-01 8.98622155e-01
8.21874380e-01 -2.90178601e-02 3.86631787e-01 -1.23212433e+00
-5.67765594e-01 7.01904669e-02 -6.09773934e-01 -1.48440197e-01
4.93963987e-01 4.41102415e-01 4.16076332e-01 -1.02227986e+00
1.92565769e-01 1.40289986e+00 7.24369705e-01 9.21802700e-01
-1.06911290e+00 -9.55434442e-01 -3.43152523e-01 1.78046584e-01
-1.24643695e+00 -1.09450424e+00 1.06930649e+00 -7.19446242e-01
6.20561898e-01 1.44340739e-01 7.21801519e-01 1.41091621e+00
-2.65504569e-01 9.23391402e-01 1.20702863e+00 -4.98015173e-02
1.73039183e-01 3.09134722e-01 -7.29922652e-02 8.22758377e-01
-2.31791988e-01 5.34724295e-02 -6.94866955e-01 -3.63261998e-01
6.32883966e-01 5.62984459e-02 7.46091828e-02 -3.07627589e-01
-6.68965697e-01 7.85573065e-01 2.13265568e-01 4.40684408e-02
-2.02951789e-01 2.13082403e-01 6.39901519e-01 3.19984764e-01
6.36912704e-01 2.61805683e-01 -9.07220505e-03 -2.79335052e-01
-1.47502017e+00 1.11381762e-01 7.63062418e-01 2.22115800e-01
3.96613568e-01 -1.35550037e-01 1.47205219e-01 7.69276917e-01
3.48830193e-01 4.79079038e-01 2.85548836e-01 -6.21099055e-01
2.44937897e-01 6.16085887e-01 1.41862288e-01 -1.36196399e+00
-6.01199210e-01 -4.56587464e-01 -8.51343095e-01 3.40585411e-01
4.22743887e-01 -5.74307926e-02 -7.00887024e-01 1.62300074e+00
4.68864858e-01 5.60795903e-01 -3.40797335e-01 8.57171953e-01
1.00588632e+00 2.48335689e-01 -1.72081143e-01 6.38647154e-02
9.02959347e-01 -7.80459881e-01 -4.02719885e-01 -2.16357917e-01
3.38847309e-01 -6.93533957e-01 7.97694325e-01 4.81278211e-01
-1.18204296e+00 -3.62958997e-01 -7.84457147e-01 -1.09449379e-01
-2.50459641e-01 2.34001547e-01 6.48808658e-01 9.30073678e-01
-9.82914627e-01 9.18895543e-01 -7.30387390e-01 -3.40971649e-01
1.08973801e+00 6.43728435e-01 -5.22260547e-01 3.56397092e-01
-7.04470456e-01 9.18976188e-01 -2.22239494e-01 5.73980510e-01
-9.21059251e-01 -4.16730672e-01 -8.63226831e-01 8.72383341e-02
-3.36175524e-02 -6.21889353e-01 1.08940136e+00 -1.18849444e+00
-2.00729489e+00 1.28172982e+00 -2.92459518e-01 -2.90659606e-01
8.90835285e-01 -4.20960844e-01 -3.46529007e-01 2.45052442e-01
-4.63151008e-01 3.51250768e-01 1.45322466e+00 -1.13113987e+00
1.30013615e-01 -7.06230342e-01 -1.37293607e-01 -8.57228935e-02
-3.57468396e-01 5.47776043e-01 2.44889483e-01 -1.22039840e-01
-8.51922482e-02 -7.46560276e-01 2.39972427e-01 3.69711846e-01
-4.58408773e-01 -8.54570270e-02 1.15678692e+00 -9.07736003e-01
7.62123585e-01 -2.17353058e+00 -1.95075154e-01 1.00721195e-01
5.61258078e-01 6.72990382e-01 -1.45840064e-01 2.42643833e-01
-9.82408300e-02 -5.95534146e-02 2.81599001e-03 -9.67118859e-01
2.06480414e-01 -3.86887431e-01 -2.98236191e-01 9.35402572e-01
2.24625215e-01 1.05991673e+00 -8.31591010e-01 -4.07099277e-01
4.86781329e-01 7.85407960e-01 -4.46823776e-01 5.21066308e-01
3.27480555e-01 7.97515392e-01 -1.12808779e-01 7.42007732e-01
9.67508554e-01 -1.23338968e-01 6.25492707e-02 1.68840691e-01
2.36645505e-01 4.14971143e-01 -8.08903396e-01 1.36034656e+00
-6.01959825e-01 7.69482493e-01 3.51918668e-01 -9.38257158e-01
1.02215755e+00 2.92040646e-01 2.52480477e-01 -3.99829626e-01
4.76028353e-01 3.76393616e-01 -1.01534888e-01 -7.53745377e-01
1.74980074e-01 -5.05564868e-01 2.06009876e-02 5.08323729e-01
1.26609206e-01 -1.03762567e-01 -4.82941419e-01 -1.25158653e-01
7.31614888e-01 -3.24144564e-03 3.38639617e-01 -3.08564119e-03
8.97386789e-01 -8.09637129e-01 3.01093310e-01 3.86132449e-01
-4.30821538e-01 5.06315768e-01 5.39579988e-01 -8.84608567e-01
-7.50707388e-01 -9.27713633e-01 -3.79789923e-03 9.85184491e-01
-1.11112759e-01 1.04251601e-01 -8.12406301e-01 -8.26886892e-01
-2.69722688e-04 5.93134582e-01 -8.87206018e-01 -1.93085194e-01
-5.78833461e-01 -4.52451944e-01 6.76586747e-01 3.78352284e-01
6.84069693e-01 -1.20476198e+00 -7.49255538e-01 2.43169256e-03
7.77831972e-02 -9.64741826e-01 -2.62762725e-01 -5.32025874e-01
-5.33563852e-01 -1.10665333e+00 -5.60263932e-01 -3.87878656e-01
7.31979430e-01 -1.63834728e-02 9.15401042e-01 3.13777059e-01
-3.07111651e-01 -5.34845069e-02 8.86647403e-02 -2.38219157e-01
-1.32068649e-01 -2.87149787e-01 1.53778821e-01 6.38788283e-01
5.74617326e-01 -9.30098653e-01 -6.60198987e-01 3.59754562e-01
-4.28959459e-01 -1.10317953e-01 2.53457725e-01 9.07493949e-01
-2.69812018e-01 -6.86167777e-01 5.06383836e-01 -5.24298370e-01
7.34786987e-01 -1.65332466e-01 -4.47311223e-01 1.99504811e-02
2.21225262e-01 -3.68834466e-01 6.21081650e-01 -3.20253074e-01
-8.13192785e-01 2.44591817e-01 -2.10778490e-01 -4.72050428e-01
-5.31282306e-01 7.23920390e-02 -1.10776842e-01 -5.42465448e-01
4.30817902e-01 4.91730392e-01 3.46338838e-01 -5.85903466e-01
1.27690971e-01 1.10258901e+00 5.95049739e-01 -2.66167104e-01
8.47767055e-01 7.58071721e-01 8.43310058e-02 -1.09275913e+00
-9.19082224e-01 -2.98730135e-01 -9.92019832e-01 -2.88377583e-01
5.26810467e-01 -5.80305636e-01 -1.31543183e+00 1.06164432e+00
-1.34838438e+00 -2.28942618e-01 4.71195392e-02 3.71690355e-02
-7.24190116e-01 6.90177381e-01 -4.90183949e-01 -1.20048988e+00
-7.51993179e-01 -7.73629725e-01 1.09407985e+00 2.44276837e-01
-4.83449936e-01 -9.23434377e-01 1.52362496e-01 7.60144830e-01
6.08485937e-01 5.71457803e-01 4.06256795e-01 -8.63667667e-01
-1.91777483e-01 -5.74646473e-01 -4.05930132e-01 3.81959826e-01
-1.33050606e-01 -9.75154564e-02 -1.43741179e+00 -4.99772847e-01
1.03141993e-01 -8.11482966e-01 9.27590728e-01 1.11755930e-01
1.30236661e+00 -5.12129724e-01 -1.72202051e-01 7.77076244e-01
8.65914047e-01 -2.10191578e-01 7.42089808e-01 -1.59521103e-01
6.59297466e-01 8.40831757e-01 2.61942972e-03 5.80593765e-01
2.77692586e-01 5.80007613e-01 5.91204107e-01 -3.24152038e-02
1.23950645e-01 -2.42167786e-01 5.11059642e-01 3.01990837e-01
-1.63399298e-02 1.14635088e-01 -7.21927404e-01 5.77370346e-01
-1.65839589e+00 -1.37577748e+00 1.94992691e-01 2.19429636e+00
4.73304927e-01 -1.06672496e-01 5.18003643e-01 1.85950369e-01
5.47235250e-01 2.98877537e-01 -6.52047038e-01 -7.07128763e-01
4.25827861e-01 2.85279542e-01 -2.81250447e-01 6.15574062e-01
-1.18278348e+00 1.01107359e+00 6.15846634e+00 4.49471563e-01
-1.71462381e+00 -9.06023663e-03 6.62330449e-01 -4.95237559e-01
4.21386808e-01 -3.98931712e-01 -4.31926399e-01 6.33246064e-01
7.38934577e-01 1.96881101e-01 5.82234740e-01 9.06046808e-01
4.08087850e-01 1.64607406e-01 -1.21742749e+00 1.37771642e+00
3.54687601e-01 -1.17529881e+00 -4.33045596e-01 -3.84200402e-02
3.41652572e-01 -3.60151986e-04 1.91406816e-01 1.21071197e-01
4.83875535e-02 -1.52060580e+00 6.36776745e-01 5.59680521e-01
8.42276931e-01 -8.10301185e-01 8.07453930e-01 3.54329228e-01
-7.80087173e-01 -1.40512452e-01 -1.51834697e-01 -4.13268000e-01
2.50615060e-01 7.27592826e-01 -1.00173283e+00 2.18927283e-02
2.59774506e-01 7.04024136e-01 -4.42726582e-01 9.43551600e-01
-3.71479243e-01 3.97037327e-01 -3.90493840e-01 -1.40964255e-01
6.68967590e-02 1.35591790e-01 5.55258334e-01 1.35488820e+00
4.15180363e-02 -2.26093397e-01 -6.48718998e-02 1.09567010e+00
6.56659203e-03 -1.81801006e-01 -4.64104265e-01 -2.74341911e-01
8.45502317e-02 1.54161084e+00 -2.29846328e-01 -1.81112453e-01
-1.07226186e-01 1.13590086e+00 4.95855957e-01 1.26512181e-02
-6.74174368e-01 -4.80841637e-01 8.80820155e-01 4.64506716e-01
1.35535732e-01 -3.27387512e-01 -5.05162060e-01 -1.38531530e+00
1.73626482e-01 -7.20875025e-01 -3.90778221e-02 -2.90978134e-01
-1.38285708e+00 5.97767115e-01 -4.79971141e-01 -8.62298727e-01
-4.79041219e-01 -5.09828627e-01 -1.18459737e+00 7.31267691e-01
-1.47035444e+00 -1.51214099e+00 -6.22340739e-01 6.20549798e-01
1.97983965e-01 -2.81171262e-01 7.32681096e-01 2.36682430e-01
-5.73333561e-01 8.16560745e-01 -3.02006483e-01 4.34766531e-01
3.84223521e-01 -8.03471386e-01 6.64180934e-01 5.96621633e-01
-1.71413615e-01 7.38395035e-01 6.36322081e-01 -3.00690621e-01
-1.06091237e+00 -7.03096747e-01 1.20237100e+00 -5.46218812e-01
4.87315625e-01 -7.57024944e-01 -6.76905274e-01 4.64629203e-01
2.90799271e-02 3.65353286e-01 7.96586514e-01 2.32134968e-01
-3.46979499e-01 7.72070065e-02 -1.71090865e+00 2.99254656e-01
1.15438151e+00 -6.87278032e-01 -5.79014778e-01 1.33523658e-01
-2.48678893e-01 -2.37798691e-01 -3.28561455e-01 1.84196625e-02
9.00347531e-01 -1.42070901e+00 9.41421330e-01 -8.31381917e-01
6.51100338e-01 7.56609254e-03 3.01199496e-01 -1.17883563e+00
-1.35466531e-02 -1.00510323e+00 -3.63842010e-01 1.07462776e+00
3.11992746e-02 -7.27193594e-01 1.00642622e+00 7.25077569e-01
3.68576825e-01 -9.99215305e-01 -1.05714905e+00 -5.76602399e-01
-1.50531158e-01 -1.07139610e-01 5.43850422e-01 1.14847767e+00
3.43529493e-01 2.87781566e-01 -8.21489990e-01 -6.49195164e-02
9.38032746e-01 3.35160077e-01 1.30247724e+00 -1.51199841e+00
-2.11699054e-01 -3.19298774e-01 -8.19492161e-01 -1.02827072e+00
5.20030320e-01 -7.43332148e-01 -2.55762517e-01 -1.16133940e+00
4.17913347e-01 5.38287163e-02 1.63638238e-02 6.68668509e-01
1.86862960e-01 3.94263178e-01 1.99124455e-01 1.82555709e-02
-6.41967475e-01 7.79919505e-01 1.01527894e+00 -4.37407494e-02
-1.66517735e-01 1.01011880e-01 -4.64087278e-01 1.11955607e+00
8.31899881e-01 -3.32987458e-01 -5.93606085e-02 -2.26950601e-01
5.75768426e-02 -7.98634663e-02 1.00177181e+00 -9.33962464e-01
1.53065011e-01 -2.16212064e-01 6.30476356e-01 -5.35836816e-01
8.74842703e-01 -6.26868069e-01 -3.27026129e-01 5.85965276e-01
-1.76118091e-01 -5.18395782e-01 -5.26630394e-02 5.08805774e-02
-8.68744478e-02 -9.95128155e-02 1.41127229e+00 -2.33660534e-01
-1.26680881e-01 1.70908913e-01 9.62442309e-02 -7.58201480e-02
9.54783261e-01 -4.50028509e-01 -2.01114386e-01 -8.44717085e-01
-7.90969014e-01 -5.16270027e-02 4.43177998e-01 3.27784359e-01
8.00107181e-01 -1.11731339e+00 -8.27368617e-01 5.47582686e-01
-3.16388130e-01 -3.97543609e-01 2.39588588e-01 9.80328202e-01
-5.26011825e-01 1.63103312e-01 -4.57632929e-01 -4.10678118e-01
-1.53868830e+00 2.89476067e-01 6.49470210e-01 -1.61693841e-01
-4.29436058e-01 1.11959243e+00 7.78559223e-02 -7.05489635e-01
4.19999398e-02 2.33947024e-01 -7.49752149e-02 -6.65099695e-02
9.32101548e-01 5.72780192e-01 -1.80403143e-01 -8.54362667e-01
-5.81775844e-01 5.42904735e-01 -2.09691748e-01 1.76125318e-01
1.65689385e+00 -3.04963794e-02 -2.70430535e-01 2.05352390e-03
1.38772321e+00 -1.23239696e-01 -1.24808800e+00 -4.89821136e-02
-1.61766306e-01 -9.41835582e-01 3.52404676e-02 -7.76849747e-01
-1.10695207e+00 1.35909879e+00 3.21681798e-01 -4.39108908e-02
8.37867558e-01 -2.39994109e-01 1.09462917e+00 3.58168781e-01
4.13783073e-01 -1.00917828e+00 2.84417719e-01 4.46547478e-01
1.05941701e+00 -1.30018103e+00 -1.19604498e-04 -3.23385239e-01
-4.17257547e-01 1.22061646e+00 5.39441228e-01 -2.01886341e-01
5.83542883e-01 8.37070495e-02 1.37080282e-01 -3.59218687e-01
-4.05281305e-01 2.61471510e-01 2.89106786e-01 4.91962105e-01
2.31586233e-01 3.46953073e-03 -1.57492146e-01 5.79230189e-01
-3.60620081e-01 3.77003461e-01 2.00634912e-01 5.89757562e-01
-3.83861959e-01 -7.27255762e-01 -1.79627851e-01 1.45158708e-01
-4.55052197e-01 1.34488583e-01 -1.13504183e+00 2.94445038e-01
2.76583675e-02 9.32862401e-01 -2.46967915e-02 -5.38134634e-01
5.31246550e-02 2.79480755e-01 6.16940975e-01 -4.07145739e-01
-8.27176630e-01 -4.71584648e-01 1.57499909e-01 -7.51856744e-01
-9.50193703e-02 -6.68949604e-01 -8.84202957e-01 -6.85192108e-01
-2.07392231e-01 -3.54366958e-01 6.91379845e-01 1.09241009e+00
5.79959035e-01 -2.78574824e-01 1.05267894e+00 -1.48494792e+00
-8.41478407e-01 -1.14559114e+00 -4.22195613e-01 7.32461691e-01
5.75735152e-01 -7.05160141e-01 -4.56867874e-01 -2.91664571e-01] | [13.423590660095215, 1.6530399322509766] |
eb2c56d9-4148-4acb-b2f0-dd8c0c0b0e53 | key-value-information-extraction-from-full | 2304.13530 | null | https://arxiv.org/abs/2304.13530v1 | https://arxiv.org/pdf/2304.13530v1.pdf | Key-value information extraction from full handwritten pages | We propose a Transformer-based approach for information extraction from digitized handwritten documents. Our approach combines, in a single model, the different steps that were so far performed by separate models: feature extraction, handwriting recognition and named entity recognition. We compare this integrated approach with traditional two-stage methods that perform handwriting recognition before named entity recognition, and present results at different levels: line, paragraph, and page. Our experiments show that attention-based models are especially interesting when applied on full pages, as they do not require any prior segmentation step. Finally, we show that they are able to learn from key-value annotations: a list of important words with their corresponding named entities. We compare our models to state-of-the-art methods on three public databases (IAM, ESPOSALLES, and POPP) and outperform previous performances on all three datasets. | ['Christopher Kermorvant', 'Mélodie Boillet', 'Solène Tarride'] | 2023-04-26 | null | null | null | null | ['handwriting-recognition'] | ['computer-vision'] | [ 1.77565739e-01 1.89000756e-01 -3.90016645e-01 -2.18837157e-01
-1.08975554e+00 -1.05697668e+00 9.71055567e-01 3.63920778e-01
-6.26767695e-01 9.00903046e-01 3.32591861e-01 -3.51749480e-01
-2.14077994e-01 -6.06229901e-01 -6.34590566e-01 -2.80210942e-01
4.12544996e-01 1.05568099e+00 6.84191287e-01 -3.60295437e-02
9.27713513e-01 9.51880872e-01 -1.21370637e+00 6.96365952e-01
1.00194359e+00 8.80752921e-01 -3.68889309e-02 6.46980703e-01
-8.36435437e-01 1.01584756e+00 -9.17724073e-01 -8.01882267e-01
-1.95410520e-01 -1.08899996e-01 -1.30803323e+00 4.65804845e-01
5.47730982e-01 -2.07052022e-01 -7.40005523e-02 7.51195848e-01
3.57425660e-01 -2.56311417e-01 9.64328945e-01 -7.00780213e-01
-8.03687811e-01 9.05194819e-01 -4.17654335e-01 -1.42409518e-01
5.52956045e-01 -3.00299317e-01 1.06636643e+00 -1.12204754e+00
1.14078724e+00 7.76603818e-01 8.40738833e-01 2.64776796e-01
-9.72794950e-01 -5.55739664e-02 1.44887283e-01 3.15466672e-01
-1.12805688e+00 -4.35718477e-01 4.44796920e-01 -5.16980946e-01
1.34597778e+00 2.33021691e-01 2.39637434e-01 1.11269486e+00
-5.74762486e-02 1.42632627e+00 1.36611927e+00 -9.37466919e-01
9.47154984e-02 3.75691563e-01 8.11668277e-01 8.01709890e-01
3.55145723e-01 -6.62036419e-01 -6.66923881e-01 3.89103428e-03
6.37517214e-01 -5.01078308e-01 -3.16719711e-01 -4.02950406e-01
-1.29898059e+00 3.57544154e-01 -3.56744766e-01 7.78132200e-01
-5.49938500e-01 -1.92337289e-01 4.35803592e-01 -2.51289569e-02
2.48169973e-01 6.69084013e-01 -7.34221578e-01 -4.06346500e-01
-1.49037838e+00 5.76130375e-02 1.53304946e+00 1.46206188e+00
4.61814314e-01 -4.95129168e-01 -5.89041471e-01 7.96555698e-01
1.22644953e-01 2.95238286e-01 5.80107510e-01 -4.44231272e-01
8.41663599e-01 8.36541653e-01 3.01623672e-01 -6.70111656e-01
-3.74573648e-01 -1.61135003e-01 -2.61705399e-01 -2.22797543e-01
6.79524481e-01 1.20835632e-01 -1.26459348e+00 7.11629391e-01
-1.52283356e-01 -4.50635999e-01 -2.68127844e-02 4.08539951e-01
6.68183863e-01 3.55909735e-01 -1.54523596e-01 -1.81274995e-01
1.61201394e+00 -1.19086397e+00 -1.02807009e+00 -1.34061292e-01
2.83243120e-01 -9.78190303e-01 7.26550937e-01 7.75908411e-01
-1.07560551e+00 -2.92484164e-01 -8.51720929e-01 -2.65147686e-01
-8.97023320e-01 8.43046784e-01 5.17263830e-01 9.27866936e-01
-8.48570943e-01 8.38853061e-01 -7.91640520e-01 -6.03812575e-01
4.78625536e-01 3.18886727e-01 -3.76812965e-01 3.74785960e-01
-7.15971828e-01 1.06727993e+00 4.11388159e-01 1.49091288e-01
-5.61015129e-01 -2.98952222e-01 -4.99032706e-01 7.50000179e-02
6.13231897e-01 -1.07264794e-01 1.42210722e+00 -3.43451172e-01
-1.54763031e+00 1.06268668e+00 -2.97472686e-01 -4.32785630e-01
7.06330955e-01 -5.36306798e-01 -2.69407183e-01 2.87450880e-01
-1.68405786e-01 5.69735393e-02 5.95776320e-01 -1.28440261e+00
-6.52735114e-01 -4.11234707e-01 -3.66037846e-01 -2.01741368e-01
-4.69668299e-01 3.69458824e-01 -1.03546727e+00 -6.86995149e-01
-8.31822082e-02 -5.31348169e-01 8.27890709e-02 -4.93958414e-01
-9.10464346e-01 -3.14439207e-01 5.20824015e-01 -1.18637943e+00
1.41433394e+00 -1.64079607e+00 2.27888793e-01 2.85788149e-01
5.68397716e-02 5.30662358e-01 9.17243734e-02 7.75992155e-01
9.73990262e-02 4.24643189e-01 -2.31104225e-01 -4.97735620e-01
4.35630292e-01 2.55606491e-02 -4.40041572e-01 1.17904030e-01
4.51592058e-01 9.34047699e-01 -6.03154004e-01 -7.87015259e-01
-5.84741794e-02 2.31309578e-01 1.20126922e-02 -7.61660337e-02
-3.26770276e-01 -5.19679449e-02 -4.50742632e-01 8.63155544e-01
4.80945945e-01 -2.19628394e-01 4.09369051e-01 -7.40204304e-02
-4.99253660e-01 4.73890960e-01 -1.38468933e+00 1.42289281e+00
-3.80358607e-01 7.72389710e-01 -2.37387970e-01 -6.51533723e-01
1.13135028e+00 3.20217788e-01 8.38522688e-02 -3.54793042e-01
6.06266670e-02 6.75793529e-01 -7.39638865e-01 -6.20324433e-01
9.35866356e-01 4.98567492e-01 -2.38098383e-01 3.16554666e-01
5.77419519e-01 -9.45272744e-02 7.12766290e-01 9.57073942e-02
1.16902614e+00 3.77952844e-01 5.14851570e-01 -8.82532075e-02
6.76820576e-01 2.66173720e-01 2.61073202e-01 1.00784814e+00
1.88466185e-03 7.69230068e-01 9.71852779e-01 -1.11635208e-01
-1.15844190e+00 -7.28480101e-01 -1.21170148e-01 9.41650629e-01
-3.23963284e-01 -5.08649886e-01 -9.71679151e-01 -1.18078411e+00
-8.77995640e-02 6.49483204e-01 -6.25600159e-01 6.65897191e-01
-8.48375380e-01 -5.82854390e-01 9.16340470e-01 8.12021196e-01
4.31948811e-01 -1.22136569e+00 -2.80053169e-01 3.53366196e-01
-3.69396545e-02 -1.24048960e+00 -2.50334233e-01 6.21076763e-01
-7.36391723e-01 -1.12980974e+00 -1.18270481e+00 -9.04829979e-01
5.84312558e-01 -6.63546026e-01 1.28058589e+00 -3.30624610e-01
-2.09945112e-01 5.10919750e-01 -4.36655849e-01 -3.98270130e-01
-3.35911155e-01 7.88236737e-01 -4.65549439e-01 -4.34682183e-02
5.08306563e-01 -1.60100475e-01 1.32048503e-01 3.13129693e-01
-7.63387203e-01 -3.01308751e-01 1.08924532e+00 8.76012683e-01
4.80689198e-01 -2.48859644e-01 1.51659071e-01 -1.29351068e+00
6.80603445e-01 9.84991491e-02 -6.36030316e-01 9.32158053e-01
-6.52205765e-01 3.48071426e-01 3.87466848e-01 -2.68178940e-01
-1.36263692e+00 3.72448087e-01 -2.28081480e-01 2.61277836e-02
-6.20342076e-01 4.77532029e-01 -4.19232428e-01 3.27653438e-02
4.74979639e-01 6.29943669e-01 -4.48333561e-01 -1.06990659e+00
4.01882440e-01 1.11459053e+00 6.77250683e-01 -5.20169556e-01
5.98073781e-01 2.79778391e-01 -2.78410375e-01 -9.53874171e-01
-6.18007660e-01 -6.40645921e-01 -1.48713660e+00 4.43257205e-02
6.63321733e-01 -3.21318030e-01 -5.74613333e-01 1.00237346e+00
-1.52782822e+00 -8.45720842e-02 -4.79280204e-01 2.00565144e-01
-5.77522516e-01 7.56639898e-01 -9.24600005e-01 -8.10431898e-01
-3.54656637e-01 -7.04022050e-01 1.33436000e+00 3.57665807e-01
-2.64980227e-01 -9.64889944e-01 1.50334463e-01 3.19252372e-01
1.65983781e-01 -8.99338499e-02 8.84115100e-01 -1.47808301e+00
-4.78337437e-01 -3.75023842e-01 -2.39429533e-01 2.80820996e-01
7.50140548e-02 3.21572214e-01 -9.01467383e-01 2.67373145e-01
-4.15828377e-01 -9.32656378e-02 1.20749378e+00 -1.56667590e-01
8.47390831e-01 -3.45875621e-01 -5.87362111e-01 2.58418262e-01
1.29059958e+00 2.43230611e-01 1.00350428e+00 6.87002242e-01
5.74800134e-01 6.00009143e-01 5.88779747e-01 3.39650899e-01
1.17849492e-01 5.85773706e-01 -9.36590582e-02 4.83729653e-02
-2.16704458e-01 -1.69488728e-01 3.09867769e-01 7.22938001e-01
-3.01448554e-01 -4.60971296e-01 -1.24181521e+00 8.61245692e-01
-1.71297240e+00 -8.69574070e-01 -4.82631803e-01 1.85475218e+00
1.11172879e+00 2.75845408e-01 1.65526569e-01 4.10167098e-01
6.90708995e-01 5.75678945e-02 -7.27861300e-02 -3.81448746e-01
-3.10303330e-01 5.16196549e-01 7.50918806e-01 1.88827693e-01
-1.42916071e+00 1.18589103e+00 7.20945358e+00 9.62945104e-01
-8.10139775e-01 -1.21383518e-01 2.61708915e-01 4.79236960e-01
2.65870225e-02 -9.96361524e-02 -1.50800228e+00 2.13679120e-01
9.41445887e-01 2.03148454e-01 -9.22511220e-02 8.93554151e-01
-3.48366529e-01 -3.26825008e-02 -1.10638726e+00 4.68871802e-01
3.84178251e-01 -1.37369537e+00 1.46599725e-01 -5.10639101e-02
5.43093979e-01 -2.91379541e-01 -2.50532389e-01 8.66597425e-03
7.90111199e-02 -7.84422278e-01 8.86303425e-01 1.03058386e+00
4.75141793e-01 -5.06841421e-01 8.93978119e-01 3.14306282e-02
-9.69096661e-01 1.20300807e-01 -7.32453391e-02 5.39635837e-01
1.49610877e-01 5.05983531e-01 -9.05912995e-01 8.57708156e-01
2.84969658e-01 4.87308055e-01 -1.02603877e+00 1.34363878e+00
-5.16752839e-01 6.55043840e-01 -6.13465942e-02 -4.21915144e-01
2.01751620e-01 1.33268222e-01 3.74061793e-01 1.83917248e+00
1.04053929e-01 -3.13243270e-01 -2.35076413e-01 6.11435950e-01
-6.49135932e-02 4.42824900e-01 -2.77509540e-01 -4.60105628e-01
2.59778768e-01 1.23954701e+00 -1.13220561e+00 -6.53927982e-01
-3.42233509e-01 1.33123076e+00 1.69612169e-01 3.56677510e-02
-5.10284603e-01 -1.11591351e+00 -7.21490383e-02 -4.47209403e-02
1.02520192e+00 -3.26609433e-01 -4.50303942e-01 -1.20518935e+00
3.73132080e-01 -9.86892521e-01 2.68320471e-01 -4.79397446e-01
-1.15421712e+00 6.75624013e-01 -1.81876272e-01 -8.87213409e-01
-3.49865198e-01 -1.22584116e+00 -1.92710787e-01 6.31860673e-01
-1.44989014e+00 -1.35756087e+00 -2.54291091e-02 3.96963179e-01
5.41445613e-01 -1.98681250e-01 8.87045801e-01 3.38395089e-01
-5.95230520e-01 5.47832489e-01 3.68367344e-01 9.88689125e-01
6.75178826e-01 -1.70288658e+00 6.81652427e-01 8.52947712e-01
6.17738724e-01 6.76558912e-01 2.79415786e-01 -7.53767610e-01
-1.29535115e+00 -7.78634608e-01 1.54883742e+00 -8.74251187e-01
7.74462342e-01 -5.26989579e-01 -8.37415457e-01 7.42582858e-01
3.92531067e-01 -4.34284866e-01 5.96242905e-01 2.15051949e-01
-3.84284109e-01 2.73889691e-01 -8.57683420e-01 1.40514433e-01
8.52677941e-01 -4.77243900e-01 -1.10197711e+00 4.07293111e-01
2.11972728e-01 -4.53560263e-01 -9.44174290e-01 -9.53909233e-02
6.53292179e-01 -7.26561487e-01 6.19091630e-01 -7.85923839e-01
4.54370886e-01 -5.99167608e-02 9.00864452e-02 -8.60563874e-01
-1.68030653e-02 -7.87266791e-01 -3.54098022e-01 1.98210227e+00
9.23206389e-01 -4.20915157e-01 7.13359356e-01 4.35593903e-01
-1.87694523e-02 -5.37823558e-01 -7.64772594e-01 -9.86330390e-01
-5.38844094e-02 -3.50265861e-01 6.51602685e-01 8.47558141e-01
1.30539596e-01 2.62285292e-01 -2.25877866e-01 -1.14729583e-01
3.49095315e-01 2.20523849e-01 4.38886523e-01 -1.23993099e+00
-2.16902077e-01 -5.92712641e-01 -4.58656669e-01 -1.05821180e+00
1.16206661e-01 -6.44480348e-01 1.57453686e-01 -1.87566555e+00
2.85610229e-01 -1.06539212e-01 -1.44139826e-01 9.93841469e-01
9.68834758e-02 9.48045775e-02 2.19320327e-01 3.47702265e-01
-6.79165661e-01 1.79948062e-02 6.50566220e-01 -3.03113014e-01
-2.36158773e-01 -2.09120121e-02 -3.98918509e-01 8.26576889e-01
4.04043436e-01 -4.01773840e-01 5.20419955e-01 -3.57070982e-01
1.02646984e-01 -1.82515264e-01 -1.74667276e-02 -9.71457243e-01
6.20202899e-01 1.41742632e-01 6.22885525e-01 -9.73557651e-01
-5.48122004e-02 -5.68628132e-01 -4.13508862e-01 2.80723274e-01
-3.10469300e-01 -1.49347425e-01 9.23621804e-02 2.26343364e-01
-2.68611491e-01 -7.50731885e-01 3.71516973e-01 -1.38512492e-01
-8.93470705e-01 -3.76253337e-01 -8.17744434e-01 -1.65710151e-01
8.12483788e-01 -6.01018928e-02 -4.11071688e-01 1.07553937e-01
-9.19321179e-01 -1.83363095e-01 3.38212490e-01 4.28881019e-01
3.29092860e-01 -8.83949399e-01 -5.78179538e-01 -1.56723976e-01
4.71181013e-02 -4.73967910e-01 -3.73219460e-01 7.18389452e-01
-7.77141154e-01 9.48652208e-01 -2.49895692e-01 -2.97935337e-01
-1.35023820e+00 5.19911766e-01 7.11962804e-02 -9.30183828e-01
-3.28987598e-01 5.42733490e-01 -6.87992573e-01 -4.99830872e-01
2.87095129e-01 -5.93420625e-01 -6.08046174e-01 8.06760311e-01
5.22731543e-01 5.27259588e-01 6.08956516e-01 -4.22976315e-01
-5.75913072e-01 7.29204357e-01 -3.81524771e-01 -3.23655158e-01
1.40287387e+00 1.10257857e-01 -3.58955681e-01 4.75778908e-01
8.58047247e-01 5.14202714e-01 -7.20336974e-01 -1.76542208e-01
9.66462672e-01 -1.52429596e-01 -2.56003559e-01 -1.11331701e+00
-8.42319012e-01 7.19078660e-01 2.05834851e-01 2.17669845e-01
1.06215477e+00 5.37069328e-02 4.48186964e-01 8.98927867e-01
4.19053167e-01 -1.53742874e+00 -1.62973702e-01 8.52989912e-01
7.63255477e-01 -6.92028940e-01 1.88866049e-01 -6.19945884e-01
-5.57838082e-01 1.49926102e+00 1.11666255e-01 1.40186787e-01
2.79657453e-01 5.50543189e-01 -1.35432556e-01 -3.94905880e-02
-4.65841085e-01 -3.91250700e-01 6.17197573e-01 5.12271404e-01
5.51865578e-01 -3.85106832e-01 -6.12478137e-01 8.63402188e-01
1.06256627e-01 1.96188539e-01 6.78476810e-01 1.48739636e+00
-3.16674471e-01 -1.42877078e+00 -4.91476029e-01 4.10160989e-01
-7.87910759e-01 -2.61986911e-01 -1.17027628e+00 1.18901324e+00
-1.34600282e-01 6.28481746e-01 -1.10654220e-01 -3.19776572e-02
6.57414854e-01 6.65335596e-01 6.90942466e-01 -6.05103314e-01
-9.14244413e-01 -2.42136233e-02 4.26216543e-01 -3.73234063e-01
-3.48087341e-01 -9.39875543e-01 -1.01914692e+00 1.52319625e-01
-6.50164366e-01 3.29219550e-01 8.90379846e-01 8.69487524e-01
4.22098368e-01 4.22601819e-01 2.65251249e-01 -6.57002449e-01
-5.47526717e-01 -1.06681657e+00 -6.68802977e-01 -6.68659210e-02
-7.83688799e-02 -2.56115645e-01 4.21621371e-04 4.34603661e-01] | [11.719536781311035, 2.831719160079956] |
4b52bd75-275f-482e-ad6c-0457399ee99b | optimal-compression-for-minimizing | 2211.02012 | null | https://arxiv.org/abs/2211.02012v1 | https://arxiv.org/pdf/2211.02012v1.pdf | Optimal Compression for Minimizing Classification Error Probability: an Information-Theoretic Approach | We formulate the problem of performing optimal data compression under the constraints that compressed data can be used for accurate classification in machine learning. We show that this translates to a problem of minimizing the mutual information between data and its compressed version under the constraint on error probability of classification is small when using the compressed data for machine learning. We then provide analytical and computational methods to characterize the optimal trade-off between data compression and classification error probability. First, we provide an analytical characterization for the optimal compression strategy for data with binary labels. Second, for data with multiple labels, we formulate a set of convex optimization problems to characterize the optimal tradeoff, from which the optimal trade-off between the classification error and compression efficiency can be obtained by numerically solving the formulated optimization problems. We further show the improvements of our formulations over the information-bottleneck methods in classification performance. | ['Weiyu Xu', 'Ao Tang', 'Jingchao Gao'] | 2022-11-03 | null | null | null | null | ['data-compression'] | ['time-series'] | [ 7.39739120e-01 2.75881529e-01 -5.49730837e-01 -5.65759063e-01
-1.11166203e+00 -2.76824057e-01 -8.02639499e-02 6.82719469e-01
-5.85226297e-01 4.66812223e-01 -8.32556859e-02 -2.70926118e-01
-7.02288508e-01 -6.56892121e-01 -5.81038594e-01 -7.01614261e-01
-1.22488052e-01 7.62937605e-01 -4.36182261e-01 4.70133007e-01
4.11803186e-01 4.29330915e-01 -1.75156260e+00 1.12316445e-01
8.78630459e-01 1.41296577e+00 2.06012860e-01 8.09338331e-01
9.79452282e-02 4.66440469e-01 -5.58580101e-01 -5.02824545e-01
4.85171676e-01 -4.71111596e-01 -1.20446050e+00 3.56246024e-01
2.29478538e-01 -2.89539039e-01 -2.35283941e-01 1.17977464e+00
2.39109352e-01 3.73617373e-02 8.85314047e-01 -1.40428030e+00
-3.87615353e-01 5.80655158e-01 -4.00840372e-01 2.56828249e-01
9.94917974e-02 -5.16247690e-01 1.23158896e+00 -3.36044490e-01
4.32000995e-01 1.00412524e+00 5.28241515e-01 3.33065331e-01
-1.35083175e+00 -2.61730671e-01 -4.19493109e-01 2.48498455e-01
-1.43576038e+00 -5.39656341e-01 3.69373828e-01 -3.67891461e-01
6.94871187e-01 5.64475894e-01 4.39639539e-01 1.63937166e-01
1.18784450e-01 6.01310551e-01 8.28721702e-01 -7.35994101e-01
3.18313509e-01 2.95921028e-01 7.04150259e-01 7.52480328e-01
6.92684591e-01 -1.07335769e-01 -5.26975155e-01 -2.26869017e-01
1.23364948e-01 -3.67605016e-02 -1.66539952e-01 -2.07788259e-01
-5.92158556e-01 9.56322312e-01 1.34180933e-01 -5.71205020e-02
-9.57793891e-02 1.10848648e-02 4.22415763e-01 6.34171724e-01
5.95337689e-01 4.48442161e-01 -2.99429297e-01 -1.18686333e-01
-1.00807655e+00 1.95654735e-01 1.07903647e+00 1.05922484e+00
5.75855970e-01 -4.86562937e-01 2.28352845e-02 1.01539660e+00
1.03342026e-01 4.54663545e-01 3.99667650e-01 -1.23461556e+00
6.29363954e-01 3.48733723e-01 2.75992379e-02 -8.70946884e-01
-2.21865833e-01 -6.13680243e-01 -7.58054078e-01 -4.14170921e-01
5.03527462e-01 1.09931074e-01 -3.27467799e-01 1.68379557e+00
1.09866224e-01 -3.79665762e-01 2.73355007e-01 5.70369363e-01
1.28599897e-01 6.59916580e-01 -1.58398107e-01 -9.19581890e-01
1.16900539e+00 -6.62130237e-01 -8.83499086e-01 1.08192191e-02
1.18451297e+00 -5.43042064e-01 8.69230747e-01 1.73832521e-01
-1.17687774e+00 -1.55048370e-01 -1.13437438e+00 -2.96959847e-01
-7.59599805e-02 -6.07063957e-02 4.72313017e-01 7.97646940e-01
-7.57223487e-01 8.76316190e-01 -8.66187990e-01 -2.69781291e-01
3.50571305e-01 5.39253116e-01 -1.44722268e-01 -2.51743048e-01
-5.52585900e-01 6.60874009e-01 4.87345338e-01 -3.38319004e-01
-4.38325882e-01 -7.22634375e-01 -6.07178271e-01 3.47622693e-01
1.55501962e-01 -3.54615778e-01 1.42543173e+00 -7.92307556e-01
-8.48174334e-01 1.16428840e+00 -3.61487687e-01 -6.55134618e-01
2.00984135e-01 1.07587175e-02 1.68996572e-01 3.30738813e-01
-2.41232440e-02 3.63985687e-01 3.48879397e-01 -8.89909327e-01
-9.11887348e-01 -5.95522821e-01 -3.26369479e-02 2.73097903e-01
-7.77945876e-01 -1.63113445e-01 -4.59697664e-01 -4.96235549e-01
3.18832099e-01 -9.64505911e-01 -3.54276188e-02 1.59639403e-01
-3.05803627e-01 -2.58362472e-01 6.27864361e-01 -5.42557418e-01
1.31707156e+00 -2.21401381e+00 1.73205391e-01 2.73763597e-01
2.74426758e-01 -4.36555110e-02 -5.61407879e-02 2.73995399e-01
2.16237292e-01 3.08332175e-01 -3.02487642e-01 -5.60719728e-01
-1.26644984e-01 3.47910702e-01 -1.32234246e-01 7.31026471e-01
-9.99542922e-02 5.49154520e-01 -6.26812398e-01 -5.81170261e-01
-2.64066100e-01 9.29067954e-02 -7.80698180e-01 4.08054918e-01
8.50040689e-02 1.23262033e-02 -4.60335612e-01 4.22492951e-01
7.59609461e-01 -5.15771806e-01 3.83678764e-01 -1.28661200e-01
2.59588718e-01 3.49347651e-01 -9.41251338e-01 1.26466417e+00
-4.73169327e-01 5.55267751e-01 4.26705509e-01 -1.39758992e+00
5.57235658e-01 1.18528781e-02 7.76678503e-01 -3.45694155e-01
2.53434122e-01 1.53461963e-01 -3.12618643e-01 -3.46179217e-01
3.77103657e-01 -3.54139417e-01 -1.10514522e-01 8.55097890e-01
-1.97332725e-01 3.27254795e-02 2.92860746e-01 1.55847713e-01
9.87006485e-01 -8.49318027e-01 3.19344848e-01 -4.60926116e-01
5.89819103e-02 4.62142639e-02 3.19071352e-01 8.89471233e-01
4.73922938e-02 2.42347956e-01 6.53485417e-01 -8.23627114e-02
-1.30323303e+00 -5.60180247e-01 -7.33563721e-01 9.65450525e-01
1.06636725e-01 -5.40532470e-01 -1.02338684e+00 -4.29974735e-01
1.48897424e-01 5.71856320e-01 -4.68504488e-01 -2.65561640e-01
-8.06645527e-02 -9.60574687e-01 3.13504219e-01 2.57877380e-01
4.01895314e-01 -2.88824499e-01 -8.62872899e-01 -2.45552778e-01
-3.31641078e-01 -1.16846633e+00 -3.68944764e-01 5.30531108e-01
-1.15356970e+00 -1.07287812e+00 -1.88903391e-01 -6.86561465e-01
8.34370673e-01 4.09335613e-01 8.36544454e-01 4.06242818e-01
-2.48112500e-01 4.47336167e-01 -4.01632220e-01 -4.03056234e-01
-6.43172324e-01 4.07221526e-01 -1.75773911e-02 -2.07638159e-01
2.98213720e-01 -3.79855871e-01 -3.79983902e-01 2.82163531e-01
-1.15214097e+00 1.68686911e-01 3.87563527e-01 8.25787127e-01
9.10985947e-01 2.79643357e-01 4.35945839e-01 -9.58199680e-01
6.84186220e-01 -5.50425053e-01 -6.25069678e-01 4.98590112e-01
-1.17025578e+00 6.53492272e-01 5.65585434e-01 -4.43552919e-02
-5.06876111e-01 1.94054455e-01 -1.77423149e-01 -2.64739811e-01
1.23347454e-01 5.00330687e-01 -1.35299876e-01 -9.63334069e-02
3.04415733e-01 2.03775957e-01 2.85325974e-01 -6.75927520e-01
2.18494192e-01 1.35772085e+00 4.50173855e-01 -6.45511806e-01
-1.00563280e-02 3.20571363e-01 3.85172427e-01 -7.23099113e-01
-1.43413758e+00 -5.71500182e-01 -7.41258383e-01 1.07559271e-01
4.70992237e-01 -4.77510870e-01 -7.77995706e-01 1.09994389e-01
-8.01515102e-01 -3.47581506e-02 -6.74254060e-01 3.40048552e-01
-1.00609398e+00 4.53340143e-01 -6.82351470e-01 -9.49331641e-01
-3.00157636e-01 -1.05951965e+00 9.51755166e-01 -1.98674649e-01
1.11033157e-01 -8.51277947e-01 -1.57918692e-01 4.74476069e-01
3.92688289e-02 -3.10904887e-02 1.25673723e+00 -8.52896988e-01
-3.24017942e-01 -5.54953218e-01 -5.09776235e-01 3.79473358e-01
-2.89307050e-02 -5.65910578e-01 -7.18928516e-01 -6.42817616e-01
2.33373716e-01 -5.30256927e-01 8.49004924e-01 3.30704421e-01
1.73666990e+00 -8.78576636e-01 -4.53261346e-01 1.00745142e+00
1.56109834e+00 -8.75131786e-02 1.67223662e-01 -1.71656117e-01
3.16575229e-01 7.25700259e-01 7.67444074e-01 8.08048189e-01
2.06939474e-01 7.03413785e-01 1.56295627e-01 4.26927626e-01
1.56985283e-01 -2.76442617e-01 -2.09967971e-01 1.09109628e+00
3.50811422e-01 -3.52534235e-01 -6.36554718e-01 3.19141209e-01
-1.69835782e+00 -6.88386798e-01 9.39029679e-02 2.77084780e+00
9.20100331e-01 -1.18194222e-01 1.13441825e-01 6.33530378e-01
6.34257734e-01 -3.06610823e-01 -7.39597082e-01 -4.67085898e-01
-6.46780059e-02 5.99516667e-02 9.72509086e-01 6.37888253e-01
-9.88746583e-01 3.70110810e-01 8.16896248e+00 9.15598571e-01
-6.79224372e-01 2.28853300e-01 9.77032304e-01 -3.05217057e-01
-2.34381869e-01 -1.03315718e-01 -9.57349658e-01 4.61439967e-01
1.58649504e+00 -5.94405532e-01 6.04039431e-01 7.89862931e-01
-5.88040762e-02 -9.89519879e-02 -1.54637349e+00 1.17433929e+00
-4.83115576e-02 -1.19506574e+00 -4.91906814e-02 4.56550896e-01
5.56985259e-01 -2.32588112e-01 1.23125045e-02 -1.38281196e-01
1.53811678e-01 -9.45279598e-01 5.27463138e-01 2.06154644e-01
1.12689936e+00 -7.48276174e-01 5.19184291e-01 6.74449086e-01
-8.17676842e-01 -5.65811872e-01 -6.47812009e-01 -1.33709893e-01
-8.01223367e-02 9.23075438e-01 -5.72782397e-01 3.24334592e-01
2.04405934e-01 6.19028032e-01 -3.84755373e-01 1.00358582e+00
5.11894107e-01 4.76090193e-01 -5.67143142e-01 7.79705122e-02
-2.56887704e-01 -2.01943755e-01 2.80868500e-01 1.34384656e+00
3.18948120e-01 2.46846423e-01 1.10028364e-01 4.57356334e-01
-3.75031590e-01 3.35706711e-01 -5.45481384e-01 -2.02741966e-01
7.77747631e-01 7.55272746e-01 -7.43336618e-01 -3.38060528e-01
-1.60030767e-01 9.39115822e-01 8.23096216e-01 -2.31653615e-03
-2.45144710e-01 -2.79132545e-01 4.92717981e-01 8.46536085e-02
4.85168695e-02 -1.73707888e-01 -5.90825140e-01 -9.69314814e-01
1.11201748e-01 -5.98181665e-01 8.51968348e-01 -2.04850957e-01
-1.15687358e+00 2.23620459e-01 3.37335199e-01 -8.54731798e-01
-8.29848498e-02 -5.56559324e-01 1.29081219e-01 5.99119484e-01
-1.26486135e+00 -3.59137565e-01 -9.72327590e-03 2.21760303e-01
3.97826791e-01 4.04494144e-02 8.26770186e-01 4.70782936e-01
-6.36627674e-01 1.10431349e+00 6.54035270e-01 -2.40857169e-01
1.21083915e-01 -9.89007950e-01 -3.66163552e-01 6.34700775e-01
-7.53312707e-02 1.53240189e-01 6.43937826e-01 -4.72688109e-01
-1.48053777e+00 -1.06557906e+00 1.06760085e+00 -5.13003506e-02
1.82449564e-01 -5.13792709e-02 -6.82091534e-01 5.69460511e-01
-4.19534862e-01 -6.50194660e-02 1.04250872e+00 5.24667092e-03
-4.27978158e-01 -2.52610356e-01 -1.49274230e+00 -7.30477124e-02
1.06871557e+00 -4.53198493e-01 -1.34550959e-01 8.85512352e-01
7.75783420e-01 -1.92777261e-01 -1.16373658e+00 2.30008975e-01
5.66168606e-01 -6.26243174e-01 7.42910564e-01 -8.39164317e-01
5.63497007e-01 3.78971040e-01 -8.09888363e-01 -1.03385115e+00
6.60452934e-04 -4.52469975e-01 -1.50657475e-01 7.76481211e-01
3.09703439e-01 -5.66205025e-01 8.78290355e-01 9.68730271e-01
3.03221017e-01 -1.05883861e+00 -1.04766393e+00 -8.82869244e-01
1.35283068e-01 -4.32339400e-01 3.00034434e-01 7.40348220e-01
2.84265667e-01 6.95652291e-02 -3.34491730e-01 -9.18095633e-02
8.79478514e-01 2.58001715e-01 2.07422063e-01 -1.15705180e+00
-4.72746730e-01 -3.10059607e-01 -5.07649779e-01 -1.46196008e+00
1.83523342e-01 -1.34103823e+00 -1.21483631e-01 -1.12947524e+00
8.52714300e-01 -8.74368787e-01 -2.32891187e-01 2.68952549e-01
1.09112889e-01 1.21464496e-02 2.85070151e-01 7.41829336e-01
-6.05452657e-01 3.10930520e-01 9.15473998e-01 2.92504206e-02
4.63045128e-02 1.16027363e-01 -7.81659782e-01 2.41145134e-01
5.95759749e-01 -8.63706827e-01 -4.02180672e-01 -4.65390444e-01
3.00210148e-01 3.62731189e-01 -1.53484941e-01 -8.41410160e-01
2.81886131e-01 -4.84950803e-02 -9.22372863e-02 -4.61418182e-01
2.73627758e-01 -9.15802896e-01 -4.35004123e-02 6.97689235e-01
-1.23461580e+00 1.25100493e-01 -2.28113100e-01 7.10850954e-01
-5.01704514e-02 -7.88189888e-01 1.15696430e+00 1.19467057e-01
1.11929476e-01 3.12438071e-01 -2.99376756e-01 2.53011346e-01
1.06289268e+00 1.05895242e-02 -6.11737184e-02 -4.27176863e-01
-7.26946712e-01 2.62450069e-01 3.71391773e-01 -5.24772517e-02
5.22888124e-01 -1.21144259e+00 -6.70515120e-01 2.37444177e-01
-1.60484035e-02 -1.67599618e-01 -8.51767138e-02 6.70717359e-01
-5.35219014e-01 7.79351830e-01 -3.95858549e-02 -5.22515297e-01
-1.54337084e+00 5.68348646e-01 3.26532245e-01 -4.18610066e-01
-3.39113772e-01 5.78367710e-01 -2.51526445e-01 -8.29412118e-02
6.33520603e-01 -1.12684615e-01 2.75996387e-01 -9.00939107e-02
4.67607915e-01 5.74039936e-01 3.06560487e-01 -3.82687926e-01
-2.48241634e-03 3.74897331e-01 3.63896675e-02 -1.76545724e-01
1.20240593e+00 -4.61034149e-01 -2.59172171e-01 3.86071980e-01
1.91102076e+00 -3.98760617e-01 -8.74265969e-01 -5.49938202e-01
-8.56776088e-02 -7.26476312e-01 2.64056951e-01 -4.22537297e-01
-1.05006826e+00 7.58402228e-01 6.28294468e-01 4.90141630e-01
1.30965781e+00 1.29002094e-01 6.40057683e-01 6.81126475e-01
2.51914799e-01 -9.76242483e-01 -2.08821476e-01 2.08229795e-01
5.71667969e-01 -1.05282784e+00 2.20013842e-01 -7.52131402e-01
-1.89051911e-01 1.07220149e+00 4.10172492e-02 1.39655977e-01
1.13825071e+00 4.68418360e-01 -5.07709801e-01 -1.51904956e-01
-9.86225843e-01 1.36656627e-01 2.07915634e-01 3.57812881e-01
3.71663451e-01 3.78453225e-01 -7.20898569e-01 5.91186523e-01
-5.89446068e-01 -1.31880179e-01 2.15513885e-01 8.91166449e-01
-6.23483241e-01 -1.00831664e+00 -1.92144006e-01 1.07989240e+00
-7.29785681e-01 5.50852194e-02 -3.02088588e-01 1.82282090e-01
-3.65043804e-02 1.03270674e+00 4.76950496e-01 -3.17493170e-01
5.54870954e-03 1.54393658e-01 6.33312881e-01 -5.39715946e-01
4.72550318e-02 -2.69556493e-01 -1.34812044e-02 -3.63714248e-01
-2.84882188e-01 -6.39169037e-01 -1.02108586e+00 -5.84886134e-01
-5.90963960e-01 4.22810525e-01 8.97746325e-01 1.05010271e+00
4.25408691e-01 1.01958983e-01 7.92933941e-01 -2.64064610e-01
-1.17810106e+00 -6.33047104e-01 -7.65595973e-01 5.70477426e-01
4.01209205e-01 -2.68344998e-01 -5.02904892e-01 1.46706020e-02] | [7.71693229675293, 4.194887638092041] |
972f15f1-1b65-42e4-8f46-c67ff7f9e94f | deep-learning-based-identification-of-sub | 2207.09598 | null | https://arxiv.org/abs/2207.09598v1 | https://arxiv.org/pdf/2207.09598v1.pdf | Deep learning-based identification of sub-nuclear structures in FIB-SEM images | Three-dimensional volumetric imaging of cells allows for in situ visualization, thus preserving contextual insights into cellular processes. Despite recent advances in machine learning methods, morphological analysis of sub-nuclear structures have proven challenging due to both the shallow contrast profile and the technical limitation in feature detection. Here, we present a convolutional neural network, supervised deep learning-based approach which can identify sub-nuclear structures with 90% accuracy. We develop and apply this model to C. elegans gonads imaged using focused ion beam milling combined with scanning electron microscopy resulting in the accurate identification and segmentation of all sub-nuclear structures including entire chromosomes. We discuss in depth the architecture, parameterization, and optimization of the deep learning model, as well as provide evaluation metrics to assess the quality of the network prediction. Lastly, we highlight specific aspects of the model that can be optimized for its broad application to other volumetric imaging data as well as in situ cryo-electron tomography. | ['Vignesh Kasinath', 'Petrus H. Zwart', 'Danielle Jorgens', 'Abby Dernburg', 'Fan Wu', 'Harald F. Hess', 'C. Shan Xu', 'Song Pang', 'Eric J. Roberts', 'Niraj Gupta'] | 2022-07-19 | null | null | null | null | ['electron-tomography', 'morphological-analysis'] | ['medical', 'natural-language-processing'] | [ 3.07086289e-01 1.03523560e-01 4.15540665e-01 -2.69988596e-01
-4.17724639e-01 -7.77152240e-01 2.07266167e-01 5.87614298e-01
-8.28908145e-01 6.23626769e-01 -4.18635428e-01 -2.97423601e-01
-1.67306006e-01 -5.35178721e-01 -5.15827298e-01 -9.38046575e-01
-3.87908459e-01 9.46716785e-01 5.66708706e-02 1.91484496e-01
3.55751514e-01 1.00403893e+00 -1.14811468e+00 -3.67362541e-03
3.49743754e-01 8.52649689e-01 5.01139820e-01 9.43913400e-01
1.31573066e-01 8.02181587e-02 -5.03144383e-01 -7.31756240e-02
-1.05993010e-01 -2.07751319e-01 -8.98572624e-01 7.61306882e-02
5.09731114e-01 -2.61989534e-01 1.14527069e-01 7.34093428e-01
7.00088084e-01 -2.44365752e-01 7.74234414e-01 -6.05200231e-01
-3.52505088e-01 1.73167035e-01 -4.38153177e-01 6.35200024e-01
-1.07576743e-01 1.11262448e-01 7.88590431e-01 -7.54029751e-01
1.03716278e+00 9.50801492e-01 7.60474443e-01 4.44431782e-01
-1.63095534e+00 -3.59826624e-01 -3.20464373e-01 -1.23565555e-01
-8.67932618e-01 -3.72191250e-01 4.58687842e-01 -6.72868133e-01
9.37340319e-01 -9.46636125e-02 1.10593700e+00 3.16620469e-01
8.16528499e-01 5.25542498e-01 1.11476529e+00 -3.15054685e-01
4.02040422e-01 -5.40364206e-01 1.63644757e-02 1.02155197e+00
1.95379123e-01 -1.57635704e-01 -3.46709281e-01 -4.47012447e-02
1.06281745e+00 -1.12484664e-01 -4.12824042e-02 -5.21724463e-01
-1.03835917e+00 5.90686560e-01 1.63575262e-01 4.89209414e-01
1.28719592e-02 1.42809540e-01 6.16343439e-01 -1.05792373e-01
3.64141554e-01 6.78758085e-01 -7.59689569e-01 1.92889310e-02
-1.01673496e+00 3.27204823e-01 3.49340439e-01 1.99343681e-01
6.16619885e-01 -1.96051113e-02 3.68769169e-01 6.40005171e-01
2.98594475e-01 -2.98423637e-02 3.62644672e-01 -1.32883227e+00
-1.52262852e-01 8.34295094e-01 -4.34773356e-01 -6.57896876e-01
-1.11151147e+00 -4.22791660e-01 -8.73935342e-01 5.98853767e-01
7.16830373e-01 -1.37178481e-01 -9.25817728e-01 1.39707601e+00
3.88406992e-01 -2.99332708e-01 -4.16959584e-01 7.25060999e-01
7.90941656e-01 2.46334761e-01 -5.63087165e-02 -6.91680238e-02
1.28266799e+00 -2.17108920e-01 -5.56833267e-01 -7.11975619e-02
7.06427276e-01 -3.46130043e-01 6.35786474e-01 2.90103525e-01
-1.10918283e+00 -4.84574139e-02 -1.13473928e+00 -4.11079675e-01
-6.34154081e-01 3.63606542e-01 7.50210404e-01 5.26072443e-01
-9.85452175e-01 1.01455200e+00 -1.46217108e+00 -6.51393890e-01
8.67413044e-01 8.47554266e-01 -6.38930559e-01 3.21153820e-01
-2.65972883e-01 6.27976656e-01 3.74013603e-01 1.14774075e-03
-9.61947143e-01 -9.30753589e-01 -8.74544799e-01 2.94100851e-01
-3.82411569e-01 -5.99291265e-01 9.49019432e-01 -5.17980494e-02
-1.37727475e+00 1.48364341e+00 -1.04470298e-01 -3.65907490e-01
2.70513058e-01 3.27396035e-01 4.13408458e-01 3.68066847e-01
1.17325282e-03 9.58801568e-01 -1.24281906e-01 -1.18724155e+00
-4.88223433e-01 -9.75417554e-01 -2.01697782e-01 1.48172975e-02
-1.52477786e-01 5.48121296e-02 -3.89259219e-01 -1.35281175e-01
5.17183423e-01 -6.44004107e-01 -3.32718492e-01 4.17415977e-01
-4.35305655e-01 3.17904174e-01 1.14757931e+00 -5.72347999e-01
4.09505725e-01 -1.93393385e+00 4.23655599e-01 -5.99510856e-02
5.68940520e-01 1.04819432e-01 1.41363218e-01 2.97863275e-01
-4.45895121e-02 1.44214183e-01 -2.95063317e-01 -6.53203309e-01
-3.96715790e-01 2.66134609e-02 2.30453074e-01 9.96948957e-01
1.54376715e-01 9.57994878e-01 -5.89114785e-01 -5.52219212e-01
4.13602471e-01 6.02055967e-01 -2.56806880e-01 -6.13612719e-02
1.65781286e-02 6.92145765e-01 -8.96728933e-02 9.24127817e-01
5.43377697e-01 -4.26620603e-01 5.96562147e-01 -1.51906043e-01
-2.28268772e-01 -3.80810015e-02 -5.09455144e-01 1.57729864e+00
-1.67282030e-01 1.03937352e+00 6.97459280e-01 -1.11007237e+00
6.87423706e-01 -5.76373897e-02 5.58672845e-01 -5.58634460e-01
2.68380702e-01 2.04428256e-01 2.44444441e-02 -1.25499234e-01
2.13676438e-01 -3.20364535e-01 2.34846041e-01 4.00674015e-01
4.58697647e-01 -4.04582322e-01 3.61894071e-01 7.00776652e-02
9.01438534e-01 2.62316316e-01 2.40156382e-01 -7.01993883e-01
2.03132063e-01 -1.72615387e-02 6.67041898e-01 2.31215745e-01
-1.75798938e-01 8.42280388e-01 6.87657118e-01 -7.39460289e-01
-1.47182453e+00 -6.95908546e-01 -5.88282943e-01 8.38239551e-01
-1.73990820e-02 -1.71710581e-01 -7.76197314e-01 -2.21625596e-01
7.99625143e-02 -1.62495568e-01 -1.00036347e+00 1.63474530e-01
-6.22011602e-01 -1.25909579e+00 8.19870055e-01 5.18900514e-01
1.71718299e-01 -8.69355202e-01 -1.08802557e+00 1.71900511e-01
4.07461494e-01 -8.98072600e-01 4.28402781e-01 8.42937648e-01
-1.13718367e+00 -1.30483067e+00 -6.45835042e-01 -8.00211251e-01
8.46757054e-01 -2.58638114e-01 9.42709923e-01 3.94051999e-01
-9.70964789e-01 -2.36261450e-02 2.17475533e-01 -2.99226910e-01
-2.00497985e-01 1.60915926e-01 -1.98991392e-02 -8.08837652e-01
1.16806693e-01 -6.46894932e-01 -5.71445346e-01 5.10391742e-02
-9.06172991e-01 -8.71571228e-02 2.97758132e-01 9.78921413e-01
9.72896099e-01 3.42112482e-02 1.43375084e-01 -1.06727433e+00
3.53654891e-01 -1.45833641e-01 -7.63188303e-01 -7.69829303e-02
-4.50764000e-01 -3.84706885e-01 6.14718497e-01 1.06800906e-01
-7.34924853e-01 1.60885558e-01 -2.54126370e-01 9.97434929e-02
-3.86928916e-01 6.46839738e-01 -5.77252768e-02 -5.70747435e-01
4.24718231e-01 2.18430132e-01 2.12780416e-01 -3.79825890e-01
-2.63825715e-01 -2.23684125e-03 3.90784174e-01 -5.50481558e-01
4.58853096e-01 9.64846194e-01 6.98258996e-01 -9.99883354e-01
-4.03732181e-01 -2.07866803e-01 -1.10433722e+00 -2.26239726e-01
8.36351156e-01 -4.77180153e-01 -1.04058576e+00 5.55013120e-01
-8.75252068e-01 -5.89237690e-01 -3.71164307e-02 1.91355020e-01
-7.50646055e-01 5.30762613e-01 -1.34439301e+00 -5.34779966e-01
-3.39500189e-01 -1.31964588e+00 1.36924982e+00 3.98502409e-01
-1.19599529e-01 -1.35883534e+00 2.88770914e-01 2.47955829e-01
1.89718783e-01 3.34540308e-01 1.40669918e+00 -5.10662913e-01
-2.99260676e-01 -1.12777106e-01 -2.65529335e-01 -2.61325866e-01
-2.92487303e-03 3.72260273e-01 -1.01622975e+00 -5.34162760e-01
-3.91317308e-01 -7.13797510e-01 1.01810777e+00 8.72860968e-01
1.17903388e+00 3.72697502e-01 -6.84941053e-01 1.10279047e+00
1.52252734e+00 3.00246030e-01 4.69648093e-01 6.63128018e-01
3.98264557e-01 6.84716105e-01 2.73729622e-01 2.82665521e-01
-9.40075610e-03 4.32039142e-01 5.89366436e-01 -4.04739290e-01
9.96027514e-02 2.71917462e-01 -4.25215334e-01 2.94194311e-01
-1.79204553e-01 -1.14998715e-02 -1.11274731e+00 5.78630984e-01
-1.26726019e+00 -6.23494506e-01 1.19449422e-01 1.76622927e+00
6.33572519e-01 -4.82874885e-02 8.90347138e-02 3.58986557e-01
4.89484519e-01 -1.46997795e-01 -6.32900894e-01 -2.29730681e-01
-2.90742099e-01 2.66791463e-01 4.08491999e-01 4.04731631e-01
-1.16643846e+00 7.22910643e-01 7.80529118e+00 5.33199191e-01
-1.18683457e+00 -3.43237162e-01 1.17348111e+00 -1.72348600e-02
4.54528481e-02 -1.52376592e-01 -9.32489932e-01 2.50010788e-01
4.49872255e-01 2.31156528e-01 2.40095556e-01 3.46875966e-01
2.11396843e-01 -3.00590992e-01 -1.10896921e+00 5.73615730e-01
-3.14305276e-01 -1.86789930e+00 -2.60147899e-01 5.81009626e-01
4.06209648e-01 1.63441837e-01 1.29735172e-01 -1.70713335e-01
8.40493962e-02 -1.17190218e+00 3.35518420e-01 1.49689496e-01
9.01959479e-01 -9.26568270e-01 7.88081586e-01 3.27819169e-01
-9.32617843e-01 -9.65007320e-02 -6.94103420e-01 -1.51846066e-01
1.60933584e-02 6.95727289e-01 -8.11540067e-01 1.73076645e-01
8.57375622e-01 4.60648388e-01 -5.71544886e-01 1.02470064e+00
4.16700065e-01 3.84278715e-01 -5.28541148e-01 1.31578043e-01
1.94873791e-02 -4.56138730e-01 1.76200300e-01 1.33210063e+00
1.79169953e-01 -7.98490793e-02 -9.32955593e-02 1.14894319e+00
-1.44034671e-02 -7.94286653e-02 -4.56400096e-01 -3.00943673e-01
3.60077322e-01 1.70373440e+00 -1.55949438e+00 2.14982897e-01
1.86860096e-02 5.24893165e-01 7.54747331e-01 2.09682792e-01
-2.28563622e-01 -4.08300966e-01 5.00580192e-01 1.99859053e-01
1.70283437e-01 -4.08454537e-01 -5.21941662e-01 -7.19020247e-01
-4.86724049e-01 -3.52613211e-01 1.27171069e-01 -5.03732204e-01
-7.19014943e-01 1.46113276e-01 -3.07189792e-01 -4.20087546e-01
8.08895081e-02 -1.06250763e+00 -6.83847427e-01 6.13859296e-01
-1.16572690e+00 -1.18860126e+00 4.80797179e-02 -3.03329706e-01
3.29145610e-01 -1.24949262e-01 1.04727232e+00 1.51974246e-01
-8.36796999e-01 1.62109241e-01 4.13557053e-01 1.64510965e-01
1.95232064e-01 -1.60202169e+00 3.36210489e-01 4.41377252e-01
-3.38823467e-01 7.84081221e-01 7.27084696e-01 -5.93641698e-01
-1.27046645e+00 -1.01089966e+00 5.55807948e-01 -1.41391307e-01
4.01439637e-01 -4.18319613e-01 -6.99642777e-01 9.34135318e-01
1.33576989e-01 1.22754805e-01 1.09250462e+00 9.02590975e-02
2.05692470e-01 3.00620586e-01 -1.55776763e+00 4.31989610e-01
5.12715280e-01 -1.41601145e-01 3.95373181e-02 2.43257821e-01
1.04993366e-01 -5.88639975e-01 -1.15840590e+00 3.45014095e-01
7.84849763e-01 -1.12627065e+00 9.00128007e-01 -5.13408542e-01
7.43543208e-01 -2.26293251e-01 5.24738058e-02 -1.15971005e+00
-5.29219449e-01 -9.92643386e-02 2.13381872e-02 1.07621181e+00
2.86250532e-01 -2.99643695e-01 1.38948727e+00 1.79140300e-01
-2.44499967e-01 -1.19074237e+00 -1.12208760e+00 -3.36488962e-01
5.44592857e-01 2.49273673e-01 1.90211505e-01 7.90856361e-01
2.23280966e-01 1.82401896e-01 2.74841905e-01 -7.65373781e-02
7.63485372e-01 1.89419210e-01 3.34520608e-01 -1.46920037e+00
-1.56493917e-01 -3.88252288e-01 -7.64301777e-01 -7.76573956e-01
2.37540841e-01 -5.86956620e-01 -6.11139238e-02 -1.41379881e+00
6.97462559e-01 -1.64628953e-01 1.99563086e-01 3.99771661e-01
9.16920602e-02 7.21934319e-01 -1.38133243e-01 1.11836813e-01
-4.81768399e-01 2.40129337e-01 1.24187410e+00 -1.00780994e-01
2.15978026e-01 -3.93286258e-01 -3.00425768e-01 8.20783198e-01
8.19028795e-01 -2.50321180e-01 3.76527458e-02 -4.18948054e-01
2.78975993e-01 -5.32140322e-02 2.42726922e-01 -1.06023860e+00
2.25690484e-01 1.54238403e-01 1.26391459e+00 -6.83343768e-01
6.30995393e-01 -6.23215199e-01 -2.53980923e-02 6.17311418e-01
-4.16973084e-01 -1.30723998e-01 3.61623377e-01 3.57309908e-01
-2.47552665e-03 -1.72451913e-01 1.26893723e+00 -5.97496867e-01
-1.66813195e-01 1.80594712e-01 -9.04512763e-01 -4.44903046e-01
9.10852015e-01 -5.02250254e-01 -5.13097346e-01 4.73413616e-02
-9.97492075e-01 2.10506946e-01 1.03829420e+00 -7.20381796e-01
6.68515742e-01 -8.61205578e-01 -2.39258945e-01 1.60300195e-01
-1.07246421e-01 3.78260285e-01 3.50434065e-01 7.82309711e-01
-1.28513551e+00 5.19323051e-01 -6.40109837e-01 -8.13125789e-01
-1.48029947e+00 8.57436135e-02 6.55660868e-01 -3.84333521e-01
-5.12999177e-01 9.38417733e-01 2.48933896e-01 -8.93127263e-01
-3.79923955e-02 -2.18871862e-01 -6.41304553e-01 -2.91544497e-01
3.06025177e-01 3.75091076e-01 1.03483550e-01 -5.81479847e-01
-3.05757016e-01 7.45134771e-01 -1.13208890e-01 2.52396166e-01
1.66911578e+00 -3.71868074e-01 -4.41221684e-01 4.41666543e-01
1.10179448e+00 -3.92954350e-01 -1.34654593e+00 3.26404840e-01
-1.13380469e-01 -1.25799432e-01 3.10374737e-01 -7.20791996e-01
-1.27589428e+00 1.15661514e+00 6.73429966e-01 5.54302670e-02
8.98172438e-01 -7.05534220e-02 3.89289856e-01 4.48687106e-01
2.87756264e-01 -9.13268387e-01 -1.50244787e-01 5.74102283e-01
2.34938487e-01 -1.05877495e+00 4.09365863e-01 -4.01890755e-01
1.95543706e-01 1.45322967e+00 5.96846819e-01 -2.04602182e-01
4.71432477e-01 9.12990451e-01 6.14337921e-02 -7.45343983e-01
-8.09570968e-01 2.45636508e-01 -3.90804917e-01 7.71592259e-01
9.07048702e-01 -3.56566101e-01 -7.38352537e-02 2.63102591e-01
3.16137224e-02 3.53142917e-02 6.33002639e-01 1.19001245e+00
-6.42190278e-01 -8.30930352e-01 -1.31605819e-01 5.98827600e-01
-9.81315732e-01 3.29317719e-01 -5.49674153e-01 6.78768814e-01
3.26884724e-02 2.62955070e-01 4.37030971e-01 2.79342942e-02
-1.00345410e-01 -9.50914621e-02 8.67860854e-01 -7.38065660e-01
-4.83003318e-01 3.59861910e-01 -1.40843004e-01 -3.27259839e-01
-2.59617120e-01 -5.91933787e-01 -1.56730855e+00 -2.60755748e-01
-3.81913275e-01 -6.08571758e-03 9.21636820e-01 9.93071437e-01
3.19332540e-01 6.23885870e-01 9.65362042e-02 -1.48091912e+00
1.26746997e-01 -9.23021972e-01 -1.14512813e+00 -8.52105990e-02
2.16729477e-01 -5.97532392e-01 -3.21630359e-01 1.81523964e-01] | [14.32059383392334, -3.13067364692688] |
dfd4e7da-d833-4067-a119-2e77bf036320 | recent-advances-of-local-mechanisms-in | 2306.01929 | null | https://arxiv.org/abs/2306.01929v1 | https://arxiv.org/pdf/2306.01929v1.pdf | Recent Advances of Local Mechanisms in Computer Vision: A Survey and Outlook of Recent Work | Inspired by the fact that human brains can emphasize discriminative parts of the input and suppress irrelevant ones, substantial local mechanisms have been designed to boost the development of computer vision. They can not only focus on target parts to learn discriminative local representations, but also process information selectively to improve the efficiency. In terms of application scenarios and paradigms, local mechanisms have different characteristics. In this survey, we provide a systematic review of local mechanisms for various computer vision tasks and approaches, including fine-grained visual recognition, person re-identification, few-/zero-shot learning, multi-modal learning, self-supervised learning, Vision Transformers, and so on. Categorization of local mechanisms in each field is summarized. Then, advantages and disadvantages for every category are analyzed deeply, leaving room for exploration. Finally, future research directions about local mechanisms have also been discussed that may benefit future works. To the best our knowledge, this is the first survey about local mechanisms on computer vision. We hope that this survey can shed light on future research in the computer vision field. | ['Yilong Yin', 'Qiangchang Wang'] | 2023-06-02 | null | null | null | null | ['person-re-identification', 'fine-grained-visual-recognition'] | ['computer-vision', 'computer-vision'] | [ 1.83836415e-01 -2.76518524e-01 -5.43176353e-01 -4.69777256e-01
-9.18369219e-02 -2.23672643e-01 6.85516357e-01 3.12462598e-02
-5.33678114e-01 4.58153754e-01 3.16589862e-01 4.90583271e-01
-9.58026052e-02 -6.11745417e-01 -9.92932245e-02 -9.58978295e-01
1.62537873e-01 -1.50093604e-02 4.77086365e-01 -7.24716112e-02
6.10725105e-01 5.75235724e-01 -2.03744221e+00 5.67939699e-01
6.04784250e-01 5.65829039e-01 5.24072230e-01 3.93838167e-01
-1.04590751e-01 6.32887483e-01 -5.60012758e-01 -1.23181857e-01
-1.42319342e-02 -5.39828598e-01 -5.60376406e-01 1.92516565e-01
7.47424185e-01 -1.28337696e-01 -5.52586198e-01 1.27185750e+00
8.06420982e-01 2.98868746e-01 8.87041152e-01 -9.30107296e-01
-9.74127650e-01 2.11517006e-01 -5.41587949e-01 9.57637131e-01
1.78274736e-01 2.07904905e-01 7.42918789e-01 -1.17976558e+00
4.44035232e-01 1.25665784e+00 3.44122976e-01 9.84327197e-01
-1.00324810e+00 -3.99194986e-01 6.74278796e-01 7.52682388e-01
-1.32313108e+00 -5.44713259e-01 7.34342456e-01 -4.40933555e-01
9.10075426e-01 2.78815359e-01 7.03414440e-01 1.19058001e+00
3.06462735e-01 1.04325724e+00 1.32736683e+00 -5.84320486e-01
2.10469559e-01 1.67112812e-01 6.95497394e-01 7.86894619e-01
3.30955625e-01 3.17474723e-01 -9.17086005e-01 2.06928570e-02
7.38851547e-01 3.32314938e-01 -1.92679957e-01 -4.60299999e-01
-1.02079940e+00 8.48451614e-01 5.82667708e-01 7.27776766e-01
-1.71552330e-01 -1.46633342e-01 3.07689130e-01 2.05518365e-01
-5.36281569e-03 2.75971800e-01 -2.65195698e-01 3.25589448e-01
-7.57586718e-01 -5.54660847e-03 3.37268174e-01 6.15551114e-01
8.96366477e-01 2.09923580e-01 -5.72893143e-01 1.36419904e+00
3.21459740e-01 3.58953655e-01 7.64020026e-01 -5.68759620e-01
7.13343248e-02 4.18153733e-01 -3.69202107e-01 -7.91369259e-01
-4.20337260e-01 -3.49863231e-01 -9.50972855e-01 3.16641599e-01
1.67402118e-01 1.46221101e-01 -8.49566281e-01 1.49602664e+00
-8.82959962e-02 8.59162807e-02 -3.54878366e-01 1.02713406e+00
1.19756794e+00 4.71114755e-01 3.92628521e-01 -3.72164041e-01
1.44611382e+00 -1.38207626e+00 -6.39811277e-01 -6.75120175e-01
2.83859111e-02 -5.88282585e-01 7.40003228e-01 1.16224721e-01
-8.09288859e-01 -8.76570880e-01 -8.91266763e-01 -8.45443606e-02
-6.21323407e-01 2.30497345e-01 7.53895819e-01 5.71939766e-01
-9.09263015e-01 4.01298910e-01 -6.55039132e-01 -1.09995461e+00
5.14163792e-01 1.02100790e-01 -1.13823555e-01 -1.71943739e-01
-1.06939375e+00 1.12434161e+00 1.63794026e-01 -4.73183542e-02
-1.14255393e+00 -3.66312474e-01 -6.62810147e-01 -8.69584531e-02
2.20781550e-01 -8.66001487e-01 1.14391065e+00 -8.85041416e-01
-1.24679840e+00 1.24347794e+00 -5.80258548e-01 -2.70238608e-01
-1.49227723e-01 9.68677748e-04 -4.53458488e-01 2.49001890e-01
1.47895887e-01 7.33347297e-01 9.77902234e-01 -1.13424897e+00
-7.44922340e-01 -7.00952768e-01 -1.05848975e-01 2.56785393e-01
-5.12826979e-01 4.44659650e-01 -6.63564384e-01 -8.19217622e-01
-7.89057463e-02 -5.67478895e-01 -3.72205764e-01 7.57214576e-02
-8.36496614e-03 -6.56037092e-01 6.43886685e-01 -1.92646563e-01
1.14032066e+00 -2.00732112e+00 7.98696205e-02 -1.38121545e-01
2.48751074e-01 4.51408952e-01 -2.03965053e-01 3.86920512e-01
-1.02713527e-02 -2.01543957e-01 -1.22757656e-02 -1.26965746e-01
-3.96784186e-01 8.66461545e-02 -3.54698807e-01 6.62812531e-01
-1.55773684e-01 1.25761902e+00 -7.43788540e-01 -5.30949295e-01
7.45281041e-01 3.37539315e-01 2.48450208e-02 -6.77738413e-02
3.75328451e-01 1.19347736e-01 -5.95148921e-01 9.68466938e-01
5.05747378e-01 -2.01317251e-01 -1.54437684e-02 -2.14567825e-01
-4.46367592e-01 -1.29349232e-01 -9.33462977e-01 1.48365045e+00
-1.04153097e-01 6.98391438e-01 6.00320026e-02 -1.30833018e+00
9.67465937e-01 2.87964437e-02 1.50104180e-01 -8.92586648e-01
1.36727110e-01 -2.44230047e-01 -2.16945983e-03 -4.40916508e-01
1.44055843e-01 -9.82212052e-02 7.99812451e-02 1.65477499e-01
3.91698837e-01 4.74564105e-01 5.58108203e-02 6.80600107e-02
5.50624013e-01 -1.97838813e-01 8.80183816e-01 -3.17780018e-01
6.97984159e-01 -9.72492546e-02 4.77242589e-01 1.20503998e+00
-9.46155012e-01 5.44200778e-01 -2.70486236e-01 -8.00692439e-01
-5.78638911e-01 -1.21378124e+00 -1.89399555e-01 1.63762021e+00
6.71565115e-01 -4.30577219e-01 -5.70180714e-01 -4.88530129e-01
-2.26872787e-01 4.36154544e-01 -7.98913002e-01 -5.37893593e-01
-3.49369437e-01 -9.31748688e-01 4.68072683e-01 8.04731131e-01
7.38552809e-01 -1.41337621e+00 -6.55386984e-01 -1.71931788e-01
-9.11938250e-02 -7.37540722e-01 -3.80996644e-01 6.42977282e-02
-1.18298101e+00 -1.00865114e+00 -9.38991606e-01 -1.27240825e+00
6.52793705e-01 1.14189291e+00 9.19211447e-01 1.08528316e-01
-7.04593897e-01 8.45825493e-01 -8.56945515e-02 -3.32881689e-01
5.92403173e-01 -2.35239983e-01 2.96943843e-01 -7.20248511e-03
1.02699268e+00 -3.02838475e-01 -6.73996031e-01 4.45569038e-01
-4.23622280e-01 -4.27796900e-01 8.31254065e-01 8.82824481e-01
5.19974828e-01 -8.03268328e-02 4.55082744e-01 -5.44515371e-01
6.70549154e-01 -1.95181847e-01 -2.07073048e-01 4.66702282e-01
-4.53134388e-01 -2.16577977e-01 4.32736158e-01 -4.56707805e-01
-1.53090501e+00 -4.05046903e-02 2.60311186e-01 -1.77265480e-01
-7.41933644e-01 1.31108373e-01 -1.35778695e-01 -4.04355377e-01
9.99803424e-01 7.62971580e-01 -2.71918029e-01 -6.55628026e-01
4.21247065e-01 4.86897320e-01 3.90769601e-01 -2.42222115e-01
4.33151394e-01 7.74868548e-01 -2.24617913e-01 -1.36042130e+00
-9.80691552e-01 -9.20244277e-01 -8.82725298e-01 -3.17359716e-01
8.84717584e-01 -8.57649028e-01 -4.70686734e-01 6.95509791e-01
-1.00780284e+00 -4.41652117e-03 -1.27094597e-01 3.72837782e-01
-4.37581748e-01 5.21870375e-01 -6.52316272e-01 -7.98463225e-01
-1.31433681e-01 -8.97720695e-01 6.86883271e-01 9.07324195e-01
-1.99375317e-01 -1.02585816e+00 2.14989573e-01 4.39256728e-01
3.93157512e-01 -6.11435711e-01 2.96146005e-01 -4.51782167e-01
-5.81303120e-01 -3.02912146e-02 -2.36416116e-01 1.28502250e-01
1.81765765e-01 -2.48306200e-01 -1.37668347e+00 -3.08749408e-01
7.69566000e-02 -3.81539792e-01 1.68358266e+00 7.74640441e-01
1.17884064e+00 8.72607306e-02 -8.60592365e-01 6.05241597e-01
1.19256282e+00 2.79846311e-01 6.38291359e-01 3.64579231e-01
4.37223524e-01 9.18222308e-01 5.08139789e-01 2.48541847e-01
2.59368509e-01 5.95069766e-01 1.11991383e-01 -4.47974801e-02
-5.86496830e-01 -6.29351363e-02 3.60934943e-01 5.65382123e-01
-4.95904028e-01 1.46351621e-01 -4.75892842e-01 4.72797632e-01
-1.95126736e+00 -1.48105133e+00 4.58254628e-02 2.18112350e+00
4.78107393e-01 -1.86849639e-01 1.02314085e-01 -3.95700634e-01
1.10929787e+00 3.85781169e-01 -6.12682164e-01 -1.19992100e-01
-4.57409024e-01 -5.80300242e-02 1.58818170e-01 4.70738351e-01
-1.43786848e+00 1.25858080e+00 7.53215599e+00 9.42912579e-01
-1.05408168e+00 1.71966985e-01 4.48934555e-01 -1.02058528e-02
2.70781845e-01 -1.16910964e-01 -1.10209858e+00 2.62029171e-01
4.54858869e-01 -1.31985871e-02 4.11053121e-01 1.09294963e+00
1.44393137e-02 -2.32133076e-01 -8.48837256e-01 1.52616107e+00
6.72884107e-01 -1.25805318e+00 1.71827003e-01 -2.78404076e-03
7.30858624e-01 1.59423932e-01 1.38910353e-01 3.94353122e-01
-1.80963889e-01 -8.89006972e-01 2.08404288e-01 8.22022557e-01
2.93814421e-01 -6.09476507e-01 4.69483286e-01 4.92860615e-01
-1.32589972e+00 -2.72496134e-01 -1.06107736e+00 -2.60503531e-01
-1.45498887e-01 4.55785066e-01 -8.74227658e-02 1.35405943e-01
1.08414781e+00 9.26137030e-01 -9.26991820e-01 1.59799349e+00
-4.28591192e-01 3.91041249e-01 1.35339215e-01 -2.86629945e-01
2.01950036e-03 -1.53601304e-01 5.64258397e-01 1.46698833e+00
-1.62923694e-01 1.49224356e-01 4.23147708e-01 9.10323799e-01
3.79393339e-01 -2.25982759e-02 -8.01749885e-01 1.77658871e-01
3.37481797e-01 1.46734595e+00 -6.74112916e-01 -4.90631670e-01
-8.55266690e-01 1.28964639e+00 4.38801885e-01 5.46567321e-01
-3.66096258e-01 -5.17244339e-01 6.92805886e-01 -2.88175642e-01
1.24479353e-01 -2.04584241e-01 -4.39552337e-01 -1.49403095e+00
-3.95213425e-01 -5.03712833e-01 7.79583752e-01 -5.74131548e-01
-1.66974831e+00 3.88399094e-01 4.21211831e-02 -1.00555050e+00
2.00636890e-02 -8.47797275e-01 -8.34894419e-01 7.66238689e-01
-1.42408645e+00 -1.13381016e+00 -4.14740801e-01 8.49958360e-01
9.80817020e-01 -6.25476420e-01 7.99195230e-01 -7.34643638e-02
-5.92960119e-01 5.12032092e-01 -1.56849518e-01 9.71393511e-02
9.74227250e-01 -8.39741766e-01 -2.14569308e-02 8.66552770e-01
3.24039191e-01 1.19695580e+00 4.15018171e-01 -5.42650640e-01
-1.07273102e+00 -8.77005100e-01 8.77346218e-01 -4.00992095e-01
2.55260020e-01 -2.54605800e-01 -9.11705554e-01 4.96630996e-01
4.09282327e-01 1.20167829e-01 8.91279221e-01 3.81267965e-01
-4.76462692e-01 -2.99204826e-01 -1.01156938e+00 7.33351350e-01
1.18414724e+00 -6.20756745e-01 -1.10793614e+00 9.72971991e-02
4.54909243e-02 3.21661204e-01 -1.79158092e-01 5.84693924e-02
5.80336392e-01 -1.19223106e+00 1.30326056e+00 -6.40544295e-01
-2.13510484e-01 -4.99388635e-01 -8.03918466e-02 -1.21340561e+00
-1.24097979e+00 -8.97393599e-02 -3.43041360e-01 1.09779239e+00
-1.09342135e-01 -5.88893056e-01 5.08569479e-01 3.24678332e-01
-1.05099492e-01 -4.44580406e-01 -8.32837462e-01 -8.07917297e-01
-5.86797856e-02 -1.77355215e-01 -8.13315958e-02 6.53109193e-01
1.15391806e-01 6.65956914e-01 -5.65104425e-01 -1.26227699e-02
9.89906073e-01 2.87640601e-01 3.87342513e-01 -1.23293889e+00
-7.15472773e-02 -7.92892158e-01 -5.56864858e-01 -1.31904221e+00
6.17392771e-02 -7.84096301e-01 -1.08969934e-01 -1.58931053e+00
8.76263678e-01 3.60638589e-01 -7.42949843e-01 5.70834160e-01
-2.37586647e-01 5.55917919e-01 2.69060820e-01 5.29606819e-01
-8.55141759e-01 5.78347743e-01 1.31896317e+00 -3.95098567e-01
-1.93481550e-01 1.52724177e-01 -1.13615978e+00 1.07091153e+00
1.04744720e+00 -2.27443147e-02 -4.01266336e-01 -3.10216814e-01
-4.91600662e-01 -7.86052108e-01 5.63432097e-01 -1.03824472e+00
6.61415696e-01 -5.85179925e-01 8.91019583e-01 -6.50614500e-01
4.78675961e-01 -4.26351190e-01 -8.76004457e-01 5.85864902e-01
-2.40811750e-01 -3.59352320e-01 1.17117073e-02 6.23690546e-01
-3.61443698e-01 -3.99569988e-01 1.24799752e+00 -6.22088313e-01
-1.71468055e+00 4.36444104e-01 -1.00380075e+00 -2.85989732e-01
1.41393864e+00 -5.32948911e-01 -4.62921858e-01 -2.03973025e-01
-8.42928171e-01 1.53529542e-02 1.05847672e-01 6.19666874e-01
9.38793242e-01 -1.38037860e+00 -3.58902335e-01 3.25420022e-01
6.47293031e-01 -8.48670483e-01 6.81801319e-01 6.22154653e-01
2.13678464e-01 9.36154068e-01 -4.75136548e-01 -5.08929193e-01
-1.42224348e+00 1.04888225e+00 4.60640341e-01 3.48347068e-01
-5.11523366e-01 9.13418472e-01 8.78644168e-01 -1.83596060e-01
3.46519530e-01 8.57477784e-02 -7.50864923e-01 2.50227004e-02
1.22936285e+00 6.23823285e-01 -2.33793139e-01 -6.09201431e-01
-8.05468976e-01 9.48118091e-01 -2.21593365e-01 2.20587388e-01
9.88900781e-01 -4.85525221e-01 -3.01609486e-01 5.58742881e-01
7.15932071e-01 -1.29683480e-01 -1.01037169e+00 -3.82730395e-01
-1.55168042e-01 -6.16403997e-01 1.80917848e-02 -8.89869809e-01
-9.70287859e-01 1.12031329e+00 8.52488756e-01 -1.52359992e-01
1.25388682e+00 5.23459435e-01 1.17121764e-01 4.62469459e-01
5.33851087e-01 -1.31835735e+00 5.25142312e-01 6.93256974e-01
9.59038734e-01 -1.20485544e+00 -7.64681846e-02 -3.20674688e-01
-9.10451293e-01 1.23394024e+00 1.13465595e+00 -4.39408839e-01
7.71059573e-01 -1.25642151e-01 8.09568614e-02 -1.12011664e-01
-8.17649245e-01 -5.82457364e-01 5.83474040e-01 1.27491975e+00
5.50569713e-01 2.14022920e-02 -2.13774279e-01 6.48507476e-01
3.77935350e-01 -1.08471841e-01 1.17532447e-01 6.39492095e-01
-1.09441650e+00 -1.02452254e+00 -6.43069446e-01 3.01116198e-01
-1.80009324e-02 -1.17642477e-01 -6.38309717e-01 3.20017487e-01
5.07203102e-01 1.16404355e+00 4.38516252e-02 -2.54589945e-01
1.46932006e-01 1.49438214e-02 8.06565762e-01 -7.50975966e-01
-1.85951710e-01 1.28781483e-01 -2.46954992e-01 -5.05285680e-01
-6.31971478e-01 -7.10649669e-01 -8.20078194e-01 -8.21612477e-02
-1.37760893e-01 -1.56043306e-01 1.79963753e-01 9.65107560e-01
2.52149016e-01 3.86800915e-01 9.93789360e-02 -8.66173565e-01
-3.53453875e-01 -9.79040265e-01 -7.97325671e-01 1.55285731e-01
2.36954972e-01 -9.98579144e-01 -1.75639093e-01 8.57580304e-02] | [9.840896606445312, 1.9012864828109741] |
d96eaa64-ce93-4ef7-b5b7-d69e75c2bc15 | a-novel-algorithm-for-exact-concave-hull | 2206.11481 | null | https://arxiv.org/abs/2206.11481v1 | https://arxiv.org/pdf/2206.11481v1.pdf | A Novel Algorithm for Exact Concave Hull Extraction | Region extraction is necessary in a wide range of applications, from object detection in autonomous driving to analysis of subcellular morphology in cell biology. There exist two main approaches: convex hull extraction, for which exact and efficient algorithms exist and concave hulls, which are better at capturing real-world shapes but do not have a single solution. Especially in the context of a uniform grid, concave hull algorithms are largely approximate, sacrificing region integrity for spatial and temporal efficiency. In this study, we present a novel algorithm that can provide vertex-minimized concave hulls with maximal (i.e. pixel-perfect) resolution and is tunable for speed-efficiency tradeoffs. Our method provides advantages in multiple downstream applications including data compression, retrieval, visualization, and analysis. To demonstrate the practical utility of our approach, we focus on image compression. We demonstrate significant improvements through context-dependent compression on disparate regions within a single image (entropy encoding for noisy and predictive encoding for the structured regions). We show that these improvements range from biomedical images to natural images. Beyond image compression, our algorithm can be applied more broadly to aid in a wide range of practical applications for data retrieval, visualization, and analysis. | ['Murat Can Çobanoğlu', 'Kevin Christopher VanHorn'] | 2022-06-23 | null | null | null | null | ['data-compression'] | ['time-series'] | [ 5.29124141e-01 -2.08732024e-01 6.43405947e-04 -6.55767918e-02
-8.33327830e-01 -8.17101777e-01 1.68940336e-01 5.33893287e-01
-3.02786499e-01 6.59888208e-01 -9.69786420e-02 -4.28813338e-01
-1.79645568e-01 -7.46994734e-01 -6.06981993e-01 -6.88237548e-01
-2.90971518e-01 2.42375970e-01 3.80437672e-01 -8.50860029e-02
3.49489272e-01 9.68317151e-01 -1.52435291e+00 2.07683578e-01
8.17045867e-01 1.07303178e+00 2.56210804e-01 7.58739889e-01
-4.80502322e-02 7.87727013e-02 -4.82615948e-01 -1.07919745e-01
2.40520462e-01 -2.43438348e-01 -5.42996466e-01 2.34235227e-02
1.08473688e-01 -2.78224021e-01 -2.16078549e-03 8.36890459e-01
6.08227193e-01 -9.23570469e-02 6.79692328e-01 -8.47596645e-01
-3.03116590e-01 -9.07602534e-03 -8.64992380e-01 1.65838763e-01
5.27180657e-02 1.81847140e-02 8.08912992e-01 -8.04538429e-01
8.30563307e-01 9.38778043e-01 5.08570969e-01 1.97198853e-01
-1.32998955e+00 -2.62677431e-01 -1.33684680e-01 -3.23242880e-02
-1.46487951e+00 -6.26159787e-01 4.78506029e-01 -3.36762518e-01
7.08618522e-01 6.83464825e-01 6.17113709e-01 2.75043808e-02
4.13273036e-01 5.18137634e-01 1.05819058e+00 -2.42358580e-01
4.66768265e-01 -7.87893459e-02 -1.46071047e-01 7.17297137e-01
3.69094849e-01 -6.95886314e-02 -4.94907171e-01 -2.47583330e-01
9.47522283e-01 5.41959144e-02 -5.82744777e-01 -5.54259896e-01
-1.25160456e+00 5.19345939e-01 1.69884756e-01 -1.03310812e-02
-9.40594375e-02 1.86731160e-01 3.55765760e-01 9.82856080e-02
3.21111828e-01 5.10623038e-01 -2.03420341e-01 -1.79940134e-01
-1.05107200e+00 3.31493884e-01 6.28054559e-01 1.10284495e+00
4.93534207e-01 -1.80111706e-01 -6.67031035e-02 5.88942230e-01
-8.32916237e-03 4.22385305e-01 2.74547696e-01 -1.26132727e+00
3.60992141e-02 5.56309164e-01 1.70542940e-01 -1.04394770e+00
-4.50499415e-01 -3.18632513e-01 -8.07351589e-01 5.64041436e-01
4.78329122e-01 2.67070562e-01 -9.05357778e-01 1.48265064e+00
4.28011447e-01 -1.82024568e-01 4.44923714e-02 9.20967519e-01
7.66804218e-01 5.80075204e-01 -2.34123841e-01 -4.42052424e-01
1.68248105e+00 -4.84129608e-01 -6.16732657e-01 -6.57567428e-03
6.59956396e-01 -5.45956612e-01 9.16462958e-01 5.22474706e-01
-1.31515276e+00 2.40095984e-02 -1.14511549e+00 -4.86738384e-01
-4.91481155e-01 -9.29868221e-02 2.60701448e-01 3.82022172e-01
-1.14821994e+00 4.70734715e-01 -8.36870074e-01 -2.43870914e-01
7.16646075e-01 4.40063030e-01 -3.86114478e-01 4.84494045e-02
-5.31646967e-01 6.41988993e-01 2.17474811e-02 -2.92886913e-01
-3.13914061e-01 -6.37331545e-01 -6.36997938e-01 3.20703387e-01
1.14545353e-01 -5.90689838e-01 8.61239433e-01 -2.75040925e-01
-8.91336560e-01 9.77066457e-01 -4.85667646e-01 -6.72441304e-01
3.67682546e-01 2.58267760e-01 1.71482012e-01 4.59462792e-01
-4.57713157e-02 7.11244226e-01 4.17505562e-01 -1.31176841e+00
-4.78405863e-01 -5.25980234e-01 -3.30088615e-01 3.89136404e-01
-4.01855648e-01 -1.34682238e-01 -6.04031503e-01 -4.94429648e-01
3.02834600e-01 -7.00156152e-01 -2.74459779e-01 8.35293293e-01
-5.33621646e-02 2.85772771e-01 1.05940628e+00 -6.00688457e-01
1.23182690e+00 -2.19273233e+00 3.85757275e-02 3.01049381e-01
5.87859452e-01 1.16793260e-01 2.35452384e-01 4.32869613e-01
3.92116696e-01 2.95132816e-01 -4.80376244e-01 -2.16947585e-01
-3.62048179e-01 1.32434189e-01 -3.71876299e-01 6.84660435e-01
7.73272514e-02 9.78400469e-01 -7.26625562e-01 -9.19397414e-01
1.11842304e-01 6.58634067e-01 -4.71932203e-01 -1.81452900e-01
-1.29867911e-01 4.26250547e-01 -4.63417977e-01 8.24000895e-01
6.84109747e-01 -4.80936527e-01 1.16977416e-01 -2.48037592e-01
-3.15229356e-01 -2.22985744e-01 -8.51656199e-01 1.47864532e+00
-1.50063962e-01 8.23756278e-01 5.82023561e-01 -7.89131224e-01
9.38446403e-01 1.13911247e-02 7.95521379e-01 -4.53633428e-01
5.58752827e-02 2.72576094e-01 -1.54564232e-01 -1.55257642e-01
7.46253848e-01 7.10827261e-02 7.44388551e-02 4.08884525e-01
-6.01503611e-01 -4.41914082e-01 2.94451684e-01 1.76352978e-01
1.24687231e+00 -5.98236993e-02 5.63851595e-01 -4.22408283e-01
-1.92567427e-02 2.56488681e-01 4.58112121e-01 3.69475871e-01
-4.40470546e-01 1.04391253e+00 5.13870358e-01 -2.49453098e-01
-1.29791677e+00 -9.27514851e-01 -5.73995650e-01 7.28945673e-01
6.09711766e-01 -4.82793123e-01 -6.07453167e-01 2.71706712e-02
1.35857582e-01 2.46886343e-01 -3.03854346e-01 8.70876089e-02
-6.52692676e-01 -6.14408851e-01 4.02623385e-01 4.66111571e-01
3.03023964e-01 -7.09416866e-01 -1.36654842e+00 1.69628054e-01
-2.27290206e-02 -1.12531948e+00 -5.32676160e-01 2.84931272e-01
-1.01265645e+00 -8.76893997e-01 -8.03072393e-01 -7.29723990e-01
8.00131798e-01 5.27258515e-01 8.91384602e-01 2.43713856e-01
-6.32301390e-01 2.90442765e-01 -1.07557967e-01 -5.15817702e-01
-2.14265391e-01 -2.76111096e-01 -2.60000587e-01 -3.80922914e-01
-2.27768213e-01 -6.34146631e-01 -9.02764559e-01 3.14226210e-01
-1.07522333e+00 1.99707702e-01 5.18368661e-01 8.30128551e-01
1.29134834e+00 -8.71802196e-02 3.06576848e-01 -7.10670054e-01
5.89780807e-01 -3.38667482e-01 -6.57724917e-01 1.20705962e-01
-6.36581361e-01 1.01390615e-01 7.14895308e-01 -1.72626823e-01
-5.24583995e-01 2.24941358e-01 2.61740666e-02 -5.58790088e-01
8.05182308e-02 4.76216465e-01 1.24376878e-01 -3.57570142e-01
6.99675381e-01 4.94507432e-01 4.87536579e-01 -1.75718427e-01
3.45838845e-01 5.62883914e-01 7.56327033e-01 -4.74748194e-01
2.94393808e-01 1.10939121e+00 4.98484492e-01 -1.01021004e+00
-1.30108356e-01 -7.09517837e-01 -3.65524560e-01 -1.39163584e-01
6.11968219e-01 -5.24921238e-01 -6.94059670e-01 -2.88589075e-02
-1.01411557e+00 -1.84853047e-01 -4.51909035e-01 8.99802968e-02
-8.75830889e-01 5.10635376e-01 -5.17018080e-01 -8.15626204e-01
-3.91519368e-01 -1.26222849e+00 1.50695097e+00 1.28919914e-01
-1.06543742e-01 -7.36855984e-01 -1.99165925e-01 4.40389896e-03
3.17805976e-01 5.80143392e-01 9.32943165e-01 -2.10873976e-01
-7.94435382e-01 2.64119282e-02 -3.63164812e-01 -4.19291466e-01
6.25361651e-02 1.08156197e-01 -8.16368639e-01 -4.47292417e-01
-2.59157717e-01 -1.85096234e-01 9.11749661e-01 5.81862688e-01
1.42129397e+00 -3.24637502e-01 -8.02691817e-01 9.64934886e-01
1.45872438e+00 3.17369074e-01 7.55512893e-01 5.31492084e-02
3.06009233e-01 5.87150514e-01 5.33452868e-01 4.67955709e-01
1.31664753e-01 7.04082787e-01 5.27903140e-01 -4.38566744e-01
-1.94777489e-01 1.70638502e-01 -1.61200389e-01 5.37136555e-01
7.98682049e-02 -4.08530027e-01 -8.26137662e-01 6.53776109e-01
-1.74175310e+00 -8.69808197e-01 5.47154620e-02 2.42705798e+00
8.37148547e-01 -4.19214331e-02 1.06536992e-01 1.13939382e-01
6.10339046e-01 -1.06829993e-01 -8.76092374e-01 -4.74985361e-01
-2.66198874e-01 -3.62667255e-02 6.84813559e-01 3.55777830e-01
-9.69319403e-01 4.22416210e-01 6.93648863e+00 8.75599325e-01
-1.13042963e+00 -2.41954669e-01 9.41176593e-01 -1.76745325e-01
-5.48903883e-01 -1.34876490e-01 -6.92881107e-01 3.27164322e-01
6.50430501e-01 -3.33018869e-01 2.38692209e-01 6.08517826e-01
3.64413232e-01 -2.02011690e-01 -9.23584223e-01 1.19502676e+00
-2.04324678e-01 -1.70274830e+00 -1.05885305e-01 5.74354768e-01
5.78388035e-01 -2.59373963e-01 2.22195044e-01 -4.34720725e-01
2.93247327e-02 -1.10244071e+00 6.51746929e-01 2.81426013e-01
1.31715977e+00 -7.14206219e-01 2.84184277e-01 3.59494209e-01
-1.34170055e+00 4.50743847e-02 -3.90104592e-01 2.84632653e-01
3.34337115e-01 7.50261843e-01 -8.12733173e-01 3.34406942e-01
5.64952672e-01 3.95024627e-01 -2.06795216e-01 1.15783739e+00
7.80636132e-01 2.36533403e-01 -6.59452319e-01 -4.72465977e-02
-1.36798874e-01 -2.93481976e-01 7.23442793e-01 1.31449878e+00
4.58926409e-01 4.24677700e-01 -2.20968369e-02 8.57081831e-01
-3.12514082e-02 2.36055508e-01 -9.76947069e-01 5.47459498e-02
8.06767404e-01 1.11082423e+00 -1.29035521e+00 -2.35528544e-01
-3.55799109e-01 7.62075722e-01 4.53413010e-01 3.02628964e-01
-7.04335630e-01 -6.16471291e-01 7.46132970e-01 5.47304571e-01
4.33527023e-01 -5.47712445e-01 -8.67053688e-01 -7.54932225e-01
2.60255523e-02 -5.84687829e-01 2.97759444e-01 -6.00167394e-01
-6.61421299e-01 3.29198211e-01 8.35667029e-02 -1.45953810e+00
-6.13836758e-02 -5.62569737e-01 -2.73474127e-01 5.65307438e-01
-1.51316237e+00 -8.12661767e-01 -3.12460423e-01 1.18626766e-01
2.63714224e-01 3.02846104e-01 6.71084642e-01 1.54205933e-01
-1.26180053e-01 3.88592929e-01 5.97133338e-01 -1.72162265e-01
5.06555617e-01 -1.11815107e+00 3.04116964e-01 7.27913022e-01
-5.84265217e-02 6.22082710e-01 7.31593430e-01 -5.14201224e-01
-1.73336136e+00 -1.12072587e+00 4.49381202e-01 -4.38270308e-02
4.39079404e-02 -1.92139402e-01 -8.76947224e-01 3.11879128e-01
-1.24905109e-01 8.56272131e-02 5.31242490e-01 -5.03678143e-01
3.92915402e-03 -4.31817509e-02 -1.37148273e+00 9.37414169e-01
8.47843170e-01 -1.81993395e-01 1.53881446e-01 2.69787818e-01
5.01059949e-01 -7.54457951e-01 -1.02049470e+00 4.22667742e-01
8.03991199e-01 -1.01197577e+00 1.07518852e+00 -1.95805967e-01
3.72327447e-01 -6.42857909e-01 -3.48389149e-02 -9.22332823e-01
-3.09074730e-01 -7.10928142e-01 -2.05495760e-01 6.99818015e-01
2.84657061e-01 -5.87544560e-01 7.98311532e-01 5.93632400e-01
-3.19181353e-01 -1.33911073e+00 -1.06171846e+00 -6.79298043e-01
6.63832901e-03 -2.32172340e-01 6.36577904e-01 5.10510087e-01
2.85545021e-01 -1.25620618e-01 1.07866660e-01 -7.10269511e-02
6.00917935e-01 6.01773679e-01 6.55988514e-01 -9.22174513e-01
-4.78904732e-02 -6.33880973e-01 -7.06184328e-01 -1.50274253e+00
-4.77922708e-01 -7.58144617e-01 2.46799245e-01 -1.54843402e+00
2.63783574e-01 -6.70406580e-01 1.83128864e-01 3.55384618e-01
2.86031533e-02 5.15958369e-01 5.12628630e-02 4.92740750e-01
-4.04406697e-01 2.75265962e-01 1.47996211e+00 -3.37289199e-02
-3.32280129e-01 -4.00875837e-01 -7.87678123e-01 4.11100686e-01
7.57818460e-01 -9.19880271e-02 -5.14115572e-01 -3.63191128e-01
7.69516220e-03 1.98441222e-01 2.75333732e-01 -9.45311308e-01
3.24376822e-01 -2.07676232e-01 3.67900670e-01 -5.98202527e-01
5.86633384e-01 -6.54832661e-01 2.36160696e-01 3.44276875e-01
-1.68793634e-01 1.39933184e-01 2.77638793e-01 6.87026799e-01
-2.64356490e-02 -5.35890833e-02 1.01440823e+00 -1.45739704e-01
-4.58371907e-01 3.76263589e-01 -2.42203444e-01 -9.75195039e-03
1.23913705e+00 -8.59906852e-01 -4.53625292e-01 -6.06020093e-01
-5.51401138e-01 1.94770649e-01 1.01693487e+00 -3.08553219e-01
9.80780542e-01 -1.14425468e+00 -7.32845187e-01 2.81882197e-01
2.58223861e-02 3.62743169e-01 -1.13173634e-01 8.36047947e-01
-9.93748128e-01 3.89568627e-01 -7.35459924e-02 -9.21048343e-01
-1.38899565e+00 4.97095913e-01 1.69616789e-01 -1.38324082e-01
-7.61656940e-01 5.18029451e-01 6.19334698e-01 2.18080595e-01
7.97272287e-03 -5.13613880e-01 1.12678073e-02 -3.87467086e-01
5.31487286e-01 3.41530442e-01 2.27787957e-01 -6.49519563e-01
-2.24503264e-01 7.84117997e-01 1.35581046e-01 -1.73775613e-01
1.09073138e+00 -3.28056514e-01 -7.77355656e-02 3.01027626e-01
1.11038327e+00 1.77624792e-01 -1.38867748e+00 2.47014925e-01
-4.25169736e-01 -4.57946002e-01 5.35600483e-02 -3.06229919e-01
-1.04058444e+00 8.10332000e-01 4.52856213e-01 4.15122688e-01
1.35966814e+00 1.50864035e-01 9.41308796e-01 1.00390926e-01
4.36926067e-01 -7.92577505e-01 -3.09625149e-01 6.85349181e-02
8.64987195e-01 -8.24760079e-01 4.69387740e-01 -6.88630044e-01
-3.84624630e-01 1.18967927e+00 1.97827697e-01 1.93883538e-01
6.22171819e-01 9.81031239e-01 -1.35424346e-01 -3.47355872e-01
-9.04768407e-01 -1.92632854e-01 1.08046733e-01 6.90401077e-01
3.36522073e-01 -1.64646227e-02 -4.77189511e-01 1.71782166e-01
-1.43874153e-01 -9.36426744e-02 4.67216074e-01 1.29490924e+00
-8.07797968e-01 -7.26607203e-01 -4.81745303e-01 8.81170809e-01
-3.99057239e-01 3.15174162e-02 -1.99533522e-01 5.56797147e-01
-2.20765591e-01 7.20388591e-01 2.27640361e-01 -1.20587215e-01
7.70947635e-02 -2.75034547e-01 4.42805022e-01 -2.90937066e-01
-1.59198433e-01 2.94065267e-01 -1.01782888e-01 -6.16201639e-01
-1.13533005e-01 -6.64967597e-01 -1.66548729e+00 -3.61645162e-01
-3.18112999e-01 -3.06335632e-02 5.97488523e-01 5.15405357e-01
8.86926770e-01 2.58887202e-01 2.87122071e-01 -7.46466279e-01
-3.18382114e-01 -2.51631826e-01 -6.62577868e-01 2.55909741e-01
4.53518033e-01 -4.68677282e-01 -1.33988634e-01 5.33696651e-01] | [14.488420486450195, -3.1596810817718506] |
59f00095-9049-42e5-b1c0-3c959b682e78 | qbye-mlpmixer-query-by-example-open | 2206.13231 | null | https://arxiv.org/abs/2206.13231v1 | https://arxiv.org/pdf/2206.13231v1.pdf | QbyE-MLPMixer: Query-by-Example Open-Vocabulary Keyword Spotting using MLPMixer | Current keyword spotting systems are typically trained with a large amount of pre-defined keywords. Recognizing keywords in an open-vocabulary setting is essential for personalizing smart device interaction. Towards this goal, we propose a pure MLP-based neural network that is based on MLPMixer - an MLP model architecture that effectively replaces the attention mechanism in Vision Transformers. We investigate different ways of adapting the MLPMixer architecture to the QbyE open-vocabulary keyword spotting task. Comparisons with the state-of-the-art RNN and CNN models show that our method achieves better performance in challenging situations (10dB and 6dB environments) on both the publicly available Hey-Snips dataset and a larger scale internal dataset with 400 speakers. Our proposed model also has a smaller number of parameters and MACs compared to the baseline models. | ['Chul Lee', 'Han Suk Shim', 'Qianhui Wan', 'Waseem Gharbieh', 'Jinmiao Huang'] | 2022-06-23 | null | null | null | null | ['keyword-spotting'] | ['speech'] | [ 1.15938470e-01 -1.07859552e-01 -3.71528864e-01 -4.46069568e-01
-8.02595496e-01 -4.65162188e-01 6.03028297e-01 -3.59979272e-01
-7.86049128e-01 2.41839394e-01 3.56125832e-01 -4.99543101e-01
2.78877951e-02 -1.24442898e-01 -9.41120207e-01 -2.85060376e-01
4.33141172e-01 6.04328573e-01 -1.01233356e-01 -5.25484495e-02
-9.94390473e-02 2.80873895e-01 -1.59440672e+00 2.62816966e-01
1.87452987e-01 1.38051009e+00 4.47105646e-01 1.02217090e+00
-2.47041464e-01 6.72471702e-01 -7.09744155e-01 -5.35334706e-01
3.14717352e-01 2.39924103e-01 -7.06250131e-01 -3.33911687e-01
8.97087336e-01 -4.39353883e-01 -8.30785275e-01 8.46414030e-01
9.90996957e-01 8.90869647e-02 3.18064213e-01 -1.05196846e+00
-1.07077348e+00 1.13006496e+00 2.83933664e-03 4.36275244e-01
2.97119796e-01 2.41030142e-01 1.08741355e+00 -9.70548749e-01
1.05955765e-01 1.37461102e+00 7.39319563e-01 5.74498534e-01
-1.00271857e+00 -8.84326994e-01 3.63933772e-01 5.06062090e-01
-1.70080841e+00 -6.25704229e-01 4.66749668e-01 1.85682643e-02
1.48653686e+00 5.51004529e-01 3.84692311e-01 1.94324422e+00
-2.37640560e-01 9.83338475e-01 6.45253837e-01 -3.42728436e-01
-7.32446685e-02 4.74447161e-01 2.96842068e-01 2.94284314e-01
-2.42423341e-01 -1.98354170e-01 -9.54130650e-01 -1.27218783e-01
4.24760073e-01 3.72199118e-02 -4.54358697e-01 2.23210886e-01
-1.37169325e+00 5.23846209e-01 7.44449720e-02 1.87762350e-01
-3.51460576e-01 4.51857448e-01 3.78532857e-01 1.93371251e-01
2.07333535e-01 3.79764259e-01 -8.31836224e-01 -4.45668519e-01
-1.12671733e+00 1.12632751e-01 1.15462637e+00 1.34944785e+00
3.82522315e-01 -1.63615972e-01 -8.11519027e-01 1.12523437e+00
4.82920319e-01 7.35412419e-01 9.07517672e-01 -7.58861125e-01
5.26541293e-01 7.44908005e-02 -5.11905514e-02 -6.00295663e-01
-1.39907300e-01 -5.70937395e-01 -6.13168776e-01 -4.74580288e-01
6.30001351e-02 -1.40941711e-02 -1.16655624e+00 1.36838162e+00
-2.12227389e-01 2.50873297e-01 -1.16452709e-01 5.44126451e-01
1.12623572e+00 8.66027296e-01 2.79212035e-02 1.20095477e-01
1.66203177e+00 -1.40139496e+00 -1.08091640e+00 -2.60127008e-01
-5.89118861e-02 -6.66858494e-01 1.67002189e+00 5.60154676e-01
-8.66909385e-01 -7.66039252e-01 -1.16283858e+00 -4.53809828e-01
-8.62376034e-01 3.81365210e-01 3.06385398e-01 1.00000453e+00
-1.27856493e+00 2.47592494e-01 -4.78627861e-01 -5.35632551e-01
4.05456483e-01 5.41197121e-01 8.94417986e-02 3.28444630e-01
-1.24372029e+00 7.22894967e-01 3.52686018e-01 -6.23567924e-02
-1.03171742e+00 -7.22534955e-01 -4.92567033e-01 5.41534722e-01
5.98936558e-01 -4.88640636e-01 1.88797593e+00 -6.57612145e-01
-2.12929654e+00 6.71749115e-01 -2.26329759e-01 -6.79753959e-01
2.98127472e-01 -6.82174146e-01 -6.17985666e-01 1.17752999e-02
-2.19769716e-01 8.82289886e-01 1.17978501e+00 -7.68248618e-01
-5.99473417e-01 -1.29613364e-02 9.50955972e-02 3.72398138e-01
-9.37268496e-01 2.02417076e-01 -1.38291883e+00 -8.96537364e-01
-6.64356947e-01 -8.35910678e-01 1.54062346e-01 -9.05307084e-02
-7.45736480e-01 -4.90858853e-01 8.01091492e-01 -6.86185956e-01
1.36592412e+00 -2.03510380e+00 -1.56554997e-01 4.90144528e-02
2.02810705e-01 5.50202608e-01 -7.16784671e-02 2.03748375e-01
2.39985719e-01 1.42008036e-01 1.81571558e-01 -7.70506561e-01
3.34503442e-01 7.97956064e-02 -4.55064893e-01 1.32018998e-01
-4.06955391e-01 1.18680334e+00 -4.68983024e-01 -4.25954551e-01
2.68655550e-03 7.00498044e-01 -2.84641862e-01 3.22130114e-01
-4.78093982e-01 5.81880249e-02 -1.51491329e-01 8.89520049e-01
2.62682080e-01 -4.41139221e-01 -8.31055418e-02 -2.11436436e-01
2.05273032e-01 4.65400517e-01 -1.12702477e+00 2.00156355e+00
-6.48057759e-01 8.83923590e-01 2.77004212e-01 -6.88688815e-01
5.67976654e-01 5.62440157e-01 9.19766203e-02 -7.91332245e-01
3.43628645e-01 6.04915479e-03 -4.46446985e-01 -6.17772460e-01
6.81936383e-01 6.21128023e-01 6.98847771e-02 4.84376490e-01
5.37582934e-01 2.71274209e-01 -8.05917978e-02 1.59070585e-02
1.09350264e+00 -2.26363122e-01 2.31601410e-02 -2.03257680e-01
5.25088072e-01 -6.57408595e-01 1.02506310e-01 1.38929474e+00
-2.02186853e-01 5.09354174e-01 5.50791509e-02 -2.87005395e-01
-7.44209051e-01 -7.66591072e-01 -1.46102324e-01 1.53915870e+00
-3.27101722e-02 -6.98943794e-01 -7.83393085e-01 -5.94270408e-01
6.29831702e-02 5.71599483e-01 -3.42252463e-01 2.30826121e-02
-2.38170788e-01 -4.66583401e-01 1.10806441e+00 5.19487083e-01
5.80310225e-01 -1.21030116e+00 -5.31465411e-01 4.08980623e-02
-1.61269516e-01 -1.38764882e+00 -1.04584801e+00 4.63392913e-01
-2.81480938e-01 -5.69620192e-01 -1.04472780e+00 -7.77305722e-01
-4.80242111e-02 1.01688802e-01 1.28807545e+00 -4.02130395e-01
-3.91676575e-02 7.85793304e-01 -2.61595309e-01 -9.54834342e-01
-1.15613975e-01 5.65544486e-01 2.47619808e-01 2.70029157e-01
9.10733402e-01 -4.22676086e-01 -6.21521890e-01 1.96218029e-01
-6.40259624e-01 -4.09287930e-01 8.71503115e-01 6.29852653e-01
5.61142147e-01 -2.94379145e-01 1.96140721e-01 -6.29878461e-01
1.10944271e+00 -3.74371171e-01 -5.43411434e-01 4.94590074e-01
-8.96954894e-01 2.33955592e-01 3.06722641e-01 -1.13316774e+00
-7.27478147e-01 -8.84632766e-02 -2.90409923e-01 -6.86182320e-01
-2.43941814e-01 2.16403887e-01 -5.03357470e-01 -3.52813154e-02
4.28424686e-01 2.61063218e-01 -2.96975493e-01 -1.00350881e+00
7.29622960e-01 1.40869606e+00 7.33723104e-01 -1.93573117e-01
5.21991074e-01 2.18582153e-01 -7.24540114e-01 -8.56180608e-01
-2.93011248e-01 -7.11901248e-01 -1.18710510e-01 2.48915091e-01
7.96988606e-01 -1.19172287e+00 -1.18769264e+00 5.17197311e-01
-1.27341628e+00 -3.96233469e-01 -1.71787530e-01 3.75475943e-01
-3.10872644e-01 2.71544904e-01 -6.28235340e-01 -6.77675426e-01
-9.34896052e-01 -1.40276396e+00 1.28417337e+00 3.26043755e-01
-3.72075856e-01 -4.84530836e-01 -5.56516461e-02 2.38262579e-01
8.70282769e-01 -9.22937453e-01 5.34927845e-01 -1.26968586e+00
-3.79019976e-01 -1.51088178e-01 -1.50186047e-01 4.37655389e-01
-1.45116702e-01 -4.50133830e-01 -1.56735551e+00 5.66554591e-02
9.70947444e-02 -3.69368613e-01 9.91271138e-01 2.95315981e-01
1.97999418e+00 -5.74418008e-01 -4.36588138e-01 9.21321154e-01
9.18078661e-01 3.05944443e-01 3.42879176e-01 2.64294505e-01
9.24054325e-01 1.52900577e-01 -4.09748219e-02 2.34980077e-01
6.52890921e-01 9.17191207e-01 2.72335947e-01 3.46169109e-03
-1.33951530e-01 -2.46469796e-01 2.11246789e-01 6.45991862e-01
5.21652699e-01 -5.62917709e-01 -7.62836993e-01 6.08590662e-01
-1.77221107e+00 -6.78595841e-01 5.49313903e-01 1.76809788e+00
1.23761165e+00 2.34838545e-01 6.35589063e-02 3.83789302e-03
4.62593257e-01 2.65190929e-01 -7.20541596e-01 -5.57462394e-01
-1.63518220e-01 4.71904874e-01 7.31345713e-01 5.40793419e-01
-1.25449908e+00 9.06912744e-01 6.95879221e+00 1.12382352e+00
-1.22910738e+00 3.79177511e-01 6.43967748e-01 -6.14612103e-01
-9.14345160e-02 -6.78014338e-01 -1.27912080e+00 7.33534038e-01
1.46676147e+00 3.06178361e-01 9.29633498e-01 1.01984882e+00
-1.81310490e-01 3.97892147e-01 -1.47540188e+00 1.82657111e+00
4.85093236e-01 -1.22330511e+00 -1.08730271e-01 5.43818064e-02
1.87008157e-01 4.09741908e-01 3.95492733e-01 6.52958870e-01
2.52501160e-01 -1.39679050e+00 9.25069690e-01 5.57270765e-01
1.07230043e+00 -3.46644878e-01 4.88162249e-01 1.47339657e-01
-7.73799837e-01 -1.75536752e-01 -5.30816242e-02 2.62832791e-01
-6.06281683e-02 3.18466991e-01 -9.92324173e-01 -2.28567183e-01
1.36978090e+00 3.72780293e-01 -7.11200774e-01 8.55780065e-01
-7.84870312e-02 6.77869737e-01 -6.52710855e-01 -3.76456678e-01
6.66746870e-02 4.32806969e-01 4.63380396e-01 1.51229239e+00
7.82618895e-02 -3.53312939e-01 -1.47475660e-01 6.79418683e-01
-7.36076891e-01 5.55487685e-02 -3.17998499e-01 -2.87468225e-01
7.02850640e-01 1.31404316e+00 -2.67925054e-01 -4.98598456e-01
-6.64767861e-01 1.32918513e+00 2.27210373e-01 7.52971232e-01
-7.87862360e-01 -5.62711000e-01 8.67038548e-01 -2.49979615e-01
8.76732469e-01 2.68418062e-02 -1.60546854e-01 -1.01264441e+00
2.79076368e-01 -1.35227561e+00 1.06249139e-01 -6.89365745e-01
-1.13232732e+00 8.67583752e-01 1.08127512e-01 -6.51598155e-01
-1.94474399e-01 -7.99832940e-01 -3.07251155e-01 9.74065661e-01
-1.52761114e+00 -1.14548969e+00 -2.42275372e-01 6.34252787e-01
9.25570130e-01 -4.93734092e-01 1.06745732e+00 6.31837130e-01
-6.05825186e-01 1.23763573e+00 2.66428113e-01 1.72417443e-02
9.16318655e-01 -1.24559796e+00 8.29489172e-01 6.05055392e-01
6.31258905e-01 9.03175652e-01 5.55060267e-01 -3.40072036e-01
-1.52754724e+00 -8.72391582e-01 9.44218755e-01 -9.09192502e-01
5.26389301e-01 -9.37522233e-01 -6.74883544e-01 8.09744298e-01
6.91432714e-01 5.51271029e-02 8.28783810e-01 1.87010899e-01
-4.76177156e-01 -4.46349919e-01 -7.75960505e-01 6.46332622e-01
1.01763773e+00 -1.14873099e+00 -4.45126563e-01 2.85741538e-01
1.27384269e+00 -4.27177668e-01 -7.31509566e-01 2.95236558e-01
1.01525843e+00 -2.35933512e-01 1.30505860e+00 -4.35392857e-01
-2.61436880e-01 1.54119462e-01 -3.65631133e-01 -1.00271082e+00
-9.58293676e-02 -1.10877776e+00 -8.47745121e-01 1.21362209e+00
5.29661894e-01 -3.33845705e-01 6.44449353e-01 5.60005784e-01
5.18748835e-02 -8.04705381e-01 -1.11282492e+00 -5.44085383e-01
-3.52259666e-01 -8.52174819e-01 8.65441024e-01 4.48278040e-01
-3.03664565e-01 6.60389125e-01 -5.75523436e-01 5.39192408e-02
3.45909655e-01 -5.19520283e-01 7.23510265e-01 -9.47014689e-01
-6.23955965e-01 -5.56919992e-01 6.19480088e-02 -1.65218616e+00
2.44187325e-01 -5.19039333e-01 6.94015771e-02 -1.06978202e+00
-3.59592661e-02 -1.85075134e-01 -5.76986551e-01 5.21303535e-01
-2.11376980e-01 2.72210330e-01 2.05260515e-01 1.33592933e-01
-8.46831620e-01 6.05220258e-01 4.30532604e-01 -8.66129935e-01
-3.04890245e-01 1.18463211e-01 -9.08357620e-01 4.89490926e-01
6.12071276e-01 -3.10097218e-01 -5.02923965e-01 -6.96888030e-01
1.66404247e-01 -5.35647690e-01 8.55088905e-02 -9.75808680e-01
5.12846828e-01 3.38936210e-01 2.17113107e-01 -7.64825046e-01
6.77942753e-01 -9.73635495e-01 -2.34557971e-01 -1.48371458e-01
-7.07279086e-01 6.51698560e-02 4.00798589e-01 8.10783505e-01
2.38293827e-01 -1.63793191e-01 1.06880322e-01 -2.12868273e-01
-6.89823031e-01 2.93253839e-01 -5.42255759e-01 -2.52850503e-01
4.66033936e-01 5.36594316e-02 -1.70239627e-01 -6.39999390e-01
-5.82299471e-01 3.60533595e-01 -1.19366504e-01 8.89719307e-01
5.32012045e-01 -1.10453141e+00 -2.52911001e-01 2.47984961e-01
3.23979944e-01 -2.76011705e-01 -2.56172776e-01 5.12452781e-01
-2.39342138e-01 1.10625088e+00 2.01906234e-01 -5.82957685e-01
-1.12686813e+00 7.19427884e-01 5.04136562e-01 -2.55437374e-01
-4.94969755e-01 1.45814073e+00 -7.54013509e-02 -5.28483987e-01
1.45743024e+00 -8.89848888e-01 -3.80802453e-01 -1.14843547e-01
8.78518522e-01 1.74887329e-01 3.84040415e-01 -5.13987958e-01
-3.51597756e-01 2.30571091e-01 -2.81486690e-01 -3.98791373e-01
8.34546983e-01 -2.65955329e-01 2.97345936e-01 6.42597377e-01
1.28225791e+00 -1.42642632e-01 -1.04025698e+00 -6.60657167e-01
-8.20703432e-02 3.66084687e-02 4.42426145e-01 -1.11997592e+00
-7.43058860e-01 8.14286828e-01 1.09435058e+00 -3.11431494e-02
1.01547730e+00 1.33305892e-01 9.54436719e-01 1.02958810e+00
-1.09089948e-01 -1.24106526e+00 -4.23183385e-03 4.43063587e-01
9.24714208e-01 -1.35488391e+00 -2.08088979e-01 1.42557919e-01
-5.74733138e-01 8.99725556e-01 4.64293361e-01 4.45233136e-01
8.81006718e-01 3.63758057e-01 2.30575472e-01 -3.06533456e-01
-6.92941129e-01 -1.59871057e-01 6.72363818e-01 5.28485477e-01
1.39556667e-02 -5.27335517e-02 1.56205043e-01 1.12309837e+00
-3.17586750e-01 -1.30280942e-01 -2.87316795e-02 6.58719301e-01
6.95895776e-02 -1.08050215e+00 -3.94839853e-01 5.20840287e-01
-9.26754951e-01 -5.36268651e-01 -4.41503406e-01 3.15495968e-01
-3.16868946e-02 1.10186195e+00 6.55984282e-02 -6.93202615e-01
2.01151922e-01 4.79621947e-01 1.31304190e-01 -6.01691902e-01
-8.97628427e-01 1.88048527e-01 1.09598547e-01 -8.25917244e-01
-3.07059854e-01 -1.51528940e-01 -5.53456187e-01 8.34309086e-02
-2.57768452e-01 -3.99822183e-02 9.76912260e-01 8.58491063e-01
7.05621421e-01 7.30580389e-01 4.57825363e-02 -9.17850018e-01
-7.79666901e-01 -1.61005139e+00 -4.16354328e-01 1.46004958e-02
5.65136790e-01 -3.92222345e-01 2.59048529e-02 1.53228551e-01] | [14.247093200683594, 6.429457187652588] |
40521dac-790c-48d2-9d80-04cdcfa9c396 | seqsleepnet-end-to-end-hierarchical-recurrent | 1809.10932 | null | http://arxiv.org/abs/1809.10932v3 | http://arxiv.org/pdf/1809.10932v3.pdf | SeqSleepNet: End-to-End Hierarchical Recurrent Neural Network for Sequence-to-Sequence Automatic Sleep Staging | Automatic sleep staging has been often treated as a simple classification
problem that aims at determining the label of individual target polysomnography
(PSG) epochs one at a time. In this work, we tackle the task as a
sequence-to-sequence classification problem that receives a sequence of
multiple epochs as input and classifies all of their labels at once. For this
purpose, we propose a hierarchical recurrent neural network named SeqSleepNet.
At the epoch processing level, the network consists of a filterbank layer
tailored to learn frequency-domain filters for preprocessing and an
attention-based recurrent layer designed for short-term sequential modelling.
At the sequence processing level, a recurrent layer placed on top of the
learned epoch-wise features for long-term modelling of sequential epochs. The
classification is then carried out on the output vectors at every time step of
the top recurrent layer to produce the sequence of output labels. Despite being
hierarchical, we present a strategy to train the network in an end-to-end
fashion. We show that the proposed network outperforms state-of-the-art
approaches, achieving an overall accuracy, macro F1-score, and Cohen's kappa of
87.1%, 83.3%, and 0.815 on a publicly available dataset with 200 subjects. | ['Oliver Y. Chén', 'Navin Cooray', 'Fernando Andreotti', 'Huy Phan', 'Maarten De Vos'] | 2018-09-28 | null | null | null | null | ['sleep-stage-detection', 'sleep-staging'] | ['medical', 'medical'] | [ 6.17480695e-01 6.54364675e-02 -1.66949436e-01 -6.02804542e-01
-7.42151678e-01 -2.39593014e-01 1.02873996e-01 1.43537462e-01
-7.11950302e-01 6.10745013e-01 1.50991336e-01 -3.06002349e-01
6.74457997e-02 -1.18918456e-01 -2.73231566e-01 -8.30911994e-01
-1.29676433e-02 1.35955602e-01 1.04778253e-01 1.41955256e-01
1.80694893e-01 1.75007924e-01 -1.65340126e+00 5.00465810e-01
6.75625622e-01 1.39317608e+00 2.52312750e-01 8.74994338e-01
1.91360071e-01 9.81592536e-01 -6.41996205e-01 -2.77649611e-01
-1.95824042e-01 -7.21262217e-01 -8.12398195e-01 8.36002380e-02
1.56052455e-01 3.42801571e-01 -1.75159946e-01 1.07735968e+00
6.86845958e-01 3.66509110e-01 4.04921472e-01 -7.11530745e-01
-8.73615500e-03 5.36749542e-01 -1.44513309e-01 7.68671036e-01
1.23856448e-01 -8.62133652e-02 1.26118100e+00 -6.37384892e-01
2.79254466e-01 5.48057377e-01 6.95730090e-01 9.52985168e-01
-1.27734208e+00 -6.28796339e-01 2.23453924e-01 3.58132690e-01
-1.18341720e+00 -7.40306199e-01 6.75091684e-01 -3.98339123e-01
1.07008922e+00 3.09237331e-01 6.14770949e-01 1.22843206e+00
6.31258428e-01 5.58156133e-01 9.72547293e-01 -3.52289379e-01
3.49777699e-01 -5.20326234e-02 5.31823754e-01 5.60146272e-01
-3.77276301e-01 -1.29698858e-01 -5.84695697e-01 -1.02711711e-02
1.89617664e-01 2.90484399e-01 -1.43864438e-01 3.27988416e-01
-8.31665576e-01 5.27855933e-01 5.22880018e-01 6.22860372e-01
-7.89750993e-01 1.23452144e-02 6.25575900e-01 2.33013153e-01
7.27652431e-01 2.06560224e-01 -4.19356942e-01 -3.43484879e-01
-1.50100589e+00 -2.33855397e-01 7.32131243e-01 5.15746713e-01
3.83997858e-01 -1.50381267e-01 -5.17955065e-01 1.03446090e+00
9.61151645e-02 -8.30438510e-02 1.17250443e+00 -6.46385729e-01
3.58193547e-01 6.24732673e-01 -9.93087068e-02 -5.83206594e-01
-1.04030883e+00 -1.06606448e+00 -9.10368323e-01 -1.65772617e-01
-7.97844082e-02 -1.27482817e-01 -9.08974111e-01 1.73972821e+00
-6.66607916e-02 5.30858397e-01 3.89694124e-02 7.86347032e-01
7.02015162e-01 5.74558258e-01 3.59200805e-01 -4.16783631e-01
1.64315450e+00 -1.24858713e+00 -7.15519667e-01 -6.16568863e-01
5.80823720e-01 -3.34009528e-01 8.27371240e-01 3.32808346e-01
-1.33188212e+00 -7.13036239e-01 -9.70810592e-01 -4.02290225e-02
-1.14432432e-01 6.86080039e-01 -1.67914957e-01 4.68848437e-01
-1.29492593e+00 8.05477142e-01 -1.04844189e+00 -3.34098607e-01
2.65765816e-01 6.90130413e-01 2.92966887e-03 4.66197312e-01
-9.65094388e-01 7.78447747e-01 3.04705560e-01 3.75299245e-01
-8.06455612e-01 -5.34041703e-01 -7.42744625e-01 3.97629261e-01
2.82844119e-02 -7.10291386e-01 1.41263700e+00 -1.22451413e+00
-1.68162298e+00 1.07526851e+00 -5.98457873e-01 -7.73469448e-01
1.48394763e-01 1.22764697e-02 -7.60051966e-01 2.42238671e-01
9.55803096e-02 2.77017772e-01 1.14203203e+00 -5.49451172e-01
-9.62251961e-01 -4.82648939e-01 -1.85009524e-01 1.25389874e-01
-3.28797191e-01 2.08207279e-01 -3.14874381e-01 -6.87946320e-01
1.16434269e-01 -1.05709183e+00 -2.56733745e-01 -7.94467986e-01
-3.04335445e-01 -3.95478636e-01 1.95602894e-01 -9.68483806e-01
1.60966182e+00 -2.46992588e+00 3.28539371e-01 -1.28749937e-01
3.25263709e-01 3.56247842e-01 -2.73346752e-02 1.76170036e-01
-3.16936582e-01 -2.28870720e-01 -3.02777112e-01 -1.07479441e+00
-2.26900280e-01 -1.22826785e-01 1.37523124e-02 6.83670878e-01
6.67939112e-02 8.66084695e-01 -9.91575062e-01 -5.04227132e-02
1.12777643e-01 3.97388905e-01 -3.54162484e-01 4.96733874e-01
3.18802707e-02 5.69192350e-01 -7.14375600e-02 1.79960161e-01
7.21776485e-02 -4.52146471e-01 1.59515604e-01 1.38804451e-01
-1.75524905e-01 8.53677630e-01 -5.29604554e-01 1.89792871e+00
-7.66112268e-01 5.85946500e-01 -1.83721885e-01 -1.02364719e+00
8.87903810e-01 5.92731833e-01 3.81330669e-01 -7.67801762e-01
3.58965814e-01 2.87433237e-01 9.64846313e-02 -4.97123599e-01
3.41000348e-01 -5.45907795e-01 -2.59144694e-01 3.91776562e-01
3.00431877e-01 7.28984714e-01 1.67313904e-01 -1.68296441e-01
1.45754266e+00 -3.20739120e-01 2.73754328e-01 -1.87611908e-01
9.32090461e-01 -5.26993871e-01 7.97548473e-01 6.85218692e-01
-1.37938365e-01 6.49922729e-01 4.32242095e-01 -4.26793039e-01
-8.29962909e-01 -5.48182607e-01 6.37717247e-02 1.37207496e+00
-1.82345495e-01 -6.57564938e-01 -6.77158594e-01 -7.80604303e-01
-3.50224286e-01 7.73501694e-01 -9.06343699e-01 -5.75947404e-01
-3.72592241e-01 -6.22563660e-01 4.79302347e-01 5.89278936e-01
2.47991458e-01 -1.46651030e+00 -7.72519410e-01 4.96020943e-01
-1.52888343e-01 -1.09399748e+00 -6.32524788e-01 5.98394990e-01
-9.01993155e-01 -7.63218701e-01 -6.21720433e-01 -8.65128636e-01
5.71314991e-01 -1.67537898e-01 1.04340637e+00 -4.48523322e-03
6.36396324e-03 -9.61124673e-02 -3.97539288e-01 6.78739920e-02
-3.35161209e-01 6.01611674e-01 1.36804402e-01 7.40423322e-01
3.25609684e-01 -8.61997664e-01 -8.85120988e-01 3.02102119e-01
-6.08287573e-01 -2.52944678e-02 6.51867628e-01 8.53182495e-01
5.40205300e-01 -1.92034945e-01 7.19307601e-01 -9.48604584e-01
5.19816518e-01 -5.32124460e-01 -4.02375668e-01 1.29841730e-01
-4.57747310e-01 -7.95730054e-02 9.94250238e-01 -3.45459104e-01
-6.25124514e-01 1.61029071e-01 -5.39904058e-01 -5.42003691e-01
-1.47952825e-01 4.14894044e-01 9.36059579e-02 2.16863915e-01
5.19947648e-01 6.09163463e-01 -1.57199487e-01 -7.15972185e-01
-5.99429496e-02 9.38479960e-01 5.60677588e-01 1.91113383e-01
2.29968354e-01 5.07485075e-03 -2.82867491e-01 -7.99101412e-01
-1.30833757e+00 -8.61490369e-01 -7.69544661e-01 -2.49041438e-01
1.10378301e+00 -9.01186883e-01 -8.26561749e-01 3.81820202e-01
-8.06406975e-01 -4.86099213e-01 -2.03443632e-01 3.69953662e-01
-5.68693697e-01 -1.42303154e-01 -5.68765819e-01 -7.67823637e-01
-8.48596275e-01 -9.37369645e-01 1.01511431e+00 3.55238050e-01
-4.59552288e-01 -9.58497405e-01 7.24757537e-02 2.71322280e-01
3.23729813e-01 -1.92978591e-01 6.74574852e-01 -9.31026816e-01
1.34110630e-01 -2.64372319e-01 1.48689687e-01 5.91477156e-01
1.25949770e-01 -6.91914737e-01 -1.19341063e+00 -4.77646977e-01
5.33653200e-01 -5.68693727e-02 1.02759147e+00 4.40072268e-01
1.38031018e+00 -2.82676607e-01 -1.64992392e-01 6.90786302e-01
1.17458129e+00 3.44489604e-01 4.65372175e-01 2.56622046e-01
5.02992451e-01 2.37831593e-01 2.51894414e-01 2.36726135e-01
2.90810794e-01 6.21356308e-01 1.44519880e-01 1.35545805e-01
-1.45784363e-01 1.14138119e-01 5.03406525e-01 9.98182237e-01
2.10091010e-01 -5.26587069e-02 -6.04085028e-01 5.47951639e-01
-1.84748805e+00 -9.02322114e-01 1.56828165e-01 2.26219702e+00
6.73665941e-01 3.88196200e-01 4.05625880e-01 2.67702729e-01
7.30926514e-01 3.22810322e-01 -6.44539058e-01 -6.12029314e-01
2.96194851e-01 3.11999053e-01 2.51201093e-01 3.04255992e-01
-1.12168741e+00 7.36490488e-01 5.06967974e+00 6.06362641e-01
-1.35466111e+00 3.90983552e-01 6.10809803e-01 -3.83921772e-01
3.00775439e-01 -3.38742107e-01 -1.01516771e+00 9.94294822e-01
1.73236728e+00 1.21530801e-01 6.17929816e-01 5.68367779e-01
4.90307927e-01 7.38073438e-02 -1.01756978e+00 1.11276162e+00
2.64383107e-01 -1.12283564e+00 -4.44026172e-01 -2.26075560e-01
5.14152586e-01 2.31324434e-01 -1.18170165e-01 4.05221939e-01
-4.16765749e-01 -9.43358183e-01 8.01376522e-01 7.38040268e-01
6.24077857e-01 -8.03880632e-01 7.99505889e-01 4.92395520e-01
-1.19138730e+00 -4.31255668e-01 -1.14475116e-01 -1.50164276e-01
6.66144043e-02 6.59395158e-01 -8.37458611e-01 4.80578423e-01
5.29767573e-01 1.18986952e+00 -5.95675170e-01 1.35982001e+00
-5.45995235e-01 9.74242806e-01 4.08901609e-02 -1.62946314e-01
3.76905620e-01 1.82424635e-02 3.41179937e-01 1.37876642e+00
1.26267195e-01 1.39159217e-01 -3.20317782e-02 4.70166028e-01
-3.42344731e-01 -1.20375149e-01 -9.68095958e-02 1.05056919e-01
1.82118386e-01 1.41720951e+00 -7.90825546e-01 -3.01044703e-01
-2.03872085e-01 1.25510883e+00 5.45188844e-01 2.17343956e-01
-6.55948699e-01 -5.18071651e-01 4.04323548e-01 -1.78579882e-01
5.12987435e-01 1.08941384e-01 -3.74751031e-01 -1.14560294e+00
1.25207797e-01 -5.53865552e-01 5.91056406e-01 -4.33562040e-01
-1.25084293e+00 1.06386936e+00 -7.07415521e-01 -1.20416272e+00
-3.77511412e-01 -2.53570974e-01 -8.09834599e-01 1.07217014e+00
-1.34427047e+00 -7.03968763e-01 -1.46111861e-01 4.11408812e-01
8.31386089e-01 5.70200244e-03 8.32448006e-01 4.89266068e-01
-8.90004873e-01 6.69330418e-01 1.45286873e-01 -1.15564084e-02
3.99583638e-01 -1.26064873e+00 4.67242181e-01 8.23215663e-01
8.35309774e-02 5.38632214e-01 4.43692565e-01 -5.21841526e-01
-8.31996977e-01 -1.33807063e+00 1.53081703e+00 -1.49462312e-01
5.69520831e-01 -5.87243319e-01 -8.41971338e-01 7.46809721e-01
-1.13754809e-01 5.91681572e-03 7.64507473e-01 2.05498576e-01
2.73807555e-01 -3.26452821e-01 -7.34935462e-01 2.44530126e-01
8.99717689e-01 -7.31554508e-01 -6.38613820e-01 1.99129462e-01
4.87759888e-01 -3.87787730e-01 -7.68042147e-01 2.00827166e-01
5.00070214e-01 -7.83118129e-01 5.89683473e-01 -4.82595712e-01
3.18397999e-01 -1.04515642e-01 2.52921104e-01 -1.26817858e+00
-4.79228824e-01 -7.79811621e-01 1.83337592e-02 9.61298883e-01
6.04297996e-01 -4.87775683e-01 8.33380759e-01 2.68284768e-01
-5.10129154e-01 -1.06581032e+00 -1.10104394e+00 -5.10379076e-01
-5.42929947e-01 -4.25785691e-01 1.18547060e-01 3.97407591e-01
-7.95093272e-03 8.16193640e-01 -5.31328440e-01 7.69734681e-02
3.87327820e-01 1.94005996e-01 -3.91205437e-02 -1.19277275e+00
-2.11081892e-01 -3.79972845e-01 -4.67344552e-01 -8.90026808e-01
4.09590006e-01 -1.16071904e+00 2.59584606e-01 -1.31122291e+00
8.30561221e-02 -2.06914067e-01 -9.24020171e-01 5.31254411e-01
-3.96467805e-01 4.30328995e-01 4.17806841e-02 1.21863507e-01
-8.42489660e-01 4.85009223e-01 6.93281054e-01 7.70706013e-02
-5.56739271e-01 8.43365014e-01 -5.54665029e-01 6.57717645e-01
9.43954706e-01 -8.19857955e-01 -2.89531887e-01 -1.08559012e-01
1.15549583e-02 1.75374538e-01 1.62826598e-01 -1.33748996e+00
4.78425890e-01 5.86036384e-01 3.25588942e-01 -5.29973030e-01
5.43566287e-01 -5.64924538e-01 -6.67769602e-03 5.98052144e-01
-7.13535309e-01 7.49070719e-02 2.01225411e-02 4.47390079e-01
2.73176562e-02 -4.35877144e-01 8.34957004e-01 2.27805208e-02
-4.26408648e-01 1.88023806e-01 -5.18017471e-01 -3.01223099e-02
6.74025595e-01 5.74343875e-02 8.08470994e-02 -5.48063964e-02
-1.32345021e+00 1.66125119e-01 9.15223733e-02 3.72905016e-01
4.91181016e-01 -9.69742954e-01 -5.23437381e-01 5.06657660e-01
1.37947900e-02 -4.36379582e-01 3.81801903e-01 1.16389978e+00
7.48484656e-02 4.22071904e-01 -1.38683498e-01 -5.58404684e-01
-1.41035366e+00 3.18277538e-01 5.52851796e-01 -5.00576794e-01
-7.80603111e-01 9.87633884e-01 -8.25785249e-02 -6.97940448e-03
4.82727617e-01 -4.09102112e-01 -5.91243923e-01 3.88232529e-01
6.97435617e-01 2.81828195e-01 6.49950206e-01 -7.66311824e-01
-4.64079350e-01 3.72919858e-01 -7.58699030e-02 -7.72813857e-02
1.54974043e+00 -1.23658434e-01 -2.45573029e-01 9.69474614e-01
1.39841390e+00 -2.88404167e-01 -1.03792858e+00 -1.81521714e-01
2.88143039e-01 9.05990452e-02 7.52542028e-03 -9.72544312e-01
-1.03158748e+00 8.61155272e-01 6.98458791e-01 2.66075194e-01
1.41615355e+00 -4.04586233e-02 8.70468616e-01 9.63545777e-03
-3.81537415e-02 -7.79358864e-01 -2.29065269e-01 6.04438186e-01
5.19944429e-01 -8.22681010e-01 -4.84417409e-01 9.21930075e-02
-5.96583247e-01 1.02544510e+00 1.71156526e-01 -2.79153347e-01
5.73144317e-01 -1.75247073e-01 -5.55943958e-02 -2.11209103e-01
-1.00151861e+00 -1.74018636e-01 6.04254544e-01 -5.47151230e-02
2.87018359e-01 9.47086960e-02 -5.26750028e-01 1.07006907e+00
-3.90334845e-01 1.86353371e-01 1.48295149e-01 6.60688221e-01
-4.48704749e-01 -1.07220626e+00 1.03306748e-01 8.21509540e-01
-9.19095814e-01 -1.42814383e-01 -5.84458783e-02 -6.90063834e-02
3.29611218e-03 1.11057615e+00 2.16428876e-01 -6.66273296e-01
4.12457168e-01 5.92593491e-01 1.15641847e-01 -9.24027443e-01
-1.25893438e+00 -1.02841489e-01 1.65799960e-01 -5.77422738e-01
-3.55807245e-01 -7.49220431e-01 -1.15646315e+00 2.88691819e-01
-5.81792109e-02 5.13717532e-01 6.15411401e-01 1.16686571e+00
5.02973378e-01 1.04165578e+00 8.83810937e-01 -1.05167627e+00
-3.49258959e-01 -1.18993580e+00 -5.61937869e-01 3.37387443e-01
8.48071992e-01 -3.54574949e-01 -3.27834278e-01 9.42525174e-03] | [13.507197380065918, 3.5224976539611816] |
12966173-430a-4da6-a3a5-c6e4134e0456 | continual-learning-through-human-robot | 2305.16332 | null | https://arxiv.org/abs/2305.16332v1 | https://arxiv.org/pdf/2305.16332v1.pdf | Continual Learning through Human-Robot Interaction -- Human Perceptions of a Continual Learning Robot in Repeated Interactions | For long-term deployment in dynamic real-world environments, assistive robots must continue to learn and adapt to their environments. Researchers have developed various computational models for continual learning (CL) that can allow robots to continually learn from limited training data, and avoid forgetting previous knowledge. While these CL models can mitigate forgetting on static, systematically collected datasets, it is unclear how human users might perceive a robot that continually learns over multiple interactions with them. In this paper, we developed a system that integrates CL models for object recognition with a Fetch mobile manipulator robot and allows human participants to directly teach and test the robot over multiple sessions. We conducted an in-person study with 60 participants who interacted with our system in 300 sessions (5 sessions per participant). We conducted a between-participant study with three different CL models (3 experimental conditions) to understand human perceptions of continual learning robots over multiple sessions. Our results suggest that participants' perceptions of trust, competence, and usability of a continual learning robot significantly decrease over multiple sessions if the robot forgets previously learned objects. However, the perceived task load on participants for teaching and testing the robot remains the same over multiple sessions even if the robot forgets previously learned objects. Our results also indicate that state-of-the-art CL models might perform unreliably when applied to robots interacting with human participants. Further, continual learning robots are not perceived as very trustworthy or competent by human participants, regardless of the underlying continual learning model or the session number. | ['Kerstin Dautenhahn', 'Chrystopher L. Nehaniv', 'Patrick Holthaus', 'Zachary De Francesco', 'Ali Ayub'] | 2023-05-22 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [-4.19002503e-01 7.31855869e-01 1.33841947e-01 -3.96773219e-01
5.38482368e-02 -4.48158920e-01 -1.77789144e-02 1.20473891e-01
-7.02260554e-01 9.01912570e-01 -4.61454511e-01 -3.92235160e-01
-3.75705719e-01 -1.94918692e-01 -8.84925842e-01 -1.99742407e-01
-4.46972191e-01 6.78795159e-01 2.21114922e-02 -2.24039048e-01
1.17700197e-01 6.41493440e-01 -1.69646323e+00 1.47745252e-01
8.83952737e-01 9.11961570e-02 7.95209646e-01 5.09289801e-01
4.86635685e-01 9.72339153e-01 -6.57668412e-01 6.45661876e-02
1.40556306e-01 6.07269816e-02 -1.01983428e+00 2.38913625e-01
1.30021691e-01 -1.06915832e+00 -9.41638574e-02 6.10991061e-01
2.56681681e-01 4.45152909e-01 3.62426400e-01 -1.85355353e+00
-8.22121620e-01 5.36666632e-01 -1.14060761e-02 -4.39145684e-01
8.71848285e-01 5.54661810e-01 2.39560932e-01 -7.93241382e-01
3.25387508e-01 1.58233583e+00 9.85261321e-01 7.29973733e-01
-1.21429992e+00 -7.98275769e-01 1.54189706e-01 3.16385299e-01
-1.36909008e+00 -4.09334153e-01 3.46309781e-01 -4.77015585e-01
1.18869686e+00 -2.70692617e-01 8.90256822e-01 1.17647946e+00
8.44225228e-01 6.65347576e-01 7.87183642e-01 -5.56029916e-01
3.27121973e-01 9.47804391e-01 3.45007181e-01 6.74279392e-01
8.10289919e-01 1.13663062e-01 -7.28274047e-01 -6.24477044e-02
7.74214685e-01 1.89839080e-01 -1.87676370e-01 -6.96499109e-01
-9.74815369e-01 3.72016847e-01 3.13747913e-01 3.72037202e-01
-8.13581049e-01 1.03533924e-01 -1.04612976e-01 8.64186585e-01
-2.78168380e-01 5.37511230e-01 -7.36442268e-01 -4.13525909e-01
2.72501975e-01 3.01681627e-02 1.07344973e+00 1.36435890e+00
9.17080760e-01 -3.91390800e-01 2.75681764e-01 6.63050950e-01
4.20737863e-01 4.41737592e-01 6.16258800e-01 -1.39333224e+00
-2.67767847e-01 4.11280721e-01 7.44260073e-01 -9.84758317e-01
-6.41611516e-01 1.14654861e-01 1.70050152e-02 6.23937428e-01
1.71269756e-02 -1.89959660e-01 -4.85854268e-01 1.56707835e+00
2.71206737e-01 -5.24466872e-01 1.96311623e-01 7.20901370e-01
2.84097284e-01 2.54257560e-01 4.87508327e-01 -3.25561792e-01
1.09425294e+00 -9.02585447e-01 -9.17358100e-01 -8.04732382e-01
8.88084412e-01 -3.28430593e-01 1.94562721e+00 7.17369914e-01
-5.50515950e-01 -1.01962268e+00 -1.37453794e+00 2.39517033e-01
3.25193331e-02 -4.37228680e-02 8.31773162e-01 8.93441498e-01
-1.20582926e+00 7.18525708e-01 -1.21528065e+00 -1.03094089e+00
-7.57065788e-02 7.09673345e-01 -5.26364565e-01 -4.00912672e-01
-6.76634371e-01 1.47629178e+00 2.87555277e-01 1.76112592e-01
-1.31704533e+00 -2.24719524e-01 -8.88622880e-01 -2.83900291e-01
1.30559847e-01 -4.79271561e-01 1.80044186e+00 -9.55423474e-01
-1.55376875e+00 1.95022181e-01 2.01721534e-01 -6.40677139e-02
2.31996700e-01 -9.15166855e-01 -7.59516507e-02 -5.36302365e-02
3.77444208e-01 8.94887805e-01 5.09322464e-01 -1.69881141e+00
-3.20820689e-01 -4.62489069e-01 2.78983802e-01 5.35621703e-01
-7.21955776e-01 -6.64836645e-01 3.50221932e-01 1.07640631e-01
9.79130790e-02 -1.29196763e+00 1.24509104e-01 1.67866156e-01
5.48668146e-01 -4.06051993e-01 1.17288280e+00 -3.35454106e-01
3.45035076e-01 -2.40243506e+00 -1.64941609e-01 -6.59589767e-02
-1.27176061e-01 -7.96092227e-02 -3.35599244e-01 4.42832679e-01
2.76243687e-01 -2.43454993e-01 4.36492532e-01 -3.55886698e-01
-1.75208002e-01 3.91076207e-01 2.73977429e-01 4.13022518e-01
-2.45819286e-01 6.04633510e-01 -9.93072212e-01 -1.29606634e-01
1.43309638e-01 3.92679542e-01 -2.88608432e-01 4.31307793e-01
1.94476038e-01 3.81316692e-01 -6.88483939e-02 4.34982091e-01
4.88310575e-01 -6.89045787e-02 3.78613114e-01 2.94382721e-01
1.69868916e-01 -1.89730257e-01 -7.78353274e-01 1.66441071e+00
-4.29274857e-01 4.99064088e-01 2.12536439e-01 -5.15547469e-02
7.40855396e-01 5.41842937e-01 2.94361502e-01 -5.61493874e-01
1.04634553e-01 2.21730173e-01 6.85569122e-02 -9.94069397e-01
6.73955441e-01 -2.96564400e-01 4.20814566e-02 7.53273249e-01
9.19060037e-02 -2.61483610e-01 -5.05229175e-01 1.65384009e-01
1.39230263e+00 9.47721750e-02 6.81883097e-02 -1.00960359e-01
-6.45126030e-02 3.41493160e-01 2.55583197e-01 9.44278181e-01
-8.87240827e-01 -2.06063613e-01 -2.45993555e-01 -5.23966372e-01
-6.10870063e-01 -9.37621891e-01 4.81174499e-01 1.31185031e+00
4.61816788e-01 2.03533322e-02 -5.57674766e-01 -4.53511447e-01
1.31406426e-01 1.50873649e+00 -2.34700620e-01 -7.39683926e-01
-1.06346034e-01 -3.93414982e-02 3.73008966e-01 2.34849826e-01
5.24958730e-01 -1.43341148e+00 -1.28601599e+00 2.02405661e-01
-3.31106842e-01 -6.89031065e-01 -2.02104941e-01 3.09541166e-01
-1.08536959e+00 -9.65534091e-01 -1.17846452e-01 -1.22667229e+00
9.25520301e-01 1.01942468e+00 5.46211123e-01 3.30605179e-01
-6.41774759e-02 1.55706155e+00 -4.93912965e-01 -6.08495414e-01
-4.97658134e-01 -5.57950512e-02 8.33709896e-01 -7.65983701e-01
5.63862205e-01 -5.68888664e-01 -3.74315768e-01 6.51500225e-01
-5.05554259e-01 -2.36353189e-01 9.83067572e-01 6.84940696e-01
-2.56911188e-01 4.50108111e-01 9.08449829e-01 -5.36505282e-01
1.15292823e+00 -5.87042630e-01 1.66239217e-01 4.12329972e-01
-9.72544074e-01 -3.97241980e-01 -4.24118489e-02 -1.05807590e+00
-1.26227295e+00 1.84935838e-01 6.04796350e-01 -2.74572879e-01
-2.81242549e-01 5.02760053e-01 -9.44519192e-02 -2.13285625e-01
1.18986285e+00 -3.11360836e-01 5.30404270e-01 -8.86717252e-03
7.10545704e-02 1.05245829e+00 3.98058951e-01 -8.17649841e-01
5.95604241e-01 -4.85496707e-02 -8.71055424e-01 -8.27640176e-01
-2.77701288e-01 -9.39494893e-02 -7.35048473e-01 -7.56501198e-01
1.49080604e-01 -1.07930696e+00 -1.40097094e+00 6.45888984e-01
-1.22453487e+00 -1.21438384e+00 -2.02653557e-01 8.84199321e-01
-6.44777000e-01 5.56635819e-02 -7.07177937e-01 -1.30184090e+00
-5.20977527e-02 -9.72707987e-01 3.91845942e-01 3.22517276e-01
-8.80446434e-01 -5.22868991e-01 -2.84211606e-01 5.23752391e-01
4.60268199e-01 -3.99682045e-01 7.88987219e-01 -3.63147527e-01
-2.18150020e-01 -2.58039236e-01 3.73857379e-01 3.21735471e-01
6.82875574e-01 -7.93287873e-01 -7.28193283e-01 -9.90708828e-01
5.21538079e-01 -1.02735615e+00 -3.41612756e-01 -1.45728171e-01
1.82804540e-01 -3.78140509e-01 -4.60984319e-01 -5.65217078e-01
7.91953743e-01 6.61523819e-01 6.23067915e-01 4.65638280e-01
3.35612804e-01 8.81125629e-01 1.28126562e+00 5.35502024e-02
5.48320711e-01 -3.97629924e-02 1.06029741e-01 6.23627305e-01
3.71553123e-01 -3.30778092e-01 9.17875051e-01 5.02370715e-01
1.42077565e-01 1.09371163e-01 -9.24108803e-01 4.77911741e-01
-1.96102595e+00 -4.92635876e-01 1.70126885e-01 2.15234017e+00
5.96738517e-01 2.41031736e-01 -2.34198064e-01 -2.40068883e-02
6.25455797e-01 -8.90573621e-01 -9.57836270e-01 -1.67174399e-01
6.44807756e-01 -4.23828363e-01 2.73740888e-01 5.78687966e-01
-3.63583893e-01 9.67426181e-01 6.60771036e+00 -1.77066624e-01
-7.16409743e-01 1.00871161e-01 3.97287130e-01 -3.94348018e-02
1.57342270e-01 4.23422344e-02 -2.94706017e-01 -3.22337478e-01
9.76072550e-01 -7.78779611e-02 5.17462313e-01 1.29901385e+00
2.35423788e-01 -7.46455848e-01 -1.48369920e+00 6.66561365e-01
9.82275382e-02 -4.89166491e-02 -3.19695890e-01 -2.49504983e-01
2.01513872e-01 -4.46785241e-01 5.20892218e-02 9.62403953e-01
3.69357884e-01 -9.27842915e-01 9.00664389e-01 4.18323010e-01
3.53597820e-01 -5.95562279e-01 7.23000407e-01 1.03256154e+00
-4.13357288e-01 -6.65882885e-01 -3.02804559e-01 -6.82831526e-01
-3.11788172e-02 -1.85854942e-01 -1.68877721e+00 -1.21870674e-01
1.25362253e+00 9.07120109e-02 -7.47185349e-01 6.21899545e-01
-1.49913505e-01 1.87753215e-01 -5.31440601e-02 -2.99906969e-01
-4.02075201e-01 1.86446026e-01 1.54299423e-01 3.26025695e-01
3.60776275e-01 3.97074252e-01 1.21970572e-01 5.86099446e-01
4.94734645e-01 -2.96423167e-01 -9.01673973e-01 8.10687840e-02
8.41892004e-01 8.21628511e-01 -7.79925764e-01 6.80784062e-02
-3.43657508e-02 1.36241734e+00 4.69120443e-01 4.30199474e-01
-4.69459414e-01 -3.37164372e-01 3.46982688e-01 1.64893284e-01
-3.41926694e-01 -8.42639685e-01 -2.37746149e-01 -4.00598377e-01
2.32036650e-01 -9.39788997e-01 -3.60851251e-02 -1.32528722e+00
-1.33854997e+00 4.23955679e-01 1.59013152e-01 -1.11786604e+00
-1.12041004e-01 -4.01748151e-01 2.51614694e-02 2.96142310e-01
-7.50319898e-01 -1.04688227e+00 -7.57350266e-01 5.06356537e-01
7.93577492e-01 -1.54262438e-01 1.06016588e+00 -1.06921963e-01
-1.07876278e-01 4.78342801e-01 -3.09081852e-01 -6.46843255e-01
1.21140540e+00 -4.60714340e-01 -3.50878119e-01 -1.09815381e-01
-4.99083877e-01 1.50702643e+00 8.74442995e-01 -1.37900400e+00
-1.79915369e+00 -6.32666171e-01 3.11877936e-01 -7.37473369e-01
1.15698002e-01 -2.51201153e-01 -1.21132815e+00 1.30942786e+00
1.90994143e-01 -6.15712404e-01 6.76215291e-01 3.66055548e-01
9.18382257e-02 -1.92856371e-01 -1.61942172e+00 6.16241276e-01
1.17400432e+00 -5.05358934e-01 -8.63349617e-01 2.47088775e-01
1.12988734e+00 -8.67700875e-02 -4.83815521e-01 3.18159282e-01
1.02940083e+00 -6.79506779e-01 4.36778873e-01 -7.87300318e-02
-1.65801570e-01 -3.20988089e-01 1.47238806e-01 -1.41800880e+00
-6.23774171e-01 -5.87239444e-01 3.44678521e-01 9.37500894e-01
2.14633241e-01 -7.96746731e-01 5.36920667e-01 1.39739788e+00
-2.17891291e-01 -8.15710649e-02 -5.65910935e-01 -1.00831747e+00
-2.00734586e-01 -4.60750550e-01 2.22972140e-01 7.43572891e-01
9.15692449e-01 5.74018359e-01 -1.28209859e-01 4.67674285e-01
3.88700575e-01 -9.46022034e-01 1.03076029e+00 -1.49544752e+00
-2.95536611e-02 3.39155972e-01 -1.62656039e-01 -7.83412158e-01
1.65691346e-01 -3.67673010e-01 7.87682593e-01 -1.47138917e+00
3.06770831e-01 -6.55445755e-01 1.11972794e-01 9.11542058e-01
5.59989847e-02 -5.37006974e-01 6.26011118e-02 6.21400476e-01
-7.31619060e-01 8.47786427e-01 9.77243781e-01 -1.07447900e-01
-7.50805140e-01 -6.73987642e-02 -7.72007406e-01 6.62564397e-01
8.35247815e-01 -5.46180069e-01 -1.13409519e+00 -4.62046415e-01
3.46437246e-02 2.95890328e-02 5.21367848e-01 -1.54123747e+00
8.98610711e-01 -2.50145078e-01 4.54917848e-01 -1.49891436e-01
4.29274619e-01 -1.38489830e+00 6.97145522e-01 8.82630408e-01
-4.00886834e-02 -6.53896201e-03 6.97046757e-01 5.88779449e-01
4.83951598e-01 -4.13324833e-01 3.09758812e-01 -1.63927004e-01
-5.85637093e-01 -5.45338571e-01 -1.28911412e+00 -9.84735072e-01
1.48796117e+00 -3.93868208e-01 -1.82168230e-01 -8.35569501e-01
-8.44439268e-01 5.58764398e-01 6.81770325e-01 8.83368313e-01
9.81842160e-01 -1.14099121e+00 -7.71104731e-03 7.34298080e-02
1.26515508e-01 1.03043236e-01 2.76361018e-01 4.08925027e-01
-3.59487921e-01 2.30984628e-01 -6.32704198e-01 -5.63631177e-01
-1.17044294e+00 4.24820334e-01 1.21177621e-01 6.52163923e-01
-4.36836421e-01 5.11826575e-01 -1.39564469e-01 -9.31984127e-01
7.39202499e-01 -8.33622962e-02 1.26039550e-01 -2.56010920e-01
2.85703927e-01 5.76715708e-01 -2.58929059e-02 -1.34230733e-01
-1.98750049e-01 -2.74822153e-02 -3.98758262e-01 -4.16057497e-01
1.31930101e+00 -5.81752360e-01 -5.00994362e-02 1.24552333e+00
4.72288251e-01 -5.91932058e-01 -1.49871492e+00 2.97454745e-02
-2.39840541e-02 -4.08510566e-01 -5.76926351e-01 -1.02782607e+00
-2.58168131e-01 2.94848293e-01 9.28016663e-01 -3.06268871e-01
6.19111121e-01 -6.53008670e-02 3.91824186e-01 1.56040037e+00
1.47232735e+00 -1.44555616e+00 1.10040510e+00 5.52971184e-01
1.30003870e+00 -1.22097838e+00 7.45262355e-02 -2.88583249e-01
-8.75745535e-01 9.53605592e-01 1.41008174e+00 4.03678834e-01
7.27937043e-01 -9.69734937e-02 5.13734296e-02 -3.46145272e-01
-6.24059796e-01 6.61748886e-01 -5.22534609e-01 9.91560459e-01
2.81698823e-01 1.74413040e-01 -4.68267351e-02 9.74043489e-01
-4.10935640e-01 5.21687746e-01 1.09134436e+00 1.63792527e+00
-9.14154530e-01 -6.25450611e-01 -6.65343106e-01 8.38699862e-02
6.43140733e-01 7.36084402e-01 -4.02433157e-01 7.74993300e-01
1.47786751e-01 1.63394535e+00 -1.15669891e-01 -8.31025064e-01
7.47665882e-01 4.30191964e-01 6.04125023e-01 -1.01463497e+00
-3.94940436e-01 -5.43999851e-01 -8.04715324e-03 -4.01632279e-01
-1.57021731e-01 -1.00587869e+00 -1.71203578e+00 -4.48938370e-01
-4.28218126e-01 2.23146707e-01 6.01833582e-01 6.27977669e-01
3.59032243e-01 2.52415001e-01 4.51023132e-01 -9.07791495e-01
-4.88618284e-01 -1.56381202e+00 -4.77134734e-01 7.67109841e-02
1.59281343e-01 -8.53326201e-01 -3.48470241e-01 7.09762350e-02] | [4.639204978942871, 0.9508499503135681] |
9f2c1bbd-cbd0-47ab-9e10-a4d4d978701a | recurrent-homography-estimation-using | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Cao_Recurrent_Homography_Estimation_Using_Homography-Guided_Image_Warping_and_Focus_Transformer_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Cao_Recurrent_Homography_Estimation_Using_Homography-Guided_Image_Warping_and_Focus_Transformer_CVPR_2023_paper.pdf | Recurrent Homography Estimation Using Homography-Guided Image Warping and Focus Transformer | We propose the Recurrent homography estimation framework using Homography-guided image Warping and Focus transformer (FocusFormer), named RHWF. Both being appropriately absorbed into the recurrent framework, the homography-guided image warping progressively enhances the feature consistency and the attention-focusing mechanism in FocusFormer aggregates the intra-inter correspondence in a global->nonlocal->local manner. Thanks to the above strategies, RHWF ranks top in accuracy on a variety of datasets, including the challenging cross-resolution and cross-modal ones. Meanwhile, benefiting from the recurrent framework, RHWF achieves parameter efficiency despite the transformer architecture. Compared to previous state-of-the-art approaches LocalTrans and IHN, RHWF reduces the mean average corner error (MACE) by about 70% and 38.1% on the MSCOCO dataset, while saving the parameter costs by 86.5% and 24.6%. Similar to the previous works, RHWF can also be arranged in 1-scale for efficiency and 2-scale for accuracy, with the 1-scale RHWF already outperforming most of the previous methods. Source code is available at https://github.com/imdumpl78/RHWF. | ['Hui-Liang Shen', 'Junwei Li', 'Zehua Sheng', 'Beinan Yu', 'Lun Luo', 'Runmin Zhang', 'Si-Yuan Cao'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['homography-estimation'] | ['computer-vision'] | [ 5.46332262e-02 1.58332177e-02 -3.64532620e-02 6.31073341e-02
-1.06719089e+00 -2.97069460e-01 5.13012469e-01 -3.68529320e-01
-5.54666854e-02 4.47935939e-01 5.58831990e-01 1.60499483e-01
-2.79185951e-01 -7.90875554e-01 -6.49172187e-01 -8.52592409e-01
2.44437814e-01 -6.06389642e-02 3.52400839e-01 -2.56528705e-01
3.42932194e-01 2.63980657e-01 -1.50101054e+00 2.45814770e-01
9.01685953e-01 1.19562566e+00 3.72061342e-01 1.89344764e-01
2.65823245e-01 4.59613621e-01 6.39806166e-02 -4.29642797e-01
4.03350174e-01 -2.42949590e-01 -5.38037002e-01 1.12379849e-01
7.47201145e-01 -3.76459539e-01 -4.61595923e-01 8.67503285e-01
6.89659774e-01 1.13939509e-01 1.85077086e-01 -9.11525905e-01
-5.93492627e-01 2.67162859e-01 -9.03898478e-01 -2.41435654e-02
4.20122087e-01 9.70019624e-02 1.05775011e+00 -1.33501923e+00
8.66219759e-01 1.06278110e+00 6.72166944e-01 -6.30878061e-02
-1.31760335e+00 -7.35154569e-01 1.07890584e-01 5.33998847e-01
-1.49854672e+00 -3.82429957e-01 1.00953197e+00 -2.69119501e-01
8.76389265e-01 2.46164933e-01 7.17927575e-01 9.72359180e-01
3.23301464e-01 6.96646452e-01 1.23771036e+00 -2.57049441e-01
-2.78659016e-01 -4.55118895e-01 -7.37473369e-02 6.69911742e-01
-8.31645504e-02 3.14909697e-01 -8.39968085e-01 9.66405421e-02
1.08995295e+00 6.43978044e-02 -7.00126410e-01 -3.96732956e-01
-1.64301980e+00 5.78018725e-01 7.64760435e-01 3.70253175e-01
-4.61480260e-01 3.18529680e-02 3.15342665e-01 2.03488320e-01
4.93980557e-01 4.35857207e-01 -2.18780875e-01 -3.77650037e-02
-9.16805208e-01 -4.18584533e-02 1.62380531e-01 1.11009812e+00
9.72222269e-01 1.08244978e-02 -2.33682618e-01 8.32486749e-01
1.29918367e-01 6.93970203e-01 4.11897719e-01 -1.02165854e+00
5.12738466e-01 6.17208958e-01 -6.98960498e-02 -1.39594030e+00
-4.10448879e-01 -8.62085938e-01 -1.09235013e+00 -2.03324184e-01
1.62573338e-01 3.36435825e-01 -7.86402643e-01 1.49511158e+00
3.38422865e-01 2.36357108e-01 -4.57077354e-01 8.99825335e-01
7.21383274e-01 6.93789124e-01 -3.90194982e-01 -4.03252363e-01
1.36637938e+00 -1.26374888e+00 -8.90382946e-01 -1.15436614e-02
3.17288965e-01 -1.29359841e+00 1.09332561e+00 4.77585971e-01
-1.15948784e+00 -5.82093179e-01 -8.44279945e-01 -4.21318352e-01
-2.27367252e-01 3.85936797e-01 3.86952281e-01 1.07557930e-01
-1.19041681e+00 2.75548249e-01 -7.39307165e-01 -3.92079830e-01
1.09961726e-01 1.90587014e-01 -5.35372913e-01 -3.25022459e-01
-9.17279065e-01 5.87242007e-01 -1.25047751e-02 1.94547385e-01
-4.36254025e-01 -1.17192042e+00 -8.16620350e-01 1.30412236e-01
5.22426605e-01 -7.87599683e-01 8.43743980e-01 -3.08034301e-01
-1.45702755e+00 5.74949503e-01 -4.05270427e-01 -1.87338263e-01
7.08149135e-01 -3.65761310e-01 -3.40721369e-01 1.86851203e-01
8.75900015e-02 6.69402719e-01 8.70238960e-01 -1.28611171e+00
-5.19987345e-01 -2.73907840e-01 -1.22098446e-01 2.21386567e-01
-3.36131692e-01 -3.22831362e-01 -8.45252693e-01 -9.91272449e-01
5.08034050e-01 -1.03466380e+00 -4.09255810e-02 -4.63445038e-02
-3.73121589e-01 1.36130035e-01 1.08805358e+00 -8.83920074e-01
1.63530326e+00 -2.23916578e+00 3.56819034e-01 1.55589387e-01
3.34532231e-01 3.93718407e-02 -7.97681287e-02 6.61255717e-01
-4.48984392e-02 -4.05141711e-01 -2.49465257e-01 -4.52919930e-01
-5.73144667e-02 -1.06971331e-01 -3.91340852e-01 5.96997380e-01
-1.99675888e-01 1.09722900e+00 -7.19389558e-01 -3.26655924e-01
6.15563691e-01 8.52555096e-01 -6.57389998e-01 -8.49698484e-02
3.67222488e-01 5.16231120e-01 -2.66233921e-01 7.99159706e-01
9.68049169e-01 -4.19865340e-01 8.37858766e-02 -9.07274961e-01
-3.54240447e-01 1.83084793e-02 -1.21530211e+00 1.95405746e+00
-7.21736372e-01 5.68219841e-01 2.85283197e-03 -6.39814615e-01
9.38596904e-01 1.11300200e-01 6.33416355e-01 -1.14943147e+00
9.78460610e-02 4.69698817e-01 -3.85764867e-01 -1.64428473e-01
6.47940218e-01 2.23001599e-01 -5.92783540e-02 5.89038245e-02
-1.51120415e-02 1.67273171e-02 9.83412564e-02 1.87487617e-01
9.22594786e-01 1.22008815e-01 4.34682965e-01 -4.76929843e-01
8.42703462e-01 -3.44907492e-01 5.88466883e-01 3.01121563e-01
4.91241440e-02 1.03200245e+00 7.43157119e-02 -4.39912856e-01
-9.56038535e-01 -1.09309673e+00 -3.31935525e-01 7.27271676e-01
4.40767407e-01 -7.16085434e-01 -4.14928883e-01 -2.49757230e-01
-3.10503021e-02 3.82059306e-01 -6.10080063e-01 7.61142597e-02
-7.21172333e-01 -4.99154061e-01 -4.99149188e-02 5.40069103e-01
1.08624303e+00 -6.89247370e-01 -4.87932056e-01 1.86738119e-01
-5.30004203e-01 -1.28952789e+00 -7.97098815e-01 -3.13811332e-01
-9.01315331e-01 -9.33395743e-01 -1.02263951e+00 -5.51339269e-01
6.17007434e-01 5.91100693e-01 9.82345641e-01 1.46776810e-01
-2.70485371e-01 2.21732005e-01 -4.88116801e-01 3.34365368e-01
3.14797759e-01 2.78679818e-01 -2.71126658e-01 3.35383534e-01
-2.54372686e-01 -8.79021645e-01 -1.14669764e+00 6.15711689e-01
-7.50223279e-01 4.33329105e-01 6.96827471e-01 1.02658689e+00
7.69996107e-01 -3.10009092e-01 1.42140120e-01 -5.59301674e-01
1.64393276e-01 -1.95025131e-01 -5.20426393e-01 1.70655757e-01
-7.71326244e-01 -2.53846258e-01 5.61382771e-01 -2.44338408e-01
-1.05903494e+00 1.18959509e-02 -2.49217376e-02 -5.50853848e-01
2.78204083e-01 2.46450707e-01 -8.37272331e-02 -3.32314193e-01
2.62932628e-01 3.59697193e-01 -1.00995272e-01 -6.45947695e-01
1.71718806e-01 1.70654327e-01 6.60290182e-01 -1.45117909e-01
8.29405725e-01 7.77874947e-01 1.25701413e-01 -6.50860727e-01
-6.80084229e-01 -5.23016870e-01 -6.08986080e-01 -2.95964956e-01
7.80913293e-01 -9.74401236e-01 -7.52698839e-01 5.44959724e-01
-9.66370940e-01 -1.29086196e-01 -1.69457301e-01 4.72433209e-01
-4.52186674e-01 4.38715577e-01 -5.85399270e-01 -3.00898403e-01
-3.53607923e-01 -1.40640604e+00 1.44013727e+00 -1.31916450e-02
5.62033802e-02 -8.23220670e-01 -2.36502036e-01 3.99488479e-01
7.02652574e-01 2.38763332e-01 6.01986289e-01 2.23252028e-01
-9.36636508e-01 -9.48270857e-02 -3.62216294e-01 8.74281451e-02
-6.63318578e-03 -2.13850841e-01 -7.51494527e-01 -4.19115514e-01
-1.54734492e-01 2.54220348e-02 7.84263670e-01 4.47053552e-01
9.47321475e-01 -1.06023870e-01 -2.28245273e-01 9.01968837e-01
1.59412563e+00 9.51954573e-02 9.88138795e-01 6.35695517e-01
8.11213315e-01 3.74460280e-01 6.56735241e-01 5.17974257e-01
4.68372226e-01 1.22965169e+00 3.32292348e-01 -3.01779389e-01
-4.51469570e-01 -3.78884524e-01 1.75220802e-01 1.05018711e+00
-2.68143296e-01 -2.45519681e-03 -8.49318564e-01 5.61657190e-01
-2.02109361e+00 -9.26539898e-01 -1.75148681e-01 2.11255789e+00
4.88925099e-01 -2.95113534e-01 -8.10644701e-02 1.55175194e-01
5.93625546e-01 4.88652915e-01 -3.87282073e-01 9.11615640e-02
-4.12876844e-01 1.31152971e-02 5.10776460e-01 8.46482098e-01
-9.46731150e-01 8.68435502e-01 5.39709806e+00 1.08994913e+00
-1.20971727e+00 1.52604401e-01 5.61929941e-01 -1.07270084e-01
-3.24154049e-01 1.56318665e-01 -7.37546146e-01 3.94171000e-01
3.78962159e-01 -1.94542661e-01 5.30615807e-01 5.16070247e-01
1.32913545e-01 -2.36958936e-01 -5.37354112e-01 1.29387236e+00
2.41069853e-01 -1.52485156e+00 4.81963195e-02 2.53043652e-01
8.11480641e-01 -4.33085561e-02 2.10860342e-01 2.28398871e-02
-2.16110647e-01 -7.07222521e-01 7.09687173e-01 5.79439759e-01
9.36594307e-01 -7.35867321e-01 7.48064876e-01 6.75958171e-02
-1.71625936e+00 -1.30094467e-02 -3.70000750e-02 1.94034770e-01
6.01327240e-01 9.47171271e-01 4.82335910e-02 1.16932213e+00
9.79068577e-01 1.05479026e+00 -6.87427521e-01 8.67887616e-01
-2.38527104e-01 2.01231524e-01 -4.11538482e-01 5.73825419e-01
2.23678112e-01 -3.95022780e-01 5.94230771e-01 9.79859591e-01
6.77577913e-01 2.08720133e-01 1.56102866e-01 4.66767877e-01
2.29324847e-01 7.30185136e-02 -3.13278109e-01 5.20868361e-01
4.53891933e-01 1.43924820e+00 -5.85998416e-01 -1.93145588e-01
-3.22908312e-01 1.06231058e+00 -5.36983050e-02 4.26586628e-01
-9.67899144e-01 -5.00211596e-01 3.95255953e-01 3.18763644e-01
5.48181534e-01 -2.17650414e-01 -3.33523780e-01 -1.24854791e+00
3.46144319e-01 -8.36782932e-01 2.81503618e-01 -9.11829352e-01
-9.80845869e-01 9.66356695e-01 9.04856026e-02 -1.66288221e+00
-8.30410197e-02 -3.92252088e-01 -2.56484568e-01 6.75873041e-01
-1.56707406e+00 -1.42576849e+00 -6.91004813e-01 6.28513455e-01
5.23536742e-01 1.65629238e-01 4.01658922e-01 5.51252186e-01
-5.11336148e-01 7.60419905e-01 9.64986011e-02 -2.00176731e-01
8.95722985e-01 -6.46394134e-01 2.31571645e-01 8.49621892e-01
-7.09302425e-02 7.25821733e-01 5.45392096e-01 -4.88115609e-01
-1.56074333e+00 -8.09478700e-01 1.04151273e+00 -2.14337468e-01
6.57207072e-01 -3.88924748e-01 -8.76703978e-01 5.34765661e-01
4.56301183e-01 2.60733932e-01 1.50868043e-01 -1.45538643e-01
-5.86841702e-01 -3.54195476e-01 -9.67564046e-01 7.06540108e-01
1.50537503e+00 -5.22107899e-01 -1.45513639e-01 8.35043564e-02
5.59820235e-01 -6.56201243e-01 -1.21071291e+00 7.40149498e-01
7.23593414e-01 -1.44033587e+00 1.11542881e+00 4.46371138e-01
5.58497548e-01 -5.33173978e-01 -3.01615655e-01 -1.16587937e+00
-7.38889396e-01 -8.21441293e-01 -1.57431632e-01 1.13709307e+00
1.34596258e-01 -9.58761096e-01 4.19826478e-01 8.88259932e-02
-2.76335269e-01 -1.02948487e+00 -1.16581249e+00 -8.31943452e-01
-3.68226677e-01 -2.30727911e-01 5.56012869e-01 8.71113181e-01
-2.69656152e-01 2.11346954e-01 -6.24104857e-01 1.28562480e-01
7.72659242e-01 3.90859663e-01 7.47967482e-01 -7.53708184e-01
-2.21971676e-01 -3.85767698e-01 -4.03486252e-01 -1.27575004e+00
-2.59332538e-01 -6.47943199e-01 -2.94870406e-01 -1.40264916e+00
1.11479521e-01 -1.82548910e-02 -1.97540317e-02 4.31307256e-01
8.60799383e-03 6.83228612e-01 5.47101617e-01 4.06294346e-01
-3.81530166e-01 8.66182446e-01 1.51896107e+00 1.57625034e-01
3.02569568e-02 -5.02928317e-01 -4.61781323e-01 5.05691051e-01
5.22689939e-01 -1.97221879e-02 -3.12183887e-01 -5.52226424e-01
2.90197097e-02 1.03865489e-02 4.17354673e-01 -1.01829922e+00
3.57640803e-01 1.64848030e-01 2.63402611e-01 -9.39880490e-01
5.17460406e-01 -7.62969017e-01 6.03652060e-01 2.81743407e-01
1.05555795e-01 4.49205697e-01 1.48633301e-01 4.43640441e-01
-5.54631472e-01 5.37812173e-01 8.27988744e-01 1.52189389e-01
-7.46033192e-01 3.20477515e-01 7.52905905e-02 -1.43964052e-01
9.27764297e-01 -3.66079867e-01 -5.90016127e-01 -4.52910811e-01
-4.75168318e-01 1.33874103e-01 7.07112193e-01 3.40342581e-01
6.04193449e-01 -1.43793869e+00 -4.83338416e-01 5.30254126e-01
1.88534245e-01 -1.42425075e-01 7.93796718e-01 1.59699953e+00
-3.97418529e-01 5.99378288e-01 -2.22797081e-01 -7.68932343e-01
-1.14655864e+00 3.89541268e-01 1.93614792e-02 -5.83899140e-01
-9.84984279e-01 6.33891582e-01 5.49130797e-01 -1.45202607e-01
-4.89650620e-03 -1.98103622e-01 3.45413461e-02 4.90903631e-02
4.85154718e-01 6.49153769e-01 2.26700246e-01 -8.29918087e-01
-3.26387942e-01 1.28077567e+00 -1.62297443e-01 -6.83398247e-02
1.33169627e+00 -3.29120815e-01 -6.74475282e-02 3.97571176e-03
1.39239597e+00 2.85298407e-01 -1.27401161e+00 -2.27483571e-01
-2.91098148e-01 -7.47426808e-01 2.81582952e-01 -5.97156286e-01
-1.47439587e+00 7.23468781e-01 6.00753427e-01 -6.04167655e-02
1.63661385e+00 -1.07448518e-01 9.63891089e-01 -2.13440821e-01
6.47413075e-01 -7.95663118e-01 1.40857682e-01 5.13978481e-01
1.40600801e+00 -1.01120901e+00 3.38018835e-01 -7.84192860e-01
-5.77587724e-01 8.92429411e-01 4.42434967e-01 -3.72389883e-01
6.41887069e-01 2.39288554e-01 -1.53329909e-01 -2.09095910e-01
-6.24638498e-01 -3.63255218e-02 5.91063678e-01 1.43870592e-01
4.33153331e-01 -2.32598960e-01 -5.17832518e-01 2.51529455e-01
-2.33564645e-01 -6.03307709e-02 1.50757059e-02 6.64281189e-01
-1.64109599e-02 -1.11711895e+00 -4.21021253e-01 1.62343085e-01
-4.12576236e-02 -1.83973104e-01 3.71532924e-02 9.08398628e-01
-2.97587812e-02 6.80860341e-01 -1.69041660e-02 -7.25501895e-01
6.55631840e-01 -3.79400939e-01 2.52027184e-01 4.03379649e-03
-5.98131001e-01 6.22267365e-01 -2.60409936e-02 -1.22727978e+00
-4.78360862e-01 -5.65217495e-01 -9.21002030e-01 -4.41070557e-01
-2.60655254e-01 -2.18178079e-01 2.66144276e-01 5.56406260e-01
8.52993786e-01 4.19424593e-01 6.90367699e-01 -1.11780453e+00
-1.34139732e-01 -8.66443932e-01 -3.68554890e-01 3.88741851e-01
2.45336473e-01 -9.48687315e-01 -6.10585928e-01 -1.24987930e-01] | [10.58215618133545, -2.03245210647583] |
6ef8acbb-b2cd-4145-94cc-f46ab8793b69 | a-survey-on-medical-document-summarization | 2212.01669 | null | https://arxiv.org/abs/2212.01669v1 | https://arxiv.org/pdf/2212.01669v1.pdf | A Survey on Medical Document Summarization | The internet has had a dramatic effect on the healthcare industry, allowing documents to be saved, shared, and managed digitally. This has made it easier to locate and share important data, improving patient care and providing more opportunities for medical studies. As there is so much data accessible to doctors and patients alike, summarizing it has become increasingly necessary - this has been supported through the introduction of deep learning and transformer-based networks, which have boosted the sector significantly in recent years. This paper gives a comprehensive survey of the current techniques and trends in medical summarization | ['Adam Jatowt', 'Sriparna Saha', 'Anubhav Jangra', 'Raghav Jain'] | 2022-12-03 | null | null | null | null | ['document-summarization'] | ['natural-language-processing'] | [-1.36443466e-01 -2.33663302e-02 -5.70125103e-01 -1.51697546e-01
-6.34785354e-01 -2.69736767e-01 3.27832460e-01 1.12726760e+00
-5.23116112e-01 8.50899696e-01 8.52702200e-01 -2.12570414e-01
-2.47460544e-01 -9.55921650e-01 -7.29684457e-02 -5.84215283e-01
-2.00884610e-01 6.80285156e-01 -1.14431672e-01 8.65383297e-02
2.80207217e-01 6.65576756e-01 -9.86832201e-01 4.67853844e-01
7.02500880e-01 8.51112962e-01 3.06379884e-01 3.40850741e-01
-3.69672239e-01 8.22120190e-01 -7.91329980e-01 -6.44285619e-01
-1.40625656e-01 -5.01652837e-01 -1.04320538e+00 -3.13860863e-01
-5.18416092e-02 -4.42785978e-01 -3.17069709e-01 8.49519730e-01
5.66593051e-01 -2.14987937e-02 2.00057209e-01 -4.95908946e-01
-6.69524312e-01 3.34923357e-01 -5.75253367e-01 3.40646595e-01
3.06702226e-01 -3.70934397e-01 8.01575780e-01 -9.50258523e-02
9.73346412e-01 9.99705195e-01 6.53226852e-01 3.05451274e-01
-8.64353657e-01 -4.98498797e-01 -4.57034767e-01 3.23922187e-01
-9.46028709e-01 -1.90049574e-01 4.78129029e-01 -6.17504939e-02
1.06530213e+00 3.28238249e-01 1.17077124e+00 7.24939644e-01
7.93849826e-01 6.14717722e-01 5.34839153e-01 -5.03088117e-01
-1.17606938e-01 1.64750367e-01 -1.26332939e-01 5.82581222e-01
7.27181613e-01 -3.80930930e-01 -3.17611158e-01 -1.22751057e-01
6.08237684e-01 6.42945290e-01 -5.60082018e-01 -1.22434057e-01
-1.32631397e+00 7.81915128e-01 4.30760682e-01 9.56962109e-01
-6.76392078e-01 3.55704129e-02 9.03090894e-01 1.77994758e-01
4.34982151e-01 5.76685011e-01 -2.08585650e-01 -4.93459910e-01
-9.64802265e-01 1.96730837e-01 7.43368387e-01 3.10682267e-01
5.09570651e-02 -2.17943043e-01 2.28291541e-01 7.86356688e-01
6.64086267e-02 4.52731788e-01 6.25828862e-01 -1.08540654e+00
2.87896693e-01 8.09494853e-01 3.38159199e-03 -1.41661620e+00
-5.58724344e-01 -4.83309656e-01 -1.24082363e+00 -1.21216670e-01
1.97664306e-01 1.65619388e-01 -6.85496628e-01 1.21067834e+00
-1.90721247e-02 -3.50004226e-01 -4.14508209e-02 5.80181718e-01
6.39174759e-01 6.17617249e-01 -8.71368591e-03 -1.51186854e-01
1.39615726e+00 -2.60719866e-01 -1.32624686e+00 4.33672667e-01
6.49275899e-01 -8.06201220e-01 6.07513487e-01 6.44927144e-01
-1.05210233e+00 -1.40271068e-01 -8.90797734e-01 -1.59847185e-01
-6.22398376e-01 -7.09962130e-01 7.71708071e-01 5.48438549e-01
-1.12396002e+00 8.91083956e-01 -9.76549804e-01 -7.18637884e-01
1.02701914e+00 2.17283741e-01 -6.66580498e-01 -3.46603930e-01
-1.19534731e+00 1.08876371e+00 2.95377612e-01 -7.31562153e-02
-2.38646299e-01 -6.24792516e-01 -5.44513881e-01 2.01744244e-01
-8.67843255e-02 -1.03963268e+00 1.13088000e+00 -3.84914458e-01
-9.44764018e-01 9.43662107e-01 7.89849609e-02 -6.59230113e-01
2.13036746e-01 -1.34802133e-01 -5.94551444e-01 3.42872471e-01
-3.99948843e-02 4.45069522e-01 1.18008696e-01 -5.57288647e-01
-1.02706170e+00 -6.03433192e-01 -2.67426610e-01 6.94714785e-02
-6.84150755e-01 1.35676488e-01 -3.10630947e-01 -5.37655294e-01
-1.57469526e-01 -3.40605080e-01 -2.28278175e-01 -3.45938765e-02
-8.66041332e-02 1.47686489e-02 8.55692923e-01 -1.08138001e+00
1.12845600e+00 -1.96236002e+00 1.43075408e-03 -1.69159146e-03
6.45232201e-01 5.38797677e-01 4.51377511e-01 9.17787194e-01
1.01310588e-01 3.65739256e-01 5.00059361e-03 -2.04176039e-01
-5.45288682e-01 3.62321824e-01 -1.79599792e-01 3.72179270e-01
-2.77937680e-01 9.62065637e-01 -9.91602957e-01 -5.88474095e-01
4.48354304e-01 8.84768307e-01 -1.36913091e-01 -1.20664395e-01
1.88409016e-01 5.31285048e-01 -5.06651103e-01 5.61977983e-01
1.77709877e-01 -6.60733163e-01 3.15184057e-01 1.98428005e-01
-1.52492493e-01 4.06213999e-01 -4.16079223e-01 1.54454362e+00
-5.87096810e-01 9.56184983e-01 -2.08185185e-02 -1.02031898e+00
7.11779118e-01 7.11704195e-01 9.33363378e-01 -1.10198247e+00
2.85292864e-01 8.92198458e-02 -1.12741470e-01 -5.52791178e-01
3.15737933e-01 -3.30116928e-01 1.47698745e-01 2.42461860e-01
-3.86225283e-01 4.33643945e-02 6.79475218e-02 4.08835649e-01
1.09618044e+00 -4.04474825e-01 5.64335942e-01 -4.41711061e-02
2.53521204e-01 2.51482636e-01 4.02619123e-01 2.20181346e-01
-7.66073242e-02 3.47618461e-01 2.03328833e-01 -5.99654973e-01
-1.04208112e+00 -9.66531932e-01 -3.87838721e-01 1.07171439e-01
-1.31166935e-01 -3.49982798e-01 -4.99413788e-01 -2.19868600e-01
1.69077143e-01 4.18805510e-01 -3.83669972e-01 3.95107791e-02
-5.10426641e-01 -4.15580422e-01 2.57972240e-01 6.30822778e-01
5.01683652e-01 -1.03210723e+00 -1.11128271e+00 4.23281848e-01
-3.46404195e-01 -7.92010844e-01 -1.10859476e-01 1.19470082e-01
-1.32993710e+00 -1.02650797e+00 -1.17167222e+00 -7.46071458e-01
3.83284986e-01 1.90132886e-01 9.86138761e-01 -1.52303353e-02
-9.75986421e-01 4.35953379e-01 -2.79563934e-01 -7.40147412e-01
-3.52130443e-01 1.10292956e-02 -1.78778335e-01 -4.37432349e-01
1.46108210e-01 -5.16120434e-01 -6.39488399e-01 -4.97177213e-01
-8.25776339e-01 -1.82851002e-01 4.99495417e-01 6.85808182e-01
5.54709077e-01 1.74273565e-01 8.95602524e-01 -1.16792226e+00
7.90031552e-01 -3.83212060e-01 -3.30539465e-01 -7.62773631e-03
-1.08767128e+00 -1.78464428e-01 6.00136399e-01 4.19914097e-01
-7.89689779e-01 -5.25008559e-01 -1.83467254e-01 -1.20427318e-01
-7.61311688e-03 8.10672402e-01 1.28134280e-01 1.52885139e-01
3.02107334e-01 -1.79202572e-01 3.45868528e-01 -7.33063400e-01
-1.44851031e-02 1.08973777e+00 5.68047941e-01 3.22805077e-01
1.66980818e-01 5.04695535e-01 6.26719967e-02 -1.13554966e+00
-4.84154820e-01 -6.10846877e-01 -3.05639714e-01 -1.03642456e-02
1.10502481e+00 -5.63316226e-01 -9.28306520e-01 2.70755827e-01
-1.09158802e+00 2.78940350e-01 -5.27002394e-01 4.61981237e-01
1.25589073e-01 4.46046591e-01 -6.63416922e-01 -3.44799250e-01
-6.49408638e-01 -8.87526035e-01 4.11352724e-01 2.65364736e-01
-6.07812464e-01 -1.42822099e+00 2.00068712e-01 3.17937672e-01
7.98027515e-01 5.61518371e-01 1.27557445e+00 -6.13536954e-01
-5.21405876e-01 -9.10432100e-01 -2.89485514e-01 3.57911885e-01
7.23735392e-01 -2.53978640e-01 -6.07051015e-01 -1.49756864e-01
1.07301004e-01 1.02778457e-01 7.28485227e-01 7.24664867e-01
1.30152249e+00 -2.81837553e-01 -1.00826323e+00 4.39232677e-01
1.21173370e+00 4.65739667e-01 6.67307138e-01 3.92256975e-01
2.54976630e-01 3.93551946e-01 2.29865164e-01 2.47281536e-01
5.34807324e-01 2.87175208e-01 4.40658420e-01 -3.68885815e-01
-2.86032975e-01 1.57639291e-02 -6.06779397e-01 1.11734736e+00
-9.52921137e-02 -3.04153442e-01 -1.13339329e+00 6.74848735e-01
-1.63169980e+00 -1.00073230e+00 -1.03125684e-01 2.19030619e+00
7.38790035e-01 -1.17575389e-03 -1.16717346e-01 1.91852733e-01
5.04121363e-01 1.88938174e-02 -2.86521494e-01 -5.93000293e-01
1.53888166e-01 5.09195864e-01 4.58780080e-01 2.47766107e-01
-8.09505045e-01 3.16988945e-01 7.53776598e+00 3.14577311e-01
-1.44065595e+00 1.20272852e-01 8.15119207e-01 -4.77346331e-02
-1.75665364e-01 -6.69667065e-01 -3.13259482e-01 5.19549072e-01
1.04977024e+00 -6.98316276e-01 5.10563403e-02 6.91767573e-01
5.19723177e-01 -5.53454697e-01 -8.59027445e-01 1.15893185e+00
-1.07236721e-01 -2.05575776e+00 1.57777533e-01 4.96694863e-01
4.96377259e-01 1.40735954e-01 2.57962584e-01 -4.44402933e-01
-3.58106196e-02 -1.04984653e+00 -2.30638832e-01 6.13456368e-01
8.67120922e-01 -9.59367275e-01 1.20679915e+00 2.20632777e-01
-6.92318678e-01 -1.56574532e-01 -1.19973920e-01 -1.84993092e-02
2.63455898e-01 8.61726046e-01 -9.70317185e-01 7.31609702e-01
9.32340980e-01 7.96748877e-01 -9.04087257e-03 1.51980197e+00
1.91780806e-01 4.65272367e-01 -9.37190503e-02 -7.60201067e-02
7.66140074e-02 -1.09710857e-01 2.65017152e-01 1.02055264e+00
2.04134166e-01 -8.47000331e-02 -1.48262799e-01 2.96957612e-01
-3.44286352e-01 2.98653573e-01 -6.65532291e-01 -4.39768493e-01
3.52429152e-01 1.13450503e+00 -1.05630755e+00 -3.11148584e-01
-4.52859432e-01 8.47818375e-01 -1.81026563e-01 -1.72143206e-01
-4.26473916e-01 -8.30219448e-01 6.05630636e-01 4.38992798e-01
6.23363145e-02 -1.81047738e-01 -1.03457592e-01 -7.13486850e-01
-3.14089894e-01 -1.03233826e+00 5.65872908e-01 -2.44256616e-01
-8.87574971e-01 5.28949976e-01 -1.61726147e-01 -8.73228550e-01
-3.63381982e-01 -4.23790663e-01 -1.05606735e-01 8.37749898e-01
-1.47486079e+00 -6.61054492e-01 1.13929987e-01 3.60207289e-01
2.10682586e-01 -3.30313332e-02 1.25283062e+00 6.34470165e-01
-1.47463232e-01 1.83470428e-01 5.86766183e-01 1.27526280e-02
8.00791919e-01 -9.10487354e-01 1.86007038e-01 2.18217596e-01
2.14343548e-01 7.41424084e-01 4.84001219e-01 -4.66628551e-01
-1.33131135e+00 -7.31979668e-01 1.18036699e+00 -4.32277173e-01
3.20867419e-01 7.59357512e-02 -7.59683728e-01 6.05744958e-01
3.09051722e-01 -6.17177367e-01 8.36908937e-01 1.16140261e-01
2.57197693e-02 -5.40908515e-01 -1.27488267e+00 3.16309005e-01
5.37066221e-01 -3.77774715e-01 -6.12218618e-01 5.39125919e-01
5.52459776e-01 -2.42802441e-01 -8.63247395e-01 1.13685377e-01
4.48744267e-01 -8.45589101e-01 7.67400920e-01 -4.92215455e-01
4.41410542e-01 2.51565576e-01 2.83856392e-01 -1.26833773e+00
-2.56501317e-01 -4.06301618e-01 4.37385477e-02 8.06740940e-01
3.14224333e-01 -1.02282393e+00 8.31025243e-01 6.45637453e-01
-1.95365027e-01 -9.50807154e-01 -8.78584325e-01 -4.77068871e-01
-7.96447098e-02 -1.03185691e-01 5.64921379e-01 1.00716674e+00
5.09570956e-01 1.45025507e-01 -6.67110831e-02 -5.20677745e-01
2.65146494e-01 1.53494537e-01 2.68269598e-01 -1.46882808e+00
-4.49970961e-02 -6.61604822e-01 -7.32390404e-01 -7.30311871e-01
-5.76787353e-01 -9.76205766e-01 -6.00213110e-01 -2.43412638e+00
3.55454713e-01 -1.84143946e-01 -3.71762484e-01 6.41550779e-01
2.64457762e-01 3.94266695e-01 -4.78715375e-02 3.14834118e-01
-4.21601653e-01 -9.75321233e-02 1.39463222e+00 -9.27193537e-02
-9.73006412e-02 -7.95334876e-02 -9.79457438e-01 4.50715870e-01
8.44429195e-01 -1.49095044e-01 -5.44690266e-02 -3.67861241e-01
4.61430281e-01 7.48709589e-02 -6.69385642e-02 -1.22687078e+00
1.20805666e-01 2.25727305e-01 7.71291673e-01 -6.77106202e-01
2.71569461e-01 -1.03597343e+00 3.26135278e-01 9.19909477e-01
-2.21219555e-01 1.71612158e-01 3.35995287e-01 3.17937821e-01
-4.91173059e-01 6.55462071e-02 6.63963258e-01 -1.86070770e-01
-3.48480552e-01 3.23094577e-02 -3.53442460e-01 -2.92463362e-01
1.05313313e+00 -2.30423123e-01 -1.91396758e-01 -6.28987193e-01
-7.28612185e-01 2.36990601e-01 4.04334873e-01 2.66174793e-01
5.70414603e-01 -7.90116191e-01 -3.85674179e-01 4.24487069e-02
-2.14036107e-01 -6.37023672e-02 2.16121033e-01 9.88510966e-01
-9.04271603e-01 1.00656962e+00 -3.25962245e-01 -3.60431314e-01
-1.49953806e+00 6.66823745e-01 1.18159182e-01 -4.36275184e-01
-1.37370229e+00 7.28139937e-01 -2.81554848e-01 2.23621264e-01
4.51074183e-01 -3.38690907e-01 -2.72378594e-01 5.35603128e-02
9.59651709e-01 6.24350071e-01 1.19494669e-01 -1.31936789e-01
-3.53457570e-01 1.97085425e-01 -5.60320675e-01 1.83170959e-01
1.74831045e+00 -1.38613850e-01 -4.78469193e-01 5.79987228e-01
1.10665894e+00 6.29318431e-02 -2.59961545e-01 2.36437827e-01
-2.56161653e-02 -4.53817338e-01 2.90017635e-01 -1.10291553e+00
-1.09851682e+00 1.01547825e+00 5.52468300e-01 6.17669284e-01
1.11549306e+00 5.40597364e-02 1.39697611e+00 2.27700919e-01
5.49357295e-01 -9.41869020e-01 -1.04148062e-02 1.17069431e-01
6.26968324e-01 -8.58225882e-01 2.22942054e-01 7.87854940e-02
-3.53917748e-01 1.11885750e+00 -4.48452771e-01 3.04046363e-01
6.73665822e-01 4.26131666e-01 2.11928681e-01 -3.57893914e-01
-3.47215444e-01 4.27620769e-01 -2.45989859e-01 6.77053988e-01
7.78444409e-01 -5.14292419e-02 -4.54603553e-01 8.79260004e-02
-1.16360188e-02 5.17251968e-01 2.76070267e-01 9.48315799e-01
-5.88250637e-01 -1.39255548e+00 -3.88893366e-01 9.02593672e-01
-1.01116228e+00 -4.55684625e-02 -1.36305004e-01 8.34700882e-01
-8.99196863e-02 8.13720047e-01 1.06919803e-01 1.38929114e-01
1.19095214e-01 9.83503610e-02 5.59062541e-01 -5.33389926e-01
-4.27654564e-01 -1.05305925e-01 7.40132928e-02 -4.75852609e-01
-1.84426516e-01 -5.47883749e-01 -1.56450641e+00 -6.37217820e-01
-1.43744305e-01 2.66871423e-01 9.66732740e-01 6.11138105e-01
8.57628703e-01 8.47712040e-01 2.12850273e-01 -4.08416599e-01
-3.80497068e-01 -5.74850678e-01 -6.35178149e-01 -3.43408398e-02
3.43390733e-01 -2.05651715e-01 1.67913899e-01 2.37517189e-02] | [8.032444953918457, 7.175463676452637] |
5d085f47-1a06-47f0-b305-a5f3167d7af6 | a-fully-unsupervised-instance-segmentation | 2306.14875 | null | https://arxiv.org/abs/2306.14875v1 | https://arxiv.org/pdf/2306.14875v1.pdf | A Fully Unsupervised Instance Segmentation Technique for White Blood Cell Images | White blood cells, also known as leukocytes are group of heterogeneously nucleated cells which act as salient immune system cells. These are originated in the bone marrow and are found in blood, plasma, and lymph tissues. Leukocytes kill the bacteria, virus and other kind of pathogens which invade human body through phagocytosis that in turn results immunity. Detection of a white blood cell count can reveal camouflaged infections and warn doctors about chronic medical conditions such as autoimmune diseases, immune deficiencies, and blood disorders. Segmentation plays an important role in identification of white blood cells (WBC) from microscopic image analysis. The goal of segmentation in a microscopic image is to divide the image into different distinct regions. In our paper, we tried to propose a novel instance segmentation method for segmenting the WBCs containing both the nucleus and the cytoplasm, from bone marrow images. | ['Amartya Bhattacharya', 'Shrijeet Biswas'] | 2023-06-26 | null | null | null | null | ['blood-cell-count', 'instance-segmentation'] | ['computer-vision', 'computer-vision'] | [ 2.08495036e-01 -3.32922071e-01 -9.71547812e-02 2.42177412e-01
2.33107343e-01 -6.16598904e-01 2.40465060e-01 5.63970625e-01
-6.10091388e-01 9.18146729e-01 9.87638086e-02 -1.02177776e-01
4.65253621e-01 -9.44592834e-01 4.32023197e-01 -8.85400474e-01
5.72856545e-01 9.20222104e-01 4.12462533e-01 5.22115648e-01
5.30708075e-01 1.05812800e+00 -9.13945377e-01 -3.97282057e-02
5.12886405e-01 4.58465040e-01 -3.01369786e-01 7.99993634e-01
-9.31560755e-01 7.25247681e-01 -9.33861673e-01 3.29772905e-02
-4.63978723e-02 -6.96251333e-01 -4.18822229e-01 4.70677614e-01
-3.72622907e-01 -2.29099050e-01 2.69863963e-01 1.27032065e+00
3.28227341e-01 3.77484038e-02 1.02565777e+00 -9.53590214e-01
-2.79675007e-01 1.23435244e-01 -1.21952569e+00 5.55030763e-01
1.81680098e-01 2.03237757e-01 -1.33472800e-01 -5.63591063e-01
5.24283469e-01 1.48351681e+00 2.74234116e-01 8.34712088e-01
-1.08698845e+00 -6.24545515e-01 -5.55476189e-01 -1.29090905e-01
-1.16676521e+00 -1.25967741e-01 2.16292143e-01 -7.82887816e-01
5.07834136e-01 4.91452336e-01 5.74401677e-01 2.85197765e-01
7.06639469e-01 7.73312151e-01 1.12223899e+00 -3.13984841e-01
4.72215921e-01 6.41858205e-02 5.17998099e-01 6.19157314e-01
8.41432631e-01 -3.03101271e-01 -6.65052906e-02 -3.44179124e-01
9.60781634e-01 4.67800766e-01 -2.05185354e-01 2.29918547e-02
-9.25781846e-01 7.32253075e-01 -2.45464951e-01 6.81620657e-01
-3.05324107e-01 -1.48974016e-01 2.82496840e-01 -1.94890380e-01
-1.64890196e-02 4.14858274e-02 -1.42612785e-01 8.80733877e-02
-4.96312350e-01 -3.24128181e-01 5.50311506e-01 -1.94290891e-01
5.52527070e-01 -3.74768078e-02 2.86621274e-03 6.69670820e-01
1.30587682e-01 8.07202697e-01 7.83206344e-01 -5.85682511e-01
-4.09077495e-01 1.15578628e+00 4.60250750e-02 -9.44250524e-01
-4.95039374e-01 -7.15334341e-02 -9.49356437e-01 3.97709876e-01
1.09996271e+00 8.43390599e-02 -8.81039202e-01 1.16441023e+00
7.93499053e-01 2.35960916e-01 -6.01887330e-02 1.02483869e+00
8.54724884e-01 4.69657898e-01 2.30217293e-01 -2.40000367e-01
1.80726194e+00 -6.10294104e-01 -7.47416854e-01 2.96278089e-01
4.17267591e-01 -9.65917706e-01 7.35038280e-01 3.29901129e-01
-7.30105758e-01 -4.20511991e-01 -7.05627799e-01 2.35862732e-01
-4.85731661e-01 -1.62526861e-01 6.34296000e-01 8.93704355e-01
-5.33564925e-01 2.27806151e-01 -9.97421980e-01 -7.50949800e-01
3.37080985e-01 2.81851381e-01 -3.53749037e-01 1.79937512e-01
-4.19970721e-01 5.83280802e-01 5.34471236e-02 1.04559466e-01
-3.39673460e-01 -2.73975462e-01 -4.49057251e-01 -2.11569056e-01
2.15795055e-01 -8.87656808e-01 5.87493658e-01 -6.61444008e-01
-1.17926836e+00 1.40688288e+00 -4.15098816e-01 -1.78098962e-01
3.17312956e-01 1.24989621e-01 -1.57947645e-01 5.30273974e-01
-8.13045427e-02 3.19126755e-01 5.09189665e-01 -9.29295897e-01
-7.60557652e-01 -7.55037010e-01 -6.68335140e-01 -4.03169602e-01
6.53835118e-01 1.47559971e-01 -2.22224966e-01 -6.76479936e-01
2.75086969e-01 -6.09582245e-01 -2.73404539e-01 -2.45409161e-01
-5.43170154e-01 -1.75996631e-01 1.04211617e+00 -6.24658108e-01
6.37177646e-01 -1.91139364e+00 -2.47655958e-01 2.69144267e-01
4.60321248e-01 4.24504131e-01 2.99871624e-01 -1.75335586e-01
2.34217152e-01 2.70263314e-01 -2.32316121e-01 2.59979129e-01
-1.75067201e-01 1.35427713e-01 -1.71875641e-01 8.11563015e-01
-9.05873254e-02 6.68209970e-01 -7.09392905e-01 -7.96135306e-01
4.12092686e-01 6.03817523e-01 3.70073132e-02 1.58551067e-01
-6.58666268e-02 1.17144632e+00 -7.41693735e-01 8.25799525e-01
7.44117975e-01 -4.50766325e-01 -4.51724567e-02 5.40717803e-02
2.50790894e-01 -2.79443443e-01 -1.00864077e+00 6.54677153e-01
3.20477962e-01 5.98652661e-01 3.44196916e-01 -6.67296350e-01
7.45972931e-01 1.59004986e-01 4.43376631e-01 -4.69660014e-01
7.00398624e-01 6.40192404e-02 -7.50313550e-02 -5.21352291e-01
9.73252133e-02 -5.82309783e-01 2.28786096e-01 1.07761490e+00
-6.59457743e-01 7.90788382e-02 4.63696599e-01 9.21298936e-02
7.95773506e-01 -3.97794843e-01 5.24523377e-01 -2.20874473e-01
1.03620541e+00 -4.11793701e-02 6.44603074e-01 1.75151050e-01
-5.14395773e-01 4.23593253e-01 5.97554982e-01 -6.39362574e-01
-6.10918343e-01 -1.13660097e+00 -1.79882199e-01 5.29242575e-01
3.75003010e-01 6.85557008e-01 -8.30201268e-01 -3.43519360e-01
1.53333381e-01 1.97857395e-01 -5.00417829e-01 3.21561426e-01
-6.51824474e-01 -1.03837597e+00 4.46314424e-01 6.97172433e-02
7.39344716e-01 -9.29440141e-01 -9.58973110e-01 1.19063474e-01
1.77643858e-02 -8.72597277e-01 -1.62972137e-01 -1.17289566e-01
-1.05122638e+00 -1.59566021e+00 -8.86178315e-01 -6.16867363e-01
9.52815413e-01 1.71416283e-01 6.88833177e-01 8.13941240e-01
-1.03949475e+00 1.49534762e-01 6.26136689e-03 -3.15566957e-01
-6.66692436e-01 -5.00493467e-01 3.01735289e-02 6.39838278e-02
8.92488599e-01 -1.98987320e-01 -7.45994925e-01 2.60209829e-01
-9.78650749e-01 -9.29097086e-02 2.95153320e-01 4.48903471e-01
7.06446111e-01 3.05413902e-01 2.69594043e-01 -1.25950956e+00
3.33325207e-01 -1.11041114e-01 -5.63387871e-01 1.67295784e-01
-8.52266252e-02 -8.09801221e-02 3.85128647e-01 -1.64122522e-01
-8.49633694e-01 -3.86594146e-01 2.14578629e-01 2.24187717e-01
-7.83345222e-01 -2.92225778e-01 -1.52333394e-01 1.22551862e-02
1.31968647e-01 3.09364438e-01 1.66162133e-01 -4.70585048e-01
-3.50172073e-01 6.95866942e-01 7.98934340e-01 -2.76335299e-01
3.61810595e-01 1.19294441e+00 3.39047551e-01 -1.15646756e+00
-3.15512329e-01 -7.63779223e-01 -3.42181802e-01 -1.09624773e-01
1.11579704e+00 -3.81962806e-01 -1.08438408e+00 6.51887953e-01
-1.05717659e+00 -1.67398930e-01 -7.26111000e-03 6.65217936e-01
-4.45726402e-02 9.07501817e-01 -1.24961400e+00 -9.21993017e-01
-5.11245012e-01 -8.21981311e-01 6.58854783e-01 9.89658594e-01
-4.52388644e-01 -1.35529602e+00 4.02387828e-01 5.39527059e-01
3.22487265e-01 4.63962734e-01 1.37029350e+00 -7.35272110e-01
-4.18078035e-01 -7.01494887e-02 -2.09224015e-01 -5.44884577e-02
3.70314926e-01 4.12108034e-01 -7.53963053e-01 3.00224330e-02
6.28250182e-01 2.07585573e-01 9.32188153e-01 7.23139644e-01
6.01260543e-01 1.47323027e-01 -7.39189804e-01 5.28107345e-01
1.43772876e+00 8.55492353e-01 7.19193757e-01 9.86394361e-02
3.41023862e-01 9.37828958e-01 5.44952929e-01 4.02952760e-01
-5.13913214e-01 -2.04706371e-01 -2.98762210e-02 -5.28383791e-01
-2.53695369e-01 4.40783113e-01 -2.75048196e-01 7.01117873e-01
1.68749124e-01 -3.79736781e-01 -9.02398407e-01 4.25969422e-01
-1.27289736e+00 -8.38968515e-01 -3.68173838e-01 1.98187292e+00
6.87036514e-01 8.59137475e-02 2.10495740e-02 1.27654254e-01
1.27953553e+00 -6.12849295e-01 -5.99629223e-01 -1.28142029e-01
-3.60953838e-01 7.58042097e-01 4.34254855e-01 7.80014575e-01
-7.12418079e-01 5.60994327e-01 6.46837950e+00 6.88704312e-01
-1.45435369e+00 -1.46137550e-01 1.01251483e+00 1.09808877e-01
-1.16871201e-01 -1.16279781e-01 -7.28025675e-01 7.36381352e-01
-4.03713249e-02 1.25666767e-01 -1.52660638e-01 1.36123896e-01
2.45357409e-01 -1.30650210e+00 -8.10965002e-01 9.51085746e-01
-2.15278760e-01 -1.12718856e+00 1.27508985e-02 4.36623162e-03
4.53410029e-01 -6.79063678e-01 -6.04754426e-02 -3.88952717e-02
1.72063783e-01 -1.12704039e+00 -1.28026083e-01 4.14471418e-01
3.34917814e-01 -6.37927830e-01 1.08786130e+00 2.83724457e-01
-9.69595432e-01 5.15780985e-01 -4.90034193e-01 -2.69275397e-01
2.17732444e-01 8.14191043e-01 -8.31469417e-01 -3.96060586e-01
2.05258876e-01 -4.22214009e-02 -4.84731108e-01 1.14407003e+00
-3.15127783e-02 3.99780989e-01 -3.82226020e-01 -2.58874804e-01
7.73148984e-03 -7.54923463e-01 3.59159350e-01 1.03992414e+00
-9.89845544e-02 4.98763144e-01 1.79193933e-02 8.87998700e-01
1.89056456e-01 3.04881036e-01 -2.61112094e-01 -1.55249193e-01
2.11498901e-01 1.30233312e+00 -1.92868257e+00 -6.33831263e-01
-7.50180567e-04 5.84613264e-01 -3.17014307e-01 4.55429524e-01
-6.23023689e-01 -4.32347685e-01 3.96405905e-01 2.79675364e-01
-2.29976371e-01 1.01140335e-01 -5.66822708e-01 -8.24745774e-01
-7.80568659e-01 -4.56469417e-01 5.82890868e-01 -3.28330994e-01
-9.55062449e-01 1.00616902e-01 -4.14003491e-01 -7.14329004e-01
1.12989761e-01 -6.41312897e-01 -7.51975477e-01 8.10969174e-01
-1.09420919e+00 -5.03504753e-01 -3.55227083e-01 5.41348100e-01
2.74744421e-01 -1.48978874e-01 6.84205890e-01 -3.40583175e-02
-7.63204515e-01 4.57408046e-03 -2.26553045e-02 3.90757591e-01
6.81773782e-01 -1.08805776e+00 -2.29233325e-01 7.34991133e-01
-3.20206165e-01 8.00251961e-01 8.48549664e-01 -9.43871081e-01
-7.76599824e-01 -4.80904788e-01 6.76973045e-01 1.44470587e-01
2.39838779e-01 9.95814726e-02 -6.97279751e-01 5.90682141e-02
1.54706761e-01 -1.95360050e-01 1.25060582e+00 -6.62796795e-01
1.15651287e-01 2.51524180e-01 -1.59791315e+00 6.37555242e-01
-1.62856519e-01 -2.05683202e-01 -5.45721531e-01 2.15256423e-01
1.55089498e-02 -2.17957810e-01 -4.13689494e-01 -1.25892714e-01
5.78876913e-01 -1.29887295e+00 8.05408537e-01 -4.65827823e-01
-1.45325616e-01 -5.20103455e-01 3.15714717e-01 -6.08935714e-01
1.36598393e-01 -3.96708101e-01 3.29415023e-01 8.19110274e-01
-3.03100705e-01 -1.04736257e+00 1.37925327e+00 7.17318773e-01
4.04954910e-01 -2.55751818e-01 -7.89025247e-01 -4.29925799e-01
-1.97496742e-01 3.80108416e-01 5.63403487e-01 9.94082808e-01
3.79643179e-02 3.90829563e-01 2.77358383e-01 -4.24771875e-01
1.00854158e+00 4.05501932e-01 9.00499403e-01 -1.28353262e+00
-2.41758928e-01 -7.01392889e-01 -3.23895097e-01 -6.05865836e-01
-1.82343245e-01 -3.04503709e-01 -5.78353763e-01 -1.53398132e+00
5.03703594e-01 -4.09006029e-01 -1.46327736e-02 2.21166432e-01
-2.65184879e-01 4.03618127e-01 -1.39812632e-02 2.66330272e-01
-1.91156745e-01 -3.70419979e-01 1.40693712e+00 -5.82854711e-02
-1.16756715e-01 -9.92309004e-02 -3.95261973e-01 1.03242934e+00
1.05862916e+00 -5.33388555e-01 -5.56290969e-02 5.06979041e-02
-1.49449989e-01 2.52819359e-01 -7.23362193e-02 -5.65798759e-01
6.90325052e-02 -5.18473029e-01 6.52421057e-01 -9.97943342e-01
-4.69667092e-03 -7.22956002e-01 5.17501652e-01 1.41673458e+00
6.77536950e-02 -2.75561094e-01 -1.19845212e-01 3.50991100e-01
-1.78034052e-01 -3.89398277e-01 1.34556699e+00 -5.88668644e-01
-4.50948715e-01 -2.23415241e-01 -1.12485588e+00 -1.36139333e-01
1.40410149e+00 -7.55401969e-01 -7.08521485e-01 -1.28523260e-01
-5.92543483e-01 1.49673656e-01 8.69722426e-01 -4.65455443e-01
5.54846048e-01 -9.62525129e-01 -4.53074336e-01 2.39522174e-01
-4.42009419e-01 2.96984911e-02 3.22283268e-01 1.27294064e+00
-1.52392292e+00 5.31432331e-01 -3.82189542e-01 -6.45094872e-01
-1.56640410e+00 6.69948578e-01 3.73296112e-01 -5.07751405e-01
-4.22458023e-01 9.93519783e-01 6.18226767e-01 2.43967906e-01
-1.41209923e-02 -4.29507136e-01 -3.96287024e-01 -2.53097534e-01
9.54560399e-01 8.30255210e-01 -4.70251650e-01 -7.57522523e-01
-3.83123785e-01 7.87736595e-01 6.36367053e-02 7.81372041e-02
5.29586732e-01 -1.67833120e-01 -7.77269483e-01 5.05848765e-01
1.03408742e+00 4.28498119e-01 -4.38119590e-01 7.36971211e-04
-5.83178364e-03 -7.04155624e-01 -4.52492386e-01 -4.34423834e-01
-1.16278446e+00 1.05887246e+00 4.80127454e-01 3.10360044e-01
8.74422669e-01 5.88436686e-02 1.38041711e+00 -1.88159451e-01
5.17065465e-01 -1.07514894e+00 3.27520698e-01 8.35315660e-02
-8.43444094e-02 -1.02526987e+00 1.69702083e-01 -5.91871381e-01
-2.42379129e-01 1.20359647e+00 4.61060464e-01 -2.84635991e-01
4.39210147e-01 7.15875506e-01 4.71629918e-01 -1.92211926e-01
-5.17136276e-01 -1.44827425e-01 -2.68721312e-01 8.94924939e-01
5.59672475e-01 1.10165142e-01 -6.23466849e-01 4.05854255e-01
3.48265797e-01 -5.88770732e-02 4.53175783e-01 9.20106947e-01
-1.24167192e+00 -9.82262015e-01 -1.30062985e+00 3.38122159e-01
-1.08608592e+00 5.66014290e-01 -5.69816053e-01 6.95808947e-01
2.45686576e-01 1.00662374e+00 5.74207723e-01 2.95063108e-01
-2.35341221e-01 -1.23057507e-01 6.13919795e-01 -3.97659659e-01
-5.53628266e-01 6.70495272e-01 -3.65700513e-01 -4.42448668e-02
-5.01437008e-01 -2.92780221e-01 -1.86085796e+00 -5.45005083e-01
-8.66496861e-02 4.16130155e-01 4.16059613e-01 8.36540580e-01
-4.85654265e-01 7.26123154e-02 2.93714792e-01 -1.47328436e-01
-5.27516529e-02 -8.58533919e-01 -1.28495896e+00 4.89045262e-01
2.36983702e-01 -4.36878562e-01 -5.33937871e-01 3.66273582e-01] | [14.835330963134766, -3.132556200027466] |
f53dbe51-8ffc-4150-8818-d5e044be48a5 | aclm-a-selective-denoising-based-generative | 2306.00928 | null | https://arxiv.org/abs/2306.00928v1 | https://arxiv.org/pdf/2306.00928v1.pdf | ACLM: A Selective-Denoising based Generative Data Augmentation Approach for Low-Resource Complex NER | Complex Named Entity Recognition (NER) is the task of detecting linguistically complex named entities in low-context text. In this paper, we present ACLM Attention-map aware keyword selection for Conditional Language Model fine-tuning), a novel data augmentation approach based on conditional generation to address the data scarcity problem in low-resource complex NER. ACLM alleviates the context-entity mismatch issue, a problem existing NER data augmentation techniques suffer from and often generates incoherent augmentations by placing complex named entities in the wrong context. ACLM builds on BART and is optimized on a novel text reconstruction or denoising task - we use selective masking (aided by attention maps) to retain the named entities and certain keywords in the input sentence that provide contextually relevant additional knowledge or hints about the named entities. Compared with other data augmentation strategies, ACLM can generate more diverse and coherent augmentations preserving the true word sense of complex entities in the sentence. We demonstrate the effectiveness of ACLM both qualitatively and quantitatively on monolingual, cross-lingual, and multilingual complex NER across various low-resource settings. ACLM outperforms all our neural baselines by a significant margin (1%-36%). In addition, we demonstrate the application of ACLM to other domains that suffer from data scarcity (e.g., biomedical). In practice, ACLM generates more effective and factual augmentations for these domains than prior methods. Code: https://github.com/Sreyan88/ACLM | ['Dinesh Manocha', 'S Ramaneswaran', 'Sonal Kumar', 'Manan Suri', 'Utkarsh Tyagi', 'Sreyan Ghosh'] | 2023-06-01 | null | null | null | null | ['named-entity-recognition-ner'] | ['natural-language-processing'] | [ 9.59074497e-02 2.22263932e-01 4.44618352e-02 -3.47121060e-01
-1.39057481e+00 -6.68195724e-01 5.00970602e-01 3.15468341e-01
-1.06563330e+00 9.76745725e-01 9.61145818e-01 -3.00761014e-01
3.80680859e-01 -4.59471703e-01 -7.76285529e-01 -2.28347585e-01
2.95672297e-01 4.27820355e-01 -3.00530583e-01 -3.85694653e-01
1.97441764e-02 2.33100072e-01 -9.81824696e-01 4.36274260e-01
1.10109413e+00 3.57672304e-01 5.32107174e-01 3.48997712e-01
-5.16106546e-01 6.65748835e-01 -8.08486462e-01 -6.60465717e-01
2.81918943e-02 -3.13902467e-01 -9.15470302e-01 -3.55715871e-01
7.27349579e-01 2.19231248e-01 -7.46797025e-02 8.76953125e-01
9.23650622e-01 3.01687419e-01 4.22526032e-01 -6.86658740e-01
-8.05800319e-01 1.05158126e+00 -3.48400354e-01 3.71418387e-01
1.30977973e-01 1.60995781e-01 1.01272917e+00 -1.72995889e+00
6.77928925e-01 1.25579417e+00 9.99249101e-01 8.67291629e-01
-1.29313231e+00 -7.96366692e-01 2.79391736e-01 -1.84308738e-01
-1.44100237e+00 -7.45313525e-01 3.70361835e-01 -2.00230494e-01
1.23610926e+00 2.58136064e-01 -5.77561259e-02 1.54083216e+00
-1.23476814e-02 7.50488460e-01 9.44430470e-01 -5.35140991e-01
2.80472729e-03 7.31154159e-02 3.05462420e-01 4.31983739e-01
3.06788743e-01 -2.24432454e-01 -7.14232504e-01 -2.51683384e-01
3.06857198e-01 -3.18750232e-01 -3.79571199e-01 5.23090243e-01
-1.57410741e+00 6.64400578e-01 4.04631674e-01 5.20376205e-01
-7.96933770e-01 -6.97407573e-02 4.33295518e-01 7.43008219e-03
5.83366275e-01 1.01200402e+00 -1.09164631e+00 -4.51336578e-02
-9.79067922e-01 -1.27499193e-01 8.36119115e-01 1.01716518e+00
6.81074083e-01 2.40843579e-01 -7.19851851e-01 9.68073845e-01
-2.57616155e-02 5.28383493e-01 6.57700062e-01 -4.89090562e-01
7.60241568e-01 5.17483592e-01 5.74751645e-02 -6.83961928e-01
-5.87444007e-01 -7.28586316e-01 -1.04494500e+00 -4.10069346e-01
1.72124550e-01 -6.32654727e-01 -1.12835860e+00 2.18688798e+00
1.85334340e-01 1.01122759e-01 3.23327959e-01 5.27175128e-01
1.14166164e+00 6.50911391e-01 7.34754980e-01 -5.77140646e-03
1.54741836e+00 -8.56840730e-01 -1.11740828e+00 -6.77775741e-01
8.14947784e-01 -7.83144534e-01 1.45926511e+00 -1.23830333e-01
-8.79596710e-01 -4.68529433e-01 -6.31445587e-01 -3.69206816e-01
-7.22352147e-01 4.60845441e-01 4.65222329e-01 5.49478352e-01
-1.04510295e+00 2.51939476e-01 -6.78949058e-01 -3.07171345e-01
2.88796961e-01 1.68152601e-01 -5.69099665e-01 -1.41253427e-01
-1.56513214e+00 1.07802987e+00 5.98430634e-01 1.78530589e-01
-5.93029201e-01 -1.27682948e+00 -1.29231966e+00 1.22654818e-01
3.14589351e-01 -7.39400208e-01 1.10429692e+00 -5.61849773e-01
-8.04242074e-01 7.81604469e-01 -5.02606750e-01 -5.21029234e-01
6.37999699e-02 -5.87704897e-01 -5.31630337e-01 -2.67651767e-01
6.47617579e-01 9.24093485e-01 1.78723350e-01 -9.30480242e-01
-3.69331568e-01 -1.67521626e-01 -3.71448457e-01 3.32922399e-01
-5.37009001e-01 1.60319030e-01 -4.71565783e-01 -1.13200784e+00
-1.79401323e-01 -6.61061704e-01 -5.83446622e-01 -5.81403136e-01
-6.97276950e-01 -2.35957980e-01 3.42689514e-01 -9.58127916e-01
1.37395883e+00 -2.15235138e+00 -1.17251948e-01 -2.11546928e-01
8.52029920e-02 2.96059489e-01 -4.00129825e-01 4.73690569e-01
-2.17322901e-01 6.29020512e-01 -5.45549989e-01 -7.59006798e-01
-6.67698309e-02 9.81336385e-02 -2.99253523e-01 -2.06154101e-02
7.46064007e-01 1.15293038e+00 -8.10661614e-01 -5.07559180e-01
-1.81415185e-01 7.07494318e-01 -4.26120132e-01 5.81664219e-03
-1.95851192e-01 5.35756826e-01 -1.30042180e-01 4.35673237e-01
5.28996408e-01 -2.08329737e-01 1.20289931e-02 -3.43420625e-01
-1.02825105e-01 7.69084752e-01 -1.15319109e+00 1.81904078e+00
-7.41607487e-01 5.36664367e-01 2.44841110e-02 -3.64660770e-01
6.96693420e-01 5.46552718e-01 5.76443523e-02 -5.40612519e-01
-1.58436865e-01 2.41625071e-01 -3.44940096e-01 -2.53002197e-01
7.07767189e-01 -2.56500781e-01 -4.29175973e-01 1.55368164e-01
4.35487598e-01 2.94577152e-01 2.74455786e-01 3.47242504e-01
1.15042913e+00 -2.08522737e-01 4.54239994e-01 -4.19919282e-01
2.92885333e-01 2.34022047e-02 9.13473189e-01 9.51810956e-01
6.62475154e-02 5.90181470e-01 2.42084768e-02 -5.65928109e-02
-9.91998196e-01 -8.72578621e-01 -1.81517586e-01 1.26449227e+00
-3.71656835e-01 -5.46191633e-01 -7.04179704e-01 -7.89746523e-01
-2.20676303e-01 1.09361112e+00 -5.67990780e-01 -4.51872535e-02
-7.95391500e-01 -1.09999561e+00 9.83254731e-01 6.45110607e-01
5.53475082e-01 -1.48868215e+00 1.45622432e-01 4.25110996e-01
-7.39960074e-01 -1.43765414e+00 -9.14971828e-01 4.89326119e-01
-5.87040424e-01 -6.53786063e-01 -6.38848007e-01 -9.94978189e-01
5.86707890e-01 -1.25180885e-01 1.59257293e+00 -2.52873331e-01
-1.01871714e-01 2.56586850e-01 -2.83337772e-01 -5.71025252e-01
-5.31208336e-01 5.08188188e-01 8.83750692e-02 -1.01313010e-01
5.41195869e-01 -3.92819464e-01 -3.43889475e-01 -1.17379175e-02
-8.27661455e-01 -6.29597306e-02 1.03932428e+00 1.17001069e+00
7.53533483e-01 -3.13036740e-01 9.35912073e-01 -1.23132122e+00
7.77295768e-01 -5.70512533e-01 -1.66293845e-01 2.69672006e-01
-4.65049088e-01 3.13205689e-01 5.32738805e-01 -5.76376319e-01
-1.28942037e+00 4.97796871e-02 -4.52499777e-01 3.63036171e-02
-4.95020956e-01 8.39499891e-01 -6.47946358e-01 5.05525351e-01
9.61713076e-01 2.24386647e-01 -6.48582757e-01 -8.26117516e-01
4.83596146e-01 5.74246645e-01 6.90932095e-01 -5.21870136e-01
5.73560774e-01 1.17146961e-01 -4.92020965e-01 -8.90510499e-01
-1.15927064e+00 -4.86145914e-01 -5.19916236e-01 2.68133819e-01
9.91135955e-01 -1.29624486e+00 -1.62521526e-01 1.91739112e-01
-1.37574375e+00 -2.12183535e-01 -4.86092448e-01 5.33195734e-01
1.23209208e-01 1.68804929e-01 -8.23144734e-01 -6.47510350e-01
-7.61658013e-01 -6.34703994e-01 1.12226462e+00 1.49062544e-01
-4.09500867e-01 -8.56197655e-01 -6.89052418e-02 3.04177791e-01
3.92732620e-01 1.45375296e-01 8.19270730e-01 -1.23225224e+00
-1.15816556e-01 -6.74261302e-02 1.48844030e-02 3.04324448e-01
1.33077083e-02 -6.11502826e-01 -1.01180172e+00 -1.56990021e-01
-5.88416122e-02 -1.65511817e-01 1.01267958e+00 1.80714726e-01
7.59846330e-01 -5.81920445e-01 -2.72327960e-01 3.41076970e-01
1.23577714e+00 -1.18550144e-01 5.28469980e-01 2.18740448e-01
8.31802845e-01 5.75617194e-01 4.60431159e-01 2.15335429e-01
4.77765739e-01 3.05805922e-01 -4.82356222e-03 -5.65548301e-01
-4.00474846e-01 -4.40994650e-01 3.90293032e-01 1.03274333e+00
4.52782989e-01 -4.19527948e-01 -1.18374562e+00 9.90109801e-01
-1.46796477e+00 -7.50779092e-01 -1.87651634e-01 1.97351015e+00
1.43329668e+00 1.47956893e-01 -3.09289068e-01 -3.45821917e-01
9.20561135e-01 -1.11883424e-01 -4.37494457e-01 -3.21912281e-02
-7.12635219e-01 5.07774830e-01 4.82403725e-01 5.66342235e-01
-1.32334495e+00 1.27717221e+00 5.06045437e+00 8.84878039e-01
-7.57039011e-01 5.05077958e-01 5.57561874e-01 1.20923445e-01
-4.20926273e-01 -1.41577572e-01 -1.33356786e+00 3.38043213e-01
1.14037263e+00 -3.48522477e-02 1.78707466e-02 7.01873660e-01
1.97009131e-01 2.23591045e-01 -1.00154364e+00 8.06365788e-01
2.50072908e-02 -1.37637138e+00 1.65943116e-01 -1.29938781e-01
7.41995573e-01 4.88719672e-01 -3.23213935e-02 6.46989584e-01
4.14814413e-01 -9.98764932e-01 5.17096341e-01 4.55390871e-01
9.41526830e-01 -6.23885334e-01 1.01877844e+00 4.07761961e-01
-1.05550754e+00 1.37557849e-01 -1.75494477e-01 3.26450497e-01
3.13977122e-01 8.64364028e-01 -9.90064859e-01 4.41327155e-01
7.44488716e-01 3.89791846e-01 -7.55840898e-01 8.80579948e-01
-3.85691196e-01 8.91071022e-01 -4.06983823e-01 -1.20978141e-02
1.73932761e-01 3.40918660e-01 7.07560062e-01 1.96402514e+00
1.70380436e-02 5.27356081e-02 6.70550764e-02 9.62757587e-01
-7.11455703e-01 4.87478018e-01 -5.34474730e-01 -2.47689888e-01
7.42924511e-01 1.17365444e+00 -3.65708977e-01 -3.49946201e-01
-4.29684311e-01 1.13280356e+00 6.04028463e-01 4.70299125e-01
-4.45080280e-01 -4.97505128e-01 6.53063059e-01 -1.21070251e-01
2.84507692e-01 -1.57549471e-01 -5.87469876e-01 -1.37491250e+00
2.12142751e-01 -9.67733145e-01 6.06770456e-01 -6.85526609e-01
-1.60804105e+00 9.45484459e-01 -4.26591605e-01 -8.73604298e-01
-1.54626891e-01 -4.15809423e-01 -3.93522263e-01 1.20232391e+00
-1.48217738e+00 -1.22439373e+00 7.30626956e-02 5.09957194e-01
7.71656513e-01 -1.07304193e-01 1.12464774e+00 6.56675756e-01
-5.27440667e-01 9.31227148e-01 -1.70324311e-01 6.46750927e-01
1.01666534e+00 -1.40302694e+00 9.35962439e-01 1.13089502e+00
3.40252787e-01 1.01744890e+00 4.76953894e-01 -9.50340450e-01
-8.12701344e-01 -1.47921777e+00 1.61993504e+00 -7.97294259e-01
4.62560028e-01 -7.48064756e-01 -1.06369472e+00 8.05549085e-01
3.79043341e-01 -1.97270721e-01 9.76858616e-01 4.09328192e-01
-3.91382664e-01 2.72895008e-01 -1.07702875e+00 8.79006207e-01
9.86975849e-01 -6.47581637e-01 -7.74083555e-01 3.73986363e-01
1.23282135e+00 -2.50879467e-01 -8.27900350e-01 4.60827500e-01
-8.77992436e-02 -2.16793239e-01 8.06702018e-01 -1.02997220e+00
1.00919649e-01 -2.17548192e-01 -2.73604989e-01 -1.40706515e+00
-2.88170069e-01 -5.60790598e-01 9.26768631e-02 1.68997610e+00
1.06324422e+00 -3.81952912e-01 4.38958168e-01 5.90887845e-01
-4.51691717e-01 -6.00831151e-01 -9.37904716e-01 -5.53889573e-01
3.11722755e-01 -4.72983479e-01 4.08036411e-01 1.43021274e+00
-2.05250308e-01 8.45530391e-01 -2.71268755e-01 3.98192108e-01
4.24786627e-01 -3.65831822e-01 2.12950021e-01 -9.28736448e-01
-7.64050428e-03 -2.19775751e-01 1.19752251e-02 -8.92223656e-01
4.25697178e-01 -1.11314857e+00 2.56926805e-01 -1.56699491e+00
1.40656948e-01 -3.18074256e-01 -3.85535777e-01 1.00395989e+00
-6.54268801e-01 1.35957733e-01 1.44845620e-01 -2.62398757e-02
-4.86587912e-01 6.09442472e-01 7.76204228e-01 -3.69747728e-02
-2.52696306e-01 -3.10174376e-01 -1.04811072e+00 5.78855455e-01
8.96522880e-01 -6.81271195e-01 1.59703538e-01 -6.34431660e-01
2.98785478e-01 -2.27313221e-01 2.48584867e-01 -7.42302835e-01
2.50273317e-01 2.09069714e-01 4.92653638e-01 -6.23553216e-01
1.49004772e-01 -4.97850120e-01 -1.90327615e-01 1.98821396e-01
-5.77288985e-01 2.59555429e-01 5.54878414e-01 4.48376119e-01
-1.88826814e-01 -1.45713612e-01 6.46600664e-01 -3.55874032e-01
-7.63465524e-01 1.20870724e-01 -4.82292533e-01 7.31332123e-01
2.53818333e-01 3.35755974e-01 -3.49214196e-01 -2.39421070e-01
-9.60086823e-01 2.90431947e-01 -8.39830562e-03 5.59785843e-01
5.32598615e-01 -1.30112493e+00 -1.17378807e+00 3.79131995e-02
1.60586700e-01 4.21110094e-02 2.32852668e-01 6.95076287e-01
1.17737148e-02 4.73120511e-01 2.96170622e-01 -1.11165851e-01
-9.31036532e-01 4.71860439e-01 4.10280317e-01 -4.62769747e-01
-4.24376905e-01 9.75331128e-01 3.37643236e-01 -1.00293207e+00
2.26727009e-01 -2.58445591e-01 -1.73344880e-01 1.83892623e-02
5.69177568e-01 1.07342303e-01 4.48744267e-01 -6.88910961e-01
-4.82681811e-01 -2.07987539e-02 -2.21614361e-01 -2.55893439e-01
1.39194202e+00 -1.90979123e-01 -9.14775580e-02 3.85246813e-01
1.02808142e+00 4.52215493e-01 -6.81056201e-01 -5.20067453e-01
4.80428845e-01 2.41622820e-01 6.25260919e-02 -1.33744955e+00
-6.90614462e-01 6.62782609e-01 3.66572201e-01 -4.06254947e-01
1.05739129e+00 2.40795210e-01 8.59466732e-01 5.90540707e-01
-5.64305205e-03 -9.80955958e-01 -2.02524915e-01 8.98559749e-01
1.01110601e+00 -1.35026467e+00 -4.31302398e-01 -4.03622538e-01
-8.22086394e-01 5.96243918e-01 8.53323340e-01 3.73131752e-01
5.62364042e-01 4.01684284e-01 2.88700342e-01 -6.97558001e-02
-6.97242558e-01 -3.52287084e-01 4.68203962e-01 5.62555254e-01
7.60311067e-01 -1.23031430e-01 -4.53758776e-01 1.11855996e+00
-2.40017265e-01 -5.40099919e-01 4.35604721e-01 7.04074442e-01
-1.39126495e-01 -9.48183060e-01 -2.91510314e-01 3.95877004e-01
-8.66377473e-01 -1.04203606e+00 -4.41683292e-01 6.55864775e-01
1.63462669e-01 1.03124404e+00 -8.57777223e-02 -8.09195563e-02
6.01411998e-01 4.98718053e-01 -2.25105107e-01 -9.20119941e-01
-1.01948476e+00 6.72608092e-02 3.99888277e-01 -2.98553646e-01
-1.56426698e-01 -6.60820425e-01 -1.31040347e+00 1.34032950e-01
-3.28307092e-01 3.49711388e-01 6.80778146e-01 7.91252911e-01
8.40498567e-01 5.64605713e-01 1.65377706e-01 -4.05506462e-01
-3.55744064e-01 -1.38057244e+00 -1.70999840e-01 4.86195177e-01
3.18576336e-01 -2.89296985e-01 -3.54699850e-01 5.95921800e-02] | [9.776643753051758, 9.523374557495117] |
2e6e3bcd-362a-46cf-849a-9f94545d4a46 | optical-flow-based-online-moving-foreground | 1811.07256 | null | http://arxiv.org/abs/1811.07256v1 | http://arxiv.org/pdf/1811.07256v1.pdf | Optical Flow Based Online Moving Foreground Analysis | Obtained by moving object detection, the foreground mask result is unshaped
and can not be directly used in most subsequent processes. In this paper, we
focus on this problem and address it by constructing an optical flow based
moving foreground analysis framework. During the processing procedure, the
foreground masks are analyzed and segmented through two complementary
clustering algorithms. As a result, we obtain the instance-level information
like the number, location and size of moving objects. The experimental result
show that our method adapts itself to the problem and performs well enough for
practical applications. | ['Jiagang Zhu', 'Wei Zou', 'Junjie Huang', 'Zheng Zhu'] | 2018-11-18 | null | null | null | null | ['moving-object-detection'] | ['computer-vision'] | [ 4.50973541e-01 -4.50221956e-01 -7.02649206e-02 9.08926129e-03
6.67302758e-02 -6.22273743e-01 6.19016767e-01 -1.92459468e-02
-5.11724353e-01 6.69251680e-01 -4.87704873e-01 -3.74659568e-01
-1.55648887e-02 -8.17532539e-01 -8.57127756e-02 -9.08235013e-01
-6.87343031e-02 3.89331847e-01 1.11680508e+00 3.99435699e-01
5.29038250e-01 8.61276567e-01 -1.62251341e+00 1.86697543e-01
6.56045914e-01 7.79705346e-01 2.76344031e-01 1.06999660e+00
-6.21083379e-01 1.04768431e+00 -7.61818111e-01 8.90340135e-02
2.86569178e-01 -8.49912226e-01 -9.24842477e-01 6.32195115e-01
-9.68436822e-02 -2.50674844e-01 -4.10199687e-02 1.04288721e+00
-6.68969527e-02 5.42743325e-01 4.86744314e-01 -1.15306807e+00
7.89885670e-02 4.13058788e-01 -6.43023908e-01 9.63241458e-01
5.61293103e-02 -1.33426458e-01 4.30116981e-01 -8.58976007e-01
8.01238894e-01 1.17383027e+00 1.38353571e-01 5.02725601e-01
-8.52926850e-01 -1.73340634e-01 4.36173588e-01 3.73620003e-01
-1.39801121e+00 -3.89015615e-01 8.48447919e-01 -4.93198633e-01
3.30182195e-01 6.82174146e-01 7.90844738e-01 2.77458668e-01
5.33618852e-02 9.59559560e-01 8.63325953e-01 -5.22184968e-01
5.11346102e-01 -1.48069151e-02 4.41113651e-01 6.03713930e-01
4.37769026e-01 -2.69502699e-01 -1.08509615e-01 1.43953338e-02
8.77971351e-01 7.83228800e-02 -4.85874027e-01 -5.45100808e-01
-1.37048197e+00 4.36399370e-01 1.73137209e-03 8.94991696e-01
-1.78756922e-01 1.21592984e-01 -6.42998377e-03 -2.26053432e-01
4.57071900e-01 -1.93150014e-01 -6.28453791e-02 -1.63995892e-01
-1.47602284e+00 1.12385266e-01 6.77003205e-01 8.65052521e-01
8.84192109e-01 -3.93979959e-02 -9.84617770e-02 1.80016145e-01
2.22291172e-01 4.89102215e-01 2.34026089e-01 -1.05092895e+00
3.01731229e-02 7.82393873e-01 3.10513496e-01 -1.16454911e+00
-4.67766494e-01 8.07753578e-02 -4.93660450e-01 3.59285265e-01
7.13093579e-01 -5.26434183e-02 -1.01922011e+00 1.01547158e+00
7.75362968e-01 5.26543379e-01 -1.25439733e-01 7.95004547e-01
6.91179156e-01 8.80983233e-01 -2.28038821e-02 -7.14934111e-01
1.06273794e+00 -1.12360120e+00 -1.17447531e+00 -3.61781754e-02
2.86080301e-01 -1.03693366e+00 3.72875631e-01 4.34934199e-01
-1.28311455e+00 -7.08473563e-01 -8.21505308e-01 1.37175843e-01
-4.06971216e-01 1.01892250e-02 7.10250139e-01 6.92955434e-01
-1.00897145e+00 5.89404821e-01 -1.28081942e+00 -3.08376402e-01
2.23365888e-01 3.59491706e-01 3.23838107e-02 -6.57323301e-02
-5.01385570e-01 6.54356897e-01 8.05498838e-01 6.59445524e-01
-5.92270613e-01 -2.80302279e-02 -5.28297484e-01 -1.25559479e-01
4.82372999e-01 -4.61615324e-01 1.07822680e+00 -1.01778650e+00
-1.53779757e+00 6.42860770e-01 -6.88309014e-01 -1.40980691e-01
7.85263479e-01 1.09312661e-01 -3.98444653e-01 6.33331180e-01
-7.77600706e-02 3.09935898e-01 9.83131886e-01 -1.50853717e+00
-1.23810172e+00 -9.64108184e-02 -1.02835126e-01 2.33895890e-02
-2.94730999e-03 3.83987486e-01 -8.27246845e-01 -4.98132765e-01
2.65212953e-01 -6.53456151e-01 -3.48421633e-01 -9.92595851e-02
-1.92299992e-01 -1.00770444e-02 1.40045822e+00 -5.53024173e-01
1.50225008e+00 -2.28819323e+00 2.33795732e-01 3.35558146e-01
4.36850905e-01 6.03756487e-01 4.68046248e-01 4.73414063e-02
2.93125987e-01 9.90765542e-02 -6.59506321e-01 -2.92055517e-01
-5.40074348e-01 4.29850668e-01 8.49447325e-02 7.82085240e-01
1.32700995e-01 6.10212326e-01 -1.03920412e+00 -1.00445890e+00
7.81413972e-01 2.99577385e-01 -1.58721745e-01 2.26660654e-01
1.03345536e-01 5.13069451e-01 -5.89134991e-01 6.83193088e-01
1.08936298e+00 -3.72931175e-02 2.60646641e-02 1.03726551e-01
-4.27638710e-01 -3.37436497e-01 -1.65227139e+00 1.09055924e+00
5.18712923e-02 8.27893376e-01 5.23235083e-01 -8.15344572e-01
5.97802699e-01 5.58094755e-02 7.77411699e-01 -8.73201415e-02
3.28908563e-01 1.07319534e-01 2.26775065e-01 -4.71656591e-01
6.70009553e-01 -1.98433660e-02 4.02457952e-01 3.96486759e-01
-2.38429472e-01 1.91018149e-01 6.68856084e-01 1.35506079e-01
9.54213381e-01 1.51546061e-01 3.63337636e-01 -4.35800761e-01
1.03417909e+00 3.58202726e-01 6.35834992e-01 6.09914839e-01
-4.45346475e-01 5.46167672e-01 2.56934941e-01 -3.65053058e-01
-6.53556287e-01 -1.01878202e+00 -2.25796908e-01 8.02854598e-01
7.12360442e-01 -1.13380693e-01 -1.18490767e+00 -6.24364793e-01
-2.14024439e-01 4.95957226e-01 -6.26254499e-01 2.00875118e-01
-1.10570180e+00 -8.90807807e-01 -3.47120240e-02 3.16194206e-01
2.98630923e-01 -1.22391045e+00 -1.08823812e+00 4.85782981e-01
-2.17471525e-01 -1.18859339e+00 -3.79519939e-01 -6.10811375e-02
-1.18025959e+00 -1.23219287e+00 -8.35640728e-01 -7.39811540e-01
8.05547714e-01 8.36104870e-01 7.71555126e-01 4.68359172e-01
-2.85778552e-01 3.30785602e-01 -5.49354851e-01 -2.46735364e-01
-5.53196549e-01 -1.79066718e-01 -2.36706793e-01 6.23137891e-01
1.40337572e-01 -2.66384929e-01 -5.38460553e-01 6.21299803e-01
-1.13673019e+00 -9.84301493e-02 2.91321278e-01 1.80815965e-01
4.78148878e-01 5.52536547e-01 1.49018737e-02 -8.12221587e-01
2.58246690e-01 -1.10952497e-01 -8.86914015e-01 1.24031030e-01
-9.89939049e-02 -3.32708210e-01 3.96242678e-01 -4.85090315e-01
-1.25324810e+00 2.27137357e-01 3.53988230e-01 -3.82978708e-01
-4.12128001e-01 -1.24876246e-01 -3.68752360e-01 -2.49686405e-01
7.36182109e-02 2.50294775e-01 -2.87306249e-01 -5.59443712e-01
3.34149808e-01 4.82013553e-01 5.83260179e-01 -1.90011580e-02
8.57983589e-01 1.05866766e+00 4.30172421e-02 -9.83608186e-01
-2.72319108e-01 -8.95369649e-01 -1.04953730e+00 -5.84998190e-01
1.08464634e+00 -2.79126167e-01 -4.37968910e-01 5.79547942e-01
-1.26139677e+00 -2.75066078e-01 -1.90831810e-01 5.28798163e-01
-3.50327909e-01 8.06417942e-01 -6.24593556e-01 -1.32787633e+00
5.08789495e-02 -1.12943494e+00 8.10531318e-01 2.37717390e-01
9.14872289e-02 -1.17966139e+00 -1.25005350e-01 1.36754915e-01
1.40224040e-01 2.81068504e-01 4.52004910e-01 -3.67426515e-01
-1.00493228e+00 -2.84181237e-02 -9.23651531e-02 8.02605525e-02
3.48192722e-01 5.86528838e-01 -7.91930437e-01 1.96189024e-02
2.92030871e-01 8.17435682e-01 9.23089504e-01 6.77926779e-01
9.68339801e-01 1.89494342e-01 -7.04294145e-01 4.52104956e-01
1.21164823e+00 6.34782612e-01 7.15243340e-01 2.90429324e-01
8.16147327e-01 6.20880127e-01 8.75682771e-01 2.85606772e-01
-6.74495250e-02 4.45910484e-01 3.28475356e-01 -3.14564943e-01
-2.70691782e-01 3.67238849e-01 1.95656344e-01 7.59315372e-01
-3.28069329e-01 -6.00802839e-01 -9.62721109e-01 6.18863344e-01
-2.12274933e+00 -1.13042021e+00 -7.22130001e-01 1.94006050e+00
3.37785393e-01 1.80630282e-01 1.80452093e-01 5.48306286e-01
1.24537086e+00 1.09909222e-01 -1.93419471e-01 -1.25790611e-01
-8.50334540e-02 -3.04819141e-02 4.03229952e-01 7.09342182e-01
-1.28665221e+00 1.13593054e+00 7.01190233e+00 7.38604724e-01
-8.31761837e-01 1.17847417e-02 4.05933827e-01 9.78007242e-02
8.64726380e-02 1.58089608e-01 -7.06570804e-01 5.37510693e-01
5.87673306e-01 -2.11435959e-01 2.84841985e-01 5.59886575e-01
3.23581755e-01 -7.19857991e-01 -8.52134287e-01 9.23739731e-01
-1.06683351e-01 -1.09910023e+00 5.47233187e-02 2.56097317e-02
6.07491553e-01 -5.88351548e-01 -4.05433238e-01 -2.90427804e-01
-9.67661738e-02 -6.18206382e-01 8.60645711e-01 4.83820766e-01
2.69832667e-02 -7.11878717e-01 7.83749282e-01 4.09594297e-01
-1.42256868e+00 -1.38928562e-01 -3.30192298e-01 -2.22457364e-01
7.41246819e-01 7.49858499e-01 -7.49146879e-01 6.48722470e-01
4.47847456e-01 3.73096496e-01 -5.56693912e-01 1.45042491e+00
-1.23989815e-02 6.05815232e-01 -3.02963465e-01 -1.09421864e-01
2.79888213e-01 -3.78591627e-01 5.81933558e-01 1.36712313e+00
2.40488723e-01 1.92697063e-01 4.02687579e-01 7.27404952e-01
4.61187005e-01 3.10768127e-01 -2.98172235e-01 7.84088597e-02
9.47692469e-02 1.43861914e+00 -1.92017531e+00 -7.37246513e-01
-3.21643680e-01 1.04841352e+00 -1.95321545e-01 3.70921016e-01
-9.20807600e-01 -2.48817965e-01 2.88467258e-01 1.22194923e-01
5.90663731e-01 -4.89696473e-01 -2.34301999e-01 -1.24097943e+00
-9.78372097e-02 -2.92409807e-01 2.54015446e-01 -3.99838150e-01
-7.57752419e-01 5.01025796e-01 2.01790377e-01 -1.12462890e+00
-2.13981830e-02 -7.28980958e-01 -7.82058537e-01 4.77983743e-01
-1.37012851e+00 -7.92103171e-01 -3.25580150e-01 7.25709081e-01
6.11727476e-01 2.67133653e-01 1.56488374e-01 3.94687206e-01
-8.27568054e-01 -2.54080564e-01 1.08107321e-01 2.61886448e-01
1.60229608e-01 -1.14173245e+00 2.49510914e-01 1.66217256e+00
1.70118794e-01 4.20820773e-01 6.86076522e-01 -6.87555194e-01
-1.08883011e+00 -9.86173332e-01 7.02364802e-01 -4.30985183e-01
6.37988985e-01 -3.33992988e-01 -9.63675559e-01 4.13909465e-01
8.29290897e-02 2.35913336e-01 4.52036411e-01 -4.63230103e-01
6.28790200e-01 1.78248614e-01 -1.06761706e+00 3.84593934e-01
1.07393026e+00 1.61593944e-01 -4.97358859e-01 8.76418948e-02
4.56364810e-01 -3.40836942e-01 -4.65097725e-01 2.98053294e-01
1.21686295e-01 -9.97538447e-01 8.33549142e-01 -4.38559532e-01
-2.03187615e-01 -7.90956795e-01 7.05702007e-02 -5.57363510e-01
-3.86013687e-01 -6.21342003e-01 -4.47069108e-01 1.34548450e+00
3.67202088e-02 -3.81530672e-01 8.17098916e-01 4.58006024e-01
7.20139220e-03 -1.83749005e-01 -1.07607841e+00 -7.70514131e-01
-2.98578560e-01 -2.95450240e-01 2.38909706e-01 8.72944295e-01
-3.56484413e-01 -6.30637407e-02 -1.15288436e-01 3.36120158e-01
7.96893001e-01 6.58618808e-01 7.43693471e-01 -1.34540260e+00
5.52683435e-02 -6.69711709e-01 -5.20760596e-01 -9.95541811e-01
-2.98846364e-02 -4.01802421e-01 1.48438826e-01 -1.43076587e+00
1.26292542e-01 -2.37081379e-01 -1.53876290e-01 1.25170462e-02
-4.55445766e-01 2.82098979e-01 4.84340489e-01 2.37128094e-01
-8.03672969e-01 3.64110544e-02 1.21485031e+00 6.09299242e-02
-4.85560030e-01 3.35511535e-01 -4.43726964e-02 9.87807274e-01
8.82661700e-01 -3.58447939e-01 -2.18684852e-01 -7.95537010e-02
-5.68795860e-01 -5.44949956e-02 2.77106613e-01 -1.09226263e+00
2.11868912e-01 -4.99726504e-01 3.65023732e-01 -8.11221838e-01
3.49247009e-02 -1.09584880e+00 1.77538544e-01 5.87795854e-01
1.82710081e-01 8.79591182e-02 8.37493241e-02 5.96444845e-01
-1.15931414e-01 -6.34015441e-01 8.22268426e-01 -2.22378865e-01
-1.00946951e+00 1.54988393e-01 -9.91231859e-01 -2.35509872e-01
1.46101844e+00 -6.20523274e-01 1.19201476e-02 -1.84140243e-02
-1.07396269e+00 7.97750950e-02 6.28893614e-01 7.63765424e-02
4.42968339e-01 -1.10570025e+00 -3.47748965e-01 1.37641700e-02
-3.31061661e-01 2.21130788e-01 1.47455126e-01 8.57901216e-01
-1.01157141e+00 2.42290393e-01 -9.94439200e-02 -8.06174219e-01
-1.55900109e+00 1.22253609e+00 1.44948483e-01 1.90234724e-02
-7.22921073e-01 4.69258338e-01 2.85328358e-01 3.06322306e-01
7.97480494e-02 -6.23944998e-01 -3.13346595e-01 2.31499955e-01
9.03270423e-01 9.35612142e-01 -2.00821504e-01 -8.10685217e-01
-4.60870564e-01 7.33430445e-01 2.93664962e-01 -1.90835014e-01
9.29395080e-01 -5.50962865e-01 -3.85671586e-01 6.49686933e-01
9.67053235e-01 1.80263862e-01 -1.25558555e+00 5.31501174e-02
2.92227328e-01 -7.93034136e-01 3.35764848e-02 -1.16203122e-01
-1.23620033e+00 8.26508641e-01 5.61333716e-01 7.04879880e-01
1.34647048e+00 9.12204501e-04 5.90467632e-01 1.15634523e-01
2.58235097e-01 -1.13656008e+00 -2.79696167e-01 2.32403964e-01
1.71556026e-01 -8.13222289e-01 5.05296588e-02 -9.81347859e-01
-2.31573761e-01 1.29157245e+00 3.74480933e-01 7.46378228e-02
4.92043138e-01 5.15730679e-01 1.79844618e-01 6.69118464e-02
-2.85650492e-01 -5.17160296e-01 1.30899757e-01 6.16794884e-01
5.26617318e-02 -1.38307124e-01 -4.53446954e-01 -4.40813713e-02
3.17985982e-01 -1.14498325e-01 7.39158273e-01 1.36544430e+00
-1.01329720e+00 -1.06707752e+00 -8.96908045e-01 -9.85136777e-02
-5.07053196e-01 3.26633483e-01 -5.50754964e-01 9.08789277e-01
3.04607391e-01 1.13527393e+00 1.48531258e-01 8.00884515e-02
1.31945401e-01 -2.87354011e-02 5.01580179e-01 -2.95134187e-01
-3.27681661e-01 4.64009613e-01 -2.66979545e-01 -4.88240004e-01
-1.03566921e+00 -8.19160104e-01 -1.54389715e+00 -2.77802765e-01
-6.19278669e-01 1.99158266e-01 3.03325295e-01 7.51944602e-01
-1.29150063e-01 7.04885066e-01 4.94134873e-01 -1.08582628e+00
3.19859654e-01 -5.20683825e-01 -6.37956083e-01 3.48362893e-01
3.82530719e-01 -7.17284679e-01 -4.02954102e-01 4.83032495e-01] | [8.942069053649902, -0.7875672578811646] |
4f049b61-cd6f-4a22-a827-82cbcb56b2c8 | cyclegan-with-a-blur-kernel-for-deconvolution | 1908.09414 | null | https://arxiv.org/abs/1908.09414v3 | https://arxiv.org/pdf/1908.09414v3.pdf | CycleGAN with a Blur Kernel for Deconvolution Microscopy: Optimal Transport Geometry | Deconvolution microscopy has been extensively used to improve the resolution of the wide-field fluorescent microscopy, but the performance of classical approaches critically depends on the accuracy of a model and optimization algorithms. Recently, the convolutional neural network (CNN) approaches have been studied as a fast and high performance alternative. Unfortunately, the CNN approaches usually require matched high resolution images for supervised training. In this paper, we present a novel unsupervised cycle-consistent generative adversarial network (cycleGAN) with a linear blur kernel, which can be used for both blind- and non-blind image deconvolution. In contrast to the conventional cycleGAN approaches that require two deep generators, the proposed cycleGAN approach needs only a single deep generator and a linear blur kernel, which significantly improves the robustness and efficiency of network training. We show that the proposed architecture is indeed a dual formulation of an optimal transport problem that uses a special form of the penalized least squares cost as a transport cost. Experimental results using simulated and real experimental data confirm the efficacy of the algorithm. | ['Sang-Eun Lee', 'Sungjun Lim', 'Jong Chul Ye', 'Hyoungjun Park', 'Sunghoe Chang'] | 2019-08-26 | null | null | null | null | ['image-deconvolution'] | ['computer-vision'] | [ 3.27239424e-01 -2.94340223e-01 5.67553759e-01 -2.08824083e-01
-5.92172444e-01 -5.01632750e-01 4.71196324e-01 -4.41553921e-01
-6.50703669e-01 1.18395519e+00 -3.54568541e-01 -4.38467152e-02
-1.46417826e-01 -3.67579401e-01 -9.53128278e-01 -1.26805246e+00
4.94634539e-01 5.55102378e-02 2.05022603e-01 1.48415208e-01
2.44016856e-01 5.76719463e-01 -1.17716479e+00 -2.60591954e-01
1.01246047e+00 8.43260169e-01 5.14044762e-01 7.21720695e-01
9.41129327e-02 8.90148342e-01 -4.55506593e-01 -2.34995320e-01
3.36223036e-01 -8.03550303e-01 -6.85381114e-01 8.25951025e-02
1.84906185e-01 -4.01744962e-01 -2.35951588e-01 1.34455621e+00
8.04274023e-01 1.03483565e-01 5.02722919e-01 -9.54478323e-01
-6.92768395e-01 1.76749155e-02 -4.47553545e-01 2.95250952e-01
-2.05533504e-01 2.71887571e-01 3.49979520e-01 -7.48556018e-01
5.72015405e-01 8.77309501e-01 6.61418498e-01 7.76903570e-01
-1.51960182e+00 -5.09748816e-01 -4.00268346e-01 7.95528386e-03
-1.24865258e+00 -3.99574101e-01 6.78819299e-01 -5.26484966e-01
4.46055949e-01 -2.40676235e-02 4.65073586e-01 7.65686750e-01
1.41237840e-01 1.91136390e-01 1.65873754e+00 -3.09152454e-01
2.72654712e-01 2.03355495e-02 -1.92826882e-01 6.78541541e-01
3.90297711e-01 2.43807688e-01 -1.03305317e-01 -6.95211440e-02
1.28829765e+00 1.14417158e-01 -7.98572600e-01 -2.87431151e-01
-1.10173833e+00 7.06755996e-01 6.53677583e-01 3.34313363e-01
-3.08487803e-01 3.00652057e-01 -1.04030371e-01 -1.15597755e-01
4.18901652e-01 5.06346703e-01 -1.03810960e-02 2.77323604e-01
-1.10610950e+00 1.52496368e-01 6.24721766e-01 5.80398619e-01
9.31725621e-01 2.84961581e-01 -1.15667202e-01 5.89186132e-01
2.42148727e-01 4.80240762e-01 4.44897503e-01 -1.15860772e+00
-1.26566201e-01 5.42498231e-01 3.42890263e-01 -7.70733178e-01
-2.34786317e-01 -4.41959709e-01 -1.03418911e+00 5.89514852e-01
7.85638452e-01 -2.62385309e-01 -9.27183509e-01 1.68493509e+00
2.58551508e-01 5.05207658e-01 1.80192858e-01 1.19703996e+00
4.53553438e-01 6.26320839e-01 -2.61724710e-01 -3.06061894e-01
1.01323855e+00 -9.13026392e-01 -8.80671382e-01 5.89066297e-02
8.14891830e-02 -8.56420219e-01 7.70430267e-01 1.81089878e-01
-1.14376187e+00 -3.17436963e-01 -1.03661442e+00 -1.42786369e-01
-3.38459283e-01 1.95864126e-01 4.09915507e-01 5.97875714e-01
-1.18093622e+00 8.83181810e-01 -7.56862581e-01 -2.43019179e-01
4.92695868e-01 5.24394631e-01 -4.97683853e-01 -1.00082487e-01
-8.33555937e-01 8.03171396e-01 2.93772936e-01 4.26487207e-01
-1.13693738e+00 -6.81919575e-01 -5.17696261e-01 1.30053470e-02
-2.93938001e-03 -7.78897345e-01 8.05571258e-01 -9.19835389e-01
-1.94538033e+00 7.49748409e-01 -3.23683731e-02 -5.28757691e-01
6.87599599e-01 -8.36545154e-02 1.34965286e-01 3.56291920e-01
-1.56868309e-01 5.00240147e-01 9.43202972e-01 -1.29004288e+00
-2.37252623e-01 -2.50502229e-01 9.38572958e-02 -2.17154786e-01
-1.17962910e-02 1.63290083e-01 1.25539117e-02 -5.13966143e-01
-2.08576187e-01 -8.78928125e-01 -2.32821956e-01 1.93243444e-01
-3.27937245e-01 3.40389043e-01 8.69062603e-01 -7.20419228e-01
6.11116230e-01 -1.94643724e+00 2.51971275e-01 -1.28247157e-01
2.50261545e-01 4.40445602e-01 1.12323187e-01 1.22735538e-01
-6.98661134e-02 -1.13121450e-01 -6.77524030e-01 -4.08631861e-01
-3.44769031e-01 2.78396666e-01 -7.88845718e-02 9.68378305e-01
3.66430700e-01 8.59061956e-01 -8.13446045e-01 -2.91398078e-01
2.36456856e-01 9.79660451e-01 -4.26490217e-01 6.03966475e-01
-1.37581415e-02 1.01443434e+00 -4.31072665e-03 2.06728116e-01
9.13863480e-01 -5.82938075e-01 -5.38671352e-02 -4.03347045e-01
-3.07657242e-01 -4.00567412e-01 -9.49472666e-01 1.61886644e+00
-1.45585611e-01 5.21355808e-01 4.51741487e-01 -1.11095357e+00
7.14613974e-01 3.49963218e-01 3.04427028e-01 -2.98117101e-01
2.67310172e-01 5.47628403e-01 -1.50732428e-01 -4.54020649e-01
9.01069567e-02 -4.83340681e-01 6.09506726e-01 2.98264951e-01
1.69673413e-01 -1.17668711e-01 1.07170507e-01 -9.87745672e-02
9.24467087e-01 3.82185400e-01 8.42212960e-02 -4.25387353e-01
8.02406728e-01 -1.92145422e-01 6.45251691e-01 4.74613994e-01
-6.71700537e-02 9.03787851e-01 3.41474801e-01 -2.65695065e-01
-1.35495913e+00 -5.30556202e-01 -8.38024542e-02 2.37252802e-01
2.99263388e-01 4.56886172e-01 -1.19652367e+00 -4.06546682e-01
-4.32088196e-01 5.86151108e-02 -5.14649808e-01 1.61758438e-01
-6.30994678e-01 -1.13217711e+00 6.94593549e-01 2.39307985e-01
7.57357776e-01 -7.21842945e-01 -5.30163586e-01 1.40108362e-01
-1.80387273e-02 -1.31086957e+00 -3.82470697e-01 2.56322503e-01
-6.99863791e-01 -1.25622940e+00 -1.29826820e+00 -7.26285517e-01
9.77684379e-01 2.82497287e-01 5.28198183e-01 1.12356082e-01
-2.17899874e-01 4.37576771e-02 -1.06801763e-01 2.63668671e-02
-4.36368912e-01 -3.69790345e-01 -3.86013389e-02 3.38942587e-01
-2.57285327e-01 -6.66653693e-01 -9.28330719e-01 2.96049625e-01
-1.27548087e+00 4.71428484e-02 7.37269878e-01 1.21512520e+00
7.06178904e-01 4.36215661e-02 5.58933377e-01 -8.41123402e-01
4.47056979e-01 -2.37976015e-01 -1.14572370e+00 9.47161242e-02
-7.56532192e-01 2.75944591e-01 9.51606512e-01 -3.34363550e-01
-1.10452342e+00 2.26523012e-01 5.79300150e-02 -6.96104348e-01
-1.05430454e-01 1.87468961e-01 -2.82884054e-02 -8.58327866e-01
4.28437680e-01 4.86424953e-01 3.03460270e-01 -5.33995569e-01
1.36794791e-01 3.25322360e-01 7.27685988e-01 -3.14382106e-01
1.04342926e+00 8.69964361e-01 3.85276616e-01 -6.95487559e-01
-4.73168880e-01 -3.67236406e-01 -4.34553266e-01 -1.26301646e-01
1.10511434e+00 -6.66527867e-01 -8.14562023e-01 8.74236047e-01
-1.42467356e+00 -3.85731310e-01 -9.08520222e-02 4.94851083e-01
-6.51399314e-01 6.26262546e-01 -7.11143911e-01 -8.50277364e-01
-4.12203014e-01 -1.44694257e+00 7.59784102e-01 5.38507462e-01
5.03948450e-01 -1.15246117e+00 -1.17980596e-02 2.42864922e-01
7.41853714e-01 5.97200692e-01 6.45667553e-01 -2.03076765e-01
-9.62280035e-01 -5.98423108e-02 -5.81627131e-01 7.57647634e-01
3.85346860e-02 -1.99943528e-01 -1.15599692e+00 -5.74406981e-01
4.29963708e-01 -2.63856769e-01 7.38215983e-01 4.52025622e-01
9.96530950e-01 -4.09172684e-01 -7.19213411e-02 1.08334029e+00
2.10409617e+00 1.10751420e-01 7.94664681e-01 2.03800827e-01
7.05905318e-01 3.22219491e-01 5.94468005e-02 1.53918013e-01
-8.55816826e-02 4.65278774e-01 7.05660760e-01 -3.82076442e-01
-2.38181755e-01 1.67386949e-01 2.74756223e-01 7.62011945e-01
-4.34063882e-01 -2.51353830e-01 -4.87851918e-01 5.91316342e-01
-1.79960227e+00 -8.82525802e-01 -3.15072268e-01 2.19636416e+00
8.76069546e-01 -4.64378536e-01 -2.31877282e-01 1.81745067e-01
9.24980223e-01 -1.10922202e-01 -4.94989544e-01 -5.41516161e-03
-3.71196985e-01 3.56241137e-01 7.64238477e-01 4.41798449e-01
-9.13423419e-01 5.80929816e-01 6.52832460e+00 7.82326758e-01
-1.27737558e+00 3.12516809e-01 4.32196021e-01 1.65594369e-01
-1.08006643e-03 -4.05134298e-02 -6.22026265e-01 8.05271804e-01
6.91341579e-01 1.25533596e-01 8.20008934e-01 5.18047154e-01
4.10140127e-01 -1.95046544e-01 -8.61760259e-01 1.02397370e+00
-1.53364679e-02 -1.40880191e+00 -4.92271855e-02 2.38866642e-01
9.64329541e-01 -9.35193673e-02 7.66024962e-02 -4.86770451e-01
1.05002187e-01 -1.03018832e+00 4.23019916e-01 6.93855107e-01
9.06292558e-01 -6.34972930e-01 9.64974403e-01 2.46522963e-01
-7.05453873e-01 9.21644643e-02 -4.95481849e-01 1.48271009e-01
3.04761380e-01 6.46392405e-01 -3.99160057e-01 5.57650030e-01
6.44075871e-01 5.07462800e-01 -2.69571334e-01 1.20682716e+00
-1.63357615e-01 4.53509957e-01 -1.70905679e-01 1.62219942e-01
2.40138426e-01 -3.94819081e-01 7.17787743e-01 1.05502427e+00
4.68778133e-01 -2.43287534e-02 -3.33600521e-01 1.24714518e+00
-1.89352527e-01 -2.26689219e-01 -4.65873539e-01 -7.75946900e-02
3.46719921e-02 1.46290612e+00 -7.30485260e-01 -9.40598994e-02
-3.22096378e-01 1.15080655e+00 2.07033843e-01 6.03158057e-01
-8.37825418e-01 -5.07069111e-01 5.70531249e-01 9.43604633e-02
5.69900811e-01 -1.36915982e-01 6.96337074e-02 -1.24072266e+00
-1.59528568e-01 -5.66904485e-01 -1.86256707e-01 -9.99116421e-01
-1.22840059e+00 5.66016853e-01 -4.59925503e-01 -1.21183944e+00
2.09743921e-02 -7.78672397e-01 -5.34701288e-01 1.17706859e+00
-2.22887611e+00 -1.18860543e+00 -4.78579998e-01 7.41346061e-01
1.56959280e-01 -1.14620896e-02 7.27241933e-01 4.37228382e-01
-6.71575904e-01 1.47261888e-01 5.98782063e-01 6.21120408e-02
7.16560006e-01 -1.38042295e+00 -2.03460768e-01 1.24054146e+00
-4.88051891e-01 7.74817467e-01 7.99252272e-01 -3.29606444e-01
-1.30192697e+00 -1.24077737e+00 5.68319619e-01 4.35576141e-02
5.89618504e-01 1.01729289e-01 -9.59247589e-01 5.36265850e-01
6.67733848e-01 4.05700266e-01 4.18913871e-01 -1.05084777e+00
-2.33487207e-02 -1.77657977e-01 -1.45759904e+00 2.49498427e-01
6.14019454e-01 -5.06713748e-01 -1.88578203e-01 2.69761592e-01
5.88376403e-01 -3.55451643e-01 -7.70322680e-01 1.92057952e-01
3.13498080e-01 -1.03181410e+00 8.33929420e-01 -2.58977294e-01
5.23251534e-01 -8.54688108e-01 1.20490357e-01 -1.34200168e+00
-2.35944837e-01 -8.79261792e-01 -7.63362423e-02 1.00207436e+00
-6.67909011e-02 -7.44633496e-01 6.20597363e-01 2.03508630e-01
-1.40401781e-01 -6.56713724e-01 -8.38386953e-01 -8.72445464e-01
-5.90520389e-02 2.23495722e-01 3.79050583e-01 7.95296550e-01
-4.92828965e-01 1.73474908e-01 -5.88992059e-01 2.73018211e-01
1.09332478e+00 5.84796164e-03 6.10906303e-01 -1.08158290e+00
-3.21324766e-01 -2.47882843e-01 -5.20807683e-01 -1.03992248e+00
6.00136444e-03 -7.59721160e-01 3.36258352e-01 -1.33882642e+00
2.61981159e-01 -1.81909814e-01 -1.04021892e-01 5.23137487e-02
-1.46755666e-01 4.98902410e-01 -1.62974417e-01 3.73871595e-01
-2.41919726e-01 3.75851959e-01 1.42209530e+00 -8.19678325e-03
2.02413514e-01 -1.25276536e-01 -4.74637806e-01 5.88015676e-01
7.04017758e-01 -4.91422653e-01 -2.70948559e-01 -6.01497173e-01
1.30213290e-01 -2.08754942e-01 8.08962345e-01 -1.07953489e+00
5.42799056e-01 -7.02891946e-02 3.19215804e-01 -9.62846279e-02
4.26829427e-01 -9.99478102e-01 4.28181201e-01 5.19621909e-01
-2.02763807e-02 -1.28396854e-01 -1.32539496e-01 7.59343266e-01
-2.75045931e-01 -2.86560297e-01 1.33172369e+00 -3.09253782e-01
-3.28625947e-01 3.10870171e-01 -1.91071630e-01 -1.67580768e-01
8.95929098e-01 -1.90014720e-01 -5.38784206e-01 -1.38940111e-01
-4.82181609e-01 -2.34824955e-01 7.17480898e-01 -3.32386374e-01
6.28191471e-01 -1.13140666e+00 -5.05335093e-01 1.16629928e-01
-3.19183201e-01 2.80270785e-01 1.95358351e-01 1.20109570e+00
-9.94738936e-01 3.58499527e-01 -3.83830070e-01 -5.23967981e-01
-7.91263402e-01 5.58536708e-01 7.27740943e-01 -2.84668297e-01
-3.77390862e-01 7.60154903e-01 2.76067197e-01 -1.26494378e-01
-1.40505150e-01 -2.29361460e-01 -1.93619296e-01 -5.58061659e-01
6.10862374e-01 5.94680429e-01 -1.31462976e-01 -6.24350429e-01
-1.29770324e-01 7.51854300e-01 3.72677028e-01 -4.97161299e-02
1.48581541e+00 -3.03958088e-01 -4.93861943e-01 2.41967365e-02
1.22647500e+00 -1.66554898e-01 -1.65333426e+00 -8.49973336e-02
-2.70914972e-01 -3.87632132e-01 2.22991765e-01 -4.62732077e-01
-1.31732547e+00 1.01092803e+00 6.30281329e-01 1.60516784e-01
1.14411438e+00 -4.21481997e-01 8.67415726e-01 4.21654768e-02
2.47419015e-01 -6.99109137e-01 -9.41580459e-02 2.40533262e-01
6.89800143e-01 -1.18307531e+00 -1.65831238e-01 -3.04112196e-01
-1.46363035e-01 1.10999858e+00 3.76785457e-01 -3.72161925e-01
4.41356152e-01 3.25879455e-01 5.15987687e-02 2.19485182e-02
-2.47383818e-01 -1.87372267e-01 -2.87699681e-02 6.16949379e-01
2.52997100e-01 -3.42212200e-01 -3.77266437e-01 3.06884795e-01
3.79481703e-01 3.77039254e-01 6.10481203e-01 7.40846574e-01
-2.54119754e-01 -9.17089403e-01 -3.17753404e-01 -8.57826415e-03
-8.08094323e-01 -4.35628928e-02 -8.75009298e-02 4.07181829e-01
2.37951770e-01 7.48221576e-01 -2.22456709e-01 9.67457984e-03
9.13900957e-02 -8.61655846e-02 5.97393632e-01 -3.09155136e-01
-4.84129369e-01 1.62028357e-01 -6.57220602e-01 -4.63468194e-01
-1.00605226e+00 -2.02198833e-01 -1.17630768e+00 -4.06447709e-01
-6.31374300e-01 1.83892667e-01 7.50898540e-01 9.01344061e-01
3.31332177e-01 4.79342431e-01 5.55844426e-01 -9.64538574e-01
-3.61702085e-01 -9.72149432e-01 -6.98903322e-01 2.83953607e-01
5.81873834e-01 -5.75676203e-01 -6.16676927e-01 5.33563435e-01] | [12.308128356933594, -2.6420042514801025] |
526c6a13-6859-4931-b708-7fe2abec9c6f | weakly-supervised-multi-task-learning-for | 1910.12326 | null | https://arxiv.org/abs/1910.12326v1 | https://arxiv.org/pdf/1910.12326v1.pdf | Weakly Supervised Multi-Task Learning for Cell Detection and Segmentation | Cell detection and segmentation is fundamental for all downstream analysis of digital pathology images. However, obtaining the pixel-level ground truth for single cell segmentation is extremely labor intensive. To overcome this challenge, we developed an end-to-end deep learning algorithm to perform both single cell detection and segmentation using only point labels. This is achieved through the combination of different task orientated point label encoding methods and a multi-task scheduler for training. We apply and validate our algorithm on PMS2 stained colon rectal cancer and tonsil tissue images. Compared to the state-of-the-art, our algorithm shows significant improvement in cell detection and segmentation without increasing the annotation efforts. | ['Yao Nie', 'Alireza Chamanzar'] | 2019-10-27 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [ 5.40359616e-01 2.05778956e-01 -2.32911915e-01 -1.58123016e-01
-1.40857339e+00 -6.90924287e-01 1.90504566e-01 8.89413655e-01
-7.99422204e-01 9.18237329e-01 -4.67450023e-01 -6.23680174e-01
2.04557508e-01 -5.56975782e-01 -3.75505030e-01 -1.08284593e+00
1.94737688e-01 7.14568019e-01 3.65370184e-01 1.05910726e-01
2.59327918e-01 5.69532394e-01 -5.70233643e-01 3.15523118e-01
8.14386487e-01 6.98138475e-01 2.76706666e-01 1.03359056e+00
-1.04499586e-01 3.99307728e-01 -3.18786919e-01 -8.94091874e-02
1.34251593e-02 -2.74781466e-01 -1.04178405e+00 9.15347710e-02
3.70857358e-01 -1.02853902e-01 2.21542478e-01 1.02618337e+00
7.76186585e-01 -3.86482060e-01 7.83484519e-01 -7.41499007e-01
1.38006493e-01 2.06488341e-01 -9.78093147e-01 4.42968249e-01
-4.01158005e-01 6.32145405e-02 7.56998122e-01 -3.90545577e-01
8.74781489e-01 4.25385952e-01 7.03538001e-01 5.88137269e-01
-1.46351445e+00 -4.16021824e-01 -5.04731655e-01 -4.36369926e-01
-1.56575334e+00 -2.52086699e-01 2.24240392e-01 -4.68250662e-01
9.49799716e-01 -5.33313602e-02 5.97149014e-01 5.90427876e-01
5.23759425e-01 1.00642395e+00 8.92405748e-01 -4.78089392e-01
2.57480979e-01 -9.38226730e-02 1.09851412e-01 7.57576108e-01
3.38317126e-01 -4.01535541e-01 -6.77582771e-02 1.54496521e-01
9.04214680e-01 1.69559404e-01 4.11073230e-02 8.22328776e-03
-1.23741698e+00 6.53252244e-01 3.94214302e-01 4.17860299e-01
-1.02785617e-01 4.79341656e-01 6.60252035e-01 -2.33303592e-01
5.14039993e-01 2.39193648e-01 -2.19360501e-01 -9.34724957e-02
-1.42579699e+00 -9.26941708e-02 4.89470810e-01 4.44604158e-01
4.46485996e-01 -5.29838860e-01 -4.24624711e-01 4.07921135e-01
2.24038467e-01 3.16131115e-02 5.07004142e-01 -8.07001591e-01
-2.30470210e-01 8.92684698e-01 -9.60019454e-02 -4.04150456e-01
-1.01773763e+00 -9.16083634e-01 -8.68342698e-01 3.28056306e-01
9.37905133e-01 -1.82553113e-01 -1.34866941e+00 1.24078000e+00
4.43487674e-01 1.91156089e-01 -2.26583958e-01 8.86553526e-01
4.32066053e-01 1.92072451e-01 3.14282835e-01 -8.06024596e-02
1.62868643e+00 -8.42798412e-01 -7.80719280e-01 -9.85210761e-02
1.52018857e+00 -5.48278391e-01 5.70453584e-01 2.27641255e-01
-9.09757257e-01 -6.95138648e-02 -8.24336052e-01 -4.18325275e-01
-4.05977786e-01 4.07617003e-01 7.87193656e-01 6.82731628e-01
-1.12119687e+00 2.01732606e-01 -1.23762941e+00 -4.74904209e-01
1.10118532e+00 7.62743235e-01 -3.00896794e-01 1.94389239e-01
-4.14306998e-01 6.05147839e-01 2.92036146e-01 -1.50551513e-01
-6.82430208e-01 -8.14020336e-01 -5.61024487e-01 -9.85504389e-02
9.83492807e-02 -8.25730860e-01 1.44494081e+00 -5.74042380e-01
-1.36989141e+00 1.48268342e+00 -4.64350075e-01 -5.51218510e-01
5.63380480e-01 3.71591896e-01 2.81855822e-01 3.81527156e-01
2.46766225e-01 1.07462382e+00 1.28114626e-01 -9.03576314e-01
-1.09506226e+00 -3.98686945e-01 -4.07793522e-01 1.85160846e-01
6.38646036e-02 -1.58371449e-01 -7.66148806e-01 -3.20445329e-01
4.67118025e-02 -7.43493915e-01 -6.68458760e-01 3.82596314e-01
-3.63327533e-01 3.35451178e-02 8.54981363e-01 -5.87032199e-01
6.38553679e-01 -2.20581818e+00 -6.49686083e-02 2.10926890e-01
4.42570865e-01 1.66407228e-01 2.76511073e-01 -2.47144580e-01
3.50177377e-01 3.10554296e-01 -2.28397205e-01 -6.22684896e-01
-2.49306634e-01 -1.29420295e-01 2.25324720e-01 9.61743712e-01
1.22476235e-01 1.04214525e+00 -1.05162632e+00 -1.09785914e+00
1.80688769e-01 3.47116500e-01 -4.56120491e-01 -5.40335886e-02
-1.89932331e-01 7.52809882e-01 -2.05995962e-01 1.05144560e+00
5.47657669e-01 -6.09487593e-01 3.18948954e-01 -5.77428006e-03
7.23751858e-02 -1.25548378e-01 -6.92360938e-01 1.76963270e+00
-1.09981693e-01 5.83749771e-01 4.60708052e-01 -8.25485468e-01
5.27378559e-01 3.89601350e-01 8.16460013e-01 -6.30157411e-01
3.62419069e-01 5.22413790e-01 2.66234912e-02 -2.26001889e-01
3.23869854e-01 -4.26230818e-01 -1.86987981e-01 2.91331887e-01
7.26716891e-02 -2.53453612e-01 3.72880936e-01 1.10434048e-01
1.18973458e+00 -2.42598131e-02 3.87196660e-01 -4.26284671e-01
4.92014199e-01 4.54543144e-01 7.69385815e-01 5.52092075e-01
-5.64743817e-01 7.53399312e-01 8.46725762e-01 -3.05004627e-01
-8.51632476e-01 -9.19164777e-01 -3.72189224e-01 9.82475877e-01
3.79074141e-02 1.11299723e-01 -7.38066137e-01 -7.53636241e-01
-3.83270197e-02 2.26488903e-01 -6.73778713e-01 3.54763597e-01
-6.32317960e-01 -1.18412781e+00 9.58830178e-01 4.66149539e-01
3.46816182e-01 -5.86087108e-01 -7.50275016e-01 4.34135765e-01
-9.15064737e-02 -1.09454954e+00 -2.96820402e-01 5.47116101e-01
-8.46017420e-01 -1.09844232e+00 -9.91563618e-01 -1.04879761e+00
1.06577134e+00 1.26658995e-02 9.12095010e-01 1.84248909e-01
-7.67726898e-01 -3.01569968e-01 -5.86275905e-02 -6.34150803e-01
-3.08788329e-01 4.96241212e-01 -4.90796000e-01 -2.50636071e-01
4.27711934e-01 3.19807325e-03 -8.45176756e-01 5.60554788e-02
-8.20276320e-01 3.16435337e-01 8.57410252e-01 1.09470487e+00
1.26569748e+00 -3.21937688e-02 5.21158040e-01 -1.16452122e+00
2.83888549e-01 -1.86811283e-01 -7.09730685e-01 9.94812697e-03
-1.90523162e-01 -2.28356540e-01 3.74492198e-01 7.66263828e-02
-7.07181454e-01 6.62451088e-01 -3.52264166e-01 1.30336449e-01
-2.14626685e-01 4.51284289e-01 4.58612591e-01 -3.94807309e-01
5.74648619e-01 8.06061029e-02 3.97477746e-01 -9.68703628e-03
8.59508011e-03 6.06496453e-01 7.70536840e-01 -1.07519679e-01
3.95398252e-02 9.89076316e-01 5.22987425e-01 -6.88013136e-01
-7.44765639e-01 -8.43865037e-01 -6.16334677e-01 -2.31375679e-01
1.14019847e+00 -8.63909125e-01 -8.68704796e-01 7.59079516e-01
-9.77115691e-01 -7.77574480e-01 -2.66821623e-01 3.38534623e-01
-6.80685163e-01 1.00118935e-01 -1.09820294e+00 -4.80977535e-01
-5.99464655e-01 -1.24145961e+00 1.72686589e+00 3.32137734e-01
-1.86127901e-01 -1.19980311e+00 2.46008739e-01 3.02713096e-01
3.53820413e-01 6.20622396e-01 8.17677498e-01 -6.34781122e-01
-5.65480769e-01 -5.03830552e-01 -3.37009460e-01 -4.20370430e-01
1.10674784e-01 1.30256698e-01 -8.17458391e-01 -4.31045085e-01
-4.90084916e-01 -2.93408424e-01 1.02940106e+00 8.81859541e-01
1.16933715e+00 2.84636676e-01 -1.24798822e+00 1.00078714e+00
1.64725411e+00 -3.24479342e-02 3.82317603e-01 3.83051753e-01
6.22070193e-01 3.11707735e-01 5.06855071e-01 1.09335639e-01
3.01465720e-01 3.75488043e-01 3.36524516e-01 -8.35778177e-01
-1.96041316e-01 1.79352701e-01 -4.06571180e-01 9.29957107e-02
2.39794910e-01 -5.31664193e-01 -1.45826495e+00 1.10141122e+00
-1.70051670e+00 -5.84826827e-01 -3.02641630e-01 1.60834765e+00
1.03984416e+00 2.28249654e-01 1.65337641e-02 2.24821448e-01
5.92776537e-01 -3.71543109e-01 -6.41335368e-01 -1.90920964e-01
9.33991671e-02 2.66020298e-01 7.36842811e-01 3.89884084e-01
-1.22455192e+00 9.40629184e-01 6.95250034e+00 1.14155161e+00
-1.38714218e+00 1.47274792e-01 1.29931331e+00 -1.99627429e-01
2.12200806e-01 -2.15578273e-01 -9.17981386e-01 2.78997153e-01
5.56087375e-01 1.27111241e-01 -2.33762845e-01 3.14928740e-01
1.81794360e-01 -5.51276982e-01 -9.88353193e-01 8.41652095e-01
-3.00549418e-01 -1.69850731e+00 -4.17461157e-01 5.23144603e-01
7.31538415e-01 1.83389157e-01 -9.55501273e-02 3.38439308e-02
3.21707189e-01 -1.18853164e+00 1.47945181e-01 5.15502274e-01
1.04775202e+00 -6.62930131e-01 1.33988523e+00 3.84988368e-01
-8.03283453e-01 2.86021680e-01 -5.77954315e-02 2.47350574e-01
2.96871156e-01 8.29800844e-01 -1.44420791e+00 3.14010948e-01
1.80131808e-01 4.92360234e-01 -5.29401720e-01 1.28505611e+00
2.19341248e-01 5.95544398e-01 -4.81239021e-01 -9.29876231e-03
2.48742491e-01 2.40699828e-01 1.47657797e-01 1.61249101e+00
3.42951804e-01 -7.02885613e-02 1.96422487e-01 6.84260964e-01
-6.96987007e-03 -4.12377343e-02 -6.44352883e-02 -5.65476231e-02
2.15337157e-01 1.76705539e+00 -1.56639242e+00 -1.90974504e-01
-1.00495219e-01 8.54516566e-01 3.93616706e-01 4.82545607e-02
-6.42716587e-01 -4.22564834e-01 2.26625159e-01 1.48819104e-01
1.38406783e-01 -2.79266685e-01 -9.62348342e-01 -4.79491264e-01
-6.85521245e-01 -3.74437213e-01 5.58756948e-01 -8.50615278e-02
-1.05531943e+00 1.19771086e-01 -4.72776324e-01 -9.88955379e-01
-1.89895332e-02 -6.95164561e-01 -7.22310901e-01 7.85289943e-01
-1.71487451e+00 -1.25804734e+00 -3.02532583e-01 3.27728614e-02
3.27250093e-01 2.58405089e-01 8.03548932e-01 2.41164759e-01
-7.10621834e-01 7.38823295e-01 6.62935153e-02 1.64390847e-01
5.64964771e-01 -1.63511169e+00 2.06210300e-01 5.58631539e-01
-5.27617633e-01 3.15009177e-01 4.18036729e-01 -5.53571999e-01
-1.09620273e+00 -1.33724260e+00 7.65621722e-01 -3.20308387e-01
3.82147938e-01 -3.35397810e-01 -5.74127138e-01 5.43117464e-01
1.89707354e-01 2.58268625e-01 1.19220698e+00 -2.07479313e-01
6.18079066e-01 8.51697773e-02 -1.39558864e+00 5.65524638e-01
5.24273634e-01 -2.01990202e-01 8.03455636e-02 5.11764050e-01
3.31471145e-01 -1.09103072e+00 -8.66491079e-01 2.40154266e-01
2.43209407e-01 -5.34206986e-01 6.93043053e-01 6.75349012e-02
1.88269183e-01 -5.24848938e-01 3.06818277e-01 -1.01898563e+00
-4.32592988e-01 -4.10432905e-01 2.67755657e-01 9.07182038e-01
7.35880017e-01 -4.15444791e-01 1.27622092e+00 2.43377909e-01
-4.39669251e-01 -1.07723176e+00 -1.08948278e+00 -2.93252319e-01
2.69003719e-01 1.43896103e-01 1.95711136e-01 6.35987103e-01
3.09364140e-01 1.92912996e-01 3.47456992e-01 1.91309731e-02
6.38940275e-01 3.04083042e-02 5.81697941e-01 -9.82104838e-01
3.34213227e-02 -7.58362174e-01 -5.66203833e-01 -7.73066163e-01
6.27685264e-02 -1.07110691e+00 1.38464227e-01 -1.77823305e+00
3.45008284e-01 -4.13898945e-01 -3.44583303e-01 7.30114043e-01
-3.42632920e-01 7.80444801e-01 -2.59283423e-01 1.89892665e-01
-9.33469713e-01 -8.51410255e-02 1.42530143e+00 -1.53468713e-01
-6.40968457e-02 -1.72684729e-01 -6.46211624e-01 4.53829557e-01
8.66346121e-01 -5.74466407e-01 5.78617267e-02 -3.31824601e-01
3.97209704e-01 1.13980480e-01 4.00332659e-01 -1.19172537e+00
6.04472160e-01 -3.57311852e-02 5.67914069e-01 -1.00516558e+00
1.24315076e-01 -5.23794055e-01 -1.88991711e-01 6.66197360e-01
-2.69858241e-01 -5.43966353e-01 4.31097120e-01 4.36080545e-01
-1.09770045e-01 -1.26586482e-01 9.59334195e-01 -3.03081810e-01
-2.24678665e-01 1.99782133e-01 -6.59680188e-01 -3.05849552e-01
1.42974305e+00 -5.69225311e-01 -6.73348129e-01 1.79678768e-01
-6.69834793e-01 6.02825880e-01 7.00037420e-01 -5.22342324e-01
1.96696773e-01 -8.52727115e-01 -8.60073507e-01 8.91534686e-02
1.10267745e-02 6.86359227e-01 2.91821480e-01 1.34527993e+00
-1.13737142e+00 5.72623789e-01 -1.03754781e-01 -1.03404081e+00
-1.05912662e+00 2.65175432e-01 7.73577631e-01 -8.09108913e-01
-3.41348559e-01 1.10858834e+00 5.73312454e-02 -4.23006833e-01
-3.54038477e-02 -3.61245513e-01 -2.66613960e-01 -8.76464099e-02
1.42086685e-01 1.83603436e-01 1.91802412e-01 -3.91947508e-01
-4.32566851e-01 3.15562248e-01 -4.78177190e-01 -2.25990452e-02
9.92780089e-01 -5.74171916e-02 -2.73875862e-01 2.68173099e-01
1.02254736e+00 -1.67710796e-01 -1.41079521e+00 1.66888073e-01
-1.20437751e-02 -3.62350233e-02 4.81891125e-01 -8.66975904e-01
-1.06558895e+00 7.34392762e-01 7.39830077e-01 5.28181680e-02
1.05186236e+00 -1.05024382e-01 9.92298663e-01 3.48907933e-02
3.54070783e-01 -1.14868009e+00 -2.73982257e-01 4.26541120e-01
2.46077217e-02 -1.31625295e+00 1.05136171e-01 -5.88583410e-01
-2.88192898e-01 1.00334096e+00 4.99877244e-01 -6.43619969e-02
4.33434963e-01 6.38888597e-01 3.57810169e-01 -3.89602929e-01
-7.50753403e-01 -2.95563966e-01 -1.98291183e-01 5.31774819e-01
7.63212025e-01 1.50373429e-01 -4.13020194e-01 4.26270455e-01
1.46507844e-01 4.88939315e-01 4.54781592e-01 9.93960142e-01
-4.61232424e-01 -9.67895746e-01 1.89828163e-03 6.50242865e-01
-1.08546424e+00 7.87204951e-02 -2.45468140e-01 5.64504027e-01
2.52312683e-02 5.69291532e-01 3.03038120e-01 1.70622244e-01
-2.28841797e-01 -2.54420042e-01 3.27191949e-01 -7.18863666e-01
-7.01378226e-01 3.85402352e-01 -1.98591083e-01 -1.22302882e-01
-3.00088853e-01 -6.16304755e-01 -1.93959272e+00 3.46850716e-02
-2.89221019e-01 -2.25911781e-01 6.60199463e-01 8.71343851e-01
3.34933043e-01 9.15042758e-01 3.03905934e-01 -8.19387317e-01
-4.30792570e-02 -8.23541760e-01 -6.55629158e-01 9.62528139e-02
5.46445370e-01 -1.80351377e-01 -8.80593807e-02 1.98508054e-01] | [14.987380981445312, -3.0876870155334473] |
e30acfbe-4524-44f0-ae60-68d0f178d1d9 | decompose-and-realign-tackling-condition | 2306.14408 | null | https://arxiv.org/abs/2306.14408v1 | https://arxiv.org/pdf/2306.14408v1.pdf | Decompose and Realign: Tackling Condition Misalignment in Text-to-Image Diffusion Models | Text-to-image diffusion models have advanced towards more controllable generation via supporting various image conditions (e.g., depth map) beyond text. However, these models are learned based on the premise of perfect alignment between the text and image conditions. If this alignment is not satisfied, the final output could be either dominated by one condition, or ambiguity may arise, failing to meet user expectations. To address this issue, we present a training-free approach called "Decompose and Realign'' to further improve the controllability of existing models when provided with partially aligned conditions. The ``Decompose'' phase separates conditions based on pair relationships, computing scores individually for each pair. This ensures that each pair no longer has conflicting conditions. The "Realign'' phase aligns these independently calculated scores via a cross-attention mechanism to avoid new conflicts when combing them back. Both qualitative and quantitative results demonstrate the effectiveness of our approach in handling unaligned conditions, which performs favorably against recent methods and more importantly adds flexibility to the controllable image generation process. | ['Ying-Cong Chen', 'Yijun Li', 'Guibao Shen', 'Luozhou Wang'] | 2023-06-26 | null | null | null | null | ['image-generation'] | ['computer-vision'] | [ 6.64963543e-01 4.05001253e-01 -3.18130143e-02 -3.28931361e-01
-6.46516502e-01 -8.42358708e-01 1.02338350e+00 9.77852568e-02
-3.24539214e-01 7.35417545e-01 3.62314612e-01 -3.62647116e-01
-1.82183638e-01 -5.38645327e-01 -6.08351886e-01 -7.48876333e-01
2.76936710e-01 4.41072822e-01 2.54086763e-01 -1.52685434e-01
4.16960120e-01 2.69107729e-01 -1.66137016e+00 4.52517837e-01
9.49514687e-01 7.13129699e-01 3.36151838e-01 7.05886483e-01
-6.02815114e-02 7.33668566e-01 -5.43384433e-01 -3.90824229e-01
2.85011142e-01 -4.20212716e-01 -7.75040925e-01 4.14516151e-01
5.73378861e-01 -4.75553185e-01 2.39653513e-01 1.05788791e+00
4.29249197e-01 -2.21207842e-01 5.55124760e-01 -1.29346061e+00
-6.98172867e-01 5.56400597e-01 -5.20476460e-01 -2.91633029e-02
4.30364788e-01 4.71815228e-01 1.20100188e+00 -7.46277630e-01
1.02627742e+00 1.03048003e+00 2.09941462e-01 4.53308970e-01
-1.50350654e+00 -4.63394284e-01 6.04702055e-01 -8.32042545e-02
-1.06180418e+00 -6.60022616e-01 6.22596920e-01 -6.00603402e-01
1.06739867e+00 3.67785811e-01 7.26555824e-01 1.01616430e+00
1.27094954e-01 6.04856431e-01 1.16725528e+00 -4.82295543e-01
1.86882585e-01 2.54938751e-01 -4.53619987e-01 3.91668022e-01
2.46063158e-01 2.11883396e-01 -5.59540331e-01 1.98086441e-01
7.92351007e-01 -4.90456253e-01 -4.37955022e-01 -5.78989089e-01
-1.47054613e+00 5.65354705e-01 3.60806644e-01 3.81789237e-01
-4.67539370e-01 -1.40720025e-01 6.41304553e-02 3.91215295e-01
3.16517174e-01 9.31484580e-01 -2.29660720e-01 -1.77410603e-01
-1.11043322e+00 3.90390158e-01 5.68158150e-01 9.70591187e-01
7.29857802e-01 -1.31238461e-01 -4.33399409e-01 6.46378219e-01
1.75093576e-01 2.01183259e-01 3.03258985e-01 -1.05705214e+00
5.66300213e-01 6.89238667e-01 5.02943337e-01 -1.02945697e+00
-2.79370546e-01 -5.29433131e-01 -4.84880954e-01 5.17328918e-01
5.07883608e-01 -1.33298352e-01 -1.14185715e+00 2.16013670e+00
2.35022932e-01 -1.33082092e-01 -4.49935012e-02 1.03467226e+00
3.40260088e-01 5.48041046e-01 6.54097944e-02 -3.37244779e-01
1.11326993e+00 -9.53931391e-01 -7.31398761e-01 -4.16739553e-01
4.62158233e-01 -8.63039792e-01 1.10626507e+00 4.16045934e-01
-1.40301287e+00 -3.50429416e-01 -1.17951727e+00 3.18400242e-04
-1.61026224e-01 -6.61378130e-02 3.82830530e-01 5.37711442e-01
-1.43181896e+00 5.63795030e-01 -5.58847427e-01 -1.79257035e-01
2.44049385e-01 3.58180970e-01 -3.01292390e-01 -7.87720457e-02
-1.12682283e+00 9.69240248e-01 5.31362817e-02 1.02033660e-01
-6.17342353e-01 -5.33228695e-01 -6.83567584e-01 1.58316508e-01
5.05005658e-01 -8.32376480e-01 1.29835236e+00 -1.38698900e+00
-1.34920681e+00 8.09572101e-01 -1.41443238e-01 -1.65742457e-01
9.34116542e-01 -1.67962015e-01 -9.54690203e-02 8.30067843e-02
2.56096333e-01 1.22317433e+00 9.21629429e-01 -1.74730229e+00
-6.32303774e-01 -1.14487313e-01 2.25620940e-01 5.38598776e-01
-2.42558286e-01 -8.34827423e-02 -7.73654521e-01 -4.86073405e-01
1.20595172e-01 -1.00347817e+00 -2.89416492e-01 1.77322313e-01
-5.50025761e-01 8.75523090e-02 4.76229727e-01 -2.66003489e-01
1.41878283e+00 -1.90354037e+00 4.23013300e-01 3.81596148e-01
2.54845589e-01 4.35610116e-02 -3.10396194e-01 4.71285731e-01
-2.59018838e-01 3.30915362e-01 -2.15801835e-01 -5.64272821e-01
-5.94952591e-02 2.13038906e-01 -6.99282959e-02 1.30769923e-01
6.19783759e-01 7.12201297e-01 -8.87549400e-01 -5.72228849e-01
1.25409111e-01 3.01372737e-01 -8.37054729e-01 1.18228905e-01
-5.01973748e-01 4.92739826e-01 -2.30024368e-01 3.87384593e-01
5.93444347e-01 -3.24073046e-01 2.74779946e-01 -2.21825287e-01
-4.93836373e-01 2.52837509e-01 -1.26501238e+00 1.50513816e+00
-4.04708087e-01 4.61937279e-01 1.86060771e-01 -5.90833783e-01
6.26023352e-01 2.79241264e-01 4.86143947e-01 -7.12896109e-01
1.23605318e-01 9.70614403e-02 1.83206543e-01 -5.25196791e-01
5.50107598e-01 -1.16492629e-01 2.52367109e-01 6.03849471e-01
-1.45489722e-01 -3.76201600e-01 4.74606752e-01 4.11552995e-01
8.87436986e-01 5.42955220e-01 1.88510984e-01 -1.54007331e-01
2.67077744e-01 1.08475707e-01 4.47365135e-01 9.01968658e-01
-2.17261791e-01 9.68366265e-01 9.03732300e-01 -8.76986701e-03
-1.12197363e+00 -1.04613125e+00 2.07944289e-01 8.02463531e-01
1.72117621e-01 -4.08577323e-01 -7.70819068e-01 -5.76295555e-01
-2.36728549e-01 8.24668109e-01 -8.29600036e-01 6.73861988e-03
-4.13137287e-01 -6.21258140e-01 5.48170097e-02 4.28844422e-01
3.95348042e-01 -1.06192374e+00 -9.21267986e-01 2.28228331e-01
-3.32397461e-01 -9.30867970e-01 -4.28129673e-01 1.27351329e-01
-4.38072741e-01 -8.15042615e-01 -6.87632561e-01 -6.07305169e-01
1.14055920e+00 2.81503201e-01 1.12330246e+00 2.97193617e-01
-7.18955472e-02 2.22324848e-01 -3.07527006e-01 -1.13023609e-01
-4.49377239e-01 3.90115529e-02 -2.97735453e-01 9.35486052e-03
-1.79389358e-01 -4.43409979e-01 -8.26579332e-01 4.51599360e-01
-1.17335391e+00 6.20846629e-01 7.54965365e-01 8.49836349e-01
4.90028590e-01 -7.23536313e-02 6.09056354e-01 -8.85911167e-01
8.90048385e-01 -3.46552402e-01 -4.72105384e-01 2.88503468e-01
-9.60720658e-01 2.31479108e-01 4.02377814e-01 -6.54751241e-01
-1.18885124e+00 2.54964650e-01 4.96958829e-02 -5.24830163e-01
-1.25706628e-01 5.52065790e-01 -2.87536085e-01 4.15600955e-01
6.58185124e-01 -4.73018810e-02 3.99311595e-02 7.31160194e-02
4.36294794e-01 3.73750687e-01 4.60160464e-01 -6.85827792e-01
8.62406254e-01 4.61516559e-01 -4.20500904e-01 -2.10116237e-01
-5.93998253e-01 -1.22403346e-01 -6.71393454e-01 -5.37023067e-01
9.83077168e-01 -6.49825573e-01 -3.56916636e-01 3.79847765e-01
-1.09360135e+00 -4.95195329e-01 -2.28633955e-01 2.69165337e-01
-4.96092677e-01 1.20850675e-01 -3.66925985e-01 -7.46333897e-01
6.63966611e-02 -1.51455450e+00 1.11865377e+00 2.63276994e-01
-5.95575869e-01 -8.39418769e-01 -1.27429187e-01 2.27962211e-01
5.73115349e-01 1.49813235e-01 9.55805659e-01 -3.51687312e-01
-6.65129900e-01 -9.12351236e-02 -3.32374603e-01 4.06007888e-03
3.18097621e-01 2.13540152e-01 -9.15556788e-01 -3.03849697e-01
-1.98854595e-01 -1.13297157e-01 6.58538699e-01 1.66718677e-01
6.52890801e-01 -4.80883658e-01 -1.21671580e-01 1.00145116e-01
1.24273074e+00 1.79736003e-01 6.93929017e-01 5.80592155e-01
5.51989734e-01 8.91991675e-01 5.07787108e-01 3.33967119e-01
3.97884071e-01 6.65122569e-01 6.02869511e-01 -4.66421992e-01
-1.44713759e-01 -2.61846721e-01 1.85485870e-01 3.05924028e-01
1.14727572e-01 -6.13412380e-01 -8.41360331e-01 6.18190348e-01
-1.88578701e+00 -9.07872260e-01 -1.92834139e-02 2.33291340e+00
9.62253332e-01 5.02006412e-01 -6.97371662e-02 1.56160161e-01
7.45197594e-01 1.88451752e-01 -5.06612539e-01 -4.82207388e-01
-9.01297629e-02 -1.27943978e-01 2.42027566e-01 8.37417901e-01
-7.42687643e-01 1.03876424e+00 6.08007336e+00 4.03486609e-01
-1.27793610e+00 -2.67348647e-01 7.55625725e-01 -2.81801194e-01
-9.21233058e-01 3.12966704e-01 -3.60749066e-01 3.67283314e-01
2.61203796e-01 4.62275036e-02 3.22613180e-01 3.88849884e-01
4.51939613e-01 -3.96694481e-01 -1.15766764e+00 4.69678313e-01
1.04527511e-01 -1.26268935e+00 1.59535319e-01 1.72537491e-01
8.05175006e-01 -3.46591353e-01 2.78156906e-01 1.91935599e-02
4.43617582e-01 -9.08198774e-01 1.09259462e+00 3.21864694e-01
8.38196814e-01 -4.83833849e-01 4.37395602e-01 3.42378169e-01
-8.07051241e-01 1.05217688e-01 2.85652727e-01 -3.29758562e-02
3.83260816e-01 4.62693036e-01 -9.01765466e-01 4.97220635e-01
4.45787370e-01 2.73775071e-01 -5.70875406e-01 7.21240103e-01
-3.69294137e-01 1.76847264e-01 -2.53966421e-01 1.02638371e-01
2.73345232e-01 -2.11581558e-01 6.39492571e-01 1.04662788e+00
2.96956629e-01 -7.23937973e-02 1.18069043e-02 9.67250049e-01
2.32505783e-01 1.01516478e-01 -5.38760006e-01 2.64102146e-02
3.93690705e-01 1.12697947e+00 -9.66059923e-01 -4.18262899e-01
-2.92581707e-01 1.17850471e+00 3.44349563e-01 4.39115524e-01
-8.50783825e-01 2.84587331e-02 3.13531220e-01 2.41644472e-01
4.39069539e-01 -2.17299700e-01 -6.12301290e-01 -1.09255540e+00
9.28100124e-02 -1.00729716e+00 1.04130134e-01 -1.09058726e+00
-9.68331575e-01 6.64512694e-01 1.45451903e-01 -1.25408804e+00
-3.59051168e-01 -2.86219180e-01 -6.20864272e-01 7.19763458e-01
-1.60815561e+00 -1.19442070e+00 -1.82085916e-01 4.60332990e-01
4.55444694e-01 4.12350088e-01 5.34354687e-01 1.85155809e-01
-5.33282936e-01 5.37694693e-01 -3.58290017e-01 -4.51003075e-01
1.08180213e+00 -1.29154754e+00 2.39787757e-01 1.17094100e+00
-7.06756487e-03 7.06338763e-01 1.05437136e+00 -7.84004807e-01
-1.00516629e+00 -7.49729693e-01 1.09572935e+00 -3.00729871e-01
4.39726502e-01 -2.73738295e-01 -8.15776587e-01 4.02094632e-01
3.07328969e-01 -5.09685934e-01 5.03959358e-01 5.22128977e-02
-4.08292145e-01 2.33350620e-02 -1.04183388e+00 1.10125375e+00
1.07085383e+00 -2.08711699e-01 -2.60075510e-01 1.31156668e-01
5.43808579e-01 -5.11675954e-01 -6.61834240e-01 4.27213222e-01
5.23286998e-01 -1.21411359e+00 6.28931582e-01 -4.10200030e-01
9.45334375e-01 -4.00038868e-01 5.65699786e-02 -1.33765173e+00
-4.26540911e-01 -7.56974816e-01 3.53581429e-01 1.39699817e+00
9.41790223e-01 -5.33030808e-01 5.27742386e-01 9.65953588e-01
-2.25379452e-01 -8.00809622e-01 -8.07259202e-01 -5.21965444e-01
1.52943991e-02 -3.91456813e-01 7.18332827e-01 1.00829089e+00
1.97790995e-01 3.57915491e-01 -4.07856643e-01 6.27439395e-02
1.39233813e-01 1.68637067e-01 6.49209678e-01 -9.06658113e-01
-4.07428205e-01 -8.18626583e-01 3.42860706e-02 -1.08936393e+00
-2.29698032e-01 -5.41227818e-01 3.10934514e-01 -1.77519608e+00
6.46871626e-02 -6.06825888e-01 -1.66244164e-01 7.77224600e-01
-4.47077513e-01 2.31658742e-01 4.39837873e-01 4.12447989e-01
-4.40675825e-01 2.30918109e-01 1.48350108e+00 1.15829691e-01
-3.32202703e-01 -2.92527378e-01 -1.08035398e+00 3.01969051e-01
8.90150130e-01 -1.12597764e-01 -6.58245683e-01 -7.61980891e-01
4.48268056e-01 2.92742178e-02 7.43530169e-02 -7.51619041e-01
2.57856995e-01 -4.29280192e-01 2.87021041e-01 -5.72636165e-02
2.15854764e-01 -6.73360288e-01 2.55607694e-01 3.53138655e-01
-4.22425032e-01 2.37753451e-01 9.74970981e-02 3.21502924e-01
5.20767160e-02 2.40475405e-02 7.26227641e-01 4.54078019e-02
-4.00845557e-01 6.23797737e-02 -3.85052502e-01 -1.86220035e-01
1.07261324e+00 -6.31546855e-01 -4.66870815e-01 -6.68468177e-01
-6.82762027e-01 4.92288351e-01 8.19455326e-01 5.61571479e-01
4.16936040e-01 -1.03371537e+00 -6.10018075e-01 3.47196013e-01
1.30528167e-01 1.53891921e-01 1.73860520e-01 9.38053429e-01
-1.97446346e-01 1.90165430e-01 -1.00681387e-01 -6.55427873e-01
-1.15474248e+00 4.68991101e-01 3.04719210e-01 -4.48231727e-01
-4.35921907e-01 7.41285503e-01 3.13079476e-01 -6.93610236e-02
1.18017621e-01 -3.61449689e-01 -2.10529849e-01 2.67166853e-01
2.28693619e-01 -2.12567389e-01 1.13424346e-01 -6.07375205e-01
-2.71594673e-01 4.56300825e-01 -2.65053600e-01 -7.61121750e-01
1.14105392e+00 -3.53509158e-01 1.07324235e-01 1.33832157e-01
6.37164533e-01 1.90432727e-01 -1.80456185e+00 1.22431412e-01
-3.28944534e-01 -5.46871006e-01 -1.26649454e-01 -1.15897751e+00
-1.00519931e+00 5.23335397e-01 3.62798780e-01 2.41386563e-01
1.18019831e+00 -9.25850198e-02 3.82227600e-01 -1.23360269e-01
-8.49681944e-02 -1.12187362e+00 2.48119697e-01 1.84933856e-01
9.50152755e-01 -1.13397074e+00 -3.92102413e-02 -4.09931898e-01
-9.23998713e-01 8.14028025e-01 9.02340174e-01 1.78567827e-01
1.79526165e-01 4.53641623e-01 2.59790123e-01 1.03167165e-02
-1.11615145e+00 -1.24377586e-01 1.13038383e-01 5.31319559e-01
4.59510624e-01 -2.43223295e-01 -4.82080430e-01 4.89424169e-02
-3.59325737e-01 -9.33950320e-02 4.28652376e-01 1.03010702e+00
-3.48121524e-01 -1.25093174e+00 -4.12592202e-01 2.92220712e-01
-2.09890559e-01 -1.60904855e-01 -5.86353660e-01 7.32737601e-01
2.95432240e-01 1.05347443e+00 2.11041540e-01 -2.10711718e-01
2.56258577e-01 -1.05058253e-01 4.88701046e-01 -6.01052642e-01
-6.24863327e-01 4.35537189e-01 3.03612649e-01 -3.85102570e-01
-3.07420850e-01 -7.13979721e-01 -1.07898629e+00 -2.61146370e-02
-5.67719996e-01 -7.41310716e-02 5.31650782e-01 1.01473653e+00
4.94476259e-01 7.16816485e-01 4.99648809e-01 -9.55240607e-01
-4.38299984e-01 -5.40740490e-01 1.05015084e-01 6.65982723e-01
2.39347100e-01 -6.73946798e-01 -3.55178118e-01 2.49394581e-01] | [11.336766242980957, -0.1843075007200241] |
3ed5987e-0574-4b9d-b5a8-60d8fe31c72e | semeval-2015-task-18-broad-coverage-semantic | null | null | https://aclanthology.org/S15-2153 | https://aclanthology.org/S15-2153.pdf | SemEval 2015 Task 18: Broad-Coverage Semantic Dependency Parsing | null | ["Zde{\\v{n}}ka Ure{\\v{s}}ov{\\'a}", 'Jan Haji{\\v{c}}', "Silvie Cinkov{\\'a}", 'Daniel Zeman', 'Stephan Oepen', 'Yusuke Miyao', 'Marco Kuhlmann', 'Dan Flickinger'] | 2015-06-01 | null | null | null | semeval-2015-6 | ['semantic-dependency-parsing'] | ['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] | [-7.264480113983154, 3.6558444499969482] |
9b0b81e4-a021-47b1-9935-ce800585eca5 | eth2vec-learning-contract-wide-code | 2101.02377 | null | https://arxiv.org/abs/2101.02377v2 | https://arxiv.org/pdf/2101.02377v2.pdf | Eth2Vec: Learning Contract-Wide Code Representations for Vulnerability Detection on Ethereum Smart Contracts | Ethereum smart contracts are programs that run on the Ethereum blockchain, and many smart contract vulnerabilities have been discovered in the past decade. Many security analysis tools have been created to detect such vulnerabilities, but their performance decreases drastically when codes to be analyzed are being rewritten. In this paper, we propose Eth2Vec, a machine-learning-based static analysis tool for vulnerability detection, with robustness against code rewrites in smart contracts. Existing machine-learning-based static analysis tools for vulnerability detection need features, which analysts create manually, as inputs. In contrast, Eth2Vec automatically learns features of vulnerable Ethereum Virtual Machine (EVM) bytecodes with tacit knowledge through a neural network for language processing. Therefore, Eth2Vec can detect vulnerabilities in smart contracts by comparing the code similarity between target EVM bytecodes and the EVM bytecodes it already learned. We conducted experiments with existing open databases, such as Etherscan, and our results show that Eth2Vec outperforms the existing work in terms of well-known metrics, i.e., precision, recall, and F1-score. Moreover, Eth2Vec can detect vulnerabilities even in rewritten codes. | ['Shingo Okamura', 'Jason Paul Cruz', 'Naoto Yanai', 'Nami Ashizawa'] | 2021-01-07 | null | null | null | null | ['vulnerability-detection'] | ['miscellaneous'] | [-2.71821707e-01 5.30236214e-02 -2.53093213e-01 -8.53260010e-02
-7.80882895e-01 -1.18058109e+00 6.87667310e-01 1.40627027e-01
6.31646737e-02 3.05277348e-01 3.38681430e-01 -1.28199589e+00
4.67036098e-01 -9.26653266e-01 -5.09227574e-01 -4.78773206e-01
-2.97561556e-01 1.17727421e-01 2.27853507e-01 -4.29321349e-01
4.75793540e-01 -1.04343928e-01 -1.09940886e+00 3.28440785e-01
6.49427056e-01 5.63768685e-01 -4.85783279e-01 5.37709236e-01
-3.44873220e-01 1.19157529e+00 -6.78176880e-01 -1.15981829e+00
4.48953092e-01 -3.93409505e-02 -7.89119422e-01 -1.12725878e+00
3.99584919e-02 -2.75492430e-01 -6.88959241e-01 1.66897905e+00
6.54402971e-02 -7.16585219e-01 4.40531760e-01 -1.47775638e+00
-7.05930054e-01 1.34138525e+00 -5.80584526e-01 3.65050286e-02
3.83464605e-01 6.41658306e-01 1.31208825e+00 -3.79450202e-01
9.05275047e-01 1.17135191e+00 8.24057639e-01 5.23528636e-01
-8.48587036e-01 -9.90291953e-01 -5.20416796e-01 3.34973127e-01
-7.69887447e-01 -4.26347507e-03 9.43813682e-01 -7.89454162e-01
1.36861932e+00 3.59033644e-01 3.08955461e-01 1.27259350e+00
1.16517556e+00 6.52818799e-01 1.02054191e+00 -7.46044889e-02
1.09156683e-01 7.96373263e-02 4.55564082e-01 6.79787159e-01
5.12794018e-01 7.39698291e-01 2.87887812e-01 -8.48231852e-01
1.21211238e-01 -2.75297780e-02 -1.15655959e-01 -2.65416503e-01
-1.16342199e+00 1.34110677e+00 2.99281664e-02 6.91882789e-01
6.39388710e-02 4.16522920e-01 1.56741607e+00 9.95702863e-01
-3.64646837e-02 6.45274699e-01 -1.01893187e+00 -5.13343930e-01
-5.16570270e-01 3.69918384e-02 1.23589432e+00 7.32662022e-01
5.25417864e-01 6.53566793e-02 1.39668673e-01 1.21430689e-02
5.68498909e-01 5.34181058e-01 6.61199629e-01 -6.30471885e-01
6.96694970e-01 1.04494214e+00 -4.80037779e-01 -1.18992913e+00
-6.64716139e-02 -8.79267380e-02 -5.54812968e-01 5.95507026e-01
1.39477491e-01 -1.15473114e-01 -3.63844126e-01 1.26580334e+00
1.75106704e-01 -1.01416618e-01 4.17451710e-01 2.96183616e-01
7.04421997e-01 5.53311944e-01 -2.04882830e-01 -1.07416309e-01
1.25120378e+00 -7.15967000e-01 -5.80957532e-01 3.15703422e-01
1.14067423e+00 -7.55045354e-01 7.53853142e-01 3.68010283e-01
-4.61015284e-01 5.10597602e-02 -1.03064048e+00 3.60115081e-01
-7.65596747e-01 -8.95478845e-01 9.40518618e-01 1.25285995e+00
-6.66794717e-01 6.17375135e-01 -5.87600231e-01 2.14297816e-01
3.30626965e-01 3.42614055e-01 -4.54453081e-01 5.10933876e-01
-1.34706819e+00 7.80154884e-01 6.61957562e-01 -4.74736392e-01
-1.29788482e+00 -4.51264769e-01 -1.10761774e+00 2.79858679e-01
3.53262365e-01 -1.86143294e-02 1.27861273e+00 -1.11652052e+00
-1.06434584e+00 6.18807435e-01 5.97204089e-01 -5.01742125e-01
4.70248729e-01 2.98771143e-01 -8.74580085e-01 -3.54688793e-01
-2.04312429e-01 -3.20020556e-01 7.14030266e-01 -1.07965672e+00
-4.32644904e-01 -2.23697916e-01 5.41177094e-01 -9.21527147e-01
-8.78449008e-02 4.92738903e-01 2.17847571e-01 -6.36803806e-01
-4.92747009e-01 -1.02571344e+00 -2.05073655e-01 -8.32525373e-01
-2.87931919e-01 -3.64635706e-01 1.22496235e+00 -1.05126929e+00
1.29563248e+00 -2.09246302e+00 -1.49453387e-01 7.06022441e-01
5.02787590e-01 6.96794093e-01 -4.24558997e-01 6.19814157e-01
-1.95813283e-01 4.63388532e-01 -3.67633581e-01 4.92955893e-01
3.88660163e-01 -9.77933779e-02 -6.71361685e-01 6.05762899e-01
-5.64662635e-01 1.39447820e+00 -1.06411004e+00 -4.15170252e-01
5.75250238e-02 3.34122449e-01 -5.79694152e-01 1.46691337e-01
-5.81186116e-01 2.07345620e-01 -4.33579862e-01 1.00904918e+00
6.23103499e-01 -2.54981760e-02 5.63238680e-01 1.34116158e-01
-2.88982857e-02 3.88261341e-02 -5.22787631e-01 1.37982476e+00
-6.90167546e-01 8.47725093e-01 -1.47654012e-01 -6.41183078e-01
8.04671168e-01 6.42860770e-01 5.10644853e-01 -7.23934889e-01
6.89553797e-01 4.06316847e-01 1.23698279e-01 -6.47118390e-01
1.57970741e-01 4.05896932e-01 -5.92338204e-01 1.06408918e+00
-2.17814177e-01 1.76278815e-01 7.64069334e-02 2.77532876e-01
1.63479459e+00 -1.05318781e-02 4.63603109e-01 -1.66865423e-01
9.72041786e-01 -2.05573961e-01 7.30057061e-01 5.43535650e-01
-2.57405221e-01 -3.09171140e-01 1.08335757e+00 -1.13039887e+00
-1.00059199e+00 -6.48709536e-01 6.12444952e-02 7.15891600e-01
-1.90825254e-01 -8.19462538e-01 -9.30426538e-01 -1.72512591e+00
3.23039949e-01 5.73761582e-01 -6.25011146e-01 -2.29515418e-01
-7.63567984e-01 -3.89497340e-01 1.01455355e+00 4.88576651e-01
2.67496169e-01 -1.36288810e+00 -7.40022838e-01 3.29311371e-01
6.15477413e-02 -5.02567708e-01 -5.19336224e-01 -1.66489750e-01
-4.88271266e-01 -1.97285974e+00 -1.94314700e-02 -7.15221345e-01
3.09405029e-01 -2.80857980e-01 1.07084537e+00 4.34862971e-01
-3.74789387e-01 -5.88262416e-02 -4.26873565e-01 -2.94605792e-02
-1.20283508e+00 -1.24592923e-01 -1.80294961e-01 -2.99877971e-01
7.22043097e-01 -5.50544024e-01 -2.44917512e-01 1.66452065e-01
-1.00882781e+00 -5.81510067e-01 5.79945147e-01 8.12475681e-01
-6.62452355e-02 3.77107551e-03 2.82118708e-01 -1.15379775e+00
8.53854597e-01 -6.01360142e-01 -1.36532950e+00 5.47653317e-01
-1.12964547e+00 4.12841648e-01 9.60890174e-01 -1.78280398e-01
-6.76392734e-01 -4.52851415e-01 -3.55823964e-01 -4.23111349e-01
3.69234651e-01 7.25520074e-01 -2.88380235e-01 -5.54129243e-01
4.35101807e-01 1.55181944e-01 -1.03793755e-01 -1.88485473e-01
5.00878215e-01 8.51314843e-01 3.73670191e-01 -3.89705032e-01
1.25160217e+00 4.32018451e-02 -1.16811186e-01 6.47054687e-02
1.69267297e-01 -8.97100288e-03 -8.63944516e-02 1.18835479e-01
4.40132499e-01 -2.60682195e-01 -1.41788292e+00 5.35648346e-01
-1.37850583e+00 -5.72246909e-02 1.42881826e-01 1.65361702e-01
-3.69977474e-01 9.11368072e-01 -7.81863987e-01 -4.60508943e-01
-9.33034599e-01 -1.58548689e+00 5.08242249e-01 -3.08384657e-01
-5.06524563e-01 -1.03413188e+00 8.17201197e-01 3.23167264e-01
5.76732814e-01 5.97925067e-01 1.57070577e+00 -1.01248705e+00
-6.44186258e-01 -4.36267287e-01 -6.59947768e-02 3.75081867e-01
9.27115455e-02 1.76261619e-01 -8.43097150e-01 -7.44445324e-01
6.97196275e-02 4.90815677e-02 7.03792691e-01 -2.06531450e-01
1.10303020e+00 -9.30432558e-01 -3.59049410e-01 9.89182889e-01
1.54545164e+00 7.72850573e-01 8.11213732e-01 6.84364855e-01
5.22919118e-01 2.55008250e-01 1.37634501e-01 2.89769024e-01
1.31513163e-01 2.79973805e-01 9.02505577e-01 4.37687725e-01
5.15248537e-01 -1.88583881e-01 7.23388374e-01 8.00000012e-01
-1.14074320e-01 2.48658314e-01 -1.26266217e+00 3.10990036e-01
-1.67339063e+00 -1.22607601e+00 -8.25826451e-02 1.81482446e+00
6.83917463e-01 6.55339286e-02 -9.73713994e-02 2.97550321e-01
3.42448503e-01 3.18566173e-01 -3.30984265e-01 -8.92543375e-01
1.51506186e-01 3.23054969e-01 7.87092626e-01 5.01820385e-01
-1.16771638e+00 8.61456275e-01 5.69677830e+00 5.44122159e-01
-1.04144824e+00 4.46412951e-01 7.37766027e-02 2.87811399e-01
-1.09860229e+00 5.35302103e-01 -1.64627597e-01 8.65295410e-01
9.72667754e-01 -5.66510439e-01 6.55956566e-01 1.05386174e+00
-5.18113077e-01 6.55820191e-01 -1.06493878e+00 9.63584483e-01
-6.55841827e-02 -1.54172075e+00 -2.65509486e-01 3.26413810e-01
7.96462417e-01 1.43270105e-01 7.57536814e-02 7.76221097e-01
8.80377889e-01 -9.58549380e-01 5.00852764e-01 2.26634159e-03
7.50927150e-01 -1.04346418e+00 1.44983530e+00 -3.13286692e-01
-1.05585289e+00 -5.95298588e-01 -1.18252719e-02 5.04055202e-01
-2.85276771e-01 4.06489760e-01 -6.40875220e-01 6.57148540e-01
5.55735052e-01 5.06586134e-01 -6.60157859e-01 5.52674413e-01
-4.63656276e-01 5.80342352e-01 4.28186774e-01 -1.51806921e-01
3.88653606e-01 -1.11508995e-01 6.22514844e-01 1.16057169e+00
2.79328018e-01 -4.01657104e-01 -1.59021646e-01 9.12354767e-01
-5.57644367e-02 8.12331885e-02 -1.00941169e+00 -4.06962752e-01
2.99076468e-01 1.10135007e+00 -3.83069158e-01 -2.86510408e-01
-9.85730052e-01 4.10082698e-01 -1.43817306e-01 2.09950030e-01
-1.04537678e+00 -1.03540957e+00 8.70181084e-01 -4.18611318e-01
2.83509672e-01 6.42547160e-02 -7.37116784e-02 -1.45401597e+00
2.26146355e-02 -1.56785870e+00 5.98382354e-01 -1.01957105e-01
-9.72488940e-01 7.67193556e-01 -4.27515656e-01 -1.15587902e+00
-3.11096042e-01 -5.72952330e-01 -1.08621013e+00 3.71157795e-01
-1.25879741e+00 -1.17011952e+00 3.27125847e-01 6.05110049e-01
1.55598551e-01 -9.86051440e-01 9.37176824e-01 8.68231356e-02
-2.96605408e-01 9.26020086e-01 4.38532561e-01 7.89980650e-01
2.60654628e-01 -1.04101920e+00 1.15264916e+00 7.67035425e-01
9.42133591e-02 9.56016004e-01 6.21653020e-01 -1.06472278e+00
-1.97169089e+00 -9.55873549e-01 9.90608752e-01 -4.90022391e-01
1.16459799e+00 -3.15015018e-01 -7.68140197e-01 1.02575469e+00
4.52559739e-01 -2.47812539e-01 6.75984859e-01 8.05725828e-02
-1.31618643e+00 -2.39317734e-02 -1.25370455e+00 4.21712965e-01
6.36887014e-01 -1.06003249e+00 -8.26452076e-01 1.49912909e-01
7.88298249e-01 1.41742732e-02 -9.58049834e-01 3.37355077e-01
9.86079156e-01 -9.81734872e-01 6.94998741e-01 -8.66232812e-01
4.01911259e-01 -2.03675359e-01 -2.84835458e-01 -1.15886629e+00
-1.10861221e-02 -1.05834365e+00 -5.95220327e-01 1.13116157e+00
3.55694830e-01 -1.12171519e+00 4.80358422e-01 2.03893945e-01
2.46380940e-01 -4.73706096e-01 -1.10079777e+00 -1.13780093e+00
4.18875217e-01 -4.01984155e-01 1.53297329e+00 1.42704177e+00
6.83689833e-01 -3.34925473e-01 -6.74029812e-02 -9.10127833e-02
6.91181242e-01 5.62345147e-01 8.07602465e-01 -1.19162786e+00
-3.85760128e-01 -6.68895662e-01 -8.29834580e-01 7.98815712e-02
7.27434397e-01 -1.34989059e+00 -5.54031253e-01 -7.86916912e-01
3.70984674e-01 -1.94591638e-02 -5.49131691e-01 6.25969291e-01
2.11044937e-01 -3.03813994e-01 9.57033113e-02 1.29844740e-01
-9.55652595e-02 1.82275489e-01 7.05187380e-01 -1.10009515e+00
9.48105752e-02 -1.70887187e-01 -4.29856718e-01 6.60296559e-01
8.35445940e-01 -7.28318393e-01 -1.89219788e-02 -2.09072724e-01
6.85273349e-01 2.34682918e-01 1.57184854e-01 -5.71239531e-01
-9.49391276e-02 -3.60365361e-01 -3.89105260e-01 -3.98671746e-01
-6.85405791e-01 -9.91707206e-01 4.00125653e-01 1.06042182e+00
-1.84623316e-01 4.69883204e-01 -6.01228327e-02 3.67259592e-01
-1.59793422e-01 -4.32534695e-01 4.26416755e-01 -3.58098716e-01
-7.65258193e-01 4.51956332e-01 -5.50974488e-01 -9.27936360e-02
1.32926869e+00 1.53416738e-01 -7.11933672e-01 -2.06947073e-01
-7.31847510e-02 1.10530846e-01 8.66624415e-01 8.95333469e-01
5.89046240e-01 -1.29532266e+00 -6.35307491e-01 5.03692269e-01
2.88744450e-01 -9.50946271e-01 -2.11850122e-01 6.97095037e-01
-7.56160915e-01 5.85393190e-01 -3.86437595e-01 -2.79955566e-02
-1.53196073e+00 1.16260612e+00 1.22261547e-01 -6.00734115e-01
-5.45979738e-01 2.59150445e-01 -5.92402555e-02 -7.70512342e-01
1.65745154e-01 -3.94579619e-01 -2.12951273e-01 -7.23605901e-02
6.76980853e-01 4.09212828e-01 4.02661860e-02 -6.26313150e-01
-4.95391101e-01 6.17095470e-01 -2.05071658e-01 1.41339213e-01
1.26756012e+00 8.38350713e-01 -8.45788896e-01 -3.71471882e-01
1.60029829e+00 3.18941653e-01 -4.54659730e-01 8.93194303e-02
4.00941759e-01 -6.98445082e-01 -2.23957881e-01 -6.78224921e-01
-1.54740167e+00 7.68436491e-01 5.47038674e-01 2.58133471e-01
5.24869025e-01 1.47800129e-02 1.28454471e+00 6.03626251e-01
8.65019739e-01 -7.75409102e-01 -2.64926761e-01 9.13882315e-01
5.19969702e-01 -1.06768012e+00 -5.55023074e-01 2.72229791e-01
-4.28894043e-01 1.28855181e+00 3.68692756e-01 1.63921922e-01
6.05555952e-01 5.28798103e-01 3.85550082e-01 -4.27497774e-01
-5.65317929e-01 4.96478975e-01 -2.44643509e-01 6.07558846e-01
1.73027411e-01 3.32975358e-01 -4.28990960e-01 7.09163964e-01
-2.91542649e-01 -2.30120450e-01 5.70480585e-01 9.93096471e-01
-2.46813536e-01 -1.55798173e+00 -3.67480308e-01 3.24519813e-01
-7.63730109e-01 -2.25400880e-01 -4.74068075e-01 6.58706546e-01
2.54209079e-02 9.12259877e-01 -4.91023630e-01 -9.87629116e-01
5.15904203e-02 6.13659658e-02 3.73578742e-02 -1.54302940e-01
-1.22354555e+00 -6.30789578e-01 3.37357819e-02 -9.48340595e-01
1.82467788e-01 -5.75153649e-01 -1.03310430e+00 -7.16672957e-01
-4.16576564e-01 4.96933609e-01 6.09380841e-01 6.07854605e-01
5.35022393e-02 1.60577044e-01 1.03549790e+00 8.73699635e-02
-9.97919679e-01 -6.82718992e-01 -3.55854481e-01 6.24318957e-01
4.75496382e-01 -2.31447026e-01 -6.44886196e-01 -1.50812656e-01] | [6.862297058105469, 7.3234663009643555] |
a6064a9d-73ab-4025-8cbd-2d1b98f47e82 | deep-learning-based-phase-reconstruction-for | 1811.09010 | null | http://arxiv.org/abs/1811.09010v1 | http://arxiv.org/pdf/1811.09010v1.pdf | Deep Learning Based Phase Reconstruction for Speaker Separation: A Trigonometric Perspective | This study investigates phase reconstruction for deep learning based monaural
talker-independent speaker separation in the short-time Fourier transform
(STFT) domain. The key observation is that, for a mixture of two sources, with
their magnitudes accurately estimated and under a geometric constraint, the
absolute phase difference between each source and the mixture can be uniquely
determined; in addition, the source phases at each time-frequency (T-F) unit
can be narrowed down to only two candidates. To pick the right candidate, we
propose three algorithms based on iterative phase reconstruction, group delay
estimation, and phase-difference sign prediction. State-of-the-art results are
obtained on the publicly available wsj0-2mix and 3mix corpus. | ['Zhong-Qiu Wang', 'Ke Tan', 'DeLiang Wang'] | 2018-11-22 | null | null | null | null | ['speaker-separation'] | ['speech'] | [-6.41693324e-02 -3.64675701e-01 -6.08194433e-02 -1.32199094e-01
-1.41684651e+00 -5.87875366e-01 2.08134741e-01 -3.99495754e-03
-2.71624386e-01 5.34109890e-01 2.78755903e-01 -3.56083177e-02
-4.19335127e-01 -4.54358459e-02 -3.32109660e-01 -1.23621750e+00
-3.97882909e-01 3.34579170e-01 1.81453358e-02 -1.31758571e-01
-1.01051159e-01 5.78245521e-02 -1.51511967e+00 -1.60477564e-01
9.37118053e-01 1.14138365e+00 1.37695283e-01 8.22329640e-01
1.25334278e-01 4.03414011e-01 -9.41527724e-01 -9.46388990e-02
3.03511947e-01 -8.05535734e-01 -3.30837637e-01 -4.29006308e-01
3.75490725e-01 -3.51138636e-02 -3.56630504e-01 8.86404872e-01
1.05092335e+00 1.01269394e-01 7.38296032e-01 -1.10011160e+00
2.25225538e-02 9.43442941e-01 -5.71091413e-01 1.92181751e-01
3.01940680e-01 -7.54766911e-02 9.57696378e-01 -5.98316133e-01
1.60307094e-01 5.61857700e-01 9.70921099e-01 2.10428253e-01
-1.10116601e+00 -9.25524056e-01 -1.64541006e-01 5.57551146e-01
-1.55659437e+00 -9.48380351e-01 1.17189884e+00 -3.78710926e-01
7.69292474e-01 2.45932326e-01 6.39335334e-01 6.28695667e-01
-1.76912591e-01 4.93224174e-01 1.09967136e+00 -7.10489810e-01
2.69999117e-01 -3.77832204e-01 1.06321968e-01 3.16847175e-01
-1.88482672e-01 3.42421740e-01 -1.22499406e+00 -6.22543693e-02
3.42501909e-01 -8.28865647e-01 -9.02637005e-01 -3.88420969e-02
-1.13780630e+00 4.37123597e-01 8.50854814e-02 4.68338221e-01
-3.73713553e-01 1.17804736e-01 2.00649239e-02 3.65155250e-01
4.92362022e-01 3.54557097e-01 -4.55382586e-01 -3.26851636e-01
-1.39032817e+00 6.14145175e-02 9.78840292e-01 4.95858759e-01
5.47424555e-01 6.26722753e-01 3.59933591e-03 8.98820102e-01
4.41579342e-01 8.94594014e-01 5.07390440e-01 -7.75679410e-01
1.28387913e-01 -7.37834692e-01 2.05422252e-01 -5.34713507e-01
-3.61888230e-01 -7.49472558e-01 -5.69037139e-01 -1.38005391e-01
8.08285654e-01 -5.32128870e-01 -8.10542762e-01 1.95578504e+00
4.61316437e-01 6.36080742e-01 5.30157387e-02 1.04937840e+00
8.18477690e-01 6.77070558e-01 -4.51940984e-01 -8.82265627e-01
1.46890175e+00 -8.17626536e-01 -1.12901139e+00 -3.58163655e-01
3.35063525e-02 -1.09334803e+00 2.43468940e-01 7.71120667e-01
-1.04856682e+00 -6.07345283e-01 -1.24342024e+00 4.46851909e-01
3.51025224e-01 -1.45812184e-01 3.30491006e-01 8.12530041e-01
-9.19241786e-01 5.89049459e-01 -7.70752490e-01 4.67494726e-01
-3.16436678e-01 1.51461482e-01 -6.08118586e-02 2.43019432e-01
-1.23638892e+00 5.65679848e-01 -1.80092797e-01 2.19349235e-01
-8.66912425e-01 -8.89982939e-01 -5.26772916e-01 -6.65067583e-02
-3.96159152e-03 -4.58413273e-01 1.66357982e+00 -9.09147561e-01
-1.79977477e+00 6.00450695e-01 -6.83449566e-01 -7.42486417e-01
2.51778811e-01 -1.33660734e-01 -8.97177398e-01 4.58374321e-01
-1.61077097e-01 -1.13908917e-01 1.46714795e+00 -1.21183944e+00
-5.63900530e-01 -1.56568900e-01 -6.09983623e-01 3.09132874e-01
4.84230034e-02 1.51934311e-01 -6.50992291e-03 -6.55207634e-01
6.92156732e-01 -6.90625310e-01 -5.43164201e-02 -4.45323706e-01
-4.12894756e-01 8.99396539e-02 3.82956207e-01 -1.17928493e+00
1.10802948e+00 -2.41452479e+00 1.08305544e-01 -8.88619870e-02
2.53811747e-01 1.04561612e-01 4.30196561e-02 2.19772547e-01
-3.38891149e-01 -4.52778846e-01 -2.09455892e-01 -4.46436793e-01
1.28616564e-04 -3.88412118e-01 -3.52779180e-01 7.20171273e-01
-2.38337845e-01 3.36860389e-01 -6.99404359e-01 -3.44818026e-01
8.14589858e-02 7.12960124e-01 -2.84851998e-01 2.61206210e-01
2.67348915e-01 7.37657547e-01 4.63608429e-02 2.34446466e-01
1.01616323e+00 2.89609581e-01 8.88806581e-02 -4.49424833e-01
-3.85310709e-01 8.82484794e-01 -1.54213405e+00 1.61853838e+00
-4.85428274e-01 9.81266856e-01 5.74213624e-01 -9.35696661e-01
7.35186458e-01 9.25741434e-01 7.38642514e-01 -5.21669805e-01
1.31530404e-01 5.54349661e-01 5.39949179e-01 -4.00336802e-01
1.40360013e-01 -5.14295220e-01 2.00048223e-01 5.99686265e-01
5.55146456e-01 -3.37641776e-01 9.22685117e-02 -4.23902303e-01
6.10465169e-01 -3.91364276e-01 1.44647881e-01 -3.52939479e-02
2.98787683e-01 -5.58846176e-01 5.80477595e-01 3.33012551e-01
-4.95390028e-01 8.44526529e-01 1.08415693e-01 3.73057544e-01
-4.10060793e-01 -1.38223147e+00 -1.48154631e-01 8.15878332e-01
-3.42322364e-02 -2.38196716e-01 -8.21829975e-01 4.77399528e-02
-2.92665899e-01 7.16993570e-01 6.58424422e-02 3.90232652e-02
-8.76900971e-01 -5.79942465e-01 5.33350229e-01 1.44952312e-01
2.46870503e-01 -4.86929387e-01 -3.27102959e-01 5.03332913e-01
-6.75090849e-01 -1.18348598e+00 -6.69422209e-01 6.44389987e-01
-5.16361773e-01 -8.22741449e-01 -7.53579676e-01 -9.36844587e-01
2.54950643e-01 3.85903865e-01 8.80771041e-01 -7.07208693e-01
7.83732459e-02 3.14624071e-01 -4.34673131e-02 -4.38372493e-01
-3.14834297e-01 -4.66482401e-01 4.64892715e-01 2.92201281e-01
3.69779319e-02 -9.87935781e-01 -7.58177936e-01 1.69681802e-01
-1.11855187e-01 -2.12694511e-01 1.76977724e-01 4.32829410e-01
4.30619597e-01 4.83135849e-01 5.74108958e-01 9.31588933e-02
4.39767152e-01 -9.00977105e-03 -4.75070864e-01 -1.08461380e-01
-2.42480397e-01 -9.24106762e-02 5.60521185e-01 -9.17376816e-01
-1.16669714e+00 -1.19602889e-01 -3.92568082e-01 -3.00634563e-01
-8.90976340e-02 4.62861866e-01 -2.48374730e-01 -1.17276415e-01
6.15931749e-01 5.29386282e-01 -2.61314750e-01 -5.54422736e-01
2.99698293e-01 8.42959762e-01 8.83144677e-01 -1.85026288e-01
9.51326489e-01 2.57255524e-01 -2.39468530e-01 -1.29617321e+00
-5.17219484e-01 -6.26557410e-01 -3.53345484e-01 -1.19725458e-01
5.92552543e-01 -1.14990604e+00 -6.45351350e-01 1.03241122e+00
-1.04062915e+00 -2.47141093e-01 -2.11876959e-01 1.22426987e+00
-4.67283577e-01 5.36756933e-01 -7.23087668e-01 -1.17028832e+00
-5.35549343e-01 -9.54981327e-01 6.25073552e-01 3.26191217e-01
-3.12523454e-01 -6.40775383e-01 3.17344308e-01 2.06676975e-01
4.49378222e-01 -1.25792593e-01 5.59103966e-01 -6.16885006e-01
-3.57956856e-01 -1.48790315e-01 6.96253717e-01 3.39702874e-01
1.69740960e-01 -9.26137939e-02 -1.40774429e+00 -2.19859242e-01
8.80638480e-01 2.69943893e-01 6.41531050e-01 1.22718096e+00
3.70597124e-01 -4.27264780e-01 -2.86733866e-01 7.54107893e-01
1.05182838e+00 3.18346083e-01 2.80721426e-01 -3.21424872e-01
4.81877208e-01 4.40584600e-01 3.47787023e-01 3.53044659e-01
2.16210529e-01 5.41831732e-01 -6.62573278e-02 5.10051474e-02
-5.32738388e-01 -4.18311618e-02 2.39728078e-01 1.46865213e+00
1.62575226e-02 -3.91333133e-01 -5.27859807e-01 6.72953486e-01
-1.17276931e+00 -9.01688933e-01 -3.59376758e-01 2.72129035e+00
1.10553885e+00 9.33807790e-02 4.06592973e-02 7.55971968e-01
7.40018487e-01 3.96003693e-01 -3.40318501e-01 -9.72552896e-02
-3.27788860e-01 7.30667830e-01 5.30065179e-01 7.13742971e-01
-9.10484254e-01 3.39341611e-01 6.93190718e+00 1.03677285e+00
-1.50810909e+00 3.70918602e-01 1.34058654e-01 -2.54890054e-01
-2.15965271e-01 -2.37006545e-02 -5.95463753e-01 3.71453583e-01
1.15496445e+00 -4.20948446e-01 7.41116941e-01 2.61533976e-01
1.39636666e-01 -1.34229675e-01 -1.02792323e+00 1.24320245e+00
-2.27631293e-02 -8.14617932e-01 -8.44298363e-01 -3.92538369e-01
3.95751446e-01 5.06898314e-02 2.54765958e-01 2.10541077e-02
-1.73570842e-01 -7.08159506e-01 1.09587574e+00 2.97173500e-01
7.60796070e-01 -7.70905972e-01 7.41136968e-02 4.42110479e-01
-1.49346280e+00 1.06720954e-01 2.86524296e-01 9.23954472e-02
4.27982748e-01 9.33805764e-01 -8.63041282e-01 6.77463353e-01
4.96959597e-01 2.30790228e-01 2.94962049e-01 1.42985702e+00
-6.35651469e-01 1.27963889e+00 -5.73085725e-01 3.58072460e-01
-4.71469611e-01 -4.87991683e-02 9.82740283e-01 1.02564967e+00
5.83873808e-01 1.06025301e-01 -3.42576385e-01 5.43155074e-01
1.96947753e-01 -2.16462746e-01 7.34696835e-02 2.00031064e-02
8.33103418e-01 1.13071406e+00 -5.67398071e-01 1.44849807e-01
-6.25219792e-02 7.17702091e-01 -4.17576820e-01 6.74200892e-01
-7.96300948e-01 -6.67428911e-01 8.55813444e-01 4.96780649e-02
6.64459527e-01 -4.80034739e-01 6.20481966e-04 -9.17938411e-01
1.63760129e-02 -8.84821296e-01 -6.72563910e-03 -6.05885565e-01
-1.10404646e+00 6.19212627e-01 -1.36442676e-01 -1.37068391e+00
-5.98838091e-01 -2.55722553e-01 -7.32014716e-01 1.13285863e+00
-1.67829049e+00 -7.66862869e-01 3.39233488e-01 4.90190119e-01
2.58720189e-01 7.18653426e-02 1.03044355e+00 4.71360773e-01
-1.52713150e-01 7.41347969e-01 3.48112434e-01 -4.56010252e-02
9.31326568e-01 -1.15951765e+00 2.12786943e-01 9.66447830e-01
3.11330557e-01 4.46005195e-01 1.25687218e+00 -1.35984331e-01
-1.31886721e+00 -4.80271637e-01 1.07959604e+00 1.03137329e-01
4.56880569e-01 -5.36715806e-01 -7.24048316e-01 1.98400125e-01
5.10977626e-01 -9.04668123e-02 8.95002365e-01 -1.70904677e-02
-4.67404187e-01 -4.99800622e-01 -9.24043059e-01 2.76058406e-01
5.73925495e-01 -7.19314635e-01 -5.44094443e-01 2.57489502e-01
7.53846228e-01 -7.19122469e-01 -7.95682907e-01 2.63112813e-01
6.11344874e-01 -1.08996952e+00 1.16815507e+00 1.94130242e-01
-3.01414818e-01 -5.81023991e-01 -1.18255906e-01 -1.74418187e+00
-1.04712948e-01 -1.51766872e+00 -4.82036859e-01 1.23147571e+00
1.65547177e-01 -9.21721220e-01 2.88143426e-01 -6.06554747e-02
-2.62671351e-01 -1.74401596e-01 -1.32634473e+00 -8.69691551e-01
-2.70958066e-01 -7.54162490e-01 5.40392280e-01 7.33153880e-01
1.39387861e-01 4.42073941e-01 -4.42977697e-01 6.88128412e-01
8.59804451e-01 3.56262952e-01 3.60686511e-01 -1.18092000e+00
-9.37844694e-01 -5.85549951e-01 9.81564373e-02 -1.37488258e+00
2.98476182e-02 -3.70119154e-01 5.20308852e-01 -1.45212746e+00
-4.71112788e-01 -3.19884151e-01 -4.54274267e-01 -8.14409032e-02
-2.07875893e-01 9.03964043e-02 4.57187369e-02 4.55399081e-02
4.50977571e-02 4.79409933e-01 9.63292420e-01 -1.84155524e-01
-1.84531495e-01 6.35662019e-01 -5.03966749e-01 8.31424057e-01
4.55588073e-01 -6.21332109e-01 -2.92591482e-01 -2.50948727e-01
-6.43504038e-02 7.43871450e-01 -7.28861839e-02 -1.23099756e+00
3.95611048e-01 1.37155801e-01 8.58411118e-02 -8.62889290e-01
8.42455268e-01 -4.38087881e-01 4.01859015e-01 4.58444655e-01
-6.01694845e-02 -6.80273592e-01 2.50117630e-01 3.36358279e-01
-4.40417856e-01 -2.62054265e-01 7.93958724e-01 2.43080497e-01
-1.77262098e-01 1.19346306e-01 -4.77339536e-01 1.95814088e-01
4.71139520e-01 -3.51169966e-02 -2.78930515e-01 -7.57793427e-01
-4.14611816e-01 -2.99998045e-01 -2.65809923e-01 -2.48955041e-02
2.93161243e-01 -1.24322140e+00 -8.82806659e-01 2.60035485e-01
-3.23184341e-01 -1.36408374e-01 7.33707070e-01 1.09912813e+00
-1.57437623e-01 4.88887876e-01 8.80950168e-02 -7.43238151e-01
-1.18452585e+00 2.68858463e-01 7.57663965e-01 4.28975075e-02
-1.40598953e-01 1.46613276e+00 1.40235424e-01 -8.25223699e-02
3.51707727e-01 -3.95780951e-01 -2.76653826e-01 2.85150707e-01
6.94494784e-01 3.87300611e-01 2.76599884e-01 -9.01297390e-01
-3.59618694e-01 9.00402725e-01 6.03179038e-01 -6.46728992e-01
1.10311055e+00 -2.57485956e-01 -3.91591415e-02 6.77312434e-01
1.38601995e+00 8.68955493e-01 -1.06620049e+00 -3.55146050e-01
-4.38719749e-01 -2.49791890e-01 2.30179459e-01 -6.22730970e-01
-1.08316231e+00 9.19932008e-01 8.21286440e-01 4.54152793e-01
1.46235526e+00 -2.59759367e-01 1.06552815e+00 -3.86728421e-02
1.16305344e-01 -7.74229765e-01 -1.04294869e-03 4.46278542e-01
4.30220336e-01 -7.31950760e-01 -2.48403594e-01 -2.50153899e-01
-2.56859839e-01 9.65608895e-01 5.77979684e-02 1.66695729e-01
8.44616055e-01 4.84222651e-01 4.02952701e-01 2.23508418e-01
-4.10441607e-01 -1.02544762e-01 5.79756856e-01 9.04095292e-01
6.69612169e-01 2.29633436e-01 -1.48498684e-01 7.38961935e-01
-8.70738447e-01 -5.21384120e-01 2.05349326e-01 4.78862345e-01
-5.52454174e-01 -9.79293406e-01 -7.92455137e-01 -3.29957977e-02
-6.48088872e-01 -1.64504379e-01 1.69665247e-01 2.24861190e-01
-8.61536488e-02 1.47370124e+00 9.58481357e-02 -2.40293339e-01
3.39762300e-01 2.85029322e-01 8.45091939e-01 -4.22493070e-02
-3.66868973e-01 8.69411826e-01 2.00218558e-01 -1.27041176e-01
-3.65498185e-01 -7.81797588e-01 -1.28687310e+00 -1.10215701e-01
-6.28051698e-01 4.42582399e-01 9.23915625e-01 8.37036371e-01
-3.82155329e-02 7.25038588e-01 8.98432136e-01 -7.51037717e-01
-5.55736661e-01 -1.11941957e+00 -8.97907376e-01 -1.40353158e-01
1.01409543e+00 -1.53392449e-01 -8.41383040e-01 6.32223189e-02] | [15.078801155090332, 5.73229455947876] |
905f688c-2f59-4b74-8825-4d150abc592f | cl-monoise-cross-lingual-lexical | null | null | https://aclanthology.org/2021.wnut-1.56 | https://aclanthology.org/2021.wnut-1.56.pdf | CL-MoNoise: Cross-lingual Lexical Normalization | Social media is notoriously difficult to process for existing natural language processing tools, because of spelling errors, non-standard words, shortenings, non-standard capitalization and punctuation. One method to circumvent these issues is to normalize input data before processing. Most previous work has focused on only one language, which is mostly English. In this paper, we are the first to propose a model for cross-lingual normalization, with which we participate in the WNUT 2021 shared task. To this end, we use MoNoise as a starting point, and make a simple adaptation for cross-lingual application. Our proposed model outperforms the leave-as-is baseline provided by the organizers which copies the input. Furthermore, we explore a completely different model which converts the task to a sequence labeling task. Performance of this second system is low, as it does not take capitalization into account in our implementation. | ['Rob van der Goot'] | null | null | null | null | emnlp-wnut-2021-11 | ['lexical-normalization'] | ['natural-language-processing'] | [ 1.08529590e-01 -1.06505610e-01 -4.52116802e-02 -4.85844076e-01
-7.17672884e-01 -7.08446264e-01 5.81271410e-01 4.17247623e-01
-1.14538872e+00 7.91825056e-01 3.21291775e-01 -4.24757034e-01
5.00567019e-01 -5.23191869e-01 -4.66848731e-01 -2.21551001e-01
5.79866707e-01 4.18166965e-01 3.63946378e-01 -4.00890887e-01
2.61638671e-01 2.97000289e-01 -1.06570280e+00 3.60160917e-01
8.59063327e-01 1.11658804e-01 2.24741265e-01 5.14352679e-01
-5.75752437e-01 4.53153521e-01 -7.08725393e-01 -8.36591125e-01
2.32068330e-01 -5.31824410e-01 -9.54696894e-01 -2.01581046e-01
3.08995515e-01 1.15857676e-01 2.04890206e-01 1.21095109e+00
5.15268862e-01 9.20896158e-02 7.74157584e-01 -1.05348086e+00
-5.46847224e-01 1.11981940e+00 -5.53056180e-01 2.04175070e-01
9.06291083e-02 -3.06844655e-02 9.00394678e-01 -7.53545046e-01
8.12975287e-01 1.19029295e+00 7.78680027e-01 6.36980474e-01
-1.15396714e+00 -5.68247378e-01 2.32760027e-01 5.15409629e-04
-1.56076360e+00 -5.85691988e-01 3.26072246e-01 -3.64042163e-01
1.27453339e+00 2.31437728e-01 1.66911289e-01 9.60440457e-01
1.57271065e-02 5.86540461e-01 1.10454798e+00 -1.01947081e+00
2.84163933e-02 1.92339346e-01 2.89945275e-01 2.47979224e-01
2.00665981e-01 -5.42768419e-01 -1.33009270e-01 2.03524128e-01
3.05216819e-01 -2.13485643e-01 1.36475889e-02 -7.27503980e-03
-1.27027071e+00 5.71532130e-01 -7.60522112e-02 8.79484296e-01
-2.70553052e-01 -1.87508702e-01 6.94275856e-01 3.71653289e-01
5.65569162e-01 5.18336773e-01 -7.81251073e-01 -2.43187442e-01
-1.17624617e+00 2.29789779e-01 8.49813879e-01 1.00811124e+00
8.53595972e-01 -3.28720957e-01 -2.16922387e-01 9.89951789e-01
1.87107533e-01 3.07608455e-01 8.17308128e-01 -5.43687642e-01
6.90089583e-01 3.33155125e-01 -4.78430092e-02 -6.39903307e-01
-4.76802289e-01 -1.76941529e-01 -6.49773300e-01 5.26515581e-02
4.96775925e-01 -2.97279626e-01 -8.96173358e-01 1.80945218e+00
7.36492127e-02 -2.95348525e-01 1.67580217e-01 4.92294729e-01
5.52636445e-01 5.70564747e-01 4.11109477e-01 -2.95227468e-01
1.46768832e+00 -1.28812182e+00 -9.35419142e-01 -2.62632728e-01
1.17556810e+00 -1.39270663e+00 1.39463305e+00 3.52752596e-01
-9.99484599e-01 -4.69083101e-01 -1.03699195e+00 -3.37325186e-01
-8.89764905e-01 9.73292142e-02 2.11196408e-01 8.00066173e-01
-9.17505801e-01 7.96128809e-01 -8.58106434e-01 -9.91727293e-01
-2.87274867e-01 -4.21649814e-02 -3.75800610e-01 1.94056362e-01
-1.26092350e+00 1.26851618e+00 6.55747294e-01 -2.03653887e-01
2.48993989e-02 -6.01220489e-01 -7.73305357e-01 -2.72430152e-01
3.15499723e-01 -3.26248914e-01 1.40979493e+00 -1.01415277e+00
-1.54915309e+00 1.17392159e+00 -3.49033684e-01 -4.36743468e-01
7.82820344e-01 -3.65723550e-01 -6.31737828e-01 -5.90749025e-01
1.89364582e-01 4.74625051e-01 4.36633021e-01 -8.05608690e-01
-5.70284188e-01 -2.18530476e-01 -8.01424161e-02 1.55654281e-01
-5.23540854e-01 6.28182650e-01 -7.08531618e-01 -1.00001693e+00
-2.17976853e-01 -1.12074864e+00 -4.77779627e-01 -6.51125550e-01
-4.64957505e-01 -2.26999208e-01 3.59766781e-01 -8.84484947e-01
1.73731291e+00 -2.16148639e+00 -4.09429260e-02 8.61622915e-02
-3.36139232e-01 5.78817189e-01 -3.41806114e-01 6.82121098e-01
-2.81689167e-01 6.98796809e-01 -4.30352062e-01 -1.00397503e+00
1.65843055e-01 2.81935781e-01 4.27234545e-02 2.58957088e-01
1.91325590e-01 7.95302391e-01 -7.95244992e-01 -5.14542818e-01
6.35728016e-02 4.17160362e-01 -6.06094956e-01 6.45216852e-02
-9.13829803e-02 1.53411686e-01 1.31159753e-01 2.70460486e-01
7.62820482e-01 4.07836914e-01 1.90793812e-01 -6.05989024e-02
-6.98314428e-01 7.60863960e-01 -1.25759971e+00 1.98066676e+00
-5.96659541e-01 2.94675499e-01 -1.23959340e-01 -6.20166183e-01
7.09047914e-01 2.20033363e-01 1.18114337e-01 -5.57154298e-01
2.11458042e-01 3.92422616e-01 3.54862250e-02 -4.42795932e-01
8.08293521e-01 -7.32025877e-02 -2.90046662e-01 6.33825362e-01
1.45326450e-01 -1.80052578e-01 8.33503962e-01 1.73236161e-01
8.91138375e-01 2.63892591e-01 7.01064408e-01 -4.32907850e-01
6.88969553e-01 -1.48950070e-01 6.16940379e-01 6.37172520e-01
-1.83099866e-01 9.15311277e-01 3.67068648e-01 -1.54475480e-01
-1.33409011e+00 -6.23718441e-01 1.59365073e-01 1.28685904e+00
-3.97328854e-01 -8.89749587e-01 -1.11415625e+00 -7.47000813e-01
-4.36201870e-01 8.75408828e-01 -3.60577047e-01 8.38677138e-02
-8.64512324e-01 -8.55290055e-01 8.54438603e-01 2.62662113e-01
1.95804253e-01 -1.13657010e+00 3.84326018e-02 4.28099245e-01
-3.34233791e-01 -1.31567502e+00 -8.38883400e-01 2.75881231e-01
-5.17205000e-01 -6.92269444e-01 -6.53424799e-01 -7.92111158e-01
3.73056322e-01 -2.43519563e-02 1.11401129e+00 1.30193055e-01
8.63155499e-02 -1.69949204e-01 -6.18114054e-01 -7.61546612e-01
-7.88145244e-01 8.36747050e-01 5.34049608e-02 -1.46790966e-01
8.35841060e-01 -4.11751777e-01 -5.12012430e-02 1.57466173e-01
-1.11840606e+00 1.08166840e-02 5.49490988e-01 3.42367738e-01
5.32860577e-01 -2.95388430e-01 3.38490725e-01 -1.32205546e+00
7.70954192e-01 -2.18867064e-01 -3.79300147e-01 2.82170773e-01
-5.19152880e-01 1.52318507e-01 9.09558952e-01 -1.37859792e-01
-9.48795259e-01 2.75393009e-01 -7.27949739e-01 2.40493536e-01
-5.27192414e-01 5.80214798e-01 -5.28384686e-01 3.07453394e-01
6.64410949e-01 -1.24908954e-01 -1.78516269e-01 -9.19752538e-01
5.18569410e-01 8.30875456e-01 5.23165405e-01 -3.17775756e-01
9.04770553e-01 -1.26884058e-01 -5.11521280e-01 -9.43557978e-01
-7.70527780e-01 -6.19174540e-01 -1.18737698e+00 3.35320413e-01
8.69291723e-01 -7.66511202e-01 -1.62811533e-01 6.75107121e-01
-1.59465325e+00 -3.01386356e-01 -1.06506675e-01 5.52562773e-01
-1.18163757e-01 6.82969451e-01 -7.35690594e-01 -4.34695780e-01
-3.41313303e-01 -9.09163296e-01 5.71850717e-01 9.50370543e-03
-6.81796432e-01 -9.14663196e-01 2.90924966e-01 1.77307427e-01
5.56350648e-01 -9.63122174e-02 6.25049412e-01 -9.10680950e-01
-1.05305659e-02 -9.65185985e-02 -8.28766599e-02 6.65783823e-01
3.08605164e-01 4.09134775e-01 -9.05668557e-01 -1.46307319e-01
-1.96391955e-01 -7.10568503e-02 8.48336756e-01 -3.46816182e-02
8.98364604e-01 -1.29612744e-01 7.84541592e-02 6.96223617e-01
1.14610708e+00 -2.75645349e-02 8.29635441e-01 6.39784038e-01
7.79941142e-01 6.14438295e-01 5.01229644e-01 2.41369203e-01
7.16908395e-01 6.28576040e-01 -3.98134775e-02 -4.25302804e-01
-2.13211268e-01 -2.43247941e-01 5.41168511e-01 1.48491001e+00
-5.05113006e-02 -4.16680276e-01 -9.19329405e-01 5.49552500e-01
-1.78111303e+00 -7.11022556e-01 -4.31510389e-01 2.05719829e+00
1.09059656e+00 2.69743890e-01 1.07352123e-01 1.04978867e-01
7.01467931e-01 8.38592798e-02 1.42746031e-01 -8.72380912e-01
-2.15437338e-01 2.79899746e-01 8.04607630e-01 7.27415204e-01
-1.29813874e+00 1.48662889e+00 5.75801945e+00 9.20064092e-01
-1.25592768e+00 2.86250204e-01 2.59288371e-01 1.67756513e-01
-1.30245358e-01 7.35028312e-02 -1.10929883e+00 5.09423375e-01
1.14724541e+00 -1.18365072e-01 3.77091438e-01 6.04905069e-01
3.41495365e-01 9.25716981e-02 -1.04937983e+00 7.07971692e-01
2.04232723e-01 -8.26318443e-01 -1.06557785e-02 -3.39765906e-01
5.21259308e-01 2.69682229e-01 -3.89597297e-01 4.39303368e-01
3.20891440e-01 -8.04298997e-01 8.63947511e-01 2.23125070e-01
5.38726091e-01 -8.50684285e-01 8.96648109e-01 3.30140412e-01
-1.04838908e+00 6.22399032e-01 -3.72469813e-01 -2.38262743e-01
4.58893806e-01 5.57891786e-01 -7.38992095e-01 4.99940187e-01
4.58365083e-01 4.77078855e-01 -8.71656835e-01 1.01804435e+00
-4.73268479e-01 6.81207359e-01 -3.68491143e-01 -3.65459472e-02
1.49431497e-01 -1.31528333e-01 3.52240771e-01 2.06086135e+00
3.86530787e-01 -2.91592747e-01 2.59262145e-01 2.91108876e-01
-1.98035792e-01 9.21121240e-01 -3.09211433e-01 -1.21047802e-01
2.47642577e-01 1.30745959e+00 -9.61617887e-01 -4.08773094e-01
-5.71800292e-01 1.26978076e+00 5.44833362e-01 1.78730726e-01
-6.91543341e-01 -7.74326980e-01 5.94617069e-01 6.38572425e-02
2.17881829e-01 -3.38748634e-01 -3.55460823e-01 -1.50193846e+00
4.29089256e-02 -9.33719218e-01 3.80993754e-01 -4.16022778e-01
-1.33406126e+00 8.85476291e-01 -9.92263630e-02 -9.37827587e-01
-3.42598200e-01 -6.20348394e-01 -2.82379568e-01 1.12506294e+00
-1.81804800e+00 -1.12148738e+00 1.05351731e-01 4.39536482e-01
4.76186872e-01 2.47868318e-02 8.55266571e-01 8.83652925e-01
-7.63634562e-01 9.34861541e-01 6.33150041e-02 3.42369944e-01
1.51701975e+00 -1.25339925e+00 9.35612917e-01 1.41323352e+00
1.93791792e-01 8.91803801e-01 9.55765903e-01 -6.04888856e-01
-7.10708618e-01 -1.26543391e+00 1.93301332e+00 -5.45500755e-01
9.52035785e-01 -7.03134418e-01 -1.18064427e+00 7.24513113e-01
5.98928392e-01 -1.79646030e-01 8.14657629e-01 2.39477083e-01
-2.91442454e-01 8.93650576e-02 -8.47856164e-01 7.66003549e-01
9.80264068e-01 -3.82240772e-01 -6.52183473e-01 2.18296900e-01
8.31028461e-01 -3.40720087e-01 -5.92073917e-01 2.67818458e-02
2.66225129e-01 -6.30755723e-01 5.36256373e-01 -5.48889875e-01
3.70137870e-01 -4.53042239e-01 8.05383474e-02 -1.60366821e+00
-3.99280876e-01 -6.41110301e-01 7.24191248e-01 2.00787544e+00
8.39596510e-01 -6.27284408e-01 4.60314393e-01 4.33785856e-01
-3.10202688e-01 1.20725468e-01 -5.62844276e-01 -9.16936278e-01
2.78476924e-01 -5.93340218e-01 7.11294055e-01 1.15509522e+00
-2.03601122e-02 4.82205212e-01 -4.30879951e-01 -1.10047288e-01
2.72493273e-01 -3.88469815e-01 9.30020273e-01 -1.04540038e+00
-1.63063005e-01 -5.35585940e-01 -4.25374620e-02 -7.17238903e-01
2.82108068e-01 -1.10283959e+00 3.25933009e-01 -1.19907665e+00
-8.25623944e-02 -3.99211794e-01 -3.48172396e-01 6.53864861e-01
-2.43524209e-01 5.58081269e-01 5.30854225e-01 -7.91348889e-02
-5.44653296e-01 2.52320558e-01 7.53278375e-01 1.58627033e-01
-3.98011863e-01 -2.12763965e-01 -7.77513325e-01 9.16214883e-01
1.04349780e+00 -7.94841826e-01 1.09989427e-01 -7.47010410e-01
3.15324783e-01 -7.23390460e-01 -2.32670859e-01 -8.67565811e-01
2.24949375e-01 -1.84479147e-01 -4.98130880e-02 -3.93402189e-01
-2.44940698e-01 -7.08280683e-01 1.01841256e-01 2.02733755e-01
-3.10499847e-01 4.36678708e-01 2.59618521e-01 -2.12733701e-01
-1.45032540e-01 -7.50931621e-01 7.91233659e-01 -2.74954915e-01
-5.93159139e-01 7.88087770e-02 -7.04941392e-01 2.57611394e-01
7.08095133e-01 2.93501206e-02 -2.16469988e-01 1.34027869e-01
-4.21667635e-01 -4.00144234e-03 8.87194872e-01 5.69120228e-01
-3.10041130e-01 -1.12610626e+00 -8.59499991e-01 9.30392519e-02
2.09257200e-01 -2.93254703e-01 2.28773579e-02 6.67104721e-01
-8.42970490e-01 3.56922895e-01 -2.77368456e-01 5.18499734e-03
-1.34148645e+00 6.33146346e-01 3.96041945e-02 -5.28812408e-01
-2.42782801e-01 4.83513772e-01 -2.87596524e-01 -8.79430354e-01
9.11061689e-02 -5.07572234e-01 -4.05564129e-01 4.10984367e-01
7.29563594e-01 2.14186847e-01 7.77344882e-01 -8.78343165e-01
-4.60609347e-01 5.97276568e-01 -3.86081457e-01 -2.91217208e-01
1.28481317e+00 -3.49385858e-01 -4.69049484e-01 6.43768907e-01
1.25704682e+00 5.17986000e-01 -5.86350799e-01 -2.97856361e-01
3.85504782e-01 -1.10788688e-01 -2.06115887e-01 -6.75787032e-01
-6.55164540e-01 8.65636706e-01 2.88014621e-01 1.51105106e-01
9.08508778e-01 -3.47537905e-01 7.68657386e-01 4.28769290e-01
2.81069223e-02 -1.55852103e+00 -6.39303923e-01 1.17522025e+00
6.48124397e-01 -1.23518741e+00 -1.43551171e-01 -4.02612805e-01
-5.63680589e-01 1.10601413e+00 6.76823378e-01 9.27335173e-02
5.14470994e-01 4.07695681e-01 5.00636995e-01 4.24441189e-01
-4.77337867e-01 -4.15690899e-01 5.60260527e-02 4.62465823e-01
1.04613435e+00 9.26175490e-02 -1.12593222e+00 7.00031102e-01
-6.18059874e-01 2.22697686e-02 7.79983342e-01 9.94007528e-01
-1.57969877e-01 -1.84151244e+00 -3.20948809e-01 1.62734166e-02
-1.00557208e+00 -6.37857020e-01 -4.05508071e-01 7.42908418e-01
3.01041663e-01 6.99749053e-01 8.20964426e-02 -1.86726674e-01
6.20758414e-01 2.57888198e-01 2.24902868e-01 -9.37359273e-01
-9.06842649e-01 1.16799615e-01 7.25459903e-02 -3.33464652e-01
-4.27415639e-01 -8.95425379e-01 -1.12358046e+00 -3.07871521e-01
4.90270779e-02 1.33415177e-01 7.26333380e-01 1.01967883e+00
9.99849364e-02 3.71094644e-01 1.31119505e-01 -7.80627310e-01
-2.39323288e-01 -1.22896588e+00 -3.13014865e-01 5.26962698e-01
9.00430903e-02 -1.08801790e-01 -2.79501617e-01 3.06518584e-01] | [10.272357940673828, 10.035526275634766] |
356b9ae0-e427-4a91-920d-e62ec6d30220 | k-nearest-neighbor-optimization-via | 1906.04559 | null | https://arxiv.org/abs/1906.04559v1 | https://arxiv.org/pdf/1906.04559v1.pdf | k-Nearest Neighbor Optimization via Randomized Hyperstructure Convex Hull | In the k-nearest neighbor algorithm (k-NN), the determination of classes for test instances is usually performed via a majority vote system, which may ignore the similarities among data. In this research, the researcher proposes an approach to fine-tune the selection of neighbors to be passed to the majority vote system through the construction of a random n-dimensional hyperstructure around the test instance by introducing a new threshold parameter. The accuracy of the proposed k-NN algorithm is 85.71%, while the accuracy of the conventional k-NN algorithm is 80.95% when performed on the Haberman's Cancer Survival dataset, and 94.44% for the proposed k-NN algorithm, compared to the conventional's 88.89% accuracy score on the Seeds dataset. The proposed k-NN algorithm is also on par with the conventional support vector machine algorithm accuracy, even on the Banknote Authentication and Iris datasets, even surpassing the accuracy of support vector machine on the Seeds dataset. | ['Jasper Kyle Catapang'] | 2019-06-11 | null | null | null | null | ['unsupervised-spatial-clustering'] | ['time-series'] | [-1.00229368e-01 -6.38863891e-02 -5.61421096e-01 -5.14333904e-01
-4.01594907e-01 -5.85546553e-01 4.89411831e-01 4.24805850e-01
-5.90404034e-01 9.90977585e-01 -1.11827426e-01 -5.45013726e-01
-8.59975994e-01 -9.75126088e-01 1.34996116e-01 -8.63654792e-01
3.00425440e-01 6.99406147e-01 1.66222140e-01 -1.98868215e-01
9.39526260e-01 7.70562708e-01 -1.78669131e+00 4.27039444e-01
7.34154403e-01 1.10715055e+00 -6.28354073e-01 5.70682049e-01
-7.19719566e-04 4.37987238e-01 -6.43911541e-01 -2.93281078e-01
4.06961739e-01 -3.49309713e-01 -5.44940352e-01 -3.14210206e-01
3.76402527e-01 -3.25428024e-02 -5.55214919e-02 8.38075817e-01
6.00851476e-01 2.55687386e-01 1.05778730e+00 -1.43657112e+00
-5.81723154e-01 6.03673577e-01 -3.42539728e-01 6.99232891e-02
3.02937239e-01 -6.00003600e-01 6.41663492e-01 -9.93962824e-01
4.36411262e-01 5.51103413e-01 9.57537234e-01 3.17988396e-01
-9.25410509e-01 -7.22970963e-01 -5.59320152e-01 5.18471837e-01
-1.97447336e+00 -1.33634433e-01 6.29709840e-01 -3.43302399e-01
6.48825943e-01 6.69306636e-01 8.56709719e-01 2.73900449e-01
3.19770962e-01 2.99321085e-01 1.39685571e+00 -9.14075315e-01
6.89239740e-01 3.81734550e-01 7.00709522e-01 4.80616361e-01
6.04374528e-01 2.96659976e-01 -3.17039341e-01 -7.03345716e-01
5.34583092e-01 2.52696037e-01 -5.73851110e-04 -4.14294451e-01
-8.74499202e-01 8.94519389e-01 2.34903902e-01 8.38342249e-01
-1.64389804e-01 -7.80193746e-01 3.01549852e-01 4.26691085e-01
1.75434068e-01 5.52287400e-01 -5.74684203e-01 1.40912354e-01
-1.40507841e+00 1.32849291e-01 9.99324918e-01 3.59180033e-01
5.72914302e-01 -2.37533629e-01 -1.90743014e-01 6.70189738e-01
8.81956890e-02 2.23268077e-01 7.84961760e-01 -5.64904451e-01
-7.03042001e-03 1.00442612e+00 9.93811712e-02 -1.14282525e+00
-5.35750508e-01 -6.36753798e-01 -9.18685436e-01 4.17543083e-01
7.58796096e-01 -4.74054046e-04 -7.62095988e-01 9.40098047e-01
5.46856165e-01 3.78161848e-01 3.12637627e-01 6.60902560e-01
8.79884422e-01 4.28866744e-01 -2.46778890e-01 -4.29379344e-01
8.55568886e-01 -6.54050529e-01 -4.12850529e-01 4.87160951e-01
8.83965492e-01 -8.87968898e-01 6.20258093e-01 8.09243143e-01
-5.01406252e-01 -5.56345403e-01 -1.01034451e+00 6.24705791e-01
-5.46288729e-01 3.17652136e-01 6.65149927e-01 9.60604668e-01
-8.45934868e-01 6.97548270e-01 -1.58176616e-01 -4.04607028e-01
7.03043416e-02 5.96778989e-01 -7.76485384e-01 -1.97206721e-01
-1.04988468e+00 9.60034430e-01 4.26889628e-01 1.75895728e-03
-9.64811444e-03 -4.69622910e-01 -5.68929553e-01 -1.57339443e-02
-2.18077332e-01 -3.37104708e-01 4.91621405e-01 -7.31896102e-01
-1.33888531e+00 8.03670526e-01 -9.79487374e-02 -1.01248957e-01
3.49427551e-01 2.96211004e-01 -7.97293127e-01 -1.88695401e-01
-4.15218733e-02 7.16003105e-02 2.82784432e-01 -1.11551392e+00
-7.16099739e-01 -6.23411119e-01 -2.68536896e-01 8.13478902e-02
-3.25860530e-01 -3.21843237e-01 4.05471213e-02 -4.93110836e-01
6.04520798e-01 -1.04013729e+00 -4.99177396e-01 -4.01604921e-01
-4.54580694e-01 -5.33083141e-01 6.78526163e-01 -3.41107786e-01
1.59820163e+00 -2.15451503e+00 -5.14680743e-01 1.13739622e+00
-8.59856009e-02 5.75814664e-01 1.90633625e-01 3.40021044e-01
-5.46455979e-01 -2.72130728e-01 -1.57836094e-01 5.19955039e-01
-3.93133342e-01 6.68339580e-02 2.62904689e-02 7.72921443e-01
-4.45968151e-01 3.43470961e-01 -5.35718262e-01 -6.18097663e-01
3.36915374e-01 3.65469992e-01 -2.79941410e-01 -2.77813256e-01
8.05517018e-01 -5.04701808e-02 -4.95490372e-01 9.48666871e-01
6.41146064e-01 -1.40436381e-01 1.84120581e-01 -1.93127081e-01
-3.01925540e-02 -6.49171591e-01 -1.70517468e+00 9.14395750e-01
1.13622278e-01 3.03851694e-01 -4.81897265e-01 -1.30361640e+00
1.55374813e+00 4.32414353e-01 4.47383970e-01 -4.67428714e-01
3.00332308e-01 5.79550207e-01 1.01455197e-01 -3.45521301e-01
2.24211946e-01 -1.51629537e-01 2.70136278e-02 2.45506585e-01
-3.58205289e-01 1.93813760e-02 -1.24722272e-02 -2.27287710e-01
8.72282982e-01 -4.32872504e-01 8.50082457e-01 -5.39368153e-01
1.00186920e+00 3.70400280e-01 6.29198849e-01 7.94694483e-01
-2.57882833e-01 3.97271037e-01 1.21869043e-01 -1.01136661e+00
-7.60012329e-01 -7.45199382e-01 -9.06691730e-01 5.18753290e-01
6.26263469e-02 -5.68328239e-02 -8.00722539e-01 -7.87002742e-01
4.73379523e-01 8.50357056e-01 -8.18651021e-01 -1.12618379e-01
-2.14899689e-01 -7.44821966e-01 4.01282459e-01 1.21460304e-01
3.77018213e-01 -6.61090791e-01 -2.99373944e-03 7.29458258e-02
3.36881816e-01 -4.34159845e-01 4.13517281e-02 1.96221381e-01
-9.72315431e-01 -1.44268107e+00 -4.34530377e-01 -5.86047709e-01
9.00680900e-01 -8.92192721e-02 7.88287640e-01 3.07077974e-01
-3.06694448e-01 -4.94183227e-02 -4.17558700e-01 -2.09826931e-01
-1.47248149e-01 1.18534945e-01 2.90195435e-01 5.75626008e-02
9.37934935e-01 -2.94792801e-01 -5.56877673e-01 6.25473917e-01
-5.62272489e-01 -7.22705483e-01 5.99219799e-01 9.81715441e-01
7.60771215e-01 4.97364759e-01 8.56289566e-01 -1.04214931e+00
4.73516345e-01 -5.79388738e-01 -4.92953539e-01 4.39879149e-01
-1.28122520e+00 -2.92226315e-01 6.10714912e-01 -4.39773530e-01
-4.58404571e-01 2.78106421e-01 1.54058129e-01 -1.85517464e-02
-2.41987571e-01 2.85574257e-01 1.18646652e-01 -5.93481183e-01
1.01185668e+00 9.57044289e-02 -1.71358883e-01 -8.74336511e-02
-1.30670995e-01 1.00934172e+00 3.97922099e-01 -2.26237491e-01
6.41439915e-01 1.68819800e-01 3.75510782e-01 -4.76165205e-01
-5.24600923e-01 -5.80027878e-01 -7.80652821e-01 5.58237508e-02
5.12374520e-01 -4.00778681e-01 -5.58063328e-01 4.99490201e-01
-5.85793972e-01 3.21031064e-01 -4.02722090e-01 9.44962859e-01
-2.31735840e-01 2.42192015e-01 -1.33535713e-01 -6.84612691e-01
-3.49279791e-01 -9.03473258e-01 3.72700781e-01 2.63630033e-01
-5.92076421e-01 -9.83921826e-01 2.16231018e-01 5.53783178e-01
3.07756305e-01 1.84587806e-01 8.93714845e-01 -1.44599962e+00
3.28650206e-01 -1.03301930e+00 -2.28040777e-02 1.20236479e-01
2.51890123e-01 2.19788283e-01 -8.54420364e-01 -2.58151263e-01
-1.33049950e-01 2.47624084e-01 5.46758771e-01 6.19488537e-01
1.00640154e+00 6.33385852e-02 -4.93457377e-01 5.30050159e-01
1.53182137e+00 6.91232502e-01 6.76985264e-01 5.98654687e-01
1.21108189e-01 4.03567404e-01 8.37685525e-01 5.65272570e-01
2.11562127e-01 4.18972254e-01 1.62483364e-01 -1.57176301e-01
3.38781446e-01 1.75806060e-01 -4.41324890e-01 4.84958559e-01
-6.12896830e-02 -1.20013924e-02 -1.03093171e+00 3.68016392e-01
-1.61844313e+00 -8.99009705e-01 -4.04295892e-01 2.49826241e+00
5.28334200e-01 -8.68069902e-02 -6.92454576e-02 1.06304967e+00
8.81380916e-01 -3.88385653e-01 -3.54713887e-01 -6.87732816e-01
-1.54923216e-01 4.21615988e-01 6.00358486e-01 5.91839731e-01
-1.18629169e+00 5.79441965e-01 6.92473698e+00 1.02263594e+00
-9.72853780e-01 -3.94468427e-01 7.91125000e-01 2.61691138e-02
1.08223394e-01 6.58999160e-02 -1.02264404e+00 4.53268170e-01
7.92579293e-01 -1.66140258e-01 5.63556179e-02 9.80154812e-01
8.70509371e-02 -3.73922676e-01 -8.22361231e-01 1.18767178e+00
3.59251678e-01 -1.43094027e+00 -3.34454253e-02 1.41028641e-02
9.74201441e-01 -6.07523918e-01 -1.51603222e-01 1.34819031e-01
8.02364573e-02 -1.20451403e+00 -7.03897253e-02 7.57715225e-01
6.58471048e-01 -1.17721856e+00 1.25925815e+00 4.69140410e-01
-7.99493372e-01 -3.26976687e-01 -3.77977282e-01 -2.89756030e-01
-4.75362420e-01 1.14578819e+00 -1.08793390e+00 3.24646711e-01
7.00860560e-01 1.66255653e-01 -5.61145008e-01 1.29915774e+00
3.00711215e-01 6.61018848e-01 -3.37352842e-01 -1.17639221e-01
1.69001311e-01 -2.37099871e-01 1.14896610e-01 8.44067395e-01
6.17636144e-01 5.40645778e-01 3.60073708e-02 -5.49655296e-02
3.27360690e-01 6.15224242e-01 -6.14771128e-01 4.58577603e-01
8.64983261e-01 1.08502555e+00 -8.17079067e-01 -5.60487449e-01
-1.00980259e-01 6.07693315e-01 -1.17175020e-01 -1.76025685e-02
-3.12952787e-01 -8.46232653e-01 1.62531331e-01 1.75349668e-01
1.42269418e-01 4.02600080e-01 -6.13724947e-01 -5.76373398e-01
-1.30978063e-01 -8.15556049e-01 7.90401042e-01 -5.30960977e-01
-1.36776686e+00 6.63311958e-01 -2.22185150e-01 -1.56690085e+00
-3.10825527e-01 -6.95727825e-01 -5.24609447e-01 9.01987672e-01
-8.66925359e-01 -9.11887467e-01 -2.08481878e-01 8.48388433e-01
-1.66852653e-01 -6.97945297e-01 1.11319840e+00 1.36132076e-01
-3.50431114e-01 1.05367434e+00 4.89457875e-01 1.22898974e-01
9.67328012e-01 -8.75608563e-01 -3.51813972e-01 3.41497213e-01
-1.35324106e-01 5.24179578e-01 4.67107326e-01 -5.62369704e-01
-8.93014312e-01 -6.85781717e-01 1.40247548e+00 -4.41286623e-01
1.72959507e-01 5.45341671e-01 -7.85070717e-01 2.57559508e-01
-3.29330295e-01 4.38010007e-01 1.47315848e+00 1.89950213e-01
-2.51456738e-01 -3.85230690e-01 -2.06749654e+00 1.76904961e-01
3.78867686e-01 -3.36936861e-01 -6.51859403e-01 3.86564046e-01
-1.48348331e-01 -1.58624589e-01 -1.35770738e+00 4.43100274e-01
1.04271972e+00 -1.14881551e+00 9.94031727e-01 -6.51761830e-01
-3.71796377e-02 -7.19145238e-01 -6.12095296e-01 -9.76834118e-01
-7.09722698e-01 7.10745081e-02 4.50271904e-01 1.10964847e+00
6.34459019e-01 -7.96429276e-01 1.26270080e+00 9.21089172e-01
3.40622425e-01 -9.32749927e-01 -1.24593997e+00 -4.67656136e-01
4.84832972e-02 -3.14683765e-01 8.74072969e-01 1.66457129e+00
1.75996020e-01 -2.54266649e-01 -2.15800092e-01 2.09711924e-01
6.32469118e-01 1.16114579e-01 5.58293104e-01 -1.76108825e+00
-1.73352689e-01 -2.46667877e-01 -1.01684463e+00 5.94984107e-02
9.25087333e-02 -9.15183187e-01 -5.96957684e-01 -1.27943254e+00
1.72315389e-01 -8.62495184e-01 -7.01549590e-01 4.99483317e-01
-6.76543415e-02 5.60160816e-01 -2.51919389e-01 2.42938504e-01
-3.89985219e-02 -2.82208413e-01 1.11183155e+00 1.17710665e-01
-4.80338752e-01 4.32946503e-01 -8.70133817e-01 8.31017733e-01
8.31330836e-01 -5.35825312e-01 -3.41172487e-01 2.52624482e-01
-4.54405211e-02 3.42686176e-01 -5.18391049e-03 -1.11414802e+00
6.76161647e-01 -4.85608399e-01 1.05182612e+00 -8.00822318e-01
2.04172507e-02 -1.11796129e+00 4.94848341e-01 6.46616399e-01
-2.14837626e-01 2.74499245e-02 -1.39395550e-01 2.90832013e-01
-2.96437204e-01 -4.89680588e-01 7.63786972e-01 2.66211659e-01
-6.28476799e-01 -4.11215574e-02 -8.03894103e-02 -4.92015868e-01
1.45278978e+00 -9.26344633e-01 -1.63906828e-01 1.20592704e-02
-8.81193876e-01 -1.97169244e-01 5.51920950e-01 -3.46687138e-02
7.17578709e-01 -1.57050335e+00 -8.05335343e-01 5.36807299e-01
3.02597344e-01 -5.32829046e-01 2.35788122e-01 5.87202847e-01
-5.28616369e-01 4.18470621e-01 -1.68083712e-01 -3.10970217e-01
-1.54007769e+00 3.63606244e-01 2.11059362e-01 -1.21959195e-01
-2.50266910e-01 7.47968316e-01 -5.39037824e-01 -7.20043361e-01
-1.18833661e-01 5.91562353e-02 -7.93799818e-01 2.22462714e-01
5.40549994e-01 9.59529698e-01 4.40896958e-01 -6.04434311e-01
-6.45002484e-01 9.61215079e-01 3.65042724e-02 2.44287699e-01
1.16623878e+00 4.44835544e-01 -4.10113037e-01 3.02263349e-01
9.76462185e-01 2.02924252e-01 -1.59725800e-01 -2.13051707e-01
-1.08586907e-01 -6.44528210e-01 1.59083039e-01 -9.45523620e-01
-9.21633899e-01 1.31350175e-01 1.00779998e+00 7.05481395e-02
9.31037247e-01 -4.50938016e-01 2.05605954e-01 7.44463325e-01
3.94084066e-01 -1.18618250e+00 -8.88837516e-01 3.65871131e-01
4.37691182e-01 -1.21619046e+00 4.33666408e-01 -2.42879003e-01
-4.84738648e-01 1.52572119e+00 2.78649330e-01 -3.78378242e-01
1.19027972e+00 2.89029837e-01 1.66387647e-01 1.10734694e-01
-5.22538126e-01 4.91825670e-01 5.30142844e-01 6.06865168e-01
4.15164351e-01 3.66677225e-01 -8.68086696e-01 8.68871689e-01
-4.31353480e-01 3.99824649e-01 3.38944525e-01 8.09866130e-01
-6.08172059e-01 -1.14280999e+00 -8.09932292e-01 9.34857428e-01
-2.48681605e-01 -9.25778151e-02 -5.79510450e-01 8.76983166e-01
4.33907837e-01 1.14047790e+00 1.22376055e-01 -6.90703392e-01
3.60175043e-01 2.42607310e-01 1.57999665e-01 -7.54881129e-02
-7.58774281e-01 -5.27642429e-01 1.48971844e-02 -3.23324263e-01
-3.02923441e-01 -6.83390856e-01 -1.16367269e+00 -7.06350207e-01
-7.32477188e-01 8.64132047e-01 6.31425500e-01 9.31104064e-01
6.43999055e-02 -1.15042642e-01 9.14485693e-01 -3.80461693e-01
-5.81003666e-01 -5.68083823e-01 -1.07863820e+00 1.25279754e-01
-2.35351305e-02 -5.58412373e-01 -8.33498120e-01 -3.74261826e-01] | [8.381767272949219, 4.331751823425293] |
6e9e274d-7cec-48f1-a8b2-4026ff20be51 | investigation-of-english-to-hindi-multimodal | null | null | https://aclanthology.org/2022.wat-1.15 | https://aclanthology.org/2022.wat-1.15.pdf | Investigation of English to Hindi Multimodal Neural Machine Translation using Transliteration-based Phrase Pairs Augmentation | Machine translation translates one natural language to another, a well-defined natural language processing task. Neural machine translation (NMT) is a widely accepted machine translation approach, but it requires a sufficient amount of training data, which is a challenging issue for low-resource pair translation. Moreover, the multimodal concept utilizes text and visual features to improve low-resource pair translation. WAT2022 (Workshop on Asian Translation 2022) organizes (hosted by the COLING 2022) English to Hindi multimodal translation task where we have participated as a team named CNLP-NITS-PP in two tracks: 1) text-only and 2) multimodal translation. Herein, we have proposed a transliteration-based phrase pairs augmentation approach, which shows improvement in the multimodal translation task. We have attained the second best results on the challenge test set for English to Hindi multimodal translation with BLEU score of 39.30, and a RIBES score of 0.791468. | ['Sivaji Bandyopadhyay', 'Partha Pakray', 'Riyanka Manna', 'Md Faizal Karim', 'Rahul Singh', 'Sahinur Rahman Laskar'] | null | null | null | null | wat-2022-10 | ['transliteration'] | ['natural-language-processing'] | [ 4.27893639e-01 -1.49769828e-01 -2.48156637e-01 -2.14127913e-01
-1.67339563e+00 -8.82365406e-01 9.18628335e-01 -2.33150303e-01
-5.35934150e-01 1.06564558e+00 2.30594710e-01 -6.30416512e-01
6.97685838e-01 -3.59212130e-01 -9.06415939e-01 -3.74505430e-01
7.11998999e-01 1.08438218e+00 -2.94242859e-01 -5.18154144e-01
5.19271456e-02 2.34599095e-02 -7.35092402e-01 8.20154250e-01
9.83753681e-01 5.47866821e-01 2.62315303e-01 6.25149727e-01
-1.11502104e-01 -4.71915416e-02 -2.27320030e-01 -8.69969785e-01
1.78349167e-01 -9.58628297e-01 -1.05893111e+00 -3.31359476e-01
4.52173352e-01 1.31487355e-01 1.53183281e-01 8.20426762e-01
6.95559561e-01 -2.25610703e-01 7.79734015e-01 -1.15389812e+00
-8.11416626e-01 6.62193418e-01 -6.63383305e-01 -2.16383621e-01
5.06266057e-01 3.71223353e-02 1.00489128e+00 -1.68951690e+00
8.99783254e-01 1.15447128e+00 3.87695193e-01 6.97961330e-01
-1.29141605e+00 -4.06609684e-01 -5.85630834e-01 1.64941907e-01
-1.21719527e+00 -4.73340064e-01 1.95255950e-01 -2.08848104e-01
1.15417790e+00 3.88193250e-01 1.48796767e-01 1.41084051e+00
3.35969210e-01 8.71463954e-01 1.46821201e+00 -7.78685927e-01
-2.36572221e-01 3.24753910e-01 -4.07597572e-01 4.03231293e-01
-3.24181706e-01 -1.86567023e-01 -7.59950638e-01 6.26224130e-02
3.73728752e-01 -3.87220979e-01 -9.85309109e-02 1.57067999e-01
-2.15357924e+00 8.63161922e-01 1.78012326e-01 3.13488305e-01
-2.88775951e-01 -3.13520618e-02 5.19591928e-01 8.05895448e-01
5.47672927e-01 5.86499274e-01 -5.03101468e-01 -5.74285865e-01
-7.22234547e-01 7.88368657e-02 6.98242128e-01 1.11421585e+00
6.03123069e-01 -3.17000031e-01 -3.29769105e-01 1.12316704e+00
1.07170649e-01 1.14489520e+00 5.01075804e-01 -4.80744034e-01
1.32374549e+00 4.18024719e-01 1.51391625e-01 -6.37701750e-01
-1.76678911e-01 1.97186377e-02 -8.84643555e-01 -2.89366424e-01
4.13243204e-01 -4.01198953e-01 -8.61004412e-01 1.60113060e+00
-2.82478426e-02 -5.50050914e-01 3.47976059e-01 1.14963305e+00
8.33845973e-01 1.11330688e+00 -1.36681527e-01 -3.16971332e-01
1.20932579e+00 -1.50425041e+00 -5.75650215e-01 -1.56596541e-01
7.28673041e-01 -1.56309569e+00 1.30475211e+00 -3.98991955e-03
-1.34735870e+00 -5.54397583e-01 -7.67275393e-01 -1.70062348e-01
-3.86769176e-01 5.68596482e-01 3.24509978e-01 3.86210114e-01
-1.27446449e+00 2.69736081e-01 -6.12258911e-01 -8.71265054e-01
-2.19480038e-01 5.54339290e-01 -7.61718571e-01 -3.47129256e-01
-1.32176840e+00 1.24132156e+00 1.36104763e-01 2.27488533e-01
-6.14441097e-01 -2.00496390e-01 -6.51807249e-01 -2.84644455e-01
-5.16757593e-02 -7.83628702e-01 1.10803854e+00 -1.02911520e+00
-1.74720681e+00 1.17958522e+00 -4.27105695e-01 -2.90354639e-01
7.00590909e-01 -1.51375696e-01 -4.51174974e-01 1.16101742e-01
2.40784854e-01 1.13936460e+00 7.90312171e-01 -1.01004207e+00
-5.19400835e-01 -3.25927198e-01 -4.30306286e-01 6.35165095e-01
-2.91421115e-01 6.37885749e-01 -4.91394728e-01 -4.84232843e-01
5.69809228e-02 -1.44705701e+00 4.96898107e-02 -5.75596154e-01
-3.54355097e-01 -5.41371070e-02 4.68610287e-01 -1.00471973e+00
7.80614734e-01 -1.85928845e+00 6.60501599e-01 -2.08726183e-01
-3.13121319e-01 1.98564768e-01 -6.75962150e-01 9.14243639e-01
6.43960387e-02 2.04461560e-01 -3.21360260e-01 -7.62029052e-01
2.68738240e-01 3.80226150e-02 -3.49225134e-01 6.27912134e-02
3.88956457e-01 1.47958815e+00 -5.52343488e-01 -4.28906500e-01
-2.23718926e-01 3.00636828e-01 5.15814461e-02 3.29074293e-01
-1.25163704e-01 7.02533364e-01 -3.33532281e-02 7.48721242e-01
7.15053260e-01 -4.16089445e-02 1.01699837e-01 -6.68016374e-02
-2.60067403e-01 1.79843113e-01 -2.44653821e-01 2.10991549e+00
-5.28174520e-01 7.57565856e-01 -1.86443463e-01 -5.56582689e-01
9.58108664e-01 6.89573646e-01 1.64897159e-01 -7.23331630e-01
5.75313345e-02 7.71189332e-01 1.82482472e-04 -4.70828235e-01
6.92295194e-01 -2.90176362e-01 -3.59308898e-01 5.12515128e-01
1.07944556e-01 -3.60008657e-01 3.38502288e-01 1.28002716e-02
8.18319023e-01 4.69349086e-01 -5.55424280e-02 -5.60932048e-02
2.90116817e-01 4.57412720e-01 2.51441866e-01 3.81218135e-01
-7.70602524e-02 1.07993090e+00 1.63856715e-01 -2.93651789e-01
-1.25442040e+00 -9.43375230e-01 2.49946818e-01 1.08860660e+00
-9.40712169e-02 -1.19247243e-01 -7.39287734e-01 -7.07520366e-01
-5.54654419e-01 4.75845337e-01 -3.88914943e-01 1.23665221e-02
-8.66329849e-01 -9.53512669e-01 6.51182830e-01 3.93383056e-01
4.79342133e-01 -1.15993738e+00 1.00057505e-01 2.06417367e-01
-1.06337798e+00 -1.59207046e+00 -8.46626401e-01 2.50437483e-02
-7.14172065e-01 -3.27991903e-01 -1.38174474e+00 -1.26020384e+00
5.53980827e-01 8.37051943e-02 1.17070198e+00 -4.32761878e-01
1.51996985e-01 4.70563211e-02 -6.50568306e-01 -2.33977318e-01
-6.78912580e-01 6.32458925e-01 1.04603119e-01 -6.60994463e-03
5.32136142e-01 -1.10946834e-01 -1.62598774e-01 6.48526609e-01
-4.53877389e-01 5.18222690e-01 9.89404142e-01 1.05294573e+00
4.57836986e-01 -1.04848266e+00 6.61997557e-01 -2.94910729e-01
7.05658853e-01 -3.27492319e-02 -2.91237980e-01 5.84635675e-01
-3.96189898e-01 -3.38361800e-01 5.33374429e-01 -4.57128435e-01
-8.49527240e-01 1.16065003e-01 4.49035019e-02 -1.64007489e-02
-1.19548760e-01 6.99292958e-01 -2.25828007e-01 -1.60081666e-02
5.03926277e-01 4.21934336e-01 -1.70702040e-02 -3.21118772e-01
3.73568594e-01 8.85700166e-01 4.46467876e-01 -5.71826398e-01
9.29366708e-01 -1.81926116e-01 3.26629318e-02 -4.73074973e-01
-2.66226739e-01 -3.06704462e-01 -9.27358627e-01 4.23842296e-02
1.15390682e+00 -1.00554383e+00 -3.01112801e-01 3.24078590e-01
-1.54911280e+00 -3.02196920e-01 3.78791243e-01 8.06604981e-01
-6.40501320e-01 1.53425410e-01 -8.34108889e-01 -3.76812965e-01
-6.31574333e-01 -1.38583732e+00 1.38332987e+00 -1.15752414e-01
-3.72802109e-01 -7.30397642e-01 3.41972321e-01 9.61629748e-01
5.17949522e-01 1.12098798e-01 1.04100823e+00 -6.81113541e-01
-4.32285339e-01 -1.29259691e-01 -5.49075425e-01 1.96180955e-01
-1.51640758e-01 -2.86946207e-01 -5.30036271e-01 -3.83272558e-01
-4.02907699e-01 -6.94100618e-01 6.68674052e-01 -6.95734993e-02
4.00532736e-04 9.53651033e-03 -1.23439617e-01 1.56364501e-01
1.17155111e+00 1.01511970e-01 6.69947267e-01 2.46945962e-01
7.53663123e-01 5.33831000e-01 7.57448912e-01 -1.58752352e-01
5.13430178e-01 1.10823047e+00 8.71680453e-02 -3.59271795e-01
-1.17324419e-01 -2.21289560e-01 8.30145955e-01 1.47366703e+00
-3.22764575e-01 -3.20835948e-01 -1.13343632e+00 6.06152177e-01
-2.14570737e+00 -5.84158897e-01 -3.93156350e-01 2.09297037e+00
1.03869486e+00 -3.95146728e-01 1.48895338e-01 -3.91206145e-01
7.83155203e-01 -4.15377736e-01 8.30284134e-02 -8.57890546e-01
-3.64204228e-01 1.67769164e-01 1.53453976e-01 5.17203927e-01
-8.27821970e-01 1.45357311e+00 5.50384665e+00 7.62477696e-01
-1.20578849e+00 4.25877392e-01 5.07263958e-01 1.58545285e-01
-2.04160333e-01 -6.35876730e-02 -7.33288586e-01 4.09624100e-01
1.29556179e+00 -4.15696343e-03 5.76923132e-01 2.09254026e-01
2.01393619e-01 1.38488427e-01 -1.18783307e+00 9.82273400e-01
3.20505977e-01 -1.19150293e+00 2.27586061e-01 1.49837822e-01
1.11469030e+00 3.48806411e-01 2.08613992e-01 6.00043774e-01
-3.08063865e-01 -1.22039139e+00 4.21866417e-01 2.68422455e-01
1.18410647e+00 -8.46022844e-01 1.14574945e+00 1.88856468e-01
-8.67652297e-01 4.55284476e-01 -3.85999471e-01 2.23042935e-01
3.56385529e-01 1.18973084e-01 -1.05227697e+00 8.27752948e-01
3.03827703e-01 3.83591861e-01 -4.79158551e-01 6.53614283e-01
-1.55351549e-01 6.35367930e-01 -9.85385552e-02 -3.30003947e-01
3.96168381e-01 -4.57236528e-01 5.74997485e-01 1.34660137e+00
6.04012847e-01 -4.15865570e-01 2.78794318e-01 5.12115777e-01
-5.41863203e-01 5.56729794e-01 -5.61781466e-01 -4.67050374e-01
-6.23268671e-02 1.27860367e+00 -5.10189533e-01 -2.19788015e-01
-4.19854552e-01 1.80766714e+00 4.89095449e-02 4.60519582e-01
-9.28576589e-01 -2.72124290e-01 2.05286682e-01 -2.93213308e-01
-4.43585739e-02 -3.60399872e-01 -3.31642270e-01 -1.39230704e+00
1.53962925e-01 -1.24551034e+00 -2.29664639e-01 -8.26846302e-01
-1.08307564e+00 1.08737254e+00 -3.44558477e-01 -1.55710268e+00
-4.51733649e-01 -5.91102779e-01 -2.90668398e-01 1.34122610e+00
-1.17824531e+00 -1.95786965e+00 2.04466105e-01 3.67394477e-01
7.54250824e-01 -5.73395133e-01 1.28598011e+00 5.31852424e-01
-3.21446270e-01 9.13312316e-01 2.86645710e-01 4.82644374e-03
1.42198086e+00 -1.09901035e+00 6.44340992e-01 6.96355641e-01
3.27764004e-01 6.33210778e-01 4.73571777e-01 -5.69936991e-01
-1.74482846e+00 -1.02837074e+00 1.85240054e+00 -7.07961559e-01
6.39922380e-01 -4.57871109e-01 -4.34273720e-01 5.63385665e-01
8.97849977e-01 -5.34116685e-01 7.20805347e-01 -9.17788297e-02
-1.39632717e-01 2.09129676e-01 -9.56499100e-01 8.99260998e-01
5.87283254e-01 -5.84541619e-01 -5.28164327e-01 7.32395291e-01
8.80304217e-01 -4.52372611e-01 -8.39136481e-01 4.68936622e-01
5.58190823e-01 -2.15732843e-01 5.24801970e-01 -7.48358786e-01
1.02038848e+00 -3.96089703e-01 -5.13287306e-01 -1.51252294e+00
1.78381532e-01 -1.04763198e+00 4.53400999e-01 1.16863871e+00
1.23861063e+00 -3.35747719e-01 5.26932001e-01 3.46642584e-01
-5.78827932e-02 -6.03590846e-01 -1.03586078e+00 -8.09977293e-01
4.72856373e-01 -9.41693486e-07 3.00007850e-01 1.00363100e+00
3.17873418e-01 1.08024251e+00 -7.56223798e-01 -2.88858414e-01
1.73894584e-01 3.10893029e-01 9.66161370e-01 -5.72606742e-01
-2.79730111e-01 -3.56497258e-01 -8.83386284e-02 -8.67468596e-01
5.50187081e-02 -1.22119641e+00 1.85864165e-01 -1.63720214e+00
6.51552975e-01 3.29599798e-01 1.22656196e-01 5.92564285e-01
-2.46881768e-01 8.31249535e-01 3.66050363e-01 4.25289303e-01
-4.98860955e-01 5.46386778e-01 1.44088566e+00 -2.65050441e-01
4.34276536e-02 -1.25748785e-02 -2.59745270e-01 -3.39132212e-02
8.78409863e-01 -3.84408653e-01 6.76336605e-03 -8.60564590e-01
3.31532538e-01 2.26223975e-01 -4.58831973e-02 -5.65573275e-01
-5.69918659e-03 -1.72433004e-01 2.73995966e-01 -6.45693362e-01
5.23781240e-01 -7.09512234e-01 -1.76451728e-02 2.61263698e-01
-4.37810123e-01 8.31125379e-01 2.18122393e-01 -3.45219932e-02
-5.16467094e-01 1.28613161e-02 4.98801887e-01 -1.24479115e-01
-2.54772663e-01 1.35301039e-01 -3.65386933e-01 -2.87361834e-02
7.01420903e-01 8.56960490e-02 -3.55939090e-01 -4.63367313e-01
-6.37380540e-01 1.24722168e-01 4.70121413e-01 6.71400845e-01
5.61006308e-01 -1.62029505e+00 -1.37798607e+00 -1.29615903e-01
3.60918969e-01 -7.77953923e-01 -1.63322270e-01 1.26354957e+00
-5.40213168e-01 5.43864429e-01 -4.05098140e-01 -7.38622069e-01
-1.38058484e+00 2.02693745e-01 -1.49797857e-01 -4.69376296e-01
-6.52399436e-02 7.40188956e-01 -3.88810664e-01 -1.07905543e+00
-2.67899513e-01 1.57598644e-01 1.57585055e-01 -8.45229477e-02
4.49197918e-01 1.87046170e-01 2.32568637e-01 -1.01114571e+00
-3.27287912e-01 7.63565779e-01 -1.10011719e-01 -9.23782527e-01
9.90159690e-01 -1.73016757e-01 -6.05564058e-01 3.75952303e-01
1.48091638e+00 -9.93495435e-02 -3.28136265e-01 -1.45963982e-01
1.14416361e-01 -1.17298067e-01 -5.83457887e-01 -1.40990305e+00
-4.19569582e-01 1.28146291e+00 5.74824095e-01 -3.66957307e-01
8.60845506e-01 -3.22213411e-01 1.24333310e+00 5.89675725e-01
4.40247476e-01 -1.16118407e+00 -1.08615071e-01 1.02995503e+00
1.27929163e+00 -1.73618889e+00 -5.19646287e-01 -1.84772968e-01
-9.98474181e-01 1.23674369e+00 5.17303109e-01 4.81020600e-01
-1.83252960e-01 -3.22594643e-02 6.46740735e-01 4.36666071e-01
-7.17127383e-01 -1.14093656e-02 6.22837722e-01 5.17136574e-01
7.55698323e-01 3.07165653e-01 -7.71049321e-01 3.03396553e-01
-1.49725348e-01 -1.56292498e-01 2.09180657e-02 6.21437967e-01
-8.57435465e-02 -1.56866145e+00 -4.86730754e-01 -1.17920883e-01
-4.09398228e-01 -5.55809200e-01 -8.97152781e-01 5.85058868e-01
-2.69432962e-01 1.14662325e+00 -3.72695833e-01 -6.86571836e-01
2.10349694e-01 2.96193004e-01 6.52932644e-01 -5.75263441e-01
-9.56221521e-01 2.93529958e-01 3.45982909e-01 -2.71925390e-01
-2.73441523e-01 -4.93670136e-01 -1.07254410e+00 -2.85526246e-01
-4.39578481e-02 3.81424874e-01 1.19631720e+00 9.91295218e-01
4.51189995e-01 -3.10552064e-02 6.00448072e-01 -8.59521925e-01
-4.42750752e-01 -1.34976614e+00 1.27702370e-01 2.83087492e-01
-1.47829741e-01 1.16195045e-01 5.73949777e-02 1.82762995e-01] | [11.493674278259277, 1.5339449644088745] |
77deba4c-086d-497e-b1cf-0b1bde02f4fa | local-radon-descriptors-for-image-search | 1710.04097 | null | http://arxiv.org/abs/1710.04097v1 | http://arxiv.org/pdf/1710.04097v1.pdf | Local Radon Descriptors for Image Search | Radon transform and its inverse operation are important techniques in medical
imaging tasks. Recently, there has been renewed interest in Radon transform for
applications such as content-based medical image retrieval. However, all
studies so far have used Radon transform as a global or quasi-global image
descriptor by extracting projections of the whole image or large sub-images.
This paper attempts to show that the dense sampling to generate the histogram
of local Radon projections has a much higher discrimination capability than the
global one. In this paper, we introduce Local Radon Descriptor (LRD) and apply
it to the IRMA dataset, which contains 14,410 x-ray images as well as to the
INRIA Holidays dataset with 1,990 images. Our results show significant
improvement in retrieval performance by using LRD versus its global version. We
also demonstrate that LRD can deliver results comparable to well-established
descriptors like LBP and HOG. | ['Morteza Babaie', 'Amin Khatami', 'M. E. Shiri', 'H. R. Tizhoosh'] | 2017-10-11 | null | null | null | null | ['medical-image-retrieval', 'medical-image-retrieval'] | ['computer-vision', 'medical'] | [ 2.91469336e-01 -4.72724587e-01 -7.16172457e-02 -4.03413355e-01
-1.34830022e+00 -1.16063058e-01 7.90286422e-01 3.33613485e-01
-6.71089530e-01 5.06608725e-01 3.91792119e-01 -3.81215326e-02
-4.21916664e-01 -1.15911305e+00 -3.44932854e-01 -9.77493465e-01
-1.81290165e-01 2.69062072e-01 4.24144506e-01 -7.50772431e-02
4.17548299e-01 8.10470402e-01 -1.35454440e+00 2.52612591e-01
1.94592968e-01 8.67967844e-01 2.18669847e-01 5.46322882e-01
1.75338909e-01 5.41284204e-01 -5.59830189e-01 3.06008279e-01
3.39962542e-01 -7.21325278e-01 -9.02821600e-01 -1.40913591e-01
3.92985672e-01 -3.61297518e-01 -3.14603657e-01 9.16892648e-01
9.38097417e-01 3.00652385e-01 7.52715647e-01 -4.17223185e-01
-8.31254721e-01 -1.04246214e-01 -8.29766393e-01 6.02410972e-01
6.49292588e-01 -4.50510681e-01 7.53302693e-01 -9.07933414e-01
9.34358180e-01 1.15550828e+00 7.54850328e-01 9.70372111e-02
-9.65882838e-01 -1.98740497e-01 -8.08901429e-01 2.11511359e-01
-1.55482066e+00 -8.84545892e-02 5.33028245e-01 -1.79666519e-01
7.36886799e-01 5.41327596e-01 2.49121949e-01 4.41627592e-01
7.42589116e-01 3.76053602e-01 1.86253953e+00 -7.91794002e-01
5.15510105e-02 -8.99836943e-02 2.88526446e-01 9.54851508e-01
9.00603905e-02 5.50010093e-02 -3.44791800e-01 -4.30229098e-01
9.44773793e-01 2.88594902e-01 -3.44880551e-01 -1.24420412e-01
-1.22695768e+00 8.97326589e-01 5.90712190e-01 9.47551250e-01
-6.72663629e-01 2.12452170e-02 2.68361330e-01 3.82015318e-01
5.91953754e-01 1.97935149e-01 1.45273358e-01 1.26409695e-01
-8.40517819e-01 1.02224290e-01 4.96429116e-01 2.01774269e-01
6.32651746e-01 -5.85993528e-01 -3.44835788e-01 1.12843108e+00
4.49471511e-02 6.83707952e-01 8.16592991e-01 -5.64833820e-01
-3.47817764e-02 1.40348271e-01 -2.19484329e-01 -1.43453860e+00
-3.21633369e-01 -6.73510954e-02 -7.28546798e-01 8.85162577e-02
2.35783249e-01 6.93331242e-01 -1.06593859e+00 1.16408670e+00
2.27863267e-01 -1.55863732e-01 6.53253645e-02 7.80174613e-01
9.33951318e-01 5.85486710e-01 -1.31887019e-01 -3.33995037e-02
1.70779335e+00 -7.51481414e-01 -5.98319411e-01 3.22360963e-01
4.12963986e-01 -1.06817317e+00 7.21915722e-01 3.29324633e-01
-8.67486894e-01 -6.16573334e-01 -8.39889586e-01 -1.67562604e-01
-5.15897870e-01 1.02466688e-01 2.76459664e-01 7.06861854e-01
-1.09320402e+00 4.97789264e-01 -6.86342835e-01 -9.04979646e-01
1.00509837e-01 1.83099866e-01 -6.63521409e-01 -5.07568300e-01
-9.90464747e-01 9.77317452e-01 8.00520629e-02 -8.50205496e-02
-2.73393124e-01 -3.16105872e-01 -7.39664674e-01 -1.75964445e-01
-2.27472186e-01 -2.45940477e-01 5.84014654e-01 -6.64501488e-02
-1.20728469e+00 1.25488997e+00 2.74099968e-02 -4.28727597e-01
3.40521693e-01 5.02417050e-02 -3.56140584e-01 8.06023300e-01
1.19785950e-01 4.92122829e-01 6.36000156e-01 -8.09360981e-01
-2.91968018e-01 -5.59953272e-01 -3.42770606e-01 1.22486852e-01
-1.69823319e-02 2.68451989e-01 -3.67362440e-01 -6.79883718e-01
5.15427887e-01 -9.71194804e-01 -5.74794970e-02 -1.34046534e-02
-9.19437706e-02 -2.20552355e-01 8.63710165e-01 -8.79434586e-01
9.18707311e-01 -2.25024390e+00 -2.69859612e-01 5.04104376e-01
-8.68422911e-02 9.52085033e-02 -2.08657846e-01 6.42573416e-01
-2.71014720e-01 -8.84808898e-02 -2.97175497e-01 1.61620975e-01
-3.67096722e-01 2.99333662e-01 -1.68718621e-01 9.39264357e-01
-2.74432719e-01 9.10404682e-01 -7.33185947e-01 -8.28700185e-01
4.86538351e-01 8.51305783e-01 -4.35990207e-02 -3.86959314e-01
8.64775598e-01 3.91214877e-01 -6.32871985e-01 6.08381331e-01
8.57578337e-01 -9.07193031e-03 -9.10063386e-02 -2.78372914e-01
9.07414127e-03 -1.40534833e-01 -6.28742278e-01 1.51729369e+00
-4.95373189e-01 5.98334253e-01 -3.71486843e-01 -1.08001864e+00
1.16088450e+00 2.43842706e-01 9.86350358e-01 -1.30412114e+00
-6.92080855e-02 3.84716004e-01 -6.96613073e-01 -4.15321529e-01
3.37060422e-01 -5.43154299e-01 -6.34198114e-02 5.15946329e-01
-2.23456845e-01 -1.09088853e-01 1.42617911e-01 -2.48919845e-01
1.34957206e+00 -1.93425462e-01 7.36054242e-01 -5.90803206e-01
8.44970882e-01 -1.24212265e-01 -1.39322028e-01 7.89691091e-01
-1.03548385e-01 1.04148316e+00 -5.35281049e-03 -5.33086836e-01
-6.61788464e-01 -1.18479717e+00 -7.65657008e-01 5.79729855e-01
1.45845771e-01 -1.18029810e-01 -5.05982459e-01 -6.16414547e-01
-1.64951652e-01 1.65548772e-01 -6.58648074e-01 -5.03562726e-02
-7.89505184e-01 -9.16520000e-01 4.79769707e-01 2.60295868e-01
8.98200154e-01 -1.23322034e+00 -9.73098397e-01 1.57883181e-03
-2.68101811e-01 -9.98494864e-01 -5.20107448e-01 5.75055033e-02
-1.02048802e+00 -1.00458729e+00 -1.14155948e+00 -9.41084981e-01
8.32857609e-01 4.78404909e-01 9.42501545e-01 8.18471983e-02
-9.64039087e-01 7.34930634e-01 -5.58839440e-01 1.19386323e-01
-3.15577537e-01 -3.34586054e-01 -3.93975794e-01 -2.00904414e-01
3.78121465e-01 -1.30742699e-01 -7.55330443e-01 3.82120550e-01
-1.22078705e+00 -6.22747242e-01 7.52583325e-01 9.32811379e-01
9.01005447e-01 2.94287086e-01 9.59223136e-02 -9.49226975e-01
7.64150858e-01 3.36412503e-03 -3.25136125e-01 2.86944300e-01
-4.80957776e-01 1.68534294e-01 1.69670746e-01 1.87329918e-01
-9.78995621e-01 -1.57085434e-01 -3.04930210e-01 -4.55535436e-03
-3.40468362e-02 5.51058948e-01 6.15962625e-01 -4.01178360e-01
7.62826681e-01 4.43813205e-01 1.53671890e-01 -5.34984648e-01
2.16048732e-02 6.97025955e-01 5.47657669e-01 -5.03124237e-01
6.44235432e-01 9.62444663e-01 3.95362765e-01 -1.53834963e+00
-3.74411345e-01 -1.19858313e+00 -5.15433073e-01 -4.14429000e-03
1.05687106e+00 -5.82508147e-01 -3.33972633e-01 2.91469336e-01
-7.15378225e-01 1.01403587e-01 -5.72909832e-01 6.75549269e-01
-7.18915343e-01 6.53763473e-01 -7.26459026e-01 -5.07102191e-01
-4.41798598e-01 -1.08631134e+00 1.46798539e+00 -5.73771633e-02
4.50301804e-02 -1.09147787e+00 5.42879641e-01 2.70205975e-01
6.64411128e-01 5.40960789e-01 6.94437921e-01 -3.50214303e-01
-3.67971212e-01 -4.75244731e-01 -4.42100286e-01 4.19428051e-01
3.28584641e-01 -5.12058318e-01 -9.01417255e-01 -4.13772911e-01
4.03989553e-01 -1.94400504e-01 1.19130492e+00 5.66037476e-01
1.07705784e+00 1.16627611e-01 -3.70087862e-01 3.31102997e-01
1.76548398e+00 3.05046111e-01 1.13541925e+00 4.25231844e-01
9.23881009e-02 4.71499443e-01 8.26988161e-01 2.12040856e-01
-1.93788409e-01 7.34069645e-01 -2.01257914e-01 -3.75923395e-01
-3.93977195e-01 -6.37052208e-02 1.19273350e-01 8.41211259e-01
-3.44017774e-01 1.42274186e-01 -8.11744750e-01 3.74506414e-01
-1.44944680e+00 -1.08670509e+00 1.74755510e-03 2.44497061e+00
4.72483069e-01 -4.37671542e-01 -2.34305710e-01 2.36366853e-01
6.63427472e-01 4.24082637e-01 3.48165512e-01 -2.76638776e-01
3.10998354e-02 8.50194514e-01 7.09381580e-01 4.24443156e-01
-1.24403870e+00 3.77529889e-01 6.87176132e+00 1.08710515e+00
-1.26061237e+00 2.27352500e-01 7.53754497e-01 7.50923991e-01
4.76243272e-02 -1.51894212e-01 -3.33399236e-01 5.38365431e-02
7.42590547e-01 1.35290787e-01 -3.74311283e-02 5.12555838e-01
8.32253471e-02 -5.55558503e-01 -6.34123683e-01 1.20593214e+00
4.55432326e-01 -8.85831177e-01 -3.26544642e-02 6.05810761e-01
6.57704890e-01 3.34346038e-03 2.55027145e-01 -3.24792564e-02
-4.02409226e-01 -1.11873615e+00 1.42353505e-01 7.04499960e-01
9.15849090e-01 -6.44840717e-01 9.69613433e-01 9.05107483e-02
-1.17670882e+00 2.88034409e-01 -5.68373322e-01 5.01980484e-01
-1.47028118e-01 6.79629922e-01 -8.16321909e-01 8.14644754e-01
7.16711938e-01 5.27949154e-01 -7.24159420e-01 1.17137814e+00
2.39211768e-01 1.36161432e-01 -5.20744264e-01 2.39104435e-01
3.82142633e-01 -5.21981835e-01 3.29846919e-01 1.37545741e+00
5.77407300e-01 1.26029849e-01 1.21790402e-01 2.57027209e-01
1.41707107e-01 5.81395030e-01 -9.14880157e-01 1.40447304e-01
-1.32045865e-01 1.12296295e+00 -1.16215551e+00 -3.44372392e-01
-3.49724859e-01 1.08776176e+00 -2.92014420e-01 1.81448996e-01
-5.58910847e-01 -5.00156462e-01 -9.62338746e-02 1.10094793e-01
3.11745524e-01 -1.94128424e-01 4.20682877e-01 -6.63262546e-01
-1.80495724e-01 -6.11595750e-01 6.09573960e-01 -6.70308352e-01
-1.20862329e+00 6.53643489e-01 2.63217062e-01 -1.35182881e+00
-1.45868808e-01 -6.34371579e-01 -3.35920841e-01 7.92065144e-01
-1.64039278e+00 -1.02409840e+00 -4.05883878e-01 5.60117900e-01
2.66962618e-01 1.25684932e-01 9.02690113e-01 3.61179352e-01
3.02063823e-01 1.25864461e-01 5.09117365e-01 2.12955266e-01
9.98192906e-01 -1.06875646e+00 -2.52369195e-01 4.26561147e-01
3.73660803e-01 7.58087635e-01 5.19388318e-01 -3.44399244e-01
-1.19741654e+00 -7.37049103e-01 9.61349487e-01 -2.93921113e-01
1.56027675e-01 3.93277466e-01 -6.80226445e-01 3.25001210e-01
1.60994306e-01 5.47545493e-01 5.36712587e-01 -5.31087518e-01
-1.45748496e-01 -2.90706724e-01 -1.54107833e+00 -7.70972017e-03
5.59982479e-01 -8.23161066e-01 -6.01478636e-01 5.30206084e-01
-9.27775130e-02 -3.49842250e-01 -1.23066258e+00 3.54732752e-01
5.79031408e-01 -1.35554409e+00 1.53379905e+00 2.33020201e-01
2.62695819e-01 -2.05844015e-01 -4.53658223e-01 -8.69273126e-01
-1.64812580e-01 -1.36666587e-02 6.29008412e-01 7.08981454e-01
-2.61379421e-01 -8.87011647e-01 4.61618543e-01 -1.71602145e-01
2.37986401e-01 -7.48200297e-01 -1.15159571e+00 -1.01452589e+00
-3.72308120e-02 8.78999978e-02 2.20088214e-01 7.11233199e-01
-2.60938227e-01 -6.88636750e-02 -1.26701206e-01 -3.25810075e-01
6.88180983e-01 2.97911257e-01 4.38120544e-01 -9.10669923e-01
-2.43577585e-01 -3.04001898e-01 -9.03560877e-01 -7.58893609e-01
-2.96059221e-01 -9.76058900e-01 1.88252553e-01 -1.69842994e+00
4.67354536e-01 -3.93748850e-01 -6.79290712e-01 2.50605077e-01
1.25878766e-01 1.11788094e+00 -3.22252028e-02 4.39772606e-01
-3.88633072e-01 1.54440254e-01 1.43816257e+00 -1.82590216e-01
5.19674532e-02 -1.13185853e-01 -2.69628257e-01 4.40341622e-01
6.27056897e-01 -5.51828325e-01 -1.33053258e-01 1.25897348e-01
-2.35805020e-01 1.31324068e-01 3.97590250e-01 -1.20876408e+00
7.12860897e-02 3.43435168e-01 4.62704241e-01 -6.99006021e-01
5.94488502e-01 -7.95313239e-01 6.31898418e-02 8.60449076e-01
-2.25982100e-01 2.45246723e-01 -2.14995801e-01 7.02987909e-01
-7.50308990e-01 -5.17519057e-01 9.81097758e-01 -4.51628476e-01
-6.83007181e-01 9.36756879e-02 -1.63615495e-01 -3.28500539e-01
1.13287807e+00 -3.42665881e-01 -2.66885936e-01 -1.36751875e-01
-5.35627246e-01 -4.89758879e-01 3.49703580e-01 1.35677576e-01
9.59713578e-01 -1.31053913e+00 -6.25725687e-01 3.28548819e-01
2.70583808e-01 -5.90708673e-01 2.23858744e-01 1.13078964e+00
-1.21790993e+00 7.95972764e-01 -3.44371915e-01 -7.02972412e-01
-1.74217093e+00 4.84743536e-01 -2.12045666e-02 -6.41633451e-01
-1.11305332e+00 4.03919399e-01 1.48047239e-01 -2.84980744e-01
-4.23244804e-01 -3.53625923e-01 -3.40175033e-01 1.05786212e-01
5.79771101e-01 4.61289853e-01 3.02350581e-01 -9.47484612e-01
-5.84611773e-01 1.32766807e+00 1.22329630e-01 -2.92587042e-01
1.43786252e+00 5.14527969e-02 -5.34977973e-01 2.42275551e-01
1.90118575e+00 3.11691105e-01 -2.21068189e-01 -1.75836310e-01
8.06937739e-02 -7.94799328e-01 2.91580528e-01 -3.33588511e-01
-9.60804164e-01 8.15975845e-01 1.05923855e+00 2.67588854e-01
1.26083231e+00 2.64886886e-01 8.35788369e-01 4.89152133e-01
6.10919893e-01 -8.90693665e-01 1.79843426e-01 1.93716675e-01
9.48016047e-01 -1.03212154e+00 5.51578343e-01 -3.34682703e-01
-2.37922251e-01 1.28020203e+00 -3.72845739e-01 -4.35026497e-01
8.10109496e-01 -1.19446613e-01 1.36934593e-01 -4.88752365e-01
3.57101262e-02 -9.76572484e-02 4.14863884e-01 4.97419387e-01
5.08755565e-01 -9.30695795e-03 -8.20133209e-01 -5.07841527e-01
1.47721201e-01 4.72726263e-02 1.69337884e-01 1.34925866e+00
-5.09673297e-01 -1.24516082e+00 -9.11130965e-01 5.43387353e-01
-6.41445994e-01 1.00382932e-01 9.52247158e-02 1.02651966e+00
-2.12205142e-01 4.87004280e-01 -4.18451801e-02 1.81890115e-01
1.28834173e-01 -1.34814354e-02 8.50397289e-01 -3.31456572e-01
-3.51024300e-01 2.42877483e-01 -1.21711940e-01 -6.57320082e-01
-9.06034887e-01 -5.97967505e-01 -1.31410670e+00 7.97965974e-02
-2.17743978e-01 2.07761392e-01 8.49117696e-01 6.57132745e-01
-2.02944264e-01 4.29874510e-01 5.41933000e-01 -6.23563945e-01
-6.60138845e-01 -7.08398581e-01 -9.29199696e-01 5.26035488e-01
2.95963109e-01 -5.19110739e-01 -3.20154101e-01 -1.05013192e-01] | [14.2034273147583, -1.3899301290512085] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.