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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
26bdd9b2-ad0a-443a-9777-5a2bc531c4b5 | can-everybody-sign-now-exploring-sign | 2012.10941 | null | https://arxiv.org/abs/2012.10941v2 | https://arxiv.org/pdf/2012.10941v2.pdf | Can Everybody Sign Now? Exploring Sign Language Video Generation from 2D Poses | Recent work have addressed the generation of human poses represented by 2D/3D coordinates of human joints for sign language. We use the state of the art in Deep Learning for motion transfer and evaluate them on How2Sign, an American Sign Language dataset, to generate videos of signers performing sign language given a 2D pose skeleton. We evaluate the generated videos quantitatively and qualitatively showing that the current models are not enough to generated adequate videos for Sign Language due to lack of detail in hands. | ['Xavier Giro-i-Nieto', 'Amanda Duarte', 'Lucas Ventura'] | 2020-12-20 | null | null | null | null | ['sign-language-production'] | ['natural-language-processing'] | [-2.72730827e-01 -3.94866476e-03 -5.51234782e-02 -1.40964046e-01
-5.87410212e-01 -6.12532794e-01 6.99451447e-01 -1.45087099e+00
-5.15699029e-01 7.09538400e-01 6.84772015e-01 -2.11188614e-01
2.13058740e-01 -2.15217456e-01 -8.30857515e-01 -5.02818763e-01
-1.75307870e-01 4.62331682e-01 5.19396126e-01 -3.99461180e-01
-2.78987829e-02 6.62835717e-01 -1.43207312e+00 4.04341698e-01
2.75447398e-01 1.68799624e-01 -4.60251778e-01 1.33930326e+00
3.20855230e-01 9.73789454e-01 -8.00226986e-01 -3.28910679e-01
5.83136201e-01 -1.06563938e+00 -5.71988881e-01 -1.28435239e-01
1.27415311e+00 -1.03995788e+00 -9.75205243e-01 6.51155710e-01
8.64238143e-01 -1.06435448e-01 1.01218975e+00 -1.24743664e+00
-4.52768028e-01 3.40445489e-01 -3.00458193e-01 -3.15104574e-01
6.88599527e-01 7.92250395e-01 7.06959903e-01 -8.90129685e-01
1.60098028e+00 1.39047277e+00 7.38644838e-01 1.21191859e+00
-5.47213495e-01 -5.71680367e-01 -1.95451379e-01 1.46891654e-01
-9.08978283e-01 -2.01745868e-01 6.04142010e-01 -6.58240199e-01
1.03307354e+00 -9.09702331e-02 1.53979588e+00 1.50246739e+00
6.61044791e-02 1.19329476e+00 6.59455895e-01 -4.07270551e-01
-4.50250916e-02 -9.17325258e-01 -1.26431197e-01 7.87575245e-01
3.75540763e-01 4.27544206e-01 -6.52973235e-01 6.30008057e-02
1.39838398e+00 -5.99236727e-01 -3.04997265e-01 -6.30210698e-01
-1.47851336e+00 5.62996924e-01 4.31018591e-01 1.10346682e-01
-5.28477430e-01 8.51708591e-01 2.71562338e-01 2.07378432e-01
-4.47843045e-01 1.50514051e-01 -1.29635692e-01 -5.17563045e-01
-8.83370340e-01 8.85566711e-01 8.96275103e-01 1.16429293e+00
-2.47843400e-01 4.96121764e-01 -4.07236189e-01 3.28357592e-02
6.61679864e-01 8.18708777e-01 4.85871971e-01 -1.34189391e+00
6.11940503e-01 3.22042465e-01 3.86524200e-01 -3.69832993e-01
-2.87787557e-01 2.76537448e-01 -2.63773739e-01 9.68675673e-01
1.10216808e+00 -5.30694842e-01 -1.76733184e+00 1.30812311e+00
-3.49252857e-02 -1.05618071e-02 1.39919475e-01 1.39427030e+00
9.42349672e-01 2.42493719e-01 2.66578585e-01 5.87636471e-01
9.13390219e-01 -1.00555074e+00 -6.68910682e-01 1.12536907e-01
4.91652578e-01 -7.88618982e-01 1.11059523e+00 3.86621714e-01
-1.38327515e+00 -3.62898290e-01 -7.94084549e-01 -3.06453437e-01
-4.33076248e-02 4.50302422e-01 6.08869076e-01 4.49980736e-01
-8.77123594e-01 5.44703841e-01 -1.08880305e+00 -5.50366879e-01
4.76459861e-01 1.59585968e-01 -5.56677580e-01 1.37866307e-02
-1.04845965e+00 9.29857850e-01 -8.47097021e-03 2.06214622e-01
-8.58806014e-01 -5.03657579e-01 -8.55524838e-01 -4.75114673e-01
-4.35097694e-01 -1.10360718e+00 1.65022349e+00 -7.36420572e-01
-1.66463315e+00 1.12873423e+00 1.90130115e-01 -4.43429023e-01
1.46108723e+00 -4.67415631e-01 -5.98310754e-02 5.22171795e-01
-4.43333358e-01 1.19623232e+00 8.55769098e-01 -1.02770412e+00
-3.47700417e-01 4.77643264e-03 -3.65425795e-02 3.62802818e-02
3.85314107e-01 2.66403019e-01 -2.89710492e-01 -8.32011402e-01
1.02898657e-01 -1.14427185e+00 -1.50778875e-01 7.71132350e-01
-2.77558386e-01 -1.36520267e-02 9.94725466e-01 -1.11724269e+00
6.72972441e-01 -1.44159210e+00 4.39529032e-01 1.82887197e-01
-1.86947674e-01 6.72734022e-01 -4.59418297e-01 4.29574519e-01
2.07259953e-01 -1.28651345e-02 -7.22363442e-02 -6.42255023e-02
1.75539091e-01 3.36291045e-01 -3.38062376e-01 4.37737226e-01
1.70811296e-01 1.30916286e+00 -1.02513266e+00 -4.05593038e-01
2.81562656e-01 1.06854808e+00 -5.98385870e-01 1.63169861e-01
-2.73203582e-01 6.67665124e-01 -4.40896809e-01 9.31420982e-01
3.08822036e-01 2.78958738e-01 -1.62006135e-03 -4.03726876e-01
1.70916557e-01 -1.00260740e-03 -9.87785041e-01 1.75492156e+00
8.28613639e-02 8.22073042e-01 -2.93204729e-02 -2.49959260e-01
3.99159849e-01 6.30867600e-01 4.62649912e-01 -1.99587673e-01
2.87506074e-01 6.69646442e-01 4.04767752e-01 -7.39641130e-01
-3.62059027e-02 -1.05651524e-02 1.46725744e-01 3.75683933e-01
9.92196277e-02 -6.73984051e-01 4.30943310e-01 -1.38834000e-01
1.02852023e+00 9.57293391e-01 -9.73640289e-03 1.92408368e-01
2.38388598e-01 2.68806040e-01 3.81248668e-02 5.56873262e-01
-4.23086524e-01 1.17690992e+00 3.24330598e-01 -5.87151289e-01
-1.52792788e+00 -1.39759541e+00 5.61827421e-01 3.26362222e-01
-4.88547832e-01 1.90696437e-02 -8.63079071e-01 -6.31871998e-01
2.32898563e-01 3.85983884e-01 -3.33600700e-01 2.01354593e-01
-1.30456364e+00 2.77297269e-03 1.12240016e+00 1.12070298e+00
3.22753519e-01 -1.51278961e+00 -1.22345173e+00 -8.82021114e-02
1.72778890e-01 -1.05649412e+00 -7.40235150e-01 -7.07913578e-01
-8.37796330e-01 -1.40800238e+00 -1.73970735e+00 -1.06198788e+00
8.00178468e-01 -3.01755339e-01 8.09947193e-01 -1.60373345e-01
-5.65858841e-01 6.31052494e-01 -4.17541593e-01 -2.27778360e-01
-7.83080518e-01 -4.40422297e-01 3.87019813e-02 -5.65714657e-01
1.36214420e-01 -5.88146389e-01 -8.33660841e-01 3.24973375e-01
-8.63624632e-01 1.71797752e-01 6.18190706e-01 7.41744459e-01
7.14825168e-02 -1.08782971e+00 1.04363775e-02 4.46533002e-02
5.44978499e-01 4.26504672e-01 -5.51589370e-01 1.19087175e-01
4.01426882e-01 1.46751463e-01 1.59719795e-01 -5.61531425e-01
-7.90336609e-01 6.61190808e-01 -2.78585702e-01 -4.63539571e-01
-1.49170831e-01 -1.99161447e-03 1.52116209e-01 -3.46113592e-01
8.70294213e-01 1.44369930e-01 3.42326075e-01 -3.04605424e-01
7.80865550e-01 3.33638728e-01 8.79120350e-01 -4.39791650e-01
8.64870250e-01 6.32121801e-01 3.08223993e-01 -6.10881090e-01
-3.04745194e-02 -8.90515968e-02 -1.19832051e+00 -4.29696500e-01
1.08508754e+00 -7.61155486e-01 -8.59978855e-01 8.80213261e-01
-1.42924511e+00 -5.32459378e-01 -5.48221767e-01 7.87324071e-01
-1.16582584e+00 5.30709088e-01 -8.13273489e-01 -6.20096028e-01
-3.90050471e-01 -1.00810337e+00 1.37513626e+00 -2.45943218e-02
-6.24005258e-01 -5.10632098e-01 3.49627614e-01 4.83483411e-02
2.80124098e-01 7.53357828e-01 5.48192918e-01 1.61502704e-01
-7.57876992e-01 -5.55225551e-01 4.17917967e-02 4.05728489e-01
2.02408503e-03 4.60402250e-01 -5.84258437e-01 -3.56859624e-01
-1.12327051e+00 -6.55696630e-01 9.06857431e-01 5.72411180e-01
3.48230660e-01 -2.82295793e-01 3.66481114e-03 6.17615104e-01
7.24624455e-01 6.95382729e-02 8.79589558e-01 -2.36860234e-02
7.90101767e-01 5.87013483e-01 2.73908883e-01 3.62404168e-01
1.10153235e-01 7.49550939e-01 -6.15414046e-03 2.77926796e-03
-1.12802982e+00 -7.37115085e-01 6.51754320e-01 5.34514189e-01
-9.41644192e-01 7.32511356e-02 -1.03745174e+00 6.25800312e-01
-1.84356356e+00 -1.02771246e+00 -2.73530394e-01 1.85539031e+00
6.93262637e-01 -2.88837582e-01 6.77022219e-01 2.46509969e-01
2.97868311e-01 -1.95601001e-01 -5.23140907e-01 -2.70393267e-02
-2.49464765e-01 2.99327016e-01 4.37600315e-01 5.82674503e-01
-8.24133754e-01 1.32499480e+00 7.71170902e+00 -8.20221305e-02
-1.17687559e+00 -4.17593688e-01 -4.08230901e-01 -2.13513121e-01
-6.89036101e-02 -3.42125334e-02 -5.97074986e-01 5.85012920e-02
4.00051713e-01 -1.19405530e-01 -3.05881687e-02 6.53480828e-01
3.63384783e-01 2.75994748e-01 -1.21408033e+00 9.71722782e-01
1.65099263e-01 -1.16609967e+00 6.29768610e-01 4.83301915e-02
1.00150096e+00 4.82062949e-03 -1.50123954e-01 5.37000224e-02
5.95184863e-01 -1.00893426e+00 1.01345408e+00 9.02468741e-01
9.22935128e-01 -2.04078242e-01 4.46211934e-01 2.19453499e-02
-9.74854708e-01 1.98726833e-01 1.19606689e-01 -4.27931584e-02
8.26134026e-01 -4.46834981e-01 -8.37449312e-01 1.35706365e-01
3.80033582e-01 6.67367220e-01 -4.71186221e-01 1.37640476e+00
-7.61080027e-01 6.43145323e-01 -4.88518715e-01 -1.11060373e-01
4.67219770e-01 1.78650096e-01 7.81423628e-01 1.22563052e+00
6.96137905e-01 -2.82784432e-01 -8.67972597e-02 6.80065453e-01
3.55795711e-01 -3.44771855e-02 -7.99278259e-01 -1.77509829e-01
-1.94143981e-01 4.56303000e-01 -3.17708611e-01 -5.21584868e-01
-1.49608612e-01 1.20991302e+00 -2.77052432e-01 6.38658762e-01
-6.55493915e-01 -3.21415067e-01 7.36701667e-01 1.95017338e-01
4.05210048e-01 -8.42690825e-01 6.41064420e-02 -1.35883045e+00
3.20467114e-01 -7.48702288e-01 3.55804145e-01 -1.17330492e+00
-1.10984230e+00 3.06366563e-01 1.17246397e-01 -1.68041348e+00
-1.09873760e+00 -1.18775380e+00 -6.87807381e-01 7.35124946e-01
-1.05427313e+00 -1.71601987e+00 -5.01183867e-01 6.25538468e-01
2.81833529e-01 -2.51073867e-01 7.18219459e-01 2.01679572e-01
4.29154038e-01 5.97551525e-01 -4.12241518e-01 5.68025410e-01
8.58370304e-01 -1.05698740e+00 1.02727664e+00 8.03024650e-01
2.79693633e-01 3.88109744e-01 7.31239676e-01 -7.04952776e-01
-1.38998425e+00 -5.24643958e-01 9.73777175e-01 -9.48839009e-01
3.92135113e-01 -5.20978123e-02 -2.88585573e-01 7.27659762e-01
3.64551060e-02 2.83461183e-01 -5.30048460e-02 -1.01307046e+00
-3.88880074e-01 4.57967997e-01 -1.21471250e+00 1.04670048e+00
1.62873876e+00 -2.19007641e-01 -9.53474343e-01 3.43765706e-01
9.27509293e-02 -7.34656513e-01 -6.33623898e-01 4.01236892e-01
1.46253443e+00 -6.61475897e-01 1.06039131e+00 -1.31394374e+00
5.63765883e-01 -5.59834063e-01 6.04433492e-02 -1.13352489e+00
1.23323888e-01 -7.39370108e-01 -3.89000326e-01 5.42591691e-01
1.20730646e-01 -1.32363066e-01 1.21815634e+00 5.50875545e-01
8.95169154e-02 -1.89388841e-01 -1.10440874e+00 -1.02893019e+00
4.43165511e-01 -2.12622032e-01 2.75122046e-01 3.74242812e-01
-2.00556129e-01 -2.99740493e-01 -5.67204535e-01 -3.96903068e-01
7.16194391e-01 -6.83703646e-02 1.34125698e+00 -8.08002472e-01
-2.06697732e-01 -6.43541932e-01 -1.00054276e+00 -1.02138853e+00
1.43972069e-01 -7.50165045e-01 -3.90453450e-02 -1.93290949e+00
-2.78616130e-01 7.28043079e-01 2.54005283e-01 6.12801492e-01
2.56401062e-01 3.75787646e-01 6.64319158e-01 2.33193949e-01
-3.10084205e-02 3.86285871e-01 1.87945604e+00 -7.12990761e-02
-1.89307019e-01 1.47482064e-02 2.47713014e-01 8.47082198e-01
3.68579626e-01 2.98427008e-02 -9.82988328e-02 -4.82462168e-01
-1.16758168e-01 -8.06878880e-02 9.68342125e-01 -1.16939485e+00
1.10696279e-01 1.51233628e-01 5.76491714e-01 -7.20018148e-01
3.91710043e-01 -6.30395949e-01 1.04133546e-01 1.06601870e+00
-2.56264478e-01 2.15820987e-02 6.49896637e-02 1.66600659e-01
-3.50545585e-01 2.22150534e-01 7.71898627e-01 -4.02608424e-01
-1.01683652e+00 1.46788985e-01 -4.81384456e-01 6.60069063e-02
7.39302754e-01 -4.29093510e-01 -1.26947403e-01 -9.17971969e-01
-8.94398928e-01 -3.55153829e-02 3.70086789e-01 6.30263746e-01
8.18797469e-01 -1.67208111e+00 -1.01879358e+00 2.01551780e-01
4.58435863e-02 -2.81264752e-01 -9.07685459e-02 4.57362592e-01
-1.37909365e+00 5.83445907e-01 -7.84477949e-01 -5.77820420e-01
-1.52645171e+00 -2.05510437e-01 5.52319407e-01 9.88715366e-02
-1.02050424e+00 5.68953097e-01 -4.07985926e-01 -4.95512933e-01
3.41135114e-01 -8.53016794e-01 2.53593534e-01 -4.06195492e-01
3.67566466e-01 3.67672652e-01 -4.71502125e-01 -9.08995390e-01
-3.23491067e-01 1.29913914e+00 5.06169200e-01 -7.32603610e-01
1.02973819e+00 7.70679772e-01 4.74848241e-01 -1.56715810e-01
8.31755817e-01 -1.60076976e-01 -1.66004300e+00 3.12792331e-01
-1.56303793e-01 -5.43273747e-01 -6.64182901e-01 -9.96491253e-01
-7.92385995e-01 9.85846639e-01 6.68630242e-01 -9.39131558e-01
5.58323979e-01 -1.72408983e-01 1.03231823e+00 7.77244389e-01
5.12059450e-01 -1.00599563e+00 6.23313040e-02 7.99256980e-01
1.49180806e+00 -8.78712833e-01 -1.74080387e-01 -1.48839757e-01
-6.96350336e-01 1.38102841e+00 5.54184556e-01 -5.56554794e-01
5.82700551e-01 2.73667216e-01 7.63826907e-01 1.04886768e-02
-1.88261703e-01 -3.41760457e-01 7.97024548e-01 9.55346167e-01
7.75325239e-01 -7.21084923e-02 -6.96538687e-01 1.03972532e-01
-4.82903451e-01 5.14450371e-01 4.77599949e-01 1.25714946e+00
7.95567110e-02 -1.16331267e+00 -3.30131531e-01 -1.33789256e-01
1.85381025e-02 3.56791258e-01 -1.04000187e+00 1.29817653e+00
-6.84477985e-02 2.06952393e-01 -2.44803682e-01 -1.05385929e-01
8.29214990e-01 3.57611686e-01 1.42905152e+00 -2.94371754e-01
-3.62898797e-01 -1.40860304e-01 2.98177809e-01 -6.53544605e-01
-5.53671539e-01 -7.01844633e-01 -1.45923615e+00 2.76438206e-01
4.49069738e-01 -5.65687239e-01 7.18049943e-01 4.50466514e-01
1.95510566e-01 1.73598960e-01 -3.47875446e-01 -1.53331304e+00
-9.36234713e-01 -1.05872273e+00 -4.06996012e-01 8.82256925e-01
3.54605108e-01 -5.58151126e-01 -3.88002157e-01 3.89120698e-01] | [9.19231128692627, -6.515130996704102] |
674bd818-2d4d-44cd-99d2-968865556007 | the-brain-tumor-segmentation-brats-challenge-2 | 2305.17033 | null | https://arxiv.org/abs/2305.17033v2 | https://arxiv.org/pdf/2305.17033v2.pdf | The Brain Tumor Segmentation (BraTS) Challenge 2023: Focus on Pediatrics (CBTN-CONNECT-DIPGR-ASNR-MICCAI BraTS-PEDs) | Pediatric tumors of the central nervous system are the most common cause of cancer-related death in children. The five-year survival rate for high-grade gliomas in children is less than 20\%. Due to their rarity, the diagnosis of these entities is often delayed, their treatment is mainly based on historic treatment concepts, and clinical trials require multi-institutional collaborations. The MICCAI Brain Tumor Segmentation (BraTS) Challenge is a landmark community benchmark event with a successful history of 12 years of resource creation for the segmentation and analysis of adult glioma. Here we present the CBTN-CONNECT-DIPGR-ASNR-MICCAI BraTS-PEDs 2023 challenge, which represents the first BraTS challenge focused on pediatric brain tumors with data acquired across multiple international consortia dedicated to pediatric neuro-oncology and clinical trials. The BraTS-PEDs 2023 challenge focuses on benchmarking the development of volumentric segmentation algorithms for pediatric brain glioma through standardized quantitative performance evaluation metrics utilized across the BraTS 2023 cluster of challenges. Models gaining knowledge from the BraTS-PEDs multi-parametric structural MRI (mpMRI) training data will be evaluated on separate validation and unseen test mpMRI dataof high-grade pediatric glioma. The CBTN-CONNECT-DIPGR-ASNR-MICCAI BraTS-PEDs 2023 challenge brings together clinicians and AI/imaging scientists to lead to faster development of automated segmentation techniques that could benefit clinical trials, and ultimately the care of children with brain tumors. | ['Leonie Mikael', 'Maryam Fouladi', 'Michelle Deutsch', 'Peter de Blank', 'Miriam Bornhorst', 'Russel Taki Shinohara', 'Nakul Sheth', 'Lubdha M. Shah', 'Ibraheem Salman Shaikh', 'Andres Rodriguez', 'Zachary Reitman', 'Sanjay P Prabhu', 'Julija Pavaine', 'Khanak K Nandolia', 'Ahmed W Moawad', 'Aaron S McAllister', 'Nazanin Maleki', 'Hongwei Bran Li', 'Hollie Anne Lai', 'Dominic LaBella', 'Blaise V Jones', 'Keyvan Farahani', 'Ivan Ezhov', 'Austin J. Borja', 'Marius George Linguraru', 'Spyridon Bakas', 'Arastoo Vossough', 'Brian Rood', 'Adam Resnick', 'Roger Packer', 'Ali Nabavizadeh', 'Margot Lazow', 'Benjamin Kann', 'Lindsey Hoffman', 'Mariam Aboian', 'Anna Zapaishchykova', 'Walter Wiggins', 'Benedikt Wiestler', 'Jeffrey B Ware', 'Chunhao Wang', 'Karthik Viswanathan', 'Wenxin Tu', 'Jeffrey D Rudie', 'Tina Poussaint', 'Marie Piraud', 'Bjoern Menze', 'Zeke Meier', 'Koen van Leemput', 'Florian Kofler', 'Elaine Johansen', 'Anastasia Janas', 'Juan Eugenio Iglesias', 'Shuvanjan Haldar', 'Ariana Familiar', 'James Eddy', 'Farouk Dako', 'Gian-Marco Conte', 'Verena Chung', 'Evan Calabrese', 'Timothy Bergquist', 'Ujjwal Baid', 'Sina Bagheri', 'Hannah Anderson', 'Udunna Anazodo', 'Maruf Adewole', 'Jake Albrecht', 'Syed Muhammed Anwar', 'Zhifan Jiang', 'Debanjan Haldar', 'Xinyang Liu', 'Nastaran Khalili', 'Anahita Fathi Kazerooni'] | 2023-05-26 | null | null | null | null | ['tumor-segmentation', 'brain-tumor-segmentation'] | ['computer-vision', 'medical'] | [-5.79013769e-03 2.37428233e-01 -2.81694859e-01 -2.96508104e-01
-1.20785892e+00 -5.06456256e-01 3.71072739e-01 7.19828069e-01
-6.96700335e-01 4.02161717e-01 8.95069987e-02 -5.06043732e-01
-1.88260943e-01 -4.46287006e-01 -5.20298719e-01 -8.32568586e-01
-1.72392830e-01 1.25195551e+00 4.32787240e-01 2.51798719e-01
-5.60968593e-02 6.02587283e-01 -6.73077166e-01 2.55758762e-01
1.07900119e+00 7.24491954e-01 5.91338754e-01 5.50843179e-01
-5.97525090e-02 2.75569439e-01 -1.59814939e-01 1.27678588e-01
-1.94444567e-01 -1.27356544e-01 -1.06533492e+00 -3.98599237e-01
5.04945040e-01 8.85704532e-02 -1.44356981e-01 9.05204356e-01
6.64922655e-01 -1.31040469e-01 9.20056939e-01 -1.19161677e+00
9.72070470e-02 6.62187815e-01 -8.31234276e-01 7.87699640e-01
1.58654451e-02 2.60810196e-01 3.91841352e-01 -7.20991850e-01
5.35435975e-01 4.63352531e-01 1.00514662e+00 5.20591319e-01
-1.07125616e+00 -7.34410048e-01 2.20990106e-01 1.25773668e-01
-1.26037014e+00 2.16887295e-01 -2.53574252e-01 -9.25018132e-01
1.23414099e+00 2.65549064e-01 1.19519770e+00 7.79838562e-01
5.23249745e-01 7.18565345e-01 1.07384527e+00 1.50894016e-01
3.65026712e-01 -7.48238444e-01 4.90333557e-01 6.35520697e-01
3.26712191e-01 -1.32746160e-01 -3.07610482e-01 1.39912739e-01
4.20381010e-01 -1.40792727e-01 -6.25049531e-01 -2.34429985e-01
-1.20462298e+00 6.98406339e-01 4.86086041e-01 5.46847701e-01
-3.05991918e-01 2.70020247e-01 5.63232958e-01 -1.07455961e-01
7.09707260e-01 -5.34679033e-02 -2.85008430e-01 -1.07652307e-01
-1.21437669e+00 3.97159457e-02 1.68436080e-01 7.11709321e-01
-3.56728345e-01 -2.13330030e-01 9.06856880e-02 9.74580348e-01
2.13579565e-01 3.16998184e-01 1.18378961e+00 -2.10238770e-01
2.77364254e-01 2.75039136e-01 -6.59318745e-01 -1.30769297e-01
-1.27125788e+00 -6.73104405e-01 -7.16527522e-01 1.81376353e-01
3.14894795e-01 -8.06799084e-02 -1.40256429e+00 1.56081855e+00
2.57047176e-01 1.29911751e-01 -3.09538722e-01 8.09203327e-01
1.22799659e+00 3.02635655e-02 2.01123878e-01 -2.73873322e-02
1.12379622e+00 -1.04394257e+00 -1.21237757e-02 -1.61890853e-02
1.15216267e+00 -3.58407348e-01 5.19357085e-01 5.15820980e-01
-1.21299314e+00 4.40332979e-01 -7.11622894e-01 3.94828737e-01
-4.00109380e-01 -2.44622454e-01 8.72673988e-01 7.78645873e-01
-1.60585368e+00 1.37876526e-01 -1.44525135e+00 -6.07535183e-01
1.02071023e+00 7.48424768e-01 -8.03358257e-01 -2.24196509e-01
-6.03225946e-01 1.25262988e+00 6.02583349e-01 -3.62200797e-01
-1.37358534e+00 -1.62030399e+00 -4.37756360e-01 -4.06725585e-01
1.04464352e-01 -6.10476077e-01 1.10238039e+00 -3.39745134e-01
-9.75096881e-01 1.20257771e+00 2.19635308e-01 -5.72487295e-01
5.90909481e-01 4.57641304e-01 -2.10712403e-01 2.94578105e-01
4.55115438e-01 9.78889406e-01 1.70517072e-01 -5.07450700e-01
-8.58032823e-01 -8.43263805e-01 -8.83630097e-01 3.56809914e-01
2.82156318e-01 9.79144033e-03 -1.01151146e-01 -5.12630343e-01
4.78904516e-01 -9.83009994e-01 -3.17594260e-01 -4.73741889e-01
-1.64716691e-01 -7.72675201e-02 7.85581410e-01 -6.76022589e-01
3.33278716e-01 -1.67450976e+00 2.65654594e-01 1.73859254e-01
5.66825032e-01 1.49541825e-01 3.02070496e-03 -3.12146306e-01
-7.47112155e-01 1.86325535e-01 -2.90757209e-01 -1.53528020e-01
-5.98067164e-01 -2.74159729e-01 4.34112400e-01 7.37457633e-01
-2.32077867e-01 1.10826242e+00 -1.14871383e+00 -8.45345780e-02
2.38900512e-01 2.82737374e-01 -3.68483603e-01 -9.62279513e-02
7.11521059e-02 7.38889158e-01 -2.04050735e-01 5.11349499e-01
5.28598487e-01 5.96352434e-03 -3.45955312e-01 1.96727157e-01
-3.51515085e-01 -5.34260599e-03 -5.99393547e-02 2.17905664e+00
-6.15647100e-02 6.03827596e-01 2.69521981e-01 -1.02079237e+00
2.84381509e-01 5.54182589e-01 1.10087574e+00 -5.35586536e-01
2.86576062e-01 5.68979383e-01 4.31517661e-01 -1.57506913e-01
-4.66625988e-01 -3.75123471e-01 4.56432521e-01 3.51008296e-01
2.25800917e-01 -9.86859322e-01 2.69077957e-01 1.99593216e-01
1.63673329e+00 -3.63797426e-01 -2.88610846e-01 -7.34971583e-01
2.42145717e-01 2.91780055e-01 4.00088429e-01 5.11980236e-01
-4.97010767e-01 8.96800160e-01 2.96937674e-01 -2.32229054e-01
-7.72745371e-01 -1.40326619e+00 -8.64546061e-01 6.26931250e-01
-6.34984970e-01 1.69547752e-01 -8.84838879e-01 -7.49271035e-01
-2.39598423e-01 6.99336648e-01 -8.49741578e-01 1.01314448e-01
-2.00108662e-01 -1.44024909e+00 9.20572519e-01 3.80995721e-01
2.45697811e-01 -6.36200547e-01 -8.36503327e-01 4.94664341e-01
-2.28752773e-02 -1.03572619e+00 -2.64732510e-01 5.12362659e-01
-8.19266200e-01 -1.33637357e+00 -1.51349485e+00 -1.06293952e+00
7.54723966e-01 -2.14964584e-01 7.62880206e-01 -1.20024636e-01
-6.81500614e-01 5.66100597e-01 -1.85740784e-01 -7.19194591e-01
-2.99346864e-01 3.40432853e-01 -1.82535529e-01 -7.26222157e-01
3.02467495e-01 -7.98825860e-01 -5.31211615e-01 1.42076358e-01
-6.58110201e-01 6.15900457e-01 4.10215080e-01 7.60069370e-01
7.73665011e-01 -1.59925625e-01 5.33840835e-01 -6.10822856e-01
5.14777720e-01 -6.11135960e-01 -6.94074988e-01 2.01193482e-01
-4.93941486e-01 -5.61222911e-01 1.63736179e-01 -2.62645096e-01
-5.21103859e-01 -2.82744586e-01 -2.57811785e-01 -1.11704379e-01
-3.71012449e-01 8.08442533e-01 3.46137941e-01 -3.75638992e-01
4.27558392e-01 -2.70238947e-02 -6.62591076e-03 2.85153668e-02
-9.74846706e-02 7.42958337e-02 7.03733087e-01 -5.02355874e-01
1.58700690e-01 4.35706258e-01 4.57879484e-01 -6.81594789e-01
-5.21567285e-01 -4.49319273e-01 -5.14200211e-01 -2.82393783e-01
1.27910674e+00 -6.46017194e-01 -2.08271548e-01 1.07069445e+00
-7.69749939e-01 -8.27607751e-01 -2.19708666e-01 7.12126911e-01
-6.15393341e-01 -1.90896064e-01 -7.87328482e-01 5.20925708e-02
-9.84075904e-01 -1.99144804e+00 7.73739815e-01 2.40913004e-01
-2.19544485e-01 -1.21792006e+00 4.26060230e-01 4.51440781e-01
6.29589379e-01 5.41866302e-01 1.29522491e+00 -1.01537657e+00
1.68329719e-02 -2.21201077e-01 -2.89172500e-01 -1.08368143e-01
1.15684597e-02 -5.92522882e-02 -5.80159426e-01 -4.48602647e-01
-3.61845434e-01 -3.32813591e-01 7.43137658e-01 1.03428960e+00
1.25291669e+00 6.89787567e-01 -4.97670680e-01 8.25914145e-01
1.24577475e+00 6.89770520e-01 1.83414593e-01 3.42949271e-01
8.29764903e-01 5.12082100e-01 -6.59811124e-02 -1.41004577e-01
5.93928099e-01 4.12962198e-01 5.57235301e-01 3.34191695e-02
-4.78839606e-01 1.42148182e-01 -3.97701621e-01 8.28829587e-01
-2.79429127e-02 2.75009833e-02 -1.93531716e+00 8.84839475e-01
-1.26458859e+00 -4.51444626e-01 -5.27620792e-01 2.26613641e+00
5.94503999e-01 1.51571751e-01 2.64631249e-02 3.20509709e-02
6.15714729e-01 -1.82924926e-01 -5.68788171e-01 -2.80867159e-01
-3.95521335e-02 5.40605903e-01 6.61322534e-01 3.75064224e-01
-7.02869952e-01 7.76793540e-01 6.13088560e+00 9.66601193e-01
-1.55524909e+00 8.11187863e-01 1.09415543e+00 -5.39750695e-01
-1.04269437e-01 -3.09016556e-01 -4.19318706e-01 3.08793038e-01
1.05823898e+00 -3.67840022e-01 1.10836893e-01 2.41935685e-01
1.67411149e-01 -5.80223560e-01 -1.12668848e+00 8.44857693e-01
-1.19859733e-01 -1.36721456e+00 -5.07208526e-01 2.74959505e-01
9.37506497e-01 1.18356693e+00 3.75475109e-01 1.21228270e-01
4.07190889e-01 -1.54447246e+00 5.93546152e-01 3.02140951e-01
1.28347266e+00 -7.38033652e-01 7.33648539e-01 9.26886126e-02
-8.57648075e-01 4.60692286e-01 4.89614494e-02 7.70542443e-01
1.53917357e-01 6.47461832e-01 -1.18185520e+00 3.88078839e-01
7.41900921e-01 6.10139072e-01 -3.77849758e-01 1.56144130e+00
-4.06262316e-02 7.30049491e-01 -4.87984240e-01 2.45631531e-01
3.79880637e-01 -3.38968374e-02 4.41159546e-01 1.13681400e+00
3.65049720e-01 4.81424630e-01 -2.53789406e-02 6.58779681e-01
1.24093600e-01 3.10700148e-01 -6.08822554e-02 -9.09527019e-02
8.93573165e-02 1.33217061e+00 -1.32862103e+00 -1.99606586e-02
-3.05166900e-01 3.30902904e-01 3.86573017e-01 6.31450340e-02
-6.02667332e-01 3.26911986e-01 7.57895038e-02 4.76271898e-01
-1.23171225e-01 2.79322803e-01 -5.94106138e-01 -9.08096790e-01
-7.18646407e-01 -7.06132233e-01 4.42295283e-01 -9.76040721e-01
-7.54980147e-01 5.63010812e-01 2.98056483e-01 -2.91149348e-01
-5.59512712e-02 -5.53653181e-01 -1.16474938e+00 8.67983520e-01
-1.19650662e+00 -1.08703077e+00 -3.72314006e-01 5.15306830e-01
3.58733743e-01 -3.24628592e-01 8.18001509e-01 -7.05889314e-02
-7.60794342e-01 6.31446481e-01 -7.52689317e-02 -7.35761672e-02
9.80181694e-02 -1.31237435e+00 1.79507285e-01 5.73238909e-01
-5.56892455e-01 -1.01372421e-01 4.50074881e-01 -7.92316675e-01
-7.60804057e-01 -1.09250319e+00 -7.21317232e-02 -1.18649870e-01
1.02928853e+00 2.18181342e-01 -7.23835886e-01 7.73092747e-01
1.85668886e-01 -1.76159944e-02 8.45955312e-01 -3.84965092e-01
-3.19188684e-02 3.16216409e-01 -1.57254565e+00 2.21805602e-01
6.65890098e-01 -1.76541954e-01 -3.83478165e-01 7.02716768e-01
4.43872362e-01 -8.76254559e-01 -1.34869134e+00 8.64361346e-01
1.89646557e-01 -6.52723610e-01 5.55423856e-01 -2.29617283e-01
2.71879911e-01 1.10869244e-01 2.27005795e-01 -1.86247540e+00
-1.52202532e-01 -4.62026820e-02 5.75054884e-01 5.64443946e-01
5.36556423e-01 -8.77234101e-01 1.17263436e+00 9.70728993e-01
-8.12619328e-01 -1.29659212e+00 -1.37769127e+00 -5.19975841e-01
1.07722831e+00 -6.05381489e-01 4.66032803e-01 8.23024571e-01
-2.72985827e-02 -4.11133736e-01 9.66999352e-01 -1.82600692e-02
6.93409979e-01 -6.23340666e-01 8.09896663e-02 -1.00257921e+00
-1.01922154e-01 -1.22062600e+00 -1.01585841e+00 -5.94906956e-02
-6.70505464e-02 -1.28775263e+00 -1.88508511e-01 -2.05898356e+00
5.33947825e-01 -5.83017826e-01 -4.78371203e-01 5.49718857e-01
1.60629660e-01 -6.71288595e-02 -2.11400628e-01 -1.52509853e-01
-1.63710654e-01 1.57187939e-01 1.13390958e+00 -2.93778837e-01
1.75401047e-01 4.30444181e-02 -9.62689221e-02 8.24772716e-01
6.59656763e-01 -5.25472939e-01 -6.44497871e-01 -3.87290835e-01
-4.07584071e-01 1.66123107e-01 1.31718069e-01 -1.29789913e+00
2.40841031e-01 -2.14129195e-01 3.72203648e-01 -1.07308722e+00
2.00719804e-01 -5.49500406e-01 3.41020793e-01 7.72859454e-01
6.50933385e-02 3.21786344e-01 3.93221170e-01 -1.55279741e-01
-7.16841668e-02 -2.84399509e-01 9.88056004e-01 -2.47892663e-01
-2.62271464e-01 9.60734904e-01 -7.83399522e-01 5.02594054e-01
1.51193380e+00 6.19851751e-04 -6.99490726e-01 -1.83554664e-01
-1.01561677e+00 4.64919060e-01 4.54625964e-01 6.17396310e-02
5.22323728e-01 -8.34087849e-01 -8.94453287e-01 -2.17979327e-01
-6.68014437e-02 7.77690530e-01 6.27633631e-01 1.77094924e+00
-1.06058145e+00 8.15750241e-01 -2.47222230e-01 -8.74296248e-01
-1.11756945e+00 1.87427700e-01 9.78851736e-01 -4.96391058e-01
-6.97903574e-01 1.56002808e+00 5.10513604e-01 -3.16061795e-01
1.56578511e-01 -4.77425247e-01 2.91235792e-03 -8.39498863e-02
5.59621036e-01 1.87193096e-01 6.56869411e-01 -8.77736926e-01
-4.60491061e-01 2.13658094e-01 -5.70128918e-01 -4.12114680e-01
1.63633859e+00 5.02349854e-01 -2.52448142e-01 1.04426205e-01
1.05338705e+00 -4.54989582e-01 -7.35848784e-01 3.12283128e-01
9.85258818e-03 5.01011498e-02 5.73230565e-01 -1.17509365e+00
-1.60471368e+00 7.00298011e-01 9.85212386e-01 -3.59163493e-01
1.10832202e+00 1.23941682e-01 5.36046386e-01 -5.52790463e-01
7.06792831e-01 -6.81366682e-01 -3.15642118e-01 3.61750215e-01
9.88070965e-01 -1.00274575e+00 -2.57930487e-01 -4.47827727e-01
-2.97404289e-01 9.90939617e-01 7.63674021e-01 7.00935647e-02
7.39993513e-01 5.40851831e-01 7.20485598e-02 -5.42336226e-01
-6.70034170e-01 1.76098332e-01 3.15945655e-01 8.68475735e-01
6.28852487e-01 6.48988366e-01 -3.55673075e-01 7.58199751e-01
-6.59122050e-01 -2.19430640e-01 5.13743639e-01 8.79771948e-01
-3.60762656e-01 -1.00232780e+00 -1.70230284e-01 9.64860380e-01
-4.12637174e-01 -2.80602425e-01 -2.45706767e-01 6.70671642e-01
1.71451181e-01 5.95697343e-01 5.96896894e-02 2.69966871e-02
1.22940645e-01 -2.03179289e-02 7.70299017e-01 -7.19266534e-01
-8.01703393e-01 1.39144436e-01 -6.64376393e-02 -2.56050646e-01
-7.13194758e-02 -1.03350592e+00 -1.73000920e+00 -2.03008518e-01
-5.31935766e-02 2.72586018e-01 1.25500011e+00 1.03402793e+00
-1.12567656e-01 7.88981378e-01 2.30955277e-02 -5.65825164e-01
2.34205633e-01 -9.76061046e-01 -5.75015187e-01 -1.14552870e-01
-2.13039860e-01 -7.51931727e-01 -1.42235070e-01 -9.14601803e-01] | [14.568521499633789, -2.494520425796509] |
4792a885-a7d8-4bc9-8223-c8a2eff51ffb | top-two-algorithms-revisited | 2206.05979 | null | https://arxiv.org/abs/2206.05979v2 | https://arxiv.org/pdf/2206.05979v2.pdf | Top Two Algorithms Revisited | Top Two algorithms arose as an adaptation of Thompson sampling to best arm identification in multi-armed bandit models (Russo, 2016), for parametric families of arms. They select the next arm to sample from by randomizing among two candidate arms, a leader and a challenger. Despite their good empirical performance, theoretical guarantees for fixed-confidence best arm identification have only been obtained when the arms are Gaussian with known variances. In this paper, we provide a general analysis of Top Two methods, which identifies desirable properties of the leader, the challenger, and the (possibly non-parametric) distributions of the arms. As a result, we obtain theoretically supported Top Two algorithms for best arm identification with bounded distributions. Our proof method demonstrates in particular that the sampling step used to select the leader inherited from Thompson sampling can be replaced by other choices, like selecting the empirical best arm. | ['Emilie Kaufmann', 'Rianne de Heide', 'Dorian Baudry', 'Rémy Degenne', 'Marc Jourdan'] | 2022-06-13 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [ 1.32078022e-01 1.88703537e-01 -9.86640215e-01 -1.08001903e-01
-1.29811907e+00 -1.17134798e+00 5.14290035e-01 -1.68699846e-01
-8.48033577e-02 1.01767755e+00 -4.17359807e-02 -7.56738544e-01
-7.75252461e-01 -6.02584302e-01 -9.29336965e-01 -9.00573790e-01
-1.43452855e-02 1.12897575e+00 -3.75003159e-01 2.52733201e-01
1.53674707e-01 4.54929560e-01 -1.14787054e+00 2.61296988e-01
5.16995013e-01 1.01195025e+00 -3.90155822e-01 6.66848242e-01
1.47677958e-02 4.78466928e-01 -3.63189816e-01 -6.20289087e-01
6.11873567e-01 -7.81275868e-01 -7.53923118e-01 2.03641161e-01
-2.59647612e-03 -4.84687418e-01 1.92842424e-01 8.65893960e-01
3.06929022e-01 8.75534639e-02 7.55612850e-01 -1.19198370e+00
-3.74375731e-01 1.44636583e+00 -1.01063061e+00 3.43592912e-02
2.42570922e-01 -1.85880318e-01 1.09759140e+00 -2.55696476e-01
1.13366589e-01 1.28054094e+00 5.87756038e-01 3.65656734e-01
-1.45668864e+00 -7.31420100e-01 1.81716785e-01 -2.54164904e-01
-1.12297082e+00 -6.18018031e-01 6.52816474e-01 -4.32327628e-01
1.04881264e-01 7.58269131e-01 4.85395610e-01 1.18702221e+00
-5.26073098e-01 1.13662374e+00 1.20922959e+00 -8.60163152e-01
5.10452211e-01 4.90432143e-01 5.12500823e-01 5.08926548e-02
6.48231089e-01 4.28844929e-01 -2.67710954e-01 -1.07772219e+00
6.93743587e-01 -6.88522235e-02 -6.23562038e-02 -6.31186128e-01
-8.26821685e-01 1.22295272e+00 -1.10441357e-01 -5.84898032e-02
-7.75431931e-01 2.29438499e-01 1.43585458e-01 2.81369060e-01
3.30874324e-01 1.66728958e-01 -4.15440321e-01 1.54566720e-01
-8.42731237e-01 2.77659804e-01 6.96375549e-01 9.86412883e-01
4.33280081e-01 -1.31315783e-01 -6.87339842e-01 4.90555376e-01
2.13635236e-01 7.62947679e-01 1.51886389e-01 -8.81408453e-01
4.90794063e-01 -1.00250021e-01 1.05044913e+00 -3.53549451e-01
-3.22309911e-01 -7.61429012e-01 -4.44504172e-01 -1.74081296e-01
7.92850018e-01 -4.43738878e-01 -6.64635718e-01 1.84443402e+00
4.43787813e-01 -4.32571694e-02 -2.71977425e-01 6.36106014e-01
-1.41488761e-01 1.81509644e-01 -1.92504883e-01 -5.72353482e-01
1.28898096e+00 -8.18771362e-01 -3.80034536e-01 -3.28873277e-01
5.50119579e-01 -5.68760037e-01 3.38736326e-01 4.97580886e-01
-1.30960381e+00 -2.18043640e-01 -6.71588600e-01 7.95154810e-01
2.45956242e-01 3.17338139e-01 9.13083732e-01 1.33989310e+00
-6.41893387e-01 4.20120329e-01 -6.48386240e-01 3.25224772e-02
4.64327574e-01 4.38737065e-01 2.11810656e-02 1.68053254e-01
-8.18911493e-01 3.15036893e-01 1.61460832e-01 7.89715946e-02
-8.06095958e-01 -4.29194123e-01 -2.24774227e-01 -5.03533185e-02
5.15533268e-01 -8.66214514e-01 1.59468853e+00 -1.40321124e+00
-1.69332445e+00 5.26522458e-01 -4.54729229e-01 -6.95858538e-01
7.42550254e-01 -4.71515134e-02 1.58078849e-01 5.99073768e-02
2.52511561e-01 -3.25573832e-01 1.14518821e+00 -1.13035083e+00
-8.41666579e-01 -6.26358092e-01 2.45821252e-02 -2.34357938e-01
3.29895355e-02 7.71091357e-02 -7.82445539e-03 -4.57226783e-01
6.15681373e-02 -1.23116910e+00 -3.39681119e-01 -8.05060506e-01
-6.68990850e-01 -1.70378357e-01 -1.06677234e-01 -1.64659441e-01
1.09947133e+00 -1.97409368e+00 -1.44409969e-01 6.95451140e-01
-3.08158308e-01 -1.90328404e-01 -2.41096988e-01 5.40221572e-01
-3.14196348e-01 1.99359849e-01 2.43477538e-01 -3.40067565e-01
1.98813528e-01 -3.16086471e-01 -7.07289040e-01 9.07053649e-01
-3.33849669e-01 7.91460335e-01 -7.39473641e-01 -1.80781614e-02
-3.40758488e-02 -2.55248219e-01 -2.70890087e-01 -2.45127920e-03
-2.55245790e-02 3.83517832e-01 -8.91775608e-01 3.79658729e-01
8.37891936e-01 -3.72859359e-01 3.95853698e-01 1.99929312e-01
1.05659395e-01 1.46407485e-01 -1.45026958e+00 7.43295729e-01
-2.93190271e-01 -6.92323297e-02 2.90843934e-01 -1.16671610e+00
5.38577020e-01 5.25454819e-01 3.55183154e-01 -1.78410366e-01
4.66705829e-01 3.75953406e-01 -1.19562661e-02 -1.71830028e-01
-2.33340710e-02 -3.26114535e-01 -6.88084513e-02 1.01712382e+00
-4.91297960e-01 4.06160414e-01 5.75151443e-02 -8.86335075e-02
6.19246244e-01 -2.21563876e-01 3.76546890e-01 -2.98834652e-01
2.70638853e-01 -2.86545493e-02 1.93957761e-01 1.67943621e+00
7.40493461e-02 5.03748834e-01 6.52549505e-01 -2.56474942e-01
-9.16531265e-01 -8.50212336e-01 -1.79177910e-01 1.37074113e+00
-3.25259954e-01 4.01542872e-01 -5.73697209e-01 -6.48551583e-01
3.71388733e-01 8.32716942e-01 -1.09365737e+00 2.39898980e-01
-3.50281954e-01 -7.98139393e-01 3.19621176e-01 3.97352934e-01
-1.47246346e-01 -4.05838907e-01 -4.52353239e-01 3.97043854e-01
-3.89226004e-02 -7.35155523e-01 -5.57367027e-01 3.98445636e-01
-8.02385628e-01 -1.13089192e+00 -9.37655330e-01 -2.88157046e-01
4.71750498e-01 5.19942820e-01 8.53494346e-01 -3.69402856e-01
5.27790606e-01 2.57549047e-01 -3.97188067e-01 -6.07416451e-01
-4.69989657e-01 8.98437798e-02 2.30208244e-02 6.07056201e-01
-9.32706054e-03 -2.56912887e-01 -5.79372704e-01 4.58697557e-01
-4.99009520e-01 -3.46589476e-01 4.68897104e-01 9.06236708e-01
4.30568665e-01 -7.90050402e-02 8.79846454e-01 -8.97281289e-01
4.22804236e-01 -7.03023732e-01 -1.06945264e+00 5.56384504e-01
-3.16214859e-01 2.64025927e-01 5.92913210e-01 -8.04223418e-01
-7.51550198e-01 1.11608259e-01 3.79227787e-01 -2.68587232e-01
-6.12807758e-02 4.36445862e-01 4.20864075e-02 1.93994507e-01
4.50548500e-01 9.70053226e-02 1.33830294e-01 -6.46260083e-01
2.97292203e-01 7.84936309e-01 1.56650215e-01 -9.34761465e-01
5.49545228e-01 7.24478245e-01 6.34217113e-02 -2.70282835e-01
-1.05445492e+00 -2.66475916e-01 -5.98458908e-02 3.70473295e-01
1.46031350e-01 -6.98090374e-01 -1.03786671e+00 2.12929979e-01
-9.71845329e-01 -3.07792634e-01 -3.49845529e-01 7.06970572e-01
-8.74516368e-01 6.00048117e-02 -4.85638604e-02 -1.81796062e+00
-3.05774122e-01 -1.03363049e+00 8.52601469e-01 1.47269219e-01
-3.37752759e-01 -7.15590596e-01 -2.27148719e-02 3.27478468e-01
1.35899112e-01 1.84215590e-01 9.49735582e-01 -1.07606184e+00
-3.61168593e-01 -5.51886499e-01 1.32614359e-01 -2.49348775e-01
1.32584006e-01 -5.74134216e-02 -7.38767147e-01 -5.89010417e-01
-2.89589493e-03 -1.20522887e-01 6.47519767e-01 1.27133834e+00
1.05210185e+00 -5.99770904e-01 -6.93293035e-01 3.77268404e-01
9.37093258e-01 5.60027838e-01 1.52792677e-01 6.05234206e-01
-1.35670111e-01 4.57520187e-01 5.09962440e-01 7.65084147e-01
-1.00787193e-01 6.68078482e-01 3.21150243e-01 3.57058972e-01
5.73030889e-01 -3.27021360e-01 1.00790799e-01 -1.74464926e-01
1.12396859e-01 -3.56862813e-01 -5.58443546e-01 8.68848503e-01
-1.94622970e+00 -1.10249209e+00 -5.82781173e-02 2.96671176e+00
9.16290462e-01 9.62227136e-02 9.98401284e-01 -8.89456365e-03
9.82914090e-01 -2.09419236e-01 -8.43321204e-01 -2.56385773e-01
-1.86965257e-01 2.79259682e-01 1.06224656e+00 5.34552395e-01
-1.11303008e+00 2.56675929e-01 7.20085001e+00 1.07419288e+00
-6.26770854e-01 1.46936283e-01 9.60419059e-01 -5.24563670e-01
-3.08477402e-01 2.23745793e-01 -1.06515110e+00 5.07698476e-01
1.11320388e+00 -3.92530978e-01 5.55848837e-01 7.98600614e-01
3.11553866e-01 6.30232915e-02 -1.23808444e+00 7.44550228e-01
-4.31431383e-01 -1.31316817e+00 -1.53320834e-01 2.25390211e-01
1.03063154e+00 -3.11827123e-01 3.66739303e-01 -1.01441450e-01
9.74683106e-01 -9.97632802e-01 1.02803683e+00 2.70857900e-01
4.52132970e-01 -1.11165512e+00 5.83269894e-01 6.22924984e-01
-6.59410059e-01 -6.61778390e-01 -1.37819037e-01 3.40993330e-02
5.36692142e-03 6.77338362e-01 -4.00160700e-01 7.00012922e-01
2.93509871e-01 -7.63439089e-02 6.54493794e-02 1.14618075e+00
-2.24562157e-02 9.66747642e-01 -5.92721701e-01 -5.84376156e-02
3.48818004e-01 -2.78758705e-01 5.49738407e-01 6.00591362e-01
5.52651525e-01 9.06112045e-03 1.59921408e-01 4.62956131e-01
1.02623574e-01 -2.98897978e-02 -2.69007236e-01 -6.25047386e-02
9.38248515e-01 7.06273794e-01 -5.28167903e-01 -2.75735974e-01
-1.42748743e-01 2.58418769e-01 9.25033316e-02 4.47487921e-01
-8.86545599e-01 -7.83096701e-02 3.68310481e-01 2.43547201e-01
7.01775730e-01 3.24793726e-01 -5.65157533e-01 -5.69481134e-01
-7.11619109e-02 -1.32227767e+00 7.20004141e-01 -1.61135852e-01
-1.28792703e+00 2.33790562e-01 3.83896291e-01 -9.84648108e-01
-6.00558102e-01 -3.86863351e-01 -4.91717100e-01 1.06113172e+00
-8.79079998e-01 -1.03519130e+00 6.10517323e-01 2.92150915e-01
2.37931818e-01 -1.45839691e-01 4.61597085e-01 -3.24568659e-01
-6.69414461e-01 9.75305080e-01 8.19135189e-01 -3.12223856e-04
1.50503069e-01 -9.07567084e-01 1.55283242e-01 4.57016289e-01
9.04126465e-02 7.90012598e-01 8.81119013e-01 -3.71083021e-01
-1.55846035e+00 -7.11847126e-01 2.45500848e-01 -3.05261433e-01
9.13137376e-01 -2.36636605e-02 -3.87232006e-01 1.09093618e+00
9.90694854e-04 -3.84412229e-01 6.41757607e-01 4.89554286e-01
-3.05202395e-01 -7.93128908e-02 -1.05877852e+00 4.24564928e-01
6.90935791e-01 -1.50048509e-01 -4.03284021e-02 6.33550167e-01
2.13924929e-01 -2.32875273e-01 -5.94688356e-01 1.63955167e-01
9.39381361e-01 -8.85763943e-01 9.87948120e-01 -1.07351947e+00
-1.30194798e-01 3.72783579e-02 -7.40299374e-02 -1.16149592e+00
-4.91120279e-01 -1.37267387e+00 -1.66679025e-01 1.17901170e+00
6.13305390e-01 -1.16991270e+00 9.00824547e-01 6.15309060e-01
5.99445999e-01 -6.25088215e-01 -1.26926649e+00 -1.14930141e+00
4.91811782e-01 -3.69679004e-01 1.17313600e+00 4.40827519e-01
-1.24084026e-01 1.57969907e-01 -8.05643320e-01 2.78512776e-01
9.17566895e-01 7.70616353e-01 9.42852497e-01 -1.21422982e+00
-9.08002794e-01 -7.82221258e-01 3.30567598e-01 -1.31823564e+00
8.05203915e-02 -4.16028917e-01 -1.66525364e-01 -8.71787012e-01
5.86435020e-01 -7.01866508e-01 -4.41157818e-01 2.72184879e-01
-1.20616920e-01 -1.68251708e-01 -6.66619018e-02 7.92550892e-02
-2.51073867e-01 9.31876749e-02 8.80684614e-01 -2.06739575e-01
-4.93296474e-01 1.03364336e+00 -1.28282595e+00 5.45436144e-01
9.18197632e-01 -7.04057992e-01 -4.06905740e-01 -2.15146825e-01
4.48986620e-01 7.12208092e-01 2.30591699e-01 -2.63129771e-01
3.63224521e-02 -5.88079751e-01 1.18571788e-01 -7.82224178e-01
-4.31684628e-02 -7.07579851e-01 7.26583183e-01 5.44805706e-01
-6.59888446e-01 -2.43365213e-01 3.80272865e-02 6.96553290e-01
5.53845942e-01 -6.86931252e-01 8.23450506e-01 -2.11933516e-02
4.76147801e-01 1.71963021e-01 -2.26491913e-01 -6.83362633e-02
9.75342035e-01 -2.56109256e-02 -3.34130317e-01 -8.32931697e-01
-6.99276924e-01 1.77340195e-01 8.27788338e-02 -4.23942283e-02
-2.14045234e-02 -1.26439691e+00 -1.02243865e+00 1.62680149e-01
-1.93709344e-01 -3.59996200e-01 3.03322971e-02 1.15306747e+00
2.84281164e-01 6.83486998e-01 4.48188365e-01 -3.07269901e-01
-1.08688343e+00 9.10495460e-01 3.35892409e-01 -4.35688406e-01
4.22160700e-02 7.90133357e-01 1.09966211e-01 1.03994668e-01
6.32940531e-02 5.35326004e-02 2.18244314e-01 2.55204231e-01
8.31202447e-01 6.13706112e-01 -5.13017327e-02 -3.82895768e-01
-3.88887823e-01 4.09922898e-01 -2.39543766e-01 -4.33314294e-01
1.15853870e+00 -3.23441923e-01 -9.25271288e-02 3.16168666e-01
9.35540676e-01 4.00966078e-01 -9.65751231e-01 -3.36570472e-01
9.89153013e-02 -6.06190622e-01 -1.57999530e-01 -5.51223338e-01
-9.26592946e-01 2.40751937e-01 3.23415101e-01 8.97834718e-01
8.22269380e-01 2.44850710e-01 3.84785414e-01 1.31712854e-01
8.00856948e-01 -6.37899339e-01 -5.52261114e-01 1.00715779e-01
7.34209359e-01 -1.06024122e+00 9.82421488e-02 -2.74454709e-02
-3.65300983e-01 9.05960143e-01 -1.72384471e-01 -3.12071685e-02
5.69880128e-01 -8.62680078e-02 -2.13893414e-01 3.47475231e-01
-5.64370573e-01 -3.22594404e-01 2.66693234e-01 6.11995220e-01
-7.18783364e-02 4.60887045e-01 -3.13689381e-01 1.04703796e+00
-4.02376533e-01 -6.71573058e-02 4.11137670e-01 5.25911689e-01
-3.65406603e-01 -1.16631389e+00 -9.98755395e-01 7.37213969e-01
-7.91164458e-01 1.17581218e-01 -2.42264658e-01 6.60934627e-01
-5.41248679e-01 1.43790901e+00 -3.75870280e-02 1.21290915e-01
2.15516016e-01 2.83827130e-02 6.00582361e-01 -2.05248669e-01
-4.27723646e-01 2.94325382e-01 1.37248188e-01 -7.14078769e-02
-2.05600873e-01 -8.98252428e-01 -3.97318542e-01 -4.82127279e-01
-7.43018329e-01 7.40713418e-01 3.17243129e-01 1.01502001e+00
3.08474243e-01 -1.01842731e-01 1.20796645e+00 -7.42335379e-01
-1.47378743e+00 -8.88330698e-01 -9.19142127e-01 -3.58668379e-02
6.52397692e-01 -5.99798620e-01 -6.69173777e-01 -4.80311453e-01] | [4.539419174194336, 3.2879409790039062] |
914f6e01-be49-45b7-8454-b4e4060403af | fusing-saliency-maps-with-region-proposals | 1804.03905 | null | http://arxiv.org/abs/1804.03905v1 | http://arxiv.org/pdf/1804.03905v1.pdf | Fusing Saliency Maps with Region Proposals for Unsupervised Object Localization | In this paper we address the problem of unsupervised localization of objects
in single images. Compared to previous state-of-the-art method our method is
fully unsupervised in the sense that there is no prior instance level or
category level information about the image. Furthermore, we treat each image
individually and do not rely on any neighboring image similarity. We employ
deep-learning based generation of saliency maps and region proposals to tackle
this problem. First salient regions in the image are determined using an
encoder/decoder architecture. The resulting saliency map is matched with region
proposals from a class agnostic region proposal network to roughly localize the
candidate object regions. These regions are further refined based on the
overlap and similarity ratios. Our experimental evaluations on a benchmark
dataset show that the method gets close to current state-of-the-art methods in
terms of localization accuracy even though these make use of multiple frames.
Furthermore, we created a more challenging and realistic dataset with multiple
object categories and varying viewpoint and illumination conditions for
evaluating the method's performance in real world scenarios. | ['Patric Jensfelt', 'Hakan Karaoguz'] | 2018-04-11 | null | null | null | null | ['unsupervised-object-localization'] | ['computer-vision'] | [ 2.19586402e-01 1.01723202e-01 -2.29741022e-01 -4.89571780e-01
-7.70979166e-01 -4.40066546e-01 7.07788348e-01 4.04114217e-01
-5.48215568e-01 4.92359579e-01 1.45422682e-01 2.91041225e-01
1.58058479e-01 -4.52648759e-01 -7.79241979e-01 -4.80031431e-01
3.05187441e-02 3.45854998e-01 1.09518993e+00 -1.67890742e-01
6.60730660e-01 4.61181015e-01 -1.71470749e+00 4.07778203e-01
5.87994277e-01 1.05074871e+00 7.43570268e-01 3.16083401e-01
1.87280178e-01 7.19311118e-01 -5.30572236e-01 5.78342602e-02
3.00439090e-01 -2.75252193e-01 -8.56966317e-01 5.03134906e-01
6.43908620e-01 -1.90128997e-01 8.93301796e-03 1.17756152e+00
3.40804994e-01 3.40892375e-01 5.13336897e-01 -9.70831811e-01
-5.79668701e-01 3.60972285e-01 -6.48405433e-01 5.68940520e-01
2.73219496e-01 -8.39280188e-02 9.10793006e-01 -1.34041345e+00
7.92247951e-01 1.21576691e+00 3.88788700e-01 1.87663600e-01
-1.32143760e+00 -2.29844898e-01 4.44359988e-01 2.71835923e-01
-1.62971771e+00 -5.44905007e-01 7.97244132e-01 -2.78752089e-01
8.17501962e-01 -2.24054515e-01 3.96678060e-01 5.85008442e-01
8.72879028e-02 8.43987107e-01 1.06337416e+00 -6.37450159e-01
5.07062733e-01 4.52289850e-01 -1.01304106e-01 7.00721145e-01
3.02595608e-02 1.56243313e-02 -5.67470908e-01 9.39217880e-02
8.63091111e-01 9.71619636e-02 -6.06180839e-02 -9.91137922e-01
-1.72646940e+00 7.60690928e-01 1.05097830e+00 4.49584574e-01
-5.37309408e-01 4.72682826e-02 6.33391812e-02 -2.57493168e-01
4.17635441e-01 5.79979718e-01 -2.18337625e-01 4.73037541e-01
-1.36709332e+00 1.02241628e-01 3.96666646e-01 9.22493100e-01
1.02703404e+00 -1.13987535e-01 -3.47179055e-01 7.71012068e-01
3.87077212e-01 3.84849831e-02 4.52383161e-01 -1.00889730e+00
2.06583411e-01 6.66003585e-01 3.27080697e-01 -1.10465133e+00
-2.63912857e-01 -6.11548483e-01 -3.75402421e-01 3.97347331e-01
3.06442618e-01 2.88910031e-01 -1.21216607e+00 1.47458982e+00
3.39541435e-01 2.81153321e-01 -3.95323560e-02 1.29498494e+00
7.30408370e-01 4.74030137e-01 -4.40370739e-02 1.12259060e-01
1.19079304e+00 -1.28202796e+00 -3.76855582e-01 -5.50474107e-01
-5.14780171e-03 -8.69305015e-01 8.08943033e-01 1.11464597e-01
-1.21341324e+00 -8.79545510e-01 -8.99095893e-01 -5.26305800e-03
-5.92144787e-01 4.41057712e-01 4.14107472e-01 1.68405190e-01
-1.36778593e+00 2.90256560e-01 -6.37279332e-01 -6.72641158e-01
5.55617809e-01 3.63821447e-01 -3.31783324e-01 -2.43252199e-02
-8.55700493e-01 1.13426459e+00 7.00552642e-01 -3.84346046e-03
-1.25070655e+00 -4.18228328e-01 -9.52613950e-01 4.43968251e-02
2.65324503e-01 -4.21179414e-01 1.27319586e+00 -1.34431076e+00
-1.19147480e+00 9.33898807e-01 -4.15279031e-01 -6.00100219e-01
1.58642292e-01 -9.33334231e-02 -1.37740359e-01 4.54674929e-01
5.25984347e-01 1.33389390e+00 8.91616881e-01 -1.50889361e+00
-9.70600009e-01 7.75905624e-02 2.39487842e-01 3.34198445e-01
1.22079983e-01 2.98310250e-01 -6.48946762e-01 -6.93672717e-01
2.92164981e-01 -8.09738576e-01 -4.76519138e-01 1.11255243e-01
-3.37786466e-01 -6.04528040e-02 8.36682022e-01 -2.29466230e-01
7.34516621e-01 -2.09751678e+00 5.32390550e-02 1.41811609e-01
1.37800127e-01 1.95909485e-01 -1.29450798e-01 1.76236674e-01
-4.70415689e-02 -2.12376878e-01 -1.67728364e-01 -3.25582027e-01
-2.67278403e-01 -1.10739917e-01 -2.41178960e-01 6.59213424e-01
4.23591942e-01 8.79114747e-01 -1.16092837e+00 -7.11265743e-01
4.94795948e-01 3.45213979e-01 -5.63582957e-01 1.54570401e-01
-2.05642313e-01 5.16312659e-01 -3.59272391e-01 7.97619104e-01
5.00209510e-01 -4.40213740e-01 -2.51064986e-01 -2.72587299e-01
-3.10184866e-01 2.67412096e-01 -1.06057525e+00 1.94266081e+00
-2.41344184e-01 7.60307848e-01 -2.07169816e-01 -1.04853451e+00
8.48022223e-01 4.90107499e-02 3.36398304e-01 -5.92316389e-01
6.58911318e-02 1.56014889e-01 5.72917871e-02 -9.87168178e-02
6.80001855e-01 2.44458154e-01 -1.78705472e-02 2.99160957e-01
4.23463166e-01 9.36599076e-02 3.13835412e-01 2.84072638e-01
8.78980517e-01 1.46140888e-01 4.50521111e-01 -5.56659222e-01
5.16253412e-01 1.81130186e-01 4.10678238e-01 7.88169444e-01
-4.31146443e-01 1.06104827e+00 6.48390651e-02 -5.14937460e-01
-1.03230095e+00 -1.29728794e+00 -1.27825335e-01 1.17516267e+00
8.83563817e-01 -1.48073167e-01 -7.28467584e-01 -7.72504687e-01
-2.67716348e-01 5.76051831e-01 -7.27589011e-01 4.81105074e-02
-3.39539170e-01 -3.88395548e-01 -8.41775909e-02 5.11536837e-01
5.53671300e-01 -1.34091055e+00 -1.01176310e+00 2.64920175e-01
-1.42297298e-01 -1.27238560e+00 -4.51553136e-01 9.35552567e-02
-6.65449440e-01 -1.01442552e+00 -7.98218071e-01 -1.18994117e+00
1.08926582e+00 6.51647329e-01 1.25051045e+00 -2.01998472e-01
-3.23209435e-01 2.74150819e-01 -3.85149568e-01 -3.39182526e-01
-1.06353432e-01 1.83342099e-02 -2.09197626e-02 2.15338901e-01
2.64582306e-01 -2.62625158e-01 -9.95030880e-01 5.72188318e-01
-7.48511672e-01 1.63330585e-01 8.93802464e-01 6.78992152e-01
8.46166372e-01 1.08567469e-01 6.16207421e-01 -5.05929053e-01
2.02851728e-01 -4.12161291e-01 -6.75579429e-01 2.71977305e-01
-2.90924430e-01 1.10249110e-02 2.12644294e-01 -3.64320666e-01
-1.01633060e+00 5.39510310e-01 4.32369858e-01 -4.50802237e-01
-3.96135747e-01 8.80866721e-02 1.31804883e-01 -2.55851835e-01
5.93996644e-01 2.99772531e-01 -3.24424267e-01 -3.70400935e-01
4.92666155e-01 4.21926171e-01 7.27795541e-01 -6.35028332e-02
7.64607847e-01 6.50268793e-01 -3.04122984e-01 -4.02341932e-01
-9.19499815e-01 -8.00106823e-01 -1.09700382e+00 -3.01605165e-01
8.71301830e-01 -1.12665772e+00 -2.69447863e-01 3.52959111e-02
-1.23948681e+00 -1.36759147e-01 -2.36587286e-01 5.49426794e-01
-5.96431613e-01 5.91811948e-02 -2.09111527e-01 -6.04296863e-01
-1.44182388e-02 -1.30412245e+00 1.35475183e+00 4.65608418e-01
-5.90616539e-02 -8.19969058e-01 -2.00788766e-01 2.21208017e-02
4.53194797e-01 1.42190635e-01 3.26506764e-01 -5.18085957e-01
-8.86704743e-01 -4.98810597e-02 -5.57478309e-01 1.10362187e-01
2.96429783e-01 -3.00130755e-01 -1.02996850e+00 -3.01589191e-01
-2.20878527e-01 -1.35652483e-01 9.65979457e-01 4.86785740e-01
1.07664669e+00 -1.22892097e-01 -5.76701164e-01 2.39608631e-01
1.57253075e+00 2.19542789e-03 4.33750421e-01 4.17313427e-01
5.99700570e-01 6.18841588e-01 1.02445650e+00 2.70061076e-01
3.99574280e-01 9.37481284e-01 7.30985641e-01 -5.30027151e-01
-3.01786035e-01 -1.39608294e-01 3.29181314e-01 2.26724267e-01
1.80774495e-01 -7.12179840e-02 -8.02845180e-01 1.17580903e+00
-1.92021477e+00 -9.38558817e-01 2.88095444e-01 2.29040551e+00
6.47033751e-01 1.21495254e-01 2.48762324e-01 -2.54994690e-01
1.06697464e+00 7.04913288e-02 -4.84121948e-01 -1.11269958e-01
2.90156212e-02 -3.81957777e-02 5.56961834e-01 2.72464991e-01
-1.47989261e+00 1.28234351e+00 6.69223166e+00 6.26406491e-01
-1.14814126e+00 1.58755213e-01 7.45269775e-01 -1.00238219e-01
8.10764432e-02 4.74581644e-02 -7.97716498e-01 4.50036168e-01
5.22278905e-01 1.52962551e-01 1.36114553e-01 9.59427476e-01
1.61181688e-01 -5.97508907e-01 -1.20494485e+00 7.48872161e-01
3.12561959e-01 -1.46556914e+00 -2.23424152e-01 -1.70067996e-01
9.99551952e-01 4.82076913e-01 1.39642790e-01 -2.12084856e-02
3.11947554e-01 -8.59540641e-01 1.08923590e+00 5.25527358e-01
4.22769487e-01 -6.11218333e-01 7.78177083e-01 1.66527346e-01
-1.20704567e+00 -7.34670609e-02 -5.70527434e-01 1.28374398e-01
1.09878302e-01 4.34263319e-01 -1.12555587e+00 2.55698860e-01
9.09862161e-01 7.65922070e-01 -1.00398624e+00 1.44303286e+00
-2.13251933e-01 3.05863380e-01 -3.36233258e-01 1.27583265e-01
5.07991970e-01 2.31320858e-01 4.16374713e-01 1.23470962e+00
1.88706011e-01 -1.95187435e-01 5.15180469e-01 1.01527703e+00
1.05211303e-01 6.05059136e-03 -5.40406406e-01 5.64789891e-01
5.17011523e-01 1.27649915e+00 -1.35080469e+00 -4.95842308e-01
-3.93115968e-01 1.00570333e+00 3.54878008e-01 4.21813041e-01
-8.31169009e-01 -2.72973299e-01 3.07695597e-01 2.08401620e-01
8.66677165e-01 -4.27871868e-02 2.48969402e-02 -1.08475506e+00
-6.19509220e-02 -3.49781096e-01 1.62539884e-01 -1.04619300e+00
-1.03657830e+00 6.69531703e-01 1.03026882e-01 -1.50346589e+00
-2.40165964e-01 -3.69062006e-01 -6.03572488e-01 7.15535223e-01
-1.74585199e+00 -1.16268778e+00 -3.04547518e-01 4.50210512e-01
9.10269320e-01 -2.07796931e-01 5.73937237e-01 -6.16773106e-02
-2.39035159e-01 2.82541275e-01 3.97577658e-02 1.50729030e-01
7.43392527e-01 -1.25629306e+00 4.67632234e-01 1.16696107e+00
4.07580763e-01 6.40580297e-01 6.28661096e-01 -4.69896525e-01
-9.26547766e-01 -1.40069735e+00 7.68091857e-01 -5.78781664e-01
4.60282475e-01 -5.67477822e-01 -7.43402660e-01 3.56480926e-01
4.55913275e-01 5.81571758e-01 1.72820792e-01 -2.58484870e-01
-1.75122261e-01 -1.12703197e-01 -1.20058572e+00 4.37393188e-01
7.98273921e-01 -3.71069461e-01 -7.19665766e-01 3.56037915e-01
6.71015978e-01 -2.79499680e-01 -5.36886752e-01 4.50991541e-01
3.13083291e-01 -9.92001534e-01 1.09812760e+00 -1.61814779e-01
3.11340153e-01 -8.78124058e-01 -1.37940124e-01 -1.19450676e+00
-5.51731229e-01 -2.32286006e-01 8.01341683e-02 1.15356958e+00
4.14555967e-01 -1.96234599e-01 7.10044444e-01 1.48057967e-01
-1.83065042e-01 -7.14783072e-01 -7.37420976e-01 -5.73141456e-01
-5.39121985e-01 -3.88462134e-02 2.81006932e-01 6.11491859e-01
-1.73279852e-01 2.58640319e-01 -1.56862870e-01 4.13979411e-01
6.92066729e-01 5.20557404e-01 5.54507852e-01 -9.88777399e-01
-2.86733769e-02 -3.86375159e-01 -8.46490622e-01 -8.43301892e-01
1.41509473e-01 -7.21877396e-01 5.47933996e-01 -1.61845326e+00
3.19524378e-01 -3.41665775e-01 -8.25927079e-01 5.69249868e-01
-1.20391928e-01 7.77623296e-01 2.04534248e-01 3.66611600e-01
-1.33132601e+00 4.55964863e-01 8.56198251e-01 -1.90791294e-01
-9.07396749e-02 -1.32769614e-01 -6.45303845e-01 7.65843689e-01
7.68431962e-01 -4.86652941e-01 -3.81528527e-01 -3.18476558e-01
-3.86330187e-01 -3.92706633e-01 6.72656119e-01 -1.28732252e+00
3.86980534e-01 -1.47404596e-01 6.56356871e-01 -7.49355793e-01
3.06999534e-01 -8.08373094e-01 -3.37588012e-01 2.73423016e-01
-4.98007059e-01 -6.79230550e-04 1.72443539e-01 5.45250833e-01
-3.91023338e-01 -6.34659454e-02 1.03733575e+00 -2.02114746e-01
-1.46303785e+00 4.41435166e-02 -3.20104927e-01 -2.05301896e-01
1.29809189e+00 -3.75997066e-01 -8.96426588e-02 -3.35692286e-01
-5.26042521e-01 9.36340913e-02 7.29238272e-01 5.90327144e-01
8.40852141e-01 -1.21313000e+00 -6.03934705e-01 2.38776073e-01
5.03503561e-01 4.42460589e-02 -8.50724503e-02 7.70146251e-01
-3.72960001e-01 5.89132845e-01 -3.94623190e-01 -1.02408731e+00
-1.05834711e+00 9.12809849e-01 2.81051457e-01 1.04104862e-01
-4.82557237e-01 7.43013799e-01 8.71587873e-01 -1.23917028e-01
2.61625707e-01 -4.14605677e-01 -3.71496767e-01 -1.76009580e-01
5.37126958e-01 -9.23606828e-02 -3.65974903e-02 -1.03081286e+00
-6.88563466e-01 6.66057706e-01 -2.17710137e-01 -8.94937441e-02
1.08494210e+00 -3.71676296e-01 1.24117881e-02 3.43125194e-01
1.03201115e+00 -1.80402651e-01 -1.42817283e+00 -4.91446674e-01
9.81230065e-02 -6.42101228e-01 1.66014329e-01 -8.40489388e-01
-1.07604110e+00 5.39029181e-01 8.52146447e-01 -1.19132027e-01
1.16892469e+00 3.37013423e-01 2.55247861e-01 2.03740895e-01
5.26346505e-01 -1.02266431e+00 2.75569320e-01 3.49384934e-01
7.30407417e-01 -1.59581661e+00 1.18783578e-01 -3.40079457e-01
-7.21784949e-01 7.33372033e-01 7.69181669e-01 -5.12458742e-01
5.64932227e-01 -6.99888319e-02 1.05497450e-01 -1.41673401e-01
-6.29117966e-01 -6.20650947e-01 6.18807197e-01 6.75613761e-01
3.69784713e-01 -3.10330033e-01 -1.45427436e-02 5.79543673e-02
1.09499261e-01 -2.40608796e-01 3.60304683e-01 8.73829901e-01
-8.06796908e-01 -7.86749005e-01 -3.96909982e-01 1.33283630e-01
-5.54421544e-01 -8.43073279e-02 -3.82282674e-01 7.45094538e-01
2.46901110e-01 9.59485412e-01 2.43005395e-01 -8.32596570e-02
2.56021544e-02 -4.16040778e-01 3.49573851e-01 -1.03893709e+00
-3.80773723e-01 3.00529778e-01 -3.56534421e-01 -7.62571692e-01
-8.00029993e-01 -5.41404009e-01 -1.23764026e+00 5.14078021e-01
-4.44470942e-01 2.78482199e-01 7.11721241e-01 9.11084712e-01
5.23003340e-01 4.92056638e-01 6.70418680e-01 -1.50138426e+00
-2.75006462e-02 -8.01091969e-01 -2.65453041e-01 3.84999901e-01
4.42629218e-01 -9.14608598e-01 6.00404330e-02 3.15188110e-01] | [9.576698303222656, 0.22442106902599335] |
208cfa49-d3ac-4899-820d-f05a6c612700 | an-error-propagation-spiking-neural-network | 2104.05241 | null | https://arxiv.org/abs/2104.05241v1 | https://arxiv.org/pdf/2104.05241v1.pdf | An error-propagation spiking neural network compatible with neuromorphic processors | Spiking neural networks have shown great promise for the design of low-power sensory-processing and edge-computing hardware platforms. However, implementing on-chip learning algorithms on such architectures is still an open challenge, especially for multi-layer networks that rely on the back-propagation algorithm. In this paper, we present a spike-based learning method that approximates back-propagation using local weight update mechanisms and which is compatible with mixed-signal analog/digital neuromorphic circuits. We introduce a network architecture that enables synaptic weight update mechanisms to back-propagate error signals across layers and present a network that can be trained to distinguish between two spike-based patterns that have identical mean firing rates, but different spike-timings. This work represents a first step towards the design of ultra-low power mixed-signal neuromorphic processing systems with on-chip learning circuits that can be trained to recognize different spatio-temporal patterns of spiking activity (e.g. produced by event-based vision or auditory sensors). | ['Giacomo Indiveri', 'Germain Haessig', 'Matteo Cartiglia'] | 2021-04-12 | null | null | null | null | ['event-based-vision'] | ['computer-vision'] | [ 6.58835948e-01 -5.90135872e-01 4.62764233e-01 -2.97867358e-01
-4.73428331e-02 -2.50065327e-01 2.28913024e-01 3.17211479e-01
-8.21333826e-01 6.88659012e-01 -6.11517251e-01 -1.54558867e-01
1.68917865e-01 -9.97195899e-01 -8.71853948e-01 -6.92479730e-01
-8.44795853e-02 -7.99193699e-03 9.96419787e-01 -1.31778598e-01
3.49331141e-01 5.92978120e-01 -1.89856184e+00 7.90332019e-01
2.04220220e-01 1.42927635e+00 3.87163520e-01 8.93332541e-01
-1.50744632e-01 5.24234354e-01 -3.74222904e-01 1.86856203e-02
-1.31692942e-02 -6.65402174e-01 3.19400400e-01 -7.33349860e-01
3.00686926e-01 4.30140644e-01 -4.66875225e-01 1.01560366e+00
6.12133324e-01 -4.32062417e-01 6.16947532e-01 -1.13175285e+00
-3.56212944e-01 5.23907781e-01 1.01119637e-01 5.43386638e-01
-2.65652072e-02 3.77870739e-01 2.56510526e-01 -9.19430554e-01
4.54111367e-01 5.73896825e-01 1.13476336e+00 8.34918857e-01
-1.68014371e+00 -8.61031771e-01 -2.06713036e-01 3.46431404e-01
-1.22491407e+00 -4.50042993e-01 5.98829985e-01 -2.43331611e-01
1.61138368e+00 -6.68925643e-02 1.16120899e+00 1.00651312e+00
1.19421542e+00 4.32784200e-01 1.24077296e+00 -1.33761615e-01
1.06730568e+00 -3.01792622e-01 8.00634250e-02 5.73101580e-01
5.53787172e-01 6.96795285e-02 -1.24453688e+00 6.93665221e-02
9.36136603e-01 2.06783623e-01 -1.59739897e-01 1.44655243e-01
-8.66282463e-01 1.77292168e-01 7.68744767e-01 6.05317533e-01
-5.85795343e-01 9.57220793e-01 1.66676044e-01 3.57366711e-01
-2.40034983e-01 4.49186862e-01 -3.66308481e-01 -6.95365742e-02
-1.38509130e+00 9.58795249e-02 7.56378114e-01 3.32365990e-01
7.11848915e-01 5.89561939e-01 3.29063297e-03 4.04197991e-01
2.81955779e-01 5.17714798e-01 8.26436818e-01 -7.51519322e-01
-5.16695142e-01 5.97658932e-01 -3.27080339e-01 -6.04177535e-01
-6.80742681e-01 -2.35718474e-01 -9.88703132e-01 9.91627157e-01
3.14031094e-01 -6.50349408e-02 -1.00129604e+00 1.54540408e+00
-4.92482811e-01 5.82565069e-01 1.73594244e-02 7.05409884e-01
5.99701405e-01 6.36615515e-01 -1.21467285e-01 -2.44426772e-01
1.24457443e+00 -2.86975861e-01 -6.85958326e-01 -7.20420778e-01
3.69719565e-02 -2.33853847e-01 4.34296191e-01 2.41997987e-01
-1.45519531e+00 -4.10442531e-01 -1.47861707e+00 1.06636159e-01
-7.32611358e-01 -1.27616242e-01 6.21504307e-01 4.90781426e-01
-1.33110058e+00 9.71047640e-01 -1.27013707e+00 -2.40973368e-01
5.01801074e-01 5.27374148e-01 1.37341917e-01 4.89830285e-01
-8.21247935e-01 6.68045580e-01 2.44254917e-01 -8.23620409e-02
-6.14556134e-01 -7.29968905e-01 -3.17079663e-01 2.60658950e-01
-6.89389527e-01 -6.84424639e-01 1.16147268e+00 -1.22943592e+00
-1.73822236e+00 8.66955459e-01 -4.52456355e-01 -9.17383134e-01
-1.20643608e-01 6.73290670e-01 -4.84500408e-01 -1.31565556e-01
-3.79826486e-01 7.48741746e-01 7.93760478e-01 -6.06028080e-01
-4.90205079e-01 -4.45571721e-01 -7.96625614e-01 -5.60661256e-01
-3.34339261e-01 -1.30495340e-01 1.57505646e-01 -5.95541000e-01
5.49035788e-01 -6.99128270e-01 -1.38897792e-01 8.72990727e-01
4.10272628e-01 1.15696639e-01 4.99741971e-01 1.88163489e-01
9.40430582e-01 -2.15204930e+00 -2.04209015e-01 2.17616677e-01
-8.52839127e-02 2.99228191e-01 -6.86535910e-02 2.68407941e-01
1.32166475e-01 -5.17835200e-01 -5.89866519e-01 -8.95678401e-02
-3.57560694e-01 1.58464447e-01 -2.60840356e-01 1.73713669e-01
5.33048451e-01 1.11434031e+00 -6.93709850e-01 1.96357191e-01
-1.13605991e-01 5.98966241e-01 -3.37191164e-01 -6.64742664e-02
-1.74489945e-01 1.85622349e-01 3.22041214e-01 5.85455120e-01
5.02836645e-01 -1.04186118e-01 1.12741232e-01 -3.90854597e-01
-7.41732895e-01 3.99867892e-01 -1.16738439e+00 1.82612813e+00
-3.80859613e-01 1.16238236e+00 1.47715151e-01 -7.45445907e-01
1.28029501e+00 2.14400962e-01 1.52977645e-01 -1.06150901e+00
3.70834023e-01 8.50157022e-01 1.40710548e-01 6.23134114e-02
5.94787821e-02 -1.46705657e-01 2.60273159e-01 3.89007270e-01
3.35121572e-01 -1.80496067e-01 2.00501055e-01 -4.33232844e-01
1.56181979e+00 -2.31260389e-01 -3.28876197e-01 -4.05569285e-01
1.62641615e-01 -2.23517284e-01 5.40429294e-01 8.25281084e-01
-1.69963792e-01 5.61085582e-01 -8.30824748e-02 -5.06023169e-01
-8.57983172e-01 -1.42523766e+00 -4.26263154e-01 8.64871562e-01
1.15675226e-01 3.49239670e-02 -4.61048514e-01 6.95090592e-01
-6.80850670e-02 2.50291407e-01 -2.78576702e-01 -4.43719983e-01
-3.74200255e-01 -7.64140725e-01 1.02387977e+00 7.94402778e-01
4.09807920e-01 -1.27891755e+00 -1.31433141e+00 1.00162601e+00
7.01898694e-01 -8.09819162e-01 1.20304175e-01 1.37419510e+00
-1.07891035e+00 -5.90032935e-01 -7.03169703e-01 -1.32712018e+00
5.72885871e-01 -2.98938304e-01 6.54542208e-01 -3.75534505e-01
-7.93633461e-01 1.61354437e-01 2.88121402e-01 -6.75768018e-01
-6.24594884e-03 -2.15679243e-01 2.09282115e-01 1.83421448e-02
6.79747939e-01 -1.29545462e+00 -7.78338134e-01 -1.51022643e-01
-9.07453775e-01 9.12179518e-03 5.75192928e-01 6.81716084e-01
1.07323468e+00 -3.82007569e-01 7.04739034e-01 -4.09748197e-01
5.82967341e-01 -2.74291128e-01 -7.23574221e-01 1.76592886e-01
-6.68401301e-01 3.73403907e-01 7.54709125e-01 -7.56016970e-01
-5.91369987e-01 4.60502177e-01 -3.08272302e-01 -4.91209142e-02
-4.65659285e-03 3.51125687e-01 6.71295166e-01 -6.50031388e-01
1.10184765e+00 5.79074562e-01 -1.21308379e-01 2.41973437e-02
-3.19926172e-01 4.40369099e-01 9.03250754e-01 -1.32607952e-01
2.35582873e-01 6.83925509e-01 2.68977970e-01 -6.12647772e-01
1.99071512e-01 -1.64535880e-01 -1.47008553e-01 -3.61506522e-01
6.28577769e-01 -8.48524690e-01 -7.76754022e-01 1.08996034e+00
-1.29967451e+00 -6.39678717e-01 -4.08121884e-01 4.24874306e-01
-6.98792338e-01 -4.76954252e-01 -1.19902611e+00 -8.10466051e-01
-5.81378639e-01 -5.60817599e-01 4.48441267e-01 7.41910994e-01
-1.78839296e-01 -7.07662404e-01 4.77914453e-01 -8.99602652e-01
1.05971909e+00 2.85072904e-02 5.37622452e-01 -9.57288593e-02
-4.98792738e-01 -1.33638054e-01 -5.79372644e-02 -5.41423075e-02
-1.08559698e-01 1.52823582e-01 -1.33940637e+00 -1.15086511e-01
-4.93036211e-03 -4.04749870e-01 1.27782357e+00 7.32961833e-01
1.01262093e+00 1.63581893e-01 -4.82154429e-01 7.27502108e-01
1.73990488e+00 2.43907049e-01 8.09588373e-01 -7.09702298e-02
-9.63288993e-02 3.23606543e-02 -6.24707818e-01 4.57532704e-01
5.14034815e-02 2.57114708e-01 1.03008099e-01 1.70842409e-01
-3.17752242e-01 -2.75761664e-01 5.16913950e-01 1.02769125e+00
9.66974124e-02 7.14823678e-02 -7.74260998e-01 5.75289726e-01
-1.75533390e+00 -1.17621946e+00 -4.23170596e-01 2.25186634e+00
1.01049602e+00 4.82754290e-01 -2.79399425e-01 4.22506005e-01
6.74487770e-01 -4.15013790e-01 -1.05659556e+00 -9.26177979e-01
-5.19201934e-01 1.14640570e+00 6.08668149e-01 -7.37938806e-02
-3.92448008e-01 5.41669607e-01 6.67955017e+00 3.16186100e-01
-1.75648522e+00 -4.49273773e-02 1.67821571e-01 -2.60758579e-01
-2.52736509e-01 -2.94256777e-01 -6.25291586e-01 6.47186756e-01
1.50437939e+00 -7.94495940e-02 6.73764825e-01 4.66217279e-01
-7.22281113e-02 -1.78962618e-01 -1.13447154e+00 1.19346201e+00
-3.42840075e-01 -1.91459334e+00 -2.65462577e-01 -5.58850646e-01
8.64602149e-01 4.65720326e-01 7.25387931e-02 4.36351039e-02
-1.32222474e-01 -8.53631020e-01 7.34285414e-01 8.22876811e-01
7.70999491e-01 -3.95893723e-01 4.69199628e-01 1.99534252e-01
-1.51931560e+00 -1.37461096e-01 -5.80107450e-01 -6.96975768e-01
4.29447629e-02 8.99003088e-01 -1.90031797e-01 -6.33410335e-01
1.00132275e+00 5.80174029e-01 -2.58092433e-01 1.61026597e+00
1.21589273e-01 4.72317904e-01 -6.62214339e-01 -6.53320312e-01
-9.16943401e-02 2.05431983e-01 9.67703611e-02 1.45416892e+00
7.34336138e-01 3.31572033e-02 -4.81989235e-01 1.36137140e+00
-3.87798637e-01 -3.89354944e-01 -5.55301607e-01 -2.14987114e-01
8.06619823e-01 1.22388101e+00 -1.06705332e+00 -1.42870963e-01
-3.54415864e-01 1.19596875e+00 8.34857300e-02 3.63964945e-01
-4.11234945e-01 -8.47349465e-01 7.50223815e-01 3.29188168e-01
3.52133036e-01 -4.69448715e-01 -7.64618814e-01 -8.07793677e-01
-5.16651012e-02 -7.00647235e-02 -2.14195132e-01 -8.96511078e-01
-1.10737729e+00 5.45566618e-01 -1.02301788e+00 -9.07884598e-01
-1.59582824e-01 -1.02169025e+00 -1.17096734e+00 6.86130941e-01
-1.52262557e+00 -4.12172318e-01 -2.51835972e-01 6.04713142e-01
1.95530191e-01 2.76193526e-02 1.08930445e+00 2.90387899e-01
-1.21511735e-01 4.63717252e-01 3.15890521e-01 -4.79935668e-02
4.38176006e-01 -8.39354455e-01 5.57361543e-01 7.08114147e-01
3.04909408e-01 2.31146872e-01 4.45805639e-01 -2.46479854e-01
-1.60517418e+00 -1.13689947e+00 8.83779049e-01 3.89140755e-01
7.44711041e-01 -5.02022803e-01 -1.27772057e+00 2.40244970e-01
8.10524672e-02 3.92726392e-01 6.61938965e-01 -5.34226656e-01
-5.28018117e-01 -4.77833748e-01 -1.36113870e+00 5.85722744e-01
1.04296827e+00 -7.75569379e-01 -4.16506171e-01 -2.02455133e-01
-1.93440856e-03 -1.90326702e-02 -3.87753457e-01 2.81358719e-01
8.11995447e-01 -9.05951977e-01 5.21859407e-01 -7.67274052e-02
7.18559474e-02 -5.81849217e-01 6.51992261e-02 -1.18115342e+00
-6.25688791e-01 -3.82721364e-01 -2.25194454e-01 6.42261088e-01
6.09741271e-01 -9.70005631e-01 8.56652558e-01 3.18652987e-01
-2.46714234e-01 -5.96250236e-01 -1.43937695e+00 -8.91190588e-01
-8.37476999e-02 -4.94900912e-01 -9.41424593e-02 2.18073487e-01
2.95252591e-01 -9.17048976e-02 4.24515784e-01 6.33449256e-02
4.97181445e-01 -3.76361609e-02 -3.24812770e-01 -1.24866700e+00
-3.71677071e-01 -9.55431879e-01 -1.14942002e+00 -8.37596536e-01
-1.69365913e-01 -9.75264609e-01 6.89390063e-01 -1.28086865e+00
-3.57113481e-01 -1.09542236e-01 -7.98355341e-01 5.24523616e-01
4.07045215e-01 8.67786527e-01 -1.80157870e-01 -1.14459982e-02
-2.16436923e-01 2.52118111e-01 2.80597925e-01 -1.77373216e-01
-1.41130522e-01 -1.15716211e-01 1.41421854e-01 6.31551504e-01
8.99136305e-01 -8.06413770e-01 -9.50049423e-03 -4.71126497e-01
6.37736619e-01 -2.75304377e-01 4.96032536e-01 -2.07012033e+00
1.34880757e+00 1.84005097e-01 7.14741468e-01 -2.53846571e-02
3.81558239e-01 -6.72686815e-01 2.89454341e-01 1.13550031e+00
-4.00913060e-01 2.37827286e-01 6.30480647e-01 4.83214229e-01
-1.27299458e-01 -3.40294391e-01 9.11581337e-01 -7.28835762e-02
-8.74585450e-01 1.20030038e-01 -1.31367123e+00 -1.70075580e-01
8.15937400e-01 -5.78478277e-01 -3.36694211e-01 1.89454257e-01
-6.20664656e-01 -3.16720963e-01 3.38318855e-01 -4.06505391e-02
1.09719396e+00 -1.38650882e+00 -4.57387507e-01 7.07376301e-01
1.41615435e-01 -6.40403032e-01 -3.03617362e-02 5.37566662e-01
-5.53750932e-01 1.05761506e-01 -1.05464411e+00 -7.35871434e-01
-8.03217947e-01 7.52523616e-02 7.09213078e-01 3.43668193e-01
-9.91199911e-02 1.13280511e+00 -7.36204624e-01 3.80468518e-02
2.25404352e-01 -5.66463113e-01 3.56658250e-01 -1.12908773e-01
8.18125248e-01 7.35611990e-02 3.41780037e-01 2.19207808e-01
-6.07893407e-01 6.59722328e-01 3.79146338e-01 -3.78094435e-01
1.41728127e+00 5.15189111e-01 -9.43566486e-02 1.25063908e+00
8.20996821e-01 -6.74917638e-01 -1.38882613e+00 -2.24483654e-01
7.65788257e-02 2.31587842e-01 2.23259553e-01 -8.61110032e-01
-1.07957709e+00 1.14449239e+00 1.31991291e+00 -2.15576533e-02
1.32942832e+00 -3.50970030e-01 8.94747198e-01 7.89246082e-01
7.68609643e-01 -1.32390356e+00 1.88194022e-01 6.34346485e-01
5.64769685e-01 -7.81082690e-01 -5.96209705e-01 3.54445547e-01
2.64873713e-01 1.69179130e+00 5.37983179e-01 -7.28569090e-01
9.13643241e-01 1.17290759e+00 -1.80416375e-01 -2.07259376e-02
-1.18336749e+00 -3.88356298e-01 1.36943599e-02 7.68073261e-01
6.08357668e-01 7.52943754e-02 -3.83306026e-01 6.09747052e-01
4.04577076e-01 7.38537073e-01 5.17315149e-01 1.15284932e+00
-9.21751618e-01 -8.21703196e-01 -1.62503533e-02 6.46428883e-01
-9.32736844e-02 -3.32562089e-01 -2.14066193e-01 -3.98294717e-01
2.12384224e-01 4.87555593e-01 6.93211496e-01 -5.29435575e-01
2.34450832e-01 4.19921070e-01 8.07547092e-01 -4.79532510e-01
-8.42056811e-01 -4.05232698e-01 -5.73441803e-01 -5.33456206e-01
-1.97946697e-01 -2.09409148e-01 -1.99392629e+00 -2.51508683e-01
1.70264784e-02 -4.20400262e-01 1.30235481e+00 6.16478324e-01
7.22852588e-01 9.05756235e-01 1.90866068e-01 -1.08377767e+00
-4.28689606e-02 -4.43901747e-01 -5.85903704e-01 -6.46736994e-02
1.16092414e-01 -1.22790240e-01 -3.27310711e-01 1.66637018e-01] | [8.21544075012207, 2.475943088531494] |
88e01397-796c-4d89-ae8d-29e5c1b8817e | gsb-group-superposition-binarization-for | 2305.07931 | null | https://arxiv.org/abs/2305.07931v3 | https://arxiv.org/pdf/2305.07931v3.pdf | GSB: Group Superposition Binarization for Vision Transformer with Limited Training Samples | Affected by the massive amount of parameters, ViT usually suffers from serious overfitting problems with a relatively limited number of training samples. In addition, ViT generally demands heavy computing resources, which limit its deployment on resource-constrained devices. As a type of model-compression method,model binarization is potentially a good choice to solve the above problems. Compared with the full-precision one, the model with the binarization method replaces complex tensor multiplication with simple bit-wise binary operations and represents full-precision model parameters and activations with only 1-bit ones, which potentially solves the problem of model size and computational complexity, respectively. In this paper, we find that the decline of the accuracy of the binary ViT model is mainly due to the information loss of the Attention module and the Value vector. Therefore, we propose a novel model binarization technique, called Group Superposition Binarization (GSB), to deal with these issues. Furthermore, in order to further improve the performance of the binarization model, we have investigated the gradient calculation procedure in the binarization process and derived more proper gradient calculation equations for GSB to reduce the influence of gradient mismatch. Then, the knowledge distillation technique is introduced to alleviate the performance degradation caused by model binarization. Experiments on three datasets with limited numbers of training samples demonstrate that the proposed GSB model achieves state-of-the-art performance among the binary quantization schemes and exceeds its full-precision counterpart on some indicators. | ['Hui Kong', 'Le Zhang', 'Cheng-Zhong Xu', 'Tian Gao'] | 2023-05-13 | null | null | null | null | ['model-compression'] | ['methodology'] | [ 2.50898987e-01 -4.52237815e-01 -3.21847707e-01 -2.44938031e-01
-4.52145308e-01 9.32954773e-02 4.02395397e-01 2.32146963e-01
-6.17766261e-01 6.56864643e-01 -2.81911731e-01 -3.04198772e-01
-1.44325286e-01 -8.16276312e-01 -4.99453515e-01 -9.65133905e-01
5.51115572e-01 1.89706579e-01 2.68629760e-01 -7.64031485e-02
4.71386075e-01 2.86093920e-01 -1.54836869e+00 -2.86294799e-02
1.19134748e+00 1.47078967e+00 5.47825277e-01 6.10375777e-02
-2.81257600e-01 4.82854694e-01 -7.64730871e-01 -5.83393931e-01
2.28053808e-01 -3.18553984e-01 -2.09591806e-01 -1.99899793e-01
1.21323623e-01 -5.37351370e-01 -5.79924524e-01 1.34123170e+00
6.99090898e-01 1.21731251e-01 5.04083455e-01 -1.11657858e+00
-4.72394019e-01 6.08415246e-01 -5.24724126e-01 3.89905393e-01
-3.15279841e-01 1.10446766e-01 6.21699214e-01 -8.04781914e-01
-9.03930515e-02 1.24507904e+00 6.46976829e-01 2.09379196e-01
-9.97860134e-01 -9.51482713e-01 -7.87955150e-02 6.70942366e-01
-2.00288057e+00 -3.68978679e-01 6.64407849e-01 -7.70899653e-02
1.10410988e+00 1.91338554e-01 6.49401188e-01 5.13715267e-01
1.17214039e-01 5.59098482e-01 8.60746145e-01 -4.62718129e-01
2.86034852e-01 2.78061647e-02 1.98625207e-01 5.82185745e-01
6.40923619e-01 -1.18977606e-01 -4.75063503e-01 6.21588305e-02
8.86287749e-01 1.05712287e-01 -4.66119468e-01 1.06336270e-02
-8.41485739e-01 6.74439073e-01 5.27609229e-01 3.61885667e-01
-3.55293661e-01 3.27185780e-01 3.78613830e-01 -1.93799809e-01
9.70295966e-02 1.89264014e-01 -2.37908199e-01 -2.88344473e-01
-1.20443940e+00 7.01776221e-02 3.50065857e-01 9.43401754e-01
7.19867408e-01 4.62444782e-01 -3.26613307e-01 9.71148312e-01
3.09822321e-01 5.91937423e-01 1.16466308e+00 -4.12183642e-01
6.80177748e-01 5.24028659e-01 -1.38925046e-01 -1.21577883e+00
-2.06381813e-01 -5.88007092e-01 -1.28174877e+00 -3.55222881e-01
1.63644627e-01 2.44841039e-01 -1.11581588e+00 1.37660837e+00
1.43942297e-01 1.81559801e-01 -5.17318100e-02 8.13350141e-01
8.03219795e-01 1.01849604e+00 1.71519011e-01 -3.83102477e-01
1.26786876e+00 -9.44800496e-01 -1.06187189e+00 -1.53149292e-01
5.46213984e-01 -6.62165165e-01 1.06986332e+00 5.91139853e-01
-1.03211224e+00 -7.95823634e-01 -1.40538859e+00 -1.75941780e-01
-3.46918792e-01 3.10390025e-01 7.35221803e-01 8.84434283e-01
-6.96862459e-01 8.55221689e-01 -9.75216508e-01 9.78277549e-02
4.83760893e-01 4.80182886e-01 1.08292662e-01 -1.97595358e-01
-1.37348330e+00 9.17699575e-01 7.31035888e-01 3.75539422e-01
-4.31358516e-01 -4.75633293e-01 -7.61561334e-01 5.74793637e-01
2.16972575e-01 -1.90691873e-01 1.14032757e+00 -5.00061870e-01
-1.39830172e+00 1.71080213e-02 -1.95877254e-01 -5.51657319e-01
1.27540812e-01 5.17042615e-02 -4.63808954e-01 -4.08956455e-03
-3.95503163e-01 4.43587989e-01 9.10945058e-01 -7.00646758e-01
-4.63984400e-01 -4.13251251e-01 -2.64643073e-01 2.39068091e-01
-8.14204395e-01 -3.36145341e-01 -7.51895547e-01 -6.91223979e-01
4.55199897e-01 -4.92835462e-01 -1.23680025e-01 -2.59651065e-01
-1.57335326e-01 -2.24180609e-01 6.31576359e-01 -7.38426626e-01
1.96912742e+00 -2.26923037e+00 -2.24380344e-01 2.61733592e-01
-1.80846453e-01 8.75248909e-01 2.18314365e-01 1.28874362e-01
1.43752143e-01 2.36325517e-01 -2.49599829e-01 -4.48587120e-01
3.27776838e-03 4.74397272e-01 -1.68099090e-01 2.31440529e-01
1.26523376e-01 5.74287295e-01 -6.20836139e-01 -8.19303513e-01
1.77065492e-01 6.51322961e-01 -5.48547506e-01 -1.02665812e-01
8.28233287e-02 -1.94788486e-01 -4.09083843e-01 6.18950248e-01
1.04827642e+00 -2.14962825e-01 3.93856391e-02 -6.42004490e-01
9.76353232e-03 5.02714932e-01 -1.30309486e+00 1.44427598e+00
-3.81179482e-01 2.60127574e-01 -1.23924188e-01 -1.15061975e+00
1.15789306e+00 1.81779549e-01 1.59456000e-01 -9.72756922e-01
4.76303548e-01 3.04867297e-01 2.22651269e-02 -3.18641841e-01
6.97073877e-01 -1.89192653e-01 1.64211228e-01 -1.18314423e-01
-6.03966564e-02 -2.31068358e-01 2.30116636e-01 4.30853404e-02
5.36622584e-01 -4.68707457e-02 3.23441267e-01 3.37648229e-03
4.74562526e-01 -2.59889513e-01 8.58611345e-01 4.77454096e-01
-2.35032424e-01 4.48703408e-01 1.99026108e-01 -1.10372625e-01
-8.74679625e-01 -5.98935366e-01 -4.70449239e-01 4.82689440e-01
4.73618358e-01 -7.66185641e-01 -7.94819295e-01 -1.24005519e-01
-8.28432366e-02 7.09215522e-01 -8.02972019e-02 -6.75678313e-01
-4.66094881e-01 -1.21679950e+00 7.15448618e-01 6.01924062e-01
1.14389443e+00 -6.52735770e-01 -5.90414107e-01 2.29307607e-01
-2.55264670e-01 -8.31493556e-01 -3.31259966e-01 2.73444206e-01
-9.84909534e-01 -6.32820785e-01 -6.90617383e-01 -6.78310990e-01
3.71519953e-01 2.86193550e-01 5.01064479e-01 2.91615129e-01
6.90990537e-02 -5.66114724e-01 -4.11276430e-01 -3.37595642e-01
-9.12033170e-02 9.77430642e-02 -1.40901171e-02 4.61119181e-03
4.59602326e-01 -3.97515237e-01 -6.36288762e-01 8.53939131e-02
-1.09298408e+00 1.17303967e-01 7.84109116e-01 1.01914155e+00
7.85829961e-01 6.96511745e-01 4.35200632e-01 -3.14095467e-01
4.84143943e-01 -1.76977903e-01 -7.86945403e-01 1.51844919e-01
-1.02986217e+00 1.21510804e-01 8.22850227e-01 -5.76406360e-01
-9.14343953e-01 -2.37171724e-01 -1.33335888e-01 -5.09192765e-01
4.50171798e-01 6.57643437e-01 -3.99218529e-01 -2.20224127e-01
3.30794603e-01 4.44179803e-01 -8.76173973e-02 -7.24492550e-01
3.93505320e-02 1.01463175e+00 4.56700534e-01 -2.80565113e-01
3.83099526e-01 -3.55732627e-02 8.79529789e-02 -6.60443485e-01
-3.29191417e-01 -1.11060135e-01 -1.31257594e-01 1.97879076e-01
5.86121798e-01 -9.68882918e-01 -6.22922003e-01 6.94861889e-01
-1.14703333e+00 -1.51835591e-01 1.05402134e-01 6.56945288e-01
-8.52361843e-02 6.87772512e-01 -7.36495733e-01 -7.73177028e-01
-5.54374695e-01 -1.34448004e+00 7.08323002e-01 4.78414446e-01
2.47292370e-01 -4.24187481e-01 -5.54665983e-01 2.84608215e-01
5.13952792e-01 -3.23459953e-01 1.02224326e+00 -5.11125982e-01
-5.63573539e-01 -3.29528302e-01 -3.70450765e-01 7.05888987e-01
-1.22852691e-01 -1.75270185e-01 -8.02106559e-01 -1.34624898e-01
2.91321933e-01 -1.32411748e-01 7.51567423e-01 1.67029351e-01
1.54321110e+00 -2.89128751e-01 -2.97218293e-01 8.13544154e-01
1.54644203e+00 5.37572622e-01 9.03510511e-01 2.40816802e-01
7.51617312e-01 -1.55678511e-01 7.38450289e-01 6.65467143e-01
2.40225762e-01 7.57444382e-01 3.52602065e-01 1.28418133e-01
-1.69599801e-01 -2.81182945e-01 3.61629836e-02 1.16872418e+00
5.74879572e-02 -3.32398027e-01 -6.92010343e-01 3.06944937e-01
-1.49576771e+00 -6.42678261e-01 -1.25593215e-01 2.42402649e+00
1.15634310e+00 3.28073531e-01 -4.52150047e-01 7.27720737e-01
7.59427428e-01 5.95823899e-02 -4.67997342e-01 -5.09122372e-01
-1.41417310e-01 4.03987974e-01 8.54341507e-01 2.36987099e-01
-7.90881872e-01 8.21180522e-01 5.36855412e+00 1.64845860e+00
-1.42731750e+00 9.35182124e-02 6.50159478e-01 -8.98426678e-03
3.49542908e-02 -1.07325211e-01 -1.23019886e+00 1.00043511e+00
1.03678107e+00 -2.23182961e-02 6.34289384e-01 8.53940248e-01
6.08947873e-02 -1.99769676e-01 -7.45985746e-01 1.34620380e+00
9.88983810e-02 -9.83502030e-01 3.51045668e-01 -8.39179382e-02
5.01279235e-01 -2.45021403e-01 1.07562058e-01 5.00747442e-01
-5.10362089e-01 -9.81629431e-01 7.07587779e-01 3.11233044e-01
9.03464675e-01 -8.59705925e-01 8.63025963e-01 3.57728988e-01
-1.15979207e+00 -2.27876678e-01 -8.15218985e-01 -2.41871983e-01
1.20619208e-01 7.57541001e-01 -6.01340115e-01 4.35253471e-01
6.54315889e-01 3.19081664e-01 -6.47654057e-01 1.29668748e+00
-8.40570033e-02 6.39858007e-01 -6.55734539e-01 -1.81494042e-01
1.04945496e-01 -2.16259465e-01 7.79341627e-03 8.83782804e-01
6.02668881e-01 3.41191173e-01 -1.91128448e-01 6.89492464e-01
6.56125546e-02 8.06003362e-02 -2.37145852e-02 -1.08855173e-01
9.59985197e-01 1.01303351e+00 -5.24866462e-01 -6.79530919e-01
-1.46848992e-01 7.27456927e-01 1.50468975e-01 2.64269859e-01
-1.05520558e+00 -8.18294883e-01 1.41029432e-01 -3.15580480e-02
7.53185272e-01 -1.18583806e-01 -5.41605830e-01 -1.04852414e+00
1.16977893e-01 -8.39025378e-01 5.14906570e-02 -9.38911438e-01
-5.75697303e-01 4.66947734e-01 1.44645071e-03 -1.09777057e+00
1.84338968e-02 -3.92022520e-01 -2.94841826e-01 9.75847542e-01
-1.64294946e+00 -6.33671463e-01 -4.71784204e-01 5.00388503e-01
1.96028903e-01 1.44915819e-01 6.46538496e-01 7.74743617e-01
-9.31970239e-01 1.12853372e+00 3.16419423e-01 -2.53597856e-01
3.71328652e-01 -5.47657430e-01 4.57373820e-02 8.62784863e-01
-1.84961155e-01 6.10149860e-01 5.26212752e-01 -5.45842886e-01
-1.38382387e+00 -9.32592273e-01 9.97881413e-01 4.42330748e-01
2.35534385e-01 -1.68367580e-01 -1.13698256e+00 2.66177744e-01
-2.61270881e-01 -1.23248294e-01 2.82768220e-01 -5.04769146e-01
-6.58922791e-02 -5.58409095e-01 -1.26556027e+00 5.27932823e-01
6.13746524e-01 -2.45654717e-01 -4.27350879e-01 1.64785922e-01
7.49068737e-01 -4.28466231e-01 -9.77147520e-01 6.41877234e-01
2.44391918e-01 -9.10636008e-01 8.75312448e-01 1.14801325e-01
1.73601255e-01 -4.60210443e-01 -3.98511350e-01 -9.42676842e-01
-4.02354598e-01 -2.92067379e-01 -3.37144196e-01 1.57510495e+00
2.05055821e-05 -6.74020410e-01 6.40585780e-01 6.36323631e-01
-2.68910617e-01 -1.00666308e+00 -1.09563231e+00 -7.72553027e-01
-1.22146860e-01 -1.92694172e-01 1.05716717e+00 5.74902534e-01
5.06179295e-02 2.89725214e-01 -2.76516318e-01 -9.28201154e-02
3.01351011e-01 -9.06649381e-02 3.95240337e-01 -7.65563130e-01
-3.75690132e-01 -5.45443535e-01 -4.98242199e-01 -1.40449870e+00
-3.23682547e-01 -7.02667236e-01 4.20884676e-02 -1.35254478e+00
-5.38852476e-02 -8.16420436e-01 -4.25344229e-01 4.89189833e-01
-2.65387416e-01 3.47269893e-01 3.13649893e-01 3.46220315e-01
-2.74093091e-01 8.59838963e-01 1.36159480e+00 -1.64548174e-01
-6.33351654e-02 -2.49752820e-01 -5.06761134e-01 4.40232635e-01
7.01778710e-01 -3.87968093e-01 -4.35901523e-01 -4.95776266e-01
-3.46583985e-02 4.57307650e-03 1.46855041e-01 -1.27188456e+00
2.36884892e-01 5.55953979e-02 4.21576262e-01 -7.43821263e-01
7.06909716e-01 -8.44914436e-01 4.82840762e-02 5.90268970e-01
5.47548644e-02 -1.97384530e-03 3.01000506e-01 3.09793502e-01
-3.68789017e-01 -6.64178014e-01 7.65721440e-01 1.71378121e-01
-6.82235658e-01 1.67791784e-01 -2.14173168e-01 -2.41504073e-01
6.80187762e-01 -3.12073201e-01 -2.67552108e-01 -5.29780611e-02
-1.10318512e-01 7.31225982e-02 3.39898914e-01 -7.03713074e-02
4.81816709e-01 -1.38233507e+00 -9.93344486e-02 6.14802182e-01
-2.93146312e-01 4.50351864e-01 4.58396941e-01 8.34507346e-01
-6.04071319e-01 6.49565816e-01 2.23752465e-02 -3.88186991e-01
-9.48584557e-01 7.11668313e-01 1.92416683e-01 -3.10724378e-01
-2.13148504e-01 8.11392546e-01 -1.23968385e-01 2.16007397e-01
4.82747227e-01 -7.10863471e-01 -2.09085882e-01 -5.24997823e-02
5.99315464e-01 5.62284529e-01 3.77384335e-01 -6.04627550e-01
-2.60165960e-01 5.07395089e-01 -1.35351524e-01 1.73870176e-01
8.12547266e-01 -1.05736665e-01 7.44504528e-03 1.13137715e-01
1.18353128e+00 -5.15372932e-01 -1.08056879e+00 -1.84180737e-01
-4.43598896e-01 -4.14462715e-01 4.66020197e-01 -7.05800235e-01
-1.22089589e+00 1.18642938e+00 7.36033320e-01 -1.93951838e-02
1.45600057e+00 -6.33393764e-01 1.16887510e+00 3.59401762e-01
3.03354293e-01 -1.20698988e+00 -2.56123096e-01 3.88639063e-01
4.44269538e-01 -9.06170428e-01 2.73093432e-01 -3.64677072e-01
-3.88699323e-01 7.96056747e-01 6.92663968e-01 6.47004172e-02
6.04421437e-01 3.56558412e-02 -2.49537215e-01 2.66287863e-01
-2.87612677e-01 1.81345835e-01 1.11206882e-01 4.16286230e-01
9.26467776e-02 -5.32879308e-03 -7.80917943e-01 1.03850400e+00
-3.87528926e-01 2.00525612e-01 2.81588376e-01 7.92462349e-01
-4.71682161e-01 -1.02566242e+00 -4.80509162e-01 5.90532541e-01
-4.53500956e-01 -5.72575748e-01 3.59229445e-01 5.68555892e-01
3.89946997e-01 1.08701587e+00 1.62355319e-01 -6.48816466e-01
2.21303388e-01 1.13658078e-01 5.44906795e-01 -1.60969302e-01
-4.31033403e-01 1.75157085e-01 -2.30246767e-01 -4.04065609e-01
-1.08971223e-01 -2.52985567e-01 -1.49361908e+00 -4.10826623e-01
-8.52020085e-01 1.04716472e-01 9.92707014e-01 8.87033761e-01
4.83275771e-01 5.80836117e-01 3.71013463e-01 -8.84887815e-01
-1.07459676e+00 -1.20018530e+00 -7.07328498e-01 3.27876300e-01
2.12922711e-02 -8.03192317e-01 -2.97191381e-01 -2.59381920e-01] | [8.622784614562988, 2.986327886581421] |
dc63e9d6-4948-4d68-81f9-5ebe26836af5 | frequency-and-temporal-convolutional | 1910.07364 | null | https://arxiv.org/abs/1910.07364v2 | https://arxiv.org/pdf/1910.07364v2.pdf | Frequency and temporal convolutional attention for text-independent speaker recognition | Majority of the recent approaches for text-independent speaker recognition apply attention or similar techniques for aggregation of frame-level feature descriptors generated by a deep neural network (DNN) front-end. In this paper, we propose methods of convolutional attention for independently modelling temporal and frequency information in a convolutional neural network (CNN) based front-end. Our system utilizes convolutional block attention modules (CBAMs) [1] appropriately modified to accommodate spectrogram inputs. The proposed CNN front-end fitted with the proposed convolutional attention modules outperform the no-attention and spatial-CBAM baselines by a significant margin on the VoxCeleb [2, 3] speaker verification benchmark, and our best model achieves an equal error rate of 2:031% on the VoxCeleb1 test set, improving the existing state of the art result by a significant margin. For a more thorough assessment of the effects of frequency and temporal attention in real-world conditions, we conduct ablation experiments by randomly dropping frequency bins and temporal frames from the input spectrograms, concluding that instead of modelling either of the entities, simultaneously modelling temporal and frequency attention translates to better real-world performance. | ['Sarthak Yadav', 'Atul Rai'] | 2019-10-16 | null | null | null | null | ['text-independent-speaker-recognition'] | ['speech'] | [ 1.95765808e-01 -1.62166134e-01 4.25672010e-02 -6.01394892e-01
-1.07666314e+00 -3.14985335e-01 7.09695458e-01 -1.16327226e-01
-5.81366122e-01 3.23109776e-01 4.98873383e-01 -2.75398493e-01
7.96367675e-02 -1.60176516e-01 -5.64029336e-01 -6.99262321e-01
-2.94020891e-01 -1.10061407e-01 -8.50003883e-02 -8.25143307e-02
1.88051403e-01 3.74488711e-01 -1.76537156e+00 3.62972617e-01
5.18300593e-01 1.18469095e+00 -1.59613907e-01 1.05438292e+00
1.89450741e-01 9.38735723e-01 -1.01971018e+00 -2.77380615e-01
-1.04116462e-01 -3.59248221e-01 -7.02695310e-01 -3.15930814e-01
8.63378584e-01 -3.35124344e-01 -4.59451050e-01 6.19592905e-01
9.14400280e-01 2.89880693e-01 4.95533139e-01 -1.14542747e+00
-8.24272335e-01 7.51537979e-01 -3.16380769e-01 7.40813673e-01
3.75674129e-01 3.22103202e-01 1.06168056e+00 -8.50972831e-01
1.39571931e-02 1.22104943e+00 8.41679394e-01 5.52851379e-01
-1.03346395e+00 -7.17198431e-01 2.47540504e-01 6.28828585e-01
-1.54381073e+00 -1.12642539e+00 8.93612385e-01 -2.07724810e-01
1.50004983e+00 4.45266634e-01 2.88591564e-01 1.39392662e+00
4.53910641e-02 7.10578859e-01 5.91823697e-01 -4.90495473e-01
9.73360315e-02 3.86350527e-02 3.58639002e-01 1.17768921e-01
-4.28335249e-01 3.25709194e-01 -1.13357127e+00 3.87310423e-02
3.02210212e-01 -4.52171385e-01 -3.43562514e-01 4.95746821e-01
-1.19921005e+00 6.76392436e-01 4.29639041e-01 4.41430181e-01
-3.64260256e-01 3.44135404e-01 6.83976591e-01 2.23139927e-01
6.51379228e-01 -1.11759603e-01 -4.39454913e-01 -3.22056055e-01
-1.36574757e+00 1.08550921e-01 4.62854385e-01 8.37771833e-01
1.80803165e-01 6.91620946e-01 -5.98357916e-01 8.52880299e-01
3.72638196e-01 3.84669065e-01 6.74205601e-01 -4.09506351e-01
3.72480869e-01 -1.72099546e-01 -1.24444522e-01 -4.10068899e-01
-3.52770329e-01 -6.80717289e-01 -6.78258717e-01 -1.65239051e-01
2.47938067e-01 -2.08708525e-01 -1.22552669e+00 1.86039388e+00
1.90929815e-01 4.88261580e-01 1.21057093e-01 9.51186657e-01
1.15190399e+00 8.09490919e-01 3.79784554e-01 -5.93658648e-02
1.60211039e+00 -9.97902811e-01 -1.04410160e+00 -8.04573447e-02
2.42083013e-01 -8.90140533e-01 8.85294080e-01 2.20063269e-01
-9.75196123e-01 -9.77767110e-01 -1.14558685e+00 -3.48143607e-01
-3.21807176e-01 1.81512460e-01 2.59277761e-01 9.55964327e-01
-1.36768353e+00 3.56414497e-01 -7.02632606e-01 -3.60469967e-01
3.22230726e-01 3.65983278e-01 -1.86907515e-01 3.36806804e-01
-1.37736177e+00 8.36435914e-01 -3.81026380e-02 2.88545936e-01
-1.18417490e+00 -9.71176088e-01 -1.03376496e+00 3.31114471e-01
-2.53765136e-01 -3.42492491e-01 1.52018404e+00 -9.47015047e-01
-1.91068804e+00 5.70081711e-01 -5.71171761e-01 -8.78243864e-01
2.49666438e-01 -3.12986642e-01 -8.93307030e-01 3.36735286e-02
-2.72034824e-01 7.74653018e-01 1.14486134e+00 -5.80253184e-01
-4.81152058e-01 -6.41286895e-02 -8.53290930e-02 -1.23224325e-01
-4.61567432e-01 3.99499834e-01 -1.31391674e-01 -8.67474914e-01
-2.32855082e-01 -5.68234026e-01 2.84497887e-01 -2.84828514e-01
-3.58589500e-01 -3.90308827e-01 1.07005489e+00 -9.47161734e-01
1.20369446e+00 -2.46548700e+00 -6.44987896e-02 -2.69199669e-01
-1.64888576e-01 3.24528307e-01 -3.41617227e-01 3.54983479e-01
-4.71239924e-01 1.25790447e-01 -1.30263418e-01 -8.03843260e-01
1.52915657e-01 -2.37161443e-01 -3.13931346e-01 6.40378356e-01
5.25162578e-01 7.46179521e-01 -4.32364732e-01 -1.36742622e-01
2.84054726e-01 1.04964244e+00 -6.30480528e-01 5.05601950e-02
5.97629398e-02 5.06732285e-01 2.92593688e-01 5.05567014e-01
6.64078236e-01 1.47896275e-01 -2.38300398e-01 -2.25594759e-01
-1.11976348e-01 9.16388988e-01 -7.99060106e-01 1.73134732e+00
-6.29140913e-01 1.22111130e+00 8.85252059e-02 -7.96358347e-01
7.67792046e-01 9.10881579e-01 2.51565784e-01 -8.05943549e-01
5.79759292e-02 -3.10021751e-02 3.99094999e-01 -2.91170895e-01
3.65991265e-01 -2.71963984e-01 5.06970249e-02 3.14532131e-01
5.76769710e-01 1.87665820e-01 -2.57150143e-01 2.92198509e-02
1.03028107e+00 -2.38216117e-01 -2.92760618e-02 -3.73096496e-01
9.47164655e-01 -6.45306349e-01 4.01716501e-01 6.24280751e-01
-7.07843423e-01 8.11218262e-01 2.25751892e-01 -3.79990697e-01
-9.44361806e-01 -7.88658440e-01 -3.71682107e-01 1.50715363e+00
-3.82324100e-01 -4.31561261e-01 -8.29051614e-01 -4.77893203e-01
-8.39240253e-02 8.48304868e-01 -8.66399765e-01 -6.70354255e-03
-5.85718095e-01 -5.56525230e-01 1.01576924e+00 6.31163180e-01
5.17670155e-01 -1.16478574e+00 -4.37275648e-01 3.75008374e-01
-3.12086076e-01 -1.23739660e+00 -7.56562173e-01 3.53242725e-01
-3.93412679e-01 -4.51474398e-01 -9.62642908e-01 -5.75540185e-01
-4.30822186e-02 5.58224209e-02 9.39218819e-01 -1.55921429e-01
-2.13534251e-01 2.39857092e-01 -4.40436125e-01 -6.46839797e-01
-1.78429544e-01 2.66871035e-01 7.36807138e-02 3.55879873e-01
5.16317785e-01 -5.64555347e-01 -7.36670732e-01 1.71539471e-01
-7.83575118e-01 -2.91172922e-01 1.36763960e-01 9.40493166e-01
-4.88976836e-02 -3.88007402e-01 9.91951108e-01 -1.98620245e-01
4.07886565e-01 -2.42682859e-01 -4.20920461e-01 -9.60020944e-02
-1.27915621e-01 1.96138844e-02 4.06264812e-01 -5.57777166e-01
-1.07147598e+00 -8.03793520e-02 -5.54548979e-01 -5.10183692e-01
-3.83600324e-01 3.31170201e-01 -2.18503669e-01 1.37700722e-01
5.88060856e-01 4.13097918e-01 -2.13232026e-01 -4.73972470e-01
2.63003141e-01 1.00456607e+00 5.88972092e-01 -1.81313396e-01
3.68847251e-01 3.31291676e-01 -6.08164370e-01 -9.57511961e-01
-6.02985024e-01 -5.57898939e-01 -4.98308450e-01 -1.37778878e-01
1.00837386e+00 -1.12030268e+00 -8.10427189e-01 6.41648054e-01
-1.42321074e+00 -2.52766013e-01 -1.36462465e-01 6.08893156e-01
-3.70171249e-01 5.37848212e-02 -6.54471517e-01 -1.02415109e+00
-4.74432170e-01 -1.23109198e+00 1.36389136e+00 6.96742684e-02
-3.47899109e-01 -7.60271251e-01 2.23966036e-02 4.68057364e-01
9.50289547e-01 -1.80670992e-01 6.49316013e-01 -8.98417830e-01
-3.23401362e-01 -6.87101707e-02 4.05882597e-02 3.28420252e-01
8.53692442e-02 1.18002653e-01 -2.00019765e+00 -3.80580723e-01
2.40279689e-01 6.54637218e-02 1.07489073e+00 6.42497122e-01
1.08355331e+00 -3.39970022e-01 9.87701938e-02 6.25713944e-01
9.65445638e-01 3.63615304e-01 6.81754172e-01 8.08599889e-02
3.94645333e-01 4.88708705e-01 -2.13868991e-02 3.65311444e-01
3.03921878e-01 9.86413538e-01 2.66405314e-01 -2.62670398e-01
-5.47537804e-01 1.79503843e-01 6.44802153e-01 8.60506713e-01
9.38818604e-02 -5.82304955e-01 -8.63456130e-01 9.63770509e-01
-1.50530720e+00 -1.23214018e+00 1.85408369e-01 2.07181716e+00
6.12344623e-01 3.45964320e-02 2.96607077e-01 5.29884934e-01
8.37618828e-01 3.96962374e-01 -4.87081856e-01 -6.36961162e-01
-1.43964052e-01 5.04687548e-01 1.70215681e-01 6.90065742e-01
-1.29761314e+00 6.92671239e-01 6.58808661e+00 8.55681598e-01
-1.56971717e+00 3.67449701e-01 6.64302289e-01 -4.60597247e-01
1.16621584e-01 -4.57270831e-01 -8.61283481e-01 4.12435919e-01
1.73093271e+00 -3.10877394e-02 2.61082947e-01 5.06639063e-01
3.98420006e-01 2.19035596e-01 -1.31868362e+00 1.06448436e+00
2.98786670e-01 -1.14388669e+00 -1.43803477e-01 8.59515965e-02
4.33557063e-01 1.89777136e-01 4.03390527e-01 4.40675557e-01
-9.97378305e-02 -1.31442750e+00 1.13385594e+00 2.99727678e-01
8.79826665e-01 -7.82645047e-01 8.65517795e-01 -1.13106181e-03
-1.32410693e+00 -7.99544081e-02 1.28862247e-01 -7.74700418e-02
2.17346489e-01 2.63243169e-01 -9.09013629e-01 5.14536917e-01
8.80963802e-01 4.84497726e-01 -3.36186945e-01 9.46764827e-01
5.95614985e-02 1.09977341e+00 -3.55144054e-01 1.88974366e-01
2.69532979e-01 6.56254053e-01 5.89735329e-01 1.63209713e+00
1.05917864e-01 -1.71232820e-01 -6.65481269e-01 7.33886957e-01
-6.86388761e-02 -1.03107534e-01 -2.89983064e-01 9.93788987e-02
4.04991120e-01 1.06968927e+00 -3.35370570e-01 -4.13567722e-01
-5.50123632e-01 8.39243472e-01 1.25718936e-01 6.39520645e-01
-1.26700211e+00 -6.09707832e-01 9.61242437e-01 -1.57438353e-01
7.47601271e-01 -9.28988978e-02 -3.26584727e-01 -9.96737182e-01
6.62619397e-02 -8.40413630e-01 2.93541551e-01 -6.09441817e-01
-1.22880876e+00 1.08770347e+00 -1.85808808e-01 -1.09470546e+00
-4.65163648e-01 -4.50774610e-01 -7.98053145e-01 1.32457745e+00
-1.58930278e+00 -1.19617641e+00 -1.19032592e-01 6.69369817e-01
8.96187603e-01 -2.08540410e-01 9.08813894e-01 5.05344927e-01
-7.57714808e-01 1.17976725e+00 -1.30951688e-01 2.78609961e-01
7.73657262e-01 -1.00566709e+00 8.87089014e-01 1.05214345e+00
5.02562165e-01 5.66922605e-01 5.70170403e-01 -6.84727952e-02
-1.14126551e+00 -1.04088116e+00 1.11589003e+00 -3.23536605e-01
3.08359236e-01 -8.05243969e-01 -1.04201794e+00 6.27498627e-01
8.30219686e-01 1.63096130e-01 6.36094391e-01 2.01247275e-01
-6.22229755e-01 -2.69519687e-01 -1.12604201e+00 2.51240373e-01
7.04652965e-01 -9.88224626e-01 -4.93707508e-01 1.62630692e-01
8.06206882e-01 -3.03848594e-01 -7.17670798e-01 4.83546078e-01
5.74686289e-01 -1.04031730e+00 1.04579365e+00 -4.74081218e-01
9.97721106e-02 -2.29748756e-01 -1.94135457e-01 -1.16614771e+00
-2.77715951e-01 -7.60323703e-01 -9.34867933e-02 1.38421631e+00
4.96750534e-01 -4.92038250e-01 4.84607249e-01 2.15265661e-01
-3.29473138e-01 -2.76693076e-01 -1.56940091e+00 -6.48120284e-01
1.35725951e-02 -7.84442604e-01 6.33859158e-01 7.67141044e-01
4.33685258e-03 3.58473659e-01 -3.61304671e-01 3.90508622e-01
2.83555269e-01 -4.25578922e-01 4.67024207e-01 -8.17930818e-01
-7.66120851e-02 -6.67759180e-01 -6.69714987e-01 -1.02987492e+00
2.15796620e-01 -5.45223892e-01 2.33255535e-01 -1.09025049e+00
-1.02736028e-02 8.52389708e-02 -4.76267874e-01 4.00784612e-01
-1.28329545e-01 4.20977414e-01 1.99387312e-01 -5.79580627e-02
-3.62516195e-01 7.33716965e-01 6.44187570e-01 -3.52712214e-01
-7.26265758e-02 -2.19985783e-01 -4.11656588e-01 2.99175471e-01
5.08014679e-01 -1.89788595e-01 -2.40168527e-01 -5.33022165e-01
-6.60962880e-01 1.00455046e-01 5.25281966e-01 -1.27601385e+00
5.11328280e-01 4.55690563e-01 5.34455061e-01 -6.40771449e-01
5.83673954e-01 -4.80835021e-01 5.08144684e-02 2.18551427e-01
-5.62627673e-01 -6.40332028e-02 9.48397458e-01 3.24999452e-01
-6.20620430e-01 4.91316542e-02 8.03137779e-01 3.75809699e-01
-4.18195397e-01 5.36531433e-02 -7.44928658e-01 -3.10410589e-01
5.23604512e-01 -1.35089427e-01 -3.17235559e-01 -7.11254060e-01
-9.02140498e-01 -2.87612110e-01 -2.99226373e-01 5.73695183e-01
3.65627497e-01 -1.31070244e+00 -1.01774156e+00 5.69631040e-01
5.55026494e-02 -4.47422802e-01 6.36160254e-01 7.89164662e-01
-1.87133029e-01 8.74069035e-01 5.44006936e-02 -9.43984389e-01
-1.21980321e+00 3.51919740e-01 5.98128736e-01 5.05440906e-02
-4.63676393e-01 1.19215882e+00 3.84899199e-01 -1.87905416e-01
7.45811641e-01 -6.25288785e-01 -1.10437535e-01 1.47697059e-02
7.76414812e-01 1.32038310e-01 6.12127066e-01 -1.01259553e+00
-7.37522066e-01 3.94120961e-01 -2.93643147e-01 -4.58008200e-01
1.21011341e+00 -8.68284851e-02 5.24227262e-01 3.77543300e-01
1.25006270e+00 7.16698309e-03 -1.17414486e+00 -2.40737408e-01
-1.46787405e-01 -4.95329686e-02 4.29075420e-01 -9.90280986e-01
-1.12354791e+00 1.27744567e+00 8.12447369e-01 2.85653442e-01
1.28427398e+00 -1.93147793e-01 4.84045386e-01 -1.07958078e-01
-2.71783173e-01 -5.93673527e-01 -9.18234736e-02 6.13682151e-01
1.24851310e+00 -1.12487733e+00 -3.59138906e-01 -2.81027481e-02
-4.70740169e-01 1.11090660e+00 4.69547451e-01 1.81955755e-01
7.66129792e-01 3.86385590e-01 3.88608843e-01 7.30716065e-02
-9.58453774e-01 -1.45162597e-01 5.81244171e-01 5.23849010e-01
9.09748852e-01 -1.71512201e-01 1.98404461e-01 4.91865546e-01
-3.41087729e-01 -3.00999254e-01 3.16173285e-02 6.94878221e-01
-7.51036452e-03 -7.63458431e-01 -5.69396079e-01 -6.86799213e-02
-7.08189189e-01 -5.00890613e-01 -9.60134417e-02 6.80184841e-01
-4.29911502e-02 1.27536404e+00 3.94381642e-01 -2.99039245e-01
2.76345640e-01 4.02008682e-01 2.26778716e-01 -2.18934879e-01
-1.01865673e+00 4.30320680e-01 3.36692631e-02 -3.71937603e-01
-5.31322062e-01 -6.90660715e-01 -8.13129425e-01 -2.92433172e-01
-4.07842666e-01 -2.27430407e-02 8.16943049e-01 8.90046954e-01
5.78753233e-01 1.03506303e+00 4.65680242e-01 -1.01035559e+00
-2.75706679e-01 -1.70886278e+00 -2.54050732e-01 2.67678231e-01
1.02871299e+00 -5.23188889e-01 -5.48065484e-01 3.44177604e-01] | [14.418307304382324, 6.0310378074646] |
536c4727-f2eb-4299-a58b-e5d1d0ef7dda | hybrid-machine-learning-models-for-crop-yield | 2005.04155 | null | https://arxiv.org/abs/2005.04155v1 | https://arxiv.org/pdf/2005.04155v1.pdf | Hybrid Machine Learning Models for Crop Yield Prediction | Prediction of crop yield is essential for food security policymaking, planning, and trade. The objective of the current study is to propose novel crop yield prediction models based on hybrid machine learning methods. In this study, the performance of the artificial neural networks-imperialist competitive algorithm (ANN-ICA) and artificial neural networks-gray wolf optimizer (ANN-GWO) models for the crop yield prediction are evaluated. According to the results, ANN-GWO, with R of 0.48, RMSE of 3.19, and MEA of 26.65, proved a better performance in the crop yield prediction compared to the ANN-ICA model. The results can be used by either practitioners, researchers or policymakers for food security. | ['Bertalan Beszedes', 'Saeed Nosratabadi', 'Karoly Szell', 'Felde Imre', 'Amir Mosavi', 'Sina Ardabili'] | 2020-03-08 | null | null | null | null | ['crop-yield-prediction', 'crop-yield-prediction'] | ['computer-vision', 'miscellaneous'] | [-1.65801436e-01 -1.90122962e-01 -6.62868023e-01 2.60864615e-01
1.00103855e-01 -3.78725737e-01 -1.28817767e-01 4.70528305e-01
-2.26597369e-01 1.11267090e+00 -2.69013017e-01 -8.04876983e-01
-2.64840722e-01 -1.13838696e+00 -2.09984645e-01 -9.26048338e-01
-1.43282458e-01 -1.23582415e-01 -3.96448344e-01 -3.95909458e-01
5.72538674e-01 7.21505582e-01 -1.25842106e+00 -5.71609318e-01
1.33446181e+00 1.18584657e+00 4.06468719e-01 5.75500190e-01
1.25712886e-01 4.06545460e-01 -2.23721087e-01 -2.40528539e-01
2.34187230e-01 -5.44693291e-01 -3.36552203e-01 -5.62979639e-01
-1.02009583e+00 -3.24476808e-01 3.58617783e-01 9.53033447e-01
4.82043117e-01 -1.68537527e-01 7.68232942e-01 -1.10048747e+00
-1.15726686e+00 7.13350773e-01 -7.50720859e-01 1.56176135e-01
-7.51733184e-02 3.09056733e-02 4.35150355e-01 -7.34990001e-01
1.82932556e-01 9.98470426e-01 6.82039201e-01 -1.03566907e-01
-8.45439315e-01 -7.99105525e-01 -2.73143023e-01 4.43618476e-01
-1.38928199e+00 1.07214965e-01 5.65750241e-01 -1.40245125e-01
9.57488298e-01 3.78305346e-01 1.12963831e+00 4.50266115e-02
5.52103519e-01 4.57106888e-01 1.11025882e+00 -7.38894343e-01
2.00019985e-01 5.15261516e-02 2.68706791e-02 6.38322055e-01
1.12051642e+00 6.61883652e-01 -5.95065504e-02 -6.82182536e-02
6.51766658e-01 -1.02870718e-01 1.21341139e-01 2.49924988e-01
-8.82379949e-01 1.26392281e+00 5.61149776e-01 5.04954159e-01
-8.80149901e-01 -1.82567194e-01 -2.16789823e-02 3.64239186e-01
6.72788262e-01 8.57733846e-01 -8.56882036e-01 1.38525099e-01
-7.52469063e-01 8.42335820e-02 1.04711401e+00 5.46308041e-01
2.89930493e-01 6.08312547e-01 3.00278515e-01 2.78745830e-01
5.45737267e-01 1.09235275e+00 4.75985676e-01 -6.24808013e-01
-1.25411917e-02 7.12439656e-01 4.87138689e-01 -1.49008918e+00
-6.52040362e-01 -3.40174794e-01 -9.29165423e-01 1.80943117e-01
4.29243773e-01 -5.31710148e-01 -5.36753297e-01 1.08763754e+00
1.57306418e-01 -3.57755795e-02 7.25847006e-01 6.74150586e-01
8.43820632e-01 1.03366292e+00 4.48262066e-01 -5.45134664e-01
1.11931384e+00 -8.77671599e-01 -9.38816249e-01 3.28702897e-01
8.37959230e-01 -8.27359915e-01 3.68978709e-01 2.15653971e-01
-9.65391755e-01 -3.57470274e-01 -1.23488891e+00 5.89108586e-01
-8.26366425e-01 6.33983374e-01 8.93118560e-01 8.29297006e-01
-6.32604063e-01 6.11335814e-01 -7.09949136e-01 -5.58616161e-01
8.60190392e-03 2.34711841e-01 4.85571288e-02 3.76792341e-01
-1.00024235e+00 1.47853339e+00 1.10270667e+00 5.96762538e-01
-2.37318411e-01 -4.24104631e-01 -6.32648408e-01 2.88068026e-01
6.59902617e-02 -2.45390162e-01 7.24111021e-01 -7.57463753e-01
-1.56713045e+00 6.49214745e-01 -4.69093882e-02 -5.92967808e-01
-7.06210956e-02 -7.19074234e-02 -6.59517765e-01 -5.36281290e-03
-6.03649616e-02 4.10517454e-01 1.90879673e-01 -1.16518211e+00
-7.88132489e-01 -6.12863958e-01 -4.48977888e-01 7.30983540e-02
-1.70420781e-01 -1.17864562e-02 8.52736056e-01 -7.55091786e-01
5.29809713e-01 -9.96594608e-01 -4.55764562e-01 -4.38721418e-01
-2.65816525e-02 2.71205585e-02 6.79281056e-01 -1.20576000e+00
1.58615625e+00 -1.51682591e+00 -3.47243220e-01 5.84097743e-01
-5.12384713e-01 6.66723073e-01 7.00879693e-02 3.28832656e-01
1.84295569e-02 2.91204363e-01 -5.29923178e-02 8.84750068e-01
-2.67316639e-01 3.28823850e-02 1.23628117e-01 9.27320123e-02
5.40252388e-01 1.17042363e+00 -1.07577348e+00 -2.87258595e-01
3.52761835e-01 -2.21428629e-02 5.86000495e-02 9.01123434e-02
1.28862515e-01 4.98987362e-02 -9.45114017e-01 1.26171851e+00
9.88571763e-01 -5.92641272e-02 2.75723755e-01 -2.53002018e-01
-5.82607687e-01 -7.12998033e-01 -9.78357613e-01 1.03624213e+00
-1.17935270e-01 9.74898115e-02 -3.37306768e-01 -1.12879848e+00
1.58551168e+00 3.81918877e-01 3.35403204e-01 -6.29740179e-01
4.37125891e-01 7.61897147e-01 2.12693095e-01 -7.33619452e-01
2.87376136e-01 3.75801861e-01 3.39754462e-01 2.42968902e-01
-1.17221102e-01 3.27767283e-02 4.42719311e-01 -6.15675449e-01
2.67169297e-01 3.84296507e-01 7.69164026e-01 -7.45388269e-01
7.54733086e-01 4.87148315e-01 3.95205528e-01 2.84805954e-01
-1.64866865e-01 -1.94696561e-01 2.94619679e-01 -6.62162483e-01
-1.03666711e+00 -5.03856301e-01 -1.72270700e-01 1.00841939e+00
2.32795760e-01 6.05158865e-01 -5.58240950e-01 -1.05728701e-01
1.19623877e-01 1.02518320e+00 -3.54299724e-01 -3.24708402e-01
-3.51090610e-01 -1.45204794e+00 3.49264413e-01 6.08938515e-01
8.59821916e-01 -1.22220874e+00 -1.14152431e+00 6.32753611e-01
-4.07186821e-02 -6.84252262e-01 2.08721921e-01 3.39970648e-01
-1.20429492e+00 -8.96873534e-01 -9.37760115e-01 -7.03265727e-01
5.98557770e-01 1.97478294e-01 9.07306135e-01 4.64941651e-01
-1.32525163e-02 -2.95195162e-01 -8.60774219e-01 -1.06275177e+00
-6.19778216e-01 2.28126749e-01 -2.62621641e-01 -4.94244009e-01
6.98382497e-01 -4.16724801e-01 -7.32886791e-01 1.18605673e-01
-6.35041296e-01 -1.25891432e-01 7.67731726e-01 1.01323164e+00
3.38555396e-01 3.20732862e-01 1.22518396e+00 -1.35039657e-01
5.60807526e-01 -7.71579742e-01 -1.08625960e+00 8.03094864e-01
-1.32354045e+00 8.31853226e-02 3.04595530e-01 -4.55169380e-01
-1.05274069e+00 1.35602251e-01 1.65569112e-01 2.82814384e-01
2.00275138e-01 9.73178744e-01 5.02290912e-02 -2.58764684e-01
3.49572778e-01 1.00837283e-01 1.24881670e-01 -3.56148541e-01
-9.76486132e-02 4.92508560e-01 2.22342417e-01 -1.79814160e-01
5.29519439e-01 -3.67351264e-01 4.98142719e-01 -5.77312171e-01
-4.58207399e-01 -5.31788282e-02 -4.75315899e-01 -2.23009974e-01
9.06235456e-01 -6.91463292e-01 -8.47647250e-01 6.19136810e-01
-1.00140035e+00 -1.34996012e-01 6.25648350e-02 9.62981641e-01
-2.03413263e-01 -8.90966505e-02 -8.01380500e-02 -1.31947148e+00
-9.46981966e-01 -1.06941330e+00 3.77114862e-02 8.06266367e-01
1.04959957e-01 -1.14938211e+00 -3.86590272e-01 -5.20528900e-03
6.71988964e-01 7.71649122e-01 9.57140803e-01 -4.59677160e-01
1.11769326e-01 -3.14100832e-01 -3.59383643e-01 8.15823302e-02
1.52803034e-01 5.23639262e-01 -6.39043272e-01 6.51636422e-02
-1.54666081e-01 -1.97667554e-01 3.20900023e-01 1.00070703e+00
7.32210994e-01 -4.99496162e-01 -2.74980575e-01 4.17767167e-01
1.79391587e+00 1.27439034e+00 5.68483472e-01 7.45125711e-01
-1.29386127e-01 4.86390442e-01 1.00351095e+00 6.24970376e-01
2.60999560e-01 -6.97623417e-02 5.77780545e-01 -2.30576590e-01
6.51888430e-01 -1.91904008e-01 2.08489504e-02 6.79491758e-01
-5.59307933e-01 -2.79580742e-01 -8.44025552e-01 3.93755615e-01
-1.97469866e+00 -9.08986926e-01 -2.32556462e-01 1.91755915e+00
3.45284700e-01 -1.07126668e-01 1.88157428e-02 4.96622086e-01
5.26226819e-01 -2.85824150e-01 -6.34647489e-01 -9.42599118e-01
-4.52500254e-01 4.10230279e-01 1.15873110e+00 -1.77697428e-02
-9.00432229e-01 8.60592604e-01 6.11646175e+00 6.78858817e-01
-1.06481862e+00 -3.37702096e-01 9.04120743e-01 5.53061008e-01
7.38861933e-02 -2.50281151e-02 -5.75274348e-01 2.69866973e-01
8.88190448e-01 -5.81254601e-01 6.59538090e-01 5.30709922e-01
5.24142504e-01 -5.40630996e-01 -2.28820652e-01 3.51409137e-01
-3.15707624e-01 -9.85455275e-01 -1.21592499e-01 -1.03620715e-01
1.00025988e+00 -5.05367517e-01 -8.11922327e-02 1.01012483e-01
5.19500077e-01 -1.10523808e+00 4.68870014e-01 7.20823169e-01
2.78422505e-01 -9.98244107e-01 1.30039358e+00 3.67979467e-01
-1.40584183e+00 -4.72556084e-01 -7.77325213e-01 -2.87232786e-01
-1.23342909e-01 4.79545027e-01 -2.68004894e-01 1.11636627e+00
6.72814608e-01 1.72526181e-01 -3.39140147e-01 9.17919397e-01
8.11414123e-02 7.24268377e-01 -4.84833181e-01 -4.87802297e-01
5.46232164e-01 -7.32279897e-01 -4.23297323e-02 7.62805223e-01
9.98052001e-01 5.34815967e-01 -2.12149352e-01 8.24963987e-01
4.56407487e-01 7.65631437e-01 -3.09291214e-01 -4.91709709e-01
6.98737562e-01 1.06140983e+00 -1.03741682e+00 -9.17060673e-02
-2.47360125e-01 3.32241505e-01 -4.06211734e-01 2.94508904e-01
-7.31556535e-01 -4.60306346e-01 -5.91102690e-02 -6.17791831e-01
1.86607569e-01 3.15721184e-02 -8.38398278e-01 -6.21317446e-01
-3.51077378e-01 -5.17677903e-01 3.46539587e-01 -7.38757014e-01
-7.62070417e-01 2.71587640e-01 1.13994673e-01 -9.71338034e-01
-2.20968381e-01 -7.88477242e-01 -7.72522390e-01 1.11560309e+00
-1.73621976e+00 -1.22186506e+00 -2.61097342e-01 -1.43122494e-01
-4.94938232e-02 -6.48816228e-01 1.19202948e+00 -3.04689437e-01
-7.42417395e-01 3.84111404e-01 7.50344574e-01 -2.07254291e-01
-2.65204966e-01 -8.13433349e-01 3.99459004e-02 5.98054588e-01
-5.34347236e-01 -1.92126542e-01 6.95414066e-01 -7.49499917e-01
-1.27010202e+00 -7.59157419e-01 1.15565872e+00 4.97373223e-01
5.84456503e-01 8.68056715e-01 -5.78523338e-01 2.12495968e-01
2.76008397e-01 -4.10910308e-01 5.73852658e-01 -3.16596776e-01
4.49287862e-01 -1.71907306e-01 -1.61495507e+00 4.56346452e-01
1.62105516e-01 4.73074496e-01 5.15414812e-02 -3.00803850e-03
5.31901956e-01 -4.66347598e-02 -1.51112008e+00 7.97504902e-01
1.12195086e+00 -4.70820844e-01 9.61568534e-01 -4.82345134e-01
4.30428773e-01 6.85428008e-02 -5.30610904e-02 -1.13509429e+00
-5.31390488e-01 -8.67417082e-02 -2.15662837e-01 9.17100132e-01
7.44772911e-01 -8.58777225e-01 6.02773547e-01 6.94114864e-01
5.23161948e-01 -9.41104472e-01 -5.21529555e-01 -7.12337375e-01
1.80560321e-01 2.83697844e-01 1.16386092e+00 9.57641006e-01
-2.12755520e-02 -1.69328526e-01 -2.68847257e-01 1.04576573e-01
3.33689332e-01 1.26262888e-01 2.30034381e-01 -1.22230840e+00
3.58595669e-01 -7.55505502e-01 -1.33931667e-01 1.37185827e-02
-1.03880212e-01 -4.39435095e-01 -5.35788417e-01 -1.37784708e+00
-4.42453213e-02 -2.85812259e-01 -4.82069492e-01 6.12656713e-01
-5.80201507e-01 -8.16774294e-02 3.64249289e-01 -3.11096143e-02
6.13359690e-01 5.34219980e-01 1.47190642e+00 4.76006120e-02
-7.98330605e-01 3.20610046e-01 -7.73103237e-01 5.32557011e-01
1.51368070e+00 -4.69035000e-01 -2.24525347e-01 -2.07771614e-01
2.67432511e-01 4.28127199e-01 -5.70619758e-03 -8.94150496e-01
-3.07121545e-01 -6.51739299e-01 6.02910042e-01 -9.04209197e-01
-3.13899636e-01 -1.05596817e+00 5.39938867e-01 1.21029353e+00
-9.39597562e-02 5.74427247e-01 3.56397629e-01 -1.09074858e-03
9.75793675e-02 -7.90664673e-01 3.12206745e-01 -1.15392786e-02
-5.66237152e-01 -9.87653737e-04 -3.14965725e-01 -7.50528455e-01
1.42153108e+00 -4.61098284e-01 -2.49164611e-01 -4.28749062e-02
-6.90542698e-01 3.58929217e-01 9.62678343e-03 1.16963021e-01
4.50899661e-01 -1.41430688e+00 -1.06585371e+00 6.75777048e-02
-1.05687670e-01 -4.75098729e-01 -1.55272275e-01 8.28395128e-01
-1.29802346e+00 7.14728892e-01 -6.49824739e-01 1.02844864e-01
-9.24997449e-01 5.35618186e-01 1.38995245e-01 -6.10839248e-01
1.44374892e-01 2.89558381e-01 -6.14770353e-01 -1.55064136e-01
-1.49715275e-01 -3.57489765e-01 -7.86647618e-01 3.20802294e-02
1.76866427e-01 1.08842766e+00 -2.56347954e-01 -5.35006166e-01
-5.31869866e-02 5.13275027e-01 5.03705204e-01 1.64127350e-01
1.41693830e+00 -3.59577872e-02 -2.47198194e-01 2.65021503e-01
7.11006343e-01 -6.93464041e-01 -6.64340615e-01 1.00918189e-01
5.27288318e-01 -2.57796735e-01 3.20542186e-01 -1.36840260e+00
-1.13003087e+00 4.51390207e-01 1.14987040e+00 3.62818927e-01
1.65845954e+00 -9.21026111e-01 6.63713515e-01 6.05408609e-01
1.11055553e-01 -1.28309071e+00 -6.17405474e-01 3.27523053e-01
8.95864785e-01 -1.56041503e+00 1.67527720e-01 5.59273027e-02
-4.99956995e-01 1.59135544e+00 3.97030562e-01 -1.41414389e-01
1.01598406e+00 2.52326339e-01 1.20352879e-01 4.55189019e-01
-4.57483172e-01 -3.26210618e-01 1.06374137e-01 5.44253230e-01
4.92042392e-01 5.79343855e-01 -1.08266008e+00 7.19777286e-01
1.36608388e-02 4.93784457e-01 -3.01652937e-03 9.96866703e-01
-7.66571224e-01 -8.42796445e-01 -7.28139758e-01 7.19358385e-01
-4.93138045e-01 -2.44887337e-01 -8.60985145e-02 8.43824267e-01
4.39457297e-02 1.26438570e+00 -2.76733607e-01 -2.20120758e-01
8.57884064e-02 -1.59003258e-01 2.99793482e-01 3.88042450e-01
-6.56930447e-01 1.62456736e-01 -1.64126113e-01 6.04312075e-03
-7.80902326e-01 -4.69017178e-01 -1.08355820e+00 -7.85935342e-01
-1.10521340e+00 4.84241098e-01 1.02846730e+00 8.84302258e-01
5.55448085e-02 1.82995349e-01 1.16960561e+00 -5.24469554e-01
-6.46888912e-01 -1.19611061e+00 -1.00828242e+00 -3.76129866e-01
-3.20569336e-01 -4.89182025e-01 -1.08276635e-01 -6.72815666e-02] | [9.3503999710083, -1.5936752557754517] |
4d446f20-d6c5-4e03-8579-39fb39b5fec1 | two-to-five-truths-in-non-negative-matrix | 2305.05389 | null | https://arxiv.org/abs/2305.05389v1 | https://arxiv.org/pdf/2305.05389v1.pdf | Two to Five Truths in Non-Negative Matrix Factorization | In this paper, we explore the role of matrix scaling on a matrix of counts when building a topic model using non-negative matrix factorization. We present a scaling inspired by the normalized Laplacian (NL) for graphs that can greatly improve the quality of a non-negative matrix factorization. The results parallel those in the spectral graph clustering work of \cite{Priebe:2019}, where the authors proved adjacency spectral embedding (ASE) spectral clustering was more likely to discover core-periphery partitions and Laplacian Spectral Embedding (LSE) was more likely to discover affinity partitions. In text analysis non-negative matrix factorization (NMF) is typically used on a matrix of co-occurrence ``contexts'' and ``terms" counts. The matrix scaling inspired by LSE gives significant improvement for text topic models in a variety of datasets. We illustrate the dramatic difference a matrix scalings in NMF can greatly improve the quality of a topic model on three datasets where human annotation is available. Using the adjusted Rand index (ARI), a measure cluster similarity we see an increase of 50\% for Twitter data and over 200\% for a newsgroup dataset versus using counts, which is the analogue of ASE. For clean data, such as those from the Document Understanding Conference, NL gives over 40\% improvement over ASE. We conclude with some analysis of this phenomenon and some connections of this scaling with other matrix scaling methods. | ['Nicholas A. Lines', 'Ryan Kaliszewski', 'Rod Gomez', 'Brian Baughman', 'Neil P Molino', 'John M. Conroy'] | 2023-05-06 | null | null | null | null | ['graph-clustering', 'spectral-graph-clustering', 'topic-models'] | ['graphs', 'graphs', 'natural-language-processing'] | [-1.32861768e-03 3.45951080e-01 -9.83754098e-02 1.94137450e-02
-4.41173464e-01 -8.38168442e-01 9.21753168e-01 4.61285919e-01
-4.77314711e-01 4.80225921e-01 6.09793603e-01 -5.10459304e-01
-5.75757325e-01 -9.03805017e-01 -3.73870164e-01 -7.80972660e-01
-5.61904728e-01 5.99010468e-01 1.56172872e-01 -2.55436242e-01
2.61673659e-01 -4.59772497e-02 -1.34066153e+00 4.27422374e-01
6.21675670e-01 3.41726393e-01 1.12594463e-01 6.06278002e-01
-4.68904376e-01 2.54876941e-01 -4.59153056e-01 -2.40279570e-01
3.70546550e-01 -3.17152470e-01 -8.83632541e-01 1.66943762e-02
4.53446805e-01 6.25325978e-01 -2.44177267e-01 1.10394156e+00
2.06207827e-01 2.04621479e-01 1.13214254e+00 -1.62304747e+00
-4.38612878e-01 1.11086285e+00 -1.19349396e+00 4.73449588e-01
3.60932767e-01 -3.10683161e-01 1.37353384e+00 -7.89577305e-01
9.20829773e-01 1.46488440e+00 1.08118200e+00 -8.09183121e-02
-1.94732690e+00 -7.54202306e-01 7.60989338e-02 4.06446941e-02
-1.50895262e+00 1.70739759e-02 5.24439096e-01 -6.51553750e-01
8.95955443e-01 5.21092772e-01 6.16923630e-01 6.82320118e-01
1.79375306e-01 3.64694417e-01 1.21890450e+00 -5.95473349e-01
1.87157094e-01 1.83700085e-01 4.58111405e-01 3.57241184e-01
5.10789514e-01 -4.45443273e-01 -5.74769080e-01 -7.60573983e-01
4.99841124e-01 -1.34357288e-01 -5.03524877e-02 -4.17748064e-01
-1.56784308e+00 1.17785871e+00 3.56167823e-01 8.34214449e-01
-2.48948932e-01 1.05806783e-01 5.60351491e-01 4.94124621e-01
8.00688744e-01 6.95113242e-01 -2.04346344e-01 -7.68041611e-02
-1.05070698e+00 7.75609985e-02 9.97518241e-01 5.28934419e-01
9.94897008e-01 -3.32900316e-01 1.66288838e-01 8.19199622e-01
3.03664595e-01 3.61779243e-01 4.88850206e-01 -8.96751106e-01
4.06271279e-01 7.29044974e-01 -2.43037373e-01 -1.47534001e+00
-8.12239289e-01 -3.59257281e-01 -1.02703571e+00 -7.75858089e-02
3.73115242e-01 -6.45839497e-02 -6.32323921e-01 1.67826366e+00
2.20576987e-01 8.03138390e-02 -1.73490375e-01 3.91869873e-01
4.24872786e-01 6.77484632e-01 8.34890977e-02 -5.35099030e-01
1.48967361e+00 -4.31685537e-01 -9.06360447e-01 1.46930635e-01
8.96721959e-01 -1.02932882e+00 1.13493586e+00 4.72506404e-01
-5.34704328e-01 -2.79336005e-01 -8.66867244e-01 2.27855250e-01
-6.58102572e-01 -2.54083902e-01 9.68948841e-01 8.64226878e-01
-1.54628360e+00 5.16774237e-01 -7.44643986e-01 -9.09913719e-01
-1.78310182e-02 4.08098191e-01 -5.40455759e-01 1.95180364e-02
-1.19752812e+00 4.74728227e-01 3.75746995e-01 -6.49019957e-01
5.18627055e-02 -9.36044157e-01 -4.79793042e-01 1.66809261e-02
2.11535513e-01 -6.71594143e-01 3.63481313e-01 -6.52491331e-01
-8.09245288e-01 7.43856907e-01 -1.13798767e-01 -5.65660655e-01
1.84302032e-01 5.09690531e-02 -4.79868233e-01 1.80365518e-01
6.21059239e-01 5.89704931e-01 5.97765446e-01 -1.36970448e+00
-2.99051017e-01 -4.60108101e-01 -3.18452626e-01 2.66509410e-02
-7.25195765e-01 -1.03475548e-01 -2.24674404e-01 -8.57528210e-01
4.81563807e-01 -1.30163622e+00 -3.01063150e-01 -6.21177435e-01
-5.28351963e-01 -5.74363232e-01 9.01337385e-01 -4.79314893e-01
1.68949389e+00 -2.20583725e+00 2.74116397e-01 9.30921912e-01
7.44444966e-01 -3.56017441e-01 -7.27069452e-02 1.08755910e+00
-5.66440165e-01 6.50029004e-01 -7.11917654e-02 -2.47340903e-01
1.49804980e-01 1.47851378e-01 -4.03290272e-01 5.77317536e-01
-2.48227283e-01 5.33281684e-01 -7.25891352e-01 -4.85356629e-01
-2.70431209e-02 4.79410350e-01 -5.92974842e-01 -5.00084817e-01
1.22336499e-01 -2.04538703e-01 1.76649734e-01 -4.28445563e-02
4.34537798e-01 -6.18746936e-01 5.14708638e-01 -3.13389808e-01
-7.14491308e-02 3.89356762e-02 -1.49713826e+00 1.56823683e+00
-2.74923779e-02 9.69552040e-01 9.17641297e-02 -6.77182734e-01
7.03538716e-01 3.21990609e-01 7.37127841e-01 -1.28590375e-01
-4.74162772e-02 -5.95509820e-02 4.11694407e-01 5.43949492e-02
6.98070109e-01 -1.95057556e-01 9.80714113e-02 1.08372164e+00
-1.78832207e-02 -1.50528431e-01 5.99278927e-01 1.22984958e+00
1.40180206e+00 -7.54846334e-01 2.08395235e-02 -1.16423666e+00
1.93609849e-01 5.15690306e-03 8.13716501e-02 5.06947458e-01
2.24282846e-01 4.72514391e-01 7.81518340e-01 -1.64301500e-01
-1.14456105e+00 -1.06922245e+00 -3.18918169e-01 1.24937546e+00
-2.58911073e-01 -1.43898785e+00 -6.44694626e-01 -4.27181989e-01
1.83866739e-01 6.33056641e-01 -1.08789396e+00 1.35409877e-01
-6.23871051e-02 -1.32426202e+00 3.52254450e-01 1.28636584e-01
-3.93054867e-03 -4.10854936e-01 1.71058014e-01 -1.79001033e-01
-3.78358901e-01 -8.04952741e-01 -4.11021352e-01 3.61598521e-01
-9.97814357e-01 -1.05428815e+00 -3.89615536e-01 -4.55150515e-01
6.86912239e-01 4.97181296e-01 1.17549992e+00 1.01660550e-01
-2.61685461e-01 6.10878050e-01 -4.84200239e-01 -3.10359955e-01
-4.40136909e-01 4.67481792e-01 5.61180294e-01 -1.44315809e-01
6.08975530e-01 -9.84892130e-01 -4.72784907e-01 3.60875160e-01
-1.29239154e+00 -1.58654302e-01 3.68421406e-01 6.01225197e-01
1.41962245e-01 4.44678217e-01 4.55505848e-01 -1.17695737e+00
1.13079071e+00 -7.29015648e-01 9.27972347e-02 -1.32846549e-01
-1.10062122e+00 1.90529376e-01 2.46228173e-01 -3.85127872e-01
-4.15174335e-01 -2.96657771e-01 4.63302135e-01 -2.38527924e-01
1.90269813e-01 6.08902693e-01 4.33835208e-01 1.79372430e-01
1.23145044e+00 -1.77608579e-01 4.21931595e-02 -3.60757977e-01
6.48919642e-01 6.00239277e-01 8.71238410e-02 -5.14855087e-01
9.16040123e-01 7.49049187e-01 1.37201071e-01 -1.07000220e+00
-3.49736869e-01 -9.95536625e-01 -8.95227134e-01 8.61323625e-03
8.17076504e-01 -9.05844390e-01 -5.90741873e-01 -3.06345791e-01
-6.15257919e-01 -1.22788899e-01 -4.35588807e-01 5.12280345e-01
-3.93161297e-01 6.06935084e-01 -4.82144475e-01 -7.28280187e-01
-9.01587903e-02 -5.47092855e-01 8.45984519e-01 -3.84107023e-01
-6.88912153e-01 -1.46889496e+00 5.34045160e-01 1.95674866e-01
2.71199077e-01 3.66138756e-01 1.22504461e+00 -7.45080054e-01
1.64111435e-01 -1.02753406e-02 -3.30029875e-01 -9.85825732e-02
1.50025025e-01 3.99667807e-02 -7.82739937e-01 -5.71054041e-01
-2.78978616e-01 2.88513064e-01 1.02173412e+00 3.69454026e-01
3.72105330e-01 -2.99192101e-01 -8.51062894e-01 -1.85431575e-03
1.34900367e+00 -1.45477042e-01 6.48148954e-01 3.29274327e-01
8.14679444e-01 6.76110089e-01 1.29311070e-01 3.86395156e-01
3.53542656e-01 4.61835861e-01 -1.01886153e-01 -1.94602802e-01
-6.71086982e-02 -8.85031670e-02 3.75248641e-01 1.18930149e+00
-2.55253673e-01 -9.19877738e-02 -1.35403442e+00 5.87903321e-01
-1.90419173e+00 -9.64268684e-01 -7.40060091e-01 1.96764839e+00
7.84284770e-01 1.28195271e-01 5.79207242e-01 5.02136528e-01
7.28872359e-01 3.53542864e-02 1.12070590e-01 -1.70133457e-01
-1.40368238e-01 -6.19824007e-02 5.62100410e-01 7.49456406e-01
-9.64943230e-01 6.95622504e-01 6.64004183e+00 1.04915690e+00
-7.10975528e-01 2.28396803e-01 3.48291159e-01 -1.09396979e-01
-5.00751317e-01 1.70453772e-01 -3.18640321e-01 3.66845101e-01
1.30737853e+00 -4.05695707e-01 2.93059200e-01 5.56293130e-01
1.34472862e-01 -3.28585118e-01 -9.48043764e-01 1.16821384e+00
1.85802400e-01 -1.17792594e+00 1.12806104e-01 6.70074046e-01
9.66298044e-01 2.11636871e-01 6.11048900e-02 7.83472434e-02
7.44334936e-01 -8.92645299e-01 1.70645699e-01 2.76454806e-01
6.10574722e-01 -8.20618391e-01 6.03060722e-01 1.86585173e-01
-1.30809987e+00 -3.22494586e-03 -4.47903007e-01 9.84124374e-03
9.47413221e-02 1.10891628e+00 -1.22339213e+00 6.28844559e-01
8.82651806e-01 7.75377393e-01 -9.05066490e-01 6.34851456e-01
3.08741331e-01 9.49546754e-01 -5.97296655e-01 2.12875530e-01
2.51565546e-01 -6.40612721e-01 8.26508999e-01 1.45279408e+00
4.24025692e-02 1.01533897e-01 1.90176830e-01 6.53978586e-01
1.04901321e-01 5.35907030e-01 -6.89697981e-01 -2.92196423e-01
9.51021835e-02 1.48492432e+00 -1.58789444e+00 -3.96766961e-01
-2.57908553e-01 6.75952554e-01 7.09808543e-02 2.91507900e-01
-3.67349416e-01 -3.51424813e-01 4.16759014e-01 6.81241453e-01
5.53772859e-02 -5.18201172e-01 -3.09029013e-01 -1.02227223e+00
-5.25368214e-01 -7.70010948e-01 4.86866713e-01 -4.51443195e-01
-1.63280320e+00 5.53252459e-01 1.88110933e-01 -8.94416332e-01
-9.59090814e-02 -3.68466020e-01 -4.20898169e-01 7.06687748e-01
-5.46930313e-01 -8.65036726e-01 1.10889420e-01 5.59061348e-01
1.17354974e-01 -9.73416194e-02 7.35595584e-01 3.47967058e-01
-2.10344538e-01 2.17994437e-01 4.32500780e-01 -4.32395414e-02
9.80702102e-01 -1.82498026e+00 4.44789320e-01 5.83216012e-01
6.50681138e-01 1.14620996e+00 9.89563227e-01 -8.96749020e-01
-1.08488774e+00 -7.78121591e-01 7.10317671e-01 -9.09095287e-01
1.12875700e+00 -5.42848766e-01 -7.95736551e-01 7.63317883e-01
5.47987938e-01 -5.61660945e-01 1.29535007e+00 8.06097925e-01
-5.45637071e-01 3.25580090e-01 -8.17658901e-01 5.31197250e-01
1.13915086e+00 -5.83781064e-01 -5.01986086e-01 6.52676702e-01
7.59227157e-01 4.63080913e-01 -1.19231260e+00 1.65601358e-01
3.99510294e-01 -9.44179416e-01 8.89431000e-01 -5.26277244e-01
-5.89043759e-02 -3.36857527e-01 -2.31520891e-01 -1.46085238e+00
-8.09762657e-01 -7.57385075e-01 2.91162878e-01 1.17095351e+00
6.04544163e-01 -6.51796281e-01 7.87355840e-01 4.23626393e-01
2.42013112e-01 -2.88652897e-01 -8.14170539e-01 -7.60321558e-01
2.22454503e-01 -4.95371521e-01 7.00526685e-02 1.58550429e+00
5.21056235e-01 8.15087140e-01 9.38032344e-02 -2.42980540e-01
5.87162435e-01 -1.86800420e-01 9.23368156e-01 -1.88509345e+00
-3.38399149e-02 -4.75903422e-01 -6.42577410e-01 -4.25979197e-01
8.68017897e-02 -1.24663484e+00 -5.85580289e-01 -1.54395688e+00
4.69816417e-01 -4.84113216e-01 -5.27891144e-03 3.65052789e-01
-3.14172693e-02 3.72942448e-01 3.75524700e-01 5.06011546e-01
-6.26479745e-01 2.21088231e-02 7.46605873e-01 -9.46044028e-02
-4.13792342e-01 -4.26850617e-01 -7.86042213e-01 6.57801509e-01
5.94468534e-01 -6.12481594e-01 -5.47186732e-01 1.38662085e-01
1.01666176e+00 -3.90471458e-01 -8.87022987e-02 -9.59869683e-01
3.44165415e-01 2.91209012e-01 3.42839539e-01 -5.07387280e-01
1.27896979e-01 -8.37161779e-01 5.08093834e-01 4.66493905e-01
-2.82047719e-01 5.56457102e-01 4.26244348e-01 1.00899291e+00
-5.21437777e-03 1.89846814e-01 1.96805462e-01 -9.06680450e-02
-3.14077169e-01 -1.25591755e-01 -5.39346397e-01 1.54664010e-01
8.33045304e-01 -6.61221743e-02 -3.43013346e-01 -5.91596186e-01
-1.05650938e+00 4.99953553e-02 5.41573584e-01 4.53544438e-01
9.10898969e-02 -1.36543667e+00 -6.78601921e-01 -8.03268254e-02
-7.35819191e-02 -5.29198468e-01 1.23125494e-01 1.24341702e+00
-2.13144660e-01 5.04210293e-01 1.14820488e-01 -7.42241561e-01
-1.55339086e+00 7.21153557e-01 -2.86737740e-01 -3.19293380e-01
-3.79617274e-01 5.98753810e-01 2.71189660e-01 -3.12744409e-01
-1.08205356e-01 -2.49078408e-01 -1.39191881e-01 8.16376984e-01
4.07036334e-01 7.07838118e-01 -6.18516468e-03 -7.42039204e-01
-3.42902005e-01 6.20702744e-01 -1.32371485e-01 -5.78183293e-01
1.41147542e+00 -3.37478638e-01 -8.27141881e-01 9.80457962e-01
1.15569925e+00 3.85465413e-01 -4.69234049e-01 -1.49011403e-01
2.09238514e-01 -2.37347394e-01 1.82643775e-02 -4.32483077e-01
-6.58555627e-01 6.36836410e-01 4.69747245e-01 1.18082464e+00
5.79798102e-01 2.23170832e-01 2.28729516e-01 4.63553876e-01
2.35733211e-01 -1.03161263e+00 7.22649917e-02 2.59917885e-01
8.19166481e-01 -1.00014257e+00 4.21493053e-01 -6.99603856e-01
-5.80489755e-01 1.00312269e+00 1.11043200e-01 -1.49959683e-01
1.24311411e+00 3.91639955e-02 -2.77314112e-02 -7.46598661e-01
-6.07945442e-01 -7.28678852e-02 3.41365784e-01 4.94560689e-01
5.32787502e-01 3.03163588e-01 -4.77429777e-01 3.38611752e-01
-6.95311785e-01 -6.80371165e-01 5.75838327e-01 4.84320104e-01
-3.83025020e-01 -1.18058681e+00 -6.36136234e-01 8.91907394e-01
-3.83960962e-01 -3.14684004e-01 -8.26699674e-01 8.90712440e-01
-2.34694546e-03 1.12107074e+00 2.61837721e-01 -6.51453674e-01
-1.81921586e-01 4.01726991e-01 1.66015364e-02 -8.89281154e-01
-6.72147214e-01 2.88355112e-01 -1.12147771e-01 -4.21351194e-01
-6.98055208e-01 -9.12337244e-01 -1.19767201e+00 -6.72082305e-01
-5.52689731e-01 7.15061128e-01 5.22962213e-01 4.66931939e-01
5.85755110e-01 4.28664148e-01 3.01995754e-01 -6.58440709e-01
1.17444366e-01 -1.29461932e+00 -1.02476132e+00 5.86168289e-01
-2.06863001e-01 -7.02432096e-01 -9.01147664e-01 2.95275778e-01] | [7.266862392425537, 5.1336870193481445] |
a61847b3-70cd-4693-a30a-76a63e3de9b5 | realtime-multi-person-2d-pose-estimation | 1611.08050 | null | http://arxiv.org/abs/1611.08050v2 | http://arxiv.org/pdf/1611.08050v2.pdf | Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields | We present an approach to efficiently detect the 2D pose of multiple people
in an image. The approach uses a nonparametric representation, which we refer
to as Part Affinity Fields (PAFs), to learn to associate body parts with
individuals in the image. The architecture encodes global context, allowing a
greedy bottom-up parsing step that maintains high accuracy while achieving
realtime performance, irrespective of the number of people in the image. The
architecture is designed to jointly learn part locations and their association
via two branches of the same sequential prediction process. Our method placed
first in the inaugural COCO 2016 keypoints challenge, and significantly exceeds
the previous state-of-the-art result on the MPII Multi-Person benchmark, both
in performance and efficiency. | ['Shih-En Wei', 'Zhe Cao', 'Tomas Simon', 'Yaser Sheikh'] | 2016-11-24 | realtime-multi-person-2d-pose-estimation-1 | http://openaccess.thecvf.com/content_cvpr_2017/html/Cao_Realtime_Multi-Person_2D_CVPR_2017_paper.html | http://openaccess.thecvf.com/content_cvpr_2017/papers/Cao_Realtime_Multi-Person_2D_CVPR_2017_paper.pdf | cvpr-2017-7 | ['2d-human-pose-estimation'] | ['computer-vision'] | [ 1.37120178e-02 2.05726415e-01 -1.43395752e-01 -4.15035456e-01
-5.65017939e-01 -5.82655847e-01 6.42533004e-01 1.62538990e-01
-6.18038535e-01 2.65076399e-01 4.86355960e-01 3.94081205e-01
6.55243844e-02 -4.94038910e-01 -9.22173023e-01 -2.19856009e-01
-4.49645787e-01 1.04529452e+00 3.42041403e-01 4.93300669e-02
-1.40538827e-01 2.93256313e-01 -1.50010324e+00 2.62796342e-01
1.17680229e-01 9.21743453e-01 -3.56529266e-01 8.49634826e-01
2.62071192e-01 3.91986936e-01 -4.46550578e-01 -9.59767938e-01
5.46143591e-01 -2.42704954e-02 -9.32844162e-01 1.67516127e-01
1.32986379e+00 -4.36989069e-01 -4.82514024e-01 7.04212308e-01
4.59110439e-01 -4.45119701e-02 6.01238847e-01 -1.03372633e+00
-2.14187264e-01 4.46290672e-01 -8.66913021e-01 4.37248081e-01
6.58451736e-01 -9.99571979e-02 1.02953136e+00 -7.36724555e-01
6.87298000e-01 1.58274543e+00 1.16247940e+00 4.97385204e-01
-1.33047771e+00 -3.16174835e-01 3.24863106e-01 -5.04182465e-02
-1.40596080e+00 -3.59261304e-01 2.07111150e-01 -5.78912556e-01
9.53589857e-01 1.06074646e-01 9.73735809e-01 8.46847355e-01
1.28049314e-01 1.09666395e+00 7.80088842e-01 -1.89786717e-01
-1.02122121e-01 -4.70048010e-01 1.83456048e-01 1.02625930e+00
3.26726139e-01 1.30614517e-02 -9.34366584e-01 -5.55539846e-01
8.38692963e-01 -3.35665490e-03 2.56644756e-01 -6.49640918e-01
-1.11167800e+00 5.87245822e-01 5.56769669e-01 -2.14329392e-01
-3.86017054e-01 5.07894337e-01 3.74035120e-01 -1.03200659e-01
4.39473838e-01 -5.77831455e-02 -5.77876329e-01 -1.22317411e-01
-9.49285328e-01 7.89273441e-01 7.48998463e-01 7.78651059e-01
5.30556619e-01 -7.60542274e-01 -3.49101335e-01 6.23418152e-01
3.87763321e-01 5.43793559e-01 -8.27363320e-03 -1.03386045e+00
5.10392725e-01 6.60388112e-01 4.99762110e-02 -8.67342889e-01
-5.35351276e-01 -3.07286412e-01 -4.22529250e-01 1.52770847e-01
5.60222268e-01 -1.66181311e-01 -1.27325058e+00 1.68830073e+00
8.83063316e-01 8.16528127e-02 -4.65233266e-01 8.87395620e-01
7.61550605e-01 2.06879899e-01 3.30836326e-01 5.87412655e-01
1.84506667e+00 -1.18260956e+00 -4.65628877e-02 -6.03670478e-01
4.95853946e-02 -5.04941702e-01 4.46831703e-01 3.11381370e-01
-1.23993349e+00 -7.04111934e-01 -6.90272689e-01 -3.00681829e-01
-2.45189339e-01 1.59634039e-01 9.20012414e-01 7.90249646e-01
-9.02274251e-01 6.63510382e-01 -1.02587330e+00 -4.94826972e-01
5.83205760e-01 6.18515670e-01 -6.73096716e-01 9.99832153e-02
-6.88679159e-01 6.19132936e-01 1.42517626e-01 2.77038030e-02
-6.90788507e-01 -7.72053421e-01 -9.71944630e-01 -9.85071957e-02
4.34254318e-01 -1.07355547e+00 1.29957747e+00 -6.59092188e-01
-1.22522342e+00 1.32521367e+00 -1.35002345e-01 -5.78313529e-01
8.09385955e-01 -8.83154035e-01 -8.91530737e-02 5.10425508e-01
3.43795210e-01 1.24775040e+00 9.55423117e-01 -9.59658980e-01
-9.68913615e-01 -6.37349725e-01 -3.21155638e-02 3.38517994e-01
1.67818576e-01 2.69780278e-01 -1.24380147e+00 -7.42228508e-01
1.17098741e-01 -1.16115868e+00 -2.48854071e-01 3.32023114e-01
-4.61266220e-01 -4.79730725e-01 3.56077343e-01 -1.00193310e+00
7.09117889e-01 -1.89783180e+00 3.71021330e-01 3.69509965e-01
3.02312016e-01 -2.14457840e-01 1.09684385e-01 2.43665278e-01
1.01364583e-01 -2.30162203e-01 -9.57535058e-02 -8.18587661e-01
5.84580116e-02 8.38799030e-02 7.53075406e-02 7.70917356e-01
9.18377116e-02 1.16037285e+00 -6.42918944e-01 -7.10314333e-01
2.56585032e-02 4.15065497e-01 -5.77338338e-01 1.84481040e-01
-4.19695154e-02 3.25397879e-01 -2.52416462e-01 8.32702816e-01
5.20742953e-01 -2.88257569e-01 2.79001594e-01 -1.98553964e-01
2.05519676e-01 9.93726850e-02 -1.31674778e+00 1.88271666e+00
2.84937680e-01 8.26876909e-02 3.49753886e-01 -4.79983598e-01
2.93198645e-01 1.33064240e-01 6.27357006e-01 -2.91904032e-01
-1.09310254e-01 -1.40602365e-01 -2.66567856e-01 -6.41665459e-02
3.42598617e-01 1.78314492e-01 -3.54157865e-01 5.45676202e-02
2.10581779e-01 3.83876354e-01 3.49318206e-01 3.19919050e-01
1.05496228e+00 5.13216257e-01 2.67891437e-01 -3.91374171e-01
2.92899579e-01 -1.69431895e-01 4.67839837e-01 9.45542693e-01
-1.93841890e-01 7.57067323e-01 4.41792369e-01 -8.17668378e-01
-9.19079125e-01 -1.40036106e+00 8.22818279e-02 1.40356266e+00
2.05398470e-01 -8.55016649e-01 -6.29809976e-01 -9.28491473e-01
4.66218412e-01 -3.10686469e-01 -1.03798687e+00 3.06280881e-01
-7.71651268e-01 -4.32931304e-01 6.00204408e-01 9.02180970e-01
3.58211696e-01 -9.47632432e-01 -8.73589516e-01 3.33588086e-02
-1.35641143e-01 -1.14543474e+00 -7.47503638e-01 -5.58879152e-02
-6.66401505e-01 -1.21260250e+00 -9.30719554e-01 -6.64465487e-01
5.89526653e-01 -2.01331720e-01 1.58613873e+00 1.40007675e-01
-6.70307994e-01 7.53331184e-01 -7.38487095e-02 -2.16386259e-01
2.34351248e-01 1.64867073e-01 1.29642738e-02 -1.47745878e-01
2.32686684e-01 -3.25213641e-01 -1.01861465e+00 1.76193357e-01
-3.26915175e-01 -3.79517674e-02 7.77274489e-01 5.56272030e-01
7.18058884e-01 -3.91000718e-01 -1.46771267e-01 -7.90721178e-01
-9.67985764e-02 -2.81731606e-01 -3.70873749e-01 2.78326839e-01
-1.83024764e-01 3.19810770e-02 -1.38270995e-02 -1.95730790e-01
-6.09621048e-01 5.86747706e-01 -1.59219593e-01 -2.85061866e-01
-3.57148945e-01 -2.59373728e-02 -3.15973386e-02 -1.86907321e-01
2.60384023e-01 -3.98338074e-03 -7.82402381e-02 -8.10183525e-01
3.62567961e-01 1.52651995e-01 1.04840326e+00 -8.66654158e-01
6.29824042e-01 6.05801940e-01 2.04372525e-01 -6.05995715e-01
-8.97657573e-01 -8.28020871e-01 -9.20315146e-01 -1.14374503e-01
1.05260122e+00 -1.23242283e+00 -9.18455422e-01 5.98602593e-01
-8.51497233e-01 -2.87970781e-01 -1.81602761e-01 2.04189360e-01
-3.82104963e-01 4.76603419e-01 -9.68142450e-01 -5.65442920e-01
-6.13777339e-01 -7.01868176e-01 1.66104162e+00 2.76941150e-01
-3.25489372e-01 -4.73884135e-01 1.96634158e-01 5.55585623e-01
-7.51384571e-02 2.98617899e-01 3.91018331e-01 -4.49463099e-01
-3.95850897e-01 -3.69815022e-01 -1.69375613e-01 -5.64480722e-02
-3.28606814e-01 -3.44284475e-01 -6.59662068e-01 -4.32023197e-01
-8.05708170e-01 -4.74276096e-01 9.98963416e-01 3.88516277e-01
1.06163418e+00 -1.56017274e-01 -7.21858978e-01 5.89988828e-01
9.74442124e-01 -6.21242166e-01 4.52182680e-01 1.58220857e-01
8.13863516e-01 6.83333158e-01 3.61241937e-01 3.85806382e-01
6.81201875e-01 9.43162858e-01 2.19493896e-01 -1.23582527e-01
-3.53991568e-01 -4.67551023e-01 1.30008191e-01 5.67243658e-02
-4.07395363e-01 1.51603296e-01 -9.60513175e-01 5.63768506e-01
-1.94087434e+00 -8.13260257e-01 -4.57346849e-02 2.13149929e+00
6.24026954e-01 1.30542278e-01 8.74109507e-01 -4.94132847e-01
6.88232005e-01 4.87038456e-02 -3.44258666e-01 1.99967384e-01
1.21433675e-01 3.86660010e-01 7.35491335e-01 3.30008507e-01
-1.64390159e+00 7.75544286e-01 7.88520050e+00 4.90414560e-01
-4.41020131e-01 3.63022052e-02 7.08558559e-01 -3.28463525e-01
3.75122488e-01 -2.73608774e-01 -1.17352426e+00 3.98034811e-01
7.35397220e-01 6.29889369e-01 2.35762924e-01 8.03600848e-01
-4.90364581e-01 -2.43398860e-01 -1.14763963e+00 1.11640096e+00
2.51278460e-01 -1.05102718e+00 -1.98510408e-01 2.62745291e-01
4.63583648e-01 1.86179742e-01 -1.57575667e-01 1.65922761e-01
2.14316815e-01 -9.59333181e-01 1.04780281e+00 5.82924724e-01
4.77251530e-01 -8.85780573e-01 5.67521751e-01 1.51427433e-01
-1.52530563e+00 -4.68687788e-02 -1.07450135e-01 -1.01286974e-02
1.98911056e-01 2.42230907e-01 -5.93060195e-01 3.79282176e-01
1.35031188e+00 4.30464298e-01 -9.39844608e-01 1.13909447e+00
-2.13807061e-01 4.70317513e-01 -7.68093050e-01 3.45602751e-01
-1.07798494e-01 8.38160887e-02 4.74340677e-01 1.25250340e+00
-3.07346061e-02 7.11433217e-02 6.92294002e-01 2.98570842e-01
-2.27038398e-01 -1.52986214e-01 -9.66873858e-03 3.09485257e-01
1.34458914e-01 1.22158194e+00 -9.38970327e-01 -5.24859369e-01
-4.23312366e-01 1.28549755e+00 5.82331955e-01 -1.84731811e-01
-7.07849383e-01 1.28396749e-01 7.43894279e-01 4.02070820e-01
8.50698233e-01 -2.32820958e-01 -6.23889491e-02 -9.69569325e-01
9.19951051e-02 -7.76347578e-01 8.70959580e-01 -3.84176791e-01
-1.54382288e+00 1.97612181e-01 4.37418103e-01 -6.65473878e-01
-1.55160442e-01 -6.57182574e-01 -3.06246877e-01 7.50147104e-01
-8.89584363e-01 -1.63976097e+00 -1.54561788e-01 5.49821079e-01
2.64779627e-01 1.14878550e-01 7.39154458e-01 2.73586422e-01
-4.11163896e-01 6.91220701e-01 -4.57417786e-01 4.97798949e-01
7.43432164e-01 -1.60093319e+00 8.81003797e-01 8.53329480e-01
3.55234772e-01 7.08873391e-01 6.62511587e-01 -1.13884485e+00
-1.39499593e+00 -7.49139249e-01 8.75159562e-01 -9.22848940e-01
2.39887819e-01 -7.16881454e-01 -4.22164857e-01 1.00665379e+00
1.69627592e-01 3.17885339e-01 6.87615991e-01 7.36340582e-01
-4.92437601e-01 8.11160952e-02 -1.15911925e+00 2.97525257e-01
1.35821903e+00 -2.01892748e-01 -8.62855196e-01 2.63903797e-01
2.40114093e-01 -8.62128377e-01 -8.98331821e-01 4.21164542e-01
1.00069892e+00 -8.96538258e-01 1.50307488e+00 -5.27535439e-01
2.46005893e-01 -2.30261490e-01 -7.62619898e-02 -6.57102764e-01
-6.42601848e-01 -5.14979661e-01 -6.58944249e-01 1.06460941e+00
5.32383956e-02 -1.84515923e-01 1.32344174e+00 7.95826912e-01
2.54713356e-01 -8.33970547e-01 -9.56451893e-01 -6.05527997e-01
-2.10250199e-01 -4.95705120e-02 4.38257068e-01 2.00984180e-01
-3.35016727e-01 3.58345866e-01 -5.91951966e-01 3.28706533e-01
1.02597201e+00 1.79735273e-01 1.06691420e+00 -1.28364146e+00
-8.17458272e-01 -2.59255707e-01 -8.35850418e-01 -1.24933875e+00
-9.93485302e-02 -6.06880665e-01 7.95914829e-02 -1.35989702e+00
7.19260216e-01 -1.60165772e-01 8.63543246e-03 8.63989294e-01
-4.65158790e-01 7.10968673e-01 4.10518616e-01 2.77399004e-01
-9.82482195e-01 2.00240612e-01 5.55554390e-01 -1.64278552e-01
-3.43251694e-03 1.66382730e-01 -5.14583170e-01 7.67229319e-01
2.74149984e-01 -4.99773860e-01 3.29561718e-02 -4.40497994e-01
8.43655318e-02 -1.35409802e-01 7.37196982e-01 -1.30397069e+00
1.91562578e-01 3.35894853e-01 1.10856164e+00 -9.91817653e-01
6.08765960e-01 -5.70125043e-01 2.50166774e-01 6.02592409e-01
-1.19765602e-01 2.80865580e-01 1.31804302e-01 7.71268070e-01
1.67832360e-01 1.50883391e-01 6.93874955e-01 -2.71144897e-01
-8.87646198e-01 5.69194078e-01 8.32386613e-02 7.99472034e-02
9.04517353e-01 -6.65933713e-02 2.02589156e-03 -6.96290508e-02
-9.33729947e-01 4.28015500e-01 5.14950871e-01 4.09179211e-01
2.49996215e-01 -1.13127184e+00 -7.15272248e-01 1.41033098e-01
1.30525336e-01 -1.60017163e-02 3.47257286e-01 7.02037573e-01
-6.09785914e-01 1.46938086e-01 -1.35781705e-01 -8.14983547e-01
-1.49179077e+00 3.65001380e-01 3.49759251e-01 -5.43436646e-01
-1.13694072e+00 1.17690146e+00 1.08746469e-01 -3.92696023e-01
2.93208450e-01 6.99140653e-02 1.10839680e-01 -7.85989538e-02
7.12140441e-01 3.92710894e-01 -1.09741176e-02 -8.51264000e-01
-7.71219850e-01 7.76472867e-01 -5.93552478e-02 -2.64595836e-01
1.18777382e+00 -9.35942307e-02 -6.30581677e-02 7.61881769e-02
8.39714587e-01 2.30288804e-01 -1.67803586e+00 -1.91083774e-01
-7.58862346e-02 -5.33006251e-01 -3.12163353e-01 -1.07810438e+00
-8.61313581e-01 3.83477151e-01 8.21952522e-01 -2.13735119e-01
5.93002081e-01 5.82662106e-01 8.06942284e-01 1.45662650e-01
3.86150628e-01 -1.10802758e+00 -5.15797138e-02 3.71458441e-01
6.47246599e-01 -1.09036303e+00 3.85201573e-01 -5.88434398e-01
-6.17980957e-01 9.52476442e-01 6.27659023e-01 -4.37258750e-01
5.25934041e-01 3.17254514e-01 -1.85325444e-01 -3.89138192e-01
-3.41506898e-01 -2.88233727e-01 8.73790503e-01 6.59531951e-01
3.04306328e-01 1.95249870e-01 -2.12965067e-02 7.05549240e-01
-2.30365723e-01 -2.19571546e-01 -3.88729095e-01 1.07132721e+00
-1.95431992e-01 -1.15729165e+00 -5.34797430e-01 5.68543315e-01
-7.37561822e-01 1.34180367e-01 -4.94477034e-01 7.97614813e-01
4.76539671e-01 6.22773945e-01 2.71994144e-01 -9.52934846e-02
5.03093004e-01 9.37481970e-02 9.60648894e-01 -6.78786635e-01
-8.37160110e-01 2.06384942e-01 1.35175511e-01 -9.16495621e-01
-5.08672833e-01 -9.57291067e-01 -1.07825756e+00 -2.44880706e-01
2.27728486e-01 -3.96753326e-02 2.10836545e-01 9.49332118e-01
4.44987327e-01 -7.08801858e-03 -5.06604686e-02 -1.20983613e+00
-3.23128521e-01 -8.77867758e-01 -5.41278660e-01 6.34264708e-01
1.32613152e-01 -7.64382482e-01 3.50301236e-01 5.40952981e-02] | [7.1550188064575195, -0.7913190126419067] |
557072ca-8a10-4fbc-aa53-5c876bb64855 | deep-kernelized-dense-geometric-matching | 2202.00667 | null | https://arxiv.org/abs/2202.00667v3 | https://arxiv.org/pdf/2202.00667v3.pdf | DKM: Dense Kernelized Feature Matching for Geometry Estimation | Feature matching is a challenging computer vision task that involves finding correspondences between two images of a 3D scene. In this paper we consider the dense approach instead of the more common sparse paradigm, thus striving to find all correspondences. Perhaps counter-intuitively, dense methods have previously shown inferior performance to their sparse and semi-sparse counterparts for estimation of two-view geometry. This changes with our novel dense method, which outperforms both dense and sparse methods on geometry estimation. The novelty is threefold: First, we propose a kernel regression global matcher. Secondly, we propose warp refinement through stacked feature maps and depthwise convolution kernels. Thirdly, we propose learning dense confidence through consistent depth and a balanced sampling approach for dense confidence maps. Through extensive experiments we confirm that our proposed dense method, \textbf{D}ense \textbf{K}ernelized Feature \textbf{M}atching, sets a new state-of-the-art on multiple geometry estimation benchmarks. In particular, we achieve an improvement on MegaDepth-1500 of +4.9 and +8.9 AUC$@5^{\circ}$ compared to the best previous sparse method and dense method respectively. Our code is provided at https://github.com/Parskatt/dkm | ['Mårten Wadenbäck', 'Ioannis Athanasiadis', 'Michael Felsberg', 'Johan Edstedt'] | 2022-02-01 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Edstedt_DKM_Dense_Kernelized_Feature_Matching_for_Geometry_Estimation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Edstedt_DKM_Dense_Kernelized_Feature_Matching_for_Geometry_Estimation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['geometric-matching'] | ['computer-vision'] | [-7.13204965e-02 3.80693525e-02 2.46148244e-01 -3.41957301e-01
-9.88435686e-01 -3.42331260e-01 5.59355199e-01 -2.19855811e-02
-3.50543499e-01 5.13743639e-01 3.78683925e-01 1.85484424e-01
-2.18216762e-01 -7.20160365e-01 -1.01018214e+00 -5.48986197e-01
-7.08396137e-02 5.04522502e-01 3.56818169e-01 6.23822026e-02
5.73173225e-01 5.74556291e-01 -1.46856189e+00 -3.02869640e-02
7.48616576e-01 1.27209020e+00 1.68547809e-01 5.55471241e-01
3.12828630e-01 7.88100362e-01 4.31180522e-02 -5.72774053e-01
6.05683625e-01 8.48811716e-02 -7.61430800e-01 4.45325486e-03
1.18357241e+00 -4.84400004e-01 -5.60050428e-01 9.81837690e-01
5.32250404e-01 1.42952144e-01 7.05524087e-01 -9.20495808e-01
-4.75764394e-01 2.23617747e-01 -1.16052699e+00 2.39026204e-01
5.08287430e-01 -1.89201489e-01 9.77385998e-01 -1.55764627e+00
6.98213160e-01 1.09517956e+00 9.87605512e-01 -3.76465265e-03
-9.87344623e-01 -7.09823191e-01 1.77137420e-01 2.26907939e-01
-1.85640967e+00 -5.30981839e-01 6.17226839e-01 -4.85427171e-01
1.18612242e+00 2.22159505e-01 5.47016442e-01 6.41691983e-01
1.79991588e-01 5.36948323e-01 1.02414584e+00 -3.89887422e-01
-1.17581328e-02 -1.30670339e-01 -1.22881211e-01 1.13627863e+00
2.28426769e-01 4.07284439e-01 -7.25290954e-01 -1.67602688e-01
1.11029327e+00 1.23975255e-01 -3.11073691e-01 -6.30070925e-01
-1.45010698e+00 8.89616072e-01 5.46383262e-01 -1.47801533e-01
-3.64776164e-01 4.09239799e-01 1.24854185e-01 9.27910656e-02
6.13458872e-01 2.15184346e-01 -3.64145160e-01 -7.81443715e-02
-1.11137354e+00 3.38434309e-01 6.52810454e-01 1.25531602e+00
1.16322613e+00 -6.75765350e-02 2.29012012e-01 8.19002986e-01
1.89680159e-01 4.24214154e-01 9.62715521e-02 -1.09730017e+00
4.32925731e-01 4.16097701e-01 1.50730379e-03 -1.26925766e+00
-3.72179031e-01 -3.73605877e-01 -1.02656722e+00 2.00362951e-01
3.27963412e-01 6.58079013e-02 -7.59626210e-01 1.25687790e+00
5.63489318e-01 7.85224199e-01 -2.61808574e-01 8.47651124e-01
8.84516358e-01 4.04842108e-01 -4.73197579e-01 2.20431224e-01
1.02960753e+00 -1.10036469e+00 -3.36382501e-02 -2.75952388e-02
4.51292574e-01 -1.29744494e+00 5.77661216e-01 5.30676544e-01
-1.21151543e+00 -4.57050145e-01 -8.73317063e-01 -3.15346897e-01
-1.23958215e-01 1.29283652e-01 9.36796904e-01 2.88200080e-01
-1.15370858e+00 6.10842109e-01 -8.11789632e-01 -3.45432669e-01
5.33543050e-01 4.74409103e-01 -6.69986606e-01 -4.13269341e-01
-5.21728456e-01 7.14661837e-01 8.58415887e-02 6.54201629e-03
-6.61625087e-01 -1.08641124e+00 -1.30954182e+00 -3.23416561e-01
3.09403330e-01 -1.04190826e+00 1.18732619e+00 -3.81836027e-01
-1.26539218e+00 1.15336001e+00 -1.87195286e-01 -4.28297490e-01
4.70717072e-01 -6.37312353e-01 1.68246835e-01 2.14951456e-01
3.06720197e-01 8.09925497e-01 8.83677602e-01 -1.05787587e+00
-6.76522136e-01 -3.46038640e-01 -1.67879596e-01 3.05304229e-01
1.88064322e-01 -4.81885374e-02 -5.75125396e-01 -5.82501411e-01
5.22431433e-01 -8.59438241e-01 -3.53522182e-01 2.81487137e-01
-4.71116871e-01 7.78464228e-02 5.79479337e-01 -3.80636632e-01
8.13652873e-01 -2.05871940e+00 1.18879057e-01 4.80478466e-01
6.10086858e-01 -2.37117872e-01 1.90817118e-01 3.60944182e-01
-1.52282557e-02 -3.95033866e-01 -2.03987718e-01 -7.13392377e-01
-1.46807626e-01 9.91868507e-03 -5.12127616e-02 1.10804522e+00
7.15986341e-02 8.41038883e-01 -7.36908615e-01 -5.23418069e-01
6.70099974e-01 6.18265688e-01 -8.37776721e-01 9.42515209e-02
2.55693525e-01 3.79901052e-01 -2.46114165e-01 9.42318916e-01
9.83082950e-01 -4.61979210e-01 -4.08045709e-01 -3.93118143e-01
-4.25453335e-01 -1.09595485e-01 -1.58089244e+00 2.21116877e+00
-4.32653427e-01 2.87321389e-01 -4.87918109e-02 -1.05145001e+00
1.03253269e+00 4.76539880e-02 7.06439793e-01 -4.70229715e-01
2.26921424e-01 3.78999174e-01 -3.96357089e-01 7.49281899e-04
5.84592462e-01 -6.18549762e-03 -2.16293111e-02 9.72583294e-02
1.47298738e-01 -6.45623147e-01 -2.49484181e-01 2.85758674e-01
1.18056393e+00 2.98307210e-01 4.92210895e-01 -3.86315435e-01
4.81636494e-01 -4.31348048e-02 3.77489656e-01 5.88118732e-01
3.99611071e-02 1.20957041e+00 1.95264608e-01 -4.25578296e-01
-1.19204867e+00 -1.13603866e+00 -3.97619188e-01 4.56324130e-01
3.51040393e-01 -4.49536592e-01 -5.01799822e-01 -2.33668983e-01
2.68781513e-01 4.88668010e-02 -6.92275345e-01 1.89461365e-01
-6.19583905e-01 -3.86292070e-01 3.03514004e-01 6.36576235e-01
6.20087147e-01 -5.39442837e-01 -5.80700517e-01 -8.66148919e-02
1.71791673e-01 -1.36265075e+00 -7.87355602e-01 6.41845986e-02
-8.58929873e-01 -1.10904634e+00 -8.09140861e-01 -8.73325884e-01
7.82534361e-01 4.30921078e-01 1.36387885e+00 6.19298890e-02
-6.07551694e-01 4.49703515e-01 -4.16985661e-01 -1.71058714e-01
4.23019618e-01 6.83722347e-02 7.91230127e-02 -1.13585904e-01
3.59584510e-01 -8.85089397e-01 -1.09494019e+00 3.47766310e-01
-4.14928228e-01 2.54100919e-01 7.30000198e-01 7.68614590e-01
9.87070501e-01 -4.88489985e-01 3.54114436e-02 -8.80452752e-01
5.80620766e-02 -5.80548942e-01 -6.90078855e-01 -1.13132954e-01
-5.67561984e-01 2.21347995e-02 3.32863152e-01 3.95070808e-03
-6.69507265e-01 4.49862063e-01 -3.35754216e-01 -7.98868358e-01
-2.53198266e-01 1.83070913e-01 3.27905804e-01 -5.01237214e-01
5.97930431e-01 2.51747489e-01 2.40688934e-03 -4.47627008e-01
3.79538298e-01 3.01974565e-01 6.42153144e-01 -6.63055301e-01
1.11416304e+00 8.53489220e-01 1.84426203e-01 -7.12327480e-01
-8.68936956e-01 -9.03508008e-01 -9.33957398e-01 -2.06603277e-02
7.62222826e-01 -1.28259218e+00 -7.15226710e-01 4.88582581e-01
-9.52630043e-01 -1.44620702e-01 -2.87232816e-01 6.79267943e-01
-8.87245178e-01 6.23286843e-01 -4.38092053e-01 -3.50310832e-01
-4.57765311e-01 -1.03219724e+00 1.48695731e+00 1.69919178e-01
-1.86732680e-01 -8.96113038e-01 3.64627689e-01 2.56003946e-01
3.86656731e-01 5.37206411e-01 3.22330207e-01 -1.72612503e-01
-8.87155890e-01 -2.21373349e-01 -6.35592580e-01 1.48240328e-01
2.28411448e-03 -1.28493473e-01 -9.22349811e-01 -3.10053617e-01
-2.40333036e-01 -4.62904841e-01 8.05682361e-01 6.22866690e-01
1.14826465e+00 2.82830726e-02 -2.35299557e-01 1.28780508e+00
1.75801945e+00 -4.59370047e-01 6.87929332e-01 2.65048236e-01
8.60100925e-01 2.86657840e-01 7.09966004e-01 8.36543918e-01
5.49740136e-01 8.23733449e-01 5.31080246e-01 -1.87553659e-01
-1.92232743e-01 -2.30041727e-01 -4.79345620e-02 7.70503044e-01
-4.09123063e-01 3.39231133e-01 -8.37941647e-01 5.23831546e-01
-1.86765969e+00 -8.07189345e-01 -2.34263077e-01 2.23154879e+00
6.00354433e-01 -3.90315242e-02 6.81989640e-02 -5.74733876e-02
4.60987806e-01 4.35853451e-02 -2.56156027e-01 -1.79440781e-01
-3.23097073e-02 8.48922551e-01 7.37623751e-01 6.93244517e-01
-1.19156063e+00 9.63246047e-01 5.25379705e+00 8.48015666e-01
-8.94559622e-01 8.12360644e-02 5.03463924e-01 -1.93229645e-01
-2.43941918e-01 -4.00629789e-02 -1.02390671e+00 1.72452539e-01
5.19199371e-01 1.02053963e-01 3.91398072e-02 9.60835159e-01
-3.58174630e-02 -4.98823702e-01 -1.22932887e+00 1.45173478e+00
3.33223909e-01 -1.68428171e+00 -2.52240986e-01 6.93298727e-02
9.32233572e-01 3.18482578e-01 6.55693859e-02 4.16141190e-02
3.17814887e-01 -1.20583737e+00 8.91516566e-01 5.72986126e-01
9.80111241e-01 -7.55103886e-01 6.81903839e-01 1.67153507e-01
-1.55971444e+00 4.10807759e-01 -5.80497563e-01 -1.06094308e-01
1.42190978e-01 8.37530732e-01 -5.05889535e-01 7.26589799e-01
9.84846532e-01 1.06969810e+00 -4.85627711e-01 1.28300083e+00
7.45847300e-02 5.84303997e-02 -5.48046350e-01 2.95799524e-01
3.49187136e-01 -1.59248069e-01 3.25231642e-01 1.22342432e+00
4.71046835e-01 1.96093395e-01 2.24086478e-01 6.88402534e-01
-5.88116832e-02 1.70929909e-01 -8.26577187e-01 7.23970771e-01
4.31858957e-01 1.22217214e+00 -6.69605076e-01 -1.38951719e-01
-7.78067946e-01 1.08837116e+00 6.24247491e-01 -2.99664792e-02
-7.58092344e-01 -2.95883536e-01 6.09245002e-01 4.07858223e-01
6.40554190e-01 -4.82460707e-01 -2.69558400e-01 -1.25242424e+00
5.43503277e-02 -5.75510383e-01 3.32499057e-01 -6.27124310e-01
-1.30433166e+00 4.48719859e-01 1.18766136e-01 -1.38825369e+00
-9.00310799e-02 -5.39562285e-01 -4.60880756e-01 9.90448475e-01
-1.76946402e+00 -1.15080357e+00 -7.38234222e-01 1.00955105e+00
7.09101975e-01 -2.16789041e-02 7.56963253e-01 4.02610213e-01
-2.05984816e-01 6.34578049e-01 9.07118767e-02 8.15359578e-02
7.23615348e-01 -1.18927312e+00 5.35145819e-01 7.25388050e-01
3.46485287e-01 3.65144134e-01 3.84350121e-01 -5.52431583e-01
-1.56750906e+00 -9.46246505e-01 6.20673835e-01 -5.23268700e-01
5.63486457e-01 -2.57832825e-01 -6.23290122e-01 6.64193869e-01
1.27964720e-01 5.73450744e-01 4.79382634e-01 -4.00504358e-02
-5.49507022e-01 7.66271427e-02 -1.33119047e+00 2.24546269e-01
1.28483427e+00 -3.29041511e-01 -3.60995978e-01 2.19246909e-01
5.00593245e-01 -9.11972344e-01 -1.14513719e+00 5.87028205e-01
5.67992151e-01 -1.40034032e+00 1.30134296e+00 1.12667615e-02
5.63241184e-01 -2.21714646e-01 -5.22796512e-01 -8.73805046e-01
-3.45121235e-01 -6.20617270e-01 -3.16835046e-01 6.81774735e-01
1.73077002e-01 -5.18835068e-01 1.20343637e+00 1.56362981e-01
-4.12496567e-01 -1.15939260e+00 -9.49349225e-01 -6.44974530e-01
5.62611781e-02 -4.19208050e-01 2.51202226e-01 8.91752779e-01
-3.89907449e-01 -3.09821721e-02 -3.87231141e-01 3.59423965e-01
9.30338264e-01 2.52691388e-01 8.79792094e-01 -1.09693205e+00
-4.19837832e-01 -2.71462470e-01 -7.81997204e-01 -1.40947354e+00
-7.43462984e-03 -8.58954549e-01 -7.28211403e-02 -1.34395647e+00
2.56000996e-01 -6.85798585e-01 8.39130133e-02 1.30935445e-01
4.61382270e-02 6.82836175e-01 -1.15354240e-01 1.87881589e-01
-4.68440145e-01 4.98074114e-01 1.09784722e+00 2.32020408e-01
3.27282660e-02 -7.26785213e-02 -5.43293834e-01 7.97026813e-01
6.58896267e-01 -2.99394101e-01 -2.89147466e-01 -4.79957074e-01
1.78063363e-01 -1.15767695e-01 6.17412031e-01 -1.10737443e+00
3.63767654e-01 -6.06968664e-02 5.34209847e-01 -8.33385706e-01
6.08371615e-01 -7.10625887e-01 1.54269531e-01 2.45377079e-01
2.72409022e-01 3.40919673e-01 1.72076806e-01 5.54561079e-01
-2.84297109e-01 -5.69558032e-02 8.98258507e-01 -4.88887817e-01
-1.02302873e+00 8.09083581e-01 1.95314139e-01 2.02959791e-01
1.07658422e+00 -6.10323370e-01 -1.22214638e-01 -2.18816951e-01
-6.28103137e-01 1.33736402e-01 5.75401604e-01 1.43487945e-01
1.18125653e+00 -1.37391102e+00 -8.98914397e-01 3.16180319e-01
1.58734724e-01 5.41542768e-01 3.12028885e-01 1.21200061e+00
-9.53345835e-01 3.71240199e-01 -1.35708719e-01 -8.62491786e-01
-1.28459680e+00 2.23588184e-01 2.64633566e-01 3.25904302e-02
-1.04298210e+00 1.40245163e+00 2.57668972e-01 -3.91751796e-01
3.84676814e-01 -3.99643660e-01 2.25463465e-01 -2.31549308e-01
3.45037073e-01 4.46974277e-01 3.02897185e-01 -6.90322638e-01
-5.21428049e-01 1.26142776e+00 -2.61824310e-01 1.33083150e-01
1.53605902e+00 -2.70832386e-02 -6.73116371e-02 -2.28381753e-02
1.40677679e+00 1.92203522e-01 -1.30975807e+00 -3.79055440e-01
-2.71120518e-01 -1.02397811e+00 5.04686497e-02 -2.37401977e-01
-1.06523573e+00 6.23120487e-01 5.28151333e-01 -3.43651116e-01
9.49086487e-01 3.48585129e-01 7.90831327e-01 2.00349241e-01
5.31660318e-01 -7.57194698e-01 2.40204230e-01 4.97729093e-01
1.06373775e+00 -1.52832842e+00 4.46944743e-01 -8.33504140e-01
-3.66958588e-01 1.01894689e+00 6.61824942e-01 -7.89434433e-01
1.07522810e+00 4.57613587e-01 -3.02381158e-01 -5.99869370e-01
-4.29290324e-01 -2.43973747e-01 3.76606017e-01 5.31928241e-01
4.74437386e-01 -1.42092139e-01 1.35618627e-01 3.59207881e-03
-5.04021645e-01 -1.49161726e-01 3.08110446e-01 8.69403780e-01
-4.46465462e-01 -9.22196031e-01 -3.66976112e-01 5.57640135e-01
-2.16289759e-01 -3.28037709e-01 -5.19882366e-02 8.92095804e-01
7.77642205e-02 4.01812702e-01 1.98302627e-01 -4.28313226e-01
4.45673347e-01 -2.83649534e-01 8.97387028e-01 -9.22992170e-01
-5.05255342e-01 5.09637259e-02 -1.00438885e-01 -9.75741029e-01
-4.49102551e-01 -9.15839374e-01 -1.11542785e+00 -6.42199159e-01
-3.11753362e-01 -1.53320938e-01 4.12396371e-01 6.50152922e-01
2.96948314e-01 1.83388554e-02 6.55307174e-01 -1.14420056e+00
-4.85972703e-01 -7.41432488e-01 -5.46362758e-01 1.95125163e-01
2.42952392e-01 -9.20333028e-01 -3.82386535e-01 -1.45446375e-01] | [8.407602310180664, -2.4740753173828125] |
db3bcecb-dea7-4d91-8823-48822f5dae09 | correcting-real-word-spelling-errors-a-new | 2302.06407 | null | https://arxiv.org/abs/2302.06407v1 | https://arxiv.org/pdf/2302.06407v1.pdf | Correcting Real-Word Spelling Errors: A New Hybrid Approach | Spelling correction is one of the main tasks in the field of Natural Language Processing. Contrary to common spelling errors, real-word errors cannot be detected by conventional spelling correction methods. The real-word correction model proposed by Mays, Damerau and Mercer showed a great performance in different evaluations. In this research, however, a new hybrid approach is proposed which relies on statistical and syntactic knowledge to detect and correct real-word errors. In this model, Constraint Grammar (CG) is used to discriminate among sets of correction candidates in the search space. Mays, Damerau and Mercer's trigram approach is manipulated to estimate the probability of syntactically well-formed correction candidates. The approach proposed here is tested on the Wall Street Journal corpus. The model can prove to be more practical than some other models, such as WordNet-based method of Hirst and Budanitsky and fixed windows size method of Wilcox-O'Hearn and Hirst. | ['Vahid Khatibi Bardsiri', 'Amid Khatibi Bardsiri', 'Seyed MohammadSadegh Dashti'] | 2023-02-09 | null | null | null | null | ['spelling-correction'] | ['natural-language-processing'] | [ 1.37087047e-01 -9.48885605e-02 4.80534919e-02 -2.64615536e-01
-4.90859777e-01 -3.05577695e-01 5.36119044e-01 8.82759631e-01
-8.91982138e-01 1.01299620e+00 1.49525508e-01 -5.58689713e-01
-2.62681007e-01 -6.86312616e-01 -2.87223488e-01 -2.37776279e-01
4.07560915e-01 4.66506481e-01 5.68829894e-01 -3.71768624e-01
9.91322160e-01 3.13864410e-01 -1.61826086e+00 1.20613672e-01
1.12638688e+00 1.28557414e-01 7.26598680e-01 7.07339048e-01
-6.44214332e-01 3.84756476e-01 -6.60362601e-01 -1.71985030e-01
-8.16378742e-02 -5.90146720e-01 -6.34095550e-01 -3.14982593e-01
1.35501280e-01 3.62281829e-01 3.48448932e-01 1.34206069e+00
3.90234381e-01 3.31455737e-01 6.87599599e-01 -6.10799253e-01
-4.72517729e-01 6.71366274e-01 -1.87059656e-01 3.90268028e-01
7.67380416e-01 -2.41303906e-01 7.06038773e-01 -1.02507889e+00
7.37689734e-01 1.11479020e+00 5.39702475e-01 4.93415862e-01
-7.49057055e-01 -4.66033578e-01 -2.28628479e-02 4.67394978e-01
-1.31825578e+00 9.82063115e-02 2.15898469e-01 -4.04827148e-01
1.12611473e+00 4.43999946e-01 4.47650790e-01 4.24293250e-01
4.34105188e-01 5.72981238e-01 1.38616288e+00 -1.31529498e+00
2.04397783e-01 1.86771750e-01 4.42152083e-01 5.76286376e-01
6.12196505e-01 -6.35085106e-02 -3.15936863e-01 -1.69148117e-01
1.94941521e-01 -6.11430332e-02 -2.68709421e-01 2.26080179e-01
-8.67580533e-01 7.99991012e-01 -2.60668516e-01 9.99247491e-01
-3.17101687e-01 -2.00741589e-01 3.60681891e-01 8.37651491e-02
3.91104788e-01 5.84595263e-01 -4.81687307e-01 -3.91794145e-01
-1.22055888e+00 5.46759844e-01 7.00471520e-01 7.01245666e-01
3.11340004e-01 4.31706160e-02 -2.64864624e-01 5.86148202e-01
5.45141280e-01 3.66457611e-01 8.84386003e-01 -1.21481545e-01
4.18713242e-01 7.94714808e-01 2.28062853e-01 -9.07544494e-01
-2.54193604e-01 -2.37797439e-01 -1.23347677e-01 2.49799833e-01
6.76184475e-01 8.15203786e-02 -9.60187376e-01 1.21435785e+00
9.37526226e-02 -2.19822600e-01 5.47485724e-02 7.12442636e-01
5.13709545e-01 6.89409256e-01 2.65514255e-01 -4.73543555e-01
1.16755676e+00 -6.39743090e-01 -9.60698068e-01 -1.34784937e-01
7.48990297e-01 -1.28570533e+00 1.03776813e+00 6.44635618e-01
-9.85288024e-01 -3.14380854e-01 -9.70959306e-01 2.19310388e-01
-5.77144802e-01 1.67885292e-02 3.60057324e-01 9.53517556e-01
-7.03657031e-01 6.95321143e-01 -4.71001744e-01 -6.84784651e-01
-3.05435210e-01 5.01072779e-02 -5.18987663e-02 -6.04295991e-02
-1.05219555e+00 1.22772682e+00 7.12399840e-01 -6.41710758e-02
-3.55877161e-01 -1.39004782e-01 -6.59227073e-01 -1.10414982e-01
3.01291436e-01 -7.73916394e-02 1.17017686e+00 -1.10636437e+00
-1.11677265e+00 8.13387215e-01 -4.58188236e-01 -5.28084457e-01
4.02459592e-01 -9.60953161e-02 -6.36150062e-01 -1.49649799e-01
8.90039206e-02 -1.00844465e-01 4.16261554e-01 -9.38388407e-01
-8.22520792e-01 -2.57664204e-01 -3.18201810e-01 2.37478793e-01
3.72965285e-03 8.39534283e-01 6.23637550e-02 -7.66452789e-01
1.62306771e-01 -7.04150259e-01 -2.44145125e-01 -5.50310194e-01
-4.91725840e-02 -4.28719521e-01 8.85763690e-02 -8.45727563e-01
1.67208493e+00 -1.92516112e+00 -1.32017285e-01 4.04263675e-01
-1.73488542e-01 9.06555772e-01 1.61574692e-01 6.53083920e-01
-1.33848473e-01 2.56155521e-01 -3.51517588e-01 -4.12077196e-02
-2.81273341e-03 2.89134204e-01 6.31891936e-02 1.76784948e-01
-2.83137500e-01 2.68829465e-01 -9.79047358e-01 -4.89438087e-01
1.89467490e-01 9.97835472e-02 -2.65994579e-01 -2.06731707e-02
-1.45116389e-01 4.04167622e-02 -2.55996346e-01 5.47129869e-01
5.56462467e-01 3.52625638e-01 1.45852610e-01 4.83324736e-01
-6.41681254e-01 6.63498461e-01 -1.44260144e+00 1.02674162e+00
-3.35316867e-01 2.99908191e-01 -3.42401713e-01 -6.96614623e-01
1.12164330e+00 3.61636639e-01 -2.36760244e-01 -6.76736593e-01
2.75659263e-01 7.86059916e-01 1.15187839e-01 -4.97877419e-01
8.69567633e-01 -1.57550946e-01 4.90139611e-02 2.39922911e-01
-1.99893743e-01 1.52499685e-02 6.70432985e-01 -1.06448442e-01
8.88086557e-01 1.17429256e-01 9.50603962e-01 -4.47247803e-01
7.25421727e-01 1.61222115e-01 6.19283319e-01 8.18763196e-01
-1.86289340e-01 5.03727853e-01 2.02175424e-01 -3.02553982e-01
-8.29456210e-01 -6.47222102e-01 -1.45699605e-01 9.83255982e-01
3.19966637e-02 -5.72082996e-01 -8.38159919e-01 -4.75865930e-01
-2.77165025e-01 1.29739630e+00 -2.65853226e-01 1.12234928e-01
-5.38927794e-01 -6.13833964e-01 4.00935203e-01 -7.55249038e-02
1.30392268e-01 -1.32118809e+00 -7.70112574e-01 4.31128234e-01
5.99385202e-02 -5.93167067e-01 -2.59260088e-01 2.07349658e-01
-9.66333091e-01 -1.03119147e+00 -4.53219295e-01 -7.20843375e-01
6.00300491e-01 -5.80608845e-02 7.65922308e-01 5.17666399e-01
-2.36369118e-01 -7.61141405e-02 -1.04747772e+00 -8.22704911e-01
-9.60829556e-01 -1.97389796e-01 9.96722258e-04 -3.75301033e-01
7.88766921e-01 -1.51710749e-01 -1.51025474e-01 -6.34345710e-02
-8.36114228e-01 -4.65857536e-01 5.36900818e-01 6.68984473e-01
4.89626735e-01 -1.01817973e-01 3.55893999e-01 -9.65483427e-01
8.62725616e-01 -1.74970642e-01 -5.80409646e-01 3.10427964e-01
-1.12221980e+00 6.36723563e-02 5.04123449e-01 -2.23296970e-01
-9.96158957e-01 -1.44151866e-01 -5.18587649e-01 3.14029366e-01
-4.66801703e-01 5.56407511e-01 -1.03781503e-02 -2.80247211e-01
7.88161993e-01 3.71425211e-01 -3.44094574e-01 -7.09526300e-01
-2.89158016e-01 7.37681746e-01 1.38209894e-01 -1.79337442e-01
6.03083014e-01 -2.30444223e-01 -1.17090888e-01 -6.66063428e-01
-4.61708248e-01 -6.62373185e-01 -4.45778906e-01 -3.05099815e-01
8.70987177e-01 -3.63701582e-01 -1.95053861e-01 4.53607857e-01
-1.39268506e+00 1.99533999e-01 3.00142705e-03 7.96153128e-01
-1.21178210e-01 7.44241714e-01 -3.68294477e-01 -1.14311063e+00
-1.51426107e-01 -8.39738429e-01 5.63083410e-01 3.64856154e-01
-4.01431203e-01 -8.31884742e-01 1.72373965e-01 1.17448702e-01
1.33731753e-01 -3.54688913e-02 1.05185401e+00 -1.15239918e+00
-6.76274449e-02 -4.50755715e-01 1.82184502e-01 4.12073433e-01
-1.84330881e-01 1.17660299e-01 -2.76275605e-01 5.45192845e-02
1.56539738e-01 3.19985926e-01 6.65727854e-01 1.80891827e-01
6.35853112e-01 -1.76301613e-01 -1.46557853e-01 3.22333537e-02
1.70870805e+00 5.91844797e-01 9.42731619e-01 7.67358541e-01
3.29514503e-01 2.71806985e-01 1.13278151e+00 4.50023592e-01
5.98275065e-02 4.40625370e-01 2.10885093e-01 4.15288776e-01
3.74583006e-02 -2.74797618e-01 2.96675622e-01 1.00039601e+00
-2.76905119e-01 -4.27378446e-01 -1.14788401e+00 6.26924098e-01
-1.65578604e+00 -1.00472438e+00 -7.25297809e-01 2.35477829e+00
7.06057370e-01 6.01295292e-01 -2.71966040e-01 6.08175695e-01
9.84761477e-01 -5.05207060e-03 5.10293901e-01 -1.18874216e+00
-1.36892319e-01 4.41019058e-01 6.63613319e-01 8.71351421e-01
-7.22569168e-01 1.02268362e+00 5.39340830e+00 9.16439176e-01
-8.29306006e-01 1.67449504e-01 -7.91743770e-02 4.34565544e-01
-3.85039181e-01 2.64197528e-01 -1.12961352e+00 6.88356698e-01
9.89728689e-01 -2.04695955e-01 1.87613532e-01 4.56927866e-01
3.87712061e-01 -8.25855553e-01 -5.62936664e-01 6.35877490e-01
4.26696748e-01 -8.62142146e-01 -6.81321993e-02 -2.59006828e-01
6.51931703e-01 -4.34736103e-01 -5.34353018e-01 2.61669487e-01
9.93695483e-02 -7.73450315e-01 7.48826027e-01 7.14101672e-01
2.26188660e-01 -7.55978048e-01 1.01651025e+00 7.29670703e-01
-7.11045444e-01 4.34311740e-02 -5.21999657e-01 -5.96730828e-01
1.41652673e-01 6.51113510e-01 -8.75137746e-01 5.81119478e-01
1.58668280e-01 4.05708421e-03 -6.33293271e-01 1.52258718e+00
-5.06694436e-01 9.18824792e-01 -1.75168410e-01 -7.23308563e-01
2.45567352e-01 -1.80873722e-01 7.07221568e-01 1.42519414e+00
5.53617597e-01 1.53373167e-01 -1.13365263e-01 4.24498320e-01
5.67401052e-01 7.71795452e-01 -3.81876677e-01 9.32295748e-04
6.32684708e-01 6.85789704e-01 -9.15759504e-01 -3.03025067e-01
-3.07297081e-01 7.74496555e-01 -3.32525223e-02 2.04836186e-02
-5.20364583e-01 -5.70212960e-01 3.21359225e-02 2.72543430e-01
2.36303881e-01 -2.57371515e-01 -3.61836344e-01 -7.60824382e-01
3.09757646e-02 -8.47831786e-01 2.41236374e-01 -5.09487510e-01
-7.96441078e-01 4.84958321e-01 1.12545937e-01 -1.01274478e+00
-1.37193322e-01 -7.84480274e-01 -6.85257375e-01 1.19652796e+00
-1.34716797e+00 -6.98185086e-01 1.22660667e-01 2.18846917e-01
8.52625847e-01 -1.46503299e-01 8.23141277e-01 3.30982000e-01
-3.08903098e-01 4.21839982e-01 3.01099926e-01 -1.23605534e-01
7.48717248e-01 -1.34626341e+00 1.32189527e-01 1.27051079e+00
4.15566526e-02 5.08435905e-01 1.21067238e+00 -1.19429636e+00
-8.09512496e-01 -6.64486170e-01 2.01965833e+00 -1.20211497e-01
4.89298970e-01 -1.11274328e-02 -1.13964868e+00 4.04960364e-01
2.70968258e-01 -5.52809238e-01 5.26534081e-01 -2.53224671e-01
-1.35903373e-01 1.65137187e-01 -9.99123931e-01 5.16393721e-01
6.31182253e-01 -1.06896646e-01 -1.28250170e+00 3.40140522e-01
2.86037117e-01 -3.40932488e-01 -2.90971339e-01 8.84240866e-02
2.19070286e-01 -9.77122903e-01 2.75454223e-01 -4.35386926e-01
1.58957452e-01 -4.26175743e-01 2.12449790e-03 -1.35980105e+00
-1.63135365e-01 -4.17513490e-01 6.79533601e-01 1.32923377e+00
6.41565979e-01 -5.90685487e-01 1.23485200e-01 3.74388695e-01
-3.51712853e-01 -2.67221361e-01 -9.38143790e-01 -7.98543334e-01
-3.49888317e-02 -4.23617810e-01 3.70539188e-01 8.05757105e-01
2.98283219e-01 -6.17820323e-02 -3.39387320e-02 -1.07476257e-01
4.15953785e-01 -3.70623708e-01 3.58830303e-01 -1.31564975e+00
-2.15899885e-01 -4.42044437e-01 -4.21125084e-01 -3.48494977e-01
3.25403363e-02 -8.68631303e-01 3.14778537e-01 -1.80108356e+00
-1.67420924e-01 -3.78362000e-01 -2.36289531e-01 1.68791503e-01
-3.04551363e-01 -7.38386214e-02 4.62656558e-01 -1.79051980e-02
-3.88406813e-01 1.02086358e-01 9.26708102e-01 2.92444557e-01
-2.51040637e-01 2.37592403e-02 -3.74812603e-01 9.22774434e-01
8.35504889e-01 -9.12026167e-01 -5.24499565e-02 -2.89773405e-01
6.06063008e-01 1.25273196e-02 -8.35185274e-02 -1.02369499e+00
4.09622639e-01 -3.78956616e-01 -4.18797228e-03 -4.76395756e-01
-2.83643186e-01 -5.81656873e-01 1.50762275e-01 6.30497038e-01
-2.71049708e-01 5.58132946e-01 1.23994403e-01 4.93310839e-01
-2.12684780e-01 -1.20511866e+00 8.05371702e-01 -4.28971291e-01
-8.04893076e-01 -3.15784067e-01 -8.83657098e-01 -9.02177766e-02
1.05872405e+00 -4.86558110e-01 -4.57602069e-02 5.38229309e-02
-5.69930792e-01 -1.81199983e-01 3.93842906e-01 1.98532715e-01
6.46231771e-01 -8.92864525e-01 -7.09065378e-01 2.37322077e-02
1.81219786e-01 -4.37745601e-01 3.00761294e-02 8.75164330e-01
-1.00437486e+00 4.81399834e-01 -2.00914189e-01 -2.90637501e-02
-1.61092079e+00 5.21808326e-01 3.08346115e-02 -3.62953961e-01
-3.66068393e-01 7.28001297e-01 -5.83251297e-01 -1.01472944e-01
7.57399648e-02 -3.41374308e-01 -6.02736890e-01 -4.37542051e-03
6.94279730e-01 4.46200490e-01 2.54536182e-01 -7.13419557e-01
-5.23419797e-01 4.60078478e-01 -7.40763023e-02 -2.14772001e-01
9.82764184e-01 -1.39038011e-01 -3.41437519e-01 7.19195843e-01
5.85937023e-01 5.70574820e-01 -1.18047990e-01 -6.29292009e-03
6.77307725e-01 -6.87051475e-01 -1.31725416e-01 -1.06795597e+00
-7.36154392e-02 7.54275143e-01 6.00437343e-01 1.97542623e-01
9.57246602e-01 -4.38982666e-01 5.72275937e-01 3.49437296e-01
4.91837591e-01 -1.52964258e+00 -6.53641582e-01 7.62720346e-01
6.37118578e-01 -9.38601434e-01 -1.54833928e-01 -4.21826959e-01
-5.16534746e-01 1.20937932e+00 4.44001943e-01 -8.29660371e-02
5.11785805e-01 8.66466984e-02 -2.56253202e-02 1.68904930e-01
-5.08053839e-01 -4.10832226e-01 1.59637958e-01 4.87005383e-01
7.83787370e-01 2.76263684e-01 -1.95831621e+00 5.14692962e-01
-5.57517670e-02 1.56766176e-01 9.57286119e-01 1.06196237e+00
-9.82308447e-01 -1.53917897e+00 -8.28083456e-01 4.96192157e-01
-5.78884602e-01 -4.77365345e-01 -3.81964922e-01 8.64268243e-01
3.83343399e-01 1.04507554e+00 -3.04730862e-01 -1.32461846e-01
4.26264316e-01 3.76822710e-01 4.37853307e-01 -9.08471286e-01
-1.09079623e+00 1.05181701e-01 2.96286643e-01 -5.53481132e-02
-4.11234289e-01 -9.16037083e-01 -1.37417078e+00 -2.27282628e-01
-4.52826083e-01 6.20518804e-01 9.12820160e-01 1.31775761e+00
-2.69198895e-01 2.50320077e-01 7.73046240e-02 -1.47503778e-01
-5.99935412e-01 -1.03440356e+00 -6.99088335e-01 2.79255033e-01
-1.01953875e-02 -5.06125987e-01 -4.12063330e-01 -1.95196018e-01] | [10.843220710754395, 10.580338478088379] |
f0aad60e-91c8-4002-9a0e-fca38fe46f9d | justifying-corpus-based-choices-in-referring | null | null | https://aclanthology.org/R13-1040 | https://aclanthology.org/R13-1040.pdf | Justifying Corpus-Based Choices in Referring Expression Generation | null | ['Helmut Horacek'] | 2013-09-01 | justifying-corpus-based-choices-in-referring-1 | https://aclanthology.org/R13-1040 | https://aclanthology.org/R13-1040.pdf | ranlp-2013-9 | ['referring-expression-generation'] | ['computer-vision'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.296783447265625, 3.758841037750244] |
90a52c84-641f-4186-8277-3800698d5bf7 | 3d-coded-3d-correspondences-by-deep-1 | 1806.05228 | null | http://arxiv.org/abs/1806.05228v2 | http://arxiv.org/pdf/1806.05228v2.pdf | 3D-CODED : 3D Correspondences by Deep Deformation | We present a new deep learning approach for matching deformable shapes by
introducing {\it Shape Deformation Networks} which jointly encode 3D shapes and
correspondences. This is achieved by factoring the surface representation into
(i) a template, that parameterizes the surface, and (ii) a learnt global
feature vector that parameterizes the transformation of the template into the
input surface. By predicting this feature for a new shape, we implicitly
predict correspondences between this shape and the template. We show that these
correspondences can be improved by an additional step which improves the shape
feature by minimizing the Chamfer distance between the input and transformed
template. We demonstrate that our simple approach improves on state-of-the-art
results on the difficult FAUST-inter challenge, with an average correspondence
error of 2.88cm. We show, on the TOSCA dataset, that our method is robust to
many types of perturbations, and generalizes to non-human shapes. This
robustness allows it to perform well on real unclean, meshes from the the SCAPE
dataset. | ['Thibault Groueix', 'Bryan C. Russell', 'Matthew Fisher', 'Mathieu Aubry', 'Vladimir G. Kim'] | 2018-06-13 | null | null | null | null | ['3d-dense-shape-correspondence', '3d-surface-generation', '3d-point-cloud-matching'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 2.21025631e-01 3.93597364e-01 3.27740043e-01 -4.36056703e-01
-8.38918030e-01 -8.31778347e-01 7.96584487e-01 -9.63949971e-03
-7.90824518e-02 2.76867092e-01 2.51833797e-01 1.81072488e-01
2.76107434e-02 -9.19389546e-01 -1.15736771e+00 -4.82818961e-01
-1.84027180e-01 1.14422405e+00 2.47620985e-01 -2.67673314e-01
-2.29070336e-02 1.06526709e+00 -1.20797169e+00 2.76751131e-01
2.68076718e-01 1.03303802e+00 -3.60971928e-01 5.04239738e-01
2.91310042e-01 -2.44233206e-01 -8.61625373e-02 -3.94233406e-01
7.43097305e-01 5.44552431e-02 -1.05876911e+00 2.60870576e-01
1.01734769e+00 -3.66728067e-01 -3.44026297e-01 7.05935657e-01
5.07327080e-01 -1.69307906e-02 9.13309991e-01 -8.54222715e-01
-6.52069509e-01 3.01787764e-01 -4.72852319e-01 -4.32779759e-01
2.66513288e-01 1.27294153e-01 8.91798556e-01 -1.04300582e+00
1.15436995e+00 1.47444201e+00 1.06243443e+00 5.76093853e-01
-1.57377040e+00 -4.03369009e-01 -3.30816686e-01 -5.73657274e-01
-1.36605120e+00 -6.01431608e-01 7.52662241e-01 -8.03818524e-01
8.97273779e-01 2.00778618e-01 6.47623062e-01 6.56921685e-01
2.97626793e-01 3.74174953e-01 7.16774940e-01 -2.38321021e-01
-1.36105614e-02 -5.47136247e-01 -4.88683313e-01 7.66070902e-01
1.35178557e-02 3.82480860e-01 2.13554781e-02 -2.65709400e-01
1.25247729e+00 -3.08757305e-01 9.40973312e-02 -9.71311331e-01
-1.22519410e+00 3.96049082e-01 7.60106385e-01 2.67020553e-01
-2.96977103e-01 5.14947891e-01 5.70334047e-02 4.40511964e-02
5.58229208e-01 2.69656241e-01 -5.64070165e-01 3.68251652e-02
-7.72504747e-01 5.85952461e-01 7.93389022e-01 7.69647598e-01
7.68837333e-01 2.15602014e-02 3.01082954e-02 6.47330284e-01
4.00252789e-01 9.42359865e-01 -1.34195819e-01 -1.19824231e+00
3.60994726e-01 6.42725408e-01 2.09543988e-01 -1.23735046e+00
-6.97395802e-01 -1.55190945e-01 -7.92371094e-01 5.25288880e-01
5.07032692e-01 -7.71946758e-02 -1.18336284e+00 1.67128193e+00
4.54749137e-01 1.88883290e-01 -4.77835178e-01 7.82270551e-01
8.80011320e-01 1.66431844e-01 -3.75685662e-01 5.16037047e-01
9.81746078e-01 -4.87356454e-01 -9.30540785e-02 -1.48045361e-01
3.47064078e-01 -8.27673018e-01 7.75518000e-01 1.32259279e-02
-1.49275506e+00 -3.81729484e-01 -9.84115839e-01 -1.18545443e-01
-1.05022965e-02 1.18387222e-01 2.20040157e-01 2.70846218e-01
-1.28075230e+00 1.17174077e+00 -1.21606326e+00 -4.40297842e-01
6.40469551e-01 7.39734232e-01 -7.15245664e-01 3.02314103e-01
-6.54153705e-01 9.63549614e-01 -5.54131605e-02 1.91014986e-02
-6.57232702e-01 -9.15689111e-01 -8.71518195e-01 1.61651317e-02
-3.73210683e-02 -1.02951074e+00 1.05911803e+00 -7.74335861e-01
-1.53897405e+00 1.37880170e+00 6.61430284e-02 -2.69243538e-01
7.48645902e-01 4.00997326e-02 6.11164421e-02 -2.02327012e-03
-2.03144133e-01 8.40164483e-01 9.67338681e-01 -1.45859587e+00
1.07904285e-01 -4.13810164e-01 -2.22326517e-01 -9.47199389e-03
1.83804885e-01 -2.20853359e-01 -4.33124363e-01 -8.77081633e-01
4.08386439e-01 -1.18087316e+00 -2.14432791e-01 6.40330791e-01
-4.66699272e-01 -1.45480810e-02 7.52479136e-01 -7.79216170e-01
4.77592528e-01 -1.98658562e+00 4.12807435e-01 6.48531735e-01
2.98100442e-01 1.76406875e-01 -4.47404921e-01 2.69785613e-01
-1.77669346e-01 1.92178428e-01 -6.28000855e-01 -5.65526307e-01
1.01563126e-01 1.81096196e-01 -6.98025525e-02 7.65846312e-01
1.98007792e-01 1.30449581e+00 -5.90991557e-01 -9.95703042e-02
1.36589244e-01 6.50296509e-01 -7.16489434e-01 1.28682315e-01
-2.46338144e-01 6.37722552e-01 -1.89138830e-01 3.81149501e-01
8.58231902e-01 -1.89913854e-01 8.66649114e-03 -4.33002710e-01
2.03944549e-01 1.05427533e-01 -1.31918037e+00 1.80747402e+00
-1.26565591e-01 2.62257785e-01 4.77974355e-01 -7.14607239e-01
1.16516042e+00 2.22368076e-01 9.02867913e-01 -1.75974280e-01
2.36013129e-01 3.72302264e-01 -1.35140121e-02 -8.75568241e-02
4.59740357e-03 -4.15303528e-01 5.22501394e-02 6.76245093e-01
-1.15504377e-02 -5.86196482e-01 -4.07583326e-01 -2.45685115e-01
9.47886825e-01 4.68941867e-01 6.17113933e-02 -5.20514309e-01
3.53527904e-01 -3.26423287e-01 2.83216923e-01 1.12838253e-01
1.81682378e-01 1.03957784e+00 1.84037134e-01 -8.29808593e-01
-1.46829593e+00 -1.29963613e+00 -2.46355072e-01 4.63866413e-01
3.98241729e-02 -1.95412144e-01 -8.17147493e-01 -4.94994253e-01
7.14640319e-01 1.33369222e-01 -8.51729512e-01 -5.72729334e-02
-1.05309045e+00 -2.18469322e-01 5.68363488e-01 8.66389632e-01
2.01372519e-01 -1.01332867e+00 -2.21732125e-01 2.31311157e-01
1.92567989e-01 -1.01534474e+00 -7.49929547e-01 -2.94179589e-01
-1.08032644e+00 -9.49994087e-01 -5.12929976e-01 -8.25862050e-01
9.40651178e-01 -3.76328588e-01 1.19766247e+00 5.27080774e-01
-2.74657369e-01 4.34048921e-01 2.00944245e-01 3.34856659e-02
-6.69004798e-01 8.94932449e-02 9.93551463e-02 1.40055120e-01
-2.80410469e-01 -8.57120693e-01 -5.79750896e-01 4.09472913e-01
-8.49240303e-01 -4.71014623e-03 1.82218969e-01 6.57468975e-01
8.27295244e-01 -5.70286155e-01 2.52917171e-01 -6.63051844e-01
1.56517312e-01 -1.33198813e-01 -5.78572452e-01 1.69104245e-02
-1.42043039e-01 3.79592031e-01 4.00921255e-01 -2.88791895e-01
-5.66671610e-01 4.79342192e-01 -2.87164658e-01 -5.42740166e-01
-7.69778937e-02 6.49481118e-02 -1.40603483e-01 -6.39012635e-01
8.03335547e-01 -1.70475617e-01 4.22239572e-01 -7.65060127e-01
5.57955801e-01 2.59963632e-01 8.28271270e-01 -9.88787711e-01
1.35415018e+00 8.53791595e-01 4.58562762e-01 -6.12234652e-01
-3.15994620e-01 3.50544341e-02 -1.39031208e+00 -4.08138940e-03
7.28437185e-01 -5.37108064e-01 -8.07775259e-01 6.14745080e-01
-1.24589610e+00 -5.36266506e-01 -4.52675819e-01 1.94022320e-02
-1.02322268e+00 3.14861178e-01 -6.59911275e-01 -2.26001531e-01
-5.40910542e-01 -1.05615115e+00 1.51334250e+00 -3.31509382e-01
-3.69977146e-01 -1.08900571e+00 1.97804749e-01 1.13554649e-01
5.32190502e-01 8.68225336e-01 1.01858854e+00 -4.51579452e-01
-6.42480075e-01 -3.79678667e-01 -5.70980795e-02 2.84601092e-01
3.04075092e-01 2.23429978e-01 -6.58722043e-01 -7.13672101e-01
-4.41407174e-01 -2.32039154e-01 6.31302834e-01 2.08281949e-01
1.02763128e+00 -3.19747031e-01 -4.49207366e-01 9.84407663e-01
1.20881236e+00 -2.25164175e-01 6.02329731e-01 1.06561951e-01
1.00430214e+00 2.73069143e-01 -2.54949033e-02 2.58744389e-01
3.99089992e-01 1.00071228e+00 6.54943585e-01 -3.01303923e-01
-4.89899069e-01 -1.85764149e-01 -2.32791640e-02 6.89582050e-01
-4.61423993e-01 3.34499687e-01 -1.19857955e+00 4.84975547e-01
-1.68452835e+00 -5.88676453e-01 6.04761988e-02 2.27958465e+00
8.63910854e-01 -5.33673391e-02 1.23812884e-01 -2.03890204e-01
4.77716357e-01 1.16443217e-01 -6.13672972e-01 -2.65414268e-01
-5.37717752e-02 6.41222179e-01 3.05019915e-01 8.63924980e-01
-1.28354192e+00 1.00025403e+00 6.94367552e+00 3.86425614e-01
-1.25380135e+00 -1.67132884e-01 3.11096489e-01 1.82981327e-01
-6.24317825e-01 -2.19062477e-01 -3.85301888e-01 6.78525269e-02
4.86382961e-01 -1.56634673e-01 6.78562999e-01 4.65157390e-01
-1.62084788e-01 6.40420914e-01 -1.52866936e+00 6.32268727e-01
1.13071986e-01 -1.59392965e+00 1.34185538e-01 1.04270510e-01
8.66230786e-01 1.46025822e-01 -5.66012971e-03 -1.29935071e-01
3.33055556e-01 -1.33350754e+00 7.17344463e-01 7.98243463e-01
1.19054365e+00 -4.79596585e-01 3.67266536e-01 2.96581268e-01
-1.38024449e+00 4.90063846e-01 -2.60271490e-01 1.88707247e-01
4.29699570e-02 1.73859283e-01 -8.90385687e-01 5.41637063e-01
5.69103897e-01 7.33416140e-01 -5.25389671e-01 1.04358554e+00
3.66148762e-02 1.19361289e-01 -6.64363384e-01 6.61860526e-01
-1.61867201e-01 -1.50828511e-01 8.14900339e-01 8.82334352e-01
2.95601487e-01 1.24415204e-01 2.63943970e-01 1.20273912e+00
-4.31572884e-01 -6.90131076e-03 -7.01886892e-01 2.49680907e-01
5.07948577e-01 1.27104378e+00 -6.20756090e-01 -9.18673128e-02
3.26165296e-02 8.40506375e-01 4.89446908e-01 2.35827431e-01
-4.14926767e-01 -2.36417865e-03 7.71741569e-01 5.87783694e-01
4.88560230e-01 -4.33076918e-01 -5.89340627e-01 -9.27622974e-01
2.12747946e-01 -6.86470091e-01 -1.18962824e-02 -6.71223998e-01
-1.42575049e+00 5.85141182e-01 -6.67546093e-02 -1.19187188e+00
-2.49318317e-01 -6.60198331e-01 -5.84910870e-01 9.79686677e-01
-8.38728309e-01 -1.65267229e+00 -2.03091055e-01 4.03708905e-01
-7.04565197e-02 -3.74366180e-04 8.85561764e-01 1.20941535e-01
1.30479066e-02 7.27018535e-01 -1.46638289e-01 2.74479896e-01
6.72722220e-01 -1.14632988e+00 1.20083749e+00 5.39371312e-01
1.67707607e-01 4.84342307e-01 4.54681695e-01 -7.27563858e-01
-1.51724231e+00 -1.11503375e+00 6.44932687e-01 -8.41520309e-01
4.48307991e-01 -4.04854804e-01 -1.04117203e+00 1.11854053e+00
-1.81727737e-01 5.40776193e-01 8.49851742e-02 -2.14717895e-01
-4.52630073e-01 1.40948407e-02 -1.46040320e+00 4.66932207e-01
1.24116313e+00 -3.32961559e-01 -6.47800624e-01 2.14946747e-01
5.11340261e-01 -1.06616998e+00 -1.33650362e+00 6.55338764e-01
8.26869488e-01 -4.63888019e-01 1.34346581e+00 -8.20239365e-01
2.67564565e-01 -2.96913505e-01 -1.95281014e-01 -1.44674194e+00
-5.68347871e-01 -6.95319176e-01 4.92706569e-03 7.88587332e-01
2.65373468e-01 -4.17759478e-01 9.01922941e-01 8.08577836e-01
-1.50092065e-01 -1.03645015e+00 -1.19054246e+00 -7.99633861e-01
8.12741339e-01 -8.84430408e-02 9.73399103e-01 9.17885125e-01
-3.02789420e-01 -7.11688623e-02 -1.38826460e-01 1.20475449e-01
6.66251123e-01 2.30602413e-01 9.46632028e-01 -1.59015977e+00
-7.28705972e-02 -5.41247070e-01 -7.36061811e-01 -1.06000245e+00
2.93922275e-01 -1.34947944e+00 4.95001711e-02 -1.43250465e+00
-7.20706061e-02 -6.19467974e-01 2.89846808e-01 8.54073644e-01
4.69743870e-02 3.93447936e-01 2.81415433e-01 1.34411961e-01
2.63907090e-02 4.48250026e-01 1.63897681e+00 -3.95044237e-02
-4.12299782e-02 -1.19965382e-01 -2.89348990e-01 9.04121161e-01
5.00202239e-01 -3.17196190e-01 2.96927720e-01 -7.26205885e-01
9.95630100e-02 9.92355421e-02 4.68812793e-01 -8.82291198e-01
1.45659477e-01 -4.16956358e-02 5.86077213e-01 -5.11453688e-01
4.56364095e-01 -8.70829403e-01 5.11171341e-01 3.89369756e-01
-2.15583473e-01 1.50859952e-01 3.89103174e-01 2.91693985e-01
3.53254050e-01 2.13292703e-01 1.03609979e+00 6.07313067e-02
5.55788651e-02 7.45522320e-01 2.55485862e-01 1.05097361e-01
7.64953554e-01 -2.40174159e-01 -2.10491672e-01 -3.31906199e-01
-1.09417367e+00 6.94279522e-02 1.00301027e+00 4.58684266e-01
6.21464252e-01 -1.89303970e+00 -1.04238951e+00 6.59718513e-01
-2.27624059e-01 3.18046987e-01 -1.63771793e-01 6.93935812e-01
-7.05462992e-01 -3.23111787e-02 -3.01180631e-01 -8.89838219e-01
-1.13367367e+00 1.91489100e-01 8.81594956e-01 -1.69929657e-02
-9.22690034e-01 8.07110846e-01 4.78007970e-03 -1.00541091e+00
-5.05841896e-03 -5.15653789e-01 2.73661256e-01 -3.78139675e-01
-3.78239481e-03 2.87975162e-01 4.26903903e-01 -1.17129970e+00
-5.29851377e-01 1.26815236e+00 2.88337052e-01 -1.21678397e-01
1.56950164e+00 3.03688407e-01 -4.04461145e-01 8.26356411e-02
1.41932154e+00 1.48601174e-01 -1.52436233e+00 -3.61726522e-01
-9.25986245e-02 -4.62442636e-01 -3.51505041e-01 -8.10262799e-01
-1.40237951e+00 7.17534065e-01 4.23957050e-01 -1.08648784e-01
5.28281331e-01 2.38504931e-01 1.03436482e+00 3.07963938e-01
4.90545034e-01 -5.29692411e-01 3.21724303e-02 6.87102854e-01
1.60743034e+00 -8.37472796e-01 3.68262567e-02 -4.69626874e-01
-9.72794071e-02 1.37764239e+00 3.33511353e-01 -7.17823863e-01
1.01308346e+00 5.94986260e-01 2.61388477e-02 -5.11301696e-01
-3.70484293e-01 2.35360473e-01 8.93586218e-01 5.64730942e-01
3.13263535e-01 3.01984698e-01 1.43853039e-01 4.44967896e-01
-5.92939675e-01 -8.45401287e-02 1.63761854e-01 7.40142822e-01
-3.69868368e-01 -1.26058531e+00 -4.52204973e-01 4.68300104e-01
-1.83717474e-01 2.71323919e-01 -6.10879779e-01 7.84026980e-01
-1.60380341e-02 1.96207613e-01 3.61172616e-01 -5.11732697e-01
7.38239050e-01 2.77691651e-02 8.99045169e-01 -7.06258535e-01
-7.40329921e-01 4.62946035e-02 -2.54630949e-02 -9.05892491e-01
-1.83983907e-01 -8.23043823e-01 -1.49482334e+00 -3.93386543e-01
1.06904618e-01 -4.28583294e-01 3.85881215e-01 8.48584771e-01
5.77588975e-01 6.23517521e-02 4.97166038e-01 -1.63265061e+00
-7.93994486e-01 -7.15276837e-01 -3.33951324e-01 8.93294573e-01
3.92674387e-01 -8.72939467e-01 -3.03871930e-01 2.16843858e-01] | [8.754236221313477, -3.558008909225464] |
a9118cd9-3ad5-495c-b65d-dd72dc655c20 | multi-task-learning-for-mental-health-using | 1712.03538 | null | http://arxiv.org/abs/1712.03538v1 | http://arxiv.org/pdf/1712.03538v1.pdf | Multi-Task Learning for Mental Health using Social Media Text | We introduce initial groundwork for estimating suicide risk and mental health
in a deep learning framework. By modeling multiple conditions, the system
learns to make predictions about suicide risk and mental health at a low false
positive rate. Conditions are modeled as tasks in a multi-task learning (MTL)
framework, with gender prediction as an additional auxiliary task. We
demonstrate the effectiveness of multi-task learning by comparison to a
well-tuned single-task baseline with the same number of parameters. Our best
MTL model predicts potential suicide attempt, as well as the presence of
atypical mental health, with AUC > 0.8. We also find additional large
improvements using multi-task learning on mental health tasks with limited
training data. | ['Dirk Hovy', 'Margaret Mitchell', 'Adrian Benton'] | 2017-12-10 | null | null | null | null | ['gender-prediction'] | ['computer-vision'] | [ 9.35357064e-02 4.86120433e-01 -5.32039404e-01 -6.64869189e-01
-1.40086102e+00 -1.66313145e-02 4.11261767e-01 5.57422876e-01
-6.93732738e-01 7.89239585e-01 4.71970022e-01 -1.16317468e-02
-5.01500852e-02 -5.04014075e-01 -2.48293608e-01 -2.60216475e-01
4.28281352e-02 8.99064183e-01 -4.01343048e-01 -4.06036824e-02
1.38027489e-01 -1.40651762e-01 -6.76456213e-01 7.58881688e-01
7.31050074e-01 7.59287000e-01 -4.48888719e-01 4.81872350e-01
4.18587059e-01 8.34078074e-01 -3.17422748e-01 -1.09507942e+00
-7.05936402e-02 -1.90908276e-02 -1.00293803e+00 -3.01837325e-01
2.81942725e-01 -7.21365213e-01 3.10391057e-02 3.54490161e-01
8.91665578e-01 -3.60705972e-01 1.08405268e+00 -1.48916173e+00
-3.81536454e-01 3.96997809e-01 -4.33764815e-01 -4.35793474e-02
5.35289347e-01 2.86118448e-01 8.76804233e-01 -8.10951948e-01
1.18852884e-01 1.55329180e+00 1.49316096e+00 1.03276193e+00
-1.52945471e+00 -9.41267014e-01 -1.63655862e-01 1.61202565e-01
-1.16355383e+00 -7.79637694e-01 1.04785845e-01 -4.71804291e-01
1.40849400e+00 -3.78697693e-01 4.15289253e-01 1.76919961e+00
6.28049850e-01 4.84896779e-01 8.39787364e-01 -9.93849058e-03
-5.11354432e-02 -8.68823100e-03 1.88471913e-01 9.29241836e-01
2.70756721e-01 -4.21947062e-01 -6.43064916e-01 -8.22613180e-01
4.70556885e-01 2.43148386e-01 6.26040399e-01 3.29714298e-01
-1.02092934e+00 9.13943827e-01 -2.21184805e-01 -1.62156358e-01
-5.15530586e-01 3.51376653e-01 7.25149512e-01 -1.41094411e-02
8.32976580e-01 1.09583236e-01 -7.55758703e-01 -1.73643440e-01
-1.01629806e+00 4.06101733e-01 7.46032536e-01 1.71677768e-01
5.43678820e-01 -6.40037805e-02 -6.90635085e-01 1.17501688e+00
1.52321994e-01 6.84758186e-01 4.58912164e-01 -1.05694449e+00
5.18113375e-01 6.58712089e-01 -1.11062028e-01 -4.68421459e-01
-9.72138822e-01 -3.21563870e-01 -8.99460077e-01 -1.15232788e-01
3.44301075e-01 -7.56614804e-01 -6.20116949e-01 1.90439582e+00
-7.87182152e-02 1.93978131e-01 -2.20372602e-02 -3.77431847e-02
4.86641675e-01 -1.60533690e-03 5.71879923e-01 -7.32835904e-02
1.47763765e+00 -5.89626074e-01 -6.44645452e-01 -7.24788427e-01
1.12954605e+00 -1.88598603e-01 7.56514192e-01 5.84625542e-01
-1.29641175e+00 5.74757047e-02 -2.79862612e-01 -2.77777523e-01
-3.56935114e-02 2.86508203e-01 8.64319563e-01 1.00056720e+00
-1.21553695e+00 4.35514301e-01 -7.31461048e-01 -4.31936294e-01
8.50101471e-01 2.68397480e-01 -4.16108072e-01 -1.85314924e-01
-1.23556733e+00 1.10075533e+00 9.07286704e-02 -4.04591590e-01
-1.32552207e+00 -9.94856298e-01 -8.88996363e-01 3.22023809e-01
1.74724355e-01 -1.10090685e+00 1.27778602e+00 -5.43287516e-01
-6.13414228e-01 1.20607328e+00 -4.05165881e-01 -6.11595392e-01
4.99350905e-01 -2.66958654e-01 -1.46917239e-01 2.03088343e-01
4.36729759e-01 7.82074749e-01 6.02585495e-01 -7.58871734e-01
-5.02267301e-01 -7.40157247e-01 -2.90445596e-01 3.04698437e-01
-7.00675726e-01 3.34394306e-01 1.47812754e-01 -1.97497785e-01
-4.83471036e-01 -5.57645619e-01 -2.64688432e-01 8.13884884e-02
-2.81452388e-01 -5.33238947e-01 1.05658360e-01 -7.92436898e-01
1.05444086e+00 -1.77680600e+00 -3.28273833e-01 -3.35211426e-01
5.47164202e-01 5.68109341e-02 -1.63259804e-01 2.01165617e-01
-1.81366205e-01 5.10252297e-01 -1.24005303e-01 -1.03384733e+00
-1.67979822e-01 -1.38253391e-01 2.79537797e-01 2.43498161e-01
5.66935003e-01 1.12214875e+00 -7.08418846e-01 -6.68980300e-01
-3.72848630e-01 4.66114402e-01 -8.76726151e-01 2.54298933e-02
3.42258871e-01 -2.92374697e-02 -2.64103770e-01 7.85185516e-01
2.92055160e-01 -2.45736629e-01 1.84397966e-01 2.41382524e-01
3.38939965e-01 4.69453126e-01 -5.61701834e-01 1.22501040e+00
-5.61304748e-01 -1.71344176e-01 3.00807416e-01 -1.01868129e+00
7.26948619e-01 7.32922375e-01 7.18526125e-01 -4.91185397e-01
-1.75461829e-01 -5.00728339e-02 1.32715702e-02 -6.20215952e-01
-1.27402678e-01 -7.95270085e-01 -3.69240165e-01 6.99143410e-01
2.47196838e-01 2.21826851e-01 -4.68781590e-01 2.75287032e-01
1.35896635e+00 -3.45860869e-01 4.82159942e-01 -1.29323334e-01
5.32888919e-02 -3.48159671e-01 9.94949341e-01 8.48828733e-01
-4.50727195e-01 2.84974784e-01 1.09729397e+00 -6.68033719e-01
-8.57112169e-01 -8.72784138e-01 -3.87861311e-01 1.35933065e+00
-9.08222198e-01 -5.62604368e-01 -5.53449929e-01 -8.45366001e-01
2.92211533e-01 7.78195798e-01 -1.05659187e+00 -5.85523009e-01
-7.89851770e-02 -1.49188483e+00 1.37030232e+00 5.06383657e-01
4.03890043e-01 -9.26664114e-01 -2.61943102e-01 -1.43531496e-02
-5.86144090e-01 -9.91516531e-01 -1.13613382e-01 2.41441697e-01
-9.86852229e-01 -1.09310746e+00 -8.26776624e-01 -4.88210261e-01
1.41027510e-01 -4.20616418e-01 1.34122455e+00 1.93451077e-01
-2.57106900e-01 3.77720237e-01 6.56212866e-02 -5.08114100e-01
-2.59783864e-01 2.26338446e-01 2.79166430e-01 -1.54746935e-01
5.66064119e-01 -4.41665500e-01 -6.54379189e-01 -7.98151046e-02
-4.08573210e-01 7.69279664e-03 5.55994451e-01 9.68769789e-01
9.95679013e-03 -4.28553313e-01 1.31128645e+00 -9.36490178e-01
7.85535395e-01 -1.01780593e+00 4.27771628e-01 2.79567122e-01
-9.59792256e-01 -3.51359062e-02 2.17385948e-01 -2.91807801e-01
-9.70271647e-01 -2.51597092e-02 -4.06674534e-01 2.35266387e-02
-5.02462566e-01 3.22013110e-01 -1.07427184e-02 4.14178312e-01
5.54119647e-01 9.93877426e-02 6.55162483e-02 -5.03464878e-01
-3.40491086e-01 5.48674405e-01 1.62456453e-01 -5.62532187e-01
1.04495049e-01 2.35257626e-01 3.11095357e-01 -4.51386034e-01
-1.06497490e+00 -3.00288349e-01 -6.29058301e-01 1.18012562e-01
1.08808839e+00 -1.21558952e+00 -1.06235039e+00 7.53776014e-01
-9.79169190e-01 -8.40016901e-01 3.45905900e-01 2.93898165e-01
-5.40466368e-01 2.25051910e-01 -8.18165183e-01 -1.24959028e+00
-8.61013353e-01 -7.95409203e-01 1.48703110e+00 -5.35976171e-01
-4.83628392e-01 -1.36581898e+00 1.62472382e-01 7.93144643e-01
1.47192016e-01 1.25433818e-01 1.29052651e+00 -1.01442766e+00
5.77048540e-01 -1.73573077e-01 -3.71998191e-01 3.51317450e-02
1.00591958e-01 -5.45359612e-01 -9.63992834e-01 -1.37821123e-01
-9.55652148e-02 -1.20271826e+00 1.36216760e+00 5.44221699e-01
1.09279764e+00 -3.06634843e-01 -3.91920090e-01 5.80945969e-01
9.62682724e-01 -1.77561402e-01 4.83966768e-01 5.93723468e-02
7.26774275e-01 6.13854110e-01 1.78633898e-01 8.50488424e-01
1.15050828e+00 5.89895487e-01 3.54150027e-01 -3.11173171e-01
1.75887898e-01 -7.70599246e-02 4.24411505e-01 -2.48006344e-01
-1.19013228e-02 7.05609620e-02 -1.63484108e+00 5.98280251e-01
-1.55185914e+00 -1.11741614e+00 -9.54344571e-02 2.22811532e+00
8.88841212e-01 1.54429495e-01 6.44096613e-01 2.52646878e-02
4.48071808e-01 -1.36179313e-01 -7.69326508e-01 -5.93184471e-01
-8.92253295e-02 1.53246567e-01 2.59218633e-01 4.85763431e-01
-1.17920995e+00 8.81737411e-01 7.97015619e+00 5.45825958e-01
-7.69461632e-01 6.04446650e-01 1.43780816e+00 -4.55537468e-01
-2.07643569e-01 -3.44435006e-01 -1.03274286e+00 2.61561483e-01
1.28759396e+00 -1.03617042e-01 2.16464013e-01 4.93345141e-01
3.81951332e-01 -8.59695449e-02 -1.07065058e+00 7.92433381e-01
3.38715404e-01 -5.41574001e-01 -6.64172620e-02 2.12848768e-01
4.71500278e-01 9.95203108e-02 2.41582394e-01 7.04344392e-01
3.38876665e-01 -1.40259492e+00 3.99081856e-01 6.68041170e-01
1.31474960e+00 -8.94601405e-01 7.74494946e-01 4.90137845e-01
-6.66884661e-01 -5.42539895e-01 -7.95318633e-02 -4.39512700e-01
5.00070900e-02 6.09453321e-01 -1.29071558e+00 8.61217827e-02
5.92143357e-01 8.71772707e-01 -6.31253004e-01 6.08744502e-01
6.23661429e-02 8.16336393e-01 -1.51369078e-02 3.36737126e-01
1.38335124e-01 3.70653719e-01 -8.12742934e-02 1.32774949e+00
2.14337856e-01 8.61816779e-02 1.77236930e-01 8.34753871e-01
-3.70627552e-01 9.45784226e-02 -5.71090460e-01 1.21893525e-01
1.33430585e-01 1.38143325e+00 -1.38964415e-01 -4.83658910e-01
-2.79884607e-01 9.53214407e-01 5.22622705e-01 -6.02355339e-02
-5.38394332e-01 3.24929327e-01 6.51407480e-01 2.66317815e-01
-3.67086053e-01 2.15692878e-01 -8.16481590e-01 -1.17278326e+00
-5.58057308e-01 -7.44899988e-01 7.66541302e-01 -4.52507913e-01
-1.45453560e+00 4.17801412e-03 -1.46287575e-01 -4.42606926e-01
-5.68757653e-01 -2.68749863e-01 -8.26785803e-01 9.85110164e-01
-1.35930812e+00 -1.37410522e+00 1.96306646e-01 6.86577380e-01
5.70022821e-01 -1.70135006e-01 1.19593334e+00 2.22200558e-01
-1.01575637e+00 1.08956325e+00 -1.87452495e-01 1.92719012e-01
1.36850965e+00 -1.34375298e+00 3.88363481e-01 1.05080836e-01
-8.80841911e-01 4.00223792e-01 7.11757839e-02 -1.14566672e+00
-9.08820450e-01 -1.27932060e+00 1.45641983e+00 -7.46688962e-01
2.64513522e-01 -5.85290670e-01 -8.12109351e-01 8.90133142e-01
-8.26712325e-02 -5.59053719e-01 1.28151345e+00 1.10599101e+00
-5.05385935e-01 3.28973718e-02 -1.65031719e+00 2.75121689e-01
8.38231325e-01 -5.58702350e-01 -3.90135020e-01 6.38711691e-01
2.71802872e-01 1.63861260e-01 -1.09082615e+00 3.28259587e-01
9.31817055e-01 -9.06876028e-01 1.24801278e+00 -9.84369278e-01
1.00125217e+00 1.15330648e+00 1.82795212e-01 -1.21180236e+00
-6.17294729e-01 -1.88823879e-01 8.47746357e-02 9.63952601e-01
6.75271034e-01 -6.05177164e-01 6.24768734e-01 1.12385237e+00
2.52232522e-01 -9.13034201e-01 -9.11516190e-01 -3.22466701e-01
7.63462067e-01 -4.36558843e-01 2.62932181e-01 1.11850166e+00
3.83743376e-01 5.88156462e-01 -6.94931030e-01 -3.65051553e-02
7.86574602e-01 -6.04969978e-01 1.88324302e-01 -1.54503572e+00
-1.15614824e-01 -3.12282026e-01 3.10555268e-02 -9.96512640e-03
6.24068975e-01 -9.07409787e-01 -4.32247102e-01 -1.75274324e+00
1.17141652e+00 -1.80540770e-01 -4.62528169e-01 1.42353547e+00
-7.45188951e-01 3.42366606e-01 -1.12450510e-01 -5.41209653e-02
-7.46592641e-01 5.26058614e-01 6.25306308e-01 4.25472409e-02
8.26431215e-02 1.78072438e-01 -1.07265639e+00 8.28247905e-01
1.04011703e+00 -7.72533178e-01 -1.71799392e-01 -4.82293546e-01
4.53609675e-01 6.13016367e-01 4.64348793e-01 -8.01391363e-01
-1.77061230e-01 -1.87123477e-01 8.52175176e-01 -2.33874053e-01
8.12795460e-01 3.05982754e-02 -4.01631117e-01 7.72296429e-01
-7.47432232e-01 1.38767928e-01 2.45486096e-01 1.50567591e-01
5.10613203e-01 -1.49022579e-01 6.03614569e-01 -3.61102164e-01
1.16329089e-01 2.97696471e-01 -6.07134879e-01 1.71104610e-01
5.38229644e-01 1.23483390e-01 -3.07093531e-01 -6.91968083e-01
-8.79250646e-01 4.45217550e-01 1.33593351e-01 2.14715794e-04
6.25902712e-01 -1.06441712e+00 -1.11756957e+00 4.80183288e-02
4.42470424e-02 -6.52018845e-01 4.92045075e-01 1.02418268e+00
2.46993229e-01 3.16689849e-01 -4.29808885e-01 -5.92947938e-02
-1.32235312e+00 3.24222714e-01 5.41197658e-01 -6.42155588e-01
-1.48664176e-01 5.91671348e-01 2.08778396e-01 -5.31420529e-01
1.35158479e-01 9.57463086e-02 -3.51035237e-01 2.45636702e-01
6.70182467e-01 6.46545708e-01 9.62938145e-02 -4.33765620e-01
-4.64754224e-01 4.59369607e-02 -2.32311994e-01 -3.06379735e-01
1.56438291e+00 8.97174925e-02 -2.40340158e-02 4.51752871e-01
9.53362763e-01 -4.27891582e-01 -9.96464491e-01 9.18873847e-02
-2.98570991e-02 1.59157049e-02 2.09958658e-01 -1.26772714e+00
-6.56223774e-01 1.06200647e+00 4.39842105e-01 -2.48175278e-01
9.48387146e-01 -9.73225310e-02 5.92998624e-01 5.15132070e-01
1.24747448e-01 -9.37176585e-01 3.93761039e-01 6.24283016e-01
5.27911663e-01 -1.52590966e+00 -9.28849131e-02 8.73889253e-02
-1.15407264e+00 7.90202618e-01 7.59472609e-01 1.01611391e-01
4.49744105e-01 4.51725215e-01 -9.21470970e-02 -1.79477289e-01
-1.41263068e+00 2.01033391e-02 3.01248357e-02 7.37495065e-01
6.73464537e-01 1.01738714e-01 -3.05407852e-01 1.24257851e+00
9.34318081e-03 1.88163117e-01 3.94017339e-01 3.55900079e-01
-2.81704247e-01 -1.05326402e+00 -3.37098986e-01 1.17356825e+00
-1.22308493e+00 -5.55832207e-01 -7.21429706e-01 1.64001912e-01
2.98437417e-01 1.05247009e+00 1.81768551e-01 -3.35487247e-01
-6.97714416e-03 7.20693827e-01 2.64368743e-01 -8.27647507e-01
-1.08967316e+00 -3.25845219e-02 4.94412541e-01 -7.62933731e-01
8.89784992e-02 -9.83872116e-01 -9.71709251e-01 -1.32786661e-01
3.58434021e-01 -2.90811390e-01 3.06514412e-01 1.14808273e+00
4.12191451e-01 3.77783597e-01 4.00283217e-01 -4.93252337e-01
-6.24027789e-01 -1.28209507e+00 -5.62066495e-01 3.84058774e-01
3.33841264e-01 -5.26491761e-01 -1.29065111e-01 -1.96851909e-01] | [8.741467475891113, 10.212217330932617] |
f8606d56-7560-4c85-9e9c-f583a2fe705d | stacked-capsule-autoencoders | 1906.06818 | null | https://arxiv.org/abs/1906.06818v2 | https://arxiv.org/pdf/1906.06818v2.pdf | Stacked Capsule Autoencoders | Objects are composed of a set of geometrically organized parts. We introduce an unsupervised capsule autoencoder (SCAE), which explicitly uses geometric relationships between parts to reason about objects. Since these relationships do not depend on the viewpoint, our model is robust to viewpoint changes. SCAE consists of two stages. In the first stage, the model predicts presences and poses of part templates directly from the image and tries to reconstruct the image by appropriately arranging the templates. In the second stage, SCAE predicts parameters of a few object capsules, which are then used to reconstruct part poses. Inference in this model is amortized and performed by off-the-shelf neural encoders, unlike in previous capsule networks. We find that object capsule presences are highly informative of the object class, which leads to state-of-the-art results for unsupervised classification on SVHN (55%) and MNIST (98.7%). The code is available at https://github.com/google-research/google-research/tree/master/stacked_capsule_autoencoders | ['Geoffrey E. Hinton', 'Sara Sabour', 'Adam R. Kosiorek', 'Yee Whye Teh'] | 2019-06-17 | stacked-capsule-autoencoders-1 | http://papers.nips.cc/paper/9684-stacked-capsule-autoencoders | http://papers.nips.cc/paper/9684-stacked-capsule-autoencoders.pdf | neurips-2019-12 | ['unsupervised-mnist'] | ['methodology'] | [-3.39213848e-01 2.18270570e-01 2.14409772e-02 -3.62592727e-01
-4.55309391e-01 -6.45783961e-01 5.64009786e-01 -5.10663018e-02
-4.72851694e-02 1.79726124e-01 4.25014317e-01 3.25471401e-01
-2.39700414e-02 -7.31760144e-01 -1.16160345e+00 -7.61642337e-01
-6.45760521e-02 8.78399551e-01 2.15335965e-01 1.12503864e-01
-1.27902582e-01 5.46405494e-01 -1.40830517e+00 3.82758498e-01
1.65578142e-01 1.04132855e+00 5.02618074e-01 6.59308851e-01
2.93043613e-01 9.32183087e-01 -3.90104741e-01 -4.46449995e-01
2.26136804e-01 -1.11084066e-01 -7.52749681e-01 6.53357148e-01
3.97907048e-01 -5.40714741e-01 -6.50717378e-01 9.69061136e-01
1.54834837e-01 -1.62027657e-01 7.85710573e-01 -8.43616962e-01
-5.97936869e-01 6.53333962e-01 -2.69908062e-03 -1.24733821e-01
1.01234399e-01 -4.80312519e-02 1.07114983e+00 -1.02546275e+00
8.06973875e-01 1.05148339e+00 6.54448628e-01 5.22765517e-01
-1.02847481e+00 -2.40309656e-01 2.20827401e-01 -7.27152079e-02
-1.42760372e+00 -4.52165127e-01 9.36152220e-01 -7.67229795e-01
8.24590683e-01 6.05984293e-02 1.06591892e+00 1.02723980e+00
5.44835687e-01 9.27215576e-01 6.44014895e-01 -1.74525976e-01
3.39311212e-01 2.17118710e-01 1.06138671e-02 8.64698708e-01
4.34883773e-01 -3.65955420e-02 -3.99669230e-01 4.92334403e-02
1.16382158e+00 4.21809018e-01 -2.67561495e-01 -6.86323225e-01
-1.40941238e+00 8.60272467e-01 8.22346747e-01 1.65767267e-01
-5.17377794e-01 2.09141970e-01 -1.75603654e-03 -8.03846568e-02
2.54260540e-01 5.34861982e-01 -5.65360725e-01 3.82646531e-01
-6.91031873e-01 1.70637667e-01 1.11446476e+00 1.09771156e+00
4.44530636e-01 -8.97058249e-02 3.52155954e-01 6.62370801e-01
7.95804381e-01 2.69290090e-01 5.39011002e-01 -1.05883038e+00
-6.43335655e-02 4.92809445e-01 -2.72859365e-01 -6.65670156e-01
-3.15176249e-01 -8.15854371e-01 -5.55627048e-01 1.73010901e-01
1.60281613e-01 7.45033622e-02 -9.93988514e-01 1.30679846e+00
3.30810726e-01 -6.87991679e-02 -1.59316406e-01 8.87818754e-01
1.19823217e+00 1.64198712e-01 -3.73627990e-01 3.45268011e-01
1.57833302e+00 -1.22908819e+00 -5.47912180e-01 -4.54434156e-01
1.93816632e-01 -8.40399027e-01 2.88035214e-01 4.92348790e-01
-1.14496422e+00 -5.42136908e-01 -1.15076160e+00 -2.92947233e-01
-3.88853699e-01 3.15128148e-01 9.50158954e-01 2.33603373e-01
-1.10322094e+00 6.02818787e-01 -1.38381100e+00 -1.27726421e-01
6.18190348e-01 6.94002867e-01 -5.06497860e-01 3.15389745e-02
-4.10431385e-01 7.36417115e-01 4.42698240e-01 2.18992710e-01
-1.34607625e+00 -4.80614215e-01 -1.12493384e+00 3.69273722e-02
2.94804037e-01 -8.85743856e-01 1.44221604e+00 -9.27670062e-01
-1.38997447e+00 8.21767986e-01 -6.87498972e-02 -3.65101039e-01
3.61493111e-01 -1.31618962e-01 -1.67259321e-01 3.20081711e-01
-8.86322260e-02 1.13416636e+00 8.95689607e-01 -1.58226073e+00
-1.55622900e-01 -3.00205350e-01 1.30883172e-01 1.29145548e-01
8.04474056e-02 -2.38955930e-01 -9.31431472e-01 -7.56056964e-01
4.78303492e-01 -1.12966418e+00 -6.86399341e-02 4.48877901e-01
-6.33067846e-01 -6.42682090e-02 5.02113402e-01 -8.38984132e-01
5.65660238e-01 -2.14493012e+00 3.21231574e-01 -8.69750977e-03
4.86377418e-01 -3.73104453e-01 4.94642630e-02 2.10494429e-01
-2.32366532e-01 -3.17923546e-01 -1.45056531e-01 -7.43839502e-01
-5.37696332e-02 3.38368654e-01 -3.02412212e-02 6.93492174e-01
7.41426498e-02 7.69459486e-01 -5.35433769e-01 -4.19623733e-01
3.66430670e-01 7.50133455e-01 -8.69729459e-01 3.28248829e-01
-4.36455816e-01 4.01531637e-01 -3.64401847e-01 9.27838683e-01
5.88613570e-01 -7.47345388e-01 2.22474650e-01 -5.19258440e-01
8.94244537e-02 5.33579051e-01 -1.06334448e+00 1.69867706e+00
-1.10845797e-01 5.81547379e-01 1.53826490e-01 -7.54607499e-01
4.20967877e-01 3.61111641e-01 6.78058803e-01 -1.75842687e-01
4.62447315e-01 8.80379900e-02 2.12960541e-01 -3.50665271e-01
1.20855801e-01 1.25537843e-01 1.15374804e-01 1.55282676e-01
5.68076193e-01 -2.20335945e-01 1.30592465e-01 9.47550684e-02
7.94995546e-01 2.76634872e-01 3.65024626e-01 -1.79826587e-01
1.23149373e-01 -7.79632758e-03 3.74075025e-01 4.19378936e-01
1.51614919e-01 8.31915200e-01 2.32182518e-01 -6.33981943e-01
-9.97519016e-01 -1.20303333e+00 -3.37417096e-01 4.77056950e-01
1.28775612e-01 -4.84022021e-01 -8.39498222e-01 -4.53352809e-01
1.39989570e-01 3.82927150e-01 -8.65798831e-01 -6.59464719e-03
-4.23574984e-01 -4.38480854e-01 -1.11676320e-01 9.12890375e-01
2.06480682e-01 -1.07859516e+00 -4.92724717e-01 1.38667598e-01
6.33948520e-02 -1.18623495e+00 -4.04422402e-01 4.55670178e-01
-1.07383394e+00 -1.26570046e+00 -4.64679331e-01 -9.73665655e-01
1.18875849e+00 1.84868544e-01 1.17445421e+00 2.16202930e-01
-2.42321476e-01 4.48978156e-01 -1.28560662e-01 -4.93622482e-01
-2.00234190e-01 -1.81910232e-01 1.37168363e-01 -1.97794810e-02
6.81057647e-02 -5.84452331e-01 -7.19766021e-01 2.88198888e-01
-8.27406824e-01 3.47662009e-02 8.48466516e-01 6.82877541e-01
1.11627817e+00 6.35995669e-03 -3.17979693e-01 -6.62966728e-01
-1.64111152e-01 -3.99057746e-01 -7.79571533e-01 -4.84616272e-02
-5.27245738e-03 2.06175357e-01 5.35238504e-01 -6.06011271e-01
-7.03208864e-01 5.20367444e-01 -1.53254256e-01 -8.21168065e-01
-2.79277593e-01 4.02529657e-01 -1.84075132e-01 -1.12957440e-01
2.98675805e-01 3.04776669e-01 1.39373854e-01 -8.41876984e-01
7.16495067e-02 2.92057842e-01 6.19127989e-01 -3.35348904e-01
7.26384699e-01 8.54821324e-01 -2.98799515e-01 -6.77935362e-01
-9.20774758e-01 -3.76047581e-01 -8.92362356e-01 -2.12359890e-01
1.07620609e+00 -1.27972054e+00 -8.51343811e-01 4.65455413e-01
-1.03261173e+00 -5.17545283e-01 -2.12587580e-01 7.50160158e-01
-5.16088426e-01 9.48481783e-02 -9.35263395e-01 -2.57749736e-01
-2.43464679e-01 -1.41466868e+00 1.22926450e+00 1.55408904e-01
-9.57848281e-02 -9.83056486e-01 -1.55763492e-01 5.34154892e-01
-1.49787311e-02 4.97939214e-02 4.57535684e-01 -7.58149862e-01
-1.13009274e+00 -2.77291596e-01 2.15661898e-01 3.16712856e-01
7.82245547e-02 -1.07892945e-01 -9.40636277e-01 -4.70877618e-01
2.49471307e-01 -1.55956835e-01 7.48263896e-01 7.26119637e-01
1.41580188e+00 -4.66999084e-01 -4.86913055e-01 8.31815064e-01
1.38249469e+00 1.20708779e-01 4.74909872e-01 1.63342759e-01
7.29526162e-01 2.74962544e-01 3.26930918e-02 3.12110186e-01
4.85900670e-01 6.49481893e-01 8.36736441e-01 1.30738884e-01
-3.55167359e-01 -1.89737245e-01 2.90084690e-01 1.08156192e+00
-7.71616101e-02 -3.32501344e-02 -7.22582996e-01 5.05717337e-01
-1.50796008e+00 -7.32048750e-01 1.07736057e-02 1.94299603e+00
6.23589337e-01 2.46798009e-01 -9.41274017e-02 -3.19464326e-01
2.61132121e-01 5.84193096e-02 -5.66860914e-01 3.03985119e-01
1.17581658e-01 2.93523017e-02 4.95504856e-01 5.85378647e-01
-1.41040587e+00 7.06965089e-01 6.14751101e+00 2.50451028e-01
-1.08248663e+00 1.49917483e-01 3.68078738e-01 -1.46902934e-01
-9.55137983e-02 -6.46148156e-03 -8.22949290e-01 4.96660233e-01
6.68264985e-01 3.53132784e-01 6.70351148e-01 1.18137670e+00
-3.61653149e-01 -3.73943374e-02 -1.42167425e+00 8.02971840e-01
3.79791319e-01 -1.46525931e+00 -1.27790317e-01 2.90011495e-01
6.97223842e-01 4.70707297e-01 -3.12865935e-02 8.70337114e-02
3.55096042e-01 -7.80940235e-01 1.20312691e+00 6.64784968e-01
2.67373890e-01 -3.31274718e-01 7.09544659e-01 2.19933003e-01
-1.17162216e+00 1.50215020e-02 -4.84610260e-01 2.29709238e-01
-8.63280222e-02 5.63341141e-01 -9.05290961e-01 3.53528023e-01
8.36429894e-01 9.42508519e-01 -6.53440416e-01 1.16287243e+00
-3.73787135e-01 3.31011295e-01 -5.50895512e-01 1.92525446e-01
-1.55654952e-01 -4.88627760e-04 4.93496478e-01 9.06551898e-01
4.78438698e-02 -4.50275838e-03 3.10668260e-01 9.92570579e-01
-3.06092769e-01 -4.55051899e-01 -3.13780308e-01 1.16504030e-03
2.26409301e-01 1.33157742e+00 -8.91718745e-01 -3.75383854e-01
-5.35952330e-01 9.76004779e-01 2.89541006e-01 2.02459365e-01
-7.96449780e-01 1.47515833e-01 6.20515704e-01 3.03770751e-02
9.59112704e-01 -3.91070276e-01 -1.81217879e-01 -1.39446008e+00
2.03759253e-01 -8.72888863e-01 2.88225234e-01 -9.45852935e-01
-1.16246116e+00 6.75309241e-01 -7.18769580e-02 -1.51973152e+00
-1.07274406e-01 -1.18270671e+00 -3.55494380e-01 1.60885677e-01
-1.02238703e+00 -1.26490557e+00 -3.53992283e-01 4.59900498e-01
7.82560885e-01 1.87995788e-02 1.04247797e+00 1.30948186e-01
-4.16444600e-01 1.50327489e-01 -1.24673024e-02 6.12514555e-01
3.83995026e-01 -1.22156835e+00 1.03726380e-01 5.58412433e-01
5.08234024e-01 8.64129543e-01 6.55232012e-01 -7.16452479e-01
-1.56405604e+00 -9.67569172e-01 6.14188075e-01 -5.72950721e-01
6.09935999e-01 -6.79873168e-01 -6.04618490e-01 1.21636176e+00
2.50357509e-01 4.87543344e-01 7.75566220e-01 4.66745421e-02
-4.47571397e-01 -3.65894125e-03 -7.62893558e-01 4.50926363e-01
8.90169084e-01 -5.59908748e-01 -6.32264197e-01 7.84270227e-01
5.60294926e-01 -8.53708506e-01 -1.17624462e+00 2.10455999e-01
7.10782349e-01 -8.62447560e-01 1.15073586e+00 -3.10583472e-01
6.28311098e-01 -3.32979530e-01 -2.75092483e-01 -1.09652317e+00
-5.02883554e-01 -3.88172418e-02 -6.21335804e-01 4.89930451e-01
4.31893796e-01 -4.45886731e-01 7.44051635e-01 4.52521980e-01
-4.80801105e-01 -1.01859093e+00 -5.42819679e-01 -7.17673063e-01
-2.20567495e-01 -1.68973163e-01 3.94635558e-01 8.45319510e-01
-2.82714427e-01 1.12391062e-01 3.35527547e-02 5.67167521e-01
7.04234183e-01 2.53027231e-01 5.14569640e-01 -1.20783317e+00
-8.21582735e-01 -3.38173956e-01 -5.40412724e-01 -1.10860944e+00
4.70478907e-02 -9.11812007e-01 6.79829568e-02 -1.61684132e+00
4.02489245e-01 -1.25740690e-03 -1.09761924e-01 7.12530911e-01
2.67827392e-01 2.43670687e-01 1.06546417e-01 4.62889463e-01
-6.96391761e-01 5.12337387e-01 1.21557534e+00 -3.13173890e-01
4.67727818e-02 5.63462488e-02 -5.44619679e-01 1.05566978e+00
7.03520656e-01 -5.06628096e-01 -9.14524347e-02 -7.43085027e-01
-1.20047525e-01 1.84759852e-02 5.86524546e-01 -1.24890995e+00
4.50031072e-01 3.06373924e-01 1.01572776e+00 -7.64622271e-01
8.17986131e-01 -1.01005197e+00 4.72293019e-01 5.86093068e-01
-7.91940913e-02 -2.59376522e-02 7.29616955e-02 4.56069261e-01
-2.00550362e-01 -2.84523726e-01 6.63300753e-01 -4.73980844e-01
-3.61256242e-01 4.69918430e-01 -3.38708401e-01 -4.05292094e-01
7.99353421e-01 -1.50835454e-01 -6.88014627e-02 -3.88503969e-01
-1.07789552e+00 -6.92359608e-05 5.74234664e-01 4.24450368e-01
5.43996155e-01 -1.25743163e+00 -4.34510976e-01 2.71592557e-01
5.91210164e-02 5.37759900e-01 2.42746815e-01 8.58230531e-01
-6.87974811e-01 4.48274851e-01 -8.91665816e-02 -7.62174964e-01
-1.09597790e+00 7.24453688e-01 5.79242647e-01 -9.66708660e-02
-7.87096381e-01 1.06312001e+00 3.47143203e-01 -3.69914442e-01
3.00764978e-01 -5.41032255e-01 3.75500582e-02 -1.08691692e-01
3.49868476e-01 -1.95073783e-01 1.33117601e-01 -7.13214874e-01
-4.51154709e-01 5.18576443e-01 -3.28604013e-01 5.81266694e-02
1.66188860e+00 4.43287343e-01 -2.54952788e-01 2.63427764e-01
1.34404552e+00 1.42114311e-01 -1.54132736e+00 -1.92831919e-01
-3.25369745e-01 -1.40156269e-01 1.76201239e-01 -7.06760228e-01
-1.50611031e+00 5.34057438e-01 2.95306355e-01 -1.69317767e-01
8.26351166e-01 5.56792080e-01 4.16271001e-01 4.76309329e-01
4.12562251e-01 -6.78145885e-01 2.73946732e-01 4.35148984e-01
1.14113033e+00 -1.07278764e+00 2.82255918e-01 -6.27587497e-01
-4.97917086e-01 1.03264964e+00 5.11344969e-01 -4.46848631e-01
9.79435623e-01 5.59251547e-01 -1.31780908e-01 -6.12029672e-01
-7.29384780e-01 -2.02790797e-02 5.38651168e-01 2.41741940e-01
3.95527154e-01 3.08866978e-01 3.91684353e-01 5.88419080e-01
-5.55675805e-01 -2.93741941e-01 2.95163900e-01 9.00400579e-01
-2.00818285e-01 -9.63517129e-01 -3.23507071e-01 5.24560213e-01
-4.15787548e-01 6.59924606e-03 -3.82916182e-01 8.29893827e-01
2.58078843e-01 4.69824165e-01 3.59059244e-01 -3.02764356e-01
6.69034570e-02 -1.41584858e-01 7.89547205e-01 -6.99077010e-01
-4.38078582e-01 4.90330458e-01 -4.32180874e-02 -6.72688484e-01
-3.70088220e-01 -9.45402265e-01 -1.19937205e+00 2.14240730e-01
-3.41259569e-01 -1.16544731e-01 7.18831182e-01 8.46016049e-01
3.30972761e-01 7.74245083e-01 2.47102723e-01 -1.24003196e+00
-3.58228892e-01 -9.40659761e-01 -4.29683834e-01 3.87587696e-01
4.15738016e-01 -8.09823513e-01 -2.31209412e-01 3.44286472e-01] | [8.14825439453125, -2.9956214427948] |
923eb631-17bf-4637-9dcd-c9feac494ab1 | on-the-use-of-bert-for-automated-essay | 2205.03835 | null | https://arxiv.org/abs/2205.03835v2 | https://arxiv.org/pdf/2205.03835v2.pdf | On the Use of BERT for Automated Essay Scoring: Joint Learning of Multi-Scale Essay Representation | In recent years, pre-trained models have become dominant in most natural language processing (NLP) tasks. However, in the area of Automated Essay Scoring (AES), pre-trained models such as BERT have not been properly used to outperform other deep learning models such as LSTM. In this paper, we introduce a novel multi-scale essay representation for BERT that can be jointly learned. We also employ multiple losses and transfer learning from out-of-domain essays to further improve the performance. Experiment results show that our approach derives much benefit from joint learning of multi-scale essay representation and obtains almost the state-of-the-art result among all deep learning models in the ASAP task. Our multi-scale essay representation also generalizes well to CommonLit Readability Prize data set, which suggests that the novel text representation proposed in this paper may be a new and effective choice for long-text tasks. | ['Hui Lin', 'Ruobing Li', 'Chuan Wang', 'Yongjie Wang'] | 2022-05-08 | null | https://aclanthology.org/2022.naacl-main.249 | https://aclanthology.org/2022.naacl-main.249.pdf | naacl-2022-7 | ['automated-essay-scoring'] | ['natural-language-processing'] | [-1.57731637e-01 -1.97141960e-01 -1.97968230e-01 -5.84811449e-01
-9.63274598e-01 -3.70281458e-01 4.55801576e-01 4.59752500e-01
-6.77543163e-01 9.57739115e-01 2.83425122e-01 -3.38370919e-01
-1.52460665e-01 -7.12657094e-01 -4.90392178e-01 -2.41600543e-01
5.51135778e-01 6.11066878e-01 4.22480628e-02 -5.01679599e-01
5.06245732e-01 1.05356328e-01 -9.90087986e-01 1.85251608e-01
1.25172842e+00 9.61743534e-01 2.53824502e-01 5.06482482e-01
-5.75979412e-01 9.91599023e-01 -9.88077581e-01 -8.34973812e-01
1.29100978e-01 -3.28759938e-01 -8.43959868e-01 -4.06595767e-01
5.33166587e-01 -5.64006209e-01 -5.53112149e-01 8.24532390e-01
5.32553911e-01 3.77033174e-01 7.80149341e-01 -7.46591508e-01
-1.09041417e+00 6.55963182e-01 -7.65164375e-01 2.00448021e-01
2.28767633e-01 -1.90418735e-01 1.64522815e+00 -9.42077160e-01
3.04331779e-01 1.03596067e+00 7.36077964e-01 3.25535208e-01
-8.52242529e-01 -7.45480001e-01 2.49058202e-01 4.95780647e-01
-7.87289381e-01 -6.31216839e-02 9.97239530e-01 -2.60609966e-02
8.84908259e-01 -6.46648882e-03 4.07958657e-01 1.19435883e+00
4.14240599e-01 1.64565599e+00 1.20904398e+00 -6.60721958e-01
-2.47210905e-01 -4.66312580e-02 6.66963339e-01 9.58767176e-01
1.79180160e-01 -3.41686696e-01 -7.48824239e-01 -1.66933954e-01
6.08035803e-01 4.81328294e-02 -1.15249351e-01 2.58183349e-02
-1.00574565e+00 1.15853560e+00 4.03867275e-01 2.99281448e-01
-2.74507344e-01 2.23600581e-01 5.76770604e-01 5.71698964e-01
8.52945685e-01 8.13931227e-01 -5.44665098e-01 -3.99219245e-01
-1.29089963e+00 3.87220174e-01 8.44966888e-01 6.11364007e-01
4.35278326e-01 1.03926092e-01 -6.60065770e-01 1.29894483e+00
-8.28520954e-02 1.76556095e-01 8.88804972e-01 -6.99047863e-01
7.61551857e-01 7.11160898e-01 6.72625601e-02 -9.66803074e-01
-4.96534824e-01 -6.76035821e-01 -8.07392478e-01 -1.17078468e-01
5.52675426e-01 -1.41504854e-01 -6.11959755e-01 1.51536775e+00
-4.54838216e-01 -8.23401809e-02 -9.95280296e-02 8.28840435e-01
1.02314126e+00 7.46075988e-01 -3.85225825e-02 1.15228146e-01
1.25539052e+00 -1.44547665e+00 -7.21877813e-01 -3.49386871e-01
5.92816532e-01 -7.07332850e-01 1.41753924e+00 5.06347477e-01
-1.28017163e+00 -5.67571998e-01 -1.12116277e+00 -4.34341192e-01
-2.96674073e-01 5.29422522e-01 8.41409624e-01 4.96662974e-01
-8.74126971e-01 7.10511863e-01 -3.67173046e-01 -2.47246802e-01
5.36915481e-01 2.56280422e-01 1.55954920e-02 -5.13286181e-02
-1.46919370e+00 1.07951856e+00 1.33293331e-01 3.76120606e-03
-4.92904812e-01 -4.86008346e-01 -5.78616142e-01 5.01907945e-01
1.36912972e-01 -6.42914116e-01 1.60841382e+00 -9.08315003e-01
-2.04711246e+00 9.05124187e-01 -4.26959693e-02 -6.09330535e-01
3.31569463e-01 -4.19608861e-01 -7.10957646e-02 1.36479894e-02
2.52884179e-02 4.05354172e-01 6.34690166e-01 -5.93473196e-01
-4.46731448e-01 -2.77146399e-01 2.83433020e-01 3.21272284e-01
-1.06725907e+00 9.48650315e-02 -1.28118888e-01 -8.40917230e-01
-3.84538621e-01 -6.88451529e-01 -3.59514803e-02 -7.44849816e-02
-1.71108246e-01 -8.99240911e-01 3.34417790e-01 -9.12954092e-01
1.45314455e+00 -1.64158642e+00 3.81781831e-02 -2.06541538e-01
1.42247558e-01 4.86827344e-01 -4.16316450e-01 3.55143666e-01
3.81834716e-01 -1.04898229e-01 -7.92851597e-02 -8.00036430e-01
2.90665299e-01 -9.76523459e-02 -4.27736133e-01 1.29845440e-01
6.88425153e-02 1.24299562e+00 -8.99404585e-01 -4.30371284e-01
-2.22663283e-01 -1.24352574e-01 -1.77818090e-01 3.27098757e-01
-2.83092171e-01 2.62268540e-02 -6.65687323e-01 5.12354136e-01
4.46459353e-01 -5.12938440e-01 -7.13112876e-02 5.22387922e-01
2.23932892e-01 7.30034947e-01 -4.58634228e-01 2.09409332e+00
-5.87108195e-01 8.75021040e-01 -1.78545460e-01 -1.10221279e+00
1.23624992e+00 1.65239379e-01 1.11848757e-01 -9.60978866e-01
1.12361819e-01 3.56150270e-01 5.64333983e-02 -2.40695074e-01
1.00238299e+00 -1.61699638e-01 -2.53240526e-01 9.18486178e-01
1.84185013e-01 -1.93755120e-01 2.38780871e-01 3.31081539e-01
1.02282858e+00 -5.18077202e-02 3.99357766e-01 -2.62034357e-01
6.60733879e-01 -1.57346413e-01 3.31278056e-01 9.37529206e-01
-3.21424097e-01 5.52271605e-01 5.19475222e-01 -3.19322795e-01
-9.53644454e-01 -8.32595229e-01 -1.03332140e-01 1.53351259e+00
-1.61793023e-01 -1.61972314e-01 -5.08711874e-01 -9.95682240e-01
1.65293217e-01 5.61825752e-01 -4.88531172e-01 -1.48567811e-01
-5.71830690e-01 -8.15275848e-01 7.91913867e-01 7.57818878e-01
7.10665643e-01 -1.19555354e+00 -2.87076741e-01 5.46672702e-01
-2.95554817e-01 -8.06484461e-01 -6.89364970e-01 1.91779092e-01
-1.00505519e+00 -7.68887341e-01 -1.24886072e+00 -8.40816557e-01
2.49369621e-01 2.58320987e-01 1.26565850e+00 1.59755722e-01
1.57608241e-02 2.28803709e-01 -4.40739036e-01 -6.51275158e-01
-9.80836079e-02 6.23007119e-01 -1.74594223e-01 -1.04146205e-01
6.99420810e-01 -1.83644786e-01 -4.58075851e-01 -1.06219672e-01
-5.14850080e-01 -1.11895494e-01 6.87577128e-01 1.32309365e+00
9.04622823e-02 -3.09444338e-01 1.23766291e+00 -8.83769870e-01
1.51961923e+00 -2.88274258e-01 -1.92913979e-01 4.78277266e-01
-7.88492441e-01 1.05150379e-01 1.01982880e+00 -3.83070827e-01
-1.27604616e+00 -5.89339912e-01 -1.63926840e-01 -4.06804048e-02
1.86165586e-01 1.02569008e+00 2.49166220e-01 -6.94180802e-02
4.66439098e-01 2.35497251e-01 -3.57919149e-02 -6.26457632e-01
4.30955626e-02 8.57359767e-01 2.39821717e-01 -7.80198514e-01
4.09215271e-01 2.76283547e-02 -1.56163901e-01 -3.80921811e-01
-1.49878383e+00 -6.50187910e-01 -4.42121208e-01 1.22331895e-01
4.30099398e-01 -9.15515423e-01 -6.25089645e-01 6.95921779e-01
-1.28571904e+00 -3.28659385e-01 -1.09990433e-01 3.70604157e-01
-3.98405999e-01 5.61114848e-01 -1.23589659e+00 -6.23632967e-01
-8.10239255e-01 -9.35875654e-01 8.94199371e-01 5.20696282e-01
-2.48222411e-01 -1.24407291e+00 2.63470411e-01 6.48570061e-01
4.98109043e-01 -4.33354676e-01 9.73548770e-01 -9.14354444e-01
-2.49791801e-01 -4.00947422e-01 -5.11992276e-01 5.58703482e-01
-2.21152753e-01 -3.53475720e-01 -9.37289059e-01 -2.64031649e-01
-1.05488539e-01 -1.07447219e+00 1.57115638e+00 4.13695186e-01
1.55924559e+00 -1.06179833e-01 1.25643119e-01 3.56387854e-01
9.92705226e-01 -2.14700282e-01 4.60987836e-01 6.63771570e-01
4.87822622e-01 5.47557831e-01 6.61526740e-01 5.07926106e-01
6.58492804e-01 5.32491744e-01 -4.67362488e-03 -1.12553090e-02
3.29183117e-02 -1.59231499e-01 5.35050750e-01 1.12415695e+00
-5.39338738e-02 -4.13935006e-01 -8.78385842e-01 5.33059180e-01
-2.01048374e+00 -8.46517980e-01 -1.02175087e-01 1.76250517e+00
1.12915313e+00 1.44692451e-01 1.05316110e-01 1.06698247e-02
4.39659983e-01 4.96869087e-01 -5.03080904e-01 -8.35052073e-01
-3.23411703e-01 7.20969796e-01 2.53774315e-01 2.27704436e-01
-1.12932730e+00 9.70757961e-01 6.23490095e+00 1.12417328e+00
-8.74230206e-01 3.35342795e-01 6.31249070e-01 -1.56890638e-02
-1.82681769e-01 -3.55193853e-01 -9.32005703e-01 4.36510295e-01
9.42803502e-01 -4.12519932e-01 4.27419841e-02 8.15748274e-01
-7.73389414e-02 7.49641582e-02 -8.75447154e-01 8.27336252e-01
4.08186644e-01 -1.16229296e+00 5.67413270e-02 -9.10873115e-02
9.38393891e-01 -1.02410324e-01 3.25986087e-01 9.67814863e-01
4.15549785e-01 -1.15758586e+00 5.12679756e-01 4.37027335e-01
6.60456002e-01 -7.90642083e-01 1.01595354e+00 4.94861066e-01
-7.58117735e-01 -2.69257903e-01 -9.23577130e-01 -5.14809370e-01
-7.43592829e-02 6.95018888e-01 -7.01521039e-01 5.63422859e-01
1.93146750e-01 1.01936376e+00 -7.59612978e-01 1.11261559e+00
-4.49547797e-01 8.69683266e-01 1.08083531e-01 -7.13826537e-01
5.92576444e-01 -1.92331165e-01 2.09659785e-01 1.23829627e+00
3.31715643e-01 -7.58475289e-02 1.57904223e-01 8.19338083e-01
-7.96571612e-01 3.88766527e-01 -2.29552761e-01 -2.31031373e-01
1.52178034e-01 1.29134214e+00 -2.40670726e-01 -5.06357551e-01
-5.90079010e-01 9.87861812e-01 8.96906912e-01 3.62498730e-01
-6.64728940e-01 -6.52668655e-01 3.09841245e-01 -2.28782207e-01
8.88624042e-02 -1.21137671e-01 -7.63490379e-01 -1.43886495e+00
5.12277754e-03 -7.03047156e-01 4.11824673e-01 -5.71093559e-01
-1.83717489e+00 3.39750081e-01 -5.16383111e-01 -1.02525008e+00
-2.10008085e-01 -8.37091148e-01 -9.53725696e-01 9.67473328e-01
-2.03166986e+00 -1.12532377e+00 -1.34429753e-01 4.21407372e-01
8.45900416e-01 -6.60991967e-01 9.17789578e-01 1.17190517e-01
-5.35639942e-01 1.01413286e+00 7.30937600e-01 3.28305066e-01
1.18807733e+00 -1.49757886e+00 2.75772065e-01 4.94715333e-01
1.80435181e-01 6.14858925e-01 2.84475267e-01 -2.91970044e-01
-8.86743724e-01 -7.96861589e-01 1.17626739e+00 -4.36414123e-01
8.96993220e-01 -2.72883445e-01 -8.99951518e-01 6.09027326e-01
4.78077084e-01 -4.29571480e-01 8.25341225e-01 6.22710526e-01
-4.57461715e-01 -2.37478048e-01 -8.11316073e-01 4.75216210e-01
5.40414214e-01 -6.56216800e-01 -8.73622537e-01 4.03912842e-01
4.92580056e-01 -3.76646847e-01 -6.58239663e-01 2.07307503e-01
6.24389708e-01 -6.37938380e-01 8.57102692e-01 -7.71180511e-01
1.08333409e+00 5.95546305e-01 2.61689633e-01 -1.65510094e+00
-5.48792005e-01 -2.70247817e-01 -4.39956248e-01 1.11328971e+00
3.49693298e-01 -5.62800288e-01 1.04654300e+00 4.86739546e-01
-2.56261885e-01 -1.07572997e+00 -8.21319401e-01 -8.28635216e-01
7.40980983e-01 2.78175548e-02 4.99164402e-01 9.86115992e-01
3.16174805e-01 6.45372689e-01 -4.80699927e-01 -5.99106312e-01
5.52002370e-01 4.95758384e-01 5.31841397e-01 -1.66442418e+00
-2.85018831e-01 -1.05528998e+00 4.10928279e-02 -1.57448375e+00
8.14757645e-01 -1.06242847e+00 -2.67016143e-02 -1.76857364e+00
6.46686435e-01 -3.74964833e-01 -7.46175826e-01 3.67783338e-01
-6.44733548e-01 1.46805853e-01 2.03404516e-01 2.09617332e-01
-8.08272660e-01 9.67795730e-01 1.55502439e+00 -4.67931330e-01
1.22992247e-01 4.76060398e-02 -9.60036635e-01 4.80051875e-01
9.68253732e-01 -4.34524387e-01 -1.42819330e-01 -6.25786006e-01
2.95168310e-01 9.16323215e-02 6.09089509e-02 -5.39781153e-01
3.94121706e-01 -9.28205103e-02 4.72784042e-01 -4.79752481e-01
3.84872794e-01 -3.40851903e-01 -1.04135501e+00 1.85546324e-01
-6.62537694e-01 7.62443466e-04 1.29041791e-01 5.46225250e-01
-4.96077091e-01 -7.84480453e-01 5.13147056e-01 -3.63942772e-01
-4.06673849e-01 3.09923619e-01 -2.36280844e-01 2.19758093e-01
5.41248560e-01 -1.57281850e-02 -5.87095022e-01 -5.69181919e-01
-1.61826715e-01 3.89362514e-01 1.36092350e-01 4.50497717e-01
6.33167624e-01 -1.16158473e+00 -1.00656962e+00 -1.42357364e-01
3.10266875e-02 -2.19751716e-01 2.46438950e-01 6.79121256e-01
-4.65201586e-01 9.99652326e-01 -2.56038278e-01 -1.33358240e-01
-9.78359461e-01 -6.79495512e-03 -1.50825465e-02 -1.23588955e+00
-4.50980335e-01 1.15741158e+00 2.05167264e-01 -7.77493715e-01
2.98890829e-01 -1.57013342e-01 -3.85939896e-01 8.68517011e-02
5.56289017e-01 3.11635077e-01 2.60837287e-01 -5.81256747e-02
5.73262312e-02 1.72266662e-01 -6.54006362e-01 -4.77321539e-03
1.52249420e+00 4.44004647e-02 -9.89730433e-02 6.33391023e-01
9.85152900e-01 -8.10479671e-02 -1.00885952e+00 -4.14243400e-01
1.95506588e-01 -3.75419259e-01 1.61379933e-01 -1.04274130e+00
-8.73005509e-01 1.31667614e+00 -1.19610548e-01 -1.81803808e-01
8.69557917e-01 -4.77899551e-01 1.29454899e+00 8.03984463e-01
3.53837907e-01 -1.48364556e+00 6.80730581e-01 1.05726039e+00
8.99159133e-01 -1.33999336e+00 1.52687952e-02 2.83863395e-01
-7.33723223e-01 1.41464329e+00 7.50784099e-01 -3.48866433e-01
3.58127296e-01 -7.57951438e-02 -1.85223743e-01 -6.86765760e-02
-8.31114829e-01 1.55691341e-01 4.39752996e-01 1.61286741e-01
8.20034146e-01 1.08693823e-01 -6.44781649e-01 1.18242848e+00
-2.11958051e-01 1.03385419e-01 9.49189544e-01 6.41538918e-01
-5.62233150e-01 -1.27508759e+00 -1.32258192e-01 9.54286635e-01
-6.73321784e-01 -3.79542381e-01 -5.84016681e-01 5.22746801e-01
-6.01815403e-01 7.68600881e-01 -1.96573719e-01 -9.00873169e-02
1.38574108e-01 4.04980570e-01 7.02370882e-01 -8.04073215e-01
-9.70761120e-01 -4.69689488e-01 1.57491058e-01 -1.41690914e-02
-1.86907351e-01 -4.67387885e-01 -1.02162266e+00 -6.36863053e-01
-5.65471470e-01 1.66603923e-01 5.22068083e-01 9.65688944e-01
1.42114341e-01 6.48307264e-01 2.78277844e-01 -5.63927412e-01
-1.19696760e+00 -1.40465081e+00 -9.45499182e-01 2.82629490e-01
1.91182792e-01 -6.53543234e-01 -1.67756826e-01 -3.61539304e-01] | [11.329475402832031, 9.348170280456543] |
5a742783-284e-4d04-83c7-3b3f616f31db | deep-cardiosound-an-ensembled-deep-learning | 2204.07420 | null | https://arxiv.org/abs/2204.07420v2 | https://arxiv.org/pdf/2204.07420v2.pdf | Deep CardioSound-An Ensembled Deep Learning Model for Heart Sound MultiLabelling | Heart sound diagnosis and classification play an essential role in detecting cardiovascular disorders, especially when the remote diagnosis becomes standard clinical practice. Most of the current work is designed for single category based heard sound classification tasks. To further extend the landscape of the automatic heart sound diagnosis landscape, this work proposes a deep multilabel learning model that can automatically annotate heart sound recordings with labels from different label groups, including murmur's timing, pitch, grading, quality, and shape. Our experiment results show that the proposed method has achieved outstanding performance on the holdout data for the multi-labelling task with sensitivity=0.990, specificity=0.999, F1=0.990 at the segments level, and an overall accuracy=0.969 at the patient's recording level. | ['Yonghong Peng', 'Steven Davenport', 'Li Guo'] | 2022-04-15 | null | null | null | null | ['sound-classification'] | ['audio'] | [-4.33394015e-02 2.45066181e-01 -4.91996408e-02 -3.10942322e-01
-1.17694032e+00 -5.69809914e-01 -1.03527941e-01 4.50489938e-01
-1.48774371e-01 5.08147895e-01 6.00558743e-02 -4.24021691e-01
-1.82419971e-01 -4.63744044e-01 1.06853060e-01 -6.56086326e-01
-2.15002507e-01 3.33505452e-01 2.80100316e-01 3.27291518e-01
-4.59534712e-02 2.22688541e-01 -1.00738239e+00 2.70714134e-01
4.98779148e-01 1.23821688e+00 -2.23781064e-01 1.22874808e+00
3.41402531e-01 1.04576182e+00 -8.30510616e-01 -2.63566434e-01
-1.66727930e-01 -7.53524065e-01 -7.76942313e-01 -3.30714464e-01
4.17752206e-01 -2.57088274e-01 1.67636961e-01 8.19923520e-01
1.15133095e+00 -1.59718305e-01 6.48539245e-01 -9.39640760e-01
-2.70027727e-01 6.67434573e-01 -3.17210644e-01 4.12982374e-01
1.04808301e-01 -2.61880338e-01 1.17504525e+00 -5.01401067e-01
1.47924334e-01 5.95926166e-01 1.24017584e+00 2.20991388e-01
-8.02355945e-01 -7.00144947e-01 -5.17172992e-01 1.18684098e-01
-1.49138916e+00 -1.49601921e-01 6.73432827e-01 -6.48647070e-01
4.82034177e-01 4.13986683e-01 5.22004366e-01 4.76641566e-01
2.30530024e-01 1.34748995e-01 1.23777056e+00 -6.26499414e-01
9.72928852e-02 1.51618138e-01 2.30936900e-01 8.38047266e-01
4.42708172e-02 -1.02789961e-01 -2.43040055e-01 -2.86910951e-01
7.54114628e-01 -1.80973366e-01 -5.07217795e-02 3.04990083e-01
-1.16293502e+00 8.98260236e-01 1.76724628e-01 6.80723011e-01
-1.65410578e-01 -4.65460084e-02 6.11977458e-01 2.02110767e-01
6.07570767e-01 5.33581316e-01 -4.62412804e-01 -1.66526511e-01
-1.02135360e+00 3.46776433e-02 8.10070753e-01 3.65382373e-01
1.80920158e-02 -9.22302715e-03 -5.23080945e-01 1.23519289e+00
2.30832934e-01 6.62712574e-01 4.00989205e-01 -1.12795758e+00
-1.45941377e-01 1.06181607e-01 -1.10572249e-01 -1.12589622e+00
-1.07151723e+00 -9.02683198e-01 -9.43931282e-01 -2.84951478e-01
2.06190228e-01 -4.20688719e-01 -4.73033160e-01 1.52246070e+00
3.40765178e-01 4.96162683e-01 6.32752180e-02 8.64228606e-01
1.38274896e+00 4.76179332e-01 4.26571608e-01 -4.41653639e-01
1.64040136e+00 -8.69976819e-01 -8.47584128e-01 2.34393820e-01
7.45443642e-01 -7.98509002e-01 6.57557011e-01 3.97938281e-01
-6.03822470e-01 -8.20545912e-01 -7.67095447e-01 4.03790087e-01
1.58038989e-01 5.52889049e-01 2.59040415e-01 8.85977089e-01
-8.43693972e-01 3.46795201e-01 -6.02372468e-01 -9.69609097e-02
4.24930811e-01 5.69150001e-02 -1.98794566e-02 2.05779329e-01
-1.50786090e+00 6.36689425e-01 9.49989557e-02 5.29852808e-02
-6.93870664e-01 -7.34360099e-01 -4.88703132e-01 -1.45670855e-02
-2.14589760e-01 -5.93224049e-01 1.41002631e+00 -3.48667979e-01
-1.43615472e+00 1.05823016e+00 4.21718091e-01 -1.83119819e-01
4.69705194e-01 1.33162349e-01 -8.57128739e-01 4.52216446e-01
2.61101961e-01 3.48148704e-01 5.90779126e-01 -8.19000065e-01
-1.04442286e+00 -2.43882716e-01 -1.80082664e-01 -8.96203220e-02
-2.48249710e-01 3.26653808e-01 1.88938137e-02 -6.65955067e-01
1.99391380e-01 -1.12912643e+00 -9.98737440e-02 -1.05153993e-01
-2.99710691e-01 -3.78596634e-01 2.25559041e-01 -1.03158128e+00
1.47345102e+00 -2.32171845e+00 -4.10821706e-01 1.92407236e-01
6.38129056e-01 5.27649283e-01 4.02123690e-01 -7.27452636e-02
-1.01633118e-02 2.28948951e-01 -1.46091893e-01 -2.21528318e-02
-7.55265430e-02 -8.98158252e-02 7.43000209e-02 2.98814476e-01
-2.41820678e-01 6.69533134e-01 -8.55605721e-01 -7.82980442e-01
2.42732882e-01 3.43583375e-01 -4.13740546e-01 4.39788997e-01
5.83556890e-01 5.18552601e-01 -3.46664578e-01 2.73533851e-01
3.21343631e-01 -2.08264217e-01 1.11356497e-01 -2.33291298e-01
-2.91604679e-02 -3.26772705e-02 -1.17149711e+00 1.45277238e+00
-5.43011665e-01 5.31674862e-01 -1.31062523e-01 -8.41011047e-01
1.12421072e+00 9.49011743e-01 8.12498689e-01 -3.68775249e-01
6.45019412e-02 4.16433930e-01 4.03228760e-01 -9.14270103e-01
-2.42669016e-01 -5.50864995e-01 -2.91312397e-01 3.88937205e-01
1.04624465e-01 -1.06742412e-01 -1.72145516e-01 -6.83593079e-02
1.02497804e+00 -4.53325242e-01 4.48707253e-01 -3.81964803e-01
4.70096648e-01 -2.32810885e-01 5.10351598e-01 9.36051190e-01
-6.14100993e-01 8.44812632e-01 4.37413305e-01 -5.06501257e-01
-5.90310395e-01 -8.41418445e-01 -7.31052160e-01 1.15967190e+00
-2.64346719e-01 -2.63597876e-01 -6.56020105e-01 -6.94747090e-01
-2.44373202e-01 1.88998580e-01 -3.49471807e-01 -1.44607276e-01
-4.55381930e-01 -7.68585861e-01 1.13730311e+00 7.44437814e-01
3.98820341e-01 -1.27516520e+00 -8.71796787e-01 2.91718602e-01
-6.47148132e-01 -1.12036645e+00 -2.19541416e-01 2.08104730e-01
-5.12480736e-01 -9.59453046e-01 -1.00861812e+00 -1.19548202e+00
8.53622854e-02 -3.24590355e-01 1.12395597e+00 -1.98519789e-02
-6.54674768e-01 1.03766963e-01 -5.75216711e-01 -5.72599351e-01
-5.12391925e-01 1.07232668e-01 -1.53699130e-01 4.23292629e-03
-2.30581403e-01 -3.73544216e-01 -7.39614666e-01 1.87442139e-01
-2.91902065e-01 -2.06300125e-01 2.95341730e-01 6.37504876e-01
5.13040602e-01 1.13238886e-01 1.15858030e+00 -1.01722229e+00
7.75440812e-01 -2.02683628e-01 2.26507261e-02 2.79783849e-02
-7.27544188e-01 -8.30385447e-01 4.76977348e-01 -2.57410586e-01
-5.98670125e-01 -3.79344784e-02 -6.60005748e-01 -4.53069173e-02
-5.57810366e-01 4.45382774e-01 2.99225479e-01 8.02740082e-02
8.02552521e-01 -6.74921051e-02 -1.26530051e-01 -3.12554181e-01
-3.04817446e-02 1.07316959e+00 7.79395640e-01 -2.37870634e-01
9.70609337e-02 -9.60080884e-03 1.02050312e-01 -7.37448335e-01
-1.20965517e+00 -7.76410162e-01 -5.90052545e-01 -4.43242580e-01
1.25184584e+00 -8.46123993e-01 -5.37429452e-01 6.39287055e-01
-8.99728298e-01 -1.10188238e-01 -1.02750860e-01 7.01364994e-01
-3.08545381e-01 3.93624634e-01 -8.51370931e-01 -9.54894066e-01
-8.55789661e-01 -7.73998022e-01 9.46211815e-01 5.08715212e-02
-6.02593780e-01 -1.13527966e+00 4.39519435e-01 4.84165907e-01
5.23148954e-01 4.78496164e-01 9.33259010e-01 -1.00424850e+00
3.54459971e-01 -5.19777954e-01 -1.44232213e-01 4.68299687e-01
3.35785389e-01 -1.93035603e-01 -1.19296336e+00 -5.80897145e-02
-8.06498080e-02 -3.04540396e-01 6.75569773e-01 7.21847892e-01
1.05521023e+00 2.39601955e-02 -3.24396305e-02 3.29846472e-01
9.60036874e-01 3.95223320e-01 1.72269210e-01 -1.09246515e-01
8.61325085e-01 3.87585074e-01 4.53735918e-01 4.65962350e-01
4.30663794e-01 4.84950453e-01 5.46244085e-02 -5.64128995e-01
-3.47687840e-01 6.45342320e-02 -1.83690041e-01 1.08862174e+00
-4.94365208e-02 -1.05645441e-01 -1.15422702e+00 4.76796389e-01
-1.56870639e+00 -7.71712780e-01 -4.32404399e-01 1.81797111e+00
7.99353719e-01 8.73133019e-02 3.37392747e-01 5.62517345e-01
9.59265649e-01 -4.12522666e-02 -1.18818626e-01 -4.47065532e-01
2.54942060e-01 3.49282861e-01 1.00035243e-01 4.22305733e-01
-1.59501362e+00 3.97124678e-01 6.78597021e+00 5.26109159e-01
-1.12996316e+00 2.70935684e-01 7.55665123e-01 5.30224264e-01
3.44417840e-01 -4.50384825e-01 -4.83507037e-01 5.30998886e-01
1.10093760e+00 2.99493194e-01 -1.33651808e-01 5.69989502e-01
2.56808728e-01 2.28956252e-01 -6.67639017e-01 1.19006467e+00
2.46179953e-01 -1.01481330e+00 -5.12929082e-01 -2.90568918e-01
5.37007451e-01 -1.70343906e-01 -7.08093643e-02 2.71030843e-01
-3.23238850e-01 -8.14467609e-01 4.65135127e-01 5.39494574e-01
1.29082036e+00 -5.60589254e-01 1.13592374e+00 3.26772571e-01
-1.14816451e+00 -1.41874909e-01 1.28872901e-01 7.42098968e-03
1.45177960e-01 8.24437201e-01 -1.10293663e+00 4.09500390e-01
7.45123565e-01 6.47827268e-01 -4.38300818e-01 1.21465707e+00
-2.49146029e-01 1.10032713e+00 -1.52051762e-01 2.14342847e-01
-1.74267814e-01 3.38298619e-01 2.09924728e-01 1.57409310e+00
2.63137430e-01 2.67903417e-01 6.49964988e-01 1.61784470e-01
1.88112676e-01 5.67955196e-01 -1.71869367e-01 3.70022207e-01
3.91115338e-01 1.44023752e+00 -1.01813710e+00 -4.31393385e-01
-1.04146734e-01 2.65996665e-01 -3.13081771e-01 -2.29890615e-01
-1.07951939e+00 -7.63421118e-01 -1.18143603e-01 1.90476477e-02
-5.28819114e-02 4.63038087e-01 -5.26390910e-01 -4.57159638e-01
-3.09380710e-01 -7.19423532e-01 7.65994489e-01 -6.08055294e-01
-1.23511255e+00 7.88935781e-01 -3.48439634e-01 -1.27623141e+00
-2.99685359e-01 -3.04891348e-01 -6.92465127e-01 7.35113263e-01
-1.34342599e+00 -1.07799864e+00 -3.31526697e-01 2.90045291e-01
2.62380451e-01 -1.66019499e-01 1.37016034e+00 7.84684598e-01
-4.07703131e-01 8.60475779e-01 -3.01199228e-01 5.30822754e-01
9.86589015e-01 -1.38564646e+00 -1.15075044e-01 3.11625034e-01
1.89167261e-01 9.47536379e-02 4.01177019e-01 -1.25319034e-01
-1.96310461e-01 -1.19242024e+00 1.20982516e+00 -4.15903628e-01
4.79982078e-01 3.05666924e-01 -7.35922396e-01 -6.17092215e-02
-2.38114625e-01 3.88022870e-01 1.21394062e+00 1.98860288e-01
-2.23952591e-01 -3.93004984e-01 -1.16525447e+00 -2.68920064e-01
4.77612048e-01 -6.92494035e-01 -3.82253677e-01 4.97784376e-01
4.13353801e-01 -5.77446520e-01 -1.48762703e+00 6.61707759e-01
9.69612002e-01 -6.33989036e-01 8.22848558e-01 -3.79170746e-01
2.87908524e-01 -1.97496027e-01 -4.21707705e-02 -1.13055563e+00
-6.94854259e-01 -3.52950722e-01 2.74641305e-01 1.29736650e+00
5.80579400e-01 -4.94564205e-01 2.02854082e-01 4.21817228e-02
-4.16271836e-01 -5.87680936e-01 -8.85847509e-01 -2.74862289e-01
4.33057845e-02 -5.71238935e-01 9.95319262e-02 1.11833405e+00
1.03691645e-01 5.71955383e-01 -5.75507462e-01 1.97836176e-01
4.48480546e-01 3.27178478e-01 1.05073743e-01 -1.80172181e+00
-3.84763986e-01 -3.15877348e-01 -5.57406425e-01 -1.89993903e-01
3.33879469e-03 -1.01895690e+00 1.29779398e-01 -1.53242290e+00
4.65879560e-01 -7.33866751e-01 -8.70309353e-01 7.27517784e-01
-2.71023303e-01 8.83479059e-01 4.98354286e-02 3.47262263e-01
-7.53250897e-01 -2.76790917e-01 1.26943755e+00 1.99201759e-02
-1.89201549e-01 6.14639938e-01 -7.06169784e-01 8.69867027e-01
9.13255394e-01 -6.72933340e-01 -4.25195307e-01 -5.65601327e-02
1.68436987e-03 6.19087219e-01 2.69687712e-01 -1.16172147e+00
-1.34680018e-01 2.29527697e-01 2.34955512e-02 -1.36702999e-01
-2.97408346e-02 -5.17641544e-01 1.15503103e-01 6.59837306e-01
-9.27867174e-01 -2.14332417e-01 -1.42909866e-02 2.85046488e-01
-1.81706622e-01 -1.09319814e-01 1.01726377e+00 -4.83966060e-02
-1.72436178e-01 2.05524087e-01 -4.02013332e-01 5.03256083e-01
7.33092487e-01 1.90642089e-01 -1.42926022e-01 -2.34840482e-01
-1.37323153e+00 -8.88873730e-03 -5.27315140e-01 4.09380972e-01
2.42408842e-01 -1.15060222e+00 -1.10259783e+00 1.30692013e-02
3.06185961e-01 -2.56036222e-01 4.30740535e-01 1.10522425e+00
-5.73692024e-01 4.30167317e-01 -1.26228169e-01 -9.22664106e-01
-1.44768548e+00 -3.75835150e-02 5.73004603e-01 -2.94814467e-01
-7.57367849e-01 8.94037008e-01 -3.74206975e-02 -3.20682019e-01
3.63427788e-01 -4.89852816e-01 -6.36985004e-01 2.35414013e-01
4.40463036e-01 6.43455982e-01 3.19855869e-01 -6.96611643e-01
-4.61403161e-01 7.88239956e-01 3.26607138e-01 5.12627512e-02
9.41950500e-01 7.49375746e-02 -1.30551443e-01 6.79988921e-01
1.02151859e+00 -1.55297101e-01 -3.84267598e-01 4.85347994e-02
-3.49469185e-01 3.87175791e-02 4.37580198e-01 -1.21094215e+00
-1.09410071e+00 1.11827660e+00 1.26136446e+00 6.42725527e-01
9.94080365e-01 9.93911400e-02 8.43387187e-01 4.69852053e-02
-9.31421015e-03 -8.85276914e-01 -8.32966790e-02 1.65537596e-01
4.55369651e-01 -1.14140594e+00 -3.02199125e-01 -3.54056627e-01
-7.90634692e-01 9.64410305e-01 8.75385776e-02 1.42231718e-01
1.12894344e+00 2.67617822e-01 6.15666091e-01 -9.54695717e-02
-3.39655221e-01 -1.36848809e-02 2.93246955e-01 5.82308292e-01
7.65481710e-01 6.58054471e-01 -4.35113966e-01 9.84082639e-01
-2.46159047e-01 4.66616303e-02 2.64634401e-01 6.16077304e-01
-7.14126885e-01 -6.59579217e-01 -1.19244635e-01 5.73464394e-01
-1.14928901e+00 1.16856515e-01 -6.24148585e-02 1.24313898e-01
4.65842932e-01 1.24638486e+00 -4.47835252e-02 -3.25699598e-01
3.99785489e-01 3.05349290e-01 1.09762348e-01 -7.53624439e-01
-7.58266330e-01 4.45585281e-01 1.77517742e-01 1.67237979e-03
-5.60997128e-01 -4.50299680e-01 -1.48120093e+00 3.26967597e-01
-4.07539845e-01 4.88286048e-01 2.96504736e-01 9.10728753e-01
2.11142957e-01 9.18228149e-01 8.92672241e-01 -1.18011430e-01
-4.82949257e-01 -1.04539430e+00 -8.34870517e-01 3.31132263e-01
7.36968517e-01 -4.28845525e-01 -4.51207012e-01 3.41278404e-01] | [14.360359191894531, 3.3690760135650635] |
31584e1c-edbc-4682-ad53-fc6f2aef51c4 | don-t-do-it-safer-reinforcement-learning-with | 2212.13819 | null | https://arxiv.org/abs/2212.13819v1 | https://arxiv.org/pdf/2212.13819v1.pdf | Don't do it: Safer Reinforcement Learning With Rule-based Guidance | During training, reinforcement learning systems interact with the world without considering the safety of their actions. When deployed into the real world, such systems can be dangerous and cause harm to their surroundings. Often, dangerous situations can be mitigated by defining a set of rules that the system should not violate under any conditions. For example, in robot navigation, one safety rule would be to avoid colliding with surrounding objects and people. In this work, we define safety rules in terms of the relationships between the agent and objects and use them to prevent reinforcement learning systems from performing potentially harmful actions. We propose a new safe epsilon-greedy algorithm that uses safety rules to override agents' actions if they are considered to be unsafe. In our experiments, we show that a safe epsilon-greedy policy significantly increases the safety of the agent during training, improves the learning efficiency resulting in much faster convergence, and achieves better performance than the base model. | ['Jochen Renz', 'Cheng Xue', 'Ekaterina Nikonova'] | 2022-12-28 | null | null | null | null | ['robot-navigation'] | ['robots'] | [ 1.27022276e-02 4.78945255e-01 1.06981479e-01 -2.31556460e-01
2.84055471e-01 -5.50100744e-01 6.35069847e-01 1.38209224e-01
-8.94885361e-01 1.19526088e+00 -2.33913735e-01 -4.20415252e-01
-3.32185239e-01 -1.13613188e+00 -7.65455067e-01 -8.89338493e-01
-4.59604114e-01 4.56301600e-01 5.61464429e-01 -4.71358895e-01
2.54688472e-01 6.46125078e-01 -1.77703011e+00 -1.78425312e-01
9.61766899e-01 5.56752622e-01 8.28708112e-02 5.82013428e-01
4.61187810e-01 1.25247359e+00 -8.22857857e-01 2.56127745e-01
3.76081198e-01 -4.05900866e-01 -9.00407255e-01 -1.02623381e-01
-1.49397865e-01 -6.33840680e-01 1.67452097e-02 1.19789088e+00
1.47423863e-01 4.86950725e-01 5.15887499e-01 -1.52283704e+00
8.31785575e-02 5.35528779e-01 -2.52500802e-01 -1.48209095e-01
3.56840491e-01 4.41601664e-01 3.11450154e-01 6.68984726e-02
3.25234830e-01 1.24223959e+00 3.01567405e-01 9.08490479e-01
-1.07676470e+00 -7.44164765e-01 6.29538596e-01 1.28952011e-01
-1.15503693e+00 -1.75863847e-01 3.29815418e-01 -1.47462413e-01
1.12553561e+00 4.41196442e-01 6.90467834e-01 9.74772274e-01
6.76605403e-01 5.24577260e-01 8.77398968e-01 -4.07122523e-01
8.57810438e-01 6.77040359e-03 -2.39256367e-01 5.23987234e-01
5.86479962e-01 6.42925322e-01 -9.62329730e-02 -3.11700583e-01
4.98813987e-01 -2.51153439e-01 -2.42217053e-02 -7.38459766e-01
-8.47314060e-01 7.25382626e-01 4.23579067e-01 1.28547862e-01
-4.37546521e-01 3.40753466e-01 3.26450974e-01 4.49511856e-01
-2.13023201e-01 8.20266545e-01 -3.47547293e-01 -1.12971343e-01
3.59175391e-02 4.46197659e-01 7.74128616e-01 5.63690007e-01
5.27879894e-01 1.16359673e-01 3.98804337e-01 5.55389822e-01
2.82781333e-01 4.55912262e-01 7.17562437e-02 -1.14728439e+00
1.23148590e-01 5.74540913e-01 6.41202748e-01 -7.75369704e-01
-5.34539580e-01 -3.45905386e-02 -4.21146661e-01 1.06764472e+00
2.74967790e-01 -5.11391699e-01 -7.80381203e-01 1.90016234e+00
5.91322243e-01 9.13355872e-02 4.45276499e-01 7.95983553e-01
-8.11394453e-02 6.24119461e-01 3.86248112e-01 -3.52190167e-01
6.58372641e-01 -4.78137344e-01 -5.34439981e-01 -5.05332649e-01
9.87312496e-01 -3.07875603e-01 8.27874243e-01 7.57767856e-01
-8.03194821e-01 -1.83460832e-01 -1.17108917e+00 7.39661455e-01
-3.23416680e-01 -6.21022284e-01 5.42913675e-01 5.25759757e-01
-8.39478791e-01 8.26083481e-01 -1.04084289e+00 -4.44016844e-01
2.25372445e-02 6.55587316e-01 -2.59291977e-01 3.68033767e-01
-1.19882536e+00 1.41834867e+00 8.50905895e-01 -7.99655616e-02
-1.32795250e+00 -1.81755364e-01 -9.54487324e-01 -2.15261936e-01
7.99462259e-01 -2.19474837e-01 1.39751375e+00 -8.61057043e-01
-1.47605860e+00 1.79198101e-01 6.18302107e-01 -6.89105570e-01
7.57709503e-01 -5.84038496e-01 -2.65963763e-01 -1.32028267e-01
-7.62911513e-02 6.37988031e-01 5.63703775e-01 -1.50036848e+00
-9.99958992e-01 -1.01521835e-02 5.74546635e-01 4.52345520e-01
-2.82527179e-01 -2.41193831e-01 1.23356976e-01 -1.58917338e-01
-3.28731149e-01 -1.36245382e+00 -8.79588783e-01 -1.42290294e-01
-1.76422074e-01 -6.11667037e-02 8.04673374e-01 1.87524274e-01
9.67075109e-01 -2.21716332e+00 -1.59227252e-01 4.60949630e-01
-2.58723468e-01 4.27312970e-01 -1.79807007e-01 3.29217494e-01
1.08661413e-01 -3.86406519e-02 -1.46428779e-01 3.05438787e-01
-6.48547858e-02 9.80715573e-01 -3.76130581e-01 6.14015281e-01
-2.15427473e-01 1.98097661e-01 -1.32284915e+00 -3.70303541e-01
5.16707361e-01 1.21617936e-01 -6.48439884e-01 4.30668116e-01
-9.29977670e-02 2.69996822e-01 -7.01272488e-01 1.16472550e-01
4.14010197e-01 4.75722075e-01 4.01438147e-01 6.17731512e-01
-2.27724075e-01 2.75660902e-01 -1.20429504e+00 6.82387173e-01
-5.35728753e-01 2.66844500e-02 2.74541706e-01 -7.91318417e-01
6.69441283e-01 2.89733678e-01 5.20423174e-01 -8.35681260e-01
2.90468723e-01 5.59444912e-02 1.73232064e-01 -4.82832670e-01
1.39123365e-01 -8.11271071e-02 -3.04911286e-01 6.05791807e-01
-3.39734703e-01 -2.42964223e-01 2.94168323e-01 -7.43780052e-03
1.27163529e+00 4.33887541e-02 5.21168470e-01 -4.98397022e-01
5.39545596e-01 5.01382239e-02 9.08012450e-01 1.17157614e+00
-4.23686892e-01 -4.33036119e-01 3.08559388e-01 -7.32169807e-01
-6.12421751e-01 -1.12546396e+00 1.09985568e-01 1.13872588e+00
6.10207677e-01 -2.36619353e-01 -7.94146597e-01 -1.14595783e+00
4.37635072e-02 1.20175004e+00 -5.61204910e-01 -8.15792501e-01
-7.11117744e-01 -5.19992054e-01 4.66343164e-01 4.16361004e-01
5.07668734e-01 -1.35250950e+00 -1.64382350e+00 1.57867491e-01
2.35703588e-01 -6.27192676e-01 -2.21831068e-01 6.63770616e-01
-6.45457983e-01 -1.31717968e+00 6.28369898e-02 -4.87827986e-01
1.03950250e+00 1.48624167e-01 7.87835658e-01 5.24477661e-01
1.01183794e-01 1.68014452e-01 -3.68304461e-01 -5.45264363e-01
-5.94070017e-01 -3.75309080e-01 5.99920571e-01 -5.39082050e-01
1.67265609e-01 -4.49371070e-01 -3.50968748e-01 4.88108516e-01
-8.78291309e-01 -1.12411112e-01 3.16769242e-01 6.13242090e-01
1.79909900e-01 7.15293705e-01 4.43121403e-01 -6.94852650e-01
7.10385084e-01 -2.74324536e-01 -8.36566448e-01 2.43271828e-01
-6.52487099e-01 2.90894449e-01 1.08770311e+00 -6.52729273e-01
-9.28048909e-01 2.84200430e-01 1.37346342e-01 4.91280816e-02
-4.85392064e-01 -6.80669025e-02 -2.45098785e-01 -3.30182165e-01
8.01173747e-01 2.51488443e-02 -1.56610623e-01 -3.00156977e-02
3.43901925e-02 4.06994820e-01 1.04797058e-01 -8.52826178e-01
9.07551289e-01 2.81747669e-01 9.07639265e-02 -7.09101260e-01
-6.58294082e-01 -4.17344272e-03 -1.31374672e-01 -6.46028340e-01
4.40139681e-01 -4.66885000e-01 -1.08060896e+00 2.18996719e-01
-7.36794472e-01 -7.19033480e-01 -2.23737016e-01 3.87332767e-01
-7.07354248e-01 2.41970778e-01 -2.25896597e-01 -1.17973363e+00
6.29670769e-02 -1.00686276e+00 3.39554787e-01 4.04011458e-01
-5.16207516e-01 -6.89010382e-01 2.02890038e-01 -2.19448701e-01
2.50415653e-01 2.94468373e-01 6.58053219e-01 -5.90354860e-01
-1.37990341e-01 3.12726349e-01 5.98985255e-01 2.07430288e-01
2.51162559e-01 -2.73788180e-02 -4.83504385e-01 -5.16825497e-01
-3.67602520e-02 -4.54306930e-01 3.57848972e-01 3.23692001e-02
7.92983592e-01 -8.49183857e-01 -4.70472127e-01 2.16316029e-01
1.18892241e+00 1.09887350e+00 5.86907446e-01 8.53203297e-01
2.54891098e-01 8.47739875e-01 1.08806431e+00 5.97145855e-01
-1.54720573e-02 4.33745474e-01 9.91316020e-01 2.03175128e-01
6.02631330e-01 -1.54651970e-01 6.37828588e-01 -7.58014843e-02
-6.84195897e-03 -2.61418223e-01 -8.66095722e-01 4.39700037e-01
-2.25759983e+00 -1.06658494e+00 1.66452512e-01 2.40701652e+00
7.10329533e-01 6.49879038e-01 1.01575203e-01 3.97523977e-02
4.20540005e-01 -2.26357058e-01 -6.43983901e-01 -7.70278037e-01
4.99570519e-01 -3.60237926e-01 6.49900138e-01 8.46146882e-01
-1.20572269e+00 1.02321398e+00 6.89998055e+00 3.29767197e-01
-1.00911808e+00 -1.99723080e-01 2.96771646e-01 -3.32837671e-01
1.24777362e-04 -6.49892837e-02 -4.69164133e-01 4.10992026e-01
8.03820133e-01 -2.72399098e-01 5.73819339e-01 1.17453301e+00
3.10550839e-01 -5.50324142e-01 -1.13378811e+00 3.27357411e-01
-3.94305736e-01 -5.31151116e-01 -2.73677737e-01 -6.75775558e-02
4.43591475e-01 -2.48058990e-01 -2.23560408e-01 4.73978370e-01
1.17567420e+00 -9.52295601e-01 9.91829336e-01 9.68652964e-02
-2.13111993e-02 -1.51229227e+00 5.73392570e-01 8.73642743e-01
-7.95292914e-01 -4.26999092e-01 -1.86933085e-01 -5.25262058e-01
3.31124328e-02 1.23543940e-01 -8.50131929e-01 9.00978744e-02
8.05578649e-01 3.89273055e-02 -1.97026506e-01 9.31932867e-01
-6.74652398e-01 1.08651817e-01 -5.03905535e-01 -4.54003662e-01
4.74655062e-01 -1.81098640e-01 4.82690006e-01 6.62222683e-01
6.09707050e-02 3.56616884e-01 4.74979013e-01 4.40693855e-01
6.78749084e-01 -3.28171551e-01 -1.13222456e+00 4.84312028e-01
5.10894597e-01 8.23818505e-01 -8.35784256e-01 -2.33040929e-01
3.58660854e-02 5.96597373e-01 3.51331085e-01 3.67116243e-01
-1.06874204e+00 -3.54683608e-01 9.33806121e-01 -3.73227559e-02
-1.86313078e-01 -1.74156725e-01 -9.90450475e-03 -5.98020673e-01
-1.73541978e-01 -9.22083855e-01 3.97176623e-01 -4.92000788e-01
-8.33435535e-01 7.22638547e-01 1.28805235e-01 -1.22411299e+00
-3.21628958e-01 -4.87698227e-01 -5.82222223e-01 1.92764804e-01
-1.02487755e+00 -2.21719503e-01 5.99672012e-02 6.60703003e-01
2.50252753e-01 -3.28264982e-02 8.89717162e-01 -1.56632751e-01
-5.13030231e-01 1.28969252e-01 -1.34666070e-01 -2.26529777e-01
4.76904422e-01 -1.16582537e+00 -9.47563425e-02 9.43404734e-01
-2.51167566e-01 7.20807374e-01 1.25894284e+00 -8.79628003e-01
-9.23482955e-01 -1.13489199e+00 2.84019977e-01 -1.00611173e-01
3.40008587e-01 -1.63721308e-01 -6.46145165e-01 7.32776403e-01
1.13503732e-01 -2.17476413e-01 4.11695719e-01 1.07310683e-01
-5.81584685e-02 -1.75516203e-01 -1.39210069e+00 1.13448322e+00
8.81770790e-01 -4.75833118e-02 -8.00692856e-01 2.39052892e-01
3.92762035e-01 -2.16197506e-01 -2.65493900e-01 5.59070826e-01
4.39091057e-01 -1.06854153e+00 8.14429045e-01 -7.27409780e-01
-1.01076342e-01 -6.35534108e-01 1.69261619e-01 -1.73043334e+00
-3.71611625e-01 -5.95955610e-01 1.43681215e-02 4.89527702e-01
2.72733063e-01 -7.16228187e-01 8.35780323e-01 8.11963797e-01
-4.87243533e-02 -4.83145326e-01 -1.13886511e+00 -1.23943079e+00
1.01668663e-01 -2.44359434e-01 5.10585785e-01 7.54486263e-01
6.04225099e-01 -4.09882627e-02 -5.92289388e-01 5.33091784e-01
7.31393158e-01 -2.59845942e-01 5.45132458e-01 -9.48005974e-01
-1.80363417e-01 -2.60344625e-01 -1.15098737e-01 -4.94230419e-01
3.32086086e-01 -1.38986170e-01 7.32973397e-01 -1.37309837e+00
-2.33913228e-01 -6.81291759e-01 -4.23380613e-01 9.08496201e-01
4.84010717e-03 -1.82090595e-01 2.30045691e-01 -1.79244757e-01
-9.83574927e-01 5.38194835e-01 1.01304483e+00 -1.01323932e-01
-4.34491783e-01 1.93232298e-02 -4.36619908e-01 1.13074744e+00
1.14763653e+00 -6.08987570e-01 -7.43718326e-01 -2.00524569e-01
3.28324229e-01 3.99067663e-02 -3.95538621e-02 -1.23360491e+00
1.83748364e-01 -9.73766387e-01 6.94734752e-02 -5.52358590e-02
1.54953271e-01 -1.41257393e+00 3.43676418e-01 1.41226673e+00
-4.60291862e-01 -6.63646087e-02 3.04364115e-01 2.93907136e-01
2.09476843e-01 -3.73442888e-01 1.06185710e+00 -1.29348934e-01
-6.70271397e-01 -3.34515959e-01 -1.15213883e+00 -2.77775317e-01
1.89027274e+00 -9.23219919e-02 -3.32426310e-01 -3.94942254e-01
-5.74182391e-01 8.82972896e-01 6.44106448e-01 5.01825154e-01
6.46215856e-01 -1.10917163e+00 -8.75355974e-02 2.70562798e-01
6.43584654e-02 -1.05445385e-01 -3.60631384e-02 2.59779364e-01
-6.18395030e-01 1.11135371e-01 -6.63193762e-01 -8.14142525e-02
-1.45249593e+00 8.70383024e-01 7.10157394e-01 -1.22525238e-01
-5.42584419e-01 6.38520896e-01 3.81523281e-01 -4.65217978e-01
6.25119746e-01 -1.78586945e-01 -4.06400293e-01 -4.02580857e-01
7.91402817e-01 3.64325285e-01 -1.80474162e-01 -2.57998228e-01
-7.22556651e-01 2.69298583e-01 -3.53194386e-01 -1.42618611e-01
1.27382171e+00 1.54392317e-01 1.49840936e-01 2.49719862e-02
2.48768970e-01 -1.53391913e-01 -1.68131113e+00 3.04179966e-01
1.31511288e-02 -4.87812370e-01 -6.93138912e-02 -9.01525021e-01
-7.71987617e-01 3.21803302e-01 4.10501331e-01 5.36397576e-01
1.25061965e+00 -3.69983792e-01 2.74311513e-01 8.67025554e-01
1.04233313e+00 -1.52422273e+00 1.68783799e-01 7.40265846e-01
7.00417340e-01 -9.20478106e-01 -6.15425520e-02 -2.88930625e-01
-8.60780478e-01 8.24449360e-01 1.28923988e+00 -3.73828292e-01
3.78744453e-01 5.84317505e-01 1.48363903e-01 7.79230595e-02
-9.66069043e-01 -6.91582635e-02 -4.48072433e-01 8.51176023e-01
-2.97991484e-01 2.00634673e-01 -2.89786041e-01 8.65080953e-02
-2.63623800e-02 -1.76788107e-01 8.69653404e-01 1.50880563e+00
-1.12409687e+00 -1.13553870e+00 -6.45257652e-01 1.52502313e-01
-1.52969375e-01 4.75464106e-01 -5.38845539e-01 8.90151978e-01
5.20874977e-01 1.22867334e+00 1.04836253e-02 -4.23904210e-01
5.04903376e-01 -2.91634172e-01 3.97605807e-01 -5.92424512e-01
-5.25880575e-01 -2.46927902e-01 3.01566511e-01 -9.23381865e-01
-2.41960391e-01 -5.31725526e-01 -1.77210295e+00 -3.87291580e-01
-8.87550414e-03 4.04291481e-01 -1.43490219e-02 1.03563464e+00
2.12483574e-02 6.73176944e-01 6.51794136e-01 -4.82170224e-01
-8.32370758e-01 -3.88009936e-01 -4.57500368e-01 2.19596684e-01
5.14407933e-01 -9.90128458e-01 -3.34064096e-01 -3.29880655e-01] | [4.527477264404297, 2.02911639213562] |
ac15e8db-0e9f-46ba-993d-3ff12654d22e | pugeo-net-a-geometry-centric-network-for-3d | 2002.10277 | null | https://arxiv.org/abs/2002.10277v2 | https://arxiv.org/pdf/2002.10277v2.pdf | PUGeo-Net: A Geometry-centric Network for 3D Point Cloud Upsampling | This paper addresses the problem of generating uniform dense point clouds to describe the underlying geometric structures from given sparse point clouds. Due to the irregular and unordered nature, point cloud densification as a generative task is challenging. To tackle the challenge, we propose a novel deep neural network based method, called PUGeo-Net, that learns a $3\times 3$ linear transformation matrix $\bf T$ for each input point. Matrix $\mathbf T$ approximates the augmented Jacobian matrix of a local parameterization and builds a one-to-one correspondence between the 2D parametric domain and the 3D tangent plane so that we can lift the adaptively distributed 2D samples (which are also learned from data) to 3D space. After that, we project the samples to the curved surface by computing a displacement along the normal of the tangent plane. PUGeo-Net is fundamentally different from the existing deep learning methods that are largely motivated by the image super-resolution techniques and generate new points in the abstract feature space. Thanks to its geometry-centric nature, PUGeo-Net works well for both CAD models with sharp features and scanned models with rich geometric details. Moreover, PUGeo-Net can compute the normal for the original and generated points, which is highly desired by the surface reconstruction algorithms. Computational results show that PUGeo-Net, the first neural network that can jointly generate vertex coordinates and normals, consistently outperforms the state-of-the-art in terms of accuracy and efficiency for upsampling factor $4\sim 16$. | ['Sam Kwong', 'Junhui Hou', 'Yue Qian', 'Ying He'] | 2020-02-24 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3338_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123640732.pdf | eccv-2020-8 | ['point-cloud-super-resolution'] | ['computer-vision'] | [-7.13023469e-02 6.84602186e-02 1.95531264e-01 -1.58455759e-01
-7.10298240e-01 -1.13027014e-01 5.70863307e-01 -3.73952448e-01
5.11154607e-02 7.27214158e-01 -1.79868504e-01 1.81243941e-02
-1.33441776e-01 -1.55403864e+00 -1.30100095e+00 -5.71203232e-01
3.87524664e-02 1.15532255e+00 -1.26015693e-01 -3.40193480e-01
1.72075614e-01 1.09052372e+00 -1.68907905e+00 -6.08676206e-03
1.00914240e+00 1.03594613e+00 6.65999651e-02 3.08418959e-01
-5.03571272e-01 -1.39420882e-01 -1.99589357e-01 -2.37543136e-01
5.07135689e-01 -7.13173253e-03 -5.62505424e-01 1.15230381e-01
7.81180739e-01 -4.14828211e-01 -1.03223294e-01 9.89801168e-01
2.96740711e-01 8.08969978e-03 8.52735341e-01 -1.00926661e+00
-9.76679146e-01 1.75359905e-01 -7.94845700e-01 -5.36937475e-01
5.39784916e-02 -2.65173707e-03 7.42884696e-01 -1.57882011e+00
9.58300173e-01 1.26954961e+00 1.01059687e+00 3.38995904e-01
-1.43916357e+00 -8.02794099e-01 -3.96417193e-02 -3.19959402e-01
-1.73011172e+00 1.14065856e-01 1.17015338e+00 -5.76262116e-01
6.30326986e-01 7.91997015e-02 8.52403104e-01 8.12291443e-01
-7.00250315e-03 3.31287682e-01 6.91228092e-01 -2.30039567e-01
2.48511568e-01 -2.75455832e-01 -2.99977690e-01 6.81670427e-01
2.66703784e-01 2.39030108e-01 -2.86022276e-01 -3.36444646e-01
1.84259450e+00 2.28898913e-01 -2.24148795e-01 -6.64399743e-01
-1.15601337e+00 9.45711315e-01 7.83429921e-01 1.68848157e-01
-6.72233284e-01 4.29193705e-01 -9.10061300e-02 -1.66515082e-01
6.75241053e-01 4.82016534e-01 -2.47377619e-01 7.95413852e-02
-9.58687067e-01 7.56151080e-01 5.73190510e-01 1.27221811e+00
1.29930687e+00 4.40410554e-01 1.06307648e-01 8.05212736e-01
2.82853425e-01 5.91099679e-01 7.95258284e-02 -1.31180525e+00
3.57430309e-01 7.72327840e-01 2.75463551e-01 -1.22140956e+00
-1.48418173e-01 -5.45355558e-01 -1.37337816e+00 5.04283011e-01
-3.16845663e-02 -4.37874980e-02 -1.07156527e+00 1.40306640e+00
6.06897175e-01 3.38957220e-01 -2.72322834e-01 7.21257627e-01
7.50646472e-01 8.21164608e-01 -4.33727980e-01 7.27377161e-02
9.35140073e-01 -5.47126055e-01 -2.65528858e-01 5.86307235e-02
2.47542143e-01 -5.52724063e-01 1.15524840e+00 1.96234703e-01
-1.24206471e+00 -6.40410066e-01 -9.65587378e-01 -2.32882321e-01
3.33934873e-02 -9.30384826e-03 5.97764015e-01 -3.34183052e-02
-1.09980500e+00 8.61639917e-01 -9.53368485e-01 1.44024715e-01
9.08188581e-01 3.14667344e-01 -1.69690534e-01 -2.08894253e-01
-8.07291567e-01 3.39244366e-01 -3.79588790e-02 1.51487172e-01
-6.29520237e-01 -1.33172584e+00 -7.25517452e-01 2.10035488e-01
6.98552504e-02 -1.13069642e+00 9.54452634e-01 -5.14042795e-01
-1.55764878e+00 8.03749025e-01 -1.34159669e-01 -1.83080450e-01
6.67950928e-01 -1.45791560e-01 6.24583773e-02 -5.95516115e-02
2.29159549e-01 8.23218465e-01 9.94719207e-01 -1.69077826e+00
-4.17809278e-01 -5.05418658e-01 -2.24117935e-01 1.61103904e-01
1.01703934e-01 -7.34014392e-01 -2.95023650e-01 -7.35415637e-01
4.45085019e-01 -7.21651614e-01 -4.62254167e-01 3.87448430e-01
-4.25365537e-01 -3.30973476e-01 1.03987062e+00 -3.24416697e-01
6.93405032e-01 -2.11285591e+00 1.81213960e-01 4.17916745e-01
4.95290726e-01 6.25124201e-02 2.24735755e-02 2.81805664e-01
-7.00256079e-02 2.25453123e-01 -5.74409723e-01 -4.69887376e-01
2.17281561e-02 1.63156658e-01 -3.27868462e-01 3.84476125e-01
3.17938864e-01 9.53710198e-01 -7.04047382e-01 -1.07956484e-01
2.48739213e-01 8.81692052e-01 -8.03277969e-01 1.02502145e-01
-5.27322054e-01 3.86367232e-01 -5.52698433e-01 5.11209548e-01
1.26270568e+00 -3.58273268e-01 -3.87933671e-01 -2.61131287e-01
-1.67831421e-01 -2.16832206e-01 -1.31534886e+00 1.90155566e+00
-5.14977038e-01 1.17194578e-01 2.55665541e-01 -8.35075140e-01
1.41288245e+00 2.04770267e-01 6.42926276e-01 -3.11782271e-01
-2.13207421e-03 4.11945939e-01 -4.95353341e-01 9.11872238e-02
5.14562368e-01 -4.28583264e-01 1.12645261e-01 1.43098697e-01
-1.65451080e-01 -8.03328872e-01 -4.37100977e-01 -1.93347279e-02
5.14630616e-01 3.49872082e-01 -5.90498038e-02 -3.55821222e-01
3.11579019e-01 1.02657139e-01 5.93850970e-01 3.99412841e-01
7.10818112e-01 9.51136112e-01 3.04763556e-01 -7.48322845e-01
-1.51133037e+00 -1.24154139e+00 -3.23828012e-01 3.22501808e-01
6.87641427e-02 -1.76394850e-01 -9.10412610e-01 -7.26618767e-02
2.66753197e-01 6.33453548e-01 -5.85679412e-01 8.35107043e-02
-9.60852563e-01 -2.91742682e-01 1.04445681e-01 6.84521258e-01
5.93225837e-01 -1.08555734e+00 -2.57492095e-01 3.02611887e-01
1.53339297e-01 -9.02395189e-01 -3.88386726e-01 -1.80314288e-01
-1.20707500e+00 -8.23565722e-01 -7.81154454e-01 -8.95065904e-01
8.90780270e-01 3.81540954e-02 1.19122958e+00 7.95422569e-02
-7.60240108e-02 -1.77379437e-02 6.74015358e-02 -4.07131284e-01
-2.30594501e-01 1.37643531e-01 -9.89514738e-02 8.90276209e-02
2.34511510e-01 -1.12951386e+00 -5.82570732e-01 1.35295302e-01
-9.27697659e-01 3.92475605e-01 3.34955424e-01 7.96293318e-01
1.36521232e+00 5.62036671e-02 3.08862567e-01 -9.22993124e-01
4.49734569e-01 -3.76502335e-01 -8.64206731e-01 -3.82897645e-01
-2.19044313e-01 9.68456119e-02 7.46978283e-01 -2.24765450e-01
-7.28344738e-01 2.25330293e-01 -4.19843048e-01 -1.07452047e+00
-1.70482665e-01 3.58956844e-01 -9.10967886e-02 -1.57212287e-01
7.62923419e-01 2.79755801e-01 1.24957343e-03 -6.63971007e-01
4.55988586e-01 9.26141292e-02 6.49622262e-01 -8.65644813e-01
1.16517246e+00 8.55002165e-01 4.12213743e-01 -1.00466073e+00
-5.31518161e-01 -6.30525425e-02 -8.25839221e-01 1.30229190e-01
8.12997460e-01 -8.00153077e-01 -5.65746129e-01 5.76907575e-01
-1.48386514e+00 -2.74987787e-01 -7.81738400e-01 1.35681957e-01
-8.81197572e-01 9.80903357e-02 -4.35042709e-01 -3.82593602e-01
-5.66291571e-01 -1.12968612e+00 1.48880827e+00 -6.01420142e-02
3.87056135e-02 -7.48201430e-01 9.25433338e-02 -4.38901111e-02
4.74611253e-01 8.37425113e-01 1.18474102e+00 1.37769476e-01
-1.08802211e+00 -2.71306634e-01 -3.03539604e-01 3.62466753e-01
1.11959331e-01 7.94183239e-02 -6.72497511e-01 -1.80841669e-01
1.67321116e-01 -1.16229272e-02 4.25009519e-01 5.85188746e-01
1.52214038e+00 -4.50110644e-01 -3.56899053e-01 1.26255751e+00
1.68894958e+00 -1.36650354e-01 6.72458291e-01 1.16483040e-01
1.15155923e+00 2.01358855e-01 1.29011050e-01 5.20486712e-01
3.37969035e-01 6.59297228e-01 7.95529187e-01 -1.94773182e-01
-1.19249538e-01 -4.63885844e-01 -2.89068103e-01 8.46281171e-01
-4.55883920e-01 3.02461475e-01 -8.92447472e-01 4.54893440e-01
-1.48869622e+00 -6.20788753e-01 -2.91633874e-01 2.12024665e+00
7.01469421e-01 -6.32126629e-02 -9.93753225e-02 -1.08245768e-01
7.48628438e-01 1.05425015e-01 -6.53125465e-01 -2.27785781e-01
3.61802913e-02 7.68641651e-01 2.80907750e-01 5.79597235e-01
-5.76589406e-01 1.00556350e+00 5.58038950e+00 9.74299252e-01
-1.32619500e+00 -1.06048919e-01 5.11797905e-01 1.78620256e-02
-6.90824389e-01 -3.33490700e-01 -8.77561450e-01 3.44280452e-01
4.87365335e-01 -2.05490902e-01 4.51065123e-01 1.11933267e+00
-1.46038933e-02 4.94325370e-01 -1.16730034e+00 1.20277214e+00
-1.48698643e-01 -2.13930893e+00 6.01061523e-01 3.44144374e-01
1.08027208e+00 1.85346112e-01 1.54703572e-01 1.05823740e-01
4.51837897e-01 -1.27229369e+00 7.42098331e-01 6.73736334e-01
1.37561870e+00 -9.89731193e-01 3.17158997e-01 5.57757318e-01
-1.23257554e+00 4.89870369e-01 -6.35065138e-01 8.96224529e-02
3.44834626e-01 8.10918272e-01 -8.15203369e-01 6.35225296e-01
7.97963858e-01 7.81817973e-01 -2.77467584e-03 7.91730106e-01
6.93813935e-02 1.81943983e-01 -6.55275106e-01 3.57538670e-01
3.48011822e-01 -6.69390202e-01 6.15074635e-01 6.29687011e-01
8.81035507e-01 1.67162433e-01 8.49588066e-02 1.64719236e+00
-4.44211572e-01 -9.32250917e-03 -7.91850507e-01 3.30497891e-01
7.16463804e-01 1.07578266e+00 -4.93927479e-01 -1.92653045e-01
3.64545770e-02 6.41390979e-01 4.33457196e-01 3.61878127e-01
-5.26948571e-01 -2.57290632e-01 7.56633878e-01 7.03953266e-01
5.89046419e-01 -4.78875518e-01 -6.28461123e-01 -8.81400108e-01
2.75336504e-01 -5.53878188e-01 -2.96724290e-01 -9.55027163e-01
-1.38771391e+00 9.53682899e-01 -5.54287359e-02 -1.36583686e+00
-1.80905789e-01 -5.21829069e-01 -5.66510260e-01 1.16514754e+00
-1.43161809e+00 -1.31288397e+00 -5.84851861e-01 6.28517747e-01
3.72997046e-01 2.55871862e-02 8.10646832e-01 1.37986958e-01
-1.15281567e-02 3.11335802e-01 8.83353651e-02 7.74006769e-02
3.19405086e-02 -1.00508559e+00 8.35478902e-01 4.10647780e-01
-1.12767220e-01 5.98814905e-01 2.85875022e-01 -7.35328197e-01
-1.35784209e+00 -1.29671073e+00 3.19303542e-01 -3.38443995e-01
2.27765769e-01 -4.41358805e-01 -1.26352096e+00 7.05766559e-01
-3.39498550e-01 3.55579913e-01 7.08981901e-02 -2.25152880e-01
-1.28831699e-01 -9.40346941e-02 -1.31293929e+00 5.00647068e-01
1.23124957e+00 -2.79865503e-01 -3.35258901e-01 3.72334093e-01
9.95503008e-01 -9.28990841e-01 -1.02871788e+00 5.73805809e-01
1.64543882e-01 -9.02000427e-01 1.28424430e+00 -4.01784301e-01
7.18986452e-01 -2.92691857e-01 -2.28195280e-01 -1.42855227e+00
-5.91732979e-01 -5.57297170e-01 -1.38227031e-01 9.50001657e-01
1.03180863e-01 -5.53006470e-01 1.26656616e+00 3.46160382e-01
-5.67847610e-01 -1.20430624e+00 -1.13649178e+00 -5.20361543e-01
6.47016764e-01 -4.27811474e-01 1.37276602e+00 1.05781722e+00
-9.93701339e-01 1.14420637e-01 -5.15974984e-02 2.94813156e-01
8.31961989e-01 3.23756874e-01 1.01334155e+00 -1.63874614e+00
1.07935295e-01 -4.63737905e-01 -2.54192710e-01 -1.24413502e+00
-6.23329543e-02 -8.42575192e-01 -1.90929193e-02 -1.67418897e+00
-5.00459671e-01 -9.58464146e-01 4.48660731e-01 5.84001206e-02
1.83382764e-01 1.78109676e-01 -1.14725403e-01 3.19688678e-01
2.80263364e-01 1.00796378e+00 1.79594171e+00 1.00400306e-01
-3.13554049e-01 -8.73392299e-02 -6.98988378e-01 9.40399826e-01
3.78129542e-01 -1.62028968e-01 -9.56254676e-02 -8.43680620e-01
2.98548847e-01 1.31181315e-01 4.09536153e-01 -1.05067158e+00
8.68639871e-02 -3.58103007e-01 4.23879772e-01 -1.16431332e+00
7.02121854e-01 -9.09867644e-01 5.28727412e-01 -4.61121723e-02
1.03335842e-01 1.51424989e-01 2.34280117e-02 5.19932449e-01
-7.15074465e-02 9.71464664e-02 9.08560514e-01 -4.24686104e-01
-1.89621031e-01 1.09671187e+00 2.47964263e-01 -4.63692024e-02
8.10026050e-01 -4.88136977e-01 -1.19783454e-01 -1.32906452e-01
-5.37380695e-01 -7.81858806e-04 8.29017937e-01 9.31371227e-02
9.27176714e-01 -1.88319635e+00 -8.25202465e-01 6.82286203e-01
-1.69648454e-01 1.38496864e+00 3.87124151e-01 2.39690706e-01
-9.91174757e-01 1.48049831e-01 -1.80888966e-01 -9.78886724e-01
-4.58699435e-01 3.94832522e-01 5.07909715e-01 -5.43030649e-02
-1.18417239e+00 8.48936081e-01 3.83806884e-01 -7.41141558e-01
-1.25811294e-01 -5.80365479e-01 1.47435874e-01 -3.60710621e-01
1.28781602e-01 3.96934450e-01 1.73323885e-01 -6.48957610e-01
5.85637242e-02 1.03180695e+00 6.59651831e-02 1.43259630e-01
1.82767403e+00 3.63233268e-01 -3.81249011e-01 2.23166600e-01
1.22738385e+00 -5.46292290e-02 -1.51198399e+00 -3.71124595e-01
-5.32907546e-01 -6.67430520e-01 5.38018625e-03 -2.16393158e-01
-1.35181808e+00 8.69771838e-01 1.81302398e-01 -2.16136891e-02
5.69815516e-01 3.41606885e-02 1.03140664e+00 1.69948667e-01
6.08646989e-01 -7.45485961e-01 -1.67543124e-02 6.33390844e-01
1.37308419e+00 -6.94734097e-01 2.11866256e-02 -7.60226309e-01
-1.84298769e-01 9.85552549e-01 7.41557896e-01 -8.75149250e-01
8.84437978e-01 1.81696177e-01 -2.54833221e-01 -6.47974312e-01
-2.60848224e-01 4.98308241e-01 3.17005396e-01 7.24339664e-01
1.02355190e-01 1.22401901e-01 1.81091264e-01 3.96661401e-01
-8.72031689e-01 2.26641715e-01 3.71328086e-01 6.44139826e-01
-2.56487519e-01 -9.30360973e-01 -5.84226429e-01 6.20215714e-01
1.64143607e-01 6.56521469e-02 2.29158439e-02 8.63291323e-01
2.12017328e-01 -8.54218602e-02 5.97060621e-01 -1.73619136e-01
5.18684983e-01 -1.33076906e-01 4.01099652e-01 -7.80912995e-01
-7.08132684e-02 4.70559783e-02 -4.69192922e-01 -4.95097041e-01
-1.14287041e-01 -5.75978279e-01 -1.45918548e+00 -4.94397491e-01
7.59643912e-02 1.81129336e-01 5.89981973e-01 5.74692369e-01
6.80864632e-01 4.55619842e-01 7.03944266e-01 -1.62522364e+00
-4.58148479e-01 -7.50646770e-01 -7.07393467e-01 4.71913636e-01
3.29326749e-01 -8.92855167e-01 -3.22032332e-01 -1.72598258e-01] | [8.594663619995117, -3.6473848819732666] |
63f8c5c5-92df-47cd-a110-e6baace3134d | lets-gzsl-a-latent-embedding-model-for-time | 2207.12007 | null | https://arxiv.org/abs/2207.12007v1 | https://arxiv.org/pdf/2207.12007v1.pdf | LETS-GZSL: A Latent Embedding Model for Time Series Generalized Zero Shot Learning | One of the recent developments in deep learning is generalized zero-shot learning (GZSL), which aims to recognize objects from both seen and unseen classes, when only the labeled examples from seen classes are provided. Over the past couple of years, GZSL has picked up traction and several models have been proposed to solve this problem. Whereas an extensive amount of research on GZSL has been carried out in fields such as computer vision and natural language processing, no such research has been carried out to deal with time series data. GZSL is used for applications such as detecting abnormalities from ECG and EEG data and identifying unseen classes from sensor, spectrograph and other devices' data. In this regard, we propose a Latent Embedding for Time Series - GZSL (LETS-GZSL) model that can solve the problem of GZSL for time series classification (TSC). We utilize an embedding-based approach and combine it with attribute vectors to predict the final class labels. We report our results on the widely popular UCR archive datasets. Our framework is able to achieve a harmonic mean value of at least 55% on most of the datasets except when the number of unseen classes is greater than 3 or the amount of data is very low (less than 100 training examples). | ['Manik Gupta', 'Priyanka Gupta', 'Sathvik Bhaskarpandit'] | 2022-07-25 | null | null | null | null | ['generalized-zero-shot-learning', 'generalized-zero-shot-learning'] | ['computer-vision', 'methodology'] | [ 4.01341319e-01 3.91906798e-02 -6.15196452e-02 -4.18826312e-01
-7.25517929e-01 -1.72501132e-01 5.09222269e-01 4.33458894e-01
-2.65328050e-01 6.08899713e-01 5.21986596e-02 6.53721020e-02
-3.26266199e-01 -8.18803847e-01 -1.09277397e-01 -7.74347425e-01
-3.24191809e-01 1.94911346e-01 1.83625191e-01 -1.62176713e-01
4.17111255e-02 3.28703284e-01 -1.79245520e+00 1.44705743e-01
6.38826609e-01 1.07014179e+00 -1.73078403e-01 3.59546989e-01
-1.13563567e-01 5.89907765e-01 -6.90761387e-01 -1.63305670e-01
1.12138793e-01 -4.65126187e-01 -6.08162642e-01 -1.66285718e-05
-5.89845516e-02 -9.47149396e-02 -2.97360599e-01 9.66334641e-01
9.69235420e-01 2.46499866e-01 7.84971893e-01 -1.54694498e+00
-5.63431859e-01 3.87626171e-01 -4.03443396e-01 6.14219487e-01
2.06015766e-01 -1.35704592e-01 8.22371185e-01 -7.61717618e-01
4.02280182e-01 1.05871034e+00 6.57175899e-01 5.97864211e-01
-1.08136237e+00 -6.28292501e-01 -1.20602094e-01 7.00930953e-01
-1.32560396e+00 -2.34278500e-01 1.01917803e+00 -5.36557734e-01
1.14026475e+00 1.02137342e-01 5.26675940e-01 1.39791918e+00
3.27132434e-01 7.46100664e-01 8.97382319e-01 -6.22274041e-01
6.85200810e-01 1.75410703e-01 6.28282368e-01 1.65404722e-01
2.44304053e-02 -2.23403741e-02 -4.40174788e-01 -2.14490116e-01
3.80208820e-01 3.65773052e-01 -1.60035610e-01 -2.73336262e-01
-1.09708071e+00 8.50158811e-01 1.89190265e-02 7.71604002e-01
-4.06835854e-01 -3.12991232e-01 8.03454220e-01 6.25070214e-01
8.33423436e-01 3.73439759e-01 -3.83033514e-01 -1.97117493e-01
-8.04559648e-01 3.28250835e-03 6.50507152e-01 7.21253037e-01
3.22097957e-01 3.55023652e-01 -2.06321597e-01 1.01733446e+00
-2.14261129e-01 4.25086617e-02 1.18850994e+00 -5.32758653e-01
1.62590772e-01 8.38368177e-01 -1.70096889e-01 -1.11979628e+00
-3.43547255e-01 -4.21748936e-01 -9.56757069e-01 7.77188987e-02
6.17884062e-02 -1.25277787e-01 -9.68733668e-01 1.69457316e+00
9.98869315e-02 6.14229739e-01 2.24114567e-01 7.12405086e-01
9.27340686e-01 7.97718763e-01 -8.41079727e-02 -3.69925439e-01
1.23603010e+00 -4.99251634e-01 -1.04328680e+00 -8.56153741e-02
6.05890810e-01 -3.82758498e-01 8.74331355e-01 5.18257201e-01
-4.87031132e-01 -4.98218924e-01 -1.27882338e+00 1.17114328e-01
-7.84620881e-01 -1.47702739e-01 5.02166271e-01 6.13258123e-01
-7.07250655e-01 7.23153412e-01 -9.31334317e-01 -6.18254960e-01
5.17185509e-01 3.86634648e-01 -4.07992035e-01 -2.75729224e-02
-1.40519631e+00 8.34273636e-01 4.90603685e-01 -3.84799391e-02
-8.85284901e-01 -3.10794115e-01 -7.76074767e-01 1.56417176e-01
2.92112708e-01 -1.31225795e-01 7.44816482e-01 -8.26816261e-01
-1.04341447e+00 8.62718284e-01 1.42842114e-01 -6.76338971e-01
2.07187399e-01 -7.80182797e-03 -1.00873971e+00 3.77914384e-02
9.63514298e-02 2.57942468e-01 8.70075703e-01 -6.49594605e-01
-1.78286850e-01 -6.21052027e-01 -1.31547451e-01 -9.22153294e-02
-7.56398320e-01 -6.43542111e-02 6.01930954e-02 -8.08564484e-01
2.74729401e-01 -7.20440984e-01 -4.69218828e-02 -4.16582227e-02
-2.27276623e-01 -6.66327000e-01 1.08134377e+00 -3.22894067e-01
1.17835093e+00 -2.42357373e+00 6.94186753e-03 -1.57479241e-01
2.45152116e-01 4.92690057e-01 -4.97743078e-02 7.47487009e-01
-5.12973785e-01 -1.92829520e-01 -3.59044075e-01 -1.27946019e-01
-3.76178958e-02 3.81537318e-01 -3.51118416e-01 5.83152115e-01
1.59694836e-01 6.70805752e-01 -9.63782609e-01 -2.23681584e-01
4.44254309e-01 5.87353647e-01 -1.54584125e-01 5.24240471e-02
2.96416849e-01 1.34883925e-01 -2.98741102e-01 4.49495494e-01
2.63857067e-01 -2.68676400e-01 -1.50350839e-01 -1.04885504e-01
1.29430462e-02 -5.86768538e-02 -1.23112619e+00 1.57331085e+00
-1.12749137e-01 7.82186985e-01 -6.63498163e-01 -1.38145328e+00
1.03665495e+00 8.24200451e-01 8.17287087e-01 -5.51370323e-01
3.61863673e-01 1.59524158e-01 3.38977352e-02 -7.93297887e-01
-4.56964858e-02 -3.70534003e-01 -5.57538867e-02 4.69014645e-01
4.65359956e-01 3.47369105e-01 7.38548562e-02 3.04009598e-02
1.24333525e+00 -2.39112884e-01 6.03375137e-01 -7.13247880e-02
4.52636033e-01 -2.19321996e-01 6.52643800e-01 5.03187776e-01
-5.82340002e-01 6.24962032e-01 3.53979439e-01 -7.60969281e-01
-1.01162755e+00 -9.34700429e-01 -4.19703782e-01 8.94812167e-01
-9.84688252e-02 -3.80376726e-01 -6.88869357e-01 -3.54094505e-01
-1.87490702e-01 6.89445138e-01 -6.87891960e-01 -6.45595670e-01
-1.23840995e-01 -9.57799137e-01 4.57802415e-01 6.30930126e-01
3.12721878e-01 -1.32888687e+00 -8.28600883e-01 4.52134997e-01
1.23973690e-01 -9.09510076e-01 1.41203225e-01 3.86730194e-01
-9.00413394e-01 -1.09357452e+00 -7.53114045e-01 -8.48835051e-01
4.65133041e-01 -2.03076228e-02 6.58036351e-01 -4.09109950e-01
-7.54973710e-01 3.62146020e-01 -6.75660074e-01 -7.89243877e-01
-3.61169595e-03 -1.45570949e-01 3.35902840e-01 4.32488471e-01
9.93248343e-01 -8.55653226e-01 -4.01621044e-01 1.68945566e-02
-9.09640968e-01 -3.45306516e-01 2.30949834e-01 9.37316775e-01
3.81157815e-01 3.31921339e-01 1.19705880e+00 -8.31983685e-01
8.31679463e-01 -6.67401195e-01 -1.25293180e-01 4.39608544e-02
-6.22240961e-01 -1.92947179e-01 7.02181518e-01 -6.70948744e-01
-3.37678552e-01 -2.19853073e-01 -2.46074915e-01 -8.04962337e-01
-3.31956804e-01 4.01674539e-01 6.61748499e-02 2.66214281e-01
6.63386822e-01 3.84724081e-01 -1.25606686e-01 -6.01291299e-01
4.99381945e-02 1.14618909e+00 4.48236227e-01 -1.14738263e-01
4.27086264e-01 3.72650564e-01 -2.99450636e-01 -1.01898515e+00
-7.95951962e-01 -5.52705526e-01 -6.14575684e-01 -1.40054271e-01
8.08158457e-01 -5.84904253e-01 -3.81425589e-01 4.15139139e-01
-7.74515867e-01 2.51971275e-01 -6.19570196e-01 7.36871243e-01
-5.77535272e-01 2.56645411e-01 -4.47681904e-01 -1.19318962e+00
-4.79639262e-01 -1.02041757e+00 9.80216205e-01 -6.18705899e-02
-4.18448746e-01 -1.00395513e+00 1.72444850e-01 -1.29757062e-01
3.78483623e-01 7.11896896e-01 1.09465802e+00 -1.09700608e+00
-5.21754436e-02 -6.35517418e-01 2.12187797e-01 4.87195700e-01
4.21906739e-01 -5.21646142e-01 -1.24352860e+00 -5.31867683e-01
5.90642869e-01 -4.09465790e-01 6.77315295e-01 4.32594955e-01
1.30812037e+00 -4.32528881e-03 -2.45291710e-01 3.69787365e-01
1.43373811e+00 7.27890790e-01 5.37687004e-01 1.79434434e-01
4.29200143e-01 5.69431841e-01 4.77147102e-01 6.61818266e-01
-1.02795959e-02 4.40118253e-01 2.88549870e-01 1.46246120e-01
1.62107885e-01 6.32534027e-02 1.38846457e-01 1.14469361e+00
4.02300060e-02 -2.38533229e-01 -1.02994382e+00 8.37722957e-01
-1.76045501e+00 -1.04744434e+00 2.87177004e-02 2.10429645e+00
4.53160942e-01 2.43109182e-01 -1.52689323e-01 1.00772893e+00
8.08530986e-01 2.41458580e-01 -7.56947756e-01 -4.03984547e-01
-2.86460053e-02 3.22944224e-01 -8.84160176e-02 -1.77010987e-02
-1.21823621e+00 5.60250938e-01 6.06931114e+00 7.89008498e-01
-1.28041780e+00 2.55360901e-01 4.83909011e-01 -9.32328925e-02
-1.73242781e-02 -3.50564867e-01 -4.56122488e-01 6.12710297e-01
1.07244730e+00 -5.45586765e-01 1.46790534e-01 8.76450419e-01
2.32241154e-01 3.36455494e-01 -1.19586802e+00 1.52905357e+00
2.55337745e-01 -9.22957301e-01 -4.78945300e-02 -1.15937248e-01
5.02073944e-01 -4.13131714e-02 8.54543149e-02 5.85372508e-01
-4.21236485e-01 -9.76468086e-01 1.41724810e-01 3.57550651e-01
8.09292912e-01 -8.82114291e-01 9.02307808e-01 6.65580153e-01
-1.08423948e+00 -3.86957198e-01 -5.25555253e-01 -1.83695018e-01
-4.06334437e-02 5.96732080e-01 -6.36853516e-01 4.79151130e-01
7.59381831e-01 1.09145558e+00 -2.85532296e-01 9.90532637e-01
1.66834787e-01 7.57332563e-01 -1.13767907e-01 2.64249258e-02
2.36036018e-01 -2.65348796e-02 5.90566397e-01 7.11619377e-01
3.54176641e-01 3.29649538e-01 1.12403534e-01 6.42289042e-01
1.05332963e-01 1.91294864e-01 -8.95978689e-01 -2.94911623e-01
3.37901771e-01 9.98945594e-01 -7.65120924e-01 -3.74604374e-01
-3.74565840e-01 9.42167878e-01 4.22752015e-02 2.24136278e-01
-5.94586730e-01 -7.42704272e-01 4.11523432e-01 -3.32616754e-02
4.37274911e-02 4.65079360e-02 2.71436758e-02 -1.26736248e+00
-4.52300720e-02 -6.51639462e-01 5.76488197e-01 -6.79918945e-01
-1.68186986e+00 9.23148990e-01 -1.77262409e-03 -1.68220067e+00
-3.37233096e-01 -5.43147385e-01 -6.34177625e-01 7.28705645e-01
-1.25372291e+00 -5.19782186e-01 -1.33657947e-01 7.33739257e-01
9.15198624e-01 -4.77249622e-01 1.23610616e+00 5.45988381e-01
-6.04962707e-01 4.07536179e-01 3.84121776e-01 3.36024165e-01
5.72108448e-01 -1.16229284e+00 2.76230961e-01 6.18441463e-01
2.78475255e-01 5.72928011e-01 6.55881524e-01 -3.15882266e-01
-1.12055469e+00 -1.08730197e+00 9.92143750e-01 -3.41443956e-01
5.61295986e-01 -4.74477619e-01 -1.10922551e+00 5.00272036e-01
3.82634029e-02 4.05617714e-01 9.27723944e-01 4.37878817e-02
-1.11196227e-01 -3.27667028e-01 -1.05850935e+00 2.32050747e-01
7.87709653e-01 -6.79003239e-01 -1.02014303e+00 3.84878874e-01
4.87347633e-01 -5.65834716e-02 -9.54038084e-01 4.49863732e-01
3.54281783e-01 -7.41345882e-01 7.82300174e-01 -7.07426131e-01
1.58846140e-01 -7.99209848e-02 -3.00033867e-01 -1.34756529e+00
-2.09921882e-01 -2.92701930e-01 -2.63084680e-01 1.10208166e+00
-3.54702361e-02 -6.62857831e-01 6.33941948e-01 3.75084460e-01
-2.00859219e-01 -9.59792316e-01 -1.10995638e+00 -9.13789690e-01
-2.65270263e-01 -5.95438540e-01 5.37734568e-01 1.44039869e+00
6.05962351e-02 4.64265883e-01 -5.98415434e-01 -8.17681476e-02
6.46998823e-01 1.80976897e-01 2.46275619e-01 -1.62127304e+00
-5.55544011e-02 -2.00136652e-04 -1.07763219e+00 -2.31650069e-01
1.06584020e-01 -1.04033935e+00 -3.81811291e-01 -1.56089211e+00
6.00111261e-02 -1.19945683e-01 -8.56356740e-01 6.21655464e-01
1.07815176e-01 2.27568343e-01 -6.53301030e-02 8.68370607e-02
-4.31833386e-01 6.79515123e-01 8.87019753e-01 -2.17925295e-01
-1.59531415e-01 1.70133002e-02 -5.27856350e-01 6.52698159e-01
6.51073515e-01 -5.66235840e-01 -8.13893437e-01 -1.24166265e-01
-1.58422798e-01 1.98909998e-01 1.86925128e-01 -1.33737659e+00
1.26637146e-01 2.64968306e-01 3.92606318e-01 -6.50977254e-01
4.86076266e-01 -9.48839188e-01 8.15927982e-02 3.17392498e-01
-4.28040773e-01 -4.46803831e-02 -6.02619909e-02 7.89347053e-01
-4.39008474e-01 -1.86260089e-01 6.90429568e-01 -2.74341274e-02
-9.36318934e-01 4.75264817e-01 -3.16240191e-01 -3.94642306e-03
1.26384366e+00 -4.73138958e-01 -1.51125327e-01 -2.47735590e-01
-1.16558886e+00 2.14462969e-02 -6.10553026e-02 7.96782732e-01
8.39053631e-01 -1.54841816e+00 -6.11944079e-01 5.90267897e-01
5.28446615e-01 -3.87560397e-01 3.51882815e-01 7.09837258e-01
-1.05785638e-01 4.11062270e-01 -3.04063052e-01 -6.58642471e-01
-1.10828626e+00 7.31348813e-01 1.80363610e-01 9.28506404e-02
-1.13722873e+00 5.72691083e-01 -9.18855518e-02 -1.14861645e-01
4.35049593e-01 -2.32300580e-01 -7.11672068e-01 3.76413792e-01
7.26028979e-01 4.93908107e-01 2.06236735e-01 -6.32525384e-01
-3.90321553e-01 4.99217421e-01 -3.28802504e-02 8.91057700e-02
1.78900874e+00 1.10565342e-01 1.38332874e-01 1.28146124e+00
1.47812819e+00 -7.07006395e-01 -7.75402069e-01 -2.38577798e-01
5.43819629e-02 -2.66751498e-01 1.20440252e-01 -5.07233679e-01
-8.48597109e-01 1.38084722e+00 1.23993778e+00 5.77164412e-01
1.21016920e+00 -2.00590402e-01 7.65412986e-01 3.21743786e-01
7.33928382e-01 -1.07017314e+00 8.00819620e-02 3.28682840e-01
5.83184600e-01 -1.31640697e+00 -2.51867950e-01 9.77892056e-02
-5.07458031e-01 9.88512695e-01 2.49247774e-01 -4.36210185e-01
8.48318636e-01 1.22074634e-01 -8.17018077e-02 -2.31574267e-01
-8.96313906e-01 -1.34778932e-01 2.96852887e-01 5.09441018e-01
4.29700673e-01 -7.12237135e-02 -4.39622462e-01 8.06639671e-01
1.80254020e-02 1.47219539e-01 4.17848796e-01 1.09406650e+00
-5.69313526e-01 -1.03161931e+00 -1.18706815e-01 7.81112254e-01
-6.09985828e-01 1.77290916e-01 -1.32868424e-01 5.04964828e-01
2.24587932e-01 9.36442137e-01 3.44908088e-02 -4.33439523e-01
2.92828947e-01 4.59941775e-01 2.25237772e-01 -8.98186088e-01
-3.54403287e-01 3.85063104e-02 -2.87618965e-01 -4.51788396e-01
-5.16187429e-01 -3.74789029e-01 -1.14843524e+00 1.05200380e-01
-2.42992267e-01 1.42729655e-01 3.62533569e-01 9.40055907e-01
1.92914590e-01 7.19692349e-01 7.28661299e-01 -6.87406003e-01
-5.11873305e-01 -9.81380820e-01 -1.07333136e+00 6.14521205e-01
4.42353845e-01 -8.31673205e-01 -5.24921775e-01 -1.16119303e-01] | [9.911816596984863, 3.1111841201782227] |
4f64e5e0-18c8-4800-88b2-45271970f776 | early-prediction-of-respiratory-failure-in | 2105.05728 | null | https://arxiv.org/abs/2105.05728v1 | https://arxiv.org/pdf/2105.05728v1.pdf | Early prediction of respiratory failure in the intensive care unit | The development of respiratory failure is common among patients in intensive care units (ICU). Large data quantities from ICU patient monitoring systems make timely and comprehensive analysis by clinicians difficult but are ideal for automatic processing by machine learning algorithms. Early prediction of respiratory system failure could alert clinicians to patients at risk of respiratory failure and allow for early patient reassessment and treatment adjustment. We propose an early warning system that predicts moderate/severe respiratory failure up to 8 hours in advance. Our system was trained on HiRID-II, a data-set containing more than 60,000 admissions to a tertiary care ICU. An alarm is typically triggered several hours before the beginning of respiratory failure. Our system outperforms a clinical baseline mimicking traditional clinical decision-making based on pulse-oximetric oxygen saturation and the fraction of inspired oxygen. To provide model introspection and diagnostics, we developed an easy-to-use web browser-based system to explore model input data and predictions visually. | ['Gunnar Rätsch', 'Tobias M. Merz', 'Stephanie L. Hyland', 'Chris Barber', 'Xinrui Lyu', 'Martin Faltys', 'Matthias Hüser'] | 2021-05-12 | null | null | null | null | ['respiratory-failure'] | ['medical'] | [ 4.14812624e-01 -8.19644108e-02 -3.31281088e-02 -3.84105295e-01
-1.66332006e-01 -4.79629159e-01 -9.57212523e-02 8.36826921e-01
-3.84721577e-01 7.01728463e-01 -5.41864224e-02 -1.15495312e+00
-4.76783127e-01 -6.54191196e-01 1.99680194e-01 -4.31878865e-01
-1.35899425e-01 1.01458287e+00 1.75823927e-01 3.05949032e-01
1.82536185e-01 9.40925419e-01 -8.53237092e-01 7.61391103e-01
5.00133038e-01 9.58863199e-01 3.53445411e-01 1.30123901e+00
1.58833265e-02 1.06846321e+00 -6.27477765e-01 4.71562833e-01
3.42580557e-01 -7.65324712e-01 -4.64740127e-01 -2.45236054e-01
-2.56019324e-01 -4.95119154e-01 -1.98463634e-01 1.29974410e-01
9.79766190e-01 -1.70469716e-01 5.27317345e-01 -1.22949696e+00
4.68205750e-01 2.56162763e-01 2.75336981e-01 6.02722168e-01
3.53794545e-01 7.60277331e-01 1.99443951e-01 -6.91007674e-01
-1.30770147e-01 6.76785827e-01 7.84469903e-01 6.87312543e-01
-9.70059752e-01 -2.68563330e-01 -3.33359510e-01 1.54977158e-01
-1.29739130e+00 -3.03018153e-01 -6.63768873e-03 -9.04244363e-01
9.87248659e-01 4.53849792e-01 5.87856948e-01 6.34977520e-01
5.19109249e-01 -3.01374793e-01 7.69828022e-01 -2.77994037e-01
1.65815830e-01 2.71145612e-01 1.56217143e-02 8.25816870e-01
2.58158147e-01 8.63780454e-02 -3.74544978e-01 -6.10288084e-01
1.10677433e+00 9.53650117e-01 -7.15252757e-01 -2.65332907e-02
-1.62318218e+00 3.78565103e-01 8.11828226e-02 -1.73986360e-01
-8.95338297e-01 -4.60306346e-01 2.45830670e-01 1.86211482e-01
-2.58585155e-01 4.82924670e-01 -8.86902869e-01 -5.07344961e-01
-3.57118130e-01 -6.28816605e-01 9.69550073e-01 6.09634101e-01
2.63804376e-01 -3.36619198e-01 -4.12468463e-01 6.97418690e-01
1.89749807e-01 2.06347555e-01 6.65620506e-01 -9.70125973e-01
1.73063979e-01 8.91410053e-01 5.19605398e-01 -2.45560884e-01
-8.34689736e-01 -5.02622686e-02 -9.38517511e-01 3.81472111e-01
3.56919706e-01 -5.02998650e-01 -5.78957140e-01 7.30161071e-01
1.91170096e-01 7.05999359e-02 1.07754759e-01 7.99201310e-01
4.36143696e-01 1.14629619e-01 3.31878096e-01 -7.26450324e-01
1.37789929e+00 -6.46560073e-01 -4.70469177e-01 4.65981080e-04
9.53826845e-01 -3.71095657e-01 8.01999092e-01 7.15228319e-01
-7.29564369e-01 -4.10277754e-01 -5.82486093e-01 6.05201066e-01
-1.06145926e-01 -1.78814173e-01 -3.44001055e-02 3.35889757e-01
-7.55006790e-01 1.21403289e+00 -1.07993865e+00 -8.08870912e-01
2.93601245e-01 3.64351094e-01 -2.08432049e-01 -4.55372967e-02
-7.74370015e-01 1.26558053e+00 4.28000093e-01 -9.69646797e-02
-3.41337651e-01 -8.14726174e-01 -5.27162850e-01 9.89114400e-03
-2.67650727e-02 -1.37473738e+00 1.24539077e+00 -8.80665258e-02
-1.22551429e+00 5.54960251e-01 -3.50502044e-01 -4.66898799e-01
6.19828701e-01 -3.41335863e-01 -4.73866373e-01 5.68191111e-01
-5.91975927e-01 -1.74030617e-01 5.15840828e-01 -7.53141642e-01
-8.31246018e-01 -2.84889013e-01 -4.08532768e-01 3.05599779e-01
-2.57140100e-02 3.19254100e-01 2.02376410e-01 -1.30431935e-01
-1.93814449e-02 -4.76210058e-01 -6.01649225e-01 2.04727128e-01
-8.97058398e-02 1.23194560e-01 8.86240661e-01 -5.00778496e-01
1.59805489e+00 -1.81125462e+00 -7.23072410e-01 -4.27643172e-02
5.42610765e-01 4.57224876e-01 5.00351965e-01 5.94607115e-01
-4.26330000e-01 3.05319756e-01 -2.26452276e-01 1.47072881e-01
-3.25702816e-01 3.65289360e-01 -7.66109824e-02 3.16451699e-01
2.30994225e-01 6.72054052e-01 -1.02480972e+00 -6.20876074e-01
9.53008235e-01 4.47955996e-01 -1.55629173e-01 1.21883488e+00
2.87130266e-01 9.75335300e-01 -2.90871054e-01 5.00000536e-01
-9.52879861e-02 -7.41878211e-01 1.81516737e-01 2.29976118e-01
-1.57013535e-01 3.72341424e-01 -6.23451352e-01 8.13564479e-01
-2.22887471e-01 2.80436546e-01 -1.47659808e-01 -6.46082878e-01
9.24705923e-01 8.86984468e-01 9.02991891e-01 2.38886825e-03
2.29984879e-01 -1.00974226e-03 1.28900900e-01 -1.02096593e+00
-6.01936340e-01 -5.28298497e-01 7.29841769e-01 6.50625527e-01
-8.01748931e-01 -2.00679958e-01 -6.61070272e-02 -1.03594303e-01
1.56957626e+00 -1.46445721e-01 1.00567865e+00 -1.01004273e-01
3.78233522e-01 1.37931779e-01 5.97176015e-01 7.48664975e-01
-4.53709602e-01 9.19362783e-01 1.85431480e-01 -8.50645602e-01
-6.89348817e-01 -9.73838508e-01 -3.14271688e-01 6.92251265e-01
-3.89069408e-01 -2.48818010e-01 -2.13925853e-01 -6.27393782e-01
8.18894282e-02 8.88612568e-01 -2.95060545e-01 -1.79893300e-01
-6.73977792e-01 -6.58400118e-01 -2.85467785e-02 7.13157237e-01
-2.39725247e-01 -1.24620259e+00 -1.39280081e+00 8.07914674e-01
-6.35143518e-02 -7.12088645e-01 -8.06387514e-03 5.89482903e-01
-1.28999853e+00 -1.56098640e+00 -6.57597542e-01 -4.13629174e-01
6.00392222e-01 5.62404320e-02 1.17145121e+00 5.28731883e-01
-9.16309416e-01 4.19363678e-01 -1.02083243e-01 -8.30161691e-01
-8.28967929e-01 -3.41834903e-01 5.07463336e-01 -3.37779194e-01
6.42254710e-01 -5.33480644e-01 -1.35478079e+00 6.82881415e-01
-5.19570172e-01 2.49711066e-01 3.32198620e-01 4.51730192e-01
6.66184247e-01 -4.28843051e-01 6.55413806e-01 -8.55062068e-01
7.29315102e-01 -7.44062483e-01 -2.47932360e-01 2.70302087e-01
-1.41099250e+00 -3.14994514e-01 1.03796351e+00 -1.93480045e-01
-3.62977147e-01 8.90574381e-02 1.66391283e-01 -4.51419085e-01
-7.51710474e-01 9.85230356e-02 4.32497412e-01 5.38791776e-01
7.64676273e-01 2.87550651e-02 1.57748580e-01 -6.60090923e-01
-4.59563583e-01 1.14420736e+00 6.12659454e-01 6.20708205e-02
4.51083004e-01 -8.83617401e-02 2.29961321e-01 -4.09478188e-01
-5.20747304e-01 -1.06595588e+00 -1.10585225e+00 -2.92092592e-01
9.20772910e-01 -6.25717580e-01 -1.13605571e+00 9.17038992e-02
-9.14756477e-01 -6.00041687e-01 -4.10651088e-01 7.67349958e-01
-3.27894419e-01 4.13275748e-01 -3.36210102e-01 -7.60499299e-01
-6.21362686e-01 -5.60923755e-01 4.26844150e-01 9.18714106e-02
-7.99565315e-01 -1.31415415e+00 4.08616602e-01 8.60925764e-02
5.41594207e-01 6.51809454e-01 9.76582408e-01 -1.07401073e+00
-6.02287240e-02 -3.61736089e-01 -6.71842545e-02 3.63364041e-01
9.53162134e-01 3.58228773e-01 -6.47414625e-01 -2.85155803e-01
1.03711814e-01 8.02379996e-02 2.73215920e-01 9.03420448e-02
1.35236275e+00 -3.01045597e-01 -5.47532141e-01 5.20317316e-01
1.22436583e+00 7.11113036e-01 2.25367397e-01 3.91106755e-02
6.08163059e-01 3.46933991e-01 4.92293239e-01 1.22487795e+00
2.12716773e-01 1.27163574e-01 4.11775529e-01 -2.79282987e-01
3.45354617e-01 2.18837932e-01 8.87445807e-02 7.78237402e-01
-1.33295372e-01 -3.00278276e-01 -1.49054003e+00 1.67223617e-01
-1.56524134e+00 -5.37565172e-01 -5.97446859e-01 2.61892867e+00
8.93976688e-01 -1.59384370e-01 1.99753642e-01 1.22037679e-01
8.14165473e-01 -7.20168412e-01 -6.69193268e-01 -4.02696550e-01
5.10171711e-01 2.21798286e-01 1.54708564e-01 4.27865684e-01
-5.98677814e-01 1.45255744e-01 6.91463804e+00 -6.91315174e-01
-1.19174635e+00 -2.58649945e-01 6.63681090e-01 -1.44056156e-01
5.68066120e-01 -2.49433294e-01 -3.86765391e-01 3.87247980e-01
1.49627781e+00 -2.68331200e-01 3.70535791e-01 8.87554407e-01
8.61944497e-01 -1.02682739e-01 -1.67345369e+00 1.10811007e+00
-3.31615955e-01 -1.19841075e+00 -3.39361876e-01 -2.32719898e-01
-5.25236391e-02 1.98019862e-01 -6.70993805e-01 -1.32936403e-01
2.25960031e-01 -1.18282294e+00 -1.14234224e-01 9.56722140e-01
1.30844450e+00 -2.07584441e-01 8.26054573e-01 4.76941347e-01
-8.75783265e-01 -3.75235915e-01 -1.59751028e-01 -2.75940776e-01
4.19489801e-01 3.36708188e-01 -1.94721377e+00 -8.77190530e-02
4.64998245e-01 2.89413959e-01 -4.92445350e-01 1.37401891e+00
-8.61449242e-02 6.56969965e-01 -5.50899744e-01 2.90321231e-01
-4.66675818e-01 1.72332644e-01 3.70363712e-01 1.21842611e+00
4.72501129e-01 9.65953290e-01 1.56690404e-01 3.25959474e-01
3.27942848e-01 1.92280054e-01 -7.77747095e-01 4.39428911e-03
5.06395161e-01 1.08861780e+00 -5.01220167e-01 -7.27367342e-01
-3.23204428e-01 8.19509923e-01 -1.22964710e-01 2.61606961e-01
-4.35202807e-01 -2.96705574e-01 4.21062887e-01 7.73577452e-01
-1.91763565e-01 7.18246698e-02 -7.04797924e-01 -7.29605138e-01
-4.51416403e-01 -6.30262673e-01 7.60372639e-01 -8.86514902e-01
-9.78805482e-01 8.72237504e-01 -2.92506546e-01 -1.45055556e+00
-7.20202684e-01 -8.67616475e-01 -9.79907334e-01 1.22246730e+00
-1.48705316e+00 1.82585448e-01 -8.85726035e-01 8.00161898e-01
3.71701956e-01 1.12544401e-02 1.68559730e+00 -6.21861853e-02
-7.04685032e-01 -9.14263353e-02 -1.43882379e-01 -5.86381294e-02
8.86660457e-01 -1.40500712e+00 3.97114344e-02 2.85022706e-01
-9.74278688e-01 7.30353951e-01 6.04248941e-01 -6.96414113e-01
-9.26107466e-01 -1.17363286e+00 9.31301117e-01 -9.90266323e-01
4.06833142e-01 4.43840772e-01 -1.30273199e+00 3.65176141e-01
2.92829890e-02 1.68958515e-01 1.25813377e+00 -6.70038342e-01
3.57619435e-01 5.81944473e-02 -1.10115790e+00 3.44456345e-01
5.78528821e-01 -1.31436870e-01 -7.87666619e-01 6.35583043e-01
2.92091012e-01 -1.34571910e-01 -1.46191418e+00 7.24335194e-01
2.70271182e-01 -8.09672117e-01 5.79357743e-01 -9.77728963e-01
-1.51342284e-02 -2.45390385e-01 3.10256869e-01 -1.18523538e+00
-5.36330640e-01 -1.08288264e+00 -2.52723664e-01 3.34475935e-01
2.46029317e-01 -1.08595634e+00 4.90387827e-01 1.22064531e+00
-1.04616798e-01 -8.83931994e-01 -4.46041256e-01 -2.69265354e-01
-3.05237263e-01 -1.64725378e-01 3.97685975e-01 8.51613283e-01
4.94487226e-01 4.30673778e-01 -6.43411726e-02 1.70480475e-01
3.48281473e-01 -5.13741039e-02 5.75849712e-01 -1.60955036e+00
-2.12939486e-01 -3.09303761e-01 1.02671310e-01 -3.82918596e-01
-1.02381098e+00 -4.40571994e-01 8.56878906e-02 -2.21326804e+00
4.19687591e-02 -4.40942317e-01 -9.65250373e-01 8.64323258e-01
-4.69336867e-01 -2.25792248e-02 -2.39545360e-01 3.41303349e-01
-1.49843872e-01 -1.57168567e-01 1.00370681e+00 5.84730268e-01
-6.24014616e-01 4.13007021e-01 -3.64730835e-01 7.81523764e-01
1.25722313e+00 -8.96140754e-01 -3.59638423e-01 1.57632977e-01
-1.09740041e-01 6.63470268e-01 2.17075691e-01 -9.42874432e-01
2.26332054e-01 -8.07087898e-01 6.35829329e-01 -5.77905476e-01
1.97255481e-02 -1.22933388e+00 3.55969429e-01 1.07963884e+00
-2.38912571e-02 4.25186515e-01 2.29362324e-01 1.74828917e-01
7.79899210e-02 -2.22573597e-02 9.73631561e-01 -3.60716283e-01
-1.84796780e-01 5.18744349e-01 -1.09513986e+00 -2.70521175e-02
1.12587452e+00 -3.52078497e-01 -4.98263657e-01 -3.53335053e-01
-9.65466440e-01 2.34455451e-01 5.68937123e-01 2.46837139e-01
8.21785271e-01 -8.51702094e-01 -8.61318409e-01 3.85782242e-01
5.45450211e-01 -8.32464173e-02 3.77100334e-02 1.25692523e+00
-1.22198236e+00 3.08160841e-01 -5.06748147e-02 -6.55988455e-01
-1.49875081e+00 7.99354613e-01 5.65415800e-01 -1.02242716e-02
-9.52198386e-01 5.96362233e-01 2.50156254e-01 1.79696903e-01
5.05191207e-01 -3.94165635e-01 -1.43556759e-01 -3.33724052e-01
8.68686914e-01 4.36642349e-01 -2.31083296e-02 -1.28830388e-01
-7.09894180e-01 8.01751763e-02 1.74404368e-01 3.36450636e-01
1.14779127e+00 -3.07951778e-01 -2.78751459e-02 6.88748777e-01
8.16580594e-01 -4.34825271e-01 -1.25296426e+00 -1.78844750e-01
-1.07059576e-01 -3.91882300e-01 -3.32878619e-01 -1.24930263e+00
-3.82192105e-01 1.20924473e+00 9.24735129e-01 4.45482552e-01
1.06494081e+00 -1.96841940e-01 5.13744712e-01 7.71488011e-01
-2.32402608e-02 -8.08184862e-01 1.40234595e-03 2.50297040e-01
7.32673049e-01 -1.31571293e+00 -8.25082064e-02 7.75283994e-03
-7.50325561e-01 1.60108721e+00 4.61703062e-01 5.04623890e-01
6.37316048e-01 1.57655403e-01 9.20746624e-01 -2.43693609e-02
-1.31690192e+00 4.48026419e-01 3.95612232e-02 7.26811707e-01
4.49264199e-01 2.18754277e-01 1.98748827e-01 6.04217112e-01
1.61172763e-01 2.96655327e-01 3.21907014e-01 1.16504979e+00
-9.42930222e-01 -8.93261313e-01 -5.20086348e-01 8.33856761e-01
-6.70616090e-01 -3.19514185e-01 -3.39472592e-01 2.42928997e-01
-2.49620527e-01 1.17760313e+00 -7.51924003e-04 -1.81114674e-01
5.31591833e-01 7.87950814e-01 1.65840939e-01 -8.02374423e-01
-6.18355572e-01 -2.08802104e-01 7.26089701e-02 -4.15186793e-01
1.98508263e-01 -3.13344598e-01 -1.68905020e+00 6.07905053e-02
1.28797069e-01 2.00904772e-01 6.22614622e-01 5.63382387e-01
6.25242233e-01 8.52081776e-01 9.10120189e-01 -4.21929151e-01
-5.32697737e-01 -1.22394764e+00 -4.87899214e-01 1.56218037e-01
7.24650741e-01 -1.67922929e-01 -7.35362172e-01 5.16841710e-01] | [7.988918781280518, 6.160658359527588] |
ed09331d-2052-4e4f-90af-0c7ed25705a0 | comparison-of-deep-learning-models-on-time | 1911.08414 | null | https://arxiv.org/abs/1911.08414v2 | https://arxiv.org/pdf/1911.08414v2.pdf | Comparison of Deep learning models on time series forecasting : a case study of Dissolved Oxygen Prediction | Deep learning has achieved impressive prediction performance in the field of sequence learning recently. Dissolved oxygen prediction, as a kind of time-series forecasting, is suitable for this technique. Although many researchers have developed hybrid models or variant models based on deep learning techniques, there is no comprehensive and sound comparison among the deep learning models in this field currently. Plus, most previous studies focused on one-step forecasting by using a small data set. As the convenient access to high-frequency data, this paper compares multi-step deep learning forecasting by using walk-forward validation. Specifically, we test Convolutional Neural Network (CNN), Temporal Convolutional Network (TCN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Bidirectional Recurrent Neural Network (BiRNN) based on the real-time data recorded automatically at a fixed observation point in the Yangtze River from 2012 to 2016. By comparing the average accumulated statistical metrics of root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination in each time step, We find for multi-step time series forecasting, the average performance of each time step does not decrease linearly. GRU outperforms other models with significant advantages. | ['Hongqian Qin'] | 2019-11-17 | null | null | null | null | ['small-data'] | ['computer-vision'] | [-3.13037068e-01 -6.65722013e-01 1.46227568e-01 -2.79753655e-01
-2.35778213e-01 -2.52099425e-01 7.68222213e-01 -1.94924586e-02
-3.64798039e-01 9.50403273e-01 2.27971748e-01 -8.99783492e-01
-3.89192402e-01 -1.02649426e+00 -3.30570281e-01 -1.04051137e+00
-6.34431005e-01 -1.79947644e-01 -9.13037881e-02 -3.89330298e-01
5.16503811e-01 6.53798878e-01 -1.40597200e+00 1.35709226e-01
9.22095656e-01 1.15146470e+00 1.84793651e-01 7.73682237e-01
-4.35598582e-01 1.06895685e+00 -7.16176569e-01 2.23669350e-01
1.79424092e-01 -5.12830436e-01 -6.23388827e-01 -5.75732946e-01
-2.34034449e-01 -2.98955113e-01 -3.13785315e-01 4.50255483e-01
8.41092110e-01 3.22356492e-01 5.09658694e-01 -9.87515986e-01
-6.67662084e-01 4.43820894e-01 -2.88507551e-01 6.90175056e-01
-2.00001329e-01 2.87382722e-01 4.26462501e-01 -6.86675429e-01
1.22491285e-01 9.21591222e-01 1.23907971e+00 1.70200154e-01
-8.00551832e-01 -8.09629977e-01 -4.02092636e-02 1.89925343e-01
-1.02860892e+00 -4.10594754e-02 6.41710401e-01 -7.83599615e-01
1.46311355e+00 -6.31280541e-02 8.39594603e-01 9.07618463e-01
1.02571976e+00 2.20319003e-01 1.02175200e+00 -7.73448274e-02
1.60485044e-01 -2.23180518e-01 2.15693280e-01 1.81642100e-01
2.97853090e-02 6.12249553e-01 -1.58236593e-01 1.15305074e-01
4.78880614e-01 5.94220340e-01 -2.25407839e-01 6.46552980e-01
-9.67291474e-01 8.76808882e-01 6.54345870e-01 6.66334033e-01
-8.16455245e-01 -7.27889016e-02 8.28157365e-01 7.87123680e-01
8.33426297e-01 9.31987837e-02 -1.00086319e+00 -2.46350199e-01
-9.77901697e-01 3.75755408e-05 9.21208084e-01 2.79670060e-01
5.21225750e-01 9.05210555e-01 -7.53262267e-02 6.91100955e-01
4.07624960e-01 6.30298674e-01 1.22775698e+00 -4.52894628e-01
2.15542972e-01 4.24041957e-01 1.20959818e-01 -1.14332628e+00
-8.16732585e-01 -5.66964805e-01 -1.51908648e+00 1.84106112e-01
1.45725235e-01 -7.97046721e-01 -7.42936134e-01 1.10924947e+00
-9.29087326e-02 3.57661933e-01 5.83749473e-01 6.02886617e-01
1.06882536e+00 1.56301546e+00 1.30959466e-01 -5.42488337e-01
8.64165068e-01 -1.02598560e+00 -8.50176096e-01 1.49122015e-01
8.73272836e-01 -4.40238476e-01 6.55752718e-01 2.87018448e-01
-4.62336361e-01 -7.45075881e-01 -9.29378390e-01 1.39536843e-01
-1.03855717e+00 8.01714584e-02 4.73942727e-01 4.24987793e-01
-1.11380363e+00 1.03257287e+00 -8.58858645e-01 -4.38123524e-01
-1.09100200e-01 2.28602827e-01 -2.96625257e-01 3.76975358e-01
-1.55215466e+00 1.03989732e+00 2.97520578e-01 7.78478563e-01
-9.66892838e-01 -6.49075031e-01 -5.64003766e-01 1.34195253e-01
-5.06560147e-01 -1.12753890e-01 1.09763634e+00 -8.99055958e-01
-1.84026897e+00 1.51196942e-01 -1.50379375e-01 -8.28571677e-01
2.44058892e-01 -2.02222720e-01 -7.79640496e-01 -4.33904350e-01
-2.21148208e-01 1.06183253e-01 2.96968371e-01 -7.21220791e-01
-6.65916800e-01 -2.20238566e-01 -4.57374871e-01 -3.60508919e-01
-2.50121027e-01 -4.81237732e-02 8.26126754e-01 -5.21501184e-01
-4.65998836e-02 -6.71436608e-01 -3.22475284e-01 -4.56531733e-01
1.92118973e-01 -6.75716877e-01 1.11837268e+00 -1.00587869e+00
1.40052009e+00 -2.11712956e+00 -3.94149929e-01 -1.18869267e-01
-3.66641223e-01 7.48434484e-01 -9.02646035e-02 8.25883985e-01
-2.14182496e-01 3.29156786e-01 -1.96106389e-01 6.30577058e-02
-3.32594454e-01 4.32107359e-01 -6.26685858e-01 4.49914366e-01
-1.05239134e-02 7.12404191e-01 -9.43726599e-01 7.39934295e-02
5.13807058e-01 6.74818516e-01 1.98816314e-01 3.91725361e-01
2.42327377e-02 5.84599912e-01 -1.84855759e-01 3.88007313e-01
6.34658873e-01 -1.96213052e-01 -1.30522370e-01 -2.64838580e-02
-9.11287308e-01 3.10278207e-01 -6.88165605e-01 1.17814720e+00
-7.52701044e-01 1.04786718e+00 -6.83524489e-01 -1.26279128e+00
1.57315350e+00 7.39988506e-01 4.71622556e-01 -9.02874231e-01
9.53671802e-03 5.66339552e-01 -3.48333046e-02 -9.20029759e-01
1.30255729e-01 -2.52723724e-01 5.83545029e-01 1.94945961e-01
-1.04153439e-01 4.70326751e-01 9.49763581e-02 -6.88693881e-01
7.23026931e-01 1.91377684e-01 2.15962991e-01 -2.64200538e-01
5.19324064e-01 1.74320936e-02 4.79674071e-01 5.38645029e-01
-1.15635797e-01 4.20429051e-01 2.43646502e-01 -1.20472753e+00
-1.16059744e+00 -1.64733559e-01 -3.50213081e-01 9.88728464e-01
-2.82813311e-01 -4.78700688e-03 2.15751436e-02 -1.28423376e-02
-5.80722727e-02 6.50049746e-01 -7.16614008e-01 5.08420318e-02
-5.69147706e-01 -1.16358089e+00 6.42012119e-01 6.22480869e-01
8.11866939e-01 -1.51181614e+00 -7.69898832e-01 7.40809739e-01
2.31005222e-01 -4.60037589e-01 2.00389996e-01 5.43843150e-01
-1.34502089e+00 -6.24056578e-01 -8.16976488e-01 -7.34759152e-01
-6.65384009e-02 1.36852220e-01 1.01899374e+00 -6.92401007e-02
2.85656244e-01 -3.65404457e-01 -4.20491040e-01 -4.19053018e-01
-1.73419818e-01 1.45149305e-01 -1.40826300e-01 -1.99837983e-01
6.08708441e-01 -7.08628535e-01 -5.95873117e-01 6.93516880e-02
-6.61581457e-01 -2.97999352e-01 4.65744525e-01 9.49763119e-01
2.32759491e-01 3.74453850e-02 8.82521510e-01 -3.32190067e-01
6.01257443e-01 -1.04761708e+00 -7.03873217e-01 1.58845171e-01
-9.76992607e-01 -1.77568823e-01 1.01949775e+00 -3.15939546e-01
-7.95050263e-01 -3.33342820e-01 -2.66283959e-01 -4.16809708e-01
-2.45749094e-02 1.22671592e+00 7.08171487e-01 2.03018039e-01
6.27447903e-01 7.87216067e-01 4.97879311e-02 -6.94925666e-01
-2.41992265e-01 9.64981675e-01 2.13189647e-01 2.60591507e-02
1.80066884e-01 1.57175452e-01 -7.82866701e-02 -7.90950179e-01
-6.31352842e-01 -2.65059561e-01 -7.16310680e-01 -3.93347353e-01
9.41930354e-01 -1.13280082e+00 -1.16264987e+00 7.92317867e-01
-1.25533772e+00 -6.91423178e-01 1.51195943e-01 8.60562205e-01
6.35226145e-02 5.65762483e-02 -8.22745383e-01 -1.09503663e+00
-8.51007402e-01 -7.96875358e-01 6.40871108e-01 3.62158120e-01
1.87528014e-01 -1.56443191e+00 3.90530765e-01 -3.91358942e-01
1.10803103e+00 6.38077140e-01 6.15757585e-01 -7.58722544e-01
4.92826439e-02 -4.20655459e-01 -8.62910002e-02 6.34731293e-01
2.46793240e-01 5.13686478e-01 -8.70020747e-01 -2.91085243e-01
7.48421550e-02 7.31954491e-03 8.75828445e-01 8.26711476e-01
9.94880378e-01 -5.84524810e-01 -5.13583720e-02 7.05870450e-01
1.65880764e+00 8.97926927e-01 9.41506445e-01 7.67739952e-01
5.18456042e-01 2.53710657e-01 1.41897112e-01 7.06296623e-01
3.84833276e-01 -6.92448318e-02 3.75184894e-01 -3.15650627e-02
3.71158242e-01 1.70202419e-01 6.86098099e-01 1.00391090e+00
-3.65704894e-01 -5.63422859e-01 -1.25774539e+00 6.15817904e-01
-1.81262970e+00 -1.34850848e+00 -5.51853538e-01 2.04158425e+00
3.14744413e-01 -1.62385255e-01 -6.78295875e-03 4.18045849e-01
6.09034896e-01 3.57987881e-01 -5.20315170e-01 -6.61651015e-01
-2.63473332e-01 7.19102174e-02 7.14865625e-01 3.09262186e-01
-1.06639707e+00 5.50727129e-01 6.59173727e+00 2.94378847e-01
-2.04994202e+00 8.59421417e-02 6.23062909e-01 3.00262243e-01
8.92359316e-02 -3.37192237e-01 -8.00824583e-01 7.96039224e-01
1.86637104e+00 -1.75102383e-01 2.53560334e-01 8.54419053e-01
7.56687403e-01 1.66666701e-01 -6.16555274e-01 6.83417797e-01
-2.57510543e-01 -1.35132587e+00 -4.81665209e-02 2.68073045e-02
8.77014756e-01 7.06041932e-01 -1.10161893e-01 7.05400884e-01
1.26269668e-01 -1.26321971e+00 1.63918212e-01 1.00130081e+00
3.08437735e-01 -5.95889091e-01 1.46532750e+00 3.47857475e-01
-1.25294054e+00 -4.25710440e-01 -5.11327028e-01 -6.08586013e-01
-1.19460486e-01 7.20062375e-01 -3.54857236e-01 6.27184689e-01
9.36254084e-01 1.33579612e+00 -2.06939578e-01 1.06987464e+00
3.13440084e-01 1.01276577e+00 -2.80181885e-01 -3.45824659e-01
7.84849465e-01 -4.17877316e-01 -7.83474930e-03 1.23537433e+00
8.48955572e-01 1.50396898e-01 7.43357390e-02 2.48722062e-01
3.97613764e-01 1.80361897e-01 -7.49223471e-01 -2.37350598e-01
3.32101017e-01 8.41122091e-01 -2.71101296e-01 -5.06954491e-01
-5.47079623e-01 3.59668434e-01 -1.85070530e-01 4.25460339e-01
-5.71937025e-01 -6.88201308e-01 5.19002914e-01 -2.84794658e-01
2.68679857e-01 -3.64726126e-01 -4.97480899e-01 -8.32737327e-01
-1.62914693e-01 -3.93473297e-01 3.35030824e-01 -8.81182730e-01
-1.48032618e+00 9.53835666e-01 -3.48447531e-01 -1.61464536e+00
-4.85477328e-01 -6.65547967e-01 -9.93679643e-01 1.31232512e+00
-1.84230459e+00 -7.70673156e-01 -5.94143748e-01 3.44471008e-01
6.81641281e-01 -4.90107208e-01 9.21952426e-01 5.73141038e-01
-7.65792787e-01 6.74249753e-02 8.44449103e-01 2.46064380e-01
2.11441800e-01 -9.05371189e-01 2.48267025e-01 7.11318851e-01
-6.36149228e-01 4.62646335e-01 6.19769633e-01 -5.03432035e-01
-1.16489255e+00 -1.34360123e+00 1.10751784e+00 6.45127445e-02
6.83369458e-01 3.54787409e-01 -1.24544847e+00 8.70583355e-01
3.38261306e-01 2.81580091e-01 6.37638211e-01 -1.26993150e-01
2.44633369e-02 -5.16507626e-01 -8.48668516e-01 5.66147715e-02
2.65560180e-01 -3.05786371e-01 -2.77195007e-01 3.42731923e-01
6.11034334e-01 -2.13096455e-01 -1.28630030e+00 7.61221707e-01
8.15007746e-01 -9.46643412e-01 5.12087166e-01 -6.31230772e-01
7.25651443e-01 -3.85133803e-01 -1.49588212e-01 -1.43114686e+00
-5.08832872e-01 -3.11619997e-01 -5.23634180e-02 9.14169729e-01
3.57090563e-01 -9.64840174e-01 4.11428541e-01 7.44023100e-02
-5.33752978e-01 -7.92322874e-01 -7.75830805e-01 -9.12296593e-01
3.89118642e-01 -2.27207080e-01 8.37984622e-01 1.14176273e+00
-2.53379256e-01 1.94979429e-01 -8.16157639e-01 2.66726404e-01
-1.76773354e-01 2.09730178e-01 6.01344824e-01 -1.40381634e+00
3.99055272e-01 -5.08049130e-01 -3.76600206e-01 -8.44116569e-01
-7.21750781e-02 -4.97907251e-01 -2.91898083e-02 -1.94597173e+00
-4.11720753e-01 -2.03755125e-01 -7.02679217e-01 6.87570214e-01
1.42988309e-01 -9.08891782e-02 -2.44160533e-01 3.44455540e-01
1.49856880e-01 7.58628547e-01 9.57125843e-01 -9.76373628e-02
-4.13705975e-01 1.11948743e-01 3.37562114e-02 2.97342539e-01
1.08257437e+00 -3.17697287e-01 -1.07048690e-01 -5.30546248e-01
1.84832275e-01 4.67298865e-01 4.64506522e-02 -1.24797249e+00
1.95942089e-01 -2.25430727e-01 5.05034685e-01 -9.63742137e-01
-2.77171522e-01 -7.84751117e-01 4.88869488e-01 9.42820311e-01
-3.30434680e-01 6.50537014e-01 3.47965807e-01 2.42926583e-01
-5.86571574e-01 -2.79060509e-02 5.29948592e-01 -1.69814020e-01
-8.65345299e-01 4.61367339e-01 -7.32674360e-01 -5.93431473e-01
7.66784310e-01 -3.42216700e-01 -3.27724367e-01 -1.93742663e-01
-6.26943350e-01 3.12603652e-01 -3.63353521e-01 4.29375648e-01
3.28458726e-01 -1.12880814e+00 -8.68706465e-01 1.32701069e-01
-1.81107000e-01 -2.09519431e-01 4.82252419e-01 8.90608013e-01
-9.14396584e-01 5.74038744e-01 -4.25767541e-01 -5.93513846e-01
-6.84759676e-01 4.18242365e-01 9.30179894e-01 -1.86386496e-01
-6.96564496e-01 5.19526422e-01 -3.26510221e-01 -5.15710115e-01
4.01906297e-02 -5.24410486e-01 -7.06128776e-01 3.97491813e-01
5.92015326e-01 5.67244947e-01 2.54527211e-01 -4.00870889e-01
-1.65043741e-01 6.56544685e-01 4.55339402e-01 3.46517950e-01
1.79772270e+00 -3.85571942e-02 -2.93683141e-01 1.17634833e+00
1.51252699e+00 -6.86714828e-01 -1.07741702e+00 3.05056684e-02
5.96435182e-02 2.35422980e-02 1.06196307e-01 -7.30014503e-01
-1.28461850e+00 1.24267924e+00 1.11347353e+00 7.78553784e-01
1.08057380e+00 -9.32732701e-01 9.38849211e-01 5.24234056e-01
-6.80465847e-02 -5.57736576e-01 -3.72252315e-01 1.14039791e+00
8.50781620e-01 -1.44194734e+00 -4.06594574e-01 6.11382544e-01
-3.63208055e-01 1.66440940e+00 5.03055453e-01 -2.20643818e-01
1.21942973e+00 1.38993308e-01 3.87318790e-01 -4.07424867e-02
-1.12504101e+00 8.05929229e-02 -2.06969351e-01 2.62552142e-01
9.55500782e-01 -9.84475166e-02 -3.23423207e-01 1.91930175e-01
-4.36034679e-01 3.07455450e-01 4.08047438e-01 6.99274421e-01
-6.04586780e-01 -4.18241739e-01 -1.45611137e-01 5.31357288e-01
-5.03066301e-01 -1.54592738e-01 3.35384160e-01 6.77157640e-01
1.43842166e-02 1.09748960e+00 4.41195846e-01 -5.69392622e-01
5.64967468e-02 2.28024930e-01 -3.57492805e-01 -9.01219323e-02
-9.02910709e-01 -3.01313996e-01 -3.07248384e-01 -1.49438128e-01
-6.69976711e-01 -4.08962667e-01 -1.13301992e+00 -7.19552875e-01
-9.58911851e-02 4.82411057e-01 9.56911802e-01 1.15075910e+00
3.11990499e-01 7.18815565e-01 1.09931195e+00 -1.04329109e+00
-3.01085562e-01 -1.67914438e+00 -6.59221947e-01 -4.32726294e-02
7.83153236e-01 -2.76130885e-01 -6.88373864e-01 3.35097127e-02] | [6.513217449188232, 2.9082977771759033] |
91d5e2a2-3e92-4dff-acf8-ff25b000b57b | structural-encoding-and-pre-training-matter | null | null | https://aclanthology.org/2021.eacl-main.201 | https://aclanthology.org/2021.eacl-main.201.pdf | Structural Encoding and Pre-training Matter: Adapting BERT for Table-Based Fact Verification | Growing concern with online misinformation has encouraged NLP research on fact verification. Since writers often base their assertions on structured data, we focus here on verifying textual statements given evidence in tables. Starting from the Table Parsing (TAPAS) model developed for question answering (Herzig et al., 2020), we find that modeling table structure improves a language model pre-trained on unstructured text. Pre-training language models on English Wikipedia table data further improves performance. Pre-training on a question answering task with column-level cell rank information achieves the best performance. With improved pre-training and cell embeddings, this approach outperforms the state-of-the-art Numerically-aware Graph Neural Network table fact verification model (GNN-TabFact), increasing statement classification accuracy from 72.2{\%} to 73.9{\%} even without modeling numerical information. Incorporating numerical information with cell rankings and pre-training on a question-answering task increases accuracy to 76{\%}. We further analyze accuracy on statements implicating single rows or multiple rows and columns of tables, on different numerical reasoning subtasks, and on generalizing to detecting errors in statements derived from the ToTTo table-to-text generation dataset. | ['David Smith', 'Rui Dong'] | 2021-04-01 | null | null | null | eacl-2021-2 | ['table-to-text-generation', 'table-based-fact-verification'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.22470528e-01 8.84915769e-01 -4.45963115e-01 -3.82799387e-01
-1.14939356e+00 -7.09860682e-01 6.22217298e-01 1.27272046e+00
-1.95971683e-01 9.66083586e-01 6.59633815e-01 -1.10966980e+00
2.19739843e-02 -1.32061625e+00 -1.10602236e+00 5.05176604e-01
2.18238682e-02 6.25272572e-01 4.07350250e-02 -4.41990256e-01
4.48606402e-01 -3.59983221e-02 -9.36430812e-01 1.00384295e+00
1.06926858e+00 9.78145063e-01 -6.96995080e-01 7.52881408e-01
-7.01790631e-01 1.94472134e+00 -6.58081532e-01 -1.51630509e+00
-6.50228187e-02 -2.08298057e-01 -1.08976579e+00 -3.12218726e-01
1.10075545e+00 -4.21266764e-01 -3.37551653e-01 1.03061438e+00
-3.08163278e-02 -1.63151175e-01 7.51468539e-01 -1.18051755e+00
-1.34674215e+00 1.59370887e+00 -2.89499432e-01 3.66186172e-01
8.30306470e-01 -1.99934065e-01 1.74664819e+00 -8.52964878e-01
8.95348549e-01 1.12571430e+00 1.11916614e+00 2.00422451e-01
-1.21186900e+00 -3.68997186e-01 2.19958678e-01 4.10622895e-01
-9.62329030e-01 -2.45215878e-01 8.58769715e-01 -4.26373601e-01
1.57811642e+00 4.54759032e-01 3.78237486e-01 9.08663034e-01
4.20852691e-01 6.80332839e-01 1.02839208e+00 -4.91421014e-01
-2.12370791e-03 2.98376411e-01 4.66910213e-01 9.48311031e-01
1.00316548e+00 -5.24549663e-01 -6.61005676e-01 -1.25626355e-01
2.69496709e-01 -7.17186689e-01 -7.80644193e-02 -1.61517426e-01
-1.21590090e+00 1.00553322e+00 5.49415886e-01 2.28684381e-01
-2.40839019e-01 1.99861646e-01 7.15038836e-01 4.53365207e-01
5.27373731e-01 8.50120664e-01 -7.42567122e-01 -3.62685286e-02
-9.16716814e-01 6.54048741e-01 1.43785846e+00 7.92040884e-01
3.62778366e-01 5.29421456e-02 -2.29251206e-01 6.23881876e-01
5.33706956e-02 1.99385360e-01 2.85712004e-01 -8.03577125e-01
1.44248021e+00 1.11265624e+00 4.87908944e-02 -1.61556351e+00
-6.41397357e-01 -6.09356582e-01 -6.35410130e-01 -2.41094932e-01
8.99926305e-01 -1.55450240e-01 -5.64104140e-01 1.32351875e+00
4.13932912e-02 -8.17774594e-01 3.96982878e-01 4.61171955e-01
1.11106849e+00 3.86877120e-01 -8.14596713e-02 1.88393608e-01
1.45421541e+00 -6.79974914e-01 -8.55971515e-01 -3.56445521e-01
1.26027989e+00 -2.60095030e-01 1.08603108e+00 5.10642350e-01
-9.39188659e-01 -3.63449723e-01 -1.25977528e+00 -7.67179728e-01
-9.22697365e-01 1.20812990e-01 7.52895772e-01 8.62960994e-01
-1.01455510e+00 4.33343977e-01 -5.08626521e-01 1.29600361e-04
4.59284753e-01 -2.66482346e-02 -3.96763146e-01 -2.26442099e-01
-1.54310083e+00 1.12151635e+00 5.64818859e-01 -6.50107190e-02
-1.61171660e-01 -1.11778915e+00 -1.33303154e+00 1.86498716e-01
8.33360314e-01 -5.51966846e-01 1.10402358e+00 -3.24976474e-01
-8.34937215e-01 8.38832319e-01 -7.35590830e-02 -1.02160335e+00
4.87494618e-01 -5.67721575e-02 -6.76224709e-01 -2.25407369e-02
4.04725015e-01 3.96523058e-01 4.38251123e-02 -1.21520483e+00
-2.35564247e-01 -3.54075909e-01 4.38727230e-01 -5.12213968e-02
-2.26185232e-01 -1.89284787e-01 1.77946031e-01 -5.87983608e-01
2.19272465e-01 -3.54738086e-01 1.80276066e-01 -2.04256326e-01
-9.02499974e-01 -3.09913456e-01 2.04101071e-01 -1.50058413e+00
1.43121362e+00 -1.30209827e+00 -3.56852889e-01 3.28897446e-01
4.40566808e-01 -2.70748079e-01 1.90347254e-01 6.09951079e-01
-2.25983873e-01 5.84030092e-01 -2.58443773e-01 1.08243093e-01
4.01002079e-01 -6.67118877e-02 -4.41172332e-01 -1.24430329e-01
6.33527875e-01 1.22860670e+00 -6.37070119e-01 -6.11479640e-01
-2.54479587e-01 -9.00707096e-02 -9.65398729e-01 -3.63768548e-01
-6.36152983e-01 -4.22163248e-01 -4.67594191e-02 8.52182567e-01
5.41118801e-01 -5.88712931e-01 5.21080494e-01 -1.86790660e-01
3.60140920e-01 1.14371228e+00 -1.21412480e+00 1.27266395e+00
-5.35652637e-01 6.65102422e-01 -3.23132545e-01 -9.75093246e-01
9.99127924e-01 -6.15460202e-02 6.87251464e-02 -8.53902400e-01
-1.13064773e-01 2.22222045e-01 3.20023030e-01 -5.32428026e-01
8.46109450e-01 -1.05470881e-01 -3.90761077e-01 3.82735610e-01
1.05887100e-01 -4.26505804e-01 7.30821192e-01 9.49064195e-01
1.33507562e+00 -8.64525288e-02 3.37719679e-01 -4.04860348e-01
7.84661174e-01 6.88692868e-01 9.01407152e-02 6.66752636e-01
2.09249556e-01 2.31225818e-01 1.24350858e+00 -5.33332348e-01
-1.02087915e+00 -8.71414423e-01 -9.64213163e-02 8.64910007e-01
-4.46052521e-01 -8.61801624e-01 -5.67745984e-01 -1.08574641e+00
6.44216716e-01 1.52178919e+00 -1.04934835e+00 1.27766892e-01
-5.25779486e-01 -4.24432755e-01 6.89627111e-01 7.90690720e-01
6.13105237e-01 -7.04860568e-01 -6.70962706e-02 4.14331704e-01
-4.21250135e-01 -1.36813068e+00 4.83081862e-02 1.91509187e-01
-7.50846207e-01 -1.35324562e+00 6.31179959e-02 -4.67166036e-01
6.25422716e-01 -6.67335570e-01 1.71001697e+00 2.93526471e-01
1.87684804e-01 1.89619422e-01 -2.36172691e-01 -4.81123596e-01
-8.43421340e-01 1.79390892e-01 -1.99468896e-01 -5.65128505e-01
5.35458386e-01 -5.93626089e-02 1.84083030e-01 -3.56873035e-01
-7.25042462e-01 1.28781036e-01 2.53009915e-01 8.36674333e-01
1.90891802e-01 -3.32531929e-02 5.30535161e-01 -1.45008326e+00
8.81458700e-01 -3.76540363e-01 -5.21829009e-01 5.21141946e-01
-8.15136373e-01 4.67724562e-01 9.67835128e-01 1.13048710e-01
-9.68043864e-01 -6.98546827e-01 -2.29125530e-01 2.86034882e-01
3.08069795e-01 1.12167466e+00 -7.22226799e-02 2.32908845e-01
1.10573983e+00 5.80124035e-02 -1.65567562e-01 9.22115147e-02
5.74140906e-01 2.39388391e-01 5.96706450e-01 -8.50298226e-01
8.68177056e-01 1.21607356e-01 4.53284383e-02 -1.27761394e-01
-1.16929853e+00 1.57481387e-01 -6.03781819e-01 7.23746000e-03
7.42967486e-01 -7.91142285e-01 -8.18834245e-01 -9.20099691e-02
-9.46092784e-01 -4.62496668e-01 -1.18554890e-01 1.02484889e-01
-3.14280421e-01 3.15471113e-01 -8.82128954e-01 -8.09252322e-01
-1.93332747e-01 -6.31792426e-01 6.06203616e-01 -3.82394224e-01
-5.80182374e-01 -1.32174540e+00 -9.68556851e-02 8.62152100e-01
2.68338233e-01 4.39153016e-01 1.49150848e+00 -1.12793720e+00
-5.43641865e-01 -2.99892843e-01 -5.64864278e-01 1.58764020e-01
2.35367604e-02 -9.92750749e-02 -5.71862876e-01 3.21059287e-01
-3.49585593e-01 -7.32012689e-01 8.97374332e-01 -1.09589696e-01
1.21001983e+00 -9.24692690e-01 1.48585225e-02 -5.81256021e-03
1.54966807e+00 -2.74596006e-01 3.79985034e-01 5.72364330e-01
7.45061696e-01 7.24243462e-01 1.83104053e-01 1.14840977e-01
1.13227165e+00 3.09584215e-02 3.59102994e-01 4.77164119e-01
-4.57276255e-02 -9.10866141e-01 1.40343562e-01 5.29746413e-01
3.46313000e-01 -3.83873701e-01 -1.34529769e+00 6.82786882e-01
-1.27106845e+00 -9.44283247e-01 -5.95239937e-01 1.68637943e+00
1.31401718e+00 7.85362899e-01 -7.55342022e-02 5.80259442e-01
1.33583337e-01 1.77668884e-01 -2.42565393e-01 -7.37456143e-01
-2.64463156e-01 5.79052642e-02 6.03722930e-01 8.29775870e-01
-8.91622245e-01 7.30755806e-01 5.43735123e+00 4.72356170e-01
-5.14724851e-01 -1.64003611e-01 9.42336023e-01 3.03583831e-01
-9.56133902e-01 -2.77338233e-02 -8.13162923e-01 1.93505153e-01
1.19239247e+00 -2.68775463e-01 2.90777951e-01 8.19771171e-01
-4.45402980e-01 -3.73422951e-01 -1.36036777e+00 6.61326826e-01
5.44073045e-01 -2.04676890e+00 4.71103013e-01 -3.10829222e-01
8.10508013e-01 -3.28254372e-01 -1.17579989e-01 8.30054998e-01
7.87465096e-01 -1.24820638e+00 9.75756645e-01 2.88689256e-01
6.77859962e-01 -5.22533774e-01 1.06946528e+00 2.80109406e-01
-8.39875698e-01 -1.67750075e-01 -2.31407836e-01 -4.23687041e-01
-1.20667994e-01 7.98437953e-01 -1.11771667e+00 8.09908807e-01
4.20330316e-01 5.27241051e-01 -1.29849553e+00 1.88677728e-01
-4.24058348e-01 7.05136061e-01 -2.23461211e-01 -6.00940228e-01
7.66419694e-02 8.46924633e-02 5.90870390e-03 1.32713759e+00
-7.25147277e-02 1.33763239e-01 -1.05182968e-01 1.07286298e+00
-7.44313717e-01 8.24025497e-02 -5.71751118e-01 -3.56360108e-01
1.77101105e-01 9.08982635e-01 -6.34933949e-01 -8.27768207e-01
-6.08531475e-01 4.92263913e-01 7.49190569e-01 1.68853611e-01
-7.15472341e-01 -5.15987575e-01 -2.32271403e-02 3.01865906e-01
1.52559787e-01 -2.50193477e-01 -1.19715703e+00 -1.26034498e+00
4.35106277e-01 -1.09029913e+00 6.84156597e-01 -9.91708100e-01
-1.15390193e+00 3.01602691e-01 -9.46334843e-03 -5.82462966e-01
-4.85348165e-01 -1.06821990e+00 -3.28875095e-01 7.95406640e-01
-1.42030442e+00 -1.09663200e+00 -4.36666869e-02 1.89008459e-01
7.34024644e-02 -2.81931292e-02 6.88618243e-01 -6.05692854e-03
-2.95020550e-01 9.33365285e-01 -3.36446911e-01 9.88899708e-01
4.67762589e-01 -1.66162729e+00 7.46557176e-01 9.87272799e-01
4.44695294e-01 8.71719778e-01 7.58605003e-01 -1.05122280e+00
-1.64560544e+00 -9.70058322e-01 1.66997337e+00 -1.15037394e+00
1.32618272e+00 -6.34583294e-01 -1.11480927e+00 1.13502526e+00
3.60733330e-01 -2.35899687e-01 6.69631004e-01 5.58084130e-01
-1.01984072e+00 4.39453113e-04 -1.25175703e+00 6.81069314e-01
9.01480913e-01 -9.17516410e-01 -1.02623296e+00 6.38043940e-01
6.83587313e-01 -7.73714483e-01 -1.19901061e+00 1.38456643e-01
1.39099002e-01 -7.39103675e-01 6.26189351e-01 -1.00338328e+00
1.32344997e+00 -6.25154525e-02 -3.56132179e-01 -1.15537632e+00
-1.97345182e-01 -1.58133626e-01 -4.18707728e-01 1.04657292e+00
1.23733211e+00 -6.69213235e-01 1.01681864e+00 9.83252764e-01
3.40449698e-02 -7.92021453e-01 -8.24948132e-01 -5.29281378e-01
5.60374260e-01 -7.68497109e-01 5.53542793e-01 1.23383510e+00
6.46104872e-01 2.99965262e-01 3.45284492e-01 1.38287663e-01
4.00880486e-01 1.48786813e-01 6.98365450e-01 -1.00260806e+00
-2.26398408e-01 -3.75994980e-01 -1.82583153e-01 -6.19104028e-01
2.85112351e-01 -1.41929841e+00 -3.77235025e-01 -2.32761455e+00
1.71229113e-02 -1.12534583e-01 2.27967463e-02 6.78007007e-01
-3.66471857e-01 2.68462393e-02 1.71957001e-01 -4.91130799e-01
-3.78521323e-01 1.60742566e-01 9.57047224e-01 -5.05930781e-01
3.65206808e-01 -6.35963917e-01 -1.18982446e+00 4.55005109e-01
7.41616845e-01 -4.13064569e-01 -1.94880143e-01 -3.88903022e-01
1.12199211e+00 2.94855446e-01 4.76459652e-01 -9.88429904e-01
3.10152829e-01 9.77735221e-02 7.61074603e-01 -7.70028472e-01
3.91166061e-02 -5.03203988e-01 -4.41426665e-01 5.81180930e-01
-9.26962376e-01 5.64715981e-01 5.87605774e-01 4.19738233e-01
-2.93790817e-01 -1.91051394e-01 4.74670716e-02 -3.01482588e-01
-3.91996622e-01 -3.56349051e-01 -3.16897631e-01 6.62555158e-01
3.87422472e-01 -6.98658451e-02 -8.81224632e-01 -5.06107450e-01
-4.91800636e-01 3.48391265e-01 7.73493648e-02 2.53689378e-01
4.24363792e-01 -1.16982567e+00 -9.13878918e-01 3.80478390e-02
2.82052577e-01 -2.44833961e-01 -9.54674650e-03 6.79187477e-01
-8.40422034e-01 8.12415302e-01 9.15814489e-02 1.87612101e-02
-7.37843096e-01 5.84962606e-01 1.86838835e-01 -8.52985322e-01
-1.67836696e-01 1.07459497e+00 -7.36641765e-01 -9.79968607e-01
-1.12168625e-01 -1.33236980e+00 -2.53462642e-01 1.59668550e-01
2.33530298e-01 3.39770168e-01 5.22977650e-01 -4.78556119e-02
-3.46530437e-01 -3.89981549e-03 -1.36793062e-01 -1.41855434e-01
9.37865615e-01 4.47181642e-01 -2.36657932e-01 5.01747668e-01
1.02446890e+00 8.08812499e-01 -4.11877215e-01 -7.86921531e-02
3.37385207e-01 -2.22833455e-01 -1.64504513e-01 -1.39097929e+00
-6.34770095e-01 6.10210896e-01 -5.04396260e-01 5.17380714e-01
4.36975867e-01 -1.96767509e-01 5.11609018e-01 9.20267522e-01
3.73779356e-01 -9.82834697e-01 -9.07824934e-03 6.68347299e-01
1.17839563e+00 -1.40882134e+00 3.98253858e-01 -5.35601020e-01
-7.45666564e-01 1.33731520e+00 8.03435922e-01 -1.35850042e-01
4.28704292e-01 3.02373976e-01 -5.04525788e-02 -2.92424828e-01
-8.30822825e-01 3.40045363e-01 4.58204478e-01 4.73576128e-01
7.81189680e-01 4.96556871e-02 -1.99320689e-01 8.72420251e-01
-9.01055872e-01 4.32220213e-02 9.83386099e-01 6.04105294e-01
-1.82312250e-01 -8.40099573e-01 -4.19908851e-01 1.30156577e+00
-6.33853257e-01 -6.41544521e-01 -6.99001074e-01 1.18861914e+00
-1.22354247e-01 1.10533988e+00 9.64894444e-02 -4.06189859e-01
2.76046365e-01 4.70638037e-01 4.24684256e-01 -5.72985947e-01
-8.28225613e-01 -9.43055391e-01 1.11117649e+00 -3.87052208e-01
4.54228651e-03 -7.78845191e-01 -1.39335299e+00 -6.74741566e-01
-1.16926972e-02 3.37830007e-01 2.98551798e-01 8.66898656e-01
1.71031401e-01 7.68625438e-01 -2.83300370e-01 2.83790410e-01
-7.77962565e-01 -1.07701600e+00 -2.61353940e-01 3.56081516e-01
1.96047947e-01 -2.84957319e-01 -2.75290459e-01 -5.70542999e-02] | [9.66479206085205, 7.825282096862793] |
95fd253e-6416-42b0-ae9c-b4db75dc2749 | fusion-of-eeg-and-musical-features-in | 1611.10120 | null | http://arxiv.org/abs/1611.10120v1 | http://arxiv.org/pdf/1611.10120v1.pdf | Fusion of EEG and Musical Features in Continuous Music-emotion Recognition | Emotion estimation in music listening is confronting challenges to capture
the emotion variation of listeners. Recent years have witnessed attempts to
exploit multimodality fusing information from musical contents and
physiological signals captured from listeners to improve the performance of
emotion recognition. In this paper, we present a study of fusion of signals of
electroencephalogram (EEG), a tool to capture brainwaves at a high-temporal
resolution, and musical features at decision level in recognizing the
time-varying binary classes of arousal and valence. Our empirical results
showed that the fusion could outperform the performance of emotion recognition
using only EEG modality that was suffered from inter-subject variability, and
this suggested the promise of multimodal fusion in improving the accuracy of
music-emotion recognition. | ['Ken-ichi Fukui', 'Nattapong Thammasan', 'Masayuki Numao'] | 2016-11-30 | null | null | null | null | ['music-emotion-recognition'] | ['music'] | [ 2.34386161e-01 -5.62632799e-01 5.38070023e-01 -3.30755234e-01
-9.07993913e-01 -4.56109941e-01 1.41257241e-01 -8.53024870e-02
-4.06188160e-01 7.72214293e-01 2.99834102e-01 5.48464715e-01
-5.68721473e-01 -3.24264824e-01 -5.53730987e-02 -8.03662717e-01
-4.11444247e-01 -3.67653787e-01 -5.56673050e-01 -1.59082994e-01
3.04023176e-01 4.16353732e-01 -1.88125801e+00 7.09273815e-01
5.70612192e-01 1.74262738e+00 -2.39404246e-01 6.46389484e-01
5.08805178e-02 1.92362949e-01 -1.04854095e+00 -2.51317292e-01
-3.01975384e-02 -6.16457105e-01 -7.78402165e-02 -2.57756501e-01
-9.02172998e-02 5.31671166e-01 6.14291057e-03 1.05108726e+00
1.03222001e+00 2.18473077e-02 5.35456717e-01 -1.37680984e+00
-2.19352152e-02 5.92276096e-01 -4.82143432e-01 4.57087129e-01
7.52277613e-01 -1.90342829e-01 6.32166684e-01 -5.89799106e-01
-3.07228416e-02 5.98260581e-01 6.81943893e-01 3.08206707e-01
-9.90405917e-01 -1.23493683e+00 -3.65143359e-01 5.37635863e-01
-1.58908439e+00 -3.43186200e-01 1.28502202e+00 -4.60617602e-01
9.87191260e-01 7.00926363e-01 1.16644955e+00 1.31691468e+00
5.73939085e-01 2.65619248e-01 1.63483846e+00 -1.31540984e-01
1.77905083e-01 2.00668722e-01 2.32173484e-02 1.80819720e-01
-2.42991433e-01 1.12097695e-01 -1.34658587e+00 -5.44476509e-01
5.20340264e-01 -3.08661282e-01 -4.43557262e-01 5.03510237e-01
-1.27703083e+00 3.30629766e-01 7.36838207e-02 7.02984333e-01
-9.47369993e-01 -5.79954870e-02 5.27275741e-01 4.65254486e-01
2.43898481e-01 7.76670575e-01 -5.16224086e-01 -7.65661061e-01
-1.18032074e+00 -2.54387796e-01 7.33767331e-01 1.13819331e-01
2.65909880e-01 1.87413916e-01 -2.07306936e-01 7.29878604e-01
1.83789805e-02 4.13098723e-01 9.96185184e-01 -6.94114029e-01
-1.34115607e-01 4.34061140e-01 -2.08286807e-01 -1.00553966e+00
-6.57098114e-01 -5.14810145e-01 -9.56304312e-01 1.24177471e-01
-5.09917475e-02 -2.87146300e-01 -2.93602616e-01 1.76502538e+00
-1.81979522e-01 5.62332869e-01 1.40936121e-01 9.85038817e-01
7.47204065e-01 4.03767705e-01 4.98344898e-02 -7.30640829e-01
1.68653560e+00 -4.62073125e-02 -9.67128992e-01 8.51150677e-02
-2.32596561e-01 -7.15188920e-01 5.77764153e-01 1.10015571e+00
-1.15096498e+00 -7.64869452e-01 -1.07986486e+00 7.91040301e-01
-2.93353409e-01 1.06389858e-02 7.61289716e-01 9.43351686e-01
-7.11699367e-01 3.12791497e-01 -3.96170110e-01 1.36820331e-01
2.21791342e-01 6.29361570e-01 -5.97018480e-01 7.88785517e-01
-1.35513759e+00 7.01446712e-01 2.27664560e-01 3.65357876e-01
-4.77796882e-01 -5.37259400e-01 -2.77056634e-01 1.18913531e-01
-2.72857308e-01 -4.90376741e-01 7.44387388e-01 -1.35066569e+00
-1.73480356e+00 5.13567865e-01 -3.95529903e-02 -1.72705770e-01
-1.12190798e-01 2.06081979e-02 -1.00232112e+00 1.96855530e-01
-3.81416231e-01 3.89456391e-01 9.41635251e-01 -6.08913183e-01
-1.82255700e-01 -7.31327415e-01 -6.73540533e-01 2.03790456e-01
-3.84578615e-01 2.37773165e-01 2.38433868e-01 -5.79421163e-01
3.24161500e-01 -4.96592253e-01 1.77385688e-01 -8.59797180e-01
7.90916830e-02 -3.53098959e-02 2.42463127e-01 -4.84546900e-01
1.20000076e+00 -2.65930843e+00 2.45045364e-01 4.54978883e-01
-8.62429142e-02 -1.61042407e-01 -4.66013290e-02 2.81999648e-01
-4.19937670e-01 -1.21945858e-01 -7.89714530e-02 6.36257157e-02
4.07531112e-02 -6.30621687e-02 -4.35591549e-01 3.89155149e-01
7.95749500e-02 8.47728133e-01 -3.35410714e-01 -3.26860577e-01
-4.19923384e-03 7.01641679e-01 -2.26446852e-01 3.81223202e-01
3.15272093e-01 7.25542605e-01 -1.79425463e-01 7.11446941e-01
4.31767970e-01 5.27159929e-01 -1.76114082e-01 -5.93303621e-01
-3.85013595e-02 5.02944253e-02 -1.27627230e+00 1.82964051e+00
-3.05137902e-01 7.39529312e-01 1.48893416e-01 -8.48534763e-01
1.28249002e+00 8.67835939e-01 7.79847145e-01 -9.30243850e-01
5.74056685e-01 5.20455688e-02 2.93149590e-01 -6.93208039e-01
1.63500056e-01 -5.92980266e-01 -2.73150682e-01 3.10831636e-01
2.07033888e-01 -3.72383177e-01 -3.25936526e-01 -5.03918171e-01
9.41720366e-01 -1.38582647e-01 2.73417652e-01 1.82462651e-02
5.21116793e-01 -5.47827125e-01 5.27479410e-01 4.64634866e-01
-3.17892611e-01 1.81559995e-01 2.84125239e-01 -1.24823250e-01
-1.57144383e-01 -1.03039742e+00 -3.33101213e-01 1.06420708e+00
-1.60388112e-01 -4.39063221e-01 -5.17903268e-01 6.35598376e-02
-2.50697851e-01 3.51721644e-01 -5.29632628e-01 -4.55190420e-01
1.43687144e-01 -1.02535629e+00 9.06955600e-01 4.66692269e-01
3.60874116e-01 -1.30134892e+00 -8.69483054e-01 3.15870941e-01
-1.79962039e-01 -8.99944127e-01 1.32951036e-01 7.37026632e-01
-8.04161429e-01 -6.64395452e-01 -3.67549509e-01 -5.59625566e-01
-1.45357549e-01 -5.08629262e-01 8.82656574e-01 -6.08396649e-01
-6.32983387e-01 8.36442590e-01 -2.99008399e-01 -9.01171982e-01
2.59272903e-01 -3.43563199e-01 2.46171564e-01 3.98931384e-01
4.58970785e-01 -1.30394232e+00 -5.59485495e-01 1.79231972e-01
-8.19242179e-01 -3.20874244e-01 5.59868336e-01 3.97793829e-01
4.30712312e-01 3.75185221e-01 1.09550428e+00 4.29086685e-02
1.27693236e+00 -3.03249359e-01 1.07398234e-01 4.94872108e-02
-2.76143670e-01 -3.28302205e-01 4.26603891e-02 -8.35721970e-01
-9.41265881e-01 5.63306129e-03 -4.14498262e-02 -2.98796028e-01
-4.59578425e-01 6.18723929e-01 -1.15876891e-01 -1.68195754e-01
5.91421604e-01 2.88796008e-01 -2.22469807e-01 -1.49239719e-01
-3.60313207e-02 1.13820195e+00 9.03931320e-01 -6.65075660e-01
-2.41066873e-01 1.72649063e-02 -4.62394319e-02 -8.67510557e-01
-3.68194550e-01 -5.15534997e-01 -3.87002349e-01 -6.44503593e-01
9.43699479e-01 -1.05950713e+00 -9.26505566e-01 4.53967929e-01
-9.25223947e-01 2.72818238e-01 -4.47821394e-02 1.10562849e+00
-6.66251659e-01 8.32871795e-02 -5.86865008e-01 -1.27304196e+00
-7.63274491e-01 -7.04904258e-01 8.51884127e-01 1.52131274e-01
-5.66866338e-01 -3.38083476e-01 3.94761741e-01 2.41159424e-01
3.44336033e-01 3.36079150e-01 6.23363972e-01 -6.65074110e-01
2.65409440e-01 -4.33719337e-01 3.19893658e-01 3.05711716e-01
-1.10087711e-02 -2.41938710e-01 -1.54529893e+00 1.49395600e-01
6.59611881e-01 -3.65597665e-01 4.24760729e-01 2.71769285e-01
1.12156451e+00 3.11228901e-01 2.25555867e-01 5.32791257e-01
1.09079754e+00 5.16676486e-01 7.35774696e-01 -1.21190891e-01
1.07792832e-01 3.86614949e-01 1.67075351e-01 6.87751412e-01
-7.16796666e-02 3.85493010e-01 3.10781598e-01 1.77897602e-01
4.07092184e-01 4.25914586e-01 5.66132367e-01 1.13292551e+00
-3.27585727e-01 1.86083987e-01 -5.51913142e-01 1.35648564e-01
-1.60142541e+00 -1.16779828e+00 6.97177649e-02 2.00033927e+00
6.73553526e-01 -1.08993948e-01 2.46990100e-01 7.58675396e-01
4.90245223e-01 -2.79998422e-01 -4.01232600e-01 -7.58409619e-01
-3.85992229e-01 7.27096319e-01 -2.04437822e-01 -1.28382534e-01
-8.21212411e-01 2.17025384e-01 6.90417051e+00 5.39445043e-01
-1.44656408e+00 6.00748770e-02 -2.28423029e-02 -4.31503206e-01
5.64184263e-02 -7.02260196e-01 7.46638700e-02 4.25734967e-01
1.37298906e+00 -9.15699154e-02 8.08593214e-01 1.23174891e-01
6.11215718e-02 -1.47881761e-01 -1.00912666e+00 1.66952538e+00
3.91666919e-01 -4.50437397e-01 -2.43247151e-01 -1.22982800e-01
2.97255188e-01 -1.96038455e-01 2.20699102e-01 3.76912504e-01
-7.39082813e-01 -1.13179338e+00 5.20157516e-01 1.29627812e+00
4.11197990e-01 -9.30381596e-01 8.68339479e-01 4.13994521e-01
-1.09566355e+00 -2.60308743e-01 9.95711610e-02 -5.32986999e-01
-9.30218101e-02 4.60844427e-01 -4.93641496e-01 4.97305185e-01
8.08840752e-01 6.19463921e-01 -4.62926865e-01 1.21053255e+00
2.23421529e-01 5.87435782e-01 -3.35090518e-01 -1.25957042e-01
-2.42213145e-01 -2.20749557e-01 5.47529042e-01 1.32750559e+00
7.85480917e-01 4.12372321e-01 -2.79998779e-01 6.27066672e-01
3.49994421e-01 3.35638016e-01 -4.07526523e-01 -2.97147781e-01
1.44313902e-01 1.41101730e+00 -6.66542530e-01 -1.08901054e-01
-1.41554922e-01 9.11282659e-01 -2.50917733e-01 2.10571021e-01
-8.94620895e-01 -3.99315208e-01 5.87891281e-01 -5.30290782e-01
-3.98674384e-02 -1.40741780e-01 -5.82302630e-01 -1.05557930e+00
-5.70658445e-02 -7.65389800e-01 4.42685366e-01 -1.26620507e+00
-1.33212447e+00 8.90715718e-01 -2.30565876e-01 -1.18405616e+00
-1.35071188e-01 -4.86272514e-01 -7.18580365e-01 1.12012494e+00
-7.20588624e-01 -6.29358530e-01 -2.60491908e-01 1.08030653e+00
7.66928494e-02 -1.08766541e-01 1.25188982e+00 3.42978865e-01
-3.81766647e-01 3.57378095e-01 -3.83161247e-01 -2.01080248e-01
8.63890409e-01 -9.52266276e-01 -9.71937835e-01 1.74312532e-01
4.49575871e-01 1.97164059e-01 6.67506278e-01 -1.16565466e-01
-1.39476502e+00 -2.98105687e-01 5.01622498e-01 -1.47767007e-01
5.20977139e-01 -1.91877201e-01 -5.77214658e-01 -3.67689915e-02
4.87604499e-01 -5.73132932e-01 1.48401105e+00 1.72941521e-01
-1.60065204e-01 -4.51338142e-01 -1.11222970e+00 1.95995048e-01
3.78251880e-01 -1.04727960e+00 -1.09167337e+00 -3.49579215e-01
-2.23505218e-02 4.83283475e-02 -1.34279156e+00 6.25646114e-01
1.03836632e+00 -9.90894318e-01 8.27275693e-01 -3.44917923e-01
-5.05294874e-02 -1.99929476e-01 -5.57817936e-01 -1.35275126e+00
-1.16535187e-01 -4.45661098e-01 1.49447933e-01 1.05848908e+00
2.35203072e-01 -4.35618579e-01 1.87488481e-01 5.93851566e-01
7.05912933e-02 -4.32583630e-01 -1.16089225e+00 -2.42960274e-01
-3.51222664e-01 -9.42435861e-01 6.40958250e-01 8.01447451e-01
8.38937163e-01 4.97261316e-01 -2.89695591e-01 7.23367855e-02
2.19170123e-01 1.07599065e-01 1.42149433e-01 -1.46497786e+00
-2.21917614e-01 -7.18250751e-01 -1.10694861e+00 2.32683927e-01
3.02708417e-01 -9.40823257e-01 -8.58232826e-02 -9.99346316e-01
2.08371177e-01 4.02271181e-01 -1.22792733e+00 3.22050959e-01
1.03969306e-01 7.83735514e-01 4.48738150e-02 -1.61579654e-01
-4.93911803e-01 5.20711005e-01 8.13971400e-01 -7.87209496e-02
-4.83721197e-01 -3.50505374e-02 -7.15379417e-01 7.28919804e-01
8.51017296e-01 -3.31854194e-01 -3.64479721e-01 1.67090580e-01
6.14545286e-01 7.09897518e-01 2.51884520e-01 -1.52174413e+00
3.76152456e-01 2.63247579e-01 8.31120908e-01 -4.90103990e-01
8.63910973e-01 -9.74369526e-01 6.57780826e-01 4.13572639e-02
-2.74672925e-01 8.35818350e-02 4.31042016e-01 4.21936989e-01
-5.42941570e-01 1.59373343e-01 2.84488082e-01 7.92381167e-02
-3.28304768e-01 -1.09037593e-01 -5.72750270e-01 -4.08436567e-01
8.80474448e-01 -4.34074461e-01 1.05342694e-01 -3.44908834e-01
-1.27026153e+00 -4.60747808e-01 -3.55161816e-01 2.38207430e-01
6.24842942e-01 -1.33870184e+00 -5.22344291e-01 6.16582513e-01
1.05007896e-02 -9.75239933e-01 4.41649437e-01 1.30516815e+00
3.56090635e-01 5.72990403e-02 -9.16573346e-01 -5.68673730e-01
-1.34283912e+00 3.85706186e-01 4.92878884e-01 3.40138704e-01
-1.05679311e-01 7.85465956e-01 -1.96553797e-01 2.71987557e-01
4.86585408e-01 -1.99810177e-01 -7.44066775e-01 6.68837011e-01
6.84713364e-01 2.86298305e-01 3.56508702e-01 -5.71033120e-01
-5.02269685e-01 6.39927804e-01 7.12082684e-01 -6.24548256e-01
1.36648631e+00 -2.33635749e-03 -4.88034278e-01 1.22448158e+00
7.91967928e-01 1.29588142e-01 -4.02395904e-01 3.14982772e-01
-7.69874230e-02 -7.54250437e-02 1.87354088e-01 -1.35472393e+00
-8.73019397e-01 9.23645258e-01 1.12717712e+00 4.00628805e-01
1.81699681e+00 -3.32169324e-01 3.39005679e-01 3.31526190e-01
6.09844029e-01 -9.13414061e-01 -4.01695780e-02 8.07337537e-02
9.44883347e-01 -7.25973904e-01 -2.65684456e-01 2.75053799e-01
-7.96294630e-01 1.26162112e+00 2.08291098e-01 -3.56878825e-02
1.09758770e+00 4.62953776e-01 1.35839507e-01 -2.69435316e-01
-8.60439122e-01 -8.60447288e-02 7.14825630e-01 3.53103995e-01
6.62710369e-01 3.92129660e-01 -4.16995704e-01 1.66742527e+00
-5.26517332e-01 1.53475985e-01 -1.04008671e-02 5.10129809e-01
-3.21981281e-01 -7.38919377e-01 -7.39541054e-01 5.08367896e-01
-7.22808182e-01 1.21577235e-03 -6.53096199e-01 2.87442029e-01
4.49186981e-01 1.14171982e+00 3.27065699e-02 -7.89547861e-01
4.12007809e-01 8.01813066e-01 7.28049040e-01 -7.91519806e-02
-1.08516276e+00 5.29413879e-01 -5.73135763e-02 -7.44746566e-01
-8.09320271e-01 -5.07090509e-01 -1.00997341e+00 3.04840654e-01
-7.68745169e-02 5.18645465e-01 9.74444985e-01 7.47943223e-01
4.50125962e-01 8.18773031e-01 5.16909838e-01 -9.61935639e-01
-4.06871401e-02 -1.31451547e+00 -1.22098362e+00 2.43518695e-01
2.95973301e-01 -4.52847034e-01 -5.41480958e-01 -1.41857481e-02] | [13.221643447875977, 3.3581929206848145] |
badbbbf7-dced-42a9-8cd0-8f3dcda80b9f | non-contact-sensing-for-anomaly-detection-in | 2306.10808 | null | https://arxiv.org/abs/2306.10808v1 | https://arxiv.org/pdf/2306.10808v1.pdf | Non-contact Sensing for Anomaly Detection in Wind Turbine Blades: A focus-SVDD with Complex-Valued Auto-Encoder Approach | The occurrence of manufacturing defects in wind turbine blade (WTB) production can result in significant increases in operation and maintenance costs and lead to severe and disastrous consequences. Therefore, inspection during the manufacturing process is crucial to ensure consistent fabrication of composite materials. Non-contact sensing techniques, such as Frequency Modulated Continuous Wave (FMCW) radar, are becoming increasingly popular as they offer a full view of these complex structures during curing. In this paper, we enhance the quality assurance of manufacturing utilizing FMCW radar as a non-destructive sensing modality. Additionally, a novel anomaly detection pipeline is developed that offers the following advantages: (1) We use the analytic representation of the Intermediate Frequency signal of the FMCW radar as a feature to disentangle material-specific and round-trip delay information from the received wave. (2) We propose a novel anomaly detection methodology called focus Support Vector Data Description (focus-SVDD). This methodology involves defining the limit boundaries of the dataset after removing healthy data features, thereby focusing on the attributes of anomalies. (3) The proposed method employs a complex-valued autoencoder to remove healthy features and we introduces a new activation function called Exponential Amplitude Decay (EAD). EAD takes advantage of the Rayleigh distribution, which characterizes an instantaneous amplitude signal. The effectiveness of the proposed method is demonstrated through its application to collected data, where it shows superior performance compared to other state-of-the-art unsupervised anomaly detection methods. This method is expected to make a significant contribution not only to structural health monitoring but also to the field of deep complex-valued data processing and SVDD application. | ['Olga Fink', 'David Flynn', 'Jamie Blanche', 'Daniel Mitchell', 'Gaëtan Frusque'] | 2023-06-19 | null | null | null | null | ['anomaly-detection', 'unsupervised-anomaly-detection'] | ['methodology', 'methodology'] | [ 3.47040653e-01 -3.44852358e-01 8.13128650e-01 -4.54353392e-02
-2.44414315e-01 -2.15449214e-01 4.51300442e-01 3.38582128e-01
-1.16363436e-01 4.54438090e-01 -1.46021470e-01 3.20236348e-02
-7.09476888e-01 -8.73622656e-01 -1.74656898e-01 -1.22685575e+00
-5.54411888e-01 1.98205382e-01 4.35527414e-03 -3.57433975e-01
2.56099820e-01 1.08034170e+00 -1.99610746e+00 1.68020517e-01
8.69274735e-01 1.41473329e+00 1.64618909e-01 3.20851296e-01
2.43380487e-01 3.95959347e-01 -9.39525843e-01 3.04407030e-01
2.09542871e-01 -9.64802504e-02 -3.61824185e-01 2.32429802e-01
-4.60571423e-02 -2.92784035e-01 1.57770544e-01 1.00581932e+00
5.01589060e-01 4.69440043e-01 7.00775445e-01 -1.00100458e+00
-3.38621348e-01 9.42948908e-02 -4.79980737e-01 3.44173074e-01
1.83735102e-01 1.00484774e-01 7.53248692e-01 -1.22893250e+00
8.34010616e-02 6.36546016e-01 7.56983638e-01 9.75269228e-02
-8.91990006e-01 -2.36119792e-01 -2.17512965e-01 4.73395169e-01
-1.10062730e+00 -1.87886298e-01 1.24090040e+00 -6.70326591e-01
8.54692519e-01 3.06608677e-01 6.95430517e-01 8.76288533e-01
6.59035802e-01 1.92818344e-01 1.01698089e+00 -4.84486789e-01
4.39959586e-01 -5.32284260e-01 -5.20587265e-02 5.23628771e-01
4.90417391e-01 5.07754385e-01 -3.01008791e-01 -1.17727190e-01
3.95780951e-01 5.80645390e-02 -3.98839325e-01 1.48110834e-04
-7.75022089e-01 7.61480808e-01 8.01676959e-02 7.06440210e-01
-1.11769485e+00 -2.25438267e-01 3.06856036e-01 4.44874316e-01
5.80458224e-01 7.68844724e-01 -3.11979353e-01 -4.80130203e-02
-7.36648679e-01 1.45079374e-01 5.87761164e-01 5.36872149e-02
2.92931825e-01 9.12911117e-01 1.57123700e-01 7.38316059e-01
3.40287715e-01 7.48423219e-01 4.09731746e-01 -6.65727675e-01
-2.21594125e-01 2.48384491e-01 -1.45454146e-02 -1.28966117e+00
-5.61515391e-01 -7.25376785e-01 -1.01342940e+00 5.58267713e-01
3.31372581e-02 -2.24478289e-01 -7.37220883e-01 1.11745250e+00
3.80147070e-01 2.39194818e-02 8.39759111e-02 9.32962537e-01
5.24618447e-01 5.44448733e-01 -3.35810632e-01 -4.50616896e-01
1.03275681e+00 -1.93120584e-01 -1.12626326e+00 -9.89248306e-02
1.94610208e-01 -8.12841177e-01 6.97882116e-01 9.12991166e-01
-7.41104007e-01 -5.48539937e-01 -1.42680395e+00 7.35133231e-01
-4.89359975e-01 1.23495802e-01 5.62513292e-01 6.17959261e-01
-5.82855701e-01 7.84002781e-01 -1.06182396e+00 3.02495640e-02
1.43736482e-01 1.34293988e-01 -3.40275615e-01 -1.04231246e-01
-1.06883097e+00 9.41372693e-01 1.27308756e-01 6.73115849e-01
-7.22625017e-01 -7.53693104e-01 -9.04953241e-01 1.81073193e-02
2.42251500e-01 -1.71249345e-01 7.06272125e-01 -4.60589290e-01
-1.35405254e+00 1.99485570e-01 3.75542790e-01 -3.70413780e-01
-1.48156464e-01 -5.67441761e-01 -9.74651098e-01 4.80623335e-01
-1.82831943e-01 -4.38492715e-01 1.39166725e+00 -1.23347771e+00
-3.81246060e-01 -4.52825099e-01 -5.25330722e-01 -4.19262081e-01
-4.89276558e-01 -1.61031812e-01 7.63762474e-01 -1.08092451e+00
3.68453473e-01 -4.42320585e-01 -6.49175942e-02 -3.84801835e-01
-2.79023021e-01 -3.79893929e-02 1.14325428e+00 -1.01962733e+00
1.08487046e+00 -2.35126567e+00 -6.27947673e-02 5.53122401e-01
3.79933380e-02 3.19153905e-01 7.13050142e-02 8.90076876e-01
-2.86460370e-01 -5.45228004e-01 -7.38136530e-01 1.75930504e-02
-2.93398350e-01 2.15244606e-01 -2.04763025e-01 7.07070947e-01
6.63069129e-01 1.39830962e-01 -7.38806844e-01 2.62504637e-01
3.99569064e-01 3.89863402e-01 -2.57881969e-01 1.43015310e-01
1.30300999e-01 4.81348425e-01 -2.72127390e-01 8.41319561e-01
7.68427908e-01 4.58752215e-01 -3.24572653e-01 -5.80765426e-01
-4.08007026e-01 -3.94830823e-01 -1.23397446e+00 1.25801754e+00
-3.12887698e-01 5.47750294e-01 4.87239957e-01 -1.46061993e+00
1.39240599e+00 4.91418391e-01 9.27704334e-01 -5.01038551e-01
1.53817892e-01 3.40168446e-01 1.62319243e-01 -8.79896939e-01
2.65987277e-01 -3.36898178e-01 1.68582857e-01 1.62569210e-01
8.30828175e-02 -1.64299712e-01 4.95074317e-02 -3.24355990e-01
1.32091177e+00 -1.70107469e-01 1.19992226e-01 -3.64092201e-01
8.29483151e-01 -1.43917233e-01 6.60569310e-01 2.65447110e-01
8.43513831e-02 3.56865972e-01 3.01164202e-02 -4.71453458e-01
-7.01928616e-01 -1.14316368e+00 -1.80327475e-01 1.76263884e-01
-2.48762041e-01 1.57394066e-01 -3.27009469e-01 -3.49744469e-01
1.68005720e-01 9.62758183e-01 -3.73161584e-01 -6.51095748e-01
-6.50552571e-01 -8.60301197e-01 2.96596646e-01 5.41092873e-01
3.49331737e-01 -1.06189501e+00 -8.66123676e-01 2.95629770e-01
6.24598712e-02 -9.85910356e-01 2.28743851e-01 2.85069644e-01
-9.81534183e-01 -1.17377448e+00 -3.25703502e-01 -4.62972522e-01
5.29197514e-01 1.22761838e-01 8.64099503e-01 1.98188946e-01
-8.51355672e-01 7.07589924e-01 -6.30896211e-01 -7.78573632e-01
-5.24958491e-01 -5.55309594e-01 4.73752379e-01 3.43962103e-01
2.20933855e-01 -7.65952826e-01 -5.27288795e-01 8.01829621e-02
-1.00980425e+00 -7.94346213e-01 7.93456554e-01 1.01593137e+00
4.80249316e-01 9.13696408e-01 1.01980281e+00 -5.79201579e-01
8.53058696e-01 -4.68414962e-01 -5.83403289e-01 -1.94452688e-01
-6.52717054e-01 -2.66556859e-01 9.39176917e-01 -1.24771222e-01
-1.14731812e+00 -3.25490028e-01 -3.51539075e-01 -5.46983421e-01
-4.54491287e-01 1.00410044e+00 -8.29304308e-02 -5.49483970e-02
5.46297312e-01 2.49045417e-02 4.12082940e-01 -7.31992126e-01
-2.51470506e-01 4.80106473e-01 7.50348508e-01 -3.69990289e-01
1.16041696e+00 5.25754511e-01 4.03674424e-01 -1.65194714e+00
-3.68704885e-01 -6.64341331e-01 -3.99746746e-01 -5.11399567e-01
6.16136253e-01 -4.57408786e-01 -5.71281672e-01 8.16281021e-01
-8.82003248e-01 1.48118004e-01 -6.03880465e-01 8.97603214e-01
-1.45308152e-01 6.41246021e-01 -4.48465914e-01 -1.21334445e+00
-5.62253833e-01 -7.20408499e-01 6.67826712e-01 -5.45738898e-02
-8.30928832e-02 -8.67802560e-01 -8.79927427e-02 8.32813531e-02
5.85812032e-01 8.50969374e-01 1.03980124e+00 -8.12405288e-01
-1.15680285e-02 -5.60456455e-01 4.83995795e-01 1.03207052e+00
5.13881862e-01 7.74770379e-02 -9.44377482e-01 -4.67315942e-01
6.65960252e-01 1.83859989e-01 5.49319863e-01 4.74215835e-01
9.79400754e-01 9.48635954e-03 1.32772774e-01 2.25198135e-01
1.50241566e+00 3.94154370e-01 5.49983621e-01 1.27812177e-01
3.99463832e-01 8.29069257e-01 1.11696875e+00 8.33467722e-01
-6.32243395e-01 4.73300934e-01 9.18855309e-01 -1.35389686e-01
1.47416085e-01 5.55217087e-01 4.68511015e-01 1.08127940e+00
-3.97096455e-01 -1.08322361e-02 -6.92941189e-01 6.77618861e-01
-1.24962103e+00 -1.16972029e+00 -4.40113693e-01 2.18940425e+00
2.52008200e-01 1.65646255e-01 -2.51569837e-01 1.02032638e+00
4.62903529e-01 -1.13611203e-02 -3.22478265e-01 -6.06427610e-01
-9.24722478e-02 9.47459280e-01 1.39115885e-01 2.73173779e-01
-1.06771564e+00 8.35915804e-02 4.82819176e+00 5.76394558e-01
-1.07438195e+00 -1.79095045e-01 -2.54504114e-01 2.99145371e-01
-1.13717586e-01 -4.87072974e-01 -3.93158235e-02 3.90964866e-01
7.73066044e-01 2.80027866e-01 2.72435874e-01 5.79694092e-01
3.67402941e-01 -1.25953749e-01 -6.99087501e-01 5.74849904e-01
1.30896479e-01 -8.28322649e-01 -3.20313931e-01 6.89313188e-02
2.43447632e-01 -3.24383527e-01 -9.20900553e-02 -8.36863667e-02
-3.53908509e-01 -7.60338426e-01 4.74162310e-01 8.08272541e-01
4.75579530e-01 -1.01162028e+00 1.22391093e+00 1.47138700e-01
-1.10882545e+00 -5.29663622e-01 -1.40144438e-01 -2.30966121e-01
4.35676306e-01 1.41704249e+00 -7.57606268e-01 1.10005283e+00
8.55176449e-01 5.76603115e-01 -2.97328550e-03 1.03328252e+00
-1.54432163e-01 8.40646565e-01 -2.85477102e-01 3.56972426e-01
-3.68523747e-02 -4.23863083e-01 1.13844991e+00 6.33959293e-01
7.70325601e-01 -3.60311382e-02 2.78320666e-02 6.55160248e-01
5.36373198e-01 -2.64507920e-01 -6.62304699e-01 -3.37131232e-01
5.10215580e-01 1.33249116e+00 -7.01620758e-01 3.29646796e-01
-3.16008896e-01 5.74116826e-01 -6.46072507e-01 2.77473301e-01
-4.92569566e-01 -7.06044316e-01 7.60193408e-01 1.97805390e-01
6.03182316e-01 -3.69294256e-01 -1.30703285e-01 -4.33670461e-01
2.46671632e-01 -8.19469690e-01 4.22677606e-01 -6.20264649e-01
-1.68498254e+00 5.96266329e-01 4.64835688e-02 -1.40666926e+00
-2.77116925e-01 -9.45473552e-01 -8.36268663e-01 7.70590723e-01
-1.31305099e+00 -9.19326901e-01 -5.18642426e-01 4.55505461e-01
4.55305099e-01 -5.23209393e-01 1.20697618e+00 3.75986367e-01
-4.48483765e-01 6.84793591e-02 1.48269206e-01 -7.36886170e-03
3.38869750e-01 -1.20031738e+00 2.27615274e-02 1.17036521e+00
-3.36470082e-02 3.24838698e-01 1.04345238e+00 -7.52903104e-01
-1.35217083e+00 -8.64360213e-01 2.75078714e-01 1.73115045e-01
6.48242772e-01 1.02652334e-01 -9.74350393e-01 4.41746414e-02
9.95037779e-02 1.67417049e-01 8.43004763e-01 -1.50128573e-01
3.04084897e-01 -3.22311640e-01 -1.37850177e+00 2.32568718e-02
4.15636808e-01 -2.56073892e-01 -7.99114406e-01 1.20896466e-01
4.07729775e-01 6.02843519e-03 -1.20001018e+00 8.75960052e-01
3.40376705e-01 -1.00083148e+00 1.18648767e+00 -3.35214943e-01
3.89867753e-01 -4.26364511e-01 -7.46532083e-02 -1.53161931e+00
-4.38803762e-01 -4.32539552e-01 -5.38522899e-01 1.03193879e+00
-5.46434242e-03 -9.92422938e-01 3.58152360e-01 -1.84604704e-01
-8.46191704e-01 -1.09047365e+00 -1.06642509e+00 -9.15133476e-01
-4.38653082e-01 -3.70548785e-01 4.28692758e-01 9.51921344e-01
-3.81446749e-01 -1.83821201e-01 -5.30324951e-02 7.46182621e-01
6.89285517e-01 1.16289772e-01 6.77534565e-02 -1.84645486e+00
-2.72060603e-01 -5.55912359e-03 -5.78148186e-01 -9.12431851e-02
-1.98011637e-01 -3.34632337e-01 2.67546594e-01 -1.35477364e+00
-8.70355666e-01 -2.61521876e-01 -4.90651101e-01 1.06755383e-01
1.54111639e-01 2.59595990e-01 -2.58062899e-01 -2.88694352e-02
4.61499631e-01 8.14789832e-01 9.96936202e-01 -1.05401196e-01
7.71454796e-02 3.12343389e-01 -1.50188208e-01 7.99515843e-01
9.33309138e-01 -3.23266685e-01 -3.58164430e-01 -1.50037840e-01
2.23285943e-01 -1.07347064e-01 5.42367280e-01 -1.45090055e+00
1.56936962e-02 1.11423828e-01 4.32223529e-01 -8.01062465e-01
3.68004829e-01 -1.26089346e+00 -3.67314694e-03 6.61403000e-01
4.52762812e-01 3.24180126e-01 2.58801609e-01 6.97265267e-01
-4.41899359e-01 -4.77714181e-01 6.18634999e-01 9.85114872e-02
-7.33646274e-01 -1.47834927e-01 -8.68509710e-01 -3.40388149e-01
9.90424871e-01 -3.73241425e-01 -1.24020100e-01 -6.98898658e-02
-6.41786218e-01 -1.85456902e-01 1.54558029e-02 1.21944830e-01
1.06682980e+00 -1.14772868e+00 -7.06748605e-01 6.97041810e-01
-1.31274357e-01 -2.02353336e-02 6.60955191e-01 1.14149594e+00
-6.56988740e-01 -7.18039647e-02 -3.77669632e-01 -5.77612340e-01
-1.05207407e+00 3.38407487e-01 9.34162065e-02 -1.02845095e-01
-7.77188778e-01 6.76058054e-01 -4.56698239e-01 4.28366102e-03
-2.02632874e-01 -2.56238252e-01 -4.70935434e-01 2.06178144e-01
3.61726969e-01 8.05948198e-01 8.47763181e-01 -5.91285944e-01
-4.04982120e-01 3.90417963e-01 2.30924442e-01 1.25958562e-01
1.66259933e+00 2.05117449e-01 -3.37547421e-01 5.64785182e-01
6.98214710e-01 3.05503577e-01 -8.12079251e-01 2.70236015e-01
9.88316983e-02 -3.29241395e-01 4.82446879e-01 -8.16575825e-01
-1.19372725e+00 8.12776029e-01 8.55427086e-01 6.89046025e-01
1.61607862e+00 -4.62197065e-01 8.92017484e-01 3.43347400e-01
2.14106500e-01 -1.30244339e+00 2.35655814e-01 3.95170808e-01
1.06008995e+00 -7.73920298e-01 2.93499976e-01 -4.67964977e-01
-2.46937290e-01 1.42823553e+00 2.78966010e-01 -2.25523561e-01
7.84761846e-01 4.30351645e-01 1.83361337e-01 -6.82035089e-01
-1.90414891e-01 -8.54085311e-02 2.99457580e-01 8.58079374e-01
2.20225975e-01 -3.16104479e-02 -3.15401018e-01 7.46557653e-01
-1.91060621e-02 -4.30006027e-01 3.69038999e-01 1.20325744e+00
-4.42482978e-01 -9.56366360e-01 -7.58782923e-01 6.58954680e-01
-6.04317307e-01 1.35468543e-01 4.73634750e-02 6.93601370e-01
4.79079485e-02 1.16013920e+00 4.69145514e-02 -4.33165669e-01
5.67574680e-01 1.50202766e-01 1.95435271e-01 -4.35176075e-01
-4.05805767e-01 -9.04382989e-02 1.17872059e-01 -4.79205936e-01
-5.07189333e-01 -6.90272689e-01 -1.24985182e+00 1.19392335e-01
-5.69813251e-01 2.46260598e-01 8.61127794e-01 9.45889175e-01
2.79647678e-01 1.18914855e+00 9.37857032e-01 -8.74716341e-01
-8.03244531e-01 -1.09844255e+00 -1.00755131e+00 3.52096707e-01
4.07522827e-01 -1.22429979e+00 -7.59621203e-01 -1.66282356e-01] | [6.732947826385498, 2.3330180644989014] |
f8eff211-4945-41e9-bbbc-c2fbee21ae91 | pointwise-mutual-information-based-metric-and | 2305.12191 | null | https://arxiv.org/abs/2305.12191v1 | https://arxiv.org/pdf/2305.12191v1.pdf | Pointwise Mutual Information Based Metric and Decoding Strategy for Faithful Generation in Document Grounded Dialogs | A major concern in using deep learning based generative models for document-grounded dialogs is the potential generation of responses that are not \textit{faithful} to the underlying document. Existing automated metrics used for evaluating the faithfulness of response with respect to the grounding document measure the degree of similarity between the generated response and the document's content. However, these automated metrics are far from being well aligned with human judgments. Therefore, to improve the measurement of faithfulness, we propose a new metric that utilizes (Conditional) Point-wise Mutual Information (PMI) between the generated response and the source document, conditioned on the dialogue. PMI quantifies the extent to which the document influences the generated response -- with a higher PMI indicating a more faithful response. We build upon this idea to create a new decoding technique that incorporates PMI into the response generation process to predict more faithful responses. Our experiments on the BEGIN benchmark demonstrate an improved correlation of our metric with human evaluation. We also show that our decoding technique is effective in generating more faithful responses when compared to standard decoding techniques on a set of publicly available document-grounded dialog datasets. | ['Luis A. Lastras', 'Sachindra Joshi', 'Dinesh Raghu', 'Vineet Kumar', 'Yatin Nandwani'] | 2023-05-20 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [ 2.03818589e-01 5.92719316e-01 2.31515802e-02 -8.67358208e-01
-9.28115189e-01 -5.91779649e-01 1.17090917e+00 1.78677782e-01
-1.44975722e-01 8.00975382e-01 1.16963804e+00 -7.66240135e-02
2.86211222e-01 -9.25670683e-01 -2.20251441e-01 -3.74392301e-01
4.96029705e-01 8.56353939e-01 -3.53397541e-02 -7.02301562e-01
4.28294986e-01 -1.18680388e-01 -8.91857445e-01 7.52798796e-01
6.45940483e-01 6.62474275e-01 -3.71139422e-02 7.90948391e-01
-1.54938966e-01 1.43560970e+00 -9.63820159e-01 -6.90573275e-01
-1.50617734e-01 -1.11156797e+00 -1.26147699e+00 -8.09888318e-02
7.74402618e-02 -3.31888109e-01 -6.61919713e-02 8.02196562e-01
3.36850703e-01 3.26645672e-01 8.00714374e-01 -9.80940878e-01
-7.74537027e-01 1.16763604e+00 1.44451216e-01 7.52956122e-02
8.69217694e-01 -4.98955920e-02 1.25687993e+00 -7.00161517e-01
5.52790761e-01 1.50792444e+00 5.41955292e-01 9.35459971e-01
-1.27317834e+00 -3.71250808e-01 -2.30953544e-01 -1.79282576e-01
-7.61467755e-01 -6.14069104e-01 7.80638218e-01 -4.27663326e-01
1.08218789e+00 2.98380882e-01 2.13669136e-01 1.28101802e+00
2.50655323e-01 5.10639071e-01 1.33863437e+00 -5.60442269e-01
2.56214082e-01 5.50004840e-01 2.70737797e-01 6.59702361e-01
-3.71679217e-01 1.36008382e-01 -8.31527710e-01 -3.07851255e-01
1.89410821e-01 -4.35306489e-01 -1.84900016e-01 2.98411787e-01
-1.00267434e+00 1.45298529e+00 3.10446411e-01 5.24769545e-01
-3.86286139e-01 -5.47184087e-02 9.82016027e-02 2.50456005e-01
6.92831516e-01 5.67061961e-01 3.43928784e-02 -4.86395955e-01
-9.96236145e-01 4.87099081e-01 1.42311811e+00 7.76091635e-01
7.94299066e-01 3.45565099e-03 -5.99610269e-01 8.27083111e-01
5.45627654e-01 2.30906054e-01 7.40925014e-01 -9.54167128e-01
3.58791202e-01 9.58463371e-01 1.31738275e-01 -1.20387220e+00
-1.42708078e-01 -9.44313630e-02 -6.33956611e-01 -1.05807930e-01
3.34743291e-01 -2.61501104e-01 -3.93565387e-01 1.97432792e+00
3.11007947e-02 -6.67505562e-01 3.73983026e-01 9.67513084e-01
1.01804912e+00 9.16262329e-01 -1.20352469e-01 -2.42396116e-01
8.36097002e-01 -7.27322340e-01 -7.28236735e-01 -4.26217288e-01
5.29600620e-01 -8.82404745e-01 1.26848102e+00 -4.33642678e-02
-1.04032063e+00 -4.68527526e-01 -9.53803539e-01 2.01896373e-02
1.31104253e-02 -2.71492273e-01 4.65420038e-01 5.89495361e-01
-1.19323826e+00 4.43336874e-01 -3.69132280e-01 -2.73745298e-01
-2.98237592e-01 8.24513212e-02 -1.94241598e-01 2.99386531e-01
-1.44007099e+00 1.27749634e+00 2.69222200e-01 -1.98255241e-01
-9.06591773e-01 -1.76858276e-01 -7.27088988e-01 -1.37592740e-02
9.04440209e-02 -5.64795434e-01 1.75003242e+00 -7.90565491e-01
-1.75405610e+00 9.32334721e-01 -2.04729334e-01 -5.68590820e-01
4.79815394e-01 4.45967242e-02 -2.27441818e-01 -1.60421893e-01
5.13735451e-02 7.79128015e-01 6.14781797e-01 -1.35167527e+00
-4.77817655e-01 -2.09282503e-01 3.06885302e-01 2.21482709e-01
-2.49238819e-01 -8.12885612e-02 1.11410931e-01 -3.42173755e-01
-1.99875385e-02 -1.12045598e+00 1.08047582e-01 -8.90698016e-01
-7.08879173e-01 -6.04059398e-01 4.26656187e-01 -4.76285249e-01
1.21127498e+00 -1.73545599e+00 1.11009136e-01 2.15440374e-02
3.34785402e-01 6.87869340e-02 9.35286283e-03 1.02637994e+00
3.75206620e-01 8.70913193e-02 -1.52494043e-01 -2.57316560e-01
3.42570484e-01 2.42131531e-01 -7.60276198e-01 4.17455565e-03
7.28929415e-02 8.62735569e-01 -7.97593474e-01 -3.93329591e-01
-1.20047957e-01 3.39530617e-01 -6.33228481e-01 7.04648137e-01
-4.71317232e-01 4.37281132e-01 -3.06950271e-01 3.21488120e-02
6.62769228e-02 -2.29761541e-01 3.13028544e-01 -2.16272827e-02
5.34219742e-02 9.20609415e-01 -5.03320634e-01 1.26665533e+00
-5.21705270e-01 6.76409125e-01 -3.55545372e-01 -3.80053371e-01
1.36538827e+00 4.79962021e-01 -3.56142819e-02 -6.37830079e-01
1.58726573e-01 -2.22324077e-02 2.77663350e-01 -2.39331707e-01
8.60766947e-01 -6.03497505e-01 -4.94333297e-01 1.13631487e+00
1.30120084e-01 -4.86074030e-01 1.86761633e-01 8.16815138e-01
8.73184919e-01 -2.03927934e-01 3.82519245e-01 -2.81375796e-01
3.76461685e-01 2.02235896e-02 1.92760214e-01 8.31668258e-01
6.61120117e-02 3.77994865e-01 7.39151418e-01 -1.77300572e-01
-7.84869313e-01 -7.48903155e-01 3.87763262e-01 1.30015409e+00
-1.17731109e-01 -5.21056831e-01 -1.06414402e+00 -6.14652038e-01
-3.47327471e-01 1.27819824e+00 -6.46444142e-01 -3.58668387e-01
-4.02727067e-01 -5.01598299e-01 7.91738212e-01 5.31640768e-01
4.14584994e-01 -1.16231596e+00 -5.56764781e-01 2.88242728e-01
-8.25376093e-01 -1.02800465e+00 -5.53567529e-01 8.61060023e-02
-6.22055233e-01 -5.72608232e-01 -2.75195092e-01 -5.93776822e-01
3.69486749e-01 -1.37299513e-02 1.35619211e+00 5.70143312e-02
4.93023753e-01 3.66146863e-01 -6.67747200e-01 -1.76777899e-01
-1.49614418e+00 -3.81094515e-02 -1.34827688e-01 -1.96104020e-01
4.19545710e-01 -1.00916989e-01 -2.94054449e-01 2.76951969e-01
-8.97648335e-01 2.22765028e-01 1.97573647e-01 8.51957679e-01
-7.58761074e-03 -2.86867291e-01 6.36080861e-01 -1.02304900e+00
1.58552194e+00 -3.87605458e-01 -3.79966199e-02 1.36825681e-01
-8.90205801e-01 4.02139634e-01 5.03081620e-01 -3.08963567e-01
-1.31020260e+00 -4.05424416e-01 -3.33494663e-01 4.46909964e-01
6.81729568e-03 6.29861236e-01 1.63326710e-01 3.68449152e-01
8.93035054e-01 1.68690503e-01 -7.95616210e-02 -2.21203834e-01
5.08187354e-01 9.46669459e-01 5.29517710e-01 -6.07229173e-01
2.83929944e-01 4.63472381e-02 -3.48470330e-01 -3.88769776e-01
-1.14993370e+00 -1.73240453e-01 -2.49354601e-01 -3.29576939e-01
8.02048206e-01 -4.47663337e-01 -6.86380804e-01 1.95079505e-01
-1.48075557e+00 -3.90402615e-01 2.29839534e-02 3.61330956e-01
-6.47662044e-01 2.61701524e-01 -8.56674433e-01 -9.78593707e-01
-7.22898304e-01 -9.11403060e-01 7.76406169e-01 1.08847342e-01
-1.05527449e+00 -1.26904559e+00 6.31962776e-01 5.75607479e-01
6.27895951e-01 1.58639908e-01 1.23653078e+00 -1.09591067e+00
-2.03836039e-02 -4.18631345e-01 1.08105049e-01 4.94626582e-01
1.71239540e-01 -6.97729141e-02 -1.03649926e+00 -1.89055118e-03
6.23750091e-01 -7.19192266e-01 6.27509058e-01 -1.01199500e-01
1.64676696e-01 -9.86892402e-01 1.86603159e-01 -1.03832372e-01
9.79627192e-01 2.27809459e-01 5.60485959e-01 3.76332328e-02
3.98764193e-01 8.24157417e-01 4.72121358e-01 5.56266010e-01
5.58826685e-01 8.08883786e-01 2.62868792e-01 2.27202535e-01
-8.05661008e-02 -7.41599023e-01 7.07209468e-01 1.01315093e+00
1.20312244e-01 -8.32694113e-01 -7.61094213e-01 2.65380830e-01
-1.67110229e+00 -1.12084365e+00 -3.51871818e-01 2.03869867e+00
1.32010770e+00 2.31694534e-01 1.26609817e-01 -3.51815633e-02
5.39894342e-01 2.18690097e-01 -3.74139771e-02 -8.51969481e-01
2.49272678e-02 1.67659774e-01 -3.05408061e-01 1.13183725e+00
-4.80933905e-01 1.10073853e+00 6.68925667e+00 3.30870748e-01
-9.85755086e-01 2.05883324e-01 5.60811162e-01 7.00288862e-02
-6.56280935e-01 1.73105687e-01 -8.00776780e-01 4.35577720e-01
1.17967653e+00 -4.07913357e-01 2.99129933e-01 7.20667362e-01
1.50230095e-01 5.64189143e-02 -1.49561703e+00 3.41042608e-01
2.66331792e-01 -1.20805645e+00 1.88860551e-01 1.07523156e-02
6.32038772e-01 -3.11219275e-01 -4.47394066e-02 4.22023475e-01
8.02444160e-01 -9.85203087e-01 8.40774655e-01 5.07901013e-01
4.06201065e-01 -5.60887218e-01 7.83064425e-01 6.90545797e-01
-4.32085186e-01 4.02332932e-01 -3.14084202e-01 -1.66183978e-01
1.97952911e-01 4.64053571e-01 -1.74213374e+00 -9.90643539e-03
1.05547078e-01 2.90550827e-03 -3.49520385e-01 -5.02527133e-02
-5.19047201e-01 7.06316411e-01 1.52320862e-01 -4.77031171e-01
4.03527617e-01 1.10010259e-01 5.79006672e-01 1.43918097e+00
1.48842856e-01 1.93747148e-01 5.50185069e-02 1.22125840e+00
-1.03827767e-01 5.61684147e-02 -5.18030405e-01 -5.02200902e-01
5.05637348e-01 1.28145802e+00 -4.13977981e-01 -3.96139830e-01
-3.05793509e-02 9.55628693e-01 3.09019506e-01 -3.73509992e-03
-6.42364800e-01 5.55586852e-02 1.16213098e-01 -1.40660517e-02
-3.16112846e-01 -7.66797960e-02 -2.05918103e-01 -7.93600559e-01
-1.36004418e-01 -1.22795272e+00 2.66816288e-01 -8.40855539e-01
-1.43505514e+00 1.18589461e+00 1.32643953e-02 -8.02937269e-01
-1.23671806e+00 -2.70466030e-01 -6.97561085e-01 1.06240642e+00
-7.88397133e-01 -1.02916241e+00 -3.81809235e-01 5.23920178e-01
6.37231112e-01 -2.67950684e-01 1.21008432e+00 -2.68311769e-01
-5.33477776e-02 4.51419204e-01 -7.08500326e-01 7.48995170e-02
7.24787712e-01 -1.36911678e+00 5.03408551e-01 7.47474611e-01
3.44984740e-01 1.02493000e+00 1.08217025e+00 -6.52914524e-01
-1.02630186e+00 -6.69235051e-01 1.48966372e+00 -7.38163233e-01
5.29631078e-01 -2.10236147e-01 -8.82577598e-01 6.60360694e-01
7.34694362e-01 -1.00632620e+00 9.68559563e-01 1.29138485e-01
-4.08753604e-01 1.91174775e-01 -1.03393245e+00 5.60408890e-01
5.15946269e-01 -7.40922332e-01 -1.02699900e+00 3.64775181e-01
7.45100617e-01 -3.17392707e-01 -6.90662861e-01 1.81217089e-01
3.74904424e-01 -1.32160556e+00 3.93839240e-01 -5.72952151e-01
7.65710890e-01 1.09358788e-01 -5.27302861e-01 -1.56022155e+00
-4.96845543e-01 -7.43241549e-01 -3.08769122e-02 1.35700893e+00
5.14140785e-01 -3.42595577e-01 6.56187952e-01 8.93827677e-01
-2.81643271e-01 -5.41541934e-01 -6.86226308e-01 -3.99884880e-01
1.20960109e-01 -2.59886026e-01 4.56248134e-01 7.84629285e-01
6.00427747e-01 1.01557195e+00 -6.30981863e-01 -3.82130176e-01
2.29272678e-01 2.31932878e-01 8.38859141e-01 -1.00346124e+00
-4.60748345e-01 -3.62549275e-01 -1.06665649e-01 -1.05413425e+00
2.51571864e-01 -1.07481337e+00 5.20735741e-01 -1.65710914e+00
4.43543881e-01 -1.79564338e-02 2.27878734e-01 4.27106261e-01
-1.74926341e-01 9.38458219e-02 2.17589006e-01 3.87736887e-01
-3.53626370e-01 4.80816990e-01 9.94901717e-01 -6.64945468e-02
-1.93801194e-01 8.99522454e-02 -1.05639780e+00 6.87642097e-01
7.46726930e-01 -6.77653670e-01 -5.33870578e-01 -4.30754751e-01
3.55533540e-01 4.18875664e-01 2.71325856e-01 -7.97867060e-01
3.00031185e-01 -1.23946436e-01 -2.21408922e-02 -3.55746716e-01
3.84193420e-01 -2.30787084e-01 5.54614551e-02 4.42969412e-01
-1.23427939e+00 1.60008162e-01 -2.45818347e-01 3.54369789e-01
-3.55113983e-01 -5.43754578e-01 7.03823388e-01 -2.63509154e-01
-4.55733597e-01 -2.75840551e-01 -6.34813726e-01 2.81390786e-01
3.83937240e-01 6.92850426e-02 -4.55011249e-01 -1.09628081e+00
-4.80528027e-01 -3.35616231e-01 3.80784690e-01 5.48654735e-01
7.76657939e-01 -1.25791025e+00 -9.61170375e-01 -1.76357199e-02
1.04076698e-01 -6.76828921e-01 -2.26150170e-01 4.06678170e-01
-2.26011112e-01 7.06835151e-01 -7.66407931e-03 -3.61530513e-01
-1.26631713e+00 8.70331889e-04 3.12207520e-01 -4.46499765e-01
-2.41496265e-01 9.77251887e-01 7.27758929e-02 -3.80361915e-01
1.10872932e-01 -1.28625706e-01 -6.67027310e-02 -4.42675166e-02
4.21265483e-01 2.77127624e-01 2.01603964e-01 -7.79656708e-01
-1.85768053e-01 -2.51292259e-01 -2.31824413e-01 -8.95526826e-01
9.54799354e-01 -2.00195145e-02 -3.00911337e-01 6.28297329e-01
9.53637004e-01 2.56287485e-01 -1.01919758e+00 -3.46358418e-01
7.03118518e-02 -3.34102511e-01 -1.04425065e-01 -1.11782014e+00
-4.48427051e-01 7.85607934e-01 1.04769319e-01 7.20258057e-01
6.63713813e-01 8.41153488e-02 8.36682737e-01 6.22670949e-01
3.28273982e-01 -1.07943451e+00 6.88234091e-01 8.76578331e-01
1.22027147e+00 -1.13434565e+00 -2.94584692e-01 -3.88228111e-02
-1.36798429e+00 1.02242649e+00 6.02488816e-01 1.51089087e-01
2.24618055e-02 3.53357792e-01 4.67766911e-01 -3.16594452e-01
-1.07613027e+00 1.22660920e-01 3.53457808e-01 3.93956304e-01
1.06102967e+00 1.70341924e-01 -3.74793679e-01 6.19908214e-01
-8.72588456e-01 -2.56799638e-01 6.58496559e-01 5.97568214e-01
-7.34955609e-01 -1.12952077e+00 -4.71850894e-02 1.44063056e-01
-3.89604539e-01 -2.34040707e-01 -1.39310789e+00 3.44108313e-01
-3.40958863e-01 1.69495642e+00 -1.82579890e-01 -8.93345952e-01
4.05932404e-02 2.94777244e-01 4.38811570e-01 -9.99496043e-01
-9.02848065e-01 3.87929403e-03 5.08778811e-01 -3.60642582e-01
-4.54865098e-01 -3.12447906e-01 -1.42263603e+00 -5.82078457e-01
-2.93414593e-01 4.61594701e-01 5.72822332e-01 1.22918832e+00
3.71488882e-03 8.16416815e-02 7.53764689e-01 -2.95230269e-01
-9.84510779e-01 -1.40801907e+00 -1.05182081e-01 6.94786489e-01
9.81124956e-03 -1.14727184e-01 -3.21248561e-01 7.00379387e-02] | [12.625585556030273, 8.286717414855957] |
41e2e919-f282-4733-ad14-de8896d3c911 | calibrating-for-class-weights-by-modeling | 2205.04613 | null | https://arxiv.org/abs/2205.04613v2 | https://arxiv.org/pdf/2205.04613v2.pdf | Calibrating for Class Weights by Modeling Machine Learning | A much studied issue is the extent to which the confidence scores provided by machine learning algorithms are calibrated to ground truth probabilities. Our starting point is that calibration is seemingly incompatible with class weighting, a technique often employed when one class is less common (class imbalance) or with the hope of achieving some external objective (cost-sensitive learning). We provide a model-based explanation for this incompatibility and use our anthropomorphic model to generate a simple method of recovering likelihoods from an algorithm that is miscalibrated due to class weighting. We validate this approach in the binary pneumonia detection task of Rajpurkar, Irvin, Zhu, et al. (2017). | ['Philip Marx', 'Daniel Martin', 'Andrew Caplin'] | 2022-05-10 | null | null | null | null | ['pneumonia-detection'] | ['medical'] | [ 4.89878923e-01 4.29646730e-01 -5.51889181e-01 -3.89523566e-01
-9.36611950e-01 -4.92013603e-01 6.84946060e-01 4.69343662e-01
-3.16403925e-01 9.75939035e-01 2.27111891e-01 -5.16476095e-01
-4.55422461e-01 -7.12061405e-01 -6.05259299e-01 -4.92713362e-01
2.45229945e-01 8.91880155e-01 -2.17070081e-03 3.94076079e-01
4.99362946e-01 2.78193831e-01 -1.21477640e+00 1.93359837e-01
9.15735364e-01 2.96201199e-01 -2.67079026e-01 7.19344854e-01
1.98295027e-01 7.48364031e-01 -6.49795055e-01 -7.32323945e-01
1.99763462e-01 -3.53630722e-01 -8.64761591e-01 1.11428387e-01
5.03555775e-01 -1.11343220e-01 5.05882427e-02 9.69967782e-01
2.08263218e-01 -5.00665307e-01 1.34545207e+00 -1.50331211e+00
-1.66598216e-01 7.74306536e-01 -7.80070126e-01 2.26370364e-01
7.91372135e-02 5.64292371e-02 9.96991575e-01 -3.86845499e-01
4.92242575e-01 1.32687545e+00 8.10317874e-01 1.55137420e-01
-1.45767307e+00 -6.67081356e-01 -2.08545431e-01 6.72961175e-02
-1.30968499e+00 -2.61281699e-01 2.83074766e-01 -8.27033997e-01
5.21038234e-01 3.05188298e-01 4.25583512e-01 1.01761353e+00
3.45619768e-01 4.51048583e-01 1.19202054e+00 -5.77183306e-01
3.11293900e-01 2.48280138e-01 1.24489991e-02 6.08724236e-01
8.80030036e-01 3.98733020e-01 -5.27010739e-01 -8.47932160e-01
3.35145950e-01 1.95496883e-02 -2.05914959e-01 -6.01603746e-01
-1.30136847e+00 1.07799065e+00 2.25596502e-01 -9.58243012e-02
-1.53844416e-01 6.42564297e-02 1.28231794e-01 1.94521964e-01
5.57210803e-01 8.99128616e-01 -7.85065174e-01 -1.51475444e-01
-1.09824491e+00 3.04996312e-01 7.39941120e-01 5.19417584e-01
5.48309147e-01 -3.51683587e-01 -1.92952245e-01 5.11838138e-01
4.69155967e-01 3.03733706e-01 6.12579107e-01 -1.09394002e+00
2.22865477e-01 5.40825248e-01 2.71584064e-01 -7.49525607e-01
-2.18185738e-01 -4.25785899e-01 -5.09149909e-01 4.15218115e-01
6.66533947e-01 7.75900483e-02 -8.65807831e-01 1.86496794e+00
1.38114259e-01 2.27405444e-01 1.36992568e-02 4.81362909e-01
6.63707033e-02 2.27258936e-01 3.17896217e-01 -2.40114629e-01
1.13474047e+00 -6.32097483e-01 -3.33856821e-01 -4.12516773e-01
6.63777888e-01 -8.07096422e-01 1.05969143e+00 6.07516110e-01
-7.50289381e-01 -5.27848564e-02 -1.21858585e+00 5.01013696e-01
-7.01866150e-02 -8.15040097e-02 7.73730695e-01 1.09384823e+00
-5.36709249e-01 7.27356493e-01 -8.34619999e-01 -4.76593912e-01
5.54668188e-01 6.37719929e-02 -1.37387559e-01 1.61147475e-01
-8.06010962e-01 1.25063586e+00 3.69990915e-01 -5.76600969e-01
-4.96738195e-01 -8.67772162e-01 -3.01408052e-01 -6.63448349e-02
3.41952324e-01 -5.51505566e-01 1.31221521e+00 -8.83760750e-01
-9.98747349e-01 1.14832866e+00 9.75079089e-02 -5.24566174e-01
9.82629120e-01 -1.85111970e-01 -2.76038736e-01 -1.69742003e-01
4.50265855e-02 2.75320947e-01 8.66302371e-01 -1.15095341e+00
-6.84783220e-01 -6.06236696e-01 -2.89426625e-01 6.05535917e-02
1.27511144e-01 -3.44606186e-03 2.33134076e-01 -8.69614184e-01
2.59770840e-01 -1.07273650e+00 -1.35526657e-01 -1.52381778e-01
-5.01264691e-01 -1.42857693e-02 5.23289859e-01 -5.41253030e-01
1.11448109e+00 -1.67554772e+00 -1.22016065e-01 2.95460701e-01
2.66771257e-01 1.51198998e-01 2.16319218e-01 1.58476025e-01
-2.67939508e-01 4.85774666e-01 -6.80509329e-01 2.14857295e-01
-4.77142483e-02 2.79443301e-02 -3.31164628e-01 6.42535508e-01
5.39142668e-01 5.70030034e-01 -8.05883944e-01 -5.62400043e-01
3.03657115e-01 2.93950886e-01 -4.33240682e-01 1.69550836e-01
-3.85634974e-02 1.43771902e-01 -6.62160292e-02 5.15712976e-01
5.43724179e-01 -3.39671552e-01 5.02980173e-01 -8.73476639e-02
1.37687802e-01 4.66755480e-01 -1.25706387e+00 1.04039621e+00
-1.25463113e-01 5.99621475e-01 -5.38740158e-01 -8.74679506e-01
6.99209750e-01 2.16126204e-01 2.26338059e-01 6.20045997e-02
2.12274894e-01 4.72103432e-02 5.20846769e-02 -1.98226735e-01
1.63444757e-01 -3.59841228e-01 2.48654094e-02 7.18531251e-01
-6.69553131e-02 -5.23773491e-01 -1.21373095e-01 1.77922212e-02
1.09144700e+00 3.90059464e-02 7.92447925e-01 -4.29691553e-01
-1.56701449e-02 1.42192200e-01 5.93527257e-01 9.15796041e-01
-3.00781399e-01 9.28270578e-01 6.33334458e-01 -2.16563761e-01
-1.33146489e+00 -1.28785193e+00 -5.46058416e-01 6.42511964e-01
-3.42567176e-01 -1.96449414e-01 -6.88415825e-01 -7.40402460e-01
1.43075585e-01 7.43818283e-01 -7.86602020e-01 -3.89735073e-01
-1.25638127e-01 -1.22168303e+00 5.22115946e-01 4.28387552e-01
-6.66450262e-02 -6.28268719e-01 -1.01733756e+00 8.06879848e-02
-8.86136219e-02 -6.03078842e-01 -3.31165455e-02 3.24456692e-01
-1.22840703e+00 -1.55474663e+00 -4.34436768e-01 -2.24297225e-01
5.46545267e-01 -2.96407063e-02 1.48022103e+00 1.86758503e-01
-2.99107790e-01 1.06058940e-01 -1.57331750e-01 -7.22521067e-01
-7.10836172e-01 5.76718226e-02 1.85425028e-01 -4.89163220e-01
5.65084457e-01 -1.77758247e-01 -4.36836421e-01 3.86326909e-01
-8.02776098e-01 -8.68836492e-02 6.22786641e-01 8.02872479e-01
3.95858228e-01 1.61768943e-02 5.41284919e-01 -1.31558335e+00
4.44426268e-01 -5.34155786e-01 -6.28074169e-01 5.62810421e-01
-1.20910203e+00 2.68877238e-01 -3.36787552e-02 -5.74500799e-01
-8.54716003e-01 7.70584941e-02 2.25830004e-01 -1.30292669e-01
-2.03077439e-02 2.77476609e-01 3.70321907e-02 2.29436159e-03
1.00243950e+00 -4.34895039e-01 -7.99332559e-02 -4.32301104e-01
1.92609191e-01 6.75324500e-01 5.38560212e-01 -6.52284145e-01
8.96671474e-01 2.95262843e-01 -9.17109847e-02 -2.45378107e-01
-1.13667500e+00 -1.68666914e-01 -5.94345331e-01 7.22388998e-02
5.10943592e-01 -9.29702938e-01 -3.75333369e-01 2.31693566e-01
-8.06971610e-01 -2.67013431e-01 -3.88582408e-01 7.89822638e-01
-7.02884734e-01 2.08477363e-01 -1.40184239e-01 -8.41121674e-01
-3.99347059e-02 -8.83745313e-01 7.65321493e-01 2.27774099e-01
-8.56004298e-01 -7.98871756e-01 2.34541535e-01 5.16717136e-01
1.55916601e-01 3.75041962e-01 1.18341291e+00 -8.32289100e-01
-7.19360635e-02 -4.58805650e-01 -2.26627707e-01 1.23295702e-01
1.37517110e-01 4.97732431e-01 -1.12723732e+00 -2.67227262e-01
1.14085339e-01 -3.35390449e-01 5.98935187e-01 4.41887617e-01
1.25165904e+00 -3.39949101e-01 -3.11332524e-01 1.91252708e-01
1.32173765e+00 -3.69898267e-02 4.74166632e-01 4.95125145e-01
3.60074282e-01 6.95307136e-01 7.29162395e-01 4.07362342e-01
1.93862095e-01 6.18317902e-01 2.52963185e-01 -5.89463152e-02
7.35555291e-02 -4.33568239e-01 -1.17821343e-01 2.69327670e-01
2.09044561e-01 -1.30602106e-01 -1.16837084e+00 5.43203413e-01
-1.80204630e+00 -8.87098789e-01 -2.53345639e-01 2.68661499e+00
1.01449692e+00 4.80533719e-01 2.40008771e-01 3.50474834e-01
8.86341631e-01 -2.06752256e-01 -3.90692234e-01 -2.74985760e-01
9.19427909e-03 4.98575307e-02 6.41307414e-01 5.93477130e-01
-9.60020840e-01 3.62225860e-01 7.98867798e+00 6.70002043e-01
-6.81978583e-01 1.00640327e-01 1.14464104e+00 1.00571916e-01
-1.22370720e-01 1.54211238e-01 -2.54172891e-01 5.61909556e-01
1.00489628e+00 -4.10100967e-01 1.24666803e-01 7.14344800e-01
-4.61758114e-02 -3.25013995e-01 -1.36613846e+00 6.57335818e-01
-2.33183056e-02 -1.06602454e+00 -2.30587691e-01 -4.89212610e-02
7.15907574e-01 -2.87520979e-02 1.85591467e-02 2.99191475e-03
8.88926923e-01 -1.28768909e+00 7.52862573e-01 2.08789617e-01
6.97383523e-01 -7.36801624e-01 8.42389524e-01 2.28110999e-01
-4.11672831e-01 1.12995192e-01 -2.36926273e-01 -1.66482702e-01
-2.19307482e-01 7.70063877e-01 -1.26042271e+00 5.94489723e-02
6.54301703e-01 7.74130225e-02 -7.26718783e-01 1.16949439e+00
-3.62992555e-01 1.01352108e+00 -3.13930243e-01 4.00703430e-01
-2.52440870e-01 4.76107001e-02 4.34631795e-01 9.84234512e-01
9.59966481e-02 -1.91702202e-01 -2.85786808e-01 9.20294940e-01
1.36862591e-01 -2.41833180e-01 -5.31817913e-01 -3.56632136e-02
5.62726617e-01 8.95644605e-01 -8.30568850e-01 -3.01928341e-01
-2.14635462e-01 4.58734840e-01 1.50946483e-01 1.70595005e-01
-7.02337623e-01 4.52101864e-02 4.05373752e-01 2.58169807e-02
-1.62408665e-01 1.13230817e-01 -7.38621473e-01 -1.07721341e+00
-2.84823596e-01 -9.46588993e-01 6.01561248e-01 -7.62988567e-01
-1.44852424e+00 8.50037038e-02 1.62867710e-01 -1.20726442e+00
-2.90097862e-01 -4.85183477e-01 -4.51011181e-01 8.06574047e-01
-1.09010768e+00 -6.55877411e-01 -1.45072743e-01 2.95530353e-02
2.08810866e-01 -1.47145361e-01 7.58475721e-01 -1.67990606e-02
-3.55340987e-01 5.96067905e-01 1.03136845e-01 -3.17531794e-01
8.86097193e-01 -1.50166702e+00 3.51833791e-01 6.82991803e-01
1.56124473e-01 4.23424482e-01 1.17938769e+00 -7.96461999e-01
-7.57180274e-01 -8.90567243e-01 1.00404632e+00 -1.09909236e+00
5.17698050e-01 6.36544451e-02 -9.56615210e-01 5.89946210e-01
-3.04080725e-01 -4.50313509e-01 1.00205386e+00 2.17499197e-01
-6.15811348e-01 2.03401446e-01 -1.59025371e+00 3.77765179e-01
6.27205312e-01 -4.47603941e-01 -1.00847924e+00 4.56015408e-01
4.61902529e-01 -1.06571496e-01 -6.97288096e-01 5.94521999e-01
6.72009349e-01 -9.10459161e-01 9.54137921e-01 -8.74505639e-01
4.33240891e-01 -2.74465114e-01 -2.90703148e-01 -1.22729969e+00
-1.67545214e-01 -1.62328914e-01 1.10429890e-01 1.14082241e+00
4.97379631e-01 -4.88275200e-01 9.47718680e-01 7.32962012e-01
4.96613443e-01 -5.03436744e-01 -7.87150204e-01 -8.25365305e-01
3.02442998e-01 -4.10194725e-01 6.59772336e-01 1.35741174e+00
-2.30533198e-01 1.99085042e-01 -2.94975847e-01 1.23845771e-01
8.09509337e-01 -2.20942989e-01 6.18055224e-01 -1.76488650e+00
-3.66374969e-01 -4.19749618e-01 -4.47499990e-01 3.07324845e-02
-1.28250629e-01 -6.25386536e-01 7.88362324e-02 -1.18862712e+00
7.80955732e-01 -5.35783291e-01 -4.25397992e-01 4.08231199e-01
-3.59318972e-01 2.73085773e-01 1.19114935e-01 4.95072871e-01
-2.76073784e-01 1.84761733e-03 6.15940869e-01 -2.45563500e-02
1.86694227e-02 2.14171529e-01 -1.04031086e+00 1.00116420e+00
8.78721416e-01 -1.07700658e+00 -2.87671804e-01 -1.24042690e-01
4.02250677e-01 1.17200185e-02 4.10692155e-01 -7.61811554e-01
-1.84427336e-01 -4.48522955e-01 5.02425730e-01 -2.47607842e-01
-2.52318531e-01 -8.14863265e-01 4.11769301e-01 9.90014553e-01
-4.61258769e-01 1.28552347e-01 1.50211556e-02 5.13391256e-01
1.97106943e-01 -5.77920139e-01 9.49560940e-01 2.18731537e-02
-1.01135433e-01 -4.06201333e-02 -3.54073644e-01 2.30737194e-01
7.93519080e-01 -1.91992968e-01 -5.64228058e-01 -2.29162052e-01
-3.19514900e-01 2.34923940e-02 8.37659836e-01 3.29943240e-01
1.99545041e-01 -1.03131330e+00 -9.24871802e-01 -3.88284437e-02
3.37861359e-01 -2.54155993e-01 -4.23502237e-01 5.75525880e-01
-5.32002151e-01 2.57017791e-01 -1.77072689e-01 -6.41795814e-01
-1.13968229e+00 3.93075317e-01 3.72131765e-01 -2.44908467e-01
-2.27322832e-01 7.55207300e-01 7.58693367e-02 -5.70497155e-01
6.88110813e-02 6.83986172e-02 1.89245418e-01 -3.97990905e-02
3.28376144e-01 5.64930618e-01 1.60991192e-01 -2.92430997e-01
-4.93200302e-01 3.73923063e-01 -4.52974439e-02 -2.60519028e-01
1.29132617e+00 1.60391420e-01 1.04619920e-01 5.55866420e-01
8.00030231e-01 -2.34445557e-02 -1.21852362e+00 -1.50099143e-01
3.59341055e-01 -8.11601162e-01 2.77270768e-02 -1.14439547e+00
-5.84367990e-01 6.82569146e-01 1.06279945e+00 1.55991614e-01
8.40072989e-01 -1.03453752e-02 -1.25581935e-01 3.58206838e-01
1.98480949e-01 -9.67027783e-01 -6.88263997e-02 9.49631352e-03
6.73771143e-01 -1.53003979e+00 5.46312809e-01 -2.54805118e-01
-5.45163572e-01 8.21814239e-01 4.97010231e-01 4.81580012e-03
4.80410069e-01 2.88645774e-01 -4.55957800e-02 -1.26632214e-01
-6.67294085e-01 3.30027431e-01 2.37769246e-01 6.60676241e-01
5.05758703e-01 3.85386676e-01 -6.21642530e-01 4.29798782e-01
-3.67982268e-01 4.96824794e-02 6.60867572e-01 9.54793572e-01
-3.68299127e-01 -1.02664983e+00 -5.88788867e-01 9.98917162e-01
-6.47244394e-01 -3.83850895e-02 -6.02046669e-01 6.93530262e-01
-4.75840177e-03 9.28616107e-01 6.29951358e-02 -2.82301575e-01
7.95703083e-02 1.27513975e-01 4.62245882e-01 -5.95269382e-01
-3.54791015e-01 -1.75929993e-01 8.64065732e-05 -1.54100284e-01
-4.62022871e-01 -7.56023169e-01 -9.05573606e-01 -3.31404686e-01
-6.00266397e-01 5.19941673e-02 6.82901144e-01 9.25856590e-01
-1.10702375e-02 2.84013569e-01 5.48572898e-01 -4.02132630e-01
-7.24601686e-01 -7.74760664e-01 -3.49954665e-01 5.25416434e-01
1.75599769e-01 -8.25154603e-01 -7.60871470e-01 -9.25924350e-03] | [8.594366073608398, 4.663268566131592] |
6709621b-6775-49d1-888e-2dfec2f02e3d | explainability-of-predictive-process | 2202.08041 | null | https://arxiv.org/abs/2202.08041v1 | https://arxiv.org/pdf/2202.08041v1.pdf | Explainability of Predictive Process Monitoring Results: Can You See My Data Issues? | Predictive business process monitoring (PPM) has been around for several years as a use case of process mining. PPM enables foreseeing the future of a business process through predicting relevant information about how a running process instance might end, related performance indicators, and other predictable aspects. A big share of PPM approaches adopts a Machine Learning (ML) technique to address a prediction task, especially non-process-aware PPM approaches. Consequently, PPM inherits the challenges faced by ML approaches. One of these challenges concerns the need to gain user trust in the predictions generated. The field of explainable artificial intelligence (XAI) addresses this issue. However, the choices made, and the techniques employed in a PPM task, in addition to ML model characteristics, influence resulting explanations. A comparison of the influence of different settings on the generated explanations is missing. To address this gap, we investigate the effect of different PPM settings on resulting data fed into an ML model and consequently to a XAI method. We study how differences in resulting explanations may indicate several issues in underlying data. We construct a framework for our experiments including different settings at each stage of PPM with XAI integrated as a fundamental part. Our experiments reveal several inconsistencies, as well as agreements, between data characteristics (and hence expectations about these data), important data used by the ML model as a result of querying it, and explanations of predictions of the investigated ML model. | ['Manfred Reichert', 'Mervat Abuelkheir', 'Ghada ElKhawaga'] | 2022-02-16 | null | null | null | null | ['predictive-process-monitoring'] | ['time-series'] | [ 2.24854633e-01 5.51811457e-01 -1.91040143e-01 -5.16886711e-01
-1.55464828e-01 -3.53525519e-01 1.00846708e+00 5.98063588e-01
3.51031050e-02 5.80759525e-01 2.72130817e-01 -5.48323274e-01
-5.69775760e-01 -8.60358894e-01 -5.53753614e-01 -3.56014818e-01
-4.66536991e-02 7.74155438e-01 1.51782930e-01 2.94306427e-01
6.02852523e-01 5.24203181e-01 -1.54408610e+00 5.79009533e-01
5.28013110e-01 9.12598848e-01 -1.98265687e-02 6.12554729e-01
-5.31524003e-01 1.16984069e+00 -6.94809616e-01 -3.68812621e-01
2.01236382e-01 -1.58134863e-01 -6.95916831e-01 1.02032296e-01
-3.18968028e-01 1.78436145e-01 4.57415044e-01 5.55300295e-01
-2.96986908e-01 -3.63749206e-01 7.19352901e-01 -1.80296564e+00
-1.52952388e-01 1.13630605e+00 -2.87013203e-01 9.80118215e-02
4.83197272e-01 6.13417387e-01 1.02682030e+00 -5.15865624e-01
7.81624675e-01 1.18327379e+00 4.50547874e-01 2.66151041e-01
-1.22070193e+00 -3.87753993e-01 3.65412593e-01 1.54913753e-01
-8.14356208e-01 -2.55102247e-01 5.86758137e-01 -5.02585948e-01
7.02160418e-01 4.87988114e-01 5.05811512e-01 1.00577533e+00
5.30809164e-01 5.47197580e-01 1.29887843e+00 -4.58998233e-01
5.96427917e-01 6.39507592e-01 3.49046409e-01 9.84830335e-02
5.72858393e-01 -4.14604545e-02 -7.23187864e-01 -5.80979943e-01
4.19739991e-01 9.50635448e-02 -1.59560651e-01 -2.84086615e-01
-1.07535887e+00 5.92608571e-01 -1.99233115e-01 5.08130610e-01
-6.77721560e-01 3.50538152e-03 1.26660466e-01 4.85121071e-01
-3.51094306e-02 8.08535516e-01 -9.01927888e-01 -3.93052131e-01
-6.56757712e-01 3.79926980e-01 1.45863295e+00 8.94518673e-01
7.17887998e-01 -2.95879215e-01 -4.38222349e-01 4.85076420e-02
6.59487844e-01 -2.14561090e-01 4.75142032e-01 -8.92989337e-01
3.16438228e-01 1.02323079e+00 3.00102979e-01 -7.01757073e-01
-2.44795516e-01 -3.59478176e-01 -3.33142310e-01 4.38030094e-01
6.77852392e-01 5.46216108e-02 -4.72799599e-01 1.45056534e+00
4.70103696e-02 -1.34735599e-01 -3.76575929e-03 5.27487516e-01
1.62795708e-01 4.53602105e-01 4.81023431e-01 -4.76942629e-01
1.23075080e+00 -7.08864629e-01 -6.52571559e-01 -2.16522023e-01
4.05418813e-01 -5.95513344e-01 9.67266023e-01 7.20804870e-01
-1.04859364e+00 -5.02624273e-01 -8.18201542e-01 6.81965709e-01
-1.69540077e-01 -3.04606408e-01 7.07199872e-01 4.52784151e-01
-7.02867746e-01 1.09932160e+00 -8.26177120e-01 -5.14670253e-01
4.88603003e-02 3.71823937e-01 -5.15369251e-02 2.88515329e-01
-8.48849893e-01 9.89343882e-01 1.96458489e-01 -1.87417880e-01
-5.26680052e-01 -6.85765564e-01 -2.53016770e-01 5.66524923e-01
6.23807907e-01 -7.07123220e-01 1.38996565e+00 -1.01737785e+00
-1.15251803e+00 2.22802058e-01 -1.44849315e-01 -7.96549737e-01
1.08780396e+00 -1.84880987e-01 -5.78197658e-01 -4.31511283e-01
-7.83701986e-02 1.11609362e-01 6.66591048e-01 -1.43326569e+00
-8.75189900e-01 -4.91591871e-01 -1.96675375e-01 -2.61960685e-01
3.08514506e-01 2.02359539e-03 -2.18119234e-01 4.80695777e-02
1.20536022e-01 -7.35759020e-01 -5.29642045e-01 -3.85024965e-01
-4.51146513e-01 -1.63592339e-01 4.88873124e-01 -2.19867080e-01
1.29355907e+00 -1.75894964e+00 -3.05287391e-01 6.76266849e-01
2.27994978e-01 -2.41891220e-01 2.71406680e-01 8.18394959e-01
-2.86040176e-02 6.21239960e-01 1.41190112e-01 -3.99555206e-01
2.59058982e-01 3.14810753e-01 -3.62903863e-01 9.50153451e-03
3.89073104e-01 4.55623507e-01 -4.72004682e-01 -2.77104408e-01
2.90985316e-01 2.04474866e-01 -9.93844271e-02 6.40311658e-01
-6.09049439e-01 7.66822338e-01 -4.74809349e-01 6.67532563e-01
2.95134902e-01 -3.20337534e-01 3.48129004e-01 1.16345756e-01
-4.25067872e-01 2.88674951e-01 -1.29848981e+00 7.85291016e-01
-3.84542108e-01 1.23711467e-01 -2.32321948e-01 -4.93671805e-01
1.29090703e+00 4.78822738e-01 5.72034359e-01 -2.96360970e-01
-1.71283111e-01 1.28646314e-01 2.26923019e-01 -3.63795280e-01
2.82810837e-01 3.18942964e-02 3.04318011e-01 7.24900246e-01
-4.50573921e-01 3.24352056e-01 8.80526975e-02 -1.69825107e-01
1.54453170e+00 1.77824795e-01 7.49469936e-01 -1.63731754e-01
6.10358000e-01 2.54397750e-01 9.35250521e-01 1.01901078e+00
-1.63492128e-01 4.62462574e-01 1.13516176e+00 -7.86307633e-01
-8.67908835e-01 -6.63275599e-01 2.73999758e-02 5.09765387e-01
-1.06603391e-02 -6.27721846e-01 -3.84765476e-01 -7.00225353e-01
8.32832381e-02 1.35390985e+00 -6.59450412e-01 7.15369880e-02
-1.78969011e-01 -4.96375173e-01 -8.17621127e-02 3.25642437e-01
8.03538784e-02 -1.32613170e+00 -1.02309895e+00 5.48622251e-01
2.40821779e-01 -9.97434378e-01 3.30573708e-01 3.37263405e-01
-1.08810043e+00 -1.29119897e+00 4.41980839e-01 3.25662524e-01
2.67260283e-01 -2.62023449e-01 1.53504884e+00 9.72721949e-02
1.38370350e-01 2.76930302e-01 -2.94306755e-01 -8.98868978e-01
-1.09674776e+00 9.48390812e-02 -2.40347847e-01 1.22196764e-01
6.37751281e-01 -7.28279054e-01 -3.99589151e-01 4.43667948e-01
-7.96032131e-01 1.04291454e-01 7.81694531e-01 1.15941249e-01
4.78476942e-01 2.15773925e-01 3.25844407e-01 -1.41698813e+00
9.46831703e-01 -6.32648230e-01 -5.35857916e-01 5.44569552e-01
-1.41233408e+00 4.86994892e-01 4.45342094e-01 -2.92913824e-01
-1.10780251e+00 -1.12930819e-01 2.78149873e-01 -2.28273794e-01
-6.22930646e-01 7.18583226e-01 -4.13819551e-01 6.85398042e-01
4.67192799e-01 -1.73798129e-01 7.48380721e-02 -4.31763917e-01
-9.43174362e-02 4.02913481e-01 8.19825828e-02 -5.26309907e-01
6.63464069e-01 3.07955682e-01 7.01824799e-02 -1.17956467e-01
-5.41962504e-01 -2.70811498e-01 -4.54456925e-01 -2.72935838e-01
4.39860642e-01 -3.49268675e-01 -9.18881297e-01 -2.16493323e-01
-1.05228782e+00 -1.96446016e-01 -4.37025309e-01 2.38104343e-01
-5.92931807e-01 -1.43774837e-01 -3.43634367e-01 -1.27774131e+00
-4.06953812e-01 -1.24570024e+00 5.57450712e-01 1.81793541e-01
-1.10105658e+00 -8.00370336e-01 4.12034383e-03 5.12812376e-01
6.32652104e-01 3.03570747e-01 1.21648633e+00 -1.31990731e+00
-8.86755049e-01 -2.74844408e-01 5.88835403e-02 -1.06820941e-01
1.36207476e-01 5.58265686e-01 -9.52085257e-01 2.56517380e-01
2.94201761e-01 4.96641934e-01 1.84915110e-01 1.47990406e-01
1.09637260e+00 -5.62701285e-01 -1.94391191e-01 -5.63535048e-03
1.47350585e+00 2.54403383e-01 5.80656946e-01 8.66499782e-01
3.21345687e-01 9.02773798e-01 9.17597830e-01 7.94801414e-01
2.07805201e-01 5.55362761e-01 5.92878282e-01 3.98621440e-01
4.89671677e-01 -1.60057679e-01 3.51805151e-01 9.33024064e-02
-4.49938506e-01 4.41057570e-02 -1.23580492e+00 1.07170343e-01
-2.02302670e+00 -1.11362839e+00 -5.63965142e-01 2.41114426e+00
4.75406647e-01 5.83942235e-01 4.86104675e-02 4.16309416e-01
4.30769324e-01 -1.36987865e-01 -3.87974113e-01 -8.29072893e-01
3.48937958e-01 -1.74277097e-01 4.25222427e-01 4.02654558e-01
-6.68141901e-01 2.85082459e-01 5.38674927e+00 -2.81033348e-02
-9.58162665e-01 -1.59662589e-01 7.01520443e-01 2.61372983e-01
-6.63470089e-01 4.05649751e-01 -8.59891951e-01 4.79243845e-01
1.36369085e+00 -4.30706263e-01 2.77189553e-01 1.28993499e+00
7.20398426e-01 -1.21267155e-01 -1.76447284e+00 4.60001886e-01
-4.78172421e-01 -1.42454100e+00 8.29878375e-02 4.94850010e-01
3.15335333e-01 -2.21814170e-01 -3.37545633e-01 1.87132865e-01
2.87574172e-01 -1.17213655e+00 8.94698679e-01 1.00889373e+00
-2.41274178e-01 -6.65985227e-01 1.02068698e+00 4.51432347e-01
-7.95806646e-01 -4.70455050e-01 -4.31996658e-02 -4.27469224e-01
-8.63828287e-02 9.06138003e-01 -1.44502473e+00 4.83588099e-01
6.10074341e-01 2.13081658e-01 -5.80637097e-01 8.45403671e-01
-3.57857227e-01 9.87842917e-01 -2.85045803e-02 -1.81629658e-02
-2.70687472e-02 -3.32750142e-01 5.11880100e-01 9.63193893e-01
3.41941267e-01 -4.24816728e-01 -1.88657232e-02 1.35062933e+00
3.69306743e-01 1.19426593e-01 -5.31453252e-01 4.14265394e-02
5.86089373e-01 1.11639011e+00 -5.40502191e-01 -2.68899173e-01
-5.04314899e-01 2.51140863e-01 4.66478523e-03 1.88791215e-01
-5.49381852e-01 6.64463401e-01 7.42066264e-01 7.86263347e-01
-3.50269079e-02 9.14655477e-02 -9.20834184e-01 -7.44016230e-01
1.25354022e-01 -1.15516782e+00 4.03827667e-01 -7.56923199e-01
-1.21021116e+00 6.37634277e-01 -4.60259058e-02 -1.12714100e+00
-5.23288786e-01 -2.37935215e-01 -9.98301566e-01 9.13625717e-01
-1.35886502e+00 -8.31546903e-01 -4.76721495e-01 1.25119671e-01
3.92397195e-01 3.06397732e-02 7.04252183e-01 -4.86986727e-01
-4.58602220e-01 -2.67472595e-01 -2.60912746e-01 -4.16401774e-01
5.97715974e-01 -1.40800309e+00 4.29628372e-01 6.77941322e-01
2.29074135e-01 6.79293156e-01 1.11087680e+00 -7.77775228e-01
-1.36393404e+00 -1.03911638e+00 1.24174273e+00 -8.14020097e-01
7.35271811e-01 1.80562496e-01 -1.23906875e+00 8.50834012e-01
8.55929181e-02 -4.41044927e-01 7.43977249e-01 4.95605588e-01
4.14908715e-02 -2.41223291e-01 -1.21461833e+00 5.24073660e-01
6.06653035e-01 2.84812283e-02 -5.70817590e-01 -1.29323721e-01
3.53346437e-01 8.98275077e-02 -1.24623799e+00 3.92923534e-01
3.30820978e-01 -1.46370649e+00 4.42929953e-01 -5.93684793e-01
4.83890474e-01 -2.15205505e-01 4.98822331e-02 -1.04880822e+00
-2.67003834e-01 -5.73203087e-01 -3.19318622e-01 1.47994769e+00
7.76866376e-01 -6.56212151e-01 8.47453356e-01 1.70350158e+00
3.96692127e-01 -8.34016085e-01 -4.75448757e-01 -3.09468001e-01
-4.36912030e-01 -7.83876121e-01 1.16442072e+00 6.56740427e-01
1.10463379e-02 1.24107547e-01 -5.91230653e-02 4.45439517e-01
5.87583423e-01 4.51730549e-01 1.16332841e+00 -1.70572519e+00
-5.56887150e-01 -3.90367359e-01 -3.45427275e-01 -1.11491032e-01
-3.61827165e-01 -2.19890386e-01 -3.47819358e-01 -1.42189825e+00
9.77661759e-02 -6.07567370e-01 -3.16156685e-01 3.40990156e-01
-2.42998853e-01 -8.35676014e-01 5.48822939e-01 6.79979682e-01
-2.36776829e-01 2.15722043e-02 8.05752695e-01 1.71473652e-01
-5.20866871e-01 5.64216316e-01 -8.15039814e-01 9.29465950e-01
1.03400457e+00 -5.89596510e-01 -3.84668708e-01 3.40795666e-02
3.21759403e-01 3.45953286e-01 2.15982229e-01 -1.18833625e+00
3.54502946e-01 -5.48492730e-01 3.20517480e-01 -3.42491418e-01
-6.58115074e-02 -1.16922033e+00 1.01451623e+00 5.26316464e-01
-5.99831045e-01 2.35641852e-01 -2.93062657e-01 5.89869082e-01
-2.98683763e-01 -4.55633014e-01 2.69226491e-01 -3.68666977e-01
-5.21486640e-01 1.32170439e-01 -4.47386056e-01 -4.76755708e-01
1.13815999e+00 -5.34903884e-01 7.43991870e-04 -4.14587706e-01
-9.05950010e-01 1.00807138e-01 5.92384398e-01 3.35022628e-01
6.13622665e-02 -4.99421239e-01 -6.41411841e-01 1.04186662e-01
2.34377518e-01 6.00777939e-02 -4.24730659e-01 8.95641208e-01
-1.08947240e-01 4.02630150e-01 -1.43551350e-01 -4.70523238e-01
-1.22912073e+00 5.06301999e-01 1.28783852e-01 -7.71687031e-01
-5.09534299e-01 -6.70904219e-02 -2.62535751e-01 -2.29605630e-01
2.41065010e-01 -5.77468991e-01 -3.01571578e-01 -1.47236973e-01
4.04611051e-01 2.91570246e-01 1.45947531e-01 1.63173035e-01
-1.60963684e-01 -1.82169795e-01 -1.10053159e-02 -8.19416195e-02
1.55823338e+00 -9.40001458e-02 -1.57066271e-01 9.48115289e-01
1.69691592e-01 1.69934928e-01 -1.36255836e+00 -9.55591574e-02
8.26204300e-01 -4.58342969e-01 -4.20729339e-01 -1.00058043e+00
-7.79724836e-01 5.98508120e-01 8.58818367e-02 6.29124224e-01
7.39914656e-01 1.46729186e-01 4.45635095e-02 1.63418651e-01
5.56559801e-01 -8.41642976e-01 -4.53070045e-01 -9.16194841e-02
9.68350708e-01 -1.16335785e+00 -7.04675494e-03 -6.22056305e-01
-9.93364990e-01 1.34737730e+00 6.39292002e-01 4.05912846e-01
5.08791447e-01 2.74493694e-01 2.94455737e-01 -3.25367749e-01
-1.38782263e+00 3.24612767e-01 -2.09557772e-01 4.20236945e-01
4.05514777e-01 3.41074556e-01 -2.90552378e-01 9.49750662e-01
-1.72546446e-01 3.15114498e-01 7.42970765e-01 9.53080833e-01
-3.19614142e-01 -1.35058951e+00 -6.19062960e-01 7.00125575e-01
-4.27955031e-01 3.14353287e-01 -6.14750385e-01 9.95810091e-01
9.64785069e-02 1.11818147e+00 -4.10087481e-02 -2.31974348e-01
4.88460541e-01 3.68831366e-01 -1.72882959e-01 -7.05045164e-01
-9.66159880e-01 -3.45301628e-01 4.29962993e-01 -7.07169116e-01
-2.00182900e-01 -1.11917889e+00 -1.21527553e+00 -3.43395829e-01
-1.78680733e-01 4.00382489e-01 8.44385505e-01 1.17731762e+00
4.06561941e-01 5.17245233e-01 4.65328962e-01 -2.34341338e-01
-9.05520678e-01 -9.86437917e-01 -4.03657317e-01 8.10668051e-01
-4.16019440e-01 -3.54846627e-01 -5.09714663e-01 1.75661184e-02] | [8.594805717468262, 6.012921333312988] |
653dfed9-376d-43b3-a41c-34781f1ac6c5 | rlprompt-optimizing-discrete-text-prompts | 2205.12548 | null | https://arxiv.org/abs/2205.12548v3 | https://arxiv.org/pdf/2205.12548v3.pdf | RLPrompt: Optimizing Discrete Text Prompts with Reinforcement Learning | Prompting has shown impressive success in enabling large pretrained language models (LMs) to perform diverse NLP tasks, especially when only few downstream data are available. Automatically finding the optimal prompt for each task, however, is challenging. Most existing work resorts to tuning soft prompt (e.g., embeddings) which falls short of interpretability, reusability across LMs, and applicability when gradients are not accessible. Discrete prompt, on the other hand, is difficult to optimize, and is often created by "enumeration (e.g., paraphrasing)-then-selection" heuristics that do not explore the prompt space systematically. This paper proposes RLPrompt, an efficient discrete prompt optimization approach with reinforcement learning (RL). RLPrompt formulates a parameter-efficient policy network that generates the desired discrete prompt after training with reward. To overcome the complexity and stochasticity of reward signals by the large LM environment, we incorporate effective reward stabilization that substantially enhances the training efficiency. RLPrompt is flexibly applicable to different types of LMs, such as masked (e.g., BERT) and left-to-right models (e.g., GPTs), for both classification and generation tasks. Experiments on few-shot classification and unsupervised text style transfer show superior performance over a wide range of existing finetuning or prompting methods. Interestingly, the resulting optimized prompts are often ungrammatical gibberish text; and surprisingly, those gibberish prompts are transferrable between different LMs to retain significant performance, indicating LM prompting may not follow human language patterns. | ['Zhiting Hu', 'Eric P. Xing', 'Meng Song', 'Tianmin Shu', 'Han Guo', 'Yihan Wang', 'Cheng-Ping Hsieh', 'Jianyu Wang', 'Mingkai Deng'] | 2022-05-25 | null | null | null | null | ['text-style-transfoer'] | ['natural-language-processing'] | [ 3.32593888e-01 1.18059367e-01 -3.71929467e-01 -3.30011368e-01
-1.01977837e+00 -9.57559705e-01 5.97155273e-01 -1.45788109e-02
-5.90799749e-01 8.41139197e-01 2.40126863e-01 -6.83327615e-01
3.44443917e-02 -5.21914363e-01 -5.42200565e-01 -5.42927682e-01
3.64589274e-01 5.31027615e-01 -1.49872243e-01 -4.81985480e-01
2.07493961e-01 3.54794085e-01 -1.03067398e+00 2.51304626e-01
1.26676857e+00 6.54549778e-01 8.21075916e-01 6.74526751e-01
-4.44538683e-01 5.51308870e-01 -6.94733143e-01 -2.23794445e-01
2.53483504e-01 -5.21000803e-01 -8.38367999e-01 -3.98299903e-01
2.46866807e-01 -4.31044966e-01 -8.58814567e-02 8.41240823e-01
5.48526525e-01 4.39714551e-01 7.81041086e-01 -1.03850400e+00
-1.14696145e+00 9.55356836e-01 -1.88993394e-01 3.60544235e-01
3.33440185e-01 4.06296760e-01 1.43871474e+00 -9.64439631e-01
4.31261092e-01 1.28418171e+00 3.15174013e-01 9.60678339e-01
-1.37607360e+00 -5.68399072e-01 3.87021184e-01 -1.23701617e-01
-1.10872245e+00 -3.50587994e-01 5.58357179e-01 -2.94789225e-01
1.08328974e+00 3.22794884e-01 2.82880902e-01 1.37107241e+00
1.33754462e-02 1.05966628e+00 1.10546982e+00 -6.47013664e-01
2.03459144e-01 1.49688751e-01 -8.07724968e-02 5.52712083e-01
-6.79934323e-02 3.25904965e-01 -7.39335358e-01 -1.15506262e-01
6.80670261e-01 -2.34773189e-01 -3.03312629e-01 1.04436822e-01
-1.33657849e+00 9.13363695e-01 2.44930536e-01 2.55766124e-01
-1.38049498e-01 -4.75281803e-03 2.26706609e-01 6.16719306e-01
1.44838721e-01 1.17395306e+00 -9.05231297e-01 -3.84703606e-01
-8.92930806e-01 2.06236601e-01 7.72429049e-01 1.13771725e+00
6.66356444e-01 2.82125533e-01 -6.06578171e-01 1.08860445e+00
8.94234776e-02 5.21179974e-01 9.32450593e-01 -6.33249462e-01
8.83874893e-01 4.06776130e-01 1.57900676e-01 -4.92249668e-01
-3.52114916e-01 -2.11653769e-01 -7.98219860e-01 -1.43133745e-01
5.38052380e-01 -5.48459649e-01 -8.68411779e-01 2.12458754e+00
5.23373075e-02 -5.99068701e-02 2.61978358e-01 7.24946976e-01
8.38505030e-01 1.02982056e+00 3.07171315e-01 -2.06771985e-01
1.11253750e+00 -1.07220876e+00 -4.45525616e-01 -6.81644738e-01
8.12878251e-01 -6.16066217e-01 2.02217102e+00 1.93490729e-01
-9.48288083e-01 -5.12821019e-01 -6.65839493e-01 -1.75051406e-01
-3.13358516e-01 3.50095242e-01 5.44684231e-01 3.30863297e-01
-1.01205957e+00 7.04016209e-01 -5.20165026e-01 -3.33387673e-01
1.33441657e-01 4.53400880e-01 -7.08201453e-02 5.65784164e-02
-1.43176651e+00 8.78079653e-01 4.57828045e-01 -1.34682745e-01
-7.03141391e-01 -7.72269189e-01 -7.44483650e-01 1.97509885e-01
4.02394742e-01 -5.90375721e-01 1.70593905e+00 -1.14083982e+00
-2.09436941e+00 3.84031892e-01 -1.59398004e-01 -2.81068593e-01
4.07228351e-01 -3.50157052e-01 -2.61745393e-01 -7.57239014e-02
5.22155166e-02 1.01583362e+00 1.01527047e+00 -7.82522917e-01
-6.34220362e-01 1.88156843e-01 5.03614321e-02 5.89738369e-01
-8.17229688e-01 2.31107652e-01 -4.63407822e-02 -8.07698309e-01
-3.39835107e-01 -8.76101077e-01 -3.78827661e-01 -3.42240036e-01
-3.81852210e-01 -5.41629612e-01 4.59820747e-01 -5.39298475e-01
1.48016620e+00 -2.04927611e+00 7.94107169e-02 -8.32903292e-03
-2.47449204e-01 4.23700809e-01 -6.31679595e-01 5.24611533e-01
1.50210291e-01 4.18655634e-01 -7.93985724e-02 -4.75606062e-02
1.13499358e-01 3.55928898e-01 -5.17202079e-01 -7.55789205e-02
3.81076157e-01 1.21940506e+00 -1.14141202e+00 -3.54327172e-01
-1.70322567e-01 -1.56536520e-01 -6.40675724e-01 6.65064812e-01
-6.14698410e-01 5.29434979e-01 -4.82690722e-01 4.96016949e-01
-7.08995247e-03 -4.64731902e-01 7.53538832e-02 2.11725518e-01
-9.42054987e-02 5.44754326e-01 -8.84835005e-01 1.58999527e+00
-7.87118912e-01 5.16709208e-01 -1.18456364e-01 -7.35651135e-01
9.48547184e-01 3.61035168e-01 -6.48077279e-02 -7.07675576e-01
1.68992337e-02 4.14133966e-01 1.87052026e-01 -5.00591457e-01
8.55621815e-01 -1.26063824e-01 -2.19200373e-01 5.37410259e-01
-1.73555054e-02 -3.36333990e-01 2.93785542e-01 1.47560045e-01
1.04038143e+00 2.24430770e-01 5.67858696e-01 -1.39192015e-01
2.88878113e-01 -5.46839200e-02 5.76288462e-01 9.82166648e-01
-8.13866034e-02 4.04542059e-01 3.40364933e-01 4.15826887e-02
-9.03478444e-01 -9.91044343e-01 1.35282621e-01 1.72683406e+00
-8.39067996e-02 -3.17134917e-01 -5.73397100e-01 -8.11561763e-01
-4.06001918e-02 1.00049722e+00 -2.34721541e-01 -3.30519795e-01
-7.58931220e-01 -4.97918189e-01 5.97763181e-01 7.06370771e-01
7.96236694e-02 -1.49594164e+00 -4.70814675e-01 4.79527742e-01
-2.01838940e-01 -9.74426389e-01 -9.13513362e-01 5.02916992e-01
-8.81333172e-01 -4.42898333e-01 -7.48019874e-01 -8.98445368e-01
8.72316897e-01 8.21560174e-02 1.12717557e+00 -2.71081608e-02
1.53427541e-01 1.43156216e-01 -4.30290997e-01 -2.48641446e-01
-5.63298821e-01 4.58110809e-01 1.99392840e-01 -2.68606186e-01
2.04472184e-01 -2.90067673e-01 -3.95051092e-01 2.69807070e-01
-8.45738590e-01 9.24125388e-02 8.95320594e-01 1.43130028e+00
4.76148784e-01 -3.37457895e-01 9.84396577e-01 -9.78876889e-01
1.28133619e+00 -4.44113314e-01 -5.26442349e-01 4.72666293e-01
-7.28815615e-01 3.39517266e-01 1.35330451e+00 -1.00472522e+00
-1.27593684e+00 -1.07629605e-01 -9.34456438e-02 -1.70678660e-01
-2.03591853e-01 5.28027892e-01 -1.34836197e-01 1.92998290e-01
9.66133952e-01 2.48140022e-01 -2.89111882e-01 -4.60710257e-01
7.89675653e-01 8.57511997e-01 3.69531989e-01 -9.56513643e-01
7.92394638e-01 -2.91078717e-01 -5.28284609e-01 -7.76767910e-01
-8.78162384e-01 -3.51410061e-01 -2.81808764e-01 2.29764879e-01
4.57542062e-01 -6.59529448e-01 -3.70039731e-01 9.45281163e-02
-1.23456359e+00 -9.78029788e-01 -4.63463783e-01 4.11359668e-01
-5.11124432e-01 8.20021406e-02 -8.08876216e-01 -5.79967976e-01
-5.95599174e-01 -1.09496546e+00 9.27532852e-01 3.99901032e-01
-8.21745217e-01 -9.58716691e-01 -6.81155920e-02 5.58223315e-02
4.78405654e-01 -5.01749992e-01 1.29842985e+00 -1.00712943e+00
-4.75492746e-01 1.11143023e-01 -9.91543606e-02 4.01200980e-01
2.28907660e-01 -1.18571095e-01 -7.56916881e-01 -2.72658527e-01
-4.55727786e-01 -7.04003274e-01 5.00103235e-01 1.24931194e-01
1.10153663e+00 -7.98604012e-01 -1.60598353e-01 5.98959208e-01
9.46297109e-01 2.26930022e-01 1.79051682e-01 2.42494360e-01
5.57826698e-01 4.16684628e-01 8.17009091e-01 3.18449140e-01
1.70399338e-01 5.42779267e-01 -8.81875753e-02 -2.66875252e-02
-2.56689750e-02 -6.96291804e-01 7.20406353e-01 9.60677385e-01
4.43816960e-01 -4.37785536e-01 -7.76429713e-01 4.06506538e-01
-1.63863468e+00 -6.36647284e-01 3.73765856e-01 1.89368129e+00
1.48971641e+00 9.00195837e-02 -1.64095134e-01 -3.15616548e-01
4.15138364e-01 1.32987767e-01 -6.87258184e-01 -7.38892734e-01
3.38332588e-03 3.06238085e-01 2.83056080e-01 4.88916069e-01
-6.53197706e-01 1.55105186e+00 5.61396503e+00 1.07475543e+00
-1.21767426e+00 -1.99704133e-02 5.10483980e-01 -1.07396930e-01
-6.47533119e-01 8.78969654e-02 -1.14643991e+00 4.05914873e-01
9.80216146e-01 -3.16086620e-01 7.68614292e-01 8.95930350e-01
4.10356343e-01 1.63380191e-01 -1.34644413e+00 8.57582808e-01
-2.58980632e-01 -1.15672266e+00 3.44215259e-02 -4.28730935e-01
8.85775626e-01 7.55066378e-03 1.56520754e-01 7.52948344e-01
6.65210545e-01 -1.15903747e+00 8.08807075e-01 5.07406928e-02
1.10743785e+00 -5.06650269e-01 2.32528761e-01 8.07700217e-01
-8.18700254e-01 -3.78739685e-01 -4.42176849e-01 -9.11041070e-03
1.88243628e-01 2.97459632e-01 -1.19689834e+00 1.39392123e-01
3.06597173e-01 3.43150169e-01 -4.64327872e-01 6.71667516e-01
-7.82947540e-01 7.80735910e-01 -2.92021751e-01 -7.07529664e-01
3.73570800e-01 -1.00404792e-01 4.53451902e-01 1.28362501e+00
5.27469754e-01 9.55718085e-02 3.23589385e-01 8.73283625e-01
-3.26730639e-01 5.53755701e-01 -6.05731964e-01 -4.80960786e-01
7.98853278e-01 1.26848018e+00 -4.77510244e-01 -2.37890810e-01
-3.45265031e-01 9.36333179e-01 7.41583288e-01 5.66619456e-01
-5.01912773e-01 -3.65968764e-01 3.98418903e-01 -1.57600492e-01
1.21990964e-01 -3.03863049e-01 -3.49008262e-01 -1.13493466e+00
-1.63785994e-01 -1.17485952e+00 2.98330843e-01 -6.69117928e-01
-1.38069952e+00 7.33481050e-01 6.19854545e-04 -1.02472246e+00
-5.73093891e-01 -6.51930213e-01 -7.15861261e-01 1.03817403e+00
-1.49429095e+00 -8.02715778e-01 1.74411178e-01 4.35953736e-01
8.97298694e-01 -3.15838873e-01 8.34746122e-01 -2.77429074e-02
-7.06756592e-01 9.46186543e-01 2.11074263e-01 -1.15253270e-01
9.67946470e-01 -1.40087950e+00 4.83754456e-01 8.44735384e-01
3.22480351e-01 8.51847410e-01 4.92775977e-01 -6.06435716e-01
-1.28313124e+00 -1.12139189e+00 1.05787361e+00 -2.90073752e-01
7.59216070e-01 -4.17697370e-01 -8.33632827e-01 5.76333106e-01
2.02591076e-01 -2.17542112e-01 5.79409063e-01 2.19732061e-01
-1.40338883e-01 3.72104347e-03 -7.46011615e-01 1.25553668e+00
9.81650651e-01 -4.59846109e-01 -7.35404968e-01 3.97524506e-01
9.60299373e-01 -5.29851913e-01 -4.58592057e-01 1.25933930e-01
1.96977302e-01 -2.87958145e-01 7.08673477e-01 -9.81918752e-01
4.32610571e-01 8.51002559e-02 6.36803359e-02 -1.86112142e+00
-2.89618850e-01 -1.15214527e+00 -1.88633814e-01 1.50256681e+00
9.17835057e-01 -5.64745605e-01 5.86802304e-01 7.42547154e-01
-1.66420311e-01 -1.03595388e+00 -6.27305567e-01 -1.01900196e+00
2.74616241e-01 -2.87637442e-01 8.02307308e-01 8.10081601e-01
1.73561588e-01 8.22563171e-01 -3.20268631e-01 -1.71235591e-01
-1.59472898e-02 2.34689325e-01 6.65395200e-01 -9.47884440e-01
-6.19740903e-01 -5.53204358e-01 5.62176228e-01 -1.65300405e+00
3.87545556e-01 -1.29268229e+00 4.50388849e-01 -1.31587303e+00
-1.90391198e-01 -8.91304255e-01 -2.39361838e-01 7.98287094e-01
-6.07731819e-01 -4.87193793e-01 2.41317421e-01 1.64581820e-01
-3.18178684e-01 6.39386952e-01 1.45632899e+00 2.63562445e-02
-6.60564005e-01 7.81375393e-02 -1.00595725e+00 5.78400433e-01
1.08327270e+00 -4.59802628e-01 -7.06562698e-01 -5.74472487e-01
2.98413545e-01 1.57179356e-01 1.17806932e-02 -4.05212671e-01
2.59355783e-01 -6.53183043e-01 1.88288227e-01 -1.74726725e-01
7.39837661e-02 -3.82764518e-01 -4.82760936e-01 1.07952811e-01
-9.69793916e-01 2.43724972e-01 5.71453385e-02 4.64909166e-01
4.57093753e-02 -6.09493554e-01 6.87365353e-01 -2.50097126e-01
-8.00138950e-01 5.06568015e-01 -3.49306494e-01 6.47965252e-01
6.05035663e-01 2.00571269e-02 -2.92596757e-01 -2.38620698e-01
-2.46841565e-01 3.47141862e-01 1.54546261e-01 5.63457847e-01
5.65459609e-01 -1.24049616e+00 -6.39289021e-01 1.42803490e-01
9.87486467e-02 1.16693586e-01 -2.14411527e-01 6.08510554e-01
-3.62279662e-03 4.88138527e-01 1.59062743e-01 -2.61461914e-01
-9.94986117e-01 3.82898241e-01 5.11673801e-02 -6.21310174e-01
-6.20338857e-01 1.18368053e+00 3.39992225e-01 -6.77932203e-01
2.14909717e-01 -5.34547091e-01 -9.58843678e-02 3.29346433e-02
4.53046829e-01 7.51915127e-02 -7.04174191e-02 8.66230763e-03
3.62194739e-02 1.59421876e-01 -3.76910180e-01 -2.51457989e-01
1.01789951e+00 -2.72981841e-02 2.09568858e-01 3.57881784e-01
1.03186178e+00 7.62700289e-02 -1.40140378e+00 -3.88759673e-01
2.47787341e-01 -2.96479493e-01 -2.56001472e-01 -9.67456162e-01
-5.55747211e-01 9.90218997e-01 7.28572393e-03 3.78195606e-02
8.82600367e-01 -1.30728453e-01 9.79467571e-01 7.69775867e-01
2.49786571e-01 -1.38451159e+00 3.88135254e-01 8.13996136e-01
9.37848568e-01 -1.15134084e+00 -4.87700999e-01 -4.32445155e-03
-9.40614939e-01 1.13540721e+00 9.81689274e-01 3.39376032e-01
1.67728774e-02 2.35260427e-01 1.37838081e-01 3.54943067e-01
-1.00760806e+00 -4.43249680e-02 1.66269124e-01 5.27283430e-01
5.73540986e-01 2.04291150e-01 -3.16746771e-01 8.24871719e-01
-5.21613300e-01 -1.11926787e-01 3.96481186e-01 7.48727083e-01
-5.26286066e-01 -1.41473365e+00 -2.61307925e-01 7.38096356e-01
-2.98614860e-01 -6.05313361e-01 -2.52389789e-01 5.58565319e-01
-2.02480584e-01 1.02762532e+00 -3.24687511e-01 -1.25865132e-01
2.20359683e-01 1.87790662e-01 2.09630683e-01 -1.03118551e+00
-7.52111733e-01 -8.28368664e-02 1.69005126e-01 -2.94504285e-01
3.16419572e-01 -5.91257215e-01 -1.48695588e+00 2.61839591e-02
-4.13260818e-01 1.59932926e-01 2.14873075e-01 1.05324948e+00
2.91503251e-01 4.57849503e-01 6.64414048e-01 -5.45886636e-01
-1.33238745e+00 -9.64644670e-01 -4.06008214e-01 3.22983116e-01
2.94100255e-01 -3.84945333e-01 -2.41875663e-01 -1.04401149e-01] | [11.355358123779297, 8.64217758178711] |
2b2e049d-0c19-4c61-b518-01bb9b59bcbc | discriminative-diffusion-models-as-few-shot | 2305.10722 | null | https://arxiv.org/abs/2305.10722v1 | https://arxiv.org/pdf/2305.10722v1.pdf | Discriminative Diffusion Models as Few-shot Vision and Language Learners | Diffusion models, such as Stable Diffusion, have shown incredible performance on text-to-image generation. Since text-to-image generation often requires models to generate visual concepts with fine-grained details and attributes specified in text prompts, can we leverage the powerful representations learned by pre-trained diffusion models for discriminative tasks such as image-text matching? To answer this question, we propose a novel approach, Discriminative Stable Diffusion (DSD), which turns pre-trained text-to-image diffusion models into few-shot discriminative learners. Our approach uses the cross-attention score of a Stable Diffusion model to capture the mutual influence between visual and textual information and fine-tune the model via attention-based prompt learning to perform image-text matching. By comparing DSD with state-of-the-art methods on several benchmark datasets, we demonstrate the potential of using pre-trained diffusion models for discriminative tasks with superior results on few-shot image-text matching. | ['Xin Eric Wang', 'William Yang Wang', 'Sugato Basu', 'Pradyumna Narayana', 'Arjun Akula', 'Varun Jampani', 'Tsu-Jui Fu', 'Weixi Feng', 'Xuehai He'] | 2023-05-18 | null | null | null | null | ['text-matching'] | ['natural-language-processing'] | [ 4.77586150e-01 -8.10557045e-03 -2.53624141e-01 -3.50579888e-01
-1.00833166e+00 -3.55797470e-01 1.28353691e+00 -2.83552818e-02
-2.33240828e-01 2.48259112e-01 6.95512831e-01 -1.44445136e-01
1.06612220e-01 -7.74904191e-01 -6.84759319e-01 -4.08638388e-01
3.87492031e-01 7.92353034e-01 3.58664125e-01 -2.71820247e-01
2.63107628e-01 2.31230393e-01 -1.48259544e+00 6.53625786e-01
1.01284337e+00 9.58426952e-01 5.02598226e-01 9.20469403e-01
-4.80672956e-01 1.23746121e+00 -5.33312559e-01 -3.79160732e-01
2.38518685e-01 -9.31017935e-01 -4.37097520e-01 1.49070024e-01
7.08523154e-01 -7.27683842e-01 -6.78583324e-01 7.01796174e-01
7.94908524e-01 2.88110942e-01 1.27148175e+00 -1.20287359e+00
-1.45025730e+00 4.49528605e-01 -7.85718381e-01 3.73480171e-01
4.08882022e-01 6.65531158e-01 1.01514363e+00 -1.14845252e+00
1.25015724e+00 1.59855998e+00 3.32429886e-01 9.44650233e-01
-1.50729263e+00 -5.49485803e-01 5.49129508e-02 1.65156856e-01
-9.33764458e-01 -6.68492973e-01 8.64255488e-01 -8.16190541e-01
1.07931232e+00 -1.11732349e-01 4.74464387e-01 1.69194281e+00
2.73829550e-01 1.29424286e+00 9.31279540e-01 -4.07842278e-01
-6.84590358e-03 -3.86247374e-02 -3.75175506e-01 9.14452553e-01
-1.48189589e-01 4.27524805e-01 -7.31512547e-01 -2.98219472e-02
8.13169479e-01 1.55974388e-01 2.71047559e-02 -3.93198609e-01
-1.29004431e+00 1.18934298e+00 4.93353754e-01 2.65971869e-01
-3.76843512e-01 3.39657336e-01 4.09185797e-01 4.09960121e-01
7.73531199e-01 5.36819994e-01 6.48416206e-02 -6.92107305e-02
-1.19647038e+00 2.68203616e-01 5.67622185e-01 1.01941812e+00
7.58679092e-01 2.62118280e-01 -1.21609545e+00 9.17118728e-01
9.11778286e-02 7.00289547e-01 7.00066805e-01 -7.05314279e-01
4.71850574e-01 4.37997639e-01 -1.41204178e-01 -7.11005569e-01
2.15189070e-01 8.75021964e-02 -9.40016448e-01 1.83796927e-01
1.26373470e-01 -2.23176524e-01 -1.44022477e+00 1.50130415e+00
2.14155912e-01 2.74852626e-02 -1.75365973e-02 8.45025361e-01
8.88949752e-01 8.31020415e-01 3.79668355e-01 9.49301869e-02
8.50903392e-01 -1.47558045e+00 -5.46898544e-01 -3.30645770e-01
3.99020821e-01 -9.29972649e-01 1.31817210e+00 -3.62284720e-01
-1.10683310e+00 -8.35742116e-01 -8.02497566e-01 -3.83274257e-01
-1.43501937e-01 -1.55050427e-01 4.08863366e-01 1.29495040e-01
-1.12974215e+00 5.70802808e-01 -5.75545430e-01 -5.09168744e-01
5.81418455e-01 -1.83101594e-01 -1.23905325e-02 -4.95226055e-01
-1.11407864e+00 7.01228023e-01 -1.19992130e-01 -6.47349536e-01
-1.60887218e+00 -9.24391448e-01 -8.55309367e-01 4.84200008e-02
1.22593388e-01 -1.25748539e+00 1.32985795e+00 -1.15189087e+00
-1.48134875e+00 9.47755694e-01 -1.61041200e-01 -3.90019774e-01
8.80164385e-01 -2.31966656e-03 -9.31489840e-02 4.21625197e-01
4.01242167e-01 1.24969959e+00 1.63732064e+00 -1.13958132e+00
-4.46325570e-01 -2.23022923e-02 -5.95846325e-02 5.28822094e-02
-5.16996086e-01 3.51931527e-02 -5.73101401e-01 -1.08007562e+00
-5.22729218e-01 -6.40743554e-01 -4.99507666e-01 4.64859158e-01
-2.18620032e-01 -5.26430845e-01 1.00025380e+00 -4.62633371e-01
1.03465199e+00 -1.92399359e+00 2.73676097e-01 -2.48822212e-01
3.54610920e-01 1.16616540e-01 -8.22416365e-01 6.70119941e-01
2.62656718e-01 -1.42481923e-01 -4.57823090e-02 -6.66709542e-01
2.23657727e-01 3.47927026e-02 -5.10944247e-01 4.55170162e-02
5.17130971e-01 1.63508773e+00 -1.06738102e+00 -8.83438945e-01
3.08702707e-01 5.66388905e-01 -4.64186728e-01 7.95710325e-01
-7.53735006e-01 2.41166458e-01 -6.08902872e-01 5.55242479e-01
2.82154500e-01 -6.15850568e-01 -2.77421236e-01 -9.98718198e-03
3.11184436e-01 -3.41382772e-01 -4.17034209e-01 2.02506161e+00
-5.37230968e-01 8.77117991e-01 -2.60358274e-01 -7.37552822e-01
9.09435034e-01 7.49158487e-02 4.30911273e-01 -1.28732562e+00
-9.13648158e-02 -4.16673422e-02 -4.38359708e-01 -7.05538630e-01
5.54752886e-01 -1.87934726e-01 -7.89401680e-02 9.92144704e-01
4.11239088e-01 -5.84929585e-01 2.90273935e-01 9.01080310e-01
1.17741740e+00 1.43684745e-01 -1.96368039e-01 2.90271789e-02
-6.17847079e-03 1.39477849e-02 -1.33656368e-01 1.16907334e+00
-3.74473073e-02 9.20652628e-01 2.45638669e-01 -3.17107409e-01
-1.39471364e+00 -1.10485125e+00 3.29837590e-01 1.40801537e+00
3.45255993e-02 -3.19118857e-01 -6.51873589e-01 -9.07779992e-01
2.18822703e-01 6.22651517e-01 -1.00854886e+00 -2.59966373e-01
-1.98079482e-01 -3.09689015e-01 3.66033852e-01 5.99182725e-01
1.81469381e-01 -1.29132497e+00 -2.96350598e-01 2.87613809e-01
-9.99586657e-03 -1.06279767e+00 -1.09940231e+00 -4.85169366e-02
-7.64075518e-01 -7.46474504e-01 -1.52742136e+00 -6.95301473e-01
8.24102044e-01 5.48608005e-01 1.39374602e+00 4.07431601e-03
-6.38078034e-01 8.94850791e-01 -3.98277938e-01 -1.34418771e-01
-7.69208670e-01 6.68206718e-03 -4.02925372e-01 1.32993773e-01
1.39193580e-01 -2.92982489e-01 -1.00102890e+00 1.57645673e-01
-1.29996872e+00 1.90433964e-01 8.50362062e-01 1.04831302e+00
4.66908365e-01 -6.57960355e-01 4.05460149e-01 -9.10171509e-01
1.09382319e+00 -5.38458645e-01 -3.18575382e-01 3.72970879e-01
-8.03903997e-01 4.63637054e-01 4.30847347e-01 -8.37063074e-01
-1.18244004e+00 1.84632111e-02 -3.79483104e-02 -1.01629806e+00
1.66853786e-01 3.35437328e-01 4.14690405e-01 1.30206253e-02
9.65387523e-01 4.42208529e-01 -1.17299698e-01 -2.15093598e-01
9.24866021e-01 4.05821264e-01 3.14042091e-01 -6.48583829e-01
8.81770432e-01 5.52074552e-01 -2.97573864e-01 -6.40045643e-01
-9.93831456e-01 -5.48141062e-01 -4.54370201e-01 -2.30866626e-01
1.07973659e+00 -1.02858424e+00 3.39797884e-02 5.43147087e-01
-1.23015893e+00 -8.74475062e-01 -6.73009813e-01 -9.25429091e-02
-7.59593248e-01 3.94983560e-01 -7.17067063e-01 -5.01854300e-01
-7.80997872e-01 -9.40257072e-01 1.50479615e+00 7.10046962e-02
-3.05732220e-01 -1.26883912e+00 3.80331606e-01 1.68057382e-01
7.88977385e-01 8.25314820e-02 1.00630414e+00 -3.94623488e-01
-9.41948295e-01 -1.82293788e-01 -5.54481506e-01 1.66307658e-01
-7.52253011e-02 -4.70404066e-02 -8.17561328e-01 -3.96950662e-01
-3.47305298e-01 -9.66087401e-01 1.34966242e+00 4.13443297e-01
1.05313206e+00 -2.76368707e-01 -5.15611172e-01 6.22125506e-01
1.32480597e+00 -1.90188840e-01 6.60024345e-01 8.39741342e-03
8.45237851e-01 4.66667861e-01 5.49645483e-01 5.04627228e-01
4.32137609e-01 6.06963634e-01 -4.01239730e-02 -4.43170220e-01
-8.94013345e-01 -9.39558327e-01 3.50035906e-01 5.99959791e-01
3.26256692e-01 -4.24291730e-01 -5.84968448e-01 6.82207704e-01
-1.82454371e+00 -1.26696754e+00 1.72320604e-01 1.59633255e+00
8.64460826e-01 5.57397157e-02 -1.20088300e-02 -5.89412153e-01
5.56931078e-01 7.30453134e-01 -7.65662730e-01 -1.78901657e-01
-1.59542695e-01 2.05647141e-01 4.46117260e-02 4.27536368e-01
-7.40235269e-01 1.23778057e+00 6.41119766e+00 1.22790122e+00
-1.16415083e+00 3.90779197e-01 9.06306565e-01 -2.46490985e-02
-8.42372596e-01 -7.44884536e-02 -6.83814287e-01 3.26049954e-01
7.14832544e-01 -4.00671542e-01 2.19722763e-01 8.60409856e-01
1.55596197e-01 9.18980911e-02 -1.19317722e+00 9.23182368e-01
5.00344217e-01 -1.72210979e+00 7.02742100e-01 -1.19402073e-01
1.54139042e+00 2.79191524e-01 3.44817817e-01 5.17873049e-01
8.64533782e-01 -9.80363071e-01 6.69875503e-01 7.09460378e-01
1.23126185e+00 -1.09952487e-01 4.87418585e-02 2.50499904e-01
-1.00723469e+00 -6.13581389e-02 -4.99259144e-01 4.20614809e-01
2.40416735e-01 4.50063318e-01 -8.72402012e-01 1.50110364e-01
2.52662748e-01 9.99947608e-01 -6.17311120e-01 7.99046934e-01
-1.99555174e-01 3.88398439e-01 1.71591252e-01 -2.18286023e-01
5.55812597e-01 1.08580664e-01 3.26357663e-01 1.36474884e+00
4.99585867e-01 -1.20391212e-01 1.37789011e-01 1.33157289e+00
-3.25261027e-01 7.17005655e-02 -8.33142221e-01 -5.47792614e-01
5.17957881e-02 1.25191855e+00 -5.00528216e-01 -8.36194396e-01
-4.15124506e-01 1.57186818e+00 5.37458479e-01 5.95781624e-01
-5.84909737e-01 -2.63116539e-01 4.18686062e-01 3.97059441e-01
5.41272402e-01 -1.65132686e-01 7.11601675e-02 -1.19088256e+00
-3.47969234e-01 -8.88877392e-01 4.20093983e-01 -1.10205078e+00
-1.80689764e+00 7.57698655e-01 -1.80799589e-01 -1.02163482e+00
-6.18301272e-01 -2.28678033e-01 -9.16221678e-01 8.54347587e-01
-1.56302106e+00 -1.53813374e+00 -5.20399988e-01 8.72501493e-01
1.22955549e+00 -2.81464845e-01 6.34137213e-01 4.45017964e-03
1.83721595e-02 6.66472793e-01 2.54593521e-01 1.09035335e-02
1.03505886e+00 -1.13527787e+00 9.74603176e-01 5.83414137e-01
3.91899228e-01 1.97982058e-01 5.03430724e-01 -8.43483567e-01
-1.32526064e+00 -1.19842982e+00 8.36735487e-01 -7.71326244e-01
7.80189395e-01 -4.73167151e-01 -7.10589707e-01 3.70375961e-01
6.47382557e-01 9.27643999e-02 3.00906301e-01 -2.91499346e-01
-5.38756073e-01 3.77369374e-02 -7.53379941e-01 6.96652293e-01
1.22212875e+00 -8.90617967e-01 -4.76063013e-01 4.91986066e-01
9.68638182e-01 -2.30595574e-01 -4.30945933e-01 3.40103060e-02
3.99194181e-01 -8.25282633e-01 1.04867303e+00 -7.38651037e-01
1.04329574e+00 3.13201964e-01 1.84853047e-01 -1.42153430e+00
-4.18296784e-01 -9.68097031e-01 -2.37375870e-01 1.19826305e+00
3.99096161e-01 -2.38192435e-02 7.42976844e-01 4.71146166e-01
1.51411640e-02 -5.72377980e-01 -5.53831577e-01 -5.84539413e-01
1.60926059e-01 -8.05924609e-02 3.15827519e-01 9.24222112e-01
-1.85079262e-01 7.66568959e-01 -7.02224374e-01 -6.86394870e-01
7.05692947e-01 2.24238068e-01 1.01137805e+00 -9.33352113e-01
-4.70587730e-01 -4.91475791e-01 -4.76577319e-02 -1.55736160e+00
3.83173376e-01 -1.00644410e+00 1.03426173e-01 -1.87071478e+00
5.34465373e-01 -3.93339753e-01 -1.39066428e-01 9.04478356e-02
-5.21474540e-01 1.81879863e-01 3.06087136e-01 3.54389161e-01
-9.54571605e-01 9.99709308e-01 1.95788312e+00 -8.53272557e-01
-2.74451338e-02 -3.10790718e-01 -5.66671312e-01 1.37551904e-01
1.76893786e-01 -6.10643744e-01 -6.64490283e-01 -7.52641201e-01
-5.76310046e-02 3.11045974e-01 2.33224526e-01 -8.03464174e-01
4.56922531e-01 -2.57192522e-01 5.88087022e-01 -3.70564759e-01
4.70040619e-01 -4.19929236e-01 -3.40385258e-01 4.42270815e-01
-8.21856499e-01 7.23886713e-02 -1.06489941e-01 1.00407052e+00
-2.29290903e-01 -1.56704068e-01 6.70843661e-01 -3.49646151e-01
-9.33235765e-01 8.15940022e-01 -3.83841753e-01 4.40942824e-01
1.05850530e+00 -2.17190057e-01 -4.79306012e-01 -9.38534677e-01
-4.73777205e-01 1.86726019e-01 4.70761508e-01 9.40976560e-01
1.02561283e+00 -1.47430325e+00 -9.05770302e-01 2.85090178e-01
5.86155355e-01 -4.61527467e-01 4.61132348e-01 3.99767816e-01
-5.58703877e-02 3.61653090e-01 -2.05781907e-01 -6.20853484e-01
-9.28675354e-01 8.74223769e-01 1.63972601e-01 -3.72360855e-01
-6.97413564e-01 1.14159715e+00 6.04360402e-01 1.23222666e-02
9.56065878e-02 -6.96121976e-02 3.19885343e-01 3.75523306e-02
5.10324657e-01 -8.22435319e-02 -4.45334166e-01 -4.16515172e-01
2.33024731e-01 5.57352602e-01 -4.24759686e-01 -3.26365262e-01
1.19029391e+00 -1.65201887e-01 3.30694795e-01 1.30249426e-01
1.33688402e+00 -2.70017415e-01 -1.80813920e+00 -5.48234582e-01
-1.62377626e-01 -6.90009475e-01 -5.41512780e-02 -7.99083591e-01
-1.04763639e+00 1.33627009e+00 6.32250547e-01 -1.57713056e-01
7.08931386e-01 2.90732414e-01 1.13107789e+00 2.15349123e-01
-3.91935259e-02 -1.22042060e+00 1.35476184e+00 3.83431733e-01
1.05084622e+00 -1.50053918e+00 -3.67229134e-01 2.22511083e-01
-1.05885243e+00 8.74166489e-01 8.25528681e-01 -2.06127301e-01
4.42316294e-01 1.50361761e-01 2.95224071e-01 -2.15966269e-01
-1.23590374e+00 -3.96060348e-01 6.59534872e-01 8.85393322e-01
2.55756319e-01 -2.96793491e-01 1.18909203e-01 1.21748641e-01
4.18527395e-01 1.73017070e-01 1.90383628e-01 7.17805207e-01
-4.73784566e-01 -1.05394626e+00 -1.46501679e-02 5.54455161e-01
-6.16363361e-02 -4.43869978e-01 -6.62137806e-01 3.64678591e-01
-4.24009383e-01 7.35775888e-01 2.64437050e-01 -2.64112979e-01
2.34076068e-01 -6.61335289e-02 6.15411937e-01 -6.45676017e-01
-4.23937470e-01 -1.05963033e-02 -1.09287284e-01 -5.27603567e-01
-3.07007909e-01 -2.93805331e-01 -7.18904555e-01 -2.78233707e-01
2.22761780e-02 -1.17514089e-01 3.80694747e-01 7.45044589e-01
6.16622865e-01 5.95252693e-01 5.16758144e-01 -9.91820157e-01
-7.86965668e-01 -1.09459853e+00 -2.88820446e-01 9.41677451e-01
3.45859051e-01 -4.15868014e-01 -1.63378730e-01 1.72130734e-01] | [11.188199996948242, 0.023235972970724106] |
160715ef-8ef9-49a6-8d30-e2bd83d0ca6f | unrolling-svt-to-obtain-computationally | 2212.08852 | null | https://arxiv.org/abs/2212.08852v1 | https://arxiv.org/pdf/2212.08852v1.pdf | Unrolling SVT to obtain computationally efficient SVT for n-qubit quantum state tomography | Quantum state tomography aims to estimate the state of a quantum mechanical system which is described by a trace one, Hermitian positive semidefinite complex matrix, given a set of measurements of the state. Existing works focus on estimating the density matrix that represents the state, using a compressive sensing approach, with only fewer measurements than that required for a tomographically complete set, with the assumption that the true state has a low rank. One very popular method to estimate the state is the use of the Singular Value Thresholding (SVT) algorithm. In this work, we present a machine learning approach to estimate the quantum state of n-qubit systems by unrolling the iterations of SVT which we call Learned Quantum State Tomography (LQST). As merely unrolling SVT may not ensure that the output of the network meets the constraints required for a quantum state, we design and train a custom neural network whose architecture is inspired from the iterations of SVT with additional layers to meet the required constraints. We show that our proposed LQST with very few layers reconstructs the density matrix with much better fidelity than the SVT algorithm which takes many hundreds of iterations to converge. We also demonstrate the reconstruction of the quantum Bell state from an informationally incomplete set of noisy measurements. | ['Sheetal Kalyani', 'Siva Shanmugam'] | 2022-12-17 | null | null | null | null | ['compressive-sensing', 'quantum-state-tomography'] | ['computer-vision', 'medical'] | [ 4.48508203e-01 1.30238444e-01 1.55036002e-01 -4.16531295e-01
-7.15270042e-01 -5.53061306e-01 3.42445910e-01 -2.62056440e-01
-6.18606806e-01 7.72445261e-01 -1.68093108e-02 -5.96330762e-01
-1.20931208e-01 -8.61729026e-01 -8.70399058e-01 -7.47643411e-01
-1.24453558e-02 8.16540122e-01 -2.37556070e-01 -1.27301022e-01
4.21386659e-01 3.54535580e-01 -1.10390210e+00 6.79478124e-02
4.85770017e-01 8.17635119e-01 -3.40013169e-02 8.43578219e-01
2.16105610e-01 7.69030273e-01 -4.08987164e-01 -7.20033869e-02
5.30547619e-01 -7.87822366e-01 -9.10846114e-01 -1.87094614e-01
4.69887614e-01 -5.71098089e-01 -1.02417886e+00 1.60554707e+00
1.53289109e-01 -7.06938887e-03 5.84064841e-01 -8.39028537e-01
-3.69414121e-01 8.54429066e-01 1.26729175e-01 1.10258833e-01
2.96881557e-01 1.87012609e-02 1.16074300e+00 -5.77499509e-01
5.86919904e-01 8.46227467e-01 4.66426820e-01 2.91439503e-01
-1.50719571e+00 -6.41386986e-01 -8.02636087e-01 4.55875397e-01
-1.78829479e+00 -7.84550846e-01 5.31030476e-01 -9.32946336e-03
1.08120573e+00 1.73220262e-01 7.24361241e-01 6.83326423e-01
2.11876571e-01 3.64472926e-01 1.53576374e+00 -6.20573163e-01
5.71913779e-01 1.28476679e-01 1.41158253e-01 1.08553028e+00
3.76907855e-01 3.92879277e-01 -5.13046563e-01 -2.17868388e-01
6.71133459e-01 -2.31009036e-01 -2.17033193e-01 -4.63689536e-01
-1.37354207e+00 8.06871057e-01 3.28295380e-01 3.76156718e-01
-3.80850464e-01 4.57473814e-01 1.42718181e-01 7.82049477e-01
-3.17964077e-01 5.03776431e-01 7.39735737e-02 -8.40175375e-02
-1.17569566e+00 -4.22728032e-01 1.41213012e+00 7.12748706e-01
1.43055165e+00 -2.29494758e-02 6.35883883e-02 -8.81142169e-02
3.05587441e-01 1.03910863e+00 6.90971762e-02 -1.14518690e+00
2.80917168e-01 8.56415182e-02 2.33933195e-01 -7.22849429e-01
-2.16587007e-01 -2.90809751e-01 -1.25421941e+00 -3.65193449e-02
4.24919128e-01 -3.69579464e-01 -9.18202341e-01 1.75021374e+00
7.52943335e-03 3.38814318e-01 2.92446733e-01 1.02575088e+00
2.73666799e-01 9.12826717e-01 -7.73903251e-01 -2.05710858e-01
8.79002690e-01 -3.02413791e-01 -8.23838353e-01 -3.60290468e-01
6.43020153e-01 -5.58270395e-01 2.81278163e-01 5.25819778e-01
-8.77564609e-01 -1.90897763e-01 -1.51904130e+00 3.03449668e-02
-1.02538399e-01 1.32621825e-01 7.29074299e-01 1.00349462e+00
-1.21920419e+00 9.99720395e-01 -9.81088221e-01 1.37463799e-02
-3.57281696e-03 7.06445396e-01 -8.31924558e-01 -3.36857706e-01
-1.22335029e+00 1.17489505e+00 8.17807615e-01 5.12786031e-01
-1.04417312e+00 2.06058025e-01 -9.50930595e-01 2.30825678e-01
2.75647849e-01 -5.52429557e-01 8.58946681e-01 -3.09636265e-01
-1.75175834e+00 4.97535527e-01 -2.63215005e-01 -5.26383936e-01
-1.51841953e-01 3.58721167e-01 -1.05928995e-01 3.30692351e-01
-2.31951997e-01 9.97754000e-03 9.19936240e-01 -8.60982716e-01
5.52844666e-02 -3.71533275e-01 -2.52291430e-02 -2.06530899e-01
-8.52714330e-02 -4.93162811e-01 -8.79880115e-02 4.90288854e-01
9.28347945e-01 -1.32084477e+00 -2.94707477e-01 -5.40463924e-01
-8.21974397e-01 2.64892399e-01 5.60297072e-01 -3.57386082e-01
9.72425818e-01 -1.92468548e+00 3.92391354e-01 7.47442126e-01
2.76841044e-01 2.35964403e-01 -3.17257732e-01 8.88596475e-01
1.24184564e-02 -3.80818337e-01 -4.50062811e-01 -5.57681680e-01
2.26079285e-01 4.76611376e-01 -1.74598560e-01 1.07495546e+00
-5.56641035e-02 8.33460569e-01 -7.85906255e-01 -1.13417447e-01
1.55534446e-01 2.86849260e-01 -6.39318168e-01 5.37887029e-02
1.66473966e-02 6.57722890e-01 -3.45727891e-01 1.25143841e-01
9.31431949e-01 -4.75435674e-01 5.56236684e-01 -5.09192765e-01
-1.59987718e-01 6.03830874e-01 -1.69681382e+00 1.87257898e+00
-2.96769202e-01 5.57880580e-01 3.11402470e-01 -1.22822177e+00
7.90286541e-01 4.72867340e-01 3.42958629e-01 -6.35058761e-01
5.65425813e-01 4.53205317e-01 3.18762124e-01 -3.91652912e-01
4.18456584e-01 -6.65384114e-01 -2.47587949e-01 1.10299516e+00
4.58040923e-01 -5.27017891e-01 1.43465295e-01 5.82261980e-01
1.37215590e+00 -3.62540543e-01 1.99869946e-01 -4.85597514e-02
3.42811286e-01 -2.53627867e-01 3.41493130e-01 1.15580714e+00
-1.04773670e-01 3.60914946e-01 4.17822450e-01 -1.81189492e-01
-1.34353805e+00 -1.10881972e+00 -1.56228766e-01 4.04224247e-01
5.42214699e-02 -3.85207832e-01 -6.79775119e-01 -1.78881973e-01
-3.62444907e-01 6.45843446e-01 -4.00527656e-01 -4.71036881e-01
-1.87197611e-01 -7.19945669e-01 5.58212876e-01 8.31244979e-03
8.23694050e-01 -8.54531825e-01 -1.39987215e-01 1.28906608e-01
-2.11614549e-01 -1.34833705e+00 -2.49010175e-01 5.93760908e-01
-6.42672718e-01 -8.95528078e-01 -9.01079550e-02 -5.51784694e-01
8.20732355e-01 2.50742659e-02 6.50304139e-01 -1.49904221e-01
1.66973948e-01 2.64474511e-01 -1.17522888e-01 1.60541192e-01
-7.62615025e-01 -1.26571963e-02 3.58120650e-01 2.16542795e-01
4.92202818e-01 -8.34783554e-01 -3.52658123e-01 -2.72251755e-01
-1.13037050e+00 -6.58271015e-02 9.41376567e-01 1.00215888e+00
4.47860986e-01 2.71510988e-01 -7.54733831e-02 -9.04557467e-01
4.38346624e-01 -1.90496817e-01 -8.07542801e-01 1.00411378e-01
-3.75939310e-01 8.04217815e-01 7.19865859e-01 9.19091608e-03
-5.60170472e-01 5.40928543e-01 -1.65782690e-01 -3.53227496e-01
4.90314439e-02 7.23227918e-01 3.28191072e-01 -4.83664244e-01
8.42465043e-01 7.49046385e-01 1.20661899e-01 -1.50319502e-01
4.30084467e-01 8.39573145e-01 5.02677500e-01 -3.61360699e-01
1.22455966e+00 6.57291293e-01 4.89638448e-01 -5.52940547e-01
-1.27035475e+00 -6.83544040e-01 -1.03204608e+00 2.82335490e-01
6.51189864e-01 -7.53063440e-01 -1.12789416e+00 2.68150926e-01
-1.05178583e+00 -6.30389228e-02 -1.27032161e-01 1.01196694e+00
-6.99882686e-01 6.72707558e-01 -7.85568833e-01 -9.71061289e-01
-3.18563074e-01 -1.05078006e+00 8.64695311e-01 -7.42518976e-02
2.00454086e-01 -7.77267933e-01 4.02758807e-01 3.05157930e-01
5.30198455e-01 -1.27141267e-01 6.26668394e-01 -3.44790816e-01
-9.52542126e-01 -6.73093498e-01 -1.85240403e-01 7.14765251e-01
-1.76702872e-01 -6.76282525e-01 -1.01340413e+00 -8.17483962e-01
3.83561313e-01 -5.83371341e-01 1.01763868e+00 6.63917437e-02
6.20913267e-01 -3.91209096e-01 8.62758681e-02 7.71226168e-01
1.64208257e+00 -1.75915584e-01 8.06973755e-01 -1.22929774e-01
6.44418299e-01 1.49001954e-02 -1.95384808e-02 4.79906201e-01
2.19728231e-01 1.00692697e-01 4.78515327e-01 3.42340112e-01
3.77210110e-01 -2.40688369e-01 5.16870916e-01 1.35336339e+00
-2.36097667e-02 6.31621778e-02 -5.77580750e-01 2.13017926e-01
-1.63341928e+00 -1.21189153e+00 -3.73214111e-02 2.57053065e+00
7.97161639e-01 4.74232361e-02 -4.76509839e-01 2.63579816e-01
4.37297314e-01 4.49372530e-02 -4.67061788e-01 -4.76350248e-01
-4.42588590e-02 7.09099114e-01 6.76628470e-01 7.82894313e-01
-9.14239526e-01 8.70092094e-01 6.90819550e+00 2.72959709e-01
-1.07885289e+00 1.36664957e-01 -7.61590600e-02 1.73587620e-01
-2.06519574e-01 4.82466519e-01 -5.11845529e-01 1.43190250e-01
1.45681179e+00 1.90947652e-01 1.21846449e+00 2.99615502e-01
-1.45280585e-02 -2.33790383e-01 -1.34239352e+00 1.08383632e+00
-4.08532247e-02 -1.48193693e+00 -2.64048249e-01 2.93226629e-01
8.52407634e-01 4.81027246e-01 -2.44743079e-02 4.08590287e-01
2.67818421e-01 -1.18926466e+00 5.14864564e-01 6.61777616e-01
9.06250715e-01 -5.22392273e-01 1.05201948e+00 8.00579667e-01
-7.53978848e-01 -6.32956922e-02 -7.16894686e-01 -3.16337734e-01
1.94882467e-01 6.84694588e-01 -1.01141322e+00 2.31957987e-01
1.70483202e-01 4.19223160e-01 -1.20215565e-01 8.10718179e-01
-1.82357877e-01 7.61364937e-01 -7.60481417e-01 -9.53891203e-02
4.56217140e-01 -7.58026242e-01 5.28514981e-01 8.43583643e-01
6.11240447e-01 2.36910522e-01 1.46429837e-01 1.16055012e+00
-1.51087269e-01 -3.05902928e-01 -7.94286788e-01 -5.08857787e-01
5.04678905e-01 1.08055782e+00 -2.29290813e-01 -4.22214895e-01
-1.80431366e-01 1.26608062e+00 4.07423258e-01 2.09883109e-01
-2.48286486e-01 -3.59705716e-01 3.46127748e-02 -2.22316682e-01
5.09674311e-01 -6.12965047e-01 5.68500273e-02 -1.52765763e+00
-8.87030959e-02 -7.33269989e-01 -2.48533376e-02 -4.29249287e-01
-1.13971543e+00 4.70838189e-01 -3.94553900e-01 -1.11299133e+00
-4.13161188e-01 -6.15942538e-01 -4.10165757e-01 1.11954963e+00
-1.44530690e+00 -5.76729834e-01 5.73993474e-02 6.66493714e-01
-7.64787674e-01 -3.69667262e-03 1.32017028e+00 9.12930146e-02
-5.16455948e-01 2.18805328e-01 5.50099850e-01 3.39969665e-01
3.51217508e-01 -1.43179464e+00 6.61485642e-02 9.50345337e-01
4.63680387e-01 1.12854159e+00 7.90917575e-01 -4.49032038e-01
-2.24208808e+00 -6.11494541e-01 7.68441916e-01 -1.53126031e-01
7.93861806e-01 -3.48327667e-01 -7.40264893e-01 1.01767397e+00
2.32054979e-01 3.11605215e-01 7.09023774e-01 7.66631514e-02
-4.47626501e-01 1.03727013e-01 -1.19482088e+00 -2.56743617e-02
5.38143873e-01 -1.23906481e+00 -4.66011971e-01 5.48382998e-01
-3.85267870e-03 -4.38995004e-01 -9.44670081e-01 -3.07208560e-02
3.92745435e-01 -1.14173889e+00 6.13483071e-01 -4.17087168e-01
-1.49080560e-01 -6.78176343e-01 -4.36502606e-01 -1.30673993e+00
-4.69534636e-01 -8.85165036e-01 -1.99695796e-01 1.66456923e-01
3.76315475e-01 -6.55895531e-01 1.09680533e+00 3.36787790e-01
3.17782862e-03 -1.22044496e-01 -1.43429267e+00 -6.08915210e-01
-8.71740058e-02 -3.32873225e-01 2.34982863e-01 8.70760679e-01
2.45108262e-01 7.05562294e-01 -7.37936914e-01 5.98440945e-01
8.77113521e-01 2.71779895e-01 5.41676223e-01 -9.26076412e-01
-6.57487571e-01 2.02549294e-01 -6.63627744e-01 -1.25106514e+00
9.78957489e-02 -1.27453089e+00 3.91989827e-01 -1.34325480e+00
5.08411705e-01 -2.19759360e-01 -4.50101525e-01 3.59593183e-01
3.30245912e-01 5.04969358e-01 1.49536580e-01 1.84259191e-01
-8.54620755e-01 5.90077281e-01 1.27667379e+00 -2.34666094e-01
1.13535374e-01 4.03029332e-03 -3.98530990e-01 2.77478963e-01
5.15217781e-01 -7.56854892e-01 -3.30713391e-02 -3.23550075e-01
6.09963477e-01 5.47547936e-01 2.59341121e-01 -1.24448907e+00
6.65431917e-01 2.19937205e-01 1.13405079e-01 -5.32035470e-01
8.41276169e-01 -6.49103343e-01 2.34247610e-01 6.94275022e-01
-2.50561312e-02 -3.30526382e-01 -2.63124913e-01 6.82513058e-01
-2.28357881e-01 -6.42824829e-01 7.82161832e-01 -3.08953285e-01
-5.16279697e-01 3.10863346e-01 -3.81645828e-01 -4.46602345e-01
4.40562516e-01 8.56840611e-02 -1.67483509e-01 -6.26278400e-01
-9.50009704e-01 -1.76893517e-01 3.73039335e-01 -4.73116130e-01
7.01813459e-01 -1.11251235e+00 -6.03645444e-01 4.58156586e-01
-5.62029406e-02 -3.38626593e-01 1.85429946e-01 1.01096833e+00
-6.88100517e-01 8.05328131e-01 -1.58483028e-01 -4.83429611e-01
-6.85841143e-01 3.46321166e-01 5.51508486e-01 -1.72042966e-01
-4.85592544e-01 6.52830064e-01 -6.56718075e-01 -8.40459645e-01
-1.95791200e-01 -1.67641521e-01 2.68050969e-01 -4.55294400e-01
5.19474506e-01 5.31958765e-04 2.01726139e-01 -8.16030443e-01
-1.16304994e-01 2.65231222e-01 8.40758011e-02 -3.59089702e-01
1.27963173e+00 -3.58395255e-03 -6.83984101e-01 3.35918933e-01
1.62883031e+00 -2.76260525e-02 -8.71695578e-01 -6.32540464e-01
-2.40154296e-01 -2.33002752e-01 4.85011250e-01 -4.57750797e-01
-8.14124048e-01 8.71203303e-01 5.72847247e-01 4.01008785e-01
7.28830874e-01 -8.46081674e-02 9.84665036e-01 1.46558750e+00
8.11834037e-01 -9.69903529e-01 -1.47797450e-01 8.58105898e-01
2.25958601e-01 -1.36419749e+00 -4.60001193e-02 -9.59648192e-02
-1.20094746e-01 1.25119543e+00 -1.74012527e-01 -3.45768660e-01
7.27021575e-01 1.62886217e-01 -2.18266919e-01 -4.17213440e-01
-5.04911184e-01 -2.44501710e-01 1.93640739e-01 3.06079656e-01
1.59076822e-04 2.85749942e-01 2.11511984e-01 -3.30358237e-01
-6.16946161e-01 -1.12525478e-01 9.35059965e-01 7.77870715e-01
-7.39849567e-01 -1.14592135e+00 -5.24159491e-01 5.84779620e-01
-1.73164219e-01 -3.09681952e-01 -1.25173420e-01 9.59336609e-02
-2.04164520e-01 1.00983858e+00 -1.58135653e-01 -5.63317239e-01
-3.18002328e-02 2.75977515e-02 1.01762211e+00 -9.50816095e-01
6.28737360e-02 -3.49873722e-01 7.12347254e-02 -6.60565317e-01
-3.96987528e-01 -6.01823390e-01 -1.24760866e+00 -6.70111835e-01
-5.69992900e-01 4.66759801e-01 8.60428214e-01 1.33026946e+00
8.73816088e-02 2.03303080e-02 8.10842276e-01 -6.68083012e-01
-1.10154045e+00 -1.12617075e+00 -1.00776565e+00 2.31800601e-01
6.13447189e-01 -2.01168597e-01 -5.05463481e-01 -3.96878570e-01] | [5.637819290161133, 4.911802291870117] |
bd5804fe-2df4-466c-8c63-aad6dc4abe03 | uc-owod-unknown-classified-open-world-object | 2207.11455 | null | https://arxiv.org/abs/2207.11455v1 | https://arxiv.org/pdf/2207.11455v1.pdf | UC-OWOD: Unknown-Classified Open World Object Detection | Open World Object Detection (OWOD) is a challenging computer vision problem that requires detecting unknown objects and gradually learning the identified unknown classes. However, it cannot distinguish unknown instances as multiple unknown classes. In this work, we propose a novel OWOD problem called Unknown-Classified Open World Object Detection (UC-OWOD). UC-OWOD aims to detect unknown instances and classify them into different unknown classes. Besides, we formulate the problem and devise a two-stage object detector to solve UC-OWOD. First, unknown label-aware proposal and unknown-discriminative classification head are used to detect known and unknown objects. Then, similarity-based unknown classification and unknown clustering refinement modules are constructed to distinguish multiple unknown classes. Moreover, two novel evaluation protocols are designed to evaluate unknown-class detection. Abundant experiments and visualizations prove the effectiveness of the proposed method. Code is available at https://github.com/JohnWuzh/UC-OWOD. | ['Junzhi Yu', 'Liwen Kang', 'Zhengxing Wu', 'Xingyu Chen', 'Yue Lu', 'Zhiheng Wu'] | 2022-07-23 | null | null | null | null | ['open-world-object-detection'] | ['computer-vision'] | [-2.90854406e-02 -3.70924473e-02 -1.76315233e-01 -2.28254929e-01
-7.87475586e-01 -7.29936659e-01 3.10720503e-01 1.84897706e-01
-8.02911744e-02 6.84062719e-01 -3.30939412e-01 4.85863425e-02
-2.13994414e-01 -6.38901353e-01 -3.77434999e-01 -6.69788122e-01
5.89443482e-02 7.59889305e-01 7.08873332e-01 4.66641158e-01
2.09238619e-01 4.75841820e-01 -1.74175322e+00 6.13095425e-02
1.00659859e+00 1.08108592e+00 5.20496547e-01 6.93022609e-01
-2.20895842e-01 5.37282288e-01 -5.60157180e-01 4.94881719e-02
4.89576876e-01 -3.39230627e-01 -8.57096136e-01 4.07094628e-01
6.39421582e-01 -3.57784688e-01 -7.27065280e-02 1.14594209e+00
5.66743612e-01 2.66532272e-01 8.22004795e-01 -1.69142592e+00
-9.57979679e-01 2.76399761e-01 -5.73069036e-01 5.51231802e-01
-1.48196770e-02 2.49927938e-01 1.03055131e+00 -9.92817461e-01
7.65745640e-01 1.11515796e+00 4.52806562e-01 4.44731981e-01
-1.30179155e+00 -8.16348732e-01 2.91189253e-01 6.74580634e-01
-1.70401657e+00 -2.38564298e-01 6.27654433e-01 -6.09493911e-01
1.44945294e-01 3.92509282e-01 2.31771991e-01 5.33181787e-01
-2.55524725e-01 1.02770102e+00 1.01727927e+00 -1.07116207e-01
3.42904598e-01 4.61516112e-01 7.70920157e-01 6.32719398e-01
7.78525293e-01 -7.84365609e-02 9.32070687e-02 -1.11500248e-01
1.72113970e-01 3.64284188e-01 -4.12145942e-01 -6.01513624e-01
-1.07901394e+00 3.92860532e-01 7.13322937e-01 1.62342474e-01
3.58847603e-02 -3.95146668e-01 1.26397192e-01 2.14472294e-01
3.82343650e-01 3.36102277e-01 -5.03878176e-01 4.18034166e-01
-5.30636370e-01 -6.09918609e-02 6.67904973e-01 1.05069029e+00
1.04174662e+00 -2.69164413e-01 -2.56117940e-01 1.07058072e+00
3.01960588e-01 5.57194889e-01 5.85896850e-01 -7.71531165e-01
2.70325523e-02 8.86019051e-01 1.61297366e-01 -1.09882843e+00
-4.98144686e-01 -4.94513124e-01 -5.78656316e-01 1.60748065e-01
4.61841911e-01 2.12491509e-02 -1.02244842e+00 1.13685799e+00
9.59281802e-01 8.29257965e-01 1.95084646e-01 1.12391317e+00
1.20555294e+00 7.27389991e-01 -2.03146592e-01 -2.57921278e-01
1.39859402e+00 -1.12787974e+00 -6.48927927e-01 -1.62756577e-01
5.35497546e-01 -5.87081730e-01 7.79303491e-01 1.40540823e-01
-4.81489062e-01 -6.73226535e-01 -8.73879194e-01 4.14333716e-02
-5.06446362e-01 3.92925888e-01 2.98421443e-01 4.59130377e-01
-3.84319067e-01 8.50864202e-02 -5.89109421e-01 -3.94695431e-01
8.83184910e-01 1.98147535e-01 -1.76853463e-01 -4.34193224e-01
-5.57309330e-01 5.07840037e-01 9.08559978e-01 -1.08019803e-02
-1.28005826e+00 -4.37565863e-01 -6.54728889e-01 -6.89899027e-02
7.96059310e-01 -4.31458205e-01 1.21798646e+00 -8.66995215e-01
-9.70364928e-01 8.38995516e-01 -1.23428099e-01 -1.54306322e-01
3.61412764e-01 7.50416592e-02 -6.07434630e-01 1.09776691e-01
4.11797762e-01 3.20262462e-01 8.53471160e-01 -1.57731378e+00
-1.13141787e+00 -6.82640970e-01 3.37577946e-02 7.87241384e-02
-3.98192376e-01 -8.46999660e-02 -5.78947008e-01 -5.46926856e-01
6.62630618e-01 -8.01551104e-01 -1.46879166e-01 4.22210276e-01
-6.11357391e-01 -5.57878554e-01 1.17263126e+00 -3.55422258e-01
1.04290676e+00 -2.20682669e+00 -1.97100848e-01 1.09825104e-01
8.47811162e-01 2.76248187e-01 -2.34808758e-01 -1.83929101e-01
1.23620272e-01 -2.41895929e-01 -3.17611903e-01 -3.92871015e-02
-3.44463577e-03 2.80749202e-01 -2.75873728e-02 5.92958212e-01
3.10836345e-01 6.32592440e-01 -8.31973672e-01 -5.02246857e-01
1.53920904e-01 1.06965960e-03 -2.76671387e-02 3.24292213e-01
-3.54641557e-01 3.46535444e-01 -5.41003525e-01 1.03783119e+00
1.10581303e+00 -4.41488177e-01 5.57923131e-02 -9.23497677e-02
1.87580902e-02 -4.92700458e-01 -1.91394591e+00 8.02044153e-01
-8.59721228e-02 5.24591029e-01 -4.46111942e-03 -1.05203593e+00
9.17998791e-01 -4.55592759e-02 1.14511520e-01 -2.02668190e-01
4.60744500e-01 5.01394510e-01 -3.75278182e-02 -6.87663555e-01
3.15099895e-01 1.93155199e-01 3.29914242e-01 3.25393647e-01
2.72744298e-01 1.78342193e-01 3.27396423e-01 2.00266898e-01
9.78869081e-01 -8.33147392e-02 4.54716414e-01 -2.74626166e-01
6.74431741e-01 1.90525725e-01 9.25671220e-01 9.14931178e-01
-5.12676299e-01 7.74544060e-01 9.39877108e-02 -4.69449520e-01
-6.17865503e-01 -1.23212814e+00 -4.90282565e-01 1.05572677e+00
9.13286746e-01 4.31623980e-02 -4.98424828e-01 -8.85790050e-01
1.10075973e-01 3.63660574e-01 -4.84983742e-01 -8.20125919e-03
-5.19558461e-03 -6.99707031e-01 1.83323026e-01 2.86578417e-01
3.59764278e-01 -8.19775045e-01 -3.02568853e-01 1.06809236e-01
-2.90518939e-01 -1.06819642e+00 -5.22641599e-01 5.63962162e-02
-5.47648430e-01 -1.71405470e+00 -8.27781737e-01 -1.24559939e+00
1.05280221e+00 7.63152897e-01 6.42069161e-01 1.18482724e-01
-7.71090031e-01 3.08884382e-01 -4.09839660e-01 -6.15180016e-01
-1.65469229e-01 -2.95544177e-01 2.88719445e-01 6.02202594e-01
4.15823370e-01 -1.40147269e-01 -5.03967643e-01 7.91641414e-01
-7.49952614e-01 -3.74249607e-01 5.01608372e-01 6.97327673e-01
8.15848053e-01 2.31026962e-01 4.73214835e-01 -7.55306065e-01
1.85834453e-01 -7.32335389e-01 -9.96040821e-01 3.77564698e-01
-5.39060056e-01 -2.31639013e-01 5.60720086e-01 -9.16814625e-01
-8.85857105e-01 2.16527447e-01 3.28961819e-01 -6.15745008e-01
-6.05943143e-01 1.83320362e-02 -4.33848381e-01 1.76141541e-02
6.65728092e-01 1.88361123e-01 -3.91424835e-01 -6.00060046e-01
3.35704297e-01 1.19924784e+00 7.42167234e-01 -3.86341512e-01
1.10077727e+00 7.52439082e-01 -5.17520547e-01 -9.21326935e-01
-1.07336926e+00 -1.00226438e+00 -6.92507923e-01 -2.08327055e-01
8.19342375e-01 -1.00909746e+00 -3.80170763e-01 5.31506002e-01
-1.00695002e+00 -4.38837409e-02 -4.06504065e-01 4.80887681e-01
-1.22630306e-01 4.18095052e-01 -2.16402397e-01 -8.74476969e-01
-2.87558794e-01 -9.10989583e-01 9.22079623e-01 5.82099438e-01
1.92504063e-01 -5.85605383e-01 -1.08661942e-01 6.50877535e-01
-2.04880074e-01 5.23550697e-02 2.73591578e-01 -1.15723050e+00
-7.66912699e-01 -3.14245284e-01 -6.64963663e-01 1.99953735e-01
2.03574717e-01 4.74198982e-02 -9.17977929e-01 -3.10565978e-01
-2.81439304e-01 -1.98555216e-01 7.78526604e-01 2.01689601e-01
1.19283807e+00 -1.96367905e-01 -6.36110842e-01 5.14968753e-01
1.28297555e+00 2.11777851e-01 1.23858526e-01 3.14459890e-01
6.71615958e-01 3.70524853e-01 8.37478697e-01 5.77178419e-01
4.46060479e-01 4.90686774e-01 2.71130651e-01 4.16377932e-02
-1.67515695e-01 6.04200028e-02 1.20313682e-01 3.99026066e-01
3.13385159e-01 -2.90203840e-01 -1.06481707e+00 7.68771887e-01
-1.85416794e+00 -8.55481029e-01 -6.64136469e-01 1.96296501e+00
4.47986007e-01 2.29492813e-01 4.01609428e-02 1.82141960e-01
1.29190707e+00 -3.90434235e-01 -8.27739596e-01 4.08801883e-01
-7.55678043e-02 -2.50125766e-01 1.27666414e-01 2.16220111e-01
-1.37874305e+00 6.75068080e-01 4.47470427e+00 9.80984807e-01
-7.76384532e-01 4.41631764e-01 5.02245247e-01 2.79249489e-01
2.96996683e-01 9.62221920e-02 -1.01104665e+00 4.59828168e-01
1.54241100e-01 -3.40322435e-01 7.96768442e-03 9.96173501e-01
1.90297402e-02 -1.82287246e-01 -8.68294954e-01 1.11413848e+00
2.21495047e-01 -1.14709508e+00 -1.55987203e-01 -3.44726324e-01
8.48841012e-01 1.32922515e-01 -4.04026985e-01 2.89203823e-01
4.55201566e-01 -1.70037881e-01 5.87280273e-01 2.79085577e-01
4.27470565e-01 -3.31942111e-01 6.36844993e-01 5.74959517e-01
-1.42867076e+00 -7.23173916e-01 -6.15338027e-01 6.98913708e-02
-2.25142077e-01 5.28387606e-01 -6.20022297e-01 5.01930714e-01
1.00530064e+00 7.98774600e-01 -7.14378655e-01 1.72804749e+00
-1.73634499e-01 6.80525899e-01 -4.43508446e-01 1.63067922e-01
4.84595895e-02 -1.71801299e-01 7.02604353e-01 7.59398937e-01
1.60625502e-01 4.60898697e-01 6.99263394e-01 8.86775017e-01
-1.28039584e-01 6.62050024e-02 -2.88139343e-01 2.39791647e-01
4.61697906e-01 1.48953235e+00 -1.12110865e+00 -5.27817845e-01
-3.26061189e-01 1.00447881e+00 2.06821665e-01 2.01706320e-01
-8.72510314e-01 -6.10467911e-01 6.08069241e-01 1.32814888e-02
4.87021357e-01 1.58173144e-02 4.65526767e-02 -1.39853776e+00
7.11907595e-02 -4.41080689e-01 9.52617764e-01 -8.26886594e-01
-1.60572982e+00 4.11097109e-01 -2.10933819e-01 -1.62695432e+00
8.45341682e-01 -6.54516757e-01 -6.41338348e-01 2.23102793e-01
-1.70791233e+00 -1.08474231e+00 -8.08889091e-01 5.60023427e-01
7.79734373e-01 -1.24914773e-01 4.26418602e-01 4.25803840e-01
-9.76363599e-01 3.58351141e-01 7.97057033e-01 3.01997691e-01
6.29816473e-01 -1.12126923e+00 -2.52879649e-01 9.70984817e-01
1.83993354e-01 1.84999123e-01 4.14766163e-01 -5.63229799e-01
-1.04342377e+00 -1.55187738e+00 3.47408742e-01 -4.69289511e-01
8.53132486e-01 -4.21184450e-01 -1.09063971e+00 5.64407170e-01
-4.54960972e-01 8.92435908e-01 5.12792826e-01 -2.61177957e-01
-4.13875669e-01 -2.14620978e-01 -1.20751560e+00 2.91449130e-01
9.98667419e-01 -1.51024804e-01 -7.14383364e-01 6.20709240e-01
8.94729078e-01 -3.32075238e-01 -6.53259099e-01 2.57170439e-01
2.58921325e-01 -5.09619653e-01 8.98609579e-01 -5.69284081e-01
2.14333050e-02 -9.43231761e-01 -1.32360473e-01 -9.71096456e-01
-2.42372751e-01 -1.90799192e-01 -3.37915033e-01 1.38036036e+00
2.21513391e-01 -9.06283200e-01 6.01860881e-01 2.53078133e-01
-7.65461847e-02 -2.42331967e-01 -8.20471823e-01 -1.28643870e+00
-2.86098093e-01 -3.00904810e-01 3.23584169e-01 1.23719466e+00
-4.25793678e-01 3.87740552e-01 3.91241303e-03 9.16799426e-01
8.34760666e-01 6.28955960e-01 8.61735761e-01 -1.66722870e+00
-1.65587097e-01 -1.27334937e-01 -8.54667723e-01 -9.45487618e-01
-3.99975814e-02 -1.07755661e+00 1.96843252e-01 -1.52772593e+00
4.88032252e-01 -7.04130709e-01 -5.87132335e-01 5.84582210e-01
-4.52571183e-01 5.79681873e-01 2.69813091e-01 4.99999195e-01
-1.24791932e+00 5.99840045e-01 1.01659203e+00 -4.66588587e-01
-3.21585596e-01 2.27840602e-01 -5.44712603e-01 6.36184156e-01
9.21684384e-01 -7.55752504e-01 -1.10136203e-01 -1.60674036e-01
-6.66630566e-01 -3.40504169e-01 6.37452126e-01 -1.28526652e+00
2.85317689e-01 -1.62120670e-01 3.22304994e-01 -6.84875965e-01
-4.18512970e-02 -9.77585196e-01 1.12864478e-02 7.06770837e-01
-2.28986684e-02 -6.37230754e-01 5.38006872e-02 8.14264476e-01
-6.55717868e-03 -2.72461712e-01 1.09458542e+00 7.92368054e-02
-1.30643094e+00 4.30393010e-01 -2.43054658e-01 2.58997560e-01
1.73519790e+00 -2.10412830e-01 -5.23004830e-01 1.38614580e-01
-9.55671310e-01 7.82737553e-01 1.06607445e-01 6.54324770e-01
7.70217717e-01 -1.09103441e+00 -8.67725790e-01 1.06791072e-01
8.13008070e-01 1.44398391e-01 2.76508361e-01 7.14504540e-01
-4.57904607e-01 -1.11029968e-01 9.15491283e-02 -8.05019498e-01
-1.76074171e+00 8.53494644e-01 4.22356904e-01 2.78385222e-01
-5.66835046e-01 7.34261036e-01 1.57498524e-01 -1.02880335e+00
3.40917528e-01 9.98581126e-02 -3.03845614e-01 1.48172587e-01
6.21193528e-01 6.17988467e-01 -1.94555193e-01 -6.33306801e-01
-3.06882977e-01 5.10203302e-01 -2.15108953e-02 8.53764951e-01
1.24813807e+00 -3.72678608e-01 -2.30908915e-01 4.15242344e-01
1.00242782e+00 -5.35345078e-01 -1.04658389e+00 -7.15644658e-01
2.53021806e-01 -5.88557124e-01 -9.62557122e-02 -6.43751264e-01
-8.70954096e-01 7.18899667e-01 1.26802528e+00 1.67446196e-01
1.00985193e+00 3.85748476e-01 6.74763024e-01 6.39977455e-01
2.53634304e-01 -1.00042760e+00 6.45628273e-02 3.92680377e-01
5.33741355e-01 -1.63721621e+00 6.21906407e-02 -7.30919540e-01
-2.83191442e-01 9.61846054e-01 1.16751838e+00 1.73875928e-01
7.37925291e-01 2.28813924e-02 3.36555839e-01 -2.18299460e-02
-3.76701772e-01 -7.44732559e-01 3.33152652e-01 6.25996172e-01
-3.46946985e-01 1.01343043e-01 -2.55168140e-01 5.74692070e-01
4.28137392e-01 1.12404525e-02 5.52688897e-01 8.73247564e-01
-7.30958462e-01 -6.23218834e-01 -8.57803404e-01 7.13102341e-01
-1.94881819e-02 2.45752379e-01 -3.41284513e-01 3.37416977e-01
6.72737598e-01 1.00475752e+00 1.16276748e-01 -2.42979929e-01
1.61268800e-01 -2.63319373e-01 -1.31399468e-01 -7.20749140e-01
-1.08951300e-01 -7.90218934e-02 -3.56073886e-01 -3.43094707e-01
-3.21515560e-01 -3.62096906e-01 -1.43628955e+00 2.36171737e-01
-1.07206893e+00 2.43370071e-01 2.87341356e-01 6.10575974e-01
4.23950046e-01 3.93138111e-01 8.26528132e-01 -7.67374039e-01
-4.08995956e-01 -5.33836305e-01 -7.59427190e-01 3.59281927e-01
5.10127068e-01 -8.54732513e-01 -4.84575480e-01 1.42904103e-01] | [9.263511657714844, 1.474398136138916] |
3f4adc3e-27f4-466b-8735-56ecc02fee8a | credence-counterfactual-explanations-for | 2302.04983 | null | https://arxiv.org/abs/2302.04983v1 | https://arxiv.org/pdf/2302.04983v1.pdf | CREDENCE: Counterfactual Explanations for Document Ranking | Towards better explainability in the field of information retrieval, we present CREDENCE, an interactive tool capable of generating counterfactual explanations for document rankers. Embracing the unique properties of the ranking problem, we present counterfactual explanations in terms of document perturbations, query perturbations, and even other documents. Additionally, users may build and test their own perturbations, and extract insights about their query, documents, and ranker. | ['Jaroslaw Szlichta', 'Divesh Srivastava', 'Mehdi Kargar', 'Lukasz Golab', 'Parke Godfrey', 'Joel Rorseth'] | 2023-02-10 | null | null | null | null | ['document-ranking'] | ['natural-language-processing'] | [ 4.05224681e-01 7.71571457e-01 -3.92034769e-01 -1.29604757e-01
-6.59606874e-01 -9.97531414e-01 1.30846155e+00 3.30392361e-01
4.63359989e-02 1.04289591e+00 8.92616689e-01 -9.74435031e-01
-7.17913806e-01 -5.09817123e-01 -6.89862549e-01 -7.27054030e-02
-3.44389290e-01 8.93326163e-01 -1.00625895e-01 -3.70759279e-01
8.79843175e-01 5.33494473e-01 -1.80306339e+00 6.62876964e-01
7.85948634e-01 1.81372166e-01 -1.89566270e-01 8.96744132e-01
-1.80364832e-01 5.51547110e-01 -1.00239849e+00 -7.87750006e-01
1.40224442e-01 -7.56015927e-02 -1.02109611e+00 -2.05541760e-01
3.45130146e-01 -3.29616517e-01 -4.05427128e-01 8.35872591e-01
4.07732338e-01 2.84984887e-01 8.26816618e-01 -1.52334380e+00
-1.00792563e+00 1.12372768e+00 -4.02720645e-02 2.45555088e-01
1.16480160e+00 4.83932123e-02 1.21656621e+00 -5.84522605e-01
8.00400496e-01 1.68472111e+00 -1.04840301e-01 5.49381852e-01
-1.41783369e+00 -4.89029467e-01 3.75782788e-01 2.71355599e-01
-8.36404562e-01 -2.87561059e-01 4.86827582e-01 -2.79825538e-01
6.16792977e-01 1.12265539e+00 4.68669415e-01 1.34290040e+00
5.02606258e-02 8.89711857e-01 9.69426990e-01 -7.48134911e-01
3.06471616e-01 4.24074978e-01 2.02233061e-01 3.42931062e-01
7.34712064e-01 7.04437554e-01 -7.85521507e-01 -8.75823736e-01
9.76463184e-02 -1.49576709e-01 -6.18320167e-01 -5.89078605e-01
-1.20109856e+00 6.84530079e-01 2.03552559e-01 -9.34315287e-03
-3.65364820e-01 1.08409204e-01 1.76084086e-01 3.77700806e-01
2.54314542e-01 1.39569354e+00 -6.18654370e-01 4.87789623e-02
-5.01075387e-01 8.14216554e-01 1.04948545e+00 7.80859768e-01
2.86994517e-01 -3.49453241e-01 -9.47343946e-01 2.20389441e-01
2.68303722e-01 5.09971738e-01 7.46261597e-01 -8.42854619e-01
3.59764487e-01 4.30905700e-01 6.63458943e-01 -7.25528240e-01
-1.78742837e-02 -3.83847654e-01 -1.84503317e-01 8.12427234e-03
3.71257626e-02 1.03649378e-01 -8.56399357e-01 1.36524057e+00
3.57938074e-02 -2.57774264e-01 2.76973337e-01 7.44139373e-01
4.77011144e-01 1.91349506e-01 9.99911204e-02 -5.56591392e-01
1.01407170e+00 -4.96972501e-01 -8.21588695e-01 -1.44902945e-01
8.62412453e-01 -4.77031052e-01 1.17678595e+00 3.55850846e-01
-8.84010911e-01 -5.41466773e-02 -9.89997447e-01 2.36759931e-01
-5.80017090e-01 -3.21639091e-01 1.05185604e+00 4.88957465e-01
-8.96754622e-01 8.21491659e-01 -5.83663248e-02 2.76107304e-02
1.05680814e-02 3.18173021e-01 9.75198820e-02 -2.20309243e-01
-1.59432423e+00 8.05298805e-01 3.97853553e-01 -3.35792124e-01
-7.72986650e-01 -8.88330400e-01 -5.28637528e-01 4.32548434e-01
4.39199120e-01 -9.97317374e-01 1.52784479e+00 -1.27979308e-01
-9.68005478e-01 3.99465114e-01 -1.20665647e-01 -3.83053660e-01
6.12249851e-01 -2.77267933e-01 -6.61143184e-01 -1.74574450e-01
2.03635573e-01 3.28556269e-01 4.29770380e-01 -1.60571742e+00
-3.78401875e-01 -6.23896301e-01 3.58971119e-01 5.01425207e-01
1.29649803e-01 -2.99185514e-01 -5.62654324e-02 -4.32073653e-01
8.33837986e-02 -7.62654603e-01 -4.99530584e-01 -6.83875978e-01
-1.18083143e+00 -1.40557870e-01 6.25799954e-01 -1.54622704e-01
1.28275108e+00 -1.73154402e+00 -2.97982007e-01 6.49119854e-01
1.47754475e-01 -1.98547259e-01 -1.67462081e-01 6.46478832e-01
-5.08334339e-01 1.03422713e+00 4.50615317e-01 2.95106173e-01
3.61704111e-01 -1.36787564e-01 -8.66489887e-01 -1.91266507e-01
-3.95711839e-01 8.57303679e-01 -1.03823650e+00 -2.07233042e-01
8.58255029e-02 -1.64474592e-01 -6.42858207e-01 -1.85331458e-03
-5.85620224e-01 -8.63759220e-02 -6.05419159e-01 3.79861325e-01
3.27540994e-01 -6.58729821e-02 4.89321887e-01 2.97591627e-01
5.66820726e-02 6.52898312e-01 -9.67460632e-01 1.08556461e+00
-3.19448143e-01 6.52212739e-01 -5.82620978e-01 -2.83174664e-01
3.87280613e-01 4.77799237e-01 3.61506641e-02 -3.65505248e-01
-2.64349610e-01 1.69276416e-01 -1.13573335e-01 -4.35645103e-01
7.31430292e-01 -3.90115567e-02 3.73242162e-02 1.04311168e+00
-5.90038717e-01 -3.07222486e-01 4.92493391e-01 8.34987760e-01
1.12494111e+00 -4.51165259e-01 5.07789016e-01 -9.59577262e-02
4.85251136e-02 1.07696109e-01 -2.66956687e-01 1.57560503e+00
4.23726588e-01 4.45965171e-01 6.74549997e-01 -3.89828920e-01
-7.08406627e-01 -1.16398513e+00 1.99948121e-02 1.18036544e+00
1.41866475e-01 -4.65243369e-01 -1.16485551e-01 -9.03804481e-01
5.01036286e-01 1.52696824e+00 -8.86081755e-01 -3.58148873e-01
1.82319328e-01 -4.55098033e-01 1.62824601e-01 4.16935049e-03
-1.17892258e-01 -9.51204240e-01 -2.76094049e-01 -1.28023535e-01
-3.00184965e-01 -2.64637113e-01 -4.60172355e-01 -1.76784530e-01
-1.01235259e+00 -1.46737385e+00 -3.01469594e-01 1.18549719e-01
5.89872837e-01 4.09410924e-01 1.38503540e+00 3.62008452e-01
-2.73988731e-02 5.57220578e-01 -3.07772845e-01 -8.37531805e-01
-6.20393097e-01 -3.62637013e-01 1.04621314e-01 -7.23446786e-01
1.96423382e-01 -3.79654557e-01 -5.39079785e-01 8.90987068e-02
-1.12186849e+00 -1.52551368e-01 5.91264725e-01 7.57282019e-01
1.20884642e-01 8.64174291e-02 2.09906101e-01 -1.37849319e+00
1.61940455e+00 -4.88724232e-01 -3.57933760e-01 7.64992177e-01
-1.32631230e+00 6.20782673e-01 2.16284946e-01 -4.75510836e-01
-1.18361318e+00 -4.33838934e-01 6.68260217e-01 1.50888830e-01
-1.16155840e-01 7.35181093e-01 -2.14946419e-01 5.73350251e-01
1.33553827e+00 9.49047133e-02 -3.08262408e-01 -4.37316269e-01
1.01184678e+00 4.45693016e-01 4.74312186e-01 -5.12524486e-01
9.73969579e-01 3.66697699e-01 -1.43728256e-01 -2.41231974e-02
-6.85070336e-01 -2.37902567e-01 -5.59076443e-02 -5.62948361e-02
-1.85845234e-02 -3.11390430e-01 -9.41458762e-01 -5.39139152e-01
-1.24974835e+00 6.12675659e-02 -4.90415365e-01 4.55883473e-01
-4.80896264e-01 1.06822208e-01 -5.17065520e-04 -8.89180303e-01
-7.61904344e-02 -8.70407522e-01 7.82145202e-01 4.79425713e-02
-7.48656094e-01 -8.84026289e-01 1.98430374e-01 2.49833256e-01
2.34116331e-01 8.89199898e-02 1.42335546e+00 -1.37082589e+00
-8.58703375e-01 -6.78061545e-01 4.61881310e-02 -6.20428503e-01
-1.81396708e-01 2.92588741e-01 -7.23266602e-01 -1.78269953e-01
-4.28044111e-01 -4.69046012e-02 5.59295595e-01 2.51280099e-01
1.25501466e+00 -1.11663222e+00 -8.34507585e-01 -8.85980651e-02
9.05949652e-01 2.26548165e-01 5.93662918e-01 6.66073501e-01
3.38220932e-02 7.94514477e-01 6.71859920e-01 6.48175418e-01
-7.27945343e-02 5.40963352e-01 4.71939802e-01 2.02045575e-01
3.68609101e-01 -6.83796287e-01 7.63723906e-03 -3.73784125e-01
8.00697207e-02 -6.84987903e-01 -6.25493467e-01 4.49642569e-01
-1.68365264e+00 -1.19279408e+00 -3.47198881e-02 2.17685676e+00
4.89052683e-01 2.23962188e-01 -3.04355651e-01 1.61343902e-01
6.32374108e-01 6.44923374e-02 -5.27490973e-01 -6.64650500e-01
-4.65470254e-02 -2.05952853e-01 3.31570864e-01 6.85006082e-01
-3.51939797e-01 6.23736143e-01 8.37847805e+00 2.55641729e-01
-4.94577020e-01 -6.62151396e-01 5.09790897e-01 -5.13319016e-01
-1.46417964e+00 3.58384758e-01 -1.76182643e-01 1.57513991e-01
7.82521367e-01 -1.13951087e+00 7.00406015e-01 9.48576450e-01
3.09256822e-01 4.61152084e-02 -1.59880710e+00 2.06979007e-01
-2.73608595e-01 -1.55061388e+00 8.79697263e-01 2.94154346e-01
7.53338158e-01 -4.72694039e-01 2.72434503e-01 3.23214144e-01
1.06672716e+00 -9.47226346e-01 6.77249968e-01 5.44272482e-01
6.99309170e-01 -6.00883842e-01 7.24816084e-01 4.45432723e-01
-9.90505666e-02 -3.28594863e-01 -3.41477282e-02 -3.44670296e-01
-1.32371217e-01 4.60219204e-01 -1.41592431e+00 6.61427915e-01
3.44945580e-01 -9.03893858e-02 -7.72076845e-01 7.78620839e-01
-4.41802979e-01 3.38663578e-01 1.16854593e-01 -2.43030682e-01
-3.68381470e-01 2.01245323e-01 6.75855815e-01 8.98709178e-01
3.60318094e-01 4.20830220e-01 -2.36451954e-01 6.51728690e-01
-9.21292379e-02 -2.52729654e-03 -9.93881822e-01 -7.09243566e-02
9.64892685e-01 6.09861791e-01 -1.76946923e-01 -7.36178517e-01
4.47136492e-01 6.62584305e-01 3.53074819e-03 8.77781034e-01
-1.61251709e-01 -2.70910174e-01 7.05431223e-01 1.51610389e-01
-3.61952364e-01 5.55172861e-01 -4.32804406e-01 -1.04361737e+00
-5.62153682e-02 -1.34969974e+00 6.61878288e-01 -1.24032331e+00
-8.69991899e-01 2.71214902e-01 5.54499090e-01 -7.77290404e-01
-9.33181405e-01 -3.63655627e-01 -9.45276737e-01 5.69223762e-01
-7.68902302e-01 -4.18520004e-01 9.93951634e-02 1.77895024e-01
1.41105652e-01 -2.11467966e-01 8.98947537e-01 -5.63464463e-01
-2.13979915e-01 3.96668851e-01 2.48672813e-01 -6.04823351e-01
4.02105302e-01 -1.48685610e+00 5.81314147e-01 6.22708738e-01
4.76844907e-01 1.38173640e+00 1.49795234e+00 -7.43799865e-01
-9.56584513e-01 -4.13399875e-01 1.11016786e+00 -7.40464926e-01
4.84706759e-01 -1.52717186e-02 -5.87879956e-01 6.69409931e-01
1.35723174e-01 -7.96657085e-01 8.15209031e-01 8.05492640e-01
-3.16158742e-01 2.13835448e-01 -1.11261761e+00 1.31614316e+00
9.94769931e-01 -5.91548145e-01 -8.81577134e-01 8.13009739e-01
9.10885632e-01 -2.23268226e-01 -8.49192366e-02 4.05072629e-01
6.79487586e-01 -1.20835757e+00 1.11961412e+00 -1.31465471e+00
4.53052998e-01 -1.18330605e-01 3.17340530e-02 -1.61125863e+00
-3.55056107e-01 -8.56404960e-01 -2.10509047e-01 7.50474751e-01
8.71651828e-01 -4.92826700e-01 8.19817305e-01 1.42728412e+00
1.06287330e-01 -4.73036021e-01 -5.07301867e-01 -5.49466431e-01
-4.08948511e-01 -5.49384117e-01 1.23165476e+00 7.60966003e-01
6.22251451e-01 1.41611785e-01 -1.08503126e-01 2.00798303e-01
4.52580869e-01 3.39088440e-01 8.55265796e-01 -1.23313296e+00
-4.51073319e-01 -4.65047628e-01 1.78023323e-01 -8.53735030e-01
7.68925473e-02 -8.02566707e-01 -5.40838018e-02 -1.71538997e+00
3.77581179e-01 5.53190708e-03 -3.08808893e-01 1.91648915e-01
-4.38391984e-01 -2.18137547e-01 1.63133591e-01 4.18926120e-01
-4.14890081e-01 3.52670163e-01 1.18950558e+00 -1.03911914e-01
-2.46050194e-01 3.29213798e-01 -1.45596373e+00 6.36800110e-01
6.73752069e-01 -6.46212280e-01 -7.71446466e-01 -7.46004581e-02
5.48988104e-01 3.20971787e-01 3.94903332e-01 -2.88404077e-01
1.99922416e-02 -4.83859837e-01 3.98537070e-01 -6.11976266e-01
6.37363344e-02 -4.11793500e-01 2.97531158e-01 7.25385785e-01
-1.13470447e+00 2.70125031e-01 9.68557671e-02 9.69126403e-01
-9.31934044e-02 -3.21744472e-01 -3.06830168e-01 -3.79453868e-01
-2.58270115e-01 -1.45625025e-01 -7.23891258e-01 -1.48926288e-01
4.93409067e-01 -5.85213676e-02 -7.58248508e-01 -8.67755115e-01
-7.53838718e-01 3.81121933e-01 4.10763621e-01 4.33897436e-01
6.84528053e-01 -1.16386747e+00 -4.71672684e-01 -1.70720294e-01
4.68371689e-01 -5.54278672e-01 4.55540866e-02 -3.42472941e-02
-1.16797306e-01 9.81206417e-01 2.81058371e-01 1.58910334e-01
-1.07738602e+00 7.47029603e-01 -4.18744124e-02 -6.45653725e-01
2.02232689e-01 5.74576557e-01 1.68451935e-01 -3.33940089e-01
5.54367027e-04 1.80855080e-01 -4.27253693e-01 -2.39629835e-01
6.34108603e-01 4.30724651e-01 -4.40338738e-02 4.54928428e-01
-2.45898023e-01 -5.22956371e-01 -4.83403534e-01 -7.34705329e-01
8.89466226e-01 -1.01048678e-01 -5.22810929e-02 1.07148640e-01
7.44187295e-01 2.97148705e-01 -5.89518011e-01 -1.49343073e-01
5.84159791e-01 -9.74485397e-01 -3.67944837e-01 -1.29300904e+00
-3.77440080e-02 2.34219238e-01 1.83123648e-01 6.35691702e-01
7.06723571e-01 2.26783618e-01 -1.04817465e-01 8.26369703e-01
1.19823150e-01 -7.01044917e-01 1.10100592e-02 2.09691241e-01
1.45427394e+00 -1.04902124e+00 5.01082301e-01 -2.08133474e-01
-5.72039127e-01 7.87388921e-01 3.87489140e-01 4.12690610e-01
2.77859777e-01 -3.54741693e-01 2.24241033e-01 -4.80298549e-01
-1.41309011e+00 1.20195128e-01 6.42100215e-01 7.40633667e-01
6.66849971e-01 1.54520452e-01 -5.07312536e-01 4.28841412e-01
-7.71675885e-01 -2.53003120e-01 8.89755070e-01 4.97583926e-01
-2.70451039e-01 -1.06504107e+00 -5.99440098e-01 8.35608840e-01
-4.74623322e-01 -4.57644194e-01 -1.03594196e+00 8.84127557e-01
-5.10826528e-01 1.24288559e+00 -3.94514680e-01 -4.43863273e-01
5.83765686e-01 1.35545343e-01 1.51979495e-02 -8.24305713e-01
-2.69270808e-01 -5.54324508e-01 3.58713299e-01 -5.52915990e-01
1.30822852e-01 -5.20661533e-01 -9.08924997e-01 -3.73201907e-01
-4.96687621e-01 9.91020024e-01 7.47292101e-01 9.26621556e-01
5.37453294e-01 2.81161040e-01 5.56778371e-01 -4.76540804e-01
-9.77656543e-01 -1.13788450e+00 -5.39706945e-01 7.56756902e-01
1.81451693e-01 -5.62734187e-01 -7.37469137e-01 -2.81281471e-01] | [8.784059524536133, 5.668107032775879] |
ffbcd780-926e-4e7e-aa3d-b4842fb0ca07 | classification-of-eye-state-using-eeg | 2209.01023 | null | https://arxiv.org/abs/2209.01023v1 | https://arxiv.org/pdf/2209.01023v1.pdf | Classification of eye-state using EEG recordings: speed-up gains using signal epochs and mutual information measure | The classification of electroencephalography (EEG) signals is useful in a wide range of applications such as seizure detection/prediction, motor imagery classification, emotion classification and drug effects diagnosis, amongst others. With the large number of EEG channels acquired, it has become vital that efficient data-reduction methods are developed, with varying importance from one application to another. It is also important that online classification is achieved during EEG recording for many applications, to monitor changes as they happen. In this paper we introduce a method based on Mutual Information (MI), for channel selection. Obtained results show that whilst there is a penalty on classification accuracy scores, promising speed-up gains can be achieved using MI techniques. Using MI with signal epochs (3secs) containing signal transitions enhances these speed-up gains. This work is exploratory and we suggest further research to be carried out for validation and development. Benefits to improving classification speed include improving application in clinical or educational settings. | ['Hisham Ihshaish', 'Phoebe M Asquith'] | 2022-08-31 | null | null | null | null | ['emotion-classification', 'emotion-classification'] | ['computer-vision', 'natural-language-processing'] | [ 6.06552780e-01 -2.95753926e-01 1.24012515e-01 -6.25567675e-01
-5.34077883e-01 -3.55124593e-01 3.03243756e-01 6.23571455e-01
-5.61048508e-01 9.14355516e-01 9.29515064e-02 -2.39171475e-01
-6.25856340e-01 -2.86984682e-01 3.63679938e-02 -6.07007265e-01
-7.83018172e-01 2.43934318e-02 -1.21320914e-02 1.93758786e-01
7.30314910e-01 6.86103225e-01 -1.75319970e+00 3.59983027e-01
7.37165213e-01 1.02575195e+00 3.06886077e-01 6.17891669e-01
2.35421270e-01 6.05057120e-01 -7.55070210e-01 -3.31573095e-03
2.25330442e-02 -6.24299347e-01 -7.23453701e-01 -1.58975832e-02
-5.82478464e-01 1.54470697e-01 9.65621918e-02 8.82031500e-01
7.98634887e-01 2.71084040e-01 5.29895782e-01 -1.15356767e+00
3.44119847e-01 2.17483982e-01 -5.12222290e-01 6.26373529e-01
7.02607155e-01 -2.16319233e-01 5.96251667e-01 -5.12522697e-01
1.71585381e-01 4.81600255e-01 4.11817074e-01 3.45608383e-01
-1.36526859e+00 -1.12250125e+00 -1.66959450e-01 6.58067465e-01
-1.34287107e+00 -4.53826100e-01 6.39293551e-01 -4.27725732e-01
1.24971581e+00 6.31766379e-01 8.57419610e-01 7.40238905e-01
5.25276721e-01 5.32993197e-01 1.56789660e+00 -4.03018624e-01
4.11225736e-01 3.05420607e-01 5.25299497e-02 -4.83755358e-02
-6.48525283e-02 -1.98882353e-02 -7.83416927e-01 -2.27127030e-01
3.87811154e-01 -1.01208150e-01 -4.55476731e-01 -9.87426788e-02
-1.00854611e+00 6.44064486e-01 -7.76385097e-03 6.60976052e-01
-6.47887707e-01 -2.95315802e-01 6.31018698e-01 8.62058938e-01
5.61951816e-01 9.00433302e-01 -6.33113146e-01 -9.72192466e-01
-1.24893153e+00 1.49037182e-01 8.91741216e-01 7.08157539e-01
3.47372591e-01 -2.65573233e-01 2.81374812e-01 8.79020095e-01
-3.99058163e-02 -2.05797210e-01 8.49208891e-01 -6.06877923e-01
1.73677295e-01 7.15399683e-01 -1.79068357e-01 -8.97277832e-01
-8.61837506e-01 -3.08110416e-01 -7.75469542e-01 1.20202981e-01
-1.34645894e-01 -4.76449244e-02 -5.50755441e-01 1.39317620e+00
-9.98924151e-02 1.98839530e-01 4.33776863e-02 6.46427810e-01
4.14786428e-01 2.79365093e-01 1.01395525e-01 -7.13569283e-01
1.37762439e+00 1.06040195e-01 -7.22971022e-01 -1.58718079e-01
7.34325945e-01 -9.36270416e-01 6.35623872e-01 1.06151783e+00
-9.36627150e-01 -1.00263953e-02 -9.93160307e-01 6.25627041e-01
-2.81219602e-01 -3.21406305e-01 7.70609736e-01 7.76660681e-01
-9.32985008e-01 7.63676345e-01 -1.12065995e+00 -5.63148320e-01
4.11471993e-01 1.06775498e+00 -5.55213928e-01 2.96547681e-01
-1.14900553e+00 1.10079050e+00 3.72260302e-01 -2.32439131e-01
-9.55734029e-02 -4.78626639e-01 -6.79478407e-01 -1.53576598e-01
-3.01413119e-01 3.43304463e-02 9.65828001e-01 -1.01580822e+00
-1.28256142e+00 4.88689870e-01 -1.41851097e-01 -5.55052578e-01
2.10795820e-01 1.58247769e-01 -6.35546625e-01 3.21100391e-02
-3.11163604e-01 5.23530245e-01 5.14290631e-01 -6.32042587e-01
-8.55715930e-01 -7.03525245e-01 -3.16150576e-01 3.07276666e-01
-3.88006151e-01 5.48489153e-01 1.15007691e-01 -5.10581195e-01
8.54040384e-02 -1.00687146e+00 -4.27331142e-02 -7.82786012e-01
3.31436880e-02 -1.16525836e-01 7.08634794e-01 -7.05506623e-01
1.40491295e+00 -1.96509302e+00 1.18053705e-01 6.28219187e-01
-4.32317853e-02 1.62843466e-01 2.48241991e-01 5.85568249e-01
-4.12544727e-01 -2.16576964e-01 -3.51792514e-01 1.30106196e-01
-4.21353519e-01 -7.57758915e-02 3.68797337e-04 5.30690908e-01
2.06043541e-01 3.66642684e-01 -6.86854661e-01 -3.87866907e-02
4.24245834e-01 5.95982373e-01 -2.77034312e-01 1.11583620e-01
5.21580875e-01 8.12465966e-01 -1.26582295e-01 3.20160359e-01
2.04169706e-01 6.57088310e-02 -4.38335016e-02 1.05769997e-02
-1.73884034e-01 5.18667102e-01 -1.14730787e+00 1.54070747e+00
-5.19706309e-01 1.24947667e+00 -2.52731055e-01 -1.22391045e+00
8.49301636e-01 7.05697358e-01 8.39272857e-01 -7.87774861e-01
2.37981111e-01 1.30546704e-01 6.17226720e-01 -6.80345535e-01
2.98706412e-01 1.87456999e-02 2.38424316e-01 4.65546608e-01
-1.84500173e-01 -2.78771538e-02 2.84704804e-01 -5.55074811e-02
1.25054383e+00 -1.65865555e-01 6.32529676e-01 -4.27109897e-01
3.39230120e-01 -1.82735071e-01 2.12769315e-01 6.03092052e-02
-2.13450491e-01 3.21579158e-01 3.20227712e-01 -4.91181351e-02
-5.81188202e-01 -3.58344942e-01 -6.92861676e-01 6.53398514e-01
-2.40077198e-01 -4.92027670e-01 -7.36930788e-01 -1.80948287e-01
-3.36970627e-01 7.40052760e-01 -3.90005141e-01 -3.61902595e-01
-1.69128731e-01 -1.05725324e+00 2.34966412e-01 5.21914899e-01
1.26741543e-01 -1.28088439e+00 -1.30787838e+00 6.63415492e-01
5.77780046e-02 -8.19317579e-01 1.45620674e-01 7.26439238e-01
-1.27800667e+00 -8.83953869e-01 -5.03562033e-01 -5.17868936e-01
5.46096265e-01 7.11731538e-02 6.37571335e-01 -1.49225578e-01
-5.83370030e-01 6.61312699e-01 -6.26920462e-01 -7.61671841e-01
-1.04819760e-01 -9.97474417e-02 2.97363728e-01 -2.61498958e-01
8.33936572e-01 -8.89943361e-01 -7.97510862e-01 2.30638579e-01
-9.08209980e-01 -1.42800063e-03 5.38289428e-01 6.61206841e-01
3.11259896e-01 5.60050189e-01 8.98797154e-01 -6.55541122e-01
1.05691624e+00 -4.72896010e-01 -2.59640723e-01 -1.26254454e-01
-9.29061174e-01 -1.38055548e-01 4.49173927e-01 -4.23610747e-01
-5.74617565e-01 -6.27818480e-02 -1.89670622e-01 3.51992585e-02
-5.49226344e-01 6.68674111e-01 7.99624100e-02 -4.36882615e-01
7.19867408e-01 1.95505410e-01 4.58569545e-03 -1.99183226e-01
-4.75190222e-01 1.17557108e+00 9.87903327e-02 1.10760577e-01
-2.07071509e-02 4.00419869e-02 2.83365995e-02 -1.18919063e+00
-1.26512302e-02 -7.50629842e-01 -5.01672387e-01 -1.75321385e-01
5.58763683e-01 -7.05857933e-01 -6.01682186e-01 2.07245052e-01
-7.90847242e-01 -1.81911007e-01 2.96372026e-01 9.41554844e-01
-4.61161822e-01 5.23657687e-02 -2.25880444e-01 -1.03837526e+00
-4.70471531e-01 -1.20879722e+00 6.38812780e-01 2.90630639e-01
-9.35170412e-01 -9.37349141e-01 -6.73617516e-03 1.59548119e-01
4.13922220e-01 -4.00834903e-02 6.18783057e-01 -8.67577255e-01
-9.24544409e-02 -5.10532618e-01 1.16622448e-01 2.82127976e-01
5.81724823e-01 -4.24857974e-01 -1.09502482e+00 -5.27997613e-01
8.05293694e-02 -1.14948973e-01 3.22821081e-01 4.18255240e-01
1.42021048e+00 8.31672922e-02 -2.55119205e-01 2.63203204e-01
1.06879079e+00 1.04380131e+00 8.66525114e-01 3.84330392e-01
9.44256634e-02 7.89615154e-01 6.37146652e-01 6.41850531e-01
1.73827186e-01 4.49125588e-01 -1.85726508e-02 -7.18463073e-03
3.70173991e-01 6.67604387e-01 1.75396129e-01 9.73812699e-01
-2.47041196e-01 1.13263331e-01 -8.68965209e-01 4.14327323e-01
-1.45376182e+00 -9.44918752e-01 -8.54792073e-02 2.69065070e+00
6.92078769e-01 1.94996163e-01 1.97465152e-01 8.93094540e-01
3.01463813e-01 -5.86141050e-01 -5.73890865e-01 -6.25913024e-01
3.61897022e-01 6.43480420e-01 3.52739334e-01 2.14318439e-01
-8.98674786e-01 2.69257724e-01 5.78243494e+00 4.33981538e-01
-1.45863414e+00 -1.12111099e-01 6.23181939e-01 -2.20982373e-01
3.01446289e-01 -1.43184736e-01 -2.75069982e-01 7.06481040e-01
1.32378483e+00 -1.57411948e-01 6.01234734e-01 2.76659697e-01
6.64641082e-01 -7.92326450e-01 -1.11418104e+00 1.41835415e+00
-5.31160412e-03 -6.72964275e-01 -7.58547962e-01 1.79972976e-01
3.40651035e-01 -2.28120089e-02 -3.26338589e-01 -6.74776509e-02
-4.82153594e-01 -9.68991399e-01 8.13590288e-02 2.64581680e-01
6.21942401e-01 -1.27155817e+00 8.22615862e-01 3.07055682e-01
-1.05935538e+00 6.51566982e-02 1.19559161e-01 -4.74234372e-01
-1.97658196e-01 5.19610882e-01 -9.36642289e-01 4.47037131e-01
7.18489707e-01 6.17754638e-01 -4.94477034e-01 1.56961656e+00
2.03986522e-02 8.47471535e-01 -2.10801452e-01 -2.72464812e-01
-7.92496279e-02 -1.99895754e-01 5.35252273e-01 1.31447792e+00
4.19500291e-01 3.05109233e-01 -2.79564440e-01 1.25454545e-01
4.22747940e-01 2.96475053e-01 -7.19544888e-01 -8.08624998e-02
5.25420487e-01 1.29734075e+00 -1.24903452e+00 -9.42375809e-02
-4.24294382e-01 1.11597347e+00 -1.16078928e-01 1.89748891e-02
-3.28163475e-01 -8.45467985e-01 6.19998693e-01 -1.07077666e-01
-1.81858763e-01 -1.28943324e-01 -4.02138770e-01 -7.95426607e-01
-2.26860903e-02 -1.05324268e+00 3.23739529e-01 -4.74022031e-01
-8.77442062e-01 6.75544441e-01 1.79174200e-01 -1.45353246e+00
-5.99639058e-01 -4.13157642e-01 -4.24214780e-01 6.79660678e-01
-1.07909906e+00 -3.12113732e-01 -3.91200483e-02 4.24013734e-01
5.31842947e-01 1.23634458e-01 1.18206632e+00 4.31568414e-01
-4.30147856e-01 3.78209621e-01 4.59766537e-02 -2.69958496e-01
6.77797496e-01 -1.31252217e+00 -1.46201864e-01 6.18860781e-01
3.47015858e-01 6.14727259e-01 9.32490826e-01 -4.33549583e-01
-1.28312147e+00 -7.90119290e-01 8.44686925e-01 -6.98600709e-02
5.11320651e-01 -2.44494289e-01 -9.98527408e-01 2.38267347e-01
3.85712355e-01 -6.02763057e-01 1.21884227e+00 1.35268748e-01
4.80492145e-01 -1.35803819e-01 -1.10715568e+00 3.85225773e-01
5.76776743e-01 -4.85462219e-01 -4.83242333e-01 2.69892037e-01
-3.56084138e-01 -2.70088643e-01 -1.23832917e+00 3.59834909e-01
6.99536860e-01 -9.36816812e-01 4.49922442e-01 -2.47654527e-01
-6.01950847e-02 -3.00239734e-02 3.05855155e-01 -1.72990191e+00
-8.43416601e-02 -7.49014735e-01 6.69397488e-02 9.20592964e-01
4.93792862e-01 -8.16613019e-01 7.41567671e-01 1.06280744e+00
1.39836103e-01 -9.75021780e-01 -9.00382102e-01 -6.04792953e-01
-4.33866888e-01 -8.37808967e-01 4.56445813e-01 9.71225917e-01
8.20288956e-01 4.41226184e-01 -2.85844982e-01 -1.24557473e-01
2.53416717e-01 -2.89519839e-02 2.78860420e-01 -1.34821510e+00
-1.31571680e-01 -5.15171409e-01 -1.19950294e+00 -3.15253079e-01
-2.58224577e-01 -7.49723077e-01 -1.79622725e-01 -1.32544374e+00
8.35250467e-02 -3.12768936e-01 -8.12344193e-01 4.92105693e-01
-1.44141719e-01 4.81904954e-01 -1.03427969e-01 1.14490189e-01
-2.95856982e-01 1.19543537e-01 5.53998530e-01 2.06383228e-01
-7.72831321e-01 3.25263321e-01 -5.40292025e-01 4.96961027e-01
1.23264694e+00 -5.57394922e-01 -6.88026786e-01 5.10136969e-02
1.47971615e-01 7.92363733e-02 -1.01793401e-01 -1.28563845e+00
3.39455366e-01 2.69700140e-01 6.32763386e-01 -3.96110773e-01
3.98266494e-01 -8.23834062e-01 3.63537520e-01 4.64600801e-01
-3.72830570e-01 3.87657672e-01 4.27311212e-01 2.17219725e-01
-2.99417704e-01 -2.68877447e-01 6.91420496e-01 -9.76906903e-03
-7.40171850e-01 -3.93049046e-02 -9.24390376e-01 -4.36246276e-01
1.24676108e+00 -5.25090694e-01 6.26577973e-01 -4.43045646e-01
-7.48628438e-01 6.08968688e-03 2.89320111e-01 3.23903114e-01
7.42678761e-01 -9.79178131e-01 -3.26745838e-01 5.01589954e-01
1.94605440e-01 -4.51249927e-01 1.01632304e-01 1.47268927e+00
-1.39814004e-01 5.05346060e-01 -3.16710830e-01 -5.10947347e-01
-1.71267557e+00 -9.96543840e-02 -1.57178760e-01 2.01795204e-03
-5.62809348e-01 7.87419736e-01 -3.13137561e-01 2.32509017e-01
3.50279629e-01 -1.98650435e-01 -5.60992122e-01 3.59374642e-01
9.07309651e-01 5.38305938e-01 8.57150257e-01 -4.65026617e-01
-5.29781640e-01 6.01974651e-02 -1.79408267e-01 -2.60735720e-01
1.36687982e+00 3.49961445e-02 -1.46447748e-01 6.21280789e-01
1.11715078e+00 -4.20255065e-01 -8.55164528e-01 4.58997965e-01
4.90805328e-01 -5.04586577e-01 5.23383856e-01 -9.27199125e-01
-7.79064894e-01 8.60101342e-01 1.17938077e+00 4.43436980e-01
1.74707162e+00 -3.31979454e-01 4.00264710e-01 2.89517611e-01
8.60138714e-01 -1.18487322e+00 -4.74967986e-01 1.63668841e-01
5.27293503e-01 -1.11344278e+00 1.64596245e-01 1.77889764e-01
-7.34858871e-01 1.04964793e+00 3.84053439e-02 1.34853907e-02
6.10272467e-01 5.43418109e-01 -9.24852416e-02 -2.32614547e-01
-8.48058760e-01 -3.89944278e-02 3.16098720e-01 5.35648286e-01
9.96570110e-01 3.07140380e-01 -7.45372832e-01 2.70060062e-01
-2.05537528e-01 1.12180024e-01 4.25946772e-01 1.31976807e+00
-3.14022034e-01 -1.40126622e+00 -3.61231536e-01 1.28733015e+00
-6.63018823e-01 -1.76266953e-01 -4.13798779e-01 7.11110592e-01
-9.11450610e-02 1.31186163e+00 1.35964870e-01 -5.75485528e-01
1.62605345e-01 2.62035221e-01 7.93927431e-01 -5.23427248e-01
-6.66484118e-01 1.28885940e-01 1.46792486e-01 -5.63115478e-01
-7.00962245e-01 -9.96137023e-01 -1.20073712e+00 1.28113568e-01
-6.85211480e-01 3.65787685e-01 1.06536782e+00 1.02929306e+00
5.98295748e-01 5.21139801e-01 6.75112188e-01 -7.54046142e-01
5.63044026e-02 -1.14522398e+00 -7.44320512e-01 1.14293940e-01
1.58102721e-01 -7.39516497e-01 -1.86480835e-01 1.13839671e-01] | [13.211723327636719, 3.4330949783325195] |
ba3b72d7-e3c4-4172-bd92-eea386ca2c76 | robust-object-detection-in-remote-sensing | 2210.12989 | null | https://arxiv.org/abs/2210.12989v1 | https://arxiv.org/pdf/2210.12989v1.pdf | Robust Object Detection in Remote Sensing Imagery with Noisy and Sparse Geo-Annotations (Full Version) | Recently, the availability of remote sensing imagery from aerial vehicles and satellites constantly improved. For an automated interpretation of such data, deep-learning-based object detectors achieve state-of-the-art performance. However, established object detectors require complete, precise, and correct bounding box annotations for training. In order to create the necessary training annotations for object detectors, imagery can be georeferenced and combined with data from other sources, such as points of interest localized by GPS sensors. Unfortunately, this combination often leads to poor object localization and missing annotations. Therefore, training object detectors with such data often results in insufficient detection performance. In this paper, we present a novel approach for training object detectors with extremely noisy and incomplete annotations. Our method is based on a teacher-student learning framework and a correction module accounting for imprecise and missing annotations. Thus, our method is easy to use and can be combined with arbitrary object detectors. We demonstrate that our approach improves standard detectors by 37.1\% $AP_{50}$ on a noisy real-world remote-sensing dataset. Furthermore, our method achieves great performance gains on two datasets with synthetic noise. Code is available at \url{https://github.com/mxbh/robust_object_detection}. | ['Matthias Schubert', 'Maximilian Bernhard'] | 2022-10-24 | null | null | null | null | ['robust-object-detection'] | ['computer-vision'] | [ 1.87053606e-01 -2.25141793e-01 -1.10845484e-01 -4.09647375e-01
-1.16974902e+00 -5.87203622e-01 2.61170149e-01 2.97430784e-01
-4.96059120e-01 6.03058457e-01 -4.91633892e-01 -3.71965498e-01
-3.13637964e-02 -9.18301821e-01 -7.79839933e-01 -7.08724141e-01
9.10685584e-03 3.25379252e-01 5.44838905e-01 9.53553841e-02
-2.26122037e-01 5.82574427e-01 -1.55765831e+00 -3.39978165e-03
9.09910858e-01 1.06397808e+00 6.05970144e-01 5.23077369e-01
1.68162659e-01 3.67542684e-01 -4.99611020e-01 7.26763392e-03
5.75440168e-01 1.77465249e-02 -6.02070749e-01 1.38757259e-01
5.66537738e-01 -6.85760736e-01 -1.35348171e-01 1.28292918e+00
5.07674456e-01 -4.47875913e-03 4.31089312e-01 -1.04414821e+00
-4.84501660e-01 4.71392214e-01 -8.21247756e-01 3.18463594e-01
-1.14101477e-01 1.49364561e-01 9.38305080e-01 -9.77393448e-01
3.60239632e-02 6.40646875e-01 9.71532464e-01 3.03483009e-01
-1.20869768e+00 -7.88319170e-01 2.45564893e-01 2.05823835e-02
-1.91115952e+00 -3.58310819e-01 4.11073893e-01 -5.09325862e-01
6.25597000e-01 3.28932405e-01 5.39172471e-01 5.97130299e-01
-3.91047478e-01 7.92898536e-01 7.26023376e-01 -2.43946716e-01
1.32650375e-01 1.55772254e-01 -6.39986098e-02 5.62402248e-01
5.19197881e-01 7.27131888e-02 2.10372970e-01 -8.28358382e-02
8.68036330e-01 4.37596798e-01 -3.43701124e-01 -1.79871842e-01
-1.17938375e+00 6.08055055e-01 9.73052382e-01 1.66763857e-01
-5.50129116e-01 2.16257885e-01 -3.42806689e-02 -1.65196300e-01
7.00714350e-01 2.71046489e-01 -5.58625102e-01 3.62834930e-01
-1.11711121e+00 1.41601369e-01 3.49925369e-01 1.00873601e+00
8.72599185e-01 -1.83243975e-02 5.67001849e-02 8.00664485e-01
3.98109764e-01 8.50968301e-01 3.89935747e-02 -6.80588186e-01
3.02140743e-01 7.54987955e-01 4.55651730e-01 -9.96499658e-01
-5.42099118e-01 -6.73293471e-01 -9.78355408e-01 3.41047019e-01
4.49319988e-01 -1.94819152e-01 -1.03140092e+00 1.22086179e+00
4.54669297e-01 3.67525011e-01 -1.76182181e-01 1.17674971e+00
8.54751110e-01 7.08023310e-01 2.12875843e-01 2.29697913e-01
1.11082029e+00 -4.84982520e-01 -3.58185649e-01 -4.15002376e-01
7.69279122e-01 -4.99864310e-01 8.33573520e-01 2.26685777e-01
-5.01116157e-01 -6.03745341e-01 -9.01932716e-01 2.01883018e-01
-3.27804565e-01 7.42969453e-01 4.93892789e-01 3.97727370e-01
-7.31674552e-01 5.32660365e-01 -1.10484314e+00 -3.47402126e-01
7.22284555e-01 3.16311717e-01 -1.61501974e-01 -6.78761825e-02
-7.63910174e-01 7.09500611e-01 7.29674459e-01 4.02114689e-01
-9.65347171e-01 -5.30387580e-01 -7.02100217e-01 -4.09997534e-03
5.88952899e-01 -2.62102813e-01 1.31288755e+00 -9.28903580e-01
-8.38636458e-01 7.23019838e-01 2.74209529e-01 -4.97955918e-01
4.20690984e-01 -3.67144912e-01 -3.14747840e-01 -3.32245603e-02
2.25267068e-01 6.47363663e-01 8.07415545e-01 -1.24234509e+00
-1.01067173e+00 -4.28822726e-01 4.29761149e-02 -2.32858248e-02
-1.24689572e-01 8.26844946e-02 -1.54397815e-01 -4.62064177e-01
4.72333491e-01 -8.34567606e-01 -3.58176619e-01 4.17047411e-01
-2.22152695e-01 4.40693796e-02 8.05999160e-01 -6.87375963e-01
9.97240841e-01 -2.28514218e+00 -4.70903784e-01 6.62115663e-02
1.91430286e-01 6.82639897e-01 4.95060161e-02 7.44783208e-02
3.57204974e-02 2.45925114e-01 -4.08244967e-01 7.87559838e-04
-1.88532710e-01 2.53984481e-01 -3.29261124e-01 6.83850110e-01
3.11510652e-01 6.32415414e-01 -1.04371452e+00 -4.03961122e-01
4.40454274e-01 6.10764980e-01 -1.97491407e-01 -4.29997267e-03
-2.51531333e-01 5.49195170e-01 -5.78115463e-01 1.01362705e+00
7.49532521e-01 -4.35450792e-01 -3.18436548e-02 -1.16140999e-01
-2.30458319e-01 4.96996008e-03 -1.55443215e+00 1.28437853e+00
-1.48666084e-01 6.34989202e-01 1.58853412e-01 -1.02795386e+00
9.83243406e-01 3.02026808e-01 3.79682600e-01 -3.52627397e-01
1.59993246e-01 3.93436730e-01 -2.32195660e-01 -2.61925966e-01
4.78440851e-01 1.15051061e-01 1.35900989e-01 -4.92290482e-02
-1.24533519e-01 -1.87051862e-01 -1.62473135e-02 -1.30733192e-01
9.09915626e-01 2.52460122e-01 5.85868478e-01 -1.90937109e-02
3.17247331e-01 2.10942641e-01 7.34468520e-01 9.31656897e-01
-1.67168006e-01 6.66177571e-01 -6.76712319e-02 -7.29625165e-01
-9.40424263e-01 -8.40290904e-01 -4.60411251e-01 1.05753803e+00
2.28437170e-01 -9.92542133e-02 -5.69869757e-01 -4.96557862e-01
1.23397417e-01 4.61558878e-01 -3.61665994e-01 1.89921245e-01
-1.57310054e-01 -9.43798244e-01 6.30455613e-01 7.78937101e-01
6.41027451e-01 -7.70769179e-01 -7.01868117e-01 2.69754738e-01
-1.59797311e-01 -1.04160392e+00 8.56166407e-02 1.22893177e-01
-1.01733744e+00 -1.16633987e+00 -5.69808602e-01 -4.15071785e-01
8.22892189e-01 7.99743295e-01 8.43495309e-01 3.63656253e-01
-4.33946222e-01 3.58356088e-02 -4.61633712e-01 -6.93094432e-01
-1.38907120e-01 2.91249417e-02 1.22345230e-02 -1.06399834e-01
4.62514699e-01 -4.16082829e-01 -6.60539806e-01 5.86440682e-01
-1.06413198e+00 -1.09488472e-01 7.28030622e-01 8.40341568e-01
8.44607592e-01 1.50201783e-01 2.80632436e-01 -4.95601058e-01
-1.93672270e-01 -5.04425168e-01 -1.33362877e+00 1.21902026e-01
-2.68309772e-01 -3.11455786e-01 3.32919538e-01 -2.31105134e-01
-7.44240761e-01 6.88940525e-01 -1.29886821e-01 -3.07420522e-01
-6.09585762e-01 5.55577099e-01 -1.32842988e-01 -1.81139931e-01
1.13641620e+00 1.03931077e-01 -4.39381987e-01 -7.00914979e-01
1.69824272e-01 1.04756260e+00 6.07084811e-01 -2.75653511e-01
9.70421016e-01 7.41502404e-01 -2.60197520e-01 -9.59161520e-01
-1.12637842e+00 -8.19130659e-01 -8.72106433e-01 1.78908966e-02
5.06097078e-01 -1.43018758e+00 -3.93418819e-01 4.81299013e-01
-1.05390906e+00 -3.56557101e-01 -2.33361080e-01 6.53521836e-01
-1.84502140e-01 1.66089475e-01 -1.29493490e-01 -1.04141963e+00
-3.17914575e-01 -9.14783657e-01 1.27949297e+00 2.75804609e-01
2.19519705e-01 -6.08417928e-01 -2.42527485e-01 1.96886003e-01
2.64868349e-01 2.99888939e-01 -1.09386444e-02 -6.46798491e-01
-9.32245374e-01 -6.14627182e-01 -5.98434210e-01 4.92280245e-01
1.91423953e-01 9.69969109e-02 -1.09379983e+00 -2.64669746e-01
-3.64805847e-01 -1.88010409e-01 1.11709034e+00 4.62393999e-01
1.19496787e+00 -2.09507927e-01 -6.21755123e-01 6.60969436e-01
1.61066079e+00 5.94139881e-02 4.44863886e-01 4.23888177e-01
7.85310388e-01 1.75093025e-01 6.89910293e-01 7.24938273e-01
3.29155684e-01 5.98026812e-01 8.30658853e-01 -2.45371535e-01
3.01766545e-01 -8.44240114e-02 -2.44891811e-02 -1.12055860e-01
-3.66121531e-01 -2.23303929e-01 -1.29486573e+00 8.27167809e-01
-1.97117198e+00 -1.10172212e+00 -4.19830501e-01 2.35643125e+00
6.66725159e-01 -1.23386942e-01 -6.42799065e-02 3.01729947e-01
7.89495289e-01 -2.13477314e-02 -4.66723680e-01 5.91013432e-01
-4.86457311e-02 -1.89971447e-01 1.08984065e+00 3.14860255e-01
-1.76555979e+00 9.50022161e-01 5.39934254e+00 5.68509519e-01
-9.59753096e-01 2.12124646e-01 3.03500354e-01 -1.56877916e-02
1.78200006e-01 -3.87094207e-02 -1.05273271e+00 3.56150478e-01
6.20407224e-01 1.41281933e-01 7.31822103e-02 1.26777661e+00
1.90805733e-01 -3.44052702e-01 -7.08143413e-01 9.87314284e-01
-1.90571398e-01 -1.21492028e+00 -4.11786079e-01 1.61059666e-02
6.91950381e-01 5.84017694e-01 -1.67330816e-01 1.57924846e-01
4.28371191e-01 -7.14797735e-01 6.59718692e-01 2.44874254e-01
7.83187389e-01 -4.84529763e-01 1.00220644e+00 6.88670933e-01
-1.37166774e+00 -2.94387519e-01 -6.79511428e-01 -1.47868142e-01
-3.49407457e-02 8.23185563e-01 -9.17653203e-01 5.04430771e-01
9.27629113e-01 6.92391813e-01 -6.25614941e-01 1.57598805e+00
-4.91392881e-01 6.46726608e-01 -7.17115104e-01 2.03007966e-01
1.79360062e-01 6.07887097e-02 4.46203202e-01 9.87048805e-01
5.27186692e-01 5.61980009e-01 6.34130597e-01 8.37873161e-01
1.13086421e-02 -1.17739752e-01 -7.95685470e-01 2.44172901e-01
5.87235212e-01 1.43994772e+00 -7.88716614e-01 -3.58242899e-01
-3.12517852e-01 5.38306653e-01 1.90536663e-01 1.74945176e-01
-9.54815090e-01 -2.86718220e-01 6.37645662e-01 2.29390487e-01
3.73345762e-01 -4.11003590e-01 -1.89093426e-01 -1.08003485e+00
1.18416302e-01 -3.88353288e-01 3.08209509e-01 -7.28829920e-01
-1.05319762e+00 4.29148018e-01 2.40112189e-02 -1.48121953e+00
9.33222398e-02 -6.18165135e-01 -3.31107259e-01 8.00789356e-01
-1.70506167e+00 -1.23446918e+00 -8.41547191e-01 2.11872086e-01
2.61182427e-01 1.23002611e-01 7.68339515e-01 4.83993530e-01
-5.30599952e-01 1.49135858e-01 3.58001500e-01 6.02490127e-01
4.48806226e-01 -1.12042522e+00 4.62268949e-01 1.07935441e+00
3.66276562e-01 6.48541600e-02 4.53613728e-01 -5.72336972e-01
-1.02627873e+00 -1.63565814e+00 5.83764315e-01 -5.50189137e-01
6.10740066e-01 -1.25966668e-01 -1.12938297e+00 8.73214602e-01
-5.20736933e-01 5.47552347e-01 5.59797406e-01 -1.58507392e-01
-1.94468856e-01 -4.30596888e-01 -1.06112361e+00 1.18505508e-01
8.41086030e-01 -3.67504478e-01 -3.68963063e-01 6.84960127e-01
4.13515836e-01 -6.44811392e-01 -5.77070355e-01 4.94145602e-01
3.34120482e-01 -7.77978122e-01 8.75985742e-01 -2.28770271e-01
5.80986328e-02 -7.29414761e-01 -2.85463005e-01 -1.19132173e+00
-5.07420480e-01 -5.36947586e-02 2.06109419e-01 9.11937892e-01
4.50844407e-01 -5.54937720e-01 5.43382943e-01 4.47782427e-01
-2.28705257e-01 -2.04929173e-01 -7.59690225e-01 -1.04500961e+00
-3.25795025e-01 -5.95989347e-01 5.84074914e-01 9.02050018e-01
-4.35545385e-01 -2.96577122e-02 -2.74492294e-01 9.98785377e-01
7.62261152e-01 7.06199482e-02 7.95951307e-01 -1.62044370e+00
-5.82191683e-02 -1.61476314e-01 -5.80915868e-01 -1.05594647e+00
-1.44273162e-01 -6.55074060e-01 4.21422213e-01 -1.71735060e+00
-1.01755105e-01 -8.06552052e-01 -6.90429583e-02 9.99202549e-01
-2.28964537e-01 6.71805322e-01 6.48450553e-02 4.33997571e-01
-6.26774609e-01 3.99790198e-01 6.50807202e-01 -2.21163064e-01
-2.75611281e-01 2.64554888e-01 -3.36469620e-01 1.01041174e+00
1.20336115e+00 -6.76435232e-01 7.22312331e-02 -8.03858459e-01
2.23955631e-01 -3.14615667e-01 9.89323199e-01 -1.29250312e+00
2.82501072e-01 -9.98212099e-02 5.70406735e-01 -7.98603117e-01
2.44596347e-01 -1.13066339e+00 1.62382737e-01 3.33543450e-01
1.52344331e-01 -3.90532404e-01 3.40016335e-01 6.25930429e-01
-1.21485665e-01 -2.08183825e-01 8.76453936e-01 -3.04898262e-01
-9.39424038e-01 4.19427991e-01 -1.98934644e-01 -2.44726151e-01
1.10535324e+00 -2.46838674e-01 -2.83545136e-01 -4.23594862e-02
-6.81324661e-01 3.14465344e-01 3.84175301e-01 2.18934387e-01
5.17616093e-01 -1.10117543e+00 -9.85521078e-01 1.67033166e-01
4.24993128e-01 5.45788705e-01 1.29854754e-01 8.82094085e-01
-5.47087669e-01 3.12164128e-01 1.88480038e-02 -9.41143274e-01
-1.24509478e+00 3.95207554e-01 4.10540313e-01 2.91192740e-01
-5.92481554e-01 7.34467089e-01 -1.05117355e-02 -4.49923426e-01
2.06699282e-01 -6.08559132e-01 -4.73616421e-02 5.66490702e-02
8.37769747e-01 3.12724382e-01 9.18817818e-02 -6.35056019e-01
-4.60918188e-01 4.07295048e-01 2.16314599e-01 2.21721217e-01
1.40057838e+00 -4.14326079e-02 1.81166515e-01 2.04401165e-01
5.70044637e-01 -4.38007593e-01 -1.39266849e+00 -6.50763750e-01
4.51542623e-02 -7.09923387e-01 2.58732438e-01 -6.17057204e-01
-1.03353834e+00 8.53509903e-01 8.12364161e-01 2.80955106e-01
1.11366034e+00 7.62430578e-02 3.34027171e-01 7.23950684e-01
5.09054124e-01 -9.80363190e-01 -5.00800729e-01 4.29448634e-01
6.56731009e-01 -1.88458645e+00 2.11108714e-01 -3.20235074e-01
-3.32911968e-01 8.75277817e-01 5.05477548e-01 1.29403062e-02
7.17142880e-01 1.24537162e-01 2.02291429e-01 -9.75306183e-02
-2.13956192e-01 -7.00809062e-01 2.88077623e-01 7.43252993e-01
2.00234547e-01 2.98825920e-01 5.11909910e-02 3.33108157e-01
2.09243312e-01 1.72681585e-01 3.06891650e-01 9.67523038e-01
-8.51641178e-01 -6.64394498e-01 -6.97138250e-01 4.12253499e-01
-4.39864904e-01 -2.89482415e-01 -1.53262511e-01 7.41404593e-01
2.45005444e-01 8.54306161e-01 1.83109790e-01 -1.54942259e-01
3.62651348e-01 -1.48078427e-01 5.97113138e-03 -7.78572679e-01
-3.00260872e-01 2.22942680e-02 -1.35380149e-01 -4.23730195e-01
-5.69995821e-01 -4.61603612e-01 -1.26155293e+00 -1.14383593e-01
-6.60264671e-01 -3.25216986e-02 8.78148854e-01 8.66049647e-01
3.11120540e-01 2.73246616e-01 4.08995867e-01 -1.13703454e+00
-7.07328916e-01 -9.63142633e-01 -7.01708615e-01 1.43822907e-02
4.32377398e-01 -5.80518305e-01 -2.54065275e-01 2.29157671e-01] | [9.03646469116211, -0.9182347655296326] |
dfccd936-c6ab-47cb-9faf-1302c68da208 | unsupervised-learning-of-time-varying | null | null | https://openreview.net/forum?id=HkxIIaVKPB | https://openreview.net/pdf?id=HkxIIaVKPB | Unsupervised-Learning of time-varying features | We present an architecture based on the conditional Variational Autoencoder to learn a representation
of transformations in time-sequence data. The model is constructed in a way that allows to identify sub-spaces of features indicating changes between frames without learning features that are constant within a time-sequence. Therefore, the approach disentangles content from transformations. Different model-architectures are applied to affine image-transformations on MNIST as well as a car-racing video-game task.
Results show that the model discovers relevant parameterizations, however, model architecture has a major impact on the feature-space. It turns out, that there is an advantage of only learning features describing change of state between images, over learning the states of the images at each frame. In this case, we do not only achieve higher accuracy but also more interpretable linear features. Our results also uncover the need for model architectures that combine global transformations with convolutional architectures. | ['Oswin Krause', 'Matthias Brix', 'Henrik Høeg'] | 2019-09-25 | null | null | null | null | ['carracing-v0'] | ['playing-games'] | [-1.33638561e-01 -5.65948337e-02 -5.33407591e-02 -3.15712512e-01
-2.46987030e-01 -6.40331686e-01 1.05243611e+00 -1.30153269e-01
-5.77321827e-01 3.37191463e-01 3.90417337e-01 2.38308180e-02
-9.12250131e-02 -7.24627316e-01 -9.91935313e-01 -8.40846002e-01
-1.93315685e-01 4.12131429e-01 3.66424084e-01 -4.84926581e-01
-1.06438957e-01 6.31450117e-01 -1.61815774e+00 4.40962374e-01
1.62218418e-02 7.03112900e-01 2.19260439e-01 9.16016579e-01
1.78198397e-01 1.07863152e+00 -2.99638152e-01 -2.41996780e-01
3.76590222e-01 -5.01471877e-01 -8.29887748e-01 2.29719654e-01
4.02648270e-01 -3.48255217e-01 -7.55821526e-01 9.40088034e-01
-1.52979597e-01 3.18555593e-01 8.58472049e-01 -1.32443011e+00
-6.58928156e-01 3.99042368e-01 8.84445161e-02 4.38368589e-01
1.64796442e-01 3.20302218e-01 1.31231630e+00 -4.59217370e-01
9.00350571e-01 1.41503668e+00 4.59352106e-01 4.82687622e-01
-1.48748314e+00 -1.63540244e-01 2.05484062e-01 6.33204401e-01
-9.95498359e-01 -3.37560266e-01 7.25378454e-01 -8.55795503e-01
1.12923729e+00 2.95098364e-01 9.65740383e-01 1.56225443e+00
6.67713046e-01 6.35349452e-01 8.79681051e-01 -2.97596067e-01
1.28015667e-01 -6.49290755e-02 -2.11583124e-03 6.37421727e-01
-1.46034792e-01 4.34868187e-01 -4.28681076e-01 2.54520327e-01
1.03444922e+00 1.89871609e-01 -1.87671989e-01 -9.16219950e-01
-1.46753299e+00 9.71836150e-01 3.52155536e-01 6.67047024e-01
-5.23123026e-01 5.88960588e-01 4.22877669e-01 5.62102973e-01
2.09494919e-01 4.92175400e-01 -6.24857843e-01 -3.07574838e-01
-6.74448967e-01 2.45784536e-01 6.73133075e-01 5.13529778e-01
1.02354348e+00 2.60888517e-01 -2.76243538e-01 1.73615798e-01
1.24437764e-01 2.87334293e-01 6.53097391e-01 -1.07162070e+00
1.22569412e-01 4.17813420e-01 -5.35869747e-02 -1.08536434e+00
-4.62699622e-01 -2.36120671e-01 -6.42310798e-01 4.41851705e-01
3.68170589e-01 6.55270880e-03 -9.16676044e-01 2.10016322e+00
-4.59386744e-02 4.68179822e-01 1.10780627e-01 7.20329165e-01
2.92690635e-01 6.59777462e-01 -5.51344268e-02 -1.49475485e-02
1.40257955e+00 -6.35914803e-01 -8.39446366e-01 -5.39662801e-02
3.93114775e-01 -3.39337260e-01 9.25601482e-01 1.94600105e-01
-1.08564055e+00 -8.91891718e-01 -1.03180993e+00 -1.86430588e-01
-5.97007036e-01 -6.56829076e-03 4.79967922e-01 2.82918632e-01
-1.21022546e+00 9.59184408e-01 -1.50918019e+00 -6.30116820e-01
7.92986378e-02 4.71130699e-01 -5.99057794e-01 5.19444406e-01
-1.04892790e+00 1.34468901e+00 5.30740440e-01 -1.31227285e-01
-1.35904038e+00 -6.89403892e-01 -1.00961614e+00 3.03495944e-01
1.09299995e-01 -6.97589517e-01 1.15192282e+00 -1.37116516e+00
-1.73940265e+00 6.65461600e-01 -5.82936555e-02 -7.50235975e-01
5.69482684e-01 -3.07368431e-02 -2.92719305e-01 1.21525005e-01
-2.84667492e-01 6.39985919e-01 1.13383293e+00 -1.01650012e+00
-5.26207030e-01 -1.90115571e-01 4.64387804e-01 -7.55170546e-03
-3.03819269e-01 -2.22050339e-01 -1.53887600e-01 -3.49506080e-01
-1.32233977e-01 -1.21272802e+00 -2.27080733e-01 -1.76745012e-01
-1.79509312e-01 6.25585765e-02 8.90546203e-01 -6.45537853e-01
8.24257672e-01 -2.08376718e+00 6.86693132e-01 -3.30886692e-02
2.31206238e-01 6.28182516e-02 -3.11201900e-01 4.52301502e-01
-4.74577963e-01 1.95902772e-02 -5.36296330e-02 -2.13262856e-01
1.88559759e-02 6.94765508e-01 -4.15046841e-01 5.83539367e-01
4.54495013e-01 1.18188655e+00 -6.05097532e-01 -1.03623204e-01
6.78201735e-01 7.00602114e-01 -7.66118288e-01 1.79375455e-01
-3.47259402e-01 7.23459721e-01 -3.22314203e-01 -2.00263217e-01
1.91719696e-01 -1.23129889e-01 1.13330700e-01 -3.23584050e-01
-1.52641520e-01 4.44433205e-02 -8.83104742e-01 1.85860932e+00
-4.81679052e-01 9.02809918e-01 -4.58303779e-01 -1.34775031e+00
4.54395503e-01 4.91601825e-01 7.16144443e-01 -6.15746081e-01
4.12306517e-01 -4.30409253e-01 2.19527453e-01 -7.94014275e-01
3.04538995e-01 -9.40216482e-02 1.10605005e-02 2.07181290e-01
6.49057508e-01 4.30602729e-02 2.85615623e-01 -3.88711356e-02
9.73319709e-01 5.10823846e-01 3.05534959e-01 -4.10278678e-01
7.04885662e-01 -1.95532665e-01 3.65809679e-01 4.53885972e-01
-1.12283841e-01 4.14885253e-01 8.23057652e-01 -7.96675980e-01
-1.09104037e+00 -1.06072164e+00 1.22940898e-01 1.04116118e+00
-1.53171673e-01 -4.65716094e-01 -7.42220521e-01 -3.86560500e-01
-2.56474704e-01 8.28150809e-01 -1.17585087e+00 -5.41988552e-01
-6.15189552e-01 -5.16092062e-01 2.71197408e-01 3.38813573e-01
9.23113674e-02 -1.05545318e+00 -1.13499141e+00 8.60923156e-02
-1.95781976e-01 -1.36866271e+00 -2.42216423e-01 3.13556671e-01
-8.87537003e-01 -1.04599333e+00 -2.04152688e-01 -4.89028662e-01
3.47582221e-01 -2.38887772e-01 1.12266016e+00 -4.98357028e-01
-1.99016780e-01 8.57398391e-01 -2.44899109e-01 -2.34759256e-01
-7.89065063e-01 1.47141814e-02 5.37889935e-02 4.31196839e-01
2.39222497e-01 -8.55438411e-01 -3.61482382e-01 -2.64441930e-02
-9.94754195e-01 -6.14571460e-02 4.04232770e-01 7.16421723e-01
4.71930683e-01 -2.45705888e-01 -4.47516777e-02 -6.50876045e-01
2.66207457e-01 -3.44732136e-01 -5.94862938e-01 1.71618491e-01
-3.97685140e-01 5.29692948e-01 6.09501183e-01 -5.90092957e-01
-9.99019861e-01 2.20882311e-01 5.93752190e-02 -7.91878760e-01
-3.66977334e-01 3.00275207e-01 -6.40406157e-04 3.03156823e-01
7.89866626e-01 2.62808084e-01 2.27588639e-01 -3.29758227e-01
5.35519958e-01 -2.12597072e-01 6.13436759e-01 -3.90025735e-01
8.75774980e-01 8.52293730e-01 2.93155968e-01 -7.98724234e-01
-3.63458127e-01 -1.55488208e-01 -1.03054273e+00 -6.69372380e-02
1.37344754e+00 -9.15987611e-01 -7.62623847e-01 1.04412816e-01
-1.10586500e+00 -3.80316973e-01 -6.45436108e-01 6.71733081e-01
-1.00178170e+00 2.67637193e-01 -6.71653569e-01 -3.82127702e-01
3.70969474e-01 -1.29634476e+00 7.82521188e-01 -1.27920825e-02
-2.15144783e-01 -1.24229002e+00 4.88057852e-01 -1.10484138e-01
2.78042197e-01 4.48190987e-01 9.21963990e-01 -6.27625406e-01
-7.14894295e-01 -5.55665419e-02 3.54323536e-01 3.98799479e-01
1.72908649e-01 3.08297127e-01 -1.04754198e+00 -3.35381329e-01
1.76250920e-01 -6.61698589e-03 8.35363746e-01 6.23600602e-01
1.03080451e+00 -3.12372953e-01 1.22175999e-02 6.97806478e-01
1.29427481e+00 2.75672466e-01 7.05426574e-01 3.51466507e-01
7.59162545e-01 5.05462587e-01 6.36413395e-02 2.86950648e-01
3.70011508e-01 8.56024683e-01 7.96485186e-01 3.97459753e-02
-1.38018772e-01 -2.49595821e-01 7.69744873e-01 7.42613316e-01
-3.73995841e-01 -2.12560333e-02 -6.26896679e-01 4.72638249e-01
-1.94313848e+00 -1.11771643e+00 -5.01705669e-02 2.03807878e+00
4.70968813e-01 -2.56397780e-02 1.68200880e-01 -1.58453718e-01
4.18982863e-01 4.25478727e-01 -4.42725360e-01 -4.92952645e-01
-4.51298095e-02 2.56933957e-01 3.11891019e-01 6.53836668e-01
-1.07134867e+00 1.09677541e+00 6.73491001e+00 2.38076299e-01
-1.32350302e+00 2.03965873e-01 2.36083910e-01 -1.96589321e-01
-2.13045448e-01 2.53377289e-01 -2.76844054e-01 2.55215585e-01
1.24584627e+00 -8.23196396e-02 5.61943829e-01 6.33623302e-01
2.27494985e-01 2.97645539e-01 -1.53317046e+00 8.42204452e-01
1.23673618e-01 -1.31153607e+00 2.92072892e-01 1.27576098e-01
5.24756849e-01 1.44127235e-01 2.48765811e-01 3.64260405e-01
4.95872170e-01 -9.37614977e-01 9.72992420e-01 7.79018641e-01
3.74635845e-01 -5.70074379e-01 3.79307866e-01 2.25330219e-01
-8.64438355e-01 -1.12871706e-01 -4.20187980e-01 -2.94088006e-01
-2.06856951e-02 5.38721271e-02 -6.15556419e-01 3.63460600e-01
5.77387452e-01 9.75875795e-01 -7.60968089e-01 4.07512516e-01
-2.46335551e-01 4.40383434e-01 1.94109827e-02 2.34057188e-01
2.69878745e-01 -4.12480354e-01 5.75313091e-01 1.26302385e+00
2.04807729e-01 -2.07586706e-01 3.50553095e-02 9.33727384e-01
3.71400446e-01 -2.37582296e-01 -8.10041904e-01 3.02694850e-02
-2.95655757e-01 1.17460752e+00 -5.29460430e-01 -1.86461225e-01
-5.24357915e-01 1.01794517e+00 3.26331913e-01 6.09400868e-01
-9.96774673e-01 1.38492227e-01 1.08456147e+00 -9.41108018e-02
6.46078348e-01 -5.10489821e-01 3.30677062e-01 -1.47224283e+00
-2.60127634e-01 -7.00244725e-01 2.22790033e-01 -7.69582272e-01
-8.17736208e-01 6.22803569e-01 3.37658972e-01 -1.16065967e+00
-9.50837851e-01 -7.89332151e-01 -5.19926429e-01 5.10086060e-01
-1.24905562e+00 -1.18569839e+00 -4.15137261e-02 9.18965876e-01
4.96250302e-01 -1.93482131e-01 8.74155700e-01 -9.95049477e-02
-4.74962413e-01 2.76544124e-01 1.35988533e-01 2.17363372e-01
2.24316567e-01 -1.27609956e+00 4.54391271e-01 1.11426246e+00
8.79768431e-01 5.40005088e-01 1.00398362e+00 -7.78398290e-02
-1.34364128e+00 -8.60533237e-01 5.98010242e-01 -8.55507970e-01
8.60256314e-01 -5.14449120e-01 -8.56178820e-01 1.27625120e+00
4.49623883e-01 2.15441853e-01 4.42561805e-01 -5.98935736e-03
-5.48994780e-01 -1.25806078e-01 -6.27209306e-01 7.18221247e-01
9.34635639e-01 -1.03795397e+00 -6.73803926e-01 5.75751439e-02
6.22099996e-01 -3.55449468e-01 -8.05167198e-01 -7.90273398e-02
5.46242058e-01 -1.29630435e+00 1.11094379e+00 -1.03216863e+00
5.46584547e-01 -1.70976877e-01 -2.37285942e-01 -1.59465420e+00
-7.50096858e-01 -4.39186215e-01 -2.40252480e-01 8.16065311e-01
2.28610978e-01 -5.63012838e-01 5.98271251e-01 5.91144443e-01
4.65306491e-02 -2.53588676e-01 -1.01226079e+00 -7.44690895e-01
2.39661157e-01 -4.56990123e-01 5.04231870e-01 1.00872087e+00
-1.18035816e-01 5.10781646e-01 -3.57474744e-01 1.86771452e-01
2.27951899e-01 -1.31064773e-01 7.02648818e-01 -1.26733565e+00
-4.67657149e-01 -3.71664554e-01 -9.36337411e-01 -6.41263783e-01
7.19458044e-01 -8.63008022e-01 -1.90335616e-01 -1.06241119e+00
2.03467697e-01 4.13452387e-01 -6.18366599e-01 5.25848091e-01
1.38666168e-01 -1.18467122e-01 4.22367245e-01 1.15358561e-01
-3.68261039e-01 6.39584303e-01 1.10532665e+00 -1.10537119e-01
1.03021916e-02 -2.35961705e-01 -2.22930834e-01 5.99080324e-01
7.37200856e-01 -4.31342363e-01 -7.04139352e-01 -5.81015289e-01
1.25006169e-01 2.81474702e-02 7.58947790e-01 -1.10104549e+00
-1.68263970e-03 -2.23453298e-01 4.48240340e-01 -1.96282223e-01
5.10625899e-01 -1.09957170e+00 3.97011250e-01 7.29954660e-01
-4.52660054e-01 3.82993668e-01 1.73096344e-01 7.71677792e-01
-2.91218221e-01 -7.51189375e-03 6.87051773e-01 -2.10936934e-01
-9.15207326e-01 3.86140168e-01 -7.11368620e-01 -2.72167474e-01
9.79946494e-01 -1.45187989e-01 1.89795196e-01 -6.24797285e-01
-9.42094266e-01 -3.05722803e-01 3.91580820e-01 8.23904693e-01
1.22359142e-01 -1.28366363e+00 -6.52515709e-01 4.13427472e-01
1.08385563e-01 -3.57838839e-01 3.00439686e-01 7.42440522e-01
-4.72292215e-01 5.88792682e-01 -8.30348432e-01 -9.22073781e-01
-1.06536961e+00 8.11108470e-01 5.06376147e-01 -3.46923560e-01
-6.39776230e-01 4.63657528e-01 4.67305779e-01 -3.12988937e-01
-2.33512148e-01 -9.34350729e-01 -4.02337790e-01 1.72193348e-01
1.20880507e-01 1.31682888e-01 5.91027103e-02 -1.08316028e+00
-4.37936485e-02 5.51625907e-01 -5.69184870e-02 -3.81356865e-01
1.40919781e+00 -1.02794386e-01 -1.80119311e-03 9.05245006e-01
1.50046694e+00 -5.55743277e-01 -1.77511740e+00 -1.10198811e-01
-1.81266610e-02 -1.75645173e-01 8.39120895e-02 -2.76839435e-01
-1.16478908e+00 9.48249102e-01 8.54917169e-01 1.56229600e-01
9.13468480e-01 -3.14872041e-02 1.74494177e-01 5.56885958e-01
9.17244777e-02 -8.24167907e-01 2.06800044e-01 7.38158345e-01
9.91317689e-01 -1.17131567e+00 -1.89489216e-01 2.72182703e-01
-7.22029567e-01 1.27864611e+00 2.80761808e-01 -5.06553948e-01
6.54171586e-01 -9.92070977e-03 1.94314718e-02 -2.97050387e-01
-9.44891095e-01 -2.98999399e-01 3.22331488e-01 6.33708298e-01
2.65808254e-01 7.59122819e-02 1.14235513e-01 3.03384721e-01
-3.85996163e-01 -2.04993680e-01 6.50016665e-01 5.59940279e-01
-4.92393300e-02 -1.15608311e+00 -1.58998862e-01 8.23600218e-02
-2.86032349e-01 2.72280812e-01 -1.19265877e-01 1.23353100e+00
1.56969339e-01 6.00260973e-01 4.09592450e-01 -4.50901300e-01
3.37294191e-01 3.40112120e-01 6.95718408e-01 -4.06733990e-01
-4.84719098e-01 -1.02971457e-01 -2.76334882e-01 -9.66484427e-01
-6.64273560e-01 -9.58814323e-01 -9.53956723e-01 -6.11687228e-02
4.44195755e-02 -8.78743455e-02 5.98474145e-01 1.02101445e+00
2.95763075e-01 8.61838698e-01 4.51130629e-01 -9.89457250e-01
-4.96249110e-01 -8.49966526e-01 -5.68187475e-01 7.22771704e-01
8.17009807e-01 -6.80971324e-01 -1.53191224e-01 6.21182978e-01] | [8.633235931396484, 0.42483648657798767] |
631524f7-6518-4855-97e2-81031de0c536 | predicting-3d-shapes-masks-and-properties-of | 2109.07577 | null | https://arxiv.org/abs/2109.07577v1 | https://arxiv.org/pdf/2109.07577v1.pdf | Predicting 3D shapes, masks, and properties of materials, liquids, and objects inside transparent containers, using the TransProteus CGI dataset | We present TransProteus, a dataset, and methods for predicting the 3D structure, masks, and properties of materials, liquids, and objects inside transparent vessels from a single image without prior knowledge of the image source and camera parameters. Manipulating materials in transparent containers is essential in many fields and depends heavily on vision. This work supplies a new procedurally generated dataset consisting of 50k images of liquids and solid objects inside transparent containers. The image annotations include 3D models, material properties (color/transparency/roughness...), and segmentation masks for the vessel and its content. The synthetic (CGI) part of the dataset was procedurally generated using 13k different objects, 500 different environments (HDRI), and 1450 material textures (PBR) combined with simulated liquids and procedurally generated vessels. In addition, we supply 104 real-world images of objects inside transparent vessels with depth maps of both the vessel and its content. We propose a camera agnostic method that predicts 3D models from an image as an XYZ map. This allows the trained net to predict the 3D model as a map with XYZ coordinates per pixel without prior knowledge of the image source. To calculate the training loss, we use the distance between pairs of points inside the 3D model instead of the absolute XYZ coordinates. This makes the loss function translation invariant. We use this to predict 3D models of vessels and their content from a single image. Finally, we demonstrate a net that uses a single image to predict the material properties of the vessel content and surface. | ['Alan Aspuru-Guzik', 'Yi Ru Wang', 'Haoping Xu', 'Sagi Eppel'] | 2021-09-15 | null | null | null | null | ['single-view-3d-reconstruction'] | ['computer-vision'] | [ 2.66167104e-01 1.33454889e-01 6.91990733e-01 -1.70892969e-01
-2.60259777e-01 -9.81065333e-01 5.65649271e-01 -1.56344905e-01
-2.67362595e-01 1.29256576e-01 -2.35712424e-01 -4.32363115e-02
3.42922419e-01 -1.07777452e+00 -1.16228783e+00 -6.72014475e-01
1.78252697e-01 4.51418072e-01 5.34574151e-01 1.82969049e-01
4.28425789e-01 9.76697445e-01 -1.34248400e+00 3.51417094e-01
5.82092166e-01 1.36742437e+00 3.30086112e-01 9.15403843e-01
-4.13144857e-01 4.19110179e-01 -2.96219796e-01 -7.52345800e-01
9.46756721e-01 -3.49273439e-04 -4.66149271e-01 4.42739099e-01
8.50101411e-01 -4.10087019e-01 -2.53573596e-01 8.54704440e-01
8.46865550e-02 1.74262691e-02 1.08789158e+00 -8.94225419e-01
-5.82104683e-01 -2.13140398e-01 -5.77393532e-01 -4.29331064e-01
2.66143441e-01 5.13381183e-01 4.92903948e-01 -7.54055977e-01
8.88983786e-01 1.07341850e+00 6.61971688e-01 4.13402379e-01
-1.12605226e+00 -1.97148070e-01 1.14752926e-01 -4.83867019e-01
-1.11495388e+00 -1.55673653e-01 8.75524104e-01 -8.68185043e-01
4.88202214e-01 2.22213507e-01 9.58975136e-01 7.72717118e-01
4.09804225e-01 2.78469205e-01 1.08469248e+00 -4.04789865e-01
3.98120999e-01 6.72427297e-01 -1.87135831e-01 1.02077830e+00
1.95131414e-02 2.06703305e-01 -3.79690886e-01 9.44392569e-03
1.47278237e+00 -9.65415388e-02 -4.04326379e-01 -8.21904600e-01
-1.15227675e+00 4.62746173e-01 3.47207308e-01 -4.35292780e-01
-8.64271671e-02 2.06600785e-01 -3.18153203e-02 3.80228087e-02
4.49671328e-01 6.79464519e-01 -6.10139847e-01 4.27676318e-03
-2.58841485e-01 3.10089216e-02 1.03752124e+00 1.29929292e+00
9.07096803e-01 -8.84947851e-02 2.01876506e-01 5.99858165e-01
2.68817723e-01 7.83915997e-01 7.70642161e-02 -1.42547655e+00
4.53083068e-01 6.66395426e-01 4.46003139e-01 -9.40991998e-01
-2.14745358e-01 1.43798336e-01 -3.16789836e-01 5.96464813e-01
7.46828079e-01 -2.79040098e-01 -1.10722065e+00 9.08784330e-01
6.18922055e-01 8.07415545e-02 2.47674864e-02 8.97966564e-01
1.04305482e+00 8.62163365e-01 -5.23895085e-01 3.51509690e-01
1.08077049e+00 -1.06179869e+00 -1.79460868e-02 -5.55326305e-02
4.96444851e-01 -8.64233375e-01 1.37362552e+00 4.48114932e-01
-1.29212821e+00 -3.03468615e-01 -6.58943653e-01 -1.29387617e-01
-4.86675411e-01 2.26385176e-01 7.09622085e-01 6.03969038e-01
-9.28888559e-01 6.12819374e-01 -5.40830851e-01 -5.69446161e-02
2.86156982e-01 1.15397610e-01 -4.37532961e-01 8.91757235e-02
-5.57887614e-01 7.42366850e-01 -1.04505710e-01 8.51099864e-02
-9.36610103e-01 -1.26704824e+00 -1.05884480e+00 -1.32369041e-01
3.02807927e-01 -6.31955147e-01 9.40219343e-01 -7.69855857e-01
-1.94997740e+00 1.10110426e+00 1.13556735e-01 -1.23056360e-01
1.03302383e+00 -1.48442894e-01 2.76493728e-01 3.76210511e-01
-2.68243372e-01 6.38409436e-01 8.39085460e-01 -1.75551200e+00
-3.33671689e-01 -5.57815731e-02 1.40586659e-01 2.83789843e-01
1.57082826e-01 -2.66358733e-01 -6.22009099e-01 -2.80423820e-01
1.58226490e-01 -7.73773968e-01 -1.82741389e-01 1.04146826e+00
-7.94812381e-01 6.21182978e-01 5.97675860e-01 -7.35505581e-01
1.58744618e-01 -2.21004224e+00 -1.02986105e-01 2.75475621e-01
5.52229248e-02 -2.18309298e-01 -3.04470301e-01 3.89391966e-02
2.21122086e-01 2.33664572e-01 -4.09496218e-01 -4.59224045e-01
-1.11655854e-01 -1.14532761e-01 -2.38505244e-01 4.26074386e-01
2.01320991e-01 6.84813976e-01 -5.10822713e-01 -4.10695761e-01
5.19443810e-01 6.99391544e-01 -4.41686213e-01 2.65000075e-01
-4.67880964e-01 5.34099758e-01 -5.11100650e-01 5.67028463e-01
1.00742102e+00 -2.38313293e-03 -3.75609905e-01 -4.29824352e-01
-3.69368702e-01 -2.46011510e-01 -1.20345581e+00 1.45481837e+00
-8.09279561e-01 6.41270638e-01 2.90498793e-01 -2.10282356e-01
1.22927225e+00 5.27950488e-02 6.12633526e-01 -5.29724300e-01
2.06523135e-01 1.85203433e-01 -5.10826647e-01 -5.58812737e-01
3.98692071e-01 7.38563165e-02 2.12308094e-01 3.33454430e-01
-4.38412547e-01 -1.30321145e+00 3.90366651e-02 -1.22549459e-01
8.02238405e-01 5.06973386e-01 -3.53314161e-01 2.29188725e-02
1.45372242e-01 1.57515883e-01 3.21716696e-01 5.82609475e-01
3.22920829e-01 1.22828054e+00 5.62498271e-01 -9.34249759e-01
-1.43600345e+00 -1.41844749e+00 -2.90813744e-01 2.48568058e-01
6.21233463e-01 7.60007426e-02 -7.11550117e-01 -3.32631111e-01
1.65873885e-01 4.19495791e-01 -7.93093085e-01 1.25661924e-01
-6.14887059e-01 -5.12796223e-01 -1.08111925e-01 2.19464779e-01
7.00466216e-01 -9.28787172e-01 -8.26422870e-01 -1.35860354e-01
1.09414555e-01 -1.43593681e+00 -4.49373215e-01 3.84849235e-02
-7.94270396e-01 -1.38158989e+00 -6.90489471e-01 -6.32329345e-01
8.91347051e-01 1.42040744e-01 1.22860098e+00 -2.58895546e-01
-6.60160601e-01 6.04041278e-01 -1.89100996e-01 -4.12030041e-01
-4.65345591e-01 -4.84888166e-01 -3.92387092e-01 1.80539057e-01
-5.47977805e-01 -3.20132524e-01 -6.21905744e-01 6.29295766e-01
-7.68243670e-01 4.87599313e-01 1.43303514e-01 9.91879925e-02
8.70165348e-01 3.11259553e-02 -4.01044101e-01 -8.28943372e-01
-1.44132644e-01 -9.71562639e-02 -1.12388837e+00 2.94682831e-01
1.89985543e-01 -2.44578123e-01 6.17504835e-01 -6.12252533e-01
-1.25041723e+00 4.48020667e-01 5.20298421e-01 -4.58335996e-01
-1.84876427e-01 -1.49074972e-01 -1.52384043e-01 -2.71555185e-01
6.46836340e-01 7.95960054e-02 -8.16057399e-02 -3.70345294e-01
2.76477575e-01 2.91103452e-01 3.39840442e-01 -7.45817840e-01
9.26709473e-01 9.47936416e-01 1.86808154e-01 -9.59449053e-01
-6.78354681e-01 -1.11939460e-01 -8.10862541e-01 -4.50024873e-01
9.94308889e-01 -7.67877996e-01 -8.85046899e-01 6.43289685e-01
-1.23243225e+00 -1.05773687e+00 -4.27684784e-01 4.80425566e-01
-7.68795013e-01 6.75848201e-02 -8.38176727e-01 -6.88309431e-01
-1.10897906e-01 -1.27131796e+00 1.19604027e+00 4.56769288e-01
2.44382977e-01 -9.47332144e-01 -3.87896478e-01 4.67918187e-01
3.27402532e-01 4.96099293e-01 9.40091610e-01 3.63532424e-01
-1.18976259e+00 -7.05009997e-02 -3.31500500e-01 3.75376940e-01
1.10574424e-01 6.70422554e-01 -1.19377351e+00 3.54589403e-01
2.70480718e-02 -8.70562047e-02 5.91938555e-01 7.19232678e-01
1.36282361e+00 -1.75611243e-01 -2.19499752e-01 1.08594906e+00
1.60474360e+00 3.36789817e-01 6.62253201e-01 2.79373229e-01
9.63810861e-01 9.45973277e-01 3.77512246e-01 4.22194242e-01
1.25520453e-01 6.24976575e-01 6.93942070e-01 -3.47321421e-01
-2.08300278e-01 1.17790764e-02 4.08711761e-01 2.75115788e-01
-2.08011106e-01 -3.63284141e-01 -6.68813348e-01 1.51537091e-01
-1.00314760e+00 -5.51255405e-01 -4.46609646e-01 2.42800117e+00
7.47945905e-01 -5.48758032e-03 -2.48063430e-01 -5.31158328e-01
6.22359872e-01 -3.36205095e-01 -7.49011934e-01 -6.18393958e-01
-2.39085063e-01 1.05899669e-01 8.26140285e-01 6.92977965e-01
-1.05367362e+00 8.54410827e-01 6.07972479e+00 4.08760667e-01
-1.38404751e+00 -4.86351520e-01 9.47913110e-01 -1.39499754e-01
-5.01663804e-01 -1.31286651e-01 -8.14104438e-01 4.55956072e-01
4.56916064e-01 1.17352195e-01 4.71744150e-01 7.94463992e-01
3.48657161e-01 -4.61362541e-01 -1.14272153e+00 1.01519501e+00
-6.31571785e-02 -1.39695036e+00 1.09592497e-01 -1.19718760e-02
9.62040544e-01 -3.06592267e-02 3.19702297e-01 -4.77598757e-01
2.96270758e-01 -8.18184853e-01 1.09373069e+00 8.53717685e-01
9.17534649e-01 -6.06413744e-02 5.04174173e-01 1.27025545e-01
-1.03092456e+00 2.20860913e-01 -4.08027053e-01 3.10188025e-01
1.22471683e-01 6.48412943e-01 -8.24681818e-01 1.82061955e-01
9.01889741e-01 5.28381467e-01 -5.59364319e-01 1.18683231e+00
2.96309441e-02 2.13941440e-01 -6.48726225e-01 1.41504169e-01
-8.58243108e-02 -8.13073218e-01 3.23386729e-01 8.59515846e-01
4.39402878e-01 -1.59701332e-01 -2.41283458e-02 1.37790763e+00
5.58735011e-03 1.34780690e-01 -6.54410183e-01 2.89596289e-01
2.20806137e-01 1.32819068e+00 -1.06057012e+00 -7.34501779e-02
-3.40962619e-01 8.88983667e-01 8.56923461e-02 5.07861912e-01
-8.56409669e-01 -2.56751537e-01 5.03331423e-01 4.35132682e-01
2.91681886e-01 -2.58042097e-01 -4.62010235e-01 -9.71077919e-01
3.62383008e-01 -1.27331600e-01 -3.76286685e-01 -1.44193602e+00
-1.31060851e+00 4.15586561e-01 -1.25369430e-01 -1.40797400e+00
5.57195306e-01 -1.11835492e+00 -6.11484945e-01 1.02692723e+00
-1.56458282e+00 -9.64332640e-01 -7.09235728e-01 1.68084532e-01
3.62393886e-01 2.63883948e-01 7.40476489e-01 -2.31010780e-01
-2.86827743e-01 -5.62495999e-02 1.96185604e-01 2.06705645e-01
5.49896359e-01 -1.15834868e+00 3.46916050e-01 2.73820668e-01
-3.00519049e-01 2.45164841e-01 4.57736731e-01 -4.00255740e-01
-1.28624129e+00 -1.29908800e+00 -7.16776475e-02 -8.07523847e-01
3.16818893e-01 -6.46902859e-01 -6.63150132e-01 4.60590154e-01
-2.29655847e-01 3.56112599e-01 1.59229264e-01 -6.21281862e-01
-2.57491678e-01 -5.49349003e-02 -1.40288734e+00 7.41395593e-01
8.77334177e-01 -3.30848396e-01 1.72185391e-01 5.17717123e-01
8.19918752e-01 -9.11698401e-01 -8.91055226e-01 1.69228345e-01
6.34755015e-01 -1.15476191e+00 1.01175249e+00 -6.76544830e-02
9.22510564e-01 -3.22789490e-01 4.50362312e-03 -1.32580411e+00
2.08292097e-01 -4.91163909e-01 2.69043803e-01 7.75133371e-01
7.30792999e-01 -5.43231070e-01 1.14824915e+00 1.34101129e+00
-4.84024972e-01 -7.61416256e-01 -5.26388645e-01 -5.86048067e-01
1.38278976e-01 -2.95605958e-01 5.44482768e-01 8.01721931e-01
-7.70327091e-01 -3.02245468e-01 8.42916891e-02 3.10056627e-01
5.04029572e-01 5.22766709e-01 9.15313125e-01 -1.24463582e+00
-8.27322379e-02 -3.09855521e-01 -4.36396927e-01 -9.93667006e-01
-4.27320227e-02 -6.28619373e-01 -6.94957152e-02 -1.56073117e+00
1.79075569e-01 -9.07301247e-01 7.25865543e-01 1.07236914e-01
4.08340156e-01 9.19600800e-02 2.14164574e-02 2.20771898e-02
3.15502547e-02 4.55110133e-01 1.96248412e+00 -2.10585013e-01
-5.33582211e-01 -1.31588709e-02 -1.70653373e-01 9.40582573e-01
8.17243516e-01 -1.46935835e-01 -2.06276834e-01 -7.71327198e-01
1.41938642e-01 5.03504798e-02 6.28623009e-01 -9.04846489e-01
-1.41478047e-01 -3.38155210e-01 6.26197577e-01 -3.66152883e-01
8.90773535e-01 -1.09237337e+00 2.09777087e-01 1.28552958e-01
-3.39990735e-01 -3.45820963e-01 2.62227297e-01 4.54674512e-01
3.43073785e-01 -4.03568715e-01 8.78536582e-01 -6.00008488e-01
-4.62881774e-01 4.92657483e-01 -1.13911942e-01 -4.17710841e-02
1.25008070e+00 -7.80891299e-01 -4.66319650e-01 -9.90886316e-02
-8.47885966e-01 -7.23025128e-02 1.17965519e+00 6.27519637e-02
8.57608080e-01 -7.80316114e-01 -4.21899885e-01 3.41185927e-01
-3.17537598e-02 8.23013842e-01 3.25668007e-01 2.32406333e-01
-1.41518092e+00 -1.39060885e-01 -1.01249188e-01 -6.12868845e-01
-1.01667643e+00 3.29120398e-01 9.81145620e-01 3.41855317e-01
-8.85824382e-01 1.01042616e+00 8.35695744e-01 -7.28572547e-01
7.00166225e-02 -8.11597764e-01 2.14252681e-01 -5.25547147e-01
2.92989552e-01 7.87174180e-02 -9.59015787e-02 -3.02783251e-01
1.99148610e-01 1.21804368e+00 1.76744357e-01 -7.17919320e-03
1.36698532e+00 1.10728107e-01 -1.70910731e-02 4.67321634e-01
1.20794201e+00 2.85061628e-01 -2.07008529e+00 -2.17131935e-02
-6.91797793e-01 -7.14467168e-01 -2.11401865e-01 -8.01224113e-01
-1.58147418e+00 9.29239154e-01 2.82876909e-01 -9.46733132e-02
6.09453917e-01 8.88955891e-02 6.24949455e-01 2.33907416e-01
4.42115754e-01 -9.97520864e-01 2.23645031e-01 3.71443003e-01
8.66194069e-01 -1.15434337e+00 -7.83962235e-02 -1.27162814e+00
-6.28357112e-01 1.36699605e+00 7.95421302e-01 -1.42279267e-01
7.60726750e-01 4.90995884e-01 2.98528701e-01 -3.60926211e-01
-3.40889782e-01 3.92861843e-01 1.40775353e-01 7.73808360e-01
1.16676189e-01 1.06744496e-02 8.70031714e-01 4.07903083e-02
-4.01536196e-01 -3.49249423e-01 9.95992064e-01 6.71185195e-01
-1.77355558e-01 -4.79499131e-01 -4.94116694e-01 5.35048783e-01
-3.39396037e-02 -4.11319174e-02 -2.96903074e-01 5.50314069e-01
1.50887877e-01 6.20832980e-01 3.68722945e-01 1.71950944e-02
5.44345975e-01 -4.70414311e-01 6.20411038e-01 -6.46266878e-01
-3.56333047e-01 -2.52941519e-01 -6.06378503e-02 -6.39135122e-01
-2.10993916e-01 -4.90009695e-01 -1.24960780e+00 -4.55494747e-02
-2.06197798e-01 -1.78389028e-01 1.12106943e+00 2.66154349e-01
1.67279676e-01 1.79672942e-01 8.16408575e-01 -1.17172551e+00
2.72752475e-02 -4.48158920e-01 -8.33353221e-01 4.54569042e-01
1.15636975e-01 -6.14114761e-01 -8.08969617e-01 5.69999099e-01] | [9.413822174072266, -3.0035767555236816] |
ce4ac318-c439-40fb-af6c-acf711b4f5cf | weakly-supervised-discriminative-feature | 2002.11939 | null | https://arxiv.org/abs/2002.11939v1 | https://arxiv.org/pdf/2002.11939v1.pdf | Weakly supervised discriminative feature learning with state information for person identification | Unsupervised learning of identity-discriminative visual feature is appealing in real-world tasks where manual labelling is costly. However, the images of an identity can be visually discrepant when images are taken under different states, e.g. different camera views and poses. This visual discrepancy leads to great difficulty in unsupervised discriminative learning. Fortunately, in real-world tasks we could often know the states without human annotation, e.g. we can easily have the camera view labels in person re-identification and facial pose labels in face recognition. In this work we propose utilizing the state information as weak supervision to address the visual discrepancy caused by different states. We formulate a simple pseudo label model and utilize the state information in an attempt to refine the assigned pseudo labels by the weakly supervised decision boundary rectification and weakly supervised feature drift regularization. We evaluate our model on unsupervised person re-identification and pose-invariant face recognition. Despite the simplicity of our method, it could outperform the state-of-the-art results on Duke-reID, MultiPIE and CFP datasets with a standard ResNet-50 backbone. We also find our model could perform comparably with the standard supervised fine-tuning results on the three datasets. Code is available at https://github.com/KovenYu/state-information | ['Wei-Shi Zheng', 'Hong-Xing Yu'] | 2020-02-27 | weakly-supervised-discriminative-feature-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Yu_Weakly_Supervised_Discriminative_Feature_Learning_With_State_Information_for_Person_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Yu_Weakly_Supervised_Discriminative_Feature_Learning_With_State_Information_for_Person_CVPR_2020_paper.pdf | cvpr-2020-6 | ['robust-face-recognition', 'person-identification', 'unsupervised-person-re-identification'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 7.11235031e-02 -1.12488337e-01 -2.13641033e-01 -7.74690092e-01
-4.20501590e-01 -8.55018020e-01 6.50141060e-01 -3.94645691e-01
-5.10517716e-01 6.84822857e-01 1.98778823e-01 2.17432022e-01
1.39099762e-01 -1.75084725e-01 -7.09786713e-01 -7.74208724e-01
3.34631354e-01 6.84650838e-01 -2.74208248e-01 6.73021451e-02
9.61154178e-02 5.77530086e-01 -1.66476524e+00 -4.59689982e-02
5.17993927e-01 7.36974955e-01 -1.01579554e-01 3.11160326e-01
2.72511065e-01 5.09057522e-01 -3.18577319e-01 -6.09653711e-01
6.90517545e-01 -2.49697581e-01 -9.97336805e-01 5.44098139e-01
1.14080739e+00 -3.42503846e-01 -5.03842771e-01 1.40232575e+00
5.61781049e-01 1.16631560e-01 5.26112676e-01 -1.45650709e+00
-6.59355998e-01 7.68003091e-02 -8.55247021e-01 1.09751299e-01
4.90796596e-01 1.27398193e-01 6.65352345e-01 -7.37634957e-01
8.27364802e-01 1.30874407e+00 8.28441620e-01 9.32295918e-01
-1.60111153e+00 -8.97656739e-01 4.03498828e-01 5.18793643e-01
-1.75560677e+00 -9.67691898e-01 8.63356113e-01 -4.84310269e-01
6.07745230e-01 1.25646189e-01 3.69960785e-01 1.36131966e+00
-2.91820586e-01 5.76995015e-01 1.31239057e+00 -2.23673657e-01
-1.97258815e-02 2.77552754e-01 7.10520297e-02 7.88846612e-01
1.97125122e-01 2.09172428e-01 -6.71668887e-01 -9.99006853e-02
7.62201369e-01 2.37550437e-01 -2.31851906e-01 -6.32903814e-01
-9.89324629e-01 4.53159332e-01 3.29878390e-01 8.80759582e-02
1.30209317e-02 6.39837161e-02 2.56163865e-01 3.71137083e-01
5.22835672e-01 7.18911216e-02 -4.76896435e-01 1.16707511e-01
-8.63477886e-01 -1.32964313e-01 6.83327734e-01 9.80823100e-01
9.53608692e-01 -3.92204002e-02 8.65976959e-02 8.17341924e-01
2.62125522e-01 4.50579137e-01 3.38100970e-01 -1.05338454e+00
1.46348655e-01 3.19012612e-01 2.46766686e-01 -9.07580435e-01
-3.90237927e-01 -1.54221818e-01 -1.03513551e+00 1.63059041e-01
8.12555611e-01 -2.77449526e-02 -1.08575261e+00 1.97876394e+00
3.03544074e-01 5.52990139e-01 -1.23223707e-01 1.05005264e+00
7.23456740e-01 1.44274101e-01 5.83919212e-02 -2.09264427e-01
1.39286804e+00 -8.54497850e-01 -6.56462133e-01 -3.55282664e-01
4.52162296e-01 -6.11692846e-01 6.65302694e-01 1.70264378e-01
-6.03108644e-01 -7.83400834e-01 -8.99587870e-01 1.22125126e-01
-3.00937504e-01 3.71966660e-01 6.04322195e-01 8.84426236e-01
-1.28470397e+00 5.96172094e-01 -8.71327281e-01 -7.94817030e-01
3.88989598e-01 6.21958017e-01 -9.87535715e-01 -6.15074225e-02
-8.69985521e-01 8.40554178e-01 8.90849829e-02 3.18732679e-01
-9.10514832e-01 -2.85446525e-01 -1.00122190e+00 -3.20618719e-01
3.53417426e-01 -5.04184008e-01 1.02967465e+00 -1.42326462e+00
-1.44472373e+00 1.36475956e+00 -4.70340103e-01 -1.37637779e-01
5.89450300e-01 -6.58619031e-02 -4.45318252e-01 -7.63785169e-02
2.35829473e-01 8.45152736e-01 9.92949843e-01 -1.36659729e+00
-3.43497455e-01 -7.91031718e-01 -4.47747186e-02 3.70333761e-01
-1.61415458e-01 7.55780190e-02 -5.47808409e-01 -3.43708187e-01
1.67667300e-01 -1.33840871e+00 -5.30830435e-02 9.91067961e-02
-3.22559565e-01 -2.60645658e-01 6.87231660e-01 -7.39079654e-01
5.04436076e-01 -2.22683072e+00 1.67439729e-01 1.12471670e-01
4.09251489e-02 1.04285195e-01 -2.21340194e-01 -7.75298178e-02
-3.40449393e-01 -7.71479355e-03 5.61129600e-02 -7.98455358e-01
-1.50829598e-01 3.62060964e-01 1.47371273e-02 1.20749974e+00
-4.03461233e-02 7.07750678e-01 -7.79698074e-01 -4.03443664e-01
9.58032832e-02 4.37159956e-01 -3.36432993e-01 1.65496334e-01
2.61769533e-01 1.01895380e+00 -4.07767594e-02 7.37325013e-01
8.65191758e-01 -3.64761114e-01 3.35250407e-01 -5.57603121e-01
2.58347660e-04 -5.74220978e-02 -1.49282777e+00 1.90572703e+00
-1.21046029e-01 6.11573219e-01 9.72351506e-02 -1.13959241e+00
6.38162494e-01 2.60559469e-01 4.08376515e-01 -3.80162388e-01
9.00057405e-02 -8.96655619e-02 -2.03494146e-01 -2.60274410e-01
3.05136383e-01 -5.12669608e-02 5.74984960e-02 3.36111039e-01
3.45979333e-01 2.63814241e-01 4.77569178e-02 1.23560112e-02
6.68032825e-01 4.52931970e-01 4.66766469e-02 -3.57261986e-01
5.96872032e-01 -3.62304032e-01 9.55288649e-01 5.88876724e-01
-6.75835967e-01 7.92311788e-01 2.00020596e-01 -6.42024457e-01
-9.43874359e-01 -9.11868215e-01 -2.70678818e-01 1.15574515e+00
4.65396911e-01 -1.51539609e-01 -5.95189750e-01 -8.08195293e-01
7.00288545e-03 2.44028404e-01 -7.08747804e-01 -1.21114507e-01
-3.13688666e-01 -6.30155504e-01 6.31864250e-01 5.10821640e-01
6.78725123e-01 -7.48389125e-01 1.09197512e-01 -3.45402271e-01
-1.61532044e-01 -1.20747805e+00 -7.56848276e-01 4.72317152e-02
-4.77192432e-01 -1.11121511e+00 -6.90120399e-01 -9.81730461e-01
1.10334647e+00 3.34623754e-01 9.17096853e-01 -4.58769053e-02
-2.80813426e-01 6.63935363e-01 3.19973938e-02 3.77670601e-02
-1.16780989e-01 -9.92056448e-03 7.40461111e-01 4.22540009e-01
5.55111289e-01 -5.23981392e-01 -7.51449287e-01 6.36164844e-01
-5.05028069e-01 -9.00432169e-02 1.70932010e-01 9.40383971e-01
5.85668564e-01 -1.27732426e-01 3.96403223e-01 -8.39140952e-01
1.30361825e-01 -1.41843751e-01 -6.02978289e-01 3.96186233e-01
-6.21654093e-01 6.71656430e-02 3.75399500e-01 -5.22692621e-01
-1.22404051e+00 5.85026920e-01 4.22152095e-02 -5.83305359e-01
-5.24940133e-01 -1.17069289e-01 -2.72099674e-01 -3.86785299e-01
6.30753756e-01 1.81388542e-01 9.48061123e-02 -4.97872740e-01
2.76580691e-01 7.31956303e-01 6.11515462e-01 -6.30098462e-01
9.83507276e-01 7.11063027e-01 -1.98808983e-01 -7.02223778e-01
-8.58052790e-01 -6.75726414e-01 -1.01346874e+00 -2.28824317e-01
8.47604215e-01 -1.22124958e+00 -8.09400797e-01 7.11035013e-01
-1.11526716e+00 -1.54145300e-01 -1.44634143e-01 2.55483747e-01
-4.37362403e-01 7.33968616e-01 -6.22968137e-01 -6.35018349e-01
-4.70747948e-02 -1.15976477e+00 1.07448709e+00 5.23941457e-01
-8.20164382e-02 -8.92216086e-01 4.25345078e-02 4.74373639e-01
8.26049298e-02 -7.39496797e-02 7.27017224e-02 -6.03232682e-01
-3.39794904e-01 -7.07273409e-02 -2.76261598e-01 4.59583282e-01
3.96559000e-01 -2.91129529e-01 -1.34753823e+00 -7.71202207e-01
-1.40567020e-01 -5.70845902e-01 7.87454188e-01 1.36683702e-01
1.14351487e+00 -1.97742268e-01 -2.93439776e-01 9.31577563e-01
1.11533296e+00 -1.68378115e-01 4.95764732e-01 2.22480983e-01
9.77755666e-01 7.56867588e-01 3.49856734e-01 3.24714780e-01
5.16192198e-01 9.22180295e-01 9.50422585e-02 -9.07389373e-02
-2.47461721e-01 -2.94987112e-01 4.51436102e-01 5.02616286e-01
-4.11872268e-01 1.65631846e-01 -8.13616216e-01 4.81518954e-01
-1.92203879e+00 -9.99184072e-01 2.68553257e-01 2.25894713e+00
8.53516102e-01 -2.47513905e-01 1.55036133e-02 -2.54007101e-01
1.11333561e+00 1.20260872e-01 -7.15692639e-01 1.66965723e-01
-1.22223288e-01 -2.03573883e-01 8.75513017e-01 4.54079449e-01
-1.27194047e+00 1.08311057e+00 5.52798128e+00 5.02189636e-01
-1.02263141e+00 3.60982478e-01 7.45466888e-01 -3.49440537e-02
2.37972289e-01 6.70759380e-02 -1.08795607e+00 3.18381876e-01
6.58013344e-01 6.79121241e-02 6.40237212e-01 7.29326487e-01
-5.71812019e-02 3.19489203e-02 -1.47105849e+00 1.56551731e+00
5.03242075e-01 -8.96826982e-01 -2.45020911e-01 2.04778519e-02
8.38545620e-01 -3.69777866e-02 6.31813258e-02 7.32716396e-02
3.06147724e-01 -9.28388000e-01 5.84067881e-01 4.49577749e-01
1.13935626e+00 -5.17203927e-01 6.73587739e-01 1.17309444e-01
-1.15078151e+00 3.67193744e-02 -4.58507031e-01 -8.29976518e-03
5.64005896e-02 1.14680395e-01 -5.46822727e-01 3.90353054e-01
8.43279541e-01 1.02887130e+00 -8.48684728e-01 7.08171308e-01
-2.34374449e-01 3.30305159e-01 -4.26925659e-01 5.23803234e-01
-3.11156780e-01 -2.15160996e-01 4.07943845e-01 9.23962772e-01
1.23320241e-02 -1.73898369e-01 3.51225883e-01 5.89552164e-01
-1.93304330e-01 -2.58690238e-01 -5.77374637e-01 2.69180715e-01
3.04902434e-01 1.19663405e+00 -6.38628364e-01 -2.27685899e-01
-4.92972195e-01 1.46976101e+00 4.66939479e-01 5.58256328e-01
-8.00685346e-01 2.53404826e-01 8.68390143e-01 1.75912499e-01
1.31154075e-01 -2.35002682e-01 5.98699860e-02 -1.63482559e+00
4.76244576e-02 -7.47594893e-01 4.57086295e-01 -6.06450856e-01
-1.60085201e+00 5.77390194e-01 5.00662513e-02 -1.17937815e+00
-2.59327352e-01 -6.38792574e-01 -1.90041512e-01 7.33036935e-01
-1.49444056e+00 -1.36601913e+00 -4.25916165e-01 8.95201266e-01
4.62371796e-01 -3.81897479e-01 8.34043562e-01 4.46673840e-01
-8.29212964e-01 9.43107963e-01 4.36643921e-02 4.59304601e-01
1.32132447e+00 -1.05377984e+00 3.23539644e-01 9.16968584e-01
3.49617630e-01 6.49580121e-01 6.36726141e-01 -6.31745219e-01
-1.37400484e+00 -1.04434466e+00 6.67338848e-01 -5.61554790e-01
5.77811658e-01 -5.42237937e-01 -7.06188500e-01 9.71150398e-01
1.66647747e-01 4.62745279e-01 7.44402289e-01 2.49261767e-01
-6.25900328e-01 -2.86477059e-01 -1.24694133e+00 4.54488218e-01
1.49128330e+00 -7.99869955e-01 -3.89127702e-01 4.47406054e-01
1.20944023e-01 -3.60578835e-01 -6.96637928e-01 2.25934565e-01
5.66028059e-01 -7.18180776e-01 9.62879121e-01 -6.07464969e-01
-2.92826742e-01 -5.15436471e-01 2.63632797e-02 -1.15733147e+00
-5.38815677e-01 -5.77179670e-01 2.06863865e-01 1.54412401e+00
7.61396214e-02 -8.45004201e-01 9.49500918e-01 9.81011510e-01
4.06313390e-01 -1.50191914e-02 -1.19650066e+00 -8.34846377e-01
-2.72855073e-01 9.26535018e-03 3.94938588e-01 1.10312343e+00
-1.83552355e-01 3.44580412e-01 -6.48982048e-01 4.59060460e-01
9.55500185e-01 3.45009267e-02 8.13743711e-01 -1.37095118e+00
-1.32333845e-01 -1.66629270e-01 -7.43002594e-01 -1.07570577e+00
7.87465513e-01 -9.18325603e-01 -8.11773352e-03 -9.53627765e-01
5.69543183e-01 -4.12125319e-01 -3.05862427e-01 8.16162407e-01
-6.52338238e-03 6.11065984e-01 2.16336757e-01 5.06135702e-01
-8.30378532e-01 4.17705268e-01 9.23790932e-01 -4.34994668e-01
-1.94455516e-02 -5.58207668e-02 -6.09834850e-01 8.12966466e-01
8.38774145e-01 -4.74579096e-01 -1.81362346e-01 -4.41750765e-01
2.94771120e-02 -3.00754398e-01 6.80300117e-01 -8.83286476e-01
4.98282373e-01 -1.78996623e-02 7.35586643e-01 -1.70877397e-01
3.33988994e-01 -1.03137040e+00 3.89838904e-01 2.06251174e-01
-1.96863398e-01 -2.94813178e-02 8.73459280e-02 6.65229976e-01
-4.46995869e-02 -1.78430438e-01 8.08026731e-01 -1.79944813e-01
-9.54813361e-01 5.35458565e-01 -1.72336791e-02 -1.53143406e-01
8.91297758e-01 -3.48580182e-01 -3.47507060e-01 -4.66943532e-01
-9.17688906e-01 1.23191327e-01 9.00265276e-01 6.32927954e-01
3.21141422e-01 -1.43382955e+00 -7.46203840e-01 4.03875977e-01
2.08570108e-01 -2.28560254e-01 3.52093220e-01 8.07230115e-01
-1.24176130e-01 1.46025524e-01 -3.51022065e-01 -8.01787078e-01
-1.55480015e+00 4.43654865e-01 4.72791165e-01 -1.37464236e-03
-3.31444055e-01 7.20307171e-01 3.68253946e-01 -5.15003502e-01
3.71235251e-01 3.51785004e-01 -1.92293420e-01 9.20538530e-02
4.83327657e-01 2.46951580e-01 -6.81142807e-02 -1.23160160e+00
-6.55957222e-01 7.77997196e-01 -3.30781907e-01 4.31624353e-02
1.17752647e+00 -3.93470258e-01 -1.15668938e-01 2.44204283e-01
1.23421884e+00 -9.90392640e-02 -1.58842778e+00 -2.81354964e-01
-1.75394133e-01 -5.58974862e-01 -2.24762306e-01 -5.50950587e-01
-1.32225418e+00 5.19013286e-01 1.17453408e+00 -2.17368379e-01
9.42127526e-01 1.98796630e-01 2.61713028e-01 4.62844342e-01
6.52140498e-01 -1.18619204e+00 -9.84804183e-02 3.19325507e-01
7.09435582e-01 -1.67532706e+00 1.34933561e-01 -3.82831901e-01
-6.43418252e-01 8.44972849e-01 7.43480027e-01 -6.96965307e-02
5.69552720e-01 3.72069329e-02 2.56391734e-01 -3.28909978e-02
-2.72452742e-01 -3.38540822e-01 2.74210513e-01 7.51491785e-01
3.39697957e-01 6.07297011e-02 3.69938370e-03 3.47367942e-01
-7.02867359e-02 -8.20482299e-02 2.59935707e-01 7.73092091e-01
1.83561161e-01 -1.24571836e+00 -4.71893013e-01 1.30422905e-01
-4.00401115e-01 1.51821956e-01 -4.42674756e-01 4.43041921e-01
2.33017311e-01 8.70874047e-01 1.62483707e-01 -3.90427738e-01
2.26803005e-01 2.43008539e-01 7.98727930e-01 -5.24859369e-01
-2.89995819e-01 -1.11280560e-01 -7.85760488e-03 -5.76541305e-01
-8.78893554e-01 -9.10791636e-01 -8.35980773e-01 -3.16505313e-01
-2.77767211e-01 -1.31822720e-01 3.99590552e-01 7.85781026e-01
4.56059128e-01 -8.33347961e-02 6.28593683e-01 -1.20410836e+00
-5.91853678e-01 -1.03917813e+00 -5.99810004e-01 9.92427826e-01
3.54201049e-01 -8.75600934e-01 -5.03517568e-01 5.61294615e-01] | [14.608723640441895, 1.0699182748794556] |
81f9d40a-85f6-4166-b291-5c99ad273448 | optimal-clustering-with-bandit-feedback | 2202.04294 | null | https://arxiv.org/abs/2202.04294v1 | https://arxiv.org/pdf/2202.04294v1.pdf | Optimal Clustering with Bandit Feedback | This paper considers the problem of online clustering with bandit feedback. A set of arms (or items) can be partitioned into various groups that are unknown. Within each group, the observations associated to each of the arms follow the same distribution with the same mean vector. At each time step, the agent queries or pulls an arm and obtains an independent observation from the distribution it is associated to. Subsequent pulls depend on previous ones as well as the previously obtained samples. The agent's task is to uncover the underlying partition of the arms with the least number of arm pulls and with a probability of error not exceeding a prescribed constant $\delta$. The problem proposed finds numerous applications from clustering of variants of viruses to online market segmentation. We present an instance-dependent information-theoretic lower bound on the expected sample complexity for this task, and design a computationally efficient and asymptotically optimal algorithm, namely Bandit Online Clustering (BOC). The algorithm includes a novel stopping rule for adaptive sequential testing that circumvents the need to exactly solve any NP-hard weighted clustering problem as its subroutines. We show through extensive simulations on synthetic and real-world datasets that BOC's performance matches the lower bound asymptotically, and significantly outperforms a non-adaptive baseline algorithm. | ['Vincent Y. F. Tan', 'Zixin Zhong', 'Junwen Yang'] | 2022-02-09 | null | null | null | null | ['online-clustering'] | ['computer-vision'] | [ 4.72388603e-02 -2.69448552e-02 -9.40813363e-01 -1.76681876e-01
-9.65226591e-01 -9.94185150e-01 -4.49308269e-02 2.23917007e-01
-3.20000440e-01 7.55144954e-01 -3.28032613e-01 -6.43836379e-01
-8.59256446e-01 -5.90161383e-01 -9.32403862e-01 -1.04559374e+00
-2.77131706e-01 1.42900920e+00 1.94334537e-01 6.53329849e-01
2.99163103e-01 2.54515648e-01 -1.00075912e+00 1.89495802e-01
6.91675603e-01 1.27848136e+00 4.66219224e-02 8.57199073e-01
-1.90964207e-01 6.03869498e-01 -4.81314272e-01 -4.27217960e-01
7.12721884e-01 -6.75850391e-01 -8.60639572e-01 8.54397118e-01
-3.86165828e-01 -2.63801098e-01 4.93643582e-02 1.23299408e+00
-5.46153225e-02 2.01730639e-01 9.20290649e-01 -1.29162455e+00
-4.19614345e-01 1.05442512e+00 -1.18201756e+00 4.05457139e-01
-4.47966829e-02 -4.14872952e-02 1.28398681e+00 -3.31393965e-02
3.65631461e-01 9.30224657e-01 2.57168978e-01 2.26965517e-01
-1.69713664e+00 -5.69530427e-01 4.73086745e-01 8.43904912e-02
-1.11890280e+00 -1.39045358e-01 5.98320961e-01 -3.08420151e-01
2.64921755e-01 5.41498423e-01 6.61119163e-01 4.69979972e-01
-5.31423271e-01 1.15519071e+00 8.70396376e-01 -5.49917996e-01
9.24766123e-01 2.90052980e-01 3.81058544e-01 3.85160267e-01
6.34816468e-01 -1.66714564e-01 -1.96597710e-01 -6.08620882e-01
4.60335582e-01 4.61970031e-01 -1.09273300e-01 -5.89897513e-01
-7.36801684e-01 1.12897038e+00 9.24753100e-02 7.10244849e-02
-4.93296474e-01 3.02976429e-01 -1.46156490e-01 5.95098436e-01
5.62710583e-01 2.24815607e-02 -5.09902060e-01 1.64760932e-01
-7.33321130e-01 1.32711142e-01 1.04048526e+00 9.43204105e-01
7.46484697e-01 -4.79349971e-01 -2.33001709e-01 7.10343421e-01
2.59626389e-01 6.89333022e-01 2.04909086e-01 -1.16247118e+00
9.01608706e-01 5.66512287e-01 8.29225302e-01 -7.84175634e-01
-1.34623662e-01 -5.37434280e-01 -5.28552830e-01 -9.77526754e-02
7.71263897e-01 -4.18495446e-01 -6.76132679e-01 1.66221786e+00
6.44105315e-01 -2.27510884e-01 -4.58522946e-01 7.10122585e-01
-4.33128148e-01 8.53917003e-01 -4.58712041e-01 -7.93339610e-01
9.88052905e-01 -8.55210364e-01 -5.30794024e-01 -3.78817737e-01
4.00213510e-01 -3.76524746e-01 3.02857161e-01 5.71205497e-01
-1.13910055e+00 2.25382254e-01 -7.79061079e-01 9.10451591e-01
-2.24964041e-02 -1.37953222e-01 8.28310013e-01 1.13627017e+00
-8.48827899e-01 2.92275131e-01 -8.65384161e-01 -9.59034562e-02
4.52851772e-01 6.22428477e-01 2.05034703e-01 -1.04847513e-01
-3.03889513e-01 7.37662986e-02 3.97954792e-01 -6.36891872e-02
-6.19025469e-01 -4.98252064e-01 -2.11032465e-01 3.82250622e-02
8.98849368e-01 -5.31941831e-01 1.59401333e+00 -1.54797792e+00
-1.35011542e+00 3.13152283e-01 -4.59520817e-01 -7.45689809e-01
5.38163781e-01 3.15385759e-01 1.87923126e-02 2.55512923e-01
3.13326597e-01 -7.36649334e-02 9.34855878e-01 -1.30747116e+00
-1.14798486e+00 -7.88821816e-01 -1.53803900e-01 5.07264175e-02
-3.56326342e-01 -9.40053165e-02 -6.94539011e-01 -4.07095999e-01
2.46794254e-01 -1.17664480e+00 -4.80191767e-01 -4.39432293e-01
-5.64942420e-01 -2.14191779e-01 3.16101611e-01 -2.48470038e-01
1.17422259e+00 -1.97510123e+00 2.42967345e-02 8.21287870e-01
-2.20619142e-03 -2.77042300e-01 -6.80798069e-02 5.78531146e-01
1.66150093e-01 2.32400134e-01 -1.37411833e-01 -1.90895885e-01
9.86523628e-02 -4.41955775e-02 -2.66964972e-01 8.35363209e-01
-6.08557582e-01 4.75070298e-01 -5.22302032e-01 -1.82386026e-01
-1.07739471e-01 -7.56699026e-01 -7.55397439e-01 7.17771351e-02
-6.89564764e-01 3.85066792e-02 -7.36171603e-01 2.25517690e-01
7.19475746e-01 -8.17431688e-01 6.36596382e-01 6.63087487e-01
8.89376402e-02 -1.90083638e-01 -1.64874756e+00 6.47651851e-01
1.18978798e-01 1.81802884e-01 4.33577329e-01 -1.49947774e+00
1.91362187e-01 2.90367514e-01 6.39741540e-01 -2.61954010e-01
1.67357326e-01 2.49911174e-01 -1.47830741e-03 -3.34093630e-01
-4.45609055e-02 -5.88675626e-02 -1.89225059e-02 1.09090292e+00
-3.40169817e-01 5.08997738e-01 3.22017342e-01 3.39767873e-01
1.10869145e+00 -8.38791668e-01 1.79573655e-01 -6.11855388e-02
6.85494691e-02 2.15367541e-01 5.31484783e-01 1.48436666e+00
5.58422990e-02 1.76364958e-01 7.34567642e-01 -1.37563735e-01
-7.72968888e-01 -8.58834624e-01 3.11125845e-01 1.26716888e+00
2.07151413e-01 3.17126870e-01 -8.37393701e-01 -5.83050549e-01
5.12588441e-01 4.59232360e-01 -1.05195045e+00 4.52882379e-01
9.53259617e-02 -8.45117688e-01 -4.05853748e-01 1.32383898e-01
3.91775250e-01 -6.34658635e-01 -2.78518379e-01 4.65206683e-01
-1.08591057e-01 -7.86451697e-01 -8.35638762e-01 3.16198498e-01
-9.40791011e-01 -1.30500805e+00 -5.44687629e-01 -6.25832319e-01
8.42210352e-01 7.36424685e-01 5.15874684e-01 -1.82982400e-01
1.18782297e-01 4.46676165e-01 -4.62638915e-01 -4.82111365e-01
-1.12292670e-01 2.13835418e-01 -2.47426301e-01 7.15032458e-01
3.87794197e-01 -2.28581000e-02 -7.87979066e-01 5.23551524e-01
-8.57947350e-01 -4.41256642e-01 5.13678849e-01 6.09338522e-01
6.15598559e-01 5.33288777e-01 5.76567411e-01 -1.19214332e+00
5.26716113e-01 -7.62273192e-01 -1.25405753e+00 3.78998131e-01
-4.14259344e-01 -1.37025779e-02 5.58544517e-01 -6.62382782e-01
-8.33100438e-01 1.52830198e-01 7.40551531e-01 -2.22528279e-01
6.70700669e-02 6.83702230e-01 8.18032920e-02 3.91214997e-01
1.42286137e-01 2.15621978e-01 1.39358357e-01 -4.65116024e-01
1.89905077e-01 8.37759495e-01 2.66753316e-01 -4.64033782e-01
6.54173255e-01 6.07819676e-01 -1.51451379e-01 -6.22059882e-01
-9.80031073e-01 -7.78524637e-01 -3.35487984e-02 -1.07277706e-01
3.53949964e-01 -5.07859349e-01 -1.34019458e+00 2.86980659e-01
-8.17229271e-01 -5.97370744e-01 -2.77797192e-01 4.89147753e-01
-6.64318860e-01 2.21857533e-01 -5.12007177e-01 -1.59723568e+00
9.70217958e-02 -7.53556550e-01 3.72374117e-01 1.21625990e-01
2.46632239e-03 -6.61557913e-01 1.65658474e-01 6.40793264e-01
-4.85152379e-02 -1.16184548e-01 1.06016564e+00 -1.09240949e+00
-1.00268579e+00 -4.96071398e-01 4.25351858e-02 -1.02350093e-01
1.64152771e-01 -1.07479610e-01 -2.69341946e-01 -5.78038394e-01
1.73047930e-02 -2.68899631e-02 7.04322696e-01 9.93894935e-01
1.41812146e+00 -9.22775567e-01 -7.26897120e-01 1.00931920e-01
1.32745600e+00 8.55620861e-01 -3.86307426e-02 2.92353302e-01
-1.07475102e-01 5.78359246e-01 5.27013183e-01 7.53658891e-01
-4.17339988e-02 5.15874743e-01 4.61588234e-01 3.13575417e-01
7.95068979e-01 -7.08694309e-02 1.51865110e-01 3.00153971e-01
3.15910906e-01 -6.17528439e-01 -4.22256827e-01 9.00465190e-01
-2.23452878e+00 -1.12150073e+00 -5.30265905e-02 2.76073861e+00
8.69665384e-01 1.46790087e-01 8.09715807e-01 2.26187557e-01
1.06550014e+00 -3.58147979e-01 -1.10319614e+00 -3.06558311e-01
1.60550281e-01 -8.10227767e-02 1.10178423e+00 7.03202546e-01
-9.50292349e-01 4.01988983e-01 5.80398846e+00 1.02277040e+00
-3.38519514e-01 1.10241622e-01 1.13111675e+00 -3.60199451e-01
-3.03449839e-01 2.28709150e-02 -7.70184457e-01 6.84345484e-01
1.00298905e+00 -2.85257995e-01 1.00354314e+00 8.28788459e-01
3.01299930e-01 -4.70854402e-01 -1.20916450e+00 6.83699489e-01
-3.41364712e-01 -1.43507063e+00 -3.42029959e-01 5.99038303e-01
1.24648082e+00 -5.77752143e-02 2.97649622e-01 -1.00613549e-01
1.15457499e+00 -5.30183494e-01 6.68972254e-01 8.39157701e-02
3.94733191e-01 -1.19645584e+00 4.60362643e-01 8.40224206e-01
-7.99258947e-01 -8.83996129e-01 -1.68238714e-01 4.12516557e-02
-2.56247818e-01 6.25737548e-01 -7.83202410e-01 2.55200863e-01
6.54208302e-01 8.84000137e-02 6.15652204e-02 1.43453634e+00
2.59943992e-01 1.02461672e+00 -6.58513784e-01 -4.72850651e-01
4.52595264e-01 -6.97845697e-01 2.96020061e-01 7.42859006e-01
1.72057927e-01 4.15238082e-01 4.55486864e-01 4.90606219e-01
-3.22567254e-01 1.85597703e-01 -3.74566764e-01 -2.22455978e-01
9.20070410e-01 7.05211043e-01 -1.17625487e+00 -4.76779222e-01
-2.81800658e-01 7.53858924e-01 2.78818816e-01 6.20582461e-01
-6.31790578e-01 -3.47296923e-01 4.82428640e-01 7.23338053e-02
9.69308197e-01 1.71341375e-01 -3.26809198e-01 -5.62662065e-01
-2.27156244e-02 -6.80137575e-01 1.03158128e+00 -1.17559157e-01
-1.47650361e+00 -1.04017846e-01 -2.87326500e-02 -8.57290864e-01
-4.10372794e-01 -2.58275121e-01 -4.35742021e-01 4.26326901e-01
-8.33164215e-01 -3.06231558e-01 4.80158746e-01 5.94071567e-01
5.53580523e-01 -2.52623428e-02 3.46405894e-01 -2.79250383e-01
-7.48238742e-01 4.46799427e-01 1.12004030e+00 1.59203392e-02
3.47789302e-02 -1.25149047e+00 -3.33311975e-01 5.66276014e-01
3.14494148e-02 2.87791938e-01 7.46150613e-01 -5.48715889e-01
-1.46480525e+00 -1.12712955e+00 3.63783807e-01 -1.43207520e-01
8.50028336e-01 -1.87240213e-01 -4.05708104e-01 1.02308702e+00
1.25975385e-01 -2.52934217e-01 8.26524317e-01 1.36633769e-01
-1.02636464e-01 -4.82313663e-01 -1.32334054e+00 4.51027513e-01
7.16487646e-01 1.01216197e-01 -8.51664469e-02 6.49004698e-01
6.28486574e-01 1.68940559e-01 -4.23027068e-01 -1.31518826e-01
3.14710706e-01 -8.99063468e-01 5.37032843e-01 -7.71204233e-01
-9.92233604e-02 -1.51844829e-01 -2.12353170e-02 -1.15040791e+00
-5.54494917e-01 -1.11232674e+00 -2.10047945e-01 1.04343832e+00
7.40171611e-01 -9.90618944e-01 1.31598604e+00 6.72239184e-01
7.44697630e-01 -4.81852591e-01 -1.18686569e+00 -9.65723991e-01
2.57510860e-02 -3.81151140e-01 6.06735945e-01 6.06859863e-01
2.64984369e-01 2.55210429e-01 -2.67320216e-01 3.74667078e-01
1.02330232e+00 8.81641030e-01 9.03200924e-01 -1.17202055e+00
-8.25609446e-01 -5.76468527e-01 3.41490775e-01 -1.44958675e+00
-9.50191393e-02 -5.67727745e-01 -2.32769642e-02 -1.26810932e+00
8.37718666e-01 -4.76554900e-01 -1.06494360e-01 1.54697970e-01
1.62470415e-01 -1.93602890e-01 1.42998293e-01 2.55592912e-01
-1.03986692e+00 -2.59282533e-02 8.13052297e-01 -3.74839336e-01
-6.12240493e-01 8.60794008e-01 -8.90276134e-01 4.84793663e-01
8.11850250e-01 -7.37467289e-01 -4.29896742e-01 -1.46584824e-01
1.60246089e-01 7.39817500e-01 -1.16907634e-01 -2.32964069e-01
5.64789593e-01 -5.12532592e-01 2.60039270e-01 -1.08249342e+00
-2.29960568e-02 -1.23335803e+00 4.46148098e-01 8.28484952e-01
-7.39161253e-01 -1.35648057e-01 -3.86461914e-01 1.35428727e+00
3.72660786e-01 -5.69647968e-01 7.28062332e-01 3.77627020e-03
4.06703204e-01 4.62916851e-01 -8.52803707e-01 -4.37571295e-02
1.26237094e+00 -2.44919375e-01 -1.06582955e-01 -8.66712034e-01
-7.68546581e-01 6.36484265e-01 2.03690380e-01 -1.75147057e-01
1.62908673e-01 -1.03167999e+00 -4.48973119e-01 5.60388491e-02
-1.15974106e-01 -1.94987372e-01 1.32081717e-01 8.17460120e-01
-1.60421785e-02 7.17634618e-01 6.60488904e-01 -4.96402770e-01
-1.01609969e+00 1.08661175e+00 1.75882056e-01 -4.26153123e-01
-7.65651017e-02 6.10358596e-01 4.86430749e-02 -3.06222346e-02
4.19725209e-01 -2.03734592e-01 2.25966096e-01 1.49039343e-01
5.39063454e-01 6.32628381e-01 -7.80295357e-02 8.06447193e-02
-6.18249699e-02 1.98966905e-01 -4.57819313e-01 -6.14587069e-01
1.39073706e+00 -3.51421505e-01 -1.51682660e-01 5.34740210e-01
1.00090039e+00 -8.54980201e-02 -1.27873242e+00 -6.08867466e-01
3.88505042e-01 -7.67988503e-01 -3.15980613e-01 -7.32012689e-01
-1.36565590e+00 7.72371562e-03 3.00745577e-01 1.20751214e+00
8.68342996e-01 4.73363608e-01 3.51655722e-01 4.59149241e-01
6.70300484e-01 -1.34490895e+00 9.55190808e-02 7.19239339e-02
3.73587996e-01 -9.05955374e-01 -1.87003434e-01 -2.37890035e-01
-4.21943039e-01 6.02892756e-01 1.09499425e-01 -3.92582454e-02
7.36042142e-01 -4.06703502e-02 -6.02934241e-01 -2.60559935e-02
-1.00383258e+00 -2.41628870e-01 -1.99565098e-01 4.04648125e-01
-3.76653045e-01 3.74548078e-01 -3.09107870e-01 6.90852880e-01
5.85944019e-02 -5.69354557e-02 3.94817144e-01 7.76460409e-01
-8.60781491e-01 -8.38604152e-01 -7.66076386e-01 9.11605120e-01
-8.84968698e-01 4.12519097e-01 -5.14371037e-01 4.69110399e-01
-4.70514268e-01 1.41101074e+00 3.76002133e-01 1.04891427e-01
-1.02693982e-01 -2.74597853e-01 3.94007444e-01 -4.42736566e-01
-2.38339618e-01 5.90984583e-01 -1.56504974e-01 -1.81165725e-01
-3.24630737e-01 -9.03497934e-01 -7.49192536e-01 -5.49128532e-01
-6.21226966e-01 6.92157447e-01 5.48109829e-01 9.58002031e-01
1.83332011e-01 -1.53519571e-01 1.53582740e+00 -4.93975520e-01
-1.02191949e+00 -7.87998617e-01 -1.04315650e+00 1.83700219e-01
3.58050942e-01 -3.61933202e-01 -6.46296024e-01 -9.46699455e-02] | [4.566433906555176, 3.3497769832611084] |
172b1e75-71ad-4c19-9fea-9e4865e9c779 | relationprompt-leveraging-prompts-to-generate | 2203.09101 | null | https://arxiv.org/abs/2203.09101v1 | https://arxiv.org/pdf/2203.09101v1.pdf | RelationPrompt: Leveraging Prompts to Generate Synthetic Data for Zero-Shot Relation Triplet Extraction | Despite the importance of relation extraction in building and representing knowledge, less research is focused on generalizing to unseen relations types. We introduce the task setting of Zero-Shot Relation Triplet Extraction (ZeroRTE) to encourage further research in low-resource relation extraction methods. Given an input sentence, each extracted triplet consists of the head entity, relation label, and tail entity where the relation label is not seen at the training stage. To solve ZeroRTE, we propose to synthesize relation examples by prompting language models to generate structured texts. Concretely, we unify language model prompts and structured text approaches to design a structured prompt template for generating synthetic relation samples when conditioning on relation label prompts (RelationPrompt). To overcome the limitation for extracting multiple relation triplets in a sentence, we design a novel Triplet Search Decoding method. Experiments on FewRel and Wiki-ZSL datasets show the efficacy of RelationPrompt for the ZeroRTE task and zero-shot relation classification. Our code and data are available at github.com/declare-lab/RelationPrompt. | ['Luo Si', 'Soujanya Poria', 'Lidong Bing', 'Yew Ken Chia'] | 2022-03-17 | null | https://aclanthology.org/2022.findings-acl.5 | https://aclanthology.org/2022.findings-acl.5.pdf | findings-acl-2022-5 | ['zero-shot-relation-triplet-extraction', 'relation-classification'] | ['natural-language-processing', 'natural-language-processing'] | [ 6.05840147e-01 9.26529288e-01 -5.02075315e-01 -3.69896948e-01
-8.91573846e-01 -3.45470130e-01 7.31887698e-01 3.16608846e-01
-6.92681819e-02 9.50187802e-01 3.85784328e-01 -6.29761755e-01
-5.87973557e-02 -1.00139213e+00 -6.22793317e-01 -1.97636962e-01
2.47241035e-01 8.47481012e-01 -2.17633788e-02 -4.03473526e-01
-3.16530019e-01 -1.45816058e-03 -1.23203981e+00 6.32316887e-01
8.68626714e-01 7.34079242e-01 -2.16522217e-02 4.87754345e-01
-2.79572040e-01 1.19589305e+00 -7.29045510e-01 -9.33433115e-01
3.01228091e-02 -6.39535785e-01 -1.29150999e+00 -7.30137248e-03
-4.69146855e-02 -1.11998625e-01 -4.74683344e-01 6.48428261e-01
5.79845369e-01 2.51124889e-01 8.95697474e-01 -1.40306723e+00
-8.84052753e-01 1.40110552e+00 -3.03325087e-01 2.99656451e-01
7.51100540e-01 -1.52725190e-01 1.55400324e+00 -9.55005586e-01
9.98928308e-01 1.07381773e+00 3.21440220e-01 7.16159999e-01
-1.12322426e+00 -6.27926528e-01 4.98708114e-02 4.18089449e-01
-1.44488049e+00 -6.39647782e-01 5.91509640e-01 -3.43809724e-01
1.65326631e+00 4.70091194e-01 3.29954058e-01 1.57949376e+00
-2.42677316e-01 9.46530342e-01 5.97403824e-01 -8.56566370e-01
-1.27530113e-01 7.09888935e-02 5.67141414e-01 4.68998522e-01
2.98308879e-01 -1.75187048e-02 -7.80622244e-01 -1.43123865e-01
4.38811034e-01 -6.04408860e-01 -2.87524998e-01 1.76899999e-01
-1.16012287e+00 5.48309982e-01 9.57166031e-02 2.83664376e-01
-3.08348298e-01 -2.72100449e-01 3.02045405e-01 3.58537436e-01
6.87828958e-01 5.07018447e-01 -6.07265294e-01 -4.20221090e-02
-4.84557807e-01 2.38875762e-01 1.06928062e+00 1.69665635e+00
6.24217749e-01 -4.44203198e-01 -9.49181497e-01 1.01532388e+00
6.45930991e-02 1.35381773e-01 3.75458032e-01 -5.19361138e-01
1.12174642e+00 7.97089696e-01 1.04391620e-01 -5.96093953e-01
-3.39919984e-01 -1.68142691e-01 -6.43424273e-01 -7.13748693e-01
2.13607192e-01 -4.62543607e-01 -7.63749599e-01 1.52919698e+00
6.05163991e-01 2.57303357e-01 6.54409349e-01 6.67951286e-01
1.66613162e+00 6.29172027e-01 1.51487753e-01 -3.47330749e-01
1.80040967e+00 -1.10922766e+00 -9.54400659e-01 -4.41286922e-01
1.30883980e+00 -5.27277887e-01 9.82991576e-01 -1.20623000e-01
-7.04195023e-01 -2.56466538e-01 -9.25878406e-01 -4.60799932e-01
-3.15756381e-01 1.38075635e-01 8.31655681e-01 2.99143523e-01
-3.26543450e-01 4.07592624e-01 -7.28912771e-01 -4.04319942e-01
3.01467955e-01 1.09021500e-01 -3.33270073e-01 -8.90760198e-02
-1.70289719e+00 1.01186919e+00 7.16362119e-01 1.26337260e-01
-4.25644606e-01 -6.83816612e-01 -1.32808375e+00 1.93250775e-02
9.66474891e-01 -8.08519244e-01 1.48592472e+00 -5.33543266e-02
-1.23470068e+00 7.92813838e-01 -5.63315094e-01 -5.48605442e-01
1.24816880e-01 -1.62336066e-01 -6.14411771e-01 -1.37777299e-01
2.66470850e-01 3.52724910e-01 2.13727891e-01 -1.08374953e+00
-5.36585391e-01 4.31085341e-02 2.02416107e-02 3.87237072e-01
-2.04332054e-01 4.41822946e-01 -3.37081909e-01 -5.51970959e-01
-1.94627307e-02 -5.47639489e-01 -8.92617628e-02 -7.66426682e-01
-1.08924580e+00 -8.65469754e-01 3.90332073e-01 -4.38563049e-01
1.40895629e+00 -1.83663487e+00 -9.75043178e-02 3.71819139e-02
2.53092915e-01 2.86092132e-01 -1.98127881e-01 7.13763535e-01
-3.93423349e-01 1.92559525e-01 -1.51505858e-01 -2.23213032e-01
5.71521334e-02 4.62822467e-01 -5.14313042e-01 -3.04688036e-01
6.31344736e-01 1.47412550e+00 -1.26398349e+00 -8.31070304e-01
-1.57985330e-01 2.08921701e-01 -8.51468295e-02 4.78315681e-01
-4.01302308e-01 1.76017836e-01 -6.51926637e-01 6.35282516e-01
1.56661570e-01 -6.85576975e-01 3.77425969e-01 -3.09414238e-01
4.34448868e-01 1.08240426e+00 -9.70924795e-01 1.17271912e+00
-3.63254428e-01 3.58210027e-01 -6.97039127e-01 -6.39680624e-01
1.04423416e+00 9.21915114e-01 1.80583224e-01 -4.39304441e-01
2.61613697e-01 6.82070330e-02 2.01061949e-01 -7.89479494e-01
6.88535810e-01 -1.93416387e-01 -1.46252498e-01 6.45378649e-01
4.65942174e-01 -7.39552602e-02 7.26737142e-01 5.58378637e-01
1.41941392e+00 3.17706466e-01 5.89444697e-01 3.96322310e-01
1.84472322e-01 6.49560615e-02 6.24613583e-01 6.79731727e-01
2.57627308e-01 4.45749402e-01 7.02724040e-01 -1.49609283e-01
-6.70716703e-01 -6.28984511e-01 -3.61477956e-03 9.97235358e-01
-4.85631190e-02 -1.08772242e+00 -3.43592376e-01 -9.20953572e-01
-2.22105160e-01 1.08267093e+00 -5.41209102e-01 -1.88484848e-01
-6.41905546e-01 -8.67018342e-01 7.20381320e-01 4.89021093e-01
7.42869899e-02 -1.36949980e+00 -2.55026013e-01 3.66563261e-01
-8.14632118e-01 -1.66660392e+00 -2.92612612e-01 4.57608402e-01
-3.50597054e-01 -1.01683664e+00 -2.23212361e-01 -7.62583733e-01
6.39555931e-01 4.46393639e-02 1.50962353e+00 4.20076437e-02
-3.91036309e-02 -9.88520980e-02 -8.41094017e-01 -2.77299315e-01
-4.87227499e-01 4.20484692e-01 -1.83043122e-01 -1.25663862e-01
8.33133459e-01 -4.65756267e-01 1.42321456e-03 7.83437341e-02
-5.63296318e-01 5.92871964e-01 3.52656424e-01 7.84877896e-01
3.95342112e-01 -1.38158873e-01 4.64484483e-01 -1.47145665e+00
1.01340437e+00 -6.00132883e-01 -1.52986020e-01 6.46806359e-01
-5.13931274e-01 1.93645373e-01 4.31069881e-01 -5.34052491e-01
-1.17300498e+00 -1.08942956e-01 -1.26721365e-02 -2.70807073e-02
-1.24622323e-01 7.00179577e-01 -3.59578937e-01 6.49501801e-01
8.23787034e-01 -8.16066414e-02 -5.71198344e-01 -3.35582376e-01
6.38407052e-01 8.38372946e-01 5.37341952e-01 -1.02440643e+00
7.51111865e-01 -2.46463999e-01 -1.92377105e-01 -4.99916375e-01
-1.42396879e+00 -2.31496975e-01 -5.37438154e-01 1.32964537e-01
7.19035149e-01 -8.93114030e-01 -5.10278463e-01 -8.18378478e-02
-1.42567182e+00 -5.83446264e-01 -6.13804817e-01 2.83778071e-01
-2.89739460e-01 1.74092367e-01 -7.82007873e-01 -8.96762431e-01
-5.70626259e-01 -7.25371003e-01 1.10297072e+00 1.20786518e-01
-7.81985462e-01 -8.18336904e-01 -7.73781352e-03 4.58129197e-01
-2.90620357e-01 4.93803658e-02 9.58812773e-01 -1.06185758e+00
-5.04983068e-01 -7.65824914e-02 -2.25744098e-01 -1.89611152e-01
2.79064506e-01 -6.85336888e-02 -9.06683147e-01 3.52438271e-01
-3.73000145e-01 -7.24642038e-01 5.01524389e-01 -2.56917715e-01
7.22335875e-01 -6.00296915e-01 -6.27342165e-01 3.82609129e-01
7.18313575e-01 8.72158408e-02 3.82354051e-01 1.01883724e-01
9.05619383e-01 8.65516365e-01 1.00923312e+00 2.72140056e-01
8.19705963e-01 8.61406565e-01 -2.50175804e-01 7.17223734e-02
-3.73005778e-01 -5.91269851e-01 1.16867766e-01 8.12214434e-01
-1.46220148e-01 -5.94018221e-01 -9.64417398e-01 7.02535331e-01
-1.82724011e+00 -8.85985076e-01 -3.91179651e-01 1.67531025e+00
1.70889592e+00 1.42499968e-01 -1.42145067e-01 1.97127759e-01
6.06395781e-01 2.10202229e-03 -1.75394237e-01 -1.73549950e-01
-2.08757713e-01 3.60375106e-01 1.34670049e-01 7.25372314e-01
-8.99718285e-01 1.47341752e+00 5.11942768e+00 1.04999137e+00
-4.75588530e-01 5.93197905e-02 4.39983338e-01 1.61035672e-01
-6.11503005e-01 3.51412296e-01 -1.36909163e+00 9.69101787e-02
9.48802948e-01 -4.59079146e-01 1.55170009e-01 5.39525747e-01
-1.25896066e-01 1.21431477e-01 -1.60534549e+00 7.70984888e-01
-3.75927314e-02 -1.23002243e+00 1.98266760e-01 -1.95783556e-01
4.06836122e-01 -3.69664937e-01 -4.91945595e-01 6.56258643e-01
7.59848356e-01 -1.00364661e+00 3.13401848e-01 9.95106921e-02
9.62547958e-01 -2.44129807e-01 6.30816221e-01 4.64766830e-01
-1.23857188e+00 3.28879446e-01 -1.73987761e-01 -2.39672303e-01
4.71331358e-01 7.11411178e-01 -1.42107236e+00 1.07803750e+00
1.63102284e-01 7.84763515e-01 -5.39383650e-01 4.52672929e-01
-8.83027911e-01 5.89862764e-01 -3.81738424e-01 -2.15270042e-01
-1.77753568e-01 -2.67496966e-02 3.65083009e-01 1.27210462e+00
1.17362842e-01 6.05050147e-01 3.91076952e-02 8.92361164e-01
-1.94919363e-01 1.69103637e-01 -6.26916885e-01 -3.44080985e-01
8.95546377e-01 1.44462729e+00 -6.88224614e-01 -6.13720059e-01
-4.54839379e-01 8.50393951e-01 7.79183030e-01 4.86642957e-01
-4.76521671e-01 -5.15416980e-01 2.40995839e-01 -2.33477280e-01
2.17566952e-01 1.89338997e-01 -3.52966994e-01 -1.34956145e+00
2.51044691e-01 -9.21894372e-01 4.98070031e-01 -8.27657640e-01
-1.36575198e+00 9.10221577e-01 2.27898300e-01 -1.09272361e+00
-6.69392109e-01 -3.71179342e-01 -3.98187548e-01 9.72922862e-01
-1.20037663e+00 -1.20092511e+00 -2.27633759e-01 3.02026451e-01
5.13935566e-01 -4.52626217e-03 1.25069880e+00 2.97998041e-01
-9.06687498e-01 8.86677980e-01 -8.61598790e-01 5.89024842e-01
6.04912519e-01 -1.16206849e+00 9.40013707e-01 8.72899473e-01
5.56479931e-01 7.07895339e-01 7.89638460e-01 -1.01193237e+00
-1.09195673e+00 -1.08270299e+00 1.82208931e+00 -7.90201724e-01
8.34549844e-01 -6.63204968e-01 -9.34302449e-01 1.10447800e+00
2.03340784e-01 1.05552062e-01 9.46966469e-01 4.97721583e-01
-4.70118225e-01 2.54399806e-01 -7.39630163e-01 8.57163787e-01
1.38882983e+00 -5.37049055e-01 -9.59224284e-01 6.33752942e-01
1.16413450e+00 -7.86225677e-01 -1.11378658e+00 7.37351418e-01
1.65350184e-01 -2.58029997e-01 7.17716753e-01 -7.90230930e-01
7.58716404e-01 -1.08291321e-01 4.85346839e-02 -1.24211800e+00
-1.01685025e-01 -9.02883768e-01 -9.89450634e-01 1.73462546e+00
1.20483291e+00 -3.66320103e-01 6.82675362e-01 1.01781929e+00
-6.59996271e-02 -1.10998166e+00 -7.24914014e-01 -7.24266171e-01
-2.98243165e-01 -4.30274308e-01 8.42481971e-01 1.06910169e+00
6.71985090e-01 1.38113070e+00 -3.75069141e-01 4.34589460e-02
3.48039240e-01 2.68732607e-01 8.27241361e-01 -1.03652954e+00
-6.08198225e-01 7.33213052e-02 1.45747781e-01 -1.17531204e+00
3.65640998e-01 -1.19213045e+00 8.11428130e-02 -1.88563156e+00
2.51102507e-01 -6.29549325e-01 8.23889077e-02 1.03430343e+00
-6.71278596e-01 -1.49515597e-02 -6.89693391e-02 -1.60547122e-02
-5.61085463e-01 6.91369295e-01 1.42899215e+00 -4.83735874e-02
-1.52757049e-01 -1.71902683e-03 -9.42806900e-01 2.27299616e-01
5.08329332e-01 -5.59191763e-01 -6.95713282e-01 -3.21122646e-01
3.31752509e-01 3.81142169e-01 -6.71230778e-02 -3.31002235e-01
2.64315754e-01 -1.50436878e-01 -2.64188908e-02 -5.37441969e-01
3.86741817e-01 -5.26663482e-01 -2.01015677e-02 -1.56994224e-01
-6.21714354e-01 -3.96029562e-01 -2.01393828e-01 5.28131761e-02
-1.43553346e-01 -3.56921107e-01 -9.01383385e-02 6.78429566e-03
-3.43420357e-01 3.49697083e-01 -5.57882600e-02 5.54045439e-01
7.61003435e-01 -1.60760265e-02 -6.90270543e-01 -2.67610490e-01
-8.03316593e-01 4.59354937e-01 -8.62942487e-02 5.92190444e-01
5.64412117e-01 -1.29255235e+00 -8.96871269e-01 9.10298824e-02
5.64849317e-01 4.00178462e-01 -1.92090333e-01 5.89782655e-01
1.24524958e-01 3.66033018e-01 4.00012851e-01 6.34831712e-02
-1.32647264e+00 6.24618411e-01 7.52840787e-02 -6.45105183e-01
-9.12331700e-01 1.16752243e+00 -1.83702465e-02 -4.81204718e-01
2.79717356e-01 -3.19511831e-01 -4.11431342e-01 3.17383185e-02
5.81838846e-01 -1.14488848e-01 2.40598172e-01 -4.48585123e-01
-2.62804508e-01 -5.17594926e-02 -3.64023149e-01 -2.15465739e-01
1.14703441e+00 3.42107229e-02 -1.86983123e-01 6.33919597e-01
7.31341839e-01 -1.00164704e-01 -8.78805280e-01 -6.30621672e-01
6.56615376e-01 -1.65106490e-01 -5.08674800e-01 -7.56434202e-01
-4.93661910e-01 3.92536253e-01 -4.21171427e-01 3.04167539e-01
6.47909105e-01 4.39716458e-01 1.00158596e+00 5.69300711e-01
3.68900508e-01 -7.03622878e-01 -1.51697546e-01 9.21278119e-01
7.42457032e-01 -1.16101158e+00 -8.20461512e-02 -1.16532373e+00
-9.21567380e-01 6.64828479e-01 9.16029632e-01 3.50239754e-01
3.78863901e-01 7.11034477e-01 1.36662215e-01 -3.30055892e-01
-1.18285799e+00 -4.99294668e-01 2.56057322e-01 7.43206739e-01
9.66430783e-01 1.79743648e-01 -4.20721889e-01 9.40312564e-01
-6.86375439e-01 -8.68223757e-02 3.58981758e-01 8.24639380e-01
3.12521718e-02 -1.62732673e+00 1.18145138e-01 6.69756174e-01
-3.56293052e-01 -7.00845897e-01 -6.88469172e-01 3.78754139e-01
-1.10335879e-01 1.31578660e+00 -2.07502395e-01 -4.84599024e-01
3.62386793e-01 3.58940214e-01 6.12799883e-01 -1.32687759e+00
-6.12571180e-01 -2.56152481e-01 1.11887205e+00 -1.78990766e-01
-3.06511104e-01 -4.76224482e-01 -1.30868745e+00 1.07887283e-01
-5.10138929e-01 5.77567041e-01 -1.40554175e-01 1.26986217e+00
2.48360336e-01 8.03934574e-01 1.93159372e-01 -3.71917725e-01
-2.40688860e-01 -1.43918931e+00 -2.80632883e-01 3.95190567e-01
2.68038418e-02 -7.43063271e-01 -5.67975268e-03 7.58098215e-02] | [9.464255332946777, 8.513602256774902] |
344f3725-2d8b-40ea-b82d-e6396ace8823 | towards-more-transparent-and-accurate-cancer | 2305.11728 | null | https://arxiv.org/abs/2305.11728v1 | https://arxiv.org/pdf/2305.11728v1.pdf | Towards More Transparent and Accurate Cancer Diagnosis with an Unsupervised CAE Approach | Digital pathology has revolutionized cancer diagnosis by leveraging Content-Based Medical Image Retrieval (CBMIR) for analyzing histopathological Whole Slide Images (WSIs). CBMIR enables searching for similar content, enhancing diagnostic reliability and accuracy. In 2020, breast and prostate cancer constituted 11.7% and 14.1% of cases, respectively, as reported by the Global Cancer Observatory (GCO). The proposed Unsupervised CBMIR (UCBMIR) replicates the traditional cancer diagnosis workflow, offering a dependable method to support pathologists in WSI-based diagnostic conclusions. This approach alleviates pathologists' workload, potentially enhancing diagnostic efficiency. To address the challenge of the lack of labeled histopathological images in CBMIR, a customized unsupervised Convolutional Auto Encoder (CAE) was developed, extracting 200 features per image for the search engine component. UCBMIR was evaluated using widely-used numerical techniques in CBMIR, alongside visual evaluation and comparison with a classifier. The validation involved three distinct datasets, with an external evaluation demonstrating its effectiveness. UCBMIR outperformed previous studies, achieving a top 5 recall of 99% and 80% on BreaKHis and SICAPv2, respectively, using the first evaluation technique. Precision rates of 91% and 70% were achieved for BreaKHis and SICAPv2, respectively, using the second evaluation technique. Furthermore, UCBMIR demonstrated the capability to identify various patterns in patches, achieving an 81% accuracy in the top 5 when tested on an external image from Arvaniti. | ['Valery Naranjo', 'Javier Oliver Moll', 'Adrian colomer', 'Zahra Tabatabaei'] | 2023-05-19 | null | null | null | null | ['whole-slide-images', 'medical-image-retrieval', 'medical-image-retrieval'] | ['computer-vision', 'computer-vision', 'medical'] | [ 1.57493487e-01 9.55968350e-02 -2.24793464e-01 3.14945877e-01
-1.43294764e+00 -5.83791137e-01 5.42553067e-01 8.45791221e-01
-7.20824301e-01 3.10631603e-01 1.76790711e-02 -3.55162621e-01
-3.54210883e-01 -8.85597348e-01 -2.99907416e-01 -1.11878932e+00
-3.36337239e-02 2.63395220e-01 1.29702061e-01 4.80370708e-02
1.93971112e-01 7.53969729e-01 -1.42210054e+00 5.34369111e-01
6.59822345e-01 1.16867721e+00 4.36740309e-01 1.07396102e+00
-3.59374940e-01 7.64607131e-01 -7.61287034e-01 -5.54572463e-01
-5.11938483e-02 5.38830645e-02 -8.49876165e-01 -2.39415213e-01
4.07486945e-01 -1.19181283e-01 -4.49566215e-01 8.73035669e-01
4.68561649e-01 -3.75788093e-01 9.03935969e-01 -7.81343043e-01
-6.28958046e-01 3.32464337e-01 -5.62824190e-01 2.69242615e-01
2.39026681e-01 2.68059343e-01 1.04506052e+00 -6.44800186e-01
1.19476962e+00 6.25724494e-01 7.41766334e-01 2.54734844e-01
-1.04003847e+00 -5.62993050e-01 -6.79095387e-01 3.26547652e-01
-1.60240555e+00 6.23093871e-03 2.47506052e-01 -5.05353630e-01
8.44012856e-01 7.63754010e-01 6.66372836e-01 5.28319955e-01
5.26920795e-01 6.56066597e-01 1.07972109e+00 -5.62984169e-01
1.89969331e-01 2.53067851e-01 -8.04390386e-02 7.09591687e-01
4.34806108e-01 -2.13882834e-01 -1.93725705e-01 -1.00475192e-01
4.48213875e-01 2.54857510e-01 -2.75689363e-01 2.17561677e-01
-1.20383286e+00 6.74219549e-01 7.26203740e-01 8.56093824e-01
-7.60203823e-02 -7.01166466e-02 8.29211652e-01 1.94020227e-01
2.62145221e-01 4.77175415e-01 1.59630403e-01 1.63899854e-01
-9.91583645e-01 -1.77574024e-01 3.79673153e-01 3.69466513e-01
2.07843423e-01 -5.27649820e-01 -4.69924539e-01 7.29325652e-01
7.37346038e-02 7.99352527e-01 9.36275959e-01 -4.54290748e-01
-1.95150763e-01 9.87674713e-01 -3.72610331e-01 -1.16954839e+00
-7.68699467e-01 -6.64182305e-01 -1.25062251e+00 1.07077152e-01
3.57355744e-01 6.74924672e-01 -8.93592358e-01 1.05152583e+00
1.90576971e-01 -1.81770146e-01 1.87336951e-01 8.66713345e-01
1.03649938e+00 3.93267483e-01 2.38133833e-01 9.34467986e-02
1.81586325e+00 -3.85089606e-01 -5.55836916e-01 5.31769812e-01
9.38384414e-01 -7.68726647e-01 9.28119302e-01 3.98225427e-01
-7.96491146e-01 -2.27231994e-01 -1.05814934e+00 -1.32667989e-01
-7.08600104e-01 6.44328058e-01 4.82751191e-01 4.60554808e-01
-1.46281695e+00 3.01091313e-01 -7.42554963e-01 -6.41479373e-01
6.68886900e-01 3.00015152e-01 -8.14304411e-01 -2.15557575e-01
-8.44757497e-01 6.88703656e-01 1.30121976e-01 -1.65734783e-01
-8.38639617e-01 -1.10320020e+00 -6.93210483e-01 1.72679454e-01
-1.45865068e-01 -3.74654502e-01 9.08748925e-01 -5.53754866e-01
-1.19692874e+00 1.34345818e+00 1.76819205e-01 -6.10827982e-01
5.69382548e-01 4.95647013e-01 -6.13631845e-01 8.94258678e-01
1.52069062e-01 6.60299301e-01 1.62633941e-01 -7.86771715e-01
-6.74214661e-01 -2.80715346e-01 -2.25922778e-01 -5.70457801e-02
-6.77446067e-01 -4.24494892e-01 -7.23793030e-01 -5.55470705e-01
-1.84095025e-01 -7.06988573e-01 -1.56068459e-01 3.20823878e-01
-4.35305983e-01 -5.54939173e-02 6.20282948e-01 -8.78948152e-01
1.05099106e+00 -2.32699132e+00 -1.79354399e-01 5.47442079e-01
4.55036908e-01 3.66544127e-01 -1.59910232e-01 4.87171024e-01
1.58747118e-02 2.84291744e-01 6.04266003e-02 -1.52216777e-01
-6.89003542e-02 -2.27416605e-01 2.07413211e-01 8.35862815e-01
2.28933454e-01 1.01530421e+00 -9.32003081e-01 -7.96553850e-01
3.51243734e-01 6.54905438e-01 -2.43343607e-01 -2.61117257e-02
2.94802696e-01 1.04804665e-01 4.63369563e-02 9.55220401e-01
6.17300272e-01 -7.20870018e-01 4.17883188e-01 -3.36500227e-01
5.67733496e-02 -2.92257398e-01 -7.03093767e-01 1.37840962e+00
-4.88980502e-01 1.02400672e+00 -1.92915872e-02 -6.98211670e-01
7.09734142e-01 3.01899433e-01 7.88373530e-01 -1.12751675e+00
1.40669718e-01 3.30344349e-01 -1.63459256e-01 -7.19422281e-01
2.83698231e-01 1.07722096e-01 8.04932266e-02 7.83881471e-02
-2.16397941e-02 2.09759966e-01 3.75511736e-01 5.32052517e-01
1.46316671e+00 -6.60326600e-01 4.72227186e-01 -4.07406777e-01
9.02325392e-01 4.15143758e-01 -4.11193594e-02 5.87803364e-01
-2.60456234e-01 4.50827926e-01 4.44180071e-01 -2.72509634e-01
-7.43274868e-01 -1.02278590e+00 -5.55080771e-01 4.67307389e-01
7.35730380e-02 -2.05210939e-01 -5.31659842e-01 -4.68539119e-01
6.12762421e-02 1.59395561e-01 -8.15483868e-01 2.18071118e-02
-1.75193265e-01 -7.70806611e-01 7.73890972e-01 2.35617936e-01
3.79159719e-01 -7.62926817e-01 -8.22645545e-01 -3.39461081e-02
-1.37143195e-01 -8.55421841e-01 -1.59844860e-01 4.67860736e-02
-5.53562164e-01 -1.56086361e+00 -1.07524657e+00 -6.77765369e-01
9.83873665e-01 3.22508126e-01 8.35909903e-01 3.64536852e-01
-1.34366107e+00 3.33818793e-01 -3.08368504e-01 -2.45990694e-01
-7.28928030e-01 1.40664252e-02 -3.98674935e-01 -2.60196120e-01
3.38857830e-01 8.46589282e-02 -1.03030419e+00 -3.92059982e-02
-1.27177465e+00 1.35898933e-01 1.15180540e+00 1.00972235e+00
8.88450801e-01 -1.00758955e-01 4.20211434e-01 -7.55609274e-01
1.31249055e-01 -3.30126375e-01 -5.24123669e-01 2.05746129e-01
-5.97767353e-01 -4.48488086e-01 5.93960047e-01 -1.43567786e-01
-6.30859077e-01 -1.11092888e-01 -2.94859618e-01 -3.18090022e-01
1.40154902e-02 6.60831869e-01 3.78563702e-01 -1.88747779e-01
7.26119936e-01 3.86284769e-01 4.47524935e-01 -1.44650534e-01
-1.35569796e-01 8.93226743e-01 6.16619825e-01 3.41371477e-01
4.99735445e-01 8.51557493e-01 1.32276729e-01 -7.36773074e-01
-2.73777306e-01 -1.18295968e+00 -1.89513117e-01 -3.79677862e-01
8.21331263e-01 -8.78448963e-01 -1.09725153e+00 3.39362025e-01
-5.50692439e-01 -1.20819584e-01 -2.27880985e-01 3.57832491e-01
-4.62990738e-02 3.71822208e-01 -9.57744241e-01 -5.08044720e-01
-9.04314816e-01 -1.09178936e+00 1.36739075e+00 2.42508039e-01
-2.22677886e-01 -9.62522149e-01 1.02672428e-01 2.59927452e-01
6.04384065e-01 4.50397342e-01 9.64701355e-01 -7.62967348e-01
-1.57683939e-01 -7.18366146e-01 -4.92795110e-01 -1.54880032e-01
3.00008774e-01 2.91478366e-01 -1.11410904e+00 -5.12852788e-01
-6.68099463e-01 -7.37926438e-02 7.65064538e-01 2.52493083e-01
1.21900403e+00 -2.62238175e-01 -7.66451061e-01 4.45540160e-01
2.02643919e+00 4.90607768e-02 5.84888935e-01 6.39132679e-01
1.80232048e-01 4.20161545e-01 4.51941341e-01 2.32726485e-01
6.03330322e-02 5.33213019e-01 5.29333353e-01 -5.20237684e-01
-5.35871744e-01 4.54455502e-02 -7.64128044e-02 6.17762268e-01
2.39585459e-01 -8.51050466e-02 -1.25136423e+00 8.42253566e-01
-1.06074691e+00 -6.86118126e-01 -2.31211469e-01 1.94846988e+00
8.05343390e-01 -1.16258420e-01 -2.97398686e-01 3.72400999e-01
5.48871934e-01 -3.50674152e-01 -2.94922531e-01 -1.13862455e-01
-1.18751638e-01 2.42636874e-01 7.11697578e-01 6.54080138e-02
-1.14831936e+00 3.54409158e-01 5.78889465e+00 1.13130295e+00
-1.58239770e+00 9.74502042e-02 1.03214025e+00 -1.14681259e-01
-1.51993200e-01 -7.16872275e-01 -5.25955498e-01 3.48234892e-01
8.66718650e-01 -1.92224726e-01 -1.82575211e-01 6.94594800e-01
1.68531924e-01 -3.35905075e-01 -6.74805462e-01 9.65480804e-01
1.18640028e-01 -1.96720588e+00 1.79219067e-01 3.09909105e-01
5.84935784e-01 4.56222929e-02 2.52037555e-01 5.37310652e-02
-4.76075076e-02 -1.10318518e+00 7.33939484e-02 5.62794685e-01
1.55657911e+00 -7.65422702e-01 1.30710638e+00 -9.81220305e-02
-1.12453258e+00 -3.90732400e-02 -1.90779224e-01 7.05697775e-01
-4.75403428e-01 6.85927331e-01 -1.32284737e+00 8.21845710e-01
7.75729477e-01 4.86638904e-01 -9.31272924e-01 1.09125292e+00
3.80020410e-01 4.14560199e-01 -9.36814547e-02 -6.30326942e-02
3.14949542e-01 3.72421056e-01 3.78923178e-01 1.75013578e+00
4.03978735e-01 -1.73658580e-01 -2.92629987e-01 3.11092317e-01
-7.38175958e-02 5.57463527e-01 -1.28706530e-01 -5.96359782e-02
2.78661877e-01 1.86304808e+00 -1.06860542e+00 -4.23263103e-01
-7.30005056e-02 6.94596291e-01 2.01200079e-02 9.20522362e-02
-6.39586270e-01 -5.97082138e-01 2.69172460e-01 2.22043499e-01
1.04392812e-01 4.71216828e-01 -1.05264731e-01 -6.23370230e-01
-4.28587139e-01 -8.71282279e-01 7.31829941e-01 -5.26097894e-01
-1.21162796e+00 7.10182071e-01 -4.66101259e-01 -1.56306303e+00
3.16823907e-02 -9.25351858e-01 -4.29022849e-01 6.51419997e-01
-1.73052609e+00 -1.28557217e+00 -5.52128315e-01 2.44926557e-01
2.12426379e-01 -3.65247615e-02 1.09610307e+00 3.57122093e-01
-5.10177851e-01 9.43945765e-01 5.98189592e-01 3.02059054e-01
8.07900727e-01 -1.38495779e+00 -4.13834661e-01 3.72891158e-01
-2.60625899e-01 4.36376631e-01 3.87125552e-01 -2.89008141e-01
-1.62155652e+00 -1.25874341e+00 6.69317484e-01 4.19730470e-02
6.43452644e-01 2.64663482e-03 -8.02972853e-01 -3.52235176e-02
8.66386220e-02 -1.67105850e-02 1.28224564e+00 -5.53589463e-01
-3.65941614e-01 -1.49111390e-01 -1.52749550e+00 6.50432944e-01
3.79134774e-01 -4.88902003e-01 1.89902753e-01 4.56060290e-01
2.55602419e-01 -4.07722950e-01 -1.32202339e+00 3.54616016e-01
6.51203632e-01 -6.20164692e-01 9.56963539e-01 4.80465628e-02
6.92695081e-01 -2.91562974e-01 1.74018070e-02 -8.90270352e-01
-3.27948749e-01 1.66048735e-01 1.48789391e-01 9.42695081e-01
2.54322469e-01 -5.51044405e-01 7.25252330e-01 -4.15268205e-02
9.35706310e-03 -1.03342426e+00 -1.05214536e+00 -4.87368375e-01
-7.02582300e-03 -1.51620656e-01 3.17938298e-01 6.86550796e-01
2.29585454e-01 -5.32274663e-01 5.50206423e-01 1.05664782e-01
3.08960438e-01 3.73488478e-02 4.29187924e-01 -9.78026569e-01
-2.03318402e-01 -1.01546621e+00 -1.02443266e+00 3.87878652e-04
-2.68320620e-01 -1.37000430e+00 -3.24995220e-01 -1.62637043e+00
6.32017791e-01 -3.47908229e-01 -5.55454850e-01 5.16660035e-01
-8.17893073e-02 1.10838664e+00 5.11189960e-02 5.19952834e-01
-4.95665044e-01 -1.93876922e-01 1.12392271e+00 -7.69710958e-01
2.67539501e-01 -6.35857582e-01 -6.39415562e-01 3.90030205e-01
6.01460338e-01 -1.32894173e-01 7.79322013e-02 1.20242029e-01
1.82131305e-01 -1.38299704e-01 4.41496909e-01 -1.24444795e+00
3.45532089e-01 1.66536808e-01 7.86704242e-01 -7.16789126e-01
-3.83628942e-02 -7.26402283e-01 4.34911549e-01 1.06789267e+00
-3.36914092e-01 -2.57265002e-01 5.17098606e-01 5.73464334e-01
-4.89316285e-01 -9.18940734e-03 7.35492885e-01 2.15810299e-01
-6.25442803e-01 8.38435888e-02 -6.66052043e-01 -6.35871172e-01
1.27879310e+00 -3.46818984e-01 -6.91205800e-01 1.09616565e-02
-4.56007719e-01 1.20720446e-01 5.97990453e-01 -1.48131903e-02
5.74080408e-01 -9.94677484e-01 -8.09692025e-01 -2.84427870e-02
6.71520591e-01 -6.51221573e-02 7.80586421e-01 1.22435915e+00
-1.18935132e+00 5.74194729e-01 -1.39827266e-01 -8.59523833e-01
-1.43440330e+00 5.36979914e-01 4.14364964e-01 -9.00187194e-01
-8.53388250e-01 7.23193407e-01 1.14035286e-01 -9.10366848e-02
1.83261156e-01 -1.36021629e-01 -5.53832293e-01 1.71653599e-01
8.95897686e-01 2.88777143e-01 5.37603676e-01 -4.78975803e-01
-3.94207031e-01 4.37073946e-01 -4.39812034e-01 1.33094475e-01
1.08495021e+00 1.49978548e-01 -1.42524749e-01 2.01416537e-01
1.55191016e+00 1.77014545e-01 -5.71196258e-01 1.77566032e-03
6.36937693e-02 -2.46513098e-01 3.81649844e-02 -1.17128873e+00
-1.10969949e+00 5.81691265e-01 1.07150984e+00 1.08119376e-01
1.38093507e+00 8.35427828e-03 6.73294842e-01 1.27350166e-01
1.68075383e-01 -6.80369675e-01 -4.85509224e-02 -8.56086537e-02
6.64548337e-01 -1.19310641e+00 4.87392135e-02 -3.60876083e-01
-3.09377700e-01 1.33020115e+00 1.28005937e-01 5.61935194e-02
4.04755712e-01 4.79299456e-01 3.17183912e-01 -4.48330909e-01
-6.44697309e-01 1.59708306e-01 2.85025686e-01 3.24245155e-01
4.60554600e-01 2.47273296e-01 -4.81612533e-01 4.17981207e-01
3.19353230e-02 2.58916989e-02 2.98166066e-01 8.54824960e-01
-1.19381763e-01 -5.64651430e-01 -4.17580366e-01 8.02711248e-01
-8.70950758e-01 4.84776162e-02 -3.00661176e-01 9.99844074e-01
-1.14763729e-01 6.65789366e-01 3.47889096e-01 -1.51753590e-01
-5.42115457e-02 -2.95636863e-01 -1.59969568e-01 -2.28619978e-01
-8.74568045e-01 1.34780347e-01 -3.00701678e-01 -4.28749293e-01
-1.94216743e-01 -3.44978064e-01 -1.30489028e+00 -2.33803704e-01
-2.48160064e-01 2.12577447e-01 8.56682122e-01 5.30538380e-01
4.34324801e-01 9.24772680e-01 4.77038115e-01 -3.01676869e-01
-1.84825867e-01 -8.71014655e-01 -5.40165126e-01 3.83711249e-01
3.39762419e-01 -2.87861526e-01 -2.77822345e-01 1.58655763e-01] | [15.137126922607422, -2.9589176177978516] |
9a15dee4-b671-4d0b-98aa-8ec6028021a9 | improving-task-adaptation-for-cross-domain | 2107.00358 | null | https://arxiv.org/abs/2107.00358v4 | https://arxiv.org/pdf/2107.00358v4.pdf | Cross-domain Few-shot Learning with Task-specific Adapters | In this paper, we look at the problem of cross-domain few-shot classification that aims to learn a classifier from previously unseen classes and domains with few labeled samples. Recent approaches broadly solve this problem by parameterizing their few-shot classifiers with task-agnostic and task-specific weights where the former is typically learned on a large training set and the latter is dynamically predicted through an auxiliary network conditioned on a small support set. In this work, we focus on the estimation of the latter, and propose to learn task-specific weights from scratch directly on a small support set, in contrast to dynamically estimating them. In particular, through systematic analysis, we show that task-specific weights through parametric adapters in matrix form with residual connections to multiple intermediate layers of a backbone network significantly improves the performance of the state-of-the-art models in the Meta-Dataset benchmark with minor additional cost. | ['Hakan Bilen', 'Xialei Liu', 'Wei-Hong Li'] | 2021-07-01 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Li_Cross-Domain_Few-Shot_Learning_With_Task-Specific_Adapters_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Li_Cross-Domain_Few-Shot_Learning_With_Task-Specific_Adapters_CVPR_2022_paper.pdf | cvpr-2022-1 | ['cross-domain-few-shot', 'cross-domain-few-shot-learning'] | ['computer-vision', 'computer-vision'] | [ 4.99579132e-01 6.04835264e-02 -3.87557328e-01 -5.23770630e-01
-7.79215157e-01 -4.51569945e-01 7.44882703e-01 2.28644330e-02
-5.83074212e-01 6.63437843e-01 -1.45111814e-01 1.69599831e-01
-1.81560501e-01 -5.93539476e-01 -7.62765944e-01 -4.02505606e-01
1.25893205e-01 9.11136627e-01 4.79363859e-01 -3.22942197e-01
1.02414057e-01 1.11102171e-01 -1.69168174e+00 3.31276953e-01
7.84611583e-01 1.24575901e+00 7.38620535e-02 4.90948409e-01
-9.53665376e-02 7.74001002e-01 -4.42269504e-01 -4.84033376e-01
3.88904870e-01 -1.68428868e-01 -7.99860239e-01 2.44843096e-01
4.67606932e-01 -9.38303769e-02 -1.06260747e-01 9.90437984e-01
3.36863816e-01 4.64892775e-01 9.16599393e-01 -1.38101566e+00
-4.44463104e-01 4.27001774e-01 -2.96903342e-01 2.06253514e-01
-2.37977117e-01 3.58240813e-01 1.05390251e+00 -9.09098864e-01
7.97722876e-01 7.94164777e-01 7.65622139e-01 7.51159847e-01
-1.71186340e+00 -6.16106153e-01 2.88527012e-01 3.51450950e-01
-1.36341047e+00 -5.90750456e-01 9.10312533e-01 -6.73973024e-01
1.04790807e+00 -2.12520510e-01 2.62127817e-01 1.45077002e+00
-1.28571823e-01 5.83622336e-01 7.93936670e-01 -4.82862055e-01
7.09207475e-01 3.51448029e-01 5.23999989e-01 5.30888379e-01
1.67145938e-01 4.57848683e-02 -3.36964130e-01 -3.49860728e-01
2.45632574e-01 2.21541151e-01 -3.31917331e-02 -1.04625118e+00
-1.00862932e+00 1.08232188e+00 1.99189901e-01 2.32904285e-01
-1.76028311e-01 3.52558345e-02 6.51211619e-01 6.08127117e-01
8.79936755e-01 5.10992289e-01 -7.29496002e-01 -1.80746615e-02
-9.32270348e-01 -3.32510024e-02 9.38921630e-01 1.07809365e+00
9.43786263e-01 -7.76586635e-03 -3.22362542e-01 1.17965019e+00
-2.49510631e-01 6.75366744e-02 6.44533515e-01 -5.60424984e-01
5.87776661e-01 4.98770028e-01 1.25044316e-01 -2.67573297e-01
-3.28391999e-01 -4.80017334e-01 -6.19353592e-01 6.31095096e-02
4.90721494e-01 -4.77873176e-01 -1.22201777e+00 1.97182596e+00
2.43080467e-01 6.81368828e-01 1.45611111e-02 5.88870764e-01
4.62917387e-01 4.75666642e-01 2.61443228e-01 -8.87247548e-03
1.19159639e+00 -1.27450943e+00 -2.35354021e-01 -8.03982079e-01
6.41493678e-01 -2.02014863e-01 1.26863623e+00 2.35525310e-01
-7.37833679e-01 -6.96048677e-01 -1.27907443e+00 1.31959319e-02
-7.15281308e-01 1.32348478e-01 3.59321088e-01 5.18233657e-01
-7.24483490e-01 8.00821126e-01 -5.27797103e-01 -4.84708518e-01
6.13195360e-01 2.43367225e-01 -2.49124929e-01 -2.84183204e-01
-1.25876486e+00 9.74825919e-01 4.93432045e-01 -4.81790513e-01
-1.07080793e+00 -9.45863008e-01 -9.25709367e-01 3.92414689e-01
6.17265522e-01 -6.99532688e-01 1.40200353e+00 -9.43655074e-01
-1.59704423e+00 1.01739657e+00 1.33190110e-01 -6.31965160e-01
5.17062008e-01 -3.37621570e-02 -2.90142417e-01 -1.53600788e-02
1.00696743e-01 4.12122965e-01 1.20120966e+00 -8.60550582e-01
-6.11467719e-01 -3.51554066e-01 1.49263278e-01 -1.75535455e-01
-6.84681058e-01 -1.34073004e-01 -2.31357738e-01 -5.43706000e-01
-4.50178623e-01 -8.51455092e-01 -3.80169004e-01 2.22159386e-01
2.94194333e-02 -1.90514013e-01 6.37760043e-01 -8.28070268e-02
8.39462876e-01 -2.16640472e+00 3.17960382e-01 -1.31173536e-01
8.34255144e-02 4.60627913e-01 -4.65035290e-01 2.80067503e-01
-2.30952963e-01 -3.71646225e-01 -4.46902364e-01 -4.86884117e-01
3.41346897e-02 2.08596110e-01 -4.42100614e-01 4.85971957e-01
4.85889286e-01 8.37270916e-01 -1.07892001e+00 -2.44296312e-01
1.61244497e-01 2.11360499e-01 -5.35874248e-01 3.62343371e-01
-4.53401029e-01 -2.24484671e-02 -2.84320176e-01 3.78929526e-01
4.90126073e-01 -5.18949926e-01 3.39125156e-01 5.21659106e-02
3.12195748e-01 1.15015425e-01 -1.22254133e+00 1.95842743e+00
-8.67151141e-01 3.18080515e-01 3.03092841e-02 -1.62482154e+00
8.55199277e-01 3.07446420e-01 3.35751295e-01 -3.51436347e-01
2.57376969e-01 1.31325230e-01 -1.05412178e-01 -2.16939569e-01
1.36715293e-01 -4.91986513e-01 -2.47702509e-01 5.71928501e-01
7.47545898e-01 1.32287862e-02 2.74955213e-01 -2.39685248e-03
1.19159293e+00 -4.21493091e-02 5.67282438e-01 -1.73880771e-01
4.09092009e-01 9.20879617e-02 4.86456454e-01 8.13571036e-01
-1.19599015e-01 7.31682301e-01 3.68192315e-01 -4.20684397e-01
-1.13776004e+00 -9.83303308e-01 -2.15638995e-01 1.57077444e+00
-1.22366898e-01 -2.24959373e-01 -6.16215348e-01 -1.04646790e+00
2.35327020e-01 9.50464129e-01 -9.06292796e-01 -5.30347049e-01
-3.50921184e-01 -5.97120881e-01 6.10928237e-02 6.23525023e-01
1.77253082e-01 -8.41887236e-01 -6.79983675e-01 4.58919168e-01
2.01507881e-01 -1.15040767e+00 -4.03321564e-01 7.93163002e-01
-8.53418887e-01 -1.12122238e+00 -9.98621643e-01 -6.37103558e-01
5.46360373e-01 4.12746906e-01 1.22407758e+00 -4.09801960e-01
-3.95590276e-01 3.64188969e-01 -2.46420145e-01 -3.96342635e-01
-1.81486458e-01 5.53615928e-01 7.83554986e-02 2.57719070e-01
5.95523000e-01 -8.61010432e-01 -1.34910956e-01 3.11193943e-01
-8.65942538e-01 -1.23225167e-01 2.89419591e-01 1.23505354e+00
3.37724835e-01 -3.49836588e-01 6.47200167e-01 -1.39280558e+00
4.48560506e-01 -8.60404789e-01 -6.46968007e-01 4.22588646e-01
-5.30027807e-01 2.80760318e-01 1.04561198e+00 -9.60179746e-01
-1.01074588e+00 1.89045385e-01 2.69561142e-01 -7.64396966e-01
-2.12715983e-01 2.20863760e-01 1.86338108e-02 -5.80462515e-02
1.00645006e+00 1.58235595e-01 -1.45547748e-01 -6.57803118e-01
5.15272796e-01 5.90395451e-01 2.38318726e-01 -7.04003334e-01
8.12407672e-01 4.58432019e-01 -4.90600243e-02 -5.96935272e-01
-1.42176795e+00 -7.25384295e-01 -8.92821848e-01 9.34690461e-02
5.56401730e-01 -1.06719553e+00 -9.96571109e-02 2.26412579e-01
-9.45394695e-01 -6.91178679e-01 -7.74155736e-01 2.67236799e-01
-7.68443882e-01 -3.87678705e-02 -4.08066660e-01 -4.17789191e-01
-2.54460990e-01 -7.77472734e-01 9.98778939e-01 -9.96249467e-02
-3.20051283e-01 -1.03693759e+00 2.76589572e-01 4.34926674e-02
5.62023699e-01 -2.05533039e-02 9.82240856e-01 -1.13506687e+00
-9.89912897e-02 -4.24724638e-01 -2.50605732e-01 4.91045266e-01
4.48117591e-02 -5.14464200e-01 -1.22407818e+00 -3.83927196e-01
4.96891104e-02 -7.01939046e-01 1.13158882e+00 1.93660676e-01
1.11886263e+00 -1.07209804e-02 -3.95816147e-01 8.31158340e-01
1.51581514e+00 -1.26390994e-01 1.85625821e-01 1.73653305e-01
3.79348576e-01 5.45925260e-01 6.82322443e-01 5.39092958e-01
6.82837144e-02 6.71951532e-01 2.19994128e-01 2.71851629e-01
-6.26092851e-02 -6.45730719e-02 1.08910367e-01 3.48937392e-01
1.08848080e-01 -6.17864132e-02 -8.98655474e-01 7.19855785e-01
-2.04060864e+00 -9.27756310e-01 5.19233823e-01 2.08448339e+00
8.09950709e-01 4.71307814e-01 7.84266219e-02 -1.71378419e-01
8.12516034e-01 3.63482773e-01 -8.85723352e-01 -9.79411826e-02
1.92882016e-01 5.13773024e-01 3.43960434e-01 1.88626021e-01
-1.27373600e+00 9.92892802e-01 6.34505415e+00 9.37252700e-01
-1.14899325e+00 4.28716272e-01 4.77880597e-01 -4.94033694e-01
9.75393057e-02 4.55011874e-02 -9.61297512e-01 4.78415906e-01
1.13877678e+00 -3.60058635e-01 4.25777704e-01 1.39305460e+00
-4.86784965e-01 2.26351544e-01 -1.53175974e+00 7.48674989e-01
8.10213163e-02 -1.41624475e+00 -6.87220246e-02 -8.76407549e-02
8.81032825e-01 3.46457154e-01 7.16025233e-02 8.89785826e-01
4.44666624e-01 -7.06388474e-01 6.21553242e-01 4.16916817e-01
9.71374929e-01 -5.98498583e-01 4.37432617e-01 5.93192279e-01
-1.00043881e+00 -4.64254439e-01 -6.53144777e-01 -1.70680538e-01
-1.20573327e-01 3.93415928e-01 -6.92670166e-01 2.00713843e-01
3.04595858e-01 8.48333478e-01 -4.84283775e-01 9.89006817e-01
-6.34972230e-02 6.31353676e-01 -8.75212699e-02 7.41851702e-02
3.42797488e-01 1.01173623e-02 4.15464133e-01 1.23426342e+00
1.77017629e-01 -1.11768572e-02 3.04519773e-01 8.48839283e-01
-2.82355130e-01 -1.17970407e-01 -6.51628435e-01 7.54102245e-02
2.58713633e-01 1.32035124e+00 -5.04868209e-01 -6.82217002e-01
-6.19663715e-01 8.87603581e-01 8.72668087e-01 2.87185252e-01
-8.23962629e-01 -5.22735775e-01 7.93442428e-01 2.39649579e-01
7.30052352e-01 3.94381620e-02 -1.60088435e-01 -1.56130493e+00
-1.60858527e-01 -6.12638891e-01 6.22975528e-01 -4.60441053e-01
-1.73237491e+00 4.61901546e-01 4.03580768e-03 -1.41437852e+00
-4.05571282e-01 -7.51118481e-01 -8.28922629e-01 7.91836858e-01
-1.62433207e+00 -9.62092400e-01 -1.95550621e-01 6.01908267e-01
8.26749444e-01 -3.77819747e-01 9.18865263e-01 1.98722795e-01
-5.75568974e-01 6.37663722e-01 4.07467246e-01 1.44484499e-02
9.37364101e-01 -1.10448360e+00 6.26880229e-01 5.40567219e-01
6.39119297e-02 3.99809539e-01 7.00633645e-01 -3.06553334e-01
-1.04264486e+00 -1.19922626e+00 6.83903754e-01 -4.62456077e-01
9.77934539e-01 -7.70593882e-01 -1.07474768e+00 7.08554208e-01
-8.74760225e-02 6.56647563e-01 6.01908267e-01 4.26894963e-01
-7.89595962e-01 -2.43325621e-01 -1.20453918e+00 2.36236677e-01
1.15625763e+00 -5.59538662e-01 -8.17000926e-01 4.31371808e-01
6.21086299e-01 -1.54594127e-02 -6.94095075e-01 1.62064686e-01
4.40003604e-01 -5.35515964e-01 1.03206408e+00 -1.27213931e+00
4.22795951e-01 1.48067147e-01 -2.28210330e-01 -1.59207535e+00
-4.33721960e-01 -3.10376078e-01 -3.16513270e-01 1.01482284e+00
4.54824090e-01 -5.20812571e-01 1.02206850e+00 6.51351213e-01
-4.76245396e-02 -5.13944924e-01 -1.00344503e+00 -1.11322451e+00
1.08409002e-01 -2.56481588e-01 3.98883224e-01 1.13974130e+00
-1.44994901e-02 8.25686276e-01 -6.03947818e-01 -2.52198815e-01
7.80772567e-01 2.57450461e-01 7.20382750e-01 -1.73046911e+00
-4.57008898e-01 -1.48662373e-01 -2.28747174e-01 -7.60301769e-01
4.97073412e-01 -1.02282238e+00 1.21272407e-01 -1.06730759e+00
4.41220224e-01 -2.58049130e-01 -4.98462707e-01 5.17977417e-01
-1.51920050e-01 5.07975407e-02 1.41056597e-01 5.88489696e-02
-8.48937333e-01 6.98655367e-01 8.27378929e-01 -4.01547760e-01
-9.25046355e-02 2.47971088e-01 -5.75170815e-01 6.50306702e-01
6.89705014e-01 -7.78506398e-01 -5.71108282e-01 -2.62212247e-01
-2.04903483e-02 -7.92220980e-03 3.48708153e-01 -1.12871099e+00
3.35964948e-01 -2.54861087e-01 2.28774145e-01 -4.76447567e-02
6.91593051e-01 -8.04958999e-01 -2.85186678e-01 3.74290675e-01
-5.23098707e-01 -5.72326303e-01 1.18284322e-01 9.00363505e-01
-8.03851485e-02 -6.44764245e-01 1.18197286e+00 -2.06210256e-01
-9.37464297e-01 4.03570950e-01 -2.01447532e-01 5.44667184e-01
1.22273183e+00 -1.14884734e-01 -3.33234072e-01 -1.33539438e-01
-8.61490071e-01 1.10173911e-01 3.96137327e-01 4.20865864e-01
3.07753712e-01 -1.30912411e+00 -5.33660889e-01 1.80213928e-01
6.88495994e-01 -2.88070202e-01 2.17622995e-01 4.65834916e-01
2.38116026e-01 3.03434491e-01 -4.02662307e-01 -3.98280531e-01
-9.66965556e-01 8.97117555e-01 2.69101679e-01 -4.63516653e-01
-4.04958755e-01 9.23501790e-01 1.85813785e-01 -6.19085491e-01
2.98578620e-01 -3.76936868e-02 -8.72880891e-02 3.76699179e-01
4.75466460e-01 3.77280802e-01 2.48482339e-02 -3.14439982e-01
-2.55928576e-01 3.98944408e-01 -2.72145540e-01 8.14650804e-02
1.68489802e+00 2.21548691e-01 4.08443063e-01 6.91918790e-01
1.41279411e+00 -6.32972419e-01 -1.57962251e+00 -8.68656337e-01
2.02732563e-01 -3.81234348e-01 2.20139213e-02 -7.17085958e-01
-9.67181206e-01 1.07894492e+00 4.62855995e-01 2.12637410e-02
8.16790640e-01 7.17728585e-02 6.38442218e-01 7.32059360e-01
3.83327961e-01 -1.48134363e+00 3.91206384e-01 7.89084494e-01
4.09299016e-01 -1.49007046e+00 -1.93794370e-01 -1.32188246e-01
-7.31559038e-01 1.07632816e+00 6.91725373e-01 -4.17078584e-01
8.35365653e-01 3.41648459e-02 -3.68625998e-01 -7.90024400e-02
-1.24315464e+00 -3.19432080e-01 3.72714698e-01 7.93007433e-01
1.64898410e-01 -1.64408073e-01 9.86944437e-02 7.90368259e-01
3.35958064e-01 2.50034422e-01 2.90896297e-01 1.00707495e+00
-6.67173028e-01 -1.06329596e+00 8.64039585e-02 6.84563577e-01
-2.26182237e-01 -2.40230501e-01 -1.65258661e-01 7.04078138e-01
8.07676464e-02 6.49392128e-01 1.55148506e-01 -2.15394795e-01
4.72861171e-01 5.12868226e-01 4.10184979e-01 -1.18381274e+00
-5.51479220e-01 -3.89109254e-01 2.68416554e-01 -4.12201852e-01
-1.87197790e-01 -5.52941382e-01 -5.27439237e-01 7.57120699e-02
-1.53347269e-01 7.56502897e-02 4.89859909e-01 9.39926386e-01
4.46393371e-01 3.76039654e-01 5.64753890e-01 -9.06792283e-01
-1.33489251e+00 -1.07176530e+00 -7.57931471e-01 3.56483668e-01
2.63150692e-01 -9.33863580e-01 -5.39396107e-01 -5.89656606e-02] | [9.99753189086914, 3.0009000301361084] |
ef92c397-5374-4987-a31e-bef35445f7cd | mmocr-a-comprehensive-toolbox-for-text | 2108.06543 | null | https://arxiv.org/abs/2108.06543v1 | https://arxiv.org/pdf/2108.06543v1.pdf | MMOCR: A Comprehensive Toolbox for Text Detection, Recognition and Understanding | We present MMOCR-an open-source toolbox which provides a comprehensive pipeline for text detection and recognition, as well as their downstream tasks such as named entity recognition and key information extraction. MMOCR implements 14 state-of-the-art algorithms, which is significantly more than all the existing open-source OCR projects we are aware of to date. To facilitate future research and industrial applications of text recognition-related problems, we also provide a large number of trained models and detailed benchmarks to give insights into the performance of text detection, recognition and understanding. MMOCR is publicly released at https://github.com/open-mmlab/mmocr. | ['Dahua Lin', 'Wayne Zhang', 'Kai Chen', 'Wenwei Zhang', 'Tong Gao', 'Yiqin Zhu', 'Huaqiang Wei', 'Jianyong Chen', 'Tsui Hin Lin', 'Xiaoyu Yue', 'Zhizhong Li', 'Hongbin Sun', 'Zhanghui Kuang'] | 2021-08-14 | null | null | null | null | ['key-information-extraction'] | ['natural-language-processing'] | [ 1.49639711e-01 -3.39102417e-01 -7.32826740e-02 -1.97888479e-01
-1.01636934e+00 -9.45026457e-01 6.91649079e-01 2.26601601e-01
-4.88312334e-01 2.82481909e-01 3.37415487e-01 -3.91784012e-01
5.22530556e-01 -5.79156756e-01 -2.22188741e-01 -2.71890491e-01
4.77117777e-01 5.13973892e-01 1.04152821e-01 1.25839919e-01
7.72968709e-01 3.59446675e-01 -1.38162673e+00 6.94093466e-01
8.36849213e-01 5.47253668e-01 3.50654908e-02 1.21571124e+00
-4.74827349e-01 6.76408112e-01 -7.65314877e-01 -8.55803430e-01
-5.09219095e-02 -1.14432737e-01 -1.08839297e+00 -2.55005378e-02
7.26333380e-01 -3.92033696e-01 -6.65691733e-01 8.58816087e-01
6.33540928e-01 -3.13557237e-02 5.43927073e-01 -9.85578120e-01
-1.17784131e+00 7.41300523e-01 -2.19803080e-01 3.23113114e-01
4.89874631e-01 2.93798447e-01 1.18547785e+00 -1.25364923e+00
7.28219330e-01 8.55386078e-01 5.94966710e-01 6.41820848e-01
-7.54109740e-01 -4.96303469e-01 4.11028340e-02 5.54368235e-02
-1.45174742e+00 -8.51871848e-01 2.52925605e-01 -5.32378793e-01
1.36312819e+00 4.39878464e-01 2.50311375e-01 1.21251237e+00
-1.01660369e-02 1.46912587e+00 8.92841518e-01 -5.51610470e-01
-1.90374017e-01 1.03609122e-01 5.15056431e-01 8.02974522e-01
3.38352799e-01 -4.79582518e-01 -7.37927079e-01 4.75606993e-02
3.92967820e-01 -8.81524235e-02 -3.76006246e-01 3.14988673e-01
-1.47371483e+00 4.12528932e-01 -4.84523773e-02 4.87134784e-01
8.08065087e-02 -2.20107026e-02 5.59785485e-01 1.65863425e-01
6.55300975e-01 6.07132971e-01 -7.61516929e-01 -4.61993098e-01
-1.15794790e+00 5.56214787e-02 1.13611388e+00 1.45949852e+00
3.29353660e-01 -1.19699799e-01 -3.52307975e-01 1.08201873e+00
3.79770994e-01 6.27366185e-01 4.86154646e-01 -4.64023381e-01
8.43891859e-01 9.55759227e-01 4.04877402e-02 -6.30443990e-01
-2.11152241e-01 -1.84927166e-01 -4.15130287e-01 -4.75001991e-01
4.43402857e-01 -5.59040569e-02 -1.02649212e+00 8.12232554e-01
1.74763501e-01 -2.48235360e-01 -2.09038764e-01 4.61616635e-01
1.21148503e+00 7.69025922e-01 -1.61818147e-01 2.84099162e-01
1.54136777e+00 -1.15429389e+00 -9.20027673e-01 -5.27957618e-01
1.00498724e+00 -1.38006473e+00 8.93969893e-01 1.92220658e-01
-7.29811668e-01 -4.10558283e-02 -7.01650262e-01 -6.83630228e-01
-1.13812256e+00 7.99322426e-01 4.48301762e-01 8.46327722e-01
-9.17549551e-01 3.54570419e-01 -7.87075043e-01 -8.79158676e-01
5.68306804e-01 -2.81059146e-02 -3.49534094e-01 -9.34716985e-02
-8.51214111e-01 8.64776194e-01 2.96206892e-01 1.28051847e-01
-6.33716881e-01 -6.34749591e-01 -7.71416843e-01 -1.63032815e-01
1.40806928e-01 -4.38152015e-01 1.63834751e+00 -4.91344243e-01
-1.35068381e+00 1.32690322e+00 -4.07919377e-01 -4.90501113e-02
4.93232071e-01 -6.10265315e-01 -6.46717012e-01 1.04151025e-01
1.02436105e-02 4.99292195e-01 6.75174892e-01 -5.49149990e-01
-5.94935536e-01 -4.87978905e-01 -5.85282683e-01 2.95121133e-01
-8.15740526e-01 8.21049035e-01 -8.86906445e-01 -9.84760940e-01
-2.78570682e-01 -7.20359743e-01 3.40609193e-01 2.70949394e-01
-9.97088253e-01 -2.66824991e-01 8.07659268e-01 -1.09982014e+00
1.58134031e+00 -2.06603456e+00 -1.24758087e-01 -4.19627547e-01
2.24138781e-01 4.41413254e-01 -1.72965720e-01 1.04814565e+00
7.16405511e-02 5.12084067e-01 -2.15550974e-01 -8.90750885e-01
3.24851662e-01 -4.68524337e-01 -4.43262309e-01 5.65864623e-01
1.66500330e-01 9.67902422e-01 -8.03634584e-01 -4.43117172e-01
4.03319627e-01 4.82253909e-01 1.34805739e-01 1.14039674e-01
-2.65899450e-01 -1.02897897e-01 -4.07313585e-01 9.16092813e-01
5.69290400e-01 -4.37220246e-01 -1.28053054e-01 2.04055816e-01
-3.79573137e-01 6.25422597e-01 -9.70233023e-01 1.38356757e+00
-1.78310946e-01 1.37960267e+00 1.03922924e-02 -3.24127942e-01
6.72300100e-01 3.63938749e-01 1.83525197e-02 -9.94201973e-02
2.96232283e-01 1.57288656e-01 -8.27185333e-01 -4.09273684e-01
1.01783371e+00 6.24457598e-01 3.07608098e-02 5.25775671e-01
8.94579515e-02 -3.69654447e-02 5.24370134e-01 6.08487070e-01
1.26170290e+00 4.65832800e-02 3.93953472e-01 1.41175911e-01
5.42692959e-01 2.32069790e-01 1.54148296e-01 7.90072739e-01
-4.27919686e-01 6.97578371e-01 1.86854854e-01 -1.89545766e-01
-1.07948685e+00 -7.74280965e-01 -3.87748986e-01 1.40370619e+00
-2.38593057e-01 -9.11629200e-01 -9.92511511e-01 -7.13492274e-01
1.10109150e-01 7.69138336e-01 -6.00272059e-01 3.89897972e-01
-4.26519364e-01 -7.77017832e-01 1.15214777e+00 6.50427580e-01
5.78350306e-01 -1.22266901e+00 6.56469986e-02 -1.90343097e-01
-2.79568464e-01 -1.34929061e+00 -9.82289672e-01 1.74068362e-02
-7.38941252e-01 -1.08088326e+00 -7.33525634e-01 -1.05082917e+00
6.55320227e-01 3.62982690e-01 1.01227367e+00 1.27240792e-01
-6.43656850e-01 5.69808543e-01 -4.63015199e-01 -4.30402339e-01
-3.84917945e-01 4.37273890e-01 -6.73458219e-01 -2.57561803e-01
8.01092148e-01 2.59357810e-01 -5.30395508e-01 3.04456711e-01
-8.24690461e-01 1.12495665e-02 5.11503875e-01 5.02498209e-01
3.71178746e-01 -1.26371652e-01 8.30804259e-02 -9.93646801e-01
6.85219288e-01 -1.90870881e-01 -8.34401369e-01 4.94314820e-01
-5.74296474e-01 -2.13283852e-01 4.50249404e-01 -3.87194343e-02
-1.01069188e+00 2.32153818e-01 -3.78332376e-01 2.63585061e-01
-5.27207375e-01 3.51974577e-01 -5.06227314e-02 -6.12555444e-03
4.37478662e-01 7.15319455e-01 -7.09459364e-01 -7.79953361e-01
5.60999930e-01 1.33204782e+00 3.36685091e-01 -4.74527478e-02
6.76592529e-01 2.44519249e-01 -7.52646446e-01 -1.21712565e+00
-1.00117743e+00 -1.09229076e+00 -8.62608612e-01 -1.01776063e-01
1.00733805e+00 -8.40451419e-01 -3.02309573e-01 1.17629445e+00
-1.35957634e+00 -4.18728530e-01 2.51776308e-01 -6.71097934e-02
-1.87573671e-01 4.86770838e-01 -1.15966988e+00 -8.39710772e-01
-9.28304434e-01 -8.24073195e-01 1.55852091e+00 1.48382545e-01
-2.23326817e-01 -1.06083047e+00 2.14274138e-01 8.26423109e-01
2.22191900e-01 -2.99975902e-01 4.56645101e-01 -1.03087711e+00
-5.29451907e-01 -4.57188725e-01 -2.89385438e-01 7.61049390e-02
-6.36256039e-02 6.06012940e-01 -1.06002164e+00 -1.22622833e-01
-5.23905933e-01 -2.03147203e-01 1.10200536e+00 4.50580120e-02
1.00339949e+00 -3.25027198e-01 -6.13547206e-01 5.32054842e-01
1.07730889e+00 -9.48569998e-02 6.46990478e-01 6.03857338e-01
7.52521574e-01 4.33002025e-01 4.20723021e-01 4.34564024e-01
5.41132212e-01 4.60146546e-01 5.28375544e-02 5.63254617e-02
-2.12894931e-01 -4.03341293e-01 5.16959846e-01 9.54914391e-01
6.46093071e-01 -7.44715750e-01 -1.21741056e+00 6.08471513e-01
-1.70869005e+00 -9.15154696e-01 -6.42800391e-01 1.98121834e+00
1.03986502e+00 -1.87036946e-01 7.93025568e-02 5.37087955e-02
1.21707118e+00 1.91114962e-01 -4.09791440e-01 -1.96165562e-01
-1.88855529e-01 9.32065025e-02 5.57068110e-01 2.54594862e-01
-1.61340559e+00 1.47432017e+00 7.46866512e+00 8.55481267e-01
-7.55539596e-01 2.51684953e-02 4.46356416e-01 -9.01313275e-02
1.92512408e-01 -1.54184490e-01 -1.40765893e+00 3.94600183e-01
1.02613533e+00 -3.89381871e-02 3.38213056e-01 9.59686518e-01
-5.21793626e-02 9.53320488e-02 -1.07255745e+00 9.01633203e-01
3.32344055e-01 -1.34470630e+00 1.06727676e-02 -1.09342270e-01
6.03049040e-01 6.41403973e-01 9.05758515e-02 1.20132782e-01
4.87098753e-01 -7.35206246e-01 6.76435947e-01 1.72566399e-01
7.32685506e-01 -2.62789309e-01 4.91131514e-01 3.49853449e-02
-1.30950391e+00 9.63008106e-02 -2.61958480e-01 4.34986234e-01
-7.74559081e-02 7.13012397e-01 -6.87349617e-01 2.54119426e-01
5.40913284e-01 1.05936420e+00 -1.02509344e+00 1.22445500e+00
-6.19747937e-01 7.49182403e-01 -2.38177225e-01 -4.01909232e-01
-3.91912162e-02 2.35453635e-01 5.83610594e-01 1.94004071e+00
-5.53248972e-02 -1.84881330e-01 -1.11082479e-01 7.62077451e-01
-6.01673126e-01 5.87092400e-01 -5.83256602e-01 -7.63925791e-01
7.72485912e-01 1.58942103e+00 -9.38677907e-01 -5.54579437e-01
-6.50647581e-01 1.45227873e+00 3.42792422e-01 1.93347991e-01
-5.89038551e-01 -1.09169912e+00 5.98373592e-01 -2.56446511e-01
4.45187986e-01 -4.08722311e-01 -4.41178709e-01 -1.72071302e+00
1.68809444e-01 -8.54459643e-01 7.27113068e-01 -8.87058556e-01
-1.31425595e+00 4.52581376e-01 -6.72980845e-01 -9.32053566e-01
3.06530923e-01 -9.79280114e-01 -6.17820919e-01 6.19759679e-01
-1.46839786e+00 -1.18900716e+00 -3.74202430e-01 5.57461143e-01
9.76785243e-01 -3.06828246e-02 8.60589683e-01 2.31107607e-01
-1.33872795e+00 8.43736887e-01 6.50502324e-01 1.05819774e+00
1.19691086e+00 -1.32821834e+00 1.24334538e+00 1.06006014e+00
2.76937425e-01 7.74714887e-01 2.70093620e-01 -8.03040028e-01
-1.73529100e+00 -1.28956282e+00 1.29421818e+00 -1.14961672e+00
1.24517846e+00 -8.15475345e-01 -7.76962161e-01 9.40567017e-01
2.88191676e-01 -2.75390238e-01 8.25695753e-01 1.38688892e-01
-5.55714250e-01 2.79455662e-01 -8.30445766e-01 6.46309257e-01
8.89211178e-01 -8.88856411e-01 -6.56088889e-01 7.77345657e-01
4.75667357e-01 -3.07742745e-01 -9.34696496e-01 -2.91054994e-01
5.83078206e-01 -6.14580691e-01 4.02393878e-01 -2.87926763e-01
4.80517268e-01 -1.06039606e-01 1.42396972e-01 -9.12150085e-01
-1.36379838e-01 -7.37966061e-01 -2.04880461e-01 1.72911930e+00
7.91313350e-01 -7.89297760e-01 5.84366083e-01 3.59528214e-01
-1.83376849e-01 -4.81634557e-01 -6.41105890e-01 -7.08241343e-01
2.50818670e-01 -6.05120480e-01 2.70738065e-01 1.06431556e+00
4.04753566e-01 3.18769127e-01 -1.20444693e-01 1.15323938e-01
2.88315654e-01 5.47080375e-02 6.08241498e-01 -8.26332748e-01
5.13684079e-02 -7.10483253e-01 -3.33807528e-01 -1.32100236e+00
8.28238428e-02 -1.11993790e+00 2.21702993e-01 -1.74424422e+00
4.28674698e-01 7.44140670e-02 -1.95501801e-02 8.76416802e-01
-3.54327917e-01 3.55040044e-01 8.99090767e-02 4.51205373e-01
-9.63547409e-01 4.46525186e-01 8.71602774e-01 -2.55688846e-01
1.91791579e-02 -2.21708566e-01 -6.89187586e-01 4.68762428e-01
9.54633594e-01 -6.77374184e-01 2.60313869e-01 -8.25593710e-01
2.63839327e-02 -4.10029620e-01 -8.35541934e-02 -6.19914591e-01
4.57535803e-01 9.76850539e-02 5.88745534e-01 -9.40798581e-01
4.86027496e-03 -3.03797156e-01 -6.15640640e-01 1.82531387e-01
-5.53565919e-01 2.29842722e-01 2.66353756e-01 1.95211902e-01
-6.98716789e-02 -3.75884980e-01 6.45851612e-01 6.88892649e-03
-6.53503418e-01 2.52848268e-01 -9.46120799e-01 3.40931088e-01
7.27282465e-01 7.75506794e-02 -1.08212042e+00 -1.92159772e-01
-2.87345350e-01 2.45955363e-01 6.49735510e-01 7.04804778e-01
5.47546625e-01 -8.36581767e-01 -7.67224967e-01 -4.44370545e-02
6.14364386e-01 -5.66261113e-01 -3.04249555e-01 8.04421961e-01
-6.55063510e-01 8.44989836e-01 2.71480322e-01 -2.54064798e-01
-1.54568601e+00 4.07958955e-01 2.38296568e-01 -2.98454076e-01
-5.73201180e-01 6.51395977e-01 -1.84640408e-01 -6.98468626e-01
2.49429658e-01 -7.33885393e-02 -3.25975381e-02 -1.11468993e-01
9.95760083e-01 6.99435234e-01 3.47323060e-01 -4.98291850e-01
-6.33184254e-01 3.98892313e-01 -4.42679584e-01 1.93949882e-02
1.03431058e+00 -1.59300879e-01 -3.39128435e-01 3.47251117e-01
1.06486619e+00 9.05204415e-02 -5.48177779e-01 -4.90756817e-02
3.19830865e-01 -3.14016670e-01 2.89450288e-01 -1.12764454e+00
-9.08682764e-01 9.37509894e-01 2.88935751e-01 2.82023363e-02
1.01021743e+00 -6.17582649e-02 8.11230481e-01 8.11452925e-01
-8.35474432e-02 -1.42675233e+00 -1.41773531e-02 8.79679978e-01
6.52674794e-01 -1.21672332e+00 1.35049298e-01 -5.99351764e-01
-3.41275543e-01 1.36205649e+00 5.45143843e-01 2.52774328e-01
4.87099856e-01 6.44658685e-01 2.60407895e-01 -1.01876799e-02
-8.69586229e-01 -2.99822837e-01 4.29631591e-01 4.04053807e-01
1.07249320e+00 -1.40661553e-01 -1.08562643e-02 3.27926815e-01
-9.71426591e-02 -1.32528007e-01 5.23506582e-01 1.20634341e+00
-4.95832741e-01 -1.13384759e+00 -4.13479716e-01 5.98240733e-01
-7.58174896e-01 -7.74000525e-01 -1.23629987e+00 5.86348236e-01
-6.63288355e-01 1.05007863e+00 -8.70118812e-02 -2.29464054e-01
1.55840546e-01 3.67478192e-01 3.02485377e-01 -8.39066446e-01
-8.57632816e-01 -3.78080904e-02 3.87210190e-01 -4.57328588e-01
2.75322586e-01 -8.76259565e-01 -1.12667298e+00 -6.78011596e-01
-6.27033412e-01 -1.31585732e-01 8.81249785e-01 6.67777061e-01
6.65476441e-01 9.93359163e-02 2.40850791e-01 -4.69450116e-01
-3.03094298e-01 -1.20531976e+00 -3.46933275e-01 1.09506257e-01
7.99440295e-02 3.54553647e-02 -5.26229262e-01 2.75406718e-01] | [11.95984172821045, 2.3208062648773193] |
0507e110-bad7-42ea-8ad2-0cc073b39865 | unbiased-scene-graph-generation-using | 2210.00920 | null | https://arxiv.org/abs/2210.00920v1 | https://arxiv.org/pdf/2210.00920v1.pdf | Unbiased Scene Graph Generation using Predicate Similarities | Scene Graphs are widely applied in computer vision as a graphical representation of relationships between objects shown in images. However, these applications have not yet reached a practical stage of development owing to biased training caused by long-tailed predicate distributions. In recent years, many studies have tackled this problem. In contrast, relatively few works have considered predicate similarities as a unique dataset feature which also leads to the biased prediction. Due to the feature, infrequent predicates (e.g., parked on, covered in) are easily misclassified as closely-related frequent predicates (e.g., on, in). Utilizing predicate similarities, we propose a new classification scheme that branches the process to several fine-grained classifiers for similar predicate groups. The classifiers aim to capture the differences among similar predicates in detail. We also introduce the idea of transfer learning to enhance the features for the predicates which lack sufficient training samples to learn the descriptive representations. The results of extensive experiments on the Visual Genome dataset show that the combination of our method and an existing debiasing approach greatly improves performance on tail predicates in challenging SGCls/SGDet tasks. Nonetheless, the overall performance of the proposed approach does not reach that of the current state of the art, so further analysis remains necessary as future work. | ['Yusuke Matsui', 'Misaki Ohashi'] | 2022-10-03 | null | null | null | null | ['scene-graph-generation', 'unbiased-scene-graph-generation'] | ['computer-vision', 'computer-vision'] | [ 6.21891141e-01 -1.68348048e-02 -2.91347831e-01 -5.15469432e-01
-4.26813960e-01 -4.14210558e-01 6.47099614e-01 4.97959465e-01
6.07599355e-02 7.28396475e-01 6.27644882e-02 -1.46308392e-01
-2.44937539e-01 -8.05342436e-01 -9.65577483e-01 -8.72355640e-01
-2.27597523e-02 5.00294983e-01 7.98683286e-01 6.75107837e-02
3.25791180e-01 4.25975680e-01 -1.93000710e+00 5.88192940e-01
9.58032668e-01 1.09501278e+00 3.13180298e-01 1.08499259e-01
-3.06378573e-01 8.76176894e-01 -6.88592494e-01 -3.52006853e-01
-3.61833274e-02 -4.73389953e-01 -6.12731040e-01 1.06452465e-01
6.73003793e-01 1.17802896e-01 -5.62593713e-02 1.23593724e+00
3.32586288e-01 -5.45031019e-02 7.32018769e-01 -1.47357571e+00
-7.09211648e-01 4.98663008e-01 -8.72866035e-01 3.42083305e-01
2.62817651e-01 -2.80885458e-01 1.23977876e+00 -6.03050768e-01
5.94081998e-01 1.33093488e+00 6.47480905e-01 6.21671304e-02
-1.08382285e+00 -6.45502210e-01 4.54540342e-01 7.22851157e-01
-1.26665974e+00 2.25638896e-01 1.00572407e+00 -4.35494512e-01
6.71145916e-01 3.86942416e-01 5.11941731e-01 7.88794398e-01
2.30478302e-01 9.52567101e-01 1.18052852e+00 -2.43459344e-01
5.26006594e-02 1.39409184e-01 4.36947644e-01 6.57773197e-01
5.79168499e-01 3.01048649e-03 -5.68563223e-01 -7.50940889e-02
1.61727875e-01 2.01791758e-03 -3.95191431e-01 -9.16003346e-01
-9.86765802e-01 7.82897472e-01 6.13167107e-01 1.87673196e-01
-2.91383475e-01 -1.82849467e-01 4.22367007e-01 6.60720542e-02
5.63699663e-01 2.23833561e-01 -4.86060560e-01 4.10698401e-03
-6.09274507e-01 2.73615152e-01 5.91043770e-01 1.08385992e+00
8.52970421e-01 -4.84792024e-01 -3.40038627e-01 6.85201526e-01
3.35634723e-02 2.01937243e-01 1.74895212e-01 -2.46647120e-01
3.84597749e-01 8.84402752e-01 -1.59410343e-01 -1.33645320e+00
-4.43359077e-01 -4.31630671e-01 -7.43216217e-01 -1.31438211e-01
4.90222007e-01 2.22780511e-01 -8.78549814e-01 1.62931979e+00
4.44103926e-01 1.99266270e-01 3.19918729e-02 6.62595749e-01
9.40740168e-01 6.12710357e-01 1.84426591e-01 -6.05288967e-02
1.49504614e+00 -9.93419111e-01 -6.37022614e-01 -1.34828284e-01
5.40398657e-01 -6.80996776e-01 1.08363068e+00 3.10151398e-01
-5.83331943e-01 -6.19095802e-01 -8.96243453e-01 1.35211349e-01
-6.67349160e-01 1.69785753e-01 9.29217935e-01 5.86288035e-01
-5.80095053e-01 6.14443898e-01 -8.46275032e-01 -6.65118277e-01
8.17490399e-01 1.18026912e-01 -2.33179614e-01 -2.30430678e-01
-1.02705538e+00 7.82111228e-01 8.13225567e-01 -2.26995960e-01
-5.09945631e-01 -7.80546308e-01 -7.69584835e-01 8.87103975e-02
6.03510320e-01 -3.69046420e-01 8.38088632e-01 -8.31224263e-01
-9.01691496e-01 9.16867971e-01 -2.15636581e-01 -4.05669123e-01
4.41479653e-01 -2.60363221e-01 -3.48585099e-01 -2.21803016e-03
1.45373851e-01 5.15684962e-01 6.49769723e-01 -1.29912567e+00
-1.05243325e+00 -3.77909601e-01 1.88480377e-01 2.65837580e-01
-2.34798193e-01 4.89000529e-02 -6.01118207e-01 -5.27682185e-01
3.22055548e-01 -9.92794514e-01 1.95198148e-01 -1.61605272e-02
-6.54497623e-01 -5.32126486e-01 1.16616821e+00 -4.88762349e-01
1.00038445e+00 -2.25936604e+00 1.60582792e-02 -3.86356898e-02
-2.72407196e-02 3.79280686e-01 1.87217236e-01 6.05996132e-01
-2.78137803e-01 -4.80268002e-02 -1.63846567e-01 1.13650769e-01
-1.00376792e-01 3.11768889e-01 -3.18516701e-01 5.38318038e-01
3.72436523e-01 6.16916537e-01 -1.03660619e+00 -6.42563045e-01
2.12634712e-01 2.48937711e-01 -3.71910930e-01 2.16786668e-01
-3.46720010e-01 4.21724081e-01 -5.98354340e-01 7.28358686e-01
9.05512869e-01 -4.14886296e-01 1.52999252e-01 -5.42797327e-01
5.99333027e-04 2.93705106e-01 -1.03489363e+00 1.23121357e+00
9.08719078e-02 5.67842245e-01 -5.34494102e-01 -1.42228711e+00
9.08814430e-01 -6.31313026e-02 2.45448112e-01 -4.95986342e-01
-6.31440878e-02 5.42850420e-02 2.36716598e-01 -4.75711793e-01
2.23036766e-01 4.99079488e-02 4.22285609e-02 -1.32493749e-01
-3.03285688e-01 -1.91097558e-02 4.14020956e-01 -1.61323622e-02
9.75266337e-01 4.14966375e-01 5.68533480e-01 -2.58170009e-01
6.81043983e-01 3.97657663e-01 8.53147447e-01 6.71944797e-01
-1.36957586e-01 4.71679538e-01 7.26911426e-01 -3.15528870e-01
-6.31326973e-01 -1.02255261e+00 -3.08991820e-01 1.05740893e+00
5.37853718e-01 -4.19619828e-01 -4.69089240e-01 -9.88074303e-01
1.91383436e-01 7.12761164e-01 -5.90751946e-01 -1.96992204e-01
-4.21565682e-01 -7.58021057e-01 3.53377759e-01 7.04703271e-01
6.84497356e-01 -1.13202798e+00 -6.40199959e-01 -1.29688010e-01
-9.30525884e-02 -1.05117559e+00 -9.60514620e-02 2.23068863e-01
-6.94725811e-01 -1.38115752e+00 -4.71170276e-01 -1.01809049e+00
7.00481355e-01 4.63256598e-01 1.16887856e+00 -8.18394348e-02
-2.86085486e-01 1.26918316e-01 -5.74518263e-01 -6.89331532e-01
-2.19445899e-01 -2.36197203e-01 -2.69800186e-01 1.30970869e-02
5.26851654e-01 -2.94352353e-01 -3.75462979e-01 4.01335955e-01
-7.04950988e-01 1.17433570e-01 6.98105097e-01 1.10406256e+00
8.58849585e-01 5.87949753e-01 3.56988788e-01 -1.30617356e+00
2.01491714e-01 -4.65059191e-01 -6.81250870e-01 5.14095902e-01
-4.55946952e-01 6.31368533e-02 5.87194502e-01 -4.99675661e-01
-1.28290367e+00 -4.57551740e-02 2.02823311e-01 -2.96962172e-01
-3.50507081e-01 5.97988665e-01 -5.94652653e-01 9.20455977e-02
2.56084353e-01 2.41749570e-01 -2.64450699e-01 -3.79370213e-01
1.72442898e-01 4.58160251e-01 3.88415575e-01 -5.21369100e-01
7.69203782e-01 5.06537378e-01 1.83262631e-01 -9.02455509e-01
-1.05474305e+00 -6.22447491e-01 -3.78239274e-01 1.83172196e-01
7.68277466e-01 -8.67355466e-01 -5.84712923e-01 5.70686162e-01
-9.22693789e-01 5.95849492e-02 -1.64743319e-01 3.15649927e-01
-6.14403546e-01 7.47769594e-01 -3.08115453e-01 -5.76438963e-01
-2.79918383e-03 -9.67815816e-01 1.01200378e+00 3.52892041e-01
-1.82165459e-01 -7.98331141e-01 -7.66458735e-02 3.68051678e-01
5.07040136e-02 3.99682045e-01 1.35180128e+00 -9.62929845e-01
-7.00132012e-01 -7.13925585e-02 -4.71762538e-01 2.15274155e-01
4.79162514e-01 1.07478105e-01 -7.64475286e-01 -2.87932098e-01
-2.48246804e-01 -3.28601271e-01 9.79734302e-01 3.22493643e-01
1.57955527e+00 1.15608916e-01 -9.12360787e-01 5.07043779e-01
1.29562986e+00 3.80157292e-01 5.36280453e-01 3.96713495e-01
7.76598334e-01 8.07310998e-01 1.26537013e+00 3.56615633e-01
2.73108840e-01 7.26691544e-01 4.92384315e-01 -6.71424866e-02
-3.20047170e-01 -4.16314840e-01 -1.99990962e-02 1.97455660e-02
8.48961174e-02 -4.99141246e-01 -9.56075490e-01 6.20497644e-01
-1.92976069e+00 -9.15424407e-01 -2.90534943e-01 2.12415075e+00
5.63510895e-01 2.55445033e-01 -1.11682095e-01 2.55845010e-01
9.98588622e-01 2.03520775e-01 -6.17722034e-01 -4.58296463e-02
-2.76645511e-01 -8.40546191e-03 4.57550287e-01 -2.39505768e-01
-1.44987142e+00 1.03405535e+00 5.14526939e+00 9.12844956e-01
-9.52609658e-01 -2.35648349e-01 4.40500289e-01 5.45955300e-01
-1.37845799e-01 9.55811813e-02 -9.39516962e-01 5.03911078e-01
4.18242335e-01 -7.13861063e-02 -6.63690791e-02 9.45691228e-01
-2.03001276e-01 -2.53724575e-01 -1.16704583e+00 9.13250327e-01
2.97139674e-01 -1.03287804e+00 1.85276940e-01 -5.25910296e-02
7.29461074e-01 -2.73079813e-01 1.50706187e-01 2.88087577e-01
1.61572605e-01 -7.17089593e-01 6.72786057e-01 2.46489584e-01
4.12255943e-01 -7.39999950e-01 7.60045111e-01 3.99521172e-01
-1.21488774e+00 -9.97315645e-02 -6.59789205e-01 -5.35371825e-02
-6.97795227e-02 6.99737906e-01 -9.69901204e-01 8.42579782e-01
9.96541321e-01 1.02552652e+00 -7.45232165e-01 1.12910712e+00
-4.69416499e-01 6.98316872e-01 -9.34331268e-02 -1.73928544e-01
4.65590581e-02 -9.01124999e-02 4.13060278e-01 1.00874949e+00
2.23869026e-01 -2.36485481e-01 2.44099960e-01 6.88637078e-01
1.24662191e-01 1.95282429e-01 -6.87980711e-01 -2.40242988e-01
2.03124791e-01 1.16580677e+00 -9.18703318e-01 -3.77326906e-01
-6.52673423e-01 7.58742929e-01 5.30069172e-01 7.39570931e-02
-9.61788535e-01 -1.82436571e-01 6.55158460e-01 3.73824835e-02
8.00794601e-01 1.66828439e-01 -1.08120084e-01 -8.88878763e-01
4.45278287e-02 -8.75392675e-01 7.27758765e-01 -6.03227079e-01
-1.59596777e+00 3.80647928e-01 2.44200572e-01 -1.38559473e+00
1.96311418e-02 -6.60089314e-01 -5.33960044e-01 6.67105496e-01
-1.60905480e+00 -1.28394413e+00 -5.01692832e-01 5.92623055e-01
4.18017864e-01 3.40175442e-02 6.98012114e-01 1.49157524e-01
-5.25012493e-01 4.10311192e-01 1.07115224e-01 -5.39652407e-02
8.76982450e-01 -1.32608080e+00 7.09792450e-02 8.04139674e-01
1.97913900e-01 4.37490284e-01 8.99796844e-01 -8.91802073e-01
-1.06135333e+00 -1.18143845e+00 9.26807821e-01 -2.58649468e-01
6.70099795e-01 -1.56444460e-01 -1.09694934e+00 7.35431075e-01
1.13946989e-01 1.61578298e-01 8.15413535e-01 3.05510581e-01
-4.29383159e-01 -1.71723858e-01 -1.05785024e+00 4.44469839e-01
1.21003032e+00 -1.83190107e-01 -8.96202385e-01 3.96365732e-01
4.36852545e-01 -3.96267742e-01 -4.98301059e-01 6.49999678e-01
1.53513059e-01 -1.04570591e+00 8.29705536e-01 -6.78864956e-01
4.46404219e-01 -4.90341544e-01 -1.39433682e-01 -1.28735936e+00
-3.37224752e-01 -7.73192942e-02 -4.23292406e-02 1.43375099e+00
1.42813280e-01 -6.95482373e-01 1.03166723e+00 1.48773938e-01
-2.28691444e-01 -6.62762284e-01 -6.00561440e-01 -9.48432982e-01
-2.35754222e-01 -1.07193708e-01 5.62403679e-01 7.91522980e-01
-1.53508067e-01 2.04684168e-01 -3.14959943e-01 4.07816499e-01
6.88276231e-01 8.58210325e-01 7.77999401e-01 -1.42877090e+00
-2.17724413e-01 -1.78744555e-01 -8.27790022e-01 -1.06385922e+00
2.56272197e-01 -9.27013397e-01 1.11436598e-01 -1.57528877e+00
5.41402280e-01 -5.93640327e-01 -3.65078360e-01 4.71482038e-01
-5.60506046e-01 6.88061491e-02 1.23287760e-01 6.35664240e-02
-6.38794899e-01 4.66330916e-01 1.21164811e+00 -4.09379303e-01
7.47619290e-03 1.42992139e-01 -7.38628089e-01 8.75034511e-01
8.86257768e-01 -3.63584816e-01 -5.84462702e-01 -2.03212555e-02
-1.57968793e-02 -3.28706414e-01 2.49928787e-01 -1.05408776e+00
5.57086132e-02 -1.24846853e-01 4.29971576e-01 -1.07554948e+00
2.82584101e-01 -9.24494565e-01 1.74915746e-01 5.44145942e-01
-1.20918147e-01 -1.37242049e-01 2.10461959e-01 9.42507088e-01
-4.94700849e-01 -3.96331958e-02 6.25218451e-01 8.83580819e-02
-1.19639814e+00 1.85190767e-01 -1.89099200e-02 5.75817823e-02
1.34519613e+00 -2.90195972e-01 -5.18742740e-01 -1.19960822e-01
-4.47797954e-01 2.37637222e-01 4.84593451e-01 4.55389380e-01
4.41167712e-01 -9.42097962e-01 -5.03411710e-01 4.21061330e-02
5.27548969e-01 4.00400981e-02 3.31980318e-01 6.69146359e-01
-3.62003922e-01 5.87198675e-01 -3.49455684e-01 -8.70888531e-01
-1.69603419e+00 8.57024789e-01 -3.74541879e-02 -4.29252982e-01
-6.68438077e-01 9.28519487e-01 6.84908271e-01 -1.93429217e-01
2.33466133e-01 -4.01781291e-01 -5.11954010e-01 2.59435862e-01
3.14116895e-01 2.88647294e-01 5.08083515e-02 -6.33628666e-01
-5.84290445e-01 6.99490070e-01 -2.58447081e-01 7.80824661e-01
1.28206706e+00 8.58779252e-02 -1.23382203e-01 3.95308912e-01
9.75736737e-01 1.08917588e-02 -1.20775068e+00 -3.05301547e-01
2.48850569e-01 -6.57946229e-01 -3.82569760e-01 -6.54542863e-01
-9.93788958e-01 8.70584548e-01 6.35555506e-01 1.66418135e-01
1.33910739e+00 3.66934270e-01 3.78813207e-01 2.87377059e-01
4.65514362e-01 -7.73806632e-01 2.88813803e-02 3.67281556e-01
7.11810410e-01 -1.34860027e+00 1.12459555e-01 -1.21702445e+00
-7.74016738e-01 8.23832035e-01 6.93059087e-01 1.12289917e-02
4.76167679e-01 -5.25631607e-02 -1.69848248e-01 -3.09578598e-01
-5.51947892e-01 -5.08833349e-01 3.47463280e-01 8.45375180e-01
3.86177868e-01 2.10043341e-01 -6.64567053e-01 2.70040929e-01
-1.35995343e-01 -2.60419399e-01 2.83676118e-01 1.01758146e+00
-3.61998558e-01 -1.10586905e+00 -2.11545840e-01 6.33909404e-01
-4.79081750e-01 2.67402213e-02 -3.41376394e-01 9.76819396e-01
4.63346332e-01 8.41701806e-01 1.90542698e-01 -1.95103168e-01
4.81920958e-01 1.41823981e-02 4.62561369e-01 -6.61566973e-01
-2.91260481e-01 -2.62048185e-01 1.98357210e-01 -5.25750399e-01
-6.51453435e-01 -8.83288980e-01 -1.22112870e+00 6.08046241e-02
-4.17856216e-01 4.63045314e-02 2.60790259e-01 6.44081295e-01
2.26241514e-01 6.66159928e-01 2.90726125e-01 -3.91122103e-01
-5.19728243e-01 -7.57142246e-01 -6.83808565e-01 8.54873776e-01
-5.63164130e-02 -1.16646564e+00 -2.52703518e-01 -2.74245795e-02] | [10.291558265686035, 1.7857413291931152] |
6fb17ad8-51a8-499a-b47e-ee143869a050 | fast-l1-minimization-algorithms-for-robust-1 | 1007.3753 | null | https://arxiv.org/abs/1007.3753v4 | https://arxiv.org/pdf/1007.3753v4.pdf | Fast L1-Minimization Algorithms For Robust Face Recognition | L1-minimization refers to finding the minimum L1-norm solution to an underdetermined linear system b=Ax. Under certain conditions as described in compressive sensing theory, the minimum L1-norm solution is also the sparsest solution. In this paper, our study addresses the speed and scalability of its algorithms. In particular, we focus on the numerical implementation of a sparsity-based classification framework in robust face recognition, where sparse representation is sought to recover human identities from very high-dimensional facial images that may be corrupted by illumination, facial disguise, and pose variation. Although the underlying numerical problem is a linear program, traditional algorithms are known to suffer poor scalability for large-scale applications. We investigate a new solution based on a classical convex optimization framework, known as Augmented Lagrangian Methods (ALM). The new convex solvers provide a viable solution to real-world, time-critical applications such as face recognition. We conduct extensive experiments to validate and compare the performance of the ALM algorithms against several popular L1-minimization solvers, including interior-point method, Homotopy, FISTA, SESOP-PCD, approximate message passing (AMP) and TFOCS. To aid peer evaluation, the code for all the algorithms has been made publicly available. | ['Yi Ma', 'S. Shankar Sastry', 'Arvind Ganesh', 'Zihan Zhou', 'Allen Y. Yang'] | 2010-07-21 | null | null | null | null | ['robust-face-recognition'] | ['computer-vision'] | [ 3.63462687e-01 -1.21395171e-01 -9.57698934e-03 -1.49442479e-01
-6.91440642e-01 -2.33510956e-01 2.37883344e-01 -5.08312464e-01
7.91455582e-02 8.62378240e-01 1.69873700e-01 -3.95384245e-02
-4.50470865e-01 -1.97279572e-01 -7.93483496e-01 -8.41366947e-01
-1.40017018e-01 3.28979284e-01 -8.04593801e-01 -1.81359261e-01
2.78917193e-01 7.39539564e-01 -1.58826888e+00 3.30433287e-02
8.09633255e-01 1.10778975e+00 -1.83805242e-01 2.56822735e-01
5.66487499e-02 4.77523386e-01 -1.12139262e-01 -2.71801531e-01
6.39660358e-01 -4.03568238e-01 -5.57353914e-01 4.19388533e-01
7.44906783e-01 -1.84994921e-01 -8.93738940e-02 1.26775098e+00
4.52977061e-01 1.50730014e-01 4.36346948e-01 -1.58089900e+00
-3.18771541e-01 2.53465641e-02 -8.60746264e-01 -9.98548642e-02
8.10545743e-01 -1.65012583e-01 3.52189124e-01 -1.56292677e+00
7.24908650e-01 1.40196943e+00 9.44431067e-01 3.18125516e-01
-1.14064014e+00 -5.53856671e-01 -1.07516116e-02 1.80299357e-01
-1.75783598e+00 -1.08048403e+00 6.83653474e-01 -2.15075597e-01
6.90134883e-01 5.81626594e-01 5.49016654e-01 6.45929575e-01
-3.37451473e-02 6.17678702e-01 1.15867484e+00 -4.45448190e-01
2.24186972e-01 6.85772747e-02 -1.84012756e-01 9.36795354e-01
3.88206512e-01 4.00101058e-02 -9.23645437e-01 -6.89674199e-01
6.14071727e-01 -1.15030244e-01 -5.67024589e-01 -3.13986480e-01
-1.08028960e+00 8.24287534e-01 -9.54641923e-02 1.13672584e-01
-6.14204705e-01 -1.49338499e-01 -2.46354938e-02 1.72224402e-01
5.51104426e-01 -2.76236646e-02 -9.82325524e-02 1.83213670e-02
-1.38493311e+00 3.40073824e-01 1.14280760e+00 8.58616531e-01
6.69621706e-01 5.05879700e-01 1.31775737e-01 7.89743245e-01
6.32471144e-01 8.11086833e-01 2.84360081e-01 -1.35610628e+00
2.35587373e-01 2.47385681e-01 -5.16597880e-03 -1.63523626e+00
-1.67445570e-01 -4.84253407e-01 -1.14116502e+00 8.08833167e-02
2.62822539e-01 -1.97781190e-01 -4.78459746e-01 1.51151431e+00
7.72594690e-01 9.31544721e-01 -8.48351419e-03 1.14941990e+00
7.17088163e-01 8.12293589e-01 -4.63773906e-01 -9.38756824e-01
8.79368484e-01 -7.93997347e-01 -8.38325441e-01 -3.07263043e-02
4.59046990e-01 -1.16724324e+00 2.04310372e-01 6.22671783e-01
-1.26322901e+00 -1.88710690e-01 -7.48215556e-01 2.17169419e-01
1.76254615e-01 1.54094607e-01 3.88272613e-01 6.45216584e-01
-1.06570721e+00 4.88315642e-01 -6.04428768e-01 -3.58693868e-01
4.55341667e-01 5.17703414e-01 -7.21447110e-01 -5.08310199e-01
-6.39758289e-01 7.88012564e-01 -1.87057525e-01 5.60575664e-01
-8.78448308e-01 -9.42532897e-01 -9.07428265e-01 -1.82749316e-01
3.98208231e-01 -3.92068565e-01 6.43685341e-01 -1.05130899e+00
-1.40477145e+00 9.58861291e-01 -7.31130898e-01 -1.08219273e-01
4.13081199e-01 -8.67629126e-02 -1.93124175e-01 1.04273461e-01
3.24400738e-02 1.98354676e-01 1.43258166e+00 -1.21227670e+00
-7.31881782e-02 -4.85810876e-01 -4.91661489e-01 1.97390750e-01
-3.35982502e-01 1.54890016e-01 -2.49044403e-01 -6.97439253e-01
5.28350472e-01 -1.08372736e+00 -2.30356410e-01 1.95217192e-01
-2.61507064e-01 3.20805937e-01 1.10365057e+00 -9.46925461e-01
9.94993091e-01 -2.22174096e+00 4.86258000e-01 5.83148360e-01
-5.57412282e-02 3.69569182e-01 -2.60845900e-01 2.87990451e-01
-2.24808604e-01 -2.16047525e-01 -5.91072738e-01 -5.00220001e-01
-2.91725636e-01 1.75364941e-01 -1.13202795e-01 1.25361955e+00
-1.71851993e-01 5.73538363e-01 -6.49848998e-01 -4.98829037e-01
9.27421674e-02 8.26465249e-01 -7.36540616e-01 9.05683637e-02
1.60317078e-01 6.07098699e-01 -2.89990813e-01 1.09592259e+00
1.06558633e+00 -5.88410273e-02 1.11373566e-01 -4.21861738e-01
-2.10922241e-01 -5.97748280e-01 -1.98865318e+00 1.83257961e+00
-2.05200627e-01 4.09345031e-01 1.08151078e+00 -1.38189578e+00
7.15713620e-01 4.73031342e-01 9.39134598e-01 -3.17362875e-01
1.42007425e-01 3.66929233e-01 -4.81237561e-01 -4.52965736e-01
2.30893552e-01 -8.87409747e-02 6.41660452e-01 3.64364624e-01
-1.38701797e-01 -2.78642271e-02 1.10545047e-01 2.45147914e-01
8.70152771e-01 -7.59287775e-02 4.05505657e-01 -5.74563265e-01
9.39745128e-01 3.00282575e-02 7.67390668e-01 3.77960414e-01
-6.82378486e-02 6.95080757e-01 2.14625821e-01 -2.70946771e-01
-6.84227169e-01 -6.18791342e-01 -4.80633706e-01 5.96660852e-01
-1.08371302e-02 -2.90258944e-01 -6.97014868e-01 -1.20013803e-01
1.91240698e-01 2.50504822e-01 -2.16254100e-01 1.85778290e-01
-5.73892057e-01 -9.76851702e-01 1.66992322e-01 -1.70120120e-01
4.49424356e-01 -5.69824696e-01 -8.77681896e-02 9.25351083e-02
-1.75226822e-01 -1.16660082e+00 -6.04455531e-01 -4.47389930e-01
-8.55266988e-01 -1.20041049e+00 -8.03189635e-01 -8.04934382e-01
1.06855702e+00 4.05705303e-01 8.10297310e-01 3.02155018e-01
-6.52011156e-01 8.71545017e-01 -1.75724551e-01 -1.57498896e-01
1.81880388e-02 -6.01985276e-01 5.09922802e-01 7.40931273e-01
1.81973092e-02 -4.46310014e-01 -3.55251700e-01 2.84562558e-01
-7.14367390e-01 -2.21409321e-01 3.22164029e-01 9.11965907e-01
8.67704034e-01 6.31558523e-02 1.92978799e-01 -7.85252869e-01
4.56336617e-01 -7.67722428e-01 -6.86324120e-01 1.76431447e-01
-4.24402863e-01 -3.02005291e-01 3.54079843e-01 -4.29752469e-01
-9.66583967e-01 5.63636839e-01 -1.70150562e-03 -8.43471169e-01
1.82352498e-01 6.95254207e-01 -2.18441501e-01 -8.34078074e-01
4.10754234e-01 4.93914694e-01 4.41005200e-01 -3.33956271e-01
2.38465190e-01 5.18772185e-01 4.26025808e-01 -6.50635481e-01
9.93171215e-01 8.28206003e-01 4.16583806e-01 -1.48190737e+00
-6.82970703e-01 -6.53016210e-01 -3.31398904e-01 -3.80293041e-01
3.38811666e-01 -8.91990840e-01 -7.18681157e-01 4.57455218e-01
-1.08154738e+00 1.50652915e-01 -9.02666524e-02 5.37321568e-01
-5.45108497e-01 7.09841192e-01 -3.24565619e-01 -8.99505258e-01
-2.76128858e-01 -1.14331269e+00 9.99567807e-01 3.82929780e-02
-9.72339138e-02 -9.11986768e-01 5.14875874e-02 6.67806029e-01
5.38601458e-01 4.92940575e-01 3.38481843e-01 -2.25978076e-01
-4.20424879e-01 -6.41894192e-02 4.12642919e-02 2.81065345e-01
-9.07449722e-02 -8.21280107e-02 -7.51172066e-01 -7.99910426e-01
5.22605956e-01 -2.53987402e-01 3.79314154e-01 6.67279959e-01
9.74524796e-01 -8.02947402e-01 -3.00444871e-01 1.17114186e+00
1.55777144e+00 -2.26434246e-01 3.75856459e-01 -8.11166167e-02
6.72384143e-01 6.76074684e-01 4.90229249e-01 9.47884619e-01
5.01214452e-02 5.02335966e-01 5.32609880e-01 3.37744989e-02
7.59066045e-02 2.00184748e-01 3.76988798e-01 8.76463234e-01
-1.60961986e-01 1.69116229e-01 -6.78188384e-01 3.53983790e-01
-1.92109215e+00 -1.10836923e+00 -3.53534818e-01 2.25665426e+00
5.39668322e-01 -6.98843300e-01 -1.47828788e-01 4.00239348e-01
7.25895643e-01 5.84333874e-02 -4.66981739e-01 -3.02006632e-01
-2.18666732e-01 1.79339454e-01 3.70224744e-01 6.38121784e-01
-8.50609064e-01 5.82037270e-01 6.76648426e+00 7.73207486e-01
-1.09621418e+00 2.11695924e-01 4.88799810e-01 -3.99062812e-01
-4.23048250e-02 -4.85837925e-03 -7.79417574e-01 3.84517759e-01
7.52637446e-01 -2.13045433e-01 9.23376381e-01 7.57196605e-01
4.31288689e-01 -1.06487803e-01 -1.07753921e+00 1.65103126e+00
6.90214276e-01 -1.37514365e+00 -1.19419441e-01 2.40935132e-01
1.19425547e+00 -3.18705857e-01 2.15238318e-01 -1.74211174e-01
-2.34294221e-01 -1.32920420e+00 3.91705304e-01 5.02368093e-01
7.45900989e-01 -5.45829952e-01 4.37103093e-01 3.17259699e-01
-1.02817833e+00 -2.67250109e-02 -3.57965291e-01 -9.41563770e-02
3.00956935e-01 9.64431643e-01 -2.34877944e-01 3.90455484e-01
5.23344874e-01 9.12756383e-01 7.00864196e-02 1.01224911e+00
2.60313958e-01 4.85709876e-01 -5.67011297e-01 5.45188010e-01
2.58632675e-02 -6.42095327e-01 9.42745507e-01 8.51093233e-01
6.46947563e-01 6.82039738e-01 2.69406766e-01 7.02805579e-01
5.86729199e-02 4.08270359e-01 -7.10819542e-01 7.59339854e-02
3.52609158e-01 1.29410267e+00 -3.39670867e-01 1.35375988e-02
-6.99933350e-01 7.73266792e-01 -1.04357786e-02 5.21900058e-01
-6.26091599e-01 1.45275861e-01 8.40262592e-01 1.87922955e-01
1.46445081e-01 -4.05606419e-01 -1.58716917e-01 -1.18961668e+00
4.68140580e-02 -1.39900720e+00 3.01246166e-01 -4.33798611e-01
-1.09131539e+00 2.82972038e-01 -9.09077525e-02 -1.14631546e+00
2.35363599e-02 -5.39056599e-01 -5.05279064e-01 8.44119847e-01
-1.47484457e+00 -8.77745628e-01 -3.42879295e-01 1.06836975e+00
4.30170447e-01 -5.52930057e-01 6.57695591e-01 4.95339930e-01
-8.50007176e-01 4.46029305e-01 3.38192642e-01 -2.21646830e-01
4.81930673e-01 -4.69517261e-01 -6.23270452e-01 9.65816855e-01
3.67508382e-02 5.99050641e-01 8.02771568e-01 -5.71446419e-01
-2.34330106e+00 -9.35016155e-01 7.39759922e-01 9.56820473e-02
4.15602744e-01 2.04416171e-01 -9.33404267e-01 6.81688905e-01
-3.27316076e-02 5.22255838e-01 7.73453832e-01 -2.21161276e-01
1.39832459e-02 -2.70182520e-01 -1.49939442e+00 6.65927082e-02
9.37579811e-01 -2.86509037e-01 -2.49865770e-01 7.70683110e-01
-1.50046468e-01 -5.39030194e-01 -8.71751904e-01 3.34968656e-01
4.38708335e-01 -8.34911644e-01 1.15465975e+00 -3.65340352e-01
2.57004291e-01 -3.25510651e-01 -4.38513160e-01 -1.14220548e+00
-3.14253271e-01 -1.05583549e+00 -4.60318297e-01 9.91440356e-01
-2.13609990e-02 -6.52610421e-01 9.39183533e-01 6.33121312e-01
8.17412958e-02 -7.32847571e-01 -1.27461088e+00 -7.03086436e-01
-3.37843746e-01 -2.76263386e-01 3.90958369e-01 1.16456163e+00
-7.24805966e-02 -2.58784264e-01 -7.17286587e-01 4.85132605e-01
1.34130692e+00 1.43716916e-01 6.63316727e-01 -1.03566790e+00
-1.46660596e-01 -9.84088257e-02 -4.25657243e-01 -6.53105378e-01
5.23619592e-01 -1.07478893e+00 -1.23936795e-01 -1.14040971e+00
1.85100824e-01 -4.04554904e-01 5.69876730e-02 3.08198512e-01
2.39556715e-01 4.89603847e-01 1.82155520e-01 2.61893451e-01
-3.25726122e-01 5.63690424e-01 1.07928252e+00 -2.83210874e-01
-1.48600906e-01 -2.24993154e-01 -5.51465273e-01 7.44593203e-01
5.07295609e-01 -4.80019152e-01 -2.56069124e-01 -6.00001872e-01
3.12029440e-02 3.78237396e-01 2.06220478e-01 -9.33672547e-01
5.52370906e-01 -3.92396510e-01 2.64860332e-01 -2.71079361e-01
6.90976024e-01 -9.79436100e-01 5.72817087e-01 5.01253426e-01
1.16529025e-01 -1.15631238e-01 1.18089601e-01 3.53708416e-01
-3.89545679e-01 -3.96789968e-01 9.24520552e-01 -2.02305377e-01
-5.72370768e-01 5.37796319e-01 -1.75981671e-01 7.11814538e-02
1.28071594e+00 -4.13119674e-01 1.42426550e-01 -5.03404200e-01
-7.17789352e-01 6.81658313e-02 2.66739935e-01 -1.61545202e-01
9.06876564e-01 -1.40227330e+00 -1.22789323e+00 5.18493176e-01
-4.17413831e-01 -1.25638172e-01 2.16200456e-01 1.20228851e+00
-6.88270152e-01 3.65668774e-01 -6.62630200e-02 -6.62615359e-01
-1.52976286e+00 3.95590514e-01 3.49927336e-01 3.47796977e-01
-3.58616292e-01 1.00552881e+00 -2.40053892e-01 -2.93720990e-01
3.59744221e-01 4.28654164e-01 1.70110352e-02 -5.03910035e-02
6.80939019e-01 9.51382279e-01 5.52906729e-02 -1.24786508e+00
-5.30877769e-01 9.39966977e-01 3.79143000e-01 5.85872233e-02
1.50924277e+00 -7.96434134e-02 -8.22966218e-01 -2.08017081e-01
1.49867582e+00 -3.97898220e-02 -8.45307350e-01 -2.11223438e-01
-3.06187004e-01 -8.21853518e-01 3.69488806e-01 -1.09043434e-01
-1.53963006e+00 3.32587600e-01 4.91835654e-01 -2.77980804e-01
1.21125412e+00 -4.38097984e-01 6.89867914e-01 2.60814846e-01
4.41901267e-01 -9.81014907e-01 -1.51402742e-01 3.70421141e-01
1.28017986e+00 -1.31639421e+00 5.63091338e-01 -7.43597984e-01
-2.17099383e-01 9.70684528e-01 3.54497135e-01 -8.88312832e-02
1.06349397e+00 4.91116524e-01 -1.39090598e-01 -4.61092629e-02
-4.88231659e-01 2.49737933e-01 2.98062235e-01 4.34968650e-01
3.48713964e-01 -2.60093421e-01 -4.60467070e-01 6.89833835e-02
-1.06260434e-01 1.02510527e-01 3.51184577e-01 9.93992150e-01
-1.72697023e-01 -8.89892817e-01 -1.19762254e+00 4.75278556e-01
-3.58505011e-01 1.35309950e-01 -1.98261172e-01 2.76053101e-01
7.68867731e-02 1.11934268e+00 -2.58660406e-01 1.16345637e-01
6.85970709e-02 -2.75341365e-02 6.36746943e-01 -4.13121909e-01
-2.06645235e-01 5.85123105e-03 -2.68570900e-01 -8.85613024e-01
-6.56852901e-01 -9.77699757e-01 -9.80176270e-01 -6.87533855e-01
-1.62076309e-01 1.89652488e-01 8.52841139e-01 9.33105350e-01
3.83019149e-01 -1.16881453e-01 6.57494009e-01 -1.12623513e+00
-6.67384684e-01 -4.26759124e-01 -7.58673310e-01 3.40309381e-01
3.26184601e-01 -6.19558990e-01 -5.25440812e-01 1.62732854e-01] | [12.510961532592773, 0.3726765215396881] |
99f8f411-1c12-4840-88c7-2c293d35deb9 | a-physics-informed-machine-learning-approach | 2010.02011 | null | https://arxiv.org/abs/2010.02011v1 | https://arxiv.org/pdf/2010.02011v1.pdf | A Physics-Informed Machine Learning Approach for Solving Heat Transfer Equation in Advanced Manufacturing and Engineering Applications | A physics-informed neural network is developed to solve conductive heat transfer partial differential equation (PDE), along with convective heat transfer PDEs as boundary conditions (BCs), in manufacturing and engineering applications where parts are heated in ovens. Since convective coefficients are typically unknown, current analysis approaches based on trial and error finite element (FE) simulations are slow. The loss function is defined based on errors to satisfy PDE, BCs and initial condition. An adaptive normalizing scheme is developed to reduce loss terms simultaneously. In addition, theory of heat transfer is used for feature engineering. The predictions for 1D and 2D cases are validated by comparing with FE results. It is shown that using engineered features, heat transfer beyond the training zone can be predicted. Trained model allows for fast evaluation of a range of BCs to develop feedback loops, realizing Industry 4.0 concept of active manufacturing control based on sensor data. | ['Keith D. Humfeld', 'Navid Zobeiry'] | 2020-09-28 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [ 3.63541573e-01 3.26100662e-02 2.58053783e-02 -4.91382867e-01
-8.83184820e-02 -7.53149688e-02 1.13312081e-01 3.45632255e-01
2.45784566e-01 7.61707127e-01 -4.81935352e-01 -1.11001618e-02
-4.17093933e-01 -8.49309444e-01 -7.38181114e-01 -7.40213215e-01
-1.44610330e-01 3.30206484e-01 -4.26504821e-01 -3.01950812e-01
4.27553982e-01 9.16221857e-01 -1.51485038e+00 -8.65176262e-05
8.31861973e-01 1.58295286e+00 1.33396894e-01 3.70090038e-01
1.50723308e-01 6.41476452e-01 -2.64709979e-01 5.09068847e-01
3.65867019e-01 -3.45184356e-01 -7.87873149e-01 -1.48045182e-01
-2.39863507e-02 -3.99569392e-01 1.22452728e-01 6.94479406e-01
3.52600127e-01 5.06658971e-01 9.58297789e-01 -1.06324840e+00
-8.77910376e-01 1.05443224e-01 -1.07901454e-01 -7.05997586e-01
1.29894819e-02 3.31206881e-02 3.99187624e-01 -9.68414485e-01
3.72294635e-01 8.73710692e-01 1.04769027e+00 7.99293160e-01
-1.07901776e+00 -2.97132909e-01 -3.70594770e-01 -5.06756783e-01
-1.01408541e+00 7.81232417e-02 1.14232445e+00 -5.20451725e-01
1.36620402e+00 4.55505788e-01 7.15701461e-01 4.43011045e-01
1.25621533e+00 3.31107706e-01 1.07623303e+00 -4.81029153e-01
6.35527492e-01 2.82171816e-01 3.02713718e-02 5.43613315e-01
3.64898369e-02 6.99939907e-01 6.61852434e-02 -1.64861735e-02
1.34288323e+00 -5.40886968e-02 -1.31254688e-01 -4.69688833e-01
-4.12106544e-01 7.11506188e-01 5.47036707e-01 2.06206724e-01
-4.92352009e-01 2.85904586e-01 4.78857398e-01 4.26472604e-01
7.42721975e-01 1.06711447e+00 -8.95659924e-01 5.43091856e-02
-8.04246783e-01 3.74182165e-01 1.11814952e+00 8.38468671e-01
8.99433553e-01 3.59672338e-01 6.16293736e-02 1.01102161e+00
5.17954648e-01 3.61449510e-01 4.32335958e-02 -1.26859176e+00
-2.85122871e-01 8.36606264e-01 1.77140057e-01 -6.98589623e-01
-3.79872859e-01 -2.99965233e-01 -9.98062909e-01 8.55656266e-01
-2.03543559e-01 -4.69474405e-01 -1.17844331e+00 1.12234366e+00
3.96767646e-01 -3.55741054e-01 -1.87187538e-01 1.16557109e+00
2.53533453e-01 8.10165465e-01 -2.15165187e-02 -3.38456541e-01
5.10093808e-01 -6.19078934e-01 -8.43085527e-01 9.01885033e-02
8.10868561e-01 -6.62161529e-01 7.47952104e-01 4.04592931e-01
-1.10378146e+00 -7.46640265e-01 -1.28050220e+00 -2.72355624e-03
-4.99355942e-01 7.35661089e-02 8.31539452e-01 4.20494318e-01
-8.69406521e-01 1.51983404e+00 -8.77865672e-01 -1.11826144e-01
2.50889897e-01 6.67315125e-01 1.86272383e-01 2.70923615e-01
-1.30365586e+00 9.74296272e-01 1.91797033e-01 7.01461315e-01
-7.71403193e-01 -1.45174289e+00 -7.65854120e-01 -2.56979793e-01
-2.81804472e-01 -7.08647907e-01 1.19804823e+00 -8.76514733e-01
-2.18957615e+00 4.01374698e-01 4.41821218e-01 -2.34737203e-01
5.78778207e-01 -6.65159881e-01 -3.01002681e-01 -3.10583003e-02
-4.99422014e-01 2.29703739e-01 6.66864038e-01 -1.56137788e+00
-9.51563343e-02 4.86402288e-02 -2.57119358e-01 -5.89148290e-02
4.58753854e-02 -1.92216903e-01 5.73257267e-01 -1.99322060e-01
2.21847594e-01 -7.52583861e-01 -5.35971463e-01 2.74936795e-01
-4.78542894e-01 6.48023188e-02 1.39439833e+00 -8.79416466e-01
5.51787555e-01 -1.61020422e+00 -2.18620449e-01 5.65276861e-01
-3.07402074e-01 3.31694223e-02 3.86940897e-01 7.11013019e-01
-2.95496434e-01 -2.46017292e-01 -4.93631572e-01 9.06815156e-02
2.78340392e-02 2.06797808e-01 -7.14431480e-02 4.56328660e-01
7.35156596e-01 6.88411415e-01 -4.27053392e-01 -1.61182452e-02
8.11320126e-01 6.94772482e-01 -5.50409317e-01 1.99968621e-01
-2.63179153e-01 5.42061865e-01 -5.10444403e-01 6.74054027e-01
9.04328644e-01 1.38660491e-01 -1.18604019e-01 -4.41273063e-01
-4.06647354e-01 -7.03733638e-02 -1.01591730e+00 1.55091512e+00
-1.12537861e+00 2.24442661e-01 7.06172109e-01 -1.29308152e+00
1.60006917e+00 3.25431198e-01 9.51771438e-01 -6.79528654e-01
6.50083125e-01 4.29593712e-01 -2.47913212e-01 -3.65940660e-01
5.20317137e-01 -6.22277081e-01 8.86785313e-02 -1.34556621e-01
-1.25428429e-02 -9.55057740e-01 -3.31904262e-01 -4.06211048e-01
8.79576564e-01 5.16098201e-01 -4.99303281e-01 -8.10977519e-01
5.72413743e-01 1.94216400e-01 5.32617450e-01 1.79401770e-01
9.21382476e-03 5.50323248e-01 7.67245069e-02 -4.78666484e-01
-1.44611812e+00 -7.54260302e-01 -4.97699708e-01 4.28494602e-01
4.79821861e-01 2.70929217e-01 -7.98658788e-01 1.15885697e-01
3.89528632e-01 7.08841324e-01 -5.89004159e-01 -5.89432359e-01
-8.28469217e-01 -4.42284256e-01 -3.47050816e-01 8.51095438e-01
2.98177153e-01 -9.46229100e-01 -5.82153440e-01 5.73414743e-01
8.91966105e-01 -5.07065356e-01 2.81732768e-01 5.62633693e-01
-1.26867545e+00 -9.42433059e-01 -3.59946996e-01 -7.04503953e-01
6.78077996e-01 -5.79137325e-01 1.02127612e+00 3.33067328e-02
-8.95666957e-01 5.37343979e-01 -1.06817596e-01 -5.96415758e-01
-8.89119864e-01 -2.21212193e-01 9.25192833e-02 -7.86715329e-01
-1.31445453e-01 -5.05646706e-01 -8.15719545e-01 4.30396736e-01
-5.66714108e-01 -3.08407452e-02 6.63236856e-01 8.33856046e-01
6.60271585e-01 2.85469949e-01 4.59278286e-01 -5.98065853e-01
7.06530571e-01 -2.13994458e-01 -6.55608535e-01 7.56583437e-02
-9.69361544e-01 -1.88708648e-01 7.51439810e-01 -3.39796662e-01
-1.44298708e+00 9.57680568e-02 -9.15653780e-02 -5.69068611e-01
-3.87074471e-01 4.13929224e-01 1.77662447e-02 -4.06262338e-01
5.32869518e-01 -2.93566495e-01 3.54825944e-01 -4.37166005e-01
-1.43422186e-01 3.51249367e-01 1.49640918e-01 -1.11465335e+00
4.50129092e-01 -9.35059115e-02 4.87408966e-01 -8.83883655e-01
-2.38649338e-01 -1.04772849e-02 -4.78717178e-01 -5.77363729e-01
5.22770762e-01 -4.77560133e-01 -9.78276730e-01 5.75690150e-01
-7.91291058e-01 -6.36562526e-01 -7.36667454e-01 5.83504200e-01
-9.00653005e-01 -1.89492628e-01 -1.23441780e+00 -1.21097732e+00
-7.90194511e-01 -1.00748265e+00 8.47342312e-01 2.80122191e-01
-4.56998795e-01 -1.15430272e+00 -4.93270382e-02 8.27059522e-02
7.88165689e-01 8.35745990e-01 9.90683734e-01 6.03060573e-02
-4.99980360e-01 -5.17240167e-01 7.76888877e-02 1.07239091e+00
3.78225416e-01 5.07482886e-01 -1.00613546e+00 -2.87624598e-01
3.11215997e-01 -4.56135273e-01 4.71330941e-01 8.04697275e-01
1.18840814e+00 -6.43864721e-02 -3.55876088e-01 2.59722918e-01
1.87651002e+00 4.30468410e-01 3.64923328e-01 -1.55136079e-01
6.09756052e-01 7.10811019e-01 6.72375739e-01 4.04502571e-01
-7.14502990e-01 -8.56435671e-02 7.20302880e-01 -3.18392694e-01
2.38388151e-01 8.08991194e-02 -5.36318123e-02 8.68508160e-01
-3.56274635e-01 3.33070904e-01 -6.92785800e-01 3.06069613e-01
-1.32624280e+00 -3.15786004e-01 -4.80096549e-01 2.13858008e+00
8.44779730e-01 2.67554909e-01 -6.22072935e-01 2.53492564e-01
7.31786489e-01 -3.97780150e-01 -7.01535940e-01 -1.17439771e+00
3.79759490e-01 6.36403084e-01 7.14869201e-01 5.57951987e-01
-9.23375607e-01 2.71770149e-01 6.73820591e+00 4.70751852e-01
-1.57633412e+00 -2.51123637e-01 7.36254215e-01 5.54081723e-02
-1.56353414e-01 -4.29415070e-02 -3.41527283e-01 2.73000330e-01
9.84313786e-01 6.95117787e-02 2.76389480e-01 9.67804968e-01
2.95038670e-01 -2.82164425e-01 -1.21474516e+00 4.31087017e-01
-7.56193101e-01 -1.39908528e+00 -2.85012335e-01 -2.76380144e-02
9.33372378e-01 -2.19675586e-01 -2.73975562e-02 1.87331811e-01
2.78744996e-01 -8.06989610e-01 5.39068818e-01 7.22232699e-01
7.86781609e-01 -8.59490335e-01 7.49149203e-01 -2.79339366e-02
-8.85816097e-01 -2.30798379e-01 -3.38887274e-01 -3.73408318e-01
3.19490075e-01 1.11693203e+00 -7.49962926e-01 8.18219244e-01
5.25350809e-01 3.89986664e-01 4.03641433e-01 8.21378589e-01
2.89290279e-01 4.42697555e-01 -5.91236949e-01 -1.82870597e-01
2.03956306e-01 -4.89823580e-01 1.93016976e-01 8.49773407e-01
4.94350612e-01 -1.60690427e-01 -6.44499958e-02 1.64369333e+00
2.77878821e-01 1.66131273e-01 -4.55250472e-01 -6.45209029e-02
3.31172734e-01 1.28611422e+00 -4.72698152e-01 2.77173340e-01
8.70590210e-02 6.31601512e-01 -2.70341516e-01 3.85810882e-01
-8.14529181e-01 -6.66232705e-01 8.48836482e-01 3.47765505e-01
-1.59584761e-01 -4.13645416e-01 -5.64376056e-01 -1.16018623e-01
-9.59314555e-02 2.99034834e-01 -4.77194518e-01 -7.29975939e-01
-1.43561089e+00 -5.98501861e-02 1.87173560e-02 -1.07058370e+00
-1.83258221e-01 -1.20510471e+00 -8.06902826e-01 1.12823915e+00
-1.18733037e+00 -1.00549448e+00 -2.24790685e-02 -6.97071403e-02
2.60670453e-01 2.94740528e-01 8.50788116e-01 4.33835953e-01
-6.06851339e-01 -3.36745121e-02 4.72951055e-01 -2.13324994e-01
2.88555622e-01 -1.11804962e+00 -3.37451957e-02 1.35536209e-01
-1.34865510e+00 5.64109504e-01 1.05422401e+00 -9.48041379e-01
-1.70440102e+00 -1.19517446e+00 2.79829979e-01 2.63794541e-01
4.86860991e-01 -2.64116287e-01 -1.05741167e+00 1.95946157e-01
7.26355761e-02 2.43648186e-01 2.94535339e-01 -1.08258791e-01
5.68499923e-01 -9.80354846e-02 -1.60945976e+00 -1.24328591e-01
6.69333220e-01 -3.45701218e-01 -2.00581342e-01 3.10008109e-01
6.47835314e-01 -5.46704471e-01 -1.64149773e+00 8.03645670e-01
5.09927809e-01 -4.80283082e-01 7.51805246e-01 -4.62203145e-01
9.12017465e-01 2.21226901e-01 1.07373714e-01 -1.33812690e+00
-3.39539409e-01 -5.97569823e-01 -3.46843690e-01 1.33397937e+00
4.42876965e-01 -6.41231656e-01 8.84401679e-01 1.37808037e+00
-8.13265145e-01 -1.30520725e+00 -7.58221269e-01 -8.43470812e-01
2.83004344e-01 -2.48465627e-01 3.97611171e-01 1.20178437e+00
-1.59598738e-01 -2.38119647e-01 2.25750078e-02 3.57547365e-02
3.76703799e-01 -3.29147652e-02 6.74528554e-02 -1.61438239e+00
6.98317289e-02 -5.21906428e-02 -5.56449927e-02 -4.75829452e-01
1.47398442e-01 -4.41799104e-01 4.11843240e-01 -1.61058927e+00
-5.77384591e-01 -9.73590136e-01 -4.62661415e-01 9.21088383e-02
5.51544666e-01 -4.13246274e-01 -3.74928266e-01 -1.07915670e-01
4.87081200e-01 9.21679437e-01 1.74391162e+00 -9.57620051e-03
-4.80249256e-01 -1.35728091e-01 4.19619232e-02 4.44721371e-01
1.11028695e+00 -6.34545088e-02 -4.39076453e-01 -6.89981356e-02
1.35638252e-01 2.23438069e-02 5.31191051e-01 -1.36166131e+00
7.68736843e-03 -3.51207167e-01 8.14438343e-01 -5.38460493e-01
4.37706709e-01 -1.12956548e+00 3.24109972e-01 5.53403497e-01
-3.95825535e-01 -3.17442000e-01 5.88938355e-01 2.29176924e-01
-1.88184619e-01 -1.51171565e-01 9.14237738e-01 -1.71071514e-01
-4.21275347e-01 2.61459857e-01 -2.44254395e-01 -4.93513316e-01
1.05870581e+00 -6.52769327e-01 3.35028432e-02 2.02636376e-01
-7.13607728e-01 4.21195813e-02 3.97435904e-01 1.57101840e-01
6.20925784e-01 -1.60354042e+00 -2.44493425e-01 3.69987607e-01
-4.77905452e-01 3.73350441e-01 7.20110536e-01 6.70830369e-01
-1.08285022e+00 8.41576606e-02 -4.51794356e-01 -4.73640114e-01
-4.52694416e-01 4.92527366e-01 8.93498361e-01 -1.70712210e-02
-3.68589699e-01 1.01981246e+00 -1.82071775e-01 -8.21416020e-01
-2.83247650e-01 -7.20647335e-01 5.60404241e-01 -3.76466602e-01
-8.45069811e-02 6.23054802e-01 4.43359405e-01 -1.47969767e-01
-1.05283186e-01 5.72120011e-01 1.75275058e-01 3.37439328e-01
1.57412839e+00 1.02775596e-01 -2.76971877e-01 5.27420521e-01
1.61372197e+00 -7.38986194e-01 -1.66116142e+00 3.89837921e-01
-3.78135175e-01 -7.59779289e-02 5.05815685e-01 -9.99533474e-01
-1.30733466e+00 7.12507784e-01 9.60777581e-01 2.66936421e-01
1.09306931e+00 -4.47495937e-01 9.08882439e-01 2.56332368e-01
1.01670258e-01 -2.04106832e+00 -2.60552675e-01 4.60682869e-01
1.05934966e+00 -7.53005207e-01 1.48838758e-01 -7.41266429e-01
-1.15083589e-03 1.36042488e+00 8.54803860e-01 -6.08255267e-01
1.04859507e+00 8.11906755e-01 -4.96298186e-02 -2.27903038e-01
-3.52205724e-01 7.90405631e-01 1.39364041e-02 3.92207026e-01
7.83788502e-01 -2.15175538e-03 -1.71884447e-01 8.86783525e-02
1.33503631e-01 4.39261913e-01 -1.26077101e-01 1.63601899e+00
-3.79936129e-01 -9.25790548e-01 -4.29943532e-01 6.24425709e-01
-2.35098809e-01 4.04543042e-01 -2.49348253e-01 1.15684319e+00
2.04104558e-01 6.07306480e-01 3.62219721e-01 -4.70188498e-01
4.51120287e-01 1.27700642e-02 7.70753920e-01 -1.98681742e-01
-7.79564142e-01 -2.09918961e-01 1.51869252e-01 -6.39964759e-01
-1.53618172e-01 -2.37663478e-01 -1.66496992e+00 -2.74089277e-01
-5.69747567e-01 1.16879776e-01 1.11286390e+00 6.95946038e-01
5.18847048e-01 9.11963940e-01 8.45587611e-01 -1.20153165e+00
-5.54363310e-01 -9.78081703e-01 -1.12377656e+00 3.08923483e-01
1.50899082e-01 -9.73434925e-01 -2.96376318e-01 -6.87086508e-02] | [6.2967963218688965, 3.3331828117370605] |
dbda35f7-3aa6-4eb7-b7e7-30daa3629c5c | unifying-large-language-models-and-knowledge | 2306.08302 | null | https://arxiv.org/abs/2306.08302v2 | https://arxiv.org/pdf/2306.08302v2.pdf | Unifying Large Language Models and Knowledge Graphs: A Roadmap | Large language models (LLMs), such as ChatGPT and GPT4, are making new waves in the field of natural language processing and artificial intelligence, due to their emergent ability and generalizability. However, LLMs are black-box models, which often fall short of capturing and accessing factual knowledge. In contrast, Knowledge Graphs (KGs), Wikipedia and Huapu for example, are structured knowledge models that explicitly store rich factual knowledge. KGs can enhance LLMs by providing external knowledge for inference and interpretability. Meanwhile, KGs are difficult to construct and evolving by nature, which challenges the existing methods in KGs to generate new facts and represent unseen knowledge. Therefore, it is complementary to unify LLMs and KGs together and simultaneously leverage their advantages. In this article, we present a forward-looking roadmap for the unification of LLMs and KGs. Our roadmap consists of three general frameworks, namely, 1) KG-enhanced LLMs, which incorporate KGs during the pre-training and inference phases of LLMs, or for the purpose of enhancing understanding of the knowledge learned by LLMs; 2) LLM-augmented KGs, that leverage LLMs for different KG tasks such as embedding, completion, construction, graph-to-text generation, and question answering; and 3) Synergized LLMs + KGs, in which LLMs and KGs play equal roles and work in a mutually beneficial way to enhance both LLMs and KGs for bidirectional reasoning driven by both data and knowledge. We review and summarize existing efforts within these three frameworks in our roadmap and pinpoint their future research directions. | ['Xindong Wu', 'Jiapu Wang', 'Chen Chen', 'YuFei Wang', 'Linhao Luo', 'Shirui Pan'] | 2023-06-14 | null | null | null | null | ['knowledge-graphs', 'text-generation'] | ['knowledge-base', 'natural-language-processing'] | [-1.59683645e-01 9.67971087e-01 -5.39791822e-01 -6.54289126e-02
-2.37047240e-01 -6.31287098e-01 8.55303943e-01 4.60022599e-01
1.05780903e-02 9.41524208e-01 6.07764184e-01 -6.22383296e-01
-3.58080238e-01 -1.55814159e+00 -8.87614906e-01 -1.15522936e-01
3.41189541e-02 4.91656452e-01 1.03885569e-01 -3.83260220e-01
1.21582802e-02 -1.66664850e-02 -1.23984110e+00 2.39598647e-01
1.29119384e+00 6.10833526e-01 6.52055666e-02 4.05624211e-01
-8.91090930e-01 1.41979170e+00 -2.68089026e-01 -1.07424009e+00
-1.82126686e-01 -1.53378740e-01 -1.32836604e+00 -3.61723095e-01
3.20056111e-01 -8.12879428e-02 -7.25016892e-01 8.58425736e-01
2.02333286e-01 -2.63151787e-02 4.34881866e-01 -1.55354631e+00
-1.34982324e+00 1.25554991e+00 -4.82517257e-02 -1.14188127e-01
4.72206265e-01 1.21214174e-01 1.25568664e+00 -8.05840015e-01
9.39768732e-01 1.37993610e+00 7.26932526e-01 5.58994055e-01
-7.90556431e-01 -4.75715637e-01 5.11263430e-01 5.80973566e-01
-1.20314527e+00 -2.10128754e-01 8.06074142e-01 -2.84505695e-01
1.26539183e+00 8.44738483e-02 9.55215454e-01 1.13006628e+00
-1.75513059e-01 1.28275359e+00 8.53656709e-01 -5.73704004e-01
-5.10434657e-02 2.27573797e-01 2.01986358e-01 1.06137359e+00
6.58239484e-01 -8.04555118e-02 -9.54270422e-01 -2.67432839e-01
8.71836782e-01 -1.39128983e-01 -3.84819537e-01 -2.17004955e-01
-1.37096202e+00 8.63620460e-01 5.46222568e-01 2.83168644e-01
-5.43940127e-01 2.93631643e-01 8.45364779e-02 2.63954014e-01
3.68357807e-01 7.50052273e-01 -6.06855989e-01 -2.25593388e-01
-5.69266677e-01 3.62138778e-01 1.17130196e+00 1.09306681e+00
8.60790372e-01 1.62000746e-01 -9.86503810e-02 5.70489526e-01
4.97002691e-01 4.24034745e-01 3.62324297e-01 -8.42196226e-01
8.37728143e-01 1.20585954e+00 -1.02140382e-01 -1.16747057e+00
-2.40497768e-01 -5.24293244e-01 -8.18000317e-01 -5.65421045e-01
1.46714300e-01 -5.31406812e-02 -7.91145444e-01 1.94116318e+00
4.43143338e-01 3.50000709e-01 5.58158338e-01 4.73515600e-01
1.45522690e+00 7.69172490e-01 1.97757274e-01 1.96645856e-01
1.14372826e+00 -1.21809125e+00 -8.49952996e-01 -4.50423121e-01
8.29536319e-01 -1.05833091e-01 1.31701863e+00 5.90375625e-02
-1.08122420e+00 -3.13418746e-01 -9.22613680e-01 -2.93114930e-01
-1.03839827e+00 -2.29892030e-01 1.20407808e+00 3.34686369e-01
-1.23408270e+00 4.80898827e-01 -5.56261361e-01 -2.45699763e-01
5.76387346e-01 5.47406636e-02 -3.30141574e-01 -3.96179378e-01
-2.05367565e+00 1.05813766e+00 9.27378595e-01 3.52604091e-01
-6.08824849e-01 -9.73399818e-01 -1.24005580e+00 1.21851146e-01
9.18714643e-01 -1.42871654e+00 8.61844778e-01 -3.95816565e-01
-1.36150491e+00 6.17040634e-01 -1.24404415e-01 -4.77155715e-01
1.61733538e-01 -2.28634849e-01 -6.70099258e-01 -1.03602722e-01
1.11965626e-01 5.44058740e-01 4.87141162e-01 -1.34655440e+00
-3.06418628e-01 -2.15242952e-01 5.93698621e-01 3.28722566e-01
-5.08837223e-01 -5.13259649e-01 -5.43723404e-01 -4.52288717e-01
-1.78131968e-01 -3.72305602e-01 -2.22988486e-01 -3.80558074e-01
-5.62540650e-01 -4.43330467e-01 6.85693383e-01 -7.41729736e-01
1.71532309e+00 -1.53362143e+00 2.51768440e-01 2.33526155e-01
7.95576394e-01 5.64788997e-01 -1.88679099e-01 9.45222914e-01
3.46120805e-01 4.96221125e-01 -7.31398240e-02 -5.84441982e-02
3.95153910e-01 7.35921264e-01 -5.85451484e-01 -5.12534797e-01
3.99946272e-01 2.01132631e+00 -1.44188201e+00 -5.13594389e-01
2.01233044e-01 2.00556636e-01 -3.57916683e-01 -5.54436585e-03
-6.22862399e-01 8.53598863e-02 -6.41992390e-01 6.59586787e-01
2.02532396e-01 -7.09435582e-01 4.47122991e-01 -2.40259841e-01
2.59935886e-01 4.62344140e-01 -1.14482737e+00 1.62217653e+00
-5.67525625e-01 3.48547995e-01 -2.55532533e-01 -7.40462720e-01
7.82611430e-01 3.62322867e-01 -1.06538922e-01 -3.99627358e-01
-3.27363253e-01 1.15141951e-01 -3.11116487e-01 -7.00508654e-01
7.95625925e-01 -1.39841869e-01 5.05703874e-02 5.57851195e-01
2.88736194e-01 -2.51275182e-01 3.74741644e-01 9.33397353e-01
1.02233839e+00 2.30834067e-01 5.27987301e-01 5.98223060e-02
4.53343421e-01 -6.19421117e-02 3.03264946e-01 8.64340723e-01
1.98014960e-01 -1.73652828e-01 5.20501435e-01 -4.48644400e-01
-5.77021956e-01 -1.06461823e+00 3.35345179e-01 7.14728117e-01
1.14394359e-01 -9.64318514e-01 -3.78990710e-01 -9.80842471e-01
2.97794282e-01 1.04413736e+00 -5.71688116e-01 -2.46756360e-01
-3.16642165e-01 -7.77908683e-01 9.73137200e-01 8.34139824e-01
7.59990394e-01 -1.24883068e+00 -3.48225832e-02 2.90571541e-01
-4.90752220e-01 -1.23147237e+00 1.24672867e-01 -3.85690659e-01
-8.14185083e-01 -1.27872479e+00 -1.70991823e-01 -5.32501817e-01
7.61962175e-01 1.25903010e-01 1.63714242e+00 2.66794533e-01
1.32905424e-01 7.97759831e-01 -5.55761695e-01 -5.53091824e-01
-3.90345663e-01 1.84178934e-01 -2.87520513e-02 -2.80726999e-01
3.90605211e-01 -8.13593864e-01 -3.82188231e-01 -2.14650229e-01
-1.19088984e+00 4.53741044e-01 7.66255438e-01 7.62316287e-01
3.18448603e-01 1.11085445e-01 9.66370523e-01 -1.28421867e+00
1.04627883e+00 -8.12747121e-01 -1.15937412e-01 7.66918600e-01
-9.39954519e-01 2.92211771e-01 6.30924106e-01 -8.94087851e-02
-1.22574389e+00 -8.24575603e-01 -1.75282747e-01 -1.84226871e-01
2.91477472e-01 1.36329126e+00 -2.28070378e-01 -1.11169972e-01
6.07647777e-01 3.79727632e-01 -2.21968666e-01 -2.67593324e-01
1.18525422e+00 4.03706670e-01 5.43475986e-01 -1.05629253e+00
9.62519944e-01 3.02773148e-01 -2.41823509e-01 -4.33881432e-01
-1.05551028e+00 -2.26647407e-01 -3.80732805e-01 -1.34213984e-01
4.12032843e-01 -8.34690928e-01 -5.70028305e-01 4.20087308e-01
-1.14634073e+00 -3.76550555e-01 -6.81653261e-01 1.64911240e-01
-1.69544354e-01 6.16994321e-01 -6.73290670e-01 -6.34399295e-01
-6.21556103e-01 -5.24826467e-01 8.08462143e-01 4.22151536e-01
-1.03430167e-01 -1.80295324e+00 -1.12723134e-01 7.33519077e-01
6.54408872e-01 2.55855918e-01 1.39927292e+00 -4.99755859e-01
-7.80226946e-01 -1.21356897e-01 -3.05540979e-01 2.98275501e-01
1.70944214e-01 -1.99127674e-01 -7.72007942e-01 7.80245364e-02
-4.28098738e-01 -5.76965332e-01 7.35227227e-01 -8.58773887e-02
1.11700439e+00 -7.87607908e-01 -3.18757534e-01 4.44912106e-01
1.32975459e+00 -2.30192900e-01 7.60293782e-01 3.32602233e-01
9.49477434e-01 5.42491019e-01 2.00822532e-01 2.90445499e-02
1.28642178e+00 2.27593720e-01 3.77418935e-01 -6.68093264e-02
-2.90543735e-01 -1.10270727e+00 3.00149143e-01 1.41025138e+00
-4.66501176e-01 -2.18826607e-01 -1.10912061e+00 7.77653754e-01
-2.13888955e+00 -9.02797937e-01 -2.68300354e-01 1.73444390e+00
1.18837142e+00 -8.51434767e-02 -5.41668415e-01 -1.33838177e-01
3.77217740e-01 3.72886509e-01 -6.02899551e-01 -1.89037710e-01
-4.98316914e-01 3.39594334e-01 3.68194915e-02 7.67515659e-01
-5.45411706e-01 1.40356112e+00 5.56134987e+00 9.14977133e-01
-4.90859807e-01 6.04426153e-02 1.06817439e-01 3.63173395e-01
-9.45557117e-01 4.26658213e-01 -7.18121827e-01 9.68072340e-02
7.16090858e-01 -4.98974621e-01 4.23742801e-01 7.39864588e-01
-2.89000511e-01 9.35122743e-02 -8.84027183e-01 7.91170895e-01
3.77439037e-02 -1.99772894e+00 7.34176338e-01 -1.37108386e-01
9.95233893e-01 -9.04529020e-02 -3.15092891e-01 8.36925030e-01
8.63597631e-01 -1.05265772e+00 3.61345679e-01 6.96968496e-01
4.49396193e-01 -3.01220208e-01 8.26880038e-01 4.51815844e-01
-1.27957141e+00 1.50271341e-01 -1.68022677e-01 -2.57380068e-01
4.68690932e-01 7.39378870e-01 -7.95405924e-01 1.47861409e+00
3.17876786e-01 8.44204009e-01 -6.15513504e-01 4.40580577e-01
-1.02398002e+00 6.96398556e-01 1.15008596e-02 -1.35535719e-02
3.92292798e-01 -1.29171759e-01 3.41087937e-01 1.13209653e+00
6.32855669e-03 1.26074582e-01 1.52698025e-01 1.31216335e+00
-3.21664125e-01 -9.96580161e-03 -6.23730004e-01 -6.30654931e-01
6.76818013e-01 1.28768790e+00 -1.22635469e-01 -7.98467815e-01
-6.32229984e-01 5.91988802e-01 7.10733652e-01 7.11089075e-01
-5.49407601e-01 -5.11182725e-01 3.56481314e-01 -4.02562320e-03
-1.05476025e-02 -2.80535370e-01 -9.46567729e-02 -1.61886990e+00
1.71363071e-01 -6.91545486e-01 5.88970125e-01 -1.05686414e+00
-1.53984594e+00 4.38810766e-01 5.56300767e-02 -4.57923532e-01
-4.21113670e-01 -5.86885095e-01 -4.78334129e-01 8.43311250e-01
-2.09173679e+00 -1.72048736e+00 -3.73678446e-01 7.40971386e-01
3.94532308e-02 1.40003726e-01 8.41065526e-01 5.35545424e-02
-3.37608576e-01 3.47771943e-01 -1.81959465e-01 1.77187935e-01
2.19339296e-01 -1.31666863e+00 6.90570593e-01 7.41274118e-01
5.30352414e-01 1.13366175e+00 1.13574497e-01 -9.91705775e-01
-1.69677770e+00 -1.31691325e+00 1.34169972e+00 -9.41115141e-01
1.07211733e+00 -3.30675244e-01 -1.16595626e+00 1.06382394e+00
7.00559095e-02 -2.90252835e-01 7.92445660e-01 6.01225317e-01
-5.88486195e-01 1.06179371e-01 -7.12136686e-01 8.50648701e-01
1.40159500e+00 -7.05567002e-01 -1.09374821e+00 3.85172278e-01
1.00669420e+00 -3.58380884e-01 -1.16842830e+00 3.79594326e-01
2.73953438e-01 -5.59156597e-01 9.40563679e-01 -1.05568695e+00
6.74796104e-01 -3.53464633e-01 1.08956166e-01 -1.46648920e+00
-2.83817559e-01 -7.61530280e-01 -9.91418839e-01 1.38610816e+00
6.10479832e-01 -1.05357528e+00 7.17355728e-01 7.41746783e-01
-2.73418367e-01 -1.03701758e+00 -5.89334130e-01 -7.00798035e-01
-5.80449626e-02 -7.56594837e-01 8.84197295e-01 1.27940118e+00
4.18145210e-01 5.76244652e-01 -2.24710628e-01 2.04407461e-02
3.38893801e-01 2.99864411e-01 8.65221739e-01 -1.09635353e+00
-2.86595643e-01 -2.33809084e-01 -1.39423788e-01 -1.03293312e+00
3.51230025e-01 -1.37563705e+00 -5.87328494e-01 -2.33156848e+00
2.97321707e-01 -4.62965131e-01 -1.69645399e-01 9.71007586e-01
-6.20885313e-01 -3.81490648e-01 1.18051104e-01 1.99423879e-02
-7.33381629e-01 7.73617029e-01 1.41293240e+00 -8.78338367e-02
-1.83392406e-01 -5.06945610e-01 -1.25382566e+00 6.73096538e-01
5.22374630e-01 1.27926320e-01 -8.74096513e-01 -6.59374952e-01
1.06124854e+00 -2.19262958e-01 6.60716593e-01 -3.73087943e-01
4.75363880e-01 -2.69059420e-01 -4.35015932e-02 -4.75175530e-02
1.69983596e-01 -3.63458306e-01 1.15181461e-01 1.44598752e-01
-1.11378781e-01 -6.69594109e-02 3.07769001e-01 7.06263900e-01
-5.93552589e-01 -8.55614692e-02 -2.28039771e-01 -4.70352590e-01
-1.22567332e+00 4.80699211e-01 -6.71226764e-03 4.22512114e-01
5.44579089e-01 -1.47274375e-01 -8.79155219e-01 -7.08797038e-01
-4.43141431e-01 6.51181936e-01 -1.69206280e-02 6.56231225e-01
7.12880731e-01 -1.37392986e+00 -6.35066390e-01 -7.63123296e-03
3.52455586e-01 2.68633544e-01 4.28454101e-01 8.86853099e-01
-1.30255148e-01 6.77442908e-01 2.33605236e-01 1.47402555e-01
-7.05861270e-01 5.31031847e-01 -3.06180920e-02 -8.62496793e-01
-4.10243034e-01 8.02529752e-01 9.69395339e-02 -9.19092894e-01
-1.56177312e-01 -4.25277442e-01 -2.37761825e-01 -1.81262180e-01
3.88569027e-01 2.63953537e-01 -1.00111336e-01 -9.39650312e-02
-8.65702480e-02 2.10353464e-01 -3.27423103e-02 1.55925497e-01
1.31890309e+00 -8.83549526e-02 -5.93222082e-01 2.95189142e-01
5.00231504e-01 2.88749635e-02 -6.49206161e-01 -6.48762643e-01
5.35550527e-02 -2.35867396e-01 -1.39283255e-01 -1.18736088e+00
-8.37227166e-01 8.02294016e-01 -6.33777916e-01 1.36593193e-01
8.94936979e-01 3.39991301e-01 9.82258141e-01 5.77094138e-01
6.13574326e-01 -9.08649027e-01 1.16060741e-01 7.24413097e-01
9.72099423e-01 -9.29680884e-01 -1.32879958e-01 -6.22585177e-01
-7.45920599e-01 9.97792065e-01 7.85428822e-01 5.27042687e-01
4.20612186e-01 4.24666554e-02 -2.09158123e-01 -5.56715071e-01
-1.04832411e+00 -2.75128186e-01 3.18083227e-01 7.42541730e-01
1.67055428e-01 1.31379396e-01 -2.17508927e-01 1.01943409e+00
-4.79757369e-01 4.32294697e-01 1.37411058e-01 9.46020842e-01
-1.19623899e-01 -1.30152953e+00 1.14938870e-01 7.05918014e-01
6.00270778e-02 -6.68716431e-01 -6.77547574e-01 9.48541105e-01
2.27968529e-01 7.90091097e-01 -6.59919322e-01 -5.67449331e-01
3.04850996e-01 3.80728066e-01 4.76765275e-01 -6.39623702e-01
-4.08079892e-01 -1.06557381e+00 5.80671906e-01 -3.86587262e-01
-4.23076957e-01 1.18022806e-04 -1.20332503e+00 -4.76396322e-01
-4.02710676e-01 3.72827590e-01 3.38784605e-01 1.08092952e+00
7.39687204e-01 3.82178307e-01 -1.84680432e-01 -9.97551605e-02
-2.83747464e-01 -8.21707487e-01 -5.07840335e-01 3.83051693e-01
-2.86575079e-01 -5.48412919e-01 -1.20301068e-01 -2.20684502e-02] | [9.467732429504395, 8.093534469604492] |
d96eaf39-b48d-48e5-ab1b-3ce2f0bdcd99 | financial-dynamics-economic-state | 2203.15911 | null | https://arxiv.org/abs/2203.15911v4 | https://arxiv.org/pdf/2203.15911v4.pdf | Economic state classification and portfolio optimisation with application to stagflationary environments | Motivated by the current fears of a potentially stagflationary global economic environment, this paper uses new and recently introduced mathematical techniques to study multivariate time series pertaining to country inflation (CPI), economic growth (GDP) and equity index behaviours. We begin by assessing the temporal evolution among various economic phenomena, and complement this analysis with `economic driver analysis,' where we decouple country economic trajectories and determine what is most important in their association. Next, we study the temporal self-similarity of global inflation, growth and equity index returns to identify the most anomalous historic periods, and windows in the past that are most similar to current market dynamics. We then introduce a new algorithm to construct economic state classifications and compute an economic state integral, where countries are determined to belong in one of four candidate states based on their inflation and growth behaviours. Finally, we implement a decade-by-decade portfolio optimisation to determine which equity indices and portfolio assets have been most beneficial in maximising portfolio risk-adjusted returns in various market conditions. This could be of great interest to those looking for asset allocation guidance in the current period of high economic uncertainty. | ['Max Menzies', 'Kevin Chin', 'Nick James'] | 2022-03-29 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-2.87731707e-01 -4.73117493e-02 -1.10700667e-01 1.14579983e-02
-4.04767960e-01 -7.32137740e-01 1.17687225e+00 2.20216006e-01
-3.00200284e-01 7.99698830e-01 5.30743122e-01 -1.06741774e+00
-6.98125184e-01 -9.81344819e-01 2.97135245e-02 -7.76707292e-01
-3.39337617e-01 4.15631503e-01 -1.55693218e-01 -2.97070563e-01
5.86558104e-01 5.19474983e-01 -1.33585715e+00 -4.89742249e-01
8.39692354e-01 9.52027142e-01 -2.75336355e-01 6.49164438e-01
-2.51767226e-02 5.43402851e-01 -5.45430422e-01 -5.45684218e-01
3.04973215e-01 -4.61191863e-01 -4.57222074e-01 -3.82132083e-01
-4.38008457e-01 1.31789148e-01 3.49512845e-01 1.23837745e+00
3.16796392e-01 1.68886572e-01 7.22379744e-01 -7.79595673e-01
-2.93127716e-01 6.51893675e-01 -5.25246680e-01 1.07712722e+00
1.20639689e-01 7.51794204e-02 8.56673181e-01 -6.35284841e-01
2.22410664e-01 1.12297916e+00 5.26724637e-01 -1.59201398e-01
-1.42555106e+00 -5.14444172e-01 4.22916114e-02 -1.90276474e-01
-8.72400761e-01 -5.07884145e-01 7.85036504e-01 -7.21771836e-01
1.02349138e+00 4.70421106e-01 6.44688427e-01 5.77221990e-01
5.29978395e-01 -1.78607851e-01 1.11219192e+00 -5.80517113e-01
1.05853088e-01 1.26659656e-02 -1.69918895e-01 -4.74866480e-02
4.21524882e-01 6.43406391e-01 -1.13809995e-01 -3.75036448e-01
6.12212956e-01 -2.96990812e-01 -1.25005990e-01 5.15661873e-02
-1.38904786e+00 9.55614984e-01 -5.45411892e-02 4.92425919e-01
-8.44685793e-01 -5.62213250e-02 3.35491627e-01 6.17178559e-01
8.14859211e-01 5.06629586e-01 -6.05283380e-01 -3.80188257e-01
-9.59679902e-01 4.82411742e-01 9.10542846e-01 5.44492081e-02
1.66016817e-01 3.78997684e-01 2.50273049e-01 4.52111185e-01
3.86472553e-01 7.86887348e-01 4.88978803e-01 -9.61796641e-01
5.73395848e-01 1.06853142e-01 3.71474892e-01 -1.07035732e+00
-4.74493682e-01 -6.09426618e-01 -7.37051368e-01 4.63917941e-01
7.47015536e-01 -3.01370144e-01 -2.15089425e-01 1.79793537e+00
3.20381552e-01 -1.02316424e-01 4.59818333e-01 4.16377574e-01
-3.35204571e-01 8.14218163e-01 4.81265113e-02 -8.88441265e-01
1.41910088e+00 -2.03285351e-01 -4.07047123e-01 2.16759399e-01
3.96465003e-01 -9.40076590e-01 2.73563176e-01 1.74911931e-01
-1.30704975e+00 -1.76431552e-01 -4.31113511e-01 9.37777340e-01
-3.83426994e-01 -3.84987324e-01 4.58065659e-01 7.10362434e-01
-8.95410419e-01 9.02048290e-01 -9.00144100e-01 8.89644176e-02
-2.44829699e-01 7.13796020e-02 3.98684114e-01 1.07057917e+00
-1.50588620e+00 1.20636415e+00 4.64881390e-01 8.38815644e-02
-4.34181243e-01 -6.92230105e-01 -6.47012293e-01 8.18849206e-02
-7.22633163e-03 -5.90859532e-01 1.07993722e+00 -8.17020655e-01
-1.02891934e+00 6.00104809e-01 1.68675601e-01 -9.33100820e-01
5.52542329e-01 3.56460690e-01 -8.88450921e-01 -7.04663713e-03
4.29974683e-02 -2.39455774e-01 5.66246331e-01 -7.46984661e-01
-1.02541864e+00 -4.74910945e-01 -1.55407786e-01 3.90806645e-01
-1.13397978e-01 8.03870738e-01 4.21480030e-01 -1.22137129e+00
7.85271749e-02 -7.80763805e-01 -4.52678502e-01 -1.00415373e+00
-6.01235852e-02 -1.98300704e-01 2.43698418e-01 -1.03193414e+00
1.62642419e+00 -1.66631985e+00 3.11383009e-02 6.77291751e-01
-1.19013660e-01 -9.25259069e-02 6.25849545e-01 3.83332312e-01
-5.27158976e-01 3.32490087e-01 -2.83714205e-01 1.31976187e-01
1.86870560e-01 -7.13557079e-02 -6.29849494e-01 5.59420764e-01
2.32804716e-01 5.71897864e-01 -8.24877739e-01 -2.49864184e-03
2.28702933e-01 7.47772381e-02 -2.17139259e-01 3.09044612e-03
6.63224384e-02 6.85069680e-01 -3.63723010e-01 6.17787302e-01
3.81620109e-01 1.65007040e-01 -3.17165032e-02 3.30029815e-01
-8.36084902e-01 4.57295179e-01 -1.04972136e+00 5.51114500e-01
-3.72901887e-01 3.96523565e-01 -5.26324809e-02 -1.24378490e+00
1.04156518e+00 5.49495876e-01 6.72868848e-01 -8.60089839e-01
1.35941887e-02 6.39846325e-01 3.21204901e-01 -2.64169306e-01
4.58450496e-01 -4.72845912e-01 -9.70888585e-02 7.93326139e-01
-4.23444390e-01 6.17314223e-03 3.79067779e-01 -2.19170615e-01
5.72396815e-01 -1.76394731e-01 4.51476097e-01 -9.69219446e-01
6.41603708e-01 -2.48258144e-01 5.14872789e-01 2.46219128e-01
-1.89416066e-01 -7.78937861e-02 9.38853443e-01 -5.92880189e-01
-1.25954843e+00 -1.11123478e+00 -3.97944957e-01 7.59494185e-01
-4.46915329e-01 2.45901927e-01 -2.50320643e-01 -1.18573867e-02
6.46338239e-02 1.06083739e+00 -6.16917253e-01 -5.86141646e-02
-7.25768507e-01 -1.58461249e+00 3.30473147e-02 1.16304327e-02
1.60597354e-01 -1.18128347e+00 -1.16942799e+00 5.04093468e-01
3.82486545e-02 -1.50748104e-01 -1.02753118e-02 2.65648872e-01
-9.90471244e-01 -9.20218349e-01 -1.12486279e+00 -2.91216582e-01
4.52027768e-01 -4.85352486e-01 1.32588220e+00 -5.48699796e-01
2.33681455e-01 2.29670450e-01 -1.57493085e-01 -8.87533784e-01
-6.76850975e-01 -2.34198079e-01 3.24772924e-01 6.27574101e-02
2.16673940e-01 -6.27247214e-01 -7.89412975e-01 1.37736425e-01
-6.19840503e-01 -3.52006346e-01 -1.26293167e-01 5.31111419e-01
1.75890759e-01 6.15220010e-01 8.97739708e-01 -3.13288391e-01
9.09933150e-01 -8.98395777e-01 -1.07730639e+00 2.25826398e-01
-1.19247258e+00 -4.84660044e-02 2.06332192e-01 -1.87352866e-01
-1.34912133e+00 -7.15967834e-01 -2.01974269e-02 1.07615978e-01
1.75077647e-01 1.01346099e+00 3.98630470e-01 4.03014064e-01
2.86617935e-01 1.73136164e-02 6.10863380e-02 -5.80311894e-01
2.52679475e-02 3.32736015e-01 8.82011592e-01 -7.56942630e-01
8.55874598e-01 4.63839293e-01 -3.17464881e-02 -2.63129711e-01
-3.56228739e-01 -4.73282248e-01 -3.81801248e-01 -2.73543507e-01
6.86425626e-01 -8.31446469e-01 -5.44625878e-01 6.21911168e-01
-9.31349516e-01 9.45526466e-04 -4.33937043e-01 7.67959774e-01
-6.29274130e-01 1.07303552e-01 -4.11256582e-01 -1.36355948e+00
-5.36776185e-01 -1.05615377e+00 2.41141170e-01 3.73142004e-01
-3.73384535e-01 -1.57946229e+00 5.94353020e-01 -2.89003909e-01
7.16538191e-01 7.97724843e-01 9.53325093e-01 -6.12634003e-01
1.36845261e-01 4.94470224e-02 -2.92038601e-02 1.11106597e-01
1.53155491e-01 2.27106169e-01 -4.81350541e-01 -4.64256316e-01
6.18680716e-01 5.10518670e-01 5.78804493e-01 7.18172431e-01
2.79266000e-01 -7.02719033e-01 -5.23630865e-02 6.93254113e-01
1.53705144e+00 8.22291732e-01 2.75306731e-01 1.14375019e+00
-3.02168597e-02 9.76378798e-01 5.39444745e-01 5.62052488e-01
1.89385399e-01 6.02889836e-01 4.74211335e-01 -1.03965148e-01
6.87240720e-01 1.00075103e-01 5.54840565e-01 7.35005021e-01
-6.53587222e-01 2.29133904e-01 -1.30403113e+00 1.01194179e+00
-1.46848965e+00 -1.29193807e+00 -4.01704460e-02 2.31037283e+00
6.32767975e-01 5.68684161e-01 6.93643391e-01 1.65009573e-01
6.81123853e-01 2.31089965e-01 -2.72310019e-01 -8.62707317e-01
-2.73754835e-01 2.41598353e-01 9.25572634e-01 3.80113363e-01
-1.02462614e+00 2.38643065e-01 6.80267572e+00 3.41334850e-01
-9.87672210e-01 -2.57037640e-01 1.22666883e+00 1.02228001e-01
-5.78794599e-01 2.41372615e-01 -6.58422530e-01 7.66077161e-01
1.66876936e+00 -1.12297487e+00 1.50136143e-01 4.67909008e-01
7.95746505e-01 -1.12328090e-01 -3.84157628e-01 3.14766705e-01
-5.23641706e-01 -1.16024566e+00 -2.40082905e-01 5.40926754e-01
6.55306399e-01 5.08658290e-02 3.33802283e-01 -3.65944467e-02
5.45897067e-01 -7.58625805e-01 9.54142988e-01 7.64299750e-01
4.61410880e-01 -1.23797226e+00 6.16573691e-01 2.50544429e-01
-1.21803868e+00 -4.89780486e-01 -2.68314391e-01 -2.53912866e-01
6.82256758e-01 8.37817848e-01 -3.30020964e-01 6.82425022e-01
7.16589928e-01 3.72488797e-01 -2.13726625e-01 7.77637184e-01
2.31611893e-01 7.42419064e-01 -4.30873364e-01 4.11411375e-01
4.85459179e-01 -8.33978236e-01 9.26290512e-01 9.02227819e-01
7.51321077e-01 1.47207931e-01 -4.26749945e-01 8.29174697e-01
5.33905625e-01 1.16756208e-01 -6.24460399e-01 7.77194872e-02
3.29303861e-01 7.55592585e-01 -8.14342141e-01 -3.54436696e-01
-5.37256777e-01 1.38266012e-01 -3.55149001e-01 2.62214482e-01
-5.86832225e-01 -2.91204154e-01 8.14514935e-01 6.58544153e-02
7.16157407e-02 -2.24524185e-01 -3.14896703e-01 -1.18741047e+00
5.03182365e-03 -8.63913596e-01 7.46243000e-01 -1.74729928e-01
-7.75007248e-01 3.43543112e-01 3.15648228e-01 -1.00935328e+00
-9.17197347e-01 -3.04282904e-01 -1.24133086e+00 1.38095772e+00
-1.64221072e+00 -3.95224214e-01 8.59889090e-01 3.12888741e-01
3.31306756e-01 -2.97990084e-01 5.64158142e-01 -1.21046059e-01
-6.56306863e-01 1.67360471e-03 7.87476659e-01 -2.37933602e-02
-1.60122588e-02 -1.46589255e+00 8.04808617e-01 1.16576266e+00
-8.40488449e-02 5.80905139e-01 1.01604474e+00 -7.91740477e-01
-5.40186346e-01 -6.84185505e-01 1.30083060e+00 -4.54371929e-01
1.34801447e+00 2.42220223e-01 -6.91106319e-01 4.31854367e-01
1.69182152e-01 -6.41447783e-01 3.33929390e-01 -8.97760466e-02
1.04757801e-01 9.39346626e-02 -1.02806234e+00 4.71952558e-01
2.85528451e-01 -1.98825717e-01 -9.90107775e-01 1.41682997e-01
4.83374447e-01 2.04019055e-01 -1.37077868e+00 5.28829753e-01
6.99280918e-01 -1.02969348e+00 1.16789901e+00 -6.05631709e-01
1.12250432e-01 1.12350322e-01 -3.39385085e-02 -1.58098507e+00
-4.54195946e-01 -1.09996796e+00 2.55969819e-02 1.46522439e+00
5.37421823e-01 -1.21728754e+00 4.72658515e-01 6.42147660e-01
3.41680497e-01 -6.35426462e-01 -1.28358984e+00 -9.12523627e-01
6.08475208e-01 -4.34504636e-02 1.04563606e+00 9.99234378e-01
3.26319113e-02 -3.40963632e-01 -1.78803653e-01 1.25799924e-01
8.75348628e-01 4.96506929e-01 2.53514349e-01 -1.20303547e+00
-1.12052336e-01 -1.18663824e+00 -8.19473863e-02 -1.06315143e-01
-1.48802280e-01 -6.35026872e-01 -5.54394245e-01 -8.20624471e-01
-1.19219102e-01 -4.71393764e-01 -5.84724784e-01 -2.77508289e-01
-9.36505571e-02 -1.71101063e-01 2.07407549e-01 5.28385222e-01
3.14756870e-01 2.48901799e-01 6.13832057e-01 1.68090373e-01
-3.16120118e-01 4.47158039e-01 -6.38730407e-01 8.30416739e-01
1.04303932e+00 -2.15638369e-01 -1.85328528e-01 -7.65549019e-02
6.08209431e-01 4.45060819e-01 3.08144420e-01 -8.76273751e-01
-8.95676389e-02 -7.22838044e-01 1.14083104e-01 -6.87110245e-01
-4.05407071e-01 -6.42894030e-01 7.21949935e-01 8.44255745e-01
-2.23492980e-01 8.68509591e-01 1.03793278e-01 2.44258150e-01
-2.56971002e-01 -3.46865833e-01 8.70719969e-01 -3.22662354e-01
-2.40155444e-01 6.43397123e-02 -5.19749761e-01 -3.64649892e-02
1.11550844e+00 -7.65900537e-02 3.64928357e-02 -4.07788068e-01
-9.11134541e-01 3.87216181e-01 5.51536620e-01 3.11484993e-01
-6.36363178e-02 -1.13666356e+00 -1.01417673e+00 1.88233089e-02
-2.87502259e-01 -4.49997187e-01 -1.17247000e-01 9.56462145e-01
-5.99420011e-01 8.12239349e-01 -6.11082725e-02 -9.10555348e-02
-8.54106903e-01 5.85953474e-01 7.33799875e-01 -7.37208188e-01
-6.08439505e-01 5.12000203e-01 6.09630644e-02 -8.62143282e-03
-2.15445772e-01 -1.85333386e-01 -5.29422939e-01 8.03214431e-01
6.18159652e-01 5.88169158e-01 1.22518037e-02 -1.10137045e+00
-1.65552661e-01 5.52069724e-01 3.55663419e-01 -4.69907492e-01
1.38564599e+00 -4.89759594e-01 -2.20558077e-01 6.29118145e-01
9.36401188e-01 -1.04712211e-01 -1.18930459e+00 -4.67546955e-02
6.94779158e-01 -2.76798636e-01 -4.95386831e-02 -4.59855825e-01
-1.04155374e+00 4.76420641e-01 3.57262701e-01 8.17024946e-01
1.18333256e+00 -2.37174928e-01 5.83948672e-01 -3.56246203e-01
6.19439781e-02 -1.11795282e+00 -7.61650264e-01 2.58639604e-01
9.56454039e-01 -8.17889988e-01 1.55200679e-02 2.94021726e-01
-1.46815911e-01 1.10674226e+00 -3.46050948e-01 -1.49793662e-02
1.17936862e+00 4.56040576e-02 4.52810489e-02 -3.32529962e-01
-7.55901337e-01 -1.22323878e-01 3.56206261e-02 2.83250839e-01
9.13916975e-02 3.94132167e-01 -8.26355219e-01 1.79504067e-01
-7.32814729e-01 -5.85811377e-01 4.10619348e-01 6.71912670e-01
-3.84570777e-01 -8.01700532e-01 -8.87788951e-01 6.72367036e-01
-1.14903986e+00 -1.76845849e-01 1.32648945e-01 6.08249605e-01
-1.89026281e-01 6.23406231e-01 5.74296713e-01 2.25049153e-01
2.71144599e-01 1.94419891e-01 -1.91151276e-01 -3.97984535e-02
-5.87976635e-01 4.20935988e-01 1.99680328e-01 -1.51944272e-02
-3.99984449e-01 -1.49405372e+00 -7.80622900e-01 -5.63198090e-01
-4.05254513e-01 4.19517517e-01 8.05446327e-01 8.54924977e-01
-1.85237631e-01 2.60697395e-01 1.14910245e+00 -7.79202640e-01
-9.99881983e-01 -7.23898888e-01 -6.24420166e-01 2.60375023e-01
3.47178727e-01 -5.58151901e-01 -8.85932088e-01 -3.67053831e-03] | [5.437546730041504, 4.022642612457275] |
99c30ffe-aa15-47af-9538-e0f332994009 | aspect-sentiment-multiple-opinion-triplet | 2110.07303 | null | https://arxiv.org/abs/2110.07303v1 | https://arxiv.org/pdf/2110.07303v1.pdf | Aspect-Sentiment-Multiple-Opinion Triplet Extraction | Aspect Sentiment Triplet Extraction (ASTE) aims to extract aspect term (aspect), sentiment and opinion term (opinion) triplets from sentences and can tell a complete story, i.e., the discussed aspect, the sentiment toward the aspect, and the cause of the sentiment. ASTE is a charming task, however, one triplet extracted by ASTE only includes one opinion of the aspect, but an aspect in a sentence may have multiple corresponding opinions and one opinion only provides part of the reason why the aspect has this sentiment, as a consequence, some triplets extracted by ASTE are hard to understand, and provide erroneous information for downstream tasks. In this paper, we introduce a new task, named Aspect Sentiment Multiple Opinions Triplet Extraction (ASMOTE). ASMOTE aims to extract aspect, sentiment and multiple opinions triplets. Specifically, one triplet extracted by ASMOTE contains all opinions about the aspect and can tell the exact reason that the aspect has the sentiment. We propose an Aspect-Guided Framework (AGF) to address this task. AGF first extracts aspects, then predicts their opinions and sentiments. Moreover, with the help of the proposed Sequence Labeling Attention(SLA), AGF improves the performance of the sentiment classification using the extracted opinions. Experimental results on multiple datasets demonstrate the effectiveness of our approach. | ['Yancheng He', 'Cunxiang Yin', 'Sheng-hua Zhong', 'Yuncong Li', 'Fang Wang'] | 2021-10-14 | null | null | null | null | ['extract-aspect', 'aspect-sentiment-triplet-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [ 1.77605733e-01 4.55548614e-02 -1.24488249e-01 -6.89934254e-01
-5.46934724e-01 -7.35022604e-01 6.43180072e-01 3.59049797e-01
5.80441169e-02 6.02417648e-01 6.44276381e-01 -2.96122521e-01
2.57243037e-01 -8.08545470e-01 -5.37843585e-01 -7.07366168e-01
6.19399786e-01 2.22690776e-01 -3.72222066e-02 -5.99919736e-01
4.23900247e-01 -2.16048732e-01 -1.31010818e+00 7.62979269e-01
7.05276370e-01 1.27135336e+00 -1.67642415e-01 2.81061262e-01
-8.63700211e-01 1.10854650e+00 -9.70504165e-01 -7.42533743e-01
-1.87794074e-01 -5.14642656e-01 -6.93002760e-01 3.32890898e-01
-1.44977555e-01 -9.81225669e-02 5.55873036e-01 1.07029641e+00
1.60182670e-01 -2.12283701e-01 7.84503520e-01 -1.39396489e+00
-4.86988395e-01 8.17298651e-01 -6.36453629e-01 -7.12670982e-02
4.17531461e-01 -1.25803396e-01 1.44518220e+00 -1.03113341e+00
4.39502954e-01 1.26206064e+00 3.72570962e-01 2.58627087e-01
-2.37789184e-01 -5.03874481e-01 7.91950285e-01 4.50260798e-03
-5.44144034e-01 -2.07914516e-01 8.50609481e-01 -3.70165855e-01
8.69874954e-01 2.20427528e-01 9.59009886e-01 9.75625992e-01
2.60334134e-01 1.20176673e+00 1.26454782e+00 -1.35391191e-01
1.71538442e-01 4.69294995e-01 8.86846602e-01 2.48170704e-01
2.95044124e-01 -5.48803747e-01 -6.19877577e-01 -1.13999844e-02
-1.95459217e-01 4.82106768e-02 -2.05926120e-01 3.20772201e-01
-9.64765608e-01 6.07504189e-01 1.80848673e-01 2.87613630e-01
-5.85352242e-01 -3.46462190e-01 6.68321252e-01 3.12462360e-01
5.24070263e-01 1.64478719e-01 -9.93308902e-01 -1.70485780e-01
-2.52301782e-01 1.22688718e-01 1.14607513e+00 9.31020558e-01
9.77857649e-01 -1.25504017e-01 -2.93632299e-01 5.06660402e-01
3.51302236e-01 8.07043791e-01 5.18942952e-01 -3.25181991e-01
5.44334292e-01 1.38752282e+00 6.62730187e-02 -1.10007095e+00
-3.91546726e-01 -2.74843335e-01 -6.96628809e-01 -1.81978315e-01
-2.34856471e-01 -5.30725956e-01 -9.05274034e-01 1.29842091e+00
5.83166957e-01 -4.68000978e-01 4.34682518e-01 7.15882361e-01
1.41642511e+00 8.52247715e-01 -2.29850188e-01 -3.09199810e-01
1.80488157e+00 -1.12505126e+00 -9.17291403e-01 -7.86766350e-01
4.97094840e-01 -1.04733455e+00 8.54387403e-01 3.19207549e-01
-3.60248744e-01 -2.21038893e-01 -9.24448013e-01 9.49354842e-02
-3.82349938e-01 2.77952284e-01 7.44453967e-01 2.21103191e-01
-4.94500518e-01 -6.70706108e-02 -3.89530808e-01 -9.05508474e-02
7.76292309e-02 2.12435588e-01 -1.81756914e-01 -6.02522194e-02
-1.18247712e+00 5.99420249e-01 7.40699396e-02 3.23677748e-01
-2.51906604e-01 -3.82760048e-01 -1.16088867e+00 5.02307229e-02
6.28176689e-01 -6.84365928e-01 1.29398632e+00 -1.33352339e+00
-1.09601545e+00 4.42122966e-01 -7.92655826e-01 7.04893470e-02
-8.11633691e-02 -2.63908923e-01 -6.70286596e-01 -1.36003029e-02
6.38368309e-01 2.03216612e-01 7.13997722e-01 -1.10713172e+00
-9.19354916e-01 -6.05899513e-01 3.64128023e-01 2.56088316e-01
-4.57803369e-01 2.42554933e-01 -5.51375210e-01 -3.90682429e-01
-1.23460134e-02 -8.60723794e-01 -7.55560100e-02 -5.88433206e-01
-8.04050028e-01 -3.63475412e-01 8.47853422e-01 -5.01445472e-01
1.26958334e+00 -2.09244180e+00 -2.11232483e-01 1.80504248e-01
6.47703558e-02 2.91120499e-01 -1.05809368e-01 4.49759901e-01
-7.43010268e-02 2.31991038e-01 -3.43934983e-01 -3.26425545e-02
4.48244214e-02 3.54185909e-01 -5.93979776e-01 -3.52702081e-01
4.67721254e-01 1.01737499e+00 -7.39504695e-01 -3.62339586e-01
-3.16158354e-01 2.95592576e-01 -2.87369102e-01 1.63578704e-01
-4.63910848e-01 3.55400950e-01 -9.86833274e-01 6.50144100e-01
6.35157466e-01 -3.46123993e-01 -1.08722612e-01 -6.54414594e-01
-1.22024335e-01 6.25536740e-01 -8.14325511e-01 8.03350985e-01
-5.33459365e-01 3.65847349e-01 -2.19447687e-01 -6.84170187e-01
1.05035996e+00 3.65260601e-01 2.55553126e-01 -4.68943924e-01
3.91836226e-01 3.25665861e-01 -1.84038758e-01 -6.84404731e-01
5.12986720e-01 -3.95435721e-01 -3.09935182e-01 7.05903113e-01
-1.97438464e-01 -1.70415968e-01 6.78174019e-01 3.76649410e-01
9.13021386e-01 5.03971986e-02 5.34183323e-01 2.20363677e-01
9.39327419e-01 2.79424340e-01 8.27332675e-01 1.38363495e-01
1.22126050e-01 4.45863932e-01 9.97498453e-01 -5.30723095e-01
-6.78731561e-01 -1.50278971e-01 3.27527404e-01 6.16563261e-01
2.59832025e-01 -7.85434902e-01 -2.44512022e-01 -1.18866122e+00
-2.95057833e-01 7.93847382e-01 -7.67795980e-01 4.20062020e-02
-3.29315484e-01 -8.47009480e-01 -1.97208866e-01 4.65466172e-01
5.56188643e-01 -1.29838002e+00 -9.06253532e-02 1.23998985e-01
-6.87420607e-01 -1.32008803e+00 -6.65159941e-01 5.84699512e-02
-3.83533508e-01 -1.21455264e+00 -3.61089975e-01 -7.04185367e-01
8.67380500e-01 3.40256095e-01 1.24404061e+00 1.37199208e-01
4.93622124e-01 -1.05967239e-01 -7.74043083e-01 -7.53612280e-01
-2.79570132e-01 -1.27810121e-01 -3.47620308e-01 4.79814887e-01
8.08283925e-01 -5.11419475e-01 -4.60423082e-01 1.54071078e-01
-8.99564147e-01 -2.19669603e-02 8.41645181e-01 5.29534996e-01
6.00615501e-01 1.29218593e-01 4.52210546e-01 -1.47931767e+00
8.46182942e-01 -5.55638671e-01 -2.40451589e-01 1.23337477e-01
-6.94193006e-01 1.02621548e-01 8.42923164e-01 -4.15523313e-02
-1.16796231e+00 -4.66060072e-01 -5.02307057e-01 -2.04962678e-02
-2.67201327e-02 1.06868184e+00 -3.73647332e-01 6.62019312e-01
-3.36996056e-02 5.24735272e-01 -3.39048296e-01 -7.63539672e-02
1.28282264e-01 8.51675689e-01 4.43265997e-02 -2.46692255e-01
4.54507589e-01 5.22376180e-01 -1.40671253e-01 -5.32129705e-01
-1.74003983e+00 -6.04277670e-01 -2.51671255e-01 -1.61029965e-01
6.64144814e-01 -8.98933768e-01 -5.31757057e-01 4.31851238e-01
-1.33291829e+00 5.87589920e-01 -1.27265781e-01 1.25527307e-01
-9.18984115e-02 2.28242442e-01 -3.39097768e-01 -9.63390887e-01
-8.85547101e-01 -1.19450533e+00 1.23415327e+00 3.01919103e-01
-4.93159950e-01 -8.63421738e-01 -4.91377227e-02 5.22579014e-01
8.41943771e-02 1.01078182e-01 1.03787136e+00 -1.01241744e+00
-2.96633184e-01 -3.78230572e-01 -1.56420574e-01 4.78476971e-01
5.00145555e-01 1.76692903e-01 -7.28707433e-01 2.68302470e-01
3.06897581e-01 -1.15006484e-01 7.42043674e-01 1.43696234e-01
3.85415196e-01 -7.31660962e-01 -5.55833206e-02 6.45457879e-02
1.10574162e+00 3.73421907e-01 3.02275062e-01 4.70712334e-01
7.11260796e-01 8.00246418e-01 9.91076887e-01 3.61037582e-01
7.85698712e-01 2.93387353e-01 2.94700623e-01 6.89175539e-03
-1.87124237e-02 -1.11507624e-01 9.65417862e-01 1.29473996e+00
4.32512939e-01 -2.54045516e-01 -4.65630680e-01 6.63789928e-01
-1.76653099e+00 -6.82242334e-01 -5.63471079e-01 1.33672345e+00
6.40457332e-01 3.40945601e-01 -2.04094410e-01 2.37199038e-01
5.56869805e-01 4.01442051e-01 -7.63646364e-01 -5.83712816e-01
-3.94680083e-01 -2.07113042e-01 -3.53019774e-01 3.53511363e-01
-8.43671083e-01 8.11849594e-01 4.42135000e+00 6.28657639e-01
-9.05042648e-01 -8.56244415e-02 5.96798718e-01 2.47736573e-01
-9.50070977e-01 3.57514381e-01 -1.12059748e+00 4.10169691e-01
3.45522523e-01 -4.81201738e-01 -1.95411041e-01 8.92492592e-01
3.05867314e-01 -1.24650083e-01 -7.51509368e-01 4.33309287e-01
4.76778239e-01 -8.17351282e-01 5.96810400e-01 -2.65656054e-01
7.76028514e-01 -2.08770394e-01 -1.65688336e-01 2.68516868e-01
1.52096912e-01 -3.64790261e-01 7.18286872e-01 2.36688301e-01
2.63008505e-01 -8.99664044e-01 1.48179281e+00 2.74747580e-01
-1.43567371e+00 4.01720926e-02 -1.07852176e-01 -8.21314156e-02
3.21302354e-01 1.35574615e+00 -5.93724132e-01 8.43108058e-01
7.05599964e-01 1.24796236e+00 -4.26563203e-01 6.00139976e-01
-1.06200051e+00 5.95820248e-01 4.84771058e-02 -5.95995545e-01
4.52515990e-01 -5.24021626e-01 7.54406393e-01 9.55583513e-01
2.99305260e-01 1.86377540e-01 5.27887456e-02 5.34212470e-01
4.32602875e-02 4.07056719e-01 -5.16857684e-01 -4.38452542e-01
5.23349009e-02 1.68363571e+00 -6.46414518e-01 -5.91378510e-01
-6.09716058e-01 8.13196778e-01 5.89658767e-02 3.85560721e-01
-3.09102178e-01 -5.03287196e-01 7.26893187e-01 -3.83380026e-01
6.94460273e-01 4.34213251e-01 -4.69421625e-01 -1.45568442e+00
6.35036230e-01 -1.14086640e+00 3.06033581e-01 -1.14374626e+00
-1.29833925e+00 1.08535242e+00 -6.46816313e-01 -1.46056449e+00
-2.94802248e-01 -5.16238511e-01 -9.44842637e-01 7.36101747e-01
-1.81702483e+00 -1.15305459e+00 -1.46052733e-01 2.89596885e-01
7.09770739e-01 -3.72020248e-03 5.67522168e-01 -1.74200684e-02
-5.79653919e-01 1.32325292e-01 -6.41421497e-01 2.37993449e-01
4.86786813e-01 -1.12454975e+00 3.72185200e-01 9.13092911e-01
-3.73290554e-02 7.64128029e-01 7.42763400e-01 -7.32233107e-01
-1.31834197e+00 -1.16249156e+00 1.56737423e+00 -4.11179900e-01
8.31766009e-01 1.78332515e-02 -6.20744288e-01 7.16010451e-01
3.20916742e-01 -6.48567855e-01 9.64092433e-01 3.20230752e-01
-4.57221538e-01 -1.22470461e-01 -6.21792436e-01 6.52413011e-01
5.11672616e-01 -3.21805388e-01 -7.87040949e-01 3.09073597e-01
1.18007648e+00 -2.27591116e-02 -3.94179851e-01 6.01510823e-01
4.60758030e-01 -1.12613142e+00 3.01260829e-01 -4.99650866e-01
1.04517949e+00 -5.51923037e-01 -9.18055400e-02 -1.47030163e+00
2.69296288e-01 -1.80586681e-01 -7.58484676e-02 1.70710206e+00
9.89841521e-01 -6.16653144e-01 4.16799575e-01 4.93497640e-01
-2.59373218e-01 -1.10192764e+00 -4.06272888e-01 -1.16577052e-01
-3.87497038e-01 -5.95855236e-01 9.87010121e-01 6.30452156e-01
1.38386369e-01 1.10704172e+00 -2.14493632e-01 1.81058347e-01
9.60834175e-02 1.05877209e+00 6.19979978e-01 -9.20964837e-01
-1.09317657e-02 -3.82552773e-01 -1.59671172e-01 -1.09436715e+00
1.25829592e-01 -6.53078496e-01 -1.10954270e-02 -1.94928300e+00
4.29781765e-01 6.47389591e-02 -4.44313921e-02 7.49817550e-01
-6.92365646e-01 -5.98679446e-02 1.03344768e-01 3.84107642e-02
-6.13230050e-01 8.04865241e-01 1.54451692e+00 -4.40418869e-01
8.31402317e-02 4.05895501e-01 -1.48465121e+00 1.02101362e+00
5.51923275e-01 -5.06558180e-01 -2.56161541e-01 -2.52587765e-01
9.84327853e-01 -1.39228508e-01 -2.83217192e-01 -3.01615626e-01
2.38658756e-01 -1.11452658e-02 1.35153502e-01 -9.44460571e-01
1.88253179e-01 -8.12729657e-01 -4.38327044e-01 3.45102400e-01
-6.02143407e-02 2.62406826e-01 2.51951870e-02 2.54887909e-01
-7.69475520e-01 -1.75348714e-01 1.03593599e-02 -1.42569378e-01
-6.38168037e-01 3.20138514e-01 -3.43389541e-01 1.34957477e-01
5.40693700e-01 1.54735684e-01 -4.93588120e-01 -5.62470257e-01
-4.95124370e-01 6.66509569e-01 1.42919108e-01 6.52342737e-01
7.62880981e-01 -1.24289644e+00 -6.87007606e-01 2.00109780e-01
4.71992940e-01 -3.84431928e-02 1.15356416e-01 8.40206087e-01
2.06525028e-01 4.28413004e-01 4.23457146e-01 -2.70228148e-01
-1.45754027e+00 4.59778965e-01 -2.37419885e-02 -4.94569451e-01
-2.76428282e-01 7.07802653e-01 3.61817211e-01 -7.75508285e-01
-3.85272443e-01 -2.46073648e-01 -1.03593016e+00 6.17259145e-01
6.76412761e-01 -2.70080894e-01 1.64198548e-01 -8.43618035e-01
-4.45722044e-01 9.35241520e-01 -2.97415137e-01 -1.12689443e-01
1.27424598e+00 -2.99334735e-01 -8.13875794e-01 7.14029849e-01
1.06650937e+00 3.47724676e-01 -6.30086303e-01 -1.15205370e-01
-1.61182448e-01 -9.51806381e-02 -2.83815116e-01 -9.74403143e-01
-1.25823319e+00 7.86117315e-01 -4.32864398e-01 1.78094596e-01
1.31847334e+00 1.15273148e-01 1.09823132e+00 3.42231482e-01
1.00566812e-01 -6.52880311e-01 -5.89265488e-02 9.19737399e-01
8.97291720e-01 -1.27321482e+00 -1.52867243e-01 -7.05296397e-01
-1.20376968e+00 1.04820919e+00 6.53669119e-01 3.20236504e-01
7.25989282e-01 2.03212827e-01 6.08814895e-01 -5.26276767e-01
-1.11750233e+00 -3.62130284e-01 3.31970870e-01 6.40222132e-02
3.88822287e-01 -1.06188521e-01 -6.23149812e-01 1.14623415e+00
-6.02168381e-01 -2.63148934e-01 5.62321663e-01 9.07831609e-01
-4.76836324e-01 -1.15193367e+00 -5.01097776e-02 5.78359723e-01
-5.26446462e-01 -5.38415790e-01 -9.72092271e-01 1.78244725e-01
1.36239782e-01 1.51360655e+00 -2.79221952e-01 -5.82320333e-01
6.06583595e-01 -1.50219472e-02 -3.54103476e-01 -6.79865837e-01
-8.69639158e-01 1.25651479e-01 4.31015909e-01 -3.90112102e-01
-6.52299225e-01 -4.86115575e-01 -1.45007598e+00 7.14152008e-02
-4.12737310e-01 7.91484892e-01 7.64299452e-01 1.66480637e+00
4.05001581e-01 7.07459152e-01 1.07348394e+00 5.75162619e-02
-4.11083326e-02 -9.23406482e-01 -3.26316088e-01 5.04958451e-01
5.01707137e-01 -1.36642545e-01 -5.85696757e-01 -1.03182822e-01] | [11.483474731445312, 6.640137195587158] |
218ec99d-3182-4161-a353-15c845885d80 | recognition-and-prediction-of-surgical | 2212.01683 | null | https://arxiv.org/abs/2212.01683v1 | https://arxiv.org/pdf/2212.01683v1.pdf | Recognition and Prediction of Surgical Gestures and Trajectories Using Transformer Models in Robot-Assisted Surgery | Surgical activity recognition and prediction can help provide important context in many Robot-Assisted Surgery (RAS) applications, for example, surgical progress monitoring and estimation, surgical skill evaluation, and shared control strategies during teleoperation. Transformer models were first developed for Natural Language Processing (NLP) to model word sequences and soon the method gained popularity for general sequence modeling tasks. In this paper, we propose the novel use of a Transformer model for three tasks: gesture recognition, gesture prediction, and trajectory prediction during RAS. We modify the original Transformer architecture to be able to generate the current gesture sequence, future gesture sequence, and future trajectory sequence estimations using only the current kinematic data of the surgical robot end-effectors. We evaluate our proposed models on the JHU-ISI Gesture and Skill Assessment Working Set (JIGSAWS) and use Leave-One-User-Out (LOUO) cross-validation to ensure the generalizability of our results. Our models achieve up to 89.3\% gesture recognition accuracy, 84.6\% gesture prediction accuracy (1 second ahead) and 2.71mm trajectory prediction error (1 second ahead). Our models are comparable to and able to outperform state-of-the-art methods while using only the kinematic data channel. This approach can enable near-real time surgical activity recognition and prediction. | ['Ann Majewicz Fey', 'Yi Zheng', 'Chang Shi'] | 2022-12-03 | null | null | null | null | ['gesture-recognition'] | ['computer-vision'] | [ 4.07246530e-01 2.59528518e-01 -6.20010138e-01 -1.90893114e-01
-9.94850218e-01 -3.93222392e-01 4.74617213e-01 -7.36629544e-03
-8.99011493e-01 4.09128934e-01 6.64883554e-01 -6.05396152e-01
-4.57317948e-01 -1.46857426e-01 -2.15948015e-01 -5.32727599e-01
-3.84925306e-01 6.09392822e-01 1.90938413e-01 -2.05345631e-01
2.91445076e-01 6.77559674e-01 -1.12277079e+00 5.88759005e-01
3.65103602e-01 8.30619395e-01 5.77452898e-01 1.22735119e+00
1.00404792e-01 9.90379155e-01 -2.95193315e-01 1.34166405e-01
1.71505332e-01 -4.50695813e-01 -1.03948474e+00 -3.80870968e-01
-2.36296073e-01 -3.98194581e-01 -3.68002236e-01 3.00972998e-01
6.53246641e-01 7.47037157e-02 4.55703229e-01 -7.93471336e-01
4.88547117e-01 4.60363895e-01 2.05534786e-01 -1.16485842e-02
5.25690496e-01 2.89378911e-01 3.37143064e-01 -5.81317544e-01
1.12139189e+00 6.78074479e-01 8.04261506e-01 8.59532893e-01
-6.43701196e-01 -6.49511039e-01 -8.65770057e-02 1.43747970e-01
-9.98995960e-01 -1.94673985e-01 2.01811805e-01 -5.74503660e-01
1.17352390e+00 5.32856464e-01 8.91167164e-01 1.30933297e+00
1.10911691e+00 1.02978647e+00 7.37110794e-01 -3.54435921e-01
3.58547829e-02 -5.07479966e-01 -1.06898449e-01 8.36979806e-01
-3.22817653e-01 6.81311607e-01 -5.37737608e-01 -2.56454246e-03
1.00849414e+00 2.50061452e-01 -2.48890504e-01 -1.56398132e-01
-2.05756974e+00 4.55010235e-01 4.64529425e-01 4.44781452e-01
-7.44248629e-01 3.52371603e-01 5.36577821e-01 3.70248497e-01
-1.86580807e-01 4.40391392e-01 -5.11734366e-01 -1.12141681e+00
-5.57470083e-01 3.43733374e-03 9.24849510e-01 1.06054020e+00
-2.85556912e-01 -3.56815785e-01 -3.62484992e-01 4.55120176e-01
2.00399041e-01 1.50123447e-01 1.27097118e+00 -7.84891009e-01
4.11298305e-01 3.48250628e-01 -4.77855578e-02 -4.58127290e-01
-1.18812799e+00 -1.84951589e-01 -7.18452513e-01 1.69383615e-01
3.14647675e-01 -1.08268000e-01 -1.35169864e+00 1.09310687e+00
4.64798994e-02 1.56405061e-01 1.65279061e-01 8.88087809e-01
5.68707347e-01 2.71625131e-01 3.43809277e-01 -3.70242327e-01
1.12630153e+00 -9.29702818e-01 -6.71248019e-01 -2.60376573e-01
1.39690077e+00 -7.94965684e-01 8.88890803e-01 6.75458252e-01
-6.45062029e-01 -3.68621439e-01 -6.48629248e-01 2.78750569e-01
1.42636970e-01 4.30860728e-01 7.50902474e-01 5.53380728e-01
-7.69004285e-01 7.00396478e-01 -1.75706279e+00 -6.17899299e-01
2.85431352e-02 8.16716552e-01 -6.15621209e-01 2.47779079e-02
-6.41878247e-01 1.23021126e+00 4.02802825e-01 2.83736229e-01
-9.31772709e-01 -5.04295468e-01 -7.86057591e-01 -5.64750373e-01
1.56091228e-01 -4.68834817e-01 1.58157206e+00 -5.31156778e-01
-2.00443316e+00 5.70784211e-01 -2.10585549e-01 -4.76592124e-01
6.50188446e-01 -4.52270508e-01 -2.94680893e-01 -1.67100027e-01
-4.92196351e-01 4.02808845e-01 2.64113307e-01 -5.81831574e-01
-7.29105413e-01 -3.61374199e-01 -4.77774352e-01 4.47131753e-01
7.26513043e-02 -1.91440612e-01 -4.30841178e-01 -7.88920462e-01
4.69469219e-01 -1.56433117e+00 -6.47626102e-01 3.16856466e-02
-7.75134414e-02 2.05567390e-01 2.41521925e-01 -9.49812472e-01
1.26068306e+00 -1.94273114e+00 3.56688321e-01 3.20682794e-01
-2.04641059e-01 2.06290469e-01 -2.65950531e-01 6.11516535e-01
-4.99089733e-02 -4.23892647e-01 -8.15912266e-04 -1.44845350e-02
-4.96702224e-01 5.53644598e-01 6.64892094e-03 4.66069847e-01
-3.62206459e-01 1.07476437e+00 -1.00116789e+00 -4.97358799e-01
5.59502363e-01 5.45645021e-02 -4.53730315e-01 2.95803726e-01
1.17063656e-01 1.04539382e+00 -4.30010885e-01 6.15911782e-01
-3.21288675e-01 1.81559563e-01 6.01567268e-01 -1.01807132e-01
-1.91011708e-02 2.26654813e-01 -7.15528727e-01 2.37619543e+00
-8.11948776e-01 6.66473508e-01 -2.53087372e-01 -4.62507039e-01
8.49422574e-01 4.79701161e-01 1.15684605e+00 -6.65838301e-01
3.43413502e-01 4.02167350e-01 5.02669871e-01 -9.11755800e-01
2.28494272e-01 -3.55874985e-01 -2.29143009e-01 2.99458981e-01
-3.89492661e-02 -5.58695309e-02 -1.86144009e-01 -2.94880092e-01
1.31348085e+00 4.42813873e-01 8.21614504e-01 -3.09384544e-03
2.47438535e-01 4.23512459e-01 3.45466882e-01 6.53266549e-01
-3.09436500e-01 4.58722264e-01 1.61576018e-01 -4.90084052e-01
-6.40137017e-01 -9.96956408e-01 2.41981655e-01 9.63999927e-01
6.86732158e-02 -3.67585987e-01 -2.89066941e-01 -7.65946090e-01
-2.65518010e-01 7.35400736e-01 -6.23813629e-01 -3.16450328e-01
-1.00336277e+00 -4.66618359e-01 6.21010423e-01 9.55264986e-01
-2.36706331e-01 -1.27097154e+00 -1.00540698e+00 5.70220351e-01
-2.85409719e-01 -9.00923789e-01 -3.69848877e-01 1.75773397e-01
-1.32753098e+00 -1.24872684e+00 -6.47391796e-01 -9.41008925e-01
5.02290606e-01 -4.76539403e-01 4.96945620e-01 -1.21566661e-01
-5.10867774e-01 5.44903755e-01 -6.16004586e-01 -4.72417295e-01
-8.32906961e-01 8.97178203e-02 1.49371162e-01 -5.94002306e-01
1.05013385e-01 -1.30094111e-01 -6.27147138e-01 5.95475137e-01
-7.39074051e-01 4.17883575e-01 1.28694856e+00 1.10892820e+00
4.62103695e-01 -7.70716310e-01 -5.01568466e-02 -4.96987045e-01
6.19054973e-01 -2.33643964e-01 -1.08954631e-01 3.11111659e-01
-8.51795137e-01 2.36394376e-01 1.78836644e-01 -7.72387326e-01
-7.56330609e-01 6.01354361e-01 -4.72909629e-01 -3.79673272e-01
-2.54206538e-01 7.06520617e-01 4.47308898e-01 -1.15935162e-01
5.48346937e-01 3.89669150e-01 6.17700458e-01 -3.88992250e-01
5.30872233e-02 7.18903184e-01 6.03239179e-01 -6.78993389e-02
1.13932967e-01 1.55763030e-01 4.90873419e-02 -5.58280766e-01
1.69811174e-02 -8.56247902e-01 -9.28844452e-01 -4.18593287e-01
6.02531016e-01 -5.63519418e-01 -1.14691293e+00 4.89766568e-01
-9.58321154e-01 -8.51592004e-01 -1.08131684e-01 1.23516560e+00
-1.15072978e+00 1.47690147e-01 -7.46012330e-01 -7.73755491e-01
-5.79909563e-01 -1.40585470e+00 1.27349019e+00 -3.41629475e-01
-7.95625150e-01 -6.79431915e-01 2.39469960e-01 -1.71941277e-02
2.91795760e-01 4.49379265e-01 6.66754127e-01 -7.28051722e-01
-2.03314722e-01 -7.61528611e-01 3.57582659e-01 -1.41247734e-01
1.25374869e-01 -6.36484563e-01 -2.22549722e-01 -1.77529708e-01
-2.29469210e-01 1.52896151e-01 3.04359853e-01 4.08232003e-01
1.16059399e+00 -2.21483141e-01 -8.69437397e-01 4.37756270e-01
1.04475248e+00 5.50699353e-01 8.49038363e-01 3.07707191e-01
5.69558322e-01 4.61988449e-01 1.23255241e+00 3.33259970e-01
1.08046830e-01 1.06308877e+00 3.95141840e-01 3.41302246e-01
-1.41505241e-01 -2.19782203e-01 4.41325992e-01 9.88685608e-01
-3.52878720e-01 -2.73242127e-02 -1.43769264e+00 4.47697341e-01
-1.94825578e+00 -7.45131075e-01 -1.28189087e-01 2.24131799e+00
5.53231120e-01 -1.23262458e-01 -9.15166438e-02 -4.94496524e-02
-8.15159306e-02 -3.74021798e-01 -3.15477163e-01 -3.35972786e-01
9.24482346e-01 3.77252132e-01 8.44192684e-01 5.03203034e-01
-8.44872892e-01 7.65342832e-01 6.27237129e+00 6.04193270e-01
-1.28277886e+00 9.61016417e-02 4.02208231e-02 -3.71496230e-01
4.41039145e-01 -3.73798043e-01 -3.73877674e-01 2.73804843e-01
1.20617509e+00 2.84251064e-01 1.23732515e-01 7.90993154e-01
4.05650318e-01 -1.91863120e-01 -1.22816920e+00 1.15844250e+00
-6.59287616e-04 -1.42363191e+00 -2.31062531e-01 4.93309684e-02
8.69929641e-02 2.62182355e-01 -4.11325961e-01 3.76374096e-01
-1.23599474e-03 -1.23584235e+00 4.32689905e-01 9.54868495e-01
1.11710393e+00 -4.16478187e-01 1.16793144e+00 8.10771286e-01
-1.16586924e+00 -4.48136330e-01 4.91536826e-01 9.57799852e-02
4.37427908e-01 -2.93842494e-01 -1.67355132e+00 6.26592100e-01
3.00530195e-01 5.97726703e-01 -8.01182017e-02 1.01300204e+00
9.07937437e-03 3.75239551e-01 -3.94079834e-01 -4.19548690e-01
2.89874405e-01 2.21370772e-01 4.98533905e-01 1.14180303e+00
6.29914880e-01 3.48758161e-01 1.39684558e-01 -2.02773407e-01
6.03953123e-01 4.26182449e-02 -4.67686713e-01 -1.14101484e-01
-4.65272255e-02 6.29852414e-01 -6.33399129e-01 6.38711080e-02
-7.92933404e-02 1.27933323e+00 -1.44914776e-01 8.10505301e-02
-5.81585407e-01 -3.20082366e-01 8.86841834e-01 -3.61675508e-02
-2.19516858e-01 -6.96268499e-01 -4.40876842e-01 -6.44157350e-01
-1.10334702e-01 -8.38913500e-01 4.11951214e-01 -5.13637841e-01
-4.63668615e-01 6.33643985e-01 8.46416503e-03 -2.01846790e+00
-1.16907132e+00 -1.16970968e+00 -2.90766150e-01 5.90086639e-01
-6.88467801e-01 -1.17208207e+00 -4.37554628e-01 4.42309409e-01
8.49800408e-01 -3.39791141e-02 1.24247599e+00 -7.74313211e-02
-3.03803552e-02 4.05750215e-01 2.60859150e-02 3.22328299e-01
5.45907438e-01 -7.74839938e-01 1.98094472e-01 4.83924270e-01
1.07485920e-01 6.51505589e-01 6.36278927e-01 -9.38926399e-01
-1.79828179e+00 -8.02854121e-01 7.74442255e-01 -7.67367303e-01
5.14437199e-01 8.07800591e-02 -1.44044325e-01 9.05518353e-01
-6.17668867e-01 -2.51996785e-01 8.07798803e-01 -1.49593145e-01
5.30297697e-01 2.61718512e-01 -8.74958694e-01 7.45130897e-01
1.43566060e+00 -1.22092068e-01 -4.48103547e-01 3.37522894e-01
3.84385884e-01 -1.18175793e+00 -1.17710209e+00 9.00016129e-01
1.41298628e+00 -3.30397487e-01 8.14339876e-01 -6.70845568e-01
1.98600993e-01 9.87906083e-02 4.92501780e-02 -1.23889375e+00
4.43126447e-02 -5.99701941e-01 -9.50896926e-03 -5.05454354e-02
5.06492317e-01 -4.36050832e-01 1.03911471e+00 6.29232168e-01
-5.34636736e-01 -1.08849919e+00 -1.15744185e+00 -7.53022254e-01
-3.00653666e-01 -9.51302052e-01 2.32486740e-01 4.52916086e-01
7.24136949e-01 -3.62693578e-01 -3.07497382e-01 -6.81041032e-02
-6.01096898e-02 -2.96156425e-02 7.99373269e-01 -9.66556907e-01
-1.76832378e-01 -6.03609145e-01 -8.53852987e-01 -1.05931258e+00
-2.60447711e-01 -9.42529202e-01 6.06742203e-01 -1.71099830e+00
-2.92501301e-01 -4.15442765e-01 -1.39261544e-01 5.70215881e-01
2.24301353e-01 -9.00863335e-02 2.01943472e-01 3.72465461e-01
-1.62710492e-02 3.52822751e-01 1.24110198e+00 2.95036465e-01
-4.94642109e-01 4.89709437e-01 2.04663366e-01 4.50671911e-01
6.69189215e-01 -5.27840316e-01 -3.72443378e-01 -1.24658406e-01
-3.16583008e-01 7.54723847e-01 2.50056207e-01 -1.10979342e+00
4.33644503e-01 -3.75150472e-01 2.41760090e-01 -3.66799891e-01
3.11574250e-01 -1.04873705e+00 3.82753700e-01 1.38737655e+00
-3.86514544e-01 4.57204357e-02 4.08696830e-01 6.11409366e-01
-5.64892031e-02 6.41725510e-02 1.12190783e-01 -6.42447099e-02
-1.16008866e+00 3.91042382e-01 -8.05275381e-01 -6.36166394e-01
1.20185483e+00 -4.76900458e-01 1.83202118e-01 -2.31846988e-01
-1.19820094e+00 3.65714240e-03 8.20256472e-02 9.48900998e-01
9.97531116e-01 -1.07385778e+00 -5.46883941e-01 2.94926971e-01
4.66195941e-01 -1.68738157e-01 1.45398438e-01 1.48405623e+00
-9.94183421e-01 7.65843093e-01 -2.80756146e-01 -6.09808624e-01
-1.46706617e+00 1.71812773e-01 3.15678984e-01 -6.11903071e-01
-9.57870305e-01 6.70542657e-01 -2.57871985e-01 -7.56359458e-01
4.30423945e-01 -7.62358665e-01 -1.58018321e-01 -4.40549493e-01
4.71830964e-01 1.87144756e-01 1.82189807e-01 -3.22730720e-01
-5.75248659e-01 3.81535202e-01 2.33331516e-01 -3.96286130e-01
1.04632246e+00 1.40352562e-01 3.56627196e-01 5.34041047e-01
9.74082768e-01 -4.94779855e-01 -9.56173122e-01 1.07837491e-01
4.35928166e-01 -1.98524311e-01 -3.46118100e-02 -1.29222727e+00
-5.83464146e-01 6.17687464e-01 1.14705145e+00 -7.16158271e-01
1.13170755e+00 1.82100963e-02 8.76194060e-01 3.35203648e-01
9.37632442e-01 -6.25440061e-01 -1.28234580e-01 6.04618728e-01
9.98840332e-01 -9.68239486e-01 -1.73560515e-01 -3.46756101e-01
-1.05447769e+00 1.28182805e+00 3.68060023e-01 1.45705581e-01
6.07424915e-01 4.33694899e-01 4.68393147e-01 -1.21845104e-01
-5.36631167e-01 -5.29808700e-02 7.17035234e-01 5.60486019e-01
7.68262446e-01 4.29381609e-01 -6.17538393e-01 5.29381752e-01
-2.45538652e-01 6.01776421e-01 1.92540094e-01 1.22761166e+00
-1.46936417e-01 -1.13845634e+00 -1.03394717e-01 7.90712893e-01
-1.22311518e-01 1.98570974e-02 1.20115476e-02 1.02559125e+00
-4.06658053e-01 5.19137919e-01 -2.95318048e-02 -7.92132616e-01
5.63676000e-01 3.58695328e-01 6.42301917e-01 -5.61813951e-01
-7.30473816e-01 -1.99611112e-01 3.15350682e-01 -1.34687483e+00
-1.25347942e-01 -7.08297014e-01 -1.64610636e+00 2.08094776e-01
5.18645830e-02 -9.86792296e-02 1.08339000e+00 1.10177100e+00
2.93763816e-01 7.88020134e-01 1.34488633e-02 -9.51099813e-01
-6.27651870e-01 -1.36760044e+00 -2.50322193e-01 5.60455918e-02
2.62760758e-01 -4.83876735e-01 3.26885730e-01 1.46843851e-01] | [14.047775268554688, -3.368760585784912] |
8ddd5c6c-e97f-4339-9cc0-1b05e89be5a2 | neural-machine-translation-by-jointly | 1409.0473 | null | http://arxiv.org/abs/1409.0473v7 | http://arxiv.org/pdf/1409.0473v7.pdf | Neural Machine Translation by Jointly Learning to Align and Translate | Neural machine translation is a recently proposed approach to machine
translation. Unlike the traditional statistical machine translation, the neural
machine translation aims at building a single neural network that can be
jointly tuned to maximize the translation performance. The models proposed
recently for neural machine translation often belong to a family of
encoder-decoders and consists of an encoder that encodes a source sentence into
a fixed-length vector from which a decoder generates a translation. In this
paper, we conjecture that the use of a fixed-length vector is a bottleneck in
improving the performance of this basic encoder-decoder architecture, and
propose to extend this by allowing a model to automatically (soft-)search for
parts of a source sentence that are relevant to predicting a target word,
without having to form these parts as a hard segment explicitly. With this new
approach, we achieve a translation performance comparable to the existing
state-of-the-art phrase-based system on the task of English-to-French
translation. Furthermore, qualitative analysis reveals that the
(soft-)alignments found by the model agree well with our intuition. | ['Kyunghyun Cho', 'Yoshua Bengio', 'Dzmitry Bahdanau'] | 2014-09-01 | null | null | null | null | ['bangla-spelling-error-correction'] | ['natural-language-processing'] | [ 6.51026607e-01 4.78096902e-01 -4.92339015e-01 -4.84126985e-01
-1.29237151e+00 -6.00202262e-01 6.61758065e-01 -1.72751203e-01
-2.85340011e-01 1.03990149e+00 5.10267317e-01 -8.67818177e-01
5.11374593e-01 -7.66668260e-01 -1.20130384e+00 -4.04237807e-01
4.06114161e-01 8.00173223e-01 -2.28457525e-01 -5.44910550e-01
1.85203463e-01 4.51819086e-03 -8.88697565e-01 5.48220515e-01
9.21797633e-01 4.22369510e-01 3.82895797e-01 5.72582901e-01
-2.50117153e-01 2.89296895e-01 -3.57548982e-01 -9.74492550e-01
4.10986960e-01 -1.01349795e+00 -9.28313494e-01 -1.55026510e-01
2.13861078e-01 -3.20719182e-01 -1.13798633e-01 1.18867517e+00
5.13543785e-01 -4.19871449e-01 6.75116718e-01 -6.58549845e-01
-1.15937436e+00 9.60054696e-01 -2.48424172e-01 1.68069527e-01
4.32728887e-01 -7.54037276e-02 1.29658401e+00 -1.29639733e+00
8.00374508e-01 1.05821300e+00 4.41549301e-01 7.08054066e-01
-1.21162045e+00 -3.27085078e-01 -2.20335007e-01 -1.08086905e-02
-1.15601814e+00 -6.93016052e-01 5.24151623e-01 -3.14939052e-01
1.48171449e+00 1.99190706e-01 5.20615160e-01 1.29454744e+00
7.64934599e-01 6.13945544e-01 9.14973974e-01 -8.69477510e-01
3.03792246e-02 2.38065228e-01 -3.28759462e-01 5.46569049e-01
-1.45989522e-01 2.48428002e-01 -6.29800320e-01 -1.33767143e-01
7.66231239e-01 -3.95236433e-01 -1.46181479e-01 5.12536243e-02
-1.50067830e+00 9.55440104e-01 2.33147204e-01 3.70461702e-01
-5.85844994e-01 1.16360314e-01 3.94636422e-01 6.54306531e-01
7.81553507e-01 6.84795260e-01 -6.98657572e-01 -3.45919341e-01
-1.10745370e+00 8.52853805e-02 1.04784989e+00 1.01707089e+00
7.68995047e-01 -7.16624856e-02 -4.72939849e-01 7.54625261e-01
2.83215851e-01 5.54679155e-01 5.60539663e-01 -4.97251391e-01
9.46510255e-01 5.21943986e-01 -6.92586042e-03 -4.66052026e-01
1.85455903e-01 -5.99704325e-01 -7.71373034e-01 -2.43002996e-01
1.23078063e-01 -4.09894109e-01 -7.64882743e-01 1.75055170e+00
-7.23450631e-02 -1.11436509e-01 3.10134262e-01 1.04566109e+00
2.69841313e-01 1.10690570e+00 -3.83504122e-01 -3.77587587e-01
1.11948025e+00 -1.32680404e+00 -6.81119561e-01 -5.15224457e-01
6.07034206e-01 -1.15274799e+00 1.03264284e+00 1.40577002e-04
-1.55130112e+00 -3.92860055e-01 -9.24319804e-01 -1.39095709e-01
-1.25147939e-01 2.50504643e-01 3.94402176e-01 3.33184063e-01
-1.29374552e+00 7.20334649e-01 -5.21751106e-01 -4.04032677e-01
-1.75657291e-02 5.60167134e-01 -3.61457616e-01 7.76805356e-02
-1.48522270e+00 1.54495692e+00 3.26255172e-01 3.15221727e-01
-7.59781182e-01 -1.83911398e-01 -6.20253086e-01 8.98371339e-02
-8.52194726e-02 -1.09951162e+00 1.49030709e+00 -1.61670196e+00
-1.84494233e+00 8.62624109e-01 -6.68320954e-01 -5.47033966e-01
2.95576274e-01 -1.53900281e-01 -6.94723502e-02 -1.14391789e-01
2.05051452e-01 8.55935693e-01 8.28739464e-01 -8.62472117e-01
-5.15901327e-01 -6.76148236e-02 -5.31821996e-02 3.39641392e-01
-1.78551361e-01 5.71410656e-01 -3.36328238e-01 -5.99995792e-01
-6.81444481e-02 -1.00496233e+00 -4.09889579e-01 -3.60079914e-01
-4.23622340e-01 -6.70049340e-02 6.92366213e-02 -8.42429221e-01
1.32033849e+00 -1.63774252e+00 8.01153839e-01 -1.57412022e-01
-2.40835384e-01 1.96350828e-01 -4.50110406e-01 8.17392588e-01
-4.24884856e-02 1.82632193e-01 -4.90424931e-01 -3.76375973e-01
-3.29914801e-02 2.97457546e-01 -5.04356503e-01 2.52219081e-01
5.90995431e-01 1.42171192e+00 -7.68304586e-01 -3.02088946e-01
-2.57702947e-01 3.80461484e-01 -4.66970712e-01 4.49382156e-01
-3.73042256e-01 2.82588720e-01 -3.01006883e-01 5.03398061e-01
3.36204678e-01 -7.43535534e-02 8.15898553e-02 3.20443273e-01
-9.17984620e-02 9.03609693e-01 -3.42032015e-01 1.95664108e+00
-4.67282981e-01 6.80464625e-01 -4.54230867e-02 -9.07950819e-01
1.07517982e+00 6.37551069e-01 7.14408308e-02 -6.75722122e-01
1.97193742e-01 7.34187245e-01 3.11024636e-01 -3.87313664e-01
5.02227724e-01 -3.04940194e-01 -9.13877040e-02 7.68606424e-01
2.62759537e-01 -3.23607028e-02 2.16391176e-01 -1.58688664e-01
9.98566926e-01 4.38854039e-01 3.03253293e-01 -2.70902753e-01
4.61422533e-01 2.53246367e-01 3.89541626e-01 4.43186730e-01
1.90676197e-01 7.55610883e-01 3.46602619e-01 -5.57896852e-01
-1.72977805e+00 -7.68030167e-01 2.82671332e-01 1.06918788e+00
-2.77435809e-01 -3.21523070e-01 -1.09157407e+00 -5.72860003e-01
-5.06984174e-01 7.77915120e-01 -5.05936027e-01 -1.75950751e-01
-9.60537672e-01 -7.28348613e-01 5.83336294e-01 3.70677769e-01
1.86572429e-02 -1.15575051e+00 -2.11825237e-01 6.14929855e-01
-6.27374172e-01 -9.01565194e-01 -7.24325120e-01 4.72637326e-01
-1.02253664e+00 -3.79713893e-01 -8.36655676e-01 -1.21404803e+00
7.97073603e-01 -6.58631101e-02 1.49141371e+00 6.30272180e-02
3.69953364e-01 -5.07946134e-01 -4.13108051e-01 -2.69616097e-01
-1.00126827e+00 5.94038785e-01 2.38078516e-02 -1.16927661e-01
6.72359824e-01 -5.15220702e-01 -2.02506661e-01 1.10388562e-01
-6.66098773e-01 6.27512872e-01 1.20116937e+00 9.25872684e-01
4.68151778e-01 -7.77897120e-01 5.87316632e-01 -5.22693455e-01
1.13622046e+00 -5.21761954e-01 -2.99603283e-01 3.66894037e-01
-6.54783547e-01 2.56610304e-01 7.59233534e-01 -5.09829640e-01
-5.09798765e-01 1.91101655e-01 -3.81613463e-01 -1.58553615e-01
4.62913513e-02 7.37162471e-01 1.54634640e-01 1.52237162e-01
6.56092048e-01 8.35433304e-01 1.46588981e-01 -3.68250996e-01
4.18563455e-01 8.80532444e-01 2.00871930e-01 -5.01462281e-01
7.45309770e-01 -3.17237258e-01 -1.24004789e-01 -1.33479580e-01
-6.53085470e-01 -8.21247920e-02 -9.99084830e-01 2.00095046e-02
7.79089332e-01 -9.32833254e-01 -4.09113131e-02 -1.11486278e-01
-1.79629946e+00 -2.29795933e-01 -8.16438496e-02 4.57940668e-01
-8.61412466e-01 -2.52898820e-02 -8.02167237e-01 -4.56754923e-01
-7.81064510e-01 -1.37984455e+00 1.21528876e+00 -1.15401939e-01
-5.27340174e-01 -8.90941083e-01 2.49513626e-01 1.95331171e-01
6.39489770e-01 -3.30889672e-01 9.59357798e-01 -7.49798059e-01
-4.72913235e-01 -4.90502045e-02 -1.51585564e-01 3.87414515e-01
-8.60101208e-02 -2.31562898e-01 -4.47642207e-01 -2.07351238e-01
-1.59816891e-02 -2.21746087e-01 6.21517897e-01 1.77683651e-01
3.61927807e-01 -6.89609110e-01 -9.78795886e-02 4.09735739e-01
1.29842007e+00 7.43358582e-02 6.31574988e-01 3.14496547e-01
3.67012531e-01 4.95552003e-01 3.06592673e-01 -1.73105061e-01
3.53514999e-01 7.49443889e-01 3.72025609e-01 -2.29057372e-01
3.39524597e-02 -4.85401958e-01 9.43980753e-01 1.21361828e+00
-1.36219665e-01 -3.02493423e-01 -8.20424557e-01 7.17100680e-01
-2.01627445e+00 -6.79013789e-01 -5.97317554e-02 1.93406308e+00
1.28993022e+00 6.84675723e-02 -1.89664871e-01 -3.00102800e-01
5.55802763e-01 -9.51651260e-02 -1.34498373e-01 -1.08078432e+00
-2.74718665e-02 3.47458243e-01 5.61069489e-01 8.10174286e-01
-6.47406399e-01 1.40517879e+00 7.44253731e+00 6.91092014e-01
-1.10583186e+00 3.38616222e-01 5.62206745e-01 -8.78024474e-02
-6.11952364e-01 1.66813165e-01 -8.54131222e-01 3.44136059e-01
1.46934056e+00 -1.56424925e-01 7.43740499e-01 5.54816663e-01
4.70147133e-01 4.22441483e-01 -1.45671475e+00 4.81456816e-01
2.37302974e-01 -1.49403858e+00 2.58758873e-01 2.08195299e-01
8.91068757e-01 2.64985770e-01 -1.09478369e-01 3.04775387e-01
1.71830967e-01 -1.26661658e+00 7.95614600e-01 4.42327440e-01
8.69145274e-01 -6.24492168e-01 8.93738866e-01 6.88159645e-01
-5.59948802e-01 1.48562729e-01 -7.15658963e-01 -3.25985909e-01
3.85604441e-01 3.75812531e-01 -1.09734333e+00 4.29849356e-01
2.58442964e-02 4.57790971e-01 -3.11450154e-01 6.71868265e-01
-5.85777700e-01 9.68559682e-01 7.53509030e-02 -2.95487970e-01
5.70843041e-01 -4.02713090e-01 5.44105470e-01 1.34383321e+00
7.08343565e-01 -2.58751810e-01 -1.24263782e-02 1.07056904e+00
-2.90915906e-01 3.63278180e-01 -6.83949769e-01 -2.94473588e-01
1.05658211e-01 8.73052299e-01 -1.61325812e-01 -4.46762502e-01
-6.28453970e-01 1.44562531e+00 5.34436524e-01 2.93681026e-01
-7.27223933e-01 -1.58045903e-01 5.78638971e-01 3.21240909e-02
3.02418739e-01 -3.34299445e-01 -6.05174720e-01 -1.30151212e+00
2.35879496e-01 -1.12344539e+00 -4.40062284e-01 -9.05258179e-01
-1.03856218e+00 1.10028934e+00 -4.27480340e-01 -1.38006926e+00
-8.21441531e-01 -3.77335578e-01 -4.67167288e-01 1.50210333e+00
-1.40649199e+00 -1.35308349e+00 7.11549342e-01 1.85047135e-01
8.52378130e-01 -4.30289894e-01 1.22116387e+00 2.09277704e-01
-3.12859774e-01 5.73535800e-01 5.19315079e-02 2.16134921e-01
7.44696677e-01 -9.84473169e-01 1.04134405e+00 1.05050147e+00
4.64320362e-01 7.93348968e-01 8.61431479e-01 -6.76442444e-01
-1.56156909e+00 -1.00405347e+00 2.10311460e+00 -5.75460553e-01
5.77320457e-01 -5.33684134e-01 -8.03645432e-01 6.60324574e-01
8.67124021e-01 -4.84744042e-01 6.26812756e-01 -1.52331288e-03
-1.45868421e-01 2.70071119e-01 -7.52766371e-01 6.34667873e-01
9.00114417e-01 -6.00632846e-01 -9.82417166e-01 4.75448042e-01
9.92951632e-01 -4.87194628e-01 -6.02139652e-01 1.09734587e-01
5.14285743e-01 -5.89080334e-01 6.86607778e-01 -9.35127139e-01
1.32865298e+00 -5.58201224e-02 -2.12718219e-01 -1.66402125e+00
-5.01652896e-01 -8.58930349e-01 -1.23619124e-01 7.69300103e-01
1.21752620e+00 -3.66639018e-01 6.72406614e-01 2.13412955e-01
-3.55304092e-01 -1.13972139e+00 -1.09861147e+00 -5.07773578e-01
5.61262965e-01 -1.25766635e-01 6.87605381e-01 6.49945557e-01
1.53200075e-01 8.37375581e-01 -6.63815737e-01 -1.47740589e-02
1.87585518e-01 8.49244520e-02 5.53949416e-01 -7.97800124e-01
-4.83849794e-01 -5.31892121e-01 6.22978546e-02 -1.46806765e+00
3.17182750e-01 -1.21911621e+00 2.60028243e-01 -1.80646265e+00
3.72321278e-01 3.10210697e-02 -1.09773159e-01 5.05684376e-01
-2.44822398e-01 3.44365537e-01 -4.89246659e-02 4.50599045e-01
-4.96909097e-02 4.68749374e-01 1.47936714e+00 -3.16442065e-02
-1.15690883e-02 7.08464608e-02 -8.95129144e-01 2.39667282e-01
7.01355517e-01 -8.48157227e-01 -5.54661192e-02 -1.03319979e+00
6.74335837e-01 3.07345390e-01 -1.06559336e-01 -2.55045861e-01
2.08939761e-01 -2.15358913e-01 2.01820299e-01 -3.88429999e-01
1.98263839e-01 -6.91074371e-01 -4.23877761e-02 4.26769048e-01
-6.15223348e-01 6.22146606e-01 -7.23467246e-02 1.51861086e-01
-5.30517340e-01 -3.53163749e-01 4.47963148e-01 -3.69785994e-01
-1.46619931e-01 1.18863255e-01 -6.15103126e-01 -3.05451393e-01
6.65171444e-01 -6.62430972e-02 8.10568705e-02 -4.24332201e-01
-3.92425358e-01 -1.00973129e-01 4.18169558e-01 6.82533979e-01
4.74730819e-01 -1.43949986e+00 -1.27204037e+00 3.13853502e-01
-8.24756622e-02 -3.99620593e-01 -6.12981498e-01 7.54140794e-01
-4.23440784e-01 7.55787551e-01 -2.65097558e-01 -3.49927694e-01
-9.58773732e-01 4.29332674e-01 2.91882932e-01 -4.49314892e-01
-3.29060555e-01 8.60400379e-01 -3.45854133e-01 -6.57572687e-01
-1.68177962e-01 -2.96082407e-01 1.73605949e-01 -3.53553563e-01
4.47232515e-01 -2.00271115e-01 1.17137298e-01 -8.45508695e-01
-1.91948488e-01 4.02865857e-01 -2.62295199e-03 -5.18842220e-01
1.27258015e+00 -8.83528665e-02 -6.08839095e-01 2.47887686e-01
1.11442065e+00 -1.21380702e-01 -8.01584959e-01 -3.71408194e-01
-4.72380333e-02 -1.56559870e-01 -1.75355703e-01 -9.72964108e-01
-5.87059438e-01 1.06756008e+00 7.94938058e-02 -1.83463797e-01
1.00260532e+00 -1.06095165e-01 1.01687908e+00 3.72224540e-01
4.45802629e-01 -9.35746074e-01 -2.88334876e-01 1.06064069e+00
9.77119982e-01 -1.30698335e+00 -4.20009583e-01 -2.03126267e-01
-6.32271111e-01 1.40546131e+00 2.94661850e-01 -2.14225739e-01
2.16542661e-01 3.92562598e-01 2.40873754e-01 2.94454485e-01
-1.21960676e+00 -3.00342180e-02 5.37416160e-01 2.38623157e-01
8.65037560e-01 2.36570448e-01 -8.85046482e-01 5.03956079e-01
-5.01083672e-01 9.91743952e-02 3.99176538e-01 6.80514753e-01
-5.31879187e-01 -1.76946568e+00 -1.44231573e-01 1.81114063e-01
-6.84470832e-01 -6.21697605e-01 -8.75741065e-01 1.06680118e-01
3.90694179e-02 8.93035889e-01 -3.00133750e-02 -5.89681506e-01
1.56366408e-01 4.98163640e-01 4.99016792e-01 -9.20933604e-01
-1.08213139e+00 1.64284632e-01 2.41282016e-01 -3.56622279e-01
-1.65105820e-01 -4.07349765e-01 -8.33910704e-01 -1.34370580e-01
-3.36306840e-01 3.15039456e-01 1.00028908e+00 1.15567791e+00
5.18367589e-01 2.61918843e-01 6.34418368e-01 -6.42803252e-01
-9.31400359e-01 -1.29775167e+00 1.54606462e-01 3.28504154e-03
3.53529692e-01 1.26559719e-01 -5.57471849e-02 1.48157120e-01] | [11.630714416503906, 10.28522777557373] |
efd32e57-a005-4d6c-807d-70cc41df5982 | convolutional-neural-network-compression | 2109.14710 | null | https://arxiv.org/abs/2109.14710v2 | https://arxiv.org/pdf/2109.14710v2.pdf | Convolutional Neural Network Compression through Generalized Kronecker Product Decomposition | Modern Convolutional Neural Network (CNN) architectures, despite their superiority in solving various problems, are generally too large to be deployed on resource constrained edge devices. In this paper, we reduce memory usage and floating-point operations required by convolutional layers in CNNs. We compress these layers by generalizing the Kronecker Product Decomposition to apply to multidimensional tensors, leading to the Generalized Kronecker Product Decomposition (GKPD). Our approach yields a plug-and-play module that can be used as a drop-in replacement for any convolutional layer. Experimental results for image classification on CIFAR-10 and ImageNet datasets using ResNet, MobileNetv2 and SeNet architectures substantiate the effectiveness of our proposed approach. We find that GKPD outperforms state-of-the-art decomposition methods including Tensor-Train and Tensor-Ring as well as other relevant compression methods such as pruning and knowledge distillation. | ['Vahid Partovi Nia', 'Ali Mosleh', 'Marzieh S. Tahaei', 'Marawan Gamal Abdel Hameed'] | 2021-09-29 | null | null | null | null | ['neural-network-compression', 'neural-network-compression'] | ['methodology', 'miscellaneous'] | [-2.26592690e-01 -1.12891428e-01 -5.83876781e-02 -3.56207401e-01
1.30847588e-01 -2.99355417e-01 2.17067137e-01 -1.61181599e-01
-9.17064250e-01 6.28111362e-01 -1.74807549e-01 -8.38002980e-01
-4.09605682e-01 -9.17193353e-01 -8.34203660e-01 -5.34692705e-01
-1.34289786e-01 -3.86993513e-02 4.39804465e-01 -1.53111741e-01
3.97832766e-02 5.71465433e-01 -1.35373688e+00 3.42433602e-01
3.91802371e-01 1.46559477e+00 2.32442483e-01 6.74366713e-01
-3.73522528e-02 1.01444614e+00 -2.85328776e-01 -1.06825554e+00
5.79951227e-01 3.81867707e-01 -1.02156997e+00 -2.26402029e-01
4.56445515e-01 -5.88502288e-01 -8.13131690e-01 9.93910670e-01
4.38237578e-01 -3.35208550e-02 1.27917215e-01 -1.01059520e+00
-5.13253033e-01 7.33100951e-01 -2.93821394e-01 6.88694358e-01
-3.91613990e-01 -2.74240047e-01 9.83659089e-01 -8.72858822e-01
3.10790777e-01 1.00737977e+00 1.02523625e+00 4.69247580e-01
-8.80735338e-01 -7.25116670e-01 -5.94722517e-02 4.92455810e-01
-1.28098440e+00 -5.58277905e-01 5.49298823e-01 -1.66142270e-01
1.60323024e+00 8.67506340e-02 8.02969635e-01 7.12814748e-01
7.51305446e-02 5.74520707e-01 4.59607899e-01 -1.73501268e-01
3.24313939e-01 -3.84764880e-01 1.99367031e-01 1.09552598e+00
7.78639913e-01 -2.85160989e-01 -7.17512310e-01 -1.40078053e-01
8.35279405e-01 1.65752187e-01 -1.66861370e-01 6.30596140e-03
-1.18501329e+00 5.58404982e-01 5.72760880e-01 -8.50840006e-03
-6.68590486e-01 8.58358324e-01 9.65215445e-01 4.91456419e-01
5.02118051e-01 1.76810026e-02 -7.59792745e-01 -2.92721063e-01
-8.82343769e-01 2.65629262e-01 9.95977163e-01 1.12798965e+00
5.99544525e-01 2.78445482e-01 1.55292869e-01 7.57012427e-01
1.64658606e-01 -8.59367289e-03 5.18346369e-01 -8.75209749e-01
6.49624765e-01 5.23318470e-01 -5.35549521e-01 -7.15331376e-01
-2.65427798e-01 -6.88112080e-01 -1.25247121e+00 -4.04906161e-02
-1.22807838e-01 -3.66813958e-01 -9.20858324e-01 1.20567477e+00
2.32332334e-01 5.28781295e-01 3.90297174e-01 6.34312809e-01
9.25212502e-01 2.44619578e-01 -9.13828388e-02 2.78631598e-01
1.49804974e+00 -1.14807296e+00 -4.75064665e-01 -2.80232728e-03
1.09872770e+00 -7.05156922e-01 3.91132176e-01 5.38431764e-01
-1.16665089e+00 -3.65485847e-01 -1.32426202e+00 -3.81148368e-01
-4.11869794e-01 3.57792825e-01 1.22984278e+00 7.61634648e-01
-1.33474696e+00 1.00853431e+00 -1.31411135e+00 1.08824866e-02
9.77121592e-01 7.68425167e-01 -5.93198419e-01 -6.14169426e-02
-7.86841273e-01 6.38870418e-01 8.51091206e-01 3.72384518e-01
-8.98367107e-01 -1.09807909e+00 -5.63831151e-01 2.87882626e-01
8.54971856e-02 -8.96920145e-01 1.33257365e+00 -3.92875046e-01
-1.33172226e+00 4.54246312e-01 7.36305490e-02 -1.09882641e+00
1.23323709e-01 -2.07792446e-01 -1.40634254e-01 3.05581629e-01
-3.98015767e-01 6.21889949e-01 6.74564004e-01 -3.73526692e-01
-6.15401387e-01 -1.78487018e-01 4.63766128e-01 -1.06979705e-01
-7.56213725e-01 -4.72491942e-02 -3.51355851e-01 -4.92006809e-01
5.40090740e-01 -8.21070850e-01 -2.64956087e-01 1.40480086e-01
-3.16839844e-01 -2.48872921e-01 1.05578232e+00 -4.95772600e-01
1.14998853e+00 -2.08501601e+00 2.55934335e-02 4.91149649e-02
4.88668412e-01 7.25300491e-01 -5.48814423e-03 1.23425357e-01
-1.88016683e-01 1.67485088e-01 2.62691885e-01 -7.22851157e-01
-9.65136066e-02 7.84854949e-01 -2.26634234e-01 4.54024941e-01
2.45889768e-01 7.46816993e-01 -7.13360846e-01 -3.03521693e-01
-1.72141925e-01 7.62307286e-01 -1.01728702e+00 -4.11199987e-01
-7.00850599e-03 -3.66391957e-01 -3.04049581e-01 5.87698638e-01
8.60374510e-01 -2.28690892e-01 1.37050167e-01 -6.57137930e-01
-4.79071289e-02 6.70636356e-01 -8.71101856e-01 1.80660701e+00
-5.54709375e-01 8.56589437e-01 -4.39950079e-03 -1.24255061e+00
7.78904617e-01 5.45619488e-01 2.57376432e-01 -2.33939439e-01
2.77012438e-01 4.19257909e-01 1.05528913e-01 -3.84228885e-01
7.06576943e-01 1.93382636e-01 5.29246628e-01 8.22184756e-02
6.24709427e-01 3.53681743e-01 5.44689357e-01 3.43536168e-01
1.38358235e+00 -2.61295974e-01 -2.31134087e-01 -3.17099541e-01
3.33381772e-01 -2.08442569e-01 5.13514876e-01 3.50773394e-01
1.08918428e-01 2.07572103e-01 8.54534328e-01 -7.66414881e-01
-1.14837015e+00 -6.44508958e-01 -1.62215292e-01 8.98210585e-01
-4.77132380e-01 -8.71369064e-01 -8.36139858e-01 -4.60561574e-01
-1.89213797e-01 4.73422706e-02 -4.67119396e-01 -1.71975344e-02
-4.67360497e-01 -7.71365881e-01 1.09509635e+00 9.28495347e-01
1.27697289e+00 -4.75731462e-01 -7.86426723e-01 2.28951260e-01
1.63030654e-01 -1.42221999e+00 6.43936396e-02 5.02654135e-01
-1.52181411e+00 -9.76192653e-01 -4.50047940e-01 -8.37699711e-01
7.02789247e-01 4.83548045e-01 8.95580947e-01 1.77780315e-01
-1.51495829e-01 1.70802195e-02 -4.14862186e-01 -4.19911385e-01
1.40767351e-01 3.03585261e-01 1.53068289e-01 -1.50847822e-01
3.30833644e-01 -1.00336921e+00 -7.05122232e-01 -5.41447364e-02
-1.00303996e+00 1.37815520e-01 6.69889390e-01 9.02590573e-01
4.20710027e-01 5.07511199e-01 -9.08648744e-02 -7.12333024e-01
7.18771815e-01 -2.87786067e-01 -6.79780662e-01 8.34634230e-02
-5.66873193e-01 2.73858994e-01 5.96995533e-01 -3.75589699e-01
-6.43062174e-01 3.89359519e-02 -1.32121459e-01 -9.29937005e-01
3.75316858e-01 8.42686474e-01 2.49881864e-01 -6.09448254e-01
5.74212968e-01 5.17975986e-02 -2.09179178e-01 -8.04514766e-01
6.43568635e-02 3.78830612e-01 6.54951155e-01 -6.24342382e-01
4.45824802e-01 5.56311548e-01 5.00577390e-01 -7.11323798e-01
-5.55895865e-01 -3.49786431e-01 -5.04897594e-01 2.18723923e-01
6.78066313e-01 -1.24674916e+00 -1.02865767e+00 4.57510918e-01
-1.65522969e+00 -8.52793828e-02 -3.75488810e-02 6.77780092e-01
-1.97983030e-02 9.65975896e-02 -8.57287109e-01 -4.03084099e-01
-5.94015360e-01 -1.18190086e+00 6.16443455e-01 8.60464349e-02
5.38121283e-01 -1.00520027e+00 -3.28065693e-01 3.02824020e-01
6.28069401e-01 -7.89346099e-02 7.41648495e-01 -5.73999286e-01
-8.34526718e-01 -2.45946735e-01 -6.42238438e-01 8.91974390e-01
-5.09162724e-01 -1.09725662e-01 -8.86931062e-01 -2.60437220e-01
-1.85239628e-01 -2.52993196e-01 1.27878666e+00 7.36185759e-02
1.48315775e+00 -6.75278068e-01 -9.86572355e-02 1.28012729e+00
1.64415276e+00 -1.29399583e-01 8.21123898e-01 2.39715293e-01
8.38602483e-01 -1.70309644e-03 -8.32417980e-02 5.82700312e-01
3.19144011e-01 1.24580510e-01 7.94316828e-01 7.70866573e-02
-2.01015934e-01 5.93703575e-02 -6.15115985e-02 1.23085213e+00
-9.04872894e-01 1.42528757e-01 -1.04715884e+00 6.94086075e-01
-1.83569598e+00 -7.18948007e-01 -1.81760892e-01 1.71893740e+00
7.66958654e-01 2.62565583e-01 -3.96996111e-01 5.55017412e-01
7.19746575e-02 9.07363445e-02 -2.65384108e-01 -5.17413914e-01
1.32256895e-01 8.35852742e-01 1.29883659e+00 -3.03560570e-02
-1.01271975e+00 7.85711884e-01 6.03816605e+00 8.06832194e-01
-1.19201326e+00 5.87853909e-01 4.66711044e-01 -3.88685353e-02
1.62929758e-01 -1.16535136e-02 -8.68674874e-01 -2.43770834e-02
1.18303978e+00 7.64783174e-02 6.47845626e-01 1.18295050e+00
-3.38342905e-01 1.64349064e-01 -1.08426583e+00 1.34757364e+00
-3.97061050e-01 -1.83168828e+00 -7.19664246e-02 1.51228786e-01
6.89776599e-01 5.75412989e-01 -5.83886579e-02 1.46328464e-01
2.80452609e-01 -9.56104219e-01 6.32110775e-01 2.53929138e-01
6.76021218e-01 -7.75618672e-01 1.03566504e+00 -1.30908424e-02
-1.26026762e+00 -2.78116446e-02 -9.74837482e-01 -3.08381319e-01
-3.02744895e-01 7.60207295e-01 -9.40013409e-01 5.24715483e-01
9.62484539e-01 6.56410098e-01 -4.34741557e-01 1.07641256e+00
3.68428901e-02 6.81747615e-01 -4.03103799e-01 -1.11708716e-01
4.46392685e-01 7.80564994e-02 6.37623146e-02 1.27128506e+00
3.63146663e-01 1.27123728e-01 -5.19552708e-01 5.38013518e-01
-5.97825646e-01 -3.19741130e-01 -2.83576339e-01 -3.21420252e-01
3.91808629e-01 1.26001990e+00 -5.97357571e-01 -3.86098534e-01
-6.02456212e-01 8.30369174e-01 4.79207218e-01 4.37506795e-01
-6.99144721e-01 -6.79632008e-01 1.01000059e+00 -1.31940901e-01
8.85511816e-01 -6.13227487e-01 -4.14566964e-01 -1.19701552e+00
3.71084720e-01 -4.09274071e-01 -8.81446712e-03 -5.44326842e-01
-7.71706104e-01 8.43210220e-01 9.95714962e-02 -9.63587642e-01
1.89973310e-01 -1.09914899e+00 -2.93939382e-01 6.93629146e-01
-1.82258904e+00 -1.02269030e+00 -3.75280708e-01 8.21509302e-01
1.21845953e-01 -4.03725296e-01 8.67515206e-01 7.49778152e-01
-6.21838927e-01 6.98871434e-01 9.14826468e-02 4.88540262e-01
-2.27403596e-01 -8.03883672e-01 5.34149110e-01 8.56117308e-01
-3.70956548e-02 8.82017195e-01 2.60215729e-01 -2.26184309e-01
-1.80673611e+00 -1.28482854e+00 8.93280029e-01 4.93355870e-01
9.12862360e-01 -2.45802507e-01 -6.62589192e-01 7.32828617e-01
1.41343638e-01 7.57575393e-01 5.52040040e-01 2.40299717e-01
-7.56234527e-01 -4.62930858e-01 -9.95817959e-01 4.94865507e-01
1.28604162e+00 -6.37470961e-01 -1.45941004e-01 5.73925436e-01
8.25946808e-01 -6.36533558e-01 -1.15954924e+00 3.18732679e-01
4.65479553e-01 -9.68866765e-01 1.02526259e+00 -8.29995692e-01
6.75399840e-01 -4.11593206e-02 -3.07765335e-01 -1.00927651e+00
-2.30310529e-01 -8.08072805e-01 -4.37648594e-01 7.69135952e-01
3.52265924e-01 -7.91746080e-01 1.01508212e+00 5.63983262e-01
-5.33399403e-01 -1.15628695e+00 -1.23114884e+00 -8.26890290e-01
-2.56228745e-01 -7.79352546e-01 7.52355218e-01 7.03973770e-01
-1.15869954e-01 1.08201534e-01 -2.43713055e-02 2.05932125e-01
5.68547964e-01 -7.91958511e-01 4.60470438e-01 -1.08820379e+00
-5.42772174e-01 -3.76164824e-01 -9.45334613e-01 -1.09737659e+00
-6.03484958e-02 -9.90549862e-01 -5.15245795e-01 -1.42036355e+00
-9.43186209e-02 -6.73972905e-01 -3.54941815e-01 8.63687456e-01
6.25849068e-01 5.79159856e-01 2.22266868e-01 2.19952300e-01
-4.81713831e-01 3.16922992e-01 1.08437383e+00 -8.42197016e-02
2.23935708e-01 -2.10934684e-01 -5.18167019e-01 7.58929431e-01
7.92326391e-01 -4.87223208e-01 -5.69554925e-01 -1.08430767e+00
6.73508942e-01 -2.25646377e-01 6.57704830e-01 -1.37060237e+00
7.00378239e-01 4.77859676e-01 7.24049136e-02 -5.66090941e-01
5.18110633e-01 -8.74848902e-01 1.35779396e-01 5.20651460e-01
-2.54464732e-03 3.85012597e-01 4.81948107e-01 3.80050004e-01
-2.41521373e-01 -4.46726412e-01 3.75541925e-01 -8.44371766e-02
-6.33851171e-01 6.20252788e-01 -1.23427119e-02 -2.91874379e-01
6.95935726e-01 -2.14205220e-01 -7.08751678e-01 2.81168371e-01
-5.05009949e-01 -2.05267996e-01 -3.01823169e-01 7.27865100e-03
9.20011163e-01 -1.33711421e+00 -5.66804230e-01 3.02586675e-01
-3.44183683e-01 4.67103720e-01 9.58829746e-02 1.10044897e+00
-1.34751427e+00 8.57821763e-01 -2.52748609e-01 -3.17088664e-01
-1.31530583e+00 1.70070976e-01 2.43049040e-01 -3.55185837e-01
-5.21969676e-01 1.35888898e+00 -3.36409777e-01 8.84992536e-03
4.29820359e-01 -9.09385681e-01 2.12834373e-01 -2.97473997e-01
5.39526463e-01 5.63641846e-01 7.63214052e-01 -1.84408396e-01
-3.25897872e-01 6.07788526e-02 -2.45414525e-01 1.45518392e-01
1.62586081e+00 4.64023024e-01 -5.66467941e-01 -1.66955963e-01
1.51687622e+00 -8.16449225e-01 -1.04404604e+00 -3.58947009e-01
-2.23484874e-01 -2.85234779e-01 7.72541642e-01 -1.89967200e-01
-1.74681461e+00 9.30441141e-01 5.26043892e-01 6.40365854e-02
1.30839252e+00 -2.70089060e-01 1.03259385e+00 1.11513591e+00
3.47173929e-01 -8.83210778e-01 -1.25162393e-01 8.06302428e-01
5.30513108e-01 -7.28032708e-01 2.58790940e-01 -4.72072572e-01
-1.33885399e-01 1.37757707e+00 3.39121163e-01 -3.42285186e-01
1.26504445e+00 4.59066272e-01 -5.05801141e-01 -2.28284046e-01
-1.27769136e+00 3.72820124e-02 1.39552429e-01 4.30912763e-01
5.04937410e-01 -2.95547047e-03 -3.97126764e-01 3.34628493e-01
-5.35332859e-01 3.53803307e-01 4.33687419e-01 1.14507306e+00
-5.07687069e-02 -9.62479234e-01 -3.90578876e-03 6.20900214e-01
-8.26357126e-01 -5.11684775e-01 2.78443664e-01 4.95536178e-01
4.46119189e-01 4.85967308e-01 2.76248753e-01 -8.44051838e-01
8.82173851e-02 -1.22198448e-01 5.61376631e-01 -3.47294241e-01
-7.08599985e-01 -3.84033710e-01 3.58627081e-01 -5.75615525e-01
-6.24512374e-01 -3.49682838e-01 -1.16440463e+00 -8.43842030e-01
-4.10715520e-01 -2.11632237e-01 1.35550046e+00 7.92470694e-01
6.28636062e-01 9.74742413e-01 -1.79380365e-02 -7.88835526e-01
-6.03323519e-01 -9.89726126e-01 -5.22993207e-01 -1.29985452e-01
2.98991621e-01 -5.87997556e-01 -2.13769749e-02 1.74325079e-01] | [8.5014066696167, 3.0536675453186035] |
4aa4c810-5804-4882-94c9-abba3c6b71ca | zero-shot-learning-for-joint-intent-and-slot | 2212.07922 | null | https://arxiv.org/abs/2212.07922v1 | https://arxiv.org/pdf/2212.07922v1.pdf | Zero-Shot Learning for Joint Intent and Slot Labeling | It is expensive and difficult to obtain the large number of sentence-level intent and token-level slot label annotations required to train neural network (NN)-based Natural Language Understanding (NLU) components of task-oriented dialog systems, especially for the many real world tasks that have a large and growing number of intents and slot types. While zero shot learning approaches that require no labeled examples -- only features and auxiliary information -- have been proposed only for slot labeling, we show that one can profitably perform joint zero-shot intent classification and slot labeling. We demonstrate the value of capturing dependencies between intents and slots, and between different slots in an utterance in the zero shot setting. We describe NN architectures that translate between word and sentence embedding spaces, and demonstrate that these modifications are required to enable zero shot learning for this task. We show a substantial improvement over strong baselines and explain the intuition behind each architectural modification through visualizations and ablation studies. | ['Balakrishnan Narayanaswamy', 'Rashmi Gangadharaiah'] | 2022-11-29 | null | null | null | null | ['intent-classification'] | ['natural-language-processing'] | [ 3.62120152e-01 6.52817369e-01 -3.73302341e-01 -7.08037615e-01
-7.30047464e-01 -4.28221345e-01 7.85427630e-01 2.61451393e-01
-6.46703184e-01 5.49655855e-01 7.03920960e-01 -7.50098944e-01
1.81312054e-01 -6.52779818e-01 -4.79801707e-02 -1.76807478e-01
-1.47866309e-01 6.04641795e-01 8.00123736e-02 -5.56804836e-01
7.42023736e-02 -7.26529285e-02 -1.33518445e+00 2.41103441e-01
4.46319968e-01 7.24959135e-01 8.28461796e-02 8.79771411e-01
-8.23180676e-01 9.37361181e-01 -5.81748426e-01 -2.33694762e-01
8.33770335e-02 -3.56347203e-01 -1.11269975e+00 -1.57202825e-01
2.81032264e-01 -4.84815687e-01 -2.62201488e-01 5.72311163e-01
4.46638376e-01 6.50610805e-01 7.33317018e-01 -1.28836584e+00
-6.04762733e-01 6.20674491e-01 -1.69215441e-01 1.84423506e-01
2.75729269e-01 2.60748893e-01 1.49846780e+00 -7.86356211e-01
6.53975010e-01 1.38394356e+00 5.95341384e-01 1.00039363e+00
-1.21588910e+00 -3.06255490e-01 1.21774264e-01 8.65370557e-02
-4.86023486e-01 -7.46514380e-01 5.97747326e-01 -4.47261274e-01
1.54533374e+00 3.84062678e-02 2.79785633e-01 1.30217171e+00
-1.30249873e-01 9.27342296e-01 5.45259833e-01 -7.14297295e-01
5.66852272e-01 1.29916146e-01 1.08826602e+00 7.25163639e-01
-2.11492240e-01 -1.31736755e-01 -5.60156107e-01 -2.20290661e-01
4.21783596e-01 1.66932926e-01 2.16235295e-02 -4.76691097e-01
-1.02247071e+00 1.40165615e+00 3.35398436e-01 3.14405024e-01
9.18767182e-04 1.06666155e-01 8.54932249e-01 4.92385745e-01
5.13448000e-01 7.71959841e-01 -5.53149641e-01 -5.89943707e-01
-5.91161370e-01 -6.11935928e-02 1.16035128e+00 1.04588234e+00
9.05979097e-01 9.83128771e-02 -5.72385430e-01 1.09896207e+00
4.46008518e-02 -2.96669398e-02 6.25092447e-01 -8.49751115e-01
3.01673979e-01 6.07473135e-01 1.93910345e-01 -4.63695765e-01
-7.05979824e-01 5.41401878e-02 -3.41965944e-01 3.98459844e-02
3.62819642e-01 -4.60639864e-01 -1.12594235e+00 2.06767011e+00
1.53556600e-01 1.95470564e-02 3.82160157e-01 5.69861889e-01
9.22458291e-01 6.02650642e-01 5.33430040e-01 -2.13493556e-02
1.71811986e+00 -1.06850255e+00 -9.53157067e-01 -5.82057953e-01
1.29877782e+00 -3.65290105e-01 1.74789786e+00 -4.16139036e-01
-7.47632921e-01 -3.03515255e-01 -1.06885636e+00 -6.12270772e-01
-7.25049496e-01 -4.66163546e-01 9.36591208e-01 7.06412375e-01
-8.01907003e-01 4.51068133e-01 -8.52844059e-01 -6.30490422e-01
2.52056837e-01 2.12989047e-01 -2.53258497e-01 1.60266683e-01
-1.61638117e+00 1.12186432e+00 2.20488757e-01 -4.01127100e-01
-5.35036862e-01 -8.74525905e-01 -1.44066346e+00 4.75523651e-01
4.79971349e-01 -3.32423896e-01 1.76731658e+00 -3.17236334e-01
-1.37077510e+00 5.39918661e-01 -2.85991639e-01 -6.43161237e-01
-7.92488605e-02 -1.51274920e-01 -9.71126854e-02 -6.08749278e-02
9.95353982e-02 9.96730983e-01 4.28282440e-01 -6.89883530e-01
-6.02276802e-01 -1.00496970e-01 5.12506783e-01 3.33897740e-01
-8.00623715e-01 -1.01948686e-01 -4.43514287e-02 -2.35224932e-01
-1.88947186e-01 -6.82188988e-01 -4.09562767e-01 -1.38147268e-02
-2.88911462e-01 -7.09004521e-01 1.01853442e+00 -4.22693104e-01
9.84716833e-01 -2.17028308e+00 -2.75661230e-01 -3.57723266e-01
1.99629396e-01 2.39927411e-01 -3.17062408e-01 4.32202846e-01
7.07643032e-02 8.84998739e-02 -1.87160820e-01 -6.80558562e-01
4.81643468e-01 4.07516629e-01 -4.44952428e-01 6.10115100e-03
2.47287631e-01 1.21050215e+00 -9.30913806e-01 -3.94065201e-01
4.98916417e-01 2.67486364e-01 -6.47073150e-01 4.30609941e-01
-4.82152164e-01 -8.24694261e-02 -1.89091980e-01 3.48963588e-01
9.89120305e-02 -3.90192091e-01 1.66344106e-01 -1.85126007e-01
1.59400061e-01 9.37481880e-01 -7.61240363e-01 1.90624058e+00
-8.95220816e-01 7.86332428e-01 -6.22421801e-02 -8.97678673e-01
7.14770615e-01 6.61007464e-01 1.92681849e-01 -5.13268232e-01
1.67111352e-01 -2.05696061e-01 1.27979532e-01 -4.07174855e-01
7.28302777e-01 -5.95509946e-01 -6.83160305e-01 8.90420556e-01
6.13971472e-01 -5.64228818e-02 2.73434728e-01 4.79537636e-01
1.35506284e+00 -3.26596618e-01 4.35809702e-01 -1.22641101e-01
-5.37056513e-02 2.45572507e-01 3.19785208e-01 8.92186284e-01
-4.87942129e-01 3.91090542e-01 7.87505150e-01 -6.52723789e-01
-1.22924197e+00 -9.15712476e-01 -5.93112558e-02 1.79013634e+00
-3.44944596e-02 -4.11137164e-01 -4.50164884e-01 -7.36183107e-01
-2.24719465e-01 1.23105562e+00 -7.43039370e-01 -1.61882594e-01
-2.53155291e-01 -3.32827628e-01 4.73266989e-01 6.44290626e-01
1.43782467e-01 -1.22075105e+00 -8.06341350e-01 2.53173530e-01
-7.02967793e-02 -9.90361154e-01 -4.59214717e-01 9.05932605e-01
-4.50491756e-01 -8.12249899e-01 -3.43994856e-01 -7.67385185e-01
3.30076605e-01 2.49568626e-01 1.39310539e+00 -1.06658384e-01
-5.95657527e-01 6.39584720e-01 -4.61121082e-01 -4.46831197e-01
-3.79937321e-01 1.83503777e-01 1.11972570e-01 -4.95352000e-01
1.01978087e+00 -5.55783093e-01 -2.58475900e-01 2.81815398e-02
-7.33831823e-01 5.41207381e-02 3.22998583e-01 1.24843490e+00
-2.26997927e-01 -4.82612282e-01 6.49080396e-01 -1.18694913e+00
1.06238496e+00 -3.94558787e-01 -2.60224909e-01 3.51075232e-01
-3.41687828e-01 3.23067605e-01 4.30343360e-01 -3.69330645e-01
-1.05333865e+00 -8.72196928e-02 -3.35562319e-01 -3.12391013e-01
-5.26902139e-01 3.60287696e-01 -5.38297417e-03 4.11070257e-01
7.71373570e-01 -1.51160374e-01 1.76445797e-01 -4.03942734e-01
8.67827415e-01 8.95594954e-01 3.57971668e-01 -4.66512144e-01
4.53754157e-01 1.44018769e-01 -4.71229315e-01 -9.57135081e-01
-1.21073341e+00 -7.50914752e-01 -7.00350165e-01 3.81164163e-01
1.17683232e+00 -7.71136165e-01 -7.47313499e-01 -2.71059752e-01
-1.34734416e+00 -5.58500886e-01 -7.54612327e-01 2.13499516e-01
-5.90936661e-01 1.47635683e-01 -8.92523229e-01 -1.01979506e+00
-4.46669668e-01 -9.77277458e-01 9.80735719e-01 2.13288888e-01
-8.21671426e-01 -1.35812593e+00 2.00167879e-01 1.20504811e-01
5.79996586e-01 -4.61456150e-01 1.35675108e+00 -1.33942544e+00
-1.41141966e-01 -1.28978714e-01 -2.46639132e-01 3.50585170e-02
2.65563220e-01 -6.24834597e-01 -1.18494785e+00 -8.71951878e-02
6.66156188e-02 -8.17609668e-01 7.61990190e-01 2.40382478e-01
5.12842238e-01 -2.75066525e-01 -2.10751995e-01 1.27021730e-01
9.92130220e-01 1.80159032e-01 3.14901382e-01 4.64949906e-02
5.01051426e-01 9.07127917e-01 5.38626552e-01 4.06879306e-01
2.53494531e-01 5.59955478e-01 5.27364388e-02 -3.07000309e-01
7.00186044e-02 -2.78246969e-01 1.15544155e-01 6.34967864e-01
7.22452402e-01 -2.15757310e-01 -9.71307576e-01 7.89862096e-01
-1.78778112e+00 -9.07078326e-01 4.21415597e-01 1.82512820e+00
9.71013308e-01 2.11048558e-01 3.92543375e-02 -2.57263213e-01
5.17834306e-01 5.78023791e-01 -6.89201117e-01 -7.73108363e-01
4.08406913e-01 3.31374109e-01 1.61081702e-01 8.71905327e-01
-1.27704394e+00 1.29217541e+00 6.85754633e+00 5.61433196e-01
-7.67030001e-01 2.29374290e-01 5.75497925e-01 -1.93557739e-01
-4.18263316e-01 9.83605906e-02 -7.95813322e-01 1.79404151e-02
1.28394377e+00 -1.82110742e-01 7.05212355e-02 1.25399601e+00
-1.26562163e-01 1.17999263e-01 -1.40214050e+00 7.18767703e-01
-1.21522248e-01 -1.46552086e+00 -2.19219267e-01 -1.25783771e-01
2.84453452e-01 2.23234110e-02 -2.13929102e-01 9.54725862e-01
7.76260138e-01 -1.03803039e+00 2.54591461e-03 -2.00709686e-01
6.94142044e-01 -4.52097833e-01 6.69576526e-01 2.85484970e-01
-9.49127972e-01 -1.80629194e-01 -6.52976632e-01 -3.41969401e-01
4.73480761e-01 2.44298026e-01 -1.32942808e+00 -4.42154519e-02
1.27033681e-01 4.79903907e-01 -2.02626988e-01 4.62820977e-01
-1.38714895e-01 5.36104679e-01 -1.98724449e-01 -4.18754071e-01
5.17061830e-01 1.77343309e-01 2.31631517e-01 1.27593017e+00
7.12489709e-02 1.84099764e-01 3.10560137e-01 8.34020555e-01
-8.16469565e-02 7.28723686e-03 -9.25493062e-01 -4.03424054e-01
5.51573694e-01 1.42990208e+00 -6.42117500e-01 -5.31093657e-01
-7.56877005e-01 8.78416419e-01 5.11219919e-01 2.00011879e-01
-4.32047337e-01 -6.92765236e-01 1.16622663e+00 -3.71016800e-01
2.19431058e-01 -4.15243328e-01 -4.80072677e-01 -1.20144737e+00
-3.72248381e-01 -4.75513488e-01 4.65521008e-01 -4.52805340e-01
-1.37077343e+00 6.33790672e-01 -1.48096815e-01 -8.83534372e-01
-7.82845020e-01 -8.02276909e-01 -9.79300916e-01 8.93399417e-01
-1.26251388e+00 -1.03266621e+00 9.79409963e-02 3.18518519e-01
1.06710541e+00 -2.21715912e-01 1.43623984e+00 -1.37284063e-02
-3.56318116e-01 5.73088408e-01 -2.64004350e-01 4.21448886e-01
6.40245259e-01 -1.44361424e+00 9.00034010e-01 3.92427921e-01
2.94823825e-01 9.01016235e-01 9.61501896e-01 -4.93985862e-01
-1.26152396e+00 -7.12553680e-01 1.04822361e+00 -6.94306076e-01
8.58232141e-01 -9.65589225e-01 -8.71699750e-01 9.11589146e-01
3.82439524e-01 -1.89903975e-01 1.13090634e+00 9.64073896e-01
-5.72330952e-01 5.67396522e-01 -8.36692870e-01 8.14605892e-01
8.82689774e-01 -8.71830285e-01 -1.19457924e+00 4.82795000e-01
1.40495157e+00 -8.58631954e-02 -6.18830681e-01 2.33704358e-01
4.28103924e-01 -7.02746272e-01 9.16968346e-01 -1.06522787e+00
3.49481851e-01 2.11224452e-01 -2.04711482e-01 -1.23642802e+00
-8.95746574e-02 -4.26681995e-01 -1.83640331e-01 1.09970188e+00
7.48541117e-01 -2.04974189e-01 1.06197548e+00 1.30603850e+00
-3.16944510e-01 -5.83147764e-01 -1.03508568e+00 -6.03480279e-01
6.27304465e-02 -5.98558009e-01 3.56838465e-01 1.06230688e+00
8.55790854e-01 1.14937425e+00 -6.35731697e-01 -3.69249314e-01
2.25577980e-01 1.09336637e-02 7.31649399e-01 -1.26494706e+00
-2.89803922e-01 -2.06913978e-01 -3.13451439e-01 -1.06832230e+00
4.71397400e-01 -6.85939133e-01 3.30683947e-01 -1.61346734e+00
1.43633112e-01 -4.07541186e-01 -3.55061471e-01 1.00047696e+00
-3.30350339e-01 -2.80637536e-02 1.57629222e-01 9.87124294e-02
-8.43890190e-01 7.91519165e-01 6.38041139e-01 -2.04644650e-01
-3.98542136e-01 -3.26957971e-01 -7.91408479e-01 5.78967690e-01
5.50668538e-01 -3.09494734e-01 -8.29514980e-01 -2.11431697e-01
-3.15465331e-01 8.77620727e-02 2.70931646e-02 -7.71657228e-01
2.69269168e-01 -4.48045135e-02 1.97220091e-02 -4.22728539e-01
8.29844356e-01 -5.53025603e-01 -7.64013410e-01 2.90125817e-01
-1.03434932e+00 -3.66118044e-01 1.80435106e-01 5.45639992e-01
-6.35952130e-02 -4.98759866e-01 7.11694956e-01 -2.59881705e-01
-1.04855752e+00 1.40865788e-01 -4.00418550e-01 2.50373125e-01
7.65773118e-01 -1.88268833e-02 -4.78185683e-01 -7.30731606e-01
-8.14139247e-01 3.20562810e-01 4.00570124e-01 6.60106242e-01
3.99093419e-01 -1.08559394e+00 -8.73193815e-02 3.30757976e-01
3.31374735e-01 -2.86775082e-01 3.26788455e-01 3.98755223e-01
-2.33026184e-02 6.39950991e-01 -1.79609045e-01 -3.23155642e-01
-1.10994780e+00 4.96153146e-01 3.91595922e-02 -4.82630700e-01
-7.61092901e-01 8.75849366e-01 3.75199109e-01 -1.01218450e+00
6.07678652e-01 -8.09911713e-02 -2.01187372e-01 1.61135539e-01
8.14415753e-01 -1.43381834e-01 -9.56255123e-02 -1.36463776e-01
-1.60073444e-01 -1.20888010e-01 -4.35600519e-01 -4.62669760e-01
1.38402319e+00 -1.47395521e-01 2.60876656e-01 8.91627192e-01
1.22907698e+00 -6.24153554e-01 -1.25550532e+00 -4.52136874e-01
3.27812821e-01 -1.94154561e-01 -1.85739964e-01 -6.52496338e-01
-2.03773603e-01 1.51595438e+00 4.77723032e-01 5.33615410e-01
3.65819842e-01 1.75671428e-01 9.69119847e-01 9.40254271e-01
2.83270329e-01 -1.17045927e+00 1.14135765e-01 9.65833187e-01
3.19901049e-01 -1.40246928e+00 -3.66267264e-01 -1.20531522e-01
-8.56473625e-01 9.76764977e-01 7.56896794e-01 2.65675902e-01
5.26276886e-01 4.68008757e-01 3.87081653e-01 -2.23207653e-01
-1.19322479e+00 -3.75942558e-01 -8.97174254e-02 6.71313345e-01
8.75645518e-01 -1.27907977e-01 -8.70070532e-02 5.02816200e-01
-1.59262300e-01 -2.59915471e-01 3.67334098e-01 1.21496642e+00
-8.01603734e-01 -1.33609259e+00 1.51460573e-01 7.03275084e-01
-1.00208133e-01 -4.53702748e-01 -2.37013981e-01 5.85094869e-01
-5.19682765e-01 1.10716283e+00 4.52761829e-01 -3.97263408e-01
5.46108671e-02 8.42731595e-01 -1.47316486e-01 -1.33930016e+00
-3.39871973e-01 -1.75694883e-01 4.82747704e-01 -4.76479053e-01
-4.85529862e-02 -2.56197006e-01 -1.36669683e+00 -1.15589976e-01
-2.88009673e-01 4.68955994e-01 7.24478126e-01 1.20475221e+00
3.79074186e-01 5.59510827e-01 2.40288153e-01 -8.28259885e-01
-7.00946212e-01 -1.21776497e+00 -5.92689753e-01 5.75945497e-01
3.14640105e-01 -7.55582571e-01 -4.94463027e-01 -8.02167132e-02] | [12.380186080932617, 7.671482563018799] |
31565532-5872-450e-b5c0-2048ce88dca1 | a-deep-neural-network-algorithm-for-linear | 2301.10869 | null | https://arxiv.org/abs/2301.10869v2 | https://arxiv.org/pdf/2301.10869v2.pdf | A Deep Neural Network Algorithm for Linear-Quadratic Portfolio Optimization with MGARCH and Small Transaction Costs | We analyze a fixed-point algorithm for reinforcement learning (RL) of optimal portfolio mean-variance preferences in the setting of multivariate generalized autoregressive conditional-heteroskedasticity (MGARCH) with a small penalty on trading. A numerical solution is obtained using a neural network (NN) architecture within a recursive RL loop. A fixed-point theorem proves that NN approximation error has a big-oh bound that we can reduce by increasing the number of NN parameters. The functional form of the trading penalty has a parameter $\epsilon>0$ that controls the magnitude of transaction costs. When $\epsilon$ is small, we can implement an NN algorithm based on the expansion of the solution in powers of $\epsilon$. This expansion has a base term equal to a myopic solution with an explicit form, and a first-order correction term that we compute in the RL loop. Our expansion-based algorithm is stable, allows for fast computation, and outputs a solution that shows positive testing performance. | ['Farshad Khorrami', 'Prashanth Krishnamurthy', 'Hao Fu', 'Andrew Papanicolaou'] | 2023-01-25 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-4.20163631e-01 2.36346394e-01 -5.32271974e-02 -3.45359832e-01
-6.98145747e-01 -6.09700620e-01 5.67890257e-02 -2.08185390e-01
-6.99842095e-01 6.97008669e-01 -1.09515823e-01 -7.95400143e-01
-5.55289268e-01 -8.59836698e-01 -9.66531754e-01 -8.25490415e-01
-3.95514756e-01 5.18354714e-01 -6.01792559e-02 -2.32776478e-01
3.64644825e-01 2.71021158e-01 -1.02809489e+00 -1.44069031e-01
5.87703049e-01 1.78136325e+00 -1.70423061e-01 6.67427123e-01
1.34058714e-01 8.02977502e-01 -4.73059624e-01 -5.33268452e-01
9.81453419e-01 -4.62955654e-01 -3.91951859e-01 -1.84014037e-01
-1.02963448e-02 -3.74521792e-01 1.79013759e-01 1.17802608e+00
4.45806891e-01 2.53840387e-01 6.57898545e-01 -1.07141674e+00
-5.81677854e-01 7.92721689e-01 -5.19507766e-01 1.47623822e-01
-4.68193650e-01 9.35872793e-02 1.30853760e+00 -5.00846148e-01
1.70842335e-01 1.05059397e+00 9.73958313e-01 2.93504894e-01
-1.34764242e+00 -5.97082675e-01 -1.59142718e-01 -6.17190838e-01
-1.05987418e+00 -1.64946378e-03 4.25022483e-01 -4.28258181e-01
1.06284630e+00 2.28068769e-01 7.24160373e-01 3.12598705e-01
4.53303754e-01 2.50604242e-01 9.13521588e-01 -4.39355373e-01
7.39035368e-01 1.16444774e-01 -1.83123454e-01 5.46401620e-01
2.78760135e-01 4.40891534e-01 -6.22841008e-02 -4.57349062e-01
1.30054140e+00 1.65944234e-01 1.65508866e-01 -4.01928991e-01
-4.15400922e-01 1.36235821e+00 2.35850796e-01 4.48122621e-02
-5.55588961e-01 6.83579683e-01 -1.38549386e-02 9.02753890e-01
4.46735144e-01 7.65474796e-01 -7.72912741e-01 -1.57415166e-01
-7.96061277e-01 1.72835186e-01 9.75592554e-01 5.32377362e-01
6.87005401e-01 4.83254671e-01 4.78879772e-02 8.68925214e-01
3.16439793e-02 7.80366540e-01 5.42489588e-01 -1.69169521e+00
3.17483872e-01 2.80947983e-01 4.53194618e-01 -6.56602561e-01
-2.64227748e-01 -6.53108776e-01 -7.29410291e-01 5.11089027e-01
7.39823341e-01 -7.23355055e-01 -2.17578560e-01 1.83473742e+00
3.16113769e-03 -1.79891646e-01 -2.63213012e-02 5.62890112e-01
-5.50833523e-01 5.57345152e-01 -4.72433567e-01 -6.51470244e-01
9.30236161e-01 -6.16895497e-01 -2.39265546e-01 -1.25720631e-02
5.86053610e-01 -4.41539019e-01 1.03577352e+00 5.18375874e-01
-1.48072493e+00 8.85995571e-03 -8.29008222e-01 5.16253650e-01
1.57388106e-01 -2.37902418e-01 4.96696472e-01 8.84117842e-01
-1.24225676e+00 1.09367645e+00 -5.51319897e-01 6.14542305e-01
6.52936697e-02 6.46875799e-01 2.96540052e-01 7.81107485e-01
-9.59216595e-01 5.06442785e-01 2.10899726e-01 2.20509887e-01
-3.09973538e-01 -9.75843668e-01 -4.94526923e-01 4.50356752e-01
4.84987497e-01 -3.88892442e-01 1.55936337e+00 -1.39620209e+00
-1.75827134e+00 1.90347731e-01 3.75742972e-01 -8.59467149e-01
7.81523049e-01 -3.63209844e-02 -9.68004093e-02 1.10446535e-01
-5.12434281e-02 2.38009095e-01 8.68893623e-01 -6.24665320e-01
-5.70496142e-01 -2.55813092e-01 -1.76839605e-01 -1.19466558e-01
-4.23414081e-01 -1.34142518e-01 3.50959778e-01 -9.42145884e-01
-1.54246792e-01 -1.05532730e+00 -4.67290103e-01 -3.37539941e-01
4.37345356e-01 -7.50791878e-02 -2.97248401e-02 -6.67075634e-01
1.27693510e+00 -2.19360018e+00 -3.92736256e-01 7.87935853e-01
-2.50856102e-01 -1.81431919e-01 -3.33757326e-02 3.99506956e-01
-2.02403724e-01 1.17018051e-01 -2.92856276e-01 4.49175499e-02
3.29897463e-01 5.36462627e-02 -5.15871525e-01 2.36849323e-01
-1.42259166e-01 7.70213962e-01 -5.23214579e-01 1.62480801e-01
-3.75613511e-01 -1.45966485e-01 -1.01074874e+00 2.53673568e-02
-4.64242309e-01 -4.18367326e-01 -2.27368265e-01 1.23393051e-01
4.68924850e-01 -5.46993911e-01 3.91560584e-01 4.25705910e-01
-3.78550947e-01 -7.23268241e-02 -1.54334676e+00 6.04143977e-01
-5.35044372e-01 2.57232010e-01 1.69534907e-01 -9.21779335e-01
9.94638860e-01 7.49488994e-02 4.47385252e-01 -6.14920914e-01
2.59348810e-01 6.39883280e-01 -1.14620559e-01 5.82952052e-02
3.09601188e-01 -4.71853495e-01 -3.57348323e-02 9.16785300e-01
-2.57235527e-01 1.56930044e-01 2.09042311e-01 -1.63411275e-01
1.02306759e+00 -4.25870746e-01 3.10647823e-02 -5.81582248e-01
1.29156709e-01 -2.97679514e-01 5.99065363e-01 1.01735198e+00
2.98196882e-01 1.58049285e-01 1.27702594e+00 -5.16018867e-01
-1.14671981e+00 -7.04039335e-01 1.09653538e-02 1.01780939e+00
-5.79654515e-01 1.73803508e-01 -8.10258806e-01 -3.31872433e-01
2.88349599e-01 8.13172579e-01 -7.30912626e-01 -2.02484727e-02
-6.34934664e-01 -8.23905170e-01 2.34822631e-01 8.52919161e-01
6.81624770e-01 -1.21862745e+00 -9.10620928e-01 2.50335544e-01
4.54194665e-01 -3.03155750e-01 -7.41517007e-01 4.66037810e-01
-1.01716650e+00 -6.85840547e-01 -9.32004988e-01 -6.11257672e-01
5.19408762e-01 -3.31354290e-01 9.13572133e-01 -1.15820520e-01
3.28342378e-01 3.67237628e-01 3.01584333e-01 -3.97676855e-01
-2.91031688e-01 -3.43440741e-01 -1.04919836e-01 1.51213109e-01
9.11970157e-03 -5.86226821e-01 -8.11339021e-01 2.48655096e-01
-7.95959353e-01 -7.11815953e-01 5.67377746e-01 1.16753507e+00
6.03418052e-01 1.40809178e-01 5.87619662e-01 -6.37456477e-01
8.88461351e-01 -1.55344829e-01 -1.78048718e+00 2.71050841e-01
-1.07780945e+00 6.14041626e-01 7.65458822e-01 -5.30270517e-01
-8.84686470e-01 -2.02219501e-01 2.94989735e-01 -5.22931814e-01
6.59273744e-01 6.43644750e-01 3.59821349e-01 -6.58906549e-02
4.84010458e-01 -7.08891898e-02 2.49804586e-01 -5.73086619e-01
2.15465754e-01 3.20511013e-01 3.45388234e-01 -5.82019627e-01
2.32906193e-01 8.65654722e-02 2.77430266e-01 -3.97943616e-01
-4.47868794e-01 1.44958079e-01 6.69141710e-02 9.70535800e-02
3.69032770e-01 -5.78237414e-01 -1.55080640e+00 3.38213652e-01
-6.27555609e-01 -9.30286944e-01 -7.48251796e-01 6.42087579e-01
-9.78244960e-01 -1.33432522e-01 -8.45222056e-01 -1.22622848e+00
-3.69813412e-01 -6.45302951e-01 1.59494221e-01 -2.49087196e-02
5.32531999e-02 -9.63630319e-01 2.94824153e-01 -2.00195163e-01
4.68791127e-01 -9.14329141e-02 1.16868222e+00 -7.82425880e-01
-4.01904106e-01 -2.34180838e-01 4.52312315e-03 7.17992663e-01
-4.22951192e-01 -1.34012610e-01 -2.40723997e-01 -2.25747585e-01
4.49704826e-01 -3.66743952e-02 6.06737196e-01 9.33555126e-01
8.31393778e-01 -1.07060099e+00 4.55643296e-01 5.67131162e-01
1.57815135e+00 5.77509701e-01 2.32867807e-01 4.29139912e-01
-6.53814822e-02 4.43076521e-01 3.26724559e-01 7.99219489e-01
-9.29842815e-02 3.96767221e-02 1.19094901e-01 3.49664927e-01
8.16341221e-01 -9.13222879e-02 4.82941419e-01 5.20918608e-01
-5.26024438e-02 2.63965964e-01 -9.03417051e-01 3.16222101e-01
-1.92175174e+00 -1.04549408e+00 2.72224486e-01 2.59390092e+00
8.67036223e-01 5.02913773e-01 5.49522340e-01 -2.84973025e-01
4.37911958e-01 -3.08735073e-01 -7.03802586e-01 -9.73385811e-01
2.19003558e-02 5.01523435e-01 1.15675592e+00 6.75354600e-01
-5.69079578e-01 3.18149328e-01 7.15637255e+00 7.97711790e-01
-1.03981531e+00 -1.68043017e-01 9.39430594e-01 -2.78437614e-01
-5.65050781e-01 -6.17792718e-02 -7.17847168e-01 6.54185772e-01
1.26599228e+00 -3.15270394e-01 6.60723567e-01 1.21815598e+00
9.82146338e-02 -1.34765927e-03 -8.33106637e-01 6.59531891e-01
-3.69848311e-01 -1.51931214e+00 -3.96338284e-01 4.94743198e-01
9.85565901e-01 -2.36903235e-01 5.17040312e-01 3.27460557e-01
7.14007854e-01 -6.93738639e-01 6.71120763e-01 4.51692253e-01
3.31642479e-01 -1.49985790e+00 7.22093821e-01 3.86094600e-01
-7.87964523e-01 -8.42452765e-01 -4.51140016e-01 -2.41808355e-01
-1.59417272e-01 4.67700452e-01 -5.65636754e-01 -1.02043703e-01
6.24904752e-01 -7.73344859e-02 -1.05509609e-01 7.99142599e-01
1.17648453e-01 4.76407051e-01 -5.31581759e-01 -2.41120934e-01
4.96763527e-01 -8.95326912e-01 6.17687479e-02 7.36318350e-01
7.26833105e-01 3.06006014e-01 -2.71333426e-01 9.88513350e-01
-1.14596039e-01 1.56012505e-01 -2.39847660e-01 -1.47634909e-01
3.00489247e-01 8.25718701e-01 -6.27067983e-01 -2.04043403e-01
-3.53955626e-01 3.86914253e-01 2.19116107e-01 4.57891852e-01
-4.35074598e-01 -4.17045027e-01 2.51893789e-01 4.74798866e-02
1.00502717e+00 -1.12580359e-01 -4.03675467e-01 -6.93439424e-01
3.22299749e-01 -8.98995161e-01 7.91809976e-01 -3.46186608e-01
-1.19059086e+00 2.87427813e-01 -1.69693723e-01 -9.42005813e-01
-9.78697717e-01 -5.52971303e-01 -4.92997438e-01 7.66793132e-01
-8.82732451e-01 -2.55544871e-01 7.40646780e-01 5.26651800e-01
-6.21673502e-02 -4.46848035e-01 6.48422778e-01 -9.65638831e-02
-4.18544084e-01 8.24659109e-01 7.32865572e-01 7.19242096e-02
1.88420758e-01 -1.29079723e+00 -5.36406711e-02 6.21607900e-01
-1.84722796e-01 6.80975556e-01 6.64180636e-01 -6.43019974e-01
-1.05495715e+00 -7.32671976e-01 8.72116625e-01 -2.13095755e-03
1.05968118e+00 1.39587810e-02 -5.63080132e-01 1.02043843e+00
2.38239259e-01 -1.46074668e-01 5.25613785e-01 1.18494138e-01
-2.66108900e-01 -5.57971895e-01 -1.10061455e+00 5.68105698e-01
3.03972155e-01 -1.60044283e-01 -2.10412696e-01 1.17122836e-01
7.25866497e-01 -3.46600413e-02 -1.08899856e+00 8.62629190e-02
7.78344035e-01 -9.88485217e-01 5.75320065e-01 -6.05986893e-01
2.48078391e-01 2.61138260e-01 -6.22504950e-02 -1.07685804e+00
-1.38043866e-01 -1.36349821e+00 -1.43157929e-01 6.57423675e-01
6.27295911e-01 -1.13182247e+00 8.34774017e-01 1.02429152e+00
2.33041033e-01 -8.41956556e-01 -9.40054655e-01 -1.16673672e+00
4.68336791e-01 -7.30524510e-02 3.84202272e-01 4.80239660e-01
3.28540578e-02 -7.38258436e-02 -4.57470059e-01 -1.42009124e-01
7.08879411e-01 3.13038290e-01 3.10522437e-01 -1.06596386e+00
-1.04362917e+00 -7.76253223e-01 2.15577021e-01 -1.09631348e+00
1.48454765e-02 -3.70987386e-01 -1.86749116e-01 -4.94546324e-01
-1.16786659e-01 -4.68363583e-01 -4.71899807e-01 2.90051639e-01
4.72765386e-01 -1.67695642e-01 2.00175196e-01 -8.41331556e-02
-1.79464161e-01 2.65460908e-01 6.80297136e-01 1.74471810e-01
-5.41500032e-01 3.43264014e-01 -6.69601202e-01 9.57340717e-01
9.26436424e-01 -5.78715205e-01 -2.41432086e-01 -1.45287290e-01
7.37490296e-01 6.06070280e-01 2.68460751e-01 -3.78350884e-01
1.85953856e-01 -3.76776367e-01 3.11240733e-01 -4.83305007e-01
1.59285828e-01 -4.88444000e-01 3.07673126e-01 7.00479925e-01
-7.59845018e-01 6.66900694e-01 -1.88633829e-01 3.08355331e-01
-1.34893376e-02 -6.35626078e-01 8.06353509e-01 -2.13265866e-01
1.43413752e-01 1.80630181e-02 -3.34189296e-01 1.76848769e-01
5.77296853e-01 2.48353422e-01 1.23937219e-01 -7.89496958e-01
-7.28849709e-01 3.28076154e-01 2.23939016e-01 -1.74858883e-01
3.39267671e-01 -1.38518727e+00 -3.18966240e-01 3.41050148e-01
-7.91878819e-01 -4.70961541e-01 7.79615864e-02 7.53386259e-01
-4.27574128e-01 3.42854828e-01 -9.35684964e-02 -1.99051872e-02
-7.47322023e-01 5.02966881e-01 7.05201209e-01 -5.04022002e-01
-1.42215058e-01 8.50737870e-01 -3.93036939e-02 -1.47638291e-01
2.99153090e-01 -3.75142723e-01 2.65510231e-01 6.47271723e-02
5.04218578e-01 6.70951903e-01 -5.01681156e-02 1.83228210e-01
8.71916488e-02 4.15327460e-01 3.07054520e-01 -7.74806321e-01
1.58105445e+00 6.85163513e-02 -2.33786061e-01 5.11254072e-01
1.06644619e+00 -8.45072344e-02 -1.44373143e+00 6.09560357e-03
-7.91115779e-03 -2.06464350e-01 -2.20740393e-01 -6.00516379e-01
-1.22186100e+00 5.25164604e-01 4.66866106e-01 6.19731069e-01
9.50826585e-01 -4.59514022e-01 6.04910612e-01 8.14734995e-01
2.06228912e-01 -1.66948593e+00 2.00486690e-01 6.79959655e-01
8.16567302e-01 -6.64723098e-01 -2.87097037e-01 3.83150339e-01
-7.09035516e-01 1.16676784e+00 1.99332386e-01 -6.65963292e-01
1.13829839e+00 3.39320928e-01 -2.80820895e-02 9.11203101e-02
-8.01782489e-01 4.14729744e-01 1.33311361e-01 -1.63957819e-01
-3.18460464e-02 9.61094201e-02 -5.18623233e-01 8.41330707e-01
-5.44787765e-01 -1.43039376e-01 3.88588548e-01 6.78081989e-01
-6.25084698e-01 -1.03762126e+00 -2.04359725e-01 4.77610409e-01
-6.36583567e-01 -2.59514660e-01 -7.07359016e-02 8.33648801e-01
-3.08332622e-01 6.98123813e-01 4.56794739e-01 1.57744065e-01
2.33524621e-01 1.39826164e-01 3.49446237e-01 3.61051746e-02
-7.16843009e-01 6.18393838e-01 -1.93497255e-01 -4.85131443e-01
5.01044840e-03 -7.21779227e-01 -1.14581370e+00 -5.82608640e-01
-2.23924339e-01 5.39684832e-01 3.02508742e-01 7.35401332e-01
2.70474195e-01 -1.11257136e-01 1.12115204e+00 -2.35213354e-01
-1.58651221e+00 -5.72252214e-01 -1.08632624e+00 1.35438135e-02
2.44096458e-01 -1.63595498e-01 -7.22545803e-01 -3.78851503e-01] | [4.93118953704834, 3.933338165283203] |
978719d4-bef2-42ab-abe6-73b551c455de | a-fast-keypoint-based-hybrid-method-for-copy | 1612.03989 | null | http://arxiv.org/abs/1612.03989v1 | http://arxiv.org/pdf/1612.03989v1.pdf | A Fast Keypoint Based Hybrid Method for Copy Move Forgery Detection | Copy move forgery detection in digital images has become a very popular
research topic in the area of image forensics. Due to the availability of
sophisticated image editing tools and ever increasing hardware capabilities, it
has become an easy task to manipulate the digital images. Passive forgery
detection techniques are more relevant as they can be applied without the prior
information about the image in question. Block based techniques are used to
detect copy move forgery, but have limitations of large time complexity and
sensitivity against affine operations like rotation and scaling. Keypoint based
approaches are used to detect forgery in large images where the possibility of
significant post processing operations like rotation and scaling is more. A
hybrid approach is proposed using different methods for keypoint detection and
description. Speeded Up Robust Features (SURF) are used to detect the keypoints
in the image and Binary Robust Invariant Scalable Keypoints (BRISK) features
are used to describe features at these keypoints. The proposed method has
performed better than the existing forgery detection method using SURF
significantly in terms of detection speed and is invariant to post processing
operations like rotation and scaling. The proposed method is also invariant to
other commonly applied post processing operations like adding Gaussian noise
and JPEG compression | ['Sunil Kumar', 'J. V. Desai', 'Shaktidev Mukherjee'] | 2016-12-11 | null | null | null | null | ['image-forensics'] | ['computer-vision'] | [ 3.62967104e-01 -6.96416318e-01 2.02634230e-01 1.30617663e-01
-6.41387999e-01 -7.17146695e-01 6.25715852e-01 7.02526629e-01
-6.36259675e-01 3.00366849e-01 -1.20458841e-01 -1.35981381e-01
-2.10834444e-01 -7.03378439e-01 -4.12802756e-01 -7.44965374e-01
-1.55511303e-02 -1.69984266e-01 8.64237070e-01 -2.94506401e-01
1.11171722e+00 1.00155437e+00 -1.49101067e+00 9.45312064e-03
2.26714581e-01 8.34800184e-01 3.88844371e-01 1.02073312e+00
1.33070901e-01 6.01446033e-01 -7.01193392e-01 -2.19184875e-01
6.15558088e-01 -2.13521257e-01 -4.67996925e-01 2.46403188e-01
2.19964057e-01 -4.88358080e-01 -5.21834016e-01 1.23112535e+00
4.04229671e-01 3.55794966e-01 3.92800838e-01 -1.12813234e+00
-2.36860693e-01 -1.44186124e-01 -1.08389008e+00 7.96520352e-01
5.31297922e-01 -1.10383749e-01 1.92389905e-01 -9.28024828e-01
6.74826622e-01 1.10706127e+00 6.78669930e-01 -3.60045701e-01
-6.77904487e-01 -7.19735444e-01 -7.44968772e-01 8.14527631e-01
-1.56540108e+00 -4.07204211e-01 7.27408767e-01 -2.71185264e-02
9.28432226e-01 4.59171027e-01 -1.61150098e-03 1.74253121e-01
7.17411339e-01 2.55236000e-01 1.14024472e+00 -9.02041197e-01
2.99274363e-02 1.70486987e-01 1.62120208e-01 6.36054873e-01
5.53244293e-01 9.09708291e-02 -3.55303824e-01 -3.25614125e-01
8.05563509e-01 4.19598132e-01 -3.43681961e-01 -1.11584686e-01
-1.12935328e+00 7.63267398e-01 1.72297001e-01 3.97719353e-01
-5.54406762e-01 1.86381787e-01 5.85030675e-01 3.91525954e-01
-2.01327026e-01 2.03259856e-01 5.06004319e-02 -2.21210927e-01
-1.18935347e+00 3.99305135e-01 2.15629786e-01 6.28303349e-01
6.48832738e-01 -1.06577482e-02 2.93546200e-01 4.99274701e-01
1.65952072e-01 2.99831122e-01 7.78733969e-01 -7.71261752e-01
5.36596060e-01 6.92960560e-01 2.98721910e-01 -1.63332641e+00
-2.57668626e-02 1.51242122e-01 -4.84683603e-01 6.32665753e-01
2.64652610e-01 4.62811917e-01 -9.34274614e-01 6.83394611e-01
5.13731241e-01 -6.15741536e-02 -5.97102493e-02 4.66893196e-01
2.81842053e-01 7.68777788e-01 -2.61341244e-01 -1.75620869e-01
1.63281071e+00 -3.60040098e-01 -7.14980960e-01 1.40997190e-02
3.74345303e-01 -1.44741750e+00 5.94870985e-01 6.18364394e-01
-8.27937186e-01 -4.63354498e-01 -1.38072777e+00 6.41235039e-02
-7.74302423e-01 -6.56453669e-02 2.83583552e-01 1.02369642e+00
-7.85130084e-01 6.00806236e-01 -6.62510037e-01 -4.75989521e-01
2.22349703e-01 6.09572649e-01 -7.59660065e-01 -3.77768695e-01
-6.68149114e-01 8.74230802e-01 5.20689189e-01 -7.93294758e-02
-3.45151961e-01 -1.08131483e-01 -6.42110825e-01 -6.58399835e-02
3.18553776e-01 1.74940571e-01 6.37578368e-01 -7.72813678e-01
-9.89617527e-01 8.52577746e-01 1.78145636e-02 -3.70213121e-01
6.43915117e-01 -2.26872079e-02 -5.75179398e-01 8.02178979e-01
-1.51556749e-02 5.82278408e-02 1.42481840e+00 -8.23794663e-01
-5.44607103e-01 -5.21591067e-01 -3.55003804e-01 3.84406820e-02
-1.48666471e-01 6.26898944e-01 -1.79098904e-01 -1.02286589e+00
4.12581235e-01 -6.84266508e-01 2.14462593e-01 2.08067358e-01
-1.15263984e-01 1.25882059e-01 1.74602389e+00 -1.00053740e+00
9.96586859e-01 -2.15152383e+00 -4.99311119e-01 3.40353578e-01
-8.60195011e-02 7.08182037e-01 3.31465244e-01 9.81673539e-01
-2.19842330e-01 -1.12400837e-01 -1.16627783e-01 8.29322636e-02
-4.60007697e-01 7.99555238e-03 -2.00440854e-01 1.06902897e+00
-1.10916406e-01 2.06280544e-01 -4.90060925e-01 -6.41906857e-01
6.62883997e-01 4.93358940e-01 -6.23628758e-02 -3.99662033e-02
5.41798592e-01 -1.38231725e-01 -2.68675417e-01 7.91135848e-01
1.01018834e+00 2.93557048e-01 -4.65545624e-01 -3.71771991e-01
-2.11016491e-01 -3.71343046e-01 -1.75320005e+00 9.76353586e-01
-1.79929987e-01 8.08716416e-01 2.34971911e-01 -1.08694494e+00
1.06789517e+00 3.50914448e-01 8.95623192e-02 -3.09488684e-01
3.03600699e-01 2.04485029e-01 -2.26759851e-01 -7.12173343e-01
9.63825047e-01 2.18231184e-03 2.27753147e-01 5.92635095e-01
-2.45436639e-01 7.66007090e-03 8.13567191e-02 2.56113142e-01
1.27051389e+00 -2.42875796e-02 7.36549377e-01 -3.45043018e-02
8.14812541e-01 -2.38057654e-02 2.35998183e-01 4.81999904e-01
-3.50017041e-01 5.27129412e-01 -5.04374169e-02 -3.13125670e-01
-1.09199750e+00 -6.34029210e-01 -5.94760254e-02 4.96720791e-01
4.26957726e-01 -3.11298490e-01 -4.37439084e-01 -4.40592229e-01
5.70061393e-02 1.91005930e-01 -4.32361007e-01 1.03909178e-02
-8.25280607e-01 -4.69236374e-01 5.38951099e-01 2.55962908e-01
8.33811581e-01 -8.85056376e-01 -1.16069138e+00 1.87528074e-01
2.82054991e-01 -1.07998729e+00 -3.10708106e-01 -9.13570896e-02
-1.17682517e+00 -1.27878726e+00 -7.21302569e-01 -6.44900918e-01
8.49068403e-01 8.62746239e-01 2.10598245e-01 5.77336729e-01
-1.01969349e+00 3.93328339e-01 -6.85119510e-01 -2.39385709e-01
-5.43692887e-01 -7.85370648e-01 -1.52917996e-01 4.31448631e-02
1.19918965e-01 -2.96873033e-01 -8.06239784e-01 4.65100318e-01
-1.35449648e+00 -6.65451825e-01 7.20781684e-01 6.30183160e-01
3.10360342e-01 7.79008329e-01 1.90681204e-01 -5.58773577e-01
5.69231272e-01 -1.76797181e-01 -5.14578044e-01 1.82251886e-01
-3.98622632e-01 -1.03594661e-02 5.66015899e-01 -2.45433971e-01
-8.00447345e-01 -7.37449825e-02 2.03419104e-01 -2.88414448e-01
-1.93685964e-01 -1.69467591e-02 5.79033121e-02 -8.49937975e-01
5.29110491e-01 5.07481039e-01 1.41106635e-01 -5.65535486e-01
4.36502174e-02 9.45082724e-01 7.23795176e-01 -2.05299221e-02
1.17271411e+00 7.25889921e-01 3.55590314e-01 -1.31124914e+00
4.16253865e-01 -9.59077895e-01 -6.25927746e-01 -2.31461361e-01
5.72873712e-01 -3.98649991e-01 -4.94499266e-01 8.14599633e-01
-9.80495572e-01 6.48879170e-01 3.71086746e-01 3.07637721e-01
-1.60514385e-01 1.35448837e+00 -4.06297117e-01 -1.01464343e+00
-5.48073471e-01 -1.19597173e+00 7.32516706e-01 2.82309592e-01
1.71520915e-02 -7.44558513e-01 -4.70908672e-01 3.77708107e-01
5.59989691e-01 6.25051320e-01 8.54865015e-01 -4.61696297e-01
-6.34111643e-01 -1.10351837e+00 -3.71665597e-01 3.66380632e-01
4.80502188e-01 4.04377952e-02 -4.61239547e-01 -4.77998912e-01
3.60375464e-01 1.88450590e-01 4.94579941e-01 -2.71839723e-02
7.73577332e-01 -2.24740595e-01 -5.10278702e-01 3.54655206e-01
1.82594049e+00 3.70801181e-01 9.99606133e-01 8.68714571e-01
3.24267447e-01 4.02132303e-01 8.61646056e-01 4.36454386e-01
-2.31260523e-01 6.50491476e-01 3.77199799e-01 1.80154860e-01
-1.30010664e-01 1.43556267e-01 2.90485322e-01 2.82433093e-01
-2.95199677e-02 -1.07487671e-01 -6.43385708e-01 4.92223412e-01
-1.34037089e+00 -1.08406878e+00 -3.91802460e-01 2.51099849e+00
1.55796021e-01 3.77519876e-02 4.08805348e-02 1.07226145e+00
1.13257599e+00 -2.11516209e-02 -1.13054350e-01 -5.54467201e-01
4.69199307e-02 4.72467393e-01 1.24503887e+00 2.30988041e-01
-9.98918891e-01 6.12855792e-01 5.31096601e+00 9.14262533e-01
-1.22379220e+00 7.59513453e-02 9.91553962e-02 3.60676408e-01
5.34704983e-01 2.82883555e-01 -5.12258112e-01 5.74169397e-01
4.75097448e-01 2.10954607e-01 2.79263090e-02 6.06215298e-01
3.29982877e-01 -9.48073030e-01 -3.83354187e-01 1.24976325e+00
3.06159645e-01 -1.13227713e+00 -2.82689296e-02 2.14872971e-01
2.07968995e-01 -5.37573576e-01 -1.11514926e-02 -4.94988024e-01
-4.06693548e-01 -6.92440271e-01 3.98864686e-01 1.16649397e-01
4.10471320e-01 -1.01897311e+00 8.58952820e-01 2.09738225e-01
-1.10195029e+00 -8.12126920e-02 -6.46943510e-01 4.52660732e-02
2.29032069e-01 7.52914771e-02 -1.00490606e+00 3.98640692e-01
6.58564150e-01 6.85353950e-02 -6.28623962e-01 1.35809970e+00
2.05552563e-01 4.08810228e-01 -4.67309833e-01 2.50271887e-01
2.87784964e-01 3.36755924e-02 6.41077757e-01 1.19246399e+00
6.80444717e-01 3.78315113e-02 -3.30005974e-01 1.81833476e-01
2.66929358e-01 5.54452896e-01 -4.65677589e-01 4.15175445e-02
4.71952319e-01 1.15938163e+00 -1.36835468e+00 -1.98208585e-01
-3.04691166e-01 1.49323153e+00 -4.13194597e-01 -1.08982392e-01
-2.26644352e-01 -1.20758963e+00 1.82832420e-01 6.84569359e-01
7.28530526e-01 -6.20060325e-01 3.47753137e-01 -6.80311859e-01
4.85857837e-02 -8.99679542e-01 5.86675763e-01 -6.32197440e-01
-5.83442628e-01 2.14496091e-01 2.44138539e-01 -1.22679090e+00
1.72953506e-03 -6.97076201e-01 -8.74315619e-01 7.48548627e-01
-1.22548985e+00 -1.13270795e+00 -3.38634431e-01 8.48801792e-01
7.14215338e-01 -2.45748237e-01 5.14434993e-01 1.30870879e-01
-2.16507480e-01 4.89876390e-01 2.85202056e-01 2.28425026e-01
8.33517194e-01 -8.07115853e-01 2.38523901e-01 1.41379321e+00
-1.89657547e-02 8.36378038e-01 9.25835192e-01 -9.35603082e-01
-1.83363044e+00 -4.46272224e-01 5.57527423e-01 -8.56785942e-03
3.62530142e-01 -3.31911594e-02 -9.73505735e-01 2.88494170e-01
1.72163472e-01 3.00455123e-01 3.02865654e-01 -9.74150538e-01
-1.76968604e-01 -7.15818554e-02 -1.71397531e+00 -8.19342434e-02
7.20044076e-02 -2.81800151e-01 -6.63766861e-01 1.85234487e-01
-1.88101575e-01 -1.90198272e-01 -5.59702814e-01 -6.38852939e-02
4.60357606e-01 -1.18435502e+00 1.12084222e+00 1.76916897e-01
-2.75691986e-01 -6.49229705e-01 -1.85790643e-01 -5.27445257e-01
1.68344472e-02 -9.72621739e-01 1.89396635e-01 1.21363890e+00
-3.78106952e-01 -7.41093457e-01 7.63976753e-01 1.54585660e-01
3.71992171e-01 -1.92418665e-01 -1.09356391e+00 -8.25440466e-01
-5.88683665e-01 -2.07269397e-02 1.90356761e-01 8.25562775e-01
-6.31417558e-02 -4.55422729e-01 -4.86122012e-01 3.07641059e-01
8.33817065e-01 -1.67999461e-01 7.62147248e-01 -8.56166959e-01
-3.49208266e-01 -3.50995548e-02 -1.38991380e+00 -3.68726254e-01
-6.87959731e-01 -3.56359601e-01 -3.96846086e-01 -1.09825540e+00
1.54607855e-02 -2.79445082e-01 -1.39169684e-02 9.72274169e-02
-2.06680939e-01 6.46041572e-01 2.57521689e-01 4.55305278e-01
1.46087809e-02 -1.64080143e-01 7.02163815e-01 2.45940313e-01
1.56410038e-02 -8.91641341e-03 -1.30610719e-01 6.31017447e-01
6.58478796e-01 -6.90358162e-01 -1.84079617e-01 1.05879374e-01
1.33605614e-01 7.07382038e-02 3.94783467e-01 -1.21706128e+00
3.11077565e-01 3.95999923e-02 6.12730145e-01 -8.47052276e-01
3.27985883e-01 -8.51505995e-01 1.11153081e-01 6.79325879e-01
3.19120109e-01 8.47916663e-01 1.68009922e-01 8.26286972e-01
-2.47370526e-01 -7.95588076e-01 1.03639412e+00 -4.07388389e-01
-9.61552262e-01 -2.49634162e-01 -5.21996677e-01 -7.14003921e-01
1.50779259e+00 -1.09286273e+00 -7.21990913e-02 -5.16669035e-01
-2.08242953e-01 -5.92202008e-01 7.23964453e-01 2.87493676e-01
1.00910342e+00 -8.23443055e-01 -4.39040363e-01 2.16816708e-01
2.57627051e-02 -5.86305857e-01 1.49795741e-01 6.65223777e-01
-1.37541759e+00 1.72073469e-01 -6.07367694e-01 -2.50103563e-01
-1.98229444e+00 7.21245050e-01 -1.72949851e-01 1.28736645e-01
-9.37828422e-01 5.63741267e-01 -5.99045515e-01 6.34458780e-01
-9.82376561e-02 -1.52889341e-01 -2.12789676e-03 -1.56490415e-01
9.19213116e-01 9.19222951e-01 2.84158975e-01 -1.15211344e+00
-4.59736019e-01 9.24184859e-01 -4.47434098e-01 -5.19135147e-02
1.22503424e+00 -3.03831607e-01 -1.66622370e-01 -2.38946572e-01
1.39509618e+00 3.37745845e-01 -6.37722135e-01 -1.73668284e-02
2.17281401e-01 -1.12063909e+00 4.73591119e-01 -3.53804946e-01
-7.66151547e-01 8.07109654e-01 1.06207657e+00 1.23316243e-01
1.12068927e+00 -4.20463234e-01 9.14090455e-01 1.54885486e-01
3.53572041e-01 -1.24705720e+00 6.48877919e-02 -1.95064008e-01
6.73542261e-01 -1.14825869e+00 7.13134885e-01 -4.34705347e-01
-2.61708170e-01 1.56708360e+00 -6.78170100e-02 -6.26889169e-01
4.83835995e-01 2.35085815e-01 -8.18272587e-03 -2.48016179e-01
1.80686638e-02 1.94156513e-01 -9.71607864e-02 6.57681525e-01
1.93317756e-01 -3.34442079e-01 -6.66797578e-01 -3.27577859e-01
9.13550183e-02 -2.23571137e-01 8.44732285e-01 1.62694371e+00
-7.15679288e-01 -1.41061652e+00 -1.33258879e+00 4.07237709e-01
-1.09867632e+00 2.94562548e-01 -1.77327543e-01 1.08898413e+00
-2.34844163e-02 9.84694660e-01 -2.88310111e-01 -1.42228201e-01
3.84384729e-02 -8.37850571e-02 6.75884664e-01 -1.67585716e-01
-6.99574590e-01 -2.55048759e-02 -4.53249097e-01 -4.61205930e-01
-2.89229274e-01 -8.71566236e-01 -1.08430135e+00 -3.16045135e-01
-7.06503689e-01 -1.83123548e-03 1.24681747e+00 8.25863600e-01
1.43681109e-01 -1.64722979e-01 5.12019515e-01 -7.85155952e-01
-4.92182076e-01 -7.53760040e-01 -6.14854991e-01 5.68865061e-01
4.08273607e-01 -7.82471120e-01 -3.54529500e-01 3.23228061e-01] | [12.363049507141113, 0.9572150111198425] |
a3e0d500-859b-4d15-b591-fcc2721c3b9b | machine-learning-for-predicting-epileptic | 2002.01925 | null | https://arxiv.org/abs/2002.01925v1 | https://arxiv.org/pdf/2002.01925v1.pdf | Machine Learning for Predicting Epileptic Seizures Using EEG Signals: A Review | With the advancement in artificial intelligence (AI) and machine learning (ML) techniques, researchers are striving towards employing these techniques for advancing clinical practice. One of the key objectives in healthcare is the early detection and prediction of disease to timely provide preventive interventions. This is especially the case for epilepsy, which is characterized by recurrent and unpredictable seizures. Patients can be relieved from the adverse consequences of epileptic seizures if it could somehow be predicted in advance. Despite decades of research, seizure prediction remains an unsolved problem. This is likely to remain at least partly because of the inadequate amount of data to resolve the problem. There have been exciting new developments in ML-based algorithms that have the potential to deliver a paradigm shift in the early and accurate prediction of epileptic seizures. Here we provide a comprehensive review of state-of-the-art ML techniques in early prediction of seizures using EEG signals. We will identify the gaps, challenges, and pitfalls in the current research and recommend future directions. | ["Terence O'Brien", 'Shobi Sivathamboo', 'Adnan Qayyum', 'Junaid Qadir', 'Levin Kuhlmann', 'Khansa Rasheed', 'Patrick Kwan', 'Adeel Razi'] | 2020-02-04 | null | null | null | null | ['seizure-prediction'] | ['medical'] | [ 2.52419680e-01 -7.56605789e-02 6.94610775e-02 -3.97602946e-01
-8.89524519e-01 -5.91907278e-02 1.00425355e-01 4.69808847e-01
-3.47167522e-01 7.51836181e-01 1.70621485e-01 -2.71077454e-01
-4.79364187e-01 -2.15205252e-01 -8.61664563e-02 -7.86541224e-01
-7.70881832e-01 3.96173596e-01 -1.28234074e-01 1.95059888e-02
3.88576418e-01 6.84633553e-01 -1.40343261e+00 5.74954152e-01
1.13735628e+00 1.05313802e+00 4.49011803e-01 4.76104110e-01
-3.32147777e-02 8.87593329e-01 -6.03605032e-01 1.55204311e-02
-1.86008051e-01 -7.07785726e-01 -4.00149196e-01 -2.09455639e-01
-6.36131167e-01 -2.39911061e-02 -2.07154602e-01 8.22918892e-01
8.34273160e-01 -4.43293899e-01 5.57759702e-01 -1.08919251e+00
1.76858276e-01 2.55934030e-01 -3.15868318e-01 4.99377847e-01
4.44759071e-01 -2.04386771e-01 2.95630276e-01 -7.00262070e-01
9.19715539e-02 4.51612324e-01 4.83896673e-01 7.82242537e-01
-1.03416276e+00 -9.38884318e-01 4.76795137e-02 7.48529613e-01
-1.38338637e+00 -4.16864365e-01 5.46435237e-01 -6.33525670e-01
1.01352262e+00 3.40612084e-01 1.07010972e+00 8.92689586e-01
7.40871251e-01 8.04192543e-01 1.20787323e+00 -4.12672222e-01
3.41801405e-01 -2.84809317e-03 2.37834100e-02 1.80790827e-01
6.36451021e-02 2.57266015e-01 -8.77310693e-01 -3.41134906e-01
3.25083971e-01 2.45861262e-01 -5.81738591e-01 7.20291883e-02
-1.16391063e+00 8.05706680e-01 2.04658151e-01 6.32767558e-01
-7.17006803e-01 -4.74699318e-01 2.38340259e-01 2.44445726e-01
5.51378548e-01 6.62951112e-01 -6.14642262e-01 -5.45704067e-01
-1.34868264e+00 1.25495836e-01 7.55819201e-01 1.95026264e-01
1.84864596e-01 1.51808351e-01 2.15878397e-01 6.06450677e-01
1.59604326e-02 1.19406983e-01 8.02047491e-01 -2.75209934e-01
-2.99817473e-02 7.14549720e-01 -1.34306699e-01 -7.25805402e-01
-6.49638414e-01 -7.52020061e-01 -9.50926065e-01 1.94930583e-01
2.09559482e-02 -2.05674976e-01 -6.35332346e-01 1.18928921e+00
-2.43451640e-01 5.13003647e-01 6.09294362e-02 4.21500981e-01
5.39892554e-01 2.80496389e-01 1.42432660e-01 -7.62775123e-01
1.01733923e+00 -3.18930060e-01 -8.74150574e-01 -5.30128539e-01
7.50079453e-01 -5.88932276e-01 1.87519252e-01 9.00339723e-01
-6.84602141e-01 1.04112756e-02 -9.22692358e-01 6.82095528e-01
-1.39037699e-01 -4.05342057e-02 7.99518704e-01 3.76136661e-01
-8.60853314e-01 5.11555493e-01 -1.40124762e+00 -3.03425997e-01
4.72197860e-01 8.85679543e-01 -1.85420588e-01 9.02141407e-02
-1.04069531e+00 1.17681289e+00 2.99160630e-01 1.75038442e-01
-4.64550853e-01 -6.93217814e-01 -4.84286964e-01 -2.24168211e-01
-1.25314407e-02 -2.24536583e-01 1.01388168e+00 -6.82238638e-01
-9.20649946e-01 5.89024246e-01 -5.13959765e-01 -8.82049561e-01
2.20498130e-01 -1.70840815e-01 -5.46544015e-01 1.12348227e-02
-2.19982132e-01 4.86323297e-01 4.43255693e-01 -6.18214488e-01
-1.08226967e+00 -7.34627068e-01 -7.71476388e-01 1.18854130e-02
-7.99239874e-02 4.51195329e-01 3.50346774e-01 -6.60942972e-01
2.70350426e-01 -7.44958997e-01 -4.94187295e-01 -2.44714111e-01
1.28926739e-01 -2.27237239e-01 7.07622111e-01 -8.74091804e-01
1.35589635e+00 -1.78797174e+00 3.43552195e-02 -9.26495269e-02
3.95142943e-01 3.52952033e-01 4.20289844e-01 5.24843395e-01
-2.79852748e-01 -2.69496322e-01 -1.99916914e-01 1.14303276e-01
-5.59806108e-01 -4.66243550e-02 -4.97282565e-01 4.55941945e-01
2.62769431e-01 9.62749243e-01 -8.21992815e-01 -9.09123570e-03
4.23395902e-01 7.07606971e-01 -5.85605390e-02 4.26683903e-01
8.05138946e-02 9.55833077e-01 -5.65586984e-01 6.66367650e-01
1.89520568e-01 -1.70264110e-01 4.17216308e-02 2.90646970e-01
-1.03490964e-01 3.32634538e-01 -7.86262035e-01 1.25909197e+00
5.42380884e-02 7.45797932e-01 -2.99362332e-01 -1.55836833e+00
1.00018585e+00 8.28029335e-01 8.17686141e-01 -7.91520298e-01
1.06955305e-01 7.32380986e-01 4.33777690e-01 -7.34523475e-01
-5.65506339e-01 -1.04469009e-01 3.00463766e-01 3.33716631e-01
-5.21470964e-01 2.33950689e-02 -4.76921499e-02 -2.38575771e-01
1.33666682e+00 -1.25440851e-01 6.78328454e-01 -1.21471480e-01
6.07310057e-01 2.79064439e-02 8.25539291e-01 3.23447138e-01
-1.73980519e-01 2.95172065e-01 1.78201765e-01 -7.54683554e-01
-4.41810638e-01 -6.91908240e-01 -5.18553972e-01 2.66717643e-01
-4.14247751e-01 -2.36221194e-01 -6.33665502e-01 -2.78305948e-01
-2.18471318e-01 6.62641048e-01 -3.28089774e-01 -1.76487446e-01
-3.88083458e-01 -1.32891357e+00 1.71906859e-01 4.95698005e-01
3.38859707e-02 -1.21356416e+00 -1.08595490e+00 6.95889890e-01
-1.45220160e-01 -9.67768550e-01 3.36854398e-01 5.58263183e-01
-1.07838643e+00 -1.17510331e+00 -7.21148491e-01 -8.59654427e-01
5.59126735e-01 -1.85211703e-01 7.51125932e-01 1.58925027e-01
-5.96224904e-01 7.80941769e-02 -3.93318534e-01 -8.45027626e-01
-2.41084576e-01 -2.42627978e-01 3.73539656e-01 -7.67246336e-02
9.75548267e-01 -8.59905303e-01 -8.89764547e-01 -6.24105446e-02
-5.52514493e-01 1.82519153e-01 8.02545607e-01 8.02722394e-01
6.15003884e-01 3.44191998e-01 1.26627243e+00 -5.03985286e-01
7.34128892e-01 -6.79806948e-01 -4.34481353e-01 2.45030448e-01
-9.08141792e-01 -6.14581555e-02 4.03371036e-01 -4.72955972e-01
-4.58173156e-01 1.93758443e-01 -3.20190489e-01 1.05473213e-02
-4.72375065e-01 6.56462729e-01 1.42315939e-01 -3.12807351e-01
3.90473664e-01 5.04179597e-01 -1.39718235e-01 -5.13183355e-01
-5.71529508e-01 1.14966023e+00 2.41568401e-01 3.04662716e-02
9.72035676e-02 1.90729499e-01 1.39153535e-02 -9.00247037e-01
-6.84476614e-01 -4.69845861e-01 -4.38071936e-01 -1.89244866e-01
6.74066603e-01 -7.13145554e-01 -4.22205389e-01 5.07224202e-01
-8.91802788e-01 -3.32611538e-02 2.29565188e-01 7.61795819e-01
-4.09096867e-01 8.18606988e-02 -2.72643268e-01 -9.41236973e-01
-7.79145658e-01 -1.55504298e+00 6.12341166e-01 2.54722387e-01
-5.94044745e-01 -8.35807681e-01 1.18594185e-01 1.32484317e-01
3.39816213e-01 3.85144711e-01 1.07275295e+00 -9.96293426e-01
-3.42219830e-01 -5.26046336e-01 1.96469441e-01 2.34872982e-01
4.86892194e-01 -4.61741775e-01 -8.94609928e-01 -3.22806954e-01
5.34124076e-01 -9.68157500e-02 3.95227253e-01 6.78390443e-01
1.04214430e+00 -2.91885212e-02 -6.93535268e-01 5.48066854e-01
1.09202182e+00 1.08805943e+00 4.07760441e-01 3.35442007e-01
7.36186951e-02 4.92558509e-01 4.76085246e-01 3.87357622e-01
2.87646681e-01 5.16612530e-01 1.87286690e-01 1.08007908e-01
3.34559791e-02 2.16658741e-01 -3.06948088e-02 9.35237586e-01
1.00626618e-01 2.72724122e-01 -1.22706580e+00 6.28677249e-01
-1.64407074e+00 -9.01211560e-01 -6.75521642e-02 2.37370658e+00
9.33065534e-01 7.63732269e-02 -7.41509795e-02 7.14441121e-01
3.12074840e-01 -5.60645938e-01 -5.94985783e-01 -1.86116159e-01
1.73974819e-02 5.04004240e-01 -8.46608207e-02 1.14105001e-01
-8.16751003e-01 5.25042772e-01 6.61894941e+00 5.49291074e-01
-1.65055907e+00 -1.72068089e-01 7.64920056e-01 -7.76320370e-03
1.81101739e-01 -1.59170106e-01 -7.53343046e-01 6.85653150e-01
1.19136143e+00 -4.30987716e-01 5.67016900e-01 4.42142785e-01
6.16655827e-01 -4.14351732e-01 -1.13981342e+00 1.22068131e+00
1.90277562e-01 -1.24123275e+00 -5.32608509e-01 -3.09697315e-02
4.98524427e-01 1.96853250e-01 -1.43011525e-01 -1.53150067e-01
-5.19675374e-01 -1.22380102e+00 6.66510165e-02 6.90083385e-01
6.37905121e-01 -9.87608492e-01 8.22279215e-01 7.43115008e-01
-9.48415995e-01 -2.87582248e-01 -8.52863193e-02 -3.30952555e-01
7.97406733e-02 7.53902137e-01 -1.11227739e+00 2.22988307e-01
7.66478896e-01 8.36502671e-01 -4.46509898e-01 1.67533934e+00
-3.48107636e-01 6.22872412e-01 -1.99400663e-01 -4.37917076e-02
-5.98000139e-02 -7.33494461e-02 5.58609486e-01 7.96111226e-01
5.40927649e-01 4.38277960e-01 2.90725194e-02 4.81136292e-01
5.63243449e-01 1.19116299e-01 -4.17614728e-01 -4.06993657e-01
3.09114933e-01 1.01082349e+00 -7.18178630e-01 5.57107627e-02
-5.85843801e-01 6.58338606e-01 2.34221831e-01 4.06452157e-02
-1.02439858e-01 -2.24529326e-01 4.84163761e-01 1.09714672e-01
-2.10638553e-01 2.98338681e-02 -5.21201968e-01 -1.03300810e+00
-3.18304747e-02 -1.21003652e+00 2.49697238e-01 -4.57451969e-01
-8.78235638e-01 9.95632946e-01 -1.63512662e-01 -1.08594143e+00
-7.65564859e-01 -4.58671749e-01 -5.15148163e-01 7.90890813e-01
-1.49920011e+00 -6.55216694e-01 1.21041119e-01 2.62387246e-01
7.75091708e-01 -2.61399508e-01 1.34075677e+00 8.76174271e-02
-3.56654495e-01 2.23101899e-01 1.03923239e-01 -1.20837688e-01
5.47808945e-01 -9.25789535e-01 7.59695016e-04 7.02303708e-01
2.72483051e-01 2.34240204e-01 9.34346616e-01 -6.92906260e-01
-1.15633047e+00 -8.40730131e-01 1.21686554e+00 -3.08992535e-01
7.07971215e-01 -2.15496331e-01 -1.01811457e+00 5.56730568e-01
-1.28345549e-01 -2.86212564e-01 9.60154295e-01 -2.96673924e-01
4.48002875e-01 -1.30964860e-01 -1.03175557e+00 4.00964051e-01
3.00876290e-01 -2.19727486e-01 -8.04612219e-01 3.90341818e-01
-4.48273383e-02 -2.33148128e-01 -7.48751819e-01 8.10107052e-01
6.04313016e-01 -8.15451622e-01 6.13492966e-01 -3.55337560e-01
-1.57717809e-01 1.03066452e-01 3.28883410e-01 -1.42557096e+00
1.05044544e-01 -7.40239203e-01 -1.92895740e-01 4.68286037e-01
3.06211233e-01 -6.72259331e-01 8.44619572e-01 7.53340900e-01
1.66617297e-02 -1.48345876e+00 -9.08388078e-01 -4.13734257e-01
-8.75711534e-03 -7.28519440e-01 5.32686353e-01 4.32989299e-01
6.22969866e-01 4.28102724e-02 -3.33881140e-01 -1.24552837e-02
5.88164687e-01 3.12710665e-02 -4.45091128e-02 -1.50531495e+00
7.14413077e-02 -3.42388332e-01 -8.45677137e-01 -3.83454144e-01
1.55659057e-02 -7.17718005e-01 -1.22462362e-01 -1.64910734e+00
3.24263096e-01 -1.69846937e-01 -5.64777672e-01 5.84662378e-01
-1.64437085e-01 2.11358503e-01 -2.77061820e-01 1.67185381e-01
-1.03505149e-01 1.79605559e-01 7.53967643e-01 1.50297955e-01
-5.17822564e-01 5.69577932e-01 -7.24229336e-01 8.96939993e-01
1.03041542e+00 -7.08547771e-01 -4.43168193e-01 -1.39133304e-01
2.53038853e-01 4.58073437e-01 -3.80505584e-02 -1.05022323e+00
6.50758147e-01 -6.29471540e-02 5.97686172e-01 -8.26414049e-01
3.65952164e-01 -8.46460521e-01 2.43902102e-01 8.41754258e-01
-4.62548211e-02 2.98046172e-01 3.16159964e-01 2.23219633e-01
-3.42222929e-01 6.48557097e-02 7.02400565e-01 7.18666762e-02
-7.68801153e-01 2.29808137e-01 -8.05996537e-01 -1.26127988e-01
1.31366181e+00 -1.74029022e-01 1.14139393e-01 -4.37024146e-01
-1.04960632e+00 2.63376087e-01 -8.28624144e-02 5.19184649e-01
1.08604586e+00 -8.79865170e-01 -8.52222204e-01 5.52657545e-01
6.88918829e-02 -2.83034801e-01 1.73821107e-01 1.37034154e+00
-3.62397403e-01 8.56511474e-01 -1.37957975e-01 -4.35164630e-01
-1.43293774e+00 3.26765746e-01 4.68713462e-01 -1.11272745e-01
-9.33325529e-01 7.91359305e-01 3.10995281e-02 3.02470803e-01
6.32427752e-01 8.59703273e-02 -5.01458228e-01 -3.13709527e-01
1.09924519e+00 2.81585783e-01 4.73163962e-01 -4.08933848e-01
-6.77207291e-01 2.15312138e-01 -2.61168987e-01 2.65097469e-01
1.80123329e+00 1.01806737e-01 -3.04555893e-01 3.23843420e-01
4.85724688e-01 -3.88577998e-01 -7.59002626e-01 1.97595745e-01
4.57825154e-01 -2.55460590e-02 4.74700689e-01 -1.36959147e+00
-9.69096601e-01 1.10779619e+00 9.53145444e-01 -6.49055466e-02
1.42253661e+00 2.26147380e-02 7.34600902e-01 2.06393912e-01
8.22493494e-01 -7.22944796e-01 -4.09676522e-01 -6.72671571e-02
8.13469410e-01 -1.14839935e+00 -6.84895441e-02 2.92960275e-02
-4.17828083e-01 1.25700498e+00 2.63296869e-02 -9.64383930e-02
8.76367152e-01 6.94193780e-01 6.15440309e-02 -1.42632648e-01
-9.90733624e-01 6.95664808e-03 4.75153744e-01 5.95107675e-01
7.39529073e-01 1.51976511e-01 -7.21964478e-01 9.13741529e-01
-1.31294042e-01 3.75926673e-01 1.63163438e-01 9.76815641e-01
-6.70958996e-01 -1.63560998e+00 -3.21192533e-01 1.02016258e+00
-8.52109909e-01 -2.13200971e-01 -3.37112844e-01 5.08476257e-01
2.04845890e-01 1.12487793e+00 -2.14348599e-01 -3.41473639e-01
2.70652920e-02 2.79529452e-01 5.18048942e-01 -5.65196812e-01
-2.65047222e-01 4.60777968e-01 -2.61684865e-01 -4.06587213e-01
-1.25605986e-01 -9.66602206e-01 -1.36445868e+00 3.36256027e-01
-2.85533875e-01 3.78416806e-01 7.69768238e-01 1.35246849e+00
5.30335486e-01 5.49065530e-01 4.83910024e-01 -4.76643860e-01
-3.00126284e-01 -8.38616312e-01 -6.52666867e-01 -2.29859531e-01
2.75874197e-01 -6.54457271e-01 -5.46257675e-01 -1.18812293e-01] | [13.2317533493042, 3.52274227142334] |
e65145ac-ffdb-49b1-90c8-1ae8490772d4 | cbr-ikb-a-case-based-reasoning-approach-for | 2204.08554 | null | https://arxiv.org/abs/2204.08554v1 | https://arxiv.org/pdf/2204.08554v1.pdf | CBR-iKB: A Case-Based Reasoning Approach for Question Answering over Incomplete Knowledge Bases | Knowledge bases (KBs) are often incomplete and constantly changing in practice. Yet, in many question answering applications coupled with knowledge bases, the sparse nature of KBs is often overlooked. To this end, we propose a case-based reasoning approach, CBR-iKB, for knowledge base question answering (KBQA) with incomplete-KB as our main focus. Our method ensembles decisions from multiple reasoning chains with a novel nonparametric reasoning algorithm. By design, CBR-iKB can seamlessly adapt to changes in KBs without any task-specific training or fine-tuning. Our method achieves 100% accuracy on MetaQA and establishes new state-of-the-art on multiple benchmarks. For instance, CBR-iKB achieves an accuracy of 70% on WebQSP under the incomplete-KB setting, outperforming the existing state-of-the-art method by 22.3%. | ['Andrew McCallum', 'Achille Fokoue', 'Pavan Kapanipathi', 'Rajarshi Das', 'Tahira Naseem', 'Nandana Mihindukulasooriya', 'Mudit Chaudhary', 'Ibrahim Abdelaziz', 'Srinivas Ravishankar', 'Dung Thai'] | 2022-04-18 | null | null | null | null | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-5.85637629e-01 2.66580701e-01 -3.87995005e-01 -3.76998395e-01
-1.38167763e+00 -7.30893314e-01 6.00400344e-02 1.32563874e-01
-1.96594536e-01 1.34319043e+00 1.61782235e-01 -4.97974515e-01
-4.36613142e-01 -1.03864157e+00 -9.58827436e-01 -1.32223889e-01
4.44600731e-01 1.22465181e+00 9.99287009e-01 -7.49367833e-01
5.73654175e-02 1.86929718e-01 -1.26467085e+00 8.42014670e-01
1.29414165e+00 1.26978254e+00 -2.57571369e-01 6.26519322e-01
-6.53289020e-01 1.41053486e+00 -4.45396662e-01 -1.06615269e+00
-6.57071248e-02 7.10794479e-02 -1.45570493e+00 -7.64991999e-01
5.54152966e-01 -2.33781725e-01 -4.74824339e-01 7.98791349e-01
3.59310865e-01 3.40582788e-01 6.18478239e-01 -1.10646880e+00
-1.12750483e+00 7.36183345e-01 -2.61935145e-01 5.14101863e-01
6.93451524e-01 -3.32118608e-02 1.22326231e+00 -7.00912297e-01
5.79074085e-01 1.32266688e+00 8.53725553e-01 5.77220321e-01
-1.15669692e+00 -4.09621805e-01 2.97401041e-01 1.03709483e+00
-1.47561681e+00 -3.36413890e-01 5.55780768e-01 -2.27175698e-01
1.14597070e+00 3.27833980e-01 4.15637761e-01 8.87531996e-01
-2.69565582e-01 1.01486540e+00 9.70975995e-01 -4.16151822e-01
5.51214039e-01 2.11918652e-01 4.00723040e-01 6.54466271e-01
2.45993271e-01 -6.61984444e-01 -6.35266662e-01 -6.16529465e-01
2.78342515e-01 -3.06167573e-01 -2.28497133e-01 -3.41390967e-01
-9.23989534e-01 7.87928224e-01 4.25737560e-01 -2.93725878e-02
-5.32838464e-01 3.13262254e-01 3.17588985e-01 5.11715174e-01
2.45773330e-01 4.85876322e-01 -1.02266908e+00 -3.73244941e-01
-7.43059754e-01 7.75915682e-01 1.32511318e+00 1.05640841e+00
7.14303672e-01 -4.95152265e-01 -6.22826815e-01 1.11231172e+00
6.99505210e-02 6.94254220e-01 3.43150496e-01 -1.09955502e+00
7.33291984e-01 7.34753847e-01 4.90150779e-01 -6.47196829e-01
-2.11989284e-01 -6.18633740e-02 -5.36830008e-01 -6.67733490e-01
5.25480747e-01 1.46315545e-01 -7.99923182e-01 1.39934802e+00
6.93418443e-01 1.10563047e-01 3.53974849e-01 5.73791325e-01
9.06956017e-01 5.63962936e-01 -9.99591723e-02 4.28296588e-02
1.47048509e+00 -1.28296280e+00 -6.82421863e-01 -5.28628118e-02
4.14033234e-01 -1.20111175e-01 1.35291469e+00 5.74031413e-01
-1.09532368e+00 -1.38250977e-01 -5.84972084e-01 -2.69599110e-01
-4.31850970e-01 -4.61226583e-01 5.94720542e-01 6.52936399e-01
-1.03121972e+00 1.04928017e-03 -5.36246181e-01 -1.98907912e-01
4.34399098e-01 -7.08256364e-02 -1.52438402e-01 -6.37349904e-01
-1.64708126e+00 1.06217480e+00 6.18931055e-01 1.16092479e-02
-5.72743833e-01 -1.27240539e+00 -4.58510429e-01 1.59608930e-01
1.06393886e+00 -1.24706864e+00 1.82081437e+00 -3.02001864e-01
-1.72587037e+00 3.01032871e-01 -2.50743806e-01 -5.98262668e-01
6.13779068e-01 -4.39120948e-01 -5.80194890e-01 2.99508691e-01
3.87870036e-02 2.38637716e-01 5.75860322e-01 -1.23237848e+00
-6.82955921e-01 -2.29652956e-01 5.47081411e-01 5.81557304e-02
-1.14816874e-01 -5.20386159e-01 -6.87359929e-01 -3.69390965e-01
-6.63643405e-02 -6.32049441e-01 1.24191744e-02 -1.98389784e-01
-1.54699370e-01 -7.10708201e-01 5.36371768e-01 -9.76406991e-01
1.40241563e+00 -1.58818305e+00 4.21965271e-02 2.25125998e-01
3.38684767e-02 2.96595722e-01 -6.59215748e-02 4.32111681e-01
6.07990503e-01 5.91161661e-02 -2.16409117e-01 3.06727529e-01
1.59051582e-01 5.13784528e-01 -8.20007145e-01 -3.33959311e-01
2.13398099e-01 1.28757429e+00 -1.31686890e+00 -7.44108319e-01
-3.78961533e-01 -3.01303584e-02 -8.98827493e-01 8.23098645e-02
-9.93717432e-01 9.58622321e-02 -6.30836427e-01 1.09986711e+00
4.99751955e-01 -6.33120060e-01 3.32232505e-01 -2.05267817e-01
7.26233542e-01 4.00539100e-01 -1.02861106e+00 1.84476173e+00
-5.82350492e-01 1.62216455e-01 -3.04691911e-01 -9.98015285e-01
7.59223580e-01 2.20427707e-01 1.64344832e-02 -8.70847404e-01
-5.89806557e-01 3.99930567e-01 -2.17600852e-01 -7.00965703e-01
5.99679589e-01 -1.61147028e-01 -5.69798872e-02 1.85344651e-01
1.50760919e-01 -4.31106359e-01 3.43021631e-01 3.55464548e-01
1.55369854e+00 -1.50259554e-01 5.33224404e-01 -6.47951886e-02
6.00908577e-01 3.00486088e-01 4.93911505e-01 1.24640191e+00
-1.13736771e-01 1.87600195e-01 5.52493274e-01 -5.13802648e-01
-5.83097100e-01 -1.28865230e+00 5.33378916e-03 1.28855765e+00
8.56756717e-02 -1.51431844e-01 -3.48001778e-01 -1.10908639e+00
6.57474756e-01 9.93044019e-01 -5.94494104e-01 -2.17259943e-01
-3.57078344e-01 -4.69170958e-01 7.85541654e-01 6.24731004e-01
6.40760422e-01 -1.19957721e+00 3.30736414e-02 5.51479697e-01
-6.65376842e-01 -1.23264694e+00 6.00389056e-02 -1.95748627e-01
-9.61235046e-01 -1.17961657e+00 -5.71575522e-01 -2.04865798e-01
1.25533998e-01 5.42717008e-03 1.74582577e+00 -4.25180756e-02
4.44523469e-02 7.24080861e-01 -7.01942682e-01 -3.67907107e-01
-3.59905034e-01 6.67806625e-01 -2.31347769e-01 -3.26074332e-01
3.78862828e-01 -6.63943231e-01 -7.30080903e-01 2.85651207e-01
-9.41124737e-01 -4.10144567e-01 6.36690974e-01 9.58183706e-01
6.39586329e-01 5.47322445e-02 1.24946308e+00 -1.25741601e+00
8.39669168e-01 -8.31790447e-01 -5.46626627e-01 1.06511199e+00
-7.26540685e-01 2.77043402e-01 4.87014234e-01 -4.40888584e-01
-1.45747089e+00 -7.64595270e-01 -1.20463528e-01 -2.08572298e-01
1.92917094e-01 9.05460715e-01 -4.44609970e-02 5.85129112e-02
1.07520735e+00 1.14019543e-01 -4.69018489e-01 -5.12952983e-01
7.07460523e-01 7.18660235e-01 5.55257440e-01 -1.24485373e+00
6.61787331e-01 5.33559442e-01 -4.57232982e-01 -2.65634716e-01
-1.45997083e+00 -6.56670272e-01 -2.91934073e-01 6.78683296e-02
2.87629068e-01 -9.46788907e-01 -8.45340788e-01 2.32812196e-01
-9.80256975e-01 -5.99756658e-01 -4.75899845e-01 -4.35766503e-02
-5.89958131e-01 2.65093237e-01 -5.74442267e-01 -7.94584692e-01
-5.89476228e-01 -6.25121653e-01 7.64040947e-01 2.54528165e-01
-1.41899943e-01 -1.00777757e+00 3.96576434e-01 1.33698535e+00
5.18400133e-01 -3.21215093e-01 1.25271249e+00 -8.03040802e-01
-6.52509153e-01 -9.07638073e-02 -4.10456717e-01 2.91614354e-01
1.84886847e-02 -3.47169399e-01 -9.06503558e-01 5.69603778e-02
-4.69429940e-01 -8.09147179e-01 9.67835367e-01 -8.43007043e-02
1.46036708e+00 -5.30404031e-01 -1.04503579e-01 1.55718341e-01
1.21210754e+00 1.80845112e-02 7.55418360e-01 6.79099739e-01
4.62467104e-01 3.24211329e-01 7.21481681e-01 3.90129983e-01
1.05795300e+00 5.19980848e-01 2.21189976e-01 6.39895618e-01
2.16447622e-01 -2.56305873e-01 6.20760657e-02 8.27654004e-01
-8.56794044e-02 -1.03849299e-01 -1.37355745e+00 8.45874906e-01
-2.26854515e+00 -8.94943595e-01 2.50469387e-01 1.67757571e+00
1.59036994e+00 5.34908734e-02 -4.59147282e-02 -1.68700829e-01
3.52428406e-01 -4.66275588e-02 -1.07714272e+00 -2.75792837e-01
8.60585868e-02 3.29438478e-01 2.67179668e-01 3.52598131e-01
-6.90372825e-01 1.04910362e+00 6.29392242e+00 1.08721805e+00
-3.93467873e-01 1.54840931e-01 3.24591130e-01 -1.39576644e-01
-6.43644452e-01 3.08794510e-02 -8.27787876e-01 3.54572594e-01
8.99798453e-01 -7.13315383e-02 7.82536924e-01 1.02533877e+00
-6.38052821e-01 -2.27581754e-01 -8.15449953e-01 8.20492983e-01
-3.30547094e-02 -1.77529418e+00 5.04433513e-01 -5.36547661e-01
1.02052164e+00 5.35488175e-03 -2.04782635e-01 1.10517216e+00
8.57923567e-01 -8.67342591e-01 6.94953859e-01 1.05370450e+00
5.80402911e-01 -6.08156919e-01 9.29616928e-01 5.87485611e-01
-8.17119658e-01 -2.92079926e-01 -4.73724246e-01 2.02599391e-01
1.62722781e-01 7.71693289e-01 -1.01463306e+00 1.03298283e+00
1.13216543e+00 1.82156861e-01 -6.18751764e-01 8.36112797e-01
-2.72016078e-01 9.81836855e-01 -3.18075627e-01 -1.58654690e-01
2.67407130e-02 1.18695356e-01 1.84487566e-01 1.08946431e+00
4.44061905e-02 4.46480125e-01 -3.26684639e-02 6.71481371e-01
-4.31052357e-01 2.19342075e-02 -1.23793154e-03 -6.43724925e-04
8.12552512e-01 9.04457152e-01 -6.56673312e-02 -5.60213208e-01
-3.31291795e-01 6.71099484e-01 9.99713123e-01 6.32018983e-01
-7.25492477e-01 -2.23619461e-01 4.88482296e-01 -7.77543262e-02
7.43862092e-01 3.30315351e-01 1.38031483e-01 -1.40379035e+00
4.56697732e-01 -1.18305838e+00 1.00965250e+00 -1.03569043e+00
-2.00296092e+00 3.05499792e-01 2.80595243e-01 -3.81053537e-01
-3.88705403e-01 -4.54516262e-01 1.04104280e-01 6.50696695e-01
-2.05375648e+00 -1.17402077e+00 -4.88902658e-01 7.24300504e-01
1.64959043e-01 -1.78400397e-01 8.34574103e-01 1.57489717e-01
-1.52996987e-01 6.17317200e-01 4.55590457e-01 9.94467661e-02
8.25724244e-01 -1.44997871e+00 2.53284931e-01 5.07873856e-02
1.28317446e-01 6.72838509e-01 4.47946787e-01 -6.16626740e-01
-1.50743032e+00 -1.08518076e+00 5.85184395e-01 -9.56699073e-01
1.05416095e+00 3.83740989e-03 -1.40529871e+00 7.81396747e-01
-3.26324016e-01 4.16408747e-01 7.59668052e-01 6.84665203e-01
-1.03817356e+00 -6.10824406e-01 -1.28726041e+00 5.53331733e-01
9.53405023e-01 -8.17148089e-01 -1.22968781e+00 2.78937697e-01
9.91461694e-01 -7.74426579e-01 -1.13662279e+00 6.27069056e-01
6.90199196e-01 -8.73767555e-01 1.24950790e+00 -1.15597773e+00
1.31608903e-01 -3.39415073e-01 -3.50780070e-01 -1.16895330e+00
-2.47497812e-01 -3.44342440e-01 -1.00124395e+00 1.00822699e+00
5.57097375e-01 -8.13900888e-01 8.90776575e-01 7.60558069e-01
2.13394165e-01 -9.70624149e-01 -1.12833810e+00 -7.81673968e-01
1.32351115e-01 -5.10165513e-01 9.51408207e-01 9.26700830e-01
-1.41953632e-01 2.31469333e-01 -8.87238327e-03 2.49843195e-01
1.96885288e-01 3.66578043e-01 8.85951221e-01 -1.16115618e+00
-4.56789136e-01 -1.44938663e-01 -1.77045092e-01 -1.08726883e+00
3.54377329e-01 -6.96456432e-01 -1.31671399e-01 -1.88824010e+00
3.35448563e-01 -7.30455935e-01 -5.43029368e-01 6.89005733e-01
-6.00869596e-01 -1.12140223e-01 -1.62948444e-01 1.70928121e-01
-1.26507211e+00 7.41696596e-01 1.15256178e+00 -4.39847350e-01
-2.22469375e-01 -8.46197009e-02 -9.76101577e-01 5.46226740e-01
6.96741521e-01 -4.72002387e-01 -6.41640007e-01 -2.09998533e-01
9.72615898e-01 2.32354492e-01 3.70880753e-01 -7.76647866e-01
4.31866020e-01 -5.10940611e-01 -2.65069585e-02 -6.84283257e-01
3.17476690e-01 -6.86999321e-01 3.11329383e-02 6.84330091e-02
-2.96380967e-01 -1.79070532e-01 3.72115999e-01 1.12601614e+00
-4.32472616e-01 -3.58622253e-01 1.85562998e-01 -1.80381149e-01
-9.03326929e-01 1.68822944e-01 3.44884545e-02 9.24316227e-01
6.31727338e-01 1.36210352e-01 -7.73717523e-01 -3.68800908e-01
-4.32699084e-01 6.41022921e-01 6.44443631e-02 1.80687413e-01
3.38260651e-01 -1.21629572e+00 -5.63074231e-01 -5.04066885e-01
5.47209263e-01 6.99365556e-01 6.33716166e-01 7.26717949e-01
-6.46669090e-01 5.91425061e-01 1.96199074e-01 -2.65141010e-01
-5.97617865e-01 4.16923612e-01 4.31427896e-01 -8.82174730e-01
-3.73096436e-01 8.54851365e-01 -3.23152095e-01 -8.90323400e-01
1.32829681e-01 -5.31091273e-01 -5.08524105e-02 3.48396339e-02
4.50147778e-01 5.53632140e-01 4.46330935e-01 3.45304966e-01
-2.59185165e-01 -4.84190732e-02 -5.07100403e-01 -8.29121470e-02
1.23185003e+00 9.92007703e-02 -2.59235084e-01 7.57801831e-01
5.04774690e-01 -9.81728658e-02 -9.46378946e-01 -6.64521337e-01
3.72541755e-01 -2.91327745e-01 -2.62902558e-01 -1.44220066e+00
-6.18142188e-01 4.23533857e-01 -1.68726385e-01 1.36266544e-01
7.92868555e-01 3.28243017e-01 9.07701075e-01 1.54474652e+00
5.97216606e-01 -1.13918364e+00 2.46108651e-01 8.02493393e-01
1.15715098e+00 -1.33204579e+00 3.82716805e-02 -1.56833634e-01
-7.29640722e-01 8.59848857e-01 8.04892123e-01 2.87132829e-01
6.31734967e-01 -1.67685181e-01 8.59386250e-02 -1.99134365e-01
-1.22688758e+00 -2.11380776e-02 5.04939616e-01 7.02861309e-01
-2.31920332e-01 9.41823050e-03 2.38831401e-01 9.07225966e-01
-2.82223463e-01 3.35238129e-01 -2.37932522e-02 1.07885075e+00
-2.95508802e-01 -8.35786581e-01 -3.76246184e-01 8.36612165e-01
-2.00217366e-01 -2.63324589e-01 -2.70285249e-01 8.87879491e-01
-3.86802256e-01 8.14717114e-01 -1.63527519e-01 6.35290071e-02
5.67264855e-01 6.55605018e-01 6.25362098e-01 -5.00085533e-01
-4.59948152e-01 -9.98098910e-01 3.03817302e-01 -4.49173063e-01
-3.42779607e-01 -4.45582330e-01 -1.15693486e+00 -3.54793787e-01
-3.79518747e-01 4.28774655e-01 1.69102654e-01 9.94817555e-01
6.03863358e-01 3.80264044e-01 -1.44020960e-01 2.64788002e-01
-1.09945405e+00 -1.04312968e+00 -4.64772284e-01 3.76991272e-01
2.32557178e-01 -8.03991377e-01 -8.06599930e-02 -2.59977719e-03] | [10.495563507080078, 7.936329364776611] |
f7ba79ba-71d3-46a4-900b-182fd22229e4 | chinese-grammatical-error-diagnosis-using-1 | null | null | https://aclanthology.org/W15-4415 | https://aclanthology.org/W15-4415.pdf | Chinese Grammatical Error Diagnosis Using Ensemble Learning | null | ['Wenying Han', 'Xiaolong Wang', 'Yang Xiang', 'Qinghua Hong'] | 2015-07-01 | null | null | null | ws-2015-7 | ['grammatical-error-detection'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.45961856842041, 3.7029452323913574] |
edbc1cc3-5f6e-4c52-a799-4ccb82b233f6 | jokr-joint-keypoint-representation-for | 2106.09679 | null | https://arxiv.org/abs/2106.09679v1 | https://arxiv.org/pdf/2106.09679v1.pdf | JOKR: Joint Keypoint Representation for Unsupervised Cross-Domain Motion Retargeting | The task of unsupervised motion retargeting in videos has seen substantial advancements through the use of deep neural networks. While early works concentrated on specific object priors such as a human face or body, recent work considered the unsupervised case. When the source and target videos, however, are of different shapes, current methods fail. To alleviate this problem, we introduce JOKR - a JOint Keypoint Representation that captures the motion common to both the source and target videos, without requiring any object prior or data collection. By employing a domain confusion term, we enforce the unsupervised keypoint representations of both videos to be indistinguishable. This encourages disentanglement between the parts of the motion that are common to the two domains, and their distinctive appearance and motion, enabling the generation of videos that capture the motion of the one while depicting the style of the other. To enable cases where the objects are of different proportions or orientations, we apply a learned affine transformation between the JOKRs. This augments the representation to be affine invariant, and in practice broadens the variety of possible retargeting pairs. This geometry-driven representation enables further intuitive control, such as temporal coherence and manual editing. Through comprehensive experimentation, we demonstrate the applicability of our method to different challenging cross-domain video pairs. We evaluate our method both qualitatively and quantitatively, and demonstrate that our method handles various cross-domain scenarios, such as different animals, different flowers, and humans. We also demonstrate superior temporal coherency and visual quality compared to state-of-the-art alternatives, through statistical metrics and a user study. Source code and videos can be found at https://rmokady.github.io/JOKR/ . | ['Daniel Cohen-Or', 'Amit H. Bermano', 'Sagie Benaim', 'Rotem Tzaban', 'Ron Mokady'] | 2021-06-17 | null | null | null | null | ['motion-retargeting'] | ['computer-vision'] | [ 1.31862983e-01 -2.43674085e-01 -2.98203051e-01 1.36832008e-02
-3.76554996e-01 -1.04481506e+00 8.55564177e-01 -3.51619810e-01
-8.99326578e-02 4.49186862e-01 3.83367270e-01 1.33199692e-01
7.28175938e-02 -4.16799664e-01 -6.47496879e-01 -7.53120899e-01
1.34403659e-02 8.03255737e-02 3.64604264e-01 -1.47128582e-01
8.50632340e-02 6.85177207e-01 -1.43294787e+00 1.66009828e-01
4.14730072e-01 7.19874918e-01 1.95552751e-01 7.26009309e-01
4.40838724e-01 3.94753486e-01 -4.95455384e-01 -1.85642406e-01
6.29485488e-01 -3.62010986e-01 -7.22542465e-01 5.25178790e-01
9.12471175e-01 -4.61881161e-01 -6.58092678e-01 9.10843790e-01
3.00478578e-01 3.72217506e-01 6.28727078e-01 -1.34313774e+00
-7.88191676e-01 6.12300746e-02 -7.78252244e-01 9.14593861e-02
6.18108332e-01 4.04771864e-01 8.89223993e-01 -7.65334249e-01
9.82663810e-01 1.16942644e+00 4.84533906e-01 4.84352261e-01
-1.59566236e+00 -5.71119606e-01 2.12165102e-01 6.62847757e-02
-1.39898491e+00 -6.24284387e-01 1.02608800e+00 -6.64943933e-01
4.17032719e-01 1.90936476e-01 7.53844142e-01 1.33787155e+00
1.68754205e-01 7.28790343e-01 7.84455657e-01 -2.68655866e-01
6.59160092e-02 -1.60231307e-01 -3.68202090e-01 4.51838374e-01
2.01995656e-01 2.08644927e-01 -5.57296395e-01 -7.04749897e-02
1.20487142e+00 2.04876121e-02 -5.54593742e-01 -1.12132490e+00
-1.51727247e+00 6.45252168e-01 1.66531101e-01 2.46153206e-01
-2.12918058e-01 1.77257538e-01 1.35801435e-01 1.38413295e-01
1.54060304e-01 5.92612743e-01 -3.88355434e-01 4.67658900e-02
-9.15986359e-01 4.05548483e-01 6.78253412e-01 1.07643759e+00
6.86250627e-01 1.52672138e-02 -1.68368801e-01 5.08371592e-01
1.99398160e-01 3.83288741e-01 3.79251450e-01 -1.31682038e+00
2.13346407e-01 3.85896266e-01 4.35677111e-01 -1.40604329e+00
-2.89337754e-01 -2.19876561e-02 -6.94647610e-01 2.98418522e-01
7.03612089e-01 -5.91924973e-02 -1.01889765e+00 1.91046619e+00
4.67355251e-01 2.33978689e-01 -5.28263338e-02 9.67411160e-01
6.71063542e-01 4.05618668e-01 -1.02045678e-01 -1.58844858e-01
1.27517343e+00 -9.66471136e-01 -5.57395041e-01 -2.28317857e-01
4.35508788e-01 -9.40486848e-01 1.02480125e+00 1.78186536e-01
-1.16848242e+00 -5.00861645e-01 -9.95836079e-01 -1.40591845e-01
-1.91480726e-01 2.03797579e-01 4.17755157e-01 2.85403907e-01
-1.02260482e+00 6.51403129e-01 -7.74047494e-01 -5.24288237e-01
2.20693246e-01 2.47037381e-01 -7.41987169e-01 -1.17788643e-01
-9.19362366e-01 6.33903325e-01 2.37102941e-01 -6.25550672e-02
-9.34687436e-01 -7.11196423e-01 -9.61394846e-01 -7.35442713e-02
4.63092417e-01 -8.13903093e-01 1.28754568e+00 -1.36702168e+00
-1.62711179e+00 9.11482692e-01 7.54086971e-02 3.08497269e-02
6.18682504e-01 -3.42224389e-01 -3.69586706e-01 3.88164699e-01
7.22638294e-02 9.14228559e-01 1.26934707e+00 -1.20730710e+00
-5.21133006e-01 -2.26722099e-02 3.11458915e-01 3.15665096e-01
-1.67308763e-01 -5.40269762e-02 -8.32535088e-01 -1.12776184e+00
-7.05348328e-02 -1.32659423e+00 3.87658291e-02 4.30852085e-01
-3.47168148e-01 2.59114861e-01 1.16353381e+00 -5.66801727e-01
9.73983109e-01 -2.50105381e+00 6.31065786e-01 6.57723173e-02
1.88325033e-01 1.15584627e-01 -3.69592398e-01 3.96238446e-01
-3.67662579e-01 -6.48140558e-04 -1.75868765e-01 -2.42099091e-01
-1.54130727e-01 2.18896076e-01 -1.58595979e-01 8.04571390e-01
2.21044928e-01 7.65217960e-01 -1.02780831e+00 -2.29955077e-01
1.56191304e-01 5.13532698e-01 -5.67436039e-01 2.90055394e-01
-1.39962897e-01 7.93249905e-01 -1.64368421e-01 4.75264549e-01
5.91248155e-01 -3.35052699e-01 2.66805112e-01 -5.97876906e-01
9.08074751e-02 5.98838665e-02 -1.47581577e+00 1.78547919e+00
-8.50192457e-02 8.39316845e-01 1.71746120e-01 -7.20683575e-01
5.39620340e-01 3.34237367e-01 6.63733959e-01 -4.44307745e-01
-3.94072197e-03 -8.32763389e-02 3.39448266e-02 -4.91148442e-01
6.25687242e-01 5.39731458e-02 3.47755514e-02 4.07149553e-01
1.30501553e-01 -3.31201613e-01 3.20562363e-01 2.37975672e-01
8.18293333e-01 5.30885279e-01 4.02494222e-01 -2.35301480e-01
1.16293333e-01 -1.79721028e-01 5.96494615e-01 5.35171807e-01
-1.95352733e-01 9.06419218e-01 3.30781758e-01 -3.99296880e-01
-1.11744189e+00 -1.17149031e+00 1.18168518e-01 9.59350407e-01
4.79381204e-01 -3.57031614e-01 -4.06060815e-01 -6.62541091e-01
4.84119095e-02 3.94937903e-01 -7.61761665e-01 -1.08266450e-01
-6.32773459e-01 -2.22697273e-01 3.15781683e-01 4.76952493e-01
2.94307888e-01 -7.81305492e-01 -8.21058750e-01 -6.29756972e-02
-2.85110354e-01 -1.29960680e+00 -9.81530249e-01 -1.03438586e-01
-7.01461375e-01 -1.14343190e+00 -9.48377073e-01 -6.92652285e-01
6.57725811e-01 4.95958805e-01 1.01463473e+00 -9.20324847e-02
-3.37279439e-01 8.17925453e-01 -3.65081102e-01 2.98065692e-01
-2.38503069e-01 -2.51393020e-01 2.68180251e-01 1.85622245e-01
-1.80785447e-01 -7.12435842e-01 -8.48344505e-01 5.79719841e-01
-1.16495466e+00 2.19679087e-01 2.35602871e-01 7.01565385e-01
2.99455166e-01 -2.01229945e-01 3.78622450e-02 -5.26803672e-01
2.57761568e-01 -3.82470310e-01 -5.56017101e-01 4.69116569e-02
7.58550912e-02 -2.27762666e-02 4.68084514e-01 -9.73765194e-01
-8.51544023e-01 2.62324542e-01 4.27302301e-01 -7.59663701e-01
-2.65588820e-01 2.39681855e-01 -2.22885653e-01 -1.40141189e-01
5.82751393e-01 6.18138611e-02 -3.12806927e-02 -3.13112408e-01
6.35536313e-01 1.81850538e-01 7.03361988e-01 -5.69984376e-01
1.01470864e+00 7.58053362e-01 -1.48426011e-01 -9.73626316e-01
-4.21256632e-01 -3.07986885e-01 -9.13431048e-01 -2.18074262e-01
8.95423293e-01 -8.10803771e-01 -3.13145280e-01 4.26190734e-01
-1.10431933e+00 -5.17160535e-01 -3.55384082e-01 5.33803821e-01
-6.77708983e-01 5.88811517e-01 -3.60857368e-01 -2.86815763e-01
2.42694646e-01 -1.25249016e+00 1.16954672e+00 2.77533710e-01
-5.18795252e-01 -1.02066410e+00 9.06302184e-02 -6.83730319e-02
1.42822862e-01 3.66012216e-01 7.78240561e-01 -4.43658143e-01
-6.61437571e-01 -5.63091636e-02 -2.40818132e-02 3.75278629e-02
5.41525841e-01 4.05179173e-01 -7.39883661e-01 -5.76390445e-01
-3.44377577e-01 -1.10973217e-01 4.78244096e-01 4.17842358e-01
6.74930692e-01 -4.17015225e-01 -3.87758195e-01 7.47338355e-01
1.03876853e+00 1.83821276e-01 6.70498550e-01 3.24043840e-01
9.44357574e-01 7.61875451e-01 4.79413658e-01 4.01887715e-01
2.01310128e-01 1.12748814e+00 3.67477685e-01 -1.44667149e-01
-2.16701314e-01 -1.04682393e-01 4.26444620e-01 4.77465987e-01
-8.93870518e-02 -2.87301421e-01 -6.66422665e-01 7.21767664e-01
-1.79683161e+00 -9.79990304e-01 3.68472010e-01 2.32911730e+00
5.76020062e-01 -1.94748446e-01 2.88483948e-01 -2.18487829e-01
6.94084406e-01 3.10188979e-01 -6.01294518e-01 -6.52352395e-03
-6.87052235e-02 -1.78696185e-01 3.37853342e-01 3.31233084e-01
-1.32904410e+00 8.83480549e-01 6.36114788e+00 5.84357560e-01
-1.31871927e+00 -1.97532669e-01 3.88487548e-01 -2.83356935e-01
-2.68747211e-01 1.20574139e-01 -3.94739777e-01 3.48038942e-01
2.85321295e-01 -1.02911443e-01 4.16117460e-01 5.77405691e-01
2.99631506e-01 -2.23268457e-02 -1.42133391e+00 9.68771517e-01
6.94706589e-02 -1.25968659e+00 5.02720810e-02 -6.77228495e-02
8.96470666e-01 -2.89347500e-01 2.36901164e-01 -6.05428517e-02
2.75337309e-01 -8.65638077e-01 9.33418453e-01 5.40085912e-01
8.60278904e-01 -3.59802067e-01 2.12396622e-01 -1.44032137e-02
-1.35765517e+00 2.16599796e-02 -4.70252037e-02 9.00236294e-02
1.76508099e-01 1.88168243e-01 -4.01785731e-01 6.34519696e-01
8.32731843e-01 9.92554486e-01 -5.02712786e-01 9.09813762e-01
-1.17168725e-01 1.87798366e-01 -2.92073727e-01 4.57492352e-01
2.18045730e-02 -2.14206383e-01 9.87300456e-01 1.06846774e+00
3.00773650e-01 1.71846226e-02 3.61144215e-01 7.20725000e-01
-3.76999564e-02 -6.84887692e-02 -8.69897425e-01 -1.63188249e-01
4.71010268e-01 1.18897879e+00 -8.44779968e-01 -1.91798598e-01
-4.56620187e-01 1.22480500e+00 1.78553194e-01 7.70887375e-01
-7.93333054e-01 -2.09273234e-01 9.41955030e-01 2.10647896e-01
6.99367225e-01 -6.04527116e-01 1.32699713e-01 -1.39164746e+00
6.16885647e-02 -9.81903315e-01 2.77292520e-01 -9.17484522e-01
-1.04956853e+00 3.23373616e-01 3.95588547e-01 -1.57860994e+00
-3.07176471e-01 -5.24437308e-01 -5.71984112e-01 5.71152985e-01
-1.00548160e+00 -1.30474079e+00 -3.40774715e-01 5.31338096e-01
4.26534295e-01 2.81007588e-02 5.23517728e-01 2.45924413e-01
-4.46331322e-01 4.92018998e-01 7.47927278e-02 5.59599809e-02
1.02441883e+00 -8.88992727e-01 3.18670124e-01 1.04597867e+00
1.94158062e-01 6.86953843e-01 9.48870301e-01 -5.41390300e-01
-1.46370566e+00 -1.04223549e+00 2.34378651e-01 -5.54847300e-01
6.07697010e-01 -2.93262541e-01 -8.22038114e-01 7.88792133e-01
2.40228444e-01 1.74560234e-01 4.45307940e-01 -2.57608503e-01
-5.68414271e-01 1.55949160e-01 -8.57748151e-01 1.21517158e+00
1.05828333e+00 -4.61058617e-01 -3.45091581e-01 9.90290567e-02
7.08372593e-01 -5.28840184e-01 -8.40382695e-01 3.80368799e-01
8.41520607e-01 -1.00965822e+00 1.12359571e+00 -5.79912066e-01
4.83223736e-01 -6.68184280e-01 -2.68605918e-01 -1.33439016e+00
-4.89006817e-01 -8.02030623e-01 -3.12113762e-01 1.00407231e+00
4.45167460e-02 -3.60448360e-01 5.44370174e-01 5.38104057e-01
1.52225822e-01 -4.94314849e-01 -8.04584265e-01 -8.33915055e-01
-1.61823243e-01 -9.68448445e-02 3.47205102e-01 1.20525694e+00
-1.30530149e-01 2.31095612e-01 -6.64233446e-01 2.33096391e-01
1.87718302e-01 2.73023367e-01 1.01428890e+00 -9.34358597e-01
-4.95805055e-01 -4.63693053e-01 -5.89954376e-01 -1.27925980e+00
4.54588793e-02 -5.41332603e-01 1.39256762e-02 -1.18585098e+00
6.73791915e-02 -7.34538659e-02 1.78940296e-01 6.04980409e-01
-9.42352489e-02 4.57405984e-01 4.53792423e-01 3.28566670e-01
-3.86594892e-01 6.28590405e-01 1.43767846e+00 -1.06881022e-01
-4.80712712e-01 -4.22403693e-01 -6.39724851e-01 8.83825541e-01
6.27375901e-01 -2.83977598e-01 -4.05563384e-01 -5.53902388e-01
-2.97853295e-02 -1.34605903e-03 5.46970189e-01 -9.22682345e-01
-4.20817956e-02 -3.44859928e-01 5.04231393e-01 -3.48303616e-02
4.84897405e-01 -9.71843600e-01 4.61458981e-01 2.42097303e-01
-2.45031819e-01 2.58518934e-01 4.45797175e-01 7.51084924e-01
-3.87650095e-02 8.09589699e-02 9.53376114e-01 5.94535470e-02
-7.04606593e-01 3.95296782e-01 -3.49066347e-01 8.68615210e-02
1.22293663e+00 -5.28771102e-01 -1.67868927e-01 -7.15243757e-01
-7.66462147e-01 -7.51414225e-02 9.61195707e-01 6.29741371e-01
5.00192761e-01 -1.47453952e+00 -4.81023788e-01 3.02233160e-01
1.87150806e-01 -4.79355082e-02 2.94543892e-01 1.05322933e+00
-4.58454192e-01 5.68448491e-02 -4.22944337e-01 -7.94326365e-01
-1.33112943e+00 6.25654161e-01 3.24523807e-01 5.88603280e-02
-6.95432365e-01 4.51297909e-01 7.48470068e-01 -2.00540856e-01
3.79568823e-02 -2.34526843e-01 -1.65920570e-01 1.76235531e-02
3.28402400e-01 2.84051359e-01 -3.16043139e-01 -9.66157854e-01
-3.37924719e-01 7.69999981e-01 -1.44446313e-01 -1.69387683e-01
1.07162762e+00 -2.46597484e-01 2.29884014e-01 3.11410904e-01
1.18359041e+00 3.39721829e-01 -1.72114599e+00 -1.86090082e-01
-2.96280563e-01 -7.78071940e-01 -3.74650598e-01 -3.42591703e-01
-1.07373703e+00 5.51515698e-01 5.11902273e-01 9.39572901e-02
1.22781956e+00 1.06165353e-02 4.10607070e-01 1.07025117e-01
1.78878307e-01 -7.52146423e-01 4.18064952e-01 4.69895422e-01
1.02830601e+00 -1.09334338e+00 1.02428317e-01 -3.14588308e-01
-6.75169230e-01 1.11972904e+00 6.10409856e-01 -2.69715369e-01
4.70954478e-01 2.01508012e-02 1.02401428e-01 -4.93051596e-02
-4.51954246e-01 -1.66845545e-02 5.54687262e-01 6.88771248e-01
3.44704330e-01 -9.26364586e-02 5.56470677e-02 1.14545457e-01
-1.14754979e-02 -1.33984610e-01 6.31361485e-01 1.08373809e+00
5.74304201e-02 -1.03344107e+00 -4.45991904e-01 1.55488297e-01
-2.85931081e-01 6.96771815e-02 -4.77670461e-01 1.09609079e+00
-2.57573519e-02 6.98489070e-01 2.13708580e-01 -3.04458499e-01
2.30827853e-01 -2.92581737e-01 7.22400725e-01 -5.93764365e-01
-9.66782570e-02 3.44361067e-01 -6.41655624e-02 -7.72421896e-01
-6.82871103e-01 -6.14347577e-01 -8.07057679e-01 -2.22622231e-01
-1.85984209e-01 -2.00548664e-01 1.46734238e-01 7.90891707e-01
5.62529027e-01 3.55903327e-01 3.59308988e-01 -1.50742972e+00
-1.25607446e-01 -7.05020845e-01 -5.50953567e-01 7.41956711e-01
6.40325427e-01 -1.02630460e+00 -2.74222940e-01 5.76993108e-01] | [10.624373435974121, -0.8155587911605835] |
19e8e18f-2e14-4b14-b423-51f56cefc298 | mapping-it-differently-a-solution-to-the | null | null | https://aclanthology.org/2016.gwc-1.57 | https://aclanthology.org/2016.gwc-1.57.pdf | Mapping it differently: A solution to the linking challenges | This paper reports the work of creating bilingual mappings in English for certain synsets of Hindi wordnet, the need for doing this, the methods adopted and the tools created for the task. Hindi wordnet, which forms the foundation for other Indian language wordnets, has been linked to the English WordNet. To maximize linkages, an important strategy of using direct and hypernymy linkages has been followed. However, the hypernymy linkages were found to be inadequate in certain cases and posed a challenge due to sense granularity of language. Thus, the idea of creating bilingual mappings was adopted as a solution. A bilingual mapping means a linkage between a concept in two different languages, with the help of translation and/or transliteration. Such mappings retain meaningful representations, while capturing semantic similarity at the same time. This has also proven to be a great enhancement of Hindi wordnet and can be a crucial resource for multilingual applications in natural language processing, including machine translation and cross language information retrieval. | ['Pushpak Bhattacharyya', 'Diptesh Kanojia', 'Laxmi Kashyap', 'Jaya Saraswati', 'Rajita Shukla', 'Meghna Singh'] | null | null | null | null | gwc-2016-1 | ['transliteration'] | ['natural-language-processing'] | [-1.83449507e-01 -1.93348210e-02 -4.38435912e-01 -2.00726658e-01
-3.01075667e-01 -7.27606237e-01 9.66831923e-01 4.03660208e-01
-7.77120352e-01 1.15162480e+00 4.76707935e-01 -3.26402336e-01
-3.46100748e-01 -1.05217350e+00 -4.57549319e-02 -3.19208413e-01
3.39907557e-01 7.62472510e-01 1.10394076e-01 -8.94791722e-01
3.45503181e-01 3.60765129e-01 -1.20594490e+00 1.11613482e-01
7.48792827e-01 6.96027279e-02 6.64868534e-01 -6.56212494e-02
-7.04135537e-01 7.57725775e-01 -3.34100217e-01 -6.51888371e-01
2.10326567e-01 -5.92427433e-01 -1.23823857e+00 -3.80599529e-01
3.05315405e-02 3.88558060e-01 -2.16270894e-01 1.32382452e+00
4.58719641e-01 -1.53076192e-02 2.91611195e-01 -1.08201838e+00
-5.48604608e-01 9.39227760e-01 -2.93263316e-01 9.56306681e-02
7.36057222e-01 -6.42934263e-01 1.20721984e+00 -8.10061455e-01
1.19374490e+00 1.10505164e+00 5.17729521e-01 3.28801334e-01
-9.78900015e-01 -7.03074694e-01 -4.40406263e-01 3.96684617e-01
-1.52621627e+00 3.61566208e-02 5.23806334e-01 -3.17240834e-01
1.09168363e+00 2.00153068e-01 6.84435785e-01 7.32919693e-01
4.44505751e-01 1.60878375e-01 1.20365715e+00 -1.02861047e+00
-2.09555954e-01 6.08254075e-01 1.58718526e-01 2.58728147e-01
3.23395759e-01 -2.33025298e-01 -3.25814843e-01 4.07672040e-02
5.86591244e-01 2.72438470e-02 7.25076199e-02 -2.46374041e-01
-1.37715399e+00 9.02329445e-01 4.24762636e-01 1.06306779e+00
-2.77234316e-01 -3.87201667e-01 6.35002792e-01 6.67466879e-01
1.16470695e-01 7.41361141e-01 -2.17679664e-01 8.47075880e-03
-6.66094840e-01 3.27450901e-01 7.81048119e-01 1.00414538e+00
9.60064530e-01 -3.32930923e-01 4.71733570e-01 9.93899107e-01
1.89335003e-01 3.46225202e-01 7.62508810e-01 -4.27319467e-01
4.70325798e-01 9.58941519e-01 -2.34419271e-01 -1.10040963e+00
-3.35721761e-01 -1.21666126e-01 -4.55546588e-01 -2.19007321e-02
7.74561986e-02 3.41199070e-01 -6.35979772e-01 1.75680149e+00
2.81373173e-01 -5.16833842e-01 2.86308199e-01 6.94994986e-01
7.63275802e-01 5.63835382e-01 3.08861792e-01 -8.87738690e-02
1.50170147e+00 -5.54318070e-01 -8.59595835e-01 -1.41367361e-01
8.14037740e-01 -1.31794190e+00 8.30755353e-01 -1.22670367e-01
-9.70833242e-01 -5.66198766e-01 -1.14906669e+00 -9.95812416e-02
-1.07630825e+00 -4.12182629e-01 4.65241224e-01 5.65889776e-01
-1.12121117e+00 3.35803241e-01 -2.41189182e-01 -1.18165886e+00
-4.17990386e-01 2.28476301e-01 -7.37383604e-01 -3.85757893e-01
-1.94120300e+00 1.66497815e+00 1.00549507e+00 -1.99772358e-01
-4.37784903e-02 -2.31771603e-01 -8.39918852e-01 -3.42462331e-01
1.11712411e-01 -4.67808425e-01 5.80049574e-01 -7.44348705e-01
-6.91569805e-01 1.40201330e+00 1.71814099e-01 -4.24564630e-01
2.86235720e-01 1.68910056e-01 -1.09246624e+00 -1.54810622e-01
9.36963737e-01 6.93701386e-01 8.16525593e-02 -9.78291452e-01
-8.84266496e-01 -3.59171152e-01 2.64522463e-01 4.94126797e-01
-4.67379719e-01 5.44328213e-01 -3.63546729e-01 -7.69182205e-01
3.44189882e-01 -8.80410492e-01 -1.42374709e-01 -7.12293327e-01
9.61723551e-02 1.54943513e-02 4.81659025e-01 -1.04959762e+00
1.39980078e+00 -1.81608593e+00 5.69386184e-02 5.40159583e-01
8.56124833e-02 2.20891714e-01 -5.65780066e-02 1.17986894e+00
-1.82455972e-01 6.21707812e-02 -2.20435500e-01 5.68220258e-01
-2.34936491e-01 7.01392174e-01 -1.68852940e-01 2.30081037e-01
-2.62025982e-01 8.04469943e-01 -1.05102074e+00 -6.03910625e-01
4.40972835e-01 2.96155065e-01 -8.87764767e-02 -3.09932441e-01
2.82091051e-01 1.80168450e-01 -1.61409944e-01 3.07780117e-01
4.33626294e-01 4.39568013e-01 7.91658282e-01 -3.71822193e-02
-3.65721464e-01 5.69870293e-01 -1.18230879e+00 1.79466236e+00
-7.59135723e-01 4.00629908e-01 -2.20255926e-01 -9.86229479e-01
1.26157963e+00 6.52412355e-01 7.25047827e-01 -1.06726861e+00
2.64145080e-02 7.16590643e-01 6.93215206e-02 -4.43373591e-01
9.45437729e-01 -3.88654649e-01 -2.85459101e-01 3.70880961e-01
1.64764926e-01 -1.91683978e-01 7.63908803e-01 1.93082586e-01
5.74061751e-01 2.74420232e-01 1.01122153e+00 -7.74512231e-01
7.04461336e-01 5.38962901e-01 5.11390328e-01 1.96066678e-01
2.56741822e-01 2.17835858e-01 -2.58697532e-02 -5.60636640e-01
-1.13999724e+00 -9.58956242e-01 -2.49259681e-01 8.54081094e-01
2.99153715e-01 -6.37050092e-01 -5.76702535e-01 -3.75469536e-01
-3.99054289e-01 8.20740223e-01 -2.62143433e-01 1.55929485e-02
-8.16426098e-01 -6.26279831e-01 6.61230624e-01 2.35671625e-01
3.10241073e-01 -1.32693982e+00 -6.83623195e-01 6.71444595e-01
-4.19897169e-01 -1.08761287e+00 -1.59740970e-01 1.25773460e-01
-5.58823287e-01 -1.05505073e+00 -4.38289762e-01 -1.27974653e+00
4.44591552e-01 4.18210268e-01 1.21903229e+00 1.06687702e-01
-3.41870815e-01 -4.60842736e-02 -7.23009765e-01 -3.51655185e-01
-6.77290976e-01 4.23730344e-01 1.31468788e-01 -6.13555372e-01
8.92575443e-01 -5.32392621e-01 1.13974899e-01 3.32205266e-01
-1.31711102e+00 -5.71933091e-02 2.46584430e-01 7.32417643e-01
2.31709480e-01 1.18458070e-01 5.79213440e-01 -1.05211329e+00
8.71192515e-01 -4.71780181e-01 -3.73954892e-01 5.70126891e-01
-6.15776598e-01 4.15821820e-02 4.14265990e-01 5.60287982e-02
-9.60349679e-01 -2.77629197e-01 -3.72711957e-01 4.78397131e-01
-5.65806627e-02 9.11165833e-01 -4.25003201e-01 -1.25581428e-01
6.34646714e-01 2.73713693e-02 -1.25298887e-01 -4.19404685e-01
5.07544219e-01 7.23767579e-01 3.59708220e-01 -6.28954113e-01
4.19368654e-01 2.77106822e-01 1.24039659e-02 -8.71404350e-01
-3.25525373e-01 -9.75895822e-01 -1.21812820e+00 3.68931727e-03
9.03884113e-01 -9.45628285e-01 5.45221381e-02 -8.81537646e-02
-1.17693841e+00 4.44322735e-01 -2.52663821e-01 6.16726875e-01
-2.13541657e-01 3.23151916e-01 -2.38509312e-01 -1.97718844e-01
-2.15055868e-01 -8.97284746e-01 3.20077151e-01 -7.44710937e-02
-7.51930356e-01 -1.40587413e+00 4.08042580e-01 2.19101474e-01
2.76776820e-01 1.02122389e-01 1.35500968e+00 -8.09748888e-01
1.64554879e-01 -2.45274320e-01 -1.64913565e-01 1.33833244e-01
5.74848413e-01 -2.91212440e-01 -4.42158073e-01 -2.22707629e-01
1.45052850e-01 -1.09329000e-01 2.06217945e-01 -3.14845651e-01
-4.39672954e-02 -2.72958189e-01 -2.27310777e-01 9.27358940e-02
1.91425347e+00 5.15514374e-01 7.41178751e-01 8.80915999e-01
6.46203697e-01 9.14408445e-01 7.20268667e-01 -5.56176864e-02
5.28795898e-01 8.03937614e-01 -1.44670010e-01 -3.00084651e-01
-1.80402800e-01 -2.58126259e-01 1.30989745e-01 1.54351473e+00
-5.80452122e-02 2.61475712e-01 -1.38312864e+00 8.50260854e-01
-1.69047558e+00 -8.82754624e-01 -2.37665713e-01 2.15018582e+00
8.37600410e-01 -4.71194424e-02 1.01543881e-01 1.61111727e-01
8.08290124e-01 -5.02553396e-02 3.35048199e-01 -6.13904297e-01
-3.13287079e-01 5.20642638e-01 5.52031338e-01 8.65906835e-01
-4.90116030e-01 1.41539240e+00 5.78361702e+00 6.82767272e-01
-1.06676722e+00 3.61856878e-01 -3.12965274e-01 4.42202628e-01
-5.31335831e-01 4.44478005e-01 -7.67206073e-01 3.41700166e-01
5.52956343e-01 -4.13035691e-01 4.87430602e-01 4.27020401e-01
-1.20138250e-01 4.53996062e-02 -7.39189863e-01 7.96211004e-01
1.91372156e-01 -1.14039135e+00 4.35665965e-01 -4.48965393e-02
7.49811351e-01 6.98334724e-02 -5.90625763e-01 2.07541212e-02
3.96982580e-01 -7.75235951e-01 6.81202412e-01 1.39585529e-02
9.15885866e-01 -8.61965835e-01 9.55336690e-01 7.13185146e-02
-1.40969110e+00 3.28406721e-01 -6.14162385e-01 -2.83541441e-01
2.29607046e-01 2.53857970e-01 -8.51474583e-01 1.10578525e+00
4.53886807e-01 6.15032971e-01 -4.10325527e-01 7.09440410e-01
-3.79143149e-01 -4.16497104e-02 -7.61146247e-02 -4.44955677e-02
6.17534995e-01 -5.99044859e-01 4.32303697e-01 1.34975314e+00
5.26278436e-01 -3.44576985e-01 3.68715703e-01 3.37364078e-01
1.35350570e-01 8.28913927e-01 -1.03614712e+00 -5.04056998e-02
6.64461017e-01 9.46783960e-01 -8.88571441e-01 -2.82672316e-01
-6.17076039e-01 1.00726759e+00 7.70075396e-02 2.49680623e-01
-5.33956766e-01 -5.99077404e-01 6.63550854e-01 2.88762391e-01
-4.39592868e-01 -2.83797592e-01 -3.57845992e-01 -1.04021657e+00
-7.53018958e-03 -9.61058617e-01 5.74142754e-01 -5.67040682e-01
-1.03098321e+00 8.78674746e-01 2.55365282e-01 -1.26181102e+00
-4.63880956e-01 -6.44473493e-01 -1.64252296e-01 1.20180249e+00
-1.15905857e+00 -1.35153782e+00 4.46709991e-02 6.43284976e-01
4.33214545e-01 -3.83021504e-01 1.16959989e+00 6.90341115e-01
7.82790184e-02 1.61843836e-01 4.09517586e-02 8.93762559e-02
7.32239723e-01 -9.01012838e-01 5.37272394e-01 8.69899273e-01
5.13683498e-01 9.68664587e-01 7.64553607e-01 -1.02584314e+00
-7.86441267e-01 -7.86638975e-01 2.06064796e+00 -2.99656093e-01
1.10014606e+00 -1.71477199e-01 -7.84856260e-01 5.90277910e-01
6.47751391e-01 -7.93611884e-01 8.63237381e-01 -2.44631320e-02
-2.70965695e-01 -8.10901225e-02 -9.72562432e-01 8.17259312e-01
1.03765488e+00 -8.06637347e-01 -1.06844735e+00 3.31503928e-01
4.72237498e-01 9.81120840e-02 -9.10358906e-01 8.83215442e-02
4.67549056e-01 -5.77646315e-01 9.38126743e-01 -6.39421761e-01
2.75739104e-01 -4.49344128e-01 -3.92736346e-01 -1.33573186e+00
-5.38994610e-01 -2.35149667e-01 9.37661469e-01 1.46627295e+00
5.87456524e-01 -7.77293980e-01 1.44598499e-01 3.24819118e-01
1.61982700e-01 -3.37618925e-02 -9.16755736e-01 -9.26568866e-01
2.10029721e-01 -2.39783168e-01 6.78147495e-01 1.52542996e+00
4.28258181e-01 5.50386131e-01 -4.05473500e-01 -4.45132852e-01
3.14190388e-01 -1.53423145e-01 3.97177190e-01 -1.33530629e+00
3.38548869e-01 -4.81502771e-01 -6.78170562e-01 -1.84354514e-01
2.91524559e-01 -1.28502440e+00 -1.74769834e-01 -1.79919553e+00
-4.01026383e-03 -5.28378129e-01 -3.04795027e-01 3.79115015e-01
-4.01155949e-02 4.93621141e-01 3.58876735e-01 5.59580624e-01
1.02802381e-01 1.60924159e-02 8.73531759e-01 -1.98376235e-02
-1.17427289e-01 -4.93115872e-01 -7.95076132e-01 5.38303435e-01
8.68656754e-01 -8.09366465e-01 -4.97569352e-01 -5.22033513e-01
7.84852862e-01 -2.31660664e-01 -1.82249337e-01 -8.92007411e-01
2.48028532e-01 -3.15668792e-01 -1.53670967e-01 -1.98750496e-01
1.31710351e-01 -1.15488052e+00 5.89885890e-01 6.41032338e-01
-2.39590243e-01 7.30912209e-01 -3.11343838e-02 -1.02197967e-01
-7.46424079e-01 -4.97750580e-01 7.60691226e-01 -5.88162720e-01
-1.34179616e+00 -2.11869389e-01 -3.56068492e-01 9.31582302e-02
1.10257864e+00 -5.00217974e-01 1.59508258e-01 8.20524469e-02
-4.22322392e-01 3.39167379e-02 8.31661046e-01 7.99253583e-01
3.31817389e-01 -1.48963189e+00 -7.86138535e-01 1.38073415e-01
4.54833388e-01 -8.43295097e-01 -2.71272302e-01 4.82216656e-01
-9.12587225e-01 6.28922820e-01 -1.05853820e+00 5.19066155e-02
-1.17950487e+00 5.73591471e-01 -1.80386871e-01 -3.00225407e-01
-6.45207465e-01 1.24780610e-01 -8.47023427e-02 -8.94942582e-01
-1.82492375e-01 1.33266896e-01 -7.22087681e-01 3.57319534e-01
3.29351187e-01 3.42071801e-01 2.91027755e-01 -1.34415174e+00
-5.84255874e-01 7.20968008e-01 6.43787012e-02 -5.08333206e-01
1.07963383e+00 -4.28472966e-01 -8.97398412e-01 6.09511793e-01
1.03482306e+00 2.55155742e-01 2.31540918e-01 -3.61405402e-01
6.22859061e-01 -4.82242465e-01 -4.80352461e-01 -6.07424259e-01
-5.64771354e-01 5.85074902e-01 2.30377480e-01 1.69552058e-01
8.07432711e-01 -2.13010490e-01 8.12758446e-01 2.69034535e-01
7.99405873e-01 -1.38586092e+00 -7.77371645e-01 8.50043237e-01
6.54942453e-01 -8.87380362e-01 -1.58245131e-01 -4.37610030e-01
-6.75472915e-01 1.09827662e+00 2.71980137e-01 1.17079884e-01
6.08426869e-01 1.99703693e-01 5.79493642e-01 -2.11503759e-01
8.45674798e-02 -3.32500130e-01 1.67745858e-01 5.78817368e-01
8.59235466e-01 2.72198498e-01 -1.32573438e+00 -1.22334450e-01
-5.38492084e-01 -6.52956367e-02 2.79882014e-01 9.94494975e-01
-2.70647466e-01 -1.71455395e+00 -4.30306762e-01 1.17028013e-01
-5.73032200e-01 -5.18466115e-01 -5.78068972e-01 1.02163053e+00
3.95335555e-01 8.28619838e-01 -1.98513046e-01 -3.31509441e-01
2.70701468e-01 3.22468616e-02 4.44442600e-01 -7.04766870e-01
-8.75286162e-01 -2.65827715e-01 2.47297183e-01 -2.51553833e-01
-6.30279481e-01 -3.25929999e-01 -1.18449557e+00 -4.24850553e-01
-6.15690500e-02 5.93335629e-01 7.59355187e-01 1.19682205e+00
-2.38193572e-01 1.66047007e-01 2.63909191e-01 -1.72090948e-01
-1.25241235e-01 -8.52288485e-01 -6.80267036e-01 7.41232872e-01
-4.93813574e-01 -4.80407834e-01 2.68814266e-01 6.28527552e-02] | [10.495444297790527, 9.65776252746582] |
cc23efb6-f448-438b-ac69-04880947e8c1 | memory-selection-network-for-video | null | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/2319_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123600171.pdf | Memory Selection Network for Video Propagation | Video propagation is a fundamental problem in video processing where guidance frame predictions are propagated to guide predictions of the target frame. Previous research mainly treats the previous adjacent frame as guidance, which, however, could make the propagation vulnerable to occlusion, large motion, and inaccurate information in the previous adjacent frame. To tackle this challenge, we propose a memory selection network, which learns to select suitable guidance from all previous frames for effective and robust propagation. Experimental results on video object segmentation and video colorization tasks show that our method consistently improves performance and can robustly handle challenging scenarios in video propagation. | ['Xiaojuan Qi', 'Jiaya Jia', 'Huaijia Lin', 'Ruizheng Wu'] | null | null | null | null | eccv-2020-8 | ['video-propagation'] | ['computer-vision'] | [ 2.87820011e-01 -3.94546717e-01 -4.83031631e-01 -5.00701964e-01
-4.27123368e-01 -4.05948788e-01 5.80034442e-02 -9.45088342e-02
-5.39115012e-01 6.26816392e-01 2.91499514e-02 -2.01423585e-01
4.80265826e-01 -7.47555435e-01 -7.97586143e-01 -7.15867460e-01
-1.33983299e-01 -1.82782024e-01 1.10661018e+00 2.92253476e-02
2.54799098e-01 3.33542258e-01 -9.26599503e-01 4.58581388e-01
7.39624679e-01 9.78247464e-01 4.74279553e-01 9.38482285e-01
-6.15484677e-02 1.15431166e+00 -5.54715931e-01 -1.22401416e-01
2.87554324e-01 -4.70758349e-01 -8.39697897e-01 2.97654986e-01
7.41193712e-01 -1.04277205e+00 -6.23981178e-01 1.20931876e+00
1.54076651e-01 5.54244876e-01 3.20465118e-01 -1.25210249e+00
-5.53953767e-01 6.97240055e-01 -7.18789935e-01 7.20469952e-01
5.21482587e-01 3.06426913e-01 5.25544822e-01 -7.70623386e-01
8.83416355e-01 1.29741859e+00 4.29251313e-01 7.81626582e-01
-7.58524060e-01 -6.52493834e-01 9.36814547e-01 3.82641971e-01
-1.16868424e+00 -4.23897624e-01 6.06332779e-01 -2.95771658e-01
5.59828818e-01 8.74372646e-02 8.51912081e-01 8.30618501e-01
2.80103415e-01 1.12920809e+00 4.70890999e-01 5.84497722e-03
1.48920208e-01 -2.78410941e-01 7.84712061e-02 8.81123424e-01
-1.15240030e-01 1.70270845e-01 -6.20771706e-01 7.42624104e-02
9.35438573e-01 4.77964543e-02 -7.33560622e-01 7.28991777e-02
-9.98460948e-01 5.85028112e-01 5.80657184e-01 -3.05160075e-01
-3.57315779e-01 3.17040861e-01 8.54053497e-02 1.77912101e-01
4.42286491e-01 -1.61888689e-01 -3.77707809e-01 8.16374924e-03
-1.13206029e+00 3.12783599e-01 5.61857939e-01 1.13022685e+00
7.77651131e-01 3.08856279e-01 -5.01639783e-01 5.64335763e-01
5.76396346e-01 4.30248111e-01 -3.04403584e-02 -1.45760822e+00
6.17962360e-01 3.23396415e-01 1.29073963e-01 -1.16511679e+00
-2.68924803e-01 -2.48423591e-03 -5.70434630e-01 2.91449964e-01
4.69241440e-01 -4.36665595e-01 -1.24263334e+00 1.49494767e+00
4.50141937e-01 7.70592868e-01 -2.75066018e-01 1.44852650e+00
9.37504709e-01 1.10041511e+00 4.45546538e-01 -4.03407425e-01
4.91691053e-01 -1.42415714e+00 -4.97549862e-01 -4.28572953e-01
1.62915379e-01 -8.86330605e-01 5.51100254e-01 1.15667678e-01
-1.32797515e+00 -6.97177231e-01 -9.07871366e-01 7.08182752e-02
3.06296319e-01 -1.43458575e-01 6.59760296e-01 2.04039648e-01
-1.24049020e+00 3.79840940e-01 -9.99332190e-01 -1.90859720e-01
6.66254222e-01 2.61549413e-01 -2.50370689e-02 -4.54180539e-01
-9.61375296e-01 2.92903185e-01 5.47529936e-01 3.85834426e-01
-1.03022110e+00 -6.41433001e-01 -6.69744015e-01 -2.67265916e-01
4.63149339e-01 -8.17798555e-01 1.13773334e+00 -1.27434635e+00
-1.46904469e+00 1.14208117e-01 -4.46801364e-01 -6.30766749e-01
6.46727443e-01 -4.88461316e-01 -2.06974521e-01 4.43079621e-01
-1.52285129e-01 1.17990804e+00 1.21857285e+00 -1.05620849e+00
-1.08622205e+00 -6.62141815e-02 3.54069352e-01 2.95267165e-01
-7.07581490e-02 3.26971970e-02 -1.30696249e+00 -8.72199476e-01
3.06198537e-01 -9.71330166e-01 -6.18429124e-01 2.73451526e-02
-1.79262832e-01 6.00340776e-02 1.09165275e+00 -5.87778151e-01
1.26118076e+00 -1.96654487e+00 2.26488620e-01 2.36471623e-01
2.42390469e-01 1.91612750e-01 -2.37539113e-01 -2.64348090e-01
5.03542125e-01 1.68164410e-02 -2.44412497e-02 -4.08978090e-02
-4.62512344e-01 1.63728490e-01 -3.13395798e-01 3.58923376e-01
2.25826785e-01 8.05031240e-01 -9.43310976e-01 -5.72673321e-01
1.12827435e-01 7.20215380e-01 -7.43537188e-01 4.67400879e-01
-4.92745548e-01 6.38986886e-01 -6.71286404e-01 8.55324090e-01
6.91753328e-01 -1.55677661e-01 -1.63268328e-01 -2.04960495e-01
9.05853435e-02 -1.96766898e-01 -1.06272805e+00 1.56739688e+00
1.69472188e-01 9.17774260e-01 1.56533733e-01 -5.01846552e-01
6.33848548e-01 3.81078795e-02 4.93654370e-01 -5.81217110e-01
1.26627699e-01 -1.64095998e-01 -1.36336982e-01 -5.25400877e-01
7.17974365e-01 4.83752191e-01 5.60990155e-01 3.06688905e-01
-2.89696574e-01 1.39835462e-01 4.78513420e-01 4.63972896e-01
8.65380585e-01 4.76777196e-01 -3.80533636e-01 1.72612503e-01
5.07810235e-01 -3.23700197e-02 1.19891870e+00 8.30780327e-01
-5.65538645e-01 9.34063137e-01 1.35567218e-01 -6.43033564e-01
-6.04599833e-01 -8.43531013e-01 3.56751859e-01 1.49832833e+00
7.74562895e-01 -4.87957031e-01 -9.46377873e-01 -9.10683930e-01
-3.80242199e-01 3.90860736e-01 -2.32976541e-01 -8.01590756e-02
-1.02118647e+00 -8.42967391e-01 4.48306184e-03 8.90046239e-01
5.50567567e-01 -9.52833772e-01 -6.17546618e-01 5.16349256e-01
-3.36917520e-01 -1.36820114e+00 -7.94014513e-01 -3.52603793e-01
-1.11415482e+00 -1.23557854e+00 -9.00252163e-01 -7.42955506e-01
8.99643600e-01 5.72968483e-01 1.17706597e+00 6.91936791e-01
4.74299379e-02 3.93148690e-01 -5.16941249e-01 1.42315015e-01
-3.34537148e-01 -2.47685060e-01 -2.67506748e-01 4.45899833e-03
3.19201380e-01 -1.12626143e-02 -9.45332348e-01 4.65466887e-01
-8.46675515e-01 9.55516919e-02 2.66035169e-01 5.18443406e-01
8.04695964e-01 1.78068995e-01 1.54788047e-01 -1.06542313e+00
2.98344404e-01 -3.41964930e-01 -7.65011609e-01 4.42160130e-01
-2.96723485e-01 -3.77770573e-01 3.51830453e-01 -3.41810852e-01
-1.19559824e+00 2.28496715e-01 -1.67959094e-01 -3.90484273e-01
-1.67533815e-01 2.40966737e-01 4.91686836e-02 -5.29779017e-01
3.26053739e-01 1.71226561e-01 -5.05627811e-01 -1.11402281e-01
3.26321065e-01 -3.32825854e-02 5.81632674e-01 -3.40947300e-01
5.21259010e-01 4.63222086e-01 -2.17859432e-01 -6.61524475e-01
-5.54395378e-01 -2.01671794e-01 -6.95403099e-01 -6.50653064e-01
8.38774979e-01 -9.69975829e-01 -2.73748368e-01 4.90373224e-01
-1.40895581e+00 -5.56835353e-01 3.64114016e-01 4.83331949e-01
-2.40863964e-01 4.15571600e-01 -1.04720902e+00 -5.33188045e-01
-2.87480980e-01 -1.61541045e+00 7.80945480e-01 7.84095466e-01
-1.66053802e-01 -1.15178871e+00 -3.34997982e-01 1.96428150e-01
3.70474577e-01 9.96186212e-02 4.86059606e-01 5.26447110e-02
-1.39375126e+00 -2.67684199e-02 -2.50711024e-01 2.03517690e-01
-1.02588862e-01 6.23385608e-01 -5.97541809e-01 -3.43348384e-01
-2.71457821e-01 3.67965214e-02 1.34876859e+00 7.27395713e-01
1.08207881e+00 -2.10636988e-01 -4.68875974e-01 9.58611369e-01
1.15745318e+00 4.76881057e-01 4.90734160e-01 2.44834960e-01
1.01289177e+00 1.33577049e-01 9.45232213e-01 3.18906277e-01
4.59588110e-01 4.37476277e-01 3.26129675e-01 -7.64249116e-02
-4.40830976e-01 -1.32580161e-01 5.82259834e-01 5.46622157e-01
-2.74064690e-01 -5.45413733e-01 -7.38589704e-01 8.28020200e-02
-2.03823209e+00 -8.61989081e-01 -1.46838114e-01 1.85029018e+00
5.89261949e-01 3.31042469e-01 7.22339302e-02 -3.70384216e-01
6.46598876e-01 2.76591420e-01 -5.67702413e-01 1.20398104e-01
-2.56528229e-01 -4.02262330e-01 5.47093868e-01 8.63680482e-01
-1.34648788e+00 1.30581617e+00 7.45062494e+00 4.52419192e-01
-1.33722341e+00 3.70627828e-02 9.91791546e-01 -2.15399355e-01
-3.21135491e-01 -9.25997365e-03 -1.03632832e+00 5.19276738e-01
2.85863101e-01 1.49210557e-01 2.87192553e-01 6.56572342e-01
4.44424689e-01 -5.08053243e-01 -9.76676881e-01 9.18057442e-01
4.10604626e-02 -1.46005368e+00 3.52494717e-01 -5.65769732e-01
1.14636087e+00 8.60987902e-02 2.23166838e-01 -6.50481507e-02
1.79057553e-01 -6.98437214e-01 7.30374217e-01 4.95306164e-01
1.67002976e-01 -6.81562364e-01 4.66180325e-01 1.08092248e-01
-1.23905265e+00 -3.91493440e-02 -5.52958786e-01 9.24363658e-02
5.66240191e-01 3.13456059e-01 -6.57596350e-01 3.74590307e-02
8.42950881e-01 9.87625718e-01 -6.40542150e-01 1.47106969e+00
-4.54232424e-01 9.02838230e-01 -2.20750481e-01 1.84750423e-01
4.50460255e-01 -2.60382175e-01 6.68241203e-01 1.34908223e+00
1.79375052e-01 3.72826099e-01 7.42307067e-01 5.82868934e-01
-3.53094526e-02 -5.58250211e-02 -1.47861943e-01 2.37324148e-01
3.31285536e-01 8.76983523e-01 -1.19180763e+00 -3.26659769e-01
-5.57914853e-01 1.18902004e+00 1.83909267e-01 1.17099988e+00
-1.00972211e+00 1.93423346e-01 4.86764282e-01 7.41843507e-02
3.38547438e-01 -4.66558844e-01 -1.36553673e-02 -9.95572209e-01
-3.62200029e-02 -4.28432167e-01 5.23728907e-01 -6.38978958e-01
-8.86005759e-01 8.50295961e-01 -2.28964031e-01 -1.03520930e+00
-2.52745599e-01 -3.87192547e-01 -7.38891125e-01 5.03180027e-01
-1.62495327e+00 -9.21757638e-01 -3.39903861e-01 7.74037540e-01
1.01999724e+00 3.39742526e-02 1.70816377e-01 3.85447830e-01
-8.09265554e-01 4.71973419e-01 -2.00362369e-01 1.64548695e-01
6.82374895e-01 -9.27009642e-01 2.57101506e-01 1.36729622e+00
1.70353726e-01 3.99666220e-01 6.26341581e-01 -9.10306633e-01
-1.21916080e+00 -1.39497459e+00 2.14940950e-01 -1.66010410e-01
3.56463611e-01 1.24747783e-01 -1.04527152e+00 5.92735946e-01
1.78717390e-01 1.75133944e-01 3.90576303e-01 -3.09927791e-01
-1.28193691e-01 -5.19021265e-02 -6.80677652e-01 7.63690650e-01
1.03397942e+00 -2.40071744e-01 -2.82625742e-02 2.84067094e-01
8.01972866e-01 -9.24675584e-01 -3.44121784e-01 3.77108008e-01
4.94890779e-01 -9.54828918e-01 9.90467191e-01 -4.79708225e-01
3.59143049e-01 -6.34709954e-01 2.32420396e-02 -1.06323969e+00
-2.39790499e-01 -7.97596574e-01 -5.21348596e-01 1.08131850e+00
5.25038183e-01 1.51898146e-01 1.02280164e+00 1.21674109e+00
7.86027759e-02 -6.85004532e-01 -4.79401946e-01 -2.07496092e-01
-3.15210193e-01 -6.67871833e-01 3.31772447e-01 4.68446463e-01
-4.44696605e-01 -1.05920896e-01 -6.13298893e-01 3.54796916e-01
6.85721517e-01 2.13315025e-01 6.40757442e-01 -6.01816654e-01
-2.29189977e-01 -4.85129923e-01 -3.70463729e-01 -1.79666293e+00
1.68515846e-01 -4.02310759e-01 3.96018445e-01 -1.58128488e+00
1.51913702e-01 -5.00709295e-01 -4.02528018e-01 1.67897284e-01
-7.53540158e-01 6.14350915e-01 4.84805673e-01 1.57048389e-01
-1.18064761e+00 2.54770130e-01 1.39059556e+00 -4.16009158e-01
-4.62100178e-01 2.54039317e-01 -3.59394044e-01 1.08515275e+00
6.10920131e-01 -4.26347852e-01 -6.04382753e-01 -9.14721251e-01
7.03092441e-02 4.04213890e-02 3.00663114e-01 -8.72208118e-01
7.14325607e-01 -5.07199883e-01 7.59211361e-01 -6.31899536e-01
2.09125355e-01 -7.79010236e-01 -1.05232529e-01 1.81834623e-01
-2.74570078e-01 1.12668507e-01 5.42008244e-02 8.23878407e-01
-3.94057393e-01 -4.46609072e-02 6.94150507e-01 -2.93242335e-02
-1.58938968e+00 9.08519268e-01 -5.36379933e-01 7.31438398e-02
1.03348625e+00 -3.73196095e-01 -1.52935654e-01 -6.18153453e-01
-7.48513401e-01 4.93570745e-01 5.09384036e-01 6.02133870e-01
1.19934464e+00 -1.23448086e+00 -8.12826276e-01 1.81684285e-01
-4.18192863e-01 2.09721610e-01 3.02689791e-01 7.72097588e-01
-8.29905450e-01 -4.27691191e-02 -3.25716175e-02 -8.50996614e-01
-1.37487161e+00 3.86579067e-01 2.72895545e-01 6.22601211e-01
-7.21844137e-01 1.47278297e+00 2.94543892e-01 5.89819193e-01
5.98288894e-01 -4.00926381e-01 -3.15679163e-01 -2.32367903e-01
1.01129758e+00 3.91336262e-01 -4.13135052e-01 -8.45052063e-01
-1.80107072e-01 7.12355316e-01 -3.08651000e-01 1.06856033e-01
9.82856870e-01 -6.67060554e-01 -1.28894046e-01 2.49302790e-01
9.74138141e-01 -2.38520533e-01 -2.10052109e+00 -1.91624403e-01
-2.25319892e-01 -9.40849900e-01 2.55671501e-01 -4.96521413e-01
-1.78558385e+00 6.49667621e-01 4.88184452e-01 -1.08852305e-01
1.27830124e+00 -2.82058269e-01 1.14908385e+00 2.32979536e-01
4.14756656e-01 -1.09530783e+00 -3.43288928e-02 5.89815378e-01
5.93916297e-01 -1.32684171e+00 -6.43889904e-02 -7.35561907e-01
-6.53878152e-01 1.43893218e+00 1.13504493e+00 1.18815769e-02
7.60724664e-01 4.22080219e-01 5.10860682e-01 2.29814321e-01
-7.32026696e-01 7.07011744e-02 4.33503866e-01 5.61867833e-01
5.28510988e-01 -2.78779417e-01 4.67491969e-02 1.06835380e-01
3.90830219e-01 5.62224574e-02 4.26783502e-01 8.89688730e-01
-6.54609084e-01 -8.07160497e-01 -4.70408559e-01 2.33580843e-01
-5.96851945e-01 -1.22761518e-01 -2.66038895e-01 3.18287462e-01
9.22047943e-02 1.13671553e+00 2.32920027e-03 -3.14685166e-01
-2.63123661e-02 -4.33386326e-01 5.05642235e-01 -2.74995685e-01
-4.04799491e-01 4.94195908e-01 -3.04760218e-01 -1.09444094e+00
-8.39637876e-01 -3.95455152e-01 -1.64956427e+00 -3.63630891e-01
-2.98642725e-01 -5.54455481e-02 6.29142374e-02 9.25080240e-01
2.26610214e-01 5.63749909e-01 2.82328993e-01 -1.00984704e+00
2.09620014e-01 -4.08754438e-01 -3.38140070e-01 5.64117014e-01
5.71541071e-01 -5.06406724e-01 -1.13551028e-03 5.03092408e-01] | [9.192179679870605, -0.14194990694522858] |
b14d38a9-2fbe-4a15-969c-e3359b8e0b1b | findings-of-the-shared-task-on-multilingual | 2209.07841 | null | https://arxiv.org/abs/2209.07841v1 | https://arxiv.org/pdf/2209.07841v1.pdf | Findings of the Shared Task on Multilingual Coreference Resolution | This paper presents an overview of the shared task on multilingual coreference resolution associated with the CRAC 2022 workshop. Shared task participants were supposed to develop trainable systems capable of identifying mentions and clustering them according to identity coreference. The public edition of CorefUD 1.0, which contains 13 datasets for 10 languages, was used as the source of training and evaluation data. The CoNLL score used in previous coreference-oriented shared tasks was used as the main evaluation metric. There were 8 coreference prediction systems submitted by 5 participating teams; in addition, there was a competitive Transformer-based baseline system provided by the organizers at the beginning of the shared task. The winner system outperformed the baseline by 12 percentage points (in terms of the CoNLL scores averaged across all datasets for individual languages). | ['YIlun Zhu', 'Daniel Zeman', 'Jakub Sido', 'Ondřej Pražák', 'Martin Popel', 'Maciej Ogrodniczuk', 'Michal Novák', 'Anna Nedoluzhko', 'Miloslav Konopík', 'Zdeněk Žabokrtský'] | 2022-09-16 | null | https://aclanthology.org/2022.crac-mcr.1 | https://aclanthology.org/2022.crac-mcr.1.pdf | crac-acl-2022-10 | ['coreference-resolution'] | ['natural-language-processing'] | [-1.65208936e-01 3.06502610e-01 -3.77875388e-01 -4.21815425e-01
-1.57320273e+00 -9.30987716e-01 9.09166515e-01 2.36383274e-01
-6.40417099e-01 1.01000774e+00 7.36670434e-01 -1.97833017e-01
-6.76831678e-02 -3.25999171e-01 -4.93707120e-01 -4.05247748e-01
2.19692122e-02 1.34800148e+00 3.53054971e-01 -6.61294341e-01
2.18067378e-01 2.78084129e-01 -1.21185148e+00 8.00827324e-01
9.79989946e-01 2.85554588e-01 5.38606942e-01 3.34817588e-01
-1.05841890e-01 5.99973679e-01 -6.28101766e-01 -7.53266335e-01
-1.16430610e-01 -2.97117084e-01 -1.56480503e+00 -9.30162370e-01
7.89862692e-01 4.68343079e-01 -2.00930133e-01 1.23736274e+00
5.72223485e-01 2.32655138e-01 1.24617614e-01 -1.09935498e+00
3.32248426e-04 1.49421191e+00 -1.54003367e-01 2.72484511e-01
5.83315432e-01 -6.23016000e-01 1.58908796e+00 -7.71534562e-01
1.19292760e+00 1.34267986e+00 6.18510723e-01 5.85740149e-01
-1.17784607e+00 -8.86654437e-01 -1.20922774e-01 5.41047990e-01
-1.35709023e+00 -6.91870987e-01 4.00082290e-01 -2.71699160e-01
1.38844728e+00 4.25623089e-01 2.55520139e-02 1.04612195e+00
-2.40534529e-01 6.77630603e-01 9.18590605e-01 -7.94877946e-01
-4.24245238e-01 3.11119437e-01 3.32774639e-01 2.70127177e-01
1.06198452e-01 1.59163579e-01 -5.17740607e-01 -2.57736713e-01
2.19686911e-01 -9.54613388e-01 -6.10340297e-01 -4.07449037e-01
-1.29930091e+00 7.98733354e-01 3.96335065e-01 1.03242278e+00
5.38948327e-02 -5.79866350e-01 7.07687855e-01 5.43926835e-01
2.98618138e-01 8.00651789e-01 -7.18919694e-01 -7.74236843e-02
-8.60222280e-01 3.96659702e-01 9.70298886e-01 9.65146899e-01
4.63700533e-01 -5.86835206e-01 -3.21837552e-02 1.09379101e+00
1.70020517e-02 3.92335176e-01 3.34598690e-01 -1.23338544e+00
9.73914802e-01 4.85216796e-01 2.10660964e-01 -7.30284452e-01
-5.80722451e-01 -4.53409255e-01 -4.47326213e-01 -2.40266681e-01
5.82666755e-01 -2.45174572e-01 -2.22905412e-01 1.87682045e+00
-6.15919083e-02 -1.08124077e-01 4.60121006e-01 8.01913857e-01
1.07293320e+00 3.49689007e-01 2.28772755e-03 -3.77520829e-01
1.39268315e+00 -9.76322472e-01 -8.10596049e-01 1.43463343e-01
1.05670738e+00 -1.21809232e+00 4.60443556e-01 1.61909550e-01
-1.03368211e+00 -5.34923732e-01 -9.32393610e-01 -1.85114928e-02
-2.19814077e-01 -1.35972634e-01 4.03352827e-01 2.50041693e-01
-1.07047498e+00 4.79732305e-01 -4.91791457e-01 -6.53986990e-01
-5.06103635e-01 4.02488261e-01 -7.06625760e-01 3.57115380e-02
-1.68555450e+00 1.42259407e+00 6.20163977e-01 -2.89029121e-01
-5.39564371e-01 -7.83498049e-01 -7.30592906e-01 -9.03390571e-02
1.54974446e-01 -2.16494009e-01 1.48056006e+00 -8.47176254e-01
-1.12403500e+00 1.52381563e+00 -1.49882451e-01 -6.90443695e-01
3.84011149e-01 -4.72625107e-01 -9.36422586e-01 -1.51925832e-01
5.68825662e-01 6.59788668e-01 -2.49572337e-01 -1.04511881e+00
-1.18942571e+00 -1.60881132e-01 9.94933099e-02 4.76977408e-01
3.23682457e-01 6.60667658e-01 -3.99732858e-01 -3.34586263e-01
-9.20160413e-02 -1.00687265e+00 6.86962754e-02 -1.29561937e+00
-2.54900366e-01 -6.32120132e-01 6.58607185e-01 -7.42935717e-01
1.04357553e+00 -2.05920029e+00 3.63843024e-01 -8.02545175e-02
-2.62713552e-01 3.54594797e-01 -2.06605062e-01 5.89072883e-01
-5.20228565e-01 -1.66059345e-01 1.31496549e-01 -3.67242455e-01
2.45736004e-03 -4.89177182e-02 -4.34209585e-01 2.23571092e-01
-3.12064976e-01 3.52372527e-01 -1.17880881e+00 -5.68745494e-01
1.80010907e-02 2.50691533e-01 -3.80862981e-01 1.53955236e-01
-6.74081296e-02 5.12946188e-01 1.07513638e-02 1.69882979e-02
4.59168762e-01 3.58978987e-01 8.58012974e-01 -4.20711130e-01
-6.05604827e-01 9.98798013e-01 -9.77534652e-01 2.05605268e+00
-4.69790608e-01 6.51848793e-01 2.20520750e-01 -8.51433933e-01
1.13374352e+00 9.56900418e-01 3.03017110e-01 -7.98317909e-01
-2.37978935e-01 7.14847505e-01 2.95741022e-01 7.05840662e-02
8.35636139e-01 1.73819512e-01 -5.97912073e-01 4.56016779e-01
4.64687496e-01 4.80188727e-02 2.86559880e-01 4.44309533e-01
6.76371872e-01 9.98420045e-02 1.26528651e-01 -8.04528534e-01
9.89333451e-01 5.60854793e-01 1.01141751e+00 3.43246520e-01
-1.30311236e-01 4.29441899e-01 2.57678568e-01 -4.47526664e-01
-8.45152557e-01 -7.26468086e-01 -3.11705589e-01 1.26496327e+00
6.63914857e-03 -6.94334090e-01 -7.11957753e-01 -7.49445438e-01
-5.57390034e-01 9.06313419e-01 -2.37640634e-01 3.31203938e-01
-1.15842938e+00 -3.50113720e-01 9.37054574e-01 2.52271861e-01
3.14690381e-01 -1.03985190e+00 -1.40698776e-01 3.18362147e-01
-1.14543021e+00 -1.29946995e+00 -7.32636690e-01 3.40693504e-01
-4.09925073e-01 -1.47695172e+00 -4.46018279e-01 -1.29393089e+00
6.94825649e-02 2.90545784e-02 1.31411397e+00 -1.66149095e-01
1.56814769e-01 -5.94458543e-02 -3.33172560e-01 -1.27749026e-01
-4.64340717e-01 5.99787652e-01 4.36167754e-02 -4.66254324e-01
7.00074613e-01 -1.81176066e-01 7.56079257e-02 2.14549646e-01
3.07533704e-03 -1.36109173e-01 4.91844304e-03 8.86110067e-01
2.80675232e-01 -3.95061284e-01 2.62730569e-01 -9.44255531e-01
5.75825870e-01 -1.44810289e-01 -6.92685544e-01 5.07364810e-01
-2.90398777e-01 1.50161594e-01 9.78917331e-02 1.45563662e-01
-1.50806522e+00 -1.46464929e-01 -7.86793008e-02 7.79676065e-02
-1.53020427e-01 3.19363862e-01 -4.63318408e-01 1.30023539e-01
4.31077331e-01 -3.24099064e-01 -4.65076208e-01 -9.03553486e-01
3.14785033e-01 8.10816228e-01 1.17907381e+00 -9.82518375e-01
2.23308772e-01 -2.97487557e-01 -4.20714587e-01 -5.08920789e-01
-8.65862191e-01 -7.83799767e-01 -1.06372726e+00 1.86059758e-01
7.13860154e-01 -1.19948113e+00 -4.40036982e-01 3.15785706e-01
-1.55213523e+00 -2.51369596e-01 1.11773998e-01 7.60090172e-01
-5.48895895e-01 1.91741809e-01 -6.81731284e-01 -1.17362864e-01
-7.28835523e-01 -1.20719779e+00 5.20438194e-01 -4.37728278e-02
-8.33541572e-01 -1.20188546e+00 7.49543548e-01 6.14509940e-01
9.32689980e-02 7.73964226e-02 9.89226818e-01 -1.22808540e+00
1.58803403e-01 1.74266100e-01 -5.94888292e-02 -1.04689948e-01
-1.31572932e-01 -1.09515153e-01 -8.00015271e-01 -4.30510104e-01
-4.04224247e-01 -2.10572973e-01 6.03787720e-01 6.18392564e-02
-3.43299173e-02 8.46104622e-02 -7.65753984e-01 3.99899364e-01
1.30027473e+00 2.36572474e-01 3.96267653e-01 8.14824343e-01
3.49680543e-01 7.55652249e-01 7.47671008e-01 -2.62560219e-01
5.72535157e-01 1.28990483e+00 -7.98943266e-03 1.89622924e-01
-2.97106087e-01 -1.76891863e-01 3.28689128e-01 1.17152500e+00
-3.32927316e-01 1.34345695e-01 -1.23796678e+00 8.71846914e-01
-1.92527139e+00 -1.08699632e+00 -6.03066146e-01 2.24347949e+00
1.00541151e+00 6.51374012e-02 8.75022784e-02 -8.47336501e-02
1.20415843e+00 -1.61295161e-01 3.82474512e-01 -6.66457117e-01
-5.08905649e-01 3.82726967e-01 2.76250780e-01 1.06219816e+00
-1.18823028e+00 1.32135391e+00 6.64016581e+00 2.69827425e-01
-8.39810789e-01 2.96850890e-01 -2.98651218e-01 1.04125611e-01
-1.10844731e-01 3.27511579e-01 -1.21299279e+00 8.54891688e-02
1.26597226e+00 -5.37388861e-01 7.08223656e-02 3.34250540e-01
-3.39000732e-01 9.03134122e-02 -1.10046208e+00 4.71985132e-01
7.84591809e-02 -1.16383553e+00 -1.48606881e-01 -1.64929494e-01
7.15081334e-01 6.85626209e-01 -3.69596273e-01 4.63921547e-01
1.03112423e+00 -7.67144382e-01 5.32728434e-01 2.17575029e-01
7.55977035e-01 -9.67256308e-01 1.22081172e+00 1.67214766e-01
-1.21284389e+00 3.32756817e-01 -1.23326637e-01 2.15941578e-01
3.34774554e-01 3.05226520e-02 -6.15654528e-01 1.18084753e+00
7.57716835e-01 5.51388979e-01 -1.40940517e-01 1.05005765e+00
-3.71179879e-01 2.16034070e-01 -1.15437649e-01 5.53244054e-01
3.79168838e-01 1.08974837e-01 7.84502506e-01 1.69699121e+00
1.35632575e-01 -1.14962913e-01 1.47848532e-01 1.96832925e-01
-1.13167495e-01 2.96304911e-01 -2.51523495e-01 4.59489346e-01
1.11547244e+00 1.18171918e+00 -1.23866670e-01 -1.81905985e-01
-4.20258492e-01 6.57191694e-01 5.30697346e-01 -1.71905253e-02
-4.52060908e-01 -8.03688109e-01 6.92684174e-01 -3.02312791e-01
1.95879698e-01 2.12044671e-01 3.32348734e-01 -9.84039783e-01
-4.93041635e-01 -1.19226480e+00 9.97350991e-01 -3.95481080e-01
-9.43815649e-01 1.01892471e+00 1.67847186e-01 -7.73016214e-01
-5.22582889e-01 -3.25606972e-01 -6.54463232e-01 1.22475898e+00
-1.12163281e+00 -1.16667163e+00 1.52252585e-01 8.37201536e-01
3.61174077e-01 -5.45970082e-01 1.40608299e+00 4.97130424e-01
-4.37318623e-01 8.30296338e-01 1.26585990e-04 5.72632015e-01
1.52681577e+00 -1.29928553e+00 3.88130456e-01 6.47353649e-01
3.50489676e-01 7.12175965e-01 1.01039660e+00 -3.77121836e-01
-7.15619326e-01 -8.18588912e-01 2.07412195e+00 -4.61509854e-01
8.97870064e-01 -2.05742326e-02 -7.85695910e-01 1.17287934e+00
8.39681029e-01 -4.46273267e-01 6.50190413e-01 7.31540799e-01
-5.67407489e-01 4.86643380e-03 -1.01827800e+00 1.38747662e-01
9.63075936e-01 -6.47574902e-01 -1.38721144e+00 1.58414081e-01
6.43756092e-01 -6.63792193e-01 -1.22921348e+00 5.14466584e-01
2.11525142e-01 -6.97527766e-01 6.16228819e-01 -6.30870938e-01
-1.53901204e-01 -4.44741308e-04 -4.93612766e-01 -1.48638809e+00
-5.61468244e-01 -6.58007324e-01 4.91290301e-01 1.73005629e+00
7.10350633e-01 -4.39180434e-01 2.44549304e-01 7.20215822e-03
-3.73895943e-01 2.00673312e-01 -1.06783068e+00 -6.34789228e-01
4.93437618e-01 -1.44531563e-01 5.84446430e-01 1.31700599e+00
7.28400111e-01 9.15828407e-01 2.86744796e-02 1.08551025e-01
8.52981091e-01 2.76047617e-01 6.35924280e-01 -1.50309491e+00
8.82618129e-02 -5.69242418e-01 -1.09766953e-01 -4.38280016e-01
6.04486585e-01 -1.33946550e+00 -4.28022183e-02 -1.18173528e+00
2.69794613e-01 -6.10467911e-01 -4.33084041e-01 5.61369181e-01
1.00654075e-02 2.04098318e-02 4.27041978e-01 5.26158929e-01
-5.67573130e-01 -1.59681827e-01 6.26272500e-01 -2.28133559e-01
-1.97055377e-03 -6.68462515e-02 -7.28563607e-01 5.62242329e-01
6.94787860e-01 -5.13749421e-01 4.91322614e-02 -5.77742517e-01
-7.17825145e-02 1.30644411e-01 -1.50233001e-01 -9.02707100e-01
5.28557122e-01 1.49201468e-01 -2.98556149e-01 -8.89066041e-01
1.60124734e-01 -4.40770119e-01 3.70422989e-01 5.51238716e-01
-4.35024410e-01 1.99679777e-01 3.59523118e-01 -3.52685630e-01
-7.25082397e-01 -1.68274462e-01 6.38834059e-01 -1.09536484e-01
-9.87753689e-01 -3.04208457e-01 -1.15449133e-03 5.31707644e-01
7.74641395e-01 4.05520082e-01 -5.48338830e-01 7.45087266e-02
-8.48753512e-01 5.13580680e-01 2.41038889e-01 7.91447103e-01
-1.98925331e-01 -1.29729033e+00 -1.25240409e+00 -2.56172150e-01
2.35435754e-01 -5.76752961e-01 8.59379992e-02 8.86489213e-01
-1.60113558e-01 1.30372751e+00 -4.72374380e-01 -3.83170933e-01
-1.83552396e+00 2.74556428e-01 6.70070052e-01 -7.66298115e-01
-4.99512464e-01 7.28590846e-01 -1.15210503e-01 -1.01368737e+00
3.63622874e-01 3.84887874e-01 -5.06592274e-01 2.76901990e-01
7.05293477e-01 3.09939086e-01 4.42018807e-01 -1.14148247e+00
-8.06626379e-01 4.64330614e-01 -4.43484098e-01 -3.82585913e-01
1.38181221e+00 -5.90610504e-02 -3.21312487e-01 3.93513113e-01
1.08440399e+00 6.72301129e-02 -3.28205228e-01 -6.98390663e-01
6.93075716e-01 1.87514007e-01 -3.58958393e-02 -1.08588099e+00
-8.92271459e-01 6.89790726e-01 3.24074924e-01 -2.90994704e-01
5.30231476e-01 2.60370493e-01 5.07222176e-01 2.44296864e-01
7.94785202e-01 -1.07272184e+00 -7.25223839e-01 1.13460815e+00
9.64845777e-01 -9.09595251e-01 -3.93862784e-01 -3.19090843e-01
-6.66662276e-01 9.77344275e-01 5.76695502e-01 1.80423662e-01
2.40592092e-01 4.15589750e-01 3.70795935e-01 -1.21589795e-01
-8.65430892e-01 -5.76216877e-02 2.39829063e-01 7.25491643e-01
1.06301641e+00 3.59259814e-01 -7.23722160e-01 7.03140378e-01
-4.99322593e-01 -3.61167163e-01 1.28418073e-01 4.52744722e-01
-4.56009150e-01 -1.71351266e+00 -3.68433595e-01 -4.17376131e-01
-7.61650980e-01 -2.46131688e-01 -5.51250458e-01 1.02159214e+00
1.92420781e-01 9.94898081e-01 3.12735111e-01 -3.80434155e-01
5.54726720e-01 7.92675242e-02 6.54876053e-01 -6.11983776e-01
-9.39793587e-01 -2.21097976e-01 8.86173248e-01 -4.48151916e-01
-5.42895555e-01 -1.09986174e+00 -1.24654603e+00 -3.91556293e-01
-2.42731661e-01 1.16983163e+00 1.97598457e-01 9.33148980e-01
2.91084051e-02 8.43338519e-02 4.17127311e-01 -4.63420808e-01
-2.48062342e-01 -1.27736127e+00 -3.02169323e-01 4.63508457e-01
-5.85454144e-02 -5.64988911e-01 -2.40187109e-01 -1.28493890e-01] | [9.300512313842773, 9.591814041137695] |
edf68c79-4018-4854-a092-6c172b0fe51d | using-semantic-role-labeling-to-improve | null | null | https://aclanthology.org/2022.lrec-1.329 | https://aclanthology.org/2022.lrec-1.329.pdf | Using Semantic Role Labeling to Improve Neural Machine Translation | Despite impressive progress in machine translation in recent years, it has occasionally been argued that current systems are still mainly based on pattern recognition and that further progress may be possible by using text understanding techniques, thereby e.g. looking at semantics of the type “Who is doing what to whom?”. In the current research we aim to take a small step into this direction. Assuming that semantic role labeling (SRL) grasps some of the relevant semantics, we automatically annotate the source language side of a standard parallel corpus, namely Europarl, with semantic roles. We then train a neural machine translation (NMT) system using the annotated corpus on the source language side, and the original unannotated corpus on the target language side. New text to be translated is first annotated by the same SRL system and then fed into the translation system. We compare the results to those of a baseline NMT system trained with unannotated text on both sides and find that the SRL-based system yields small improvements in terms of BLEU scores for each of the four language pairs under investigation, involving English, French, German, Greek and Spanish. | ['Reinhard Rapp'] | null | null | null | null | lrec-2022-6 | ['semantic-role-labeling'] | ['natural-language-processing'] | [ 6.60931945e-01 4.64027107e-01 -3.22764754e-01 -5.05235136e-01
-8.48893821e-01 -1.02798486e+00 1.02841437e+00 2.62697577e-01
-5.18756807e-01 1.21705234e+00 5.51096439e-01 -7.49509752e-01
1.84202299e-01 -7.11180985e-01 -6.53292298e-01 -4.58468229e-01
6.65582776e-01 1.18255460e+00 1.97903350e-01 -7.49772727e-01
1.55195221e-01 2.19206717e-02 -1.15119636e+00 5.01615584e-01
8.50236714e-01 5.36698043e-01 1.73893765e-01 3.67962956e-01
-4.42859560e-01 7.48250782e-01 -4.86494452e-01 -6.63898349e-01
4.29229021e-01 -7.75935471e-01 -1.32275653e+00 8.18219222e-03
7.42730945e-02 3.25835735e-01 1.11211181e-01 9.72848833e-01
1.70240909e-01 -9.90883708e-02 4.07571852e-01 -8.97711039e-01
-6.34722769e-01 8.13398778e-01 -1.83005691e-01 8.87202471e-02
6.00656092e-01 -1.78006798e-01 1.33522379e+00 -7.62332976e-01
9.49111879e-01 1.32941794e+00 2.38551229e-01 8.29000652e-01
-1.29071307e+00 -3.08618575e-01 -7.17304200e-02 8.62914100e-02
-7.79613972e-01 -3.68452489e-01 4.98545349e-01 -9.53919739e-02
1.21679354e+00 1.95809886e-01 3.06710273e-01 1.00842738e+00
3.81628200e-02 7.41841674e-01 1.37481248e+00 -1.05630231e+00
9.24161524e-02 4.06070381e-01 6.68889331e-03 3.22953790e-01
4.69295420e-02 -2.10711658e-01 -5.54685116e-01 -1.82925418e-01
5.85540831e-01 -6.39597774e-01 -1.55247703e-01 -2.06107587e-01
-1.81384778e+00 7.28558660e-01 -6.12531640e-02 8.08105171e-01
-3.33146006e-01 -1.05255254e-01 7.39548326e-01 1.03088355e+00
7.21529186e-01 8.90995860e-01 -1.03950977e+00 -4.02309388e-01
-7.79856622e-01 8.30572248e-02 1.23576856e+00 7.99421847e-01
4.86354947e-01 -4.53828096e-01 1.79340795e-01 1.00171709e+00
1.18559510e-01 3.46079469e-01 6.52728021e-01 -8.89975846e-01
8.09604585e-01 9.22081828e-01 5.00468969e-01 -2.95059443e-01
6.97762817e-02 7.70044848e-02 -1.49836406e-01 -2.73929060e-01
6.98173821e-01 -1.91869497e-01 -7.54138172e-01 1.71396732e+00
2.10852385e-01 -5.28318644e-01 7.22518444e-01 9.45921540e-01
2.90215224e-01 8.60426486e-01 1.98750079e-01 -2.48318866e-01
1.46151066e+00 -1.16300309e+00 -5.25757968e-01 -3.91648173e-01
8.08685958e-01 -9.55577493e-01 1.25018620e+00 1.20022453e-01
-1.00881791e+00 -3.36858481e-01 -9.23926413e-01 -1.56109974e-01
-5.72793782e-01 1.43181175e-01 4.30578202e-01 5.80939829e-01
-1.23858798e+00 5.58842003e-01 -6.08136237e-01 -9.68724847e-01
-1.30728751e-01 4.58245158e-01 -5.35069525e-01 -2.65523344e-01
-1.56481802e+00 1.24435592e+00 4.92386431e-01 -1.68298244e-01
-4.59124625e-01 -2.85150230e-01 -7.65137374e-01 -2.44696766e-01
3.98066223e-01 -7.97786653e-01 1.54144204e+00 -1.66752982e+00
-1.59367836e+00 1.28320038e+00 -3.93378645e-01 -4.72550154e-01
5.61787963e-01 -3.39449309e-02 -3.95988435e-01 7.12899044e-02
6.55618429e-01 5.49155653e-01 4.24707621e-01 -1.14424396e+00
-8.83653104e-01 -4.86997038e-01 3.49472046e-01 5.27437925e-01
-4.78651896e-02 7.06544518e-01 -1.18133441e-01 -6.65217757e-01
1.18382022e-01 -1.07895660e+00 -5.93123361e-02 -3.48824531e-01
3.24291326e-02 -5.25240839e-01 4.39608753e-01 -8.58253837e-01
8.30185235e-01 -1.87341022e+00 3.63607973e-01 6.48611784e-02
-4.00922149e-01 1.67433679e-01 -2.33375847e-01 7.48007715e-01
-2.95830697e-01 1.49018988e-01 -5.23683488e-01 -1.06583260e-01
-6.07048301e-03 7.70011485e-01 -3.72948289e-01 1.53063878e-01
9.56839919e-02 9.43338513e-01 -1.20141935e+00 -1.98530599e-01
-1.15428567e-01 -4.48205099e-02 -2.49334164e-02 3.20779867e-02
-6.10442579e-01 7.11528242e-01 -4.39726174e-01 4.94109869e-01
1.13751456e-01 -9.58724841e-02 5.81819355e-01 4.62277293e-01
-1.09249696e-01 1.01428866e+00 -7.14447260e-01 1.69949532e+00
-6.98717117e-01 5.93954682e-01 2.47933809e-02 -1.14408946e+00
9.00144279e-01 8.24899852e-01 2.15092897e-01 -6.35127187e-01
-5.65112680e-02 8.91807497e-01 3.84324223e-01 -3.41179878e-01
3.17306936e-01 -6.29381478e-01 -1.86857283e-01 9.76410210e-01
-4.36852723e-02 -9.39661264e-02 4.22756165e-01 -7.43843988e-02
9.27812755e-01 4.19290006e-01 5.46921730e-01 -4.98587996e-01
1.00890613e+00 7.88319945e-01 1.80188760e-01 3.11142802e-01
-1.24841807e-02 3.88878495e-01 5.18174291e-01 -5.36216259e-01
-1.46907616e+00 -7.11775720e-01 7.31754228e-02 1.39770555e+00
-8.46894458e-03 -3.40509832e-01 -9.07603562e-01 -1.04048657e+00
-3.66343111e-01 1.14360690e+00 -3.85634065e-01 2.57439226e-01
-1.10152102e+00 -6.29334092e-01 6.31092489e-01 4.34680372e-01
4.26911175e-01 -1.25236821e+00 -4.15307581e-01 6.01270199e-01
-3.70419025e-01 -1.16931069e+00 -1.31106541e-01 1.67928547e-01
-9.21529353e-01 -8.88099015e-01 -5.18012047e-01 -8.81726384e-01
5.69837093e-01 6.18955642e-02 1.33730817e+00 -5.92313940e-04
5.28821409e-01 1.64069742e-01 -6.35286331e-01 -4.44092780e-01
-1.05821323e+00 2.52628475e-01 -1.86805967e-02 1.64050311e-02
7.33156681e-01 -4.24941987e-01 -1.45175368e-01 3.80972475e-01
-7.16671407e-01 1.24152847e-01 5.46473444e-01 8.54117751e-01
2.21966490e-01 -2.39769951e-01 5.75851023e-01 -1.17297268e+00
6.84135258e-01 -4.65868711e-01 -3.10346901e-01 4.39015985e-01
-5.49138367e-01 3.17363590e-01 9.34020102e-01 -1.60776988e-01
-1.20499909e+00 -7.81454220e-02 -2.97433525e-01 2.65184045e-01
-3.98785025e-01 3.75922143e-01 -4.73478913e-01 2.88180619e-01
7.70544052e-01 4.61310565e-01 1.63408741e-01 -3.85051757e-01
3.45063239e-01 9.40331697e-01 2.83790469e-01 -8.75711501e-01
6.53814077e-01 3.06055963e-01 -2.33571753e-01 -5.45756757e-01
-8.88276696e-01 -5.44867277e-01 -7.51195192e-01 2.45328888e-01
7.23218560e-01 -7.93700278e-01 2.42621335e-03 1.42429769e-01
-1.44977152e+00 -4.09153879e-01 -2.20701143e-01 2.67947882e-01
-8.85366082e-01 1.88375801e-01 -4.40349877e-01 -3.23471278e-01
-3.08121890e-01 -9.86712039e-01 1.30155230e+00 -2.22569346e-01
-6.68637097e-01 -1.30800235e+00 2.70583034e-02 6.99095249e-01
2.36229137e-01 -1.98114485e-01 1.52053976e+00 -1.32057703e+00
-1.66609332e-01 -8.94291103e-02 -2.62239158e-01 5.11219680e-01
2.45615974e-01 -7.01626480e-01 -4.76725698e-01 -2.50242520e-02
2.81257421e-01 -2.86994070e-01 2.25296125e-01 -1.15657046e-01
2.47860432e-01 -3.75806779e-01 -1.52011335e-01 -1.61415070e-01
1.30345976e+00 2.35892564e-01 4.94891524e-01 7.13372052e-01
3.89596909e-01 1.08493495e+00 8.26346517e-01 -3.42820197e-01
5.94609678e-01 7.47130156e-01 -7.40246028e-02 2.88905092e-02
-8.20795819e-02 -3.49267632e-01 5.69757819e-01 6.30401075e-01
-2.25777581e-01 -1.66196495e-01 -1.02539074e+00 7.02548802e-01
-2.11747742e+00 -6.84511423e-01 -8.53911415e-02 1.81659400e+00
1.18927038e+00 2.78657582e-02 7.63297901e-02 5.00578731e-02
6.87229216e-01 -2.98701748e-02 -2.05223501e-01 -7.41375148e-01
-3.82073596e-02 2.91330606e-01 4.98379052e-01 6.71379924e-01
-8.52431297e-01 1.35002184e+00 5.85694075e+00 5.28888464e-01
-1.01666546e+00 4.81185794e-01 4.31485236e-01 3.17187607e-01
-3.80958259e-01 4.30880010e-01 -6.08462393e-01 3.36938918e-01
1.23218691e+00 -3.81767035e-01 7.63746202e-01 5.49908102e-01
4.20667976e-01 -1.54314205e-01 -1.35024738e+00 6.08645678e-01
2.37433538e-01 -1.06053269e+00 1.83450505e-01 -7.25716501e-02
6.60796463e-01 1.54924616e-01 -6.08983099e-01 2.40171745e-01
3.47333908e-01 -8.67415190e-01 8.84828091e-01 -7.11742863e-02
7.14214325e-01 -4.39068407e-01 1.02445483e+00 7.67255664e-01
-6.52106166e-01 -2.53474899e-02 -2.66273886e-01 -3.84414107e-01
8.68935138e-02 3.59991789e-02 -1.10934556e+00 7.34248281e-01
3.04803848e-01 7.57014394e-01 -2.76336670e-01 3.20348829e-01
-8.18031490e-01 8.30680192e-01 -2.09702492e-01 -1.73702464e-01
4.91680324e-01 -4.43823218e-01 7.06557631e-01 9.87014651e-01
1.64069846e-01 5.49018122e-02 1.35199681e-01 5.29831350e-01
-2.20580712e-01 4.92673129e-01 -6.77134037e-01 -3.22466582e-01
1.23641513e-01 8.86527181e-01 -8.36598992e-01 -7.29604781e-01
-6.07172012e-01 1.32499218e+00 6.34089336e-02 1.82351530e-01
-2.84846723e-01 -7.66766816e-02 3.71893525e-01 3.29747945e-01
-1.54366419e-01 -1.67395845e-01 -4.62462008e-01 -1.37998950e+00
1.95195049e-01 -1.27461863e+00 4.25142974e-01 -9.05103147e-01
-1.13015854e+00 6.15637183e-01 5.88798104e-03 -9.68256116e-01
-6.36889398e-01 -8.66864920e-01 -2.73249209e-01 1.27303553e+00
-1.39319658e+00 -1.34471250e+00 6.08683884e-01 2.97304451e-01
1.10180581e+00 -3.58815432e-01 1.00455678e+00 7.00089559e-02
1.27427243e-02 1.59835204e-01 -5.79198860e-02 1.39712006e-01
8.03121507e-01 -1.33069623e+00 6.98312819e-01 9.13519025e-01
3.69349122e-01 8.24711442e-01 9.63211656e-01 -6.60212934e-01
-1.14272976e+00 -9.03117836e-01 1.90214729e+00 -8.80941987e-01
1.19649065e+00 -4.54182327e-01 -8.00133228e-01 8.07147086e-01
5.87178111e-01 -4.44266915e-01 5.85821033e-01 -9.86112580e-02
-1.66006759e-01 3.65964770e-01 -1.02739084e+00 5.55921078e-01
1.01626205e+00 -6.38980448e-01 -1.38718057e+00 5.62838674e-01
8.85766268e-01 -1.88981906e-01 -6.19380116e-01 9.22794417e-02
3.82550776e-01 -3.88430655e-01 4.74672496e-01 -9.55491245e-01
5.43888211e-01 -4.54566538e-01 -3.41199398e-01 -1.52173316e+00
2.83934891e-01 -7.08914220e-01 6.16908014e-01 8.32373619e-01
9.46543634e-01 -8.35351050e-01 5.81433713e-01 5.44968247e-01
-1.19281173e-01 -5.16551435e-01 -9.89873469e-01 -7.82875121e-01
3.53229105e-01 -3.11202049e-01 6.00791276e-01 9.39597249e-01
2.48723358e-01 9.91074324e-01 -6.63175210e-02 -2.19988346e-01
1.36698842e-01 3.88962120e-01 6.72392488e-01 -1.25138092e+00
-3.68599176e-01 -1.60463020e-01 3.74015905e-02 -9.30029273e-01
4.67028886e-01 -1.16571355e+00 4.67950143e-02 -1.85953486e+00
5.80494106e-02 -2.57487029e-01 1.24171577e-01 8.15734029e-01
-6.73717484e-02 2.49770045e-01 -1.80004984e-02 5.29947460e-01
-3.09674621e-01 1.65485352e-01 1.13592231e+00 5.19126728e-02
-8.95182341e-02 -2.28353553e-02 -8.57503235e-01 6.63977563e-01
7.62875021e-01 -6.47296548e-01 -3.32914650e-01 -6.04627550e-01
4.09074098e-01 6.40714169e-02 7.36071020e-02 -3.37238163e-01
-2.16620509e-02 -3.63700479e-01 -1.32884398e-01 1.99787721e-01
-8.59654918e-02 -9.04517889e-01 -5.59941418e-02 3.28893602e-01
-5.32125652e-01 2.40256101e-01 4.91304472e-02 2.18819529e-01
-4.52046305e-01 -5.61866224e-01 5.88669956e-01 -5.09262443e-01
-7.47217238e-01 -3.44616950e-01 -3.65355760e-01 -7.98021778e-02
8.12052369e-01 -3.40620905e-01 -7.38740489e-02 -2.99723566e-01
-7.40164757e-01 9.95992422e-02 8.01009834e-01 6.14744961e-01
-4.17248830e-02 -9.54922795e-01 -7.68450558e-01 1.35955498e-01
3.95387560e-01 -4.14779067e-01 -5.31088710e-01 7.97906518e-01
-5.25388300e-01 7.34905899e-01 -2.54071385e-01 -8.38902742e-02
-1.14398634e+00 3.44363958e-01 2.42424369e-01 -4.41522032e-01
-3.18935573e-01 3.72220695e-01 -8.94661248e-02 -8.94824088e-01
-4.23971683e-01 5.16317971e-02 -2.18363658e-01 -2.81685926e-02
2.89775938e-01 -2.33907178e-02 2.43332446e-01 -1.04130065e+00
-3.18967313e-01 3.50343674e-01 6.80522919e-02 -6.52794361e-01
1.30280137e+00 -2.26947889e-01 -6.91649139e-01 4.37710583e-01
8.96388590e-01 2.42549330e-01 -5.54868400e-01 -2.25547299e-01
6.81917131e-01 -2.75864363e-01 -5.03367662e-01 -1.16410089e+00
-2.87106782e-01 7.76162386e-01 -5.10000922e-02 1.63266644e-01
9.98190463e-01 3.36520195e-01 8.65522683e-01 4.64986801e-01
7.50590861e-01 -1.05645001e+00 -3.57164294e-01 8.11004341e-01
8.35507631e-01 -1.18695009e+00 -4.55037206e-01 -5.58195114e-01
-7.93405354e-01 1.10324943e+00 2.40427449e-01 3.23907621e-02
1.12410359e-01 -1.20808199e-01 3.51764470e-01 -7.19369128e-02
-8.60284925e-01 -2.60177821e-01 -1.58427015e-03 3.12247872e-01
6.32256150e-01 1.61834821e-01 -1.00432312e+00 3.02446634e-01
-4.12526697e-01 -1.94137152e-02 5.36617517e-01 1.00977111e+00
-3.15537959e-01 -2.04408050e+00 -3.18914354e-01 1.23174660e-01
-7.46917486e-01 -4.17855471e-01 -8.89210701e-01 8.38617742e-01
1.15867913e-01 1.09827864e+00 -1.10296346e-01 9.59136039e-02
4.21314716e-01 6.23895407e-01 5.14117122e-01 -1.28186607e+00
-7.85834789e-01 1.95861291e-02 7.49767005e-01 -2.85754472e-01
-7.25952327e-01 -8.61854196e-01 -1.39924443e+00 1.09025024e-01
1.53370038e-01 5.74690580e-01 7.88428366e-01 1.33463264e+00
1.05175920e-01 -2.17623293e-01 3.67599368e-01 -2.87360013e-01
-6.15147591e-01 -1.02603781e+00 -1.11370750e-01 5.04665136e-01
-9.87586565e-03 -2.96951503e-01 -1.62454933e-01 3.27322334e-01] | [10.524209022521973, 9.538413047790527] |
c8897d1b-9e3c-4259-8472-052184b505aa | alba-reinforcement-learning-for-video-object | 2005.13039 | null | https://arxiv.org/abs/2005.13039v2 | https://arxiv.org/pdf/2005.13039v2.pdf | ALBA : Reinforcement Learning for Video Object Segmentation | We consider the challenging problem of zero-shot video object segmentation (VOS). That is, segmenting and tracking multiple moving objects within a video fully automatically, without any manual initialization. We treat this as a grouping problem by exploiting object proposals and making a joint inference about grouping over both space and time. We propose a network architecture for tractably performing proposal selection and joint grouping. Crucially, we then show how to train this network with reinforcement learning so that it learns to perform the optimal non-myopic sequence of grouping decisions to segment the whole video. Unlike standard supervised techniques, this also enables us to directly optimize for the non-differentiable overlap-based metrics used to evaluate VOS. We show that the proposed method, which we call ALBA outperforms the previous stateof-the-art on three benchmarks: DAVIS 2017 [2], FBMS [20] and Youtube-VOS [27]. | ['Laura Sevilla-Lara', 'Timothy Hospedales', 'Shreyank N Gowda', 'Panagiotis Eustratiadis'] | 2020-05-26 | null | null | null | null | ['one-shot-visual-object-segmentation', 'unsupervised-video-object-segmentation'] | ['computer-vision', 'computer-vision'] | [ 7.89119303e-02 -9.49210580e-03 -3.87912989e-01 -3.02376837e-01
-8.39069963e-01 -6.91826046e-01 4.57329929e-01 -1.80744246e-01
-8.03871453e-01 4.92336273e-01 -7.59590715e-02 -1.41927958e-01
8.26798826e-02 -3.06478769e-01 -1.08163095e+00 -6.15658641e-01
-5.18922508e-01 5.74267387e-01 5.81345618e-01 3.98071289e-01
1.41514659e-01 3.70098442e-01 -1.44073129e+00 -1.26231432e-01
7.06568956e-01 1.01572430e+00 2.96446532e-01 1.03212512e+00
2.44155586e-01 8.54887605e-01 -5.66988707e-01 -1.86232999e-01
6.02092326e-01 -4.66988385e-01 -1.11118782e+00 7.16823816e-01
8.58683944e-01 -7.05440104e-01 -3.64015043e-01 9.39997077e-01
2.18835831e-01 6.80911124e-01 7.62087405e-01 -1.46494508e+00
-2.46363699e-01 6.02810502e-01 -6.51304066e-01 3.79548013e-01
4.84197959e-03 5.71810722e-01 1.14068604e+00 -6.51252687e-01
9.59673107e-01 1.26968336e+00 6.12471819e-01 5.99078953e-01
-1.18045926e+00 -3.62064958e-01 5.15178621e-01 2.03266308e-01
-1.26617754e+00 -4.94499415e-01 4.52252001e-01 -7.71441817e-01
7.48570383e-01 3.42317708e-02 7.67262876e-01 9.11085308e-01
-2.78307945e-01 1.45866227e+00 5.26015520e-01 -1.08111292e-01
4.36441302e-01 -3.82678181e-01 2.16883555e-01 8.33124995e-01
9.94476527e-02 -2.12394476e-01 -3.26588511e-01 5.08218408e-02
8.47859144e-01 8.44506919e-02 -2.31781617e-01 -6.64138734e-01
-1.46921623e+00 7.07054317e-01 3.05378765e-01 -2.30695549e-02
-4.50416237e-01 7.31940746e-01 5.27176917e-01 2.72208806e-02
3.53480965e-01 4.82868433e-01 -4.89788145e-01 -1.27985939e-01
-1.42562675e+00 5.12191117e-01 7.87358046e-01 1.06644368e+00
7.54937768e-01 1.48571283e-01 -4.58030939e-01 4.72821504e-01
5.89573205e-01 2.10305274e-01 1.75370798e-01 -1.75130904e+00
3.07051092e-01 3.21902223e-02 4.47204441e-01 -4.21288520e-01
-3.12307060e-01 -4.95261885e-02 -4.22667980e-01 2.69550472e-01
6.19580626e-01 -5.15084386e-01 -1.41501296e+00 1.88347769e+00
4.61416215e-01 8.91471446e-01 -9.48312506e-02 1.09585607e+00
7.29301155e-01 6.03252172e-01 4.23858970e-01 -2.77523249e-01
1.03485823e+00 -1.55381358e+00 -6.67952001e-01 -1.63268134e-01
4.32387561e-01 -3.13278824e-01 7.06651568e-01 1.81730643e-01
-1.30146933e+00 -6.43564403e-01 -1.04913878e+00 6.71241358e-02
-5.99622205e-02 1.75256550e-03 6.26901388e-01 2.91146696e-01
-1.18110597e+00 1.03422427e+00 -1.29289007e+00 -2.76697487e-01
8.48278284e-01 5.65347373e-01 -2.17376538e-02 1.92747399e-01
-8.35953236e-01 3.72558296e-01 5.37919939e-01 -6.28091628e-03
-1.56409800e+00 -6.83552027e-01 -8.70616496e-01 -1.69908311e-02
9.83458161e-01 -8.54899168e-01 1.53274858e+00 -1.12389135e+00
-1.62345481e+00 8.64698589e-01 -1.62989765e-01 -8.09108794e-01
7.91615009e-01 -5.24627745e-01 8.14124644e-02 4.51100260e-01
3.40492666e-01 1.30459583e+00 8.77500176e-01 -1.19221091e+00
-8.82848024e-01 -6.21116161e-02 2.98981011e-01 3.78607661e-01
1.07740998e-01 1.62695393e-01 -1.05490553e+00 -6.18331850e-01
-1.68851048e-01 -8.73645186e-01 -5.06854415e-01 9.06895697e-02
-6.41033113e-01 -5.43294966e-01 9.03940082e-01 -7.39029169e-01
9.01747525e-01 -2.00120735e+00 4.98302311e-01 -1.97633758e-01
3.79010797e-01 3.01546842e-01 -1.99671596e-01 -9.58217159e-02
3.60787183e-01 1.99912623e-01 -4.49924111e-01 -9.12829161e-01
1.13244869e-01 2.59007990e-01 -5.17101660e-02 7.45069921e-01
2.79681414e-01 1.17630064e+00 -9.98707950e-01 -7.72945106e-01
1.92510650e-01 8.41876119e-02 -7.26592362e-01 4.32979345e-01
-6.52950227e-01 4.46990371e-01 -2.87126213e-01 6.75390601e-01
4.12113488e-01 -5.25942922e-01 6.29867613e-02 8.04505646e-02
5.57592660e-02 -1.34259891e-02 -1.33936846e+00 2.00181198e+00
1.71833560e-01 8.06053162e-01 1.69887438e-01 -1.08205986e+00
1.93311542e-01 1.75879464e-01 8.16709936e-01 -8.06923583e-02
2.07292333e-01 -1.72259435e-01 -3.04976374e-01 -6.11196101e-01
4.90548849e-01 2.06783310e-01 1.11958243e-01 3.70156199e-01
5.53693712e-01 1.66981354e-01 5.66607714e-01 3.82427126e-01
9.84672546e-01 5.31089365e-01 8.96338075e-02 -1.52670473e-01
1.45911604e-01 -3.23137417e-02 8.49020958e-01 1.07829428e+00
-5.71836293e-01 6.17411613e-01 5.26954591e-01 -1.24740928e-01
-8.88373017e-01 -1.14915609e+00 3.50286156e-01 1.27757716e+00
5.55924416e-01 -1.89581349e-01 -1.04384184e+00 -1.12637591e+00
-3.34000960e-02 4.65970188e-01 -6.54349566e-01 2.19859630e-01
-7.21839070e-01 -2.80926853e-01 3.97232652e-01 6.94774866e-01
4.95901227e-01 -1.35704541e+00 -8.35900009e-01 2.49144047e-01
-1.26276061e-01 -1.31074405e+00 -8.99389327e-01 1.95396379e-01
-7.82958746e-01 -1.13140881e+00 -9.13685441e-01 -7.72690296e-01
3.80568802e-01 2.91722476e-01 1.27223921e+00 6.03948198e-02
-2.30700657e-01 5.49242079e-01 -1.45647556e-01 -1.01177156e-01
-1.69848055e-01 1.89128712e-01 -3.70433033e-02 7.97293112e-02
1.04879864e-01 -2.46963620e-01 -8.42516959e-01 3.13325405e-01
-8.33582938e-01 9.89372060e-02 2.68602520e-01 5.20158708e-01
8.31306159e-01 -3.48472118e-01 5.08750021e-01 -8.76029015e-01
8.39690119e-02 -5.19339502e-01 -7.47937083e-01 3.21825236e-01
-5.58197051e-02 -1.65091857e-01 3.19019645e-01 -4.84306186e-01
-6.91604793e-01 3.24857593e-01 1.02037378e-01 -9.66953695e-01
-4.63051885e-01 2.37830143e-04 -7.36774653e-02 -4.45994586e-02
3.01130116e-01 -8.71598870e-02 -1.31691277e-01 -4.02079374e-01
5.75905263e-01 3.35278571e-01 8.20628047e-01 -3.83660048e-01
5.98048985e-01 6.65270925e-01 -2.50410736e-01 -5.61714351e-01
-9.63934124e-01 -1.00961745e+00 -9.49462831e-01 -4.02897656e-01
1.45719588e+00 -9.78487372e-01 -8.88858557e-01 3.82731736e-01
-1.12802720e+00 -1.07120371e+00 -5.20186186e-01 3.98169577e-01
-1.08544505e+00 5.60285032e-01 -6.60756946e-01 -8.10921431e-01
-1.28412575e-01 -1.22035384e+00 1.27551734e+00 4.65937346e-01
-1.13721959e-01 -9.77363586e-01 3.76863107e-02 2.18720555e-01
-7.91504011e-02 4.96644527e-01 1.01196349e-01 -6.30977750e-01
-9.51233745e-01 2.20753968e-01 -2.11362794e-01 3.68195593e-01
-7.43342414e-02 2.14856818e-01 -6.89920366e-01 -3.00569654e-01
-2.90449798e-01 -4.35771376e-01 1.36675441e+00 8.25105548e-01
1.28531897e+00 -2.75904506e-01 -3.72504205e-01 9.70473588e-01
1.33044219e+00 2.68368989e-01 2.55430728e-01 1.77718893e-01
1.05352700e+00 4.58309621e-01 8.81230116e-01 3.26725692e-01
4.20419574e-01 7.17282712e-01 6.40597105e-01 -6.23900630e-02
-1.50065243e-01 -2.86090840e-02 3.76025349e-01 2.61462182e-01
-1.81163669e-01 -5.03376603e-01 -6.83462083e-01 7.53738880e-01
-2.33730030e+00 -1.04187155e+00 2.21838295e-01 2.01887751e+00
5.81311226e-01 9.99945402e-02 6.46272719e-01 -4.32898343e-01
8.19450200e-01 4.49217707e-01 -9.50314283e-01 1.35257006e-01
1.26790896e-01 -1.04265578e-01 7.78924644e-01 3.75785589e-01
-1.79204011e+00 1.35190403e+00 7.04550648e+00 6.58975542e-01
-8.29817772e-01 1.90597773e-01 8.22681725e-01 -2.25765496e-01
1.44218877e-01 -2.71609053e-02 -7.29939401e-01 3.69846910e-01
7.80828595e-01 1.63287729e-01 6.32862389e-01 8.35777521e-01
2.51322329e-01 -2.06871480e-01 -1.34038031e+00 8.35791886e-01
5.62205054e-02 -1.54842365e+00 -2.71430850e-01 -1.20943762e-01
1.10960758e+00 4.53578472e-01 -1.23399705e-01 3.19683015e-01
6.65887415e-01 -8.46265614e-01 1.00416815e+00 5.47033191e-01
4.31693941e-01 -5.73493242e-01 3.30281556e-01 9.91983637e-02
-1.13502824e+00 -4.30344827e-02 -1.09602258e-01 2.88802683e-01
6.73335910e-01 7.89105147e-02 -6.69242024e-01 3.57548207e-01
6.39586449e-01 9.91110384e-01 -4.02907193e-01 1.53411901e+00
-1.41320661e-01 8.36238205e-01 -3.61097783e-01 2.02722222e-01
6.47277594e-01 -2.22583786e-01 7.00375438e-01 1.39313519e+00
5.41123897e-02 -8.50289986e-02 8.46284568e-01 8.06662202e-01
-4.39198524e-01 -2.11238459e-01 -2.20543966e-01 -8.67289454e-02
2.50066221e-01 1.23850358e+00 -1.30139172e+00 -9.45477903e-01
-2.99646914e-01 1.11580026e+00 2.92600840e-01 6.84536457e-01
-1.03697634e+00 -1.02236234e-01 7.16806710e-01 -1.35827214e-01
9.29679692e-01 -4.11766529e-01 5.08755110e-02 -1.18630302e+00
-2.45700955e-01 -6.05473876e-01 5.55363178e-01 -5.04112661e-01
-1.11337149e+00 2.64058143e-01 1.88180938e-01 -1.07421255e+00
-3.74909341e-01 -4.91437554e-01 -7.16404200e-01 3.47514868e-01
-1.47776341e+00 -8.74561787e-01 -1.30019307e-01 4.83412027e-01
1.06282949e+00 9.60721746e-02 5.70890531e-02 1.92061767e-01
-7.32414484e-01 3.52496803e-01 -5.77886961e-02 3.62881184e-01
5.88311434e-01 -1.60049796e+00 6.17669880e-01 1.04721034e+00
3.69769126e-01 2.43220761e-01 7.19917357e-01 -6.56649590e-01
-1.15851927e+00 -1.41003191e+00 1.82090908e-01 -3.37667882e-01
6.02408528e-01 -3.76002818e-01 -8.44478548e-01 9.84215438e-01
3.94091100e-01 4.42696422e-01 2.82801092e-01 -2.43717313e-01
-5.49744554e-02 2.74803132e-01 -9.86600935e-01 7.33279705e-01
1.45407403e+00 -7.49762803e-02 -4.57262248e-01 5.91176331e-01
1.16339195e+00 -5.19528627e-01 -6.74067497e-01 2.69301593e-01
2.49772087e-01 -8.17680299e-01 9.61942136e-01 -9.83282447e-01
2.26060063e-01 -5.12854576e-01 -6.34514987e-02 -9.90764380e-01
-1.00303836e-01 -1.03173149e+00 -5.12999535e-01 9.14183021e-01
1.86450556e-01 -1.57859176e-01 1.02274728e+00 4.72904682e-01
-3.18665028e-01 -7.80649722e-01 -7.51306117e-01 -8.37768316e-01
-2.51259893e-01 -3.29097897e-01 1.92555830e-01 7.18907118e-01
-5.07947624e-01 2.57353727e-02 -3.82147759e-01 2.26600930e-01
9.19873178e-01 4.96334918e-02 8.72872353e-01 -9.48915184e-01
-6.17066562e-01 -5.90547264e-01 -3.76531184e-01 -1.57051611e+00
3.40752870e-01 -6.84100688e-01 5.47112226e-01 -1.56037903e+00
2.16265336e-01 -1.62980318e-01 -2.83186972e-01 3.86105120e-01
-3.95463973e-01 3.82138044e-01 4.67301309e-01 3.02028865e-01
-1.74803972e+00 4.67261225e-01 1.26838517e+00 -1.12062059e-01
-4.37679321e-01 1.46207377e-01 -3.43969345e-01 8.80725265e-01
5.56946397e-01 -4.68506217e-01 -2.49299198e-01 -2.67091155e-01
-2.57149547e-01 2.44644329e-01 5.59255600e-01 -9.66810524e-01
3.31149727e-01 -3.76873344e-01 2.34479100e-01 -8.05307090e-01
2.45005384e-01 -4.18867916e-01 -2.13908136e-01 4.07509327e-01
-4.62596923e-01 -1.90286294e-01 -1.51367318e-02 7.56885290e-01
-1.12581059e-01 -3.09556007e-01 7.45851338e-01 -1.70732617e-01
-1.24639642e+00 7.84164846e-01 -4.01889145e-01 4.19170022e-01
1.39433169e+00 -1.67596340e-01 6.07889518e-02 -2.77283281e-01
-9.98273730e-01 7.48262763e-01 4.35787857e-01 3.25760543e-01
3.68169576e-01 -1.15120077e+00 -4.66255844e-01 -2.97157198e-01
-2.64392883e-01 3.84641975e-01 3.48173678e-01 9.22263443e-01
-5.71325004e-01 7.98132047e-02 -1.02376357e-01 -9.83546972e-01
-1.18094110e+00 6.67208016e-01 3.32789928e-01 -1.99194759e-01
-7.19132960e-01 8.65216672e-01 3.07246864e-01 -2.14971714e-02
6.42049015e-01 -3.65180075e-01 -2.94917166e-01 1.28239924e-02
3.64408404e-01 4.59151477e-01 -5.29232204e-01 -5.25741637e-01
-2.42253929e-01 4.33905810e-01 -1.37252793e-01 -2.98740774e-01
1.22510660e+00 -3.00697356e-01 1.71357229e-01 5.85241318e-01
1.26103473e+00 -4.85368311e-01 -2.21114826e+00 -2.36258298e-01
1.88016042e-01 -3.40217799e-01 -2.84603924e-01 -3.67614716e-01
-1.13086832e+00 6.25956476e-01 3.51555973e-01 3.01651925e-01
8.24642599e-01 1.08808115e-01 6.51163697e-01 5.14600933e-01
1.22245461e-01 -1.26559889e+00 2.81364739e-01 3.66411090e-01
3.93626750e-01 -1.43642735e+00 -1.31745577e-01 -2.44182006e-01
-9.06260550e-01 9.11775947e-01 6.77740514e-01 -5.70259273e-01
4.71971601e-01 2.29701726e-03 -1.21385634e-01 -1.37750328e-01
-6.45284951e-01 -5.61137080e-01 4.53552723e-01 4.91426349e-01
2.89863702e-02 -5.94529919e-02 9.45953429e-02 2.10039705e-01
2.52664596e-01 5.03336452e-02 4.89210337e-01 8.61738324e-01
-5.25957882e-01 -6.12932742e-01 -1.57073021e-01 5.64165831e-01
-6.36290669e-01 2.65165538e-01 -1.62000507e-01 8.30239296e-01
1.15840912e-01 6.99701130e-01 2.43246734e-01 -1.39094308e-01
-1.15415797e-01 6.19232804e-02 3.23918819e-01 -7.45396078e-01
-3.28799754e-01 2.63606250e-01 -2.63026673e-02 -8.82082343e-01
-9.37902749e-01 -9.58821416e-01 -1.35854340e+00 1.12417951e-01
-2.48643070e-01 3.43295047e-03 4.68495548e-01 1.19326460e+00
2.12200612e-01 8.43624353e-01 3.85036469e-01 -1.40066433e+00
-3.80904496e-01 -7.19267964e-01 -4.13730383e-01 5.07825434e-01
5.71673870e-01 -7.00439215e-01 -2.16761768e-01 3.07272077e-01] | [9.139674186706543, -0.07904629409313202] |
9c2cd868-a097-417e-9fea-4e419726ff79 | chinese-word-sense-embedding-with-sememewsd | 2206.14388 | null | https://arxiv.org/abs/2206.14388v1 | https://arxiv.org/pdf/2206.14388v1.pdf | Chinese Word Sense Embedding with SememeWSD and Synonym Set | Word embedding is a fundamental natural language processing task which can learn feature of words. However, most word embedding methods assign only one vector to a word, even if polysemous words have multi-senses. To address this limitation, we propose SememeWSD Synonym (SWSDS) model to assign a different vector to every sense of polysemous words with the help of word sense disambiguation (WSD) and synonym set in OpenHowNet. We use the SememeWSD model, an unsupervised word sense disambiguation model based on OpenHowNet, to do word sense disambiguation and annotate the polysemous word with sense id. Then, we obtain top 10 synonyms of the word sense from OpenHowNet and calculate the average vector of synonyms as the vector of the word sense. In experiments, We evaluate the SWSDS model on semantic similarity calculation with Gensim's wmdistance method. It achieves improvement of accuracy. We also examine the SememeWSD model on different BERT models to find the more effective model. | ['Zeli Guan', 'Ang Li', 'Zhe Xue', 'Junping Du', 'Yangxi Zhou'] | 2022-06-29 | null | null | null | null | ['word-sense-disambiguation'] | ['natural-language-processing'] | [-2.04891726e-01 -2.86173612e-01 -2.14025661e-01 -2.58491695e-01
1.09755903e-01 -5.68289161e-01 4.94018346e-01 6.61482990e-01
-9.59086835e-01 4.44016874e-01 7.21949220e-01 -1.24835864e-01
-1.95798546e-01 -1.06318188e+00 3.62663388e-01 -3.50540936e-01
4.02027011e-01 2.54499733e-01 3.25438470e-01 -8.52572441e-01
6.46383166e-01 -5.57099283e-02 -1.30693638e+00 -2.17885408e-03
8.37966800e-01 4.54027265e-01 6.47823811e-01 1.85021032e-02
-9.86967564e-01 -7.69058615e-02 -7.28367746e-01 -2.77834594e-01
1.62622988e-01 -1.53276041e-01 -9.11076367e-01 -5.78797340e-01
4.79041524e-02 5.30501187e-01 -2.11989969e-01 1.52840567e+00
4.72643822e-01 5.35856843e-01 6.72973692e-01 -1.11426425e+00
-1.05235016e+00 8.65180314e-01 -3.59615177e-01 7.06304073e-01
4.59010035e-01 -2.80559391e-01 1.35974741e+00 -7.56211042e-01
4.90781009e-01 1.60778046e+00 4.37947959e-01 5.29024601e-01
-8.64124238e-01 -8.81803274e-01 2.37446487e-01 4.45315033e-01
-1.51813364e+00 4.45350319e-01 6.00911438e-01 -2.67481476e-01
1.25950897e+00 2.30800793e-01 6.90641403e-01 8.21027339e-01
2.59705305e-01 1.58849105e-01 9.14325118e-01 -4.01793480e-01
2.44622558e-01 5.88812083e-02 1.04970646e+00 4.77828234e-01
9.25490975e-01 -3.23124647e-01 -2.80909717e-01 -2.13740766e-01
2.35779122e-01 4.29633528e-01 -2.93297589e-01 -1.61957070e-02
-1.37281895e+00 9.82462108e-01 6.30479991e-01 9.37843919e-01
-1.00515626e-01 3.05250175e-02 5.14273047e-01 4.10530418e-01
1.85478091e-01 1.07267892e+00 -5.12754440e-01 1.88720629e-01
-3.73529196e-01 1.74797133e-01 5.88387609e-01 6.77275062e-01
1.25661492e+00 -1.74931541e-01 -5.44332974e-02 1.01632524e+00
6.30158842e-01 6.72247648e-01 1.58021963e+00 -3.08146417e-01
1.39819801e-01 1.06527019e+00 -2.82041967e-01 -1.55138934e+00
-4.21193004e-01 -1.57507166e-01 -3.91696811e-01 -2.49682859e-01
-5.47801614e-01 -6.09248579e-02 -7.60132432e-01 1.64608157e+00
2.90277004e-01 7.07323253e-01 3.31232578e-01 8.78580868e-01
1.24333918e+00 7.74147511e-01 5.82674325e-01 -8.32993090e-02
1.77187073e+00 -5.96418142e-01 -9.32253957e-01 -5.56160390e-01
9.42984521e-01 -7.78989315e-01 1.29674256e+00 -5.89414127e-02
-2.39177346e-02 -3.81781638e-01 -1.37747860e+00 -3.39317061e-02
-1.31204998e+00 -4.87198770e-01 6.74224973e-01 5.57899415e-01
-6.28374517e-01 5.13321161e-01 -3.00947607e-01 -7.16556072e-01
-2.05164298e-01 8.28867778e-02 -5.08669734e-01 1.34635083e-02
-2.10776544e+00 1.31025267e+00 1.28157246e+00 -7.28760362e-01
-1.53607309e-01 -7.54824698e-01 -1.39332736e+00 2.01312676e-01
1.40964106e-01 -6.41353726e-01 6.11221671e-01 -6.66879356e-01
-7.01151252e-01 8.41821134e-01 -7.20148459e-02 -3.78082275e-01
-8.57079387e-01 2.20797621e-02 -1.16892922e+00 -2.35180274e-01
4.99590695e-01 3.48592132e-01 3.59251589e-01 -9.81132269e-01
-5.32330036e-01 -1.82119891e-01 1.19796045e-01 4.82041568e-01
-8.18906069e-01 -1.18223891e-01 9.98927429e-02 -8.85732591e-01
1.97497487e-01 -5.96207440e-01 -3.32886815e-01 -5.20115614e-01
-1.20502941e-01 -7.47985244e-01 7.74325490e-01 -2.40272433e-01
1.78973508e+00 -2.44676805e+00 -2.05785811e-01 2.60250330e-01
5.30708909e-01 5.35966694e-01 -4.74360555e-01 5.37355721e-01
-4.12257522e-01 3.09546679e-01 -2.01779634e-01 2.18799815e-01
1.82062879e-01 5.70289195e-01 -2.63926983e-01 -1.06599480e-01
-1.26803383e-01 4.98638213e-01 -1.44587719e+00 -4.21410412e-01
4.36290801e-01 2.44221892e-02 -4.17875081e-01 1.85721423e-02
6.01498224e-02 -6.17054164e-01 -7.55076289e-01 -9.20615941e-02
6.60570145e-01 5.19786626e-02 3.73184055e-01 -5.03436089e-01
1.26295388e-01 3.82642955e-01 -1.47000790e+00 1.70576346e+00
-8.36285114e-01 3.64126801e-01 -7.62522936e-01 -7.19855189e-01
1.11674082e+00 1.89577490e-01 2.34914169e-01 -3.65363151e-01
2.76922286e-01 2.93692410e-01 9.98637527e-02 -6.75530314e-01
8.23600709e-01 -5.99811316e-01 -2.75997311e-01 4.59177524e-01
4.64251041e-01 -2.56283075e-01 1.92508817e-01 5.24794698e-01
1.13440585e+00 -6.20906234e-01 8.97957861e-01 -7.98286676e-01
5.81905484e-01 -1.79569468e-01 5.23549020e-01 2.08403260e-01
-2.68037885e-01 2.05502555e-01 8.56351182e-02 -3.49722207e-01
-4.58847255e-01 -1.30415261e+00 -1.50762215e-01 9.54007745e-01
7.42856205e-01 -1.19848013e+00 -4.82511938e-01 -6.17139935e-01
1.34378448e-01 1.05507147e+00 -4.14337188e-01 -6.03837550e-01
-3.49310860e-02 -7.45496690e-01 2.87523270e-01 2.19607174e-01
3.86717200e-01 -1.13076210e+00 -3.03883851e-01 2.28203878e-01
3.25889178e-02 -7.38555849e-01 -5.82459092e-01 2.22948566e-02
-2.33791053e-01 -1.02792072e+00 -1.17413409e-01 -9.81427014e-01
2.77850866e-01 4.09240782e-01 1.09563434e+00 1.32674411e-01
-2.25386009e-01 2.03689888e-01 -7.78367996e-01 -3.67376417e-01
-2.60296091e-02 5.51950128e-04 6.16650939e-01 -1.90080434e-01
1.40200162e+00 -5.08626580e-01 -4.23703671e-01 -1.04067884e-01
-1.25590730e+00 -6.14413679e-01 7.93975145e-02 6.76021755e-01
4.05546606e-01 1.72697917e-01 4.67362881e-01 -7.08764791e-01
1.24367118e+00 -7.94520438e-01 -1.48073912e-01 1.31365836e-01
-1.05345130e+00 5.50753593e-01 3.96559030e-01 -2.49514058e-01
-6.78760171e-01 -5.28084993e-01 -7.34663308e-02 -2.04990372e-01
-7.15994015e-02 7.12816238e-01 -1.93996817e-01 1.56822518e-01
6.60088539e-01 1.40341356e-01 -3.20136070e-01 -4.95172888e-01
7.69642234e-01 9.16934788e-01 1.33088842e-01 -3.27282071e-01
7.50214458e-01 3.84717546e-02 -2.78058916e-01 -8.60694706e-01
-1.06208479e+00 -9.79992986e-01 -3.26052159e-01 3.86288375e-01
1.35569334e+00 -7.35070229e-01 -4.05799121e-01 -2.14180470e-01
-1.35701168e+00 6.22241795e-01 -1.06760114e-01 7.61664271e-01
5.80927655e-02 7.13188827e-01 -7.66802132e-02 -2.21657559e-01
-4.64146346e-01 -6.67619944e-01 7.51444101e-01 2.98176527e-01
-7.38326192e-01 -1.38299394e+00 5.59789419e-01 -2.01368228e-01
3.04393798e-01 -3.71820591e-02 1.21544445e+00 -1.35785627e+00
2.72040427e-01 -8.75069275e-02 -1.28945917e-01 4.54484314e-01
5.85603058e-01 -4.85408247e-01 -5.74559987e-01 -1.50122359e-01
-1.80586539e-02 1.39840364e-01 1.03487837e+00 -1.32891133e-01
9.52625990e-01 -3.13340396e-01 -5.11965930e-01 3.47206920e-01
1.89613438e+00 1.20493457e-01 4.04530525e-01 3.88951600e-01
8.00890982e-01 2.46892571e-01 6.40122950e-01 2.63723940e-01
2.41952166e-01 2.37187088e-01 1.15336433e-01 3.75175357e-01
-3.88281532e-02 -3.89205694e-01 1.21831864e-01 1.50019670e+00
3.69180143e-01 -1.52311355e-01 -9.91500616e-01 6.18907690e-01
-1.41166902e+00 -9.02824819e-01 -1.71734050e-01 1.93502533e+00
7.83912420e-01 8.53479579e-02 -4.48066026e-01 1.51982546e-01
8.71205151e-01 6.23587966e-01 -3.30865197e-02 -6.79212511e-01
-7.25151747e-02 5.88905692e-01 4.30551916e-01 8.44404042e-01
-6.29838526e-01 1.49236858e+00 5.46695471e+00 1.09520102e+00
-9.61393416e-01 3.44039768e-01 -2.42271483e-01 2.85527706e-01
-9.09579039e-01 1.88976854e-01 -7.01013386e-01 7.22584367e-01
5.93562722e-01 -7.58033097e-01 3.23578656e-01 7.79178500e-01
-1.99106812e-01 2.70278901e-01 -7.86371827e-01 1.43669808e+00
4.53808427e-01 -9.91545439e-01 6.88484430e-01 -5.43798089e-01
5.06799519e-01 3.84863392e-02 -3.49936843e-01 4.78951752e-01
5.22539735e-01 -9.71965551e-01 -1.68157563e-01 2.52939939e-01
5.15970826e-01 -8.04599524e-01 9.62491810e-01 -8.68257582e-02
-1.52614939e+00 -5.06207161e-03 -1.01011264e+00 -2.45936826e-01
-5.96358664e-02 6.49512768e-01 -6.51867270e-01 5.62122524e-01
3.62107635e-01 1.02742457e+00 -5.69570661e-01 6.20588720e-01
-5.36532760e-01 3.19195837e-01 -1.11849427e-01 -4.87369567e-01
3.97240996e-01 -1.84895322e-01 6.57251358e-01 1.34066355e+00
4.63912189e-01 2.98705757e-01 2.67167091e-01 6.69395208e-01
-2.19326820e-02 6.10011876e-01 -7.73300409e-01 -2.61969239e-01
1.10490155e+00 1.23332632e+00 -2.84113765e-01 -4.86489117e-01
-2.37926573e-01 1.01901078e+00 9.42541808e-02 1.81871921e-01
-4.15190041e-01 -1.06542802e+00 1.34914184e+00 -2.33020961e-01
-1.64892867e-01 -1.29909590e-01 -1.77971274e-01 -1.21434104e+00
-3.92815083e-01 -3.91077548e-01 8.24864745e-01 -9.38602388e-01
-1.79644489e+00 5.58683515e-01 -1.04874019e-02 -1.05327630e+00
1.75038010e-01 -8.99949729e-01 -9.33206260e-01 9.89928901e-01
-1.40315998e+00 -4.90875304e-01 -2.95546502e-01 5.60126185e-01
5.32382369e-01 -3.84538859e-01 1.17874992e+00 1.66200280e-01
-1.57454953e-01 4.64671224e-01 -8.19421038e-02 8.00870508e-02
8.55059445e-01 -1.39858961e+00 4.59424138e-01 5.35297632e-01
2.21534088e-01 1.26113451e+00 9.07316625e-01 -8.08427095e-01
-8.46058965e-01 -1.11696255e+00 1.42370689e+00 -5.40859818e-01
1.01807594e+00 1.19472079e-01 -9.39763963e-01 5.44375539e-01
5.00554800e-01 -3.38075638e-01 1.13433254e+00 1.81828409e-01
-6.07644200e-01 -7.94553161e-02 -1.04619586e+00 8.83072138e-01
1.06931913e+00 -4.85095292e-01 -1.83655858e+00 1.84997961e-01
1.63621271e+00 4.46151316e-01 -6.66528165e-01 1.04867004e-01
-3.04683354e-02 -3.04565012e-01 1.03469396e+00 -9.15112436e-01
2.77367920e-01 -6.21159256e-01 -5.46800733e-01 -2.06500745e+00
-5.28297186e-01 -1.29664317e-01 3.56402814e-01 1.13769162e+00
3.89060587e-01 -9.47419047e-01 6.85766637e-02 9.88515988e-02
-7.10162381e-03 -3.13664645e-01 -8.83611619e-01 -9.66550469e-01
1.75786033e-01 -3.03371221e-01 1.18171978e+00 1.72556090e+00
5.90374231e-01 8.61305475e-01 1.05734095e-01 -4.67700958e-02
2.19383508e-01 -8.21364149e-02 3.20187360e-02 -1.40635443e+00
1.55230582e-01 -5.16216815e-01 -1.07913458e+00 -4.47451711e-01
6.06504738e-01 -1.41387165e+00 -2.78719336e-01 -1.67786181e+00
1.66609377e-01 3.08701657e-02 -1.09614074e+00 3.89234334e-01
-6.76396012e-01 -5.41194119e-02 2.08313704e-01 -3.67164016e-01
-4.90966499e-01 5.78339517e-01 9.59942937e-01 -3.82175416e-01
5.50975353e-02 -8.22119415e-01 -8.15041423e-01 7.49843359e-01
7.91172445e-01 -6.91890717e-01 -7.05458045e-01 -5.70081055e-01
5.78534245e-01 -6.43416703e-01 6.14541173e-02 -9.77737367e-01
-5.80568686e-02 -4.26952571e-01 -8.28232616e-02 1.75211970e-02
1.56163409e-01 -9.44143176e-01 -3.64538789e-01 6.47628903e-01
-3.69495839e-01 5.77050805e-01 -2.61229696e-03 4.91680771e-01
-3.39888424e-01 -6.86504304e-01 5.82422435e-01 -3.44370216e-01
-1.45398998e+00 1.47485286e-01 -3.38857889e-01 3.63952667e-01
8.45011175e-01 -2.86727957e-02 -2.19482526e-01 1.64704517e-01
-4.73816127e-01 2.19167545e-01 2.57198274e-01 9.88321543e-01
8.23096037e-01 -1.79346585e+00 -5.03339767e-01 1.02080382e-01
6.53655887e-01 -6.38518274e-01 8.95476267e-02 -1.40721560e-01
-4.70908433e-01 1.91802323e-01 -2.84308702e-01 8.89320076e-02
-1.15615094e+00 9.56167161e-01 2.14675039e-01 8.51634443e-02
-3.95317823e-01 8.40315402e-01 2.31372103e-01 -6.92026973e-01
-4.31496263e-01 -3.24935883e-01 -9.17247236e-01 2.39168942e-01
7.43122041e-01 1.96715042e-01 -1.43394187e-01 -6.43926442e-01
-5.85823715e-01 7.58245289e-01 1.18941180e-01 -1.00982480e-01
8.15771341e-01 -4.87223081e-02 -6.57991350e-01 5.57511568e-01
1.56501615e+00 1.13730885e-01 2.55950540e-01 -3.18739355e-01
2.07607761e-01 -5.93471289e-01 4.14172076e-02 -5.13854325e-01
-6.58680499e-01 7.02315927e-01 9.75534081e-01 5.93974255e-02
6.81872129e-01 -8.36056545e-02 1.21935952e+00 6.65200293e-01
4.20202643e-01 -1.10937059e+00 -5.82444742e-02 9.89121675e-01
5.83296895e-01 -9.31002676e-01 -3.04114252e-01 -4.54073727e-01
-6.11452341e-01 9.47091341e-01 8.27474773e-01 -3.87250692e-01
1.11062467e+00 -2.93289423e-01 3.64379019e-01 -4.19519901e-01
-3.77589226e-01 -4.14582074e-01 3.30839723e-01 5.72871387e-01
4.63270366e-01 3.90006393e-01 -1.29494596e+00 8.88600469e-01
-6.27133667e-01 -4.77490664e-01 4.46433872e-01 5.03409505e-01
-1.00734758e+00 -1.24938083e+00 -3.35581577e-03 4.57854450e-01
2.33791005e-02 -6.73599541e-01 -3.73908669e-01 2.73271769e-01
4.46954310e-01 9.91452336e-01 1.79404244e-01 -8.75478268e-01
8.14532191e-02 2.49404609e-01 -5.85753992e-02 -1.25697529e+00
-4.74217921e-01 -7.82827973e-01 -1.55027732e-01 -5.13562977e-01
-2.74022296e-02 -2.13893689e-02 -1.73053694e+00 -1.78208590e-01
-2.41869867e-01 6.47054791e-01 3.62023443e-01 1.25457001e+00
2.13599667e-01 4.86821055e-01 5.52436233e-01 2.06435665e-01
-5.13839602e-01 -1.12369204e+00 -8.98564696e-01 1.10350716e+00
-5.97163327e-02 -8.41166198e-01 -4.16438401e-01 -4.08364683e-01] | [10.224637031555176, 8.840865135192871] |
da24208a-8697-4ea4-a8ec-18f351bd0f83 | deepfakeson-phys-deepfakes-detection-based-on | 2010.00400 | null | https://arxiv.org/abs/2010.00400v3 | https://arxiv.org/pdf/2010.00400v3.pdf | DeepFakesON-Phys: DeepFakes Detection based on Heart Rate Estimation | This work introduces a novel DeepFake detection framework based on physiological measurement. In particular, we consider information related to the heart rate using remote photoplethysmography (rPPG). rPPG methods analyze video sequences looking for subtle color changes in the human skin, revealing the presence of human blood under the tissues. In this work we investigate to what extent rPPG is useful for the detection of DeepFake videos. The proposed fake detector named DeepFakesON-Phys uses a Convolutional Attention Network (CAN), which extracts spatial and temporal information from video frames, analyzing and combining both sources to better detect fake videos. This detection approach has been experimentally evaluated using the latest public databases in the field: Celeb-DF and DFDC. The results achieved, above 98% AUC (Area Under the Curve) on both databases, outperform the state of the art and prove the success of fake detectors based on physiological measurement to detect the latest DeepFake videos. | ['Ruben Tolosana', 'Javier Hernandez-Ortega', 'Julian Fierrez', 'Aythami Morales'] | 2020-10-01 | null | null | null | null | ['heart-rate-estimation'] | ['medical'] | [-2.30242431e-01 -1.90978572e-02 5.74306175e-02 1.70620829e-01
-1.85898781e-01 -3.48922282e-01 6.70379877e-01 -2.84966439e-01
-4.66227531e-01 5.06594837e-01 1.80777505e-01 1.97875977e-01
3.71600658e-01 -4.30398762e-01 -6.42207265e-01 -7.61416852e-01
-1.97537810e-01 -2.29764089e-01 1.11752056e-01 -5.22097386e-03
4.48280841e-01 5.57713628e-01 -1.45304203e+00 5.07330120e-01
4.80038583e-01 1.24013543e+00 -6.98382258e-01 8.10530245e-01
3.43571305e-01 8.76760066e-01 -9.28556025e-01 -4.94822294e-01
3.58001173e-01 -5.57743251e-01 -2.36119315e-01 -2.05573574e-01
7.90745854e-01 -9.98495102e-01 -7.79166758e-01 1.04270148e+00
6.76875234e-01 -4.23914611e-01 3.39484453e-01 -1.25622439e+00
-4.33294684e-01 2.56979167e-02 -4.08803314e-01 8.94685388e-01
4.56194311e-01 5.89201748e-01 2.84926325e-01 -8.39972079e-01
7.11845994e-01 1.05020082e+00 8.00794661e-01 6.48227274e-01
-6.94129825e-01 -5.83692908e-01 -6.98326170e-01 7.41312206e-01
-1.16459560e+00 -3.40488911e-01 8.55851233e-01 -4.74922001e-01
7.92505622e-01 1.97170854e-01 9.12146866e-01 1.59984744e+00
5.96992254e-01 8.03635657e-01 1.37272084e+00 -2.81508207e-01
8.92099962e-02 3.93790841e-01 3.61851268e-02 7.19407260e-01
5.13756514e-01 3.46242279e-01 -9.06956255e-01 -1.88513085e-01
6.06001914e-01 -2.41640806e-01 -6.33181572e-01 -1.72383562e-02
-1.16560090e+00 5.52634180e-01 2.54267067e-01 6.31707132e-01
-5.48135936e-01 3.52909505e-01 8.29056859e-01 4.84028876e-01
5.68437278e-01 5.45855641e-01 -8.62293541e-02 -2.23042548e-01
-1.05635452e+00 -1.82321873e-02 8.39837313e-01 1.65777355e-01
1.36653483e-01 6.43707216e-02 -4.29831207e-01 1.17007904e-01
2.56148219e-01 5.49349070e-01 6.17300212e-01 -6.42451525e-01
2.34013394e-01 2.25274250e-01 3.18776041e-01 -1.55447555e+00
-2.87607729e-01 -1.57833040e-01 -4.55220431e-01 2.61807859e-01
5.60129881e-01 -1.31443664e-01 -5.85872531e-01 1.00045395e+00
3.18571061e-01 6.68490171e-01 -6.14613853e-02 1.56023955e+00
1.21418440e+00 4.21811253e-01 3.03022135e-02 -1.75726891e-01
1.42600238e+00 -6.58687174e-01 -9.35914755e-01 1.55170903e-01
4.96083856e-01 -4.23228770e-01 6.19780719e-01 7.75658369e-01
-5.35219967e-01 -5.41433275e-01 -1.23615456e+00 7.83373192e-02
-4.88703668e-01 4.35994506e-01 2.57386893e-01 1.28173387e+00
-9.19875503e-01 7.85573125e-01 -5.90106666e-01 -4.71698046e-01
5.95164955e-01 -4.90700752e-02 -5.16696095e-01 -6.74638897e-02
-1.44517648e+00 1.02158511e+00 2.41734926e-02 5.97272992e-01
-1.15529442e+00 -4.84401375e-01 -4.20605958e-01 -3.14528011e-02
1.37742847e-01 -2.60358661e-01 5.21818280e-01 -1.36280489e+00
-1.51874089e+00 1.26121938e+00 3.59464526e-01 -8.27463150e-01
1.19713652e+00 -3.66031617e-01 -5.67313790e-01 8.98151159e-01
-4.36497688e-01 4.48083937e-01 1.32117403e+00 -6.80202603e-01
1.58660680e-01 -2.96139508e-01 -3.34798008e-01 -1.19112730e-01
-4.69876766e-01 1.72462612e-01 -1.97324064e-02 -3.79320145e-01
-4.89846379e-01 -6.56581581e-01 6.33368850e-01 2.25747451e-01
-3.69860977e-01 -1.20360002e-01 1.20107567e+00 -1.28196132e+00
9.80734527e-01 -2.11530066e+00 -3.12108189e-01 1.01117194e-01
5.86237609e-01 9.44025576e-01 -1.00907408e-01 1.80663675e-01
-2.53506452e-02 -4.11651507e-02 3.87734413e-01 -3.85557450e-02
-1.17442720e-01 -3.10633451e-01 -1.40350252e-01 1.16081536e+00
6.80163950e-02 1.12255716e+00 -8.25663984e-01 -2.14735046e-01
5.80482304e-01 6.65521622e-01 -1.77666303e-02 7.63249546e-02
2.19030946e-01 2.90949464e-01 -2.03962743e-01 7.58148015e-01
9.19362426e-01 1.24910221e-01 -1.24231629e-01 -5.37326038e-01
4.94482331e-02 -1.02104619e-01 -4.67073768e-01 1.23180294e+00
1.42780438e-01 1.34696221e+00 -2.68187281e-02 -1.00069928e+00
1.05636334e+00 5.21677911e-01 6.23801112e-01 -8.92193377e-01
6.13755345e-01 2.89150417e-01 -4.23124880e-02 -1.09827685e+00
1.24525718e-01 6.24300949e-02 3.70209634e-01 8.62303972e-02
1.80166066e-01 4.61795747e-01 -1.94202974e-01 1.46587223e-01
1.17856848e+00 1.79133162e-01 9.83193740e-02 -2.38067582e-01
6.97915673e-01 -1.81367889e-01 1.52327299e-01 8.68550062e-01
-1.02368152e+00 3.81183267e-01 9.04809713e-01 -7.71435916e-01
-8.59407961e-01 -5.99075019e-01 -1.83904748e-02 2.10972995e-01
3.08324754e-01 -1.01924472e-01 -9.07680213e-01 -8.07626009e-01
2.52868384e-01 8.70523527e-02 -1.08646429e+00 -3.26738954e-01
-2.63017058e-01 -7.72130549e-01 1.03396702e+00 -8.37572850e-04
9.62339401e-01 -1.20369291e+00 -1.18979084e+00 1.65863261e-01
-2.60002851e-01 -1.41700780e+00 -2.30738968e-02 -4.27390426e-01
-5.07442236e-01 -1.46089149e+00 -9.92940307e-01 -4.30015065e-02
1.08102988e-02 2.88445167e-02 7.60539234e-01 2.13056393e-02
-1.03766096e+00 5.82403302e-01 -3.83258730e-01 -2.40358531e-01
-4.74605650e-01 -5.28748572e-01 -4.28254753e-02 5.78376114e-01
8.60810995e-01 -5.40858284e-02 -8.74878287e-01 2.25853160e-01
-6.45081282e-01 -3.54422539e-01 5.17424166e-01 4.96188372e-01
-8.46980289e-02 -4.28973913e-01 3.92822504e-01 -4.06931072e-01
4.71793443e-01 -3.13323677e-01 -6.52975798e-01 3.40468585e-02
-2.67926753e-01 -4.24410284e-01 2.41524622e-01 -3.77930671e-01
-5.34945548e-01 -2.31935978e-01 2.71940768e-01 -1.05724740e+00
-2.51242727e-01 1.33793667e-01 3.36075187e-01 -5.48722029e-01
8.82409632e-01 3.45798403e-01 2.59314924e-01 -1.99949101e-01
-2.08173200e-01 7.89088547e-01 6.34542763e-01 2.07369342e-01
3.08545232e-01 7.64534175e-01 1.68834552e-01 -9.65838253e-01
-2.91104764e-01 -6.91885173e-01 -4.83105212e-01 -7.97643900e-01
9.21248674e-01 -9.46402073e-01 -1.12122500e+00 1.14795756e+00
-1.28365886e+00 -6.35764226e-02 2.60952979e-01 6.56123817e-01
-3.18678707e-01 9.34548616e-01 -8.21044743e-01 -1.05183613e+00
-6.17600799e-01 -6.50358438e-01 9.49227571e-01 1.14976108e-01
1.01519741e-01 -9.00432110e-01 1.31972045e-01 5.75003088e-01
5.75902045e-01 9.12018597e-01 7.01488629e-02 -5.30144095e-01
-5.56241870e-01 -5.63436449e-01 -3.66793007e-01 6.54014885e-01
-2.10194886e-01 5.18173315e-02 -1.43630326e+00 -1.17084466e-01
4.55504097e-02 -2.79881239e-01 1.08552361e+00 3.60887378e-01
9.70340312e-01 -1.60069615e-01 -1.09914981e-01 3.91751081e-01
1.39468706e+00 -1.69404134e-01 1.32165265e+00 3.19305062e-01
5.73129714e-01 5.84177911e-01 4.28954095e-01 5.82775891e-01
-1.57790810e-01 7.51114726e-01 6.97627902e-01 -1.64972126e-01
-3.84378105e-01 3.06972951e-01 8.58801067e-01 1.15931585e-01
-2.65661538e-01 -3.19748670e-01 -7.04823732e-01 5.31973362e-01
-1.49987721e+00 -1.15839088e+00 -5.40377676e-01 2.09886527e+00
2.37478882e-01 -2.86860198e-01 3.15531462e-01 9.31362584e-02
8.47550154e-01 1.55685768e-01 -3.39188159e-01 -5.77026546e-01
-2.59165049e-01 1.86057702e-01 8.03385913e-01 1.52388457e-02
-1.33456433e+00 7.17034936e-01 5.77452946e+00 3.97725463e-01
-1.65909171e+00 2.95698017e-01 5.32032728e-01 -1.30946010e-01
5.39404809e-01 -6.80895030e-01 -2.91095167e-01 9.62584734e-01
1.21392202e+00 2.00749069e-01 4.39102560e-01 5.28129518e-01
6.71771169e-01 -3.85308206e-01 -6.43810272e-01 1.32989216e+00
6.91257060e-01 -1.28233862e+00 -3.67077231e-01 1.43518686e-01
2.31641948e-01 1.60508901e-01 -7.04566166e-02 1.14594623e-02
-6.33090019e-01 -9.44340169e-01 3.54046702e-01 7.49939740e-01
7.33802199e-01 -4.63894397e-01 1.21662092e+00 -3.42954509e-02
-5.68273187e-01 -1.58936735e-02 -3.58846933e-01 1.53462410e-01
-1.71781838e-01 7.02239990e-01 -9.47697222e-01 2.60206133e-01
8.64868760e-01 9.78827536e-01 -5.46718061e-01 1.23187542e+00
-1.23810127e-01 5.97676456e-01 -2.18880847e-01 -9.78170931e-02
1.33263767e-01 2.16618210e-01 6.77734733e-01 1.35234451e+00
4.41447049e-01 -1.36276290e-01 -3.23386878e-01 9.38681066e-01
-2.23242249e-02 2.84249103e-03 -6.79120302e-01 -3.37227672e-01
-4.74101342e-02 1.31253195e+00 -5.79519510e-01 -5.21571934e-01
-3.06067884e-01 1.24347854e+00 -4.11518127e-01 6.30861819e-02
-1.12336767e+00 -3.65032494e-01 6.86945021e-01 1.63449854e-01
2.03022331e-01 1.52277902e-01 1.87208623e-01 -1.56710231e+00
8.61400068e-02 -6.81979060e-01 4.17294681e-01 -1.04315603e+00
-1.03983235e+00 3.50829542e-01 -3.59652102e-01 -1.38407767e+00
1.42119437e-01 -9.85538781e-01 -4.57124501e-01 6.83731019e-01
-1.68076861e+00 -8.60496342e-01 -1.03343225e+00 7.18759656e-01
1.76900283e-01 -5.97869270e-02 4.82787639e-01 3.66546392e-01
-4.86958504e-01 5.87573767e-01 -1.87915415e-01 4.76174653e-01
9.05555308e-01 -8.08818877e-01 2.17559740e-01 9.98561680e-01
-1.98245332e-01 3.79584134e-02 5.91045439e-01 -6.76303804e-01
-1.53640914e+00 -6.74121857e-01 6.25152111e-01 -4.98800755e-01
5.34843624e-01 -2.38685608e-01 -8.31491768e-01 1.24800488e-01
2.97125548e-01 4.10197496e-01 2.96856850e-01 -7.56115794e-01
-5.41028917e-01 4.50766971e-03 -1.57712758e+00 -2.00428963e-01
3.50672573e-01 -5.68924844e-01 -3.73673528e-01 3.42187017e-01
-1.90632388e-01 -2.99148649e-01 -7.84232438e-01 2.39500746e-01
7.91361749e-01 -1.33966374e+00 8.36268723e-01 -4.88966405e-01
5.79275310e-01 -2.36338556e-01 4.59360510e-01 -1.17498994e+00
1.60174578e-01 -6.71413422e-01 -4.78952467e-01 4.19371396e-01
-3.54612023e-01 -7.36417055e-01 8.43897223e-01 2.75155246e-01
1.82429820e-01 -3.31037045e-02 -1.14778173e+00 -7.89948404e-01
-3.70467752e-01 -8.31018388e-02 -5.58499470e-02 1.24360859e+00
1.21761665e-01 -2.98478276e-01 -9.64936376e-01 9.08395946e-02
7.57922053e-01 -2.91042477e-01 7.85071135e-01 -1.13300025e+00
-2.25091517e-01 -2.32650802e-01 -1.19829035e+00 -3.27804416e-01
-1.92547873e-01 -4.17482853e-01 -2.18752682e-01 -8.74047160e-01
6.62997961e-02 3.88581544e-01 -4.02341217e-01 3.66677642e-01
9.17344615e-02 9.41280484e-01 2.69253880e-01 9.62865651e-02
-3.75080943e-01 1.99053556e-01 1.10858512e+00 -5.83102442e-02
1.47672802e-01 -6.62263513e-01 1.09370686e-02 2.88518429e-01
6.77924514e-01 -3.30798268e-01 3.15734804e-01 1.06669478e-01
-6.04290590e-02 1.83799431e-01 1.09314620e+00 -1.31300581e+00
-1.82241857e-01 3.73805046e-01 6.96962118e-01 -2.72228390e-01
3.17455679e-01 -5.60190380e-01 -1.41775817e-01 1.04004431e+00
-2.72026788e-02 -3.60881239e-01 2.94060260e-01 6.59555614e-01
-1.70593038e-01 1.15510181e-01 1.02840078e+00 -1.61521107e-01
-8.63070667e-01 -1.11710034e-01 -6.11314535e-01 -3.95530134e-01
1.06330800e+00 -4.47132468e-01 -8.89616668e-01 -4.02150780e-01
-7.49060869e-01 -3.59140277e-01 9.37159806e-02 4.76269513e-01
6.42011881e-01 -9.89033699e-01 -8.32606733e-01 2.51370251e-01
2.50567019e-01 -1.23730433e+00 4.53028083e-01 1.42044079e+00
-1.07217491e+00 3.27886164e-01 -8.06072474e-01 -5.62002659e-01
-1.55859065e+00 5.59300661e-01 9.03491735e-01 1.74229115e-01
-7.75652766e-01 5.40115237e-01 -3.24819088e-01 2.17859417e-01
1.80150475e-02 -2.84532964e-01 -3.93538058e-01 1.03314310e-01
7.80428648e-01 6.61513150e-01 3.20245624e-01 -6.09377086e-01
-5.32401145e-01 2.14313418e-01 2.88669467e-01 3.38610321e-01
9.22728598e-01 1.48557061e-02 -1.39263138e-01 3.79026458e-02
1.35719585e+00 -7.35622421e-02 -1.19387996e+00 1.38738006e-01
-2.90173739e-01 -9.85903561e-01 3.58865887e-01 -1.10906518e+00
-1.25863326e+00 1.12136412e+00 1.16491640e+00 2.51760632e-01
9.38451588e-01 -3.37323606e-01 7.99555361e-01 1.19281307e-01
2.42548957e-01 -9.29333389e-01 3.92156929e-01 -1.85891297e-02
6.88869357e-01 -1.14136565e+00 -3.41534577e-02 -3.36533755e-01
-5.06318033e-01 1.55686498e+00 4.75217581e-01 -5.03588021e-01
4.65754300e-01 -1.25587881e-01 2.44189411e-01 -3.79450887e-01
-3.59924704e-01 -8.90612043e-03 2.07674459e-01 4.47596133e-01
7.02617466e-02 2.31814291e-02 -6.90695882e-01 8.11492056e-02
3.34889024e-01 4.74469662e-01 9.68388081e-01 3.27873677e-01
-2.63142139e-01 -3.19180280e-01 -4.74932045e-01 2.89016128e-01
-7.98446953e-01 1.48752168e-01 -9.19818044e-01 8.83402407e-01
1.96093664e-01 8.27578723e-01 -2.56310795e-02 -2.82874495e-01
6.62298640e-04 4.89900121e-03 5.48712432e-01 3.11885960e-02
-9.14421141e-01 -7.89278094e-03 1.15646303e-01 -1.04882944e+00
-5.97601652e-01 -5.33189237e-01 -4.31286007e-01 -4.34892088e-01
-1.26159951e-01 -1.88988671e-01 7.35836923e-01 9.01357591e-01
4.33176965e-01 1.76249608e-01 5.64090490e-01 -8.46458316e-01
-2.34096259e-01 -1.08238053e+00 -7.81939387e-01 6.55121207e-01
5.34125388e-01 -6.32997692e-01 -6.51203096e-01 -3.21291648e-02] | [12.660526275634766, 1.1719478368759155] |
75985bd7-87cc-43a7-ab14-5696c94cd9be | cdlt-a-dataset-with-concept-drift-and-long | 2306.02346 | null | https://arxiv.org/abs/2306.02346v1 | https://arxiv.org/pdf/2306.02346v1.pdf | CDLT: A Dataset with Concept Drift and Long-Tailed Distribution for Fine-Grained Visual Categorization | Data is the foundation for the development of computer vision, and the establishment of datasets plays an important role in advancing the techniques of fine-grained visual categorization~(FGVC). In the existing FGVC datasets used in computer vision, it is generally assumed that each collected instance has fixed characteristics and the distribution of different categories is relatively balanced. In contrast, the real world scenario reveals the fact that the characteristics of instances tend to vary with time and exhibit a long-tailed distribution. Hence, the collected datasets may mislead the optimization of the fine-grained classifiers, resulting in unpleasant performance in real applications. Starting from the real-world conditions and to promote the practical progress of fine-grained visual categorization, we present a Concept Drift and Long-Tailed Distribution dataset. Specifically, the dataset is collected by gathering 11195 images of 250 instances in different species for 47 consecutive months in their natural contexts. The collection process involves dozens of crowd workers for photographing and domain experts for labelling. Extensive baseline experiments using the state-of-the-art fine-grained classification models demonstrate the issues of concept drift and long-tailed distribution existed in the dataset, which require the attention of future researches. | ['Xinge You', 'Chuanwu Yang', 'Jiamiao Xu', 'Yu Wang', 'Ruxin Wang', 'Yufeng Shi', 'Shuo Ye'] | 2023-06-04 | null | null | null | null | ['fine-grained-visual-categorization'] | ['computer-vision'] | [-7.97310546e-02 -6.53421938e-01 -8.70082248e-03 -7.63602197e-01
1.56061873e-01 -7.60138333e-01 7.37738550e-01 1.59500256e-01
-5.25396585e-01 7.33650386e-01 -1.72868207e-01 5.24501279e-02
-1.79036453e-01 -6.61874175e-01 -5.20649374e-01 -1.15326512e+00
1.40237194e-02 4.70844060e-01 4.61189181e-01 1.72789656e-02
4.97756869e-01 3.75231177e-01 -2.12222171e+00 2.69361854e-01
8.82202685e-01 1.12781191e+00 2.90107757e-01 4.70311165e-01
-4.64076668e-01 5.66196680e-01 -1.20580089e+00 -4.76549029e-01
1.07564718e-01 4.98678759e-02 -7.42536664e-01 3.74438643e-01
5.47658682e-01 -7.57734478e-02 -1.06624871e-01 1.62415147e+00
2.74051577e-01 -2.93737054e-02 9.87218797e-01 -1.66671169e+00
-1.05669093e+00 3.33615094e-01 -9.61767435e-01 7.43221045e-01
-3.70564044e-01 1.78108305e-01 4.29958552e-01 -3.59890610e-01
3.40385795e-01 1.63281000e+00 6.43991888e-01 6.12073004e-01
-7.92195618e-01 -8.87358248e-01 7.49348819e-01 5.82927048e-01
-1.63173211e+00 -1.41906038e-01 6.09431028e-01 -8.64238501e-01
3.45380604e-01 5.47058657e-02 5.86028874e-01 1.04770911e+00
3.08603823e-01 5.07178009e-01 1.23673475e+00 -3.00210238e-01
4.16201770e-01 2.61057854e-01 5.13163507e-01 2.99980074e-01
6.22765243e-01 2.32999906e-01 -3.20680052e-01 2.55475864e-02
4.15099353e-01 4.46881831e-01 -4.51468788e-02 -2.96558112e-01
-1.03714192e+00 6.88483179e-01 6.92314029e-01 3.21364105e-01
-8.80143717e-02 2.77562113e-03 5.22370219e-01 2.89307624e-01
4.34555292e-01 -9.83894244e-02 -3.95101339e-01 4.85071167e-02
-6.69512391e-01 3.29793036e-01 2.46261418e-01 1.07710755e+00
7.62527227e-01 -2.09497899e-01 -1.82780683e-01 8.49178314e-01
2.17373565e-01 5.69988668e-01 9.96113062e-01 -6.82270825e-01
2.13898897e-01 7.23675549e-01 1.37920305e-01 -1.23441887e+00
-2.12014288e-01 -4.38327700e-01 -1.11172593e+00 2.04184756e-01
6.15386009e-01 4.36187759e-02 -1.18718791e+00 1.62955141e+00
3.63461912e-01 3.71125266e-02 -3.37880403e-01 9.03069198e-01
7.66039491e-01 4.25778478e-01 3.48233640e-01 -3.13178778e-01
1.22678125e+00 -6.08614624e-01 -8.09001863e-01 -1.60650283e-01
3.10770865e-03 -4.12859440e-01 1.23032248e+00 3.58827442e-01
-2.70568043e-01 -9.68914330e-01 -1.03598666e+00 4.81571466e-01
-6.38777316e-01 -2.56527930e-01 5.80154359e-01 6.59778655e-01
-8.00172627e-01 2.67615259e-01 -4.49229360e-01 -4.30859983e-01
7.69665658e-01 -1.13947600e-01 -2.09244013e-01 -4.66343641e-01
-1.01063311e+00 4.74379569e-01 5.82207739e-01 2.53188223e-01
-1.02000248e+00 -4.14731443e-01 -1.81671754e-01 -2.60762960e-01
2.04334095e-01 -4.15347278e-01 1.31397414e+00 -1.14530706e+00
-5.76717019e-01 1.09034383e+00 -2.59315316e-02 -2.31995940e-01
6.27451718e-01 1.46588832e-01 -6.38393402e-01 -1.62614360e-01
4.20374572e-01 5.00077724e-01 1.25990260e+00 -1.41286564e+00
-1.16054738e+00 -6.48630738e-01 4.39087301e-02 -2.66840532e-02
-2.87668288e-01 -2.52381623e-01 -4.08642262e-01 -7.39078641e-01
-2.77702630e-01 -7.87243009e-01 -6.19569086e-02 -5.85324578e-02
-1.23852724e-02 -5.42377889e-01 8.85763049e-01 1.66781358e-02
1.36440122e+00 -2.27473664e+00 -2.36350551e-01 -3.06806892e-01
3.24900627e-01 2.67865598e-01 1.58153608e-01 9.59395394e-02
2.57087722e-02 2.08658278e-01 -1.90027341e-01 2.06050143e-01
-1.10625885e-01 4.14554834e-01 -3.70155931e-01 5.86224437e-01
-3.45217474e-02 4.80868429e-01 -9.95235860e-01 -6.17506385e-01
2.01022133e-01 -6.80747405e-02 -6.23869859e-02 2.69332796e-01
-3.13038945e-01 3.23745430e-01 -6.25487089e-01 8.45960617e-01
8.53032112e-01 -2.26134196e-01 -9.84516069e-02 -9.44012180e-02
3.25563247e-03 -8.74485850e-01 -1.06158352e+00 1.32841957e+00
1.54863477e-01 6.38576031e-01 -2.41807222e-01 -1.08834291e+00
9.26969707e-01 -1.75913602e-01 1.00202575e-01 -6.95167005e-01
2.93052405e-01 -2.80935839e-02 8.65008980e-02 -6.02917552e-01
3.83154303e-01 -3.10644865e-01 -3.21208537e-01 1.77766457e-01
-1.35568678e-01 -2.08892927e-01 2.91900635e-01 1.97657254e-02
6.93550587e-01 -2.79073656e-01 2.30216861e-01 -5.07230639e-01
4.20053810e-01 3.38787109e-01 7.76223421e-01 8.23392332e-01
-8.31890523e-01 3.66872728e-01 1.80564910e-01 -8.40913713e-01
-7.73042262e-01 -9.72717166e-01 -4.26377714e-01 1.27631688e+00
6.36040628e-01 1.08992808e-01 -8.43706727e-01 -6.58279836e-01
1.28342286e-01 4.29031223e-01 -1.02073300e+00 -3.61945182e-01
-6.78427070e-02 -1.16342294e+00 4.87685919e-01 4.28277880e-01
8.84504139e-01 -1.15203071e+00 -6.75152600e-01 1.30174786e-01
-1.77494705e-01 -6.99272215e-01 -3.26040119e-01 1.84555054e-01
-6.66571081e-01 -1.36016083e+00 -8.16010654e-01 -7.52723157e-01
8.00861657e-01 6.82959795e-01 1.29057658e+00 1.37875780e-01
-5.80036342e-01 1.03316631e-03 -5.90232372e-01 -1.00288558e+00
-2.08387002e-02 -2.63824373e-01 1.80732369e-01 1.28035933e-01
8.15657556e-01 -2.08138034e-01 -7.20849812e-01 6.19267821e-01
-1.02679265e+00 -5.35299778e-01 3.27933997e-01 9.74499881e-01
5.06580472e-01 9.84785795e-01 6.17255807e-01 -8.09974909e-01
7.71500468e-01 -5.46138763e-01 -6.79551363e-01 4.94670451e-01
-6.78929090e-01 -1.57032371e-01 5.36723673e-01 -6.06814563e-01
-1.31087673e+00 -3.41639876e-01 4.21350986e-01 -4.60237563e-01
-6.02658331e-01 -4.30248640e-02 -2.20316440e-01 1.32947266e-01
8.03861201e-01 1.32387578e-01 -4.50495183e-01 -4.41970259e-01
3.91127616e-01 1.16904962e+00 7.66719103e-01 -5.66989541e-01
8.06321859e-01 5.29267728e-01 -4.68303710e-01 -8.09782088e-01
-9.78518546e-01 -6.60420597e-01 -7.80855775e-01 -3.85259539e-01
7.12720633e-01 -9.40087497e-01 -5.64049721e-01 1.05678570e+00
-8.50690365e-01 -4.25291620e-02 -1.86318532e-01 -4.17235214e-03
-2.08273798e-01 3.09896082e-01 -7.53530860e-02 -7.92555511e-01
-9.92169902e-02 -1.02765584e+00 1.02237940e+00 6.55802429e-01
8.12442601e-02 -8.28438401e-01 -1.45016447e-01 3.43341082e-01
2.06892744e-01 3.53238843e-02 9.48729515e-01 -1.40536904e-01
-3.50616455e-01 -1.11070737e-01 -5.09554267e-01 3.12637031e-01
3.55720967e-01 2.67852426e-01 -1.12743771e+00 -4.54387277e-01
6.57025129e-02 -4.50929135e-01 7.54003644e-01 5.65436780e-01
1.44543099e+00 -7.28193158e-03 -4.62209940e-01 5.46703458e-01
1.47854102e+00 6.80284500e-01 2.24122837e-01 4.49463457e-01
6.72929943e-01 7.61551261e-01 1.16727304e+00 5.91927528e-01
4.11429852e-01 3.23853940e-01 5.99677861e-01 2.69655824e-01
3.06735188e-02 -1.47041783e-01 -2.66151547e-01 5.04625797e-01
-3.94033566e-02 -3.92602801e-01 -9.68518198e-01 9.04604614e-01
-1.82187927e+00 -9.31230247e-01 -3.70135368e-03 1.98764157e+00
7.77058423e-01 1.28793612e-01 1.68749854e-01 2.94025481e-01
1.18787849e+00 1.09006092e-01 -9.59543228e-01 -1.54678896e-01
-7.96587095e-02 -5.10948777e-01 6.28412187e-01 -8.19620769e-03
-1.20016146e+00 9.23123956e-01 6.10353708e+00 9.58755434e-01
-1.22247875e+00 -3.89289334e-02 7.63660252e-01 2.65632719e-01
1.59055009e-01 -4.56176281e-01 -9.69915152e-01 8.56799006e-01
4.95242745e-01 -4.17298555e-01 2.55373091e-01 8.68309975e-01
-1.23441793e-01 -2.90621191e-01 -7.27882624e-01 1.43528724e+00
-2.06794925e-02 -9.64011192e-01 -1.71999615e-02 -1.82617784e-01
8.37815762e-01 -3.23140286e-02 -1.17530972e-02 2.43340373e-01
5.91615856e-01 -8.74617040e-01 9.54351306e-01 4.75600332e-01
8.98085654e-01 -7.81327724e-01 8.70167315e-01 7.34953642e-01
-1.10873747e+00 -3.79295439e-01 -9.21766281e-01 -1.81064904e-01
-3.12311381e-01 6.49424851e-01 -6.84016943e-01 2.54015833e-01
1.43330920e+00 6.41440809e-01 -9.26226616e-01 1.07870495e+00
3.38422246e-02 5.01646340e-01 8.23460296e-02 -1.30206928e-01
2.65405744e-01 4.70845625e-02 6.66913912e-02 1.05677247e+00
-1.12656116e-01 -3.71297710e-02 2.28329122e-01 2.88132012e-01
7.07952380e-02 -3.57342720e-01 -3.84705901e-01 -5.15830889e-02
6.69406235e-01 1.08537352e+00 -1.00026584e+00 -3.88633251e-01
-2.08103538e-01 7.35992312e-01 2.65645534e-01 3.37111324e-01
-8.09226632e-01 -3.68151248e-01 6.73683882e-01 -8.62111617e-03
3.66427243e-01 -2.99029723e-02 -2.60719538e-01 -1.20383632e+00
-2.35802773e-03 -8.71909320e-01 6.63426399e-01 -4.33485955e-01
-1.99763215e+00 8.76396894e-01 2.03322813e-01 -1.10280645e+00
-1.75888091e-01 -4.13570285e-01 -3.20823997e-01 7.80294895e-01
-1.60528266e+00 -9.77258325e-01 -9.61558998e-01 1.08017302e+00
8.60459030e-01 -2.17495069e-01 4.42052096e-01 2.87232608e-01
-5.08890152e-01 3.92668158e-01 3.01326662e-01 1.17619097e-01
8.54534984e-01 -1.20792961e+00 3.40668052e-01 8.03029776e-01
-1.56946138e-01 4.53075737e-01 7.34275222e-01 -6.57804668e-01
-8.14311862e-01 -1.43298757e+00 3.67434829e-01 -4.37148571e-01
5.72063863e-01 -2.38773301e-01 -9.02418435e-01 2.73455441e-01
-1.77783772e-01 3.20217073e-01 4.17191267e-01 -1.25136256e-01
-1.92289203e-01 -5.04862547e-01 -1.35559440e+00 1.58088759e-01
1.16803908e+00 -4.11081761e-01 -9.95759428e-01 2.85173416e-01
4.86797571e-01 3.08265649e-02 -5.01025438e-01 2.18414247e-01
4.79278535e-01 -9.30864811e-01 8.02411079e-01 -6.02368474e-01
-1.83525961e-02 -6.37052238e-01 -3.05100292e-01 -1.57651615e+00
-6.37806237e-01 3.70739140e-02 2.29494452e-01 1.38203251e+00
-3.29119891e-01 -5.53391695e-01 5.60284734e-01 4.37801212e-01
1.27749503e-01 -1.45333007e-01 -6.65417314e-01 -8.93226445e-01
1.39317855e-01 -2.08985373e-01 8.17815006e-01 8.43582690e-01
-5.42647123e-01 1.90207392e-01 -1.74709126e-01 2.11489439e-01
1.02656233e+00 4.11706984e-01 6.54490948e-01 -1.51493323e+00
1.71333417e-01 -3.01437229e-01 -7.09293246e-01 -7.75105417e-01
-2.73825023e-02 -3.09654713e-01 4.76352841e-01 -1.23738742e+00
4.12257314e-01 -9.14961100e-01 -3.69739652e-01 1.58556759e-01
-4.29395318e-01 3.97639155e-01 1.72554567e-01 6.27290964e-01
-7.44409740e-01 3.95644456e-01 1.27163863e+00 -4.98845190e-01
1.94507703e-01 1.92302510e-01 -9.63594973e-01 8.93288314e-01
6.23040617e-01 -4.16099101e-01 -8.81929934e-01 -5.13178766e-01
-3.02601695e-01 -4.17674422e-01 1.77213892e-01 -9.59465206e-01
2.09944338e-01 -5.07152796e-01 6.15186691e-01 -7.14619398e-01
-1.22681998e-01 -1.07768595e+00 2.30566949e-01 4.87532645e-01
6.56499788e-02 -4.17736061e-02 1.06983565e-01 8.41473758e-01
-5.05698800e-01 -1.15134813e-01 1.03357494e+00 -4.22571570e-01
-1.28970671e+00 6.10263348e-01 -2.33381942e-01 3.38945538e-01
1.35057783e+00 -4.53997105e-01 -6.33496523e-01 1.20903812e-01
-2.27992758e-01 1.80740759e-01 5.88043809e-01 6.78289711e-01
4.30839628e-01 -1.14087498e+00 -5.61423063e-01 2.76236564e-01
5.02137601e-01 3.52261752e-01 5.74902833e-01 1.07926078e-01
-3.69917691e-01 2.70152241e-01 -6.86407387e-01 -9.93642747e-01
-1.27528048e+00 1.08513999e+00 2.35622540e-01 2.05585957e-01
-2.73497283e-01 9.95183766e-01 7.61247635e-01 -2.19908908e-01
4.62941855e-01 -3.36142600e-01 -4.24416959e-01 5.19631267e-01
7.60069907e-01 4.02930856e-01 4.09217365e-02 -5.65293550e-01
-5.36481857e-01 7.35479295e-01 -3.05595964e-01 4.19678539e-01
1.25877976e+00 -5.28720856e-01 5.34206592e-02 7.40856230e-01
8.38819623e-01 -5.16949892e-01 -1.51795900e+00 -2.96152145e-01
-2.57402174e-02 -8.15998852e-01 -1.24715686e-01 -9.36958909e-01
-1.03517759e+00 9.42404151e-01 1.05647671e+00 4.25094545e-01
1.34135509e+00 -1.74164139e-02 2.66833246e-01 1.68864280e-01
9.17365968e-01 -1.17098558e+00 -2.52603013e-02 2.50966460e-01
6.88894868e-01 -1.50471878e+00 -3.62175405e-02 -1.99997202e-01
-7.58903325e-01 8.59340429e-01 7.46768832e-01 9.19476226e-02
7.70279944e-01 1.34144962e-01 5.18504858e-01 -2.45509043e-01
-6.84533417e-01 -3.79016325e-02 7.32443109e-02 1.07298994e+00
2.37419065e-02 3.38631630e-01 -4.61060479e-02 4.15908098e-01
-7.90258348e-02 1.47849575e-01 3.29113632e-01 8.90932202e-01
-5.41610599e-01 -8.24626625e-01 -6.11942887e-01 4.98139769e-01
-2.90675491e-01 2.68777013e-01 -3.45191747e-01 6.56147480e-01
8.23929667e-01 1.21331918e+00 3.21324468e-01 -2.81960338e-01
3.59637529e-01 -3.24574351e-01 3.71458590e-01 -3.23111057e-01
-3.43826205e-01 -3.60338151e-01 -3.03673744e-01 -3.07233572e-01
-6.56186044e-01 -7.44734943e-01 -9.16843593e-01 -5.41808784e-01
-2.90957123e-01 5.14179356e-02 5.10612071e-01 7.42422879e-01
1.15347635e-02 5.73656917e-01 7.88610995e-01 -5.84004223e-01
-5.02020836e-01 -9.84048009e-01 -9.59717155e-01 7.55502045e-01
3.92405748e-01 -1.14038014e+00 -4.63738918e-01 4.40718293e-01] | [9.684576034545898, 2.102139949798584] |
d50f521b-b03e-4bc3-9c8c-cb7fd20d6ca8 | sample-efficient-deep-reinforcement-learning-3 | 2301.12579 | null | https://arxiv.org/abs/2301.12579v2 | https://arxiv.org/pdf/2301.12579v2.pdf | Sample Efficient Deep Reinforcement Learning via Local Planning | The focus of this work is sample-efficient deep reinforcement learning (RL) with a simulator. One useful property of simulators is that it is typically easy to reset the environment to a previously observed state. We propose an algorithmic framework, named uncertainty-first local planning (UFLP), that takes advantage of this property. Concretely, in each data collection iteration, with some probability, our meta-algorithm resets the environment to an observed state which has high uncertainty, instead of sampling according to the initial-state distribution. The agent-environment interaction then proceeds as in the standard online RL setting. We demonstrate that this simple procedure can dramatically improve the sample cost of several baseline RL algorithms on difficult exploration tasks. Notably, with our framework, we can achieve super-human performance on the notoriously hard Atari game, Montezuma's Revenge, with a simple (distributional) double DQN. Our work can be seen as an efficient approximate implementation of an existing algorithm with theoretical guarantees, which offers an interpretation of the positive empirical results. | ['Csaba Szepesvari', 'Botao Hao', 'Nived Rajaraman', 'Nevena Lazic', 'Sridhar Thiagarajan', 'Dong Yin'] | 2023-01-29 | null | null | null | null | ['montezumas-revenge'] | ['playing-games'] | [-8.03561136e-02 4.54564303e-01 -2.78473884e-01 4.90381382e-02
-1.12744749e+00 -6.33235157e-01 6.74780846e-01 6.29754215e-02
-8.68080497e-01 1.26322186e+00 7.45406076e-02 -3.89350235e-01
-2.90035129e-01 -9.14708376e-01 -1.06830823e+00 -8.51669610e-01
-6.14594877e-01 1.15593946e+00 1.25041276e-01 -2.53421307e-01
2.11579114e-01 2.92664647e-01 -1.36630929e+00 -3.98310930e-01
6.07785165e-01 7.39575267e-01 4.49481577e-01 8.65505993e-01
2.93177754e-01 8.37489128e-01 -4.34886813e-01 6.90256432e-02
4.67935711e-01 -6.01759315e-01 -1.06833005e+00 1.22084834e-01
-3.80096793e-01 -9.42778468e-01 -2.93091655e-01 1.14135075e+00
4.49506313e-01 4.68821049e-01 2.42069438e-01 -1.22582889e+00
1.93606898e-01 1.12390792e+00 -4.68445778e-01 -7.31315911e-02
3.96460712e-01 4.00329620e-01 1.06271601e+00 -1.06961332e-01
8.32621932e-01 1.16501117e+00 1.77771360e-01 3.46694589e-01
-1.29321146e+00 -3.81311655e-01 1.94997534e-01 2.73634009e-02
-1.05148959e+00 -2.72832394e-01 3.48580301e-01 1.22542754e-01
7.80258834e-01 -1.22235633e-01 8.06088746e-01 1.14072335e+00
2.60552406e-01 1.14766955e+00 1.46330643e+00 -4.44935083e-01
9.60823536e-01 -1.59952462e-01 -2.53238261e-01 7.58089244e-01
2.30882674e-01 7.13991702e-01 -5.51927805e-01 -4.28254157e-01
7.09495306e-01 -1.71461821e-01 -2.03017920e-01 -8.49913955e-01
-1.10054862e+00 9.89479125e-01 2.49406829e-01 -4.31897901e-02
-4.96830553e-01 7.03162313e-01 4.08632487e-01 6.50398016e-01
2.47321092e-02 4.87510890e-01 -3.70306283e-01 -6.65390611e-01
-5.59494972e-01 7.48595774e-01 1.18739402e+00 9.01918471e-01
8.39804232e-01 -6.91919923e-02 1.03900284e-01 1.81957632e-01
1.20318204e-01 5.01898527e-01 2.62228966e-01 -1.59347069e+00
4.43038732e-01 -3.02349657e-01 8.84502351e-01 -3.66131186e-01
-5.28730512e-01 -4.30392563e-01 -4.15854007e-01 6.23543143e-01
5.63748062e-01 -5.30479014e-01 -4.45135325e-01 2.04837298e+00
5.73077381e-01 2.74231166e-01 3.05263668e-01 7.05857635e-01
-3.63674015e-01 5.60843766e-01 -2.37548724e-01 -3.74301732e-01
1.00238621e+00 -8.61913145e-01 -5.49381793e-01 -3.63341153e-01
7.20029533e-01 3.59245166e-02 1.13395429e+00 7.17735350e-01
-1.24644983e+00 -3.42992805e-02 -1.01733792e+00 2.49107167e-01
2.56155208e-02 -5.17116964e-01 1.02453339e+00 5.26626706e-01
-1.07187521e+00 1.02648938e+00 -1.36773098e+00 -1.96960941e-01
3.59275818e-01 2.96533436e-01 -1.72379240e-01 -5.47023751e-02
-9.59468365e-01 9.28405046e-01 7.04587042e-01 -1.32188827e-01
-1.71227646e+00 -2.74503917e-01 -8.14052820e-01 -5.37952082e-03
1.28739178e+00 -5.36610484e-01 1.97387493e+00 -6.65153146e-01
-2.15156460e+00 2.62526900e-01 -4.76656035e-02 -9.49173570e-01
8.01357388e-01 -2.99636006e-01 5.00078201e-01 1.17642842e-01
6.84297830e-02 5.12731969e-01 5.58258593e-01 -1.23458779e+00
-7.89461076e-01 -3.35972130e-01 4.97856379e-01 5.50872624e-01
3.36510420e-01 -4.10345495e-01 -7.00792074e-02 4.95610610e-02
-2.47968689e-01 -1.13810253e+00 -8.67486238e-01 -2.70593613e-01
-2.01996669e-01 -2.68077016e-01 -1.79279968e-02 -4.06451598e-02
7.10061610e-01 -1.73595929e+00 2.68616289e-01 3.64500612e-01
2.07766548e-01 -2.94579864e-01 -1.46805480e-01 6.72897637e-01
5.18836379e-01 -2.18334630e-01 -3.10314000e-01 -4.45450723e-01
5.04662037e-01 4.55325663e-01 -5.14294565e-01 6.09670103e-01
-4.93396372e-01 1.05931437e+00 -1.44845271e+00 -2.66946673e-01
8.81405249e-02 -2.95439541e-01 -8.03558230e-01 2.42458314e-01
-5.87382078e-01 4.01921272e-01 -6.32737517e-01 1.08379506e-01
3.95352274e-01 -1.45812616e-01 6.13288522e-01 6.99254215e-01
-1.11884214e-01 4.59489197e-01 -1.32845628e+00 2.11355591e+00
-4.89666402e-01 2.49104932e-01 2.41867632e-01 -8.32438171e-01
1.87543154e-01 1.97565913e-01 3.13793957e-01 -6.13322020e-01
1.95746079e-01 2.20904097e-01 -6.21331111e-02 -1.17774330e-01
6.75393641e-01 -5.89978993e-02 -2.78344274e-01 8.88288677e-01
-6.92080483e-02 -2.91514039e-01 7.04149157e-02 1.87257320e-01
1.27743387e+00 5.55057585e-01 6.37484968e-01 -2.83708543e-01
-8.16714540e-02 1.28617033e-01 3.68036509e-01 1.59187078e+00
-2.57673383e-01 -6.16245084e-02 8.75060916e-01 -2.75496781e-01
-1.01327074e+00 -1.15478969e+00 2.86045641e-01 9.88950789e-01
2.15839148e-01 -4.65644807e-01 -7.92711854e-01 -6.04722798e-01
-1.29721224e-01 8.82882595e-01 -6.76782906e-01 9.85250175e-02
-6.25304580e-01 -2.54864633e-01 4.02054548e-01 3.07440579e-01
6.09080970e-01 -1.31109405e+00 -1.29939032e+00 4.07229692e-01
-4.68054004e-02 -6.43072963e-01 -2.51853704e-01 5.95734715e-01
-8.26341271e-01 -9.09455657e-01 -3.64267230e-01 -3.35317940e-01
1.80449724e-01 1.24299042e-01 1.02927315e+00 -2.05359206e-01
2.25294814e-01 5.32462716e-01 -2.12051108e-01 -2.72526473e-01
-4.24826652e-01 1.79958731e-01 1.51403710e-01 -5.72939277e-01
-2.97811031e-02 -5.97848535e-01 -5.81425309e-01 -1.27002820e-01
-7.80640781e-01 -3.40564661e-02 4.08777893e-01 1.00895011e+00
5.73046446e-01 3.57830197e-01 2.94744432e-01 -8.80580962e-01
6.68204606e-01 -6.49676263e-01 -1.22829425e+00 -8.92622545e-02
-4.54419345e-01 5.76508939e-01 7.20694780e-01 -2.90333867e-01
-9.20327008e-01 1.93902645e-02 7.69201294e-02 -1.64296404e-01
-1.00074850e-01 5.20999372e-01 1.84310377e-02 4.27596197e-02
5.34483671e-01 4.74415064e-01 3.51753175e-01 -1.50399461e-01
4.06880528e-01 2.18617037e-01 5.14347315e-01 -1.21932161e+00
8.14772427e-01 4.88967746e-01 2.13498697e-01 -4.82751578e-01
-7.74859428e-01 2.37479527e-02 -2.31571347e-01 2.91494261e-02
4.22060281e-01 -8.09748828e-01 -1.48142111e+00 3.46753865e-01
-7.28995502e-01 -1.18562269e+00 -6.62244380e-01 3.80359352e-01
-1.25094891e+00 4.04872030e-01 -5.33264220e-01 -1.29972112e+00
1.76216468e-01 -1.20433760e+00 9.58223164e-01 1.54577672e-01
8.36421102e-02 -7.86051095e-01 3.11304182e-01 -1.29002944e-01
3.10402691e-01 2.27410644e-01 4.98451740e-01 -5.35374463e-01
-9.62560236e-01 3.16780746e-01 3.23155135e-01 -1.15087934e-01
-2.59983122e-01 -6.79334819e-01 -7.54640639e-01 -6.92206562e-01
8.71560872e-02 -9.06905472e-01 6.60474837e-01 4.09529865e-01
1.20812237e+00 -5.23641169e-01 -2.24669591e-01 3.29797477e-01
1.40392780e+00 2.79881656e-01 5.14319539e-01 7.05026448e-01
3.71705480e-02 2.69545674e-01 1.00346768e+00 1.02690685e+00
5.40939510e-01 7.10311651e-01 5.99692523e-01 4.97299612e-01
6.42195225e-01 -5.38843691e-01 6.40567005e-01 3.90831791e-02
2.64444221e-02 -9.44964215e-02 -5.06991386e-01 4.89809632e-01
-2.04665828e+00 -1.04286182e+00 6.43236816e-01 2.48691678e+00
1.00630200e+00 2.41395578e-01 4.35143232e-01 -2.06757024e-01
1.48170933e-01 2.59704649e-01 -8.56109321e-01 -3.73787850e-01
2.37247497e-01 5.29023409e-01 8.58584762e-01 9.59233701e-01
-8.33290756e-01 1.05430806e+00 6.35018539e+00 8.73556256e-01
-4.61534441e-01 2.56935805e-01 4.17383999e-01 -3.02953213e-01
-2.47185901e-01 1.64289504e-01 -6.42847359e-01 3.56643289e-01
1.27916491e+00 -3.57862920e-01 1.21988261e+00 1.23126709e+00
1.36792213e-01 -7.04356372e-01 -1.48569214e+00 7.28531539e-01
-4.87685621e-01 -1.28564453e+00 -6.22494698e-01 5.32530308e-01
5.99243283e-01 1.19494088e-01 -3.86508442e-02 6.69609666e-01
1.25870526e+00 -9.71384227e-01 8.24922860e-01 6.61739558e-02
5.61900139e-01 -1.26473844e+00 5.91620266e-01 7.70342946e-01
-8.99161935e-01 -3.28630954e-01 -5.00196159e-01 -3.45097065e-01
2.19330639e-01 1.57801971e-01 -9.26939547e-01 5.42340517e-01
4.41121012e-01 8.51188377e-02 2.16999054e-01 8.75182390e-01
-3.43457103e-01 5.30429184e-01 -7.86988854e-01 -2.56213903e-01
6.86766803e-01 -4.63134199e-01 6.24412894e-01 6.04055822e-01
2.15713620e-01 1.77898228e-01 6.80212140e-01 8.46673548e-01
2.92689335e-02 -4.30451691e-01 -8.64044249e-01 6.83869720e-02
7.07610428e-01 8.53296220e-01 -5.08850515e-01 -2.11189523e-01
8.67490023e-02 9.46080446e-01 6.49980366e-01 1.66319832e-01
-8.48271728e-01 -2.64769047e-01 6.10041380e-01 -4.15782750e-01
3.90771031e-01 -3.62851024e-01 5.97626083e-02 -1.02074003e+00
1.07851140e-02 -1.06546474e+00 3.86506557e-01 -4.14695174e-01
-7.78692663e-01 2.50268400e-01 1.66930526e-01 -8.36723268e-01
-8.11977625e-01 -2.84175724e-01 -3.10662180e-01 4.59317923e-01
-1.49337482e+00 -5.59357822e-01 9.33315977e-02 4.32333112e-01
4.72046256e-01 8.43749717e-02 7.93330252e-01 -4.72219735e-01
-2.73200423e-01 4.48568434e-01 3.76255631e-01 -2.99052894e-01
2.24432245e-01 -1.75576758e+00 3.32038164e-01 7.60164440e-01
-1.18373754e-02 2.98858762e-01 9.99660909e-01 -5.49008250e-01
-1.78667855e+00 -7.11889207e-01 1.95624128e-01 -2.72754192e-01
9.79338288e-01 -4.78257447e-01 -4.91598755e-01 1.08943379e+00
3.09738278e-01 -2.88934529e-01 7.02375844e-02 1.96333230e-01
7.78152198e-02 5.32102734e-02 -1.19687569e+00 9.74849999e-01
1.17245567e+00 -2.32712954e-01 -6.38074934e-01 5.18428326e-01
8.12993824e-01 -9.98935997e-01 -4.83584255e-01 -2.36276999e-01
3.94082963e-01 -9.64108706e-01 7.36952603e-01 -6.36516988e-01
1.65252119e-01 -2.77319074e-01 -1.68594256e-01 -1.56962669e+00
-2.89316531e-02 -1.52693665e+00 -3.65714103e-01 5.31490266e-01
1.23494975e-01 -9.23854947e-01 9.01444495e-01 4.73679721e-01
1.13004759e-01 -6.26253426e-01 -1.21275210e+00 -9.88257408e-01
2.52904654e-01 -5.06225407e-01 7.18546808e-01 2.11768255e-01
2.06730142e-01 -1.22593986e-02 -5.64830005e-01 4.85336602e-01
1.00702691e+00 2.19398379e-01 1.04997957e+00 -8.16371620e-01
-1.04313254e+00 -2.34060809e-01 1.67805448e-01 -1.44531226e+00
4.15146023e-01 -5.41403592e-01 4.38892424e-01 -1.13586068e+00
1.90486982e-01 -6.21605277e-01 -1.36026576e-01 3.39149058e-01
2.07164466e-01 -5.02568543e-01 3.28557700e-01 -1.62659481e-01
-1.07098913e+00 7.99175382e-01 1.28302264e+00 2.69011915e-01
-4.02187020e-01 1.89870179e-01 -5.68602145e-01 5.72133124e-01
8.08822215e-01 -4.86945540e-01 -6.33422256e-01 -1.72217980e-01
3.84669691e-01 6.84563816e-01 2.37647027e-01 -8.48209679e-01
3.31325352e-01 -5.54851234e-01 -1.75839081e-01 -5.37479877e-01
3.37810755e-01 -7.27190435e-01 6.92108795e-02 1.00213170e+00
-6.82104945e-01 -6.14946671e-02 -7.55749121e-02 8.75980914e-01
3.98248643e-01 -4.52798635e-01 7.29179442e-01 -3.11076373e-01
-5.98074675e-01 3.18501055e-01 -6.65863752e-01 2.41222963e-01
1.02043998e+00 1.00421533e-01 -1.79556906e-01 -7.46995866e-01
-5.32502890e-01 6.05156541e-01 5.63063383e-01 -3.48349541e-01
3.96505803e-01 -1.05586660e+00 -4.42821324e-01 -1.30281284e-01
-2.47821927e-01 4.08193439e-01 7.41295069e-02 7.24738717e-01
-3.30191493e-01 3.34830105e-01 -1.58941194e-01 -2.91221410e-01
-6.06366217e-01 6.06362879e-01 5.41038513e-01 -7.34385848e-01
-9.14297223e-01 5.45592248e-01 3.65230516e-02 -4.78838891e-01
3.39548349e-01 -4.34049666e-01 2.78136641e-01 -3.20710987e-01
7.21742153e-01 4.86081094e-01 -3.55197757e-01 3.42935652e-01
-4.70975302e-02 -2.93617342e-02 -2.07879692e-01 -8.40943336e-01
1.42079234e+00 -2.46812880e-01 1.55850261e-01 5.21805167e-01
7.70160139e-01 -6.01486079e-02 -1.79641151e+00 -3.75865698e-01
1.35474011e-01 -4.16485995e-01 -4.71193530e-02 -6.62537396e-01
-5.85411370e-01 4.04351860e-01 3.32177192e-01 1.71722740e-01
8.30830395e-01 -9.99031588e-02 6.85123622e-01 1.12322938e+00
1.30792320e+00 -1.16800821e+00 -5.13708889e-02 7.01752722e-01
6.16440058e-01 -1.06624985e+00 -7.46713206e-02 4.08555180e-01
-8.22871864e-01 8.67979288e-01 3.65319520e-01 -4.77015406e-01
1.47458985e-01 4.19078887e-01 -6.07535720e-01 -1.14526814e-02
-1.08414543e+00 -3.68700832e-01 -6.83535159e-01 6.69888377e-01
-4.65710968e-01 3.11079144e-01 1.15202613e-01 3.58939797e-01
-4.22488213e-01 1.42247304e-01 9.54943776e-01 1.17619860e+00
-8.56288850e-01 -1.12770128e+00 -1.59157619e-01 1.32640645e-01
-3.11495125e-01 1.61926016e-01 1.29869461e-01 1.10576010e+00
-6.42980397e-01 8.61287057e-01 7.15833455e-02 2.17289060e-01
-2.20069468e-01 -8.91732723e-02 7.97583878e-01 -4.59948331e-01
-2.14483321e-01 4.09116074e-02 2.19942182e-01 -1.30790722e+00
-1.11748599e-01 -7.57965088e-01 -1.42530262e+00 -5.42441905e-01
6.75802752e-02 3.79711509e-01 4.40038651e-01 1.23855257e+00
2.13188186e-01 4.48606722e-02 8.34765255e-01 -1.02420282e+00
-1.36981571e+00 -5.84034860e-01 -7.92819560e-01 -1.74355447e-01
5.72497129e-01 -8.01711619e-01 -3.36067975e-01 -6.69808507e-01] | [4.155027389526367, 2.054785966873169] |
1b5a36af-8909-41a7-8140-66c3b86e7223 | noiser-noise-is-all-you-need-for-enhancing | 2211.04700 | null | https://arxiv.org/abs/2211.04700v2 | https://arxiv.org/pdf/2211.04700v2.pdf | NoiSER: Noise is All You Need for Low-Light Image Enhancement | In this paper, we present an embarrassingly simple yet effective solution to a seemingly impossible mission, low-light image enhancement (LLIE) without access to any task-related data. The proposed solution, Noise SElf-Regression (NoiSER), simply learns a convolutional neural network equipped with a instance-normalization layer by taking a random noise image, $\mathcal{N}(0,\sigma^2)$ for each pixel, as both input and output for each training pair, and then the low-light image is fed to the learned network for predicting the normal-light image. Technically, an intuitive explanation for its effectiveness is as follows: 1) the self-regression reconstructs the contrast between adjacent pixels of the input image, 2) the instance-normalization layers may naturally remediate the overall magnitude/lighting of the input image, and 3) the $\mathcal{N}(0,\sigma^2)$ assumption for each pixel enforces the output image to follow the well-known gray-world hypothesis \cite{Gary-world_Hypothesis} when the image size is big enough, namely, the averages of three RGB components of an image converge to the same value. Compared to existing SOTA LLIE methods with access to different task-related data, NoiSER is surprisingly highly competitive in enhancement quality, yet with a much smaller model size, and much lower training and inference cost. With only $\sim$ 1K parameters, NoiSER realizes about 1 minute for training and 1.2 ms for inference with 600x400 resolution on RTX 2080 Ti. As a bonus, NoiSER possesses automated over-exposure suppression ability and shows excellent performance on over-exposed photos. | ['Shuicheng Yan', 'Yi Yang', 'Mingliang Xu', 'Xiaojie Jin', 'Suiyi Zhao', 'Zhao Zhang'] | 2022-11-09 | null | null | null | null | ['low-light-image-enhancement'] | ['computer-vision'] | [ 5.37545562e-01 -2.02062324e-01 3.85145247e-01 -5.03132224e-01
-5.10843575e-01 -1.98203370e-01 1.01024829e-01 -1.91188514e-01
-7.15444863e-01 7.30421782e-01 -6.23950303e-01 -3.99447411e-01
-1.52504118e-02 -9.64631855e-01 -8.77691031e-01 -1.00630176e+00
3.39577764e-01 -3.08639944e-01 1.73641518e-01 -2.55476922e-01
-3.38850506e-02 4.83402967e-01 -1.96723402e+00 1.32088261e-02
7.48632252e-01 1.63437307e+00 2.78274983e-01 7.70145714e-01
-5.28330095e-02 8.64742339e-01 -6.65409207e-01 -4.01005059e-01
4.97620225e-01 -3.22299451e-01 -4.28535640e-01 -1.77124679e-01
7.00745761e-01 -2.35472560e-01 -5.38557649e-01 1.52512932e+00
6.49368227e-01 2.78164864e-01 2.39381209e-01 -1.09872997e+00
-5.69908679e-01 9.93902013e-02 -6.43771768e-01 1.53535932e-01
-2.06520855e-01 6.60694182e-01 5.44384837e-01 -5.71761668e-01
2.88684130e-01 8.01285028e-01 6.46993041e-01 5.84339857e-01
-1.35928297e+00 -9.65704560e-01 -1.27110153e-01 1.60501242e-01
-1.52253950e+00 -2.44084224e-01 7.17719197e-01 6.68274760e-02
8.10437977e-01 3.76852155e-01 4.95232731e-01 8.48410189e-01
2.81160176e-01 1.72054872e-01 1.72501469e+00 -3.58741224e-01
2.45276541e-01 7.94123337e-02 -6.93532601e-02 5.31140924e-01
1.69816911e-01 1.76632419e-01 -3.98040414e-01 2.78238416e-01
9.54205453e-01 1.04687013e-01 -3.32721978e-01 3.62718910e-01
-8.45855176e-01 2.32367203e-01 7.01718688e-01 2.56865650e-01
-2.83707380e-01 4.08361882e-01 9.38925147e-02 4.05612916e-01
2.52702832e-01 3.20099473e-01 -4.94483471e-01 -3.21890227e-02
-7.98377454e-01 -5.47401868e-02 3.54838222e-01 6.61051095e-01
1.19880235e+00 4.21502620e-01 -1.52109250e-01 7.09870577e-01
-1.94389254e-01 9.61288810e-01 2.87452698e-01 -1.09930539e+00
2.53284335e-01 6.52099073e-01 8.28558430e-02 -9.60617423e-01
-5.86934030e-01 -4.33483779e-01 -1.39920282e+00 8.51814866e-01
6.13268137e-01 -2.63718039e-01 -1.03303695e+00 1.83676028e+00
-5.78636117e-03 4.02958356e-02 9.86440331e-02 8.81320715e-01
8.25647593e-01 6.17398739e-01 7.00137857e-03 -4.13546264e-01
1.45857871e+00 -5.71048319e-01 -6.20671928e-01 -4.14418757e-01
-9.13685262e-02 -8.25793922e-01 1.56620479e+00 5.67915082e-01
-1.27737963e+00 -9.62998807e-01 -1.13286841e+00 -2.94048041e-01
-4.19894606e-01 2.69793928e-01 5.47477484e-01 5.73926151e-01
-1.08109546e+00 7.41702795e-01 -4.78569448e-01 7.86276758e-02
3.31413299e-01 4.75533485e-01 -2.59027600e-01 -2.48218626e-01
-1.01471925e+00 6.89546883e-01 6.65654466e-02 3.86075139e-01
-6.70922101e-01 -6.43711746e-01 -5.21258295e-01 7.68845007e-02
4.84004766e-01 -3.49817932e-01 8.85538936e-01 -1.29861653e+00
-1.56155407e+00 9.04072821e-01 -1.70204997e-01 -2.33231217e-01
3.41093391e-01 1.20764762e-01 -6.73993766e-01 1.07570982e-03
-1.76779136e-01 6.04767084e-01 1.23972476e+00 -1.33298719e+00
-5.64033329e-01 -3.76060903e-01 5.98540790e-02 -6.94948155e-03
-1.83238998e-01 3.95429172e-02 -6.38307214e-01 -5.37039161e-01
2.93428600e-01 -6.49240673e-01 -2.23611102e-01 2.65773296e-01
-2.99502462e-01 1.75725609e-01 5.83442688e-01 -4.10900533e-01
1.20027542e+00 -2.32622790e+00 -5.03560960e-01 3.46656978e-01
3.09099227e-01 3.57122779e-01 1.23730535e-02 -4.08169359e-01
-4.02269512e-01 -4.66180965e-02 -3.18168461e-01 -8.69341567e-02
-2.85810918e-01 2.20776036e-01 -8.57935697e-02 4.23532724e-01
1.22232735e-01 7.84642398e-01 -7.17015147e-01 -2.19125941e-01
4.16571796e-01 7.45156765e-01 -1.72252417e-01 2.17619032e-01
-8.51580426e-02 2.43974552e-01 2.84978598e-02 4.38117325e-01
1.08532119e+00 -3.10966700e-01 1.84445351e-01 -7.64663517e-01
-3.45471770e-01 -1.25078529e-01 -1.39266455e+00 1.28447509e+00
-4.83038068e-01 8.43311965e-01 2.18210414e-01 -7.64247954e-01
1.08267629e+00 -1.33839240e-02 3.13217551e-01 -1.36597741e+00
4.51988697e-01 9.14422497e-02 -1.60689298e-02 -5.12133658e-01
2.04863861e-01 -2.45532736e-01 1.96631446e-01 3.81755531e-01
-1.22040622e-02 -2.47735396e-01 -2.20917948e-02 -2.55117089e-01
1.02266192e+00 7.09982961e-02 1.01059712e-01 -7.32086087e-03
5.14087617e-01 -3.42470199e-01 4.60492909e-01 9.48204279e-01
-2.59069085e-01 6.27497196e-01 4.40217853e-01 -5.71352005e-01
-1.08025646e+00 -1.03566241e+00 -1.90192431e-01 8.54753792e-01
3.06738168e-01 -6.12064358e-03 -9.32178557e-01 -2.73087829e-01
-3.88205320e-01 7.12440431e-01 -4.95805979e-01 -1.53445870e-01
-6.70508265e-01 -9.80542302e-01 5.16379893e-01 4.54140812e-01
9.70816135e-01 -1.13611281e+00 -8.12065899e-01 -8.88467394e-03
-1.31124988e-01 -1.13411176e+00 -1.72895715e-01 6.93901837e-01
-6.62697852e-01 -1.03741777e+00 -3.20002317e-01 -5.30718625e-01
8.85590613e-01 4.29055125e-01 1.07933712e+00 2.68581569e-01
-6.07273042e-01 4.33413945e-02 1.69700861e-01 -3.30040395e-01
-5.85857928e-02 -6.25504255e-01 -7.55154015e-03 1.32392824e-01
4.21118617e-01 -7.60774493e-01 -9.82803524e-01 3.50021571e-01
-1.02803755e+00 -8.04946292e-03 5.98135889e-01 8.05428982e-01
9.20701027e-01 5.80533803e-01 8.39628093e-03 -5.63082039e-01
1.93446726e-01 1.70012429e-01 -8.92769873e-01 2.09789276e-01
-7.86803484e-01 -4.15704995e-02 9.05448377e-01 -5.77601552e-01
-1.20258904e+00 1.15045737e-02 -1.82786897e-01 -5.46680987e-01
-3.34244192e-01 -1.74742639e-01 -2.96760082e-01 -1.66258752e-01
1.00415087e+00 3.57312202e-01 6.47815466e-02 -3.22075158e-01
2.05250919e-01 4.54260141e-01 1.09510994e+00 -4.31460559e-01
9.62772489e-01 7.50795186e-01 2.60253400e-01 -7.08245873e-01
-8.04491341e-01 1.22041814e-02 -3.56453538e-01 -2.66022205e-01
9.21531856e-01 -9.50863957e-01 -1.37979019e+00 7.46054888e-01
-7.82656550e-01 -5.65050185e-01 -6.02762341e-01 2.81334192e-01
-2.60696620e-01 1.58655077e-01 -4.69437063e-01 -9.40963030e-01
-4.84945059e-01 -1.14535069e+00 5.78398228e-01 6.96114719e-01
3.13254327e-01 -5.24628699e-01 -6.05941892e-01 2.31394455e-01
6.23004079e-01 1.38493180e-01 9.62549984e-01 1.17428407e-01
-6.30203128e-01 -1.70263186e-01 -8.91108930e-01 9.16450322e-01
1.12682410e-01 9.05614533e-03 -1.31983638e+00 -6.85956255e-02
4.50832337e-01 -2.21711412e-01 6.86070502e-01 5.29577971e-01
1.68197155e+00 -1.50805458e-01 2.37909034e-01 9.15607095e-01
1.87332869e+00 2.29105830e-01 1.06520772e+00 3.30604345e-01
4.60450172e-01 2.00479656e-01 3.05026799e-01 3.04268986e-01
1.78521663e-01 4.48413879e-01 6.72994792e-01 -7.24575698e-01
-4.32563603e-01 3.50841172e-02 1.87810570e-01 1.52391225e-01
-2.59191304e-01 -6.29781038e-02 -4.66807336e-01 5.15252016e-02
-1.29959667e+00 -9.10899818e-01 -3.01590651e-01 2.38592243e+00
9.87188458e-01 3.72180462e-01 -3.89621526e-01 2.26066977e-01
5.97902179e-01 1.48874268e-01 -8.55569482e-01 -2.16160685e-01
-6.27579749e-01 7.57369995e-01 8.19944382e-01 2.47030690e-01
-7.84021676e-01 7.44478703e-01 5.99033499e+00 9.87863958e-01
-1.54594970e+00 1.20527312e-01 9.09772575e-01 -1.79737523e-01
-7.85750821e-02 -1.48660034e-01 -5.73480129e-01 6.10424697e-01
6.97180986e-01 1.87291220e-01 8.91083300e-01 6.85822666e-01
2.93030798e-01 -5.20060360e-01 -7.30412841e-01 1.36255014e+00
1.09052211e-01 -1.10045910e+00 -2.93753177e-01 -2.26900011e-01
6.09033406e-01 -4.76772152e-02 2.82355070e-01 1.32870361e-01
2.22364411e-01 -1.11421239e+00 5.57673097e-01 6.90244853e-01
1.29114091e+00 -6.43184483e-01 6.71715498e-01 1.76804915e-01
-9.46250975e-01 -3.46942246e-01 -5.37846923e-01 -1.38375401e-01
-1.68563738e-01 7.78013229e-01 -1.60233378e-01 3.38946700e-01
1.25179636e+00 1.24757148e-01 -6.24805689e-01 5.86910367e-01
-3.64902079e-01 5.27236640e-01 -4.91619468e-01 3.36656243e-01
-2.38272045e-02 -4.08128113e-01 1.28561229e-01 9.01973188e-01
3.35323811e-01 3.81611258e-01 -1.51948392e-01 9.21007156e-01
-2.04853877e-01 -1.72313005e-01 -1.62765250e-01 4.22297716e-01
9.99504030e-02 1.44657063e+00 -7.29101002e-01 -3.03343832e-01
-3.77036542e-01 1.00590229e+00 6.29952103e-02 6.07037067e-01
-8.99653733e-01 -5.92199445e-01 6.38898790e-01 1.49348438e-01
8.24676380e-02 1.06960252e-01 -7.33946323e-01 -9.38117921e-01
2.73849964e-01 -8.57115686e-01 1.57119453e-01 -1.27752936e+00
-1.04171669e+00 8.46275330e-01 -4.94506478e-01 -1.19991529e+00
3.16911131e-01 -8.37276518e-01 -5.50982118e-01 1.02651691e+00
-1.87246442e+00 -8.80186975e-01 -8.38731229e-01 9.94180322e-01
1.13071635e-01 7.67986923e-02 6.65985465e-01 5.04790187e-01
-7.81067133e-01 6.47761166e-01 1.74915064e-02 1.00026332e-01
7.52603590e-01 -1.16664755e+00 1.79100052e-01 8.70752156e-01
-2.50604361e-01 3.81674081e-01 6.17985129e-01 -3.09328794e-01
-1.35488904e+00 -9.87233043e-01 4.40635413e-01 -1.95376799e-01
4.58731800e-01 -2.96890974e-01 -9.90059733e-01 3.08620572e-01
4.46035415e-02 4.56713229e-01 2.72335321e-01 -3.63419652e-01
-4.41303223e-01 -8.17610145e-01 -1.27933002e+00 6.46243811e-01
9.51499939e-01 -5.85851371e-01 4.07500602e-02 2.33469367e-01
5.95366120e-01 -5.67497492e-01 -8.00432622e-01 3.26785326e-01
3.44266236e-01 -1.28525054e+00 1.04875803e+00 -2.15796396e-01
3.76946777e-01 -5.74056625e-01 -3.82267952e-01 -8.16559374e-01
-1.54868603e-01 -6.68436825e-01 2.29481086e-01 1.01529682e+00
3.29116195e-01 -5.95882356e-01 6.14986658e-01 7.49626696e-01
5.19256070e-02 -7.08471119e-01 -1.04356420e+00 -4.84612465e-01
-1.86636865e-01 -7.74367929e-01 5.14750361e-01 6.22285008e-01
-6.38302684e-01 6.05671220e-02 -3.39254707e-01 3.88567418e-01
7.83659160e-01 -5.88855296e-02 6.45939767e-01 -1.03774953e+00
-1.96283296e-01 -2.92117268e-01 -1.93707585e-01 -8.32077444e-01
-2.36340001e-01 -3.39041084e-01 1.44334346e-01 -1.15655613e+00
6.70720562e-02 -6.10499799e-01 -5.09622097e-01 7.95961857e-01
-2.33442128e-01 7.04403877e-01 6.24089725e-02 -1.94551237e-02
-3.78591180e-01 1.87352732e-01 1.25519073e+00 -2.27784440e-01
-1.00931101e-01 6.10912591e-03 -7.19500184e-01 8.32523048e-01
5.12718797e-01 -2.82197922e-01 -1.79038316e-01 -5.52041888e-01
5.39538562e-01 -1.74855784e-01 6.76944852e-01 -1.17767358e+00
3.42042595e-01 -1.96007669e-01 8.09384644e-01 -3.82495463e-01
4.22732353e-01 -1.11303937e+00 3.01220655e-01 1.78661257e-01
3.97323966e-02 -8.11301023e-02 2.51183033e-01 1.43875942e-01
-5.79695851e-02 -1.71604186e-01 1.29488516e+00 -2.34741002e-01
-6.89666986e-01 1.70838922e-01 -7.14474171e-02 4.88748495e-03
8.51053238e-01 -5.10338843e-01 -5.56351542e-01 -2.90419906e-01
-5.51637471e-01 -2.00432599e-01 4.78621662e-01 1.38819497e-02
5.91589868e-01 -1.06789732e+00 -2.96625853e-01 6.05621755e-01
-1.51044458e-01 1.22175798e-01 7.80213535e-01 6.88897729e-01
-3.25709194e-01 -2.35301867e-01 -2.90106922e-01 -5.77636361e-01
-1.01163590e+00 4.71385449e-01 7.18310595e-01 1.46925608e-02
-6.74261987e-01 8.11385691e-01 3.48353535e-02 -1.38725191e-01
2.42011845e-01 -3.11091483e-01 1.54154927e-01 -1.87055767e-01
7.46642113e-01 4.47259545e-01 2.45203301e-01 -3.74371350e-01
-5.42787760e-02 7.31553495e-01 2.39314109e-01 1.59722313e-01
1.39461958e+00 -1.67793706e-01 -3.10135931e-01 2.21686780e-01
1.16164112e+00 -1.87075630e-01 -1.61123216e+00 -2.66929835e-01
-6.82051778e-01 -3.93740356e-01 3.51208270e-01 -9.55518544e-01
-1.42478359e+00 8.96810353e-01 1.16181564e+00 2.00834554e-02
1.80102611e+00 -4.28636044e-01 6.17028177e-01 3.85788441e-01
6.00932576e-02 -1.40227127e+00 1.15771957e-01 2.68452734e-01
5.32198608e-01 -1.32146859e+00 8.70928764e-02 -2.31838867e-01
-5.58504403e-01 1.10166395e+00 9.31081355e-01 2.04393119e-02
6.23492897e-01 6.41884208e-01 3.49566370e-01 -1.94705069e-01
-4.87647921e-01 -2.95364946e-01 2.26094592e-02 5.81114173e-01
3.63932811e-02 -1.35265648e-01 2.34244019e-01 4.88865376e-01
-1.33384094e-01 8.57190192e-02 3.18268001e-01 5.46558440e-01
-3.20320100e-01 -7.25751281e-01 -4.56265479e-01 4.76806551e-01
-4.00104702e-01 -2.77852654e-01 2.46002585e-01 6.90907359e-01
5.92982233e-01 9.51939166e-01 4.13794279e-01 -3.23551476e-01
6.86144352e-01 -8.88661444e-02 3.39925826e-01 -2.06810236e-02
-7.14586318e-01 1.50580466e-01 -5.26078820e-01 -7.86235154e-01
-2.90155053e-01 -1.75432637e-01 -1.37273669e+00 -4.03419405e-01
-1.79762766e-01 -3.17129493e-01 8.88092756e-01 6.95023537e-01
1.44601390e-01 6.98451221e-01 8.21133316e-01 -7.09490001e-01
-3.11395198e-01 -7.30355740e-01 -6.70172036e-01 5.49773395e-01
2.95100361e-01 -3.84117812e-01 -6.07375562e-01 2.93163173e-02] | [10.778270721435547, -2.301703929901123] |
f547b8dd-13eb-444b-ad72-7a9eddfff3cb | internal-video-inpainting-by-implicit-long | 2108.01912 | null | https://arxiv.org/abs/2108.01912v3 | https://arxiv.org/pdf/2108.01912v3.pdf | Internal Video Inpainting by Implicit Long-range Propagation | We propose a novel framework for video inpainting by adopting an internal learning strategy. Unlike previous methods that use optical flow for cross-frame context propagation to inpaint unknown regions, we show that this can be achieved implicitly by fitting a convolutional neural network to known regions. Moreover, to handle challenging sequences with ambiguous backgrounds or long-term occlusion, we design two regularization terms to preserve high-frequency details and long-term temporal consistency. Extensive experiments on the DAVIS dataset demonstrate that the proposed method achieves state-of-the-art inpainting quality quantitatively and qualitatively. We further extend the proposed method to another challenging task: learning to remove an object from a video giving a single object mask in only one frame in a 4K video. | ['Qifeng Chen', 'Tengfei Wang', 'Hao Ouyang'] | 2021-08-04 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Ouyang_Internal_Video_Inpainting_by_Implicit_Long-Range_Propagation_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Ouyang_Internal_Video_Inpainting_by_Implicit_Long-Range_Propagation_ICCV_2021_paper.pdf | iccv-2021-1 | ['video-inpainting'] | ['computer-vision'] | [ 3.33581269e-01 -6.52055070e-02 -5.99621795e-02 -3.26755643e-01
-8.24180901e-01 -4.39318568e-01 2.87996352e-01 -4.42109734e-01
-4.15516883e-01 9.78793323e-01 4.02544588e-02 4.44020331e-02
1.24520093e-01 -4.39447224e-01 -1.25135708e+00 -4.68701690e-01
2.55118776e-03 -7.57622123e-02 1.73914447e-01 1.73451826e-01
1.04228579e-01 6.88974440e-01 -1.11998343e+00 4.42624539e-01
7.05677390e-01 1.01416719e+00 1.99667841e-01 7.76978433e-01
2.05447853e-01 1.16380155e+00 -5.65527618e-01 -3.41563195e-01
6.87275827e-01 -4.50528890e-01 -9.36653793e-01 5.97912908e-01
1.10045958e+00 -8.46588731e-01 -8.22476149e-01 9.23317432e-01
1.39616638e-01 2.90158153e-01 2.64966160e-01 -8.66738021e-01
-5.85396945e-01 -1.93776675e-02 -6.98007822e-01 2.11406916e-01
4.26435143e-01 3.54415029e-01 6.84286296e-01 -8.93196762e-01
9.25363123e-01 1.18111932e+00 7.82378972e-01 4.93523896e-01
-1.42461860e+00 -4.86678272e-01 3.71898115e-01 2.71629810e-01
-1.09184897e+00 -6.72334611e-01 1.04964054e+00 -4.35180575e-01
7.41174281e-01 1.13953859e-01 6.30783916e-01 9.73544776e-01
2.13263601e-01 8.56418252e-01 1.04075491e+00 -3.07135075e-01
5.30446284e-02 -4.23401803e-01 -2.72984743e-01 8.22125852e-01
-4.18951586e-02 2.87897944e-01 -5.76534271e-01 -5.94163798e-02
1.25385559e+00 3.33347581e-02 -5.58098316e-01 -2.17024714e-01
-1.13350308e+00 5.79984426e-01 3.95939201e-01 -1.49035558e-01
-3.06728780e-01 5.45056880e-01 2.52596259e-01 3.62999737e-01
6.84164405e-01 1.96597934e-01 -4.69075501e-01 4.67448123e-02
-1.28702414e+00 2.78434873e-01 6.34862661e-01 1.00623226e+00
8.71074140e-01 2.79445916e-01 -2.32086182e-01 6.60273671e-01
3.44186053e-02 2.49590740e-01 -1.64449643e-02 -1.83366704e+00
3.20158333e-01 -1.22951239e-01 5.01286983e-01 -9.69785154e-01
5.04089966e-02 -3.39122385e-01 -7.22835124e-01 4.87413824e-01
4.37075168e-01 -1.68049425e-01 -9.56138790e-01 1.70261252e+00
3.39488417e-01 7.69168794e-01 -2.18367264e-01 1.08518553e+00
5.16141534e-01 7.67919362e-01 -2.86927670e-01 -5.66650152e-01
9.13083494e-01 -1.44065893e+00 -9.75369692e-01 -3.44198525e-01
6.03627712e-02 -9.04431403e-01 6.46622002e-01 5.29496431e-01
-1.58566451e+00 -7.27884471e-01 -9.48698759e-01 -4.29351002e-01
2.92452872e-01 2.99297478e-02 4.10690099e-01 1.96797363e-02
-1.07196379e+00 8.26043546e-01 -9.31621552e-01 -6.07922897e-02
6.19145691e-01 1.01374120e-01 -5.44026554e-01 -3.06294888e-01
-9.53683913e-01 7.36448526e-01 1.13929436e-01 2.25829095e-01
-1.31303537e+00 -9.27241743e-01 -9.24729109e-01 -3.52465399e-02
4.93739605e-01 -9.89620209e-01 1.06185484e+00 -1.35120368e+00
-1.64632535e+00 6.87019289e-01 -3.33416045e-01 -4.24885482e-01
8.31453919e-01 -7.44472563e-01 -7.23765567e-02 5.89173198e-01
9.71316099e-02 7.37973273e-01 1.36826766e+00 -1.50852072e+00
-3.80142868e-01 7.93289393e-02 2.39929497e-01 1.42847270e-01
1.03502370e-01 3.38583440e-02 -8.38943720e-01 -1.15404546e+00
-4.59531546e-02 -7.51900136e-01 -3.03411007e-01 6.64215088e-01
-1.23669721e-01 4.90550101e-01 1.18871081e+00 -1.10601580e+00
8.75135005e-01 -2.02032590e+00 4.16008234e-01 -3.30724388e-01
2.15357497e-01 3.53622824e-01 -3.25448900e-01 -8.56580585e-02
1.68253891e-02 -3.09381962e-01 -5.00985682e-01 -7.82456756e-01
-4.46818471e-01 4.69193012e-01 -3.79786372e-01 6.04021013e-01
5.03228188e-01 8.72254729e-01 -1.02825236e+00 -3.79500419e-01
4.52192813e-01 8.10851038e-01 -7.87907839e-01 4.00169820e-01
-2.36358598e-01 7.07105696e-01 -9.74646062e-02 5.24419785e-01
9.01170135e-01 -1.41687229e-01 -1.33697763e-01 -2.58198321e-01
3.63285020e-02 -3.12027670e-02 -1.02026868e+00 2.18747473e+00
-3.37247193e-01 9.44704950e-01 5.97334743e-01 -9.50665712e-01
5.66988826e-01 1.51335180e-01 6.37705863e-01 -3.87903720e-01
4.58605140e-02 -5.72391786e-02 -4.47636992e-01 -5.64008713e-01
4.59867120e-01 -1.82924733e-01 4.32524532e-01 1.92333877e-01
3.48041773e-01 -3.10171004e-02 1.96555972e-01 1.84391409e-01
1.17073953e+00 5.83524942e-01 -3.57704517e-03 -5.84071465e-02
4.95105356e-01 -4.19468671e-01 8.35017323e-01 6.50874972e-01
-3.67263198e-01 1.24463844e+00 2.47572586e-01 -6.32928371e-01
-1.05660808e+00 -9.45045233e-01 5.04724868e-03 5.55557549e-01
1.80316135e-01 -2.54318416e-01 -6.90129280e-01 -7.92566955e-01
-1.88260391e-01 2.86060512e-01 -6.36604130e-01 5.83876409e-02
-8.86410356e-01 -2.12787554e-01 3.13261926e-01 4.34564859e-01
6.24751806e-01 -9.69919980e-01 -5.65927982e-01 5.09594023e-01
-5.91196656e-01 -1.59910405e+00 -8.62026513e-01 -1.04648508e-02
-9.51087594e-01 -1.14177358e+00 -8.59145999e-01 -8.81956518e-01
6.62698805e-01 4.49380189e-01 1.29034686e+00 2.57592708e-01
-3.43947947e-01 4.04430687e-01 -1.21640034e-01 3.00276577e-01
-3.72699469e-01 -4.73959416e-01 -2.35486269e-01 1.62017792e-01
-3.09988052e-01 -7.24968731e-01 -7.59875357e-01 1.91616714e-01
-1.19865143e+00 2.05698758e-01 3.51696402e-01 1.01705873e+00
6.62877440e-01 -6.21572845e-02 2.01619834e-01 -5.84791481e-01
1.32566795e-01 -1.29895493e-01 -6.53056979e-01 1.03029311e-01
-1.03695609e-01 -2.04332069e-01 5.82549095e-01 -4.79805738e-01
-1.24795902e+00 3.49538773e-01 -3.10022812e-02 -1.20689440e+00
-1.55353084e-01 5.95361181e-02 -3.93157490e-02 -4.44797337e-01
2.08618000e-01 1.96780220e-01 9.03289244e-02 -4.17316079e-01
5.34692228e-01 -1.04216903e-01 9.37268794e-01 -7.39140093e-01
7.91226804e-01 8.89699996e-01 1.05142087e-01 -6.67444825e-01
-1.02092636e+00 -2.60838598e-01 -8.12070668e-01 -3.20099533e-01
9.00381625e-01 -1.16077495e+00 -4.96189594e-01 5.26994944e-01
-1.45240092e+00 -5.10677397e-01 -3.69834363e-01 4.48628724e-01
-9.27881181e-01 7.47943103e-01 -1.09854150e+00 -3.93806249e-01
-1.63150236e-01 -1.13813567e+00 1.09638417e+00 -7.42885051e-03
1.02532409e-01 -1.10448813e+00 3.35860029e-02 2.71173984e-01
3.04488450e-01 2.88075328e-01 3.36092025e-01 2.26332709e-01
-9.53123987e-01 2.68756717e-01 -3.44258577e-01 7.56976426e-01
2.77020574e-01 1.64171178e-02 -8.89739811e-01 -5.28245389e-01
2.96496153e-01 -2.37051979e-01 1.08257580e+00 6.71572268e-01
1.20485711e+00 -3.58589321e-01 -1.26506329e-01 1.20210540e+00
1.50957143e+00 6.53624237e-02 8.86922479e-01 2.55834013e-01
9.33364689e-01 4.66897011e-01 5.98656952e-01 5.20412326e-01
9.15407017e-03 7.14628160e-01 5.28813720e-01 -3.75208765e-01
-5.63728154e-01 -5.04943095e-02 4.67653692e-01 5.03049254e-01
-1.26288995e-01 -1.29685178e-01 -2.90261924e-01 6.38361156e-01
-1.98035693e+00 -1.07242441e+00 1.25499859e-01 2.00090909e+00
9.73464549e-01 4.77264784e-02 -1.45079508e-01 -8.52957293e-02
6.46240413e-01 3.78707796e-01 -3.83825809e-01 -9.74782705e-02
-2.18482316e-01 2.31739670e-01 5.19207597e-01 1.02563334e+00
-1.23458588e+00 9.82423604e-01 7.03981018e+00 6.70937777e-01
-1.06614792e+00 2.08035216e-01 8.70170236e-01 -2.88276613e-01
-9.74722952e-02 1.74730018e-01 -4.13738549e-01 3.76620620e-01
4.16309595e-01 2.76767641e-01 5.89632690e-01 3.60766441e-01
4.65872347e-01 -9.08592269e-02 -1.15114415e+00 1.00553095e+00
2.71107495e-01 -1.57051277e+00 4.77252435e-03 -1.10709168e-01
1.10069466e+00 -2.35968426e-01 -1.22861922e-01 -1.16682217e-01
6.57618046e-02 -8.57266247e-01 8.36310506e-01 7.37217784e-01
7.36144125e-01 -6.29884422e-01 2.91927814e-01 -1.03819501e-02
-1.04641962e+00 1.39951020e-01 -3.70817751e-01 -2.49750420e-01
5.39038062e-01 7.43601799e-01 -1.83689162e-01 5.80291927e-01
7.69299865e-01 1.06817329e+00 -3.64762008e-01 1.21279466e+00
-2.39135697e-01 4.32235956e-01 -2.16553241e-01 1.05348432e+00
2.23278344e-01 -3.68237972e-01 7.05967009e-01 1.12685227e+00
3.52481216e-01 3.03041071e-01 1.66285709e-01 9.84953821e-01
-2.46258482e-01 -3.40700209e-01 -6.20059490e-01 3.97943974e-01
1.77332647e-02 1.15177464e+00 -6.31615222e-01 -4.73185658e-01
-6.18432879e-01 1.56272125e+00 3.57486397e-01 7.57035077e-01
-9.80932176e-01 -1.45280302e-01 7.95170784e-01 9.13982764e-02
6.68795705e-01 -3.94288510e-01 -7.90764987e-02 -1.44720566e+00
2.02223346e-01 -6.64652228e-01 8.28653798e-02 -9.61856186e-01
-1.22765887e+00 4.23141479e-01 -2.14821801e-01 -1.31340396e+00
-7.56320059e-02 -4.13869083e-01 -5.75257540e-01 6.63065135e-01
-1.88486767e+00 -1.09667337e+00 -4.13768679e-01 7.34213829e-01
8.10280025e-01 1.29505828e-01 3.60350102e-01 5.77710509e-01
-1.94974616e-01 2.22347513e-01 1.44964337e-01 1.65352449e-01
1.07399237e+00 -8.51096451e-01 2.76367396e-01 1.15941250e+00
-3.28061618e-02 3.44357878e-01 7.48745441e-01 -6.22897565e-01
-1.30364060e+00 -1.28223622e+00 4.73733097e-01 -2.53627032e-01
3.43185991e-01 -1.97212264e-01 -1.12352383e+00 1.01525497e+00
5.22683859e-01 8.16643536e-01 8.93332660e-02 -4.53486323e-01
-4.29106086e-01 -7.05700517e-02 -1.16231287e+00 3.34346175e-01
1.11911154e+00 -4.72082824e-01 -4.13572878e-01 2.76889175e-01
8.84680331e-01 -6.38078928e-01 -6.64361119e-01 4.42288220e-01
5.04899800e-01 -1.07449365e+00 1.16727114e+00 -5.52150726e-01
7.82220781e-01 -5.45736730e-01 -7.26234615e-02 -1.14998531e+00
-3.42207938e-01 -1.12622702e+00 -4.85404968e-01 1.02554905e+00
-2.52983332e-01 -2.47107908e-01 8.42585444e-01 4.69492674e-01
-2.53707170e-01 -5.58241367e-01 -8.79880071e-01 -7.55744278e-01
1.14670210e-02 -3.84202063e-01 8.85175616e-02 9.92009223e-01
-6.46254599e-01 -7.64982179e-02 -1.01294470e+00 2.37588748e-01
9.09132421e-01 -6.66388050e-02 6.96128309e-01 -7.07435787e-01
-6.02067649e-01 -1.16311848e-01 -3.52455199e-01 -1.47756958e+00
5.02611220e-01 -2.96690792e-01 2.21838370e-01 -1.28820395e+00
2.11636752e-01 -1.54025197e-01 -2.59701788e-01 2.09259078e-01
-2.13338733e-01 6.50992692e-01 3.22176635e-01 1.01998702e-01
-7.10688293e-01 6.08315289e-01 1.61518312e+00 -2.72227198e-01
3.55932973e-02 -2.18047038e-01 -3.67060065e-01 7.59704709e-01
3.68988365e-01 -4.25578237e-01 -3.79550517e-01 -7.75310159e-01
-1.40636593e-01 5.83103716e-01 6.22165263e-01 -9.86212611e-01
6.88861758e-02 -3.32085043e-01 7.13038087e-01 -4.35052931e-01
5.74763477e-01 -9.57511067e-01 2.64442027e-01 2.64535159e-01
-3.01970363e-01 -1.98889356e-02 4.08688962e-01 8.33603442e-01
-4.73292172e-01 -1.15492105e-01 1.01132965e+00 -1.14770852e-01
-6.47977591e-01 4.61758584e-01 -1.95694908e-01 1.13711134e-01
9.46023285e-01 -7.09018335e-02 -1.63876861e-01 -5.63786209e-01
-1.00176966e+00 5.02489433e-02 6.84453785e-01 3.03617626e-01
8.80004883e-01 -1.34956992e+00 -7.74890542e-01 3.39592576e-01
-3.67492914e-01 8.85255858e-02 3.06324571e-01 7.62324929e-01
-8.85641038e-01 2.29608957e-02 -3.88171524e-01 -6.61907673e-01
-1.26466334e+00 6.49551213e-01 4.68383938e-01 -1.54949874e-01
-9.16642368e-01 9.28282976e-01 6.21946871e-01 5.58330715e-02
3.05937111e-01 -2.10870236e-01 3.30615908e-01 -3.88508439e-01
6.31157815e-01 1.33523390e-01 -3.67603861e-02 -7.75041401e-01
-1.83669463e-01 6.54548705e-01 -9.19581279e-02 -1.17896058e-01
1.32768929e+00 -3.55162770e-01 -1.20668337e-01 1.53033838e-01
1.38873434e+00 2.32824087e-02 -2.14634132e+00 -2.73830056e-01
-3.99752110e-01 -1.11576426e+00 7.34993145e-02 -4.92236912e-01
-1.53884661e+00 7.16751695e-01 4.46441799e-01 -3.03059161e-01
1.29092956e+00 -3.48883748e-01 9.84374404e-01 1.92384839e-01
7.41058215e-02 -8.98157179e-01 3.90577823e-01 4.52082008e-01
1.04867053e+00 -1.35675013e+00 2.31488809e-01 -5.85577130e-01
-2.21183226e-01 1.33105946e+00 5.09896576e-01 -5.22775054e-01
6.77189171e-01 3.59987110e-01 1.16735674e-01 2.22514153e-01
-8.74410868e-01 8.20903778e-02 3.18265378e-01 5.37599504e-01
3.54625225e-01 -3.50890398e-01 -2.02069536e-01 -2.69144066e-02
3.69408160e-01 3.18307459e-01 8.06565344e-01 1.06885290e+00
-1.80300385e-01 -9.95331407e-01 -3.96109551e-01 6.48591891e-02
-7.25254893e-01 -8.04766193e-02 -3.74490023e-02 5.87866902e-01
7.39474893e-02 7.55342066e-01 -3.55917565e-03 2.79744179e-03
2.17783190e-02 -1.38823688e-01 9.48133469e-01 -2.42337868e-01
-4.24809486e-01 5.06687164e-01 -4.05837856e-02 -9.91567671e-01
-8.31294000e-01 -5.46676338e-01 -9.79766667e-01 -1.59579590e-01
-5.79508170e-02 -2.40260646e-01 2.93368161e-01 8.60900760e-01
3.76132429e-01 6.79653525e-01 5.11394262e-01 -1.32618022e+00
-5.27473651e-02 -5.91059029e-01 -4.55483139e-01 7.24222481e-01
8.35216582e-01 -6.13523185e-01 -5.26925147e-01 5.42401671e-01] | [10.753185272216797, -1.4257851839065552] |
0a8a0bf4-2e9a-4998-8fdf-951763477ccd | identifying-causal-structure-in-dynamical | 2006.03906 | null | https://arxiv.org/abs/2006.03906v2 | https://arxiv.org/pdf/2006.03906v2.pdf | Identifying Causal Structure in Dynamical Systems | Mathematical models are fundamental building blocks in the design of dynamical control systems. As control systems are becoming increasingly complex and networked, approaches for obtaining such models based on first principles reach their limits. Data-driven methods provide an alternative. However, without structural knowledge, these methods are prone to finding spurious correlations in the training data, which can hamper generalization capabilities of the obtained models. This can significantly lower control and prediction performance when the system is exposed to unknown situations. A preceding causal identification can prevent this pitfall. In this paper, we propose a method that identifies the causal structure of control systems. We design experiments based on the concept of controllability, which provides a systematic way to compute input trajectories that steer the system to specific regions in its state space. We then analyze the resulting data leveraging powerful techniques from causal inference and extend them to control systems. Further, we derive conditions that guarantee the discovery of the true causal structure of the system. Experiments on a robot arm demonstrate reliable causal identification from real-world data and enhanced generalization capabilities. | ['Dominik Baumann', 'Karl H. Johansson', 'Sebastian Trimpe', 'Friedrich Solowjow'] | 2020-06-06 | null | null | null | null | ['causal-identification'] | ['reasoning'] | [ 3.23651254e-01 6.41176999e-02 -5.33531308e-01 5.42145148e-02
-4.16124575e-02 -5.37632942e-01 7.15261459e-01 1.20415032e-01
2.33934388e-01 9.77771342e-01 -1.64990261e-01 -6.61704123e-01
-1.03781545e+00 -6.81550384e-01 -7.88503289e-01 -9.12519276e-01
-4.61233228e-01 1.84158936e-01 1.11599892e-01 -1.77099794e-01
3.22811425e-01 5.97029030e-01 -1.47857857e+00 -2.66479552e-01
9.59346116e-01 6.72005415e-01 2.45306551e-01 5.97536266e-01
3.83799702e-01 8.62316430e-01 -3.93365502e-01 5.95901966e-01
1.78681657e-01 -5.61404407e-01 -4.52428132e-01 6.54529482e-02
-6.62104636e-02 -1.19373083e-01 -3.74884337e-01 1.13332057e+00
9.46552679e-03 1.96925044e-01 6.41786039e-01 -1.50484574e+00
-2.96001464e-01 6.00769818e-01 -1.09661214e-01 -7.59747252e-02
2.30822489e-01 9.34208632e-02 8.09212029e-01 -4.37102616e-01
3.20139736e-01 1.29506648e+00 5.34210384e-01 4.98548836e-01
-1.61600637e+00 -5.97445965e-01 3.73299181e-01 2.91236956e-02
-1.23615003e+00 -4.14669007e-01 8.43193114e-01 -7.69213617e-01
3.74806494e-01 2.64588892e-01 4.81413066e-01 1.02753222e+00
4.58319396e-01 3.83599877e-01 1.21255279e+00 -3.47881675e-01
3.62388223e-01 2.03680340e-02 1.82732463e-01 6.82159543e-01
5.40355563e-01 7.27539480e-01 -3.61028016e-01 -3.22622716e-01
1.02565742e+00 2.08951579e-03 -3.95427257e-01 -6.93307817e-01
-9.53823864e-01 8.21148396e-01 4.91947055e-01 2.79050171e-01
-4.81578529e-01 2.61751413e-01 1.75202582e-02 5.78621745e-01
8.71142372e-03 7.97262013e-01 -3.34853172e-01 2.05586150e-01
-3.62060487e-01 3.65012437e-01 8.22573841e-01 9.27763641e-01
2.79428422e-01 8.84455368e-02 3.35587293e-01 1.89418018e-01
3.54624212e-01 4.03891861e-01 -1.56979654e-02 -7.71459758e-01
2.02896908e-01 7.57726431e-01 3.70753437e-01 -1.29122508e+00
-3.66463959e-01 -2.63483375e-01 -8.80705774e-01 2.64774442e-01
4.75470990e-01 -2.39161327e-01 -7.28742003e-01 1.77705443e+00
4.06835765e-01 3.29128616e-02 5.87344766e-02 8.80616605e-01
-3.55848759e-01 5.99501550e-01 -1.55761868e-01 -6.74598634e-01
6.28616631e-01 -1.68432862e-01 -7.39397645e-01 1.10305354e-01
4.10130620e-01 -2.66946137e-01 8.53183389e-01 4.97915834e-01
-6.31487072e-01 -4.15940285e-01 -1.17795455e+00 7.25129485e-01
4.12786938e-02 -1.47757471e-01 3.77793521e-01 3.64509910e-01
-6.18399680e-01 1.05786896e+00 -1.13264167e+00 -3.69312704e-01
-6.55799285e-02 5.06463230e-01 -4.36623022e-02 2.17486307e-01
-1.14195263e+00 9.67748702e-01 5.72733760e-01 3.72291446e-01
-1.08430588e+00 -7.58343041e-01 -3.36727560e-01 -1.48034528e-01
8.62863123e-01 -4.48944598e-01 1.05377686e+00 -7.48346508e-01
-1.36284733e+00 -2.77652085e-01 4.17673811e-02 -4.49102461e-01
4.28882569e-01 -1.63422465e-01 -3.40599388e-01 -6.39426410e-02
-1.30903795e-01 -1.63698241e-01 1.19636762e+00 -1.37378883e+00
-5.94749928e-01 -2.54014999e-01 6.25604987e-02 -2.83905864e-01
-3.09935004e-01 -3.76273632e-01 3.33947212e-01 -2.73987472e-01
1.33169532e-01 -1.11077702e+00 -7.55424976e-01 -1.38587594e-01
-7.39025593e-01 -1.96167782e-01 8.60114872e-01 -8.78860503e-02
1.48396921e+00 -2.11400509e+00 4.66253191e-01 6.63152397e-01
3.30411643e-01 1.38122991e-01 2.16292471e-01 8.24905634e-01
-3.19604874e-01 3.01992595e-01 -3.87781076e-02 3.62552881e-01
-2.52689332e-01 4.48466718e-01 -6.01346850e-01 5.76504171e-01
4.50454861e-01 2.59854734e-01 -9.83458579e-01 -1.78363875e-01
4.14366007e-01 -2.92132818e-03 -5.55950284e-01 3.61740291e-01
-3.13303083e-01 7.87054420e-01 -1.00212324e+00 -1.08893332e-03
-5.29532619e-02 -1.96261734e-01 5.82595944e-01 4.34091315e-02
-3.29711616e-01 2.42485911e-01 -1.25075638e+00 7.92173564e-01
-6.36195898e-01 3.48617166e-01 -6.53149635e-02 -1.36171722e+00
9.01535988e-01 3.42713118e-01 6.12232327e-01 -2.09619641e-01
3.11752319e-01 9.71412659e-02 3.80674601e-01 -5.73231876e-01
6.98630512e-02 -1.86265275e-01 -1.94189891e-01 1.10200822e-01
-1.78813487e-01 1.01648010e-02 1.66063681e-01 5.41322082e-02
1.22636175e+00 -2.13062853e-01 5.68745673e-01 -7.11126506e-01
4.65555429e-01 3.21718186e-01 6.45690262e-01 7.61310160e-01
4.57794443e-02 -1.52141303e-01 8.13715518e-01 -2.30247930e-01
-1.09459436e+00 -9.93988931e-01 -1.24511719e-01 3.28822106e-01
1.80484399e-01 -3.16192001e-01 -4.03830290e-01 -4.38900739e-01
2.77256966e-01 8.59885991e-01 -8.32781434e-01 -6.12067342e-01
-4.00221407e-01 -5.98370075e-01 4.96852212e-02 4.03710216e-01
-1.14110284e-01 -4.14895505e-01 -6.85640156e-01 4.08225268e-01
1.37911275e-01 -7.52579451e-01 7.36102536e-02 3.02756518e-01
-1.07952785e+00 -1.45923066e+00 -4.82594110e-02 -2.79930174e-01
8.89953911e-01 -5.66714667e-02 5.39720953e-01 4.30911742e-02
-2.85478175e-01 2.24069968e-01 2.83588842e-02 -2.90972084e-01
-9.56763029e-01 -1.59040555e-01 7.41678596e-01 -5.79387210e-02
-2.82498538e-01 -5.89792490e-01 -2.31972158e-01 6.02259696e-01
-5.07143795e-01 -6.85530677e-02 5.08994818e-01 1.03626025e+00
3.19146276e-01 7.05555975e-01 9.18946087e-01 -7.38657713e-01
7.61477530e-01 -6.45917833e-01 -1.27600706e+00 1.58459634e-01
-1.16719842e+00 3.58725816e-01 9.96334970e-01 -7.30512023e-01
-8.53951097e-01 3.75846177e-01 6.74473941e-01 -7.25336432e-01
-5.59747331e-02 8.53599072e-01 -6.42671287e-02 2.88987428e-01
7.92719185e-01 -5.31994775e-02 3.51939917e-01 -3.68976444e-01
2.90817291e-01 2.91019738e-01 3.64405930e-01 -6.60951793e-01
9.38354194e-01 1.27329588e-01 5.80397308e-01 -9.49952066e-01
-6.38332844e-01 -3.38526428e-01 -6.91444516e-01 -4.88367707e-01
3.73471081e-01 -3.87809306e-01 -1.06298339e+00 -7.02880621e-02
-8.68828237e-01 -1.85327649e-01 -5.33048809e-02 4.54259306e-01
-6.39231682e-01 -2.58728802e-01 -2.12970242e-01 -1.36595094e+00
4.05989081e-01 -9.05221939e-01 4.73873198e-01 8.68440121e-02
-4.82495010e-01 -9.93780077e-01 1.69733480e-01 -2.86332428e-01
6.00718893e-02 4.06990230e-01 1.15035546e+00 -5.02449214e-01
-5.06794870e-01 -3.12349945e-01 1.73387900e-01 1.97792172e-01
4.18030798e-01 3.28601539e-01 -4.51439291e-01 -3.20315897e-01
1.55362800e-01 1.11199245e-01 3.00009131e-01 4.62902427e-01
6.64784789e-01 -7.00299025e-01 -7.86942959e-01 -2.14838013e-01
1.46947920e+00 5.29232204e-01 2.00155109e-01 1.43387644e-02
8.06471884e-01 9.62831259e-01 7.31934726e-01 2.86166042e-01
-1.50432855e-01 5.98124862e-01 4.55557585e-01 2.60362715e-01
4.19360131e-01 -5.22999406e-01 3.26498538e-01 7.09464490e-01
-1.61645755e-01 -6.56185746e-02 -9.66458857e-01 5.88060975e-01
-2.05493069e+00 -9.17584002e-01 -5.27512968e-01 2.44445658e+00
8.12696874e-01 1.64617315e-01 2.01611385e-01 3.57430458e-01
8.14553380e-01 -5.78451633e-01 -6.78252280e-01 -1.33743003e-01
3.32970858e-01 -2.23716155e-01 6.53077602e-01 4.76396650e-01
-9.48654115e-01 4.48453128e-01 6.39178228e+00 2.51618207e-01
-1.01963794e+00 -4.51136738e-01 2.70343482e-01 2.13086560e-01
1.43296421e-01 1.37413472e-01 -6.83000803e-01 3.07334512e-01
1.24326146e+00 -6.35907948e-01 3.41250658e-01 8.64957571e-01
7.35045433e-01 -1.65779650e-01 -1.41283751e+00 4.61908609e-01
-6.97416306e-01 -1.02531004e+00 -1.23216130e-01 3.55309427e-01
8.55913639e-01 -5.02141654e-01 8.12616479e-03 -1.49661481e-01
6.21464372e-01 -9.53938484e-01 4.88523930e-01 6.47980690e-01
3.31126004e-01 -8.53219390e-01 4.26865429e-01 6.12634897e-01
-7.92187214e-01 -5.64414561e-01 -1.73978657e-01 -3.68992239e-01
1.51637450e-01 8.02493691e-01 -1.08512783e+00 6.62669897e-01
2.24703237e-01 7.24456191e-01 -1.73663646e-01 8.80163491e-01
-3.42305988e-01 7.83952832e-01 -3.94268245e-01 -4.02279466e-01
-1.15767885e-02 -1.57737717e-01 8.42962205e-01 4.60361123e-01
2.13904962e-01 8.42734948e-02 3.91622186e-01 9.70523715e-01
5.65523267e-01 -3.24038476e-01 -1.14572978e+00 -3.52550209e-01
5.51008523e-01 6.76631272e-01 -5.21702349e-01 -1.56372413e-01
-4.00939025e-02 3.23250800e-01 1.31190658e-01 3.01126719e-01
-7.66320884e-01 -1.61578372e-01 7.40430236e-01 2.71176361e-02
-6.72628954e-02 -5.78430295e-01 -3.02170783e-01 -8.17647636e-01
-1.26809165e-01 -8.71200800e-01 1.18108399e-01 -6.82770088e-02
-1.31861699e+00 1.76110551e-01 3.54076296e-01 -1.41160893e+00
-8.37729931e-01 -5.37864983e-01 -2.67058611e-01 6.74231768e-01
-7.99960732e-01 -3.83196563e-01 1.15681753e-01 5.35612583e-01
2.13748500e-01 2.24526182e-01 6.84388101e-01 1.28547803e-01
-8.06824028e-01 -9.18553993e-02 1.65324897e-01 -1.45056024e-01
4.58737582e-01 -1.17594612e+00 -1.38893157e-01 1.01974416e+00
-1.78263485e-01 9.50304151e-01 1.21938622e+00 -8.37587893e-01
-1.77317834e+00 -1.12776458e+00 4.61527467e-01 -4.31999892e-01
1.41081834e+00 -2.53190577e-01 -9.20262277e-01 5.09178400e-01
-3.91209185e-01 -3.22441220e-01 4.59295884e-02 3.96026909e-01
-8.55801255e-03 -1.07586540e-01 -7.59765923e-01 9.22473371e-01
9.01462674e-01 -2.39666939e-01 -5.92093706e-01 1.25694081e-01
6.43106043e-01 6.38651699e-02 -9.84263062e-01 4.41526651e-01
5.82716823e-01 -4.85021144e-01 7.40387678e-01 -8.32968056e-01
4.51773018e-01 -5.78022242e-01 9.96548831e-02 -1.62424529e+00
-2.03078240e-01 -8.95099103e-01 -1.16471894e-01 1.07402432e+00
5.28700054e-01 -7.62945473e-01 4.70925242e-01 6.31559134e-01
5.47970422e-02 -4.07319784e-01 -7.41852403e-01 -1.32280540e+00
7.12509528e-02 -3.90112162e-01 1.46154463e-01 1.05506837e+00
4.72262114e-01 5.19883275e-01 -3.85851324e-01 6.37975872e-01
8.57462466e-01 1.86757848e-01 6.73658907e-01 -1.37257433e+00
-1.96880594e-01 -4.22027141e-01 -4.63287920e-01 -6.15238428e-01
1.37738436e-01 -3.22468758e-01 5.12509525e-01 -1.02097225e+00
-2.10923791e-01 -7.96583831e-01 -1.45570219e-01 3.18870366e-01
-1.12348929e-01 -4.68061984e-01 -1.37187792e-02 2.16506884e-01
8.48003253e-02 4.96153116e-01 1.07340276e+00 8.05143118e-02
-4.77974504e-01 3.84629011e-01 -2.70373821e-01 6.80447340e-01
1.06620383e+00 -5.07812381e-01 -9.12805796e-01 1.14280432e-01
1.55970231e-01 4.85395104e-01 6.55274332e-01 -9.37037885e-01
2.72316843e-01 -6.72264397e-01 -2.22853664e-02 -2.33064264e-01
-1.14213467e-01 -1.26008046e+00 4.95074332e-01 9.29822445e-01
-5.47927141e-01 -1.04230121e-01 1.29527882e-01 1.00418460e+00
-1.29426330e-01 -9.63204205e-02 6.81440890e-01 3.33342195e-01
-4.74051774e-01 2.00896673e-02 -6.67554975e-01 -4.43448514e-01
1.09842241e+00 2.45546162e-01 -3.70022506e-02 -3.15680563e-01
-7.83126473e-01 4.88824159e-01 8.25129896e-02 5.19468844e-01
3.59756261e-01 -1.05773294e+00 -2.46412948e-01 -5.61910048e-02
-3.50508764e-02 -3.57107043e-01 -3.43004689e-02 9.80771482e-01
1.74371660e-01 7.52652705e-01 -1.75702386e-02 -7.18783200e-01
-1.03416693e+00 9.49683011e-01 3.44121069e-01 1.19133532e-01
-5.81194639e-01 1.33683890e-01 1.50932863e-01 -1.46060169e-01
-6.82271719e-02 -5.65784574e-01 -1.09518282e-02 -1.15286790e-01
3.80896389e-01 2.72466481e-01 -1.59369767e-01 -1.05539910e-01
-2.63809919e-01 9.94906053e-02 1.66503504e-01 -1.07437134e-01
1.10341895e+00 -2.37238228e-01 3.83688174e-02 8.58001709e-01
8.62559795e-01 -3.52025747e-01 -1.39901733e+00 3.62292049e-03
4.57801938e-01 -4.40246224e-01 2.30880558e-01 -4.96003717e-01
-7.28652537e-01 5.31587958e-01 4.52857852e-01 8.44423413e-01
1.02250183e+00 -1.81395248e-01 -8.24580193e-02 5.78088343e-01
5.42312324e-01 -7.95719981e-01 -1.47329941e-01 2.98066676e-01
8.47112298e-01 -9.26662624e-01 -1.09522760e-01 -6.30900741e-01
-1.14846811e-01 1.26240313e+00 3.64097565e-01 -5.83942533e-01
7.65037119e-01 4.17547733e-01 -4.20925587e-01 -1.56021595e-01
-1.17437088e+00 -7.74480104e-02 2.59959966e-01 4.85333443e-01
8.01365897e-02 2.38191724e-01 -3.19300711e-01 2.90214121e-01
1.04001379e-02 -3.64831393e-03 7.14768887e-01 9.78519261e-01
-3.39248598e-01 -1.04240799e+00 -4.86739129e-01 3.64123553e-01
-7.02995732e-02 2.21522540e-01 -4.02786702e-01 1.18321466e+00
-4.99682814e-01 1.25318015e+00 -8.08421075e-02 -5.46980202e-01
6.83739960e-01 -9.33174267e-02 4.90875274e-01 -4.72461283e-01
8.00321922e-02 1.16166228e-03 2.59335637e-01 -6.43780053e-01
-3.67471427e-01 -9.64065075e-01 -1.10576403e+00 -3.91754210e-01
-5.96551538e-01 2.12108329e-01 5.02874970e-01 1.02519035e+00
2.93324947e-01 8.39580238e-01 1.00696695e+00 -6.00733876e-01
-9.35331941e-01 -7.49040782e-01 -5.10771573e-01 -1.84816837e-01
6.43293321e-01 -1.18002880e+00 -4.50421631e-01 3.06803614e-01] | [4.902138710021973, 2.3609161376953125] |
9b2440e8-0422-4bb9-bad4-6fd79226fc1e | an-application-of-pseudo-log-likelihoods-to | 2201.09377 | null | https://arxiv.org/abs/2201.09377v1 | https://arxiv.org/pdf/2201.09377v1.pdf | An Application of Pseudo-Log-Likelihoods to Natural Language Scoring | Language models built using semi-supervised machine learning on large corpora of natural language have very quickly enveloped the fields of natural language generation and understanding. In this paper we apply a zero-shot approach independently developed by a number of researchers now gaining recognition as a significant alternative to fine-tuning for evaluation on common sense tasks. A language model with relatively few parameters and training steps compared to a more recent language model (T5) can outperform it on a recent large data set (TimeDial), while displaying robustness in its performance across a similar class of language tasks. Surprisingly, this result is achieved by using a hyperparameter-free zero-shot method with the smaller model, compared to fine-tuning to the larger model. We argue that robustness of the smaller model ought to be understood in terms of compositionality, in a sense that we draw from recent literature on a class of similar models. We identify a practical cost for our method and model: high GPU-time for natural language evaluation. The zero-shot measurement technique that produces remarkable stability, both for ALBERT and other BERT variants, is an application of pseudo-log-likelihoods to masked language models for the relative measurement of probability for substitution alternatives in forced choice language tasks such as the Winograd Schema Challenge, Winogrande, and others. One contribution of this paper is to bring together a number of similar, but independent strands of research. We produce some absolute state-of-the-art results for common sense reasoning in binary choice tasks, performing better than any published result in the literature, including fine-tuned efforts. We show a remarkable consistency of the model's performance under adversarial settings, which we argue is best explained by the model's compositionality of representations. | ['Ali Emami', 'Darren Abramson'] | 2022-01-23 | null | null | null | null | ['timedial'] | ['natural-language-processing'] | [ 5.57699144e-01 4.23408777e-01 -3.54845315e-01 -3.09872836e-01
-1.11721361e+00 -6.42488122e-01 1.20683408e+00 1.39289916e-01
-7.94262588e-01 9.60299313e-01 4.68318462e-01 -7.94707835e-01
-2.83066630e-01 -8.61676693e-01 -4.85341698e-01 -4.28739607e-01
2.88634777e-01 8.69978368e-01 2.79963762e-01 -9.43197906e-01
2.96692014e-01 1.65860415e-01 -1.35170901e+00 3.39473724e-01
4.96572226e-01 4.80965853e-01 -1.09750181e-01 6.54297411e-01
-1.55388981e-01 1.06854427e+00 -3.80840778e-01 -8.47825170e-01
2.43036583e-01 -2.68972784e-01 -1.10685289e+00 -3.69227767e-01
2.72290140e-01 -7.67034069e-02 -2.01604784e-01 8.73554289e-01
8.34201396e-01 2.31418863e-01 6.50744438e-01 -1.10160100e+00
-5.72573721e-01 9.61938679e-01 -1.78026617e-01 3.45117450e-01
5.86936116e-01 4.06884998e-01 1.34434795e+00 -4.84721541e-01
8.95933867e-01 1.63641691e+00 7.40526915e-01 8.69681120e-01
-1.65262580e+00 -4.53667760e-01 -1.36285260e-01 1.66050643e-01
-1.29992580e+00 -5.03399909e-01 3.28481734e-01 -4.98702973e-01
1.52829456e+00 1.88320950e-01 2.25806698e-01 1.57275748e+00
2.46473491e-01 5.86573899e-01 1.54457021e+00 -8.21602702e-01
3.81861776e-01 2.27327645e-01 1.60984054e-01 7.59919405e-01
1.56962037e-01 4.51374263e-01 -4.63079572e-01 -6.96665287e-01
2.95770109e-01 -3.45350653e-01 5.34831583e-02 -4.12811011e-01
-1.19351578e+00 1.19491243e+00 5.35516776e-02 4.77395296e-01
-8.00439063e-03 1.83460012e-01 5.19645870e-01 4.88077909e-01
4.40270275e-01 6.17141187e-01 -4.86801445e-01 -2.28407040e-01
-1.14929771e+00 5.66229820e-01 1.00746965e+00 6.52657628e-01
4.73180830e-01 2.34877691e-03 -5.07824183e-01 7.91795671e-01
3.12202483e-01 2.76006281e-01 8.60793412e-01 -9.13534284e-01
3.31559956e-01 3.10907483e-01 8.76603648e-02 -5.01393080e-01
-4.17731166e-01 -1.64514109e-01 -5.20388961e-01 4.83145148e-01
6.64991140e-01 -3.23128887e-02 -7.50150442e-01 2.16731930e+00
1.84230506e-02 -1.48945495e-01 1.90859705e-01 4.83915150e-01
3.43496859e-01 2.40373060e-01 1.88528955e-01 -6.43959567e-02
1.64194083e+00 -6.95008159e-01 -3.51953596e-01 -3.90140146e-01
8.96233499e-01 -8.13207865e-01 1.36377501e+00 3.34459394e-01
-1.10333931e+00 -2.47195587e-01 -1.31850743e+00 -1.23563424e-01
-5.90236604e-01 -6.55200541e-01 9.20171380e-01 1.12413478e+00
-1.13715434e+00 6.94773793e-01 -5.39496064e-01 -5.86113751e-01
2.78341621e-01 2.14496642e-01 -2.41645098e-01 -8.32907185e-02
-1.78360569e+00 1.36772978e+00 4.87217456e-01 -6.59063280e-01
-6.98528171e-01 -5.71728051e-01 -9.95766461e-01 -2.51566675e-02
7.11923242e-01 -8.36138606e-01 1.55222464e+00 -6.82083607e-01
-1.51808107e+00 1.33884180e+00 -7.29886889e-02 -9.24723923e-01
6.70757055e-01 1.16992116e-01 -3.52664709e-01 -6.97649419e-02
2.60111749e-01 6.15954161e-01 7.39183664e-01 -7.86139488e-01
-2.24756688e-01 -3.35230827e-02 2.75329471e-01 -7.69743845e-02
2.84866989e-01 4.20368999e-01 1.92060262e-01 -7.97855377e-01
-3.30134720e-01 -8.68260741e-01 -3.80464375e-01 -8.52618739e-02
-1.23779774e-01 -4.44689959e-01 -3.00040115e-02 -9.18750688e-02
1.24102485e+00 -1.91686130e+00 -8.45910087e-02 8.65157787e-03
2.10677519e-01 3.19156826e-01 -1.19818196e-01 6.13685846e-01
-3.90242100e-01 4.43749398e-01 -7.13677943e-01 -3.19966435e-01
5.40269911e-01 2.92784572e-01 -7.36764193e-01 3.56948972e-01
3.46768886e-01 1.14863527e+00 -1.03923595e+00 -4.58754748e-01
1.45754263e-01 3.61870937e-02 -6.49147868e-01 -1.17203362e-01
-4.31629896e-01 -2.28030846e-01 -9.32138041e-02 2.11020544e-01
2.91365981e-01 -2.67030120e-01 2.50816762e-01 3.56885135e-01
9.67896655e-02 6.13496482e-01 -1.13547456e+00 1.82975304e+00
-2.87977993e-01 2.74961472e-01 -3.28559428e-01 -8.30546021e-01
6.83297276e-01 4.52130377e-01 -6.38345852e-02 -7.42486298e-01
9.69429687e-02 3.68454576e-01 3.08395267e-01 -2.81993657e-01
4.58155721e-01 -7.30475843e-01 -5.03204763e-01 9.48138535e-01
3.24899048e-01 -5.92593431e-01 3.24042469e-01 4.84319985e-01
1.12708449e+00 2.05389202e-01 8.16534460e-01 -4.31371748e-01
4.50902611e-01 1.11562066e-01 1.98939040e-01 1.38938558e+00
-3.18965226e-01 4.26169902e-01 5.27739942e-01 -5.04823446e-01
-1.25505030e+00 -1.17667377e+00 -2.93561429e-01 1.29160786e+00
-3.61552536e-01 -6.18198514e-01 -6.74885571e-01 -3.11602831e-01
-2.07975417e-01 1.27707899e+00 -6.05125964e-01 -3.32902223e-01
-4.41649258e-01 -1.05356824e+00 1.26235259e+00 3.27808648e-01
1.43281743e-01 -1.09490263e+00 -6.97198033e-01 3.94149214e-01
9.05547291e-02 -9.49352205e-01 -1.91591121e-02 4.06339139e-01
-4.09343034e-01 -9.25221562e-01 -5.64757764e-01 -4.61311787e-01
-1.53793143e-02 -1.27981424e-01 1.34055126e+00 -1.64536256e-02
-3.17024112e-01 3.67484689e-01 -1.40614003e-01 -4.02416766e-01
-8.97815108e-01 -3.31495479e-02 2.36895710e-01 -4.67715591e-01
7.03615069e-01 -4.75488514e-01 -9.35521573e-02 1.01939842e-01
-1.17639554e+00 -1.00452274e-01 4.01875794e-01 1.21250916e+00
2.69072428e-02 -3.79995555e-01 4.18098330e-01 -1.10710371e+00
1.01055515e+00 -5.55694401e-01 -3.81095469e-01 3.62346888e-01
-8.62272799e-01 5.77900171e-01 4.82599884e-01 -2.68443346e-01
-1.12922800e+00 -4.26491201e-01 -2.10296497e-01 -2.92977821e-02
-1.60482049e-01 2.65272379e-01 1.75835177e-01 1.34462401e-01
1.16446674e+00 1.59834370e-01 6.05586655e-02 -2.00917408e-01
8.27865660e-01 4.82965082e-01 4.26576316e-01 -9.96404290e-01
7.51920521e-01 4.98194396e-01 -5.00524715e-02 -5.00654697e-01
-7.85331547e-01 -2.72974074e-01 -4.07179922e-01 5.50744712e-01
7.49429822e-01 -8.09215188e-01 -3.79185230e-01 2.85841107e-01
-1.17529941e+00 -5.20748675e-01 -5.12362838e-01 3.40497792e-01
-8.62201035e-01 5.14895022e-01 -5.83603323e-01 -9.35773909e-01
-3.17899197e-01 -1.05741382e+00 9.92513537e-01 -2.19268024e-01
-7.31703997e-01 -1.29858804e+00 4.80528474e-01 3.33905131e-01
5.81953168e-01 -1.54898148e-02 1.28184509e+00 -1.12773848e+00
-1.57220185e-01 -1.21886410e-01 2.34731771e-02 1.01825804e-01
-3.34809870e-01 -2.95270145e-01 -1.07210970e+00 -2.54533857e-01
2.66332209e-01 -8.27490270e-01 1.06316304e+00 -2.06363797e-02
5.57411492e-01 -1.80909932e-01 -1.21199079e-01 3.98970127e-01
1.35033619e+00 -5.00817336e-02 7.48973131e-01 4.43133295e-01
1.30012542e-01 5.00921488e-01 4.22779977e-01 3.27956885e-01
2.13721067e-01 7.57666051e-01 -1.52923539e-01 1.51763290e-01
-1.85976438e-02 -3.49273711e-01 3.56260836e-01 3.44854265e-01
1.46993369e-01 -3.01927507e-01 -1.14023972e+00 3.86935651e-01
-1.72674251e+00 -1.33023989e+00 4.09972638e-01 2.23037434e+00
1.04721296e+00 6.76431179e-01 9.50948671e-02 1.42868489e-01
5.64450383e-01 5.40257275e-01 -3.26252699e-01 -8.49210978e-01
-4.62015957e-01 8.28863502e-01 3.35862666e-01 8.66894364e-01
-7.87134469e-01 1.15822101e+00 7.18939781e+00 1.31416595e+00
-8.73918593e-01 3.94532025e-01 6.28799200e-01 -1.43717676e-01
-5.60786545e-01 3.96922946e-01 -6.62053108e-01 1.88093841e-01
1.18411052e+00 -5.12639880e-01 6.52997136e-01 7.40892708e-01
-2.06446201e-01 -1.44850025e-02 -1.23564410e+00 8.58922362e-01
3.43200594e-01 -1.39319921e+00 1.35781458e-02 -1.93608373e-01
5.79726636e-01 3.06783259e-01 -3.93088348e-02 6.73147380e-01
8.65325212e-01 -1.36137974e+00 8.51237118e-01 2.81775773e-01
8.68894041e-01 -2.75339723e-01 4.63795811e-01 5.62105894e-01
-6.76628053e-01 -1.06592208e-01 -4.12532717e-01 -5.26817143e-01
1.31393537e-01 1.69068858e-01 -9.75811779e-01 2.43170381e-01
1.35019794e-01 -1.23388311e-02 -5.60661793e-01 4.85059887e-01
-2.79926807e-01 6.41555786e-01 -2.85695314e-01 -2.30666921e-01
4.34514850e-01 2.17168376e-01 6.64791703e-01 1.35704517e+00
-1.17754884e-01 6.38408735e-02 2.55457629e-02 1.20592797e+00
5.20732179e-02 1.36328582e-02 -8.31084430e-01 -1.33735552e-01
5.11984229e-01 9.70860958e-01 -4.16189313e-01 -6.17607594e-01
-4.86501426e-01 7.14169443e-01 3.57916921e-01 -2.28740834e-02
-6.73883438e-01 -2.39677176e-01 4.14749831e-01 1.81027763e-02
1.13283336e-01 -1.67332426e-01 -2.93080151e-01 -1.25926161e+00
-1.53330013e-01 -1.16617405e+00 6.34735048e-01 -6.79351985e-01
-1.59155166e+00 5.69104791e-01 3.04960310e-01 -8.75574589e-01
-9.15413916e-01 -8.33708942e-01 -6.50748968e-01 1.03458023e+00
-1.27781856e+00 -8.39021444e-01 4.11334515e-01 5.43604910e-01
6.09665632e-01 -4.59706038e-01 1.31435478e+00 -1.17575064e-01
-2.06011936e-01 7.98313558e-01 -1.04982868e-01 -2.28950232e-02
8.30710888e-01 -1.32427096e+00 9.45279539e-01 7.22553909e-01
3.46062809e-01 9.90707636e-01 7.03772366e-01 -4.33893085e-01
-1.11356616e+00 -5.34340203e-01 1.17556965e+00 -8.14094067e-01
1.06503248e+00 -4.59969401e-01 -6.80874646e-01 6.21294320e-01
2.94073999e-01 -2.27487862e-01 7.93401718e-01 7.52536133e-02
-7.98517227e-01 4.42623407e-01 -1.21026254e+00 7.99409330e-01
9.71286833e-01 -8.93617809e-01 -1.42262077e+00 3.98390591e-01
8.82298112e-01 -1.59702718e-01 -4.72934932e-01 2.00493783e-01
6.42715514e-01 -9.78458226e-01 9.68919456e-01 -1.13413870e+00
4.38045025e-01 4.77532335e-02 -4.50558811e-01 -1.13746202e+00
-4.09411371e-01 -1.18004251e+00 2.57716030e-01 9.62745845e-01
5.15672386e-01 -8.74525726e-01 3.52909297e-01 7.61167765e-01
2.87389427e-01 -7.36239076e-01 -1.13025129e+00 -6.64876521e-01
6.04990542e-01 -7.71519303e-01 4.53780055e-01 9.25219059e-01
2.29002386e-01 7.41718471e-01 -3.56924266e-01 -5.21458328e-01
5.93120813e-01 -1.43514678e-01 7.10093558e-01 -1.04822898e+00
-7.14630663e-01 -5.47755957e-01 -5.27949870e-01 -6.76611900e-01
3.72595072e-01 -1.30157542e+00 -2.99179852e-01 -1.07540655e+00
2.57777452e-01 -1.86462060e-01 -2.47641891e-01 4.97783989e-01
-2.43842542e-01 1.69146642e-01 2.35268220e-01 9.21840295e-02
-3.75769049e-01 2.18603924e-01 7.20108747e-01 -2.90563434e-01
1.32755488e-01 -2.33347088e-01 -1.02343762e+00 8.77656698e-01
6.72992289e-01 -6.10162258e-01 -3.97497863e-01 -2.58180767e-01
5.75116873e-01 -1.41456351e-02 4.33870047e-01 -8.41798425e-01
2.02106550e-01 -1.64470255e-01 -4.96349344e-03 9.48781446e-02
4.13648039e-01 -1.43103510e-01 -2.20176652e-01 7.47029662e-01
-7.15884268e-01 1.21045619e-01 1.66812047e-01 4.59109634e-01
2.64091283e-01 -4.63522553e-01 8.30312908e-01 -6.90917969e-01
-9.29037988e-01 -1.25186905e-01 -5.47325432e-01 6.10318780e-01
7.19052374e-01 7.32786581e-03 -6.23192608e-01 -3.43391299e-01
-6.04604900e-01 -1.03049055e-01 4.91124153e-01 4.44242477e-01
3.06450725e-01 -1.10273206e+00 -1.04727924e+00 1.17956705e-01
2.70693034e-01 -5.61311603e-01 -4.23526242e-02 7.61895299e-01
-3.50886196e-01 5.62785387e-01 -1.44852772e-01 -1.86075091e-01
-7.69759476e-01 8.80886018e-01 2.84168184e-01 -7.46261179e-01
-3.60486269e-01 8.60504210e-01 5.16688228e-02 -6.40013099e-01
-8.11360851e-02 -1.59317881e-01 1.72106773e-01 -1.79147944e-02
6.64269924e-01 1.96901277e-01 -1.41007006e-01 -5.01319230e-01
-4.33324873e-01 2.45399892e-01 -1.64662644e-01 -5.55311918e-01
1.10840535e+00 1.18651055e-01 -1.22355171e-01 6.83323741e-01
8.84218574e-01 1.51256099e-01 -6.47884369e-01 -5.66175878e-01
2.47332454e-01 -1.86236531e-01 -2.04795748e-01 -1.05539501e+00
-6.40782937e-02 9.25038278e-01 5.56244731e-01 9.65888277e-02
4.63063359e-01 3.40542234e-02 6.00568056e-01 6.72570705e-01
6.11807227e-01 -1.00661731e+00 -1.68994352e-01 6.52546704e-01
6.63462460e-01 -1.21679389e+00 5.91730587e-02 -1.44652249e-02
-7.44585156e-01 8.70904446e-01 8.68642852e-02 -2.64496893e-01
2.20567778e-01 3.45497131e-01 -8.46602768e-02 -1.45047039e-01
-1.17345452e+00 -3.37796211e-01 1.10109681e-02 5.82290590e-01
4.68749553e-01 2.27464154e-01 -4.77753460e-01 7.41786599e-01
-5.67137361e-01 -7.87613541e-02 3.50686699e-01 8.83042097e-01
-4.19431746e-01 -1.49119556e+00 -3.99460346e-01 2.82340586e-01
-5.46058774e-01 -6.63333297e-01 -4.99442220e-01 7.66132653e-01
7.55923912e-02 1.04817247e+00 -2.29230136e-01 -1.24694914e-01
6.93324581e-02 5.78597605e-01 6.21361077e-01 -8.65857363e-01
-7.10276186e-01 -2.93377399e-01 3.48842323e-01 -4.68442798e-01
-2.32820317e-01 -3.80653232e-01 -8.96793544e-01 -3.77476454e-01
-2.48356517e-02 1.45773008e-01 3.59222203e-01 1.18824315e+00
7.19823614e-02 1.90198883e-01 1.30852163e-01 -7.38316357e-01
-1.29881120e+00 -9.07319427e-01 -4.56808895e-01 6.49597526e-01
1.07750734e-02 -6.69614613e-01 -3.84307653e-01 -3.90263200e-01] | [10.723957061767578, 8.514583587646484] |
b6ee53dd-8fb7-4559-bff9-d5c6d9090dd5 | blood-pressure-estimation-from | null | null | https://doi.org/10.3390/s19153420 | https://www.mdpi.com/1424-8220/19/15/3420/pdf | Blood Pressure Estimation from Photoplethysmogram Using a Spectro-Temporal Deep Neural Network | Blood pressure (BP) is a direct indicator of hypertension, a dangerous and potentially deadly condition. Regular monitoring of BP is thus important, but many people have aversion towards cuff-based devices, and their limitation is that they can only be used at rest. Using just a photoplethysmogram (PPG) to estimate BP is a potential solution investigated in our study. We analyzed the MIMIC III database for high-quality PPG and arterial BP waveforms, resulting in over 700 h of signals after preprocessing, belonging to 510 subjects. We then used the PPG alongside its first and second derivative as inputs into a novel spectro-temporal deep neural network with residual connections. We have shown in a leave-one-subject-out experiment that the network is able to model the dependency between PPG and BP, achieving mean absolute errors of 9.43 for systolic and 6.88 for diastolic BP. Additionally we have shown that personalization of models is important and substantially improves the results, while deriving a good general predictive model is difficult. We have made crucial parts of our study, especially the list of used subjects and our neural network code, publicly available, in an effort to provide a solid baseline and simplify potential comparison between future studies on an explicit MIMIC III subset. | ['Mitja Luštrek', 'Nejc Mlakar', 'Gašper Slapničar'] | 2019-08-04 | null | null | null | sensors-2019-8 | ['blood-pressure-estimation', 'photoplethysmography-ppg'] | ['medical', 'medical'] | [-5.35632558e-02 -1.55670587e-02 3.02977622e-01 -8.70124876e-01
-5.93267441e-01 -3.23528647e-01 -3.80235910e-02 9.22114402e-02
-5.47905028e-01 9.84153450e-01 2.00127438e-01 -3.99703950e-01
-1.75514221e-01 -6.15812540e-01 -3.67594004e-01 -5.71701169e-01
-7.84411132e-01 1.55644059e-01 -1.07326724e-01 1.07975956e-02
-7.38375783e-02 6.00055814e-01 -1.15347815e+00 7.75586441e-03
9.19992328e-01 1.05340171e+00 -5.10206044e-01 9.02104974e-01
3.39737058e-01 2.33987167e-01 -8.68343413e-01 -2.98335016e-01
5.29307723e-01 -6.62970304e-01 -3.65450263e-01 -2.29220733e-01
8.31643522e-01 -4.42122102e-01 -2.93731838e-01 4.98230308e-01
1.35445154e+00 2.02697620e-01 1.69320196e-01 -5.75689197e-01
-3.79739344e-01 4.68899846e-01 -1.88342929e-01 6.22892857e-01
-1.21058583e-01 6.22738481e-01 4.73958999e-01 -4.22197342e-01
2.75269374e-02 7.55502820e-01 9.26846147e-01 4.88684088e-01
-1.77724564e+00 -4.37745154e-01 -2.71829635e-01 3.86922853e-03
-1.44573951e+00 -8.38504195e-01 7.43813992e-01 -4.65617359e-01
8.48815322e-01 7.36629307e-01 1.13283002e+00 8.13324332e-01
3.37942749e-01 -2.34893449e-02 1.54154897e+00 -3.70320559e-01
1.28131211e-01 4.99451488e-01 4.05522376e-01 2.00092435e-01
2.23682076e-01 2.43736163e-01 5.40201478e-02 -3.26491594e-01
8.46249461e-01 -1.85350299e-01 -4.53170091e-01 1.01245858e-01
-6.49362028e-01 5.35719872e-01 2.06159398e-01 4.47264016e-01
-4.75494057e-01 8.55013430e-02 3.84207785e-01 3.94861519e-01
4.14770246e-01 5.42585492e-01 -4.14949179e-01 -4.96162832e-01
-1.13977432e+00 3.52873862e-01 1.17099619e+00 1.35838687e-01
1.91852719e-01 1.71285365e-02 -3.92891049e-01 1.07007647e+00
1.29814386e-01 6.15507126e-01 3.52921247e-01 -1.33431244e+00
1.84450835e-01 2.43354172e-01 2.71180630e-01 -1.28424573e+00
-1.01933944e+00 -7.15587854e-01 -9.38457251e-01 2.94551432e-01
7.94241965e-01 -4.73486573e-01 -7.16396570e-01 1.62745380e+00
-8.07308406e-02 -2.89892592e-03 -4.89815295e-01 1.11421514e+00
7.67976403e-01 1.27636552e-01 3.86825502e-01 -3.66517186e-01
1.47800374e+00 -3.15939486e-01 -5.37704766e-01 -2.93667406e-01
1.71093449e-01 -3.37112069e-01 8.56786489e-01 4.38467443e-01
-1.15296292e+00 -8.03769827e-01 -8.78343523e-01 1.01944223e-01
-2.10235968e-01 1.20805748e-01 4.17697787e-01 1.20993769e+00
-1.15612972e+00 1.30838072e+00 -8.85228753e-01 -3.05139482e-01
3.68660063e-01 3.42163742e-01 -1.65582180e-01 4.24496800e-01
-1.54219675e+00 1.20846701e+00 -1.46531850e-01 7.06989110e-01
-2.77205296e-02 -1.08514702e+00 -6.13400936e-01 1.37391791e-01
-2.10188270e-01 -6.00288987e-01 8.72045040e-01 -2.14124382e-01
-1.53037596e+00 8.60584855e-01 -1.64205372e-01 -4.93721455e-01
7.36956239e-01 -2.53190815e-01 -8.26360881e-01 3.26554894e-01
-3.90744150e-01 2.15968430e-01 6.87061071e-01 -5.75590491e-01
-3.04851793e-02 -5.24399877e-01 -5.07872224e-01 -2.46816516e-01
9.30074751e-02 -1.17154503e-02 -4.03302871e-02 -1.80381984e-01
1.45738851e-02 -8.32335770e-01 -5.75872660e-01 5.86396679e-02
-1.92082245e-02 3.89658779e-01 -2.00481117e-01 -1.36135411e+00
1.33921313e+00 -1.80460286e+00 -3.99502724e-01 5.96342027e-01
5.79570174e-01 6.39404118e-01 2.52937138e-01 2.00056314e-01
-4.41627741e-01 3.56487364e-01 -2.38275111e-01 -5.26168197e-02
8.92182533e-03 4.38082702e-02 -6.86911196e-02 6.93524182e-01
1.10935144e-01 1.03825974e+00 -4.28629756e-01 -2.15799421e-01
5.90839744e-01 6.97601795e-01 -2.22627938e-01 -8.58070254e-02
3.47264349e-01 7.04991221e-01 7.68692940e-02 1.41761182e-02
7.22039104e-01 -4.07730378e-02 5.00943884e-02 -3.35810721e-01
-2.69904971e-01 3.38835657e-01 -1.07392812e+00 1.17199588e+00
-3.52509916e-01 7.88296044e-01 -1.80378973e-01 -8.99708986e-01
1.34416163e+00 5.56908309e-01 6.01675093e-01 -9.55331326e-01
2.00455025e-01 2.92665184e-01 6.93678796e-01 -7.28911936e-01
-1.91095769e-01 -4.72945809e-01 4.85244632e-01 2.75976658e-01
-4.00219142e-01 8.49956274e-02 3.94703716e-01 -2.60595679e-01
7.88015723e-01 -6.45560622e-02 3.01722467e-01 -6.16411209e-01
5.00067532e-01 -4.77666199e-01 3.42481732e-01 9.63022351e-01
-7.46737838e-01 7.90272832e-01 6.85386419e-01 -7.58119464e-01
-9.01921391e-01 -8.69016647e-01 -9.26953077e-01 1.57966018e-01
-5.84525466e-01 -1.18482545e-01 -2.83247411e-01 1.39252678e-01
4.62891281e-01 7.69654632e-01 -6.49959445e-01 -2.21379787e-01
-8.12279284e-01 -1.14293075e+00 6.03999376e-01 5.10168314e-01
3.24734449e-01 -8.23531151e-01 -7.03039467e-01 7.12181866e-01
4.34359163e-02 -7.25563169e-01 2.97486335e-02 1.66132942e-01
-1.02465904e+00 -9.80547071e-01 -9.60384130e-01 -4.11428586e-02
1.12425879e-01 -4.55096245e-01 1.15808344e+00 -2.39031360e-01
-6.76345825e-01 1.31460249e-01 1.84924126e-01 -4.07416999e-01
-3.61033618e-01 -2.73078620e-01 1.48575142e-01 -6.07520528e-03
6.81769490e-01 -1.23135781e+00 -1.05606890e+00 4.60841954e-02
7.56803527e-03 -3.75941515e-01 2.69521356e-01 2.70335257e-01
2.48711109e-01 -4.79274541e-01 6.43919051e-01 -7.17004836e-01
8.66087556e-01 -1.01006322e-01 -6.44188166e-01 -2.29161948e-01
-8.19367886e-01 -4.03215110e-01 5.10634184e-01 -1.30076602e-01
-5.14180779e-01 -3.06407720e-01 -4.39877063e-01 -8.54772180e-02
-3.49787682e-01 4.05966699e-01 3.64901513e-01 -1.50861368e-01
1.26448750e+00 1.74682468e-01 6.23752236e-01 -6.80396438e-01
5.70012815e-03 6.11165464e-01 6.19968474e-01 -5.94935417e-01
3.87909412e-01 1.94741786e-01 3.63499880e-01 -1.12019801e+00
-5.31373143e-01 -2.75434911e-01 -7.63112187e-01 -2.68736899e-01
6.17008090e-01 -7.73207426e-01 -1.27019882e+00 3.07769954e-01
-6.48058295e-01 -3.98545265e-01 -3.36819142e-01 6.89521551e-01
-3.18122566e-01 4.50012207e-01 -5.78397334e-01 -1.02658045e+00
-6.49264753e-01 -7.01550603e-01 2.86051631e-01 2.98100233e-01
-4.79244500e-01 -1.03427684e+00 3.07997823e-01 5.65805472e-02
1.17249537e+00 7.72631407e-01 5.93699872e-01 -5.99285781e-01
6.69839233e-02 -6.29796267e-01 -2.47283071e-01 7.05294013e-01
5.00341117e-01 -1.62665639e-03 -1.16191864e+00 -1.71562940e-01
3.31536889e-01 1.71584666e-01 6.77433133e-01 8.32091093e-01
1.14860129e+00 -1.62610039e-01 2.16922015e-02 4.69439566e-01
1.31196058e+00 3.74865681e-01 9.71916795e-01 5.78328967e-02
4.69813526e-01 4.07406449e-01 -3.17268580e-01 4.13817734e-01
1.02434702e-01 6.90234661e-01 -3.30043167e-01 -4.50054288e-01
-5.64977713e-03 4.97677743e-01 -7.39544034e-02 4.74439263e-01
-5.65723777e-01 5.22341847e-01 -7.47626781e-01 1.11789495e-01
-1.40803862e+00 -1.00028539e+00 -5.82422972e-01 2.44589686e+00
1.04787683e+00 1.29787728e-01 5.46626210e-01 7.23880380e-02
4.45191056e-01 1.57432929e-02 -3.87662053e-01 -6.27747536e-01
1.32189497e-01 6.35147929e-01 4.30024177e-01 6.75310493e-01
-1.02675366e+00 1.02758184e-02 7.12256622e+00 -2.78609991e-01
-1.41071033e+00 -2.47512907e-01 8.71435761e-01 -2.08292246e-01
4.57685113e-01 -2.92417049e-01 -5.39223552e-01 8.68288517e-01
1.79101539e+00 -3.75979602e-01 3.73720407e-01 6.48443997e-01
8.87078762e-01 -1.54577166e-01 -1.16059518e+00 9.61356461e-01
-1.83930248e-01 -1.03747773e+00 -7.39505708e-01 -1.23888984e-01
5.04309162e-02 2.96759065e-02 -3.49397659e-01 1.99288130e-01
-3.90023232e-01 -1.09243476e+00 -1.67120267e-02 1.00753546e+00
8.01909566e-01 -1.59463793e-01 6.67001963e-01 -1.59214348e-01
-6.71120048e-01 2.55308151e-01 -3.93270850e-01 -3.50017488e-01
2.71742105e-01 1.04158735e+00 -4.68734503e-01 2.76890099e-01
5.92422664e-01 5.34396291e-01 -7.99176216e-01 1.48728979e+00
-5.17209135e-02 9.08306837e-01 -6.08079851e-01 1.17420763e-01
-3.71519923e-01 -3.83979797e-01 4.68336225e-01 1.18997300e+00
7.81343132e-02 1.73962206e-01 -8.09340775e-02 1.32752919e+00
6.57661080e-01 8.77674446e-02 -1.93724245e-01 1.15782544e-01
1.23527370e-01 1.09631574e+00 -3.48534107e-01 -4.10769969e-01
-5.21788716e-01 2.84697860e-01 -2.60838009e-02 4.61059570e-01
-6.64187789e-01 -7.56329834e-01 6.37371600e-01 2.92156368e-01
-1.12247989e-01 -8.62740632e-03 -7.60016680e-01 -9.55438852e-01
3.37903976e-01 -7.87653685e-01 3.26829076e-01 -4.20342922e-01
-1.18920338e+00 3.92642021e-01 1.05355140e-02 -8.45674336e-01
-2.82513499e-01 -5.15145779e-01 -7.07882106e-01 1.95360398e+00
-1.25512362e+00 -2.63713598e-01 -4.38389182e-01 1.45009100e-01
-2.14495510e-01 3.45383823e-01 1.08324981e+00 7.34938502e-01
-6.71448946e-01 4.92676675e-01 -3.92103523e-01 1.72362834e-01
8.07847917e-01 -1.27725458e+00 4.47517991e-01 6.52746379e-01
-4.00307059e-01 1.14343762e+00 9.66177762e-01 -5.16238630e-01
-9.34719563e-01 -8.94406796e-01 1.03842258e+00 -4.87981290e-01
4.00829971e-01 -1.50401071e-01 -9.81109500e-01 5.16137242e-01
-1.85442045e-01 2.15473324e-01 7.53402293e-01 1.53441995e-01
1.93557233e-01 -5.24624050e-01 -1.14819026e+00 5.37086666e-01
5.91060996e-01 -2.85562485e-01 -5.87319016e-01 3.11869472e-01
2.60934204e-01 -6.19302630e-01 -1.56277347e+00 2.64328361e-01
9.04360771e-01 -1.12196052e+00 9.63315308e-01 -5.43933392e-01
3.36579144e-01 -8.24535489e-02 3.83004367e-01 -1.31872630e+00
-5.78017235e-01 -7.81148612e-01 -1.68158859e-03 1.08891273e+00
4.91922438e-01 -1.21093881e+00 7.97313571e-01 1.67380321e+00
-1.98833182e-01 -6.52560472e-01 -7.73937106e-01 -7.51358747e-01
1.77700073e-01 -3.46603811e-01 2.20365584e-01 5.81028402e-01
1.71823993e-01 2.29334980e-02 -6.29943132e-01 1.40393022e-02
6.74204111e-01 1.30885690e-01 7.19005466e-01 -1.31805778e+00
-4.93284613e-01 -5.23521543e-01 -4.44466323e-01 -4.90422040e-01
-5.43996215e-01 -7.70071030e-01 -2.36019075e-01 -1.35189235e+00
-1.44439429e-01 -3.84390116e-01 -4.50258702e-01 5.15767217e-01
-2.93319792e-01 3.66394758e-01 -8.07247460e-02 -1.71554074e-01
5.22981226e-01 1.39733672e-01 1.08454227e+00 1.53443247e-01
-8.69703650e-01 3.72972161e-01 -8.44933510e-01 2.43237495e-01
1.14536560e+00 -4.29120511e-01 -2.01241046e-01 2.70327181e-01
-1.53095111e-01 1.94651946e-01 5.93316913e-01 -1.17106462e+00
-1.86633855e-01 1.39445454e-01 9.19224501e-01 -2.06392393e-01
5.13382852e-01 -6.92954659e-01 3.49500299e-01 8.15586925e-01
-3.34468752e-01 -3.92163366e-01 5.60223401e-01 -1.75733268e-02
1.12215430e-01 -1.46908984e-01 8.11784863e-01 -4.57520068e-01
-1.12804316e-01 1.78573176e-01 -3.10473323e-01 1.29419833e-01
3.68269563e-01 -2.29207546e-01 -3.38448942e-01 -4.19484675e-01
-1.29428363e+00 1.34330481e-01 -1.57486796e-01 7.91161954e-02
1.62613183e-01 -1.04886079e+00 -1.01944637e+00 4.14200604e-01
-1.52461916e-01 -5.82895517e-01 4.12616074e-01 1.39454794e+00
-6.25591159e-01 6.65294647e-01 -4.30011570e-01 -5.83668172e-01
-1.06563079e+00 3.56897742e-01 9.46586132e-01 2.25053698e-01
-1.29902148e+00 4.29769665e-01 -4.75208461e-01 -2.98284367e-02
3.62101197e-01 -4.25534606e-01 -2.42522463e-01 -1.14612766e-01
8.18320453e-01 6.02006555e-01 2.78883636e-01 -6.32313490e-02
-3.49220723e-01 3.15936416e-01 1.47368312e-01 -1.87850893e-02
1.37447250e+00 -2.18795702e-01 -3.10050130e-01 6.15737438e-01
1.04394460e+00 -2.85094026e-02 -8.23835433e-01 -1.05647575e-02
-2.89558887e-01 -4.95889485e-01 -1.26401126e-01 -1.11335599e+00
-1.05315709e+00 9.40631926e-01 1.07047009e+00 3.78527284e-01
9.35184598e-01 -5.67072630e-01 4.48683739e-01 1.57951921e-01
-6.91627711e-02 -8.53667617e-01 -6.87033415e-01 -6.47464246e-02
1.02630532e+00 -8.49600434e-01 2.91188151e-01 -2.54874825e-01
-2.42448270e-01 1.19665384e+00 3.50152493e-01 -2.68552303e-01
8.90869081e-01 -5.85460737e-02 4.38280821e-01 -1.86299622e-01
-4.05440986e-01 1.05997786e-01 3.75045270e-01 7.10018516e-01
8.27947676e-01 7.06389323e-02 -8.24544430e-01 5.01532555e-01
-4.41265285e-01 4.47236568e-01 6.25149846e-01 5.57686925e-01
-4.04747546e-01 -9.64502931e-01 -1.70496389e-01 9.73399758e-01
-6.65643096e-01 -1.22095510e-01 6.72150552e-02 8.45437646e-01
-2.84215212e-01 8.82813096e-01 -6.54182434e-02 3.87425832e-02
6.52829111e-01 7.44429171e-01 5.27198851e-01 -3.56375813e-01
-6.35003567e-01 -3.75470426e-03 2.80608743e-01 -6.07057393e-01
-2.04691023e-01 -6.02162123e-01 -7.25135267e-01 -2.11206853e-01
2.14230612e-01 3.03673018e-02 6.49567842e-01 6.97836399e-01
4.83370036e-01 6.39240026e-01 4.97765332e-01 -7.55384564e-01
-8.81379962e-01 -1.15227985e+00 -9.43289399e-01 3.50516975e-01
5.07097900e-01 -3.32662046e-01 -4.88334805e-01 7.44261891e-02] | [14.096305847167969, 3.001804828643799] |
f8875770-00e1-4fb3-9c99-fbbf014ffcb6 | adn-artifact-disentanglement-network-for | 1908.01104 | null | https://arxiv.org/abs/1908.01104v4 | https://arxiv.org/pdf/1908.01104v4.pdf | ADN: Artifact Disentanglement Network for Unsupervised Metal Artifact Reduction | Current deep neural network based approaches to computed tomography (CT) metal artifact reduction (MAR) are supervised methods that rely on synthesized metal artifacts for training. However, as synthesized data may not accurately simulate the underlying physical mechanisms of CT imaging, the supervised methods often generalize poorly to clinical applications. To address this problem, we propose, to the best of our knowledge, the first unsupervised learning approach to MAR. Specifically, we introduce a novel artifact disentanglement network that disentangles the metal artifacts from CT images in the latent space. It supports different forms of generations (artifact reduction, artifact transfer, and self-reconstruction, etc.) with specialized loss functions to obviate the need for supervision with synthesized data. Extensive experiments show that when applied to a synthesized dataset, our method addresses metal artifacts significantly better than the existing unsupervised models designed for natural image-to-image translation problems, and achieves comparable performance to existing supervised models for MAR. When applied to clinical datasets, our method demonstrates better generalization ability over the supervised models. The source code of this paper is publicly available at https://github.com/liaohaofu/adn. | ['Wei-An Lin', 'S. Kevin Zhou', 'Jiebo Luo', 'Haofu Liao'] | 2019-08-03 | null | null | null | null | ['medical-image-generation', 'metal-artifact-reduction'] | ['medical', 'medical'] | [ 4.25968856e-01 1.62082687e-01 1.90274920e-02 -3.50648016e-01
-1.18131506e+00 -1.34818302e-02 1.90764859e-01 -7.40872994e-02
-7.99569711e-02 7.25429058e-01 3.55162352e-01 -2.73774236e-01
-1.47284493e-01 -6.35181606e-01 -7.54683375e-01 -7.27892935e-01
1.00456655e-01 4.30537969e-01 -2.42502149e-02 1.29319593e-01
-1.02626406e-01 4.60241646e-01 -7.26991951e-01 4.81544733e-01
1.03877318e+00 8.97700369e-01 8.78959224e-02 4.32423711e-01
1.81303725e-01 1.30392790e+00 -5.05842626e-01 9.49100032e-02
1.47958472e-01 -6.68522835e-01 -9.01886284e-01 1.29683599e-01
3.04066062e-01 -6.71226025e-01 -7.18517601e-01 9.67687964e-01
6.04122758e-01 -2.40426034e-01 8.24086547e-01 -1.07239258e+00
-8.83987725e-01 7.01388478e-01 -3.47628325e-01 3.77182454e-01
4.50359322e-02 1.71690211e-01 4.94824767e-01 -1.03249097e+00
5.59726477e-01 6.46976173e-01 8.90440762e-01 8.94609451e-01
-1.45970547e+00 -6.56854033e-01 -3.11186522e-01 -9.41880420e-02
-1.13462472e+00 -1.47131681e-01 9.17595327e-01 -5.59149742e-01
6.22924447e-01 2.23253325e-01 4.95956808e-01 1.51429796e+00
4.31910127e-01 9.00992632e-01 1.21434546e+00 -3.06529254e-01
2.41096839e-01 -2.35846080e-03 2.51042750e-02 8.09155166e-01
2.98582554e-01 3.51585239e-01 -5.80579400e-01 -3.19105834e-01
1.18877971e+00 2.19527394e-01 -6.63964272e-01 -4.96085554e-01
-1.15086555e+00 7.18292236e-01 5.44998288e-01 3.30325216e-01
-3.60278815e-01 5.43554723e-01 3.29531372e-01 1.42076552e-01
5.49722552e-01 5.75145185e-01 -2.24997029e-01 1.28954589e-01
-9.97872889e-01 -4.38714214e-02 4.13148671e-01 9.13004220e-01
2.76723176e-01 4.44587409e-01 -1.67884037e-01 7.61777282e-01
1.16930418e-01 4.62491184e-01 8.79269361e-01 -8.84376645e-01
2.56113291e-01 2.06576273e-01 -6.36746138e-02 -6.13340735e-01
-4.97599810e-01 -4.96307075e-01 -1.10954475e+00 3.60150695e-01
1.68332472e-01 -3.02146245e-02 -1.30306673e+00 1.58917594e+00
-1.46217525e-01 3.65521967e-01 -2.00770959e-01 9.20997977e-01
1.06463885e+00 2.39666447e-01 1.05831861e-01 -1.58528000e-01
1.11444545e+00 -9.57035482e-01 -8.79949510e-01 -9.00998712e-02
5.85401893e-01 -7.15410471e-01 1.19514978e+00 5.93937635e-01
-1.27671921e+00 -3.30398679e-01 -1.27399397e+00 -1.09035298e-01
-7.67260790e-02 1.24805821e-02 6.97735369e-01 5.61607540e-01
-9.56249058e-01 8.07554722e-01 -1.39383841e+00 -5.86379841e-02
6.50198221e-01 2.56267905e-01 -7.39209130e-02 -7.83265606e-02
-1.18343580e+00 9.57293272e-01 -2.20393986e-01 6.55807331e-02
-1.11058569e+00 -1.00702274e+00 -8.21448743e-01 -2.22558409e-01
1.69637263e-01 -8.85165036e-01 1.60331309e+00 -1.18129730e+00
-1.51902854e+00 6.55104101e-01 1.03380322e-01 -4.61741686e-01
7.44612932e-01 -3.87633622e-01 -2.57385761e-01 3.95994931e-01
1.19773142e-01 4.11707312e-01 7.70637870e-01 -1.39410222e+00
2.55574673e-01 -2.23819837e-02 -4.26500916e-01 -9.03806016e-02
-3.48954827e-01 -4.62349690e-02 -9.27472394e-03 -1.16534531e+00
3.39655995e-01 -9.42356646e-01 -4.80100989e-01 5.20213425e-01
-4.15989429e-01 3.88181418e-01 4.36568648e-01 -6.02777481e-01
8.12140226e-01 -2.03278685e+00 -2.02350784e-03 1.49233401e-01
6.66908503e-01 -2.66771261e-02 8.89956281e-02 2.65098095e-01
-5.69095790e-01 -2.58673774e-03 -8.74564528e-01 -3.79698217e-01
-4.19678837e-01 1.79514095e-01 -3.28107737e-02 5.12318790e-01
2.61352658e-01 1.08113670e+00 -1.04311085e+00 -4.87786531e-01
1.22296050e-01 4.80066359e-01 -4.43103254e-01 1.74254403e-01
1.12455070e-01 9.13915455e-01 -4.86370176e-01 4.75858152e-01
6.60906851e-01 -7.75972188e-01 7.67893866e-02 -3.02503914e-01
3.45238984e-01 4.09951508e-01 -6.38903320e-01 2.10807133e+00
-6.49763882e-01 6.92357421e-01 -3.38009089e-01 -7.96441972e-01
4.53443825e-01 7.37388790e-01 1.06129169e+00 -5.73547184e-01
2.86006093e-01 4.16893423e-01 -8.97412822e-02 -7.12833524e-01
3.15156430e-02 -5.53667009e-01 2.24012583e-01 6.65116370e-01
-8.98632705e-02 -2.88036376e-01 -4.93964761e-01 2.49813572e-01
1.27514911e+00 2.53008809e-02 2.46580932e-02 -1.98035568e-01
6.25338629e-02 6.10045046e-02 4.15662616e-01 7.96805859e-01
-2.75280923e-01 1.21410620e+00 9.63162035e-02 -5.55232286e-01
-1.12811255e+00 -1.59286618e+00 -3.15279126e-01 1.72901839e-01
5.60740046e-02 -4.24423069e-02 -8.89205456e-01 -6.94139957e-01
-2.19151661e-01 4.97210681e-01 -6.73574686e-01 -3.32582891e-01
-6.55446410e-01 -9.22309220e-01 7.75320590e-01 9.10624206e-01
4.24955755e-01 -1.03944468e+00 -5.32370508e-01 2.47476995e-01
-4.76228297e-01 -1.10138047e+00 -5.84184468e-01 2.46603891e-01
-1.30684400e+00 -1.04594827e+00 -8.54730546e-01 -7.50529766e-01
1.01187098e+00 1.33041963e-01 1.15556145e+00 3.28756541e-01
-7.54061520e-01 3.35250467e-01 -2.51585007e-01 -3.55869859e-01
-6.45067275e-01 -3.55041504e-01 1.01639763e-01 -2.15399697e-01
1.07263483e-01 -7.49659657e-01 -9.10522819e-01 1.59358189e-01
-1.28558552e+00 3.51354539e-01 6.43204033e-01 1.00878263e+00
6.18239462e-01 -1.57934457e-01 3.70890528e-01 -1.17855203e+00
4.73522425e-01 -3.64700437e-01 -1.46146476e-01 -7.03254864e-02
-7.46393800e-01 1.48502946e-01 5.47591329e-01 -6.76655710e-01
-1.09355712e+00 9.46005136e-02 -3.84910032e-02 -5.83999693e-01
-1.08457983e-01 4.28193629e-01 2.29589209e-01 -1.66332453e-01
1.08877599e+00 3.32214087e-01 4.09374060e-03 -5.12928426e-01
1.29369557e-01 4.37397540e-01 4.43602026e-01 -5.92026830e-01
8.44974279e-01 7.85706997e-01 6.95115700e-02 -4.15892988e-01
-7.38715172e-01 -4.89556827e-02 -5.44986784e-01 -7.30670337e-03
8.60658944e-01 -8.13639104e-01 1.85058061e-02 4.06969130e-01
-1.06057549e+00 -5.13426125e-01 -4.75597590e-01 7.40685761e-01
-8.33561063e-01 5.78204811e-01 -1.19961405e+00 -3.61541808e-01
-5.48669636e-01 -1.36358380e+00 7.99465537e-01 -2.71614641e-01
-3.29353869e-01 -8.34451139e-01 1.11465178e-01 2.71523416e-01
5.63642025e-01 5.15124142e-01 1.04828215e+00 -2.17484683e-01
-7.13474035e-01 -3.20195854e-02 -2.10350230e-01 5.26381612e-01
5.92605293e-01 -4.68111187e-01 -9.99936759e-01 -1.79801583e-01
5.31823277e-01 -3.27252030e-01 8.16422343e-01 6.39566958e-01
1.69920576e+00 -1.96723580e-01 -4.03572172e-02 7.56748497e-01
1.41006517e+00 1.38031736e-01 8.78032029e-01 3.69189084e-01
6.53075159e-01 1.53889507e-01 4.39328328e-02 2.53424615e-01
-2.71977223e-02 4.47568357e-01 2.93394744e-01 -5.89708149e-01
-6.33657575e-01 -1.81277275e-01 6.61403462e-02 9.90578055e-01
1.37897683e-02 7.98616558e-02 -9.38291550e-01 5.14024079e-01
-1.65602958e+00 -6.16816401e-01 -3.01345348e-01 1.93470967e+00
1.02661014e+00 5.87945543e-02 -2.74931282e-01 7.01758564e-02
3.54809582e-01 -7.53003359e-02 -5.51440537e-01 -5.61799146e-02
1.31411031e-01 5.87333083e-01 4.87792134e-01 3.45681608e-01
-9.04109120e-01 5.22514164e-01 6.63482237e+00 5.58613598e-01
-1.17260408e+00 5.63763857e-01 7.22209692e-01 -1.98923677e-01
-3.14665318e-01 -3.39561611e-01 1.85271576e-01 3.74605268e-01
7.13872492e-01 2.60804057e-01 1.99682370e-01 6.73243821e-01
5.28499544e-01 2.30836093e-01 -1.35075438e+00 8.95006478e-01
1.14109501e-01 -1.39269483e+00 7.69107640e-02 -1.85011789e-01
8.34862947e-01 1.10879555e-01 2.92010397e-01 -2.03437552e-01
1.29520327e-01 -1.19581544e+00 5.29433310e-01 5.31790793e-01
9.50812638e-01 -3.11420113e-01 6.96075082e-01 -4.78489511e-02
-5.77249944e-01 3.04665655e-01 -1.36585996e-01 3.49036902e-01
1.03853762e-01 7.14162827e-01 -6.68061435e-01 5.27989626e-01
6.44983113e-01 5.46031475e-01 -4.11576003e-01 1.06655705e+00
-3.84043217e-01 7.22952425e-01 -6.25810698e-02 3.37013692e-01
4.92950715e-02 2.97861785e-01 4.86366600e-01 9.97983396e-01
3.06090474e-01 1.70541212e-01 -5.62733673e-02 1.18293035e+00
-9.91120338e-02 1.16331466e-02 -5.93648255e-01 2.81828225e-01
6.65671797e-03 8.49275351e-01 -6.15645409e-01 -4.23873514e-01
-4.10935014e-01 1.29465401e+00 7.21758753e-02 4.72467542e-01
-1.07024872e+00 -2.09766358e-01 2.42720306e-01 4.27103102e-01
-1.75821587e-01 -1.65982723e-01 -8.30432892e-01 -1.35177422e+00
4.74823564e-02 -9.46779013e-01 2.88334966e-01 -1.18918049e+00
-1.53412223e+00 9.48926568e-01 9.35677160e-03 -1.70037127e+00
1.01537094e-01 -7.09764481e-01 -6.01424217e-01 6.29526258e-01
-1.41233265e+00 -1.07657516e+00 -4.73205894e-01 6.21046603e-01
5.72304428e-01 1.50638046e-02 9.13849771e-01 4.37334836e-01
-2.33685777e-01 6.89759433e-01 1.97640553e-01 3.61179203e-01
7.93016613e-01 -1.17098451e+00 1.78565919e-01 9.11357820e-01
3.75211537e-02 5.68691552e-01 6.89919949e-01 -7.14592516e-01
-1.12768745e+00 -1.14167297e+00 4.93454933e-01 -5.65767705e-01
6.91677034e-01 -6.55123517e-02 -1.00602365e+00 7.06529081e-01
2.72896290e-01 1.66099742e-01 8.86490047e-01 -3.57319951e-01
-4.07965243e-01 5.11493348e-02 -1.24623179e+00 6.76749825e-01
9.13005054e-01 -4.67691153e-01 -5.40015280e-01 6.90437496e-01
6.33986294e-01 -6.32988214e-01 -9.55204844e-01 3.59853357e-01
2.78862864e-01 -7.23357618e-01 1.00253356e+00 -5.84045947e-01
1.08419347e+00 -1.44279614e-01 8.84548575e-02 -1.25363898e+00
-4.09717292e-01 -4.24054235e-01 -1.15447268e-01 3.73174965e-01
5.74245811e-01 -5.21244586e-01 9.65070009e-01 7.46531725e-01
-4.32038724e-01 -9.15985644e-01 -9.51981127e-01 -9.31074321e-01
3.60163361e-01 -6.17297590e-01 4.70357150e-01 1.16716862e+00
-1.55282244e-02 -1.53033733e-01 -5.13177514e-01 4.28105116e-01
9.85444427e-01 1.62652731e-02 1.97410405e-01 -7.59201109e-01
-6.72871768e-01 -3.79025847e-01 -2.37386808e-01 -8.29521775e-01
-2.98785299e-01 -1.04289186e+00 2.93241233e-01 -1.53082252e+00
4.64917809e-01 -4.03635085e-01 -5.59202015e-01 6.25937939e-01
-5.71240038e-02 5.85761011e-01 -3.76864597e-02 3.91862303e-01
-1.75741650e-02 5.13900161e-01 1.53022432e+00 -3.84487808e-01
-6.36345372e-02 -2.15250835e-01 -6.25558734e-01 6.76591218e-01
9.82981086e-01 -8.73433113e-01 -4.31130320e-01 -8.50878358e-01
-1.27021641e-01 2.45603278e-01 5.34974158e-01 -9.78837907e-01
1.60312444e-01 -1.50310025e-01 5.26043534e-01 -2.74358630e-01
1.82893977e-01 -7.65329719e-01 3.25993061e-01 8.40431571e-01
-4.95269358e-01 -7.94159472e-02 2.70133048e-01 4.09063101e-01
-1.12971343e-01 -3.02616060e-01 9.40890193e-01 -3.95912260e-01
9.89850890e-03 5.10974944e-01 -8.07619870e-01 -9.61212665e-02
4.23449367e-01 -1.38594344e-01 -1.39056474e-01 -4.39568102e-01
-9.66254532e-01 -1.16402403e-01 2.59544790e-01 1.79860011e-01
1.08807302e+00 -1.55168331e+00 -7.45399714e-01 1.33183762e-01
1.92474425e-01 2.06492782e-01 3.10284138e-01 1.01800752e+00
-8.16982508e-01 2.07854539e-01 -1.00923061e-01 -6.28410399e-01
-9.13836360e-01 4.95817333e-01 6.55524731e-01 -4.20276165e-01
-1.04807174e+00 7.27255940e-01 5.61605096e-01 -2.16697529e-01
5.88135701e-03 -4.76538301e-01 5.42959988e-01 -7.85388410e-01
5.18227756e-01 2.21721277e-01 2.33227015e-01 -1.86943397e-01
-1.73093259e-01 3.70085955e-01 -3.04822266e-01 -7.90643618e-02
1.50436223e+00 1.05203263e-01 -6.99045602e-03 3.06773841e-01
1.18349004e+00 -1.97350711e-01 -9.52315152e-01 -4.27561015e-01
-1.98665470e-01 -4.02844995e-01 2.58397341e-01 -1.04961646e+00
-1.41020501e+00 8.77430618e-01 9.06893849e-01 -2.67569214e-01
1.28391254e+00 -1.54399887e-01 9.77478147e-01 2.20570400e-01
3.30172360e-01 -8.07844818e-01 6.04350269e-01 1.22949360e-02
1.01548052e+00 -1.20360422e+00 1.21262886e-01 -6.10617101e-01
-6.20159686e-01 1.05223048e+00 5.96169829e-01 -3.14162433e-01
7.90628195e-01 4.81093794e-01 4.30397123e-01 -4.43650782e-01
-3.53628427e-01 3.72171253e-01 1.91246673e-01 5.54248154e-01
5.27272761e-01 8.19523185e-02 -2.83344328e-01 5.53967178e-01
-4.94903736e-02 3.38240296e-01 8.86419833e-01 1.15704763e+00
-3.79405841e-02 -1.11592555e+00 -4.23816413e-01 4.87165689e-01
-5.72743952e-01 -3.93565446e-01 -9.30376127e-02 6.30769908e-01
-6.65430874e-02 6.22075319e-01 -3.22696328e-01 -9.21624899e-02
2.89064229e-01 -5.59580065e-02 7.28427410e-01 -7.68559039e-01
-5.73476017e-01 2.24395886e-01 -1.79913610e-01 -5.53423524e-01
-3.91853571e-01 -3.90912205e-01 -1.42157209e+00 4.55761626e-02
-2.86377996e-01 -9.80871394e-02 3.77447516e-01 7.47637630e-01
1.75728887e-01 9.15509224e-01 6.44040406e-01 -6.62874043e-01
-4.50800866e-01 -1.01190472e+00 -5.63293815e-01 7.21101642e-01
4.77428198e-01 -6.68510616e-01 -3.49514484e-01 2.55813748e-01] | [13.554051399230957, -2.5182135105133057] |
56e12a85-55e7-4e1e-b2fb-c8713e639e3f | padl-language-directed-physics-based | 2301.13868 | null | https://arxiv.org/abs/2301.13868v1 | https://arxiv.org/pdf/2301.13868v1.pdf | PADL: Language-Directed Physics-Based Character Control | Developing systems that can synthesize natural and life-like motions for simulated characters has long been a focus for computer animation. But in order for these systems to be useful for downstream applications, they need not only produce high-quality motions, but must also provide an accessible and versatile interface through which users can direct a character's behaviors. Natural language provides a simple-to-use and expressive medium for specifying a user's intent. Recent breakthroughs in natural language processing (NLP) have demonstrated effective use of language-based interfaces for applications such as image generation and program synthesis. In this work, we present PADL, which leverages recent innovations in NLP in order to take steps towards developing language-directed controllers for physics-based character animation. PADL allows users to issue natural language commands for specifying both high-level tasks and low-level skills that a character should perform. We present an adversarial imitation learning approach for training policies to map high-level language commands to low-level controls that enable a character to perform the desired task and skill specified by a user's commands. Furthermore, we propose a multi-task aggregation method that leverages a language-based multiple-choice question-answering approach to determine high-level task objectives from language commands. We show that our framework can be applied to effectively direct a simulated humanoid character to perform a diverse array of complex motor skills. | ['Xue Bin Peng', 'Sanja Fidler', 'Yunrong Guo', 'Jordan Juravsky'] | 2023-01-31 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 3.02453935e-01 1.44580938e-02 -1.69985324e-01 -2.05255806e-01
-8.50162685e-01 -8.43172669e-01 7.88238049e-01 -9.35726538e-02
-3.52387071e-01 6.45577729e-01 4.44550142e-02 -4.49350208e-01
2.72449732e-01 -7.79180288e-01 -7.65955210e-01 -4.35307950e-01
1.16526857e-02 4.55050409e-01 3.23022246e-01 -6.26392663e-01
1.66074708e-01 7.69285917e-01 -1.68192267e+00 2.22974226e-01
7.98226774e-01 2.11642534e-01 5.91772914e-01 1.28786039e+00
7.71680102e-02 1.10867941e+00 -6.17844522e-01 1.37896553e-01
1.59274265e-01 -7.32155085e-01 -7.11017191e-01 -9.27506164e-02
1.37058899e-01 -4.89107102e-01 -6.15560114e-02 7.35091269e-01
3.93706650e-01 4.13462192e-01 6.99792087e-01 -1.35349798e+00
-1.08735435e-01 2.09699675e-01 9.25098360e-02 -4.25207585e-01
9.26351786e-01 1.06172800e+00 9.08978760e-01 -1.43562317e-01
7.56212831e-01 1.40769136e+00 1.47081539e-01 1.17429507e+00
-1.36846602e+00 -5.65943003e-01 -3.44477296e-02 -4.22439992e-01
-8.18436146e-01 -2.33498782e-01 4.98325258e-01 -6.78974807e-01
8.90479147e-01 2.92481810e-01 1.04060149e+00 1.25619793e+00
4.70612556e-01 7.63619125e-01 9.73175049e-01 -4.06234026e-01
3.24674457e-01 -1.15733847e-01 -4.11036164e-01 8.56185317e-01
-3.02590877e-01 3.34912896e-01 -4.45342869e-01 -1.89112023e-01
1.33605802e+00 -6.13145292e-01 -7.51226991e-02 -1.21909693e-01
-1.45932853e+00 8.49143624e-01 1.19576212e-02 1.36686042e-01
-5.04187405e-01 7.65848458e-01 2.72859514e-01 2.84623832e-01
-2.23392978e-01 1.23973215e+00 -4.54880744e-01 -6.88911974e-01
-5.45022786e-01 9.95810926e-01 1.08598495e+00 8.49994659e-01
3.63449901e-01 4.78286356e-01 -5.39659142e-01 3.94303888e-01
1.45820454e-01 6.34058356e-01 2.69559085e-01 -1.76913548e+00
8.86237994e-02 3.82523865e-01 4.20718133e-01 -7.01212883e-01
-2.07577735e-01 1.48010626e-01 -3.87864292e-01 1.04869044e+00
4.67221528e-01 -4.42327023e-01 -6.80764973e-01 2.14740348e+00
3.87403876e-01 -6.04693219e-02 2.20056266e-01 1.16412020e+00
3.73502582e-01 1.03075671e+00 2.73110539e-01 7.81501755e-02
1.35748231e+00 -4.82253969e-01 -3.05260777e-01 -1.85121566e-01
5.39080322e-01 -5.81274450e-01 1.70570076e+00 4.26140368e-01
-1.37197304e+00 -5.59136212e-01 -7.51185894e-01 1.18674219e-01
1.10253371e-01 -2.80270070e-01 5.87392449e-01 2.92730838e-01
-1.05047798e+00 7.44323671e-01 -9.39752281e-01 -1.92242116e-01
4.12749313e-02 5.45023978e-01 -2.52330452e-01 5.01407981e-01
-1.09944034e+00 9.87421811e-01 2.44739980e-01 -6.64748132e-01
-1.36594474e+00 -7.07040131e-01 -1.01278746e+00 -9.87578928e-02
4.44223225e-01 -9.90520895e-01 1.74179554e+00 -9.46960032e-01
-2.22700071e+00 6.13425553e-01 1.87523708e-01 -3.62919509e-01
4.07688171e-01 -2.02114880e-01 1.51953727e-01 2.92743653e-01
1.56397760e-01 1.26331604e+00 8.44635129e-01 -1.04051828e+00
-6.12541735e-01 2.67576009e-01 5.19265234e-01 4.60145563e-01
-7.80808702e-02 2.70248294e-01 -3.77360255e-01 -6.97301269e-01
-7.58526921e-01 -1.19972813e+00 -3.46776903e-01 5.45549870e-01
-4.59818810e-01 -1.96336374e-01 7.88003147e-01 -3.40865940e-01
7.14254498e-01 -1.74575663e+00 7.44040668e-01 -5.83243482e-02
5.24459686e-03 2.56473273e-01 -2.79568136e-01 5.24016798e-01
3.26631099e-01 8.78000185e-02 -2.38003165e-01 -1.37239486e-01
1.90604091e-01 2.62358010e-01 -3.27187508e-01 6.14372417e-02
4.90529507e-01 9.29380894e-01 -1.11923981e+00 -6.04861438e-01
2.13354245e-01 3.90470803e-01 -9.07324255e-01 8.36528838e-01
-9.66138005e-01 1.04100430e+00 -8.28856885e-01 1.86977476e-01
-4.56825942e-01 3.48839499e-02 -7.91177750e-02 3.13813388e-01
-2.28671551e-01 2.36680001e-01 -7.99030781e-01 1.66443717e+00
-8.75196934e-01 5.77714384e-01 3.80975127e-01 -3.63988489e-01
7.22265720e-01 2.97831267e-01 4.03075784e-01 -1.25109509e-01
1.40585855e-01 -1.90999098e-02 2.41062880e-01 -6.84401572e-01
3.75068069e-01 -3.25519383e-01 -7.44470417e-01 7.57129192e-01
-3.38371158e-01 -1.41322315e+00 2.27344036e-01 2.17463627e-01
1.23372769e+00 8.22260797e-01 2.80165344e-01 -2.68496364e-01
5.61490178e-01 4.68489677e-01 3.38387728e-01 5.71191251e-01
8.15017521e-02 3.20281714e-01 5.75468004e-01 -4.56981100e-02
-1.26928794e+00 -9.32746172e-01 6.69665754e-01 1.39066815e+00
-2.47540399e-01 -7.95210898e-02 -9.33545887e-01 -7.90465772e-02
-9.31875557e-02 1.05038917e+00 -2.29195103e-01 -1.85646981e-01
-9.76343274e-01 1.96423128e-01 8.93449426e-01 4.31022853e-01
2.02088878e-01 -1.60839021e+00 -1.22470272e+00 2.89969951e-01
2.24244632e-02 -1.10731268e+00 -9.42142427e-01 -1.34051010e-01
-5.78522742e-01 -8.79180849e-01 -8.00965309e-01 -9.12830472e-01
4.40056473e-01 -3.17897230e-01 1.03666794e+00 2.10030563e-02
-2.06824169e-01 6.59367442e-01 -1.21143989e-01 -2.43622556e-01
-1.22295296e+00 1.02484098e-03 2.18914419e-01 -4.01638627e-01
-3.45428348e-01 -5.18461049e-01 -3.26386958e-01 1.37776658e-01
-1.03006387e+00 6.17146611e-01 3.87728572e-01 7.01253057e-01
3.40934187e-01 -3.88537288e-01 4.45805848e-01 -3.87865245e-01
1.21345186e+00 -4.09931690e-02 -7.78203189e-01 -7.07753468e-03
-2.09554862e-02 4.26306397e-01 1.19145823e+00 -9.66109514e-01
-7.52791107e-01 4.23371434e-01 -2.86774606e-01 -4.36286032e-01
-4.03763384e-01 2.50906795e-01 -1.47403926e-01 -3.98821197e-03
6.98824227e-01 3.93481016e-01 2.81638265e-01 8.83147195e-02
6.88969195e-01 5.24480462e-01 8.55970860e-01 -1.31953025e+00
8.53306234e-01 3.33610103e-02 1.60255462e-01 -9.02742386e-01
-1.98371679e-01 5.55466153e-02 -3.72881025e-01 -3.86650324e-01
1.26597154e+00 -7.68738627e-01 -1.35676062e+00 5.79540372e-01
-1.26085389e+00 -1.04228663e+00 -4.24238831e-01 2.65010089e-01
-1.26478410e+00 3.40064205e-02 -6.73147678e-01 -7.30717301e-01
-3.34049225e-01 -1.63050866e+00 1.32841611e+00 2.58932114e-01
-9.80476916e-01 -7.05597103e-01 6.69048280e-02 9.69861597e-02
4.15700734e-01 5.38484514e-01 1.02940476e+00 -3.42050612e-01
-5.96160650e-01 1.92789771e-02 4.62902099e-01 1.42121673e-01
8.27157795e-02 1.92208678e-01 -2.95249492e-01 -2.82423377e-01
-3.71566355e-01 -8.86089444e-01 -1.47078171e-01 1.67450517e-01
8.94887984e-01 -4.11640406e-01 -3.90961841e-02 4.83783096e-01
9.29813325e-01 2.92116195e-01 5.06813884e-01 3.62488218e-02
6.93863094e-01 6.99617326e-01 6.36664212e-01 2.79625416e-01
3.73411357e-01 9.46985483e-01 8.83505195e-02 1.81559190e-01
-3.22032534e-02 -5.07016838e-01 6.93398476e-01 5.42146325e-01
-4.09489013e-02 -7.79121667e-02 -8.39213789e-01 3.69697332e-01
-1.60265148e+00 -1.00862849e+00 3.28898840e-02 2.03012514e+00
1.33757699e+00 2.04508871e-01 4.43392903e-01 -2.55706251e-01
4.97601122e-01 -1.28935426e-01 -8.86815131e-01 -8.14661205e-01
4.22743618e-01 5.08893490e-01 1.00412384e-01 5.98174810e-01
-8.08093488e-01 1.38620508e+00 5.87789392e+00 7.90158689e-01
-1.17927170e+00 -3.67364347e-01 3.23121101e-01 6.33505434e-02
-4.47021604e-01 -9.72395614e-02 -5.68488777e-01 2.54793972e-01
1.02575469e+00 -3.94684374e-01 6.93152010e-01 7.19218433e-01
8.28147590e-01 -2.60463264e-02 -1.42274928e+00 5.48530579e-01
-2.99329281e-01 -1.40731215e+00 1.20426074e-01 -2.56779373e-01
5.55648506e-01 -4.98706788e-01 -4.09094393e-02 3.23514909e-01
1.00275934e+00 -1.26781321e+00 7.48762608e-01 4.38629687e-01
1.18000138e+00 -7.33497143e-01 -1.96601808e-01 8.66045535e-01
-1.17806160e+00 1.69166937e-01 3.02785486e-01 -3.34613770e-01
3.83561730e-01 -3.56215060e-01 -8.14643443e-01 -2.33513713e-01
3.91347855e-02 1.74441501e-01 2.01287940e-01 3.68452966e-01
-6.46444082e-01 5.09092629e-01 -3.45636934e-01 -6.89954698e-01
3.06889623e-01 -1.29810452e-01 7.28725493e-01 1.04026496e+00
1.85111642e-01 3.71168494e-01 4.94831681e-01 1.21124661e+00
2.49508113e-01 -7.33742565e-02 -8.98472011e-01 -3.85300308e-01
4.40341830e-01 1.03902650e+00 -3.92528921e-01 -3.09574276e-01
2.30903551e-02 1.13041985e+00 1.93696275e-01 2.07111135e-01
-9.69971657e-01 -5.09445727e-01 1.14926803e+00 2.46225789e-01
-1.81859672e-01 -8.30873609e-01 -2.05674708e-01 -9.46961284e-01
-3.89911324e-01 -1.48028541e+00 -3.27200741e-01 -1.11079431e+00
-7.43200719e-01 3.69783282e-01 1.51687041e-01 -1.10033667e+00
-1.03033161e+00 -3.86178613e-01 -8.86198759e-01 1.02120459e+00
-7.25108683e-01 -1.05889475e+00 -9.58462134e-02 6.01002634e-01
8.99999440e-01 -3.19761664e-01 9.94846821e-01 -3.09025586e-01
-1.52445257e-01 2.50656456e-01 -6.81982815e-01 1.58583820e-01
4.73960698e-01 -1.16544044e+00 6.35535717e-01 7.27956176e-01
-2.19269484e-01 5.42617977e-01 1.08350265e+00 -7.17427969e-01
-1.96456039e+00 -9.14794505e-01 4.26911622e-01 -3.77013534e-01
8.14048409e-01 -3.11606854e-01 -5.20784140e-01 4.70150888e-01
6.81849048e-02 -4.79803413e-01 2.08275616e-01 -6.88568354e-01
1.38631225e-01 2.67885953e-01 -1.04156363e+00 1.22416377e+00
7.95190692e-01 -6.59978509e-01 -6.97070777e-01 2.59242892e-01
9.60423768e-01 -6.60269320e-01 -7.32709587e-01 8.40032026e-02
5.84376514e-01 -5.08917034e-01 9.61209655e-01 -8.87886047e-01
8.81525934e-01 -4.09902424e-01 5.53456694e-02 -1.40236044e+00
-9.56333578e-02 -1.19539666e+00 2.54789770e-01 7.89628744e-01
1.89481109e-01 -3.74010772e-01 7.72872448e-01 8.66980553e-01
-1.00776054e-01 -6.11373425e-01 -5.46511471e-01 -6.10791326e-01
5.30520499e-01 -6.65634215e-01 3.72834891e-01 4.64260131e-01
2.25783020e-01 4.19339657e-01 -5.00683486e-01 1.27901435e-01
2.52204657e-01 2.16670781e-01 1.22175884e+00 -6.81979954e-01
-9.11599219e-01 -6.04202569e-01 -1.77337021e-01 -1.38064253e+00
6.37459278e-01 -8.08347046e-01 5.16368866e-01 -1.27854097e+00
-1.36310101e-01 -3.71764272e-01 5.93808949e-01 5.63866496e-01
-1.01688996e-01 -2.14172959e-01 6.19956255e-01 1.78274274e-01
-1.54688388e-01 4.53395724e-01 1.69557154e+00 -3.24493200e-02
-3.99154991e-01 1.34694695e-01 -4.48892266e-01 7.05043495e-01
8.01485121e-01 -2.87595928e-01 -5.36525130e-01 -1.93493515e-01
1.60809204e-01 7.35257506e-01 4.70982194e-01 -1.10430539e+00
3.10095161e-01 -8.21904600e-01 -5.91644868e-02 1.03947654e-01
5.96214592e-01 -5.28226078e-01 7.58027136e-02 9.15440321e-01
-8.48230720e-01 1.83353081e-01 2.54821122e-01 1.30998656e-01
-3.13808257e-03 1.84722003e-02 8.54949296e-01 -2.93549418e-01
-6.92637801e-01 1.71329707e-01 -1.13565826e+00 1.67170897e-01
1.37189221e+00 9.10579115e-02 2.23836843e-02 -9.31966603e-01
-5.48904240e-01 4.29302961e-01 7.95973837e-01 4.78141546e-01
7.34869123e-01 -1.07573020e+00 -8.05340648e-01 4.79715019e-02
-7.61039853e-02 -1.91931590e-01 -3.59809726e-01 1.55731246e-01
-8.88767838e-01 9.21929181e-02 -4.43506032e-01 -4.40772831e-01
-1.20932949e+00 2.31168985e-01 2.74669379e-01 -1.90929905e-01
-6.35463834e-01 7.10823298e-01 1.48216814e-01 -5.91900826e-01
-3.77087593e-02 -4.84289646e-01 6.62248582e-02 -5.93750477e-01
4.35904145e-01 -1.09401502e-01 -7.98251748e-01 -3.98260772e-01
-1.88174754e-01 5.95897317e-01 4.87523496e-01 -7.00768232e-01
1.08601046e+00 4.03296590e-01 1.79153904e-01 5.02008557e-01
7.34235704e-01 -5.19775115e-02 -1.73263454e+00 5.47936320e-01
-3.98535013e-01 -2.03810140e-01 -2.98919737e-01 -6.39044642e-01
-5.29050469e-01 8.20695341e-01 1.44771012e-02 -1.81869656e-01
8.81477773e-01 -1.22287929e-01 9.72802699e-01 6.34885192e-01
7.68755913e-01 -9.52168047e-01 5.05989790e-01 5.90784073e-01
9.90594983e-01 -8.82441223e-01 -2.41830781e-01 -1.69661209e-01
-1.04642856e+00 1.16663611e+00 7.77190626e-01 -3.16643566e-01
4.59888913e-02 8.99943590e-01 1.15171693e-01 -1.13568455e-02
-9.91478384e-01 -1.82266220e-01 2.25039661e-01 8.28895807e-01
3.13907951e-01 9.69065279e-02 -5.54784313e-02 1.99883178e-01
-4.10752952e-01 1.02531619e-01 7.65239894e-01 1.03853238e+00
-5.53133667e-01 -1.45958769e+00 -3.73355031e-01 2.21235245e-01
-2.36245573e-01 5.56116998e-02 -3.59107256e-01 7.75269628e-01
-2.67109245e-01 8.72416556e-01 -1.85848087e-01 -4.29765910e-01
2.45987102e-01 3.35238352e-02 6.30819559e-01 -1.03582609e+00
-8.60505939e-01 -3.28786194e-01 2.50001937e-01 -6.80382609e-01
-8.63791034e-02 -6.42875552e-01 -1.66341305e+00 -2.50068218e-01
3.21951270e-01 7.01159704e-03 4.87505704e-01 8.03778529e-01
1.85295135e-01 5.38537920e-01 4.30023909e-01 -1.32201993e+00
-6.73892915e-01 -5.58155298e-01 4.12453711e-02 5.02210736e-01
1.62694931e-01 -2.87536085e-01 2.79133879e-02 4.32981551e-01] | [4.894867897033691, 0.7266707420349121] |
55b12ccd-9408-470f-a665-a31c1bad3643 | revisiting-low-resource-neural-machine | 1905.11901 | null | https://arxiv.org/abs/1905.11901v1 | https://arxiv.org/pdf/1905.11901v1.pdf | Revisiting Low-Resource Neural Machine Translation: A Case Study | It has been shown that the performance of neural machine translation (NMT) drops starkly in low-resource conditions, underperforming phrase-based statistical machine translation (PBSMT) and requiring large amounts of auxiliary data to achieve competitive results. In this paper, we re-assess the validity of these results, arguing that they are the result of lack of system adaptation to low-resource settings. We discuss some pitfalls to be aware of when training low-resource NMT systems, and recent techniques that have shown to be especially helpful in low-resource settings, resulting in a set of best practices for low-resource NMT. In our experiments on German--English with different amounts of IWSLT14 training data, we show that, without the use of any auxiliary monolingual or multilingual data, an optimized NMT system can outperform PBSMT with far less data than previously claimed. We also apply these techniques to a low-resource Korean-English dataset, surpassing previously reported results by 4 BLEU. | ['Biao Zhang', 'Rico Sennrich'] | 2019-05-28 | revisiting-low-resource-neural-machine-1 | https://aclanthology.org/P19-1021 | https://aclanthology.org/P19-1021.pdf | acl-2019-7 | ['low-resource-neural-machine-translation'] | ['natural-language-processing'] | [ 3.18679661e-01 -1.52792661e-02 -3.58271807e-01 -2.43383452e-01
-1.55700970e+00 -6.94153130e-01 7.80923724e-01 -2.11752206e-01
-1.05557668e+00 1.35215712e+00 3.80010724e-01 -1.08901668e+00
2.71538496e-01 -2.41746306e-01 -8.45294058e-01 -4.78354007e-01
2.28201300e-01 9.93760705e-01 -9.80725735e-02 -4.51063246e-01
1.65435687e-01 1.15650609e-01 -8.49943817e-01 2.58995205e-01
9.98126626e-01 5.96551113e-02 4.25572693e-01 5.90544760e-01
3.57633904e-02 2.64753193e-01 -6.80139542e-01 -5.57977676e-01
4.41420496e-01 -6.50470555e-01 -9.04609323e-01 -3.84533316e-01
5.74344993e-01 -9.76518914e-02 5.78773841e-02 7.74868846e-01
6.96365893e-01 -5.86337410e-02 5.61614454e-01 -5.66642582e-01
-7.62018740e-01 8.09196115e-01 -3.21441263e-01 4.59697098e-01
2.24250406e-01 1.76136509e-01 1.02639878e+00 -1.15888965e+00
8.55368137e-01 1.16088915e+00 7.78951228e-01 6.46565020e-01
-1.26403630e+00 -5.62829852e-01 -4.32610884e-02 -3.59069631e-02
-1.17550862e+00 -8.52525055e-01 4.18608524e-02 1.09592184e-01
1.60656941e+00 2.76760697e-01 3.96764219e-01 1.28004336e+00
3.97162765e-01 6.05097473e-01 1.53455412e+00 -8.80966604e-01
-2.20563024e-01 2.96442837e-01 -4.00230348e-01 4.00764763e-01
3.28701228e-01 -3.38187888e-02 -5.84760427e-01 -5.17171696e-02
5.54097235e-01 -6.93918467e-01 -2.68758059e-01 1.62426233e-01
-1.78848648e+00 6.62660003e-01 8.18492249e-02 7.23699152e-01
-2.51207501e-01 -9.47486460e-02 3.58056754e-01 8.10055971e-01
7.70906210e-01 8.43895733e-01 -9.68652964e-01 -4.71974581e-01
-1.05375588e+00 -1.40614137e-01 9.14135218e-01 1.02941227e+00
5.96614540e-01 2.91181624e-01 1.72565803e-01 1.07302880e+00
-2.30100408e-01 7.94072628e-01 8.04148674e-01 -7.04424858e-01
1.06384468e+00 -5.05100153e-02 7.64439702e-02 -3.88248503e-01
-1.54270932e-01 -3.89689565e-01 -5.28414905e-01 -2.85196573e-01
6.04473948e-01 -3.87420177e-01 -9.47700143e-01 1.87298441e+00
-1.89797327e-01 -2.75703132e-01 3.75788927e-01 7.00571477e-01
2.70756274e-01 7.51535654e-01 -7.90562257e-02 -5.70745647e-01
1.01600075e+00 -1.01835835e+00 -5.85092366e-01 -4.59393412e-01
1.05622077e+00 -1.25029755e+00 1.43631256e+00 3.21262807e-01
-1.24640846e+00 -2.77590513e-01 -9.15804207e-01 -5.03543653e-02
-3.59243870e-01 6.85899407e-02 6.65568948e-01 6.37621582e-01
-1.35684812e+00 7.87936389e-01 -8.68532896e-01 -9.38230097e-01
-1.34641171e-01 6.14599228e-01 -6.12443268e-01 -2.49250963e-01
-1.24730921e+00 1.56311059e+00 3.88621122e-01 2.38016382e-01
-5.34973860e-01 -3.63622725e-01 -5.90523839e-01 -3.49091440e-01
3.62095743e-01 -6.12339318e-01 1.45443940e+00 -1.13352895e+00
-1.64089918e+00 9.19333994e-01 -2.92812854e-01 -3.44486654e-01
5.03166795e-01 -2.90312380e-01 -5.39812386e-01 -1.24859966e-01
2.22126916e-01 4.72767979e-01 4.27214801e-01 -8.45465064e-01
-6.20953500e-01 -1.26400128e-01 -2.17499003e-01 5.20247221e-01
-3.12255740e-01 5.05370498e-01 -4.74331290e-01 -5.84302545e-01
-8.97877105e-03 -1.25449276e+00 -2.67611235e-01 -9.18715358e-01
-2.38856077e-01 2.12991424e-02 2.73852468e-01 -9.32409048e-01
9.87494946e-01 -1.73537493e+00 2.90300459e-01 -1.76481918e-01
-4.19932812e-01 3.06483954e-01 -5.25448442e-01 6.91733599e-01
8.87670070e-02 4.83624756e-01 -2.67105967e-01 -3.27316791e-01
-1.20169401e-01 5.24360597e-01 -1.06000006e-01 2.59798229e-01
2.71476746e-01 1.16532457e+00 -8.15545499e-01 -6.17779970e-01
-1.53203815e-01 7.22997338e-02 -4.04424757e-01 -1.98132858e-01
-1.22319572e-02 5.02068102e-01 -1.73251346e-01 7.19709635e-01
2.68205166e-01 7.52447546e-02 4.86264199e-01 2.57264942e-01
-4.02006775e-01 9.01891172e-01 -5.13639152e-01 1.92751765e+00
-6.80174351e-01 7.17269599e-01 8.22897106e-02 -7.72865832e-01
4.87933159e-01 4.47861671e-01 1.34832203e-01 -8.72292578e-01
-9.58966166e-02 1.00228214e+00 5.93175650e-01 -3.12598795e-01
6.14653647e-01 -4.21876103e-01 -1.78600624e-01 7.58604765e-01
2.13665977e-01 -3.65185499e-01 2.57779360e-01 3.17003168e-02
9.52391028e-01 3.55517596e-01 6.68984577e-02 -6.55135334e-01
1.41198277e-01 3.45838368e-01 5.46287119e-01 7.37603128e-01
-4.82027233e-02 4.95603621e-01 5.56957759e-02 -1.23223230e-01
-1.41951585e+00 -8.88369203e-01 -1.01098999e-01 1.19224334e+00
-4.22923326e-01 -3.41417372e-01 -7.74336755e-01 -5.56801200e-01
-4.10136014e-01 8.88331234e-01 -2.79477566e-01 2.36852661e-01
-1.24685824e+00 -1.31552958e+00 7.59458244e-01 3.12022895e-01
1.45337313e-01 -1.07053626e+00 -1.06853411e-01 3.53225708e-01
-4.22228128e-01 -1.38187861e+00 -5.77027977e-01 5.26278853e-01
-1.13752043e+00 -5.08626938e-01 -9.49204743e-01 -8.92066121e-01
4.15196538e-01 2.17613682e-01 1.37195253e+00 -4.94592031e-03
2.83100814e-01 -6.80024223e-03 -2.67522901e-01 -2.84012377e-01
-8.90101969e-01 8.25792253e-01 4.39393431e-01 -7.63124287e-01
5.08276999e-01 -5.62022269e-01 -1.53406650e-01 3.39794517e-01
-5.47144949e-01 1.45749211e-01 1.22604632e+00 1.03907275e+00
3.27600807e-01 -5.93800962e-01 7.32931197e-01 -1.01665366e+00
7.62944877e-01 -3.15977395e-01 -1.63212493e-01 3.57780099e-01
-8.98842692e-01 7.39204288e-02 6.81439698e-01 -5.11295140e-01
-8.80498767e-01 -4.97267753e-01 -1.92960724e-01 8.45729485e-02
8.52173716e-02 5.93131065e-01 -9.77386162e-02 -5.03897704e-02
7.31531739e-01 3.52971375e-01 -1.21997707e-01 -5.70171893e-01
2.67505199e-01 7.88552225e-01 3.10125977e-01 -8.53276312e-01
8.66001129e-01 -1.14367343e-01 -1.85425639e-01 -6.74893022e-01
-7.40882814e-01 -1.14120834e-01 -8.48876536e-01 3.43245775e-01
5.23656905e-01 -9.85146344e-01 1.62058920e-01 1.47783414e-01
-9.19605374e-01 -6.94469273e-01 -6.35197088e-02 8.76454651e-01
-6.16601944e-01 2.60885835e-01 -1.07636714e+00 -2.67739415e-01
-4.80371833e-01 -1.25234675e+00 7.70031333e-01 -3.07338893e-01
-3.40372324e-01 -1.06382668e+00 1.36726409e-01 4.36316103e-01
5.59832633e-01 -2.66047060e-01 1.01645637e+00 -8.80883455e-01
-3.08115363e-01 -2.93006301e-02 -3.89017947e-02 4.73878890e-01
2.03546092e-01 -2.66225010e-01 -6.79992139e-01 -5.43171585e-01
-1.47571927e-02 -4.77267772e-01 6.37535274e-01 4.47959043e-02
3.06813300e-01 -2.99795121e-01 -1.07025929e-01 5.25536954e-01
1.44615984e+00 -5.83897978e-02 3.12794685e-01 6.43147290e-01
5.62232077e-01 4.57739651e-01 5.17411232e-01 -3.84109348e-01
3.57308030e-01 7.08471298e-01 -2.45123953e-01 -2.18440041e-01
-1.58744484e-01 -3.02566141e-01 7.53209174e-01 1.65827000e+00
-3.72038931e-01 -2.75719613e-01 -9.67129707e-01 6.69328809e-01
-1.71171975e+00 -6.93910837e-01 -1.54847950e-01 2.21904325e+00
1.20211267e+00 3.04341227e-01 1.89498097e-01 -3.85311097e-01
4.74396139e-01 -4.59348150e-02 -2.35899895e-01 -9.81039524e-01
-3.49187702e-01 4.25996840e-01 8.26436996e-01 5.99820793e-01
-6.27395034e-01 1.54040408e+00 7.73502016e+00 7.85510898e-01
-1.00448632e+00 5.91127336e-01 3.23197663e-01 -2.43236065e-01
-3.25404108e-01 1.54971540e-01 -6.73787832e-01 2.21157625e-01
1.63742542e+00 -1.54175207e-01 6.70859873e-01 3.79418761e-01
1.70421377e-01 -5.88666275e-02 -1.18065214e+00 4.51975524e-01
1.54683620e-01 -8.98308575e-01 -1.78717263e-02 1.46995336e-01
8.58074844e-01 6.65162086e-01 6.78153187e-02 5.91797590e-01
4.79082823e-01 -1.04165387e+00 5.23370743e-01 -2.16119662e-01
1.03188455e+00 -6.73032939e-01 6.84217632e-01 5.08408844e-01
-5.94166279e-01 3.47858936e-01 -6.34588659e-01 -1.99119315e-01
5.55286780e-02 1.75167426e-01 -1.13495207e+00 7.23786294e-01
3.76242965e-01 3.84115398e-01 -4.67513919e-01 5.50085545e-01
-3.24571222e-01 9.10512507e-01 -4.65224713e-01 -8.33438337e-02
6.28734112e-01 -2.15103477e-01 5.24979830e-01 1.54586899e+00
5.19399226e-01 -2.29952380e-01 -2.40463931e-02 3.27401459e-01
-1.47105142e-01 6.14840686e-01 -7.15180635e-01 -2.17539921e-01
2.36973971e-01 9.65184689e-01 -4.40075964e-01 -2.93680698e-01
-5.64814270e-01 1.31057143e+00 5.00235379e-01 3.60321015e-01
-4.71893966e-01 -9.99478716e-03 4.17920053e-01 -1.18767478e-01
6.71950227e-04 -6.32291853e-01 -2.15750113e-01 -1.54441667e+00
1.08043581e-01 -1.32530785e+00 1.47630647e-01 -6.28797591e-01
-9.72638309e-01 9.60160911e-01 4.70012240e-02 -1.04804611e+00
-7.32351124e-01 -7.22579062e-01 -2.81537831e-01 1.38686502e+00
-1.48599112e+00 -1.25526631e+00 6.53549552e-01 2.41783321e-01
8.08566868e-01 -3.33772212e-01 1.17064023e+00 2.88639188e-01
-5.55906534e-01 7.99015164e-01 4.78649169e-01 -3.95174921e-02
1.21384954e+00 -1.25894153e+00 9.17034805e-01 1.17815852e+00
5.24035335e-01 8.84099841e-01 8.31173539e-01 -5.88751078e-01
-1.44522178e+00 -6.75206542e-01 1.40893900e+00 -9.16074395e-01
8.43130171e-01 -5.32494426e-01 -7.30893493e-01 9.88655925e-01
6.60089970e-01 -6.00855708e-01 7.09249556e-01 5.80578685e-01
-2.57820457e-01 2.12385520e-01 -8.59904945e-01 9.00766790e-01
1.14906657e+00 -5.68969011e-01 -8.76013935e-01 5.67350388e-01
7.52134085e-01 -3.17488223e-01 -8.66096735e-01 3.94306093e-01
5.87982118e-01 -5.06006598e-01 6.03769064e-01 -8.89399648e-01
3.51598263e-01 -4.29037085e-04 -2.70070046e-01 -1.76412821e+00
-2.53828228e-01 -7.86082745e-01 4.84454423e-01 9.12110031e-01
1.09232688e+00 -7.94663072e-01 5.32623351e-01 3.15615207e-01
-3.72148871e-01 -6.57838225e-01 -1.07016754e+00 -9.93719101e-01
8.59824419e-01 -2.82484144e-01 2.59646595e-01 1.14023209e+00
9.31158513e-02 6.97937071e-01 -5.62410116e-01 -2.06078321e-01
2.14072287e-01 -3.06452867e-02 7.57728398e-01 -8.04193795e-01
-4.45916653e-01 -3.58682662e-01 2.49085858e-01 -8.95761013e-01
2.68472612e-01 -1.02347946e+00 2.19500780e-01 -1.45252872e+00
2.96672523e-01 -4.64974612e-01 -3.06773096e-01 5.96476674e-01
-2.98305541e-01 6.04756117e-01 2.73360223e-01 6.71145797e-01
-1.97319523e-01 1.77224100e-01 1.33474803e+00 6.53544217e-02
-7.13626742e-02 -2.18663007e-01 -8.60957682e-01 4.18296456e-01
8.45621228e-01 -6.25825584e-01 -2.06072897e-01 -9.78831589e-01
2.19695821e-01 -1.11577868e-01 -3.69758934e-01 -6.71759486e-01
-4.02521379e-02 -2.77442545e-01 2.07341388e-01 -1.95838779e-01
2.94351786e-01 -5.26434362e-01 -1.78810414e-02 4.65089351e-01
-2.85403371e-01 6.90464139e-01 4.68988240e-01 -6.65937876e-03
-4.54335511e-02 -3.50640774e-01 5.53099334e-01 -4.57067460e-01
-5.14353275e-01 -2.49975156e-02 -5.45582533e-01 2.17526570e-01
4.18712378e-01 -1.54269546e-01 -2.23440066e-01 -2.39540607e-01
-4.81409639e-01 -2.86246240e-02 8.41440976e-01 4.31184202e-01
2.37996501e-05 -1.12619877e+00 -1.20336771e+00 9.06960964e-02
2.91857077e-03 -4.19356912e-01 -2.72470623e-01 1.17972410e+00
-4.79486048e-01 8.08274865e-01 -2.22644791e-01 -3.32032412e-01
-1.03814030e+00 3.93491387e-01 1.90132946e-01 -3.72734606e-01
-5.45765042e-01 5.39543331e-01 -2.68862784e-01 -6.91731453e-01
-2.29244530e-01 -2.24727735e-01 3.64966005e-01 -1.95631474e-01
7.69593790e-02 -8.99443030e-02 3.77952933e-01 -6.58338308e-01
-2.67877102e-01 3.61712843e-01 -3.93563002e-01 -7.29774296e-01
1.17763317e+00 -6.95573837e-02 -1.41756326e-01 6.42256320e-01
9.88085032e-01 5.01117706e-01 -5.21283865e-01 -3.30433398e-01
1.81139544e-01 -1.96332291e-01 -1.73250228e-01 -1.01426983e+00
-3.96680415e-01 7.87362158e-01 1.70794874e-01 -2.53054947e-01
9.90547717e-01 -3.34811211e-01 9.64307129e-01 8.05090725e-01
7.28119791e-01 -1.29902375e+00 -4.60877955e-01 7.46604800e-01
6.72045350e-01 -1.20908499e+00 -8.95666610e-03 -1.56300552e-02
-5.56264758e-01 8.56528878e-01 3.89367342e-01 3.24684009e-02
8.72108936e-02 2.80790091e-01 4.06982154e-01 3.14519346e-01
-1.01967371e+00 -1.29143253e-01 6.90284967e-02 3.41190517e-01
8.02967370e-01 9.54733118e-02 -6.97793484e-01 9.06052291e-02
-5.37675619e-01 -2.96184480e-01 5.04384041e-01 8.91705096e-01
-2.23471209e-01 -1.61017644e+00 -3.39089513e-01 3.43757838e-01
-9.63946402e-01 -6.97488070e-01 -5.97337484e-01 1.27060783e+00
-1.83283761e-01 7.95816541e-01 -1.07335791e-01 -2.96902090e-01
1.71030954e-01 4.97451186e-01 7.26582229e-01 -9.07679796e-01
-7.53006756e-01 2.70596743e-01 6.09187305e-01 -2.71734327e-01
-3.80957782e-01 -8.00317049e-01 -6.42711818e-01 -3.62532943e-01
-3.91795188e-01 4.30704206e-01 7.75808096e-01 1.07170796e+00
2.12974519e-01 1.01240158e-01 2.88426816e-01 -7.11061776e-01
-7.48149216e-01 -1.49565077e+00 -1.21968940e-01 7.10563222e-03
5.31882159e-02 -7.97583461e-02 -3.24599892e-01 -1.89687088e-01] | [11.562766075134277, 10.348076820373535] |
1ec1b417-9066-42d6-ba55-8b28270377cf | semeval-2020-task-5-counterfactual | 2008.00563 | null | https://arxiv.org/abs/2008.00563v1 | https://arxiv.org/pdf/2008.00563v1.pdf | SemEval-2020 Task 5: Counterfactual Recognition | We present a counterfactual recognition (CR) task, the shared Task 5 of SemEval-2020. Counterfactuals describe potential outcomes (consequents) produced by actions or circumstances that did not happen or cannot happen and are counter to the facts (antecedent). Counterfactual thinking is an important characteristic of the human cognitive system; it connects antecedents and consequents with causal relations. Our task provides a benchmark for counterfactual recognition in natural language with two subtasks. Subtask-1 aims to determine whether a given sentence is a counterfactual statement or not. Subtask-2 requires the participating systems to extract the antecedent and consequent in a given counterfactual statement. During the SemEval-2020 official evaluation period, we received 27 submissions to Subtask-1 and 11 to Subtask-2. The data, baseline code, and leaderboard can be found at https://competitions.codalab.org/competitions/21691. The data and baseline code are also available at https://zenodo.org/record/3932442. | ['Stan Matwin', 'Xiaodan Zhu', 'Xiaoyu Yang', 'Stephen Obadinma', 'Qiong Zhang', 'Huasha Zhao'] | 2020-08-02 | null | https://aclanthology.org/2020.semeval-1.40 | https://aclanthology.org/2020.semeval-1.40.pdf | semeval-2020 | ['counterfactual-inference'] | ['miscellaneous'] | [ 3.81707191e-01 5.50683439e-01 -2.70305336e-01 -5.91207981e-01
-7.67610073e-01 -5.64778030e-01 1.41371906e+00 2.23477021e-01
-2.46990457e-01 1.42701530e+00 9.25712824e-01 -6.14061236e-01
-2.86843106e-02 -5.92092514e-01 -9.28921103e-01 -8.57754722e-02
-4.29846764e-01 3.44216377e-01 -1.85283110e-01 -1.60533324e-01
4.26874131e-01 -3.54142226e-02 -1.55142868e+00 1.01253140e+00
4.93581653e-01 9.76075590e-01 1.05808610e-02 6.87901199e-01
1.20918870e-01 1.37836754e+00 -8.11966956e-01 -6.94079101e-01
2.19554082e-01 -4.21770900e-01 -1.34094799e+00 -5.25419474e-01
3.68805945e-01 -9.57814530e-02 -1.31011773e-02 1.03600788e+00
3.78179580e-01 7.40797371e-02 6.58680916e-01 -1.59985626e+00
-6.21998489e-01 9.47951555e-01 -1.19058512e-01 4.03927892e-01
8.14646959e-01 3.85420889e-01 1.07574105e+00 -8.46480846e-01
6.82969093e-01 1.50970626e+00 2.30703950e-01 7.04582632e-01
-1.29954791e+00 -8.53942156e-01 3.49070787e-01 6.29911482e-01
-9.66767669e-01 -8.69138300e-01 7.40004361e-01 -4.17036235e-01
1.39472091e+00 7.83595622e-01 6.12781107e-01 1.51134050e+00
6.66841507e-01 9.84415472e-01 1.44871807e+00 -4.20640677e-01
5.11953056e-01 -1.54529780e-01 2.04671577e-01 9.12856013e-02
3.45971316e-01 7.97176659e-01 -6.05157852e-01 -1.79278612e-01
7.55230784e-02 -5.66842020e-01 -4.57395732e-01 -1.87543593e-02
-1.58469212e+00 5.81818879e-01 2.94204772e-01 2.41547421e-01
-7.62113214e-01 3.34134489e-01 6.04100347e-01 5.44921398e-01
3.47235590e-01 8.51105809e-01 -5.18439293e-01 -1.55941874e-01
-7.80263245e-01 1.06448638e+00 8.45535696e-01 6.10738099e-01
3.04316193e-01 -4.49351579e-01 -5.50587296e-01 3.58494669e-01
6.98811114e-02 4.35125917e-01 7.83031106e-01 -1.10323715e+00
6.70841813e-01 5.26646197e-01 5.25219202e-01 -7.57628083e-01
-4.91051197e-01 -1.64262652e-01 -5.96943557e-01 2.90006906e-01
2.25315407e-01 -3.38351369e-01 -5.82069278e-01 1.79363489e+00
1.15427032e-01 6.96465448e-02 2.68904656e-01 8.74133468e-01
8.66787493e-01 3.72986108e-01 3.35473865e-01 -6.91339970e-01
1.21162474e+00 -3.64132702e-01 -9.20513093e-01 -3.42037141e-01
7.85436451e-01 -7.93193460e-01 1.08140528e+00 3.56328785e-01
-1.01500154e+00 -2.74940819e-01 -1.17888451e+00 2.45914608e-01
-3.00740391e-01 -3.16738218e-01 7.60911703e-01 3.55210066e-01
-7.85534084e-01 3.45873028e-01 -2.45523989e-01 1.35748699e-01
3.22468758e-01 -6.37834370e-02 -1.79839194e-01 2.96189219e-01
-1.89952159e+00 1.11456883e+00 8.46531272e-01 -1.03231952e-01
-1.15351355e+00 -8.50775182e-01 -8.00916314e-01 -9.75129455e-02
8.28526855e-01 -6.43527091e-01 1.92677760e+00 -1.07534814e+00
-1.10486138e+00 1.08361650e+00 -3.45446467e-02 -1.12689364e+00
9.17253792e-01 -3.03926677e-01 -8.94055605e-01 -3.94386590e-01
5.25638580e-01 5.01676083e-01 4.17603135e-01 -9.80777740e-01
-8.13421726e-01 -2.35323638e-01 3.60692203e-01 3.67265701e-01
6.17815971e-01 2.00568229e-01 4.88863498e-01 -6.19813740e-01
-3.70670646e-01 -6.54424787e-01 -3.40285823e-02 -8.10768545e-01
-9.64602172e-01 -5.15908182e-01 4.54963475e-01 -1.87040016e-01
1.21977091e+00 -1.63104391e+00 -6.11877978e-01 -3.21777403e-01
-5.90121886e-03 6.77930489e-02 3.46676521e-02 2.91211367e-01
-9.64121521e-01 3.48576725e-01 -1.89088166e-01 1.82683453e-01
2.87287205e-01 -3.78242493e-01 -1.00074136e+00 1.40402153e-01
1.25395864e-01 1.02093959e+00 -1.10563982e+00 -4.31323230e-01
4.67359632e-01 -4.61361825e-01 -1.86766610e-01 1.56567216e-01
-6.90348446e-01 -5.67129552e-02 -3.48672718e-01 9.11968201e-02
6.37658715e-01 1.09564647e-01 3.96643639e-01 -2.97804102e-02
-3.65797192e-01 1.39946318e+00 -1.20616257e+00 1.24106836e+00
-3.56342524e-01 5.79351544e-01 -5.37339687e-01 -9.51025367e-01
4.35047179e-01 8.18515122e-01 9.89904627e-03 -1.00706458e+00
1.62263680e-02 1.75308868e-01 1.05033666e-02 -4.26537842e-01
6.50237918e-01 -8.38723958e-01 -5.10390162e-01 6.25943244e-01
-1.92822754e-01 -3.78579021e-01 6.98529005e-01 2.59181887e-01
1.18499029e+00 3.20066869e-01 1.16185784e+00 -5.12708068e-01
5.80397964e-01 2.36123338e-01 7.42981076e-01 8.86958122e-01
-3.34010661e-01 2.02261761e-01 7.29656935e-01 -8.55363846e-01
-4.99954283e-01 -1.35318136e+00 -2.32221615e-02 5.32732069e-01
-2.30046526e-01 -3.07514131e-01 -4.00363177e-01 -9.75845218e-01
-1.39202073e-01 1.82604778e+00 -7.82814741e-01 -5.76315150e-02
-4.47441310e-01 -5.45243919e-01 5.38656831e-01 2.38795832e-01
7.50396311e-01 -1.59023857e+00 -1.03507292e+00 1.97034627e-01
-8.64531398e-01 -6.13795578e-01 -3.53467405e-01 -1.51783362e-01
-5.22372663e-01 -1.68162477e+00 -1.51668653e-01 1.23830870e-01
-2.05256622e-02 3.45917605e-03 1.51409578e+00 -2.07219109e-01
7.46829212e-02 -6.72274902e-02 -1.79182701e-02 -1.05797291e+00
-5.89561820e-01 -6.74279273e-01 3.34472328e-01 -2.87938148e-01
6.05676234e-01 -4.21746224e-01 -4.39496279e-01 4.20404822e-02
-5.50628185e-01 5.19596696e-01 3.50480169e-01 5.67852080e-01
4.27833378e-01 -1.48185745e-01 8.14486265e-01 -1.02061749e+00
9.81732190e-01 -4.57525879e-01 -6.02930486e-01 3.27263057e-01
-6.26135886e-01 -5.79033233e-02 7.88308918e-01 1.86317042e-01
-1.70047736e+00 -3.23249847e-01 1.48042515e-01 2.39102587e-01
-6.89175069e-01 6.38053834e-01 -4.20479387e-01 1.31684172e+00
1.06097353e+00 6.40264675e-02 -5.60169816e-01 -1.26174316e-01
5.33751726e-01 5.49142897e-01 9.08947766e-01 -7.58644938e-01
2.44308233e-01 3.33922774e-01 -2.89683610e-01 -1.91798031e-01
-1.27748382e+00 -1.39201477e-01 -1.59297571e-01 -4.12239909e-01
5.88538766e-01 -8.77572834e-01 -9.61905658e-01 7.01081827e-02
-1.45175505e+00 -4.67828482e-01 -6.07796550e-01 4.97246414e-01
-9.73955452e-01 -1.29805163e-01 2.13845409e-02 -1.03471661e+00
-3.75599176e-01 -4.68807042e-01 4.80003417e-01 -9.03239623e-02
-9.17408109e-01 -6.77415609e-01 3.15101743e-01 3.27907085e-01
-5.88030182e-02 5.32585859e-01 6.30422771e-01 -8.07791770e-01
-1.33949518e-01 -9.34611261e-02 1.79979764e-02 1.08626477e-01
1.56486034e-01 -2.69957453e-01 -9.72498596e-01 5.73941842e-02
2.19403714e-01 -5.16860783e-01 9.40047681e-01 4.05079633e-01
9.22304213e-01 -9.82994556e-01 -3.40393662e-01 -2.25948706e-01
1.02698493e+00 4.16891515e-01 6.60792112e-01 4.70835298e-01
-1.66477621e-01 8.44376087e-01 9.20776010e-01 2.61718392e-01
5.10706604e-01 6.63465321e-01 4.51215208e-01 4.77366745e-01
-2.95207351e-01 -4.24185574e-01 7.44870961e-01 -1.66733146e-01
-9.44114625e-02 -2.33247072e-01 -1.13329232e+00 9.73628640e-01
-1.94387519e+00 -1.75783312e+00 -4.66016948e-01 2.23102593e+00
1.04142773e+00 5.28460741e-01 8.85817260e-02 1.59575194e-01
6.89694941e-01 3.21638823e-01 -3.86924535e-01 -7.14714408e-01
-1.67803958e-01 -1.11479647e-01 -4.16393206e-03 5.81522465e-01
-1.17313731e+00 7.34929442e-01 5.82692432e+00 7.01879025e-01
-7.23531663e-01 1.04095094e-01 7.55953193e-01 -6.30135119e-01
-4.97213840e-01 1.75924361e-01 -3.31723243e-01 5.51556289e-01
1.27977920e+00 -9.76278007e-01 5.80623001e-02 5.48392594e-01
6.90068901e-01 -3.47627312e-01 -1.56567609e+00 3.48563403e-01
-3.31749856e-01 -1.77357495e+00 3.11167181e-01 -1.92793310e-01
7.75886059e-01 9.38833952e-02 -1.34624243e-01 5.69501102e-01
8.49580824e-01 -1.04225099e+00 1.50066435e+00 5.96153021e-01
7.61716247e-01 -6.13743424e-01 8.60326052e-01 4.06371355e-01
-7.01161265e-01 -6.14698045e-02 3.11630731e-03 -8.90727639e-01
2.81767637e-01 9.17976081e-01 -1.00688124e+00 8.14229071e-01
5.35720110e-01 5.27675867e-01 -2.34429583e-01 6.08524203e-01
-1.03869355e+00 7.93268263e-01 1.98232174e-01 -2.16063350e-01
-1.56420395e-02 4.19996232e-01 1.02785516e+00 1.38620973e+00
-1.42459661e-01 1.59749731e-01 -3.33372563e-01 1.21835220e+00
-2.79237300e-01 -1.21821627e-01 -7.34839737e-01 5.73440865e-02
6.12871587e-01 6.54163122e-01 -1.12514727e-01 -7.96547771e-01
-1.22099882e-02 5.17503500e-01 2.16919586e-01 3.52471620e-01
-7.07627952e-01 -1.24482913e-02 7.32646465e-01 2.31710058e-02
-2.93976545e-01 4.98818606e-01 -4.79098022e-01 -1.22586226e+00
1.98191479e-01 -9.96791303e-01 6.91426158e-01 -1.01345932e+00
-1.15844965e+00 3.37240279e-01 3.59761000e-01 -1.06284189e+00
-7.52909064e-01 -5.18943965e-01 -9.75801170e-01 8.09930146e-01
-1.13441026e+00 -7.58271873e-01 2.57825315e-01 3.88940126e-01
4.91264582e-01 -9.21764150e-02 1.08474624e+00 -3.94793093e-01
-2.20200107e-01 -2.50736205e-03 -6.66711569e-01 -5.00414632e-02
6.93708837e-01 -1.45327663e+00 6.38726532e-01 1.19653642e+00
8.53239298e-02 7.31951714e-01 1.31828356e+00 -8.75374198e-01
-5.59653759e-01 -1.07941246e+00 1.61677897e+00 -6.36980712e-01
6.62737250e-01 -1.65340602e-01 -4.08469230e-01 9.36363995e-01
3.14833909e-01 -4.53962058e-01 5.93021691e-01 4.74708229e-01
-4.08391714e-01 1.50411949e-01 -9.58312392e-01 9.08859789e-01
1.04211318e+00 -3.87296021e-01 -1.77632046e+00 3.40790123e-01
7.52104282e-01 -4.13646609e-01 -1.42595693e-01 2.72342712e-01
5.71884990e-01 -1.11125576e+00 6.73222721e-01 -1.02810478e+00
1.13199997e+00 -4.73930538e-01 -2.57076651e-01 -1.40823114e+00
-2.89059263e-02 -7.52426207e-01 -2.68163323e-01 8.93988132e-01
8.12211037e-01 -6.48307145e-01 2.23761484e-01 8.64500225e-01
-9.00114328e-02 -3.20808500e-01 -1.41159964e+00 -9.39603627e-01
1.83670819e-01 -1.16095877e+00 1.00863409e+00 1.14459038e+00
5.48992097e-01 5.53364158e-01 -4.43168767e-02 -2.83374131e-01
5.99320948e-01 5.27411878e-01 6.28233612e-01 -7.41068780e-01
-3.97017188e-02 -6.75252140e-01 5.18476069e-02 -4.68113750e-01
2.45630801e-01 -1.12606788e+00 -6.40809536e-02 -1.57864416e+00
3.97892028e-01 9.72729996e-02 -4.98347551e-01 7.87947893e-01
-2.25594163e-01 -2.10794523e-01 3.56803358e-01 5.96799925e-02
-7.72457540e-01 5.71804345e-01 8.26979399e-01 -2.02272683e-01
-1.81472711e-02 -1.45569409e-03 -1.03202975e+00 1.01962686e+00
9.18540657e-01 -4.58783299e-01 -2.25231811e-01 4.42584157e-02
3.90962660e-01 1.24357648e-01 8.31311822e-01 -6.58540606e-01
-1.48300365e-01 -6.09705448e-01 1.24423251e-01 -5.05819738e-01
3.63056101e-02 -2.91770041e-01 2.32825279e-01 6.44055963e-01
-8.06924820e-01 1.71194691e-02 3.77626598e-01 4.89426255e-01
-2.91618913e-01 -5.55014275e-02 1.78356320e-01 -3.19679290e-01
-8.84719670e-01 -4.61078525e-01 -5.61824977e-01 3.67375374e-01
1.04216218e+00 1.63359240e-01 -6.24254048e-01 -2.84180939e-01
-6.45111024e-01 5.01637578e-01 -7.73547450e-03 6.94692910e-01
7.57729292e-01 -1.41570795e+00 -1.22965348e+00 -4.21720803e-01
2.11030364e-01 -3.61703664e-01 3.84285986e-01 7.93875158e-01
-1.33282701e-02 1.04701054e+00 -2.38427103e-01 1.77921176e-01
-1.15314901e+00 5.70231736e-01 5.25761962e-01 -7.23261654e-01
-2.87840188e-01 8.70974302e-01 4.85330433e-01 -4.98350590e-01
-3.50417554e-01 -1.86878547e-01 -2.41365999e-01 -9.33075100e-02
1.08654892e+00 4.56568986e-01 1.54394642e-01 -3.83160323e-01
-8.40593934e-01 -7.16193974e-01 -1.84645489e-01 -4.97224331e-01
1.24967527e+00 1.44125193e-01 -2.08491638e-01 6.39775515e-01
6.37876511e-01 -2.72330314e-01 -7.66745865e-01 -4.01439611e-03
4.13630366e-01 -3.81437272e-01 -6.79966733e-02 -1.68477869e+00
-1.36640236e-01 4.32765871e-01 9.30638239e-02 3.00925523e-01
1.03939307e+00 -2.87331659e-02 2.04059005e-01 2.48185366e-01
5.19846737e-01 -1.18967366e+00 -4.37322885e-01 5.02569795e-01
1.94445086e+00 -1.07821643e+00 3.65036651e-02 -9.23535377e-02
-7.04642296e-01 6.64394736e-01 5.87406814e-01 3.31830457e-02
1.97207525e-01 1.38391867e-01 1.05886124e-01 -3.52260977e-01
-1.62475431e+00 1.16016947e-01 4.03591573e-01 3.18653822e-01
1.11243904e+00 1.00919163e+00 -9.23335135e-01 9.47266936e-01
-8.74446273e-01 3.58578056e-01 8.71256232e-01 5.29491365e-01
1.24705955e-01 -6.84089959e-01 -4.05742079e-01 6.31850421e-01
-5.19769490e-01 -3.69663507e-01 -8.01410377e-01 7.59243906e-01
2.41584212e-01 1.36788213e+00 -6.45989850e-02 -1.87257439e-01
5.99026024e-01 2.50514865e-01 4.37166750e-01 -7.10326910e-01
-5.06983995e-01 -5.61640024e-01 9.65054452e-01 -1.04958534e+00
-5.12955189e-01 -1.07209826e+00 -1.34854805e+00 -3.93218935e-01
1.12950906e-01 3.17451298e-01 2.96025008e-01 1.15739965e+00
1.55792937e-01 6.22107804e-01 1.55863166e-01 -2.74222404e-01
-7.79798388e-01 -1.11692822e+00 -2.15407848e-01 4.47605342e-01
4.09680218e-01 -5.83322525e-01 -2.34687179e-01 2.77324110e-01] | [9.995975494384766, 8.068673133850098] |
33a7b431-54f0-4462-be14-041fc4fb7a7a | waveform-boundary-detection-for-partially | 2211.00226 | null | https://arxiv.org/abs/2211.00226v1 | https://arxiv.org/pdf/2211.00226v1.pdf | Waveform Boundary Detection for Partially Spoofed Audio | The present paper proposes a waveform boundary detection system for audio spoofing attacks containing partially manipulated segments. Partially spoofed/fake audio, where part of the utterance is replaced, either with synthetic or natural audio clips, has recently been reported as one scenario of audio deepfakes. As deepfakes can be a threat to social security, the detection of such spoofing audio is essential. Accordingly, we propose to address the problem with a deep learning-based frame-level detection system that can detect partially spoofed audio and locate the manipulated pieces. Our proposed method is trained and evaluated on data provided by the ADD2022 Challenge. We evaluate our detection model concerning various acoustic features and network configurations. As a result, our detection system achieves an equal error rate (EER) of 6.58% on the ADD2022 challenge test set, which is the best performance in partially spoofed audio detection systems that can locate manipulated clips. | ['Ming Li', 'Weiqing Wang', 'Zexin Cai'] | 2022-11-01 | null | null | null | null | ['boundary-detection'] | ['computer-vision'] | [ 3.57970029e-01 -9.70329270e-02 -7.51335472e-02 2.67012775e-01
-1.18449056e+00 -6.72720909e-01 3.08727413e-01 1.76624745e-01
-1.27451241e-01 2.76755452e-01 1.13700703e-01 -2.33813480e-01
4.91098106e-01 -3.11881810e-01 -7.50487447e-01 -6.49188638e-01
-2.57604212e-01 -1.74156308e-01 5.07884085e-01 -1.38531625e-01
1.22137539e-01 4.17164624e-01 -1.59224164e+00 8.58845353e-01
3.62330735e-01 1.17894018e+00 7.75558203e-02 1.11933434e+00
5.50439000e-01 4.57417220e-01 -1.41957724e+00 -4.72197950e-01
3.56808901e-02 -1.73030794e-01 -4.95438486e-01 -2.12301910e-01
6.39329314e-01 -6.31650567e-01 -5.97985625e-01 1.26375186e+00
8.22865546e-01 -4.16561276e-01 2.98886240e-01 -1.50065041e+00
1.29670098e-01 9.17713165e-01 -5.85519791e-01 5.97830772e-01
6.45468056e-01 1.12219667e-02 6.37385130e-01 -8.15504491e-01
2.64335901e-01 1.41138589e+00 6.26421690e-01 4.38809127e-01
-8.75077546e-01 -1.29518867e+00 -5.19305050e-01 4.71862555e-01
-1.49918902e+00 -9.23929214e-01 1.16041148e+00 -3.74581635e-01
5.83298743e-01 4.00150567e-01 2.67995775e-01 1.45009935e+00
1.06945202e-01 9.39469218e-01 4.84480023e-01 -5.54731727e-01
-1.85207501e-01 5.80630749e-02 -2.02628583e-01 3.00628334e-01
2.01036349e-01 4.18558091e-01 -8.40398371e-01 -3.74893576e-01
7.96431974e-02 -5.73547304e-01 -5.62972963e-01 3.94231707e-01
-9.93026018e-01 6.31289780e-01 -5.71897589e-02 4.40582663e-01
-2.74597056e-04 3.57411623e-01 8.65794003e-01 5.90091825e-01
4.08989221e-01 6.01887822e-01 -8.97034258e-02 -1.50257662e-01
-1.24446738e+00 1.72302380e-01 5.02537251e-01 3.96990746e-01
4.25521517e-03 5.40776789e-01 -9.61346254e-02 7.90738642e-01
1.18381977e-01 6.77983224e-01 3.38492602e-01 -5.89067996e-01
6.58885777e-01 -3.45737427e-01 3.12252074e-01 -1.56624019e+00
-2.98392087e-01 -5.49334407e-01 -4.90609497e-01 -2.37728477e-01
2.62521684e-01 -4.58167285e-01 -5.07329166e-01 1.40168881e+00
2.29028717e-01 5.19142985e-01 -1.51860341e-01 9.42371726e-01
7.85918117e-01 7.91974723e-01 -3.57501447e-01 -2.56846070e-01
1.54111302e+00 -7.85999894e-01 -1.12932551e+00 1.31444335e-02
5.06325662e-01 -1.10347283e+00 5.16022384e-01 9.74007487e-01
-7.96984494e-01 -7.42802680e-01 -1.30171764e+00 4.05174375e-01
-1.65591270e-01 2.11431459e-01 -2.33726129e-02 1.56820667e+00
-7.70639360e-01 1.87893525e-01 -5.17781317e-01 2.50364244e-01
1.41505614e-01 7.70326108e-02 -2.08270907e-01 2.77172893e-01
-1.69160485e+00 1.54385686e-01 3.62018168e-01 -6.21845312e-02
-1.44746876e+00 -5.77960432e-01 -6.91005468e-01 1.69303060e-01
1.80529714e-01 1.84872881e-01 1.27290022e+00 -6.11209571e-01
-1.18839145e+00 7.98244715e-01 1.10440306e-01 -9.82971191e-01
6.35202348e-01 -2.60866165e-01 -1.33532083e+00 8.33900750e-01
-7.33097568e-02 3.92025560e-01 1.64693916e+00 -1.14699829e+00
-7.73398042e-01 1.03730664e-01 -2.31215701e-01 -5.17677665e-01
-6.22088730e-01 6.33029938e-01 -5.04263565e-02 -1.07401621e+00
-4.30963263e-02 -8.40669453e-01 6.34084404e-01 -3.30427974e-01
-6.98275983e-01 4.20700423e-02 1.29176223e+00 -1.18109691e+00
1.53238964e+00 -2.59798908e+00 -6.74595118e-01 -7.16554299e-02
1.56433210e-02 9.34051275e-01 -4.84532639e-02 3.99371177e-01
-2.57599354e-01 3.69201124e-01 1.90659370e-02 -4.52558517e-01
-5.62977530e-02 -7.20813096e-01 -8.28119516e-01 7.96569824e-01
1.28693864e-01 1.78471297e-01 -8.19085479e-01 -3.17108274e-01
-2.58162469e-02 6.26994848e-01 -5.27238965e-01 1.22125126e-01
1.27631128e-01 3.21787685e-01 1.62480325e-02 7.30930448e-01
1.01226211e+00 6.56625926e-01 -1.98196217e-01 -4.12759721e-01
1.00360394e-01 5.29096246e-01 -1.06831253e+00 1.29096580e+00
-2.15268850e-01 1.20289183e+00 5.80282211e-01 -8.28552067e-01
1.01221728e+00 9.21906173e-01 4.18118984e-01 -3.43327850e-01
2.26691157e-01 3.07099730e-01 5.23137078e-02 -7.91891217e-01
5.83373010e-01 1.59204975e-01 -3.32124144e-01 2.66782254e-01
6.43115565e-02 -1.19279653e-01 -1.33104473e-01 3.90156591e-03
1.34074330e+00 -5.27314484e-01 -2.50145465e-01 1.06689297e-01
6.01660013e-01 -6.41549468e-01 3.82916093e-01 9.16801095e-01
-5.81337512e-01 6.62268877e-01 3.85715961e-01 -9.55260471e-02
-7.97625124e-01 -8.15940678e-01 -2.72515714e-01 1.17308688e+00
1.46594331e-01 -4.66080189e-01 -7.20759869e-01 -5.32228768e-01
-1.02240913e-01 4.50450242e-01 -6.61445558e-02 -4.89319086e-01
-6.02087617e-01 -2.74215907e-01 1.67667460e+00 -1.21464275e-01
5.90691924e-01 -1.03517962e+00 -5.35994947e-01 4.57810462e-01
-6.79685593e-01 -1.67768371e+00 -6.11585617e-01 -1.27358794e-01
-7.26917684e-02 -9.47601795e-01 -4.40124840e-01 -9.84501302e-01
-2.25538328e-01 4.59612966e-01 4.87413794e-01 -4.12748791e-02
-2.00599506e-01 -5.88145666e-02 -6.05904698e-01 -4.83224958e-01
-8.81433010e-01 -2.83684820e-01 3.22107315e-01 4.21741575e-01
-3.06694955e-02 -3.85031253e-01 -4.26648110e-01 5.27365744e-01
-9.35653448e-01 -4.72723573e-01 -2.44568996e-02 6.04736984e-01
-1.35935590e-01 4.90911812e-01 9.12851632e-01 -2.44779572e-01
7.50600636e-01 -3.00943375e-01 -4.93353188e-01 -1.06755391e-01
1.41105160e-01 -5.86075485e-01 3.54096830e-01 -7.43683398e-01
-7.83917606e-01 -2.07880810e-01 -5.15066206e-01 -6.02172554e-01
-3.22028011e-01 1.63298875e-01 -3.87709826e-01 -1.45097822e-01
6.86007082e-01 1.32585436e-01 -2.99200922e-01 -5.78476369e-01
-1.83651492e-01 1.40872097e+00 8.78795147e-01 -1.12616830e-01
7.01285958e-01 5.49303234e-01 -3.38431984e-01 -1.30057251e+00
-2.78857023e-01 -5.99353135e-01 -6.30732924e-02 -2.67513186e-01
5.12640953e-01 -9.69190717e-01 -8.26289415e-01 1.02189493e+00
-1.62048244e+00 2.45900944e-01 5.25749743e-01 4.61356640e-01
-2.20581785e-01 9.05387998e-01 -1.03481030e+00 -1.23736739e+00
-4.27381843e-01 -1.30777311e+00 1.32626343e+00 -4.36964959e-01
-2.14102909e-01 -3.56028408e-01 -1.23006761e-01 5.36824048e-01
2.72715539e-01 4.77545351e-01 2.87337095e-01 -1.02517176e+00
1.01257572e-02 -8.11654329e-01 7.46355429e-02 4.55554903e-01
4.33313772e-02 1.83983110e-02 -1.59976840e+00 -5.55404007e-01
2.47023866e-01 -2.84445822e-01 1.00402141e+00 1.94822714e-01
1.20145237e+00 -3.35064918e-01 -4.45492566e-01 4.27122056e-01
6.49770796e-01 5.98726928e-01 5.74954331e-01 -7.25266412e-02
4.58957314e-01 5.13876438e-01 6.46512866e-01 5.96003473e-01
-3.43544900e-01 9.87614810e-01 6.88708961e-01 2.49509647e-01
-2.84721464e-01 -3.57349664e-01 8.09993386e-01 7.18679428e-01
7.36900747e-01 -7.96949089e-01 -8.72796714e-01 5.69295526e-01
-1.10780323e+00 -1.18352330e+00 -1.41437858e-01 2.03646302e+00
5.71960032e-01 4.22091603e-01 3.88086528e-01 1.18279350e+00
1.54076362e+00 4.59889621e-01 -2.18118504e-02 -5.71884274e-01
-7.62436837e-02 1.31775156e-01 2.45235234e-01 5.31503141e-01
-1.65043914e+00 7.29805946e-01 5.86004925e+00 1.35255158e+00
-1.38718987e+00 3.38421792e-01 3.00296336e-01 -6.61045164e-02
2.46637493e-01 -6.15943193e-01 -7.66261637e-01 8.24559629e-01
1.27417898e+00 3.00348550e-01 6.56244159e-02 4.53804046e-01
4.50441331e-01 3.25750589e-01 -7.27434218e-01 1.14665079e+00
3.41931164e-01 -1.08977962e+00 -1.90769076e-01 9.28197876e-02
2.47856885e-01 -3.99024606e-01 4.46396738e-01 1.05746850e-01
-3.85751098e-01 -9.13217723e-01 1.03719068e+00 -2.61070341e-01
1.05831766e+00 -1.10470974e+00 7.22411275e-01 3.17949116e-01
-1.17036974e+00 -2.55808890e-01 -1.38058066e-01 1.61642671e-01
3.35923970e-01 6.57654107e-01 -1.39885128e+00 2.42568538e-01
6.51374042e-01 2.66657084e-01 -1.32715836e-01 1.36636555e+00
-1.50915816e-01 1.19252896e+00 -2.96934783e-01 2.34931633e-01
-7.96025246e-02 8.94014478e-01 1.12603474e+00 1.59095335e+00
5.70295215e-01 -6.52713656e-01 1.97529886e-02 4.06649977e-01
-1.90530807e-01 -2.80541986e-01 -6.78164661e-01 -1.55024752e-01
9.30376232e-01 7.95550764e-01 -4.81686801e-01 2.74784155e-02
1.96399242e-01 7.01749265e-01 -4.72421139e-01 1.06963396e-01
-1.24197197e+00 -8.27153921e-01 5.96712470e-01 1.32002369e-01
5.44220626e-01 6.28984049e-02 4.23438996e-01 -8.33127081e-01
-8.21805298e-02 -1.10599363e+00 2.85930574e-01 -5.80749989e-01
-7.42019653e-01 5.79509377e-01 -2.93085098e-01 -1.51325500e+00
-8.96008089e-02 -3.37188303e-01 -3.73553962e-01 2.32823551e-01
-1.27795029e+00 -7.67172635e-01 8.88257325e-02 4.06090140e-01
6.09682620e-01 -3.27918917e-01 5.10933459e-01 8.06325138e-01
-4.26572829e-01 1.26135552e+00 -7.46156350e-02 6.43606901e-01
8.31940651e-01 -5.77271938e-01 6.44261062e-01 1.06347263e+00
3.15969020e-01 8.45400915e-02 1.15036225e+00 -6.37425065e-01
-1.11958122e+00 -1.01554585e+00 8.38650048e-01 -7.17199147e-02
7.98779666e-01 -6.00233436e-01 -7.51122415e-01 1.51450351e-01
1.52993634e-01 6.57426715e-02 5.98035574e-01 -6.01140320e-01
-5.67150593e-01 -1.23703219e-01 -1.31132829e+00 3.62929702e-02
5.59675097e-01 -9.61414337e-01 -3.68430972e-01 2.97674596e-01
1.18888497e+00 -3.37320685e-01 -4.72139806e-01 5.04963815e-01
6.56988680e-01 -1.02042592e+00 1.08097351e+00 -2.81203091e-01
6.39020205e-02 -2.14673012e-01 -2.17794269e-01 -1.24621999e+00
1.88819632e-01 -1.26400757e+00 -3.37538511e-01 1.55669761e+00
2.03395203e-01 -2.54639089e-01 5.58264673e-01 -4.54124033e-01
-2.50320464e-01 6.49329275e-02 -1.38754940e+00 -1.11240530e+00
-3.41359168e-01 -9.50531960e-01 7.01417506e-01 9.12009597e-01
2.33586147e-01 -1.90165520e-01 -1.08212280e+00 6.25326455e-01
4.46350932e-01 -4.15757626e-01 6.63847804e-01 -7.84755111e-01
-3.67176741e-01 -2.58861184e-01 -7.50128567e-01 -1.06238532e+00
2.31260836e-01 -4.53505576e-01 6.60464214e-03 -3.64085823e-01
-4.79496717e-01 3.71512175e-02 -2.31088117e-01 -1.14716068e-02
1.95066184e-01 6.51830137e-01 1.97828159e-01 -3.60505283e-02
-3.94645065e-01 1.86516434e-01 8.00833464e-01 -5.55872858e-01
3.50692333e-03 3.27601731e-01 -6.46781251e-02 7.59344339e-01
7.74770260e-01 -9.19826388e-01 5.59977256e-02 -1.87603578e-01
-7.56748244e-02 6.58558071e-01 3.69316876e-01 -1.36651897e+00
1.36745885e-01 4.14996505e-01 -1.52299881e-01 -8.12030435e-01
7.01829791e-01 -8.31824958e-01 -8.66015330e-02 9.00226235e-01
-5.70934772e-01 -1.81546465e-01 2.59307057e-01 6.34089172e-01
-4.50277090e-01 -2.91675478e-01 7.86280155e-01 4.37092751e-01
-1.74579322e-01 -2.08900839e-01 -7.43974328e-01 -1.76000893e-01
7.50037730e-01 3.05991294e-03 -5.88648796e-01 -6.36806130e-01
-6.87723815e-01 -3.28062564e-01 -2.66380068e-02 4.86248314e-01
8.62835050e-01 -1.17587268e+00 -7.65514910e-01 3.93688351e-01
-5.04713804e-02 -7.20438004e-01 1.87068194e-01 5.89565814e-01
-6.03756845e-01 5.49069166e-01 3.98239270e-02 -5.33042610e-01
-1.68886840e+00 8.26705396e-01 2.90611297e-01 1.00698590e-01
-2.95372993e-01 1.06806409e+00 -2.35951841e-01 1.37531281e-01
6.65767908e-01 -2.30466709e-01 -2.27920040e-01 3.03829819e-01
9.69721854e-01 6.82202935e-01 3.18767041e-01 -9.19917881e-01
-4.94877338e-01 -6.60187528e-02 9.52751562e-02 -1.78038672e-01
7.29575813e-01 -1.66920666e-02 1.88793182e-01 1.19719706e-01
1.56480348e+00 5.17257512e-01 -7.94800878e-01 -5.71623258e-02
-1.96207777e-01 -6.61826015e-01 3.28984886e-01 -4.64419723e-01
-1.07529652e+00 1.24331319e+00 8.34401906e-01 9.07523155e-01
9.41648960e-01 -1.42774537e-01 1.23850822e+00 5.25030345e-02
2.81554520e-01 -6.84085727e-01 4.83516097e-01 2.57671714e-01
8.85213554e-01 -8.77187073e-01 -4.69013512e-01 -4.71176475e-01
-2.15257034e-01 1.06876969e+00 3.15150857e-01 -8.35766494e-02
7.21521616e-01 5.90920568e-01 -4.12418656e-02 4.65291888e-02
-5.00025988e-01 2.93709368e-01 1.45073429e-01 7.75644183e-01
9.35042351e-02 1.31984293e-01 5.37502766e-02 5.36786914e-01
-5.76783180e-01 -5.73530793e-01 6.21926308e-01 7.76821613e-01
-6.78998113e-01 -6.85421050e-01 -1.10394609e+00 7.45594501e-02
-1.10141873e+00 -1.43975010e-02 -4.29686666e-01 2.26331696e-01
1.26056567e-01 1.81164587e+00 -1.46888243e-02 -8.86596143e-01
-1.53243002e-02 -5.19577377e-02 1.00523308e-01 -1.88462317e-01
-6.11110449e-01 5.02926648e-01 2.43703485e-01 -2.96781182e-01
-1.31790221e-01 -5.77455103e-01 -8.27237248e-01 -2.39948854e-01
-8.24097335e-01 2.31205732e-01 7.40432501e-01 7.62135208e-01
2.28653774e-01 4.13374364e-01 9.83126581e-01 -8.17558110e-01
-5.17047584e-01 -1.01916909e+00 -7.18578100e-01 3.23432714e-01
1.22352684e+00 -6.58407152e-01 -8.41085553e-01 -9.44003686e-02] | [14.130279541015625, 5.810825347900391] |
94e82b5b-a32b-4732-a193-3f0c1807a5ec | atst-audio-representation-learning-with | 2204.12076 | null | https://arxiv.org/abs/2204.12076v3 | https://arxiv.org/pdf/2204.12076v3.pdf | ATST: Audio Representation Learning with Teacher-Student Transformer | Self-supervised learning (SSL) learns knowledge from a large amount of unlabeled data, and then transfers the knowledge to a specific problem with a limited number of labeled data. SSL has achieved promising results in various domains. This work addresses the problem of segment-level general audio SSL, and proposes a new transformer-based teacher-student SSL model, named ATST. A transformer encoder is developed on a recently emerged teacher-student baseline scheme, which largely improves the modeling capability of pre-training. In addition, a new strategy for positive pair creation is designed to fully leverage the capability of transformer. Extensive experiments have been conducted, and the proposed model achieves the new state-of-the-art results on almost all of the downstream tasks. | ['Xiaofei Li', 'Xian Li'] | 2022-04-26 | null | null | null | null | ['instrument-recognition', 'speaker-identification', 'spoken-command-recognition'] | ['audio', 'speech', 'speech'] | [ 3.88749212e-01 2.46274322e-01 -5.94928861e-01 -6.79895401e-01
-1.29049051e+00 -3.75218481e-01 3.46691608e-01 -1.21779703e-01
-8.71723071e-02 7.13363290e-01 3.04415405e-01 -3.20239872e-01
1.94071412e-01 -4.63558584e-01 -6.94381714e-01 -5.74183404e-01
1.86354578e-01 3.31392616e-01 5.15989900e-01 -2.20184550e-01
-2.89590091e-01 -7.95331299e-02 -1.38804328e+00 4.66816097e-01
9.41863000e-01 1.19032586e+00 2.50821233e-01 2.97839582e-01
-1.52905822e-01 1.18664110e+00 -2.79924124e-01 -3.01506788e-01
8.68646577e-02 -6.07929349e-01 -9.12628293e-01 1.56355202e-02
4.88353878e-01 -1.30870178e-01 -3.52947623e-01 7.75792122e-01
7.81635165e-01 2.11896196e-01 3.13951373e-01 -1.47321761e+00
-6.28088713e-01 1.28319955e+00 -5.73570728e-01 1.35197580e-01
8.86507928e-02 -2.33597197e-02 1.33587515e+00 -1.19314873e+00
8.00475180e-02 1.23967361e+00 7.56264806e-01 5.91056228e-01
-1.07632327e+00 -1.18308437e+00 3.40524584e-01 3.75739962e-01
-1.28732979e+00 -4.28520113e-01 9.19518530e-01 -2.13321939e-01
6.99758828e-01 -2.57524997e-01 6.72285855e-01 1.08144522e+00
-5.29981852e-01 1.49482799e+00 1.20179713e+00 -5.85052907e-01
-9.83368233e-02 2.26343721e-01 2.59548545e-01 6.27785802e-01
-3.96813869e-01 1.70839310e-01 -1.09855044e+00 4.78046574e-02
3.95277172e-01 -2.30953172e-01 -2.86277413e-01 -4.25838709e-01
-1.15176094e+00 7.12867141e-01 3.83809090e-01 2.02688411e-01
1.27594709e-01 -1.13767058e-01 5.37666380e-01 5.05938172e-01
7.82825232e-01 1.77715778e-01 -8.23157847e-01 -2.47873932e-01
-9.54543054e-01 -1.82864696e-01 5.56767464e-01 1.11503386e+00
6.30594552e-01 2.76167452e-01 -4.43390131e-01 1.18934298e+00
3.49214435e-01 4.56019908e-01 7.62755394e-01 -6.29315138e-01
6.63493633e-01 6.71493828e-01 -4.73446935e-01 -1.79570824e-01
3.20735611e-02 -9.06546474e-01 -6.43613338e-01 -4.03290391e-01
-7.70614892e-02 -2.35953391e-01 -9.12047565e-01 1.99947369e+00
3.65592003e-01 8.93315315e-01 2.21820980e-01 5.79465091e-01
1.26698506e+00 7.92148888e-01 -1.01421773e-02 -2.43758515e-01
7.99794376e-01 -1.49450374e+00 -6.24568462e-01 -3.06822926e-01
5.10815203e-01 -6.37051344e-01 1.18464589e+00 5.10192156e-01
-1.05440021e+00 -7.91720629e-01 -8.33197236e-01 -1.70960575e-01
1.35541722e-01 3.15887809e-01 4.71399873e-01 4.71520066e-01
-1.16429460e+00 4.55816478e-01 -6.51652992e-01 -6.04957342e-02
5.71049988e-01 4.51604337e-01 -6.16836175e-03 -1.05388246e-01
-1.38999093e+00 5.34993351e-01 4.11178946e-01 -4.12038006e-02
-1.12712657e+00 -1.06418240e+00 -7.93125987e-01 2.62860388e-01
3.48422408e-01 -3.33913684e-01 1.89077783e+00 -8.55032384e-01
-1.90546310e+00 7.40970850e-01 -1.03139095e-01 -4.90357101e-01
3.33457410e-01 -4.09030408e-01 -4.89829212e-01 6.59349784e-02
2.82351494e-01 9.57784355e-01 8.40362132e-01 -1.11847699e+00
-8.28878403e-01 -1.06571831e-01 -1.91268057e-01 3.31661165e-01
-9.12440121e-01 -7.90454596e-02 -2.87083179e-01 -9.79680061e-01
5.24395779e-02 -8.29606771e-01 -1.45401344e-01 -3.82090926e-01
-3.20289403e-01 -7.37550199e-01 8.72228444e-01 -1.71471283e-01
1.33098400e+00 -2.29318357e+00 3.51834483e-02 2.21390761e-02
1.21373639e-01 6.75383091e-01 -4.60598975e-01 4.91219550e-01
-4.04320598e-01 -1.97657794e-01 -2.05828547e-01 -5.57793796e-01
-1.06623314e-01 5.78677878e-02 -7.53497720e-01 -7.74116144e-02
2.74489194e-01 8.05510581e-01 -1.22288895e+00 -6.40775084e-01
-1.40897050e-01 1.82597160e-01 -6.47190273e-01 6.00362599e-01
-2.28670612e-01 6.13992333e-01 -3.02928418e-01 5.28493047e-01
4.42640722e-01 -3.50753248e-01 7.32301101e-02 -7.63414428e-02
-2.26484630e-02 8.59131336e-01 -9.36195254e-01 1.92810023e+00
-4.73777980e-01 4.39366788e-01 -3.29272419e-01 -1.36587286e+00
1.03906739e+00 7.27379382e-01 4.72514957e-01 -5.45832455e-01
1.00014701e-01 2.58583337e-01 -5.19737266e-02 -3.80814642e-01
1.89152539e-01 -4.02759105e-01 6.39293194e-02 5.18180907e-01
6.93568766e-01 -2.29682773e-01 4.87083122e-02 3.86072129e-01
9.58660960e-01 2.07153723e-01 1.16054013e-01 8.38945713e-03
4.70497221e-01 -1.86350569e-01 9.00548756e-01 5.46452165e-01
-2.74674267e-01 4.65746880e-01 1.91065729e-01 1.69923186e-01
-5.02677083e-01 -1.19575882e+00 1.03949770e-01 1.37051511e+00
5.59172630e-02 -5.92926443e-01 -5.09531021e-01 -9.08251703e-01
-3.19095398e-03 6.25502586e-01 -3.47248435e-01 -5.98125041e-01
-3.96409720e-01 -2.63010979e-01 6.23252571e-01 8.59290600e-01
6.86343729e-01 -8.95788610e-01 1.25206694e-01 2.24137709e-01
-3.08537632e-01 -1.22904158e+00 -5.69861293e-01 3.35771799e-01
-1.13377416e+00 -8.25892091e-01 -4.98415649e-01 -1.33933830e+00
3.82176250e-01 5.71306586e-01 1.04212093e+00 -3.83402050e-01
3.15068334e-01 1.88383922e-01 -6.47664547e-01 -5.64631641e-01
-3.38948816e-01 4.17518407e-01 2.43061021e-01 3.12485278e-01
3.37984711e-01 -7.82288611e-01 -1.49207368e-01 3.27001393e-01
-5.41167498e-01 7.92825222e-02 6.16294503e-01 1.16708577e+00
5.85200489e-01 9.35385302e-02 1.22499990e+00 -1.00386322e+00
4.07447755e-01 -4.13088471e-01 -2.01652706e-01 2.70671040e-01
-6.61906302e-01 8.76777172e-02 8.18072140e-01 -7.32991874e-01
-1.35169673e+00 5.67752868e-02 -1.58042803e-01 -8.35786521e-01
3.02234199e-02 6.13068044e-01 -2.06657708e-01 1.13368914e-01
3.19147468e-01 1.85937434e-01 -5.54492138e-02 -8.66072416e-01
4.87030089e-01 1.04741824e+00 6.40337586e-01 -6.26854360e-01
9.02646124e-01 1.38403356e-01 -4.34787303e-01 -3.81569713e-01
-1.65340054e+00 -6.27051473e-01 -5.73090017e-01 -5.52950948e-02
1.53349102e-01 -1.43692994e+00 -5.03965080e-01 7.01965451e-01
-7.36713529e-01 -6.58456266e-01 -7.63923645e-01 7.40729928e-01
-4.72602457e-01 1.44243196e-01 -7.31080472e-01 -6.24884546e-01
-3.14378709e-01 -1.04012918e+00 8.06113899e-01 2.50853747e-01
-1.09006017e-01 -8.37711692e-01 2.24250495e-01 6.19612634e-01
1.23476163e-01 -6.80694222e-01 7.68996477e-01 -1.02610183e+00
-4.68629628e-01 1.65022641e-01 -2.20594294e-02 9.04295981e-01
2.62250304e-01 -3.76455098e-01 -1.29083276e+00 -4.62240726e-01
-6.60170540e-02 -1.00696540e+00 1.03648460e+00 5.82133569e-02
1.29429007e+00 2.02323925e-02 -3.02201957e-01 3.94998282e-01
8.26591372e-01 1.30839601e-01 1.34142032e-02 -1.48927020e-02
7.81493425e-01 3.79590601e-01 8.93518448e-01 3.17552745e-01
6.30520523e-01 5.04549682e-01 8.38402435e-02 -2.50576466e-01
-4.06462371e-01 -8.56591165e-01 6.23603106e-01 1.47434855e+00
3.51500213e-01 -9.01733874e-04 -7.23559320e-01 8.32074523e-01
-1.82416677e+00 -7.80669153e-01 1.00450583e-01 1.78488100e+00
1.40182436e+00 1.88170031e-01 1.91150025e-01 5.06349683e-01
7.00977921e-01 1.61888555e-01 -6.13958657e-01 -2.55921856e-02
7.06093982e-02 6.32316768e-01 1.66185781e-01 2.04646185e-01
-1.13182878e+00 1.26949596e+00 6.88752508e+00 1.22050488e+00
-1.23520124e+00 3.12096030e-01 3.51570845e-01 -1.94124524e-02
-2.04202861e-01 -2.50357203e-02 -1.02931535e+00 3.73526335e-01
9.33968008e-01 -3.21565896e-01 1.55992955e-01 9.41851318e-01
2.59957790e-01 2.77999371e-01 -1.24467254e+00 8.93767297e-01
8.90112072e-02 -9.95567441e-01 4.75706011e-02 -4.97390062e-01
8.87434900e-01 9.22072679e-02 3.13397497e-01 6.86842859e-01
6.09031260e-01 -5.77320814e-01 6.58434510e-01 -2.20235288e-02
1.03236306e+00 -7.90968478e-01 4.68984246e-01 4.24220145e-01
-1.36849415e+00 -3.32075894e-01 5.38476296e-02 -1.29073054e-01
1.18709885e-01 5.82378030e-01 -1.24422669e+00 4.24414337e-01
7.09499419e-01 1.37381852e+00 -6.29966557e-01 1.05340683e+00
-7.98928022e-01 1.55660474e+00 -1.39849916e-01 1.29589424e-01
2.88065165e-01 1.14498779e-01 2.91966289e-01 1.15085828e+00
2.43239537e-01 4.25950773e-02 4.93549556e-01 5.29678226e-01
-4.49407101e-01 1.55994013e-01 -3.73463184e-01 -8.14873204e-02
7.93627679e-01 1.14262545e+00 -1.22293621e-01 -4.83792692e-01
-6.00485206e-01 5.65206409e-01 5.52891314e-01 1.71506613e-01
-6.01747990e-01 -2.46187687e-01 4.08975393e-01 -1.31403074e-01
4.25358951e-01 2.42403105e-01 -1.33673429e-01 -1.26811934e+00
-7.86361843e-02 -9.34248984e-01 4.98419881e-01 -8.67223144e-01
-1.45958209e+00 3.87672633e-01 -1.70437489e-02 -1.70030189e+00
-2.68243968e-01 -1.71798587e-01 -7.49881446e-01 5.40578067e-01
-1.93436229e+00 -1.35071945e+00 8.62495676e-02 7.60033369e-01
7.39620566e-01 -5.76951623e-01 6.83249414e-01 4.89731431e-01
-6.03304863e-01 9.58370149e-01 1.78072453e-01 2.58320987e-01
1.03100753e+00 -1.29869771e+00 3.09280097e-01 7.40102232e-01
4.09016579e-01 3.34408015e-01 1.84174076e-01 -3.58523428e-01
-1.04826176e+00 -1.25218976e+00 1.14113665e+00 -1.42307714e-01
8.25477064e-01 -2.65772790e-01 -9.00887430e-01 9.45401669e-01
2.39599854e-01 1.82961166e-01 1.10485446e+00 3.12573969e-01
-5.49610257e-01 -4.57331866e-01 -7.11826622e-01 2.99639970e-01
1.14074707e+00 -8.11454654e-01 -9.00993645e-01 2.28732049e-01
1.16459954e+00 -5.10960877e-01 -7.59736896e-01 6.34322047e-01
2.81339496e-01 -6.93819344e-01 8.89915347e-01 -7.35649824e-01
4.39763010e-01 -1.93577081e-01 3.55313458e-02 -1.79024553e+00
-2.55201459e-01 -7.66721129e-01 -1.97663561e-01 1.70248210e+00
4.09582794e-01 -3.63068074e-01 8.95298243e-01 8.50610144e-04
-6.65194631e-01 -9.88964796e-01 -8.35550189e-01 -9.92414713e-01
-4.29763645e-02 -4.86338973e-01 6.18583977e-01 1.12195945e+00
3.95545900e-01 9.90469813e-01 -3.31489086e-01 4.27073129e-02
5.93164742e-01 3.40217292e-01 6.37728274e-01 -1.29902291e+00
-3.69537890e-01 -1.47199137e-02 -1.80890843e-01 -1.50709641e+00
5.56250453e-01 -1.33741534e+00 1.39518812e-01 -1.16069913e+00
2.59991139e-01 -7.54781842e-01 -6.95454180e-01 8.45451653e-01
-4.39661384e-01 2.38400668e-01 2.44488537e-01 -7.30945691e-02
-7.86423147e-01 9.71478224e-01 1.35718763e+00 -2.47350127e-01
-3.01815689e-01 3.76005113e-01 -7.68157840e-01 6.57408893e-01
7.69447446e-01 -4.62945521e-01 -9.09181476e-01 -4.24055040e-01
-1.88010141e-01 9.72337369e-03 1.65227175e-04 -9.32929099e-01
2.74937034e-01 -1.15198791e-02 -1.10287890e-01 -7.91805208e-01
3.59717757e-01 -6.82656348e-01 -4.81858283e-01 1.12921886e-01
-9.35701966e-01 -3.29222977e-01 -1.45269185e-02 5.01145840e-01
-6.75656676e-01 -1.23220593e-01 9.34340835e-01 1.99745730e-01
-7.75845408e-01 4.31174785e-01 4.02600169e-02 5.64747214e-01
7.88688540e-01 1.89881653e-01 1.10113593e-02 -3.80740166e-01
-7.23145247e-01 6.32130504e-01 -1.63871810e-01 5.49613833e-01
6.20256841e-01 -1.72613585e+00 -7.56936967e-01 3.99975330e-01
2.73785144e-01 2.60234624e-01 9.42924023e-02 7.12795973e-01
4.32080448e-01 3.57280344e-01 1.02722786e-01 -7.44809449e-01
-1.35340333e+00 3.77127260e-01 2.83642840e-02 -3.85208815e-01
-4.90181327e-01 1.32500088e+00 1.24142051e-01 -7.72169411e-01
5.33297002e-01 -3.94005567e-01 -2.44370401e-01 8.49661306e-02
4.30572599e-01 2.65218645e-01 7.06821978e-02 -4.99544501e-01
-8.38813186e-02 4.00046200e-01 -2.00984240e-01 -1.01814769e-01
1.50094748e+00 -1.91569459e-02 2.25596890e-01 7.88395405e-01
1.25282347e+00 -1.15587801e-01 -1.27231646e+00 -1.02777553e+00
-1.85751006e-01 -2.71586686e-01 1.47851452e-01 -7.64341295e-01
-1.30038118e+00 1.25087416e+00 3.32812607e-01 -6.83921874e-02
1.24376762e+00 8.94716829e-02 1.20394325e+00 3.48514616e-01
2.50231087e-01 -1.17876053e+00 5.22599757e-01 7.36913323e-01
4.05642509e-01 -1.30311954e+00 -3.76127511e-01 -6.66343868e-01
-7.75062978e-01 7.47046113e-01 8.51283669e-01 2.03396678e-01
7.44764805e-01 4.18022752e-01 1.28472030e-01 2.66850471e-01
-1.01227915e+00 -3.23103160e-01 2.23740295e-01 5.81166029e-01
6.21010780e-01 -2.27934625e-02 3.39780971e-02 9.91758645e-01
-2.49274909e-01 4.51292306e-01 1.48774207e-01 8.76109064e-01
-4.43852663e-01 -1.52855408e+00 2.79494692e-02 4.60874230e-01
-2.60416567e-01 -3.44388306e-01 -3.36321771e-01 3.17507297e-01
8.00650716e-02 1.10043037e+00 -1.63470060e-01 -6.37350798e-01
4.24300790e-01 2.49761388e-01 3.43770057e-01 -1.07328928e+00
-7.14220941e-01 2.20696881e-01 8.93038437e-02 -1.99646130e-01
-8.60610187e-01 -5.01794934e-01 -1.41841638e+00 3.24389130e-01
-6.06240511e-01 5.49835622e-01 4.44942638e-02 1.04829371e+00
3.25773031e-01 5.86620748e-01 1.17956376e+00 -5.32072663e-01
-8.94413233e-01 -1.03260827e+00 -7.48262525e-01 2.60817558e-01
3.88771951e-01 -7.45864749e-01 -2.18756244e-01 1.38940305e-01] | [15.214885711669922, 5.187015533447266] |
808fc97a-09d4-4e4e-8d52-52d66762366e | an-attention-free-long-short-term-memory-for | 2209.09548 | null | https://arxiv.org/abs/2209.09548v1 | https://arxiv.org/pdf/2209.09548v1.pdf | An Attention Free Long Short-Term Memory for Time Series Forecasting | Deep learning is playing an increasingly important role in time series analysis. We focused on time series forecasting using attention free mechanism, a more efficient framework, and proposed a new architecture for time series prediction for which linear models seem to be unable to capture the time dependence. We proposed an architecture built using attention free LSTM layers that overcome linear models for conditional variance prediction. Our findings confirm the validity of our model, which also allowed to improve the prediction capacity of a LSTM, while improving the efficiency of the learning task. | ['Ludovic De Villelongue', 'Hugo Inzirillo'] | 2022-09-20 | null | null | null | null | ['time-series-prediction'] | ['time-series'] | [-4.31631118e-01 -6.76543638e-02 4.14311774e-02 -3.63042563e-01
-1.84773728e-01 -1.71635985e-01 6.59836173e-01 -1.18673742e-01
-3.87526363e-01 6.04095697e-01 2.49104649e-01 -6.75599217e-01
-3.33211541e-01 -7.02843070e-01 -6.12497866e-01 -6.24700487e-01
-4.75442111e-01 2.30668053e-01 1.15057692e-01 -3.01628053e-01
2.10620269e-01 6.36640191e-01 -1.49330997e+00 2.41229624e-01
5.69747090e-01 1.17866397e+00 3.14296573e-01 6.77082181e-01
-8.65726173e-02 1.42039609e+00 -5.38042605e-01 -1.30825236e-01
3.31719941e-03 -1.79213479e-01 -4.53394800e-01 -7.06079662e-01
6.64497120e-03 -3.38425130e-01 -6.57376885e-01 4.34046060e-01
1.89829752e-01 4.26785856e-01 5.35101414e-01 -1.22205067e+00
-7.72906303e-01 7.39372492e-01 -1.15827180e-01 7.13523984e-01
-2.00315028e-01 -1.63432270e-01 8.57323050e-01 -4.90692437e-01
2.27664009e-01 1.02006269e+00 9.91470933e-01 4.94342953e-01
-9.38698888e-01 -6.41169071e-01 3.50601196e-01 5.71639001e-01
-8.55458915e-01 -2.32148111e-01 9.39555705e-01 -3.67649466e-01
1.59591734e+00 3.53714347e-01 7.40032911e-01 1.25533581e+00
7.73462951e-01 7.24786401e-01 9.31977928e-01 -3.44053149e-01
1.33764669e-01 -1.17984407e-01 1.99821129e-01 1.57286108e-01
-2.86305070e-01 2.79196441e-01 -1.00123044e-02 -3.34193334e-02
9.21175957e-01 4.74655837e-01 1.23941511e-01 2.63742179e-01
-9.94981766e-01 8.54597270e-01 5.82220256e-01 1.06137872e+00
-7.90152192e-01 4.59260792e-01 6.36675060e-01 4.81788903e-01
9.42125976e-01 3.97937387e-01 -1.04157472e+00 -3.12006295e-01
-9.36179698e-01 -5.09101748e-02 6.18984461e-01 2.66810596e-01
2.16992766e-01 6.79800928e-01 -4.28763539e-01 4.66743678e-01
-1.23330489e-01 5.33095658e-01 8.94018292e-01 -6.86277926e-01
1.63685143e-01 4.07126665e-01 -1.22294769e-01 -1.00905621e+00
-8.08905005e-01 -7.65527904e-01 -1.06447911e+00 -7.35181645e-02
2.74613053e-01 -3.45817417e-01 -9.31457639e-01 1.86829412e+00
-4.29182768e-01 4.80165690e-01 -1.70955881e-01 6.23626828e-01
1.79216981e-01 1.04527974e+00 3.84866983e-01 -5.48186064e-01
7.92503774e-01 -8.27048898e-01 -1.00051129e+00 7.07418099e-02
7.41683066e-01 -4.21259135e-01 8.77150536e-01 3.45582336e-01
-8.36693883e-01 -7.63067186e-01 -6.21347964e-01 -1.74786896e-02
-5.08546054e-01 5.11602238e-02 9.11004305e-01 3.46030742e-01
-1.24558163e+00 1.02205908e+00 -1.18113554e+00 -1.80622280e-01
1.70517400e-01 3.50862622e-01 -9.06588137e-02 5.31218767e-01
-1.51722181e+00 1.22445500e+00 4.64445412e-01 2.53258407e-01
-6.87603593e-01 -6.38398528e-01 -5.02127051e-01 5.45877576e-01
-1.02804087e-01 -3.59878659e-01 1.21615636e+00 -1.28877711e+00
-1.72268760e+00 1.36797875e-01 -2.35054284e-01 -9.76143241e-01
3.36439580e-01 -2.74538159e-01 -7.89640427e-01 -2.34906405e-01
-3.68673533e-01 3.05132687e-01 9.15855944e-01 -3.60259324e-01
-2.49386102e-01 -3.46544832e-01 -2.71290243e-01 -5.77856898e-01
-4.88108754e-01 -1.65565293e-02 3.38194996e-01 -8.72107208e-01
-1.41579434e-01 -8.68197501e-01 -4.62922752e-01 -5.87570608e-01
1.94820538e-01 -5.50214708e-01 8.42675030e-01 -1.07827997e+00
1.67449212e+00 -2.08439875e+00 -2.82174438e-01 4.04271632e-02
-2.13818923e-02 3.80120277e-01 -2.80104220e-01 6.13536894e-01
-5.06129265e-01 1.33242995e-01 9.86044481e-02 -4.30062175e-01
-2.03331746e-02 2.69875467e-01 -1.09984612e+00 2.29382977e-01
2.47777745e-01 1.17497075e+00 -4.69549924e-01 6.18747063e-02
3.85951012e-01 6.66927814e-01 -3.08542967e-01 2.96051979e-01
-3.55116457e-01 3.71950567e-01 -2.30638340e-01 5.06892689e-02
4.78952259e-01 -4.05352414e-01 4.00575623e-02 9.29986760e-02
-6.17462099e-01 4.30322230e-01 -4.15975720e-01 1.39510071e+00
-6.17035866e-01 9.59560692e-01 -7.58669972e-01 -1.16457081e+00
1.00850320e+00 6.01845980e-01 6.95419431e-01 -1.02417243e+00
1.68807492e-01 1.65259883e-01 3.12015146e-01 -6.70077980e-01
3.82502228e-01 -2.19480902e-01 2.29783371e-01 3.01926434e-01
1.80333793e-01 4.03949618e-01 -1.69086382e-01 -3.55303437e-01
9.92017865e-01 3.20258111e-01 2.02058554e-01 -3.14493537e-01
3.57650101e-01 -2.19982505e-01 2.48331666e-01 6.25881732e-01
-2.75922995e-02 1.59400970e-01 6.10115826e-01 -1.09733248e+00
-1.18872857e+00 -4.91049111e-01 -5.03014065e-02 1.42845297e+00
-7.84623623e-01 -3.05372179e-01 -3.04450929e-01 -3.25458169e-01
-3.23755175e-01 1.10522509e+00 -9.93544638e-01 -1.35387763e-01
-7.56666064e-01 -7.30872214e-01 4.21462357e-01 7.02513635e-01
1.14290290e-01 -1.36377442e+00 -7.05326915e-01 6.48206592e-01
-1.93614829e-02 -7.41881013e-01 -5.53756878e-02 5.93606234e-01
-1.33739948e+00 -6.18287563e-01 -9.27528024e-01 -3.77746552e-01
-1.28973881e-02 -1.45957291e-01 8.98386359e-01 -3.19240689e-01
1.52644306e-01 9.31263342e-02 -2.51119137e-01 -8.25122654e-01
-1.33022606e-01 3.58295441e-01 -1.63971875e-02 -1.87443495e-01
5.17370760e-01 -8.77822280e-01 -5.61035633e-01 4.85573001e-02
-8.06572914e-01 -1.73039630e-01 3.19095105e-01 5.90228498e-01
6.72056451e-02 -5.76764792e-02 9.43799436e-01 -4.55173761e-01
7.23245680e-01 -6.95802987e-01 -5.90495884e-01 1.02120250e-01
-7.98123956e-01 2.37272620e-01 9.98525798e-01 -7.55677283e-01
-9.25068140e-01 -4.03162301e-01 -4.99736428e-01 -7.52758980e-01
6.42918199e-02 6.86030507e-01 6.58834398e-01 2.60577112e-01
3.55129570e-01 4.51228589e-01 -9.77856889e-02 -6.62270248e-01
-6.36719912e-02 2.74371624e-01 -3.62839364e-02 5.23954583e-03
4.15659323e-02 3.51028144e-01 3.65459621e-02 -7.33918130e-01
-5.98841727e-01 -1.41648769e-01 -4.32897180e-01 -1.85501575e-01
7.41135955e-01 -7.74144650e-01 -9.22517121e-01 5.02914250e-01
-1.38301635e+00 -4.82390583e-01 -2.82642663e-01 8.60237360e-01
-4.82974678e-01 -4.48322929e-02 -9.52771008e-01 -1.22896802e+00
-2.37843439e-01 -4.47494447e-01 6.07341826e-01 -1.47051483e-01
-1.23858839e-01 -1.58643723e+00 5.12751579e-01 -4.71093059e-01
1.18194723e+00 2.35275120e-01 1.02128744e+00 -9.27865088e-01
-8.03008676e-02 -3.88406992e-01 -9.77356285e-02 3.51645291e-01
-1.30853459e-01 9.57434848e-02 -1.29497802e+00 -4.81145941e-02
4.59948719e-01 1.96230575e-01 1.04410040e+00 7.98791647e-01
1.53310883e+00 -3.37210566e-01 -2.01219618e-01 3.84985447e-01
1.19766092e+00 5.28246403e-01 8.95811737e-01 2.73392737e-01
4.94879514e-01 5.54757059e-01 8.70509818e-02 5.34401119e-01
4.79420543e-01 4.24921095e-01 3.39559615e-01 -9.43448395e-02
3.11088622e-01 -3.12737644e-01 4.50360984e-01 1.21015775e+00
-3.99903387e-01 -3.99519533e-01 -9.93451238e-01 5.41219950e-01
-2.11941814e+00 -1.33848178e+00 -3.80431086e-01 1.98148632e+00
2.83139944e-01 2.27544516e-01 1.67642981e-01 2.39289269e-01
3.53763312e-01 2.37423509e-01 -4.83763635e-01 -6.21689558e-01
-1.46964550e-01 2.83367634e-01 3.35728586e-01 2.14432344e-01
-1.00674129e+00 7.47724056e-01 7.86643124e+00 7.26682782e-01
-1.77632785e+00 3.08619410e-01 6.17922366e-01 9.45060998e-02
-2.68404961e-01 -3.60109597e-01 -3.74667823e-01 7.69238949e-01
1.91057909e+00 -2.04280004e-01 2.18639046e-01 8.01275790e-01
4.53383625e-01 3.43334883e-01 -9.41363692e-01 7.79752672e-01
-1.04772620e-01 -1.19630694e+00 -3.45115811e-02 -4.64210520e-03
5.19660532e-01 3.55623871e-01 2.20425308e-01 8.33344519e-01
1.71299540e-02 -1.07394803e+00 4.71682608e-01 1.07360685e+00
1.74637258e-01 -5.78018785e-01 8.79058242e-01 5.59676051e-01
-9.10974979e-01 -4.22518879e-01 -3.69638801e-01 -7.27405787e-01
1.84660777e-01 5.70905626e-01 -5.58641791e-01 2.81629473e-01
6.45985067e-01 8.13945949e-01 -4.30803567e-01 9.04228270e-01
2.25381911e-01 8.43277693e-01 -2.49436423e-01 -8.60225707e-02
4.84303206e-01 1.51055276e-01 3.58670950e-01 1.15366077e+00
6.92525685e-01 -1.28633723e-01 -1.35381714e-01 6.50380909e-01
3.81650448e-01 5.33172973e-02 -6.22675776e-01 -3.56407315e-01
-7.83009455e-02 7.08847344e-01 -6.62024796e-01 -3.16822171e-01
-5.11595488e-01 7.16068029e-01 4.38359410e-01 4.81735289e-01
-1.04728520e+00 -1.62308171e-01 3.34850699e-01 1.16127394e-01
4.91358280e-01 -4.46920305e-01 -2.20981911e-01 -1.33411634e+00
-7.81090483e-02 -3.51121038e-01 5.09792089e-01 -9.11459506e-01
-1.40405750e+00 1.02764380e+00 -1.08603137e-02 -1.16488063e+00
-7.22732961e-01 -7.35723853e-01 -6.93141401e-01 9.83611405e-01
-1.43995154e+00 -1.05078566e+00 2.08702102e-01 6.29143357e-01
5.75629115e-01 -9.12045091e-02 9.77424204e-01 4.45581138e-01
-3.84359926e-01 9.13347304e-02 1.89686805e-01 -2.62587041e-01
3.83688539e-01 -1.06362247e+00 4.95390892e-01 6.46236897e-01
8.55718777e-02 5.66803634e-01 7.04119146e-01 -4.78384316e-01
-1.00205767e+00 -9.44783866e-01 1.28857625e+00 -3.25084031e-01
9.25926447e-01 -2.31941149e-01 -1.28802609e+00 8.92748654e-01
2.60820091e-01 -3.97089571e-01 7.80917525e-01 2.54624277e-01
-2.43947208e-01 -1.59501106e-01 -7.39523590e-01 2.95313329e-01
6.36933982e-01 -6.29383564e-01 -7.52914429e-01 1.31766096e-01
9.54941869e-01 3.55185336e-03 -7.47627556e-01 4.10241723e-01
5.44503391e-01 -8.60417843e-01 6.58548832e-01 -7.71113634e-01
4.74491388e-01 3.88010561e-01 -6.42011538e-02 -1.22743666e+00
-8.43817890e-01 -4.17707086e-01 -7.51748741e-01 8.73503447e-01
4.37550098e-01 -1.04616892e+00 3.88800025e-01 8.08912575e-01
-1.46065250e-01 -5.25462985e-01 -1.03128541e+00 -9.21508729e-01
2.81076938e-01 -9.29918706e-01 6.69361234e-01 1.04654264e+00
-1.43265566e-02 4.28786241e-02 -6.85054243e-01 -1.62153825e-01
4.37298156e-02 6.93007708e-02 3.68219316e-02 -1.53187621e+00
-9.17544290e-02 -7.09377348e-01 -2.18763217e-01 -6.96262479e-01
4.69209582e-01 -6.49565995e-01 -4.73111808e-01 -1.33921731e+00
-2.53027827e-01 -9.93727744e-02 -1.03024769e+00 5.31599045e-01
7.66383931e-02 6.92325160e-02 2.02345982e-01 3.17997187e-01
-3.62580001e-01 7.62732387e-01 8.68414879e-01 1.81813955e-01
-2.32663542e-01 2.62082547e-01 -3.03482115e-01 6.30236685e-01
1.02075315e+00 -5.17680168e-01 -3.92536044e-01 -5.95248878e-01
4.03584152e-01 2.92407662e-01 4.10394728e-01 -1.01315260e+00
3.00527602e-01 -1.19502030e-01 7.35369742e-01 -5.02155185e-01
3.42512935e-01 -1.20924437e+00 3.24352205e-01 7.74204195e-01
-5.30865133e-01 4.16060358e-01 6.67893171e-01 2.30614528e-01
-3.27253491e-01 3.42515297e-02 3.24468374e-01 -3.02557610e-02
-6.86433911e-01 3.48041892e-01 -5.87754488e-01 -6.27700925e-01
8.35233986e-01 3.90138812e-02 -4.48400714e-02 -5.44829249e-01
-1.05810022e+00 -1.25511557e-01 -2.28284568e-01 7.17353940e-01
3.52207452e-01 -1.30504167e+00 -5.27996659e-01 2.91144222e-01
-3.71440917e-01 -9.92842197e-01 5.05152941e-01 9.27848876e-01
-2.29955107e-01 1.08014083e+00 -4.62605685e-01 -2.86980689e-01
-4.76206064e-01 1.04705024e+00 4.74790961e-01 -4.94979799e-01
-6.21851146e-01 4.24113393e-01 7.06604347e-02 -9.88313556e-02
1.47572339e-01 -8.63912582e-01 -5.46874404e-01 1.84880361e-01
6.52195513e-01 6.34495318e-01 1.53214976e-01 -1.66522682e-01
-2.91121304e-01 4.42248911e-01 1.07978791e-01 3.21973045e-03
1.78080034e+00 -9.47774947e-02 -1.76060081e-01 1.28807318e+00
1.02763832e+00 -3.74565929e-01 -1.13087082e+00 2.51701847e-02
2.94746727e-01 -1.47286922e-01 2.92948037e-01 -8.64147663e-01
-9.39460158e-01 1.10859311e+00 6.47427380e-01 9.34460580e-01
1.29018462e+00 -5.10160208e-01 8.33446324e-01 2.83659995e-01
2.05228642e-01 -8.42450380e-01 -3.12245011e-01 1.00523818e+00
1.02860820e+00 -1.30797386e+00 -5.11769533e-01 4.00977373e-01
-4.28063959e-01 1.56451559e+00 4.20508981e-01 -4.80881840e-01
1.02578068e+00 2.99698293e-01 -1.50572225e-01 3.97514217e-02
-1.38232362e+00 -2.17948899e-01 5.14961779e-01 2.88310707e-01
9.46504414e-01 6.16631657e-02 -3.13055396e-01 8.20142210e-01
-8.15285966e-02 3.68670881e-01 1.65562332e-02 2.74048448e-01
-1.20523125e-01 -7.49682605e-01 -1.02484278e-01 4.66736227e-01
-7.95047820e-01 -1.47525460e-01 2.90867984e-02 7.51601100e-01
-9.13891569e-02 7.99262464e-01 5.05038261e-01 -4.70526427e-01
1.62177309e-01 4.33775753e-01 1.77462175e-01 4.96034883e-02
-8.95556033e-01 3.90508473e-02 -2.52667427e-01 -5.68538070e-01
-3.05865467e-01 -3.87295246e-01 -7.19106615e-01 -4.07011002e-01
-1.46627322e-01 1.26028284e-01 7.87887871e-01 8.72257411e-01
5.53703368e-01 9.22315776e-01 7.53189802e-01 -8.23057055e-01
-4.24323499e-01 -1.35223472e+00 -5.31320453e-01 -3.99849601e-02
6.76396847e-01 -2.51439154e-01 -3.61525536e-01 -8.66834223e-02] | [6.841998100280762, 3.0286407470703125] |
ebeffcb6-58ca-47b4-8911-d7d0d538eac5 | detection-based-defense-against-adversarial | 1806.09186 | null | http://arxiv.org/abs/1806.09186v3 | http://arxiv.org/pdf/1806.09186v3.pdf | Detection based Defense against Adversarial Examples from the Steganalysis Point of View | Deep Neural Networks (DNNs) have recently led to significant improvements in
many fields. However, DNNs are vulnerable to adversarial examples which are
samples with imperceptible perturbations while dramatically misleading the
DNNs. Moreover, adversarial examples can be used to perform an attack on
various kinds of DNN based systems, even if the adversary has no access to the
underlying model. Many defense methods have been proposed, such as obfuscating
gradients of the networks or detecting adversarial examples. However it is
proved out that these defense methods are not effective or cannot resist
secondary adversarial attacks. In this paper, we point out that steganalysis
can be applied to adversarial examples detection, and propose a method to
enhance steganalysis features by estimating the probability of modifications
caused by adversarial attacks. Experimental results show that the proposed
method can accurately detect adversarial examples. Moreover, secondary
adversarial attacks cannot be directly performed to our method because our
method is not based on a neural network but based on high-dimensional
artificial features and FLD (Fisher Linear Discriminant) ensemble. | ['Nenghai Yu', 'Yujia Liu', 'Yiwei Zhang', 'Weiming Zhang', 'Jiayang Liu', 'Hongyue Zha', 'Dongdong Hou'] | 2018-06-21 | detection-based-defense-against-adversarial-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Liu_Detection_Based_Defense_Against_Adversarial_Examples_From_the_Steganalysis_Point_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Liu_Detection_Based_Defense_Against_Adversarial_Examples_From_the_Steganalysis_Point_CVPR_2019_paper.pdf | cvpr-2019-6 | ['steganalysis'] | ['computer-vision'] | [ 5.90716124e-01 1.13785286e-02 3.92422646e-01 -3.91684808e-02
-2.73413777e-01 -9.27134931e-01 7.91101694e-01 -5.03237128e-01
-3.99717957e-01 7.00511515e-01 -2.40362808e-01 -5.13109863e-01
3.10127527e-01 -1.15651536e+00 -8.63425970e-01 -9.20124829e-01
-1.42547235e-01 -1.29562438e-01 4.76391345e-01 -4.14632648e-01
1.53629407e-01 9.27690566e-01 -1.19105315e+00 5.13369404e-02
6.60796404e-01 8.30940485e-01 -1.65191785e-01 9.56969380e-01
8.77917111e-02 8.20156038e-01 -1.16927207e+00 -4.86836463e-01
7.02651143e-01 -6.17221355e-01 -2.49340847e-01 -2.92395800e-01
3.42467070e-01 -6.03117228e-01 -1.03628993e+00 1.83759999e+00
4.44032103e-01 -1.05136238e-01 6.46637976e-01 -1.58445835e+00
-8.78903627e-01 5.67583501e-01 -1.22156017e-01 2.88354844e-01
-4.77233119e-02 3.45221400e-01 2.38125265e-01 -4.04400051e-01
2.14665964e-01 1.59375012e+00 4.58357871e-01 1.03681159e+00
-7.76723623e-01 -1.16213763e+00 -1.01503745e-01 3.17407668e-01
-1.02763987e+00 -4.32163298e-01 1.09121919e+00 -1.92950621e-01
2.85858393e-01 5.00384927e-01 2.91175783e-01 1.61333323e+00
4.50510412e-01 5.89121878e-01 1.15750718e+00 -3.23102206e-01
1.65571049e-01 2.38214180e-01 -3.39429468e-01 6.66650832e-01
5.00543535e-01 6.70623243e-01 1.64698586e-01 -3.07150245e-01
8.21843863e-01 2.10972965e-01 -5.49929678e-01 4.01456375e-03
-9.43061948e-01 1.06185353e+00 6.40542209e-01 2.54247099e-01
-1.04049981e-01 2.33193859e-01 4.34471756e-01 6.52040780e-01
5.05340174e-02 4.29125071e-01 -2.57331163e-01 3.44571799e-01
-3.57126951e-01 -2.77020931e-02 8.23522687e-01 6.66783631e-01
3.97757888e-01 6.65714800e-01 2.25123405e-01 3.40472817e-01
3.84896725e-01 9.35613632e-01 6.01113975e-01 -5.89620531e-01
3.80694270e-01 2.49829158e-01 -3.16498339e-01 -1.49915826e+00
2.09859665e-03 -2.01503411e-01 -1.17782414e+00 8.10750484e-01
4.02455091e-01 -4.63382572e-01 -9.68256474e-01 1.60217416e+00
3.83348912e-02 3.22017550e-01 5.19735456e-01 6.43932045e-01
7.42058277e-01 6.62915528e-01 -1.96465239e-01 1.40088379e-01
8.83406043e-01 -7.46200562e-01 -6.89622104e-01 -3.42900217e-01
3.63838136e-01 -6.25080407e-01 4.85848457e-01 3.39948803e-01
-5.96437216e-01 -4.58025426e-01 -1.33715427e+00 4.78885233e-01
-6.37231588e-01 -5.56042254e-01 2.52569795e-01 1.35853493e+00
-8.34176958e-01 7.50136793e-01 -7.02743411e-01 8.04654136e-02
6.10040247e-01 5.40028691e-01 -4.24054950e-01 -1.52823329e-01
-1.76739252e+00 9.43560302e-01 6.27727509e-01 3.31225216e-01
-1.34419954e+00 -6.36552423e-02 -8.86626184e-01 -5.67495376e-02
-1.37461023e-02 -7.41889849e-02 6.83461368e-01 -1.39982998e+00
-1.29088140e+00 5.08465827e-01 5.86642206e-01 -6.95809066e-01
9.30079401e-01 1.34109572e-01 -9.86265957e-01 3.51716638e-01
-4.29078132e-01 2.58402586e-01 1.28855503e+00 -1.13553667e+00
-2.68873632e-01 -3.44243556e-01 1.67329282e-01 -2.69410670e-01
-8.03293109e-01 6.98804483e-02 2.02467754e-01 -9.77298319e-01
5.05730473e-02 -9.72207129e-01 -2.51712352e-01 3.01107734e-01
-6.47623181e-01 5.11764884e-01 1.62550640e+00 -5.83038449e-01
8.47379267e-01 -2.20307589e+00 -3.69977325e-01 3.42415065e-01
3.01450193e-01 1.13088512e+00 -2.59517342e-01 2.53043562e-01
-2.06763089e-01 5.12564600e-01 -2.58474499e-01 3.13936800e-01
3.21424231e-02 2.64702618e-01 -5.56427896e-01 8.19147766e-01
8.23229700e-02 8.05016756e-01 -8.83240998e-01 -1.91976637e-01
3.42154622e-01 7.15823054e-01 -7.65293986e-02 5.97604550e-02
1.71613827e-01 3.50009143e-01 -6.01762295e-01 4.46729809e-01
1.05175912e+00 3.06569785e-01 -2.89885420e-02 -5.47640026e-02
5.72194755e-01 -2.07702160e-01 -8.57551157e-01 4.97482032e-01
-1.50126904e-01 1.05591190e+00 -6.34042248e-02 -1.25891638e+00
1.11893713e+00 3.87375861e-01 -2.67483652e-01 -3.19476157e-01
4.18456465e-01 3.79401326e-01 3.53006005e-01 -4.30512309e-01
-8.21251869e-02 -2.80275010e-02 -2.38865260e-02 1.46064162e-01
-6.41403794e-02 1.98140457e-01 -3.86061430e-01 -4.64828536e-02
1.34625065e+00 -5.42250633e-01 2.01532006e-01 6.60113699e-04
8.93125296e-01 -3.98195446e-01 4.28305894e-01 1.00885689e+00
-3.32034081e-01 2.85070986e-01 3.37128699e-01 -4.85509932e-01
-1.08151793e+00 -1.01573813e+00 -3.90180200e-02 3.93080533e-01
3.46101195e-01 3.37741137e-01 -8.08990777e-01 -1.25375509e+00
-8.44631344e-02 4.39786375e-01 -5.35501361e-01 -7.87063181e-01
-5.03712595e-01 -5.93462050e-01 1.35334539e+00 3.89382482e-01
1.10323167e+00 -1.09986448e+00 -1.46621332e-01 1.47669211e-01
2.87663192e-01 -1.21708512e+00 -1.69462308e-01 -5.13545796e-02
-8.21647882e-01 -1.12920547e+00 -6.87352300e-01 -8.42018068e-01
8.16393554e-01 2.14832187e-01 5.14204741e-01 2.59411663e-01
-1.04777917e-01 -2.67190188e-02 -4.63402629e-01 -5.61223507e-01
-1.27752125e+00 -3.94741774e-01 3.77207309e-01 6.42907023e-02
3.33146513e-01 -7.23377943e-01 -3.74895364e-01 4.96006280e-01
-1.30950761e+00 -5.84399760e-01 6.21450365e-01 7.75424123e-01
-8.14145580e-02 5.48693120e-01 7.16726303e-01 -1.06356418e+00
4.84087408e-01 -3.85387391e-01 -6.43286109e-01 9.20994803e-02
-4.13874090e-01 1.47787556e-01 1.28501177e+00 -9.14387226e-01
-6.37726963e-01 -2.76705980e-01 -4.40923691e-01 -9.31371331e-01
-4.05674815e-01 5.78527525e-02 -6.56307518e-01 -8.82220209e-01
8.45371544e-01 4.46622908e-01 7.43648410e-02 -2.29968965e-01
3.17109860e-02 6.83430135e-01 5.03320992e-01 2.97438577e-02
1.53855574e+00 4.89469379e-01 2.64274836e-01 -8.39758396e-01
-4.16079491e-01 3.39002669e-01 -2.19989568e-01 -2.58999109e-01
5.65210104e-01 -5.59932470e-01 -5.65446138e-01 1.12024415e+00
-1.13099432e+00 -4.12087962e-02 2.18386620e-01 4.38972622e-01
-6.30998313e-02 7.70924509e-01 -6.61516190e-01 -7.79922783e-01
-1.65661782e-01 -1.20510542e+00 2.03701481e-01 1.86290041e-01
3.63866448e-01 -1.27452397e+00 -2.99047798e-01 9.33752581e-02
6.19950175e-01 8.92611802e-01 7.33906150e-01 -1.25370455e+00
-4.61957514e-01 -7.61084378e-01 -1.13753416e-02 9.43330407e-01
3.72879624e-01 1.44390091e-01 -1.12245560e+00 -5.12122393e-01
4.94539261e-01 -1.45385280e-01 7.62241483e-01 1.74152758e-02
1.25277793e+00 -9.13146675e-01 -2.68230587e-01 7.94626474e-01
1.45774841e+00 5.41446745e-01 9.40483212e-01 4.39088047e-01
9.26300883e-01 2.58827269e-01 1.71289310e-01 -1.01292646e-02
-3.80357176e-01 2.29830533e-01 1.09528661e+00 -9.52042118e-02
1.73145548e-01 -1.14384308e-01 7.22699761e-01 4.52270150e-01
2.02219114e-01 -7.87022173e-01 -6.46159112e-01 1.44513801e-01
-1.41216993e+00 -1.16069329e+00 -2.75852233e-01 2.07007790e+00
4.33181614e-01 5.62714756e-01 -3.05370212e-01 4.87109751e-01
1.28000891e+00 4.10758138e-01 -8.00043464e-01 -4.96406108e-01
-3.24752986e-01 5.64020909e-02 8.11581671e-01 3.68080109e-01
-1.33776772e+00 8.51250172e-01 6.16896439e+00 9.39186752e-01
-1.24150062e+00 4.51810099e-02 5.15762448e-01 3.03548843e-01
-1.36772692e-01 -3.05941552e-01 -5.26094675e-01 7.51209140e-01
8.19994330e-01 -1.01504378e-01 3.54021490e-01 9.09071326e-01
-1.84863940e-01 5.97892284e-01 -7.66485512e-01 7.36649632e-01
1.94866717e-01 -1.07911134e+00 2.82752901e-01 2.20571995e-01
6.93957925e-01 -2.02067390e-01 3.22236896e-01 3.04006070e-01
5.09114087e-01 -1.05999887e+00 2.12736636e-01 2.41290838e-01
5.48586190e-01 -1.07228863e+00 1.01072848e+00 5.27255237e-01
-7.07272112e-01 2.50720587e-02 -6.52651906e-01 5.12321666e-02
-2.69821048e-01 3.38243157e-01 -6.86399221e-01 2.80417085e-01
4.26720679e-01 3.39478463e-01 -4.16238159e-01 6.07414544e-01
-4.64522749e-01 7.04240322e-01 -2.47342497e-01 -2.31906891e-01
3.64930660e-01 1.14384048e-01 9.09842730e-01 8.42414439e-01
3.99355054e-01 -6.68600649e-02 -1.33791089e-01 8.10069025e-01
-2.04184532e-01 -2.99072713e-01 -1.27018678e+00 -3.07646573e-01
5.14286995e-01 9.50777173e-01 -6.45489395e-01 -2.41206706e-01
-2.17922911e-01 1.18811440e+00 -2.12114260e-01 4.45838302e-01
-8.95695984e-01 -8.66996765e-01 7.62317359e-01 -3.11735600e-01
3.33520263e-01 -4.11390048e-03 2.29018599e-01 -1.16116631e+00
-1.56992972e-01 -1.13532853e+00 7.04772696e-02 -3.86887312e-01
-1.43720114e+00 8.07107985e-01 -4.76349443e-01 -1.51064157e+00
-8.63887295e-02 -8.84509206e-01 -8.37347686e-01 7.44542599e-01
-1.33319008e+00 -7.95721292e-01 -6.67762458e-02 9.79154110e-01
1.98877648e-01 -6.49517536e-01 8.41938734e-01 1.46677509e-01
-4.90661621e-01 9.99162436e-01 4.71148103e-01 8.27992737e-01
4.68765140e-01 -8.71298075e-01 5.95797598e-01 1.47169018e+00
-7.28596598e-02 2.82740861e-01 8.98003161e-01 -5.54988384e-01
-1.17513204e+00 -1.39687264e+00 2.69092351e-01 -1.40020773e-01
6.92244411e-01 -4.12993938e-01 -9.72023785e-01 7.27506220e-01
-1.78942829e-02 4.31591243e-01 5.47387838e-01 -8.26909542e-01
-3.55696082e-01 5.36015294e-02 -1.69426191e+00 7.16683567e-01
7.95880735e-01 -4.88765448e-01 -5.30828416e-01 3.08110774e-01
8.00706863e-01 -2.60237545e-01 -5.52346647e-01 3.49820703e-01
2.94189841e-01 -1.03662550e+00 1.09806097e+00 -6.71063185e-01
4.30193126e-01 -2.88371116e-01 -2.19107509e-01 -1.17874300e+00
4.56650835e-03 -7.19904602e-01 -3.84093970e-01 1.13325024e+00
9.00482610e-02 -1.16373610e+00 6.85110509e-01 2.49369875e-01
2.20087186e-01 -2.38956660e-01 -8.28166187e-01 -1.14732504e+00
7.75789246e-02 -3.57967645e-01 7.50445604e-01 1.15199518e+00
-5.46638191e-01 -2.85295039e-01 -6.28073037e-01 8.34238768e-01
9.63094175e-01 -5.40346265e-01 6.78148746e-01 -1.17971218e+00
-1.95209712e-01 -3.94529939e-01 -1.16400075e+00 -6.57263339e-01
3.23208213e-01 -6.07781708e-01 2.34034320e-04 -8.74125183e-01
-4.79337811e-01 -2.88596869e-01 -5.90516686e-01 3.70567888e-01
-2.61480182e-01 6.36008501e-01 6.96364865e-02 9.49319378e-02
-1.39650041e-02 3.83594722e-01 1.17322052e+00 -3.88287514e-01
3.52481663e-01 3.12832505e-01 -6.17756903e-01 1.00880837e+00
9.51831937e-01 -8.82753849e-01 -4.61799890e-01 -1.69547275e-01
-1.00681432e-01 -1.41145125e-01 6.42474473e-01 -1.29405057e+00
7.43618459e-02 -5.29556535e-02 5.01153529e-01 -1.39381468e-01
1.26094595e-01 -1.16548550e+00 6.93713352e-02 9.93032515e-01
-1.23069711e-01 -1.95746139e-01 1.12710036e-01 8.15109193e-01
-1.94219455e-01 -5.23021102e-01 1.08415282e+00 -1.84681952e-01
-6.66926086e-01 4.28504109e-01 -5.17313123e-01 -8.38739797e-02
1.26529896e+00 -3.44974846e-01 -4.44319487e-01 -5.01551092e-01
-5.93596995e-01 -8.67594630e-02 5.05904257e-01 3.01589072e-01
8.52341115e-01 -1.44021177e+00 -5.68592489e-01 4.50316966e-01
-2.82763928e-01 -3.46314013e-01 1.51874289e-01 1.11532316e-01
-8.56400549e-01 1.70444131e-01 -5.21408677e-01 -2.00054362e-01
-1.45081484e+00 1.02749550e+00 4.40890521e-01 -5.49082719e-02
-4.55258876e-01 8.41895759e-01 1.87803045e-01 -3.46198767e-01
2.76683390e-01 1.48982450e-01 -2.80721605e-01 -3.92997950e-01
7.02390015e-01 3.32275748e-01 -1.92573622e-01 -6.61203146e-01
-2.98301876e-01 2.71131635e-01 -2.24455148e-01 2.01919347e-01
1.10712481e+00 9.01189893e-02 1.26113892e-02 -5.79691119e-02
1.51740694e+00 -1.82626978e-01 -1.13882351e+00 -1.94840491e-01
-5.38423955e-01 -6.47007465e-01 9.72318277e-02 -4.13136035e-01
-1.49811530e+00 1.07428896e+00 8.28724325e-01 8.62387836e-01
1.18822753e+00 -5.93058348e-01 9.71034706e-01 6.14310563e-01
3.42777759e-01 -4.77603197e-01 3.18402052e-03 2.96886206e-01
6.22891366e-01 -1.32495522e+00 -3.87002021e-01 -2.56016344e-01
-3.04317445e-01 1.33383763e+00 6.03824914e-01 -6.17734015e-01
7.07446635e-01 2.69979924e-01 3.16751540e-01 9.22205001e-02
-2.84137100e-01 3.58059287e-01 4.40062769e-02 9.00138974e-01
-4.35401112e-01 -1.37135535e-01 -4.84340591e-03 9.83529985e-02
-1.98774695e-01 -4.02359784e-01 7.76670754e-01 8.39741886e-01
-4.51763213e-01 -1.05006647e+00 -7.96036422e-01 3.42471749e-01
-7.80617177e-01 -1.32531196e-01 -3.91651720e-01 8.36128175e-01
9.65172201e-02 9.44436967e-01 -2.60869235e-01 -8.26522112e-01
7.49075785e-02 -1.94795534e-01 2.13258371e-01 -2.08496317e-01
-4.28317577e-01 -4.04956043e-01 -3.43444675e-01 -3.09759825e-01
-2.23119915e-01 -3.05537581e-01 -7.47539937e-01 -7.02262521e-01
-5.34824789e-01 4.35555412e-04 5.39645851e-01 9.77147579e-01
1.24578341e-03 5.60747027e-01 1.14791119e+00 -7.95275271e-01
-7.98566282e-01 -8.40503812e-01 -6.50411189e-01 3.25632364e-01
7.36595392e-01 -4.12628949e-01 -1.10941243e+00 -1.74354762e-01] | [5.525314807891846, 7.907380104064941] |
05c4c217-b85e-49c2-873e-ccbd1439492d | unsupervised-speech-recognition-via-segmental | 1812.09323 | null | http://arxiv.org/abs/1812.09323v1 | http://arxiv.org/pdf/1812.09323v1.pdf | Unsupervised Speech Recognition via Segmental Empirical Output Distribution Matching | We consider the problem of training speech recognition systems without using
any labeled data, under the assumption that the learner can only access to the
input utterances and a phoneme language model estimated from a non-overlapping
corpus. We propose a fully unsupervised learning algorithm that alternates
between solving two sub-problems: (i) learn a phoneme classifier for a given
set of phoneme segmentation boundaries, and (ii) refining the phoneme
boundaries based on a given classifier. To solve the first sub-problem, we
introduce a novel unsupervised cost function named Segmental Empirical Output
Distribution Matching, which generalizes the work in (Liu et al., 2017) to
segmental structures. For the second sub-problem, we develop an approximate MAP
approach to refining the boundaries obtained from Wang et al. (2017).
Experimental results on TIMIT dataset demonstrate the success of this fully
unsupervised phoneme recognition system, which achieves a phone error rate
(PER) of 41.6%. Although it is still far away from the state-of-the-art
supervised systems, we show that with oracle boundaries and matching language
model, the PER could be improved to 32.5%.This performance approaches the
supervised system of the same model architecture, demonstrating the great
potential of the proposed method. | ['Chih-Kuan Yeh', 'Jianshu Chen', 'Chengzhu Yu', 'Dong Yu'] | 2018-12-23 | unsupervised-speech-recognition-via-segmental-1 | https://openreview.net/forum?id=Bylmkh05KX | https://openreview.net/pdf?id=Bylmkh05KX | iclr-2019-5 | ['unsupervised-speech-recognition'] | ['speech'] | [ 6.42636716e-01 5.05919158e-01 -2.30611473e-01 -5.85358620e-01
-1.33850050e+00 -7.10715473e-01 3.77308697e-01 -7.73627013e-02
-5.09023368e-01 5.18450081e-01 -1.35656316e-02 -5.92466533e-01
2.25970849e-01 -4.12162393e-01 -8.86684537e-01 -6.08882546e-01
1.54321432e-01 6.95288122e-01 1.69795454e-01 1.74626634e-01
1.21340975e-01 9.35310572e-02 -1.57580233e+00 2.73052603e-01
1.11942065e+00 1.12044871e+00 3.87150466e-01 9.26380575e-01
-1.82116047e-01 5.43676198e-01 -6.09261096e-01 -2.53361464e-01
3.73805165e-02 -6.04511976e-01 -1.06008923e+00 5.05732536e-01
3.45120192e-01 -1.95542164e-02 -3.05100195e-02 1.13913918e+00
2.08101630e-01 2.72611141e-01 7.62666821e-01 -8.62296283e-01
-4.27894384e-01 9.38383698e-01 -8.95297453e-02 -1.09795660e-01
2.24202015e-02 -4.27022904e-01 9.29515719e-01 -1.12078404e+00
5.60182072e-02 8.79278958e-01 4.95495856e-01 6.52695119e-01
-1.02555954e+00 -4.02928263e-01 1.80615112e-01 1.15454711e-01
-1.58651567e+00 -7.44345605e-01 4.18950647e-01 -5.28684616e-01
1.19135797e+00 1.81151077e-01 -3.24495770e-02 7.05163717e-01
-4.09642041e-01 8.52386534e-01 1.16946816e+00 -9.05664384e-01
4.41066444e-01 4.13965851e-01 2.92295814e-01 5.80184340e-01
-3.13950926e-01 1.92264915e-01 -3.64571959e-01 1.41169831e-01
3.55487853e-01 -3.48055154e-01 -2.68353283e-01 -1.81597874e-01
-1.01467860e+00 6.12536967e-01 -5.10985702e-02 3.82177681e-01
-1.23040825e-01 -1.93857893e-01 1.81685805e-01 1.63609371e-01
5.31193018e-01 1.20572217e-01 -7.13482797e-01 -4.26360130e-01
-1.22245932e+00 -4.53466386e-01 1.03298199e+00 1.06346846e+00
8.20092797e-01 1.03030987e-01 1.74296662e-01 1.05928338e+00
3.54926676e-01 4.22162026e-01 5.87611735e-01 -7.26678967e-01
7.57727206e-01 2.01110706e-01 -6.22108728e-02 -1.25618681e-01
-3.44817303e-02 -4.52199668e-01 -7.61242688e-01 -1.72484457e-01
6.60711944e-01 -2.12376460e-01 -1.12306798e+00 1.66058421e+00
8.27971920e-02 5.60831130e-01 2.91494131e-01 5.27044117e-01
3.80276412e-01 1.06866336e+00 -1.58751607e-01 -3.62686932e-01
1.05463660e+00 -1.29939687e+00 -6.99120879e-01 -3.19456011e-01
7.96556056e-01 -6.52016461e-01 1.01249897e+00 6.07929170e-01
-9.94095504e-01 -6.96502745e-01 -1.07527947e+00 2.36603886e-01
-4.48234349e-01 5.55779934e-01 1.66245520e-01 1.03248537e+00
-1.14642668e+00 3.06877166e-01 -8.11919868e-01 -3.09719920e-01
4.57130335e-02 4.97682750e-01 -8.15010518e-02 1.21947095e-01
-1.14420521e+00 7.50603080e-01 7.10467398e-01 2.12045714e-01
-9.43830013e-01 -5.47004163e-01 -8.10140908e-01 1.95501596e-01
4.68783915e-01 -8.21806788e-02 1.54334903e+00 -1.12264645e+00
-2.05237460e+00 8.19014192e-01 -5.50624311e-01 -6.35005951e-01
2.60840982e-01 -2.54435211e-01 -3.73229444e-01 -5.32844523e-03
-2.75583178e-01 6.74180090e-01 6.84769690e-01 -1.10740101e+00
-9.59910095e-01 -1.10758714e-01 -4.30734575e-01 2.61371940e-01
-4.41944450e-01 -6.60203099e-02 -7.54287720e-01 -5.68980336e-01
1.01716883e-01 -9.37868297e-01 -7.96758011e-02 -7.02016413e-01
-5.12565136e-01 -4.42659736e-01 5.04226267e-01 -9.74299252e-01
1.40709341e+00 -2.11427522e+00 1.86900154e-01 3.28900903e-01
-2.77195662e-01 5.00519395e-01 5.50762936e-02 2.17071742e-01
-5.48556894e-02 2.21907526e-01 -1.09299672e+00 -7.64863193e-01
4.44831848e-02 2.55819708e-01 -5.05954087e-01 2.40604565e-01
2.56853640e-01 7.20620215e-01 -6.63769126e-01 -1.35076717e-01
1.99905783e-01 3.09521019e-01 -4.30258781e-01 4.39088285e-01
-1.33560717e-01 4.26519722e-01 2.10558921e-01 4.50147510e-01
5.94517827e-01 1.55074567e-01 2.90019006e-01 2.50378460e-01
-7.25325346e-02 7.41059542e-01 -1.30171955e+00 1.53742266e+00
-6.27729952e-01 6.47096395e-01 1.66691661e-01 -1.43352306e+00
1.06280899e+00 6.04540467e-01 1.35375530e-01 -3.83434027e-01
-8.62236023e-02 5.44031024e-01 -1.16209410e-01 -2.57948816e-01
3.25761259e-01 -1.52190983e-01 -2.19807476e-01 4.67100441e-01
3.41033161e-01 -8.67659301e-02 -1.36374295e-01 -9.08956826e-02
7.84618139e-01 -5.79918437e-02 6.51284903e-02 -2.90772945e-01
7.15632975e-01 -3.65841806e-01 3.35230976e-01 8.91811192e-01
-2.27178812e-01 7.01525986e-01 2.14702249e-01 1.64024085e-01
-1.01645958e+00 -1.20614541e+00 -3.04441690e-01 1.15293157e+00
-2.07577154e-01 -2.44142994e-01 -1.38657391e+00 -7.44604111e-01
-3.58942479e-01 6.58968568e-01 -2.96412915e-01 9.54335704e-02
-6.32455289e-01 -3.45682591e-01 8.52756143e-01 6.26426995e-01
5.88531077e-01 -1.06053829e+00 -7.11170584e-02 1.57175779e-01
-3.61453891e-01 -1.30211759e+00 -6.99059129e-01 5.88357747e-01
-7.53532648e-01 -5.66708326e-01 -8.00794363e-01 -1.35330641e+00
7.41305768e-01 -1.63606495e-01 8.37521076e-01 -2.77152658e-01
3.89270812e-01 4.02240396e-01 -3.66995513e-01 -8.49460512e-02
-7.21278429e-01 5.25504053e-01 2.01445282e-01 2.52428174e-01
3.72198373e-01 -2.81861216e-01 -2.39895191e-02 4.20811087e-01
-8.82730424e-01 -1.94753092e-02 5.31998098e-01 9.81223583e-01
8.13061535e-01 2.34556034e-01 1.00327837e+00 -1.02783823e+00
2.83927500e-01 -5.08994639e-01 -5.85377514e-01 4.58533049e-01
-7.80744135e-01 1.01942159e-01 7.51407743e-01 -4.21351016e-01
-1.12184298e+00 5.43255866e-01 -4.70395625e-01 -1.70367926e-01
-5.58361590e-01 4.95105773e-01 -5.75555921e-01 3.44882548e-01
1.50226608e-01 5.75519383e-01 -1.73473075e-01 -7.83956647e-01
5.65167546e-01 1.34844995e+00 9.66413796e-01 -4.70806777e-01
5.03864229e-01 2.31891647e-02 -7.06503808e-01 -9.67497587e-01
-8.13355923e-01 -7.79058456e-01 -9.41358745e-01 6.64040223e-02
8.04033875e-01 -8.57933342e-01 -4.43313599e-01 6.62245214e-01
-1.17687762e+00 -6.86578989e-01 -2.09543601e-01 6.71935916e-01
-8.05735767e-01 4.81376618e-01 -6.20851517e-01 -1.10611308e+00
-7.81723186e-02 -9.93591905e-01 9.73081768e-01 5.75962290e-03
-1.13927461e-01 -9.94720042e-01 7.56035373e-02 5.44575274e-01
2.08050087e-01 -5.02649784e-01 6.94640040e-01 -1.06999409e+00
-3.60411227e-01 -3.30119096e-02 1.51720509e-01 8.45061004e-01
2.98986822e-01 -2.38687828e-01 -1.36859262e+00 -1.58080697e-01
2.12998018e-01 -3.72828275e-01 1.11072254e+00 2.50783741e-01
1.31631541e+00 -4.38616574e-01 -1.02146894e-01 5.57602465e-01
9.44372296e-01 5.05931437e-01 6.35583043e-01 -1.29970703e-02
6.62041068e-01 8.93767238e-01 5.33352613e-01 5.44304959e-02
3.64863127e-01 6.33028507e-01 6.22888468e-02 5.69657199e-02
-1.56391272e-03 -5.96922934e-01 6.32970870e-01 1.60176361e+00
4.96541888e-01 -3.75146657e-01 -1.04743171e+00 8.20513666e-01
-1.80266285e+00 -6.06073618e-01 1.99505463e-01 2.45023465e+00
1.00907397e+00 2.23287135e-01 1.98732063e-01 3.98572803e-01
1.01136494e+00 -2.05202237e-01 -4.44215477e-01 -7.07969308e-01
-6.28656819e-02 3.81992012e-01 4.33474928e-01 9.97240126e-01
-1.25087249e+00 1.19667566e+00 6.46935987e+00 1.13478386e+00
-9.71174836e-01 1.49522468e-01 8.58072102e-01 4.66705203e-01
-2.36123949e-01 7.74510577e-02 -1.02088916e+00 5.46083212e-01
1.64962876e+00 2.30349004e-01 6.71773493e-01 7.71748841e-01
-8.15422684e-02 -5.77024631e-02 -1.21146166e+00 9.31994617e-01
3.37693244e-01 -9.63832498e-01 -2.10374773e-01 -1.70642920e-02
8.51621926e-01 1.05126686e-01 6.97552562e-02 5.35353661e-01
1.36111602e-01 -1.26953232e+00 8.20171773e-01 3.18639666e-01
9.25008953e-01 -8.38680863e-01 6.37880266e-01 7.68591642e-01
-1.24904597e+00 1.91863239e-01 -1.86492831e-01 1.41048729e-02
2.37201020e-01 5.16583920e-01 -1.08105397e+00 2.16949224e-01
4.52200681e-01 4.54656094e-01 -2.04575539e-01 9.96551156e-01
-3.60684633e-01 1.41491151e+00 -5.03162086e-01 7.11336210e-02
1.64531112e-01 -2.46915981e-01 2.33237505e-01 1.69392693e+00
3.30811083e-01 -4.79552969e-02 7.13763982e-02 7.32973456e-01
-2.17460737e-01 9.74268466e-02 -3.90935212e-01 -1.95147887e-01
8.04320574e-01 8.57709348e-01 -7.37989247e-01 -5.28876662e-01
-2.33027175e-01 1.24206662e+00 5.48911035e-01 4.91650701e-01
-6.54555142e-01 -5.94792187e-01 2.89262116e-01 -3.20725024e-01
7.64495850e-01 -2.62263745e-01 -3.76910925e-01 -9.40374494e-01
1.89068362e-01 -8.18587363e-01 1.65202186e-01 -2.56069332e-01
-1.12722051e+00 6.25857532e-01 -1.54036582e-01 -1.06626809e+00
-5.18600881e-01 -7.00035930e-01 -5.47718883e-01 9.03871179e-01
-1.46904123e+00 -7.00393438e-01 2.72011280e-01 2.46584252e-01
8.68013620e-01 -1.81013733e-01 9.86797094e-01 2.78802365e-01
-6.11994624e-01 9.87993181e-01 5.45794785e-01 2.75087059e-01
5.12658417e-01 -1.40745425e+00 5.87133527e-01 9.20429885e-01
5.97531080e-01 4.26929116e-01 1.97498664e-01 -3.77048552e-01
-1.07722485e+00 -1.12564385e+00 1.26220918e+00 -4.87849742e-01
4.88625914e-01 -7.61905909e-01 -1.08293116e+00 7.44174838e-01
1.13530435e-01 -2.77548939e-01 8.10795128e-01 -6.81068003e-02
-3.11892927e-01 -1.71818223e-03 -9.12474632e-01 2.23752007e-01
8.11076462e-01 -9.03494239e-01 -7.23975539e-01 1.11155443e-01
8.40763152e-01 -3.43560010e-01 -8.26339900e-01 4.36770022e-01
3.47471178e-01 -6.81729436e-01 6.43994510e-01 -4.28239703e-01
2.08253004e-02 -2.38702521e-01 -3.69191200e-01 -1.49462509e+00
2.17645332e-01 -6.73522472e-01 -2.31859103e-01 1.73473597e+00
8.67732227e-01 -7.26018548e-01 6.73947811e-01 2.91138411e-01
-5.16268194e-01 -8.01127553e-01 -1.15779650e+00 -9.97859776e-01
3.48125637e-01 -8.68169427e-01 5.01057565e-01 7.67624021e-01
1.84894115e-01 3.70081782e-01 -2.74122775e-01 3.86311740e-01
3.78537476e-01 -1.83735505e-01 4.77056772e-01 -7.26821721e-01
-6.22022569e-01 -4.18728173e-01 -5.91345057e-02 -1.88936222e+00
5.02301455e-01 -1.00191057e+00 8.31301630e-01 -1.28803110e+00
-6.41116723e-02 -5.90486526e-01 -4.84172642e-01 3.01010162e-01
-1.09683193e-01 5.05780578e-02 4.00277860e-02 1.62528500e-01
-6.29221261e-01 5.13511717e-01 5.08651316e-01 -1.84302688e-01
-2.94315606e-01 2.47687623e-01 -5.07391572e-01 6.86250091e-01
9.42868471e-01 -4.09203440e-01 -4.98826504e-01 -3.19288701e-01
-3.27218086e-01 -1.51692526e-02 -6.64327368e-02 -9.50756013e-01
5.16625524e-01 1.48200482e-01 -1.62343845e-01 -6.92313790e-01
1.53131336e-01 -6.86744809e-01 -2.97850877e-01 1.91093162e-01
-4.91972357e-01 -3.53697419e-01 3.00377160e-01 5.45900881e-01
-4.69260365e-01 -4.12111342e-01 5.61921716e-01 3.13239783e-01
-6.73169315e-01 -6.17645420e-02 -7.18591452e-01 2.46648297e-01
7.26674020e-01 -2.37057701e-01 -1.22876137e-01 -3.59432638e-01
-8.57927382e-01 9.34452266e-02 2.44249087e-02 2.87438780e-01
4.98145580e-01 -1.21210647e+00 -6.10503316e-01 3.68848383e-01
3.24268900e-02 8.47003758e-02 -9.65124667e-02 6.87924623e-01
-2.36131862e-01 7.34318018e-01 4.65673655e-01 -8.50390434e-01
-8.49513113e-01 2.96623528e-01 4.06806529e-01 -2.07526237e-01
-1.72641590e-01 9.69653428e-01 3.00304472e-01 -9.80357945e-01
7.25525260e-01 -4.25480634e-01 -5.64806797e-02 -1.72068521e-01
4.17655677e-01 2.93875545e-01 3.95570517e-01 -8.17003369e-01
-3.54233027e-01 5.34335732e-01 1.13495057e-02 -5.25797188e-01
9.02656198e-01 -3.02520573e-01 2.44621351e-01 8.25838804e-01
1.25592363e+00 1.53854847e-01 -1.44665396e+00 -3.53876293e-01
4.13149625e-01 3.18113677e-02 -1.52656317e-01 -8.44367445e-01
-7.59615123e-01 1.09420049e+00 6.13037109e-01 2.22786829e-01
1.15167332e+00 1.99991405e-01 9.86714602e-01 4.20280039e-01
2.04394624e-01 -1.20007706e+00 -1.84193090e-01 1.01290858e+00
3.70673418e-01 -1.18439579e+00 -7.60579169e-01 -4.93080825e-01
-6.20975196e-01 8.95704269e-01 4.49080467e-01 7.12079331e-02
7.56146252e-01 4.30145144e-01 1.28364503e-01 3.94864321e-01
-6.69382870e-01 -4.07605231e-01 4.77828801e-01 6.52306736e-01
4.57144767e-01 3.70749354e-01 -4.49871123e-02 8.02655280e-01
-3.90964746e-01 -3.23166549e-01 1.89561144e-01 5.61712861e-01
-7.37662077e-01 -1.16025746e+00 -3.26442599e-01 1.83726981e-01
-2.87534177e-01 -3.10270607e-01 -5.75752020e-01 3.12545896e-01
9.99052972e-02 1.31072974e+00 3.02769899e-01 -5.72399795e-01
1.78313375e-01 6.00976884e-01 2.02763125e-01 -9.23436165e-01
-4.79937375e-01 1.79467872e-01 8.99592713e-02 -5.91314808e-02
-1.94575205e-01 -4.49520290e-01 -1.45939696e+00 2.31510028e-01
-5.25195777e-01 5.37098229e-01 7.61925578e-01 1.20734334e+00
1.63335949e-01 3.57653916e-01 8.44033122e-01 -6.16344213e-01
-6.70433819e-01 -1.16451073e+00 -5.33282936e-01 1.66570082e-01
2.66189367e-01 -2.70562530e-01 -5.32222390e-01 4.16885644e-01] | [14.540863990783691, 6.725465774536133] |
9aa87e6a-5d45-4813-be84-08465ff0ef5b | a-graph-matching-perspective-with | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Qin_A_Graph_Matching_Perspective_With_Transformers_on_Video_Instance_Segmentation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Qin_A_Graph_Matching_Perspective_With_Transformers_on_Video_Instance_Segmentation_CVPR_2022_paper.pdf | A Graph Matching Perspective With Transformers on Video Instance Segmentation | Video Instance Segmentation (VIS) needs to automatically track and segment multiple objects in videos that rely on modeling the spatial-temporal interactions of the instances. This paper presents a graph matching-based method to formulate VIS. Unlike traditional tracking-by-detection paradigm or bottom-up generative solutions, we introduce a novel, learnable graph matching Transformer to predict the instances by heuristically learning the spatial-temporal relationships. Specifically, we take advantage of the powerful Transformer and exploit temporal feature aggregation to capture the long-term temporal information across frames in an implicit way. Our model generates instance proposals per-frame and performs the data association between current and historical frames via the proposed graph matching based on the enhanced feature. Furthermore, to make the whole network optimization end-to-end differentiable, we relax the original graph matching into continuous quadratic programming and then unroll the training of it into a deep graph network. Extensive experimental results on two representative available benchmarks, including YouTube-VIS19 and OVIS, verify the effectiveness of our graph matching Transformer. | ['Jianbing Shen', 'Yilong Yin', 'Xiushan Nie', 'Xiankai Lu', 'Zheyun Qin'] | 2022-01-01 | a-graph-matching-perspective-with-1 | https://openaccess.thecvf.com/content/CVPR2022/html/Qin_A_Graph_Matching_Perspective_With_Transformers_on_Video_Instance_Segmentation_CVPR_2022_paper.html | https://openaccess.thecvf.com/content/CVPR2022/papers/Qin_A_Graph_Matching_Perspective_With_Transformers_on_Video_Instance_Segmentation_CVPR_2022_paper.pdf | cvpr-2022-6 | ['video-instance-segmentation', 'graph-matching'] | ['computer-vision', 'graphs'] | [-3.54561768e-02 1.31675079e-01 -3.60458583e-01 -4.41713482e-01
-6.83554590e-01 -4.17603642e-01 3.33963662e-01 6.56392332e-03
-2.72512525e-01 4.59434241e-01 -8.16099271e-02 5.86094223e-02
-2.35634178e-01 -7.29741156e-01 -1.05830646e+00 -6.27834320e-01
-4.09394711e-01 4.37622011e-01 5.73786318e-01 1.92935556e-01
3.95437405e-02 3.13844502e-01 -1.23806083e+00 3.57848927e-02
7.53896117e-01 1.19364238e+00 3.48998755e-01 6.80241346e-01
-5.74935600e-02 1.00612533e+00 -1.60787731e-01 -4.48878080e-01
5.67441225e-01 -4.00364280e-01 -5.91361403e-01 7.04846382e-01
6.96923554e-01 -3.92447770e-01 -7.52544105e-01 9.89457190e-01
3.03612828e-01 5.57851732e-01 1.54397935e-01 -1.44601381e+00
-4.63370979e-01 5.44314027e-01 -8.32641602e-01 3.92113000e-01
3.45968872e-01 3.56125534e-01 9.95575964e-01 -7.71781862e-01
9.37673807e-01 1.15818620e+00 6.45675361e-01 3.65112811e-01
-8.99403453e-01 -5.18181741e-01 9.37056184e-01 4.41943228e-01
-1.30861080e+00 -9.52040851e-02 9.93289232e-01 -5.88819087e-01
6.24515414e-01 8.98177642e-03 1.24919248e+00 7.14631021e-01
-2.59538386e-02 1.06887233e+00 5.85804939e-01 1.04336977e-01
-1.02440432e-01 -2.50287473e-01 1.26009077e-01 1.21880352e+00
-6.31006584e-02 1.74249336e-01 -4.72409368e-01 -1.35093722e-02
9.52623606e-01 4.37375069e-01 -3.05662900e-01 -5.07939219e-01
-1.16304219e+00 6.50211930e-01 6.80304468e-01 -7.79953897e-02
-5.51181853e-01 5.10215044e-01 3.76526177e-01 1.31787853e-02
5.04530787e-01 -9.02513340e-02 -3.59831512e-01 1.09895103e-01
-1.22467351e+00 3.54797006e-01 6.52134836e-01 1.28541362e+00
1.02112949e+00 -2.38974895e-02 -6.00796878e-01 3.90244186e-01
4.77143764e-01 1.74146965e-01 2.96861708e-01 -9.55998957e-01
5.10362506e-01 6.70800865e-01 -5.91543652e-02 -1.14697134e+00
-9.36406553e-02 -5.17520308e-01 -5.54827154e-01 -2.63234407e-01
2.67521143e-01 -1.96419731e-01 -1.21554136e+00 1.55474412e+00
7.31555223e-01 1.14260077e+00 -3.74132603e-01 1.16576302e+00
7.07988620e-01 7.90704727e-01 1.60296142e-01 -4.45256889e-01
1.03390336e+00 -1.38352835e+00 -6.47363901e-01 -8.11054334e-02
3.91539514e-01 -3.11627358e-01 5.88356614e-01 8.80366308e-04
-1.31459820e+00 -6.50378764e-01 -7.69868851e-01 3.84631124e-03
-3.87827121e-02 7.19658658e-02 6.33093536e-01 2.28235722e-02
-9.05510604e-01 6.48657560e-01 -1.02905130e+00 -3.78308669e-02
6.94529891e-01 4.75345314e-01 -1.33913904e-01 -1.12171523e-01
-9.32286024e-01 1.05699174e-01 5.90632915e-01 3.90511513e-01
-1.21704412e+00 -7.54628897e-01 -1.03697014e+00 5.02894260e-02
9.23425734e-01 -1.03079391e+00 9.06872988e-01 -1.07422268e+00
-1.31869507e+00 6.25686705e-01 -2.73429185e-01 -6.30847633e-01
8.93388748e-01 -1.85255542e-01 -3.72274704e-02 1.88226238e-01
2.10505277e-01 6.72378004e-01 8.91463935e-01 -1.09097624e+00
-9.88887906e-01 -2.26956338e-01 2.85219729e-01 2.21610725e-01
-7.42310584e-02 -1.90956041e-01 -1.32314277e+00 -5.81063628e-01
1.51391644e-02 -9.43277717e-01 -5.13100922e-01 4.27950248e-02
-5.12898147e-01 -2.84039468e-01 9.27620947e-01 -7.04920888e-01
1.27882266e+00 -2.04359078e+00 3.58769953e-01 2.50791967e-01
4.74207044e-01 1.38305292e-01 -1.04201809e-01 6.39732406e-02
2.30657399e-01 -2.00878456e-01 -1.69532210e-01 -5.20509660e-01
-1.01650290e-01 2.96444952e-01 -5.42653874e-02 5.77145159e-01
1.92352593e-01 1.31184280e+00 -1.21524870e+00 -8.04447532e-01
3.12135071e-01 4.49308604e-01 -6.25940204e-01 4.62591290e-01
-6.37933016e-01 4.19059694e-01 -7.56745338e-01 5.68642020e-01
5.41340947e-01 -5.02234936e-01 1.44640103e-01 -3.72169971e-01
7.87482262e-02 -1.03151232e-01 -1.17740488e+00 1.93708503e+00
-1.30860857e-03 5.15656054e-01 -1.24895096e-01 -1.14952898e+00
5.87653816e-01 7.02470616e-02 8.85833442e-01 -5.47621310e-01
3.37849325e-03 -1.29217148e-01 -4.19593185e-01 -6.41129196e-01
3.95189106e-01 2.49397963e-01 1.41095743e-01 -2.02960297e-02
1.86994046e-01 3.05978119e-01 5.31757176e-01 4.69429910e-01
8.28523099e-01 5.49983323e-01 -3.69932689e-02 1.24378100e-01
5.84882200e-01 2.38043349e-02 9.87943053e-01 6.27091646e-01
-1.22274891e-01 5.36598265e-01 4.19410855e-01 -5.30008972e-01
-6.16832435e-01 -9.14243162e-01 4.07531142e-01 9.27655458e-01
6.54560804e-01 -3.63740712e-01 -7.76997805e-01 -1.04220271e+00
1.54838972e-02 2.72810876e-01 -7.02192605e-01 -4.47078310e-02
-7.92611182e-01 -5.06888449e-01 -4.62760683e-03 7.06714392e-01
5.65967500e-01 -9.85082626e-01 -3.57093185e-01 4.80248809e-01
-1.83855623e-01 -1.32570028e+00 -1.05200875e+00 -2.58830637e-01
-7.82447159e-01 -1.19100440e+00 -6.96304619e-01 -9.09544885e-01
6.42812610e-01 3.33851874e-01 1.03664911e+00 3.06265444e-01
-4.03570116e-01 5.90046883e-01 -1.12759903e-01 5.60968667e-02
1.16232641e-01 7.97538087e-03 -4.97008979e-01 3.52178633e-01
3.20064723e-02 -4.31209117e-01 -8.36990654e-01 2.35072389e-01
-6.08191907e-01 2.47566715e-01 4.48021680e-01 7.19660878e-01
1.16118670e+00 2.29691584e-02 4.00601268e-01 -9.44797218e-01
1.77791923e-01 -6.39596939e-01 -9.59553182e-01 4.55470830e-01
-4.33881402e-01 -1.18807793e-01 1.93253718e-02 -6.06090724e-01
-8.43412936e-01 4.74192142e-01 2.57529706e-01 -1.22008634e+00
2.12326184e-01 6.17643774e-01 -7.11512417e-02 -8.69538933e-02
-3.93242277e-02 4.33589995e-01 -2.25513712e-01 -1.76019043e-01
5.49073279e-01 -5.83891943e-02 8.16914022e-01 -5.23958266e-01
9.73790407e-01 4.91748482e-01 -9.12741721e-02 -3.39435399e-01
-1.09083891e+00 -7.67849863e-01 -6.43051028e-01 -6.17038846e-01
1.22183752e+00 -1.06384778e+00 -8.13064337e-01 3.08785111e-01
-1.11018014e+00 -5.60491800e-01 -3.41393590e-01 4.34878141e-01
-7.38745928e-01 4.17448461e-01 -6.38402522e-01 -5.91868341e-01
-3.21620196e-01 -1.18635046e+00 1.36031425e+00 4.70452666e-01
2.69896358e-01 -1.16963840e+00 5.42752817e-02 2.77743667e-01
-1.48387745e-01 6.56870782e-01 3.97539943e-01 -3.73226851e-01
-1.34946382e+00 -2.41275012e-01 -4.09768283e-01 1.10473949e-02
1.13372602e-01 2.31743887e-01 -4.74628747e-01 -2.39563435e-01
-3.19263637e-01 1.20547779e-01 9.26560163e-01 7.75349081e-01
1.41929984e+00 -4.43757385e-01 -6.34056687e-01 9.83251989e-01
1.61026764e+00 1.58973679e-01 3.43150735e-01 1.37792408e-01
1.36310935e+00 3.85201007e-01 9.24362838e-01 4.19467598e-01
6.49582982e-01 7.70300806e-01 7.14215517e-01 -1.39467314e-01
-6.04065731e-02 -4.38254714e-01 2.23411411e-01 5.67163825e-01
-2.62431353e-01 -3.23531896e-01 -4.77916509e-01 6.88101828e-01
-2.43290162e+00 -1.09462523e+00 -1.12317719e-01 2.07571816e+00
4.57151920e-01 1.14078730e-01 3.64258975e-01 -5.81244051e-01
1.00170696e+00 2.29597896e-01 -6.59716368e-01 3.25101197e-01
1.37070090e-01 -1.67209208e-01 5.86613417e-01 5.08037448e-01
-1.26506662e+00 1.13992202e+00 5.39589977e+00 6.72959566e-01
-1.14018989e+00 3.43871489e-02 6.97934985e-01 -2.52814651e-01
-1.70176432e-01 1.67691838e-02 -7.48074532e-01 5.51863968e-01
4.41278249e-01 -4.38950986e-01 5.74784279e-01 8.80422890e-01
3.44052911e-01 3.14287007e-01 -1.22161233e+00 1.00981963e+00
-1.02075338e-01 -1.73272061e+00 6.26963601e-02 -1.27657251e-02
8.59003365e-01 8.92574787e-02 -9.03539136e-02 2.14499220e-01
3.08175027e-01 -6.05905771e-01 1.05705500e+00 7.55572855e-01
3.25366646e-01 -6.77675843e-01 4.07706618e-01 1.56752527e-01
-1.75420868e+00 -1.33284360e-01 -2.56138921e-01 2.51760185e-01
6.05415702e-01 3.83334666e-01 -6.64711416e-01 7.45170534e-01
5.52263498e-01 1.39090526e+00 -4.22798425e-01 1.41784644e+00
-1.44893661e-01 4.53565300e-01 -1.13118738e-01 3.01112056e-01
5.28454781e-01 -3.90643328e-01 6.57046258e-01 1.20811510e+00
2.60783046e-01 1.19874515e-01 8.72100413e-01 1.00746202e+00
-1.98877737e-01 -1.35245785e-01 -3.22556555e-01 -1.52279750e-01
3.27673286e-01 1.44024789e+00 -8.46293688e-01 -6.17692888e-01
-5.70201159e-01 1.01940536e+00 3.90760392e-01 5.64189613e-01
-1.24992490e+00 5.86012229e-02 4.74485934e-01 2.47837886e-01
6.51031435e-01 -2.49459520e-01 2.74562478e-01 -1.18668735e+00
2.47217834e-01 -4.94406551e-01 5.29528081e-01 -7.33783603e-01
-1.22457266e+00 4.83508855e-01 3.17644477e-02 -1.20710385e+00
-3.41185272e-01 -1.46870285e-01 -7.61877775e-01 6.08014643e-01
-1.52028525e+00 -1.30048895e+00 -6.80993319e-01 9.13257957e-01
7.97812819e-01 2.29567677e-01 -1.30714301e-03 4.55870628e-01
-8.33906293e-01 3.53652120e-01 -4.92291331e-01 4.10388619e-01
2.22319022e-01 -1.29997742e+00 3.86179119e-01 8.51200521e-01
2.26922289e-01 2.70962298e-01 4.29770440e-01 -8.92978787e-01
-1.40256381e+00 -1.68553925e+00 4.52427924e-01 -2.25590557e-01
7.83618212e-01 -3.02576810e-01 -9.74916756e-01 1.04515660e+00
1.42870218e-01 6.43491030e-01 6.00041673e-02 -3.26333493e-01
-1.33156121e-01 -1.38081819e-01 -8.41492891e-01 3.88175040e-01
1.52200198e+00 -3.65767479e-01 -2.42554978e-01 6.51042640e-01
1.06834841e+00 -8.21245193e-01 -7.70051360e-01 3.22186261e-01
2.88381904e-01 -6.49802148e-01 1.02982223e+00 -8.04966688e-01
3.08119327e-01 -4.74799603e-01 1.29221886e-01 -1.01188791e+00
-4.04687464e-01 -9.48457181e-01 -4.74462956e-01 1.20484173e+00
1.73934937e-01 -4.04495686e-01 1.04706407e+00 6.18103266e-01
-2.87171334e-01 -1.01385987e+00 -6.47036612e-01 -6.76733136e-01
-6.01445377e-01 -3.18251610e-01 6.70696259e-01 9.91128802e-01
-4.13742721e-01 8.19474906e-02 -4.08562511e-01 4.39590961e-01
8.48547697e-01 3.86405826e-01 8.98273766e-01 -9.90179777e-01
-6.61137521e-01 -2.94726998e-01 -7.84794569e-01 -1.37700319e+00
3.07730615e-01 -7.77320862e-01 1.15967147e-01 -1.60887468e+00
3.73248398e-01 -4.20261025e-01 -3.89904469e-01 3.76679152e-01
-6.05770648e-01 2.77498327e-02 3.26530427e-01 5.01966216e-02
-1.14615548e+00 5.36556125e-01 1.53067863e+00 -3.06584299e-01
-2.60774195e-01 7.89614841e-02 -3.04653436e-01 6.07613564e-01
2.86623657e-01 -4.83388454e-01 -7.24168837e-01 -4.20260072e-01
-5.79315878e-04 5.17319381e-01 6.72178805e-01 -8.50566983e-01
4.32987899e-01 -4.91987854e-01 1.48064762e-01 -6.90399885e-01
2.47530520e-01 -8.70029688e-01 2.96300381e-01 3.10788959e-01
-2.68795013e-01 2.75731320e-03 -1.17319413e-02 1.07057035e+00
-2.74757087e-01 1.31817400e-01 5.15646577e-01 -2.09183514e-01
-1.12293565e+00 1.28383589e+00 1.36422858e-01 1.88263327e-01
1.38342261e+00 -3.88567090e-01 -4.25219014e-02 -2.33448103e-01
-1.06811345e+00 6.75488353e-01 3.68530482e-01 4.61858302e-01
6.01045370e-01 -1.35949218e+00 -4.64419544e-01 -1.20176353e-01
-2.69376766e-02 3.91900569e-01 4.48587865e-01 1.05419350e+00
-3.27560008e-01 -1.22099882e-02 1.10217623e-01 -9.21479225e-01
-1.25943828e+00 8.20313394e-01 6.48968756e-01 -2.66638011e-01
-8.55171263e-01 9.71619487e-01 5.89887381e-01 4.41574231e-02
2.88197696e-01 -2.71825910e-01 -1.53682098e-01 -1.21784337e-01
2.42674485e-01 2.30509251e-01 -4.01292503e-01 -6.18721306e-01
-1.68573961e-01 5.84852517e-01 -1.62183180e-01 1.72721386e-01
1.34455061e+00 -2.63591588e-01 8.06862041e-02 2.53375798e-01
1.30912900e+00 -3.21541667e-01 -1.88027370e+00 -4.05514717e-01
-1.38451964e-01 -7.13730693e-01 -1.40468366e-02 -1.62515640e-01
-1.79222918e+00 5.31918287e-01 3.33554626e-01 2.49111652e-01
1.07166815e+00 5.13387136e-02 1.01984143e+00 5.84314875e-02
3.08487475e-01 -8.81012082e-01 1.82589829e-01 1.52730510e-01
5.28229713e-01 -1.09274507e+00 -9.89690945e-02 -7.72070944e-01
-4.70744967e-01 9.88669336e-01 8.11817050e-01 -3.33883435e-01
6.48092389e-01 -2.71009933e-02 -2.00048894e-01 -4.60741013e-01
-7.56744385e-01 -2.78539717e-01 5.64257562e-01 3.81450951e-01
1.31546512e-01 -7.84376189e-02 -1.68319613e-01 2.95729578e-01
2.06427023e-01 2.94365913e-01 1.52463242e-01 8.38718832e-01
-2.49288112e-01 -7.75277674e-01 9.55820456e-02 4.90181118e-01
-3.20987999e-01 9.77036133e-02 -4.36034799e-02 8.20110559e-01
1.80491000e-01 5.01331329e-01 2.15810359e-01 -2.65445232e-01
3.03252488e-01 -3.39680821e-01 4.38991189e-01 -6.06078148e-01
-4.64820892e-01 2.79294372e-01 -1.48773655e-01 -9.82128441e-01
-6.87989175e-01 -8.05969834e-01 -1.44683623e+00 -5.39128780e-02
-4.05048966e-01 5.34191169e-02 2.66425580e-01 1.10057199e+00
4.58810180e-01 8.69916081e-01 5.88400722e-01 -1.10979831e+00
-2.22917885e-01 -4.79001790e-01 -4.19459283e-01 6.09360814e-01
4.15372819e-01 -7.09346890e-01 -6.41166717e-02 3.57208192e-01] | [9.212193489074707, -0.08724682778120041] |
aa17b875-794e-4186-a494-5c45cb7e2bb6 | extraction-of-diagnostic-reasoning-relations | null | null | https://aclanthology.org/2022.acl-srw.33 | https://aclanthology.org/2022.acl-srw.33.pdf | Extraction of Diagnostic Reasoning Relations for Clinical Knowledge Graphs | Clinical knowledge graphs lack meaningful diagnostic relations (e.g. comorbidities, sign/symptoms), limiting their ability to represent real-world diagnostic processes. Previous methods in biomedical relation extraction have focused on concept relations, such as gene-disease and disease-drug, and largely ignored clinical processes. In this thesis, we leverage a clinical reasoning ontology and propose methods to extract such relations from a physician-facing point-of-care reference wiki and consumer health resource texts. Given the lack of data labeled with diagnostic relations, we also propose new methods of evaluating the correctness of extracted triples in the zero-shot setting. We describe a process for the intrinsic evaluation of new facts by triple confidence filtering and clinician manual review, as well extrinsic evaluation in the form of a differential diagnosis prediction task. | ['Vimig Socrates'] | null | null | null | null | acl-2022-5 | ['clinical-knowledge'] | ['miscellaneous'] | [ 2.20641375e-01 9.92479980e-01 -6.66400194e-01 -2.06161112e-01
-6.78167999e-01 -4.30291325e-01 4.77660894e-01 1.07943976e+00
1.38377538e-03 1.03030837e+00 5.80233812e-01 -8.09527814e-01
-8.08459878e-01 -9.57370281e-01 -2.74124652e-01 -1.52791187e-01
-8.87305215e-02 1.00354469e+00 1.49919137e-01 8.30831900e-02
-2.45495781e-01 2.71887004e-01 -1.14255261e+00 4.75391269e-01
7.45844126e-01 6.98698461e-01 -4.99524325e-01 3.99697095e-01
3.00495848e-02 1.27785063e+00 -3.79643172e-01 -8.28149199e-01
-2.76276439e-01 -6.22953117e-01 -1.08104873e+00 -2.24763498e-01
-1.78746954e-01 -3.46703678e-02 -2.66497523e-01 7.76498020e-01
5.26175797e-01 -1.87150061e-01 7.22795546e-01 -1.34466422e+00
-4.91865039e-01 9.53284860e-01 -1.61658730e-02 1.97637260e-01
8.69452894e-01 -1.24636836e-01 1.32432318e+00 -4.71247166e-01
1.43206501e+00 9.49488223e-01 8.25152695e-01 7.55298555e-01
-1.16868353e+00 -2.45172828e-01 -1.40518904e-01 1.01645917e-01
-1.27577019e+00 -4.52449977e-01 -1.17277481e-01 -6.91434920e-01
1.32859921e+00 5.38168490e-01 7.50549018e-01 1.03263402e+00
1.82666317e-01 2.82121003e-01 5.71677744e-01 -4.21010762e-01
1.95124745e-01 1.30856350e-01 3.12176526e-01 1.03629053e+00
8.82334590e-01 -1.24711595e-01 -4.50699031e-01 -7.47702122e-01
2.40391806e-01 3.59904431e-02 -2.76659369e-01 6.38031289e-02
-1.12821567e+00 4.65001762e-01 1.45098627e-01 2.59011000e-01
-2.65899718e-01 -1.78261310e-01 5.23316324e-01 5.82891377e-03
4.34501231e-01 6.60389721e-01 -7.85244167e-01 1.02243029e-01
-8.19298029e-01 1.83387116e-01 1.35338497e+00 1.09895599e+00
1.84785992e-01 -7.68069327e-01 -2.22561270e-01 4.79491800e-01
1.78825065e-01 -1.09700233e-01 5.36241829e-01 -7.32167184e-01
1.56741112e-01 9.93363082e-01 6.32247105e-02 -9.19649303e-01
-6.79446578e-01 -2.92707533e-01 -5.46831667e-01 -4.97640163e-01
2.22190663e-01 -2.05986798e-01 -9.80867326e-01 1.37926018e+00
5.27012587e-01 2.45878831e-01 3.82233828e-01 4.24106181e-01
1.30243373e+00 -2.79626042e-01 5.23039639e-01 -5.07439196e-01
1.82488549e+00 -4.43965495e-01 -1.01098514e+00 1.78499892e-01
1.25356090e+00 -3.54370356e-01 2.33534262e-01 1.94699988e-01
-1.04086816e+00 3.10220718e-01 -8.51675451e-01 -2.68765956e-01
-7.10958123e-01 -2.54579093e-02 9.22827184e-01 2.18507737e-01
-5.98415315e-01 7.97951579e-01 -9.68819261e-01 -7.04593658e-01
6.34874523e-01 3.15637171e-01 -5.07406533e-01 -6.57352135e-02
-1.39179122e+00 1.42407000e+00 4.91648942e-01 -3.75916630e-01
-5.07235408e-01 -1.15687835e+00 -9.00892437e-01 1.04334623e-01
8.89598370e-01 -1.45550501e+00 1.16668630e+00 -2.53778338e-01
-6.48046255e-01 1.17637610e+00 -1.52220041e-01 -4.97884005e-01
3.52993250e-01 5.73528521e-02 -9.89425361e-01 2.13155344e-01
1.75421461e-01 2.25919738e-01 6.61189333e-02 -7.19180584e-01
-8.42432916e-01 -5.33573389e-01 2.78211713e-01 8.32829997e-02
3.42120603e-02 -7.91986007e-03 -2.74538636e-01 -3.47854197e-01
1.48482069e-01 -7.50507772e-01 -4.57192183e-01 -8.83130636e-03
-7.84958184e-01 -1.07644282e-01 3.43416840e-01 -7.30021715e-01
1.65038753e+00 -1.68710542e+00 6.43566772e-02 3.71538043e-01
8.51927340e-01 -3.49297710e-02 6.06255770e-01 4.55524266e-01
-4.67883646e-01 6.41616940e-01 1.93389028e-03 1.55454531e-01
-2.63556451e-01 5.04972160e-01 1.17292203e-01 1.99630588e-01
3.83001357e-01 1.02507758e+00 -1.19456708e+00 -8.42106104e-01
-2.92681485e-01 4.00085419e-01 -4.87949342e-01 -3.02982062e-01
-3.17897290e-01 -5.82874902e-02 -5.98861396e-01 8.43488991e-01
-1.64838731e-01 -7.55439878e-01 7.89701998e-01 -5.84557891e-01
3.05398047e-01 7.94771671e-01 -8.15862119e-01 1.33584404e+00
-8.55320841e-02 1.58888437e-02 -1.31299838e-01 -5.03129125e-01
2.89739877e-01 8.19101512e-01 9.38907146e-01 -1.63305581e-01
9.08563733e-02 1.78922862e-01 2.16626197e-01 -7.80742645e-01
1.34379014e-01 -3.12035799e-01 2.20181406e-01 2.87933230e-01
2.33454630e-01 9.95542631e-02 4.99821156e-01 6.40556097e-01
1.70833063e+00 3.99416499e-03 9.59328294e-01 -2.00708359e-01
3.42548609e-01 5.27653635e-01 7.31246769e-01 2.24071264e-01
2.29781196e-01 2.97563493e-01 9.93229330e-01 -3.49417597e-01
-8.16298187e-01 -7.40771532e-01 -4.12948459e-01 3.84655565e-01
-2.47046560e-01 -1.25952685e+00 -3.27076703e-01 -8.16939235e-01
3.33986133e-01 5.97262323e-01 -8.45153987e-01 -2.48228058e-01
-6.22137785e-02 -9.36968923e-01 6.68002248e-01 3.89916271e-01
-3.03042203e-01 -4.88912612e-01 -6.55644119e-01 4.01141793e-01
-1.10730298e-01 -1.30393767e+00 3.85775864e-02 2.46519268e-01
-8.28297734e-01 -1.82673621e+00 -7.06484020e-02 -3.57354254e-01
7.31162250e-01 -6.37044728e-01 1.41358972e+00 3.59864563e-01
-7.99374819e-01 3.01972836e-01 -3.29905272e-01 -4.84622896e-01
-5.35075843e-01 -1.83956489e-01 -3.56421471e-02 -5.88138103e-01
6.92470789e-01 -3.11156541e-01 -5.20798028e-01 4.21775132e-02
-8.34825635e-01 1.37915730e-01 4.14347917e-01 8.61053586e-01
6.75544024e-01 -1.41114518e-01 3.96110415e-01 -1.52264977e+00
9.12014365e-01 -8.10697377e-01 -8.08313861e-02 6.79825604e-01
-1.18738770e+00 1.15521304e-01 3.11086532e-02 -1.23231769e-01
-8.78355682e-01 4.22969088e-02 -1.32935584e-01 1.23082913e-01
-7.48526007e-02 1.14140475e+00 -2.94003822e-02 5.04125357e-01
9.46052015e-01 -5.59511364e-01 -1.01024158e-01 -2.03727767e-01
3.22914928e-01 3.85808706e-01 3.87114227e-01 -5.79262614e-01
1.19421452e-01 3.52181286e-01 5.13795555e-01 -1.61621645e-01
-1.06124270e+00 -4.78156984e-01 -5.51341057e-01 3.25007409e-01
9.64860260e-01 -9.49060023e-01 -9.43481803e-01 -5.95863163e-01
-8.68953705e-01 1.14011526e-01 -7.50311077e-01 6.35278583e-01
-2.88485497e-01 1.22012228e-01 -8.88597965e-01 -4.80133116e-01
-4.25851196e-01 -8.73381138e-01 9.19489324e-01 -4.42085750e-02
-7.87624955e-01 -1.21673334e+00 7.66949654e-02 7.81498700e-02
-1.36837929e-01 4.28558916e-01 1.46570826e+00 -1.01065707e+00
-1.86448410e-01 -3.37082207e-01 -7.89052844e-02 -3.92122507e-01
4.19637322e-01 2.66057402e-01 -6.53195441e-01 3.19387674e-01
-4.91426855e-01 6.17942289e-02 4.16728079e-01 1.61931023e-01
7.75075018e-01 -4.48665887e-01 -1.06179285e+00 3.63575190e-01
1.34463131e+00 2.95265943e-01 6.07195377e-01 -3.56574357e-02
5.69953859e-01 5.93271375e-01 3.54847103e-01 3.91561866e-01
4.40950900e-01 4.70797330e-01 -1.17184438e-01 -1.45078167e-01
-1.68937564e-01 -2.18293309e-01 -4.72114980e-01 3.68572146e-01
-3.91451538e-01 -2.85701215e-01 -1.34695756e+00 7.12525725e-01
-1.89573300e+00 -7.92242050e-01 -3.46585780e-01 1.84684014e+00
1.44121397e+00 1.85395911e-01 -1.54900372e-01 1.98549312e-02
3.96978915e-01 -9.13791478e-01 -3.51994962e-01 -5.90991564e-02
-3.05938045e-03 2.85319567e-01 5.00062764e-01 4.23625141e-01
-5.82997143e-01 6.84773266e-01 6.70862484e+00 4.00655925e-01
-5.04261434e-01 2.91945606e-01 5.60678184e-01 -7.30124339e-02
-4.15188104e-01 1.51874691e-01 -8.08130980e-01 8.76562223e-02
1.06404281e+00 -5.72821677e-01 -1.87879041e-01 5.65429389e-01
2.24667881e-02 -1.65648118e-01 -1.53324568e+00 7.54524291e-01
-2.06439137e-01 -1.51350284e+00 -6.16077222e-02 2.15521425e-01
5.45849562e-01 -3.92167389e-01 -4.90539789e-01 -3.32081504e-02
8.95788133e-01 -9.81330335e-01 1.42027780e-01 8.60069811e-01
1.05615354e+00 -3.10229361e-02 8.66325736e-01 -1.17336102e-01
-8.75565529e-01 6.88844994e-02 1.16146229e-01 2.50336796e-01
1.81266859e-01 1.08511591e+00 -1.33781028e+00 1.11086822e+00
4.22094703e-01 9.37867045e-01 -5.75941145e-01 8.79824460e-01
-2.70069271e-01 4.87371624e-01 -2.47961223e-01 3.53486747e-01
-2.17047110e-01 1.74217820e-01 2.89882064e-01 1.19468713e+00
4.17859524e-01 5.85706353e-01 8.59569907e-02 6.76992238e-01
-1.08254895e-01 1.67278439e-01 -7.06119299e-01 -3.44764292e-01
3.35300088e-01 1.23589528e+00 -7.82942474e-01 -9.60512817e-01
-4.47640091e-01 3.85063410e-01 3.26152071e-02 2.38201171e-01
-6.07471585e-01 -7.33643696e-02 6.67724252e-01 4.18463945e-01
-2.47361481e-01 5.08327246e-01 -3.41159701e-01 -1.20580888e+00
-3.18577379e-01 -8.91953468e-01 9.29465830e-01 -7.40890384e-01
-1.36885095e+00 5.42401850e-01 5.75609580e-02 -1.03414261e+00
-4.88259077e-01 -5.35503805e-01 1.26540735e-02 6.15249455e-01
-1.02453399e+00 -1.08134663e+00 -1.00252599e-01 4.77297038e-01
-1.75364733e-01 1.01054735e-01 1.17182410e+00 6.52277648e-01
-5.87121904e-01 4.22970742e-01 -4.81211215e-01 7.54523873e-02
8.97341192e-01 -1.17422175e+00 1.58014998e-01 3.22873801e-01
-3.52756262e-01 9.98904467e-01 6.80476665e-01 -1.21959090e+00
-9.33178127e-01 -1.02697980e+00 1.37170362e+00 -7.42980182e-01
8.75819445e-01 1.33845001e-01 -8.47154617e-01 7.94164062e-01
-1.95579141e-01 1.34761229e-01 1.15605831e+00 4.10821497e-01
-3.97440493e-01 1.92187667e-01 -1.40890372e+00 5.84322453e-01
1.52595282e+00 -5.72244704e-01 -7.42739081e-01 7.66984940e-01
7.01601863e-01 -4.43241537e-01 -1.46029890e+00 4.94972467e-01
5.33719063e-01 -1.50239661e-01 9.85650599e-01 -1.42732561e+00
5.71401656e-01 -2.59649009e-01 1.01556763e-01 -1.04661453e+00
-2.19815642e-01 -4.26011175e-01 -5.01617432e-01 9.31034327e-01
1.03836274e+00 -5.95612347e-01 4.94892240e-01 1.06992984e+00
-2.43070032e-02 -9.37712252e-01 -6.18727803e-01 -3.06668878e-01
-4.76466537e-01 -6.79879934e-02 5.75727642e-01 1.52327359e+00
8.13255847e-01 6.02804422e-01 2.59422928e-01 1.01760603e-01
2.55277336e-01 -4.08421047e-02 -5.00884205e-02 -1.72790992e+00
-4.30774331e-01 -1.77292958e-01 -7.78597713e-01 -8.50666873e-03
-1.71988606e-01 -9.96384740e-01 -2.76953965e-01 -2.21601081e+00
4.73842114e-01 -5.83517611e-01 -1.56107545e-01 9.60715413e-01
-2.12964520e-01 2.93655898e-02 -4.64510500e-01 6.39787614e-02
-5.24411917e-01 -1.28201157e-01 9.19016898e-01 -2.09149122e-01
1.73452213e-01 -3.36600810e-01 -8.26772630e-01 8.28030586e-01
4.59898949e-01 -8.61070931e-01 -4.89340246e-01 -7.84925818e-02
8.79653811e-01 4.30728167e-01 2.37175509e-01 -7.15860069e-01
4.16330844e-01 -6.53789118e-02 1.48304582e-01 -1.68854833e-01
7.38659948e-02 -7.41197586e-01 7.06361890e-01 7.22185552e-01
-5.07236242e-01 1.31589800e-01 -4.25011255e-02 5.35735786e-01
-2.28893951e-01 -2.08062366e-01 8.79115611e-02 -4.61128086e-01
-2.81865060e-01 4.78036553e-02 -2.06521213e-01 3.11426669e-01
1.04222286e+00 -2.48742495e-02 -7.63027251e-01 -7.09573179e-02
-1.57613599e+00 1.99397147e-01 2.57446706e-01 2.22528324e-01
4.29103255e-01 -8.58850539e-01 -6.54254198e-01 -4.75764006e-01
2.46795624e-01 -2.07547024e-01 -4.22638841e-02 1.27949178e+00
-5.55011392e-01 4.44075584e-01 4.32276987e-02 -6.25193724e-03
-1.51989329e+00 8.98081839e-01 3.41051936e-01 -6.27584517e-01
-5.98039210e-01 6.16643071e-01 -8.24375153e-02 -4.98787574e-02
1.66742072e-01 -5.71644843e-01 -2.68844903e-01 3.69447619e-01
3.99843961e-01 5.62444963e-02 5.16088367e-01 -2.93508410e-01
-7.62380183e-01 -5.36049251e-03 -5.40832952e-02 3.53007540e-02
1.32274544e+00 1.79279283e-01 -3.74410748e-01 2.85147935e-01
7.03332782e-01 1.02335759e-01 -1.00444838e-01 -4.15786467e-02
4.48878944e-01 -2.41470076e-02 -3.54931921e-01 -1.25624466e+00
-7.63743877e-01 3.42652291e-01 1.24877594e-01 2.13128030e-01
7.67871737e-01 2.35881284e-01 2.48861730e-01 4.53179032e-01
1.82474509e-01 -8.60514224e-01 -6.12839341e-01 1.93586454e-01
5.05031645e-01 -1.00377798e+00 7.50373721e-01 -1.15883780e+00
-5.70091307e-01 1.03083181e+00 3.34490478e-01 6.26770318e-01
9.46463943e-01 7.03734398e-01 -2.19516590e-01 -7.62007952e-01
-1.36691880e+00 -4.16271776e-01 3.42810303e-01 4.34805036e-01
8.47129107e-01 3.56091768e-01 -7.98161626e-01 8.91762018e-01
1.00320816e-01 5.34599841e-01 5.04724920e-01 1.06096280e+00
1.62151128e-01 -1.24379134e+00 8.95165056e-02 1.05735219e+00
-9.78106022e-01 -5.13925254e-01 -6.70327127e-01 6.89535677e-01
5.63622892e-01 8.10471416e-01 -1.33761883e-01 -2.99475342e-01
4.14810717e-01 3.77329409e-01 5.60946882e-01 -9.31533992e-01
-7.80642450e-01 -1.08287334e-01 8.57620716e-01 -4.48106676e-01
-6.05509818e-01 -6.53399706e-01 -1.52836943e+00 5.04590711e-03
-4.58856970e-01 1.99569061e-01 3.30605596e-01 1.09298134e+00
7.58639753e-01 1.01329362e+00 -6.06829286e-01 4.30650353e-01
-1.12450458e-01 -6.99811161e-01 -3.34999859e-01 5.94748497e-01
-6.35716766e-02 -7.81337500e-01 1.44995466e-01 4.85672414e-01] | [8.475911140441895, 8.650739669799805] |
710ac821-d76e-40dc-9bd7-2c4dae2bc324 | rlogist-fast-observation-strategy-on-whole | 2212.01737 | null | https://arxiv.org/abs/2212.01737v2 | https://arxiv.org/pdf/2212.01737v2.pdf | RLogist: Fast Observation Strategy on Whole-slide Images with Deep Reinforcement Learning | Whole-slide images (WSI) in computational pathology have high resolution with gigapixel size, but are generally with sparse regions of interest, which leads to weak diagnostic relevance and data inefficiency for each area in the slide. Most of the existing methods rely on a multiple instance learning framework that requires densely sampling local patches at high magnification. The limitation is evident in the application stage as the heavy computation for extracting patch-level features is inevitable. In this paper, we develop RLogist, a benchmarking deep reinforcement learning (DRL) method for fast observation strategy on WSIs. Imitating the diagnostic logic of human pathologists, our RL agent learns how to find regions of observation value and obtain representative features across multiple resolution levels, without having to analyze each part of the WSI at the high magnification. We benchmark our method on two whole-slide level classification tasks, including detection of metastases in WSIs of lymph node sections, and subtyping of lung cancer. Experimental results demonstrate that RLogist achieves competitive classification performance compared to typical multiple instance learning algorithms, while having a significantly short observation path. In addition, the observation path given by RLogist provides good decision-making interpretability, and its ability of reading path navigation can potentially be used by pathologists for educational/assistive purposes. Our code is available at: \url{https://github.com/tencent-ailab/RLogist}. | ['Wei Yang', 'Qiang Fu', 'Xiao Han', 'Jian Cao', 'Deheng Ye', 'Jun Zhang', 'Boxuan Zhao'] | 2022-12-04 | null | null | null | null | ['multiple-instance-learning'] | ['methodology'] | [ 2.23680407e-01 3.38093102e-01 -6.05711341e-01 1.10889629e-01
-1.44950414e+00 -5.65416455e-01 1.57833453e-02 4.64930326e-01
-2.57232577e-01 7.93577850e-01 -2.73035854e-01 -6.55148566e-01
-4.23284948e-01 -8.65777850e-01 -5.95475495e-01 -1.32126021e+00
-1.20018005e-01 5.62716007e-01 3.20987552e-01 1.53891101e-01
8.08505416e-02 7.33320534e-01 -8.80393505e-01 4.76161093e-01
6.20222926e-01 7.63666928e-01 4.86247718e-01 1.07521665e+00
1.23150401e-01 8.27397585e-01 -5.41103363e-01 -1.92518590e-03
9.21658874e-02 -2.95901954e-01 -8.69922042e-01 3.96609344e-02
3.91034991e-01 -4.18157041e-01 -2.29079992e-01 1.00235939e+00
5.30147374e-01 -2.80381471e-01 7.35073507e-01 -1.02416432e+00
-1.84631422e-01 2.88589418e-01 -9.87199783e-01 5.64050436e-01
1.34666301e-02 4.34850842e-01 1.28212476e+00 -5.81239522e-01
9.10288811e-01 7.14360595e-01 5.85062921e-01 5.13050556e-01
-1.03798461e+00 -4.88882750e-01 -9.99973565e-02 2.76106030e-01
-1.36975098e+00 1.17055271e-02 4.04528111e-01 -2.51852989e-01
7.75854528e-01 4.28705692e-01 7.04747438e-01 7.42524862e-01
7.45887160e-01 1.05370641e+00 1.20220387e+00 -2.65580982e-01
2.06000835e-01 1.66821644e-01 3.41054611e-02 1.12468719e+00
1.29122883e-01 -6.63641617e-02 -1.72513410e-01 -2.00651824e-01
1.16855180e+00 6.39017940e-01 -2.96032697e-01 -4.93684113e-02
-1.30856001e+00 8.27964425e-01 8.64148498e-01 3.56807947e-01
-2.50900146e-02 2.09242716e-01 3.26364189e-01 2.87450910e-01
2.28242815e-01 3.38617951e-01 -2.54930645e-01 3.06930035e-01
-7.66116440e-01 -1.52802775e-02 3.85521948e-01 5.28096974e-01
6.52373254e-01 -6.96810842e-01 -9.92933810e-02 5.76586246e-01
7.01423064e-02 3.90583187e-01 7.26118982e-01 -7.78094590e-01
-1.22365497e-01 9.55503404e-01 -2.70233035e-01 -6.22834384e-01
-7.85291135e-01 -5.86063981e-01 -1.24945068e+00 5.09607136e-01
6.59287035e-01 3.37030590e-02 -7.85892725e-01 1.21098912e+00
6.77507579e-01 3.07838500e-01 -6.72867568e-03 7.35067964e-01
7.92779446e-01 4.51181740e-01 -8.99746269e-02 -8.84847045e-02
1.81727004e+00 -1.04976475e+00 -3.93092245e-01 -1.84236839e-02
1.20543504e+00 -4.07535434e-01 1.32960737e+00 3.08975607e-01
-9.25173342e-01 -1.63285375e-01 -7.51664102e-01 -1.28415510e-01
-3.79875004e-01 3.61019224e-01 6.09944880e-01 1.61334388e-02
-1.18893874e+00 3.27635318e-01 -1.03317583e+00 -3.90830398e-01
8.18490982e-01 4.62008387e-01 -4.86576676e-01 -7.33594447e-02
-7.89253771e-01 5.27440250e-01 -1.48971811e-01 -6.70579374e-02
-1.02023220e+00 -1.13412750e+00 -6.14414632e-01 4.00375202e-02
4.24294889e-01 -5.96412778e-01 1.19963503e+00 -7.96694577e-01
-1.24985373e+00 1.12683046e+00 -2.50904024e-01 -1.24664754e-01
5.44444919e-01 4.81156737e-01 -5.27596250e-02 5.83923042e-01
2.09525108e-01 7.00925052e-01 4.89479035e-01 -8.45828950e-01
-8.73257518e-01 -5.41962624e-01 2.36371160e-01 3.09383720e-01
-1.61814302e-01 -4.90984112e-01 -3.88625354e-01 -3.95101637e-01
-1.54628024e-01 -8.64873528e-01 -5.03642261e-01 5.18904150e-01
-5.53696334e-01 -3.21513891e-01 6.99694395e-01 -4.44842190e-01
9.23777759e-01 -2.29998016e+00 -3.29746567e-02 3.35939467e-01
4.25058782e-01 -1.04200557e-01 -3.73662896e-02 1.02879643e-01
7.71070048e-02 1.48234054e-01 -4.39478159e-02 1.40835181e-01
-3.26062739e-01 1.71405926e-01 1.25974104e-01 7.06853330e-01
1.71629041e-01 1.19373918e+00 -1.04402125e+00 -9.71196830e-01
3.66299897e-02 1.11118197e-01 -4.63567346e-01 2.47508556e-01
-8.79882872e-02 3.90377313e-01 -6.27498507e-01 1.04190016e+00
3.24408770e-01 -1.18981826e+00 3.68007272e-01 -1.22039460e-01
3.96289974e-01 -1.04796082e-01 -9.91700172e-01 1.47214651e+00
-3.24079663e-01 4.16954637e-01 9.95024145e-02 -9.65588272e-01
2.69083738e-01 3.48880082e-01 4.73230839e-01 -5.29666126e-01
-1.09011665e-01 2.40819857e-01 8.67737159e-02 -6.63096905e-01
-2.77354449e-01 -1.99580729e-01 2.15532809e-01 7.29749560e-01
-2.03941166e-01 -1.18183400e-02 2.49965057e-01 2.29487211e-01
1.58541942e+00 -4.50044006e-01 8.03363144e-01 -2.92664796e-01
4.96882588e-01 2.82870770e-01 4.90925282e-01 7.09658444e-01
-2.42017627e-01 3.95573884e-01 8.25550258e-01 -5.65320134e-01
-7.94625819e-01 -1.21139073e+00 -4.42072511e-01 8.31342816e-01
1.96694002e-01 2.01215655e-01 -3.45429808e-01 -1.01560128e+00
6.25803545e-02 -7.87147358e-02 -1.06505311e+00 2.89187700e-01
-5.51354527e-01 -8.89374375e-01 3.93309623e-01 6.03139281e-01
1.27138957e-01 -1.08000767e+00 -7.82921195e-01 7.47569576e-02
-5.05147427e-02 -6.20754957e-01 -2.24299908e-01 2.15719253e-01
-1.00464618e+00 -1.37310040e+00 -7.92921185e-01 -9.83662128e-01
1.34363389e+00 5.22409916e-01 1.00040126e+00 6.42633379e-01
-1.00509453e+00 1.85736254e-01 -1.99778214e-01 -4.71146703e-01
-3.83771986e-01 2.15384036e-01 -4.95043516e-01 -3.35770816e-01
3.46250296e-01 -3.43347460e-01 -1.13007081e+00 3.67066652e-01
-1.03091645e+00 8.33388269e-02 1.02691472e+00 1.32386601e+00
1.16315138e+00 2.04540733e-02 5.58858097e-01 -1.30090642e+00
1.42523855e-01 -5.12443244e-01 -4.29307520e-01 3.77028018e-01
-1.34966269e-01 -3.15660894e-01 7.76530564e-01 -2.00022593e-01
-6.60866976e-01 -1.95957065e-01 -2.69639604e-02 -1.60254717e-01
-2.49413282e-01 5.05295098e-01 3.36296618e-01 -1.50374085e-01
8.18949759e-01 1.49932370e-01 5.62161148e-01 2.23001674e-01
-3.16030383e-01 4.21999156e-01 6.48167655e-02 -1.48171216e-01
5.38822055e-01 8.99888277e-01 2.26789534e-01 -5.32671332e-01
-8.79438639e-01 -7.78480113e-01 -2.20445916e-01 -1.44584224e-01
4.52375531e-01 -7.92274058e-01 -9.80763316e-01 1.93564206e-01
-4.82053697e-01 -8.70509505e-01 -4.95193392e-01 3.61172616e-01
-4.42635030e-01 5.39959222e-02 -1.19208550e+00 -1.03038013e-01
-4.49650109e-01 -1.28218019e+00 1.48236537e+00 3.28218311e-01
-1.92659825e-01 -1.35202909e+00 1.70410722e-01 3.26941699e-01
2.26370126e-01 3.64926338e-01 1.09407997e+00 -4.87238824e-01
-9.18539941e-01 -3.91063124e-01 -2.41182983e-01 -2.96785414e-01
3.17736030e-01 3.01940978e-01 -9.00947750e-01 -6.17868245e-01
-4.41102147e-01 -5.01559615e-01 6.50130332e-01 7.30986118e-01
1.73643458e+00 -2.74067879e-01 -8.61575723e-01 7.05682814e-01
1.65419161e+00 -9.31324139e-02 3.46847445e-01 3.74458671e-01
3.49993974e-01 5.26322722e-01 6.78614020e-01 3.44730705e-01
1.49504453e-01 2.40130290e-01 5.05599439e-01 -7.43669868e-01
-9.41712186e-02 1.19850069e-01 -6.50530457e-02 4.06056255e-01
1.01764500e-01 -2.99681485e-01 -9.43793118e-01 6.57763541e-01
-1.64850509e+00 -7.80534327e-01 3.22980881e-01 1.66784859e+00
9.62681651e-01 -1.51002005e-01 -8.37259665e-02 1.69030100e-01
3.09220672e-01 -5.70771955e-02 -8.78259122e-01 -1.93316042e-02
2.72953123e-01 1.12079456e-01 4.10808563e-01 5.67692935e-01
-8.01108778e-01 3.87045979e-01 5.65299463e+00 1.30652213e+00
-1.24011517e+00 4.58646528e-02 1.27907908e+00 -3.73399317e-01
-2.05233827e-01 -4.87034231e-01 -9.27206099e-01 1.27382874e-01
4.58966047e-01 -8.92461464e-02 2.34019198e-02 6.89416826e-01
2.24203095e-01 -3.43872309e-01 -1.19462287e+00 6.47999167e-01
-2.96781898e-01 -1.70602369e+00 2.40384229e-02 4.04259771e-01
6.66849375e-01 -7.06081185e-03 1.78802952e-01 1.06770702e-01
4.66159493e-01 -1.27383411e+00 -2.61620134e-01 3.26465994e-01
1.08712268e+00 -5.58754802e-01 9.89621162e-01 4.02103007e-01
-9.24212039e-01 -4.71955240e-02 -4.76863295e-01 4.23642427e-01
-4.44445521e-01 7.08015025e-01 -1.42328703e+00 3.88478428e-01
6.67518497e-01 7.04352379e-01 -5.82294047e-01 6.76069558e-01
-7.87811503e-02 5.59044957e-01 -3.73393416e-01 -3.23085397e-01
3.02360058e-01 2.61538893e-01 1.56142026e-01 1.27259886e+00
4.28975701e-01 1.51083365e-01 1.05497740e-01 4.12230790e-01
4.29108031e-02 2.04580650e-01 -3.80860388e-01 2.52938420e-01
3.52594435e-01 1.93356395e+00 -1.02012408e+00 -3.49378169e-01
-3.56131345e-01 6.24195457e-01 6.07806504e-01 3.77395511e-01
-7.29332089e-01 -2.41805568e-01 2.45568603e-01 4.51998979e-01
9.78075936e-02 5.02391577e-01 -2.78421313e-01 -8.99488688e-01
-1.33655250e-01 -9.52233315e-01 8.82838309e-01 -4.20222163e-01
-1.30912530e+00 3.83606821e-01 -3.71777266e-01 -1.56903327e+00
-1.23055227e-01 -9.31477845e-01 -9.50791299e-01 3.85125905e-01
-1.85416949e+00 -1.20677555e+00 -5.10050476e-01 6.65748775e-01
5.34267068e-01 1.97585579e-02 9.57824111e-01 -3.28007117e-02
-5.95453203e-01 9.06144559e-01 2.43734241e-01 1.86041936e-01
7.39032209e-01 -1.65762746e+00 -2.25860313e-01 1.91329792e-01
-3.19352239e-01 5.33379197e-01 2.46963143e-01 -2.51817554e-01
-1.33393037e+00 -1.18712819e+00 4.63622332e-01 -2.94632822e-01
8.21690083e-01 -3.02229226e-02 -8.82407129e-01 7.29255855e-01
1.32347196e-01 6.07511222e-01 1.29731047e+00 -8.67266506e-02
1.89657509e-01 -2.79083520e-01 -1.38922596e+00 8.13030839e-01
5.48859239e-01 -2.83984929e-01 8.97427425e-02 7.78948128e-01
1.77026004e-01 -5.93703330e-01 -1.18273163e+00 2.23476633e-01
5.08882284e-01 -7.49438941e-01 9.10877585e-01 -2.77475238e-01
6.53455257e-01 -3.32691729e-01 3.82518202e-01 -1.23996437e+00
-6.16947770e-01 -4.71419580e-02 4.03535701e-02 5.21498740e-01
7.94355989e-01 -7.63556898e-01 1.16984725e+00 4.55958247e-02
3.80050726e-02 -1.56641948e+00 -9.41893578e-01 -2.81324953e-01
1.71177357e-01 3.16831023e-01 4.12170053e-01 7.87390530e-01
3.11172962e-01 -6.56149909e-02 3.39791834e-01 4.33900625e-01
7.68671632e-01 4.24092263e-01 7.12317228e-01 -7.94979095e-01
-8.02211165e-01 -5.74491739e-01 -4.18598533e-01 -7.61214077e-01
-2.58327454e-01 -1.11030269e+00 -2.86727130e-01 -1.52384984e+00
5.44976592e-01 -4.29251194e-01 -5.85083187e-01 6.86401367e-01
-4.58805174e-01 4.18681175e-01 -4.29043710e-01 4.46594954e-01
-6.91869974e-01 -1.19724259e-01 1.90881717e+00 -4.50093806e-01
1.79540724e-01 -7.12211952e-02 -7.04440653e-01 7.66718864e-01
8.50223243e-01 -3.43729079e-01 -3.34185570e-01 -1.38616219e-01
2.78354585e-01 3.53301913e-01 6.54771805e-01 -9.31272149e-01
4.97575641e-01 -3.58599633e-01 7.93032467e-01 -5.55649519e-01
5.87399416e-02 -6.20009780e-01 -5.72943455e-03 7.87816823e-01
-4.37355936e-01 -1.92957908e-01 3.42425518e-02 5.10370553e-01
-2.69128144e-01 -1.57964990e-01 9.66298342e-01 -4.65573817e-01
-3.52021188e-01 6.27728343e-01 -4.50866491e-01 -2.56669104e-01
1.34748614e+00 -3.98482770e-01 -5.96100509e-01 1.53186753e-01
-5.99120677e-01 4.91651446e-01 4.53464895e-01 -4.09057140e-01
5.57133555e-01 -1.03052270e+00 -8.08963239e-01 1.90512821e-01
2.55988181e-01 5.56314170e-01 7.50476360e-01 1.19018447e+00
-8.74883354e-01 1.34818330e-01 -1.29781544e-01 -8.99741411e-01
-1.29920578e+00 4.25841510e-01 6.70498431e-01 -1.15314460e+00
-7.62155235e-01 1.18472445e+00 8.00773442e-01 -5.32034159e-01
1.81741431e-01 -3.87924999e-01 -1.25082389e-01 -1.49302453e-01
7.68267751e-01 1.52282611e-01 5.10887764e-02 2.32062116e-01
-2.94731557e-01 6.01620436e-01 -7.93022990e-01 3.64403665e-01
1.15249264e+00 1.86069831e-01 -6.40783459e-02 3.51684242e-01
1.18661964e+00 -9.39807594e-02 -1.22811711e+00 -3.85980874e-01
-4.62113619e-01 -2.99100876e-01 -3.87718454e-02 -8.13258708e-01
-1.10017681e+00 9.02619243e-01 6.03208601e-01 2.45777220e-01
1.15178108e+00 1.99035823e-01 6.37474775e-01 3.05756599e-01
1.62660211e-01 -7.79823899e-01 2.09340408e-01 -9.76453200e-02
6.94718421e-01 -1.57786322e+00 2.84143776e-01 -3.34426373e-01
-4.21970636e-01 1.26154923e+00 7.44070530e-01 -3.14485282e-01
6.84894919e-01 7.40651965e-01 3.96558434e-01 -4.85840470e-01
-1.15310001e+00 2.19496086e-01 -1.20570123e-01 2.61447906e-01
4.21791166e-01 2.71712899e-01 5.63567691e-02 1.96190476e-01
1.34317338e-01 -1.58606358e-02 4.39445108e-01 1.01323855e+00
-5.97170770e-01 -8.30218017e-01 -2.23960489e-01 9.09180939e-01
-6.44254148e-01 7.66872093e-02 -9.35844630e-02 9.17829812e-01
-2.48171821e-01 2.79791117e-01 1.92039162e-01 2.31824964e-01
-1.71854384e-02 -5.87654769e-01 3.03263307e-01 -5.20564914e-01
-5.99515676e-01 2.24308714e-01 -5.55803299e-01 -5.55023015e-01
-1.74986884e-01 -6.30214453e-01 -1.56469166e+00 -3.02839458e-01
-1.83746293e-01 1.53580397e-01 1.11415103e-01 6.89125836e-01
1.89044461e-01 7.19107389e-01 7.71099567e-01 -4.37441379e-01
-4.58079904e-01 -7.80973315e-01 -8.85934770e-01 1.39206707e-01
7.11822629e-01 -3.96594822e-01 -5.11369824e-01 -6.27491856e-03] | [15.113755226135254, -2.9529941082000732] |
c3c1d2fc-ff0d-4786-b241-190ce0c99751 | management-and-detection-system-for-medical | 2211.02351 | null | https://arxiv.org/abs/2211.02351v1 | https://arxiv.org/pdf/2211.02351v1.pdf | Management and Detection System for Medical Surgical Equipment | Retained surgical bodies (RSB) are any foreign bodies left inside the patient after a medical procedure. RSB is often caused by human mistakes or miscommunication between medical staff during the procedure. Infection, medical complications, and even death are possible consequences of RSB, and it is a significant risk for patients, hospitals, and surgical staff. In this paper. we describe the engineering process we have done to explore the design space, define a feasible solution, and simulate, verify, and validate a state-of-the-art Cyber-Physical System that can significantly decrease the incidence of RSB and thus increase patients' survivability rate. This system might save patients' suffering and lives and reduce medical staff negligence lawsuits while improving the hospital's reputation. The paper illustrates each step of the process with examples and describes the chosen solution in detail. | ['Michael Winokur', 'Natan Levy', 'Alexandra Hadar'] | 2022-11-04 | null | null | null | null | ['medical-procedure'] | ['medical'] | [-2.33550612e-02 6.28360748e-01 -8.56183283e-03 5.44688106e-01
1.47428885e-01 -4.92747962e-01 1.45403847e-01 2.95174122e-01
-3.48446608e-01 1.02843368e+00 2.29376391e-01 -6.10540569e-01
-5.02768874e-01 -5.35762370e-01 -3.74354750e-01 -6.95002258e-01
2.05547005e-01 -1.50295906e-02 1.69076130e-01 -4.66043651e-02
4.37829465e-01 9.20032740e-01 -7.91270971e-01 -6.70199320e-02
7.73000181e-01 8.66326690e-01 1.71890944e-01 5.84074967e-02
1.44190609e-01 1.23208833e+00 -8.47218335e-01 -1.81411460e-01
3.19831580e-01 -4.49485421e-01 -9.89747643e-01 -1.80679839e-04
-8.89207482e-01 -3.43515694e-01 -1.82224706e-01 9.64465082e-01
2.34629184e-01 -2.31654555e-01 3.42464179e-01 -1.10043859e+00
-7.01306062e-03 3.65127355e-01 -9.02979448e-02 -4.74777341e-01
3.78331065e-01 2.16768861e-01 -3.91414434e-01 -2.70449370e-01
6.74415410e-01 8.24241400e-01 9.06757593e-01 1.02358115e+00
-7.85280168e-01 -5.11005640e-01 -6.56609476e-01 -6.18068576e-01
-1.24263084e+00 -1.09242514e-01 3.58582199e-01 -6.57489717e-01
7.67546892e-01 7.55180120e-01 1.19467402e+00 7.60182977e-01
1.29042161e+00 3.49261284e-01 8.78996730e-01 -4.56222832e-01
2.90336281e-01 4.60860640e-01 -3.70873744e-03 7.62012541e-01
1.20251215e+00 3.90548170e-01 1.38412155e-02 -3.54474962e-01
1.07406652e+00 5.00008941e-01 -6.84612513e-01 -7.00719178e-01
-1.12451017e+00 2.02360660e-01 2.97246039e-01 5.17343581e-01
-5.14337182e-01 1.93638906e-01 2.92102098e-01 -2.64345765e-01
-5.80840170e-01 8.81543577e-01 -3.81328970e-01 -1.85039744e-01
-1.13865644e-01 -3.94476175e-01 1.06401658e+00 7.58004308e-01
-1.18768059e-01 -3.16095680e-01 -8.46623927e-02 1.78030550e-01
2.06865445e-01 2.45333418e-01 5.70214748e-01 -6.91913724e-01
-2.23014042e-01 6.44449174e-01 5.21826267e-01 -9.24286604e-01
-4.82503116e-01 -3.62101912e-01 -8.71225119e-01 2.38182962e-01
-1.68067649e-01 -8.20368379e-02 -7.20159829e-01 9.52050984e-01
9.11058784e-02 -3.10849816e-01 1.01774700e-01 6.12377107e-01
7.64780223e-01 3.41750652e-01 -2.71783359e-02 -6.36456966e-01
1.13538909e+00 -7.31463015e-01 -1.30453706e+00 -3.73856984e-02
9.94184911e-01 -6.30462110e-01 5.57004631e-01 6.03919566e-01
-1.08101201e+00 2.36935690e-02 -8.91883850e-01 6.06766522e-01
1.24868732e-02 -1.25734791e-01 5.29585481e-01 9.73317802e-01
-6.00332499e-01 8.94261241e-01 -1.06930125e+00 -5.08476377e-01
2.31474251e-01 3.66066515e-01 -2.28748709e-01 1.86631262e-01
-7.11009145e-01 1.56450784e+00 3.71959098e-02 3.43550265e-01
-8.10338080e-01 -6.12379730e-01 -7.68107951e-01 -1.35238931e-01
4.16771710e-01 -8.98265421e-01 1.14932072e+00 -2.21584886e-01
-1.31520224e+00 9.68325198e-01 4.93777841e-01 -2.09008977e-01
5.40378451e-01 -2.85484403e-01 -4.83573169e-01 -1.72227815e-01
-1.19092964e-01 -2.83659637e-01 -4.42263521e-02 -1.43709195e+00
-2.54561305e-01 1.14158615e-02 -2.92771965e-01 -3.68348539e-01
-8.02877992e-02 -5.22098877e-03 7.18786791e-02 -4.81878608e-01
7.79129378e-03 -1.22460258e+00 -8.72690380e-01 -8.93865600e-02
-4.19971466e-01 3.77905965e-01 4.01010215e-01 -4.35363948e-01
1.37828255e+00 -2.45293951e+00 -2.56607085e-01 2.71375626e-01
-6.91005737e-02 5.60245991e-01 4.58332598e-01 8.79444718e-01
9.91778672e-02 3.73076260e-01 1.06588408e-01 1.86581269e-01
-4.26910043e-01 3.22907537e-01 6.79380074e-02 8.48089099e-01
-4.30567175e-01 5.52089334e-01 -8.46448898e-01 -4.93122131e-01
5.31830788e-01 3.63608092e-01 -2.41820723e-01 3.20175171e-01
7.30341613e-01 6.55957818e-01 -5.79269826e-01 6.69817448e-01
5.58003843e-01 -1.25251502e-01 1.13998808e-01 -1.62595734e-01
-7.92570859e-02 2.51652390e-01 -7.18209684e-01 1.34852934e+00
-3.95631850e-01 1.31620780e-01 2.66173124e-01 -4.28641528e-01
8.23026478e-01 6.27648354e-01 6.52587831e-01 -3.86094093e-01
7.08122671e-01 4.51094836e-01 -1.61355540e-01 -1.04836667e+00
1.41087338e-01 -8.27943325e-01 -3.55258793e-01 7.24054947e-02
-4.61439937e-01 -3.32013577e-01 -4.53445435e-01 1.37463212e-01
1.31185102e+00 -2.43497863e-01 1.07524729e+00 -7.10057199e-01
4.41238672e-01 2.80568391e-01 7.91851342e-01 6.35996819e-01
-4.27915961e-01 -1.73008405e-02 2.56804198e-01 -4.71692413e-01
-3.33191633e-01 -9.75686252e-01 4.11239006e-02 -3.30806285e-01
1.04951954e+00 -9.30636749e-02 -5.95311105e-01 -4.65459049e-01
2.14433908e-01 1.01353371e+00 -4.60347682e-01 -8.19582641e-01
-3.71374249e-01 -2.05019757e-01 3.20314348e-01 4.49734122e-01
8.48992467e-02 -1.02944243e+00 -1.33349526e+00 3.72340202e-01
-1.77769437e-01 -7.23093867e-01 -1.90494850e-01 9.64967161e-02
-1.01880932e+00 -1.25195920e+00 -3.30902904e-01 -6.21648312e-01
1.06857300e+00 1.72002062e-01 5.46646535e-01 5.68569660e-01
-8.99858952e-01 3.72428805e-01 -3.44700903e-01 -7.36827433e-01
-9.61829066e-01 -7.96550870e-01 2.08888829e-01 -6.97819531e-01
-2.08844408e-01 1.31661206e-01 -8.03016424e-01 4.85659719e-01
-5.77217162e-01 3.61901596e-02 5.46507120e-01 7.12059379e-01
8.54352787e-02 2.78112352e-01 3.21544200e-01 -7.95678675e-01
6.35608673e-01 -9.94304791e-02 -3.29277813e-01 1.91183940e-01
-5.58767021e-01 -2.34848157e-01 2.00562090e-01 -1.90609589e-01
-8.98522973e-01 1.94829721e-02 3.33234757e-01 -2.08241522e-01
-2.12047353e-01 1.58966035e-01 4.11474109e-02 -3.08028519e-01
4.56722230e-01 -2.76297897e-01 5.63222051e-01 -3.66927266e-01
-5.39737940e-01 6.42778337e-01 4.40028220e-01 3.26426327e-02
4.37090248e-01 4.89135563e-01 2.75378793e-01 -3.29148561e-01
-6.37899458e-01 -5.63506484e-01 2.94670202e-02 -4.96754736e-01
6.72615588e-01 -5.80478251e-01 -1.18153787e+00 2.78716832e-01
-9.88893688e-01 -2.21114963e-01 -3.49553108e-01 9.53143299e-01
-3.92230242e-01 1.31612018e-01 -7.98599601e-01 -1.00387549e+00
-3.12054545e-01 -1.20156491e+00 4.36631650e-01 2.09277228e-01
-5.75862467e-01 -7.82553494e-01 -7.28121921e-02 2.99658537e-01
3.86754274e-01 9.21383917e-01 7.16437697e-01 -4.75214384e-02
-5.91375351e-01 -6.23448014e-01 4.03631091e-01 3.81427467e-01
7.54856527e-01 -1.72798663e-01 -1.47960395e-01 -7.71913826e-01
6.26844883e-01 3.48420322e-01 9.35060903e-02 4.63119894e-01
9.35911000e-01 -5.80891132e-01 -1.18567145e+00 4.12408590e-01
1.55560410e+00 8.29064846e-01 1.02374303e+00 3.39918643e-01
1.45348653e-01 5.90744019e-01 1.26442552e+00 5.28746486e-01
-5.35443842e-01 -1.32904435e-02 7.02580810e-01 -3.01483184e-01
4.42820974e-02 -8.02507717e-03 1.79413229e-01 7.71370530e-01
-2.55437911e-01 -2.95051217e-01 -1.00139010e+00 6.49443984e-01
-1.52722239e+00 -6.14028513e-01 -3.77471209e-01 2.22855139e+00
5.00719070e-01 -6.97575882e-02 -5.14207661e-01 1.93608686e-01
8.40312719e-01 -8.21405232e-01 1.81472316e-01 -4.75071341e-01
7.21814334e-01 -5.79848625e-02 1.04308033e+00 1.52506307e-01
-5.12509465e-01 3.95826876e-01 7.11483860e+00 2.44351879e-01
-1.05900216e+00 -2.37025619e-01 1.54824883e-01 -2.38598660e-01
1.12441108e-01 1.67227566e-01 -3.05234849e-01 3.20553035e-01
4.44535881e-01 -3.16426516e-01 1.66911222e-02 7.62917161e-01
1.18267387e-01 -4.99147713e-01 -9.37918663e-01 8.99251044e-01
1.25754206e-02 -1.48906422e+00 -3.67550433e-01 1.99102908e-01
3.44334394e-01 -7.55297184e-01 -5.10979950e-01 -1.70174375e-01
6.88008070e-02 -9.46062446e-01 7.12159097e-01 7.81567454e-01
6.32804930e-01 -6.91466749e-01 1.12825012e+00 5.79176307e-01
-5.08311331e-01 -3.12566310e-01 1.19412765e-01 -2.44999245e-01
5.90683043e-01 5.20685017e-01 -9.43454802e-01 3.48055035e-01
9.10634696e-01 -6.38293549e-02 -1.11183319e-02 1.31856024e+00
-2.48268262e-01 1.32539555e-01 -6.17064983e-02 -5.46619058e-01
-1.65600702e-01 -1.41492905e-02 5.71632743e-01 6.97347522e-01
4.95000541e-01 5.61552465e-01 -3.01389486e-01 4.93630111e-01
4.59043175e-01 -7.53552765e-02 -6.52818620e-01 1.70169070e-01
3.57228994e-01 8.62502873e-01 -8.31348479e-01 -1.10810935e-01
5.26530184e-02 8.35656345e-01 -5.44231653e-01 -2.32482210e-01
-9.75710034e-01 -5.73690414e-01 4.53631312e-01 6.04587674e-01
-3.76618475e-01 1.43644080e-01 -5.28362870e-01 -5.54215968e-01
-1.30788803e-01 -4.34303999e-01 3.92939672e-02 -7.19591618e-01
-7.48626709e-01 3.56805772e-01 -3.03193748e-01 -1.71995616e+00
1.09501019e-01 -6.26185656e-01 -2.99087286e-01 6.14296377e-01
-5.94508886e-01 -5.89050651e-01 -5.16349196e-01 3.37127864e-01
-2.94386715e-01 1.13385193e-01 1.08558428e+00 -7.36903995e-02
-3.90840977e-01 1.65579543e-01 -7.05142273e-03 -1.59106612e-01
7.18443573e-01 -3.11902791e-01 -5.54969311e-01 5.74406087e-01
-1.23899138e+00 1.12314153e+00 9.17096198e-01 -1.02917588e+00
-1.52038157e+00 -7.15204358e-01 6.65731490e-01 -2.27438658e-01
4.40702051e-01 1.96746141e-02 -3.27448428e-01 5.07008016e-01
1.16397552e-01 -3.35965425e-01 1.01340032e+00 -3.21960270e-01
5.10090649e-01 3.28333639e-02 -1.66146243e+00 8.78038406e-01
9.47986305e-01 -1.38071060e-01 -8.14943671e-01 1.73574805e-01
6.64575398e-01 -6.44678295e-01 -9.60363269e-01 8.48040879e-01
5.77569902e-01 -8.75517905e-01 8.41423273e-01 -2.77632445e-01
2.10408792e-01 -2.70950288e-01 7.30047047e-01 -1.20316458e+00
-4.14283127e-01 -8.63379300e-01 5.54743528e-01 7.56910563e-01
2.76872128e-01 -8.58237088e-01 2.67308563e-01 1.19052136e+00
-3.12061638e-01 -6.26989901e-01 -8.01115513e-01 -1.07913017e+00
-3.56629550e-01 3.41130435e-01 5.60865819e-01 1.04872143e+00
1.04447520e+00 -3.14766765e-01 -4.49194908e-01 1.86683983e-01
3.56560320e-01 -1.58600911e-01 6.81137443e-01 -9.51221883e-01
-2.34621644e-01 -1.04189880e-01 -5.37482262e-01 -5.61057068e-02
-6.60350859e-01 -3.44769388e-01 3.20973366e-01 -2.10281420e+00
1.71933681e-01 -5.87605238e-01 -2.29528055e-01 4.79675651e-01
1.34029850e-01 -4.35633421e-01 3.29987973e-01 2.32686535e-01
9.33453720e-03 5.23640625e-02 1.58976865e+00 2.94090450e-01
-3.47904861e-01 -1.97887421e-04 -6.32530212e-01 8.49267006e-01
8.22923660e-01 -7.68090248e-01 -9.30576101e-02 -2.75629410e-03
1.21242940e-01 8.63292813e-01 3.01754206e-01 -9.66400743e-01
3.28962386e-01 -6.18512213e-01 6.24461286e-02 -4.35782373e-01
1.19249575e-01 -1.59748292e+00 9.67091858e-01 1.77738476e+00
1.85201973e-01 -5.12244105e-01 4.20252651e-01 3.02522123e-01
9.89389718e-02 -5.63228071e-01 1.05071366e+00 -8.23108703e-02
3.33125740e-02 -4.24378306e-01 -1.00104737e+00 -8.60642910e-01
1.81766689e+00 -4.42720026e-01 -4.54973161e-01 1.88761815e-01
-8.17663610e-01 6.71698079e-02 7.85789907e-01 1.68555066e-01
5.71191370e-01 -9.57730889e-01 -7.70903751e-02 1.27811179e-01
-1.37668261e-02 -6.76765665e-02 4.39670861e-01 1.14629984e+00
-1.26939881e+00 3.64659429e-01 -3.93763930e-01 -1.19665913e-01
-1.49508524e+00 1.10761213e+00 5.82850516e-01 -1.70524299e-01
-7.98264444e-01 6.91822112e-01 1.19476140e-01 -6.15412295e-02
1.99515328e-01 -4.69591677e-01 1.75987393e-01 -6.65986419e-01
2.68857747e-01 4.58486170e-01 -1.31919786e-01 -8.26498568e-02
-8.52897048e-01 2.08482683e-01 2.00072035e-01 1.71635985e-01
9.48441803e-01 -1.81119025e-01 -3.84965926e-01 2.27256641e-01
7.26756632e-01 2.71075547e-01 -5.53013682e-01 5.09451926e-01
-3.60739201e-01 -7.33314991e-01 -2.12149397e-01 -1.03303158e+00
-9.29869294e-01 1.99614972e-01 3.44156832e-01 1.49035901e-01
1.07914722e+00 1.20616518e-01 7.03437507e-01 2.19729200e-01
1.04990041e+00 -9.78198767e-01 2.10810438e-01 -1.47838950e-01
1.02497923e+00 -4.82735693e-01 6.47687539e-02 -9.98480499e-01
-6.32113338e-01 6.73114717e-01 4.09608424e-01 -2.33213887e-01
1.02218473e+00 7.93547690e-01 9.64443162e-02 -2.59186149e-01
-3.00120384e-01 5.34892976e-01 -3.19406271e-01 3.56550395e-01
2.98690736e-01 4.57026809e-01 -1.06099749e+00 8.70033026e-01
-9.52680930e-02 5.95760286e-01 1.02934706e+00 1.66490912e+00
-2.91404694e-01 -8.38696539e-01 -7.63808489e-01 4.11798149e-01
-6.45842016e-01 2.65010923e-01 -3.32665473e-01 8.20484400e-01
1.36609852e-01 1.10064101e+00 -3.08804691e-01 -2.74037898e-01
7.25331426e-01 -3.75214338e-01 4.74948049e-01 -3.63733709e-01
-9.45806503e-01 1.23296037e-01 3.35859805e-01 -8.89512360e-01
-2.77308226e-01 -3.65396470e-01 -1.65333116e+00 -4.15978491e-01
-7.21188247e-01 2.44373038e-01 7.66448438e-01 5.93331814e-01
3.04376900e-01 1.04804385e+00 4.27693158e-01 -4.09960635e-02
-3.48865479e-01 -6.44499838e-01 -9.42754924e-01 3.39539379e-01
2.44631782e-01 -5.11122882e-01 -5.26049256e-01 -1.11901119e-01] | [13.839810371398926, -2.955263137817383] |
d4f682fc-2ba0-4854-8a3a-586cf76b1dfb | influence-of-multiple-sequence-alignment | 1812.04162 | null | http://arxiv.org/abs/1812.04162v2 | http://arxiv.org/pdf/1812.04162v2.pdf | Influence of Multiple Sequence Alignment Depth on Potts Statistical Models of Protein Covariation | Potts statistical models have become a popular and promising way to analyze
mutational covariation in protein Multiple Sequence Alignments (MSAs) in order
to understand protein structure, function and fitness. But the statistical
limitations of these models, which can have millions of parameters and are fit
to MSAs of only thousands or hundreds of effective sequences using a procedure
known as inverse Ising inference, are incompletely understood. In this work we
predict how model quality degrades as a function of the number of sequences
$N$, sequence length $L$, amino-acid alphabet size $q$, and the degree of
conservation of the MSA, in different applications of the Potts models: In
"fitness" predictions of individual protein sequences, in predictions of the
effects of single-point mutations, in "double mutant cycle" predictions of
epistasis, and in 3-d contact prediction in protein structure. We show how as
MSA depth $N$ decreases an "overfitting" effect occurs such that sequences in
the training MSA have overestimated fitness, and we predict the magnitude of
this effect and discuss how regularization can help correct for it, use a
regularization procedure motivated by statistical analysis of the effects of
finite sampling. We find that as $N$ decreases the quality of point-mutation
effect predictions degrade least, fitness and epistasis predictions degrade
more rapidly, and contact predictions are most affected. However, overfitting
becomes negligible for MSA depths of more than a few thousand effective
sequences, as often used in practice, and regularization becomes less
necessary. We discuss the implications of these results for users of Potts
covariation analysis. | [] | 2019-01-19 | null | null | null | null | ['multiple-sequence-alignment'] | ['medical'] | [ 6.27693534e-01 -1.55479521e-01 -3.15934569e-02 -2.68881500e-01
-4.88719404e-01 -5.20038903e-01 1.43590048e-01 4.97012615e-01
-5.56917131e-01 1.15745032e+00 -4.73595224e-02 -7.74581432e-01
-4.95681390e-02 -4.29841906e-01 -1.23940206e+00 -8.90887260e-01
-7.15504959e-02 6.21394634e-01 4.49207485e-01 -3.83647591e-01
5.14352918e-01 2.41211846e-01 -1.29453444e+00 9.45715681e-02
1.11947215e+00 1.94730714e-01 4.19847876e-01 6.67890191e-01
6.83878362e-02 1.22684620e-01 -4.39577192e-01 -1.84515744e-01
-2.31904928e-02 -6.29223645e-01 -7.40365148e-01 -2.40818217e-01
2.02690810e-01 2.16097206e-01 1.03781961e-01 8.74141693e-01
3.93949032e-01 -4.42306660e-02 8.26529026e-01 -6.15053177e-01
-1.33950934e-01 3.14674228e-01 -6.42683625e-01 1.28092647e-01
2.23525971e-01 6.85734868e-01 9.87694919e-01 -3.50160122e-01
9.04434800e-01 1.11338615e+00 8.63304973e-01 2.49232322e-01
-1.71518385e+00 -2.88138062e-01 -2.19517145e-02 -8.58062580e-02
-1.07068574e+00 -8.40020180e-02 1.55399576e-01 -5.36705375e-01
1.51912618e+00 2.57424533e-01 6.42081559e-01 6.56443000e-01
6.18087888e-01 1.75459698e-01 5.60238719e-01 -4.68892395e-01
2.75527388e-01 -4.59870160e-01 1.95884258e-01 7.49991238e-01
1.94042996e-01 1.34039400e-02 -5.11651814e-01 -9.72484410e-01
4.60825861e-01 -2.03743413e-01 -2.42310241e-01 -3.23812634e-01
-1.00154138e+00 8.20760071e-01 -3.83452065e-02 3.81442085e-02
-3.37184966e-01 9.51047689e-02 3.16146106e-01 8.48986283e-02
3.39211404e-01 7.43722796e-01 -1.07259274e+00 -4.47630018e-01
-6.69970512e-01 7.06642687e-01 6.16296172e-01 3.81403118e-01
7.52957523e-01 -2.62925595e-01 7.40361631e-01 9.69107449e-01
-8.58216286e-02 5.37040353e-01 4.34566349e-01 -9.07328010e-01
1.82932124e-01 5.40682316e-01 4.49509442e-01 -5.37506700e-01
-4.07462537e-01 -1.52978212e-01 -4.21949089e-01 2.51152754e-01
8.11131299e-01 -5.46122417e-02 -7.88220942e-01 1.92671359e+00
7.41578788e-02 -1.42377898e-01 -1.79811686e-01 5.18977761e-01
-7.94535652e-02 6.67322159e-01 3.54200542e-01 -7.68436134e-01
1.23315132e+00 -3.01528603e-01 -2.23906904e-01 -2.73370802e-01
9.30862248e-01 -8.84756267e-01 1.18413842e+00 3.76565516e-01
-1.11699641e+00 -1.89458504e-01 -1.00745845e+00 2.46347338e-01
-3.10841687e-02 -4.26939189e-01 5.15198708e-01 4.54244226e-01
-6.33296907e-01 1.17694736e+00 -1.03138328e+00 -3.74992311e-01
1.76245064e-01 5.92499733e-01 -1.30015776e-01 1.19566701e-01
-9.89066958e-01 1.13505626e+00 6.28928319e-02 -3.39105427e-01
-1.57456785e-01 -8.83901298e-01 -4.85005230e-01 4.62956494e-03
5.61223272e-03 -5.54961801e-01 7.91240633e-01 -1.07365060e+00
-9.10149097e-01 6.99871957e-01 -6.78536177e-01 -4.93110657e-01
1.25980183e-01 -3.56561616e-02 2.95527019e-02 -4.30234820e-01
-1.15749240e-01 4.62956250e-01 1.92715302e-01 -6.70819104e-01
-7.65913054e-02 -6.61299050e-01 -3.77478302e-01 1.94632515e-01
5.22529364e-01 7.50155821e-02 2.36046940e-01 -3.04940790e-01
2.34171942e-01 -1.19699621e+00 -7.37033367e-01 -3.04879129e-01
-4.45191599e-02 1.67007446e-01 2.87172556e-01 -6.59607232e-01
9.57968831e-01 -1.88637614e+00 2.94128835e-01 4.40899909e-01
2.36046650e-02 2.24619582e-01 -1.99648947e-01 7.24084914e-01
-2.61065871e-01 1.68599337e-01 -6.69341743e-01 7.11673737e-01
-3.54519904e-01 1.86604783e-01 7.63477981e-02 3.84258419e-01
9.72059593e-02 6.02802157e-01 -6.44480824e-01 -6.54730126e-02
-3.39305736e-02 4.39277828e-01 -7.66461730e-01 -1.43412083e-01
-6.62324011e-01 2.55017221e-01 -2.65785456e-01 1.47819132e-01
6.06375813e-01 -6.24503195e-01 7.93565929e-01 3.80407162e-02
-2.77473450e-01 6.25439763e-01 -6.78981185e-01 1.07365704e+00
2.56304741e-02 3.13145876e-01 -3.82249415e-01 -8.71684909e-01
8.69594216e-01 -6.03064299e-02 4.51140702e-01 -1.93642288e-01
-1.83376104e-01 3.74062926e-01 8.78817081e-01 -1.38646483e-01
7.05319494e-02 -3.64748299e-01 2.58589655e-01 4.11577612e-01
-3.47302645e-01 -4.82387468e-02 2.56438106e-01 3.68868485e-02
1.42040348e+00 2.28161097e-01 4.18381631e-01 -3.56699020e-01
1.77580416e-01 4.09892648e-01 8.04255843e-01 5.02723515e-01
-2.18766872e-02 3.87578189e-01 9.73050773e-01 -6.50696576e-01
-1.52528787e+00 -7.36339927e-01 -3.58734190e-01 1.06264174e+00
6.72821552e-02 -5.15679061e-01 -9.17925656e-01 -2.42073730e-01
1.13651551e-01 6.25063896e-01 -2.11378187e-01 -4.01510417e-01
-8.12163651e-01 -1.64611685e+00 4.03407097e-01 1.27067000e-01
-1.40704513e-01 -1.18821430e+00 -6.33977413e-01 4.69743818e-01
-1.99919507e-01 -6.97005928e-01 -3.47430736e-01 5.56040406e-01
-1.07976270e+00 -1.30046809e+00 -4.66978163e-01 -5.25985599e-01
5.75208724e-01 -1.63867325e-01 1.01466429e+00 3.43164474e-01
-6.02263749e-01 -3.85680228e-01 -2.19618920e-02 -1.45006761e-01
-8.41287971e-01 -1.60751626e-01 1.99485868e-01 -7.75740147e-01
6.38447583e-01 -9.21510696e-01 -5.74542046e-01 4.35700715e-01
-7.72790134e-01 -2.21839473e-01 5.21969199e-01 1.13514769e+00
5.70906997e-01 -3.16158175e-01 3.55879068e-01 -1.00213170e+00
5.11411905e-01 -2.12561458e-01 -7.27694869e-01 4.04984146e-01
-6.69753850e-01 4.75556672e-01 6.02451265e-01 -5.16719162e-01
-6.67043865e-01 2.29771331e-01 -3.28995407e-01 1.23100325e-01
5.58538456e-03 5.12611568e-01 -4.63894531e-02 -1.56098410e-01
9.11144435e-01 4.12986368e-01 3.52684766e-01 -4.47801352e-01
-1.95460111e-01 2.85201550e-01 -1.04751304e-01 -6.01128221e-01
1.31075025e-01 -9.02038440e-02 2.14437097e-01 -1.11934531e+00
-7.47445077e-02 -1.48216158e-01 -4.91449207e-01 2.32333511e-01
5.09565771e-01 -4.33086962e-01 -1.24901009e+00 3.11627567e-01
-9.39900577e-01 -6.30306959e-01 1.36270300e-01 4.97474819e-01
-9.23290253e-01 8.47348392e-01 -6.10561967e-01 -6.63020909e-01
2.08974490e-03 -1.61894393e+00 7.32355595e-01 -5.41353822e-02
-5.79188466e-01 -6.08748078e-01 2.51197487e-01 3.25641692e-01
1.01994492e-01 4.54441160e-02 1.80936790e+00 -4.42906469e-01
-4.01806742e-01 9.32571739e-02 1.98485956e-01 6.03205264e-02
1.05851650e-01 3.07771951e-01 -1.36273488e-01 -1.80067681e-02
-2.68144935e-01 -1.90041512e-01 7.46106505e-01 7.50368416e-01
8.36454690e-01 -1.46627933e-01 -5.42168021e-01 2.44888350e-01
1.40952778e+00 3.74346286e-01 7.11124539e-01 2.32291311e-01
2.71386564e-01 6.43884122e-01 6.78703725e-01 3.17187637e-01
-2.36174777e-01 7.07521081e-01 1.40152335e-01 2.17742682e-01
4.63412464e-01 -1.22388482e-01 3.11147273e-01 5.57971597e-01
-3.20573419e-01 -1.83900803e-01 -1.18809009e+00 1.47450477e-01
-1.81550598e+00 -9.17347133e-01 -2.90687323e-01 2.48397970e+00
1.09033704e+00 3.14780891e-01 4.63553816e-01 -2.41893768e-01
6.11622512e-01 -6.97035640e-02 -1.00420189e+00 -6.85275257e-01
-3.51947963e-01 2.49721318e-01 8.65045071e-01 8.56384635e-01
-4.38223362e-01 8.81220937e-01 7.30772114e+00 7.06849873e-01
-8.54297459e-01 -3.71630162e-01 8.71549308e-01 1.08908881e-02
-3.36643040e-01 4.47132587e-01 -7.03258932e-01 6.70216858e-01
1.14559782e+00 1.99905053e-01 4.19975370e-01 5.19357681e-01
4.24018860e-01 -3.83806407e-01 -8.12017262e-01 5.22670925e-01
-5.59107184e-01 -1.47308731e+00 -1.19885847e-01 3.45017523e-01
4.50880855e-01 2.16399193e-01 -9.60995182e-02 -2.89230973e-01
5.29612243e-01 -9.95312810e-01 -3.96821573e-02 3.86730105e-01
3.56539905e-01 -9.61322486e-01 6.11214042e-01 6.44038975e-01
-5.17325342e-01 2.49671608e-01 -7.15895116e-01 -2.19120234e-01
2.26307794e-01 6.14423215e-01 -7.29635060e-01 -1.54683545e-01
2.30120584e-01 2.08688825e-01 -1.03032947e-01 6.64020896e-01
2.98617214e-01 7.67071486e-01 -6.11419261e-01 -2.00629190e-01
1.08458899e-01 -6.17073894e-01 4.67425585e-01 7.93440402e-01
1.42186387e-02 4.17967647e-01 -1.87698245e-01 6.45776391e-01
3.26316297e-01 1.11529067e-01 -1.42727688e-01 -2.81520903e-01
3.68758678e-01 4.10949320e-01 -5.92301667e-01 -1.19241476e-01
-2.17109591e-01 7.85873294e-01 3.18555504e-01 4.27255809e-01
-7.16698945e-01 -1.26200706e-01 1.13356054e+00 5.19185424e-01
2.64504611e-01 -1.07022025e-01 -6.94000646e-02 -9.02678490e-01
2.75800820e-03 -1.29424584e+00 -3.36912498e-02 -6.40784144e-01
-1.05136430e+00 8.20170194e-02 -2.86118448e-01 -5.92846990e-01
-4.58971113e-01 -7.12963104e-01 -3.19772720e-01 9.69691455e-01
-7.79362977e-01 -4.37062413e-01 7.33595729e-01 -2.16429874e-01
4.42896217e-01 6.85582161e-02 8.86932373e-01 -2.98307184e-02
-3.76070887e-01 4.43570852e-01 6.19127154e-01 -5.36971092e-01
6.04539037e-01 -8.90125096e-01 8.11564267e-01 2.97739208e-01
-1.87459603e-01 9.50791717e-01 1.37705517e+00 -1.12137473e+00
-1.12864375e+00 -4.70135927e-01 9.78272796e-01 -1.89328939e-01
5.75689256e-01 -1.24578811e-01 -1.24275529e+00 5.45244992e-01
-2.95576781e-01 -3.51831496e-01 8.75310898e-01 3.96403939e-01
-3.54665697e-01 3.56968254e-01 -1.08793974e+00 7.28725910e-01
1.03568685e+00 -2.36045599e-01 -2.12238312e-01 5.13508677e-01
5.88421524e-01 2.59715281e-02 -7.79382825e-01 6.28824055e-01
8.75253379e-01 -1.05602622e+00 1.03303397e+00 -9.97412860e-01
2.47505382e-01 -1.59255907e-01 -2.12546512e-01 -1.10865402e+00
-2.62464046e-01 -5.10730386e-01 5.09256661e-01 3.75915527e-01
9.60759938e-01 -7.01620460e-01 9.89085495e-01 5.47355294e-01
1.15933297e-02 -9.40250695e-01 -8.08614552e-01 -6.04386032e-01
2.72117168e-01 4.83511835e-02 2.66719878e-01 7.34072685e-01
3.20937335e-01 2.62744844e-01 -2.72118181e-01 -3.60269397e-02
5.15412927e-01 -2.01060981e-01 5.32397628e-01 -1.04947090e+00
-7.32024014e-01 -3.18468004e-01 -3.74058723e-01 -1.11790848e+00
-1.38593361e-01 -2.93835431e-01 1.17565140e-01 -7.53272057e-01
3.86805087e-01 -4.49880600e-01 6.38453364e-02 1.57970622e-01
-4.44287151e-01 -2.06722304e-01 -2.03688413e-01 1.97119996e-01
-1.46436810e-01 1.17120638e-01 7.74307311e-01 2.28191927e-01
-2.36667275e-01 4.84170802e-02 -1.66894138e-01 9.07887638e-01
7.45781422e-01 -4.94029343e-01 -1.09614432e-01 1.08936489e-01
5.87170720e-01 3.70225340e-01 3.53177972e-02 -5.96408069e-01
-3.58775705e-01 -5.32806993e-01 3.63511622e-01 -4.87107843e-01
4.14564461e-01 -4.14356172e-01 5.19989669e-01 8.39987576e-01
-5.33629537e-01 1.37829632e-01 2.10515738e-01 5.29708028e-01
3.05837959e-01 -3.08108538e-01 9.04072583e-01 -2.55189121e-01
1.08603761e-01 -9.00878683e-02 -7.51227498e-01 -9.51848179e-02
5.95686615e-01 -1.82309121e-01 -1.18454911e-01 -2.24581033e-01
-7.21901059e-01 -1.32011056e-01 9.46409881e-01 -1.55772358e-01
3.29425871e-01 -5.44229984e-01 -4.87149715e-01 1.29066318e-01
-1.05765983e-01 -5.10258317e-01 2.85331786e-01 8.13403726e-01
-8.36500645e-01 4.28031772e-01 -1.01285979e-01 -5.59754491e-01
-1.69782758e+00 2.87185043e-01 3.02641779e-01 -3.00593376e-01
-1.37609750e-01 7.17757523e-01 2.63976872e-01 -4.77516085e-01
-2.30360702e-01 -3.76489833e-02 3.18864733e-01 -5.40295959e-01
2.09794015e-01 2.24592716e-01 -1.19329765e-01 -6.39007926e-01
-3.53565991e-01 7.41700768e-01 -4.37781155e-01 8.78240317e-02
1.36687028e+00 9.75915045e-02 -2.66578317e-01 4.48348045e-01
8.30387712e-01 -7.24816089e-03 -1.32497382e+00 2.04997212e-01
5.19612953e-02 -1.16508402e-01 -5.69209099e-01 -8.43448818e-01
-1.65862679e-01 7.33818352e-01 5.04872382e-01 -1.18237495e-01
6.61047637e-01 4.08952907e-02 8.15278709e-01 5.87103128e-01
3.32278669e-01 -7.88032353e-01 -3.41376811e-01 5.65304279e-01
3.62235129e-01 -9.54471231e-01 2.38356560e-01 -4.69533503e-01
-4.32095140e-01 8.41014504e-01 3.41219038e-01 -4.95913252e-02
2.90356368e-01 1.32756904e-01 -1.21298626e-01 -1.15947388e-01
-9.29775119e-01 1.40472561e-01 -7.63627142e-02 3.86344731e-01
6.52208984e-01 1.15088761e-01 -8.40321660e-01 1.97904214e-01
-1.98913172e-01 -2.41152436e-01 3.40264589e-01 9.68842030e-01
-9.51490760e-01 -1.61084294e+00 -1.99342251e-01 5.76905608e-01
-6.01099789e-01 -4.90998656e-01 -6.26931071e-01 6.16173863e-01
-3.76541279e-02 5.25744438e-01 -1.24245081e-02 -1.01821147e-01
8.36780220e-02 4.57663596e-01 6.73779368e-01 -2.99823254e-01
-5.06146550e-01 2.63100445e-01 2.67795742e-01 -2.77657181e-01
-2.28184626e-01 -7.28404641e-01 -1.46096194e+00 -6.54305339e-01
-6.60010874e-01 4.25523430e-01 6.16775990e-01 1.08447123e+00
5.48449576e-01 4.97889658e-03 2.06867665e-01 -3.45960766e-01
-7.24621236e-01 -8.73024642e-01 -4.98998940e-01 1.41950458e-01
1.74938604e-01 -3.17004859e-01 -5.42093277e-01 7.46000633e-02] | [4.775681972503662, 5.36406135559082] |
f85ca0f8-39af-41f7-82dc-5e269db9581b | learning-to-estimate-3d-human-pose-from-point | 2212.12910 | null | https://arxiv.org/abs/2212.12910v1 | https://arxiv.org/pdf/2212.12910v1.pdf | Learning to Estimate 3D Human Pose from Point Cloud | 3D pose estimation is a challenging problem in computer vision. Most of the existing neural-network-based approaches address color or depth images through convolution networks (CNNs). In this paper, we study the task of 3D human pose estimation from depth images. Different from the existing CNN-based human pose estimation method, we propose a deep human pose network for 3D pose estimation by taking the point cloud data as input data to model the surface of complex human structures. We first cast the 3D human pose estimation from 2D depth images to 3D point clouds and directly predict the 3D joint position. Our experiments on two public datasets show that our approach achieves higher accuracy than previous state-of-art methods. The reported results on both ITOP and EVAL datasets demonstrate the effectiveness of our method on the targeted tasks. | ['Abdulmotaleb El Saddik', 'Haiwei Dong', 'Yufan Zhou'] | 2022-12-25 | null | null | null | null | ['3d-pose-estimation', '3d-human-pose-estimation'] | ['computer-vision', 'computer-vision'] | [-4.03488696e-01 -4.24389291e-04 5.31455278e-02 -3.23752642e-01
-4.02394235e-01 -1.84738085e-01 2.88661957e-01 -4.88639235e-01
-8.11825752e-01 2.33599558e-01 1.53932303e-01 2.55463332e-01
5.17346382e-01 -5.24851680e-01 -8.49276543e-01 -6.31797537e-02
1.36157917e-02 1.19614935e+00 2.34158665e-01 -3.43523026e-01
8.13864172e-02 8.13647032e-01 -1.20356500e+00 -1.03903688e-01
3.45047772e-01 1.26885843e+00 -2.42915764e-01 7.05093324e-01
1.89705566e-01 1.76327586e-01 -7.49765813e-01 -3.69095236e-01
6.02005661e-01 1.87068582e-01 -8.05361032e-01 2.61419147e-01
6.54558063e-01 -8.03571343e-01 -4.79401886e-01 7.41162539e-01
8.50686312e-01 -5.24829440e-02 6.91679180e-01 -1.39320087e+00
-1.97815806e-01 -3.51434529e-01 -7.60917723e-01 -2.53392369e-01
8.28962326e-01 1.64458737e-01 5.14831722e-01 -1.09630513e+00
6.42306030e-01 1.76887250e+00 1.00220490e+00 7.38298893e-01
-7.28203595e-01 -7.09434032e-01 1.64636467e-02 -5.41709810e-02
-1.36834145e+00 1.98760122e-01 9.86241817e-01 -6.55172765e-01
9.32049453e-01 -1.04965471e-01 1.23288810e+00 1.35310519e+00
2.17045501e-01 1.20832241e+00 8.55853498e-01 -2.87389278e-01
-1.46299854e-01 -6.14884675e-01 -2.74869293e-01 9.64376986e-01
3.01997870e-01 2.28539377e-01 -5.65361619e-01 -1.93266533e-02
1.46060288e+00 1.54438064e-01 -1.75467897e-02 -8.77215624e-01
-1.39184475e+00 6.21993363e-01 9.06762362e-01 -3.54572117e-01
-4.55346465e-01 6.70358598e-01 2.92833447e-01 -2.58391082e-01
7.11903870e-01 2.34500930e-01 -7.04075456e-01 -6.04861565e-02
-5.12596071e-01 8.58677626e-01 6.44520938e-01 9.56997097e-01
3.51007223e-01 -4.28369254e-01 -2.05191951e-02 4.62675959e-01
3.91340137e-01 6.96662962e-01 8.61552916e-03 -1.03759527e+00
6.54611647e-01 8.92105579e-01 4.83022839e-01 -1.11825144e+00
-9.82251763e-01 -2.69212872e-01 -6.12037182e-01 3.85925651e-01
6.40703082e-01 -1.79658309e-01 -1.13713205e+00 1.34019518e+00
6.73534393e-01 -3.04338813e-01 -4.81439501e-01 1.40891075e+00
1.05565870e+00 3.32070321e-01 -1.48975343e-01 7.48937368e-01
1.32980847e+00 -1.07705069e+00 -4.41179037e-01 -4.93473530e-01
2.69414246e-01 -3.74584913e-01 7.67329276e-01 5.28737664e-01
-1.00574720e+00 -7.43268549e-01 -9.33611512e-01 -4.46644902e-01
-2.76578277e-01 5.78302681e-01 7.23137617e-01 3.84328306e-01
-6.98379338e-01 4.13863271e-01 -1.13724589e+00 -4.64251012e-01
4.37932134e-01 5.49215853e-01 -7.01771498e-01 9.65106636e-02
-1.07411981e+00 1.08374977e+00 3.61459374e-01 5.69060445e-01
-7.58396864e-01 -2.54828185e-01 -1.04782486e+00 -3.51388961e-01
4.18014079e-01 -1.17183924e+00 1.45910895e+00 -2.75125414e-01
-1.39151573e+00 1.38614643e+00 6.57964423e-02 -2.97161758e-01
1.17672324e+00 -1.18872011e+00 3.99049073e-01 1.94090232e-01
-5.80014661e-02 1.13865411e+00 5.84567666e-01 -1.16445231e+00
-3.86835217e-01 -8.13111901e-01 6.79100901e-02 4.24427003e-01
-2.79089101e-02 -2.82971442e-01 -1.15415406e+00 -3.91618818e-01
4.71259385e-01 -1.23005724e+00 -3.48101228e-01 7.88641274e-01
-8.66300881e-01 -3.89481127e-01 5.66084206e-01 -6.43766522e-01
5.89612544e-01 -1.49597931e+00 5.87648809e-01 1.36627063e-01
5.37687063e-01 1.61686733e-01 1.97574005e-01 1.12034969e-01
2.04334095e-01 -3.44732612e-01 3.14023942e-02 -7.23683655e-01
3.11793178e-01 1.27865374e-01 3.82118642e-01 6.81157112e-01
1.59242555e-01 1.38773513e+00 -6.45144820e-01 -6.13566399e-01
4.44666654e-01 8.12270045e-01 -5.32627344e-01 5.43559492e-01
-1.15770772e-01 6.13005042e-01 -4.16371644e-01 8.36317778e-01
6.42599046e-01 -3.93183470e-01 -1.04763143e-01 -4.08393800e-01
2.89077908e-01 2.36103241e-03 -9.44211066e-01 2.31284952e+00
-1.42850712e-01 3.33069831e-01 -1.97722554e-01 -5.28550208e-01
9.20606434e-01 2.66450047e-01 6.53258085e-01 -3.87326539e-01
5.26939452e-01 4.46354076e-02 -4.51085836e-01 -3.36883575e-01
2.87629843e-01 7.55943954e-02 -4.99990523e-01 1.38983294e-01
-8.28367751e-03 -2.69573152e-01 -4.16930288e-01 -2.60841757e-01
7.52965331e-01 6.88728094e-01 3.09840709e-01 2.19309881e-01
1.87105343e-01 6.12943061e-02 3.43474388e-01 3.08137864e-01
-5.40035307e-01 9.93595660e-01 4.64155108e-01 -1.00759053e+00
-1.19272470e+00 -1.39366102e+00 2.33030915e-01 6.99006915e-01
1.68590859e-01 -3.08925331e-01 -8.87066066e-01 -8.08975875e-01
4.25361127e-01 -2.30264336e-01 -8.92919481e-01 4.20152321e-02
-8.56782079e-01 -2.56436825e-01 5.81143677e-01 1.01777256e+00
9.03246641e-01 -8.57455194e-01 -1.23660135e+00 -1.30340695e-01
-4.44121003e-01 -1.36471355e+00 -4.83825952e-01 -1.62213966e-01
-7.76492059e-01 -1.29169679e+00 -1.21161699e+00 -6.70047283e-01
6.50850475e-01 -1.97410315e-01 1.17492640e+00 -1.25270069e-01
-4.32944208e-01 4.91660357e-01 -1.34054840e-01 -7.19405472e-01
4.01054233e-01 2.17168316e-01 3.53417009e-01 -4.64017868e-01
6.35133564e-01 -1.93173230e-01 -9.53832686e-01 2.80940711e-01
-1.65908948e-01 1.67347431e-01 5.43939471e-01 6.12848699e-01
6.36161506e-01 -3.89906794e-01 -1.79816529e-01 -5.29117823e-01
3.77884090e-01 2.17281416e-01 -5.96282303e-01 -8.04099515e-02
-5.94750904e-02 -1.29809231e-01 -2.11193219e-01 -4.60792303e-01
-7.55611777e-01 7.95696735e-01 -3.45357090e-01 -7.67828941e-01
-3.96878302e-01 1.47902668e-01 -7.39443451e-02 -7.73847997e-02
6.71551585e-01 -2.00887829e-01 3.26646566e-01 -6.44382477e-01
1.15843736e-01 2.67497092e-01 8.32466722e-01 -6.47906959e-01
9.39450145e-01 7.20082998e-01 3.00313920e-01 -3.19656223e-01
-1.02907979e+00 -4.03056502e-01 -1.43316364e+00 -4.87222016e-01
1.32128930e+00 -1.15670073e+00 -1.39324415e+00 7.58083165e-01
-1.62824011e+00 -1.95629388e-01 2.83454686e-01 4.05328512e-01
-8.50785673e-01 2.17998251e-01 -7.99199224e-01 -8.84653747e-01
-5.24376929e-01 -1.15211701e+00 1.85915005e+00 -7.69933015e-02
-5.69079101e-01 -7.21555531e-01 1.20962277e-01 5.06307006e-01
-1.97456107e-01 1.03096175e+00 2.37500817e-01 -1.63794890e-01
-3.43742222e-01 -8.06322992e-01 -9.67923105e-02 2.11708974e-02
-2.62954891e-01 -4.87265646e-01 -7.05178261e-01 -2.06384242e-01
-3.90776694e-01 -7.06957579e-01 5.55750549e-01 6.39214754e-01
1.17399323e+00 9.40886065e-02 -4.93718475e-01 8.51163983e-01
9.43258226e-01 -3.51590961e-01 4.58929390e-01 5.20072818e-01
1.21175146e+00 7.40063012e-01 7.63278246e-01 6.23089790e-01
5.49400806e-01 8.96152496e-01 6.68328226e-01 -4.41644698e-01
1.25051782e-01 -5.89571774e-01 -5.99857420e-02 2.70035744e-01
-4.83626097e-01 2.26850256e-01 -1.33922541e+00 1.81370005e-01
-1.84449184e+00 -3.43620539e-01 -2.06678826e-02 1.82319808e+00
4.95398611e-01 3.70677143e-01 5.04033744e-01 5.76049648e-02
5.08594096e-01 -4.91236076e-02 -6.30933940e-01 2.23677412e-01
4.15566087e-01 2.03018352e-01 4.55720335e-01 7.61579797e-02
-1.58561718e+00 1.00232255e+00 6.66420603e+00 1.24438301e-01
-8.25617671e-01 -1.12016439e-01 2.74615079e-01 -5.80622852e-01
6.07557237e-01 -6.91271603e-01 -8.43769133e-01 1.03989959e-01
1.28536880e-01 5.54394662e-01 -1.80172741e-01 1.11644554e+00
4.74831499e-02 4.36249226e-02 -1.47327220e+00 1.56895959e+00
1.00033283e-01 -9.08306420e-01 -7.61858746e-02 1.93004951e-01
5.32352686e-01 -1.21796206e-01 1.21988051e-01 2.13969082e-01
2.06032187e-01 -1.31947601e+00 9.43452001e-01 6.36582971e-01
8.31537187e-01 -1.08422649e+00 8.46445322e-01 6.19147420e-01
-1.32746243e+00 1.86592713e-01 -2.88393676e-01 -4.22401667e-01
3.03550631e-01 3.65935504e-01 -7.22314239e-01 1.90599337e-01
1.20891464e+00 7.16705322e-01 -6.54364765e-01 9.03671443e-01
-5.42931199e-01 -1.67072505e-01 -3.82225364e-01 -1.49105325e-01
8.89237523e-02 1.87870905e-01 3.00679415e-01 8.74512911e-01
3.18606123e-02 7.11098313e-02 4.63083923e-01 1.02434528e+00
-7.78261796e-02 -3.26422513e-01 -5.70154846e-01 2.64078826e-01
1.49347022e-01 1.12246454e+00 -6.04805410e-01 -1.26029760e-01
-2.63931379e-02 1.42568398e+00 4.60943580e-01 8.62436444e-02
-7.95927525e-01 -3.07382762e-01 7.69611955e-01 1.51316941e-01
2.00794429e-01 -7.64199793e-01 -3.20632547e-01 -1.12605083e+00
2.72462755e-01 -6.20347083e-01 2.27670565e-01 -1.02809250e+00
-1.20759273e+00 3.87538582e-01 2.50024229e-01 -1.33078814e+00
-3.95557195e-01 -1.18231142e+00 -2.36626253e-01 8.76544833e-01
-9.84254122e-01 -1.45808434e+00 -5.64734638e-01 6.17714882e-01
3.38500708e-01 2.03452930e-01 5.87746322e-01 -4.96161208e-02
-1.68543905e-01 6.07075393e-01 -7.63846338e-01 8.99214685e-01
5.41028976e-01 -1.23997891e+00 1.13813293e+00 3.74548137e-01
-1.58388391e-01 6.12444580e-01 5.15469611e-01 -8.11875045e-01
-1.48988962e+00 -6.93189025e-01 8.47654819e-01 -1.20960343e+00
-2.30857171e-02 -8.00202668e-01 -2.08561063e-01 8.24887156e-01
-2.91496277e-01 9.93394330e-02 2.43838564e-01 2.15487197e-01
-4.14846122e-01 2.81372011e-01 -1.21584988e+00 5.68230689e-01
1.56619918e+00 -2.56491989e-01 -6.97186351e-01 3.16336602e-01
7.00041831e-01 -1.14658070e+00 -7.39501238e-01 6.24842167e-01
9.79471624e-01 -8.13345492e-01 1.52684224e+00 -7.55267799e-01
6.60723329e-01 -2.25116655e-01 -3.43262702e-02 -1.14000547e+00
-2.49612123e-01 -1.16358221e-01 -3.53916585e-01 1.27163753e-01
-1.08869202e-01 -7.87064731e-02 1.50489295e+00 5.74837863e-01
2.10171103e-01 -9.21024382e-01 -1.00479817e+00 -6.00822806e-01
9.12976637e-02 -4.01634365e-01 4.12578166e-01 3.29810023e-01
-3.99114758e-01 1.61956519e-01 -5.91092467e-01 1.57094136e-01
9.59803820e-01 -1.17574729e-01 1.49441850e+00 -1.61287546e+00
-1.10795334e-01 -2.75503993e-01 -7.61482894e-01 -1.48378015e+00
1.59026146e-01 -4.12634850e-01 2.30348393e-01 -1.65639985e+00
2.69971546e-02 2.37985268e-01 2.31904089e-01 3.89411032e-01
-7.37656355e-02 5.31780422e-01 2.35857904e-01 5.79381809e-02
-6.59586430e-01 5.48022091e-01 1.51655531e+00 -5.48348352e-02
1.50033832e-01 2.38792095e-02 -4.98573482e-02 9.96900916e-01
4.17086214e-01 -1.51693076e-01 -3.89777087e-02 -5.98623812e-01
1.75925061e-01 -1.89850964e-02 9.90318179e-01 -1.29718518e+00
2.01554745e-01 7.21167922e-02 1.27863896e+00 -1.41196442e+00
9.57647443e-01 -8.81670892e-01 -7.38046095e-02 1.03692412e+00
-1.73692331e-01 2.80389100e-01 -3.33111063e-02 4.09745842e-01
9.17681083e-02 5.46769381e-01 6.06744170e-01 -6.71329677e-01
-8.24626505e-01 8.36784244e-01 1.38937578e-01 4.04850999e-03
8.19189608e-01 -3.40020090e-01 1.96139723e-01 -4.49700266e-01
-7.85420597e-01 3.55318427e-01 4.02143389e-01 5.92351675e-01
1.10660112e+00 -1.66912842e+00 -5.50687134e-01 1.06236331e-01
3.27676952e-01 5.99799931e-01 -7.39080235e-02 5.53022861e-01
-9.99025226e-01 6.28850758e-01 -4.12567675e-01 -9.86887574e-01
-1.24119234e+00 3.72974187e-01 7.17060149e-01 -1.15126774e-01
-7.15635955e-01 1.01477182e+00 2.42266655e-01 -1.07664609e+00
6.29446149e-01 -3.18684846e-01 3.95719148e-02 -4.03662324e-01
2.71546870e-01 3.51690441e-01 -1.65795565e-01 -7.95899987e-01
-5.49249351e-01 9.54143286e-01 2.39559427e-01 -1.66988790e-01
1.36895990e+00 1.77689105e-01 1.33457571e-01 3.59205276e-01
1.46289611e+00 -6.79093122e-01 -1.60148120e+00 -2.51206607e-01
-2.80932307e-01 -5.94947159e-01 -3.71765047e-01 -7.27863491e-01
-1.12951434e+00 1.21774566e+00 6.71757817e-01 -6.13731086e-01
7.32261360e-01 2.33827621e-01 1.01900518e+00 6.21724784e-01
7.42156267e-01 -1.27134049e+00 6.69176340e-01 6.40698493e-01
1.16389668e+00 -1.42520678e+00 1.04055442e-01 -5.15322983e-01
-4.46653426e-01 1.16834617e+00 1.22107959e+00 -4.62719202e-01
6.27674103e-01 3.43297832e-02 2.24342972e-01 -5.66648781e-01
-2.28316173e-01 -2.15302289e-01 8.03459823e-01 7.95073032e-01
4.18272793e-01 9.33182761e-02 -3.47441398e-02 4.48517352e-01
-5.40730655e-01 1.01721145e-01 -1.47128880e-01 1.17667615e+00
-4.05520588e-01 -7.60679781e-01 -7.63858974e-01 2.42552608e-02
-2.09568858e-01 4.60582256e-01 -9.14469004e-01 1.09020674e+00
1.51817188e-01 4.26195264e-01 2.40943953e-01 -6.57107770e-01
7.69467235e-01 7.80919716e-02 8.88541520e-01 -5.89528084e-01
-5.52532673e-01 7.74963573e-02 -9.37926397e-02 -1.01666057e+00
-2.90229648e-01 -3.95086884e-01 -1.29098892e+00 -3.81912500e-01
9.82790291e-02 -5.77085078e-01 7.21741199e-01 9.20592248e-01
6.66946173e-02 3.29974443e-01 1.69278681e-02 -1.78064430e+00
-4.97640342e-01 -8.94311666e-01 -4.52614427e-01 7.21425772e-01
1.75282404e-01 -1.36385083e+00 1.08438365e-01 -3.44390184e-01] | [6.952374458312988, -0.9257334470748901] |
436d364f-a0a8-419b-bb2a-e07d080211c9 | intelligent-resource-allocation-in-joint | null | null | https://ieeexplore.ieee.org/document/9921194 | https://ieeexplore.ieee.org/document/9921194 | Intelligent Resource Allocation in Joint Radar-Communication With Graph Neural Networks | Autonomous vehicles produce high data rates of sensory information from sensing systems. To achieve the advantages of sensor fusion among different vehicles in a cooperative driving scenario, high data-rate communication becomes essential. Current strategies for joint radar-communication (JRC) often rely on specialized hardware, prior knowledge of the system model, and entail diminished capability in either radar or communication functions. In this paper, we propose a framework for intelligent vehicles to conduct JRC, with minimal prior knowledge of the system model and a tunable performance balance, in an environment where surrounding vehicles execute radar detection periodically, which is typical in contemporary protocols. We introduce a metric on the usefulness of data to help an intelligent vehicle decide what, and to whom, data should be transmitted. The problem framework is cast as a generalized form of the Markov Decision Process (MDP). We identify deep reinforcement learning algorithms (DRL) and algorithmic extensions suitable for solving our JRC problem. For multi-agent scenarios, we introduce a Graph Neural Network (GNN) framework via a control channel. This framework enables modular and fair comparisons of various algorithmic extensions. Our experiments show that DRL results in superior performance compared to non-learning algorithms. Learning of inter-agent coordination in the GNN framework, based only on the Markov task reward, further improves performance. | ['David González G.', 'Yong Liang Guan', 'Dusit Niyato', 'Yanyu Cheng', 'Joash Lee'] | 2022-10-17 | null | null | null | ieee-transactions-on-vehicular-technology-3 | ['distributional-reinforcement-learning', 'joint-radar-communication'] | ['methodology', 'robots'] | [ 3.45577449e-01 3.61077815e-01 -2.88867503e-01 -2.55214572e-01
-6.23778462e-01 -3.33419383e-01 8.52761030e-01 -4.07457445e-03
-6.31079614e-01 8.32658052e-01 -4.38188553e-01 -6.80232525e-01
-4.11641657e-01 -9.63320971e-01 -7.95481503e-01 -9.88618016e-01
-5.72179437e-01 5.07947922e-01 2.37863839e-01 -4.75935370e-01
2.89640781e-02 7.14416385e-01 -1.41460299e+00 -5.23006976e-01
4.80102837e-01 1.22280192e+00 3.58914375e-01 1.04009843e+00
3.04982692e-01 8.68487716e-01 -7.54712820e-01 -8.99175480e-02
4.33624953e-01 -2.21103981e-01 -9.52567458e-02 -1.04787372e-01
-1.96048915e-01 -3.54930460e-01 -4.63262349e-01 7.47359455e-01
3.52228731e-01 -9.01712477e-02 6.86967373e-01 -1.77110839e+00
-8.43913332e-02 6.58278406e-01 -3.75058532e-01 6.99088676e-03
-1.47710353e-01 4.11180884e-01 6.88962638e-01 -3.45356427e-02
4.44981784e-01 1.06347442e+00 4.26571012e-01 7.13906467e-01
-1.16474831e+00 -8.54441345e-01 2.86842257e-01 3.07527870e-01
-1.02781999e+00 -6.06711745e-01 7.61049688e-01 -2.32595116e-01
8.60827923e-01 -1.12124726e-01 8.40575337e-01 9.10996020e-01
9.30950582e-01 5.01115322e-01 9.24426079e-01 -6.23492822e-02
8.37187231e-01 -3.66201907e-01 2.93777790e-02 5.72163463e-01
5.94022512e-01 8.99736881e-01 -5.77097595e-01 4.22218107e-02
3.13650638e-01 -2.92779952e-01 9.06353518e-02 -6.27775192e-01
-9.28883612e-01 1.00334954e+00 3.41862828e-01 -1.65476754e-01
-7.09636390e-01 9.97132719e-01 1.50045589e-01 8.29151988e-01
-2.49964848e-01 2.39363208e-01 -3.85365486e-02 3.20406607e-03
-4.87282187e-01 3.40035468e-01 9.84105110e-01 9.76079345e-01
1.00570524e+00 4.83256787e-01 2.12161899e-01 1.91290155e-02
6.67701423e-01 1.16060817e+00 -1.42863810e-01 -1.47687745e+00
3.16080540e-01 9.01248679e-03 2.10714504e-01 -7.92338669e-01
-7.83409178e-01 -4.36897069e-01 -9.01374280e-01 8.10892820e-01
-1.28203824e-01 -8.27373207e-01 -8.28132868e-01 1.80780637e+00
1.88929334e-01 1.72757596e-01 1.05223501e+00 7.74406075e-01
2.86524922e-01 6.34504557e-01 -3.28042656e-02 -3.98677617e-01
9.05257523e-01 -3.13847184e-01 -7.90639639e-01 -3.81085932e-01
5.01640618e-01 -2.67396837e-01 -1.59523010e-01 3.43669057e-01
-7.67335951e-01 -3.13681543e-01 -1.58586979e+00 6.25836253e-01
-4.16508883e-01 -3.72135878e-01 6.07806683e-01 8.61188531e-01
-1.40145254e+00 1.23440079e-01 -9.67044890e-01 -2.02652350e-01
1.60471901e-01 6.51122928e-01 2.14640304e-01 6.58544824e-02
-1.31068933e+00 1.12580550e+00 1.98624924e-01 1.88335717e-01
-1.48093212e+00 -3.40938210e-01 -9.60300803e-01 -3.48291904e-01
6.77249551e-01 -4.53104794e-01 1.35689974e+00 -5.80653548e-01
-1.72138381e+00 9.44947377e-02 2.84955740e-01 -1.34159648e+00
2.48292238e-01 1.26572505e-01 -4.73805040e-01 2.25946486e-01
-1.08284041e-01 9.88528311e-01 1.04630983e+00 -1.51433253e+00
-9.20701444e-01 -8.38120207e-02 2.62485921e-01 1.14635035e-01
5.69654286e-01 -5.51105440e-01 3.21748946e-03 -9.83463973e-03
-3.98965299e-01 -1.10306549e+00 -6.61537230e-01 -2.43558928e-01
5.98498806e-02 -2.25197122e-01 1.23658550e+00 7.98356533e-02
5.22394359e-01 -1.86895585e+00 -9.97446477e-02 5.15817940e-01
3.85080874e-01 3.36612649e-02 -3.35275233e-01 5.74701905e-01
7.17838407e-01 -3.58894378e-01 4.94863419e-03 -1.57744840e-01
1.39215201e-01 7.53317833e-01 -2.08972111e-01 5.76563716e-01
2.63730735e-01 7.99498856e-01 -6.76499844e-01 -3.71223480e-01
1.31666332e-01 2.73794383e-01 -1.56850889e-01 3.51197994e-03
-4.25584316e-01 3.03722978e-01 -7.81096637e-01 4.10142124e-01
5.85739553e-01 3.92843746e-02 2.58944660e-01 1.23166755e-01
-2.85061449e-01 -1.93047464e-01 -9.06031370e-01 1.15527523e+00
-4.79851425e-01 7.23030210e-01 7.64061689e-01 -1.25086105e+00
1.11282754e+00 2.37998907e-02 7.02788472e-01 -1.03410268e+00
3.88211727e-01 -2.21300423e-02 3.03466380e-01 -1.23251155e-01
6.22469246e-01 -1.74938031e-02 -4.36350971e-01 5.66338122e-01
-2.38496259e-01 -2.52570659e-01 3.55604529e-01 3.57821733e-01
1.57086301e+00 -4.65117484e-01 5.21800071e-02 -1.61604300e-01
1.22673586e-01 1.78752109e-01 6.02293491e-01 1.19652963e+00
-4.07376051e-01 -5.35231888e-01 4.00149345e-01 -1.30104870e-01
-6.01971686e-01 -1.01496303e+00 1.82685494e-01 7.12304413e-01
6.48794949e-01 -9.23546404e-02 -3.81867141e-01 -3.30337524e-01
3.26473773e-01 7.65016556e-01 -5.05626142e-01 -3.38582039e-01
-3.97481531e-01 -4.57597256e-01 5.69419622e-01 2.69360960e-01
4.96996731e-01 -7.54298091e-01 -1.24913478e+00 5.63260615e-01
3.56802195e-01 -1.31156433e+00 4.45080735e-02 6.17683709e-01
-4.32587773e-01 -9.89813387e-01 -7.37243593e-02 -4.22156632e-01
1.71743035e-01 5.62821090e-01 7.20056415e-01 -7.66884536e-02
3.44214998e-02 1.00887930e+00 -3.79830837e-01 -6.96750760e-01
-7.54933059e-01 -1.52990222e-01 2.41101608e-01 7.79615045e-02
2.02213585e-01 -4.14380312e-01 -2.66418934e-01 1.34083703e-01
-7.15176404e-01 -5.79823926e-02 1.07955468e+00 4.17893738e-01
3.30876052e-01 2.39255115e-01 7.80471742e-01 -3.19550067e-01
6.21301770e-01 -4.42094386e-01 -1.26723909e+00 -3.89177501e-02
-7.21692920e-01 4.68834937e-01 4.60837930e-01 -6.97730333e-02
-7.73690164e-01 3.25959414e-01 2.19722852e-01 -2.29764089e-01
2.24622432e-02 4.60231870e-01 2.68642716e-02 -4.82429534e-01
2.80728012e-01 2.26705939e-01 6.27877176e-01 4.19062793e-01
2.85063326e-01 5.49052417e-01 4.12900150e-01 -4.88479197e-01
9.12560880e-01 4.81300503e-01 5.86758196e-01 -9.50169504e-01
-3.51385176e-01 -2.85342541e-02 -1.52111873e-01 -8.04411352e-01
9.25930679e-01 -1.07935667e+00 -1.35194218e+00 2.11451828e-01
-1.12528360e+00 -7.53953457e-01 -2.01735750e-01 8.33558619e-01
-8.62552762e-01 -3.20101194e-02 -2.50868231e-01 -1.22963285e+00
7.05556646e-02 -1.24495912e+00 8.51084352e-01 2.09957018e-01
5.08685887e-01 -9.16558206e-01 1.20206632e-01 1.50179967e-01
7.88800597e-01 2.81404108e-01 5.41472077e-01 -3.32642525e-01
-1.24742174e+00 -8.24367106e-02 -3.18158492e-02 1.27474675e-02
-2.69472927e-01 -1.24236941e-01 -4.98356223e-01 -5.23782134e-01
-2.94854909e-01 -3.76848042e-01 9.17794824e-01 5.35507977e-01
4.52509105e-01 -1.14161700e-01 -5.44247866e-01 1.95058823e-01
1.43888259e+00 5.59941947e-01 3.59582901e-01 1.16260208e-01
1.69189900e-01 5.71375370e-01 5.09053349e-01 4.66160119e-01
8.37539554e-01 5.13250172e-01 1.02352393e+00 3.29847604e-01
9.60723087e-02 1.36226818e-01 9.41379189e-01 4.82581198e-01
3.41182165e-02 -4.08400208e-01 -7.28352964e-01 1.50546119e-01
-1.96874511e+00 -9.08742666e-01 1.59894511e-01 1.92259502e+00
3.93306196e-01 4.70113724e-01 -1.73562057e-02 -2.19212666e-01
6.35526597e-01 1.83508188e-01 -8.67153287e-01 -5.10756433e-01
-5.50885983e-02 -2.17410773e-01 1.26552629e+00 7.45869875e-01
-8.58304381e-01 6.04906619e-01 6.64190149e+00 6.98678493e-01
-1.01015139e+00 4.63572219e-02 5.03708608e-02 6.35134056e-02
-2.18025580e-01 -1.38363456e-02 -1.12289608e+00 1.96206018e-01
1.57214582e+00 -2.10543349e-01 4.99493033e-01 4.65165168e-01
5.39977551e-01 -4.67108846e-01 -1.05606210e+00 7.67404854e-01
-2.69982126e-02 -1.53007472e+00 -3.02124679e-01 5.13859749e-01
4.57263678e-01 4.87294465e-01 -1.11039907e-01 6.37525678e-01
1.21573913e+00 -7.47415841e-01 7.51540542e-01 6.54922783e-01
3.44451487e-01 -8.72483552e-01 6.65134311e-01 3.89503211e-01
-1.22756720e+00 -3.42761010e-01 -2.83085585e-01 -1.85963288e-01
3.18761855e-01 5.12236178e-01 -6.32217348e-01 6.37296796e-01
3.43278229e-01 5.80920994e-01 -2.00757548e-01 6.53151572e-01
4.22082804e-02 4.17528510e-01 -5.47617257e-01 -6.32078946e-01
5.14775932e-01 -2.25795284e-01 7.70392239e-01 8.72179747e-01
2.01444834e-01 1.11767780e-02 5.51101208e-01 4.42569107e-01
1.73021555e-01 -6.32625937e-01 -1.01248789e+00 1.10072091e-01
6.00966811e-01 1.17819107e+00 -4.80466634e-01 -1.29592121e-01
-4.81145293e-01 9.13396478e-02 1.18959293e-01 5.36813259e-01
-8.38888645e-01 -2.56250411e-01 8.40878546e-01 -3.44413608e-01
5.01830161e-01 -5.80589354e-01 2.81777792e-02 -2.54829913e-01
-1.09986305e-01 -5.25435805e-01 1.56004457e-02 -4.14524972e-01
-8.91571581e-01 3.05652112e-01 -3.12235430e-02 -1.26303637e+00
-5.49327135e-01 -4.68479484e-01 -4.25307631e-01 2.33430550e-01
-1.92442155e+00 -9.03337896e-01 -7.84879029e-02 6.55769944e-01
1.67860627e-01 -5.08558273e-01 4.59034204e-01 -1.36097223e-01
-3.00601602e-01 1.46890447e-01 1.55415490e-01 -9.00746509e-02
2.19503596e-01 -1.00810361e+00 -9.73892733e-02 8.50706279e-01
-3.68995279e-01 -1.07526697e-01 9.37774777e-01 -7.30174959e-01
-2.31258106e+00 -1.32446635e+00 1.74880281e-01 -1.53824408e-02
9.48215246e-01 -2.99016684e-01 -6.05309308e-02 4.73953456e-01
5.00362754e-01 -2.30987757e-01 3.58368039e-01 -2.00239748e-01
-3.84800695e-02 -5.77237129e-01 -8.48225176e-01 5.24272382e-01
5.03763676e-01 -1.38079211e-01 -1.80532351e-01 2.66798008e-02
7.83553421e-01 -8.33760947e-03 -6.71342194e-01 3.76692951e-01
4.87906575e-01 -5.48441827e-01 4.48483944e-01 -1.95485860e-01
-2.69108713e-01 -6.30526483e-01 -4.93432462e-01 -1.31678081e+00
3.11818365e-02 -1.03024030e+00 -2.76550233e-01 6.32054210e-01
5.64569771e-01 -7.92418718e-01 8.49997401e-01 2.23838016e-01
-2.96265870e-01 -3.26920331e-01 -1.29580891e+00 -1.00959408e+00
-3.76418307e-02 -8.81186008e-01 3.10320199e-01 1.24596454e-01
-1.96297944e-01 4.94835228e-01 -3.55314285e-01 6.84761405e-01
1.21141255e+00 4.76033315e-02 9.50762391e-01 -1.32009387e+00
-1.94838986e-01 -3.85568082e-01 -6.45274997e-01 -1.22139978e+00
4.34716105e-01 -5.95536470e-01 5.41368127e-01 -1.41088402e+00
-3.55117321e-01 -6.93107247e-01 -2.99088396e-02 2.90626734e-01
6.40234292e-01 -1.77358314e-01 3.32844734e-01 1.13181621e-01
-1.08816123e+00 6.83158398e-01 9.72731292e-01 -4.53065246e-01
-8.90349001e-02 2.37273738e-01 -4.83095765e-01 2.53790975e-01
1.14340734e+00 -3.34241122e-01 -4.48815525e-01 -2.75084525e-01
2.24030033e-01 6.20593309e-01 5.85345924e-01 -1.39759851e+00
9.16190505e-01 -3.68427575e-01 -5.75872436e-02 -7.93781936e-01
6.27885461e-01 -1.08942389e+00 1.15284532e-01 1.00301981e+00
-3.81404877e-01 1.38951922e-02 -6.79785684e-02 1.24778342e+00
-1.82488766e-02 -6.62343856e-03 7.41013467e-01 3.03045124e-01
-8.61134112e-01 1.74565643e-01 -1.28694749e+00 -3.06275815e-01
1.42732000e+00 -4.91039567e-02 -4.34267282e-01 -8.20006251e-01
-6.56391203e-01 9.95607078e-01 4.24743928e-02 3.03937912e-01
6.63445950e-01 -1.12226415e+00 -7.06449687e-01 1.00206211e-01
7.32220113e-02 -4.49112922e-01 3.36949117e-02 9.12766039e-01
-9.40746069e-02 6.05399847e-01 -2.06006512e-01 -6.41233504e-01
-7.96806335e-01 3.33508939e-01 3.95475775e-01 -2.55261958e-01
-3.35419863e-01 9.79190096e-02 -1.55130580e-01 -1.97487220e-01
3.11494678e-01 -1.70249850e-01 -1.13509379e-01 -7.28692114e-02
3.20646167e-01 7.95695409e-02 -2.37547662e-02 -4.12022918e-01
-1.62928790e-01 3.29584390e-01 -1.60422608e-01 -3.84595066e-01
9.26055431e-01 -4.67508048e-01 4.31206286e-01 2.15972155e-01
6.16835296e-01 -3.90932858e-01 -1.73530555e+00 -1.83600053e-01
-1.25385970e-01 4.02382702e-01 5.28080583e-01 -5.36531985e-01
-1.18843651e+00 3.90122175e-01 6.35691524e-01 5.53740561e-01
8.62391770e-01 6.73921704e-02 5.03400326e-01 1.06658053e+00
8.12603891e-01 -1.39264464e+00 -4.62200828e-02 7.31274545e-01
5.07161498e-01 -1.13855934e+00 -1.16793975e-01 3.94817926e-02
-5.72872162e-01 1.08380008e+00 4.88452584e-01 -2.03404993e-01
9.99176025e-01 8.92082214e-01 1.84045225e-01 -2.97066033e-01
-1.33196557e+00 -7.78739333e-01 -5.89206457e-01 1.18266058e+00
-5.15465915e-01 3.24787229e-01 -3.58498865e-03 1.73293427e-01
-1.63679436e-01 -2.40901500e-01 8.60879123e-01 1.16532946e+00
-1.05632687e+00 -8.58182967e-01 -3.17575544e-01 4.42474306e-01
8.00545886e-02 4.17078495e-01 -1.38116345e-01 9.61936951e-01
-1.32980064e-01 1.50408399e+00 1.88647404e-01 -4.94354784e-01
4.55962494e-02 -5.23705661e-01 2.72872239e-01 7.85438940e-02
5.78628853e-02 -1.07330933e-01 2.54390746e-01 -7.30295479e-01
-9.34782922e-01 -7.16933548e-01 -1.44556510e+00 -4.03511673e-01
-2.82086097e-02 3.89615864e-01 9.21691895e-01 1.11528337e+00
4.81895477e-01 6.39810741e-01 8.38736713e-01 -9.57486629e-01
-6.26379848e-01 -4.28180456e-01 -6.02168560e-01 -6.32139623e-01
7.34808803e-01 -9.19404507e-01 -2.76031315e-01 -2.54512995e-01] | [5.222797870635986, 1.4087061882019043] |
39e19fdd-c146-4e7b-93da-5dd80a87fa41 | styleheat-one-shot-high-resolution-editable | 2203.04036 | null | https://arxiv.org/abs/2203.04036v2 | https://arxiv.org/pdf/2203.04036v2.pdf | StyleHEAT: One-Shot High-Resolution Editable Talking Face Generation via Pre-trained StyleGAN | One-shot talking face generation aims at synthesizing a high-quality talking face video from an arbitrary portrait image, driven by a video or an audio segment. One challenging quality factor is the resolution of the output video: higher resolution conveys more details. In this work, we investigate the latent feature space of a pre-trained StyleGAN and discover some excellent spatial transformation properties. Upon the observation, we explore the possibility of using a pre-trained StyleGAN to break through the resolution limit of training datasets. We propose a novel unified framework based on a pre-trained StyleGAN that enables a set of powerful functionalities, i.e., high-resolution video generation, disentangled control by driving video or audio, and flexible face editing. Our framework elevates the resolution of the synthesized talking face to 1024*1024 for the first time, even though the training dataset has a lower resolution. We design a video-based motion generation module and an audio-based one, which can be plugged into the framework either individually or jointly to drive the video generation. The predicted motion is used to transform the latent features of StyleGAN for visual animation. To compensate for the transformation distortion, we propose a calibration network as well as a domain loss to refine the features. Moreover, our framework allows two types of facial editing, i.e., global editing via GAN inversion and intuitive editing based on 3D morphable models. Comprehensive experiments show superior video quality, flexible controllability, and editability over state-of-the-art methods. | ['Yujiu Yang', 'Jue Wang', 'Baoyuan Wu', 'Qingyan Bai', 'Xuan Wang', 'Yanbo Fan', 'Mingdeng Cao', 'Xiaodong Cun', 'Yong Zhang', 'Fei Yin'] | 2022-03-08 | null | null | null | null | ['facial-editing', 'talking-face-generation'] | ['computer-vision', 'computer-vision'] | [ 6.13550663e-01 3.45829964e-01 -4.48492914e-02 -7.25437477e-02
-4.99545842e-01 -6.86591506e-01 6.99498236e-01 -1.06960475e+00
9.79961455e-02 5.88125527e-01 2.42919892e-01 1.55946046e-01
7.46100843e-02 -9.33164775e-01 -9.89931524e-01 -9.48227823e-01
4.35017914e-01 -6.46340400e-02 -6.82295263e-02 -3.52570444e-01
-1.69422388e-01 4.93392855e-01 -1.60410905e+00 1.71074614e-01
9.37366366e-01 9.85145032e-01 2.20927849e-01 5.32536328e-01
-4.59880196e-02 4.85761791e-01 -4.50650781e-01 -5.15431464e-01
4.21584994e-01 -7.55289912e-01 -2.78881371e-01 3.64588797e-01
2.86777556e-01 -5.53232849e-01 -4.43230599e-01 9.17764246e-01
2.97756642e-01 -2.22364873e-01 4.25747514e-01 -1.37823629e+00
-5.89169621e-01 4.47721839e-01 -5.28952837e-01 -4.74757940e-01
6.28265679e-01 4.81633365e-01 6.53435409e-01 -5.75860083e-01
9.62318003e-01 1.47083771e+00 2.72706419e-01 9.31454778e-01
-1.28423512e+00 -1.03273535e+00 2.14327220e-02 -1.13831304e-01
-1.29335165e+00 -7.04297662e-01 1.20939410e+00 -4.34092104e-01
3.33676971e-02 3.55101675e-01 9.26592410e-01 1.44918799e+00
1.14928588e-01 4.07198071e-01 1.00411808e+00 -2.87835300e-01
-1.00536868e-01 8.32793936e-02 -7.83633530e-01 8.85005713e-01
-9.53717455e-02 2.51597494e-01 -5.48022449e-01 9.64116231e-02
1.36581314e+00 4.14920077e-02 -7.46828258e-01 -2.73309261e-01
-1.30766499e+00 6.63818836e-01 2.36257181e-01 2.14840293e-01
-1.74476087e-01 6.72143623e-02 5.95686473e-02 2.39818737e-01
1.44726798e-01 3.27254206e-01 -1.51750324e-02 -2.80006919e-02
-1.02986550e+00 -2.21724939e-02 5.38124263e-01 9.49489355e-01
7.81698585e-01 4.59984750e-01 -4.11051720e-01 4.40763235e-01
1.82348251e-01 6.24348044e-01 3.99627477e-01 -1.17539370e+00
4.22229052e-01 5.26909053e-01 6.04570890e-03 -1.05585635e+00
4.97667305e-02 -2.60017812e-01 -1.07531309e+00 3.99365753e-01
2.10472107e-01 -1.60587192e-01 -7.46135533e-01 2.05548906e+00
4.85785365e-01 3.81810933e-01 1.31389368e-02 9.28020477e-01
5.07927597e-01 8.14214110e-01 -4.14857477e-01 -5.86623251e-01
1.22992694e+00 -9.45319235e-01 -9.70093369e-01 2.60811836e-01
1.32639199e-01 -7.59512961e-01 1.24817729e+00 3.94696414e-01
-1.19265294e+00 -7.94452965e-01 -1.01224792e+00 1.47102065e-02
1.76075995e-01 2.84827173e-01 1.79461434e-01 6.50834322e-01
-8.79845560e-01 5.40256798e-01 -7.21379399e-01 1.42648099e-02
2.04517558e-01 4.04603630e-01 -6.20276749e-01 8.53382647e-02
-1.13645172e+00 3.36495131e-01 1.83758244e-01 1.41962022e-01
-9.16877091e-01 -7.35168576e-01 -8.15159619e-01 8.39811042e-02
3.62561971e-01 -1.05131555e+00 6.32869184e-01 -1.31570995e+00
-2.44164538e+00 6.44710779e-01 3.94959226e-02 -6.61811680e-02
9.14447367e-01 -9.80156437e-02 -3.87165934e-01 2.79193789e-01
-1.67083248e-01 7.03391492e-01 1.39038646e+00 -1.12509608e+00
-3.75641793e-01 -1.95938155e-01 2.04485223e-01 1.15896285e-01
-6.39314294e-01 -1.42546028e-01 -5.38514256e-01 -8.31210077e-01
-2.29591325e-01 -1.09000897e+00 1.62450507e-01 2.28854164e-01
-4.03817296e-01 4.64818835e-01 1.19025409e+00 -6.07955158e-01
1.05621099e+00 -2.25139570e+00 5.82961738e-01 6.75932765e-02
2.09232524e-01 1.31833896e-01 -2.71253973e-01 1.70034349e-01
-1.09939538e-01 2.38361016e-01 -1.06917202e-01 -2.55028844e-01
-2.11478040e-01 1.03694968e-01 -4.18406785e-01 3.87555659e-01
2.92664587e-01 7.74477720e-01 -6.57019138e-01 -4.34496105e-01
1.92901969e-01 9.11302626e-01 -8.23511481e-01 3.86679202e-01
-1.96210384e-01 1.23413289e+00 -4.98774529e-01 3.61010164e-01
8.45754325e-01 2.02277657e-02 2.72148371e-01 -5.85511446e-01
-3.20229769e-01 -8.93935934e-02 -1.24187958e+00 1.91723752e+00
-4.61678237e-01 2.87494153e-01 4.73252028e-01 -6.01228774e-01
1.12630177e+00 3.68137747e-01 3.89948010e-01 -4.47405189e-01
2.18305096e-01 6.24357648e-02 -2.71647155e-01 -4.38632220e-01
2.42461354e-01 -2.76221305e-01 -2.20180061e-02 2.10631698e-01
8.48490819e-02 -1.84745207e-01 -1.01991601e-01 -1.75918743e-01
7.92960703e-01 4.23083454e-01 -1.43861428e-01 -1.10734805e-01
8.13985229e-01 -5.86057067e-01 7.46246755e-01 1.44072704e-03
4.06662643e-01 6.95914745e-01 6.45679235e-01 -2.56346136e-01
-1.22385824e+00 -7.84421504e-01 1.86471924e-01 6.47302985e-01
2.07472786e-01 -4.34498638e-01 -9.78933752e-01 -3.91375035e-01
-2.89877832e-01 1.64669216e-01 -5.20618916e-01 -2.96408832e-01
-8.01909447e-01 -1.28747836e-01 4.91528451e-01 2.63169587e-01
8.12481344e-01 -7.35238492e-01 -4.03459102e-01 -6.13450445e-02
-1.48586661e-01 -1.22859120e+00 -6.82209849e-01 -5.78285635e-01
-6.30885601e-01 -6.81308925e-01 -6.22417986e-01 -7.44235456e-01
5.01697063e-01 4.19251248e-02 5.03515482e-01 -9.20250639e-02
4.36585426e-04 2.51709610e-01 -1.75853387e-01 2.36819148e-01
-5.88366151e-01 5.77274971e-02 2.87140965e-01 5.86492419e-01
-6.69789910e-01 -9.77646291e-01 -8.36085021e-01 4.43394274e-01
-1.06856525e+00 6.12426102e-01 5.47767520e-01 9.97744977e-01
4.90580410e-01 1.04265371e-02 2.16990620e-01 -5.62309980e-01
2.53751040e-01 -5.66089004e-02 -7.32749999e-01 1.02630958e-01
-1.45701483e-01 1.67189345e-01 1.02830923e+00 -7.73665786e-01
-1.17909741e+00 2.74936080e-01 -1.45217389e-01 -9.08230722e-01
8.66835415e-02 -1.45225570e-01 -8.24764967e-01 -3.28167379e-01
2.65163392e-01 3.27421576e-01 3.25998753e-01 -2.88893133e-01
6.04110062e-01 4.40692455e-01 5.76271474e-01 -5.32087982e-01
1.25769377e+00 6.57790899e-01 1.08273722e-01 -8.40109587e-01
-1.23679653e-01 4.69315648e-01 -4.73420352e-01 -3.37517291e-01
9.39387441e-01 -1.00768185e+00 -8.96405637e-01 5.35168231e-01
-1.09355962e+00 -2.63977110e-01 -2.14966983e-01 4.04254049e-01
-6.68153048e-01 1.71370447e-01 -5.78566253e-01 -4.15754914e-01
-3.47826689e-01 -1.42350662e+00 1.18808639e+00 2.54964709e-01
9.46939290e-02 -5.12798071e-01 -1.11781567e-01 3.57070982e-01
4.12034929e-01 7.47358799e-01 6.74265862e-01 2.06483483e-01
-9.77513194e-01 2.32152998e-01 6.67030960e-02 2.22855389e-01
3.29012781e-01 3.80092919e-01 -8.43863130e-01 -4.85968411e-01
1.78946987e-01 -5.01791313e-02 6.44474506e-01 6.92503303e-02
1.06051946e+00 -7.50452518e-01 -3.18993747e-01 1.20848119e+00
1.27846754e+00 2.16569617e-01 7.34061301e-01 -1.77226495e-02
9.41897154e-01 4.59930360e-01 4.28624094e-01 4.77663457e-01
9.68585387e-02 9.76094246e-01 4.71645862e-01 4.56421543e-03
-1.01198375e-01 -5.95828116e-01 6.30929768e-01 7.86907852e-01
-3.13423008e-01 -1.68989375e-01 -3.52879316e-01 1.21601466e-02
-1.45827377e+00 -9.34385478e-01 3.10595930e-01 2.14845276e+00
7.85856426e-01 1.32320798e-04 4.92094010e-02 2.04699248e-01
9.35792387e-01 3.51199120e-01 -5.30962348e-01 -1.28767520e-01
2.59866137e-02 1.55605659e-01 1.31138653e-01 4.41850394e-01
-7.53845036e-01 8.77222061e-01 5.11269903e+00 9.18559074e-01
-1.67271817e+00 1.46792652e-02 4.65965807e-01 -2.39439875e-01
-7.50757575e-01 -1.85121950e-02 -6.23384595e-01 6.36725843e-01
6.14377499e-01 -1.11149080e-01 5.53652406e-01 6.42627239e-01
3.12117845e-01 5.79996586e-01 -9.47059214e-01 1.00840950e+00
9.89185199e-02 -1.54042089e+00 5.19919932e-01 3.01329315e-01
6.59662068e-01 -8.85220349e-01 3.12959850e-01 1.81736350e-02
-2.83139378e-01 -8.70171487e-01 8.75120997e-01 6.85377479e-01
1.56039929e+00 -8.22881877e-01 6.60407767e-02 2.69349128e-01
-1.20808482e+00 -1.53136447e-01 -5.91066703e-02 1.62996456e-01
3.01487416e-01 3.96679908e-01 -2.52039492e-01 5.97533226e-01
3.96564901e-01 6.29443586e-01 -1.29895747e-01 2.88653135e-01
-2.32741162e-01 2.52436191e-01 -2.59513646e-01 4.78827119e-01
-1.09688416e-01 -5.25653124e-01 7.12804317e-01 6.84743285e-01
7.57510066e-01 2.07303330e-01 2.71935910e-02 1.15913022e+00
-2.95536220e-01 -1.04187451e-01 -8.35058510e-01 4.38551195e-02
4.23225343e-01 1.48955059e+00 -2.73489118e-01 -6.30943701e-02
-2.14826837e-01 1.22695374e+00 -3.04007679e-02 2.46729910e-01
-1.22592628e+00 -2.82298923e-01 7.68001139e-01 2.99696356e-01
4.02022302e-01 -1.89965442e-01 1.42014503e-01 -1.50320554e+00
1.66966781e-01 -9.35830593e-01 -3.14339399e-01 -7.55639911e-01
-7.03075707e-01 8.11628342e-01 -2.96347558e-01 -1.47019088e+00
-3.19303274e-01 -3.00248146e-01 -6.54381454e-01 6.49012029e-01
-1.40988290e+00 -1.43142581e+00 -7.03828931e-01 8.42660606e-01
3.65461528e-01 -1.90930828e-01 6.75153196e-01 3.08150083e-01
-6.51327729e-01 9.52930212e-01 -3.48974392e-02 5.37798405e-02
8.18213582e-01 -6.53090894e-01 8.90453234e-02 7.61204183e-01
-2.11269885e-01 4.39702064e-01 5.20434320e-01 -3.88513654e-01
-1.87294662e+00 -1.16635609e+00 1.36276871e-01 -8.66749734e-02
4.20114219e-01 -6.09143972e-01 -6.82492673e-01 6.22652590e-01
2.30087906e-01 -5.35153076e-02 3.86800200e-01 -8.06364000e-01
-2.55336493e-01 -5.78736782e-01 -1.20687103e+00 6.57075882e-01
1.35121441e+00 -5.14091611e-01 -1.47364810e-01 -1.02264717e-01
1.06919992e+00 -5.11680067e-01 -9.91028607e-01 5.07579744e-01
8.85468125e-01 -1.01176560e+00 8.20705652e-01 -2.38096323e-02
7.99354613e-01 -4.68430191e-01 8.09353143e-02 -1.11820877e+00
-2.85484284e-01 -1.38935924e+00 4.37120385e-02 1.74659443e+00
-2.29173489e-02 -4.77785826e-01 6.99073374e-01 2.87741423e-01
-5.85847721e-02 -5.45915008e-01 -8.35264146e-01 -6.56767368e-01
-1.67711347e-01 1.56399339e-01 9.56556559e-01 9.93103802e-01
-2.52180755e-01 2.70095587e-01 -7.45110571e-01 1.80570379e-01
4.93350714e-01 2.12279454e-01 1.01851475e+00 -1.00192285e+00
-3.08659881e-01 -2.45367259e-01 -4.36489701e-01 -1.23657417e+00
2.58747667e-01 -5.04080057e-01 -3.27868849e-01 -8.26857865e-01
-5.21945171e-02 -2.31758699e-01 2.87118226e-01 2.29126081e-01
2.73707688e-01 3.83059204e-01 4.62839842e-01 2.35588372e-01
2.08454859e-02 8.77568364e-01 1.85134017e+00 -6.33111130e-03
-4.16317105e-01 -2.57926792e-01 -5.91293335e-01 6.03374898e-01
5.93696713e-01 -1.39063179e-01 -4.82206017e-01 -4.25097913e-01
1.05762862e-01 6.58686519e-01 3.92185539e-01 -1.07159233e+00
3.20496294e-03 -1.99386671e-01 3.36687207e-01 5.99767417e-02
6.25812650e-01 -9.15991426e-01 7.39102483e-01 3.39969128e-01
-2.27067649e-01 -1.05685554e-01 -4.76579331e-02 5.19843280e-01
-2.94759303e-01 4.07230198e-01 9.86739337e-01 -7.02069551e-02
-1.48089051e-01 4.53954220e-01 -2.00501513e-02 -2.61362165e-01
1.04037845e+00 -2.98569173e-01 -4.02476817e-01 -4.64076817e-01
-6.81758106e-01 -1.91521585e-01 8.04752350e-01 4.52185124e-01
5.36584139e-01 -1.73214817e+00 -5.62945306e-01 7.84938991e-01
-2.17518002e-01 -4.64417934e-02 3.46822321e-01 7.68287122e-01
-4.43785697e-01 1.38436034e-01 -7.34543562e-01 -6.56575382e-01
-1.19667864e+00 5.26307702e-01 1.80798307e-01 -1.52904391e-01
-6.80671394e-01 4.30690706e-01 5.60093582e-01 -1.87319562e-01
-1.42027736e-01 -2.99320757e-01 -8.68212283e-02 -3.41664907e-03
5.05146801e-01 1.79861248e-01 -3.35460454e-01 -7.88056433e-01
-1.01307161e-01 1.09179044e+00 2.94289947e-01 -1.09161049e-01
1.26954472e+00 -2.08490267e-01 -1.07725859e-01 6.93376213e-02
1.24392450e+00 2.39720732e-01 -1.70866406e+00 6.64371252e-02
-8.80659938e-01 -6.79393530e-01 -1.26644745e-01 -2.35866502e-01
-1.46308732e+00 8.08353841e-01 5.01264334e-01 -7.39414096e-02
1.39912391e+00 -2.67498851e-01 9.66563821e-01 -2.02053294e-01
3.93185705e-01 -6.81034565e-01 2.88232982e-01 5.88735230e-02
1.06060708e+00 -7.80784369e-01 -4.51783299e-01 -5.87839186e-01
-5.41356742e-01 1.27514410e+00 6.47839606e-01 -1.71620607e-01
3.83261949e-01 3.43616396e-01 -3.22765142e-01 2.19936684e-01
-5.81987083e-01 3.25808197e-01 1.80504620e-01 4.49519515e-01
7.76608214e-02 -7.67262578e-02 -1.89093649e-01 6.12980545e-01
-4.64857459e-01 2.41826057e-01 6.10477209e-01 4.04197007e-01
-8.09196085e-02 -1.00080025e+00 -4.33900088e-01 -1.11592457e-01
-3.65174741e-01 1.93620294e-01 -1.04068249e-01 7.61729479e-01
4.10258889e-01 6.49206460e-01 3.20113776e-03 -7.31434941e-01
2.76003748e-01 -1.86584610e-02 5.38664818e-01 -2.02298149e-01
-2.76890367e-01 2.98969656e-01 -2.32958660e-01 -8.02046001e-01
-4.48778033e-01 -2.75357336e-01 -8.63041401e-01 -5.34177184e-01
-3.30084950e-01 4.58664633e-02 3.32988292e-01 6.86982930e-01
6.18848085e-01 5.44354439e-01 1.09259093e+00 -1.00322771e+00
-1.29654080e-01 -6.96712375e-01 -5.23752987e-01 3.50477606e-01
4.95497108e-01 -6.46179795e-01 -4.27060306e-01 3.50769907e-01] | [12.593626022338867, -0.3636786639690399] |
37988d3d-7365-4346-968a-636b66da8cc6 | hyperparameter-tricks-in-multi-agent | 2102.03479 | null | https://arxiv.org/abs/2102.03479v11 | https://arxiv.org/pdf/2102.03479v11.pdf | Rethinking the Implementation Matters in Cooperative Multi-Agent Reinforcement Learning | Multi-Agent Reinforcement Learning (MARL) has seen revolutionary breakthroughs with its successful application to multi-agent cooperative tasks such as computer games and robot swarms. QMIX, a widely popular MARL algorithm, has been used to solve cooperative tasks, e.g. Starcraft Multi-Agent Challenge (SMAC), Difficulty-Enhanced Predator-Prey (DEPP). Recent variants of QMIX target relaxing the monotonicity constraint of QMIX, allowing for performance improvement in SMAC. However, in this paper, we investigate the code-level optimizations of these variants and the monotonicity constraint. We find that (1) such improvements of the variants are significantly affected by various code-level optimizations; (2) QMIX with normalized optimizations outperforms other previous works in SMAC; (3) the monotonicity constraint may improve sample efficiency in SMAC and DEPP. Last, a discussion with theoretical analysis is demonstrated about why QMIX works well in SMAC. We open-source the code at \url{https://github.com/hijkzzz/pymarl2}. | ['Haibin Wu', 'Siyang Jiang', 'Shih-wei Liao', 'Seth Austin Harding', 'Jian Hu'] | 2021-02-06 | null | null | null | null | ['smac-1', 'smac'] | ['playing-games', 'playing-games'] | [-4.00192976e-01 -2.77945220e-01 -1.31628424e-01 4.80283201e-01
-4.48757380e-01 -4.65650350e-01 5.23031771e-01 3.50267053e-01
-6.56633317e-01 1.19141746e+00 -2.00364530e-01 -2.09241092e-01
-4.41680908e-01 -5.42250156e-01 -6.43446028e-01 -9.93045151e-01
-7.32738733e-01 4.65667099e-01 5.65195382e-01 -9.60248590e-01
4.52534080e-01 8.14519823e-02 -1.59096920e+00 -6.11554459e-02
1.03395712e+00 6.15764976e-01 4.64943588e-01 9.64796782e-01
4.18663889e-01 1.00793505e+00 -6.84851944e-01 -1.69535547e-01
4.51511383e-01 -5.01581848e-01 -9.15111065e-01 -4.41334933e-01
-4.75638479e-01 2.05455497e-02 7.29369149e-02 9.82314944e-01
6.58226669e-01 1.81038737e-01 3.33953202e-01 -2.06448507e+00
-2.57454813e-01 6.62295341e-01 -9.89398360e-01 6.34462908e-02
1.74042046e-01 4.84949946e-01 9.40519571e-01 -2.68253446e-01
4.35873598e-01 1.22921360e+00 7.02527285e-01 5.90743303e-01
-6.22115970e-01 -5.55014789e-01 8.97619128e-02 4.60612625e-01
-1.29367507e+00 6.58879280e-02 4.17157143e-01 -4.51128110e-02
1.18160391e+00 1.99652120e-01 6.60128534e-01 8.42508495e-01
5.83383679e-01 9.54250157e-01 1.06313252e+00 -3.32337439e-01
7.03181207e-01 -3.77293497e-01 -3.30236793e-01 7.84697473e-01
5.38099170e-01 3.31878275e-01 -5.51350653e-01 -5.18516421e-01
6.04269803e-01 -3.62919331e-01 3.66632156e-02 -5.28584003e-01
-1.36611056e+00 1.26622677e+00 4.04314578e-01 1.60595044e-01
-4.87686098e-01 4.54539746e-01 5.21885633e-01 5.94521284e-01
2.82291353e-01 1.01471233e+00 -5.81279159e-01 -5.64696491e-01
-2.10453361e-01 7.39558578e-01 1.00874507e+00 8.40469956e-01
6.23115122e-01 1.53421953e-01 2.80053973e-01 6.60307050e-01
2.97219783e-01 5.51495910e-01 6.22114897e-01 -1.55145133e+00
3.34729970e-01 4.34434056e-01 3.00298989e-01 -6.52390063e-01
-8.72249782e-01 -2.81216860e-01 -7.67366827e-01 7.11553514e-01
3.41131330e-01 -7.17001140e-01 -1.70959681e-02 1.62101161e+00
5.10562718e-01 2.30583087e-01 6.22226834e-01 9.47071552e-01
4.31177765e-01 5.79759181e-01 -1.55573532e-01 -3.76713395e-01
1.19351149e+00 -1.42281413e+00 -2.43771240e-01 -2.69307822e-01
8.73344541e-01 -6.75584733e-01 8.48807633e-01 4.37081695e-01
-9.35981512e-01 -7.36224502e-02 -1.06620193e+00 5.89433312e-01
-2.75002003e-01 -3.17611963e-01 9.54795659e-01 5.63492000e-01
-1.03973746e+00 6.77833557e-01 -1.10884559e+00 -5.59278250e-01
4.55237441e-02 5.84983647e-01 2.59885993e-02 2.67881781e-01
-1.03155565e+00 9.08444166e-01 4.12520528e-01 -4.32785571e-01
-9.39008355e-01 -4.18816119e-01 -6.51194930e-01 -1.02697603e-01
8.66082191e-01 -5.44929683e-01 1.54082501e+00 -8.97069752e-01
-1.97027218e+00 3.30261409e-01 4.15728986e-01 -6.91415727e-01
4.80419904e-01 -2.28640452e-01 6.91562667e-02 2.70602368e-02
1.63103476e-01 6.65438890e-01 5.00629425e-01 -1.14140725e+00
-1.07458627e+00 9.04545002e-03 5.00642598e-01 7.27320492e-01
7.50244595e-03 1.12537755e-05 -2.04487089e-02 -4.85494852e-01
-7.03875422e-01 -1.17034781e+00 -6.96864486e-01 -3.45620513e-01
1.02732055e-01 -5.28898001e-01 6.56128049e-01 4.07224298e-02
7.93879211e-01 -1.83438587e+00 5.21117270e-01 -2.52397954e-01
1.80946484e-01 3.79133940e-01 -6.97371244e-01 1.01958060e+00
5.92384577e-01 -6.28884062e-02 -1.05004169e-01 -1.86426491e-01
8.93268883e-02 4.53063399e-01 2.70148993e-01 5.74293792e-01
9.67019424e-02 7.53929019e-01 -1.29852700e+00 -3.90978783e-01
5.85670955e-02 -6.74006864e-02 -8.60909998e-01 1.51726201e-01
-6.23308957e-01 3.87500644e-01 -5.39786577e-01 7.04245031e-01
5.45168519e-01 -3.04272622e-01 1.26122355e-01 5.27698934e-01
-4.45801049e-01 -1.80515528e-01 -1.22702432e+00 1.62148809e+00
-3.13159108e-01 4.38537985e-01 3.91388506e-01 -1.10943711e+00
5.67620099e-01 1.06527187e-01 8.20135832e-01 -6.88523352e-01
6.76153451e-02 2.06807837e-01 3.94382089e-01 -4.54103470e-01
6.56808794e-01 2.63311684e-01 -2.68055856e-01 5.21822751e-01
-6.05944656e-02 -3.21089685e-01 6.67092383e-01 1.04445979e-01
1.40350235e+00 1.33358106e-01 7.59015381e-01 -5.64003944e-01
4.23468798e-01 5.19422293e-01 7.71854997e-01 1.04418552e+00
-5.18473744e-01 -5.15261432e-03 7.09616184e-01 -4.43493247e-01
-8.54785681e-01 -7.79017627e-01 4.55002189e-01 1.42799568e+00
6.38101161e-01 -6.01185203e-01 -4.86786664e-01 -4.76948857e-01
1.79929167e-01 5.82698941e-01 -5.56837320e-01 -4.27189581e-02
-5.82178771e-01 -1.20570934e+00 6.30412638e-01 1.11002522e-02
7.01618254e-01 -1.33448577e+00 -1.20416570e+00 2.85072714e-01
-4.03000973e-02 -7.99434066e-01 -1.95281312e-01 1.92830905e-01
-5.39550126e-01 -1.34849143e+00 -6.94117904e-01 -6.34099305e-01
9.42739323e-02 4.37450618e-01 1.13253629e+00 4.19658989e-01
-1.55558079e-01 5.59683025e-01 -1.06525064e+00 -6.23791337e-01
-6.10840261e-01 4.14645642e-01 3.83613259e-01 -5.18982291e-01
-6.39664680e-02 -5.03489375e-01 -4.45989907e-01 5.64540386e-01
-7.20835984e-01 1.59247279e-01 6.27022266e-01 8.89252603e-01
1.11250661e-01 1.97841004e-01 7.71993935e-01 -4.62092489e-01
9.05599475e-01 -6.55315876e-01 -9.10055757e-01 1.67920128e-01
-4.61100727e-01 2.12985441e-01 7.18721151e-01 -5.26906371e-01
-6.22425020e-01 -1.48452699e-01 -1.22562155e-01 -1.37644574e-01
1.54061943e-01 5.28462350e-01 4.82263803e-01 -3.31867546e-01
6.98540926e-01 1.59664258e-01 5.24200976e-01 -1.10187680e-01
2.47100387e-02 4.31611657e-01 4.89239693e-02 -7.19970703e-01
7.02679038e-01 1.70110032e-01 8.43902081e-02 -8.57303798e-01
-4.26374793e-01 -3.42410803e-01 -6.56197220e-02 -9.69764441e-02
5.58629513e-01 -8.24849427e-01 -1.27852845e+00 5.38887382e-01
-9.46875453e-01 -9.98401284e-01 -2.14933708e-01 6.13180041e-01
-8.82027030e-01 3.31617802e-01 -6.37479663e-01 -9.24535394e-01
-3.35550934e-01 -1.23062658e+00 7.16039419e-01 6.41971767e-01
1.99045688e-01 -9.34903860e-01 6.75593972e-01 9.39581543e-02
5.49676418e-01 2.39501551e-01 4.44690973e-01 -5.56279182e-01
-2.85091370e-01 3.77232313e-01 1.42442212e-01 -7.15736821e-02
6.31570211e-03 1.32482592e-02 -2.12644041e-01 -9.14565921e-01
-3.49799484e-01 -5.23035705e-01 3.50113928e-01 3.03899407e-01
4.83530521e-01 -2.62995034e-01 -8.14147890e-02 4.26327556e-01
1.45992541e+00 3.88275683e-01 3.03742468e-01 9.08731401e-01
9.81515571e-02 1.36013970e-01 9.61847961e-01 1.07492375e+00
5.81372499e-01 6.70856357e-01 1.02354193e+00 1.41714200e-01
2.94065922e-01 2.08980888e-01 9.20537472e-01 8.51922274e-01
-4.13527519e-01 -2.28573889e-01 -9.38300550e-01 3.72568637e-01
-2.53558946e+00 -8.68769646e-01 -6.85677901e-02 1.88808858e+00
5.75778246e-01 -2.15096325e-01 5.12319505e-01 -2.85573035e-01
5.81344128e-01 4.82598087e-03 -8.44864905e-01 -4.60833967e-01
-2.20650002e-01 -5.82418172e-03 7.00925529e-01 5.87771893e-01
-9.39486742e-01 1.08830500e+00 5.61664677e+00 1.08176315e+00
-9.17182028e-01 3.17204237e-01 5.55661209e-02 -2.96502531e-01
3.19053650e-01 -1.48083791e-02 -5.93152940e-01 4.80130911e-01
6.29080653e-01 -5.49474835e-01 1.01419640e+00 9.61063921e-01
2.54591227e-01 -5.67172348e-01 -5.48094809e-01 7.74387896e-01
-7.67392516e-02 -1.27012968e+00 -4.28027630e-01 1.62458867e-01
8.96002054e-01 6.33156300e-01 -1.11992836e-01 7.55970895e-01
1.05916357e+00 -6.99996293e-01 5.52658141e-01 -1.02488615e-01
1.96898460e-01 -1.01113439e+00 1.10943627e+00 4.87937003e-01
-1.19509935e+00 -3.35445464e-01 -6.80999398e-01 -6.64063931e-01
-1.59136310e-01 -7.59366527e-02 -5.78043282e-01 7.98529088e-01
9.10027504e-01 6.43624008e-01 -6.56706095e-01 1.29535544e+00
-1.63000777e-01 4.09539163e-01 -2.76444376e-01 -7.84175515e-01
6.17487609e-01 -2.89308280e-01 8.46182168e-01 7.67951310e-01
2.47817039e-01 -1.34976162e-02 4.19779181e-01 4.19649571e-01
3.28272969e-01 7.94189945e-02 -3.01286966e-01 2.31602738e-04
4.94597495e-01 1.28864133e+00 -7.18968689e-01 3.95989567e-02
-3.85116279e-01 8.66829336e-01 5.04283905e-01 1.44813418e-01
-1.12876916e+00 -3.49094927e-01 1.05474019e+00 -4.04351652e-01
3.74178827e-01 -4.27140355e-01 7.61203766e-02 -1.08022702e+00
-3.12063932e-01 -1.06871390e+00 5.16493976e-01 -5.52681148e-01
-1.23917341e+00 4.21575308e-01 -6.74855486e-02 -1.45123088e+00
-3.50987732e-01 -6.01005793e-01 -7.44270444e-01 1.92949936e-01
-1.59151149e+00 -7.97631800e-01 -1.38146698e-01 5.61218083e-01
7.52322793e-01 -7.62675703e-01 7.67248034e-01 -5.18277138e-02
-8.08849156e-01 3.08704138e-01 6.71969771e-01 -3.53331327e-01
6.57579243e-01 -1.26809454e+00 4.88049611e-02 6.09741986e-01
-2.37110823e-01 6.20018132e-02 9.11735475e-01 -5.31449318e-01
-1.78757250e+00 -9.50331569e-01 5.91362454e-02 -7.58270025e-02
9.59074438e-01 -9.17642489e-02 -3.50251317e-01 2.49404401e-01
6.72768176e-01 -2.43530005e-01 5.09935498e-01 8.96806791e-02
1.38012394e-01 1.12507954e-01 -9.51971471e-01 7.64512420e-01
7.18209386e-01 3.37439150e-01 -1.83783397e-01 4.39568639e-01
7.37883568e-01 -3.75300914e-01 -8.06736708e-01 1.50613219e-01
3.28275055e-01 -1.07000363e+00 7.41620541e-01 -4.28657264e-01
3.67478341e-01 -4.67820853e-01 -6.27653077e-02 -1.86294377e+00
-4.26312327e-01 -9.42614853e-01 -1.84107825e-01 6.26127422e-01
2.60040671e-01 -8.63667190e-01 7.27846086e-01 -9.52334404e-02
-1.30838200e-01 -7.54716992e-01 -1.17404985e+00 -1.25012505e+00
4.32292402e-01 1.46632493e-01 4.09666717e-01 8.45101357e-01
4.26104695e-01 2.25143835e-01 -6.02073431e-01 3.34905833e-01
6.83813691e-01 -2.71101040e-03 9.96364474e-01 -9.61812854e-01
-7.59313166e-01 -5.71643114e-01 -1.37188500e-02 -7.74515212e-01
-2.77613588e-02 -6.08525753e-01 3.36716045e-03 -1.26444662e+00
1.71541274e-01 -5.51042676e-01 -1.26110539e-01 4.81707275e-01
-2.29370281e-01 6.74001575e-02 6.78272069e-01 3.74676198e-01
-1.47594380e+00 8.51144075e-01 1.15309381e+00 3.88592333e-02
-4.50203836e-01 5.63475005e-02 -4.46968824e-01 6.24863744e-01
1.46533728e+00 -5.63277125e-01 -1.75368994e-01 -3.99561703e-01
5.48721790e-01 2.46782482e-01 1.21730611e-01 -1.23691106e+00
3.96766543e-01 -7.49009252e-01 -3.89240235e-01 -1.23643853e-01
1.49984166e-01 -4.59435433e-01 -6.97880015e-02 1.25892615e+00
-1.29248695e-02 5.48982322e-01 2.87471563e-01 4.08198416e-01
-3.72731835e-02 -5.42478681e-01 8.35941911e-01 -2.97286272e-01
-8.75338078e-01 -2.39128824e-02 -1.01440144e+00 2.86459506e-01
1.45760012e+00 1.92861617e-01 -6.19606555e-01 -4.15687770e-01
-9.53899175e-02 9.76731360e-01 6.12581432e-01 3.78485322e-01
2.82414675e-01 -9.96896565e-01 -9.37655270e-01 -2.16110051e-01
-2.61904486e-03 -1.84604958e-01 7.28426725e-02 1.11853874e+00
-9.38063085e-01 1.36601970e-01 -5.66699445e-01 -3.43651980e-01
-1.35937631e+00 5.53872108e-01 4.07110423e-01 -6.95432544e-01
-2.68192172e-01 6.79933369e-01 -2.09613770e-01 -5.96828401e-01
1.49770871e-01 -4.94902916e-02 -8.97031799e-02 -3.02217364e-01
3.74488652e-01 6.59885764e-01 -3.93600851e-01 -1.89101368e-01
-4.90397811e-01 3.29051644e-01 -6.76388480e-03 -3.95172089e-02
1.67493892e+00 -4.28713486e-02 -1.73712850e-01 3.69418003e-02
6.21550798e-01 -1.53634265e-01 -1.43942261e+00 2.01158356e-02
4.01880108e-02 -1.99642833e-02 -1.69988617e-01 -7.19584286e-01
-8.93513858e-01 3.04184347e-01 4.58713889e-01 3.72311652e-01
1.05709112e+00 -1.34057194e-01 4.87504989e-01 8.06194544e-01
8.42151523e-01 -1.13067842e+00 3.36491883e-01 9.49589014e-01
8.67813110e-01 -1.32505536e+00 2.29338467e-01 -7.16972426e-02
-1.02518642e+00 1.02878594e+00 8.84176970e-01 -4.35207129e-01
3.16698432e-01 3.78557771e-01 -2.32828394e-01 -1.06129520e-01
-1.14177930e+00 -4.90240633e-01 -5.94756424e-01 7.69056499e-01
-7.57763162e-02 1.31153196e-01 -5.97623467e-01 4.06254143e-01
-2.52446413e-01 -2.73048669e-01 8.49767745e-01 1.30346668e+00
-5.74901044e-01 -1.31042933e+00 -4.82432246e-01 1.79589123e-01
-1.46579593e-01 4.09353264e-02 -4.63397861e-01 1.03969872e+00
6.46758527e-02 1.11500442e+00 -6.05159812e-02 -2.33777598e-01
4.25300002e-02 -6.19254291e-01 3.78909677e-01 -1.72074541e-01
-9.00745273e-01 -2.00570121e-01 6.95600361e-02 -5.95729887e-01
-5.69426298e-01 -6.49387956e-01 -1.37592936e+00 -6.06796503e-01
-2.12994054e-01 7.33140886e-01 3.93085092e-01 7.71374106e-01
5.36779284e-01 3.77134025e-01 5.81297696e-01 -1.01954949e+00
-5.42344868e-01 -8.22301686e-01 -5.68317235e-01 -4.56629926e-03
2.35246465e-01 -9.76680040e-01 -3.26006562e-01 -5.34929216e-01] | [3.7646656036376953, 1.9176326990127563] |
61aae092-9aad-44f6-83d6-ed19bf23663f | learning-with-label-noise-for-image-retrieval | 2112.10453 | null | https://arxiv.org/abs/2112.10453v2 | https://arxiv.org/pdf/2112.10453v2.pdf | Learning with Label Noise for Image Retrieval by Selecting Interactions | Learning with noisy labels is an active research area for image classification. However, the effect of noisy labels on image retrieval has been less studied. In this work, we propose a noise-resistant method for image retrieval named Teacher-based Selection of Interactions, T-SINT, which identifies noisy interactions, ie. elements in the distance matrix, and selects correct positive and negative interactions to be considered in the retrieval loss by using a teacher-based training setup which contributes to the stability. As a result, it consistently outperforms state-of-the-art methods on high noise rates across benchmark datasets with synthetic noise and more realistic noise. | ['Stéphane Clinchant', 'Rafael Sampaio de Rezende', 'Arnaud Sors', 'Sarah Ibrahimi'] | 2021-12-20 | null | null | null | null | ['learning-with-noisy-labels', 'learning-with-noisy-labels'] | ['computer-vision', 'natural-language-processing'] | [ 4.25917119e-01 -6.16160512e-01 -4.06550854e-01 -4.10947561e-01
-1.27463126e+00 -5.45158923e-01 6.28497899e-01 2.48428166e-01
-6.20521545e-01 6.22649014e-01 -1.54942915e-01 1.77414328e-01
-4.81058896e-01 -4.61667657e-01 -5.93596935e-01 -1.11948848e+00
3.33764330e-02 3.36865127e-01 2.50075340e-01 5.78714535e-02
2.69568712e-01 1.74637452e-01 -1.86425245e+00 1.88078165e-01
6.39558136e-01 1.17289078e+00 -1.33146048e-01 3.19946766e-01
7.22511560e-02 6.49136543e-01 -7.74778843e-01 -3.70112419e-01
3.84580076e-01 -3.40496421e-01 -5.65835476e-01 -1.16360396e-01
6.17831230e-01 -2.88033858e-02 -4.15934712e-01 1.20375752e+00
9.56977308e-01 2.46736646e-01 8.38893175e-01 -1.07934809e+00
-5.61563134e-01 5.84806621e-01 -6.15006447e-01 4.15453494e-01
1.19151928e-01 -1.62119597e-01 1.07776880e+00 -1.08889985e+00
7.53320098e-01 1.50980401e+00 2.19733804e-01 5.12383580e-01
-1.40759444e+00 -9.41690981e-01 7.76830763e-02 4.15788352e-01
-1.73591447e+00 -2.31706142e-01 7.75486112e-01 -2.32468471e-01
5.22778034e-01 6.44580841e-01 3.53393346e-01 9.92113352e-01
1.61593899e-01 9.77133930e-01 1.20778775e+00 -6.19262397e-01
1.78338706e-01 4.30339649e-02 5.87074757e-01 3.32483053e-01
2.20797390e-01 1.78342581e-01 -6.76447749e-01 -4.14266169e-01
6.79525360e-02 -1.61565751e-01 -2.00409770e-01 -2.32269287e-01
-9.10457253e-01 8.06341887e-01 4.22151446e-01 2.07790077e-01
-3.13773602e-02 3.44504148e-01 5.95023990e-01 6.56309485e-01
9.80320811e-01 2.64374971e-01 -2.24489540e-01 1.57013789e-01
-5.44394970e-01 2.59941876e-01 4.36678469e-01 6.26750112e-01
5.70952415e-01 -2.64305174e-01 -7.13427186e-01 1.40798163e+00
2.15436116e-01 7.05511093e-01 4.94209468e-01 -7.45870590e-01
8.26623589e-02 5.06567836e-01 -1.26957104e-01 -1.16736758e+00
-1.98546380e-01 -5.84832191e-01 -7.04699993e-01 -4.73567471e-03
-6.80681020e-02 2.26969391e-01 -1.06632185e+00 1.77613282e+00
3.67504776e-01 3.59482646e-01 -1.28634088e-02 9.88291383e-01
1.17540479e+00 4.04909670e-01 1.18597046e-01 -2.58816361e-01
1.15926456e+00 -1.08489740e+00 -9.33856845e-01 2.72593796e-02
6.58650160e-01 -1.03904593e+00 8.80625784e-01 6.80861115e-01
-8.04484248e-01 -4.15625691e-01 -7.61322141e-01 2.71132030e-03
-5.11939645e-01 3.08407247e-01 5.18446445e-01 6.32579148e-01
-8.86799514e-01 1.99189305e-01 -4.40724909e-01 -9.43137258e-02
4.45457518e-01 6.55836105e-01 -2.19124630e-01 -3.45407367e-01
-1.32513738e+00 7.65732348e-01 1.35946706e-01 -5.90359047e-03
-9.37637687e-01 -4.59095657e-01 -6.28361046e-01 -5.19487076e-03
6.51115656e-01 -3.18728149e-01 1.08546174e+00 -8.65952730e-01
-1.10159063e+00 8.27799916e-01 -1.14662260e-01 -4.42571282e-01
5.00190854e-01 -3.51164281e-01 -2.47820109e-01 1.29974231e-01
1.26604903e-02 9.43930149e-01 1.15709043e+00 -1.66149116e+00
-3.52527440e-01 -3.45485002e-01 -2.28280827e-01 2.05294997e-01
-5.53400278e-01 1.03408456e-01 -7.62823045e-01 -9.94978309e-01
2.31940597e-01 -1.20636868e+00 -1.40009105e-01 8.47467035e-03
-4.54471409e-01 -7.75920630e-01 1.09818029e+00 3.31572816e-03
1.29615355e+00 -2.18715692e+00 -3.34878266e-02 5.89945078e-01
2.19474971e-01 5.34312844e-01 -4.62072730e-01 2.12170154e-01
-1.44597813e-01 4.32969451e-01 1.30213469e-01 -6.08093977e-01
-1.90907300e-01 1.49253771e-01 -1.88744336e-01 4.01991427e-01
5.09032011e-02 6.92000866e-01 -1.00354958e+00 -6.62653685e-01
2.34094083e-01 6.08749688e-01 -1.38257951e-01 1.91813886e-01
-1.50991619e-01 4.68524396e-01 -5.66132724e-01 4.57766384e-01
8.98041725e-01 -1.97703183e-01 -6.61685914e-02 -1.97913557e-01
2.97064006e-01 -1.61588937e-02 -1.27889037e+00 1.23296452e+00
-9.37457383e-02 5.54925323e-01 -2.20554113e-01 -8.94194365e-01
9.22981918e-01 1.86806053e-01 4.77622777e-01 -9.10665154e-01
3.41660678e-01 3.30499679e-01 -1.17390506e-01 -6.23877525e-01
2.03946501e-01 2.93191731e-01 1.19930550e-01 4.51519787e-01
-1.46993250e-02 -1.09535426e-01 3.99718285e-01 3.39806080e-01
1.06295991e+00 -4.05953288e-01 -6.33326173e-02 -1.16214603e-01
5.28797507e-01 -3.82988989e-01 4.84421849e-01 1.33062100e+00
-9.68943313e-02 8.28004479e-01 1.06209040e-01 -2.38261402e-01
-6.72653317e-01 -5.21271348e-01 -4.60811019e-01 1.16873741e+00
6.52159333e-01 -5.21869838e-01 -6.86311483e-01 -8.71674657e-01
-7.35544413e-03 2.04748318e-01 -6.40401065e-01 -4.95402932e-01
-4.34839666e-01 -9.47208762e-01 5.73657751e-01 3.43248732e-02
4.29009140e-01 -1.00616205e+00 7.28042573e-02 -1.52512565e-01
-1.38238624e-01 -9.85861480e-01 -6.96503103e-01 3.12446833e-01
-3.68824422e-01 -1.10079181e+00 -7.41935611e-01 -9.37897682e-01
7.98793674e-01 6.50993705e-01 9.92203116e-01 5.54470658e-01
-4.07925665e-01 4.58089173e-01 -6.11686468e-01 -4.31848705e-01
-4.82609197e-02 6.36972338e-02 2.35091895e-02 1.59746438e-01
3.13995421e-01 -1.52270496e-01 -6.62330389e-01 5.08436739e-01
-1.36979985e+00 -5.31603098e-01 4.79176909e-01 1.35536504e+00
9.61062551e-01 2.98235238e-01 5.11756539e-01 -1.20463181e+00
7.72138953e-01 -3.63946795e-01 -6.89655662e-01 4.34190959e-01
-7.15695381e-01 1.59211189e-01 2.45813102e-01 -9.47399259e-01
-8.19581628e-01 7.77673274e-02 1.26262903e-01 -4.10108000e-01
1.36061281e-01 3.42565686e-01 1.47441909e-01 -6.90835774e-01
7.83995032e-01 4.56719287e-02 -3.78631592e-01 -4.85341728e-01
2.60370970e-01 7.35971570e-01 4.13579084e-02 -4.82936919e-01
7.43547618e-01 4.29620653e-01 7.46122897e-02 -6.79003060e-01
-1.04443896e+00 -7.51591980e-01 -1.09465107e-01 -1.69955567e-01
4.46005106e-01 -8.78808320e-01 -6.21435642e-01 6.03772521e-01
-1.11016977e+00 1.03458755e-01 -6.94684833e-02 6.23296976e-01
8.57771486e-02 2.68434674e-01 -5.23379207e-01 -1.06031990e+00
-5.18876731e-01 -1.55497658e+00 1.29908073e+00 1.71867728e-01
1.98117904e-02 -5.85913897e-01 -3.19285393e-02 4.47618335e-01
3.43792021e-01 -1.05027750e-01 8.55481029e-01 -9.46530342e-01
-7.58910239e-01 -2.97832072e-01 -3.10290247e-01 6.31145954e-01
-1.46057829e-01 -3.31150815e-02 -1.00449610e+00 -3.93301696e-01
-1.37718767e-01 -7.53807545e-01 1.39384270e+00 4.55415338e-01
1.46595263e+00 -9.26171616e-02 -4.93846983e-01 2.09409356e-01
1.22730160e+00 2.05979258e-01 5.76509595e-01 2.62819290e-01
4.88110662e-01 4.14678991e-01 8.51680398e-01 2.72440404e-01
4.49527167e-02 8.48929048e-01 2.71507084e-01 -3.17983657e-01
-3.74120772e-01 1.16003059e-01 -4.84523661e-02 8.53651941e-01
6.82857990e-01 -7.41313636e-01 -6.83314562e-01 4.36876744e-01
-1.95156419e+00 -4.27127749e-01 -8.97083133e-02 2.13966775e+00
1.09118414e+00 2.90837020e-01 -3.53395998e-01 3.16710472e-01
8.40987921e-01 1.29234344e-01 -6.30303741e-01 1.50298774e-01
-4.79204923e-01 2.82671869e-01 5.39178073e-01 4.92135048e-01
-1.42016125e+00 1.02838349e+00 6.96707726e+00 1.68482542e+00
-1.08529830e+00 1.34495348e-01 1.02243853e+00 -4.79285419e-03
-2.74514556e-01 -1.65332258e-01 -8.92656684e-01 4.01075244e-01
3.16164970e-01 1.06568635e-01 2.56674290e-01 7.46625304e-01
1.70669138e-01 -4.58817422e-01 -7.69272208e-01 1.24339724e+00
3.62434268e-01 -8.76182854e-01 3.49434435e-01 -2.63978869e-01
1.15638030e+00 4.23486419e-02 4.15147960e-01 -9.79773700e-03
3.31371784e-01 -9.33371723e-01 5.11307180e-01 3.64054739e-01
5.41140199e-01 -6.98273480e-01 1.07484221e+00 2.73523927e-01
-8.05924773e-01 -1.14799455e-01 -3.75524223e-01 3.66462409e-01
-3.39833498e-01 1.03441048e+00 -4.47087735e-01 1.87767088e-01
9.29153562e-01 6.07787371e-01 -9.45253253e-01 1.40474486e+00
-1.62772775e-01 9.14282501e-01 -5.03280640e-01 -1.76696479e-01
4.60039265e-02 1.27718430e-02 4.84696984e-01 1.17673087e+00
-4.55516577e-03 1.19497925e-01 5.52274585e-01 2.32128546e-01
-5.49180210e-01 3.78178775e-01 -6.52499199e-01 1.79682523e-01
5.69269955e-01 1.19629443e+00 -8.25237989e-01 -2.90596515e-01
1.50171071e-01 5.09171903e-01 2.16857105e-01 5.03870130e-01
-4.58357871e-01 -2.88217843e-01 3.40720624e-01 -2.87742734e-01
1.02463827e-01 1.93164140e-01 -2.81539351e-01 -9.62380409e-01
3.67159620e-02 -1.15786421e+00 4.57339942e-01 -5.30292749e-01
-1.61189413e+00 6.35947168e-01 6.14143088e-02 -1.32829690e+00
2.92425990e-01 -7.00417608e-02 -1.55068040e-01 4.84627783e-01
-1.59539843e+00 -8.46376956e-01 -2.67627567e-01 3.43879431e-01
5.32599926e-01 -1.07662573e-01 6.60975993e-01 6.01586998e-01
-6.02766037e-01 9.62453008e-01 4.74635422e-01 2.10488643e-02
1.13014293e+00 -9.40477014e-01 -9.92367417e-02 4.62383598e-01
4.06307101e-01 5.28117418e-01 5.71139276e-01 -5.54930508e-01
-1.10299611e+00 -1.14858174e+00 7.27570713e-01 -9.42846090e-02
2.08328217e-01 -3.60791028e-01 -8.20657670e-01 -3.88628133e-02
5.15679531e-02 3.14562619e-01 6.31257296e-01 -2.66813815e-01
-5.36972582e-01 -4.15779114e-01 -1.26787817e+00 5.17146885e-01
9.04145181e-01 -5.38077593e-01 8.04931074e-02 7.10780740e-01
6.80640459e-01 -4.10802364e-01 -5.27186155e-01 7.09395468e-01
4.22040552e-01 -4.43342507e-01 1.06229639e+00 -2.52415925e-01
-1.14342034e-01 -3.69396269e-01 -1.92617904e-03 -1.25090170e+00
-1.50660321e-01 -5.49011469e-01 2.64420927e-01 1.21461916e+00
3.55259240e-01 -4.24830198e-01 5.90911746e-01 2.76790321e-01
3.63992244e-01 -9.33784246e-01 -1.00771117e+00 -7.68079877e-01
-2.32093051e-01 -2.29502872e-01 1.62159801e-01 7.44053125e-01
-5.32156646e-01 3.75300348e-01 -6.32778049e-01 -1.89036831e-01
6.95712924e-01 -2.41810709e-01 5.28676391e-01 -1.27359343e+00
6.44895360e-02 -2.84526885e-01 -4.98498112e-01 -1.03569245e+00
4.03210342e-01 -7.15394437e-01 3.40254515e-01 -1.08297205e+00
5.06077409e-01 -7.27458298e-01 -5.49796999e-01 5.03852367e-01
-4.55753535e-01 1.01189613e+00 1.68769397e-02 4.64020878e-01
-1.19368267e+00 6.83087587e-01 1.20618820e+00 -5.91386557e-01
6.49734586e-03 -4.99486439e-02 -3.93745154e-01 5.11325419e-01
7.20252097e-01 -1.04302979e+00 -3.86901468e-01 -3.42172205e-01
1.48884431e-01 -3.92095178e-01 4.50641811e-02 -7.42482424e-01
3.01414311e-01 -2.71644443e-04 1.69395924e-01 -6.17933989e-01
4.68401611e-01 -7.54131377e-01 3.98453884e-02 3.37254524e-01
-1.06113815e+00 -4.14093286e-02 -1.25612274e-01 7.77267098e-01
-3.66544843e-01 -5.52545547e-01 7.66874313e-01 1.06012553e-01
-2.79949188e-01 2.98182487e-01 -2.14704126e-01 1.71368927e-01
9.07677054e-01 2.32342541e-01 -4.78664249e-01 -4.72969145e-01
-5.26885748e-01 4.47547555e-01 -1.21727390e-02 4.82569963e-01
6.66684508e-01 -1.43677413e+00 -7.44704247e-01 2.88657956e-02
3.43348384e-01 -2.53561109e-01 -1.04706911e-02 6.16406798e-01
-9.21346247e-02 1.67917281e-01 6.08073950e-01 -8.58081818e-01
-1.96830952e+00 2.22423345e-01 9.92524773e-02 -3.45850348e-01
-2.30795845e-01 1.06821501e+00 -1.31980786e-02 -2.21113518e-01
8.85943711e-01 2.36238867e-01 -4.09150869e-01 1.09363496e-01
5.67670465e-01 1.79730773e-01 3.72193605e-01 -6.99785769e-01
-1.80236876e-01 7.52314329e-01 -3.31214547e-01 3.70310294e-03
1.08691263e+00 -1.43338457e-01 -3.23801041e-01 4.42526668e-01
1.59195650e+00 -1.87079191e-01 -7.39360452e-01 -6.29742622e-01
2.33691454e-01 -7.51210272e-01 3.62266839e-01 -8.57176304e-01
-1.40882146e+00 5.28339982e-01 1.22976732e+00 3.01102340e-01
1.16327345e+00 6.66265115e-02 7.62180448e-01 8.34839642e-01
1.63988143e-01 -1.13619065e+00 4.53037113e-01 4.37324047e-01
6.93084598e-01 -1.55444288e+00 -2.47831549e-02 -7.88718104e-01
-4.14777279e-01 6.12731814e-01 5.03870368e-01 -2.13353440e-01
1.03806102e+00 2.04486370e-01 3.94023001e-01 -1.31818458e-01
-8.81563067e-01 -4.64639008e-01 5.92720866e-01 4.34816122e-01
2.87863493e-01 -2.05157191e-01 -9.58363950e-01 9.33561474e-02
3.19735497e-01 -1.39174491e-01 -5.39878104e-03 8.86920929e-01
-1.98278025e-01 -1.58563328e+00 -4.41671520e-01 8.02082539e-01
-6.99760973e-01 -2.84933031e-01 -6.80940270e-01 2.01015204e-01
3.22135359e-01 1.26435530e+00 -3.71811807e-01 -4.62007612e-01
2.13986278e-01 -1.80737734e-01 2.58303523e-01 -6.04363739e-01
-9.40069556e-01 3.82871836e-01 -4.18806747e-02 -3.63519907e-01
-6.54327512e-01 -1.59880906e-01 -9.26538229e-01 1.34163514e-01
-1.25770748e+00 4.61223900e-01 5.84948897e-01 8.28544140e-01
4.55023110e-01 2.49025255e-01 9.16274428e-01 -6.48379803e-01
-6.71126187e-01 -9.78417635e-01 -4.20709491e-01 8.23351502e-01
2.22617596e-01 -9.86640811e-01 -7.38363385e-01 -1.95563033e-01] | [9.433987617492676, 3.960584878921509] |
7a899998-cbdf-4af5-88db-0b35719cdc8f | identity-aware-cyclegan-for-face-photo-sketch | 2103.16019 | null | https://arxiv.org/abs/2103.16019v1 | https://arxiv.org/pdf/2103.16019v1.pdf | Identity-Aware CycleGAN for Face Photo-Sketch Synthesis and Recognition | Face photo-sketch synthesis and recognition has many applications in digital entertainment and law enforcement. Recently, generative adversarial networks (GANs) based methods have significantly improved the quality of image synthesis, but they have not explicitly considered the purpose of recognition. In this paper, we first propose an Identity-Aware CycleGAN (IACycleGAN) model that applies a new perceptual loss to supervise the image generation network. It improves CycleGAN on photo-sketch synthesis by paying more attention to the synthesis of key facial regions, such as eyes and nose, which are important for identity recognition. Furthermore, we develop a mutual optimization procedure between the synthesis model and the recognition model, which iteratively synthesizes better images by IACycleGAN and enhances the recognition model by the triplet loss of the generated and real samples. Extensive experiments are performed on both photo-tosketch and sketch-to-photo tasks using the widely used CUFS and CUFSF databases. The results show that the proposed method performs better than several state-of-the-art methods in terms of both synthetic image quality and photo-sketch recognition accuracy. | ['Weihong Deng', 'Jiani Hu', 'Yuke Fang'] | 2021-03-30 | null | null | null | null | ['sketch-recognition'] | ['computer-vision'] | [ 5.64484835e-01 6.20134622e-02 5.03926128e-02 -2.54024982e-01
-5.36550701e-01 -3.84133935e-01 8.95719111e-01 -9.57470000e-01
3.83563153e-02 6.21633589e-01 3.75069641e-02 1.29287392e-01
2.55929917e-01 -8.50712836e-01 -8.17029715e-01 -9.38484967e-01
5.81690133e-01 6.71404824e-02 -3.59954298e-01 -7.56977051e-02
3.20151038e-02 5.99060595e-01 -1.50421250e+00 3.15735251e-01
7.74417579e-01 1.09607732e+00 -3.33555967e-01 5.00429690e-01
1.42082378e-01 6.85080290e-01 -5.97992778e-01 -8.54632676e-01
5.92590988e-01 -9.04147148e-01 -1.72305435e-01 4.01960015e-01
6.64983034e-01 -4.98337388e-01 -6.20402575e-01 1.24276996e+00
6.33242071e-01 -9.29980502e-02 6.76826835e-01 -1.58822501e+00
-1.06273198e+00 2.00330526e-01 -6.09402239e-01 -3.85168612e-01
3.17264259e-01 4.24440742e-01 6.12294257e-01 -1.09614587e+00
6.13748968e-01 1.60177612e+00 4.10380036e-01 9.40771759e-01
-1.11978662e+00 -1.22961068e+00 -2.56530136e-01 2.81836241e-01
-1.64894557e+00 -8.56794953e-01 1.01796079e+00 -2.20692292e-01
3.43778312e-01 2.74424970e-01 5.44676781e-01 1.15721464e+00
-1.05266690e-01 8.90573442e-01 1.11179769e+00 -4.72671449e-01
-5.02363369e-02 1.11973308e-01 -7.32456744e-01 7.09507406e-01
-5.82274757e-02 4.79641229e-01 -3.36319625e-01 1.00749977e-01
1.27720141e+00 6.23396225e-02 -2.98549533e-01 -1.49124071e-01
-9.03186560e-01 6.72846854e-01 2.61566877e-01 5.95183633e-02
-4.98780131e-01 2.01460272e-01 -5.66301262e-03 3.18290740e-01
2.29060903e-01 1.40518785e-01 4.51393843e-01 2.74575472e-01
-9.45007443e-01 1.01587772e-01 5.52239120e-01 9.39887047e-01
6.65809572e-01 6.00197732e-01 -3.40413272e-01 1.07480669e+00
2.88470358e-01 8.18133533e-01 2.91573822e-01 -9.79088426e-01
3.48965883e-01 6.27760172e-01 -6.96047246e-02 -1.21257353e+00
4.27741826e-01 -6.26883507e-02 -1.12899911e+00 4.05677766e-01
1.56688094e-01 -4.94907051e-02 -1.01356769e+00 1.58238828e+00
3.54024500e-01 5.82169056e-01 1.55510560e-01 8.91574144e-01
1.02699280e+00 7.56965458e-01 -3.74102928e-02 -3.53640392e-02
1.01187348e+00 -1.09109461e+00 -8.35620224e-01 -7.18239620e-02
-1.05728723e-01 -9.90468740e-01 8.24618757e-01 2.75088042e-01
-1.26169193e+00 -8.79509032e-01 -1.15524459e+00 1.57711521e-01
3.06393895e-02 5.74638784e-01 2.89626032e-01 8.74485910e-01
-8.90672922e-01 5.35932124e-01 -3.47584575e-01 -5.51967733e-02
8.01315129e-01 3.94287795e-01 -5.93468666e-01 -3.30537945e-01
-1.01747167e+00 4.84162807e-01 1.50232986e-01 2.62867987e-01
-1.08788610e+00 -5.65659344e-01 -7.91307390e-01 8.70084316e-02
3.20319355e-01 -6.30201399e-01 8.08867097e-01 -1.52346802e+00
-1.92303896e+00 9.39347625e-01 3.17733735e-02 -7.94936046e-02
6.69215620e-01 1.99540213e-01 -6.47575200e-01 1.98337823e-01
-2.77934283e-01 9.70828593e-01 1.51849365e+00 -1.32285666e+00
-3.00204605e-01 -3.03943604e-01 -6.38576671e-02 6.11198135e-02
-1.53859362e-01 -1.96644533e-02 -6.43234789e-01 -9.40037727e-01
-3.20038587e-01 -9.49601114e-01 2.27701813e-01 3.38974714e-01
-4.82041001e-01 1.37455389e-02 1.03713059e+00 -8.45023513e-01
8.27886641e-01 -2.30671549e+00 5.30283414e-02 2.87460387e-01
-9.17986035e-02 7.37684965e-01 -5.97360134e-01 4.45391208e-01
-2.20346749e-01 -1.68594103e-02 -2.10377634e-01 -4.45886582e-01
-8.08161572e-02 2.78203219e-01 -3.36044461e-01 4.88326371e-01
4.68175262e-01 1.14419651e+00 -5.35125613e-01 -4.17446494e-01
2.80084759e-01 7.13403344e-01 -4.69895780e-01 4.25477773e-01
-8.11559409e-02 5.25606573e-01 -2.01732188e-01 8.41775835e-01
1.06994510e+00 1.74681190e-02 1.77724257e-01 -2.12947279e-01
3.52974325e-01 -4.94321764e-01 -1.05488765e+00 1.30614316e+00
-3.82705599e-01 4.51106459e-01 8.32049642e-03 -6.30331695e-01
1.09087467e+00 3.88242215e-01 2.50949949e-01 -7.15473473e-01
2.04271376e-01 1.66197658e-01 7.03857541e-02 -3.11329991e-01
1.25932589e-01 -1.72699302e-01 3.75454217e-01 3.50410074e-01
9.37631056e-02 -2.86882311e-01 -4.52960916e-02 -1.55292168e-01
5.63639700e-01 1.97850749e-01 1.57892525e-01 1.43278137e-01
1.04355979e+00 -4.93435264e-01 5.45002937e-01 2.65390307e-01
6.62366599e-02 8.44850779e-01 4.73515868e-01 -1.38396055e-01
-1.38010859e+00 -9.35132384e-01 3.22711259e-01 4.07336116e-01
1.44497409e-01 2.12223399e-02 -1.13473368e+00 -6.71359360e-01
-5.01028858e-02 5.39353073e-01 -6.22891426e-01 -3.01990420e-01
-5.06334722e-01 -2.79313296e-01 8.48063648e-01 3.37708175e-01
9.16043282e-01 -1.39614534e+00 2.23548889e-01 -7.55428895e-02
1.93985626e-02 -1.27246749e+00 -9.10784364e-01 -1.07725513e+00
-3.77487630e-01 -1.15799677e+00 -1.08696532e+00 -8.50093126e-01
1.12935138e+00 9.13830251e-02 5.60987294e-01 3.01133454e-01
-4.53259617e-01 3.18008274e-01 -1.84741139e-01 -3.32658261e-01
-7.99586475e-01 -3.41364622e-01 -8.66322666e-02 8.82421494e-01
2.71824538e-03 -5.68926215e-01 -6.58555984e-01 6.20368838e-01
-1.07423878e+00 8.77545327e-02 7.23211110e-01 1.10524583e+00
4.86298501e-01 -1.22450627e-01 5.85669637e-01 -8.86618495e-01
6.06023192e-01 -1.18874960e-01 -6.62007511e-01 4.96974736e-01
-6.01182818e-01 -7.36004114e-02 7.26067483e-01 -6.13839924e-01
-1.20962262e+00 1.93318129e-01 -3.19350988e-01 -1.00524688e+00
3.62489782e-02 -2.11489230e-01 -5.65337598e-01 -6.13008618e-01
2.63346761e-01 6.24943674e-01 4.44903284e-01 -2.11432770e-01
2.43275866e-01 6.77114248e-01 7.39681125e-01 -3.35955322e-01
1.11844289e+00 4.99521911e-01 1.12437636e-01 -9.29091096e-01
-1.77158430e-01 4.14680839e-02 -2.40028754e-01 -3.36900979e-01
7.05487728e-01 -8.32709491e-01 -9.04130578e-01 1.01966941e+00
-1.11065757e+00 -1.23891108e-01 -2.21074894e-01 1.89718112e-01
-5.57753742e-01 5.78260481e-01 -2.88160175e-01 -8.49615812e-01
-6.11489534e-01 -1.16458154e+00 1.15974569e+00 5.58145165e-01
3.24570537e-01 -6.67160511e-01 -2.51468152e-01 2.56698042e-01
3.43269169e-01 4.12488401e-01 7.26473510e-01 -1.56772867e-01
-8.08507085e-01 -2.71997929e-01 -4.57092434e-01 7.94837892e-01
2.59849757e-01 1.41327292e-01 -9.85990524e-01 -2.20508382e-01
-3.17519933e-01 -2.43945122e-01 6.69887781e-01 2.01419299e-03
1.18256807e+00 -6.26738966e-01 -6.83669746e-02 9.15292978e-01
1.35680437e+00 5.60806930e-01 1.34784544e+00 -2.46759981e-01
7.60783970e-01 5.32604158e-01 4.39380020e-01 2.78015614e-01
9.64123663e-03 7.74681151e-01 3.99705619e-01 -2.98062563e-01
-5.00823796e-01 -6.78946793e-01 4.23476964e-01 4.56968188e-01
-1.68980539e-01 -3.72399837e-01 -2.37033099e-01 3.43676150e-01
-1.67905295e+00 -1.22189534e+00 2.83415526e-01 2.05883908e+00
6.42546415e-01 -4.30895358e-01 -9.33499336e-02 1.61256745e-01
9.49781239e-01 2.58122921e-01 -6.00282371e-01 -3.17991704e-01
-1.94554612e-01 4.84947115e-01 1.74163356e-01 3.82755101e-01
-8.24379206e-01 1.04769027e+00 5.87830067e+00 1.15323460e+00
-1.26232755e+00 -9.12383869e-02 7.61485815e-01 2.45937809e-01
-3.05494517e-01 -1.61147699e-01 -5.46933413e-01 5.57217658e-01
2.39625141e-01 -9.97035727e-02 9.33301210e-01 8.01082909e-01
-2.06275135e-02 2.62215942e-01 -9.72916663e-01 1.29488850e+00
3.68743837e-01 -1.40919185e+00 4.37467784e-01 1.39909804e-01
1.00752437e+00 -7.74722099e-01 3.30776900e-01 1.88524090e-03
-1.04454868e-02 -1.25724876e+00 6.09752059e-01 8.39634180e-01
1.33591592e+00 -8.13205838e-01 5.57236373e-01 1.33112475e-01
-9.80219126e-01 8.60843882e-02 -3.05680007e-01 4.63622630e-01
-2.18454394e-02 1.49709463e-01 -7.44707704e-01 5.03825068e-01
1.45865783e-01 6.40726864e-01 -3.78784418e-01 6.82735741e-01
-6.70374870e-01 3.88928980e-01 -4.65883128e-02 1.33552566e-01
2.62376312e-02 -4.41075176e-01 5.18656254e-01 7.08562195e-01
4.47190821e-01 2.22846210e-01 -2.70099431e-01 1.19337630e+00
-4.39429104e-01 8.63407552e-02 -6.71341836e-01 -3.47904861e-01
4.86426860e-01 1.19405842e+00 -1.86186060e-01 -3.11612278e-01
-1.13183767e-01 1.17646861e+00 -2.86646903e-01 3.08421791e-01
-8.67626846e-01 -4.91337419e-01 8.17087948e-01 7.19989017e-02
4.17092949e-01 1.02415219e-01 2.87969466e-02 -1.06219316e+00
-1.12551572e-02 -1.23765194e+00 -1.32307634e-02 -8.44705105e-01
-1.10260296e+00 6.01596534e-01 -3.99854541e-01 -1.24896967e+00
-5.04293978e-01 -3.82296383e-01 -8.09567809e-01 1.01390195e+00
-1.48185813e+00 -1.57326853e+00 -5.42521477e-01 7.59539008e-01
5.18172562e-01 -6.64274573e-01 6.42621279e-01 5.34626722e-01
-5.60146570e-01 1.13756776e+00 -4.56443876e-02 3.37720782e-01
6.78749442e-01 -4.70043957e-01 6.50483489e-01 9.94484067e-01
1.06876150e-01 4.43118155e-01 2.37660110e-01 -5.55601716e-01
-1.53203082e+00 -1.20709395e+00 6.27118587e-01 -4.47632596e-02
-1.60415284e-03 -2.92113990e-01 -5.62500358e-01 3.47323477e-01
2.68099189e-01 1.46883592e-01 2.57510453e-01 -8.25229466e-01
-2.46878177e-01 -3.77151668e-01 -1.57439482e+00 6.76280797e-01
1.10206378e+00 -5.20622075e-01 -2.54337303e-03 4.70052510e-02
3.15144867e-01 -2.96726704e-01 -7.01074898e-01 4.36203539e-01
9.30014431e-01 -8.77748072e-01 1.11763561e+00 -3.03020358e-01
6.37718379e-01 -3.46809953e-01 5.62509522e-03 -1.20450985e+00
-9.46967751e-02 -7.76956856e-01 3.68122570e-02 1.61190593e+00
3.80770825e-02 -7.36164391e-01 8.58267009e-01 2.89935976e-01
2.36581862e-01 -4.76407140e-01 -7.04112530e-01 -8.07188570e-01
-3.40880871e-01 1.06811244e-02 1.06123865e+00 8.79238069e-01
-6.40525579e-01 1.19818367e-01 -9.24161732e-01 -8.81943628e-02
7.60240316e-01 2.89756600e-02 1.20767546e+00 -7.27918684e-01
-1.44422129e-01 -3.23473513e-01 -5.09307444e-01 -7.64126360e-01
2.79080600e-01 -6.71617687e-01 -1.50770470e-01 -1.05646718e+00
9.20399465e-03 -2.58635938e-01 1.32995799e-01 4.76188689e-01
-1.47489280e-01 6.81951582e-01 4.72072929e-01 1.58851177e-01
6.61510509e-03 7.33888507e-01 1.80310416e+00 -3.16937864e-01
1.19507290e-01 2.32752468e-02 -6.27245069e-01 3.22723329e-01
5.79635262e-01 -3.38678122e-01 -4.15370375e-01 -2.09323987e-01
-2.69688576e-01 1.66876271e-01 5.39752066e-01 -9.55050588e-01
9.03420374e-02 -1.90729201e-01 5.38301647e-01 -2.70413220e-01
6.60000563e-01 -8.23629260e-01 7.00620115e-01 4.82286513e-01
-2.22897708e-01 -2.07474127e-01 5.32920882e-02 4.02111143e-01
-4.17712063e-01 -3.84282060e-02 1.18239355e+00 -9.37551335e-02
-5.39768815e-01 5.91726780e-01 1.24594964e-01 -2.27163613e-01
1.19976199e+00 -4.13500786e-01 -2.04965353e-01 -7.66570151e-01
-3.89565170e-01 -4.36440222e-02 5.73690474e-01 5.74310899e-01
1.00750995e+00 -1.81358874e+00 -1.00073719e+00 8.19683552e-01
1.20678926e-02 -4.21392679e-01 4.12563294e-01 4.98611867e-01
-5.91404021e-01 3.17475677e-01 -5.80763459e-01 -3.55937272e-01
-1.46582031e+00 4.92058069e-01 3.42767000e-01 3.59442197e-02
-3.12757432e-01 7.53957629e-01 5.06674647e-01 -2.96749741e-01
1.82009712e-01 2.03107625e-01 -3.48974317e-02 -3.14856499e-01
5.77257752e-01 5.02511144e-01 -2.60019094e-01 -8.51190329e-01
-1.44067749e-01 7.11700141e-01 6.52587563e-02 -2.65069213e-02
1.09029543e+00 9.66805294e-02 -1.05124444e-01 -4.69596207e-01
1.23267293e+00 -1.42387918e-03 -1.48610055e+00 -1.66031301e-01
-7.42304623e-01 -9.31684911e-01 -1.91521093e-01 -7.42565215e-01
-1.57838285e+00 9.02422130e-01 6.43839240e-01 -2.93731004e-01
1.34737134e+00 -4.80977714e-01 1.00046384e+00 -1.20662540e-01
2.19583035e-01 -8.84704709e-01 2.82078266e-01 5.24181724e-02
1.21559381e+00 -9.93580043e-01 -1.07678562e-01 -5.87282419e-01
-7.73246467e-01 1.02445650e+00 6.10708952e-01 -3.59115750e-01
4.01983500e-01 -5.64455837e-02 -8.03823844e-02 2.21198291e-01
-3.90380770e-01 4.28621843e-02 6.33486331e-01 7.05198109e-01
-1.32621720e-01 1.54279394e-03 -2.15138167e-01 3.86587381e-01
2.24512815e-02 1.60592467e-01 1.97917983e-01 4.75473225e-01
1.27461463e-01 -1.39154959e+00 -5.49433470e-01 1.43249229e-01
-4.13605779e-01 -2.84048766e-02 -7.33536839e-01 7.02862799e-01
2.97655106e-01 7.49272346e-01 -1.34093001e-01 -5.34085155e-01
1.90061107e-01 -5.60289770e-02 8.41277659e-01 -2.99096674e-01
-4.17112142e-01 -9.78731588e-02 9.30930674e-03 -4.76220548e-01
-3.85893047e-01 -5.09272575e-01 -7.45874166e-01 -5.78326941e-01
-3.99826318e-01 -7.31394216e-02 7.24320412e-01 6.16290987e-01
4.63220954e-01 3.04389983e-01 1.13396871e+00 -8.78611088e-01
-5.17800868e-01 -7.76151121e-01 -6.31194472e-01 4.28688616e-01
1.12778820e-01 -4.97047722e-01 -8.89277384e-02 2.79049635e-01] | [12.4877347946167, -0.07272697240114212] |
9909093b-9b70-4665-9ec5-1b5ef331bdc4 | an-iterative-unbiased-geometric-approach-to | 2212.02421 | null | https://arxiv.org/abs/2212.02421v3 | https://arxiv.org/pdf/2212.02421v3.pdf | Score-based denoising for atomic structure identification | We propose an effective method for removing thermal vibrations that complicate the task of analyzing complex dynamics in atomistic simulation of condensed matter. Our method iteratively subtracts thermal noises or perturbations in atomic positions using a denoising score function trained on synthetically noised but otherwise perfect crystal lattices. The resulting denoised structures clearly reveal underlying crystal order while retaining disorder associated with crystal defects. Purely geometric, agnostic to interatomic potentials, and trained without inputs from explicit simulations, our denoiser can be applied to simulation data generated from vastly different interatomic interactions. The denoiser is shown to improve existing classification methods such as common neighbor analysis and polyhedral template matching, reaching perfect classification accuracy on a recent benchmark dataset of thermally perturbed structures up to the melting point. Demonstrated here in a wide variety of atomistic simulation contexts, the denoiser is general, robust, and readily extendable to delineate order from disorder in structurally and chemically complex materials. | ['Fei Zhou', 'Vasily Bulatov', 'James Chapman', 'Cheol Woo Park', 'Nicolas Bertin', 'Babak Sadigh', 'Tim Hsu'] | 2022-12-05 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [ 6.33576751e-01 -2.43061453e-01 2.65235424e-01 -2.22870439e-01
-9.64317441e-01 -7.38371491e-01 7.80143559e-01 6.60637617e-02
-4.30637509e-01 1.08371449e+00 3.20617467e-01 -1.22395366e-01
1.94635615e-02 -6.99918151e-01 -8.46053481e-01 -1.40881670e+00
-3.33112516e-02 1.08651245e+00 1.92287713e-01 -5.62849939e-01
3.75953645e-01 6.52050853e-01 -1.54972827e+00 3.33583683e-01
8.91558230e-01 5.37280977e-01 -8.05285871e-02 5.99141061e-01
3.59322041e-01 2.68464148e-01 -2.18155906e-01 4.77128811e-02
1.88037872e-01 -4.88296419e-01 -7.39090025e-01 -1.60509214e-01
3.68114442e-01 8.31230879e-02 -3.00874323e-01 9.25382555e-01
6.19472802e-01 3.65493119e-01 1.14744842e+00 -2.15007633e-01
-5.32543838e-01 3.13782483e-01 -7.84157217e-02 1.09353334e-01
4.71822560e-01 6.50879502e-01 9.78598952e-01 -8.19028258e-01
1.13388908e+00 1.20178699e+00 6.89342797e-01 6.21650875e-01
-1.92230880e+00 -4.64547873e-01 -2.73014665e-01 -1.33075342e-01
-8.47905636e-01 -5.55939794e-01 5.06549001e-01 -4.86768931e-01
1.58403611e+00 4.82210547e-01 4.34615105e-01 1.23770297e+00
7.87008524e-01 -1.26205385e-01 1.13566005e+00 -2.76791483e-01
5.33695400e-01 -8.90973330e-01 1.99106634e-01 3.20894301e-01
3.39008689e-01 2.60313362e-01 -3.21501464e-01 -8.41519296e-01
3.86185348e-01 -4.55317870e-02 -2.13049293e-01 -6.19465768e-01
-1.17728686e+00 4.96246338e-01 2.43052036e-01 6.09873086e-02
-3.20595592e-01 2.73782611e-01 5.41740477e-01 3.58782351e-01
2.94938207e-01 8.88660312e-01 -2.67258614e-01 -1.76487595e-01
-7.16697156e-01 9.25604939e-01 5.67806005e-01 3.90151083e-01
6.96993768e-01 2.19610240e-02 3.10741514e-01 4.08839911e-01
3.93595174e-02 5.44905007e-01 2.81596601e-01 -1.19919300e+00
9.72370803e-02 1.11934409e-01 3.50194424e-01 -3.40479344e-01
-4.55094784e-01 -2.57473085e-02 -1.09927821e+00 5.23215413e-01
2.51188070e-01 1.98810562e-01 -1.21980774e+00 1.48818886e+00
5.04937172e-01 -2.88273245e-01 7.76398703e-02 8.32901001e-01
6.73032463e-01 5.13521791e-01 -5.49855195e-02 -4.34029818e-01
1.09932244e+00 -4.65403140e-01 -4.61585581e-01 -1.02231875e-01
5.38996041e-01 -7.11807847e-01 6.72150135e-01 6.95961952e-01
-1.38448381e+00 -2.56306052e-01 -1.23048270e+00 -2.33812988e-01
6.88464707e-03 -7.25730360e-01 6.79947555e-01 3.85345876e-01
-7.70768583e-01 1.41259909e+00 -1.35547090e+00 -2.61306129e-02
1.22310169e-01 7.31779695e-01 -5.97109616e-01 -4.17694673e-02
-1.11187088e+00 7.27159381e-01 2.56906480e-01 -9.13821310e-02
-9.35609758e-01 -7.25296736e-01 -5.47257721e-01 -2.96485931e-01
1.60950556e-01 -9.29793894e-01 1.18484771e+00 -6.53491795e-01
-1.33599532e+00 7.08664656e-01 -6.43743575e-01 -5.00315785e-01
4.03263628e-01 2.19134361e-01 -3.19166303e-01 -8.96348506e-02
9.93081778e-02 -1.95991267e-02 7.26618648e-01 -1.20228314e+00
4.00981277e-01 -4.44099963e-01 -6.50433540e-01 2.01450035e-01
4.04733866e-01 -1.91923290e-01 8.26060176e-02 -6.76396966e-01
6.33273661e-01 -9.65294957e-01 -8.77113700e-01 -5.71402371e-01
-3.41575712e-01 1.73363224e-01 5.02916038e-01 -4.91110504e-01
8.98221254e-01 -1.75542462e+00 7.63205469e-01 8.52483273e-01
6.34444714e-01 3.19120660e-02 1.89841688e-01 8.85219634e-01
-5.64963698e-01 1.16286688e-01 -8.07121575e-01 3.65282260e-02
-6.49232119e-02 1.66636631e-01 -5.49155138e-02 5.16558588e-01
2.56204605e-02 8.77630889e-01 -7.80641139e-01 2.13690892e-01
-2.87348963e-02 3.61004800e-01 -7.83219218e-01 -4.33461107e-02
-5.96584737e-01 7.81458080e-01 -4.04148787e-01 3.29567432e-01
7.12263644e-01 -2.41702765e-01 5.77990472e-01 -3.78103810e-04
2.07156185e-02 6.45494998e-01 -1.01625872e+00 1.51323259e+00
7.70721659e-02 4.59549157e-03 3.80958110e-01 -8.03511977e-01
8.05391550e-01 2.83056080e-01 5.77860534e-01 -6.86657250e-01
-1.40709758e-01 4.10592884e-01 5.26383400e-01 -2.21789807e-01
3.21392953e-01 -6.84698761e-01 -2.62413591e-01 7.07872987e-01
-4.99216318e-02 -5.52923739e-01 4.61039841e-02 2.25810081e-01
1.57810199e+00 -4.94410880e-02 3.18337172e-01 -5.59447289e-01
3.86124849e-01 1.23810083e-01 4.40735668e-01 7.01276481e-01
1.62286554e-02 9.01321650e-01 4.28411931e-01 -7.53376782e-01
-1.83966780e+00 -1.20894384e+00 -3.18988830e-01 8.14691722e-01
-1.10433489e-01 -7.02996254e-01 -8.17512393e-01 7.45342905e-03
2.24210277e-01 1.43430635e-01 -4.91034269e-01 -4.57576603e-01
-8.00788164e-01 -1.39009261e+00 1.99385926e-01 4.57599871e-02
-3.84425037e-02 -1.23912859e+00 -4.72165318e-03 5.79402685e-01
1.79275304e-01 -7.40089297e-01 -4.11966860e-01 7.22134829e-01
-8.79865706e-01 -1.15607166e+00 -1.99707881e-01 -4.02955502e-01
4.77373004e-01 -2.40283906e-02 1.30844390e+00 1.77319780e-01
-5.28377414e-01 -1.13219246e-01 8.75427350e-02 2.48576790e-01
-9.92655396e-01 3.59436348e-02 5.37632823e-01 -5.38403451e-01
2.24128455e-01 -1.14102972e+00 -6.10514939e-01 1.24234691e-01
-9.18392479e-01 -2.65919834e-01 1.56133443e-01 9.49835479e-01
6.78185940e-01 6.48488551e-02 1.85817495e-01 -1.03013289e+00
6.93677127e-01 -2.50520438e-01 -3.95116061e-01 -1.72911778e-01
-5.83926499e-01 7.54791558e-01 8.80250633e-01 1.09321559e-02
-9.80980992e-01 1.88840121e-01 -4.53628242e-01 1.08928867e-01
-2.81997144e-01 6.42955154e-02 -2.75084913e-01 -3.35397393e-01
9.58162427e-01 3.05727512e-01 1.80356324e-01 -5.09496868e-01
7.74711892e-02 2.05339387e-01 8.92849028e-01 -1.07647228e+00
8.68289471e-01 6.71367228e-01 4.74726319e-01 -8.51368785e-01
-1.71402693e-01 -2.31624007e-01 -8.38885963e-01 2.68439174e-01
7.78870583e-01 -6.71294332e-01 -1.16661012e+00 4.74413872e-01
-9.35246944e-01 -2.71512717e-01 -3.31743024e-02 2.63596207e-01
-6.54667556e-01 8.59473109e-01 -8.51298273e-01 -5.12511969e-01
-4.68322068e-01 -1.46275532e+00 1.05343950e+00 -4.02988702e-01
-6.87183738e-01 -7.73046374e-01 4.54347551e-01 3.76209617e-01
1.30514443e-01 6.73409581e-01 1.40882444e+00 -4.32033926e-01
-6.72033608e-01 4.97421362e-02 4.79706913e-01 6.80041090e-02
1.77306533e-01 2.69386739e-01 -7.28677988e-01 -4.03498769e-01
4.26783375e-02 -3.92160267e-01 1.34392905e+00 2.83861101e-01
8.08862805e-01 -1.76117450e-01 -3.67449105e-01 5.06845951e-01
1.17957079e+00 2.43402272e-01 8.45820963e-01 2.75315523e-01
9.14470911e-01 4.19602841e-01 9.21396092e-02 2.28606015e-01
-3.62278402e-01 6.37067080e-01 2.68889129e-01 1.56836674e-01
3.61621797e-01 1.76069513e-01 4.42705154e-01 8.14129114e-01
-5.97809732e-01 -8.37015212e-02 -1.20938194e+00 9.48930010e-02
-1.66015613e+00 -1.27161610e+00 -3.49351436e-01 2.16186953e+00
1.19039524e+00 6.07650042e-01 9.20041502e-02 1.32477269e-01
4.32780057e-01 2.28093237e-01 -1.03323495e+00 -5.47254086e-01
-2.84281373e-01 8.52009714e-01 6.03022933e-01 8.18347275e-01
-1.02110672e+00 8.75424743e-01 7.62563276e+00 7.73904741e-01
-8.40146542e-01 5.46635613e-02 4.40561295e-01 -3.26850891e-01
-5.69125235e-01 1.23726055e-01 -3.80379349e-01 3.99261028e-01
1.16615891e+00 7.90393725e-02 6.29745007e-01 4.93408114e-01
3.85586977e-01 8.40568617e-02 -1.25240040e+00 5.89605212e-01
-6.16094232e-01 -1.78274751e+00 9.11745504e-02 8.00116509e-02
9.75002825e-01 2.53761142e-01 1.54236527e-02 -2.78677553e-01
6.28135502e-01 -1.55498528e+00 3.63875359e-01 5.96652627e-01
7.21401215e-01 -9.39535558e-01 3.45605701e-01 1.74823359e-01
-8.87926400e-01 5.15271425e-01 -2.76485711e-01 -3.33171874e-01
1.22104183e-01 7.46519506e-01 -5.86760104e-01 3.76551658e-01
5.86772442e-01 7.11030900e-01 -7.49196112e-02 3.27638030e-01
2.90466964e-01 4.99957561e-01 -4.38569248e-01 2.65607953e-01
2.87559599e-01 -6.90410912e-01 4.85649735e-01 7.56503820e-01
-1.26234531e-01 4.48016405e-01 4.20833714e-02 7.75168359e-01
-8.70669037e-02 -4.75407660e-01 -6.17715776e-01 -1.29835820e-02
3.34308952e-01 8.24595869e-01 -7.66310394e-01 -1.45354465e-01
1.50900083e-02 9.76551712e-01 1.77184209e-01 5.14906228e-01
-4.32563156e-01 -1.05929360e-01 1.22115171e+00 4.90619838e-01
3.59713674e-01 -4.36302513e-01 -1.87927946e-01 -9.43952739e-01
1.61869541e-01 -1.23457301e+00 -3.00784379e-01 -3.67354453e-01
-1.35776532e+00 3.12130153e-01 -1.95283532e-01 -5.74361503e-01
-3.45676720e-01 -8.01824749e-01 -8.10803771e-01 9.04181838e-01
-7.67319798e-01 -3.65491807e-01 6.26835749e-02 5.21963798e-02
-9.19975564e-02 8.44217837e-02 8.52328360e-01 -4.76102792e-02
-4.79160666e-01 2.58676946e-01 1.07819414e+00 -4.52812731e-01
7.93546379e-01 -1.33155394e+00 1.23204458e+00 4.88454878e-01
-4.64549512e-01 9.01143849e-01 1.37666571e+00 -1.07277763e+00
-1.59545958e+00 -9.83429193e-01 2.92105705e-01 -6.20323956e-01
5.64511895e-01 -8.78292382e-01 -1.37969697e+00 5.06711721e-01
-6.28167242e-02 -5.47484122e-02 4.43992615e-01 -2.00809360e-01
-3.89092296e-01 4.08911854e-01 -1.28042746e+00 5.85881174e-01
1.47462404e+00 -6.47397339e-01 -5.83904743e-01 7.42976069e-01
4.85182524e-01 -4.03107047e-01 -1.10997081e+00 4.43976730e-01
5.38438380e-01 -1.06949997e+00 1.36199033e+00 -9.11567092e-01
3.27240169e-01 -6.81621209e-02 -2.23539397e-02 -1.04477787e+00
-5.67394495e-01 -1.15118599e+00 1.54327214e-01 5.18436313e-01
3.52656960e-01 -7.05704451e-01 1.06408668e+00 7.84661949e-01
-2.89062411e-01 -5.68528593e-01 -1.23518288e+00 -7.67857194e-01
5.66191137e-01 -7.76420310e-02 3.71721238e-01 7.62365162e-01
-6.39461353e-02 2.48983800e-01 -1.21282928e-01 -9.44356842e-04
8.62755597e-01 -7.97206834e-02 5.54064572e-01 -1.35858583e+00
-3.26077431e-01 -2.61359423e-01 -3.53728503e-01 -6.37007654e-01
2.78167427e-01 -1.08356023e+00 -2.24301405e-03 -9.66515183e-01
3.50079447e-01 -2.06540465e-01 1.81138784e-01 1.34261757e-01
6.97938129e-02 2.56782860e-01 -3.51296872e-01 5.01923561e-01
-4.34855282e-01 9.39236224e-01 1.31353593e+00 -7.85438810e-03
4.96243918e-03 -3.41673940e-01 -3.61987382e-01 6.54357851e-01
5.99358976e-01 -7.42483377e-01 2.36913450e-02 3.84168893e-01
3.77486765e-01 -1.37768969e-01 3.62627327e-01 -1.07935750e+00
-1.09214641e-01 -1.27050281e-01 3.48103821e-01 -4.19666797e-01
3.32296550e-01 -5.16082048e-01 7.07204878e-01 7.39472747e-01
-9.95086432e-02 3.52350287e-02 -1.60306171e-02 7.00638533e-01
9.21361148e-03 -6.98779337e-03 9.58017945e-01 -4.64076072e-01
-1.63485274e-01 1.85484946e-01 -7.48462677e-01 1.13236725e-01
7.83591211e-01 -2.60134041e-01 -2.00074241e-01 -1.60120517e-01
-1.12599766e+00 -8.46810266e-03 1.22002184e+00 -3.23223174e-01
3.08130264e-01 -1.19879901e+00 -4.67745274e-01 3.35759759e-01
-1.55114025e-01 1.61925554e-01 3.01148653e-01 5.00817895e-01
-8.47262919e-01 4.87478942e-01 -1.19720422e-01 -6.29782259e-01
-1.29362094e+00 4.49977607e-01 5.83839536e-01 -3.44335020e-01
-4.66194004e-01 5.20780027e-01 2.11473927e-01 -6.83846712e-01
-3.59548062e-01 -6.95648670e-01 5.64739645e-01 -3.20907772e-01
1.66754246e-01 3.24181408e-01 6.24415994e-01 -6.18467867e-01
-3.48421514e-01 6.15692377e-01 -4.67646718e-01 5.13426624e-02
1.61145413e+00 2.29288474e-01 -3.97491455e-01 2.01258466e-01
1.19060040e+00 -8.08561891e-02 -1.47543287e+00 -1.27444893e-01
5.68177551e-02 2.57162243e-01 -3.71071398e-01 -4.03420836e-01
-4.21113521e-01 5.02030075e-01 3.36977869e-01 4.06826893e-03
6.22763753e-01 3.48825417e-02 9.32743907e-01 8.72819841e-01
2.86861062e-01 -8.20849895e-01 -5.74913397e-02 7.19475567e-01
6.56940460e-01 -1.14623857e+00 2.88164258e-01 -2.82449067e-01
-1.34088472e-01 1.14351177e+00 2.26121008e-01 -5.24563432e-01
3.76622409e-01 4.72520888e-01 -4.04732764e-01 -4.70205814e-01
-8.43770385e-01 3.05486232e-01 1.21456802e-01 5.32632053e-01
4.31817442e-01 -2.03878596e-01 -2.06895709e-01 2.09484592e-01
-4.06072676e-01 -6.62025750e-01 6.27175212e-01 1.27456570e+00
-5.97435713e-01 -1.54385364e+00 -6.25042856e-01 4.92482394e-01
-4.09321427e-01 -3.81441981e-01 -7.58260906e-01 6.34401023e-01
-1.21144861e-01 3.76862228e-01 2.24878807e-02 -1.04672387e-01
2.16906667e-01 4.47095573e-01 6.48649871e-01 -6.57037735e-01
-8.12635779e-01 1.04956090e-01 1.62036434e-01 -7.86705196e-01
-3.82464081e-01 -9.16701853e-01 -1.61899507e+00 -7.90582001e-01
-4.95854430e-02 3.97652298e-01 1.56233355e-01 9.78882492e-01
4.48131830e-01 4.95996356e-01 2.95492619e-01 -1.46328902e+00
-5.69628119e-01 -7.13306785e-01 -5.68211019e-01 6.92708254e-01
7.32893527e-01 -5.78715384e-01 -5.30154288e-01 -1.95421010e-01] | [5.024406433105469, 5.31587553024292] |
7cf63430-bf8f-44b6-8c60-2f0227101f1e | diversity-encouraged-learning-of-unsupervised | 1611.04899 | null | http://arxiv.org/abs/1611.04899v2 | http://arxiv.org/pdf/1611.04899v2.pdf | Diversity encouraged learning of unsupervised LSTM ensemble for neural activity video prediction | Being able to predict the neural signal in the near future from the current
and previous observations has the potential to enable real-time responsive
brain stimulation to suppress seizures. We have investigated how to use an
auto-encoder model consisting of LSTM cells for such prediction. Recog- nizing
that there exist multiple activity pattern clusters, we have further explored
to train an ensemble of LSTM mod- els so that each model can specialize in
modeling certain neural activities, without explicitly clustering the training
data. We train the ensemble using an ensemble-awareness loss, which jointly
solves the model assignment problem and the error minimization problem. During
training, for each training sequence, only the model that has the lowest recon-
struction and prediction error is updated. Intrinsically such a loss function
enables each LTSM model to be adapted to a subset of the training sequences
that share similar dynamic behavior. We demonstrate this can be trained in an
end- to-end manner and achieve significant accuracy in neural activity
prediction. | ['Yao Wang', 'Yilin Song', 'Jonathan Viventi'] | 2016-11-15 | null | null | null | null | ['activity-prediction', 'activity-prediction'] | ['computer-vision', 'time-series'] | [ 7.68003523e-01 2.24730954e-01 6.32437989e-02 -5.98605871e-01
-6.58689260e-01 -3.23775530e-01 4.20587003e-01 -2.39936903e-01
-4.95595485e-01 8.86124671e-01 2.52577126e-01 8.32453892e-02
-2.45429575e-01 -2.85558552e-01 -8.43303978e-01 -9.16258395e-01
-4.42238241e-01 6.25928342e-01 6.51189610e-02 2.18283474e-01
2.84942031e-01 4.50350761e-01 -1.56614685e+00 7.52696633e-01
8.20371389e-01 1.05871487e+00 7.35834599e-01 6.12495244e-01
1.56058982e-01 1.17530000e+00 -5.66030800e-01 4.19371068e-01
-2.32232243e-01 -7.01051176e-01 -5.69011271e-01 3.91985029e-02
2.09415630e-01 1.63272440e-01 -4.76918042e-01 5.18077612e-01
5.84651470e-01 2.48645380e-01 6.95852458e-01 -8.95306408e-01
1.17483482e-01 6.87954724e-01 -1.30299032e-01 5.03970623e-01
2.53031366e-02 4.15212363e-02 4.77297485e-01 -6.20121598e-01
5.97441852e-01 4.82959747e-01 5.55641890e-01 1.09521890e+00
-1.60316956e+00 -8.33815634e-01 2.85300106e-01 1.72292858e-01
-1.29232824e+00 -6.67408526e-01 5.59414744e-01 -2.78507203e-01
1.43046296e+00 1.36770234e-01 8.97615910e-01 1.51324987e+00
6.77026570e-01 9.40087140e-01 9.01551425e-01 -1.26501650e-01
4.73767519e-01 -4.23788540e-02 -2.37698331e-01 3.98569524e-01
-3.21650773e-01 -3.91526036e-02 -9.64722872e-01 -9.61585529e-03
3.62257540e-01 -3.98153253e-02 -4.80223835e-01 -2.98798680e-01
-1.28103149e+00 3.01760584e-01 1.88834533e-01 7.24573910e-01
-5.77076674e-01 5.63349903e-01 3.44032168e-01 1.10059395e-01
3.95356417e-01 8.69490504e-01 -7.55616307e-01 -3.16795051e-01
-1.31809139e+00 -1.84695184e-01 6.52433038e-01 4.38045502e-01
6.81128383e-01 3.20172876e-01 -2.17207983e-01 7.25357234e-01
-1.37230393e-03 8.74278620e-02 1.00618482e+00 -1.06962621e+00
1.81559220e-01 3.81019294e-01 -2.01160625e-01 -4.18618023e-01
-4.46317226e-01 -6.84942484e-01 -7.70772040e-01 4.18337435e-02
-1.31825656e-01 -3.78604650e-01 -9.98496413e-01 2.16354728e+00
-2.35055432e-01 9.15005624e-01 9.93712023e-02 1.11138612e-01
-4.90504503e-02 6.14260972e-01 2.62889504e-01 -4.86840159e-01
5.66504657e-01 -5.92359066e-01 -6.30195320e-01 -5.60964644e-01
1.08322585e+00 1.74152087e-02 5.07055998e-01 6.17517412e-01
-1.13461018e+00 -2.69192785e-01 -1.17551267e+00 5.86548746e-01
-1.07956007e-02 -3.31799164e-02 5.48374712e-01 1.95999101e-01
-1.40426862e+00 8.98735046e-01 -1.49417222e+00 -4.52209473e-01
6.80893302e-01 8.67270887e-01 -2.72182822e-01 4.21303779e-01
-8.83946836e-01 1.17342544e+00 7.89886117e-01 -1.08108101e-02
-1.43229079e+00 -7.36478746e-01 -5.06298900e-01 2.43807614e-01
-2.04293206e-01 -6.89428031e-01 1.05862200e+00 -1.49969590e+00
-1.50454569e+00 6.07529044e-01 -5.66889524e-01 -8.02431285e-01
4.65078913e-02 2.98401922e-01 -4.39709485e-01 -1.46239713e-01
-1.03301398e-01 7.73560762e-01 7.55592406e-01 -1.11404836e+00
-8.12925816e-01 -4.15851742e-01 -8.32137883e-01 4.92660433e-01
-7.33950138e-01 -1.10521704e-01 -1.07783034e-01 -6.61802113e-01
1.37897879e-01 -9.67710257e-01 -2.66156495e-01 -4.17951733e-01
-1.33459359e-01 -2.34166533e-02 8.13599527e-01 -4.14405316e-01
1.24659467e+00 -1.95100641e+00 3.63799214e-01 3.36494982e-01
9.87451077e-02 -2.69907974e-02 -2.39997789e-01 3.01350564e-01
-3.21308583e-01 -2.50981867e-01 -4.04632837e-01 -5.49410045e-01
-4.50645745e-01 3.99231046e-01 -5.21546245e-01 2.84705728e-01
1.00431912e-01 8.34486723e-01 -6.58283710e-01 -1.28819257e-01
3.88971418e-02 5.60675502e-01 -4.59440738e-01 2.57293582e-01
-3.28681856e-01 8.89063120e-01 -3.71154934e-01 1.25009492e-01
4.84735258e-02 -3.22231114e-01 4.28349078e-01 1.11875877e-01
2.28417497e-02 3.97579879e-01 -7.35992730e-01 1.97227895e+00
-8.16935360e-01 9.52077389e-01 -3.27363282e-01 -1.37876368e+00
8.85723352e-01 5.92045307e-01 1.01337576e+00 -6.97785497e-01
2.55735535e-02 2.66475648e-01 1.78075522e-01 -2.62118697e-01
-1.66239038e-01 1.02237590e-01 1.22486651e-01 7.01742947e-01
5.05165279e-01 4.56591308e-01 -1.48964487e-02 -8.63343943e-03
1.40404749e+00 2.46683002e-01 -1.14657134e-01 -2.86176115e-01
2.96753228e-01 -2.99595445e-01 6.20871365e-01 8.64362001e-01
9.01895016e-02 3.21658552e-01 8.83412808e-02 -3.37069333e-01
-8.97901356e-01 -9.84112382e-01 -1.24282598e-01 1.22269034e+00
-2.71113694e-01 -4.07404527e-02 -7.17414141e-01 -4.01285559e-01
-6.09842777e-01 1.11570811e+00 -8.50706637e-01 -7.24976480e-01
-8.00947011e-01 -8.65986943e-01 5.11162937e-01 5.51061451e-01
1.99772477e-01 -1.47793591e+00 -9.72284853e-01 8.04226518e-01
-2.34666675e-01 -6.82346940e-01 -3.53277594e-01 1.08513629e+00
-1.18804348e+00 -5.20969093e-01 -3.62403870e-01 -1.07187700e+00
7.12295532e-01 -4.18369383e-01 9.45631266e-01 -8.98446292e-02
-2.59191394e-02 1.45597517e-01 -3.46306041e-02 -3.35854888e-01
-4.04461890e-01 3.14417481e-01 2.61831313e-01 2.91333020e-01
5.30575037e-01 -1.18144655e+00 -6.15861654e-01 9.25058946e-02
-5.87500274e-01 2.01748088e-01 4.91727263e-01 8.62446010e-01
7.74213850e-01 -5.73517606e-02 8.74587715e-01 -7.45748281e-01
2.96207309e-01 -4.52733427e-01 -8.62894952e-02 4.80396569e-01
-6.25367880e-01 4.48756635e-01 6.99582219e-01 -7.73296177e-01
-1.05607951e+00 5.35834432e-01 -7.65841380e-02 -4.61112201e-01
-6.11620918e-02 5.51922381e-01 -3.74379903e-02 -6.43090978e-02
6.73328698e-01 1.02707899e+00 -1.60530508e-01 -1.82630926e-01
-1.45223485e-02 3.47412556e-01 5.84065795e-01 -2.69606352e-01
-1.88298672e-02 1.37856871e-01 -2.51638740e-01 -4.93719190e-01
-6.91414356e-01 -1.01356983e-01 -4.79230285e-01 -3.79063576e-01
7.49760687e-01 -8.77035737e-01 -6.76413953e-01 5.05258858e-01
-1.12984788e+00 -6.87307537e-01 -2.61378139e-01 5.08600712e-01
-1.13128245e+00 -1.97768956e-01 -3.58738035e-01 -8.77508998e-01
-3.76411855e-01 -9.68954563e-01 6.86868727e-01 9.93271098e-02
-5.46151042e-01 -9.85577106e-01 4.37586963e-01 -2.53420740e-01
3.99214149e-01 2.40187764e-01 8.21383953e-01 -9.21646297e-01
-3.99123907e-01 -2.13591844e-01 6.15137637e-01 1.39461696e-01
4.41589542e-02 -3.71728987e-01 -1.21977568e+00 -4.78470087e-01
3.63716483e-01 -2.84405112e-01 1.12691617e+00 6.46288395e-01
1.61924779e+00 -3.47093463e-01 -9.24931347e-01 7.37901092e-01
1.19170642e+00 6.80641890e-01 8.69778275e-01 1.64914742e-01
4.21028674e-01 3.29985261e-01 -3.93150412e-02 2.32141688e-01
1.09975278e-01 6.13765121e-01 2.32245699e-01 1.97318211e-01
8.42334554e-02 -3.93966675e-01 6.27061427e-01 8.29676390e-01
2.08330095e-01 -2.36035064e-01 -9.70347524e-01 6.70102835e-01
-1.97383344e+00 -1.31907129e+00 6.15382016e-01 2.37415743e+00
9.65369284e-01 1.57185152e-01 -2.13441506e-01 -1.06874026e-01
4.01373982e-01 -7.50039741e-02 -1.07428026e+00 -2.89003193e-01
-2.54034162e-01 4.95443046e-01 4.08646464e-01 3.29979628e-01
-8.75433266e-01 9.46187198e-01 6.74358225e+00 9.24856246e-01
-1.37725723e+00 1.47809520e-01 8.66099536e-01 -5.14925897e-01
-1.67799816e-01 -1.40924335e-01 -7.35871196e-01 7.98132479e-01
1.67387331e+00 -2.32509732e-01 8.62190843e-01 3.63429934e-01
4.09378439e-01 -1.53575480e-01 -1.42034698e+00 7.71822572e-01
1.43291265e-01 -1.47233367e+00 -8.51186216e-02 -1.46241821e-02
9.82183397e-01 5.00664890e-01 1.65613398e-01 3.33341211e-01
3.66613626e-01 -1.34317279e+00 4.64855582e-01 9.06613946e-01
7.13361204e-01 -7.61271894e-01 2.01380983e-01 8.49100053e-01
-1.00438905e+00 -4.26437289e-01 -7.36689544e-04 2.25229457e-01
9.40701962e-02 2.94474691e-01 -9.95921612e-01 -6.97172433e-02
5.88331699e-01 9.05338466e-01 -3.76470178e-01 9.90548074e-01
5.51116839e-02 6.88658774e-01 -4.04903054e-01 -1.72966957e-01
1.47551224e-02 7.58365318e-02 3.69817883e-01 1.00654364e+00
7.51313925e-01 4.36367467e-02 -4.86895181e-02 6.77214801e-01
5.53472415e-02 -3.24986756e-01 -6.00946963e-01 -9.40495580e-02
5.87577105e-01 8.65316331e-01 -5.44457138e-01 -3.21021169e-01
6.89913034e-02 1.15885413e+00 6.92433834e-01 5.09705782e-01
-6.21255875e-01 -9.55680832e-02 4.06992614e-01 -2.22207215e-02
3.31881016e-01 7.60262460e-02 -5.97624004e-01 -1.14215052e+00
-2.01469854e-01 -6.88244879e-01 2.61517406e-01 -8.36837530e-01
-8.05715621e-01 1.07077527e+00 -4.18051988e-01 -1.14668763e+00
-8.05744469e-01 -2.41804168e-01 -8.02734971e-01 7.22322524e-01
-8.05801094e-01 -8.13659370e-01 2.13531241e-01 4.96386021e-01
5.32492876e-01 -2.65712440e-01 1.10697281e+00 -1.99137925e-04
-6.03924572e-01 6.07368648e-01 4.33876574e-01 -3.62950414e-01
3.25756192e-01 -1.07225621e+00 8.39966238e-02 7.38591075e-01
4.58668530e-01 3.31249833e-01 8.84481907e-01 -4.50484008e-01
-8.15990627e-01 -1.32066381e+00 1.13194418e+00 -6.83921754e-01
4.85878766e-01 -5.37606597e-01 -1.13065028e+00 1.05481255e+00
1.46228582e-01 -2.88160294e-01 7.92788923e-01 -1.16574757e-01
1.13574356e-01 -3.08605403e-01 -8.25693369e-01 5.79089880e-01
1.07352078e+00 -4.81464237e-01 -3.82733643e-01 2.80810922e-01
2.41640866e-01 -1.45219266e-01 -7.93144405e-01 4.35996085e-01
3.78180206e-01 -6.92196786e-01 5.67365408e-01 -6.66700900e-01
9.49996412e-02 -1.57967106e-01 2.42634341e-02 -1.62650371e+00
-3.14046472e-01 -5.42703450e-01 -5.53284287e-01 7.50058889e-01
8.97967935e-01 -4.96605903e-01 1.11896765e+00 7.32160032e-01
-3.84048462e-01 -1.16736138e+00 -1.14398527e+00 -6.00803137e-01
4.57104072e-02 -4.88033533e-01 2.02778563e-01 5.76203287e-01
4.16191518e-01 2.72075087e-01 -5.27576268e-01 -1.47072403e-02
3.38375002e-01 -9.94035453e-02 1.39456898e-01 -9.93035138e-01
-4.51945662e-01 -4.19767141e-01 -3.35247755e-01 -1.19478345e+00
5.05498707e-01 -1.17151392e+00 3.77909243e-01 -1.39809155e+00
2.78984100e-01 -2.76422948e-01 -8.74342322e-01 7.18457162e-01
8.26250911e-02 2.21240893e-02 -1.24312676e-01 1.87065810e-01
-7.33755410e-01 6.82946563e-01 7.59096205e-01 -1.56899631e-01
-4.92017657e-01 4.47539054e-02 -5.11687040e-01 4.21397924e-01
9.55208004e-01 -8.48945081e-01 -6.03143930e-01 -4.63683546e-01
1.41459890e-02 2.03213647e-01 -9.92075503e-02 -1.56887388e+00
6.96874380e-01 5.62302954e-02 8.56647491e-01 -3.53845358e-01
5.92260420e-01 -6.92470551e-01 4.43258882e-01 6.21189833e-01
-8.59789670e-01 -2.60722816e-01 3.09121579e-01 7.18888044e-01
-9.66510549e-02 1.78295299e-02 8.45491648e-01 -1.29447374e-02
-6.20525360e-01 6.14251196e-01 -9.21617150e-01 -1.69393763e-01
1.15090632e+00 -6.09229326e-01 1.06333673e-01 -3.62291992e-01
-1.14375949e+00 3.61254543e-01 2.20436558e-01 2.65744120e-01
5.51428556e-01 -1.26212192e+00 -5.68110347e-01 3.60459149e-01
9.17938203e-02 -3.77781093e-01 3.97461712e-01 8.65441144e-01
-1.75030306e-02 3.98625940e-01 -3.82820785e-01 -5.79201818e-01
-9.25114870e-01 3.24577242e-01 9.72012281e-01 -4.59002674e-01
-3.09220701e-01 9.63865817e-01 2.77873993e-01 -1.95284173e-01
3.38317573e-01 1.00467473e-01 -3.72901767e-01 -2.52003372e-01
5.69560528e-01 -7.79096633e-02 2.40726665e-01 -4.64093804e-01
-4.03045803e-01 1.00287102e-01 -3.08766872e-01 -3.68059695e-01
1.65803695e+00 3.66936736e-02 -1.12605598e-02 6.24469101e-01
1.28646469e+00 -8.09609234e-01 -1.73066020e+00 4.19774540e-02
2.22300380e-01 8.33899900e-02 2.59680092e-01 -1.11290073e+00
-1.11127460e+00 8.01578522e-01 8.34767401e-01 -2.38326460e-01
1.40896332e+00 -7.12210238e-02 5.73521733e-01 4.49406624e-01
5.17836630e-01 -1.21049821e+00 2.19119459e-01 5.66265464e-01
6.31442666e-01 -6.76375270e-01 -5.27746022e-01 4.57349479e-01
-7.78772295e-01 1.03582823e+00 7.99885154e-01 -1.97061330e-01
6.57928109e-01 4.89910096e-01 -2.78471678e-01 -1.31594062e-01
-1.44343805e+00 2.79375762e-01 1.92637667e-01 7.96405494e-01
4.18617249e-01 -2.30723210e-02 6.81126639e-02 4.78049397e-01
2.25868598e-02 3.44597787e-01 4.10161465e-01 7.31331050e-01
-6.74912572e-01 -9.15412605e-01 2.06281319e-01 1.11034739e+00
-2.31748998e-01 -8.74054804e-02 -1.58543870e-01 8.47306848e-02
-3.08205970e-02 6.16509259e-01 5.29722512e-01 -6.13976657e-01
-1.02968737e-02 3.75659406e-01 6.17655158e-01 -7.09147334e-01
-5.25394380e-01 6.80714920e-02 -1.36173740e-01 -5.80990732e-01
-2.15003431e-01 -8.12191308e-01 -1.34548032e+00 1.12622492e-01
-1.79393649e-01 2.50938386e-02 4.63302225e-01 1.02657866e+00
6.59771740e-01 6.83048248e-01 8.39062572e-01 -8.50267589e-01
-2.68525302e-01 -9.28187132e-01 -6.56835198e-01 7.46571645e-02
2.83802122e-01 -3.86908412e-01 -1.73893526e-01 1.98061377e-01] | [13.191075325012207, 3.539456367492676] |
623b3893-ede0-4406-b961-1db206dcfe38 | multiple-people-tracking-using-body-and-joint | null | null | http://openaccess.thecvf.com/content_CVPRW_2019/html/BMTT/Henschel_Multiple_People_Tracking_Using_Body_and_Joint_Detections_CVPRW_2019_paper.html | http://openaccess.thecvf.com/content_CVPRW_2019/papers/BMTT/Henschel_Multiple_People_Tracking_Using_Body_and_Joint_Detections_CVPRW_2019_paper.pdf | Multiple People Tracking using Body and Joint Detections | Most multiple people tracking systems compute trajectories based on the tracking-by-detection paradigm. Consequently, the performance depends to a large extent on the quality of the employed input detections. However, despite an enormous progress in recent years, partially occluded people are still often not recognized. Also, many correct detections are mistakenly discarded when the non-maximum suppression is performed. Improving the tracking performance thus requires to augment the coarse input. Wellsuited for this task are fine-graded body joint detections, as they allow to locate even strongly occluded persons.
Thus in this work, we analyze the suitability of including joint detections for multiple people tracking. We introduce different affinities between the two detection types and evaluate their performances. Tracking is then performed within a near-online framework based on a min cost graph labeling formulation. As a result, our framework can recover heavily occluded persons and solve the data association efficiently. We evaluate our framework on the MOT16/17 benchmark. Experimental results demonstrate that our framework achieves state-of-the-art results. | ['Bodo Rosenhahn', 'Yunzhe Zou', 'Roberto Henschel'] | 2019-06-01 | null | null | null | the-ieee-conference-on-computer-vision-and-3 | ['multiple-people-tracking'] | ['computer-vision'] | [-4.43272889e-02 -2.69593775e-01 -6.89116567e-02 -7.54368380e-02
-6.55218065e-01 -5.54847062e-01 4.79861259e-01 2.45377526e-01
-7.04890311e-01 7.34388769e-01 -1.54397026e-01 1.57368943e-01
-6.99589252e-02 -6.42069817e-01 -7.15598702e-01 -5.79916239e-01
-4.64909188e-02 8.77807736e-01 6.66186869e-01 8.22528526e-02
-2.07856476e-01 5.84237516e-01 -1.79736769e+00 8.06826726e-03
9.27072763e-01 6.92797959e-01 -2.51596142e-02 5.64827502e-01
1.01786062e-01 2.28097543e-01 -5.94451368e-01 -5.75693905e-01
4.47931200e-01 -2.22411811e-01 -3.45786780e-01 4.24789131e-01
8.69765937e-01 -2.72020817e-01 -1.74720690e-01 1.24226558e+00
3.65607440e-01 1.73548356e-01 4.56331491e-01 -1.29926217e+00
-1.37000993e-01 3.73277664e-01 -7.20028937e-01 2.03313306e-01
5.46188891e-01 1.29757792e-01 7.63185859e-01 -9.70571935e-01
4.97718930e-01 1.25653970e+00 8.88124406e-01 5.74991703e-01
-1.32762563e+00 -5.73031425e-01 5.57051420e-01 8.19861889e-02
-1.69526064e+00 -5.31405449e-01 3.95879686e-01 -5.13202071e-01
3.67215782e-01 3.73682082e-01 8.22320461e-01 8.63370955e-01
-2.40214542e-01 7.39898384e-01 9.26061571e-01 -2.52179742e-01
-1.79954339e-02 1.17897160e-01 2.39811078e-01 9.91856396e-01
9.19662535e-01 1.10907689e-01 -3.24864179e-01 -2.25693583e-01
5.00822842e-01 1.41335487e-01 -2.73649931e-01 -5.54904819e-01
-1.10428643e+00 5.01653552e-01 5.57706177e-01 2.68412113e-01
-2.58263618e-01 1.26264825e-01 1.65060297e-01 -4.31761257e-02
4.32475358e-01 -1.67719908e-02 4.92470227e-02 1.54532060e-01
-1.28055823e+00 5.65550625e-01 5.57491124e-01 9.25062895e-01
5.94658852e-01 -2.18428791e-01 -5.61384141e-01 4.62962240e-01
2.64884681e-01 7.59862065e-01 -1.89293310e-01 -6.26302838e-01
5.07913768e-01 8.51557314e-01 5.78228652e-01 -1.11467767e+00
-6.95840299e-01 -8.08344126e-01 -8.95159721e-01 3.65403831e-01
1.06458628e+00 -1.61431819e-01 -8.66110861e-01 1.74976385e+00
7.63794065e-01 1.92522660e-01 -3.52878481e-01 1.20711613e+00
3.25340271e-01 1.68600067e-01 2.70804793e-01 -2.88886905e-01
1.57730091e+00 -9.76588726e-01 -9.02014554e-01 -3.32968086e-01
5.01145899e-01 -6.38158798e-01 4.92310822e-01 2.81062216e-01
-1.07800376e+00 -8.12308490e-01 -7.41784751e-01 3.02898586e-01
-2.52734095e-01 5.67520976e-01 2.97377616e-01 9.56598103e-01
-9.64949191e-01 6.18740857e-01 -9.51989949e-01 -5.35858214e-01
4.74641502e-01 4.94458109e-01 -2.02973515e-01 -9.99598056e-02
-8.90230477e-01 7.35140979e-01 3.73802453e-01 4.59619015e-01
-4.66156304e-01 -5.17851293e-01 -5.79362154e-01 6.00575935e-03
7.06045151e-01 -1.03537095e+00 9.09896493e-01 -5.69182932e-01
-8.13120604e-01 7.86563754e-01 -2.66556710e-01 -5.74116707e-01
1.27215374e+00 -4.24352884e-01 -4.25571620e-01 1.02136031e-01
2.37736911e-01 5.25280714e-01 9.06509757e-01 -1.20375943e+00
-1.02429450e+00 -4.96169269e-01 1.23306327e-01 9.87291262e-02
-2.84861773e-01 1.25439376e-01 -9.21946943e-01 -5.60961008e-01
6.06442466e-02 -1.33952785e+00 -3.98882061e-01 4.54505712e-01
-5.28484464e-01 -3.00038218e-01 6.29431784e-01 -4.39310133e-01
1.28884780e+00 -1.86270118e+00 2.11617947e-01 2.29584336e-01
4.31345999e-01 4.01054502e-01 2.04395890e-01 1.91732362e-01
4.06758517e-01 -2.32636452e-01 -9.43461955e-02 -8.32049549e-01
1.27201185e-01 -9.44720954e-02 8.77807196e-03 8.63592029e-01
5.73280901e-02 7.18507767e-01 -8.01933229e-01 -8.16836298e-01
3.61049891e-01 4.88600880e-01 -5.47570646e-01 7.28162676e-02
-9.69003886e-02 6.13006055e-01 -5.89468777e-01 8.15504134e-01
7.35347211e-01 -3.78956109e-01 1.34966657e-01 -3.30058098e-01
-2.53675789e-01 -4.61177379e-01 -1.55072904e+00 1.61778247e+00
2.49285862e-01 2.04861253e-01 1.13787644e-01 -6.20841086e-01
5.43608606e-01 2.13951036e-01 7.08081365e-01 -2.42231548e-01
7.56326169e-02 2.64261782e-01 9.98057332e-03 -2.08246320e-01
5.79649270e-01 1.61915362e-01 1.57562848e-02 -7.23541677e-02
-4.67899591e-01 8.14571381e-01 6.21772528e-01 1.19240910e-01
1.00648522e+00 4.09870654e-01 2.51837105e-01 -2.62984723e-01
7.75065124e-01 2.65722334e-01 7.08098114e-01 1.00338197e+00
-5.94623029e-01 5.04601538e-01 2.40682010e-02 -3.76579970e-01
-5.40942848e-01 -1.10487819e+00 -9.06428471e-02 1.06520438e+00
3.85594308e-01 -4.00270641e-01 -8.74404848e-01 -7.02340484e-01
1.40466154e-01 6.31509945e-02 -5.26563942e-01 7.90211931e-02
-7.91337371e-01 -7.85538971e-01 6.40247464e-01 5.40588081e-01
2.47268647e-01 -8.23815405e-01 -7.75732756e-01 2.91709691e-01
-2.47476041e-01 -1.39387894e+00 -5.24441838e-01 -3.55078280e-01
-7.23753333e-01 -1.31385398e+00 -1.10108125e+00 -5.33392072e-01
8.46759498e-01 5.73846698e-01 9.92219508e-01 5.28033614e-01
-3.41569602e-01 4.68107671e-01 -1.61701113e-01 -8.07115883e-02
-1.71239793e-01 9.50739384e-02 3.95531476e-01 2.69075096e-01
1.61801994e-01 -1.85598865e-01 -7.14540780e-01 4.66143429e-01
-5.27757525e-01 -9.04696286e-02 6.20047748e-01 3.46201211e-01
6.78753972e-01 -9.84979514e-03 4.12275314e-01 -7.51159966e-01
8.82285461e-02 -9.78088379e-02 -8.45636904e-01 4.35130239e-01
-2.64739484e-01 6.85058348e-03 4.00026649e-01 -5.58778346e-01
-8.33995461e-01 5.08508384e-01 -7.53382742e-02 -6.59315467e-01
-2.49266148e-01 -3.89159806e-02 -1.27715588e-01 -2.23269150e-01
4.59889591e-01 -5.34906723e-02 -2.86182135e-01 -5.66366196e-01
2.86044598e-01 1.22891754e-01 6.66013241e-01 -5.48410773e-01
1.12791574e+00 7.81445146e-01 1.79301143e-01 -7.42947340e-01
-9.08353209e-01 -8.02119553e-01 -7.09005654e-01 -6.53896689e-01
9.49364483e-01 -9.23684537e-01 -1.03330386e+00 2.87450224e-01
-1.11746180e+00 1.11817688e-01 -9.28303227e-02 4.26805735e-01
-1.37906417e-01 5.71622789e-01 -3.26438487e-01 -1.23146653e+00
-2.36435309e-01 -1.16581523e+00 1.36029375e+00 2.70888448e-01
-2.75608897e-02 -7.75828063e-01 2.26921719e-02 2.62934119e-01
7.99819306e-02 4.22483742e-01 1.30019590e-01 -4.06713665e-01
-9.17618513e-01 -4.20593679e-01 -2.69526243e-01 -3.55552226e-01
5.65758208e-03 -2.11104974e-01 -8.78395617e-01 -6.59637094e-01
-5.68994224e-01 7.44915903e-02 1.09498739e+00 4.21940565e-01
7.90607274e-01 4.59411629e-02 -8.62817168e-01 3.38859528e-01
1.21826839e+00 -3.80117357e-01 2.36728415e-01 1.75487921e-01
9.18316245e-01 6.66306078e-01 7.92033970e-01 3.82734060e-01
3.59535545e-01 1.13328600e+00 4.01694804e-01 -5.26279397e-02
-4.00129199e-01 -9.89747792e-02 3.81875068e-01 2.95296699e-01
-4.43752706e-01 -2.56656677e-01 -7.18544662e-01 5.28086662e-01
-2.21212149e+00 -1.03650320e+00 -5.65098166e-01 2.39087200e+00
4.30285752e-01 3.34655702e-01 7.57311523e-01 1.84975937e-01
1.06778526e+00 5.78261204e-02 -3.66976619e-01 6.35449350e-01
-2.52086259e-02 -1.90778330e-01 5.93196213e-01 3.77322346e-01
-1.40247405e+00 7.15207517e-01 5.32088804e+00 7.68534124e-01
-5.15836596e-01 1.99448481e-01 1.09771062e-02 -1.19239122e-01
3.95717382e-01 -2.41220519e-01 -1.35556221e+00 5.04809737e-01
4.02851939e-01 3.20845470e-02 1.62263840e-01 6.57573521e-01
7.27439672e-02 -1.02018937e-01 -9.98701632e-01 1.10901880e+00
-5.21952510e-02 -9.29075897e-01 -2.68310100e-01 1.34134129e-01
5.38621128e-01 -3.24056566e-01 1.32355075e-02 3.20835948e-01
-1.31450012e-01 -5.47508895e-01 9.33389544e-01 5.98340571e-01
4.47551399e-01 -5.96766233e-01 4.18543875e-01 5.26413023e-01
-1.86842263e+00 -2.96272747e-02 -4.41787422e-01 1.50680915e-01
4.53733504e-01 6.87484145e-01 -4.49810833e-01 7.75643170e-01
7.23097503e-01 4.76272255e-01 -7.83687890e-01 1.51346028e+00
-1.05172796e-02 1.96537226e-01 -5.75212836e-01 1.50218651e-01
-5.58129549e-02 -1.14641361e-01 9.02282953e-01 1.35949910e+00
3.52065086e-01 -2.16459949e-03 8.75916660e-01 7.61067033e-01
1.21589676e-01 4.43397388e-02 -3.49824965e-01 3.08320880e-01
2.75240719e-01 1.25983393e+00 -1.11405158e+00 -4.13550973e-01
-4.37121063e-01 8.53806674e-01 4.68995094e-01 2.18468502e-01
-1.18521821e+00 3.18586200e-01 7.07819641e-01 3.08618367e-01
3.71196926e-01 -1.22011639e-01 1.30484387e-01 -1.25139809e+00
3.12727004e-01 -5.85553348e-01 6.58986866e-01 -1.32492974e-01
-1.23349237e+00 5.50686777e-01 -1.80364226e-03 -1.49118185e+00
-6.94182888e-02 -4.54267234e-01 -7.99270272e-02 5.02622843e-01
-1.44375980e+00 -1.13957477e+00 -5.15907824e-01 6.67389810e-01
4.03215557e-01 3.47868115e-01 5.28125644e-01 9.31507945e-01
-8.30122948e-01 6.06703520e-01 -2.62683421e-01 1.54582262e-01
6.64427757e-01 -1.20425832e+00 2.06273735e-01 1.15759480e+00
2.29227468e-01 4.54072863e-01 9.90359366e-01 -1.02727246e+00
-1.33387208e+00 -1.37387979e+00 7.06623733e-01 -5.19242764e-01
3.65116864e-01 -4.16662037e-01 -9.16692853e-01 6.44515157e-01
-9.60214585e-02 3.68326277e-01 2.57883489e-01 1.00253917e-01
-7.84637704e-02 -4.73217107e-02 -9.58190620e-01 5.65933585e-01
1.45993972e+00 9.58242279e-04 -3.88501197e-01 5.73955834e-01
2.65245736e-01 -5.31603158e-01 -6.11309648e-01 4.09079224e-01
5.94081581e-01 -9.65613604e-01 1.24883616e+00 -4.17030782e-01
-4.67078358e-01 -7.99154401e-01 1.33332774e-01 -7.29068339e-01
-4.17115659e-01 -5.18474698e-01 -4.72616374e-01 1.20030844e+00
8.95701572e-02 -4.25847083e-01 1.03634822e+00 5.01296163e-01
1.03003636e-01 -3.35062504e-01 -9.45324123e-01 -1.15919423e+00
-4.37735230e-01 -1.47388235e-01 2.56280810e-01 5.60622036e-01
-4.87552285e-01 9.82579514e-02 -7.14123547e-01 7.55647838e-01
1.23975587e+00 4.62294728e-01 9.69445288e-01 -1.58471704e+00
-3.57998639e-01 -4.08339322e-01 -4.41246629e-01 -1.21183586e+00
-7.05208182e-02 -6.66764617e-01 1.90661773e-01 -1.40963125e+00
3.23559225e-01 -5.19114494e-01 -2.18796730e-01 3.88096184e-01
-6.20966971e-01 5.51518738e-01 4.82393503e-01 3.70385498e-01
-1.18415320e+00 3.22397500e-01 9.45619226e-01 -6.41354397e-02
-1.41395524e-01 4.04700637e-01 -2.24842057e-01 8.20386529e-01
6.61533952e-01 -6.92193568e-01 1.80133760e-01 -3.52241218e-01
9.28009599e-02 -1.01859652e-01 8.18746209e-01 -1.46728778e+00
5.37904978e-01 1.20463856e-01 5.24486065e-01 -1.02617145e+00
6.75854981e-01 -1.06890535e+00 3.73816997e-01 7.79619992e-01
5.26291654e-02 2.12597363e-02 1.29182816e-01 9.34977651e-01
1.71908721e-01 -2.57302523e-02 7.33385921e-01 5.63079342e-02
-6.13883674e-01 4.03510690e-01 -4.85053882e-02 -7.73411570e-03
1.01246977e+00 -2.52756655e-01 -1.80953071e-01 -1.67792160e-02
-9.15345788e-01 3.77934694e-01 5.03486454e-01 3.15165877e-01
3.05166841e-01 -1.31674707e+00 -6.94667041e-01 -1.15428261e-01
1.27260953e-01 -2.88118601e-01 2.33773381e-01 1.27863932e+00
-1.17136084e-01 2.75660664e-01 1.19038314e-01 -9.23127115e-01
-1.55797756e+00 7.99062967e-01 2.23910421e-01 -4.70850468e-01
-8.59207153e-01 3.76287848e-01 2.37191543e-01 -3.87788527e-02
4.54459459e-01 -3.02300841e-01 -2.39566728e-01 6.79735392e-02
6.37822688e-01 7.42163301e-01 -1.67610958e-01 -8.51675272e-01
-7.05515862e-01 7.89062500e-01 1.03443742e-01 1.70486689e-01
7.65180051e-01 -2.67575890e-01 3.66694510e-01 1.05908215e-01
5.42875290e-01 2.28288442e-01 -1.32579768e+00 -2.02923685e-01
1.13708027e-01 -4.96779203e-01 -3.62036914e-01 -3.91181052e-01
-1.14457631e+00 5.08817971e-01 8.07533443e-01 4.11011487e-01
9.36246634e-01 -1.10945422e-02 6.51567996e-01 3.82684171e-01
5.63166678e-01 -8.82868767e-01 -3.16667944e-01 2.93359198e-02
5.40336728e-01 -1.21202612e+00 2.11175039e-01 -8.41332197e-01
-2.10639983e-01 8.48218739e-01 5.11561871e-01 -1.58994377e-01
2.78735548e-01 1.39536440e-01 -1.33677363e-01 -2.12089717e-01
-1.01439230e-01 -9.23715711e-01 5.42126477e-01 3.82028401e-01
7.14900792e-02 -1.24904560e-02 -3.22672963e-01 3.83623272e-01
2.05841795e-01 -5.12218662e-02 4.21382524e-02 8.05870891e-01
-5.36735177e-01 -1.05009568e+00 -9.26531494e-01 2.15641484e-01
-4.89680797e-01 3.83015096e-01 -2.58976579e-01 1.02248073e+00
3.45531195e-01 1.13079453e+00 -2.89952219e-01 7.20300749e-02
7.15025663e-01 -1.20988809e-01 6.66428566e-01 -3.49598527e-01
-6.43244505e-01 3.18315327e-01 2.29926221e-02 -6.42689109e-01
-8.38877797e-01 -9.05916095e-01 -1.18745244e+00 -2.40750894e-01
-5.14919162e-01 5.08588813e-02 2.95888931e-01 9.41043854e-01
1.58876091e-01 6.43426478e-01 6.64228573e-02 -1.01454544e+00
-6.83727443e-01 -6.40640140e-01 -3.76332790e-01 5.32486856e-01
2.69792557e-01 -1.15728688e+00 -6.94992989e-02 -7.31094033e-02] | [6.521662712097168, -1.8310952186584473] |
3138074a-07b0-4442-9103-8299c6f4da81 | cliper-a-unified-vision-language-framework | 2303.00193 | null | https://arxiv.org/abs/2303.00193v1 | https://arxiv.org/pdf/2303.00193v1.pdf | CLIPER: A Unified Vision-Language Framework for In-the-Wild Facial Expression Recognition | Facial expression recognition (FER) is an essential task for understanding human behaviors. As one of the most informative behaviors of humans, facial expressions are often compound and variable, which is manifested by the fact that different people may express the same expression in very different ways. However, most FER methods still use one-hot or soft labels as the supervision, which lack sufficient semantic descriptions of facial expressions and are less interpretable. Recently, contrastive vision-language pre-training (VLP) models (e.g., CLIP) use text as supervision and have injected new vitality into various computer vision tasks, benefiting from the rich semantics in text. Therefore, in this work, we propose CLIPER, a unified framework for both static and dynamic facial Expression Recognition based on CLIP. Besides, we introduce multiple expression text descriptors (METD) to learn fine-grained expression representations that make CLIPER more interpretable. We conduct extensive experiments on several popular FER benchmarks and achieve state-of-the-art performance, which demonstrates the effectiveness of CLIPER. | ['Feng Zhao', 'Zhaoqing Zhu', 'Hongjing Niu', 'Hanting Li'] | 2023-03-01 | null | null | null | null | ['facial-expression-recognition'] | ['computer-vision'] | [ 9.00417417e-02 -4.35052216e-01 -4.00382757e-01 -1.04013240e+00
-1.56631321e-01 -2.48401538e-01 5.81628382e-01 -3.37263674e-01
-2.29259759e-01 5.40100038e-01 2.20662087e-01 3.81987810e-01
3.13709974e-01 -3.73885334e-01 -4.23377603e-01 -7.41637468e-01
3.27684492e-01 3.37511338e-02 -1.53337955e-01 -4.65697885e-01
-4.14705724e-01 3.50312769e-01 -1.49572527e+00 4.57612127e-01
5.61993539e-01 1.51235247e+00 -3.61591876e-01 -1.04574980e-02
-3.87067020e-01 1.21792018e+00 -5.68725646e-01 -8.76494527e-01
-3.53046328e-01 -4.64285523e-01 -4.52498287e-01 3.02513868e-01
2.82502592e-01 -3.48163843e-01 -3.96068603e-01 1.31274998e+00
7.30257109e-02 -1.20559372e-01 6.12039685e-01 -1.68521214e+00
-8.35048378e-01 2.29420200e-01 -9.05775070e-01 -1.54818401e-01
4.53219384e-01 3.40146162e-02 8.82313073e-01 -1.00986671e+00
5.63410223e-01 1.54538167e+00 4.03479010e-01 6.68437600e-01
-9.55528677e-01 -1.07687020e+00 4.69651252e-01 2.39636645e-01
-1.43104172e+00 -7.72598326e-01 9.89860773e-01 -3.09869409e-01
3.96036983e-01 2.40028813e-01 6.38838232e-01 1.43202615e+00
-1.28343806e-01 1.19895601e+00 1.26950026e+00 -1.14190087e-01
-1.01288289e-01 1.27747729e-01 9.96573195e-02 9.07746017e-01
-3.30916733e-01 -2.37490654e-01 -6.84724033e-01 3.41427070e-03
4.17638302e-01 5.35124838e-01 -3.54174703e-01 1.01871192e-01
-6.59880579e-01 6.45916879e-01 5.01747549e-01 2.28605509e-01
-2.90691078e-01 2.78440863e-01 8.22101355e-01 1.95743486e-01
6.53801203e-01 -2.34077513e-01 -1.54757962e-01 -5.10680735e-01
-5.72809398e-01 1.02608809e-02 3.88334006e-01 7.41001070e-01
9.12367702e-01 1.04552798e-01 -3.61736178e-01 1.15227008e+00
4.40218836e-01 6.35647058e-01 4.69387531e-01 -7.97887206e-01
1.06234334e-01 8.15943182e-01 -1.33642182e-01 -1.39192259e+00
-5.39449565e-02 -1.32437330e-02 -1.05716252e+00 2.91709658e-02
-9.00991186e-02 -1.39647603e-01 -8.48112524e-01 2.04924035e+00
9.21519697e-02 2.36217722e-01 -1.18911445e-01 9.70234990e-01
8.16584170e-01 6.30968690e-01 4.80316788e-01 -2.47223631e-01
1.35497570e+00 -1.27652097e+00 -1.15634823e+00 -1.04032218e-01
5.47762156e-01 -4.22760248e-01 1.13669097e+00 4.20529962e-01
-6.37955368e-01 -5.07576466e-01 -7.12762892e-01 -1.22941777e-01
-3.29168051e-01 2.93503910e-01 1.01013279e+00 4.19032156e-01
-5.69077611e-01 1.52801320e-01 -7.85643518e-01 -5.87519035e-02
8.69215012e-01 1.14626899e-01 -6.16906464e-01 -1.00778483e-01
-1.27422881e+00 5.24438143e-01 1.59754995e-02 5.08467853e-01
-6.78873599e-01 -2.84346491e-01 -1.11958694e+00 -2.37489603e-02
3.71111959e-01 -1.10226497e-01 1.25066149e+00 -2.05920601e+00
-1.70075417e+00 1.13644278e+00 -5.28054535e-01 1.42422300e-02
4.18777883e-01 -2.72665143e-01 -7.71318734e-01 2.93988138e-01
-8.99078548e-02 5.74786007e-01 1.04290199e+00 -1.06165683e+00
-2.29585677e-01 -7.30579913e-01 6.78614080e-02 -1.26812249e-01
-6.02838576e-01 5.21183908e-01 -6.79786921e-01 -5.96533716e-01
-5.31820834e-01 -7.91111171e-01 1.24998331e-01 5.16475439e-01
-2.68946022e-01 -5.65728605e-01 1.12999594e+00 -5.02948403e-01
1.36883414e+00 -2.57336378e+00 8.74821693e-02 1.07537940e-01
3.96480262e-01 3.19725275e-01 -9.62889940e-02 7.22688511e-02
-1.78466126e-01 1.70579121e-01 -1.64138511e-01 -7.00306118e-01
2.26298138e-01 4.51453716e-01 -4.06647027e-01 3.43696594e-01
4.38168466e-01 9.51713443e-01 -8.32810342e-01 -7.18415320e-01
6.95069833e-03 5.59197903e-01 -1.80855662e-01 3.03611994e-01
-3.38229597e-01 4.94820982e-01 -9.09071147e-01 1.01429725e+00
6.54709578e-01 -3.67654413e-01 4.88449186e-02 -3.60088438e-01
1.76441491e-01 -4.52268392e-01 -7.00616837e-01 1.55413604e+00
-5.08455694e-01 6.75572038e-01 2.27349713e-01 -1.07610798e+00
1.05287635e+00 1.84026182e-01 4.53504473e-01 -8.16269219e-01
4.25289690e-01 6.93163648e-02 -4.96272743e-01 -7.39012122e-01
1.68802619e-01 -2.69768208e-01 -1.26693398e-01 1.97201639e-01
2.00991165e-02 3.76156986e-01 6.88309371e-02 4.24935529e-03
7.37280369e-01 2.10118383e-01 2.81445205e-01 1.37899786e-01
7.86297083e-01 -4.86210108e-01 9.78785336e-01 -3.88008878e-02
-4.95119870e-01 3.35980237e-01 7.99680650e-01 -4.60252583e-01
-3.18455130e-01 -6.89193666e-01 -1.64458320e-01 1.44259369e+00
2.00458437e-01 -6.17520809e-01 -6.83728218e-01 -7.92858899e-01
-1.12379268e-01 1.38260096e-01 -9.32803154e-01 -3.06230307e-01
-2.16528267e-01 -4.31422889e-01 7.63357341e-01 6.19864523e-01
8.32060635e-01 -1.11978507e+00 -2.21904978e-01 -9.30535719e-02
-2.46135905e-01 -1.37922406e+00 -4.65339035e-01 -2.66220540e-01
-2.59654045e-01 -8.42195868e-01 -7.66308069e-01 -5.85195482e-01
7.69015312e-01 1.42356560e-01 1.15584266e+00 4.13735628e-01
-2.49604166e-01 2.83380210e-01 -6.17817283e-01 -5.65133214e-01
-7.31514469e-02 -4.08258080e-01 -3.80594796e-03 7.70220041e-01
7.81331360e-01 -3.83029848e-01 -3.89781654e-01 2.70570487e-01
-1.09529746e+00 -1.58338230e-02 5.02040207e-01 9.13876116e-01
7.19597280e-01 -3.08030605e-01 4.12812203e-01 -8.85635257e-01
5.27076185e-01 -4.78644729e-01 -3.00244361e-01 4.79938567e-01
-1.46863297e-01 -6.96266368e-02 7.80424178e-01 -4.58627313e-01
-1.30551887e+00 -4.14812416e-02 -2.20302358e-01 -9.66545939e-01
-1.74883321e-01 5.48102736e-01 -4.48395461e-01 -9.07802284e-02
2.16898859e-01 2.94315487e-01 2.57023275e-01 -3.27752888e-01
2.16804236e-01 9.49902654e-01 4.75529999e-01 -8.00552726e-01
3.22129160e-01 6.21262968e-01 7.88847655e-02 -7.82749057e-01
-1.19563258e+00 -2.20129177e-01 -3.00250530e-01 -2.99193621e-01
8.88320982e-01 -1.06075168e+00 -8.15435767e-01 8.80434036e-01
-1.11218536e+00 -1.19642660e-01 2.37263769e-01 -1.29573112e-02
-4.13357019e-01 3.77144963e-01 -6.28828049e-01 -8.09581041e-01
-2.20050067e-01 -1.21618450e+00 1.48810184e+00 5.40821254e-01
-1.06378153e-01 -8.46405268e-01 -1.56060159e-01 2.68592149e-01
3.54132086e-01 5.65700114e-01 6.85227215e-01 -2.44356483e-01
-8.50956440e-02 -7.75922742e-03 -5.00920594e-01 6.04478657e-01
3.00092816e-01 4.45257634e-01 -1.12368476e+00 -7.61017250e-03
-2.59237468e-01 -9.98856068e-01 8.42124343e-01 -1.12766817e-01
1.71553957e+00 -2.84124732e-01 -2.95997888e-01 7.57068992e-01
1.17391205e+00 -5.33629842e-02 6.97272718e-01 -4.42440249e-02
6.82392180e-01 6.71465993e-01 8.18796694e-01 7.65492141e-01
2.75125712e-01 9.07221615e-01 3.47938091e-01 -3.99119854e-01
2.64204413e-01 -4.18930024e-01 5.38891613e-01 5.32193780e-01
-2.10298240e-01 -5.52397743e-02 -5.59533894e-01 1.17740102e-01
-1.86734021e+00 -9.71738040e-01 1.52165428e-01 1.53974056e+00
1.01926756e+00 -2.18416929e-01 4.08536680e-02 -3.03010315e-01
5.01367748e-01 5.65907240e-01 -7.23844051e-01 -3.74735177e-01
-2.35961378e-01 2.61088550e-01 -1.41246766e-01 -3.25638540e-02
-1.03785789e+00 1.09470868e+00 5.21430826e+00 8.72143626e-01
-1.58676124e+00 -1.59176178e-02 9.32424009e-01 1.07299216e-01
-2.87915245e-02 -4.06791568e-01 -6.48745596e-01 8.14662397e-01
5.73493302e-01 -2.14928910e-01 1.61195323e-01 1.20027673e+00
1.94434822e-01 2.04963997e-01 -1.05777848e+00 1.55646574e+00
3.56762409e-01 -8.85200560e-01 1.64305046e-01 -2.17210382e-01
5.36801994e-01 -2.76873916e-01 1.46015644e-01 5.55189192e-01
-1.76828960e-03 -1.29612041e+00 5.75592220e-01 6.21110082e-01
1.05043948e+00 -6.52443528e-01 7.35850930e-01 -7.17566982e-02
-1.20329762e+00 -3.85353086e-03 -3.10167491e-01 9.28031504e-02
1.72397643e-01 3.98462832e-01 9.77008566e-02 3.89738679e-01
8.44479382e-01 1.26800311e+00 -4.95329678e-01 2.80919164e-01
-3.94517779e-01 5.08055747e-01 -1.90987214e-01 -1.19091995e-01
2.63203353e-01 -3.74306917e-01 3.14829610e-02 1.26495755e+00
4.65377346e-02 2.59036720e-01 3.81878614e-01 9.15338874e-01
-4.83286470e-01 1.96811020e-01 -5.06520271e-01 -4.86039668e-01
1.84460908e-01 1.36792707e+00 -8.64102915e-02 -2.84142733e-01
-8.19356501e-01 1.21734929e+00 3.85068178e-01 5.01257479e-01
-9.81438339e-01 -2.09192306e-01 1.28220701e+00 -2.18676046e-01
1.48298806e-02 -1.16429210e-01 4.87444609e-01 -1.36176109e+00
2.09141329e-01 -9.85478997e-01 2.37530440e-01 -8.81024003e-01
-1.52643669e+00 6.87735736e-01 -2.30990335e-01 -1.15874624e+00
-2.40384564e-01 -8.76185775e-01 -5.15353918e-01 5.26438892e-01
-1.73721671e+00 -1.39969110e+00 -8.82293582e-01 9.91385639e-01
4.55300152e-01 2.21008486e-05 8.29915345e-01 4.73045588e-01
-7.85521746e-01 9.08495188e-01 -1.71051800e-01 6.62291467e-01
8.49690437e-01 -8.39369535e-01 -4.17182148e-01 3.44916224e-01
2.71865726e-01 5.05203903e-01 3.63773406e-01 -7.86140189e-02
-1.43307197e+00 -1.12267137e+00 4.66124564e-01 -1.33752525e-01
7.72557735e-01 -3.62789065e-01 -9.48416054e-01 7.52701998e-01
-1.93707403e-02 4.30498570e-01 9.53024626e-01 1.40419379e-01
-5.94052672e-01 -4.45088238e-01 -8.14345300e-01 5.86312950e-01
1.04760313e+00 -8.51877213e-01 -2.73164570e-01 1.80384755e-01
5.31993330e-01 -2.33583480e-01 -7.02503622e-01 4.51904982e-01
6.38419092e-01 -1.08248842e+00 4.42835480e-01 -8.58548820e-01
6.91516042e-01 -1.96945935e-01 -2.12508947e-01 -1.06000197e+00
1.53937906e-01 -4.08827394e-01 -5.32314815e-02 1.45781279e+00
-6.81075752e-02 -4.32083130e-01 6.54534757e-01 6.34583771e-01
9.69527513e-02 -8.77598226e-01 -6.19992971e-01 -6.48733735e-01
-3.48663002e-01 -3.76084685e-01 7.38271177e-01 1.09544742e+00
-9.23790708e-02 3.80413175e-01 -6.24905825e-01 -3.51818919e-01
2.84555972e-01 2.10259706e-01 9.31447864e-01 -1.11332917e+00
-5.40858954e-02 -6.06319547e-01 -7.36490488e-01 -1.21878004e+00
8.71494293e-01 -7.55785823e-01 2.05712538e-04 -8.26587975e-01
4.44660723e-01 -1.14427149e-01 -5.18722534e-01 8.40978920e-01
-2.56006896e-01 2.47690961e-01 1.99174896e-01 1.60420507e-01
-9.73117828e-01 1.09857476e+00 1.18535066e+00 -2.85934091e-01
2.92961031e-01 -3.37339133e-01 -7.21624911e-01 9.88749146e-01
4.40198690e-01 -1.25452563e-01 -3.55898261e-01 -5.48820794e-01
2.22865254e-01 -7.03839213e-02 3.23309600e-01 -5.91099560e-01
-9.93407145e-02 -4.11944568e-01 3.95913363e-01 -6.52953982e-02
5.53964138e-01 -9.89313781e-01 -1.52254313e-01 -9.36627761e-02
-2.51496792e-01 7.72612588e-03 1.00147232e-01 5.67261040e-01
-8.96174729e-01 1.28832087e-01 7.57024288e-01 7.49056861e-02
-1.17406452e+00 8.84615183e-01 4.02877666e-02 -2.69110333e-02
9.35264647e-01 7.58655965e-02 -2.50519991e-01 -5.07591903e-01
-5.22997320e-01 3.76671165e-01 4.39047396e-01 7.65931427e-01
5.93130946e-01 -1.55727935e+00 -5.82405448e-01 1.15608886e-01
6.37681663e-01 -1.04998887e-01 3.56464624e-01 8.23267996e-01
-1.24701016e-01 2.53500026e-02 -4.06733632e-01 -6.36332393e-01
-1.57759571e+00 3.24792743e-01 4.50027645e-01 -7.74407461e-02
-3.48045379e-01 6.78653002e-01 4.63552445e-01 -1.05571941e-01
1.46072328e-01 8.66776556e-02 -2.50618398e-01 -4.99979667e-02
9.45051908e-01 -2.16431022e-01 -4.12533283e-01 -9.73307312e-01
-3.54823798e-01 7.15023756e-01 -1.96726456e-01 2.82579333e-01
1.15087152e+00 -6.19421676e-02 -4.44571644e-01 4.66570646e-01
1.53716421e+00 -1.27070934e-01 -1.22060263e+00 -5.08703709e-01
-6.93373680e-02 -6.66194201e-01 4.35930155e-02 -5.55015564e-01
-1.52923882e+00 1.14037609e+00 5.03176570e-01 -3.19364399e-01
1.48687434e+00 5.43351360e-02 7.85297334e-01 4.85900879e-01
3.91672313e-01 -1.05208170e+00 3.03294688e-01 2.84334958e-01
9.18497324e-01 -1.37892413e+00 -2.78017253e-01 -5.51161408e-01
-9.68001425e-01 1.23319173e+00 8.91364634e-01 2.61264563e-01
6.96912348e-01 1.79289997e-01 3.08099508e-01 -3.03346962e-01
-7.26071298e-01 -1.05665922e-01 -6.37783930e-02 3.80327255e-01
5.87738633e-01 6.63977787e-02 -1.44013271e-01 9.41088855e-01
2.03152746e-01 5.45853257e-01 -1.14032021e-02 8.09539855e-01
-5.27389273e-02 -1.05141032e+00 1.42883107e-01 2.09803805e-01
-6.25746846e-01 2.43949637e-01 -6.66028678e-01 6.44685090e-01
6.50737956e-02 8.25188100e-01 -1.00817960e-02 -3.59375358e-01
2.59713858e-01 4.95547391e-02 3.62776399e-01 -3.21270257e-01
-1.19803414e-01 -2.59781092e-01 -1.17164282e-02 -8.61849487e-01
-8.28329742e-01 -3.77546310e-01 -1.36998749e+00 -2.21417442e-01
1.11554570e-01 5.63249625e-02 3.62619281e-01 1.19081008e+00
3.74120057e-01 2.84850925e-01 9.65375125e-01 -3.97178352e-01
-4.63519633e-01 -8.02635014e-01 -7.70717323e-01 1.01324689e+00
2.55563259e-01 -1.03832018e+00 -1.75267667e-01 3.71358186e-01] | [13.66100025177002, 1.6928826570510864] |
f55ad8b4-76c5-49da-95c9-7f2c27bce820 | simple-real-time-qrs-detector-with-the-mamemi | null | null | https://www.sciencedirect.com/science/article/pii/S1746809415001032 | https://www.sciencedirect.com/science/article/pii/S1746809415001032 | Simple real-time QRS detector with the MaMeMi filter | Detection of QRS complexes in ECG signals is required to determine heart rate, and it is an important step in the study of cardiac disorders. ECG signals are usually affected by noise of low and high frequency. To improve the accuracy of QRS detectors several methods have been proposed to filter out the noise and detect the characteristic pattern of QRS complex. Most of the existing methods are at a disadvantage from relatively high computational complexity or high resource needs making them less optimized for its implementation on portable embedded systems, wearable devices or ultra-low power chips. We present a new method to detect the QRS signal in a simple way with minimal computational cost and resource needs using a novel non-linear filter. | ['David Castells-Rufas', 'Jordi Carrabina'] | 2015-08-01 | null | null | null | null | ['qrs-complex-detection'] | ['medical'] | [ 3.40010554e-01 -5.75492859e-01 1.90632075e-01 -2.88329184e-01
-1.93887800e-01 -4.66381013e-01 -4.06942546e-01 6.16309643e-01
-5.72306156e-01 6.57217681e-01 -4.25031215e-01 -1.48332953e-01
-6.23322167e-02 -7.98117697e-01 2.53684908e-01 -4.46740627e-01
4.84264269e-03 7.06860721e-02 4.76084024e-01 -1.18514104e-03
2.68012341e-02 6.19050384e-01 -1.22540319e+00 -8.12410787e-02
8.59722495e-01 7.54841030e-01 1.32627904e-01 7.55355060e-01
2.53272384e-01 1.41892329e-01 -8.82145405e-01 2.25273371e-01
-5.59586985e-03 -1.12207282e+00 -1.15704425e-01 -2.93568492e-01
-6.97089195e-01 -1.15824632e-01 2.03165591e-01 1.00028121e+00
1.07869828e+00 -3.56192559e-01 4.46976960e-01 -6.46060586e-01
4.74472612e-01 4.93254155e-01 -4.99144614e-01 6.32372260e-01
3.97017807e-01 -1.97182029e-01 6.69178963e-02 -7.61075675e-01
1.92900121e-01 5.99528015e-01 1.02582479e+00 2.55060524e-01
-1.31197572e+00 -5.31177580e-01 -6.82766974e-01 2.36715540e-01
-1.66034782e+00 -2.88550377e-01 9.30250287e-01 -1.05698332e-01
7.69834042e-01 4.68331397e-01 9.23891842e-01 4.60698962e-01
5.64819336e-01 6.00467287e-02 1.10612667e+00 -5.55686176e-01
2.83445179e-01 9.53992680e-02 3.35527092e-01 4.07599598e-01
7.22434044e-01 -2.72525430e-01 -1.74639016e-01 -2.77171165e-01
7.47547805e-01 1.53326631e-01 -4.36535329e-01 1.08078659e-01
-9.62693274e-01 6.01063907e-01 3.80163491e-02 8.98624837e-01
-4.14947331e-01 2.22570132e-02 4.30319548e-01 1.35739833e-01
1.50944665e-02 4.81298715e-01 -2.65397877e-01 -4.45296049e-01
-8.32688212e-01 -5.11234999e-03 8.09933841e-01 3.39408785e-01
3.04067105e-01 2.17777476e-01 9.14863124e-03 5.35759330e-01
1.58472657e-01 6.74992263e-01 5.60434341e-01 -2.62155145e-01
-9.19645801e-02 4.87172753e-01 1.96558721e-02 -9.79007065e-01
-1.00409031e+00 -6.64594889e-01 -8.43087673e-01 -5.49295172e-02
3.33650768e-01 -4.51615393e-01 -6.38025343e-01 1.00912380e+00
2.92016000e-01 -5.34647368e-02 -5.58876656e-02 8.38320792e-01
9.56468821e-01 6.36767328e-01 1.01274869e-04 -7.53609061e-01
1.50824153e+00 3.59667931e-03 -1.06228054e+00 -8.50707889e-02
4.04878497e-01 -8.91562402e-01 6.22929871e-01 5.57497501e-01
-7.48307109e-01 -5.26358128e-01 -1.31917596e+00 2.80240625e-01
1.22936361e-01 3.39792162e-01 1.96047470e-01 1.09101796e+00
-4.63789165e-01 6.24047816e-01 -8.67040277e-01 -3.85436058e-01
9.72594768e-02 4.04562622e-01 1.01048268e-01 4.95769173e-01
-1.05586755e+00 9.14285123e-01 2.26030946e-01 3.89475793e-01
-7.14882696e-03 -3.14045817e-01 -6.31414056e-01 7.14397430e-02
-2.46453676e-02 -3.18932146e-01 9.46394742e-01 -7.10828125e-01
-1.43222010e+00 5.56507766e-01 -2.01136172e-01 -3.70565176e-01
3.32327247e-01 -1.64912775e-01 -7.96323001e-01 2.82607257e-01
-1.57801852e-01 -4.60601479e-01 9.80590343e-01 -3.05549651e-01
-2.60263562e-01 -4.68531311e-01 -7.19857514e-01 -1.17135443e-01
8.92384350e-02 7.23129287e-02 1.26328528e-01 -3.91982704e-01
6.94504321e-01 -7.04755485e-01 -2.90247798e-01 -4.12285417e-01
-1.42088747e-02 4.46682498e-02 8.55983675e-01 -4.38569307e-01
1.41024911e+00 -2.23669791e+00 -3.71239603e-01 5.90312958e-01
1.54097956e-02 7.63684630e-01 4.18022662e-01 4.48125154e-01
1.25726014e-01 -8.97037834e-02 8.12496804e-03 6.29012167e-01
-6.34765089e-01 7.86092430e-02 7.66247362e-02 7.44567871e-01
8.86600018e-02 5.40041745e-01 -7.05449581e-01 -3.43744397e-01
4.33622539e-01 6.60659313e-01 -1.13012176e-02 6.89593479e-02
5.48156500e-01 5.81640363e-01 -4.55704629e-01 3.90731841e-01
4.85787898e-01 1.34732068e-01 3.93557400e-01 -6.69568539e-01
-2.01261163e-01 3.35695624e-01 -1.59102702e+00 1.22121966e+00
7.38514736e-02 4.62968379e-01 3.89315411e-02 -1.12983418e+00
1.24049222e+00 6.64969623e-01 4.29324389e-01 -7.56876171e-01
5.56653082e-01 4.75170135e-01 4.01419878e-01 -7.89930224e-01
-4.73841354e-02 -4.80295748e-01 5.50269447e-02 9.67755765e-02
-3.08410883e-01 -1.17786840e-01 1.79949343e-01 -3.76266420e-01
9.78496075e-01 -8.39956552e-02 7.85979569e-01 -4.41163778e-01
5.26733875e-01 -2.49174148e-01 9.20628786e-01 5.53005576e-01
-1.36600941e-01 4.86801833e-01 3.16521466e-01 -5.96445620e-01
-5.65466225e-01 -7.85344541e-01 -4.09358650e-01 2.00475946e-01
1.51431307e-01 -5.12388110e-01 -5.74771166e-01 -7.80477226e-02
-1.06877573e-01 3.70097645e-02 4.07839715e-02 -2.19621941e-01
-7.89947569e-01 -9.25973892e-01 6.32214189e-01 4.34617430e-01
4.03922260e-01 -8.64429891e-01 -1.49266863e+00 8.14430714e-01
1.15198987e-02 -9.54826117e-01 -3.15555446e-02 3.30368429e-01
-1.18324327e+00 -9.19431865e-01 -6.21289849e-01 -6.85229242e-01
6.23391330e-01 -6.37738332e-02 7.25597382e-01 2.39859998e-01
-1.00342429e+00 3.67415361e-02 -3.04670393e-01 -9.51577365e-01
-1.58667490e-01 -5.86942062e-02 1.10606842e-01 4.20500524e-02
4.15304124e-01 -5.61272562e-01 -9.41180527e-01 2.67921120e-01
-4.76186424e-01 -5.20465255e-01 6.98973835e-01 5.46198368e-01
3.67098629e-01 2.95057803e-01 1.08802485e+00 -8.86810482e-01
7.68614173e-01 -8.66408180e-03 -6.96901202e-01 -3.12968493e-01
-6.61137819e-01 -8.29826668e-02 7.52637029e-01 -2.24310264e-01
-5.90022087e-01 3.65914851e-01 -4.72950935e-01 3.13440830e-01
-1.58177599e-01 4.83627081e-01 -3.05740796e-02 -1.29526451e-01
8.66873562e-01 5.68408184e-02 2.53624707e-01 -4.66364741e-01
-4.54850405e-01 5.97966731e-01 4.93829966e-01 -1.72907002e-02
6.80138469e-01 1.17351212e-01 4.20339346e-01 -1.27806818e+00
-3.08076292e-01 -6.22226477e-01 -4.27763641e-01 -1.65870249e-01
7.76825309e-01 -7.05016434e-01 -6.62419856e-01 2.84956187e-01
-1.04963410e+00 3.44464540e-01 -4.35360754e-03 9.71661985e-01
3.75290476e-02 5.93421578e-01 -5.69776058e-01 -1.04243588e+00
-9.45648074e-01 -8.14954340e-01 2.02776223e-01 5.74707031e-01
-7.36591578e-01 -4.40801799e-01 5.11598662e-02 -2.96005428e-01
5.42509973e-01 5.99046648e-01 7.31672287e-01 -1.93563193e-01
4.44289595e-02 -6.20296359e-01 1.11184888e-01 9.41212177e-02
3.77962381e-01 -3.86029556e-02 -7.99045622e-01 -2.98479706e-01
4.74163949e-01 2.17596173e-01 3.35130244e-01 6.11991704e-01
6.53894424e-01 3.14015269e-01 -4.83615100e-01 4.13062543e-01
1.67575240e+00 7.51907587e-01 8.06648254e-01 -4.99171391e-02
2.27900282e-01 2.14209352e-02 7.31687069e-01 4.52034831e-01
-3.35882515e-01 2.80823350e-01 -1.57082424e-01 -4.85847205e-01
2.89389193e-01 2.95714885e-01 -2.62452085e-02 9.18701947e-01
-4.61501151e-01 -1.70352012e-02 -8.27722788e-01 4.56600815e-01
-1.63977540e+00 -8.49080563e-01 -9.50801969e-01 2.44877529e+00
7.84022152e-01 2.76310235e-01 3.32566589e-01 8.72186363e-01
6.42095327e-01 -4.22537416e-01 -2.76041806e-01 -6.65415287e-01
7.36750364e-02 7.95871496e-01 2.45128721e-01 7.87981227e-02
-8.18238437e-01 -1.11591499e-02 7.15693665e+00 1.02263130e-02
-1.47214556e+00 -1.92231424e-02 -1.18339378e-02 1.19329952e-02
3.18792105e-01 -1.48725390e-01 -5.79997480e-01 5.59294522e-01
9.54460621e-01 -7.17496201e-02 -2.83193469e-01 6.66030705e-01
2.47245505e-01 -4.96241629e-01 -8.55791807e-01 1.51283801e+00
-3.10993880e-01 -7.90412426e-01 -6.08031392e-01 -3.54088217e-01
6.98300228e-02 -5.05144835e-01 -4.29061085e-01 -1.63058341e-01
-1.12021840e+00 -7.35716105e-01 1.59962595e-01 4.68133986e-01
6.80833340e-01 -9.88053918e-01 1.01018453e+00 2.47927576e-01
-1.24323308e+00 8.55973735e-03 -4.44807500e-01 -5.65418839e-01
2.03434750e-01 1.19478858e+00 -7.63296127e-01 1.79556057e-01
6.20968938e-01 3.26227903e-01 -2.14155272e-01 1.32037449e+00
-1.20075531e-01 9.14388895e-01 -5.44033468e-01 -3.26819092e-01
-3.46626282e-01 -9.15987939e-02 6.19705558e-01 1.22999358e+00
4.89756703e-01 3.12782407e-01 6.23640455e-02 6.64246142e-01
4.42272991e-01 3.93277556e-01 -4.18947250e-01 4.83324751e-02
3.42621028e-01 1.28125322e+00 -1.28110933e+00 -2.06530735e-01
-4.52431887e-01 6.31676376e-01 -4.93043512e-01 -4.20475155e-02
-7.40865946e-01 -1.15258789e+00 8.58305469e-02 5.94752729e-01
-1.93016848e-03 -2.87363261e-01 -4.04676497e-01 -8.62842321e-01
2.47837067e-01 -7.76121020e-01 3.41992557e-01 -1.54497504e-01
-6.50578797e-01 6.30060732e-01 -4.20224130e-01 -1.39346623e+00
-2.19205111e-01 -2.21959502e-01 -6.87638640e-01 1.15618801e+00
-9.82135773e-01 -2.06684619e-01 -4.21639889e-01 6.27744198e-01
2.56874323e-01 1.83801606e-01 1.18936539e+00 4.42941993e-01
-3.81446660e-01 3.04211050e-01 -1.68949068e-01 1.66414976e-01
7.58358359e-01 -9.59641516e-01 -9.27126408e-02 9.56981182e-01
-3.37470949e-01 5.35720706e-01 8.45939755e-01 -6.00407600e-01
-1.43959904e+00 -4.16968524e-01 1.01879323e+00 1.27524376e-01
-4.00384590e-02 -3.76465827e-01 -8.85209262e-01 -1.79754704e-01
-9.30435658e-02 6.91112131e-02 8.81400704e-01 -1.55590624e-01
4.46497738e-01 -6.80826724e-01 -1.26174176e+00 2.78177649e-01
2.89591134e-01 -2.81972647e-01 -6.95498526e-01 -2.02500701e-01
-9.42211300e-02 -2.41120175e-01 -9.36503351e-01 6.13446057e-01
8.83080006e-01 -9.14313734e-01 5.66673160e-01 2.99702793e-01
-3.81818831e-01 -6.82474792e-01 6.88755870e-01 -9.72866714e-01
-6.11919045e-01 -9.62517679e-01 3.96309793e-01 1.07760048e+00
2.74552494e-01 -7.96237290e-01 6.76393747e-01 3.19413006e-01
2.16389090e-01 -4.05305713e-01 -8.09336185e-01 -5.69526255e-01
-8.76502156e-01 1.09193146e-01 7.70993903e-02 5.97769678e-01
4.47935373e-01 7.04767466e-01 -2.91170388e-01 -2.61597596e-02
5.98441184e-01 1.01933658e-01 1.86250821e-01 -1.57550395e+00
-1.89305827e-01 -2.78837066e-02 -9.04454112e-01 -2.23530993e-01
-8.01403224e-01 -2.91326195e-01 6.98243082e-03 -1.49449289e+00
-1.44526169e-01 -9.19277072e-02 -3.82946789e-01 5.06011434e-02
-2.92018265e-01 6.35218084e-01 -1.81102216e-01 -4.54339683e-02
-6.44454658e-02 -2.42541283e-01 6.86164200e-01 2.72266984e-01
-7.10912406e-01 2.97469884e-01 -3.07557732e-01 9.55572486e-01
8.83391142e-01 -8.69325459e-01 -6.04673326e-01 7.22006932e-02
4.56920445e-01 2.28438780e-01 -1.22030877e-01 -1.48463583e+00
8.83188471e-02 2.70651251e-01 8.77734125e-01 -8.27749968e-01
4.35360819e-02 -8.97289991e-01 5.40738523e-01 1.20452523e+00
2.81442225e-01 1.51724309e-01 1.51769504e-01 2.53155529e-01
-1.83045000e-01 -3.84915620e-01 9.78900671e-01 -1.68638110e-01
-2.90881276e-01 -2.69823611e-01 -1.10238326e+00 -2.51243621e-01
9.83084083e-01 -5.90694427e-01 1.28050864e-01 -1.46039248e-01
-7.99705505e-01 -2.13859811e-01 2.36657020e-02 -6.57016635e-02
7.74777710e-01 -9.17208791e-01 -5.31883955e-01 4.58663404e-01
-1.17071107e-01 -2.58403093e-01 2.11115539e-01 1.16437507e+00
-1.01693392e+00 3.00525665e-01 -6.11877441e-01 -7.40492404e-01
-1.74016976e+00 2.73622692e-01 4.19721872e-01 4.32761833e-02
-9.36646402e-01 4.93799090e-01 -5.56339920e-01 6.56305075e-01
5.49258199e-03 -4.94932890e-01 -4.68168169e-01 9.39371884e-02
7.60420501e-01 6.88158333e-01 2.87132889e-01 -1.52592599e-01
-7.77886331e-01 9.95707095e-01 3.34449142e-01 1.98758945e-01
1.00577700e+00 -3.04249208e-02 -2.71562010e-01 6.38765156e-01
8.34459662e-01 2.10069150e-01 -2.16805518e-01 1.50590271e-01
4.20619845e-02 -2.51642108e-01 -2.18114704e-02 -3.99423689e-01
-6.56312704e-01 8.20293427e-01 1.09900868e+00 3.79556954e-01
1.51107097e+00 -5.46317816e-01 7.40278125e-01 2.83956856e-01
5.02064228e-01 -1.14013040e+00 -1.75931871e-01 -2.56764349e-02
5.48401415e-01 -6.58956528e-01 4.40525293e-01 -5.45159936e-01
-1.63450837e-01 1.49466288e+00 1.30686447e-01 -3.56592864e-01
9.81948614e-01 6.54220402e-01 3.80876660e-01 -5.96807860e-02
-1.94937453e-01 -2.53079265e-01 1.02680743e-01 6.89675331e-01
9.09298182e-01 -2.53118463e-02 -1.52281559e+00 7.25476563e-01
2.40906328e-01 4.18497235e-01 6.93713367e-01 1.21536815e+00
-5.74420094e-01 -1.20357621e+00 -4.88690883e-01 7.12057173e-01
-1.15901458e+00 3.15431237e-01 -2.03607425e-01 3.05357009e-01
1.78843632e-01 1.15402877e+00 -2.07915470e-01 -2.14752436e-01
5.70793867e-01 2.27652863e-01 5.51046550e-01 -4.84537125e-01
-7.76640475e-01 6.09769166e-01 4.51227948e-02 -4.85478908e-01
-3.97994041e-01 -3.86041552e-01 -1.59073341e+00 1.72175229e-01
-4.95987177e-01 3.20467263e-01 7.49892831e-01 6.95912302e-01
3.42130482e-01 6.03850782e-01 3.60670000e-01 -2.73441374e-01
-3.23119611e-01 -9.66484725e-01 -1.06014061e+00 2.87025928e-01
2.81639576e-01 -2.49817282e-01 -6.80795982e-02 -1.69716366e-02] | [14.099815368652344, 3.1734707355499268] |
8dead145-a7f0-4f22-a305-862800f778d6 | combining-subgoal-graphs-with-reinforcement | 1811.01700 | null | http://arxiv.org/abs/1811.01700v1 | http://arxiv.org/pdf/1811.01700v1.pdf | Combining Subgoal Graphs with Reinforcement Learning to Build a Rational Pathfinder | In this paper, we present a hierarchical path planning framework called SG-RL
(subgoal graphs-reinforcement learning), to plan rational paths for agents
maneuvering in continuous and uncertain environments. By "rational", we mean
(1) efficient path planning to eliminate first-move lags; (2) collision-free
and smooth for agents with kinematic constraints satisfied. SG-RL works in a
two-level manner. At the first level, SG-RL uses a geometric path-planning
method, i.e., Simple Subgoal Graphs (SSG), to efficiently find optimal abstract
paths, also called subgoal sequences. At the second level, SG-RL uses an RL
method, i.e., Least-Squares Policy Iteration (LSPI), to learn near-optimal
motion-planning policies which can generate kinematically feasible and
collision-free trajectories between adjacent subgoals. The first advantage of
the proposed method is that SSG can solve the limitations of sparse reward and
local minima trap for RL agents; thus, LSPI can be used to generate paths in
complex environments. The second advantage is that, when the environment
changes slightly (i.e., unexpected obstacles appearing), SG-RL does not need to
reconstruct subgoal graphs and replan subgoal sequences using SSG, since LSPI
can deal with uncertainties by exploiting its generalization ability to handle
changes in environments. Simulation experiments in representative scenarios
demonstrate that, compared with existing methods, SG-RL can work well on
large-scale maps with relatively low action-switching frequencies and shorter
path lengths, and SG-RL can deal with small changes in environments. We further
demonstrate that the design of reward functions and the types of training
environments are important factors for learning feasible policies. | ['Long Qin', 'Cong Hu', 'Yue Hu', 'Junjie Zeng', 'Quanjun Yin'] | 2018-11-05 | null | null | null | null | ['optimal-motion-planning'] | ['robots'] | [-1.50485441e-01 3.78122002e-01 -2.54355490e-01 1.25234872e-01
-6.26278877e-01 -4.59656358e-01 2.20933959e-01 -2.44747400e-02
-4.50966537e-01 1.26452887e+00 9.33707952e-02 -5.33250093e-01
-4.33566153e-01 -1.02091360e+00 -7.63363957e-01 -7.42806554e-01
-6.83333397e-01 5.42367101e-01 6.26843929e-01 -6.12901211e-01
3.40022683e-01 6.66410744e-01 -1.33520865e+00 -4.34505075e-01
1.25330293e+00 4.58545953e-01 7.76754975e-01 4.46895450e-01
5.53242713e-02 7.28789508e-01 -2.47395739e-01 4.20110643e-01
4.85144347e-01 -2.74841577e-01 -7.31626928e-01 -1.25343770e-01
-5.51327884e-01 -5.65443516e-01 -4.56569433e-01 8.56541693e-01
2.49988526e-01 6.70932472e-01 3.49715382e-01 -1.40314388e+00
-8.00277144e-02 5.46329021e-01 -2.98648328e-01 -1.02427945e-01
4.60883766e-01 5.74356318e-01 7.51167595e-01 -4.34130609e-01
5.91218412e-01 1.51540804e+00 4.47046578e-01 5.66502333e-01
-9.13269043e-01 -4.47487682e-01 6.89890027e-01 2.73599088e-01
-1.27133858e+00 1.60294888e-03 3.88033211e-01 -5.04544228e-02
1.12834001e+00 3.97152714e-02 7.75985599e-01 7.07459807e-01
6.76910341e-01 6.89238131e-01 9.19720888e-01 -1.79420924e-03
6.43513501e-01 -5.17588079e-01 -5.04419267e-01 7.67964542e-01
4.05153871e-01 6.98621273e-01 -3.00729368e-02 -6.31285533e-02
8.45464766e-01 -1.73288047e-01 -2.59958178e-01 -4.92201179e-01
-1.41365588e+00 8.04221570e-01 5.83956897e-01 8.98349062e-02
-6.58197045e-01 3.81896019e-01 7.11417496e-02 1.98360324e-01
-3.55958194e-01 7.90994287e-01 -3.17759722e-01 -2.94515401e-01
-2.42620245e-01 8.73390555e-01 6.30225718e-01 1.23858261e+00
8.53357077e-01 4.39527303e-01 -7.63243139e-02 3.06902230e-01
3.63546878e-01 7.70226479e-01 2.91703939e-01 -1.27607262e+00
7.10967243e-01 3.47393632e-01 6.97579205e-01 -1.04628241e+00
-8.52826178e-01 -1.64810568e-01 -5.00872016e-01 5.84183216e-01
4.18546438e-01 -4.14513141e-01 -9.74074662e-01 1.87453938e+00
4.13341671e-01 5.23771681e-02 4.47270185e-01 8.49255562e-01
2.85951674e-01 8.85177732e-01 -9.79695562e-03 -3.11726958e-01
7.88930655e-01 -1.08101881e+00 -6.99574411e-01 -5.39893329e-01
6.03335321e-01 -8.57343078e-02 1.12680221e+00 2.90288866e-01
-9.76530612e-01 -3.37060511e-01 -9.98613358e-01 6.07601881e-01
-1.71020865e-01 -4.30296183e-01 6.63164914e-01 1.96963996e-01
-1.06120288e+00 7.47876406e-01 -9.95751023e-01 -3.12801212e-01
2.74446346e-02 3.85988146e-01 -7.96323456e-03 -1.30849451e-01
-1.26793993e+00 1.02096081e+00 8.51893961e-01 -1.00057766e-01
-1.41812825e+00 -2.42841825e-01 -1.02479279e+00 8.21023732e-02
9.29471552e-01 -3.81135076e-01 1.34031916e+00 -4.62424278e-01
-1.87594426e+00 -3.95133793e-01 -1.72622472e-01 -2.82731384e-01
3.84836406e-01 -5.85153773e-02 -2.67086029e-01 7.12115839e-02
2.94893444e-01 7.72153676e-01 6.31309092e-01 -1.51908183e+00
-1.02931213e+00 -2.21322905e-02 1.08050555e-01 7.65275657e-01
2.49442771e-01 -6.12980008e-01 -4.32310045e-01 -2.94880569e-01
3.01511317e-01 -1.25772107e+00 -1.13718295e+00 -5.15637577e-01
-3.19934130e-01 -1.10657439e-01 5.76934040e-01 -3.74252617e-01
1.06355333e+00 -1.83202529e+00 3.40260059e-01 5.02444863e-01
-1.28911465e-01 4.87204529e-02 -4.30162013e-01 8.76150966e-01
4.77155775e-01 -1.73589796e-01 -1.93435535e-01 7.74219334e-02
-1.33404182e-02 6.71182454e-01 -3.67043734e-01 5.29528558e-02
-2.18151599e-01 9.66598153e-01 -1.42916083e+00 -1.85410187e-01
1.33081377e-01 -7.62670487e-02 -4.78545845e-01 1.14874609e-01
-5.36384344e-01 8.15397799e-01 -1.01034570e+00 5.23716271e-01
4.68081146e-01 1.24186806e-01 3.08496505e-01 4.13312465e-01
-6.63698971e-01 2.09959105e-01 -1.35458767e+00 1.57233703e+00
-2.93603092e-01 7.96980709e-02 1.19547710e-01 -7.48624206e-01
1.00731099e+00 5.71830794e-02 5.14616072e-01 -8.00105572e-01
-5.05514815e-02 4.78906065e-01 -8.00326467e-02 -3.64706546e-01
7.42297351e-01 2.20336482e-01 -3.15261334e-01 3.42027009e-01
-5.34487367e-01 -3.30408931e-01 2.50805229e-01 8.15554857e-02
1.22114074e+00 4.21348393e-01 4.72859979e-01 -2.07527786e-01
4.79419619e-01 5.01533210e-01 9.66312706e-01 9.63839948e-01
-2.58563071e-01 -1.61120236e-01 2.99855113e-01 -4.10115331e-01
-6.45581722e-01 -1.16316938e+00 5.12076974e-01 7.84503460e-01
9.49828386e-01 -2.56963074e-01 -5.48722148e-01 -4.30757076e-01
5.35706840e-02 9.82128382e-01 -2.51164585e-01 -1.59228638e-01
-9.78419840e-01 -4.18843538e-01 1.70492560e-01 4.50447023e-01
6.31538212e-01 -1.29433298e+00 -1.15241599e+00 6.65776968e-01
-3.33273619e-01 -9.42071974e-01 -5.00341475e-01 7.92117268e-02
-9.17572975e-01 -1.12899768e+00 -3.72247964e-01 -6.90486193e-01
8.21070552e-01 6.71197712e-01 5.52509546e-01 6.27950430e-02
2.20363393e-01 3.72609496e-01 -5.71887314e-01 -1.82904735e-01
-3.29886973e-01 -1.24997251e-01 2.23184288e-01 -4.61321950e-01
-1.97425902e-01 -5.16838431e-01 -4.86964613e-01 6.64243996e-01
-5.75607002e-01 2.30034903e-01 8.50564122e-01 7.20850050e-01
9.27711308e-01 6.05223119e-01 8.11033428e-01 -3.58644426e-01
8.41714263e-01 -3.88514459e-01 -9.40152049e-01 1.97774455e-01
-6.75138891e-01 1.83351740e-01 9.88127470e-01 -4.25303847e-01
-1.11536324e+00 1.04899190e-01 1.36312831e-03 -1.17815197e-01
-1.07405275e-01 6.55392945e-01 -3.30160946e-01 -2.26234674e-01
5.13629854e-01 4.17991698e-01 2.67726392e-03 5.71392551e-02
4.92390454e-01 1.12170063e-01 3.67322028e-01 -7.19948351e-01
9.78058040e-01 3.57005596e-01 3.36125940e-01 -5.72374463e-01
-1.71828076e-01 -1.98162153e-01 -3.41239929e-01 -3.89549106e-01
6.51017368e-01 -6.79325521e-01 -9.86901522e-01 4.25586849e-01
-7.60622442e-01 -9.84965742e-01 -2.45473251e-01 5.69483817e-01
-1.08173239e+00 4.04690206e-01 -3.48325610e-01 -1.06580722e+00
3.61460336e-02 -1.28316188e+00 4.65864897e-01 5.03232718e-01
-1.57824848e-02 -7.62990832e-01 4.67361533e-04 -2.15053037e-01
4.08707470e-01 5.39912462e-01 7.33643413e-01 -1.95018366e-01
-1.09951031e+00 2.57526040e-01 1.46061555e-01 -4.53934282e-01
2.15864420e-01 -5.37230134e-01 -1.11719631e-01 -7.28975594e-01
-2.78647810e-01 -2.98393399e-01 2.92768747e-01 4.65096474e-01
7.33810544e-01 -6.45274758e-01 -7.65948892e-01 4.05622572e-01
1.34433663e+00 1.03918314e+00 6.74930215e-01 7.46027231e-01
4.08185154e-01 6.80521905e-01 1.45719337e+00 4.78917181e-01
9.10975337e-01 7.72567868e-01 8.30464602e-01 1.07340641e-01
3.13445836e-01 -8.09378862e-01 5.06899536e-01 3.63398492e-01
-1.81905180e-01 -6.43561482e-02 -8.00442398e-01 4.84913319e-01
-2.40307307e+00 -9.50079799e-01 -6.78242519e-02 2.27562356e+00
3.95441830e-01 2.21592009e-01 2.28234515e-01 -1.97144717e-01
4.21697050e-01 1.05891442e-02 -9.50906932e-01 -1.41304567e-01
2.54332244e-01 -3.33110303e-01 8.40482235e-01 8.80705297e-01
-7.52686560e-01 1.34943521e+00 5.56360197e+00 6.09709263e-01
-7.70148158e-01 -1.62242740e-01 6.06490672e-03 2.79592216e-01
-3.62527221e-01 1.87563285e-01 -8.41836274e-01 3.88415277e-01
5.82269728e-01 -1.65420324e-01 9.37772512e-01 9.57201481e-01
6.67937756e-01 -3.46080512e-01 -6.59133196e-01 5.02173424e-01
-5.19574583e-01 -1.18513501e+00 -2.37459615e-01 8.20789486e-02
7.29563773e-01 -7.37523660e-02 -2.11426452e-01 6.32559955e-01
9.82650518e-01 -9.24878955e-01 9.23027694e-01 1.91359743e-01
5.04241943e-01 -1.14740348e+00 5.61788321e-01 8.41462672e-01
-1.56906497e+00 -4.87273365e-01 -5.08875370e-01 -8.24638829e-02
7.83712924e-01 -4.14831191e-02 -9.07934308e-01 9.78622794e-01
6.54767096e-01 4.45217162e-01 6.38995767e-02 1.16818213e+00
-7.08649516e-01 7.77460560e-02 -3.19244832e-01 -4.06716287e-01
7.40848005e-01 -4.80201781e-01 8.82479787e-01 6.44627035e-01
5.37024140e-01 4.11243290e-01 8.68239045e-01 8.39050114e-01
9.42899823e-01 -1.74674302e-01 -7.41020262e-01 1.38078898e-01
8.94510627e-01 8.25997949e-01 -7.81050742e-01 -8.10381374e-04
2.29703952e-02 5.89002311e-01 3.69738519e-01 7.94363618e-01
-8.86871099e-01 -4.72676784e-01 6.32734418e-01 -1.49412602e-01
2.59829402e-01 -7.21450508e-01 -3.06187887e-02 -6.04233325e-01
-2.91732520e-01 -8.40740860e-01 2.20985264e-01 -5.46567440e-01
-6.02774024e-01 6.36830389e-01 3.08388770e-01 -1.32751632e+00
-5.77381313e-01 -2.59417892e-01 -6.20899320e-01 6.81858540e-01
-1.70911539e+00 -8.62939298e-01 -3.44981939e-01 8.17970872e-01
6.07852638e-01 -1.36972964e-01 6.24646902e-01 -3.82244617e-01
-2.58465737e-01 1.41981483e-01 -3.43842804e-02 -3.60683829e-01
8.80565196e-02 -1.09397733e+00 5.25158107e-01 9.13848102e-01
-4.16633189e-01 3.84428382e-01 7.71851122e-01 -1.15058959e+00
-1.55030775e+00 -1.15075362e+00 3.31534863e-01 1.92280948e-01
3.24107766e-01 4.77705412e-02 -8.47489476e-01 6.81285918e-01
-3.40710521e-01 -5.71391344e-01 1.00333601e-01 -2.49889433e-01
3.10698450e-01 3.03470716e-03 -1.08832610e+00 1.18055463e+00
1.20291364e+00 3.98948401e-01 -3.06941688e-01 2.72220910e-01
9.29895043e-01 -8.28338563e-01 -4.75262076e-01 5.71446836e-01
3.27965528e-01 -7.69600689e-01 9.40702200e-01 -4.23170239e-01
-2.61728257e-01 -7.78964877e-01 -6.22980557e-02 -1.78798330e+00
-6.60383701e-01 -1.00139105e+00 -9.18304324e-02 5.36271632e-01
4.39694285e-01 -9.91231918e-01 7.05075920e-01 3.74058306e-01
-6.43195808e-01 -8.87189567e-01 -9.68875408e-01 -1.29333794e+00
-1.10698067e-01 -3.59341949e-01 9.90707755e-01 4.02431697e-01
1.49847642e-01 -1.05308115e-01 -5.07260323e-01 5.44592023e-01
6.26193821e-01 8.48864838e-02 9.99977648e-01 -7.17489481e-01
-3.42282146e-01 -3.05508256e-01 1.74640238e-01 -1.36895478e+00
1.27009064e-01 -4.43872750e-01 7.30296493e-01 -1.96693516e+00
-5.45162618e-01 -1.03435302e+00 1.27671555e-01 6.01513386e-01
-1.42493725e-01 -8.44740748e-01 3.51290315e-01 4.41799074e-01
-5.40726423e-01 9.52677667e-01 1.84220576e+00 1.30569831e-01
-9.14444029e-01 2.56707788e-01 -5.08282483e-01 6.17730975e-01
1.01388848e+00 -2.19845071e-01 -9.91244793e-01 -2.65934914e-01
-4.07610759e-02 6.58946097e-01 -9.33203921e-02 -1.07280374e+00
2.48975128e-01 -1.09515274e+00 -2.24473834e-01 -6.96615040e-01
2.94139594e-01 -8.14921796e-01 3.45962703e-01 1.03120446e+00
-4.43484001e-02 2.41600528e-01 2.69120932e-01 9.73494232e-01
8.72708410e-02 -2.74188519e-01 4.87643898e-01 -2.52331078e-01
-1.19318533e+00 4.00741071e-01 -8.73367012e-01 -4.55965996e-02
1.52617252e+00 -3.76666099e-01 -2.72980630e-01 -6.73563957e-01
-4.49688077e-01 1.02338696e+00 4.80957836e-01 3.02858204e-01
9.97853458e-01 -1.27966535e+00 -3.59281868e-01 3.66672091e-02
-1.30780652e-01 4.35290277e-01 1.85838833e-01 6.83346629e-01
-5.37398994e-01 5.59094131e-01 -3.16099703e-01 -3.35255623e-01
-6.23484671e-01 5.87775767e-01 2.36245751e-01 -5.96079826e-01
-9.64240074e-01 4.79169309e-01 2.12153584e-01 -6.79241061e-01
1.95332631e-01 -3.25812846e-01 -3.33168745e-01 -6.29570127e-01
5.91697276e-01 4.26460773e-01 -3.80315572e-01 -3.16755563e-01
-2.07718670e-01 4.72098738e-01 2.05574319e-01 -4.12663668e-01
1.11121476e+00 -2.78223515e-01 2.27671295e-01 9.11629051e-02
1.75821707e-01 -1.58388302e-01 -1.95059931e+00 9.89490896e-02
1.10921443e-01 -6.06461465e-01 -2.03174382e-01 -7.57781148e-01
-6.20472550e-01 2.53725350e-01 1.11103199e-01 1.59333516e-02
9.27443445e-01 -3.01080346e-01 9.40121233e-01 7.70378351e-01
1.27872741e+00 -1.14606094e+00 3.17456484e-01 9.08519566e-01
9.32891905e-01 -7.93559313e-01 -2.19324380e-01 -4.65916127e-01
-1.01680756e+00 9.34812486e-01 9.78245199e-01 2.52035037e-02
8.27162638e-02 7.95319155e-02 -6.98780566e-02 -3.84519324e-02
-6.25648201e-01 -4.23697680e-01 -1.14338413e-01 1.09700704e+00
-5.80982029e-01 3.50659430e-01 -3.29475671e-01 3.03896934e-01
-2.22033083e-01 -1.86755434e-01 6.07295752e-01 1.07486820e+00
-1.07440257e+00 -1.01631665e+00 -5.41936934e-01 6.61625937e-02
4.44484413e-01 4.22740012e-01 1.08760111e-01 1.11983800e+00
-9.03514996e-02 1.17340219e+00 -1.86451763e-01 -4.97203916e-01
5.19820333e-01 -5.05250156e-01 3.93990189e-01 -6.08956635e-01
-7.42936283e-02 -4.17658426e-02 3.05795163e-01 -1.07405424e+00
-4.06780429e-02 -7.01808691e-01 -2.07643247e+00 -2.87113041e-01
-4.07312997e-02 3.01387250e-01 3.00092548e-01 1.00547802e+00
2.64735162e-01 5.52187920e-01 6.38242424e-01 -1.08423781e+00
-8.21573794e-01 -4.53371853e-01 -3.65823120e-01 4.22024466e-02
1.67493016e-01 -1.09254336e+00 -4.77300957e-02 -5.84403336e-01] | [4.844281196594238, 1.4902842044830322] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.