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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
da89219b-9797-42d6-94dc-facab9d7197f | nndetection-for-intracranial-aneurysms | 2305.13398 | null | https://arxiv.org/abs/2305.13398v1 | https://arxiv.org/pdf/2305.13398v1.pdf | nnDetection for Intracranial Aneurysms Detection and Localization | Intracranial aneurysms are a commonly occurring and life-threatening condition, affecting approximately 3.2% of the general population. Consequently, detecting these aneurysms plays a crucial role in their management. Lesion detection involves the simultaneous localization and categorization of abnormalities within medical images. In this study, we employed the nnDetection framework, a self-configuring framework specifically designed for 3D medical object detection, to detect and localize the 3D coordinates of aneurysms effectively. To capture and extract diverse features associated with aneurysms, we utilized TOF-MRA and structural MRI, both obtained from the ADAM dataset. The performance of our proposed deep learning model was assessed through the utilization of free-response receiver operative characteristics for evaluation purposes. The model's weights and 3D prediction of the bounding box of TOF-MRA are publicly available at https://github.com/orouskhani/AneurysmDetection. | ['Chengcheng Zhu', 'Mahmud Mossa-Basha', 'Shaojun Xia', 'Negar Firoozeh', 'Maysam Orouskhani'] | 2023-05-22 | null | null | null | null | ['medical-object-detection'] | ['computer-vision'] | [-5.40746868e-01 -1.36184096e-01 1.89770192e-01 -5.32421112e-01
-8.39405894e-01 -5.20615876e-01 3.31111968e-01 3.02929491e-01
-4.11661476e-01 3.96644890e-01 2.92198569e-01 -5.35099328e-01
-1.55775189e-01 -5.23606598e-01 -5.36126554e-01 -7.90761054e-01
-9.15699482e-01 6.80891931e-01 1.52296990e-01 2.73496747e-01
4.34715658e-01 1.12253916e+00 -1.11535633e+00 2.43399262e-01
7.92999089e-01 1.52473986e+00 2.14446425e-01 5.96229076e-01
4.23667654e-02 7.23603070e-01 -4.50600892e-01 -9.44884196e-02
3.35929185e-01 3.32504153e-01 -4.02452677e-01 -2.52317414e-02
3.14865798e-01 -7.47743130e-01 -4.75036025e-01 8.01113069e-01
7.06761003e-01 -7.68645927e-02 9.16741908e-01 -8.58372033e-01
-3.69850487e-01 3.42766523e-01 -4.00846869e-01 1.14150476e+00
-1.08928509e-01 9.73832756e-02 5.34453869e-01 -1.33383262e+00
2.86545306e-01 8.57310116e-01 3.45076114e-01 1.87836871e-01
-6.11283302e-01 -6.14417136e-01 -1.16620816e-01 2.13189855e-01
-1.33341753e+00 -5.51918566e-01 5.38259447e-01 -8.61611366e-01
6.36219382e-01 2.62624413e-01 7.85183370e-01 1.07301319e+00
6.75265729e-01 7.16328382e-01 9.26529884e-01 -3.99440005e-02
2.06645012e-01 -7.62368962e-02 1.27500549e-01 7.61322677e-01
4.70790893e-01 3.74658078e-01 -4.23598439e-02 -5.25183380e-01
1.27182317e+00 1.96492121e-01 6.05370626e-02 -6.16870999e-01
-1.01975369e+00 9.70604479e-01 4.47295964e-01 9.02713239e-02
-6.15094781e-01 -1.76696256e-01 4.54210609e-01 -2.20924869e-01
4.34643894e-01 4.77165848e-01 -4.00760889e-01 -6.21009022e-02
-4.34635133e-01 2.60378212e-01 2.34234512e-01 6.19811952e-01
-3.46190482e-01 -2.66050398e-02 -3.43259834e-02 7.37780809e-01
2.25574255e-01 5.14363050e-01 7.40171671e-01 -6.79029346e-01
8.58375579e-02 5.04116654e-01 1.47467300e-01 -9.80201840e-01
-1.07527924e+00 -1.00376725e+00 -7.08105922e-01 1.49891824e-01
5.81281424e-01 -1.91415608e-01 -1.09727001e+00 1.28240442e+00
4.74462599e-01 9.26782284e-03 -3.23604554e-01 1.30969632e+00
1.12553954e+00 7.68496618e-02 1.86358750e-01 5.77812409e-03
1.56924224e+00 -6.81467354e-01 -6.34350255e-02 1.54115623e-02
6.65747583e-01 -5.30081093e-01 4.75534856e-01 1.48620546e-01
-1.10704136e+00 -4.85589765e-02 -6.28683269e-01 1.52729169e-01
-3.15563411e-01 2.71218717e-01 6.69715166e-01 3.28226089e-01
-6.76648498e-01 -5.35509847e-02 -1.13110137e+00 6.40847683e-02
8.13615680e-01 1.87016562e-01 -4.31284994e-01 2.49104396e-01
-1.07973969e+00 1.23447967e+00 1.49504200e-01 4.25059982e-02
-9.80942428e-01 -7.89746523e-01 -6.48535073e-01 -7.63636082e-02
-9.60280001e-03 -5.84521294e-01 1.16340327e+00 -1.85676783e-01
-9.32342529e-01 1.10264659e+00 -6.68854192e-02 -6.68722391e-01
8.18917692e-01 -2.88465351e-01 -3.73877168e-01 4.65467542e-01
1.73831567e-01 3.81150633e-01 9.57968533e-01 -9.01054919e-01
-7.60218143e-01 -7.35256791e-01 -1.95556387e-01 -5.14543168e-02
9.44111198e-02 2.21017614e-01 1.66392416e-01 -1.04815829e+00
6.48786664e-01 -7.62228251e-01 -5.98456085e-01 6.17972948e-02
-5.21607697e-01 1.77963451e-01 5.40175080e-01 -1.04302442e+00
7.71791399e-01 -2.08031011e+00 -2.31847480e-01 3.82694691e-01
9.96436179e-01 1.91061497e-01 3.29635590e-01 -2.86934435e-01
-3.48401695e-01 2.69963704e-02 -1.52507290e-01 4.62816358e-02
-4.55426723e-01 -5.08100092e-01 -1.58046678e-01 8.21174920e-01
1.70842811e-01 8.69759560e-01 -5.19179106e-01 -4.08186108e-01
3.64941418e-01 5.41481078e-01 -2.21733600e-01 2.01920673e-01
4.16396379e-01 9.65367377e-01 -8.48104596e-01 9.11512315e-01
5.59770465e-01 -3.01658601e-01 -5.73821783e-01 8.21995176e-03
-7.96163082e-02 2.39520893e-02 -7.82970607e-01 1.25626075e+00
-8.07886571e-02 4.06650871e-01 -5.62020838e-02 -1.14118612e+00
9.80744064e-01 4.26907420e-01 9.70948279e-01 -7.42822468e-01
6.22704268e-01 5.24376154e-01 3.12802315e-01 -8.38185668e-01
-7.46404454e-02 1.28452197e-01 4.98751272e-03 2.34645680e-01
-4.62612897e-01 3.60100508e-01 -1.53211281e-01 4.68043871e-02
1.15710807e+00 -6.21460438e-01 6.33859038e-01 -2.23810032e-01
5.29215634e-01 8.00565109e-02 3.03439438e-01 9.16558206e-01
-6.69117272e-01 6.85511589e-01 4.10298258e-01 -9.79712188e-01
-1.10108578e+00 -1.41757572e+00 -1.04595459e+00 4.23612475e-01
3.29888053e-02 2.82916367e-01 -4.22533751e-01 -5.71503639e-01
3.74985218e-01 4.28363949e-01 -8.55868280e-01 -1.44552156e-01
-6.93172991e-01 -8.37494731e-01 1.73139676e-01 6.05369508e-01
1.78982496e-01 -7.02360094e-01 -1.31996763e+00 2.25927636e-01
9.28107556e-03 -1.03727078e+00 -6.59004897e-02 -2.02140026e-02
-9.46116924e-01 -1.28892398e+00 -1.03650725e+00 -7.48002350e-01
9.14888501e-01 -6.78850487e-02 6.74663246e-01 -2.88100857e-02
-9.33649063e-01 4.82646674e-02 -2.80222595e-01 -5.06560862e-01
-3.44309837e-01 2.08246242e-02 1.81604326e-01 -1.04951605e-01
1.57771289e-01 -6.08058214e-01 -1.12637198e+00 3.58838111e-01
-3.49066854e-01 3.13199572e-02 8.16266894e-01 6.62062585e-01
5.20853281e-01 -3.13670427e-01 7.75395274e-01 -4.94646937e-01
4.46942478e-01 -6.91535354e-01 -6.38570726e-01 -2.08389580e-01
-8.66493657e-02 -3.78122330e-01 3.99408340e-01 -4.73979414e-01
-5.34392595e-01 1.55409440e-01 -2.11079475e-02 -6.14840806e-01
-5.16851246e-01 4.39096749e-01 3.39444339e-01 -2.01926008e-01
5.84144354e-01 5.23956157e-02 7.36123472e-02 -4.02801543e-01
4.06208560e-02 7.68675506e-01 6.13972187e-01 -4.26907450e-01
3.46317858e-01 5.53558707e-01 1.34043708e-01 -5.74620545e-01
-6.02266788e-01 -5.30334115e-01 -5.38919270e-01 -3.94249797e-01
6.84166849e-01 -7.19397604e-01 -5.14404356e-01 3.42150271e-01
-9.42334592e-01 3.55344057e-01 4.18837182e-03 9.29654598e-01
-4.70019341e-01 -8.77061859e-03 -6.23229802e-01 -7.82531500e-01
-7.37779915e-01 -1.64725125e+00 9.32170749e-01 6.69685379e-02
-2.50557400e-02 -7.19409227e-01 -3.10386330e-01 2.42914796e-01
7.93551266e-01 5.78660309e-01 1.03897715e+00 -1.13861299e+00
-4.48092729e-01 -8.17941725e-01 -3.77140790e-01 -1.13343045e-01
1.03351725e-02 -1.94692537e-01 -7.37112582e-01 -2.11559340e-01
-6.71242084e-03 1.42360583e-01 5.83609819e-01 1.17648065e+00
1.37970102e+00 -1.10510699e-02 -4.18401361e-01 7.04891920e-01
8.72031808e-01 6.50316238e-01 9.44197848e-02 6.27210379e-01
5.21704555e-01 3.84433717e-01 2.16137245e-01 8.68892848e-01
2.29141593e-01 3.82612705e-01 6.68735743e-01 2.52008047e-02
1.40651718e-01 3.07473272e-01 -4.12912369e-01 5.39737821e-01
-5.64452484e-02 1.63115948e-01 -1.42232168e+00 5.60276270e-01
-1.52831256e+00 -5.59869349e-01 -1.14053845e-01 2.10920954e+00
2.58509606e-01 3.09711576e-01 1.71728298e-01 -3.59227806e-01
8.12972009e-01 -1.81224152e-01 -8.74477506e-01 5.50154336e-02
-2.52341125e-02 -1.38041809e-01 6.20233178e-01 1.30364508e-03
-1.63476121e+00 3.93780708e-01 6.20106411e+00 3.21543187e-01
-1.47856557e+00 1.35811836e-01 9.25831556e-01 -2.92233676e-01
3.65744054e-01 -4.55865562e-01 -5.14075279e-01 5.22231877e-01
7.49223053e-01 -2.45931000e-01 -2.36519232e-01 1.04251420e+00
5.55718362e-01 -8.94502029e-02 -6.56414747e-01 9.37605381e-01
-1.92396790e-01 -1.38804924e+00 -6.24889657e-02 5.37970625e-02
2.29214579e-01 1.89975366e-01 3.69784534e-01 2.07126252e-02
-7.78845996e-02 -1.01841295e+00 7.92335987e-01 5.52839816e-01
7.10060596e-01 -6.01021111e-01 8.35842669e-01 1.30058810e-01
-8.51611912e-01 -3.49157453e-01 1.37066608e-02 2.00456142e-01
4.60528553e-01 7.08649695e-01 -1.09397733e+00 -5.98316034e-03
9.82055187e-01 5.30457735e-01 -4.36225593e-01 1.76757741e+00
-2.22864076e-02 4.76689130e-01 -2.57211477e-01 4.16174293e-01
1.05337553e-01 1.72301322e-01 1.14498866e+00 8.06013405e-01
4.22471017e-01 2.46430576e-01 2.18316868e-01 8.52502465e-01
3.34680378e-01 3.33944112e-01 -3.55167329e-01 3.55461657e-01
5.59006810e-01 1.17031205e+00 -8.98377955e-01 -3.82228017e-01
-2.18196377e-01 1.81304906e-02 2.78438598e-01 1.21421389e-01
-9.46862996e-01 -2.38653868e-02 6.53654993e-01 3.29391688e-01
1.92720830e-01 -1.65300086e-01 -7.10811377e-01 -1.12963879e+00
1.65294126e-01 -4.64608639e-01 3.44305336e-01 -5.12251019e-01
-1.32669377e+00 7.11045742e-01 1.27404466e-01 -1.28383291e+00
-1.30764410e-01 -7.76372135e-01 -6.67244315e-01 7.61111438e-01
-1.02592850e+00 -8.72047722e-01 -3.19297075e-01 3.04671317e-01
4.10258681e-01 -4.45125192e-01 7.22114503e-01 2.86754459e-01
-8.88377845e-01 4.39194560e-01 -1.19256449e-03 4.31869328e-01
3.73760909e-01 -9.04254198e-01 4.88363028e-01 6.99702084e-01
-5.77236056e-01 6.04126632e-01 6.79898024e-01 -6.46427274e-01
-8.30441833e-01 -1.11014080e+00 3.03654194e-01 -4.29263949e-01
7.43742228e-01 -5.49560785e-02 -8.59527767e-01 6.64340258e-01
-6.46910489e-01 3.87477309e-01 4.13162977e-01 -4.58404243e-01
-9.15589854e-02 2.97127247e-01 -1.31835091e+00 3.47746968e-01
9.38430369e-01 -5.76146925e-03 -5.88465214e-01 4.98964250e-01
3.29031795e-01 -8.54721308e-01 -1.13567126e+00 7.26683617e-01
7.29625046e-01 -6.61609232e-01 1.27451515e+00 -1.01757324e+00
6.31490827e-01 2.36584619e-01 9.87610593e-02 -1.27051795e+00
-3.59630853e-01 3.47216614e-03 -3.41682017e-01 2.48737603e-01
4.94504482e-01 -9.89495575e-01 7.06198156e-01 8.95914495e-01
-4.32794005e-01 -1.18899930e+00 -1.18510389e+00 -4.35093999e-01
4.15166110e-01 -2.85421133e-01 5.47989130e-01 6.69512987e-01
-1.30958974e-01 -4.72443134e-01 1.67318106e-01 7.35428452e-01
7.58787811e-01 7.23123923e-02 2.76753247e-01 -1.43145871e+00
1.53322861e-01 -8.31536114e-01 -9.55461085e-01 -4.79222327e-01
-7.36603066e-02 -8.92342091e-01 -3.56081963e-01 -1.34589028e+00
1.48192912e-01 -6.80012584e-01 -4.10801619e-01 1.03525952e-01
8.99861977e-02 -1.42558321e-01 -2.55321741e-01 5.26227713e-01
-1.35511652e-01 2.50395298e-01 1.34355092e+00 -1.65956060e-03
-3.29015739e-02 3.06620449e-01 -3.78394872e-01 8.48812461e-01
1.23632371e+00 -4.43641692e-01 -4.08977531e-02 -3.53318185e-01
-5.07550895e-01 4.64471161e-01 6.34538949e-01 -8.28031659e-01
1.26031235e-01 3.82505432e-02 8.18408549e-01 -6.51424706e-01
3.91019493e-01 -6.46086156e-01 -3.02396387e-01 5.29806972e-01
-4.18108791e-01 2.08582878e-01 1.51610240e-01 1.33353382e-01
4.46770601e-02 -2.13332269e-02 9.04097259e-01 -3.14720571e-01
-6.28090143e-01 6.21149063e-01 -6.25418484e-01 1.15539268e-01
1.43418825e+00 -1.18920654e-02 -3.02560002e-01 -2.16044858e-01
-1.13278496e+00 4.73578423e-02 -1.04008771e-01 4.13349211e-01
1.04840696e+00 -8.73385668e-01 -8.72003436e-01 4.60852146e-01
2.17043564e-01 2.37073228e-01 4.70473707e-01 1.33914804e+00
-8.22251976e-01 6.63740456e-01 -4.45506483e-01 -6.94719911e-01
-9.33974624e-01 4.11395997e-01 8.23862493e-01 1.17852077e-01
-1.20636952e+00 9.35384154e-01 5.14602423e-01 -2.91856080e-01
3.90521377e-01 -4.16137010e-01 -4.78096724e-01 -3.15523386e-01
8.13318491e-01 2.76804090e-01 3.33373934e-01 -7.84815013e-01
-4.87947166e-01 6.27972111e-02 -5.41332185e-01 1.39361218e-01
1.27994895e+00 -7.25565627e-02 1.53738335e-02 -9.50557664e-02
1.03324866e+00 -3.33230495e-01 -8.66437733e-01 -2.01285094e-01
-9.45351124e-02 -5.73483825e-01 4.10206795e-01 -7.27871776e-01
-1.40201867e+00 7.23335385e-01 1.00067472e+00 2.04986557e-02
5.24763942e-01 3.56415838e-01 8.08172882e-01 -8.82495195e-02
4.81205881e-01 -3.39307159e-01 -4.78170335e-01 1.58996284e-01
1.09616780e+00 -1.46507812e+00 -1.29683718e-01 -4.06589568e-01
-2.88644135e-01 1.16603589e+00 7.06819892e-01 -2.53476620e-01
1.02134526e+00 3.32943857e-01 3.08029413e-01 -4.73165214e-01
-4.24701661e-01 1.27590850e-01 2.83360481e-01 3.64369482e-01
2.25564510e-01 5.36246121e-01 -1.82432264e-01 8.22784662e-01
-3.40056419e-01 -3.06278884e-01 2.80741543e-01 1.14341891e+00
-6.16179943e-01 -3.02481860e-01 -5.72070956e-01 1.14671123e+00
-5.24503529e-01 4.06352952e-02 3.72042693e-02 5.22245944e-01
-2.02833906e-01 5.32297611e-01 1.88809603e-01 -9.32564586e-02
4.17839348e-01 -2.48426527e-01 4.14599515e-02 -2.08180785e-01
-1.99741215e-01 1.73363127e-02 -1.08492307e-01 -4.50762212e-01
1.85922429e-01 -9.22584474e-01 -1.41047573e+00 8.38210359e-02
-2.58923005e-02 -6.52936995e-02 9.39601839e-01 8.53514969e-01
5.17086446e-01 4.78689611e-01 6.54123306e-01 -8.37444186e-01
-6.89369142e-01 -8.98638368e-01 -6.32260799e-01 3.35019469e-01
5.98675787e-01 -1.35663044e+00 -5.19423604e-01 -2.77745694e-01] | [14.477801322937012, -2.2161643505096436] |
245c203a-228e-48b7-b2cd-9b7a58fa9e22 | m3fas-an-accurate-and-robust-multimodal | 2301.12831 | null | https://arxiv.org/abs/2301.12831v2 | https://arxiv.org/pdf/2301.12831v2.pdf | M3FAS: An Accurate and Robust MultiModal Mobile Face Anti-Spoofing System | Face presentation attacks (FPA), also known as face spoofing, have brought increasing concerns to the public through various malicious applications, such as financial fraud and privacy leakage. Therefore, safeguarding face recognition systems against FPA is of utmost importance. Although existing learning-based face anti-spoofing (FAS) models can achieve outstanding detection performance, they lack generalization capability and suffer significant performance drops in unforeseen environments. Many methodologies seek to use auxiliary modality data (e.g., depth and infrared maps) during the presentation attack detection (PAD) to address this limitation. However, these methods can be limited since (1) they require specific sensors such as depth and infrared cameras for data capture, which are rarely available on commodity mobile devices, and (2) they cannot work properly in practical scenarios when either modality is missing or of poor quality. In this paper, we devise an accurate and robust MultiModal Mobile Face Anti-Spoofing system named M3FAS to overcome the issues above. The innovation of this work mainly lies in the following aspects: (1) To achieve robust PAD, our system combines visual and auditory modalities using three pervasively available sensors: camera, speaker, and microphone; (2) We design a novel two-branch neural network with three hierarchical feature aggregation modules to perform cross-modal feature fusion; (3). We propose a multi-head training strategy. The model outputs three predictions from the vision, acoustic, and fusion heads, enabling a more flexible PAD. Extensive experiments have demonstrated the accuracy, robustness, and flexibility of M3FAS under various challenging experimental settings. | ['Haoliang Li', 'Anderson Rocha', 'Shiqi Wang', 'Yibing Liu', 'Kexin Zheng', 'Chenqi Kong'] | 2023-01-30 | null | null | null | null | ['face-anti-spoofing'] | ['computer-vision'] | [ 3.14018816e-01 -3.19652617e-01 -9.35488120e-02 -1.76431820e-01
-6.10590339e-01 -4.19019699e-01 5.07851958e-01 -2.71789044e-01
-1.25233665e-01 3.63756388e-01 8.55088905e-02 -3.62051666e-01
-9.58209559e-02 -6.03622556e-01 -4.75967973e-01 -1.00527692e+00
1.30324125e-01 -2.59583533e-01 2.44512334e-01 -1.91241339e-01
7.84335658e-03 7.58007109e-01 -1.79901421e+00 2.16329172e-01
8.49970102e-01 1.36808217e+00 -9.45743993e-02 3.98538500e-01
2.26482246e-02 2.68767238e-01 -5.81479669e-01 -5.34960389e-01
2.01665685e-01 -2.68549502e-01 -2.78595716e-01 -7.35854357e-03
7.70374313e-02 -5.70208967e-01 -3.82327229e-01 1.01001751e+00
7.47384012e-01 -2.69606948e-01 2.96335220e-01 -1.63374007e+00
-3.70778263e-01 1.08219564e-01 -8.12277138e-01 -1.12886094e-01
5.53011239e-01 9.64160934e-02 2.39923552e-01 -9.46322262e-01
-8.90052915e-02 1.47290874e+00 6.81243181e-01 8.31459582e-01
-7.08356678e-01 -1.21619475e+00 -5.00633642e-02 2.25889027e-01
-1.44261503e+00 -1.00271845e+00 1.01776147e+00 -1.71941191e-01
4.99644250e-01 3.35866541e-01 4.43422914e-01 1.44480073e+00
4.74850237e-02 5.19211113e-01 1.17970347e+00 -3.03576142e-01
1.41408771e-01 4.33766186e-01 -2.23614737e-01 5.90359032e-01
3.74220908e-01 -1.14793773e-03 -6.37918949e-01 -4.78667229e-01
4.77608383e-01 2.62039810e-01 -3.49271148e-01 -1.22936465e-01
-6.32466733e-01 7.76961327e-01 1.99633881e-01 3.84924829e-01
-2.88608015e-01 -3.33387017e-01 3.16767067e-01 9.40507427e-02
1.83783188e-01 -1.51299894e-01 -1.32021949e-01 1.91062629e-01
-7.40849078e-01 -6.71218038e-02 6.37699664e-01 5.68495512e-01
5.53847194e-01 1.49522424e-01 2.36126408e-01 8.85414362e-01
8.37302864e-01 9.62018013e-01 5.14670074e-01 -4.85952020e-01
4.48667765e-01 3.81972641e-01 -1.24518216e-01 -1.56811762e+00
-3.90878528e-01 -3.46130738e-03 -9.11109865e-01 -1.31622748e-02
1.67881057e-01 -2.10685402e-01 -6.37615800e-01 1.72612560e+00
5.60098529e-01 3.79754186e-01 4.95694764e-02 6.92364335e-01
8.60997260e-01 5.17004013e-01 9.44573060e-02 -5.49336493e-01
1.42366874e+00 -4.29017037e-01 -8.70588481e-01 -1.27681047e-01
3.02250177e-01 -9.19045627e-01 8.12203884e-01 4.66311634e-01
-7.33684540e-01 -4.21674460e-01 -1.14358926e+00 3.05703074e-01
-4.51593190e-01 -3.10558919e-02 5.80993950e-01 1.43026149e+00
-9.68473613e-01 6.58666044e-02 -6.67688310e-01 -3.73493731e-01
5.46174943e-01 6.64328396e-01 -6.03989303e-01 -1.62972420e-01
-1.18629372e+00 4.45996463e-01 -1.84374861e-03 3.46736044e-01
-7.51773477e-01 -1.74859926e-01 -8.66947412e-01 -1.99833587e-02
2.96653777e-01 -2.54101783e-01 7.87250876e-01 -7.62302816e-01
-1.68760419e+00 6.68802500e-01 -2.82048911e-01 -6.23523677e-03
2.15202376e-01 -2.07272246e-01 -9.39521909e-01 2.87439555e-01
-1.54455796e-01 2.93895334e-01 1.24866402e+00 -1.29441440e+00
-4.60724235e-01 -7.69762516e-01 -8.99879187e-02 -2.32270509e-01
-1.11789107e+00 4.05655324e-01 -2.95500189e-01 -5.86022973e-01
1.17768191e-01 -7.83640385e-01 2.18814686e-01 8.77648685e-03
-3.67729038e-01 -5.83096631e-02 1.56396353e+00 -8.05906892e-01
1.25865114e+00 -2.30341935e+00 -3.63146722e-01 2.61159956e-01
-2.17621829e-02 9.42626774e-01 -7.96042383e-02 1.81538627e-01
5.38848415e-02 2.13395327e-01 -9.15008262e-02 -4.64825332e-01
-2.30894625e-01 4.66689877e-02 -3.30362290e-01 6.48829758e-01
1.97316766e-01 3.21580917e-01 -5.16037941e-01 -6.85898066e-01
3.82787347e-01 9.49513018e-01 -5.84891677e-01 2.61529535e-01
2.21035853e-01 4.91824061e-01 -5.68852603e-01 1.13973153e+00
1.11557674e+00 1.29493639e-01 5.54759987e-03 -1.79377094e-01
1.77326992e-01 -2.85328105e-02 -1.32237864e+00 1.05902755e+00
-3.22797179e-01 1.83593661e-01 6.00947440e-01 -8.14946055e-01
9.59420323e-01 6.15076184e-01 5.65859735e-01 -6.50737941e-01
3.53272170e-01 3.18711251e-01 -3.11263978e-01 -8.75660181e-01
9.66783613e-02 -6.24291785e-02 5.32561094e-02 2.76037008e-01
-4.11070138e-02 4.20194864e-01 -4.90039170e-01 -1.29960865e-01
9.24922109e-01 -2.62081802e-01 7.72053525e-02 1.55697331e-01
1.09320974e+00 -8.38260889e-01 7.92480111e-01 3.67626339e-01
-5.98927796e-01 3.39284480e-01 1.53400555e-01 -1.16691373e-01
-3.76455367e-01 -8.25116158e-01 -9.75644290e-02 9.64386940e-01
2.65477121e-01 -3.73895586e-01 -8.24876845e-01 -7.85089195e-01
-7.47698313e-03 1.17241830e-01 -1.18233673e-01 -2.21759170e-01
-6.63497806e-01 -6.45096838e-01 9.30620968e-01 2.39872307e-01
8.47578943e-01 -8.47763360e-01 -3.49562943e-01 -5.14035560e-02
-3.04477632e-01 -1.32667661e+00 -2.34922811e-01 -4.34862494e-01
-5.00128150e-01 -1.18668330e+00 -5.31387687e-01 -6.92833185e-01
4.49398190e-01 7.55017638e-01 1.59992203e-01 2.63816923e-01
-7.55643174e-02 3.71937782e-01 -1.73691839e-01 -5.16873062e-01
-1.76830500e-01 -1.71446532e-01 6.19758308e-01 6.22059107e-01
5.07388055e-01 -5.96082091e-01 -5.61810493e-01 5.44364333e-01
-9.14128363e-01 -5.23043334e-01 5.42441070e-01 6.65575087e-01
1.72185935e-02 1.90127581e-01 7.42292881e-01 -4.01335984e-01
6.10745490e-01 -5.39959788e-01 -4.62169856e-01 2.95951158e-01
-4.27957803e-01 -5.34892023e-01 4.36754942e-01 -5.03646135e-01
-1.07754922e+00 2.59391358e-03 -3.87190908e-01 -5.32987714e-01
-4.14312840e-01 5.04117087e-02 -9.49779153e-01 -5.43033957e-01
4.04705912e-01 1.50870681e-01 3.22132766e-01 -4.51317817e-01
-7.23369047e-02 1.40305221e+00 4.71879482e-01 -2.46724591e-01
9.82209384e-01 5.40690064e-01 3.19202244e-02 -1.26230967e+00
-1.67954892e-01 -4.52463359e-01 -2.28416950e-01 -3.37934464e-01
5.83741367e-01 -9.21302557e-01 -1.24456251e+00 9.82840240e-01
-1.11444557e+00 5.37947118e-01 6.91141963e-01 5.38079679e-01
-6.02968559e-02 8.62669408e-01 -5.98117173e-01 -1.30626929e+00
-4.47760016e-01 -1.28622723e+00 9.47133362e-01 5.84310532e-01
1.29844919e-01 -6.37505949e-01 -3.67805302e-01 7.42914081e-01
7.07413435e-01 2.92828560e-01 4.81410533e-01 -5.31195164e-01
-2.35630617e-01 -3.73459786e-01 -2.75730342e-01 4.01885122e-01
4.97369111e-01 9.07450467e-02 -1.53971374e+00 -5.24211168e-01
3.71093333e-01 -4.01279807e-01 4.97518539e-01 1.81979343e-01
1.26930213e+00 -5.69476247e-01 -5.42358279e-01 6.76500022e-01
9.48423564e-01 4.34964538e-01 5.42429149e-01 1.30107448e-01
7.16717243e-01 8.98859918e-01 3.49996716e-01 5.57049394e-01
3.02577823e-01 6.76631153e-01 5.93626618e-01 9.11272690e-02
2.14258045e-01 -2.53302962e-01 7.08981335e-01 7.36127257e-01
1.07651107e-01 -3.72937709e-01 -6.43762946e-01 2.90214866e-01
-1.52628160e+00 -9.26827371e-01 2.09938928e-01 2.29717803e+00
3.75490338e-01 -7.48329014e-02 2.26833090e-01 5.81681848e-01
9.03692245e-01 2.29686573e-01 -4.02702332e-01 -1.84269041e-01
-1.30293071e-01 -7.03079477e-02 1.94474101e-01 2.21089706e-01
-1.26554990e+00 6.33181512e-01 5.36431026e+00 8.09451044e-01
-1.40507948e+00 2.57088989e-01 4.57173884e-01 1.01474255e-01
-1.65283278e-01 -3.44435573e-01 -8.96406353e-01 7.48807609e-01
9.02298927e-01 3.60213488e-01 3.20403129e-01 8.00972700e-01
9.57155302e-02 2.26411328e-01 -6.44688845e-01 1.46053147e+00
4.61434662e-01 -8.09463501e-01 -2.98954155e-02 3.28533888e-01
1.10244058e-01 -4.18184966e-01 3.06498379e-01 1.95915215e-02
-2.11810946e-01 -9.96207476e-01 2.83281416e-01 1.28583685e-01
7.75599897e-01 -8.18568468e-01 6.24965906e-01 3.11958611e-01
-1.09265661e+00 -4.40243244e-01 -1.21876955e-01 2.44805679e-01
1.44302338e-01 4.39140052e-01 -4.40693170e-01 5.67315698e-01
8.24249268e-01 2.92103052e-01 -4.14624065e-01 8.53602529e-01
-1.18884347e-01 6.01712584e-01 -5.90957522e-01 5.98336197e-02
-1.48804754e-01 1.93339705e-01 5.23508728e-01 1.00515270e+00
5.00027120e-01 1.17743224e-01 6.86027780e-02 3.20609391e-01
-8.00949708e-02 5.66676408e-02 -7.99080968e-01 1.25403069e-02
7.26103783e-01 1.27557778e+00 -3.79346937e-01 2.56114990e-01
-5.36987960e-01 7.99467623e-01 -2.21966833e-01 2.37176612e-01
-7.51459599e-01 -3.74918610e-01 8.63642752e-01 1.56947479e-01
8.92196074e-02 -5.71552105e-02 -7.89444074e-02 -1.12666130e+00
1.82539463e-01 -1.04400313e+00 5.06714821e-01 -3.10484767e-01
-1.18948638e+00 5.16208231e-01 -2.57614017e-01 -9.83135819e-01
-1.18390784e-01 -6.79340422e-01 -4.75007117e-01 6.62702322e-01
-1.57773793e+00 -1.39743268e+00 -3.43035370e-01 1.03753901e+00
1.50527745e-01 -3.98071021e-01 8.30403388e-01 6.80261195e-01
-9.33073640e-01 1.06786144e+00 -2.53188223e-01 2.16080472e-01
7.12025404e-01 -3.32848877e-01 -6.12992607e-02 8.83010328e-01
-1.15128048e-01 7.22072244e-01 3.56043965e-01 -4.89424467e-01
-1.84528053e+00 -7.64792442e-01 8.15048218e-01 -1.28036842e-01
3.12659979e-01 -4.53825414e-01 -9.17961717e-01 2.46188074e-01
-1.41643360e-01 1.83234781e-01 8.30151439e-01 -2.11497292e-01
-5.05707026e-01 -4.65142459e-01 -1.64041030e+00 3.11006069e-01
6.89283133e-01 -8.68249595e-01 -1.00873865e-01 1.89169049e-01
3.94819468e-01 -8.30636322e-02 -6.26126885e-01 5.58385134e-01
7.95379102e-01 -1.06040907e+00 9.77900743e-01 -2.03459948e-01
-1.16397694e-01 -2.44617715e-01 -3.56808007e-01 -6.59805238e-01
4.68749292e-02 -8.04970682e-01 -4.01873052e-01 1.71568346e+00
-1.05854489e-01 -1.05376565e+00 7.03713059e-01 4.86344129e-01
2.85403579e-01 -6.23536289e-01 -1.18101156e+00 -6.94472730e-01
-3.97940993e-01 -4.33867395e-01 9.46887672e-01 9.06952858e-01
-1.43289519e-02 -1.12101257e-01 -6.89705849e-01 6.59821272e-01
7.02402115e-01 -4.39723998e-01 7.32698560e-01 -1.24109352e+00
-3.30180414e-02 -2.51608640e-01 -4.54846293e-01 -7.83425987e-01
1.04904428e-01 -2.93429047e-01 -2.23356858e-01 -8.83242428e-01
-5.80663979e-02 -4.29532647e-01 -3.63464952e-01 5.44020474e-01
8.55284706e-02 3.21646571e-01 1.15593314e-01 2.46557429e-01
-2.62429774e-01 5.88271141e-01 6.99976325e-01 -5.45131564e-02
-8.62862989e-02 1.88342348e-01 -7.61769235e-01 8.49876344e-01
7.31086612e-01 -2.47736976e-01 -3.08674425e-01 -3.36848348e-01
-2.49128729e-01 2.69455433e-01 4.03392404e-01 -1.15254033e+00
3.56145352e-01 -5.29151149e-02 4.84072864e-01 -3.05710614e-01
6.36803448e-01 -1.07146287e+00 -2.58997884e-02 5.78346968e-01
2.90630341e-01 3.62872742e-02 1.63408101e-01 4.92389709e-01
-1.72672302e-01 9.17479321e-02 8.90046835e-01 2.49587968e-01
-4.29370344e-01 3.29905093e-01 -2.68393189e-01 -5.61628282e-01
1.15534711e+00 -3.06909472e-01 -5.54229200e-01 -4.53754693e-01
-1.98300913e-01 2.31590122e-02 4.28797960e-01 6.28100097e-01
8.04709494e-01 -1.27887154e+00 -3.83366436e-01 8.11316490e-01
-3.06854788e-02 -3.49490792e-01 3.23573530e-01 7.31555879e-01
-1.48032933e-01 4.67125177e-01 -1.55620381e-01 -5.33790827e-01
-1.63318777e+00 6.32105529e-01 1.87567875e-01 2.60836989e-01
-1.53932631e-01 6.57726109e-01 -8.77447203e-02 -3.32913160e-01
4.58535939e-01 3.87013346e-01 -3.71612221e-01 1.28068775e-01
1.07480049e+00 4.32005435e-01 2.12717235e-01 -1.17087173e+00
-6.47790551e-01 6.38134241e-01 -3.91307287e-02 1.32828936e-01
1.03166854e+00 -3.23799729e-01 -1.22373648e-01 -1.77324623e-01
1.20585454e+00 1.75531656e-01 -9.14659798e-01 -2.67603904e-01
-2.10972667e-01 -7.43253946e-01 -4.69555408e-02 -2.95987040e-01
-1.11986208e+00 1.01576483e+00 9.84359384e-01 2.76686490e-01
1.33746004e+00 -2.46387407e-01 1.06785476e+00 1.45278752e-01
4.93343294e-01 -8.71050298e-01 8.90867412e-02 1.74930375e-02
5.27460933e-01 -1.27658272e+00 -2.45487288e-01 -6.32514358e-01
-2.60344982e-01 9.70968187e-01 6.29924059e-01 4.90326762e-01
8.51263046e-01 2.63789237e-01 8.88956040e-02 -1.03702145e-02
-2.77931571e-01 1.65846720e-01 2.75857486e-02 8.99486363e-01
1.20080546e-01 -1.34073988e-01 -9.78264492e-03 6.98690057e-01
8.84815399e-03 -9.76018757e-02 1.35833845e-01 1.04393268e+00
-3.54105800e-01 -1.02375340e+00 -8.62093627e-01 1.82339072e-01
-8.92133057e-01 5.10557234e-01 -2.33985439e-01 3.58366638e-01
3.15153956e-01 1.60257900e+00 -4.42578077e-01 -8.25608969e-01
1.00001447e-01 -3.00642420e-02 1.23322479e-01 -1.11903183e-01
-3.28651518e-01 1.20442227e-01 -2.67297387e-01 -6.13038063e-01
-5.67919493e-01 -5.63721061e-01 -8.03377450e-01 -6.34037197e-01
-5.54018855e-01 -2.42279433e-02 9.20268834e-01 1.02390492e+00
4.43631262e-01 -2.57891286e-02 1.16746676e+00 -8.71676922e-01
-4.98908579e-01 -8.33522201e-01 -5.29030144e-01 2.12523252e-01
6.06885374e-01 -8.19520175e-01 -4.40443754e-01 -2.39615217e-01] | [13.077737808227539, 1.1708847284317017] |
f7a7d885-938c-4d56-80c3-da6108bce0e4 | the-impact-of-subword-pooling-strategy-for | 2302.11365 | null | https://arxiv.org/abs/2302.11365v2 | https://arxiv.org/pdf/2302.11365v2.pdf | Impact of Subword Pooling Strategy on Cross-lingual Event Detection | Pre-trained multilingual language models (e.g., mBERT, XLM-RoBERTa) have significantly advanced the state-of-the-art for zero-shot cross-lingual information extraction. These language models ubiquitously rely on word segmentation techniques that break a word into smaller constituent subwords. Therefore, all word labeling tasks (e.g. named entity recognition, event detection, etc.), necessitate a pooling strategy that takes the subword representations as input and outputs a representation for the entire word. Taking the task of cross-lingual event detection as a motivating example, we show that the choice of pooling strategy can have a significant impact on the target language performance. For example, the performance varies by up to 16 absolute $f_{1}$ points depending on the pooling strategy when training in English and testing in Arabic on the ACE task. We carry out our analysis with five different pooling strategies across nine languages in diverse multi-lingual datasets. Across configurations, we find that the canonical strategy of taking just the first subword to represent the entire word is usually sub-optimal. On the other hand, we show that attention pooling is robust to language and dataset variations by being either the best or close to the optimal strategy. For reproducibility, we make our code available at https://github.com/isi-boston/ed-pooling. | ['Elizabeth Boschee', 'Scott Miller', 'Chris Jenkins', 'Steven Fincke', 'Shantanu Agarwal'] | 2023-02-22 | null | null | null | null | ['event-extraction'] | ['natural-language-processing'] | [-1.96034864e-01 -2.95162201e-01 -2.45368496e-01 -2.67831624e-01
-1.24713600e+00 -7.94234812e-01 6.62270844e-01 5.02847552e-01
-8.45762014e-01 6.04209483e-01 2.66342759e-01 -5.08149207e-01
3.16285729e-01 -6.10911667e-01 -6.88937783e-01 -5.29639304e-01
9.82323885e-02 -1.03023089e-02 1.74110815e-01 -9.94864181e-02
2.74259180e-01 4.01054293e-01 -1.33378279e+00 2.73721784e-01
7.96939433e-01 6.82681978e-01 2.67823488e-01 3.73518109e-01
-4.71141875e-01 4.05372262e-01 -6.60064816e-01 -5.13053894e-01
9.68844593e-02 -3.67753327e-01 -7.23112345e-01 -3.63237108e-03
1.94155097e-01 7.34846890e-02 -1.28959849e-01 1.02967596e+00
5.62569201e-01 1.47351012e-01 5.99690437e-01 -6.97987437e-01
-6.62442148e-01 8.77981067e-01 -5.07518589e-01 4.20543402e-01
1.41780794e-01 3.80948782e-02 1.26418376e+00 -1.41034508e+00
7.63273954e-01 1.11387575e+00 6.05012596e-01 1.89174846e-01
-1.21435547e+00 -5.43885648e-01 2.75137812e-01 2.23085046e-01
-1.65304196e+00 -6.65876091e-01 3.11416000e-01 -3.79423201e-01
1.33446598e+00 -1.47997718e-02 2.96062857e-01 1.02870154e+00
3.89767230e-01 7.14891613e-01 1.07634103e+00 -7.80130327e-01
1.35494918e-01 1.39933243e-01 5.67434728e-01 4.65514332e-01
4.37755972e-01 -3.75237554e-01 -4.98789907e-01 -1.69673990e-02
2.31910810e-01 -4.07246768e-01 -2.65320271e-01 2.32884452e-01
-1.08234525e+00 9.54681933e-01 7.28787705e-02 6.27362847e-01
-4.04810965e-01 -1.30692035e-01 5.70138693e-01 9.21361819e-02
5.04975557e-01 5.12983084e-01 -7.28685856e-01 -3.85313258e-02
-9.29558575e-01 1.32991165e-01 7.62409508e-01 7.49291658e-01
7.64150739e-01 -5.80553636e-02 -2.34165341e-01 1.13761485e+00
1.03582203e-01 1.76119298e-01 7.24478424e-01 -3.66671532e-01
5.14154375e-01 3.61221999e-01 1.91660761e-03 -5.70853829e-01
-4.31436747e-01 -2.44458064e-01 -2.73761958e-01 -2.19433099e-01
7.64173269e-01 -5.64207852e-01 -1.10248590e+00 1.79448152e+00
4.20192368e-02 -3.56169909e-01 6.96326420e-02 5.15870512e-01
6.67047262e-01 9.34485495e-01 4.34961319e-01 -2.32786208e-01
1.84750545e+00 -6.70902848e-01 -7.38061726e-01 -6.84430957e-01
7.59747684e-01 -1.03912818e+00 1.17456436e+00 3.09943587e-01
-1.01732349e+00 -3.44333738e-01 -1.15534675e+00 -3.70305806e-01
-9.09389555e-01 1.99944749e-01 2.86608934e-01 5.94083250e-01
-6.09875619e-01 3.86742294e-01 -8.40184629e-01 -6.86334968e-01
-9.62372811e-04 4.09377702e-02 -3.93463552e-01 -2.35983565e-01
-1.44334507e+00 1.12859273e+00 7.14265525e-01 -1.39150679e-01
-5.03302991e-01 -6.57957315e-01 -9.02729213e-01 1.25446007e-01
3.51941675e-01 7.93020893e-03 1.21136451e+00 -9.06126022e-01
-1.08165157e+00 1.06535625e+00 -3.20713490e-01 -4.65563208e-01
6.25034645e-02 -2.92259097e-01 -4.83800411e-01 -1.35531113e-01
2.92691737e-01 5.34085512e-01 3.56604397e-01 -7.87064373e-01
-5.20562291e-01 -3.01136136e-01 -1.07908629e-01 7.44796321e-02
-2.40120873e-01 7.32684493e-01 -5.57629704e-01 -9.22399163e-01
-1.22802570e-01 -7.02077746e-01 -1.03007099e-02 -4.84190702e-01
-3.02731425e-01 -5.57414353e-01 3.17415714e-01 -1.01708829e+00
1.37709475e+00 -2.35019183e+00 6.98436098e-03 -8.83649960e-02
-4.29690897e-01 1.22079574e-01 -2.69019715e-02 5.15407622e-01
-3.29904705e-01 3.65309566e-01 -2.45160401e-01 -3.52878153e-01
9.77734998e-02 1.28171399e-01 -1.92389324e-01 5.93071222e-01
5.52949607e-01 8.42272520e-01 -6.49522364e-01 -4.91050154e-01
-5.23713306e-02 6.19123518e-01 -3.95716935e-01 -1.18188411e-01
-5.87411299e-02 -4.38400880e-02 -7.49376640e-02 5.16111672e-01
3.23815674e-01 -1.26966462e-01 3.00932199e-01 -1.07482523e-02
-5.74486494e-01 8.61825049e-01 -1.19481349e+00 1.72618675e+00
-6.57630384e-01 7.39377081e-01 -1.56995341e-01 -7.95692265e-01
8.39029133e-01 5.94292700e-01 1.01968504e-01 -6.63480520e-01
2.61858910e-01 5.68614542e-01 3.08038980e-01 -1.62935197e-01
6.56142533e-01 -3.49531889e-01 -4.29865748e-01 4.23435152e-01
4.95165706e-01 1.63438037e-01 5.70458770e-01 -6.94012735e-03
8.35394204e-01 3.57720964e-02 7.40412295e-01 -5.27103782e-01
3.18966180e-01 1.38090655e-01 6.92184508e-01 5.55685639e-01
-2.01917514e-01 7.13486314e-01 6.14942133e-01 -2.22483389e-02
-9.22479987e-01 -8.74120593e-01 -5.33360243e-01 1.31365907e+00
-3.38845849e-01 -5.73930085e-01 -8.23048055e-01 -4.17683303e-01
-2.19594643e-01 1.24543917e+00 -3.38548034e-01 4.11087647e-02
-6.96351111e-01 -1.07081234e+00 6.74705267e-01 3.95543426e-01
2.18288973e-01 -1.21372581e+00 -5.60757816e-01 4.35252845e-01
-2.95604587e-01 -1.27170622e+00 -4.90599841e-01 6.41777933e-01
-3.68416101e-01 -7.09723771e-01 -8.42927992e-01 -7.87212729e-01
4.07339126e-01 -3.10420282e-02 1.19887853e+00 -2.87911117e-01
-2.50618309e-01 1.89162597e-01 -3.69271249e-01 -5.67550540e-01
-1.69361025e-01 4.04239595e-01 -8.02857131e-02 -7.40494207e-02
7.26230145e-01 -1.05423644e-01 -2.55931944e-01 -1.17776982e-01
-7.17938066e-01 -3.24193805e-01 3.95100921e-01 4.84103292e-01
7.17831373e-01 -1.28455341e-01 5.55750668e-01 -7.75281727e-01
6.37755096e-01 -7.07416236e-01 -4.67382610e-01 3.41168016e-01
-2.21483514e-01 1.38442069e-01 5.53018332e-01 -3.78054380e-01
-9.40857053e-01 -1.82166353e-01 -1.16464689e-01 1.99360754e-02
-2.74408400e-01 8.51968408e-01 -1.91123262e-01 4.77369130e-01
5.91800869e-01 9.61485878e-02 -5.28329194e-01 -5.71405113e-01
3.75131249e-01 6.65958703e-01 3.16557676e-01 -5.55795372e-01
1.13842241e-01 9.29005370e-02 -6.29749835e-01 -9.52503085e-01
-7.94093430e-01 -4.03736353e-01 -6.04416132e-01 1.90982986e-02
1.25177193e+00 -9.71204042e-01 -5.40523231e-02 5.06811619e-01
-1.36433423e+00 -3.02310795e-01 -1.61283702e-01 5.63156307e-01
-2.32591137e-01 2.05608550e-02 -7.19337106e-01 -6.80125535e-01
-1.86311066e-01 -1.19768107e+00 8.22842777e-01 1.69580430e-01
-4.73484337e-01 -8.54533494e-01 5.65696321e-02 -6.89082071e-02
1.52644679e-01 3.09310574e-02 9.91560221e-01 -1.06110990e+00
-2.15602145e-01 -1.26404136e-01 -7.32207596e-02 3.55247259e-01
1.39781237e-01 6.73710322e-03 -8.85197043e-01 -8.50646272e-02
-6.15523420e-02 -1.69811472e-01 9.50165212e-01 3.96243632e-01
7.59930313e-01 3.53462286e-02 -1.25693634e-01 2.32212052e-01
1.45176804e+00 2.12564841e-01 5.18846691e-01 3.77118021e-01
4.53897744e-01 6.73487186e-01 4.72490281e-01 1.20125942e-01
3.25577319e-01 4.79059368e-01 -2.47986451e-01 2.64949560e-01
-1.70138359e-01 -3.57578881e-02 8.24087918e-01 8.74942899e-01
2.16150209e-01 -4.40007240e-01 -1.20021760e+00 9.83731687e-01
-1.46818304e+00 -7.87876427e-01 9.68430340e-02 2.40405798e+00
1.01939178e+00 2.35950291e-01 -2.23373547e-02 9.30881351e-02
7.41147816e-01 2.86730587e-01 -1.85958385e-01 -7.05908298e-01
-2.87186861e-01 6.07733905e-01 7.07239330e-01 5.13627708e-01
-1.09056115e+00 1.24157608e+00 5.66824818e+00 9.00135577e-01
-1.24345052e+00 3.73454720e-01 5.43038547e-01 -1.75097689e-01
-2.44648345e-02 -1.60978690e-01 -1.21873057e+00 4.31946486e-01
1.41803217e+00 -3.96158278e-01 2.64364243e-01 4.71007019e-01
2.12892354e-01 -3.56255233e-01 -9.33226466e-01 6.92850590e-01
1.61977619e-01 -9.58398461e-01 -1.22935161e-01 2.73719560e-02
5.94240487e-01 3.96480352e-01 -2.62178779e-01 3.78799438e-01
2.60047793e-01 -9.64171171e-01 8.28864634e-01 1.38948962e-01
6.13378346e-01 -9.22205687e-01 5.24720609e-01 2.35445306e-01
-1.25124109e+00 2.69330323e-01 -1.51829422e-01 1.86791644e-01
4.21228468e-01 4.25175190e-01 -4.52904999e-01 5.28628469e-01
6.49847925e-01 4.92138684e-01 -5.48882365e-01 8.16112399e-01
-4.45895940e-01 7.90541649e-01 -3.83961886e-01 5.04601076e-02
4.53770995e-01 1.88699015e-03 5.07370353e-01 1.77035570e+00
3.71593744e-01 6.31307438e-02 1.16486587e-01 6.06232524e-01
-1.30230933e-01 6.95406497e-01 -4.66334522e-01 -4.12181228e-01
3.45693856e-01 1.06761432e+00 -9.49732006e-01 -4.23952162e-01
-8.86710286e-01 7.65089929e-01 5.47324598e-01 4.35508639e-01
-7.58488297e-01 -7.74063349e-01 7.54128754e-01 -4.16153297e-02
6.98394477e-01 -4.99520868e-01 -3.53717804e-01 -1.12102842e+00
3.17261182e-02 -8.69601905e-01 4.73210484e-01 -5.66019356e-01
-1.24937284e+00 6.53090537e-01 -4.09813747e-02 -6.93911016e-01
-3.31722170e-01 -9.28245783e-01 -4.71454293e-01 1.19199061e+00
-1.31410921e+00 -7.71524191e-01 2.89755195e-01 4.31002170e-01
7.68259287e-01 5.18521816e-02 8.45015585e-01 5.35954475e-01
-8.48761082e-01 5.47643423e-01 -1.84548602e-01 4.27432030e-01
8.39950264e-01 -1.04253745e+00 3.70437592e-01 1.27859032e+00
4.76201862e-01 7.31058717e-01 5.50211787e-01 -6.21588767e-01
-1.17486846e+00 -9.23455298e-01 1.44649673e+00 -3.10772121e-01
1.03825021e+00 -4.86104965e-01 -1.09138703e+00 9.71315205e-01
5.84015906e-01 -1.29272968e-01 7.30203807e-01 2.70844966e-01
-4.65839922e-01 1.95195511e-01 -8.61754596e-01 7.66970575e-01
6.77751422e-01 -6.79715455e-01 -7.71643937e-01 1.81168377e-01
7.05409884e-01 -1.53835982e-01 -8.93997014e-01 1.22702375e-01
4.29261386e-01 -6.72367930e-01 6.38449490e-01 -6.57984138e-01
3.23238671e-01 -1.27322823e-01 -5.30737698e-01 -1.26329303e+00
-1.92754224e-01 -2.81336784e-01 2.78354555e-01 1.43870938e+00
8.52672756e-01 -8.15652192e-01 -1.59911159e-02 4.99139071e-01
-1.34797201e-01 -5.99936843e-01 -9.30742562e-01 -7.05791354e-01
4.78578687e-01 -8.75815749e-01 2.66100854e-01 1.00539744e+00
-4.93567102e-02 3.87877494e-01 5.24110496e-02 2.67852604e-01
2.43441403e-01 -2.21249178e-01 1.55269012e-01 -7.96537519e-01
-1.66964591e-01 -5.36703944e-01 -2.90001724e-02 -7.22068548e-01
2.98210472e-01 -1.06111467e+00 1.87170148e-01 -1.42726600e+00
2.08376325e-03 -1.46543443e-01 -5.66877961e-01 7.58808792e-01
-2.99470156e-01 2.69221514e-01 2.79315263e-01 -1.06868677e-01
-1.62395418e-01 6.86315298e-02 6.75796151e-01 1.47312760e-01
-3.06032807e-01 -3.79877985e-01 -6.73807263e-01 6.65316701e-01
1.07641983e+00 -5.61419666e-01 1.12083830e-01 -5.73182583e-01
2.40643978e-01 -3.16669047e-01 1.10173203e-01 -6.89755321e-01
3.89703512e-02 -4.77694646e-02 1.79149076e-01 -4.06323999e-01
1.64557174e-01 -4.03508067e-01 -2.03484118e-01 2.90406734e-01
-1.25106022e-01 3.38651747e-01 6.59195900e-01 2.50162464e-02
-2.58248299e-01 -5.44187188e-01 8.04649293e-01 -3.47095162e-01
-8.54964316e-01 -1.20338928e-02 -6.15169644e-01 3.66933554e-01
7.31454432e-01 1.12900153e-01 -1.82397053e-01 1.49246119e-02
-6.98983490e-01 1.36347273e-02 2.45184600e-01 6.78591609e-01
-2.07142457e-02 -1.13820052e+00 -8.94029558e-01 1.40518993e-01
1.85385451e-01 -4.63729382e-01 -9.19304565e-02 8.77520978e-01
-2.60530382e-01 6.15888715e-01 6.78063631e-02 -1.90215483e-01
-1.07949471e+00 3.67812783e-01 3.44280928e-01 -2.87875086e-01
-3.40337932e-01 6.66893005e-01 4.13172990e-02 -1.88553125e-01
7.09281564e-02 -3.88927758e-01 -7.09060729e-02 6.63861811e-01
5.80154121e-01 2.11063161e-01 3.93748075e-01 -8.54861259e-01
-5.69137752e-01 4.71483409e-01 -2.45876998e-01 -4.85200256e-01
1.16190493e+00 -3.83750238e-02 -1.33293837e-01 9.79105711e-01
1.10660660e+00 2.11678088e-01 -7.15802193e-01 -1.27725020e-01
3.00454050e-01 7.77903944e-02 1.53095007e-01 -7.38123298e-01
-8.62901628e-01 8.28266680e-01 3.25577855e-01 3.18740517e-01
8.65108132e-01 1.76407501e-01 5.52167416e-01 7.75418729e-02
3.30834329e-01 -1.12774253e+00 -5.15728354e-01 7.96022236e-01
8.57445836e-01 -1.16797030e+00 -8.68829414e-02 -1.47951633e-01
-6.34434938e-01 9.67287004e-01 2.76944995e-01 -2.70492733e-01
7.87304223e-01 1.83978781e-01 1.32369578e-01 -7.62738436e-02
-6.28686190e-01 -4.27760929e-01 3.99009317e-01 -1.19679607e-01
1.00003958e+00 1.96276531e-01 -8.37143481e-01 7.06772327e-01
-2.17481747e-01 -2.88293034e-01 3.00557762e-01 9.25742030e-01
-3.96171868e-01 -1.25160956e+00 -4.89060402e-01 2.44485602e-01
-8.97184432e-01 -5.21003485e-01 -2.26396203e-01 1.09719431e+00
2.66293615e-01 8.91940355e-01 3.69002074e-01 2.00863302e-01
3.58517081e-01 9.32093799e-01 4.03512806e-01 -8.74167800e-01
-8.00026655e-01 1.75514862e-01 2.32505426e-01 -3.29688460e-01
-2.70726740e-01 -1.06783020e+00 -1.41786432e+00 -1.35797292e-01
-7.91122019e-02 -7.37657622e-02 7.00241685e-01 1.04592001e+00
2.49280542e-01 3.03198785e-01 7.16822073e-02 -6.64618909e-01
-2.34431371e-01 -1.17047572e+00 -4.26265121e-01 1.38507769e-01
8.51713195e-02 -5.40522575e-01 -3.93880963e-01 3.86262201e-02] | [10.230374336242676, 9.893712997436523] |
5b5eb4a8-22bc-443f-b064-3efcac6100e9 | topic-ontologies-for-arguments | 2301.09759 | null | https://arxiv.org/abs/2301.09759v1 | https://arxiv.org/pdf/2301.09759v1.pdf | Topic Ontologies for Arguments | Many computational argumentation tasks, like stance classification, are topic-dependent: the effectiveness of approaches to these tasks significantly depends on whether the approaches were trained on arguments from the same topics as those they are tested on. So, which are these topics that researchers train approaches on? This paper contributes the first comprehensive survey of topic coverage, assessing 45 argument corpora. For the assessment, we take the first step towards building an argument topic ontology, consulting three diverse authoritative sources: the World Economic Forum, the Wikipedia list of controversial topics, and Debatepedia. Comparing the topic sets between the authoritative sources and corpora, our analysis shows that the corpora topics-which are mostly those frequently discussed in public online fora - are covered well by the sources. However, other topics from the sources are less extensively covered by the corpora of today, revealing interesting future directions for corpus construction. | ['Martin Potthast', 'Benno Stein', 'Johannes Kiesel', 'Yamen Ajjour'] | 2023-01-23 | null | null | null | null | ['topic-coverage'] | ['natural-language-processing'] | [-1.04058616e-01 7.78309226e-01 -8.75658393e-01 -5.65451384e-02
-1.15324330e+00 -1.03902888e+00 1.02214801e+00 8.16151857e-01
-5.10823488e-01 1.08892262e+00 9.36027527e-01 -7.07244933e-01
-3.97032559e-01 -7.83076108e-01 -6.11480415e-01 -5.04302263e-01
3.27788562e-01 1.05610240e+00 6.14529133e-01 -6.98620081e-01
7.45006800e-01 -4.41604972e-01 -1.43902099e+00 6.08552814e-01
1.21502972e+00 4.78120804e-01 -1.93320751e-01 9.44166183e-02
-8.01555276e-01 6.88108683e-01 -1.03411496e+00 -9.56766307e-01
-3.15350652e-01 -3.12191874e-01 -1.21390188e+00 -3.98483366e-01
2.41055265e-01 2.37938836e-01 2.92124331e-01 9.68305945e-01
1.84001729e-01 -4.52429980e-01 8.37345302e-01 -1.03359222e+00
-1.73535086e-02 1.36212540e+00 -3.37686509e-01 4.40317154e-01
3.82683069e-01 -6.40336812e-01 1.62649930e+00 -5.09197950e-01
1.41568255e+00 1.60793889e+00 5.36124766e-01 9.51595306e-02
-7.01786458e-01 -4.39120859e-01 2.58827209e-01 4.32735942e-02
-2.10873991e-01 -8.71549323e-02 1.07388949e+00 -7.23125041e-01
4.31571782e-01 1.22069985e-01 7.74319708e-01 1.34391773e+00
9.04939622e-02 8.13022256e-01 1.24639130e+00 -6.48167491e-01
1.52207553e-01 4.37381387e-01 7.24124670e-01 -4.95217331e-02
8.77839863e-01 -6.34233952e-01 -6.08430088e-01 -8.40228617e-01
-4.55005392e-02 -8.05258036e-01 -4.10895854e-01 -4.02061194e-01
-1.17500949e+00 1.45238125e+00 -6.00417852e-02 8.56688440e-01
-3.17357093e-01 -4.60622787e-01 1.03509295e+00 4.20388222e-01
1.06781662e+00 6.65733933e-01 -8.65149736e-01 -3.00291300e-01
-7.53329217e-01 7.28479981e-01 1.47501826e+00 4.24904823e-01
3.23219508e-01 -5.50675154e-01 1.44984871e-01 9.77523267e-01
4.88199592e-01 3.86186242e-01 2.78991193e-01 -8.28282118e-01
1.29036653e+00 8.72082531e-01 3.45272899e-01 -9.05437887e-01
-2.49463230e-01 -3.18708777e-01 2.16890633e-01 -1.29680157e-01
9.07385588e-01 -4.96678561e-01 -3.60492826e-01 1.38654721e+00
6.00916862e-01 -1.06623483e+00 2.75183558e-01 6.23922408e-01
1.20695221e+00 5.67406893e-01 4.64119226e-01 -2.96311051e-01
1.68968129e+00 -6.43212855e-01 -8.63308191e-01 -1.04466371e-01
8.99320960e-01 -1.02642071e+00 9.00654376e-01 2.57914990e-01
-1.19637442e+00 1.02339514e-01 -1.21308625e+00 -1.39984071e-01
-6.89027607e-01 -1.67026713e-01 7.27663815e-01 6.28785670e-01
-4.21841055e-01 3.02043468e-01 -2.29125842e-01 -4.13769722e-01
4.72991079e-01 -4.04475778e-01 -1.68481329e-03 3.87198538e-01
-1.72943425e+00 1.25369537e+00 6.15363717e-01 -7.15084314e-01
-3.59814972e-01 -5.89928210e-01 -5.71235418e-01 -5.19262217e-02
6.11821234e-01 -8.42185244e-02 1.13973582e+00 -8.64427745e-01
-1.04498351e+00 1.34758031e+00 8.77325535e-02 -5.68004489e-01
6.05796635e-01 -4.38233882e-01 -2.28274301e-01 2.24999905e-01
6.64640009e-01 8.18229243e-02 4.14316863e-01 -1.19765544e+00
-9.01723146e-01 -1.64797068e-01 5.10172069e-01 3.02471191e-01
-3.11722398e-01 6.55988634e-01 1.97503701e-01 -6.81892216e-01
3.40517014e-02 -6.83996439e-01 1.82217643e-01 -4.77425486e-01
-4.37294453e-01 -9.82795119e-01 9.31373417e-01 -6.40895009e-01
1.01060736e+00 -1.84081829e+00 5.81235029e-02 1.99948743e-01
1.40669063e-01 -1.37816519e-01 6.41378343e-01 7.47263551e-01
-9.16518969e-04 5.32227099e-01 1.46445796e-01 5.02900839e-01
1.87676862e-01 7.70216510e-02 -8.76866758e-01 4.37914252e-01
-2.91712880e-01 6.18831277e-01 -9.31717157e-01 -8.18943739e-01
-2.80602723e-01 1.12429909e-01 -3.95650685e-01 -4.76936907e-01
-6.56425238e-01 8.11265036e-02 -1.04488575e+00 5.11744261e-01
4.20738570e-03 -2.64724165e-01 4.36859488e-01 -4.32508476e-02
-3.16639423e-01 1.50140762e+00 -7.63196707e-01 1.42209053e+00
-2.85775419e-02 1.00309741e+00 2.75443196e-01 -1.17762256e+00
7.85894215e-01 7.14542985e-01 5.79853833e-01 -3.82373989e-01
4.16590244e-01 7.13811517e-01 4.44354504e-01 -1.99758187e-01
4.88244414e-01 1.33597389e-01 -3.90777707e-01 1.06024933e+00
-1.23833805e-01 -3.35343659e-01 7.46538877e-01 4.60552394e-01
5.54435968e-01 1.78861037e-01 3.40132952e-01 -9.21842933e-01
3.00403208e-01 8.85897279e-01 5.37313819e-01 5.97220957e-01
4.42454107e-02 1.43764839e-01 1.06985283e+00 -5.55511475e-01
-1.22161877e+00 -5.87453663e-01 -8.73131037e-01 1.23585129e+00
-2.60149129e-02 -6.48595273e-01 -1.04978681e+00 -8.93375158e-01
-7.83026740e-02 6.99343503e-01 -7.90035069e-01 8.74271572e-01
-7.45577037e-01 -8.72657776e-01 3.03001136e-01 -8.15767348e-02
4.70314533e-01 -1.05894589e+00 -9.88161802e-01 4.97694522e-01
-8.44372034e-01 -8.11076939e-01 2.99759179e-01 2.97837764e-01
-7.20863879e-01 -1.67989838e+00 -5.92656612e-01 -5.87263942e-01
2.90828403e-02 -1.00594021e-01 1.40872562e+00 1.21493213e-01
1.35017306e-01 2.82816272e-02 -5.19670248e-01 -1.18232572e+00
-6.92838848e-01 5.23757458e-01 -4.99381930e-01 -8.53837371e-01
3.98363888e-01 -2.33609304e-01 -2.28363365e-01 1.39354929e-01
-4.30317521e-01 -2.76208013e-01 5.33249527e-02 8.61018240e-01
-4.53468002e-02 -1.04449101e-01 7.52131402e-01 -1.61449194e+00
9.56853867e-01 -7.95470655e-01 -2.08164915e-01 3.47386986e-01
-4.10650730e-01 -7.17540681e-02 -7.19769597e-02 -2.28318319e-01
-1.27552307e+00 -1.15589571e+00 -1.57862321e-01 8.99833500e-01
4.61211801e-03 8.50533426e-01 -5.47053590e-02 4.93122578e-01
9.66401339e-01 -7.48339117e-01 -2.05513000e-01 -3.33877355e-01
2.41702899e-01 5.04314661e-01 -3.31487879e-02 -1.35195780e+00
5.32111883e-01 4.74020660e-01 -5.47378302e-01 -7.27310359e-01
-1.40405500e+00 -3.55966181e-01 -1.62805378e-01 -1.76738709e-01
7.11919248e-01 -7.91405916e-01 -2.44783536e-01 -1.70276791e-01
-1.39842784e+00 -1.67859167e-01 -3.61992657e-01 4.45905238e-01
-4.12558198e-01 6.19855039e-02 -6.10855043e-01 -6.90225363e-01
-5.17606020e-01 -1.08295083e+00 7.89659321e-01 3.35891582e-02
-9.75475729e-01 -1.44669640e+00 3.52067202e-01 5.11034787e-01
2.02089220e-01 5.20955861e-01 1.53421855e+00 -1.18156171e+00
2.35436484e-01 9.56377760e-02 9.99541208e-03 -4.18820739e-01
-1.50999919e-01 1.62399173e-01 -7.22672403e-01 3.13461781e-03
1.29935399e-01 -5.95963240e-01 7.87983716e-01 5.32659888e-01
3.13662976e-01 -3.26852500e-01 -7.32967377e-01 -3.68720531e-01
1.05870533e+00 8.70027468e-02 4.71351594e-01 1.32727885e+00
-2.02464759e-01 1.18907893e+00 8.47956240e-01 4.30511273e-02
5.83140671e-01 4.64063287e-01 2.00037017e-01 8.51670802e-02
4.20573950e-02 -1.03195474e-01 1.06804937e-01 7.30937243e-01
-2.72064358e-01 -4.02319223e-01 -1.06717372e+00 1.05236042e+00
-1.87065041e+00 -1.18482816e+00 -5.06715715e-01 1.62474060e+00
1.01325691e+00 7.49918938e-01 1.79009855e-01 4.04414654e-01
6.70570731e-01 4.52783763e-01 -1.41893467e-02 -4.83004272e-01
-5.23297846e-01 2.78305143e-01 2.15979088e-02 5.17113388e-01
-1.30118191e+00 7.80304849e-01 6.46553946e+00 6.92898870e-01
-4.67440337e-01 3.36502343e-01 6.76981568e-01 1.32717937e-01
-6.27008200e-01 5.93792379e-01 -7.98143029e-01 4.96268719e-01
8.59597743e-01 -5.28498411e-01 -5.89632273e-01 1.10678792e+00
-7.97184929e-02 -3.66535306e-01 -5.84751785e-01 -9.12685599e-03
-6.46204203e-02 -1.57171702e+00 -2.45839506e-01 3.46298039e-01
9.15513217e-01 1.48220956e-01 -2.21010342e-01 2.69505113e-01
9.40937102e-01 -5.12349725e-01 9.79546309e-01 -2.79393673e-01
9.31339711e-02 -3.66433918e-01 9.99794185e-01 2.44115055e-01
-4.13029760e-01 1.46746650e-01 -3.09642524e-01 7.13857636e-02
4.38767403e-01 8.53228152e-01 -3.67059439e-01 5.28719902e-01
8.91809285e-01 4.55883056e-01 -1.11808442e-01 7.51872301e-01
-6.55616522e-01 1.14282262e+00 -2.43601412e-01 -3.15491170e-01
4.50341433e-01 -6.64540455e-02 8.47927809e-01 1.09723473e+00
-1.30895860e-02 1.13488436e-01 3.23601276e-01 2.20717058e-01
-7.28821456e-02 5.99518239e-01 -5.23573220e-01 -7.15203807e-02
6.27755046e-01 1.08878183e+00 -9.33839083e-01 -3.61357987e-01
-4.02650744e-01 -1.73485100e-01 2.25431442e-01 7.27005973e-02
-6.64718270e-01 -3.40123624e-01 3.26455861e-01 2.51341611e-01
-1.81659997e-01 1.28401861e-01 -4.07028377e-01 -9.09374058e-01
6.26479760e-02 -1.07967448e+00 8.48583579e-01 -3.71497333e-01
-1.30939662e+00 4.38035607e-01 5.52014887e-01 -8.46029103e-01
-3.25792819e-01 -6.32860899e-01 -8.05543959e-01 5.37321150e-01
-1.54362869e+00 -1.04615819e+00 2.37589896e-01 1.53530791e-01
6.07044637e-01 -1.01550385e-01 6.39954865e-01 -1.37034869e-02
9.73939449e-02 -3.74086462e-02 -4.21087891e-02 2.49733195e-01
8.65990043e-01 -1.20402718e+00 2.25641951e-01 1.94848686e-01
1.41795099e-01 6.29805505e-01 1.22563922e+00 -8.97271097e-01
-5.32951593e-01 3.94438952e-03 1.26664937e+00 -6.50082231e-01
1.31895530e+00 9.81082246e-02 -9.57618713e-01 5.18630087e-01
9.75924432e-01 -1.04679024e+00 8.31659317e-01 8.38904321e-01
-4.41703022e-01 6.05304897e-01 -9.44946229e-01 3.92165154e-01
6.73001707e-01 -2.05271482e-01 -1.46508515e+00 7.73508966e-01
4.59263772e-01 -6.06813967e-01 -9.31134403e-01 9.71935540e-02
4.59893405e-01 -6.86565340e-01 7.58277535e-01 -8.00689876e-01
8.80068839e-01 2.40446910e-01 -6.47584498e-02 -1.33556712e+00
2.55678833e-01 -3.34356427e-01 3.42133403e-01 1.49588954e+00
1.04413748e+00 -9.18830514e-01 8.74808371e-01 3.38973343e-01
-1.23559259e-01 -5.17922461e-01 -9.67319310e-01 -8.18437636e-02
8.70252788e-01 -2.10714534e-01 3.48814160e-01 1.31802559e+00
5.67637861e-01 5.67711771e-01 4.05517310e-01 -5.95243096e-01
7.01225221e-01 6.45122111e-01 6.83449984e-01 -2.13707376e+00
2.06493378e-01 -5.68222284e-01 1.66105643e-01 -5.61706007e-01
2.98949689e-01 -5.86786985e-01 -2.38196984e-01 -1.87379491e+00
3.04007828e-01 -6.38104022e-01 3.05424422e-01 2.28157267e-01
-2.14156434e-01 -3.77062976e-01 -2.09825858e-01 5.24679422e-01
-2.19446525e-01 1.22598179e-01 1.17969418e+00 -1.14194840e-01
1.30883406e-03 -2.38620698e-01 -1.01238644e+00 1.38016248e+00
7.98749745e-01 -6.42543972e-01 -2.35880375e-01 -2.85924405e-01
6.86303675e-01 -2.00666890e-01 -1.22499473e-01 -5.48022032e-01
-5.09512983e-03 -3.22949886e-01 -6.96774423e-02 -8.15457582e-01
1.29971609e-01 -3.63469958e-01 -2.74963647e-01 4.32571203e-01
-6.07914329e-01 9.97725651e-02 9.10587236e-02 3.60351920e-01
-4.45909500e-01 -6.25434339e-01 4.74619389e-01 -3.48331422e-01
-1.82260647e-01 -3.28187674e-01 -4.17497039e-01 1.09738207e+00
7.27055490e-01 5.15362397e-02 -1.18501699e+00 -4.63143945e-01
-5.54037094e-01 2.93690097e-02 3.77915472e-01 3.07058990e-01
-1.87944710e-01 -9.01999533e-01 -1.13176882e+00 -8.05115104e-01
-7.82911777e-02 -3.05257916e-01 -3.48272622e-01 7.24483907e-01
-3.67224306e-01 6.54774547e-01 3.66472602e-02 -1.99343264e-01
-9.98686731e-01 1.02940924e-01 -1.78176597e-01 -9.04903412e-01
-8.21347177e-01 4.69358146e-01 1.47781357e-01 -2.35761032e-01
6.96448758e-02 4.80500907e-02 -8.42196524e-01 7.25839078e-01
3.31675678e-01 1.32013604e-01 -6.82372004e-02 -6.42535388e-01
-1.74607605e-01 2.90178269e-01 -3.77360135e-01 -5.33552885e-01
1.50775981e+00 1.68082356e-01 -2.89324492e-01 6.34235322e-01
4.96008784e-01 4.55317020e-01 -4.75090295e-01 -5.65004237e-02
6.50073111e-01 -4.26886044e-02 -1.70991674e-01 -8.76510680e-01
-5.10240018e-01 6.02337301e-01 -2.90372044e-01 8.90683770e-01
2.34145895e-01 3.47313225e-01 3.69384766e-01 1.63844809e-01
4.59963620e-01 -1.56352890e+00 -1.11874796e-01 7.30931878e-01
1.05524814e+00 -8.95195723e-01 5.16121626e-01 -8.64294648e-01
-6.60602212e-01 1.05859292e+00 2.47533068e-01 -1.91662032e-02
7.85794199e-01 1.42965183e-01 2.11607233e-01 -1.06673253e+00
-7.15046167e-01 -3.00277043e-02 1.75969392e-01 2.24161804e-01
9.03008223e-01 -6.16268143e-02 -1.27648151e+00 5.11037469e-01
-6.94330335e-01 -7.67366648e-01 5.81128597e-01 9.32654202e-01
-7.23553896e-01 -1.24210513e+00 -5.15501916e-01 5.41623592e-01
-1.07854342e+00 -1.12975791e-01 -9.14185762e-01 1.37368977e+00
-1.50800288e-01 1.11052763e+00 3.77462842e-02 4.72178578e-01
4.60813865e-02 3.08190376e-01 1.86419740e-01 -6.46582067e-01
-9.69556093e-01 1.13664038e-01 1.21102881e+00 9.01306868e-02
-9.51256990e-01 -1.03821981e+00 -9.52651799e-01 -1.64898023e-01
-7.44453371e-01 1.17569864e+00 9.30499434e-01 1.24010360e+00
-1.82860374e-01 2.59134293e-01 -4.80069034e-02 -5.36001146e-01
-1.88152313e-01 -1.15845382e+00 -2.68291712e-01 3.90554696e-01
-2.51035988e-01 -1.17451084e+00 -5.81341982e-01 -1.65646821e-01] | [9.29742431640625, 9.738357543945312] |
f77ea7f8-5119-4d2d-9ab1-601a7218f5ec | svdiff-compact-parameter-space-for-diffusion | 2303.11305 | null | https://arxiv.org/abs/2303.11305v4 | https://arxiv.org/pdf/2303.11305v4.pdf | SVDiff: Compact Parameter Space for Diffusion Fine-Tuning | Diffusion models have achieved remarkable success in text-to-image generation, enabling the creation of high-quality images from text prompts or other modalities. However, existing methods for customizing these models are limited by handling multiple personalized subjects and the risk of overfitting. Moreover, their large number of parameters is inefficient for model storage. In this paper, we propose a novel approach to address these limitations in existing text-to-image diffusion models for personalization. Our method involves fine-tuning the singular values of the weight matrices, leading to a compact and efficient parameter space that reduces the risk of overfitting and language drifting. We also propose a Cut-Mix-Unmix data-augmentation technique to enhance the quality of multi-subject image generation and a simple text-based image editing framework. Our proposed SVDiff method has a significantly smaller model size compared to existing methods (approximately 2,200 times fewer parameters compared with vanilla DreamBooth), making it more practical for real-world applications. | ['Feng Yang', 'Dimitris Metaxas', 'Peyman Milanfar', 'Han Zhang', 'Yinxiao Li', 'Ligong Han'] | 2023-03-20 | null | null | null | null | ['text-based-image-editing'] | ['computer-vision'] | [ 2.44645774e-01 -1.75975502e-01 -7.01311529e-02 -3.09918731e-01
-7.04209447e-01 -4.15550530e-01 5.24782717e-01 -9.90577564e-02
-4.70576018e-01 5.07987618e-01 2.99747497e-01 -1.97497010e-01
-3.12697925e-02 -3.95149767e-01 -3.22842330e-01 -5.66532314e-01
5.17130435e-01 2.84165114e-01 1.39681175e-01 -2.78503653e-02
1.61718577e-01 7.97022656e-02 -1.25965369e+00 1.87057078e-01
1.19326437e+00 6.59777045e-01 6.11166775e-01 6.88167274e-01
-1.88717782e-01 5.14522076e-01 -4.97659504e-01 -5.53631961e-01
2.32423291e-01 -6.00497544e-01 -4.98252839e-01 2.52091646e-01
4.51160014e-01 -4.92897123e-01 -3.74483943e-01 9.28162217e-01
9.51602995e-01 2.36566693e-01 4.22156453e-01 -1.10173118e+00
-1.16511989e+00 3.97475600e-01 -7.91779339e-01 1.39208347e-01
2.22486615e-01 8.65138024e-02 5.45098782e-01 -6.81670904e-01
6.52926326e-01 1.20349526e+00 6.08218312e-01 8.69681537e-01
-1.41693187e+00 -5.98537087e-01 1.77238524e-01 1.62037015e-01
-1.32065690e+00 -7.95323253e-01 5.82215786e-01 -5.57844758e-01
6.94724143e-01 4.68189359e-01 7.26729274e-01 1.23092604e+00
-4.08091657e-02 6.50964916e-01 1.21070683e+00 -4.18525606e-01
1.20985352e-01 3.95919710e-01 -3.04170638e-01 6.18117809e-01
8.83217454e-02 -2.88946241e-01 -6.75954938e-01 -3.82409066e-01
1.02221501e+00 -9.80226230e-03 -1.77887455e-01 -3.09183031e-01
-1.46909404e+00 6.50169671e-01 -2.45216079e-02 1.80229276e-01
-3.40191722e-01 7.11617172e-02 3.31950307e-01 1.65999725e-01
6.03066027e-01 2.32623935e-01 -3.03123951e-01 -1.91865489e-01
-1.16378415e+00 2.32480109e-01 3.03660125e-01 9.76623297e-01
3.61851156e-01 6.87635168e-02 -6.02495372e-01 1.24352443e+00
3.03043842e-01 6.42813444e-01 8.64000499e-01 -1.07556295e+00
4.55890238e-01 3.77631605e-01 1.49660900e-01 -1.06257761e+00
-2.90685534e-01 -3.31211910e-02 -9.72436249e-01 -2.01808974e-01
1.35085240e-01 -1.88789979e-01 -9.64043736e-01 1.70004547e+00
4.80945170e-01 -4.46838774e-02 -1.97528690e-01 7.47905791e-01
5.73587596e-01 6.37932420e-01 1.13201961e-01 -3.73436213e-01
1.26914942e+00 -1.07751215e+00 -1.04370320e+00 -2.80747026e-01
3.74088585e-01 -9.72601891e-01 1.60103381e+00 2.26289600e-01
-1.12450171e+00 -3.76756757e-01 -7.88431287e-01 -1.72633484e-01
3.60527448e-03 2.23596424e-01 7.84333944e-01 9.26576078e-01
-1.16622066e+00 4.76779044e-01 -6.43570840e-01 -5.63542545e-01
3.86308283e-01 2.71205664e-01 -2.18711689e-01 -8.66215825e-02
-8.36364567e-01 6.30908489e-01 1.28061742e-01 -2.26709738e-01
-4.82684314e-01 -7.77484298e-01 -5.42299926e-01 -1.17144398e-01
2.27840543e-01 -7.75281072e-01 1.06431508e+00 -9.27430987e-01
-1.86283541e+00 7.00069189e-01 -3.34975660e-01 3.33625078e-03
6.45719469e-01 -1.94050863e-01 -3.95442665e-01 3.06783244e-03
1.34806540e-02 7.92187989e-01 1.29765832e+00 -1.09092963e+00
-2.48039290e-01 -3.11548889e-01 -1.67362645e-01 3.37187380e-01
-9.74544466e-01 1.45367727e-01 -1.11924708e+00 -1.08120298e+00
-1.65994000e-02 -1.00766766e+00 -3.62264037e-01 2.68602133e-01
-3.20025653e-01 1.97684661e-01 6.55007958e-01 -8.26382518e-01
1.54192042e+00 -2.26934123e+00 4.90153879e-02 -5.72490692e-03
4.43958014e-01 3.44773918e-01 -4.33798611e-01 3.01515847e-01
1.45575255e-01 2.08209604e-01 -1.17504738e-01 -5.87081790e-01
-1.40819475e-01 1.36132970e-01 -5.82429506e-02 1.31471813e-01
-3.73846620e-01 8.40355873e-01 -7.07262218e-01 -7.80031979e-01
4.47129235e-02 6.49724007e-01 -7.10368812e-01 9.15911868e-02
-6.64383322e-02 6.00137591e-01 -3.52859795e-01 2.96416104e-01
7.39142001e-01 -2.59188145e-01 2.10814282e-01 -2.24153072e-01
1.60358235e-01 -2.17809796e-01 -1.17515981e+00 1.68041480e+00
-5.15214622e-01 5.41456342e-01 1.12166293e-01 -3.57094377e-01
6.35554373e-01 3.11588883e-01 6.28151417e-01 -7.59353399e-01
-2.24465481e-03 -3.58813852e-02 -3.13548207e-01 -6.48578286e-01
6.99396908e-01 -1.40224202e-02 3.25251251e-01 6.91118300e-01
-1.76876560e-01 -1.38419554e-01 3.95660728e-01 3.91052157e-01
6.87870920e-01 9.35212299e-02 -1.45203441e-01 -3.48954163e-02
2.60183424e-01 -3.86437356e-01 6.35698378e-01 6.47651851e-01
-1.52801424e-01 7.86086261e-01 2.68697977e-01 -1.77537873e-01
-1.15782416e+00 -7.54470229e-01 4.86603342e-02 9.97145534e-01
-1.57162756e-01 -8.20300519e-01 -1.14075124e+00 -4.48688060e-01
-2.43059844e-01 7.35295415e-01 -4.32249486e-01 -6.63523003e-02
-2.66505629e-01 -1.13827348e+00 4.50946718e-01 2.64057755e-01
4.51322317e-01 -7.60534704e-01 -1.98601961e-01 1.72775045e-01
-6.27673566e-01 -1.15974224e+00 -1.19687283e+00 -6.15978479e-01
-1.16016293e+00 -6.50476635e-01 -1.13901353e+00 -5.49463332e-01
1.00876045e+00 4.77503538e-01 6.74639821e-01 3.29079293e-02
-1.28344193e-01 4.84606832e-01 -2.22718984e-01 -2.67391682e-01
-5.18519700e-01 -7.62236491e-02 3.21378946e-01 3.42878014e-01
4.84229214e-02 -3.36068690e-01 -8.33329976e-01 3.77498955e-01
-1.19325674e+00 3.74895573e-01 4.58263099e-01 7.66118228e-01
6.83472276e-01 -2.14338955e-02 4.01766211e-01 -9.32673156e-01
1.15678954e+00 -7.03866929e-02 -4.38052446e-01 4.09043700e-01
-1.11790884e+00 -4.06846255e-02 3.70107025e-01 -8.60721350e-01
-1.39109921e+00 4.13240157e-02 7.00320229e-02 -3.04212004e-01
2.80735344e-01 3.67579430e-01 1.19260214e-01 -2.59454161e-01
6.67019725e-01 1.94626674e-01 1.22397423e-01 -5.80693722e-01
5.98510206e-01 7.86167979e-01 2.21943647e-01 -4.26131129e-01
6.64594591e-01 3.09697747e-01 -3.85384530e-01 -9.25452828e-01
-4.42659348e-01 -1.83089644e-01 -5.56772053e-01 -1.73300728e-01
6.56863511e-01 -9.02364910e-01 -3.92942697e-01 7.71126270e-01
-9.12178516e-01 -5.00843346e-01 -3.03418696e-01 4.09845918e-01
-4.51768756e-01 8.20008695e-01 -7.02418208e-01 -5.83956778e-01
-5.86746991e-01 -1.14466083e+00 7.56947458e-01 2.25725099e-01
-2.83991963e-01 -8.54539037e-01 7.91042745e-02 7.97978580e-01
6.83042645e-01 -2.51575351e-01 9.35372591e-01 1.05621265e-02
-3.52329731e-01 -2.18571082e-01 -2.16343820e-01 2.99037904e-01
6.99719563e-02 -3.20755690e-02 -6.81556463e-01 -4.73256022e-01
1.53739765e-01 -1.13982782e-01 5.45180917e-01 4.82411861e-01
1.21473050e+00 -4.58893657e-01 -3.29791933e-01 6.42884076e-01
1.13776171e+00 7.44900294e-03 6.59182847e-01 2.46980965e-01
8.16506803e-01 3.10524136e-01 3.61247897e-01 7.07441986e-01
5.77305138e-01 8.15903902e-01 -3.86415929e-01 -2.94480085e-01
-3.51392299e-01 -2.59828150e-01 4.15504336e-01 1.28972447e+00
-8.98115709e-02 -2.24151656e-01 -6.36372447e-01 6.49228632e-01
-1.81239474e+00 -7.51954496e-01 -1.27852902e-01 2.42243981e+00
1.17663789e+00 -3.08715075e-01 1.48339570e-01 -3.00766319e-01
6.82179928e-01 2.23366022e-02 -5.59644043e-01 -1.80552319e-01
-1.41889855e-01 -2.55005419e-01 4.62413192e-01 4.01049495e-01
-7.01991379e-01 1.00557947e+00 7.31369209e+00 9.35166657e-01
-9.26473558e-01 5.76707542e-01 5.40432334e-01 -5.64293504e-01
-4.66300070e-01 -5.54236025e-02 -6.00686669e-01 5.33177912e-01
7.59870768e-01 -1.21236145e-01 8.25752616e-01 4.57493722e-01
5.77293456e-01 -1.77890211e-01 -6.43715203e-01 1.31909633e+00
3.50337386e-01 -1.14301670e+00 2.50785977e-01 3.42372395e-02
9.03291285e-01 -1.40028045e-01 2.45323032e-01 -1.64019600e-01
1.78514332e-01 -6.01113498e-01 7.43042648e-01 5.64783752e-01
1.09636402e+00 -3.83895308e-01 1.19934522e-01 2.24835858e-01
-7.19022691e-01 -8.55407491e-02 -5.00100076e-01 4.39394325e-01
4.28471327e-01 7.47407079e-01 -3.61912996e-01 9.12093148e-02
8.45284998e-01 3.32491845e-01 -7.36742079e-01 8.64869237e-01
1.37229353e-01 5.63231230e-01 -3.66077036e-01 1.40904054e-01
-3.01229000e-01 -3.20503354e-01 5.00109255e-01 9.76509690e-01
5.15535176e-01 3.15552652e-02 -1.86472479e-02 7.34724581e-01
-9.14509147e-02 5.00505626e-01 -4.19869661e-01 -2.96818703e-01
3.94217014e-01 1.28140557e+00 -6.76259220e-01 -3.86528581e-01
-4.49779630e-01 1.45994353e+00 2.99097478e-01 4.90328550e-01
-8.17061841e-01 -3.07727844e-01 3.96187544e-01 2.86861271e-01
1.62207454e-01 -3.94094288e-01 -2.65908659e-01 -1.44325650e+00
1.21535972e-01 -1.20165300e+00 1.96691468e-01 -9.24958467e-01
-1.21960258e+00 6.04955971e-01 -8.10585022e-02 -8.72402430e-01
6.25930429e-02 -1.97957620e-01 -2.49249294e-01 8.27361584e-01
-1.13512886e+00 -1.26700187e+00 -2.53750712e-01 8.13501894e-01
5.35781324e-01 -2.45675549e-01 7.54438460e-01 5.73684216e-01
-8.04246247e-01 9.08751190e-01 4.39616323e-01 -3.48400861e-01
1.15263808e+00 -8.20339739e-01 4.92217779e-01 8.76773179e-01
4.02871482e-02 8.27973843e-01 5.30317008e-01 -8.94889235e-01
-1.32272291e+00 -8.96426797e-01 8.57734501e-01 -4.96792376e-01
3.62504572e-01 -4.04097050e-01 -7.05412269e-01 5.46258271e-01
3.15337390e-01 -5.04902482e-01 8.88146996e-01 3.41313891e-02
-1.62975073e-01 -2.23033220e-01 -1.11587811e+00 1.14100456e+00
1.01884270e+00 -4.19686258e-01 5.54321222e-02 5.88456511e-01
5.90558350e-01 -4.13052291e-01 -9.25358474e-01 -1.41600370e-01
4.63312447e-01 -7.63716698e-01 9.28251386e-01 -3.59108269e-01
1.70932651e-01 -2.46149406e-01 1.60709575e-01 -1.23576462e+00
-6.87165141e-01 -1.19679379e+00 -2.63143599e-01 1.48708045e+00
4.20716375e-01 -5.80346525e-01 5.71218848e-01 1.11948669e+00
3.92196268e-01 -3.76455754e-01 -7.13752449e-01 -5.52564740e-01
-1.51264697e-01 -3.06726187e-01 5.90725422e-01 1.01778758e+00
-1.12336539e-01 4.68775868e-01 -8.76973331e-01 -2.99466044e-01
7.94440567e-01 -2.37887308e-01 8.05927932e-01 -9.00069654e-01
-2.96660274e-01 -3.08531374e-01 8.18671957e-02 -1.11630774e+00
-3.14665705e-01 -8.06496561e-01 -1.40378758e-01 -1.57405436e+00
4.84893799e-01 -6.17878556e-01 -4.08021808e-02 4.68107402e-01
-4.54055011e-01 3.87354910e-01 4.67169583e-01 5.39471388e-01
-4.17873383e-01 6.58535957e-01 1.63175595e+00 -1.11587502e-01
-4.52490032e-01 -2.40272641e-01 -8.71767521e-01 5.17331898e-01
8.21005404e-01 -5.53903222e-01 -6.82784855e-01 -8.69539618e-01
3.15648764e-01 -2.12088540e-01 1.02421321e-01 -7.75219023e-01
2.37431884e-01 -3.07689697e-01 4.23007220e-01 -2.29499876e-01
2.44503543e-01 -6.13905430e-01 3.64064902e-01 6.96404204e-02
-3.42467099e-01 1.39333799e-01 2.22524151e-01 6.56748474e-01
2.57813901e-01 -3.23748648e-01 7.71018684e-01 -1.22799180e-01
-3.79644722e-01 3.92459750e-01 -6.20947301e-01 -3.43114906e-03
9.84202683e-01 -2.39969313e-01 -1.84702128e-02 -6.31571949e-01
-5.57448983e-01 1.22781061e-01 6.10716403e-01 7.22876549e-01
6.25499070e-01 -1.40368128e+00 -5.78204274e-01 3.09289753e-01
3.36005050e-03 -4.01351392e-01 7.21229255e-01 9.93633807e-01
-3.90177637e-01 1.72323152e-01 -5.63385300e-02 -3.11130285e-01
-1.42358041e+00 6.05650902e-01 3.62162068e-02 -1.21165387e-01
-7.32859969e-01 6.84411466e-01 1.07913740e-01 -4.35842335e-01
1.26764312e-01 4.42155357e-03 -1.00237139e-01 1.60239451e-02
6.42581463e-01 5.26611924e-01 -1.92413181e-01 -6.21876419e-01
-7.31889578e-03 6.15053356e-01 -3.88462543e-01 -4.44465548e-01
1.24409330e+00 -6.16813660e-01 -2.67832845e-01 2.03535140e-01
6.94700360e-01 1.35827854e-01 -1.26819444e+00 -3.62881452e-01
-4.73128945e-01 -6.58341229e-01 3.23164165e-01 -8.68663549e-01
-1.17165160e+00 6.83776081e-01 8.02592397e-01 -8.64583924e-02
1.18084204e+00 -3.12277019e-01 1.17646599e+00 2.12581843e-01
2.89192200e-02 -1.63980389e+00 1.22812517e-01 1.94626123e-01
9.49180424e-01 -1.07375288e+00 2.49568000e-01 -2.29813933e-01
-1.09253442e+00 7.50035465e-01 5.46282053e-01 6.69892132e-01
5.19893229e-01 7.71423429e-02 3.93653125e-01 -4.46967408e-02
-5.37693024e-01 2.09595948e-01 3.24489981e-01 8.49796653e-01
4.30440009e-01 -4.31759842e-02 -4.48060066e-01 5.18925250e-01
1.32931128e-01 2.91677862e-01 4.64421600e-01 7.40884423e-01
-3.85404117e-02 -1.42001212e+00 -5.04388154e-01 6.33868217e-01
-4.94692922e-01 -3.29747260e-01 -2.09196731e-01 1.27352715e-01
-2.05040246e-01 1.06043208e+00 -2.24563867e-01 -2.52915174e-01
2.87906975e-01 4.06371467e-02 5.94427824e-01 -3.26974988e-01
-4.04316097e-01 3.33495170e-01 -1.50518730e-01 -4.81040865e-01
-4.15984571e-01 -7.17744291e-01 -8.00567925e-01 -5.96714377e-01
-4.33390111e-01 -2.04306543e-01 6.95537627e-01 6.52186990e-01
9.80917454e-01 3.91272902e-01 3.33712310e-01 -7.68655956e-01
-4.19943422e-01 -9.66600060e-01 -5.71660697e-01 6.24475539e-01
-4.94589806e-02 -5.45714855e-01 1.46147296e-01 5.15851736e-01] | [11.270611763000488, -0.4633926749229431] |
783db7d4-0407-40c1-b8ae-7632b5e1d121 | bridging-the-gap-between-sign-language | null | null | https://aclanthology.org/W15-5102 | https://aclanthology.org/W15-5102.pdf | Bridging the gap between sign language machine translation and sign language animation using sequence classification | null | ['Matt Huenerfauth', 'Sarah Ebling'] | 2015-09-01 | null | null | null | ws-2015-9 | ['sign-language-translation'] | ['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] | [-7.422554969787598, 3.6515443325042725] |
9221d922-a67c-48d5-94a8-db217fda655d | unsupervised-instance-discriminative-learning | 2206.13016 | null | https://arxiv.org/abs/2206.13016v1 | https://arxiv.org/pdf/2206.13016v1.pdf | Unsupervised Instance Discriminative Learning for Depression Detection from Speech Signals | Major Depressive Disorder (MDD) is a severe illness that affects millions of people, and it is critical to diagnose this disorder as early as possible. Detecting depression from voice signals can be of great help to physicians and can be done without any invasive procedure. Since relevant labelled data are scarce, we propose a modified Instance Discriminative Learning (IDL) method, an unsupervised pre-training technique, to extract augment-invariant and instance-spread-out embeddings. In terms of learning augment-invariant embeddings, various data augmentation methods for speech are investigated, and time-masking yields the best performance. To learn instance-spread-out embeddings, we explore methods for sampling instances for a training batch (distinct speaker-based and random sampling). It is found that the distinct speaker-based sampling provides better performance than the random one, and we hypothesize that this result is because relevant speaker information is preserved in the embedding. Additionally, we propose a novel sampling strategy, Pseudo Instance-based Sampling (PIS), based on clustering algorithms, to enhance spread-out characteristics of the embeddings. Experiments are conducted with DepAudioNet on DAIC-WOZ (English) and CONVERGE (Mandarin) datasets, and statistically significant improvements, with p-value 0.0015 and 0.05, respectively, are observed using PIS in the detection of MDD relative to the baseline without pre-training. | ['Abeer Alwan', 'Jonathan Flint', 'Vijay Ravi', 'Jinhan Wang'] | 2022-06-27 | null | null | null | null | ['unsupervised-pre-training'] | ['methodology'] | [-3.76724428e-03 -9.37278196e-02 -2.77376503e-01 -6.27409160e-01
-9.44568276e-01 -1.82780445e-01 6.06751144e-01 3.27839226e-01
-3.34502339e-01 4.51921046e-01 3.31808120e-01 -1.12401389e-01
-2.04105467e-01 -3.94890040e-01 8.57127756e-02 -9.03536379e-01
-3.78886491e-01 3.53869617e-01 -1.73652261e-01 -1.39284521e-01
7.81445727e-02 6.65006876e-01 -1.35145748e+00 3.78069997e-01
9.90039468e-01 6.24941587e-01 1.52584881e-01 4.15002316e-01
-1.40029684e-01 4.04153943e-01 -8.04495037e-01 -1.63006991e-01
-8.82365629e-02 -5.61011851e-01 -4.51093644e-01 3.74379426e-01
-2.05639496e-01 -1.98469117e-01 -2.23173231e-01 8.72277915e-01
9.69676912e-01 1.49186298e-01 8.82617474e-01 -1.04137635e+00
-4.73362416e-01 3.11216205e-01 -6.45147085e-01 4.46176678e-01
2.27529079e-01 -1.33463770e-01 6.74592257e-01 -1.10650241e+00
2.32264578e-01 1.32487726e+00 5.97604156e-01 8.10261726e-01
-1.28753865e+00 -6.47053480e-01 -1.36011569e-02 3.31447035e-01
-1.56189620e+00 -8.33942711e-01 1.06830454e+00 -4.28253353e-01
8.95563364e-01 1.81814000e-01 4.27045375e-01 1.31989527e+00
4.87055704e-02 7.54857302e-01 9.24035072e-01 -4.85135525e-01
4.57088143e-01 4.89744395e-01 1.54950609e-02 3.50113690e-01
-1.12808067e-02 -4.59241867e-03 -5.24755895e-01 -3.97021234e-01
4.54831809e-01 3.13459039e-01 -2.40187019e-01 -2.71376371e-01
-7.87772298e-01 1.00198901e+00 5.58739938e-02 6.05083644e-01
-7.32422769e-01 -4.41331774e-01 5.10819435e-01 1.62396222e-01
8.19160640e-01 2.21581519e-01 -4.81780469e-01 -2.92352438e-01
-1.07218373e+00 4.75949980e-03 4.93621618e-01 3.88774544e-01
2.80478269e-01 2.76521921e-01 -2.66791672e-01 1.43349838e+00
2.26066425e-01 5.85700989e-01 9.21791434e-01 -5.52081943e-01
2.71462798e-01 7.79797256e-01 -2.32797056e-01 -9.89679396e-01
-4.56876665e-01 -4.01111186e-01 -1.12699091e+00 -3.15093398e-01
-1.34256676e-01 -3.65682900e-01 -7.37364292e-01 1.75294578e+00
4.88802850e-01 3.34459156e-01 2.13732868e-01 6.05984807e-01
6.18617237e-01 5.60263455e-01 -6.55867718e-03 -6.19100153e-01
1.40525699e+00 -5.81097424e-01 -1.03668261e+00 -2.82353401e-01
6.30749464e-01 -5.23522377e-01 8.04957092e-01 5.37884057e-01
-5.60185730e-01 -4.79117215e-01 -9.01741922e-01 3.52821797e-01
-2.46434003e-01 3.56611580e-01 4.14080143e-01 9.83418584e-01
-7.77153790e-01 5.07391512e-01 -9.60684478e-01 -4.16692138e-01
4.27524686e-01 2.92314798e-01 -4.05786425e-01 -1.51328117e-01
-1.20952868e+00 6.01056993e-01 -1.21599123e-01 3.04999035e-02
-6.02999508e-01 -6.84678078e-01 -9.55414176e-01 -2.49792367e-01
-2.59322315e-01 -1.36346385e-01 8.46856654e-01 -7.66627908e-01
-1.33684695e+00 8.57528090e-01 -5.74458539e-01 -5.41710615e-01
1.64132997e-01 1.04463827e-02 -7.94882596e-01 6.56893432e-01
1.03874654e-01 4.98500824e-01 1.21971560e+00 -1.04984462e+00
-3.31240684e-01 -7.69786119e-01 -4.77032959e-01 2.35553477e-02
-7.85343289e-01 2.23380476e-01 -3.05052139e-02 -7.48315334e-01
2.28862673e-01 -7.60136545e-01 -9.32564437e-02 -2.72332847e-01
-3.09451401e-01 -3.76047671e-01 8.59781981e-01 -1.00847232e+00
1.63847876e+00 -2.63396192e+00 -1.94624871e-01 6.76733851e-02
1.84879050e-01 5.16375184e-01 -1.84895173e-01 5.30170739e-01
-4.05056357e-01 -5.97398765e-02 -4.96910721e-01 -5.71195245e-01
-4.85244244e-02 2.21237183e-01 3.25120203e-02 6.44787192e-01
6.04889095e-01 2.82827020e-01 -6.65715039e-01 -2.84761161e-01
2.15386719e-01 6.56692982e-01 -7.21237540e-01 2.91881382e-01
4.16613400e-01 4.30214107e-01 -2.44656682e-01 6.57880366e-01
8.41428876e-01 1.57199949e-01 2.26593316e-01 -6.87404126e-02
1.17438972e-01 3.50925773e-01 -1.24288511e+00 1.20101261e+00
-3.52853954e-01 4.77097154e-01 9.72435996e-02 -1.48240697e+00
1.16742098e+00 6.67042077e-01 5.21556020e-01 -5.52157044e-01
7.69962817e-02 2.90872395e-01 1.85123056e-01 -8.25819910e-01
1.51389673e-01 -3.54530364e-01 1.33064687e-01 7.29854032e-02
-7.25783035e-02 2.10682675e-01 5.46576567e-02 6.51612356e-02
1.20448792e+00 -5.82775950e-01 4.01144594e-01 -2.00999081e-01
5.08240044e-01 -4.08397943e-01 6.01904452e-01 2.01246664e-01
-5.89196324e-01 7.44684994e-01 6.18611991e-01 2.20054965e-02
-5.22122681e-01 -8.48416448e-01 -5.14554262e-01 7.06414938e-01
-3.57544452e-01 -1.48677751e-01 -7.11094379e-01 -6.68429434e-01
7.91023821e-02 8.51077080e-01 -6.18277192e-01 -5.49608350e-01
-3.33958507e-01 -1.07005346e+00 5.05266547e-01 5.35256684e-01
3.37283850e-01 -9.84583259e-01 -3.94488752e-01 2.69121468e-01
-2.32158422e-01 -9.39188838e-01 -4.81670201e-01 4.23616648e-01
-1.02241802e+00 -9.44322467e-01 -9.41238701e-01 -8.15520167e-01
7.29102254e-01 2.91705817e-01 6.66592419e-01 -3.23135376e-01
-4.15947437e-01 4.62102413e-01 -5.09747028e-01 -2.95897216e-01
-3.69371414e-01 -9.86619219e-02 4.83080149e-01 4.31095958e-01
9.16154087e-01 -6.01448357e-01 -6.58364654e-01 2.51279585e-02
-7.29783595e-01 -6.84455633e-01 6.02251172e-01 9.45017517e-01
3.28073710e-01 1.07802346e-01 1.17373562e+00 -7.99397886e-01
8.25066745e-01 -6.85878634e-01 1.50815159e-01 -2.67478883e-01
-5.53414643e-01 -1.51727885e-01 4.30296361e-01 -6.67364180e-01
-7.10098088e-01 -4.68240492e-02 -4.36908662e-01 -6.46746755e-01
-4.92140234e-01 3.60923886e-01 -2.41234913e-01 3.97038400e-01
6.55848980e-01 3.55065674e-01 3.23624402e-01 -6.34544253e-01
-3.43793482e-02 1.28526568e+00 1.10478535e-01 -1.68033645e-01
3.68659914e-01 3.08989704e-01 -5.03650129e-01 -1.55123663e+00
-4.48113531e-01 -8.88059676e-01 -4.44191486e-01 1.94046780e-01
6.69087231e-01 -8.96273017e-01 -1.73895389e-01 3.41122657e-01
-1.01921082e+00 5.46268485e-02 -2.10749507e-01 7.66160905e-01
-1.88057378e-01 3.37416321e-01 -3.12588155e-01 -1.20158327e+00
-3.34403843e-01 -9.62909877e-01 1.08179605e+00 2.72838771e-02
-6.10518813e-01 -1.02555203e+00 3.02474320e-01 1.77622527e-01
2.08207384e-01 8.82746205e-02 1.10146952e+00 -1.39106500e+00
3.61642003e-01 -4.80968356e-01 2.63127759e-02 8.47305417e-01
6.78544998e-01 -1.28146484e-01 -1.19678342e+00 -3.84233683e-01
4.41324383e-01 1.40874274e-03 5.77331126e-01 5.96934378e-01
1.05931246e+00 -2.11986601e-01 -3.32045078e-01 1.58962220e-01
1.03652894e+00 5.56048512e-01 5.67116559e-01 6.55835494e-03
3.65231872e-01 6.02236509e-01 3.86929780e-01 8.43163371e-01
2.71060944e-01 5.37386835e-01 -2.59260479e-02 -1.73067153e-01
-5.05687222e-02 1.15741357e-01 5.73571801e-01 1.10322392e+00
3.82615656e-01 1.40841857e-01 -6.72159612e-01 1.03496158e+00
-1.32344222e+00 -9.13187623e-01 -6.00797608e-02 2.31347561e+00
9.90571260e-01 -7.28871971e-02 4.72092509e-01 8.37445974e-01
8.27386677e-01 4.89137769e-02 -4.46974337e-01 -4.55946058e-01
8.72654095e-02 4.66758907e-01 -1.92750901e-01 2.12677956e-01
-1.08120012e+00 3.23910624e-01 5.44689655e+00 8.33221138e-01
-1.30809581e+00 2.44514883e-01 6.96085393e-01 -8.41100663e-02
-3.59186083e-02 -6.02424443e-01 -6.73451066e-01 6.99427009e-01
1.24044597e+00 3.83098312e-02 9.20503214e-02 7.07157791e-01
7.51767397e-01 1.10239461e-02 -1.06876314e+00 1.25603426e+00
2.53281474e-01 -6.76396191e-01 -2.13366598e-01 4.10403870e-02
4.46764380e-01 -3.51360589e-01 1.28526717e-01 4.78567749e-01
-4.46177274e-01 -7.05676854e-01 2.17402503e-01 2.28996158e-01
5.33050060e-01 -1.09747493e+00 9.51621413e-01 3.30235004e-01
-1.05468285e+00 -1.03805371e-01 -2.88887978e-01 2.69142717e-01
-1.81503862e-01 1.01979661e+00 -9.25652385e-01 3.88964951e-01
5.24126291e-01 7.26885021e-01 -3.92951727e-01 9.59101498e-01
-7.10219815e-02 9.28742886e-01 -1.23890869e-01 -9.26891640e-02
-4.34268191e-02 -1.36228964e-01 5.48100233e-01 1.32969379e+00
4.31576252e-01 8.14884529e-02 -4.78585735e-02 4.97887075e-01
4.87448648e-02 2.13891059e-01 -6.93414748e-01 -2.74381399e-01
6.22868299e-01 1.15308166e+00 -4.52568889e-01 -2.23337844e-01
-2.33251423e-01 1.16971409e+00 -7.40373284e-02 2.84326226e-01
-6.82767689e-01 -6.79404318e-01 6.77987039e-01 1.44477919e-01
3.85469675e-01 -2.81425323e-02 -7.29551837e-02 -7.90124655e-01
1.39255626e-02 -1.21333253e+00 2.64114261e-01 -1.14376940e-01
-1.38370323e+00 5.70984006e-01 -1.17753744e-01 -1.23553371e+00
-3.92863423e-01 -2.45063588e-01 -5.81320524e-01 7.67008126e-01
-1.40324807e+00 -5.91914654e-01 6.89225271e-02 5.71288526e-01
8.10241818e-01 -2.81125247e-01 1.12098551e+00 6.47746086e-01
-9.74530280e-01 8.59465539e-01 2.52631992e-01 1.22016713e-01
7.40858018e-01 -1.11447680e+00 -1.99143246e-01 5.32004058e-01
1.73030838e-01 8.21986139e-01 5.28726101e-01 -3.55691791e-01
-1.22609782e+00 -1.25982225e+00 1.17554879e+00 -2.59324819e-01
3.71546328e-01 -2.95898318e-01 -1.06894791e+00 6.03969730e-02
1.32260010e-01 -2.03966379e-01 1.03409433e+00 5.94127886e-02
-1.25959963e-01 -3.67563605e-01 -1.49912632e+00 2.96134800e-01
4.65797842e-01 -5.70172071e-01 -8.94134343e-01 2.83638299e-01
4.44638848e-01 2.05360144e-01 -7.70803869e-01 3.29578459e-01
2.02941567e-01 -8.25464189e-01 9.05348599e-01 -3.24817508e-01
2.45509490e-01 3.03351358e-02 -3.18993598e-01 -1.53507996e+00
-7.70191848e-02 -4.58837628e-01 -2.72302777e-01 1.53375173e+00
3.19361210e-01 -7.38532662e-01 5.44359088e-01 1.48058414e-01
-9.58353281e-02 -8.38673949e-01 -1.03591454e+00 -8.86450827e-01
-1.16002552e-01 -4.04244632e-01 3.47105920e-01 1.14380980e+00
2.07093269e-01 4.03511941e-01 -1.26934305e-01 3.37770522e-01
3.28183621e-01 -2.58199215e-01 3.75891387e-01 -1.26586866e+00
-6.19668998e-02 -1.59644648e-01 -4.13066238e-01 -6.60760939e-01
8.12120214e-02 -7.20794499e-01 -1.84394438e-02 -1.50627267e+00
4.61910293e-02 -2.00738996e-01 -4.80532855e-01 5.08952022e-01
-2.71531641e-01 4.56672609e-02 -1.97733968e-01 -1.41567709e-02
-2.41472304e-01 7.40403891e-01 6.91120744e-01 -2.39563152e-01
-5.43671072e-01 2.93762803e-01 -7.14338481e-01 6.39051378e-01
8.29707384e-01 -7.25616992e-01 -5.78108490e-01 1.04115158e-01
-5.65385401e-01 3.13158706e-03 1.17746048e-01 -9.40324366e-01
-1.39830753e-01 2.14332059e-01 2.71114945e-01 -6.30765021e-01
5.10017753e-01 -5.62607944e-01 -2.86223292e-01 6.43091142e-01
-2.68980235e-01 -7.50473142e-02 4.34702516e-01 5.17570376e-01
-3.86090249e-01 -3.89598280e-01 9.54854786e-01 2.65758514e-01
-3.49110931e-01 7.20519572e-02 -5.94742656e-01 1.47222206e-01
7.36779511e-01 -4.97639552e-03 3.37042928e-01 -3.04404885e-01
-7.47651875e-01 -6.24792241e-02 -2.31724158e-01 2.98166841e-01
7.48276234e-01 -1.50820637e+00 -7.74058402e-01 4.68932509e-01
1.46305576e-01 -4.41262454e-01 4.64742780e-01 1.20281756e+00
-3.64584029e-02 2.98235059e-01 3.77439320e-01 -6.84424281e-01
-1.34567010e+00 3.91586363e-01 8.78252536e-02 -6.08140752e-02
-5.14730752e-01 7.90933788e-01 -5.39392941e-02 -5.62201917e-01
4.24998999e-01 -3.43478590e-01 -4.60570961e-01 5.01654208e-01
8.12854290e-01 4.81232136e-01 4.53783035e-01 -3.95906001e-01
-7.96272695e-01 2.82599181e-01 -1.93985507e-01 -8.61583576e-02
1.45627117e+00 3.41261248e-03 -4.44420017e-02 5.75745344e-01
1.34557140e+00 1.57661662e-01 -6.48286760e-01 -3.82722914e-02
5.44903651e-02 -2.23355740e-01 9.48841944e-02 -5.25741279e-01
-9.98590410e-01 1.17705095e+00 1.24386740e+00 5.52398264e-01
1.19569397e+00 -1.54778346e-01 6.23086274e-01 1.86531827e-01
1.72192249e-02 -1.07730174e+00 1.74871281e-01 1.49419919e-01
8.39228630e-01 -1.16279054e+00 -3.40863168e-01 -1.33076180e-02
-7.29987919e-01 8.44867110e-01 2.94718921e-01 -1.04580447e-01
8.46872807e-01 5.88118173e-02 5.35261184e-02 -4.64451648e-02
-6.91189647e-01 -1.67319879e-01 1.75053477e-01 8.81576836e-01
5.41292191e-01 1.56406581e-01 -3.24345529e-01 1.04250538e+00
9.78477672e-02 -2.60868788e-01 1.70083180e-01 7.06302941e-01
-4.65131581e-01 -1.17724657e+00 -4.71783698e-01 6.30246818e-01
-4.40049708e-01 -9.58838537e-02 -3.43052328e-01 6.68078423e-01
1.68147534e-01 1.32890737e+00 9.42829475e-02 -3.26207697e-01
5.01610160e-01 4.81849194e-01 1.90238699e-01 -8.36154282e-01
-3.51298124e-01 4.41748410e-01 4.06097136e-02 -1.53477818e-01
-4.14400190e-01 -7.81033397e-01 -1.30957627e+00 -5.17266691e-02
-2.80040383e-01 3.07023585e-01 6.05505586e-01 9.86159265e-01
6.09898388e-01 7.81544924e-01 1.26967847e+00 -5.54317534e-01
-7.38310575e-01 -1.21600270e+00 -8.66070628e-01 3.52035522e-01
5.02880692e-01 -6.82648718e-01 -7.68128037e-01 -4.14175577e-02] | [14.219378471374512, 6.095572471618652] |
3119dd66-ef66-4b80-a81e-931e570ef0a0 | exploiting-rigidity-constraints-for-lidar | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Dong_Exploiting_Rigidity_Constraints_for_LiDAR_Scene_Flow_Estimation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Dong_Exploiting_Rigidity_Constraints_for_LiDAR_Scene_Flow_Estimation_CVPR_2022_paper.pdf | Exploiting Rigidity Constraints for LiDAR Scene Flow Estimation | Previous LiDAR scene flow estimation methods, especially recurrent neural networks, usually suffer from structure distortion in challenging cases, such as sparse reflection and motion occlusions. In this paper, we propose a novel optimization method based on a recurrent neural network to predict LiDAR scene flow in a weakly supervised manner. Specifically, our neural recurrent network exploits direct rigidity constraints to preserve the geometric structure of the warped source scene during an iterative alignment procedure. An error awarded optimization strategy is proposed to update the LiDAR scene flow by minimizing the point measurement error instead of reconstructing the cost volume multiple times. Trained on two autonomous driving datasets, our network outperforms recent state-of-the-art networks on lidarKITTI by a large margin. The code and models will be available at https://github. com/gtdong-ustc/LiDARSceneFlow. | ['Zhiwei Xiong', 'Xiaoyan Sun', 'HanLin Li', 'Yueyi Zhang', 'Guanting Dong'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['scene-flow-estimation'] | ['computer-vision'] | [ 1.63240999e-01 -1.63114473e-01 -3.52519870e-01 -5.32024622e-01
-6.12291753e-01 -3.57505172e-01 4.57359403e-01 -5.17727733e-01
-4.31838065e-01 7.10261822e-01 1.92164063e-01 -4.53817487e-01
2.85070017e-02 -8.28755200e-01 -9.22547996e-01 -4.63206172e-01
3.71111035e-01 6.16613150e-01 -3.28892693e-02 -2.40153968e-02
3.62187624e-01 7.54586875e-01 -1.51093030e+00 -3.68206680e-01
9.19508755e-01 6.00110888e-01 2.85138637e-01 6.65442884e-01
-2.01087639e-01 1.01230729e+00 -9.98862609e-02 -2.45249942e-01
5.93596518e-01 -1.30760685e-01 -6.55706704e-01 2.71065421e-02
1.14188337e+00 -6.32407546e-01 -9.43168998e-01 8.63909900e-01
4.44557011e-01 3.87622118e-01 3.99222612e-01 -1.09869337e+00
-2.82170355e-01 4.27223951e-01 -5.68953753e-01 1.00391030e-01
-2.18695506e-01 2.91529059e-01 1.06297159e+00 -1.26556325e+00
7.49608397e-01 1.10900497e+00 6.72765493e-01 3.74017060e-01
-1.43980491e+00 -9.25374866e-01 1.89840227e-01 4.47151721e-01
-1.43107235e+00 -6.90089822e-01 1.20025456e+00 -4.54709709e-01
9.44480360e-01 -1.36281326e-01 5.95035315e-01 9.65605259e-01
4.93191294e-02 6.52693450e-01 4.91121262e-01 9.25781429e-02
-1.17190935e-01 -2.34824225e-01 2.87405923e-02 9.68947470e-01
2.53288537e-01 4.26370919e-01 -7.01242507e-01 1.79736048e-01
7.90526986e-01 1.17948162e-03 -2.17426568e-01 -6.98109746e-01
-1.12293625e+00 1.10012531e+00 7.46093154e-01 -1.84709370e-01
-1.55897990e-01 5.99188685e-01 3.10837090e-01 -7.72596076e-02
5.84554255e-01 2.12782130e-01 -1.62914857e-01 -3.19160998e-01
-1.00445461e+00 1.86676696e-01 6.61572218e-01 9.18233991e-01
1.17857230e+00 4.80079114e-01 1.78936109e-01 6.42224908e-01
3.85494500e-01 6.94554985e-01 1.63821772e-01 -1.46327019e+00
7.57998466e-01 3.04878980e-01 -1.07343145e-01 -1.12044477e+00
-3.50438595e-01 -4.29799497e-01 -9.80226934e-01 2.08981708e-01
2.70398527e-01 -1.76333755e-01 -7.71368444e-01 1.58240664e+00
4.07262355e-01 7.33945906e-01 -9.23683345e-02 9.44974482e-01
8.17676306e-01 7.80967295e-01 -4.86523658e-01 -5.16964048e-02
6.22112215e-01 -1.17501402e+00 -6.52022362e-01 -4.32717621e-01
5.52034795e-01 -8.15716565e-01 8.87839556e-01 8.35372582e-02
-1.12717867e+00 -6.98569834e-01 -1.04876852e+00 -5.39950430e-01
7.13003576e-02 2.77471900e-01 5.76487362e-01 2.89199352e-01
-9.03413296e-01 7.36159384e-01 -1.09966862e+00 -3.09676342e-02
6.01233482e-01 2.83662826e-01 -3.00861076e-02 -7.51895607e-02
-6.08884215e-01 7.96296835e-01 5.54258153e-02 6.71734691e-01
-9.28750515e-01 -9.48769331e-01 -1.25375128e+00 -2.04356059e-01
4.09699023e-01 -8.21688831e-01 1.26504397e+00 -2.85895109e-01
-1.75550807e+00 5.28936386e-01 -6.62245810e-01 -6.23995841e-01
7.93080151e-01 -6.23761654e-01 1.91524491e-01 -1.17211476e-01
1.06309578e-01 1.04504836e+00 7.52964258e-01 -1.18895137e+00
-4.41706449e-01 -1.10157199e-01 -2.50964135e-01 1.26196653e-01
1.34221315e-01 -6.96870148e-01 -2.92358488e-01 -3.95592898e-01
2.48288304e-01 -1.13200927e+00 -4.77494538e-01 2.19095901e-01
-4.64961082e-01 6.80958852e-02 1.22745931e+00 -2.74547696e-01
8.10943782e-01 -1.87971997e+00 2.10075364e-01 -1.93730486e-03
2.93769389e-01 1.69163987e-01 -2.65437692e-01 -7.17787966e-02
8.32805783e-03 -1.25283763e-01 -6.65759802e-01 -8.37804854e-01
-1.95461363e-01 4.40778524e-01 -7.65398979e-01 7.62958288e-01
4.27055389e-01 1.11788380e+00 -7.69677103e-01 -4.07160282e-01
6.76558971e-01 7.55683362e-01 -7.18084037e-01 2.33059824e-01
-2.71300912e-01 7.23067343e-01 -2.21969023e-01 4.76598650e-01
7.75766253e-01 -1.80608869e-01 -1.44710317e-01 -2.53335953e-01
-4.97863859e-01 5.60478508e-01 -1.07558298e+00 2.15824962e+00
-7.23413765e-01 1.22284198e+00 -2.67160714e-01 -9.19116199e-01
1.04754388e+00 -1.99224234e-01 7.64814079e-01 -6.17210031e-01
9.57796425e-02 1.00183956e-01 -1.69559762e-01 -2.90450960e-01
8.58363569e-01 1.52441844e-01 2.73238152e-01 3.49981338e-01
-2.04759836e-01 -4.98686105e-01 8.47133920e-02 1.59030169e-01
7.86921501e-01 6.16044283e-01 -1.22665182e-01 1.48637995e-01
4.88315672e-01 1.08615451e-01 8.64291787e-01 6.22288704e-01
-1.49683118e-01 8.86415899e-01 1.02995716e-01 -6.22095406e-01
-1.22238934e+00 -9.95227218e-01 -3.38060826e-01 3.87761772e-01
2.25059450e-01 -3.25839072e-01 -1.82967186e-01 -4.34467584e-01
1.84244633e-01 7.21033216e-01 -2.24625677e-01 -2.67826039e-02
-1.23953283e+00 -4.52712297e-01 4.85823423e-01 5.37670374e-01
6.77902460e-01 -7.72712171e-01 -7.70530701e-01 2.87396878e-01
-4.15767789e-01 -1.48077142e+00 -6.88004494e-01 -8.92248936e-03
-1.03422844e+00 -7.91878700e-01 -1.47745952e-01 -4.44476902e-01
4.81384546e-01 5.06593823e-01 9.87155855e-01 7.35716745e-02
-4.72230822e-01 3.96527676e-03 2.74631947e-01 -1.86055511e-01
4.54529263e-02 3.70933771e-01 -3.00118532e-02 -7.91007057e-02
2.20464841e-01 -8.95800471e-01 -5.74200034e-01 3.71955991e-01
-5.03480792e-01 2.65677840e-01 2.93608189e-01 7.87082434e-01
7.78894126e-01 -4.18457121e-01 -1.11750558e-01 -7.14073360e-01
5.25735244e-02 -2.74386436e-01 -1.11275804e+00 -3.54298621e-01
-5.05477250e-01 3.37343931e-01 4.36460197e-01 -2.76809037e-01
-1.03308213e+00 5.16113281e-01 -9.12457183e-02 -9.48342323e-01
-8.89080111e-03 3.51103395e-01 1.31538674e-01 -3.26507866e-01
4.63467121e-01 1.29770562e-01 1.18997628e-02 -3.23695064e-01
5.19467235e-01 9.52879414e-02 7.69109964e-01 -4.69913572e-01
1.26079154e+00 8.06361377e-01 4.96793866e-01 -8.71022224e-01
-1.00224400e+00 -5.48185050e-01 -9.77997541e-01 -4.23716486e-01
7.12745368e-01 -1.06720328e+00 -8.52311254e-01 4.82323319e-01
-1.44005370e+00 -6.44584417e-01 -4.72428799e-01 7.95536637e-01
-6.38741672e-01 3.42621595e-01 -4.53282654e-01 -6.21424198e-01
-1.86058715e-01 -1.24799192e+00 1.11507332e+00 1.15372203e-01
5.34697026e-02 -8.24027300e-01 3.36905092e-01 4.38241512e-01
3.39012504e-01 3.36143434e-01 2.57860392e-01 2.05948770e-01
-1.31599712e+00 2.24188000e-01 -3.76546651e-01 1.66586563e-01
1.30448446e-01 3.80289555e-01 -9.36046660e-01 -8.53757933e-02
-3.19017544e-02 -2.64904231e-01 1.20779502e+00 6.14968479e-01
1.10067248e+00 -2.29886159e-01 -8.73452649e-02 1.35919404e+00
1.43183720e+00 -1.99711069e-01 6.21128082e-01 1.71324641e-01
1.31693673e+00 5.77232361e-01 4.96781915e-01 3.36810619e-01
6.95813596e-01 7.20702589e-01 6.87329829e-01 1.01485632e-01
-3.14037025e-01 -4.70802218e-01 3.96737039e-01 1.01557994e+00
-2.35849749e-02 -6.73315674e-02 -1.00919378e+00 4.95025724e-01
-2.07704806e+00 -9.34874058e-01 -2.95148283e-01 2.07824564e+00
4.59468126e-01 2.01692268e-01 -2.16700703e-01 -2.44987205e-01
4.30136472e-01 6.48678184e-01 -9.26153064e-01 -2.12095022e-01
-1.25537321e-01 1.56303182e-01 7.89893031e-01 1.11753583e+00
-8.15684140e-01 1.37033558e+00 5.53438950e+00 3.00865561e-01
-1.34641635e+00 9.52840596e-02 3.21632832e-01 -4.64069724e-01
-3.54051799e-01 2.67635256e-01 -1.15676033e+00 1.95295617e-01
8.67997646e-01 -7.34637752e-02 3.35649937e-01 6.23427570e-01
5.49378395e-01 2.69642491e-02 -7.18886256e-01 9.79717493e-01
4.32176962e-02 -1.73257887e+00 5.69967218e-02 2.32901111e-01
6.43810034e-01 6.34018898e-01 5.85831814e-02 8.92633423e-02
4.62592334e-01 -9.14490461e-01 6.54438794e-01 6.41943634e-01
6.97008073e-01 -6.46164477e-01 3.43370974e-01 2.30828092e-01
-1.52614582e+00 1.65144339e-01 -4.38914686e-01 -2.04608649e-01
7.01048613e-01 7.13901877e-01 -7.34290242e-01 4.78760660e-01
5.29780209e-01 1.48651886e+00 -3.71851951e-01 8.86655748e-01
-3.62946838e-01 5.95130861e-01 -5.61841369e-01 2.89665401e-01
4.00458932e-01 -7.04114437e-01 8.61476123e-01 9.68671739e-01
3.27806741e-01 -8.47361237e-02 2.70522505e-01 1.27898765e+00
-2.52291650e-01 -2.24007204e-01 -9.77117836e-01 2.67405987e-01
3.20621192e-01 1.26774704e+00 -3.72851491e-01 -2.74284780e-02
-3.26756805e-01 6.22438192e-01 4.92553979e-01 5.08453131e-01
-9.72708702e-01 -1.52183175e-01 9.49747920e-01 -3.50003578e-02
3.40586513e-01 -9.30125594e-01 -5.18579781e-01 -1.35750937e+00
2.25872368e-01 -1.24175914e-01 5.66162132e-02 -8.11025918e-01
-8.37478876e-01 3.83388072e-01 -8.41445550e-02 -1.29115069e+00
-3.74701887e-01 -4.23166811e-01 -7.17957258e-01 6.54569924e-01
-2.02033734e+00 -1.16172898e+00 -6.99619174e-01 5.21004140e-01
7.75297999e-01 -5.22222519e-02 1.67352527e-01 2.88765579e-01
-8.81538332e-01 1.71599239e-01 -2.56396115e-01 1.69676408e-01
6.03110671e-01 -9.55586970e-01 7.41607010e-01 1.10018861e+00
3.03140938e-01 3.50916207e-01 6.25762343e-01 -6.30560637e-01
-1.39283979e+00 -1.43919802e+00 9.43322122e-01 -3.75086844e-01
7.77456939e-01 -4.33528006e-01 -9.50429857e-01 9.61413741e-01
4.96626012e-02 2.88131803e-01 6.73767477e-02 -6.14537857e-02
-3.73224884e-01 -3.24097931e-01 -7.10613012e-01 6.32943332e-01
1.47122180e+00 -5.80394208e-01 -1.63636863e-01 3.24142635e-01
9.08128083e-01 -8.67231607e-01 -5.37629962e-01 6.01143837e-01
5.59755087e-01 -7.59624422e-01 1.07148135e+00 -3.34417015e-01
6.55423701e-01 -4.72341061e-01 -1.41938701e-01 -9.85904694e-01
-7.50216916e-02 -9.31388497e-01 -4.42575246e-01 7.74354279e-01
3.87792617e-01 -7.43849039e-01 1.11283994e+00 2.17698336e-01
-3.00749362e-01 -6.08835936e-01 -1.13173580e+00 -6.42744482e-01
-3.26268072e-03 -6.61000311e-01 3.01746547e-01 7.50775635e-01
-8.70172918e-01 6.08869135e-01 -5.96163034e-01 3.13877940e-01
8.77712429e-01 2.26507381e-01 1.11479139e+00 -1.11236155e+00
3.65129001e-02 -4.66784805e-01 -3.87846649e-01 -1.43214226e+00
7.67605007e-01 -8.30706239e-01 2.46847868e-01 -1.19988561e+00
-2.44420603e-01 -3.46697867e-01 2.54111975e-01 4.36855853e-01
7.63363168e-02 3.20420980e-01 3.28641236e-01 4.15167928e-01
-2.31601626e-01 1.02933490e+00 1.22586143e+00 -1.63835108e-01
-4.55388486e-01 1.36872325e-02 -2.14063451e-01 6.52982891e-01
8.93354893e-01 -5.61281443e-01 -4.09205198e-01 -9.27829325e-01
3.16931337e-01 -1.43672138e-01 5.88670671e-01 -1.09769082e+00
5.94276726e-01 -1.81379393e-01 8.13010335e-02 -1.04995191e+00
7.00897813e-01 -7.11325347e-01 1.25823557e-01 4.89164054e-01
-2.82171875e-01 3.64919305e-01 3.47416103e-01 5.65053642e-01
-1.62780166e-01 -8.80797878e-02 7.82422245e-01 1.25461206e-01
-5.32923639e-01 7.91514218e-01 -7.51010925e-02 3.82391959e-02
5.47452331e-01 -1.99764922e-01 -4.16943222e-01 -4.56999004e-01
-2.26087570e-01 3.17173988e-01 2.94384748e-01 5.28293014e-01
7.90574551e-01 -1.37731433e+00 -9.25774157e-01 2.03048468e-01
-1.19892523e-01 7.55611181e-01 1.57057092e-01 9.11752641e-01
-6.80366337e-01 3.97418886e-01 -1.98237300e-02 -1.04533780e+00
-1.12678373e+00 -7.36535117e-02 5.97063363e-01 -1.08472660e-01
-1.03840363e+00 7.41723239e-01 -2.86822841e-02 -6.88661098e-01
1.09168500e-01 -2.85245419e-01 9.12362263e-02 -3.63429524e-02
1.37745261e-01 4.55308199e-01 -3.85576785e-02 -9.23796892e-01
-1.46942094e-01 1.11374438e+00 2.17266250e-02 -1.21146113e-01
1.38368106e+00 -2.08383784e-01 -6.49287403e-02 5.54631174e-01
1.45547986e+00 -4.03761566e-02 -1.69910896e+00 -5.64911306e-01
-2.50481248e-01 -6.40127063e-01 4.15503919e-01 4.48224164e-04
-1.60080838e+00 1.01287067e+00 4.51758236e-01 -5.00294447e-01
7.32963741e-01 -3.56054008e-01 9.71541822e-01 7.49040544e-01
-2.91920435e-02 -8.27319682e-01 1.66249409e-01 9.58015084e-01
8.59149516e-01 -1.42096210e+00 1.91188917e-01 -3.62098515e-01
-4.29890603e-01 1.23095822e+00 6.96271479e-01 -4.26369518e-01
6.89922869e-01 3.20081264e-01 9.25860181e-02 -1.55463815e-01
-8.06608081e-01 -2.48253226e-01 2.25425035e-01 2.08943591e-01
2.10246414e-01 -1.73319042e-01 2.19019875e-01 -5.91109455e-01
-6.06870413e-01 -2.89439205e-02 6.73205853e-01 7.71277487e-01
-2.37900183e-01 -8.90788674e-01 -2.14900915e-02 2.45707139e-01
1.33120298e-01 -1.36209726e-01 -1.61565393e-01 7.00368047e-01
-1.78458259e-01 6.39068365e-01 5.24501622e-01 -3.39062184e-01
4.39654589e-01 -2.96849817e-01 3.65883440e-01 -3.93241078e-01
-2.07876071e-01 -6.07555956e-02 -2.22235499e-03 -1.00163507e+00
-6.44401729e-01 -9.59200740e-01 -1.20928085e+00 -6.17957413e-01
-2.12097615e-01 -3.08205932e-01 7.56530583e-01 8.67197633e-01
4.13501203e-01 3.87167722e-01 8.00437212e-01 -1.18673885e+00
-1.65491298e-01 -6.90954924e-01 1.00264661e-02 2.99217533e-02
8.48394871e-01 -6.80643737e-01 -5.28500557e-01 -1.13479793e-01] | [8.508454322814941, -2.0702407360076904] |
971fe461-7100-471f-ae75-e6131a6053f0 | relaxing-instrument-exclusion-with-common | 2301.02052 | null | https://arxiv.org/abs/2301.02052v2 | https://arxiv.org/pdf/2301.02052v2.pdf | Relaxing Instrument Exogeneity with Common Confounders | Instruments can be used to identify causal effects in the presence of unobserved confounding, under the famous relevance and exogeneity (unconfoundedness and exclusion) assumptions. As exogeneity is difficult to justify and to some degree untestable, it often invites criticism in applications. Hoping to alleviate this problem, we propose a novel identification approach, which relaxes traditional IV exogeneity to exogeneity conditional on some unobserved common confounders. We assume there exist some relevant proxies for the unobserved common confounders. Unlike typical proxies, our proxies can have a direct effect on the endogenous regressor and the outcome. We provide point identification results with a linearly separable outcome model in the disturbance, and alternatively with strict monotonicity in the first stage. General doubly robust and Neyman orthogonal moments are derived consecutively to enable the straightforward root-n estimation of low-dimensional parameters despite the high-dimensionality of nuisances, themselves non-uniquely defined by Fredholm integral equations. Using this novel method with NLS97 data, we separate ability bias from general selection bias in the economic returns to education problem. | ['Christian Tien'] | 2023-01-05 | null | null | null | null | ['selection-bias'] | ['natural-language-processing'] | [-2.21875057e-01 5.29779214e-03 -9.75648761e-01 -2.21397921e-01
-6.65745497e-01 -4.66221154e-01 2.78099388e-01 -2.19465673e-01
-4.02702332e-01 1.36866164e+00 5.89725971e-01 -5.67152798e-01
-8.50137591e-01 -6.52244925e-01 -7.82332122e-01 -5.63735366e-01
-7.26441070e-02 4.57435697e-01 -6.84351265e-01 3.54149222e-01
9.95422602e-02 4.38642532e-01 -1.37252879e+00 -6.30986929e-01
1.49774146e+00 4.31418180e-01 -1.70311630e-01 3.16801220e-01
1.87595353e-01 8.20582986e-01 -2.03356177e-01 -4.93365049e-01
1.99556097e-01 -2.78362274e-01 -5.07638335e-01 -5.90610020e-02
2.90855497e-01 -2.84303635e-01 -6.58556670e-02 1.13793361e+00
6.13611817e-01 2.71729641e-02 1.25399470e+00 -1.28940845e+00
-1.08579588e+00 5.79674244e-01 -6.19031191e-01 7.91661665e-02
3.80424708e-01 -1.98754087e-01 9.45268929e-01 -9.83977675e-01
2.78612196e-01 1.37441003e+00 6.92352891e-01 2.59896249e-01
-1.38322711e+00 -8.89467537e-01 -5.04714176e-02 -1.68274164e-01
-1.27294946e+00 -4.35225517e-01 4.18415636e-01 -7.38058031e-01
9.74615812e-02 4.49414730e-01 2.56967247e-01 1.28122783e+00
2.97158390e-01 4.80113924e-01 1.39539957e+00 -5.30846417e-01
9.10238475e-02 2.20614731e-01 3.20588768e-01 5.24840713e-01
9.42950845e-01 6.11959577e-01 -2.52495855e-01 -5.55460155e-01
1.19942355e+00 1.23217009e-01 -2.78264940e-01 -4.18137640e-01
-1.29067719e+00 1.12920582e+00 -1.60351530e-01 4.78220731e-03
-8.89773369e-01 1.69180989e-01 9.80150849e-02 5.24760008e-01
3.39263976e-01 3.73882264e-01 -6.67725980e-01 2.43538141e-01
-6.27649903e-01 2.99507916e-01 6.99629068e-01 7.51703978e-01
5.08716226e-01 1.40536591e-01 -4.14991528e-01 3.77495050e-01
8.63686725e-02 1.03436220e+00 5.44769764e-01 -9.71137881e-01
3.11558217e-01 3.35249454e-01 6.45792782e-01 -7.88297415e-01
-5.36788225e-01 -6.70914233e-01 -1.09868097e+00 -5.24978302e-02
7.42791474e-01 -6.17318988e-01 -6.06214166e-01 2.18346238e+00
4.33021843e-01 4.02581513e-01 -9.48144197e-02 8.40906262e-01
4.10885960e-01 1.67793408e-01 2.51801968e-01 -6.98613763e-01
1.36369860e+00 -2.81095862e-01 -9.80651081e-01 3.42861861e-01
6.66153014e-01 -4.82321441e-01 1.02227247e+00 9.94519293e-02
-1.32756114e+00 -2.94003010e-01 -2.24733174e-01 -9.51929111e-03
-3.38981561e-02 1.29397169e-01 9.46998000e-01 9.80482340e-01
-7.20137835e-01 3.75292629e-01 -2.99313486e-01 1.75805107e-01
-3.84869762e-02 6.84708238e-01 -3.14464003e-01 1.09698176e-01
-1.32005477e+00 6.26997173e-01 -2.55539089e-01 -2.56847978e-01
-7.08375990e-01 -1.33378017e+00 -7.24707842e-01 4.71992135e-01
5.01666009e-01 -1.11322188e+00 9.17076528e-01 -1.08586907e+00
-1.32284939e+00 5.29694736e-01 -3.61373663e-01 -7.86912367e-02
8.75922561e-01 -1.71424821e-01 -5.48975468e-01 -1.83039308e-01
7.72075713e-01 -3.11876982e-01 6.96878493e-01 -9.22059774e-01
-4.81661230e-01 -5.70009768e-01 -2.88623452e-01 6.94464380e-03
-7.38423988e-02 2.40232512e-01 2.64253527e-01 -9.25468028e-01
3.35899815e-02 -7.82819033e-01 -1.87348872e-01 -6.70638084e-01
-4.26821202e-01 -1.76280886e-01 -6.65462017e-02 -6.30924940e-01
1.10272813e+00 -1.95593822e+00 -6.85493201e-02 3.27576727e-01
3.98505591e-02 -3.26517075e-01 -9.05197114e-02 1.12259537e-01
-6.31307840e-01 2.32889354e-01 -8.06172043e-02 -8.12037569e-03
3.90002131e-01 -7.18596950e-02 -3.76572013e-01 1.10364807e+00
-1.62903033e-02 9.37978148e-01 -7.06629932e-01 -3.61033291e-01
-6.39647469e-02 1.97752967e-01 -6.64907575e-01 4.56622383e-03
5.33762336e-01 5.93536794e-01 -6.37885988e-01 5.81844926e-01
8.75175655e-01 -2.33596623e-01 3.73074226e-02 4.34027791e-01
-5.17762244e-01 3.96993041e-01 -1.74176383e+00 9.86095786e-01
-3.67388427e-01 1.17328361e-01 7.44327456e-02 -1.42133594e+00
4.20176476e-01 8.46235096e-01 4.85260099e-01 -5.75725257e-01
8.75917971e-02 4.58740860e-01 -1.04967341e-01 -8.17545235e-01
1.87364042e-01 -7.25630879e-01 -2.19650447e-01 1.81679338e-01
9.37701836e-02 5.66040814e-01 -1.96553677e-01 -1.77786678e-01
7.45813966e-01 -3.00910890e-01 6.22208893e-01 -9.14029837e-01
3.61205220e-01 -2.25311846e-01 8.79123092e-01 9.68361914e-01
1.76976115e-01 3.39959711e-01 8.07450056e-01 1.54783890e-01
-8.03752840e-01 -1.07513869e+00 -5.76557457e-01 8.94597054e-01
-2.65020072e-01 8.25376630e-01 -2.28906184e-01 -4.66737300e-01
7.26226687e-01 7.13802695e-01 -9.75373685e-01 9.78285149e-02
-1.10794902e-01 -9.55217898e-01 5.19570947e-01 4.97272640e-01
2.63202637e-01 -5.30003548e-01 -6.09338060e-02 9.16785300e-02
1.15412727e-01 -3.53624523e-01 -3.81563634e-01 5.33819571e-02
-8.87219250e-01 -9.52625096e-01 -1.25707531e+00 -5.01261353e-01
5.59354126e-01 1.47063866e-01 1.13702321e+00 -2.89468933e-02
3.06645006e-01 4.87063318e-01 2.82589108e-01 -7.46994138e-01
6.48709312e-02 -3.17121029e-01 5.23140252e-01 7.78153539e-02
5.08503020e-01 -5.16639650e-01 -6.74325407e-01 9.58257914e-02
-5.05356848e-01 -6.14066362e-01 5.93531668e-01 1.11961603e+00
2.70027727e-01 -6.34635566e-03 1.31630695e+00 -9.86134946e-01
3.94513249e-01 -9.95550632e-01 -9.10525501e-01 1.79242492e-01
-7.82765329e-01 5.94064631e-02 3.42795223e-01 -9.01467323e-01
-1.40886211e+00 -4.45622265e-01 4.35867101e-01 -1.26147136e-01
-1.94529921e-01 7.59892941e-01 -4.65348125e-01 1.46072030e-01
4.36448663e-01 -4.40609813e-01 -2.67671227e-01 -7.01230466e-01
6.97785169e-02 4.92545873e-01 4.54763830e-01 -8.83928478e-01
7.18845963e-01 6.09774351e-01 4.90795583e-01 -4.54366237e-01
-6.11852288e-01 -5.86468220e-01 -3.67732018e-01 5.23790479e-01
7.08521545e-01 -1.29984272e+00 -1.11765933e+00 -1.24653026e-01
-7.97215283e-01 -1.12170674e-01 -6.05651081e-01 1.49954247e+00
-5.84580719e-01 -1.60220768e-02 -4.41203862e-01 -1.30526233e+00
2.00872660e-01 -8.03674579e-01 6.37906730e-01 4.59442735e-02
-2.23209888e-01 -1.33594775e+00 1.18771151e-01 9.43632573e-02
-9.21497792e-02 2.20903218e-01 1.01878941e+00 -4.81289476e-01
-3.52816075e-01 -2.70298392e-01 -3.02132100e-01 -1.13259643e-01
-3.76034826e-02 -1.08475506e-01 -6.52049720e-01 -1.31896555e-01
3.53604764e-01 1.15874462e-01 6.19739890e-01 1.27346778e+00
9.32513416e-01 -5.82629561e-01 -2.80897468e-01 6.34643853e-01
1.50070715e+00 -2.34930944e-02 4.73927796e-01 1.96187586e-01
5.09044170e-01 9.55469906e-01 2.80288249e-01 6.47804677e-01
5.29959619e-01 9.83406827e-02 -1.88328084e-02 -5.36556423e-01
5.47595859e-01 -3.62668574e-01 1.36769205e-01 5.07323623e-01
-3.09718460e-01 -7.58685917e-02 -5.42399704e-01 9.82118070e-01
-1.65889919e+00 -1.10778236e+00 -8.30459177e-01 2.57546544e+00
7.44173110e-01 -5.14869332e-01 3.59943867e-01 -7.55452439e-02
7.68080294e-01 -5.76272309e-01 -3.17543596e-01 -2.50035405e-01
-4.18939412e-01 2.13324890e-01 1.18898344e+00 5.79289734e-01
-5.07637799e-01 1.90067530e-01 6.96129465e+00 3.96354884e-01
-7.61522889e-01 2.86519289e-01 4.70849514e-01 1.54911488e-01
-7.86495030e-01 3.81549388e-01 -6.09345794e-01 5.10102034e-01
8.33140910e-01 -5.86775362e-01 1.82677768e-02 3.58400047e-01
7.18153954e-01 -1.79074511e-01 -1.00418210e+00 4.44294244e-01
-5.23113549e-01 -7.02672184e-01 -3.70398641e-01 6.11119509e-01
1.26209188e+00 -4.67357934e-01 5.79229534e-01 2.07146779e-01
6.03343546e-01 -1.20627117e+00 7.07831740e-01 7.46922016e-01
1.13496494e+00 -9.57437396e-01 6.50362253e-01 4.79250848e-01
-5.78515828e-01 -3.23761880e-01 -6.43924296e-01 -7.01663256e-01
-1.63820028e-01 9.59471881e-01 -1.85279950e-01 5.89815021e-01
1.06121749e-01 1.60221964e-01 1.31872490e-01 1.03018975e+00
-2.96986639e-01 8.23634565e-01 -6.32357150e-02 3.02925527e-01
-8.03205296e-02 -3.59697163e-01 4.41373110e-01 8.95259500e-01
7.94093311e-01 5.27419090e-01 -3.98885339e-01 1.04350555e+00
-2.93743908e-01 3.82583022e-01 -6.99247003e-01 3.66096705e-01
3.86820495e-01 6.31018817e-01 -1.85822949e-01 -5.53706527e-01
-8.93492818e-01 5.23275554e-01 -5.80428280e-02 9.00277734e-01
-8.03056359e-01 1.58738956e-01 6.36088848e-01 2.56575704e-01
-6.37719482e-02 8.21522772e-02 -6.01289868e-01 -1.63554502e+00
-9.71594974e-02 -9.62628961e-01 5.90595365e-01 -2.00420380e-01
-1.27870846e+00 -7.12690651e-01 2.21359909e-01 -7.16430962e-01
-2.01508090e-01 -3.66938710e-01 -6.09212160e-01 1.41562545e+00
-1.68823564e+00 -6.98042095e-01 3.66342992e-01 7.73347318e-01
-1.39927929e-02 9.27070677e-02 7.69302726e-01 4.13117588e-01
-6.85477614e-01 5.32210410e-01 6.30341411e-01 -2.58004904e-01
7.71049201e-01 -1.41547072e+00 -1.75448880e-01 6.98275805e-01
-3.67727965e-01 9.60636258e-01 7.32630908e-01 -1.03295779e+00
-1.32532990e+00 -8.17641079e-01 1.31326139e+00 -3.75544727e-01
8.21407378e-01 -9.20590758e-02 -8.90200377e-01 1.03213561e+00
-2.49846667e-01 -1.13320611e-01 7.56260514e-01 6.15895092e-01
-3.17420103e-02 2.51654893e-01 -1.17418182e+00 5.30160606e-01
9.49633598e-01 -3.11290145e-01 -6.46415293e-01 2.92559743e-01
5.53928733e-01 1.19090356e-01 -9.37291384e-01 5.01718223e-01
6.74974620e-01 -7.94874012e-01 1.05445683e+00 -1.03879738e+00
1.99668273e-01 1.20549612e-01 3.49921957e-02 -1.07465339e+00
-7.47960091e-01 -7.35708177e-01 2.47136548e-01 1.48615360e+00
3.40368718e-01 -1.01687050e+00 5.18405974e-01 9.85265434e-01
4.01679218e-01 -3.48035842e-02 -1.02888525e+00 -1.20220006e+00
6.12170279e-01 -1.88780278e-01 1.05141878e+00 1.68066573e+00
7.08544906e-03 2.68812031e-01 -6.15399122e-01 4.70254451e-01
9.19704854e-01 9.94381458e-02 7.46571600e-01 -1.79048133e+00
-2.60836452e-01 -1.87457651e-01 1.51841074e-01 -7.36294210e-01
7.49939084e-01 -5.60815454e-01 -4.81810272e-01 -9.64406312e-01
4.58126903e-01 -5.54547608e-01 -2.60139316e-01 -1.48465723e-01
-4.76471990e-01 -3.02161187e-01 -3.88798773e-01 1.02516554e-01
1.04811743e-01 4.43931639e-01 1.17470860e+00 1.85094237e-01
-4.72549587e-01 3.20351243e-01 -9.12533581e-01 9.25457776e-01
5.81850767e-01 -7.06058145e-01 -4.08214152e-01 -1.84505984e-01
3.56652230e-01 7.90197492e-01 7.75424361e-01 -3.83606434e-01
-8.74207243e-02 -7.54169881e-01 5.44539392e-01 -1.79535836e-01
-2.26606503e-01 -8.83474112e-01 4.58228976e-01 3.49764049e-01
-3.19237977e-01 1.74066693e-01 -1.13450654e-01 4.23725337e-01
7.66526684e-02 -4.35689330e-01 3.86391938e-01 -7.34508336e-02
1.00933127e-01 2.26099968e-01 -2.66013592e-01 1.09492823e-01
6.40063643e-01 1.35812908e-01 -8.58737528e-02 -5.05356431e-01
-6.60193264e-01 2.10127160e-01 4.72811878e-01 -2.79766142e-01
1.92678481e-01 -1.50058341e+00 -1.05791068e+00 2.09346458e-01
-2.86124587e-01 -5.64956844e-01 4.15835947e-01 1.18386996e+00
3.51806700e-01 6.90721452e-01 1.40109211e-01 1.28236869e-02
-7.48275220e-01 1.21608281e+00 1.47037819e-01 -7.60141164e-02
-3.32906932e-01 4.15138930e-01 9.32386160e-01 -8.47365558e-02
-1.41235972e-02 -2.57790476e-01 -1.86396644e-01 1.51484996e-01
3.94161165e-01 1.04571962e+00 -3.01479548e-01 -5.91662884e-01
-1.29522994e-01 4.21719432e-01 6.54661655e-01 -3.07970226e-01
1.01974475e+00 -7.01012373e-01 -1.95054084e-01 5.98868072e-01
1.17707253e+00 6.48795784e-01 -9.80078757e-01 -1.59517467e-01
2.65027732e-01 -4.52397645e-01 -1.52945697e-01 -4.41886127e-01
-6.10144675e-01 5.84407806e-01 4.52480316e-01 3.58122170e-01
9.14555073e-01 -2.34023094e-01 8.64229575e-02 -1.10565007e-01
2.27591246e-02 -8.79096985e-01 -6.18039489e-01 1.19553186e-01
6.80218518e-01 -1.33142197e+00 -1.70333870e-02 -3.77174169e-01
-2.75683314e-01 4.43380356e-01 -5.78723475e-02 -3.53188336e-01
6.15938783e-01 2.15584673e-02 -1.31746054e-01 -8.09764192e-02
-5.71539402e-01 -2.43263692e-01 5.78930616e-01 6.71388805e-01
7.18373775e-01 5.44264615e-01 -1.08981955e+00 7.84384668e-01
-1.07819386e-01 -1.21191628e-02 6.97227120e-01 2.29726180e-01
-8.42187740e-03 -8.13564479e-01 -1.03000903e+00 5.22867918e-01
-1.29658282e+00 -2.49350503e-01 1.28735721e-01 1.20612407e+00
1.28242001e-01 7.77047813e-01 1.00947551e-01 6.59178734e-01
5.13327301e-01 3.30990478e-02 9.56386551e-02 -4.26538020e-01
-1.54538721e-01 3.56295079e-01 -2.26894170e-01 -9.19077024e-02
-6.68298662e-01 -9.90903080e-01 -9.04370964e-01 -6.04916930e-01
-7.59137392e-01 4.72171992e-01 2.93275923e-01 9.51166809e-01
3.77790742e-02 2.83753783e-01 8.17337155e-01 -2.25118011e-01
-1.04239333e+00 -8.04698586e-01 -1.21697390e+00 3.42720270e-01
7.48137712e-01 -9.71318185e-01 -6.59482837e-01 -2.63537556e-01] | [7.9434814453125, 5.104445934295654] |
ce90d043-e117-4c29-9f98-31652182dc55 | the-effects-of-political-martyrdom-on | 2305.18004 | null | https://arxiv.org/abs/2305.18004v1 | https://arxiv.org/pdf/2305.18004v1.pdf | The Effects of Political Martyrdom on Election Results: The Assassination of Abe | In developed nations assassinations are rare and thus the impact of such acts on the electoral and political landscape is understudied. In this paper, we focus on Twitter data to examine the effects of Japan's former Primer Minister Abe's assassination on the Japanese House of Councillors elections in 2022. We utilize sentiment analysis and emotion detection together with topic modeling on over 2 million tweets and compare them against tweets during previous election cycles. Our findings indicate that Twitter sentiments were negatively impacted by the event in the short term and that social media attention span has shortened. We also discuss how "necropolitics" affected the outcome of the elections in favor of the deceased's party meaning that there seems to have been an effect of Abe's death on the election outcome though the findings warrant further investigation for conclusive results. | ['Miu Nicole Takagi'] | 2023-05-29 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [-2.92668074e-01 2.50408232e-01 -3.89298707e-01 -2.78522938e-01
-4.02280003e-01 -4.07566696e-01 1.21566403e+00 4.81636435e-01
-8.60340118e-01 8.94389987e-01 1.24162757e+00 -9.48615372e-01
2.99775749e-01 -7.74075568e-01 -4.13539588e-01 -6.57984734e-01
3.07849050e-01 -3.79575305e-02 -5.87011993e-01 -8.27932477e-01
6.85138643e-01 3.60561162e-02 -1.15422261e+00 -2.55380750e-01
4.16860133e-01 1.78884611e-01 -2.07703307e-01 2.29298264e-01
2.69926161e-01 7.54510939e-01 -7.44906604e-01 -4.37080830e-01
3.80312145e-01 -1.07765220e-01 -5.09891629e-01 -1.91558570e-01
5.51069975e-01 -1.29057929e-01 -5.16648114e-01 9.92811620e-01
4.79619503e-01 5.95699251e-02 3.01757038e-01 -5.07457972e-01
-5.35608940e-02 8.16710353e-01 -7.41187572e-01 5.98910213e-01
2.65754879e-01 1.57934412e-01 9.68748450e-01 -7.74266541e-01
1.09614789e+00 1.23956192e+00 6.42460585e-01 7.48988539e-02
-1.20270324e+00 -1.03822410e+00 3.53902966e-01 -1.60916615e-02
-1.11748564e+00 -7.63198793e-01 5.75241148e-01 -7.33202398e-01
5.62127769e-01 4.87387508e-01 9.94758844e-01 9.39380407e-01
6.27463222e-01 -3.97752486e-02 1.45544672e+00 -1.94703698e-01
6.21339194e-02 2.22593471e-01 2.99116939e-01 -1.02997698e-01
6.55311346e-01 -1.58683397e-02 -2.81097651e-01 -5.72633147e-01
-1.20373107e-01 -2.97319293e-01 -9.53881629e-03 4.98443693e-01
-1.03788352e+00 1.26053035e+00 4.93113965e-01 6.17098272e-01
-7.85948992e-01 -9.62433114e-04 7.23493934e-01 3.65451097e-01
1.10269380e+00 4.44115162e-01 -3.19047958e-01 -2.74904102e-01
-8.99666429e-01 3.05101156e-01 6.17199779e-01 -1.54264018e-01
5.44112921e-01 -4.05105591e-01 1.20149396e-01 2.86291122e-01
8.17891434e-02 5.65780282e-01 4.18243073e-02 -5.89899778e-01
5.61625481e-01 3.54471922e-01 2.44740561e-01 -1.77028739e+00
-5.32113433e-01 -5.89365363e-01 -5.82336009e-01 7.33952364e-03
5.38106382e-01 -5.62073648e-01 -4.97284800e-01 1.82035208e+00
3.62925947e-01 -3.34807336e-01 -2.29541883e-01 8.50172102e-01
6.04577243e-01 7.37124622e-01 6.37024105e-01 -6.06029510e-01
1.42409110e+00 1.09380692e-01 -7.41914213e-01 -4.60276842e-01
2.95254767e-01 -9.73091006e-01 5.63719034e-01 -9.11287963e-02
-6.43578708e-01 -2.44634777e-01 -8.26187909e-01 1.42790332e-01
-3.56714606e-01 -2.83849150e-01 8.59320998e-01 7.80505538e-01
-5.14860988e-01 3.74310970e-01 -6.44044042e-01 -7.64429331e-01
1.94023609e-01 5.66442274e-02 -1.18952900e-01 1.21654533e-01
-1.41258001e+00 1.16293108e+00 -1.70169890e-01 -6.52679205e-02
-1.59313083e-01 -1.47004217e-01 -9.04497504e-01 -2.50345588e-01
3.28392759e-02 -4.53903884e-01 7.11198568e-01 -8.24065566e-01
-8.84115279e-01 1.05601144e+00 -3.62046622e-02 -4.15659428e-01
5.22413611e-01 1.56324178e-01 -3.92791539e-01 -3.23347270e-01
8.15731883e-01 1.28650397e-01 2.58844942e-01 -5.26645958e-01
-5.85250616e-01 -8.81041288e-01 2.13431597e-01 3.39392096e-01
-3.40452403e-01 6.08438611e-01 2.85758913e-01 -5.90979040e-01
2.28362083e-01 -1.10185838e+00 -3.55434000e-01 -1.05737162e+00
-4.86292630e-01 -2.64278471e-01 5.93778372e-01 -7.28845537e-01
1.42909706e+00 -2.08890486e+00 -6.44969523e-01 3.67261544e-02
-1.84648022e-01 -2.87415653e-01 4.41960484e-01 8.06024194e-01
-5.65268397e-02 4.62985396e-01 3.36709201e-01 -9.82176214e-02
-2.36122683e-01 4.60514165e-02 -4.43801016e-01 1.26222467e+00
-4.29128230e-01 3.68819147e-01 -7.55915642e-01 -6.18263558e-02
5.76280318e-02 1.82785124e-01 -3.95280689e-01 -7.91183233e-01
3.47060293e-01 5.68486691e-01 -4.19466287e-01 5.75280488e-01
7.29644477e-01 1.97890684e-01 5.68088055e-01 -1.27218187e-01
-9.82831657e-01 9.06412959e-01 -2.59811223e-01 9.31393266e-01
-2.24215999e-01 1.24647319e+00 4.59039599e-01 -5.39323986e-01
7.04252183e-01 3.56240332e-01 1.43739685e-01 -8.78102005e-01
6.40799224e-01 2.07349241e-01 2.57366061e-01 -3.81327778e-01
1.08823073e+00 -8.76716852e-01 -7.26703286e-01 5.23598969e-01
-1.03743255e+00 -2.43722841e-01 1.14314400e-01 2.79375553e-01
7.89256752e-01 -4.98962194e-01 3.93556774e-01 -6.97540879e-01
1.17508069e-01 2.79970735e-01 8.03862154e-01 5.95976591e-01
-2.31944531e-01 7.67135695e-02 3.84296924e-01 -6.44247711e-01
-1.00404811e+00 -1.37451276e-01 -4.07559127e-01 1.04136312e+00
9.10544619e-02 -3.16392571e-01 -2.56510675e-01 -2.23827660e-01
-1.01345845e-01 1.26527870e+00 -7.16412544e-01 2.62668729e-01
-5.19886255e-01 -1.32559502e+00 4.02281396e-02 -3.30517888e-01
4.29948568e-01 -5.64373672e-01 -6.48277998e-01 1.47764802e-01
-5.38286269e-01 -9.54425037e-01 -1.67932495e-01 -1.92901134e-01
-5.09796858e-01 -1.03532875e+00 -4.05839622e-01 -1.46347210e-01
6.14796817e-01 2.34887436e-01 8.08528066e-01 -2.56110102e-01
6.36310354e-02 7.70744160e-02 3.15636918e-02 -8.10043275e-01
-3.26757699e-01 2.25383177e-01 3.82838219e-01 -1.96764916e-01
3.66322190e-01 -6.73895717e-01 -6.92160010e-01 7.98677281e-03
-3.24812859e-01 -3.42958085e-02 -1.33944554e-02 6.51513636e-02
-2.81484067e-01 -2.36432813e-02 7.10483372e-01 -9.97509003e-01
6.67837620e-01 -9.66763735e-01 -4.58137631e-01 -7.15723872e-01
-5.68836391e-01 -8.00269425e-01 4.47988622e-02 2.14226730e-03
-1.19060266e+00 -5.42845726e-01 2.26313174e-01 6.58309937e-01
1.74236730e-01 9.86696780e-01 4.62394178e-01 3.95197719e-01
5.82597554e-01 -4.61807340e-01 -9.51334611e-02 -2.91968614e-01
-1.72916204e-01 7.71016777e-01 4.52457100e-01 -3.04101795e-01
6.37591004e-01 8.56056154e-01 -3.55532914e-01 -9.53589380e-01
-8.48516166e-01 -4.87923682e-01 1.31705359e-01 -5.79912424e-01
9.66352105e-01 -1.52957392e+00 -8.69360328e-01 2.34493077e-01
-1.17345750e+00 1.03113309e-01 1.37786597e-01 8.77931237e-01
2.14682207e-01 -7.35924300e-03 -6.47937477e-01 -9.39219296e-01
-1.69437423e-01 -6.19834960e-01 4.11344022e-02 4.01452899e-01
-6.64547563e-01 -7.10000098e-01 4.37728912e-01 8.30091536e-01
4.90579396e-01 7.55656362e-01 5.57723284e-01 -5.80521822e-01
-7.26165846e-02 -4.59847659e-01 4.58252840e-02 -2.44570151e-02
1.87561736e-01 6.83830306e-02 -5.93774855e-01 -3.04779589e-01
1.45281464e-01 2.37613562e-02 6.20448291e-01 6.85606420e-01
1.63102448e-02 -6.32629812e-01 -4.42637324e-01 -2.90610164e-01
1.22028518e+00 -1.03098392e-01 6.31270647e-01 1.12755644e+00
3.27766960e-04 6.78724527e-01 7.99463093e-01 9.09845293e-01
7.19982266e-01 5.22321105e-01 2.84724981e-01 -4.85364161e-02
3.23753059e-02 -1.46273822e-01 7.41014838e-01 7.48435080e-01
-1.93898812e-01 9.44895297e-03 -8.33963096e-01 8.81955922e-01
-1.46109879e+00 -1.02952456e+00 -7.72874773e-01 2.02250743e+00
5.61372042e-01 3.23491812e-01 -7.36443773e-02 -9.48992893e-02
1.15269125e+00 5.85660756e-01 -2.92484779e-02 -6.59240663e-01
-3.39728534e-01 -9.49819833e-02 8.33750427e-01 6.81374073e-01
-9.74958003e-01 6.97674811e-01 6.36896086e+00 4.07435954e-01
-1.23963737e+00 7.45502040e-02 1.18403387e+00 -3.15796077e-01
-4.09919918e-01 5.19484520e-01 -5.49624681e-01 4.01842535e-01
9.02094722e-01 -5.12098372e-01 -3.22238028e-01 4.62883204e-01
1.05161583e+00 -6.92589223e-01 -1.30294994e-01 1.84930727e-01
-1.42084479e-01 -1.22545493e+00 -5.72932482e-01 6.72993660e-01
8.77306581e-01 4.16231036e-01 5.15327789e-02 3.30386788e-01
2.92502135e-01 -7.14839816e-01 9.16527092e-01 1.79743275e-01
5.34025729e-01 -7.10633099e-01 8.41380000e-01 2.92684972e-01
-3.89947951e-01 -7.23432451e-02 -1.65235609e-01 -1.23322463e+00
5.24977982e-01 1.00729179e+00 -1.11031365e+00 2.14618534e-01
3.53467464e-01 4.22486663e-01 -1.97638199e-01 3.70071769e-01
-3.30442637e-01 9.50471878e-01 -2.58256823e-01 1.31671084e-02
7.24865913e-01 -2.80390769e-01 1.11950207e+00 7.65351892e-01
1.96274713e-01 3.87133539e-01 -1.54551372e-01 7.80560300e-02
-8.95635411e-02 4.40264940e-01 -6.17353678e-01 -2.67102234e-02
5.26062071e-01 1.19024026e+00 -8.15584362e-01 -4.70393151e-01
-3.79200101e-01 1.77656144e-01 -2.36457407e-01 -2.55713388e-02
-6.21216595e-01 -2.91865203e-03 6.91251814e-01 7.55063832e-01
-1.32818118e-01 5.15578240e-02 -5.91369271e-01 -1.07738256e+00
-4.74069566e-01 -8.16477597e-01 2.56723553e-01 -3.75636816e-01
-7.31627822e-01 1.54359743e-01 -2.98074305e-01 -5.63606620e-01
1.95835173e-01 2.92605668e-01 -8.91917884e-01 6.97118044e-01
-1.07997417e+00 -8.17435801e-01 3.19918633e-01 -5.69642298e-02
3.18603039e-01 2.22246811e-01 5.97710371e-01 1.95508465e-01
-5.19196808e-01 -4.49766070e-02 -1.02590181e-01 -2.69535314e-02
7.25155830e-01 -5.88150263e-01 2.70184010e-01 8.02105367e-01
-6.13893390e-01 5.99673212e-01 1.41190279e+00 -9.87517178e-01
-8.54444921e-01 -7.63839543e-01 1.75670111e+00 -3.29398513e-01
8.87122989e-01 -1.20411150e-01 -4.46920050e-04 8.05371523e-01
5.65389991e-01 -7.72308886e-01 6.78398848e-01 5.43739259e-01
1.65366344e-02 2.88748652e-01 -1.17903090e+00 6.90907836e-01
3.62632006e-01 -3.58643979e-01 -6.25225544e-01 3.90382439e-01
1.16655216e-01 2.63817579e-01 -6.29489124e-01 2.87225783e-01
7.65956879e-01 -8.58526111e-01 6.19956493e-01 -4.61304277e-01
5.17662108e-01 -6.26198277e-02 -3.46358299e-01 -1.13058317e+00
-5.38766682e-01 -6.48186147e-01 1.06274426e+00 1.08620775e+00
5.65632880e-01 -7.93346465e-01 5.33649623e-01 8.65324974e-01
5.09218276e-02 -3.29561591e-01 -1.18862832e+00 -3.69847715e-02
1.20478615e-01 -2.08652511e-01 2.63907373e-01 1.62262011e+00
4.25623864e-01 6.24659956e-01 -4.91626769e-01 2.36182660e-01
2.20949754e-01 1.77440926e-01 1.04248810e+00 -1.15501523e+00
2.43331522e-01 -3.42082292e-01 -3.21372539e-01 -2.53584117e-01
1.20789865e-02 -6.09956205e-01 -4.50844467e-01 -1.18424714e+00
5.72832286e-01 -3.40428472e-01 1.57546088e-01 9.02999267e-02
-4.74026203e-02 3.46785367e-01 3.25404525e-01 1.91610113e-01
-1.54668093e-01 3.54913563e-01 1.10340703e+00 -2.61943489e-01
-2.70026147e-01 1.22682385e-01 -1.16601086e+00 7.46613443e-01
7.27088690e-01 -5.26922166e-01 4.15684998e-01 -6.69511333e-02
8.78493369e-01 1.57712445e-01 1.05461545e-01 -5.19948959e-01
1.56694558e-02 -2.62132704e-01 1.77691057e-01 -5.77764630e-01
2.96852410e-01 -4.47032392e-01 4.37602371e-01 8.33383441e-01
1.59012079e-02 6.52883500e-02 2.88679540e-01 2.49609619e-01
-2.79505670e-01 1.91596940e-01 5.21262527e-01 -1.24203369e-01
1.84522361e-01 3.73790739e-03 -1.15276635e+00 -2.73402274e-01
7.77594447e-01 1.11842975e-01 -2.44436517e-01 -7.69500017e-01
-8.52050245e-01 3.08219753e-02 7.53825665e-01 2.90088028e-01
-2.92797297e-01 -1.19396639e+00 -1.33300686e+00 -6.32752597e-01
-2.23752663e-01 -7.58616447e-01 7.74712265e-02 1.30588233e+00
-2.66021550e-01 4.64955717e-01 1.60661638e-01 2.52834827e-01
-1.53866959e+00 -2.60764416e-02 -1.61831573e-01 -2.03572884e-01
-3.93863440e-01 5.46652675e-01 3.30238640e-01 -3.39076638e-01
-3.75196129e-01 9.46497917e-02 -3.28906804e-01 9.42774475e-01
4.48459476e-01 6.69939101e-01 -2.23243326e-01 -1.05019808e+00
-5.84759295e-01 -7.66467378e-02 -3.54141444e-01 -2.43612632e-01
1.59344721e+00 -5.11940181e-01 -3.79037827e-01 7.55352437e-01
1.01327825e+00 6.67352498e-01 -3.73932838e-01 2.26036862e-01
-5.99316508e-02 -7.37466455e-01 3.17826003e-01 -6.08962536e-01
-7.26553261e-01 9.10773799e-02 4.09540720e-02 5.34254789e-01
4.67269331e-01 6.38953000e-02 5.74789405e-01 -1.30662039e-01
1.01878151e-01 -1.28131318e+00 -6.28875375e-01 4.55664068e-01
9.71463203e-01 -8.87534678e-01 6.10377073e-01 -1.77582964e-01
-3.53610665e-01 7.57784426e-01 5.02661243e-02 -5.70126437e-02
5.57093263e-01 -2.86677569e-01 1.30000010e-01 -5.90208590e-01
-6.49502516e-01 2.72695810e-01 -3.41351002e-01 -2.07763001e-01
7.58072615e-01 6.78472698e-01 -1.44709897e+00 4.02224034e-01
-6.86273396e-01 -2.98875332e-01 1.05076361e+00 5.52082896e-01
-5.19427896e-01 -7.87529886e-01 -7.71202326e-01 4.48728383e-01
-1.00398242e+00 -3.25344771e-01 -4.94967490e-01 1.05964482e+00
2.84894943e-01 9.33467209e-01 3.19942802e-01 -1.95187002e-01
-6.25189841e-02 -1.63537443e-01 2.19301544e-02 -5.04119635e-01
-7.89340734e-01 1.08639374e-01 1.02356422e+00 -3.79221626e-02
-5.39389968e-01 -1.49279737e+00 -7.07213461e-01 -1.10823298e+00
-7.00860918e-01 3.30321610e-01 1.16295230e+00 9.56912160e-01
2.73592144e-01 -6.30573481e-02 8.70557487e-01 -5.27266383e-01
-2.09671497e-01 -1.30836344e+00 -4.87910450e-01 1.18978873e-01
2.70302981e-01 -1.91662207e-01 -6.08781874e-01 -4.79940832e-01] | [8.806431770324707, 9.932908058166504] |
b8f65a60-c402-4ab2-8a12-6e8291872dff | re-centric-recommendations-for-the | 2306.01774 | null | https://arxiv.org/abs/2306.01774v1 | https://arxiv.org/pdf/2306.01774v1.pdf | RE-centric Recommendations for the Development of Trustworthy(er) Autonomous Systems | Complying with the EU AI Act (AIA) guidelines while developing and implementing AI systems will soon be mandatory within the EU. However, practitioners lack actionable instructions to operationalise ethics during AI systems development. A literature review of different ethical guidelines revealed inconsistencies in the principles addressed and the terminology used to describe them. Furthermore, requirements engineering (RE), which is identified to foster trustworthiness in the AI development process from the early stages was observed to be absent in a lot of frameworks that support the development of ethical and trustworthy AI. This incongruous phrasing combined with a lack of concrete development practices makes trustworthy AI development harder. To address this concern, we formulated a comparison table for the terminology used and the coverage of the ethical AI principles in major ethical AI guidelines. We then examined the applicability of ethical AI development frameworks for performing effective RE during the development of trustworthy AI systems. A tertiary review and meta-analysis of literature discussing ethical AI frameworks revealed their limitations when developing trustworthy AI. Based on our findings, we propose recommendations to address such limitations during the development of trustworthy AI. | ['Christian Berger', 'Jennifer Horkoff', 'Beatriz Cabrero-Daniel', 'Krishna Ronanki'] | 2023-05-29 | null | null | null | null | ['ethics'] | ['miscellaneous'] | [ 2.74343491e-01 5.80052137e-01 1.15677036e-01 -5.13126135e-01
-1.57914311e-02 -5.55430651e-01 4.82514739e-01 2.35740587e-01
-3.77356350e-01 5.11754394e-01 3.20898086e-01 -4.25311446e-01
-3.97794038e-01 -3.70464057e-01 -3.94396782e-01 -1.55262411e-01
7.21207023e-01 8.05064738e-02 -3.04471999e-01 -2.23368078e-01
5.12018681e-01 1.81077003e-01 -1.29018772e+00 5.13859726e-02
9.37043488e-01 8.15710366e-01 -5.37256360e-01 2.70044893e-01
1.60451576e-01 1.11444914e+00 -6.02477074e-01 -7.40686417e-01
3.79013896e-01 -6.13420486e-01 -9.15177822e-01 -3.35324645e-01
-1.79397509e-01 -7.41938591e-01 4.29396749e-01 8.80241096e-01
1.74369678e-01 -5.80514193e-01 6.84222102e-01 -1.84222138e+00
-9.65724528e-01 8.90798986e-01 9.79024097e-02 -3.07402402e-01
4.84506041e-01 2.90991783e-01 7.48124957e-01 -4.12570804e-01
7.23436594e-01 7.11587727e-01 5.21597803e-01 8.32915723e-01
-6.00285530e-01 -9.64075327e-01 -2.20662072e-01 -1.49227038e-04
-1.31954193e+00 -8.65147769e-01 5.63574731e-01 -7.62636423e-01
1.09195518e+00 1.78962097e-01 1.17411101e+00 9.83353555e-01
6.41228497e-01 6.08344749e-02 1.03722990e+00 -8.80431831e-01
5.24599373e-01 7.32370138e-01 3.13464165e-01 -1.61920458e-01
1.35034490e+00 -1.30003169e-02 -4.24215764e-01 -3.64250213e-01
1.71024889e-01 -6.16048455e-01 2.49086589e-01 9.10703372e-03
-1.01610434e+00 5.39343178e-01 -3.13925058e-01 9.08773541e-01
-5.40460110e-01 2.87844032e-01 6.22281432e-01 3.05018961e-01
1.97073072e-01 7.83935547e-01 7.75189251e-02 -6.26087546e-01
-6.96329832e-01 1.09875441e-01 1.06304383e+00 8.87123823e-01
1.37574866e-01 3.25257033e-01 3.04992348e-01 4.49566320e-02
1.12379718e+00 5.60854018e-01 -3.08581978e-01 -1.38579309e+00
-2.77579993e-01 8.54615092e-01 5.16697884e-01 -1.11778867e+00
4.13563810e-02 6.81824535e-02 -2.25571841e-01 4.16409999e-01
2.32438952e-01 -3.09442759e-01 -3.57971370e-01 1.23512614e+00
2.51552910e-01 -6.67264640e-01 4.99022603e-01 8.91541541e-01
6.11296117e-01 3.67800027e-01 4.64858264e-01 -2.96516448e-01
1.15954947e+00 -1.47391027e-02 -1.21846175e+00 -4.07274514e-01
7.71366656e-01 -4.52628762e-01 7.29178965e-01 7.46807992e-01
-1.18413258e+00 1.59064949e-01 -1.43688583e+00 6.62794784e-02
-9.16289911e-02 -2.57307947e-01 6.83836877e-01 1.33158696e+00
-9.53112841e-01 5.95147014e-02 -4.98792887e-01 -5.30086100e-01
3.33237231e-01 -4.29971069e-02 -3.95990133e-01 5.19824475e-02
-1.34510660e+00 1.52543664e+00 2.78316289e-01 6.15565419e-01
-4.88626719e-01 -4.30443466e-01 -7.88148224e-01 -4.36278760e-01
2.55383998e-01 -5.87181866e-01 1.15183401e+00 -1.52238619e+00
-1.26446509e+00 6.41400397e-01 6.25635684e-01 -5.38478136e-01
3.16626251e-01 -4.43246394e-01 -6.06361210e-01 1.26004115e-01
2.01637104e-01 2.49481961e-01 2.19848990e-01 -1.34460258e+00
-5.06082892e-01 -1.62074357e-01 -1.27650201e-01 -5.85259050e-02
-3.37816775e-01 7.75092304e-01 4.57588851e-01 -2.73272432e-02
-5.08634448e-01 -7.53057420e-01 -1.05204940e-01 6.09097034e-02
3.67947876e-01 -1.99854910e-01 2.92695105e-01 -4.15009856e-01
1.65248144e+00 -2.03765559e+00 -3.86518687e-01 3.65392715e-01
4.03285414e-01 2.33337641e-01 2.21692309e-01 9.01821136e-01
4.35652733e-01 5.43643534e-01 -1.30950466e-01 5.00240445e-01
7.00073242e-01 2.06374004e-01 -2.39715919e-01 6.30081534e-01
8.85911882e-02 6.39566302e-01 -1.00708461e+00 -4.96589571e-01
1.54544413e-01 7.59099185e-01 -1.61691368e-01 1.54049955e-02
2.91319102e-01 -3.07882037e-02 -4.88774300e-01 4.89495844e-01
3.53196263e-01 1.69951439e-01 4.69660610e-01 2.64716804e-01
-3.81721407e-01 3.34796041e-01 -8.71794760e-01 1.31079459e+00
-2.50939906e-01 6.32314086e-01 1.54827341e-01 -2.54791051e-01
1.21143448e+00 1.13511527e+00 4.57484871e-01 -6.32040501e-01
5.24922252e-01 7.94822454e-01 7.38987029e-01 -5.01553476e-01
8.16630125e-02 -2.57818878e-01 -3.50239575e-01 1.13349247e+00
-2.96342611e-01 -8.16212595e-01 1.23574939e-02 1.51486933e-01
7.46069729e-01 5.76127946e-01 7.62391567e-01 -4.47760671e-01
4.00213748e-01 1.86429575e-01 8.63578081e-01 1.30787283e-01
-6.20994270e-01 1.95657507e-01 4.13986295e-01 -7.79101133e-01
-9.21227217e-01 -1.64341509e-01 -2.41091773e-01 1.12638250e-01
-2.15991102e-02 -9.19972599e-01 -1.05650103e+00 -5.64722180e-01
-4.03613150e-01 1.53736365e+00 -7.32636154e-01 -4.18749690e-01
-4.63741496e-02 4.13605981e-02 6.60840034e-01 2.23021880e-01
3.31928432e-01 -1.35363770e+00 -2.13513660e+00 2.58381665e-01
2.18253165e-01 -1.18250740e+00 2.77109325e-01 -1.92994341e-01
-7.68944295e-03 -1.27146673e+00 -4.87307273e-02 6.42905459e-02
6.79830432e-01 -3.05060148e-01 8.99887562e-01 5.09936273e-01
-6.55928440e-03 5.98730385e-01 -6.51004672e-01 -1.14633775e+00
-1.11185825e+00 -5.81602097e-01 3.58731486e-02 -5.74150205e-01
1.15168822e+00 -2.13907450e-01 -3.12694728e-01 1.37062237e-01
-1.16495061e+00 4.12421264e-02 5.26106834e-01 4.50787768e-02
-2.68199086e-01 -1.51686281e-01 8.07300508e-01 -6.59269750e-01
1.09043932e+00 -4.63023841e-01 -2.03775227e-01 5.43198168e-01
-1.27535772e+00 -3.82574588e-01 2.21761063e-01 5.51816486e-02
-7.96117485e-01 -5.54260612e-01 1.12399034e-01 3.05384070e-01
-5.85709929e-01 5.40177882e-01 -1.79125294e-02 -1.74490139e-01
7.12875426e-01 -5.15813470e-01 3.24149400e-01 5.26459515e-01
-1.14000857e-01 9.61666405e-01 -2.99739808e-01 -7.55988657e-01
5.90397954e-01 -5.87655865e-02 -4.76268739e-01 -5.99860728e-01
-9.88055691e-02 1.90636039e-01 -4.26603049e-01 -8.17804873e-01
8.78466427e-01 -4.84537035e-01 -6.35731995e-01 -6.35463595e-02
-1.10683131e+00 -3.07125241e-01 -4.40159380e-01 7.45694816e-01
-5.05944610e-01 2.73330986e-01 3.51112373e-02 -1.46286356e+00
-9.69014168e-01 -1.09483540e+00 3.75035316e-01 -1.24381088e-01
-1.37780571e+00 -5.53770185e-01 2.72092581e-01 4.43764508e-01
6.52235925e-01 7.49755025e-01 6.67796791e-01 -6.58626556e-01
1.60818920e-01 -5.12416720e-01 3.80125165e-01 5.43674827e-01
3.70250195e-01 7.41446376e-01 -4.85533118e-01 3.47131908e-01
4.16838169e-01 -3.62272382e-01 -6.89257681e-01 3.60375904e-02
-2.11227417e-01 -8.26573670e-01 1.09850995e-01 -2.39158601e-01
1.26436734e+00 9.00936246e-01 7.25745499e-01 6.54849350e-01
1.93350554e-01 1.41980684e+00 1.15054798e+00 6.51690304e-01
4.98323292e-01 2.98686385e-01 1.12183705e-01 5.42537093e-01
6.02060795e-01 4.25612986e-01 5.60803354e-01 7.15540290e-01
-1.70607597e-01 1.59587711e-01 -1.65575767e+00 6.44714057e-01
-1.75873041e+00 -5.26289165e-01 -2.20273077e-01 1.88638616e+00
7.00451076e-01 2.79452384e-01 2.35611156e-01 2.15934500e-01
3.28044519e-02 -3.64484102e-01 -2.23018043e-02 -1.13792408e+00
4.89516884e-01 -2.20294997e-01 -1.00402422e-01 1.91189721e-01
-3.81090522e-01 6.80654347e-01 6.65738916e+00 -2.29690745e-01
-6.04591608e-01 -3.08598410e-02 1.77273989e-01 -6.56102924e-03
-1.13621950e+00 3.86615366e-01 -4.15421724e-02 2.35086113e-01
1.26066244e+00 -9.37670052e-01 3.50649375e-03 6.66293085e-01
3.45817357e-01 -2.14105234e-01 -1.12737548e+00 3.77652347e-01
1.26967236e-01 -9.64425683e-01 -1.44837379e-01 5.13128154e-02
2.50381976e-01 -6.82120025e-01 -1.73650771e-01 -4.19212520e-01
3.59522998e-01 -1.21910167e+00 1.25519359e+00 2.72817463e-01
5.31008244e-01 -8.86503577e-01 1.18977392e+00 1.06830753e-01
-3.89374673e-01 1.87186584e-01 -2.87621439e-01 -4.76337224e-01
1.46466762e-01 1.91336423e-01 -7.08240509e-01 4.49289173e-01
8.73030722e-01 5.61682999e-01 -1.00429475e-01 3.45095783e-01
-1.83665663e-01 4.71106261e-01 -3.68899077e-01 -1.95998564e-01
3.06079209e-01 -6.87603712e-01 4.52003181e-01 9.52763975e-01
1.93269879e-01 3.62694263e-01 -8.52992773e-01 1.08764136e+00
7.35564113e-01 2.26030841e-01 -1.11664569e+00 -6.60286009e-01
7.47452557e-01 9.63723898e-01 -6.83796406e-01 2.03473553e-01
-6.93158329e-01 2.23282740e-01 -6.52884305e-01 2.65003294e-01
-6.02639019e-01 -1.47441268e-01 5.40319085e-01 3.50561291e-01
-4.39362317e-01 -1.82374846e-02 -5.45804620e-01 -3.18411350e-01
2.15996847e-01 -1.42563295e+00 -1.15208991e-03 -6.32100821e-01
-7.66475022e-01 8.09740007e-01 5.40566146e-01 -1.21584618e+00
-4.02729779e-01 -2.22238153e-01 -1.03923254e-01 4.37101424e-01
-8.00590754e-01 -1.56263649e+00 -1.42751724e-01 -1.19984187e-01
-2.69920856e-01 2.45740078e-02 1.04958975e+00 -2.05700263e-01
-3.21510226e-01 3.07430863e-01 -7.39909708e-01 -3.05878818e-01
8.01201522e-01 -7.97038972e-01 1.79395452e-01 1.07162523e+00
-4.37889367e-01 1.08025467e+00 9.31944013e-01 -1.06404865e+00
-1.39077389e+00 -2.05502808e-01 1.08394396e+00 -5.28204501e-01
6.52030110e-01 -1.23279966e-01 -6.46369874e-01 7.32558191e-01
9.40016389e-01 -5.67347348e-01 1.35787046e+00 -5.11257887e-01
-1.33999348e-01 -1.27976075e-01 -1.39583361e+00 6.99655056e-01
5.31762183e-01 -3.36307824e-01 -9.27147567e-01 -2.52727240e-01
4.80007619e-01 1.71403378e-01 -1.20647991e+00 3.37880164e-01
1.03834963e+00 -1.04250348e+00 1.83374181e-01 -8.82828310e-02
2.11202517e-01 -4.11025405e-01 5.77299334e-02 -5.52540302e-01
-3.54958773e-02 -1.06527126e+00 4.37670439e-01 1.29540086e+00
3.50576073e-01 -9.63663399e-01 1.74191341e-01 1.93320000e+00
-6.35114014e-02 -2.55301356e-01 -9.62599814e-01 -1.32864028e-01
2.71834582e-01 -7.62934685e-01 8.52404416e-01 1.17940450e+00
8.03285897e-01 -2.47256115e-01 -2.48291254e-01 -2.31961489e-01
5.47291219e-01 -6.05641305e-01 7.16760993e-01 -1.30021656e+00
5.50559878e-01 -3.30076337e-01 -4.34040636e-01 4.56517816e-01
-2.91075936e-04 -3.43022853e-01 1.36829972e-01 -2.08762860e+00
-1.87972486e-01 -1.88558549e-01 8.27095956e-02 4.63253409e-01
2.19456181e-01 -3.51118416e-01 4.05039161e-01 1.10053554e-01
-1.56705439e-01 2.27433756e-01 6.79212093e-01 3.36710513e-01
-5.17731905e-02 -4.87938553e-01 -1.36020577e+00 6.21482015e-01
7.67289698e-01 -7.16691494e-01 -7.14905441e-01 -2.73359030e-01
1.28460681e+00 -2.65164018e-01 5.86329736e-02 -9.99220610e-01
4.62808102e-01 -6.27387106e-01 -3.16611856e-01 -5.96130043e-02
-3.35959285e-01 -1.92051959e+00 1.07398760e+00 5.13256848e-01
-3.69762450e-01 7.62466267e-02 3.69849652e-01 -3.75114888e-01
-8.80211741e-02 -6.87521160e-01 1.84859604e-01 1.87621936e-01
-2.91307062e-01 -3.62432837e-01 -8.96855235e-01 -1.22555017e-01
1.69092405e+00 -9.41567779e-01 -2.26223350e-01 -4.37756330e-01
5.62440865e-02 2.15475306e-01 1.13242650e+00 2.57456034e-01
8.24649036e-01 -1.08537364e+00 -8.14153492e-01 1.61671322e-02
1.89238444e-01 -5.95247149e-02 1.27953082e-01 7.90751517e-01
-9.25175667e-01 4.47838753e-01 -8.12268853e-01 1.27675429e-01
-8.45902562e-01 3.17221940e-01 2.32663900e-01 3.75876188e-01
-6.47593915e-01 1.41774192e-01 -1.56102031e-01 7.43010193e-02
1.22594729e-01 -1.87378019e-01 -4.64323342e-01 -3.62873077e-02
5.77708125e-01 3.69029462e-01 -3.74033123e-01 -8.80307138e-01
-8.96473706e-01 6.24284506e-01 2.13256046e-01 -6.99496806e-01
1.43324423e+00 -1.53878391e-01 -6.33938849e-01 6.51203513e-01
2.20000222e-01 1.10112159e-02 -7.27919757e-01 7.24931657e-01
3.82176697e-01 -4.16014969e-01 -1.81352675e-01 -9.80388284e-01
-5.01555264e-01 4.95370328e-01 1.55607443e-02 3.73186678e-01
1.01809931e+00 -5.29872715e-01 3.32333386e-01 3.84252751e-03
6.06673539e-01 -1.57088566e+00 -5.20800531e-01 -1.54389903e-01
1.24309075e+00 -7.09102452e-01 3.05186123e-01 -3.70445162e-01
-1.39019525e+00 1.35129619e+00 7.74349988e-01 2.00721726e-01
6.90481842e-01 6.91835225e-01 5.48587501e-01 -3.23899269e-01
-6.85952246e-01 4.80158478e-01 4.81671989e-02 7.71450281e-01
7.59081006e-01 -6.67901635e-02 -1.18718469e+00 1.00511670e+00
2.28304677e-02 6.47113621e-01 9.45534945e-01 1.20746505e+00
4.98128906e-02 -1.07360911e+00 -3.31485927e-01 -1.36074439e-01
-7.67466962e-01 7.68809319e-02 -1.27151787e+00 1.22599745e+00
-1.69512164e-02 1.51598442e+00 -1.20266557e-01 -4.61842209e-01
3.10296237e-01 3.51413749e-02 1.85223624e-01 -4.71152723e-01
-1.44945121e+00 -4.35197055e-02 6.75502598e-01 -4.95854884e-01
-8.71101737e-01 -5.98751307e-01 -1.23601830e+00 -2.97486454e-01
-6.34325296e-02 7.39736259e-01 6.89404905e-01 1.04287279e+00
4.24262315e-01 1.79542676e-01 -7.60929659e-02 2.30024412e-01
-2.77222604e-01 -7.25420713e-01 -2.71960050e-01 1.00602567e-01
5.41827059e-04 -1.64368376e-01 -4.45259064e-01 -2.98163109e-02] | [9.060908317565918, 6.346818923950195] |
28085484-cc08-47a9-aec0-57257b24f35d | scale-invariant-scale-channel-networks-deep | 2106.06418 | null | https://arxiv.org/abs/2106.06418v2 | https://arxiv.org/pdf/2106.06418v2.pdf | Scale-invariant scale-channel networks: Deep networks that generalise to previously unseen scales | The ability to handle large scale variations is crucial for many real world visual tasks. A straightforward approach for handling scale in a deep network is to process an image at several scales simultaneously in a set of scale channels. Scale invariance can then, in principle, be achieved by using weight sharing between the scale channels together with max or average pooling over the outputs from the scale channels. The ability of such scale channel networks to generalise to scales not present in the training set over significant scale ranges has, however, not previously been explored. In this paper, we present a systematic study of this methodology by implementing different types of scale channel networks and evaluating their ability to generalise to previously unseen scales. We develop a formalism for analysing the covariance and invariance properties of scale channel networks, and explore how different design choices, unique to scaling transformations, affect the overall performance of scale channel networks. We first show that two previously proposed scale channel network designs do not generalise well to scales not present in the training set. We explain theoretically and demonstrate experimentally why generalisation fails in these cases. We then propose a new type of foveated scale channel architecture}, where the scale channels process increasingly larger parts of the image with decreasing resolution. This new type of scale channel network is shown to generalise extremely well, provided sufficient image resolution and the absence of boundary effects. Our proposed FovMax and FovAvg networks perform almost identically over a scale range of 8, also when training on single scale training data, and do also give improved performance when learning from datasets with large scale variations in the small sample regime. | ['Tony Lindeberg', 'Ylva Jansson'] | 2021-06-11 | null | null | null | null | ['scale-generalisation'] | ['computer-vision'] | [ 4.99587715e-01 -2.01283749e-02 4.26842183e-01 -4.15796727e-01
-1.75652310e-01 -8.70328248e-01 5.83756089e-01 -1.45607382e-01
-8.47940743e-01 5.73266864e-01 -2.21890792e-01 -7.13349134e-02
-3.03784907e-01 -7.22644746e-01 -7.66007483e-01 -5.97942412e-01
-3.38958204e-01 9.93881598e-02 9.29690003e-01 -4.72557336e-01
6.28297031e-02 9.30132806e-01 -1.49082792e+00 3.54600906e-01
3.32985699e-01 1.10181510e+00 1.80210490e-02 1.15463531e+00
-9.93114896e-03 -3.09552345e-02 -8.50033164e-01 -9.86993760e-02
7.87369132e-01 -4.04182941e-01 -5.91314137e-01 1.39982942e-02
7.60266542e-01 -6.48522153e-02 -1.23490490e-01 1.19753993e+00
8.22722256e-01 -3.18548493e-02 5.27712047e-01 -8.68148386e-01
-5.36938965e-01 3.98235261e-01 -4.06360030e-01 6.06384754e-01
-8.72458369e-02 5.66159235e-03 7.07470298e-01 -4.71958876e-01
6.53478563e-01 1.37493122e+00 7.77940154e-01 4.36578959e-01
-1.55431485e+00 -5.22796512e-01 -9.29435343e-02 -2.79975086e-01
-1.24970853e+00 -2.04906434e-01 5.09080470e-01 -1.27317086e-01
1.03411198e+00 9.50011164e-02 5.38482308e-01 8.92517447e-01
4.83474195e-01 1.06222026e-01 1.96104622e+00 -4.15634692e-01
4.35604513e-01 2.63756383e-02 -2.53708929e-01 3.42318028e-01
3.32546741e-01 6.75199460e-03 -3.06445628e-01 2.97557622e-01
1.42784894e+00 -2.73422807e-01 -1.10206939e-01 -5.35396934e-01
-1.30844653e+00 7.31341779e-01 1.07777500e+00 4.91229594e-01
-1.04357660e-01 3.80020261e-01 3.34290355e-01 8.88538778e-01
3.22101295e-01 7.36810148e-01 -6.90983415e-01 1.70301288e-01
-8.22021484e-01 2.69419905e-02 6.73806369e-01 8.55526328e-01
7.00453520e-01 -1.49560258e-01 3.99546474e-02 5.34131825e-01
-2.51005411e-01 6.46630287e-01 6.43239439e-01 -8.99207413e-01
2.82466352e-01 4.50863779e-01 1.23009486e-02 -6.41570270e-01
-1.09785998e+00 -3.13399106e-01 -9.48762417e-01 7.95361817e-01
8.91970098e-01 -2.83579051e-01 -1.17793810e+00 1.86924553e+00
7.26833418e-02 -3.03787827e-01 2.11957693e-01 7.77130723e-01
5.72588086e-01 1.35822937e-01 -9.14580002e-02 -1.76647469e-01
1.45803344e+00 -2.05994323e-01 -3.04693222e-01 -3.16205561e-01
1.35883778e-01 -8.76190841e-01 1.11577034e+00 3.59039009e-01
-1.15713143e+00 -9.19300616e-01 -1.25016260e+00 4.45380099e-02
-7.92435229e-01 8.60546082e-02 6.52283490e-01 6.82547927e-01
-1.47762358e+00 8.81367087e-01 -6.68463647e-01 -8.79190743e-01
2.84885049e-01 7.83299267e-01 -4.51994598e-01 4.48232770e-01
-1.23486578e+00 8.75945389e-01 4.47976470e-01 7.30064809e-02
2.84382626e-02 -1.39472455e-01 -5.80728352e-01 8.99851844e-02
3.37589458e-02 -6.24547660e-01 1.04633009e+00 -1.30315650e+00
-1.32588434e+00 7.78541088e-01 4.33179885e-02 -6.02402389e-01
5.35431206e-01 3.09288893e-02 -3.13988298e-01 3.14379305e-01
-1.81352168e-01 1.06754160e+00 1.18307948e+00 -9.11315620e-01
-5.41256607e-01 -3.93149555e-01 3.69900852e-01 1.08580433e-01
-2.66146272e-01 9.37222540e-02 -2.15086892e-01 -5.21144092e-01
2.74433374e-01 -1.15965569e+00 -2.59978712e-01 2.81233847e-01
1.02289394e-01 6.59748986e-02 4.38059539e-01 -2.52410322e-01
9.73177433e-01 -2.14344645e+00 1.61914811e-01 4.08815771e-01
1.10764869e-01 1.58755302e-01 -3.80552709e-01 3.39541018e-01
-1.56345814e-01 7.37715960e-02 -6.15814365e-02 8.97717103e-02
-7.08792508e-02 2.99751610e-01 -1.60853900e-02 4.88881379e-01
4.05669600e-01 8.54676247e-01 -3.89411390e-01 -2.16379285e-01
5.12180328e-02 5.74912727e-01 -4.18866903e-01 -2.23770291e-01
2.13983297e-01 3.03149015e-01 -1.31524906e-01 1.37892127e-01
6.84801340e-01 -2.46475652e-01 -9.33613107e-02 -2.36737713e-01
-9.95341241e-02 -4.33502048e-01 -1.51045132e+00 1.45352280e+00
-4.52782899e-01 8.39898169e-01 2.42001384e-01 -8.08043182e-01
9.36720490e-01 5.70564866e-02 3.40482295e-02 -7.98881292e-01
2.01300547e-01 4.96016830e-01 4.75824088e-01 -2.30191648e-01
4.08324189e-02 -5.57286084e-01 -1.68266684e-01 2.43035913e-03
2.09369764e-01 -4.59483862e-01 2.23341167e-01 -2.01014981e-01
1.01987708e+00 -2.62159228e-01 6.01812065e-01 -3.77350807e-01
4.05752927e-01 -6.08210862e-01 4.71265502e-02 8.85565877e-01
-5.26943386e-01 7.26095855e-01 7.42971122e-01 -5.23206472e-01
-1.43250227e+00 -1.06936431e+00 -4.32641953e-01 1.06437826e+00
1.72805175e-01 1.43231198e-01 -8.18000078e-01 -4.44954991e-01
1.57465916e-02 -2.37689838e-01 -1.04794419e+00 2.58939639e-02
-5.89008451e-01 -7.48944998e-01 5.45448720e-01 7.81589925e-01
8.27776194e-01 -1.24823356e+00 -1.02509522e+00 1.26514733e-01
6.33338749e-01 -1.22895908e+00 -4.84999008e-02 6.91428363e-01
-8.71013403e-01 -8.80317330e-01 -8.42138886e-01 -7.28045881e-01
4.88972634e-01 -2.69220863e-02 7.78757334e-01 -2.61118203e-01
-4.41039264e-01 3.21674973e-01 -1.95032984e-01 -4.51847434e-01
-3.41029435e-01 1.69971570e-01 2.07157865e-01 -1.31504118e-01
9.07812715e-02 -8.26259792e-01 -8.78482163e-01 5.20700276e-01
-1.23000646e+00 -2.13637486e-01 1.04586732e+00 6.96171403e-01
4.28998023e-01 2.17292756e-01 6.39470279e-01 -6.47168398e-01
6.30269408e-01 3.23818535e-01 -5.65511286e-01 4.21457142e-02
-4.63944286e-01 3.64097893e-01 8.73021424e-01 -6.33278906e-01
-7.01252282e-01 1.15937866e-01 -4.04048376e-02 -8.43628077e-04
-2.96695977e-01 -1.70772761e-01 -2.18094531e-02 -6.99479878e-01
1.22498131e+00 -1.58754021e-01 1.55451432e-01 -2.20267385e-01
5.92001736e-01 4.50644672e-01 2.43522406e-01 -1.61408231e-01
9.84727144e-01 9.37376797e-01 5.72997510e-01 -1.00967574e+00
-3.37753564e-01 -5.51911056e-01 -1.05669916e+00 5.99892549e-02
9.75071609e-01 -6.50987267e-01 -3.35163951e-01 4.26267087e-01
-6.83969796e-01 -5.75479567e-01 -5.46672761e-01 4.43051815e-01
-8.42675447e-01 1.26401916e-01 -6.19412959e-01 -4.01672274e-01
-6.03172630e-02 -1.05268121e+00 1.13864017e+00 5.36861718e-01
5.01743378e-03 -1.13225114e+00 -2.52343595e-01 -4.15498227e-01
8.74775350e-01 2.52462000e-01 7.17296898e-01 -6.30270898e-01
-1.62185416e-01 -1.98527843e-01 -6.77307367e-01 4.89061505e-01
6.25072122e-02 -1.36301354e-01 -1.04881108e+00 -7.16924369e-01
-1.55441046e-01 -2.24579051e-01 1.09386158e+00 5.66167474e-01
5.56091011e-01 1.43576428e-01 1.42480498e-02 6.66048646e-01
1.64778769e+00 -1.06171638e-01 5.68437457e-01 5.54395497e-01
2.80117065e-01 5.75347066e-01 5.84135577e-02 4.00942899e-02
-4.84342992e-01 6.70413673e-01 2.12166503e-01 -4.01705235e-01
-3.78690392e-01 3.26297492e-01 1.30810946e-01 8.80969763e-02
-2.95929909e-01 1.63578957e-01 -4.89993155e-01 2.13220239e-01
-1.35555410e+00 -7.66650736e-01 2.39241317e-01 2.25024986e+00
6.55858994e-01 6.28612638e-01 2.57994205e-01 4.57908064e-02
5.31390548e-01 7.14878067e-02 -6.04939222e-01 -7.48084247e-01
-4.38626528e-01 6.79811358e-01 1.02108002e+00 2.97734708e-01
-1.34139025e+00 8.96077037e-01 6.72369576e+00 4.66855526e-01
-1.48716152e+00 -6.00199215e-02 2.66321033e-01 -1.39158547e-01
2.03024715e-01 -2.95669194e-02 -7.20455468e-01 1.60045475e-02
9.25294399e-01 1.05672255e-01 4.53970164e-01 6.78909123e-01
1.43440261e-01 -7.60496110e-02 -8.98182273e-01 7.48242140e-01
-1.76473230e-01 -8.06237400e-01 -1.47471607e-01 1.12072796e-01
7.74674892e-01 2.22803786e-01 2.50582218e-01 2.64722735e-01
-1.13698639e-01 -9.91672039e-01 4.80711639e-01 3.40299875e-01
9.95225310e-01 -4.96845722e-01 6.96884215e-01 8.42962638e-02
-1.22428143e+00 -5.13086021e-01 -8.23772371e-01 -1.38843656e-01
-1.84356421e-01 1.31944954e-01 -5.13127983e-01 4.24091183e-02
7.89196491e-01 1.24144584e-01 -1.23475921e+00 1.12577224e+00
4.44423184e-02 1.50707573e-01 -6.47292376e-01 -3.37478258e-02
2.97121972e-01 2.19783261e-01 3.58671904e-01 1.45301044e+00
3.23795229e-01 -3.00724387e-01 -1.54659137e-01 3.54330629e-01
1.27364486e-01 2.08494082e-01 -5.78371644e-01 3.62250119e-01
1.87820941e-01 1.31778920e+00 -1.41427147e+00 -1.95538938e-01
-7.25902617e-01 1.11794388e+00 3.37599605e-01 7.31289148e-01
-3.99133414e-01 -5.73774397e-01 3.21135283e-01 2.57067770e-01
7.34076023e-01 -3.23507220e-01 -2.68969297e-01 -9.91116226e-01
4.35337126e-02 -6.12209618e-01 2.49625728e-01 -7.36848414e-01
-1.08573830e+00 6.08365595e-01 2.89655954e-01 -1.11757112e+00
-6.44260496e-02 -1.14465022e+00 -4.83429015e-01 1.01014173e+00
-1.52793944e+00 -1.12457883e+00 -2.06273377e-01 4.95047212e-01
3.26784521e-01 -3.13142203e-02 6.39247894e-01 -2.71928877e-01
1.43725097e-01 6.11245930e-01 1.38783827e-01 4.89216074e-02
9.41199303e-01 -1.69564390e+00 6.09207034e-01 7.37794042e-01
6.15926459e-02 9.12390351e-01 7.91378558e-01 -2.75119901e-01
-9.10841107e-01 -8.90039623e-01 3.10109764e-01 -2.39112705e-01
8.03743660e-01 -5.68587840e-01 -9.18287873e-01 3.60049337e-01
1.29106939e-01 3.46175969e-01 2.72995859e-01 5.00186756e-02
-5.19778728e-01 -3.22357595e-01 -1.10953724e+00 4.85017121e-01
9.18435395e-01 -6.08904421e-01 -5.60780883e-01 3.15642282e-02
5.75880170e-01 -2.21900851e-01 -9.48555112e-01 4.13280189e-01
7.30307341e-01 -1.26171744e+00 9.29571390e-01 -4.74498630e-01
5.41111231e-02 -2.71734148e-01 -1.21904783e-01 -1.28101254e+00
-3.65754724e-01 -4.97372240e-01 3.87576073e-01 7.42010713e-01
4.88561362e-01 -8.69909763e-01 5.84364533e-01 2.18613297e-01
2.80935973e-01 -5.03764927e-01 -1.29242361e+00 -1.05569041e+00
4.67839867e-01 -6.80053160e-02 2.00444311e-01 3.20313603e-01
-3.04697871e-01 3.28598917e-01 1.10561170e-01 3.05079311e-01
4.27020848e-01 4.49823812e-02 7.05830634e-01 -1.26033449e+00
-5.09235263e-01 -7.78117955e-01 -1.11093831e+00 -8.67850363e-01
-3.60415608e-01 -4.54244018e-01 -1.49285197e-01 -1.41758752e+00
-5.88323176e-02 -1.08872943e-01 -3.04509342e-01 1.22302279e-01
-1.35068774e-01 7.46558785e-01 2.46124268e-01 3.29589814e-01
-2.99012601e-01 -1.26007825e-01 1.54528594e+00 3.66064727e-01
-3.31849515e-01 1.43668741e-01 -4.65501606e-01 7.55719543e-01
8.43509614e-01 1.04475975e-01 -5.05279541e-01 -4.64540906e-02
4.95992810e-01 -2.83923209e-01 4.96245950e-01 -1.51989710e+00
7.30453581e-02 1.43499911e-01 9.01762664e-01 1.30046964e-01
1.74968064e-01 -8.30611289e-01 -1.66542932e-01 3.15331042e-01
-5.25104344e-01 1.44293085e-01 4.61152285e-01 5.49602866e-01
-2.81173959e-02 -7.39162788e-02 1.26189196e+00 -2.13439181e-01
-8.77270818e-01 2.15362608e-02 -1.31849408e-01 -1.24724112e-01
8.88564706e-01 -4.76117700e-01 -2.11947829e-01 -3.23526621e-01
-1.34840930e+00 -2.59975791e-01 7.23372996e-01 3.33524972e-01
2.57965416e-01 -1.16770840e+00 -3.70802790e-01 4.87397760e-01
9.43735316e-02 -1.31478786e-01 2.08497435e-01 7.33883679e-01
-6.50896549e-01 3.39148581e-01 -5.46940029e-01 -5.41863143e-01
-1.19142354e+00 7.78356791e-01 7.21131444e-01 -5.85088469e-02
-5.46445429e-01 8.74379992e-01 4.68295962e-01 -1.13513656e-01
2.43844627e-03 -9.30142283e-01 -2.52937257e-01 1.50031880e-01
4.61982042e-01 -2.21306667e-01 1.59123212e-01 -5.99921465e-01
-2.85917129e-02 1.17105687e+00 1.14039797e-02 -2.73064822e-01
9.94217277e-01 -2.27974385e-01 1.61729753e-01 5.39651811e-01
1.40220869e+00 -1.96651489e-01 -1.68874669e+00 -2.02640802e-01
-2.35812485e-01 -3.48520726e-02 -3.25994164e-01 -6.20473266e-01
-6.82037055e-01 9.51300263e-01 1.16946065e+00 6.02702200e-01
1.41683781e+00 7.90650547e-02 2.46694013e-01 3.50791991e-01
2.14551732e-01 -1.22944403e+00 1.31059840e-01 3.13923478e-01
9.96031880e-01 -1.23687387e+00 -1.51194958e-02 -3.06191921e-01
-2.72308797e-01 1.49820185e+00 4.62751061e-01 -5.25165081e-01
5.96506834e-01 3.36221308e-01 2.20232353e-01 -1.69798851e-01
-3.31205755e-01 -5.46754003e-01 3.46398622e-01 7.46177077e-01
3.77017081e-01 -1.48142185e-02 -4.47553009e-01 3.48352119e-02
-2.48187691e-01 -2.17537880e-01 5.35206676e-01 9.77295876e-01
-8.49239349e-01 -9.13289607e-01 -4.85182822e-01 3.04432034e-01
-4.21914309e-01 -9.49108377e-02 -5.23912430e-01 1.18557060e+00
3.26186925e-01 5.17684281e-01 7.16706812e-02 -3.12137902e-02
6.14584804e-01 2.26276532e-01 6.54401422e-01 -4.60664839e-01
-4.60558563e-01 3.10553610e-01 -2.60372490e-01 -5.90129733e-01
-5.59731364e-01 -5.34650922e-01 -1.24953294e+00 1.68100223e-01
-1.80661693e-01 -3.36542577e-01 8.19575310e-01 6.92833006e-01
6.54756278e-02 5.66379309e-01 3.66645068e-01 -1.04137838e+00
-7.71984935e-01 -1.01322436e+00 -7.83610284e-01 6.45218194e-01
5.17598927e-01 -5.67467511e-01 -5.27098060e-01 -4.25357744e-02] | [9.218524932861328, 2.3135428428649902] |
42421043-6d5f-4ec2-81cf-a717d9af27e5 | mesh-sort-simple-and-effective-of-location | 2302.14415 | null | https://arxiv.org/abs/2302.14415v3 | https://arxiv.org/pdf/2302.14415v3.pdf | Mesh-SORT: Simple and effective location-wise tracker with lost management strategies | Multi-Object Tracking (MOT) has gained extensive attention in recent years due to its potential applications in traffic and pedestrian detection. We note that tracking by detection may suffer from errors generated by noise detectors, such as an imprecise bounding box before the occlusions, and observed that in most tracking scenarios, objects tend to move and lost within specific locations. To counter this, we present a novel tracker to deal with the bad detector and occlusions. Firstly, we proposed a location-wise sub-region recognition method which equally divided the frame, which we called mesh. Then we proposed corresponding location-wise loss management strategies and different matching strategies. The resulting Mesh-SORT, ablation studies demonstrate its effectiveness and made 3% fragmentation 7.2% ID switches drop and 0.4% MOTA improvement compared to the baseline on MOT17 datasets. Finally, we analyze its limitation on the specific scene and discussed what future works can be extended. | ['ZongTan Li'] | 2023-02-28 | null | null | null | null | ['pedestrian-detection', 'human-detection'] | ['computer-vision', 'computer-vision'] | [ 5.46337850e-02 -2.74351001e-01 6.73335418e-02 -7.65615106e-02
-6.24604106e-01 -4.38826948e-01 4.97641832e-01 9.98182669e-02
-5.61788678e-01 8.29429507e-01 -2.43564427e-01 -4.28065658e-02
1.28422186e-01 -7.00201631e-01 -8.82080257e-01 -6.12550616e-01
-7.84372464e-02 4.55743015e-01 1.20492160e+00 2.29900941e-01
6.29489198e-02 5.75210333e-01 -1.83455408e+00 1.30771875e-01
6.07127905e-01 1.00723732e+00 1.46224990e-01 6.36028409e-01
-2.08882540e-02 4.75124061e-01 -9.56640959e-01 -4.88657683e-01
4.26271230e-01 5.43547533e-02 -1.78056613e-01 1.95438284e-02
1.10755038e+00 -3.94172132e-01 -4.07352269e-01 9.83828425e-01
6.90416098e-01 8.62388834e-02 3.92079026e-01 -1.71968913e+00
-2.27548331e-02 2.08796278e-01 -9.83996511e-01 5.90551674e-01
9.37838852e-02 3.35615396e-01 3.57825994e-01 -6.78551435e-01
5.14303923e-01 1.59493196e+00 1.03628230e+00 6.14211738e-01
-1.02390695e+00 -9.42233920e-01 1.41144440e-01 3.30940634e-01
-1.44512582e+00 -4.93900716e-01 2.98238456e-01 -3.79427671e-01
5.62165618e-01 4.83780473e-01 5.53618252e-01 8.71526957e-01
3.33794177e-01 9.44576859e-01 8.74646544e-01 -3.44168842e-01
-1.70377642e-02 1.76205367e-01 1.67879730e-01 5.30022204e-01
7.19657302e-01 3.62781256e-01 -2.97538161e-01 -1.07846417e-01
6.09773397e-01 -9.27483290e-02 1.35662347e-01 -4.42751110e-01
-9.82256234e-01 5.63600302e-01 3.76389533e-01 2.91320551e-02
6.00511357e-02 3.99037927e-01 4.47162986e-01 -3.11741810e-02
4.98561084e-01 -1.64595321e-01 -1.95474207e-01 -2.02497929e-01
-1.18755877e+00 4.74236012e-01 3.83069158e-01 1.21382034e+00
4.26020235e-01 -1.33174688e-01 -5.45665205e-01 5.54150403e-01
3.08466852e-01 7.42226660e-01 -8.77491385e-02 -9.45806921e-01
5.98006427e-01 4.82350111e-01 1.73266292e-01 -1.05472314e+00
-5.16775787e-01 -4.53973711e-01 -6.50870502e-01 4.29689556e-01
8.50812793e-01 -1.18722014e-01 -7.84366012e-01 1.48438215e+00
6.59539223e-01 7.14951396e-01 -5.51612139e-01 9.55100060e-01
6.64576828e-01 1.40648961e-01 2.23969772e-01 -1.81035727e-01
1.52362454e+00 -8.57725263e-01 -8.01405191e-01 -2.30125770e-01
4.76456732e-01 -9.06192541e-01 3.82854253e-01 3.58022064e-01
-1.24906242e+00 -7.52557039e-01 -9.56861854e-01 3.28419447e-01
-3.58708143e-01 6.54568672e-02 3.80439192e-01 1.10842967e+00
-9.84676719e-01 4.96774435e-01 -9.72069204e-01 -6.57942533e-01
8.68488133e-01 4.24300671e-01 2.00041592e-01 5.69276251e-02
-7.88559496e-01 8.96676183e-01 3.40613216e-01 -9.85433608e-02
-6.93198860e-01 -8.01879048e-01 -4.99101758e-01 -1.71321526e-01
5.35370946e-01 -7.07895756e-01 1.01983929e+00 -4.23183501e-01
-9.44550872e-01 7.43674159e-01 -4.49421406e-01 -8.58075678e-01
1.00152755e+00 -5.86550772e-01 -6.17289364e-01 -1.40896529e-01
2.24564910e-01 6.81458175e-01 7.48850167e-01 -1.11415946e+00
-1.14147735e+00 -2.32580423e-01 -1.73866659e-01 -2.04095110e-01
-1.20586336e-01 4.18877870e-01 -7.38427520e-01 -7.87932456e-01
3.62353139e-02 -1.02655506e+00 -2.77791500e-01 3.73365521e-01
-3.06219190e-01 -1.33509323e-01 1.22123671e+00 -3.46175253e-01
1.40966117e+00 -2.10079741e+00 -4.74464059e-01 -1.36803444e-02
3.36679280e-01 5.79977751e-01 1.85178965e-01 3.26633127e-03
3.21600497e-01 3.80647881e-03 1.78101614e-01 -6.29590154e-01
-1.53666988e-01 -2.61425897e-02 -2.57712305e-01 7.86126852e-01
1.29811969e-02 7.10426927e-01 -9.12167370e-01 -8.45228910e-01
6.00817859e-01 5.25227606e-01 -4.38368499e-01 -2.54948974e-01
-1.27431452e-01 3.84413153e-01 -3.71472239e-01 9.71024096e-01
1.28516281e+00 2.76668444e-02 -4.00955200e-01 -3.51587892e-01
-4.19901639e-01 -4.30421382e-02 -1.70218623e+00 1.27520978e+00
6.18964992e-02 8.68546367e-01 3.95968333e-02 -6.30054593e-01
6.47931933e-01 5.51431328e-02 6.64866805e-01 -7.31484711e-01
1.83350787e-01 -1.07799256e-02 -1.50003672e-01 -2.16069013e-01
8.22183251e-01 3.08093041e-01 1.43969461e-01 -7.64409825e-02
-5.20869792e-01 6.08875275e-01 1.66558683e-01 2.38488138e-01
1.20611989e+00 1.50757626e-01 7.21330196e-02 -2.14122221e-01
4.84131128e-01 1.75877854e-01 8.23551238e-01 1.12917805e+00
-8.75758231e-01 7.15804815e-01 -1.04336008e-01 -4.58754003e-01
-7.21600056e-01 -1.06195974e+00 -4.35651630e-01 8.31407845e-01
5.60007811e-01 -4.78769809e-01 -5.32423496e-01 -5.28539002e-01
3.37875575e-01 3.64091903e-01 -2.96714753e-01 1.62566677e-02
-1.12178433e+00 -8.61659110e-01 8.37773144e-01 6.73731267e-01
7.44593620e-01 -8.19535971e-01 -8.84596586e-01 3.59474868e-01
-3.00187916e-02 -1.40043938e+00 -4.79808658e-01 -2.88538009e-01
-9.06391442e-01 -1.20433390e+00 -6.70805156e-01 -2.57131785e-01
4.13360834e-01 6.74532235e-01 1.13976502e+00 1.55685365e-01
-5.29765427e-01 2.35470921e-01 -2.00690418e-01 -5.86363196e-01
-2.35663071e-01 -1.33409932e-01 2.91467220e-01 -6.05676025e-02
4.62886035e-01 -2.39366576e-01 -8.23515713e-01 7.14804590e-01
-3.59350055e-01 -2.52014518e-01 4.73720968e-01 3.65578532e-01
4.92389590e-01 1.59301206e-01 1.95543483e-01 -5.50949454e-01
-1.24175966e-01 -3.09408277e-01 -9.86692369e-01 5.63820936e-02
-3.80293667e-01 -2.90420443e-01 1.43763751e-01 -5.22273123e-01
-8.80314231e-01 1.36887923e-01 7.37776682e-02 -4.45648164e-01
-2.21380606e-01 -5.79452395e-01 -3.45321804e-01 -3.89374346e-01
3.73732358e-01 -1.88550353e-01 -2.15017244e-01 -5.84417343e-01
2.55106032e-01 4.45408493e-01 4.23151284e-01 -4.92401838e-01
9.58848715e-01 9.43351924e-01 1.70861274e-01 -7.32182741e-01
-5.09392738e-01 -7.46436536e-01 -4.33057427e-01 -5.03190041e-01
7.76838005e-01 -1.00954378e+00 -1.10363209e+00 4.29120839e-01
-1.28436196e+00 1.43360108e-01 -1.93048924e-01 3.63814324e-01
-1.68387920e-01 6.20186031e-01 -4.20770794e-01 -1.28731358e+00
-1.65769845e-01 -1.03755403e+00 1.17463434e+00 3.51921916e-01
9.07911360e-03 -5.11169851e-01 -1.15017585e-01 1.94616571e-01
4.47587043e-01 3.40352714e-01 4.42653038e-02 -3.57751161e-01
-1.15889907e+00 -5.24948724e-03 -5.12532592e-01 -1.29263982e-01
-1.86070070e-01 7.22202193e-03 -1.02999973e+00 -6.12426043e-01
-3.78837705e-01 1.92320853e-01 1.11020660e+00 6.94496214e-01
9.20017481e-01 6.39115795e-02 -1.10244656e+00 5.50244153e-01
1.33629763e+00 1.58026710e-01 6.40526593e-01 7.08870232e-01
6.65455222e-01 2.57199973e-01 9.29420114e-01 3.92345130e-01
1.31562233e-01 1.35727870e+00 4.63988841e-01 9.77911577e-02
-6.95980251e-01 4.83936332e-02 3.94214272e-01 1.82300061e-01
5.88734075e-02 -3.81921798e-01 -5.21497369e-01 5.58629274e-01
-2.00597024e+00 -1.13858843e+00 -6.62447393e-01 2.30535030e+00
3.31060797e-01 6.70076013e-01 8.17874312e-01 1.14189491e-01
1.07745898e+00 2.26510596e-02 -4.84791338e-01 2.10318044e-01
-2.54211068e-01 -4.83774319e-02 1.14705360e+00 3.92124563e-01
-1.33915818e+00 9.49336946e-01 6.37212086e+00 1.03386939e+00
-5.77546120e-01 3.73083353e-01 2.04002947e-01 -2.73002863e-01
5.12822449e-01 -1.14427656e-01 -1.73411405e+00 7.67161131e-01
8.33590329e-01 2.39003092e-01 -1.78696975e-01 6.87350094e-01
3.88457358e-01 -2.96031594e-01 -7.80102432e-01 1.09388328e+00
-1.56294644e-01 -1.08547091e+00 -1.86675012e-01 1.99077114e-01
3.48851025e-01 -4.73892987e-02 -3.99592780e-02 3.56863946e-01
2.78495371e-01 -5.05790770e-01 8.23335826e-01 4.30881411e-01
5.11435390e-01 -5.97382307e-01 4.92617011e-01 2.85991013e-01
-1.77018023e+00 1.16961241e-01 -5.90280414e-01 -6.02062531e-02
4.90943432e-01 7.72368073e-01 -6.62700176e-01 6.29565775e-01
8.72426212e-01 5.60832381e-01 -9.32734668e-01 1.82865977e+00
3.71360004e-01 7.07591057e-01 -7.21189201e-01 6.56180382e-02
-1.29066378e-01 1.19426817e-01 1.05930173e+00 1.44673955e+00
2.87253112e-01 -2.97042310e-01 3.88624191e-01 7.27096617e-01
2.75756925e-01 -2.89301723e-01 -3.82699400e-01 7.87759423e-01
6.52361214e-01 1.09566569e+00 -9.40362990e-01 -4.63334262e-01
-5.90751231e-01 7.47851670e-01 -4.07862477e-02 1.69840410e-01
-1.46892571e+00 -5.33112921e-02 8.06207716e-01 5.34038246e-01
5.76303780e-01 -1.24945212e-03 -3.22085440e-01 -9.81816471e-01
2.50796616e-01 -3.71788025e-01 5.20572901e-01 -2.50911623e-01
-9.28531051e-01 3.14037263e-01 2.83049375e-01 -1.47592854e+00
2.85837680e-01 -4.18837041e-01 -4.46651489e-01 4.50148642e-01
-1.39075398e+00 -1.03006625e+00 -3.92629057e-01 2.62477160e-01
6.60414755e-01 2.03148052e-02 1.55685190e-02 1.13016713e+00
-8.10541928e-01 1.00524986e+00 8.45307261e-02 1.58447757e-01
8.83262992e-01 -9.09725785e-01 5.99444032e-01 1.08504295e+00
-3.84586304e-02 2.94577509e-01 9.32273924e-01 -9.15454686e-01
-1.02459896e+00 -1.47759748e+00 8.57235551e-01 -7.10423648e-01
3.21801335e-01 -4.46546942e-01 -7.96357810e-01 6.65645540e-01
-3.61577831e-02 4.02186245e-01 2.74170395e-02 -2.34197706e-01
5.73443659e-02 -3.27186227e-01 -1.41950500e+00 5.47634423e-01
1.49785662e+00 2.19472736e-01 -3.94432038e-01 2.83855468e-01
5.18005788e-01 -5.92272580e-01 -5.06444156e-01 5.19812346e-01
7.17653394e-01 -1.07333410e+00 1.40554023e+00 -2.17620358e-01
-6.45004570e-01 -7.99948394e-01 7.94151351e-02 -6.06194794e-01
-4.88258272e-01 -5.99814534e-01 -5.76578617e-01 1.66755533e+00
-1.01225212e-01 -4.91073698e-01 1.06808412e+00 5.02571881e-01
-1.41235977e-01 -3.04634541e-01 -1.26774156e+00 -1.26640880e+00
-1.70551017e-01 -5.56213677e-01 4.05205131e-01 4.39886391e-01
-6.54560387e-01 -7.65566230e-02 -5.46765864e-01 4.69748110e-01
1.02257514e+00 -2.50974268e-01 9.76873517e-01 -1.23159301e+00
-1.73929911e-02 -4.91911620e-01 -7.21890092e-01 -1.45821321e+00
-2.69788057e-01 -4.53726709e-01 7.40665421e-02 -1.09093142e+00
2.15418667e-01 -7.27188528e-01 -3.08298707e-01 8.28865319e-02
-3.74122977e-01 4.33910161e-01 4.19683784e-01 1.53553069e-01
-1.17733467e+00 2.29774192e-01 8.97844911e-01 -2.29849353e-01
-1.11148402e-01 3.75160426e-01 -1.95549697e-01 6.08171523e-01
6.29815221e-01 -8.71425033e-01 2.01590657e-01 -2.46605605e-01
-1.61675483e-01 -3.38854849e-01 8.22012305e-01 -1.70510149e+00
5.13341010e-01 1.02521107e-01 4.40117240e-01 -1.26082957e+00
4.18573201e-01 -1.06249118e+00 3.43753874e-01 7.38036931e-01
2.82989472e-01 1.21582657e-01 7.00385869e-01 7.32600451e-01
1.57481700e-01 7.85667673e-02 8.64455640e-01 3.75434421e-02
-8.37768734e-01 1.86321020e-01 -2.28724808e-01 5.30946404e-02
1.44387221e+00 -6.55140638e-01 -3.44535232e-01 1.11426972e-01
-4.89903778e-01 4.69698966e-01 4.28880990e-01 5.21637261e-01
4.34459090e-01 -1.48186970e+00 -7.66193032e-01 8.06957409e-02
-9.02163982e-02 -2.59796798e-01 8.54431465e-02 9.95741606e-01
-2.97860295e-01 5.38097620e-01 5.92612810e-02 -9.85068977e-01
-1.72202849e+00 6.37927353e-01 2.39678353e-01 -1.95733726e-01
-9.21456873e-01 6.85951352e-01 7.14610517e-02 2.12466359e-01
6.01576388e-01 -4.83468235e-01 -2.04995628e-02 -1.11615844e-01
7.63043940e-01 1.05776191e+00 -2.31674369e-02 -8.33617866e-01
-8.70319903e-01 8.36953759e-01 -1.70649111e-01 1.32016242e-01
5.96249402e-01 -3.94323438e-01 4.72598374e-01 1.31244034e-01
7.29033113e-01 2.89112721e-02 -1.47920215e+00 -1.11661680e-01
2.09145010e-01 -8.22892129e-01 -1.32722378e-01 -5.87015033e-01
-1.12140584e+00 4.75677162e-01 1.26067865e+00 1.53328478e-01
8.18871617e-01 -1.10056944e-01 1.18893969e+00 -8.57305601e-02
5.32932878e-01 -8.53113890e-01 -3.24172765e-01 2.78369933e-01
2.34344810e-01 -1.34350920e+00 1.90389767e-01 -6.61127627e-01
-9.94993374e-02 8.19954693e-01 7.60927081e-01 -7.46243820e-02
5.50029159e-01 4.48310584e-01 -1.73447296e-01 -1.37654215e-01
-3.99208099e-01 -5.40213764e-01 1.98087379e-01 7.14081645e-01
1.29911676e-01 -1.08328342e-01 -2.83332169e-01 1.91813633e-01
1.64472908e-01 4.61427607e-02 5.93509823e-02 1.02442813e+00
-6.02081835e-01 -1.06969821e+00 -9.71310437e-01 4.26696926e-01
-5.35818040e-01 1.98383912e-01 9.92266312e-02 1.04262710e+00
6.33919895e-01 1.20167041e+00 2.08908364e-01 -2.98826844e-01
7.70197690e-01 -2.68235117e-01 4.89878178e-01 -1.29291266e-01
-7.52187312e-01 1.61540598e-01 5.36322929e-02 -8.77988756e-01
-5.95207632e-01 -1.03404140e+00 -1.07865655e+00 -4.58120465e-01
-6.27220750e-01 -1.77490219e-01 2.02094793e-01 7.33248293e-01
4.83290017e-01 7.49438405e-01 6.45600036e-02 -7.21956313e-01
-2.62427181e-01 -8.12639832e-01 -2.45066583e-01 4.95200515e-01
5.59876382e-01 -1.14674866e+00 -3.85341913e-01 -7.74422288e-02] | [6.522732257843018, -1.9880743026733398] |
f5c9dafa-4557-4123-a521-af51973f9b1d | enhancing-adversarial-training-via | 2306.14275 | null | https://arxiv.org/abs/2306.14275v3 | https://arxiv.org/pdf/2306.14275v3.pdf | Enhancing Adversarial Training via Reweighting Optimization Trajectory | Despite the fact that adversarial training has become the de facto method for improving the robustness of deep neural networks, it is well-known that vanilla adversarial training suffers from daunting robust overfitting, resulting in unsatisfactory robust generalization. A number of approaches have been proposed to address these drawbacks such as extra regularization, adversarial weights perturbation, and training with more data over the last few years. However, the robust generalization improvement is yet far from satisfactory. In this paper, we approach this challenge with a brand new perspective -- refining historical optimization trajectories. We propose a new method named \textbf{Weighted Optimization Trajectories (WOT)} that leverages the optimization trajectories of adversarial training in time. We have conducted extensive experiments to demonstrate the effectiveness of WOT under various state-of-the-art adversarial attacks. Our results show that WOT integrates seamlessly with the existing adversarial training methods and consistently overcomes the robust overfitting issue, resulting in better adversarial robustness. For example, WOT boosts the robust accuracy of AT-PGD under AA-$L_{\infty}$ attack by 1.53\% $\sim$ 6.11\% and meanwhile increases the clean accuracy by 0.55\%$\sim$5.47\% across SVHN, CIFAR-10, CIFAR-100, and Tiny-ImageNet datasets. | ['Mykola Pechenizkiy', 'Yulong Pei', 'Lu Yin', 'Vlaod Menkovski', 'Li Shen', 'Meng Fang', 'Tianlong Chen', 'Shiwei Liu', 'Tianjin Huang'] | 2023-06-25 | null | null | null | null | ['adversarial-robustness'] | ['adversarial'] | [ 2.43168883e-03 -9.78555083e-02 1.41692504e-01 -3.18268597e-01
-9.67143416e-01 -7.33682692e-01 3.60081345e-01 -3.55336696e-01
-5.81622243e-01 8.40021133e-01 -1.87192112e-02 -4.95515645e-01
-8.40989202e-02 -7.61728108e-01 -9.03007090e-01 -7.21644342e-01
-1.62159845e-01 -1.78672910e-01 1.36590451e-01 -5.18815815e-01
-1.72967628e-01 5.36760807e-01 -7.53281951e-01 -8.33584741e-02
1.00857699e+00 1.12526584e+00 -6.32323384e-01 4.22104210e-01
1.64071813e-01 7.27692068e-01 -9.61839378e-01 -1.00529480e+00
6.91688538e-01 -1.52246505e-01 -4.52132642e-01 -5.72036982e-01
7.20761597e-01 -3.44707608e-01 -8.26191962e-01 1.37022233e+00
7.57296979e-01 2.31981963e-01 2.83616781e-01 -1.31682909e+00
-8.31810534e-01 7.30478406e-01 -4.55557287e-01 3.24497581e-01
-2.82517463e-01 3.41636479e-01 6.93078935e-01 -7.14500666e-01
2.74726748e-01 1.15774906e+00 1.00550389e+00 9.63621557e-01
-1.05640626e+00 -1.32327509e+00 4.12418723e-01 -1.09014407e-01
-1.21864510e+00 -4.70859557e-01 8.35471094e-01 -2.47136146e-01
7.75073647e-01 2.63260782e-01 2.13194683e-01 1.44025409e+00
1.64876893e-01 6.38272583e-01 9.63454068e-01 4.79577258e-02
1.57711864e-01 -2.35935356e-02 7.01182857e-02 5.82765400e-01
1.39850616e-01 4.18055564e-01 -3.48676629e-02 6.98128203e-03
5.85613608e-01 2.09821656e-01 -2.64871985e-01 1.39917925e-01
-7.61639357e-01 8.23193491e-01 8.10733914e-01 1.35510698e-01
-4.68481565e-03 4.57051724e-01 6.36964202e-01 3.45106661e-01
5.21285772e-01 6.07863545e-01 -4.18645203e-01 1.00452348e-03
-7.30462968e-01 4.96244133e-01 5.25265872e-01 7.25631475e-01
2.11715668e-01 9.84394014e-01 -6.19582422e-02 8.35023880e-01
7.76942745e-02 6.13563180e-01 3.46947074e-01 -1.01878023e+00
8.00749898e-01 3.19295794e-01 -2.28049099e-01 -1.03926182e+00
-2.12391600e-01 -8.64163637e-01 -1.12510538e+00 4.91587400e-01
5.12764931e-01 -5.49392581e-01 -1.12773001e+00 2.29515195e+00
6.36850521e-02 3.30188662e-01 2.37234920e-01 5.83268821e-01
6.27588928e-01 5.56038618e-01 1.80911526e-01 4.61136922e-02
6.94736123e-01 -9.45596755e-01 -6.35579586e-01 -4.05522585e-01
1.29729182e-01 -6.17013454e-01 1.01274955e+00 4.30388570e-01
-1.05786920e+00 -5.23947656e-01 -1.27918112e+00 3.85110259e-01
-3.74806941e-01 -4.74591464e-01 6.18622541e-01 1.03966415e+00
-6.18620753e-01 8.58419240e-01 -8.62251461e-01 3.26310128e-01
9.26988006e-01 4.23568517e-01 -3.30180675e-01 -1.27554417e-01
-1.46313238e+00 8.09496105e-01 3.25854063e-01 1.04963146e-01
-1.17682624e+00 -1.12293029e+00 -7.76307225e-01 -6.12319745e-02
4.01944816e-01 -3.08455020e-01 9.65777874e-01 -9.08625841e-01
-1.43541491e+00 3.32974434e-01 2.33704612e-01 -8.44479501e-01
7.31962442e-01 -5.17854393e-01 -8.40632975e-01 -1.51897147e-01
-2.22327426e-01 4.56870377e-01 8.96724582e-01 -1.24769521e+00
-1.66866153e-01 -4.45550770e-01 1.73579141e-01 -1.51560262e-01
-7.19220281e-01 -3.65873314e-02 -3.15845013e-01 -1.33182025e+00
-1.14789732e-01 -9.72349465e-01 -4.96928185e-01 -2.07983360e-01
-4.46168274e-01 2.31793091e-01 8.69051576e-01 -6.32634640e-01
1.34518909e+00 -2.27116466e+00 -4.86775711e-02 1.50932491e-01
3.03575397e-01 8.87983501e-01 -4.26066518e-02 8.03425163e-02
-3.02964061e-01 6.08292937e-01 -4.88321126e-01 -2.80090272e-01
8.37726519e-02 3.69044274e-01 -7.15238869e-01 4.83676642e-01
2.13513181e-01 9.81631994e-01 -9.08460915e-01 -7.27912709e-02
6.27948791e-02 6.12379253e-01 -6.31757081e-01 -3.92639711e-02
4.08603922e-02 4.10032719e-01 -4.49487716e-01 6.72315061e-01
7.90251493e-01 2.32361943e-01 -2.99979031e-01 -4.04222235e-02
2.90196061e-01 -1.08650997e-01 -1.20574999e+00 1.53853595e+00
-1.39580831e-01 5.45720518e-01 3.65293175e-02 -1.07568312e+00
7.14158714e-01 2.95013458e-01 4.85017776e-01 -7.25125313e-01
3.05123657e-01 1.22043543e-01 1.46289647e-01 -1.89470470e-01
3.12741131e-01 -3.45440358e-01 -3.33625734e-01 7.05337711e-03
1.68163657e-01 2.08055563e-02 -8.17859471e-02 1.85467869e-01
1.22044075e+00 -1.39160994e-02 -4.95757982e-02 -4.85642850e-02
4.04934734e-01 -2.46735945e-01 9.71141338e-01 7.43427753e-01
-5.48005104e-01 5.56373239e-01 3.09808642e-01 -3.10197443e-01
-8.84169519e-01 -1.35698485e+00 -2.09031910e-01 8.53057504e-01
1.24537600e-02 -2.14657292e-01 -9.48019981e-01 -8.94036651e-01
1.30831942e-01 7.72155821e-01 -5.93707681e-01 -5.84425569e-01
-8.41971099e-01 -9.99082983e-01 1.26623178e+00 7.22746134e-01
9.09953952e-01 -8.56465876e-01 1.04047805e-01 1.37969568e-01
2.14046717e-01 -1.14608896e+00 -5.48547149e-01 1.24334432e-01
-8.11443329e-01 -6.66910112e-01 -6.66453660e-01 -5.13924301e-01
6.03229463e-01 6.68953061e-02 9.30337787e-01 -3.34794745e-02
-1.61217257e-01 -4.61292751e-02 -2.27031305e-01 -6.40210807e-01
-4.77759033e-01 -8.66885483e-02 4.54895765e-01 -1.05993874e-01
-1.26903892e-01 -8.05314183e-01 -5.72862267e-01 3.48519385e-01
-1.09185100e+00 -6.29106581e-01 3.09947491e-01 1.00394297e+00
6.24047399e-01 2.61061132e-01 9.73097324e-01 -1.04384935e+00
6.11429274e-01 -6.01978660e-01 -5.93077242e-01 -7.45693687e-03
-7.63642430e-01 3.01200505e-02 1.27842510e+00 -7.33519793e-01
-9.24812317e-01 -3.69649142e-01 -7.15505838e-01 -9.35492575e-01
1.05187707e-01 2.55468071e-01 -3.00575882e-01 -4.00969028e-01
1.01390815e+00 -6.06812350e-02 -1.80230632e-01 -3.59446406e-01
3.95499587e-01 1.36561885e-01 8.91493618e-01 -5.97235322e-01
1.52142918e+00 3.64290982e-01 -1.71906039e-01 -3.85532975e-01
-9.60227489e-01 1.74871787e-01 -4.67613563e-02 -1.13880962e-01
6.30181491e-01 -8.85028958e-01 -6.58574700e-01 7.26771891e-01
-6.44239247e-01 -4.24098194e-01 -3.76751631e-01 2.09504366e-01
-3.37394774e-02 3.17891747e-01 -5.25318742e-01 -6.20181203e-01
-6.93154454e-01 -1.33144021e+00 1.89778104e-01 3.30521315e-01
1.93575457e-01 -9.93816316e-01 -1.99878171e-01 3.53242010e-01
7.21278846e-01 9.07964051e-01 5.75356781e-01 -7.95476973e-01
-3.80048692e-01 -6.58273458e-01 -2.82630697e-02 1.02094853e+00
-6.25600070e-02 -1.03934370e-01 -1.08747923e+00 -4.23803270e-01
1.61791891e-01 -3.18992525e-01 8.35068643e-01 1.43273085e-01
1.44285214e+00 -6.03719532e-01 4.73481528e-02 1.12543762e+00
1.53394985e+00 3.58622223e-01 6.87025428e-01 2.79119372e-01
8.78822386e-01 -6.61839694e-02 3.44833851e-01 1.35175943e-01
-1.03882357e-01 4.36132520e-01 8.02046120e-01 -3.97698022e-02
-1.83554757e-02 -2.47885704e-01 5.75431049e-01 6.38197839e-01
-6.49218038e-02 -2.16911897e-01 -8.00813377e-01 3.95137906e-01
-1.40427148e+00 -1.14048076e+00 2.38831729e-01 2.04534221e+00
9.47939038e-01 7.22465277e-01 -8.82482156e-02 3.51270199e-01
3.71806890e-01 3.85034263e-01 -8.45967829e-01 -4.50028777e-01
-9.44301113e-02 5.53801835e-01 8.66582215e-01 3.01694512e-01
-1.18484879e+00 1.03489840e+00 5.56104708e+00 9.95874822e-01
-1.22268367e+00 2.87210435e-01 6.40696824e-01 -2.65946805e-01
-1.04996964e-01 -2.48722956e-01 -7.23309994e-01 6.35999620e-01
8.05330575e-01 -1.72406793e-01 6.72642589e-01 9.97900546e-01
-1.92018524e-01 5.26498556e-01 -7.97982633e-01 7.28868008e-01
3.13595235e-02 -1.37703300e+00 -3.59105431e-02 -7.84952119e-02
1.04105651e+00 1.99367300e-01 5.12020528e-01 7.28980541e-01
7.39072144e-01 -1.27042079e+00 8.03133428e-01 3.08545172e-01
7.88805664e-01 -9.97556567e-01 6.71289861e-01 3.12458813e-01
-1.00746238e+00 -3.62126350e-01 -1.83269098e-01 2.85114944e-01
8.52060914e-02 4.65917557e-01 -4.17318106e-01 5.23564160e-01
8.98996830e-01 4.65900570e-01 -5.40420771e-01 7.35859752e-01
-3.28404516e-01 9.11990166e-01 -1.86034307e-01 2.84524679e-01
3.88464391e-01 -4.89749908e-02 8.40013564e-01 9.22304451e-01
9.31031629e-02 1.01265863e-01 -3.44032906e-02 7.84884334e-01
-5.97610235e-01 -2.64127761e-01 -6.03511989e-01 5.20876795e-02
6.11155689e-01 1.00784445e+00 -1.77631900e-01 -7.16125071e-02
-2.13092893e-01 7.46068716e-01 2.93804705e-01 5.07028341e-01
-1.36300516e+00 -5.18091142e-01 1.02762735e+00 -2.56551504e-01
9.89652947e-02 -2.72336662e-01 -4.45312172e-01 -1.07003641e+00
1.80387244e-01 -1.21727824e+00 3.01064670e-01 -2.44421139e-01
-1.34832656e+00 8.49940836e-01 -2.10048988e-01 -1.16508591e+00
3.41692939e-02 -5.91208756e-01 -8.17117214e-01 9.12138283e-01
-1.43987918e+00 -9.10520375e-01 -2.18204319e-01 7.75646269e-01
4.34954703e-01 -4.94679689e-01 7.09631264e-01 6.26446605e-01
-1.08147025e+00 1.47394490e+00 2.97113925e-01 4.87317979e-01
6.61471903e-01 -1.21732104e+00 7.38535523e-01 1.26789761e+00
-1.22441696e-02 7.40486145e-01 6.81052566e-01 -3.42597842e-01
-1.30831516e+00 -1.51065016e+00 1.66564658e-01 -5.37155747e-01
8.94704342e-01 -3.10711712e-01 -1.15556157e+00 6.79955304e-01
8.19517896e-02 4.12789941e-01 4.70966280e-01 -3.08850616e-01
-6.58297420e-01 -5.92389941e-01 -1.38054562e+00 9.83035088e-01
1.13733339e+00 -4.22334820e-01 -3.85590285e-01 1.67549074e-01
1.13656211e+00 -7.30778694e-01 -1.26032984e+00 6.31006062e-01
2.46438950e-01 -6.36098444e-01 1.42930460e+00 -7.88137913e-01
3.21216166e-01 -1.67068336e-02 -3.41395706e-01 -1.38637757e+00
-1.13495506e-01 -1.03104150e+00 -1.57008126e-01 1.40553510e+00
4.82943743e-01 -9.19774055e-01 6.83126509e-01 5.80742121e-01
-5.53733349e-01 -9.88277316e-01 -1.02561665e+00 -1.03625214e+00
6.79520786e-01 -7.34263420e-01 6.18597448e-01 1.00964689e+00
-5.28288305e-01 -4.86123487e-02 -4.66105431e-01 3.30645949e-01
7.24580944e-01 -6.24136925e-01 8.45941007e-01 -7.93974876e-01
-2.67854989e-01 -6.03855848e-01 -3.33767563e-01 -5.64121664e-01
2.11791918e-01 -9.09297109e-01 -1.72797069e-01 -1.01874030e+00
-4.14765269e-01 -6.53288722e-01 -6.68186486e-01 6.69968724e-01
-4.69811738e-01 3.31594348e-01 3.32743973e-01 -1.13349468e-01
-3.09342295e-01 4.49157238e-01 1.18475676e+00 -2.16242343e-01
4.45960835e-02 8.70221779e-02 -9.96229351e-01 7.08291054e-01
1.07658303e+00 -6.16337657e-01 -3.91822308e-01 -5.74809015e-01
1.06304139e-01 -1.35126457e-01 4.55946654e-01 -1.30845654e+00
1.33341411e-02 -9.21566933e-02 3.75128031e-01 -9.58711430e-02
3.11395198e-01 -7.37717390e-01 9.82402787e-02 4.52567548e-01
-2.24399865e-01 2.04290658e-01 6.15169287e-01 7.24646509e-01
-9.95014310e-02 -6.85534254e-03 1.13163197e+00 4.16506119e-02
-5.20152569e-01 6.20549679e-01 2.56585032e-01 6.13859713e-01
9.83832240e-01 4.87919897e-02 -4.90496457e-01 -1.28945664e-01
-5.50828874e-01 3.14291716e-01 1.92421287e-01 5.42142987e-01
5.33923209e-01 -1.43117094e+00 -6.48623228e-01 1.77862272e-01
-3.15646827e-01 7.76993111e-02 4.29517061e-01 5.61004221e-01
-2.33842865e-01 -4.74539176e-02 -1.59523994e-01 -1.99004486e-01
-1.01683176e+00 6.68651938e-01 6.97080076e-01 -3.76435965e-01
-5.24595082e-01 1.12301576e+00 -7.05997199e-02 -3.77484709e-01
5.34901440e-01 -7.60256201e-02 3.23698595e-02 -3.41021061e-01
3.35841507e-01 4.89842296e-01 1.51307002e-01 -4.88730669e-01
-2.39791155e-01 3.40231508e-01 -1.94472820e-01 4.05022129e-02
1.37282598e+00 3.80740434e-01 2.31430903e-01 4.48423922e-02
1.35643792e+00 1.92739785e-01 -1.60109532e+00 -3.80888730e-01
-3.97123963e-01 -2.89743870e-01 -3.68798301e-02 -8.61714482e-01
-1.67524958e+00 8.43932867e-01 6.32057011e-01 2.27774441e-01
1.23556232e+00 -5.26151359e-01 1.10475993e+00 2.80726820e-01
2.31564403e-01 -8.89717519e-01 1.59509465e-01 6.38077140e-01
8.92780662e-01 -1.25381422e+00 2.50815749e-02 -2.61007007e-02
-6.44690692e-01 6.46381080e-01 7.75489807e-01 -5.02819538e-01
7.21814930e-01 3.46406251e-01 8.13720599e-02 1.20935142e-01
-3.35239708e-01 2.70364314e-01 2.89315999e-01 4.39751685e-01
7.44811376e-04 -5.18022552e-02 3.91415842e-02 9.00524974e-01
-3.90396237e-01 -3.03372502e-01 1.05788708e-01 8.55691433e-01
-5.71622550e-02 -1.08831036e+00 -4.34009075e-01 3.86802167e-01
-1.23501837e+00 -1.56233460e-01 -1.01234764e-01 9.72020686e-01
1.84643075e-01 8.09741616e-01 -4.33656007e-01 -7.19085157e-01
5.69010496e-01 9.76237953e-02 2.42548093e-01 -1.29019976e-01
-1.00258815e+00 -2.83839703e-01 -1.55520812e-01 -6.32494271e-01
9.03224349e-02 -4.45207804e-01 -1.29542899e+00 -4.85488266e-01
-2.02324986e-01 -4.36042286e-02 5.14998198e-01 8.56040001e-01
2.54209459e-01 9.35220599e-01 8.73132825e-01 -6.56988919e-01
-1.04872763e+00 -8.06668937e-01 -2.24600106e-01 5.48526525e-01
3.49788070e-01 -5.72141528e-01 -5.49571037e-01 -1.64686084e-01] | [5.603680610656738, 7.899422645568848] |
26dbecc4-9075-413c-941a-9b9dc6973787 | semantics-preserving-sketch-embedding-for | 2211.13015 | null | https://arxiv.org/abs/2211.13015v2 | https://arxiv.org/pdf/2211.13015v2.pdf | Semantics-Preserving Sketch Embedding for Face Generation | With recent advances in image-to-image translation tasks, remarkable progress has been witnessed in generating face images from sketches. However, existing methods frequently fail to generate images with details that are semantically and geometrically consistent with the input sketch, especially when various decoration strokes are drawn. To address this issue, we introduce a novel W-W+ encoder architecture to take advantage of the high expressive power of W+ space and semantic controllability of W space. We introduce an explicit intermediate representation for sketch semantic embedding. With a semantic feature matching loss for effective semantic supervision, our sketch embedding precisely conveys the semantics in the input sketches to the synthesized images. Moreover, a novel sketch semantic interpretation approach is designed to automatically extract semantics from vectorized sketches. We conduct extensive experiments on both synthesized sketches and hand-drawn sketches, and the results demonstrate the superiority of our method over existing approaches on both semantics-preserving and generalization ability. | ['Xiaoyan Sun', 'Zihan Chen', 'Chi Zhang', 'Chaoqun Wang', 'Xuejin Chen', 'Binxin Yang'] | 2022-11-23 | null | null | null | null | ['face-generation'] | ['computer-vision'] | [ 4.33897465e-01 3.18516977e-02 -3.16637039e-01 -6.63800597e-01
-3.54930758e-01 -5.95722675e-01 8.04315209e-01 -5.12616515e-01
1.93348452e-01 5.05261779e-01 3.91887307e-01 1.03989057e-01
-3.26447971e-02 -9.28346515e-01 -7.22410202e-01 -3.47370356e-01
4.24956352e-01 2.03325495e-01 -2.42796123e-01 -2.94035703e-01
1.27316967e-01 6.36421621e-01 -1.47311068e+00 1.66326940e-01
6.03042901e-01 1.02879417e+00 1.69376135e-01 3.78387332e-01
-6.13262117e-01 5.13489723e-01 -5.00708461e-01 -7.92785347e-01
4.07263190e-01 -6.26828611e-01 -3.70326400e-01 4.39809471e-01
7.23923802e-01 -4.12546545e-01 -6.34419739e-01 9.50824797e-01
2.99737543e-01 -7.33945817e-02 6.37066245e-01 -1.70228493e+00
-1.44648564e+00 4.20173764e-01 -4.14778143e-01 -5.57331860e-01
3.31651926e-01 2.18201846e-01 1.18566477e+00 -1.23591828e+00
9.79233205e-01 1.45799804e+00 4.60319161e-01 8.40777576e-01
-1.41204965e+00 -9.73904550e-01 3.14980358e-01 -1.64755687e-01
-1.41417170e+00 -4.05821741e-01 1.40420759e+00 -1.62678406e-01
4.21132147e-01 1.30581781e-01 8.14572275e-01 1.23901188e+00
-3.92342180e-01 9.27632451e-01 8.67171943e-01 -2.24993378e-01
9.65693220e-02 3.15092027e-01 -4.75637347e-01 9.85354781e-01
9.20128301e-02 4.16282490e-02 -6.89705431e-01 -9.96056572e-02
1.42303848e+00 2.93167651e-01 -2.33660564e-01 -7.54043341e-01
-1.13470459e+00 9.48778272e-01 5.46732187e-01 1.54559091e-01
-2.71508366e-01 6.04060888e-01 2.99656570e-01 3.05287808e-01
3.50653112e-01 4.31845456e-01 8.37306213e-03 2.30397955e-01
-8.62482488e-01 3.68861079e-01 5.04197359e-01 1.38582671e+00
5.99397123e-01 3.47723424e-01 -2.81725824e-01 9.87816513e-01
5.73348254e-02 8.51904750e-01 2.89576203e-01 -9.36534226e-01
5.15758634e-01 7.22861111e-01 5.50195426e-02 -1.23447084e+00
3.41596991e-01 2.57553030e-02 -9.49377775e-01 1.86771274e-01
2.09619431e-03 3.35612774e-01 -7.40913808e-01 1.93098831e+00
6.13919497e-02 2.77827948e-01 3.94876562e-02 9.38414752e-01
7.51905978e-01 5.69183886e-01 1.83830068e-01 3.02052736e-01
1.32223368e+00 -8.20658207e-01 -8.96316826e-01 -4.25315164e-02
-1.35702297e-01 -6.07114911e-01 1.55960023e+00 -2.18978584e-01
-1.19696927e+00 -7.71742582e-01 -1.34456789e+00 -2.48729885e-01
-2.62311101e-01 4.39949691e-01 6.81872249e-01 4.87361312e-01
-8.03359926e-01 5.57839870e-01 -3.49624902e-01 -1.34024143e-01
9.05515015e-01 2.02689961e-01 -6.05939209e-01 -7.54369572e-02
-1.04395998e+00 4.63743806e-01 9.51815322e-02 -9.79993939e-02
-6.86107457e-01 -8.67244780e-01 -1.01104128e+00 1.96737871e-01
1.64506793e-01 -8.23706627e-01 1.04382241e+00 -1.40051270e+00
-1.67894220e+00 9.16792691e-01 -2.19535947e-01 1.96950715e-02
5.73399365e-01 1.63731053e-02 -3.00555795e-01 2.55437285e-01
4.56032008e-02 9.94368434e-01 1.23550439e+00 -1.58999574e+00
-3.49489540e-01 -2.01774150e-01 8.32295790e-02 1.73114818e-02
-6.34738386e-01 -2.45537221e-01 -4.87788916e-01 -1.22193527e+00
-2.53644343e-02 -6.42264664e-01 -4.21636999e-02 7.86584020e-01
-3.89765531e-01 -8.63009170e-02 8.72502625e-01 -3.36731970e-01
8.85725856e-01 -2.12801528e+00 3.48734498e-01 3.34272444e-01
2.13416815e-01 2.07530975e-01 -5.91468573e-01 6.75228059e-01
-5.34580983e-02 1.35301813e-01 -2.96776742e-01 -4.04731870e-01
3.20421249e-01 3.68863404e-01 -1.12517059e+00 4.42371070e-02
6.85993135e-01 1.34342229e+00 -1.00963438e+00 -3.90026331e-01
2.03969330e-01 7.16851890e-01 -5.64873815e-01 4.71782804e-01
-3.20590883e-01 2.27764711e-01 -7.25580633e-01 6.26633883e-01
5.79226613e-01 -2.53607571e-01 2.69072711e-01 -2.24464133e-01
3.99503917e-01 -8.31128657e-02 -9.42354798e-01 2.02142549e+00
-7.49367654e-01 5.31757116e-01 -1.15001127e-01 -8.15980434e-01
1.23245764e+00 3.50141674e-01 1.51392445e-01 -6.97562635e-01
-2.12146193e-01 1.83667406e-01 -6.81394696e-01 -2.38104731e-01
4.06940997e-01 -4.87521529e-01 -2.48602945e-02 6.09144986e-01
-3.24713513e-02 -5.73396683e-01 -3.60075057e-01 1.77149683e-01
5.00199437e-01 4.42932576e-01 3.21461968e-02 -1.27165526e-01
5.57982445e-01 -2.82063931e-01 3.11290890e-01 4.24983978e-01
2.33852953e-01 7.83732593e-01 5.49074590e-01 -6.60727799e-01
-1.31431544e+00 -1.35055780e+00 2.13056043e-01 7.28037953e-01
3.36506009e-01 -3.86618435e-01 -7.41912127e-01 -7.49535143e-01
2.43985206e-01 5.41969359e-01 -6.64755344e-01 -3.08071971e-01
-6.84039176e-01 9.42918360e-02 5.79687595e-01 8.97943497e-01
5.03225863e-01 -1.23607886e+00 -3.00893813e-01 5.74604906e-02
1.02855884e-01 -1.22653747e+00 -8.97433162e-01 -7.53018498e-01
-8.44195485e-01 -8.23389232e-01 -9.52895224e-01 -1.06183541e+00
1.04445386e+00 3.71029735e-01 9.83223200e-01 2.68569142e-01
-5.42483032e-01 4.83599037e-01 -6.83357045e-02 -2.04466224e-01
-3.28731328e-01 -2.13606209e-01 -1.21496268e-01 3.34424078e-01
7.03630550e-03 -6.90488338e-01 -6.92368627e-01 1.72872409e-01
-1.22563756e+00 5.86709797e-01 4.97205019e-01 1.10202157e+00
4.41737980e-01 -2.71358401e-01 8.55482399e-01 -8.46985996e-01
7.26896584e-01 -4.23451066e-02 -5.91138244e-01 4.86035198e-01
-4.25531805e-01 4.44530249e-01 9.01651144e-01 -5.64792335e-01
-1.06270289e+00 1.38655558e-01 -8.43442231e-02 -8.98219407e-01
1.39817342e-01 -7.21312240e-02 -4.24641252e-01 4.48941365e-02
1.84021801e-01 4.10160959e-01 2.66536713e-01 -2.53848135e-01
8.99047196e-01 4.67141241e-01 6.51478887e-01 -9.92442429e-01
1.17197895e+00 7.63708651e-01 8.54946859e-03 -7.24324524e-01
-5.22288859e-01 3.35918993e-01 -3.27208400e-01 2.12584864e-02
5.93004048e-01 -8.05726945e-01 -6.78885520e-01 1.86891139e-01
-1.20318639e+00 -3.47017497e-02 -5.73858380e-01 -7.62429014e-02
-7.85774767e-01 2.65936643e-01 -4.20541227e-01 -7.04656959e-01
-4.05830741e-01 -1.19559562e+00 1.47735751e+00 5.91415018e-02
-2.40940899e-01 -8.46498191e-01 -3.58427107e-01 7.37802014e-02
5.47509611e-01 3.56794894e-01 1.34718823e+00 1.67153135e-03
-7.87417591e-01 -2.51345128e-01 -6.25052512e-01 2.81359911e-01
3.96180332e-01 -4.59191464e-02 -8.76622260e-01 -9.84826311e-02
-4.74492401e-01 -3.99112821e-01 6.66556120e-01 -1.43780082e-01
1.44137561e+00 -4.59748358e-01 -1.54976398e-01 6.59515679e-01
1.53149474e+00 -1.19136311e-01 4.94991630e-01 -2.66594619e-01
7.64845014e-01 6.21795118e-01 2.85323650e-01 3.61843556e-01
2.72725284e-01 7.70312667e-01 1.89064384e-01 -2.50440627e-01
-4.71976340e-01 -1.21281183e+00 2.13644817e-01 5.11425793e-01
1.44485727e-01 6.54431060e-02 -3.44001740e-01 4.36516762e-01
-1.69063723e+00 -8.17612588e-01 7.00348556e-01 2.06194735e+00
8.13027322e-01 -1.59260303e-01 -1.16818197e-01 2.06408743e-02
6.69311106e-01 4.01517093e-01 -5.05036712e-01 -2.28675082e-01
-1.20742016e-01 5.46136498e-01 1.02477252e-01 3.81820112e-01
-7.52285182e-01 1.23747289e+00 6.27564192e+00 8.53198647e-01
-1.17372906e+00 -1.84195206e-01 4.55425292e-01 5.07880375e-02
-9.88942981e-01 -9.53449681e-02 -2.98486531e-01 4.10079360e-01
1.11060873e-01 -1.84509605e-01 6.37212336e-01 7.94749796e-01
-1.65177479e-01 6.78375483e-01 -1.22497058e+00 1.30420864e+00
1.65244609e-01 -1.86009324e+00 7.75341928e-01 -2.14798823e-01
7.47878850e-01 -9.10991073e-01 3.04966033e-01 3.94895487e-02
2.03944430e-01 -1.27930069e+00 8.95852029e-01 5.62753916e-01
1.48516524e+00 -8.40676665e-01 1.78583309e-01 -1.62188604e-01
-1.33133566e+00 5.13292216e-02 -4.42344576e-01 -2.83842031e-02
5.35000749e-02 4.79139313e-02 -4.69295859e-01 4.55034852e-01
1.34305492e-01 8.38971853e-01 -3.19227546e-01 1.22421838e-01
-5.45030475e-01 1.05019167e-01 9.64074358e-02 -9.09077823e-02
2.84754008e-01 -3.78916383e-01 2.17689157e-01 9.34334695e-01
3.23556930e-01 1.24445878e-01 -2.28879377e-02 1.48757946e+00
-4.60452348e-01 8.30679312e-02 -1.06887412e+00 -4.81052041e-01
5.98171651e-01 1.17923570e+00 -4.82591540e-01 -5.20219088e-01
-3.87737960e-01 1.38432741e+00 3.37344408e-01 5.33985674e-01
-8.19421053e-01 -5.45834780e-01 1.13920736e+00 -3.91125679e-04
2.91329592e-01 -2.35380158e-01 -4.82044339e-01 -1.21585989e+00
3.71402383e-01 -5.65968156e-01 -2.62006104e-01 -8.86484563e-01
-1.47617638e+00 5.77209294e-01 -1.78512186e-01 -1.02827442e+00
-2.48875543e-02 -6.65546477e-01 -6.97172284e-01 8.18975747e-01
-1.44084549e+00 -1.67113376e+00 -2.55915999e-01 4.86112654e-01
6.56594872e-01 -4.14272338e-01 9.82648671e-01 7.49540702e-02
-1.66868687e-01 8.29809308e-01 -2.71091938e-01 3.52350265e-01
4.24093515e-01 -9.39306080e-01 7.63250709e-01 4.05596524e-01
4.40528959e-01 7.11126804e-01 2.70673007e-01 -4.66552824e-01
-1.80013800e+00 -1.18721128e+00 8.04106355e-01 -3.96432966e-01
5.17453313e-01 -6.99294150e-01 -6.74140930e-01 5.77896655e-01
1.89262509e-01 1.07918970e-01 5.54332495e-01 -3.92366827e-01
-9.50255811e-01 -1.96040109e-01 -1.03376877e+00 1.06243932e+00
1.54230416e+00 -8.46380532e-01 -6.58404648e-01 1.19842164e-01
7.43248522e-01 -1.33417293e-01 -7.75189102e-01 2.83889741e-01
9.95734274e-01 -6.33944035e-01 1.38157046e+00 -8.62092733e-01
8.46036732e-01 -2.22028330e-01 -2.13347703e-01 -1.11837578e+00
-1.01855107e-01 -5.93375564e-01 9.88141075e-02 1.19175899e+00
5.75899407e-02 -4.86802518e-01 8.08131993e-01 6.14293575e-01
3.18712562e-01 -7.44293928e-01 -6.15203559e-01 -8.41078758e-01
1.41544668e-02 -1.99602053e-01 1.21236110e+00 8.84494603e-01
-2.91677266e-01 2.38213927e-01 -6.41818762e-01 -2.00785547e-01
6.49809718e-01 7.75159180e-01 9.26002800e-01 -1.00999510e+00
-1.34045810e-01 -6.91371560e-01 -4.59885180e-01 -1.05456448e+00
6.59744799e-01 -1.03755844e+00 -2.25331753e-01 -1.31192970e+00
1.80349290e-01 -5.17655373e-01 -4.74680029e-02 5.03237963e-01
-2.37172216e-01 5.27069092e-01 5.30220330e-01 1.06467642e-01
-2.10320443e-01 9.87732351e-01 1.66978633e+00 -2.18309045e-01
1.71769291e-01 -4.26519781e-01 -8.24030340e-01 4.95903134e-01
5.32351553e-01 -1.88966751e-01 -7.03072488e-01 -6.40183747e-01
4.05148463e-03 1.99687377e-01 6.44341886e-01 -4.83668357e-01
-1.91769645e-01 -3.44833225e-01 5.42036176e-01 -6.65973574e-02
5.73605835e-01 -9.13487136e-01 2.28458449e-01 2.95035720e-01
-6.87723160e-01 -1.16558727e-02 -1.96674820e-02 5.44463813e-01
-3.23894322e-01 1.88486397e-01 7.66210616e-01 3.14960070e-02
-5.55866599e-01 6.64818048e-01 2.60187924e-01 -4.71778959e-02
8.76191139e-01 -2.80347377e-01 4.36891243e-02 -5.59445202e-01
-3.89360756e-01 -1.30762374e-02 7.20745027e-01 9.57716107e-01
1.14614487e+00 -2.08218193e+00 -5.88781655e-01 8.33529115e-01
5.00085831e-01 -3.76419038e-01 1.35926262e-01 6.86032027e-02
-4.27038610e-01 3.53925645e-01 -4.57004517e-01 -1.82445765e-01
-9.64113593e-01 6.50628746e-01 -4.55872342e-02 3.36778671e-01
-1.04012370e+00 6.82129204e-01 5.43127775e-01 -4.06717151e-01
2.42003053e-01 -3.32009137e-01 3.29688311e-01 -3.18827868e-01
6.30187869e-01 3.54642458e-02 -5.74638367e-01 -4.86891657e-01
-6.24631569e-02 8.89490306e-01 2.61863738e-01 -2.85142571e-01
1.16247475e+00 2.13578075e-01 7.19946995e-02 2.15345562e-01
1.24693716e+00 -1.99587625e-02 -1.59054625e+00 -3.61689031e-01
-5.12936227e-02 -9.65062499e-01 -2.89372802e-01 -5.07757902e-01
-1.34511399e+00 1.06080103e+00 2.06458196e-01 -2.68459022e-01
1.03181887e+00 -6.73625469e-02 1.07755125e+00 2.76098043e-01
3.81794870e-01 -8.77475917e-01 4.45299327e-01 3.26809131e-05
1.37657905e+00 -1.07369685e+00 -3.06316733e-01 -4.37744021e-01
-9.14838433e-01 1.20452249e+00 4.20137346e-01 -4.57536161e-01
4.05624866e-01 9.18144360e-02 -2.23452915e-02 -1.51302129e-01
-6.06277525e-01 4.33013700e-02 3.58386904e-01 4.83205736e-01
2.23845005e-01 2.09004834e-01 3.75825986e-02 6.59025073e-01
-1.22336157e-01 1.68766424e-01 1.00779995e-01 6.96127951e-01
-1.30670428e-01 -1.41128409e+00 -4.48080078e-02 8.38451013e-02
4.75433841e-02 -2.45663337e-02 -6.18076801e-01 7.97690392e-01
-1.07282557e-01 4.13347423e-01 -1.13142645e-02 -2.25352034e-01
4.44573283e-01 2.15698883e-01 7.76402175e-01 -4.41362411e-01
-2.38461539e-01 -2.78307885e-01 -3.25256199e-01 -6.76970780e-01
-1.38068244e-01 -2.10753798e-01 -1.31005347e+00 -2.25582719e-01
9.30662230e-02 9.21819545e-03 7.09976494e-01 4.98112768e-01
4.88701195e-01 3.62650812e-01 8.11413705e-01 -7.54660785e-01
-6.10476494e-01 -4.08646643e-01 -6.65662050e-01 8.76597941e-01
2.38423273e-01 -7.27695167e-01 -5.81289977e-02 1.83689639e-01] | [11.865050315856934, 0.18135066330432892] |
2937c15b-cee2-405f-88cd-36fc84b27b89 | graph-sparsification-for-gcn-towards-optimal | 2306.01725 | null | https://arxiv.org/abs/2306.01725v1 | https://arxiv.org/pdf/2306.01725v1.pdf | Graph Sparsification for GCN Towards Optimal Crop Yield Predictions | In agronomics, predicting crop yield at a per field/county granularity is important for farmers to minimize uncertainty and plan seeding for the next crop cycle. While state-of-the-art prediction techniques employ graph convolutional nets (GCN) to predict future crop yields given relevant features and crop yields of previous years, a dense underlying graph kernel requires long training and execution time. In this paper, we propose a graph sparsification method based on the Fiedler number to remove edges from a complete graph kernel, in order to lower the complexity of GCN training/execution. Specifically, we first show that greedily removing an edge at a time that induces the minimal change in the second eigenvalue leads to a sparse graph with good GCN performance. We then propose a fast method to choose an edge for removal per iteration based on an eigenvalue perturbation theorem. Experiments show that our Fiedler-based method produces a sparse graph with good GCN performance compared to other graph sparsification schemes in crop yield prediction. | ['Tim Eadie', 'Gene Cheung', 'Saghar Bagheri'] | 2023-06-02 | null | null | null | null | ['crop-yield-prediction', 'crop-yield-prediction'] | ['computer-vision', 'miscellaneous'] | [ 2.28819221e-01 3.72839272e-01 -3.59060735e-01 -3.34885009e-02
3.20832968e-01 -4.87954050e-01 -1.19844317e-01 7.38004744e-01
1.56793237e-01 5.12222350e-01 -5.90888448e-02 -7.52548158e-01
-3.96634161e-01 -1.54413247e+00 -1.00128937e+00 -6.62419260e-01
-7.07597375e-01 2.75221048e-03 -4.67564762e-02 -4.39326286e-01
-1.25991195e-01 7.65141010e-01 -1.26822448e+00 -1.96461409e-01
9.64196980e-01 8.40649128e-01 4.26487088e-01 7.61125386e-01
1.01689391e-01 4.02556509e-01 3.56715769e-02 -2.79684484e-01
5.77481270e-01 -2.75627434e-01 -4.94809598e-01 1.37403056e-01
1.41892031e-01 -4.93216991e-01 -4.62503940e-01 1.27787340e+00
3.19015324e-01 7.95916319e-02 7.09037542e-01 -1.18982935e+00
-6.22055173e-01 1.14884663e+00 -1.00477016e+00 -2.58635223e-01
-1.09016977e-01 -1.06075957e-01 9.88705397e-01 -4.83910054e-01
5.09116232e-01 9.46267188e-01 9.57059979e-01 -2.18114167e-01
-1.42502797e+00 -7.27848172e-01 3.20362270e-01 -9.83394459e-02
-1.34187186e+00 -1.07947998e-02 7.54451931e-01 -2.08737448e-01
8.71018887e-01 1.90987751e-01 9.95306313e-01 2.20399171e-01
4.66824800e-01 4.52063471e-01 2.75420517e-01 -4.37083930e-01
1.21767789e-01 -6.98392630e-01 -7.73907900e-02 1.04988527e+00
7.96895087e-01 8.43267590e-02 -2.22936422e-01 -1.28540084e-01
1.09801638e+00 1.01316109e-01 -1.70589507e-01 -5.76574147e-01
-9.42050159e-01 1.03035283e+00 1.00500214e+00 1.50544673e-01
-6.88945651e-01 3.89111787e-01 2.26299793e-01 1.90468892e-01
5.88057637e-01 4.63189483e-01 -5.10588109e-01 4.56006557e-01
-1.06467295e+00 3.28896075e-01 1.12684703e+00 1.09535682e+00
1.08448982e+00 2.03883216e-01 6.55216798e-02 4.42453682e-01
-1.77248418e-02 7.90063083e-01 -1.73927560e-01 -6.57169223e-01
4.23825741e-01 9.13729012e-01 -1.38206080e-01 -1.78367627e+00
-8.05239797e-01 -3.08589429e-01 -1.69207847e+00 -3.25270258e-02
3.18793833e-01 -4.81625080e-01 -1.00999558e+00 1.62126124e+00
2.94845551e-01 2.95289487e-01 3.33746895e-02 7.50655413e-01
6.29845679e-01 8.79740953e-01 6.93630949e-02 -5.59607893e-02
9.94525313e-01 -4.56312597e-01 -4.72907901e-01 -2.59134680e-01
1.11236644e+00 -4.42056268e-01 5.74073315e-01 3.99099998e-02
-7.03037858e-01 -2.19788879e-01 -1.02974188e+00 3.14065486e-01
-4.17119652e-01 5.46716392e-01 1.25645947e+00 3.97235394e-01
-1.07610786e+00 1.00258219e+00 -1.04095471e+00 -5.19890368e-01
3.04614931e-01 4.19522762e-01 -4.98655081e-01 3.24612744e-02
-1.05457020e+00 8.11081171e-01 8.24624240e-01 6.28541470e-01
-6.10914290e-01 -9.07450974e-01 -1.07188475e+00 7.50354886e-01
5.02419889e-01 -2.41791666e-01 5.53239763e-01 -7.59971976e-01
-1.26222420e+00 4.08471614e-01 7.76285753e-02 -8.52045000e-01
-6.45776317e-02 -7.04664588e-02 -1.35018885e-01 3.35415490e-02
-1.16420470e-01 4.95182455e-01 6.85048044e-01 -8.91879439e-01
-4.67070431e-01 -3.54426682e-01 -5.38671296e-03 9.17975008e-02
-3.94605696e-01 -4.95110273e-01 9.74984393e-02 -6.91358745e-01
5.34622073e-01 -1.09099424e+00 -7.53799021e-01 -1.16037026e-01
-5.14057338e-01 1.82992399e-01 5.95467091e-01 -7.42257237e-01
1.27753520e+00 -1.94781017e+00 1.26554206e-01 6.38071954e-01
5.95233679e-01 2.70254374e-01 -4.37398374e-01 6.27446115e-01
-1.11170158e-01 -3.95762920e-02 -3.25736582e-01 5.36814332e-01
-1.74403593e-01 1.16305396e-01 -2.59675115e-01 4.74382430e-01
4.75679606e-01 9.39517677e-01 -1.07736576e+00 1.05085957e-03
2.30601117e-01 2.96522677e-01 -4.96499479e-01 8.48930702e-02
-1.58047780e-01 -3.15977037e-01 -5.78364670e-01 5.71034551e-01
1.13822353e+00 -4.74831045e-01 6.92257643e-01 -3.40990990e-01
-1.52995065e-01 -3.36485803e-01 -1.20369959e+00 1.14229691e+00
-3.73706728e-01 3.92422259e-01 2.40539268e-01 -1.40173900e+00
1.28256488e+00 -8.42452142e-03 5.86575389e-01 3.39502953e-02
-5.99777624e-02 2.24029541e-01 3.67059484e-02 1.98021561e-01
5.55819988e-01 4.99523193e-01 9.38641503e-02 2.32377753e-01
1.31695792e-02 -1.53491437e-01 6.00611091e-01 3.33572835e-01
1.12718844e+00 1.66194111e-01 5.36736071e-01 -8.17031741e-01
6.90968707e-02 6.15309551e-02 5.37544847e-01 6.42730415e-01
6.61481619e-02 1.83547571e-01 1.03678179e+00 -7.60166824e-01
-9.41612303e-01 -5.54984927e-01 4.02166218e-01 1.06138897e+00
-4.14440148e-02 -3.43599647e-01 -5.20043850e-01 -3.38266373e-01
4.31854397e-01 4.67922688e-01 -6.56376243e-01 -3.06092381e-01
-3.11258793e-01 -1.00950038e+00 2.57028669e-01 5.53753018e-01
4.99716997e-01 -1.02021277e+00 -5.30495942e-01 4.61470664e-01
1.68181926e-01 -8.95952642e-01 -3.35903406e-01 6.70588374e-01
-8.80521536e-01 -8.86454523e-01 -7.34567106e-01 -7.85112977e-01
8.77401531e-01 5.59048414e-01 1.10040843e+00 2.54418582e-01
-1.14820473e-01 -3.46323341e-01 -6.27049267e-01 -4.55912530e-01
-2.92278618e-01 7.49143600e-01 -1.53036624e-01 -3.76919478e-01
3.78486663e-01 -7.91277170e-01 -5.74483573e-01 -2.69121051e-01
-8.26062679e-01 3.58787864e-01 7.45918393e-01 7.75256813e-01
5.11507869e-01 4.81573403e-01 4.75054741e-01 -8.44727695e-01
3.44518751e-01 -5.93185365e-01 -1.27180350e+00 5.33893168e-01
-6.27932787e-01 3.10096532e-01 8.65177035e-01 -3.51541728e-01
-4.94167358e-01 7.61211097e-01 4.50641870e-01 -1.71641573e-01
1.46013007e-01 1.18578148e+00 1.00090325e-01 -5.05368471e-01
5.37654102e-01 3.75648551e-02 1.39991501e-02 -1.03557803e-01
4.91636395e-01 -6.68823421e-02 3.23999912e-01 -8.10913444e-02
8.85421813e-01 1.82461366e-02 7.15843678e-01 -8.88239443e-01
-7.78251112e-01 -2.63190746e-01 -7.71342933e-01 -1.87579602e-01
4.55586553e-01 -9.96383548e-01 -9.77379262e-01 4.01267380e-01
-1.03441978e+00 -6.38132095e-01 5.34530021e-02 6.32312059e-01
-2.45761633e-01 1.76638424e-01 -5.00839531e-01 -4.73762184e-01
-6.53738499e-01 -7.78658807e-01 8.17322671e-01 2.44306460e-01
1.81201324e-01 -9.73029315e-01 -1.33156016e-01 -6.02439642e-01
3.92640829e-01 6.01870298e-01 9.91496861e-01 -1.54337615e-01
-4.41963673e-01 -3.03750157e-01 -8.09372842e-01 -1.09813727e-01
3.48753363e-01 1.57823756e-01 -3.08161557e-01 -4.23559129e-01
-6.81763649e-01 -8.15026369e-03 9.54187810e-01 1.04154599e+00
1.05907226e+00 -3.65690529e-01 -3.99158001e-01 9.13325012e-01
1.66401660e+00 -1.81108236e-01 3.96533638e-01 -8.02565962e-02
1.09317744e+00 3.39530528e-01 3.98531765e-01 6.61721289e-01
2.35461578e-01 -1.02006318e-02 6.56642139e-01 -4.61053759e-01
1.63385525e-01 -3.46260041e-01 1.96375996e-01 7.21435249e-01
-2.51215786e-01 -4.24946845e-01 -1.01097393e+00 7.32734323e-01
-2.13845325e+00 -7.82983243e-01 -1.63789243e-01 2.13027024e+00
6.62159681e-01 1.38617745e-02 -1.25841245e-01 9.27427132e-03
7.24864483e-01 3.41941059e-01 -5.70129037e-01 -3.93666863e-01
-1.90982729e-01 3.75126958e-01 1.68746066e+00 4.84875500e-01
-1.02164149e+00 1.41012037e+00 6.00568771e+00 7.41906881e-01
-1.05236876e+00 -7.12695479e-01 6.09058619e-01 4.14416403e-01
-2.17946604e-01 2.68528938e-01 -8.22228909e-01 -2.03207403e-01
6.82099283e-01 -3.30081969e-01 6.53771996e-01 9.51433063e-01
2.30959520e-01 -1.49796665e-01 -6.66152000e-01 4.27314192e-01
-3.59373093e-01 -1.45276237e+00 1.53757855e-01 -3.04834694e-02
8.69319856e-01 -1.00396737e-01 -3.89724255e-01 1.77118555e-02
9.85266685e-01 -1.03949797e+00 8.24896097e-02 3.29269201e-01
6.44634187e-01 -1.05432594e+00 6.46356285e-01 1.56756550e-01
-2.01204395e+00 1.88632861e-01 -9.34943676e-01 -2.26112276e-01
-1.19672209e-01 1.17966723e+00 -9.72116292e-01 9.46267605e-01
5.38202643e-01 9.32845950e-01 -4.72595274e-01 7.77389467e-01
-3.97871174e-02 9.07776177e-01 -6.13978565e-01 -1.48941442e-01
4.88870233e-01 -5.94556808e-01 2.60559976e-01 1.09416628e+00
7.61274755e-01 2.68200845e-01 4.11987007e-01 6.93745732e-01
-1.26848042e-01 2.80021042e-01 -8.93758714e-01 -5.05537152e-01
5.02746046e-01 1.39468551e+00 -1.10063088e+00 2.76521128e-02
-2.91167408e-01 7.33267605e-01 5.11841357e-01 3.90681893e-01
-3.59702587e-01 -7.58337915e-01 3.08130175e-01 1.15749026e-02
5.97454667e-01 -3.71327102e-01 -2.45701283e-01 -8.90625000e-01
-1.19956747e-01 -7.08740652e-01 1.26046494e-01 -5.86676598e-01
-8.95975828e-01 3.15070838e-01 -2.14783534e-01 -8.52845550e-01
-1.25947550e-01 -6.72041059e-01 -5.43276668e-01 9.38709378e-01
-1.51393497e+00 -1.34069800e+00 -5.85496843e-01 1.15589365e-01
-1.17180221e-01 -5.21518514e-02 1.08922875e+00 -1.65139318e-01
-5.82408547e-01 3.64324689e-01 9.50243697e-02 1.70787007e-01
1.85141921e-01 -1.05980301e+00 8.19571018e-01 1.11795783e+00
-1.38991058e-01 1.89395592e-01 7.14543283e-01 -1.07063961e+00
-1.64480674e+00 -1.35697567e+00 8.50801706e-01 5.45215845e-01
9.38785732e-01 -1.19270861e-01 -9.37129259e-01 8.11770618e-01
-1.01048946e-01 2.00629637e-01 3.76313068e-02 3.43035102e-01
-7.88959414e-02 -3.72921318e-01 -8.17496717e-01 6.10517085e-01
8.90198946e-01 5.95819391e-03 4.66999322e-01 4.96436059e-01
8.53266418e-01 -3.73553723e-01 -1.07540047e+00 7.22425938e-01
6.64174914e-01 -3.74466747e-01 6.58646166e-01 -4.56585705e-01
5.25213480e-01 -5.22979647e-02 -1.19657896e-01 -1.66188776e+00
-9.33995247e-01 -9.07172740e-01 -1.27894357e-01 7.60502517e-01
4.81238991e-01 -4.26126033e-01 9.84457493e-01 1.89466864e-01
2.55726367e-01 -7.02646077e-01 -2.07568914e-01 -6.89890027e-01
2.31453218e-03 3.58767919e-02 1.06686997e+00 8.47321391e-01
-4.65897806e-02 5.02997600e-02 -4.63670164e-01 7.67084658e-01
5.41698754e-01 3.65459919e-01 7.84029305e-01 -1.44472182e+00
7.15902895e-02 -4.31478769e-01 -6.08131349e-01 -7.06552327e-01
1.80831835e-01 -8.82005215e-01 1.40889501e-02 -1.57696772e+00
1.33904770e-01 -5.00538588e-01 1.46017037e-02 8.30261827e-01
-4.94475722e-01 -1.84793979e-01 7.55306035e-02 -4.05439734e-01
1.26895934e-01 3.68585497e-01 1.18965292e+00 -1.28583517e-02
-6.72076285e-01 8.39588419e-02 -5.41255534e-01 4.95120525e-01
9.98670340e-01 -3.99674237e-01 -4.60798234e-01 -3.99532199e-01
7.85261869e-01 3.01288664e-01 3.26522410e-01 -9.01701093e-01
-1.10407032e-01 -5.51147521e-01 3.50235909e-01 -9.07575369e-01
-3.47746015e-01 -7.36652672e-01 2.70686716e-01 7.27007329e-01
-2.48580098e-01 9.05019045e-02 5.19834816e-01 5.78580379e-01
2.09223792e-01 -2.46189028e-01 4.94327605e-01 6.05158322e-02
-6.29995525e-01 7.75516391e-01 -2.27303445e-01 -5.01517296e-01
7.92979598e-01 2.37443537e-01 -1.76082570e-02 -4.23074573e-01
-6.27336323e-01 4.65117961e-01 1.49360284e-01 1.86228380e-01
3.80728930e-01 -1.26618230e+00 -1.12797022e+00 2.52687752e-01
-8.06743056e-02 -7.91308284e-03 3.50113325e-02 5.39054453e-01
-1.12164247e+00 3.48065048e-01 -3.93763483e-01 -3.29190731e-01
-9.97582436e-01 4.29983258e-01 4.87978943e-02 -7.40209639e-01
-6.61258876e-01 8.06283653e-01 6.84348196e-02 -4.00170594e-01
-1.52616771e-02 -6.74472094e-01 -1.51090398e-01 -5.26956059e-02
1.36316180e-01 1.95504650e-01 2.31258154e-01 -2.75301576e-01
-1.05347447e-02 2.13772103e-01 7.85744563e-02 4.36663598e-01
1.65048623e+00 1.18728079e-01 -3.51429760e-01 1.45286947e-01
7.99905121e-01 -4.22977865e-01 -1.22851276e+00 -7.16418549e-02
-2.79292534e-03 -3.53795499e-01 4.61846292e-01 -4.01300937e-01
-1.58140934e+00 4.99688059e-01 5.72554231e-01 6.19453728e-01
1.37444687e+00 -6.21091545e-01 7.33359456e-01 7.56460845e-01
7.26871043e-02 -9.68241930e-01 -5.94598472e-01 5.76896191e-01
7.81306803e-01 -1.25305259e+00 4.85713422e-01 -7.53163636e-01
-2.99470395e-01 1.32014000e+00 4.51605976e-01 -4.16993827e-01
1.14798617e+00 4.33921337e-01 -3.54047865e-01 -2.28647202e-01
-3.89411360e-01 -4.52766895e-01 1.88626140e-01 5.99963129e-01
4.84962165e-01 8.31611454e-01 -1.63595840e-01 1.90511644e-01
-4.00604546e-01 3.70870717e-02 4.32793379e-01 7.34789610e-01
-6.43590569e-01 -8.34787369e-01 -5.18347658e-02 1.12770391e+00
-2.41484866e-02 -4.41531211e-01 -1.98419482e-01 6.18254006e-01
-3.35823357e-01 5.40509641e-01 -1.32412195e-01 -3.18777233e-01
8.28610584e-02 -2.98785448e-01 4.82651412e-01 -7.19984949e-01
-2.25093216e-01 -1.27581954e-01 -1.38528988e-01 -5.74206173e-01
-3.53681266e-01 -3.77164304e-01 -1.11941648e+00 -8.71398270e-01
-7.87300229e-01 -1.74179599e-01 4.88570899e-01 8.01258266e-01
5.74859679e-01 4.21728134e-01 6.64666176e-01 -1.07677841e+00
-3.75449866e-01 -9.76690233e-01 -1.20735192e+00 -1.50815263e-01
2.38243476e-01 -4.71912622e-01 -1.96162552e-01 5.61362505e-02] | [9.34361457824707, -1.5169011354446411] |
1249eebf-3955-4f7c-b15d-0bc90c42a08e | towards-designing-a-chatgpt-conversational | 2304.09866 | null | https://arxiv.org/abs/2304.09866v1 | https://arxiv.org/pdf/2304.09866v1.pdf | Towards Designing a ChatGPT Conversational Companion for Elderly People | Loneliness and social isolation are serious and widespread problems among older people, affecting their physical and mental health, quality of life, and longevity. In this paper, we propose a ChatGPT-based conversational companion system for elderly people. The system is designed to provide companionship and help reduce feelings of loneliness and social isolation. The system was evaluated with a preliminary study. The results showed that the system was able to generate responses that were relevant to the created elderly personas. However, it is essential to acknowledge the limitations of ChatGPT, such as potential biases and misinformation, and to consider the ethical implications of using AI-based companionship for the elderly, including privacy concerns. | ['Hend Al-Khalifa', 'Abeer Alessa'] | 2023-04-18 | null | null | null | null | ['misinformation'] | ['miscellaneous'] | [-3.50311697e-01 9.35697556e-01 1.19020678e-01 -3.43431979e-01
2.49294713e-01 8.58906135e-02 1.42348632e-01 1.25054181e-01
-6.87241912e-01 1.67638814e+00 8.16082656e-01 -7.74201155e-02
1.17456436e-01 -5.51300168e-01 2.25553617e-01 -1.79911748e-01
-1.06709965e-01 9.48040411e-02 -1.01832211e-01 -5.85354090e-01
1.32895678e-01 2.09707528e-01 -1.53698003e+00 1.67953014e-01
1.11321855e+00 2.70511776e-01 5.46032414e-02 4.63707179e-01
2.91539520e-01 1.07245851e+00 -1.04175186e+00 -4.62200642e-01
-5.23526788e-01 -4.34642851e-01 -8.04076493e-01 -4.56970006e-01
-1.67763263e-01 -1.24413133e+00 -3.47688884e-01 4.66581762e-01
1.12654400e+00 1.80601507e-01 3.53663564e-01 -1.64446294e+00
-4.91539419e-01 3.97804916e-01 4.62114781e-01 -9.74712614e-03
1.13705122e+00 -2.10394673e-02 8.24798197e-02 -5.79622626e-01
3.97747427e-01 1.42434347e+00 1.01330841e+00 1.22876287e+00
-8.60516846e-01 -7.41554499e-01 -1.16504319e-01 1.60848141e-01
-1.17490613e+00 -9.39860821e-01 2.38469616e-01 -1.10344507e-01
1.00625825e+00 3.44465822e-01 1.29300439e+00 1.41280115e+00
3.25387120e-01 5.16514540e-01 7.54224300e-01 -5.14676154e-01
6.81514740e-01 6.26249433e-01 3.87366265e-01 3.10987816e-03
7.29958415e-01 -3.64054620e-01 -4.69646722e-01 -8.45051944e-01
3.42389822e-01 1.07750192e-01 -3.69670451e-01 3.54958773e-02
-9.20573890e-01 8.34650636e-01 2.76207507e-01 4.06961650e-01
-5.30549288e-01 -3.01236808e-01 4.42532450e-01 2.73046732e-01
4.25452203e-01 1.21336840e-01 -1.57492533e-01 -7.47485816e-01
5.12502231e-02 1.54610634e-01 1.49939215e+00 9.98619854e-01
8.25514421e-02 -4.36132580e-01 -2.55707532e-01 8.52933228e-01
6.26147807e-01 3.80852431e-01 4.38744783e-01 -1.09736657e+00
6.69280142e-02 4.89265233e-01 2.24168822e-01 -1.01396227e+00
-5.66121519e-01 4.52557087e-01 -5.05552828e-01 2.98826456e-01
3.14524382e-01 -9.19341505e-01 1.89200416e-02 1.61241972e+00
1.58869058e-01 -7.51405239e-01 4.49281126e-01 4.96185511e-01
1.21757591e+00 -1.03406876e-01 3.71373445e-01 -4.79304999e-01
1.24778914e+00 -4.37795937e-01 -1.39081931e+00 -9.37534392e-01
8.82705688e-01 -2.62457520e-01 9.88934457e-01 -1.92631930e-01
-9.26195383e-01 9.26129073e-02 -1.04078138e+00 8.61548632e-02
-1.66780740e-01 -1.00090213e-01 5.64880610e-01 1.33179641e+00
-1.38332248e+00 5.46178281e-01 -3.30718547e-01 -1.62691700e+00
3.86940658e-01 4.59077418e-01 -4.70983595e-01 1.19710289e-01
-1.39272761e+00 1.42021668e+00 -4.31384519e-02 1.93817422e-01
3.39750290e-01 -8.58387575e-02 -8.12845588e-01 -2.92958796e-01
-4.60571349e-01 -8.67779255e-01 1.21149075e+00 -6.30997479e-01
-1.50201094e+00 8.17974389e-01 -1.61973629e-02 -3.98653805e-01
6.65969193e-01 -6.20774150e-01 -4.34592307e-01 2.96994209e-01
5.26572227e-01 7.18201458e-01 -1.37284592e-01 -7.72844434e-01
-1.79946557e-01 -4.73369747e-01 -1.00383617e-01 6.28061116e-01
-6.98879600e-01 4.96907353e-01 4.19170171e-01 -1.84481665e-01
-2.39093170e-01 -7.12342143e-01 -1.56362861e-01 6.68748319e-01
4.57561910e-02 8.73708576e-02 9.50810492e-01 -1.08949614e+00
1.20802355e+00 -2.20470095e+00 -5.91994584e-01 -2.75717735e-01
3.07112962e-01 3.24726045e-01 4.72980022e-01 8.30154955e-01
4.82291698e-01 1.42585218e-01 2.62299448e-01 -5.01691580e-01
-2.03919619e-01 8.25921595e-02 6.09820187e-01 3.64319056e-01
-2.24615574e-01 5.89038014e-01 -9.83941972e-01 -7.56457627e-01
1.77654862e-01 6.95220530e-01 -9.09141228e-02 3.36967319e-01
4.13361281e-01 2.51936466e-01 -3.93356174e-01 4.04501945e-01
5.31111956e-01 3.56762916e-01 3.34445149e-01 4.17596579e-01
-8.55205357e-02 3.78621906e-01 -3.37050766e-01 1.03013015e+00
1.30787835e-01 5.23111641e-01 2.56792605e-01 -9.59955007e-02
1.16737282e+00 5.18655598e-01 2.09162399e-01 -3.76236945e-01
6.92332506e-01 2.35764787e-01 -9.75702831e-04 -1.11837530e+00
4.27898318e-01 -1.30187869e-01 8.02955180e-02 6.64571702e-01
-6.56700969e-01 1.69798270e-01 -4.54805404e-01 3.82588893e-01
1.05688393e+00 -9.69448984e-02 7.04701483e-01 -4.19900715e-01
2.10950196e-01 -3.07859570e-01 3.60678405e-01 4.76613164e-01
-8.98244441e-01 3.25253874e-01 3.24290425e-01 -1.45739600e-01
-6.07754409e-01 -5.21289527e-01 2.30048701e-01 5.41460514e-01
-1.43979877e-01 -5.35034001e-01 -8.80018294e-01 -4.14052784e-01
-6.25361083e-03 1.11428511e+00 -3.55126947e-01 -7.02118218e-01
1.45265078e-02 -3.71477008e-01 7.03262806e-01 3.52866709e-01
9.80215251e-01 -1.36723030e+00 -1.03488183e+00 2.67616540e-01
-8.53764415e-01 -8.79121006e-01 -3.31371605e-01 -1.83753505e-01
-1.14142776e+00 -7.26855397e-01 -1.05540955e+00 -1.00696957e+00
6.25733733e-01 5.30968010e-01 4.02647465e-01 2.87264943e-01
6.40559420e-02 6.50882959e-01 -5.57840288e-01 -4.27823395e-01
-4.37444359e-01 -3.27932201e-02 2.94608057e-01 -4.64317441e-01
6.94310546e-01 -9.08260167e-01 -6.69379234e-01 5.59476256e-01
-2.54097749e-02 -1.70238167e-01 8.46469402e-02 4.50735390e-01
-8.55473697e-01 -5.32514036e-01 1.11703300e+00 -4.65599179e-01
1.23568022e+00 -6.04514539e-01 1.00646448e+00 3.93501073e-02
-4.08536851e-01 -4.75878447e-01 5.74676692e-02 -6.46652818e-01
-1.11465406e+00 -2.61394650e-01 -1.00582637e-01 7.26601303e-01
-1.87132031e-01 8.18238705e-02 -6.39641643e-01 -2.32767642e-01
8.67318630e-01 -4.44173664e-01 8.88050258e-01 -3.38021338e-01
-3.23313922e-01 1.63844538e+00 2.02983111e-01 -1.43019602e-01
-2.32739002e-01 1.06644712e-01 -6.62553787e-01 -1.41825223e+00
-1.00195529e-02 -6.12214133e-02 -3.66329253e-01 -8.44029725e-01
6.41852498e-01 -9.23399806e-01 -1.15163350e+00 8.44427168e-01
-9.95228529e-01 -6.36940897e-01 -3.20861675e-02 5.74536383e-01
-4.15587902e-01 4.27399397e-01 -4.91218984e-01 -1.47090089e+00
-7.38538206e-01 -1.07442968e-01 1.50680631e-01 4.85812277e-01
-1.39335954e+00 -7.84950495e-01 -7.28343874e-02 5.44893324e-01
6.98505461e-01 5.40270865e-01 4.59217608e-01 -4.00138438e-01
4.98118430e-01 -4.99237925e-01 -1.02631606e-01 1.05822869e-01
5.62165260e-01 -5.44058144e-01 -8.09217751e-01 -3.47543359e-01
3.57882112e-01 -6.46678448e-01 -3.45154226e-01 1.01646066e-01
-2.61890262e-01 -6.85402572e-01 -6.84578300e-01 -2.96397209e-01
3.78946990e-01 4.37445283e-01 1.36010933e+00 6.52312934e-01
1.54739007e-01 1.08547795e+00 8.51316631e-01 7.45055318e-01
8.05970669e-01 3.23822051e-01 -1.90750007e-02 1.56404346e-01
3.86198759e-02 -2.61920244e-01 6.60661578e-01 2.83069350e-02
3.49931419e-02 2.16246042e-02 -5.56966662e-01 4.12489712e-01
-2.10222411e+00 -7.88617432e-01 -2.66908228e-01 2.19643831e+00
5.24527013e-01 -1.17992319e-01 6.58611059e-01 3.61930937e-01
9.02018666e-01 -3.85533035e-01 -5.36923468e-01 -4.78422791e-01
-1.56485915e-01 -3.35453689e-01 1.69407457e-01 3.07678372e-01
-4.57401276e-01 5.26975513e-01 7.36606979e+00 -6.09649599e-01
-3.90233040e-01 -8.32498744e-02 5.89215994e-01 1.79375410e-01
3.26484777e-02 -1.26288002e-02 -4.29760128e-01 3.56225222e-01
8.92208099e-01 -5.28804123e-01 1.45362347e-01 7.44107664e-01
6.17880404e-01 -7.47339725e-01 -9.27721024e-01 7.82283723e-01
4.06707913e-01 -3.50370526e-01 -7.27841020e-01 9.68820229e-02
-2.11415336e-01 -5.43550313e-01 -6.91055179e-01 6.87096566e-02
9.04433727e-02 -6.60518348e-01 6.95141315e-01 6.33096814e-01
4.89067495e-01 -5.78027487e-01 9.65611637e-01 2.98595577e-01
-5.19071043e-01 -1.79086953e-01 1.62909627e-02 -1.08890808e+00
4.01663125e-01 -8.74000639e-02 -9.61892128e-01 -3.49306464e-01
1.08113027e+00 5.29011607e-01 -5.39137304e-01 1.12754524e+00
-6.38775602e-02 5.91476336e-02 -5.52281976e-01 -1.03081334e+00
-6.51619494e-01 -1.02552660e-01 2.43977323e-01 5.78997493e-01
4.17809188e-01 4.56704736e-01 -6.28756702e-01 4.36593950e-01
5.51665068e-01 1.24608807e-01 -1.04695952e+00 2.63591949e-02
7.51087904e-01 7.07612574e-01 -4.63785589e-01 7.01353606e-03
-1.96433276e-01 1.26521087e+00 -1.66077930e-02 2.29208320e-01
-2.58696258e-01 -4.95073229e-01 7.01841950e-01 4.97585565e-01
-4.35021847e-01 2.88940519e-02 -6.46013021e-01 -7.53194809e-01
2.23612159e-01 -9.07398701e-01 2.99097419e-01 -9.17468429e-01
-8.94425869e-01 2.74222463e-01 3.02500129e-02 -1.08475029e+00
-1.66745931e-01 1.18370526e-01 -8.49708259e-01 5.89847207e-01
-1.94005698e-01 -8.24600995e-01 -6.41219914e-01 3.39454949e-01
-3.03701967e-01 4.76969257e-02 1.29141533e+00 2.88030297e-01
-4.13457841e-01 8.66222084e-01 -2.36719534e-01 -4.07045335e-01
1.01901400e+00 -5.04363418e-01 2.79618084e-01 2.69473642e-01
-1.59246933e+00 7.35027790e-01 9.75200057e-01 -1.10469174e+00
-9.50571954e-01 -3.95224333e-01 1.57777929e+00 -2.08346188e-01
1.92092821e-01 -4.21716392e-01 -4.32642937e-01 5.75445473e-01
1.13013968e-01 -9.33772564e-01 7.60819137e-01 1.86782941e-01
3.51765364e-01 7.91781619e-02 -2.11315227e+00 1.19491768e+00
1.27043986e+00 -5.92318892e-01 -5.77002108e-01 2.25658387e-01
6.03720248e-01 2.71934777e-01 -9.10503983e-01 -1.12596832e-01
1.06376183e+00 -1.06627059e+00 7.00282156e-01 1.62182182e-01
-1.15441643e-02 3.39321077e-01 1.98168591e-01 -8.62267673e-01
-1.62378311e-01 -8.68000627e-01 1.19432285e-01 1.61959672e+00
2.90037513e-01 -1.36653507e+00 7.81080127e-01 2.03918195e+00
1.58131495e-01 -9.43851471e-02 -8.44198823e-01 -4.78438407e-01
-1.61653623e-01 2.93540001e-01 4.48262572e-01 6.01403654e-01
1.45489335e+00 7.57817388e-01 -8.38999450e-01 -3.29298377e-01
3.97269189e-01 -1.20360839e+00 8.66761088e-01 -1.32976890e+00
2.02145338e-01 1.80312201e-01 -5.00312805e-01 -3.67731661e-01
-2.57387370e-01 1.81137532e-01 1.67589992e-01 -2.01784468e+00
1.16586469e-01 -1.68102369e-01 6.91741645e-01 4.65054154e-01
-1.21862911e-01 -3.08662206e-01 1.83500517e-02 -5.16221561e-02
-5.07217050e-02 7.32873380e-01 8.23742270e-01 3.06081027e-01
-7.87220955e-01 4.12017673e-01 -1.38257158e+00 3.53059590e-01
1.44992387e+00 -1.76131517e-01 -4.19920117e-01 1.44441620e-01
-5.87749295e-03 2.70056337e-01 4.62651849e-01 -1.19304419e+00
2.49328539e-01 2.72791684e-01 2.05752477e-01 -4.18880433e-01
7.42448628e-01 -7.40368724e-01 4.53411043e-01 9.03801262e-01
3.86448056e-02 -3.23444903e-01 1.48706302e-01 1.47512093e-01
5.45494318e-01 -2.89863437e-01 3.99904251e-01 1.48230165e-01
1.71100453e-01 -5.42078257e-01 -1.32313561e+00 -3.35114002e-01
1.17242682e+00 -6.47233546e-01 -5.46834826e-01 -1.23478985e+00
-6.90752566e-01 5.24816751e-01 9.38766241e-01 4.61751312e-01
8.71242106e-01 -1.31435919e+00 -2.35499471e-01 2.19939873e-02
3.86879832e-01 -5.70983052e-01 1.80741996e-01 6.78162992e-01
-5.76877654e-01 5.77071682e-02 -7.25112259e-01 4.28849012e-01
-1.71646488e+00 2.40309626e-01 2.80169100e-01 6.38834417e-01
-1.12027133e+00 2.63149381e-01 -4.94646966e-01 -2.97879845e-01
7.47113824e-01 -1.15112392e-02 -5.24035573e-01 3.29671502e-02
9.02428329e-01 8.84849489e-01 -5.02573669e-01 -5.69998920e-01
-5.83417535e-01 -1.87455356e-01 -9.62046087e-02 -4.81879264e-01
1.12392497e+00 -7.11128116e-01 -2.17977524e-01 6.19175196e-01
6.88409626e-01 -5.04187882e-01 -7.36082375e-01 2.56839663e-01
4.37677875e-02 -4.63795185e-01 -5.48884869e-01 -6.23701632e-01
-1.57286197e-01 3.42837751e-01 7.44475007e-01 2.77330369e-01
7.19513774e-01 -1.59825549e-01 1.02524483e+00 7.26648152e-01
6.46145046e-01 -1.18781102e+00 -1.65492371e-02 1.42968059e-01
1.16495240e+00 -1.00489140e+00 3.60519029e-02 -4.88193303e-01
-8.36973369e-01 8.70063007e-01 8.40302467e-01 3.53498548e-01
8.05734098e-01 1.08895518e-01 2.18441993e-01 1.58991650e-01
-3.78909171e-01 2.25034878e-01 -7.97276556e-01 1.38374054e+00
5.66726387e-01 2.34599367e-01 -1.09934819e+00 8.77078652e-01
-3.24040800e-01 6.06743395e-01 6.79480076e-01 1.58284271e+00
-5.86573482e-01 -1.10737419e+00 -6.19890153e-01 4.96636927e-01
1.21472152e-02 2.45113745e-01 -1.15931249e+00 4.53699261e-01
-2.82532692e-01 1.78317130e+00 -4.68157530e-01 -6.31533980e-01
4.18310314e-01 7.52452090e-02 2.85148695e-02 -3.48367065e-01
-7.73669183e-01 -5.97670853e-01 1.30786836e+00 -1.94205508e-01
-3.94685298e-01 -1.04902005e+00 -1.06479549e+00 -8.65417302e-01
-6.60313815e-02 1.25106648e-01 2.91302264e-01 8.84280562e-01
4.12346572e-01 -3.03039551e-01 3.41642469e-01 -7.31294096e-01
1.59839746e-02 -1.49118602e+00 -5.95897734e-01 -1.65517315e-01
9.45866555e-02 -2.90283978e-01 -5.58838189e-01 -4.95113522e-01] | [12.843080520629883, 7.7910566329956055] |
893c9d14-7ff5-48ad-8dba-1ed5834e14bf | probing-schema-linking-information-from-pre | null | null | https://openreview.net/forum?id=uKGVHs4EMy | https://openreview.net/pdf?id=uKGVHs4EMy | Probing Schema Linking Information from Pre-trained Language Models for Text-to-SQL Parsing | The importance of building text-to-SQL parsers which can be applied to new databases has long been acknowledged, and a critical step to achieve this goal is schema linking, i.e., properly recognizing mentions of unseen columns or tables when generating SQLs. In this work, we propose a novel framework to elicit relational structures from large-scale pre-trained language models (PLMs) via a probing procedure based on Poincaré distance metric, and use the induced relations to augment current graph-based parsers for better schema linking. Compared with commonly-used rule-based methods for schema linking, we found that probing relations can robustly capture semantic correspondences, even when surface forms of mentions and entities differ. Moreover, our probing procedure is entirely unsupervised and requires no additional parameters. Extensive experiments show that our framework outperforms strong baselines on three benchmarks, and sets new state-of-the-art performance on two of them. We empirically verify that our probing procedure can indeed find desired relational structures through qualitative analysis. For reproducibility, we will release our code and data upon the publication of this paper. | ['Anonymous'] | 2021-11-16 | null | null | null | acl-arr-november-2021-11 | ['text-to-sql'] | ['computer-code'] | [ 2.38538906e-01 6.36791706e-01 -4.65461403e-01 -4.82571989e-01
-1.17018330e+00 -1.04304779e+00 6.69781864e-01 7.53127217e-01
-5.49337603e-02 4.88911033e-01 3.43115032e-01 -5.58417618e-01
-1.24998674e-01 -1.30508590e+00 -1.13706529e+00 1.53989837e-01
-3.24003585e-02 8.97601604e-01 4.38595116e-01 -3.49883765e-01
1.53278425e-01 5.04514873e-01 -1.27888823e+00 5.90544999e-01
8.38132203e-01 6.46951139e-01 -2.66295075e-01 2.50328660e-01
-6.46855533e-01 7.05079734e-01 -2.40961999e-01 -1.10547924e+00
1.31267831e-01 -2.03666791e-01 -1.26510835e+00 -2.01917186e-01
4.96657729e-01 2.00443879e-01 -2.72194952e-01 1.09385335e+00
1.43609447e-02 -2.56918073e-01 3.79961044e-01 -1.03776586e+00
-7.22864747e-01 1.21916401e+00 -1.77065924e-01 3.47083025e-02
7.52876759e-01 -3.31598192e-01 1.68830907e+00 -9.59555387e-01
1.10170782e+00 1.27449584e+00 6.70038939e-01 3.94250035e-01
-1.41311002e+00 -4.81271595e-01 7.80247226e-02 3.75870652e-02
-1.30330586e+00 -6.30519629e-01 7.50547826e-01 -2.80355096e-01
1.08132958e+00 4.03607190e-01 4.83755507e-02 1.05371618e+00
-2.70088106e-01 6.21834517e-01 8.73716712e-01 -5.40277243e-01
-1.56690064e-03 3.78230780e-01 1.97661683e-01 9.11031306e-01
7.13776648e-01 -3.38017166e-01 -4.66426313e-01 -2.60852486e-01
5.65255582e-01 -5.04838467e-01 1.11622058e-01 -7.28244364e-01
-1.35698199e+00 7.58657932e-01 2.70327777e-01 3.46860975e-01
-9.17517319e-02 -1.68488264e-01 3.86607111e-01 2.02638745e-01
2.87116796e-01 7.66782701e-01 -5.66390514e-01 -1.20895907e-01
-6.47784948e-01 1.76221743e-01 1.18199337e+00 1.54016817e+00
8.72179627e-01 -7.33368874e-01 1.67865932e-01 7.04288125e-01
1.92475677e-01 2.75076598e-01 1.79080982e-02 -8.74238253e-01
9.99107718e-01 1.11484277e+00 -5.60799241e-02 -1.10261858e+00
-4.13971812e-01 -2.84334719e-01 -4.92837071e-01 -5.91560304e-01
4.67669517e-01 1.92341655e-01 -4.49115038e-01 1.61832607e+00
3.85938406e-01 -2.18014300e-01 3.47568989e-01 5.10574281e-01
9.81484175e-01 3.37461233e-01 1.29917055e-01 -1.87861938e-02
1.45039487e+00 -6.25956059e-01 -5.35459340e-01 -4.56196070e-01
1.06967664e+00 -6.24978602e-01 1.19317818e+00 4.03630063e-02
-1.18771625e+00 -3.88465017e-01 -9.18649316e-01 -2.94394255e-01
-6.94988012e-01 -7.53238425e-02 8.97171676e-01 5.02318621e-01
-7.51164913e-01 6.00335777e-01 -8.54392111e-01 -9.36980128e-01
3.53117943e-01 9.21699032e-02 -6.28995776e-01 -6.24616183e-02
-1.21266913e+00 8.11397076e-01 8.18263531e-01 -7.88755864e-02
-2.25466192e-01 -6.82194054e-01 -1.11834824e+00 2.08588075e-02
1.02599096e+00 -7.46349037e-01 1.18056583e+00 -1.84114605e-01
-1.06356263e+00 1.15284741e+00 -3.32021773e-01 -4.78971153e-01
2.24153057e-01 -2.41079882e-01 -5.73517382e-01 5.36682904e-02
3.89368594e-01 4.41688478e-01 -1.15909204e-01 -1.45952046e+00
-4.18063939e-01 -3.69199216e-01 2.70252943e-01 -6.83154538e-02
-2.80082315e-01 1.67815894e-01 -8.23596358e-01 -4.62740928e-01
5.65615177e-01 -6.91314876e-01 -7.52151385e-02 -2.88166583e-01
-9.28578317e-01 -4.17834491e-01 2.68827021e-01 -4.80751067e-01
1.38746905e+00 -1.81820655e+00 5.63370995e-02 4.61034656e-01
1.35391906e-01 1.07712619e-01 -7.47798458e-02 9.09433663e-01
-8.27788040e-02 5.84724963e-01 -4.29721415e-01 -1.77840222e-04
3.65375221e-01 4.12940413e-01 -5.97583473e-01 -1.35043427e-01
5.67270935e-01 1.23349369e+00 -9.08708274e-01 -8.49986911e-01
-2.24902421e-01 -2.59002168e-02 -5.30751407e-01 2.76759088e-01
-5.81852257e-01 -3.46991792e-02 -3.06378663e-01 7.63119638e-01
4.36591595e-01 -5.48648715e-01 7.37328291e-01 -4.92012233e-01
2.35618591e-01 1.01766491e+00 -1.33327246e+00 1.81909776e+00
-3.36896360e-01 2.91798767e-02 -4.25040603e-01 -9.42785501e-01
1.12564135e+00 -2.18141563e-02 2.47762501e-01 -7.17279911e-01
-3.44419241e-01 3.36401552e-01 -3.02356392e-01 -6.13154054e-01
4.57163572e-01 1.50250286e-01 -4.35593635e-01 3.60140920e-01
1.53020695e-01 -1.74891412e-01 5.58736324e-01 6.26583695e-01
1.29376769e+00 3.14044863e-01 3.74785960e-01 -1.03976041e-01
5.39968967e-01 3.27644378e-01 5.61701834e-01 7.18202710e-01
3.73553485e-01 3.63056898e-01 9.06820714e-01 -2.76402682e-01
-9.79763210e-01 -1.22132325e+00 -9.13940296e-02 9.52163875e-01
9.89158377e-02 -9.14126217e-01 -5.54970264e-01 -1.20832813e+00
6.15220033e-02 8.25975418e-01 -5.15651643e-01 1.64842606e-01
-8.71879935e-01 -5.23943365e-01 7.28502929e-01 7.38399923e-01
9.33040008e-02 -8.89754534e-01 -1.28286388e-02 2.77924240e-01
-3.20661932e-01 -1.62928295e+00 2.28423625e-03 2.38385156e-01
-8.14820647e-01 -1.41166139e+00 2.15942010e-01 -8.74066472e-01
6.77905679e-01 -5.44211790e-02 1.64498556e+00 1.47940710e-01
1.15530394e-01 2.24328816e-01 -2.83838958e-01 -1.41626507e-01
-6.87536061e-01 5.97517371e-01 -4.29544032e-01 -4.04645711e-01
5.90425313e-01 -4.73768175e-01 -8.98649842e-02 2.68422097e-01
-1.04723942e+00 2.75737792e-01 5.91217875e-01 4.37434047e-01
6.80488288e-01 -3.24601568e-02 4.14012879e-01 -1.81454682e+00
4.77190942e-01 -4.40964401e-01 -6.68372154e-01 7.49539137e-01
-7.30790138e-01 5.20286262e-01 6.22390866e-01 2.22233295e-01
-9.23149288e-01 8.43294114e-02 -8.68264958e-02 2.22841978e-01
-2.76604950e-01 9.23562944e-01 -4.82599795e-01 2.41613299e-01
6.19027913e-01 -1.39092073e-01 -4.14977282e-01 -6.16876543e-01
8.04225028e-01 3.64157796e-01 8.45331490e-01 -1.05528557e+00
1.21597588e+00 4.22714680e-01 1.15069114e-01 -2.15660229e-01
-1.03028679e+00 -4.12885278e-01 -1.04820335e+00 4.29523379e-01
5.55280149e-01 -8.56029570e-01 -5.05026758e-01 -1.38957798e-01
-1.27365041e+00 -1.44034296e-01 -2.03433454e-01 4.61523607e-02
-2.42813960e-01 4.19649661e-01 -7.03129649e-01 -3.77777845e-01
-8.26923177e-02 -8.00951004e-01 1.11307108e+00 3.62075842e-03
-4.52232182e-01 -1.14816427e+00 1.58595234e-01 5.61294079e-01
1.57295447e-02 2.27559060e-01 1.40996671e+00 -1.11979008e+00
-8.65190208e-01 -1.15837567e-01 -5.16735673e-01 -1.77148908e-01
1.98507726e-01 8.56501833e-02 -6.47762895e-01 1.42349169e-01
-5.38915932e-01 -3.74688804e-01 6.16928637e-01 -5.15766323e-01
9.73823488e-01 -4.37527955e-01 -5.60528815e-01 5.45446992e-01
1.41796255e+00 -9.91777424e-03 6.37469649e-01 5.45863867e-01
8.59211147e-01 8.87824059e-01 6.19973838e-01 2.20660982e-03
1.04711246e+00 6.78746998e-01 3.06070566e-01 1.44245261e-02
-9.53597277e-02 -8.32322538e-01 -3.62053812e-02 9.79197025e-01
3.41384947e-01 -1.13931887e-01 -1.19138789e+00 6.01374269e-01
-1.81631136e+00 -6.94519460e-01 -2.97879845e-01 2.03586364e+00
1.27514708e+00 4.85069007e-01 -1.98740605e-02 -5.05983382e-02
5.12937963e-01 -3.73594910e-02 -3.12235951e-01 -2.11629540e-01
-2.53153503e-01 4.62432563e-01 4.96439308e-01 5.19522905e-01
-1.07161021e+00 1.26974189e+00 5.97499514e+00 4.00096893e-01
-6.62200391e-01 -1.46730408e-01 2.94064432e-01 3.11200112e-01
-8.78041685e-01 4.19917136e-01 -1.03824186e+00 3.94466631e-02
9.08831596e-01 -2.54857153e-01 3.55887800e-01 7.37289131e-01
-1.73430905e-01 1.49672583e-01 -1.55751777e+00 6.01137996e-01
-1.14749588e-01 -1.59957659e+00 1.07550643e-01 -1.52076453e-01
3.69688213e-01 -8.74501318e-02 -3.77728105e-01 3.85113209e-01
6.87522411e-01 -8.83028448e-01 5.45175016e-01 4.55395788e-01
5.87200761e-01 -4.48802501e-01 6.63883388e-01 1.21756494e-02
-1.32262969e+00 3.35802883e-01 -2.84259558e-01 3.25688064e-01
2.99451575e-02 6.47200346e-01 -9.68805730e-01 1.06335545e+00
5.21991313e-01 7.61834562e-01 -1.12475693e+00 6.56885326e-01
-5.58514774e-01 5.15039504e-01 -2.70526916e-01 5.68822175e-02
-9.86034796e-02 -6.34036772e-03 2.60925025e-01 1.40243554e+00
1.69465691e-01 4.16739061e-02 1.20173194e-01 1.18614209e+00
-5.11231422e-01 2.68819094e-01 -7.92069912e-01 -3.50070119e-01
8.58064830e-01 1.31366491e+00 -6.87519968e-01 -4.30641770e-01
-5.85980713e-01 5.75143218e-01 7.22859502e-01 2.03335419e-01
-6.11598492e-01 -5.47746956e-01 3.31343740e-01 2.49991834e-01
3.83476824e-01 -2.56779581e-01 -5.09181082e-01 -1.38075924e+00
4.58337992e-01 -1.06351519e+00 7.72982180e-01 -7.64019310e-01
-1.33273911e+00 6.11616671e-01 8.07732791e-02 -8.01178634e-01
-2.69436121e-01 -4.30195093e-01 -2.97036588e-01 6.29729390e-01
-1.48225629e+00 -1.33443999e+00 -2.27284670e-01 3.15815210e-01
9.65475589e-02 1.34241283e-01 9.82274950e-01 3.46529722e-01
-6.52440369e-01 6.08802915e-01 -4.43178266e-01 6.81715846e-01
8.07098329e-01 -1.43682015e+00 9.26971018e-01 1.21693599e+00
8.38564217e-01 1.09084809e+00 6.12105310e-01 -8.24618936e-01
-1.73259819e+00 -1.20319235e+00 1.45020068e+00 -8.05376709e-01
1.05047989e+00 -7.73024857e-01 -1.34811318e+00 1.08337116e+00
1.31440222e-01 9.33123380e-03 6.73191845e-01 5.15928507e-01
-9.11630929e-01 -2.00186968e-01 -7.85294056e-01 5.24187863e-01
1.56201303e+00 -8.16405594e-01 -8.94922733e-01 5.42768657e-01
9.41014946e-01 -6.97072029e-01 -1.29049778e+00 6.45257175e-01
2.17583805e-01 -8.11416209e-01 8.97487521e-01 -9.69849467e-01
4.54564631e-01 -3.48091453e-01 -2.82686532e-01 -9.37343836e-01
-1.92076892e-01 -6.25965357e-01 -2.73638874e-01 1.75287795e+00
9.00441527e-01 -6.30325258e-01 8.67490768e-01 6.80793703e-01
-8.64819214e-02 -6.41091883e-01 -4.97278184e-01 -7.93057919e-01
5.65751605e-02 -5.27076840e-01 8.68311584e-01 1.14241672e+00
3.37527931e-01 4.93869931e-01 2.48344004e-01 5.98378956e-01
5.42670131e-01 4.10522699e-01 9.21932518e-01 -1.28552270e+00
-2.37512305e-01 -2.57145107e-01 -1.63180292e-01 -9.09066319e-01
3.41869324e-01 -1.26314604e+00 -1.64706409e-01 -1.93636227e+00
1.88870654e-01 -7.60282099e-01 -1.26093075e-01 7.35264778e-01
-2.46341735e-01 3.46424282e-02 -4.23461050e-02 2.08287224e-01
-7.53329933e-01 9.53685045e-02 7.52303183e-01 -1.37296036e-01
-7.32649714e-02 -2.38576472e-01 -1.20836651e+00 4.98286635e-01
6.07264996e-01 -7.17236638e-01 -4.09538627e-01 -4.30126905e-01
7.45134115e-01 -2.15345323e-02 2.89270073e-01 -7.64411449e-01
4.03318346e-01 -1.84599146e-01 -5.40452078e-02 -5.39407790e-01
-1.09613590e-01 -7.03868270e-01 1.91183507e-01 5.33216558e-02
-6.03312790e-01 3.09599340e-01 1.81160852e-01 3.96954685e-01
-2.88207412e-01 -2.10240260e-01 3.22926879e-01 -1.05170004e-01
-6.48178935e-01 8.19391906e-02 1.75001174e-01 6.64810240e-01
7.47418702e-01 2.74878234e-01 -6.72987521e-01 -1.44204810e-01
-4.86955822e-01 1.72722876e-01 5.60038209e-01 5.12552619e-01
3.52382302e-01 -1.31981778e+00 -5.34437537e-01 4.93405089e-02
5.47395051e-01 2.24292144e-01 -4.99900371e-01 6.00789189e-01
-4.40388888e-01 6.08630836e-01 2.46214122e-01 -3.89075130e-01
-1.16159034e+00 7.43712008e-01 4.19615768e-02 -6.01144552e-01
-4.46487159e-01 5.88238001e-01 -1.28169090e-01 -8.11813354e-01
1.92787215e-01 -6.53887928e-01 -8.53222460e-02 -1.18853614e-01
1.12284683e-01 -2.77707484e-02 4.54955190e-01 -4.27836418e-01
-5.92223942e-01 4.75760132e-01 -1.50564715e-01 3.56478691e-02
1.38296592e+00 1.66138005e-03 -4.59846646e-01 4.14024949e-01
1.01653862e+00 4.10344094e-01 -6.22262537e-01 -6.08231723e-01
8.40834737e-01 -2.64871061e-01 -5.82018197e-01 -1.01069176e+00
-7.70729005e-01 5.98402619e-01 -2.44835809e-01 6.01244211e-01
8.18127394e-01 5.47038674e-01 8.60044062e-01 6.76916361e-01
4.39177513e-01 -7.70160496e-01 -1.22941613e-01 4.73476976e-01
6.61946476e-01 -1.29480386e+00 -4.67669927e-02 -1.07749283e+00
-4.75209534e-01 1.22353435e+00 6.12287760e-01 2.79790014e-01
3.29252094e-01 4.07950312e-01 1.00211278e-02 -4.48531806e-01
-8.52977276e-01 -3.76096159e-01 4.37755495e-01 5.83952844e-01
6.59084082e-01 -1.03613541e-01 -1.11013703e-01 7.41303146e-01
-3.95183325e-01 -2.29052216e-01 3.62637579e-01 9.66248631e-01
-9.56701040e-02 -1.75278592e+00 -1.65375061e-02 3.27127308e-01
-3.90087575e-01 -3.74814659e-01 -6.63167715e-01 9.95053589e-01
-1.22522697e-01 9.48230028e-01 -1.59496069e-01 -3.62433016e-01
7.95467019e-01 3.66349071e-01 5.59203506e-01 -1.00699151e+00
-5.30118883e-01 -3.00151467e-01 6.23250902e-01 -7.36328423e-01
-4.08109695e-01 -7.29457796e-01 -1.37069643e+00 -2.78397948e-01
-1.61988020e-01 3.68502051e-01 4.88961279e-01 8.76783371e-01
6.96051121e-01 2.71132559e-01 3.06193739e-01 2.28462502e-01
-3.77546400e-01 -5.73272824e-01 -9.44978967e-02 7.50042379e-01
-3.69057238e-01 -4.59462672e-01 -6.95933960e-03 2.26010486e-01] | [9.519392967224121, 8.118215560913086] |
7c1a616b-8445-4824-aa13-196917f2de3e | winter-wheat-crop-yield-prediction-on | 2306.11946 | null | https://arxiv.org/abs/2306.11946v1 | https://arxiv.org/pdf/2306.11946v1.pdf | Winter Wheat Crop Yield Prediction on Multiple Heterogeneous Datasets using Machine Learning | Winter wheat is one of the most important crops in the United Kingdom, and crop yield prediction is essential for the nation's food security. Several studies have employed machine learning (ML) techniques to predict crop yield on a county or farm-based level. The main objective of this study is to predict winter wheat crop yield using ML models on multiple heterogeneous datasets, i.e., soil and weather on a zone-based level. Experimental results demonstrated their impact when used alone and in combination. In addition, we employ numerous ML algorithms to emphasize the significance of data quality in any machine-learning strategy. | ['Prof. Mohand Tahar Kechadi', 'Dr. David Lillis', 'Yogesh Bansal'] | 2023-06-20 | null | null | null | null | ['crop-yield-prediction', 'crop-yield-prediction'] | ['computer-vision', 'miscellaneous'] | [ 1.00936718e-01 -3.49102229e-01 -6.82955503e-01 -2.46404976e-01
-4.23477292e-01 -5.04143775e-01 2.58178353e-01 9.01383162e-01
-2.34272648e-02 9.69957829e-01 -1.66294277e-01 -8.07933509e-01
-3.44321400e-01 -1.58868361e+00 -5.48208475e-01 -7.69611657e-01
-1.72175735e-01 -1.56066984e-01 5.77326268e-02 -4.30820614e-01
-3.87010127e-02 6.92622781e-01 -1.67763734e+00 8.54408816e-02
9.20611918e-01 7.99887955e-01 6.04607165e-01 8.39051723e-01
-3.14513445e-02 4.36343908e-01 -3.43567759e-01 1.83779828e-03
1.49364159e-01 -3.09598297e-01 -1.77572757e-01 8.41584057e-02
-1.77155361e-01 -8.01458433e-02 2.93356985e-01 8.44413936e-01
3.25722188e-01 -3.44386011e-01 7.59659827e-01 -1.10012829e+00
-3.27892542e-01 6.10450864e-01 -8.83172512e-01 -1.09505504e-01
8.00789800e-03 -8.01010057e-02 8.69018376e-01 -5.04395425e-01
1.07483856e-01 1.01685536e+00 5.89373589e-01 -5.39852083e-01
-1.34909260e+00 -6.97068870e-01 4.32766639e-02 1.80013273e-02
-1.06617510e+00 -2.28265896e-01 3.97332788e-01 -4.97892708e-01
7.83784807e-01 3.13917488e-01 8.55337620e-01 6.70800433e-02
1.05020368e+00 7.56822467e-01 1.27076089e+00 -7.95025110e-01
2.82886446e-01 -1.75034344e-01 9.12496448e-02 4.60942715e-01
8.56968105e-01 5.36883295e-01 -3.58443707e-01 -1.84149891e-01
6.53969765e-01 7.40576461e-02 2.06960469e-01 -4.36405450e-01
-1.00275612e+00 1.17183721e+00 6.38411760e-01 2.12928399e-01
-7.48463869e-01 -2.04034641e-01 2.37158030e-01 4.82736140e-01
7.04038143e-01 3.85978341e-01 -1.05746770e+00 2.99127340e-01
-9.81955290e-01 3.60778630e-01 7.57814467e-01 7.07797885e-01
6.18205965e-01 2.05057979e-01 3.57476473e-01 6.31425261e-01
4.34909165e-01 1.14742470e+00 6.34455532e-02 -5.04539669e-01
1.72548756e-01 7.45759547e-01 1.72870010e-01 -1.22500551e+00
-5.45566440e-01 -2.60295749e-01 -1.08095157e+00 1.76074237e-01
1.99475929e-01 -2.18541086e-01 -7.50853598e-01 1.33885622e+00
-4.25346568e-02 5.80123952e-03 5.17132699e-01 2.62929022e-01
8.20315540e-01 7.24730730e-01 5.42681336e-01 -3.90064776e-01
1.24182296e+00 -2.31409177e-01 -7.45999992e-01 -1.21590361e-01
1.15482080e+00 -7.46397555e-01 5.75071752e-01 2.50935435e-01
-6.82118893e-01 -5.26499629e-01 -9.80212808e-01 7.65482366e-01
-1.14065802e+00 5.42483985e-01 8.78520012e-01 6.68179214e-01
-7.17813790e-01 4.42759007e-01 -9.82210100e-01 -4.89371896e-01
-1.46298349e-01 1.00992166e-01 -4.81949389e-01 -2.18990296e-01
-1.09585047e+00 1.24493420e+00 7.47677922e-01 3.00948411e-01
-1.25878677e-01 -7.60319829e-01 -1.00082934e+00 1.29991710e-01
-1.05364740e-01 -1.40028700e-01 6.05427682e-01 -3.93970400e-01
-1.02337408e+00 7.62207687e-01 -2.97401190e-01 -4.15441424e-01
-7.20091388e-02 -1.63115934e-01 -7.60621250e-01 -4.85003740e-01
2.29816958e-01 3.55392605e-01 1.05104946e-01 -1.17073143e+00
-1.02639174e+00 -5.44757724e-01 -3.93714666e-01 -1.00952685e-02
6.66598603e-02 1.20615192e-01 5.50947845e-01 -5.53735018e-01
5.81178248e-01 -1.03895628e+00 -4.13205773e-01 -5.14866531e-01
9.75412279e-02 1.34857401e-01 6.50741339e-01 -9.80892777e-01
1.19577420e+00 -1.96223474e+00 -1.54025301e-01 3.17660511e-01
-4.36482430e-01 1.37423322e-01 -1.41473606e-01 4.87317026e-01
2.59429845e-03 6.64594397e-02 -3.98557901e-01 5.79359651e-01
-2.58338541e-01 4.05846655e-01 -2.09843623e-03 3.38194996e-01
7.62939453e-01 6.93501055e-01 -8.92878473e-01 -2.25254864e-01
5.94915450e-01 1.02040380e-01 1.94536313e-01 1.05709814e-01
-1.62520647e-01 -1.64954618e-01 -3.16645890e-01 1.14652395e+00
1.01056170e+00 2.27263689e-01 4.46639240e-01 -1.02014273e-01
-5.83328843e-01 -3.68032604e-01 -1.09177840e+00 1.02782488e+00
-4.93156731e-01 3.69277209e-01 -2.23715067e-01 -1.01725721e+00
1.33665824e+00 3.33534747e-01 4.44267303e-01 -5.33900440e-01
-2.33175531e-01 4.50515211e-01 2.76095331e-01 -2.21498281e-01
5.97950637e-01 2.86093771e-01 -2.61537313e-01 8.69087055e-02
-4.66375127e-02 -3.18595678e-01 3.62227529e-01 -3.45687419e-01
6.02258205e-01 4.73962516e-01 8.81504714e-01 -9.86605823e-01
1.88771039e-01 4.44228381e-01 6.56969011e-01 3.25344980e-01
-5.37125953e-02 2.93007940e-01 5.23985326e-01 -4.18127447e-01
-7.21871257e-01 -6.86055601e-01 -5.33561289e-01 1.28756070e+00
-2.21189678e-01 2.57812232e-01 2.44930275e-02 -2.14791059e-01
5.35844028e-01 8.79211962e-01 -5.20370901e-01 -2.87340004e-02
-1.32379487e-01 -1.82900035e+00 1.40991807e-01 6.62791431e-01
4.90514040e-01 -1.11470389e+00 -8.09818149e-01 5.84286690e-01
-8.90578777e-02 -7.08109558e-01 6.76945925e-01 9.28893745e-01
-9.45522010e-01 -8.26686800e-01 -5.75263083e-01 -4.90460694e-01
2.33223334e-01 3.44548732e-01 1.48806739e+00 -8.37408975e-02
-2.59007871e-01 -2.63461769e-01 -4.72464710e-01 -1.12752926e+00
-4.95469213e-01 5.59308410e-01 -1.98791429e-01 -5.78609347e-01
8.00946653e-01 -2.61595666e-01 -1.35108471e-01 5.34152538e-02
-7.59480178e-01 -2.19498277e-01 7.71759868e-01 7.23742366e-01
6.15686953e-01 6.03457689e-01 9.39001143e-01 -6.57386124e-01
4.16427314e-01 -9.02991116e-01 -9.32044804e-01 8.88391137e-01
-6.09084606e-01 -1.00736298e-01 -7.63041005e-02 4.38431054e-02
-7.83988416e-01 3.01242352e-01 2.11883232e-01 6.89862192e-01
-4.62159067e-01 1.29443693e+00 -4.34058696e-01 1.23355776e-01
2.97987521e-01 4.79371939e-03 -9.55631211e-02 -2.12141052e-01
9.69908084e-04 4.81686890e-01 2.60632098e-01 -3.04108083e-01
4.42377329e-01 -2.83023804e-01 6.36690855e-01 -1.09444082e+00
-5.59646726e-01 -5.14840841e-01 -1.18216622e+00 -2.03789338e-01
4.62720037e-01 -1.15872121e+00 -2.02611566e-01 7.48054564e-01
-5.76682866e-01 -4.96756822e-01 1.65196538e-01 8.33039761e-01
-3.69563818e-01 -4.23907220e-01 -3.78527828e-02 -8.34412932e-01
-2.36917496e-01 -8.24905872e-01 7.75480092e-01 3.05046707e-01
-1.13796495e-01 -1.06162882e+00 4.37274203e-02 -3.21930170e-01
2.20209047e-01 8.63389790e-01 1.14817643e+00 -1.17222428e-01
7.64240175e-02 -4.78629857e-01 -6.36042714e-01 -1.17263153e-01
7.00058460e-01 5.55028677e-01 -8.72625887e-01 -3.57320100e-01
-4.63742137e-01 -2.20028549e-01 8.33106339e-01 1.07508039e+00
3.87274325e-01 3.96924525e-01 -2.97234803e-01 2.59398401e-01
1.80480611e+00 4.14544016e-01 4.51476365e-01 5.10420442e-01
1.18725695e-01 7.93724597e-01 1.26485145e+00 4.10543710e-01
3.12336951e-01 1.25949815e-01 3.50508928e-01 -5.54682314e-01
3.07095468e-01 5.34604080e-02 -1.66075826e-02 6.32105410e-01
-5.51794097e-02 -3.09385031e-01 -1.14665627e+00 5.61348855e-01
-1.98068726e+00 -9.68326092e-01 -3.80352110e-01 1.96527398e+00
5.53196013e-01 2.74926610e-02 -4.07843255e-02 6.93175673e-01
1.56767502e-01 1.57996520e-01 -4.57160205e-01 -6.65032387e-01
-5.94678402e-01 2.34378338e-01 1.35913658e+00 1.91993192e-01
-1.28048110e+00 9.80630875e-01 6.67107344e+00 2.81717330e-01
-9.88094985e-01 -6.23459697e-01 7.86743224e-01 4.46157336e-01
-1.09440975e-01 -1.04521483e-01 -9.32945848e-01 -3.17252018e-02
9.69774425e-01 3.42576168e-02 -2.72120684e-02 5.51776767e-01
5.81603825e-01 -8.70906711e-01 -5.36467135e-01 -9.03765634e-02
-4.29664165e-01 -9.13931727e-01 -3.81374389e-01 1.15204461e-01
7.74204791e-01 -9.46875364e-02 -1.59306511e-01 3.05860639e-01
9.88939226e-01 -9.14238334e-01 1.95614517e-01 5.81671834e-01
6.50413156e-01 -1.09559906e+00 1.22931707e+00 3.20483446e-01
-1.67026162e+00 -1.58332614e-03 -5.95539510e-01 -2.61615008e-01
-2.69415259e-01 8.10336292e-01 -4.39874411e-01 1.10653770e+00
9.10336137e-01 8.13596547e-01 -6.51455879e-01 6.99753821e-01
1.30142421e-01 8.69192958e-01 -4.89182591e-01 1.93587858e-02
3.89331996e-01 -2.28520438e-01 -2.47222632e-01 1.23469651e+00
5.98388433e-01 2.96491563e-01 2.78789312e-01 2.79188812e-01
4.24562335e-01 3.52643490e-01 -7.16230810e-01 1.70847461e-01
4.98550236e-01 8.92981112e-01 -9.30575609e-01 2.82230288e-01
-4.41525340e-01 3.29149902e-01 -2.72926897e-01 1.43889055e-01
-2.00810954e-01 -4.40600574e-01 6.73258066e-01 -2.12804556e-01
2.06597865e-01 -2.42542669e-01 -5.24599016e-01 -7.29319572e-01
-3.42806369e-01 -5.70769131e-01 2.02643320e-01 -6.08746231e-01
-8.96583378e-01 9.93316155e-03 1.32744238e-01 -8.61873269e-01
-2.44686037e-01 -6.85508788e-01 -4.42582607e-01 1.38805866e+00
-1.91972291e+00 -1.37965477e+00 -2.19231978e-01 2.46238634e-02
2.41666123e-01 -4.87149894e-01 1.57146490e+00 -3.46534669e-01
-6.03046715e-01 1.49582773e-01 6.80866063e-01 -1.63435310e-01
3.49838704e-01 -1.24371541e+00 3.95041615e-01 8.94988000e-01
-1.32668868e-01 -2.76318073e-01 4.25762713e-01 -8.33442986e-01
-1.14711046e+00 -1.33143854e+00 1.25504088e+00 1.59131065e-01
5.70749164e-01 1.68691128e-01 -7.92584121e-01 4.66877520e-01
-8.64274502e-02 -5.05774282e-02 7.49135852e-01 4.40635264e-01
1.61711723e-01 -3.66349816e-01 -1.40230453e+00 1.18692406e-01
1.35117546e-01 -3.93803231e-02 1.20444410e-01 3.83544974e-02
4.52277094e-01 -4.57768627e-02 -1.41185534e+00 9.46173012e-01
8.65680158e-01 -6.21306360e-01 9.31163609e-01 -3.74268681e-01
3.06854934e-01 -9.89590734e-02 -7.17835426e-01 -1.61251140e+00
-7.12390125e-01 3.32548767e-01 9.26644132e-02 1.16991758e+00
4.65051740e-01 -5.00078261e-01 4.88603979e-01 4.34056193e-01
4.81360942e-01 -3.77468288e-01 -3.46580803e-01 -5.75253606e-01
4.16663080e-01 -3.28036994e-02 1.15585637e+00 7.84563959e-01
-2.78451324e-01 -2.39471868e-01 -6.85971752e-02 6.61424339e-01
5.59242845e-01 5.03659725e-01 5.32396436e-01 -1.64464986e+00
5.00856102e-01 -5.19064724e-01 -4.65231478e-01 2.47158647e-01
1.16157480e-01 -3.05795610e-01 -1.42032370e-01 -1.44326830e+00
7.79249221e-02 -8.09275806e-01 -6.48317397e-01 6.10591054e-01
-5.65734446e-01 -1.58344880e-01 1.07638583e-01 -1.84757009e-01
6.68558419e-01 -2.82552354e-02 7.84099996e-01 -1.45376159e-03
-5.91384828e-01 6.58379257e-01 -4.50333208e-01 4.28781837e-01
1.52120471e+00 -5.47508717e-01 -3.47889602e-01 -3.37751627e-01
2.61777490e-01 3.36672455e-01 1.01227805e-01 -7.25821912e-01
-4.20891285e-01 -8.72029424e-01 7.26241469e-01 -1.30681157e+00
-2.08315104e-01 -9.21921015e-01 4.12475556e-01 7.60174930e-01
-3.04852992e-01 4.27677393e-01 7.52749920e-01 2.30528906e-01
-1.78947940e-01 -1.70788586e-01 6.00541055e-01 -4.74820798e-03
-8.20613563e-01 -3.69267017e-02 -4.89122868e-01 -7.75017619e-01
1.00553393e+00 4.90957983e-02 1.79136723e-01 1.00201674e-01
-4.63260323e-01 2.47409090e-01 1.79717675e-01 5.64726293e-01
1.63227618e-01 -1.14840150e+00 -1.46031392e+00 4.80535686e-01
3.81706983e-01 -3.76810968e-01 -3.92938346e-01 3.03704351e-01
-8.75040591e-01 7.90764809e-01 -5.86313903e-01 -5.35806835e-01
-1.33402824e+00 2.90234208e-01 6.09910525e-02 -6.35483682e-01
-1.91355199e-01 4.70070541e-01 -4.17745769e-01 -5.91821611e-01
-1.32167101e-01 -6.71461284e-01 -5.11950433e-01 4.00316119e-01
1.19707532e-01 1.96989655e-01 3.70086312e-01 -5.36141515e-01
-2.40209505e-01 2.73124695e-01 4.65652704e-01 -7.31004821e-03
1.48793411e+00 5.74709959e-02 -8.35079327e-02 9.00406003e-01
5.54981887e-01 -6.13203704e-01 -8.40226114e-01 -2.68434912e-01
6.44069254e-01 -3.73210669e-01 5.16776502e-01 -1.04647303e+00
-1.02412748e+00 5.58489919e-01 9.17191267e-01 3.37708384e-01
1.34450424e+00 -7.13899553e-01 1.61058053e-01 4.57171053e-01
2.88261086e-01 -7.61988699e-01 -9.16032851e-01 1.56957880e-01
6.91242754e-01 -1.85727668e+00 4.33908492e-01 -1.27760082e-01
-4.02860016e-01 1.09009349e+00 1.59055665e-01 5.71333878e-02
1.22442877e+00 6.44528151e-01 5.82789592e-02 3.95387322e-01
-6.42907739e-01 -3.51620495e-01 -1.29738092e-01 8.68033409e-01
9.74026918e-01 1.05190527e+00 -2.20595166e-01 1.81222484e-01
-5.21369949e-02 2.27971092e-01 3.32267702e-01 1.41456664e+00
-6.81190193e-01 -1.23822927e+00 -6.94379866e-01 9.80836153e-01
-1.20587699e-01 -1.20253935e-01 -1.47375807e-01 8.52836788e-01
1.35800183e-01 1.17081165e+00 -4.33501266e-02 -2.32151002e-02
3.27519506e-01 -1.90833788e-02 3.47042620e-01 -3.77135128e-01
-5.72826028e-01 9.16208029e-02 -1.92061692e-01 5.64820245e-02
-7.96924651e-01 -1.10334671e+00 -5.78078866e-01 -8.88932407e-01
-8.23609769e-01 -5.25206663e-02 9.53810930e-01 6.42810583e-01
-1.05731720e-02 5.28112769e-01 8.81450057e-01 -6.03719831e-01
-2.03231871e-01 -1.19276488e+00 -1.05676281e+00 -4.94405270e-01
1.07973859e-01 -6.03172302e-01 1.19743675e-01 1.67328224e-01] | [9.361527442932129, -1.5976794958114624] |
9709fd56-4c90-4ce9-8ae7-54957e7bb478 | object-counting-and-instance-segmentation | 1903.02494 | null | https://arxiv.org/abs/1903.02494v2 | https://arxiv.org/pdf/1903.02494v2.pdf | Object Counting and Instance Segmentation with Image-level Supervision | Common object counting in a natural scene is a challenging problem in computer vision with numerous real-world applications. Existing image-level supervised common object counting approaches only predict the global object count and rely on additional instance-level supervision to also determine object locations. We propose an image-level supervised approach that provides both the global object count and the spatial distribution of object instances by constructing an object category density map. Motivated by psychological studies, we further reduce image-level supervision using a limited object count information (up to four). To the best of our knowledge, we are the first to propose image-level supervised density map estimation for common object counting and demonstrate its effectiveness in image-level supervised instance segmentation. Comprehensive experiments are performed on the PASCAL VOC and COCO datasets. Our approach outperforms existing methods, including those using instance-level supervision, on both datasets for common object counting. Moreover, our approach improves state-of-the-art image-level supervised instance segmentation with a relative gain of 17.8% in terms of average best overlap, on the PASCAL VOC 2012 dataset. Code link: https://github.com/GuoleiSun/CountSeg | ['Guolei Sun', 'Hisham Cholakkal', 'Fahad Shahbaz Khan', 'Ling Shao'] | 2019-03-06 | object-counting-and-instance-segmentation-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Cholakkal_Object_Counting_and_Instance_Segmentation_With_Image-Level_Supervision_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Cholakkal_Object_Counting_and_Instance_Segmentation_With_Image-Level_Supervision_CVPR_2019_paper.pdf | cvpr-2019-6 | ['object-counting', 'image-level-supervised-instance-segmentation'] | ['computer-vision', 'computer-vision'] | [ 1.95550278e-01 -2.75808007e-01 -5.01645923e-01 -5.79188168e-01
-8.23594809e-01 -4.73029524e-01 5.45053661e-01 4.12516713e-01
-8.95652235e-01 5.86461246e-01 -5.80018222e-01 1.78667475e-02
1.78976774e-01 -8.59779477e-01 -9.66705084e-01 -3.41561317e-01
3.18149358e-01 9.85233068e-01 6.44007862e-01 5.72899401e-01
5.90682149e-01 2.00540155e-01 -1.63968050e+00 7.19700903e-02
9.73777533e-01 9.00560915e-01 6.32508218e-01 8.12845826e-01
-3.36191267e-01 1.01962650e+00 -6.06748462e-01 -4.31965500e-01
6.89456835e-02 -2.95846581e-01 -9.49057400e-01 2.86589116e-01
1.17905617e+00 -5.68353057e-01 7.42771430e-03 1.37027657e+00
-6.47009956e-03 -3.93577404e-02 9.48247910e-01 -1.17344940e+00
-5.38744032e-01 1.95361346e-01 -1.24086189e+00 6.56500101e-01
-5.38350567e-02 1.70095354e-01 8.74469161e-01 -7.37520993e-01
3.03520501e-01 1.06162882e+00 4.63536710e-01 3.58411700e-01
-1.19990385e+00 -7.46443570e-01 1.12919010e-01 1.79404870e-01
-1.44229853e+00 6.19619116e-02 4.81632799e-01 -5.58616400e-01
8.45571637e-01 6.08416311e-02 6.04325950e-01 2.51736760e-01
-3.77255857e-01 1.12433505e+00 1.38874710e+00 -4.35961843e-01
2.32957453e-01 3.30730043e-02 4.74400461e-01 9.56150711e-01
6.41394019e-01 -3.17829758e-01 -3.70596200e-01 1.84472963e-01
9.94309187e-01 7.55237192e-02 2.54237473e-01 -1.77624077e-01
-1.12359464e+00 8.42367291e-01 6.66419208e-01 1.49255916e-01
-2.26247802e-01 5.39891958e-01 2.60757685e-01 -4.94879544e-01
7.71941006e-01 9.93140712e-02 -4.76417840e-01 -1.21961959e-01
-1.36398733e+00 2.55294383e-01 8.70504677e-01 1.09642458e+00
9.70835805e-01 -2.19750419e-01 -3.49093378e-01 9.80811536e-01
-1.36344731e-02 7.43667603e-01 1.38344184e-01 -1.20684075e+00
5.53771615e-01 7.80893326e-01 6.87787533e-02 -8.76762390e-01
-2.93377727e-01 -3.37636769e-01 -4.18008864e-01 -8.44339207e-02
9.04479325e-01 2.93356329e-01 -1.20774221e+00 1.31034660e+00
4.23367560e-01 5.33702970e-01 -6.08033061e-01 8.28583360e-01
8.77795994e-01 4.68255669e-01 3.09845120e-01 8.21581036e-02
1.55347109e+00 -1.05373025e+00 -4.55246031e-01 -4.96136725e-01
5.46591222e-01 -4.34115887e-01 1.19872439e+00 2.37018183e-01
-9.89594400e-01 -5.29157996e-01 -8.99391711e-01 -1.81493506e-01
-4.57716584e-01 7.26837814e-01 8.79386604e-01 7.14887261e-01
-5.79621017e-01 3.88061792e-01 -9.82266366e-01 -2.59453535e-01
1.08392060e+00 3.33760202e-01 2.05297749e-02 -3.87358278e-01
-3.54732394e-01 6.63690567e-01 7.02578843e-01 -2.61991709e-01
-7.64881670e-01 -8.59277785e-01 -9.67829525e-01 -1.19495362e-01
6.29317224e-01 -5.47222376e-01 1.24731088e+00 -5.83404124e-01
-7.89437592e-01 1.43012953e+00 -3.76112044e-01 -4.86016750e-01
4.23315316e-01 -1.35334894e-01 3.63696277e-01 3.26813668e-01
7.98436761e-01 1.22654390e+00 3.86044621e-01 -1.45971751e+00
-1.03690743e+00 -6.47939622e-01 8.25546775e-03 1.28255546e-01
-1.93524271e-01 -1.31472543e-01 -8.01873326e-01 -3.48728061e-01
6.81267902e-02 -7.96746671e-01 -8.39441791e-02 2.97814440e-02
-4.03522015e-01 -6.43640161e-01 6.85618341e-01 -5.85527539e-01
8.55492890e-01 -1.71353102e+00 -2.24548131e-01 -2.92881012e-01
1.98240608e-01 1.43150970e-01 1.23320125e-01 -3.07539880e-01
4.60964471e-01 1.44130168e-02 -6.04429603e-01 -7.99109638e-01
-1.85463026e-01 4.42062646e-01 9.22398716e-02 6.78928435e-01
4.02083248e-01 9.94183481e-01 -1.13314283e+00 -1.11744118e+00
6.54380858e-01 3.57853383e-01 -5.59397399e-01 2.20622700e-02
-4.49087620e-01 5.48617661e-01 -2.22257420e-01 9.68215346e-01
8.85895610e-01 -4.33058798e-01 -1.33728758e-01 -1.72658533e-01
-3.35423350e-02 -3.38299684e-02 -1.01618445e+00 1.49400377e+00
-4.09074157e-01 5.01167893e-01 -2.84234673e-01 -9.78075027e-01
5.96425891e-01 -3.63497168e-01 4.34769243e-01 -7.09676743e-01
2.46476546e-01 3.02782387e-01 -1.72343180e-01 -9.12786350e-02
5.49883962e-01 9.68140885e-02 -1.60428733e-01 1.82662189e-01
4.60748464e-01 -6.91859663e-01 8.91913235e-01 4.12396818e-01
7.64293432e-01 2.75318697e-02 4.33685303e-01 -2.39785492e-01
3.88227105e-01 3.85358751e-01 2.98319310e-01 9.65402842e-01
-5.25875747e-01 8.27949762e-01 5.64393222e-01 -2.23150223e-01
-9.69076037e-01 -9.33862150e-01 -5.56613624e-01 1.20841551e+00
4.09972191e-01 -1.19339421e-01 -1.14373422e+00 -7.60039866e-01
3.36012915e-02 7.79016376e-01 -6.40003204e-01 5.09989381e-01
-5.20427465e-01 -7.91887999e-01 4.96527344e-01 8.73234451e-01
9.66720045e-01 -1.10924888e+00 -3.35124135e-01 -8.27331170e-02
-3.66685778e-01 -1.89182174e+00 -6.51808679e-01 1.41616642e-01
-9.69960511e-01 -1.29700804e+00 -7.71726429e-01 -7.06126511e-01
8.94821763e-01 2.85022110e-01 1.45571125e+00 2.55672425e-01
-8.45069468e-01 6.57273233e-01 -3.82527038e-02 -7.09782660e-01
1.58434480e-01 1.35258764e-01 -3.02062064e-01 -3.45380276e-01
8.58647645e-01 -1.48423120e-01 -7.00305223e-01 4.11858708e-01
-7.76765168e-01 -1.27299815e-01 4.01969641e-01 5.79424679e-01
1.06349432e+00 -3.92882805e-03 3.61938924e-01 -1.01240373e+00
9.74591076e-02 -3.17267478e-01 -9.26432788e-01 -3.67021188e-02
-2.34762892e-01 -2.97382861e-01 1.36776447e-01 -3.44890088e-01
-9.62852597e-01 3.92968714e-01 1.95279375e-01 -3.19577754e-01
-4.94418293e-01 -1.21291362e-01 8.39726254e-02 -4.01776768e-02
5.23379147e-01 3.99196267e-01 -3.65770966e-01 -2.62481242e-01
3.22820634e-01 4.44239765e-01 7.58588254e-01 -6.85736358e-01
5.31064093e-01 7.72718072e-01 -1.16535172e-01 -8.28043103e-01
-1.58969808e+00 -1.19379759e+00 -1.09918225e+00 -2.59468943e-01
1.28954339e+00 -1.07852733e+00 -8.43228161e-01 6.29123330e-01
-1.17262805e+00 -4.61446881e-01 -2.68669486e-01 5.02393067e-01
-7.91156173e-01 2.51073688e-01 -7.16078937e-01 -8.78533185e-01
-1.33196414e-01 -9.17143762e-01 1.54047334e+00 4.55120027e-01
1.10048912e-01 -8.55000257e-01 -1.13183670e-01 9.58695710e-01
-1.10670529e-01 5.35498224e-02 3.37499112e-01 -4.30530995e-01
-8.88268471e-01 -2.35338017e-01 -9.05360222e-01 4.89113122e-01
-1.05699256e-01 -2.33915806e-01 -9.42861021e-01 1.30546421e-01
-2.41583511e-01 -5.02200961e-01 1.28728688e+00 8.75394642e-01
1.74899673e+00 9.83267948e-02 -3.06087166e-01 5.60614705e-01
1.63657439e+00 -2.54744530e-01 5.22704244e-01 2.13149592e-01
1.10319531e+00 4.21876788e-01 9.08236682e-01 4.00955588e-01
6.25461936e-01 4.68837291e-01 6.33984864e-01 -1.29249409e-01
-3.93767029e-01 -3.60334784e-01 -3.85793716e-01 4.69117761e-01
-3.93763393e-01 1.92428939e-02 -9.34613943e-01 9.56759274e-01
-1.74938715e+00 -8.88995767e-01 -4.47635651e-01 2.06812620e+00
8.37928236e-01 7.42124096e-02 4.16030198e-01 8.80934522e-02
9.82110083e-01 -4.14083153e-02 -6.04099512e-01 -2.35995017e-02
2.12966308e-01 1.80217579e-01 1.15469563e+00 4.07095730e-01
-1.49969792e+00 1.17095327e+00 5.54824829e+00 1.23192036e+00
-3.93907487e-01 4.57410663e-01 1.06848764e+00 -1.49978191e-01
4.58779901e-01 -2.08992675e-01 -1.26838541e+00 6.28246188e-01
4.45598722e-01 3.15772116e-01 2.62336135e-01 1.12537968e+00
-2.08699003e-01 -7.80940771e-01 -9.50184703e-01 1.25093162e+00
2.29942739e-01 -1.11662567e+00 -1.56403676e-01 1.41698673e-01
8.93676460e-01 9.85729694e-02 -5.24014197e-02 3.66818547e-01
2.72589356e-01 -1.06499076e+00 6.95758641e-01 1.92039713e-01
8.58969748e-01 -8.23504329e-01 8.42074692e-01 6.77245259e-01
-1.38059950e+00 1.58804491e-01 -6.02015078e-01 -1.45714357e-01
1.53062508e-01 7.72953272e-01 -6.17781043e-01 5.40303579e-03
8.09425712e-01 5.99973440e-01 -7.93745041e-01 1.11716092e+00
-3.58760297e-01 8.35637331e-01 -4.69447732e-01 -2.42063613e-03
2.89113402e-01 -2.19685093e-01 5.38908765e-02 1.43716919e+00
-1.20641246e-01 2.32299358e-01 4.71347064e-01 1.25255466e+00
-3.59768808e-01 4.23665978e-02 -3.42046648e-01 1.56992853e-01
3.92146289e-01 1.38150740e+00 -1.43327963e+00 -6.11390889e-01
-2.97429889e-01 1.01940477e+00 6.26892686e-01 -4.27503511e-02
-1.11769128e+00 -5.41426167e-02 7.01024905e-02 3.37864369e-01
4.50239301e-01 -4.05684292e-01 -5.87049484e-01 -9.21286881e-01
6.63974434e-02 -2.12204516e-01 4.25662339e-01 -5.54276764e-01
-1.42430615e+00 -5.81009164e-02 3.91241193e-01 -8.51293683e-01
1.58402115e-01 -7.67350852e-01 -4.08636600e-01 5.55999696e-01
-1.62729156e+00 -1.35696077e+00 -5.90360940e-01 4.47622836e-01
8.96415710e-01 1.04146227e-01 3.94445986e-01 4.01043028e-01
-3.67513776e-01 3.97615641e-01 -1.67016998e-01 5.15722692e-01
2.86302626e-01 -1.59518993e+00 7.93910474e-02 5.57285309e-01
4.73716915e-01 3.92130971e-01 2.88348377e-01 -6.74234629e-01
-8.80669415e-01 -1.24348617e+00 4.78830457e-01 -8.63926232e-01
4.03507352e-01 -3.73464406e-01 -7.10224330e-01 6.05801761e-01
-2.13326290e-01 6.28050625e-01 3.04890186e-01 2.60949321e-02
-1.82495221e-01 1.75036386e-01 -1.41363466e+00 1.41824812e-01
1.09196472e+00 -3.54745418e-01 -2.46206865e-01 7.12924957e-01
5.55555999e-01 -5.21696031e-01 -7.56390870e-01 3.46126646e-01
2.30930328e-01 -9.44660544e-01 1.10726798e+00 1.03916965e-01
6.78517401e-01 -3.61967027e-01 -1.71104372e-01 -7.40183413e-01
-1.82086490e-02 4.75133210e-01 -1.90186441e-01 1.24533689e+00
1.21733412e-01 -4.52353209e-01 1.19865620e+00 3.45085591e-01
1.18027709e-01 -7.66359508e-01 -9.08423245e-01 -8.02474856e-01
1.33998916e-01 -6.29791975e-01 9.11494195e-02 7.19583094e-01
-5.56617558e-01 8.13512653e-02 -1.23689272e-01 1.50652975e-01
1.15738070e+00 -4.12600823e-02 9.73802626e-01 -9.63213742e-01
-1.04820602e-01 -5.07211864e-01 -7.76758909e-01 -1.30771232e+00
2.43258581e-01 -7.76408911e-01 2.04012722e-01 -1.62555850e+00
1.03340006e+00 -3.97185266e-01 8.53438079e-02 3.33770186e-01
-4.41226274e-01 9.04326141e-01 3.74774128e-01 2.98644334e-01
-1.06432879e+00 2.11812273e-01 1.36143565e+00 -3.08907509e-01
3.55463549e-02 3.85244526e-02 -3.66021633e-01 1.02351022e+00
9.28206563e-01 -7.49072015e-01 -1.93161875e-01 -4.23090041e-01
-2.43113980e-01 -3.40868682e-01 7.52114475e-01 -1.21556401e+00
2.45774820e-01 -2.38533929e-01 5.55367708e-01 -1.17179883e+00
4.05736893e-01 -4.71139997e-01 -6.12043381e-01 4.14317518e-01
-2.67378967e-02 -6.07836008e-01 2.91542590e-01 7.07404315e-01
-1.84840620e-01 -5.37709832e-01 9.86953378e-01 -4.87608522e-01
-8.77190411e-01 5.14807463e-01 1.32843062e-01 3.11338633e-01
1.12737346e+00 -5.20200968e-01 -4.33105975e-01 6.83088005e-02
-4.00460929e-01 2.70437598e-01 3.32136720e-01 6.19959980e-02
3.35878015e-01 -1.03224552e+00 -6.89221501e-01 -2.07676202e-01
2.09011868e-01 5.11714637e-01 4.55871634e-02 8.59020054e-01
-6.21472299e-01 6.27448797e-01 7.46921301e-02 -1.29986572e+00
-1.26922607e+00 2.65510231e-01 2.74763495e-01 -3.46281052e-01
-3.43449771e-01 9.59362388e-01 4.73916888e-01 -7.28907049e-01
1.88209098e-02 -4.80246156e-01 -1.55992135e-01 -2.68831588e-02
5.36121130e-01 5.50474286e-01 -4.36316021e-02 -7.31826663e-01
-4.11813259e-01 7.67956138e-01 -4.91171479e-02 3.87139767e-02
1.14237010e+00 6.27842396e-02 -5.30039854e-02 5.50326347e-01
1.20651889e+00 -4.09017295e-01 -1.34398115e+00 -2.36501887e-01
-2.63036527e-02 -9.56735313e-01 1.28147066e-01 -6.79806113e-01
-1.25477469e+00 9.15832102e-01 6.00047469e-01 -1.09169386e-01
7.86485851e-01 5.60906708e-01 4.66741383e-01 2.47229129e-01
6.47099078e-01 -1.14864433e+00 4.42193151e-01 5.09393334e-01
3.33952069e-01 -1.94424188e+00 3.08996677e-01 -7.58720934e-01
-5.37526250e-01 6.04783595e-01 1.00565660e+00 -4.46199626e-01
4.37338382e-01 2.62811095e-01 -3.77515286e-01 -4.49575096e-01
2.37861872e-02 -7.64888823e-01 3.97147059e-01 7.55398154e-01
4.81911808e-01 3.53535295e-01 -3.30993712e-01 3.37140858e-01
-4.50763963e-02 5.20013720e-02 3.72368187e-01 7.53759027e-01
-6.73081219e-01 -5.48465729e-01 -5.68524957e-01 9.87609208e-01
-6.32054210e-01 -1.30396307e-01 -3.73967998e-02 9.73793805e-01
3.34848136e-01 8.23578298e-01 5.30703306e-01 4.15458113e-01
1.09974355e-01 -3.70835096e-01 9.25335228e-01 -9.69850898e-01
-8.11070949e-02 -2.49678150e-01 -1.58270121e-01 -4.11815912e-01
-7.93927908e-01 -6.74864054e-01 -1.50067246e+00 -3.89731586e-01
-7.60452569e-01 -5.37065268e-01 8.65511477e-01 9.04768884e-01
-3.28259170e-01 5.33922732e-01 5.05131483e-02 -1.27799737e+00
-1.13049775e-01 -1.11854732e+00 -8.15771997e-01 4.92233634e-01
-1.37559846e-01 -8.84735167e-01 -3.13304394e-01 2.56305993e-01] | [9.054641723632812, 0.4598081409931183] |
92e06fbf-87b8-47b2-8498-e3af351b2920 | when-does-return-conditioned-supervised | 2206.01079 | null | https://arxiv.org/abs/2206.01079v3 | https://arxiv.org/pdf/2206.01079v3.pdf | When does return-conditioned supervised learning work for offline reinforcement learning? | Several recent works have proposed a class of algorithms for the offline reinforcement learning (RL) problem that we will refer to as return-conditioned supervised learning (RCSL). RCSL algorithms learn the distribution of actions conditioned on both the state and the return of the trajectory. Then they define a policy by conditioning on achieving high return. In this paper, we provide a rigorous study of the capabilities and limitations of RCSL, something which is crucially missing in previous work. We find that RCSL returns the optimal policy under a set of assumptions that are stronger than those needed for the more traditional dynamic programming-based algorithms. We provide specific examples of MDPs and datasets that illustrate the necessity of these assumptions and the limits of RCSL. Finally, we present empirical evidence that these limitations will also cause issues in practice by providing illustrative experiments in simple point-mass environments and on datasets from the D4RL benchmark. | ['Joan Bruna', 'Romain Laroche', 'Jacob Buckman', 'Alberto Bietti', 'David Brandfonbrener'] | 2022-06-02 | null | null | null | null | ['d4rl'] | ['robots'] | [-1.07130203e-02 2.06131816e-01 -6.51371777e-01 -8.51162821e-02
-7.36548901e-01 -5.99043965e-01 7.66727805e-01 2.29406074e-01
-6.07556224e-01 1.22029686e+00 -5.19399121e-02 -6.58527970e-01
-7.07306743e-01 -6.99321270e-01 -1.00497341e+00 -8.44616294e-01
-8.12243283e-01 4.85120028e-01 2.22596332e-01 -1.58761039e-01
4.30001557e-01 5.19620955e-01 -1.70090437e+00 -1.47032127e-01
5.73860347e-01 7.82950044e-01 1.52059674e-01 7.54059255e-01
2.11352438e-01 8.05752277e-01 -4.26678836e-01 1.24145187e-01
6.23525381e-01 -4.95264232e-01 -6.46932542e-01 1.16451368e-01
-2.13243872e-01 -5.59679806e-01 -2.33026832e-01 8.08443010e-01
4.00515974e-01 4.67676193e-01 6.64775431e-01 -1.31078279e+00
-5.93443541e-03 5.57241917e-01 -4.49281365e-01 3.02730590e-01
4.22740012e-01 3.97107154e-01 9.20247436e-01 -2.90603727e-01
4.56455559e-01 1.29310060e+00 4.85810727e-01 4.64235097e-01
-1.14300060e+00 -3.05698752e-01 4.44213033e-01 -8.18715151e-03
-8.22466552e-01 -2.86318749e-01 4.97316599e-01 -2.97789395e-01
9.38479304e-01 9.33645293e-02 6.01047575e-01 1.00957513e+00
1.32673666e-01 9.96600747e-01 1.40655720e+00 -4.86192614e-01
7.10058987e-01 -5.04650287e-02 -1.77839428e-01 5.09172261e-01
3.79894257e-01 9.48193729e-01 -3.04840237e-01 -3.76787692e-01
8.41815948e-01 -2.38355175e-01 -3.28479037e-02 -9.68581975e-01
-1.03440440e+00 1.04689002e+00 2.55779028e-02 -1.46496892e-01
-3.83772522e-01 3.01453680e-01 3.36523712e-01 6.12636745e-01
2.95978844e-01 2.79312164e-01 -4.35325414e-01 -3.49291712e-01
-6.03586614e-01 6.25805914e-01 1.00833356e+00 9.56845462e-01
5.17267227e-01 2.09266338e-02 -1.68580249e-01 3.10715795e-01
2.13421911e-01 4.64290708e-01 4.76921797e-02 -1.23655248e+00
6.30345702e-01 -1.26135200e-01 8.12942505e-01 -3.32322687e-01
-2.88545161e-01 -3.58035475e-01 -1.39643848e-01 4.57800329e-01
5.02405167e-01 -5.31886220e-01 -4.75208849e-01 1.70767140e+00
3.44111860e-01 2.44019508e-01 2.69009769e-01 7.35239089e-01
-1.68395579e-01 6.44147098e-01 3.33484821e-02 -6.40719771e-01
2.21058264e-01 -6.83893144e-01 -5.10843873e-01 -4.52501811e-02
6.35236263e-01 -3.08045328e-01 1.09593654e+00 4.76108164e-01
-1.13751185e+00 -2.32686996e-01 -1.02789080e+00 7.79761434e-01
-6.08085170e-02 -2.27359101e-01 8.89315724e-01 6.53301656e-01
-8.82956982e-01 1.05520654e+00 -1.08337879e+00 -3.41525644e-01
7.02710748e-02 1.49808928e-01 3.01657617e-01 1.50571659e-01
-9.66523528e-01 1.03932774e+00 3.75553906e-01 -3.09538603e-01
-1.44687116e+00 -4.43283737e-01 -4.51237649e-01 -1.46280155e-01
8.73575270e-01 -8.51227492e-02 1.75548184e+00 -7.54040003e-01
-1.59859800e+00 3.66954476e-01 3.11498880e-01 -8.56798232e-01
8.16596806e-01 -3.53900522e-01 -1.09899752e-01 2.74998546e-01
-3.48040350e-02 3.00121635e-01 7.09492981e-01 -1.29923165e+00
-9.22846854e-01 -6.38810471e-02 3.66214097e-01 3.43723267e-01
4.24667299e-02 -1.36417061e-01 -2.43416186e-02 -3.84966552e-01
-4.15951222e-01 -1.01948380e+00 -4.55602825e-01 -2.11124137e-01
-1.94526594e-02 -4.76271451e-01 5.37977040e-01 -1.41727757e-02
9.75269675e-01 -1.94609666e+00 2.43679255e-01 2.85562396e-01
-4.66728389e-01 2.25194637e-03 -1.19217932e-01 9.80461121e-01
-3.71536799e-02 -1.74651593e-02 -2.89724231e-01 6.94793314e-02
4.44441617e-01 6.07383668e-01 -7.54601359e-01 7.65396476e-01
8.62110704e-02 5.74388504e-01 -1.20009196e+00 -1.98479906e-01
1.53651074e-01 -2.75008768e-01 -5.20679533e-01 2.10940450e-01
-7.28530467e-01 3.20416689e-01 -6.53082252e-01 3.95658672e-01
1.96546972e-01 1.62835434e-01 3.98709923e-01 4.21331495e-01
-3.44336212e-01 3.30384254e-01 -1.24104011e+00 1.31174803e+00
-2.54872203e-01 1.30059913e-01 1.29763588e-01 -1.28487909e+00
6.21533632e-01 1.65596485e-01 7.80398726e-01 -6.76415622e-01
-1.02935359e-02 1.24896877e-01 -1.20353557e-01 -5.00557363e-01
2.66757786e-01 -4.14748162e-01 -1.17024347e-01 9.54069734e-01
-1.50946844e-02 -2.70826668e-01 3.48062724e-01 1.33728728e-01
1.09130442e+00 6.58631921e-01 2.25293785e-01 -4.82105762e-01
1.57030255e-01 8.38119835e-02 5.56095779e-01 1.19282377e+00
-4.09861177e-01 -2.05733314e-01 7.45291889e-01 -9.89100710e-02
-1.00600517e+00 -1.24590123e+00 2.16976064e-03 1.03013742e+00
1.56982094e-01 -2.26887330e-01 -3.36019278e-01 -9.55586970e-01
4.84579861e-01 9.38907385e-01 -4.68366623e-01 -1.27607509e-01
-4.24627423e-01 -7.84295142e-01 2.69952446e-01 4.87197340e-01
1.30878091e-01 -1.02261817e+00 -9.55683410e-01 1.34940624e-01
3.59114438e-01 -8.32114697e-01 -3.54619548e-02 4.91106868e-01
-1.03012276e+00 -1.30916417e+00 -4.16789830e-01 -3.78062755e-01
4.67339218e-01 3.96665297e-02 1.04401791e+00 -3.60488929e-02
3.12604383e-02 1.12977064e+00 -4.57914472e-01 -5.64815998e-01
-5.00035167e-01 -1.42125068e-02 3.45462263e-01 -2.94786751e-01
-1.11755664e-02 -4.88530606e-01 -4.27974939e-01 2.42995620e-01
-8.51779282e-01 -1.94501385e-01 6.07822537e-01 6.25978589e-01
5.45430362e-01 2.90037453e-01 8.86812031e-01 -7.71527231e-01
8.73301387e-01 -6.79499686e-01 -1.06735575e+00 1.13595508e-01
-7.76135504e-01 3.77285212e-01 5.79040527e-01 -4.23747838e-01
-1.02086949e+00 -7.59789208e-03 1.71756014e-01 -2.79052764e-01
-5.92512153e-02 4.15146858e-01 2.68385381e-01 4.23555337e-02
4.68789518e-01 2.74420083e-01 2.21106216e-01 -4.71318483e-01
2.27859393e-01 3.05707633e-01 2.27425620e-01 -1.29044759e+00
8.49727571e-01 3.27149779e-01 2.01373592e-01 -5.81781507e-01
-8.74534130e-01 -1.22593343e-01 -3.38373870e-01 -4.14110184e-01
3.29363227e-01 -5.97171724e-01 -9.21758056e-01 6.03686972e-03
-3.95350248e-01 -1.05346715e+00 -6.59552455e-01 5.58927000e-01
-1.41445422e+00 3.16659629e-01 -5.02757668e-01 -1.26976156e+00
2.63625741e-01 -8.28965664e-01 5.85051894e-01 9.61021557e-02
2.77328074e-01 -1.05348432e+00 3.63594890e-01 -3.48813444e-01
2.54920542e-01 3.24295580e-01 9.34150040e-01 -3.67885053e-01
-6.31791174e-01 2.54553169e-01 3.06079298e-01 2.84834981e-01
-8.06716084e-02 -1.31919593e-01 -4.84420031e-01 -8.12144458e-01
5.49374670e-02 -6.47361755e-01 7.49513149e-01 3.49858105e-01
1.17863464e+00 -3.79006624e-01 -2.41631463e-01 2.87434191e-01
1.64561236e+00 4.17350590e-01 2.59928524e-01 6.23967588e-01
-1.05524167e-01 5.91439843e-01 1.35968626e+00 8.39915872e-01
1.55697256e-01 3.37822497e-01 6.06385052e-01 1.63122460e-01
3.56227785e-01 -4.92492110e-01 6.49253488e-01 3.36754590e-01
-7.05669895e-02 -5.59030883e-02 -6.46043599e-01 5.46274245e-01
-2.00615525e+00 -1.12319648e+00 3.52382988e-01 2.51366782e+00
9.26632583e-01 3.37477922e-01 6.39494359e-01 1.32600695e-01
4.93872613e-01 7.45124593e-02 -8.75561416e-01 -5.68487942e-01
2.48868570e-01 3.30686480e-01 9.55026031e-01 5.29401720e-01
-1.02646744e+00 6.78155422e-01 7.82910013e+00 5.63181460e-01
-7.30375230e-01 -1.05848022e-01 3.53356272e-01 -2.50066668e-01
-1.17969647e-01 1.71953425e-01 -7.35336125e-01 3.79497290e-01
1.12070835e+00 -2.97407717e-01 7.52892137e-01 9.07432556e-01
3.99948031e-01 -4.89718407e-01 -1.25292397e+00 5.05725265e-01
-3.94171327e-01 -9.71495450e-01 -4.26203221e-01 3.08176856e-02
8.12771022e-01 1.47859722e-01 6.20890483e-02 6.46075904e-01
8.56351078e-01 -7.88312018e-01 6.42148256e-01 4.69978422e-01
5.13058722e-01 -1.17046142e+00 2.71292776e-01 7.01164722e-01
-7.29109645e-01 -5.95088422e-01 -3.22655767e-01 -2.70742953e-01
7.86717907e-02 2.42640868e-01 -6.74570262e-01 6.43300414e-01
5.15843213e-01 7.20073700e-01 -1.24756344e-01 1.13655055e+00
-3.17971289e-01 7.74997354e-01 -4.63059455e-01 -3.04360509e-01
6.18921340e-01 -2.78396845e-01 5.49025595e-01 9.88804638e-01
8.65873843e-02 -1.02168091e-01 5.71438432e-01 6.05992675e-01
5.22383332e-01 -2.74040967e-01 -7.26463795e-01 -1.08958781e-01
2.75893033e-01 7.92504489e-01 -5.84172547e-01 -1.46249905e-01
-3.07498544e-01 3.99600893e-01 2.84465224e-01 3.44528168e-01
-8.97790074e-01 -1.92396790e-01 5.68611145e-01 7.14767864e-03
4.75112408e-01 -5.93409240e-01 2.73383588e-01 -8.40880692e-01
-1.06993124e-01 -8.81152272e-01 4.51694667e-01 -2.10625976e-01
-1.36517179e+00 -2.70437628e-01 6.19844675e-01 -1.25829852e+00
-4.97169703e-01 -5.64855754e-01 -4.63009000e-01 3.46836060e-01
-1.75496840e+00 -2.49284625e-01 3.58092517e-01 4.57968205e-01
4.42202240e-01 -1.06896296e-01 4.90935445e-01 -4.21899036e-02
-5.44334590e-01 9.76669714e-02 5.34801483e-01 -3.19288254e-01
4.44504350e-01 -1.47264767e+00 1.27541199e-02 6.38384759e-01
-1.44717261e-01 3.02809298e-01 9.16519105e-01 -6.52410448e-01
-1.70065820e+00 -9.49038327e-01 1.39287159e-01 -2.75133729e-01
7.76806235e-01 -3.95152010e-02 -4.56415623e-01 8.45348358e-01
1.86427347e-02 -2.18676314e-01 3.24903458e-01 -5.31822070e-02
3.11595201e-01 -7.70262107e-02 -1.10120451e+00 5.82353234e-01
1.07899785e+00 -2.17888579e-01 -6.49307728e-01 4.95221049e-01
5.35988569e-01 -4.01021212e-01 -6.94329798e-01 4.63770449e-01
3.16967130e-01 -1.06866550e+00 9.35343027e-01 -9.19484735e-01
1.47820860e-01 -1.43073142e-01 -1.26477703e-01 -1.40786517e+00
1.33484863e-02 -8.51029694e-01 -4.64715540e-01 8.55871201e-01
7.09438771e-02 -6.55508399e-01 5.60496330e-01 3.14817101e-01
2.68103294e-02 -9.23657954e-01 -8.89484406e-01 -1.28383982e+00
4.70760882e-01 -4.81220812e-01 4.27559227e-01 6.19595468e-01
1.62628636e-01 -1.31873354e-01 -3.62033188e-01 7.10481554e-02
9.51708734e-01 2.24609792e-01 6.60922050e-01 -7.19136178e-01
-6.69263124e-01 -2.59277344e-01 1.53353199e-01 -1.14014268e+00
3.24076563e-01 -6.76410675e-01 2.37697512e-01 -1.26334667e+00
1.11195989e-01 -8.72958899e-01 -3.90509486e-01 3.08539569e-01
1.35775506e-01 -5.20466745e-01 1.67596862e-01 6.37909397e-02
-9.15148795e-01 7.54947662e-01 1.23252439e+00 2.12050706e-01
-2.87449747e-01 4.52157050e-01 -4.45153415e-01 5.71779013e-01
1.07185793e+00 -4.22760576e-01 -8.45889568e-01 -8.57080892e-02
1.47715956e-01 4.43703145e-01 3.20065707e-01 -8.06289256e-01
4.78193127e-02 -8.27959657e-01 8.68440568e-02 -8.38691711e-01
8.31906945e-02 -7.69632280e-01 -2.61357009e-01 8.41138661e-01
-6.08966172e-01 1.75532550e-01 1.12924553e-01 9.96471167e-01
2.42037892e-01 -3.92965138e-01 7.77267218e-01 -1.94263399e-01
-6.76911473e-01 3.22819054e-01 -6.41185760e-01 2.44546741e-01
1.36900043e+00 1.89673513e-01 4.06251438e-02 -6.21290445e-01
-6.58131838e-01 5.53515136e-01 3.36631238e-01 2.96722561e-01
4.60400760e-01 -1.03852642e+00 -2.96709687e-01 -2.30219718e-02
-2.07229018e-01 -3.28601182e-01 -3.33670944e-01 7.19884396e-01
-1.52260408e-01 4.82183695e-01 -1.45697653e-01 -4.17254329e-01
-7.34868646e-01 9.33646679e-01 4.71601158e-01 -3.22174281e-01
-6.71710372e-01 8.26899186e-02 -3.91555369e-01 -4.30153638e-01
5.11366487e-01 -3.28347236e-01 1.76043153e-01 -3.23527336e-01
2.34804139e-01 4.28733408e-01 -2.18272746e-01 2.97724269e-02
-3.01712565e-02 1.40046388e-01 3.15104336e-01 -4.73747253e-01
1.53604257e+00 -2.53226310e-02 3.68535906e-01 5.89446604e-01
5.18210649e-01 -2.28569135e-01 -1.82274759e+00 -2.74520367e-01
3.46182853e-01 -4.81451213e-01 -4.39973362e-02 -7.29556024e-01
-7.23973155e-01 5.42026520e-01 5.60945272e-01 2.63616323e-01
9.42240655e-01 -1.38378918e-01 3.09184521e-01 6.24422669e-01
9.15629327e-01 -1.51130712e+00 1.53726190e-01 5.65541863e-01
7.20297098e-01 -9.16185796e-01 2.33912721e-01 1.09797128e-01
-6.45810246e-01 9.54790354e-01 5.35378575e-01 -4.73627955e-01
5.60385823e-01 3.51829529e-01 -5.18557847e-01 6.25149980e-02
-1.10955310e+00 -5.75356781e-01 -4.36809421e-01 6.78648770e-01
-7.74245858e-02 5.11810072e-02 -5.83074510e-01 2.92609543e-01
3.39603797e-02 3.43945734e-02 4.48889852e-01 1.48838413e+00
-6.82490528e-01 -1.34888923e+00 -3.43548119e-01 3.16865146e-01
-3.96011919e-01 5.10564148e-01 -1.40440226e-01 1.10823023e+00
-3.68535966e-01 8.78620386e-01 7.72377476e-02 -6.60276338e-02
2.45796874e-01 2.52558738e-02 9.45995390e-01 -4.33100790e-01
-2.49158770e-01 -5.18975146e-02 3.44145805e-01 -6.97815120e-01
-5.55922985e-01 -8.18327844e-01 -1.38473690e+00 -3.78625035e-01
2.12035775e-02 3.57375562e-01 3.86531204e-01 1.03046000e+00
4.80533466e-02 2.27959543e-01 9.78917718e-01 -8.75002980e-01
-1.45499218e+00 -6.13432348e-01 -7.42595673e-01 2.10811615e-01
4.08157974e-01 -9.93024588e-01 -4.89933014e-01 -3.72758090e-01] | [4.234809875488281, 2.2541377544403076] |
ccbbd1ee-e17c-4589-89df-579c96ff627a | outfin-a-multi-device-and-multi-modal-dataset | 2205.14921 | null | https://arxiv.org/abs/2205.14921v1 | https://arxiv.org/pdf/2205.14921v1.pdf | OutFin, a multi-device and multi-modal dataset for outdoor localization based on the fingerprinting approach | In recent years, fingerprint-based positioning has gained researchers attention since it is a promising alternative to the Global Navigation Satellite System and cellular network-based localization in urban areas. Despite this, the lack of publicly available datasets that researchers can use to develop, evaluate, and compare fingerprint-based positioning solutions constitutes a high entry barrier for studies. As an effort to overcome this barrier and foster new research efforts, this paper presents OutFin, a novel dataset of outdoor location fingerprints that were collected using two different smartphones. OutFin is comprised of diverse data types such as WiFi, Bluetooth, and cellular signal strengths, in addition to measurements from various sensors including the magnetometer, accelerometer, gyroscope, barometer, and ambient light sensor. The collection area spanned four dispersed sites with a total of 122 reference points. Each site is different in terms of its visibility to the Global Navigation Satellite System and reference points number, arrangement, and spacing. Before OutFin was made available to the public, several experiments were conducted to validate its technical quality. | ['Mohammad H. Mahoor', 'Fahad Alhomayani'] | 2022-05-30 | null | null | null | null | ['outdoor-localization'] | ['robots'] | [-7.37843812e-02 -4.40949649e-01 -3.63781780e-01 -3.67029637e-01
-5.82094073e-01 -6.75567925e-01 3.51451129e-01 2.19814330e-01
-3.44590724e-01 1.03264093e+00 5.90825826e-03 -4.24266994e-01
-2.35163391e-01 -1.17925179e+00 -4.71250117e-01 -5.40072978e-01
9.58124734e-03 3.58904973e-02 1.65652171e-01 -2.20081341e-02
3.63818109e-01 5.53382874e-01 -1.30411804e+00 -5.37370145e-01
8.60706329e-01 1.16456723e+00 5.42481318e-02 2.19180807e-01
1.87539041e-01 3.66453938e-02 -8.05475831e-01 -1.95354313e-01
1.16924912e-01 -2.64391806e-02 -1.15893282e-01 -5.90813398e-01
2.17792660e-01 -3.21235508e-01 -2.64106005e-01 5.72225451e-01
8.30463529e-01 7.29533285e-02 9.07471254e-02 -1.35962570e+00
-3.28662723e-01 2.77915359e-01 -4.13179457e-01 1.21136434e-01
8.23070109e-01 -3.98943573e-01 4.25418049e-01 -4.00336564e-01
3.46698016e-01 5.43489516e-01 1.25569940e+00 -1.12926230e-01
-9.25958633e-01 -8.56920540e-01 -4.09435123e-01 -2.39805192e-01
-1.81370723e+00 -4.27930683e-01 4.87440169e-01 -2.70840645e-01
3.77691954e-01 4.58947152e-01 5.40530980e-01 1.36354387e+00
5.74654937e-02 -8.23078156e-02 1.09375465e+00 -3.60451818e-01
4.31798041e-01 2.51809895e-01 -4.50326828e-03 2.17339665e-01
9.21326280e-01 -2.98283771e-02 -6.53325796e-01 -7.51926303e-01
6.33393586e-01 2.77394354e-01 -3.87859404e-01 -3.83038700e-01
-1.05275393e+00 1.36481464e-01 5.60278952e-01 6.25730634e-01
-4.28101003e-01 1.36329725e-01 -2.33623251e-01 -4.71169017e-02
-2.55399775e-02 2.46222273e-01 -2.82467097e-01 -6.65648520e-01
-1.02544522e+00 -5.37384599e-02 7.42952049e-01 9.85189378e-01
8.32523346e-01 -2.82450944e-01 1.92459971e-01 8.61789584e-01
6.45410836e-01 1.18000519e+00 3.87545764e-01 -8.38348329e-01
6.65002584e-01 2.80366659e-01 5.30168891e-01 -1.66483772e+00
-4.79768306e-01 -5.23799360e-01 -7.86036134e-01 -5.40888488e-01
5.54475725e-01 -5.35544574e-01 -1.88593641e-01 1.43188119e+00
1.93759859e-01 4.84436035e-01 -5.12771010e-01 5.49761236e-01
4.24111366e-01 3.05240631e-01 -3.42802852e-01 1.21617913e-01
1.00202703e+00 -3.85817826e-01 -3.44350815e-01 -7.27005228e-02
4.33233410e-01 -7.00720251e-01 6.31369054e-01 1.54302225e-01
-4.77732658e-01 -4.96783465e-01 -1.18021178e+00 5.82792640e-01
-5.08213758e-01 5.17779514e-02 3.84772807e-01 1.48548806e+00
-9.91644800e-01 2.02556610e-01 -8.33326221e-01 -7.61837602e-01
7.47790262e-02 2.59777725e-01 -2.03650385e-01 -1.56808451e-01
-1.16542006e+00 5.51300049e-01 -2.79344827e-01 2.70887345e-01
-9.58518591e-03 -2.64001787e-01 -6.48888290e-01 -6.96205944e-02
-1.25438318e-01 -4.10509229e-01 9.28456068e-01 -1.48381040e-01
-1.27948678e+00 1.13817878e-01 -4.58077937e-01 -2.62735814e-01
2.07395688e-01 1.55156836e-01 -1.09362519e+00 -3.16128671e-01
6.61138892e-01 -1.31309405e-01 -9.43178404e-03 -9.98581529e-01
-6.95231974e-01 -4.75930005e-01 -1.01565830e-01 -2.98024207e-01
-1.96855888e-01 -3.43667537e-01 -3.42675269e-01 -4.67291474e-01
5.10602713e-01 -1.21462715e+00 -6.51139319e-02 -4.65876371e-01
-4.63372082e-01 4.58212823e-01 3.44702184e-01 -2.95059741e-01
1.45395303e+00 -2.00486946e+00 -9.82703805e-01 8.94881308e-01
-1.85456976e-01 -8.04458112e-02 3.48641545e-01 7.68539727e-01
5.04059970e-01 3.01959664e-01 4.95041050e-02 -2.61623442e-01
-3.05019021e-02 1.74470037e-01 -4.94086891e-02 5.74154854e-01
-6.22040749e-01 3.90026152e-01 -9.36417639e-01 -1.54739052e-01
3.00057113e-01 4.78815347e-01 -9.91500542e-02 -3.57971489e-01
6.42707944e-01 6.62944674e-01 -7.64755964e-01 9.93388295e-01
1.10381234e+00 -2.34547686e-02 2.35967129e-01 2.55712271e-01
-6.14526331e-01 4.73806739e-01 -1.36077130e+00 1.38291812e+00
-4.64207172e-01 6.71979010e-01 -1.17486157e-01 -6.34751976e-01
1.07058394e+00 3.05960774e-01 6.05715752e-01 -1.01819837e+00
5.27389869e-02 5.93064904e-01 -5.94594061e-01 -3.14418793e-01
6.58781350e-01 4.50200737e-01 -3.48070800e-01 3.37088227e-01
-5.06411850e-01 4.91532534e-01 1.05813928e-01 -1.36781111e-01
1.21956182e+00 -1.59763709e-01 1.59882635e-01 -1.21964373e-01
4.47850108e-01 -1.88544691e-01 5.03734171e-01 1.07766426e+00
-4.75990139e-02 3.40953976e-01 -1.58799678e-01 -1.86325740e-02
-3.49985480e-01 -1.04414070e+00 -5.61456680e-01 5.00012815e-01
6.82171464e-01 -4.77842510e-01 -2.86921889e-01 -7.59851858e-02
4.42672521e-01 1.63357675e-01 -3.89475793e-01 5.12819588e-01
-2.41622820e-01 -5.50036311e-01 1.08957219e+00 2.15597883e-01
1.01847064e+00 -2.06591442e-01 -2.23571748e-01 2.94650674e-01
-6.55992508e-01 -8.81780982e-01 -2.49262467e-01 -2.62325644e-01
-7.14717507e-01 -1.18305540e+00 -7.89845586e-01 -2.88505375e-01
4.42354023e-01 9.88206863e-01 6.24973893e-01 5.64578436e-02
4.15230244e-01 3.92919540e-01 -2.44903937e-01 -2.79067010e-01
4.10776794e-01 4.23436463e-01 1.98637083e-01 2.17456445e-01
2.40702257e-01 -7.84886003e-01 -7.89915085e-01 1.05687213e+00
-2.31031448e-01 -6.78830445e-01 2.62194276e-01 1.44294485e-01
3.88755143e-01 7.30183646e-02 3.66547376e-01 -3.10538441e-01
5.31202734e-01 -1.04523385e+00 -7.05165267e-01 1.43182175e-02
-7.58205652e-01 -6.16995156e-01 3.27459544e-01 2.52356350e-01
-6.86346412e-01 -3.01796407e-01 -9.32582170e-02 6.41036093e-01
-3.97380888e-01 6.76956892e-01 -3.86216521e-01 -6.13889039e-01
6.60786629e-01 8.34691375e-02 -3.16422015e-01 -6.34252667e-01
-2.69049019e-01 1.22170961e+00 5.22855341e-01 -4.36603189e-01
8.75672102e-01 4.89491612e-01 -3.78438681e-02 -1.07085276e+00
-2.33221233e-01 -6.01463377e-01 -2.02435657e-01 -2.00970814e-01
1.17595784e-01 -1.04304647e+00 -8.37886691e-01 5.22902787e-01
-7.53717482e-01 7.96272457e-02 5.48619330e-01 7.01886714e-01
1.68788463e-01 3.69546384e-01 -1.28495678e-01 -8.96182418e-01
1.41205043e-01 -7.42992699e-01 7.21880317e-01 7.51492441e-01
-3.27744186e-01 -7.45554984e-01 2.55473405e-01 4.80216563e-01
1.01494277e+00 5.83940685e-01 -6.60699531e-02 2.77907941e-02
-6.11569405e-01 -7.19076157e-01 -6.77516162e-02 -3.28892052e-01
5.10291576e-01 -9.03623104e-02 -7.81240463e-01 -3.32441986e-01
-1.78271204e-01 4.54928100e-01 1.88890975e-02 4.67652678e-01
9.08026636e-01 -1.34297922e-01 -8.17198157e-01 6.97911799e-01
1.37671280e+00 4.78131890e-01 8.54875028e-01 9.98537779e-01
2.88621634e-01 -9.28605720e-02 6.04404569e-01 5.00597656e-01
8.30156922e-01 1.07426727e+00 3.44177932e-01 1.28733158e-01
3.24232876e-01 -2.92609185e-01 4.46460098e-02 2.92644858e-01
-3.74312103e-01 -4.56022948e-01 -1.06392217e+00 4.88983959e-01
-1.44742632e+00 -1.02500677e+00 -5.43417335e-01 2.64299154e+00
1.10396214e-01 -4.36863936e-02 2.08168656e-01 2.86237091e-01
9.41817343e-01 8.69498402e-02 -1.43714935e-01 2.76694372e-02
-1.37730807e-01 -9.39165503e-02 1.06552970e+00 3.57024610e-01
-8.88012230e-01 7.65588731e-02 6.11879539e+00 4.38698947e-01
-1.45147300e+00 3.84698287e-02 2.84452707e-01 2.25375503e-01
-3.54095012e-01 -3.29519100e-02 -7.64191687e-01 1.24577188e+00
1.12501550e+00 1.16562217e-01 3.48652512e-01 8.48915100e-01
5.41552544e-01 -7.25027978e-01 -3.57727557e-01 1.04571140e+00
-2.86253512e-01 -1.39108551e+00 -6.70027375e-01 5.12890577e-01
6.57742262e-01 4.74239647e-01 -1.18039092e-02 -1.09575897e-01
-2.51957208e-01 -5.63121200e-01 5.80955565e-01 5.90466380e-01
6.30783081e-01 -7.25691438e-01 9.06994760e-01 2.33749390e-01
-1.41895127e+00 -3.97306159e-02 -4.26169597e-02 -4.06535983e-01
3.08860630e-01 7.22438157e-01 -5.96387506e-01 8.39541793e-01
8.13054740e-01 4.03245986e-01 -5.58328331e-01 1.63009560e+00
-1.11977331e-01 7.69966185e-01 -8.24915349e-01 -2.12425217e-01
6.48872107e-02 -2.90471822e-01 1.39668792e-01 7.40192413e-01
1.18718266e+00 -3.02752793e-01 -1.82129472e-01 2.87255645e-01
-1.51619501e-02 -1.88185155e-01 -8.04870903e-01 4.20435488e-01
1.36212742e+00 1.00702727e+00 -5.20194113e-01 2.41975874e-01
-3.98857474e-01 4.16541845e-01 -5.46811998e-01 6.30626619e-01
-9.09423172e-01 -8.20533156e-01 5.79009056e-01 5.68543553e-01
8.55596829e-03 -7.72499382e-01 -4.61145908e-01 -8.96961808e-01
2.01646879e-01 -3.03454131e-01 -1.48984015e-01 -5.93135774e-01
-7.06214070e-01 1.88523471e-01 -2.71446019e-01 -1.46102774e+00
-9.95549336e-02 -6.14121370e-02 -5.07197678e-01 1.15174508e+00
-1.23393142e+00 -6.67366982e-01 -9.55493212e-01 5.55392504e-01
-4.65904742e-01 -5.25736175e-02 9.41786528e-01 9.93584514e-01
-6.05665565e-01 7.82094836e-01 9.72259223e-01 1.99435234e-01
6.31833494e-01 -7.08710670e-01 5.44734001e-01 7.44967401e-01
7.25272223e-02 1.13841426e+00 4.24206942e-01 -5.75194061e-01
-1.31074643e+00 -8.95851433e-01 1.08079541e+00 -3.91955256e-01
4.32994217e-01 -3.33948165e-01 -1.21759973e-01 5.59947610e-01
-3.10879111e-01 -3.86477378e-03 1.01023650e+00 2.36463144e-01
1.88531458e-01 -6.73466027e-01 -1.44129694e+00 4.41674441e-01
7.80015826e-01 -4.73308414e-01 1.29008800e-01 -1.88977346e-01
-1.32991463e-01 -3.99708599e-01 -9.68957663e-01 1.78399161e-01
1.05915153e+00 -1.12127459e+00 9.15702224e-01 7.46354699e-01
-3.76755714e-01 -6.47436142e-01 -3.02716881e-01 -1.06253946e+00
-2.57707477e-01 -5.00865161e-01 1.56695023e-01 1.66740739e+00
3.67197186e-01 -1.41706705e+00 9.62938607e-01 6.33147895e-01
2.58548707e-01 -3.96203220e-01 -1.17068517e+00 -8.91142011e-01
-6.82273567e-01 -4.74019736e-01 1.34016585e+00 8.25187087e-01
-9.62563902e-02 -1.49364369e-02 -4.00135040e-01 4.88774419e-01
5.37561357e-01 -5.58355413e-02 1.22718883e+00 -1.39466918e+00
4.54259105e-02 -1.29537389e-01 -5.57805538e-01 -1.35099494e+00
-7.14521945e-01 -3.92866939e-01 -2.15955228e-01 -1.46140242e+00
-4.97851968e-01 -1.16306031e+00 -2.59048730e-01 3.19582641e-01
5.72329104e-01 5.96598506e-01 -3.52339894e-01 3.57660830e-01
-4.20638144e-01 1.22962348e-01 5.51596403e-01 -7.77279809e-02
-2.65337497e-01 7.01198578e-01 -8.20812166e-01 4.21504915e-01
1.05283022e+00 -5.74523509e-01 -2.67813236e-01 -5.74944377e-01
3.37168574e-01 2.03076512e-01 3.05978507e-01 -1.48648882e+00
3.56350929e-01 -8.68866593e-02 5.87320685e-01 -6.16026938e-01
2.60589272e-01 -9.71673369e-01 7.34147310e-01 4.31513727e-01
4.83183682e-01 -4.02133279e-02 5.10670096e-02 6.21544242e-01
-1.68687716e-01 5.37526608e-02 1.72295123e-01 3.54260504e-01
-3.94799829e-01 1.16640173e-01 -6.05209947e-01 -3.99862498e-01
8.32542121e-01 -8.90628278e-01 -6.54304624e-01 -5.35632730e-01
-7.80823976e-02 1.53041586e-01 8.98896277e-01 2.41484061e-01
1.20836496e-01 -1.47589183e+00 4.09892239e-02 1.87873766e-01
2.13110209e-01 -4.63549405e-01 2.32504040e-01 1.06616509e+00
-6.42292023e-01 8.08510125e-01 -2.51990020e-01 -4.99575883e-01
-1.00696301e+00 -2.42172390e-01 3.58019531e-01 2.75786966e-01
-2.59677283e-02 4.63958383e-01 -8.89886141e-01 -5.17200708e-01
2.97663271e-01 -3.78068924e-01 3.16559523e-02 -8.34668875e-02
4.13997203e-01 9.05272901e-01 2.33269036e-01 -8.98975968e-01
-9.77111936e-01 8.02896202e-01 8.04647028e-01 -1.12721235e-01
8.31565559e-01 -6.92337036e-01 6.33816747e-03 2.53612906e-01
8.59112203e-01 7.23882377e-01 -5.86208045e-01 -5.07343560e-02
3.49831730e-01 -8.95068228e-01 -3.71836036e-01 -6.05415702e-01
-7.19947815e-01 1.76632881e-01 8.68602455e-01 7.35217929e-01
7.24160433e-01 -4.29830283e-01 8.68673921e-01 3.79463971e-01
1.31215096e+00 -8.25083673e-01 -7.21765757e-01 5.17912269e-01
1.82400227e-01 -1.11648834e+00 -1.95314005e-01 -1.73552334e-01
3.17226470e-01 8.43885005e-01 2.04493672e-01 1.65976197e-01
8.06933045e-01 4.76214625e-02 1.03112444e-01 1.91448539e-01
3.65423352e-01 3.64096165e-02 -2.08581775e-01 9.02478576e-01
5.37895441e-01 1.00506887e-01 -5.33121526e-01 6.90319717e-01
-4.00149643e-01 2.38132223e-01 4.56075996e-01 1.12022972e+00
-3.76427472e-01 -1.46775639e+00 -7.95544267e-01 5.33754408e-01
-4.42183554e-01 2.88061559e-01 -1.20954156e-01 6.93008780e-01
3.88347030e-01 1.48775613e+00 -2.69016802e-01 -5.40795088e-01
3.46799791e-01 -4.33006793e-01 1.14687957e-01 -1.45115837e-01
-1.90164819e-01 -5.25903642e-01 3.77690673e-01 -6.07281685e-01
-2.93229133e-01 -6.73658073e-01 -9.94224787e-01 -6.76578641e-01
-3.29809219e-01 6.19890511e-01 1.16432738e+00 7.16032922e-01
6.88886285e-01 -4.45482358e-02 9.63644326e-01 -9.86514449e-01
1.49571568e-01 -7.92667091e-01 -8.98581743e-01 -2.15945184e-01
1.93153679e-01 -8.35627973e-01 -1.62639201e-01 -6.94844306e-01] | [6.375899791717529, 0.9599154591560364] |
4477ee06-af68-4e2a-922c-399e52be245f | rucola-russian-corpus-of-linguistic | 2210.12814 | null | https://arxiv.org/abs/2210.12814v1 | https://arxiv.org/pdf/2210.12814v1.pdf | RuCoLA: Russian Corpus of Linguistic Acceptability | Linguistic acceptability (LA) attracts the attention of the research community due to its many uses, such as testing the grammatical knowledge of language models and filtering implausible texts with acceptability classifiers. However, the application scope of LA in languages other than English is limited due to the lack of high-quality resources. To this end, we introduce the Russian Corpus of Linguistic Acceptability (RuCoLA), built from the ground up under the well-established binary LA approach. RuCoLA consists of $9.8$k in-domain sentences from linguistic publications and $3.6$k out-of-domain sentences produced by generative models. The out-of-domain set is created to facilitate the practical use of acceptability for improving language generation. Our paper describes the data collection protocol and presents a fine-grained analysis of acceptability classification experiments with a range of baseline approaches. In particular, we demonstrate that the most widely used language models still fall behind humans by a large margin, especially when detecting morphological and semantic errors. We release RuCoLA, the code of experiments, and a public leaderboard (rucola-benchmark.com) to assess the linguistic competence of language models for Russian. | ['Ekaterina Artemova', 'Ivan Smurov', 'Alena Pestova', 'Max Ryabinin', 'Tatiana Shamardina', 'Vladislav Mikhailov'] | 2022-10-23 | null | null | null | null | ['linguistic-acceptability'] | ['natural-language-processing'] | [-5.89644313e-02 2.86036134e-01 1.08968183e-01 -8.09630990e-01
-1.37387145e+00 -8.44685853e-01 4.02162135e-01 4.49022204e-01
-5.15797257e-01 7.94229448e-01 2.32338414e-01 -7.85233617e-01
2.48001050e-02 -7.80922174e-01 -7.56215215e-01 -1.47754729e-01
3.35569650e-01 5.07312834e-01 -1.71599999e-01 -4.72026318e-01
3.22078407e-01 4.47224900e-02 -1.36513412e+00 6.99223936e-01
1.45799959e+00 8.31759155e-01 8.11647028e-02 4.18880373e-01
-5.79252467e-02 5.63376904e-01 -8.05024326e-01 -1.08402240e+00
-2.19936728e-01 -2.99108803e-01 -9.22060251e-01 -3.13683838e-01
3.46447736e-01 8.32271799e-02 3.63106459e-01 1.07758188e+00
4.21081245e-01 2.28223968e-02 7.23747790e-01 -8.85219693e-01
-1.03511310e+00 1.06559265e+00 1.82573065e-01 6.69589713e-02
7.93319285e-01 2.39243105e-01 1.24455261e+00 -1.04985046e+00
7.87326455e-01 1.67276740e+00 6.27441168e-01 6.70882225e-01
-1.05864525e+00 -5.05420804e-01 1.57900110e-01 -4.26875427e-02
-1.16701531e+00 -5.13079524e-01 5.35363019e-01 -5.49596190e-01
1.40083098e+00 3.47180337e-01 4.11947280e-01 1.45247102e+00
2.21769035e-01 6.29067302e-01 1.34757471e+00 -7.72891641e-01
2.66568601e-01 5.07140160e-01 4.58935320e-01 5.95684588e-01
2.74784654e-01 5.36218174e-02 -5.28799295e-01 -3.13198954e-01
-2.10029915e-01 -1.04629195e+00 -1.07724816e-01 1.30016759e-01
-6.33405566e-01 1.04680562e+00 -3.13675925e-02 1.86547130e-01
7.07982406e-02 -4.02431250e-01 5.12363195e-01 5.49179316e-01
7.55270362e-01 6.97048724e-01 -5.69804251e-01 -3.38612109e-01
-5.54871500e-01 3.88231128e-01 7.96460867e-01 8.71667325e-01
2.69291133e-01 1.32008508e-01 2.10450828e-01 1.19965541e+00
5.73203266e-01 7.99149871e-01 5.43053389e-01 -8.23524952e-01
6.45489693e-01 7.37949848e-01 -8.85943398e-02 -7.27234006e-01
-2.47732103e-01 9.86339711e-03 -3.28473449e-01 2.16600567e-01
6.92044556e-01 -9.57836732e-02 -5.07022560e-01 1.92399943e+00
4.15777713e-02 -1.02545369e+00 4.31427509e-01 3.24130625e-01
1.06439686e+00 7.61072397e-01 4.08887327e-01 -2.94580609e-01
1.22480059e+00 -5.71264923e-01 -4.64582056e-01 -4.66249377e-01
1.26416111e+00 -9.35715616e-01 1.84553754e+00 5.61712444e-01
-1.38023221e+00 -5.34494221e-01 -9.58532631e-01 -5.82417212e-02
-6.00226343e-01 1.88311964e-01 5.54259539e-01 1.16160309e+00
-9.30784702e-01 2.01565981e-01 -3.35400313e-01 -3.01157504e-01
1.67389780e-01 2.47921556e-01 -4.74735945e-01 -1.03009693e-01
-1.55479705e+00 1.25917101e+00 3.14921081e-01 -5.86417085e-03
-4.51542467e-01 -3.18444908e-01 -1.31501818e+00 -3.50423783e-01
1.59962580e-01 -2.45398074e-01 1.29886115e+00 -1.12696469e+00
-1.46842420e+00 1.33753145e+00 -1.51759282e-01 -2.04410493e-01
3.33293438e-01 -4.51844782e-01 -7.41988599e-01 -4.30935919e-01
2.17563957e-01 2.61343986e-01 3.22151303e-01 -8.90473545e-01
-5.56764960e-01 -3.66643906e-01 2.30878681e-01 8.63316506e-02
-1.55535519e-01 5.88347614e-01 3.78082716e-03 -7.19589591e-01
-8.00341442e-02 -1.01594687e+00 4.68649305e-02 -8.13480258e-01
-3.94846559e-01 -7.30324566e-01 1.70524478e-01 -8.28069627e-01
1.61956084e+00 -2.16888356e+00 -1.30059049e-01 2.61979997e-01
-1.64604560e-01 2.53207564e-01 -1.04763009e-01 3.67884219e-01
-7.34370947e-02 6.35474384e-01 -3.63383561e-01 -3.11457068e-01
1.17596015e-01 -1.43142138e-02 -4.29463357e-01 3.05478185e-01
2.37566456e-01 7.79366314e-01 -8.46050560e-01 -4.40070003e-01
9.47408229e-02 7.66905099e-02 -6.14561498e-01 1.69131175e-01
-1.89343065e-01 1.52185813e-01 -3.00656527e-01 7.00847328e-01
4.21086848e-01 3.41438353e-01 1.84869796e-01 2.74407536e-01
-1.65713415e-01 1.03789461e+00 -8.85631084e-01 1.33913946e+00
-7.85713673e-01 4.67234015e-01 -2.38830999e-01 -6.30672991e-01
9.63570833e-01 8.22221115e-02 -4.87702668e-01 -8.07261229e-01
4.47234139e-02 8.21878552e-01 2.89572835e-01 -3.37410003e-01
7.59734333e-01 8.23089574e-03 -7.90524960e-01 2.78043628e-01
-9.63571593e-02 -6.24491453e-01 4.13523257e-01 2.57910520e-01
8.38255227e-01 7.52082989e-02 5.45935333e-01 -6.12940431e-01
7.75250733e-01 9.54351574e-02 5.15107393e-01 7.55094469e-01
-2.17358232e-01 3.29186171e-01 5.26002347e-01 -2.05061615e-01
-7.68951774e-01 -1.13168895e+00 -5.00955224e-01 1.34192622e+00
-1.60512432e-01 -5.83939075e-01 -1.06371248e+00 -9.32601154e-01
-2.68575400e-01 1.39408052e+00 -4.37140107e-01 -1.69364110e-01
-6.02734566e-01 -7.41250634e-01 7.00112760e-01 4.34639484e-01
2.24529766e-02 -1.61156225e+00 -3.00212979e-01 1.16136268e-01
-3.42407048e-01 -9.07145441e-01 7.05257617e-03 8.03148188e-03
-3.93487781e-01 -1.03429270e+00 5.71298897e-02 -7.48996198e-01
4.57847714e-01 -5.78723609e-01 1.50871611e+00 1.27982855e-01
2.00720146e-01 4.60522212e-02 -5.52021444e-01 -5.71109354e-01
-1.33803797e+00 3.30607034e-02 1.15575179e-01 -5.76112568e-01
6.30882084e-01 8.09776224e-03 1.59764066e-01 4.71420512e-02
-6.46387577e-01 -3.43224943e-01 3.36151630e-01 8.00586879e-01
3.19409281e-01 -2.90596306e-01 8.07699859e-01 -1.45766187e+00
1.09524417e+00 -2.46903002e-01 -5.98972082e-01 2.96254694e-01
-5.42784750e-01 -7.33141825e-02 7.41100430e-01 -4.76809256e-02
-1.39666545e+00 -2.45661438e-01 -8.07736695e-01 6.45000935e-01
-3.40073109e-01 6.64732814e-01 -6.28636241e-01 3.25876951e-01
1.01222551e+00 -3.03605974e-01 -3.22581589e-01 -3.44708890e-01
4.79157716e-01 1.11422527e+00 3.44504446e-01 -8.16437781e-01
3.67824316e-01 -1.97303891e-01 -6.43706739e-01 -9.34963584e-01
-9.13709402e-01 -5.12387007e-02 -3.68536770e-01 -4.23654988e-02
5.17171085e-01 -7.99099684e-01 -3.97658229e-01 6.34636283e-01
-1.33045936e+00 -4.26954627e-01 -1.38834998e-01 3.03964645e-01
-5.27716815e-01 3.29586148e-01 -8.65489662e-01 -9.30995524e-01
-4.11130250e-01 -1.35658002e+00 9.02731776e-01 -2.80242652e-01
-9.34586763e-01 -9.27109122e-01 1.07175529e-01 4.03754264e-01
8.32804367e-02 -8.36746991e-02 1.60638702e+00 -7.60345817e-01
1.13015689e-01 -1.16924010e-01 2.20797598e-01 7.08494544e-01
-3.87372077e-01 2.50401378e-01 -1.10381758e+00 -6.90838546e-02
2.76137851e-02 -9.12804127e-01 6.37423158e-01 3.63727391e-01
9.24267232e-01 -2.25487128e-01 1.00699760e-01 2.27917492e-01
9.92083967e-01 3.06100070e-01 6.81940377e-01 4.56982732e-01
3.31667721e-01 1.02860701e+00 8.56277943e-01 9.67021883e-02
5.22335470e-01 5.92689097e-01 8.03886503e-02 2.22729415e-01
9.48666185e-02 -2.82552540e-01 8.45678687e-01 1.16583979e+00
2.30542853e-01 -5.18924177e-01 -1.10509586e+00 5.39950073e-01
-1.49926996e+00 -6.58376575e-01 -3.50542694e-01 2.29391575e+00
1.08042300e+00 4.99540597e-01 -1.80091430e-02 2.66062498e-01
5.25206268e-01 -2.45667361e-02 8.08650702e-02 -1.30881608e+00
-5.23422360e-01 1.48140490e-01 -1.90617636e-01 9.54374969e-01
-9.89267051e-01 1.29404759e+00 6.12707376e+00 1.14263237e+00
-6.81728780e-01 -8.58692080e-03 9.98159587e-01 3.84197116e-01
-7.04495609e-01 -4.29426953e-02 -1.05170977e+00 4.15703565e-01
1.30663121e+00 -2.25947440e-01 1.69463232e-01 9.91540968e-01
1.97299615e-01 -3.38765562e-01 -1.26440525e+00 6.85652733e-01
2.52907336e-01 -7.93064654e-01 9.08684954e-02 -2.50826031e-01
5.30630708e-01 -8.38489532e-02 1.79356992e-01 7.26908803e-01
3.80590379e-01 -1.17777801e+00 1.17837870e+00 1.25664109e-02
9.93473291e-01 -8.31777632e-01 8.98979664e-01 3.87082636e-01
-7.06343412e-01 -1.24956910e-02 -4.29287493e-01 -2.97880709e-01
1.00064285e-01 8.66146624e-01 -5.96405566e-01 1.62746429e-01
8.45760763e-01 3.02907228e-01 -8.79732430e-01 2.21958101e-01
-5.47582150e-01 1.08859754e+00 -2.56972969e-01 -3.25990260e-01
4.01967466e-02 -2.83181489e-01 4.14064944e-01 1.45365107e+00
2.02517271e-01 -4.69442084e-02 -8.28078464e-02 8.64830554e-01
-1.81454107e-01 7.01103449e-01 -8.83715987e-01 -1.19553931e-01
5.55310249e-01 9.40784931e-01 -2.94395328e-01 -3.40572268e-01
-4.25155580e-01 5.19687772e-01 7.12093532e-01 2.80103926e-02
-5.35860956e-01 -4.10003781e-01 3.74982566e-01 5.81212901e-02
-3.21057141e-01 -1.13613028e-02 -5.64912498e-01 -8.67419302e-01
2.57510513e-01 -1.31503475e+00 3.65192801e-01 -7.00477242e-01
-1.48124421e+00 1.01605976e+00 -5.42658344e-02 -7.51862705e-01
-6.33108675e-01 -8.70371819e-01 -2.64080703e-01 1.17332208e+00
-1.02841926e+00 -1.02735662e+00 1.17004834e-01 2.24065661e-01
8.22704852e-01 -1.96873337e-01 1.06617284e+00 -4.75486219e-02
-4.19690579e-01 8.35346699e-01 -1.09098591e-01 -3.23809460e-02
7.87008941e-01 -1.19092011e+00 7.71733701e-01 8.69368553e-01
1.41835054e-02 8.97728443e-01 6.97731256e-01 -8.86149228e-01
-8.39061856e-01 -9.83385205e-01 1.71527696e+00 -1.04011095e+00
6.73409641e-01 -6.36375010e-01 -1.01274323e+00 6.42456293e-01
1.35970443e-01 -4.68587458e-01 7.21241832e-01 4.39773619e-01
-3.47351015e-01 1.31902844e-01 -1.10466480e+00 7.70861864e-01
1.01069486e+00 -7.35296965e-01 -9.41497684e-01 2.89050400e-01
5.77443540e-01 -3.96463752e-01 -6.40138566e-01 5.15002251e-01
4.43676442e-01 -1.09210300e+00 5.03996432e-01 -5.36293149e-01
7.64489114e-01 1.65041611e-01 -3.07532698e-01 -1.58181977e+00
6.77788183e-02 -4.03403848e-01 4.31772768e-01 1.43279159e+00
1.16467488e+00 -7.33306587e-01 2.72034615e-01 8.67565691e-01
-4.29642111e-01 -7.49955416e-01 -8.96293759e-01 -6.37770355e-01
6.43445134e-01 -1.00416160e+00 3.94957215e-01 6.30227149e-01
2.94529587e-01 5.61093926e-01 1.81313068e-01 -7.60059059e-02
2.01029018e-01 6.30151480e-02 5.59085786e-01 -9.92328942e-01
-2.91011781e-01 -3.32781523e-01 -2.01456830e-01 -7.05881715e-01
6.46064401e-01 -9.77897167e-01 3.37093651e-01 -1.17893302e+00
6.29205927e-02 -6.38448536e-01 1.97823286e-01 2.35270396e-01
-4.15962279e-01 5.96750937e-02 9.79183614e-02 -1.52456030e-01
-4.23972249e-01 4.78605658e-01 5.67399204e-01 3.99380326e-02
-2.67170072e-01 -1.30071506e-01 -8.60931218e-01 1.15620768e+00
9.63150144e-01 -4.19839621e-01 -3.99550140e-01 -4.40045893e-01
5.69026589e-01 -2.93538481e-01 -3.99490669e-02 -7.66834974e-01
-4.16747421e-01 -8.86899084e-02 1.07250169e-01 -2.15755805e-01
1.64707340e-02 -1.73569158e-01 -1.73246190e-01 2.33694226e-01
-5.58694124e-01 5.29934168e-01 2.84612507e-01 5.42842522e-02
-2.69299567e-01 -5.45918226e-01 7.43968010e-01 -1.38147071e-01
-6.98505938e-01 -2.98754126e-01 -6.33530676e-01 7.29054034e-01
7.45089769e-01 -9.35219005e-02 -5.67869723e-01 -2.98572034e-01
-5.26413560e-01 -2.54520357e-01 6.52181029e-01 6.61866426e-01
5.03167808e-01 -7.97539949e-01 -7.70045400e-01 3.93130779e-01
4.89998490e-01 -2.40287021e-01 8.03365409e-02 4.52016979e-01
-5.84456623e-01 4.32876468e-01 1.36808053e-01 -2.63513088e-01
-1.24899054e+00 2.98053682e-01 9.22272950e-02 -1.71311960e-01
-1.98184848e-01 9.16725934e-01 3.40807408e-01 -8.41555834e-01
-1.99236330e-02 -3.94411564e-01 -2.54047036e-01 -3.17470998e-01
3.87933999e-01 3.04174334e-01 4.60386992e-01 -1.01586163e+00
-4.01941448e-01 9.86485258e-02 3.29729393e-02 -2.89417654e-01
8.44457388e-01 -9.25577730e-02 -2.43783176e-01 6.63258195e-01
8.93100142e-01 6.71220183e-01 -3.23377132e-01 2.59331644e-01
2.65972137e-01 -1.73707426e-01 -3.46302271e-01 -1.17731726e+00
-3.42128038e-01 8.99805188e-01 1.57773063e-01 1.18489966e-01
7.72536337e-01 1.39559165e-01 5.27064204e-01 5.37304163e-01
4.98803049e-01 -1.42262495e+00 -3.29417765e-01 8.69826913e-01
1.08617949e+00 -1.25050366e+00 -4.20783371e-01 -8.38067472e-01
-8.42161775e-01 6.39000952e-01 7.10604608e-01 1.99328780e-01
3.90848458e-01 1.54608771e-01 4.86278296e-01 1.62430666e-02
-7.50253022e-01 2.01319709e-01 1.14161164e-01 6.69040740e-01
9.80847836e-01 3.85987163e-01 -8.94700289e-01 1.15187263e+00
-7.73382068e-01 -4.82835352e-01 6.76240146e-01 5.48974156e-01
-3.22756708e-01 -1.32630825e+00 -2.01714516e-01 5.02374172e-01
-7.43571699e-01 -6.54312491e-01 -6.27628863e-01 8.31196904e-01
3.93076874e-02 1.39815462e+00 -3.35273176e-01 -2.63577342e-01
6.25963926e-01 2.12281182e-01 1.47690073e-01 -8.13611090e-01
-7.86518574e-01 -2.15610161e-01 8.21347356e-01 -5.07059753e-01
-8.39685798e-02 -7.70151198e-01 -1.07224524e+00 -3.08578432e-01
-3.00897598e-01 3.46204847e-01 6.75288856e-01 9.28104639e-01
9.31445435e-02 3.40046249e-02 3.12665164e-01 -2.48536214e-01
-7.31387556e-01 -1.09429824e+00 -6.08590245e-01 6.93904161e-01
-2.47102186e-01 -3.26380253e-01 -6.67454004e-01 -1.19447827e-01] | [10.760862350463867, 9.598322868347168] |
452c5896-1fa4-42d2-bf43-d6b9e24ba53a | federated-multi-view-learning-for-private | 2105.01603 | null | https://arxiv.org/abs/2105.01603v1 | https://arxiv.org/pdf/2105.01603v1.pdf | Federated Multi-View Learning for Private Medical Data Integration and Analysis | Along with the rapid expansion of information technology and digitalization of health data, there is an increasing concern on maintaining data privacy while garnering the benefits in medical field. Two critical challenges are identified: Firstly, medical data is naturally distributed across multiple local sites, making it difficult to collectively train machine learning models without data leakage. Secondly, in medical applications, data are often collected from different sources and views, resulting in heterogeneity and complexity that requires reconciliation. This paper aims to provide a generic Federated Multi-View Learning (FedMV) framework for multi-view data leakage prevention, which is based on different types of local data availability and enables to accommodate two types of problems: Vertical Federated Multi-View Learning (V-FedMV) and Horizontal Federated Multi-View Learning (H-FedMV). We experimented with real-world keyboard data collected from BiAffect study. The results demonstrated that the proposed FedMV approach can make full use of multi-view data in a privacy-preserving way, and both V-FedMV and H-FedMV methods perform better than their single-view and pairwise counterparts. Besides, the proposed model can be easily adapted to deal with multi-view sequential data in a federated environment, which has been modeled and experimentally studied. To the best of our knowledge, this framework is the first to consider both vertical and horizontal diversification in the multi-view setting, as well as their sequential federated learning. | ['Lifang He', 'Yong Chen', 'Lichao Sun', 'Hao Peng', 'Sicong Che'] | 2021-05-04 | null | null | null | null | ['multi-view-learning', 'data-integration'] | ['computer-vision', 'knowledge-base'] | [-9.04686525e-02 -2.62469407e-02 -3.96250606e-01 -3.24491620e-01
-8.08697045e-01 -7.52571106e-01 3.33332151e-01 5.97376943e-01
-4.07001644e-01 5.37754476e-01 4.29946601e-01 -3.85876745e-01
-3.71402174e-01 -7.52793729e-01 -4.83124763e-01 -8.36597860e-01
-1.88300550e-01 9.20499563e-02 -9.37329680e-02 1.39981493e-01
-1.51476696e-01 3.87440652e-01 -1.13854611e+00 6.22492194e-01
6.67875171e-01 7.90498853e-01 -7.91799352e-02 3.59280556e-01
4.14567851e-02 5.15386403e-01 -1.42166883e-01 -9.38784599e-01
5.95633268e-01 -1.06416970e-01 -8.26200545e-01 -5.86941652e-02
2.62420923e-01 -3.80309552e-01 -7.69712403e-02 9.37075615e-01
8.48736763e-01 -2.62476355e-01 6.70378655e-02 -1.45388043e+00
-3.78455698e-01 2.67304480e-01 -5.59683621e-01 1.14364708e-02
1.56607941e-01 4.48665358e-02 9.39886749e-01 -5.45309067e-01
8.78449023e-01 8.30081284e-01 6.81415200e-01 5.16063392e-01
-1.17729926e+00 -6.00086629e-01 1.42401129e-01 6.13298640e-02
-1.10498607e+00 -2.69464433e-01 5.85486650e-01 -4.38827246e-01
5.35471201e-01 6.06355071e-01 6.10884666e-01 9.63521838e-01
5.07451296e-01 8.93138766e-01 1.36984813e+00 -5.84849864e-02
3.42221498e-01 4.96994525e-01 -3.86092882e-03 4.59710002e-01
4.63909060e-01 1.31100401e-01 -5.49617171e-01 -8.94302309e-01
2.53985673e-01 7.25742400e-01 -4.92980301e-01 -9.84525144e-01
-1.20856774e+00 7.43821561e-01 3.40737820e-01 2.30222449e-01
-3.66116554e-01 -5.56027174e-01 8.75721216e-01 3.82363379e-01
3.13985407e-01 -3.98811186e-03 -6.84873044e-01 2.67044663e-01
-8.06441069e-01 1.46070242e-01 9.18916821e-01 8.99082780e-01
6.06706083e-01 -5.30658245e-01 -2.05754749e-02 3.82821351e-01
1.56278983e-01 3.56463313e-01 5.30346751e-01 -5.99517882e-01
8.67429376e-01 7.48220742e-01 -3.62007841e-02 -1.30848420e+00
-4.01207447e-01 -3.98394346e-01 -1.35144508e+00 -5.89772593e-03
4.30641830e-01 -9.71860290e-02 -2.90772080e-01 1.89536250e+00
7.60147572e-01 -2.30603412e-01 3.73642772e-01 6.12218499e-01
5.75246513e-01 1.38628930e-01 -1.60537332e-01 -6.58749700e-01
1.47716439e+00 -6.28581047e-01 -9.63406622e-01 2.73120105e-01
8.58116865e-01 -4.47120994e-01 7.05740392e-01 6.28102064e-01
-7.95254171e-01 -8.41849372e-02 -8.56055975e-01 1.10200159e-01
-4.54921633e-01 -3.16323489e-01 4.56187338e-01 7.85507381e-01
-9.15596366e-01 2.79484361e-01 -8.53203833e-01 -4.71619636e-01
6.92800224e-01 2.62130529e-01 -9.47094619e-01 -4.30577338e-01
-1.10689664e+00 4.28086549e-01 2.06221461e-01 -9.69750062e-03
-7.34518051e-01 -8.21534455e-01 -6.46650553e-01 -7.85180330e-02
5.70560038e-01 -8.49086463e-01 7.30633855e-01 -5.77441394e-01
-6.51763737e-01 7.92142212e-01 -8.82076565e-03 -4.12787080e-01
8.95032287e-01 -1.13608144e-01 -5.62014461e-01 -9.49900299e-02
1.45352455e-02 -1.85530618e-01 4.84842420e-01 -1.32586586e+00
-5.35836875e-01 -1.09222758e+00 -1.50863156e-01 1.33742556e-01
-6.70409024e-01 -2.48990193e-01 -2.43876725e-01 -5.78649938e-01
-1.36284664e-01 -7.76720762e-01 -3.74204218e-01 3.41333330e-01
-3.52101237e-01 2.92723864e-01 1.01994991e+00 -8.32298219e-01
1.42913795e+00 -2.21134901e+00 1.11129835e-01 8.13208669e-02
4.44714993e-01 4.56228644e-01 1.73519686e-01 8.87057602e-01
1.09105222e-01 2.02385023e-01 -4.78025943e-01 -4.39272135e-01
-4.46716249e-01 3.61164719e-01 -1.15733020e-01 6.67414725e-01
-4.21741277e-01 6.78356051e-01 -9.03172672e-01 -5.62228024e-01
-4.32857573e-02 3.40880066e-01 -7.05470324e-01 3.10737222e-01
2.98805505e-01 5.34826517e-01 -4.93195593e-01 9.19954062e-01
9.97413814e-01 -3.14677924e-01 6.26113832e-01 -1.98320940e-01
-4.33134921e-02 -3.83931696e-01 -1.27157485e+00 1.68825042e+00
-4.84760851e-01 -1.11185193e-01 6.02818191e-01 -7.71653891e-01
5.64128876e-01 6.22748792e-01 7.64922142e-01 -5.17190754e-01
-1.57510236e-01 1.73417762e-01 -2.94353962e-01 -6.28488958e-01
2.98919499e-01 -2.28187278e-01 -1.13410667e-01 6.81143820e-01
-1.01919681e-01 6.17291331e-01 -5.06221473e-01 3.44762743e-01
9.68439817e-01 -2.48847887e-01 8.54858696e-01 -1.86731428e-01
6.59038842e-01 -3.34872961e-01 8.95970941e-01 4.15711671e-01
-3.87576520e-01 5.13895333e-01 4.06478226e-01 -6.42800152e-01
-6.55657828e-01 -9.05445218e-01 -2.71334708e-01 5.83795667e-01
8.72541964e-02 -5.91569066e-01 -4.14030343e-01 -1.26458943e+00
2.07301810e-01 2.66446263e-01 -4.85043883e-01 -2.77992859e-02
-3.78441870e-01 -8.30370784e-01 3.28190774e-01 2.66519874e-01
3.10793519e-01 -6.03280962e-01 -9.12650287e-01 1.71358019e-01
-3.51464719e-01 -8.51483643e-01 -6.32843196e-01 -5.17663658e-02
-1.03233910e+00 -1.42392707e+00 -5.00390828e-01 -2.37645462e-01
5.91937125e-01 4.87342387e-01 8.77948821e-01 -7.07418919e-02
-4.63718802e-01 4.69842613e-01 -2.57859915e-01 -4.11533892e-01
-4.06819165e-01 -1.24972463e-01 1.65174529e-01 6.02389157e-01
1.81746393e-01 -4.57815140e-01 -8.64437640e-01 4.04159606e-01
-1.42508268e+00 -1.19296886e-01 4.52776045e-01 1.18515527e+00
8.16739500e-01 1.73959918e-02 6.65128589e-01 -1.42306709e+00
4.37550664e-01 -8.12181473e-01 -4.45653439e-01 6.95784092e-01
-1.08431292e+00 -2.03728601e-01 8.57456207e-01 -6.29897192e-02
-1.01031780e+00 5.90300187e-02 2.14433800e-02 -3.85073423e-01
4.21535484e-02 4.47698742e-01 -6.09354496e-01 4.94239740e-02
2.83640653e-01 2.64970034e-01 3.18937242e-01 -6.10694230e-01
4.61891055e-01 9.52261865e-01 1.39705494e-01 -1.46973968e-01
3.98015320e-01 8.56704891e-01 -2.62559801e-02 -4.07866448e-01
-3.03767353e-01 -5.99079609e-01 -5.17268658e-01 7.60428682e-02
7.61078179e-01 -1.03014576e+00 -5.99721253e-01 6.62437022e-01
-8.17267299e-01 2.95754135e-01 -2.76978165e-01 3.77923310e-01
-3.27834934e-01 7.43689060e-01 -3.83186191e-01 -7.80455410e-01
-7.71973789e-01 -1.00303257e+00 6.52115881e-01 -1.79736033e-01
1.39971673e-01 -1.07153881e+00 2.67209917e-01 5.95341325e-01
3.83740336e-01 5.70543110e-01 8.76022995e-01 -1.01302350e+00
-4.82380152e-01 -2.74938613e-01 1.47489458e-01 3.24792027e-01
6.35732710e-01 -6.44866228e-01 -9.02829349e-01 -1.10601664e+00
6.01692379e-01 -3.70439947e-01 3.82938921e-01 7.35329390e-02
1.11909771e+00 -8.98754060e-01 -4.81248230e-01 7.77762949e-01
1.73354805e+00 -9.79218110e-02 3.17718089e-01 2.21267492e-01
6.11016035e-01 7.16569245e-01 7.19343424e-01 8.96203160e-01
6.26464427e-01 6.47309661e-01 8.05755019e-01 -2.31081367e-01
5.00627220e-01 -2.85491943e-01 -9.42761749e-02 7.88434267e-01
3.68122041e-01 -2.88947791e-01 -7.93459177e-01 7.11361647e-01
-2.11942959e+00 -9.15832698e-01 4.08258401e-02 2.67134023e+00
6.35568440e-01 -4.87111211e-01 1.41102165e-01 -2.17460841e-02
5.24656057e-01 2.47653514e-01 -6.48088574e-01 -3.74598145e-01
-1.58641599e-02 -2.86147535e-01 6.54646456e-01 -5.62043749e-02
-1.04371417e+00 3.04088145e-02 5.28922129e+00 4.05734181e-01
-1.07714951e+00 4.85703349e-01 7.89671600e-01 -3.36003363e-01
-5.41870832e-01 -1.42018721e-01 -5.12176275e-01 4.78085816e-01
7.18443990e-01 -2.97161102e-01 1.36542380e-01 7.16960967e-01
-9.74552929e-02 2.05954820e-01 -1.09329629e+00 1.09426153e+00
-9.54210199e-03 -1.30891740e+00 3.54748592e-02 5.63913405e-01
6.97922468e-01 3.81415412e-02 -1.16719073e-02 -1.96449384e-01
1.13418914e-01 -6.68101490e-01 2.00421616e-01 3.61665606e-01
6.66632652e-01 -7.99431920e-01 8.38486493e-01 7.21543133e-01
-1.09327018e+00 -4.41579968e-01 -8.15550014e-02 4.63008285e-01
1.96653739e-01 6.48213208e-01 -6.08704567e-01 1.48685980e+00
9.13731039e-01 7.14884639e-01 -4.26131725e-01 9.32197332e-01
5.50323784e-01 1.98360920e-01 -7.09552094e-02 5.01291513e-01
-1.53974921e-01 -3.86571395e-03 4.66705173e-01 8.80386710e-01
3.45807195e-01 2.52035982e-03 2.07187444e-01 2.19910547e-01
-4.15690839e-02 4.63661164e-01 -1.00493109e+00 2.63446599e-01
5.15832067e-01 1.29383373e+00 -1.91993505e-01 3.00795175e-02
-8.54422748e-01 7.82891750e-01 3.34312856e-01 8.41754209e-03
-4.16013002e-01 1.19688272e-01 8.00245523e-01 2.93989271e-01
2.13979915e-01 1.80000171e-01 -1.05065830e-01 -1.49458718e+00
2.70010680e-01 -1.22321033e+00 1.28264892e+00 5.85116260e-02
-1.65069532e+00 6.62713349e-01 -7.87213221e-02 -1.49776554e+00
-1.18595548e-01 -1.08521327e-01 -2.01434135e-01 7.14971066e-01
-1.36722469e+00 -1.24634695e+00 -2.31552441e-02 1.05681741e+00
4.69929315e-02 -2.32952625e-01 1.11852467e+00 4.00061995e-01
-4.49808866e-01 8.85759950e-01 4.41000730e-01 -1.54101804e-01
8.58082294e-01 -9.02448118e-01 -4.64512482e-02 8.59775186e-01
4.93125394e-02 7.81964660e-01 1.75217792e-01 -6.32417917e-01
-1.79231298e+00 -1.41101444e+00 7.98597574e-01 -3.69234949e-01
1.40112489e-01 -5.43205619e-01 -1.09315407e+00 7.03221798e-01
1.80877134e-01 4.84327525e-01 1.27380705e+00 -3.58394794e-02
-4.89314467e-01 -4.87748981e-01 -1.80544555e+00 1.79017201e-01
7.88601995e-01 -6.33818030e-01 -1.93375543e-01 2.29420617e-01
5.80843866e-01 -1.83585048e-01 -1.24566698e+00 5.15483260e-01
5.24624646e-01 -1.22697365e+00 8.15564930e-01 -6.84257567e-01
-9.63167921e-02 -2.29072317e-01 -4.09950644e-01 -9.94533062e-01
-9.48629901e-02 -7.44517744e-01 -2.99546242e-01 1.41865599e+00
6.26434013e-02 -1.03682101e+00 6.91611826e-01 8.26366425e-01
4.20437306e-01 -9.86859858e-01 -1.40233302e+00 -5.78590214e-01
3.93790985e-03 9.62165147e-02 8.12440217e-01 1.26192307e+00
4.38441485e-02 -4.23801750e-01 -8.01871240e-01 3.52074504e-01
8.38296294e-01 4.34251279e-01 8.04106116e-01 -1.03733730e+00
-6.11327589e-01 1.43793285e-01 -4.44701225e-01 -3.57101172e-01
-4.45747614e-01 -1.03307474e+00 -5.60298085e-01 -1.33824694e+00
5.31795502e-01 -5.21498561e-01 -6.65034533e-01 5.06188929e-01
-1.73172534e-01 -1.43342689e-01 2.46058345e-01 4.17871714e-01
-3.41695160e-01 3.46957773e-01 9.54762280e-01 6.34251684e-02
-8.44832510e-02 3.98409337e-01 -9.39836383e-01 2.27045253e-01
6.16071522e-01 -5.58867097e-01 -7.11295307e-01 -3.43748748e-01
1.64006557e-02 5.67192078e-01 3.34595919e-01 -5.11237442e-01
3.70904684e-01 -1.87279209e-01 2.45440658e-02 -5.65318525e-01
-8.56650099e-02 -1.31828356e+00 6.55231237e-01 7.73845196e-01
-1.51181743e-01 2.47328892e-01 3.63330469e-02 1.11451364e+00
-3.46294314e-01 3.58731776e-01 5.95616758e-01 -3.47571492e-01
-2.86049068e-01 6.15879536e-01 1.50408074e-01 3.45517434e-02
1.35414839e+00 -5.26307151e-02 -2.81537205e-01 -1.07471131e-01
-6.40751600e-01 5.25000453e-01 6.45045996e-01 4.75493252e-01
5.46963155e-01 -1.32795489e+00 -6.34451628e-01 5.53501844e-01
3.93833905e-01 1.39908761e-01 7.05416620e-01 1.05807817e+00
3.93525921e-02 2.90279865e-01 -1.24797627e-01 -5.86870074e-01
-1.58749020e+00 1.15921235e+00 9.45706591e-02 -7.11359084e-01
-8.45567226e-01 3.93384904e-01 4.45180476e-01 -7.44675159e-01
3.25813085e-01 9.42194834e-02 2.30501369e-01 1.96488932e-01
6.22790396e-01 3.72249544e-01 3.07460636e-01 -5.87100327e-01
-5.23599207e-01 2.93034554e-01 -2.54904866e-01 1.91022128e-01
1.47771156e+00 -4.14922446e-01 -3.19752395e-01 2.62474447e-01
1.49627316e+00 1.82435542e-01 -9.71440077e-01 -4.83729601e-01
-1.64444253e-01 -7.06634462e-01 -2.20156118e-01 -8.20107818e-01
-1.36352766e+00 7.05504537e-01 7.48715460e-01 1.19809201e-02
1.45407712e+00 -2.93368816e-01 8.93955231e-01 3.39195393e-02
8.46899450e-01 -6.19985878e-01 -2.74715483e-01 -4.47984904e-01
6.13018394e-01 -1.16161478e+00 2.15683341e-01 -1.23841323e-01
-9.12879407e-01 7.14619040e-01 4.20430899e-01 5.88114202e-01
9.35682416e-01 2.20242903e-01 1.56200498e-01 -1.44653171e-01
-9.26432371e-01 5.03378034e-01 -5.95477447e-02 4.66343731e-01
7.63234198e-02 1.02355264e-01 -3.86206269e-01 8.41296017e-01
3.73046249e-01 2.04236180e-01 3.36410999e-01 1.34074998e+00
3.15234810e-01 -1.53155780e+00 -3.34448099e-01 5.02628922e-01
-7.64763653e-01 1.35027185e-01 -1.92613170e-01 5.76449573e-01
2.86037445e-01 7.86530316e-01 -3.79724324e-01 -3.24384630e-01
4.10621256e-01 9.96005237e-02 6.77102432e-02 -3.34568590e-01
-1.06567347e+00 -8.66100118e-02 -2.86003053e-01 -7.67169356e-01
-4.27917212e-01 -9.19794977e-01 -8.43759120e-01 -1.80919454e-01
-3.48735243e-01 2.37980127e-01 5.99469423e-01 5.06914675e-01
8.15535426e-01 5.45438118e-02 1.05356252e+00 2.10183449e-02
-1.05583525e+00 -2.17980027e-01 -6.38220370e-01 6.13160968e-01
6.31249309e-01 -2.23632287e-02 -2.43480280e-01 -2.58104831e-01] | [6.121460437774658, 6.472764015197754] |
48ddbbe8-1033-43aa-99a5-058da8a8bc75 | rangenet-fast-and-accurate-lidar-semantic | null | null | http://www.ipb.uni-bonn.de/wp-content/papercite-data/pdf/milioto2019iros.pdf | http://www.ipb.uni-bonn.de/wp-content/papercite-data/pdf/milioto2019iros.pdf | RangeNet++: Fast and Accurate LiDAR Semantic Segmentation | Perception in autonomous vehicles is often carried out through a suite of different sensing modalities. Given the massive amount of openly available labeled RGB data and the advent of high-quality deep learning algorithms for image-based recognition, high-level semantic perception tasks are pre-dominantly solved using high-resolution cameras. As a result of that, other sensor modalities potentially useful for this task are often ignored. In this paper, we push the state of the art in LiDAR-only semantic segmentation forward in order to provide another independent source of semantic information to the vehicle. Our approach can accurately perform full semantic segmentation of LiDAR point clouds at sensor frame rate. We exploit range images as an intermediate representation in combination with a Convolutional Neural Network (CNN) exploiting the rotating LiDAR sensor model. To obtain accurate results, we propose a novel post-processing algorithm that deals with problems arising from this intermediate representation such as discretization errors and blurry CNN outputs. We implemented and thoroughly evaluated our approach including several comparisons to the state of the art. Our experiments show that our approach outperforms state-of-the-art approaches, while still running online on a single embedded GPU. The code can be accessed at https://github.com/PRBonn/lidar-bonnetal | ['Jens Behley', 'Ignacio Vizzo', 'Cyrill Stachniss', 'Andres Milioto'] | 2019-11-04 | null | null | null | ieeersj-international-conference-on | ['robust-3d-semantic-segmentation', 'lidar-semantic-segmentation'] | ['computer-vision', 'computer-vision'] | [ 4.13148195e-01 -1.22951522e-01 4.39983271e-02 -7.17402458e-01
-9.01104689e-01 -5.81346810e-01 5.70951045e-01 5.87261766e-02
-8.00197601e-01 3.66030812e-01 -6.27556324e-01 -3.94250602e-01
2.45294288e-01 -1.05634499e+00 -1.10386884e+00 -5.74838221e-01
3.62205595e-01 6.15664542e-01 6.38838708e-01 -1.83142155e-01
2.70491898e-01 7.91172862e-01 -2.06523061e+00 1.97925344e-01
8.96810651e-01 1.51019764e+00 4.08691615e-01 4.92184967e-01
-3.91047329e-01 3.81348073e-01 -5.26171923e-02 -3.19147229e-01
5.23265183e-01 1.24701016e-01 -6.13047421e-01 1.21852569e-01
7.49051452e-01 -3.57589573e-01 -1.78210646e-01 1.29796505e+00
1.93502486e-01 -3.58195081e-02 2.01477125e-01 -1.27616000e+00
-5.43230213e-02 -4.02335823e-02 -4.03211296e-01 -2.13411629e-01
1.60767678e-02 1.85796887e-01 6.24498844e-01 -9.68694925e-01
5.02845526e-01 1.29565787e+00 7.24473417e-01 2.94925719e-01
-1.13378572e+00 -6.94611073e-01 -2.84538046e-02 4.45462584e-01
-1.42327189e+00 -4.19554055e-01 8.99650872e-01 -3.23708028e-01
7.34544933e-01 -1.75895821e-02 6.71293497e-01 8.96124840e-01
-3.52778137e-02 5.02762973e-01 1.20701540e+00 -2.32223254e-02
4.62884039e-01 9.21832472e-02 2.25701981e-04 7.06525862e-01
4.78077024e-01 2.92706251e-01 -4.14115697e-01 1.98129684e-01
6.04556680e-01 2.24520430e-01 2.86552683e-03 -6.35930657e-01
-9.53461289e-01 8.53173018e-01 7.79037356e-01 -1.21763758e-01
-2.81875551e-01 4.78686959e-01 2.15836808e-01 5.95579557e-02
4.10432070e-01 -1.87934652e-01 -5.88309944e-01 -9.20166299e-02
-9.73710179e-01 1.35269433e-01 5.88731825e-01 8.87575150e-01
1.31748676e+00 -1.24496944e-01 6.00610137e-01 5.78204572e-01
5.02960443e-01 9.05035675e-01 7.58160129e-02 -1.41078866e+00
4.22003001e-01 5.44283628e-01 8.96734670e-02 -7.05431283e-01
-3.33522081e-01 -1.63592070e-01 -7.13406324e-01 6.89135730e-01
4.95265633e-01 1.87896639e-01 -1.18117225e+00 1.09378552e+00
4.23306048e-01 2.02821046e-01 1.45446777e-01 1.03340161e+00
8.20152223e-01 5.19583046e-01 -9.14148241e-02 2.53895074e-01
1.33940268e+00 -7.67512202e-01 -2.12334082e-01 -5.00325859e-01
3.39996755e-01 -7.21140623e-01 8.17728341e-01 5.32264590e-01
-6.83784068e-01 -6.43400431e-01 -1.08128583e+00 -4.85913455e-01
-7.14772940e-01 3.03474162e-02 6.22706056e-01 5.17131090e-01
-1.09508204e+00 7.15137839e-01 -1.12664032e+00 -5.43007731e-01
7.71973968e-01 2.87527412e-01 -3.51695627e-01 -4.34834152e-01
-7.85693765e-01 9.61663902e-01 3.99944961e-01 2.85443544e-01
-6.45018578e-01 -6.50901735e-01 -1.00516474e+00 -3.10439706e-01
6.16701245e-01 -5.81230998e-01 1.32657528e+00 -8.74629676e-01
-1.44654202e+00 9.40947175e-01 -2.50796378e-01 -5.75044990e-01
7.51415908e-01 -2.77903378e-01 1.30282268e-01 3.72690260e-01
-2.31116060e-02 1.01424086e+00 7.62803435e-01 -1.53747737e+00
-7.68426657e-01 -6.80121303e-01 8.05645436e-03 5.55799492e-02
2.72054225e-01 -3.40661943e-01 -5.28485775e-01 8.40008333e-02
4.40103799e-01 -1.05245578e+00 -3.35145026e-01 4.27020431e-01
-9.23342481e-02 -6.60169125e-02 1.15681100e+00 -3.26719403e-01
1.10752985e-01 -2.02597666e+00 -2.23081112e-01 -7.85233155e-02
8.87601729e-03 4.57069278e-01 9.06004086e-02 1.36772916e-01
2.73088932e-01 -9.32632312e-02 -6.35278583e-01 -6.86271012e-01
-6.04292564e-02 6.27066433e-01 -3.48519146e-01 6.24193370e-01
3.12516212e-01 1.04364550e+00 -8.10810685e-01 -4.68253434e-01
7.90577292e-01 9.51463997e-01 -2.79097140e-01 1.77137870e-02
-5.07106066e-01 6.95020735e-01 -3.79867762e-01 8.28930080e-01
1.12510359e+00 3.77223827e-03 -1.95261627e-01 -2.68328398e-01
-3.76627058e-01 2.68203914e-01 -1.04680991e+00 2.05805612e+00
-4.75498229e-01 6.71119273e-01 3.29755545e-01 -1.22748899e+00
9.75789309e-01 -1.74870446e-01 4.76120144e-01 -8.20424259e-01
3.29720736e-01 4.57039088e-01 -3.83159876e-01 -3.72154385e-01
6.26869857e-01 5.28273247e-02 3.10600176e-02 -1.19398512e-01
-1.49914891e-01 -6.95179105e-01 -3.53762647e-03 -9.51725990e-02
7.73237288e-01 5.33397317e-01 -6.89584762e-02 -9.16088149e-02
4.42478865e-01 4.41372126e-01 5.26392579e-01 5.47708333e-01
-2.25797206e-01 7.09575951e-01 7.80474395e-02 -4.60547328e-01
-1.06061697e+00 -9.84231293e-01 -3.87061179e-01 6.74471319e-01
5.65462589e-01 -4.01075631e-02 -6.98816121e-01 -2.88710535e-01
2.14725837e-01 5.26708126e-01 -3.01632762e-01 2.22582936e-01
-5.17897069e-01 -4.68940407e-01 4.27421600e-01 6.14114583e-01
9.36555505e-01 -7.89357662e-01 -1.29449868e+00 9.21700969e-02
8.07712004e-02 -1.77783513e+00 4.08790231e-01 3.55822682e-01
-1.08448100e+00 -1.14463556e+00 -3.58979315e-01 -3.58304560e-01
4.23119217e-01 5.69242537e-01 1.02504265e+00 -1.00844242e-01
-3.31611246e-01 3.60608250e-01 -2.89864600e-01 -5.39071262e-01
-1.98140696e-01 -8.33996013e-02 -2.86543459e-01 -3.83391902e-02
5.71920455e-01 -5.62495589e-01 -6.81751490e-01 6.22479953e-02
-1.00601852e+00 2.18807429e-01 5.72962344e-01 3.79913598e-01
8.29765141e-01 -1.54268026e-01 -1.49676248e-01 -6.22048914e-01
-1.81111917e-01 -1.27661869e-01 -1.24517870e+00 -2.58183122e-01
-5.17983854e-01 3.09397858e-02 4.71910954e-01 4.73755179e-03
-8.58909070e-01 6.55465841e-01 -4.12821889e-01 -6.28502190e-01
-6.83704436e-01 2.19625458e-01 -1.16846435e-01 -3.47302854e-01
2.92922020e-01 1.59284711e-01 1.94612995e-01 -4.94356304e-01
5.58086216e-01 6.99916661e-01 6.91004813e-01 -3.27806920e-01
8.48795474e-01 1.10703671e+00 2.69092321e-01 -1.12443888e+00
-9.25571263e-01 -6.27141774e-01 -8.34479570e-01 -2.78778106e-01
1.00265467e+00 -1.00973165e+00 -8.17297697e-01 6.32320225e-01
-1.38292253e+00 -4.51491207e-01 -2.73925394e-01 3.29066068e-01
-6.46546423e-01 3.36993843e-01 -2.69141704e-01 -8.70851219e-01
-1.02315046e-01 -1.40064716e+00 1.48405123e+00 3.13800901e-01
4.39077169e-01 -5.08427978e-01 -2.47554466e-01 6.62513077e-01
4.22939777e-01 3.23165238e-01 2.71841794e-01 -2.49793246e-01
-1.23814261e+00 3.55701409e-02 -6.10687613e-01 4.31967288e-01
-4.22511965e-01 -3.09364758e-02 -1.29317689e+00 9.96349826e-02
4.11881916e-02 -3.07342470e-01 1.25830555e+00 2.35318527e-01
1.27958989e+00 2.41370246e-01 -2.37992316e-01 9.22510624e-01
1.74732685e+00 -6.04869910e-02 7.05380797e-01 2.50663579e-01
9.59981501e-01 5.82613230e-01 7.85237730e-01 1.46011636e-01
7.60288835e-01 7.53784835e-01 1.07433009e+00 -7.15185106e-02
-9.28975865e-02 -4.35621850e-02 5.47554307e-02 4.57408756e-01
-7.80574754e-02 5.25872186e-02 -1.20253718e+00 5.06289661e-01
-1.97206867e+00 -7.41493106e-01 -4.66713369e-01 2.28367567e+00
3.46067965e-01 1.54992923e-01 -2.25178108e-01 2.39900455e-01
3.76708567e-01 -1.35618541e-03 -7.02223718e-01 -3.75273079e-01
4.75052930e-02 2.94684231e-01 1.13264179e+00 5.77224910e-01
-1.22073293e+00 1.15275145e+00 4.83273411e+00 5.94383597e-01
-1.42654049e+00 2.44604930e-01 4.31332916e-01 1.54338211e-01
-1.10767573e-01 -2.24730256e-03 -8.21349978e-01 4.03034300e-01
8.93536091e-01 2.71045774e-01 4.19408411e-01 9.00726199e-01
1.64667755e-01 -6.45736039e-01 -9.40878928e-01 1.24854338e+00
-3.17472662e-03 -1.26628792e+00 -3.15602750e-01 2.28024080e-01
5.54312170e-01 8.22113276e-01 -5.80382869e-02 -3.42710279e-02
1.50209323e-01 -8.98415744e-01 9.07152176e-01 4.60716993e-01
7.11237311e-01 -7.19278872e-01 7.17327893e-01 4.54865396e-01
-1.31097579e+00 -3.82765308e-02 -5.79599202e-01 -1.57081813e-01
-6.62455661e-03 8.85209441e-01 -5.67947268e-01 6.48890853e-01
9.33483064e-01 8.19427729e-01 -5.77859938e-01 9.08569455e-01
-3.03258240e-01 3.17919403e-01 -7.96665311e-01 2.44897246e-01
2.93912560e-01 -3.58054876e-01 2.08553210e-01 1.00890291e+00
3.97936523e-01 2.01923102e-01 2.86824167e-01 9.88391161e-01
-4.91068102e-02 -3.20232868e-01 -7.22816885e-01 2.01789171e-01
3.39784920e-01 1.35293078e+00 -1.02126181e+00 -3.42351824e-01
-4.32219863e-01 1.00552571e+00 1.51151210e-01 2.80941427e-01
-8.56072068e-01 -1.04516529e-01 8.42541039e-01 9.06164125e-02
6.02072954e-01 -7.22075939e-01 -6.52587652e-01 -1.08108711e+00
1.46926671e-01 -2.71476597e-01 -1.64097786e-01 -9.29109693e-01
-9.06604946e-01 3.89347643e-01 -6.89831600e-02 -1.14347827e+00
-3.81718352e-02 -8.38120461e-01 -8.41376707e-02 7.13193417e-01
-2.17079687e+00 -1.23457849e+00 -7.66889513e-01 5.18862009e-01
6.24580801e-01 3.47633094e-01 6.39920592e-01 3.39494854e-01
-1.20506749e-01 -5.20088300e-02 9.05450583e-02 5.50121665e-02
3.32523286e-01 -1.02517474e+00 4.27552164e-01 9.56219971e-01
9.46919844e-02 4.81401831e-02 5.85988343e-01 -4.11602199e-01
-1.69788742e+00 -1.28050971e+00 6.32254660e-01 -3.46056402e-01
5.27740121e-01 -4.22562301e-01 -8.53941917e-01 4.83181119e-01
-9.69292223e-02 4.64838594e-01 6.43211529e-02 -4.58695889e-01
-3.68717581e-01 -3.75874966e-01 -1.15820098e+00 1.57113418e-01
1.02149439e+00 -5.57092249e-01 -3.87353033e-01 4.14451100e-02
6.69041514e-01 -6.35656834e-01 -4.31422830e-01 7.14960456e-01
3.51188213e-01 -1.27620316e+00 1.11551440e+00 2.55245298e-01
5.31121731e-01 -5.69907486e-01 -2.61879981e-01 -9.47202623e-01
4.47557449e-01 -1.55890167e-01 1.22633494e-01 7.30155051e-01
5.70447408e-02 -8.09317172e-01 8.09276402e-01 5.44961751e-01
-2.91871548e-01 -4.78489071e-01 -1.25520837e+00 -6.78335130e-01
-7.02299401e-02 -9.72192705e-01 3.77023786e-01 5.73917449e-01
-6.43125296e-01 1.22659318e-01 1.19777784e-01 3.72316003e-01
9.61325824e-01 4.84099805e-01 8.41632426e-01 -1.48650813e+00
1.67916507e-01 -3.14449787e-01 -8.17191780e-01 -1.16050422e+00
1.15704037e-01 -7.55207241e-01 3.47660869e-01 -1.70534098e+00
-2.36137092e-01 -5.13500750e-01 7.77434781e-02 4.31481719e-01
2.96270758e-01 8.04198503e-01 3.89970064e-01 5.65722324e-02
-5.86384475e-01 5.11309683e-01 9.36214626e-01 -1.21271193e-01
1.40299514e-01 -1.53957769e-01 -3.14985782e-01 1.02192342e+00
1.01801348e+00 -4.79738683e-01 -1.14285983e-01 -6.24945402e-01
1.94098130e-01 -2.26077333e-01 9.21191156e-01 -1.34861684e+00
4.05741960e-01 -1.21567426e-02 2.35240176e-01 -9.64274824e-01
7.56276548e-01 -1.16286933e+00 3.91617566e-02 4.85116988e-01
1.36326700e-01 -3.57573748e-01 2.75203347e-01 4.17461455e-01
-2.17906326e-01 -3.56080905e-02 8.98671627e-01 -1.45993501e-01
-1.09843540e+00 3.12785000e-01 -2.35439286e-01 -2.34643370e-01
9.10467923e-01 -4.32493001e-01 -2.19872743e-01 -1.04617357e-01
-3.76459986e-01 1.52942404e-01 8.04798782e-01 4.04373646e-01
6.50881708e-01 -9.23401713e-01 -4.43910956e-01 3.65383238e-01
1.71247244e-01 6.79087102e-01 5.17498329e-02 7.61680484e-01
-8.76395047e-01 4.61798608e-01 -1.85963437e-01 -1.13822544e+00
-1.12268877e+00 3.58169794e-01 3.47963274e-01 2.81529337e-01
-7.49879003e-01 6.59178317e-01 -1.49138153e-01 -5.65702558e-01
1.60145029e-01 -6.21163607e-01 2.19277740e-01 -9.33518186e-02
2.81698018e-01 2.98229814e-01 4.19466674e-01 -8.55644584e-01
-5.25912762e-01 8.91871274e-01 4.34194058e-01 -1.21420279e-01
1.27627420e+00 -2.47877747e-01 -1.02137782e-01 5.16709447e-01
1.27868152e+00 -3.98001969e-01 -1.58390582e+00 -1.51382029e-01
-8.78820568e-02 -5.50418973e-01 3.65533501e-01 -5.12016475e-01
-1.29604244e+00 1.30199230e+00 8.15671384e-01 6.90970421e-02
1.10825384e+00 7.56843090e-02 8.71749043e-01 4.75623757e-01
7.70324230e-01 -1.00704825e+00 -3.36150646e-01 6.73018217e-01
4.74587113e-01 -1.79078376e+00 5.22822738e-02 -7.30391800e-01
-3.16460341e-01 1.08209682e+00 3.27934712e-01 -2.72149116e-01
6.03268147e-01 4.07931834e-01 3.87492210e-01 -9.22910050e-02
-2.44381025e-01 -6.79159820e-01 -9.53025278e-03 5.92336714e-01
6.45704716e-02 3.01444493e-02 -3.41700539e-02 -4.65999842e-02
-1.73035145e-01 2.04531595e-01 3.74473959e-01 9.50226068e-01
-6.23864532e-01 -1.05371058e+00 -5.85828245e-01 1.02676041e-01
-1.24847636e-01 2.33714725e-03 -1.76907584e-01 5.86943388e-01
4.58922505e-01 9.89845395e-01 3.54243785e-01 -1.99288175e-01
2.42716610e-01 -2.76269242e-02 5.05696714e-01 -3.66016030e-01
-1.39228672e-01 -7.89677650e-02 -2.24406913e-01 -1.17157817e+00
-6.93103850e-01 -6.73181713e-01 -1.39609551e+00 -2.80575275e-01
6.40710518e-02 -3.42789441e-01 1.43254936e+00 9.92202997e-01
3.48841071e-01 3.53013575e-01 2.83853829e-01 -1.49394965e+00
-3.35039794e-01 -5.48639894e-01 -2.42719635e-01 2.44887531e-01
3.90528888e-01 -6.32980108e-01 -1.55910358e-01 1.18215837e-01] | [8.361729621887207, -2.453817129135132] |
12e4a94b-4348-44c3-8d47-e6694d6a37ac | recursive-deep-learning-framework-for | 2301.10874 | null | https://arxiv.org/abs/2301.10874v1 | https://arxiv.org/pdf/2301.10874v1.pdf | Recursive deep learning framework for forecasting the decadal world economic outlook | Gross domestic product (GDP) is the most widely used indicator in macroeconomics and the main tool for measuring a country's economic ouput. Due to the diversity and complexity of the world economy, a wide range of models have been used, but there are challenges in making decadal GDP forecasts given unexpected changes such as pandemics and wars. Deep learning models are well suited for modeling temporal sequences have been applied for time series forecasting. In this paper, we develop a deep learning framework to forecast the GDP growth rate of the world economy over a decade. We use Penn World Table as the source of our data, taking data from 1980 to 2019, across 13 countries, such as Australia, China, India, the United States and so on. We test multiple deep learning models, LSTM, BD-LSTM, ED-LSTM and CNN, and compared their results with the traditional time series model (ARIMA,VAR). Our results indicate that ED-LSTM is the best performing model. We present a recursive deep learning framework to predict the GDP growth rate in the next ten years. We predict that most countries will experience economic growth slowdown, stagnation or even recession within five years; only China, France and India are predicted to experience stable, or increasing, GDP growth. | ['Rohitash Chandra', 'John Hawkins', 'Rodney Beard', 'Tianyi Wang'] | 2023-01-25 | null | null | null | null | ['temporal-sequences'] | ['reasoning'] | [-9.26921189e-01 -3.11863065e-01 -1.11761011e-01 6.00526072e-02
-1.50042996e-01 -4.58888739e-01 1.08497488e+00 9.82188806e-03
-2.85968989e-01 1.12173343e+00 6.04826152e-01 -1.03790188e+00
2.22987011e-01 -1.09049988e+00 -2.15258896e-01 -6.65494502e-01
-5.05712211e-01 3.47650260e-01 -3.39023530e-01 -4.28054124e-01
1.54032364e-01 4.57047850e-01 -1.01757419e+00 -3.23351771e-01
8.72614861e-01 9.10040557e-01 1.72622323e-01 5.43482184e-01
-4.97983515e-01 1.00452781e+00 -8.54119360e-01 -1.25237986e-01
1.41818434e-01 -1.74209505e-01 -7.42267072e-01 -5.83664298e-01
-5.07796049e-01 -3.96008044e-01 -5.38656712e-01 8.24521124e-01
4.01942521e-01 1.35379523e-01 6.23558104e-01 -7.85025716e-01
-9.28209186e-01 6.37136757e-01 -7.40148842e-01 6.56741738e-01
1.70489877e-01 1.06058404e-01 4.01237905e-01 -6.93965316e-01
3.69143337e-01 1.30678308e+00 7.45947182e-01 7.64097273e-02
-9.68744338e-01 -7.86938310e-01 1.99203834e-01 7.91659951e-02
-9.81496572e-01 -1.52537197e-01 5.46415031e-01 -6.49742544e-01
1.56131458e+00 -1.22218028e-01 7.44542241e-01 1.12780225e+00
9.73105073e-01 3.35615754e-01 1.21763253e+00 -3.79865915e-01
9.70037803e-02 -4.91059154e-01 -8.82064402e-02 2.84124672e-01
2.28477810e-02 4.32870299e-01 2.01366648e-01 -6.88068867e-02
8.86934102e-01 3.88477296e-01 1.76208571e-01 7.17571557e-01
-1.17426515e+00 8.78191113e-01 4.93039310e-01 7.86786675e-01
-8.76100838e-01 3.57453339e-02 6.89478278e-01 7.05042958e-01
1.07724595e+00 4.28735605e-03 -9.28734243e-01 -5.90680182e-01
-8.51744890e-01 2.82425493e-01 8.16481352e-01 1.49001792e-01
3.24077815e-01 7.89597869e-01 2.51783848e-01 6.14503682e-01
-3.02863703e-03 8.38317692e-01 9.79960620e-01 -5.76549530e-01
4.71082896e-01 3.29496294e-01 3.03038448e-01 -1.35954785e+00
-8.20451021e-01 -5.18264472e-01 -1.47600758e+00 1.42730996e-01
2.94629812e-01 -6.96467698e-01 -1.19316471e+00 1.62340879e+00
-3.03712428e-01 2.69378334e-01 2.52097428e-01 3.76871198e-01
4.71088201e-01 1.32950187e+00 1.70536622e-01 -6.21151149e-01
8.86298716e-01 -5.98904490e-01 -8.33471954e-01 -1.00365639e-01
7.45789528e-01 -5.20153821e-01 4.29641485e-01 3.58314455e-01
-8.89824808e-01 -6.62245214e-01 -2.79530317e-01 2.21370429e-01
-8.28241169e-01 -1.98019758e-01 7.31708348e-01 3.62540781e-01
-1.14125848e+00 6.01849020e-01 -1.21996701e+00 -3.17767769e-01
-6.65433332e-02 1.01966932e-01 -1.94372237e-01 4.70189244e-01
-1.64971828e+00 1.12494719e+00 7.06878603e-01 2.67444868e-02
-6.22048378e-01 -3.86829108e-01 -7.80158818e-01 2.76713580e-01
-4.89317209e-01 -4.37100202e-01 1.09229302e+00 -8.71135771e-01
-1.38507009e+00 4.71983075e-01 -1.51815116e-01 -8.48900616e-01
1.63200125e-01 1.07319675e-01 -1.04400575e+00 -6.81659818e-01
-4.95262444e-02 6.79230466e-02 1.83390155e-01 -4.83246952e-01
-7.88489163e-01 -5.46584904e-01 -2.43419915e-01 -3.26258540e-01
-2.32477471e-01 3.49762946e-01 3.62771720e-01 -8.99593353e-01
-6.59329370e-02 -8.52260411e-01 -4.15217757e-01 -1.22578824e+00
-1.20819183e-02 -5.78442216e-01 8.19290340e-01 -1.37617528e+00
1.75919521e+00 -1.87827659e+00 -2.49098390e-01 1.30053982e-03
-1.02410465e-01 3.96879911e-01 9.39264297e-02 6.93692088e-01
-3.67404133e-01 3.42311770e-01 -1.65736780e-01 -2.23918498e-01
-1.34136513e-01 4.50879395e-01 -8.61770928e-01 3.50326598e-01
3.98766287e-02 8.97814631e-01 -8.93194020e-01 3.25220436e-01
3.51623178e-01 5.07888317e-01 5.38528487e-02 7.95111209e-02
-7.07228333e-02 6.86739624e-01 -1.99556708e-01 5.16840100e-01
6.85802996e-01 -2.08684772e-01 1.00130634e-02 4.89498138e-01
-7.55610526e-01 5.01363635e-01 -6.06511295e-01 1.15265453e+00
-7.39022493e-01 9.01250541e-01 -5.80542505e-01 -1.28807378e+00
1.32100844e+00 6.82630360e-01 3.21109653e-01 -1.35058224e+00
7.39246830e-02 6.13608837e-01 2.13081658e-01 -3.81785214e-01
4.37120646e-01 -2.46161804e-01 -1.78993106e-01 5.24479270e-01
-2.50283778e-01 2.63946176e-01 2.61579931e-01 -1.70822069e-01
8.68390799e-01 -2.03510597e-01 5.61375737e-01 -3.32503170e-01
3.74528497e-01 -1.64345518e-01 8.00156236e-01 1.96255848e-01
-2.41317660e-01 7.31283128e-02 7.79148817e-01 -1.47843623e+00
-1.23670089e+00 -8.48747253e-01 7.02654012e-03 9.04656708e-01
-8.50022376e-01 2.23270521e-01 -1.96576938e-01 -1.63867563e-01
-5.50008416e-02 8.83250415e-01 -6.96355343e-01 3.27710629e-01
-8.61599505e-01 -1.15264964e+00 3.67597640e-01 4.44080234e-01
7.77038097e-01 -1.42328870e+00 -4.86263722e-01 6.33230865e-01
-1.73086941e-01 -6.95766687e-01 1.55669093e-01 1.80013910e-01
-9.01242077e-01 -6.08800292e-01 -1.06886363e+00 -6.84399247e-01
1.53049916e-01 -2.18947902e-01 1.32041299e+00 -5.07279873e-01
3.37453991e-01 -1.85953572e-01 -8.07239190e-02 -7.54897296e-01
-4.66172040e-01 1.64956033e-01 3.22830588e-01 -4.61755484e-01
4.29454058e-01 -7.21041262e-01 -6.62932277e-01 -3.16629976e-01
-5.34748971e-01 -1.15298428e-01 9.04855058e-02 6.28139496e-01
1.21263877e-01 3.84855151e-01 9.91247356e-01 -4.70289171e-01
8.93828273e-01 -1.05765843e+00 -7.45510578e-01 -1.06258072e-01
-8.67916286e-01 -2.48106927e-01 9.14668024e-01 -2.15650931e-01
-1.14259732e+00 -3.83596957e-01 -3.02633256e-01 -9.79753956e-02
-3.99715379e-02 1.48514450e+00 6.83691323e-01 5.79481661e-01
2.59493619e-01 4.48030144e-01 -2.61291087e-01 -6.32972717e-01
-5.06948307e-02 6.81175053e-01 8.13352883e-01 -3.00270885e-01
4.23846006e-01 3.14865053e-01 -1.85639009e-01 -4.86269653e-01
-2.63332516e-01 -1.49280384e-01 -4.99202669e-01 -2.03108951e-01
6.44595325e-01 -1.19842446e+00 -6.12103820e-01 9.80705261e-01
-1.38976836e+00 -3.72173876e-01 2.31297314e-01 5.91171205e-01
-2.09202841e-01 -1.42512530e-01 -9.46487963e-01 -1.03012240e+00
-6.59521520e-01 -7.39896834e-01 3.65720630e-01 3.26513439e-01
-8.52434114e-02 -1.90943563e+00 6.88204229e-01 -4.70589906e-01
9.38614786e-01 1.01912880e+00 1.00485122e+00 -4.62065190e-01
2.34514281e-01 -1.02429956e-01 -2.73602366e-01 3.11038017e-01
5.18358707e-01 2.58963704e-01 -7.74989784e-01 -5.69385529e-01
5.85259404e-03 5.51849827e-02 9.71813798e-01 9.63423848e-01
9.65578735e-01 -5.77570677e-01 -1.38851583e-01 6.56857729e-01
1.46516335e+00 9.04543579e-01 7.52497613e-01 8.07562828e-01
4.31461394e-01 2.85573930e-01 2.56810099e-01 7.58709610e-01
6.64981604e-01 -1.21170580e-02 4.60371763e-01 -2.33227432e-01
4.50954646e-01 -1.25904307e-01 5.90029061e-01 1.38871801e+00
-6.20065689e-01 -3.20087343e-01 -1.68850160e+00 8.99782360e-01
-1.59216595e+00 -1.11223757e+00 -2.56004006e-01 2.09523392e+00
5.81446648e-01 2.36456871e-01 3.70021224e-01 1.40097708e-01
2.86759466e-01 5.13464332e-01 -3.66407245e-01 -1.04803050e+00
-2.59769112e-01 3.98458689e-01 6.72056794e-01 4.21424448e-01
-1.03305137e+00 9.66247916e-01 7.11108255e+00 1.95687398e-01
-1.78877664e+00 -7.61860535e-02 1.29445493e+00 3.40021133e-01
-2.06423104e-01 -1.93800762e-01 -4.01958138e-01 7.77585745e-01
1.81685448e+00 -8.46252024e-01 2.74740249e-01 5.74320853e-01
8.21162224e-01 1.62633076e-01 -2.74373382e-01 7.91145027e-01
-6.03878915e-01 -1.28802586e+00 -1.62386641e-01 3.61828119e-01
1.13413560e+00 6.50800705e-01 2.71637529e-01 7.04031289e-01
8.92183959e-01 -1.19799685e+00 3.31918627e-01 7.51159787e-01
7.93005764e-01 -1.11948037e+00 9.69788134e-01 4.41862106e-01
-1.21353698e+00 -5.16755104e-01 -4.33111876e-01 -8.94595504e-01
3.44195753e-01 9.53548551e-01 -5.16605496e-01 5.20682096e-01
7.72481382e-01 9.27773297e-01 8.57172254e-03 5.43717861e-01
1.42139018e-01 9.08107281e-01 -1.80770993e-01 4.14030403e-01
7.74289429e-01 -2.90719688e-01 1.18725911e-01 1.22724569e+00
1.02090657e+00 4.96769734e-02 1.71986222e-01 3.40021312e-01
6.58748671e-02 -9.51331705e-02 -7.87089646e-01 -1.33328304e-01
4.21640933e-01 5.19431353e-01 -4.54765856e-01 -4.79573667e-01
-6.82804286e-01 5.23908973e-01 7.31022432e-02 6.68177724e-01
-7.12852538e-01 -2.65851259e-01 7.04817116e-01 -1.54674485e-01
-1.54002637e-01 -6.48638606e-01 -2.61785507e-01 -1.04407990e+00
-3.64971429e-01 -7.10481703e-01 4.28142965e-01 -5.44572115e-01
-1.04304051e+00 5.91837764e-01 -3.40575606e-01 -9.64299560e-01
-9.14201140e-01 -6.17848575e-01 -1.08117485e+00 1.31720161e+00
-1.69510841e+00 -8.85302782e-01 2.18917772e-01 4.75063175e-01
6.09765708e-01 -5.41111588e-01 9.27951276e-01 3.01930934e-01
-7.22406626e-01 -1.70790210e-01 7.55695999e-01 3.59138459e-01
2.06872135e-01 -1.29779923e+00 1.29617572e+00 8.83276045e-01
-3.71734858e-01 5.24804592e-01 6.83367491e-01 -6.94094300e-01
-8.24703753e-01 -1.03611720e+00 1.46648061e+00 1.07730649e-01
1.08228350e+00 1.23462543e-01 -1.07187426e+00 1.07830095e+00
3.75943840e-01 -1.62484005e-01 2.79454261e-01 2.43905455e-01
-1.52947605e-01 -1.03705242e-01 -8.99198353e-01 1.75643653e-01
2.87957489e-01 -3.81207824e-01 -4.80059028e-01 1.37186959e-01
6.52345479e-01 -2.58349597e-01 -1.20654511e+00 3.24573129e-01
6.53381288e-01 -1.00300002e+00 6.61360025e-01 -6.28498137e-01
4.88610953e-01 3.50447655e-01 4.04140167e-03 -1.97475183e+00
-8.04108858e-01 -5.54914534e-01 -4.74411726e-01 1.14179337e+00
1.55949339e-01 -1.24220383e+00 3.69513273e-01 4.54168141e-01
-6.56483695e-02 -5.15596688e-01 -1.08320940e+00 -1.05837023e+00
8.76266420e-01 -2.12463170e-01 1.14737523e+00 1.41857803e+00
2.13272162e-02 -1.56589866e-01 -5.31484187e-01 -1.74206663e-02
9.10762250e-02 8.25523362e-02 6.16728246e-01 -1.33553767e+00
3.04474056e-01 -8.80586684e-01 -2.51872391e-01 -7.54790723e-01
1.66884467e-01 -3.58433485e-01 -7.02268064e-01 -1.63937664e+00
-2.99247671e-02 -2.74350971e-01 -6.92217529e-01 4.08186495e-01
1.21601522e-01 -2.89980054e-01 -2.96693370e-02 1.76390782e-01
3.25845361e-01 4.66060460e-01 9.95730817e-01 -1.12086147e-01
-3.18224639e-01 5.41704185e-02 -3.13294500e-01 7.27844357e-01
1.12356353e+00 -6.94579631e-02 -1.74106583e-02 -7.45324194e-01
5.54722309e-01 5.97164214e-01 -7.96609595e-02 -8.82192314e-01
2.19324678e-02 -5.24575353e-01 4.47935253e-01 -7.55662084e-01
-2.77334481e-01 -5.05749166e-01 4.39522982e-01 9.98507679e-01
1.29752502e-01 9.28690732e-01 5.14902890e-01 7.83955604e-02
-6.30868971e-01 5.09761572e-01 5.01751900e-01 -3.18861872e-01
-8.28317046e-01 5.76887786e-01 -7.11301446e-01 -2.75787234e-01
5.35955429e-01 5.03732748e-02 -3.78605515e-01 -4.89025265e-01
-8.31781149e-01 3.82731199e-01 1.23257540e-01 7.63984621e-01
2.58348495e-01 -1.31373847e+00 -1.10882175e+00 1.62800640e-01
-4.64595258e-01 -4.00652349e-01 1.18118696e-01 7.11674690e-01
-6.32828832e-01 1.05364466e+00 -4.28353488e-01 -3.34800519e-02
-6.01515055e-01 5.60946047e-01 4.78591144e-01 -9.18670893e-01
-5.94126284e-01 3.55550110e-01 1.42579228e-01 -3.27629626e-01
-7.94541314e-02 -6.39697731e-01 -4.53346163e-01 3.84667873e-01
7.12994814e-01 5.85019112e-01 -9.63976793e-03 -7.45819569e-01
-8.50549713e-02 2.34503344e-01 1.07292436e-01 8.52003619e-02
1.69598114e+00 -3.15277845e-01 -3.45663369e-01 9.87309933e-01
1.17453468e+00 -5.57041466e-01 -8.46759558e-01 -6.37446046e-02
3.38066190e-01 -1.27325371e-01 2.47523665e-01 -7.96444118e-01
-1.27960873e+00 9.43297267e-01 6.67213321e-01 8.83986890e-01
1.18603158e+00 -5.64585507e-01 1.27361143e+00 -1.08762451e-01
4.82227564e-01 -8.46647561e-01 -5.53515851e-01 1.22309208e+00
8.18548143e-01 -1.22128427e+00 -3.64482015e-01 8.00692737e-01
-1.34623006e-01 1.33746815e+00 3.32747847e-02 -1.60149053e-01
1.03547406e+00 1.94606349e-01 1.56171471e-01 3.35344300e-02
-1.15555608e+00 -6.66860715e-02 -2.97405630e-01 4.58781898e-01
9.25091028e-01 6.18501902e-01 -3.83618027e-01 2.31315032e-01
-4.90778357e-01 1.19728073e-01 4.83051300e-01 3.70956004e-01
-4.52770263e-01 -7.55313337e-01 -6.11519694e-01 5.74757576e-01
-1.01982987e+00 -1.44395545e-01 1.62554845e-01 7.52954066e-01
-1.30431905e-01 7.77426600e-01 6.34899735e-01 -1.28476977e-01
-7.10033402e-02 1.18562043e-01 -1.95636153e-01 -3.55264805e-02
-5.46391606e-01 -8.37353617e-02 -6.92436695e-02 -1.07436635e-01
-2.49843523e-01 -7.67764509e-01 -1.32335615e+00 -1.16631341e+00
2.25184768e-01 -6.89250603e-02 7.32983768e-01 1.06969607e+00
1.71293914e-01 3.93305272e-01 1.16489458e+00 -8.55581701e-01
-1.93547443e-01 -1.30145335e+00 -6.49068594e-01 1.46996155e-01
7.83916831e-01 -4.60665911e-01 -4.64854717e-01 -1.18101770e-02] | [6.32613468170166, 3.0599498748779297] |
9b5b517d-2426-46d5-8d64-6151ee7e0b7e | integrating-deep-features-for-material | 1511.06522 | null | http://arxiv.org/abs/1511.06522v6 | http://arxiv.org/pdf/1511.06522v6.pdf | Integrating Deep Features for Material Recognition | We propose a method for integration of features extracted using deep
representations of Convolutional Neural Networks (CNNs) each of which is
learned using a different image dataset of objects and materials for material
recognition. Given a set of representations of multiple pre-trained CNNs, we
first compute activations of features using the representations on the images
to select a set of samples which are best represented by the features. Then, we
measure the uncertainty of the features by computing the entropy of class
distributions for each sample set. Finally, we compute the contribution of each
feature to representation of classes for feature selection and integration. We
examine the proposed method on three benchmark datasets for material
recognition. Experimental results show that the proposed method achieves
state-of-the-art performance by integrating deep features. Additionally, we
introduce a new material dataset called EFMD by extending Flickr Material
Database (FMD). By the employment of the EFMD with transfer learning for
updating the learned CNN models, we achieve 84.0%+/-1.8% accuracy on the FMD
dataset which is close to human performance that is 84.9%. | ['Yan Zhang', 'Takayuki Okatani', 'Mete Ozay', 'Xing Liu'] | 2015-11-20 | null | null | null | null | ['material-recognition'] | ['computer-vision'] | [ 1.93164960e-01 -2.18167230e-01 -1.53346136e-01 -4.76651281e-01
-8.10792983e-01 -1.87309608e-01 6.55793488e-01 2.77497172e-02
-4.34275627e-01 5.90162456e-01 5.29490001e-02 4.70568627e-01
-3.92949164e-01 -1.11098468e+00 -1.14173567e+00 -8.00616741e-01
-1.78020403e-01 2.62499720e-01 6.11111410e-02 3.32599968e-01
3.04540128e-01 7.14189947e-01 -1.92966330e+00 7.43194699e-01
4.58498240e-01 2.07652497e+00 2.29071364e-01 4.65891391e-01
-2.06660796e-02 6.61831439e-01 -6.15123630e-01 -1.65782779e-01
3.82321388e-01 1.69322178e-01 -9.17996407e-01 2.00312048e-01
8.17763627e-01 -3.51274282e-01 -6.21523380e-01 8.77813697e-01
2.08854601e-01 5.62584698e-01 1.06474757e+00 -1.06664193e+00
-1.03352320e+00 5.32443821e-01 -1.92552254e-01 -4.38088598e-03
1.09758303e-01 -1.53415516e-01 9.80116844e-01 -1.16688371e+00
5.55396259e-01 1.07751429e+00 4.77320522e-01 4.74543810e-01
-9.80484664e-01 -4.23706800e-01 3.54918465e-02 1.87192798e-01
-1.66059709e+00 -3.43270361e-01 7.57596314e-01 -4.77035820e-01
1.19181001e+00 2.07038403e-01 7.08300292e-01 7.22833216e-01
3.70277166e-02 1.15296006e+00 5.53933799e-01 -3.97453517e-01
1.23422250e-01 3.13309520e-01 4.30600315e-01 8.15575480e-01
2.58900195e-01 -1.04856178e-01 -5.50634742e-01 -1.39154550e-02
5.87049603e-01 1.38787791e-01 1.41902072e-02 -1.80493712e-01
-1.06707907e+00 6.19809210e-01 9.02042389e-01 3.44233245e-01
-7.33023524e-01 3.73954207e-01 3.36831570e-01 1.60111263e-01
4.52272832e-01 1.86342925e-01 -3.98032725e-01 4.23103273e-01
-6.89942360e-01 2.25747749e-01 5.41227639e-01 8.72909129e-01
9.19572234e-01 -1.08742885e-01 -6.01543546e-01 1.00243056e+00
2.74335086e-01 4.90639746e-01 3.22771132e-01 -5.04442275e-01
3.73658121e-01 8.11023891e-01 9.23922732e-02 -1.32436073e+00
-1.60069838e-01 -4.87984985e-01 -1.01847649e+00 5.59671409e-02
1.83932856e-01 3.89705718e-01 -1.08100283e+00 1.49864197e+00
9.69554856e-02 1.08234659e-01 1.25011370e-01 6.23090625e-01
1.05925679e+00 7.19793200e-01 -1.82620242e-01 3.41235906e-01
1.01857996e+00 -1.14693666e+00 -2.45776653e-01 9.49307680e-02
2.82475173e-01 -5.41316032e-01 6.36907279e-01 5.07850587e-01
-8.83462250e-01 -9.92656291e-01 -1.12618577e+00 1.64196998e-01
-5.92018247e-01 6.89491570e-01 4.24423665e-01 2.02527151e-01
-7.73029923e-01 8.55825961e-01 -6.48097992e-01 -6.94666877e-02
9.64893460e-01 7.73863256e-01 -4.29834217e-01 -2.28675976e-01
-9.32909429e-01 5.87678313e-01 6.05564952e-01 1.29613459e-01
-1.17378712e+00 -5.57154238e-01 -7.13902891e-01 4.98335287e-02
-8.73655453e-02 -3.54441911e-01 7.67044902e-01 -1.31740952e+00
-1.13377535e+00 8.26622188e-01 2.11799547e-01 -4.30057257e-01
2.75284588e-01 -4.63809162e-01 -4.49180424e-01 1.10573158e-01
-1.15347981e-01 8.17012131e-01 8.68570030e-01 -1.25490701e+00
-6.86039448e-01 -1.09846368e-01 -9.75239426e-02 -1.49627319e-02
-5.96973300e-01 -6.67459294e-02 -5.19995391e-01 -3.44053537e-01
1.06850430e-01 -7.88185537e-01 8.47875625e-02 2.39646450e-01
-5.90807557e-01 -4.89550799e-01 7.42045760e-01 -5.77695727e-01
7.88381279e-01 -2.40245223e+00 1.33467957e-01 3.83114755e-01
2.01567840e-02 3.32502067e-01 -4.40209180e-01 -2.58109104e-02
2.05421001e-01 4.45022695e-02 1.08473472e-01 -4.04825687e-01
7.17278570e-02 -7.98952207e-02 1.42684923e-02 3.40029895e-01
5.78812480e-01 9.31423724e-01 -5.86394012e-01 -2.44000942e-01
4.30646151e-01 8.38161469e-01 -3.44917685e-01 3.60854954e-01
-2.24033669e-01 -3.63715217e-02 -5.05717158e-01 7.00546622e-01
8.60501826e-01 -3.09348762e-01 -1.70184359e-01 -6.38499200e-01
2.44788244e-01 1.97642297e-01 -1.31704295e+00 1.57016015e+00
-2.93795884e-01 5.60421288e-01 -4.90228027e-01 -8.86203229e-01
1.25726318e+00 -8.10659155e-02 2.81987041e-01 -4.17230546e-01
4.08431739e-01 2.29566425e-01 -1.64112061e-01 -5.58770478e-01
5.92313170e-01 3.39632839e-01 7.91980699e-02 3.18412155e-01
6.98456109e-01 1.16886497e-01 1.30630046e-01 -2.27386624e-01
9.51810956e-01 -2.22744849e-02 1.30342841e-01 -8.92952159e-02
4.58170682e-01 -4.82572049e-01 1.45398185e-01 8.42705905e-01
-2.74157263e-02 5.93800426e-01 9.64499486e-04 -9.26566184e-01
-8.34582686e-01 -1.03010130e+00 -1.69779256e-01 8.95297527e-01
-1.47523046e-01 1.59412192e-03 -4.42648947e-01 -8.53355110e-01
2.58940041e-01 3.85831207e-01 -1.14157295e+00 -4.46139276e-01
-1.48333177e-01 -5.78075707e-01 3.34719419e-01 7.37224579e-01
8.30086350e-01 -1.43467057e+00 -4.97807503e-01 6.35684431e-02
1.14955440e-01 -8.69519293e-01 -4.20405937e-04 3.70761126e-01
-5.70831716e-01 -1.08483040e+00 -4.44361687e-01 -1.09707701e+00
7.03658402e-01 2.24016428e-01 1.03065419e+00 1.03441298e-01
-3.05746824e-01 3.14498186e-01 -3.53558868e-01 -9.89280939e-02
-1.96412250e-01 2.23695055e-01 -8.04060772e-02 5.23789287e-01
4.67990607e-01 -1.82480589e-01 -5.36157310e-01 -2.37434935e-02
-9.46358979e-01 -1.07964791e-01 6.71136200e-01 8.12002838e-01
7.80655086e-01 1.65874273e-01 6.60822451e-01 -4.78493184e-01
4.52378660e-01 -5.53378224e-01 -2.75953203e-01 3.80580038e-01
-1.32228479e-01 9.94277839e-03 5.02968609e-01 -5.61096072e-01
-9.38520074e-01 2.43654057e-01 2.34304830e-01 -7.27180183e-01
-4.18202460e-01 5.36663830e-01 -2.13354826e-02 -1.59877673e-01
5.97236931e-01 2.14257404e-01 -1.98032781e-01 -4.06916618e-01
2.66588420e-01 8.38820040e-01 5.37891448e-01 -6.26881719e-01
4.63889509e-01 2.31925145e-01 -8.89631137e-02 -6.75843298e-01
-1.00387847e+00 -2.49792472e-01 -8.29961896e-01 -4.27465528e-01
8.24748874e-01 -8.55612874e-01 -5.07240474e-01 7.87452817e-01
-1.01333857e+00 -2.43419446e-02 -4.59064394e-01 4.88620907e-01
-5.41215777e-01 -1.81658819e-01 -4.93791759e-01 -8.66942823e-01
-4.97791141e-01 -1.14455307e+00 1.10908782e+00 3.24717671e-01
-1.83372840e-03 -8.11319053e-01 -2.73803532e-01 1.42963275e-01
4.67771202e-01 3.84052604e-01 8.92676353e-01 -9.19810593e-01
-7.83672929e-01 -7.26498783e-01 -3.41421276e-01 9.12334621e-01
4.67399776e-01 1.49724796e-01 -1.37728560e+00 -2.28190526e-01
-3.94885629e-01 -6.38279796e-01 1.43068957e+00 5.62300026e-01
1.55808043e+00 -1.33339286e-01 -3.30161989e-01 4.11105692e-01
1.57617247e+00 2.40819141e-01 7.88108408e-01 2.74163038e-01
5.47141075e-01 3.23949814e-01 5.09522855e-01 5.71874917e-01
3.70786758e-03 4.15669918e-01 5.88304818e-01 1.82318345e-01
-2.02048808e-01 -2.12085508e-02 2.34089226e-01 7.30853379e-01
-8.11065808e-02 -3.16373974e-01 -5.78455508e-01 7.58918226e-01
-1.66444433e+00 -7.70238459e-01 3.91882092e-01 2.19266891e+00
5.09216607e-01 2.13982299e-01 7.84411430e-02 2.54994005e-01
7.88849950e-01 6.71743825e-02 -6.21203244e-01 -9.16162133e-02
-1.03972860e-01 4.71276432e-01 1.63822621e-01 1.28175959e-01
-1.58394408e+00 7.22229660e-01 5.89534283e+00 8.71133745e-01
-9.96886551e-01 -2.76481599e-01 8.01761210e-01 -4.94044498e-02
4.07477431e-02 -6.30902290e-01 -8.20619702e-01 1.85046583e-01
6.91003144e-01 1.55229166e-01 5.66360652e-01 9.84797001e-01
-5.14412105e-01 -4.37859409e-02 -1.29412067e+00 9.27208781e-01
3.72547209e-01 -1.55739999e+00 2.91434914e-01 -1.42697126e-01
9.74935114e-01 1.48081705e-01 3.26285660e-01 2.38402054e-01
1.34685412e-01 -1.11022568e+00 8.28266680e-01 1.00012994e+00
7.14083254e-01 -9.20750260e-01 9.10614669e-01 1.00279171e-02
-1.31409812e+00 -2.23901376e-01 -7.68583655e-01 2.48788312e-01
-5.28931558e-01 6.75154626e-01 -8.87741745e-01 5.35822749e-01
8.27985287e-01 8.83739889e-01 -8.19412172e-01 1.19505310e+00
1.85713217e-01 3.84330899e-01 -3.29508364e-01 -3.31991732e-01
1.84848174e-01 1.67727545e-01 1.58671811e-01 1.17699885e+00
4.27554607e-01 -5.44354975e-01 1.77356943e-01 1.04869819e+00
-5.70310235e-01 3.27379815e-02 -7.15142906e-01 -2.07327440e-01
5.43887496e-01 1.38988388e+00 -6.46312296e-01 -5.26512682e-01
-2.91467518e-01 8.10282409e-01 5.85260689e-01 1.71276122e-01
-5.73854625e-01 -5.41683614e-01 4.99972761e-01 -3.49961221e-01
6.72956645e-01 1.54822797e-01 1.93982333e-01 -8.25740039e-01
2.41861299e-01 -5.43806732e-01 4.15237129e-01 -7.09648430e-01
-1.55995297e+00 8.84640694e-01 -1.67375281e-01 -1.37252033e+00
1.04802288e-01 -9.06519055e-01 -3.68551672e-01 5.91756225e-01
-1.22563636e+00 -1.20385253e+00 -6.10990644e-01 4.93273228e-01
4.03403074e-01 -4.59889352e-01 8.94150317e-01 2.84570515e-01
-2.54521072e-01 5.48208892e-01 2.38470823e-01 4.11696136e-01
3.91495407e-01 -8.89676094e-01 2.83716261e-01 3.23931903e-01
1.08256087e-01 2.24450365e-01 -4.83453572e-02 -4.44603026e-01
-1.32940733e+00 -1.55908716e+00 5.69349468e-01 -1.43131092e-01
1.49435490e-01 -1.57585219e-01 -7.41586566e-01 7.30478883e-01
1.25967845e-01 4.30211782e-01 7.65179157e-01 -1.02913737e-01
-5.88191330e-01 -3.23469639e-01 -1.45167089e+00 1.12659715e-01
9.25298154e-01 -5.34996212e-01 -3.77339900e-01 3.29898566e-01
8.84258151e-01 -1.03019647e-01 -9.32791233e-01 4.48667198e-01
6.98580861e-01 -6.94676995e-01 9.77851212e-01 -7.70199418e-01
4.94422108e-01 -1.93872362e-01 -1.07744181e+00 -1.28780019e+00
-7.06573606e-01 2.88752943e-01 -2.10439831e-01 1.34638345e+00
3.57071102e-01 -2.88368195e-01 5.51303864e-01 3.41039330e-01
-9.44033712e-02 -9.60524380e-01 -8.70846450e-01 -7.73629487e-01
-1.32488653e-01 -5.58132827e-01 8.35946918e-01 6.20503843e-01
-6.79436028e-01 9.77634918e-03 -3.40789109e-01 3.59575376e-02
6.25405312e-01 2.23361135e-01 5.17688632e-01 -1.23040199e+00
-1.08160511e-01 -2.81120658e-01 -8.56814921e-01 -5.94041228e-01
2.10767999e-01 -1.18912482e+00 1.59502849e-01 -1.60284257e+00
5.22529423e-01 -4.08270836e-01 -1.05366766e+00 5.26248157e-01
2.67994672e-01 3.32837105e-01 2.60007024e-01 -9.29885358e-02
-9.42146122e-01 6.76747024e-01 1.11557055e+00 -7.17939258e-01
-6.41031936e-02 -8.75013396e-02 -4.81045663e-01 4.59308773e-01
8.27076793e-01 -4.02788222e-01 1.99034717e-02 -4.68405873e-01
3.91710922e-02 -4.71679360e-01 4.81172502e-01 -1.57709372e+00
-8.89662728e-02 -7.22962543e-02 1.17813325e+00 -8.40212166e-01
4.58836794e-01 -8.59796703e-01 2.63954163e-01 3.89451474e-01
-6.97887003e-01 -5.29739678e-01 2.47210398e-01 4.49175060e-01
-2.01301828e-01 -4.84564632e-01 8.12052310e-01 -1.14560023e-01
-6.33446038e-01 4.98394907e-01 -2.44995579e-01 -3.12396526e-01
8.97072375e-01 5.02861617e-03 -2.16169804e-01 -9.17728618e-02
-7.45175242e-01 -2.02762604e-01 2.31161397e-02 5.69747806e-01
1.14360023e+00 -1.90980303e+00 -5.52348793e-01 5.06768763e-01
1.93476334e-01 1.18121855e-01 1.94016531e-01 1.65898457e-01
-1.78075314e-01 2.39878386e-01 -4.69777107e-01 -7.22831428e-01
-1.08043981e+00 5.21966457e-01 4.41160053e-01 -2.59347647e-01
-1.12189695e-01 1.04882073e+00 -9.11280513e-03 -4.70540345e-01
3.99180084e-01 -5.41145921e-01 -1.81428686e-01 2.40887538e-01
5.67020774e-01 2.83375442e-01 2.05126017e-01 -7.01339364e-01
-4.60915715e-01 5.71274698e-01 -2.80481130e-01 2.64468372e-01
1.83279037e+00 3.28252614e-01 -1.91297397e-01 4.42382127e-01
1.72744751e+00 -4.73021299e-01 -1.27735353e+00 -5.47066271e-01
-1.80176705e-01 -5.72232127e-01 -3.79673182e-03 -8.77793908e-01
-1.49563730e+00 6.42842889e-01 9.63634193e-01 2.91596446e-02
1.25726891e+00 5.73357344e-02 4.83647019e-01 8.07408273e-01
2.28603050e-01 -1.05092633e+00 5.07050812e-01 5.35254955e-01
1.25401580e+00 -1.14755332e+00 -7.50968978e-02 -2.73070298e-02
-2.98460633e-01 1.34556985e+00 7.62285888e-01 -6.28248036e-01
1.11229920e+00 -3.24376337e-02 -3.94277036e-01 -2.40124121e-01
-8.00768733e-01 -1.35220796e-01 8.12003732e-01 5.86946666e-01
4.94995043e-02 2.92609692e-01 4.93378490e-01 7.03951776e-01
1.35603040e-01 1.56655297e-01 -4.55368422e-02 8.43067646e-01
-5.84534943e-01 -5.13590157e-01 -2.29380161e-01 8.93595278e-01
-1.21541426e-01 1.09147616e-01 -4.14780080e-01 6.43306375e-01
3.14586878e-01 6.38628542e-01 4.16094482e-01 -1.02015376e+00
1.74283579e-01 2.22945660e-01 6.80053711e-01 -4.21137333e-01
-6.09708428e-01 -2.06766501e-01 1.10189386e-01 -2.79038966e-01
-5.78938365e-01 -4.83653575e-01 -1.04839528e+00 3.30280885e-02
-6.27739787e-01 -1.63803920e-01 6.18289828e-01 8.40665162e-01
4.19396609e-01 6.15963280e-01 8.12294424e-01 -1.09272456e+00
-4.29841727e-01 -1.13020194e+00 -6.84039652e-01 5.78293443e-01
2.73697793e-01 -9.14324462e-01 -2.72039682e-01 -7.52266496e-03] | [9.654777526855469, 1.8687775135040283] |
a2fa6cb0-d37c-44d8-9f48-dec3b65569d4 | coloristanet-for-photorealistic-video-style | 2212.09247 | null | https://arxiv.org/abs/2212.09247v2 | https://arxiv.org/pdf/2212.09247v2.pdf | ColoristaNet for Photorealistic Video Style Transfer | Photorealistic style transfer aims to transfer the artistic style of an image onto an input image or video while keeping photorealism. In this paper, we think it's the summary statistics matching scheme in existing algorithms that leads to unrealistic stylization. To avoid employing the popular Gram loss, we propose a self-supervised style transfer framework, which contains a style removal part and a style restoration part. The style removal network removes the original image styles, and the style restoration network recovers image styles in a supervised manner. Meanwhile, to address the problems in current feature transformation methods, we propose decoupled instance normalization to decompose feature transformation into style whitening and restylization. It works quite well in ColoristaNet and can transfer image styles efficiently while keeping photorealism. To ensure temporal coherency, we also incorporate optical flow methods and ConvLSTM to embed contextual information. Experiments demonstrates that ColoristaNet can achieve better stylization effects when compared with state-of-the-art algorithms. | ['Weifeng Ge', 'Wenqiang Zhang', 'Yingtao Zhang', 'Boan He', 'Ruize Xu', 'Xiaowen Qiu'] | 2022-12-19 | null | null | null | null | ['video-style-transfer'] | ['computer-vision'] | [ 4.72878784e-01 -3.86732548e-01 9.59167853e-02 -3.96522969e-01
-3.90033387e-02 -6.57805741e-01 5.47764778e-01 -7.48159707e-01
-2.85628825e-01 7.37071574e-01 2.19473526e-01 2.69055609e-02
3.79134238e-01 -7.77761757e-01 -7.43105650e-01 -7.24653125e-01
8.43595624e-01 -1.24241665e-01 9.54682752e-02 -3.19008261e-01
4.26464498e-01 6.43658936e-01 -1.30947328e+00 2.84852654e-01
9.91488039e-01 8.77270818e-01 1.25554278e-01 6.25015974e-01
-3.84993523e-01 9.91178811e-01 -5.80188870e-01 -4.99745220e-01
4.50516760e-01 -6.68781996e-01 -7.22213268e-01 4.89728957e-01
8.24166656e-01 -6.25549138e-01 -7.05200791e-01 1.26072335e+00
3.61870527e-01 1.17445275e-01 7.03721166e-01 -1.36142457e+00
-1.26860332e+00 1.01962537e-01 -9.92242515e-01 -1.03147820e-01
1.59949154e-01 2.68830389e-01 6.58336401e-01 -6.68887019e-01
8.52164865e-01 1.61256993e+00 3.54203403e-01 6.92138076e-01
-1.32525599e+00 -9.44926977e-01 1.63563833e-01 -1.51813865e-01
-9.92272913e-01 -3.90540928e-01 1.18660474e+00 -2.25783288e-01
2.36172855e-01 3.41616064e-01 8.42718422e-01 1.28338861e+00
3.69109273e-01 9.28098619e-01 1.39561880e+00 -3.68454963e-01
4.43796292e-02 6.22332208e-02 -3.53402644e-01 6.81956887e-01
6.23491481e-02 1.29307553e-01 -6.17665648e-01 2.59531230e-01
1.39788198e+00 2.03607142e-01 -2.26219818e-01 -4.60135400e-01
-1.33353174e+00 5.97159803e-01 4.02110249e-01 1.06685773e-01
6.24225251e-02 2.45353773e-01 4.18520033e-01 4.75526333e-01
6.72193348e-01 4.64965969e-01 -4.31223586e-02 9.30178072e-03
-1.14800310e+00 1.40710041e-01 4.15017009e-01 1.38023329e+00
9.05896842e-01 4.86396641e-01 -3.95387262e-01 8.21812451e-01
9.38228741e-02 7.26755738e-01 3.21757048e-01 -1.23313451e+00
2.40545213e-01 6.25679016e-01 -1.19204134e-01 -1.03453028e+00
2.14371979e-01 -1.65205300e-01 -1.09704399e+00 4.89216179e-01
2.81029314e-01 -1.53280869e-01 -9.50581372e-01 1.61252105e+00
3.48691903e-02 1.38621315e-01 -2.20447019e-01 8.01311076e-01
7.00214148e-01 6.54273391e-01 -5.08920476e-02 7.00575300e-03
1.24534667e+00 -1.09437501e+00 -1.02683306e+00 -1.88036934e-01
1.21382743e-01 -1.30164158e+00 1.36999011e+00 3.35684597e-01
-1.19385469e+00 -7.34411538e-01 -1.03890026e+00 -4.79472667e-01
-1.86801642e-01 2.08975449e-01 5.90544343e-01 5.61523438e-01
-1.04316151e+00 8.00851643e-01 -6.48679197e-01 -4.02233958e-01
5.54864466e-01 1.08631268e-01 -4.16395426e-01 1.18938863e-01
-9.22438979e-01 6.10731363e-01 1.78269535e-01 6.35326058e-02
-4.63315219e-01 -6.90585673e-01 -9.04918849e-01 -2.98654228e-01
1.46077350e-01 -1.07439148e+00 9.08223987e-01 -1.54748631e+00
-2.03742385e+00 9.72320557e-01 -3.92146349e-01 1.01686761e-01
8.08495700e-01 -3.37243587e-01 -3.96954596e-01 1.92423891e-02
1.14369087e-01 9.34035659e-01 1.22429669e+00 -1.40951419e+00
-8.09229553e-01 -8.24888796e-02 -7.97126559e-04 3.49619716e-01
-5.94904542e-01 6.39821738e-02 -8.24830770e-01 -1.26365352e+00
1.05089471e-01 -9.88373995e-01 -3.14377993e-02 3.54026467e-01
-4.71109599e-01 3.06582093e-01 1.08169663e+00 -6.78997993e-01
1.07051134e+00 -2.11142087e+00 2.93790370e-01 8.63930136e-02
2.43085936e-01 1.15747638e-01 -2.78099298e-01 1.74869344e-01
-1.81402788e-01 -5.36228344e-02 -1.56697556e-01 -5.19721985e-01
-1.43831477e-01 2.28801101e-01 -5.39454699e-01 3.45844150e-01
2.82639682e-01 1.05234039e+00 -9.05569494e-01 -5.48268139e-01
2.81004280e-01 4.12091315e-01 -7.70262182e-01 2.91915506e-01
-1.10951684e-01 7.13009357e-01 -3.72382879e-01 4.73981947e-01
9.41517889e-01 -1.62328854e-02 -1.01623654e-01 -6.07645631e-01
-9.57013518e-02 -1.94256961e-01 -1.09796131e+00 2.15014386e+00
-4.42839146e-01 7.13408947e-01 -1.67449266e-01 -4.68257517e-01
1.01739800e+00 -7.26032779e-02 3.92721534e-01 -5.38187563e-01
4.18717951e-01 1.01913586e-01 -3.24064940e-01 -2.60817409e-01
7.00902045e-01 -2.23415822e-01 1.35879755e-01 4.95853603e-01
-6.03827694e-03 -4.69304740e-01 1.17364280e-01 1.40180945e-01
6.98689222e-01 7.78275192e-01 -1.12374261e-01 -4.74081665e-01
6.26223564e-01 -2.03881323e-01 7.62587547e-01 3.06872308e-01
3.85531448e-02 9.98996079e-01 4.18642819e-01 -5.42284608e-01
-1.27963984e+00 -1.11509383e+00 3.06820214e-01 8.57324362e-01
5.10956585e-01 -2.05370277e-01 -9.64257121e-01 -7.00831592e-01
-1.84987545e-01 5.45589864e-01 -7.03075469e-01 -3.79429340e-01
-7.10039318e-01 -3.22357506e-01 4.95591432e-01 4.86090392e-01
1.10163569e+00 -1.23507190e+00 -1.55608624e-01 7.93025121e-02
-2.10184619e-01 -9.69660819e-01 -1.36347890e+00 -2.80753732e-01
-8.03829610e-01 -8.00998509e-01 -9.60155249e-01 -9.98253226e-01
8.88155103e-01 4.39872056e-01 9.75259542e-01 -1.12661585e-01
-1.66249931e-01 1.19336024e-01 -2.41840199e-01 -1.53563544e-01
-4.75527853e-01 -3.41968164e-02 -3.26604284e-02 4.16005135e-01
2.07956374e-01 -7.92030871e-01 -8.55030417e-01 2.51265198e-01
-1.31510961e+00 3.33382934e-01 4.76483405e-01 8.93488646e-01
4.10614908e-01 -4.60854322e-02 -3.67558585e-03 -1.07763529e+00
5.80934107e-01 3.34394693e-01 -3.59914809e-01 3.28525990e-01
-4.98984754e-01 1.78670898e-01 7.85481513e-01 -5.59471428e-01
-1.43012738e+00 6.40701354e-02 3.10390264e-01 -8.42334270e-01
-1.34461492e-01 -3.32777143e-01 -4.08945560e-01 -2.80030757e-01
4.14760709e-01 4.21706855e-01 2.36696303e-01 -4.67570543e-01
6.56894803e-01 5.46811938e-01 5.89547694e-01 -7.27603257e-01
1.34503341e+00 8.71852934e-01 -9.94855091e-02 -7.57293999e-01
-8.56204510e-01 -4.03306745e-02 -8.56769383e-01 -9.79940668e-02
9.10241127e-01 -9.14232016e-01 -6.18236423e-01 7.87645996e-01
-1.31186724e+00 -3.07511479e-01 -5.40158033e-01 1.67144328e-01
-6.98031306e-01 6.45246804e-01 -7.58922458e-01 -3.58434290e-01
-3.27092379e-01 -1.12273586e+00 1.10642231e+00 3.64574432e-01
-3.30340296e-01 -9.75346208e-01 -6.64219931e-02 1.96780235e-01
4.33120281e-01 2.22326785e-01 7.26381600e-01 3.83746982e-01
-5.60422361e-01 3.18832487e-01 -5.42893708e-01 6.25226974e-01
7.33798325e-01 2.17600912e-01 -1.00903857e+00 -3.97023499e-01
-5.36901169e-02 2.29606424e-02 9.16846395e-01 1.41509295e-01
1.20043564e+00 -2.58889914e-01 7.60304779e-02 1.13558745e+00
1.26572847e+00 4.12240356e-01 9.28552151e-01 4.51636821e-01
1.21505177e+00 6.64178550e-01 3.64713311e-01 1.72540456e-01
1.11401908e-01 4.15172756e-01 -8.39993060e-02 -7.08044052e-01
-6.18549407e-01 -5.76704204e-01 5.02201140e-01 8.35453510e-01
-4.76792309e-04 -1.38528600e-01 -1.42077297e-01 2.28490397e-01
-1.75786376e+00 -9.44071710e-01 1.76143482e-01 1.94425499e+00
9.83675063e-01 1.77990331e-03 -4.11707815e-03 -4.34981845e-02
8.19995224e-01 3.71649861e-01 -6.63757920e-01 -5.15293121e-01
-4.39683288e-01 1.19987480e-01 5.40610671e-01 3.05825830e-01
-9.99413073e-01 1.43094265e+00 5.82786655e+00 1.08042240e+00
-1.36327088e+00 -7.06151724e-02 8.01484346e-01 -4.63910848e-02
-6.11696839e-01 2.51756590e-02 -4.01352167e-01 5.83151042e-01
-3.24646607e-02 -4.68970016e-02 6.92139626e-01 4.78708029e-01
4.03655171e-01 1.80706888e-01 -1.06734264e+00 1.32914174e+00
3.48080277e-01 -1.18941224e+00 5.02240717e-01 -2.89681256e-02
9.76534486e-01 -6.84832036e-01 2.63880342e-01 -7.59650096e-02
3.79053175e-01 -8.23433876e-01 9.44598138e-01 6.80515110e-01
1.12435412e+00 -7.09769070e-01 2.06029549e-01 -2.69251168e-01
-1.17786109e+00 2.67343313e-01 -3.97072792e-01 9.11184847e-02
1.15002446e-01 4.70544726e-01 -7.20003694e-02 6.94162190e-01
6.77214205e-01 1.14027023e+00 -6.94132447e-01 7.69768298e-01
-3.37512732e-01 2.50810564e-01 -2.50876509e-02 2.81206548e-01
2.40205631e-01 -8.19341302e-01 4.30869907e-01 9.97236431e-01
2.97269583e-01 -1.75926149e-01 3.03689092e-02 1.10459876e+00
-1.89294204e-01 1.67266935e-01 -7.14645982e-01 1.33798912e-01
1.73673868e-01 1.25314510e+00 -8.55526924e-01 -3.49884421e-01
-2.79095948e-01 1.79731762e+00 -5.85668907e-03 6.20277524e-01
-6.95109665e-01 -4.81875360e-01 8.24377537e-01 7.98876863e-03
3.37463692e-02 -1.41251996e-01 -6.80187881e-01 -1.52348459e+00
9.41075608e-02 -9.09771025e-01 -6.87205791e-02 -1.06837535e+00
-1.43874264e+00 6.74070299e-01 -3.21224868e-01 -1.63112712e+00
2.26457641e-01 -5.99927723e-01 -9.89692330e-01 8.01295757e-01
-1.48286903e+00 -1.52359021e+00 -4.44432139e-01 6.59677505e-01
6.69177055e-01 -2.97002882e-01 3.81852895e-01 2.70286441e-01
-5.53042948e-01 7.20076323e-01 1.64767250e-01 1.98038310e-01
1.35627854e+00 -1.34946799e+00 5.49789488e-01 1.05839765e+00
-6.27765581e-02 7.69995511e-01 7.51043737e-01 -7.52356350e-01
-1.31428003e+00 -1.30433893e+00 4.44616228e-01 -2.55970180e-01
5.05581260e-01 -3.73597175e-01 -7.80724168e-01 5.18690407e-01
5.54975152e-01 -2.23449767e-01 3.40627193e-01 -4.02646124e-01
-5.27035952e-01 -3.61391664e-01 -9.61926401e-01 1.21499467e+00
1.44823861e+00 -5.23087859e-01 -6.02444768e-01 8.21458548e-02
6.20576143e-01 -2.69242167e-01 -6.44554079e-01 2.02503324e-01
7.23732471e-01 -9.30704832e-01 8.77604127e-01 -5.27698159e-01
7.59009659e-01 -5.46295404e-01 9.33287367e-02 -1.47581613e+00
-5.45827627e-01 -1.11406100e+00 2.96745420e-01 1.71915126e+00
-1.04859043e-02 -5.10366678e-01 8.43466282e-01 5.04425108e-01
2.07935527e-01 -2.36304343e-01 -3.83407980e-01 -8.36037576e-01
2.01225415e-01 9.77905989e-02 8.41132343e-01 1.02453899e+00
-4.85070556e-01 3.28162313e-01 -7.81674385e-01 -2.26366982e-01
8.43819022e-01 3.76538455e-01 9.37274516e-01 -9.17999804e-01
-1.14671290e-01 -6.14934742e-01 -1.78817049e-01 -1.01031339e+00
4.14236546e-01 -6.69596910e-01 -2.41484240e-01 -1.11961746e+00
1.96524397e-01 -2.57992238e-01 -2.45885611e-01 3.46051991e-01
-1.38515636e-01 5.79752922e-01 3.68771851e-01 2.21864328e-01
-2.41510943e-01 9.27617908e-01 2.12277699e+00 -3.15928549e-01
-3.04956943e-01 -3.18358988e-01 -9.44707215e-01 7.52001107e-01
7.57826149e-01 -1.75841302e-01 -5.89603066e-01 -6.69072926e-01
-1.93387896e-01 -2.19757810e-01 1.95906609e-01 -7.67244339e-01
5.65649346e-02 -3.61045450e-01 7.12905109e-01 -3.53046685e-01
2.22677812e-01 -7.15987265e-01 1.75868988e-01 2.77769268e-01
-3.55551928e-01 2.59991795e-01 8.14215317e-02 4.95913088e-01
-2.82492280e-01 9.18551832e-02 1.13200116e+00 -3.82790789e-02
-5.95996499e-01 5.68546951e-01 -1.17171511e-01 2.62420718e-03
8.04023087e-01 -5.13928711e-01 -1.70665950e-01 -4.38203722e-01
-3.39722246e-01 3.97495627e-02 1.01616395e+00 7.53809929e-01
6.33480191e-01 -1.72520101e+00 -5.44800758e-01 5.05561888e-01
3.56445685e-02 -1.92094147e-01 3.27034831e-01 4.02298212e-01
-9.23971474e-01 -1.62611324e-02 -6.79966390e-01 -3.32551897e-01
-1.18208683e+00 5.08095622e-01 1.65807098e-01 5.53768314e-02
-7.31016397e-01 7.01008737e-01 8.38685453e-01 -3.18966866e-01
1.15109526e-01 -2.25656807e-01 -1.35345906e-01 -5.28071374e-02
4.26782817e-01 3.21561575e-01 -3.13440442e-01 -5.70294023e-01
6.52979091e-02 1.02848256e+00 -2.06129551e-01 -2.35381708e-01
1.11611378e+00 -4.18622196e-01 -4.06416982e-01 3.58647108e-01
1.24137020e+00 1.32914394e-01 -1.66109121e+00 -2.85554349e-01
-3.61452609e-01 -7.48472154e-01 4.46696430e-02 -4.80709106e-01
-1.40765595e+00 9.41781819e-01 6.01425350e-01 -2.55843729e-01
1.32910252e+00 -6.46161497e-01 1.09099483e+00 6.61447644e-02
1.56574264e-01 -1.16026402e+00 2.03354001e-01 4.78273898e-01
1.03683829e+00 -1.05035758e+00 -7.43979868e-03 -5.74784160e-01
-7.83927381e-01 1.25356102e+00 8.09297740e-01 -4.43944037e-01
5.48034251e-01 4.90908511e-02 3.29675496e-01 9.14877653e-02
-2.49634936e-01 -1.89713277e-02 4.77855206e-01 4.15515453e-01
3.66760671e-01 -1.96852699e-01 -7.72946253e-02 1.36633158e-01
-3.61402988e-01 3.23686749e-02 5.70507526e-01 6.89086020e-01
-1.04286969e-01 -1.38481724e+00 -4.05059814e-01 2.60299921e-01
-2.57009447e-01 -2.42185250e-01 -4.58232224e-01 5.94222844e-01
8.69090930e-02 8.22405577e-01 1.52925670e-01 -5.24026036e-01
4.59352255e-01 -1.36400029e-01 6.18640542e-01 -3.53378415e-01
-4.86306697e-01 5.42613268e-01 -3.21148157e-01 -7.08541691e-01
-4.34612513e-01 -3.65214169e-01 -7.28275418e-01 -6.98486805e-01
9.25198942e-02 -1.73198089e-01 1.64664611e-01 7.62448847e-01
1.97188765e-01 7.12148368e-01 9.36105847e-01 -9.28806365e-01
-1.09532535e-01 -7.81750441e-01 -9.76319551e-01 8.45952451e-01
1.57873273e-01 -5.73298752e-01 -3.75073880e-01 5.87914526e-01] | [11.507486343383789, -0.6756829619407654] |
05d984f1-db5d-41fd-9672-529259154532 | universal-differentiable-renderer-for | 2003.09852 | null | https://arxiv.org/abs/2003.09852v3 | https://arxiv.org/pdf/2003.09852v3.pdf | Multiview Neural Surface Reconstruction by Disentangling Geometry and Appearance | In this work we address the challenging problem of multiview 3D surface reconstruction. We introduce a neural network architecture that simultaneously learns the unknown geometry, camera parameters, and a neural renderer that approximates the light reflected from the surface towards the camera. The geometry is represented as a zero level-set of a neural network, while the neural renderer, derived from the rendering equation, is capable of (implicitly) modeling a wide set of lighting conditions and materials. We trained our network on real world 2D images of objects with different material properties, lighting conditions, and noisy camera initializations from the DTU MVS dataset. We found our model to produce state of the art 3D surface reconstructions with high fidelity, resolution and detail. | ['Yoni Kasten', 'Dror Moran', 'Yaron Lipman', 'Meirav Galun', 'Matan Atzmon', 'Lior Yariv', 'Ronen Basri'] | 2020-03-22 | null | http://proceedings.neurips.cc/paper/2020/hash/1a77befc3b608d6ed363567685f70e1e-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/1a77befc3b608d6ed363567685f70e1e-Paper.pdf | neurips-2020-12 | ['3d-shape-representation'] | ['computer-vision'] | [ 4.99699652e-01 -1.76170617e-02 4.83731776e-01 -3.68004054e-01
-6.43811822e-01 -3.68939877e-01 5.45893312e-01 -4.59556073e-01
-4.62405607e-02 9.49754193e-02 -6.34798631e-02 -1.33624136e-01
4.11370933e-01 -8.77158344e-01 -1.20090389e+00 -4.37392056e-01
2.49903589e-01 6.92625463e-01 1.61731709e-02 1.86207108e-02
6.48920387e-02 7.30927765e-01 -1.69841850e+00 3.19170177e-01
2.47370958e-01 1.19871044e+00 3.04985136e-01 9.70471263e-01
4.24400643e-02 5.32762170e-01 -2.39654943e-01 -1.18882835e-01
5.73692620e-01 -6.89302385e-02 -4.84909415e-01 4.63072360e-01
1.13898599e+00 -9.34709609e-01 -5.74967444e-01 8.46369565e-01
3.26907158e-01 1.14542224e-01 7.50358522e-01 -6.96723819e-01
-8.77519667e-01 -3.04500043e-01 -5.16246855e-01 -1.91020638e-01
4.52129990e-01 1.16570316e-01 6.08556330e-01 -9.56495464e-01
1.05119216e+00 1.22595942e+00 6.15522683e-01 6.46371603e-01
-1.50910628e+00 -1.23617478e-01 2.49621749e-01 -3.26023698e-01
-1.26292717e+00 -5.05074024e-01 1.02904224e+00 -5.28483987e-01
1.03919744e+00 -9.04710405e-03 8.92481446e-01 1.23396492e+00
2.02892900e-01 3.51251036e-01 8.19753110e-01 -4.25926298e-01
3.61025929e-01 -1.12039387e-01 -3.22685540e-01 8.75546753e-01
-1.80467740e-01 1.60518244e-01 -4.63377565e-01 -3.25007498e-01
1.63024700e+00 9.91050750e-02 -5.65049350e-01 -6.74155354e-01
-1.04724109e+00 4.25545156e-01 2.96311378e-01 -2.92179316e-01
-5.47947288e-01 4.33737129e-01 -1.36829063e-01 1.55803621e-01
7.65669644e-01 1.30837068e-01 -4.22796249e-01 3.33586574e-01
-4.49066073e-01 2.10484505e-01 8.23065102e-01 8.83872211e-01
9.88105237e-01 4.23378676e-01 3.17927957e-01 7.59908617e-01
4.20293212e-01 7.65664458e-01 -1.91747963e-01 -1.80059469e+00
3.71744424e-01 3.63388360e-01 2.86201984e-01 -9.82775092e-01
-9.91180614e-02 -2.91977853e-01 -7.49441147e-01 6.70210004e-01
3.59142572e-02 3.83065790e-02 -1.00322461e+00 1.48227763e+00
4.56416398e-01 4.96751577e-01 -2.78723001e-01 1.00496399e+00
8.93233538e-01 7.66731560e-01 -5.05895019e-01 6.49485216e-02
7.99624324e-01 -5.13991356e-01 -4.61809903e-01 -3.11905891e-01
6.55198991e-02 -6.40246928e-01 8.90752912e-01 5.69294989e-01
-1.67395389e+00 -5.28961658e-01 -9.82005298e-01 -3.80385995e-01
1.37809858e-01 -2.25738753e-02 4.13194597e-01 5.94376996e-02
-1.45820212e+00 6.37769997e-01 -8.37541759e-01 3.80908251e-02
2.40523979e-01 1.85752228e-01 -4.05579180e-01 -2.68783361e-01
-6.41287625e-01 6.61458313e-01 -3.45220298e-01 2.10494488e-01
-1.34348810e+00 -7.42396593e-01 -8.66290629e-01 2.12721061e-02
1.34159833e-01 -1.34542048e+00 1.32381475e+00 -9.44820344e-01
-1.69010997e+00 1.18505383e+00 -1.40510634e-01 1.26262203e-01
3.68991435e-01 -8.89989734e-02 1.20187178e-02 2.82485396e-01
-2.59829283e-01 4.31354851e-01 9.48866248e-01 -2.08208132e+00
-2.81664766e-02 -5.27244508e-01 8.01026747e-02 4.67047781e-01
2.78938293e-01 -3.20188016e-01 -6.60050333e-01 -3.11514586e-01
5.61083019e-01 -6.71249509e-01 -2.59114176e-01 4.38167542e-01
-2.73965180e-01 3.88342083e-01 6.31345928e-01 -7.15089679e-01
2.65527397e-01 -2.00410104e+00 5.43130577e-01 1.91904441e-01
3.02003682e-01 -1.73775822e-01 -2.77708262e-01 6.97670728e-02
-6.98504448e-02 -5.09331562e-02 -1.81134060e-01 -8.41588080e-01
-2.87512541e-01 2.75666356e-01 -2.58649766e-01 6.86643779e-01
-9.30944681e-02 6.87426269e-01 -5.57962179e-01 -9.93490368e-02
5.89410067e-01 1.11994755e+00 -7.58376539e-01 6.82038188e-01
-6.28468096e-01 6.42279208e-01 -2.96250015e-01 5.62939167e-01
7.81210363e-01 -4.60743964e-01 -6.36850819e-02 -3.28724325e-01
4.97547798e-02 1.39101371e-01 -1.24360621e+00 2.17170048e+00
-6.73579097e-01 6.40432358e-01 6.95433021e-01 -4.55117643e-01
9.83024657e-01 3.52976561e-01 5.07860899e-01 -6.59269214e-01
4.11609203e-01 -1.80800363e-01 -7.42841482e-01 -5.66021621e-01
4.91462409e-01 -3.10420722e-01 5.27220488e-01 4.33980435e-01
1.06395856e-01 -7.18229055e-01 -6.99869990e-01 7.59211555e-02
8.71229768e-01 5.47460318e-01 -2.26037815e-01 1.21525079e-01
2.09703714e-01 -4.63313460e-01 2.74292767e-01 4.67647880e-01
3.13440919e-01 1.12673330e+00 1.29031420e-01 -8.81066322e-01
-1.44267046e+00 -1.43602026e+00 -7.65437558e-02 5.77786148e-01
2.12596938e-01 -1.44746721e-01 -5.44006646e-01 8.63587335e-02
-1.25937387e-01 5.24387598e-01 -5.65198064e-01 1.18812360e-01
-7.64860928e-01 -3.97412591e-02 -2.91337781e-02 2.49840498e-01
2.57247448e-01 -7.45628417e-01 -5.64875364e-01 -1.75092630e-02
-1.98571146e-01 -1.39801025e+00 -1.00806639e-01 -9.77026448e-02
-9.83340502e-01 -1.03184128e+00 -3.77293974e-01 -6.20863140e-01
8.50760877e-01 3.52660626e-01 1.64800060e+00 2.02096790e-01
-2.02019483e-01 7.34962642e-01 3.23412657e-01 -2.17941359e-01
-5.70836902e-01 -4.81688321e-01 -1.41793892e-01 1.23335667e-01
-4.50414509e-01 -1.01646698e+00 -6.40437484e-01 4.79379259e-02
-9.96251583e-01 5.11858046e-01 -1.50056958e-01 4.29531574e-01
1.00330281e+00 -3.92395735e-01 -3.29935104e-01 -7.24414647e-01
6.71442896e-02 -3.44601065e-01 -8.25652480e-01 4.71486332e-04
2.51062512e-02 -1.64177105e-01 6.06369913e-01 -3.77913624e-01
-1.32500041e+00 2.57643193e-01 -3.15128595e-01 -8.61055315e-01
-5.21661878e-01 -1.53657600e-01 -2.49147817e-01 -2.56791532e-01
7.25136280e-01 1.06506504e-01 -2.04853222e-01 -6.41887724e-01
3.40903819e-01 2.72638321e-01 7.43650317e-01 -8.05588186e-01
6.09850526e-01 1.03537953e+00 2.62648284e-01 -8.99336636e-01
-7.61613786e-01 5.66769019e-02 -5.84516585e-01 -4.28471237e-01
8.20445836e-01 -1.11333573e+00 -8.45621169e-01 6.08876705e-01
-1.48555005e+00 -6.87100887e-01 -2.90115565e-01 2.62206942e-01
-9.16564345e-01 2.09034920e-01 -7.66981840e-01 -7.13916063e-01
-2.21892059e-01 -1.25184548e+00 1.57441390e+00 -7.67232627e-02
8.36922973e-02 -1.01403916e+00 1.26657829e-01 3.43990713e-01
3.25589150e-01 6.90259457e-01 1.02465332e+00 4.50556248e-01
-1.23670268e+00 -5.64514063e-02 -1.62529245e-01 5.36134958e-01
8.49049613e-02 1.07378572e-01 -1.32666981e+00 -1.44805685e-01
4.51233625e-01 -3.08803260e-01 5.83248377e-01 7.04903245e-01
1.43740022e+00 3.22437449e-03 4.88430113e-02 1.53612041e+00
1.79889023e+00 -6.09725900e-02 4.82213527e-01 4.86737378e-02
1.13802087e+00 3.65129769e-01 -2.94429064e-01 4.14227724e-01
5.87369978e-01 5.43331563e-01 1.06224585e+00 -1.69344783e-01
-2.09259868e-01 -2.74556160e-01 2.51078662e-02 6.31773770e-01
-3.56405258e-01 -2.87999094e-01 -8.60118151e-01 1.67256251e-01
-1.37854540e+00 -6.02729201e-01 -1.87501252e-01 2.18286586e+00
4.26505893e-01 -1.09932022e-02 -4.85264719e-01 -2.98636228e-01
3.22746724e-01 3.99127841e-01 -6.64662480e-01 -4.64101702e-01
-2.23401263e-01 1.17195295e-02 3.33383620e-01 8.99572372e-01
-6.57653928e-01 6.63411736e-01 7.34565401e+00 -7.10292757e-02
-1.21050251e+00 -8.52402151e-02 6.01664722e-01 -2.85606235e-01
-7.82812774e-01 -2.45801657e-01 -4.08374518e-01 1.09491132e-01
6.80252910e-01 4.22090679e-01 1.05729127e+00 6.14132643e-01
-6.60435529e-03 9.95278955e-02 -1.26792002e+00 1.13422191e+00
4.34170514e-01 -1.57931924e+00 3.52690578e-01 6.74512461e-02
1.00028718e+00 5.29330134e-01 1.32397011e-01 -3.78769547e-01
6.11843288e-01 -9.52136159e-01 8.08592081e-01 8.48844469e-01
1.04105949e+00 -3.81003410e-01 9.12554637e-02 5.35072684e-01
-7.41545200e-01 3.06367606e-01 -2.60365427e-01 -1.35301813e-01
4.48211104e-01 3.77685159e-01 -4.19519365e-01 2.74256408e-01
7.79655278e-01 7.72127450e-01 2.07486860e-02 5.18213749e-01
-2.08956107e-01 2.23772094e-01 -4.22416627e-01 7.25659430e-01
-1.30277928e-02 -4.50168997e-01 6.02141917e-01 5.03464878e-01
2.70299017e-01 3.72416764e-01 8.21865425e-02 1.18114805e+00
-4.21425283e-01 -2.78160363e-01 -1.05147862e+00 6.49476469e-01
5.88838458e-02 9.71296251e-01 -4.68371451e-01 -2.79285818e-01
-3.76402259e-01 1.04667532e+00 6.17866099e-01 9.25825179e-01
-3.93973410e-01 5.04438162e-01 7.75723696e-01 3.52723002e-01
2.47671902e-01 -5.87661564e-01 -3.72364849e-01 -1.37035692e+00
7.64647126e-02 -4.86443907e-01 -2.40137011e-01 -1.44069755e+00
-1.13859832e+00 5.50630927e-01 -2.02400967e-01 -9.13737655e-01
-7.79005662e-02 -8.26763272e-01 -5.60517848e-01 1.22427642e+00
-1.47922087e+00 -1.10296524e+00 -4.85786498e-01 5.86653829e-01
4.43549484e-01 2.37972111e-01 8.79642963e-01 1.16219081e-01
-1.50881425e-01 -2.05247834e-01 -1.19103072e-02 -1.79875851e-01
2.61831164e-01 -1.03845179e+00 7.65785396e-01 4.76391166e-01
-5.95463105e-02 5.92790365e-01 6.19462192e-01 -5.31339169e-01
-1.90999568e+00 -7.93960392e-01 2.72011518e-01 -6.76879942e-01
1.65327385e-01 -5.58281124e-01 -9.54707742e-01 1.09468377e+00
1.44015595e-01 3.78313005e-01 3.59429240e-01 -1.60433337e-01
-4.65445697e-01 6.81852326e-02 -1.16847205e+00 4.10140634e-01
1.15789163e+00 -7.38942683e-01 -2.22239479e-01 2.32028931e-01
6.16504967e-01 -1.05773842e+00 -9.33916330e-01 1.77998602e-01
6.37592673e-01 -1.19888437e+00 1.47567189e+00 -4.02931154e-01
6.62895203e-01 -1.16044216e-01 -6.02366090e-01 -1.54051924e+00
-2.37056553e-01 -3.43694806e-01 -3.23652983e-01 5.93053818e-01
-2.13740077e-02 -3.98370415e-01 7.69995153e-01 8.72854233e-01
-2.77197570e-01 -8.83023024e-01 -8.43430579e-01 -1.44550160e-01
4.81581874e-02 -5.99972427e-01 5.95053613e-01 7.73981214e-01
-1.06838262e+00 2.97584981e-01 -4.49877530e-01 4.97112513e-01
9.80284810e-01 2.14978233e-01 7.96397984e-01 -1.30411220e+00
-5.11242926e-01 -1.12032451e-01 1.21835405e-02 -1.37693107e+00
2.64074802e-01 -7.55171895e-01 1.63909141e-02 -1.84547985e+00
3.84841524e-02 -2.97632575e-01 4.10346001e-01 -3.61916348e-02
1.93698987e-01 2.11877480e-01 1.56895816e-01 1.23821311e-01
-2.76673555e-01 5.39822459e-01 1.61463892e+00 3.53012867e-02
-9.49186534e-02 -1.14460789e-01 -3.06761175e-01 1.12801123e+00
4.59813356e-01 -1.90865770e-01 -3.42158526e-01 -1.40657914e+00
6.11810207e-01 4.20366198e-01 6.23988807e-01 -5.60071647e-01
5.17697781e-02 -1.83090031e-01 7.12149918e-01 -6.62823558e-01
1.01444232e+00 -1.14005482e+00 4.08718079e-01 -6.82570562e-02
-2.26133436e-01 9.95890349e-02 5.36022894e-02 4.33962613e-01
3.04924548e-01 1.50305912e-01 6.59356236e-01 -5.90140522e-01
-2.61374235e-01 6.42370522e-01 -3.80253623e-04 3.82032618e-02
3.22286457e-01 -3.07866842e-01 -1.65767178e-01 -4.54804897e-01
-6.42936170e-01 -3.03636454e-02 1.17887330e+00 1.69365972e-01
9.38545704e-01 -1.34394443e+00 -5.70870936e-01 7.44464636e-01
-1.97856516e-01 7.73071170e-01 3.51682276e-01 -1.00968316e-01
-8.05520177e-01 -3.17531109e-01 -7.57868290e-02 -9.64045763e-01
-8.37213933e-01 3.10639411e-01 8.49089026e-01 2.72848606e-01
-9.10750628e-01 7.80253708e-01 4.39625055e-01 -8.88242900e-01
2.52104074e-01 -5.18327057e-01 3.03757727e-01 -6.19958580e-01
3.83292556e-01 3.08545798e-01 5.28999045e-02 -7.66022444e-01
1.78106837e-02 9.78949964e-01 6.20036900e-01 -1.40144065e-01
1.66440821e+00 -1.69008791e-01 -6.47086427e-02 6.92101419e-01
1.48268509e+00 -3.70007098e-01 -1.80170560e+00 -2.84401745e-01
-8.93916190e-01 -5.47949076e-01 4.01929200e-01 -5.45539200e-01
-1.21636927e+00 9.78521168e-01 4.08590674e-01 -9.05444100e-02
9.21952426e-01 4.21574228e-02 7.96954036e-01 4.45426494e-01
4.21946585e-01 -7.56735384e-01 2.76717506e-02 5.94117224e-01
9.70169604e-01 -1.18923295e+00 -2.39015035e-02 -3.20977896e-01
-2.40320802e-01 9.94251013e-01 5.41560769e-01 -4.24956739e-01
7.71142304e-01 6.50218308e-01 1.93013668e-01 -5.14100730e-01
-7.27821827e-01 3.91541123e-01 1.74407318e-01 5.95801592e-01
1.49937212e-01 -4.64737862e-02 8.60963345e-01 -8.78380015e-02
-3.65619548e-02 -2.49034129e-02 5.10941803e-01 6.96160972e-01
-3.08996916e-01 -5.81595004e-01 -4.09819931e-01 2.23555207e-01
-2.29065806e-01 8.91419873e-02 -1.99246034e-01 4.16120589e-01
1.29852161e-01 3.80982101e-01 3.48055571e-01 -2.19645917e-01
5.92901528e-01 -1.42543703e-01 9.15056288e-01 -6.49627209e-01
-3.10901612e-01 1.40878811e-01 -1.31737977e-01 -8.38847220e-01
-3.34246665e-01 -5.92948258e-01 -1.22989237e+00 -1.93301484e-01
8.28925893e-02 -4.82700795e-01 1.11250257e+00 6.24464095e-01
3.37627798e-01 4.69987601e-01 8.61847758e-01 -1.68146300e+00
-2.20043465e-01 -5.23553848e-01 -5.55233061e-01 5.32397687e-01
8.31976116e-01 -4.89198625e-01 -5.15438676e-01 1.96345642e-01] | [9.224832534790039, -3.1154415607452393] |
6fe2893a-6cda-41b4-9548-ab1396476b1b | a-survey-on-biomedical-text-summarization | 2304.08763 | null | https://arxiv.org/abs/2304.08763v1 | https://arxiv.org/pdf/2304.08763v1.pdf | A Survey on Biomedical Text Summarization with Pre-trained Language Model | The exponential growth of biomedical texts such as biomedical literature and electronic health records (EHRs), provides a big challenge for clinicians and researchers to access clinical information efficiently. To address the problem, biomedical text summarization has been proposed to support clinical information retrieval and management, aiming at generating concise summaries that distill key information from single or multiple biomedical documents. In recent years, pre-trained language models (PLMs) have been the de facto standard of various natural language processing tasks in the general domain. Most recently, PLMs have been further investigated in the biomedical field and brought new insights into the biomedical text summarization task. In this paper, we systematically summarize recent advances that explore PLMs for biomedical text summarization, to help understand recent progress, challenges, and future directions. We categorize PLMs-based approaches according to how they utilize PLMs and what PLMs they use. We then review available datasets, recent approaches and evaluation metrics of the task. We finally discuss existing challenges and promising future directions. To facilitate the research community, we line up open resources including available datasets, recent approaches, codes, evaluation metrics, and the leaderboard in a public project: https://github.com/KenZLuo/Biomedical-Text-Summarization-Survey/tree/master. | ['Sophia Ananiadou', 'Benyou Wang', 'Zheheng Luo', 'Qianqian Xie'] | 2023-04-18 | null | null | null | null | ['text-summarization'] | ['natural-language-processing'] | [ 5.42449057e-01 2.21982747e-01 -5.03062725e-01 -2.56150782e-01
-1.20625985e+00 -2.31304526e-01 3.15797389e-01 1.16034079e+00
-3.67106378e-01 1.03402126e+00 9.48950291e-01 -4.13762443e-02
-1.19793281e-01 -3.18989933e-01 -3.60459313e-02 -6.77692652e-01
-1.55631034e-02 5.04792511e-01 -1.77885145e-01 5.88113442e-02
5.07408857e-01 2.65565246e-01 -8.84671509e-01 6.26313984e-01
1.28126419e+00 4.52841103e-01 2.69740611e-01 8.43487799e-01
-2.78718233e-01 6.01566970e-01 -8.34747791e-01 -2.76625156e-01
-6.10942900e-01 -6.96430087e-01 -1.06486332e+00 -3.38394463e-01
9.91600677e-02 9.72690657e-02 -4.86379325e-01 1.00916243e+00
1.17577732e+00 -1.69197544e-01 7.16732383e-01 -8.65361333e-01
-5.62081158e-01 7.41714656e-01 -4.61546451e-01 4.79286611e-01
5.68287790e-01 -6.14409670e-02 7.38999486e-01 -5.78479469e-01
9.29093719e-01 1.00429761e+00 5.16850889e-01 9.18832541e-01
-8.62609625e-01 -3.34956557e-01 1.46004949e-02 2.49141648e-01
-1.10383117e+00 -7.30617881e-01 4.94965166e-01 -2.74309099e-01
1.05063689e+00 7.66409159e-01 4.96540695e-01 1.08522367e+00
7.94628859e-01 1.19627810e+00 3.97621751e-01 -2.08765835e-01
1.29681423e-01 -2.12542892e-01 5.59334576e-01 3.95421654e-01
6.95551217e-01 -7.56937802e-01 -5.61769724e-01 -6.93909407e-01
7.58169545e-03 1.66487291e-01 -5.55258095e-01 4.91452038e-01
-1.57660580e+00 5.61754763e-01 -5.34045659e-02 6.10716403e-01
-5.96557796e-01 -2.82660484e-01 1.02705610e+00 1.62718259e-02
6.23579979e-01 6.66799724e-01 -4.75326717e-01 -8.21194798e-02
-1.17709410e+00 3.84550333e-01 8.77596796e-01 9.94156301e-01
-1.37072250e-01 -2.82428801e-01 -6.18878961e-01 1.00233436e+00
1.54165914e-02 3.40126663e-01 9.52232659e-01 -6.23065889e-01
5.40279150e-01 5.74281633e-01 -3.37100208e-01 -9.88944590e-01
-8.33610892e-01 -4.51490641e-01 -1.50799894e+00 -1.06051719e+00
-9.25565213e-02 -2.11287633e-01 -6.18824303e-01 1.28572810e+00
2.55079240e-01 -2.84918193e-02 3.30898196e-01 3.50974619e-01
1.60604620e+00 7.29883015e-01 2.81047404e-01 -8.11892807e-01
1.93167961e+00 -7.50254631e-01 -1.30214179e+00 -9.95389372e-02
9.15192962e-01 -9.17165101e-01 4.03757751e-01 4.28446412e-01
-1.25676084e+00 -2.44827531e-02 -7.11490393e-01 -4.81851965e-01
-2.71749407e-01 2.79316783e-01 3.17505389e-01 1.79652438e-01
-9.32660341e-01 6.84443533e-01 -1.16598606e+00 -7.88648248e-01
7.61632383e-01 7.15540349e-02 -2.02065349e-01 -1.30992472e-01
-1.20210397e+00 7.93860018e-01 5.03859341e-01 1.85181201e-02
-4.00522500e-01 -9.16707277e-01 -8.21220875e-01 1.23129785e-03
2.40301803e-01 -1.35133862e+00 1.29659033e+00 2.11489782e-01
-1.18233526e+00 1.04120636e+00 -5.14597535e-01 -4.65975314e-01
2.98960060e-01 -1.88785940e-01 -4.07615095e-01 4.38551396e-01
3.12445402e-01 4.38341498e-01 -9.55397338e-02 -6.30848527e-01
-5.50014377e-01 -4.51839298e-01 -6.99680686e-01 1.13032125e-01
-2.80298889e-01 2.42507622e-01 -4.05549377e-01 -9.47302699e-01
-1.65500551e-01 -5.35901010e-01 -6.72281742e-01 -3.49316776e-01
-1.06415188e+00 -5.16899228e-01 1.53007805e-01 -8.80259931e-01
1.92019188e+00 -1.74988186e+00 1.14450127e-01 -4.26886231e-01
5.46255767e-01 4.60955530e-01 -1.04997106e-01 1.00018275e+00
-1.32262424e-01 3.41648549e-01 -4.23745841e-01 -4.29706156e-01
-3.58278245e-01 -9.75230560e-02 -2.23030612e-01 4.66959566e-01
1.09589614e-01 1.21070886e+00 -1.15887177e+00 -9.03438210e-01
-7.83957019e-02 2.47812003e-01 -2.51100540e-01 1.08769372e-01
-6.80218041e-02 6.21752262e-01 -8.82352948e-01 7.56705344e-01
4.16781545e-01 -5.11495948e-01 1.74552158e-01 -2.12715968e-01
-3.00902426e-02 4.92368311e-01 -4.37801808e-01 1.83714843e+00
7.40480125e-02 2.99534768e-01 -8.74993280e-02 -1.21991956e+00
6.40148580e-01 7.11807370e-01 1.00747013e+00 -1.92179322e-01
1.56539857e-01 3.11450303e-01 -2.63184160e-01 -8.37178528e-01
4.95069742e-01 1.14835473e-02 -1.71209291e-01 2.83948869e-01
-2.00300142e-01 -5.78106642e-02 5.97407401e-01 5.00458121e-01
1.42669404e+00 -3.96319717e-01 1.22799838e+00 -2.61637688e-01
7.29154229e-01 3.25873613e-01 6.57225728e-01 7.63405800e-01
-2.75701165e-01 6.77844048e-01 6.01692379e-01 -1.84730217e-01
-5.86962581e-01 -5.65168500e-01 -3.42213780e-01 5.84409535e-01
-2.60754704e-01 -9.87247109e-01 -7.73600817e-01 -5.19540012e-01
-1.78147420e-01 6.16057634e-01 -4.30738598e-01 -3.18897814e-01
-6.44507647e-01 -1.26283848e+00 9.53262091e-01 1.66126668e-01
2.22883269e-01 -1.11357629e+00 -5.17703056e-01 4.02441651e-01
-7.60266304e-01 -1.13584900e+00 -7.54674494e-01 -1.66299745e-01
-1.29643023e+00 -1.15345836e+00 -1.14347506e+00 -8.47624600e-01
6.64080560e-01 1.17946178e-01 1.05796289e+00 8.12663212e-02
-7.43495464e-01 2.87562519e-01 -3.85013938e-01 -8.51855814e-01
-7.13783324e-01 4.44770575e-01 -7.57573098e-02 -5.63858569e-01
2.76670128e-01 -2.76719391e-01 -8.58188868e-01 -3.41936618e-01
-1.14405358e+00 3.03853601e-01 7.59221375e-01 8.05267453e-01
6.64051950e-01 -4.48038697e-01 1.07327008e+00 -1.12463498e+00
1.10888588e+00 -6.99846745e-01 8.05870146e-02 4.27879989e-01
-6.29896283e-01 1.46256685e-01 5.63784719e-01 4.10506465e-02
-7.86706209e-01 -2.58885711e-01 -5.70930898e-01 3.69289458e-01
-2.17054367e-01 1.09839690e+00 6.98483437e-02 5.93608379e-01
6.06431186e-01 5.05527437e-01 5.30597344e-02 -7.18657732e-01
1.50347337e-01 9.70354080e-01 3.90823007e-01 -2.62576342e-01
2.90994756e-02 3.92501980e-01 3.46898288e-02 -9.92844582e-01
-9.95895922e-01 -8.69129121e-01 -3.45551461e-01 2.00216055e-01
9.16762888e-01 -6.30309939e-01 -5.84269166e-01 3.05688947e-01
-1.47450268e+00 1.99640960e-01 -2.20358402e-01 4.09805208e-01
-3.62730473e-01 9.68022883e-01 -9.42049682e-01 -3.61442715e-01
-1.32355368e+00 -1.02824366e+00 1.19841003e+00 2.85447538e-01
-7.41170645e-01 -1.05391169e+00 3.26674193e-01 3.74733746e-01
2.64443040e-01 4.90841269e-01 1.05957484e+00 -1.10079181e+00
1.96792021e-01 -3.18179458e-01 9.96953845e-02 -6.13599364e-03
5.77151120e-01 -1.48264706e-01 -5.87288558e-01 -2.85822272e-01
7.92566314e-02 -4.63457778e-02 1.14553452e+00 7.80348957e-01
1.40361845e+00 -6.22984171e-01 -9.34275270e-01 3.72004837e-01
9.00199175e-01 1.48962021e-01 4.64024007e-01 4.36301380e-02
4.47571635e-01 5.47395289e-01 5.09591341e-01 5.01302242e-01
5.73460698e-01 2.29757726e-01 -2.46238559e-01 -9.71740857e-02
-9.34106037e-02 7.83966556e-02 9.46508199e-02 1.62000024e+00
1.02240123e-01 -4.71558779e-01 -9.44372714e-01 5.03416121e-01
-1.96106875e+00 -8.01777780e-01 -2.26853281e-01 1.83281314e+00
1.43851614e+00 -2.42747694e-01 -1.23001605e-01 -8.91411975e-02
6.46454692e-01 1.12755097e-01 -6.37422860e-01 -2.44190261e-01
-1.05764940e-01 2.74390099e-03 1.74336344e-01 2.28135303e-01
-9.86124635e-01 6.12742543e-01 6.22285414e+00 9.85925317e-01
-9.31326628e-01 -7.31837898e-02 8.01652908e-01 -1.15799010e-01
-1.02328517e-01 -4.56509173e-01 -8.71063769e-01 4.80043650e-01
1.03011334e+00 -9.03254867e-01 -3.44880998e-01 5.04842281e-01
7.16820896e-01 -2.09956348e-01 -1.12747383e+00 1.03487027e+00
2.43957981e-01 -1.65491867e+00 4.88410741e-01 -2.08030194e-02
6.17246151e-01 1.33073315e-01 -2.76289672e-01 6.21475652e-02
-1.00224480e-01 -7.89020896e-01 -1.27153739e-01 7.36193180e-01
8.59067857e-01 -2.36188680e-01 9.38706398e-01 4.70143259e-01
-9.89224672e-01 3.56803924e-01 -4.56830412e-01 3.99161756e-01
3.67545366e-01 9.80429709e-01 -8.90622437e-01 1.22659016e+00
3.37461174e-01 1.12348270e+00 -5.23573875e-01 1.29213727e+00
9.71066728e-02 6.06629610e-01 9.82820615e-02 -1.16832174e-01
-1.31878898e-01 7.13745691e-03 7.18468666e-01 1.76207876e+00
2.95493096e-01 4.23641801e-01 1.71108633e-01 3.66899580e-01
-1.85720399e-01 5.77645957e-01 -2.64454275e-01 -4.67319995e-01
3.98054153e-01 1.25179887e+00 -7.96956778e-01 -7.17728972e-01
-1.92750156e-01 7.23883152e-01 -6.91400468e-02 1.28464803e-01
-4.47978139e-01 -6.53258264e-01 4.44911957e-01 -1.62917510e-01
-4.85772461e-01 1.39684066e-01 -3.38236004e-01 -1.38573074e+00
-1.74546763e-01 -1.02696514e+00 9.18433368e-01 -4.21586126e-01
-1.36530972e+00 5.16294539e-01 1.55811906e-01 -1.18206513e+00
-2.91858017e-01 -2.60266393e-01 -3.50526482e-01 5.54577112e-01
-1.50049174e+00 -6.40780091e-01 -7.86386430e-02 3.24500352e-02
8.76492381e-01 -1.10000491e-01 1.03310454e+00 3.32063556e-01
-8.97145271e-01 4.32493687e-01 3.67807150e-01 1.42752051e-01
1.09325051e+00 -1.05488169e+00 4.95191365e-01 4.25567389e-01
-3.70416194e-01 1.02484334e+00 7.95917153e-01 -8.72358799e-01
-1.37310696e+00 -1.23171830e+00 1.46631873e+00 -2.94721663e-01
3.74236941e-01 4.14490625e-02 -1.09906268e+00 4.03738201e-01
4.14981306e-01 -4.79214191e-01 1.18228209e+00 -2.56242424e-01
3.45054716e-01 4.64958251e-02 -9.77926493e-01 7.34624267e-01
8.12371254e-01 -3.85247432e-02 -7.51473188e-01 6.89700544e-01
7.17502654e-01 -4.44208473e-01 -1.13615727e+00 4.66278136e-01
4.25671488e-01 -2.93921351e-01 7.33570516e-01 -8.22523892e-01
5.67458272e-01 -3.59502397e-02 5.00805914e-01 -1.33593595e+00
-1.83507994e-01 -8.10133636e-01 -1.39719382e-01 1.01078546e+00
2.93577254e-01 -6.77364588e-01 6.24218047e-01 3.00642461e-01
-5.08050680e-01 -1.17290020e+00 -8.48129570e-01 -2.56740361e-01
2.51768112e-01 -8.04208145e-02 2.29518786e-01 8.53464842e-01
7.40770340e-01 5.76354563e-01 -1.03832580e-01 -3.76159251e-01
4.30204749e-01 2.27226410e-03 4.45311278e-01 -1.21237051e+00
1.21874601e-01 -7.63562977e-01 -1.71498567e-01 -1.02315402e+00
-5.21760546e-02 -1.33114743e+00 -1.11113943e-01 -2.32054472e+00
8.10268760e-01 1.98171794e-01 -2.34717116e-01 5.66044092e-01
-6.84318006e-01 -1.21245615e-01 -1.33964092e-01 3.36329907e-01
-8.96404088e-01 4.11980987e-01 1.28030133e+00 -3.68520528e-01
-2.27431536e-01 6.37230799e-02 -1.24603963e+00 4.70575511e-01
1.02349353e+00 -7.17287302e-01 -6.02832660e-02 -1.73684284e-01
2.66817719e-01 3.27230364e-01 -2.17092991e-01 -6.63399041e-01
6.68556452e-01 -1.34637788e-01 2.13496283e-01 -9.59437191e-01
-2.08810106e-01 -1.38712034e-01 -9.47847031e-03 9.34752405e-01
-7.29580402e-01 1.13821246e-01 3.83588552e-01 4.36513990e-01
-3.73372704e-01 -2.64654666e-01 5.04651129e-01 -2.68962771e-01
3.93215269e-02 4.30360407e-01 -7.55414367e-01 3.05739999e-01
6.75946176e-01 2.78891116e-01 -4.46371526e-01 -2.25054726e-01
-6.80423141e-01 7.68366933e-01 -2.39740033e-02 2.57326156e-01
7.71954894e-01 -9.97563064e-01 -1.38613999e+00 -3.66828889e-01
3.26082140e-01 1.65730923e-01 5.29109001e-01 1.29940283e+00
-5.62177002e-01 1.02519715e+00 1.54482499e-01 -3.96398813e-01
-1.66404045e+00 5.33863366e-01 3.50501947e-02 -7.59404957e-01
-7.69934297e-01 3.79481316e-01 2.41584495e-01 -1.41282052e-01
1.63413748e-01 -6.93359375e-01 -5.29799759e-01 2.27393523e-01
9.74378347e-01 4.73917663e-01 4.25843090e-01 -2.41966099e-01
-5.11854112e-01 2.36696869e-01 -4.49210882e-01 3.31811339e-01
1.46169436e+00 -2.84671664e-01 -6.18440926e-01 5.11478841e-01
1.19646692e+00 1.30809080e-02 -6.79419339e-02 -3.67590845e-01
5.13809383e-01 3.30632925e-01 -2.77571410e-01 -8.23256433e-01
-6.96119666e-01 7.45438576e-01 -4.54602428e-02 1.85476691e-02
1.19163871e+00 1.44502193e-01 1.02572095e+00 4.90072221e-01
-1.25961944e-01 -9.37988162e-01 -1.19460829e-01 4.19111878e-01
1.02597904e+00 -9.81854379e-01 5.76975405e-01 -4.53605443e-01
-5.63060582e-01 1.22655320e+00 -8.69794413e-02 3.23052049e-01
5.80922365e-01 2.21969187e-01 4.33045663e-02 -3.21760654e-01
-1.00163686e+00 1.62494063e-01 4.78102475e-01 2.74241269e-01
1.00918925e+00 -6.25405908e-02 -1.01082349e+00 1.00462008e+00
-1.09195925e-01 2.35328972e-01 5.18750548e-01 1.01501298e+00
-3.57100189e-01 -1.23343182e+00 -2.84332961e-01 9.84468877e-01
-1.02387118e+00 -4.52531397e-01 -5.02328575e-01 2.49643147e-01
-4.50819492e-01 1.17410779e+00 -4.84017134e-01 5.97949103e-02
3.88862073e-01 3.39618884e-02 2.40051538e-01 -9.42750394e-01
-5.46032250e-01 2.96930879e-01 7.98994526e-02 -3.38612944e-01
-4.00246948e-01 -8.56003463e-01 -1.42844999e+00 -8.48751888e-02
-1.31545603e-01 5.74590445e-01 3.62070531e-01 8.11873615e-01
8.13535571e-01 8.44641209e-01 1.79465041e-01 -3.96352112e-01
-5.48494041e-01 -1.25499976e+00 -3.00763905e-01 -4.48244326e-02
4.20255184e-01 7.68820420e-02 -8.98844600e-02 3.50932419e-01] | [12.188308715820312, 9.500997543334961] |
45f384d4-d1af-4e81-93fb-02df5500b775 | compositional-diversity-in-visual-concept | 2305.19374 | null | https://arxiv.org/abs/2305.19374v1 | https://arxiv.org/pdf/2305.19374v1.pdf | Compositional diversity in visual concept learning | Humans leverage compositionality to efficiently learn new concepts, understanding how familiar parts can combine together to form novel objects. In contrast, popular computer vision models struggle to make the same types of inferences, requiring more data and generalizing less flexibly than people do. Here, we study these distinctively human abilities across a range of different types of visual composition, examining how people classify and generate ``alien figures'' with rich relational structure. We also develop a Bayesian program induction model which searches for the best programs for generating the candidate visual figures, utilizing a large program space containing different compositional mechanisms and abstractions. In few shot classification tasks, we find that people and the program induction model can make a range of meaningful compositional generalizations, with the model providing a strong account of the experimental data as well as interpretable parameters that reveal human assumptions about the factors invariant to category membership (here, to rotation and changing part attachment). In few shot generation tasks, both people and the models are able to construct compelling novel examples, with people behaving in additional structured ways beyond the model capabilities, e.g. making choices that complete a set or reconfiguring existing parts in highly novel ways. To capture these additional behavioral patterns, we develop an alternative model based on neuro-symbolic program induction: this model also composes new concepts from existing parts yet, distinctively, it utilizes neural network modules to successfully capture residual statistical structure. Together, our behavioral and computational findings show how people and models can produce a rich variety of compositional behavior when classifying and generating visual objects. | ['Brenden M. Lake', 'Reuben Feinman', 'Yanli Zhou'] | 2023-05-30 | null | null | null | null | ['program-induction'] | ['computer-code'] | [ 2.64365673e-01 2.50337064e-01 3.60811017e-02 -4.08389777e-01
1.01042204e-01 -8.53259504e-01 8.86948287e-01 2.57051408e-01
-7.78594464e-02 1.94488153e-01 2.82862633e-01 -2.01825961e-01
-1.57515138e-01 -8.42552662e-01 -8.29891086e-01 -4.98211920e-01
-8.60260352e-02 7.42296398e-01 3.06307584e-01 -1.74687862e-01
3.94559890e-01 2.61445403e-01 -2.07075644e+00 6.37545884e-01
8.57407451e-01 3.21626306e-01 4.10491884e-01 7.22445011e-01
-1.81923267e-02 8.55146587e-01 -4.31746811e-01 -5.65673590e-01
2.14951918e-01 -4.64242220e-01 -6.24508500e-01 2.87446827e-02
8.08523774e-01 -4.34277266e-01 -2.36138597e-01 9.16714668e-01
1.17566861e-01 3.83613110e-01 7.29534924e-01 -1.13870132e+00
-1.25059319e+00 1.16026866e+00 -2.37591267e-01 1.27795592e-01
7.81675279e-01 8.86884809e-01 1.15697873e+00 -8.20661485e-01
7.75444329e-01 1.64977717e+00 8.00165296e-01 7.91722655e-01
-2.11019492e+00 -6.06634378e-01 3.00012797e-01 1.16658591e-01
-1.09493828e+00 -3.20520371e-01 7.36290991e-01 -9.18230176e-01
9.90906179e-01 2.12172508e-01 1.27237570e+00 1.33528423e+00
-1.08567942e-02 8.11243176e-01 1.45267189e+00 -2.75982171e-01
4.13537621e-01 1.98369727e-01 6.21526003e-01 1.11838436e+00
5.53715944e-01 4.79358822e-01 -8.01537454e-01 -2.86016643e-01
6.39068067e-01 2.04344332e-01 1.26558021e-02 -5.97338140e-01
-1.39200819e+00 6.71729743e-01 4.93619472e-01 1.77927479e-01
-1.17029540e-01 3.90906245e-01 1.96379974e-01 1.65159419e-01
-3.81837159e-01 1.15668356e+00 -2.23510891e-01 1.72261655e-01
-6.80802107e-01 7.41719902e-01 8.12153101e-01 1.07373714e+00
9.37386572e-01 1.10567011e-01 -3.12077463e-01 5.39659202e-01
2.09975258e-01 4.89173472e-01 5.02144098e-01 -1.23700809e+00
3.62260230e-02 6.89076483e-01 -3.14780742e-01 -1.03191209e+00
-4.76096392e-01 -2.60121346e-01 -4.02308494e-01 4.72007930e-01
5.87000728e-01 7.96081591e-03 -1.02891541e+00 2.09615302e+00
-5.90133667e-02 -3.54852408e-01 -3.23423982e-01 5.15252531e-01
6.44949496e-01 5.66629589e-01 3.47069621e-01 1.73315018e-01
1.49520218e+00 -6.46384418e-01 -8.07524398e-02 -4.16655660e-01
3.86195123e-01 -6.79828078e-02 1.45395911e+00 4.11995828e-01
-1.39056480e+00 -1.02208376e+00 -9.94483173e-01 -1.31116569e-01
-3.94785494e-01 -3.90422136e-01 1.31458366e+00 7.47334659e-01
-1.17523956e+00 7.11282432e-01 -6.56137645e-01 -5.02499104e-01
7.70181715e-01 1.69601306e-01 -1.90137312e-01 -1.02093160e-01
-6.32312536e-01 8.55967343e-01 4.54172790e-01 -4.80575413e-01
-1.18407404e+00 -1.04034722e+00 -8.35603297e-01 2.50515908e-01
3.53932530e-01 -1.43384433e+00 1.30437803e+00 -1.22306466e+00
-1.19091213e+00 8.36389363e-01 -1.70589894e-01 -2.30453730e-01
1.33097008e-01 2.79584676e-01 1.06212460e-01 5.59050813e-02
4.07222584e-02 1.06083834e+00 9.72559512e-01 -1.45350254e+00
-2.45039865e-01 -5.40102065e-01 2.55860448e-01 2.45070774e-02
7.78645948e-02 -6.20941073e-02 2.29205579e-01 -9.06809926e-01
5.29944375e-02 -1.02612734e+00 -1.68301642e-01 3.54504734e-01
-2.24774227e-01 -4.14388865e-01 1.73163235e-01 -4.31079537e-01
8.79480958e-01 -1.94631302e+00 6.41249299e-01 3.35153967e-01
6.76096678e-01 -2.36370414e-01 -1.14884920e-01 2.90857881e-01
3.64678800e-02 4.15893316e-01 -2.20348909e-01 -1.95288330e-01
4.24811602e-01 1.77485630e-01 -4.53801483e-01 -5.73187098e-02
2.34894186e-01 1.14975691e+00 -9.54722881e-01 -1.69262037e-01
-5.57623357e-02 -1.83840264e-02 -1.05391335e+00 2.62287185e-02
-4.97156918e-01 1.26845285e-01 -8.02432373e-02 6.22093081e-01
1.67943999e-01 -2.87869394e-01 2.67103493e-01 3.93535271e-02
9.51609537e-02 6.67757094e-02 -9.99023616e-01 1.49011326e+00
-1.40715241e-01 7.49630868e-01 -4.38559592e-01 -9.75378096e-01
6.64881647e-01 -2.57693559e-01 -3.42708170e-01 -2.56968677e-01
1.03808090e-01 -2.38488153e-01 5.66120744e-01 -6.87761068e-01
3.22201490e-01 -4.09648567e-01 -2.03315422e-01 7.21055567e-01
3.42542052e-01 -5.64917564e-01 2.16117293e-01 3.15982789e-01
1.13699508e+00 1.31950557e-01 3.72451425e-01 -3.41846049e-01
-1.58678472e-01 1.30525663e-01 2.85561115e-01 1.32258093e+00
-1.41279832e-01 2.98647851e-01 5.07434070e-01 -6.00937903e-01
-1.20969713e+00 -1.63328040e+00 1.71027735e-01 1.65651083e+00
-6.05876707e-02 -1.85724884e-01 -5.87272644e-01 -9.10269171e-02
1.55436873e-01 1.10768783e+00 -8.90493929e-01 -5.40993273e-01
-3.25924993e-01 -4.39064652e-01 6.72856510e-01 6.38891697e-01
2.44013652e-01 -1.28387904e+00 -1.02282643e+00 -6.92621171e-02
1.36149228e-01 -6.30821526e-01 -2.32581869e-01 1.45797580e-01
-1.00276613e+00 -9.33408201e-01 -2.60789841e-01 -9.01216686e-01
8.66634250e-01 3.83261532e-01 1.24765098e+00 3.83293778e-01
-7.27055669e-01 6.86334550e-01 -9.19117555e-02 -4.53924537e-01
-5.78936994e-01 -3.25278580e-01 2.16721728e-01 -4.56989706e-01
2.24941462e-01 -9.33210492e-01 -1.19791135e-01 -8.16771612e-02
-8.12040269e-01 5.58460593e-01 6.91346645e-01 8.49020004e-01
-4.08595428e-02 -1.82894096e-02 3.17688614e-01 -9.14476812e-01
7.36233115e-01 -6.18129015e-01 -2.40672901e-01 2.55046248e-01
-5.04712760e-01 4.40373421e-01 6.32372320e-01 -1.03968453e+00
-1.32270706e+00 -6.27972260e-02 6.55497491e-01 -2.47566596e-01
-4.79747117e-01 3.86179209e-01 -5.71916662e-02 1.57854557e-01
1.28411186e+00 4.25836802e-01 9.60102584e-03 -2.46884361e-01
8.87796581e-01 -1.72267646e-01 5.35409510e-01 -1.24216890e+00
1.20901871e+00 5.27524590e-01 -2.22537830e-01 -8.12575638e-01
-4.73968118e-01 -7.74750067e-03 -8.00456941e-01 -7.43498728e-02
1.03203678e+00 -7.57467091e-01 -8.70261848e-01 4.38991100e-01
-1.16824937e+00 -5.92981577e-01 -4.71957684e-01 2.01965645e-01
-6.97089493e-01 1.47891521e-01 -6.61459684e-01 -5.48498809e-01
2.67955571e-01 -1.15582418e+00 6.65123761e-01 2.89257973e-01
-1.01291716e+00 -7.38162816e-01 1.07750356e-01 1.90528557e-01
3.09748977e-01 2.68388450e-01 1.60249650e+00 -6.57095492e-01
-9.15554106e-01 2.79842019e-01 -8.71322379e-02 -2.13636473e-01
-3.14516574e-01 2.36354649e-01 -7.97501564e-01 -6.64566411e-03
-1.88793153e-01 -5.61696112e-01 8.81267786e-01 8.20169449e-02
1.27331233e+00 -4.56867069e-01 -4.49837625e-01 7.29588449e-01
1.02314627e+00 3.06435436e-01 4.51770842e-01 -6.53584599e-02
7.70858526e-01 1.00163460e+00 -3.25753152e-01 1.94307461e-01
4.89515096e-01 4.11621571e-01 -2.35859916e-01 5.03336906e-01
-1.22942857e-01 -7.14278579e-01 5.11582255e-01 2.68727362e-01
-4.11846489e-01 3.70991051e-01 -9.87114727e-01 3.29987466e-01
-1.61970556e+00 -1.64338791e+00 2.77646989e-01 1.93025529e+00
1.17228472e+00 2.75954843e-01 2.23579645e-01 -2.85795659e-01
5.27451515e-01 7.02342913e-02 -7.77127028e-01 -4.33502674e-01
-3.54598574e-02 5.94261348e-01 -1.16808288e-01 3.03769410e-01
-4.80090827e-01 1.00434208e+00 7.32078981e+00 3.68824691e-01
-4.54193056e-01 -2.56668001e-01 3.87589425e-01 -2.28564769e-01
-7.38420606e-01 3.05810720e-01 -5.37648082e-01 1.65992573e-01
7.87400126e-01 -2.60128736e-01 1.09468496e+00 9.33744073e-01
-3.29409033e-01 -9.73420441e-02 -1.82458711e+00 9.17173147e-01
4.43223476e-01 -1.60389459e+00 5.69159925e-01 -1.44264981e-01
8.42220783e-01 -4.80226815e-01 2.21609876e-01 7.14441478e-01
1.01952279e+00 -1.24937081e+00 1.32258320e+00 8.93635571e-01
2.94013917e-01 -2.89695442e-01 -2.40056768e-01 3.97871882e-01
-8.90702426e-01 -6.11203015e-01 -2.57783800e-01 -6.58151984e-01
-3.48947763e-01 6.95353374e-03 -5.49139798e-01 -3.88488561e-01
8.52331817e-01 3.65567237e-01 -1.18970585e+00 7.02646613e-01
-3.78427893e-01 2.96651691e-01 -8.93634977e-04 -3.53646696e-01
-4.09390748e-01 9.50183123e-02 7.16290116e-01 1.02184939e+00
2.69185930e-01 3.12286913e-01 2.11754650e-01 1.94911432e+00
1.30967140e-01 -4.51696068e-01 -8.68269622e-01 -1.03339277e-01
6.50060833e-01 8.80930126e-01 -7.32007027e-01 -6.41240954e-01
-3.16890150e-01 6.68689370e-01 5.74883640e-01 6.05421901e-01
-6.18340373e-01 -1.27534240e-01 5.67456484e-01 5.54079153e-02
2.74314523e-01 -4.79554892e-01 -5.30071855e-01 -1.22444117e+00
-2.87420988e-01 -1.23281968e+00 3.90724801e-02 -1.21334779e+00
-1.50693166e+00 2.81694140e-02 5.26820123e-01 -4.56151545e-01
-2.19797775e-01 -7.98038483e-01 -9.11371052e-01 5.22654772e-01
-4.30924773e-01 -1.45531750e+00 -3.72025520e-01 5.86606681e-01
5.29150665e-01 -3.52619201e-01 5.67899883e-01 -5.77778578e-01
-3.18223029e-01 5.98999798e-01 -5.09902835e-01 6.48427382e-02
3.68068278e-01 -1.27749741e+00 4.95813191e-01 7.97061503e-01
3.30309302e-01 1.46576262e+00 8.03613722e-01 -6.60924852e-01
-1.49380040e+00 -6.03167057e-01 2.59425551e-01 -1.05647004e+00
7.25923955e-01 -7.02269554e-01 -9.42611814e-01 9.86630559e-01
1.00197392e-02 -4.52344298e-01 9.14549589e-01 4.18575972e-01
-1.18319011e+00 2.49732658e-01 -9.16098058e-01 1.10147679e+00
1.65130198e+00 -8.20967138e-01 -1.37573326e+00 -1.15487225e-01
1.05613077e+00 1.60146520e-01 -4.15409833e-01 -5.66136464e-02
7.65314698e-01 -1.20524096e+00 1.14326811e+00 -1.02781570e+00
7.72557616e-01 -3.79151523e-01 -1.64706156e-01 -1.41664970e+00
-8.59459758e-01 -4.90099192e-01 -2.35712126e-01 9.49738145e-01
3.63988250e-01 -5.43228805e-01 3.17183018e-01 1.01883221e+00
-2.27007508e-01 -3.43006790e-01 -2.90956289e-01 -6.76230431e-01
1.38978451e-01 -5.03921628e-01 6.09382272e-01 9.40305829e-01
3.39880794e-01 4.92951959e-01 3.06205928e-01 -1.17751315e-01
8.46907616e-01 2.85217404e-01 1.00245810e+00 -1.64613008e+00
-8.83062601e-01 -1.02947521e+00 -3.03386956e-01 -7.54656494e-01
2.61417449e-01 -1.23851323e+00 4.92300689e-02 -1.09734035e+00
8.07298303e-01 -3.28764200e-01 1.95848688e-01 5.06531239e-01
-2.88816839e-01 -3.30432206e-01 6.61168516e-01 3.97238791e-01
-3.71825606e-01 1.30116120e-01 9.85532880e-01 -3.53794247e-01
-3.43116581e-01 -2.70982563e-01 -1.41735148e+00 1.15505600e+00
4.96407866e-01 -1.75259501e-01 -5.61534584e-01 -4.35296267e-01
6.99710786e-01 -2.94124633e-01 1.02548814e+00 -1.14094794e+00
2.33347908e-01 -4.04774934e-01 8.98983061e-01 -5.12861572e-02
2.02419516e-02 -5.10548294e-01 2.57022470e-01 6.96650445e-01
-6.18170559e-01 5.71764484e-02 2.33173758e-01 8.41883421e-01
5.95808685e-01 -1.25163019e-01 7.11030006e-01 -6.17475748e-01
-9.91750419e-01 8.57280940e-02 -7.31850505e-01 6.45154342e-02
1.01661038e+00 -5.51133335e-01 -6.84563816e-01 -2.23626018e-01
-9.91939366e-01 -7.47661144e-02 8.16365242e-01 4.77634102e-01
5.89179695e-01 -1.20804417e+00 -3.70533079e-01 3.85179073e-01
3.75037283e-01 -1.60852134e-01 2.07095519e-01 3.59200805e-01
-3.08427662e-01 -1.87672451e-01 -5.24423182e-01 -6.13123238e-01
-8.57457519e-01 1.04585993e+00 1.89815193e-01 2.27618933e-01
-5.17556429e-01 9.94502842e-01 6.21915638e-01 -5.20937681e-01
-2.45672911e-02 -8.77795637e-01 5.63295484e-02 1.06763050e-01
6.06618106e-01 2.97066271e-01 -8.38456988e-01 -1.37042746e-01
1.25829145e-01 4.65596884e-01 -2.02057697e-02 -1.38745651e-01
1.21541560e+00 9.62325484e-02 -4.38985169e-01 7.11305261e-01
5.66221476e-01 -6.67957142e-02 -1.36687636e+00 -1.56465799e-01
-3.34829302e-03 -3.83575797e-01 -6.38367832e-01 -7.59253740e-01
-2.91407645e-01 9.26722288e-01 2.41645053e-01 -1.88339520e-02
5.70303082e-01 4.71751958e-01 2.09795445e-01 8.74775708e-01
6.06951773e-01 -9.35926020e-01 7.48463988e-01 4.49117690e-01
9.15456712e-01 -1.10271525e+00 4.19006012e-02 -1.48365255e-02
-5.88377118e-01 1.04540706e+00 8.67693424e-01 -2.88359076e-01
3.63568664e-01 8.99926107e-03 -6.92923367e-01 -3.43112707e-01
-9.86603737e-01 -1.92776427e-01 3.16611677e-01 1.17766786e+00
6.56124130e-02 3.27636957e-01 3.71507198e-01 8.45302641e-01
-6.91602945e-01 -3.50020260e-01 6.54457211e-01 6.24327064e-01
-7.34239340e-01 -4.78710055e-01 -4.03017968e-01 6.30667031e-01
2.56095260e-01 -2.22724304e-01 -3.98941457e-01 7.83798635e-01
4.22501355e-01 5.65841317e-01 2.84828395e-01 -4.38647062e-01
2.29322478e-01 4.12060469e-01 8.90777886e-01 -1.06801891e+00
-3.79166335e-01 -5.69623053e-01 -2.30919033e-01 -5.16165435e-01
-2.21117318e-01 -1.00907207e+00 -1.31694257e+00 -4.82591331e-01
2.29256198e-01 -6.07472777e-01 2.05025449e-02 8.34394634e-01
1.20142072e-01 6.01136804e-01 1.25947684e-01 -1.08747053e+00
-5.83551884e-01 -7.46474564e-01 -2.98074216e-01 9.14183021e-01
1.83346957e-01 -8.05829704e-01 -5.23574054e-01 4.98613566e-01] | [9.562214851379395, 6.845407485961914] |
ea79e23e-8636-4386-be31-86af3c19bcd2 | adversarial-transfer-learning-for-chinese | null | null | https://aclanthology.org/D18-1017 | https://aclanthology.org/D18-1017.pdf | Adversarial Transfer Learning for Chinese Named Entity Recognition with Self-Attention Mechanism | Named entity recognition (NER) is an important task in natural language processing area, which needs to determine entities boundaries and classify them into pre-defined categories. For Chinese NER task, there is only a very small amount of annotated data available. Chinese NER task and Chinese word segmentation (CWS) task have many similar word boundaries. There are also specificities in each task. However, existing methods for Chinese NER either do not exploit word boundary information from CWS or cannot filter the specific information of CWS. In this paper, we propose a novel adversarial transfer learning framework to make full use of task-shared boundaries information and prevent the task-specific features of CWS. Besides, since arbitrary character can provide important cues when predicting entity type, we exploit self-attention to explicitly capture long range dependencies between two tokens. Experimental results on two different widely used datasets show that our proposed model significantly and consistently outperforms other state-of-the-art methods. | ['Yubo Chen', 'Pengfei Cao', 'Jun Zhao', 'Shengping Liu', 'Kang Liu'] | 2018-10-01 | null | null | null | emnlp-2018-10 | ['chinese-named-entity-recognition'] | ['natural-language-processing'] | [ 9.78101045e-02 -1.96811512e-01 -9.13471058e-02 -5.32848001e-01
-7.48251796e-01 -8.40124488e-01 3.03791076e-01 9.32295248e-02
-1.00363350e+00 9.01553452e-01 3.25889140e-01 -4.39174265e-01
4.80984360e-01 -8.63647640e-01 -5.08860350e-01 -3.72712672e-01
1.56052411e-01 1.97966069e-01 5.31210005e-01 -7.72375315e-02
2.05572665e-01 3.67159635e-01 -5.91669679e-01 2.17806786e-01
1.36886418e+00 7.26022840e-01 3.03361416e-01 5.10160625e-01
-5.90509355e-01 5.11172056e-01 -7.06950188e-01 -3.72459143e-01
1.64995819e-01 -3.59269887e-01 -1.06011522e+00 -3.90757143e-01
5.75873218e-02 3.88982496e-03 -1.90380782e-01 1.22202420e+00
6.18344307e-01 2.48406708e-01 6.71562374e-01 -6.74134433e-01
-1.03644764e+00 7.68396854e-01 -4.84035224e-01 5.27979136e-01
-3.10993139e-02 -1.41474336e-01 9.94796216e-01 -7.70985663e-01
6.32739305e-01 7.76699781e-01 7.35529065e-01 9.56755102e-01
-6.18574917e-01 -9.15847123e-01 4.29463208e-01 7.10095391e-02
-1.40061259e+00 -1.34079456e-01 5.63606441e-01 -1.77390158e-01
9.12539065e-01 1.76935270e-01 -3.59931700e-02 9.64309990e-01
6.65612891e-02 9.33214247e-01 1.09908533e+00 -3.98154736e-01
1.79239899e-01 2.51689497e-02 5.42438090e-01 2.51182556e-01
2.50135422e-01 -3.34634840e-01 6.27986118e-02 1.00251205e-01
5.67341268e-01 5.04656695e-02 -3.37343961e-01 2.87727982e-01
-9.88531947e-01 6.94390118e-01 3.06298882e-01 8.30682456e-01
-2.35616252e-01 -1.12445720e-01 4.71098989e-01 1.37232214e-01
5.11838078e-01 3.98578554e-01 -9.24590886e-01 -1.53416514e-01
-7.45819271e-01 -3.55205268e-01 9.16198730e-01 1.22600222e+00
7.83639252e-01 1.35398924e-01 -4.46891636e-01 6.81423366e-01
5.46759851e-02 3.80687654e-01 7.50440061e-01 -1.63068473e-01
7.02991605e-01 4.83270079e-01 2.10943267e-01 -6.41482651e-01
-2.98798680e-01 -2.28571132e-01 -8.54234755e-01 -2.94579446e-01
4.34845477e-01 -7.81621456e-01 -1.21995413e+00 1.64177358e+00
2.22751185e-01 5.17480731e-01 2.23080695e-01 8.03080022e-01
8.08711708e-01 7.07465708e-01 6.57632768e-01 2.25229576e-01
1.40615070e+00 -1.00213480e+00 -7.79178441e-01 -5.54677486e-01
6.81917965e-01 -8.21675122e-01 9.64124322e-01 -1.72505319e-01
-5.90924025e-01 -5.14952421e-01 -7.99819887e-01 -1.66209668e-01
-7.87115097e-01 2.77417094e-01 7.40502894e-01 8.75578880e-01
-6.53653622e-01 5.42277694e-01 -9.47134793e-01 -2.05759794e-01
4.06377792e-01 4.07655835e-01 -3.59768480e-01 -1.32062316e-01
-1.70247233e+00 7.34964609e-01 7.23660946e-01 4.70828265e-01
-3.57275695e-01 -6.19953990e-01 -1.01625752e+00 2.15844736e-01
4.28819001e-01 -1.88239694e-01 1.11379731e+00 -9.30858791e-01
-1.30571568e+00 6.38732553e-01 -1.84855491e-01 -3.38813394e-01
4.06310946e-01 -5.70804000e-01 -8.06306362e-01 -1.20628893e-01
1.57130942e-01 5.20819485e-01 4.71797317e-01 -9.60707605e-01
-6.86710000e-01 -1.53262675e-01 -1.10401087e-01 8.75435993e-02
-6.23530626e-01 4.13003474e-01 -4.93197650e-01 -7.94248819e-01
-3.18425506e-01 -5.89018047e-01 -5.27304173e-01 -5.69623947e-01
-7.59203553e-01 -5.20310342e-01 7.06787288e-01 -9.21496272e-01
1.47624528e+00 -2.03159618e+00 -3.16843420e-01 -2.17843801e-02
-7.77347311e-02 8.61984313e-01 -2.95199811e-01 3.76481682e-01
6.46311715e-02 7.88599372e-01 -4.33005542e-01 -1.33051693e-01
1.17424224e-02 1.99062407e-01 -1.83954433e-01 2.34923914e-01
6.82240844e-01 8.86205196e-01 -8.96821082e-01 -6.05829477e-01
-1.45373628e-01 3.90997857e-01 -2.28727415e-01 3.36518466e-01
4.49116109e-03 5.06582618e-01 -9.26087141e-01 4.01744038e-01
1.01217282e+00 1.43329784e-01 5.92872128e-02 1.19701356e-01
-2.36891389e-01 4.32958364e-01 -1.25250244e+00 1.45167220e+00
-5.29451847e-01 3.69046003e-01 -1.05273917e-01 -9.17674482e-01
8.44924510e-01 3.50480139e-01 -7.75489211e-02 -4.64106441e-01
8.54937211e-02 2.09359959e-01 1.29188403e-01 -3.73085052e-01
5.91769516e-01 -1.13252789e-01 -4.25354928e-01 1.71176374e-01
6.58592209e-02 3.93510014e-01 2.40342960e-01 -1.13219284e-01
1.25341237e+00 1.14044391e-01 3.45903933e-01 -1.79649889e-01
6.95501626e-01 -3.04211341e-02 1.37468135e+00 8.08217824e-01
-4.93762016e-01 7.56818295e-01 2.21965790e-01 -2.00232074e-01
-8.26702297e-01 -8.43293011e-01 -1.13999717e-01 1.02965999e+00
2.16174915e-01 -8.60249326e-02 -1.08513474e+00 -1.25585508e+00
-3.79594475e-01 7.23679006e-01 -6.57911181e-01 1.28857672e-01
-1.00128365e+00 -7.33392656e-01 9.65804994e-01 9.89678800e-01
6.90919936e-01 -1.58763242e+00 -1.03641108e-01 3.22760701e-01
-1.33403957e-01 -1.50874352e+00 -1.08027899e+00 1.52853891e-01
-6.17512226e-01 -9.75142300e-01 -9.94640112e-01 -1.26131177e+00
7.70812035e-01 7.22067505e-02 8.88275206e-01 -2.35604662e-02
-1.38047338e-01 5.39277419e-02 -6.28352880e-01 -4.91054714e-01
-2.85753012e-01 6.53913021e-01 -2.12530598e-01 1.59467515e-02
8.98536801e-01 -1.23024009e-01 -5.14675081e-01 2.21064135e-01
-8.80527318e-01 -3.08522701e-01 8.76320899e-01 8.06511581e-01
4.19659346e-01 1.31046832e-01 8.33959639e-01 -1.55188024e+00
5.44638336e-01 -5.51068008e-01 -2.80910313e-01 5.32666028e-01
-2.18923539e-01 8.80344138e-02 1.03460121e+00 -5.77618957e-01
-1.58713698e+00 1.11266457e-01 -3.43064696e-01 4.39607836e-02
-7.45721221e-01 4.71951336e-01 -6.98521495e-01 2.78654158e-01
2.05742672e-01 2.10706696e-01 -8.73387396e-01 -7.62577772e-01
2.15465128e-01 9.19807076e-01 4.16329384e-01 -5.04881561e-01
7.60565341e-01 1.14458971e-01 -6.88742042e-01 -8.22815716e-01
-9.35322165e-01 -7.87275553e-01 -9.81738448e-01 4.14695829e-01
1.29853439e+00 -8.46357763e-01 -2.67038733e-01 7.51453280e-01
-1.39331281e+00 -2.79861867e-01 5.44555206e-03 5.24924338e-01
1.01521939e-01 5.04223704e-01 -9.59708512e-01 -7.10713148e-01
-6.15626752e-01 -7.66188681e-01 7.31177926e-01 8.38914156e-01
1.26807705e-01 -1.07499492e+00 -1.44803720e-02 1.65009454e-01
3.98461699e-01 1.83486879e-01 6.65640533e-01 -1.34022057e+00
-2.18036905e-01 -2.91671425e-01 -3.35011840e-01 4.49103236e-01
2.71235555e-01 -2.96244651e-01 -7.56366611e-01 6.35627098e-03
-2.00441808e-01 -2.41940115e-02 1.01730072e+00 -2.14841720e-02
9.67393219e-01 -1.33498922e-01 -3.78570080e-01 4.79675025e-01
1.48486555e+00 2.98393518e-01 6.60441399e-01 2.40375698e-01
9.94424939e-01 6.01537883e-01 7.13859200e-01 1.11239053e-01
3.19393843e-01 1.39187677e-02 -6.06030710e-02 -2.20851272e-01
1.36932746e-01 -2.25369126e-01 3.10958505e-01 9.96698618e-01
1.09201372e-02 -4.37347978e-01 -1.08428752e+00 7.19044507e-01
-1.61121929e+00 -6.86061382e-01 -2.35795602e-01 1.92668819e+00
1.10556030e+00 2.48030350e-01 -4.19415683e-01 -3.21156710e-01
1.31644058e+00 1.33709013e-01 -5.73505938e-01 -4.80970383e-01
-2.98653170e-02 7.03520119e-01 8.21892440e-01 1.74749032e-01
-1.64177418e+00 1.51522100e+00 5.28406763e+00 1.03863871e+00
-1.00512409e+00 1.09371103e-01 5.16645908e-01 8.20409417e-01
-2.75220662e-01 -7.80766308e-02 -1.20571601e+00 4.66919333e-01
8.58758509e-01 -3.96839865e-02 -7.30677322e-02 7.48153865e-01
-8.54085013e-02 1.24804027e-01 -8.11480045e-01 3.81348997e-01
-1.59040883e-01 -7.93146610e-01 -1.90874964e-01 -1.70797989e-01
8.07819903e-01 1.33502617e-01 -3.32048535e-01 5.35984337e-01
6.17462754e-01 -9.79719698e-01 3.26965094e-01 3.69611204e-01
7.73238957e-01 -8.76846373e-01 1.04595459e+00 4.00685132e-01
-1.45295966e+00 2.12345168e-01 -6.66433096e-01 7.78867975e-02
9.49295089e-02 3.92380029e-01 -5.24595559e-01 6.58876896e-01
5.65386593e-01 5.48755407e-01 -4.21282411e-01 1.09574616e+00
-6.72837317e-01 1.10231125e+00 -1.79000244e-01 -2.52155095e-01
4.66709435e-01 -1.69086799e-01 2.03003094e-01 1.76935649e+00
2.29172543e-01 3.37071508e-01 1.09299213e-01 7.25003898e-01
-3.64172339e-01 4.12324011e-01 -2.06424251e-01 -2.98960567e-01
6.05421841e-01 1.35621428e+00 -1.04621005e+00 -1.55562386e-01
-7.33263075e-01 1.45406234e+00 5.61443686e-01 3.33465785e-01
-9.02651310e-01 -1.12953269e+00 7.27552354e-01 -3.44593853e-01
7.91429520e-01 -3.21210593e-01 -2.44777694e-01 -1.43415940e+00
-3.81318815e-02 -4.42385882e-01 4.02292907e-01 -2.19015777e-01
-1.58147919e+00 6.64007485e-01 -5.66015959e-01 -1.04391730e+00
1.36507079e-01 -6.34222269e-01 -1.20185626e+00 1.04035032e+00
-1.95327556e+00 -1.30350065e+00 1.24768522e-02 4.36441541e-01
6.28395140e-01 6.05815388e-02 9.26280439e-01 4.45704132e-01
-9.43062603e-01 9.00688708e-01 4.19421822e-01 1.18646538e+00
9.88983154e-01 -1.51114225e+00 8.21616471e-01 1.21136487e+00
8.24885294e-02 7.84764290e-01 2.24004790e-01 -9.07737792e-01
-8.79984796e-01 -1.33517694e+00 1.22540426e+00 -1.47401735e-01
5.80905914e-01 -5.96088946e-01 -1.08101821e+00 7.32093155e-01
2.82673568e-01 2.41460264e-01 1.01292002e+00 -5.42613212e-03
-2.61461496e-01 1.82255268e-01 -1.03503156e+00 5.53776026e-01
9.38374996e-01 -2.93646783e-01 -9.55963194e-01 -1.53443411e-01
8.27005565e-01 -3.66616160e-01 -7.30429471e-01 2.11068839e-01
1.02829546e-01 -4.83696610e-01 6.69460714e-01 -9.13163424e-01
2.15171337e-01 -1.70120522e-01 2.25711867e-01 -1.37278438e+00
-2.08611965e-01 -5.42842925e-01 3.97655815e-01 1.95771492e+00
6.23566270e-01 -6.46617532e-01 7.44458795e-01 8.91435921e-01
-3.88235271e-01 -4.49442327e-01 -6.35874867e-01 -7.49780774e-01
5.44394076e-01 -2.39378914e-01 6.90177500e-01 1.24452138e+00
-2.65596390e-01 3.29211056e-01 -2.49675021e-01 4.17622536e-01
2.36238793e-01 1.55939092e-03 2.18142033e-01 -1.08994126e+00
3.36142741e-02 -1.22991450e-01 -2.27723703e-01 -1.04825544e+00
5.31318843e-01 -6.48195744e-01 3.52297366e-01 -1.52477169e+00
5.49406931e-02 -7.52773166e-01 -7.36548245e-01 6.14269435e-01
-7.76639462e-01 1.24847844e-01 2.35805571e-01 -3.18046957e-02
-7.03660071e-01 3.53038162e-01 1.04984641e+00 5.01881838e-02
-2.59875596e-01 2.14474853e-02 -7.42424846e-01 8.30460310e-01
9.92636263e-01 -6.51717365e-01 1.25690520e-01 -7.06131220e-01
-3.57264131e-02 -2.78538913e-01 -3.20442647e-01 -8.52427900e-01
3.79359663e-01 -3.32979679e-01 5.28945625e-01 -4.18124586e-01
-3.01007241e-01 -7.49964416e-01 -5.31743586e-01 1.86199203e-01
-1.98865816e-01 -1.89809665e-01 1.79194152e-01 7.03600168e-01
-2.86583334e-01 -6.17108464e-01 5.92369139e-01 -2.14592829e-01
-1.19858265e+00 4.13223833e-01 -5.40020287e-01 7.45127916e-01
1.09521186e+00 -9.75256041e-02 -1.44927740e-01 1.44704819e-01
-5.14429212e-01 5.08687377e-01 9.96683463e-02 6.25099301e-01
3.70979309e-01 -1.08173621e+00 -7.48985469e-01 5.34250438e-02
-9.65549499e-02 2.40718678e-01 4.32953686e-01 2.45267287e-01
-4.78095412e-01 4.17821586e-01 -1.35131881e-01 1.40685469e-01
-9.93267179e-01 5.28434277e-01 1.23534247e-01 -5.55240929e-01
-3.52241307e-01 8.56955707e-01 2.57450074e-01 -7.06856072e-01
-3.40334997e-02 -3.71005058e-01 -6.62977278e-01 -1.99129563e-02
6.06869102e-01 1.98771596e-01 -1.35084033e-01 -8.00033987e-01
-5.65845251e-01 5.89657903e-01 -3.57137382e-01 1.90871358e-01
1.28070498e+00 -1.08828627e-01 8.14958215e-02 1.73849821e-01
1.11377442e+00 3.56984347e-01 -1.14311063e+00 -3.43378901e-01
5.90393245e-01 7.12765148e-03 -2.24918425e-01 -6.97489977e-01
-1.04450846e+00 1.20524311e+00 2.18052626e-01 1.78217608e-02
1.04142833e+00 -2.16458336e-01 1.49587154e+00 3.14060271e-01
1.50805101e-01 -1.37319767e+00 -4.01308924e-01 9.88304019e-01
2.00561509e-01 -1.38207340e+00 -5.51796734e-01 -6.69935405e-01
-8.71689737e-01 1.08165038e+00 1.01934636e+00 -3.00460815e-01
6.41907215e-01 3.47850353e-01 2.20395103e-01 5.28512657e-01
-1.65468752e-01 -5.52264273e-01 2.26865694e-01 7.67257631e-01
6.15825415e-01 1.02661224e-02 -7.59078920e-01 1.28440392e+00
1.87899590e-01 -2.30629086e-01 4.06266123e-01 9.34582174e-01
-3.94934922e-01 -1.45718968e+00 -8.68875757e-02 2.88342834e-01
-1.25649691e+00 -5.56612313e-01 -4.13020998e-01 8.09742033e-01
1.39167011e-01 9.64178681e-01 1.19459517e-01 -1.15138076e-01
2.27815002e-01 1.99537560e-01 -3.13057825e-02 -8.64223242e-01
-9.11441207e-01 -1.52492151e-01 4.15164195e-02 -1.08260311e-01
-2.29071841e-01 -5.57285428e-01 -1.73618770e+00 7.62454048e-02
-6.66830719e-01 5.36603808e-01 4.28850234e-01 1.01439464e+00
4.02124196e-01 4.49337035e-01 5.71677566e-01 -3.05739552e-01
-5.01868308e-01 -1.03004038e+00 -5.54980755e-01 5.32646358e-01
4.95724343e-02 -1.75493702e-01 -1.17629729e-01 -3.68152149e-02] | [9.809525489807129, 9.854578018188477] |
e028eef3-2a06-4273-a910-efd1b36cf7a4 | deformable-graph-convolutional-networks | 2112.14438 | null | https://arxiv.org/abs/2112.14438v1 | https://arxiv.org/pdf/2112.14438v1.pdf | Deformable Graph Convolutional Networks | Graph neural networks (GNNs) have significantly improved the representation power for graph-structured data. Despite of the recent success of GNNs, the graph convolution in most GNNs have two limitations. Since the graph convolution is performed in a small local neighborhood on the input graph, it is inherently incapable to capture long-range dependencies between distance nodes. In addition, when a node has neighbors that belong to different classes, i.e., heterophily, the aggregated messages from them often negatively affect representation learning. To address the two common problems of graph convolution, in this paper, we propose Deformable Graph Convolutional Networks (Deformable GCNs) that adaptively perform convolution in multiple latent spaces and capture short/long-range dependencies between nodes. Separated from node representations (features), our framework simultaneously learns the node positional embeddings (coordinates) to determine the relations between nodes in an end-to-end fashion. Depending on node position, the convolution kernels are deformed by deformation vectors and apply different transformations to its neighbor nodes. Our extensive experiments demonstrate that Deformable GCNs flexibly handles the heterophily and achieve the best performance in node classification tasks on six heterophilic graph datasets. | ['Hyunwoo J. Kim', 'Jihwan Park', 'Sungdong Yoo', 'Jinyoung Park'] | 2021-12-29 | null | null | null | null | ['node-classification-on-non-homophilic'] | ['graphs'] | [-1.85525000e-01 1.75300911e-01 -1.15311123e-01 -3.38860095e-01
2.59646684e-01 -7.02691078e-01 4.51893628e-01 3.52253735e-01
-9.52592939e-02 3.03920209e-01 1.65001988e-01 -2.81419992e-01
-2.08567232e-01 -1.53835130e+00 -6.11374140e-01 -9.34349954e-01
-4.83791202e-01 6.02803051e-01 1.95028245e-01 -2.81855613e-01
-1.16514884e-01 9.83209789e-01 -1.02690995e+00 1.51690707e-01
7.41068065e-01 6.04474664e-01 -1.84906013e-02 7.22685933e-01
-2.57590175e-01 4.39762533e-01 -4.48191285e-01 -4.11153555e-01
1.53901502e-01 -1.60968363e-01 -6.18500710e-01 -3.59030515e-01
4.56740439e-01 -1.14527509e-01 -9.40824032e-01 1.09345007e+00
2.93721735e-01 2.53137976e-01 5.95219135e-01 -1.33977032e+00
-1.34259820e+00 8.11272740e-01 -3.26253235e-01 2.65421778e-01
3.02250609e-02 -9.91927162e-02 1.25876904e+00 -5.56411147e-01
6.73453152e-01 1.48286712e+00 7.95188725e-01 4.64734137e-01
-1.30447733e+00 -6.06857955e-01 4.44912881e-01 -1.73683867e-01
-1.35763884e+00 1.68746367e-01 8.86348248e-01 -2.40784422e-01
9.38900352e-01 2.99069792e-01 7.89264798e-01 9.83488500e-01
3.39710653e-01 2.80731410e-01 3.35278213e-01 9.50826555e-02
-1.24757864e-01 -4.58695590e-01 1.73901677e-01 9.06567097e-01
3.93025696e-01 -2.04149812e-01 -1.18314095e-01 -2.75372833e-01
1.11644220e+00 7.36734807e-01 -2.43912250e-01 -4.27484304e-01
-1.06300473e+00 1.03690612e+00 1.35757005e+00 5.31405985e-01
-1.07813425e-01 6.29619777e-01 3.75644118e-01 5.44053793e-01
5.24304926e-01 2.08309591e-01 -1.94679350e-01 3.51486802e-01
-2.54318863e-01 -7.83313438e-02 9.10748422e-01 6.59324408e-01
1.05211842e+00 -1.61273196e-01 -2.16514960e-01 5.82227111e-01
3.76302600e-01 3.40030015e-01 3.32331568e-01 -5.12692183e-02
4.72236753e-01 1.21779883e+00 -5.18943489e-01 -1.78646898e+00
-5.37206411e-01 -4.80956703e-01 -1.27480876e+00 -1.88425586e-01
2.37751633e-01 1.08843753e-02 -1.05019712e+00 1.77457118e+00
1.85603932e-01 5.47878027e-01 -8.80608931e-02 7.12386310e-01
1.29487753e+00 7.45692611e-01 2.56036490e-01 3.70514810e-01
1.07121515e+00 -7.39646196e-01 -4.52576667e-01 -6.74106255e-02
8.67456257e-01 -4.01428729e-01 9.16409314e-01 -5.40301263e-01
-6.43865049e-01 -3.43830973e-01 -8.39762330e-01 -1.32968873e-01
-7.63120353e-01 -4.30403054e-01 9.04603839e-01 3.40474188e-01
-1.33800113e+00 8.18810523e-01 -1.08077073e+00 -5.55630505e-01
4.33413357e-01 5.61764777e-01 -4.60915774e-01 -1.01723112e-01
-1.32215643e+00 3.03143471e-01 2.35192731e-01 2.33727783e-01
-6.00252569e-01 -5.16443193e-01 -9.87241387e-01 5.87473869e-01
-1.02565130e-02 -5.34180105e-01 3.46335977e-01 -8.16202760e-01
-1.11539698e+00 7.26273239e-01 2.23380193e-01 -3.24587673e-01
1.95643947e-01 2.05406025e-01 -4.09454286e-01 1.06448360e-01
-1.30873233e-01 4.65136707e-01 5.68946004e-01 -8.79552186e-01
4.58983355e-04 -4.12942201e-01 2.79757768e-01 1.60617098e-01
-7.49486327e-01 -2.76268899e-01 -5.57681739e-01 -8.28103483e-01
3.34798396e-01 -1.13361228e+00 -1.78861588e-01 2.67811984e-01
-3.71927828e-01 -4.28889811e-01 1.16510642e+00 -1.38302445e-01
1.25630832e+00 -2.24315238e+00 4.00156528e-01 5.31735003e-01
9.09329474e-01 3.05777013e-01 -4.15451080e-01 7.79066861e-01
-1.02309309e-01 3.19186956e-01 -7.54842609e-02 7.91973621e-02
-2.30323955e-01 6.35318398e-01 -1.20041028e-01 5.74208260e-01
8.35202262e-02 1.42752373e+00 -1.10217023e+00 -3.25090259e-01
1.16058789e-01 8.41852307e-01 -5.03747225e-01 2.27720529e-01
-8.41038004e-02 1.35587394e-01 -8.07521999e-01 3.75228822e-01
7.72013366e-01 -6.97693825e-01 3.52678269e-01 -2.63785303e-01
4.08057302e-01 2.59094220e-02 -8.36831450e-01 1.39570832e+00
-3.41080099e-01 7.03596234e-01 9.95782912e-02 -1.11515903e+00
1.12881505e+00 8.47031623e-02 4.96091783e-01 -2.97123551e-01
-8.28697998e-03 -1.12726338e-01 1.55143157e-01 -3.18948179e-01
9.69396830e-02 1.87380731e-01 -7.52058476e-02 5.19187093e-01
8.95763412e-02 2.79842615e-01 -5.07534742e-02 5.20491660e-01
1.44886363e+00 -5.29361546e-01 -3.12567949e-02 -3.82205784e-01
3.20765734e-01 -5.13598859e-01 3.81702900e-01 5.19615233e-01
-1.09311873e-02 5.75675905e-01 8.02171469e-01 -9.04890656e-01
-6.23805881e-01 -1.24731207e+00 1.03727110e-01 1.33544135e+00
4.37438846e-01 -6.21738374e-01 -4.33186114e-01 -9.29879963e-01
4.36126500e-01 -7.30996057e-02 -1.00317609e+00 -4.97760564e-01
-8.40350091e-01 -8.06759536e-01 6.51296139e-01 7.13728964e-01
3.94897550e-01 -1.04315615e+00 -9.64207482e-03 1.19087033e-01
2.33652726e-01 -9.09914434e-01 -6.61995590e-01 1.27384096e-01
-7.83931553e-01 -1.25446892e+00 -4.24375296e-01 -1.07199156e+00
1.11994934e+00 5.48391104e-01 1.11588776e+00 8.22520494e-01
-2.03571588e-01 2.06970334e-01 -3.73048902e-01 2.41153836e-01
-2.28838027e-01 3.58587444e-01 -2.06372827e-01 1.17133118e-01
3.13670099e-01 -8.60652685e-01 -6.63342714e-01 4.24009830e-01
-8.67684126e-01 -1.28164232e-01 4.85462666e-01 9.20895159e-01
5.08292735e-01 1.11407854e-01 1.96470454e-01 -1.34553564e+00
8.57586026e-01 -7.33645380e-01 -4.27685827e-01 3.66569012e-01
-4.34238076e-01 1.29070848e-01 8.38554442e-01 -6.13091528e-01
-4.87156570e-01 -1.85201883e-01 9.55640078e-02 -4.91215497e-01
1.71772987e-01 6.98170543e-01 -5.83513975e-02 -4.86612588e-01
6.22400701e-01 -7.61960149e-02 4.36543934e-02 -3.06556106e-01
4.01873320e-01 2.21367449e-01 1.52914017e-01 -3.68383527e-01
1.06259859e+00 4.55372512e-01 2.34738514e-01 -7.52516925e-01
-3.54596525e-01 -1.59835875e-01 -7.05878377e-01 -9.29371081e-03
8.68220747e-01 -5.82421243e-01 -7.60769367e-01 3.57638270e-01
-8.92170072e-01 -5.19905031e-01 -8.95736739e-02 1.29076511e-01
1.38409242e-01 2.90555477e-01 -9.11473155e-01 -1.09984174e-01
-4.41249013e-01 -1.05666268e+00 8.01884592e-01 3.62194121e-01
-1.30871683e-02 -1.57138193e+00 -1.44078853e-02 -3.07911038e-01
6.64093614e-01 5.96547663e-01 1.24403143e+00 -6.88353837e-01
-6.54190421e-01 -4.42794472e-01 -5.04184365e-01 -9.36386138e-02
3.15983146e-01 2.01999232e-01 -4.82512265e-01 -6.57022595e-01
-7.25431323e-01 -4.24423208e-03 9.75777864e-01 2.30092287e-01
1.30090380e+00 -3.40281039e-01 -8.33704948e-01 1.11905897e+00
1.18726921e+00 -2.09019154e-01 5.25232196e-01 -2.00901195e-01
1.29291773e+00 3.08524966e-01 -1.78454876e-01 -4.18462865e-02
2.78637916e-01 3.61479104e-01 7.60325015e-01 -2.01336443e-01
-2.38157600e-01 -4.50713068e-01 -1.19085815e-02 7.01262176e-01
-2.05933586e-01 -7.10613012e-01 -7.78224289e-01 2.88587719e-01
-1.87241483e+00 -6.93053007e-01 -1.42320797e-01 1.84405756e+00
2.73567051e-01 -2.50484794e-02 -1.22127533e-01 -3.97650361e-01
9.26870406e-01 6.59644902e-01 -5.59252799e-01 -2.78273493e-01
-8.30472112e-02 1.36788800e-01 5.78197122e-01 4.21138674e-01
-9.40841556e-01 1.04416251e+00 5.85756874e+00 3.98895293e-01
-1.33219576e+00 -9.39204767e-02 5.46113729e-01 9.33683813e-02
-6.96447313e-01 -4.44779396e-02 -4.36935127e-01 4.09177512e-01
5.30710697e-01 -3.41906995e-02 6.54469669e-01 7.56384850e-01
-4.52350557e-01 6.05940640e-01 -9.85031903e-01 8.95510375e-01
-1.08336017e-01 -1.53453660e+00 3.17283303e-01 2.14854106e-01
7.21997857e-01 3.05299610e-01 1.40037984e-01 1.60890773e-01
6.13016844e-01 -1.48733020e+00 3.80942672e-02 3.04218441e-01
8.72238040e-01 -7.62321532e-01 6.78671360e-01 7.58686885e-02
-1.69614100e+00 5.60189523e-02 -6.73553050e-01 -4.25741710e-02
-1.73014686e-01 4.61645395e-01 -8.21934760e-01 3.62862378e-01
6.94195807e-01 9.70753849e-01 -6.39658034e-01 5.38496256e-01
-2.74790645e-01 4.50448334e-01 -3.28334868e-01 -3.04469615e-01
4.52713519e-01 -3.41112196e-01 4.48108077e-01 1.12401128e+00
2.35772848e-01 5.72928973e-02 1.97149649e-01 9.98727977e-01
-5.44063628e-01 -4.91355658e-02 -7.89192021e-01 -4.38920408e-01
5.37259102e-01 1.53954768e+00 -1.07363462e+00 3.09423842e-02
-4.33679432e-01 1.05950999e+00 9.98370111e-01 6.72544479e-01
-6.03837192e-01 -5.79723835e-01 9.88852262e-01 3.64544570e-01
2.94781804e-01 -3.99779439e-01 1.52393147e-01 -1.11908758e+00
-1.73123166e-01 -3.86014968e-01 6.32715523e-01 -1.99016690e-01
-1.62091899e+00 7.46887863e-01 -3.07001501e-01 -7.68724442e-01
2.49745712e-01 -6.91066921e-01 -1.02872181e+00 8.11932087e-01
-1.11607230e+00 -1.41296792e+00 -6.61639869e-01 7.08629310e-01
-1.98591143e-01 -1.50836064e-02 9.46089327e-01 3.16626638e-01
-4.59152251e-01 8.01082969e-01 8.10649842e-02 7.63560593e-01
4.55584347e-01 -1.18038440e+00 7.40552306e-01 4.96820688e-01
2.87999928e-01 1.04096246e+00 1.96919635e-01 -7.95382679e-01
-1.53310597e+00 -1.34956753e+00 6.74503028e-01 -2.18541011e-01
7.50378549e-01 -7.92272627e-01 -1.21521389e+00 1.03101885e+00
-2.72050619e-01 9.02894914e-01 6.55649722e-01 3.07425261e-01
-6.39271975e-01 -1.88791856e-01 -1.03437054e+00 6.45866752e-01
1.52353859e+00 -7.37948775e-01 2.51181107e-02 4.22018081e-01
8.87809455e-01 -4.62759286e-01 -1.03889465e+00 3.62317115e-01
4.35052216e-01 -7.12806940e-01 1.07790983e+00 -8.86468828e-01
1.23989843e-01 -1.33571684e-01 6.84594139e-02 -1.46910965e+00
-7.27935672e-01 -5.73345721e-01 -1.91978812e-01 9.76013124e-01
2.61921793e-01 -1.10991991e+00 9.69180763e-01 4.85480160e-01
3.83058563e-02 -8.81208420e-01 -8.15138817e-01 -4.44046080e-01
1.08649753e-01 2.45241269e-01 9.60763693e-01 1.16751623e+00
-2.03851402e-01 3.06190550e-01 -7.65859559e-02 3.90310735e-01
3.09277236e-01 3.91381502e-01 7.09217191e-01 -1.60357392e+00
-8.97958279e-02 -5.97126842e-01 -9.94950831e-01 -1.13490558e+00
4.19278920e-01 -1.44398201e+00 -4.46265072e-01 -1.62687194e+00
7.00916648e-02 -8.36099029e-01 -4.71686423e-01 5.69200635e-01
-2.13770583e-01 2.32509732e-01 -5.17163426e-02 1.35928601e-01
-4.74179983e-01 4.26217794e-01 1.47426438e+00 -2.68769801e-01
-3.02092761e-01 -1.11142181e-01 -5.65946519e-01 4.48957890e-01
6.86137617e-01 -3.91518593e-01 -5.08065224e-01 -6.95281506e-01
5.00796258e-01 -1.43217698e-01 5.47658682e-01 -7.58890927e-01
2.16804788e-01 -4.70550172e-02 3.98340911e-01 -3.25333774e-01
2.17498969e-02 -8.74831259e-01 4.01372641e-01 6.26097441e-01
-3.70319217e-01 2.58889824e-01 -1.59505457e-01 9.94077802e-01
-2.19775476e-02 1.84291333e-01 6.77455664e-01 -8.41334760e-02
-4.19130564e-01 9.96067762e-01 -4.38048877e-02 -4.25354317e-02
1.03246975e+00 5.51253883e-03 -5.36656737e-01 -4.03587162e-01
-7.08249152e-01 3.39736730e-01 6.68602049e-01 4.67976540e-01
5.65276980e-01 -1.45264864e+00 -4.71172869e-01 3.94348383e-01
1.91380247e-01 3.10874850e-01 1.96407944e-01 5.01997411e-01
-8.20414543e-01 1.74683928e-02 -1.49091959e-01 -5.19295931e-01
-1.24622512e+00 5.40174961e-01 5.44589877e-01 -3.21089089e-01
-9.77642238e-01 1.14716387e+00 3.75675172e-01 -6.78010285e-01
1.33683860e-01 -4.16338205e-01 -2.36532435e-01 -1.63023070e-01
1.24985576e-01 1.41466230e-01 -1.42218351e-01 -6.90165102e-01
-3.07495862e-01 6.07864857e-01 -1.97933361e-01 6.79599583e-01
1.52135956e+00 1.88591018e-01 -4.96628493e-01 8.97463337e-02
1.52189147e+00 -4.82550934e-02 -1.12799561e+00 -3.76118094e-01
-4.06093985e-01 -5.24669170e-01 -1.21290363e-01 -1.24793470e-01
-1.59624720e+00 8.37607920e-01 3.11215043e-01 6.94568038e-01
6.84771597e-01 3.42519999e-01 9.24381077e-01 5.05629897e-01
-2.94501521e-02 -6.05753422e-01 2.96723187e-01 5.47806680e-01
6.98652267e-01 -1.00270379e+00 -1.29250363e-01 -6.88910604e-01
-1.40470430e-01 1.36381114e+00 8.13741565e-01 -5.39413035e-01
9.60420012e-01 2.65623420e-01 -1.69298619e-01 -6.91515565e-01
-3.78271937e-01 4.21987996e-02 2.53286600e-01 7.35295475e-01
4.37169760e-01 3.79761308e-01 5.18397763e-02 1.52093813e-01
3.41245495e-02 -7.29903042e-01 3.63137685e-02 5.74975729e-01
-2.54856974e-01 -1.01875901e+00 1.45017833e-01 6.25096858e-01
-1.41894922e-01 -7.89647847e-02 -6.37748539e-01 7.87028790e-01
-1.06717668e-01 4.65491921e-01 3.57185364e-01 -5.52997351e-01
4.58171554e-02 -2.31322035e-01 2.45587751e-01 -7.53613770e-01
-6.36881053e-01 -3.58530968e-01 -3.00766617e-01 -5.20997047e-01
-7.93148205e-02 -1.42273009e-01 -1.64811194e+00 -6.88184381e-01
-1.48691505e-01 6.84682801e-02 1.25856370e-01 5.60432613e-01
6.39113128e-01 6.01555824e-01 5.37335336e-01 -6.91186965e-01
-2.68416137e-01 -7.83004820e-01 -6.89264119e-01 8.75512719e-01
2.87770420e-01 -6.20920718e-01 -4.55249220e-01 -5.01956701e-01] | [7.075328350067139, 6.221421718597412] |
87580b98-9b30-44d0-b298-34adf1a77787 | porter-5-fast-state-of-the-art-ab-initio | null | null | https://doi.org/10.1101/289033 | https://www.biorxiv.org/content/early/2018/10/05/289033.full.pdf | Porter 5: fast, state-of-the-art ab initio prediction of protein secondary structure in 3 and 8 classes | Motivation: Although secondary structure predictors have been developed for decades, current ab initio methods have still some way to go to reach their theoretical limits. Moreover, the continuous effort towards harnessing ever-expanding data sets and more sophisticated, deeper Machine Learning techniques, has not come to an end.
Results: Here we present Porter 5, the latest release of one of the best performing ab initio secondary structure predictors. Version 5 achieves 84% accuracy (84% SOV) when tested on 3 classes, and 73% accuracy (77% SOV) on 8 classes, on a large independent set, significantly outperforming all the most recent ab initio predictors we have tested.
Availability: The web and standalone versions of Porter5 are available at http://distilldeep.ucd.ie/porter/. | ['Mirko Torrisi', 'Gianluca Pollastri', 'Manaz Kaleel'] | 2018-10-05 | null | null | null | biorxiv-2018-10 | ['protein-secondary-structure-prediction'] | ['medical'] | [ 1.30982352e-02 -9.74272043e-02 -3.88910830e-01 -4.47350740e-01
-1.32846010e+00 -5.77004790e-01 4.54378039e-01 3.16309363e-01
-3.94539058e-01 1.43786025e+00 8.84909183e-02 -7.54447401e-01
4.83067222e-02 -3.74982148e-01 -5.41401267e-01 -1.04162431e+00
-8.98151845e-02 8.15709531e-01 4.46570575e-01 -3.51445019e-01
5.20768940e-01 4.40013170e-01 -1.58074307e+00 3.79378706e-01
1.09619677e+00 5.36295533e-01 3.63776892e-01 6.51818156e-01
1.53576538e-01 4.34941888e-01 -2.44394779e-01 -4.01523322e-01
-1.25549018e-01 -5.16301453e-01 -1.22561800e+00 -5.53656638e-01
3.95919055e-01 1.06340975e-01 1.54397011e-01 5.52858174e-01
6.12662494e-01 -1.95018142e-01 6.63542032e-01 -4.96059537e-01
-6.39226556e-01 2.51369774e-01 -2.81331986e-01 3.01442236e-01
5.96163273e-01 3.06966126e-01 1.49875402e+00 -9.49919343e-01
9.35645819e-01 8.18454087e-01 5.76124787e-01 4.98092741e-01
-1.70861137e+00 -6.06619358e-01 -3.08106065e-01 3.42954814e-01
-1.15673542e+00 -3.88662070e-01 3.59775573e-01 -3.43313634e-01
1.63669491e+00 2.54167080e-01 8.04187775e-01 9.74348605e-01
4.89841253e-01 6.74487472e-01 1.53649271e+00 -3.64062876e-01
2.16454431e-01 -5.58702946e-02 4.50648367e-01 6.53284132e-01
3.11039001e-01 9.79067385e-02 -5.27634978e-01 -6.52869403e-01
1.79589406e-01 -4.59829062e-01 -3.33774209e-01 -3.37852955e-01
-9.81617033e-01 8.63874376e-01 3.47590357e-01 2.90507972e-01
-4.23192263e-01 -4.25683588e-01 1.05473109e-01 2.07017154e-01
4.42588747e-01 6.57214284e-01 -1.02921605e+00 -5.15158892e-01
-8.18855047e-01 6.04490519e-01 9.65148211e-01 6.23165607e-01
8.29852045e-01 -3.26712310e-01 6.69661343e-01 7.30753899e-01
8.91883150e-02 2.97251940e-01 3.82935345e-01 -5.56046486e-01
2.07080975e-01 4.62975621e-01 2.81386584e-01 -2.87515342e-01
-6.42015517e-01 -2.37614304e-01 -3.42504084e-01 2.27520585e-01
5.92420399e-01 -3.25182639e-02 -7.85962164e-01 1.43781459e+00
2.79748797e-01 -3.43845725e-01 1.27452880e-01 8.69629264e-01
7.07363307e-01 8.00407290e-01 1.80349752e-01 -4.62076724e-01
9.98489857e-01 -5.08839190e-01 -2.69499391e-01 1.22294053e-01
9.37434554e-01 -1.04704881e+00 8.60087276e-01 7.72840738e-01
-1.05807519e+00 -2.66368061e-01 -1.07504666e+00 -4.09196198e-01
-3.14613611e-01 -1.89899474e-01 1.01906443e+00 6.69821978e-01
-8.69648516e-01 8.01470935e-01 -9.35506761e-01 -4.37730551e-01
1.54671818e-01 6.51176274e-01 -6.67070687e-01 -4.89029742e-04
-1.16753054e+00 1.07296395e+00 3.41323227e-01 -2.13795885e-01
-6.35069907e-01 -6.82371497e-01 -2.40801394e-01 -2.19738126e-01
2.73198724e-01 -3.57260615e-01 1.20204556e+00 -6.21574581e-01
-1.20178366e+00 1.07318377e+00 -4.05127555e-01 -4.15220469e-01
2.12852299e-01 -4.42785650e-01 -3.35673988e-01 1.73483063e-02
-1.32995322e-01 4.92885441e-01 -2.20426425e-01 -1.00665689e+00
-6.10647023e-01 -6.60775661e-01 -1.47955328e-01 1.47293687e-01
2.09859446e-01 2.13321403e-01 2.99385548e-01 -2.36154318e-01
2.44526163e-01 -1.04773498e+00 -3.04908603e-01 -3.65818471e-01
-3.99528921e-01 -4.14987147e-01 5.33695161e-01 -8.12692344e-01
1.03778410e+00 -1.51802862e+00 4.18047369e-01 9.15335417e-02
2.02956975e-01 5.72245717e-01 1.27992854e-01 9.22906697e-01
-4.57777828e-01 1.01968028e-01 -4.41488564e-01 7.69931823e-02
-3.01817000e-01 1.67384431e-01 -9.96639580e-02 5.21781623e-01
5.53128906e-02 5.06704092e-01 -8.60882819e-01 -2.13712081e-01
1.19949296e-01 5.76548100e-01 -6.83156013e-01 -2.06486657e-01
-2.32600987e-01 4.35570449e-01 -3.54109079e-01 5.61663866e-01
5.24882555e-01 -4.24414605e-01 6.58867419e-01 1.87169075e-01
-5.76507449e-01 7.71096826e-01 -7.63513923e-01 1.57823122e+00
1.18517883e-01 2.87118852e-01 -1.19353412e-02 -1.00536859e+00
1.01506352e+00 1.95149004e-01 7.80077755e-01 -4.36370701e-01
-7.92971998e-02 4.73029941e-01 6.31858885e-01 -1.36486962e-01
8.82673264e-02 -3.43278736e-01 3.29937249e-01 1.88485697e-01
6.90105408e-02 -4.13269848e-02 2.45977849e-01 2.77241528e-01
1.04490948e+00 4.12895113e-01 6.34522974e-01 -6.79721653e-01
6.40119493e-01 5.63442171e-01 6.71998560e-01 3.11172634e-01
-3.77615601e-01 3.96341503e-01 5.52449822e-01 -6.82380915e-01
-1.35757864e+00 -9.57688749e-01 -7.81675220e-01 1.06263888e+00
-2.93962836e-01 -7.73336291e-01 -7.96296120e-01 -4.52250630e-01
-7.68303052e-02 5.54425776e-01 -2.46010214e-01 1.46518067e-01
-4.10829723e-01 -1.22933698e+00 1.68103471e-01 2.08515197e-01
4.47150767e-02 -8.55548024e-01 8.80184025e-02 3.75771433e-01
-1.67598501e-02 -5.42041779e-01 8.74240473e-02 5.60870111e-01
-1.09360671e+00 -1.19941115e+00 -4.39087749e-01 -7.09750652e-01
6.27766028e-02 1.61064476e-01 1.11150491e+00 3.34344596e-01
-5.42887628e-01 -3.53797048e-01 -3.24451834e-01 -3.30824196e-01
-4.14901078e-01 4.15345639e-01 3.14916342e-01 -8.55915308e-01
8.97361040e-01 -8.80085528e-01 -6.05600595e-01 3.48062456e-01
-4.00041342e-01 2.44005788e-02 6.68525755e-01 1.01617181e+00
6.16606534e-01 -4.00272071e-01 5.52706897e-01 -9.35266495e-01
1.72581732e-01 -4.09276038e-01 -6.43849313e-01 9.89013761e-02
-9.15587723e-01 1.15958914e-01 7.16289043e-01 2.76692837e-01
-9.23510730e-01 1.98893234e-01 -9.12183523e-01 5.78329325e-01
-3.92320812e-01 4.54137683e-01 -1.49580657e-01 1.47680625e-01
8.54209661e-01 3.04129153e-01 3.85509757e-03 -7.73486912e-01
7.05943182e-02 6.54835165e-01 9.32563618e-02 -7.95357466e-01
5.35737395e-01 2.55035669e-01 6.29073381e-02 -1.07117605e+00
-8.31656277e-01 -6.13357008e-01 -1.02509785e+00 2.60445058e-01
8.79982293e-01 -5.88252366e-01 -7.64861763e-01 2.56720752e-01
-6.83099687e-01 -2.17453182e-01 5.20282924e-01 4.11518216e-01
-6.64925039e-01 6.65240228e-01 -5.10773659e-01 -5.32364428e-01
-4.42283988e-01 -1.42350459e+00 9.09710109e-01 3.98795381e-02
-2.61303306e-01 -7.50395715e-01 4.65357333e-01 7.97915101e-01
3.57814273e-03 8.89821872e-02 1.16510916e+00 -8.68243217e-01
-3.90909314e-01 -2.09275428e-02 8.89725015e-02 7.44324476e-02
-6.20035408e-03 4.43517596e-01 -8.15179050e-01 -2.64288187e-01
-5.43919683e-01 -5.37333429e-01 7.54011512e-01 3.46918017e-01
7.20617175e-01 -1.51254274e-02 -5.26971221e-01 4.00324583e-01
1.29861522e+00 4.60288316e-01 7.07758605e-01 4.68253314e-01
2.58753747e-01 4.60326493e-01 7.72396743e-01 1.74903691e-01
1.40399024e-01 7.40795791e-01 1.97662070e-01 -8.63529518e-02
1.94748163e-01 -1.51502788e-01 2.12881014e-01 7.77987182e-01
-7.89142191e-01 7.86238611e-02 -1.37336075e+00 2.78612971e-02
-1.61861086e+00 -1.24930751e+00 -6.85970902e-01 2.18859410e+00
1.04948545e+00 3.85887057e-01 3.73624504e-01 1.98550254e-01
2.59318024e-01 -1.28947467e-01 -8.26674581e-01 -4.77144539e-01
-1.12767689e-01 4.86883670e-01 2.58801311e-01 5.36397099e-01
-9.42797482e-01 1.05653656e+00 7.43701029e+00 6.74259603e-01
-1.04663062e+00 -2.07446501e-01 7.05135107e-01 -2.38691717e-02
-1.39308795e-01 4.31258231e-01 -9.90632057e-01 4.87570137e-01
1.32523561e+00 -9.53092873e-02 2.54386514e-01 9.61052835e-01
3.55495393e-01 -2.31016979e-01 -7.73949742e-01 4.94763911e-01
-4.81974363e-01 -1.55744374e+00 -2.14699030e-01 3.94150347e-01
4.83624548e-01 5.08073211e-01 -6.36438653e-02 3.01055133e-01
3.56605351e-01 -1.03507614e+00 2.28744745e-01 3.05041254e-01
5.38310409e-01 -1.05279446e+00 6.85568333e-01 4.81603682e-01
-8.95904720e-01 2.74959356e-01 -6.05920374e-01 -3.42517555e-01
4.24648933e-02 6.76385641e-01 -8.36511612e-01 5.56922615e-01
8.88588488e-01 6.51393712e-01 -5.91621339e-01 7.23233521e-01
-1.75232396e-01 9.46208417e-01 -2.76436061e-01 -4.26775664e-01
3.50405872e-01 -5.09413660e-01 3.66112649e-01 9.22070861e-01
-3.37141931e-01 4.24731642e-01 2.68996805e-01 3.23331058e-01
2.90017188e-01 3.57494980e-01 -3.43440652e-01 -1.82345897e-01
4.07756269e-01 1.11485648e+00 -4.38608378e-01 -2.70469964e-01
-5.56859791e-01 5.46148956e-01 6.40498042e-01 1.36012822e-01
-8.26658845e-01 -4.10197318e-01 1.14422739e+00 7.09882006e-02
-1.97981335e-02 -4.74295974e-01 -2.90001109e-02 -1.16355586e+00
-2.36762956e-01 -1.18907475e+00 4.80939865e-01 -7.53610015e-01
-1.20173872e+00 2.77711391e-01 -3.13986957e-01 -5.82879066e-01
4.86926809e-02 -1.13453078e+00 -3.50333542e-01 8.35058212e-01
-1.06758475e+00 -7.50782311e-01 3.40621293e-01 7.50364959e-02
3.32289100e-01 -3.34476978e-02 1.14399493e+00 9.21358392e-02
-7.41591871e-01 3.39911669e-01 8.05583954e-01 -2.37536222e-01
8.55126560e-01 -1.39252973e+00 3.59281272e-01 3.34372997e-01
-1.06155515e-01 9.30880070e-01 9.58912313e-01 -5.93804657e-01
-1.49703765e+00 -1.88365787e-01 1.00933135e+00 -6.09149814e-01
1.02205420e+00 -4.49381232e-01 -1.24157929e+00 7.78960824e-01
4.14314866e-02 -5.49752176e-01 1.26469243e+00 6.39451385e-01
-3.13791782e-01 2.21781060e-01 -1.02861905e+00 4.03601348e-01
1.19075239e+00 -2.84236848e-01 -7.58285999e-01 6.46019280e-01
4.73787516e-01 -2.15012804e-01 -1.28797126e+00 3.82800341e-01
7.86793351e-01 -1.42563355e+00 1.18653786e+00 -1.06833279e+00
4.77330059e-01 -8.15366954e-02 -2.56060064e-01 -8.67157638e-01
-5.36637366e-01 -5.62811077e-01 1.63211599e-01 7.74859011e-01
7.97794580e-01 -7.66510785e-01 1.14192545e+00 4.94262934e-01
-4.25347298e-01 -1.26980734e+00 -9.88321543e-01 -6.69246256e-01
6.86073065e-01 -1.43637404e-01 4.18115616e-01 8.19904208e-01
3.71249825e-01 5.88897705e-01 -2.53764302e-01 -9.76044983e-02
5.53018749e-01 2.20985129e-01 7.63158619e-01 -1.34751987e+00
-2.89783239e-01 -4.46336240e-01 -4.41099733e-01 -8.96131217e-01
3.34991999e-02 -1.03739703e+00 -2.83846855e-01 -1.45024931e+00
6.23992682e-01 -2.62994140e-01 -2.25239262e-01 6.95165813e-01
-3.72988075e-01 3.05366099e-01 -2.26597324e-01 4.42473799e-01
-1.99152172e-01 4.79991436e-01 1.02157927e+00 1.63120970e-01
8.96219090e-02 -5.69885038e-02 -7.36018717e-01 5.40765643e-01
9.92336214e-01 -3.14018339e-01 -8.71519931e-03 3.10687155e-01
-6.12028986e-02 -1.88516229e-01 -7.60475770e-02 -9.80129004e-01
-2.66660601e-01 -4.25480098e-01 5.52902460e-01 -8.52171361e-01
4.73838478e-01 -3.76805395e-01 2.85306931e-01 8.45683515e-01
-2.30232906e-02 1.58257157e-01 1.57319248e-01 2.29060665e-01
2.33656820e-02 -1.99349150e-01 9.69755769e-01 -9.91038457e-02
-5.70151329e-01 2.56002337e-01 -7.71690011e-01 -2.42244646e-01
1.01263452e+00 -1.42842144e-01 -4.67548400e-01 -2.74975691e-02
-8.99938047e-01 3.00994068e-02 7.61703372e-01 -2.42827879e-03
2.50796616e-01 -6.85037255e-01 -5.47422767e-01 -1.16234280e-01
1.39442459e-01 -6.11003101e-01 7.58877993e-02 9.13751423e-01
-7.60965824e-01 9.06866670e-01 -2.09761277e-01 -5.25455534e-01
-1.57838500e+00 6.51201487e-01 3.66619557e-01 -2.19548956e-01
-6.81545317e-01 7.00575292e-01 -1.64719701e-01 -5.40831804e-01
-2.25422278e-01 5.09099104e-02 -8.88761356e-02 -9.06137899e-02
5.05916297e-01 4.39792693e-01 2.39275113e-01 -6.95009470e-01
-4.77985650e-01 4.43440706e-01 -4.53530282e-01 2.23363101e-01
1.73172510e+00 3.68819118e-01 -1.23082131e-01 3.63169581e-01
1.06357217e+00 2.21324321e-02 -8.75550985e-01 1.75468162e-01
3.48907322e-01 -2.12596923e-01 4.99470439e-03 -1.11035836e+00
-2.17413470e-01 7.76800931e-01 3.25885653e-01 4.25741151e-02
7.31682956e-01 -2.50218529e-02 8.19353044e-01 8.86925876e-01
5.33529341e-01 -7.45952904e-01 -4.12785351e-01 6.93478882e-01
5.18056870e-01 -1.33998513e+00 4.38904315e-01 -3.96713108e-01
-4.05628562e-01 1.04112637e+00 4.71519291e-01 2.99341269e-02
4.48538572e-01 1.11815773e-01 6.27959371e-02 -2.78072834e-01
-1.09878647e+00 -1.83651298e-01 5.02925515e-02 6.32974386e-01
1.13609290e+00 7.20441341e-02 -8.77719402e-01 2.09245607e-01
-3.76529336e-01 -2.60974735e-01 2.55404979e-01 9.86818433e-01
-1.02481580e+00 -1.54826069e+00 -2.17141122e-01 4.24458921e-01
-5.84319830e-01 -2.33776510e-01 -8.19981456e-01 1.01743770e+00
-3.10602278e-01 7.17099786e-01 -4.13751215e-01 -1.59904480e-01
1.11965530e-01 5.60204804e-01 5.67271173e-01 -5.51028192e-01
-5.76384187e-01 -8.17783400e-02 4.42220509e-01 -4.12949860e-01
-4.16657090e-01 -9.47460353e-01 -1.60365939e+00 -8.00640106e-01
-4.84115869e-01 6.68354154e-01 6.42909110e-01 7.79603004e-01
3.20291489e-01 -6.34073466e-02 4.55191225e-01 -8.31760645e-01
-5.15746534e-01 -9.23583269e-01 -4.22309786e-01 8.11997801e-03
-3.06257531e-02 -7.77014792e-01 -3.20940077e-01 -9.42905620e-02] | [4.774402618408203, 5.513051509857178] |
a42e4157-1b6c-4e40-8c1c-3ee52d822a24 | flexible-sampling-for-long-tailed-skin-lesion | 2204.03161 | null | https://arxiv.org/abs/2204.03161v1 | https://arxiv.org/pdf/2204.03161v1.pdf | Flexible Sampling for Long-tailed Skin Lesion Classification | Most of the medical tasks naturally exhibit a long-tailed distribution due to the complex patient-level conditions and the existence of rare diseases. Existing long-tailed learning methods usually treat each class equally to re-balance the long-tailed distribution. However, considering that some challenging classes may present diverse intra-class distributions, re-balancing all classes equally may lead to a significant performance drop. To address this, in this paper, we propose a curriculum learning-based framework called Flexible Sampling for the long-tailed skin lesion classification task. Specifically, we initially sample a subset of training data as anchor points based on the individual class prototypes. Then, these anchor points are used to pre-train an inference model to evaluate the per-class learning difficulty. Finally, we use a curriculum sampling module to dynamically query new samples from the rest training samples with the learning difficulty-aware sampling probability. We evaluated our model against several state-of-the-art methods on the ISIC dataset. The results with two long-tailed settings have demonstrated the superiority of our proposed training strategy, which achieves a new benchmark for long-tailed skin lesion classification. | ['ZongYuan Ge', 'Paul Bonnington', 'Xin Wang', 'Xin Zhao', 'Zhen Yu', 'Lin Wang', 'Yicheng Wu', 'Lie Ju'] | 2022-04-07 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 2.46685103e-01 -2.47215822e-01 -7.03255236e-01 -6.73553586e-01
-1.03806376e+00 -1.07784666e-01 3.07869107e-01 2.70923495e-01
-4.82617706e-01 8.86593044e-01 2.73812674e-02 -2.73076236e-01
-4.95294243e-01 -6.67011738e-01 -5.05776227e-01 -1.07125556e+00
1.14691094e-01 7.60345995e-01 5.16063690e-01 3.37941498e-01
2.04062805e-01 6.07801788e-02 -1.48569572e+00 3.73616546e-01
1.38141215e+00 9.37346876e-01 3.18711430e-01 5.46246350e-01
-1.88099667e-01 2.28901401e-01 -5.93957365e-01 -2.55615801e-01
-1.49758279e-01 -2.24277526e-01 -4.51570213e-01 -5.55596612e-02
4.98478800e-01 -2.17920214e-01 2.84551643e-02 1.08400810e+00
8.12977850e-01 -1.80369383e-03 1.09884465e+00 -1.20161343e+00
-1.64159074e-01 3.32215190e-01 -7.26518691e-01 1.32500395e-01
9.52297263e-03 1.28766865e-01 7.84828961e-01 -5.20255864e-01
4.09036577e-01 1.06570756e+00 5.57211697e-01 7.26029873e-01
-1.13118589e+00 -8.54663491e-01 4.15326953e-01 3.06551456e-01
-1.31340337e+00 1.29936442e-01 5.74083328e-01 -4.37000751e-01
1.27232298e-01 2.98353076e-01 4.41890836e-01 1.27453125e+00
3.89006019e-01 1.11092401e+00 1.26548946e+00 -1.77907228e-01
4.26452279e-01 3.24414313e-01 2.45201960e-01 4.48781550e-01
2.12093040e-01 1.05677500e-01 -2.12658003e-01 -5.84967792e-01
4.39752251e-01 3.00834626e-01 -2.41634622e-01 -5.15594959e-01
-8.43099058e-01 8.70030165e-01 3.93972516e-01 -1.85516879e-01
-2.75895536e-01 -1.72080249e-01 6.95582867e-01 -5.95802255e-02
4.31743741e-01 -1.40370980e-01 -7.07459033e-01 -6.87567368e-02
-8.21987748e-01 4.21201050e-01 7.57374585e-01 7.46328175e-01
3.49998385e-01 -6.20593548e-01 -9.59567070e-01 1.03189373e+00
1.12890385e-01 2.84521997e-01 5.66207051e-01 -1.51762545e-01
4.14754391e-01 3.72983158e-01 -9.10666585e-02 -5.37515223e-01
-3.36058199e-01 -9.10396993e-01 -1.06397688e+00 -1.65128514e-01
4.70154017e-01 -2.29610950e-01 -1.16425419e+00 1.79088140e+00
7.19357014e-01 2.84489870e-01 -3.21982980e-01 7.45689332e-01
4.41112578e-01 4.50577945e-01 5.58094978e-01 -1.45132154e-01
1.50032103e+00 -9.06710505e-01 -5.88941991e-01 9.61692482e-02
5.89347720e-01 -5.56919575e-01 1.55058563e+00 6.10291481e-01
-5.48568130e-01 -3.72756481e-01 -7.00392008e-01 4.17807579e-01
-5.16338348e-02 7.66042396e-02 3.23166847e-01 6.50604546e-01
-3.08048338e-01 5.34150243e-01 -7.33276188e-01 -1.41518354e-01
8.41589510e-01 -7.70808384e-02 2.41505340e-01 -5.15663266e-01
-1.04964507e+00 3.75600994e-01 4.99834448e-01 -2.84865767e-01
-8.81569505e-01 -1.24888551e+00 -4.25736457e-01 9.37460288e-02
5.03549695e-01 -8.00736964e-01 1.20848548e+00 -7.51601338e-01
-1.27692819e+00 6.75983012e-01 5.61386906e-02 -1.20314844e-01
8.14153969e-01 -7.13201538e-02 -2.82135814e-01 -2.51455847e-02
9.56870243e-02 4.19843256e-01 8.16376507e-01 -1.02247155e+00
-7.77849615e-01 -3.08134764e-01 -2.18758434e-01 2.03239277e-01
-6.70478761e-01 -3.64995033e-01 -5.17400622e-01 -7.99297631e-01
-3.73388290e-01 -8.60508204e-01 -3.49927753e-01 2.22782806e-01
-5.39140284e-01 -5.15587032e-01 4.38959122e-01 -1.49312094e-01
1.61956263e+00 -2.12098670e+00 -7.80962631e-02 2.63408571e-01
6.05935752e-02 3.18073958e-01 -2.04870105e-01 1.86347365e-01
-5.90242036e-02 -1.06329270e-01 -2.29809761e-01 -1.39600471e-01
-1.62337795e-02 1.35949224e-01 6.44271895e-02 3.11100304e-01
1.53131932e-01 4.03423339e-01 -1.05658901e+00 -8.73628020e-01
7.09769726e-02 2.88609982e-01 -7.83964634e-01 4.07298177e-01
-5.26194453e-01 3.87569666e-01 -6.85719550e-01 7.01979756e-01
8.28146935e-01 -4.71733570e-01 -2.04404593e-02 -1.48151547e-01
4.08978552e-01 -1.76236823e-01 -9.86759722e-01 1.44717860e+00
-4.26100075e-01 -9.64509174e-02 -2.18971446e-01 -9.61629152e-01
6.86243773e-01 1.09245768e-02 4.42470312e-01 -3.94956440e-01
1.08857796e-01 2.60634303e-01 1.49306923e-01 -6.50312126e-01
-1.04627781e-01 -4.03221726e-01 7.44213015e-02 9.54751819e-02
-2.46677816e-01 -6.74159303e-02 7.88850859e-02 -6.89913556e-02
9.07105684e-01 -1.37216613e-01 2.71193385e-01 -2.91854978e-01
4.79401946e-01 -2.99360722e-01 7.54194558e-01 7.13346422e-01
-4.16912645e-01 5.67134500e-01 7.97238469e-01 -3.83752704e-01
-6.54411674e-01 -1.28174400e+00 -7.41039038e-01 1.24281502e+00
-3.17453295e-02 -2.45006174e-01 -7.12893724e-01 -1.12006223e+00
1.32115856e-01 7.25632310e-01 -8.26486588e-01 -2.95081735e-01
-7.47340266e-04 -1.17623615e+00 2.09907368e-01 4.73112434e-01
2.13170886e-01 -8.19861114e-01 -5.13862908e-01 -1.96370259e-02
-1.17588192e-01 -7.08811343e-01 -8.03860903e-01 2.63036579e-01
-7.82845020e-01 -1.24672318e+00 -1.23223150e+00 -7.45321691e-01
8.97627056e-01 -1.13964766e-01 1.08430922e+00 -7.53444899e-03
-6.69572532e-01 -6.98821843e-02 -3.21222514e-01 -5.35507202e-01
-5.68418130e-02 3.48421812e-01 -3.34379315e-01 3.01977787e-02
4.17954117e-01 -2.60116696e-01 -1.05007398e+00 4.28038716e-01
-1.03310084e+00 -1.46666929e-01 7.80813634e-01 1.21776617e+00
9.75501657e-01 2.39666909e-01 7.86904573e-01 -1.31476974e+00
6.50375068e-01 -7.71490395e-01 -2.81513482e-01 4.25897747e-01
-6.50501132e-01 -9.85222533e-02 6.47902668e-01 -8.90736997e-01
-9.07532036e-01 -7.25276023e-03 -1.41221210e-01 -3.16680878e-01
-1.96817622e-01 6.52834475e-01 -1.96080506e-01 4.09685105e-01
6.42427504e-01 1.61157638e-01 -6.61337301e-02 -3.07021677e-01
1.64799998e-03 8.74408603e-01 2.23601043e-01 -8.97611678e-01
5.80357850e-01 1.53446510e-01 -3.65062691e-02 -5.02626657e-01
-1.28283823e+00 -5.53955138e-01 -2.10914358e-01 5.01980484e-02
6.02457702e-01 -8.49955976e-01 -5.48603475e-01 7.62684643e-01
-5.37409902e-01 -5.61725616e-01 -2.82473326e-01 3.92172188e-01
-4.99108881e-01 3.06043446e-01 -3.95068944e-01 -6.36992097e-01
-5.29836297e-01 -1.24437988e+00 1.26771629e+00 5.08430719e-01
-1.15795322e-01 -1.16510975e+00 1.28341213e-01 2.47371808e-01
4.22791511e-01 1.73351467e-01 1.36113942e+00 -8.00652564e-01
-9.96427089e-02 -3.19459379e-01 -2.52815932e-01 2.86588311e-01
3.48113537e-01 -9.80227366e-02 -7.30030477e-01 -6.10969663e-01
-3.94385129e-01 -6.11948013e-01 8.99057209e-01 5.78720570e-01
2.12427092e+00 8.21535513e-02 -5.38926899e-01 7.59444058e-01
1.24771130e+00 -2.23599404e-01 3.64052147e-01 1.47630095e-01
2.94075817e-01 4.95393306e-01 1.04070330e+00 7.35420287e-01
2.81677753e-01 3.43807310e-01 4.68814373e-01 -1.28943682e-01
1.04468748e-01 -4.22889054e-01 -2.20354676e-01 4.63877439e-01
5.71977198e-01 -2.46865124e-01 -9.00808930e-01 4.80858475e-01
-1.56764746e+00 -6.08157218e-01 2.18864337e-01 2.34017301e+00
1.44768095e+00 3.86155158e-01 3.10207814e-01 -8.08058530e-02
6.99555337e-01 -5.27036190e-02 -1.04943717e+00 1.00926936e-01
3.38466883e-01 2.35520020e-01 4.08508807e-01 2.03353271e-01
-1.06656802e+00 6.15452111e-01 5.99265766e+00 1.46121597e+00
-1.18268359e+00 -2.21730158e-01 1.01279652e+00 -1.67291671e-01
-3.18013608e-01 -4.17727411e-01 -1.03374290e+00 8.43839467e-01
7.28031099e-01 -1.20799273e-01 -1.22258611e-01 8.16533625e-01
-3.19974776e-03 -1.52825534e-01 -1.16052151e+00 9.33941841e-01
3.10561024e-02 -9.72941935e-01 1.89992428e-01 -5.77542484e-02
7.97495842e-01 -1.80287793e-01 3.47378105e-01 5.66481769e-01
1.86004058e-01 -9.15174961e-01 1.58269733e-01 5.54450214e-01
9.88416553e-01 -8.34741712e-01 8.42528522e-01 6.59908712e-01
-7.28377044e-01 -1.40333772e-01 -3.84660602e-01 4.58576649e-01
-2.25518242e-01 1.07236469e+00 -9.05854702e-01 3.46926421e-01
6.06244266e-01 3.76381069e-01 -6.37497783e-01 1.56405365e+00
7.12101460e-02 8.84493709e-01 -2.14944378e-01 -4.04561460e-01
4.78250459e-02 9.26139206e-02 1.33521691e-01 1.08349514e+00
2.41684198e-01 -1.11574613e-01 4.94479001e-01 4.90576088e-01
1.87576085e-01 2.01188073e-01 1.77279770e-01 1.92864060e-01
5.87454379e-01 1.22727799e+00 -5.87527037e-01 -3.67621899e-01
-2.74086297e-01 6.70148611e-01 1.95245832e-01 3.08803737e-01
-1.00994122e+00 -4.83945101e-01 4.09653425e-01 2.08067358e-01
3.25020403e-01 3.41499031e-01 -2.34627016e-02 -1.01188385e+00
-1.00070074e-01 -9.68528688e-01 8.93873870e-01 -3.18855435e-01
-1.88120782e+00 3.78487706e-01 1.15994900e-01 -1.25940847e+00
6.92930371e-02 -6.13547146e-01 -7.61828065e-01 7.60666072e-01
-1.88712764e+00 -9.17442799e-01 -5.24927914e-01 7.16949463e-01
5.69380581e-01 -5.23280585e-03 8.29528272e-01 4.28678215e-01
-6.41887009e-01 1.22317207e+00 3.36811930e-01 -4.79203798e-02
1.25443351e+00 -1.41008484e+00 -2.73260176e-01 -8.91212467e-03
-5.54485619e-01 4.00387704e-01 5.13064384e-01 -6.40127242e-01
-1.00498164e+00 -1.29869163e+00 3.15919846e-01 -1.42712116e-01
5.33311248e-01 -2.11731151e-01 -1.06579030e+00 2.41140932e-01
-2.32079789e-01 2.51778454e-01 1.15480077e+00 3.19724351e-01
-3.99958432e-01 -3.92258674e-01 -1.32582259e+00 5.72849274e-01
6.69198275e-01 -8.92878026e-02 -3.47969383e-01 5.73545933e-01
5.92467368e-01 -4.73832756e-01 -8.27686727e-01 5.63155830e-01
5.49582541e-01 -5.31659782e-01 8.05179417e-01 -7.81031787e-01
5.32436371e-01 4.56053875e-02 -1.76482439e-01 -1.59203398e+00
-3.25635880e-01 -1.98027954e-01 -3.08195036e-02 1.00692773e+00
1.49178997e-01 -5.89446545e-01 1.18087757e+00 4.68977779e-01
7.31760636e-02 -1.37081063e+00 -8.35073411e-01 -6.26450300e-01
4.87251610e-01 -3.64565253e-02 5.53564131e-01 6.65226996e-01
-2.57562965e-01 1.54997453e-01 -1.38359234e-01 1.06307089e-01
8.81083012e-01 1.35764048e-01 7.24068582e-01 -1.28187132e+00
-4.91422981e-01 -3.38558823e-01 -1.29352272e-01 -1.03788114e+00
-9.10043716e-03 -8.26732814e-01 2.19987527e-01 -1.18679845e+00
7.65251935e-01 -9.00997579e-01 -6.31975651e-01 3.08074802e-01
-8.50905776e-01 -2.74590909e-01 -3.21322769e-01 -5.32543026e-02
-7.31185615e-01 6.49286985e-01 1.62284386e+00 -5.69347963e-02
-6.16915748e-02 4.85372543e-01 -7.55117655e-01 5.69612145e-01
7.10746467e-01 -4.82583553e-01 -7.86703289e-01 -1.17067985e-01
-1.00070499e-01 5.24967909e-02 1.53245613e-01 -9.71237302e-01
2.50129759e-01 -5.61229408e-01 4.82193440e-01 -9.43199635e-01
-2.88556647e-02 -7.81677485e-01 -3.84843737e-01 8.27241182e-01
-5.20208955e-01 -5.82468808e-01 1.33817509e-01 9.07150567e-01
-9.39848647e-03 -2.18718812e-01 1.11757600e+00 1.99306622e-01
-1.14091128e-01 7.65689373e-01 -1.08427316e-01 4.36849743e-01
1.27059877e+00 6.96314871e-02 -2.91169971e-01 2.26433687e-02
-5.84261894e-01 7.23671913e-01 1.68755278e-01 2.52662301e-01
4.72969353e-01 -1.24106789e+00 -8.74450266e-01 2.64102399e-01
4.75308090e-01 4.82416391e-01 6.91529572e-01 8.97876740e-01
-2.06974074e-01 -9.48918052e-03 -9.80961621e-02 -9.59351778e-01
-1.31325066e+00 6.72646582e-01 3.67069006e-01 -5.01806021e-01
-3.21960121e-01 9.25125599e-01 3.69583726e-01 -7.77186990e-01
6.62501574e-01 -3.05071384e-01 5.47521152e-02 5.02662323e-02
6.14851594e-01 2.23610416e-01 -1.43011987e-01 2.42466509e-01
-3.60603094e-01 6.25451863e-01 -5.82270086e-01 3.86653274e-01
1.22312737e+00 2.24748343e-01 2.26250082e-01 3.63419503e-01
1.20310998e+00 -2.66338170e-01 -1.31638527e+00 -4.56470907e-01
-1.48597047e-01 -5.57158113e-01 -7.67127424e-02 -1.17360520e+00
-9.78027225e-01 9.45692241e-01 8.27061296e-01 -2.39325330e-01
1.34638524e+00 -4.30632792e-02 8.25001955e-01 1.17864981e-01
1.55475333e-01 -1.12294638e+00 2.94298798e-01 2.20958009e-01
5.17709136e-01 -1.20725012e+00 5.61146438e-02 -5.80013692e-01
-5.42818367e-01 8.10525894e-01 8.31812620e-01 -1.38417967e-02
1.03658199e+00 2.92971492e-01 6.11433163e-02 1.53786629e-01
-9.93180573e-01 1.47996753e-01 3.31395924e-01 6.66998684e-01
3.76168370e-01 2.90629297e-01 -4.40737188e-01 7.06607521e-01
3.31426673e-02 2.57413983e-01 7.27068484e-02 7.15455413e-01
-4.28243458e-01 -1.22558641e+00 -2.41846055e-01 8.94388735e-01
-4.38971102e-01 5.67118786e-02 1.18604727e-01 5.85020959e-01
1.29485488e-01 4.32716310e-01 -3.21843885e-02 -1.20503508e-01
4.44288552e-01 -4.60771695e-02 4.63048667e-01 -7.00534105e-01
-3.98734480e-01 1.76317737e-01 -2.81124145e-01 -3.36740166e-01
1.35273010e-01 -5.91487169e-01 -1.09410858e+00 5.15237637e-02
-3.65960479e-01 1.33093357e-01 3.99674833e-01 5.40594578e-01
1.55810520e-01 8.49612474e-01 7.24497676e-01 -2.65579671e-01
-1.49197686e+00 -8.36847067e-01 -5.94295800e-01 5.60286641e-01
2.61110425e-01 -7.11777985e-01 -4.48500544e-01 -3.07871938e-01] | [15.290389060974121, -2.608619213104248] |
d6f396c2-8fbc-40bd-8de9-872b26a8713c | interaction-compass-multi-label-zero-shot | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Huynh_Interaction_Compass_Multi-Label_Zero-Shot_Learning_of_Human-Object_Interactions_via_Spatial_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Huynh_Interaction_Compass_Multi-Label_Zero-Shot_Learning_of_Human-Object_Interactions_via_Spatial_ICCV_2021_paper.pdf | Interaction Compass: Multi-Label Zero-Shot Learning of Human-Object Interactions via Spatial Relations | We study the problem of multi-label zero-shot recognition in which labels are in the form of human-object interactions (combinations of actions on objects), each image may contain multiple interactions and some interactions do not have training images. We propose a novel compositional learning framework that decouples interaction labels into separate action and object scores that incorporate the spatial compatibility between the two components. We combine these scores to efficiently recognize seen and unseen interactions. However, learning action-object spatial relations, in principle, requires bounding-box annotations, which are costly to gather. Moreover, it is not clear how to generalize spatial relations to unseen interactions. We address these challenges by developing a cross-attention mechanism that localizes objects from action locations and vice versa by predicting displacements between them, referred to as relational directions. During training, we estimate the relational directions as ones maximizing the scores of ground-truth interactions that guide predictions toward compatible action-object regions. By extensive experiments, we show the effectiveness of our framework, where we improve the state of the art by 2.6% mAP score and 5.8% recall score on HICO and Visual Genome datasets, respectively. Code is available at https://github.com/hbdat/iccv21_relational_direction. | ['Ehsan Elhamifar', 'Dat Huynh'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['multi-label-zero-shot-learning'] | ['computer-vision'] | [ 3.80890667e-01 3.97644863e-02 -2.46127352e-01 -4.63250577e-01
-8.04119766e-01 -6.32339001e-01 6.19589925e-01 -5.14207259e-02
-1.93220630e-01 5.93193948e-01 2.37198219e-01 1.34062409e-01
-1.36875078e-01 -5.95782697e-01 -1.11086023e+00 -9.10414457e-01
1.97310112e-02 5.65916419e-01 5.31141102e-01 1.87218353e-01
1.02018692e-01 2.07126200e-01 -1.55941021e+00 8.17262173e-01
4.87675697e-01 9.48188245e-01 3.84065032e-01 7.58239985e-01
2.25799844e-01 9.45639610e-01 -3.82238716e-01 -1.97047576e-01
1.95104390e-01 -5.58168709e-01 -9.15806592e-01 2.46859804e-01
4.64996755e-01 -3.53994757e-01 -1.66376799e-01 9.38567340e-01
2.24357456e-01 2.36903876e-01 7.90643871e-01 -1.39471877e+00
-9.60462272e-01 2.32641697e-01 -7.61641443e-01 7.19020516e-02
1.09565780e-01 1.79649293e-01 1.26278591e+00 -9.81885552e-01
8.11105728e-01 1.08235407e+00 2.66923994e-01 6.09359384e-01
-1.22132051e+00 -3.53756309e-01 2.96117157e-01 3.59674066e-01
-1.43563104e+00 -4.44776267e-01 4.38347816e-01 -8.12312365e-01
1.09157968e+00 2.85919338e-01 3.16361874e-01 9.41643834e-01
3.06055844e-02 8.92545760e-01 7.37166703e-01 -3.97791356e-01
2.14154527e-01 -1.51090980e-01 1.16041936e-01 7.74392724e-01
-1.22133426e-01 -6.51996732e-02 -5.71519911e-01 -2.67610718e-02
7.23390996e-01 2.35993207e-01 -1.80871129e-01 -6.59597337e-01
-1.40306115e+00 5.98200142e-01 6.43472672e-01 2.10660189e-01
-2.23025545e-01 3.99908155e-01 1.53971359e-01 -2.74347931e-01
3.46476406e-01 2.01978251e-01 -4.98771131e-01 9.11981910e-02
-3.71631116e-01 6.06844723e-02 1.74757391e-01 1.09670711e+00
7.09337533e-01 -7.08232462e-01 -2.83305079e-01 9.36351895e-01
1.80997089e-01 1.82917923e-01 1.47382736e-01 -9.97560561e-01
3.63364041e-01 6.63605452e-01 2.10618675e-01 -9.78027225e-01
-3.19099486e-01 3.56929637e-02 -7.27898359e-01 -3.42856348e-03
3.67898613e-01 1.21293582e-01 -8.85327876e-01 1.95863414e+00
5.43192029e-01 4.68453974e-01 -1.40901461e-01 9.67429578e-01
6.58911467e-01 6.87145829e-01 1.64848447e-01 2.20119003e-02
1.38045299e+00 -1.39163685e+00 -6.26855016e-01 -3.40475470e-01
9.86237288e-01 -3.27361643e-01 9.81309950e-01 7.19168335e-02
-1.03779411e+00 -4.99457955e-01 -8.66293788e-01 -2.17136085e-01
-5.41077733e-01 3.32017571e-01 5.41220069e-01 -9.17380527e-02
-8.21314514e-01 3.98053139e-01 -9.93825257e-01 -4.18748081e-01
5.31510234e-01 2.93043017e-01 -4.68087643e-01 6.36868402e-02
-9.32173014e-01 8.06962669e-01 2.97813892e-01 -2.71684751e-02
-1.06677914e+00 -5.82827508e-01 -8.60227287e-01 2.45879814e-02
7.66779661e-01 -3.05804849e-01 1.09977293e+00 -7.63491333e-01
-9.34359848e-01 9.70739603e-01 -3.10666054e-01 -7.50705376e-02
3.19480270e-01 -2.84921527e-01 -1.23077348e-01 4.35228199e-02
3.00393492e-01 1.01574171e+00 1.80358425e-01 -1.24503815e+00
-7.42489755e-01 -4.01694149e-01 2.21929103e-01 3.52573574e-01
5.83860138e-03 1.70930922e-01 -7.73901582e-01 -3.94931018e-01
1.42528430e-01 -1.09027243e+00 -6.87175319e-02 4.03655201e-01
-5.54434896e-01 -3.65771085e-01 7.25644648e-01 -4.74799931e-01
7.85474241e-01 -2.24079442e+00 4.69512343e-01 -2.18076974e-01
9.39298049e-02 4.37996117e-03 -2.40041956e-01 2.71728277e-01
-5.04748896e-03 1.24576889e-01 -3.07429314e-01 -1.14608422e-01
-6.69991598e-02 2.40370840e-01 -1.57662898e-01 5.89840829e-01
3.48296404e-01 1.18844080e+00 -9.30919230e-01 -3.86726469e-01
1.93454280e-01 4.32999849e-01 -4.89287168e-01 3.02166969e-01
-5.11244655e-01 6.16325140e-01 -3.49051327e-01 7.26484895e-01
4.56160545e-01 -7.71254361e-01 3.99473339e-01 -1.97978914e-01
-1.15594743e-02 2.79387832e-02 -9.85750973e-01 1.78978622e+00
-1.21971428e-01 4.19037461e-01 -2.18820333e-01 -1.01921248e+00
4.73323941e-01 2.09433660e-01 4.53337461e-01 -4.84839439e-01
-2.97958474e-03 -1.64982140e-01 -7.47990087e-02 -5.87310910e-01
1.89101454e-02 4.32366133e-02 -1.82533383e-01 4.72083151e-01
1.91313997e-01 3.89455169e-01 9.57223475e-02 1.76786393e-01
1.13641918e+00 4.45141166e-01 5.18651009e-01 -1.84368212e-02
3.10839415e-01 -1.18063465e-01 7.59423077e-01 6.11606658e-01
-3.00971657e-01 8.00806820e-01 7.10455060e-01 -3.63015652e-01
-8.71798813e-01 -1.10751581e+00 -4.22840565e-02 1.54701233e+00
5.00969946e-01 -2.93878883e-01 -4.76828367e-01 -9.65604365e-01
-1.45862654e-01 5.31031311e-01 -8.59266520e-01 1.52019347e-04
-5.77424645e-01 -6.08782768e-01 2.91298896e-01 8.51776898e-01
8.46575350e-02 -1.11207175e+00 -6.42923415e-01 -2.09763795e-01
-1.97626188e-01 -1.13839519e+00 -6.76848710e-01 2.07233623e-01
-2.65757889e-01 -1.28515172e+00 -7.15123117e-01 -7.67589629e-01
9.50808108e-01 4.21071410e-01 9.52203751e-01 2.24422038e-01
-4.74147439e-01 1.53590038e-01 -3.65296394e-01 -1.50439873e-01
9.79385823e-02 -2.31613129e-01 7.96224514e-04 1.55434564e-01
2.57008195e-01 -3.05900484e-01 -5.83629668e-01 7.24328518e-01
-7.79638708e-01 5.60957789e-01 5.34010649e-01 1.05525327e+00
9.04887736e-01 -3.91219854e-01 3.90305728e-01 -9.71399307e-01
-1.96909368e-01 -6.02877200e-01 -6.82744145e-01 8.15650344e-01
4.18604165e-02 -8.67366865e-02 2.32184991e-01 -4.19249266e-01
-1.13015258e+00 5.51340580e-01 2.07868338e-01 -5.08105934e-01
-4.27495509e-01 2.75993139e-01 -3.74380141e-01 1.78780690e-01
4.87041533e-01 8.81664455e-02 -3.92825902e-01 -2.62023747e-01
6.49252653e-01 5.41553438e-01 4.65501398e-01 -5.32306612e-01
3.48418891e-01 6.57360315e-01 -6.58041686e-02 -6.00212514e-01
-1.19196820e+00 -6.95879877e-01 -1.06972253e+00 -2.25544423e-01
1.33745229e+00 -7.97367990e-01 -8.02773654e-01 3.17213714e-01
-1.21380532e+00 -5.83167374e-01 -3.57072093e-02 2.95567632e-01
-8.19834113e-01 1.28021449e-01 -6.48900807e-01 -6.51333749e-01
1.18644424e-01 -1.19814897e+00 1.46016252e+00 7.32251108e-02
-2.58106202e-01 -7.48616755e-01 1.27874851e-01 5.17956555e-01
-1.06816806e-01 1.53002888e-01 9.46580887e-01 -6.22206450e-01
-9.61185515e-01 4.48940620e-02 -3.64837348e-01 -7.92762637e-02
1.09329991e-01 -1.47945732e-01 -9.77071166e-01 -1.94953885e-02
-4.34607476e-01 -7.14569092e-01 8.79232287e-01 4.51565951e-01
1.39991343e+00 -1.60746634e-01 -6.77156985e-01 4.39692348e-01
1.21731567e+00 3.85502934e-01 5.80949605e-01 -1.69235080e-01
9.15765524e-01 8.70057166e-01 9.28298175e-01 3.88094515e-01
3.24164689e-01 1.14145827e+00 3.11671495e-01 -7.49280229e-02
-1.53911188e-01 -2.11423665e-01 1.53971627e-01 3.88331503e-01
2.93790530e-02 -5.39132953e-01 -1.02633667e+00 6.07079268e-01
-2.32869172e+00 -9.27865863e-01 2.30207993e-03 1.99108648e+00
6.97193682e-01 -1.87969759e-01 8.21730569e-02 -4.18775767e-01
8.10879230e-01 1.00242294e-01 -7.71271467e-01 2.29787558e-01
-5.19340821e-02 -2.90399462e-01 3.33738625e-01 5.79329371e-01
-1.34419203e+00 9.88026559e-01 5.48707724e+00 6.61177099e-01
-6.47669375e-01 3.57567310e-01 8.32201183e-01 -4.15147275e-01
-2.77680829e-02 -3.47275822e-03 -8.31349790e-01 3.35252732e-01
5.58980823e-01 2.98229873e-01 3.60351801e-01 7.52039731e-01
-1.22124769e-01 -1.61045492e-01 -1.42913163e+00 7.64322579e-01
1.98259443e-01 -1.39819598e+00 -1.96283430e-01 8.63404199e-02
7.32614338e-01 -8.43992271e-03 -8.41706768e-02 1.64856598e-01
4.44568992e-01 -9.41159666e-01 7.26637065e-01 6.42036557e-01
6.86658800e-01 -3.24637920e-01 4.53309566e-01 4.73692328e-01
-1.26994634e+00 3.50394547e-02 -2.74644732e-01 1.01670269e-02
3.03390563e-01 2.36793324e-01 -5.66162586e-01 3.88052315e-01
6.12866580e-01 8.30941856e-01 -4.29134130e-01 7.77142823e-01
-3.28327715e-01 1.08297303e-01 -5.08328527e-02 1.03116810e-01
2.04054788e-02 -1.50371492e-01 1.41849518e-01 9.63756263e-01
1.68236792e-01 4.63746756e-01 2.74620295e-01 9.57895756e-01
-7.95424283e-02 -1.22636840e-01 -6.66533291e-01 -8.76453519e-03
3.41830462e-01 1.19299769e+00 -9.27482367e-01 -3.78259510e-01
-7.27638125e-01 1.15385747e+00 7.31395066e-01 3.18290710e-01
-1.23698425e+00 -3.55990343e-02 7.04420626e-01 -5.15410118e-02
3.77364039e-01 5.86169399e-02 -1.53700680e-01 -1.20059371e+00
8.12445357e-02 -4.83757138e-01 5.06080866e-01 -9.41746235e-01
-1.31813645e+00 2.82755643e-01 9.50277224e-02 -1.20261109e+00
-4.65394445e-02 -6.22862458e-01 -4.13015455e-01 5.11341870e-01
-8.32866669e-01 -1.37279880e+00 -2.06774041e-01 6.43854141e-02
6.59472466e-01 1.07302129e-01 8.08083236e-01 2.62181610e-01
-7.25929856e-01 4.32364851e-01 -6.03316501e-02 8.66675563e-03
6.07484937e-01 -1.09127748e+00 3.13808262e-01 6.67110860e-01
3.94820631e-01 3.72185320e-01 2.15566784e-01 -7.82405913e-01
-1.06170225e+00 -1.23999655e+00 9.76423204e-01 -6.85629845e-01
5.16718507e-01 -7.11155355e-01 -9.54723060e-01 1.12428093e+00
1.20683592e-02 4.71160442e-01 8.34415734e-01 2.49425218e-01
-3.79146457e-01 1.64318368e-01 -8.48872006e-01 6.58011675e-01
1.54212415e+00 -5.72245240e-01 -3.03983063e-01 6.22094452e-01
6.51114106e-01 -3.68039310e-01 -6.41754150e-01 4.23620671e-01
5.42149961e-01 -8.69593382e-01 1.07515121e+00 -8.41567397e-01
5.10711908e-01 -5.71278036e-01 -3.92373115e-01 -9.75300550e-01
-6.06790960e-01 8.94028768e-02 -2.08469689e-01 1.15385938e+00
5.43209851e-01 -3.17031741e-01 7.03880608e-01 6.78751469e-01
-1.14977427e-01 -1.07289338e+00 -8.04190695e-01 -7.83279657e-01
-1.70778066e-01 -1.84974000e-01 3.29842210e-01 1.01429462e+00
2.67395735e-01 5.02323806e-01 -6.10560715e-01 3.33319694e-01
4.38673347e-01 1.96600989e-01 5.78007221e-01 -8.25362444e-01
-6.87291801e-01 -2.70124912e-01 -4.05792594e-01 -1.22369742e+00
2.76321828e-01 -9.23114121e-01 4.17861044e-01 -1.57641292e+00
7.93671489e-01 -3.83078039e-01 -4.27867711e-01 8.71115506e-01
-1.29226804e-01 3.63960773e-01 1.80329531e-01 3.02390784e-01
-1.20959163e+00 5.45620859e-01 1.05470765e+00 -2.40774408e-01
6.85337977e-03 -3.25182050e-01 -5.03919303e-01 8.18581641e-01
6.60719395e-01 -4.21753705e-01 -5.34017861e-01 -6.26674473e-01
1.93617155e-03 1.68505684e-01 5.06197512e-01 -8.92148614e-01
3.20691794e-01 -4.71521407e-01 4.64168757e-01 -6.50899351e-01
6.25700772e-01 -5.86887598e-01 3.99882257e-01 2.65726358e-01
-7.66411424e-01 -3.16832513e-01 -9.06780362e-02 7.02810943e-01
5.99177442e-02 -7.76682049e-02 7.90380955e-01 -1.42594799e-01
-7.49518275e-01 3.48343730e-01 2.72088917e-03 -1.63664296e-01
1.64527786e+00 1.84389025e-01 -5.54691434e-01 -1.30686164e-01
-8.56530666e-01 4.02190685e-01 4.70967501e-01 4.64153945e-01
5.58312774e-01 -1.40351725e+00 -4.04758394e-01 3.38292196e-02
4.58805859e-01 1.90064069e-02 5.84240735e-01 9.10290539e-01
-1.83689296e-01 5.62249780e-01 -1.30648807e-01 -7.10106790e-01
-1.50230122e+00 8.41951668e-01 3.18037689e-01 -1.69507209e-02
-4.22434002e-01 1.14886987e+00 1.07220674e+00 -4.97275591e-01
3.53147775e-01 -1.05199434e-01 -6.60339817e-02 -3.77436578e-02
6.95009291e-01 2.95807332e-01 -3.05223495e-01 -8.87889922e-01
-5.79207838e-01 6.70046508e-01 -1.59051687e-01 -1.22470900e-01
1.25342095e+00 -1.97837297e-02 -9.76177081e-02 6.41774356e-01
1.22351050e+00 -4.24823284e-01 -1.54933572e+00 -2.81424999e-01
1.12980112e-01 -6.28352821e-01 -3.98160070e-01 -9.35910702e-01
-9.35152352e-01 9.15771782e-01 5.20460606e-01 -3.95444669e-02
9.97683227e-01 7.64488876e-01 4.72139239e-01 3.00571740e-01
3.24133992e-01 -7.51920044e-01 3.65576386e-01 3.41543704e-01
9.58149016e-01 -1.33177304e+00 -2.54562974e-01 -7.51309574e-01
-8.86544228e-01 4.45001334e-01 1.02041316e+00 1.77027985e-01
5.14216900e-01 2.69122899e-01 -9.88278911e-02 -3.49372536e-01
-1.07829332e+00 -3.38326037e-01 4.39473838e-01 4.24028963e-01
4.82247621e-01 2.97420770e-01 -1.71629712e-01 5.34170151e-01
5.39135516e-01 -8.23335052e-02 5.81149459e-02 9.00546968e-01
-4.35529172e-01 -9.64323759e-01 -8.91702175e-02 4.50492650e-01
-2.65329212e-01 6.70316890e-02 -4.60803092e-01 4.00206923e-01
4.49369878e-01 7.92228103e-01 3.79281551e-01 -4.63920146e-01
1.96555659e-01 6.34873807e-02 4.57760185e-01 -8.04499030e-01
5.87424496e-03 1.64594918e-01 1.19854407e-02 -6.27064049e-01
-3.80379945e-01 -7.10213602e-01 -1.54663622e+00 1.66738063e-01
-3.68024528e-01 -2.07919985e-01 2.08975330e-01 1.00120485e+00
5.54969788e-01 5.39299548e-01 3.44873279e-01 -8.41485023e-01
-2.10075602e-01 -8.32560897e-01 -4.34333444e-01 5.73192894e-01
1.45874530e-01 -9.01889443e-01 -3.16037685e-02 3.25586379e-01] | [9.165985107421875, 1.0153249502182007] |
0feddc9e-dfd3-4cde-9b41-4e919d46cad2 | a-fusion-model-towards-a-virtual-physical-and | 2305.09992 | null | https://arxiv.org/abs/2305.09992v1 | https://arxiv.org/pdf/2305.09992v1.pdf | A Fusion Model: Towards a Virtual, Physical and Cognitive Integration and its Principles | Virtual Reality (VR), Augmented Reality (AR), Mixed Reality (MR), digital twin, Metaverse and other related digital technologies have attracted much attention in recent years. These new emerging technologies are changing the world significantly. This research introduces a fusion model, i.e. Fusion Universe (FU), where the virtual, physical, and cognitive worlds are merged together. Therefore, it is crucial to establish a set of principles for the fusion model that is compatible with our physical universe laws and principles. This paper investigates several aspects that could affect immersive and interactive experience; and proposes the fundamental principles for Fusion Universe that can integrate physical and virtual world seamlessly. | ['Sanghyuk Lee', 'Yifan Lu', 'Yun Xue', 'Hao Lan Zhang'] | 2023-05-17 | null | null | null | null | ['mixed-reality'] | ['computer-vision'] | [-3.18170369e-01 -3.31182152e-01 2.15228066e-01 3.14824671e-01
1.51380196e-01 -6.04150236e-01 9.59120929e-01 -3.75775099e-01
1.13223186e-02 7.17568517e-01 5.37565649e-01 -4.36414957e-01
-2.33656272e-01 -1.26180601e+00 -1.55065119e-01 -9.11919102e-02
-9.47591588e-02 -4.77123380e-01 5.79653263e-01 -9.13825631e-01
3.59794229e-01 2.41039276e-01 -1.76993668e+00 -2.78339684e-02
1.22298968e+00 7.42498279e-01 7.48901010e-01 8.45468044e-01
-4.30212975e-01 7.83793569e-01 -4.17337596e-01 -2.80861229e-01
-2.30189443e-01 -5.13203561e-01 -5.03994405e-01 -5.81709802e-01
-4.82842267e-01 -4.59698051e-01 -6.66309595e-01 1.12155211e+00
5.44618607e-01 4.92805809e-01 1.59551933e-01 -1.01197720e+00
-1.58677220e+00 3.70614022e-01 -5.60209394e-01 5.33883691e-01
1.16260505e+00 -3.35776657e-01 2.52579629e-01 -6.44051373e-01
1.04539871e+00 1.06035399e+00 3.60696971e-01 2.99322098e-01
-4.68683124e-01 -5.06955624e-01 6.41241223e-02 7.90056884e-02
-1.58620799e+00 -2.06271857e-01 5.74067175e-01 -3.10702652e-01
7.44891822e-01 8.18759680e-01 1.22634852e+00 9.15984452e-01
6.28880322e-01 1.03477150e-01 1.07914829e+00 -5.73994815e-01
1.17760532e-01 6.08399093e-01 4.78221655e-01 1.27584606e-01
5.79640567e-01 4.14566100e-01 -1.03801645e-01 5.08697443e-02
1.37186980e+00 6.25711083e-02 -5.97829103e-01 -1.08615652e-01
-1.11831212e+00 1.50318399e-01 2.09995881e-01 8.89274061e-01
-4.08086061e-01 -8.13679546e-02 2.61062473e-01 1.92005858e-01
6.73655048e-02 1.51151657e-01 3.20449122e-03 -7.65124023e-01
1.76077962e-01 2.23467294e-02 5.14917195e-01 7.81942487e-01
8.34883153e-02 1.25373408e-01 3.73706162e-01 5.87612271e-01
7.40244091e-01 4.05855238e-01 6.05745137e-01 -7.70745158e-01
4.69308533e-02 3.96125466e-01 3.94514173e-01 -9.48504746e-01
-1.78803280e-01 -3.13380629e-01 -4.47862238e-01 3.06065649e-01
-4.61212784e-01 -1.65579692e-01 -6.62982583e-01 1.30778420e+00
6.45276546e-01 7.03276396e-01 3.80205363e-01 8.69972408e-01
1.85075176e+00 7.89505303e-01 9.00050178e-02 -2.78874576e-01
1.29562020e+00 -3.18111122e-01 -1.59557950e+00 2.83050746e-01
4.82514381e-01 -9.58786666e-01 9.98550534e-01 3.98831099e-01
-1.17953885e+00 -7.67837405e-01 -1.41475022e+00 3.30602825e-02
-7.93226182e-01 -6.02253616e-01 9.87580359e-01 1.46960282e+00
-1.08707404e+00 2.66998678e-01 -3.94708753e-01 -6.31874442e-01
-4.00560439e-01 7.26078376e-02 -3.52953076e-01 2.51083255e-01
-1.44807792e+00 1.27997196e+00 5.63648522e-01 6.56205788e-02
3.20301205e-02 -2.40176648e-01 -7.89130926e-01 -4.11032408e-01
1.67317539e-01 -8.36241841e-01 9.83547091e-01 -5.07915914e-01
-1.75911140e+00 5.49777329e-01 2.76279509e-01 2.67820626e-01
3.95998687e-01 -6.42038107e-01 -1.57600045e+00 -1.83500022e-01
-2.75932312e-01 -3.43670666e-01 -3.54594886e-01 -1.80184376e+00
-7.23861396e-01 -3.94371927e-01 4.20594931e-01 4.06986415e-01
-4.24224399e-02 3.28264803e-01 -3.88325572e-01 -1.53181106e-01
5.27515769e-01 -3.69485855e-01 -1.39280573e-01 -5.34880400e-01
1.24745511e-01 1.42263189e-01 1.01424205e+00 -7.60278583e-01
1.57181239e+00 -2.21975327e+00 -1.20485509e-02 6.97566196e-02
4.94234204e-01 2.14743868e-01 4.31135029e-01 6.75608158e-01
3.62888649e-02 3.34974110e-01 7.34592497e-01 2.60381013e-01
-4.28533405e-02 1.42644539e-01 -1.86834693e-01 3.55289340e-01
-9.39963698e-01 8.36518884e-01 -1.09536779e+00 -4.69280124e-01
8.30888748e-01 8.45875859e-01 -1.93732142e-01 6.02880307e-02
4.07850385e-01 7.07157671e-01 -5.45555115e-01 7.23347962e-01
1.11837709e+00 2.17060000e-01 1.77190796e-01 1.63376052e-02
-7.29593158e-01 -1.60927862e-01 -1.62785494e+00 1.56100309e+00
-6.11626089e-01 4.10539955e-01 -5.45410737e-02 6.12684190e-02
1.02223313e+00 8.19342792e-01 2.04968929e-01 -1.33040774e+00
7.23067284e-01 3.36753905e-01 -2.42675841e-01 -1.12008083e+00
1.27461529e+00 -3.42741728e-01 1.55738711e-01 1.22030005e-01
-1.62360415e-01 7.79519901e-02 -3.92586112e-01 2.43611142e-01
7.45095074e-01 5.97784519e-01 6.67362869e-01 1.64273158e-02
2.17494994e-01 -5.43517470e-01 4.43513364e-01 3.42265457e-01
-4.35143530e-01 4.76876527e-01 -1.67071790e-01 -8.54965523e-02
-6.45973384e-01 -1.64857793e+00 -1.67748570e-01 5.23635626e-01
1.44148219e+00 -5.37524819e-01 -2.16887832e-01 9.44974273e-02
-5.16374409e-01 1.02976441e+00 -6.07246876e-01 -1.21509328e-01
-1.14362322e-01 -2.70853847e-01 1.28927389e-02 3.98773432e-01
7.03835964e-01 -7.70181000e-01 -6.57877445e-01 1.70060322e-02
-7.98000619e-02 -1.06707454e+00 3.83564532e-01 -5.59551775e-01
-8.50925207e-01 -6.30830407e-01 -3.09080750e-01 -3.06588411e-01
2.59067900e-02 9.60876703e-01 8.46845388e-01 2.48060763e-01
8.70899633e-02 5.65174520e-01 -1.01757860e+00 -2.19036549e-01
-9.47093815e-02 -7.01939642e-01 8.96697864e-02 -3.52981776e-01
-2.20998064e-01 -7.87878513e-01 -4.51918781e-01 1.41637567e-02
-8.25545609e-01 3.68283033e-01 -3.92840058e-02 -5.05982302e-02
2.48750091e-01 1.67593464e-01 6.92233622e-01 -5.93835771e-01
7.62079239e-01 -8.60438824e-01 -1.58579096e-01 3.80990267e-01
1.61017478e-01 -8.43517780e-01 1.63910612e-01 -3.55631113e-01
-1.75207090e+00 -9.50453341e-01 -1.84351891e-01 -9.22066495e-02
-2.85146236e-02 5.75799763e-01 -7.87325323e-01 -2.71568149e-01
5.82271695e-01 3.84748615e-02 -4.44647133e-01 -5.82694113e-01
6.75063729e-01 1.03435206e+00 6.16675317e-01 -2.58397013e-01
4.28153127e-01 4.85351890e-01 -3.83176684e-01 -9.28502917e-01
1.35182947e-01 -1.62121788e-01 -3.14628780e-01 -9.21828151e-01
9.41244245e-01 -8.35746646e-01 -7.38633454e-01 -1.45828575e-01
-1.04187608e+00 1.85952753e-01 -3.40461552e-01 1.25152218e+00
-3.01935911e-01 3.68280143e-01 -4.08161283e-01 -1.48609900e+00
9.45570506e-03 -7.31427610e-01 3.47465008e-01 9.70666885e-01
-1.62315816e-01 -1.04507685e+00 3.51019233e-01 3.09007198e-01
3.87577176e-01 7.73786366e-01 1.17687762e-01 6.72091991e-02
-7.37199426e-01 -2.49936998e-01 -2.91826665e-01 -1.10213816e-01
4.29742724e-01 6.38786316e-01 -7.18774438e-01 3.73974681e-01
2.69443572e-01 3.85379851e-01 1.54525777e-02 1.78679273e-01
2.51032412e-01 3.84918094e-01 -5.27745664e-01 5.85155368e-01
1.77376759e+00 1.25032413e+00 1.43747652e+00 5.79933763e-01
4.78350252e-01 1.49352774e-01 9.05933499e-01 5.21022797e-01
5.92659533e-01 3.26958776e-01 2.11126208e-01 2.58295648e-02
-8.03129971e-02 -2.23434716e-01 8.28824565e-03 1.22174585e+00
-9.16477859e-01 -1.46438554e-01 -7.55880892e-01 7.88293630e-02
-1.85069752e+00 -1.13788426e+00 -7.61458516e-01 2.26036072e+00
-1.13230333e-01 1.94911942e-01 -2.10947394e-01 -2.67020985e-02
8.13647687e-01 -1.41891718e-01 2.20647752e-01 -5.54308712e-01
-2.18649000e-01 6.64799586e-02 -5.58772385e-02 3.60788792e-01
-6.24472797e-01 6.13234699e-01 6.94538355e+00 4.00411844e-01
-9.42354977e-01 6.20429099e-01 -4.78584804e-02 2.55217731e-01
-8.70478690e-01 2.75116295e-01 -3.08744144e-02 5.06530702e-01
7.48356223e-01 -6.27733469e-01 3.15011859e-01 4.85124916e-01
3.43220472e-01 -4.53655005e-01 -2.29547918e-01 1.13046777e+00
-1.08537272e-01 -1.43784988e+00 -2.16860488e-01 3.24399114e-01
8.56197715e-01 -3.05243939e-01 7.79967979e-02 1.99805647e-01
4.23824102e-01 -6.48197174e-01 7.76454926e-01 1.07621360e+00
7.07808971e-01 -7.75774181e-01 8.03898811e-01 1.69679206e-02
-1.52633202e+00 5.68278879e-02 -2.23074749e-01 -4.97530431e-01
7.83048391e-01 1.70372397e-01 2.29758173e-01 1.41466570e+00
6.57932460e-01 1.82635084e-01 -9.42615047e-02 1.39744997e+00
4.24829751e-01 -2.28619307e-01 -1.31429732e-01 6.15274301e-03
-2.46250436e-01 -6.44547939e-01 5.14026165e-01 6.30796909e-01
8.02789867e-01 7.70621479e-01 -2.03556627e-01 7.89836943e-01
3.92412663e-01 2.66321957e-01 -1.00558877e+00 6.14377744e-02
7.41945386e-01 9.87314105e-01 -1.04915285e+00 -4.48401153e-01
-6.40138566e-01 6.78011179e-01 -4.92380828e-01 3.14469665e-01
-1.30080616e+00 -2.73176759e-01 4.18103099e-01 -2.72826571e-02
-3.75796556e-01 -5.00027478e-01 -6.64977670e-01 -1.32374907e+00
-2.96449274e-01 -1.84754997e-01 1.23357944e-01 -1.24969363e+00
-6.75095260e-01 6.06230855e-01 -6.96148053e-02 -1.47440386e+00
4.40542847e-01 -5.25353593e-04 -7.34714866e-01 8.81784141e-01
-9.55608368e-01 -1.18748200e+00 -6.61214054e-01 1.94194525e-01
-1.02013394e-01 2.74268657e-01 8.43538523e-01 5.69547176e-01
-5.48159540e-01 -1.61205471e-01 3.50122631e-01 -6.01633906e-01
1.15052134e-01 -8.55400324e-01 3.85506243e-01 7.81193912e-01
-3.49550635e-01 8.78744841e-01 1.04774046e+00 -1.06632173e+00
-1.65866482e+00 8.00568424e-03 4.16512221e-01 -4.64937329e-01
4.92803693e-01 3.63348648e-02 -6.84501231e-01 7.85504043e-01
3.47966254e-01 -9.87458527e-02 9.13301528e-01 2.71043450e-01
6.47538528e-03 2.35155240e-01 -1.44567883e+00 1.10467756e+00
1.46304500e+00 -3.62139642e-01 -7.93891788e-01 -2.15018436e-01
1.12869108e+00 -5.09877980e-01 -1.06317449e+00 3.56624246e-01
9.83026624e-01 -1.44422710e+00 1.02498877e+00 -1.13968272e-02
-5.89178018e-02 -5.48765600e-01 -5.43248236e-01 -9.29630339e-01
-4.08185393e-01 -6.23574615e-01 -4.07132447e-01 1.39580739e+00
-2.28674680e-01 -1.27026796e+00 9.99685898e-02 7.74227917e-01
-2.71690637e-01 -2.02064216e-01 -8.46339703e-01 -7.48324871e-01
-4.87557590e-01 -4.22852278e-01 1.16114271e+00 1.35106885e+00
6.64403856e-01 1.04264393e-01 -2.58639842e-01 2.68564135e-01
-9.74954572e-03 -3.25176805e-01 6.93831921e-01 -1.35280669e+00
-8.58682171e-02 -2.78995305e-01 -7.06356347e-01 -6.55714035e-01
-7.97801614e-01 -6.45943210e-02 -6.59974158e-01 -2.29534507e+00
-8.91930889e-04 -4.80169982e-01 -4.14912462e-01 -7.56017029e-01
-5.04153669e-02 8.34785104e-02 3.95987332e-01 9.97220203e-02
-5.70070744e-01 7.20273316e-01 1.78304696e+00 7.45561481e-01
-7.32294500e-01 -5.48165739e-01 -9.03856039e-01 5.85430622e-01
6.24846280e-01 4.89348680e-01 -6.60418272e-01 -7.73531720e-02
4.59607571e-01 6.37528777e-01 1.23370394e-01 -1.16787875e+00
1.15123644e-01 -4.35874462e-01 2.67380983e-01 -4.73256737e-01
6.37223780e-01 -9.04330552e-01 1.13609159e+00 2.76026130e-01
6.43842578e-01 1.38976067e-01 7.42214099e-02 2.42091537e-01
7.45089203e-02 -1.33199515e-02 3.66400659e-01 -1.13763750e-01
-1.10863781e+00 -3.11640859e-01 -2.46087387e-01 -6.79416955e-01
1.61989617e+00 -1.12243938e+00 -6.26216233e-01 -2.39876464e-01
-8.93602490e-01 -3.16249788e-01 7.10635126e-01 5.92293680e-01
7.55313635e-01 -1.41146803e+00 -8.78265202e-02 -7.25574419e-02
-1.46578595e-01 -5.83257794e-01 1.15602195e+00 4.28816736e-01
-1.00800633e+00 1.24663651e-01 -6.60545170e-01 1.24181777e-01
-1.23986483e+00 7.57076383e-01 2.61937648e-01 4.10177976e-01
-8.41272831e-01 5.61653316e-01 2.64479160e-01 -2.59148985e-01
-2.63835222e-01 -5.90575933e-02 -8.89034092e-01 -2.02584684e-01
8.01992893e-01 8.28932166e-01 -5.52075446e-01 -9.93033886e-01
-2.07675323e-01 5.48906088e-01 3.70357305e-01 -5.15421867e-01
8.80593777e-01 -8.27510893e-01 -2.13251278e-01 9.01736677e-01
6.57797098e-01 2.96205133e-01 -3.18693429e-01 3.82339269e-01
-5.08210540e-01 -1.19962025e+00 2.07710415e-01 -8.39410245e-01
-5.33173859e-01 3.09997141e-01 1.05387020e+00 8.33854675e-01
1.01431549e+00 -7.32385442e-02 8.06998432e-01 -5.50150812e-01
9.09897387e-01 -1.03583121e+00 -2.84078807e-01 8.13578069e-02
7.63599992e-01 -5.32281399e-01 -7.95158073e-02 -1.05138445e+00
-4.51894253e-01 8.27579975e-01 8.90325129e-01 -5.73996780e-03
1.05279613e+00 4.14947957e-01 -2.17804890e-02 -3.04378182e-01
-4.51019168e-01 -3.99695218e-01 -1.33160993e-01 1.03415573e+00
8.18419456e-01 4.81734157e-01 -8.54200244e-01 9.78020608e-01
-2.57605463e-01 4.73307550e-01 8.51349771e-01 1.40500784e+00
-7.48330593e-01 -9.04317975e-01 -9.81159449e-01 9.43910778e-02
-2.29682446e-01 5.28618336e-01 2.79911682e-02 8.40765476e-01
6.16280556e-01 1.12648678e+00 8.11655223e-02 -9.86980438e-01
5.49642920e-01 -5.42557001e-01 6.92297578e-01 -1.67839214e-01
-4.36558545e-01 -9.66909751e-02 2.14718059e-01 -3.26182365e-01
-1.66322902e-01 -2.94939369e-01 -1.55484343e+00 -9.02556121e-01
-7.45712399e-01 1.33195758e-01 8.71142805e-01 8.05701196e-01
2.60482907e-01 1.05302048e+00 2.15068981e-01 -6.12103581e-01
4.95623022e-01 -6.64120853e-01 -9.35034335e-01 -1.64278194e-01
-5.15943393e-02 -9.41333532e-01 1.36398673e-01 -3.54045421e-01] | [8.202964782714844, -1.5629075765609741] |
1da0bbc0-4ca3-47be-a7a5-eb18f16a63e2 | cross-lingual-alzheimer-s-disease-detection | 2303.07650 | null | https://arxiv.org/abs/2303.07650v1 | https://arxiv.org/pdf/2303.07650v1.pdf | Cross-lingual Alzheimer's Disease detection based on paralinguistic and pre-trained features | We present our submission to the ICASSP-SPGC-2023 ADReSS-M Challenge Task, which aims to investigate which acoustic features can be generalized and transferred across languages for Alzheimer's Disease (AD) prediction. The challenge consists of two tasks: one is to classify the speech of AD patients and healthy individuals, and the other is to infer Mini Mental State Examination (MMSE) score based on speech only. The difficulty is mainly embodied in the mismatch of the dataset, in which the training set is in English while the test set is in Greek. We extract paralinguistic features using openSmile toolkit and acoustic features using XLSR-53. In addition, we extract linguistic features after transcribing the speech into text. These features are used as indicators for AD detection in our method. Our method achieves an accuracy of 69.6% on the classification task and a root mean squared error (RMSE) of 4.788 on the regression task. The results show that our proposed method is expected to achieve automatic multilingual Alzheimer's Disease detection through spontaneous speech. | ['Wei-Qiang Zhang', 'Jinpeng Li', 'Yu Pu', 'Xuchu Chen'] | 2023-03-14 | null | null | null | null | ['alzheimer-s-disease-detection'] | ['medical'] | [ 5.18944487e-03 1.71053752e-01 3.56125504e-01 -6.43755734e-01
-1.50871837e+00 -6.01054728e-02 5.10896862e-01 -1.56498760e-01
-6.90886199e-01 6.07184649e-01 4.74363416e-01 -2.09496513e-01
3.17001075e-01 -2.45138973e-01 -1.63655698e-01 -4.28441703e-01
-3.60650927e-01 4.60886389e-01 4.75095548e-02 -1.62683979e-01
-4.30647820e-01 1.13928117e-01 -1.07833254e+00 5.70999503e-01
9.08468127e-01 1.25630808e+00 3.48479837e-01 7.49169946e-01
9.24728364e-02 5.75274706e-01 -6.91305876e-01 -5.58446586e-01
-2.56904494e-02 -3.42560291e-01 -9.39464509e-01 -6.07123114e-02
2.87187129e-01 -3.57291400e-01 -4.45228428e-01 1.07467246e+00
9.16406095e-01 -3.53927553e-01 6.88123524e-01 -8.54211628e-01
-4.68187481e-01 5.90962768e-01 1.17172949e-01 5.03501713e-01
4.20961201e-01 -4.76159975e-02 1.00759459e+00 -9.26235557e-01
3.37107778e-01 1.38596189e+00 7.88383663e-01 7.03214586e-01
-9.57543731e-01 -5.10994971e-01 -4.63301476e-05 6.94452465e-01
-1.22958863e+00 -1.02742970e+00 4.41740215e-01 -6.02220774e-01
9.01540160e-01 2.38971010e-01 7.73733854e-01 1.46277225e+00
1.69414982e-01 7.01065779e-01 1.27857471e+00 -4.29714620e-01
4.53914940e-01 2.45060354e-01 3.71554554e-01 4.91446316e-01
-3.29069704e-01 -6.31254688e-02 -4.24944937e-01 -5.33362508e-01
2.17655987e-01 -6.29178286e-01 -1.78574473e-01 3.93178016e-01
-1.36872196e+00 9.53308702e-01 9.78532508e-02 3.66333485e-01
-5.34676552e-01 -3.35353285e-01 6.19885266e-01 6.05668247e-01
6.17635906e-01 -6.92478940e-02 -7.84639657e-01 -2.20786721e-01
-8.64901483e-01 1.45856977e-01 8.11573625e-01 5.33414721e-01
-1.72178552e-01 9.94187146e-02 7.82232545e-03 1.47165430e+00
8.43724310e-01 9.62706804e-01 1.01411700e+00 -6.99762523e-01
7.22232223e-01 1.90742105e-01 -3.07522446e-01 -4.68851238e-01
-6.39157891e-01 -1.38380721e-01 -7.04262614e-01 -2.59832829e-01
3.78304839e-01 -4.45808470e-01 -7.80113518e-01 1.71841240e+00
-8.50566328e-02 -1.34774491e-01 2.83449531e-01 7.24459708e-01
8.75127077e-01 4.96470928e-01 3.20461929e-01 -4.24378633e-01
1.81492388e+00 -6.24933362e-01 -7.98650503e-01 -6.38914227e-01
8.61328840e-01 -6.34384751e-01 8.92672062e-01 4.44257528e-01
-8.67045343e-01 -3.50625604e-01 -1.01203334e+00 1.17122419e-01
-1.70238703e-01 7.46231914e-01 1.68710247e-01 9.94749308e-01
-1.13933218e+00 7.41030276e-02 -1.01734841e+00 -4.34036762e-01
3.45955342e-01 3.43815237e-01 -5.98746061e-01 1.36780336e-01
-1.55227351e+00 9.63428199e-01 1.57308236e-01 1.06907971e-01
-4.50425267e-01 -4.52745020e-01 -6.95935905e-01 -4.81250823e-01
-3.23619902e-01 -4.58952665e-01 1.30983448e+00 -7.91528523e-01
-1.58172798e+00 1.09345615e+00 -1.23476863e-01 -7.28959918e-01
5.06254196e-01 -2.62710661e-01 -1.10891771e+00 2.54324108e-01
3.12378705e-01 5.48776567e-01 5.30825198e-01 -3.88333350e-01
-6.90065682e-01 -1.07954848e+00 -7.46535122e-01 -4.92021814e-02
-4.06289846e-01 4.31401551e-01 9.09419656e-02 -7.54164457e-01
2.18528226e-01 -1.11866510e+00 -9.00937691e-02 9.49857850e-03
-4.88238275e-01 -2.73792386e-01 4.37186301e-01 -1.55476689e+00
1.17670679e+00 -2.14425039e+00 -1.02385819e-01 1.35314941e-01
8.97261724e-02 2.85462409e-01 -2.08090529e-01 -9.01105553e-02
-1.89290479e-01 -9.09544826e-02 -2.91965544e-01 -3.62950057e-01
7.09642991e-02 -1.01236850e-02 -1.49058312e-01 3.21059406e-01
3.13009858e-01 6.68907166e-01 -3.79819512e-01 -2.49112770e-01
-2.07589552e-01 5.70088327e-01 -6.01887226e-01 1.08864106e-01
2.17005715e-01 4.97136861e-01 -4.49478358e-01 3.49375784e-01
4.01507884e-01 2.94608474e-01 1.31709114e-01 -4.31013890e-02
8.12754110e-02 9.40862060e-01 -8.82309318e-01 1.45977294e+00
-2.84303963e-01 4.14655447e-01 1.41968979e-02 -9.97185111e-01
9.45570052e-01 5.98179042e-01 3.06938022e-01 -7.93727577e-01
3.29808630e-02 5.28279543e-01 3.46497804e-01 -7.56168723e-01
-3.38820070e-01 -1.02497172e-03 -8.99491087e-02 1.32836059e-01
-6.53118500e-03 3.13632935e-01 1.38539046e-01 -1.80819742e-02
1.27286458e+00 -4.89839733e-01 5.13191760e-01 -3.48626286e-01
7.98900485e-01 -4.30281341e-01 5.88903785e-01 4.08445179e-01
-5.59300423e-01 4.36744928e-01 4.88894373e-01 -1.56798661e-01
-8.46321404e-01 -1.28022087e+00 -4.25673872e-01 7.67818272e-01
-9.99026358e-01 -6.32845640e-01 -8.93847108e-01 -8.24474692e-01
-3.03792983e-01 8.34303916e-01 -3.21431190e-01 -1.21990740e-01
-5.24766743e-01 -1.09971178e+00 8.87481868e-01 5.50693870e-01
5.52796125e-01 -8.35857868e-01 -1.21400833e-01 2.25354731e-01
-5.66475332e-01 -1.40287137e+00 -5.29524744e-01 2.01834053e-01
-8.78942728e-01 -9.13445234e-01 -1.00159419e+00 -9.93916988e-01
9.48639587e-02 -5.50494432e-01 7.53697455e-01 -7.99901664e-01
-1.50378108e-01 4.91199851e-01 -3.80154639e-01 -4.59376693e-01
-7.84818709e-01 1.96006522e-01 3.88920009e-01 2.65507698e-02
8.36524427e-01 -5.19743502e-01 -4.36827362e-01 1.89510942e-01
-2.58311242e-01 -4.60675776e-01 7.51942873e-01 5.38489401e-01
4.73387241e-01 -4.29937929e-01 1.04142261e+00 -2.87458062e-01
6.82633042e-01 -5.20485818e-01 -7.60047976e-03 1.96208358e-01
-3.77548665e-01 -9.54603925e-02 5.03736138e-01 -3.52362454e-01
-6.26792789e-01 3.89012426e-01 -8.94101024e-01 3.82353693e-01
-4.14676249e-01 3.85453939e-01 -5.06624758e-01 4.88626868e-01
5.84062457e-01 5.07883966e-01 1.35932878e-01 -7.02521384e-01
-1.71090942e-02 1.52306521e+00 3.93674761e-01 -1.40275106e-01
8.93424228e-02 1.25510478e-02 -6.36121631e-01 -1.38603187e+00
-6.59994066e-01 -3.12650174e-01 -7.54603505e-01 3.08436379e-02
1.08575511e+00 -1.11422670e+00 -2.70742565e-01 8.41015220e-01
-1.16115749e+00 -1.46633461e-01 1.53667480e-01 9.26941216e-01
-6.50150359e-01 4.14633870e-01 -6.65470123e-01 -6.66290164e-01
-6.98639810e-01 -1.35653126e+00 1.15844190e+00 -4.87152874e-01
-6.34127796e-01 -8.51232231e-01 2.34697551e-01 7.12029636e-01
2.40436688e-01 -4.08841282e-01 9.25918579e-01 -1.23196745e+00
3.76776516e-01 -3.17214280e-01 -1.28208831e-01 8.21938574e-01
2.74491817e-01 -7.32923627e-01 -1.01937079e+00 -9.71724689e-02
3.28602344e-01 -3.41923326e-01 7.32789040e-01 3.92122030e-01
7.00697064e-01 -4.09592956e-01 -4.53890711e-02 -6.40364438e-02
5.05471766e-01 3.86136740e-01 8.15518439e-01 2.23582670e-01
2.04145446e-01 4.17023897e-01 1.08387172e-01 3.76111150e-01
7.13653624e-01 9.79012370e-01 -1.73967808e-01 4.61703897e-01
-2.62027532e-01 3.61834973e-01 1.06085169e+00 1.40433133e+00
2.54994243e-01 9.47174802e-02 -1.12046325e+00 3.16663980e-01
-1.32383001e+00 -6.15018129e-01 -3.22806001e-01 2.12130570e+00
1.12856615e+00 -1.09116301e-01 5.39554894e-01 4.37435418e-01
5.40556014e-01 -8.70469362e-02 -1.29522473e-01 -1.51968524e-01
-7.45954141e-02 8.75881463e-02 1.45534221e-02 3.31771284e-01
-1.35884416e+00 4.95562643e-01 6.16652393e+00 5.03535271e-01
-1.22706711e+00 4.13827717e-01 4.84380752e-01 -1.03334725e-01
3.95578682e-01 -7.67554402e-01 -8.08158517e-01 7.75200665e-01
1.91146362e+00 2.69075751e-01 2.01597169e-01 6.32921398e-01
3.35784882e-01 1.00181878e-01 -1.15308392e+00 1.05300999e+00
3.05239320e-01 -4.09993410e-01 -1.44215245e-02 3.91957872e-02
2.66416352e-02 4.50719804e-01 8.65219906e-02 2.89866298e-01
-2.60392964e-01 -8.79466712e-01 7.38241553e-01 6.76108301e-01
6.50977969e-01 -5.65042198e-01 7.94172406e-01 9.07614902e-02
-1.02298307e+00 -1.12034470e-01 -2.09521800e-01 4.05546904e-01
2.06517622e-01 6.90541744e-01 -1.10724688e+00 1.16646819e-01
6.85737073e-01 5.68511665e-01 -5.47556281e-01 9.68929291e-01
-3.16042006e-02 9.92774010e-01 -2.75156885e-01 -4.61721979e-02
-8.53773654e-02 -2.92546842e-02 7.65186608e-01 1.44189811e+00
5.34821093e-01 -1.02709807e-01 4.02614847e-02 6.65318012e-01
2.33350679e-01 5.39538026e-01 -2.19556883e-01 -3.35138917e-01
3.02020162e-01 7.56722450e-01 -1.71024352e-01 -1.92976803e-01
-5.15709400e-01 1.13399088e+00 3.28685381e-02 1.26916654e-02
-5.49384594e-01 -3.73482794e-01 5.27966797e-01 -2.01007761e-02
2.19024777e-01 -2.98718780e-01 -1.81848779e-01 -1.05046988e+00
2.62377739e-01 -1.12612343e+00 2.93381602e-01 -6.42702162e-01
-1.58171070e+00 9.89614367e-01 -3.98447514e-01 -9.62055385e-01
-4.22775924e-01 -9.02771831e-01 -1.89957947e-01 9.97264922e-01
-1.13176584e+00 -1.09900463e+00 1.16495855e-01 5.64005673e-01
6.24784648e-01 -8.48975003e-01 1.28204513e+00 8.93943787e-01
-5.73241174e-01 6.83708906e-01 4.53904979e-02 4.26001668e-01
8.79802287e-01 -1.16249657e+00 4.33668941e-01 3.23887795e-01
1.69829577e-01 4.19814169e-01 4.96904314e-01 -6.49636686e-01
-7.73048341e-01 -1.32253361e+00 1.56199408e+00 -4.12464082e-01
9.96503115e-01 -3.61671716e-01 -6.34069622e-01 5.09831905e-01
-2.66119659e-01 -1.56314224e-01 8.91414046e-01 -1.24464534e-01
-3.25896621e-01 -1.28533348e-01 -1.15539956e+00 3.89806241e-01
7.95346737e-01 -8.26469541e-01 -7.17962861e-01 4.85140294e-01
4.35498744e-01 -1.67126451e-02 -1.21804583e+00 2.04390511e-01
6.59001946e-01 -5.33124149e-01 9.91032779e-01 -7.86279142e-01
2.92145342e-01 1.83013320e-01 -7.42124319e-01 -1.62105024e+00
-1.67649817e-02 -1.95811629e-01 2.01598778e-01 1.11990678e+00
6.60381913e-01 -8.90941739e-01 1.60341337e-01 3.08654428e-01
-1.51928186e-01 -5.92396498e-01 -1.33960247e+00 -1.00325632e+00
3.33905280e-01 -8.99735749e-01 -8.95585120e-02 5.73684275e-01
3.04967593e-02 4.83756602e-01 -9.74317454e-03 2.84797013e-01
4.30730104e-01 -9.21162128e-01 -5.19142905e-03 -1.61745048e+00
-4.73942831e-02 -1.22640304e-01 -8.11982453e-01 -5.76959014e-01
3.41330498e-01 -1.18456185e+00 -6.05705120e-02 -1.25869465e+00
1.38442367e-01 -1.48227796e-01 -1.98464647e-01 5.36658347e-01
9.68186632e-02 1.62356287e-01 7.16115758e-02 6.60460293e-02
-3.11401963e-01 7.96623051e-01 5.39546907e-01 -3.51166219e-01
-4.71606523e-01 5.63178957e-01 -6.13856673e-01 8.02697182e-01
9.29336965e-01 -4.49823022e-01 -1.86189681e-01 -5.86019680e-02
-3.15340430e-01 -8.37122649e-02 6.21491730e-01 -1.16121304e+00
-1.93510652e-01 5.61989129e-01 2.25270078e-01 -2.99738944e-01
5.50913990e-01 -5.14070451e-01 -2.51173168e-01 5.73639631e-01
-4.78200614e-01 -7.56821111e-02 1.69785708e-01 4.50580388e-01
1.69838127e-03 -2.10034996e-01 8.08304667e-01 7.93458149e-02
-1.63112566e-01 -9.19134822e-03 -8.94014835e-01 3.19602370e-01
5.33967972e-01 2.63698637e-01 -2.58104742e-01 -2.74291009e-01
-1.39744067e+00 -1.65350348e-01 -4.46327239e-01 6.03817463e-01
4.64193583e-01 -1.53497493e+00 -1.17559016e+00 2.91699857e-01
1.25670284e-01 -8.75113070e-01 1.46839947e-01 1.37320781e+00
-1.77599892e-01 8.27613890e-01 -8.98363665e-02 -4.57265496e-01
-1.56918728e+00 -2.08340973e-01 5.01466811e-01 -8.46036896e-02
-6.36759102e-01 6.59579873e-01 -2.36172035e-01 -4.48616177e-01
4.40636337e-01 -2.85771191e-01 -2.70666182e-01 3.57517332e-01
9.65021431e-01 5.30643344e-01 6.03597045e-01 -7.69524395e-01
-6.79361522e-01 4.98855747e-02 -4.58673716e-01 -4.28273320e-01
1.58784640e+00 -1.68985575e-01 -1.66540116e-01 7.02487946e-01
1.36370158e+00 -6.61336482e-02 -3.31528425e-01 -2.35344395e-01
3.48454207e-01 4.14034367e-01 3.50173503e-01 -1.15660036e+00
-9.41760600e-01 8.58851910e-01 1.46763086e+00 2.40499288e-01
6.74426436e-01 2.86372632e-01 1.10842037e+00 5.89322150e-01
7.38648325e-02 -1.32139158e+00 -2.03666344e-01 6.21020913e-01
1.38081634e+00 -1.14307117e+00 -4.30801004e-01 -2.05220968e-01
-8.08063149e-01 1.24766445e+00 5.07705212e-02 5.48506901e-02
8.94983888e-01 1.61060661e-01 2.70384938e-01 5.26250377e-02
-5.73603988e-01 -1.48023218e-01 6.22970283e-01 8.83447528e-01
6.53623819e-01 2.26563573e-01 -4.09516215e-01 1.46649528e+00
-7.97981262e-01 -1.18283592e-01 2.29645714e-01 4.40622121e-01
-4.99450952e-01 -9.95110691e-01 -3.99193704e-01 7.67290831e-01
-7.61453390e-01 -1.91377625e-01 -6.92702413e-01 3.30238044e-01
-2.33886018e-03 1.17520797e+00 -3.84989101e-03 -2.63401568e-01
3.65968019e-01 7.41625905e-01 2.69715160e-01 -5.31947315e-01
-9.57051367e-02 4.79330495e-02 5.57768226e-01 -4.15845364e-01
-3.64298731e-01 -1.08013546e+00 -1.11078823e+00 3.63160372e-01
2.44849861e-01 -8.32316279e-03 8.00684869e-01 1.22833669e+00
3.84942383e-01 6.75535858e-01 4.47777361e-01 -3.72783929e-01
-8.67781460e-01 -1.37785900e+00 -6.46357656e-01 -1.21533930e-01
4.13408130e-01 -3.86006385e-01 -3.59574646e-01 2.12112948e-01] | [13.906970977783203, 5.361222743988037] |
9a89efc9-c180-4757-a357-e5cdda065fd0 | fine-tashkeel-finetuning-byte-level-models | 2303.14588 | null | https://arxiv.org/abs/2303.14588v1 | https://arxiv.org/pdf/2303.14588v1.pdf | Fine-Tashkeel: Finetuning Byte-Level Models for Accurate Arabic Text Diacritization | Most of previous work on learning diacritization of the Arabic language relied on training models from scratch. In this paper, we investigate how to leverage pre-trained language models to learn diacritization. We finetune token-free pre-trained multilingual models (ByT5) to learn to predict and insert missing diacritics in Arabic text, a complex task that requires understanding the sentence semantics and the morphological structure of the tokens. We show that we can achieve state-of-the-art on the diacritization task with minimal amount of training and no feature engineering, reducing WER by 40%. We release our finetuned models for the greater benefit of the researchers in the community. | ['Rami Al-Rfou', 'Gheith Abandah', 'Bashar Al-Rfooh'] | 2023-03-25 | null | null | null | null | ['feature-engineering', 'arabic-text-diacritization'] | ['methodology', 'natural-language-processing'] | [ 1.35063142e-01 1.23402931e-01 -1.64122477e-01 -4.55054402e-01
-8.66684377e-01 -8.59569550e-01 3.09651792e-01 4.37205076e-01
-6.70434833e-01 7.35093415e-01 9.34026986e-02 -9.50207353e-01
4.06108707e-01 -7.26288259e-01 -8.78222883e-01 -1.81451961e-01
-4.38644364e-02 3.59330565e-01 2.53743261e-01 -9.11808014e-01
2.41394922e-01 2.61380136e-01 -7.65368104e-01 8.85113895e-01
1.26694655e+00 2.60174364e-01 -4.10519075e-03 6.00265205e-01
-1.13334283e-01 7.00406253e-01 -7.96149611e-01 -7.81430483e-01
1.63512215e-01 -3.28772932e-01 -8.89015019e-01 -2.50152886e-01
3.76617700e-01 -2.19666094e-01 2.50803176e-02 6.78887725e-01
2.94685103e-02 -2.34584242e-01 7.16876030e-01 -5.88801563e-01
-8.81503105e-01 1.14735639e+00 -5.21388590e-01 1.34125531e-01
2.05758870e-01 -7.28394240e-02 1.22456932e+00 -1.10211146e+00
6.17482662e-01 1.12797487e+00 6.36971474e-01 3.80540431e-01
-8.62616062e-01 -4.42491174e-01 2.60582417e-01 3.88940334e-01
-1.32036483e+00 -4.72018898e-01 3.94216895e-01 -1.30559355e-01
1.64408755e+00 2.34415367e-01 3.21056396e-01 4.24419940e-01
2.84011781e-01 9.65116322e-01 9.92002547e-01 -1.31618845e+00
-4.03124124e-01 -4.88096662e-02 3.44886124e-01 1.04910469e+00
1.40440017e-01 -4.29445535e-01 -5.01428425e-01 3.98550779e-01
2.73815155e-01 -4.44010854e-01 2.14845791e-01 1.13515131e-01
-1.17334580e+00 7.79525220e-01 1.94291785e-01 4.51161683e-01
8.21237862e-02 -1.42070344e-02 5.80939233e-01 5.98700166e-01
2.99650818e-01 4.97368068e-01 -8.87098372e-01 -3.62388909e-01
-9.13515031e-01 -1.12498596e-01 8.35938334e-01 9.32190657e-01
8.23727012e-01 1.45890594e-01 2.71725625e-01 7.60298431e-01
1.99963331e-01 7.49837816e-01 4.67407942e-01 -5.00674248e-01
9.16025758e-01 6.86143339e-01 -7.43019208e-02 -4.48836505e-01
-1.21501714e-01 -2.03658491e-02 -1.61783457e-01 -1.12241901e-01
8.18478763e-01 -6.08115554e-01 -9.96733785e-01 1.27559161e+00
-1.43645555e-02 -6.20930254e-01 3.31790328e-01 2.07896277e-01
3.26786816e-01 9.05647814e-01 -1.58458427e-01 -4.17641550e-02
1.18586242e+00 -1.14261234e+00 -6.51963949e-01 -6.54765666e-01
1.30870855e+00 -1.17515624e+00 1.40712202e+00 4.60643619e-01
-1.31897473e+00 -3.04579288e-01 -1.40776598e+00 -3.75978976e-01
-6.32163405e-01 4.43055332e-01 5.62197447e-01 1.01276517e+00
-7.73955047e-01 6.22524679e-01 -1.19281626e+00 -2.63490766e-01
-7.23021105e-03 4.39674377e-01 -3.86625767e-01 -2.33957157e-01
-1.25380170e+00 1.34161091e+00 6.45531714e-01 2.42226675e-01
-3.66507113e-01 -5.82856834e-01 -1.12758899e+00 -1.21438339e-01
1.84898838e-01 2.70729535e-03 1.30040240e+00 -1.07297361e+00
-1.68060863e+00 8.92537475e-01 -3.97500128e-01 -4.34236050e-01
5.83858113e-04 -4.95619595e-01 -2.95014679e-01 -2.06479579e-01
-2.34907582e-01 3.74732912e-01 6.44353390e-01 -9.49721992e-01
-7.82953322e-01 -2.04895958e-01 1.10211268e-01 3.09636950e-01
-4.72136617e-01 4.88353163e-01 -3.12045336e-01 -7.47007549e-01
7.64084980e-03 -1.06337416e+00 1.29959255e-01 -6.40328646e-01
-2.15791523e-01 -3.57723027e-01 5.42765141e-01 -1.18508232e+00
1.45943654e+00 -1.93471551e+00 -1.55726317e-02 2.80172825e-01
-3.62951010e-01 5.85213304e-01 -3.04575384e-01 7.11097479e-01
1.01155162e-01 1.84097007e-01 -2.39964172e-01 -2.48538479e-01
2.44571760e-01 4.26227003e-01 -3.25938255e-01 2.04271480e-01
6.14332676e-01 9.30145979e-01 -9.28592384e-01 -2.80704349e-01
4.28015031e-02 2.05884859e-01 -6.41500950e-01 -2.69490778e-02
-3.94205332e-01 1.44063100e-01 8.08330253e-02 6.94602013e-01
5.50868452e-01 2.94745833e-01 7.01851368e-01 -8.23798701e-02
-1.75133213e-01 1.00318718e+00 -8.92603099e-01 1.63624716e+00
-8.80174994e-01 6.02793276e-01 -2.68057644e-01 -8.68704021e-01
6.59470677e-01 1.75958320e-01 -3.31153125e-02 -5.97899675e-01
9.48044360e-02 7.48969853e-01 5.20078421e-01 -1.37626782e-01
7.16059685e-01 -1.15379155e-01 -3.44725430e-01 8.03136766e-01
2.58551747e-01 -2.16773987e-01 6.89867079e-01 7.98273459e-02
7.73364067e-01 2.71172255e-01 3.86225551e-01 -2.37944856e-01
5.20300806e-01 7.59847388e-02 1.58786550e-01 5.20586014e-01
2.39558652e-01 2.26561010e-01 3.36318403e-01 -3.58533353e-01
-8.94884884e-01 -1.13683355e+00 -6.54005334e-02 1.68885601e+00
-4.91072327e-01 -8.33791196e-01 -9.74732041e-01 -1.11357558e+00
-1.64249837e-01 1.00808847e+00 -3.67828459e-01 1.36214755e-02
-1.52969396e+00 -9.59334493e-01 9.24508572e-01 5.84708929e-01
1.25957683e-01 -1.04418743e+00 -7.81627372e-02 2.21386388e-01
-2.32594982e-01 -1.03264201e+00 -4.78234977e-01 5.23066223e-01
-6.52971447e-01 -8.23460519e-01 -2.52895206e-01 -1.26873326e+00
6.78431690e-01 -2.05519348e-01 1.14083529e+00 2.06786975e-01
1.14439875e-01 -2.25657851e-01 -4.95602995e-01 -6.83466733e-01
-8.26557755e-01 5.63981354e-01 -2.97628224e-01 -3.71367097e-01
7.30545282e-01 -9.65957120e-02 2.09066600e-01 -6.74656034e-02
-9.06959236e-01 -1.23068683e-01 5.16222060e-01 7.39586890e-01
3.73922765e-01 -1.33631244e-01 3.99345487e-01 -1.40991843e+00
3.46883565e-01 -9.60468873e-02 -4.04999852e-01 5.59423387e-01
-4.34514374e-01 4.08859909e-01 7.70550370e-01 -1.86645731e-01
-1.10329247e+00 -2.44044326e-02 -6.10423088e-01 6.05302334e-01
1.16224483e-01 1.09455395e+00 -3.79797071e-01 -7.85709359e-03
6.95247650e-01 7.83907026e-02 -1.86963007e-01 -4.28037345e-01
7.62819052e-01 7.49219775e-01 3.24884325e-01 -8.08579028e-01
6.76867902e-01 1.32170334e-01 -2.27350652e-01 -8.73879075e-01
-1.07023060e+00 -8.47642124e-02 -1.02680457e+00 4.49334621e-01
3.73744398e-01 -9.73094404e-01 -1.53965995e-01 7.28420198e-01
-1.15650129e+00 -7.95737505e-01 -1.53625593e-01 8.40265304e-02
-2.00794101e-01 6.43310845e-01 -1.11324763e+00 -1.46575391e-01
-3.86508584e-01 -9.86171603e-01 6.65406883e-01 -3.60800326e-02
-2.21467584e-01 -1.32034743e+00 5.90831637e-02 2.01832131e-01
4.25866663e-01 -3.06160599e-01 1.38561559e+00 -6.74803615e-01
-4.41249847e-01 -1.57580346e-01 -1.86313108e-01 6.30740464e-01
4.77453083e-01 1.76151901e-01 -6.88941419e-01 -2.43320569e-01
-2.82777786e-01 -4.22388345e-01 5.71689487e-01 1.35420719e-02
6.05480373e-01 -4.03240800e-01 7.57092685e-02 4.47731942e-01
1.11143196e+00 -6.04765080e-02 7.34581590e-01 6.98407888e-01
8.15383613e-01 4.75827903e-01 7.42569625e-01 1.89267874e-01
8.14913571e-01 3.31978887e-01 9.18803737e-02 6.33170083e-02
-3.40452343e-01 -1.67595074e-01 8.35421741e-01 1.41343999e+00
1.27893865e-01 -2.31737599e-01 -1.16431999e+00 8.02168250e-01
-1.58994234e+00 -2.69074202e-01 -2.97838124e-03 1.81136513e+00
1.67259610e+00 3.50646138e-01 -1.91652864e-01 1.93122149e-01
3.46207649e-01 -8.60031396e-02 1.25805989e-01 -1.02381361e+00
-1.87473372e-01 9.07290697e-01 5.86845100e-01 9.54353213e-01
-1.10071480e+00 1.86048603e+00 6.45191336e+00 9.54826057e-01
-1.29778838e+00 7.91106969e-02 1.64483890e-01 3.70587587e-01
-3.80845666e-01 3.37511986e-01 -1.28538847e+00 1.45582110e-02
1.37951255e+00 4.56512719e-02 4.53178823e-01 3.38963836e-01
-5.70272282e-03 -2.78815385e-02 -1.11843050e+00 3.29075187e-01
3.57259333e-01 -1.05764341e+00 2.75180250e-01 -3.18764448e-01
7.64822960e-01 1.89958677e-01 3.99371758e-02 5.05517662e-01
6.46712899e-01 -1.15003741e+00 6.49467587e-01 2.34700844e-01
8.15975070e-01 -9.61353421e-01 8.70402694e-01 5.44706024e-02
-8.04631412e-01 1.49718642e-01 -2.95870453e-01 -4.25982028e-01
8.24802965e-02 3.02404612e-01 -1.23153186e+00 4.99982685e-01
1.56860813e-01 9.19138730e-01 -1.02265966e+00 6.58301890e-01
-8.61951590e-01 1.15999854e+00 -4.49664026e-01 -1.54303417e-01
4.69741195e-01 -1.04985118e-01 8.49975497e-02 1.70808911e+00
2.04370603e-01 -2.94408321e-01 1.54643208e-01 1.31200939e-01
-1.04599886e-01 4.78966057e-01 -1.76907644e-01 -2.55238980e-01
2.61792272e-01 9.20755863e-01 -6.03432059e-01 -2.90990174e-01
-5.17286897e-01 1.06438422e+00 6.60269976e-01 4.71513811e-03
-5.52186072e-01 -9.38180447e-01 3.94331247e-01 -7.46677443e-03
7.29300797e-01 -8.16278696e-01 -4.69984353e-01 -1.29773748e+00
8.30609873e-02 -1.27352107e+00 4.53693092e-01 -3.75194013e-01
-1.23302722e+00 5.79897106e-01 -2.27184609e-01 -7.65304208e-01
-5.83213449e-01 -1.10240364e+00 -4.11811560e-01 1.02152288e+00
-1.73579156e+00 -1.75099754e+00 4.49331224e-01 4.20661062e-01
4.34651673e-01 -1.48401290e-01 1.13707626e+00 2.16443524e-01
-3.62731844e-01 1.07582068e+00 3.96514595e-01 6.70355201e-01
1.12744224e+00 -1.50972462e+00 7.93987989e-01 1.24939072e+00
4.21447545e-01 1.00800931e+00 4.16920513e-01 -5.18008232e-01
-1.34787202e+00 -1.07127154e+00 1.79791009e+00 -7.87597656e-01
1.17062962e+00 -4.51677978e-01 -8.16427290e-01 1.17086995e+00
6.41138434e-01 -2.32148707e-01 7.87210047e-01 5.53559899e-01
-4.73202586e-01 -8.78737718e-02 -5.73674977e-01 7.67272651e-01
4.77900028e-01 -5.88466167e-01 -8.93675506e-01 2.95004755e-01
6.45821869e-01 -6.43662512e-01 -8.47675204e-01 9.42717344e-02
5.16594112e-01 -3.29840064e-01 7.12582350e-01 -8.16216469e-01
3.91697109e-01 -2.19550520e-01 -1.78509951e-01 -1.67942667e+00
8.12629908e-02 -5.94265461e-01 -1.04190998e-01 1.04943299e+00
1.02008796e+00 -2.70618796e-01 6.60766780e-01 1.29483759e-01
-5.08140147e-01 -4.47257906e-01 -5.63499272e-01 -7.72756398e-01
9.01290894e-01 -3.79793823e-01 4.34294373e-01 8.08709502e-01
5.74194670e-01 7.00223923e-01 -6.05249032e-02 6.15094230e-02
5.82370833e-02 -3.32304351e-02 5.28334677e-01 -6.58814192e-01
-2.36582935e-01 -9.34993774e-02 -5.83194867e-02 -1.10345221e+00
4.82144475e-01 -1.19943082e+00 1.41993508e-01 -1.38907743e+00
-3.15512747e-01 -6.13351464e-01 -1.48061156e-01 1.06642103e+00
-3.21936429e-01 3.64017874e-01 3.55638236e-01 -2.38912225e-01
-5.94725549e-01 2.58795857e-01 8.98288429e-01 -1.65499732e-01
-3.09849143e-01 -2.39458606e-01 -5.92032313e-01 7.06088960e-01
9.58631456e-01 -5.53807676e-01 -1.68592304e-01 -1.17592275e+00
4.77325648e-01 -4.66941476e-01 -4.84145492e-01 -7.00569808e-01
-2.45941356e-02 -2.35144660e-01 2.45986938e-01 -4.70602900e-01
6.15199767e-02 -3.91103834e-01 -8.80823493e-01 4.45460916e-01
-1.62158608e-01 5.50046563e-01 7.16319442e-01 -2.51146138e-01
-2.40744799e-01 -6.85035169e-01 6.10430360e-01 -2.04558611e-01
-7.05693603e-01 -2.34304175e-01 -8.06183398e-01 4.15402412e-01
4.75908428e-01 2.19736770e-01 -3.93563688e-01 -2.61241626e-02
-6.36149764e-01 -1.27265215e-01 4.18252289e-01 3.21936607e-01
3.51665646e-01 -6.73117995e-01 -9.10175860e-01 3.99607390e-01
3.91911902e-02 -2.61949420e-01 -4.06455904e-01 5.44820309e-01
-1.14755559e+00 6.89609349e-01 -3.39562893e-01 -1.71909064e-01
-1.14532411e+00 1.58249125e-01 2.76660413e-01 -6.87088847e-01
-1.14815481e-01 7.54469991e-01 -4.47659254e-01 -9.45641100e-01
-3.12093049e-02 -4.77322698e-01 -7.04905018e-02 -3.05576739e-03
4.63781297e-01 -7.65511720e-03 8.92124474e-01 -5.51468611e-01
-3.86729896e-01 7.95890093e-02 -6.91485584e-01 -2.61101127e-01
1.27336669e+00 -9.57966596e-02 -5.18295467e-01 3.72941911e-01
9.79185283e-01 5.71751237e-01 -8.13608110e-01 -2.33825892e-01
5.06592393e-01 -8.48178640e-02 -2.02138081e-01 -1.16246462e+00
-5.10830820e-01 1.16201532e+00 1.08249612e-01 -2.96924412e-01
7.94633865e-01 -3.63513678e-01 1.24426532e+00 1.01406705e+00
2.09932312e-01 -1.30496752e+00 9.06096697e-02 1.60300446e+00
4.75479633e-01 -1.06633806e+00 -2.18577206e-01 -5.01457274e-01
-6.33550286e-01 1.32284594e+00 5.55307090e-01 -3.51426661e-01
8.59293580e-01 4.77191210e-01 5.18796444e-01 2.54751295e-01
-5.33529758e-01 -1.63621023e-01 3.35160792e-01 6.24536157e-01
1.03212702e+00 1.23703800e-01 -5.74715257e-01 7.28608191e-01
-6.06146216e-01 -3.63745004e-01 7.39699841e-01 1.46573651e+00
-3.87185425e-01 -1.88327765e+00 -2.15027764e-01 3.06662470e-01
-6.87796175e-01 -9.52847123e-01 -3.10492843e-01 7.78717637e-01
1.13126732e-01 8.50200772e-01 -7.08963424e-02 -1.28770456e-01
3.40389907e-01 3.63790154e-01 9.92873311e-01 -1.06648576e+00
-7.52515852e-01 -6.66018501e-02 4.30330753e-01 -2.62749735e-02
-1.52112901e-01 -7.48480201e-01 -1.51559365e+00 -3.07891637e-01
-2.62370437e-01 9.78714302e-02 6.27454758e-01 1.41154671e+00
1.06390104e-01 2.43967265e-01 2.25282028e-01 -4.86720622e-01
-6.41229987e-01 -1.14080417e+00 -4.54358399e-01 -2.12607048e-02
1.66415840e-01 -4.36009577e-04 -7.04162344e-02 3.59023392e-01] | [10.768723487854004, 10.244183540344238] |
1a6db66a-21b9-4dce-a4d3-2a33fc305c6a | applying-automated-machine-translation-to | 2301.03141 | null | https://arxiv.org/abs/2301.03141v1 | https://arxiv.org/pdf/2301.03141v1.pdf | Applying Automated Machine Translation to Educational Video Courses | We studied the capability of automated machine translation in the online video education space by automatically translating Khan Academy videos with state of the art translation models and applying Text-to-Speech synthesis to build engaging videos in target languages. We also analyzed and established a reliable translation confidence estimator based on round-trip translations in order to efficiently manage translation quality and reduce human translation effort. Finally, we developed a deployable system to deliver translated videos to end users and collect user corrections for iterative improvement. | ['Linden Wang'] | 2023-01-09 | null | null | null | null | ['text-to-speech-synthesis'] | ['speech'] | [ 8.03626105e-02 1.40519366e-01 -3.11188251e-01 -2.54720390e-01
-1.65370822e+00 -8.71683002e-01 4.11164284e-01 -1.98260441e-01
-1.63146839e-01 9.02992666e-01 3.85036618e-01 -6.84697270e-01
9.70382616e-02 -3.99170697e-01 -1.18077302e+00 8.05978701e-02
3.63168299e-01 5.43396652e-01 -1.19077131e-01 -4.84846652e-01
3.23329270e-01 2.04686731e-01 -1.30935752e+00 5.66075206e-01
1.42259037e+00 9.09059867e-02 2.62249976e-01 1.34253168e+00
2.00127214e-01 7.95883715e-01 -9.10301983e-01 -1.21976042e+00
2.68744975e-01 -1.04739904e+00 -6.81184053e-01 2.37308536e-02
9.78431284e-01 -6.75996423e-01 -4.35719669e-01 1.13701689e+00
7.73339808e-01 8.64023715e-02 1.16268434e-01 -1.04103398e+00
-9.99569774e-01 9.05627549e-01 2.32929528e-01 3.30009311e-01
1.27832770e+00 1.72494963e-01 4.24044371e-01 -1.03483987e+00
1.08231330e+00 1.02672982e+00 5.89108467e-01 3.09575558e-01
-7.08994865e-01 -9.91825044e-01 -5.36153376e-01 5.01896322e-01
-1.18158746e+00 -7.01778352e-01 3.34252834e-01 -4.34593201e-01
8.64336967e-01 3.69994253e-01 1.14799142e+00 1.39413071e+00
8.40191662e-01 4.95473802e-01 1.04804111e+00 -7.54938185e-01
-1.74269810e-01 2.66985387e-01 -7.61607528e-01 1.15683913e+00
-1.60791799e-01 -1.25062987e-01 -1.12343466e+00 2.37357721e-01
8.29833984e-01 -5.50135911e-01 4.63598259e-02 4.46015924e-01
-1.51293564e+00 6.04277432e-01 -5.18361926e-01 -6.57711774e-02
-1.53390110e-01 -1.18282475e-01 2.50543088e-01 1.10516250e+00
6.91692412e-01 6.39425457e-01 -4.41099256e-01 -1.10185277e+00
-1.33594525e+00 1.25989377e-01 9.11954880e-01 1.69884944e+00
4.66204286e-01 2.44068857e-02 -2.81391591e-01 5.87590396e-01
2.46252352e-03 8.55453670e-01 7.23394573e-01 -1.37546158e+00
7.72565722e-01 3.24196219e-01 -2.27001775e-02 -8.11334491e-01
2.59703994e-01 -4.33443785e-01 7.83719346e-02 -3.38055879e-01
1.74338698e-01 -3.94123435e-01 -6.83233500e-01 1.07149172e+00
1.29858479e-01 2.77527183e-01 -7.52489865e-02 7.66406953e-01
7.77009726e-01 5.93361616e-01 -4.79384989e-01 -5.14602721e-01
9.20941830e-01 -1.38615847e+00 -1.12700677e+00 1.94795832e-01
1.10302973e+00 -1.91149330e+00 1.16085243e+00 4.41809118e-01
-1.65300596e+00 -6.82050228e-01 -8.07863235e-01 -1.89392135e-01
1.31931782e-01 5.79110861e-01 9.63019729e-02 8.59290421e-01
-1.48413742e+00 8.14317226e-01 -9.35749650e-01 -3.18958849e-01
8.07745606e-02 4.26786333e-01 -5.70161581e-01 -1.51387751e-01
-1.03645837e+00 1.15302980e+00 -1.97024435e-01 -1.81132331e-01
-9.87568080e-01 -7.37056792e-01 -9.03695643e-01 -3.17961663e-01
-2.29981467e-02 -5.05148947e-01 1.77341080e+00 -1.29540265e+00
-2.32038760e+00 7.20891714e-01 -1.56443134e-01 -9.33112875e-02
8.92359376e-01 -4.64744806e-01 -3.46806884e-01 3.59017372e-01
2.03243300e-01 4.03079480e-01 6.41392171e-01 -2.74488300e-01
-6.47008061e-01 4.48164754e-02 -2.84175038e-01 4.83185947e-01
-3.69817704e-01 4.74660575e-01 -5.14621377e-01 -6.77916169e-01
-9.34641510e-02 -1.16900456e+00 9.21255071e-03 -4.28694814e-01
2.06067935e-01 1.77710980e-01 5.03280640e-01 -1.64367807e+00
1.25579953e+00 -1.53763509e+00 4.87500131e-01 -6.04032911e-02
-1.62011623e-01 2.09860235e-01 -4.20783609e-01 6.35531604e-01
2.78211802e-01 -5.21877818e-02 5.59962928e-01 -2.83995688e-01
-2.11217895e-01 3.83512513e-03 -1.21499285e-01 5.21002173e-01
1.33856148e-01 7.66053557e-01 -1.09512901e+00 -9.06061530e-01
2.36883402e-01 3.72719258e-01 -9.42527831e-01 5.58236599e-01
2.15796605e-01 4.16405976e-01 -1.34774819e-01 8.76206636e-01
5.59040233e-02 4.51062232e-01 8.38558450e-02 3.36651057e-01
-2.65615404e-01 4.62802857e-01 -9.22472477e-01 2.13847136e+00
-7.73158073e-01 1.37891722e+00 -1.90941453e-01 -4.82472450e-01
9.71700311e-01 7.11153746e-01 4.26999897e-01 -7.28556216e-01
2.46147871e-01 4.38893229e-01 -1.66461468e-01 -8.98076773e-01
9.55609918e-01 3.93274605e-01 1.47368237e-01 3.98158431e-01
7.84819484e-01 -5.44232249e-01 3.81899029e-01 6.34921789e-02
8.93676579e-01 8.19554210e-01 -8.98154378e-02 -6.78304881e-02
2.12400749e-01 3.85187119e-01 2.01897264e-01 4.75219637e-01
-3.31055760e-01 6.31035805e-01 1.01930067e-01 -2.73707002e-01
-1.31200421e+00 -7.86494613e-01 4.35438216e-01 1.43777895e+00
-3.37167025e-01 -6.86444759e-01 -1.23581290e+00 -5.06861806e-01
-8.30020845e-01 6.72834098e-01 7.92120397e-02 -1.62725523e-01
-6.36824608e-01 2.98428714e-01 6.20639682e-01 1.86431065e-01
-1.08417068e-02 -6.40256047e-01 -1.45911068e-01 4.67500657e-01
-7.99483955e-01 -1.17952228e+00 -1.07643569e+00 -5.53803563e-01
-8.19596827e-01 -7.35748827e-01 -7.01326907e-01 -1.31737626e+00
5.05440533e-01 1.71799853e-01 1.05648196e+00 -1.35691419e-01
-8.15904960e-02 4.64361489e-01 -5.06795406e-01 -4.42433476e-01
-1.08974349e+00 2.06413686e-01 4.34835702e-01 -7.42293835e-01
3.04944515e-01 -2.60843486e-01 -1.33651838e-01 4.43025887e-01
-4.56208915e-01 4.86361474e-01 6.41788304e-01 5.91538906e-01
3.23852479e-01 -4.31845605e-01 2.67699301e-01 -4.59390789e-01
9.34541047e-01 3.15530971e-02 -6.89600706e-01 4.70085919e-01
-4.21131730e-01 -3.38232517e-01 6.26926363e-01 -6.84472859e-01
-8.28783751e-01 4.23337407e-02 -1.38257310e-01 -6.62025034e-01
3.56188595e-01 4.31323171e-01 4.16734666e-01 -3.94055009e-01
6.69320643e-01 2.11144045e-01 -5.38061783e-02 7.26611242e-02
3.82921547e-01 1.05623138e+00 5.50379515e-01 -5.67479968e-01
6.02883935e-01 -4.70120996e-01 -3.54708314e-01 -8.45350504e-01
-3.65643144e-01 -1.76285535e-01 -7.84912765e-01 -7.59697974e-01
7.77138948e-01 -1.36034703e+00 -6.21626258e-01 5.93018271e-02
-1.29595160e+00 -2.36569569e-01 -1.07500358e-02 1.15804470e+00
-1.07904840e+00 9.94455367e-02 -1.12352788e+00 -3.99383992e-01
-2.80435503e-01 -1.63760662e+00 8.79896760e-01 3.16642553e-01
-4.87554103e-01 -7.95830011e-01 3.72172654e-01 1.06061268e+00
1.08712003e-01 -3.10452521e-01 1.59704253e-01 -4.94116038e-01
-5.38654387e-01 -2.59251058e-01 2.97294110e-01 2.44996220e-01
-5.31418696e-02 7.20616758e-01 -1.70383319e-01 -3.08181971e-01
-2.61408240e-01 -2.61486739e-01 -2.16156974e-01 2.06203461e-01
3.72611761e-01 -7.15770125e-01 2.82458007e-01 3.91089052e-01
7.65264213e-01 -1.21986367e-01 4.24798936e-01 2.34338328e-01
4.80610281e-01 3.90974313e-01 1.06439984e+00 1.50600910e-01
2.66347110e-01 7.13562012e-01 -4.41477895e-01 2.55207449e-01
-4.01990235e-01 -8.85005295e-01 1.02817309e+00 2.25098181e+00
-4.39385802e-01 -3.32185179e-02 -8.55862737e-01 6.33047044e-01
-1.63661373e+00 -1.01955628e+00 -1.85353070e-01 1.91374469e+00
1.12403905e+00 -2.96784419e-04 7.88192526e-02 -4.68626738e-01
6.25575423e-01 -7.21126616e-01 4.36853826e-01 -9.58054066e-01
3.41905266e-01 5.38860857e-01 7.42258072e-01 7.69998431e-01
-4.93467003e-01 1.29288316e+00 7.95264912e+00 7.78200150e-01
-8.93980682e-01 5.86517811e-01 3.05341065e-01 -2.17616171e-01
-1.47779435e-01 -1.25980571e-01 -6.84682310e-01 3.34050328e-01
1.71278620e+00 -4.66901630e-01 6.26361907e-01 9.15004313e-01
7.21421599e-01 3.09302151e-01 -9.16748226e-01 7.76121616e-01
3.60099435e-01 -1.65014052e+00 1.02237977e-01 -5.12356963e-03
1.45713127e+00 -4.36030924e-02 -1.34311885e-01 3.60403717e-01
4.06380802e-01 -6.77715480e-01 8.62016916e-01 4.80397522e-01
8.17028105e-01 -8.50993216e-01 7.94078469e-01 2.86245257e-01
-8.29306126e-01 2.95429677e-01 -3.73509973e-01 -3.43502700e-01
-2.01743349e-01 -6.92192614e-02 -1.36804485e+00 4.95744377e-01
2.58336633e-01 6.47831798e-01 -6.73902392e-01 8.83191943e-01
-4.65857267e-01 9.49021041e-01 7.52733201e-02 -1.17142297e-01
9.08342898e-02 -6.13667369e-01 5.18195868e-01 1.33241653e+00
1.10270107e+00 -2.32736655e-02 2.65075952e-01 3.44301820e-01
-2.00135693e-01 4.83235508e-01 -6.79883003e-01 -3.69716257e-01
5.07768273e-01 9.89453793e-01 -4.52074438e-01 -4.26084220e-01
-3.24562937e-01 1.14650857e+00 -8.11344832e-02 2.81333804e-01
-1.05742991e+00 -3.85192394e-01 6.13143921e-01 1.85537949e-01
-1.62424147e-01 -5.59976339e-01 -1.12106405e-01 -1.37308133e+00
-1.47340029e-01 -1.49601543e+00 -1.81650788e-01 -9.58904326e-01
-4.85437125e-01 4.32011992e-01 -2.33837888e-01 -1.71189296e+00
-5.50744176e-01 -2.50061631e-01 -3.01087379e-01 6.73467100e-01
-8.10492456e-01 -1.08807921e+00 8.10876787e-02 5.37555218e-01
1.18026888e+00 -4.27387118e-01 7.84362793e-01 6.18293703e-01
-3.99891526e-01 1.04817069e+00 8.80380049e-02 3.46761383e-02
1.31568849e+00 -6.27695858e-01 3.50869447e-01 1.24013209e+00
3.78672779e-01 5.35993576e-01 1.15159023e+00 -9.92391109e-01
-1.86519599e+00 -9.05406058e-01 1.33609855e+00 -8.45165849e-01
7.73440480e-01 -3.18591036e-02 -3.33700269e-01 9.05685246e-01
6.87591255e-01 -6.02784991e-01 8.11842322e-01 -3.14363807e-01
2.20787853e-01 1.80346578e-01 -1.01119447e+00 8.60673249e-01
1.07806087e+00 -7.84713089e-01 -5.98915994e-01 7.46756256e-01
9.86288488e-01 -1.15392303e+00 -1.25832629e+00 -1.56488910e-01
6.17690861e-01 -2.96236575e-01 5.09112000e-01 -7.00642228e-01
1.26099062e+00 4.52753790e-02 -8.26246198e-03 -1.50766838e+00
-6.57586977e-02 -1.44394755e+00 1.61570370e-01 1.03282213e+00
6.00460112e-01 -6.76237186e-03 8.23677778e-01 5.35121977e-01
-6.29378200e-01 -3.10368478e-01 -9.99791861e-01 -5.90798378e-01
-1.28350362e-01 -3.03897530e-01 2.10602313e-01 1.07328761e+00
5.60307503e-01 1.61364749e-01 -6.00239336e-01 9.68521386e-02
2.54421264e-01 -2.82162428e-01 9.36814010e-01 -3.53951454e-01
-3.40536922e-01 -2.73916423e-01 -6.71475589e-01 -5.58542550e-01
1.78679571e-01 -8.68670583e-01 1.38265252e-01 -1.49788988e+00
1.04555935e-01 4.38080311e-01 3.98923814e-01 1.80205867e-01
-1.65160656e-01 8.23500276e-01 1.86456859e-01 -3.61978933e-02
-6.13232791e-01 1.70746848e-01 1.72157466e+00 -5.53650521e-02
-2.92259544e-01 9.95500684e-02 -1.87074900e-01 3.49382132e-01
4.72566009e-01 -5.63168943e-01 -3.64575028e-01 -6.23509705e-01
4.37417805e-01 5.80511332e-01 -2.63010383e-01 -9.85082865e-01
2.55541086e-01 -5.16507149e-01 3.33083272e-01 -3.32580924e-01
-1.24538787e-01 -6.40728593e-01 3.00811261e-01 4.48785216e-01
-6.37092769e-01 7.91350067e-01 2.47314513e-01 -1.51086852e-01
-3.08221906e-01 -8.93939510e-02 3.82216156e-01 6.41220957e-02
-1.55906528e-01 -2.18695283e-01 -8.32381487e-01 -9.62307453e-02
1.09536493e+00 -1.81072474e-01 -3.11436057e-01 -8.46186519e-01
-5.96624315e-01 -3.10078301e-02 5.20621538e-01 6.93139136e-01
6.97471857e-01 -1.45740986e+00 -1.09768736e+00 4.37825471e-01
-1.96191341e-01 -7.82853723e-01 -6.77826852e-02 8.96439314e-01
-1.41808724e+00 5.00740945e-01 -6.63125753e-01 -5.86495876e-01
-1.75881040e+00 -3.74872424e-02 -2.08565909e-02 7.42208511e-02
-2.84336686e-01 8.51441324e-01 -9.34672415e-01 -6.57867372e-01
8.39798748e-02 -5.10716021e-01 3.21975738e-01 -3.33851933e-01
6.29651427e-01 6.04342520e-01 5.19836605e-01 -6.95371807e-01
7.31633604e-02 2.29679152e-01 2.61112928e-01 -4.82204646e-01
1.02568364e+00 -2.80859530e-01 2.61529863e-01 2.53679842e-01
9.72790718e-01 4.14551526e-01 -8.78500402e-01 1.94714844e-01
-4.75622356e-01 -8.30774665e-01 -5.15878797e-02 -7.29687095e-01
-5.33017159e-01 4.98742461e-01 5.39883018e-01 -5.29396772e-01
9.09393132e-01 -4.93552715e-01 8.96179020e-01 4.99317676e-01
4.55773264e-01 -1.52440572e+00 5.84210176e-03 4.17466193e-01
4.94651318e-01 -1.19588566e+00 2.35698193e-01 -2.19246179e-01
-7.16298997e-01 1.35735798e+00 3.60922754e-01 9.73408893e-02
6.35998994e-02 1.84091642e-01 4.13472593e-01 4.43528593e-01
-9.66572404e-01 4.73915368e-01 6.15497410e-01 5.23574650e-01
8.82387042e-01 3.62749159e-01 -6.29695714e-01 1.37978420e-01
-9.89516139e-01 4.95760679e-01 1.19746888e+00 8.04705679e-01
-3.64767045e-01 -1.24900854e+00 -4.61748183e-01 1.12695806e-01
-6.04399443e-01 -5.56865670e-02 -5.71019888e-01 3.73811305e-01
-5.03358655e-02 1.20498264e+00 -1.87363967e-01 -6.77095532e-01
4.30837572e-01 1.57021970e-01 9.75325525e-01 -6.40244484e-01
-1.10723913e+00 1.34990364e-01 3.29312503e-01 -7.06799090e-01
-3.41916025e-01 -5.72219133e-01 -6.05884135e-01 -8.13559830e-01
-4.15867925e-01 2.89207011e-01 1.12611926e+00 1.09263813e+00
4.86826420e-01 5.54544091e-01 5.39599836e-01 -6.89368963e-01
-4.07860875e-01 -1.10377276e+00 2.74635762e-01 2.36925613e-02
3.44382152e-02 6.17519692e-02 2.43171006e-01 5.76403975e-01] | [14.48471450805664, 7.17218017578125] |
fa357f42-95f3-494d-9a99-57f7870868f7 | successive-projection-algorithm-robust-to | 1908.04109 | null | https://arxiv.org/abs/1908.04109v1 | https://arxiv.org/pdf/1908.04109v1.pdf | Successive Projection Algorithm Robust to Outliers | The successive projection algorithm (SPA) is a fast algorithm to tackle separable nonnegative matrix factorization (NMF). Given a nonnegative data matrix $X$, SPA identifies an index set $\mathcal{K}$ such that there exists a nonnegative matrix $H$ with $X \approx X(:,\mathcal{K})H$. SPA has been successfully used as a pure-pixel search algorithm in hyperspectral unmixing and for anchor word selection in document classification. Moreover, SPA is provably robust in low-noise settings. The main drawbacks of SPA are that it is not robust to outliers and does not take the data fitting term into account when selecting the indices in $\mathcal{K}$. In this paper, we propose a new SPA variant, dubbed Robust SPA (RSPA), that is robust to outliers while still being provably robust in low-noise settings, and that takes into account the reconstruction error for selecting the indices in $\mathcal{K}$. We illustrate the effectiveness of RSPA on synthetic data sets and hyperspectral images. | ['Nicolas Gillis'] | 2019-08-12 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 6.98912740e-01 -5.15521288e-01 1.82024449e-01 3.09571717e-03
-8.71647894e-01 -6.46699429e-01 1.25849783e-01 -7.36546190e-03
-3.99206519e-01 5.32805979e-01 -1.59577176e-01 -4.49246705e-01
-1.03612614e+00 -8.45829070e-01 -3.67885172e-01 -1.31640387e+00
-5.47213964e-02 2.59677291e-01 -4.41867262e-01 -8.50025266e-02
9.29057747e-02 6.54138863e-01 -1.67945921e+00 -1.56350583e-02
1.15459609e+00 1.16140187e+00 3.79278332e-01 5.24996042e-01
4.98977304e-03 5.55829704e-01 -4.27345842e-01 -2.20481604e-01
7.79965460e-01 -5.08946478e-01 -5.17325640e-01 6.07156217e-01
2.96806097e-01 2.90833503e-01 1.00324340e-01 1.60591567e+00
4.18677777e-01 3.25104266e-01 6.10260129e-01 -1.30381656e+00
-4.19257402e-01 5.00278234e-01 -1.04225743e+00 -6.62818030e-02
4.49135900e-02 -6.88771904e-02 1.06812859e+00 -1.18314743e+00
5.35902739e-01 8.17386806e-01 6.88808620e-01 6.04382157e-02
-1.31638551e+00 -7.66505301e-01 9.69122276e-02 8.84943232e-02
-1.62923861e+00 -3.45602423e-01 7.56203532e-01 -6.14058495e-01
5.48886001e-01 9.54674125e-01 5.86067379e-01 1.63386375e-01
-2.78622180e-01 4.99552816e-01 1.35457742e+00 -6.51741207e-01
5.36423862e-01 -1.99189276e-01 3.48514587e-01 3.69230151e-01
4.17770803e-01 -8.33719075e-02 -4.94944662e-01 -5.33974349e-01
3.44799608e-01 1.30623236e-01 -7.20046580e-01 -4.62820411e-01
-1.09057581e+00 9.19554591e-01 6.10900670e-02 8.69061500e-02
-7.42066026e-01 -2.99323052e-01 3.92160757e-04 3.09600294e-01
2.99082965e-01 3.23373675e-01 -2.05781639e-01 3.19247752e-01
-1.26184928e+00 -2.04650848e-03 3.22102875e-01 7.92701483e-01
9.47167754e-01 2.84030795e-01 1.97312478e-02 9.82617915e-01
2.72420883e-01 9.54485476e-01 2.51952261e-01 -1.10289192e+00
6.38631523e-01 6.80842876e-01 3.74727488e-01 -1.15821338e+00
-5.58687270e-01 -5.06212533e-01 -1.22842836e+00 2.86987573e-01
2.11380512e-01 1.52588859e-02 -6.77825034e-01 1.63318074e+00
1.85330540e-01 1.12167679e-01 4.15771276e-01 9.49873626e-01
5.33570468e-01 7.41811693e-01 -3.76085728e-01 -1.02694297e+00
1.03272152e+00 -5.09034634e-01 -5.73313177e-01 -3.07212472e-01
2.42095336e-01 -9.31213915e-01 9.99876976e-01 7.77294874e-01
-9.98850644e-01 -1.53595313e-01 -1.08630705e+00 4.38513488e-01
-1.51917458e-01 4.38916922e-01 4.93023723e-01 9.31999385e-01
-9.98383582e-01 4.28239584e-01 -5.30085325e-01 -1.17662407e-01
-4.83963033e-03 6.20634258e-01 -4.58391219e-01 -1.93041459e-01
-8.20639193e-01 3.44350517e-01 4.01038677e-01 5.37300408e-01
-3.96757752e-01 -5.11651456e-01 -6.84545934e-01 -4.14638706e-02
5.01475513e-01 -3.02134722e-01 4.87502068e-01 -1.08308101e+00
-1.16565800e+00 7.97768295e-01 -3.89614552e-01 -3.08097512e-01
2.77413189e-01 7.08605051e-02 -6.07517600e-01 4.17463809e-01
9.12257358e-02 1.06242254e-01 1.26855600e+00 -1.27581728e+00
-3.18754435e-01 -7.60105073e-01 -3.05907071e-01 1.46604136e-01
-5.07461429e-01 9.31395888e-02 -3.30036789e-01 -7.73939967e-01
1.17735684e+00 -1.22298646e+00 -4.98297215e-01 -3.25282365e-01
-5.63006878e-01 2.33920097e-01 5.76636791e-01 -6.76026344e-01
1.24808693e+00 -2.25103354e+00 4.62739319e-01 9.59376276e-01
7.56163597e-02 3.67069840e-01 -1.56378329e-01 4.69841778e-01
-4.96557564e-01 -8.19376856e-02 -8.51768732e-01 -2.32405990e-01
-1.32532522e-01 7.59104788e-02 -2.93535739e-01 8.86515498e-01
-2.21192956e-01 1.76932916e-01 -7.88090110e-01 -3.41434404e-02
2.62953281e-01 3.20487827e-01 -2.35641614e-01 -1.77722588e-01
-1.75084114e-01 1.88067943e-01 -5.14971167e-02 8.42171848e-01
1.15047038e+00 -1.79360718e-01 1.20379075e-01 -2.76448220e-01
-5.71479440e-01 -5.26613891e-01 -1.95927334e+00 1.26805246e+00
1.89838812e-01 2.68310159e-01 6.54141843e-01 -1.04334378e+00
1.00521290e+00 2.36303583e-01 1.05733645e+00 -4.22338754e-01
3.79475839e-02 4.23823655e-01 -1.21599019e-01 -2.78075218e-01
6.33258700e-01 -3.55881870e-01 4.11980957e-01 6.22965276e-01
-3.60158622e-01 -2.93529600e-01 5.30985177e-01 2.04559103e-01
8.76286566e-01 -4.23434824e-01 1.74525857e-01 -5.66506386e-01
9.89622951e-01 -3.07822395e-02 6.90759182e-01 5.95761597e-01
-1.97380483e-01 8.39527965e-01 1.97107911e-01 -5.10261804e-02
-7.35208750e-01 -7.86579072e-01 -1.88026220e-01 8.13583970e-01
5.46956845e-02 -4.99140948e-01 -5.35571992e-01 -1.53675839e-01
-1.70607448e-01 4.54604834e-01 -3.59214962e-01 1.28594771e-01
-1.32571995e-01 -1.51511812e+00 4.93526533e-02 1.08158059e-01
4.27071601e-01 -7.75575459e-01 -2.50740200e-01 1.27070457e-01
-3.42303634e-01 -9.13258493e-01 -2.48858377e-01 5.35040736e-01
-8.56465220e-01 -1.16114652e+00 -5.78020155e-01 -3.90598953e-01
1.04616570e+00 8.32832217e-01 7.03609109e-01 -2.45220289e-01
-7.21064806e-02 4.50864762e-01 -4.48485911e-01 -3.01721722e-01
-1.42089233e-01 -3.26859266e-01 1.76854938e-01 5.12682378e-01
1.92410678e-01 -4.95131940e-01 -5.47004521e-01 3.83774936e-01
-1.33995700e+00 -1.42586261e-01 2.79786289e-01 7.85841107e-01
1.17859221e+00 7.21256733e-01 3.73301245e-02 -9.43544090e-01
4.21026707e-01 -1.69794768e-01 -8.38829994e-01 3.50201517e-01
-7.02782452e-01 -1.38111264e-01 6.28613710e-01 -6.23082183e-02
-6.00082874e-01 4.55863595e-01 3.74351069e-02 -5.15342653e-01
3.13406289e-01 7.85338819e-01 -3.17731559e-01 -2.56960273e-01
8.87938917e-01 4.68360633e-01 3.24212834e-02 -5.25432289e-01
3.11524451e-01 4.95350569e-01 5.19552410e-01 -4.53718543e-01
1.01518130e+00 9.15227354e-01 3.96704495e-01 -1.25473297e+00
-7.18591511e-01 -8.75876248e-01 -5.79339743e-01 -8.13586116e-02
3.73321384e-01 -9.67807591e-01 -4.85510796e-01 3.72913241e-01
-5.14298201e-01 7.97244012e-02 -3.83114785e-01 5.50832927e-01
-5.37098527e-01 6.94830954e-01 -1.46928117e-01 -1.13796711e+00
-5.61078966e-01 -1.21427464e+00 6.25769913e-01 -1.48279697e-01
8.82929489e-02 -5.02929926e-01 -1.30267069e-01 4.75349456e-01
7.34025091e-02 2.09855959e-01 8.64358902e-01 -1.47040024e-01
-3.11544210e-01 -2.75338531e-01 -3.54090482e-01 7.65475690e-01
9.55203250e-02 1.90127566e-01 -7.53016591e-01 -6.74002588e-01
3.29792678e-01 1.80954456e-01 9.55694377e-01 5.84752679e-01
1.14252329e+00 -3.82103056e-01 -3.92620154e-02 7.38902211e-01
1.64978123e+00 2.07301855e-01 7.47527599e-01 3.48538786e-01
5.54745674e-01 5.65698862e-01 5.54602861e-01 6.56618357e-01
-1.00302525e-01 2.44410008e-01 6.15978539e-01 -1.61198556e-01
4.19332355e-01 5.90094030e-01 3.57016146e-01 7.48267114e-01
-2.47126237e-01 -2.00473756e-01 -7.80892670e-01 4.07967359e-01
-1.75286269e+00 -1.01234567e+00 -5.26827574e-01 2.55825353e+00
5.24887681e-01 -2.64738649e-01 4.35256772e-02 8.69377553e-01
7.90763199e-01 2.44352669e-01 -4.97987121e-01 2.95457747e-02
-8.32021952e-01 4.46062922e-01 8.92921865e-01 5.61105728e-01
-1.14022422e+00 4.08497274e-01 4.85262203e+00 6.59749329e-01
-1.04500473e+00 1.77927185e-02 1.95355937e-01 -3.98479998e-02
-4.45702434e-01 7.73575380e-02 -4.87565488e-01 3.40237916e-01
5.82624912e-01 2.08416536e-01 4.90453184e-01 4.84224558e-01
4.31665152e-01 -3.02895635e-01 -4.22450423e-01 1.29232264e+00
2.45060116e-01 -1.07468510e+00 -9.02906582e-02 7.14712543e-03
1.02283061e+00 -6.35973960e-02 3.35404217e-01 -4.55743045e-01
1.32847711e-01 -8.33790839e-01 7.05918074e-01 3.11339080e-01
5.82009494e-01 -8.75790656e-01 4.35707808e-01 3.67457926e-01
-1.13494658e+00 -3.59390408e-01 -4.89059597e-01 1.41596094e-01
-6.98072314e-02 1.10249007e+00 -3.63233447e-01 8.47098291e-01
6.76201761e-01 6.31183505e-01 -3.82339209e-01 1.01067924e+00
-1.39059946e-01 4.35034126e-01 -5.36192358e-01 4.28372443e-01
2.89818436e-01 -1.05719459e+00 7.16836810e-01 8.54964375e-01
8.44523549e-01 4.67707485e-01 3.74940723e-01 4.87819791e-01
2.08339170e-01 5.88874161e-01 -3.32710981e-01 -3.20161164e-01
3.23001385e-01 1.01597643e+00 -9.28471625e-01 -3.66555527e-02
-2.07885072e-01 8.58433723e-01 -1.28676504e-01 5.42588413e-01
-3.93231124e-01 -9.61371586e-02 6.27833307e-01 5.68971522e-02
3.20259035e-01 -4.20579225e-01 -4.17092413e-01 -1.18671405e+00
2.01495081e-01 -1.01578820e+00 8.00815761e-01 -7.35654593e-01
-1.09389424e+00 4.90049362e-01 -3.25420946e-01 -1.56513774e+00
3.55431080e-01 -6.38589442e-01 5.63197322e-02 8.84343863e-01
-1.43005621e+00 -8.03935707e-01 -2.46987462e-01 8.40272248e-01
1.52109880e-02 -1.43084541e-01 9.86105084e-01 2.00249523e-01
-8.95218074e-01 1.81186661e-01 7.40658939e-01 -2.65361965e-01
3.42637867e-01 -1.04952514e+00 -4.34708595e-01 1.38531876e+00
1.81699082e-01 7.49420166e-01 8.38385403e-01 -3.86354566e-01
-1.54304731e+00 -1.17353690e+00 5.61741233e-01 7.85002261e-02
5.68470836e-01 8.18028897e-02 -6.23545885e-01 4.10409153e-01
-6.53915480e-02 -7.66004100e-02 1.21733212e+00 -1.11254998e-01
-3.84298444e-01 -4.02571321e-01 -1.28277051e+00 4.37530577e-01
6.16333604e-01 -4.49132979e-01 4.39281762e-02 8.44876885e-01
2.14177385e-01 -2.87700474e-01 -8.67423117e-01 6.23232663e-01
2.29634359e-01 -1.13192677e+00 1.23626733e+00 -1.28607288e-01
-5.08138165e-02 -7.61126757e-01 -5.31564951e-01 -1.01444650e+00
-3.33723217e-01 -8.40759814e-01 7.44684860e-02 7.62681723e-01
5.52096963e-01 -6.95112348e-01 8.71482432e-01 5.63466787e-01
-4.20366116e-02 -2.04769179e-01 -1.04620528e+00 -9.24506187e-01
-1.03001647e-01 -6.46410882e-01 5.01266718e-01 1.07760799e+00
-1.74661353e-01 -1.41847387e-01 -5.74906945e-01 7.11482584e-01
8.01263034e-01 5.36285937e-01 4.37395722e-01 -1.38991249e+00
-3.39279234e-01 -1.91649243e-01 -9.74516645e-02 -5.24497509e-01
2.92927660e-02 -8.47660422e-01 7.56911980e-03 -1.36974752e+00
3.99041437e-02 -7.43377984e-01 -4.34758931e-01 4.09082890e-01
-6.60224259e-02 4.96371299e-01 3.60063195e-01 4.01828974e-01
-1.26259491e-01 1.82185471e-01 9.75897968e-01 -3.45285445e-01
-4.93248522e-01 1.09274656e-01 -7.61738837e-01 6.69164717e-01
6.34450972e-01 -3.44202727e-01 -4.37221855e-01 -3.77581447e-01
6.96503222e-01 1.17939301e-01 4.85454462e-02 -9.22750056e-01
2.64838487e-01 -4.09654200e-01 1.25745445e-01 -9.67860937e-01
4.67677981e-01 -1.22290683e+00 7.12624788e-01 3.29774171e-01
-7.41767790e-03 7.71750957e-02 -4.26775962e-02 4.50136960e-01
-1.32917136e-01 -4.65990365e-01 8.95517945e-01 -1.09167747e-01
-4.57173944e-01 3.38464081e-01 -2.88319319e-01 -4.92078036e-01
9.59357977e-01 -3.73766661e-01 -4.51321527e-02 -1.88706860e-01
-8.62622261e-01 1.81637015e-02 5.04956424e-01 -1.67586118e-01
5.08567214e-01 -1.01206601e+00 -7.89927363e-01 5.19175768e-01
1.87403977e-01 4.75518182e-02 4.40703839e-01 1.14779592e+00
-5.71958721e-01 1.55449688e-01 3.58314365e-01 -6.01956964e-01
-1.35445452e+00 5.72927594e-01 8.65504667e-02 -1.10918726e-03
-3.51782411e-01 1.10564411e+00 -3.78272980e-01 -3.16505671e-01
1.48194030e-01 -2.91239291e-01 -7.05451816e-02 4.43845272e-01
7.61242867e-01 6.52836084e-01 4.48393971e-01 -9.65427876e-01
-1.78493679e-01 6.46142304e-01 4.95623708e-01 -2.74156064e-01
1.40389514e+00 -1.89617872e-01 -8.85176778e-01 1.49628490e-01
1.04197133e+00 3.24925214e-01 -9.29786861e-01 -3.25509697e-01
-9.54926461e-02 -6.16842270e-01 5.39405867e-02 -5.56727052e-01
-1.38060379e+00 5.84272206e-01 7.88955390e-01 5.12652695e-01
1.63483119e+00 -6.37403846e-01 4.36039716e-01 3.49986404e-01
3.04196805e-01 -1.16367316e+00 -3.15549046e-01 4.59770918e-01
8.11324060e-01 -1.01819921e+00 4.61124629e-01 -7.39076734e-01
-4.15975213e-01 1.06177449e+00 3.48493597e-03 1.95689619e-01
7.21830249e-01 7.66046643e-02 1.70531183e-01 -2.50573784e-01
4.73530740e-02 -4.54368412e-01 1.32404357e-01 4.31359053e-01
6.62283227e-02 3.57492357e-01 -4.38030303e-01 3.12628120e-01
-3.44254345e-01 -3.02150309e-01 5.19667983e-01 1.00436854e+00
-5.53283095e-01 -1.11977303e+00 -1.11029088e+00 6.00505888e-01
-1.36516273e-01 -2.43482113e-01 -3.72508973e-01 2.29720712e-01
2.71798491e-01 1.18755114e+00 -4.97601360e-01 -1.57184914e-01
1.74342975e-01 1.85772091e-01 2.28207335e-01 -4.40399379e-01
-3.60885829e-01 3.99959207e-01 -3.38155270e-01 -2.71280408e-01
-8.77431989e-01 -1.10170722e+00 -1.22977436e+00 -1.10163614e-01
-5.68819702e-01 4.66883212e-01 6.15176857e-01 7.89509654e-01
1.69703379e-01 -2.85184756e-02 8.78866673e-01 -3.80204588e-01
-4.80926782e-01 -6.13402605e-01 -1.21479297e+00 2.09564120e-01
2.41457552e-01 -3.86355758e-01 -6.12758696e-01 1.23970531e-01] | [10.061166763305664, -1.9949190616607666] |
161bb0e0-7acf-498d-a7d8-f76c029515d8 | on-the-role-of-conceptualization-in | 2003.03239 | null | https://arxiv.org/abs/2003.03239v2 | https://arxiv.org/pdf/2003.03239v2.pdf | On the Role of Conceptualization in Commonsense Knowledge Graph Construction | Commonsense knowledge graphs (CKGs) like Atomic and ASER are substantially different from conventional KGs as they consist of much larger number of nodes formed by loosely-structured text, which, though, enables them to handle highly diverse queries in natural language related to commonsense, leads to unique challenges for automatic KG construction methods. Besides identifying relations absent from the KG between nodes, such methods are also expected to explore absent nodes represented by text, in which different real-world things, or entities, may appear. To deal with the innumerable entities involved with commonsense in the real world, we introduce to CKG construction methods conceptualization, i.e., to view entities mentioned in text as instances of specific concepts or vice versa. We build synthetic triples by conceptualization, and further formulate the task as triple classification, handled by a discriminatory model with knowledge transferred from pretrained language models and fine-tuned by negative sampling. Experiments demonstrate that our methods can effectively identify plausible triples and expand the KG by triples of both new nodes and edges of high diversity and novelty. | ['Mutian He', 'Kun Xu', 'Yangqiu Song', 'Dong Yu'] | 2020-03-06 | null | null | null | null | ['triple-classification'] | ['graphs'] | [ 2.41225958e-01 8.21734309e-01 -3.13380539e-01 -1.21119857e-01
-3.98003876e-01 -8.28610897e-01 7.98681259e-01 4.60067153e-01
-1.96350599e-03 1.10762417e+00 4.93319035e-01 -2.40944579e-01
-2.48224378e-01 -1.33996177e+00 -8.98774385e-01 -3.71706754e-01
-7.99556896e-02 9.23157156e-01 1.77606091e-01 -4.88039941e-01
1.61469821e-02 5.33464663e-02 -1.55699027e+00 2.86529720e-01
1.25461507e+00 8.65327001e-01 -2.76885957e-01 -5.56368791e-02
-4.03572828e-01 1.15541458e+00 -6.04689717e-01 -1.10024691e+00
1.72978372e-03 -3.03302288e-01 -1.08550477e+00 -1.20321684e-01
5.40032089e-01 2.35982180e-01 -3.40057522e-01 1.24074256e+00
7.63991326e-02 9.77370590e-02 7.83613861e-01 -1.62762058e+00
-1.20967758e+00 1.67068326e+00 -1.10007785e-01 -8.21389258e-02
6.90508306e-01 -4.63106669e-02 1.72103167e+00 -9.47640240e-01
9.93407071e-01 1.62625933e+00 8.13951850e-01 4.00183350e-01
-1.44618535e+00 -5.12844086e-01 2.28704810e-02 4.55251634e-01
-1.43629205e+00 -4.08700049e-01 8.28192055e-01 -1.97091475e-01
8.99254739e-01 4.81048942e-01 7.74536848e-01 1.39881015e+00
-3.71899694e-01 7.35238016e-01 9.49202776e-01 -4.70502675e-01
3.23163480e-01 3.37383002e-01 1.61694765e-01 5.81058681e-01
9.08741117e-01 -3.83831829e-01 -5.82392097e-01 -3.21816415e-01
2.50221401e-01 -3.31849784e-01 -5.22590935e-01 -6.86890602e-01
-1.45531642e+00 8.09416473e-01 5.79409719e-01 3.17375809e-01
-3.58716637e-01 7.14544132e-02 4.95355129e-01 3.23025137e-01
2.31364697e-01 8.06800067e-01 -6.01456642e-01 1.79454491e-01
-7.24529386e-01 4.27261204e-01 1.20135558e+00 1.38665354e+00
7.40652084e-01 -2.09288910e-01 3.41635197e-02 5.98271489e-01
4.21154536e-02 5.09143591e-01 4.24266785e-01 -6.90414548e-01
6.05940163e-01 1.04742014e+00 -1.17962874e-01 -1.26365566e+00
-3.41284335e-01 -3.38280529e-01 -9.44988430e-01 -6.87609613e-01
1.25032783e-01 2.01252684e-01 -7.60128617e-01 1.96376216e+00
4.83732730e-01 -1.02346964e-01 4.35357481e-01 4.46898997e-01
1.11126769e+00 1.90464020e-01 1.81674227e-01 -2.91979164e-01
1.32912338e+00 -4.25952584e-01 -6.87608361e-01 -1.51791453e-01
7.84736276e-01 -5.23236319e-02 1.31999111e+00 9.89636555e-02
-8.93134713e-01 2.94727404e-02 -8.50259721e-01 -3.30264628e-01
-9.87570465e-01 -4.93047744e-01 9.97869730e-01 4.61344630e-01
-8.55303288e-01 6.20479405e-01 -1.90103859e-01 -4.39021260e-01
5.85587084e-01 3.07968305e-03 -4.34794724e-01 -5.24642691e-02
-2.05467820e+00 1.04652131e+00 1.26994741e+00 -6.93679675e-02
-4.50410604e-01 -8.68188441e-01 -1.21978545e+00 5.84426858e-02
1.07074523e+00 -1.23072982e+00 6.52344286e-01 -6.70946479e-01
-6.65568709e-01 1.10864639e+00 1.72465935e-01 -6.00976527e-01
3.55868071e-01 1.19797401e-01 -7.68940210e-01 -2.90671196e-02
4.94799912e-01 3.98338318e-01 7.30269194e-01 -1.44427514e+00
-2.75385141e-01 -4.14426893e-01 5.97248197e-01 1.56155676e-01
-3.16405267e-01 -3.32862496e-01 4.31541093e-02 -6.06885135e-01
3.91381353e-01 -7.30822265e-01 1.09597757e-01 -2.20182613e-01
-1.20993662e+00 -4.28464204e-01 5.35054326e-01 -3.29065621e-01
1.41831028e+00 -1.71556413e+00 9.58271548e-02 5.96729934e-01
7.16247976e-01 -2.94123173e-01 1.86082006e-01 5.00598967e-01
-2.01334789e-01 6.81902707e-01 -1.44813254e-01 2.92621017e-01
4.24839467e-01 4.75470811e-01 -6.14110529e-01 -1.38243899e-01
1.63138539e-01 1.38889098e+00 -1.44916415e+00 -7.74044096e-01
-2.49353781e-01 -2.07364932e-01 -3.42066765e-01 -2.79305667e-01
-5.05357087e-01 -2.13630259e-01 -3.86845857e-01 1.01055431e+00
3.94617170e-01 -5.50763071e-01 4.09229040e-01 -6.47279680e-01
5.29924572e-01 5.23532212e-01 -1.13244939e+00 1.24909711e+00
-1.59272879e-01 2.05184549e-01 -4.97064084e-01 -1.06587446e+00
6.30389571e-01 2.38358695e-02 9.31187272e-02 -1.83652207e-01
-9.94152427e-02 2.10962161e-01 -1.04222029e-01 -5.00556707e-01
7.37280488e-01 -6.98552310e-01 -3.45284373e-01 1.90105364e-01
3.19184184e-01 -5.55147231e-01 5.47759533e-01 8.31871331e-01
1.08328080e+00 -2.78130352e-01 7.52586901e-01 -2.88409740e-01
2.34954700e-01 2.06409082e-01 6.47891283e-01 8.28519821e-01
9.40147340e-02 3.26640606e-02 8.03290248e-01 -2.36249983e-01
-7.87335873e-01 -1.38075662e+00 -1.87065929e-01 8.36813271e-01
3.94123763e-01 -7.83741951e-01 -3.38179283e-02 -8.77322495e-01
3.61253619e-01 1.26769757e+00 -9.51450467e-01 -4.18347895e-01
-1.20022453e-01 -5.03226757e-01 6.92341864e-01 3.75374675e-01
2.88168490e-01 -1.12837923e+00 -2.03945369e-01 1.34651646e-01
-6.28781021e-01 -1.26125228e+00 -5.37356548e-02 2.93034524e-01
-6.45787239e-01 -1.34998262e+00 -7.70069733e-02 -4.99649107e-01
4.65717971e-01 -1.00418046e-01 2.05221295e+00 -2.36423742e-02
-2.22222418e-01 3.23298186e-01 -5.54228604e-01 -3.66283953e-01
-4.33002859e-01 -2.33496889e-01 1.85687631e-01 -3.30559403e-01
6.61378920e-01 -9.85033691e-01 -4.90382314e-02 -2.18622521e-01
-1.13662183e+00 3.24674509e-02 2.88298219e-01 9.10847843e-01
4.57322806e-01 2.84195781e-01 7.76455820e-01 -1.36919618e+00
8.88974249e-01 -8.76381397e-01 5.33762909e-02 7.46579289e-01
-6.64023042e-01 3.40689629e-01 6.34974778e-01 -4.76512462e-01
-6.94348514e-01 -6.46197915e-01 3.61337095e-01 -2.75443196e-01
2.53750443e-01 9.72620130e-01 -5.13145387e-01 2.99264818e-01
1.05583894e+00 3.21051985e-01 -5.40686727e-01 9.89567712e-02
1.14074016e+00 2.70112544e-01 5.01904190e-01 -9.93183374e-01
1.05884850e+00 4.14052218e-01 5.95744438e-02 -6.08203709e-01
-1.15505588e+00 -1.79637715e-01 -6.55729353e-01 4.16895002e-02
4.80295509e-01 -8.21372211e-01 -4.73861635e-01 -4.91108606e-03
-1.05707526e+00 2.11476803e-01 -9.02173638e-01 8.84583816e-02
-5.02122939e-01 4.34162885e-01 -4.25427437e-01 -5.78166723e-01
-3.24861795e-01 -3.47187847e-01 8.54212821e-01 -1.92066714e-01
-5.55973589e-01 -1.23867226e+00 -9.62794125e-02 2.17865959e-01
1.24657132e-01 6.03056014e-01 1.70817971e+00 -1.07599688e+00
-4.87003058e-01 -3.05293482e-02 -2.46647865e-01 1.66835159e-01
2.56635964e-01 -2.91076452e-02 -6.30206168e-01 1.47231251e-01
-2.79947698e-01 -9.71697211e-01 8.34770620e-01 -1.36864915e-01
1.06843197e+00 -8.30790460e-01 -6.44800663e-01 1.70079470e-01
1.32728446e+00 -3.82056415e-01 6.70311391e-01 1.89216316e-01
7.35704482e-01 6.61479294e-01 2.58045971e-01 5.66373408e-01
7.35145688e-01 2.14790627e-01 4.29015785e-01 2.42088884e-01
1.39053255e-01 -7.36593127e-01 7.80128539e-02 6.16873622e-01
-9.49576497e-02 -3.29628408e-01 -1.00848711e+00 7.50432134e-01
-1.64832830e+00 -1.40149570e+00 -1.26088068e-01 1.88915849e+00
1.47190642e+00 2.60551840e-01 3.14545482e-02 8.87444168e-02
6.78468823e-01 3.32966708e-02 -7.92862415e-01 2.21091628e-01
-6.98998630e-01 2.88256496e-01 4.15615998e-02 2.39158317e-01
-8.55379879e-01 1.17734826e+00 5.90915394e+00 9.46074307e-01
-5.61256468e-01 -2.31975541e-01 1.92419589e-01 6.40212446e-02
-1.18156528e+00 2.68180072e-01 -4.27849054e-01 3.14214379e-01
5.38735092e-01 -7.42882133e-01 3.17331463e-01 8.36777627e-01
-4.42587346e-01 2.31043503e-01 -1.40471292e+00 9.67671394e-01
1.55903831e-01 -1.38070929e+00 8.51158381e-01 -2.66604602e-01
7.22663224e-01 -3.01768005e-01 -2.42614642e-01 6.79161966e-01
8.03401411e-01 -9.11079407e-01 8.27820122e-01 4.62565303e-01
6.06627643e-01 -4.39312398e-01 4.79570150e-01 3.72244388e-01
-1.10461366e+00 -1.58821568e-02 -5.22823572e-01 2.32158899e-01
-1.36226490e-01 1.04455340e+00 -8.29851925e-01 8.83358896e-01
5.47733665e-01 7.88997173e-01 -6.97720408e-01 4.49764699e-01
-5.88951051e-01 3.40013623e-01 -1.76205069e-01 -2.68357545e-01
-4.96873893e-02 -8.14916268e-02 5.90526879e-01 1.12110448e+00
1.65484577e-01 1.40469134e-01 1.60238028e-01 1.46005642e+00
-6.23003483e-01 8.12319070e-02 -9.95705545e-01 -4.75189894e-01
8.62509489e-01 1.16095603e+00 -5.75473487e-01 -6.87244236e-01
-1.53059378e-01 8.09383869e-01 6.67154670e-01 4.15884465e-01
-7.33384490e-01 -6.22430801e-01 3.89188111e-01 1.85246989e-01
-1.22376524e-01 1.60242274e-01 -6.91322535e-02 -1.62746799e+00
3.54626089e-01 -8.03391635e-01 6.27132535e-01 -9.77773070e-01
-1.96074104e+00 4.46768045e-01 1.20879956e-01 -9.85202909e-01
-2.94248223e-01 -4.91246283e-01 -3.54573756e-01 5.30338645e-01
-1.25534594e+00 -1.47876310e+00 -1.96721435e-01 8.02179337e-01
2.06745025e-02 3.09491068e-01 7.70026922e-01 -2.13990584e-01
-5.03887646e-02 5.62171996e-01 -2.30845496e-01 1.33297026e-01
4.39030141e-01 -1.65276909e+00 3.28265190e-01 5.83055079e-01
3.38993311e-01 1.23072886e+00 8.12686443e-01 -1.03478873e+00
-1.16416621e+00 -1.08427620e+00 1.34596729e+00 -7.62488663e-01
1.39973962e+00 -5.30572295e-01 -1.08694530e+00 9.83168423e-01
-1.71760526e-02 7.32715949e-02 9.13202763e-01 8.08145642e-01
-1.05401397e+00 2.64795274e-01 -1.23972809e+00 9.31630254e-01
1.74366772e+00 -8.16719472e-01 -1.27894199e+00 5.99203289e-01
1.05545104e+00 -2.54485253e-02 -8.19928169e-01 5.22227526e-01
2.22362623e-01 -6.47437751e-01 8.95767689e-01 -1.26382041e+00
6.87354207e-01 -2.47937471e-01 -2.63734967e-01 -1.64218867e+00
-1.97895378e-01 -4.75959778e-01 -5.92881620e-01 1.31350517e+00
7.96495140e-01 -6.13523483e-01 6.09171152e-01 7.88908780e-01
-4.79760468e-02 -5.77364147e-01 -9.02549803e-01 -9.36312497e-01
1.02614716e-01 -3.62987906e-01 7.19495654e-01 1.63806653e+00
7.63779819e-01 7.34952748e-01 -1.51778562e-02 -1.97982714e-01
8.35762858e-01 4.49355662e-01 3.25406849e-01 -1.61549044e+00
-1.20781012e-01 -3.47382456e-01 -6.74456418e-01 -6.50682211e-01
5.65326810e-01 -1.34467852e+00 -9.41414461e-02 -1.52339995e+00
5.55697441e-01 -6.30464256e-01 -1.02883177e-02 7.50848889e-01
-3.88265699e-01 -7.80186430e-02 -4.87226732e-02 5.82408383e-02
-7.14033067e-01 8.62499297e-01 1.22094500e+00 -5.30137420e-01
6.54925546e-03 -5.69846272e-01 -1.19818854e+00 8.98141444e-01
4.58791673e-01 -3.50746840e-01 -7.53064454e-01 1.49973268e-02
1.08300209e+00 -6.28856197e-02 6.20699704e-01 -5.84714770e-01
2.25865424e-01 -3.01489711e-01 8.74425173e-02 -2.92893916e-01
2.00658187e-01 -8.24660301e-01 2.02560037e-01 9.35884714e-02
-5.92057526e-01 -1.64908171e-01 -1.56813279e-01 7.58891225e-01
-2.89663345e-01 -1.22691110e-01 3.22615504e-01 -5.16016781e-01
-7.44648695e-01 9.78720337e-02 2.48948678e-01 9.33160186e-01
8.44923854e-01 -4.47323322e-01 -6.91592216e-01 -3.47177774e-01
-1.02125645e+00 1.92713499e-01 3.74178290e-01 3.62149686e-01
7.38967478e-01 -1.49056530e+00 -7.95609355e-01 -3.08928907e-01
8.41503024e-01 4.39091958e-02 4.70286682e-02 6.48905993e-01
-2.09725834e-02 1.84108630e-01 1.04259036e-01 -7.81308934e-02
-8.51417601e-01 8.60545635e-01 1.33683294e-01 -2.59516388e-01
-4.25070912e-01 8.56443644e-01 4.87776548e-02 -6.80063665e-01
-1.20920986e-02 -7.31522679e-01 -1.22344375e-01 3.00216883e-01
1.10797316e-01 1.38267756e-01 1.00066988e-02 -4.91741508e-01
-3.89688343e-01 -2.85561420e-02 1.27325105e-02 4.08587754e-01
1.14503706e+00 -1.50581151e-02 -7.10206211e-01 7.04731524e-01
7.53671825e-01 2.50557691e-01 -1.40170991e-01 -5.41496754e-01
4.90551114e-01 -3.33411783e-01 -4.57926214e-01 -1.09622133e+00
-5.35329461e-01 3.36876065e-01 -3.54306519e-01 5.60151041e-01
7.54180074e-01 6.61690414e-01 4.72945541e-01 8.29786062e-01
7.80669451e-01 -8.78099799e-01 2.85717025e-02 6.11302853e-01
1.16330671e+00 -1.21950877e+00 -6.21479675e-02 -7.90932059e-01
-5.67989051e-01 9.83089924e-01 5.27722478e-01 3.35134327e-01
5.26883304e-01 -8.58702976e-03 -4.26301450e-01 -5.45954525e-01
-8.11324239e-01 -3.69361818e-01 3.12331229e-01 7.27554798e-01
2.22447962e-01 3.63374472e-01 -1.93756565e-01 7.74772108e-01
-5.87629378e-01 -2.32318595e-01 5.91197371e-01 6.14719272e-01
-3.86612564e-01 -6.58000588e-01 4.36068885e-02 8.68367910e-01
-6.38371110e-02 -8.92577708e-01 -9.91882205e-01 9.79691923e-01
3.52644920e-01 7.55088389e-01 -4.15376037e-01 -4.04575884e-01
3.62441033e-01 2.48387471e-01 4.71809804e-01 -7.57063270e-01
-3.20293874e-01 -9.74500895e-01 5.60499370e-01 -2.89774388e-01
-4.57791179e-01 -4.41252232e-01 -1.21709085e+00 -4.73034650e-01
-6.07050836e-01 3.25055897e-01 5.37707610e-03 1.02996588e+00
1.93810418e-01 2.14375630e-01 1.74712092e-01 -9.40466747e-02
-7.42053807e-01 -9.15561199e-01 -1.02352893e+00 9.84568954e-01
-8.33278429e-03 -7.79026985e-01 -6.33813620e-01 8.49601179e-02] | [9.657221794128418, 8.16015911102295] |
79a96d58-3134-4efb-bddd-85f8af1af292 | zero-1-to-3-zero-shot-one-image-to-3d-object | 2303.11328 | null | https://arxiv.org/abs/2303.11328v1 | https://arxiv.org/pdf/2303.11328v1.pdf | Zero-1-to-3: Zero-shot One Image to 3D Object | We introduce Zero-1-to-3, a framework for changing the camera viewpoint of an object given just a single RGB image. To perform novel view synthesis in this under-constrained setting, we capitalize on the geometric priors that large-scale diffusion models learn about natural images. Our conditional diffusion model uses a synthetic dataset to learn controls of the relative camera viewpoint, which allow new images to be generated of the same object under a specified camera transformation. Even though it is trained on a synthetic dataset, our model retains a strong zero-shot generalization ability to out-of-distribution datasets as well as in-the-wild images, including impressionist paintings. Our viewpoint-conditioned diffusion approach can further be used for the task of 3D reconstruction from a single image. Qualitative and quantitative experiments show that our method significantly outperforms state-of-the-art single-view 3D reconstruction and novel view synthesis models by leveraging Internet-scale pre-training. | ['Carl Vondrick', 'Sergey Zakharov', 'Pavel Tokmakov', 'Basile Van Hoorick', 'Rundi Wu', 'Ruoshi Liu'] | 2023-03-20 | null | null | null | null | ['single-view-3d-reconstruction', 'image-to-3d'] | ['computer-vision', 'computer-vision'] | [ 4.43605721e-01 3.42333347e-01 -1.03859568e-03 -3.17553848e-01
-5.69191933e-01 -8.81063938e-01 9.51381624e-01 -8.17060947e-01
-9.91285741e-02 2.23571584e-01 1.80044428e-01 1.01307712e-01
2.85429984e-01 -8.06865931e-01 -9.85133588e-01 -6.07515156e-01
5.41598082e-01 6.76031053e-01 3.91600847e-01 5.23593687e-02
1.99413881e-01 6.74879909e-01 -1.28301775e+00 3.63988996e-01
2.27585196e-01 8.22938740e-01 4.31591809e-01 8.62541199e-01
1.57861218e-01 6.13919377e-01 -3.00306171e-01 -4.02211010e-01
6.57570839e-01 -3.11325312e-01 -3.96972358e-01 8.87215137e-01
6.05592668e-01 -7.85712481e-01 -6.08173668e-01 1.02675283e+00
3.59087735e-01 -5.92738716e-03 7.54907668e-01 -8.71451735e-01
-1.15771806e+00 -6.85442518e-03 -6.24945581e-01 -5.18465079e-02
7.02501237e-01 5.00107348e-01 7.75305748e-01 -9.81637776e-01
1.37711251e+00 1.38552547e+00 1.96190953e-01 7.35612035e-01
-1.68633556e+00 -4.15026963e-01 4.71226275e-01 -3.34609509e-01
-1.17101133e+00 -4.67914194e-01 1.02824104e+00 -4.58553076e-01
7.59256482e-01 -1.27614304e-01 9.06475246e-01 1.77300632e+00
2.36366197e-01 6.93514764e-01 1.38443744e+00 -1.52765751e-01
6.18616939e-01 6.96759447e-02 -6.99589193e-01 5.34186602e-01
-2.40804218e-02 1.37708008e-01 -6.84885740e-01 4.78335395e-02
1.35122931e+00 2.41787344e-01 -3.51273745e-01 -1.06156421e+00
-1.49139810e+00 7.59118915e-01 5.03972650e-01 -2.14532897e-01
-2.78165966e-01 2.26872563e-01 -3.45958859e-01 2.42621124e-01
6.65411472e-01 2.08629668e-01 -5.82873940e-01 4.88942564e-02
-7.84565270e-01 2.24916667e-01 7.33254015e-01 1.26817560e+00
6.29903495e-01 8.86505693e-02 1.68982685e-01 4.01749402e-01
2.20078737e-01 8.64730299e-01 1.38871968e-01 -1.66859031e+00
2.83794940e-01 3.07809919e-01 3.40276927e-01 -4.87757623e-01
5.10331355e-02 -2.46318713e-01 -6.01928949e-01 6.39814615e-01
3.85288864e-01 6.94522485e-02 -1.20319366e+00 1.61006260e+00
4.26666915e-01 6.41927645e-02 -1.54008627e-01 8.24901640e-01
2.34025896e-01 5.06290495e-01 -6.67292356e-01 -9.66987833e-02
9.13455546e-01 -9.20912862e-01 -2.50310093e-01 -3.89610797e-01
-4.32282574e-02 -6.18280470e-01 1.33256292e+00 6.71284437e-01
-1.21290684e+00 -3.37636471e-01 -8.12944770e-01 -2.13148221e-01
-2.18801931e-01 -2.89943337e-01 4.28822368e-01 6.14319801e-01
-1.10823154e+00 3.35017681e-01 -1.00866652e+00 -4.83720183e-01
6.31529450e-01 -7.68387765e-02 -5.62237978e-01 -5.85601330e-01
-3.29578012e-01 6.55545890e-01 -2.50183970e-01 -2.99058676e-01
-1.49598813e+00 -9.13171649e-01 -7.40342736e-01 -2.70072520e-01
5.20215213e-01 -1.26624513e+00 1.05809438e+00 -8.17151725e-01
-1.95389009e+00 1.13145208e+00 -2.37916950e-02 -9.52704325e-02
8.70526850e-01 -1.86496284e-02 -3.01753506e-02 4.25562263e-01
1.36710420e-01 9.98081744e-01 1.28950644e+00 -1.79111218e+00
-1.13314807e-01 -6.38910353e-01 3.71075809e-01 2.89324790e-01
3.52188535e-02 -3.66714448e-01 -7.11362362e-01 -7.36811101e-01
1.38973668e-01 -1.29733622e+00 -2.07399428e-01 8.14848304e-01
-3.92989814e-01 5.34709871e-01 1.02604973e+00 -2.58269131e-01
3.03667486e-01 -2.00235987e+00 5.99735439e-01 -2.66020671e-02
6.84625581e-02 -3.73970896e-01 -8.68604705e-02 2.70881116e-01
1.53373569e-01 -8.09961036e-02 -3.10820460e-01 -8.18543911e-01
-1.36583880e-01 4.21331048e-01 -3.76568139e-01 6.07505679e-01
-1.95347294e-02 8.12281787e-01 -8.97451580e-01 -1.01356305e-01
5.44112921e-01 6.84939325e-01 -9.21205580e-01 3.04648757e-01
-5.51663339e-01 9.42740977e-01 -2.43064031e-01 6.38536155e-01
7.42864788e-01 -6.18240654e-01 2.83111408e-02 -2.64396936e-01
2.09435478e-01 -2.36184448e-01 -9.57525432e-01 2.67136717e+00
-7.28575587e-01 4.35337216e-01 1.16896920e-01 -3.47049505e-01
5.61884999e-01 -9.07261893e-02 3.83598566e-01 -3.09030890e-01
-3.61184403e-02 -7.77629018e-02 -3.56862277e-01 -3.33768427e-01
3.33723187e-01 -3.00372660e-01 6.55049235e-02 7.48474956e-01
9.90428180e-02 -9.70955670e-01 -2.08361834e-01 5.97602844e-01
9.87800837e-01 4.34496075e-01 -1.52937621e-01 4.39294316e-02
5.75531507e-03 -5.99704444e-01 1.81253865e-01 5.72381914e-01
1.62494287e-01 1.30256522e+00 2.35668734e-01 -2.89335400e-01
-1.34578419e+00 -1.60659146e+00 2.61337180e-02 3.84585142e-01
2.87065595e-01 -4.30467399e-03 -5.99056780e-01 -8.49018335e-01
4.20556255e-02 9.37753320e-01 -9.09231365e-01 6.44753873e-02
-5.95986880e-02 -3.27643692e-01 5.50609604e-02 2.16218174e-01
4.65729177e-01 -6.80840969e-01 -6.68025494e-01 -1.15124978e-01
2.79428028e-02 -1.48002827e+00 -6.86226189e-01 -5.34178540e-02
-8.97128820e-01 -9.36829686e-01 -1.03366733e+00 -4.31586713e-01
9.25870001e-01 6.24532104e-01 1.05203998e+00 -4.72789407e-01
-2.52704471e-01 9.90431666e-01 -8.33169594e-02 -6.60250932e-02
-4.57872152e-01 -2.67480999e-01 6.02656305e-02 3.10135126e-01
-1.60249054e-01 -9.92528021e-01 -1.16860676e+00 4.44960028e-01
-1.24130380e+00 2.16014206e-01 4.81229633e-01 6.44914031e-01
6.81035101e-01 -2.83733070e-01 4.09249496e-03 -8.53562415e-01
3.02102417e-01 -2.01981917e-01 -6.24491155e-01 1.68465555e-01
-5.49262941e-01 1.11545846e-01 3.65490675e-01 -7.42148340e-01
-1.32191408e+00 3.23637366e-01 3.78974676e-01 -1.10640085e+00
-6.99365512e-02 -1.81798249e-01 -2.08325282e-01 -2.57052213e-01
6.92043722e-01 2.76358008e-01 1.29839867e-01 -3.91839623e-01
9.81284916e-01 2.23978788e-01 3.87815773e-01 -5.73912203e-01
9.06138599e-01 1.44076383e+00 -7.96380825e-03 -8.21392894e-01
-8.01875830e-01 8.77931342e-02 -1.06066096e+00 -2.38464400e-01
1.15609169e+00 -1.08576047e+00 -4.17110860e-01 6.12237632e-01
-1.07206726e+00 -7.24025667e-01 -6.60401165e-01 1.87970147e-01
-9.68609929e-01 3.58646423e-01 -4.31838065e-01 -5.32727778e-01
1.61794364e-01 -1.24115002e+00 1.53632808e+00 -1.39015391e-01
-6.56331405e-02 -1.04278445e+00 -3.99831645e-02 4.79477465e-01
2.16611743e-01 2.27898225e-01 6.75391197e-01 2.18689516e-01
-1.15365207e+00 1.72507148e-02 8.16528033e-03 6.13977253e-01
2.06183761e-01 5.91748096e-02 -1.06759787e+00 -4.23476696e-01
2.06759691e-01 -5.99166751e-01 8.04521918e-01 3.91482711e-01
1.02883625e+00 6.21723896e-03 -2.53810763e-01 9.77448761e-01
1.26560497e+00 -1.33666024e-01 5.31929970e-01 5.92365451e-02
6.88373804e-01 3.48781496e-01 2.08066583e-01 4.84634399e-01
4.48515058e-01 5.56892157e-01 6.75564587e-01 7.07883760e-02
-4.98159766e-01 -6.49742842e-01 3.47008348e-01 5.15840113e-01
4.56366688e-02 -3.71035993e-01 -4.32299972e-01 4.10468280e-01
-1.32421875e+00 -8.82993519e-01 5.53289771e-01 2.17750478e+00
5.86177111e-01 2.74208397e-01 -1.67382374e-01 -3.42776358e-01
3.48221064e-01 4.72610682e-01 -1.14252245e+00 9.41715613e-02
-2.79588342e-01 -1.13508455e-01 5.73050618e-01 6.76437438e-01
-5.79071581e-01 9.45423663e-01 6.56288385e+00 3.49573791e-01
-1.10978711e+00 1.20487586e-01 6.88780248e-01 -4.89016891e-01
-8.88037860e-01 2.43437767e-01 -4.99725848e-01 3.25008065e-01
3.51054102e-01 1.47201121e-01 8.87240946e-01 6.87421203e-01
6.78645307e-03 -1.12553395e-01 -1.22507417e+00 1.16530824e+00
4.45573509e-01 -1.44543242e+00 4.46818560e-01 5.05689144e-01
1.27794158e+00 2.76240945e-01 6.28111541e-01 -3.17983150e-01
6.61773205e-01 -5.97031355e-01 9.51418817e-01 7.73840606e-01
9.44263220e-01 -1.93627924e-01 -2.07687110e-01 5.10210097e-01
-6.26525044e-01 -3.15866321e-02 -3.39500338e-01 2.00928986e-01
5.96662045e-01 4.88960564e-01 -5.00799477e-01 2.64584422e-01
6.84661865e-01 1.00278175e+00 -5.23867726e-01 3.26681226e-01
-3.57294053e-01 -2.60371827e-02 -4.50244606e-01 4.77548867e-01
9.02082585e-03 -3.29996139e-01 5.88873625e-01 5.28622925e-01
6.24953270e-01 1.78837955e-01 -6.88635260e-02 1.23297906e+00
-2.99147606e-01 -6.45366371e-01 -1.04125154e+00 1.65688023e-01
2.37479448e-01 1.08406460e+00 -8.09839904e-01 -2.80312836e-01
-5.32065690e-01 1.67742968e+00 2.77072161e-01 6.72026753e-01
-6.35259569e-01 3.00552845e-01 6.12199485e-01 3.65642935e-01
8.66622865e-01 -5.74055970e-01 -1.16780654e-01 -1.72831726e+00
-7.97739774e-02 -7.63566315e-01 -1.28965229e-01 -1.56757152e+00
-1.47008193e+00 4.34005111e-01 1.92794994e-01 -1.26064086e+00
-1.89412355e-01 -7.28313029e-01 -3.51548791e-01 5.13612151e-01
-1.30942523e+00 -1.46143699e+00 -3.90226454e-01 9.36846673e-01
8.46630812e-01 -1.20214246e-01 6.79587543e-01 -2.02393472e-01
3.66082452e-02 6.42950162e-02 1.06626786e-01 -3.74926269e-01
8.55161011e-01 -1.11662340e+00 6.28572881e-01 7.13164508e-01
3.69748563e-01 5.52291274e-01 5.08478999e-01 -5.44140279e-01
-1.65391886e+00 -8.20676208e-01 -1.20184543e-02 -1.06396770e+00
6.53577626e-01 -7.34012425e-01 -4.07318354e-01 9.15122807e-01
2.68073171e-01 4.72806513e-01 2.96439201e-01 -3.65833372e-01
-6.23434186e-01 -7.38140643e-02 -1.14189899e+00 8.15923989e-01
1.47418833e+00 -9.02913153e-01 -2.65513182e-01 3.24530244e-01
8.10225964e-01 -5.29938579e-01 -9.20454502e-01 8.71044118e-03
7.55100012e-01 -1.23573136e+00 1.33964646e+00 -3.37874889e-01
7.00006604e-01 -2.02686831e-01 -5.45250118e-01 -1.50527394e+00
-1.11086980e-01 -8.18046510e-01 -2.06471488e-01 7.47240663e-01
2.81206101e-01 -4.78879809e-01 8.92750978e-01 7.39064693e-01
2.48648703e-01 -4.38733190e-01 -8.08572710e-01 -8.43943059e-01
1.47190392e-01 -3.93328607e-01 3.89261127e-01 8.00041020e-01
-6.15046024e-01 4.84997451e-01 -4.68511075e-01 2.10302398e-01
1.00418448e+00 2.42640391e-01 1.20080626e+00 -9.59514201e-01
-7.06030250e-01 -7.60251954e-02 -2.96733856e-01 -1.59133911e+00
1.17638864e-01 -6.55214131e-01 -3.55904162e-01 -1.32646894e+00
1.07409656e-01 -1.09390169e-01 2.75142282e-01 -8.76126066e-02
2.40114957e-01 4.10039544e-01 3.24726373e-01 3.20053756e-01
-4.07108396e-01 9.21140373e-01 1.86893356e+00 -2.61508077e-01
-7.98711106e-02 -1.42728519e-02 -5.22704601e-01 8.86631012e-01
2.42913142e-01 -4.58333224e-01 -9.43580925e-01 -8.29498231e-01
4.09859240e-01 1.88187972e-01 3.89685959e-01 -6.36024773e-01
1.89763367e-01 -3.18982661e-01 6.63426757e-01 -3.02550703e-01
9.84026253e-01 -1.08332813e+00 2.89250016e-01 1.90518871e-01
-3.69800866e-01 8.06223974e-02 -1.75054505e-01 1.30963254e+00
2.90243417e-01 2.11108074e-01 7.92447567e-01 -5.81100523e-01
-3.34446967e-01 6.36111200e-01 -1.09413572e-01 1.97765082e-01
1.03658211e+00 -3.16310465e-01 -3.48443776e-01 -8.08283031e-01
-8.75183582e-01 -2.72984862e-01 1.29085457e+00 4.93405998e-01
8.43798876e-01 -1.19171107e+00 -3.28322113e-01 5.04083097e-01
3.49184394e-01 3.73120397e-01 3.23477656e-01 3.59432727e-01
-5.80448270e-01 -1.39006168e-01 -1.44290805e-01 -8.99082899e-01
-8.43219817e-01 8.47545087e-01 1.93071872e-01 1.89633667e-02
-9.22305465e-01 7.99361885e-01 7.71966100e-01 -6.25286520e-01
-5.02879657e-02 -5.72063267e-01 7.59182811e-01 -4.65794981e-01
2.20774382e-01 -8.65781233e-02 -3.10463488e-01 -4.51237470e-01
-3.22604328e-02 8.06822002e-01 -1.53743804e-01 -7.96618581e-01
1.33830011e+00 -5.05276918e-01 3.09277207e-01 7.70140409e-01
1.19755161e+00 2.65835911e-01 -2.20159793e+00 -2.74285465e-01
-8.12623799e-01 -9.86917078e-01 6.05596900e-02 -8.66986513e-01
-1.10916364e+00 1.02526379e+00 3.51092994e-01 -1.98119536e-01
7.28221178e-01 1.87979832e-01 5.46146393e-01 4.31117535e-01
6.89515889e-01 -9.37916100e-01 6.42192006e-01 1.54041991e-01
1.02135623e+00 -1.28534412e+00 2.37000346e-01 -2.60700703e-01
-8.66539180e-01 8.93187702e-01 4.03203845e-01 -3.52722824e-01
1.00183630e+00 1.77580029e-01 1.16848201e-02 -1.88880101e-01
-8.28827083e-01 2.48804465e-01 6.40447363e-02 9.45482731e-01
-2.03313589e-01 -2.19355643e-01 6.87173426e-01 -1.12011172e-01
-4.14682813e-02 3.24384905e-02 7.98987806e-01 8.85392785e-01
4.49839905e-02 -9.20835078e-01 -5.07719666e-02 -3.21203321e-02
2.27589421e-02 3.38363238e-02 -6.26556754e-01 8.73698235e-01
-2.54391313e-01 6.49441957e-01 9.42722559e-02 -2.65129745e-01
2.67487943e-01 -2.15495795e-01 1.16970229e+00 -7.08328724e-01
1.34797364e-01 1.68013468e-01 -2.53432184e-01 -7.66268432e-01
-6.18919432e-01 -7.43415236e-01 -7.10741997e-01 -2.07646042e-01
-3.02477665e-02 -7.14580297e-01 6.26637101e-01 7.26105034e-01
4.75868344e-01 3.64698887e-01 7.78019190e-01 -1.46332073e+00
-5.04917383e-01 -6.65979207e-01 -9.03108895e-01 6.25450611e-01
4.42610383e-01 -7.41088331e-01 -6.95178926e-01 3.64582300e-01] | [9.255990982055664, -3.115262508392334] |
da1e5078-e64d-4b12-be80-34d20743a662 | hit-a-hierarchically-fused-deep-attention | 2105.14600 | null | https://arxiv.org/abs/2105.14600v1 | https://arxiv.org/pdf/2105.14600v1.pdf | HIT: A Hierarchically Fused Deep Attention Network for Robust Code-mixed Language Representation | Understanding linguistics and morphology of resource-scarce code-mixed texts remains a key challenge in text processing. Although word embedding comes in handy to support downstream tasks for low-resource languages, there are plenty of scopes in improving the quality of language representation particularly for code-mixed languages. In this paper, we propose HIT, a robust representation learning method for code-mixed texts. HIT is a hierarchical transformer-based framework that captures the semantic relationship among words and hierarchically learns the sentence-level semantics using a fused attention mechanism. HIT incorporates two attention modules, a multi-headed self-attention and an outer product attention module, and computes their weighted sum to obtain the attention weights. Our evaluation of HIT on one European (Spanish) and five Indic (Hindi, Bengali, Tamil, Telugu, and Malayalam) languages across four NLP tasks on eleven datasets suggests significant performance improvement against various state-of-the-art systems. We further show the adaptability of learned representation across tasks in a transfer learning setup (with and without fine-tuning). | ['Md Shad Akhtar', 'Tanmoy Chakraborty', 'Sourabh Kumar Bhattacharjee', 'Ayan Sengupta'] | 2021-05-30 | null | null | null | null | ['deep-attention', 'deep-attention'] | ['computer-vision', 'natural-language-processing'] | [-3.19755450e-03 -3.60281765e-02 -9.61971506e-02 -4.25270945e-01
-1.06371117e+00 -6.29510641e-01 4.10995960e-01 3.36634606e-01
-6.18795812e-01 3.89566302e-01 5.75321853e-01 -7.65905619e-01
3.82821709e-02 -5.42362213e-01 -4.13423866e-01 -3.53918493e-01
7.15431422e-02 2.86322206e-01 -1.28967976e-02 -6.15530849e-01
3.13175589e-01 1.53301768e-02 -9.98145878e-01 6.03717208e-01
1.26440406e+00 3.98825437e-01 6.14709795e-01 6.23576283e-01
-5.69703281e-01 1.09622669e+00 -3.29070836e-01 -7.41147697e-01
-7.25489035e-02 -2.09737509e-01 -9.60578620e-01 -2.83086240e-01
1.89647824e-01 1.53479233e-01 -7.58277923e-02 9.38731551e-01
4.65225697e-01 2.23708935e-02 6.56418979e-01 -6.76588356e-01
-1.22097886e+00 1.18449295e+00 -7.01874673e-01 4.96277720e-01
3.39292079e-01 -1.30651206e-01 1.50933993e+00 -1.18954325e+00
5.18788934e-01 1.37475610e+00 6.91826463e-01 4.26366061e-01
-8.94548535e-01 -4.28776920e-01 1.64856464e-01 1.38121188e-01
-1.23855412e+00 -4.68882114e-01 6.23686135e-01 -6.52046800e-01
1.63086545e+00 -7.36708418e-02 8.61250386e-02 7.53191352e-01
3.41396064e-01 8.43305171e-01 7.90954173e-01 -8.35140824e-01
-2.13898897e-01 2.81071872e-01 3.70871097e-01 9.18140352e-01
2.03981012e-01 -6.20793283e-01 -6.18097842e-01 -1.30354688e-01
2.93039978e-01 -1.58475816e-01 -2.78742313e-01 -2.47038957e-02
-1.13611650e+00 1.23121452e+00 3.06296736e-01 6.64927781e-01
-2.49823779e-01 -3.06893652e-03 6.88561320e-01 2.88847744e-01
7.38930523e-01 4.88502115e-01 -9.15841639e-01 -3.50618333e-01
-8.25324297e-01 -1.09538212e-01 6.73330426e-01 1.02527082e+00
6.77423656e-01 2.61583418e-01 -1.64297700e-01 1.30003810e+00
5.48807800e-01 4.66130674e-01 1.01269388e+00 -2.59056121e-01
1.11326909e+00 7.97895074e-01 -4.12766337e-01 -8.54885161e-01
-2.89142281e-01 -2.92780668e-01 -4.78134304e-01 -2.82961071e-01
2.19292223e-01 -2.04343319e-01 -7.10846543e-01 1.65896285e+00
-3.03461333e-03 -4.57481831e-01 3.18265587e-01 4.65053469e-01
9.89676654e-01 9.03651595e-01 1.99228019e-01 1.19389877e-01
1.55370200e+00 -1.33137643e+00 -6.54450178e-01 -7.32001781e-01
8.28552783e-01 -9.00567710e-01 1.46681416e+00 -7.32433498e-02
-1.28785920e+00 -4.17387813e-01 -9.92935479e-01 -7.06842363e-01
-6.60482407e-01 4.68025625e-01 5.07554293e-01 5.69605827e-01
-9.16780353e-01 3.11095119e-01 -7.83744872e-01 -5.22444189e-01
2.67961174e-01 2.05266684e-01 -3.48549962e-01 -2.46950448e-01
-1.13817656e+00 1.02846313e+00 3.58058363e-01 -1.07086010e-01
-3.61645460e-01 -7.68499732e-01 -1.40108407e+00 2.42505774e-01
-7.04973191e-02 -3.64339083e-01 1.10501707e+00 -8.02347183e-01
-1.42151141e+00 1.06809688e+00 -3.25313717e-01 -3.60871047e-01
8.76703486e-02 -3.43799859e-01 -2.87560999e-01 -8.16773474e-02
2.95840979e-01 3.62949848e-01 7.40702271e-01 -7.85251856e-01
-5.41248918e-01 -2.78131068e-01 1.12967268e-01 2.17199787e-01
-6.64839089e-01 6.61634326e-01 -3.60696137e-01 -7.49960363e-01
-2.28309661e-01 -6.74059331e-01 -1.04332767e-01 -4.54459339e-01
9.19120014e-03 -5.34466922e-01 5.65161884e-01 -1.14198256e+00
1.55094695e+00 -2.38584042e+00 3.42958421e-01 -2.80837327e-01
-8.24300498e-02 3.38765413e-01 -3.50842774e-01 6.74595296e-01
-3.91304353e-03 1.68934435e-01 -5.14356554e-01 -4.53175277e-01
5.34559861e-02 3.17271382e-01 -3.80739570e-01 2.97817826e-01
6.50483131e-01 1.11261213e+00 -9.64197993e-01 -4.58115786e-01
3.42876464e-02 4.07766908e-01 -7.23511517e-01 1.19865626e-01
-2.59265713e-02 -9.65805593e-05 -3.51624101e-01 6.54431403e-01
3.05456609e-01 -1.11532085e-01 2.53735870e-01 1.08284391e-01
-2.55319893e-01 8.30053389e-01 -6.68639004e-01 1.96588767e+00
-1.02819502e+00 6.07998967e-01 1.54101044e-01 -8.06707501e-01
7.67017066e-01 2.70851970e-01 -6.00964092e-02 -5.49891829e-01
3.52985144e-01 3.53666991e-01 3.78103197e-01 -6.70731068e-01
9.96913016e-01 -1.81028731e-02 -4.19619054e-01 5.27180076e-01
3.75431746e-01 -2.60472715e-01 3.94243419e-01 2.91159093e-01
1.14029074e+00 1.86771959e-01 7.46568441e-01 -6.55407667e-01
8.49960208e-01 -1.52051911e-01 3.58827561e-01 3.04205537e-01
-2.52428874e-02 5.81634045e-01 5.80575824e-01 -1.24099113e-01
-9.14246559e-01 -6.55864775e-01 -5.73729426e-02 1.92577875e+00
-3.75289232e-01 -6.11820400e-01 -5.49801350e-01 -7.36231744e-01
-2.85960492e-02 8.56297553e-01 -7.45947838e-01 -8.36184695e-02
-8.14898074e-01 -7.56741881e-01 6.58323705e-01 5.55273294e-01
1.94924578e-01 -1.35565567e+00 -3.78603637e-01 3.78767580e-01
-3.11139554e-01 -1.08490670e+00 -6.48030281e-01 3.71285200e-01
-4.38874304e-01 -8.22178483e-01 -6.60819650e-01 -1.15460861e+00
5.14133096e-01 2.39930823e-01 1.23999095e+00 1.16611898e-01
-1.16588026e-01 2.35905051e-01 -5.53335547e-01 -5.12515187e-01
-5.14735639e-01 4.37294751e-01 -4.01397079e-01 -2.03231394e-01
5.30532360e-01 -1.98070090e-02 -5.24333604e-02 -3.09517562e-01
-7.72269249e-01 -1.39480591e-01 5.68507075e-01 9.16403174e-01
9.22756419e-02 -4.49774772e-01 5.56596100e-01 -1.11798036e+00
9.17271078e-01 -8.11942279e-01 -3.13237458e-01 3.00347447e-01
-2.28537664e-01 2.76874632e-01 7.10497737e-01 -2.91993916e-01
-1.20679998e+00 -8.40736702e-02 -3.25546682e-01 2.09387183e-01
3.46802056e-01 9.24549341e-01 -2.27099150e-01 2.57115126e-01
6.12057447e-01 1.91558972e-01 -1.92947507e-01 -5.72468340e-01
6.66998982e-01 1.00790465e+00 3.42415869e-01 -8.51844251e-01
6.47913575e-01 1.30522370e-01 -7.35426068e-01 -1.06093931e+00
-9.29891467e-01 -6.40482545e-01 -7.51705766e-01 3.47495943e-01
9.92671311e-01 -1.08198452e+00 -8.50283876e-02 3.04234952e-01
-1.40637469e+00 -3.44674081e-01 -2.05962926e-01 3.39371830e-01
-1.40409052e-01 3.74198109e-01 -9.61713672e-01 -4.41891402e-01
-5.67192972e-01 -1.12753999e+00 1.13178647e+00 -1.25983313e-01
-2.78074354e-01 -1.33045387e+00 1.06153429e-01 3.56569290e-01
5.51923573e-01 -2.89878607e-01 1.41896570e+00 -7.84326553e-01
-1.66541189e-02 7.04988390e-02 -4.93538260e-01 4.49511081e-01
6.83164224e-02 6.12803828e-03 -8.33005548e-01 -3.09422672e-01
-1.96976870e-01 -4.21792775e-01 8.78867507e-01 7.95713142e-02
7.69643366e-01 -3.41653049e-01 3.37028801e-02 3.97179306e-01
1.40370274e+00 -6.55342862e-02 3.85729611e-01 3.68662775e-01
8.21902454e-01 5.91879129e-01 2.39757106e-01 3.19417477e-01
7.36524343e-01 2.51155704e-01 1.37762576e-01 -1.85138695e-02
-1.90265849e-01 -4.66933437e-02 7.35110462e-01 1.73592854e+00
2.38368556e-01 -2.17932716e-01 -1.18526495e+00 1.02451766e+00
-1.71839046e+00 -5.42985797e-01 -1.83556244e-01 1.75201583e+00
1.20036042e+00 -7.26271123e-02 -2.55861938e-01 -8.38835388e-02
4.66619164e-01 2.22259104e-01 -1.73935592e-01 -9.58034456e-01
1.61235072e-02 4.21794981e-01 3.15997779e-01 5.30467868e-01
-9.79593515e-01 1.31014955e+00 5.61100960e+00 8.28160763e-01
-1.07425416e+00 3.96749020e-01 1.46353140e-01 3.93770516e-01
-4.96009737e-01 -8.40200260e-02 -8.23811650e-01 2.75769651e-01
1.12957561e+00 -4.67557788e-01 2.96189249e-01 9.37184632e-01
-1.60277352e-01 2.75091499e-01 -9.59572434e-01 7.07844377e-01
4.86442208e-01 -1.02220392e+00 -4.14456241e-03 -2.06173524e-01
7.70203352e-01 5.57583451e-01 -3.23744193e-02 8.02105248e-01
5.01567364e-01 -9.38898265e-01 8.07591736e-01 -1.79615542e-01
8.60412717e-01 -7.18962371e-01 9.57279801e-01 2.93391407e-01
-1.30623746e+00 -3.62119973e-02 -5.81462920e-01 -2.61465251e-01
-3.96978296e-02 1.97730482e-01 -7.38820553e-01 5.07098317e-01
5.38304567e-01 9.41266239e-01 -8.43688786e-01 5.96404254e-01
-5.54510653e-01 7.96101213e-01 1.19321674e-01 -3.15993339e-01
5.68090498e-01 -2.91345245e-03 2.71746635e-01 1.84501088e+00
2.83050567e-01 -4.80331853e-02 1.05353311e-01 7.40697682e-01
-2.35131040e-01 7.00710595e-01 -5.37851095e-01 -3.10139239e-01
2.92821825e-01 1.51496577e+00 -5.20356119e-01 -3.52410406e-01
-8.90191197e-01 8.49077761e-01 8.71509910e-01 1.60469741e-01
-5.65781415e-01 -7.76321590e-01 6.74603701e-01 -1.68275848e-01
5.47079206e-01 -5.36579609e-01 -2.24221766e-01 -1.34876084e+00
-1.68795571e-01 -9.62676466e-01 4.97073442e-01 -6.67145193e-01
-1.44196463e+00 8.17258894e-01 -2.36053392e-01 -9.49602425e-01
-2.50471801e-01 -9.74213958e-01 -5.03222823e-01 1.10587180e+00
-1.74900270e+00 -1.42526531e+00 8.28489214e-02 5.46166182e-01
1.04119492e+00 -4.47692662e-01 9.56658244e-01 3.42195958e-01
-4.75017428e-01 6.75623417e-01 2.15035558e-01 5.68519473e-01
6.15166545e-01 -1.33017933e+00 5.68019271e-01 9.82449651e-01
3.49756926e-01 9.96762812e-01 3.20053786e-01 -3.29564154e-01
-1.46613300e+00 -1.19418263e+00 1.42870212e+00 -4.18937713e-01
1.25522089e+00 -6.33571506e-01 -1.12774134e+00 8.65967214e-01
7.28298008e-01 -6.02520816e-02 8.04367483e-01 2.54088700e-01
-5.92638254e-01 3.20035398e-01 -8.07687044e-01 5.48062682e-01
5.49472511e-01 -7.62759030e-01 -1.17259228e+00 1.50350213e-01
9.45931971e-01 -3.95571589e-01 -7.28609443e-01 1.62910821e-03
3.13096493e-01 -3.94594848e-01 5.25083423e-01 -6.12493157e-01
8.08365166e-01 -1.19955383e-01 -3.51700068e-01 -1.43635726e+00
-4.89732295e-01 -4.66389596e-01 2.44145408e-01 1.53413880e+00
7.15471625e-01 -5.13253033e-01 1.70122176e-01 2.47200966e-01
-4.72652912e-01 -5.55666685e-01 -7.67423391e-01 -5.81872284e-01
6.61966026e-01 -4.89042252e-01 3.84094626e-01 1.12858152e+00
4.68516976e-01 8.66632462e-01 -1.25282267e-02 -8.36331323e-02
1.93912640e-01 -1.99986733e-02 3.94290984e-01 -9.76251364e-01
-2.40242958e-01 -4.76947725e-01 -2.96276718e-01 -8.43433976e-01
6.96781218e-01 -1.53114116e+00 9.80835631e-02 -1.68373883e+00
3.00826102e-01 -1.99275613e-01 -2.72958130e-01 8.25306594e-01
-2.61208653e-01 6.30825683e-02 2.37980217e-01 -7.45776743e-02
-4.89117175e-01 6.61439121e-01 7.82845259e-01 -2.93799639e-01
-2.24346574e-02 -4.12079722e-01 -9.84450758e-01 7.49870002e-01
7.93603420e-01 -5.39341450e-01 -2.64445782e-01 -1.01638114e+00
4.63994980e-01 -3.47483456e-01 -3.54581565e-01 -5.77556670e-01
1.51141793e-01 3.36938687e-02 -6.94716945e-02 -3.74604881e-01
-6.34344593e-02 -5.66968143e-01 -7.69059479e-01 3.01624149e-01
-4.77552176e-01 4.07675415e-01 4.47439641e-01 1.18752934e-01
-2.70257175e-01 -6.43758833e-01 7.45010316e-01 -3.22025985e-01
-7.40972579e-01 5.94647266e-02 -5.61493635e-01 6.31523192e-01
5.69986105e-01 5.56949936e-02 -3.46151888e-01 1.38747022e-01
-1.76808268e-01 2.77917743e-01 1.38110161e-01 8.82077336e-01
5.71625471e-01 -1.10409379e+00 -1.04968643e+00 5.08562088e-01
4.40457225e-01 -3.03162515e-01 -1.82067245e-01 4.97560978e-01
-4.32431668e-01 4.91639018e-01 -9.52014476e-02 -2.20384941e-01
-1.20390356e+00 3.53367895e-01 1.04137529e-02 -5.00371158e-01
-3.88984382e-01 9.91195202e-01 1.32061154e-01 -9.35937345e-01
-5.07058986e-02 -6.15530550e-01 -4.44837332e-01 1.55675292e-01
5.30037880e-01 2.33605802e-01 3.15508246e-01 -9.10484076e-01
-4.57348436e-01 5.63937962e-01 -3.38789076e-01 2.43228935e-02
1.45376992e+00 -2.27132156e-01 -5.36271513e-01 6.53667927e-01
1.23903179e+00 4.76006061e-01 -5.92893302e-01 -3.42635691e-01
3.89094323e-01 -1.27492681e-01 -4.11992380e-03 -6.60199404e-01
-8.11287642e-01 1.25708842e+00 -7.16811195e-02 -9.63163842e-03
7.25177467e-01 4.78479378e-02 7.75105774e-01 2.90712357e-01
1.42722189e-01 -1.17120278e+00 -1.99134205e-03 1.29516828e+00
9.34862792e-01 -1.20070851e+00 -1.37062028e-01 -2.61244506e-01
-8.36842000e-01 1.09804273e+00 4.95426655e-01 -4.74424735e-02
7.19171405e-01 3.93451154e-01 1.17362052e-01 -9.99489725e-02
-8.69084954e-01 -2.88780302e-01 4.91253883e-01 3.67083102e-01
1.25717199e+00 -1.26194835e-01 -4.92071480e-01 9.13153350e-01
-3.49243194e-01 -5.23596108e-01 5.97053826e-01 9.85458016e-01
-4.63384360e-01 -1.08490324e+00 -2.24286504e-03 3.73474985e-01
-9.20758426e-01 -7.74762392e-01 -1.64668471e-01 8.57149780e-01
6.83588535e-02 9.34175372e-01 -2.94038989e-02 7.36559033e-02
1.44796312e-01 2.39381433e-01 2.76001722e-01 -1.21809542e+00
-1.09005034e+00 -6.27517104e-02 6.78328723e-02 -2.27705896e-01
-1.87430426e-01 -5.75156033e-01 -1.42852175e+00 -1.76024828e-02
-3.06356996e-01 1.84671089e-01 7.14381516e-01 9.39360201e-01
2.52044380e-01 7.20362484e-01 1.78034276e-01 -6.14566445e-01
-4.93660927e-01 -1.31015193e+00 -5.09980857e-01 3.39686722e-01
1.77751675e-01 -2.81087101e-01 -3.14440966e-01 1.35669932e-01] | [10.649577140808105, 9.695432662963867] |
847d9597-f05e-492a-a689-2a253dcd5ed0 | document-enhancement-using-visibility | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Kligler_Document_Enhancement_Using_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Kligler_Document_Enhancement_Using_CVPR_2018_paper.pdf | Document Enhancement Using Visibility Detection | This paper re-visits classical problems in document enhancement. Rather than proposing a new algorithm for a specific problem, we introduce a novel general approach. The key idea is to modify any state- of-the-art algorithm, by providing it with new information (input), improving its own results. Interestingly, this information is based on a solution to a seemingly unrelated problem of visibility detection in R3. We show that a simple representation of an image as a 3D point cloud, gives visibility detection on this cloud a new interpretation. What does it mean for a point to be visible? Although this question has been widely studied within computer vision, it has always been assumed that the point set is a sampling of a real scene. We show that the answer to this question in our context reveals unique and useful information about the image. We demonstrate the benefit of this idea for document binarization and for unshadowing. | ['Sagi Katz', 'Netanel Kligler', 'Ayellet Tal'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['document-enhancement'] | ['computer-vision'] | [ 8.05518925e-01 2.38588095e-01 2.63856649e-01 -3.36153284e-02
-2.94530272e-01 -7.88859665e-01 7.00107038e-01 2.46013194e-01
-8.39731842e-02 1.81687176e-01 -1.56628609e-01 -5.95271766e-01
-1.64096206e-01 -9.43518817e-01 -3.91104907e-01 -1.05074024e+00
-2.72822920e-02 3.79083723e-01 6.21992946e-01 -6.00801170e-01
5.60702860e-01 1.00642157e+00 -1.59770727e+00 1.93586648e-01
6.74949467e-01 5.14283597e-01 3.78351003e-01 1.00480282e+00
-2.22026154e-01 6.98205605e-02 -8.72171521e-01 -5.83403766e-01
3.91856790e-01 -4.51146811e-01 -9.79102135e-01 8.06214869e-01
3.87148291e-01 7.13078901e-02 1.28840730e-01 1.26508367e+00
2.76905969e-02 3.66909690e-02 6.99442506e-01 -1.03732717e+00
-6.85072541e-01 -7.20684305e-02 -7.81633854e-01 1.42920345e-01
5.20245790e-01 -2.59143800e-01 8.57337475e-01 -7.95775235e-01
6.99429870e-01 9.41702425e-01 6.63048387e-01 7.71593451e-02
-1.28373885e+00 1.41776085e-01 1.83177561e-01 1.44294620e-01
-1.31513608e+00 2.21201237e-02 7.61069536e-01 -3.93874645e-01
4.34165418e-01 8.93853664e-01 8.80651951e-01 4.37525243e-01
-5.72145246e-02 6.90007031e-01 1.23090529e+00 -1.03665292e+00
1.62142113e-01 4.98057902e-01 4.77798611e-01 6.19841993e-01
4.99599934e-01 5.32898866e-03 -2.14273989e-01 -2.18661502e-02
7.14021087e-01 -4.64674830e-03 -6.59731925e-01 -6.67677760e-01
-9.55959201e-01 7.17862129e-01 3.65519226e-01 5.75649679e-01
-1.09347001e-01 -2.87007242e-01 -1.72967762e-01 2.85276115e-01
3.58224392e-01 3.05276960e-01 -5.29184714e-02 1.33059785e-01
-1.03264654e+00 1.99035883e-01 9.07781005e-01 7.44968295e-01
7.69381344e-01 -1.44959897e-01 2.71594912e-01 3.12779635e-01
1.84094355e-01 5.98246694e-01 -1.11580037e-01 -8.72453034e-01
-2.89970338e-02 5.94027460e-01 1.49273545e-01 -1.26052117e+00
-2.37152025e-01 -5.84064722e-01 -5.28083622e-01 8.94510686e-01
6.52702689e-01 5.09040356e-01 -6.69718266e-01 1.10990345e+00
4.55659300e-01 -1.87489808e-01 -3.57667776e-03 8.31106365e-01
4.16135132e-01 6.12640679e-01 -8.64872217e-01 -2.97241420e-01
1.39383364e+00 -7.62194812e-01 -6.78731978e-01 2.12455094e-01
2.28718027e-01 -9.94707406e-01 8.79908442e-01 1.14956844e+00
-1.03389168e+00 -4.77273226e-01 -1.28000462e+00 -2.73649246e-02
-6.14386916e-01 -1.55906901e-01 6.25054419e-01 1.05140889e+00
-1.22473228e+00 7.04061747e-01 -2.13591918e-01 -7.66557336e-01
-1.11129984e-01 9.19197947e-02 -2.86251158e-01 -2.42247805e-02
-5.20324290e-01 1.06575358e+00 1.06469169e-01 -1.26294270e-02
-3.62782210e-01 -5.76057285e-02 -4.78212357e-01 2.40954887e-02
6.05284393e-01 -4.87962693e-01 1.01366508e+00 -1.02624035e+00
-1.20678318e+00 1.21251392e+00 -4.65373844e-01 -1.76591843e-01
6.50059283e-01 -4.27268445e-02 -2.84546435e-01 4.56581682e-01
-2.54148155e-01 -6.72286898e-02 1.22137320e+00 -1.84881496e+00
-4.29123372e-01 -5.76023281e-01 3.17010880e-01 -3.90039049e-02
-1.42482027e-01 1.79163843e-01 -5.78809619e-01 -6.20953619e-01
5.16824961e-01 -6.95400059e-01 -1.87312290e-01 5.03414981e-02
-2.72521377e-01 1.63034424e-01 9.15805280e-01 -6.28692389e-01
8.79402757e-01 -2.06066871e+00 2.33012885e-01 5.99516332e-01
2.89801002e-01 2.75670052e-01 6.06963523e-02 5.88774323e-01
-2.08032593e-01 1.62314937e-01 -5.88343859e-01 -4.37486619e-01
-6.09618388e-02 2.26697445e-01 -5.62788188e-01 9.18455899e-01
3.82478945e-02 4.55169886e-01 -8.48712444e-01 -5.46046555e-01
4.95261371e-01 6.70442462e-01 -1.37256011e-01 -2.81129181e-01
5.65168932e-02 3.10273498e-01 -2.26852417e-01 6.40434682e-01
1.22046769e+00 -1.15612619e-01 4.12907638e-02 8.32957029e-02
-4.32364881e-01 -2.07493931e-01 -1.52901423e+00 1.30640256e+00
1.14594121e-02 7.89178431e-01 3.10880512e-01 -1.00530946e+00
1.01030087e+00 1.91721782e-01 4.20554996e-01 -5.08277297e-01
9.43291634e-02 8.92593041e-02 -3.89857411e-01 -4.29410070e-01
9.89905596e-01 -1.00882955e-01 4.86394763e-01 4.26872551e-01
-3.42143595e-01 -2.93395758e-01 3.02040833e-03 1.22331142e-01
1.08643746e+00 6.37907684e-02 5.57115674e-01 -1.67357191e-01
5.89418888e-01 2.48445645e-01 4.41854447e-02 8.86908054e-01
3.90740782e-02 1.03706300e+00 3.90588284e-01 -1.26964301e-01
-1.02600038e+00 -1.18615353e+00 -5.86962640e-01 7.70228267e-01
3.74246687e-01 -6.08365655e-01 -8.46947789e-01 -4.72493410e-01
-2.55354084e-02 4.12455678e-01 -7.39508152e-01 3.67242604e-01
-3.88309956e-01 -9.13860083e-01 -6.18639551e-02 3.10743768e-02
1.83104411e-01 -7.07906008e-01 -8.58819485e-01 -7.69307688e-02
-5.60158268e-02 -1.02482748e+00 -7.48936087e-02 1.97836816e-01
-1.10838532e+00 -1.20640421e+00 -1.03421056e+00 -3.81590009e-01
8.55752528e-01 1.10129154e+00 1.23062873e+00 5.76050758e-01
-4.51406389e-01 6.26406133e-01 -7.18777418e-01 -4.75444168e-01
-4.79789436e-01 -4.65897590e-01 -2.51925886e-01 1.45185426e-01
4.42135334e-01 -4.39079940e-01 -3.08685452e-01 2.45769486e-01
-1.21307909e+00 -3.52663577e-01 4.36042786e-01 3.46444368e-01
4.66735363e-01 3.92351329e-01 -2.70095736e-01 -7.65178502e-01
5.13719022e-01 1.12465568e-01 -7.01982319e-01 3.48621279e-01
-5.03597081e-01 -5.83981872e-02 2.37336054e-01 4.06787694e-02
-7.42121100e-01 1.23854712e-01 5.26647195e-02 -1.25584528e-01
-1.52738780e-01 1.21083006e-01 -2.65245140e-01 -3.82558048e-01
5.58765769e-01 3.80823463e-01 -3.27457115e-02 -6.41434908e-01
6.79042399e-01 5.58843970e-01 7.06416905e-01 -1.99566200e-01
1.26120746e+00 1.22930610e+00 5.12882710e-01 -1.34585536e+00
-4.56616253e-01 -1.06276810e+00 -1.05611444e+00 -2.79508263e-01
7.75114357e-01 -1.67640060e-01 -7.87818968e-01 7.96072856e-02
-1.53444195e+00 2.25245640e-01 -4.89816457e-01 8.34656060e-02
-5.31168759e-01 9.53859448e-01 1.27682611e-01 -1.29313409e+00
5.17253280e-02 -9.38985109e-01 1.06604493e+00 -4.59296480e-02
8.55529085e-02 -9.66634989e-01 2.43777499e-01 1.77844107e-01
6.78735897e-02 2.71457374e-01 5.85136354e-01 -2.20284984e-01
-7.01539397e-01 -3.50392222e-01 -2.95633256e-01 4.13890123e-01
8.72053429e-02 3.64248991e-01 -1.26445866e+00 -1.48858756e-01
5.09582818e-01 5.59275508e-01 9.60843027e-01 3.09875816e-01
8.90532970e-01 2.48971116e-02 -3.56198490e-01 6.52951598e-01
1.70589066e+00 7.43284076e-02 7.44670510e-01 7.59812117e-01
4.47989881e-01 8.69761109e-01 6.58176661e-01 2.11629555e-01
2.37566549e-02 9.72111106e-01 9.05881345e-01 -4.43649828e-01
-6.01963401e-02 2.40660727e-01 1.00948080e-01 3.84582430e-01
-6.61351621e-01 -1.65713012e-01 -7.21301138e-01 3.17907453e-01
-1.79570353e+00 -1.06984162e+00 -7.81286538e-01 2.50646544e+00
2.13988915e-01 1.95438609e-01 2.96574712e-01 8.34216177e-01
7.19556987e-01 1.49000600e-01 1.74183846e-01 -6.50723934e-01
-4.30469692e-01 9.19883624e-02 6.32576942e-01 8.20966244e-01
-8.95986795e-01 4.32911277e-01 6.77734947e+00 5.17238200e-01
-9.28456366e-01 1.16037890e-01 6.40438963e-03 3.65077794e-01
-4.61316735e-01 2.98030555e-01 -7.07380891e-01 8.06634799e-02
1.96925908e-01 -1.28788888e-01 2.43044928e-01 7.91336536e-01
5.56919314e-02 -4.19605970e-01 -8.55127692e-01 8.72781515e-01
7.22640038e-01 -1.06454623e+00 5.66823743e-02 5.12366414e-01
3.94933373e-01 -5.42173684e-01 2.40525633e-01 -3.99595469e-01
4.09188978e-02 -9.89957452e-01 7.37223268e-01 6.30456388e-01
2.22652450e-01 -6.03028715e-01 6.01466656e-01 3.46591026e-01
-1.04482234e+00 1.25771955e-01 -4.37934577e-01 -2.05137074e-01
2.08951190e-01 7.54757226e-01 -8.31823528e-01 8.29177141e-01
4.67898935e-01 4.58540469e-01 -8.74952674e-01 1.51779628e+00
-2.38277495e-01 3.64987925e-02 -2.33842865e-01 1.05055556e-01
2.41143480e-01 -5.65906465e-01 9.87470150e-01 1.33027172e+00
2.79452652e-01 -8.76224935e-02 -1.37656987e-01 7.54529655e-01
4.38052833e-01 7.05443472e-02 -9.90031421e-01 4.68124241e-01
-1.39918521e-01 1.32145941e+00 -1.23155820e+00 -2.71943986e-01
-4.26661313e-01 1.16236222e+00 -9.17041525e-02 3.32232267e-01
-4.06594247e-01 -4.42503095e-01 1.34449989e-01 3.46055388e-01
6.20757401e-01 -4.16325778e-01 -4.70195740e-01 -1.03303003e+00
2.55897820e-01 -6.50915504e-01 1.09040432e-01 -1.04594612e+00
-7.93919623e-01 5.10356009e-01 3.39856222e-02 -1.44571066e+00
1.29183400e-02 -1.09035850e+00 -6.99863434e-01 8.40819061e-01
-1.52743399e+00 -8.02841842e-01 -5.12288749e-01 5.67450523e-01
3.30909759e-01 1.10032432e-01 9.04665947e-01 -1.55400380e-01
2.42788978e-02 -1.28884956e-01 5.36070704e-01 -1.87821910e-01
3.12787592e-01 -1.80139482e+00 3.05020839e-01 1.31938231e+00
5.78729212e-01 6.72013462e-01 1.35576987e+00 -2.99748152e-01
-1.30026233e+00 -4.85893041e-01 8.96271884e-01 -9.59866285e-01
4.71620977e-01 -3.55594039e-01 -9.51969385e-01 5.69268405e-01
3.45679045e-01 -2.43360564e-01 4.39219415e-01 2.00312912e-01
-1.60080954e-01 1.26422450e-01 -1.03988922e+00 4.91737992e-01
8.94729197e-01 -2.67196625e-01 -9.43041801e-01 4.17575032e-01
3.68515760e-01 -2.15836838e-01 -4.16843951e-01 1.01402000e-01
4.29073483e-01 -1.28297698e+00 1.32819545e+00 -2.78064430e-01
7.57019743e-02 -5.25669515e-01 -3.16284299e-01 -1.14813042e+00
-1.64097965e-01 -7.28141665e-01 -9.12776624e-04 8.74079227e-01
2.36660000e-02 -5.05158722e-01 8.91590297e-01 -1.34517208e-01
-2.31803194e-01 -4.50854987e-01 -9.69928443e-01 -9.75249112e-01
-1.09039634e-01 -6.33101821e-01 6.00335956e-01 8.70254993e-01
5.39305881e-02 1.66926384e-01 -2.43827268e-01 5.10670602e-01
1.00325048e+00 4.43467647e-01 1.04179943e+00 -1.65927434e+00
-3.60144883e-01 -4.24413413e-01 -6.49197757e-01 -9.83139515e-01
-4.46663886e-01 -6.95869982e-01 -2.17146873e-01 -1.73018658e+00
1.20852023e-01 -1.85930341e-01 4.06730026e-02 -7.61227310e-02
-5.84069937e-02 4.51906651e-01 4.53065932e-01 2.79748470e-01
-4.73985285e-01 1.53234387e-02 1.20878088e+00 -1.47711173e-01
-8.66804868e-02 3.44750643e-01 -6.31066382e-01 9.53129351e-01
4.95323211e-01 -4.05677259e-01 -2.13131711e-01 -3.37710410e-01
4.51343298e-01 -2.05166474e-01 8.14229608e-01 -8.23614419e-01
7.54708722e-02 8.59982241e-03 3.32863092e-01 -1.07850850e+00
6.18745387e-01 -1.28280914e+00 -2.27991398e-02 4.04028088e-01
2.61819005e-01 1.43906638e-01 -1.25456005e-01 5.43940842e-01
-3.93410586e-02 -1.05768585e+00 5.58740735e-01 -3.82580161e-01
-6.21692717e-01 -1.97289884e-01 -2.59615332e-01 -5.24999797e-01
1.08548450e+00 -8.85113716e-01 -4.48354214e-01 -3.47441167e-01
-9.69553113e-01 -3.45028162e-01 1.08467340e+00 4.60139699e-02
7.18445957e-01 -8.96301270e-01 -5.23742855e-01 1.29397035e-01
2.38662824e-01 -1.48940310e-01 -8.38915184e-02 7.67860591e-01
-8.23162854e-01 4.69450206e-01 7.99583197e-02 -8.84980321e-01
-1.83151913e+00 8.65771115e-01 2.38493234e-01 -4.85477485e-02
-1.02318525e+00 6.42535508e-01 7.31921494e-02 -5.22811292e-03
1.36523679e-01 -3.68137300e-01 -3.04338068e-01 -4.05104421e-02
7.05112040e-01 3.72106701e-01 3.84454519e-01 -7.10867643e-01
-2.65128404e-01 9.53059912e-01 7.99948573e-02 -4.03057158e-01
1.17534804e+00 -2.11257890e-01 -2.34991416e-01 5.03648877e-01
9.62830842e-01 6.00321531e-01 -7.94416964e-01 -2.27810889e-01
-8.13112706e-02 -1.00571644e+00 6.84856847e-02 -4.45800126e-01
-7.37516284e-01 1.01191080e+00 5.47110736e-01 1.20530152e+00
1.21315718e+00 1.68146059e-01 1.07763946e-01 4.51581448e-01
3.42353046e-01 -8.44658077e-01 9.27102715e-02 2.81295508e-01
1.06494486e+00 -1.14926767e+00 4.99007851e-01 -7.80249774e-01
-3.35508853e-01 1.41964507e+00 -3.21350902e-01 -8.84773135e-02
5.18339932e-01 9.54485983e-02 -2.14266833e-02 -4.25356120e-01
-1.48106039e-01 -5.76809406e-01 1.06381461e-01 1.10413980e+00
1.02784343e-01 -9.68205035e-02 -2.09544867e-01 -1.54421534e-02
-4.14802909e-01 -9.59973633e-02 8.43005121e-01 1.22233069e+00
-9.45246994e-01 -1.15307176e+00 -1.21704972e+00 -3.08375545e-02
-2.81045765e-01 -1.74949393e-02 -6.35321975e-01 1.18940210e+00
1.09709002e-01 9.15655434e-01 -3.40915993e-02 -8.17856193e-02
4.24409926e-01 -2.92099148e-01 7.36211598e-01 -6.09295130e-01
-3.51373017e-01 8.28886963e-03 -2.99344242e-01 -1.98525861e-01
-6.96375966e-01 -8.17150593e-01 -9.58392978e-01 -3.55319589e-01
-4.95522439e-01 2.87150592e-01 1.11748898e+00 7.83344448e-01
-2.47701183e-01 6.20174825e-01 7.52538562e-01 -8.88724029e-01
-2.67348826e-01 -4.91890728e-01 -9.89075005e-01 2.47467637e-01
5.44135451e-01 -6.56810045e-01 -6.37454987e-01 1.84986219e-01] | [10.841329574584961, -2.6225645542144775] |
ca920c60-7773-48c2-bb6f-03b669f46edc | accelerated-functional-brain-aging-in-major | 2205.04871 | null | https://arxiv.org/abs/2205.04871v1 | https://arxiv.org/pdf/2205.04871v1.pdf | Accelerated functional brain aging in major depressive disorder: evidence from a large scale fMRI analysis of Chinese participants | Major depressive disorder (MDD) is one of the most common mental health conditions that has been intensively investigated for its association with brain atrophy and mortality. Recent studies reveal that the deviation between the predicted and the chronological age can be a marker of accelerated brain aging to characterize MDD. However, current conclusions are usually drawn based on structural MRI information collected from Caucasian participants. The universality of this biomarker needs to be further validated by subjects with different ethnic/racial backgrounds and by different types of data. Here we make use of the REST-meta-MDD, a large scale resting-state fMRI dataset collected from multiple cohort participants in China. We develop a stacking machine learning model based on 1101 healthy controls, which estimates a subject's chronological age from fMRI with promising accuracy. The trained model is then applied to 1276 MDD patients from 24 sites. We observe that MDD patients exhibit a $+4.43$ years ($\text{$p$} < 0.0001$, $\text{Cohen's $d$} = 0.35$, $\text{95\% CI}:1.86 - 3.91$) higher brain-predicted age difference (brain-PAD) compared to controls. In the MDD subgroup, we observe a statistically significant $+2.09$ years ($\text{$p$} < 0.05$, $\text{Cohen's $d$} = 0.134483$) brain-PAD in antidepressant users compared to medication-free patients. The statistical relationship observed is further checked by three different machine learning algorithms. The positive brain-PAD observed in participants in China confirms the presence of accelerated brain aging in MDD patients. The utilization of functional brain connectivity for age estimation verifies existing findings from a new dimension. | ['Tao Jia', 'Jiang Qiu', 'Wenyu Chen', 'YunSong Luo'] | 2022-05-08 | null | null | null | null | ['age-estimation', 'age-estimation'] | ['computer-vision', 'miscellaneous'] | [-1.09935842e-01 -3.52332331e-02 -2.53617465e-01 -5.89284301e-01
-4.50244188e-01 2.06086218e-01 2.02229232e-01 2.66314238e-01
-7.65016973e-01 7.36595750e-01 1.00446669e-02 -3.73827726e-01
-1.16601847e-01 -7.86434829e-01 -4.17788953e-01 -4.30057585e-01
-6.95115924e-01 2.52067417e-01 -4.23060477e-01 -1.86338499e-02
4.47033674e-01 2.53542066e-01 -1.03614986e+00 -1.48815721e-01
1.36604583e+00 8.07091415e-01 2.11855054e-01 -1.96519643e-01
2.42840692e-01 -1.49291426e-01 -1.88437030e-01 -4.12545562e-01
-1.96749270e-02 -4.82750118e-01 -3.58011812e-01 -3.33111882e-01
2.93702602e-01 -3.88161212e-01 -1.15304515e-01 1.07033706e+00
7.29227066e-01 -1.00010529e-01 9.99701679e-01 -1.01586616e+00
-6.33144200e-01 7.39404023e-01 -1.21718216e+00 5.48039734e-01
-1.82091189e-03 1.45558462e-01 3.98938149e-01 -9.52195823e-01
4.25427020e-01 9.75963414e-01 7.34504104e-01 5.76688290e-01
-1.18241489e+00 -1.24623704e+00 1.63822442e-01 3.23094845e-01
-1.51586926e+00 -4.28875178e-01 4.51599091e-01 -7.07672060e-01
9.30604935e-01 -2.72751123e-01 8.60239565e-01 8.88644099e-01
6.96514249e-01 -2.64584124e-01 1.23166788e+00 -1.30452245e-01
2.98819304e-01 -3.28749061e-01 3.43719989e-01 7.63169110e-01
5.69027364e-01 -3.02332286e-02 -2.07021862e-01 -3.16366434e-01
6.27264738e-01 -5.45312054e-02 7.13177444e-03 -5.29025793e-02
-9.75292325e-01 7.37322211e-01 3.39635491e-01 4.31569666e-01
-4.37929273e-01 -1.19206138e-01 5.82452118e-01 9.57417395e-03
7.84276307e-01 -9.30736735e-02 -5.83236396e-01 -1.30892321e-02
-8.80132556e-01 1.53277099e-01 -3.92473973e-02 1.52235419e-01
3.48216981e-01 4.77591008e-01 2.25746959e-01 1.24548972e+00
3.96896809e-01 8.10241342e-01 6.65643811e-01 -8.72281849e-01
4.24351752e-01 7.22828925e-01 -3.00161511e-01 -9.47206259e-01
-9.23655212e-01 -4.67778534e-01 -1.24718916e+00 1.38382018e-01
4.80559886e-01 -3.84974539e-01 -4.35159147e-01 2.28363776e+00
-6.57194555e-02 -1.51360452e-01 -3.44985783e-01 7.44996130e-01
1.03443623e-01 -3.77692580e-02 4.68998849e-01 -5.36239803e-01
1.63326633e+00 -1.43097937e-01 -3.01371478e-02 -6.04241192e-01
6.76321507e-01 6.39247298e-02 9.04047966e-01 4.05502528e-01
-1.32101285e+00 -5.51685274e-01 -8.76059949e-01 3.38107914e-01
-1.84524536e-01 2.38551423e-01 5.36479175e-01 1.04230368e+00
-9.30555046e-01 5.35826623e-01 -1.03692210e+00 -5.15365899e-01
7.14895368e-01 3.61254066e-01 -3.49853277e-01 -2.17732891e-01
-1.18694437e+00 1.06045294e+00 -5.44765890e-02 6.39044717e-02
-5.73718905e-01 -9.72977400e-01 -6.49751723e-01 -1.01222835e-01
-2.00003415e-01 -8.08353186e-01 4.52869892e-01 -7.59567976e-01
-6.92725480e-01 1.17712474e+00 -3.29144388e-01 -3.86715889e-01
3.53029445e-02 -8.03744793e-02 -7.10979044e-01 3.49368781e-01
4.59272683e-01 6.50902212e-01 5.94369769e-01 -5.63839316e-01
-7.19914958e-02 -1.58599520e+00 -4.42388594e-01 2.39191521e-02
-3.21836174e-01 2.27906108e-01 3.87557685e-01 -5.15644133e-01
4.58260685e-01 -9.28072572e-01 -2.99540937e-01 3.82572934e-02
-1.38544351e-01 -6.19089790e-02 1.47501096e-01 -1.13893545e+00
1.38158023e+00 -1.71674192e+00 -4.26194012e-01 1.82679072e-01
5.88708758e-01 -1.01452522e-01 2.01603904e-01 -1.04774870e-01
-5.11002183e-01 5.12518466e-01 -5.71054101e-01 1.40094414e-01
-1.31469980e-01 -5.72422624e-01 4.35031712e-01 9.58897352e-01
1.55840203e-01 8.05101514e-01 -4.70374376e-01 -2.21934661e-01
-2.46978879e-01 4.73958760e-01 -5.40556014e-01 -4.93559003e-01
6.04687512e-01 1.97126865e-01 -2.97249287e-01 5.45724094e-01
1.01159573e+00 -8.51196572e-02 4.44770962e-01 9.82694477e-02
-3.86494637e-01 -3.28227766e-02 -5.81607223e-01 1.73089719e+00
2.22629189e-01 3.37735176e-01 3.12612325e-01 -1.43626750e+00
1.09124446e+00 8.08240250e-02 4.67623413e-01 -1.07812929e+00
2.85450041e-01 2.70864576e-01 1.02269745e+00 -1.72016665e-01
-1.77771717e-01 -4.41236883e-01 -2.97576450e-02 5.50166368e-01
-3.92346561e-01 2.86191136e-01 9.00723562e-02 1.01814702e-01
1.27136838e+00 -2.92656660e-01 1.36505246e-01 -9.78257120e-01
5.44733286e-01 -3.04874986e-01 8.83309066e-01 3.56645465e-01
-6.23435438e-01 1.59812018e-01 8.62976491e-01 -1.05536312e-01
-9.33292687e-01 -1.14519095e+00 -7.84883499e-01 4.95391756e-01
-4.78913665e-01 6.56596292e-03 -8.25576901e-01 -4.74400550e-01
2.21446306e-01 6.76579654e-01 -7.05804944e-01 -5.70583761e-01
-2.49659717e-01 -1.44881785e+00 6.58001304e-01 6.29950225e-01
7.53027618e-01 -6.45848215e-01 -5.92289567e-01 2.34630778e-02
1.82705186e-02 -5.74611366e-01 -2.22839445e-01 -1.74457461e-01
-1.36798441e+00 -1.02941990e+00 -1.01823068e+00 -3.50823373e-01
9.25759912e-01 3.94966528e-02 9.49922204e-01 1.30031168e-01
-4.39341307e-01 1.65249243e-01 6.56916723e-02 -3.82462651e-01
1.15601085e-01 9.27550346e-02 6.41511500e-01 -3.10008794e-01
6.31945848e-01 -8.75345051e-01 -1.26490569e+00 1.31773978e-01
-4.64306027e-01 -1.72562569e-01 6.25746608e-01 4.90837812e-01
2.31841087e-01 -2.71081239e-01 1.26699865e+00 -5.48766375e-01
2.42782325e-01 -8.69254112e-01 -1.73070371e-01 -7.24091083e-02
-1.22788560e+00 -3.18748593e-01 1.69336826e-01 -3.29689533e-01
-9.99135911e-01 -6.29538953e-01 -8.31330270e-02 2.25051284e-01
-2.64514387e-01 7.30137289e-01 -2.41544068e-01 7.05393791e-01
4.38394845e-01 -1.58311874e-01 3.28965336e-01 -3.67824435e-01
-2.92664289e-01 4.02772367e-01 9.33596939e-02 -6.60902500e-01
9.35044512e-02 3.76708657e-01 1.28246009e-01 -9.37804699e-01
-5.67707084e-02 8.07151869e-02 -7.23721862e-01 -6.89547434e-02
8.16323042e-01 -1.04172719e+00 -8.50971520e-01 6.16923094e-01
-4.75681722e-01 -5.28746903e-01 6.44647300e-01 9.64287639e-01
-1.25008821e-01 3.75343233e-01 -5.58071315e-01 -6.54296696e-01
-9.01263535e-01 -8.96085918e-01 3.42200637e-01 8.22761804e-02
-5.91933608e-01 -7.73721993e-01 7.93012679e-02 4.76874381e-01
3.25424433e-01 4.16822553e-01 1.56480002e+00 -4.58475083e-01
4.04994547e-01 -2.25494549e-01 -6.07533991e-01 1.26604825e-01
1.91631988e-01 -6.81003630e-02 -3.55830371e-01 -1.42060190e-01
1.93272755e-02 5.35429157e-02 6.21860802e-01 9.57052410e-01
1.08032429e+00 2.21339837e-01 -4.04947996e-01 -6.29047826e-02
1.20138800e+00 7.34301090e-01 9.48159575e-01 1.11955598e-01
3.00967604e-01 6.66420281e-01 1.31270066e-01 5.90681314e-01
4.15162742e-01 3.96102458e-01 1.84131861e-01 1.82158798e-01
2.94816017e-01 3.86660457e-01 4.89659011e-01 4.03808802e-01
-2.71351725e-01 4.50279415e-01 -1.23674202e+00 5.42450428e-01
-1.14473307e+00 -1.05004632e+00 -6.29071891e-01 2.42618108e+00
6.24975026e-01 2.25960195e-01 4.21536297e-01 -4.48188819e-02
1.00128937e+00 -1.14434332e-01 -9.87846494e-01 -6.45979702e-01
-9.70337093e-02 6.85465515e-01 -7.91079830e-03 -7.49362409e-02
-3.49483073e-01 2.44056433e-01 4.85762453e+00 2.38775417e-01
-1.21790695e+00 2.30192393e-01 1.24876726e+00 -4.48370129e-01
1.16888590e-01 -1.83016546e-02 -3.61050874e-01 6.56208694e-01
1.55372787e+00 -5.16692638e-01 1.64892673e-01 5.22322953e-01
7.81271994e-01 -7.65073717e-01 -7.12658703e-01 8.37459922e-01
2.50182927e-01 -8.84858072e-01 -5.44730484e-01 3.93611014e-01
3.36467743e-01 -8.13194439e-02 2.63928086e-01 3.98502499e-01
-1.77008554e-01 -9.26327765e-01 4.21801448e-01 6.80163145e-01
1.34942746e+00 -7.30443358e-01 4.15104479e-01 3.94451953e-02
-1.00487077e+00 -4.52126414e-02 -1.67271703e-01 -5.54512560e-01
2.11121783e-01 1.25003600e+00 -4.64936286e-01 2.59469688e-01
9.21718955e-01 5.96722186e-01 -6.31748855e-01 5.24818838e-01
2.27273315e-01 4.55789864e-01 -2.84473803e-02 4.54298466e-01
-1.96669415e-01 -5.94047070e-01 -1.15989640e-01 6.76631331e-01
8.42331529e-01 4.62830424e-01 -5.28941572e-01 1.17110586e+00
2.99492944e-03 1.82012990e-01 -3.78899604e-01 -1.75974011e-01
3.14972997e-01 1.27992892e+00 -1.01406717e+00 -4.34778810e-01
-5.32693028e-01 5.23346603e-01 2.22738504e-01 4.17379402e-02
-7.90557265e-01 -4.19245929e-01 5.99984884e-01 6.05959237e-01
-3.21413785e-01 -3.34947795e-01 -7.16302454e-01 -9.78792846e-01
-1.24679111e-01 -5.46909928e-01 2.26318404e-01 -7.50262141e-01
-1.23563683e+00 2.11721614e-01 1.48786709e-01 -5.52293181e-01
-3.61584648e-02 -3.34422976e-01 -7.31140137e-01 1.36832213e+00
-8.02195251e-01 -4.24432814e-01 -1.04890011e-01 3.91575456e-01
7.24716932e-02 -5.15244119e-02 8.38509500e-01 3.50036442e-01
-1.25847912e+00 3.43756378e-01 -5.95053397e-02 6.24172017e-02
8.18340778e-01 -7.96731234e-01 -6.13078661e-02 6.87540948e-01
-8.02214980e-01 1.06659007e+00 3.16291526e-02 -1.18049502e+00
-9.73326623e-01 -7.36951530e-01 1.04030526e+00 -2.47859672e-01
5.30807734e-01 -1.65404916e-01 -6.62950337e-01 3.95402372e-01
9.04762670e-02 -4.53034937e-01 9.42167282e-01 3.27290297e-01
-2.99215078e-01 -1.87617660e-01 -1.58276808e+00 5.51571667e-01
1.29328537e+00 -2.81008005e-01 -4.04795915e-01 -2.50116408e-01
-5.86207211e-02 4.54269439e-01 -1.35589826e+00 5.91028988e-01
1.02362084e+00 -1.12498748e+00 9.74648595e-01 -3.36486995e-01
5.97523034e-01 3.46671253e-01 1.67357624e-01 -1.02333748e+00
-3.46339077e-01 2.82626092e-01 2.96820104e-01 1.27550089e+00
4.36043590e-01 -7.25657701e-01 7.70656168e-01 1.29034138e+00
-2.95311779e-01 -7.00482130e-01 -1.07510185e+00 -6.56657636e-01
7.22439408e-01 -3.65088940e-01 3.24332774e-01 9.97771323e-01
2.85289973e-01 6.96243867e-02 3.11004311e-01 1.28376603e-01
6.02207899e-01 -4.76424128e-01 4.10805084e-02 -1.49093485e+00
3.93110067e-01 -5.17935932e-01 -2.39576846e-02 9.77458507e-02
3.51935089e-01 -8.51904809e-01 -5.99749982e-01 -1.71211410e+00
6.63677931e-01 -4.81916726e-01 -4.08558011e-01 5.68422556e-01
-6.72062859e-02 -5.55069633e-02 -3.94775391e-01 -2.31966600e-01
9.65642333e-02 4.97559845e-01 6.52666450e-01 -1.70527190e-01
-2.08103746e-01 -3.50815326e-01 -9.81583178e-01 7.06751406e-01
1.35632515e+00 -3.48352790e-01 -4.45243180e-01 -1.80185959e-01
8.94059166e-02 1.65415123e-01 3.45243633e-01 -1.03430223e+00
-5.41257918e-01 -2.57186174e-01 1.03495431e+00 -3.80013078e-01
3.18074971e-01 -2.51567066e-01 2.75937796e-01 1.24841762e+00
2.29960427e-01 6.69196308e-01 2.17338830e-01 -5.87285273e-02
6.01125062e-01 -1.05534904e-01 7.94094205e-01 -1.92461982e-01
-7.68282935e-02 3.88642430e-01 -7.40104854e-01 2.84081567e-02
7.69155502e-01 -2.89603651e-01 -4.10991132e-01 5.80873862e-02
-6.65512145e-01 -1.94027543e-01 2.93414742e-01 1.43031195e-01
5.04858375e-01 -1.21778572e+00 -7.81911373e-01 -2.25225259e-02
-7.17845485e-02 -7.87979722e-01 1.02907288e+00 1.62212932e+00
-1.33769810e-01 3.38112473e-01 -6.18319690e-01 -3.17469016e-02
-9.85575676e-01 4.54856157e-01 1.61241829e-01 -1.25189871e-01
-3.88886899e-01 4.70153898e-01 1.11102596e-01 1.49234682e-01
-3.60324264e-01 -1.83485538e-01 -1.54041336e-03 5.76440275e-01
5.56561112e-01 9.47179437e-01 4.82120402e-02 -6.63138330e-01
-7.15430915e-01 2.77851582e-01 -2.40967825e-01 -3.65829766e-01
1.52172458e+00 -3.44520271e-01 -5.98518789e-01 3.67309570e-01
1.10380173e+00 -4.37084511e-02 -8.91040087e-01 2.22408012e-01
-6.07888438e-02 -7.63853341e-02 -1.24986961e-01 -7.08824456e-01
-1.36198831e+00 8.03055942e-01 1.34144747e+00 -3.55469733e-01
8.54283094e-01 -6.17414117e-02 5.12974501e-01 -1.65383741e-01
4.91825759e-01 -1.20321512e+00 -1.70224026e-01 2.06660613e-01
7.99299717e-01 -9.05420423e-01 1.23773642e-01 3.51700038e-01
-4.07684147e-01 6.53794885e-01 8.88471425e-01 -2.84474641e-01
9.29858029e-01 -2.96782315e-01 -3.05905133e-01 -3.58191103e-01
-2.45833069e-01 1.02195926e-01 3.75148328e-03 7.02770472e-01
7.17717111e-01 2.02429190e-01 -1.31282020e+00 1.41379583e+00
-3.01588863e-01 -5.51793985e-02 4.27005172e-01 6.07767761e-01
-2.11870208e-01 -1.15914047e+00 -3.91332060e-01 1.21141219e+00
-7.27597177e-01 -2.91122317e-01 -2.38070980e-01 8.31800282e-01
4.31221992e-01 9.10816729e-01 4.45666790e-01 2.46039964e-02
2.26343632e-01 6.99750066e-01 5.07523179e-01 -5.35723984e-01
-3.91225845e-01 -1.55132368e-01 1.58833489e-02 -1.33967936e-01
-4.55805391e-01 -1.13899529e+00 -1.56200945e+00 -5.25995910e-01
1.66680306e-01 3.79638635e-02 6.70195818e-01 6.90876305e-01
7.16006756e-01 2.83812046e-01 3.73802871e-01 -6.09825969e-01
8.48174915e-02 -1.21449423e+00 -1.10769618e+00 2.56227534e-02
-4.36201960e-01 -9.98608112e-01 -1.71240106e-01 -1.47982061e-01] | [14.106755256652832, -1.547048807144165] |
a887fba9-a137-4e18-ac34-1d5bb04a89cb | combinatory-chemistry-towards-a-simple-model | 2003.07916 | null | https://arxiv.org/abs/2003.07916v2 | https://arxiv.org/pdf/2003.07916v2.pdf | Combinatory Chemistry: Towards a Simple Model of Emergent Evolution | An explanatory model for the emergence of evolvable units must display emerging structures that (1) preserve themselves in time (2) self-reproduce and (3) tolerate a certain amount of variation when reproducing. To tackle this challenge, here we introduce Combinatory Chemistry, an Algorithmic Artificial Chemistry based on a minimalistic computational paradigm named Combinatory Logic. The dynamics of this system comprise very few rules, it is initialised with an elementary tabula rasa state, and features conservation laws replicating natural resource constraints. Our experiments show that a single run of this dynamical system with no external intervention discovers a wide range of emergent patterns. All these structures rely on acquiring basic constituents from the environment and decomposing them in a process that is remarkably similar to biological metabolisms. These patterns include autopoietic structures that maintain their organisation, recursive ones that grow in linear chains or binary-branching trees, and most notably, patterns able to reproduce themselves, duplicating their number at each generation. | ['Germán Kruszewski', 'Tomas Mikolov'] | 2020-03-17 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 3.81042242e-01 4.15592045e-01 6.02162033e-02 4.38270390e-01
1.02389967e+00 -1.05125487e+00 1.11954224e+00 5.38999774e-02
1.04884870e-01 1.07511306e+00 -2.34792784e-01 -3.74401987e-01
-1.83283865e-01 -1.03117466e+00 -6.87986851e-01 -1.13004708e+00
-4.58548456e-01 6.48603380e-01 3.84780258e-01 -5.89721084e-01
2.07086559e-02 6.40059710e-01 -1.94924057e+00 1.01852175e-02
9.43630278e-01 6.03613019e-01 1.85299441e-01 8.07559967e-01
4.80334572e-02 8.93389821e-01 -1.19624466e-01 -7.32160956e-02
3.64144474e-01 -1.05654836e+00 -5.61590075e-01 2.28075773e-01
-4.77922320e-01 4.28375155e-01 -1.58427715e-01 6.75336301e-01
-1.04572237e-01 2.34097261e-02 9.76337850e-01 -9.02139008e-01
-8.25431347e-01 6.81904852e-01 5.17127886e-02 -1.38708398e-01
1.66438326e-01 6.80987716e-01 7.57050395e-01 -4.53197062e-01
9.98485804e-01 9.97372568e-01 3.90873373e-01 8.74870479e-01
-1.57984841e+00 -2.73693129e-02 -1.29893571e-01 -1.71884313e-01
-1.26150346e+00 -5.83973587e-01 2.57915735e-01 -4.78098780e-01
1.10485196e+00 7.10650444e-01 1.34723341e+00 1.08172107e+00
9.51070189e-01 -6.77418560e-02 1.44086003e+00 -6.15011036e-01
7.97053218e-01 2.16493942e-02 -3.67233574e-01 8.76285732e-01
8.25726390e-01 4.03725177e-01 -6.56984210e-01 -4.77330759e-02
7.69068122e-01 -1.00059703e-01 -2.10238889e-01 -5.49951851e-01
-1.18797290e+00 2.52072245e-01 5.12211435e-02 7.46353984e-01
-3.94439846e-01 3.06149840e-01 -2.60118157e-01 4.06808108e-01
-8.55701566e-02 8.45957160e-01 -7.64573455e-01 1.26482800e-01
-2.91566610e-01 -8.96213353e-02 1.19486129e+00 7.83452392e-01
8.19105446e-01 3.30782719e-02 3.81592423e-01 1.90705955e-01
1.31637424e-01 4.91807729e-01 7.07231283e-01 -7.60648191e-01
-7.48065174e-01 9.56084847e-01 -1.05551995e-01 -7.29265511e-01
-4.60445106e-01 -6.51072443e-01 -1.11722398e+00 4.60327595e-01
2.39552826e-01 -1.63032755e-01 -8.24768543e-01 1.80131459e+00
4.13569540e-01 -1.83243141e-01 1.10979125e-01 3.48486751e-01
2.25212649e-01 8.72573495e-01 -1.05686300e-01 -7.47098684e-01
1.26396251e+00 -5.70907414e-01 -5.48371673e-01 2.90592074e-01
2.23431423e-01 -2.28495926e-01 6.67727530e-01 3.98006260e-01
-1.32504630e+00 -1.36242494e-01 -1.22128785e+00 3.85743350e-01
-6.85607433e-01 -4.90810156e-01 9.75425661e-01 6.51906550e-01
-1.37256193e+00 9.08123136e-01 -8.95822942e-01 -7.53519475e-01
-1.52837411e-01 3.04746449e-01 -3.96842152e-01 6.99523568e-01
-8.76616955e-01 9.11308706e-01 6.28372729e-01 -9.01151225e-02
-1.19386029e+00 -5.75970531e-01 -3.91466141e-01 1.72844931e-01
4.48019207e-01 -1.38856995e+00 8.10451269e-01 -1.00885570e+00
-1.97722900e+00 8.79678190e-01 -1.13643318e-01 -3.06824684e-01
4.10222113e-01 6.55964911e-01 -2.22113967e-01 -1.18765041e-01
-3.34078372e-01 4.64296877e-01 8.38959634e-01 -1.41834724e+00
-4.08171713e-01 -2.00346783e-01 -2.20544294e-01 -1.00620337e-01
-1.69561610e-01 -3.83595020e-01 3.00471216e-01 -3.53444815e-01
2.72002399e-01 -1.01769710e+00 -5.48972309e-01 -2.05780521e-01
-4.91195321e-01 -7.72952363e-02 3.83514017e-01 2.91724741e-01
1.07505608e+00 -1.68531895e+00 8.07183981e-01 2.70152330e-01
5.71750402e-01 -1.89580008e-01 1.69286236e-01 1.10822296e+00
7.68268406e-02 5.54169059e-01 -5.00967383e-01 3.06453675e-01
-1.12539511e-02 4.07332689e-01 -1.13050126e-01 2.04032615e-01
1.75077200e-01 9.97755647e-01 -8.54327381e-01 -1.83591262e-01
-1.98160455e-01 1.10126823e-01 -5.83190978e-01 -7.17422217e-02
-6.69069171e-01 5.46803176e-01 -3.50816458e-01 6.08115435e-01
2.48553097e-01 -4.88008738e-01 6.58453524e-01 5.06976008e-01
-8.85717392e-01 2.12325272e-03 -7.94527471e-01 1.10593998e+00
-7.49064386e-02 3.37497950e-01 1.58747703e-01 -6.71237051e-01
7.65504599e-01 1.26221195e-01 4.85904932e-01 -3.55233282e-01
2.51730710e-01 4.44102675e-01 6.54620826e-01 -2.05046490e-01
1.19331539e-01 -2.34074458e-01 5.51306047e-02 7.53617227e-01
1.33001894e-01 -4.60363716e-01 4.65273976e-01 2.04111531e-01
1.48617387e+00 1.05591789e-02 8.01140308e-01 -8.36643934e-01
6.72082007e-01 6.37762845e-02 5.91099620e-01 9.14554179e-01
1.25378653e-01 1.83279619e-01 5.55742145e-01 -8.41822743e-01
-1.47554743e+00 -1.08149171e+00 -2.20081151e-01 6.90939724e-01
2.04091206e-01 -4.65896159e-01 -7.84443319e-01 2.39202753e-01
-1.81766257e-01 4.88735348e-01 -1.01041007e+00 -5.02298117e-01
-3.45741987e-01 -8.93975317e-01 1.09815739e-01 -4.36608076e-01
3.58016819e-01 -1.54714060e+00 -1.18923128e+00 4.13630277e-01
3.95810455e-01 -4.08497483e-01 2.46657133e-02 7.90358126e-01
-8.66225123e-01 -9.53609586e-01 -7.06268698e-02 -4.60916936e-01
8.85237873e-01 -1.49353832e-01 1.12075067e+00 7.10533440e-01
-6.93569243e-01 7.96473250e-02 -3.76677245e-01 -2.88211644e-01
-9.51966703e-01 1.56870186e-01 5.27506948e-01 -1.29746735e-01
-3.39116096e-01 -1.34443533e+00 -4.14295495e-01 -4.23116289e-04
-1.09522939e+00 3.00074846e-01 5.97727537e-01 7.94200182e-01
5.23401439e-01 3.93071622e-01 3.87832493e-01 -5.07621467e-01
5.58047295e-01 -4.71460581e-01 -4.31143284e-01 5.41365504e-01
-8.87011528e-01 3.44774514e-01 9.64938521e-01 -3.69754642e-01
-7.27110684e-01 1.87340975e-01 5.32275259e-01 3.40063661e-01
-1.12711288e-01 2.27529496e-01 -2.42531776e-01 -9.93094593e-02
6.80733085e-01 9.15013731e-01 2.52556235e-01 -1.52678385e-01
4.32373881e-01 2.27900535e-01 5.09356141e-01 -7.92667627e-01
1.01966488e+00 5.68893731e-01 5.14091015e-01 -9.75407422e-01
1.72830820e-02 5.42521894e-01 -9.54965413e-01 -3.07392389e-01
7.49090970e-01 -2.03754783e-01 -1.06313372e+00 5.40262818e-01
-7.98158407e-01 -4.15765494e-01 -9.43386495e-01 -4.31640558e-02
-6.95991576e-01 5.52665591e-02 -4.19347256e-01 -1.04765296e+00
-1.59661993e-01 -4.16084260e-01 3.93054605e-01 4.79833096e-01
-2.84769833e-01 -7.72663295e-01 7.00807214e-01 -5.37722468e-01
7.47065008e-01 6.45075381e-01 1.36378419e+00 -4.23026383e-01
-9.58985686e-01 3.62710625e-01 6.10425889e-01 -4.74215329e-01
4.35605675e-01 6.73290133e-01 -3.60541224e-01 -7.88068622e-02
-8.39909390e-02 1.22543059e-01 7.11617529e-01 5.51903211e-02
3.31047833e-01 -5.28706551e-01 -4.68240708e-01 5.81429720e-01
1.32239485e+00 7.86784887e-01 5.42264104e-01 4.63716835e-01
2.19952464e-02 8.31158638e-01 -4.12497699e-01 3.86463940e-01
6.96214139e-02 1.62685007e-01 5.13765275e-01 3.11263889e-01
1.16274439e-01 -1.92413524e-01 3.88746798e-01 1.11044800e+00
-6.01053774e-01 -1.74557537e-01 -7.78591454e-01 2.88631320e-01
-1.82461751e+00 -1.21554339e+00 -1.94253474e-01 1.98880970e+00
1.14724994e+00 8.85083452e-02 1.06781922e-01 -2.98630148e-02
4.33958501e-01 -2.74424911e-01 -8.19136739e-01 -7.52873242e-01
-7.26325095e-01 4.30036455e-01 2.67525136e-01 2.86144078e-01
-5.39544761e-01 1.19545472e+00 7.91716623e+00 3.15817475e-01
-9.59289312e-01 -1.08045623e-01 3.63571316e-01 -2.41482574e-02
-6.04809165e-01 5.98978221e-01 -2.57795215e-01 4.46093619e-01
1.09785330e+00 -6.19398773e-01 1.03425932e+00 3.87951374e-01
1.77613154e-01 -2.29302302e-01 -9.06122923e-01 3.60701144e-01
-4.19982374e-01 -1.61124313e+00 2.71946400e-01 3.03433329e-01
1.02571237e+00 -4.47656363e-01 -1.81185246e-01 -5.98735809e-02
8.77666712e-01 -1.32181048e+00 8.76806080e-01 8.60240698e-01
6.65746808e-01 -4.38670248e-01 -4.99505410e-03 6.84102833e-01
-8.56438100e-01 -2.43052647e-01 -4.34488803e-01 -5.99071562e-01
3.42421234e-02 4.76589710e-01 -3.23833048e-01 3.80024672e-01
5.32636762e-01 3.51970434e-01 -5.75205982e-01 8.21752191e-01
-2.04597697e-01 2.15162233e-01 -3.44971508e-01 -5.57954967e-01
-7.59423971e-02 -8.32338214e-01 8.97003710e-01 7.15706050e-01
4.45173860e-01 5.29634953e-01 -4.08239514e-01 1.15486681e+00
2.05780223e-01 -1.83592796e-01 -8.21614623e-01 -2.51574308e-01
3.97635221e-01 1.15226007e+00 -1.22818732e+00 -3.71616006e-01
3.62006307e-01 7.56849527e-01 -3.35323922e-02 7.78169855e-02
-4.54878986e-01 -1.11212902e-01 6.63266718e-01 1.48953840e-01
2.55283237e-01 -4.28738326e-01 -2.90161639e-01 -1.19470119e+00
-3.68643612e-01 -8.74946177e-01 -2.14853987e-01 -6.79557681e-01
-9.16810751e-01 6.92592323e-01 -3.82259518e-01 -6.68637991e-01
-1.96063206e-01 -4.23251212e-01 -7.70296812e-01 1.93169132e-01
-7.37589180e-01 -8.07894289e-01 -1.13445379e-01 4.05224890e-01
1.36562154e-01 -3.90943676e-01 8.83821905e-01 -5.79238832e-01
-6.88258827e-01 8.61560330e-02 4.23769146e-01 -7.13747084e-01
-4.04200442e-02 -1.29742289e+00 2.12031230e-01 8.27031255e-01
7.43315518e-02 9.77498889e-01 9.95270610e-01 -8.25652897e-01
-1.74497914e+00 -7.40574837e-01 9.51419234e-01 -4.81480539e-01
7.36423790e-01 -6.87346280e-01 -6.14595830e-01 4.06827122e-01
5.48605084e-01 -4.96120661e-01 4.60208565e-01 -7.21342936e-02
-1.13261476e-01 1.33038566e-01 -1.04930055e+00 1.06133592e+00
1.71495450e+00 5.68158664e-02 -4.05606538e-01 3.43720645e-01
7.79515088e-01 1.10714860e-01 -5.45058489e-01 3.63456845e-01
9.02443111e-01 -1.19710243e+00 4.98630315e-01 -9.21637475e-01
5.23147941e-01 -5.90519667e-01 2.35159636e-01 -1.25219369e+00
-8.52176070e-01 -1.32018554e+00 -1.44938841e-01 7.73690820e-01
5.82837880e-01 -1.01590562e+00 1.28398001e-01 3.17025423e-01
-6.24333285e-02 -5.01203001e-01 -9.29434180e-01 -1.09768045e+00
1.61001354e-01 6.23153329e-01 8.01725447e-01 9.63837981e-01
3.43874514e-01 2.28602007e-01 -1.09830424e-01 -3.27475280e-01
5.09278953e-01 1.37990922e-01 5.00424623e-01 -1.41915393e+00
-3.42383832e-01 -7.62829065e-01 -3.65859419e-01 -4.10271645e-01
-1.65337488e-01 -9.25305188e-01 -5.13882004e-02 -9.65601146e-01
3.53489637e-01 -3.28878284e-01 -2.70555960e-03 3.92294139e-01
4.27122504e-01 -2.32914388e-01 4.16125767e-02 5.05237818e-01
-4.87145543e-01 6.29851401e-01 1.18419909e+00 1.58776492e-01
-3.87249738e-01 -4.74591821e-01 -8.92631352e-01 4.88202810e-01
7.77544320e-01 -5.67676485e-01 -3.23348105e-01 2.94905812e-01
8.08185697e-01 -1.49097502e-01 8.24879557e-02 -9.20376241e-01
3.26689512e-01 -5.56790411e-01 2.29019955e-01 -7.42952013e-03
-1.65602490e-01 -7.49810755e-01 1.08440578e+00 1.22913301e+00
-9.60793495e-02 1.73606262e-01 -1.36375561e-01 6.06448770e-01
4.25973952e-01 -3.35411966e-01 8.14630628e-01 -4.52266216e-01
-9.56653804e-02 7.42700696e-02 -1.29653931e+00 -3.82742137e-01
1.53023231e+00 -3.93315375e-01 -5.09631038e-01 -1.73674664e-03
-8.70329440e-01 -1.72531360e-03 1.22038555e+00 9.11264867e-02
7.13842288e-02 -1.02199590e+00 -5.59876740e-01 1.29720196e-01
-2.14277297e-01 -2.08224431e-01 -1.60608366e-01 5.74762225e-01
-1.11675417e+00 4.70502377e-01 -7.64700651e-01 -4.85059798e-01
-6.16791844e-01 7.55099654e-01 5.29979289e-01 -2.10615724e-01
-4.99170035e-01 6.64877415e-01 2.01633945e-01 -3.27138960e-01
-5.44343293e-01 -3.06699753e-01 -1.58642203e-01 -1.51392758e-01
2.73042947e-01 8.43238607e-02 -2.84573227e-01 -2.37240687e-01
-4.70647395e-01 5.49583793e-01 4.29787904e-01 -7.52623677e-02
1.64011908e+00 -4.46932912e-02 -1.13918149e+00 6.30730808e-01
1.90275878e-01 2.09412783e-01 -9.22568142e-01 3.22984546e-01
3.39520425e-02 -4.36448911e-03 -7.68521011e-01 -7.44890511e-01
-4.60709661e-01 1.69430584e-01 -6.31513745e-02 9.72310722e-01
1.02620971e+00 6.52593523e-02 2.21439913e-01 6.84184611e-01
8.07106853e-01 -8.46558213e-01 5.41252047e-02 6.47301078e-01
9.16000366e-01 -4.14951026e-01 -5.60901314e-02 -1.96027830e-01
-2.68601980e-02 1.22118104e+00 2.41144881e-01 -3.31164986e-01
6.51475132e-01 4.72112179e-01 -6.32600307e-01 -4.05098379e-01
-1.78092003e+00 -4.01127607e-01 -2.97286302e-01 7.41275549e-01
2.22186685e-01 2.11854234e-01 -8.55837286e-01 4.09082621e-01
-5.72939217e-01 -3.76881123e-01 8.73166263e-01 8.19770694e-01
-8.78362775e-01 -1.25746703e+00 -1.75638095e-01 1.32496888e-02
1.13258825e-03 -1.24732934e-01 -1.21620965e+00 9.09073710e-01
4.82133210e-01 8.48213911e-01 -2.70121470e-02 -3.24982196e-01
-2.69415770e-02 2.00244054e-01 6.49305403e-01 -5.17423332e-01
-7.21680462e-01 -2.30693072e-01 -1.94064379e-01 -1.16463408e-01
-2.02470124e-01 -1.07049739e+00 -1.18293798e+00 -6.36531472e-01
-1.22588687e-01 2.77280420e-01 4.94549960e-01 6.94726706e-01
5.66703439e-01 5.96059918e-01 7.17867196e-01 -5.08336663e-01
-3.28292847e-01 -6.02789700e-01 -7.24498630e-01 6.95426166e-02
1.72766790e-01 -5.35535216e-01 -6.47336245e-01 3.64290476e-01] | [5.606098175048828, 4.154456615447998] |
3cc3e71a-26c0-4556-a006-73c1e8890949 | every-time-i-fire-a-conversational-designer-1 | null | null | https://aclanthology.org/2022.lrec-1.15 | https://aclanthology.org/2022.lrec-1.15.pdf | Every time I fire a conversational designer, the performance of the dialogue system goes down | Incorporating handwritten domain scripts into neural-based task-oriented dialogue systems may be an effective way to reduce the need for large sets of annotated dialogues. In this paper, we investigate how the use of domain scripts written by conversational designers affects the performance of neural-based dialogue systems. To support this investigation, we propose the Conversational-Logic-Injection-in-Neural-Network system (CLINN) where domain scripts are coded in semi-logical rules. By using CLINN, we evaluated semi-logical rules produced by a team of differently-skilled conversational designers. We experimented with the Restaurant domain of the MultiWOZ dataset. Results show that external knowledge is extremely important for reducing the need for annotated examples for conversational systems. In fact, rules from conversational designers used in CLINN significantly outperform a state-of-the-art neural-based dialogue system when trained with smaller sets of annotated dialogues. | ['Fabio Massimo Zanzotto', 'Raniero Romagnoli', 'Andrea Favalli', 'Cristina Giannone', 'Samir Salman', 'Michele Mastromattei', 'Giancarlo Xompero'] | null | null | null | null | lrec-2022-6 | ['task-oriented-dialogue-systems'] | ['natural-language-processing'] | [-5.27141020e-02 7.27103829e-01 3.89146060e-01 -6.98422730e-01
-2.11987257e-01 -6.29925251e-01 7.19899356e-01 -3.33216190e-01
-6.25681520e-01 9.25926805e-01 5.12261212e-01 -6.00616693e-01
7.22494647e-02 -7.16802120e-01 -3.62542063e-01 1.55249566e-01
3.20325255e-01 9.70486939e-01 9.44875479e-02 -9.60295677e-01
1.16421886e-01 8.35491642e-02 -9.82310116e-01 1.02588212e+00
8.81413579e-01 6.19176209e-01 1.74133986e-01 7.82281578e-01
-4.79686469e-01 1.52683568e+00 -1.02916813e+00 -4.68477577e-01
1.72417000e-01 -9.99519408e-01 -1.17943633e+00 1.61198862e-02
2.68020391e-01 -8.18624020e-01 -2.04512313e-01 5.60637176e-01
4.31182861e-01 4.98520583e-01 5.33508360e-01 -7.38162756e-01
-4.53472197e-01 1.31035948e+00 4.72354591e-01 -2.27794483e-01
3.34873408e-01 3.83646697e-01 8.75283420e-01 -6.97226465e-01
8.30621362e-01 1.53125930e+00 7.63351381e-01 1.12116134e+00
-1.34929812e+00 -3.29906642e-01 -8.69725421e-02 -1.24180071e-01
-7.91844547e-01 -7.59851992e-01 7.32655644e-01 -4.00080442e-01
1.44893742e+00 1.49684623e-01 5.47637641e-01 1.21931756e+00
-1.45288602e-01 7.70004809e-01 1.02340949e+00 -7.88764417e-01
2.58817583e-01 5.78649163e-01 2.77137041e-01 8.09062719e-01
-2.61913031e-01 -1.15744285e-01 -7.11950481e-01 -1.80182979e-01
7.68443704e-01 -8.82393777e-01 -2.11759061e-01 1.84239894e-02
-7.88988352e-01 1.26381123e+00 -1.16781913e-01 4.74641979e-01
-2.71661699e-01 -2.37433404e-01 9.41037655e-01 7.68331945e-01
5.53271174e-01 1.37623155e+00 -8.09854746e-01 -5.54569840e-01
-4.26519185e-01 3.15355659e-01 1.79698503e+00 7.75217414e-01
2.23075435e-01 3.13869983e-01 -4.35901493e-01 1.66231596e+00
1.19361609e-01 -1.03749171e-01 6.92737341e-01 -1.29519033e+00
5.87644339e-01 7.12059319e-01 2.22278669e-01 -6.48380458e-01
-4.84088123e-01 3.66894960e-01 -4.82432097e-01 -5.84506243e-02
1.16986609e+00 -9.54410553e-01 -2.31256202e-01 1.48012447e+00
-8.89430493e-02 -9.30821419e-01 3.85891259e-01 7.22567737e-01
9.65344310e-01 7.31030464e-01 -1.92751318e-01 -1.84560582e-01
1.22595513e+00 -1.19553268e+00 -8.80369484e-01 -2.22293913e-01
7.27644205e-01 -5.06876111e-01 1.43579960e+00 6.57821417e-01
-1.12909007e+00 -6.06858909e-01 -6.59844577e-01 4.62627225e-03
-3.70693028e-01 3.14161927e-01 4.34641898e-01 7.67064810e-01
-1.09306931e+00 5.39028585e-01 -2.52841324e-01 -5.73413849e-01
-1.79176673e-01 2.80843496e-01 -7.35756159e-02 2.06274867e-01
-1.59046829e+00 1.41673172e+00 5.39340019e-01 1.33897722e-01
-7.06630290e-01 -2.29595527e-01 -8.79710197e-01 2.47114734e-03
4.62543428e-01 -1.36131451e-01 2.15766144e+00 -1.04585660e+00
-2.23440289e+00 7.34217763e-01 3.14682573e-01 -6.44314885e-01
7.29286253e-01 -3.62306923e-01 -3.31309587e-02 -7.34215379e-02
-2.29354799e-01 6.81810319e-01 3.72515768e-01 -1.09054577e+00
-3.78599524e-01 1.39914334e-01 2.93625951e-01 2.86557078e-01
-5.07315516e-01 2.21360642e-02 -2.55411208e-01 -3.90435517e-01
-5.65091908e-01 -8.90142500e-01 9.63483285e-03 -2.39029616e-01
-5.75542986e-01 -8.72000098e-01 5.17945766e-01 -7.97709942e-01
1.16322827e+00 -1.76678848e+00 -9.77044180e-03 2.32232772e-02
8.68554339e-02 6.69332623e-01 -2.06290558e-01 6.38948441e-01
3.49288017e-01 8.34266916e-02 1.51921362e-01 -2.33398005e-01
2.37566397e-01 4.38584805e-01 -3.29776928e-02 -3.14809918e-01
3.31146598e-01 5.50013423e-01 -7.79538751e-01 -3.08498979e-01
1.82310209e-01 -1.38368784e-02 -6.66328192e-01 7.21926868e-01
-9.74573374e-01 3.55151743e-01 -2.77482837e-01 -7.52544254e-02
-5.14647290e-02 3.78536507e-02 6.22976720e-01 1.41026080e-01
-7.02740103e-02 8.01120460e-01 -5.45658350e-01 1.45360112e+00
-8.79785001e-01 1.06134915e+00 1.68756455e-01 -7.26760447e-01
1.29166949e+00 5.18051147e-01 -2.20252156e-01 -7.74497628e-01
2.75664002e-01 5.74649451e-03 6.02648914e-01 -7.60698140e-01
7.30637550e-01 -1.98692009e-01 -1.43927991e-01 8.94935310e-01
2.37416431e-01 -4.44738388e-01 4.54012245e-01 2.93628927e-02
1.12238073e+00 4.02387157e-02 2.43900523e-01 -5.15169166e-02
5.02498925e-01 4.78006482e-01 2.86941856e-01 1.09242380e+00
-1.48481339e-01 -5.27100451e-02 9.19647872e-01 -6.14172995e-01
-1.16606748e+00 -3.81552637e-01 2.68202811e-01 1.61507046e+00
-7.31103361e-01 -3.16216111e-01 -1.12268126e+00 -6.47925794e-01
-2.41936013e-01 1.21182704e+00 -3.83379877e-01 9.32827815e-02
-6.80901647e-01 -1.66884094e-01 1.15289724e+00 2.47934997e-01
8.92438173e-01 -1.48345435e+00 -5.49916506e-01 7.17253864e-01
-2.47965068e-01 -1.09826767e+00 -2.55151629e-01 3.86992127e-01
-6.47673368e-01 -8.95483553e-01 -5.82663178e-01 -7.01653779e-01
2.28047594e-01 -4.42334473e-01 1.28332615e+00 1.29797868e-02
3.12966704e-01 2.22634137e-01 -5.69422722e-01 -3.96183550e-01
-1.29393005e+00 4.19770032e-01 -4.76304665e-02 -2.84874856e-01
4.21747625e-01 -9.27398652e-02 1.85281768e-01 5.98375678e-01
-3.97957414e-01 3.55501711e-01 2.39027932e-01 1.21357942e+00
-5.31619072e-01 -2.11169854e-01 8.02616775e-01 -1.37180495e+00
1.83696651e+00 2.98034716e-02 -4.69343096e-01 1.98125213e-01
-3.06683391e-01 3.02347630e-01 1.11722875e+00 -5.51190257e-01
-1.58736658e+00 -7.95417354e-02 -1.66348279e-01 1.80831477e-01
-2.91487932e-01 5.36993265e-01 4.44308482e-02 1.15256026e-01
1.33815920e+00 -2.99960114e-02 3.87313873e-01 -3.23645562e-01
5.21329880e-01 9.70004737e-01 3.00672323e-01 -1.01299953e+00
7.06795156e-02 -4.78465140e-01 -7.59420991e-01 -1.07100630e+00
-4.89501804e-01 -4.71289530e-02 -4.08121347e-01 -5.13427615e-01
7.06848562e-01 -6.41447961e-01 -8.63866687e-01 6.43002570e-01
-1.71966803e+00 -1.34465849e+00 -1.18518613e-01 1.81704551e-01
-6.63521826e-01 7.78095722e-02 -1.10972798e+00 -9.51860785e-01
-3.31212163e-01 -1.01908505e+00 1.01683371e-01 9.00918841e-02
-1.11705744e+00 -1.07992744e+00 1.79282710e-01 6.81421638e-01
6.25862718e-01 -1.65659517e-01 1.09906757e+00 -1.47431040e+00
1.43440828e-01 1.17124051e-01 -1.07664980e-01 7.96682954e-01
1.84601024e-02 -1.63615897e-01 -1.04972053e+00 4.01265681e-01
8.21280032e-02 -1.01516306e+00 2.37942323e-01 -1.15425646e-01
6.41191185e-01 -7.26592302e-01 3.02475899e-01 -2.62435585e-01
5.81081629e-01 5.27452528e-01 3.06927055e-01 2.53537565e-01
3.58764529e-01 1.20880747e+00 4.21597332e-01 4.81695771e-01
3.17429334e-01 6.77582443e-01 -4.03321236e-01 1.27699122e-01
9.80582610e-02 -9.75342914e-02 5.90732813e-01 7.92733252e-01
4.37254272e-02 -3.23299915e-01 -1.20095134e+00 4.15070951e-01
-1.92828500e+00 -5.52051425e-01 -4.13656272e-02 1.39852762e+00
1.60718656e+00 3.02397490e-01 2.92360961e-01 -1.45179197e-01
6.80589259e-01 -6.25671968e-02 -4.47902590e-01 -1.23395491e+00
2.66524792e-01 1.04890093e-02 -1.34896263e-01 6.89772606e-01
-7.72790074e-01 1.16278112e+00 6.12066031e+00 4.74693060e-01
-9.25778985e-01 1.50160892e-02 5.81868052e-01 7.26189166e-02
1.47355869e-01 -4.58883524e-01 -6.68307960e-01 2.64185697e-01
1.27496684e+00 1.25890374e-01 7.82460451e-01 1.04602289e+00
5.09470582e-01 -2.68789589e-01 -1.47358680e+00 3.67054552e-01
1.09748483e-01 -1.29757702e+00 -8.56818557e-02 -2.60579020e-01
3.90412658e-01 -1.78992823e-01 -5.12654424e-01 8.09623182e-01
8.97052050e-01 -8.46599460e-01 3.45687032e-01 4.02660578e-01
4.44837153e-01 -5.53584993e-01 8.15675735e-01 5.95425248e-01
-1.73319146e-01 4.25795317e-02 -1.37822241e-01 -4.89537835e-01
-7.91337267e-02 1.49214759e-01 -1.74767649e+00 -2.92695820e-01
3.36749822e-01 -2.99910903e-02 -3.53711098e-01 3.60068023e-01
-3.78150493e-01 7.67272949e-01 -3.19806576e-01 -8.07502568e-01
4.01085585e-01 -1.61742850e-03 1.38485521e-01 1.53936911e+00
-5.20404696e-01 2.66924024e-01 1.85712099e-01 9.06928003e-01
-1.93383947e-01 5.38598299e-02 -9.28231776e-01 -4.27812040e-01
3.85409981e-01 9.80554700e-01 -3.92517090e-01 -3.68945301e-01
-2.88378298e-01 7.50334263e-01 4.92103964e-01 1.60590783e-01
-3.14693868e-01 -5.93022525e-01 3.03733557e-01 -1.65350601e-01
-3.51734161e-02 -1.46050796e-01 -4.91744906e-01 -7.37520874e-01
-2.56715536e-01 -1.51327622e+00 -1.20210379e-01 -9.44191396e-01
-1.42500031e+00 8.92635822e-01 1.65694044e-03 -5.86936533e-01
-9.59757328e-01 -9.50628877e-01 -6.49408281e-01 8.72164369e-01
-6.88796937e-01 -7.23559856e-01 -1.89609870e-01 4.52102572e-01
1.04298854e+00 -7.65690506e-01 1.19248986e+00 -1.46932021e-01
-4.86061603e-01 5.67945898e-01 -6.04192130e-02 8.28260481e-01
8.35338831e-01 -1.22275996e+00 3.52636993e-01 1.74087837e-01
-3.53195161e-01 7.85174608e-01 7.76727498e-01 -6.32586122e-01
-1.04218566e+00 -8.51511240e-01 7.80517340e-01 -4.50493246e-01
6.96551740e-01 -5.12229443e-01 -8.52575362e-01 4.39171880e-01
7.61978090e-01 -9.19919968e-01 8.47943842e-01 3.95984679e-01
-1.56038389e-01 1.70641556e-01 -1.12215066e+00 7.64317036e-01
6.03083372e-01 -8.40255320e-01 -1.14530373e+00 5.69587171e-01
7.64049590e-01 -4.92186457e-01 -9.51137364e-01 -1.27204061e-01
4.79647160e-01 -8.85655642e-01 4.44690257e-01 -7.46864438e-01
8.68484080e-01 3.21981251e-01 1.83604300e-01 -1.68356240e+00
2.28846341e-01 -7.76609063e-01 2.11138725e-01 1.26729751e+00
7.89882541e-01 -5.89216530e-01 5.92871845e-01 1.18149590e+00
-1.24114573e-01 -5.00032783e-01 -6.13813281e-01 -5.79097450e-01
3.16232353e-01 -2.81133950e-01 2.20714491e-02 1.03637147e+00
6.91981316e-01 9.81915832e-01 -3.33275408e-01 -7.57461548e-01
-1.09115213e-01 -5.58620811e-01 1.10143840e+00 -1.16726792e+00
-4.28246796e-01 -4.01892453e-01 4.55857635e-01 -1.01710916e+00
4.41104829e-01 -4.22731578e-01 6.25828266e-01 -1.36719227e+00
-5.57948470e-01 -2.80531496e-01 6.24568641e-01 9.30913687e-01
2.60797799e-01 -3.34527284e-01 3.57046038e-01 -1.13244839e-01
-4.33108270e-01 4.65017378e-01 1.19331956e+00 -1.03497781e-01
-6.64973915e-01 -4.50831316e-02 -5.57087362e-01 9.28558886e-01
8.83931100e-01 -2.45413423e-01 -4.67554867e-01 -3.05007726e-01
1.19854562e-01 3.15856099e-01 -5.17396592e-02 -7.34971941e-01
4.65758473e-01 -1.20268486e-01 1.44448057e-01 -1.92236915e-01
4.49685007e-01 -4.56654340e-01 -5.28310537e-01 3.32836539e-01
-9.24277425e-01 -1.40104190e-01 5.10938108e-01 -1.18375123e-02
-1.87530279e-01 -5.62744379e-01 5.94174027e-01 -6.75876319e-01
-6.78555131e-01 -8.47275436e-01 -1.37279606e+00 3.07448775e-01
5.65106153e-01 -9.22802165e-02 -5.70015311e-01 -9.13393199e-01
-6.24544084e-01 3.37742746e-01 7.95204267e-02 2.90992558e-01
4.15193856e-01 -7.42116332e-01 -6.21654987e-01 8.06376189e-02
5.06073274e-02 8.37531015e-02 -2.88968563e-01 2.28968620e-01
-8.09767663e-01 7.08489478e-01 -4.52142835e-01 -1.89995438e-01
-1.11930466e+00 -2.38007352e-01 7.12066174e-01 -4.67356741e-01
-2.92167515e-01 8.85709286e-01 -2.47503191e-01 -1.17791879e+00
5.17323375e-01 -5.22987664e-01 -2.83702284e-01 2.62253672e-01
3.50858450e-01 1.67857409e-01 1.99633464e-02 1.59143984e-01
1.64904147e-01 -3.23870420e-01 -4.84678149e-01 -7.02838540e-01
1.25883663e+00 4.10987258e-01 -7.77781904e-02 7.04242826e-01
5.29108942e-01 -3.15063745e-01 -1.06436396e+00 -3.42020810e-01
2.26675421e-01 1.85597539e-01 -2.82729954e-01 -1.42413092e+00
-4.15055990e-01 7.64976621e-01 -3.84808592e-02 6.38827980e-01
6.42840862e-01 -4.39764172e-01 5.83528936e-01 1.52356935e+00
3.01105112e-01 -1.57914209e+00 2.98298389e-01 1.55866599e+00
1.19748449e+00 -1.16012990e+00 -4.32236850e-01 -1.71957649e-02
-1.23456728e+00 1.39652646e+00 1.20435512e+00 -3.53081785e-02
1.21358082e-01 1.42501712e-01 4.16971177e-01 -1.50330350e-01
-1.17487788e+00 2.06807882e-01 -1.95241079e-01 6.29666924e-01
7.65058577e-01 -5.68736717e-02 -2.53298998e-01 7.33905911e-01
-3.55285674e-01 2.20384136e-01 8.45876575e-01 8.00369024e-01
-3.28641415e-01 -1.04953969e+00 -4.46858197e-01 4.86570716e-01
5.44263422e-02 -3.34062219e-01 -1.30987144e+00 9.54396188e-01
-2.13448092e-01 1.27635586e+00 -8.59783590e-03 -4.32858199e-01
5.73100746e-01 6.97213531e-01 1.93180010e-01 -1.02854764e+00
-1.45226502e+00 -2.24042237e-01 1.36230481e+00 -1.64639324e-01
-7.22411945e-02 -3.24135333e-01 -1.01126599e+00 -2.78613895e-01
-3.26209635e-01 3.64365131e-01 4.92459834e-01 1.09906888e+00
4.89387810e-02 4.51283425e-01 3.53947610e-01 -5.94314754e-01
-8.92509639e-01 -1.65746844e+00 -1.61418661e-01 2.08614692e-01
-2.43675578e-02 -2.28269249e-01 -6.07434921e-02 5.17105199e-02] | [12.918716430664062, 7.984976768493652] |
5e988ae3-233b-4c33-bf02-9713aa8c8249 | 3d-point-cloud-generative-adversarial-network | 1905.06292 | null | https://arxiv.org/abs/1905.06292v2 | https://arxiv.org/pdf/1905.06292v2.pdf | 3D Point Cloud Generative Adversarial Network Based on Tree Structured Graph Convolutions | In this paper, we propose a novel generative adversarial network (GAN) for 3D point clouds generation, which is called tree-GAN. To achieve state-of-the-art performance for multi-class 3D point cloud generation, a tree-structured graph convolution network (TreeGCN) is introduced as a generator for tree-GAN. Because TreeGCN performs graph convolutions within a tree, it can use ancestor information to boost the representation power for features. To evaluate GANs for 3D point clouds accurately, we develop a novel evaluation metric called Frechet point cloud distance (FPD). Experimental results demonstrate that the proposed tree-GAN outperforms state-of-the-art GANs in terms of both conventional metrics and FPD, and can generate point clouds for different semantic parts without prior knowledge. | ['Dong Wook Shu', 'Sung Woo Park', 'Junseok Kwon'] | 2019-05-15 | 3d-point-cloud-generative-adversarial-network-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Shu_3D_Point_Cloud_Generative_Adversarial_Network_Based_on_Tree_Structured_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Shu_3D_Point_Cloud_Generative_Adversarial_Network_Based_on_Tree_Structured_ICCV_2019_paper.pdf | iccv-2019-10 | ['point-cloud-generation'] | ['computer-vision'] | [-5.88531382e-02 2.06373528e-01 2.30690405e-01 -1.43979192e-01
-7.85777867e-01 -5.48671246e-01 6.18512750e-01 -4.07810807e-01
4.36057150e-01 6.86499238e-01 -3.01959813e-01 -4.07098681e-01
3.79584163e-01 -1.57397425e+00 -1.00691569e+00 -5.64523697e-01
1.85964733e-01 5.03948212e-01 7.59459063e-02 -1.58848971e-01
-1.29669234e-01 8.63910973e-01 -1.23293293e+00 1.21079087e-02
1.13724613e+00 1.01639211e+00 6.48021251e-02 5.50641656e-01
-4.91600901e-01 4.84843880e-01 -9.50629711e-01 -6.53760254e-01
4.92608517e-01 -6.14088237e-01 -3.71430337e-01 -4.23669405e-02
3.99377465e-01 -2.08430156e-01 -4.27503228e-01 1.01239371e+00
5.10345161e-01 -4.71537486e-02 8.71993661e-01 -1.65714824e+00
-1.06745338e+00 2.27354348e-01 -3.90860736e-01 -3.08719158e-01
2.83709079e-01 3.89262527e-01 7.33165622e-01 -9.12054241e-01
3.29457700e-01 1.50089800e+00 8.46297324e-01 6.96810365e-01
-9.50183153e-01 -1.11510062e+00 1.26912475e-01 -2.81143993e-01
-1.36990333e+00 3.13700944e-01 1.12779152e+00 -4.34049368e-01
6.93024576e-01 1.47234216e-01 9.68457282e-01 8.93706024e-01
2.02664033e-01 6.32945061e-01 9.06751752e-01 -2.48515401e-02
1.29334956e-01 -4.13131773e-01 -6.75085545e-01 8.43182862e-01
3.72898281e-01 3.08640540e-01 -7.24319443e-02 -7.47041330e-02
1.50115526e+00 7.08585680e-02 -1.05182305e-01 -3.96977276e-01
-1.08980870e+00 9.77505207e-01 1.07382417e+00 4.87341955e-02
-4.08233643e-01 6.02387249e-01 -4.75706868e-02 1.52501985e-01
7.75766075e-01 2.45498613e-01 2.04382055e-02 5.81782870e-02
-5.26271284e-01 4.65322495e-01 3.55705678e-01 1.48947620e+00
9.62480664e-01 3.75418454e-01 -4.79269654e-01 5.40035605e-01
4.37318981e-01 8.81569803e-01 5.14454618e-02 -7.16014504e-01
6.42488539e-01 9.71445978e-01 -1.93210870e-01 -8.50458443e-01
7.64440894e-02 -4.42952663e-01 -1.19704151e+00 4.55226451e-01
-2.38598645e-01 -1.43900067e-01 -1.56171513e+00 1.42886150e+00
4.70999211e-01 6.01091623e-01 1.43553719e-01 8.10719371e-01
1.31939030e+00 8.21881056e-01 -1.65060833e-01 3.78170907e-01
8.48662794e-01 -8.49753082e-01 -2.84435660e-01 -1.17427327e-01
2.50303119e-01 -4.56017137e-01 1.01853442e+00 -2.31113836e-01
-1.10890746e+00 -8.08785021e-01 -1.05061555e+00 -5.00949798e-03
-2.72347718e-01 -2.33137995e-01 7.00246692e-01 6.10461950e-01
-1.00788617e+00 4.57325250e-01 -8.18266571e-01 -4.38741371e-02
9.95949864e-01 1.64917737e-01 -1.01360068e-01 -2.09284663e-01
-9.33129907e-01 4.62880552e-01 1.62257075e-01 -1.32447109e-01
-1.11756647e+00 -9.20493722e-01 -9.77454603e-01 6.68044314e-02
-1.43635049e-01 -1.46513271e+00 1.03277278e+00 -4.14170772e-01
-1.65074885e+00 9.86834407e-01 1.07004866e-01 -2.40811959e-01
4.75097150e-01 7.10630566e-02 -2.74766117e-01 7.08316863e-02
2.37726033e-01 8.04135442e-01 1.00283432e+00 -1.61423981e+00
-3.53496999e-01 -3.95357192e-01 7.28069022e-02 2.28976473e-01
3.70374739e-01 -5.08213580e-01 -4.33519036e-01 -1.08033156e+00
2.73490131e-01 -1.00442410e+00 -3.22749943e-01 -6.34942949e-02
-7.25059688e-01 -2.72690237e-01 1.03582656e+00 -1.61656365e-01
5.21783948e-01 -2.05288506e+00 5.00880368e-02 3.16098809e-01
3.90066922e-01 1.88237846e-01 -1.94744468e-01 2.97425270e-01
3.95731814e-02 4.47629392e-01 -5.78315258e-01 -4.96170938e-01
7.73157701e-02 3.73230070e-01 -3.93349916e-01 1.06329553e-01
5.76911688e-01 1.45893753e+00 -9.82180536e-01 -2.93163389e-01
4.63075638e-01 6.39396906e-01 -4.88144726e-01 5.00404656e-01
-3.62579972e-01 7.01217115e-01 -7.47126758e-01 8.51892948e-01
1.12639213e+00 -3.67575139e-01 -5.87537110e-01 2.34339926e-02
4.53626037e-01 5.98976389e-02 -5.30990899e-01 1.98726296e+00
-4.43812966e-01 1.51035354e-01 -4.39475477e-01 -5.39940357e-01
1.40372610e+00 7.07395226e-02 3.59903485e-01 -2.46936098e-01
5.80817349e-02 1.47512808e-01 -3.14045906e-01 2.75493890e-01
3.09136152e-01 -2.38366425e-01 -1.85832769e-01 5.94946258e-02
2.94600632e-02 -9.92806196e-01 -4.31855798e-01 6.14549108e-02
1.04669845e+00 1.74475119e-01 -2.19568759e-01 3.14939953e-02
4.64981616e-01 -1.40066132e-01 4.55930948e-01 5.43288708e-01
3.47547531e-01 1.09769464e+00 3.21670234e-01 -2.71371871e-01
-1.01854944e+00 -1.55649209e+00 2.29699999e-01 1.22588791e-01
4.04874086e-01 -3.08141321e-01 -8.29578757e-01 -1.01318717e+00
4.49597120e-01 7.07324386e-01 -5.47434688e-01 -3.40410471e-01
-5.11967957e-01 -2.95252949e-01 6.03213966e-01 6.46903694e-01
8.09703112e-01 -9.61103082e-01 -2.17575997e-01 9.22004730e-02
2.17822030e-01 -1.16703117e+00 -4.26205397e-01 -3.41364533e-01
-1.03239405e+00 -9.15058494e-01 -9.68517423e-01 -7.04873264e-01
8.58732700e-01 2.82366604e-01 1.41798878e+00 9.29975957e-02
2.59420812e-01 3.08170259e-01 -5.02246559e-01 -8.31816912e-01
-6.29663765e-01 2.04417706e-01 -3.49373728e-01 -2.54606187e-01
1.73268110e-01 -9.69212651e-01 -7.90705740e-01 2.71311820e-01
-8.17616701e-01 2.82183647e-01 6.75698400e-01 7.50102520e-01
1.13846385e+00 1.58317953e-01 4.19009149e-01 -9.07729268e-01
6.07277274e-01 -3.47684413e-01 -7.61373222e-01 -7.99131617e-02
-3.72119397e-01 -1.30086586e-01 7.46434510e-01 -6.08129278e-02
-6.15668714e-01 -1.58200353e-01 -2.97507346e-01 -1.22192478e+00
-7.46752098e-02 2.46553585e-01 -4.66476172e-01 -5.04927874e-01
4.00267273e-01 4.17064041e-01 -7.52991885e-02 -2.73732990e-01
6.60412490e-01 2.63362467e-01 6.71237886e-01 -5.53624988e-01
1.50075150e+00 3.87632191e-01 5.03559649e-01 -4.46740210e-01
-6.23278677e-01 6.71341866e-02 -4.57702458e-01 -1.32216185e-01
1.04955387e+00 -1.13681066e+00 -5.27592540e-01 8.21901977e-01
-1.35279155e+00 -3.97562653e-01 -5.08242488e-01 1.12794176e-01
-7.94707954e-01 -3.06475721e-02 -3.23751181e-01 -6.16756320e-01
-7.27192342e-01 -1.05783248e+00 1.59332240e+00 2.54372120e-01
3.75288278e-01 -9.58108068e-01 -1.77941665e-01 3.09655163e-02
3.09537977e-01 1.08759367e+00 8.30771327e-01 -6.16580360e-02
-1.22557223e+00 -3.80600870e-01 -3.43282372e-01 5.59099674e-01
3.93598288e-01 -2.34272733e-01 -8.60578239e-01 -3.37693393e-01
-1.23825073e-01 5.86424172e-02 5.51636994e-01 3.02462906e-01
1.62190080e+00 -2.44755581e-01 -3.93257231e-01 1.30168998e+00
1.38674068e+00 2.64372617e-01 9.45934772e-01 -6.51353747e-02
1.40901053e+00 -2.48189330e-01 3.46929312e-01 2.10869476e-01
5.71111023e-01 4.78517473e-01 8.42065513e-01 -1.54755846e-01
-3.97661984e-01 -9.24091518e-01 -6.96698576e-02 8.83055568e-01
-2.52239853e-01 -3.79358292e-01 -7.61367083e-01 3.71450543e-01
-1.56741226e+00 -5.49347222e-01 -4.73738834e-02 1.83680618e+00
2.82341212e-01 1.06679074e-01 -7.52194971e-02 2.98471730e-02
7.34051764e-01 2.40596950e-01 -7.59560168e-01 -5.36921546e-02
-5.36367260e-02 7.09026396e-01 5.57425380e-01 9.02645737e-02
-8.22825134e-01 1.06912577e+00 6.06805706e+00 8.19240212e-01
-1.13174033e+00 -4.29975837e-02 4.77187932e-01 3.09994817e-01
-7.78610885e-01 -1.21754058e-01 -5.59279919e-01 7.99795330e-01
3.39226395e-01 -4.10941720e-01 3.59547019e-01 1.08351922e+00
-3.05341542e-01 6.39164984e-01 -9.58756745e-01 1.33660400e+00
-1.52788246e-02 -1.52177548e+00 6.44027948e-01 2.18786329e-01
1.02118158e+00 1.59990452e-02 1.53831512e-01 2.92277396e-01
6.88926041e-01 -1.32544851e+00 5.10027409e-01 4.71549451e-01
1.20784330e+00 -1.10280573e+00 5.66434145e-01 1.68464050e-01
-1.33334315e+00 4.23439652e-01 -4.83012050e-01 2.86679506e-01
2.58151948e-01 8.40765476e-01 -8.11391711e-01 1.08812547e+00
5.47053099e-01 1.01071322e+00 -4.36314821e-01 9.96110797e-01
-6.33605301e-01 3.71986061e-01 -2.00863704e-01 5.75652122e-02
4.22875553e-01 -4.26125348e-01 7.45428085e-01 6.05181396e-01
9.37708914e-01 2.87923127e-01 1.04371585e-01 1.45954311e+00
-5.22056520e-01 -2.70184368e-01 -9.09399867e-01 3.63242105e-02
8.60849023e-01 9.72475946e-01 -4.29899454e-01 -1.76720932e-01
-9.15089622e-02 1.06897914e+00 1.22163452e-01 2.34320238e-01
-8.88585329e-01 -5.98726630e-01 9.94040370e-01 8.61214697e-02
5.11976838e-01 -2.96453953e-01 -4.44710374e-01 -9.89707232e-01
1.78617209e-01 -4.76163715e-01 7.13984966e-02 -1.23769057e+00
-1.65679455e+00 7.84759641e-01 -9.38306004e-02 -1.61255324e+00
-1.73294112e-01 -4.70877320e-01 -9.73830462e-01 1.25835013e+00
-1.59321034e+00 -1.79367089e+00 -9.05173719e-01 7.47701943e-01
7.83510283e-02 -3.56072634e-01 7.33708143e-01 -3.69352959e-02
-2.83237901e-02 7.35009432e-01 -3.49751234e-01 2.10814923e-01
-1.20733038e-03 -1.34708750e+00 1.38919866e+00 8.38149309e-01
1.39630273e-01 1.75094053e-01 1.12763517e-01 -8.02855611e-01
-1.45745671e+00 -1.76795101e+00 9.62821096e-02 -5.69445610e-01
1.83693707e-01 -3.96654576e-01 -8.32010746e-01 6.31798863e-01
-1.21221267e-01 1.72582984e-01 4.16721612e-01 -5.03185093e-01
-2.24477455e-01 -9.50857904e-03 -1.47815144e+00 3.31426173e-01
1.56014395e+00 -4.99409437e-01 -4.01550800e-01 3.89937997e-01
1.44940853e+00 -1.05450714e+00 -1.07174432e+00 7.66105533e-01
-7.60603100e-02 -9.53589380e-01 1.15196466e+00 -2.68420041e-01
6.41387641e-01 -5.49140096e-01 -8.88827667e-02 -1.85204220e+00
-4.72965509e-01 -5.41871130e-01 -2.95556337e-01 1.15269220e+00
3.09604965e-02 -7.48041689e-01 1.07648158e+00 -9.99529753e-03
-7.35253572e-01 -9.51002777e-01 -9.27752972e-01 -1.04975736e+00
4.42054152e-01 -4.03088719e-01 1.77484965e+00 7.93201447e-01
-1.02074838e+00 2.33327687e-01 -6.66267276e-02 3.30772817e-01
8.21003258e-01 2.97704935e-01 1.28646851e+00 -1.19831598e+00
2.21424811e-02 -4.80254322e-01 -9.94379342e-01 -1.12237334e+00
2.68946141e-01 -1.18509364e+00 -2.63890684e-01 -1.99449050e+00
-4.81466383e-01 -7.92864203e-01 -7.35907853e-02 1.96635604e-01
-3.35831136e-01 2.95429051e-01 3.18015516e-01 1.75006986e-02
8.00684169e-02 1.00374699e+00 1.99158335e+00 -3.42399657e-01
-3.38766985e-02 1.26711890e-01 -7.46159673e-01 3.68284166e-01
5.95156848e-01 -3.22838515e-01 -6.22140765e-01 -7.10676253e-01
-7.25150704e-02 9.34281759e-03 6.59931064e-01 -1.37024522e+00
-1.31695539e-01 -1.27758473e-01 5.81230462e-01 -1.13063669e+00
4.92859155e-01 -7.96982348e-01 5.89963257e-01 2.04835296e-01
2.69423544e-01 1.07514039e-01 2.30714440e-01 5.10420203e-01
-2.82792449e-01 4.14568275e-01 7.48836696e-01 -2.85049617e-01
-2.93089569e-01 1.12540495e+00 5.34779429e-01 9.40612629e-02
1.09958446e+00 -3.23176384e-01 -4.13170397e-01 -3.62179130e-01
-2.87063301e-01 2.77186841e-01 9.92425561e-01 4.93665785e-01
1.04104197e+00 -2.11371279e+00 -8.16956401e-01 4.45222676e-01
3.21134210e-01 1.04171026e+00 2.95847207e-01 -8.77096280e-02
-8.57089639e-01 6.44113049e-02 -2.02696025e-01 -8.69377851e-01
-6.59585059e-01 4.71213013e-01 3.66312265e-01 -1.05324857e-01
-8.98904324e-01 1.06268656e+00 4.57233667e-01 -6.46100998e-01
-2.92275667e-01 -5.23364723e-01 3.88259530e-01 -7.19847083e-01
5.73555753e-02 1.26802353e-02 8.10850337e-02 -4.98817444e-01
-2.34963208e-01 8.47058415e-01 4.02380943e-01 3.04495156e-01
1.14931965e+00 4.40700591e-01 -9.74165499e-02 1.37950242e-01
1.07398963e+00 -1.59293532e-01 -1.32110465e+00 6.59898967e-02
-7.79082060e-01 -8.60252917e-01 -8.80126003e-03 -5.59342623e-01
-1.47867513e+00 7.34227240e-01 4.29137647e-01 2.21570104e-01
1.09540224e+00 1.38631955e-01 1.24038029e+00 3.63544784e-02
7.49949515e-01 -2.70638734e-01 8.94768611e-02 4.84323591e-01
1.13092601e+00 -8.76572192e-01 -1.40773475e-01 -8.29540610e-01
-5.64275563e-01 6.89047158e-01 8.61796319e-01 -6.29617453e-01
6.15790069e-01 -2.58634519e-02 -2.16939598e-02 -3.89144003e-01
-3.24384838e-01 -1.38561562e-01 4.72181767e-01 1.14872193e+00
7.60651147e-03 3.79327625e-01 3.72832596e-01 4.72468525e-01
-8.64832222e-01 7.66295847e-03 1.03094161e-01 6.19648814e-01
4.42183763e-02 -1.15630090e+00 -1.98777109e-01 6.28908277e-01
-1.50118815e-02 -1.36446245e-02 -4.86796111e-01 7.86253870e-01
2.10967705e-01 5.12143612e-01 3.63398373e-01 -8.22273731e-01
6.03452384e-01 -3.00753623e-01 6.70812845e-01 -7.18174875e-01
-3.19722980e-01 -4.80200410e-01 -3.94217908e-01 -4.55801070e-01
-3.70486319e-01 -2.63071984e-01 -1.17944157e+00 -4.45360124e-01
-1.72801852e-01 2.75877416e-01 5.39138973e-01 4.23196465e-01
7.64625549e-01 8.19168091e-01 1.00814736e+00 -1.01119435e+00
-1.03591077e-01 -8.05022478e-01 -5.38389981e-01 5.63724458e-01
1.09882221e-01 -7.72192121e-01 -2.03282431e-01 -2.85448343e-01] | [8.852622032165527, -3.6910488605499268] |
f01cab74-cdee-4563-b47a-269cabe1d354 | neural-pruning-via-growing-regularization-1 | 2012.09243 | null | https://arxiv.org/abs/2012.09243v2 | https://arxiv.org/pdf/2012.09243v2.pdf | Neural Pruning via Growing Regularization | Regularization has long been utilized to learn sparsity in deep neural network pruning. However, its role is mainly explored in the small penalty strength regime. In this work, we extend its application to a new scenario where the regularization grows large gradually to tackle two central problems of pruning: pruning schedule and weight importance scoring. (1) The former topic is newly brought up in this work, which we find critical to the pruning performance while receives little research attention. Specifically, we propose an L2 regularization variant with rising penalty factors and show it can bring significant accuracy gains compared with its one-shot counterpart, even when the same weights are removed. (2) The growing penalty scheme also brings us an approach to exploit the Hessian information for more accurate pruning without knowing their specific values, thus not bothered by the common Hessian approximation problems. Empirically, the proposed algorithms are easy to implement and scalable to large datasets and networks in both structured and unstructured pruning. Their effectiveness is demonstrated with modern deep neural networks on the CIFAR and ImageNet datasets, achieving competitive results compared to many state-of-the-art algorithms. Our code and trained models are publicly available at https://github.com/mingsuntse/regularization-pruning. | ['Yun Fu', 'Yulun Zhang', 'Can Qin', 'Huan Wang'] | 2020-12-16 | neural-pruning-via-growing-regularization | https://openreview.net/forum?id=o966_Is_nPA | https://openreview.net/pdf?id=o966_Is_nPA | iclr-2021-1 | ['l2-regularization'] | ['methodology'] | [ 2.12026402e-01 5.42820338e-03 -3.27347338e-01 -3.13397467e-01
-3.77631575e-01 -1.81181490e-01 1.41909555e-01 2.15063125e-01
-8.32220316e-01 6.91531003e-01 -6.05018102e-02 -2.38844037e-01
-3.25475127e-01 -5.76625228e-01 -7.76379108e-01 -7.88881719e-01
-1.53897554e-01 1.01594269e-01 3.65783632e-01 -2.74180919e-01
1.90272734e-01 4.25222069e-01 -1.57080281e+00 1.73075348e-02
8.50255966e-01 1.14353013e+00 2.70435840e-01 2.58446991e-01
-5.77750197e-03 7.72410989e-01 -3.08504283e-01 -5.71679175e-01
6.33535266e-01 -5.69423661e-04 -6.76478744e-01 -8.71947557e-02
5.85545242e-01 -2.49266297e-01 -3.45592648e-01 1.23283887e+00
3.65518063e-01 1.38214707e-01 2.62327850e-01 -7.38600075e-01
-4.10143971e-01 8.52022052e-01 -8.75073731e-01 5.57762921e-01
-3.96374881e-01 -1.34382546e-01 1.22155082e+00 -9.49597061e-01
4.80607778e-01 9.06347752e-01 7.66360819e-01 5.56019902e-01
-1.08169997e+00 -7.28565276e-01 4.89448130e-01 4.39383566e-01
-1.42709720e+00 -4.38118279e-01 8.27284455e-01 -1.70580775e-01
8.54166031e-01 9.26532447e-02 7.07632959e-01 8.38907361e-01
-2.24966854e-01 8.43472898e-01 7.66299903e-01 -5.15581429e-01
1.41840145e-01 1.02621697e-01 6.22058451e-01 7.84607053e-01
5.85714936e-01 3.46810855e-02 -5.32103837e-01 3.27698439e-02
4.88037407e-01 1.97108060e-01 -4.65647101e-01 -3.08295339e-01
-5.80941975e-01 9.58788097e-01 4.76374805e-01 4.40346926e-01
-1.29253194e-01 3.35893363e-01 5.34748852e-01 5.42549312e-01
6.98132455e-01 2.40876496e-01 -6.84834838e-01 -2.03782111e-01
-9.53348696e-01 2.90100336e-01 7.59142816e-01 8.48926783e-01
7.66928554e-01 1.92919374e-01 1.01950625e-02 1.21747220e+00
-8.40260535e-02 -1.05594210e-01 4.46620286e-01 -7.50053167e-01
7.65134990e-01 4.54351068e-01 -1.90454915e-01 -9.94549572e-01
-4.04021740e-01 -8.39082539e-01 -1.00547314e+00 1.59766138e-01
3.94487619e-01 -2.62174577e-01 -7.38134265e-01 1.83712173e+00
2.59661436e-01 3.69914740e-01 -1.32614374e-01 7.91054070e-01
6.96636438e-01 3.33991766e-01 -2.27290336e-02 -3.04676205e-01
1.13987112e+00 -1.18954575e+00 -5.11124253e-01 -3.22759837e-01
7.98027337e-01 -5.57479084e-01 1.04612803e+00 6.77398801e-01
-1.18171990e+00 -3.63315016e-01 -1.19354820e+00 -2.25241914e-01
-2.22091287e-01 3.59462589e-01 8.36970568e-01 5.79482973e-01
-1.04422092e+00 1.05241287e+00 -8.90739143e-01 -1.51633143e-01
8.99941623e-01 4.71222967e-01 -4.08182219e-02 -1.65720329e-01
-1.10792458e+00 8.30940068e-01 4.81145769e-01 1.64419979e-01
-5.18665016e-01 -7.51192272e-01 -4.86533523e-01 4.07192618e-01
6.75335109e-01 -3.35925162e-01 1.34068501e+00 -9.40773785e-01
-1.49050069e+00 6.30325437e-01 -3.46918404e-02 -8.87744367e-01
6.11689270e-01 -6.40760660e-01 3.83032151e-02 5.94320707e-02
-2.41958439e-01 6.20435715e-01 9.18910027e-01 -9.86116588e-01
-5.35824776e-01 -2.76771724e-01 1.63684711e-01 1.39708892e-01
-1.01350534e+00 -6.55809194e-02 -5.45302510e-01 -8.88782382e-01
1.53164476e-01 -6.38680041e-01 -4.42009509e-01 5.95278516e-02
-2.57899106e-01 -1.99721262e-01 6.87522054e-01 -3.73256415e-01
1.43050289e+00 -2.06930757e+00 1.50750935e-01 3.53544094e-02
4.72011298e-01 6.83294117e-01 -1.42839968e-01 3.61511350e-01
-3.98603119e-02 9.04664546e-02 -3.37973148e-01 -5.06639421e-01
-2.12169275e-01 3.04234445e-01 -2.72216797e-01 4.16267067e-01
1.33718535e-01 6.95235610e-01 -6.96359634e-01 -2.79709965e-01
-8.46824273e-02 5.15654087e-01 -8.34146559e-01 -1.64192691e-01
1.12616546e-01 9.35735926e-02 -3.77770662e-01 4.90529001e-01
7.64189959e-01 -2.21169233e-01 1.58053651e-01 -2.26589501e-01
-1.03587344e-01 1.86769262e-01 -1.18496227e+00 1.47635567e+00
-3.23497117e-01 4.65238273e-01 2.46141031e-01 -1.44824052e+00
8.73771667e-01 4.29329760e-02 3.32990229e-01 -4.79167491e-01
2.78546572e-01 3.28419596e-01 1.60372987e-01 -3.30847353e-01
2.60277271e-01 -5.65253459e-02 4.09508914e-01 1.81408331e-01
1.36796027e-01 2.13272080e-01 7.27124631e-01 1.54271871e-01
1.24266505e+00 -2.01103047e-01 4.01127994e-01 -4.74269986e-01
5.16499281e-01 -2.49771550e-01 8.32440674e-01 8.18297625e-01
-7.02815428e-02 3.73406321e-01 7.24901259e-01 -6.12847030e-01
-7.54193842e-01 -6.01392210e-01 -3.94476205e-01 1.23924077e+00
1.13334052e-01 -6.90518796e-01 -6.93163693e-01 -5.95818996e-01
-1.78780651e-03 3.46537590e-01 -6.14719629e-01 -2.12320566e-01
-8.12984169e-01 -8.65285039e-01 3.97140622e-01 5.97860038e-01
5.32717049e-01 -8.56618345e-01 -8.23969364e-01 2.03995630e-01
2.11696357e-01 -1.07002747e+00 -1.94309711e-01 6.37127638e-01
-1.41200304e+00 -1.13844454e+00 -7.71135151e-01 -7.12542415e-01
6.31407619e-01 3.06383520e-01 1.07991529e+00 4.99022335e-01
-3.41573924e-01 7.15521872e-02 -5.25686562e-01 -5.25829196e-01
1.47639066e-01 4.36247319e-01 -7.64507847e-03 -2.70299405e-01
2.67833799e-01 -9.27469850e-01 -6.44620836e-01 2.11747289e-01
-8.89066160e-01 3.66906971e-02 8.25279176e-01 9.79241967e-01
6.04566097e-01 -1.06919762e-02 6.38073504e-01 -1.32255995e+00
7.75542021e-01 -4.15891111e-01 -6.70897365e-01 1.31550627e-02
-7.29479134e-01 1.58648286e-02 8.75974715e-01 -5.08240581e-01
-8.64346683e-01 -1.10697895e-01 -3.30631703e-01 -5.37927032e-01
1.18585704e-02 5.98136008e-01 2.19747558e-01 -3.39683414e-01
5.29191673e-01 1.08304415e-02 -3.30310673e-01 -8.92700493e-01
1.55440480e-01 1.60224125e-01 2.00915277e-01 -5.43338954e-01
6.72775805e-01 4.22957987e-01 6.36791512e-02 -8.73900771e-01
-1.21352518e+00 -5.68776965e-01 -5.93271017e-01 2.38437086e-01
2.32180089e-01 -7.03684568e-01 -4.33660835e-01 2.61198282e-01
-1.07640970e+00 -2.57712901e-01 -4.39128011e-01 3.73901904e-01
-1.62216574e-01 5.25863230e-01 -9.05066848e-01 -6.65846705e-01
-5.33174813e-01 -1.16165853e+00 5.21338344e-01 1.24408871e-01
3.20393384e-01 -8.09354722e-01 -4.37560268e-02 1.18591532e-01
4.89998907e-01 -3.44159305e-02 9.72360730e-01 -6.91565096e-01
-4.58172530e-01 -1.57900780e-01 -4.23556775e-01 6.93169415e-01
-2.06445411e-01 -3.01626533e-01 -8.89290929e-01 -3.98913950e-01
3.63985419e-01 -4.72780496e-01 1.48180568e+00 3.87324154e-01
1.61440897e+00 -3.82481277e-01 -1.33042127e-01 8.92506599e-01
1.53243446e+00 -1.40471220e-01 3.38462293e-01 3.80274862e-01
6.53047144e-01 3.81291777e-01 4.48319286e-01 6.06378794e-01
-1.01027481e-01 6.15302563e-01 5.88351071e-01 -1.40039520e-02
-5.55913635e-02 1.22348666e-01 4.36785677e-03 1.13929391e+00
-3.09807867e-01 -7.40551725e-02 -8.21549535e-01 3.73965383e-01
-2.01732016e+00 -7.65397251e-01 -5.86228110e-02 2.02316952e+00
9.42324936e-01 4.27373320e-01 -1.95686787e-01 3.98363084e-01
5.52525520e-01 3.38773221e-01 -5.84372580e-01 -4.63339299e-01
2.27986220e-02 5.26455820e-01 6.52277291e-01 3.30435127e-01
-1.17554080e+00 9.77172613e-01 5.50274992e+00 1.16105044e+00
-9.20708358e-01 4.02390271e-01 7.53213644e-01 -4.59886193e-01
1.39556095e-01 -8.97866338e-02 -1.10462117e+00 1.75099581e-01
6.78740561e-01 -1.23701748e-02 3.91778618e-01 1.11320388e+00
1.25288934e-01 -6.78448603e-02 -1.00490618e+00 1.04471695e+00
-8.46265852e-02 -1.42393720e+00 4.02875133e-02 -1.03089556e-01
7.31302977e-01 3.07544649e-01 1.88811682e-02 5.19287527e-01
-1.18064135e-01 -9.07362282e-01 6.33646250e-01 1.66739136e-01
5.43975472e-01 -9.30754006e-01 7.94746757e-01 5.36023319e-01
-1.14261162e+00 -4.28736925e-01 -9.01581883e-01 -2.92737752e-01
6.88143149e-02 9.78963256e-01 -3.56639326e-01 3.07274848e-01
7.81352520e-01 8.43735814e-01 -5.61266005e-01 1.33140230e+00
-3.56401712e-01 8.67843628e-01 -3.46867204e-01 -9.69126355e-03
4.00668263e-01 -2.95092046e-01 4.29501414e-01 1.16359007e+00
2.37065136e-01 5.74756637e-02 -5.72400959e-03 5.82732499e-01
-5.24542451e-01 3.56200278e-01 -3.99172425e-01 1.09381057e-01
2.70223826e-01 1.27959073e+00 -8.68169427e-01 -2.08793342e-01
-4.86390084e-01 5.68762779e-01 9.30038393e-01 3.08558553e-01
-8.20092916e-01 -4.34091032e-01 5.30170560e-01 1.36799544e-01
6.78510427e-01 -1.47706285e-01 -4.06438768e-01 -1.10271215e+00
2.14780942e-01 -6.43658698e-01 2.33807743e-01 1.24458391e-02
-1.16092026e+00 6.01506829e-01 1.07969530e-01 -9.56668019e-01
3.10618341e-01 -7.66682267e-01 -6.73223853e-01 4.44782108e-01
-1.83561552e+00 -6.52717769e-01 -2.11556062e-01 3.79402906e-01
7.28971660e-01 -8.48294422e-02 5.01720488e-01 7.30325818e-01
-8.42642009e-01 7.97469616e-01 2.78774410e-01 -7.47471303e-02
3.55738461e-01 -9.80795503e-01 1.80635229e-01 9.28021789e-01
3.38320792e-01 7.59346783e-01 5.62357605e-01 -3.34111452e-01
-1.19499743e+00 -1.02593601e+00 6.79562569e-01 1.42330959e-01
7.46288598e-01 -4.23317969e-01 -1.16696906e+00 3.51211280e-01
-2.63032001e-02 2.45262325e-01 5.35296977e-01 3.05264831e-01
-3.03431869e-01 -3.78880858e-01 -9.10113215e-01 5.59567571e-01
1.35260487e+00 4.04781178e-02 -4.56848830e-01 4.07023787e-01
7.02435672e-01 -4.05724913e-01 -5.37197530e-01 5.93986928e-01
4.10053492e-01 -1.17653024e+00 9.12227929e-01 -4.80057925e-01
5.51676154e-01 1.59553066e-01 3.24748270e-02 -8.85666311e-01
-1.29500911e-01 -5.57643652e-01 -5.39228976e-01 9.08496082e-01
3.84440124e-01 -6.86859787e-01 1.06160808e+00 2.60550410e-01
-2.40456119e-01 -1.52608323e+00 -9.76153016e-01 -8.48368824e-01
3.52178365e-02 -5.37098765e-01 3.05152535e-01 7.21036851e-01
-2.59319723e-01 2.69885212e-01 -5.30006588e-01 -2.12853149e-01
5.33988595e-01 1.44475065e-02 4.92636949e-01 -1.29873228e+00
-4.40571755e-01 -6.59191728e-01 -2.61471003e-01 -1.39938629e+00
3.43396850e-02 -8.84471238e-01 -1.19039565e-01 -1.16101944e+00
1.94432065e-01 -5.90875030e-01 -5.03571153e-01 5.83434582e-01
-9.55075845e-02 3.51098925e-01 2.72057414e-01 3.59584957e-01
-5.78678489e-01 7.51999319e-01 1.00316811e+00 1.01858901e-03
-1.72606111e-01 2.46645689e-01 -8.53055060e-01 1.04440856e+00
1.07952929e+00 -7.35607326e-01 -6.40189707e-01 -7.23880589e-01
2.21002713e-01 -3.54581594e-01 2.27979660e-01 -1.25748432e+00
2.68738955e-01 6.93592727e-02 1.23671759e-02 -4.39349681e-01
3.44330847e-01 -8.35661769e-01 -3.65187347e-01 5.44947028e-01
-4.04846042e-01 3.61083262e-02 2.64950186e-01 7.42045939e-01
-2.38571659e-01 -8.19018781e-01 1.02215207e+00 -1.72966599e-01
-6.14722908e-01 4.40437794e-01 -2.59360881e-03 2.95519233e-01
7.86315441e-01 -2.10828841e-01 -1.21702395e-01 -6.26971871e-02
-6.46112084e-01 1.05525210e-01 1.59267068e-01 1.20010875e-01
6.92822814e-01 -1.04232132e+00 -4.64132845e-01 6.69829398e-02
-1.63964897e-01 1.05890013e-01 1.95498511e-01 1.13905668e+00
-5.71329653e-01 4.13074136e-01 3.86017226e-02 -4.57173854e-01
-1.22792363e+00 3.18069160e-01 4.66290005e-02 -6.11574531e-01
-9.23803449e-01 1.34804094e+00 1.60797849e-01 -1.04100415e-02
7.35585570e-01 -5.42781115e-01 -2.74856329e-01 -1.28608681e-02
5.62615931e-01 4.27072257e-01 3.46721262e-01 -2.01531455e-01
-8.74137133e-02 4.91268605e-01 -4.35481101e-01 4.97123539e-01
1.75866222e+00 1.77313566e-01 -1.54628942e-03 2.75265306e-01
1.24795890e+00 -6.71505854e-02 -1.26539886e+00 -3.76739711e-01
2.15132385e-01 -4.84068483e-01 1.57969669e-01 -4.14856195e-01
-1.61076200e+00 9.77618992e-01 5.51624238e-01 1.93763152e-01
1.25322413e+00 -1.70612767e-01 9.75408852e-01 8.74826431e-01
3.20897400e-01 -1.19217896e+00 -2.17005927e-02 6.71431899e-01
6.68684065e-01 -1.21272039e+00 3.27028394e-01 -5.00465453e-01
-4.27595705e-01 1.05401838e+00 6.92556143e-01 -3.72061819e-01
8.13552797e-01 3.38508368e-01 -2.40423277e-01 -2.75262862e-01
-8.47589910e-01 -1.65252820e-01 1.68636426e-01 1.75052464e-01
5.56004226e-01 -2.90015489e-01 -7.68242657e-01 6.64256215e-01
-6.65370151e-02 6.14836477e-02 2.36391604e-01 9.78275537e-01
-6.16625547e-01 -1.29843330e+00 1.24706037e-01 8.07907701e-01
-7.44551539e-01 -4.11649317e-01 -7.00310618e-02 8.88956308e-01
2.21522361e-01 5.17742097e-01 -3.19802791e-01 -1.47938699e-01
2.84153491e-01 -1.28823742e-01 3.71599972e-01 -7.25722671e-01
-8.03601921e-01 -9.33807045e-02 -1.42888464e-02 -6.11938536e-01
-4.16198194e-01 -3.90366286e-01 -9.75029945e-01 -1.37968943e-01
-5.21000028e-01 1.86970428e-01 4.51801956e-01 7.54188299e-01
2.77167916e-01 4.72358406e-01 2.43177459e-01 -9.43720102e-01
-9.53584373e-01 -8.48137259e-01 -5.92565656e-01 2.46863425e-01
1.14495710e-01 -9.00630474e-01 -4.97127950e-01 -2.54967958e-01] | [8.653255462646484, 3.282565116882324] |
1f1c9755-4bb7-45b1-87b6-2c0de8d23f24 | distant-supervision-for-relation-extraction | null | null | https://aclanthology.org/D15-1203 | https://aclanthology.org/D15-1203.pdf | Distant Supervision for Relation Extraction via Piecewise Convolutional Neural Networks | null | ['Daojian Zeng', 'Yubo Chen', 'Jun Zhao', 'Kang Liu'] | 2015-09-01 | null | null | null | emnlp-2015-9 | ['relationship-extraction-distant-supervised'] | ['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] | [-7.255461692810059, 3.7228293418884277] |
24249f5c-814c-483b-84a3-a099d5e51410 | robust-constrained-hyperspectral-unmixing | 2302.08247 | null | https://arxiv.org/abs/2302.08247v1 | https://arxiv.org/pdf/2302.08247v1.pdf | Robust Constrained Hyperspectral Unmixing Using Reconstructed-Image Regularization | Hyperspectral (HS) unmixing is the process of decomposing an HS image into material-specific spectra (endmembers) and their spatial distributions (abundance maps). Existing unmixing methods have two limitations with respect to noise robustness. First, if the input HS image is highly noisy, even if the balance between sparse and piecewise-smooth regularizations for abundance maps is carefully adjusted, noise may remain in the estimated abundance maps or undesirable artifacts may appear. Second, existing methods do not explicitly account for the effects of stripe noise, which is common in HS measurements, in their formulations, resulting in significant degradation of unmixing performance when such noise is present in the input HS image. To overcome these limitations, we propose a new robust hyperspectral unmixing method based on constrained convex optimization. Our method employs, in addition to the two regularizations for the abundance maps, regularizations for the HS image reconstructed by mixing the estimated abundance maps and endmembers. This strategy makes the unmixing process much more robust in highly-noisy scenarios, under the assumption that the abundance maps used to reconstruct the HS image with desirable spatio-spectral structure are also expected to have desirable properties. Furthermore, our method is designed to accommodate a wider variety of noise including stripe noise. To solve the formulated optimization problem, we develop an efficient algorithm based on a preconditioned primal-dual splitting method, which can automatically determine appropriate stepsizes based on the problem structure. Experiments on synthetic and real HS images demonstrate the advantages of our method over existing methods. | ['Shunsuke Ono', 'Yuki Nagamatsu', 'Kazuki Naganuma'] | 2023-02-16 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 6.89159870e-01 -5.92922986e-01 8.92907754e-02 -4.39682007e-02
-6.04006350e-01 -4.98878390e-01 2.20272601e-01 -1.99961469e-01
-7.40489438e-02 7.81068861e-01 2.35793039e-01 -8.33513588e-02
-2.50156999e-01 -8.18212688e-01 -5.03702581e-01 -1.29728758e+00
3.13788086e-01 7.08030835e-02 -2.47816876e-01 -1.62960291e-01
-1.67757779e-01 4.93079752e-01 -1.71845305e+00 -7.54812732e-03
1.40687168e+00 8.56569707e-01 4.39574420e-01 2.46272013e-01
-1.23230271e-01 5.58771014e-01 -3.55597228e-01 3.07927728e-01
7.43134320e-01 -6.10376775e-01 -3.81003499e-01 7.42131233e-01
4.06293333e-01 -1.51562393e-01 -1.42907068e-01 1.64351285e+00
2.95850664e-01 3.33146125e-01 6.22861266e-01 -1.01165867e+00
-3.00746173e-01 4.46301192e-01 -1.12418211e+00 -2.24844024e-01
-2.00729564e-01 1.55578271e-01 7.51193464e-01 -7.64799476e-01
2.19373018e-01 1.08886337e+00 6.31119549e-01 1.66994724e-02
-1.52124977e+00 -6.41945243e-01 1.20560624e-01 -2.51381993e-01
-1.54027617e+00 -4.09123003e-01 9.10401046e-01 -6.04064465e-01
3.13034803e-01 6.45592153e-01 6.59942746e-01 4.48628217e-01
-2.33991593e-01 4.32422936e-01 1.43392587e+00 -4.64964062e-01
1.92200437e-01 -1.59540155e-03 7.41314888e-02 3.80112708e-01
5.17377794e-01 8.38824883e-02 -1.62128210e-01 -4.39215928e-01
6.08324111e-01 1.41891062e-01 -7.49030471e-01 -4.37282175e-01
-1.34968293e+00 6.47602499e-01 3.97302508e-01 2.35504299e-01
-6.98412180e-01 -3.79333496e-01 -1.37951011e-02 8.48279595e-02
7.08043933e-01 3.60295445e-01 -4.17771637e-02 6.03793442e-01
-1.27320182e+00 2.12709904e-02 3.85904789e-01 5.78381896e-01
1.27836573e+00 2.98255235e-01 6.35064915e-02 1.17380905e+00
4.04326975e-01 9.92713511e-01 2.94390023e-01 -9.20466483e-01
4.17328179e-01 5.43815196e-01 4.05451715e-01 -1.05923820e+00
-1.71265021e-01 -6.42013073e-01 -1.18048275e+00 2.40405247e-01
2.73447245e-01 -1.99226961e-01 -9.70337510e-01 1.61351752e+00
4.17872101e-01 2.74750501e-01 2.06393018e-01 1.09732699e+00
4.98680323e-01 1.04284477e+00 -2.33391486e-02 -8.04623723e-01
9.11090732e-01 -6.91905499e-01 -1.03232360e+00 -3.42634469e-01
2.47093439e-01 -8.04468393e-01 8.89896512e-01 2.14087188e-01
-7.52636731e-01 -3.43286663e-01 -1.19640684e+00 3.63924474e-01
-1.50762424e-01 2.32376829e-01 3.91474158e-01 4.94898468e-01
-6.54790163e-01 4.59688246e-01 -6.31499946e-01 -3.05263191e-01
-4.70045879e-02 2.30751306e-01 -1.73530579e-01 -7.09359497e-02
-9.67552960e-01 6.68234587e-01 5.74969351e-01 5.18042862e-01
-5.80868065e-01 -6.10166669e-01 -8.54983032e-01 -1.53797297e-02
3.01181614e-01 -4.80105370e-01 6.32676244e-01 -1.46121883e+00
-1.37480128e+00 6.41937137e-01 -2.47857556e-01 1.50474176e-01
2.78530359e-01 1.29452720e-01 -5.85966349e-01 9.02221128e-02
1.45098850e-01 8.20704475e-02 1.02978539e+00 -1.70173013e+00
-5.50327420e-01 -4.25362647e-01 -3.14303160e-01 4.61387426e-01
-4.38422680e-01 -1.97602570e-01 -1.45994484e-01 -7.51094520e-01
7.10043669e-01 -9.73585486e-01 -4.37593013e-01 -1.29949316e-01
-4.11252648e-01 6.74039185e-01 9.21199501e-01 -7.09238172e-01
1.10302329e+00 -2.37993336e+00 3.18254381e-01 4.74361211e-01
6.94499584e-03 2.78673261e-01 -2.53550589e-01 3.52564931e-01
-2.89217144e-01 -1.24197200e-01 -9.99350309e-01 -7.93287456e-02
-2.66495526e-01 2.90804684e-01 -1.92710504e-01 9.01967585e-01
-2.30542030e-02 2.70908654e-01 -1.00567305e+00 -1.36926755e-01
4.18015003e-01 4.27049160e-01 -2.42020171e-02 3.30495387e-01
-1.03241459e-01 6.61198080e-01 -1.86342612e-01 6.46391153e-01
1.21357429e+00 -1.53043404e-01 4.60057288e-01 -5.33409417e-01
-4.95506078e-01 -2.94661522e-01 -1.63487482e+00 1.31462717e+00
-2.19780803e-01 1.81545690e-01 8.53357434e-01 -1.16984153e+00
7.81736732e-01 3.46898407e-01 8.44812810e-01 -1.97196454e-01
-5.32284267e-02 5.22706151e-01 -1.01022892e-01 -4.85909194e-01
3.39674741e-01 -7.46455193e-01 6.00226760e-01 3.77852380e-01
-3.04059774e-01 -2.18941122e-01 2.43896902e-01 -1.66701049e-01
4.67952400e-01 7.97879510e-03 2.86656737e-01 -7.20337629e-01
7.35796690e-01 1.49971902e-01 7.46201873e-01 4.66888219e-01
2.13555489e-02 6.69447064e-01 -2.88386759e-03 -1.97601154e-01
-9.15453017e-01 -8.30658078e-01 -3.84725183e-01 7.54902303e-01
4.13857132e-01 1.86262816e-01 -6.05802417e-01 -2.07265705e-01
-1.11869201e-01 3.95163327e-01 -2.21562609e-01 9.21200439e-02
-2.26973653e-01 -1.70237064e+00 1.90707073e-01 8.71564597e-02
6.61110997e-01 -5.79895496e-01 -1.23085208e-01 2.97348708e-01
-5.36225021e-01 -8.89751732e-01 -2.76599675e-01 1.91052526e-01
-9.69266772e-01 -1.11646843e+00 -8.10191810e-01 -6.37903869e-01
9.32493150e-01 8.32408011e-01 6.68056786e-01 -9.03833583e-02
-4.25747037e-02 1.51731849e-01 -2.27049366e-01 -2.47356713e-01
-3.84534001e-01 -4.23613608e-01 1.26708925e-01 6.83468282e-01
3.15136984e-02 -5.80838323e-01 -3.55254322e-01 2.98167586e-01
-1.39588904e+00 2.23188162e-01 4.98105377e-01 9.47835684e-01
7.51562059e-01 5.60821116e-01 2.45601043e-01 -9.03654814e-01
3.63608301e-01 -6.29474759e-01 -7.62714744e-01 4.15974110e-01
-5.49131513e-01 -4.50578779e-02 4.78085876e-01 -3.99271548e-01
-1.34059238e+00 4.39321578e-01 1.77493393e-01 -4.32374895e-01
-9.37100574e-02 8.88860047e-01 -4.50048357e-01 -3.94752592e-01
7.56000578e-01 4.80017543e-01 2.87359923e-01 -4.80775386e-01
2.40795374e-01 8.67076576e-01 4.98705596e-01 -5.46450615e-01
1.14877069e+00 7.85377920e-01 1.18387148e-01 -1.36577225e+00
-8.01015496e-01 -9.05144632e-01 -5.01830757e-01 -7.38975555e-02
5.81292450e-01 -1.15719450e+00 -7.96267986e-02 6.60735726e-01
-7.41104901e-01 -2.05105036e-01 -1.73080131e-01 6.42219543e-01
-2.98796684e-01 7.76817203e-01 -4.82332826e-01 -1.11368203e+00
-2.07009628e-01 -1.24364293e+00 6.44242942e-01 1.49596259e-02
3.09035003e-01 -9.55453932e-01 -4.48086187e-02 3.91040951e-01
4.19129193e-01 5.32841027e-01 9.05873001e-01 1.06476709e-01
-4.23140794e-01 3.87835577e-02 -3.82056266e-01 5.68914473e-01
6.17002070e-01 1.57230169e-01 -1.01490688e+00 -4.37319428e-01
4.06284094e-01 -9.80650708e-02 9.29496765e-01 6.23864293e-01
8.85171235e-01 -3.94651473e-01 -1.24815539e-01 9.17905271e-01
1.82551980e+00 6.07301407e-02 5.46594560e-01 2.63983607e-01
9.13630188e-01 8.56855512e-01 5.16392112e-01 4.28231955e-01
-1.70973614e-01 3.22130740e-01 5.88162005e-01 -6.02485776e-01
9.65894833e-02 1.61337391e-01 3.11868876e-01 8.99233401e-01
-2.05218926e-01 -2.40087509e-01 -6.09005451e-01 5.07902384e-01
-1.99782777e+00 -9.74695861e-01 -5.64641714e-01 2.41389775e+00
7.72746146e-01 -7.10617065e-01 -1.73741616e-02 2.89740622e-01
1.09586406e+00 5.37868857e-01 -5.15463591e-01 3.82268280e-01
-6.38148844e-01 -1.96722105e-01 8.28085661e-01 6.42885327e-01
-1.24344242e+00 4.68291402e-01 6.33464241e+00 5.93798876e-01
-1.30742228e+00 -2.46169101e-02 3.50391328e-01 2.43402761e-03
-5.24939477e-01 6.25242665e-02 -2.11905167e-01 4.28151488e-01
4.64750141e-01 9.52780768e-02 8.25235724e-01 3.11171412e-01
5.96487284e-01 -1.42866790e-01 -4.18109268e-01 1.08041847e+00
-3.24144177e-02 -9.78278160e-01 5.65402582e-02 1.25644386e-01
1.04805529e+00 -3.50440457e-03 -7.12631419e-02 -3.42720121e-01
2.52783656e-01 -7.33266056e-01 7.62372017e-01 3.39459211e-01
7.56206393e-01 -5.45367897e-01 6.31650388e-01 5.17618299e-01
-1.06017137e+00 -1.18474349e-01 -5.86443663e-01 -8.25729147e-02
1.17978744e-01 1.20683706e+00 -2.07909077e-01 7.61580050e-01
3.76864761e-01 7.69821763e-01 -2.33733449e-02 9.52486098e-01
-2.57046700e-01 5.53147852e-01 -5.09544373e-01 5.96460104e-01
1.74934074e-01 -1.08222961e+00 7.63753951e-01 1.07314420e+00
6.07666910e-01 4.30727243e-01 3.11167330e-01 1.03372538e+00
1.57168850e-01 4.47171926e-01 -5.45013249e-01 -2.70355850e-01
2.02584267e-01 1.32536197e+00 -6.88020825e-01 -3.07974994e-01
-4.83418941e-01 7.93975830e-01 -1.76016882e-01 9.04732049e-01
-4.53758091e-01 1.54873366e-02 7.13024080e-01 2.53919750e-01
-1.34789990e-02 -1.68114722e-01 -3.86163026e-01 -1.51994216e+00
-7.38134608e-02 -1.17558777e+00 3.54990989e-01 -6.76222086e-01
-1.38222778e+00 3.75522137e-01 -1.27103716e-01 -1.35513115e+00
3.44400883e-01 -5.40534079e-01 -5.02957880e-01 1.05702889e+00
-1.66993797e+00 -1.22830057e+00 -5.20833850e-01 5.00511289e-01
2.67712027e-01 1.27315700e-01 6.03472948e-01 2.77147144e-01
-8.22758794e-01 -2.94016510e-01 6.60179317e-01 -2.74516463e-01
5.60174704e-01 -1.07954872e+00 -5.33551216e-01 1.25907111e+00
-1.17464602e-01 5.85658371e-01 8.23784828e-01 -7.65033484e-01
-1.21130645e+00 -1.27747643e+00 3.83825570e-01 2.75924295e-01
6.45235360e-01 1.76127572e-02 -1.07698107e+00 4.94884282e-01
9.63107049e-02 1.14485547e-01 9.44640338e-01 -1.05497420e-01
-1.08161300e-01 -2.91797251e-01 -1.12407053e+00 4.07094896e-01
5.67285120e-01 -4.27287370e-01 -1.70670435e-01 6.07833683e-01
3.06415349e-01 -2.02115282e-01 -8.01058769e-01 7.27185786e-01
3.57770711e-01 -9.01453435e-01 9.93261695e-01 -1.71139792e-01
1.77489787e-01 -9.09748375e-01 -4.09732044e-01 -1.39000332e+00
-6.43229663e-01 -4.02115852e-01 9.75294411e-02 1.06814265e+00
4.24134076e-01 -6.12848163e-01 4.29827332e-01 4.85380709e-01
-6.43699095e-02 -1.10752888e-01 -5.30124962e-01 -7.93250620e-01
-2.89839000e-01 -6.45500496e-02 6.54284775e-01 1.35908818e+00
-2.19452932e-01 1.17271617e-01 -6.63643837e-01 8.00268888e-01
1.04325950e+00 3.09463441e-01 4.45492923e-01 -1.36631525e+00
-1.59835204e-01 -2.59377211e-01 1.85637623e-01 -5.72699130e-01
2.23166004e-01 -6.69094563e-01 4.52237755e-01 -1.39715993e+00
3.42252195e-01 -4.86443937e-01 -4.35685277e-01 4.97514516e-01
-4.83559340e-01 3.83261681e-01 4.06570174e-02 7.06643581e-01
5.58779538e-02 6.06867075e-01 1.17673588e+00 -4.93636400e-01
-5.11574626e-01 -1.04269885e-01 -6.00562930e-01 7.34114826e-01
6.72772467e-01 -3.49400938e-01 -4.14378226e-01 -4.40246671e-01
3.80047634e-02 6.40350506e-02 1.58247456e-01 -9.62932587e-01
-7.02709556e-02 -7.23111629e-01 1.29025817e-01 -3.91077250e-01
2.19415173e-01 -1.02838635e+00 8.56642783e-01 2.44918019e-01
1.01218335e-02 -5.34822583e-01 9.49645564e-02 4.77564871e-01
-3.83131802e-01 -4.09379840e-01 1.13598025e+00 -2.20566824e-01
-6.03410304e-01 3.30088317e-01 -4.11061168e-01 -5.14444709e-01
8.08375239e-01 -2.70122111e-01 -1.04929447e-01 -3.83129716e-01
-4.96978432e-01 3.69289480e-02 7.97062874e-01 -1.87329948e-01
3.19405764e-01 -1.26950788e+00 -8.04969668e-01 2.64754653e-01
1.82184443e-01 7.93374702e-02 3.30689639e-01 9.89216268e-01
-7.45002925e-01 -1.51319979e-02 7.22074183e-03 -4.54775780e-01
-9.99127507e-01 7.14500010e-01 4.83402938e-01 -8.55174661e-02
-3.07004869e-01 5.06179750e-01 5.22787154e-01 -5.48757434e-01
-1.37497455e-01 -2.06387550e-01 -5.73931038e-02 1.20970309e-01
5.52789330e-01 4.69251007e-01 3.85956652e-02 -1.07019937e+00
1.61283072e-02 6.26380503e-01 6.74174964e-01 -2.43691597e-02
1.35184944e+00 -4.02511835e-01 -7.67846584e-01 4.40341949e-01
1.00485599e+00 1.88593149e-01 -1.34038472e+00 -2.85003275e-01
-2.78404057e-01 -5.48398554e-01 5.15558898e-01 -4.66492772e-01
-1.23589325e+00 6.29812777e-01 4.94105667e-01 1.86158821e-01
1.43452156e+00 -6.51798248e-01 5.23941338e-01 8.35952535e-02
6.16232306e-02 -1.03625512e+00 -3.96638721e-01 1.79664165e-01
7.99945533e-01 -1.33402383e+00 3.84453058e-01 -8.51583838e-01
-3.19863737e-01 9.98072982e-01 3.45962137e-01 2.36773178e-01
5.01242399e-01 1.05981559e-01 4.09447044e-01 -2.07719043e-01
9.74865705e-02 -4.35582340e-01 2.57243991e-01 4.84579712e-01
2.21832395e-01 2.46054351e-01 -2.86038518e-01 2.06585880e-02
3.29344630e-01 -2.84320176e-01 4.33147639e-01 7.66003728e-01
-6.92844331e-01 -9.19210970e-01 -1.09905303e+00 4.06775981e-01
-2.57910430e-01 -1.66149378e-01 -1.27250984e-01 3.03886652e-01
3.09117258e-01 1.10821891e+00 -2.14215770e-01 -8.71875584e-02
1.95776835e-01 -5.41347191e-02 1.29494667e-01 -6.84054315e-01
-1.09711058e-01 6.64370716e-01 -1.92318380e-01 -1.54159993e-01
-1.02061892e+00 -5.52316606e-01 -1.01995146e+00 -2.15871125e-01
-7.20746577e-01 1.97678372e-01 5.61171591e-01 9.32749748e-01
-8.30175504e-02 2.21507534e-01 7.15371549e-01 -1.00119758e+00
-5.94124794e-01 -9.73919570e-01 -1.42917752e+00 5.30125320e-01
4.91381705e-01 -5.78109920e-01 -6.58176124e-01 2.28281021e-01] | [10.053078651428223, -2.090796709060669] |
7a3a5e70-a0b5-4ddc-a42e-601f99fef0f2 | portfolio-optimization-with-relative-tail | 2303.12209 | null | https://arxiv.org/abs/2303.12209v2 | https://arxiv.org/pdf/2303.12209v2.pdf | Portfolio Optimization with Relative Tail Risk | This paper proposes analytic forms of portfolio CoVaR and CoCVaR on the normal tempered stable market model. Since CoCVaR captures the relative risk of the portfolio with respect to a benchmark return, we apply it to the relative portfolio optimization. Moreover, we derive analytic forms for the marginal contribution to CoVaR and the marginal contribution to CoCVaR. We discuss the Monte-Carlo simulation method to calculate CoCVaR and the marginal contributions of CoVaR and CoCVaR. As the empirical illustration, we show relative portfolio optimization with thirty stocks under the distress condition of the Dow Jones Industrial Average. Finally, we perform the risk budgeting method to reduce the CoVaR and CoCVaR of the portfolio based on the marginal contributions to CoVaR and CoCVaR. | ['Young Shin Kim'] | 2023-03-21 | null | null | null | null | ['portfolio-optimization'] | ['time-series'] | [-5.49654067e-01 -1.73496887e-01 3.70545059e-01 7.84650594e-02
-6.51209712e-01 -9.94747818e-01 5.11285484e-01 -1.93088979e-01
-1.83820948e-01 6.98774219e-01 -4.20115143e-02 -5.57105124e-01
-6.10967219e-01 -1.00748110e+00 -3.23676050e-01 -7.67642856e-01
3.97585239e-03 3.28183055e-01 -8.26988295e-02 -1.02315165e-01
4.18235093e-01 2.96668738e-01 -6.68312550e-01 -5.12756169e-01
4.19157743e-01 1.34372270e+00 -3.56018811e-01 3.87068897e-01
2.74295896e-01 4.34672117e-01 -7.27519989e-01 -1.12484860e+00
8.76935720e-01 -2.36585915e-01 2.55412400e-01 -4.07885432e-01
-3.35265934e-01 -3.60665292e-01 1.51278317e-01 1.27639377e+00
3.20676446e-01 1.86027303e-01 1.15363097e+00 -9.38085854e-01
-9.52214992e-04 7.66545236e-01 -8.45982790e-01 5.89957654e-01
-9.78706703e-02 -1.70588363e-02 1.32613719e+00 -7.55673826e-01
2.27234475e-02 1.10329056e+00 5.74364007e-01 -8.22839215e-02
-9.39610839e-01 -5.69538832e-01 4.83759940e-02 -5.40511787e-01
-1.06872082e+00 2.33939022e-01 3.81415874e-01 -6.57941282e-01
6.41100407e-01 3.77652943e-01 7.19626904e-01 3.68980080e-01
1.09933543e+00 1.34101376e-01 9.77375329e-01 -1.74165040e-01
4.69022214e-01 -4.53109108e-02 3.92688364e-01 -2.32244179e-01
1.01179838e+00 5.60980499e-01 -1.01366103e-01 -5.67109704e-01
7.91023552e-01 7.83434436e-02 8.34091604e-02 9.29127112e-02
-6.50785923e-01 7.93242097e-01 -4.08812404e-01 -2.31779575e-01
-5.98860979e-01 3.88681501e-01 -5.83874509e-02 2.46756822e-01
4.22148824e-01 4.49757069e-01 -3.45202088e-01 -8.95403549e-02
-6.77688420e-01 5.92078805e-01 1.03907061e+00 9.95067120e-01
2.92663604e-01 3.22378695e-01 -3.64034504e-01 1.96379274e-01
6.48288071e-01 1.20020127e+00 -7.73889571e-02 -1.20886207e+00
6.62706077e-01 1.33799657e-01 4.59605902e-01 -6.71551764e-01
8.98427889e-02 -7.95822740e-01 -3.38810593e-01 3.74489546e-01
4.26218748e-01 -5.82004011e-01 -9.72841233e-02 1.41137815e+00
-1.80411711e-01 4.89375368e-02 1.35973468e-01 3.67655098e-01
-3.06644380e-01 5.42960465e-01 -4.33113247e-01 -7.86371708e-01
1.14023876e+00 -4.64366078e-01 -8.61758828e-01 5.05398102e-02
5.36270477e-02 -8.45292926e-01 4.10106689e-01 3.46989959e-01
-1.27612340e+00 1.97150305e-01 -1.08119643e+00 1.10015845e+00
1.75318688e-01 -5.14654458e-01 2.20375881e-01 1.05772102e+00
-4.75606829e-01 6.77720249e-01 -6.67613089e-01 6.72326803e-01
8.74681748e-04 6.61051497e-02 4.73343909e-01 5.73233902e-01
-1.00134170e+00 9.18286085e-01 -2.71736532e-02 2.76805639e-01
-8.98541570e-01 -9.95064557e-01 -3.01156759e-01 3.78243715e-01
3.74139011e-01 -4.58052367e-01 1.23180687e+00 -4.10836577e-01
-1.50554931e+00 -6.38784841e-02 5.44613421e-01 -6.79067612e-01
7.98178136e-01 -5.04728377e-01 -7.13386014e-02 -1.32111818e-01
-5.08331470e-02 -7.90362120e-01 5.83351612e-01 -7.12677479e-01
-2.99522460e-01 -8.60158876e-02 -1.94256946e-01 -5.53311780e-02
1.41351819e-01 3.05339932e-01 1.11160561e-01 -1.17184532e+00
4.85923328e-03 -9.85236049e-01 -4.98799533e-01 -8.65068078e-01
-2.28720769e-01 2.71288574e-01 2.27857437e-02 -6.33804977e-01
1.25849593e+00 -2.27842689e+00 -2.06808537e-01 8.27755511e-01
-1.65284276e-01 -4.01344359e-01 3.57125223e-01 5.12928367e-01
-1.37508750e-01 2.68147767e-01 -2.60106057e-01 1.47750929e-01
4.65931892e-01 -1.79771453e-01 -9.89886582e-01 5.46245933e-01
4.06835340e-02 5.84553123e-01 -3.61736417e-01 5.50470166e-02
-1.45588160e-01 -2.48346418e-01 -5.70142627e-01 3.88103873e-01
-3.58491093e-02 -2.85227954e-01 -5.55828214e-01 5.51683128e-01
9.31662142e-01 1.55348539e-01 2.54994482e-01 2.17477888e-01
-2.62300998e-01 -1.68722838e-01 -1.43225133e+00 1.41362280e-01
-1.95627376e-01 3.98271307e-02 -2.01636013e-02 -4.01589245e-01
1.00587797e+00 3.44547890e-02 2.81165540e-01 -2.72340328e-02
4.85826641e-01 2.39070818e-01 1.24254815e-01 2.54955828e-01
4.68681604e-01 -9.94795442e-01 -4.08083409e-01 1.10274422e+00
-1.66166797e-01 -2.24898264e-01 1.02171414e-01 6.94089532e-02
6.49404824e-01 -1.77851871e-01 4.16668028e-01 -6.59492791e-01
1.73591018e-01 -6.21388018e-01 6.37163639e-01 5.67091644e-01
-1.24986574e-01 4.24176812e-01 1.18423796e+00 1.17716081e-01
-8.07912588e-01 -1.49821305e+00 -1.97683647e-01 4.58396405e-01
1.10224515e-01 4.35481817e-02 -7.04345465e-01 -3.10816079e-01
4.70792234e-01 1.05352664e+00 -4.76787388e-01 1.79759762e-03
-3.01431984e-01 -1.42651558e+00 1.51905671e-01 4.82212454e-01
4.36925106e-02 -5.47461212e-01 -6.73372626e-01 1.50444075e-01
4.02504832e-01 -4.81415778e-01 -7.26264060e-01 2.30456054e-01
-6.36303246e-01 -1.06419921e+00 -7.07001269e-01 1.97425559e-01
3.28285724e-01 -6.80897087e-02 8.39681029e-01 -5.46608150e-01
1.48392335e-01 4.98951852e-01 9.88701209e-02 -1.06291282e+00
-4.25593555e-02 -6.10863268e-01 1.78730160e-01 1.67870700e-01
-2.53394358e-02 -3.49565536e-01 -5.26225984e-01 4.66994822e-01
-2.93047130e-01 -8.60744834e-01 5.17285950e-02 5.52202702e-01
6.36743665e-01 1.60640150e-01 7.31277883e-01 -4.65546638e-01
9.79271829e-01 -5.35658658e-01 -1.41866159e+00 4.32450891e-01
-1.13133478e+00 5.59223741e-02 1.46849975e-02 -5.41160643e-01
-1.46853185e+00 -6.57396317e-01 5.32917261e-01 -3.94489110e-01
8.71530116e-01 6.38710916e-01 -1.59569815e-01 1.66122332e-01
-5.69664240e-02 -3.09757769e-01 -4.81629819e-02 -4.96615618e-01
3.15749496e-01 2.00350136e-01 3.09336454e-01 -7.12087035e-01
1.09541655e+00 4.31663066e-01 1.93585739e-01 -1.55682256e-02
-8.65889370e-01 2.75713694e-03 -2.64751345e-01 -3.46018940e-01
7.78076708e-01 -7.98638284e-01 -8.67125332e-01 3.96737754e-01
-7.10343301e-01 -1.42392799e-01 -7.31548727e-01 9.55009818e-01
-8.59542310e-01 8.17106441e-02 -4.73523229e-01 -1.47978616e+00
-4.86070096e-01 -8.89104366e-01 2.79310077e-01 2.54385203e-01
-8.68476033e-02 -1.10667527e+00 5.46969295e-01 -6.08770438e-02
3.43018115e-01 5.81069231e-01 8.08740735e-01 -9.06488836e-01
-7.94803321e-01 -4.19600725e-01 3.41285802e-02 6.39327228e-01
-2.66419798e-01 1.23644911e-01 -3.55312556e-01 -2.45078638e-01
8.45595777e-01 4.20930982e-01 4.01092678e-01 7.64712274e-01
2.72600174e-01 -4.96096939e-01 2.12328881e-01 5.57306290e-01
1.61517942e+00 7.24731743e-01 4.56528127e-01 5.78946590e-01
-1.51212150e-02 6.99559093e-01 1.05667162e+00 1.09216225e+00
-2.13808715e-01 1.48331404e-01 2.20701590e-01 7.55294263e-01
6.55192733e-01 7.42828622e-02 7.42787123e-01 9.89620626e-01
-1.27884820e-01 -9.43533331e-02 -6.71407223e-01 3.34428579e-01
-1.57736695e+00 -1.16055512e+00 -1.54913262e-01 2.25979161e+00
3.19986254e-01 4.74534988e-01 1.17907874e-01 -4.19369698e-01
7.52502143e-01 1.05611391e-01 -4.52816963e-01 -4.74037111e-01
-3.18214715e-01 2.02593356e-01 1.07477808e+00 5.30787468e-01
-6.59886718e-01 1.45059034e-01 7.89258623e+00 6.83439970e-01
-4.79268163e-01 -2.33532526e-02 7.60220885e-01 -4.36657995e-01
-7.09983587e-01 3.67803395e-01 -1.08864677e+00 5.93422830e-01
1.06451666e+00 -1.15603435e+00 2.19984427e-01 9.38144147e-01
9.43669006e-02 -2.55588293e-01 -9.27557826e-01 4.11504447e-01
-2.60872751e-01 -1.00377834e+00 -1.48255721e-01 5.36448836e-01
8.10739636e-01 -2.86119878e-01 4.27408367e-01 1.80113047e-01
4.94505405e-01 -6.97279274e-01 1.08783388e+00 1.10744441e+00
3.36131424e-01 -1.14506924e+00 1.33366835e+00 8.87197629e-02
-1.03917027e+00 -1.76506370e-01 -5.21220982e-01 -1.00460634e-01
5.19423544e-01 5.92466056e-01 -2.92444974e-01 4.50397253e-01
3.64361614e-01 1.72176614e-01 -2.06435293e-01 9.35040653e-01
1.34285182e-01 6.29627824e-01 -4.48167980e-01 -3.66768278e-02
-5.73413707e-02 -1.07126153e+00 7.14970052e-01 6.33602142e-01
7.72661984e-01 2.43449628e-01 -5.07337749e-01 1.10727012e+00
1.58048362e-01 9.64297876e-02 -3.58428895e-01 -1.21239079e-02
4.94583637e-01 8.96822929e-01 -4.79228705e-01 5.32498695e-02
-3.64186525e-01 -2.37032231e-02 -5.09717107e-01 3.69235247e-01
-7.35170305e-01 -3.62116069e-01 6.99825704e-01 -1.29393050e-02
5.02940178e-01 -7.14963451e-02 -7.49971926e-01 -8.91569614e-01
8.13385248e-02 -7.19540000e-01 5.26037514e-01 -2.97622323e-01
-1.47805679e+00 1.73200071e-01 5.06259799e-01 -1.22538209e+00
-3.95171493e-01 -6.13408923e-01 -1.25078893e+00 1.27692509e+00
-1.13182020e+00 -1.01445898e-01 4.78402048e-01 1.23885185e-01
-3.73783857e-02 -8.89820099e-01 2.17836067e-01 -3.02441746e-01
-9.16707695e-01 4.80848283e-01 8.78432691e-01 -7.44669363e-02
2.98920691e-01 -1.13453829e+00 4.48756486e-01 1.09354699e+00
-3.99947613e-01 9.07267928e-01 8.36588204e-01 -9.68144178e-01
-1.21330416e+00 -8.51237476e-01 1.50042668e-01 -6.35707378e-01
1.26999021e+00 1.50784940e-01 -4.79232073e-01 8.07448447e-01
2.01172173e-01 -4.27685678e-01 7.85471082e-01 -1.95308298e-01
-4.41476814e-02 -3.09814900e-01 -1.08538866e+00 6.69544697e-01
1.29106790e-01 -2.95635968e-01 -8.03920567e-01 -4.57697250e-02
9.62817907e-01 2.20386252e-01 -1.25770247e+00 2.47522548e-01
8.17950726e-01 -9.13210273e-01 8.38702977e-01 -2.75179565e-01
-1.54471397e-01 -1.14328139e-01 -4.32666183e-01 -1.05849123e+00
-1.71783656e-01 -1.08613002e+00 -1.43604362e-02 1.32657278e+00
4.88315254e-01 -9.50155556e-01 3.71742427e-01 7.83206820e-01
4.13733602e-01 -4.49596167e-01 -1.22138655e+00 -1.15409577e+00
4.53723073e-01 -4.57033187e-01 7.57148802e-01 2.87006557e-01
-1.31957069e-01 -2.46024475e-01 -3.41134936e-01 1.91236082e-02
1.23349762e+00 3.83723170e-01 3.66313487e-01 -8.95028055e-01
-6.06962264e-01 -5.54034948e-01 2.22872764e-01 -2.24142298e-01
-8.60076323e-02 -5.36468327e-01 -2.04817593e-01 -7.13192701e-01
4.28089321e-01 -2.19260417e-02 -6.42389476e-01 -4.15780485e-01
-1.34855151e-01 -3.71503353e-01 6.74089909e-01 2.78522015e-01
-3.57258618e-02 5.44519246e-01 8.98441374e-01 1.50657967e-01
5.51866628e-02 5.24404407e-01 -7.86539137e-01 8.95471752e-01
7.42285728e-01 -7.28468180e-01 -3.55267406e-01 3.22647303e-01
6.68228149e-01 4.92454261e-01 3.48281972e-02 -4.10226643e-01
-2.56210327e-01 -6.55974030e-01 -1.05248794e-01 -1.04076982e+00
-2.96423230e-02 -5.52025199e-01 8.09833109e-01 6.18835330e-01
-1.44838735e-01 4.07034516e-01 -9.68006477e-02 9.12518620e-01
-7.38855526e-02 -8.47524703e-01 7.43090332e-01 -7.44168460e-02
6.15626693e-01 7.96654597e-02 -6.57964170e-01 2.72240043e-01
1.22848570e+00 2.39494011e-01 -3.46220970e-01 -5.85722029e-01
-7.01005757e-01 3.96647960e-01 2.33652487e-01 -1.14956796e-01
5.25842667e-01 -1.30763268e+00 -6.50120556e-01 -7.32502937e-02
-4.41968203e-01 -4.88699019e-01 1.09325014e-01 7.79876709e-01
-5.44228852e-01 3.53844494e-01 -6.90991953e-02 1.53709620e-01
-7.63337731e-01 5.14762521e-01 7.62245953e-01 -5.27667642e-01
-1.25247568e-01 5.73828876e-01 7.08759308e-01 2.69308448e-01
1.33973239e-02 -2.70289332e-01 -9.03776214e-02 3.57823670e-01
6.54026330e-01 9.71020997e-01 -3.64837408e-01 -1.20557822e-01
-3.65475714e-01 7.12578833e-01 2.20675096e-01 -8.06721807e-01
1.20738447e+00 -2.02166766e-01 -4.27755117e-01 6.62793994e-01
7.59667516e-01 3.95300895e-01 -1.40035868e+00 1.98966503e-01
4.51916903e-01 -4.20359015e-01 -3.23214889e-01 -3.99278164e-01
-9.78634834e-01 6.88357770e-01 4.20476973e-01 1.78558648e-01
7.96894252e-01 -2.70067751e-01 3.49686474e-01 2.98229545e-01
2.67535210e-01 -1.11614048e+00 5.60468470e-04 5.39654851e-01
1.04763305e+00 -3.10263038e-01 4.22866464e-01 -1.26066908e-01
-8.89772892e-01 9.14933681e-01 7.08424225e-02 -7.59343982e-01
1.48342729e+00 6.39897823e-01 8.74278471e-02 2.69170646e-02
-9.21770751e-01 4.15621459e-01 2.63181508e-01 1.27831355e-01
-8.45302865e-02 3.85164171e-01 -7.55648136e-01 1.39702678e+00
-5.47880590e-01 -6.98421717e-01 1.03951776e+00 7.94675827e-01
-4.54850167e-01 -8.13212574e-01 -6.15184188e-01 4.49611992e-01
-1.07771003e+00 -1.80918589e-01 -1.03267670e-01 5.69833875e-01
-4.00399119e-01 6.70697510e-01 2.39747986e-01 -4.31130193e-02
6.80301785e-01 -4.83307019e-02 6.19245842e-02 -3.70601356e-01
-7.77323365e-01 9.20885146e-01 2.45054197e-02 -1.21778913e-01
1.50352091e-01 -1.29435205e+00 -7.76219130e-01 -3.12242657e-01
-6.49772644e-01 2.92115688e-01 1.88695252e-01 8.35590005e-01
-3.35316777e-01 3.78434032e-01 9.79408026e-01 -3.99046063e-01
-1.49914396e+00 -9.05960083e-01 -1.34549904e+00 -1.05465695e-01
-3.08061275e-03 -6.94732726e-01 -9.48998690e-01 -3.31889302e-01] | [4.94803524017334, 3.9578661918640137] |
6e3e3d32-8f34-44f7-9cc7-e1a68a9b68a0 | fine-grained-object-semantic-understanding | 1912.12577 | null | https://arxiv.org/abs/1912.12577v2 | https://arxiv.org/pdf/1912.12577v2.pdf | Human Correspondence Consensus for 3D Object Semantic Understanding | Semantic understanding of 3D objects is crucial in many applications such as object manipulation. However, it is hard to give a universal definition of point-level semantics that everyone would agree on. We observe that people have a consensus on semantic correspondences between two areas from different objects, but are less certain about the exact semantic meaning of each area. Therefore, we argue that by providing human labeled correspondences between different objects from the same category instead of explicit semantic labels, one can recover rich semantic information of an object. In this paper, we introduce a new dataset named CorresPondenceNet. Based on this dataset, we are able to learn dense semantic embeddings with a novel geodesic consistency loss. Accordingly, several state-of-the-art networks are evaluated on this correspondence benchmark. We further show that CorresPondenceNet could not only boost fine-grained understanding of heterogeneous objects but also cross-object registration and partial object matching. | ['Yang You', 'Chengkun Li', 'Zhoujun Cheng', 'Cewu Lu', 'Yujing Lou', 'Weiming Wang', 'Lizhuang Ma', 'Liangwei Li'] | 2019-12-29 | human-correspondence-consensus-for-3d-object | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/4107_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123670494.pdf | eccv-2020-8 | ['3d-feature-matching', '3d-point-cloud-matching'] | ['computer-vision', 'computer-vision'] | [-6.20767251e-02 2.39816420e-02 -6.63082078e-02 -6.70224130e-01
-5.22322953e-01 -6.78962052e-01 6.37358427e-01 4.98315156e-01
-3.52460265e-01 1.13806419e-01 7.92482942e-02 2.43902519e-01
-3.10431212e-01 -1.05786204e+00 -8.05194199e-01 -3.82957220e-01
3.10172290e-01 9.05170321e-01 4.88496900e-01 -2.94476092e-01
1.85259268e-01 7.02376366e-01 -1.55288005e+00 2.71388590e-01
6.11053526e-01 1.10374177e+00 3.60792726e-01 -1.35359481e-01
-3.62948209e-01 1.56578019e-01 -3.08977604e-01 -4.39449519e-01
3.42943072e-01 -1.58219263e-01 -1.12100708e+00 3.52227762e-02
7.81630576e-01 -4.94696647e-02 -3.14075470e-01 1.27992499e+00
1.20553866e-01 1.00967631e-01 7.60351002e-01 -1.41211402e+00
-1.01590431e+00 4.11888301e-01 -3.56385201e-01 -1.10190682e-01
2.86058426e-01 1.28579354e-02 1.42070925e+00 -7.70388484e-01
7.04856873e-01 1.39188147e+00 5.87627172e-01 4.62164044e-01
-1.11961937e+00 -4.36412364e-01 3.07137221e-01 4.13580269e-01
-1.46182716e+00 -5.15808277e-02 9.27110791e-01 -4.47598726e-01
5.96673727e-01 4.73587327e-02 6.83014989e-01 9.46711421e-01
-1.31447867e-01 5.05962074e-01 1.04838836e+00 -1.28991529e-01
1.77782878e-01 1.88461185e-01 3.30577433e-01 5.71859062e-01
4.21593726e-01 -3.21758278e-02 -4.69109267e-01 1.76431343e-01
7.09896564e-01 1.52894288e-01 -1.49225071e-01 -9.48791265e-01
-1.55335927e+00 8.37937415e-01 1.04195881e+00 4.76490229e-01
-1.52696505e-01 3.22342724e-01 2.01190606e-01 1.57027796e-01
4.44060445e-01 5.03366590e-01 -4.83752519e-01 1.67379871e-01
-3.23668271e-01 3.90745729e-01 6.93825006e-01 1.20815229e+00
1.07670546e+00 -5.37840784e-01 9.53194574e-02 4.88212764e-01
4.51042175e-01 5.86613536e-01 -3.91133055e-02 -1.16798556e+00
1.54789090e-01 9.87498164e-01 1.41462550e-01 -1.21072388e+00
-4.40287650e-01 -3.90537381e-01 -6.55631781e-01 3.46740484e-01
6.03195369e-01 5.67474306e-01 -8.18853855e-01 1.93326128e+00
3.19512755e-01 3.30313176e-01 -2.06617579e-01 1.25026298e+00
9.06586349e-01 1.06440097e-01 7.04041645e-02 7.20594525e-01
1.63773561e+00 -6.96713567e-01 -2.40029410e-01 -2.53869534e-01
4.90164787e-01 -5.85985065e-01 9.75852191e-01 -1.19582489e-01
-8.51365983e-01 -5.92148781e-01 -9.83273685e-01 -4.97208953e-01
-6.96191430e-01 -3.04183632e-01 8.59838605e-01 4.61156249e-01
-7.79873610e-01 7.41379559e-01 -8.64422619e-01 -7.53110170e-01
6.28023863e-01 1.84395507e-01 -6.85971618e-01 -1.67594045e-01
-8.99107456e-01 1.26647031e+00 4.01853710e-01 -9.94405746e-02
-7.17088878e-01 -1.02916145e+00 -9.22140121e-01 6.70877993e-02
2.13869750e-01 -1.16203523e+00 7.73223698e-01 -6.34742320e-01
-9.04864609e-01 1.66140819e+00 -6.60883039e-02 -2.20691219e-01
4.41398293e-01 -2.18823448e-01 -1.94355518e-01 5.44060729e-02
3.26348960e-01 8.25646639e-01 5.06801546e-01 -1.54002964e+00
-4.97638077e-01 -8.84529710e-01 3.32019478e-01 1.05359875e-01
1.34637371e-01 -1.34844810e-01 -1.99954689e-01 -4.12800550e-01
5.97328007e-01 -8.97733569e-01 1.57478973e-01 7.03164518e-01
-3.87611926e-01 -5.74890435e-01 8.08356464e-01 -4.09338892e-01
3.04198533e-01 -2.17808175e+00 3.58331829e-01 2.05231965e-01
4.58733618e-01 -2.29578748e-01 -2.56967008e-01 8.73206183e-02
-4.97276634e-02 1.52773157e-01 -3.61926615e-01 -7.71322399e-02
4.71111804e-01 3.07088822e-01 -2.75186449e-01 8.17821562e-01
1.54521719e-01 1.32842934e+00 -1.08124948e+00 -1.68058127e-01
3.81587178e-01 5.54971159e-01 -5.10627568e-01 8.54664594e-02
-3.77920002e-01 6.55029237e-01 -7.51885235e-01 5.78269303e-01
8.34629297e-01 -5.44094384e-01 -1.12004355e-01 -5.85703015e-01
2.89057314e-01 2.98522323e-01 -1.13317919e+00 2.35693026e+00
-4.46180314e-01 5.31925857e-01 -1.67141229e-01 -1.25811994e+00
9.71094489e-01 1.52547285e-03 6.34962380e-01 -6.80258274e-01
2.77174473e-01 2.26397499e-01 -1.28436223e-01 -2.87429899e-01
3.64309162e-01 -2.49388933e-01 -1.61825538e-01 5.46280861e-01
1.24575466e-01 -5.74248731e-01 -2.90109605e-01 7.71849998e-04
8.22264135e-01 1.52859986e-01 6.17408082e-02 -4.10320699e-01
4.11602139e-01 1.03882566e-01 2.17418969e-01 6.30589366e-01
-1.07600398e-01 8.56322050e-01 8.73157904e-02 -5.34399092e-01
-1.14776027e+00 -1.66862369e+00 -3.01523417e-01 8.05947781e-01
1.05668294e+00 -2.46821240e-01 -5.49347699e-01 -7.14258909e-01
4.20366317e-01 5.12084484e-01 -6.32492900e-01 -1.91209272e-01
-4.91812557e-01 -2.38868624e-01 1.50007546e-01 5.28820395e-01
5.87605476e-01 -9.28409517e-01 -4.42170769e-01 -5.79878092e-02
-2.75964230e-01 -1.46956694e+00 -4.68146533e-01 -2.39848256e-01
-7.73846626e-01 -1.35387218e+00 -4.67235804e-01 -8.26924920e-01
7.21041322e-01 7.11190522e-01 1.49972939e+00 2.26926252e-01
-3.36726636e-01 5.57168007e-01 -4.17890072e-01 -3.06298047e-01
-3.68869334e-01 1.70032069e-01 -1.71501888e-03 6.70759380e-02
7.83344746e-01 -5.96652448e-01 -5.92290998e-01 5.83532453e-01
-6.93820953e-01 -4.98517826e-02 4.05083001e-01 3.85104001e-01
6.37593925e-01 -2.70222217e-01 2.17119366e-01 -6.72536492e-01
4.06171978e-02 -3.88788104e-01 -4.58735794e-01 4.12913173e-01
-2.63353676e-01 3.16425800e-01 1.00248858e-01 -1.97151393e-01
-8.60732675e-01 -1.99998751e-01 -2.84026619e-02 -3.90437782e-01
-5.12401819e-01 8.84543806e-02 -3.56944710e-01 -2.50912368e-01
4.69055176e-01 -6.40492439e-02 -3.23426910e-02 -7.08999813e-01
8.33381712e-01 3.46784621e-01 5.72863877e-01 -8.72753441e-01
1.05435324e+00 9.37181771e-01 1.39520228e-01 -5.25500298e-01
-1.15911043e+00 -6.39968753e-01 -9.34541643e-01 1.35793522e-01
1.28634548e+00 -1.03717935e+00 -7.43525267e-01 3.69231164e-01
-1.38293397e+00 2.58064326e-02 -5.47439992e-01 4.69916046e-01
-8.58496189e-01 2.11295664e-01 -2.13069692e-01 1.07179686e-01
9.42728370e-02 -9.85376537e-01 1.45615661e+00 9.20466557e-02
-3.59146297e-01 -1.01723886e+00 -1.77789599e-01 5.45354843e-01
2.66922832e-01 1.80780634e-01 1.04113019e+00 -6.68310165e-01
-1.03578448e+00 -6.74121007e-02 -6.64830089e-01 1.47007138e-01
3.42650056e-01 -4.51297253e-01 -1.01215553e+00 -1.53403938e-01
-1.94770582e-02 -1.15868576e-01 6.84158802e-01 1.59123987e-01
1.21925712e+00 1.63618267e-01 -5.17329514e-01 7.03682959e-01
1.49552536e+00 -5.11046231e-01 6.09727085e-01 3.37367684e-01
9.42254007e-01 9.10018146e-01 4.21239913e-01 2.73018889e-03
5.82855165e-01 9.93597925e-01 5.87837696e-01 3.48763168e-02
-3.70379984e-01 -3.43529612e-01 -1.66341022e-01 5.53583443e-01
-7.82321095e-02 -6.29366711e-02 -7.51748979e-01 6.51391983e-01
-1.68201721e+00 -7.47736514e-01 -2.41104543e-01 2.03731751e+00
6.13294601e-01 -1.04426555e-01 -1.49775267e-01 -3.05302054e-01
8.82971942e-01 3.84946689e-02 -5.36230087e-01 1.78671896e-01
-1.86171368e-01 2.37964645e-01 4.60643530e-01 3.52129340e-01
-1.07986891e+00 9.68017995e-01 5.49544525e+00 6.02810264e-01
-8.81020665e-01 4.17778134e-01 3.49041909e-01 3.78178865e-01
-5.76596439e-01 1.05561726e-01 -7.63033211e-01 2.66026139e-01
3.92459750e-01 -2.75390863e-01 2.16639847e-01 7.62194812e-01
-2.77025580e-01 1.59010351e-01 -1.64487600e+00 1.12936759e+00
8.27159360e-02 -1.33714080e+00 2.26137415e-01 2.05884837e-02
6.86061382e-01 1.21512495e-01 -1.64995521e-01 -1.38786376e-01
3.28699142e-01 -1.13730776e+00 8.60936701e-01 5.22214174e-01
5.44991195e-01 -3.62053603e-01 6.19405985e-01 1.25355497e-01
-1.24629009e+00 4.06887561e-01 -5.44150472e-01 1.60108730e-01
2.38135740e-01 4.61319804e-01 -3.15342963e-01 7.03674316e-01
7.91442215e-01 9.15195107e-01 -5.49880505e-01 1.06049955e+00
-4.32283938e-01 -2.06805244e-01 -4.49468076e-01 2.65025288e-01
4.56257649e-02 -3.19459707e-01 5.07203341e-01 6.64400458e-01
3.85764241e-01 3.90279219e-02 9.29536894e-02 1.57952106e+00
-2.64950186e-01 -1.85328513e-01 -6.22145534e-01 2.38194436e-01
4.10354793e-01 1.15275109e+00 -7.95802534e-01 -1.11007482e-01
-6.32981956e-01 9.87986624e-01 3.93672556e-01 1.22401029e-01
-8.17922294e-01 -5.64093119e-04 1.21393657e+00 2.36298293e-01
1.69945553e-01 -4.40528989e-01 -5.99390626e-01 -1.25104952e+00
2.09558278e-01 -2.37830698e-01 1.12350486e-01 -9.95644808e-01
-1.65582323e+00 2.62477279e-01 5.02006523e-02 -1.13906527e+00
3.22609425e-01 -8.31557572e-01 -3.40304166e-01 9.24469829e-01
-1.59674644e+00 -1.32191992e+00 -7.10987628e-01 5.12125671e-01
1.70338050e-01 1.93332613e-01 8.23576391e-01 1.56901568e-01
2.55418658e-01 2.02380508e-01 -1.53910860e-01 2.49774724e-01
5.51457644e-01 -1.18226659e+00 5.90495765e-01 2.79162437e-01
5.09756744e-01 7.19657302e-01 5.81971705e-01 -3.85229349e-01
-1.35185277e+00 -1.06494963e+00 9.44992125e-01 -1.06422746e+00
6.98978782e-01 -4.38560903e-01 -1.13468719e+00 7.30478287e-01
-7.61161670e-02 3.54694694e-01 2.86717892e-01 1.16339393e-01
-7.68517792e-01 -4.29063067e-02 -1.16059995e+00 3.02232057e-01
1.61036468e+00 -7.91175842e-01 -9.28076446e-01 4.02928948e-01
9.51107144e-01 -3.84900033e-01 -1.16677105e+00 4.63878930e-01
4.50792164e-01 -9.92325902e-01 1.53328013e+00 -7.68166363e-01
2.48063847e-01 -3.99391145e-01 -5.19667387e-01 -1.32166445e+00
-1.93405569e-01 1.97686121e-01 2.63399124e-01 1.01443720e+00
7.64545500e-02 -5.59885442e-01 7.54969835e-01 6.89984739e-01
-2.17087239e-01 -2.36720383e-01 -9.03949440e-01 -1.12761950e+00
4.90389615e-01 -5.30897677e-01 9.20468450e-01 1.13231289e+00
-3.71721715e-01 1.19452201e-01 1.89933196e-01 4.39704210e-01
9.28734481e-01 5.88227093e-01 8.04114878e-01 -1.70124435e+00
4.75199372e-02 -6.62440479e-01 -9.99651492e-01 -1.11247957e+00
5.44112027e-01 -1.28701830e+00 -1.49737075e-01 -1.60044825e+00
3.94735575e-01 -8.58233273e-01 -3.58244628e-01 3.51440519e-01
-2.32070358e-03 4.68447268e-01 2.55297869e-01 1.46090910e-01
-5.84618509e-01 8.09965253e-01 1.41803598e+00 -5.21100819e-01
3.01078290e-01 -2.88503408e-01 -6.95559680e-01 6.64876521e-01
7.19857931e-01 -5.41958988e-01 -1.09845750e-01 -7.47337937e-01
3.25022280e-01 -4.83671844e-01 1.08326399e+00 -7.97408819e-01
1.28534213e-01 -2.34520212e-01 1.19003884e-01 -4.27565217e-01
2.78225958e-01 -1.14968491e+00 2.49765664e-01 6.94675446e-02
-2.69800544e-01 -3.56977195e-01 -6.66468590e-02 5.74230909e-01
-2.40240917e-01 -1.44097805e-01 8.43093216e-01 -3.07917386e-01
-1.14661312e+00 6.87291563e-01 5.57572424e-01 3.07620436e-01
1.11589229e+00 -2.06946984e-01 -3.87694776e-01 -4.49576117e-02
-8.10914099e-01 8.54835510e-02 1.00932503e+00 8.83821428e-01
6.20964527e-01 -1.54776239e+00 -6.46663725e-01 1.39538527e-01
6.43894672e-01 1.05581768e-01 2.42293552e-01 5.73985517e-01
-4.23957139e-01 4.04518038e-01 -3.18600327e-01 -9.31393325e-01
-1.00543344e+00 5.50483644e-01 4.51224536e-01 3.71264517e-01
-8.97649467e-01 7.80576408e-01 6.35071576e-01 -8.79091561e-01
-1.93192940e-02 -4.36228693e-01 2.29372889e-01 -2.32920825e-01
2.16662183e-01 1.63682103e-01 4.54962440e-02 -9.45211709e-01
-6.09833837e-01 1.25023723e+00 2.81319797e-01 1.42189294e-01
1.29322767e+00 -2.65553057e-01 -3.23427856e-01 4.53109711e-01
1.49090540e+00 -3.53667706e-01 -1.16846907e+00 -6.10239446e-01
9.90190804e-02 -8.32055330e-01 -1.59995183e-01 -5.33632636e-01
-1.16180325e+00 9.87144530e-01 6.14416242e-01 1.13447696e-01
6.70134604e-01 7.81486452e-01 7.80152857e-01 3.57170820e-01
8.56899142e-01 -7.02120185e-01 7.75064379e-02 4.19171751e-01
9.48663175e-01 -1.60302830e+00 -6.37020916e-02 -7.62501001e-01
-3.76215249e-01 9.41400826e-01 4.86251086e-01 -3.18109989e-01
7.49285698e-01 -3.67026150e-01 -1.66006774e-01 -6.23373806e-01
-2.68123507e-01 -4.41775531e-01 5.40930986e-01 9.31279778e-01
1.64073959e-01 2.04181552e-01 1.05059512e-01 3.23857576e-01
-3.11375111e-01 -2.62991726e-01 1.34878948e-01 5.19406199e-01
-5.33763111e-01 -1.13599157e+00 -1.61654904e-01 1.03068717e-01
-6.67066276e-02 1.34119406e-01 -3.71157259e-01 9.15721595e-01
1.85437232e-01 6.79288149e-01 4.25334811e-01 -1.37998566e-01
4.01018739e-01 -1.38297215e-01 1.03993309e+00 -7.42570639e-01
-6.41488507e-02 -5.57230413e-01 -2.12902829e-01 -7.84466267e-01
-7.37348974e-01 -4.89148974e-01 -1.30712163e+00 -4.77621138e-01
2.69455742e-03 -1.07229084e-01 8.44324470e-01 1.06987464e+00
2.92865127e-01 2.46407062e-01 3.01781029e-01 -7.66442716e-01
-4.23227191e-01 -5.96367538e-01 -6.75540149e-01 1.16365421e+00
7.28611499e-02 -1.05988777e+00 -3.03523272e-01 4.70349826e-02] | [8.014567375183105, -3.23715877532959] |
e3d3c10d-4eea-4ad4-8fa0-c0ea51c909e0 | learning-visual-question-answering-by | 1808.00300 | null | http://arxiv.org/abs/1808.00300v1 | http://arxiv.org/pdf/1808.00300v1.pdf | Learning Visual Question Answering by Bootstrapping Hard Attention | Attention mechanisms in biological perception are thought to select subsets
of perceptual information for more sophisticated processing which would be
prohibitive to perform on all sensory inputs. In computer vision, however,
there has been relatively little exploration of hard attention, where some
information is selectively ignored, in spite of the success of soft attention,
where information is re-weighted and aggregated, but never filtered out. Here,
we introduce a new approach for hard attention and find it achieves very
competitive performance on a recently-released visual question answering
datasets, equalling and in some cases surpassing similar soft attention
architectures while entirely ignoring some features. Even though the hard
attention mechanism is thought to be non-differentiable, we found that the
feature magnitudes correlate with semantic relevance, and provide a useful
signal for our mechanism's attentional selection criterion. Because hard
attention selects important features of the input information, it can also be
more efficient than analogous soft attention mechanisms. This is especially
important for recent approaches that use non-local pairwise operations, whereby
computational and memory costs are quadratic in the size of the set of
features. | ['Adam Santoro', 'Mateusz Malinowski', 'Carl Doersch', 'Peter Battaglia'] | 2018-08-01 | learning-visual-question-answering-by-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Mateusz_Malinowski_Learning_Visual_Question_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Mateusz_Malinowski_Learning_Visual_Question_ECCV_2018_paper.pdf | eccv-2018-9 | ['hard-attention'] | ['methodology'] | [ 5.32315135e-01 1.83851346e-01 2.82260656e-01 -4.55736369e-01
-5.73748648e-01 -7.02997565e-01 5.63629091e-01 5.92630029e-01
-8.40289176e-01 6.26866400e-01 3.53569537e-01 -2.77269602e-01
-3.49852741e-01 -5.67848742e-01 -5.11030853e-01 -6.13219917e-01
5.57491602e-03 3.94598246e-01 5.74127197e-01 -3.51323813e-01
6.61515415e-01 3.97316158e-01 -1.89506280e+00 3.61823469e-01
5.35605013e-01 1.00535643e+00 4.72495645e-01 8.12944710e-01
-8.98025632e-02 5.17707646e-01 -4.66292322e-01 -3.54490966e-01
2.04161197e-01 -5.38098454e-01 -1.19190669e+00 -1.41867578e-01
8.07237387e-01 2.22415373e-01 -3.20828408e-02 1.06863093e+00
5.08320153e-01 4.26535100e-01 6.48213208e-01 -7.76270390e-01
-1.09539366e+00 9.78104249e-02 -3.59865189e-01 8.01872373e-01
2.82595962e-01 2.60064125e-01 1.67393172e+00 -9.62971568e-01
5.48358977e-01 1.40972185e+00 1.26078904e-01 3.28619421e-01
-1.36252224e+00 -6.97693303e-02 4.03977841e-01 5.47333956e-01
-9.25735235e-01 -4.32495743e-01 4.72633213e-01 -2.83639938e-01
1.23598766e+00 7.12020755e-01 5.96953332e-01 5.45272052e-01
2.72063404e-01 6.13852322e-01 1.19316387e+00 -4.23454821e-01
3.54278237e-01 -1.89278889e-02 3.33948553e-01 4.75167006e-01
1.78738683e-01 -1.45243466e-01 -5.73433995e-01 -1.19271338e-01
4.71250147e-01 -7.77783804e-03 -3.56934577e-01 -1.45491764e-01
-1.31447208e+00 7.13171661e-01 8.95005286e-01 2.74901658e-01
-4.29262102e-01 6.78070411e-02 1.76668137e-01 6.32135868e-01
1.82816952e-01 9.55170929e-01 -5.02077341e-01 -2.48374622e-02
-5.65343797e-01 1.08807255e-02 4.29691106e-01 6.33176267e-01
8.43106747e-01 -3.14797372e-01 -2.35612944e-01 7.66381860e-01
6.94535151e-02 6.69291094e-02 3.32031697e-01 -1.00182366e+00
1.90080643e-01 6.73000634e-01 2.46775504e-02 -1.01710665e+00
-4.97040629e-01 -2.78088957e-01 -6.28764033e-01 5.57264447e-01
7.11748123e-01 3.06287974e-01 -1.10700929e+00 1.77125311e+00
3.46583617e-03 -6.70330942e-01 -3.38958323e-01 1.22742665e+00
5.98424971e-01 3.44817966e-01 2.66573459e-01 -7.09656328e-02
1.62071598e+00 -7.27753937e-01 -3.73339832e-01 -5.96030414e-01
7.70540088e-02 -8.59180033e-01 1.39763558e+00 3.73623431e-01
-1.25320971e+00 -6.97253048e-01 -8.92971694e-01 -6.14265144e-01
-5.75527608e-01 -5.05721509e-01 8.59604180e-01 2.53770500e-01
-1.18501365e+00 4.87086535e-01 -3.69187385e-01 -6.95453227e-01
6.18230760e-01 5.45717895e-01 -4.13640201e-01 -9.53745693e-02
-1.01886797e+00 1.16188920e+00 2.08788171e-01 4.64002900e-02
-5.68511546e-01 -3.83497745e-01 -7.13754237e-01 3.28615606e-01
3.34446162e-01 -7.43207872e-01 9.98930216e-01 -1.32040870e+00
-1.04659081e+00 9.66845036e-01 -3.09724748e-01 -2.29574561e-01
5.79790212e-02 -5.32752685e-02 1.32513657e-01 3.64405453e-01
-7.08904266e-02 1.05173314e+00 9.79337752e-01 -9.26992357e-01
-4.04392242e-01 -5.01615167e-01 3.14604521e-01 4.18169856e-01
-2.28770286e-01 2.02299893e-01 -2.57374495e-01 -6.17388248e-01
1.71617642e-01 -6.69424474e-01 -4.28784400e-01 4.27859098e-01
-7.61952400e-02 -4.54074085e-01 4.82457966e-01 -3.68340641e-01
7.02903032e-01 -2.23269343e+00 3.88662666e-01 1.36775628e-01
3.81399155e-01 9.10104811e-02 -3.20081294e-01 3.66166174e-01
-8.06174949e-02 1.56386867e-01 -5.24810493e-01 -7.58117288e-02
-8.52470547e-02 2.15443820e-01 -3.45927864e-01 3.44009221e-01
6.83010817e-01 1.04660821e+00 -9.64696467e-01 -4.11135226e-01
4.33014594e-02 2.29526341e-01 -7.62597203e-01 -8.15396160e-02
-1.06252000e-01 6.61738440e-02 -2.28901982e-01 7.48775959e-01
2.57037699e-01 -4.30056751e-01 9.21182800e-03 -1.35521606e-01
-7.40682706e-02 4.46271122e-01 -9.18459177e-01 1.34677756e+00
-3.71107575e-03 9.59879279e-01 2.15522587e-01 -1.01433730e+00
5.58893561e-01 -1.46804914e-01 2.27999330e-01 -9.46301520e-01
1.52653530e-01 -6.65659364e-03 6.10195875e-01 -4.84952778e-01
5.84243417e-01 -1.95364207e-01 1.08842008e-01 4.26270038e-01
1.57916263e-01 -2.35931844e-01 2.05572441e-01 2.17198342e-01
1.12557185e+00 2.65567303e-02 4.89614338e-01 -5.64324796e-01
2.59445548e-01 1.61283433e-01 4.38452035e-01 9.81670856e-01
-4.71153498e-01 8.64527881e-01 4.36770111e-01 -2.83356130e-01
-9.61257696e-01 -1.04644191e+00 -2.41754457e-01 1.73153353e+00
3.05781573e-01 -3.41267824e-01 -5.26021838e-01 -5.54109752e-01
3.98491742e-03 4.12857890e-01 -8.89279604e-01 -2.57105708e-01
-3.50204021e-01 -6.78560972e-01 8.66778791e-02 5.96389651e-01
4.04653810e-02 -1.57927907e+00 -1.02989328e+00 1.15385421e-01
1.36813030e-01 -5.46032369e-01 -4.66893733e-01 6.24164164e-01
-6.18651628e-01 -1.07901168e+00 -6.66405737e-01 -7.09610283e-01
8.32790077e-01 4.88491446e-01 1.22098315e+00 4.14842039e-01
-6.02848768e-01 3.81886870e-01 -3.06694269e-01 -7.42558956e-01
2.96984464e-01 -6.97924420e-02 -2.08005413e-01 -3.86782885e-02
5.30837357e-01 -3.08576316e-01 -8.27562571e-01 2.74393141e-01
-8.34774673e-01 -3.32471073e-01 8.10008645e-01 9.70852435e-01
5.96279860e-01 -4.09218043e-01 6.23936355e-01 -5.79624951e-01
6.78799808e-01 -5.19949377e-01 -1.93898410e-01 7.68475384e-02
-2.35467434e-01 2.16093287e-01 4.91984248e-01 -3.03242356e-01
-7.09478319e-01 -1.46234304e-01 -5.38498834e-02 -5.06033897e-02
-1.42187461e-01 3.54027957e-01 6.01320527e-02 -3.34648132e-01
5.95604777e-01 -6.52043447e-02 -6.95871711e-02 -2.66665846e-01
1.06170386e-01 3.70519876e-01 1.83104381e-01 -4.00135726e-01
3.80234092e-01 5.66684484e-01 3.12354136e-02 -1.03799510e+00
-9.44598794e-01 -4.87129778e-01 -6.50844276e-01 1.51587173e-01
1.07012451e+00 -3.44147325e-01 -9.55759287e-01 9.01068747e-02
-8.55475664e-01 -7.45248562e-03 -4.97801602e-01 5.45298100e-01
-5.56441188e-01 2.93930084e-01 -4.47338462e-01 -7.26987064e-01
2.18687225e-02 -1.15353847e+00 9.44273889e-01 1.31556854e-01
-5.84801555e-01 -7.10433960e-01 -2.36886382e-01 2.66587615e-01
5.53651094e-01 -2.67948776e-01 1.04179966e+00 -7.08633125e-01
-6.86103106e-01 -1.53443986e-03 -6.58266962e-01 1.26031712e-01
1.56267937e-02 2.35156226e-03 -1.20525360e+00 -2.16299161e-01
-1.30389437e-01 -5.45704007e-01 1.49819160e+00 3.50531965e-01
1.22636330e+00 1.07820369e-01 -1.38604358e-01 3.20016593e-01
1.07472765e+00 1.09822623e-01 5.75229168e-01 1.53049260e-01
4.48420405e-01 8.96877527e-01 3.86966139e-01 8.27103760e-03
4.02222276e-01 4.99018520e-01 7.74091303e-01 -3.46028388e-01
-1.18107572e-01 2.85138160e-01 -7.52376067e-03 6.23976409e-01
-3.04325670e-01 -8.22830871e-02 -7.93098986e-01 7.19760537e-01
-1.57776272e+00 -1.25450540e+00 -1.01046614e-01 2.31555629e+00
8.05895507e-01 4.68596190e-01 -8.52583796e-02 9.53283235e-02
3.57546836e-01 1.61883533e-01 -6.85894728e-01 -8.16865861e-01
-1.88831970e-01 5.51834822e-01 2.55011410e-01 5.81356347e-01
-1.04613793e+00 6.85293198e-01 7.50848055e+00 3.21004152e-01
-7.59450257e-01 -9.36041400e-02 5.86363971e-01 -3.17406118e-01
-4.13080007e-01 1.63404465e-01 -4.97435451e-01 3.68345052e-01
5.07259250e-01 1.47321641e-01 4.28135365e-01 4.51778054e-01
-1.72001585e-01 -6.23512268e-01 -1.22536469e+00 7.89010525e-01
1.69287369e-01 -9.72202182e-01 3.51713076e-02 1.13624856e-01
2.97898799e-01 3.79594155e-02 -4.17744890e-02 3.01098436e-01
1.97793648e-01 -1.28514576e+00 6.08683527e-01 6.34807467e-01
2.97525376e-01 -5.79704702e-01 5.07741749e-01 2.82442123e-01
-9.36561704e-01 -3.72490197e-01 -8.11096489e-01 -5.89409411e-01
-1.06329836e-01 4.90535796e-01 -4.21673179e-01 -2.41030976e-02
9.72013831e-01 2.64228791e-01 -8.53368700e-01 1.28923655e+00
-4.72183228e-02 4.12553579e-01 -3.40167731e-01 -4.00760978e-01
4.47341293e-01 1.86596334e-01 5.84120810e-01 1.10754001e+00
6.06032647e-02 2.57831782e-01 1.10320181e-01 7.24491894e-01
1.97894331e-02 1.66664720e-01 -5.25908351e-01 -2.50102533e-03
1.72332540e-01 1.26517463e+00 -8.86359930e-01 -3.11211050e-01
-4.46517825e-01 8.93382311e-01 6.12399995e-01 3.89208287e-01
-4.74796623e-01 -6.47807598e-01 8.50016534e-01 -3.86330448e-02
4.63968068e-01 -3.70449871e-01 -3.43045115e-01 -9.33210850e-01
-3.14975120e-02 -5.94389021e-01 5.82460999e-01 -8.79077494e-01
-1.40769029e+00 5.90164900e-01 -3.09191376e-01 -7.95754373e-01
2.36161072e-02 -7.95089006e-01 -5.18486917e-01 1.14888573e+00
-1.40345907e+00 -7.75389254e-01 -1.28356144e-01 5.99330962e-01
7.91107595e-01 3.11944127e-01 7.39083350e-01 -1.30407792e-02
-1.14980526e-01 3.51152271e-01 -2.22161427e-01 -2.04529420e-01
8.75995338e-01 -1.56637204e+00 2.33873665e-01 6.47360861e-01
3.26467872e-01 9.32837903e-01 7.41839528e-01 -3.08174461e-01
-1.30473042e+00 -4.81460631e-01 8.68788242e-01 -7.25287974e-01
5.37056565e-01 -3.74275535e-01 -1.18662846e+00 4.59358782e-01
6.05687737e-01 1.01733863e-01 7.38263249e-01 3.80904406e-01
-5.22846878e-01 1.27655156e-02 -1.01312101e+00 6.42517209e-01
1.10186851e+00 -5.71515918e-01 -1.07055688e+00 7.89087713e-02
4.62346613e-01 8.45449045e-02 -5.28699517e-01 1.82872072e-01
5.31569302e-01 -9.76309299e-01 9.24691677e-01 -8.11003506e-01
1.56788751e-01 -4.82185096e-01 -2.92065144e-01 -1.26837325e+00
-8.73803794e-01 -2.49090120e-01 2.16535151e-01 8.23439717e-01
5.02487600e-01 -7.26357937e-01 3.00463229e-01 5.44082820e-01
-1.75396800e-01 -7.55655348e-01 -9.04377282e-01 -3.99313748e-01
2.39270218e-02 -2.26920068e-01 1.32083297e-01 8.63980949e-01
-7.41947559e-04 5.46842694e-01 8.93542469e-02 -1.04399629e-01
6.16712689e-01 3.70310456e-01 1.81045339e-01 -1.38366103e+00
-3.60759169e-01 -8.74851108e-01 -5.15847862e-01 -8.46229672e-01
-2.27798253e-01 -9.23925281e-01 2.19714299e-01 -1.62539804e+00
3.11160296e-01 -2.08154738e-01 -8.18684101e-01 7.01168001e-01
-5.45340598e-01 6.57819211e-01 2.92076707e-01 1.37767047e-01
-5.98091543e-01 3.12846333e-01 1.50294137e+00 -2.62632519e-01
-5.44564798e-02 -2.63100863e-01 -1.19149947e+00 6.97122037e-01
6.86422408e-01 -2.10702226e-01 -3.95366818e-01 -5.04960299e-01
3.97761524e-01 -6.19959116e-01 6.02190435e-01 -9.66796875e-01
1.93620548e-01 -4.28982854e-01 7.52853394e-01 -3.02907497e-01
5.40189803e-01 -6.52516246e-01 -4.91146922e-01 2.10180536e-01
-5.74854374e-01 9.10685360e-02 2.23451123e-01 4.16729122e-01
-2.22541347e-01 -1.87671125e-01 8.04310143e-01 -4.25205112e-01
-1.05513752e+00 1.97877228e-01 -4.73175883e-01 1.65516019e-01
8.92672956e-01 -4.53328907e-01 -4.74761546e-01 -1.34009480e-01
-8.49696636e-01 1.62158951e-01 3.71712744e-01 4.95435864e-01
6.70235693e-01 -1.01345515e+00 -5.27403116e-01 2.38663003e-01
2.47328639e-01 -2.52807677e-01 4.43608277e-02 9.18369234e-01
-8.11209530e-02 4.35553014e-01 -6.07872725e-01 -5.10721266e-01
-1.19771087e+00 8.76993418e-01 2.26540435e-02 2.90484756e-01
-3.94112259e-01 1.16851425e+00 5.23707628e-01 1.96309090e-02
3.80558483e-02 -4.09142941e-01 -3.22210461e-01 4.09491867e-01
6.60514832e-01 8.05722997e-02 6.14115223e-02 -4.60892141e-01
-6.48777008e-01 7.86765635e-01 -1.19153753e-01 8.20154995e-02
1.31159043e+00 -5.27372174e-02 -2.92983174e-01 4.94455665e-01
7.73821354e-01 2.11004987e-02 -1.22607136e+00 -3.50143611e-02
3.81020480e-03 -5.13879061e-01 -4.64474224e-02 -7.45618880e-01
-6.87673450e-01 1.33642745e+00 2.90998310e-01 6.01078570e-01
1.18919683e+00 3.38590473e-01 3.63745153e-01 6.90609396e-01
2.35203817e-01 -1.03389204e+00 1.02393731e-01 5.93282938e-01
1.11932182e+00 -1.40421534e+00 1.34914955e-02 -1.24898158e-01
-5.69495738e-01 1.01390779e+00 7.76276350e-01 -1.75106823e-01
5.01651704e-01 1.80160452e-03 -1.79949895e-01 -3.24101686e-01
-1.11123705e+00 -7.70855963e-01 4.90489811e-01 5.74209213e-01
6.07620955e-01 -1.55345708e-01 -2.31311142e-01 9.95450988e-02
-1.75629016e-02 -3.92872483e-01 2.36633912e-01 9.51860368e-01
-9.63671267e-01 -7.56882846e-01 -3.43225896e-01 7.17981696e-01
-3.86582226e-01 -3.71392459e-01 -7.04151869e-01 6.28565192e-01
1.97242990e-01 8.18823099e-01 2.97815710e-01 7.12158084e-02
2.08898231e-01 7.42399022e-02 6.50724411e-01 -7.38670826e-01
-7.89363980e-01 -2.54218340e-01 -1.02477603e-01 -5.98715186e-01
-3.60550761e-01 -5.39351583e-01 -1.21211863e+00 1.43865094e-01
-8.24745670e-02 2.51273662e-02 3.55342150e-01 8.91521156e-01
3.68741035e-01 4.91700619e-01 1.78275377e-01 -1.01906800e+00
-4.75140333e-01 -8.11759830e-01 -1.92674577e-01 5.98873973e-01
6.47031903e-01 -8.09859812e-01 -3.36092234e-01 9.22918227e-03] | [10.0269775390625, 1.8408575057983398] |
93da8949-066f-492f-a75c-ee3ee96cd510 | imitation-learning-via-differentiable-physics | 2206.04873 | null | https://arxiv.org/abs/2206.04873v1 | https://arxiv.org/pdf/2206.04873v1.pdf | Imitation Learning via Differentiable Physics | Existing imitation learning (IL) methods such as inverse reinforcement learning (IRL) usually have a double-loop training process, alternating between learning a reward function and a policy and tend to suffer long training time and high variance. In this work, we identify the benefits of differentiable physics simulators and propose a new IL method, i.e., Imitation Learning via Differentiable Physics (ILD), which gets rid of the double-loop design and achieves significant improvements in final performance, convergence speed, and stability. The proposed ILD incorporates the differentiable physics simulator as a physics prior into its computational graph for policy learning. It unrolls the dynamics by sampling actions from a parameterized policy, simply minimizing the distance between the expert trajectory and the agent trajectory, and back-propagating the gradient into the policy via temporal physics operators. With the physics prior, ILD policies can not only be transferable to unseen environment specifications but also yield higher final performance on a variety of tasks. In addition, ILD naturally forms a single-loop structure, which significantly improves the stability and training speed. To simplify the complex optimization landscape induced by temporal physics operations, ILD dynamically selects the learning objectives for each state during optimization. In our experiments, we show that ILD outperforms state-of-the-art methods in a variety of continuous control tasks with Brax, requiring only one expert demonstration. In addition, ILD can be applied to challenging deformable object manipulation tasks and can be generalized to unseen configurations. | ['Zhongwen Xu', 'Xiao Ma', 'Siwei Chen'] | 2022-06-10 | null | null | null | null | ['deformable-object-manipulation'] | ['robots'] | [-1.03131412e-02 -8.95083044e-03 -2.47122526e-01 1.69661418e-01
-3.29171240e-01 -6.29280686e-01 5.05339682e-01 -5.76280579e-02
-5.93717754e-01 8.40260088e-01 -4.17857438e-01 -3.59391868e-01
-4.37235177e-01 -6.13103330e-01 -1.14929605e+00 -7.83715069e-01
-2.28794098e-01 6.05865359e-01 4.46626097e-01 -3.41417938e-01
9.26250815e-02 6.21851504e-01 -1.10741377e+00 -5.62586904e-01
1.29485488e+00 7.42605448e-01 6.24442101e-01 4.37474132e-01
1.29233584e-01 6.57386124e-01 -4.10157084e-01 1.19495124e-01
1.65117294e-01 -2.52599061e-01 -3.77803802e-01 -1.50862513e-02
1.19463287e-01 -4.79325175e-01 -6.22757852e-01 8.92782390e-01
5.66428185e-01 6.33503675e-01 5.08481085e-01 -1.22147751e+00
-6.39886439e-01 2.84419149e-01 -4.59860891e-01 -2.18945295e-01
1.62991777e-01 7.53976524e-01 6.18980527e-01 -6.59669995e-01
4.66663867e-01 1.47268867e+00 5.92229664e-01 6.91076875e-01
-1.36703110e+00 -7.56331742e-01 4.90426183e-01 1.58200674e-02
-1.09165299e+00 1.31998658e-01 7.34267235e-01 -4.64827001e-01
7.42949486e-01 -1.18257068e-01 8.79370332e-01 1.21600306e+00
4.14255768e-01 8.86630714e-01 1.07612514e+00 -2.99389008e-02
3.96182746e-01 -1.63314074e-01 -2.27652580e-01 1.15368700e+00
1.30013138e-01 6.52166188e-01 -1.26807451e-01 -2.01885045e-01
1.22423327e+00 5.14021516e-02 -4.70086396e-01 -7.37188935e-01
-1.19516623e+00 7.04401791e-01 5.93319654e-01 -3.52655113e-01
-2.84407705e-01 5.31755805e-01 3.13370377e-01 1.92247495e-01
1.06887687e-02 6.81800008e-01 -4.45714682e-01 -8.46151784e-02
-3.17641735e-01 5.84436655e-01 8.40358973e-01 9.85096276e-01
6.39526308e-01 2.71205068e-01 -2.80060381e-01 6.53793454e-01
3.17512661e-01 7.69455194e-01 2.51621217e-01 -1.11533868e+00
3.20742488e-01 5.09978473e-01 4.07667637e-01 -5.44177592e-01
-3.36128324e-01 -4.27103877e-01 -6.34453237e-01 7.19016492e-01
3.39351296e-01 -3.45649987e-01 -1.05673420e+00 1.87289345e+00
6.30457580e-01 4.53546584e-01 -5.59349433e-02 1.03285360e+00
2.90042222e-01 6.97731078e-01 1.06173180e-01 -9.25312862e-02
9.25231934e-01 -1.07491720e+00 -5.76400697e-01 -1.31684139e-01
3.35266262e-01 -3.54316413e-01 1.49762797e+00 2.92487562e-01
-1.01180053e+00 -5.29651284e-01 -1.08354461e+00 1.06680371e-01
6.08220622e-02 3.13781679e-01 4.29486573e-01 -6.57095611e-02
-5.76825976e-01 1.24443555e+00 -1.28897583e+00 3.45709696e-02
1.26384407e-01 6.08186126e-01 1.01817869e-01 2.95965284e-01
-1.07875514e+00 9.69564080e-01 4.24398333e-01 -1.24666849e-02
-1.32551634e+00 -9.98833120e-01 -7.59851396e-01 -1.45416215e-01
9.91228878e-01 -7.94412553e-01 1.30168927e+00 -5.81823707e-01
-2.47961736e+00 1.12821504e-01 2.57940948e-01 -1.60278082e-01
7.70610988e-01 -4.97701854e-01 -4.44499366e-02 -7.10779428e-02
-9.41781774e-02 5.11300087e-01 1.20226467e+00 -1.24395144e+00
-3.77117634e-01 -5.38408421e-02 2.57968992e-01 3.12831521e-01
-1.80619702e-01 -7.17891634e-01 -6.26561344e-01 -7.92521477e-01
-3.21308225e-01 -1.46513879e+00 -3.63906950e-01 3.14125359e-01
-4.02283011e-04 -3.75785679e-01 1.14160573e+00 -2.93169111e-01
9.98472035e-01 -2.06620550e+00 7.69885302e-01 6.60709944e-03
4.79548424e-02 4.67134327e-01 -2.18880013e-01 3.20813924e-01
3.07629883e-01 -3.20715100e-01 -2.75731951e-01 5.33499848e-03
2.37002805e-01 5.29856026e-01 -5.81603467e-01 4.21768099e-01
2.19506502e-01 1.16868246e+00 -1.35122681e+00 -3.71213436e-01
1.78297326e-01 3.48425508e-01 -6.64348781e-01 4.52290565e-01
-7.91136563e-01 9.65178609e-01 -9.47691739e-01 1.30010247e-01
3.66742939e-01 -3.34219187e-01 7.54278973e-02 2.23652069e-02
-2.51891881e-01 -1.30371545e-02 -1.06694746e+00 1.71720266e+00
-6.56721175e-01 1.64999723e-01 3.32932025e-01 -8.29810083e-01
9.66116726e-01 -2.87906546e-02 4.34785903e-01 -4.47791964e-01
1.27304047e-01 2.13337481e-01 4.71742917e-03 -5.44751525e-01
2.16636509e-01 9.78145450e-02 6.04962669e-02 2.05770195e-01
4.34692949e-02 -6.07127666e-01 5.76931015e-02 -1.09595478e-01
9.42305207e-01 8.05436909e-01 -8.71774703e-02 -2.61277288e-01
6.73562944e-01 1.11044385e-01 6.90638363e-01 7.08403409e-01
6.11751899e-02 -5.31353615e-02 3.70213956e-01 -1.63215533e-01
-8.83937538e-01 -1.36209178e+00 1.68697923e-01 9.40517962e-01
6.28775001e-01 -1.01550020e-01 -5.50500691e-01 -6.41755402e-01
4.25719798e-01 7.60129690e-01 -2.48318538e-01 -4.97623086e-01
-1.02437234e+00 -1.67791933e-01 1.94408357e-01 4.82814372e-01
4.76589352e-01 -1.20530057e+00 -6.93856597e-01 4.64465678e-01
2.91478366e-01 -1.06654286e+00 -8.35975051e-01 -9.27514806e-02
-9.05457377e-01 -9.66784000e-01 -6.81068242e-01 -7.26994514e-01
6.67077661e-01 3.23829651e-02 6.26041710e-01 7.39604887e-03
-2.40434512e-01 5.84119856e-01 -1.15476698e-01 -1.93211302e-01
-3.67376626e-01 -2.52438989e-02 2.87421316e-01 -2.61823803e-01
-4.79539603e-01 -6.59245670e-01 -6.81415856e-01 5.12930512e-01
-7.00011432e-01 2.98216604e-02 6.83630884e-01 1.14976728e+00
7.95072496e-01 -8.88899714e-02 2.96195149e-01 -3.59167188e-01
7.75072336e-01 -1.93808630e-01 -1.22070050e+00 1.93275020e-01
-5.78839839e-01 4.67646211e-01 1.11789024e+00 -1.09396958e+00
-1.05264628e+00 8.96757394e-02 2.66152769e-01 -9.36302781e-01
3.30550730e-01 3.28782380e-01 1.88923523e-01 -4.44873869e-01
4.65463161e-01 1.29143685e-01 3.91764879e-01 -4.25362498e-01
5.18111527e-01 7.69729167e-02 4.69814211e-01 -1.16112912e+00
9.68805254e-01 3.95567656e-01 2.70666033e-01 -6.59718752e-01
-7.49287426e-01 -7.94917718e-02 -2.71399170e-01 -3.74442101e-01
6.62300110e-01 -6.38762891e-01 -1.38008547e+00 4.57270592e-01
-9.29438949e-01 -9.77809608e-01 -3.71761233e-01 7.98038602e-01
-9.02938247e-01 4.44876641e-01 -7.01977491e-01 -7.39729106e-01
-1.97474837e-01 -1.38205016e+00 9.16208506e-01 3.41469169e-01
6.07629642e-02 -8.87990355e-01 6.08309656e-02 -2.61320025e-01
2.15725705e-01 3.26557070e-01 9.24096406e-01 1.20024294e-01
-7.93649077e-01 5.29799350e-02 9.42819938e-02 2.05308393e-01
8.09019431e-03 -1.11263633e-01 -3.45425934e-01 -7.97793090e-01
2.52029300e-02 -4.94955629e-01 6.26580417e-01 2.73123354e-01
1.30955696e+00 -2.56417304e-01 -4.69201267e-01 6.12987399e-01
1.17744303e+00 3.09290498e-01 2.35554591e-01 2.33493701e-01
8.42048824e-01 4.33671296e-01 8.37571502e-01 3.51368815e-01
1.29653215e-01 7.97016263e-01 3.68045598e-01 8.86523277e-02
3.40851247e-02 -5.56538939e-01 7.23103523e-01 7.36851573e-01
-5.28908074e-02 8.26554149e-02 -6.65685833e-01 1.73233747e-01
-2.22957587e+00 -6.66934550e-01 1.96375653e-01 2.17055988e+00
8.12027514e-01 1.45014092e-01 7.73044080e-02 -5.32808125e-01
5.90827346e-01 -5.02150320e-02 -1.37466633e+00 -6.97107241e-02
4.67676163e-01 2.70202994e-01 6.81337416e-01 5.00967920e-01
-8.62284601e-01 1.11894941e+00 5.76287937e+00 1.05662906e+00
-1.34051716e+00 -1.16058283e-01 -7.59053603e-02 -7.94283748e-02
-4.96865660e-02 7.10321814e-02 -6.54698133e-01 4.63087946e-01
4.12852734e-01 -1.90718397e-01 9.72204566e-01 1.05416441e+00
5.24421930e-01 2.86954958e-02 -1.08301842e+00 7.51483440e-01
-5.19177616e-01 -1.22223103e+00 -1.49829641e-01 3.41757201e-02
6.85003638e-01 1.55064864e-02 1.49518922e-01 6.27736092e-01
5.67972422e-01 -7.90653229e-01 7.40589559e-01 4.74586427e-01
6.77244186e-01 -6.14521325e-01 8.27710852e-02 6.44459426e-01
-1.18367970e+00 -3.16931486e-01 -2.33623818e-01 5.52427955e-02
3.36251438e-01 3.93984430e-02 -4.20336038e-01 4.67444301e-01
5.33181071e-01 7.24435985e-01 5.29753305e-02 9.14436519e-01
-5.63682973e-01 4.11444694e-01 -4.57231015e-01 -5.51488400e-01
4.00951713e-01 -7.94419765e-01 9.64452267e-01 6.74046576e-01
1.35088623e-01 6.29374012e-02 8.31674814e-01 1.19703710e+00
2.17425674e-01 -2.33802363e-01 -5.52603722e-01 -1.10695630e-01
2.92651683e-01 1.11013186e+00 -4.37725902e-01 -2.63090789e-01
-7.64010698e-02 8.62045348e-01 5.88781893e-01 6.00892782e-01
-1.27924490e+00 -4.50823188e-01 7.95892894e-01 -2.30114192e-01
4.61674809e-01 -8.14948022e-01 2.21540451e-01 -1.05344617e+00
1.56706542e-01 -7.65202045e-01 -1.60793722e-01 -4.82048124e-01
-1.21017671e+00 1.91923901e-01 1.49348989e-01 -1.42544901e+00
-2.04648539e-01 -8.06980848e-01 -5.71847916e-01 7.01979935e-01
-1.54056346e+00 -6.35862410e-01 -2.29458004e-01 5.17793953e-01
4.22893137e-01 -6.60889000e-02 5.31074286e-01 8.13659281e-02
-6.53904140e-01 3.65487844e-01 2.22934514e-01 -1.57664835e-01
5.21154165e-01 -1.28076780e+00 2.50173777e-01 3.06013882e-01
-4.16100204e-01 5.73064983e-01 6.01114631e-01 -7.02502191e-01
-1.93526149e+00 -1.09360528e+00 -3.47312093e-01 -6.93086982e-02
1.06223404e+00 -2.74381042e-01 -1.00920224e+00 4.14846718e-01
-2.11078897e-01 -7.15066642e-02 -2.35454947e-01 -3.87407213e-01
1.12347407e-02 -9.11969319e-02 -8.74848247e-01 1.07684183e+00
1.19206941e+00 -3.42492640e-01 -3.62644821e-01 3.63558471e-01
9.97046590e-01 -9.02224958e-01 -9.45493639e-01 5.36852479e-01
4.97079998e-01 -2.16561273e-01 1.02021194e+00 -6.57529354e-01
1.48443267e-01 -5.70856750e-01 5.07713497e-01 -1.50443757e+00
-2.71284968e-01 -1.19553792e+00 -5.28639376e-01 6.10613942e-01
1.51821658e-01 -8.74745011e-01 6.36478841e-01 3.75086218e-01
-2.27011606e-01 -1.11923862e+00 -7.01707602e-01 -1.34811008e+00
1.51250124e-01 -6.57277033e-02 4.14641052e-01 5.32365978e-01
-2.19573304e-01 4.09724355e-01 -4.03833479e-01 3.49138975e-01
8.22257400e-01 3.05916488e-01 9.97679234e-01 -9.16392326e-01
-8.33421171e-01 -4.80341822e-01 8.48450437e-02 -1.65684974e+00
4.88488317e-01 -7.81054676e-01 2.92475462e-01 -1.33633268e+00
-2.81526297e-01 -6.89976037e-01 -3.03148329e-02 3.56664360e-01
-2.95302927e-01 -6.17960155e-01 2.77992517e-01 2.70971209e-01
-2.56580442e-01 1.16148484e+00 2.08871388e+00 -1.34466350e-01
-6.10927224e-01 1.25544310e-01 2.26184875e-02 7.56604970e-01
7.95987904e-01 -4.12012190e-01 -7.49134958e-01 -3.93956989e-01
-2.85626322e-01 3.40494573e-01 5.03670335e-01 -8.63414466e-01
1.62459150e-01 -5.60142457e-01 6.73439726e-02 -2.47048303e-01
5.03830314e-01 -7.36621022e-01 3.31754982e-02 9.26660717e-01
-1.33325219e-01 2.40019686e-03 4.30678874e-01 1.02627528e+00
1.53608888e-01 -1.42844096e-01 9.13683712e-01 -9.66465159e-05
-6.15112841e-01 6.83201671e-01 -7.92186037e-02 9.96992886e-02
1.18638062e+00 1.80513691e-02 -2.04115435e-01 -6.72317296e-02
-6.38648152e-01 8.26860547e-01 5.56161642e-01 4.79522496e-01
5.80112100e-01 -1.17359865e+00 -1.83590472e-01 2.45203450e-02
-3.30686837e-01 1.53687239e-01 1.79603584e-02 9.16747630e-01
-3.82238925e-01 6.01237975e-02 -1.40037209e-01 -7.34046996e-01
-7.34277785e-01 6.73378944e-01 3.54439795e-01 -3.40816319e-01
-1.12642074e+00 4.54015374e-01 3.09833050e-01 -6.74282610e-01
3.83751571e-01 -4.08597171e-01 9.18487236e-02 -5.53191423e-01
4.59958576e-02 4.97796953e-01 -5.18441379e-01 -5.28023615e-02
-9.11886692e-02 9.25204694e-01 4.74488847e-02 -1.92837909e-01
1.29006505e+00 3.70892137e-01 1.49204461e-02 3.35966408e-01
9.63450849e-01 -3.17842394e-01 -2.04777312e+00 -5.90707175e-02
-1.41088620e-01 -2.37494901e-01 2.82719526e-02 -5.37820876e-01
-9.43577111e-01 5.44910014e-01 4.49557751e-01 -7.53190294e-02
7.44905829e-01 -2.81572759e-01 9.80638862e-01 6.81391478e-01
5.88046014e-01 -1.24684560e+00 6.11962676e-01 8.15520108e-01
1.13264120e+00 -9.21135545e-01 1.43802553e-01 -3.41573030e-01
-7.42296815e-01 1.15164649e+00 8.60880077e-01 -6.80348098e-01
6.72807932e-01 2.60780543e-01 -2.45469734e-01 -5.79648726e-02
-6.10646129e-01 -1.18634999e-01 3.36684555e-01 4.23464566e-01
-1.33537531e-01 8.38172659e-02 -3.49901140e-01 5.21255545e-02
1.91386074e-01 1.19637839e-01 1.37375161e-01 1.11049080e+00
-4.69960660e-01 -1.24143636e+00 -1.39843762e-01 1.50697425e-01
1.48829952e-01 4.28412884e-01 -9.42762122e-02 1.09418190e+00
-1.65650353e-01 5.11384487e-01 -2.85344869e-01 -2.07723334e-01
6.98140025e-01 -3.44189584e-01 7.41983593e-01 -4.67313409e-01
-4.39376324e-01 9.34156626e-02 -3.42795759e-01 -8.73275220e-01
2.27546152e-02 -4.08518106e-01 -1.64760435e+00 -2.19599336e-01
-4.42259759e-01 1.06020510e-01 4.78823751e-01 8.50425541e-01
6.09044015e-01 8.14613700e-01 5.57679176e-01 -1.17197287e+00
-1.29605496e+00 -6.41272068e-01 -2.70810604e-01 3.02139074e-01
4.18215722e-01 -1.35628688e+00 -2.13003114e-01 -2.80314058e-01] | [4.419600009918213, 1.4445416927337646] |
3465bb42-027a-40a9-b5d1-dfc1af124477 | improving-sonar-image-patch-matching-via-deep | 1709.02150 | null | http://arxiv.org/abs/1709.02150v1 | http://arxiv.org/pdf/1709.02150v1.pdf | Improving Sonar Image Patch Matching via Deep Learning | Matching sonar images with high accuracy has been a problem for a long time,
as sonar images are inherently hard to model due to reflections, noise and
viewpoint dependence. Autonomous Underwater Vehicles require good sonar image
matching capabilities for tasks such as tracking, simultaneous localization and
mapping (SLAM) and some cases of object detection/recognition. We propose the
use of Convolutional Neural Networks (CNN) to learn a matching function that
can be trained from labeled sonar data, after pre-processing to generate
matching and non-matching pairs. In a dataset of 39K training pairs, we obtain
0.91 Area under the ROC Curve (AUC) for a CNN that outputs a binary
classification matching decision, and 0.89 AUC for another CNN that outputs a
matching score. In comparison, classical keypoint matching methods like SIFT,
SURF, ORB and AKAZE obtain AUC 0.61 to 0.68. Alternative learning methods
obtain similar results, with a Random Forest Classifier obtaining AUC 0.79, and
a Support Vector Machine resulting in AUC 0.66. | ['Matias Valdenegro-Toro'] | 2017-09-07 | null | null | null | null | ['patch-matching'] | ['computer-vision'] | [ 2.23437428e-01 1.50639471e-02 2.70814151e-01 -7.37094104e-01
-8.50378633e-01 -4.39511955e-01 6.62212610e-01 4.33832109e-01
-1.00551307e+00 4.62655634e-01 -2.78573334e-01 -9.50509906e-02
-3.33057910e-01 -1.04836595e+00 -9.19875741e-01 -4.71381068e-01
-4.77135032e-01 4.09521341e-01 2.73555636e-01 -2.97572643e-01
3.62382263e-01 6.99996233e-01 -1.89484811e+00 -1.84848666e-01
7.74313748e-01 1.34739876e+00 2.42294576e-02 9.79553342e-01
1.29632711e-01 2.48248741e-01 -5.05769193e-01 -4.73127127e-01
5.76405823e-01 -1.27474174e-01 -3.68753016e-01 -7.27122307e-01
1.16838968e+00 -3.45011503e-01 -4.28460032e-01 1.03410816e+00
4.46900427e-01 7.81353191e-02 6.83034658e-01 -1.37300169e+00
-2.43690148e-01 4.14655358e-01 -6.67983443e-02 -1.90316647e-01
4.16427255e-01 -2.67771244e-01 9.26589310e-01 -7.02141106e-01
5.02739251e-01 8.88032138e-01 1.30666816e+00 2.50070155e-01
-8.06808829e-01 -9.94411051e-01 -8.48667085e-01 3.86390500e-02
-1.26192343e+00 -3.39810669e-01 2.63163716e-01 -3.14896435e-01
9.11588013e-01 6.42169714e-02 9.62649941e-01 4.23443764e-01
8.67744744e-01 4.01413023e-01 9.65005994e-01 -3.85752708e-01
2.34387130e-01 -2.45324194e-01 -2.70740271e-01 8.51075947e-01
5.36390781e-01 5.15815616e-01 -5.79384685e-01 -1.19224869e-01
7.22679794e-01 2.14555427e-01 -1.29640624e-02 -4.96385396e-01
-1.13833570e+00 8.70192766e-01 9.67206717e-01 -1.18590124e-01
-3.77594560e-01 4.52285051e-01 1.84348002e-01 9.37358201e-01
-4.23912294e-02 9.61559772e-01 -3.55589747e-01 1.30190164e-01
-1.09877145e+00 4.82390314e-01 9.59831953e-01 9.91617978e-01
1.26922536e+00 1.94280464e-02 4.25218046e-01 6.75370276e-01
3.87113750e-01 1.03309870e+00 5.21057248e-01 -8.78249824e-01
-1.33785391e-02 4.25355375e-01 2.98713654e-01 -1.12292314e+00
-8.70210469e-01 -5.35382748e-01 -5.69809079e-01 4.65016782e-01
3.39905500e-01 -1.33895241e-02 -1.05068302e+00 1.29127872e+00
-1.24935275e-02 1.36560649e-01 7.27985024e-01 1.00436771e+00
1.09156191e+00 3.35672259e-01 -2.90731549e-01 3.97153020e-01
1.20041001e+00 -7.30610549e-01 -2.91233867e-01 -5.27461886e-01
5.86956620e-01 -8.73390973e-01 3.88623327e-01 3.08598638e-01
-6.71383321e-01 -5.20445645e-01 -1.67654788e+00 1.29360884e-01
-4.61682886e-01 4.01799902e-02 6.75462961e-01 4.80967104e-01
-1.20436537e+00 7.76771426e-01 -9.25453067e-01 -5.13704181e-01
1.51480824e-01 7.01937139e-01 -8.35072875e-01 -3.08289379e-01
-1.13671505e+00 1.18584943e+00 7.86324516e-02 2.06866890e-01
-8.85515869e-01 -4.30492342e-01 -1.30794573e+00 -1.15335807e-01
-4.59227502e-01 -4.37678427e-01 1.15753818e+00 -7.91013658e-01
-1.12833047e+00 8.11791599e-01 3.76594216e-01 -8.82734776e-01
7.75409937e-01 -3.15141201e-01 -3.18153381e-01 2.24105343e-02
3.47376913e-01 1.05859804e+00 6.79225504e-01 -1.00361156e+00
-1.00527489e+00 -3.68813187e-01 -6.40589967e-02 6.34040162e-02
1.70827985e-01 -2.04794124e-01 5.09774685e-02 -1.04949944e-01
9.14488018e-01 -1.09194279e+00 -3.98494244e-01 2.49289587e-01
1.47475347e-01 1.81413498e-02 5.68753779e-01 -3.22957337e-01
2.49642968e-01 -1.90618062e+00 -4.40405548e-01 2.46665314e-01
7.93161802e-03 5.85826449e-02 -2.96328723e-01 5.12234747e-01
2.61675537e-01 -3.83113742e-01 -1.38385043e-01 -1.16911918e-01
-7.14712888e-02 7.44632110e-02 6.87318444e-02 9.39365447e-01
1.21003114e-01 7.77530193e-01 -1.16491425e+00 -2.77469099e-01
3.35771233e-01 4.37376052e-01 -2.70999968e-01 3.79866391e-01
2.46491924e-01 -8.65605474e-02 -3.07420958e-02 8.79692197e-01
7.77996123e-01 1.89561129e-01 -1.51623636e-01 -3.39304149e-01
-3.02252501e-01 -6.00785054e-02 -1.12611735e+00 1.58050048e+00
-7.82503247e-01 1.35040843e+00 1.45590842e-01 -9.14768159e-01
1.79124200e+00 6.11817054e-02 3.12114716e-01 -9.03697252e-01
2.08409652e-01 7.80747890e-01 1.87522531e-01 -3.96791637e-01
8.46275926e-01 -1.63487643e-01 -2.57030457e-01 2.56908447e-01
2.58149058e-01 -4.71450746e-01 -2.40200967e-01 -1.59509391e-01
1.24956942e+00 -8.50455612e-02 1.29607469e-01 -3.49046648e-01
2.00851977e-01 4.09034520e-01 1.83600053e-01 1.15826428e+00
-6.95552677e-02 7.57557273e-01 -1.26827478e-01 -1.03550804e+00
-1.10622919e+00 -7.94263422e-01 -4.25274581e-01 5.46511233e-01
6.11980200e-01 1.02331161e-01 -2.95887142e-01 -1.11734591e-01
3.82712454e-01 8.95791203e-02 -5.11967421e-01 -2.89611399e-01
-4.82806623e-01 -2.17461795e-01 8.11463058e-01 4.37104285e-01
6.90439284e-01 -1.11218882e+00 -1.11316621e+00 3.78637463e-01
1.91280603e-01 -1.06470156e+00 2.10562184e-01 5.02453506e-01
-7.76378989e-01 -1.00176358e+00 -7.21756697e-01 -8.62348437e-01
6.42804623e-01 4.58011121e-01 9.35611129e-01 2.30459854e-01
-4.41920727e-01 2.03313515e-01 -5.51293790e-01 -5.33196926e-01
-3.06045562e-01 -2.97904350e-02 3.89964640e-01 -3.23490620e-01
3.15327674e-01 -2.38648683e-01 -7.82050490e-01 5.49784541e-01
-6.85711741e-01 -3.31992865e-01 9.08555925e-01 1.08839917e+00
2.99954832e-01 -4.77139801e-01 3.52458894e-01 -2.78002113e-01
1.66542992e-01 -2.99796343e-01 -1.00419807e+00 1.12867542e-01
-6.73294187e-01 -5.29034659e-02 1.96193770e-01 -2.63939291e-01
-2.48516575e-01 5.18752217e-01 -1.68946892e-01 -2.98562378e-01
-8.62536356e-02 6.35531068e-01 3.95106614e-01 -9.26049709e-01
9.17075157e-01 2.07121730e-01 3.95977020e-01 1.01606054e-02
-9.43861380e-02 8.78337860e-01 7.04182327e-01 -7.12475926e-02
8.95208657e-01 5.38168252e-01 2.10539937e-01 -9.91989553e-01
-6.06790423e-01 -6.75067246e-01 -6.78497195e-01 -3.16860497e-01
6.34487629e-01 -1.11457777e+00 -5.85910380e-01 5.92688024e-01
-9.67996418e-01 -2.40002304e-01 9.98204798e-02 1.05309761e+00
-5.00048220e-01 2.11416557e-02 -1.29396215e-01 -7.25133657e-01
-5.52674711e-01 -1.10579979e+00 1.18301380e+00 6.51965678e-01
-6.50503561e-02 -7.47414291e-01 1.87858164e-01 6.67735860e-02
6.45509183e-01 3.32749993e-01 -1.08063251e-01 -9.15520966e-01
-5.63982487e-01 -1.00966966e+00 -4.31721807e-01 1.14973128e-01
4.06134129e-02 -6.77447915e-02 -9.91865754e-01 -6.30152225e-01
-6.31041765e-01 -3.97998065e-01 1.10756814e+00 4.28964853e-01
2.09568530e-01 -1.66049059e-02 -3.47942144e-01 1.00176537e+00
1.45573854e+00 4.10154730e-01 5.93507171e-01 7.39613652e-01
2.06911594e-01 6.30467415e-01 1.00266457e+00 1.32778913e-01
3.80819857e-01 4.80228543e-01 8.09162080e-01 -3.03627253e-01
2.28295207e-01 -2.82986671e-01 9.74990353e-02 7.64847845e-02
1.95589587e-02 2.65053846e-02 -1.12356007e+00 5.61932564e-01
-1.59065270e+00 -6.37329161e-01 -3.02973032e-01 2.31424832e+00
4.56803143e-02 -1.72543954e-02 -4.17999625e-01 1.54841676e-01
5.14928699e-01 -1.20513648e-01 -2.95220315e-01 -2.38166660e-01
-6.34504557e-02 2.77420104e-01 1.15597665e+00 4.77286071e-01
-1.40954494e+00 9.57742631e-01 6.03438330e+00 1.27276570e-01
-1.41256547e+00 -5.47095299e-01 -9.38642547e-02 7.25831807e-01
2.13256236e-02 3.02059669e-02 -6.24302268e-01 1.81005653e-02
7.65447497e-01 1.95915014e-01 1.55079234e-02 9.48593259e-01
-2.31235862e-01 -2.33870521e-01 -8.65004838e-01 9.80221093e-01
2.20848694e-01 -1.36148274e+00 -2.44532272e-01 -5.77678010e-02
7.18533516e-01 5.26260793e-01 -3.42839569e-01 1.48785219e-01
4.01427060e-01 -1.09758782e+00 7.12007463e-01 5.15883148e-01
9.03942466e-01 -5.01603961e-01 1.44042969e+00 3.43490504e-02
-1.05446243e+00 7.25574121e-02 -7.98288584e-01 -3.31696212e-01
-1.79061353e-01 1.65220410e-01 -1.36945987e+00 3.05868357e-01
1.03363240e+00 5.51921546e-01 -3.14144641e-01 1.76392639e+00
-2.02275351e-01 1.86069414e-01 -5.36620378e-01 -4.34295595e-01
4.87295717e-01 -1.13422692e-01 3.58509630e-01 1.22532594e+00
8.19365859e-01 -5.83440475e-02 -9.57233086e-02 2.74340779e-01
1.56875059e-01 1.62161246e-01 -9.14926827e-01 3.73466104e-01
7.30387986e-01 1.30990720e+00 -8.32153559e-01 -9.56261083e-02
-2.55129606e-01 6.15926921e-01 -1.82168856e-01 -4.66040261e-02
-3.97768885e-01 -9.87436712e-01 6.73750222e-01 -7.44888037e-02
1.61926031e-01 -4.81042057e-01 -1.18659483e-02 -8.25001776e-01
-3.42077464e-01 -3.56898338e-01 2.70746559e-01 -8.72142553e-01
-1.00424111e+00 7.14428723e-01 -4.07334447e-01 -1.72286761e+00
-1.43586650e-01 -8.37302029e-01 -6.32231057e-01 6.40649199e-01
-1.84301531e+00 -1.18477845e+00 -1.14825618e+00 5.70199564e-02
2.02801034e-01 -3.13344896e-01 8.14808667e-01 2.04962537e-01
2.90350586e-01 6.31716013e-01 8.43238458e-02 5.79504669e-01
7.95373142e-01 -1.11398065e+00 4.40274209e-01 4.10523355e-01
1.89029321e-01 5.17336547e-01 9.30566728e-01 -4.43558216e-01
-1.81075823e+00 -8.55648100e-01 8.97333682e-01 -8.19829851e-02
6.35214448e-01 3.31276245e-02 -6.46700025e-01 4.17937875e-01
-3.34416509e-01 2.98041284e-01 5.51461756e-01 -2.13507310e-01
-3.95169467e-01 -5.03765225e-01 -1.29065347e+00 -3.31788212e-02
6.36032522e-01 -2.82316685e-01 -4.46852356e-01 1.53496251e-01
3.47938329e-01 -8.00803125e-01 -9.24256325e-01 7.18448400e-01
1.44033825e+00 -1.10567045e+00 9.89534974e-01 -2.71714181e-01
3.36672217e-01 -4.19491231e-01 -4.02650088e-01 -1.27354264e+00
1.21636733e-01 -2.64049917e-01 8.21162581e-01 4.79881585e-01
5.74611187e-01 -9.55959082e-01 7.81118691e-01 5.83482049e-02
-5.03550053e-01 -1.87535450e-01 -1.14985812e+00 -9.38252270e-01
-2.14837164e-01 -3.11070889e-01 6.50580645e-01 8.49540353e-01
-2.30045900e-01 -1.01187848e-01 -3.50394487e-01 5.23194134e-01
7.12840319e-01 2.89971441e-01 1.17667961e+00 -1.61338675e+00
3.38617563e-01 -1.97332293e-01 -1.15537167e+00 -8.41960669e-01
3.36898640e-02 -7.57315874e-01 5.72317600e-01 -1.67669797e+00
-2.90635496e-01 -1.03676665e+00 -2.03224704e-01 5.42244792e-01
4.29055512e-01 8.27979743e-01 -2.56615520e-01 4.92110670e-01
-4.16787773e-01 1.58097506e-01 9.80911911e-01 -2.40015835e-01
-1.81876495e-02 3.39946210e-01 -9.13206860e-02 5.19000888e-01
6.96594894e-01 -6.04659021e-01 2.41255522e-01 -4.71188009e-01
4.80286181e-01 4.94620711e-01 5.70925415e-01 -1.46571910e+00
5.63823283e-01 1.12011373e-01 4.79693562e-01 -4.55686063e-01
4.14015710e-01 -9.13553655e-01 2.36752510e-01 1.15779829e+00
-1.53862283e-01 1.14589848e-01 -1.86004734e-04 3.68646920e-01
-5.84865272e-01 -4.70598906e-01 7.61168659e-01 -6.77988827e-02
-1.00327396e+00 2.76337624e-01 -1.11956507e-01 -4.85093594e-01
6.53254807e-01 -5.77690244e-01 -3.18733066e-01 -6.87072992e-01
-2.52759367e-01 2.40962625e-01 5.67578793e-01 5.00282228e-01
1.03812826e+00 -1.00753582e+00 -8.73956919e-01 6.11631572e-01
4.97499526e-01 3.59528475e-02 -1.12093307e-01 6.62201345e-01
-1.39725363e+00 1.97842509e-01 -4.65734243e-01 -9.33601499e-01
-1.21966338e+00 -3.77702713e-01 6.29723310e-01 5.38155556e-01
-5.35557270e-01 9.12739873e-01 -4.94655132e-01 -7.43792892e-01
-7.05448631e-03 -3.56368959e-01 -1.59650728e-01 9.79756862e-02
3.62276644e-01 2.83611536e-01 3.62391710e-01 -6.30260885e-01
-4.79975939e-01 1.01685143e+00 2.73456007e-01 -5.27534746e-02
1.42865121e+00 4.08000082e-01 -4.72699329e-02 -7.71916434e-02
1.18026257e+00 -2.17524201e-01 -1.16053426e+00 -1.33708641e-01
4.21103798e-02 -6.22408569e-01 9.15662646e-02 -5.01054287e-01
-7.66393900e-01 8.88287067e-01 9.18529153e-01 3.40777427e-01
6.00795507e-01 1.73088871e-02 6.32882953e-01 9.48442042e-01
6.16800129e-01 -6.64905965e-01 -8.93853232e-02 7.98890412e-01
7.93735266e-01 -1.56357074e+00 1.32780463e-01 1.73711374e-01
-2.59349763e-01 1.55418789e+00 5.57479143e-01 -5.97023606e-01
5.75327575e-01 4.72675383e-01 7.67004311e-01 -2.24883482e-01
-1.65060356e-01 -1.80875614e-01 1.90460578e-01 6.76214695e-01
1.69407681e-01 1.03903554e-01 -1.67285994e-01 1.31576166e-01
-8.85201991e-01 -2.81583816e-01 5.01479506e-01 1.07299638e+00
-9.62765515e-01 -6.36286199e-01 -2.95344561e-01 5.21525919e-01
-1.09242797e-01 1.03359528e-01 -1.47144631e-01 9.36842620e-01
-1.06474414e-01 7.27688909e-01 6.60984993e-01 -7.48195231e-01
3.28377932e-01 -4.33438152e-01 2.69506633e-01 -2.13247731e-01
-3.77060771e-01 -3.79816234e-01 4.61961553e-02 -3.22213769e-01
-5.22911668e-01 -6.75584912e-01 -1.29786420e+00 1.39920577e-01
-7.10508823e-01 5.11471093e-01 1.26964211e+00 7.22215891e-01
-8.08630437e-02 -8.00400674e-02 7.03292549e-01 -1.18083608e+00
-5.32016933e-01 -1.09495032e+00 -4.20192599e-01 5.97925372e-02
7.67096996e-01 -7.74265587e-01 -6.37396753e-01 -4.00802970e-01] | [7.9690752029418945, -1.7656896114349365] |
4f0a4969-8f76-429f-ba8e-c60d875f800a | an-annotated-instance-segmentation-xxl-ct | 2212.08639 | null | https://arxiv.org/abs/2212.08639v1 | https://arxiv.org/pdf/2212.08639v1.pdf | An annotated instance segmentation XXL-CT dataset from a historic airplane | The Me 163 was a Second World War fighter airplane and a result of the German air force secret developments. One of these airplanes is currently owned and displayed in the historic aircraft exhibition of the Deutsches Museum in Munich, Germany. To gain insights with respect to its history, design and state of preservation, a complete CT scan was obtained using an industrial XXL-computer tomography scanner. Using the CT data from the Me 163, all its details can visually be examined at various levels, ranging from the complete hull down to single sprockets and rivets. However, while a trained human observer can identify and interpret the volumetric data with all its parts and connections, a virtual dissection of the airplane and all its different parts would be quite desirable. Nevertheless, this means, that an instance segmentation of all components and objects of interest into disjoint entities from the CT data is necessary. As of currently, no adequate computer-assisted tools for automated or semi-automated segmentation of such XXL-airplane data are available, in a first step, an interactive data annotation and object labeling process has been established. So far, seven 512 x 512 x 512 voxel sub-volumes from the Me 163 airplane have been annotated and labeled, whose results can potentially be used for various new applications in the field of digital heritage, non-destructive testing, or machine-learning. This work describes the data acquisition process of the airplane using an industrial XXL-CT scanner, outlines the interactive segmentation and labeling scheme to annotate sub-volumes of the airplane's CT data, describes and discusses various challenges with respect to interpreting and handling the annotated and labeled data. | ['Thomas Wittenberg', 'Michael Salamon', 'Stefan Gerth', 'Andreas Hempfer', 'Nils Reims', 'Roland Gruber'] | 2022-12-16 | null | null | null | null | ['interactive-segmentation'] | ['computer-vision'] | [ 2.16910884e-01 2.20090136e-01 1.74780920e-01 -1.77905113e-01
-3.33007127e-01 -5.51544547e-01 1.37667239e-01 5.63833356e-01
-3.87050986e-01 6.42884910e-01 -2.33270466e-01 -4.74044204e-01
-3.68554622e-01 -7.93737233e-01 -1.83539122e-01 -5.22001088e-01
-5.59041262e-01 1.30527985e+00 5.70845127e-01 -8.50342363e-02
5.24461605e-02 1.16141033e+00 -1.33826280e+00 2.35863775e-01
3.03412795e-01 1.07549191e+00 3.15667689e-01 4.19663936e-01
1.07559338e-01 9.52687711e-02 -5.87458551e-01 -3.01155031e-01
5.36851704e-01 -3.65791380e-01 -9.83769357e-01 5.33393264e-01
1.45920008e-01 -5.01988411e-01 3.16766500e-01 6.23070478e-01
-1.94724221e-02 5.79776708e-03 5.43934584e-01 -9.40200448e-01
5.44647872e-01 2.90615618e-01 -4.02324945e-01 1.06219733e-02
2.10682049e-01 1.22492425e-02 8.49174321e-01 -6.92940950e-01
9.93010223e-01 7.36147821e-01 4.82765645e-01 1.61200732e-01
-1.00007761e+00 -6.32172287e-01 -9.49787647e-02 9.11241025e-03
-1.20362902e+00 4.21109438e-01 7.64495492e-01 -8.94505262e-01
5.67366123e-01 5.65504551e-01 1.39931083e+00 4.22453284e-01
4.15912211e-01 4.51391846e-01 1.25524211e+00 -3.81608069e-01
3.89517844e-01 1.70793056e-01 -1.18163079e-02 9.38455224e-01
1.77880093e-01 1.64033145e-01 -2.89626360e-01 -8.38509053e-02
9.45440948e-01 -3.34598944e-02 -4.01642621e-01 -7.29522049e-01
-1.02754915e+00 3.88262421e-01 4.87089157e-01 5.95979393e-01
-3.32618117e-01 -6.59379512e-02 6.45544469e-01 3.02253570e-02
2.74625421e-01 5.49073458e-01 -4.64481205e-01 5.21966219e-02
-9.86989796e-01 1.38235450e-01 5.12296557e-01 6.39344871e-01
5.91410518e-01 -1.82493001e-01 6.01137340e-01 3.05452079e-01
3.38659912e-01 1.00105852e-01 1.05783962e-01 -4.99384105e-01
2.73587376e-01 8.70709479e-01 -3.15822124e-01 -6.78392291e-01
-6.68037057e-01 -5.24660885e-01 -3.56811255e-01 8.16697776e-01
3.20912331e-01 5.73531464e-02 -9.65463340e-01 6.72805071e-01
6.15731120e-01 -5.36337793e-01 -4.89377916e-01 1.25488830e+00
6.54125571e-01 1.85045898e-01 3.72041389e-02 -4.34941314e-02
1.68737197e+00 -3.25397342e-01 -4.90557343e-01 -5.05681150e-02
4.39743578e-01 -6.55335128e-01 7.93326437e-01 8.76521528e-01
-1.01882350e+00 -2.69567430e-01 -1.27371919e+00 2.86199927e-01
-2.87849665e-01 4.35486948e-03 6.27314270e-01 6.02914274e-01
-3.35640728e-01 6.21530235e-01 -1.08672464e+00 -3.11791897e-01
5.42278409e-01 4.64453012e-01 -7.65112281e-01 4.26758341e-02
-7.84376860e-01 7.76946843e-01 5.05647659e-01 3.53038251e-01
-9.42528665e-01 -6.42828882e-01 -6.46633923e-01 -7.45305270e-02
5.75145781e-01 -5.36260605e-01 8.80068362e-01 -4.98504549e-01
-1.04458487e+00 1.14610744e+00 6.25915170e-01 -3.52307498e-01
7.40330577e-01 -2.55369339e-02 -3.32857639e-01 4.04148161e-01
-1.11283436e-01 2.53702521e-01 3.67887795e-01 -1.56822300e+00
-7.27749348e-01 -6.81098282e-01 1.11207031e-01 -3.11329030e-02
2.71067381e-01 -9.16141719e-02 -1.66550234e-01 -6.18729353e-01
5.95269382e-01 -7.99479783e-01 -1.09006062e-01 1.08248584e-01
-4.72741365e-01 4.06491220e-01 9.17894840e-01 -9.46745038e-01
8.44762146e-01 -1.93108690e+00 -1.44712031e-02 5.31711936e-01
2.78821290e-01 3.36309224e-02 6.03376269e-01 5.77445149e-01
-1.35950491e-01 4.66696694e-02 -7.37151921e-01 -1.35978311e-01
-2.00798720e-01 1.34437829e-01 1.31149748e-02 6.81644320e-01
-3.87735873e-01 3.25842708e-01 -8.09770584e-01 -7.84970164e-01
4.06549692e-01 3.10929924e-01 -3.83699566e-01 -1.12542212e-01
-1.94838136e-01 8.69400918e-01 -5.49470842e-01 1.04081976e+00
5.92850924e-01 4.11546469e-01 2.48814076e-01 -3.21121812e-01
-2.90637106e-01 -1.40113905e-01 -9.07907367e-01 1.75580275e+00
-2.58311659e-01 5.25054038e-01 5.10225594e-01 -6.24275923e-01
9.69981432e-01 5.48860073e-01 8.65061104e-01 -4.51642931e-01
2.92016923e-01 5.09859562e-01 -1.79110661e-01 -5.78576863e-01
6.04908526e-01 -9.62269008e-01 -3.39385569e-02 3.92893940e-01
-9.23987702e-02 -6.88200831e-01 2.39440259e-02 -1.17118426e-01
8.27397227e-01 1.97259039e-01 1.49068058e-01 -2.34912753e-01
4.20062393e-01 5.92799485e-01 4.03750747e-01 -5.06285354e-02
1.61549568e-01 6.83841944e-01 2.80709803e-01 -6.70992970e-01
-8.57927680e-01 -1.19257522e+00 -6.34545386e-01 1.99046403e-01
8.23517963e-02 -2.71764010e-01 -7.35324562e-01 -5.98596454e-01
-2.40976647e-01 4.59199756e-01 -5.14749050e-01 2.64397353e-01
-5.86394548e-01 -3.74931395e-01 6.96181133e-02 1.27739027e-01
3.62153322e-01 -9.44725037e-01 -1.39315343e+00 2.63740271e-01
1.22607745e-01 -8.41437101e-01 2.39994451e-01 3.64146411e-01
-1.18914175e+00 -1.40741372e+00 -2.63072610e-01 -4.84334052e-01
9.19092894e-01 -4.38641727e-01 1.02189159e+00 2.76929170e-01
-7.60732949e-01 3.78259987e-01 -3.72366905e-01 -4.72878218e-01
-2.80685663e-01 -1.46503523e-01 -1.71930641e-01 -1.88529611e-01
-3.61006349e-01 -5.89991927e-01 -4.54665869e-01 5.24466991e-01
-1.25185692e+00 4.24554422e-02 5.75332522e-01 6.74297869e-01
9.56701100e-01 4.82921243e-01 -1.95991382e-01 -1.00247073e+00
3.09342980e-01 -1.28811762e-01 -6.20489657e-01 2.08551973e-01
-1.33421808e-01 -4.06139016e-01 3.58247578e-01 1.57249734e-01
-9.29956734e-01 1.39673755e-01 -2.74922043e-01 -2.28320435e-01
-5.31844735e-01 5.81560433e-01 -2.87865549e-01 -8.68786871e-02
2.73834854e-01 -2.54037708e-01 -1.06693096e-01 -7.00034261e-01
-1.14758141e-01 5.17672598e-01 7.08907247e-01 -5.10026157e-01
8.70597482e-01 6.59950018e-01 3.28177720e-01 -8.47381234e-01
-3.30077797e-01 -3.65206093e-01 -1.23276842e+00 -8.78040195e-01
1.04654765e+00 -1.91237479e-01 -6.28132164e-01 5.65549210e-02
-9.25406516e-01 -2.55519629e-01 -7.39975035e-01 8.11838031e-01
-5.16603768e-01 1.55041054e-01 -4.44490701e-01 -6.55623496e-01
-1.86245054e-01 -1.28054368e+00 9.66829658e-01 -2.05493808e-01
-4.12378848e-01 -9.61533487e-01 9.91469547e-02 7.37183452e-01
1.63714103e-02 9.56863403e-01 1.27928472e+00 -5.48574567e-01
-5.48125029e-01 -6.04812562e-01 4.66153055e-01 2.19636694e-01
5.59478365e-02 -1.00362234e-01 -6.58334672e-01 -2.87130356e-01
3.34610343e-01 -1.15883671e-01 2.75153428e-01 3.25971656e-02
9.93535161e-01 8.60736296e-02 -4.45152283e-01 3.73738885e-01
1.59196472e+00 5.47359943e-01 4.59462255e-01 3.32016379e-01
3.46374184e-01 1.06934071e+00 1.05402255e+00 4.51290458e-01
-2.38355502e-01 6.64139569e-01 9.55617249e-01 -2.42011592e-01
1.09942913e-01 -1.21843535e-02 -3.04575134e-02 5.75551152e-01
-5.53290725e-01 1.33504227e-01 -1.14564168e+00 2.37250552e-01
-9.01566625e-01 -6.58881903e-01 -5.25028348e-01 2.44374418e+00
2.58854717e-01 3.30311775e-01 -3.76230255e-02 5.22277832e-01
4.14729267e-01 -2.30553940e-01 -2.40253538e-01 -3.35600287e-01
2.71718085e-01 3.74581844e-01 4.07946259e-01 4.55852330e-01
-7.19462991e-01 3.51754785e-01 6.15882587e+00 6.01497531e-01
-1.01280570e+00 1.37366310e-01 2.97686517e-01 -2.26774663e-01
-2.28716627e-01 2.73599982e-01 -4.35220480e-01 1.14527140e-02
6.03932500e-01 8.83769393e-02 5.30035701e-03 6.47252202e-01
2.52489485e-02 -7.65866399e-01 -9.05409396e-01 5.66703498e-01
2.45932164e-03 -1.09746528e+00 -5.39208716e-03 5.82800329e-01
2.22789809e-01 -3.85667264e-01 -3.78735960e-01 -2.49397531e-02
-3.29910666e-01 -8.99712086e-01 1.27438200e+00 5.89596093e-01
8.14137399e-01 -8.77650321e-01 8.14087093e-01 2.89410889e-01
-1.34144008e+00 2.84431037e-02 9.81412157e-02 2.14695573e-01
6.52118385e-01 5.30564427e-01 -9.75680232e-01 9.89241540e-01
6.31019771e-01 -1.68233827e-01 -3.53377283e-01 1.15260625e+00
-9.37727541e-02 3.29341173e-01 -2.44256839e-01 3.63213360e-01
1.26884058e-01 -4.83920634e-01 7.77832210e-01 8.58402133e-01
1.66559473e-01 1.87669724e-01 2.02018961e-01 7.86950588e-01
3.48369658e-01 2.00275674e-01 -2.60669619e-01 1.54323101e-01
-1.18134737e-01 1.36476684e+00 -1.42731190e+00 -4.13518958e-02
-5.85315563e-02 6.20611727e-01 -3.90601665e-01 -3.72688510e-02
-8.23587537e-01 -2.01974735e-01 1.95721805e-01 1.18115211e+00
-8.66240263e-02 -5.20795763e-01 -2.33201057e-01 -2.98969209e-01
6.68800250e-02 -4.89264995e-01 4.32686120e-01 -8.53524506e-01
-7.25253761e-01 1.07336092e+00 5.78704357e-01 -1.51055229e+00
3.32131758e-02 -7.14912117e-01 -3.11323613e-01 4.97555882e-01
-6.30625248e-01 -1.28186250e+00 -4.17053908e-01 2.19346255e-01
4.00065780e-01 -6.66386038e-02 7.49792397e-01 2.34884694e-01
-2.16241390e-01 -3.04249555e-01 -2.23450691e-01 6.88685570e-03
8.39523301e-02 -1.18780065e+00 -1.05269685e-01 3.99919420e-01
2.55803838e-02 2.21666157e-01 8.35362732e-01 -9.83694196e-01
-1.15563953e+00 -6.07100725e-01 5.27808845e-01 -7.98726454e-02
7.08840847e-01 -4.39996123e-01 -7.47494817e-01 7.53249288e-01
9.87551808e-02 -1.68524235e-01 8.15306365e-01 -2.28773341e-01
5.73370755e-01 -7.44152665e-02 -1.40636861e+00 1.48325175e-01
6.59194231e-01 -3.71740013e-01 -5.55514812e-01 6.15875900e-01
4.88660336e-02 -6.43959224e-01 -1.23635769e+00 4.95995194e-01
4.51698989e-01 -1.29072821e+00 7.81189144e-01 -2.88109690e-01
2.53693819e-01 -3.84292513e-01 1.48045704e-01 -9.39716458e-01
2.22265214e-01 -1.37849018e-01 6.85285091e-01 9.50578988e-01
2.30543956e-01 -4.84858006e-01 9.70462322e-01 7.09020495e-01
-5.84924579e-01 -9.14163291e-01 -1.33331990e+00 -7.17487156e-01
-1.71402663e-01 -7.06757605e-01 3.48140806e-01 5.79834402e-01
-1.23165436e-01 -3.44553858e-01 1.47450492e-01 1.04246072e-01
5.48870504e-01 2.53487080e-01 5.59277892e-01 -1.60291016e+00
-2.53784388e-01 -1.20905422e-01 -7.13295341e-01 -4.83049572e-01
-2.88189977e-01 -8.68468165e-01 -8.78690332e-02 -1.69563746e+00
-2.74227321e-01 -6.96013689e-01 3.12527150e-01 2.63983488e-01
8.17940056e-01 2.87452161e-01 2.40634177e-02 3.12294066e-01
1.55791357e-01 2.74008065e-01 1.47812879e+00 -1.52962459e-02
-1.51746161e-02 1.02313325e-01 2.63646722e-01 8.06979954e-01
5.15881360e-01 -5.92171907e-01 -2.17475668e-01 -5.03947623e-02
-5.20534739e-02 3.83635461e-01 3.97334337e-01 -1.32760024e+00
1.70653462e-01 9.82909203e-02 3.11239958e-01 -1.05077517e+00
6.45102203e-01 -1.55179703e+00 1.13172781e+00 7.61583805e-01
1.67079404e-01 1.79448053e-01 2.73027271e-01 2.96273172e-01
-4.35200334e-01 -6.21196449e-01 7.37862647e-01 -7.02056706e-01
-4.30728197e-01 2.74083734e-01 -5.18385708e-01 -4.75524455e-01
1.46365035e+00 -6.28056407e-01 2.10360229e-01 2.92204976e-01
-1.24903250e+00 -1.83817923e-01 8.24955523e-01 -2.20744297e-01
6.94221377e-01 -1.00477326e+00 -4.24956828e-01 4.60112810e-01
-4.58267070e-02 4.38071072e-01 6.28907442e-01 1.01702213e+00
-1.39240324e+00 3.65204424e-01 -7.16679871e-01 -6.35069549e-01
-1.41492379e+00 4.02224183e-01 5.25681496e-01 -4.59520698e-01
-8.22004616e-01 4.99094009e-01 1.08261950e-01 -2.12892488e-01
-4.61833589e-02 -4.98340458e-01 -1.77417725e-01 3.00816208e-01
1.92334488e-01 3.63570988e-01 5.93924761e-01 -9.18104887e-01
-2.98210412e-01 7.11834073e-01 3.83850008e-01 -3.22775066e-01
1.51256073e+00 2.53402535e-02 -2.46433660e-01 5.36969483e-01
8.85830402e-01 1.14323817e-01 -1.00578547e+00 1.96672693e-01
-3.93112972e-02 -6.60743237e-01 8.81099235e-03 -7.82497585e-01
-1.37243211e+00 8.91602993e-01 4.61524189e-01 1.68152019e-01
1.28108895e+00 9.09053162e-02 5.05048871e-01 -1.49898738e-01
9.96204555e-01 -8.77642870e-01 -2.68836021e-01 -2.69713961e-02
1.24794257e+00 -7.08554506e-01 4.41645771e-01 -6.95339084e-01
-5.83206475e-01 1.45128775e+00 2.23169014e-01 1.23879716e-01
6.44078970e-01 5.50799966e-01 -1.01393677e-01 -8.53433907e-01
-1.87659279e-01 9.60508510e-02 3.74453396e-01 4.44659501e-01
3.24477106e-01 1.71514787e-02 -3.31096560e-01 3.20105612e-01
-5.64174473e-01 -2.25124180e-01 2.65864730e-01 1.41870713e+00
-4.97012287e-01 -1.10687840e+00 -8.88213754e-01 4.30907667e-01
-3.31155092e-01 3.50918412e-01 -5.15321434e-01 1.40667796e+00
5.30021787e-01 4.29882586e-01 1.98857099e-01 -2.92511702e-01
6.05368853e-01 -1.56372190e-01 5.48358440e-01 -6.19863033e-01
-1.12964761e+00 1.70965150e-01 1.21912047e-01 -2.80866474e-01
-3.26916516e-01 -6.49921715e-01 -1.61852992e+00 -1.40291899e-01
-3.20520878e-01 3.99762928e-01 1.13594258e+00 8.19745004e-01
-6.24045253e-01 6.22752428e-01 2.06778303e-01 -1.07728827e+00
1.84449270e-01 -6.55480862e-01 -1.33159161e+00 1.92860559e-01
-5.20974211e-02 -7.44128287e-01 -2.90603787e-01 1.56468630e-01] | [13.982095718383789, -2.6482956409454346] |
e7a5853a-0e6e-4ff4-9ca8-dcec7e2068c7 | hissnet-sound-event-detection-and-speaker | 2303.07538 | null | https://arxiv.org/abs/2303.07538v1 | https://arxiv.org/pdf/2303.07538v1.pdf | HiSSNet: Sound Event Detection and Speaker Identification via Hierarchical Prototypical Networks for Low-Resource Headphones | Modern noise-cancelling headphones have significantly improved users' auditory experiences by removing unwanted background noise, but they can also block out sounds that matter to users. Machine learning (ML) models for sound event detection (SED) and speaker identification (SID) can enable headphones to selectively pass through important sounds; however, implementing these models for a user-centric experience presents several unique challenges. First, most people spend limited time customizing their headphones, so the sound detection should work reasonably well out of the box. Second, the models should be able to learn over time the specific sounds that are important to users based on their implicit and explicit interactions. Finally, such models should have a small memory footprint to run on low-power headphones with limited on-chip memory. In this paper, we propose addressing these challenges using HiSSNet (Hierarchical SED and SID Network). HiSSNet is an SEID (SED and SID) model that uses a hierarchical prototypical network to detect both general and specific sounds of interest and characterize both alarm-like and speech sounds. We show that HiSSNet outperforms an SEID model trained using non-hierarchical prototypical networks by 6.9 - 8.6 percent. When compared to state-of-the-art (SOTA) models trained specifically for SED or SID alone, HiSSNet achieves similar or better performance while reducing the memory footprint required to support multiple capabilities on-device. | ['Huang', 'Chuan-Che', 'Shuo Zhang', 'Jeremy Kemmerer', 'Mohammad Rasool Izadi', 'Berker Banar', 'N Shashaank'] | 2023-03-13 | null | null | null | null | ['sound-event-detection', 'speaker-identification'] | ['audio', 'speech'] | [-7.43380636e-02 -3.22182506e-01 2.74454296e-01 -2.56366521e-01
-8.96118343e-01 -3.09887379e-01 -1.29959360e-01 9.51874927e-02
-3.60006511e-01 2.85154670e-01 2.21245736e-01 -3.49151194e-01
8.16381574e-02 -6.14643872e-01 -3.36627185e-01 -2.15301380e-01
-1.67710811e-01 1.26382038e-01 7.13565409e-01 -1.52955428e-01
-2.83891439e-01 5.81997633e-01 -1.92204797e+00 5.59152663e-01
4.37813103e-01 1.00395834e+00 4.79853839e-01 1.10315013e+00
5.43923397e-03 6.01069570e-01 -1.04048598e+00 2.85888165e-01
-2.18154386e-01 -2.24327788e-01 -1.70021445e-01 -7.62398839e-01
5.85189521e-01 -2.88974196e-01 -3.09442669e-01 7.97256052e-01
1.40888882e+00 2.97813803e-01 2.79011756e-01 -1.17207706e+00
3.24913189e-02 8.21575701e-01 -4.99490090e-02 4.62404162e-01
5.19158721e-01 7.97529072e-02 7.91545749e-01 -9.44045544e-01
-5.05630732e-01 1.35928774e+00 9.21371222e-01 8.47200871e-01
-1.30327809e+00 -1.24864829e+00 1.75080058e-04 -6.14411309e-02
-1.57457054e+00 -1.11163139e+00 5.33308446e-01 -2.71170497e-01
1.47967732e+00 7.08586693e-01 4.16009337e-01 1.08842337e+00
-1.43240660e-01 5.11163175e-01 6.85108125e-01 -3.94324481e-01
4.77726251e-01 2.77221888e-01 3.46311361e-01 1.95113748e-01
-1.50279403e-02 -5.83277829e-02 -9.93712127e-01 -5.10226786e-01
6.08927131e-01 -3.64779562e-01 -2.87637293e-01 5.33073187e-01
-4.57613766e-01 4.40258950e-01 -1.04814567e-01 4.26754326e-01
-2.54139185e-01 2.14087218e-01 5.17895818e-01 1.88477486e-02
1.71294838e-01 4.21646595e-01 -6.13286316e-01 -5.21980822e-01
-1.15618789e+00 -2.18579620e-02 1.17916000e+00 9.26747859e-01
3.92230600e-01 7.78615355e-01 -2.80066341e-01 1.27440822e+00
2.34398425e-01 5.39158523e-01 4.55687076e-01 -6.58734739e-01
-3.07989065e-02 -1.02592483e-01 -1.02117613e-01 -5.72169542e-01
-7.75936067e-01 -6.56336665e-01 -7.07987309e-01 5.83584346e-02
-7.17891827e-02 -3.57606828e-01 -9.94584024e-01 1.92241132e+00
-5.25018387e-02 5.01169324e-01 -3.38180989e-01 3.95706087e-01
1.02888572e+00 8.26121926e-01 2.22535834e-01 -1.81097001e-01
1.48634338e+00 -5.24215519e-01 -7.87717164e-01 -5.33008397e-01
2.01362461e-01 -7.82968700e-01 1.38750625e+00 7.66564190e-01
-1.23771024e+00 -8.05702031e-01 -1.00090051e+00 1.81646466e-01
-1.54273748e-01 -1.13282606e-01 3.60077530e-01 1.34484291e+00
-1.41528857e+00 1.05092093e-01 -7.61242032e-01 -2.80749917e-01
6.57422245e-02 9.25643027e-01 2.96934217e-01 5.07259250e-01
-1.17674983e+00 4.19744611e-01 -1.11796625e-01 -1.91803575e-01
-9.02830184e-01 -1.01055968e+00 -6.36265635e-01 5.32748282e-01
6.24662936e-02 -5.18837333e-01 1.89960313e+00 -5.21051288e-01
-1.82346523e+00 2.84259439e-01 -3.84577125e-01 -5.57963967e-01
3.01704183e-02 -5.05098283e-01 -1.23438084e+00 4.61890139e-02
-8.03303570e-02 4.85162258e-01 9.38022614e-01 -1.04592586e+00
-7.61577427e-01 2.51277924e-01 -3.05483699e-01 -2.66187459e-01
-7.79843152e-01 5.85538208e-01 -5.17551422e-01 -5.56789339e-01
-6.48329034e-02 -6.01940095e-01 6.89578131e-02 -4.19044703e-01
-5.92928946e-01 -1.29808083e-01 9.68778670e-01 -6.34265423e-01
1.83625305e+00 -2.46737742e+00 -8.00987661e-01 3.54814887e-01
4.98909168e-02 5.84346533e-01 -2.29945388e-02 1.42995805e-01
-1.21690035e-01 1.30428210e-01 2.25429907e-01 -4.56387371e-01
4.89623770e-02 -3.81256230e-02 -2.75486887e-01 -8.16359818e-02
-1.59781918e-01 1.23091072e-01 -4.68396366e-01 -1.40388042e-01
3.07628602e-01 6.39603794e-01 -9.27641451e-01 4.06041354e-01
1.06130555e-01 8.49963278e-02 1.86336646e-03 2.94765860e-01
6.37557626e-01 2.77185202e-01 3.67036089e-02 -3.01246315e-01
-3.10817331e-01 1.04368818e+00 -1.36530805e+00 9.44847584e-01
-9.94075477e-01 6.13084435e-01 5.78569710e-01 -5.28662503e-01
7.01846182e-01 7.13638127e-01 1.70463726e-01 -7.68061936e-01
7.19319358e-02 3.54277670e-01 2.04806507e-01 -4.35422331e-01
2.20339268e-01 -1.37941897e-01 -1.50140956e-01 3.15819681e-01
1.50822610e-01 -2.10087113e-02 -2.43013903e-01 6.08015098e-02
1.34420586e+00 -8.76460373e-01 2.12144747e-01 -2.64159381e-01
3.29783082e-01 -9.03165340e-01 5.32518089e-01 1.28308475e+00
-1.79821983e-01 5.65724432e-01 -5.84130734e-03 2.33778328e-01
-4.03363436e-01 -1.32155263e+00 -1.55295581e-01 1.72937441e+00
-3.63613009e-01 -5.58547914e-01 -7.80021727e-01 -9.03889909e-02
-1.55568346e-01 1.17237318e+00 1.40256673e-01 -2.19964296e-01
-5.77288210e-01 -4.96083349e-01 1.17064786e+00 7.00027466e-01
3.73063207e-01 -1.01075864e+00 -4.52995121e-01 4.99495834e-01
2.49694996e-02 -1.12880552e+00 -6.09919250e-01 7.01567888e-01
-3.01388115e-01 -1.44060925e-02 -2.38449723e-01 -9.03496563e-01
2.96714492e-02 3.65186840e-01 9.93638039e-01 -1.52061433e-01
-1.78849876e-01 6.81595087e-01 -3.21378112e-02 -7.86869705e-01
-4.10373688e-01 -1.91899724e-02 6.95431411e-01 -3.40039954e-02
3.98821533e-01 -1.17654288e+00 -5.32275617e-01 3.62038136e-01
-5.78674793e-01 -2.81488061e-01 2.84779936e-01 3.52309138e-01
5.08061275e-02 4.33396071e-01 9.66913760e-01 -5.09870827e-01
6.90256536e-01 -4.16132540e-01 -1.27008885e-01 -2.81863660e-01
-9.97060463e-02 -5.10181785e-01 7.45440364e-01 -7.76871264e-01
-1.11285150e+00 -8.35424811e-02 -9.77211535e-01 -1.10313624e-01
-3.94076049e-01 2.40846761e-02 -5.00475466e-01 2.08986048e-02
6.78994775e-01 2.14783661e-02 -5.41441381e-01 -7.76212096e-01
-1.47634804e-01 1.28819847e+00 5.95301092e-01 -4.27320689e-01
4.16245252e-01 9.19565782e-02 -5.13526738e-01 -1.30552506e+00
-5.75093210e-01 -6.66660130e-01 -2.44410653e-02 -2.35890336e-02
6.45055115e-01 -1.14064360e+00 -9.95364666e-01 5.08151829e-01
-1.03633785e+00 -4.24142718e-01 -8.89969543e-02 5.13341069e-01
-1.87252924e-01 -1.27087668e-01 -9.02686059e-01 -1.47995698e+00
-5.27123928e-01 -7.27571368e-01 7.94839323e-01 3.65122229e-01
-9.39264238e-01 -5.44038653e-01 -2.91415185e-01 -1.24678634e-01
8.08761001e-01 -6.34846866e-01 9.24604595e-01 -9.45827246e-01
-1.38595477e-01 -1.95373848e-01 1.53032005e-01 5.81025660e-01
2.65681893e-01 -3.12715977e-01 -1.79888153e+00 -1.89632565e-01
3.45482737e-01 4.94912490e-02 6.28746033e-01 8.13138247e-01
1.35111570e+00 -2.40578160e-01 -3.40906292e-01 2.72874057e-01
8.21342349e-01 6.76652670e-01 5.55656850e-01 -2.15991810e-01
4.72117126e-01 2.19324276e-01 -1.56550799e-02 5.18451214e-01
7.32921436e-02 7.26303816e-01 2.53219735e-02 -1.90164164e-01
-4.21270967e-01 -1.90403715e-01 7.35726655e-01 1.15157413e+00
5.51903963e-01 -5.87398827e-01 -7.97356367e-01 5.51153660e-01
-1.36183870e+00 -8.78756046e-01 -1.62530228e-01 2.41764188e+00
9.38673615e-01 7.09978878e-01 3.97051394e-01 3.92659158e-01
9.00645196e-01 -1.49841860e-01 -4.04582381e-01 -7.28866518e-01
-1.93510074e-02 9.66016650e-01 1.63224146e-01 6.31958127e-01
-9.08850908e-01 5.77536523e-01 6.55937433e+00 1.08064914e+00
-1.25728917e+00 4.80762661e-01 2.01805249e-01 -5.30669987e-01
4.22378071e-02 -5.24383724e-01 -1.22220671e+00 5.24112582e-01
1.72319567e+00 7.87207112e-02 2.69714445e-01 7.79823542e-01
5.42868376e-01 -1.47054046e-01 -1.30601895e+00 1.14976168e+00
-9.60001573e-02 -9.01652396e-01 -1.16023228e-01 -2.59467781e-01
9.47454646e-02 -5.22553511e-02 3.13032120e-01 6.10686481e-01
1.44671604e-01 -8.93433452e-01 7.93927848e-01 3.03012639e-01
9.34058666e-01 -8.81709754e-01 3.07339072e-01 2.62989402e-01
-1.36840081e+00 -3.65258217e-01 -3.74045447e-02 -2.54288912e-01
2.16226742e-01 8.43403637e-01 -1.09836531e+00 -1.96647540e-01
1.10671031e+00 -1.49555162e-01 -2.96554267e-01 1.29658747e+00
6.91696629e-02 1.42182064e+00 -7.24093676e-01 -2.28862122e-01
-2.29907796e-01 7.49748707e-01 5.93045950e-01 1.74417663e+00
6.03656650e-01 7.96505660e-02 1.28939040e-02 5.89644849e-01
-2.30785068e-02 -1.73059732e-01 -2.16749892e-01 3.11119288e-01
1.02678454e+00 1.04269505e+00 -3.72219443e-01 -2.50655711e-01
-3.61560881e-01 7.62462258e-01 -3.48914653e-01 3.19860131e-01
-6.66979849e-01 -7.19586909e-01 8.99288654e-01 3.86809587e-01
1.51298448e-01 -6.15139976e-02 -3.08439821e-01 -4.60374832e-01
-3.92389685e-01 -1.01721478e+00 2.33756781e-01 -7.52841473e-01
-1.25420165e+00 6.48146749e-01 -3.41771841e-01 -1.05207622e+00
-1.22010028e-02 -4.36325967e-01 -9.54331219e-01 9.20238018e-01
-7.87427366e-01 -5.15196800e-01 7.25246919e-03 4.87090230e-01
6.92081690e-01 -6.85618119e-03 1.12749434e+00 8.64644289e-01
-5.98112643e-01 1.05189967e+00 -2.82144934e-01 -4.69353311e-02
7.89846838e-01 -1.04497302e+00 5.96836448e-01 6.03561461e-01
-1.05100006e-01 9.22278762e-01 8.48614097e-01 -5.06752789e-01
-1.03789711e+00 -1.14366937e+00 8.50200176e-01 -2.42384642e-01
3.50630909e-01 -1.01404572e+00 -1.01867795e+00 4.73001808e-01
1.12920560e-01 -1.57138005e-01 1.07625151e+00 4.34105098e-01
-2.79803455e-01 -4.69704747e-01 -1.03551614e+00 7.65428126e-01
1.01869953e+00 -8.85394394e-01 -2.02808142e-01 -3.03665875e-03
1.02108979e+00 -1.09969407e-01 -3.09484243e-01 2.08390638e-01
5.58679700e-01 -1.00303352e+00 1.05205989e+00 -1.71791375e-01
-5.06643891e-01 -2.17268839e-01 -3.98082674e-01 -1.15444279e+00
-5.56195796e-01 -1.13309395e+00 -2.63667434e-01 1.74131143e+00
4.47192788e-01 -7.06054568e-01 4.22590852e-01 5.45660317e-01
-5.31264544e-01 -2.58530259e-01 -9.93412971e-01 -1.00548077e+00
-3.61226529e-01 -1.22613692e+00 5.64449966e-01 3.91093999e-01
3.48150283e-02 6.07569873e-01 -3.66899699e-01 7.30731845e-01
2.91273892e-01 -7.60424376e-01 4.39758599e-01 -1.25678706e+00
-4.56985205e-01 -4.79356349e-01 -2.64121324e-01 -1.16717255e+00
-1.77744165e-01 -3.99684042e-01 3.19878310e-01 -1.06545103e+00
-1.80270776e-01 -4.21625286e-01 -6.78742707e-01 4.83726382e-01
1.01728000e-01 1.45995706e-01 -1.71515197e-01 -4.88637745e-01
-4.33504790e-01 2.21896648e-01 2.02242494e-01 6.23427629e-02
-7.99751937e-01 5.80026150e-01 -8.36610615e-01 1.06623805e+00
8.50011587e-01 -4.53431249e-01 -4.24683809e-01 -2.31447846e-01
4.11724485e-02 4.90376540e-02 4.26397175e-01 -1.65005112e+00
6.54772043e-01 2.88455099e-01 3.41246277e-01 -6.96557701e-01
8.22937429e-01 -5.73489130e-01 1.15207592e-02 3.84675264e-01
-3.41787457e-01 -4.09536332e-01 9.15513456e-01 2.45301247e-01
3.82641256e-02 -2.03616276e-01 9.77109313e-01 1.99698672e-01
-4.56125915e-01 -1.00203767e-01 -1.27451420e+00 2.98650321e-02
3.20803344e-01 3.21226008e-02 -7.54754767e-02 -8.57333779e-01
-8.47214937e-01 -2.82667994e-01 -4.13217992e-01 3.25260609e-01
6.72179341e-01 -1.09860349e+00 -2.34784737e-01 5.00956178e-01
-2.97903895e-01 -4.74196702e-01 6.04893982e-01 5.70059061e-01
4.61687818e-02 4.20189500e-01 1.85816795e-01 -5.58220208e-01
-1.51179969e+00 1.72009870e-01 6.34579897e-01 3.41852814e-01
-3.85882318e-01 1.39265060e+00 3.91392142e-01 -3.14321816e-01
7.00037956e-01 -3.86766583e-01 -1.03129983e-01 3.77992331e-03
8.42194974e-01 6.05579197e-01 4.27881211e-01 -2.64682084e-01
-5.68706393e-01 8.75159279e-02 -6.59972280e-02 -3.65756333e-01
1.07872915e+00 -1.26143336e-01 5.14451303e-02 6.27845228e-01
1.10448122e+00 6.00907326e-01 -6.39945209e-01 -2.92479217e-01
-1.31178916e-01 -6.17631413e-02 3.42105091e-01 -8.60372901e-01
-6.02676213e-01 1.08601165e+00 1.07792509e+00 5.24416625e-01
1.44243062e+00 -6.27459735e-02 1.10982478e+00 2.58826643e-01
2.54247934e-01 -1.15017033e+00 1.53914765e-01 5.92493296e-01
6.04208827e-01 -4.57068801e-01 -5.48791170e-01 -5.48183203e-01
-2.27010801e-01 9.11582172e-01 8.85000944e-01 3.01599801e-01
1.06821728e+00 9.50500786e-01 -1.49521366e-01 8.62841234e-02
-8.74084473e-01 -3.08366358e-01 2.39881769e-01 8.04865122e-01
4.66205984e-01 2.10326597e-01 4.16638017e-01 1.32432520e+00
-5.13964593e-01 -3.62298518e-01 3.35357159e-01 8.18951368e-01
-9.62982059e-01 -8.27617466e-01 -5.71499169e-01 8.07644606e-01
-6.34981871e-01 -4.92153764e-01 -1.63371190e-01 2.60265082e-01
2.98679262e-01 1.64069605e+00 2.77902156e-01 -7.15133846e-01
7.30954349e-01 2.53946632e-01 1.12730809e-01 -1.11858344e+00
-8.89802575e-01 6.12223804e-01 3.89443517e-01 -2.77186841e-01
2.86144942e-01 -4.80149031e-01 -1.43946636e+00 -1.64754242e-01
-3.64871591e-01 7.40912706e-02 4.78232175e-01 6.77795231e-01
4.97522593e-01 1.18282270e+00 6.00092292e-01 -8.76092017e-01
-2.75661945e-01 -9.15550530e-01 -7.41588950e-01 -3.83221835e-01
6.31134152e-01 -3.92832696e-01 -4.79360342e-01 -2.95645982e-01] | [14.908876419067383, 5.927661418914795] |
19254e3b-a012-4eae-a78b-4c2b95606b2e | respect-reinforcement-learning-based-edge | 2304.04716 | null | https://arxiv.org/abs/2304.04716v1 | https://arxiv.org/pdf/2304.04716v1.pdf | RESPECT: Reinforcement Learning based Edge Scheduling on Pipelined Coral Edge TPUs | Deep neural networks (DNNs) have substantial computational and memory requirements, and the compilation of its computational graphs has a great impact on the performance of resource-constrained (e.g., computation, I/O, and memory-bound) edge computing systems. While efficient execution of their computational graph requires an effective scheduling algorithm, generating the optimal scheduling solution is a challenging NP-hard problem. Furthermore, the complexity of scheduling DNN computational graphs will further increase on pipelined multi-core systems considering memory communication cost, as well as the increasing size of DNNs. Using the synthetic graph for the training dataset, this work presents a reinforcement learning (RL) based scheduling framework RESPECT, which learns the behaviors of optimal optimization algorithms and generates near-optimal scheduling results with short solving runtime overhead. Our framework has demonstrated up to $\sim2.5\times$ real-world on-chip inference runtime speedups over the commercial compiler with ten popular ImageNet models deployed on the physical Coral Edge TPUs system. Moreover, compared to the exact optimization methods, the proposed RL scheduling improves the scheduling optimization runtime by up to 683$\times$ speedups compared to the commercial compiler and matches the exact optimal solutions with up to 930$\times$ speedups. Finally, we perform a comprehensive generalizability test, which demonstrates RESPECT successfully imitates optimal solving behaviors from small synthetic graphs to large real-world DNNs computational graphs. | ['Cunxi Yu', 'Daniel Robinson', 'Yingjie Li', 'Jiaqi Yin'] | 2023-04-10 | null | null | null | null | ['edge-computing'] | ['time-series'] | [-6.62063658e-02 1.87711760e-01 -9.49642137e-02 -3.27320874e-01
-8.99564028e-02 -1.96017891e-01 -3.40468772e-02 -8.58203247e-02
-8.14895391e-01 6.32611096e-01 -4.53585595e-01 -6.61322176e-01
-1.97626323e-01 -1.08876109e+00 -1.04991055e+00 -4.92466152e-01
-2.92658716e-01 7.15444624e-01 1.82119027e-01 5.64707853e-02
-1.48864284e-01 5.13452768e-01 -1.65179467e+00 -2.42070615e-01
6.67770267e-01 1.37272668e+00 4.91239458e-01 6.92549944e-01
-1.82462811e-01 7.55962849e-01 -6.17167711e-01 -4.15660083e-01
5.21958530e-01 1.73088968e-01 -4.66638654e-01 -2.06456706e-01
7.28323996e-01 -4.49775547e-01 -7.14043140e-01 1.45433521e+00
3.30720663e-01 8.95051882e-02 -2.20896509e-02 -1.75291252e+00
-2.32987136e-01 9.70565140e-01 -4.02395517e-01 2.93384731e-01
-4.83497143e-01 5.06796502e-02 1.10326564e+00 -3.44888687e-01
6.89927936e-01 9.19050992e-01 6.16089404e-01 5.45669436e-01
-9.99472141e-01 -1.10307848e+00 1.97578937e-01 6.09773338e-01
-1.28655875e+00 -3.52585793e-01 5.03810287e-01 2.60197464e-02
1.54669261e+00 -1.63370773e-01 9.58615661e-01 7.62524426e-01
2.51732290e-01 5.81600964e-01 7.12276459e-01 -1.24587327e-01
6.77430034e-01 -4.03143317e-01 5.42983770e-01 8.84422958e-01
6.78542316e-01 1.91566706e-01 -3.36792886e-01 2.30437100e-01
3.65545183e-01 -1.99080020e-01 1.48297390e-02 2.14391172e-01
-5.49636602e-01 3.53356659e-01 7.12093711e-01 -5.13023250e-02
-2.76306093e-01 8.70780706e-01 6.84305966e-01 1.26946658e-01
1.02256224e-01 2.71284848e-01 -6.75419450e-01 -2.94381350e-01
-1.00098777e+00 2.40101233e-01 1.09476805e+00 1.62074053e+00
9.48563278e-01 5.66449225e-01 2.09509805e-01 2.46123895e-01
2.88476497e-01 7.69793570e-01 2.30114430e-01 -8.30936074e-01
4.84444976e-01 7.76374102e-01 -5.54082930e-01 -8.68133962e-01
-8.75672519e-01 -6.60397589e-01 -1.09525836e+00 1.02974355e-01
3.51023912e-01 -4.79968935e-01 -8.70018244e-01 1.80194700e+00
1.23666510e-01 4.42290306e-01 1.71121061e-01 9.24022794e-01
8.11558485e-01 6.37681603e-01 1.49127796e-01 -4.50485796e-02
1.53044164e+00 -1.23782301e+00 -6.58060372e-01 -5.88057697e-01
7.06078231e-01 -2.06600949e-01 7.49444366e-01 4.22560982e-02
-1.09855700e+00 -3.36468369e-01 -1.30655122e+00 -4.46476042e-02
-2.76494294e-01 1.02085873e-01 1.02344406e+00 7.91364670e-01
-1.36648476e+00 6.38858438e-01 -1.29602301e+00 7.42450729e-02
5.04950821e-01 7.25602388e-01 -1.63981598e-02 -1.50827944e-01
-8.89333367e-01 4.64822978e-01 6.13338888e-01 1.49289057e-01
-1.03368652e+00 -1.30584311e+00 -6.97387755e-01 5.41457951e-01
4.33022499e-01 -5.80947638e-01 1.06370831e+00 -8.54173541e-01
-1.43034840e+00 5.02238333e-01 5.91299534e-02 -7.21350968e-01
1.22458912e-01 2.15761662e-01 -4.40000862e-01 1.13434918e-01
-3.14370722e-01 6.95506871e-01 4.81156409e-01 -4.89066303e-01
-6.39973164e-01 -4.95929420e-01 3.31239015e-01 -1.51453793e-01
-6.07799530e-01 -1.87178731e-01 -5.75059533e-01 -1.31398112e-01
-2.67704546e-01 -1.29736507e+00 -4.07422572e-01 9.37935617e-03
-2.68297166e-01 8.92265420e-03 7.46377647e-01 -2.66662627e-01
1.03293014e+00 -2.15059185e+00 2.55123973e-02 2.75133669e-01
6.20839894e-01 2.26770699e-01 -1.10431053e-01 -1.06472671e-01
1.81565285e-01 -8.87916014e-02 2.18213335e-01 -2.42953643e-01
4.23888892e-01 4.92371231e-01 9.44523215e-02 2.60274589e-01
-2.00823084e-01 8.88720572e-01 -7.34860122e-01 -3.82435381e-01
-3.58622521e-02 2.86634654e-01 -8.65877271e-01 1.27146542e-01
-4.77048039e-01 -2.74991363e-01 -2.87933230e-01 4.59992707e-01
8.04316580e-01 -6.92236245e-01 6.79047048e-01 -4.24757451e-01
2.18482483e-02 -1.84734911e-02 -1.07865357e+00 1.56744075e+00
-6.05187833e-01 9.68976676e-01 3.44544709e-01 -1.06276810e+00
7.89026558e-01 -6.45802319e-02 3.49334598e-01 -1.09571528e+00
4.59565252e-01 9.08197090e-02 3.23104858e-01 -1.62089080e-01
5.84519029e-01 2.57462442e-01 5.29846363e-02 4.29673940e-01
9.61573645e-02 1.54105499e-01 7.43158281e-01 1.58611611e-01
1.67555785e+00 -4.13443685e-01 -3.49552602e-01 -7.45326996e-01
1.88327402e-01 2.34419435e-01 7.81283259e-01 5.27050555e-01
-3.18533272e-01 -2.84430087e-01 8.68054450e-01 -6.95427299e-01
-1.21303022e+00 -7.94182599e-01 -8.39623958e-02 1.02500081e+00
2.77360588e-01 -3.81217599e-01 -1.27878034e+00 -1.70132726e-01
-2.49993265e-01 6.59473300e-01 -2.92158097e-01 -2.03313366e-01
-4.75626349e-01 -9.85391021e-01 7.52258062e-01 8.06246936e-01
8.65992427e-01 -9.99999166e-01 -9.92115438e-01 2.92824745e-01
4.59443033e-01 -1.91080785e+00 -2.96599030e-01 3.55298162e-01
-8.56563926e-01 -9.91994500e-01 -1.55561427e-02 -9.56838369e-01
8.36171806e-01 -1.00276299e-01 1.33472621e+00 2.95316339e-01
-5.55289686e-01 -1.32139437e-02 -2.46380866e-02 -2.72962898e-01
-2.13384628e-01 3.18550676e-01 -2.31017023e-02 -4.49950963e-01
3.66510063e-01 -7.90483892e-01 -5.75798929e-01 1.12155080e-02
-8.62962723e-01 3.65714133e-01 5.05554736e-01 7.47937560e-01
6.64334416e-01 4.62402135e-01 3.04649919e-01 -9.12571013e-01
5.66600382e-01 -1.13611706e-01 -1.54819524e+00 2.90459305e-01
-1.02884853e+00 5.26463151e-01 1.28417397e+00 -3.19057763e-01
-6.93341732e-01 8.39946344e-02 -4.86242659e-02 -7.57095039e-01
2.02053800e-01 3.95671546e-01 -2.30081126e-01 -2.57161558e-01
4.15397972e-01 -1.87507849e-02 -2.02577427e-01 1.99279830e-01
-7.79654309e-02 2.65088975e-02 4.29561794e-01 -7.36020148e-01
4.89863157e-01 2.29609609e-01 4.85981107e-01 -6.65617287e-01
-5.39655626e-01 1.68166503e-01 1.54175870e-02 -3.30045909e-01
7.24501073e-01 -8.59357536e-01 -1.56880534e+00 4.46317434e-01
-9.60305572e-01 -8.43058228e-01 8.56232643e-02 3.69286716e-01
-1.49612695e-01 -9.85658616e-02 -8.08758616e-01 -4.58799362e-01
-7.20649123e-01 -1.70485902e+00 7.93629169e-01 5.34026921e-01
1.13946944e-01 -9.79352653e-01 -3.87535572e-01 1.48769304e-01
6.76065266e-01 5.84668256e-02 1.10689819e+00 -4.04032320e-01
-1.10176134e+00 -1.64311507e-03 -8.21545959e-01 1.02950908e-01
-7.11095572e-01 1.16658129e-01 -7.36054480e-01 -3.53270382e-01
-3.58561099e-01 -1.43665776e-01 4.09318149e-01 5.24344265e-01
1.39679503e+00 -2.56456941e-01 -3.12499762e-01 1.08087146e+00
1.82172096e+00 2.73336381e-01 4.57284749e-01 1.46179736e-01
8.39578152e-01 -9.74314585e-02 1.05491273e-01 6.67099953e-01
4.10675824e-01 4.41014260e-01 9.76618052e-01 1.69602856e-01
-1.18563898e-01 1.78315826e-02 2.43735746e-01 1.00884247e+00
6.44667149e-02 -3.41780186e-01 -1.03096163e+00 2.66720414e-01
-1.74922061e+00 -5.11886418e-01 3.54611780e-03 1.84266043e+00
5.13484001e-01 3.95924330e-01 -3.28043044e-01 -4.58918288e-02
5.81781745e-01 -4.14444283e-02 -1.06766438e+00 -5.90554893e-01
2.84720749e-01 6.51124418e-01 1.08371198e+00 3.12868804e-01
-5.52295148e-01 1.05244172e+00 5.70799923e+00 9.17719781e-01
-1.27628827e+00 7.78947175e-02 6.00731909e-01 -3.46846581e-01
-7.80890137e-02 -6.97655082e-02 -1.08445168e+00 4.83084530e-01
1.58521199e+00 -6.47731125e-01 1.04591739e+00 1.31590199e+00
-5.87690286e-02 1.32216007e-01 -1.33248365e+00 1.17668664e+00
-2.87459999e-01 -1.71691477e+00 -3.14299077e-01 1.95271447e-01
7.77508020e-01 4.56755012e-01 -3.51444960e-01 4.74188983e-01
6.16171300e-01 -8.94430220e-01 7.97434449e-01 -5.19486517e-02
8.49730074e-01 -1.11159897e+00 8.06475401e-01 2.13449121e-01
-1.39253366e+00 -5.52916229e-02 -5.95596731e-01 -2.05487698e-01
6.92123398e-02 7.79076993e-01 -7.38534570e-01 1.69405267e-01
8.77351880e-01 4.60082948e-01 -2.84380555e-01 6.12347186e-01
-1.86859891e-01 2.73553193e-01 -4.43551153e-01 -6.49297059e-01
1.65088549e-01 -3.76113147e-01 9.29017588e-02 1.00758731e+00
2.12933347e-01 9.90078598e-02 8.86576697e-02 9.65743721e-01
-6.86848283e-01 -1.64040238e-01 -1.83422580e-01 -3.59860450e-01
6.73744857e-01 1.65930486e+00 -1.00120938e+00 -4.34401602e-01
-4.37931597e-01 5.01995862e-01 7.56005228e-01 2.93029904e-01
-1.39437187e+00 -3.70384336e-01 1.04220665e+00 -1.92868814e-01
6.87767938e-02 -3.72268617e-01 -5.05081356e-01 -8.34975064e-01
2.65314663e-03 -5.82707226e-01 9.02962983e-02 -5.64216793e-01
-8.11768234e-01 9.54636991e-01 -2.94749707e-01 -5.97279191e-01
7.85485283e-02 -1.03807461e+00 -3.59019935e-01 4.42860097e-01
-1.50831234e+00 -5.06232083e-01 -7.08274364e-01 3.64888370e-01
8.17478001e-02 -2.30422109e-01 5.97966909e-01 8.63402188e-01
-1.19542003e+00 1.02569664e+00 -1.91879030e-02 1.70929015e-01
-4.39854637e-02 -1.00863647e+00 4.89656061e-01 8.43188763e-01
-9.66900140e-02 2.53652811e-01 6.55005276e-01 -3.04353118e-01
-2.37882972e+00 -1.11099386e+00 3.49974126e-01 5.93444943e-01
1.12829804e+00 -4.99932766e-01 -5.80579579e-01 7.35437274e-01
1.05588600e-01 3.60616595e-01 3.28681201e-01 5.41699566e-02
-1.08431391e-01 -6.10261083e-01 -1.02695084e+00 9.31093395e-01
1.46042180e+00 -2.20234409e-01 2.94625014e-01 5.31128764e-01
8.94891858e-01 -8.75119328e-01 -9.41602290e-01 2.77162135e-01
4.71689433e-01 -7.17618585e-01 6.82664335e-01 -3.94521445e-01
6.52804613e-01 8.54129642e-02 -1.63725197e-01 -1.04337382e+00
-8.26750398e-02 -3.34832937e-01 -1.81661010e-01 6.59238458e-01
5.72423041e-01 -7.40908861e-01 1.34050512e+00 1.20307744e+00
-6.18397474e-01 -8.08942258e-01 -9.58970070e-01 -8.68536651e-01
-3.95250082e-01 -8.24706376e-01 7.35463560e-01 6.30715847e-01
-1.16470300e-01 2.74174035e-01 2.54123867e-01 4.38008904e-01
9.42383468e-01 1.74490079e-01 6.56428695e-01 -8.43651354e-01
-5.75721622e-01 -7.57651150e-01 -6.33581996e-01 -9.20658588e-01
7.43411720e-01 -1.18438673e+00 -2.62450010e-01 -1.14266944e+00
6.30279779e-02 -6.69825256e-01 -3.75972353e-02 6.00075603e-01
2.23517790e-01 -1.64142445e-01 1.68739840e-01 -4.50436711e-01
-7.97109425e-01 3.58030915e-01 1.20060098e+00 -2.67345399e-01
2.15526551e-01 -4.71588552e-01 -3.25876802e-01 5.31774700e-01
5.42474508e-01 -5.17940998e-01 -5.77129602e-01 -9.13041413e-01
6.14418685e-01 2.59585321e-01 1.68398425e-01 -1.42192578e+00
7.94110417e-01 -3.56736337e-03 -2.07542032e-01 -2.67231226e-01
1.49801686e-01 -1.22857809e+00 3.82031441e-01 8.73743832e-01
2.38207821e-02 5.37595332e-01 6.82037592e-01 3.64301473e-01
1.09154060e-02 -3.63808215e-01 6.62455976e-01 1.06732510e-01
-1.12881148e+00 9.11433458e-01 -9.05016512e-02 4.20131594e-01
1.13400459e+00 -8.07481334e-02 -7.76517749e-01 6.25748038e-02
-5.55959344e-01 4.33241904e-01 2.44322255e-01 3.20424102e-02
3.98584932e-01 -9.45684075e-01 -3.06051314e-01 3.75726819e-01
-2.78213173e-01 1.85750529e-01 5.16607344e-01 5.35595715e-01
-1.17001545e+00 3.72970402e-01 -4.70802665e-01 -5.66610932e-01
-1.03266239e+00 5.13646305e-01 3.52527350e-01 -5.15705109e-01
-6.58634186e-01 8.43800068e-01 -2.08571777e-02 -9.27566513e-02
3.50804031e-01 -5.12711585e-01 3.88616890e-01 -4.24405217e-01
3.87381226e-01 6.42460883e-01 3.65301609e-01 1.24669835e-01
-4.58430856e-01 1.73359200e-01 -9.03117582e-02 2.71393090e-01
1.48703086e+00 4.37366098e-01 -3.28804225e-01 -3.05842698e-01
1.18752062e+00 -6.51280165e-01 -1.30448818e+00 -1.53494775e-01
7.78446207e-03 7.88099915e-02 4.79380965e-01 -3.04893434e-01
-1.95929253e+00 6.90652907e-01 4.77317542e-01 1.07548967e-01
1.39293468e+00 -3.07206273e-01 9.90215838e-01 7.07265615e-01
6.45854652e-01 -1.33004189e+00 -1.61628574e-01 6.70603931e-01
2.15221882e-01 -9.09376562e-01 2.78772619e-02 -4.50873643e-01
-2.47346759e-02 1.19699788e+00 1.20877016e+00 -3.57443333e-01
9.19335067e-01 1.22245657e+00 -3.00114989e-01 -3.57568055e-01
-1.08428919e+00 3.04657638e-01 -3.12388510e-01 1.91449881e-01
-1.32670654e-02 4.03614551e-01 -1.29275918e-01 6.25471890e-01
-5.84823668e-01 1.75433122e-02 4.83035922e-01 5.00526488e-01
-7.65912011e-02 -8.70252490e-01 1.78648829e-01 5.72990894e-01
-1.65943339e-01 -3.07154596e-01 3.72453153e-01 8.82767439e-01
-1.98166281e-01 5.37218213e-01 6.26668394e-01 -4.51346993e-01
2.22998619e-01 -2.08609834e-01 4.38346893e-01 -2.15303928e-01
-6.51730478e-01 -4.78005975e-01 2.48494819e-01 -7.69998431e-01
1.37364388e-01 -8.12268332e-02 -1.87790358e+00 -9.99783218e-01
-4.95480411e-02 -2.31909409e-01 1.08379984e+00 9.25399661e-01
6.86135948e-01 1.20689559e+00 1.83477908e-01 -7.83640206e-01
-4.05610830e-01 -5.32680094e-01 -6.05126619e-01 2.12156728e-01
-2.99154073e-01 -4.51131254e-01 -3.15396279e-01 -2.59727061e-01] | [7.15974760055542, 5.405590534210205] |
30b10d1c-3f2c-4416-9ec8-c34ce5455caa | cross-view-tracking-for-multi-human-3d-pose | 2003.03972 | null | https://arxiv.org/abs/2003.03972v3 | https://arxiv.org/pdf/2003.03972v3.pdf | Cross-View Tracking for Multi-Human 3D Pose Estimation at over 100 FPS | Estimating 3D poses of multiple humans in real-time is a classic but still challenging task in computer vision. Its major difficulty lies in the ambiguity in cross-view association of 2D poses and the huge state space when there are multiple people in multiple views. In this paper, we present a novel solution for multi-human 3D pose estimation from multiple calibrated camera views. It takes 2D poses in different camera coordinates as inputs and aims for the accurate 3D poses in the global coordinate. Unlike previous methods that associate 2D poses among all pairs of views from scratch at every frame, we exploit the temporal consistency in videos to match the 2D inputs with 3D poses directly in 3-space. More specifically, we propose to retain the 3D pose for each person and update them iteratively via the cross-view multi-human tracking. This novel formulation improves both accuracy and efficiency, as we demonstrated on widely-used public datasets. To further verify the scalability of our method, we propose a new large-scale multi-human dataset with 12 to 28 camera views. Without bells and whistles, our solution achieves 154 FPS on 12 cameras and 34 FPS on 28 cameras, indicating its ability to handle large-scale real-world applications. The proposed dataset is released at https://github.com/longcw/crossview_3d_pose_tracking. | ['Shuang Liu', 'Haizhou Ai', 'Zijie Zhuang', 'Long Chen', 'Rui Chen'] | 2020-03-09 | cross-view-tracking-for-multi-human-3d-pose-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Chen_Cross-View_Tracking_for_Multi-Human_3D_Pose_Estimation_at_Over_100_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Chen_Cross-View_Tracking_for_Multi-Human_3D_Pose_Estimation_at_Over_100_CVPR_2020_paper.pdf | cvpr-2020-6 | ['3d-multi-person-pose-estimation'] | ['computer-vision'] | [-3.76503229e-01 -6.23910129e-01 8.90289843e-02 -1.35543168e-01
-6.33786559e-01 -7.80535638e-01 2.34402210e-01 -2.50908792e-01
-6.48138583e-01 4.56379116e-01 -9.57755670e-02 4.31366414e-01
1.68858573e-01 -1.48598179e-01 -7.14106858e-01 -3.90620142e-01
-8.83407518e-02 6.93712890e-01 5.06033242e-01 -9.51161981e-03
-1.39447778e-01 5.44304371e-01 -1.43250918e+00 -1.99080765e-01
2.17945427e-01 6.16286814e-01 2.22310983e-02 9.94426131e-01
6.39728665e-01 1.53955504e-01 -6.07555091e-01 -4.90776777e-01
7.59666264e-01 -1.30589396e-01 -2.45347530e-01 5.32613337e-01
1.02661896e+00 -6.78129077e-01 -3.97632301e-01 9.54205453e-01
6.29441917e-01 1.38692603e-01 -6.41305968e-02 -1.58229423e+00
4.59776819e-02 -4.76011992e-01 -1.04870522e+00 1.53863102e-01
9.80499029e-01 2.09129408e-01 5.33198476e-01 -8.11841190e-01
7.34317958e-01 1.39926004e+00 9.42052305e-01 6.03939235e-01
-9.32644904e-01 -7.08729804e-01 2.46726125e-01 9.79069769e-02
-1.66800141e+00 -3.27617466e-01 4.56820697e-01 -5.53149521e-01
5.56115866e-01 2.20524788e-01 1.12653792e+00 1.08338201e+00
3.20549160e-01 4.29610729e-01 7.97288179e-01 -1.35228589e-01
-2.54630029e-01 -9.87469405e-02 7.15125129e-02 8.58586371e-01
6.89213216e-01 5.87735064e-02 -6.35769367e-01 -2.55012035e-01
1.11991704e+00 7.65227437e-01 -2.42729917e-01 -7.34685898e-01
-1.64275742e+00 5.87271810e-01 7.70031139e-02 -2.18186811e-01
-2.57679433e-01 2.83064276e-01 2.76879221e-01 2.95792371e-02
1.29969969e-01 -7.99049512e-02 -5.47856271e-01 -1.28612444e-01
-4.74426210e-01 6.00360096e-01 4.84751374e-01 1.36243320e+00
5.24951041e-01 -3.58422965e-01 2.43059576e-01 4.30776298e-01
3.05650085e-01 1.01961195e+00 1.90388158e-01 -1.17384887e+00
5.35553277e-01 5.42713940e-01 5.13469219e-01 -1.29420829e+00
-4.98805702e-01 -1.59109727e-01 -7.90118277e-01 1.38557240e-01
7.34158158e-01 -4.19275910e-01 -3.15538883e-01 1.74058342e+00
9.37711895e-01 1.47619471e-01 -2.16486961e-01 1.25370479e+00
3.91603380e-01 3.87954593e-01 -4.03416187e-01 -2.87305743e-01
1.74683654e+00 -9.92684245e-01 -5.92409432e-01 -5.54561377e-01
8.74900147e-02 -7.75572002e-01 4.04000044e-01 4.05485928e-01
-8.95025790e-01 -8.50628376e-01 -7.74209559e-01 1.66701481e-01
-5.36330342e-02 3.57669652e-01 3.59801292e-01 5.67976594e-01
-7.97608614e-01 1.85885832e-01 -9.37195897e-01 -7.55588949e-01
-2.83992320e-01 4.46676165e-01 -8.25973034e-01 -6.35987371e-02
-8.97064328e-01 9.30660069e-01 1.89192936e-01 2.58531004e-01
-5.90863645e-01 -2.92342126e-01 -7.99132109e-01 -3.81587535e-01
9.33322787e-01 -9.03780580e-01 1.21730351e+00 -4.18772548e-01
-1.24331439e+00 8.29331100e-01 -2.85485893e-01 -7.76656996e-03
7.70597100e-01 -8.93606603e-01 -3.67124677e-01 2.45868281e-01
2.01942086e-01 5.30657470e-01 7.17842221e-01 -1.20050752e+00
-7.58660078e-01 -8.56434703e-01 7.03020841e-02 4.60788429e-01
-3.07749659e-02 -8.44936371e-02 -1.28643167e+00 -4.62603629e-01
3.23667020e-01 -1.48946226e+00 -2.25219801e-01 3.50712717e-01
-2.71008492e-01 3.41488570e-02 6.00829124e-01 -6.12432122e-01
8.27827811e-01 -1.84969223e+00 4.44327444e-01 -6.55248240e-02
2.43080407e-01 1.46147832e-01 2.26608321e-01 3.14370692e-01
1.62629470e-01 -5.11764228e-01 4.46436524e-01 -5.40192425e-01
-1.47496790e-01 1.40339732e-01 2.08252490e-01 8.58146787e-01
-3.05350900e-01 6.38999462e-01 -7.28219867e-01 -6.52894497e-01
4.11331803e-01 7.01137304e-01 -4.75186080e-01 3.92147899e-01
3.98461610e-01 5.71310043e-01 -3.15698564e-01 6.01546526e-01
7.49821424e-01 -3.84349704e-01 4.16467130e-01 -3.79938871e-01
-4.86713462e-02 -5.52929997e-01 -1.83487201e+00 1.90921366e+00
1.25602588e-01 2.54751652e-01 4.68209088e-02 -4.55379784e-01
5.40671945e-01 4.38702285e-01 7.42375553e-01 -1.46673888e-01
1.12387404e-01 -9.15650204e-02 -2.98927307e-01 -2.76981026e-01
3.97494227e-01 1.61449969e-01 -2.53813475e-01 2.56632298e-01
3.44173871e-02 1.68901518e-01 3.17164719e-01 8.49230438e-02
7.59486854e-01 3.63281161e-01 7.07014978e-01 1.49636164e-01
5.15284717e-01 -1.78683877e-01 8.35361362e-01 5.36595106e-01
-5.02771199e-01 6.78534031e-01 1.16854705e-01 -6.67721272e-01
-9.72964585e-01 -1.23406565e+00 3.87223542e-01 7.60460615e-01
4.29290026e-01 -7.49828994e-01 -6.08932257e-01 -6.97751760e-01
1.65299952e-01 -2.34617084e-01 -4.51934040e-01 2.74844974e-01
-8.57557178e-01 -3.46340507e-01 2.15037167e-01 5.04905522e-01
4.86405015e-01 -2.62134045e-01 -1.20334530e+00 -1.14140354e-01
-4.52725440e-01 -1.53149235e+00 -1.12576473e+00 -4.31877732e-01
-7.27463663e-01 -1.63681459e+00 -9.34700191e-01 -4.40500021e-01
8.06243479e-01 7.81587660e-01 8.90425444e-01 4.73155566e-02
-3.02785337e-01 8.53823423e-01 -1.41809106e-01 -5.13471253e-02
2.49198914e-01 -4.00949568e-01 7.24993229e-01 1.27386358e-02
2.65852749e-01 -1.63745582e-01 -6.62117600e-01 8.28286886e-01
-2.21813902e-01 1.78793162e-01 3.50032508e-01 5.65236032e-01
7.19979584e-01 -9.35364664e-02 -1.98974200e-02 -4.45523798e-01
2.27875262e-02 9.38150510e-02 -9.58506286e-01 3.96738380e-01
-1.03664793e-01 -4.34826702e-01 2.98871845e-01 -6.32479727e-01
-6.42666757e-01 6.61801696e-01 2.74359077e-01 -8.04262877e-01
-2.62876928e-01 -1.19795613e-01 1.93293151e-02 7.08117150e-03
3.28682691e-01 1.20053381e-01 3.53771329e-01 -3.86061341e-01
1.14331983e-01 2.35360220e-01 7.36815631e-01 -4.49113786e-01
1.01204562e+00 5.48555315e-01 1.75823152e-01 -6.32728457e-01
-9.12605107e-01 -8.73681843e-01 -1.09726489e+00 -5.51909328e-01
9.33722973e-01 -1.37701845e+00 -1.38498104e+00 5.78657031e-01
-1.29159176e+00 2.68859923e-01 2.14660972e-01 7.02979982e-01
-4.77476627e-01 7.21207559e-01 -4.11935210e-01 -9.07870233e-01
-3.37796628e-01 -1.16362619e+00 1.33286631e+00 2.06752211e-01
-2.32613146e-01 -7.23125935e-01 2.50003040e-01 3.35827917e-01
-1.76276624e-01 4.89623308e-01 -3.85675460e-01 -2.29778752e-01
-5.21092117e-01 -4.63950306e-01 1.48053676e-01 -1.15708396e-01
1.09852105e-01 -1.89401567e-01 -3.96024555e-01 -8.16535890e-01
-2.41649952e-02 -9.21171978e-02 1.91416219e-01 4.05435711e-01
4.78872210e-01 -7.85822645e-02 -5.04774749e-01 4.25996035e-01
1.30705595e+00 4.98338304e-02 2.53805108e-02 3.01929832e-01
8.14135194e-01 4.80010450e-01 8.64100456e-01 7.74663091e-01
5.97516298e-01 1.25644076e+00 2.80214906e-01 1.64552078e-01
1.38386354e-01 -2.04348460e-01 5.73835969e-01 7.04846621e-01
-4.87242579e-01 -4.47896048e-02 -7.72723317e-01 6.58202916e-02
-2.12511587e+00 -1.13904679e+00 -1.77450702e-01 2.55160117e+00
3.13629299e-01 1.50512410e-02 6.15638673e-01 -2.80125082e-01
1.11796129e+00 -1.13020487e-01 -6.03233457e-01 6.34648204e-01
2.32553571e-01 -4.64704424e-01 5.38062334e-01 4.54232872e-01
-1.21444964e+00 5.13938606e-01 5.74359274e+00 6.50619343e-02
-6.29230499e-01 1.38809636e-01 1.15639448e-01 -6.33032620e-01
6.77195907e-01 -1.97003946e-01 -1.39483190e+00 4.74701196e-01
4.49260324e-01 4.53533791e-02 2.65906692e-01 7.24036872e-01
6.17567077e-02 -3.17499876e-01 -1.15129876e+00 1.53999400e+00
4.19545203e-01 -9.83487725e-01 -2.37917751e-01 1.53025284e-01
4.67635125e-01 -2.60573089e-01 -2.77264923e-01 2.12638266e-02
1.03484042e-01 -3.20885897e-01 9.24557328e-01 4.53433901e-01
7.52176285e-01 -7.42203414e-01 7.05673218e-01 6.28752649e-01
-1.60919511e+00 6.34612367e-02 -2.99636453e-01 -1.54180676e-01
5.15084982e-01 2.05906421e-01 -4.26472485e-01 6.32991910e-01
1.01156533e+00 8.36009324e-01 -6.72839105e-01 8.14039826e-01
7.69538339e-03 -3.08861256e-01 -5.26453614e-01 2.57919312e-01
-3.16262364e-01 -2.33466133e-01 4.76834238e-01 8.69407952e-01
5.57336152e-01 1.47353277e-01 7.25433588e-01 1.26125962e-01
4.20772493e-01 -3.69234294e-01 -5.80352366e-01 5.11460006e-01
4.73717988e-01 1.38194931e+00 -6.29161000e-01 -4.32006449e-01
-5.20095944e-01 1.19031239e+00 1.77451894e-01 1.18657202e-01
-1.35564554e+00 5.51879779e-02 8.26222181e-01 1.08916342e-01
4.43103164e-01 -7.28179455e-01 4.38630998e-01 -1.73827457e+00
3.74871701e-01 -9.64708388e-01 7.80560434e-01 -6.86340034e-01
-1.13994300e+00 5.24848402e-01 3.22009385e-01 -1.69837058e+00
-4.28036243e-01 -5.96653461e-01 -6.28074035e-02 5.47980845e-01
-8.47082496e-01 -1.22656989e+00 -5.76693356e-01 9.82432246e-01
5.74663103e-01 -8.12277347e-02 6.52225137e-01 3.78874451e-01
-6.28769994e-01 5.07296920e-01 -2.01542646e-01 2.69157052e-01
1.14250195e+00 -9.99347746e-01 5.04232764e-01 1.03549910e+00
1.98414803e-01 8.18538487e-01 7.86008239e-01 -7.36507297e-01
-1.89013028e+00 -7.57579625e-01 6.64920509e-01 -9.73187625e-01
2.57848859e-01 -5.97657263e-01 -3.27796876e-01 1.06260586e+00
-4.98643145e-02 3.37694794e-01 5.16753137e-01 3.70109640e-02
-2.54405618e-01 -1.84426889e-01 -1.05954349e+00 3.94756585e-01
1.18678725e+00 -8.60548764e-02 -4.15881425e-01 3.47780138e-01
4.82194930e-01 -9.45855081e-01 -8.10065508e-01 5.92802390e-02
8.80433023e-01 -1.15553486e+00 1.35286248e+00 -2.82475471e-01
-3.12244445e-01 -7.51881897e-01 -2.88937390e-01 -9.75798249e-01
-3.21156979e-01 -6.00296915e-01 -2.61695385e-01 9.18287337e-01
-3.71644139e-01 -7.11084306e-01 7.79942751e-01 5.86142898e-01
3.48082662e-01 -4.37141567e-01 -1.03679657e+00 -9.29220200e-01
-6.42782927e-01 -1.85727179e-01 2.37615034e-01 5.01215160e-01
-2.27095202e-01 3.29202116e-01 -9.46567953e-01 6.49823546e-01
1.12453902e+00 2.77197272e-01 1.50181890e+00 -1.29666805e+00
-4.95288551e-01 2.41589978e-01 -5.25071383e-01 -1.30188286e+00
-2.05486968e-01 -2.04603687e-01 -6.68544620e-02 -1.02368879e+00
5.91673732e-01 9.84657854e-02 2.19119459e-01 2.59944171e-01
-3.10560077e-01 3.47282261e-01 7.35813081e-01 3.50640476e-01
-1.10499847e+00 1.58248633e-01 1.06988966e+00 1.75003290e-01
8.60332176e-02 1.47136524e-01 -3.08724761e-01 9.00267482e-01
4.91038263e-01 -4.43711251e-01 -1.56664774e-01 -4.78322238e-01
-3.65191735e-02 4.71611887e-01 8.60770583e-01 -1.34978426e+00
5.02511621e-01 -2.04829782e-01 8.37363660e-01 -9.51600194e-01
8.40801299e-01 -1.17752016e+00 8.19599748e-01 7.80008197e-01
2.50490010e-01 7.51483560e-01 1.72911212e-02 7.47889876e-01
1.01400644e-01 2.36913994e-01 7.11264670e-01 -4.63325024e-01
-8.25023890e-01 6.46351755e-01 6.85406923e-02 -1.28323706e-02
1.44116735e+00 -2.89113998e-01 -9.32953134e-02 -4.35372531e-01
-7.92574525e-01 4.19364691e-01 7.65966356e-01 5.42316854e-01
6.81184947e-01 -1.51120949e+00 -4.90660876e-01 3.20793539e-01
-1.21965092e-02 -3.62696382e-03 5.67773104e-01 8.79085541e-01
-4.62557524e-01 4.72343832e-01 -4.10267651e-01 -1.07462239e+00
-1.89720333e+00 6.65711224e-01 3.18672627e-01 -2.00642198e-01
-6.88565850e-01 3.96856308e-01 7.63784200e-02 -4.00270224e-01
1.16558939e-01 6.51429594e-02 1.12844214e-01 -4.78810556e-02
6.80404007e-01 5.87355494e-01 -3.90539348e-01 -1.03525686e+00
-7.37129629e-01 1.17326927e+00 6.87347203e-02 -2.07200140e-01
1.16924894e+00 -5.33359826e-01 1.76309615e-01 4.80601758e-01
1.06663585e+00 1.04532309e-01 -1.68474674e+00 -1.71568081e-01
-4.53388959e-01 -9.36231017e-01 -6.86887085e-01 -3.01804900e-01
-1.14847612e+00 5.14340937e-01 8.24236691e-01 -2.21274674e-01
8.10135484e-01 1.82217881e-02 7.55227804e-01 5.67346692e-01
8.47913682e-01 -8.77197027e-01 3.19034457e-01 3.81757945e-01
8.02990615e-01 -1.37792337e+00 4.68114704e-01 -4.63224739e-01
-6.70834720e-01 1.11037755e+00 1.05547726e+00 -1.06527500e-01
3.61162484e-01 1.11501411e-01 1.45400852e-01 -1.14346512e-01
-5.71663201e-01 8.18718411e-03 1.89193353e-01 5.54276705e-01
1.87953860e-01 -3.81329618e-02 1.44810051e-01 2.10807636e-01
7.28406534e-02 -8.92595574e-02 4.28817332e-01 1.04519904e+00
-2.04598784e-01 -9.90465283e-01 -9.47992444e-01 -1.11478820e-01
-2.37308800e-01 6.14003003e-01 -1.22646578e-01 1.04670358e+00
1.79044932e-01 8.62285614e-01 4.68431562e-02 -3.59276593e-01
7.06032336e-01 -6.68327436e-02 7.00942218e-01 -4.60602283e-01
-4.62518632e-01 3.75579238e-01 -2.18015775e-01 -8.68054986e-01
-6.71798825e-01 -1.06824386e+00 -1.03571129e+00 -6.04227781e-01
-1.91923857e-01 -9.39753652e-02 3.18214089e-01 6.49777532e-01
5.35027683e-01 6.95657358e-02 3.87089670e-01 -1.23219597e+00
-6.25339806e-01 -6.86593533e-01 -4.85394090e-01 6.98883772e-01
4.33266491e-01 -1.07504857e+00 -1.19271472e-01 2.93250889e-01] | [7.081691741943359, -1.0655995607376099] |
8c0965dc-b8fb-4393-a206-a0ecd1f4467e | augmenting-deep-learning-adaptation-for | 2307.00883 | null | https://arxiv.org/abs/2307.00883v1 | https://arxiv.org/pdf/2307.00883v1.pdf | Augmenting Deep Learning Adaptation for Wearable Sensor Data through Combined Temporal-Frequency Image Encoding | Deep learning advancements have revolutionized scalable classification in many domains including computer vision. However, when it comes to wearable-based classification and domain adaptation, existing computer vision-based deep learning architectures and pretrained models trained on thousands of labeled images for months fall short. This is primarily because wearable sensor data necessitates sensor-specific preprocessing, architectural modification, and extensive data collection. To overcome these challenges, researchers have proposed encoding of wearable temporal sensor data in images using recurrent plots. In this paper, we present a novel modified-recurrent plot-based image representation that seamlessly integrates both temporal and frequency domain information. Our approach incorporates an efficient Fourier transform-based frequency domain angular difference estimation scheme in conjunction with the existing temporal recurrent plot image. Furthermore, we employ mixup image augmentation to enhance the representation. We evaluate the proposed method using accelerometer-based activity recognition data and a pretrained ResNet model, and demonstrate its superior performance compared to existing approaches. | ['Mohammad Arif Ul Alam', 'Md Mahmudur Rahman', 'Yidong Zhu'] | 2023-07-03 | null | null | null | null | ['activity-recognition', 'image-augmentation'] | ['computer-vision', 'computer-vision'] | [ 5.87725699e-01 -3.28053921e-01 -3.22363466e-01 -5.42470753e-01
-6.05137229e-01 -1.94351315e-01 3.87455851e-01 3.13934356e-01
-6.76196575e-01 5.71170151e-01 5.02599239e-01 7.70094842e-02
-3.93467285e-02 -6.89235985e-01 -7.02687919e-01 -4.72899765e-01
-3.26382101e-01 -3.83189231e-01 2.43759770e-02 -1.71169356e-01
-1.57682493e-01 2.41692618e-01 -1.32567632e+00 1.33652344e-01
7.69552052e-01 1.32771552e+00 -1.68634593e-01 6.96926117e-01
2.76044369e-01 6.21427000e-01 -8.27278793e-01 6.92211688e-02
1.33598059e-01 -3.99469256e-01 -2.63670534e-01 -3.79712842e-02
4.36082631e-01 -4.68636274e-01 -4.54871535e-01 4.99914140e-01
6.58906758e-01 4.24296767e-01 3.47130150e-01 -1.02688348e+00
-6.75737321e-01 2.71018326e-01 -5.69757044e-01 6.72276199e-01
4.40557361e-01 -2.67554279e-02 4.23126012e-01 -6.06591940e-01
1.29098713e-01 5.51966012e-01 1.34041500e+00 2.52631545e-01
-1.11812794e+00 -6.30274296e-01 7.91620091e-02 3.72980356e-01
-1.34489667e+00 -4.27225195e-02 1.27513373e+00 -2.71130115e-01
1.33136809e+00 8.59313384e-02 1.19925797e+00 1.56616819e+00
2.98607171e-01 3.56416255e-01 8.64975631e-01 -2.19866112e-01
3.60064298e-01 -3.70903850e-01 9.96612385e-02 3.86206508e-01
3.92815441e-01 -5.71604669e-02 -6.68203831e-01 6.52160570e-02
8.83328199e-01 6.28303349e-01 2.77505666e-02 -1.91998646e-01
-1.16095209e+00 4.29372102e-01 7.10401237e-01 4.07903910e-01
-6.96111619e-01 5.03165007e-01 5.39466918e-01 2.59806607e-02
5.60552537e-01 1.53776005e-01 -3.81813377e-01 -5.15929520e-01
-1.11393976e+00 1.92816760e-02 3.35052580e-01 4.48902130e-01
5.09975195e-01 4.88358051e-01 1.24901056e-01 9.72060859e-01
1.27520099e-01 5.92879236e-01 1.07410657e+00 -5.59371173e-01
5.10557234e-01 7.94532359e-01 -1.13930665e-01 -1.31719911e+00
-9.12859440e-01 -6.93904459e-01 -1.18075943e+00 -3.90247732e-01
-2.73915082e-02 -2.59795576e-01 -9.75028455e-01 1.52664578e+00
3.36784929e-01 6.44024849e-01 -1.79320857e-01 9.55991089e-01
6.98345244e-01 4.62990850e-01 2.02977821e-01 -5.70477657e-02
1.12738907e+00 -7.60966301e-01 -7.71018028e-01 -2.74992198e-01
4.05668259e-01 -3.05015922e-01 1.23224103e+00 2.89978057e-01
-5.12870371e-01 -9.31318521e-01 -1.73166978e+00 -8.75849351e-02
-4.47523355e-01 3.08096826e-01 6.12355649e-01 1.02011108e+00
-7.19960988e-01 6.05386317e-01 -1.40782702e+00 -6.23399973e-01
4.47496414e-01 4.79474634e-01 -1.90148875e-01 1.37632966e-01
-9.88831222e-01 5.46295047e-01 2.04732299e-01 2.18280062e-01
-4.56624359e-01 -5.18423021e-01 -1.05192411e+00 -1.37367710e-01
-1.35206059e-01 -4.75550532e-01 1.14142978e+00 -8.37357044e-01
-1.54315245e+00 2.89638519e-01 -5.01228124e-02 -9.55649376e-01
1.86305404e-01 -5.50768912e-01 -8.99997950e-01 2.44993493e-01
-5.43713272e-02 2.67475545e-01 9.52651918e-01 -4.25455183e-01
-2.43430808e-01 -4.13740784e-01 -6.77286685e-02 2.89392080e-02
-9.97020006e-01 -4.21969503e-01 -4.61754296e-03 -1.07264245e+00
1.78588435e-01 -8.71792614e-01 -2.55658299e-01 -2.60428637e-01
6.15061931e-02 6.13447428e-02 9.07918930e-01 -7.05563784e-01
1.25528920e+00 -2.18401122e+00 -4.08764690e-01 1.18665524e-01
1.49751097e-01 2.62914479e-01 -2.17915867e-02 2.19365165e-01
-2.14023069e-01 -2.15041533e-01 -3.99414450e-02 -3.20520967e-01
-3.29502016e-01 3.25728059e-01 -8.29879493e-02 6.63017333e-01
3.40196699e-01 8.31746161e-01 -6.72035873e-01 -2.26914167e-01
5.12812793e-01 8.97885859e-01 -6.27705932e-01 4.06663567e-02
1.22315310e-01 6.76563740e-01 -2.67439574e-01 5.74573576e-01
5.76219797e-01 -3.92487437e-01 1.16724052e-01 -7.34982312e-01
2.99113635e-02 3.84941190e-01 -9.71376419e-01 2.15813327e+00
-5.48353016e-01 6.64000452e-01 -6.48876429e-01 -1.61927569e+00
9.76883590e-01 2.37978503e-01 1.06054509e+00 -1.01157320e+00
2.63225645e-01 -7.05513582e-02 -3.69872004e-01 -6.79944813e-01
5.68596423e-01 1.40193060e-01 -2.34734282e-01 2.57302612e-01
1.92320362e-01 3.81279707e-01 -1.68569446e-01 -3.30188304e-01
1.34580660e+00 5.45430899e-01 3.07503104e-01 2.01760709e-01
3.68637413e-01 8.35509002e-02 6.01582825e-01 4.93973821e-01
-2.82819599e-01 6.39608204e-01 -4.06811684e-01 -8.81009340e-01
-9.01857793e-01 -9.98238385e-01 1.22449599e-01 1.10958564e+00
1.83783054e-01 -6.53829753e-01 -5.17707527e-01 -3.82034987e-01
-2.25202993e-01 8.71193875e-03 -7.28099763e-01 -3.47688317e-01
-9.29721296e-01 -1.01651776e+00 8.11846554e-01 1.04358804e+00
1.15954614e+00 -3.79721135e-01 -1.09856939e+00 5.05175650e-01
-3.01310927e-01 -1.19845057e+00 -3.95417005e-01 3.03048611e-01
-1.12699449e+00 -1.01427400e+00 -7.07892418e-01 -6.04138672e-01
3.66450876e-01 4.15768772e-01 6.21556938e-01 -2.97788292e-01
-4.49710459e-01 5.69182873e-01 -4.18816566e-01 -4.31343794e-01
3.60659868e-01 2.64244586e-01 2.32099339e-01 1.34547338e-01
4.47948605e-01 -1.14884138e+00 -9.87869203e-01 1.43615186e-01
-7.11761773e-01 -1.69457555e-01 5.09305894e-01 7.66722679e-01
6.18296683e-01 -2.13930368e-01 6.16650164e-01 -2.08432332e-01
6.32637739e-01 -4.06534880e-01 -1.02784984e-01 -9.18117631e-03
-5.74134111e-01 -1.78778137e-03 5.34382343e-01 -9.65767860e-01
-8.69779110e-01 2.13007390e-01 -3.25095147e-01 -3.32364500e-01
-4.50992733e-02 8.06803405e-01 2.77933210e-01 -1.19477272e-01
1.08711088e+00 3.03301126e-01 -1.50163427e-01 -5.73589325e-01
1.16281770e-01 6.22150719e-01 9.21785533e-01 -3.21382850e-01
5.56522191e-01 5.85654080e-01 -1.41192839e-01 -1.15285218e+00
-5.62826097e-01 -4.47343022e-01 -5.89471877e-01 -3.60587746e-01
8.87349904e-01 -1.22864592e+00 -4.96759057e-01 4.28345710e-01
-8.61759007e-01 -2.91460723e-01 -5.17836392e-01 8.66418064e-01
-3.80486965e-01 2.02727720e-01 -4.35658127e-01 -7.71188498e-01
-4.61492300e-01 -5.69709420e-01 1.13332713e+00 3.19829553e-01
-3.95736545e-01 -7.93573201e-01 3.47509027e-01 3.31076264e-01
6.46770000e-01 8.85096371e-01 1.78682134e-01 -3.48207027e-01
-1.20571174e-01 -4.90016073e-01 4.17639725e-02 3.96937758e-01
3.79725605e-01 -5.16260266e-01 -9.15324688e-01 -3.06792110e-01
8.68997201e-02 -2.63413191e-01 6.76643133e-01 3.39469880e-01
1.22115338e+00 -2.92726129e-01 -2.45784327e-01 9.09526348e-01
1.14787948e+00 3.45201641e-01 5.01624644e-01 5.54637551e-01
9.10873413e-01 -1.97794393e-01 2.87862271e-01 5.63578069e-01
6.06431186e-01 6.57822847e-01 3.75548489e-02 -3.34420592e-01
-2.00004235e-01 -2.54575223e-01 3.56183439e-01 1.13610947e+00
-3.73751193e-01 1.22741364e-01 -8.21760893e-01 4.58005786e-01
-1.82483613e+00 -9.68318224e-01 2.99273759e-01 1.87111056e+00
5.06171763e-01 -3.31357983e-03 3.85657936e-01 5.56034029e-01
3.60028654e-01 2.98357993e-01 -9.00486946e-01 -1.53798178e-01
-6.64717555e-02 5.29019058e-01 6.60837471e-01 -2.00480357e-01
-1.34719181e+00 1.93389073e-01 6.58031702e+00 2.49149516e-01
-1.63765395e+00 3.90489668e-01 2.09325895e-01 -2.65433908e-01
1.40730932e-01 -6.09992266e-01 -2.62720197e-01 4.38842446e-01
1.33527327e+00 6.14501052e-02 2.52201170e-01 8.70851219e-01
4.21357065e-01 3.14577632e-02 -8.01221430e-01 1.55960178e+00
2.68504053e-01 -1.24528015e+00 -3.23096693e-01 4.04729284e-02
5.27466357e-01 2.20114633e-01 1.09917305e-01 2.83699602e-01
-2.42468670e-01 -8.29135239e-01 4.95406717e-01 3.49646211e-01
7.30426788e-01 -5.55782974e-01 5.20395815e-01 1.96633842e-02
-1.70544660e+00 -2.31204107e-01 2.47787293e-02 -4.70903069e-01
-1.63397729e-01 5.43412626e-01 -8.48959923e-01 6.50807798e-01
1.04352200e+00 1.15335202e+00 -7.20245659e-01 8.50083113e-01
1.40857935e-01 9.67012703e-01 -6.66972160e-01 -2.61718333e-02
-5.89853898e-03 1.29751787e-01 2.55397875e-02 1.26575673e+00
5.48191667e-01 -6.76872656e-02 2.33536705e-01 3.03503960e-01
4.45727892e-02 -5.71362190e-02 -6.81033909e-01 -4.83034402e-02
2.16832265e-01 9.79892075e-01 -6.20152295e-01 -3.54398012e-01
-6.04633510e-01 1.08828902e+00 1.83762182e-02 2.92332768e-01
-1.11354125e+00 -3.96641225e-01 6.43590391e-01 1.70547739e-01
1.33518323e-01 -8.29809010e-01 -4.22535419e-01 -1.39322186e+00
2.54155368e-01 -5.15073240e-01 4.28844839e-01 -5.82174718e-01
-9.74321544e-01 4.28746969e-01 4.00122218e-02 -1.53693926e+00
-5.67346334e-01 -2.15557694e-01 -4.52643633e-01 3.42962891e-01
-1.21285248e+00 -1.36582410e+00 -8.94382298e-01 8.72610390e-01
4.96754408e-01 2.19275635e-02 8.27231884e-01 8.24258327e-01
-6.78476155e-01 7.14274108e-01 -3.12585123e-02 3.00875425e-01
4.89585072e-01 -9.32043552e-01 6.23199701e-01 1.01288712e+00
2.36982092e-01 6.41170800e-01 5.16463757e-01 -5.48837900e-01
-1.45488822e+00 -1.46717072e+00 3.37692857e-01 -1.63630828e-01
5.45514047e-01 -3.25336665e-01 -7.45938241e-01 7.48755217e-01
9.90115032e-02 3.56557578e-01 1.06213462e+00 5.16638272e-02
-2.56271005e-01 -6.20630920e-01 -1.09772456e+00 3.76443982e-01
1.24370098e+00 -7.01441944e-01 -6.43311441e-01 5.94665259e-02
5.27625024e-01 -4.42398548e-01 -1.09702611e+00 5.78372300e-01
9.15224135e-01 -5.52679121e-01 1.18413579e+00 -1.79589987e-01
-4.73168679e-02 -2.94066221e-01 -2.17363656e-01 -1.13387382e+00
-1.88663498e-01 -4.86285537e-01 -3.75225425e-01 7.95200109e-01
-7.25767091e-02 -6.22998536e-01 1.02256143e+00 3.40639472e-01
-1.83534205e-01 -4.83249873e-01 -1.07938671e+00 -8.75699103e-01
-5.21419823e-01 -7.98748791e-01 6.99543297e-01 9.75522816e-01
6.14850782e-02 4.29193467e-01 -6.28990829e-01 2.18419850e-01
5.10300457e-01 -1.53683186e-01 7.49717176e-01 -1.05123508e+00
-1.88604712e-01 8.91671255e-02 -8.37619007e-01 -1.10035932e+00
-4.96192366e-01 -4.95932341e-01 -1.34383008e-01 -1.37518072e+00
-2.76470661e-01 -8.29703137e-02 -7.37366498e-01 6.79658115e-01
1.42082438e-01 8.13116550e-01 -3.60769145e-02 2.77786721e-02
-5.98923743e-01 6.69606388e-01 6.67185605e-01 -4.43137407e-01
-5.15555978e-01 -2.89427251e-01 -3.89137417e-01 6.79974377e-01
1.05902636e+00 -2.21711352e-01 -6.71611369e-01 -4.69051629e-01
1.46018565e-01 -2.06509456e-01 6.40654027e-01 -1.86455286e+00
2.56163388e-01 8.97205546e-02 9.69667614e-01 -6.09757185e-01
6.16488457e-01 -8.60117197e-01 2.71464974e-01 5.63597381e-01
-1.85778186e-01 3.91452700e-01 4.12213683e-01 7.87897050e-01
-1.20810449e-01 5.77410936e-01 5.29046714e-01 1.54564902e-01
-6.98549271e-01 1.94732383e-01 -4.43756580e-01 -2.84878343e-01
7.57965505e-01 -6.34027779e-01 -1.90867856e-01 -2.87395447e-01
-8.17636430e-01 -2.36628279e-01 4.63248342e-02 6.44337177e-01
7.68887758e-01 -1.69413376e+00 -7.55940676e-02 3.88987184e-01
2.38248929e-01 -1.37346551e-01 3.51653486e-01 8.75506341e-01
-3.93096209e-01 2.41510555e-01 -4.90123183e-01 -7.87707150e-01
-1.00727022e+00 1.41502038e-01 3.41606855e-01 -1.47795126e-01
-8.99461031e-01 3.75619143e-01 -4.90362048e-01 -1.03239238e-01
1.83285341e-01 -9.85351026e-01 -1.67774916e-01 2.80230846e-02
5.49300730e-01 3.82771969e-01 3.79285902e-01 -3.96657467e-01
-6.00096881e-01 6.63784742e-01 2.23191261e-01 -6.09824844e-02
1.52944231e+00 -7.58584365e-02 5.87809265e-01 6.31725371e-01
1.38035297e+00 -4.40289408e-01 -1.26933539e+00 -2.34316170e-01
1.07002109e-01 -2.04154655e-01 9.83981714e-02 -6.98191464e-01
-1.03566003e+00 8.24643135e-01 1.36707413e+00 1.02671526e-01
1.45709419e+00 -5.71111679e-01 1.14797211e+00 5.38758159e-01
3.33041459e-01 -1.25677860e+00 5.88536263e-01 2.60289043e-01
5.51968753e-01 -1.16137993e+00 1.81692004e-01 3.10252924e-02
-1.31797433e-01 9.12601769e-01 5.32869935e-01 -4.08178657e-01
7.64509976e-01 1.18388325e-01 1.81722715e-02 1.33410597e-03
-3.22133362e-01 -8.24613050e-02 2.82066941e-01 1.04506397e+00
5.62045157e-01 -1.77714422e-01 -3.58772874e-01 6.89031839e-01
-3.37817311e-01 4.46736097e-01 1.76988915e-01 1.38106191e+00
-2.07914069e-01 -8.63646328e-01 -3.94662291e-01 5.26446521e-01
-4.38913077e-01 2.34741732e-01 4.84173559e-02 5.56664467e-01
3.49457532e-01 9.42723751e-01 2.47837067e-01 -9.19570446e-01
4.94822264e-01 -6.28665835e-03 5.01919091e-01 -3.30090880e-01
-5.81250489e-01 8.17645043e-02 -1.93576459e-02 -7.32671678e-01
-8.43768418e-01 -4.13843989e-01 -1.17074585e+00 -1.88371185e-02
1.35824189e-01 -1.72380134e-01 1.03900373e+00 9.37553823e-01
6.44158542e-01 8.82794321e-01 5.18498719e-01 -1.20973003e+00
-1.56576738e-01 -1.15958703e+00 -2.78622150e-01 3.29313785e-01
7.61008918e-01 -8.07790995e-01 2.24728540e-01 3.51240277e-01] | [7.578712463378906, 0.8140469789505005] |
9fbaebf0-2f23-4d5e-a855-36fba2cea208 | recognition-of-instrument-tissue-interactions | 2007.05405 | null | https://arxiv.org/abs/2007.05405v1 | https://arxiv.org/pdf/2007.05405v1.pdf | Recognition of Instrument-Tissue Interactions in Endoscopic Videos via Action Triplets | Recognition of surgical activity is an essential component to develop context-aware decision support for the operating room. In this work, we tackle the recognition of fine-grained activities, modeled as action triplets <instrument, verb, target> representing the tool activity. To this end, we introduce a new laparoscopic dataset, CholecT40, consisting of 40 videos from the public dataset Cholec80 in which all frames have been annotated using 128 triplet classes. Furthermore, we present an approach to recognize these triplets directly from the video data. It relies on a module called Class Activation Guide (CAG), which uses the instrument activation maps to guide the verb and target recognition. To model the recognition of multiple triplets in the same frame, we also propose a trainable 3D Interaction Space, which captures the associations between the triplet components. Finally, we demonstrate the significance of these contributions via several ablation studies and comparisons to baselines on CholecT40. | ['Cristians Gonzalez', 'Tong Yu', 'Didier Mutter', 'Pietro Mascagni', 'Nicolas Padoy', 'Jacques Marescaux', 'Chinedu Innocent Nwoye'] | 2020-07-10 | null | null | null | null | ['action-triplet-recognition', 'weakly-supervised-action-localization'] | ['computer-vision', 'computer-vision'] | [ 3.87019008e-01 1.87908307e-01 -3.91907126e-01 -3.15329820e-01
-6.79624617e-01 -8.02159190e-01 6.08100474e-01 8.55159089e-02
-3.44834417e-01 2.20998451e-01 6.93216026e-01 -1.50919005e-01
-2.10627675e-01 -3.80173355e-01 -7.81538069e-01 -5.82806826e-01
-1.43286794e-01 1.86140165e-01 -6.80768713e-02 -8.53538290e-02
3.19337815e-01 4.71634865e-01 -1.47140360e+00 9.87422049e-01
3.45766962e-01 1.08797348e+00 1.87892001e-03 7.43003786e-01
1.53163344e-01 1.14908373e+00 -5.45146942e-01 1.37533978e-01
1.91565707e-01 -5.09499013e-01 -9.99907017e-01 1.53060006e-02
5.57699263e-01 -2.85153538e-01 -1.82402030e-01 4.59397793e-01
3.03038538e-01 2.66745090e-01 8.41712892e-01 -9.20486212e-01
1.96081474e-01 5.33943713e-01 -1.45911932e-01 2.92182475e-01
8.13855410e-01 1.57802805e-01 5.08081436e-01 -8.37385356e-01
1.14249980e+00 1.03816676e+00 3.10376585e-01 5.87063432e-01
-8.70625138e-01 -6.28019273e-01 1.77012831e-01 2.28214681e-01
-1.01421404e+00 -3.51130188e-01 6.37934506e-01 -8.34889829e-01
9.17127371e-01 4.39596176e-01 1.44365513e+00 1.71781600e+00
5.25242925e-01 1.12722540e+00 7.99600005e-01 -4.87779140e-01
2.24903405e-01 -3.82105768e-01 1.03814542e-01 1.00887513e+00
2.18228586e-02 3.24521303e-01 -9.30574775e-01 8.61861929e-02
9.88879561e-01 1.47450686e-01 -2.31169775e-01 -1.03833795e+00
-1.87530422e+00 5.94051480e-01 3.15044314e-01 3.56612831e-01
-4.36976463e-01 3.03188413e-01 4.42904890e-01 -1.46272518e-02
-1.65603012e-01 6.55080974e-01 -8.13808441e-02 -6.40283048e-01
-4.89204139e-01 -1.50477916e-01 9.64033902e-01 1.22975051e+00
5.58258057e-01 -4.81659204e-01 -8.00234497e-01 3.55361849e-01
8.06929469e-02 -1.72445461e-01 5.22030413e-01 -1.24182320e+00
2.67973006e-01 9.52920794e-01 4.29252014e-02 -6.28049910e-01
-7.48110652e-01 -1.96494743e-01 -3.29961419e-01 -2.14157738e-02
1.23285554e-01 1.15519978e-01 -1.12326050e+00 1.34535480e+00
3.22090328e-01 4.94757414e-01 3.13128717e-02 6.41830027e-01
1.07081068e+00 5.83438985e-02 1.77373752e-01 -9.93525609e-02
1.39152515e+00 -1.33367240e+00 -7.72016108e-01 -3.03229392e-01
1.09001684e+00 -5.87387800e-01 8.29203129e-01 5.74171782e-01
-7.70579398e-01 -4.93526489e-01 -9.20702696e-01 -7.79115781e-02
-3.39808404e-01 6.79602504e-01 8.00619543e-01 2.59530574e-01
-7.78940022e-01 3.58144820e-01 -1.45765507e+00 -5.11515617e-01
2.50103295e-01 4.52488720e-01 -7.02737629e-01 4.75132465e-02
-5.72647810e-01 9.48758364e-01 5.04442632e-01 3.33111554e-01
-1.48641169e+00 -4.52040434e-01 -1.27940357e+00 -2.60231346e-01
5.62315941e-01 -7.70063043e-01 1.35286760e+00 -8.47042203e-01
-1.55650640e+00 1.26767588e+00 9.39701945e-02 -3.49134743e-01
5.45510948e-01 -5.27091682e-01 -1.94657370e-01 3.58687311e-01
-7.48211741e-02 6.14252210e-01 7.67277837e-01 -8.56303036e-01
-6.72795951e-01 -1.84725523e-01 5.26481330e-01 3.87760013e-01
-1.59863919e-01 -1.00284174e-01 -8.88352454e-01 -8.79241645e-01
-1.93364635e-01 -1.44534826e+00 -1.10445969e-01 6.64489642e-02
-4.72324938e-01 -8.65394026e-02 4.72937703e-01 -4.26187098e-01
1.25628471e+00 -2.47829604e+00 6.43675387e-01 1.19766921e-01
1.48759991e-01 -1.17719144e-01 -1.70130789e-01 2.39697143e-01
-2.65068680e-01 -3.05556506e-01 -2.94189323e-02 -3.10217023e-01
-3.36884290e-01 5.57643175e-01 2.67622024e-01 5.15358865e-01
-4.27972190e-02 8.55321944e-01 -1.10373390e+00 -5.79429090e-01
7.03170955e-01 3.57111335e-01 -8.25137079e-01 4.21663463e-01
7.84870535e-02 8.67942214e-01 -5.30212224e-01 8.51427972e-01
-8.11195001e-02 -2.14442052e-02 4.14112389e-01 -6.56715214e-01
-1.01950154e-01 -5.39553426e-02 -8.62112105e-01 2.75350976e+00
-5.24500430e-01 3.66003871e-01 -1.50297090e-01 -8.72901857e-01
3.85008335e-01 3.42353731e-01 9.41816628e-01 -3.18865001e-01
4.47191924e-01 1.37365103e-01 1.90869961e-02 -7.60752261e-01
2.77739286e-01 4.04783249e-01 -4.27400529e-01 9.75456461e-02
4.67770487e-01 -1.05301522e-01 4.01390165e-01 1.09864198e-01
1.45861995e+00 5.82760870e-01 7.16322243e-01 -2.67795384e-01
3.87496233e-01 3.08148652e-01 2.49746293e-01 8.47725272e-01
-2.41815224e-01 5.56061327e-01 5.58266222e-01 -6.76250815e-01
-2.75936693e-01 -8.78903389e-01 2.07372174e-01 8.02643538e-01
3.86968255e-01 -6.58304691e-01 -6.54961824e-01 -1.10246706e+00
-2.24622101e-01 3.22546661e-01 -1.41836417e+00 -5.60503781e-01
-9.40933287e-01 -2.65465915e-01 1.40275419e-01 7.98922777e-01
1.43747136e-01 -1.26077139e+00 -1.20114136e+00 9.47364271e-02
-4.78260785e-01 -1.32079816e+00 -6.41559958e-01 5.04431784e-01
-7.86075950e-01 -1.57556891e+00 -3.57750416e-01 -8.79175961e-01
8.05140793e-01 1.41936272e-01 1.18866181e+00 9.54023376e-03
-6.66152298e-01 9.69877958e-01 -6.15096748e-01 -3.66847575e-01
-3.95986915e-01 1.29911855e-01 -7.27256685e-02 5.18309213e-02
7.00695813e-02 -8.43249038e-02 -9.57344711e-01 4.65086639e-01
-6.63455307e-01 4.87754613e-01 9.05897021e-01 7.19540834e-01
7.35436082e-01 -8.29550564e-01 -3.41368347e-01 -8.19887578e-01
1.01568364e-01 -3.43912005e-01 -2.43286952e-01 3.41962695e-01
2.31668763e-02 1.00102350e-02 9.20673832e-03 -6.00386918e-01
-8.67093205e-01 5.80567658e-01 1.32035106e-01 -6.75801933e-01
-1.13564789e-01 5.90656459e-01 1.30453736e-01 -2.98730433e-01
7.11073160e-01 -2.37748735e-02 5.92229050e-03 -3.76423091e-01
1.74225494e-01 1.95778653e-01 8.13045502e-01 -6.07318938e-01
2.63109833e-01 5.55612564e-01 -2.10093055e-02 -3.82732093e-01
-9.43816364e-01 -8.25099349e-01 -8.33109915e-01 -4.60302860e-01
1.12110746e+00 -9.26617026e-01 -7.26771355e-01 1.80557013e-01
-9.46116984e-01 -6.14776611e-01 -3.58659983e-01 8.53542924e-01
-1.12921286e+00 6.32388815e-02 -4.79323328e-01 -2.26441607e-01
-7.97616169e-02 -1.42177308e+00 1.48128390e+00 6.12595975e-02
-5.23301423e-01 -8.03520799e-01 1.84436738e-01 3.89657557e-01
-8.65406543e-02 8.40624988e-01 4.41537261e-01 -6.38606310e-01
-5.86907566e-01 -3.88563037e-01 3.91174436e-01 1.37161404e-01
4.24435407e-01 -9.82445329e-02 -4.81563389e-01 -1.99788764e-01
-2.20822796e-01 -5.80539405e-02 5.14199555e-01 4.44352806e-01
1.48384011e+00 3.13443206e-02 -7.46714830e-01 9.43779111e-01
8.93345237e-01 4.20254678e-01 5.82608998e-01 6.11821771e-01
7.23701835e-01 3.86917979e-01 1.22326005e+00 1.88034475e-01
9.52414796e-02 1.02646935e+00 6.29837215e-01 -1.11561976e-01
-2.54117250e-01 -5.62836090e-03 4.12881732e-01 3.58830929e-01
-6.32303953e-01 5.29478490e-02 -9.97264504e-01 4.08874303e-01
-1.89468813e+00 -8.83130908e-01 2.28018939e-01 1.87545121e+00
6.10983253e-01 -5.03456257e-02 -1.07614540e-01 -1.87276512e-01
1.20537832e-01 -2.84835212e-02 -2.58882761e-01 -8.58471170e-02
3.45830262e-01 1.26859263e-01 2.63815314e-01 1.85479805e-01
-1.44765997e+00 7.68242538e-01 6.75399780e+00 4.47506398e-01
-1.00987804e+00 -7.16924369e-02 1.05444066e-01 -3.63623410e-01
3.74175161e-01 -2.21636355e-01 -5.36926031e-01 2.41915405e-01
7.01942146e-01 1.39437556e-01 2.45158955e-01 8.72943819e-01
3.90506208e-01 -3.98069441e-01 -1.67876923e+00 1.09572291e+00
3.95232916e-01 -1.56124508e+00 1.48246944e-01 2.86489166e-02
4.21393424e-01 -2.59496987e-01 -2.62242734e-01 5.20246565e-01
9.94587466e-02 -8.70487154e-01 8.50089908e-01 6.26466870e-01
9.30869043e-01 -2.19533905e-01 4.70148087e-01 1.78228945e-01
-1.30678427e+00 -3.81363064e-01 4.62363511e-01 3.56984258e-01
-8.82760063e-02 -1.55172139e-01 -1.04603350e+00 7.83091784e-01
8.47239077e-01 1.20807028e+00 -5.22238076e-01 1.06772423e+00
-3.92687887e-01 2.09186018e-01 -2.13630721e-01 4.52292562e-01
2.10446760e-01 -8.68853275e-03 5.46775341e-01 1.35576689e+00
4.47425485e-01 1.42840728e-01 3.98630232e-01 8.99186879e-02
6.62113205e-02 -6.02778308e-02 -6.39824748e-01 -2.70527322e-02
-6.98925182e-03 1.33437014e+00 -8.76169443e-01 -2.64194518e-01
-3.46557617e-01 1.11245263e+00 1.29938692e-01 1.82438672e-01
-9.81966853e-01 2.47363988e-02 5.96867502e-01 -8.38070139e-02
-4.60461527e-02 -1.92120105e-01 1.74201980e-01 -1.34609294e+00
1.52958632e-01 -1.14756060e+00 7.97696888e-01 -8.21739495e-01
-2.61963576e-01 5.80103517e-01 1.83618888e-01 -1.97008479e+00
-5.10300636e-01 -8.37430000e-01 -2.96049029e-01 2.30860144e-01
-1.11869383e+00 -1.37677932e+00 -1.06397939e+00 8.52745414e-01
7.95848906e-01 1.32518604e-01 1.17998552e+00 9.96682718e-02
-4.91006911e-01 2.64479786e-01 -3.72812986e-01 4.20925677e-01
8.35584283e-01 -1.12925875e+00 -9.90085080e-02 5.49021482e-01
2.41287991e-01 7.02902198e-01 4.68745440e-01 -5.70109129e-01
-1.54831171e+00 -9.49354410e-01 1.67091414e-01 -1.15638781e+00
4.64448631e-01 -4.11731392e-01 -2.02570990e-01 1.21255898e+00
-3.08028292e-02 2.97918648e-01 8.26714218e-01 -1.27593707e-02
4.03061882e-02 1.90746292e-01 -7.49331236e-01 4.34098959e-01
1.55850136e+00 -2.47831941e-01 -7.23728359e-01 4.25670236e-01
3.49600255e-01 -1.24500263e+00 -1.06107128e+00 6.29252076e-01
8.29598844e-01 -9.37985718e-01 1.15271080e+00 -5.98108232e-01
4.92734730e-01 -1.51223302e-01 1.02836452e-01 -1.32457983e+00
-6.36265725e-02 -4.94445086e-01 -5.99756420e-01 3.96446437e-01
4.15157117e-02 7.00375140e-02 8.47351789e-01 1.85139537e-01
-7.15314150e-01 -7.23456562e-01 -1.04589045e+00 -6.93816841e-01
-5.09061813e-01 -3.35697085e-01 2.33262151e-01 7.73695886e-01
1.96901992e-01 -1.97723746e-01 -3.27892035e-01 -1.40695482e-01
2.71742612e-01 2.42672667e-01 8.43275130e-01 -1.01379871e+00
-1.93459153e-01 -2.61989623e-01 -6.66668177e-01 -7.44763136e-01
-4.02467651e-03 -8.05943489e-01 1.93306804e-01 -1.75691748e+00
3.05902243e-01 -1.18269034e-01 -3.33765596e-01 9.09712732e-01
1.31845579e-01 3.00678581e-01 3.23142260e-02 3.24218363e-01
-7.41162419e-01 2.39919737e-01 1.39827681e+00 -2.92400271e-01
-3.01698625e-01 2.20156442e-02 -3.14090669e-01 8.29869628e-01
2.32830122e-01 -3.58754873e-01 -2.44053066e-01 -5.75684726e-01
-1.55247571e-02 3.14190894e-01 4.05229270e-01 -1.19070995e+00
1.76083535e-01 -2.54261762e-01 2.52948254e-01 -5.42148471e-01
6.15221977e-01 -1.07154608e+00 4.66683596e-01 6.48915231e-01
-3.25935245e-01 -2.35470533e-01 2.24969462e-01 6.42269313e-01
-3.80296946e-01 1.19111985e-01 4.78093863e-01 -1.96391076e-01
-1.01222503e+00 3.40595990e-01 -5.16462564e-01 -2.06185460e-01
1.51824725e+00 -3.97155404e-01 -7.96044320e-02 1.26657113e-01
-1.38866806e+00 1.82236433e-01 3.51157546e-01 9.85207200e-01
4.73979652e-01 -1.26323628e+00 -3.43129426e-01 2.60684729e-01
7.60809600e-01 1.25826374e-01 5.11219978e-01 1.44301009e+00
-6.89797461e-01 4.55042660e-01 -3.97767156e-01 -8.42531025e-01
-1.38152504e+00 5.46720266e-01 6.41137302e-01 -4.06075925e-01
-9.27683890e-01 6.66302800e-01 4.89533991e-01 -1.72814816e-01
5.10545254e-01 -7.45764434e-01 -7.01321423e-01 2.31433049e-01
3.93782377e-01 -2.38166749e-02 3.16381514e-01 -4.08286184e-01
-6.99901342e-01 6.57053947e-01 -2.40077395e-02 2.64502078e-01
1.09818947e+00 3.10805589e-01 2.77729660e-01 3.13060015e-01
1.07607591e+00 -1.41989499e-01 -1.21572983e+00 1.20377585e-01
8.94393250e-02 -4.60954607e-01 -1.03086442e-01 -1.00881147e+00
-9.39477384e-01 4.72372085e-01 8.20043325e-01 -3.61803830e-01
1.01675189e+00 1.52137533e-01 2.18112364e-01 3.93133432e-01
6.53859854e-01 -9.78926182e-01 4.10537869e-01 3.67847830e-01
1.12578869e+00 -1.12843907e+00 -2.25443676e-01 -6.47239625e-01
-6.60975516e-01 1.20707321e+00 6.96947932e-01 5.35748005e-02
4.76407439e-01 2.92054355e-01 8.25210810e-02 -5.07590234e-01
-5.09404838e-01 -5.87296821e-02 5.58672607e-01 4.64277297e-01
4.56361860e-01 4.09610234e-02 -3.00891191e-01 5.17027855e-01
6.08530343e-02 5.79996169e-01 6.20591044e-01 1.51818955e+00
1.14683442e-01 -8.68467391e-01 -1.18740372e-01 4.81672436e-01
-3.04825604e-01 1.47553802e-01 -2.16273889e-01 8.43179286e-01
4.47920039e-02 5.92979908e-01 1.13034220e-02 -3.41954410e-01
8.66198659e-01 -2.84941435e-01 7.95375943e-01 -1.07205212e+00
-8.80496979e-01 9.11987796e-02 3.71618032e-01 -1.31274164e+00
-7.88426280e-01 -8.26346517e-01 -1.19256341e+00 5.00704706e-01
2.33703330e-01 1.41894951e-01 4.37681377e-01 7.78292835e-01
2.87770629e-01 1.00329363e+00 1.78341776e-01 -1.26949191e+00
-9.91879404e-02 -7.71107137e-01 -2.75974810e-01 3.99424195e-01
3.88389796e-01 -1.24138141e+00 -2.71748543e-01 2.49096572e-01] | [14.093537330627441, -3.3978044986724854] |
8607749e-3e4e-40b5-8521-f981a779ee84 | learning-generative-embeddings-using-an | 2209.00372 | null | https://arxiv.org/abs/2209.00372v1 | https://arxiv.org/pdf/2209.00372v1.pdf | Learning Generative Embeddings using an Optimal Subsampling Policy for Tensor Sketching | Data tensors of orders 3 and greater are routinely being generated. These data collections are increasingly huge and growing. They are either tensor fields (e.g., images, videos, geographic data) in which each location of data contains important information or permutation invariant general tensors (e.g., unsupervised latent space learning, graph network analysis, recommendation systems, etc.). Directly accessing such large data tensor collections for information has become increasingly prohibitive. We learn approximate full-rank and compact tensor sketches with decompositive representations providing compact space, time and spectral embeddings of both tensor fields (P-SCT) and general tensors (P-SCT-Permute). All subsequent information querying with high accuracy is performed on the generative sketches. We produce optimal rank-r Tucker decompositions of arbitrary order data tensors by building tensor sketches from a sample-efficient sub-sampling of tensor slices. Our sample efficient policy is learned via an adaptable stochastic Thompson sampling using Dirichlet distributions with conjugate priors. | ['Rochan Avlur', 'Taemin Heo', 'Chandrajit Bajaj'] | 2022-09-01 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [-1.86141133e-01 4.01729159e-02 -4.31472361e-01 -7.46843740e-02
-5.19087017e-01 -1.06721997e+00 8.39411914e-01 -1.81043535e-01
1.20300643e-01 3.42458814e-01 9.48835075e-01 -2.20872164e-01
-7.25823462e-01 -5.32692373e-01 -5.86670935e-01 -5.90178788e-01
-6.21041775e-01 9.44461763e-01 -7.90609121e-02 2.75360912e-01
5.22526026e-01 6.35634661e-01 -1.34625101e+00 4.36113328e-01
4.56205040e-01 1.02710974e+00 -7.19822198e-02 9.88925815e-01
-2.86054343e-01 5.08285522e-01 -7.98063725e-03 -4.18467253e-01
4.52899337e-01 3.99593323e-01 -1.00263906e+00 4.89579529e-01
4.60338295e-01 -7.24201977e-01 -4.85556930e-01 7.25197077e-01
-1.11487970e-01 1.37273684e-01 8.37602913e-01 -1.51765573e+00
-9.19298172e-01 6.69229746e-01 -5.07461488e-01 -6.59092069e-02
2.42224440e-01 -9.38127041e-02 1.40928280e+00 -1.13437021e+00
9.66468453e-01 1.41137636e+00 1.85004279e-01 1.25923395e-01
-1.64041495e+00 -2.04001829e-01 -1.24479467e-02 5.27911820e-02
-1.04655409e+00 -1.92474261e-01 9.56816971e-01 -8.15511167e-01
5.53407907e-01 7.02431381e-01 8.93136621e-01 1.16072488e+00
-1.55090660e-01 1.06022429e+00 7.01448739e-01 1.03328317e-01
3.36488098e-01 -3.40875447e-01 4.52745194e-03 6.76525354e-01
5.31374097e-01 -4.93961364e-01 -7.79336274e-01 -8.39187920e-01
1.31230855e+00 7.80468345e-01 1.54902190e-01 -8.41252685e-01
-2.08536935e+00 7.08173752e-01 -6.00458682e-03 1.73605725e-01
-6.84658527e-01 6.07304633e-01 4.66508806e-01 2.50309527e-01
5.62839389e-01 4.79239613e-01 -4.46653455e-01 -5.17311633e-01
-8.19595277e-01 4.71858799e-01 7.95299768e-01 1.25677979e+00
7.34508693e-01 2.72051115e-02 -4.75744791e-02 7.72306025e-01
1.55711368e-01 5.66884041e-01 2.48997271e-01 -1.38446856e+00
6.49776280e-01 6.39774382e-01 3.41021150e-01 -1.12752569e+00
-8.12963769e-02 7.22995251e-02 -1.07726645e+00 -6.97512507e-01
3.15711111e-01 2.24961162e-01 -7.62284517e-01 1.20200360e+00
3.93518627e-01 1.07900135e-01 -3.57132852e-01 7.85663784e-01
-1.91041101e-02 8.34419489e-01 -3.39266092e-01 1.95967108e-02
1.78574908e+00 -4.77567941e-01 -7.44139791e-01 4.59503651e-01
5.57306528e-01 -9.10698473e-01 9.59764123e-01 6.86105132e-01
-8.39936197e-01 -2.65776068e-02 -5.07520020e-01 -4.22960043e-01
-2.33213961e-01 4.04199243e-01 1.11479330e+00 3.16698968e-01
-1.04904890e+00 5.51693857e-01 -1.11175072e+00 -2.99767584e-01
3.81499469e-01 1.02910489e-01 -6.71646118e-01 -4.26961362e-01
-4.62209642e-01 -7.11446032e-02 3.27618346e-02 -1.17687263e-01
-1.02826774e+00 -9.94375587e-01 -4.55365390e-01 -1.56398550e-01
4.31757361e-01 -6.73903823e-01 8.76948178e-01 -6.04877323e-02
-1.36997676e+00 5.13935685e-01 -7.84724355e-02 -8.52262899e-02
9.58597511e-02 -2.65172541e-01 6.37113163e-03 4.07272607e-01
2.76148379e-01 3.13024759e-01 1.42777848e+00 -9.89513040e-01
-1.99490279e-01 -4.24164891e-01 1.41451627e-01 1.14107477e-02
-6.71228051e-01 -9.78784915e-03 -1.38909355e-01 -8.84762466e-01
7.94089258e-01 -1.24794173e+00 -2.66819328e-01 3.21925849e-01
-8.03699255e-01 -4.50635284e-01 1.13560772e+00 -9.31665063e-01
1.11585760e+00 -1.99512208e+00 7.85255432e-01 2.94442654e-01
7.60047972e-01 -5.13443708e-01 -2.57048190e-01 1.00960672e+00
-1.51638150e-01 2.80679554e-01 9.26523283e-02 -2.78058261e-01
2.55333096e-01 4.61277872e-01 -8.12865317e-01 4.81550127e-01
4.81377952e-02 7.02019870e-01 -1.18011904e+00 -2.54925877e-01
-1.68538839e-01 2.42367610e-01 -1.02790761e+00 1.91204622e-01
-3.88457507e-01 2.80382574e-01 -8.84169638e-01 7.12431967e-01
5.07407606e-01 -6.80777371e-01 4.11360651e-01 -7.66270638e-01
-1.61434650e-01 2.71833986e-01 -1.23384440e+00 1.92976439e+00
-2.94515103e-01 4.18192595e-01 -8.55145790e-03 -8.01923573e-01
6.21546626e-01 4.55694467e-01 1.01014078e+00 2.75626987e-01
-4.37022567e-01 1.88200831e-01 -3.84770066e-01 -4.60121751e-01
1.00034046e+00 2.58968353e-01 -4.66156043e-02 1.10792387e+00
5.04148640e-02 -1.33740604e-01 3.11782181e-01 8.28639805e-01
1.22469544e+00 1.39409574e-02 -2.37414569e-01 -4.54850435e-01
-1.08582810e-01 -1.45965055e-01 1.46505043e-01 8.96164775e-02
3.50119352e-01 5.44659317e-01 9.91202950e-01 -7.81995177e-01
-1.78030097e+00 -1.10698819e+00 3.85969616e-02 1.07464802e+00
-4.60295141e-01 -7.56122410e-01 -3.86232704e-01 -3.82192731e-01
3.28601211e-01 3.73179108e-01 -6.07464194e-01 1.78146258e-01
-2.86642194e-01 -3.68398488e-01 2.40179151e-01 3.87907088e-01
6.56457767e-02 -1.48073778e-01 -5.51915616e-02 1.32991463e-01
-3.37120920e-01 -1.03691149e+00 -1.07261574e+00 -3.59854072e-01
-1.51983511e+00 -7.37255633e-01 -8.88125300e-01 -2.20487759e-01
1.16143870e+00 5.73581457e-01 8.01230550e-01 -3.99201691e-01
-3.03043425e-01 9.05064344e-01 -2.99088567e-01 3.47591221e-01
1.38046563e-01 -9.49495807e-02 6.87862933e-01 8.07836294e-01
-8.63907486e-03 -9.63123262e-01 -6.89280450e-01 3.10801446e-01
-1.37577939e+00 1.51958913e-01 4.26201224e-01 6.64678633e-01
5.26686847e-01 -1.33991137e-01 -3.74287903e-01 -5.44641972e-01
9.58306193e-01 -7.46853709e-01 -4.62903857e-01 2.32555285e-01
-2.08131820e-01 8.48899901e-01 3.09061170e-01 -7.98591554e-01
-4.94143367e-01 -2.87796140e-01 9.33895290e-01 -1.12866473e+00
3.83076936e-01 5.06113529e-01 2.34866470e-01 2.97729135e-01
4.76347476e-01 2.79005527e-01 -1.02275461e-01 -1.02159894e+00
7.42156148e-01 4.54123497e-01 8.14714376e-03 -1.37842774e+00
1.02986801e+00 7.47550905e-01 2.40710810e-01 -1.11509860e+00
-3.97059292e-01 -5.99507451e-01 -8.25265706e-01 -1.50091067e-01
7.70220697e-01 -6.13609016e-01 -6.05189919e-01 2.22031310e-01
-1.21312022e+00 4.10231724e-02 -4.84089673e-01 7.13303745e-01
-5.07316530e-01 5.51078558e-01 -7.08413422e-01 -7.48613834e-01
3.37624326e-02 -9.59015548e-01 1.42737007e+00 -6.44915521e-01
-1.83922857e-01 -8.20918083e-01 2.58513987e-01 5.22769809e-01
3.48991901e-01 1.29343182e-01 1.09249246e+00 -3.55273634e-01
-1.47094047e+00 -3.06979358e-01 -3.79584342e-01 1.66947260e-01
2.28457123e-01 3.60230058e-01 -5.09429872e-01 -3.47367227e-01
-4.69858110e-01 -2.85359383e-01 4.77917641e-01 2.15915069e-01
1.46271670e+00 -1.19199812e+00 -2.16983095e-01 5.51857233e-01
9.87179577e-01 -3.78341973e-01 2.81562150e-01 -2.73812830e-01
1.17874181e+00 6.16257191e-01 2.03998312e-01 9.30549264e-01
5.83508849e-01 5.82998931e-01 2.02546850e-01 6.45291030e-01
1.80969447e-01 -6.54962659e-01 3.80057186e-01 1.33088946e+00
-8.17864776e-01 3.71422172e-01 -9.81409073e-01 9.08847451e-01
-1.88223553e+00 -1.02846932e+00 1.24834336e-01 2.29940009e+00
6.71683073e-01 -4.31548327e-01 2.75460571e-01 1.05861597e-01
3.89228702e-01 3.38453919e-01 -5.19823551e-01 -1.62162378e-01
2.66918212e-01 -2.31130913e-01 6.61366165e-01 2.35791683e-01
-7.78263330e-01 4.59497690e-01 5.70730019e+00 6.83393598e-01
-7.26794362e-01 4.81504053e-02 3.10067236e-01 -2.46157229e-01
-1.08347130e+00 2.51075447e-01 -1.85299635e-01 2.13716075e-01
7.69297123e-01 -1.97863057e-01 1.13065469e+00 8.82935286e-01
8.95651709e-03 2.92777807e-01 -1.30609727e+00 1.34459496e+00
-4.02605027e-01 -1.76284242e+00 6.72269702e-01 8.02365482e-01
8.94571543e-01 -6.87475055e-02 3.19070339e-01 -1.44902200e-01
6.98496401e-01 -4.41230357e-01 7.53509939e-01 7.90403724e-01
8.77992153e-01 -3.30648243e-01 -1.21387139e-01 7.12177604e-02
-1.26205397e+00 -5.84813729e-02 -5.51033199e-01 2.30221719e-01
3.47025581e-02 1.07905412e+00 -8.14856946e-01 4.49111193e-01
5.96253276e-01 1.01440048e+00 -2.13178128e-01 6.01642311e-01
2.09316283e-01 6.81547165e-01 -7.01035500e-01 -7.85355419e-02
3.15292090e-01 -8.68265808e-01 8.78228426e-01 5.48863351e-01
6.00198090e-01 1.27812222e-01 2.04723760e-01 8.17963004e-01
-1.77190930e-01 -9.96255055e-02 -8.86133850e-01 -9.83875453e-01
4.30471212e-01 1.40718377e+00 -7.19166815e-01 -5.48521698e-01
-1.30712077e-01 9.57195997e-01 1.45921201e-01 6.80628121e-01
-1.70052201e-01 5.70042385e-03 1.29230952e+00 4.80017662e-01
4.22249049e-01 -1.05993605e+00 6.47185196e-04 -1.73107231e+00
2.03138173e-01 -7.28500485e-01 2.34804839e-01 -7.30326951e-01
-1.47330928e+00 2.10543960e-01 4.69838202e-01 -1.52660012e+00
-3.31169784e-01 -7.52426863e-01 -6.71765581e-02 5.92065454e-01
-7.97538102e-01 -1.28275490e+00 1.99256659e-01 7.20753193e-01
1.71273321e-01 -3.30998838e-01 7.08952963e-01 1.13667980e-01
-3.44403088e-01 9.39550251e-02 3.67199272e-01 6.75391629e-02
8.22299570e-02 -1.36960995e+00 5.20956159e-01 6.10925853e-01
2.72046864e-01 1.17300546e+00 3.81042659e-01 -7.01938689e-01
-2.45049739e+00 -8.78278434e-01 5.75296402e-01 -6.99111938e-01
1.41486716e+00 -7.10341513e-01 -5.26116788e-01 9.92959797e-01
-1.64676368e-01 3.24327350e-01 5.86758316e-01 3.13191831e-01
-1.01067972e+00 -2.56862432e-01 -1.03421617e+00 8.13687980e-01
1.20835876e+00 -1.06327415e+00 -1.50422871e-01 8.50822687e-01
9.34506595e-01 -9.69852135e-02 -1.54739606e+00 -2.10680589e-01
6.68526113e-01 -4.39310014e-01 1.00276124e+00 -1.00616944e+00
3.94211918e-01 -3.81767601e-01 -5.31409860e-01 -1.08522356e+00
-5.03524542e-01 -1.22482467e+00 -5.02341509e-01 8.17945898e-01
8.15132037e-02 -5.87945342e-01 7.88410962e-01 8.67213249e-01
1.96386397e-01 -5.15424311e-01 -9.57819819e-01 -7.64113784e-01
-4.94994849e-01 -4.26837593e-01 9.58729267e-01 1.16965163e+00
1.68261170e-01 1.29633263e-01 -5.74334145e-01 1.76001608e-01
1.26850379e+00 3.05060148e-01 8.75818849e-01 -1.22176194e+00
-2.01243818e-01 -2.15643391e-01 -4.81444150e-01 -1.31145442e+00
-3.90851945e-01 -9.37388122e-01 -7.49127924e-01 -1.37246943e+00
3.33347440e-01 -5.40242732e-01 2.11163640e-01 4.20922518e-01
4.52490330e-01 -1.22382194e-01 3.20914090e-02 7.73753285e-01
-4.63622302e-01 6.66568518e-01 1.62120569e+00 -1.68050453e-01
1.81924313e-01 -2.33529463e-01 -2.87761420e-01 2.55823165e-01
4.42569703e-01 -2.84109294e-01 -7.37745881e-01 -5.27085364e-01
7.43767619e-01 3.75067115e-01 2.68211842e-01 -4.79557365e-01
9.95088816e-02 -5.83328784e-01 -5.33828177e-02 -6.41970158e-01
5.04110813e-01 -6.67454481e-01 4.34121728e-01 1.81397378e-01
-3.84683996e-01 3.34481180e-01 -3.22854102e-01 8.58999014e-01
2.38181427e-02 2.48001188e-01 1.32461488e-02 -1.00588568e-01
-1.57243341e-01 1.02397835e+00 -1.63586959e-01 -2.34280173e-02
5.41314721e-01 -1.45624399e-01 -1.58570483e-01 -4.92849410e-01
-6.41672671e-01 -1.95109159e-01 3.76979500e-01 5.31123459e-01
7.82046139e-01 -1.87267816e+00 -6.29480183e-01 3.99014950e-01
3.56359154e-01 9.18314680e-02 2.85289824e-01 7.68468738e-01
-5.27625740e-01 5.71713746e-01 -1.71702281e-01 -6.71109080e-01
-7.43806779e-01 5.87727308e-01 -3.17297935e-01 -4.04355824e-01
-7.18558371e-01 6.16907239e-01 1.11572579e-01 -6.99101150e-01
-1.23430438e-01 -1.11943209e+00 1.63646415e-01 1.39102384e-01
5.67386091e-01 6.56953692e-01 -3.74672979e-01 -2.89186269e-01
8.03811569e-03 3.55327934e-01 -1.00629233e-01 -3.82059813e-01
1.63993204e+00 1.66674271e-01 -7.91180730e-01 6.17791831e-01
1.36125314e+00 8.26943889e-02 -1.23514187e+00 -4.21777666e-01
6.73864111e-02 -7.32586861e-01 6.54998235e-03 3.03321630e-02
-1.03468466e+00 7.85790205e-01 -1.71849310e-01 6.51304960e-01
5.43240786e-01 1.84973359e-01 7.10171521e-01 6.82631552e-01
7.75231242e-01 -8.16793561e-01 4.54109997e-01 3.31412554e-01
1.35255599e+00 -4.65169519e-01 2.78055698e-01 -1.10408105e-01
-4.97699022e-01 1.28350282e+00 -3.01935285e-01 -3.67465198e-01
1.04955506e+00 -1.61588609e-01 -6.13011897e-01 -5.85205257e-01
-1.08920860e+00 3.55705827e-01 5.86659610e-01 4.44217801e-01
8.95846039e-02 6.13661349e-01 1.47484362e-01 3.65241766e-02
-2.90411294e-01 -1.21125542e-01 5.98510623e-01 7.53662765e-01
-9.81346220e-02 -1.21856880e+00 -6.22428656e-01 9.94265795e-01
-9.40769911e-02 -1.76150072e-02 -3.65092792e-02 1.71487257e-01
-5.80854595e-01 3.24456155e-01 1.98078156e-01 -5.15549064e-01
-1.01922750e-01 3.52501422e-02 4.24422741e-01 -5.09685457e-01
3.84472698e-01 -1.57161877e-02 1.87232524e-01 -9.68198001e-01
-2.40358829e-01 -1.13876951e+00 -6.40162885e-01 -6.39604390e-01
3.64034683e-01 2.20822975e-01 1.03941572e+00 5.20745575e-01
6.30208015e-01 -1.11892365e-01 7.71140039e-01 -1.24621820e+00
-7.26615071e-01 -9.25862134e-01 -1.01060200e+00 4.06816125e-01
4.27824080e-01 -7.04972923e-01 -3.54238033e-01 3.44752014e-01] | [7.189509868621826, 4.691577911376953] |
095eaf86-95ce-4e25-bfc1-b51fb9249b14 | transfer-learning-based-detection-of-diabetic | 1905.07203 | null | https://arxiv.org/abs/1905.07203v2 | https://arxiv.org/pdf/1905.07203v2.pdf | Transfer Learning based Detection of Diabetic Retinopathy from Small Dataset | Annotated training data insufficiency remains to be one of the challenges of applying deep learning in medical data classification problems. Transfer learning from an already trained deep convolutional network can be used to reduce the cost of training from scratch and to train with small training data for deep learning. This raises the question of whether we can use transfer learning to overcome the training data insufficiency problem in deep learning based medical data classifications. Deep convolutional networks have been achieving high performance results on the ImageNet Large Scale Visual Recognition Competition (ILSVRC) image classification challenge. One example is the Inception-V3 model that was the first runner up on the ILSVRC 2015 challenge. Inception modules that help to extract different sized features of input images in one level of convolution are the unique features of the Inception-V3. In this work, we have used a pretrained Inception-V3 model to take advantage of its Inception modules for Diabetic Retinopathy detection. In order to tackle the labelled data insufficiency problem, we sub-sampled a smaller version of the Kaggle Diabetic Retinopathy classification challenge dataset for model training, and tested the model's accuracy on a previously unseen data subset. Our technique could be used in other deep learning based medical image classification problems facing the challenge of labeled training data insufficiency. | ['Misgina Tsighe Hagos', 'Shri Kant'] | 2019-05-17 | null | null | null | null | ['diabetic-retinopathy-detection'] | ['medical'] | [ 2.20034420e-01 3.51617306e-01 -8.27242509e-02 -6.03552043e-01
-4.75125194e-01 -1.76568985e-01 3.55588347e-01 1.83038235e-01
-8.06351960e-01 4.62415010e-01 1.25073761e-01 -5.50801516e-01
-1.90846458e-01 -7.57587850e-01 -6.72150195e-01 -5.27606785e-01
2.30362103e-03 5.15883148e-01 1.67018458e-01 -3.39421004e-01
-3.13555785e-02 5.54496109e-01 -1.56214225e+00 1.12091911e+00
7.31248975e-01 8.79644871e-01 1.07824638e-01 9.24419940e-01
-6.26541749e-02 1.18115199e+00 -6.07874572e-01 4.48826365e-02
3.29734266e-01 -3.47609788e-01 -1.18499136e+00 1.36467844e-01
6.90519035e-01 -5.08728921e-01 -2.04988390e-01 7.75765717e-01
7.98746705e-01 -3.41608524e-01 5.79564273e-01 -1.05328596e+00
-4.90509719e-01 1.83465093e-01 -2.76552886e-01 6.22802913e-01
-2.22039267e-01 2.65785307e-01 4.52037394e-01 -6.28336906e-01
7.31028318e-01 1.02091408e+00 9.01541889e-01 6.94043756e-01
-1.18228066e+00 -2.94305414e-01 -2.74922550e-01 4.09301907e-01
-1.05330479e+00 -1.94823220e-01 2.06203088e-01 -7.58644760e-01
1.01874995e+00 3.54020596e-01 4.81997490e-01 1.04407394e+00
1.65494531e-02 8.75206828e-01 1.12270701e+00 -5.16795695e-01
1.35085031e-01 1.23542495e-01 2.83723205e-01 8.61710906e-01
1.32347256e-01 3.46915156e-01 1.65595323e-01 -1.40074212e-02
7.36134529e-01 -2.87591182e-02 -4.12089556e-01 -2.43492961e-01
-1.03891504e+00 1.09504366e+00 1.06727266e+00 3.72898847e-01
-2.14936033e-01 -7.55669037e-03 7.39989638e-01 7.54026711e-01
4.20497566e-01 5.92091680e-01 -8.69502306e-01 2.01757476e-01
-5.77160358e-01 -7.86810592e-02 5.71339130e-01 3.19250256e-01
6.02611721e-01 -3.43413889e-01 -4.00834620e-01 1.04610407e+00
-2.52173468e-02 -8.46824199e-02 6.43237710e-01 -6.35422707e-01
2.93929636e-01 1.02407956e+00 -2.37510204e-01 -3.14129502e-01
-9.36849117e-01 -5.09513021e-01 -1.04473972e+00 9.56700921e-01
8.68697226e-01 -3.57658237e-01 -1.58918810e+00 1.04601383e+00
1.27949521e-01 -6.57786615e-03 2.72976637e-01 9.14974689e-01
1.14282668e+00 2.76675642e-01 2.82786489e-02 3.91374707e-01
1.30164325e+00 -9.22955811e-01 -1.08128749e-01 -1.06875300e-01
1.31623769e+00 -5.67249894e-01 7.47263432e-01 4.63848025e-01
-4.77709025e-01 -7.46219397e-01 -9.66658473e-01 -3.39766413e-01
-6.60858274e-01 5.16411006e-01 6.18005991e-01 3.95181298e-01
-1.16457319e+00 6.38204813e-01 -5.71362555e-01 -5.43073714e-01
1.00090170e+00 5.39426744e-01 -7.30325937e-01 -5.94824553e-01
-9.09763098e-01 1.28455853e+00 5.24346709e-01 2.34116241e-01
-7.63586104e-01 -7.77527511e-01 -6.54609501e-01 -1.92387611e-01
1.83221232e-02 -6.94478273e-01 9.03078496e-01 -1.37519860e+00
-8.93123388e-01 1.34540999e+00 5.53855956e-01 -7.56532192e-01
7.09085822e-01 4.81036194e-02 -3.64210755e-01 1.36892095e-01
-5.10281399e-02 8.52518797e-01 5.72368026e-01 -9.28174794e-01
-7.74554908e-01 -3.09279889e-01 9.44187269e-02 -2.65967429e-01
3.93695086e-02 -1.46043539e-01 -5.62519804e-02 -4.12750691e-01
-2.53067553e-01 -8.98780167e-01 -4.17547971e-01 1.97586313e-01
-1.75400272e-01 -4.03757513e-01 5.66532552e-01 -5.50264597e-01
5.07508039e-01 -2.12325788e+00 -7.30281621e-02 3.49741019e-02
3.69181156e-01 1.00348830e+00 -5.83693624e-01 1.54839784e-01
-4.82516080e-01 1.00445673e-01 1.02598995e-01 1.79948006e-02
-7.16789842e-01 3.79637122e-01 2.21690863e-01 4.73966658e-01
6.02232516e-01 9.56146121e-01 -7.56167352e-01 -2.33773276e-01
4.29736167e-01 4.74415898e-01 -4.65872824e-01 9.06379148e-02
-3.99360210e-02 5.03560722e-01 -1.68233156e-01 4.71435964e-01
5.70356011e-01 -3.12192649e-01 -2.24586904e-01 -3.11145276e-01
-6.97667748e-02 -7.53680766e-02 -6.02596998e-01 1.58486652e+00
-4.68177259e-01 8.99172544e-01 -2.76083440e-01 -1.47674251e+00
7.86601841e-01 3.47067982e-01 5.12768745e-01 -9.95683670e-01
2.79709995e-01 2.77152002e-01 5.40913045e-01 -9.68962014e-01
-2.44493499e-01 -2.68595397e-01 2.88783580e-01 -1.07929274e-01
4.42832708e-01 4.08219486e-01 1.09330982e-01 -1.20187625e-01
1.32184923e+00 -5.16451448e-02 3.13955665e-01 -2.54732490e-01
4.69315886e-01 2.97834337e-01 3.75084698e-01 7.36489832e-01
-4.10777122e-01 9.95065153e-01 6.74806714e-01 -1.23328209e+00
-1.05415726e+00 -5.73139131e-01 -5.50469518e-01 9.03992176e-01
-5.00731647e-01 -2.17660844e-01 -4.40402985e-01 -1.17000830e+00
1.88918039e-01 2.99189627e-01 -1.21930158e+00 -2.69900233e-01
-4.48823690e-01 -8.60138416e-01 6.55650377e-01 6.39995515e-01
4.25801396e-01 -1.24363649e+00 -7.90248632e-01 9.08137932e-02
2.20758274e-01 -8.48820150e-01 2.78362393e-01 4.95623440e-01
-8.30742776e-01 -1.57016230e+00 -8.95697892e-01 -1.06829023e+00
8.13785553e-01 -1.39489308e-01 1.10641515e+00 3.81621569e-01
-1.03167927e+00 1.34663865e-01 -4.93303984e-01 -8.04528832e-01
-6.31254017e-01 7.81974569e-02 -4.54499036e-01 1.34848937e-01
6.59281313e-01 -1.70891106e-01 -7.13788629e-01 1.95827067e-01
-8.64362240e-01 1.35052204e-01 9.18620408e-01 1.12551939e+00
3.34168792e-01 -1.02721497e-01 8.48159790e-01 -1.11170077e+00
2.11305603e-01 -3.48777592e-01 -4.23861712e-01 1.27516001e-01
-4.01193976e-01 1.58327579e-01 4.56306785e-01 -2.62962162e-01
-4.02382672e-01 2.71143854e-01 -4.55702394e-01 -4.19081360e-01
-4.51761335e-01 6.24139845e-01 2.38319889e-01 -4.90066670e-02
1.04390883e+00 -1.62548706e-01 3.19577426e-01 -7.22286582e-01
1.38445899e-01 8.15742373e-01 3.82579952e-01 6.26134723e-02
1.98804006e-01 3.85963887e-01 1.09402008e-01 -7.59812176e-01
-1.20528889e+00 -6.79215550e-01 -7.83114195e-01 1.56341847e-02
1.22784758e+00 -9.61932838e-01 -4.40092891e-01 5.10022283e-01
-9.50045347e-01 -6.34725869e-01 -4.45059478e-01 4.20121908e-01
-2.13266999e-01 1.74906999e-02 -4.23422933e-01 -2.21066862e-01
-2.54688770e-01 -1.13915360e+00 8.63958955e-01 1.05908826e-01
8.10565948e-02 -1.09661424e+00 2.14591414e-01 5.40042400e-01
4.88405734e-01 6.11826658e-01 1.05913115e+00 -8.52876902e-01
-9.56015512e-02 -5.08677602e-01 -5.12647152e-01 9.88500416e-01
2.24729866e-01 -3.06050509e-01 -1.10479426e+00 -4.84103382e-01
-3.29364359e-01 -7.40131021e-01 1.57666612e+00 2.92994380e-01
1.24591410e+00 -1.17014199e-01 -2.08377898e-01 6.70889854e-01
1.52831089e+00 -1.09520972e-01 1.08517790e+00 5.79550445e-01
5.96856892e-01 5.49478590e-01 2.41212294e-01 6.15556315e-02
2.51461685e-01 4.40853894e-01 6.10597432e-01 -8.42414618e-01
-5.41547298e-01 1.09546185e-01 -1.05017282e-01 5.54775372e-02
-3.27074915e-01 1.45962790e-01 -1.04534471e+00 7.37361372e-01
-1.72662330e+00 -6.61409378e-01 -4.77185458e-01 2.05558610e+00
7.91982889e-01 1.17842823e-01 1.01532690e-01 9.13475305e-02
3.23392928e-01 -4.46761966e-01 -3.81032556e-01 -4.21709239e-01
-6.43747225e-02 3.84041458e-01 5.40157437e-01 2.31216416e-01
-1.40201664e+00 5.81150055e-01 5.57090807e+00 3.99999648e-01
-1.27559686e+00 4.54245023e-02 7.40288854e-01 6.87856600e-02
4.82050478e-01 -1.90127656e-01 -5.49238622e-01 3.34899306e-01
1.04735863e+00 5.81007719e-01 -3.00102495e-02 8.04167986e-01
7.92042539e-02 -1.76450908e-01 -1.35764444e+00 9.11264420e-01
6.78221695e-03 -1.41677511e+00 8.90844017e-02 1.93918049e-01
5.57769477e-01 5.66038728e-01 -1.27139077e-01 5.93443155e-01
1.18631509e-03 -1.57315731e+00 -2.17885718e-01 5.29543579e-01
8.21851432e-01 -6.43557608e-01 1.23876965e+00 2.79412657e-01
-4.58875924e-01 -2.51093447e-01 -7.12634265e-01 -1.00719847e-01
-3.69359881e-01 5.15286326e-01 -1.31105316e+00 3.82078767e-01
8.49141657e-01 8.08630824e-01 -9.83222365e-01 1.32367158e+00
1.73155293e-01 5.28693676e-01 -4.60129976e-02 3.11517388e-01
4.80693847e-01 1.06874727e-01 7.26574808e-02 1.26594353e+00
-1.15293898e-01 -1.67898655e-01 1.59371451e-01 5.41963756e-01
-2.10415587e-01 -5.85882813e-02 -6.90736592e-01 1.35026723e-01
-5.08196533e-01 1.22952604e+00 -4.28288996e-01 -4.21465248e-01
-5.06087065e-01 9.31145787e-01 4.41572130e-01 1.77330688e-01
-4.60769475e-01 -2.98074573e-01 3.99831563e-01 1.85968876e-01
5.12234211e-01 4.26061720e-01 -2.23644137e-01 -1.09608459e+00
-4.70065355e-01 -8.77301574e-01 6.00611329e-01 -6.74693584e-01
-1.40616727e+00 7.29142666e-01 -5.20271957e-01 -1.24594200e+00
1.01425171e-01 -1.07421088e+00 -5.86969018e-01 1.01209009e+00
-1.73092365e+00 -1.34357679e+00 -5.35761893e-01 8.71234953e-01
4.48855072e-01 -5.73693156e-01 9.79888678e-01 3.60701293e-01
-4.12730038e-01 3.54648083e-01 -7.99251199e-02 5.80251634e-01
8.25702488e-01 -1.45705569e+00 5.96325211e-02 5.65456390e-01
3.73099893e-02 1.28052101e-01 3.13075811e-01 -3.26649040e-01
-7.87052989e-01 -1.53472757e+00 7.59539306e-01 -6.88918293e-01
3.15513670e-01 -3.99678089e-02 -1.01846623e+00 6.89798117e-01
7.90411904e-02 6.74001455e-01 8.30800176e-01 6.69109598e-02
-3.66292208e-01 -7.36635700e-02 -1.14671910e+00 5.42699136e-02
5.16349673e-01 -3.07472885e-01 -7.18669891e-01 5.54147243e-01
2.78283566e-01 -3.07768494e-01 -9.19279635e-01 6.27437830e-01
3.45199823e-01 -8.98112774e-01 9.36826408e-01 -1.25364125e+00
7.15785921e-01 -7.77611062e-02 5.44335358e-02 -1.28008664e+00
-1.77713707e-01 -5.32021821e-02 3.91870737e-01 9.35505211e-01
4.92208183e-01 -4.57279444e-01 6.80128992e-01 2.39770129e-01
-2.16487318e-01 -8.52290392e-01 -9.13980186e-01 -4.42974508e-01
4.82005388e-01 -5.53326458e-02 -1.38879597e-01 9.99377668e-01
-3.66868347e-01 5.77382684e-01 -2.90045977e-01 7.02504590e-02
3.69483948e-01 -9.61756986e-03 6.96668863e-01 -1.47171164e+00
-2.03239948e-01 -1.86591968e-01 -8.63274693e-01 -3.24502796e-01
-2.17647359e-01 -1.36462176e+00 -1.89472869e-01 -1.97444928e+00
9.85492617e-02 -5.57247162e-01 -5.97151816e-01 9.85983789e-01
-5.50837107e-02 4.47106808e-01 1.52303845e-01 2.40403581e-02
-3.42421293e-01 5.32292910e-02 1.28307414e+00 -3.74588192e-01
-2.49743074e-01 2.67259985e-01 -6.09994709e-01 5.99293709e-01
8.26235235e-01 -4.19601858e-01 -2.31448069e-01 -4.53310937e-01
2.65654940e-02 -2.52774626e-01 7.44123995e-01 -1.11616230e+00
-3.39924395e-02 2.98622906e-01 7.78138757e-01 -3.44029546e-01
2.08719186e-02 -8.39555740e-01 -3.89181405e-01 8.95531356e-01
-4.28030789e-01 -5.20832241e-01 4.67274606e-01 2.96509057e-01
-1.97186589e-01 -2.47687310e-01 1.09810424e+00 -3.95227075e-01
-9.56877291e-01 1.81941167e-01 -4.65673000e-01 -4.46699671e-02
9.32643771e-01 -3.28836520e-03 -5.20725131e-01 2.59623498e-01
-1.05506074e+00 2.81853616e-01 1.90905973e-01 6.42168820e-01
6.65766239e-01 -9.90580201e-01 -1.08043694e+00 3.62302780e-01
5.74923992e-01 8.21541697e-02 2.30250791e-01 9.50266480e-01
-8.93160701e-01 4.59748328e-01 -6.66650057e-01 -8.98406088e-01
-1.31513000e+00 7.71494865e-01 8.70720506e-01 -3.22430998e-01
-1.05526149e+00 8.04334641e-01 1.79686964e-01 -4.79862601e-01
1.51740685e-01 -6.91890895e-01 -4.70333368e-01 1.02324970e-01
8.18775117e-01 1.54653370e-01 5.60959935e-01 -1.67806149e-01
-3.75090241e-01 3.70271534e-01 -3.29812706e-01 5.12406290e-01
1.60252106e+00 3.05385530e-01 2.00741794e-02 1.61956966e-01
1.59446323e+00 -6.57533228e-01 -1.02210140e+00 -1.12763882e-01
1.39106780e-01 -9.53407884e-02 3.56212169e-01 -1.41252053e+00
-1.21242559e+00 1.09271789e+00 1.33039582e+00 -5.32176383e-02
1.11377668e+00 1.91467144e-02 4.06817406e-01 5.34209013e-01
1.16849646e-01 -9.04416621e-01 -8.30437839e-02 2.72925198e-01
1.04133546e+00 -1.82062876e+00 -2.48276770e-01 -3.04457545e-03
-7.40496814e-01 1.41892576e+00 6.44566119e-01 -3.36492985e-01
6.51276350e-01 1.11771494e-01 4.60975200e-01 -3.68170679e-01
-6.35900617e-01 -5.73130369e-01 4.53571826e-01 8.90571952e-01
3.63028198e-01 -1.35388881e-01 -1.58680007e-01 2.71006912e-01
2.77415156e-01 5.96637666e-01 7.49225795e-01 8.75958920e-01
-4.72398877e-01 -1.20518863e+00 5.02261110e-02 8.06355178e-01
-4.02637184e-01 -2.87598129e-02 -4.66873080e-01 8.89045000e-01
6.82088494e-01 8.07866573e-01 -6.08550571e-03 -3.63055885e-01
5.00384271e-01 -1.43102771e-02 3.65676016e-01 -7.94561625e-01
-8.40989113e-01 -2.11188883e-01 1.73610061e-01 -5.44971645e-01
-4.92109895e-01 -1.54256031e-01 -1.06122029e+00 2.18623638e-01
-9.35123209e-03 -1.35081738e-01 6.11538351e-01 9.75730717e-01
3.98943514e-01 8.55563164e-01 4.05982614e-01 -6.24273002e-01
-5.51505387e-01 -1.16775382e+00 -3.83612782e-01 5.16362190e-01
8.88303339e-01 -2.94885665e-01 -1.27703175e-01 1.43377945e-01] | [14.95710563659668, -2.4315807819366455] |
ac9e1e7f-3139-4ca1-842c-e31e4331ed67 | super-prompting-utilizing-model-independent | 2204.11922 | null | https://arxiv.org/abs/2204.11922v1 | https://arxiv.org/pdf/2204.11922v1.pdf | Super-Prompting: Utilizing Model-Independent Contextual Data to Reduce Data Annotation Required in Visual Commonsense Tasks | Pre-trained language models have shown excellent results in few-shot learning scenarios using in-context learning. Although it is impressive, the size of language models can be prohibitive to make them usable in on-device applications, such as sensors or smartphones. With smaller language models, task-specific data annotation is needed to fine-tune the language model for a specific purpose. However, data annotation can have a substantial financial and time burden for small research groups, startups, and even companies. In this paper, we analyze different prompt-based fine-tuning techniques to improve results on both language and multimodal causal transformer models. To evaluate our results, we use a dataset focusing on visual commonsense reasoning in time. Our results show that by simple model-agnostic prompt-based fine-tuning, comparable results can be reached by only using 35%-40% of the fine-tuning training dataset. The proposed approaches result in significant time and financial savings. As the proposed methods make minimal architectural assumptions, other researchers can use the results in their transformer models with minimal adaptations. We plan to release the source code freely to make it easier for the community to use and contribute to our work. | ['Marek Z. Reformat', 'Navid Rezaei'] | 2022-04-25 | null | null | null | null | ['visual-commonsense-reasoning'] | ['reasoning'] | [ 2.17421308e-01 1.45560995e-01 -3.42762798e-01 -4.29925233e-01
-6.92137539e-01 -4.95102257e-01 5.83308458e-01 2.53934592e-01
-4.24235553e-01 5.91461182e-01 1.52055085e-01 -3.56237501e-01
-1.44515922e-02 -8.92555594e-01 -3.94027472e-01 -2.20513463e-01
2.86547959e-01 3.51603776e-01 5.86725295e-01 -3.26766878e-01
4.29983169e-01 1.84182554e-01 -1.61254728e+00 4.98677731e-01
8.41431558e-01 6.53294802e-01 5.11447549e-01 6.20807767e-01
-5.92265368e-01 1.13276839e+00 -5.43542087e-01 -4.93935883e-01
9.28449258e-02 -3.32824707e-01 -6.95125997e-01 -2.10884601e-01
1.37282401e-01 -2.49348149e-01 -3.83967236e-02 7.87072659e-01
6.27279758e-01 2.46563151e-01 3.16305935e-01 -1.20732903e+00
-5.16650438e-01 9.01068091e-01 -4.34892207e-01 -2.86965985e-02
4.61575776e-01 1.21841453e-01 8.77387464e-01 -6.27918303e-01
4.89421487e-01 1.38126254e+00 7.39631116e-01 5.80074489e-01
-1.11938107e+00 -9.52001810e-01 1.49737820e-01 4.85140294e-01
-1.32112956e+00 -6.23085499e-01 1.04544914e+00 -4.35315073e-01
1.16150379e+00 2.83756614e-01 4.71089035e-01 1.17988658e+00
-1.31646022e-01 6.54163003e-01 1.24094820e+00 -7.03349650e-01
3.32720309e-01 3.83938789e-01 3.21333438e-01 7.40206242e-01
1.78201765e-01 -3.87850314e-01 -3.58558655e-01 -2.11308330e-01
4.93179858e-01 1.74163759e-01 2.56333590e-01 -3.70908380e-01
-9.80690420e-01 1.04825497e+00 5.21997325e-02 4.85348195e-01
-8.96744654e-02 1.22962892e-01 5.63758492e-01 5.27506694e-02
2.68903762e-01 3.50335985e-01 -3.96053851e-01 -6.58411920e-01
-1.06343186e+00 7.75824720e-03 7.70395577e-01 8.76150429e-01
6.74191833e-01 -8.48738030e-02 -2.65171260e-01 1.16505265e+00
5.52832745e-02 2.50421256e-01 6.28425360e-01 -9.53020513e-01
5.49281597e-01 7.15580046e-01 1.03879280e-01 -1.06925368e+00
-3.83836448e-01 2.19215825e-02 -6.12682283e-01 -9.24212486e-02
3.26141179e-01 -1.45737901e-01 -7.50359297e-01 1.64240766e+00
1.21887743e-01 2.42717862e-01 -1.77385390e-01 6.48121417e-01
7.91966677e-01 5.79932332e-01 3.69093329e-01 -3.93983871e-01
1.42874897e+00 -9.43450570e-01 -8.50616038e-01 -1.72205821e-01
6.75708175e-01 -8.76050889e-01 1.86146772e+00 2.19963863e-01
-9.29664135e-01 -4.68402147e-01 -1.00141537e+00 -1.31778687e-01
-4.54633236e-01 3.65401320e-02 8.46920013e-01 8.08845103e-01
-5.22220492e-01 6.23002291e-01 -9.64466214e-01 -9.13738608e-01
3.76602858e-01 1.12096496e-01 1.40926689e-01 -4.13463935e-02
-1.24967611e+00 9.09425795e-01 3.25613499e-01 -4.72119600e-01
-7.84325778e-01 -6.92672014e-01 -6.61862791e-01 1.48337603e-01
6.86994314e-01 -5.94184637e-01 1.33396208e+00 -6.46943152e-01
-1.49388742e+00 6.03204072e-01 -2.35840663e-01 -2.88697392e-01
5.25270581e-01 -2.16826797e-01 -3.09135407e-01 -1.57902315e-01
2.54978910e-02 4.17892903e-01 6.40202403e-01 -1.03529775e+00
-5.72923899e-01 -4.03253771e-02 6.15007997e-01 -6.77477345e-02
-7.76551068e-01 3.03350627e-01 -4.84132051e-01 -6.94678903e-01
-5.26375532e-01 -8.65613103e-01 -1.68984309e-01 -1.95526049e-01
-2.42037758e-01 -2.11614937e-01 9.45897579e-01 -5.05029619e-01
1.69645762e+00 -1.90576708e+00 -3.18247616e-01 -2.78226554e-01
-1.57691285e-01 3.68504018e-01 -1.26207313e-02 4.83805180e-01
3.49671915e-02 3.14861685e-01 -8.74756798e-02 -2.74037868e-01
2.33780537e-02 2.93411791e-01 -2.89193869e-01 -5.17582782e-02
-3.28434892e-02 7.30983257e-01 -8.57376039e-01 -7.09052265e-01
5.59553444e-01 3.14743489e-01 -6.24268591e-01 1.60692841e-01
-3.33194494e-01 -4.41988371e-02 -3.58738810e-01 6.64238155e-01
2.56524324e-01 -3.74599904e-01 2.72478878e-01 -2.77987033e-01
9.28751454e-02 9.72332135e-02 -1.33840454e+00 1.61013401e+00
-9.56788480e-01 7.04395115e-01 -3.45091343e-01 -1.07308292e+00
7.68721581e-01 2.72403121e-01 2.18562052e-01 -7.11408436e-01
1.59266610e-02 -1.25745103e-01 -4.97332849e-02 -8.53663504e-01
4.53114003e-01 -3.38076353e-01 -3.44695300e-01 5.53113520e-01
-9.53432173e-02 1.12926261e-02 4.94134367e-01 1.85583308e-01
1.01570392e+00 9.52297375e-02 6.26382053e-01 2.19002277e-01
4.14342314e-01 -3.79192419e-02 4.55525309e-01 6.66440487e-01
-1.93985417e-01 2.15339944e-01 3.46585959e-01 -2.92401135e-01
-1.17122209e+00 -7.67758429e-01 2.06471533e-01 1.61148214e+00
-1.34834856e-01 -6.22005343e-01 -7.73570776e-01 -5.73783040e-01
-3.16436380e-01 1.21748316e+00 -4.25913125e-01 -6.13947921e-02
-3.50125313e-01 -7.30156183e-01 7.31504917e-01 5.99922538e-01
5.11746347e-01 -9.34364617e-01 -7.95144975e-01 1.62472010e-01
-4.12188619e-01 -1.14722049e+00 -1.75032809e-01 3.35403048e-02
-8.69106829e-01 -1.01074636e+00 -5.07771790e-01 -4.30565536e-01
3.88280720e-01 3.04378837e-01 8.08595121e-01 -3.94719169e-02
-3.86803776e-01 3.45036834e-01 -4.52726513e-01 -6.45442903e-01
-2.41781786e-01 -2.06302553e-02 -5.66561036e-02 -2.52416968e-01
5.77754736e-01 -6.14902616e-01 -2.62208641e-01 2.43808985e-01
-6.94675326e-01 1.75758645e-01 2.26347581e-01 7.16521621e-01
1.82851881e-01 2.30187252e-01 3.91755521e-01 -9.36774850e-01
1.00347066e+00 -3.17147166e-01 -4.23140079e-01 4.69231188e-01
-8.83965254e-01 1.65514842e-01 6.17268443e-01 -7.36078799e-01
-1.26511919e+00 -2.76318789e-02 1.51684582e-01 -4.54837322e-01
-1.32698983e-01 4.71828103e-01 1.97865348e-02 2.35962108e-01
9.19224024e-01 -8.94692540e-03 -4.50409263e-01 -5.10131717e-01
4.90483195e-01 8.04781556e-01 8.21266919e-02 -5.10960281e-01
6.34713471e-01 3.23148906e-01 -2.75506079e-01 -7.57258415e-01
-9.32874501e-01 -3.60830277e-01 -4.17500526e-01 -1.02717638e-01
7.10898161e-01 -8.63004088e-01 -6.13011241e-01 -2.23333705e-02
-1.00842690e+00 -5.33953726e-01 -1.40743807e-01 2.45530963e-01
-4.62051153e-01 3.94414395e-01 -3.99105012e-01 -9.93566871e-01
-3.43646973e-01 -9.44726825e-01 8.29607725e-01 3.04695189e-01
-6.13419175e-01 -1.21447480e+00 1.35678068e-01 6.21200860e-01
5.20425677e-01 9.22270492e-02 1.07128239e+00 -6.17779136e-01
-3.92542690e-01 -1.82247236e-01 -2.53729731e-01 2.70753559e-02
2.65317857e-01 7.90389255e-02 -1.20302987e+00 -3.92052196e-02
-3.25990677e-01 -5.65170228e-01 6.53111398e-01 2.06742570e-01
1.38219428e+00 -3.27071935e-01 -4.07586634e-01 1.54413879e-01
1.29854393e+00 3.37469190e-01 4.97654378e-01 3.62690330e-01
7.90150940e-01 4.91622508e-01 9.33745205e-01 7.18400896e-01
5.47912955e-01 7.11634934e-01 1.42242573e-02 1.62339061e-01
-2.51012385e-01 -5.01094997e-01 3.37770581e-01 6.87731087e-01
-4.09940809e-01 -6.44101351e-02 -1.06297457e+00 7.48525143e-01
-2.04375625e+00 -1.29821241e+00 1.12702757e-01 2.07753539e+00
9.48530138e-01 1.22138709e-01 1.10057630e-01 1.56946033e-01
6.48906529e-01 7.11783543e-02 -4.01925385e-01 -5.95171809e-01
2.76596457e-01 1.46355584e-01 2.13995457e-01 6.28156364e-01
-7.75889814e-01 1.15782094e+00 6.37115574e+00 1.14870334e+00
-1.32548857e+00 6.21834934e-01 3.15944105e-01 -5.56907296e-01
-3.77907008e-01 2.55202323e-01 -6.22239232e-01 4.91803616e-01
1.14424002e+00 -3.70712936e-01 6.26906395e-01 9.94881928e-01
6.51499271e-01 -1.85934663e-01 -1.04325485e+00 1.30710888e+00
4.40474004e-02 -1.43999982e+00 -1.00525163e-01 -2.82362193e-01
5.30179739e-01 -1.09123357e-01 -3.92076582e-01 7.21859872e-01
2.63077945e-01 -8.35352600e-01 4.85811800e-01 5.15125096e-01
6.82687402e-01 -8.02293897e-01 6.02448761e-01 5.56496799e-01
-1.02266669e+00 -3.14179212e-01 -3.97260606e-01 -4.79519665e-01
1.91786543e-01 5.46058893e-01 -1.09344423e+00 2.09090397e-01
6.96971238e-01 1.95681646e-01 -7.44321704e-01 6.69987440e-01
-1.20952509e-01 7.99520791e-01 -1.28965646e-01 -4.20635641e-01
-1.66399539e-01 1.93591654e-01 2.44034007e-01 1.44953990e+00
3.40926170e-01 1.08894564e-01 2.18003109e-01 6.56954825e-01
1.59388632e-02 3.55399758e-01 -5.88010192e-01 -4.53805923e-01
4.22391981e-01 1.44758475e+00 -7.75949001e-01 -6.06356561e-01
-4.26236093e-01 8.23861778e-01 4.47395623e-01 2.13504121e-01
-1.05817378e+00 -5.30063987e-01 4.16678578e-01 3.49832743e-01
1.73500434e-01 -1.77002460e-01 -4.63077992e-01 -1.17560554e+00
-2.63540924e-01 -6.38778269e-01 4.00976092e-01 -1.03875530e+00
-1.10435879e+00 3.01302969e-01 3.04103434e-01 -8.93766165e-01
-5.07887483e-01 -4.12191361e-01 -5.49990118e-01 4.83523697e-01
-1.09399343e+00 -1.22761321e+00 -4.99664634e-01 6.82771742e-01
7.49304175e-01 -2.33844906e-01 9.60532725e-01 3.43193918e-01
-4.60903823e-01 6.07871175e-01 -2.75726646e-01 7.59133836e-03
9.59727585e-01 -9.31487441e-01 -6.67964742e-02 7.05209494e-01
1.89695477e-01 9.14321780e-01 9.28825259e-01 -6.92862332e-01
-1.19923306e+00 -9.34265852e-01 1.05174494e+00 -5.42306364e-01
9.12611842e-01 -5.19648314e-01 -7.75535166e-01 5.66670299e-01
4.00840402e-01 -1.67498831e-02 9.50778127e-01 5.33574641e-01
-3.03899914e-01 -2.51739144e-01 -1.12637258e+00 8.32639515e-01
1.01071525e+00 -6.27110660e-01 -7.58265913e-01 3.47041667e-01
7.85864115e-01 -8.73274580e-02 -7.44412899e-01 2.31315151e-01
5.05209923e-01 -8.04719031e-01 6.64265990e-01 -6.28186703e-01
4.75485533e-01 -2.39455476e-01 -2.97459006e-01 -9.34824467e-01
-3.43468010e-01 -5.51983774e-01 -2.89603442e-01 1.49243963e+00
4.17377710e-01 -2.78924763e-01 6.41762733e-01 8.79457831e-01
3.60969335e-01 -5.26596546e-01 -4.97483283e-01 -7.05323398e-01
-2.42518112e-01 -1.06267703e+00 3.92871082e-01 1.07199717e+00
2.87219673e-01 7.22229481e-01 -4.97432858e-01 -1.01640120e-01
3.93342912e-01 2.98330814e-01 9.01324213e-01 -1.09900153e+00
-4.12606657e-01 -2.29637399e-01 -3.57216410e-02 -4.35934544e-01
1.20894924e-01 -6.49961710e-01 -1.92372561e-01 -1.63944435e+00
6.60882890e-01 -4.76778030e-01 -3.11085403e-01 9.71365511e-01
-1.46420345e-01 2.46962041e-01 3.89764190e-01 3.20772678e-02
-6.73644960e-01 3.10165554e-01 9.74988222e-01 -2.69719869e-01
-3.77187014e-01 -1.10800132e-01 -8.53604138e-01 9.57170010e-01
9.43347335e-01 -4.57451493e-01 -8.72346878e-01 -7.41716176e-02
3.32256705e-01 -1.25027150e-01 3.04878026e-01 -8.15589488e-01
2.61900365e-01 -5.63436925e-01 1.59205481e-01 -2.44531155e-01
4.84118640e-01 -7.75347829e-01 1.51828751e-01 3.61507297e-01
-3.85587007e-01 -2.84826458e-01 3.59527439e-01 3.67210954e-01
3.28818038e-02 -4.77081686e-01 6.59601986e-01 -2.71803141e-01
-9.10621762e-01 -1.19851306e-01 -2.70691127e-01 3.04974951e-02
1.24371314e+00 -2.42326632e-01 -1.05351932e-01 -6.51917994e-01
-5.90258539e-01 -1.04876950e-01 4.01964486e-01 6.43625319e-01
3.73465270e-01 -1.23044980e+00 -2.41979331e-01 -2.60909438e-01
3.66293252e-01 -4.71098632e-01 3.63649964e-01 6.87759280e-01
-9.03687999e-02 6.43621862e-01 -1.65368840e-01 -4.11439240e-01
-1.56686366e+00 8.74473751e-01 -1.59228995e-01 -4.63128388e-01
-3.34621996e-01 4.51614767e-01 -7.30792433e-02 -3.74982148e-01
2.20342040e-01 -3.97539794e-01 -2.16686055e-01 3.41394395e-01
6.30416036e-01 3.76148880e-01 -1.96526006e-01 -6.70094490e-02
-4.05209512e-01 4.45296526e-01 -2.28835526e-03 -2.19725400e-01
1.25050724e+00 -2.33508050e-01 2.06686154e-01 8.32581699e-01
8.14980805e-01 2.89190352e-01 -1.00773549e+00 -1.34300753e-01
-7.34719113e-02 -2.71614283e-01 5.87700307e-02 -9.98045206e-01
-6.72795057e-01 1.01571763e+00 7.06568778e-01 1.36686787e-01
1.22020042e+00 -1.09560965e-02 5.59148252e-01 4.76676434e-01
6.00899994e-01 -1.36308789e+00 1.36367247e-01 3.87921005e-01
6.52674079e-01 -1.35722280e+00 6.83917925e-02 -2.43706062e-01
-7.97054052e-01 9.46487367e-01 5.65159559e-01 1.80588692e-01
4.67001557e-01 3.52683604e-01 -3.38452272e-02 9.92841423e-02
-9.47558939e-01 -2.33960167e-01 9.43052843e-02 6.98603988e-01
7.58986354e-01 1.45229191e-01 -4.27865118e-01 7.61419296e-01
-9.92792845e-02 3.90602201e-01 4.86093998e-01 8.15742552e-01
-4.45716262e-01 -1.31482589e+00 -3.73369962e-01 3.80454361e-01
-4.11917537e-01 -3.02606106e-01 -2.94186354e-01 8.06658089e-01
2.30812371e-01 1.13438952e+00 -7.11178035e-02 -4.08170521e-01
3.25607747e-01 3.88767540e-01 5.45348704e-01 -7.41280556e-01
-2.18817353e-01 -9.68299527e-03 4.87149894e-01 -6.40450239e-01
-6.04412138e-01 -5.51186681e-01 -1.42546582e+00 -4.21206534e-01
-6.94167241e-02 -2.42389396e-01 5.14888406e-01 1.06778920e+00
4.21529353e-01 5.93435645e-01 1.63453713e-01 -5.22687376e-01
-3.14478844e-01 -1.24008143e+00 -7.21223950e-02 2.32934818e-01
-1.51331931e-01 -8.50637376e-01 -1.60502899e-03 3.37423563e-01] | [10.697067260742188, 7.9060378074646] |
020dfdde-d058-4aff-891b-f9fd5baf6dd2 | safety-of-autonomous-vehicles-a-survey-on | 2305.17941 | null | https://arxiv.org/abs/2305.17941v1 | https://arxiv.org/pdf/2305.17941v1.pdf | Safety of autonomous vehicles: A survey on Model-based vs. AI-based approaches | The growing advancements in Autonomous Vehicles (AVs) have emphasized the critical need to prioritize the absolute safety of AV maneuvers, especially in dynamic and unpredictable environments or situations. This objective becomes even more challenging due to the uniqueness of every traffic situation/condition. To cope with all these very constrained and complex configurations, AVs must have appropriate control architectures with reliable and real-time Risk Assessment and Management Strategies (RAMS). These targeted RAMS must lead to reduce drastically the navigation risks. However, the lack of safety guarantees proves, which is one of the key challenges to be addressed, limit drastically the ambition to introduce more broadly AVs on our roads and restrict the use of AVs to very limited use cases. Therefore, the focus and the ambition of this paper is to survey research on autonomous vehicles while focusing on the important topic of safety guarantee of AVs. For this purpose, it is proposed to review research on relevant methods and concepts defining an overall control architecture for AVs, with an emphasis on the safety assessment and decision-making systems composing these architectures. Moreover, it is intended through this reviewing process to highlight researches that use either model-based methods or AI-based approaches. This is performed while emphasizing the strengths and weaknesses of each methodology and investigating the research that proposes a comprehensive multi-modal design that combines model-based and AI approaches. This paper ends with discussions on the methods used to guarantee the safety of AVs namely: safety verification techniques and the standardization/generalization of safety frameworks. | ['Lounis Adouane', 'Dimia Iberraken'] | 2023-05-29 | null | null | null | null | ['autonomous-vehicles'] | ['computer-vision'] | [ 6.23801388e-02 3.23749065e-01 -3.20563018e-01 -9.70171988e-02
1.51370913e-01 -3.98568749e-01 9.02516663e-01 -6.79039657e-02
-3.23669940e-01 7.95062542e-01 -3.73710543e-01 -8.39571536e-01
-5.98142862e-01 -9.51337814e-01 -2.62951612e-01 -6.75581455e-01
1.10628024e-01 3.34399670e-01 4.27143961e-01 -8.18896174e-01
2.73129702e-01 1.04885185e+00 -2.42169762e+00 -5.39772332e-01
7.13998556e-01 1.02590311e+00 4.14547846e-02 3.41243386e-01
1.68437481e-01 4.80783015e-01 -3.32706034e-01 -1.23032480e-01
2.35970482e-01 -1.07969858e-01 -5.85178614e-01 -1.71422750e-01
-5.89865148e-02 -2.70108283e-01 -1.05179965e-01 7.73856878e-01
1.75348356e-01 4.36540127e-01 8.60973775e-01 -2.07720327e+00
2.78132617e-01 3.32444394e-03 9.06325430e-02 -7.51018003e-02
1.29714847e-01 5.12584329e-01 6.23673856e-01 -3.50565195e-01
3.77017260e-01 9.06095743e-01 2.70968109e-01 8.73126090e-01
-9.64901149e-01 -5.14726460e-01 3.40300411e-01 5.82317054e-01
-1.40694034e+00 -5.57591796e-01 4.78944153e-01 -5.40351987e-01
1.11862195e+00 5.42821229e-01 7.72998989e-01 9.54469144e-01
5.93203843e-01 3.44952881e-01 8.34504783e-01 -2.57281423e-01
5.37451267e-01 2.48221084e-01 1.94096610e-01 1.27835497e-01
7.28963196e-01 7.01111019e-01 8.59455541e-02 1.08659074e-01
-6.38014898e-02 -4.90653127e-01 1.79396182e-01 -6.59322917e-01
-5.06748319e-01 8.46854866e-01 -1.21213354e-01 3.37623388e-01
-4.72929478e-01 6.96317703e-02 6.36434197e-01 2.70303816e-01
-2.94778705e-01 1.65104315e-01 -2.26909235e-01 -1.18749134e-01
-4.58098084e-01 5.78320861e-01 7.27756023e-01 8.15088868e-01
4.99359936e-01 5.82039714e-01 2.64708489e-01 1.03391655e-01
8.37584078e-01 7.49751210e-01 -3.13295990e-01 -1.04467499e+00
-9.84353721e-02 4.38151747e-01 2.37579644e-01 -1.10576475e+00
-7.24896550e-01 -8.92737582e-02 -6.26290739e-01 1.10181272e+00
6.31760806e-02 -2.23424673e-01 -6.11020267e-01 1.71579123e+00
5.14959812e-01 -2.78208941e-01 3.07620138e-01 5.48172414e-01
3.62723678e-01 6.84907079e-01 4.09615248e-01 -5.90270936e-01
1.13648415e+00 -3.31566274e-01 -1.05663741e+00 -2.58192003e-01
5.09209216e-01 -4.31393713e-01 6.84167564e-01 2.85295963e-01
-7.32596874e-01 -4.12092924e-01 -1.39984322e+00 5.56826115e-01
-5.75888276e-01 -4.76045966e-01 3.25627923e-01 1.00816810e+00
-7.94005275e-01 5.89930601e-02 -7.36474454e-01 -5.28785169e-01
-2.18458012e-01 4.49884802e-01 -1.96722716e-01 4.25359040e-01
-1.70767534e+00 1.81077242e+00 2.80671388e-01 2.70716101e-01
-1.01821232e+00 -7.28127122e-01 -9.94474828e-01 -3.45452547e-01
4.56025660e-01 -5.75618029e-01 1.21215701e+00 -4.94989097e-01
-1.55946112e+00 3.89873087e-01 1.47881821e-01 -8.21192741e-01
6.09889209e-01 -9.55718830e-02 -7.97430813e-01 -2.77956929e-02
5.68580860e-03 3.20489764e-01 5.04336238e-01 -1.55940962e+00
-1.01003325e+00 5.22486940e-02 2.40677476e-01 -4.85590994e-02
7.30520338e-02 6.26118556e-02 2.24336639e-01 1.50136977e-01
-4.72685546e-01 -1.11614895e+00 -4.90768194e-01 -2.06450522e-01
-9.48038846e-02 -2.61000365e-01 1.14927864e+00 -2.53863912e-02
1.51189017e+00 -2.04021215e+00 1.99303194e-03 5.23262978e-01
-2.23146513e-01 6.12124562e-01 1.82066023e-01 9.54687834e-01
1.68791696e-01 -1.27122015e-01 -1.95517257e-01 8.36519748e-02
1.61135525e-01 6.32402658e-01 -6.18441880e-01 4.41455126e-01
2.15163440e-01 3.85051161e-01 -5.72698116e-01 -3.92582178e-01
1.06242740e+00 4.68695939e-01 -4.33466695e-02 9.85458121e-02
-2.48345211e-01 4.80991900e-01 -6.17969692e-01 3.48677516e-01
5.56255937e-01 8.61237884e-01 -1.04488181e-02 -5.82121573e-02
-7.01633692e-01 -9.54163894e-02 -1.26545835e+00 5.98483920e-01
-5.13256669e-01 5.16650021e-01 2.98539132e-01 -9.15539980e-01
7.89885163e-01 5.86645007e-01 6.64782107e-01 -8.98608148e-01
5.04145920e-01 5.39494991e-01 1.18478395e-01 -6.86591864e-01
4.92826790e-01 -1.47110522e-01 -3.43221605e-01 5.52678891e-02
-5.33643901e-01 -4.78638500e-01 2.19056904e-01 7.26395845e-02
6.54543400e-01 4.74530086e-02 3.48923266e-01 -4.35269445e-01
1.36155796e+00 1.06573410e-01 5.60209811e-01 2.39873454e-01
-6.47127092e-01 -2.71082819e-01 6.19363666e-01 -2.86315680e-01
-8.58695686e-01 -7.56792903e-01 -3.22595298e-01 4.07945961e-01
4.62939352e-01 -2.66694516e-01 -6.08343005e-01 -5.68146169e-01
1.16368204e-01 1.45460975e+00 -6.21779084e-01 -6.15249336e-01
-4.55567241e-01 -2.25950122e-01 5.85163295e-01 2.58085072e-01
2.02332482e-01 -7.72449255e-01 -1.32421589e+00 1.54778123e-01
1.28432363e-01 -9.09054041e-01 3.75173301e-01 2.10887403e-04
-3.75555515e-01 -1.10044861e+00 3.63122821e-02 -4.23331112e-01
3.03488970e-01 2.80616194e-01 5.20731807e-01 8.73120651e-02
4.29021828e-02 5.48004270e-01 -2.30853647e-01 -6.85186386e-01
-7.94167995e-01 -1.47308081e-01 4.44950044e-01 1.55722219e-02
3.16135675e-01 -4.86102998e-01 -2.88253099e-01 8.49461138e-01
-9.25799251e-01 -1.70965374e-01 3.40375930e-01 3.55588913e-01
4.27821398e-01 4.03397411e-01 8.29690099e-01 -1.95073992e-01
5.19527316e-01 -5.31078458e-01 -9.83956754e-01 9.66910198e-02
-1.06099248e+00 -1.55065164e-01 4.87953693e-01 -5.07540852e-02
-9.70096946e-01 1.96892582e-02 -4.78806764e-01 -2.59379670e-03
-4.61313784e-01 2.55889118e-01 -6.10496342e-01 -3.87128353e-01
4.72670615e-01 -2.92360056e-02 6.04570091e-01 1.45532057e-01
2.34190285e-01 7.11536288e-01 9.94033962e-02 -3.62927884e-01
1.00953043e+00 4.57322598e-01 6.46504402e-01 -1.01521587e+00
-1.04549609e-01 -1.84534073e-01 -3.32612514e-01 -9.86217141e-01
9.20221508e-01 -2.77255595e-01 -1.16575277e+00 4.71973419e-01
-7.93848753e-01 -3.30666095e-01 -3.50331634e-01 4.04960930e-01
-9.03808296e-01 2.71245867e-01 -8.45479872e-03 -1.49837697e+00
-9.04815048e-02 -1.51033640e+00 5.75531065e-01 1.45163372e-01
-3.95025223e-01 -7.80184269e-01 2.43704244e-02 3.04493874e-01
7.05460966e-01 5.59248090e-01 7.51745045e-01 -2.90586561e-01
-4.03331101e-01 -3.76055092e-01 3.35628539e-01 3.50866288e-01
-6.60485551e-02 4.42654997e-01 -7.19405472e-01 -2.97238920e-02
1.20590046e-01 7.08140805e-02 1.76051363e-01 3.00058812e-01
2.19838023e-01 -7.14031607e-02 -4.55794930e-01 -2.95908172e-02
1.51011598e+00 8.18259478e-01 8.08674574e-01 7.83508122e-01
1.99202105e-01 1.36686122e+00 1.56274974e+00 2.74392337e-01
3.30103070e-01 9.69839931e-01 1.06866467e+00 3.89249533e-01
1.73034593e-01 8.09290707e-02 4.40602779e-01 1.08017325e-01
-2.55739808e-01 -9.77126062e-02 -9.40493107e-01 3.87156606e-01
-1.94517696e+00 -9.33706760e-01 -5.49528420e-01 2.28674865e+00
1.72403455e-01 3.62089097e-01 1.55050278e-01 5.46117485e-01
6.35421097e-01 -5.99656217e-02 -1.57964811e-01 -8.06102097e-01
1.11661464e-01 -4.07638758e-01 7.44513631e-01 8.20051551e-01
-1.01702392e+00 7.66600847e-01 6.30223656e+00 7.48105764e-01
-1.09784365e+00 -2.63402194e-01 6.32259175e-02 2.12526530e-01
-3.38529766e-01 1.41504750e-01 -8.17056358e-01 2.99189657e-01
1.50785601e+00 -3.69171619e-01 -5.31131960e-02 7.85862386e-01
8.03474963e-01 -4.37894166e-01 -7.04220057e-01 3.37512970e-01
-2.72555143e-01 -8.77018690e-01 -7.96772391e-02 1.71381727e-01
3.29377502e-01 -2.18159556e-01 2.99047921e-02 1.72272101e-01
-2.06724539e-01 -9.05474305e-01 1.25648558e+00 4.31115896e-01
4.33234543e-01 -1.22663522e+00 8.90490115e-01 2.44465351e-01
-1.20902789e+00 -2.15093538e-01 1.44186825e-01 -2.31610045e-01
9.43281054e-01 8.85898694e-02 -2.02768072e-01 8.53615761e-01
5.49625218e-01 1.19144663e-01 -6.97242096e-02 9.07647908e-01
-2.75513321e-01 1.96935475e-01 -3.30672562e-01 -3.04598570e-01
3.94204766e-01 -2.49941379e-01 9.65619743e-01 7.18747616e-01
2.72163562e-02 -8.87381379e-03 -1.59958109e-01 5.09728014e-01
1.30781698e+00 -2.90412195e-02 -1.08654666e+00 1.88123584e-01
5.05046546e-01 1.13472521e+00 -4.21093285e-01 5.94679937e-02
-4.30546671e-01 -3.84771228e-02 -5.77038586e-01 1.52931482e-01
-1.21060503e+00 -3.59424710e-01 1.28850448e+00 4.62198883e-01
-2.52885848e-01 -6.10679328e-01 -5.39882958e-01 -1.79130286e-01
-2.10140780e-01 -7.16385841e-01 1.47608459e-01 -4.60395396e-01
-6.87839627e-01 7.30750620e-01 7.33257771e-01 -1.49280679e+00
-4.46972996e-01 -6.03374779e-01 -5.10954618e-01 6.30368292e-01
-1.47143793e+00 -1.27681875e+00 -9.62744374e-03 3.83513600e-01
3.19145620e-01 -2.01772690e-01 5.69918513e-01 3.84834707e-01
-6.31290257e-01 2.86463976e-01 -2.85211116e-01 -7.87803710e-01
2.78100908e-01 -5.02403080e-01 3.29031609e-02 1.06524539e+00
-9.21246290e-01 4.75687176e-01 1.39916253e+00 -7.97656715e-01
-1.56060493e+00 -9.61634636e-01 9.75582182e-01 -3.30676138e-01
8.07566047e-01 -5.50526418e-02 -5.74791670e-01 4.49894786e-01
2.16949761e-01 -6.91508949e-01 3.12939674e-01 -4.16940838e-01
1.51961565e-01 -2.35204875e-01 -1.32038724e+00 1.01535118e+00
7.11596310e-01 -1.42980501e-01 -4.41618234e-01 -2.12921485e-01
5.29468596e-01 -2.06045341e-02 -6.66561067e-01 7.91366935e-01
6.31602168e-01 -7.63820887e-01 7.71296501e-01 -4.93116021e-01
-3.19282949e-01 -8.78313482e-01 -1.59405217e-01 -7.31881261e-01
-4.86299321e-02 -6.04904592e-01 5.20088784e-02 1.10159039e+00
4.21036988e-01 -1.19398201e+00 3.54114532e-01 1.06743455e+00
-6.34666681e-01 -6.18818045e-01 -1.39438450e+00 -9.82334197e-01
-5.34118712e-02 -9.03545618e-01 4.73798543e-01 3.78010422e-01
1.08535327e-02 6.33383244e-02 -4.17050064e-01 4.34687555e-01
5.79785109e-01 -4.51415658e-01 7.74565339e-01 -1.24521363e+00
4.84498948e-01 -5.59610128e-01 -9.60628152e-01 -6.79111332e-02
2.85185784e-01 -1.63434029e-01 1.80700362e-01 -1.62885928e+00
-6.93623662e-01 -3.00430655e-01 4.46411446e-02 3.05912644e-01
4.66851652e-01 -3.84656526e-02 4.52259071e-02 -1.09253474e-01
-1.81726038e-01 7.09716678e-01 8.15997303e-01 -8.04858506e-02
-1.14467695e-01 4.39265281e-01 -4.01243180e-01 6.16334856e-01
9.88236606e-01 -1.48454666e-01 -9.72992539e-01 1.86229542e-01
4.56589639e-01 -3.86966392e-02 4.10935432e-01 -1.23564982e+00
3.08573902e-01 -8.63983274e-01 -6.92783892e-01 -7.87556946e-01
2.89335072e-01 -1.63965511e+00 6.76865697e-01 1.03522718e+00
4.71852869e-02 -1.28924221e-01 4.85729873e-01 4.81616110e-01
-1.89488530e-01 -2.87228972e-01 1.02150059e+00 5.51129997e-01
-1.09582031e+00 -3.36356927e-04 -1.28024662e+00 -5.43603003e-01
2.16999841e+00 -5.86201370e-01 -3.87586057e-02 -1.86947152e-01
-5.21762669e-01 6.99342310e-01 3.53010446e-01 7.70613909e-01
4.20590997e-01 -1.10747778e+00 -3.95287752e-01 1.06799915e-01
2.27128312e-01 -3.12359571e-01 7.05824375e-01 9.95427430e-01
-5.26134014e-01 8.08207631e-01 -6.16981268e-01 -3.52653593e-01
-1.32010567e+00 8.60760152e-01 5.35981297e-01 4.05605078e-01
-4.50006664e-01 1.44940585e-01 7.14734495e-02 -4.57438007e-02
1.41796723e-01 2.22252071e-01 -8.32019329e-01 1.14243761e-01
5.83598733e-01 7.01551497e-01 1.55765682e-01 -1.22067750e+00
-6.97166443e-01 6.42878771e-01 4.06111062e-01 -3.41096461e-01
7.59048283e-01 -5.30745447e-01 -1.26488417e-01 3.01542401e-01
5.18439054e-01 -1.48294531e-02 -1.00768328e+00 7.36569941e-01
5.84176742e-02 -1.68658316e-01 -1.42492354e-02 -4.91258383e-01
-8.88996899e-01 5.39406121e-01 6.14889085e-01 3.99300545e-01
1.04867530e+00 -3.26190829e-01 4.95237529e-01 1.02944717e-01
7.67057240e-01 -1.40955782e+00 -5.59500277e-01 6.62730038e-01
1.03017640e+00 -6.31186068e-01 -3.31180900e-01 -6.31206393e-01
-6.71895742e-01 1.00351596e+00 8.04369867e-01 6.47944734e-02
9.45143044e-01 4.61149096e-01 1.25507832e-01 -1.96328759e-02
-6.13742292e-01 -4.68456715e-01 7.17141628e-02 9.69906807e-01
-1.79239959e-01 -7.69359758e-03 -9.65626657e-01 4.43279117e-01
5.14275357e-02 -1.63715646e-01 5.35239995e-01 1.12527215e+00
-7.45297849e-01 -1.41001070e+00 -5.15272200e-01 1.69036478e-01
1.58018968e-03 5.64193189e-01 -2.90089041e-01 1.09691262e+00
2.80928999e-01 1.56601036e+00 -2.25931808e-01 -8.28152061e-01
8.08070600e-01 -6.98754489e-02 -7.73179159e-02 -7.37253577e-02
-4.66723442e-01 -5.61154783e-01 6.24902427e-01 -6.86207652e-01
-2.82742798e-01 -8.61786306e-01 -1.55435705e+00 -5.91744602e-01
-2.46533930e-01 3.23468357e-01 9.96010423e-01 9.83856916e-01
1.65525720e-01 6.44886315e-01 7.05335736e-01 -6.82684422e-01
-6.03709519e-01 -3.36387813e-01 -3.53652447e-01 -1.43409938e-01
1.46962538e-01 -1.23737228e+00 -5.61120510e-01 -5.61402977e-01] | [5.6361494064331055, 1.347004771232605] |
c6c2dc28-730d-4816-b370-8ae1de749471 | perceptual-losses-for-real-time-style | 1603.08155 | null | http://arxiv.org/abs/1603.08155v1 | http://arxiv.org/pdf/1603.08155v1.pdf | Perceptual Losses for Real-Time Style Transfer and Super-Resolution | We consider image transformation problems, where an input image is
transformed into an output image. Recent methods for such problems typically
train feed-forward convolutional neural networks using a \emph{per-pixel} loss
between the output and ground-truth images. Parallel work has shown that
high-quality images can be generated by defining and optimizing
\emph{perceptual} loss functions based on high-level features extracted from
pretrained networks. We combine the benefits of both approaches, and propose
the use of perceptual loss functions for training feed-forward networks for
image transformation tasks. We show results on image style transfer, where a
feed-forward network is trained to solve the optimization problem proposed by
Gatys et al in real-time. Compared to the optimization-based method, our
network gives similar qualitative results but is three orders of magnitude
faster. We also experiment with single-image super-resolution, where replacing
a per-pixel loss with a perceptual loss gives visually pleasing results. | ['Li Fei-Fei', 'Alexandre Alahi', 'Justin Johnson'] | 2016-03-27 | null | null | null | null | ['nuclear-segmentation'] | ['medical'] | [ 7.88231611e-01 3.33007723e-01 2.43865281e-01 -6.15815580e-01
-9.93106604e-01 -3.96401852e-01 6.16099298e-01 -2.49918580e-01
-6.76772237e-01 6.35649741e-01 1.86806291e-01 -9.47867259e-02
2.80383795e-01 -9.24330235e-01 -1.19656658e+00 -3.91338766e-01
2.22062290e-01 1.81567445e-01 1.56058624e-01 -5.41235209e-02
1.15889840e-01 6.03827000e-01 -1.24349380e+00 6.41674876e-01
5.55508494e-01 1.02393270e+00 2.53599256e-01 9.72217739e-01
6.84726015e-02 7.70396531e-01 -4.63359743e-01 -4.71465290e-01
6.46888018e-01 -7.08814740e-01 -9.87964988e-01 3.11336130e-01
1.08401978e+00 -5.30471981e-01 -2.73020506e-01 1.22275138e+00
5.16306341e-01 8.18828717e-02 5.60239613e-01 -1.12371957e+00
-1.21542919e+00 2.91045934e-01 -4.46840465e-01 -1.06856665e-02
2.55525082e-01 2.67673671e-01 1.03110564e+00 -9.39078689e-01
7.19459593e-01 1.42992783e+00 7.31032491e-01 5.77589631e-01
-1.91932249e+00 -4.49201643e-01 -1.10339388e-01 9.62638855e-02
-1.07190251e+00 -4.52284127e-01 6.11878812e-01 -3.35986435e-01
1.01655614e+00 2.06312478e-01 3.71083200e-01 8.52096140e-01
1.40623078e-01 4.72209841e-01 1.26466751e+00 -5.81968129e-01
-1.73332959e-01 3.96931797e-01 -5.98753631e-01 6.57929957e-01
-4.21340615e-02 3.23928565e-01 -1.71988547e-01 3.01561236e-01
1.29246747e+00 -2.32700452e-01 -2.81308115e-01 -3.14874202e-01
-1.14974546e+00 6.97251320e-01 9.46935117e-01 2.34927639e-01
-3.35431337e-01 3.96744639e-01 1.48761615e-01 4.30815786e-01
5.64920843e-01 7.69358158e-01 -2.63410836e-01 2.12116599e-01
-1.13004923e+00 2.59573936e-01 4.81022298e-01 8.79460037e-01
8.75707984e-01 2.46048078e-01 -3.48176032e-01 9.49094176e-01
7.32001569e-03 3.16870183e-01 3.35294634e-01 -1.76065016e+00
2.40962178e-01 1.32802710e-01 3.67485791e-01 -9.30109084e-01
-1.91815779e-01 -3.67927819e-01 -7.01552808e-01 1.02023840e+00
6.07718289e-01 -1.28741473e-01 -1.17803073e+00 1.80156088e+00
-5.86743690e-02 -7.91738480e-02 7.01131746e-02 9.57789481e-01
5.65997899e-01 7.18363285e-01 -2.13471632e-02 9.28259641e-02
1.13735867e+00 -1.05803728e+00 -4.26066011e-01 -1.80002764e-01
1.74661100e-01 -9.85255003e-01 1.65109336e+00 4.80488479e-01
-1.68555653e+00 -7.99478531e-01 -1.00551283e+00 -4.57899213e-01
-1.94519237e-01 1.91353008e-01 1.48865029e-01 4.08747315e-01
-1.57206810e+00 9.65424597e-01 -6.15283847e-01 -3.81664187e-01
6.10072732e-01 3.57342333e-01 -4.41607207e-01 4.08343151e-02
-7.66323924e-01 9.66875911e-01 1.60947844e-01 -1.37771785e-01
-7.36956477e-01 -9.35312271e-01 -8.72230232e-01 1.09037623e-01
-9.40322652e-02 -8.57950091e-01 1.34687996e+00 -1.72219861e+00
-1.78872788e+00 1.18333244e+00 7.73863271e-02 -5.06244302e-01
6.25307858e-01 -1.68746516e-01 -2.73352295e-01 1.33483022e-01
1.53439999e-01 1.22552919e+00 9.22123551e-01 -1.52461159e+00
-6.04480982e-01 1.30281925e-01 3.92168730e-01 2.08716795e-01
-3.78755540e-01 1.88733563e-01 -3.00570577e-01 -8.62853527e-01
-2.40565136e-01 -7.80235708e-01 -2.57729739e-01 6.03555620e-01
-3.69325131e-01 3.77443671e-01 6.59734964e-01 -8.37014854e-01
5.69081783e-01 -2.01392984e+00 1.49293363e-01 -1.96410995e-03
6.12936839e-02 2.24858522e-01 -5.77804983e-01 6.15368560e-02
-3.36522102e-01 1.10913634e-01 -5.38231730e-01 -4.65793699e-01
-7.06721544e-02 1.47161216e-01 -3.59837174e-01 2.04973683e-01
5.30058146e-01 9.01347339e-01 -7.75714576e-01 -1.72758371e-01
2.97660053e-01 8.16162288e-01 -8.62175286e-01 3.26809883e-01
-2.49677062e-01 3.78495753e-01 1.83167458e-01 3.85647044e-02
5.22695005e-01 -3.09101880e-01 -2.47129858e-01 -5.39027929e-01
1.02295727e-01 1.66623518e-01 -8.26837838e-01 1.74494827e+00
-8.66560578e-01 1.07663465e+00 2.50075877e-01 -7.14983046e-01
7.76570857e-01 2.60337174e-01 2.20135152e-01 -8.52890790e-01
1.02710985e-01 1.15592822e-01 -5.00264093e-02 -1.69684619e-01
5.06247103e-01 -4.99562740e-01 2.06109151e-01 4.54332471e-01
2.40803435e-01 -5.08458793e-01 2.85796504e-02 5.36426902e-02
1.06611776e+00 4.68487144e-01 -2.49576509e-01 -8.07445124e-02
4.20891076e-01 -3.17311217e-03 3.53828549e-01 5.93878269e-01
7.08769113e-02 1.24965918e+00 3.71884853e-01 -5.15501976e-01
-1.68571365e+00 -1.14310038e+00 2.10274141e-02 9.89159822e-01
-5.13687618e-02 -2.07737554e-02 -1.05242383e+00 -4.25262541e-01
-1.35750160e-01 8.53445649e-01 -6.76695228e-01 -1.12464659e-01
-6.88584626e-01 -4.47322130e-01 5.27420640e-01 6.87889040e-01
6.51728928e-01 -1.32941318e+00 -5.35664320e-01 2.65709788e-01
-5.62975602e-03 -1.22305512e+00 -6.34926200e-01 -4.30513918e-02
-7.45048642e-01 -6.76648259e-01 -9.38510239e-01 -8.78225923e-01
1.04870152e+00 -8.20440501e-02 1.23358250e+00 -3.32512259e-02
-4.67481911e-01 1.65162459e-01 3.58841047e-02 -9.48818699e-02
-6.81035578e-01 -2.10532978e-01 -2.22175911e-01 9.42173004e-02
-1.80011258e-01 -7.65109241e-01 -7.20916092e-01 1.70427725e-01
-1.02798760e+00 3.89303774e-01 6.36217415e-01 9.48618114e-01
6.88135684e-01 -4.95760590e-02 2.73357660e-01 -1.07433510e+00
5.53830802e-01 1.52551010e-01 -6.95186973e-01 -5.79677783e-02
-6.35672390e-01 3.92585427e-01 8.60585213e-01 -4.30470675e-01
-1.30081236e+00 2.02463761e-01 -2.99692184e-01 -5.42838216e-01
-1.51637912e-01 -5.23087680e-02 -7.85111561e-02 -3.75470608e-01
1.00839221e+00 -1.26155233e-02 -1.77327059e-02 -3.72047424e-01
8.97746682e-01 2.52217174e-01 9.38272357e-01 -4.60749686e-01
9.64670062e-01 5.40740311e-01 -1.18306145e-01 -2.90087312e-01
-8.76873910e-01 2.29825020e-01 -4.98695105e-01 -1.69120617e-02
1.09918404e+00 -7.50255346e-01 -4.12245125e-01 3.26578766e-01
-1.27453434e+00 -8.56566310e-01 -8.30122232e-01 3.40738177e-01
-1.02967083e+00 -1.45159382e-02 -7.99676180e-01 -2.09735677e-01
-1.39579833e-01 -1.13170934e+00 9.99618232e-01 1.34998381e-01
-1.38234079e-01 -1.03360283e+00 -8.58931094e-02 2.36406863e-01
7.34274626e-01 3.28306705e-01 9.03672755e-01 -1.01808995e-01
-6.49193764e-01 7.55607337e-02 -7.97047615e-01 8.38585615e-01
2.11761862e-01 -8.84659588e-02 -1.14046693e+00 -3.62364233e-01
-9.70863253e-02 -4.56339687e-01 8.47446680e-01 4.82790053e-01
1.48477507e+00 -5.41340530e-01 2.02310383e-01 1.08145380e+00
1.52151597e+00 -1.05179437e-01 8.24663520e-01 3.41501087e-01
8.37800980e-01 5.42159855e-01 1.68915689e-01 -8.95794630e-02
2.85595357e-01 7.58651316e-01 2.72519946e-01 -8.24068367e-01
-8.78331661e-01 -2.38878652e-01 3.67963970e-01 1.74810871e-01
-1.70669794e-01 -8.42442811e-02 -5.59430659e-01 5.41972458e-01
-1.58047426e+00 -9.90568638e-01 1.97356313e-01 2.15610123e+00
1.09564424e+00 3.05326581e-01 8.20342228e-02 -3.81497107e-02
5.47674239e-01 5.03605679e-02 -5.67213178e-01 -6.44495726e-01
-2.07245186e-01 3.74883711e-01 7.54415512e-01 9.41201687e-01
-9.93093312e-01 1.02980721e+00 6.99747849e+00 6.89189911e-01
-1.43889344e+00 1.02427572e-01 1.08098876e+00 -2.39720494e-01
-4.45868045e-01 -1.66891709e-01 -2.89358288e-01 2.05649778e-01
8.50584984e-01 -1.21294677e-01 6.98286474e-01 4.85402763e-01
3.13862801e-01 -1.09421881e-02 -1.33266544e+00 9.18853939e-01
1.58900440e-01 -1.46924937e+00 2.12965310e-01 -1.26731191e-02
9.62666035e-01 -7.95610622e-03 4.11987811e-01 -1.24896318e-01
7.56985366e-01 -1.30890882e+00 9.20197606e-01 5.71610749e-01
1.18277478e+00 -5.84588468e-01 3.60464007e-01 -1.22407004e-01
-8.40153933e-01 1.44832388e-01 -3.46552402e-01 2.42844876e-02
2.52319664e-01 4.30629909e-01 -7.67875493e-01 1.67223841e-01
8.61575842e-01 6.34037137e-01 -5.49193621e-01 8.79181206e-01
-3.97551864e-01 4.01733637e-01 -2.40227252e-01 5.48508763e-01
1.52664378e-01 -2.48592362e-01 4.34331328e-01 1.30807197e+00
2.83565313e-01 -1.29899323e-01 -2.28522763e-01 1.51607788e+00
-3.52085143e-01 -1.73953369e-01 -5.93061030e-01 1.39727816e-01
6.81738257e-02 1.44158399e+00 -6.90500915e-01 -3.59739095e-01
-3.48892063e-01 1.50609159e+00 3.20907623e-01 6.34891868e-01
-5.86064398e-01 -6.43607259e-01 4.36831236e-01 4.00989741e-01
4.76103008e-01 -2.43438184e-02 -5.77502370e-01 -9.07290876e-01
4.26506950e-03 -6.84444726e-01 -1.16647482e-01 -1.31829739e+00
-1.18899250e+00 8.29192579e-01 -1.04843698e-01 -9.83774960e-01
-3.80892545e-01 -7.87923813e-01 -6.70680404e-01 1.17196488e+00
-1.50179899e+00 -1.14876926e+00 -3.28941464e-01 5.49018443e-01
4.87716436e-01 5.16521856e-02 7.38535702e-01 4.05968308e-01
-1.34047449e-01 6.36465430e-01 -1.08483121e-01 2.46716186e-01
7.29161322e-01 -1.52283370e+00 6.64688408e-01 9.13605809e-01
2.06216782e-01 1.72112763e-01 8.57520163e-01 -1.69409215e-01
-8.16094816e-01 -1.34874380e+00 7.59737790e-01 -5.04775524e-01
3.59699905e-01 -3.89360160e-01 -1.02056932e+00 8.14909816e-01
6.35145903e-01 3.05026114e-01 1.65291622e-01 -3.00826818e-01
-5.52753448e-01 -1.57331020e-01 -1.38846719e+00 6.27562046e-01
1.05548096e+00 -7.38839030e-01 -3.43986124e-01 2.61745036e-01
8.56321156e-01 -3.76355320e-01 -8.77739608e-01 2.12722227e-01
5.63835919e-01 -1.09039760e+00 1.16555977e+00 -5.41168630e-01
7.93081284e-01 -3.29736888e-01 3.00975051e-02 -1.60145104e+00
-4.95134830e-01 -5.17618120e-01 5.68526387e-01 1.02019846e+00
6.90528750e-01 -3.95188332e-01 7.37753391e-01 6.67061687e-01
-3.42508964e-02 -4.24245119e-01 -5.95007241e-01 -7.26492167e-01
3.72035205e-01 -1.01761520e-01 3.45594108e-01 7.07010210e-01
-5.31366825e-01 3.18784475e-01 -2.86402822e-01 -1.68232843e-01
7.19068348e-01 -7.82074556e-02 6.99590921e-01 -7.21182346e-01
-5.71599305e-01 -6.57844365e-01 -3.41542393e-01 -9.89581466e-01
2.25285478e-02 -8.20223451e-01 2.29134217e-01 -1.62169814e+00
1.16223894e-01 -4.87993449e-01 -2.33536020e-01 6.24957025e-01
-2.86957864e-02 8.36551607e-01 3.85060698e-01 -4.76824678e-02
-2.59972453e-01 3.48176777e-01 1.48078704e+00 -1.36904627e-01
-5.53792808e-03 -2.74115801e-01 -6.69965684e-01 7.58783102e-01
7.03471363e-01 -3.96254838e-01 -4.29025859e-01 -7.96908319e-01
1.97302848e-01 -2.00522140e-01 6.61269724e-01 -1.01938641e+00
4.13501076e-03 -9.98807251e-02 6.67306364e-01 -3.08382418e-02
5.05818665e-01 -7.69506216e-01 2.59236991e-01 3.00698847e-01
-7.54659772e-01 -1.03339162e-02 3.30837160e-01 1.67241007e-01
-1.73273161e-01 3.93193327e-02 1.22728181e+00 -1.20189972e-01
-3.94361436e-01 2.25475997e-01 -4.70556878e-02 1.80445332e-02
6.22085929e-01 -4.21485692e-01 -2.19999462e-01 -6.82047904e-01
-9.12073851e-01 -2.58388609e-01 7.66752124e-01 2.96714783e-01
6.81622326e-01 -1.57314074e+00 -9.34445560e-01 4.41674531e-01
-1.94797888e-01 -1.18145674e-01 -1.49132952e-01 5.94095707e-01
-7.53430426e-01 -4.85240594e-02 -6.49477363e-01 -4.82704669e-01
-1.13136268e+00 4.40678805e-01 7.15400517e-01 -1.00015722e-01
-6.23451173e-01 1.04052234e+00 5.14738321e-01 -4.76195961e-01
1.55664086e-01 -3.78352821e-01 3.25492769e-01 -4.97787088e-01
4.83636916e-01 1.94903128e-02 -1.36620015e-01 -6.25012815e-01
8.74091685e-03 6.30152643e-01 1.13265086e-02 -6.57148302e-01
1.45583439e+00 -9.91151482e-02 5.41609451e-02 2.15011597e-01
1.45550537e+00 -1.50639683e-01 -1.95052946e+00 -4.84247923e-01
-4.06291395e-01 -6.32425129e-01 2.37557873e-01 -1.07125366e+00
-1.34555042e+00 1.04386795e+00 7.44480789e-01 8.45357031e-02
1.35213625e+00 -1.91137895e-01 7.01704741e-01 9.53538939e-02
8.57967362e-02 -9.79702473e-01 4.22343493e-01 1.39254466e-01
1.30845380e+00 -1.26038563e+00 -1.48457468e-01 -3.35840404e-01
-4.34030205e-01 1.00203121e+00 5.88065922e-01 -6.29933894e-01
4.24473077e-01 3.65491420e-01 2.42619604e-01 8.44585747e-02
-4.99035716e-01 -6.23075590e-02 4.65986162e-01 6.51240289e-01
4.60684568e-01 -6.50835484e-02 1.34119570e-01 1.83507595e-02
-4.45001453e-01 1.34075716e-01 4.89608228e-01 5.95810652e-01
-3.14228505e-01 -1.33408177e+00 -3.04832101e-01 3.14866543e-01
-3.96259457e-01 -2.71669567e-01 -3.02719980e-01 6.39963329e-01
1.95473582e-01 5.68616450e-01 3.27256709e-01 -2.68633217e-01
4.12647694e-01 -2.85524994e-01 9.77096319e-01 -6.67222261e-01
-6.04253352e-01 3.81586254e-02 1.78424753e-02 -8.03786218e-01
-4.67813969e-01 -4.11767274e-01 -1.21637022e+00 -2.76207477e-01
8.30488130e-02 -3.17229122e-01 5.55773377e-01 5.27773917e-01
3.17847729e-01 7.41075516e-01 6.29474759e-01 -1.18934441e+00
-3.47876221e-01 -7.54079640e-01 -3.91747028e-01 8.35366905e-01
5.24446249e-01 -2.76949435e-01 -2.22173855e-01 6.35929704e-01] | [11.510997772216797, -0.6817985773086548] |
124df5ee-f179-4882-b36c-0568f315c695 | semeval-2012-task-7-choice-of-plausible | null | null | https://aclanthology.org/S12-1052 | https://aclanthology.org/S12-1052.pdf | SemEval-2012 Task 7: Choice of Plausible Alternatives: An Evaluation of Commonsense Causal Reasoning | null | ['Andrew Gordon', 'Zornitsa Kozareva', 'Melissa Roemmele'] | 2012-07-01 | null | null | null | semeval-2012-7 | ['commonsense-causal-reasoning'] | ['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] | [-7.195794582366943, 3.7270984649658203] |
e82d27f3-7f35-4234-bd78-b9eb40e4a3bb | exploiting-social-relations-and-sentiment-for | null | null | https://aclanthology.org/D14-1120 | https://aclanthology.org/D14-1120.pdf | Exploiting Social Relations and Sentiment for Stock Prediction | null | ['Sinno Jialin Pan', 'Jianfeng Si', 'Huayi Li', 'Bing Liu', 'Qing Li', 'Arjun Mukherjee'] | 2014-10-01 | null | null | null | emnlp-2014-10 | ['stock-market-prediction', 'stock-prediction'] | ['time-series', 'time-series'] | [-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] | [-7.2714338302612305, 3.610337495803833] |
8dd7ea04-2e8c-4dff-af79-19e29ec91d7d | a-benchmark-study-of-contrastive-learning-for | 2210.12314 | null | https://arxiv.org/abs/2210.12314v1 | https://arxiv.org/pdf/2210.12314v1.pdf | A Benchmark Study of Contrastive Learning for Arabic Social Meaning | Contrastive learning (CL) brought significant progress to various NLP tasks. Despite this progress, CL has not been applied to Arabic NLP to date. Nor is it clear how much benefits it could bring to particular classes of tasks such as those involved in Arabic social meaning (e.g., sentiment analysis, dialect identification, hate speech detection). In this work, we present a comprehensive benchmark study of state-of-the-art supervised CL methods on a wide array of Arabic social meaning tasks. Through extensive empirical analyses, we show that CL methods outperform vanilla finetuning on most tasks we consider. We also show that CL can be data efficient and quantify this efficiency. Overall, our work allows us to demonstrate the promise of CL methods, including in low-resource settings. | ['Laks V. S. Lakshmanan', 'Muhammad Abdul-Mageed', 'AbdelRahim Elmadany', 'El Moatez Billah Nagoudi', 'Md Tawkat Islam Khondaker'] | 2022-10-22 | null | null | null | null | ['dialect-identification'] | ['natural-language-processing'] | [ 1.13600641e-01 3.45245898e-02 -1.86825752e-01 -3.32314223e-01
-1.04940116e+00 -9.98993754e-01 6.60697877e-01 4.41650152e-01
-5.92305124e-01 7.27859855e-01 2.66269088e-01 -3.69597793e-01
-6.60934970e-02 -5.06352544e-01 -4.49838758e-01 -5.97217858e-01
-2.19416581e-02 5.15855730e-01 -2.55813450e-01 -6.61773562e-01
4.28508282e-01 4.54107463e-01 -1.10084689e+00 4.92673248e-01
9.57743168e-01 6.20947838e-01 -3.45847428e-01 3.80120903e-01
1.24951489e-02 7.29411602e-01 -7.40809262e-01 -9.43958580e-01
-8.71033370e-02 -2.49005377e-01 -1.02732694e+00 3.98414060e-02
4.57924932e-01 -2.51279801e-01 3.39881837e-01 9.71521497e-01
5.31071126e-01 2.20383227e-01 9.47088063e-01 -1.25233746e+00
-9.33896780e-01 1.05263913e+00 -6.85064852e-01 -3.39782313e-02
6.13011301e-01 -4.87229936e-02 1.50165677e+00 -1.32263291e+00
6.05659604e-01 1.66924262e+00 7.85139799e-01 3.36248070e-01
-1.09162819e+00 -5.62111795e-01 3.16108912e-02 3.23235467e-02
-9.56447005e-01 -5.87667823e-01 5.83041668e-01 -9.90819409e-02
9.22648609e-01 1.38887659e-01 2.48281509e-01 1.08090305e+00
-2.16532230e-01 1.43547428e+00 1.51520979e+00 -9.08814609e-01
2.37999368e-03 -4.12556343e-02 3.86871278e-01 8.24629545e-01
7.85352290e-02 -6.74841106e-01 -8.01297545e-01 -1.63140014e-01
-5.65723069e-02 -6.74680531e-01 -2.20780075e-03 2.54207730e-01
-1.15350795e+00 1.29981124e+00 -5.52758109e-04 2.65480131e-01
7.88662583e-02 -9.12363157e-02 3.91946107e-01 5.57481587e-01
8.51968944e-01 7.40724027e-01 -6.63823247e-01 -3.10237080e-01
-8.23861718e-01 3.76027912e-01 1.11297059e+00 7.39281654e-01
5.34000814e-01 7.21257105e-02 9.30312723e-02 1.22014904e+00
2.83653885e-01 7.95503020e-01 3.15132648e-01 -8.33050191e-01
5.51737428e-01 4.53673333e-01 -5.77029698e-02 -8.75039697e-01
-5.39061904e-01 1.72849640e-01 -5.04751921e-01 -1.43593490e-01
9.52928782e-01 -4.68054265e-01 -4.07434791e-01 1.50521827e+00
1.28605679e-01 -2.61527419e-01 1.49180025e-01 5.43642998e-01
6.80585265e-01 7.64414191e-01 -9.70486104e-02 -9.74261835e-02
1.58680081e+00 -9.37294781e-01 -6.84373975e-01 -4.65357691e-01
8.33886623e-01 -1.30904651e+00 1.58537936e+00 7.44798839e-01
-9.72614884e-01 4.36141044e-02 -9.58799005e-01 -2.33346298e-01
-6.11428201e-01 2.77020127e-01 1.02906358e+00 1.09055352e+00
-6.79944754e-01 3.00078124e-01 -4.95913357e-01 -6.27633691e-01
4.39640820e-01 2.37377465e-01 -3.43851537e-01 -2.97855079e-01
-1.35120583e+00 1.09229040e+00 2.19437435e-01 1.74618721e-01
-4.03416157e-01 -3.79584998e-01 -9.63166416e-01 -1.72844470e-01
5.31188130e-01 -1.34792939e-01 1.34327114e+00 -1.23502326e+00
-1.58754539e+00 1.12171030e+00 -1.12591706e-01 -4.16519105e-01
1.55392259e-01 -7.47875452e-01 -2.24443033e-01 -1.60920583e-02
3.65457870e-02 6.90124214e-01 8.86502504e-01 -1.12821293e+00
-5.65442979e-01 -2.11675867e-01 2.10755572e-01 4.21815455e-01
-6.45629585e-01 8.09001923e-01 -2.52526700e-01 -1.02839029e+00
-2.56683618e-01 -1.29944503e+00 2.47578129e-01 -2.70184696e-01
-4.37044293e-01 -5.19315183e-01 6.61793649e-01 -7.63194621e-01
1.13619709e+00 -2.05621147e+00 1.52619675e-01 1.71219409e-01
-3.45733836e-02 3.66545588e-01 -3.13741684e-01 6.15386844e-01
3.27794999e-01 6.65513426e-02 -6.39198422e-01 -4.40151751e-01
3.77736926e-01 1.91859350e-01 -4.29541558e-01 5.07631481e-01
5.72790802e-01 1.11730719e+00 -9.77363646e-01 -2.89122730e-01
-1.63267627e-01 3.75485003e-01 -4.62613612e-01 -7.12804720e-02
-3.67957801e-01 1.22227453e-01 -1.40691444e-01 8.94252658e-01
5.31516314e-01 -1.22572638e-01 2.20130265e-01 2.86529958e-01
1.51736632e-01 2.66781121e-01 -8.21222126e-01 1.55051053e+00
-3.45463336e-01 9.06103134e-01 3.27687711e-02 -8.20498109e-01
6.34763002e-01 -2.81361211e-02 -5.48496470e-02 -4.23604757e-01
1.03639685e-01 1.91332251e-01 1.51533678e-01 -4.15076226e-01
7.46525884e-01 -1.63145006e-01 -3.56953532e-01 9.14775252e-01
6.04199357e-02 -2.48365536e-01 4.03874397e-01 2.41922274e-01
6.61242068e-01 1.94851756e-02 4.24568772e-01 -4.18373495e-01
6.92402482e-01 9.22638103e-02 -5.41233011e-02 6.79798901e-01
-1.92834809e-01 2.46834740e-01 7.99838305e-01 -5.16140126e-02
-6.83160722e-01 -9.80751991e-01 -1.75640091e-01 2.03061485e+00
-3.86294097e-01 -4.19274747e-01 -1.02708006e+00 -9.79109526e-01
4.20844462e-03 6.87605858e-01 -4.45910066e-01 2.10910723e-01
-6.46834075e-01 -1.36365759e+00 1.08175814e+00 1.05302028e-01
2.45844111e-01 -1.44747567e+00 -1.43819705e-01 6.40908554e-02
-6.72887191e-02 -1.27329731e+00 -2.39256918e-01 -3.82481255e-02
-5.14587760e-01 -1.06567442e+00 -7.07074881e-01 -1.01710057e+00
3.17778975e-01 8.96297172e-02 1.32086372e+00 -1.02990955e-01
-1.10128112e-01 4.61574614e-01 -8.08948278e-01 -6.96656764e-01
-6.24011755e-01 5.45746028e-01 5.78492507e-02 6.77326396e-02
7.15734720e-01 2.75434870e-02 -5.61293960e-02 -1.21768624e-01
-9.21341717e-01 -4.01831776e-01 4.97844845e-01 8.69245350e-01
-9.63468757e-03 -2.19541028e-01 9.40458596e-01 -1.44148397e+00
1.22412932e+00 -6.46363258e-01 -2.74157017e-01 5.12036383e-01
-4.47469056e-01 4.64609750e-02 5.05505741e-01 -2.39343747e-01
-1.08626199e+00 -1.26846150e-01 -3.28408659e-01 5.36709011e-01
-2.11304441e-01 9.63097334e-01 4.72149365e-02 7.24832788e-02
6.81462109e-01 2.24388242e-02 6.85588866e-02 -3.96503389e-01
7.68831432e-01 9.26969349e-01 2.74993330e-01 -7.66944170e-01
6.25171006e-01 2.88660109e-01 -1.29940763e-01 -1.28888965e+00
-1.33566415e+00 -2.16499954e-01 -6.89796984e-01 1.54123276e-01
7.29782820e-01 -8.77983391e-01 -5.79884231e-01 8.21724176e-01
-1.00284648e+00 -5.83233297e-01 2.57308573e-01 7.67089650e-02
-3.66902322e-01 6.72633469e-01 -1.04034293e+00 -8.64826500e-01
-5.06642699e-01 -9.11580384e-01 9.64557290e-01 4.09909226e-02
-4.24316883e-01 -1.20872772e+00 -8.76687765e-02 5.23763776e-01
2.70572633e-01 4.24899720e-02 1.25273311e+00 -9.27804589e-01
-4.15396579e-02 8.61674026e-02 -2.66657531e-01 4.14607465e-01
1.91458344e-01 2.31475860e-01 -1.19237864e+00 -3.14441651e-01
-3.45193684e-01 -1.04059350e+00 8.14949632e-01 1.53595120e-01
7.01052308e-01 -3.68805319e-01 2.89410502e-01 1.59143403e-01
8.56729865e-01 -2.62661546e-01 3.98776293e-01 4.24506247e-01
6.80474222e-01 8.27530921e-01 1.02371514e+00 3.18266362e-01
5.39189935e-01 4.10460025e-01 4.37672101e-02 -5.23832291e-02
-7.42123052e-02 9.77625847e-02 6.95843816e-01 8.73197973e-01
9.11898464e-02 -3.52742285e-01 -1.27995360e+00 4.93830621e-01
-1.88438439e+00 -7.61748970e-01 -4.30084653e-02 1.83975351e+00
1.19206715e+00 -1.29764467e-01 2.50615299e-01 1.33094475e-01
5.20127833e-01 1.54822946e-01 -1.90685868e-01 -8.02923083e-01
-5.51343083e-01 4.90474582e-01 6.23998828e-02 6.92445576e-01
-1.46463847e+00 1.71770656e+00 6.60959911e+00 1.13357604e+00
-8.49668741e-01 1.91528440e-01 7.29030609e-01 5.09215146e-02
-1.22914955e-01 -2.89199471e-01 -8.60218525e-01 2.29406014e-01
7.86786020e-01 2.62750298e-01 8.56026471e-01 5.41773677e-01
-1.39982581e-01 -1.42863765e-01 -9.77013707e-01 8.61910939e-01
5.47429562e-01 -8.99006486e-01 2.03532390e-02 -2.75144637e-01
7.62576640e-01 1.87308252e-01 3.41174036e-01 3.62025082e-01
4.36273396e-01 -1.27590847e+00 5.93342245e-01 -4.60136123e-02
7.07542121e-01 -1.19578898e+00 7.66491175e-01 2.81519830e-01
-5.61758697e-01 -5.63866533e-02 -3.81405741e-01 -1.42721727e-01
-2.89291814e-02 4.01247919e-01 -8.42482209e-01 2.38156512e-01
4.63775635e-01 7.99616516e-01 -8.69775236e-01 5.18808007e-01
-7.76461899e-01 1.17404675e+00 -2.63701409e-01 -3.73128682e-01
7.04984426e-01 -2.68015504e-01 3.96277219e-01 1.78890932e+00
-2.77364124e-02 -1.01794399e-01 5.79689927e-02 2.35226959e-01
-3.43556970e-01 7.11570442e-01 -6.49236083e-01 -3.63740772e-01
3.92747611e-01 1.30484343e+00 -7.99491644e-01 -1.29322419e-02
-6.26066267e-01 9.72344160e-01 7.05205560e-01 1.66963935e-01
-4.73940611e-01 -6.83454752e-01 5.24454832e-01 -3.12787831e-01
8.64457414e-02 -3.85853827e-01 -1.75523713e-01 -1.29080260e+00
-1.84543073e-01 -1.31963754e+00 7.16579795e-01 -6.48647845e-01
-1.88522112e+00 3.70563537e-01 -7.99540654e-02 -7.04068542e-01
-3.88636291e-01 -1.13524652e+00 -1.42794505e-01 5.53559959e-01
-1.64562953e+00 -1.54789388e+00 3.64185482e-01 6.62238479e-01
6.16693020e-01 -5.45916498e-01 1.04477704e+00 5.36141219e-03
-4.73547161e-01 9.02374089e-01 2.18093261e-01 4.34080243e-01
1.31076849e+00 -1.46587110e+00 2.98483551e-01 9.55220878e-01
6.02145255e-01 6.41512215e-01 3.61273497e-01 -3.38940471e-01
-1.42147052e+00 -7.42393553e-01 1.19058669e+00 -8.58889699e-01
1.20757759e+00 -5.32435477e-01 -7.20476925e-01 7.45384514e-01
4.68686074e-01 -2.69698143e-01 9.43375766e-01 6.45546079e-01
-7.00062454e-01 1.57909513e-01 -1.03710806e+00 7.33283579e-01
6.48411632e-01 -6.11874580e-01 -6.87932312e-01 6.11759424e-01
4.68099952e-01 -3.35611075e-01 -7.04880297e-01 9.35643725e-03
7.35156536e-01 -5.99371076e-01 8.68641317e-01 -8.51115286e-01
7.92822838e-01 1.94292143e-01 -1.93052426e-01 -1.46488428e+00
4.23353352e-02 -8.36075485e-01 -1.09155446e-01 1.43871248e+00
7.25582540e-01 -6.93292201e-01 5.73306322e-01 2.51512259e-01
-2.09137350e-02 -6.57669008e-01 -6.66308939e-01 -4.70590413e-01
7.93335259e-01 -6.87485278e-01 3.25526893e-01 1.42646515e+00
2.72990376e-01 7.77583182e-01 -4.16216999e-01 4.06646542e-02
5.42194784e-01 1.36253551e-01 5.27942002e-01 -1.11058748e+00
-2.62755483e-01 -5.94981372e-01 7.33477101e-02 -6.93610132e-01
8.48789155e-01 -1.05638742e+00 -1.29981965e-01 -1.26641655e+00
7.38031715e-02 -5.79905152e-01 -9.34145302e-02 8.79025161e-01
-5.41226447e-01 6.45707190e-01 4.26854372e-01 -2.66119856e-02
-7.23323166e-01 2.16850653e-01 9.71963525e-01 -1.61059931e-01
-9.97619256e-02 -1.38875008e-01 -8.84735465e-01 8.58369410e-01
1.11196744e+00 -5.12439251e-01 -3.15635502e-01 -5.22044241e-01
8.02922368e-01 -5.21174252e-01 -9.82783288e-02 -3.25138718e-01
-1.46617234e-01 -1.87557712e-01 1.74458504e-01 -2.05461204e-01
1.72343165e-01 -3.23132664e-01 -1.04624987e+00 2.65123039e-01
-2.96344161e-01 1.98356211e-01 3.40808213e-01 2.08200887e-01
-1.55428275e-01 -5.47777772e-01 6.62316263e-01 5.23206852e-02
-6.30498230e-01 -7.33034536e-02 -7.15728104e-01 6.76123023e-01
6.59222960e-01 2.04173058e-01 -5.07797003e-01 -5.56062579e-01
-5.15801668e-01 1.18860058e-01 2.15574726e-01 5.68724155e-01
4.42452461e-01 -9.62722957e-01 -1.18276644e+00 -5.72588034e-02
3.32989246e-01 -3.43275934e-01 -3.24405313e-01 7.70475745e-01
-6.15869045e-01 3.59235704e-01 -9.32729319e-02 -3.60605240e-01
-1.43679261e+00 2.67224491e-01 -1.83316499e-01 -3.00402939e-01
-2.60808825e-01 9.20173228e-01 -2.48637363e-01 -8.86908889e-01
8.07203948e-02 -3.21696959e-02 -3.22154552e-01 5.63765526e-01
5.64529538e-01 3.49180996e-01 1.38062865e-01 -8.03714037e-01
-3.56257021e-01 3.96956563e-01 -9.87031609e-02 -4.42066222e-01
1.45959604e+00 -2.32534036e-02 -5.86582839e-01 4.89624918e-01
8.65507543e-01 5.09920061e-01 -9.70265388e-01 -1.60100222e-01
4.69451338e-01 -1.59250885e-01 -2.81978667e-01 -1.12049651e+00
-6.41583502e-01 8.71828973e-01 6.33850768e-02 2.21112683e-01
1.05348337e+00 -9.36469957e-02 8.11470628e-01 1.04345536e+00
2.94058949e-01 -1.22443092e+00 1.30001187e-01 1.17674279e+00
8.72482955e-01 -1.50052345e+00 1.43396586e-01 -4.80951220e-01
-1.09732211e+00 1.15215516e+00 -1.82039086e-02 -1.42599165e-01
6.31977677e-01 2.78929859e-01 2.30880037e-01 -1.35767877e-01
-4.09560829e-01 -3.96954596e-01 3.63165528e-01 7.46437967e-01
7.91378081e-01 2.07617119e-01 -4.58812326e-01 6.17186368e-01
-5.71358204e-01 -5.46543062e-01 7.82949269e-01 9.97032583e-01
-4.33902711e-01 -1.25769699e+00 -4.55527455e-01 3.20525527e-01
-8.97013545e-01 -6.17851019e-01 -7.92595685e-01 9.85877931e-01
-1.81401551e-01 1.16286814e+00 -1.94258600e-01 1.34306058e-01
-7.00092316e-02 4.20187324e-01 8.08204532e-01 -7.52196312e-01
-8.51725221e-01 -2.50242259e-02 6.26313269e-01 -1.40298709e-01
-5.44346809e-01 -8.72384965e-01 -1.21398675e+00 -5.55009067e-01
-1.64236635e-01 -7.76999295e-02 5.21690071e-01 1.12045991e+00
-6.89883530e-02 -3.28784645e-01 3.20272326e-01 -6.56359255e-01
-5.38842022e-01 -1.06772482e+00 -6.71384633e-01 3.93713742e-01
1.81199849e-01 -2.66918987e-01 -2.48414487e-01 1.66777655e-01] | [11.00700569152832, 9.795551300048828] |
6ec5f0ff-9ec9-47ad-bea7-e156640b48b2 | a-deep-learning-approach-with-an-attention | 1805.05036 | null | http://arxiv.org/abs/1805.05036v1 | http://arxiv.org/pdf/1805.05036v1.pdf | A Deep Learning Approach with an Attention Mechanism for Automatic Sleep Stage Classification | Automatic sleep staging is a challenging problem and state-of-the-art
algorithms have not yet reached satisfactory performance to be used instead of
manual scoring by a sleep technician. Much research has been done to find good
feature representations that extract the useful information to correctly
classify each epoch into the correct sleep stage. While many useful features
have been discovered, the amount of features have grown to an extent that a
feature reduction step is necessary in order to avoid the curse of
dimensionality. One reason for the need of such a large feature set is that
many features are good for discriminating only one of the sleep stages and are
less informative during other stages. This paper explores how a second feature
representation over a large set of pre-defined features can be learned using an
auto-encoder with a selective attention for the current sleep stage in the
training batch. This selective attention allows the model to learn feature
representations that focuses on the more relevant inputs without having to
perform any dimensionality reduction of the input data. The performance of the
proposed algorithm is evaluated on a large data set of polysomnography (PSG)
night recordings of patients with sleep-disordered breathing. The performance
of the auto-encoder with selective attention is compared with a regular
auto-encoder and previous works using a deep belief network (DBN). | ['Martin Längkvist', 'Amy Loutfi'] | 2018-05-14 | null | null | null | null | ['sleep-staging', 'automatic-sleep-stage-classification'] | ['medical', 'medical'] | [ 1.57984480e-01 1.09946094e-01 -9.05456468e-02 -7.24605620e-01
-1.54616490e-01 4.51289602e-02 1.25351161e-01 2.00201824e-01
-6.71923399e-01 7.40853906e-01 3.05764019e-01 4.16463763e-02
-3.36670697e-01 -4.80413973e-01 9.48283374e-02 -8.68751585e-01
-4.58914340e-02 6.12796366e-01 3.39076310e-01 -2.41893589e-01
1.76517054e-01 2.97177702e-01 -1.86175966e+00 1.93439528e-01
6.22758269e-01 1.02079594e+00 5.71083605e-01 6.33826375e-01
3.53780799e-02 4.27727491e-01 -7.90708601e-01 7.48862326e-02
2.28484005e-01 -8.21093798e-01 -7.36763418e-01 5.64661361e-02
6.56665638e-02 7.95989335e-02 -1.25201806e-01 1.03041875e+00
6.12075746e-01 3.25936556e-01 3.58243108e-01 -9.03243363e-01
-2.39138767e-01 1.81616157e-01 5.17495163e-02 1.00763774e+00
7.45725483e-02 -1.08745135e-01 1.01313305e+00 -3.84616375e-01
1.33752540e-01 5.51919043e-01 4.45188135e-01 7.32570767e-01
-1.02584207e+00 -4.38987225e-01 -3.20392609e-01 6.10797524e-01
-1.19745588e+00 -5.09687364e-01 7.39838302e-01 -3.36425334e-01
1.22473061e+00 3.59282672e-01 1.18501413e+00 6.94586873e-01
5.39332449e-01 1.15029939e-01 1.13716793e+00 -4.11041021e-01
4.51797009e-01 5.21775305e-01 4.48969394e-01 7.43375480e-01
3.19647998e-01 1.25665320e-02 -4.77315724e-01 1.44349962e-01
3.32735687e-01 5.92426538e-01 -2.84818977e-01 -1.40063196e-01
-6.88476443e-01 8.54678810e-01 5.55750906e-01 9.89944994e-01
-5.81247151e-01 -2.84497887e-01 3.55143517e-01 5.51158011e-01
2.76401728e-01 7.61320472e-01 -5.12008548e-01 -3.80024701e-01
-1.44828153e+00 -2.22941518e-01 7.09835470e-01 4.04560626e-01
8.57766092e-01 -8.86329412e-02 -2.42847115e-01 7.80079603e-01
5.07406779e-02 7.10283145e-02 1.31988621e+00 -6.47070885e-01
1.76060461e-02 1.16958582e+00 -1.65636167e-01 -7.86442280e-01
-6.70522153e-01 -6.39142632e-01 -8.64846230e-01 2.89038867e-01
9.58922207e-02 8.87186527e-02 -8.99060130e-01 1.36831141e+00
-9.84890684e-02 -3.16849947e-01 -3.04391421e-02 8.50761890e-01
5.60296059e-01 5.64867139e-01 -2.65658051e-01 -2.93853343e-01
1.47071099e+00 -9.71884191e-01 -7.68221974e-01 -6.26773179e-01
2.02914044e-01 -3.76931727e-01 8.59214902e-01 5.86226702e-01
-9.42239642e-01 -7.52821982e-01 -1.15236974e+00 -9.59823132e-02
-5.01175702e-01 9.53600183e-02 5.24805367e-01 6.80032194e-01
-1.09815598e+00 9.15526390e-01 -1.09416044e+00 -6.47592604e-01
4.34038013e-01 9.60812986e-01 -4.65512365e-01 2.78144479e-01
-9.29949880e-01 1.08929443e+00 3.40598047e-01 1.37517676e-01
-6.44558132e-01 -1.84549347e-01 -6.74883962e-01 4.85927552e-01
-1.86263733e-02 -7.71243691e-01 8.94685566e-01 -1.16987753e+00
-1.35766292e+00 7.92745769e-01 -3.93935114e-01 -6.51077330e-01
-1.06957950e-01 -1.70717791e-01 -4.79823053e-01 2.74626464e-01
7.08513930e-02 3.71295750e-01 1.17869973e+00 -4.78094071e-01
-6.66298211e-01 -5.76408148e-01 -1.11523136e-01 2.17908815e-01
-6.47128284e-01 -7.99843168e-04 -1.81724906e-01 -2.42510140e-01
1.14803635e-01 -1.04235280e+00 -2.09700286e-01 -3.32870275e-01
-1.10000126e-01 -1.63338140e-01 5.98713279e-01 -4.33811069e-01
1.51178348e+00 -2.15576744e+00 3.66078019e-01 -4.06695344e-02
3.97346288e-01 3.65840822e-01 3.35993856e-01 3.24657798e-01
-1.75943285e-01 -1.83275566e-01 -7.62241110e-02 -4.53919232e-01
-4.28459436e-01 5.61681390e-01 1.88719362e-01 5.06016552e-01
6.97727129e-02 4.20516700e-01 -7.66623080e-01 -3.50255281e-01
2.52990901e-01 4.63906020e-01 -6.75703943e-01 4.65930551e-01
3.90333980e-01 3.63717049e-01 -3.18039805e-01 2.24815220e-01
4.69033606e-02 -3.72313857e-01 -1.56444162e-01 -1.68232411e-01
-1.52875498e-01 6.15000784e-01 -7.94746101e-01 1.47592521e+00
-6.06451154e-01 7.80781686e-01 -3.22296530e-01 -9.32597756e-01
8.61596107e-01 2.98828870e-01 4.74071920e-01 -5.87374926e-01
4.65752512e-01 6.85838163e-02 5.62073946e-01 -6.09988689e-01
2.69361466e-01 -7.09686995e-01 1.86967641e-01 3.22235912e-01
3.77341539e-01 1.55682415e-01 3.05807322e-01 -2.37415284e-01
1.38739598e+00 -4.48910594e-01 6.67293191e-01 -3.14436346e-01
7.06025898e-01 -1.52974784e-01 8.01472306e-01 4.51148659e-01
-3.81592005e-01 6.18221462e-01 2.19336659e-01 -7.34417617e-01
-8.07887018e-01 -5.66673219e-01 -3.30721587e-02 1.02620304e+00
-7.42469132e-02 -5.63168466e-01 -6.21551275e-01 -6.53293312e-01
-2.71823555e-01 6.35393858e-01 -1.00095916e+00 -5.46826303e-01
-2.04995975e-01 -7.73413897e-01 1.35926336e-01 5.21550834e-01
2.69289851e-01 -1.47656572e+00 -1.35522747e+00 2.29452372e-01
2.25727931e-01 -6.36249721e-01 -2.06926078e-01 9.34525251e-01
-1.20567608e+00 -1.11316454e+00 -3.76243472e-01 -5.86186290e-01
8.78772974e-01 2.36095682e-01 9.15885210e-01 1.48121238e-01
-4.56442237e-01 3.99946049e-03 -5.40978014e-01 -3.37567031e-01
-3.21309686e-01 3.05979013e-01 3.33023190e-01 1.60923392e-01
8.50343764e-01 -8.53227854e-01 -8.54912579e-01 2.75521696e-01
-6.89138830e-01 -8.44690725e-02 9.06948864e-01 1.00191689e+00
5.09502053e-01 2.95111507e-01 4.78908092e-01 -6.85991168e-01
5.06800950e-01 -4.50512797e-01 -2.22286060e-01 -2.00541154e-01
-8.16514432e-01 4.05379385e-01 9.10838544e-01 -9.38692763e-02
-6.21691287e-01 4.35758941e-02 -3.11937392e-01 -4.32266116e-01
-1.79262966e-01 1.20586611e-01 8.10164139e-02 1.15236081e-01
5.96184254e-01 5.29849350e-01 9.73306373e-02 -6.79784894e-01
-2.26777479e-01 8.44820976e-01 1.59817368e-01 2.78237641e-01
5.00469983e-01 3.63449067e-01 -3.19629796e-02 -9.88778055e-01
-1.04651201e+00 -9.02535021e-01 -8.88227880e-01 1.67186782e-01
9.96955156e-01 -5.68723500e-01 -4.49147731e-01 -1.04796013e-03
-5.00733435e-01 -1.08787306e-01 -6.59349322e-01 4.84862953e-01
-4.53699142e-01 1.44952476e-01 -1.72787100e-01 -6.61488831e-01
-6.32123470e-01 -1.18071520e+00 7.92343855e-01 5.58500290e-01
-4.43493187e-01 -8.59148264e-01 4.05552953e-01 2.88508385e-01
5.96706569e-01 -3.38709116e-01 7.84865379e-01 -1.00436628e+00
-1.13007151e-01 -4.22733039e-01 1.37376860e-01 6.74672365e-01
6.16702139e-01 -4.88803566e-01 -1.05739880e+00 -3.43744844e-01
7.08011806e-01 -6.66657239e-02 9.22426224e-01 2.83747315e-01
8.85988057e-01 -3.27832043e-01 -2.65735626e-01 7.15019464e-01
1.28100955e+00 4.29906040e-01 5.60064077e-01 5.34423947e-01
2.66624600e-01 2.44723931e-01 2.13000044e-01 3.17252189e-01
2.08091095e-01 4.58678395e-01 3.59822959e-01 9.95038673e-02
-4.60009426e-02 2.29953960e-01 2.38798246e-01 9.95997846e-01
-2.72005349e-01 1.49384707e-01 -5.31807423e-01 6.43214226e-01
-1.39909482e+00 -9.87038434e-01 3.03199381e-01 2.13760757e+00
6.14026845e-01 3.42654228e-01 2.24893048e-01 7.14922190e-01
3.20691407e-01 1.69006169e-01 -4.69735801e-01 -7.60677338e-01
3.01145166e-01 5.46470642e-01 2.25703970e-01 1.47946402e-01
-8.89965713e-01 4.24633414e-01 6.03221321e+00 2.23083884e-01
-1.33476663e+00 2.26132363e-01 2.76268214e-01 -3.90840292e-01
3.09444070e-01 -7.70563856e-02 -8.63614798e-01 7.41224349e-01
1.40856886e+00 -5.61093446e-03 4.92880881e-01 1.12630713e+00
2.77653694e-01 -2.85640568e-01 -1.13691437e+00 1.07036960e+00
2.90912241e-01 -9.00650680e-01 -3.08819622e-01 7.15907589e-02
3.60444218e-01 2.26542860e-01 -2.24803969e-01 3.82996082e-01
-3.97523165e-01 -9.00725722e-01 1.88059226e-01 7.23334074e-01
4.74713504e-01 -4.88884896e-01 1.04012048e+00 3.75560492e-01
-8.08261514e-01 -5.77909350e-01 -6.19893253e-01 -3.34218711e-01
-1.60562620e-01 5.45053899e-01 -1.18620098e+00 1.30458951e-01
8.35136533e-01 7.37089217e-01 -9.47565317e-01 1.29406357e+00
-1.50780424e-01 6.91014886e-01 -2.53421396e-01 -2.84196883e-01
1.87894359e-01 -6.01445623e-02 4.59184796e-01 9.63350236e-01
4.20851529e-01 1.19451180e-01 -2.96612352e-01 6.59537494e-01
1.72467560e-01 8.28372873e-03 -5.26729763e-01 -1.33846328e-01
-1.19327910e-01 1.42086482e+00 -8.04346025e-01 -2.20218435e-01
-4.67428148e-01 1.23713636e+00 3.10508400e-01 -2.66955853e-01
-3.07863235e-01 -4.81683195e-01 5.58447421e-01 4.00192797e-01
5.09269178e-01 1.67405214e-02 -1.32960796e-01 -9.60149407e-01
-2.11518556e-01 -5.20330071e-01 4.00326878e-01 -5.80522478e-01
-1.01047134e+00 1.11815512e+00 -2.32357442e-01 -1.21498656e+00
-3.36344033e-01 -4.49122995e-01 -7.05332935e-01 7.92119980e-01
-1.35015273e+00 -5.29377759e-01 -4.24486399e-01 5.34374535e-01
8.06742311e-01 -4.04053599e-01 1.00713813e+00 3.82734150e-01
-5.85091174e-01 2.88902730e-01 1.22175422e-02 -1.04975171e-01
5.67504466e-01 -1.57487488e+00 -1.67042851e-01 6.51963174e-01
1.62430018e-01 8.13507855e-01 6.87780261e-01 -3.91871780e-01
-1.14385855e+00 -7.83316851e-01 1.17235696e+00 -5.23802638e-01
2.43282378e-01 -1.67082414e-01 -8.11643183e-01 4.68359470e-01
1.79615006e-01 1.19059451e-01 1.06893277e+00 8.94637108e-02
2.79262215e-01 -6.47526681e-01 -1.03101575e+00 -1.48557173e-02
5.68126023e-01 -2.35457510e-01 -1.14422596e+00 -7.62705058e-02
-3.30633670e-03 8.79665390e-02 -6.01004660e-01 -2.24457700e-02
4.87113684e-01 -1.37148571e+00 5.05571365e-01 -4.12189424e-01
2.20141008e-01 -2.83123136e-01 1.42913193e-01 -1.40769207e+00
-5.17264903e-01 -5.40447831e-01 -2.06288427e-01 7.18956232e-01
1.63414717e-01 -3.97288501e-01 7.83710241e-01 4.08191711e-01
-3.82642776e-01 -9.45196092e-01 -1.01719475e+00 -5.09515584e-01
-5.75446904e-01 4.00563069e-02 3.32879186e-01 4.36465919e-01
7.16105327e-02 7.75320172e-01 -3.36473674e-01 -2.02943385e-01
1.42097443e-01 2.66055703e-01 3.53223652e-01 -1.64193952e+00
-3.25568050e-01 -3.21812361e-01 -9.36482191e-01 -4.81165260e-01
-2.54709244e-01 -8.18521500e-01 7.90954530e-02 -1.66054094e+00
3.68494600e-01 -1.60051346e-01 -7.85181880e-01 6.43776953e-01
-1.46277428e-01 3.61707300e-01 4.99110147e-02 1.88850343e-01
-6.05494618e-01 4.95754302e-01 9.63950932e-01 -3.96121806e-03
-5.73237538e-01 4.37226146e-01 -8.71101022e-01 7.98224092e-01
8.05524588e-01 -8.22914541e-01 -5.35061300e-01 2.97648329e-02
3.15912575e-01 -3.93106192e-02 2.16071028e-02 -1.53371835e+00
2.74203598e-01 3.26231390e-01 6.96630001e-01 -3.95653784e-01
5.86221218e-01 -1.01577914e+00 7.94925541e-02 5.43664932e-01
-1.74521759e-01 1.11614466e-01 3.38985696e-02 3.98972958e-01
-3.46276581e-01 -5.26654720e-01 8.63179982e-01 -3.51040155e-01
-6.77776933e-01 2.22726949e-02 -4.90832984e-01 -1.53741151e-01
7.77239919e-01 -6.23376012e-01 3.29896837e-01 -1.73664793e-01
-1.04136193e+00 -1.80917263e-01 3.48180830e-01 2.88570672e-01
4.89875823e-01 -8.76424968e-01 -1.99161828e-01 6.64919674e-01
5.01124300e-02 -3.34359348e-01 1.31863162e-01 9.76622701e-01
-3.15640241e-01 5.88875949e-01 -7.35526264e-01 -4.40064996e-01
-1.31449568e+00 6.00616217e-01 3.32642555e-01 -4.53738093e-01
-9.18168962e-01 7.54981756e-01 -1.54258937e-01 3.69735301e-01
3.98445055e-02 -6.31894469e-01 -6.74935162e-01 3.01674724e-01
7.33848453e-01 3.11137944e-01 5.22439659e-01 -6.70473397e-01
-4.60804045e-01 4.34084266e-01 -1.14599086e-01 3.53861004e-01
1.73484969e+00 -1.63090974e-01 -4.44986597e-02 6.38218343e-01
1.12860644e+00 -8.34587216e-02 -8.11010718e-01 2.47293174e-01
-7.46572390e-02 -3.59457761e-01 3.48808318e-01 -7.48517513e-01
-1.15435183e+00 1.10020447e+00 1.08191431e+00 5.17648757e-01
1.49267411e+00 -1.14327863e-01 6.93353653e-01 5.43051660e-01
1.69028208e-01 -1.01768875e+00 -1.25824764e-01 3.17851663e-01
6.28882468e-01 -1.29039383e+00 6.89842850e-02 2.49025986e-01
-7.64473677e-01 1.34803271e+00 3.35203499e-01 -3.69257301e-01
8.44984293e-01 -5.60883619e-02 -5.31203747e-02 -4.89808619e-01
-6.25952721e-01 -4.20472682e-01 4.67155069e-01 3.50243390e-01
3.29347849e-01 -1.64077997e-01 -4.48920250e-01 8.97460580e-01
-4.73944664e-01 1.79089859e-01 4.84075278e-01 7.20689118e-01
-7.21679330e-01 -1.15791106e+00 -4.62952666e-02 8.90504599e-01
-8.03142428e-01 1.27501786e-02 -1.96387365e-01 5.59970796e-01
3.63246053e-01 8.56887519e-01 1.62701651e-01 -4.35454845e-01
1.70598209e-01 4.60338354e-01 3.49265605e-01 -1.14738119e+00
-9.32590127e-01 -2.21396908e-01 -1.35142684e-01 -5.95574081e-01
-4.60699946e-01 -5.76921582e-01 -1.06459332e+00 3.94556284e-01
-2.99724251e-01 5.30421138e-01 6.90758109e-01 9.82292295e-01
4.43529487e-01 7.88124979e-01 6.28367782e-01 -8.89941871e-01
-2.09241524e-01 -1.18490934e+00 -6.78855240e-01 4.23204184e-01
6.15834594e-01 -8.38067770e-01 -5.04338086e-01 8.33088253e-03] | [13.518266677856445, 3.5039796829223633] |
0142505b-94c7-4dd2-9dd7-cbd492391f66 | balancing-between-over-weighting-and-under | 1604.04007 | null | http://arxiv.org/abs/1604.04007v1 | http://arxiv.org/pdf/1604.04007v1.pdf | Balancing Between Over-Weighting and Under-Weighting in Supervised Term Weighting | Supervised term weighting could improve the performance of text
categorization. A way proven to be effective is to give more weight to terms
with more imbalanced distributions across categories. This paper shows that
supervised term weighting should not just assign large weights to imbalanced
terms, but should also control the trade-off between over-weighting and
under-weighting. Over-weighting, a new concept proposed in this paper, is
caused by the improper handling of singular terms and too large ratios between
term weights. To prevent over-weighting, we present three regularization
techniques: add-one smoothing, sublinear scaling and bias term. Add-one
smoothing is used to handle singular terms. Sublinear scaling and bias term
shrink the ratios between term weights. However, if sublinear functions scale
down term weights too much, or the bias term is too large, under-weighting
would occur and harm the performance. It is therefore critical to balance
between over-weighting and under-weighting. Inspired by this insight, we also
propose a new supervised term weighting scheme, regularized entropy (re). Our
re employs entropy to measure term distribution, and introduces the bias term
to control over-weighting and under-weighting. Empirical evaluations on topical
and sentiment classification datasets indicate that sublinear scaling and bias
term greatly influence the performance of supervised term weighting, and our re
enjoys the best results in comparison with existing schemes. | ['Gu Xiaodong', 'Wu Haibing'] | 2016-04-14 | null | null | null | null | ['text-categorization'] | ['natural-language-processing'] | [ 4.55666259e-02 -2.87630577e-02 -5.39016426e-01 -6.00373089e-01
-3.86417240e-01 -3.31462383e-01 5.18341720e-01 5.19475698e-01
-5.79071641e-01 4.91676569e-01 5.35235107e-01 -1.92345798e-01
-1.89808920e-01 -9.33367908e-01 -1.09651983e-01 -7.66253591e-01
-9.67986509e-02 -3.34051512e-02 4.04916316e-01 -3.55346054e-01
6.87207282e-01 2.87342221e-02 -1.73709619e+00 2.42206559e-01
1.22497189e+00 1.04952884e+00 -1.79894105e-01 2.33309478e-01
-7.09323049e-01 5.13225496e-01 -7.55817294e-01 -4.14831072e-01
2.14476928e-01 -1.39144093e-01 -8.61738086e-01 -1.89411104e-01
2.16449067e-01 1.26679510e-01 8.71400610e-02 1.36335170e+00
3.89538884e-01 1.28904238e-01 9.05040860e-01 -1.20391929e+00
-6.92425966e-01 8.05170298e-01 -9.90866899e-01 2.27065966e-01
2.38320053e-01 -3.41559708e-01 1.12170994e+00 -6.35799944e-01
2.09608987e-01 1.26714730e+00 9.66662049e-01 3.10667664e-01
-1.06161034e+00 -6.85304165e-01 5.99351406e-01 -5.20424619e-02
-1.31370568e+00 -1.77676585e-02 8.53430569e-01 -3.96473795e-01
7.71958351e-01 5.36641538e-01 3.71650338e-01 5.46982169e-01
2.41483197e-01 3.09589386e-01 8.83581877e-01 -8.33360791e-01
1.55584157e-01 3.13490540e-01 7.05826044e-01 3.21675867e-01
6.47000611e-01 -1.56002775e-01 -2.60945261e-01 -4.41450208e-01
1.55177236e-01 1.41725913e-01 -2.77520508e-01 -1.68759286e-01
-9.92798746e-01 9.84073818e-01 4.66772854e-01 7.40616381e-01
-1.48736715e-01 -4.97750118e-02 7.75809944e-01 3.78844649e-01
9.71565425e-01 6.15937948e-01 -5.44152081e-01 -4.49788533e-02
-9.13674474e-01 2.47407109e-01 5.98171055e-01 6.81369424e-01
6.06599629e-01 -2.02627301e-01 -6.46210015e-01 1.55402148e+00
3.01742345e-01 3.43230188e-01 1.04447639e+00 -6.88264489e-01
5.15013576e-01 1.17542148e+00 3.90059240e-02 -1.13597178e+00
-5.29370308e-01 -2.99990505e-01 -9.36305165e-01 1.36165693e-01
5.64041257e-01 7.12688491e-02 -1.06295657e+00 1.91066813e+00
2.41098955e-01 -8.30393195e-01 -3.84766340e-01 7.47435212e-01
5.67387581e-01 5.57400167e-01 3.68662894e-01 -6.35886192e-01
1.77301598e+00 -8.11741352e-01 -1.17675471e+00 -2.13024512e-01
9.34750319e-01 -1.01392686e+00 1.34417188e+00 1.46155179e-01
-9.84627306e-01 -2.50853747e-01 -1.03254700e+00 1.19871013e-02
-7.73300171e-01 -2.03099266e-01 5.81519723e-01 7.86172867e-01
-5.63204765e-01 6.78114891e-01 -1.93713605e-01 -2.34144360e-01
1.56105518e-01 2.12764457e-01 -1.07089818e-01 2.32668087e-01
-1.65944743e+00 9.76580322e-01 3.19982171e-01 -2.74100780e-01
3.04132670e-01 -8.60554457e-01 -7.10571706e-01 2.78004229e-01
1.97794557e-01 -3.18160623e-01 1.04809546e+00 -1.01364899e+00
-1.01712489e+00 7.16100991e-01 -9.06219557e-02 -1.19206317e-01
3.85697752e-01 -5.23608876e-03 -4.83641148e-01 -3.39416504e-01
7.15130121e-02 1.46844640e-01 8.31875861e-01 -1.23627019e+00
-5.86952031e-01 -4.39177185e-01 7.64305964e-02 2.25156397e-01
-1.10661781e+00 7.11321011e-02 -1.51056260e-01 -1.03959072e+00
3.67055804e-01 -6.29257917e-01 -1.88129410e-01 -1.18677557e-01
-1.19263656e-01 -5.45836508e-01 6.98108852e-01 -4.08559263e-01
2.09891701e+00 -2.06676459e+00 -2.15279758e-01 2.88755536e-01
3.04488540e-01 2.80959576e-01 5.41104898e-02 2.88558602e-01
-2.08636835e-01 4.38427269e-01 -3.60912442e-01 2.15421710e-02
1.25235215e-01 2.32825294e-01 -2.33950287e-01 2.69950509e-01
-1.28423765e-01 3.40979457e-01 -8.11143100e-01 -5.97780764e-01
-1.33255541e-01 4.03824955e-01 -5.18096268e-01 1.09012656e-01
8.82854387e-02 -5.11515796e-01 -2.44944572e-01 5.03631473e-01
1.05703223e+00 -7.91123658e-02 -1.17057534e-02 -4.87457782e-01
-2.69593507e-01 3.83160025e-01 -1.43115830e+00 8.62204969e-01
-5.00202358e-01 1.90748632e-01 -4.23201770e-02 -1.05971456e+00
9.02658939e-01 1.96745381e-01 4.52220410e-01 -5.68134010e-01
4.68135029e-01 2.56794631e-01 -1.47634283e-01 -6.57283545e-01
8.10956478e-01 -4.49199915e-01 -6.52783364e-02 3.86309743e-01
-2.64093757e-01 -1.92688733e-01 4.95087773e-01 1.98582292e-01
7.43704796e-01 -4.11772162e-01 4.38018262e-01 -6.65672243e-01
7.76207209e-01 -3.74514610e-01 5.86055756e-01 4.10003453e-01
-4.03919250e-01 4.90865707e-01 6.19998634e-01 -2.88648158e-01
-8.64636362e-01 -2.27705583e-01 -5.09008288e-01 1.68984509e+00
2.17864960e-01 -6.00674570e-01 -7.33923614e-01 -9.62390542e-01
2.63381898e-01 5.56432962e-01 -8.21998239e-01 -4.90889460e-01
-3.67712259e-01 -1.40931892e+00 3.84035081e-01 4.95442301e-01
3.98069352e-01 -7.91501820e-01 -1.07902035e-01 1.88103452e-01
-4.68872279e-01 -5.03407180e-01 -9.25873339e-01 3.59089702e-01
-8.60549629e-01 -8.25170994e-01 -1.03765845e+00 -7.39250660e-01
9.56959546e-01 4.76851046e-01 1.09032297e+00 4.81358945e-01
1.58787575e-02 -1.01777390e-01 -5.81144810e-01 -5.99348545e-01
-2.45287016e-01 2.43001759e-01 2.03605056e-01 -5.49014509e-02
5.58788955e-01 -2.61013329e-01 -6.85242116e-01 5.48764229e-01
-1.11109877e+00 -5.87348938e-01 1.59402430e-01 1.00707841e+00
1.47666931e-01 2.20531702e-01 6.06795311e-01 -9.13064897e-01
1.03159630e+00 -2.35366493e-01 -3.34108442e-01 2.58505970e-01
-1.31545150e+00 3.73895913e-01 6.04744911e-01 -7.93650270e-01
-1.08238792e+00 -6.15955770e-01 -1.60477147e-01 4.95992601e-02
3.37910384e-01 4.02937561e-01 5.69210276e-02 -1.86259866e-01
6.67042375e-01 -2.98461765e-01 -5.19857481e-02 -5.05035520e-01
2.09708735e-01 1.04244196e+00 -1.64369211e-01 -3.87190163e-01
6.05454624e-01 3.11382413e-01 -1.92992970e-01 -4.30230469e-01
-1.08973444e+00 -6.76088929e-01 -2.73562610e-01 -1.24661498e-01
4.78726387e-01 -5.19573569e-01 -4.60947901e-01 3.54183316e-01
-7.98120856e-01 2.61726230e-01 -3.12442034e-01 3.25155616e-01
-4.99826185e-02 5.29021680e-01 -2.95248389e-01 -1.03928018e+00
-6.42327726e-01 -7.82886028e-01 6.84133947e-01 1.06895424e-01
-5.74574292e-01 -1.16457057e+00 5.74081652e-02 2.45599791e-01
9.70563233e-01 -1.77376032e-01 9.72519875e-01 -6.76158965e-01
5.44858634e-01 -4.98339593e-01 -3.98475289e-01 5.91484666e-01
2.80407846e-01 5.95053397e-02 -8.81006777e-01 -2.82342166e-01
2.73925602e-01 -3.42209451e-02 1.02677369e+00 2.51328051e-01
1.32634699e+00 -4.43209499e-01 -2.46897504e-01 4.37178820e-01
1.26771605e+00 1.36541743e-02 5.10612786e-01 5.77635527e-01
6.80891454e-01 8.26494038e-01 7.59127200e-01 3.96918267e-01
3.29619318e-01 7.13697433e-01 1.32689133e-01 -1.75473168e-01
-5.16778864e-02 -1.82825397e-03 2.19675004e-01 1.20146322e+00
-2.35496104e-01 -1.39133772e-02 -8.13858032e-01 4.32570904e-01
-1.66984153e+00 -7.77253866e-01 -4.05217797e-01 2.45660734e+00
1.30621946e+00 4.37878698e-01 1.67302996e-01 7.64963508e-01
8.12323511e-01 3.33619535e-01 -1.12674288e-01 -7.54254043e-01
9.85541940e-02 -2.43741423e-01 7.23235726e-01 7.79047430e-01
-1.05568254e+00 6.31857395e-01 6.55834484e+00 1.39897168e+00
-1.10454106e+00 1.04290232e-01 5.98787427e-01 1.08223528e-01
-4.55287874e-01 -3.68474275e-02 -8.16234410e-01 7.70532489e-01
5.76405108e-01 -4.20099646e-01 4.53036539e-02 7.92159975e-01
1.13598861e-01 -1.64702490e-01 -5.73097348e-01 6.44926786e-01
1.32683605e-01 -6.35848641e-01 1.24771327e-01 -3.96978073e-02
4.29632097e-01 -5.10871172e-01 -2.62534082e-01 5.50782442e-01
8.69055018e-02 -6.34299040e-01 4.92965519e-01 2.08895892e-01
5.15853047e-01 -7.67542660e-01 1.12634826e+00 2.01459453e-01
-1.16144013e+00 -1.72544807e-01 -5.40672421e-01 -2.48850465e-01
-6.21493533e-03 1.40373504e+00 -1.15803085e-01 3.48515838e-01
6.71208024e-01 2.24303395e-01 -4.74550813e-01 9.33533251e-01
-1.05931073e-01 5.18826783e-01 -2.87290543e-01 -1.92457348e-01
-4.53713611e-02 -2.45091304e-01 4.82151389e-01 1.51427734e+00
1.08616285e-01 -6.32059723e-02 3.83476354e-02 3.27555656e-01
-3.62336449e-02 4.92451519e-01 -3.50240618e-01 3.01436812e-01
4.92530137e-01 1.24603844e+00 -6.38970733e-01 -5.69970429e-01
-3.01947087e-01 4.86261249e-01 2.68079191e-01 1.64938167e-01
-7.90217757e-01 -8.72982860e-01 4.88712579e-01 2.69126296e-01
3.88417318e-02 2.18097597e-01 -8.71991098e-01 -1.08506656e+00
1.33709162e-01 -6.13009214e-01 6.10227644e-01 -2.55128831e-01
-1.69978571e+00 6.40626073e-01 1.79145724e-01 -1.37011731e+00
3.07881862e-01 -3.53797525e-01 -5.18888414e-01 6.99319422e-01
-1.75997365e+00 -7.36175179e-01 -2.69936353e-01 1.72873199e-01
4.34775054e-01 1.77754447e-01 7.04614520e-01 5.79371870e-01
-4.64709163e-01 9.41729248e-01 -4.09021899e-02 -2.25839943e-01
1.08254623e+00 -1.32677925e+00 -9.34357122e-02 3.59800935e-01
-4.95224595e-01 8.65244210e-01 8.41822326e-01 -4.42809910e-01
-6.85154021e-01 -8.46724153e-01 1.55218947e+00 -2.40544289e-01
6.64877653e-01 -2.65637934e-01 -1.40270638e+00 -3.63268033e-02
6.82015568e-02 -6.11344762e-02 8.13330173e-01 4.00656819e-01
-7.68479168e-01 -3.93427044e-01 -1.42032111e+00 4.86963719e-01
7.29168832e-01 -1.71779096e-01 -7.97079146e-01 4.80077207e-01
6.92637086e-01 1.15281325e-02 -1.00891531e+00 6.68672383e-01
7.12493360e-01 -8.67876053e-01 9.91185844e-01 -2.72482038e-01
2.73867399e-01 -1.78624898e-01 1.70019604e-02 -1.49941826e+00
-6.69921815e-01 -3.03943545e-01 -1.16594240e-01 1.62253535e+00
5.46775579e-01 -7.39518046e-01 2.83817470e-01 7.08822131e-01
2.48494476e-01 -7.83963621e-01 -7.54549861e-01 -1.00434709e+00
5.38992643e-01 -2.72283643e-01 7.27048516e-01 1.30020821e+00
6.82681501e-01 1.95977777e-01 -2.27514267e-01 -3.01364899e-01
5.54273725e-01 -3.58849704e-01 1.09681271e-01 -1.25772238e+00
3.48447055e-01 -8.47478092e-01 -1.98548034e-01 -7.12664068e-01
-2.00455979e-01 -6.26949430e-01 4.39982004e-02 -1.34579992e+00
2.48294756e-01 -4.12182868e-01 -5.83176792e-01 6.89601004e-01
-7.25102723e-01 4.83271182e-01 -2.22932659e-02 4.12617534e-01
-3.81524116e-01 5.55016816e-01 1.15057719e+00 -9.25620198e-02
-2.34253868e-01 -6.90405592e-02 -1.04056942e+00 1.02855563e+00
8.66853774e-01 -6.84298754e-01 -1.98608398e-01 -1.81952834e-01
4.65369821e-01 -6.50910854e-01 -3.78418803e-01 -6.75235987e-01
1.47741705e-01 -2.28818834e-01 7.54730329e-02 -2.95759827e-01
-2.53175259e-01 -9.92000222e-01 -5.34689069e-01 4.40240592e-01
-6.39911234e-01 -5.54303117e-02 -7.40254968e-02 2.91511595e-01
-3.76562268e-01 -5.26145041e-01 1.06054926e+00 -4.70898002e-02
-1.59258962e-01 -1.61114827e-01 -3.71042937e-01 2.14221627e-01
8.41934681e-01 -2.24222273e-01 -3.24058771e-01 -1.64247274e-01
-4.03355837e-01 4.86367166e-01 2.09940568e-01 6.56607866e-01
1.13271393e-01 -1.52828610e+00 -4.25179958e-01 3.05908352e-01
4.23994362e-01 -3.66228908e-01 6.76884279e-02 7.59968996e-01
-1.61894321e-01 2.43882596e-01 3.27718765e-01 -2.38737136e-01
-1.50799251e+00 5.46605527e-01 1.42956778e-01 -5.71895361e-01
-1.76490247e-01 8.45249772e-01 -7.81098977e-02 -6.39332771e-01
6.17331624e-01 -5.62237978e-01 -7.28891075e-01 6.63615525e-01
8.03779125e-01 5.50429642e-01 3.40171337e-01 -5.38753331e-01
-4.32091743e-01 9.02076066e-01 -4.19398516e-01 2.23804727e-01
9.90415037e-01 -3.10857117e-01 -4.90462333e-01 6.01082087e-01
1.24074709e+00 5.18543795e-02 -5.87687373e-01 -2.64494210e-01
1.86032072e-01 -6.42827570e-01 4.61318195e-01 -7.73098171e-01
-9.31479692e-01 5.41848481e-01 4.79247451e-01 1.08993435e+00
1.16262078e+00 -3.71069729e-01 7.40164280e-01 -3.67883965e-02
1.70334026e-01 -1.58623087e+00 -2.97142882e-02 7.73678184e-01
8.27438831e-01 -1.29343724e+00 3.93533647e-01 -6.75089896e-01
-6.90242589e-01 9.92757201e-01 6.18310630e-01 4.62276768e-03
9.99925971e-01 4.04485881e-01 1.85512602e-01 -1.30946681e-01
-5.70168972e-01 -1.21979862e-01 5.81720650e-01 1.70415878e-01
9.26866233e-01 -3.32339518e-02 -1.31646597e+00 6.15787148e-01
-1.60867020e-01 -4.86656837e-03 1.66485757e-01 9.94086504e-01
-7.62708426e-01 -1.19631112e+00 -7.28217959e-01 7.00727701e-01
-7.67226815e-01 -6.64164275e-02 -2.49481305e-01 3.56981993e-01
3.38083982e-01 1.06558013e+00 3.14191468e-02 -4.08312649e-01
5.77339828e-01 3.64899844e-01 -3.90336663e-02 -3.09259742e-01
-1.00163805e+00 3.52855287e-02 1.03308760e-01 -1.90909013e-01
-6.08049273e-01 -5.71836866e-02 -1.11813104e+00 -4.20323223e-01
-1.01346743e+00 3.94135922e-01 7.14703679e-01 8.61838698e-01
1.47168979e-01 7.37918913e-01 9.05904472e-01 -5.56454718e-01
-1.01378667e+00 -1.37832832e+00 -6.83284640e-01 8.54831874e-01
3.61182630e-01 -8.74282658e-01 -1.28574729e+00 -2.15873033e-01] | [10.479713439941406, 7.276558876037598] |
f60ef3c2-e0b9-4712-9cd9-b13b332f094e | splal-similarity-based-pseudo-labeling-with | 2307.04610 | null | https://arxiv.org/abs/2307.04610v1 | https://arxiv.org/pdf/2307.04610v1.pdf | SPLAL: Similarity-based pseudo-labeling with alignment loss for semi-supervised medical image classification | Medical image classification is a challenging task due to the scarcity of labeled samples and class imbalance caused by the high variance in disease prevalence. Semi-supervised learning (SSL) methods can mitigate these challenges by leveraging both labeled and unlabeled data. However, SSL methods for medical image classification need to address two key challenges: (1) estimating reliable pseudo-labels for the images in the unlabeled dataset and (2) reducing biases caused by class imbalance. In this paper, we propose a novel SSL approach, SPLAL, that effectively addresses these challenges. SPLAL leverages class prototypes and a weighted combination of classifiers to predict reliable pseudo-labels over a subset of unlabeled images. Additionally, we introduce alignment loss to mitigate model biases toward majority classes. To evaluate the performance of our proposed approach, we conduct experiments on two publicly available medical image classification benchmark datasets: the skin lesion classification (ISIC 2018) and the blood cell classification dataset (BCCD). The experimental results empirically demonstrate that our approach outperforms several state-of-the-art SSL methods over various evaluation metrics. Specifically, our proposed approach achieves a significant improvement over the state-of-the-art approach on the ISIC 2018 dataset in both Accuracy and F1 score, with relative margins of 2.24\% and 11.40\%, respectively. Finally, we conduct extensive ablation experiments to examine the contribution of different components of our approach, validating its effectiveness. | ['Pravendra Singh', 'Suruchi Kumari', 'Divyansh Agarwal', 'Pranaw Raj', 'Md Junaid Mahmood'] | 2023-07-10 | null | null | null | null | ['skin-lesion-classification', 'medical-image-classification', 'semi-supervised-medical-image-classification', 'classification-1'] | ['medical', 'medical', 'medical', 'methodology'] | [ 4.53977287e-01 -9.27596986e-02 -6.43369615e-01 -5.60242414e-01
-1.22308731e+00 -2.47284070e-01 4.25336957e-01 3.75104934e-01
-3.83728385e-01 7.69033015e-01 -5.88377677e-02 -1.62425786e-01
-2.71472689e-02 -3.47395211e-01 -3.97869945e-01 -8.75743151e-01
2.15055957e-01 1.92813814e-01 9.53665301e-02 4.05506283e-01
1.08904839e-01 1.57032952e-01 -1.33425951e+00 3.58537614e-01
1.26951849e+00 1.31613743e+00 -2.55457789e-01 2.63777643e-01
-7.11679608e-02 9.96108174e-01 -4.07530755e-01 -3.15814316e-01
1.50123090e-01 -4.84180301e-01 -6.70455515e-01 3.66136223e-01
3.87872100e-01 -1.35442451e-01 5.26690111e-02 1.12148583e+00
4.95799601e-01 -4.26508904e-01 9.51782525e-01 -1.34927511e+00
-4.49477822e-01 3.19820762e-01 -9.98330295e-01 -1.54724857e-02
-1.39507279e-01 2.21501157e-01 6.84003055e-01 -8.61255705e-01
5.34721017e-01 8.84925723e-01 8.26622307e-01 5.09695351e-01
-1.18813646e+00 -8.14366579e-01 8.94691423e-02 1.55345544e-01
-1.52679634e+00 -4.40618336e-01 5.24116695e-01 -6.03539348e-01
3.25103819e-01 2.78222293e-01 5.77204116e-02 8.32192540e-01
-6.38256222e-02 9.67506707e-01 1.57922637e+00 -5.44231653e-01
3.69098455e-01 3.49094033e-01 4.62787420e-01 8.77107263e-01
2.95689166e-01 2.74157356e-02 -4.18041259e-01 -3.86533469e-01
3.67445856e-01 5.59099019e-02 -2.59041816e-01 -2.51319647e-01
-1.05681241e+00 7.83539236e-01 3.51940900e-01 -1.39739677e-01
-3.44110787e-01 -2.95165896e-01 4.77447242e-01 -8.09409618e-02
6.86937094e-01 1.65558904e-01 -3.43490601e-01 2.59545028e-01
-8.92782211e-01 -1.79649845e-01 5.19248843e-01 7.43080378e-01
2.77032822e-01 -3.22344929e-01 -4.85834926e-01 1.21573579e+00
2.36228257e-01 4.78638053e-01 5.83647907e-01 -5.95192134e-01
3.71148676e-01 8.62441480e-01 -9.72660929e-02 -7.53978252e-01
-4.07088846e-01 -6.27700806e-01 -1.04549956e+00 -1.90052316e-01
5.15117824e-01 -2.47158725e-02 -1.14005649e+00 1.67996287e+00
5.16124964e-01 2.50348508e-01 6.26862794e-02 8.04766834e-01
9.13345456e-01 3.21442515e-01 3.12871248e-01 -4.21368301e-01
1.24067795e+00 -1.05689836e+00 -6.68805063e-01 -2.02531293e-02
6.92309499e-01 -6.93586707e-01 8.66765797e-01 1.79404840e-01
-6.62469506e-01 -3.31192225e-01 -1.03083360e+00 2.79794574e-01
9.11208242e-02 6.06476128e-01 4.64825928e-01 6.96867466e-01
-5.08643866e-01 2.21483260e-01 -9.14093673e-01 -3.04760665e-01
8.89364779e-01 1.66933447e-01 -2.64249265e-01 -4.55235511e-01
-7.91937232e-01 5.19361198e-01 9.65679362e-02 2.88370345e-02
-6.19651854e-01 -9.39349890e-01 -7.24932432e-01 -2.03864232e-01
3.46111655e-01 -3.14484924e-01 9.51462865e-01 -8.46750975e-01
-9.43761110e-01 1.07953048e+00 -1.44551650e-01 -4.14235055e-01
5.89676857e-01 -4.67036143e-02 -3.50165755e-01 3.34126979e-01
1.89402938e-01 7.46167123e-01 4.96543348e-01 -1.31862211e+00
-8.22448313e-01 -5.82709551e-01 -4.43609357e-01 -3.44610512e-02
-5.21009624e-01 -2.57398039e-01 -5.69870234e-01 -8.05574358e-01
1.37982488e-01 -1.09385741e+00 -2.15576664e-01 2.52546251e-01
-7.65576601e-01 -2.38842189e-01 5.10812044e-01 -5.39685726e-01
1.27627838e+00 -2.17738748e+00 -2.60669053e-01 2.26167753e-01
1.70027092e-01 4.94663715e-01 -1.83632106e-01 3.65938693e-02
-3.49238142e-02 1.31991401e-01 -3.66605103e-01 -4.28728968e-01
-2.72679776e-01 -1.00754023e-01 -8.25441554e-02 5.73516428e-01
3.88202995e-01 6.62051499e-01 -7.20943332e-01 -8.34877491e-01
6.95822313e-02 4.50495481e-01 -2.95832396e-01 7.20028281e-02
7.09273890e-02 4.39291745e-01 -3.70188773e-01 1.11939669e+00
6.86421931e-01 -6.38948977e-01 3.91769558e-01 -3.94795835e-01
3.34194452e-01 -5.74347675e-02 -9.67526257e-01 1.26989555e+00
-2.61545360e-01 1.85273677e-01 -3.21259022e-01 -1.01342058e+00
8.66452754e-01 1.99307248e-01 5.43929815e-01 -6.70652628e-01
1.11557767e-01 4.01960194e-01 -1.29604191e-01 -4.90698606e-01
-1.40538186e-01 -1.42492980e-01 9.84868631e-02 4.00210232e-01
-6.28814474e-02 3.95291179e-01 1.33337602e-01 2.31461748e-01
1.02666795e+00 -1.99201718e-01 4.07818586e-01 -3.19730222e-01
6.05986834e-01 1.61468033e-02 9.61725116e-01 6.68608665e-01
-6.42185807e-01 6.26746833e-01 5.43471515e-01 -2.81773984e-01
-6.35472834e-01 -9.31364179e-01 -5.78766406e-01 6.38366997e-01
1.48401812e-01 -1.55464724e-01 -7.16294527e-01 -1.18669462e+00
1.48366347e-01 4.20286030e-01 -7.78945208e-01 -1.36102021e-01
-1.03901275e-01 -1.34623313e+00 4.88522559e-01 6.29278183e-01
5.13118565e-01 -5.51282108e-01 -3.22260648e-01 -1.80169478e-01
-2.26945564e-01 -1.29685199e+00 -4.56846893e-01 -9.60254818e-02
-7.58259892e-01 -1.47778130e+00 -7.27823138e-01 -8.69681835e-01
1.18804657e+00 2.90213346e-01 8.95883560e-01 3.05883527e-01
-5.36008298e-01 7.78797716e-02 -4.25664186e-01 -4.41535115e-01
-3.50413114e-01 1.44124046e-01 -5.93832619e-02 3.77541125e-01
4.66890723e-01 -6.01778626e-02 -7.77416468e-01 5.36724746e-01
-8.53496373e-01 2.55502582e-01 7.07227170e-01 1.13381410e+00
9.96461451e-01 2.42713820e-02 9.33591664e-01 -1.46709573e+00
1.94055140e-01 -6.02499545e-01 -4.66732591e-01 5.77784002e-01
-9.65035737e-01 -1.63397446e-01 5.11313975e-01 -3.54203224e-01
-9.60744381e-01 2.43648529e-01 1.17384665e-01 -2.27862000e-01
2.12854464e-02 5.17022133e-01 7.50787780e-02 -7.90781993e-03
5.68633616e-01 -5.11711687e-02 2.79815614e-01 -3.72519135e-01
-4.01830822e-02 1.10242617e+00 5.04621863e-01 -3.44824076e-01
5.38734734e-01 6.38771117e-01 5.18335961e-02 -4.22400475e-01
-1.43757856e+00 -5.49259484e-01 -3.79702955e-01 -1.03335992e-01
5.82636714e-01 -1.13652503e+00 -3.88717979e-01 7.38242805e-01
-4.82062131e-01 -2.29161635e-01 -1.64600518e-02 5.14252841e-01
-1.15790434e-01 3.10217768e-01 -7.49613285e-01 -8.12438846e-01
-6.24947727e-01 -1.29227424e+00 1.07354379e+00 3.99548471e-01
-1.85611039e-01 -8.15334439e-01 -1.43878073e-01 7.28719294e-01
2.36323312e-01 5.28887689e-01 1.01728380e+00 -8.69100988e-01
-1.51253358e-01 -4.00769562e-01 -4.98487651e-01 4.80016261e-01
3.81070971e-01 -4.16418873e-02 -1.02804089e+00 -4.55505222e-01
-2.23428234e-01 -6.87639475e-01 1.09040165e+00 3.34332377e-01
1.42106175e+00 -1.18877172e-01 -4.28058088e-01 5.51975965e-01
1.46251142e+00 5.24965748e-02 4.04772401e-01 8.79305881e-03
5.43290317e-01 7.44775712e-01 9.98156488e-01 4.67951953e-01
4.61023211e-01 4.91709143e-01 2.35826313e-01 -3.57404411e-01
-2.86415726e-01 -2.35824272e-01 -1.28023937e-01 7.61532843e-01
3.48442346e-01 -2.20181774e-02 -1.03089488e+00 5.63112259e-01
-1.83904636e+00 -3.46917808e-01 -1.99966595e-01 2.26443267e+00
1.13285327e+00 1.42289221e-01 6.74225911e-02 3.19209635e-01
8.71364355e-01 -1.84904039e-01 -7.99844742e-01 3.50623190e-01
-1.16201751e-01 1.91081405e-01 4.91857439e-01 1.39890745e-01
-1.43740177e+00 4.58016813e-01 6.00842381e+00 1.02529442e+00
-1.20615959e+00 2.64402516e-02 1.29848635e+00 -1.36606172e-01
1.06351033e-01 -3.87431741e-01 -8.22553813e-01 7.04010725e-01
6.91841781e-01 1.02805302e-01 -6.01883009e-02 7.02623308e-01
-9.91921499e-02 -2.76621312e-01 -9.97059047e-01 9.54211295e-01
3.28774273e-01 -1.25055814e+00 -8.60737115e-02 -2.93521062e-02
1.03391969e+00 -1.40533105e-01 2.24976033e-01 2.82562196e-01
4.66823913e-02 -1.01973176e+00 4.33664650e-01 3.17297220e-01
1.15056479e+00 -6.18085682e-01 1.06528091e+00 2.60392398e-01
-7.93239653e-01 -5.40156253e-02 -1.06488541e-01 2.90621132e-01
-2.42691740e-01 9.95540023e-01 -7.11052060e-01 4.17668223e-01
5.32182395e-01 7.60044754e-01 -9.20345664e-01 1.03676939e+00
-1.93751544e-01 1.01215374e+00 -1.96447708e-02 7.35311583e-02
-4.81875688e-02 4.63474020e-02 -2.99381595e-02 1.00636733e+00
-2.18950976e-02 5.65465428e-02 2.92389333e-01 4.78189975e-01
-2.56905019e-01 2.76426673e-01 6.82263747e-02 5.88835999e-02
7.57528007e-01 1.37979519e+00 -8.01051319e-01 -3.32687467e-01
-3.19922566e-01 5.04697204e-01 2.78132826e-01 2.08915323e-01
-8.46016109e-01 -2.99154043e-01 2.24312589e-01 -3.67903872e-03
5.71778305e-02 5.05482614e-01 -3.73864949e-01 -1.03179991e+00
9.43090469e-02 -9.17806983e-01 6.90559506e-01 -3.00245821e-01
-1.68853128e+00 5.14508367e-01 -2.89084226e-01 -1.36508548e+00
9.41580087e-02 -4.45717812e-01 -2.16253862e-01 5.62153459e-01
-1.83790863e+00 -1.20899963e+00 -5.35104513e-01 2.66586453e-01
3.51715803e-01 -3.16844374e-01 7.82306612e-01 4.32143688e-01
-1.00332212e+00 9.79064524e-01 2.72074491e-01 2.86230534e-01
1.04320216e+00 -1.09326339e+00 -1.01941831e-01 6.45603418e-01
-1.54645175e-01 4.90363538e-01 1.51887998e-01 -4.24212456e-01
-1.05635703e+00 -1.37673068e+00 7.00757086e-01 -2.68440574e-01
3.01839679e-01 -1.50589451e-01 -9.52249527e-01 3.23576331e-01
-3.91041219e-01 5.35144210e-01 1.27253103e+00 -1.50753669e-02
-6.16505027e-01 -3.15649122e-01 -1.43487501e+00 3.43227088e-01
7.10686803e-01 -3.28424066e-01 -2.61632800e-02 4.95582044e-01
4.64788049e-01 -3.79415393e-01 -9.32315230e-01 8.44751000e-01
7.05671906e-01 -9.01586235e-01 7.19848037e-01 -5.47595143e-01
6.24483168e-01 -3.16591650e-01 -1.59604132e-01 -1.08140218e+00
1.37347716e-03 -8.83871391e-02 5.20594418e-02 1.42406642e+00
4.12172854e-01 -6.49715006e-01 8.38593662e-01 5.42905927e-01
2.09595621e-01 -1.27125931e+00 -7.93489456e-01 -5.98464251e-01
-9.86865815e-03 -1.62758380e-01 3.41281772e-01 1.07434380e+00
-1.04163289e-01 3.03481460e-01 -2.81561464e-01 4.07146886e-02
9.65036690e-01 2.47210309e-01 6.04107141e-01 -1.02781236e+00
-1.49060622e-01 -8.94998685e-02 -2.62646705e-01 -4.83745039e-01
1.61207989e-01 -9.88328099e-01 5.40957116e-02 -1.29283619e+00
8.51286173e-01 -8.88264716e-01 -5.53222537e-01 5.98830521e-01
-7.42504358e-01 7.12008357e-01 9.15020034e-02 4.89300132e-01
-8.14021051e-01 3.46122175e-01 9.65428174e-01 -2.18867764e-01
8.76654759e-02 -2.89245173e-02 -9.42476273e-01 7.06101537e-01
8.00716996e-01 -5.48263967e-01 -3.54213595e-01 -2.52197832e-01
-2.63944924e-01 -6.72116131e-02 3.00541699e-01 -9.05795813e-01
7.35676661e-02 -1.23521298e-01 4.99952435e-01 -5.77903450e-01
-8.83963853e-02 -4.93618369e-01 -1.35677204e-01 6.93575621e-01
-6.43466473e-01 -4.62087542e-01 2.14809012e-02 7.01172948e-01
-2.00909242e-01 -5.64891659e-03 1.14574301e+00 3.44933778e-01
-2.81183213e-01 2.37157404e-01 7.72000179e-02 2.32032433e-01
1.20880985e+00 2.10406169e-01 -6.35107815e-01 -5.68665005e-02
-2.12581769e-01 5.77751935e-01 3.99145514e-01 2.77280450e-01
4.20538336e-01 -1.36496043e+00 -8.80261362e-01 1.28383696e-01
6.00506186e-01 -4.37213853e-02 3.32829922e-01 1.10708010e+00
-4.47610527e-01 3.68249357e-01 4.38281745e-02 -8.36030304e-01
-1.40400100e+00 4.07967806e-01 1.87448069e-01 -5.36029935e-01
-2.39359975e-01 8.21767390e-01 2.87534326e-01 -5.56817174e-01
4.21257794e-01 -1.05866238e-01 -1.84067320e-02 -8.07833597e-02
6.52270555e-01 3.62560391e-01 1.78105786e-01 -5.82374096e-01
-5.94357908e-01 4.65583473e-01 -3.70411873e-01 3.95680726e-01
1.16528320e+00 5.44374362e-02 4.01797965e-02 2.26152480e-01
1.30055094e+00 -1.82580426e-01 -1.14404798e+00 -5.42354763e-01
1.03011668e-01 -4.90049809e-01 3.65168671e-03 -1.21199143e+00
-1.24719298e+00 7.18964696e-01 9.12102878e-01 -2.75586545e-01
1.37955987e+00 -1.42997056e-01 7.91997731e-01 -6.16764501e-02
2.02838227e-01 -1.01839948e+00 1.58393010e-01 -1.63840041e-01
4.03158069e-01 -1.68443155e+00 2.38416672e-01 -8.50697637e-01
-8.99450243e-01 6.63053751e-01 6.03958428e-01 2.02312872e-01
6.03097796e-01 1.68677762e-01 3.97407323e-01 -3.34718078e-02
-9.07074213e-01 -8.58961642e-02 3.07764411e-01 4.31372374e-01
4.80875194e-01 9.27883163e-02 -5.90489507e-01 7.26913035e-01
3.41882259e-01 2.56947935e-01 2.46756509e-01 9.64137435e-01
-8.78788158e-02 -9.87756073e-01 -3.35459203e-01 8.37843239e-01
-7.26532698e-01 9.61269513e-02 -3.16202521e-01 6.36636198e-01
6.62592649e-02 1.11510646e+00 -8.70714709e-02 -3.37149262e-01
1.76115662e-01 -8.51921737e-02 3.40003401e-01 -4.73913103e-01
-4.10019845e-01 1.54789835e-01 -7.00936317e-02 -3.64194989e-01
-6.97413564e-01 -5.08545578e-01 -1.24153936e+00 1.54076219e-01
-4.94428039e-01 5.04241325e-02 4.77764338e-01 7.81842649e-01
4.24670219e-01 4.80659008e-01 9.38326180e-01 -1.05465367e-01
-1.00024545e+00 -8.91423881e-01 -6.18330777e-01 7.72567093e-01
1.10153541e-01 -7.62590945e-01 -3.38557303e-01 1.06318288e-01] | [15.0922212600708, -2.4067208766937256] |
0ee4bd36-4953-4542-98d9-344fb9fda67c | self-paced-contrastive-learning-with-hybrid | 2006.02713 | null | https://arxiv.org/abs/2006.02713v2 | https://arxiv.org/pdf/2006.02713v2.pdf | Self-paced Contrastive Learning with Hybrid Memory for Domain Adaptive Object Re-ID | Domain adaptive object re-ID aims to transfer the learned knowledge from the labeled source domain to the unlabeled target domain to tackle the open-class re-identification problems. Although state-of-the-art pseudo-label-based methods have achieved great success, they did not make full use of all valuable information because of the domain gap and unsatisfying clustering performance. To solve these problems, we propose a novel self-paced contrastive learning framework with hybrid memory. The hybrid memory dynamically generates source-domain class-level, target-domain cluster-level and un-clustered instance-level supervisory signals for learning feature representations. Different from the conventional contrastive learning strategy, the proposed framework jointly distinguishes source-domain classes, and target-domain clusters and un-clustered instances. Most importantly, the proposed self-paced method gradually creates more reliable clusters to refine the hybrid memory and learning targets, and is shown to be the key to our outstanding performance. Our method outperforms state-of-the-arts on multiple domain adaptation tasks of object re-ID and even boosts the performance on the source domain without any extra annotations. Our generalized version on unsupervised object re-ID surpasses state-of-the-art algorithms by considerable 16.7% and 7.9% on Market-1501 and MSMT17 benchmarks. | ['Hongsheng Li', 'Dapeng Chen', 'Rui Zhao', 'Yixiao Ge', 'Feng Zhu'] | 2020-06-04 | null | http://proceedings.neurips.cc/paper/2020/hash/821fa74b50ba3f7cba1e6c53e8fa6845-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/821fa74b50ba3f7cba1e6c53e8fa6845-Paper.pdf | neurips-2020-12 | ['unsupervised-person-re-identification'] | ['computer-vision'] | [ 3.34405184e-01 -2.18963972e-03 -6.63119793e-01 -4.10244852e-01
-9.67372358e-01 -5.17116308e-01 7.45119214e-01 2.98174918e-01
-2.95640558e-01 7.88617969e-01 -1.64699614e-01 2.95275450e-01
-2.51629651e-01 -5.15602589e-01 -6.64105117e-01 -8.62904787e-01
1.28718987e-01 1.10411012e+00 3.81016552e-01 1.45395055e-01
2.27575660e-01 1.72958910e-01 -1.85492957e+00 4.23852742e-01
1.07602644e+00 1.15567827e+00 3.90568018e-01 3.72458473e-02
-4.30566818e-01 6.45307004e-01 -3.86990547e-01 -2.52727941e-02
1.51428446e-01 -4.67916131e-01 -9.80524719e-01 3.10899705e-01
4.00811195e-01 1.90914184e-01 -3.43417898e-02 1.19450533e+00
3.50553781e-01 2.75805444e-01 9.71393585e-01 -1.26718545e+00
-9.99274433e-01 5.77557325e-01 -9.20555472e-01 2.00613052e-01
-2.01284543e-01 -1.65999278e-01 7.05905557e-01 -1.06282330e+00
8.11046124e-01 1.08179605e+00 3.41273636e-01 8.50830793e-01
-1.41424382e+00 -1.12635827e+00 5.30793667e-01 4.60916013e-01
-1.47681761e+00 -3.89886677e-01 9.16001022e-01 -5.79980373e-01
6.39854968e-01 -1.99903876e-01 8.95396098e-02 1.03463662e+00
-3.26258898e-01 7.49422193e-01 1.28635526e+00 -4.69229132e-01
3.44866693e-01 4.59517747e-01 4.72792506e-01 3.68267626e-01
1.37175277e-01 1.75600126e-01 -4.71517444e-01 -8.87301713e-02
5.24617910e-01 1.32047012e-01 6.03827089e-02 -8.31378281e-01
-1.38566208e+00 7.09033966e-01 4.57426071e-01 4.53029007e-01
-1.42863944e-01 -3.24882597e-01 4.34534013e-01 2.61438251e-01
6.66083515e-01 4.35014665e-01 -5.59774578e-01 6.25148192e-02
-8.19837511e-01 6.97025657e-02 4.59212184e-01 1.18968511e+00
1.00309098e+00 -5.76115921e-02 -1.36214092e-01 1.21678448e+00
7.59552419e-02 4.04896677e-01 8.03639352e-01 -7.42593050e-01
3.65823030e-01 9.03223813e-01 -1.00003080e-02 -5.62575221e-01
-4.53144014e-01 -8.64248097e-01 -9.90345895e-01 1.18593976e-01
4.50263232e-01 1.73952967e-01 -1.10835624e+00 1.73437476e+00
5.02697766e-01 6.30838871e-01 3.08893353e-01 7.56977260e-01
8.56327295e-01 5.58869064e-01 3.52574855e-01 -3.35219145e-01
1.15374267e+00 -1.15813172e+00 -4.30701107e-01 -2.97238618e-01
4.90378648e-01 -3.74025464e-01 7.75719047e-01 2.91717023e-01
-7.23651111e-01 -1.01248229e+00 -9.88002419e-01 4.61466074e-01
-3.97486776e-01 9.26906765e-02 2.02828392e-01 3.04410785e-01
-5.44528186e-01 4.36265945e-01 -4.27976847e-01 -3.06426853e-01
6.97710872e-01 3.80724669e-01 -4.57247466e-01 -2.02093154e-01
-9.84101474e-01 8.43295038e-01 9.19689238e-01 -4.20658499e-01
-1.13662434e+00 -9.70618963e-01 -4.99962628e-01 -6.28537685e-02
5.79221129e-01 -3.72869581e-01 1.18017292e+00 -1.23801768e+00
-1.40547764e+00 1.25576186e+00 -1.83221638e-01 -4.53461736e-01
2.59451687e-01 -2.90985918e-04 -6.92513466e-01 1.71345230e-02
3.42410862e-01 8.38433862e-01 1.11284935e+00 -1.69527245e+00
-1.21921051e+00 -6.01876080e-01 -5.62189460e-01 3.00348043e-01
-5.65976620e-01 -3.47730309e-01 -4.82199609e-01 -7.23040700e-01
1.77907974e-01 -1.02145517e+00 -6.50844723e-02 -3.87297243e-01
-9.34654176e-02 -5.66302121e-01 1.03872526e+00 -7.75166824e-02
1.10084486e+00 -2.13508415e+00 1.20498978e-01 1.10838348e-02
4.59402472e-01 4.82538462e-01 -2.43767008e-01 -1.25108108e-01
-4.77043271e-01 -2.54672945e-01 -2.94836938e-01 -1.81498185e-01
-1.53388381e-01 1.16321817e-01 -3.01479071e-01 3.75297129e-01
1.97092533e-01 7.01417446e-01 -1.15470278e+00 -5.18621922e-01
1.71243146e-01 5.77593073e-02 -2.91467160e-01 3.07650059e-01
-3.37923080e-01 8.65578175e-01 -3.26039284e-01 6.53425694e-01
7.06746876e-01 -6.16022229e-01 2.69095302e-01 -2.37126708e-01
1.57522693e-01 -5.01246974e-02 -1.23921120e+00 1.71387458e+00
-1.56704083e-01 2.41109967e-01 -2.38008976e-01 -1.48422444e+00
1.21073759e+00 5.36880605e-02 6.62420392e-01 -1.03325808e+00
-1.62130356e-01 3.35321546e-01 -1.21956870e-01 -1.65183544e-01
2.23472878e-01 -2.79451191e-01 -9.52017903e-02 4.19844300e-01
4.45329219e-01 3.97254884e-01 5.29734381e-02 -3.72962765e-02
7.39445329e-01 4.45225537e-02 4.36153948e-01 -4.85195935e-01
6.69768453e-01 2.43555799e-01 7.58483052e-01 6.94274366e-01
-2.83884764e-01 5.33481658e-01 6.12983145e-02 -2.45244190e-01
-9.38477039e-01 -1.17778707e+00 -3.50138485e-01 1.63307583e+00
5.59138238e-01 3.17321792e-02 -6.22677326e-01 -1.09219277e+00
3.03979874e-01 7.80514121e-01 -6.89136147e-01 -4.48910236e-01
-4.39165324e-01 -8.12168300e-01 2.89653897e-01 6.31487548e-01
6.28146112e-01 -1.02538919e+00 6.37520552e-02 3.77650470e-01
-2.79794931e-01 -9.68874991e-01 -4.23518717e-01 5.59210420e-01
-9.59942102e-01 -1.24247324e+00 -9.51146841e-01 -1.20417285e+00
7.24319220e-01 3.76127422e-01 1.27455544e+00 -4.65946138e-01
-4.34214994e-02 3.14057797e-01 -5.06089270e-01 -2.39986166e-01
-6.44096792e-01 3.55143040e-01 2.94880271e-01 3.17770571e-01
6.47498310e-01 -4.52381492e-01 -3.54781777e-01 7.54918754e-01
-7.45630264e-01 -1.01164289e-01 5.00218630e-01 1.14686811e+00
9.69695389e-01 3.16197157e-01 1.20307946e+00 -1.18561804e+00
1.71668693e-01 -8.00553620e-01 -6.14220917e-01 3.16165060e-01
-1.22895312e+00 2.70985544e-01 4.55594003e-01 -8.69192481e-01
-1.29454815e+00 3.79871100e-01 3.57819498e-01 -5.30617297e-01
-4.39494461e-01 2.07396001e-01 -1.82599515e-01 -3.00700665e-02
9.02629316e-01 3.03165764e-01 -1.23741008e-01 -7.04331279e-01
4.73734796e-01 8.93558204e-01 9.28712308e-01 -8.71584594e-01
9.00198281e-01 4.57616836e-01 -3.59143853e-01 -2.90058255e-01
-1.05575860e+00 -9.30058956e-01 -1.01879680e+00 2.83310958e-03
7.06473529e-01 -1.26806569e+00 -3.95059884e-01 6.21061265e-01
-8.33460748e-01 -3.54870349e-01 -4.94688362e-01 1.51677698e-01
-5.04588306e-01 1.82287410e-01 -2.74214804e-01 -5.04679024e-01
-1.52992934e-01 -1.01694131e+00 9.51548517e-01 3.07328999e-01
1.01338886e-02 -8.05186689e-01 1.35717601e-01 4.46418256e-01
2.38494590e-01 6.79546371e-02 9.61889625e-01 -1.09669423e+00
-2.76051462e-01 1.65498510e-01 -4.13482398e-01 4.39287364e-01
3.63454491e-01 -6.92334235e-01 -1.07138479e+00 -3.89549762e-01
-1.80125996e-01 -4.92497057e-01 9.83074546e-01 1.14664361e-02
1.08647585e+00 -8.34586024e-02 -7.66033590e-01 3.97230059e-01
1.32986367e+00 5.08792520e-01 2.60454208e-01 4.95882750e-01
8.35994303e-01 5.13505697e-01 9.48842347e-01 4.65847224e-01
5.55815399e-01 8.23325396e-01 3.06488156e-01 -8.57342780e-02
-4.02909905e-01 -1.59981668e-01 1.28104314e-01 6.15358591e-01
2.23272651e-01 -1.40394573e-03 -1.10991538e+00 9.05169964e-01
-1.96220624e+00 -7.19901145e-01 2.94105802e-02 2.14099813e+00
1.06672549e+00 1.80567443e-01 3.79351228e-01 4.45058048e-02
1.14426708e+00 -1.47873044e-01 -1.04199731e+00 2.89860010e-01
-1.22174568e-01 1.93211436e-03 5.17214656e-01 1.03447422e-01
-1.42500007e+00 1.02185035e+00 5.33495998e+00 1.13243210e+00
-1.02130997e+00 3.88887823e-01 5.96693695e-01 1.63962901e-01
1.55384183e-01 -2.56786436e-01 -1.15691686e+00 5.87910354e-01
8.42263699e-01 -3.21303993e-01 3.07290345e-01 1.35944498e+00
-3.91750365e-01 -1.76213775e-03 -1.41015553e+00 1.30530834e+00
2.10277677e-01 -1.33497798e+00 9.70821269e-03 -9.00142640e-02
1.12462401e+00 9.12577510e-02 9.05229226e-02 6.18379474e-01
4.72411990e-01 -7.62039185e-01 7.45944977e-01 1.87415794e-01
1.02450013e+00 -8.15055192e-01 6.09424531e-01 4.69894230e-01
-1.28328574e+00 -4.18458998e-01 -3.92862618e-01 3.16635281e-01
-2.41395950e-01 5.80946624e-01 -9.02156353e-01 5.69786549e-01
9.05035913e-01 1.03256786e+00 -7.77124643e-01 9.00310278e-01
2.27384761e-01 6.81482375e-01 -5.29516721e-03 4.61366832e-01
9.69097018e-02 1.53818041e-01 3.93225849e-01 1.22158718e+00
4.63906154e-02 -2.12190207e-02 3.75976145e-01 7.81301379e-01
-2.53468931e-01 -1.28642723e-01 -3.41798544e-01 2.77229194e-02
6.36947870e-01 9.98925507e-01 -8.05495799e-01 -7.42912769e-01
-2.58466005e-01 9.72832382e-01 5.25253236e-01 2.96318084e-01
-7.86918163e-01 -1.18508756e-01 7.02452838e-01 2.03426883e-01
3.88203412e-01 1.18020900e-01 -3.53172123e-01 -1.08662128e+00
-2.83713102e-01 -9.83002245e-01 7.83165991e-01 -3.35385352e-01
-1.90853167e+00 6.15576565e-01 8.37006718e-02 -1.62823272e+00
-2.98120379e-01 -3.98607850e-01 -3.81387956e-03 4.86628801e-01
-1.72540760e+00 -1.05728614e+00 -5.40328920e-01 9.18943703e-01
7.28220761e-01 -6.20997488e-01 8.62103045e-01 5.77434063e-01
-4.64720488e-01 7.78538823e-01 6.49105549e-01 1.55040056e-01
1.25011349e+00 -1.16187644e+00 1.47368774e-01 4.15484488e-01
1.58340678e-01 2.94231087e-01 2.50992656e-01 -6.63121343e-01
-9.65938091e-01 -1.55748463e+00 4.65940297e-01 -6.12214148e-01
5.51018238e-01 -3.10514778e-01 -1.19071722e+00 4.91951555e-01
-1.98442280e-01 2.60925919e-01 3.97915483e-01 1.60366252e-01
-5.74555933e-01 -5.59432030e-01 -1.08105159e+00 1.46639645e-01
1.16516650e+00 -3.91282916e-01 -8.85300696e-01 2.86474258e-01
7.17035234e-01 -3.53782266e-01 -1.00722480e+00 6.39343977e-01
2.67377943e-01 -5.64533412e-01 1.06930542e+00 -5.18106103e-01
1.07375719e-01 -4.92225975e-01 -6.53773174e-03 -1.32815266e+00
-8.06886673e-01 -9.90082771e-02 -1.61536127e-01 1.73101389e+00
3.18291843e-01 -6.31913424e-01 9.00137484e-01 2.95333982e-01
-1.56478539e-01 -2.33059451e-01 -9.02399182e-01 -1.12510014e+00
2.24494591e-01 -2.23998249e-01 4.03635591e-01 1.40028894e+00
-8.59026536e-02 5.14585793e-01 -3.15201402e-01 5.69230020e-02
9.56646621e-01 5.56346238e-01 6.62960827e-01 -1.69078803e+00
-7.54380971e-02 -4.03435647e-01 -1.91676080e-01 -9.35879827e-01
5.39900482e-01 -1.27608109e+00 1.27362981e-01 -1.06396735e+00
6.09244227e-01 -9.62217867e-01 -7.93874204e-01 6.44625187e-01
-3.30129415e-01 3.08231235e-01 2.94120461e-01 5.03893673e-01
-1.18872559e+00 5.04945338e-01 9.87012088e-01 -4.98913437e-01
-3.64390433e-01 -8.05293396e-02 -9.24872100e-01 4.57205176e-01
5.58091700e-01 -8.75199795e-01 -5.59969187e-01 -1.35563314e-01
-4.85523850e-01 -2.17037842e-01 2.99342930e-01 -1.18765283e+00
3.46534014e-01 -1.66303992e-01 4.25772220e-01 -6.09865546e-01
-1.74174067e-02 -9.70509470e-01 2.03989148e-01 2.32815802e-01
-5.57101727e-01 -4.78449643e-01 1.63125485e-01 9.57880139e-01
-3.31900865e-01 -2.96437256e-02 1.29203129e+00 -1.79377750e-01
-1.23727751e+00 2.88117260e-01 -3.56793753e-03 4.48436350e-01
1.27788830e+00 -2.82786518e-01 -3.68142188e-01 1.59870565e-01
-8.33815575e-01 3.31152022e-01 4.63357925e-01 7.96334028e-01
3.19948912e-01 -1.46359193e+00 -7.63649642e-01 2.19724923e-01
6.37447119e-01 1.98798910e-01 3.56315583e-01 5.20638406e-01
2.66463786e-01 3.62146139e-01 -3.25355142e-01 -1.08086824e+00
-1.11630881e+00 9.79102552e-01 2.20950320e-01 -4.43071544e-01
-5.29834270e-01 7.26032734e-01 6.53357983e-01 -5.40161610e-01
4.25106943e-01 7.85400122e-02 -2.50397742e-01 2.95157462e-01
5.38727283e-01 5.00259221e-01 1.64301202e-01 -6.01144612e-01
-4.67658103e-01 6.63895130e-01 -3.24121743e-01 3.74461412e-01
1.26112580e+00 -3.57301027e-01 1.97338313e-01 4.32156831e-01
1.13949478e+00 -5.90052307e-01 -1.40331483e+00 -9.06003296e-01
4.04130876e-01 -1.78457677e-01 -2.57444143e-01 -1.11563039e+00
-1.00102437e+00 6.13006711e-01 1.12594676e+00 -1.34579912e-01
1.27915299e+00 3.67524415e-01 7.04001665e-01 3.82844210e-01
5.92536747e-01 -1.44284296e+00 5.19376516e-01 5.30133665e-01
5.57982385e-01 -1.67736232e+00 -2.22509816e-01 -3.45099568e-01
-9.07162726e-01 8.06521595e-01 9.12826061e-01 -6.45001829e-02
7.19652534e-01 1.73020303e-01 4.79040444e-02 3.85456309e-02
-6.14631832e-01 -3.16640615e-01 4.57661569e-01 9.91469324e-01
9.39801335e-02 1.06110878e-01 1.40267983e-01 9.34098065e-01
3.12137216e-01 5.36962450e-02 -3.50950509e-02 5.53788066e-01
-5.12217999e-01 -1.15791059e+00 -5.56252658e-01 4.68914926e-01
-1.74284294e-01 1.08365111e-01 -1.65091828e-01 7.53262043e-01
5.20144403e-01 8.59297812e-01 1.36718694e-02 -5.11490643e-01
2.75618285e-01 2.99564630e-01 3.09448242e-01 -7.96139657e-01
-4.19783145e-01 6.56974912e-02 -1.53437868e-01 -4.04675305e-01
-6.34627402e-01 -5.87438583e-01 -1.43086457e+00 9.37616229e-02
-3.57913017e-01 1.92576066e-01 2.74552196e-01 8.95682275e-01
6.17774904e-01 4.44724768e-01 6.41155601e-01 -8.84399474e-01
-5.55714726e-01 -9.22870219e-01 -6.67555809e-01 7.24850595e-01
2.68693686e-01 -1.18538082e+00 -1.57880053e-01 4.23845559e-01] | [10.241888999938965, 2.9809975624084473] |
135393ff-6a12-4ff4-884d-97b70983bcd7 | logic-rules-powered-knowledge-graph-embedding | 1903.03772 | null | http://arxiv.org/abs/1903.03772v1 | http://arxiv.org/pdf/1903.03772v1.pdf | Logic Rules Powered Knowledge Graph Embedding | Large scale knowledge graph embedding has attracted much attention from both
academia and industry in the field of Artificial Intelligence. However, most
existing methods concentrate solely on fact triples contained in the given
knowledge graph. Inspired by the fact that logic rules can provide a flexible
and declarative language for expressing rich background knowledge, it is
natural to integrate logic rules into knowledge graph embedding, to transfer
human knowledge to entity and relation embedding, and strengthen the learning
process. In this paper, we propose a novel logic rule-enhanced method which can
be easily integrated with any translation based knowledge graph embedding
model, such as TransE . We first introduce a method to automatically mine the
logic rules and corresponding confidences from the triples. And then, to put
both triples and mined logic rules within the same semantic space, all triples
in the knowledge graph are represented as first-order logic. Finally, we define
several operations on the first-order logic and minimize a global loss over
both of the mined logic rules and the transformed first-order logics. We
conduct extensive experiments for link prediction and triple classification on
three datasets: WN18, FB166, and FB15K. Experiments show that the rule-enhanced
method can significantly improve the performance of several baselines. The
highlight of our model is that the filtered Hits@1, which is a pivotal
evaluation in the knowledge inference task, has a significant improvement (up
to 700% improvement). | ['Pengwei Wang', 'Nisansa de Silva', 'Lianwen Jin', 'Fangzhao Wu', 'Dejing Dou'] | 2019-03-09 | null | null | null | null | ['triple-classification'] | ['graphs'] | [-1.43776983e-01 3.66636217e-01 -5.32705784e-01 -4.05424714e-01
-1.92283958e-01 -4.60099608e-01 3.94847602e-01 2.40072325e-01
-2.10777536e-01 7.18132317e-01 1.29406944e-01 -4.77483690e-01
-4.95513707e-01 -1.47873664e+00 -9.78141665e-01 -3.02083969e-01
1.20413443e-02 2.05450609e-01 5.02047479e-01 -3.61566484e-01
-3.96838278e-01 7.79969022e-02 -1.26782763e+00 4.23829257e-01
1.01506400e+00 1.20926893e+00 -3.48471910e-01 -8.42114240e-02
-3.45428973e-01 1.07712877e+00 -6.63952604e-02 -9.83859122e-01
3.56809003e-04 -3.97941738e-01 -9.26575541e-01 -3.25545162e-01
3.60547081e-02 1.18279201e-03 -6.44472361e-01 1.18778944e+00
6.22100523e-03 -9.09759477e-02 4.05193567e-01 -1.46933484e+00
-8.77456844e-01 1.08064377e+00 -1.32569700e-01 -2.23633185e-01
4.08305854e-01 -1.23352967e-01 1.44429064e+00 -1.01832867e+00
5.91692030e-01 1.34178329e+00 5.53594232e-01 2.00081915e-01
-1.14165747e+00 -6.64330840e-01 1.25206888e-01 7.82939851e-01
-1.47613692e+00 -2.84696482e-02 8.65460396e-01 -3.04042816e-01
9.69093800e-01 2.61256158e-01 7.88677335e-01 5.61322629e-01
-8.56317207e-02 7.65253961e-01 8.37829590e-01 -6.48086488e-01
-7.06986059e-03 4.89929527e-01 4.96891677e-01 1.26115668e+00
5.21854460e-01 3.99705432e-02 -4.66467053e-01 4.53410186e-02
2.81778842e-01 -3.04299369e-02 -3.68507057e-01 -6.00826442e-01
-1.33336663e+00 1.03606451e+00 9.51800585e-01 2.14133814e-01
-1.61641791e-01 2.37576589e-01 4.95575160e-01 3.58490467e-01
2.15166450e-01 2.82535136e-01 -6.40085280e-01 2.05594331e-01
-1.71352789e-01 -1.79367065e-02 9.33540881e-01 8.90536129e-01
1.07982123e+00 -3.35245699e-01 -9.51611102e-02 6.52177453e-01
3.96883726e-01 3.41299593e-01 2.34228343e-01 -4.72070813e-01
4.48046952e-01 1.38421094e+00 -2.03411758e-01 -1.43123198e+00
-4.72534895e-02 -1.17529504e-01 -5.79865932e-01 -2.78163671e-01
4.92722280e-02 -4.28254940e-02 -5.16099989e-01 1.63036501e+00
4.61605281e-01 2.75657237e-01 2.24299550e-01 6.35961592e-01
8.79396975e-01 5.67132354e-01 -1.40352249e-01 -4.36019190e-02
1.45938563e+00 -1.00455761e+00 -7.29277730e-01 5.06098047e-02
9.14605141e-01 -1.36750221e-01 1.04738414e+00 -3.65768857e-02
-6.11329615e-01 -2.58920729e-01 -1.15964031e+00 -8.53956938e-02
-7.55157232e-01 -1.22815864e-02 1.03701496e+00 3.65020901e-01
-6.01484120e-01 4.69166249e-01 -6.39224231e-01 -2.62698025e-01
4.91628051e-01 3.52690935e-01 -5.76778531e-01 -2.69064963e-01
-1.85569692e+00 8.95292342e-01 1.16163743e+00 5.30673675e-02
-2.04874456e-01 -8.40586782e-01 -1.14712071e+00 3.32164437e-01
9.84745085e-01 -7.43330896e-01 6.11058295e-01 -6.25950336e-01
-1.23286605e+00 6.79908037e-01 9.17063057e-02 -4.69957918e-01
1.90503180e-01 -2.36259457e-02 -8.47516596e-01 -9.29395929e-02
-8.15934241e-02 3.68856221e-01 2.22123310e-01 -1.03733099e+00
-7.16854811e-01 -2.24719092e-01 5.22893727e-01 -5.14330491e-02
-8.21475208e-01 -2.98510939e-01 -7.55348563e-01 -3.57371211e-01
-2.63924827e-03 -7.93628812e-01 2.09503934e-01 -3.95405339e-03
-5.21950841e-01 -5.61406434e-01 6.89129174e-01 -5.96864820e-01
1.47469211e+00 -2.13945651e+00 2.71843135e-01 5.66673636e-01
2.84077555e-01 2.89896548e-01 2.19813764e-01 2.37011313e-01
-4.15737182e-02 2.88803428e-01 -2.25224882e-01 4.04862732e-01
4.75978345e-01 5.62559009e-01 -4.91746008e-01 -6.66086599e-02
2.53203422e-01 1.33540785e+00 -1.07656574e+00 -6.04453981e-01
1.91927657e-01 3.51655483e-01 -6.83713436e-01 -9.08668041e-02
-5.02925694e-01 -3.49447727e-01 -5.27809501e-01 5.61603308e-01
3.52296859e-01 -3.75646383e-01 6.75955176e-01 -7.55290747e-01
3.70907903e-01 3.60330313e-01 -1.15433204e+00 1.34074903e+00
-4.19824511e-01 2.03867510e-01 -6.54637098e-01 -1.10590303e+00
9.35281813e-01 1.55243859e-01 3.17521662e-01 -5.79331040e-01
-2.03407276e-03 2.34196156e-01 -4.34904210e-02 -3.35959762e-01
2.35203370e-01 -2.39419654e-01 -1.47174045e-01 8.91904384e-02
1.29353225e-01 3.56254578e-01 3.64925534e-01 4.16859895e-01
9.42727268e-01 3.27451490e-02 3.14243555e-01 -1.04738496e-01
6.99809134e-01 2.24078521e-02 8.31593096e-01 1.39525861e-01
2.00485528e-01 -4.44664568e-01 8.67004097e-01 -5.96654236e-01
-3.97256166e-01 -1.05599642e+00 4.87813130e-02 7.10529983e-01
2.68403500e-01 -8.73694956e-01 -2.92688876e-01 -1.28074026e+00
4.00654674e-01 6.32886827e-01 -6.21426105e-01 -7.01549232e-01
-3.50556493e-01 -5.48969805e-01 5.78893423e-01 7.09060311e-01
8.06280911e-01 -9.19916272e-01 2.47986168e-02 9.96874198e-02
-3.21071893e-01 -1.28125966e+00 -2.93031067e-01 4.76963110e-02
-6.35750830e-01 -1.38781238e+00 3.34149152e-01 -9.37663674e-01
5.90389490e-01 -1.36918485e-01 1.03906226e+00 3.59250277e-01
2.76559517e-02 4.87436838e-02 -4.25321877e-01 -1.81169391e-01
-2.23278746e-01 -4.53468598e-02 1.42529920e-01 1.90662637e-01
7.51745343e-01 -4.16053057e-01 -1.34169772e-01 2.19476208e-01
-9.54356730e-01 1.47999659e-01 4.76329029e-01 9.05710518e-01
7.60885417e-01 5.69157898e-01 5.85686684e-01 -1.01532388e+00
4.76189613e-01 -2.84374893e-01 -6.56513989e-01 7.23500848e-01
-9.86223638e-01 4.18013781e-01 5.99459052e-01 -5.00164442e-02
-7.83353567e-01 -6.35077059e-02 1.90570965e-01 -4.84664530e-01
3.74915898e-01 1.17026615e+00 -7.02259064e-01 -1.15806647e-01
3.90742481e-01 7.01672137e-02 -2.17715189e-01 -2.80256599e-01
8.08018208e-01 3.32282841e-01 4.02629912e-01 -7.05094039e-01
1.00444949e+00 2.51192898e-01 7.83639401e-02 -3.50787789e-01
-1.09191203e+00 -2.26442311e-02 -4.28249091e-01 2.46380702e-01
6.97181463e-01 -7.38867044e-01 -1.04560256e+00 -8.74035433e-02
-9.40773487e-01 -1.46018136e-02 -3.02711219e-01 5.19569695e-01
-1.94707453e-01 4.30068105e-01 -4.26837981e-01 -3.38249236e-01
-2.65841424e-01 -9.09191549e-01 4.65564519e-01 1.87704951e-01
9.52413026e-03 -1.28509808e+00 -7.29086772e-02 2.31456533e-01
-7.03881728e-03 2.80604661e-01 1.59229815e+00 -6.74218178e-01
-6.50932610e-01 -1.89444438e-01 -5.05125284e-01 5.49133956e-01
3.02897394e-01 -6.37594536e-02 -5.02500236e-01 -1.24837393e-02
-5.29994547e-01 -4.15896535e-01 1.03009856e+00 -3.15454930e-01
1.06676662e+00 -5.61155856e-01 -5.74606121e-01 6.44371808e-01
1.58953834e+00 -7.03066140e-02 5.85867584e-01 3.12097460e-01
9.43718016e-01 3.08383107e-01 5.79982579e-01 4.97412570e-02
7.61523545e-01 8.05469513e-01 2.22906455e-01 1.84299588e-01
9.17158648e-02 -7.40530431e-01 4.04322445e-01 9.46331978e-01
-1.63494851e-02 1.34077176e-01 -8.13935041e-01 5.37931859e-01
-2.06567144e+00 -1.02043664e+00 3.04938331e-02 1.99846101e+00
1.33419120e+00 2.63468385e-01 -6.43100142e-02 3.41638625e-01
5.33841789e-01 -1.02485314e-01 -2.73012370e-01 -1.51087224e-01
-9.14800018e-02 2.47922912e-01 3.28842342e-01 6.55928731e-01
-9.97403860e-01 1.09559381e+00 5.07678461e+00 8.04291606e-01
-9.63849962e-01 -2.50719711e-02 4.65377606e-02 1.23440839e-01
-6.17205620e-01 3.22175741e-01 -7.58475125e-01 5.39887011e-01
6.32339835e-01 -5.12867391e-01 5.29673040e-01 7.70729840e-01
-3.88619035e-01 3.43958974e-01 -1.49199510e+00 7.59658515e-01
-6.46692365e-02 -1.49743414e+00 2.49106169e-01 4.23747301e-02
6.25516176e-01 -3.81398290e-01 -2.24685878e-01 8.58935833e-01
4.27394807e-01 -1.00120997e+00 4.18376714e-01 5.55484951e-01
8.28041971e-01 -9.53319132e-01 8.66752625e-01 7.88984373e-02
-1.35648191e+00 -5.83371259e-02 -4.15722013e-01 4.97284606e-02
-5.30679449e-02 8.73912036e-01 -7.22704828e-01 1.15125811e+00
5.55508256e-01 9.86111820e-01 -5.71314752e-01 5.65654695e-01
-7.31760323e-01 4.88197476e-01 -2.39638016e-01 -2.06778690e-01
6.01436570e-02 -2.49564990e-01 2.00165719e-01 1.15156627e+00
-3.74285411e-03 1.34143785e-01 3.23143452e-01 9.63850498e-01
-5.80333173e-01 5.86720034e-02 -6.05289400e-01 -3.28403801e-01
4.81404334e-01 1.29283547e+00 -3.89656514e-01 -4.29879695e-01
-6.35487735e-01 8.28404069e-01 6.33807182e-01 3.49744678e-01
-1.11167264e+00 -7.64870524e-01 6.32674873e-01 -1.71642020e-01
5.48984051e-01 1.31779641e-01 -3.62513550e-02 -1.33266103e+00
4.63452339e-01 -6.21567607e-01 6.75552905e-01 -5.13028264e-01
-1.32173729e+00 3.36118340e-01 2.37162374e-02 -7.75677860e-01
7.92992301e-03 -7.65417576e-01 -3.87830496e-01 6.34346604e-01
-1.79118562e+00 -1.25467980e+00 -2.21056044e-01 7.16480076e-01
-3.29225987e-01 -1.04954131e-01 9.31409299e-01 5.57073474e-01
-5.26733696e-01 8.94659281e-01 -1.92622662e-01 5.16128182e-01
4.09868807e-01 -1.19769287e+00 -1.73030034e-01 6.49618864e-01
3.63833249e-01 8.96630049e-01 2.55593538e-01 -6.52656019e-01
-1.65300047e+00 -1.48098361e+00 1.10219038e+00 -5.49524069e-01
9.65389609e-01 -2.35456005e-01 -1.01409304e+00 1.08693397e+00
-1.31591976e-01 4.54707712e-01 6.90039575e-01 4.04211193e-01
-7.32802689e-01 -4.88618135e-01 -8.89563382e-01 5.26078522e-01
1.07401371e+00 -6.81251705e-01 -9.42323923e-01 2.94551730e-01
1.12016785e+00 -1.90392971e-01 -1.15432858e+00 7.17019320e-01
5.30934632e-01 -5.17560124e-01 1.00065172e+00 -9.87196624e-01
5.32274306e-01 -5.67054927e-01 -3.27541381e-01 -1.30516016e+00
-3.45438093e-01 -1.41098127e-01 -7.12760448e-01 1.32447863e+00
6.63865566e-01 -6.74481273e-01 6.37460053e-01 5.05793452e-01
7.82259107e-02 -1.07014704e+00 -7.77571619e-01 -1.03210163e+00
-1.63230389e-01 -2.48324439e-01 7.75382578e-01 1.31195807e+00
5.16087830e-01 6.34006202e-01 -2.57204115e-01 1.41612276e-01
4.31347907e-01 4.78161395e-01 5.55243492e-01 -1.33450150e+00
-2.89572477e-01 -3.23247403e-01 -8.04229259e-01 -6.57667637e-01
4.46886420e-01 -1.61699390e+00 -3.15453261e-01 -1.85223019e+00
2.37748384e-01 -4.42647159e-01 -6.86530709e-01 1.02987802e+00
-2.60573596e-01 -4.61060293e-02 8.11774954e-02 -3.04156333e-01
-6.78689897e-01 8.35365415e-01 1.01709604e+00 -3.80227983e-01
6.08912222e-02 -5.25353670e-01 -1.02529097e+00 7.09017277e-01
4.89891291e-01 -3.62823546e-01 -5.70393801e-01 -3.77996981e-01
6.56523228e-01 -3.49642217e-01 5.80567181e-01 -6.49441481e-01
3.14877868e-01 -2.26514295e-01 6.37241919e-03 -2.18621299e-01
-1.24416407e-02 -1.14149320e+00 1.30520120e-01 4.80666518e-01
-6.98428750e-02 -2.96972066e-01 1.39243349e-01 7.81524122e-01
-3.84517342e-01 6.78861886e-02 4.68558490e-01 2.05504492e-01
-1.00200355e+00 1.97043315e-01 5.20040870e-01 8.32360461e-02
1.14881265e+00 9.79614109e-02 -4.30010527e-01 3.70750725e-02
-6.21109903e-01 5.12145340e-01 2.13994190e-01 5.21682680e-01
7.17658758e-01 -1.91906238e+00 -4.54685479e-01 1.66841596e-01
4.82268959e-01 -8.77683908e-02 -1.43704250e-01 8.74060631e-01
-2.12065622e-01 5.92804074e-01 4.44866084e-02 -2.12381214e-01
-9.92524028e-01 7.61581123e-01 3.19500059e-01 -4.96346414e-01
-4.68769193e-01 6.97884738e-01 -6.75548092e-02 -7.18194842e-01
1.72896072e-01 -6.00717604e-01 -6.82966784e-02 -6.61965013e-02
3.23653817e-01 2.71950722e-01 8.40017796e-02 -2.50371486e-01
-7.25568593e-01 4.18086618e-01 1.07796397e-02 3.29255700e-01
1.19526041e+00 3.51632476e-01 -5.84493160e-01 3.45060647e-01
1.19608116e+00 1.37717128e-01 -5.53959489e-01 -6.91268802e-01
2.37185284e-01 -3.86275262e-01 5.70871457e-02 -8.94266963e-01
-1.17290521e+00 6.19135261e-01 1.25716582e-01 7.79504254e-02
9.58697557e-01 2.15652600e-01 8.07882488e-01 8.47289383e-01
3.84218842e-01 -8.41298699e-01 -8.08317363e-02 6.67624116e-01
6.15388036e-01 -1.13689923e+00 -4.97135846e-03 -8.05474281e-01
-4.96990353e-01 8.87975633e-01 7.06186771e-01 1.81537420e-01
6.56132460e-01 -2.72733849e-02 -4.59986329e-01 -4.11186397e-01
-7.08930731e-01 -1.55355752e-01 5.80198705e-01 2.45193347e-01
3.43569547e-01 2.84273654e-01 -3.93891275e-01 1.13868606e+00
-1.52430907e-01 3.85765821e-01 4.62023579e-02 6.60142958e-01
-4.52207267e-01 -1.19310617e+00 1.66256055e-01 6.15221560e-01
-1.41243404e-02 -1.74186930e-01 -4.62532878e-01 8.46676767e-01
3.10566217e-01 6.97355032e-01 -3.39339912e-01 -8.91022384e-01
6.73522890e-01 3.80357116e-01 4.45078164e-01 -6.63378060e-01
-1.59440681e-01 -6.50048435e-01 2.33707115e-01 -4.54640090e-01
-2.14565560e-01 -1.52793825e-01 -1.65336132e+00 -4.79003489e-01
-5.40837705e-01 4.04432058e-01 1.12394691e-01 1.03654373e+00
4.08768803e-01 6.13124251e-01 4.24349099e-01 3.01717162e-01
-4.15547669e-01 -5.47925353e-01 -5.54452717e-01 4.93934214e-01
-1.62623167e-01 -8.22974265e-01 -1.50969118e-01 3.89069840e-02] | [8.83074951171875, 7.819396495819092] |
779a2855-b005-46af-9d9a-21c1ca859c35 | 4dhumanoutfit-a-multi-subject-4d-dataset-of | 2306.07399 | null | https://arxiv.org/abs/2306.07399v1 | https://arxiv.org/pdf/2306.07399v1.pdf | 4DHumanOutfit: a multi-subject 4D dataset of human motion sequences in varying outfits exhibiting large displacements | This work presents 4DHumanOutfit, a new dataset of densely sampled spatio-temporal 4D human motion data of different actors, outfits and motions. The dataset is designed to contain different actors wearing different outfits while performing different motions in each outfit. In this way, the dataset can be seen as a cube of data containing 4D motion sequences along 3 axes with identity, outfit and motion. This rich dataset has numerous potential applications for the processing and creation of digital humans, e.g. augmented reality, avatar creation and virtual try on. 4DHumanOutfit is released for research purposes at https://kinovis.inria.fr/4dhumanoutfit/. In addition to image data and 4D reconstructions, the dataset includes reference solutions for each axis. We present independent baselines along each axis that demonstrate the value of these reference solutions for evaluation tasks. | ['Stefanie Wuhrer', 'Anilkumar Swamy', 'Gregory Rogez', 'Rim Rekik', 'Sergi Pujades', 'Julien Pansiot', 'Mathieu Marsot', 'Vincent Leroy', 'Christophe Legras', 'Martin Humenberger', 'Jean-Sebastien Franco', 'Edmond Boyer', 'Laurence Boissieux', 'Matthieu Armando'] | 2023-06-12 | null | null | null | null | ['virtual-try-on'] | ['computer-vision'] | [-2.89024532e-01 -2.04398707e-01 3.64034176e-02 -2.68685408e-02
-3.23289424e-01 -6.50516093e-01 8.22939992e-01 -5.06220281e-01
-2.80570447e-01 1.45517200e-01 6.88206911e-01 3.04366738e-01
2.14106098e-01 -4.11946803e-01 -5.80391347e-01 -4.46224630e-01
-8.68340861e-03 5.20366251e-01 1.02938570e-01 -4.02414203e-01
-8.19471925e-02 6.49327517e-01 -1.30264497e+00 3.10888886e-01
-1.32442474e-01 8.03821981e-01 1.48571776e-02 1.05234218e+00
5.31958163e-01 8.54050815e-01 -5.90022087e-01 -5.48868120e-01
6.95743084e-01 -5.54367900e-01 -5.28243482e-01 3.62150073e-01
7.14652717e-01 -4.72183138e-01 -9.85790014e-01 4.82359380e-01
9.06986952e-01 5.35575628e-01 4.18616802e-01 -1.34663653e+00
-6.59299612e-01 2.76203930e-01 -6.76558614e-01 1.46072179e-01
1.16243923e+00 5.34341872e-01 6.54819369e-01 -1.00159657e+00
1.18752754e+00 1.54696393e+00 8.06533158e-01 5.97369969e-01
-1.06189525e+00 -5.37862182e-01 -3.14018786e-01 1.77804872e-01
-1.37539029e+00 -6.21803343e-01 8.68235171e-01 -7.42800832e-01
6.01326168e-01 4.44939137e-01 1.34072995e+00 1.88756776e+00
-7.71021587e-04 1.15155113e+00 6.79961145e-01 -1.27748474e-01
1.29540667e-01 -3.04389745e-01 -1.47238880e-01 4.31151479e-01
-9.91260484e-02 5.27301356e-02 -7.44943559e-01 -1.92378372e-01
1.47802329e+00 1.02861561e-02 -2.94618726e-01 -6.36516273e-01
-1.84768307e+00 4.55593586e-01 2.34241605e-01 1.16264969e-01
-4.40417260e-01 3.75222266e-01 4.53199804e-01 1.08318038e-01
1.20540582e-01 3.18405122e-01 -1.31624430e-01 -4.73376691e-01
-6.26303136e-01 9.31973994e-01 4.04521525e-01 1.20346034e+00
2.60428697e-01 5.70530574e-05 -2.05557883e-01 6.51736557e-01
-8.94026831e-03 6.94927931e-01 3.57183903e-01 -1.50046229e+00
4.54158217e-01 2.90919393e-01 4.83360440e-01 -1.28864610e+00
-5.44739723e-01 2.10161000e-01 -8.35538328e-01 9.90872458e-02
5.41738153e-01 5.09106256e-02 -7.31619060e-01 1.36084294e+00
6.18378580e-01 4.89382476e-01 -1.50915444e-01 1.28157580e+00
1.05521810e+00 5.74791014e-01 -1.30644843e-01 2.25039944e-01
1.41591477e+00 -1.00260198e+00 -8.71736705e-01 7.54504744e-03
5.09229362e-01 -5.89125633e-01 1.26628828e+00 6.55743778e-01
-1.24322629e+00 -7.39722252e-01 -8.12920511e-01 -4.60061669e-01
8.13361406e-02 9.10816193e-02 4.31225955e-01 4.29964662e-01
-9.30444658e-01 4.95897174e-01 -7.21661985e-01 -5.45766175e-01
2.84658730e-01 -1.97524205e-02 -9.92675483e-01 -1.49589218e-02
-8.73946190e-01 6.46064520e-01 1.50872171e-01 1.33796424e-01
-9.20066535e-01 -4.35022950e-01 -1.04727161e+00 -7.23843098e-01
3.00707698e-01 -8.40681851e-01 1.37232482e+00 -4.94634748e-01
-1.33003938e+00 1.23619258e+00 7.97350481e-02 -2.04167441e-01
1.20593715e+00 -5.57052374e-01 -3.71233195e-01 1.12940118e-01
1.61002830e-01 4.83404130e-01 7.33032525e-01 -1.08910561e+00
-2.92627752e-01 -5.10609984e-01 -1.18580371e-01 2.35535219e-01
1.70171082e-01 1.08413562e-01 -8.95272791e-01 -1.12969398e+00
2.70821787e-02 -1.22600913e+00 -2.40936503e-01 9.14425924e-02
-7.09474862e-01 2.32724138e-02 1.00531101e+00 -8.52033436e-01
1.04648054e+00 -2.28892279e+00 7.03850269e-01 1.26023769e-01
5.57608843e-01 1.76495686e-02 -1.33896872e-01 4.30767626e-01
-5.72448596e-03 5.20032384e-02 -1.25799850e-01 -8.30635369e-01
-1.22823846e-02 3.61600891e-02 1.78103104e-01 8.33171904e-01
-2.99599677e-01 1.14231837e+00 -8.99606407e-01 -3.88960809e-01
5.31741798e-01 6.04858220e-01 -3.80377561e-01 3.61071974e-01
1.07615545e-01 1.10807455e+00 -2.27865711e-01 6.26783371e-01
4.99431282e-01 1.21505380e-01 -1.20211773e-01 -2.08940715e-01
1.63997598e-02 -4.02394265e-01 -1.33578479e+00 2.25623202e+00
9.71697122e-02 7.69362271e-01 -1.74503043e-01 -1.80998147e-01
7.82687426e-01 4.83468980e-01 8.85278583e-01 -5.52678347e-01
3.50271761e-01 -2.72912215e-02 -4.93047684e-01 -9.79346991e-01
9.04451251e-01 1.73986927e-01 -3.87668848e-01 4.96109426e-01
-1.18681967e-01 -2.28132591e-01 4.88505252e-02 1.37498781e-01
1.10979426e+00 4.06160057e-01 4.64159131e-01 1.47640467e-01
1.88659176e-01 3.07806041e-02 3.42434019e-01 4.42710966e-01
-5.90924919e-01 1.18104637e+00 2.21983567e-01 -8.05403709e-01
-1.59067810e+00 -1.12662208e+00 2.44642422e-01 8.11700702e-01
3.58246535e-01 -5.20000756e-01 -5.84568441e-01 -3.00490290e-01
-2.67469529e-02 2.65363097e-01 -9.87151265e-01 2.53867060e-01
-9.28818643e-01 -1.36979789e-01 8.30099404e-01 6.02350116e-01
5.01387477e-01 -1.17671764e+00 -1.08417106e+00 -1.92882836e-01
-5.43345451e-01 -1.26982057e+00 -7.05428123e-01 -7.23093569e-01
-5.65385997e-01 -1.20435476e+00 -1.09755790e+00 -4.02210206e-01
1.62689418e-01 5.97785115e-01 1.10823858e+00 -2.23891810e-01
-2.21996918e-01 8.22525024e-01 -6.77741826e-01 -8.72386396e-02
-2.25441143e-01 -3.26109439e-01 3.57550561e-01 2.08029315e-01
2.55877408e-03 -4.87693042e-01 -8.75711501e-01 5.72129667e-01
-6.59995198e-01 1.90972507e-01 -4.18798514e-02 4.49087292e-01
5.61490178e-01 -4.63381499e-01 -2.31031254e-01 -6.18209183e-01
5.03913283e-01 -5.91701627e-01 6.94517046e-02 -1.19583001e-02
6.28388524e-01 -4.86489266e-01 1.80952311e-01 -6.55440986e-01
-8.24070215e-01 2.21138850e-01 -6.90838546e-02 -9.88944530e-01
-4.37993139e-01 -2.12958038e-01 -2.66154438e-01 2.73897767e-01
1.01869595e+00 -2.14346424e-01 -1.07438877e-01 -5.94486892e-01
6.14360750e-01 3.69112790e-01 1.02105570e+00 -3.47332925e-01
7.41862774e-01 8.63376856e-01 -1.92619100e-01 -9.56489742e-01
-1.76879883e-01 -4.34681475e-01 -1.14716220e+00 -7.18316913e-01
1.22019339e+00 -9.96629894e-01 -6.13883913e-01 7.85121262e-01
-1.21205735e+00 -6.43337429e-01 -5.49444377e-01 5.11739850e-01
-8.00083578e-01 3.87875289e-01 -6.00475252e-01 -9.08034384e-01
5.82734570e-02 -1.03085923e+00 1.41338658e+00 -1.34404808e-01
-8.37315023e-01 -9.79246557e-01 3.78653735e-01 5.13376296e-01
-3.11042219e-01 1.17331839e+00 6.62069842e-02 -2.17211574e-01
-3.02279204e-01 -3.91515702e-01 3.36732954e-01 -1.33826490e-02
1.80725213e-02 8.92047137e-02 -9.71607149e-01 -1.79312199e-01
-1.68267265e-01 -3.63401175e-01 4.60562438e-01 4.53230053e-01
5.50713301e-01 -2.44186193e-01 -9.86053422e-02 7.34108031e-01
8.72919261e-01 2.28557348e-01 1.03001988e+00 4.85160261e-01
1.24507022e+00 7.88892686e-01 6.13799751e-01 8.24920833e-01
5.79784572e-01 1.12229514e+00 1.71155974e-01 6.26134276e-02
-2.25799292e-01 -3.98939103e-01 4.85552520e-01 6.11121833e-01
-6.77365065e-01 -2.56003886e-01 -1.07419598e+00 5.47893941e-01
-1.89969230e+00 -1.38618565e+00 -5.91722965e-01 2.02519464e+00
9.77107212e-02 -5.33979952e-01 8.19269061e-01 4.36524749e-01
6.09801590e-01 3.75051528e-01 -5.31376719e-01 5.07513024e-02
-5.04576027e-01 -1.69855803e-01 5.50240241e-02 2.31437787e-01
-9.91089761e-01 7.97108650e-01 6.03330135e+00 3.94656152e-01
-7.99303114e-01 1.51767775e-01 3.13352197e-01 -4.66816723e-01
-1.04214527e-01 -2.72299647e-01 -2.10357174e-01 3.60350311e-01
4.60417539e-01 1.63630303e-02 2.64560282e-01 6.20279729e-01
5.70560038e-01 -1.92170039e-01 -1.14633644e+00 1.66559565e+00
2.25605980e-01 -1.16338539e+00 -2.31988192e-01 9.93181765e-02
7.78604090e-01 -2.61520803e-01 9.72030312e-02 -1.09096617e-01
2.16875002e-01 -8.88687849e-01 1.46832228e+00 6.55399621e-01
8.98450732e-01 -4.87274647e-01 3.52059156e-01 3.24109226e-01
-1.17215681e+00 2.62232255e-02 4.53350097e-02 -1.49726599e-01
8.09717774e-01 5.91776632e-02 -4.31406558e-01 5.55111945e-01
9.86704469e-01 1.16920483e+00 -6.82648778e-01 8.77707839e-01
7.05296695e-02 -6.36412650e-02 -1.19622573e-01 3.81249398e-01
-9.61059481e-02 -3.33897650e-01 7.35400498e-01 1.14062464e+00
5.38193822e-01 3.45839232e-01 1.19260848e-01 3.57463300e-01
1.14791311e-01 -3.84075046e-02 -1.02007508e+00 1.48509964e-01
2.90409982e-01 8.37721169e-01 -6.68294191e-01 -2.81176507e-01
5.05610742e-02 1.49222612e+00 -2.33341325e-02 3.14089686e-01
-9.99681413e-01 -2.42729597e-02 8.36558700e-01 4.55834031e-01
-1.05268441e-01 -7.80501485e-01 -1.85911700e-01 -1.38072658e+00
2.52790991e-02 -8.71974528e-01 4.05930817e-01 -1.15675271e+00
-1.11052656e+00 6.87912941e-01 4.14378196e-01 -1.44333589e+00
-3.98927122e-01 -3.61920208e-01 -2.89594054e-01 3.05055052e-01
-2.93599606e-01 -1.30483389e+00 -7.60698617e-01 9.95671928e-01
6.91961586e-01 -1.18849307e-01 4.15881813e-01 3.18293601e-01
-6.16835237e-01 3.66240472e-01 -2.30329722e-01 3.07586581e-01
6.50553823e-01 -9.74104702e-01 1.10087121e+00 7.37031996e-01
3.54664832e-01 3.45255256e-01 8.70454192e-01 -7.39547551e-01
-1.63712633e+00 -9.20769691e-01 4.80853021e-01 -1.28220522e+00
3.28409582e-01 -6.78277254e-01 -5.85156560e-01 1.11565363e+00
1.09679624e-01 1.84536919e-01 6.11938775e-01 -4.49368864e-01
-1.57500565e-01 3.60549718e-01 -1.20990944e+00 9.96616960e-01
1.83693326e+00 -3.32433730e-01 -2.73034424e-01 1.75135955e-01
4.53384817e-01 -9.67880070e-01 -1.04373717e+00 8.28824863e-02
1.00898862e+00 -1.50187147e+00 1.18288863e+00 -4.47526604e-01
4.21461701e-01 -3.72920215e-01 -3.20571870e-01 -1.13296425e+00
-3.22154641e-01 -9.83741283e-01 -4.61165398e-01 8.38938355e-01
-3.05278122e-01 -9.13195908e-02 7.97328711e-01 6.55646205e-01
1.18825153e-01 -4.25805688e-01 -7.41849065e-01 -6.61403239e-01
-3.00536960e-01 -9.65402246e-01 6.59852207e-01 1.03199029e+00
-2.66806990e-01 1.84842855e-01 -1.11405516e+00 -2.73572832e-01
6.85539424e-01 -4.47189867e-01 1.71488786e+00 -1.06395304e+00
-2.76398689e-01 -2.08585173e-01 -8.32459688e-01 -1.21113896e+00
-5.72947860e-02 -5.49137354e-01 -2.87719697e-01 -1.40024829e+00
-1.23159796e-01 -3.63161683e-01 6.13981426e-01 2.09960535e-01
4.55033667e-02 6.26872241e-01 6.84550285e-01 6.00839794e-01
-5.17768383e-01 6.49270594e-01 1.55125976e+00 1.02760591e-01
-3.53812724e-01 -2.22227842e-01 -1.29167929e-01 6.69978380e-01
6.90238118e-01 -2.04484314e-01 -4.12526786e-01 -5.05067527e-01
-2.73180306e-02 3.47681642e-01 8.97201717e-01 -1.18777680e+00
-1.43125057e-02 -2.38029972e-01 5.45579910e-01 -6.06542647e-01
9.61614907e-01 -5.87389827e-01 1.13007569e+00 2.45914534e-01
-1.46038681e-01 6.74992979e-01 -1.03506349e-01 6.24968529e-01
1.31689370e-01 4.52527314e-01 5.04351914e-01 -5.55141389e-01
-9.63186383e-01 3.80185336e-01 -2.54495084e-01 1.16683826e-01
1.36728358e+00 -7.76280642e-01 1.35037586e-01 -6.13301039e-01
-9.51276958e-01 -3.08676772e-02 9.16040123e-01 7.52514720e-01
9.68139827e-01 -1.76198232e+00 -8.74102950e-01 1.29465953e-01
1.94717199e-01 5.05459197e-02 8.22444201e-01 7.16029406e-01
-1.02954376e+00 -4.65791784e-02 -6.37980223e-01 -7.48531163e-01
-1.41709864e+00 5.54591358e-01 2.30671421e-01 1.83686242e-01
-1.15083587e+00 7.97400832e-01 1.74876720e-01 -4.05375302e-01
3.62940668e-03 -1.25064954e-01 -8.22345838e-02 -2.73352508e-02
7.52146065e-01 9.69614983e-01 -3.75261098e-01 -1.46273351e+00
-1.90994188e-01 7.34705806e-01 5.25385737e-01 -6.36116743e-01
1.07495201e+00 -2.67331481e-01 2.50229865e-01 7.08499908e-01
1.08059371e+00 9.44419280e-02 -1.44659209e+00 4.85521890e-02
-1.93099856e-01 -1.11554766e+00 -7.30920494e-01 -1.61118537e-01
-1.08332586e+00 6.33800507e-01 4.52212483e-01 -1.00709878e-01
8.32486272e-01 1.34965584e-01 1.03904939e+00 5.78410998e-02
7.25760281e-01 -7.24079311e-01 4.11963850e-01 2.47375444e-01
1.36267316e+00 -1.00028014e+00 -5.97675405e-02 -1.70598477e-01
-1.22072530e+00 7.83698857e-01 5.51015317e-01 -3.71991664e-01
4.05161709e-01 9.99806896e-02 3.08527082e-01 -3.13421249e-01
-3.33828956e-01 9.75375250e-02 1.76103026e-01 8.89645696e-01
2.04255834e-01 2.02431470e-01 5.32367900e-02 4.87644315e-01
-7.62635112e-01 -9.34721380e-02 7.44813621e-01 9.72176790e-01
3.29246998e-01 -8.56175184e-01 -9.06777501e-01 -6.75336048e-02
-4.40167356e-03 6.73648119e-01 -6.75548553e-01 9.14636910e-01
3.44503701e-01 8.83387387e-01 6.65946901e-02 -8.58648598e-01
9.20250237e-01 -3.97596478e-01 7.19529986e-01 -2.10594296e-01
-7.90953219e-01 6.16262332e-02 1.20460056e-01 -8.48627031e-01
-5.28153002e-01 -1.04655957e+00 -9.93974388e-01 -8.42651546e-01
3.91140550e-01 -2.65464872e-01 3.50155115e-01 3.29360306e-01
3.55857253e-01 1.06949888e-01 3.90200227e-01 -1.73981082e+00
1.48167700e-01 -8.66855681e-01 -7.43456066e-01 1.00007629e+00
4.58693922e-01 -8.91807735e-01 1.29695833e-01 3.96013796e-01] | [7.202971935272217, -0.7425932884216309] |
d042c308-a1e3-4358-b6c6-38f8bd49537c | multi-task-pre-training-of-modular-prompt-for | 2210.07565 | null | https://arxiv.org/abs/2210.07565v3 | https://arxiv.org/pdf/2210.07565v3.pdf | Multitask Pre-training of Modular Prompt for Chinese Few-Shot Learning | Prompt tuning is a parameter-efficient approach to adapting pre-trained language models to downstream tasks. Although prompt tuning has been shown to match the performance of full model tuning when training data is sufficient, it tends to struggle in few-shot learning settings. In this paper, we present Multi-task Pre-trained Modular Prompt (MP2) to boost prompt tuning for few-shot learning. MP2 is a set of combinable prompts pre-trained on 38 Chinese tasks. On downstream tasks, the pre-trained prompts are selectively activated and combined, leading to strong compositional generalization to unseen tasks. To bridge the gap between pre-training and fine-tuning, we formulate upstream and downstream tasks into a unified machine reading comprehension task. Extensive experiments under two learning paradigms, i.e., gradient descent and black-box tuning, show that MP2 significantly outperforms prompt tuning, full model tuning, and prior prompt pre-training methods in few-shot settings. In addition, we demonstrate that MP2 can achieve surprisingly fast and strong adaptation to downstream tasks by merely learning 8 parameters to combine the pre-trained modular prompts. | ['Xuanjing Huang', 'Xipeng Qiu', 'Qin Zhu', 'Zhengfu He', 'Tianxiang Sun'] | 2022-10-14 | null | null | null | null | ['machine-reading-comprehension'] | ['natural-language-processing'] | [ 5.54748416e-01 2.80796856e-01 -8.84366706e-02 -5.38701594e-01
-1.03811979e+00 -6.12060010e-01 5.78456819e-01 3.01078916e-01
-6.90946400e-01 5.01433194e-01 5.62450290e-01 -5.72129607e-01
2.07235426e-01 -5.94093919e-01 -5.07153571e-01 -3.63376558e-01
4.17211562e-01 3.70211899e-01 6.76712990e-01 -7.18155444e-01
2.09616557e-01 -2.08057076e-01 -1.42557657e+00 4.62179095e-01
1.00122750e+00 5.00732541e-01 7.49873281e-01 1.08775675e+00
-3.58973980e-01 6.19058073e-01 -4.52202022e-01 -2.66401768e-01
-9.88957360e-02 -6.40529394e-01 -1.07474172e+00 -4.14900959e-01
3.70291680e-01 -4.96394694e-01 -1.99618682e-01 7.14946985e-01
8.22688937e-01 7.02178776e-01 7.59064257e-01 -6.94064081e-01
-1.14691651e+00 1.01603520e+00 -2.51140714e-01 7.13241577e-01
4.10399660e-02 3.94934982e-01 1.10051835e+00 -1.20046568e+00
4.00520504e-01 1.05738115e+00 5.49532652e-01 1.15749860e+00
-1.39416456e+00 -4.05952930e-01 4.71925229e-01 3.63666445e-01
-8.68138969e-01 -6.89557850e-01 4.08951879e-01 -5.19641519e-01
1.53247201e+00 -1.50527552e-01 1.91738725e-01 1.41388988e+00
2.09032255e-03 8.74380350e-01 7.88011372e-01 -9.74127650e-01
3.86801481e-01 -1.30496249e-01 7.26208627e-01 5.07490158e-01
-1.76093787e-01 1.64806038e-01 -8.22035789e-01 -1.93227336e-01
4.60735947e-01 -9.97320339e-02 -3.79390806e-01 1.80765405e-01
-1.07315111e+00 8.13237667e-01 1.60568565e-01 3.13600808e-01
-1.64506271e-01 3.77760194e-02 5.09185314e-01 6.13737702e-01
4.90184963e-01 7.64864028e-01 -1.03535450e+00 -2.96156973e-01
-5.76342285e-01 -7.46390522e-02 6.99084103e-01 1.32006550e+00
8.13712597e-01 2.78322492e-02 -1.03257012e+00 1.09970224e+00
-1.01574443e-01 1.09380737e-01 9.77825582e-01 -7.75885820e-01
5.98857462e-01 3.08460742e-01 2.05324925e-02 4.53065373e-02
-4.22115654e-01 -1.21182866e-01 -5.51667154e-01 -2.25903228e-01
4.51107919e-01 -5.83959043e-01 -1.16390431e+00 2.08497286e+00
1.28750339e-01 7.18683004e-02 1.86752811e-01 6.10403657e-01
8.93978179e-01 8.35974514e-01 5.73481083e-01 -2.02606767e-01
1.30941558e+00 -1.55832994e+00 -6.47389412e-01 -5.81291318e-01
1.07412863e+00 -6.82602048e-01 1.80192327e+00 -5.89701235e-02
-1.13262486e+00 -8.01952064e-01 -9.13460374e-01 -4.95786697e-01
-4.09931749e-01 -1.42438948e-01 2.38873214e-01 1.30673528e-01
-1.05115426e+00 6.27283931e-01 -6.79427564e-01 -6.91462100e-01
1.33070663e-01 -4.47011665e-02 6.16316572e-02 -2.36981869e-01
-1.42997813e+00 1.02503228e+00 5.45240164e-01 -6.22901917e-01
-1.13274753e+00 -1.22013223e+00 -7.57104635e-01 5.68096340e-01
4.84017611e-01 -9.49460626e-01 2.20971274e+00 -4.27736551e-01
-1.88870585e+00 6.91934049e-01 -4.08520967e-01 -2.21865758e-01
1.03986144e-01 -5.42181611e-01 -1.07625909e-01 3.55709046e-02
3.53347026e-02 8.21011364e-01 9.19726014e-01 -5.98974884e-01
-6.51139259e-01 -3.83011368e-03 1.01336941e-01 2.19997078e-01
-6.61337554e-01 2.54022926e-01 -3.37089539e-01 -5.34476280e-01
-3.59040648e-01 -4.95040864e-01 -3.60682279e-01 -3.71450156e-01
-1.35984838e-01 -6.83851421e-01 4.28172141e-01 -4.66160476e-01
1.47253072e+00 -1.99765038e+00 2.47684330e-01 -4.72545803e-01
1.93540100e-02 5.37583411e-01 -7.18183517e-01 5.67403972e-01
8.71640593e-02 -7.04314634e-02 -2.33970791e-01 -3.62192780e-01
2.09088195e-02 2.36184344e-01 -3.61759603e-01 -2.06724837e-01
4.21240717e-01 1.34076595e+00 -1.30804193e+00 -2.38559425e-01
-6.28161579e-02 1.62755400e-02 -6.99837565e-01 6.77854061e-01
-5.62834263e-01 2.00024143e-01 -1.84556574e-01 4.40721840e-01
8.47929120e-02 -4.71019030e-01 -1.45045936e-01 3.44755501e-01
1.01554289e-01 4.51253533e-01 -4.47643608e-01 2.14817786e+00
-6.29719913e-01 4.98154879e-01 -1.33142218e-01 -7.99998999e-01
8.95555437e-01 5.51717758e-01 -2.95905113e-01 -8.46850991e-01
6.89684004e-02 9.93877873e-02 6.43596649e-02 -7.60775864e-01
5.56468904e-01 -1.27147257e-01 -1.92906722e-01 8.90772700e-01
7.05171883e-01 -1.64978668e-01 2.24277005e-01 4.09406304e-01
1.38275754e+00 2.19077185e-01 6.36914730e-01 -3.08150381e-01
3.08546126e-01 1.24494083e-01 3.24850202e-01 1.11246109e+00
-4.91404921e-01 5.43564439e-01 2.39347443e-01 -6.02919422e-02
-1.15668559e+00 -1.03890657e+00 3.47370416e-01 2.54424787e+00
-1.99386328e-01 -5.76296628e-01 -9.11428273e-01 -6.69116318e-01
-2.31971577e-01 1.10494924e+00 -7.17462599e-01 -6.20598793e-01
-6.11008108e-01 -5.47492385e-01 4.46652114e-01 9.18827832e-01
1.49865821e-01 -1.12820506e+00 -7.78924048e-01 5.51125526e-01
-1.17784761e-01 -8.98769498e-01 -8.30359399e-01 6.67300165e-01
-9.10779119e-01 -7.61400282e-01 -1.04967093e+00 -1.05130398e+00
6.38978004e-01 6.46041751e-01 1.42388093e+00 1.16163850e-01
-9.67165753e-02 3.79547626e-01 -7.07535326e-01 -3.92998159e-01
-4.31468278e-01 5.51918924e-01 -3.13327909e-02 -5.72204173e-01
8.38125706e-01 -5.34499586e-01 -4.02449995e-01 2.32951328e-01
-7.10702598e-01 2.25841954e-01 6.25426114e-01 1.33826458e+00
2.42052063e-01 -8.41388345e-01 8.79330516e-01 -1.05465198e+00
1.18464625e+00 -5.38200557e-01 -1.76015675e-01 7.58012414e-01
-6.00463808e-01 1.92387670e-01 7.96614408e-01 -8.06293905e-01
-1.48288274e+00 -1.95676774e-01 -1.13125995e-01 -1.77527949e-01
-3.88781399e-01 4.65950727e-01 1.21942572e-01 3.16855550e-01
1.34895205e+00 1.86428383e-01 -3.88300896e-01 -8.36071372e-01
9.04078186e-01 6.21488094e-01 6.62073314e-01 -8.58655155e-01
5.44141829e-01 -1.39795467e-01 -6.87543094e-01 -6.71242774e-01
-1.39065385e+00 -6.62452757e-01 -8.15521598e-01 2.14172855e-01
9.35066283e-01 -7.89755404e-01 -3.35634388e-02 3.18174630e-01
-1.29236066e+00 -1.12149358e+00 -4.51510161e-01 1.79251447e-01
-6.61265254e-01 5.36044575e-02 -8.92476559e-01 -4.25181627e-01
-6.35714352e-01 -7.37201571e-01 8.62636149e-01 4.58282739e-01
-5.40670455e-01 -1.07602942e+00 3.30753833e-01 -1.39240804e-03
8.11732411e-01 -6.79605365e-01 1.33666432e+00 -1.12924945e+00
6.07260466e-02 2.65195310e-01 -1.92719981e-01 1.57844111e-01
-2.88414443e-03 -3.19675863e-01 -1.22653198e+00 -2.02189967e-01
1.69145986e-02 -9.09767091e-01 1.03878534e+00 2.78935313e-01
9.13466573e-01 -1.33811608e-01 -2.28022099e-01 6.19110644e-01
1.12552989e+00 7.92530403e-02 4.15078439e-02 2.49287486e-01
4.66671079e-01 6.97347999e-01 6.71447098e-01 2.44880185e-01
3.36684853e-01 4.08859253e-01 -2.02999383e-01 2.45939046e-01
-3.95832568e-01 -6.24467671e-01 3.93082738e-01 1.07296085e+00
7.25804269e-02 4.33401093e-02 -1.03584588e+00 6.26919806e-01
-1.99890351e+00 -7.67713487e-01 1.91768512e-01 1.78490174e+00
1.29745460e+00 1.69471398e-01 -9.33680609e-02 -4.43026662e-01
6.56098843e-01 -3.62420455e-02 -8.39881539e-01 -7.18116283e-01
1.67472325e-02 5.49665451e-01 6.11112006e-02 6.60680234e-01
-7.43502915e-01 1.44742978e+00 6.69638729e+00 7.63608694e-01
-8.03959250e-01 7.79739916e-01 3.39049816e-01 -3.29984874e-01
-4.42775697e-01 1.18142299e-01 -1.06940031e+00 1.86556607e-01
1.19489086e+00 -5.85594356e-01 3.07885736e-01 9.17948842e-01
1.31593019e-01 6.68847114e-02 -1.34611201e+00 4.76867944e-01
2.37441454e-02 -1.23837674e+00 -6.84151426e-02 -5.94442427e-01
1.12176263e+00 3.72891366e-01 -1.66662470e-01 1.15932083e+00
6.37954831e-01 -8.04388404e-01 4.59997654e-01 1.88896850e-01
7.82118678e-01 -3.04122925e-01 3.27692896e-01 8.61635029e-01
-8.69982541e-01 -4.31793720e-01 -8.64674747e-01 -4.42204922e-01
4.45971459e-01 2.22334545e-02 -8.14855397e-01 4.51506004e-02
5.29687226e-01 1.83970526e-01 -5.54024100e-01 8.78734767e-01
-8.41586709e-01 8.77094150e-01 3.93611938e-02 -3.24238449e-01
4.22282487e-01 2.73264617e-01 2.81331986e-01 1.48901451e+00
2.50332445e-01 6.22546375e-01 7.35520571e-02 7.42008507e-01
-8.59771669e-02 1.52427077e-01 -4.69831645e-01 -5.22959046e-02
7.30918705e-01 1.06973314e+00 -1.53505638e-01 -8.08991790e-01
-6.59341991e-01 1.04600823e+00 1.01803899e+00 6.57542229e-01
-4.10040319e-01 -6.07363403e-01 5.49742043e-01 -1.36493251e-01
3.30346346e-01 -3.30213070e-01 -3.22185874e-01 -1.28390145e+00
-4.80682701e-01 -6.36666715e-01 5.56130886e-01 -8.16778541e-01
-1.46882844e+00 4.65986192e-01 1.30917519e-01 -6.54935062e-01
-5.65484166e-01 -6.06516957e-01 -1.30142462e+00 1.09335625e+00
-1.54536331e+00 -9.62233841e-01 -3.10708880e-01 5.13776302e-01
1.37912810e+00 -2.92879969e-01 1.11046529e+00 -1.72709376e-01
-5.47214031e-01 8.34234595e-01 2.62354966e-02 -1.26376465e-01
1.09994435e+00 -1.27007663e+00 9.20588672e-01 1.06041765e+00
-1.07499063e-01 7.97749400e-01 6.01660907e-01 -5.85029662e-01
-9.44735467e-01 -9.90782440e-01 1.19863760e+00 -4.44619626e-01
8.55511010e-01 -4.09810215e-01 -1.29853988e+00 9.51670766e-01
6.44320726e-01 -2.64553934e-01 9.67203975e-01 4.09908712e-01
-5.21290779e-01 3.14518660e-01 -5.03269911e-01 8.34231257e-01
1.18819165e+00 -6.54166698e-01 -1.45003951e+00 3.67409736e-01
1.41409016e+00 -2.35221818e-01 -6.31115794e-01 1.21439204e-01
8.46330076e-02 -7.09476352e-01 7.00440645e-01 -1.28260148e+00
5.67429602e-01 4.17733699e-01 -4.81326543e-02 -1.81222951e+00
-7.98649609e-01 -9.96325552e-01 -5.26653230e-01 1.20861316e+00
7.58689880e-01 -3.31613630e-01 3.93334150e-01 7.41931796e-01
-5.83773673e-01 -7.39054322e-01 -6.05107069e-01 -8.04960489e-01
5.93761325e-01 -2.55670965e-01 4.92770344e-01 7.69698441e-01
5.12064576e-01 1.08501875e+00 -1.85881853e-01 -3.42665493e-01
3.38881969e-01 -1.67156950e-01 6.61324680e-01 -9.85700369e-01
-6.66995525e-01 -6.70541763e-01 6.15688443e-01 -1.50994039e+00
1.67119578e-01 -8.78524363e-01 4.25538987e-01 -1.45868111e+00
3.48646492e-01 -2.03928977e-01 -4.32305366e-01 6.94848359e-01
-1.01211500e+00 -2.77942955e-01 1.56362548e-01 1.49415687e-01
-5.21045387e-01 6.26718402e-01 1.34318507e+00 -1.51295900e-01
-5.81958532e-01 2.27321517e-02 -9.44286287e-01 5.71395755e-01
9.19977725e-01 -2.92918354e-01 -7.20217347e-01 -8.97366226e-01
1.21069610e-01 -2.58883029e-01 -2.17375815e-01 -5.80577016e-01
6.86608732e-01 -2.12916851e-01 1.97781831e-01 -2.70522833e-01
2.06925794e-01 -8.46194625e-02 -8.97324204e-01 3.80123973e-01
-9.16626811e-01 5.43516390e-02 2.10626364e-01 5.66216528e-01
6.19954802e-02 -7.63887346e-01 7.72689760e-01 -4.11587119e-01
-1.26348996e+00 1.82359740e-01 -4.22680736e-01 6.42484725e-01
7.72762537e-01 1.57932471e-02 -7.67337978e-01 -3.00809503e-01
-9.89454865e-01 4.97007757e-01 2.04431787e-01 6.11097276e-01
6.05882108e-01 -1.10139143e+00 -7.99611151e-01 1.87738225e-01
6.10260189e-01 -6.36516046e-03 4.25914049e-01 8.14491510e-01
1.27990380e-01 4.59620327e-01 -1.50221989e-01 -3.88086349e-01
-1.04239988e+00 7.35716462e-01 -1.46216556e-01 -2.43271023e-01
-7.20158339e-01 1.45949864e+00 3.51885796e-01 -5.50816834e-01
4.13614482e-01 -3.78373981e-01 -1.88022733e-01 1.30937308e-01
9.42781746e-01 2.94369847e-01 -9.76302028e-02 1.81989521e-01
7.92282522e-02 3.35516244e-01 -4.30081785e-01 -2.87201315e-01
1.15353072e+00 -2.66734034e-01 2.91988641e-01 6.55246556e-01
9.00334656e-01 -5.31591117e-01 -1.51396024e+00 -7.98955977e-01
3.01150918e-01 4.22474071e-02 -1.85565323e-01 -8.86086106e-01
-1.60136312e-01 1.45949948e+00 6.40077442e-02 -1.59412362e-02
9.30389524e-01 1.70253605e-01 1.00098801e+00 8.33339810e-01
1.88705057e-01 -1.48200178e+00 4.99173105e-01 1.13940525e+00
5.29199600e-01 -1.30730605e+00 -5.23697257e-01 5.40476292e-02
-7.26871789e-01 1.13357460e+00 1.20795846e+00 7.98838586e-02
4.86234456e-01 8.56857821e-02 1.83663681e-01 1.52662948e-01
-1.52244556e+00 -3.18293303e-01 2.60267645e-01 6.89014852e-01
5.58390319e-01 -2.78899539e-02 -2.09768459e-01 9.63225365e-01
-1.37160614e-01 4.73181121e-02 3.91164690e-01 1.08844447e+00
-1.09276557e+00 -9.37018514e-01 -9.49363559e-02 4.05398548e-01
-1.52873755e-01 -5.31772673e-01 -3.51145744e-01 4.53804642e-01
-2.51936585e-01 9.64735806e-01 1.94105227e-02 -2.43258834e-01
3.45563859e-01 6.74482167e-01 5.20980835e-01 -1.34741902e+00
-7.93595433e-01 -1.67625129e-01 -1.61529314e-02 -3.92792612e-01
2.62781501e-01 -1.03374220e-01 -1.04927564e+00 1.98788214e-02
-3.18449140e-01 1.34277433e-01 2.23477438e-01 1.20599616e+00
4.49908048e-01 7.30668187e-01 1.51602417e-01 -8.75223935e-01
-1.36754656e+00 -1.55739081e+00 -1.08582631e-01 3.34340513e-01
2.53153980e-01 -4.04500693e-01 -3.18156987e-01 1.11104868e-01] | [10.894769668579102, 8.12389087677002] |
a61b2286-bb05-432d-b2b7-fd13b54ee47c | ongoing-eeg-artifact-correction-using-blind | 2306.16910 | null | https://arxiv.org/abs/2306.16910v1 | https://arxiv.org/pdf/2306.16910v1.pdf | Ongoing EEG artifact correction using blind source separation | Objective: Analysis of the electroencephalogram (EEG) for epileptic spike and seizure detection or brain-computer interfaces can be severely hampered by the presence of artifacts. The aim of this study is to describe and evaluate a fast automatic algorithm for ongoing correction of artifacts in continuous EEG recordings, which can be applied offline and online. Methods: The automatic algorithm for ongoing correction of artifacts is based on fast blind source separation. It uses a sliding window technique with overlapping epochs and features in the spatial, temporal and frequency domain to detect and correct ocular, cardiac, muscle and powerline artifacts. Results: The approach was validated in an independent evaluation study on publicly available continuous EEG data with 2035 marked artifacts. Validation confirmed that 88% of the artifacts could be removed successfully (ocular: 81%, cardiac: 84%, muscle: 98%, powerline: 100%). It outperformed state-of-the-art algorithms both in terms of artifact reduction rates and computation time. Conclusions: Fast ongoing artifact correction successfully removed a good proportion of artifacts, while preserving most of the EEG signals. Significance: The presented algorithm may be useful for ongoing correction of artifacts, e.g., in online systems for epileptic spike and seizure detection or brain-computer interfaces. | ['Nobukazu Nakasato', 'Kazutaka Jin', 'Yosuke Kakisaka', 'Rie Tsuda', 'Rie Sakuraba', 'Takafumi Sato', 'Izumi Itabashi', 'Kanoko Kozawa', 'Suguru Asagi', 'Harald Bornfleth', 'Arndt Ebert', 'Toshiyuki Taura', 'Yano Shumpei', 'Yoshiaki Nakao', 'Nicole Ille'] | 2023-06-29 | null | null | null | null | ['seizure-detection', 'eeg', 'eeg'] | ['medical', 'methodology', 'time-series'] | [ 3.31945837e-01 -4.17574972e-01 7.29906023e-01 -2.00160854e-02
-7.71142662e-01 -6.37526214e-01 5.93610331e-02 4.60063964e-01
-4.75195944e-01 1.08527720e+00 1.31701171e-01 -5.39656021e-02
-5.22900701e-01 -7.28014261e-02 -4.53736246e-01 -6.09170079e-01
-5.41856527e-01 -1.07957549e-01 1.60399422e-01 8.54762048e-02
5.82455814e-01 5.67022026e-01 -1.56279731e+00 3.40759814e-01
1.04842889e+00 8.68706107e-01 1.43898100e-01 4.39499348e-01
1.54192209e-01 4.22084391e-01 -1.01902199e+00 2.53956020e-01
8.34412724e-02 -6.95395947e-01 -3.25379312e-01 -2.21904159e-01
-1.32649049e-01 -2.27166843e-02 1.94926068e-01 1.05584908e+00
9.60831165e-01 -1.99775830e-01 5.18376887e-01 -1.13662660e+00
-3.73979807e-02 3.09900790e-01 -5.41143596e-01 6.07893050e-01
6.40377223e-01 1.27469366e-02 -3.29558030e-02 -1.05848074e+00
3.70446652e-01 5.35343364e-02 9.71721411e-01 2.61648238e-01
-1.34447014e+00 -1.13421369e+00 -4.30002749e-01 4.69153970e-01
-1.68740523e+00 -5.29541492e-01 6.08628035e-01 -7.55661190e-01
1.18017924e+00 5.25686979e-01 1.08186173e+00 8.48618805e-01
7.18681455e-01 6.56567737e-02 1.25951827e+00 -3.98940414e-01
4.30445910e-01 1.35067567e-01 5.22032499e-01 -4.35628630e-02
5.02940178e-01 6.83864579e-02 -9.60405588e-01 -5.32192647e-01
6.45081520e-01 -2.92484581e-01 -8.08131039e-01 1.61752805e-01
-1.22332263e+00 2.45966583e-01 -2.23202407e-01 8.39502275e-01
-8.34297717e-01 -2.93823779e-01 3.15662831e-01 3.81184071e-01
6.01264656e-01 7.88860321e-01 -3.99682641e-01 -7.18891621e-01
-1.52788317e+00 2.57446114e-02 4.78515595e-01 8.14478099e-01
3.09496909e-01 1.87500417e-02 -3.23616415e-01 7.75218308e-01
-1.23385832e-01 4.66366947e-01 8.40800881e-01 -3.08365881e-01
1.31545588e-01 4.08642977e-01 2.24778399e-01 -6.55292869e-01
-8.81265461e-01 -5.14102519e-01 -8.53760481e-01 3.56659472e-01
2.14499503e-01 -3.58192682e-01 -6.81285143e-01 1.30528820e+00
-3.74424160e-01 3.99258405e-01 -3.00046802e-01 7.94577360e-01
4.70464349e-01 1.46787956e-01 -1.93675101e-01 -7.77458727e-01
1.32766986e+00 -1.18141554e-01 -1.38995194e+00 3.41297425e-02
2.53468335e-01 -1.02772856e+00 4.57196265e-01 1.06866455e+00
-1.45607710e+00 -2.65238464e-01 -1.01503181e+00 6.03553891e-01
-1.40471444e-01 2.54928768e-01 2.57215649e-01 7.79068172e-01
-1.14419746e+00 6.59321189e-01 -9.29997981e-01 -2.07874298e-01
3.47149879e-01 6.91835403e-01 -4.84527647e-01 4.51090664e-01
-1.02939963e+00 1.16726744e+00 1.11331731e-01 1.06148282e-02
-3.54700595e-01 -9.55268264e-01 -3.73643190e-01 -8.48408416e-02
-4.03939009e-01 -2.43865013e-01 7.24387586e-01 -1.08703208e+00
-1.23395002e+00 6.61402881e-01 -3.35085988e-01 -5.92328191e-01
5.63610077e-01 -2.98002452e-01 -8.05059016e-01 1.84628546e-01
7.96941891e-02 -2.44621456e-01 1.02709484e+00 -6.21633768e-01
-3.40558261e-01 -6.21515572e-01 -9.33494091e-01 -1.79084152e-01
-1.36367917e-01 4.71934438e-01 1.31738320e-01 -1.02732825e+00
1.69872686e-01 -6.04788780e-01 2.24647775e-01 -4.83493835e-01
6.34962916e-02 2.39067510e-01 2.00699672e-01 -1.27103055e+00
1.52191734e+00 -2.22682643e+00 -9.54995155e-02 5.17293334e-01
2.21753538e-01 2.81357691e-02 2.48133257e-01 4.72028166e-01
-7.16888785e-01 -2.40657419e-01 -6.45644486e-01 -4.23880905e-04
-3.04849982e-01 -6.29328609e-01 -1.73936144e-01 8.65497530e-01
-1.49438336e-01 5.82010627e-01 -6.81918263e-01 1.36548370e-01
4.52298731e-01 7.71476626e-01 -3.96959819e-02 2.78379738e-01
8.86784494e-01 7.43393481e-01 4.04692024e-01 2.91275501e-01
7.00395048e-01 3.61799538e-01 -2.92521298e-01 -1.50576979e-01
-4.83575583e-01 1.82144597e-01 -1.30051231e+00 1.67296875e+00
-1.68927684e-01 1.12012374e+00 -1.58643220e-02 -4.98669237e-01
9.68461573e-01 8.96234393e-01 7.47471988e-01 -7.10591197e-01
3.00524771e-01 6.91316724e-01 1.22692212e-01 -6.61790431e-01
-1.18169457e-01 -7.85073861e-02 5.71165144e-01 3.75435978e-01
7.35978857e-02 -1.36409536e-01 4.01736163e-02 -1.13199987e-01
1.32169211e+00 -6.92233071e-02 5.29025078e-01 -7.09454298e-01
3.87830168e-01 -2.69416541e-01 1.46468550e-01 5.36256075e-01
-1.53761849e-01 7.83615649e-01 7.61276111e-02 -2.20212042e-01
-3.54206890e-01 -9.38040733e-01 -6.07194960e-01 2.08258048e-01
-4.73659039e-02 -2.99521685e-01 -1.13462937e+00 5.97554212e-03
-2.17445552e-01 7.20650077e-01 -5.26006281e-01 -2.60474473e-01
-2.79564083e-01 -8.92130077e-01 2.25262493e-01 3.39420378e-01
8.81500244e-02 -1.20622528e+00 -1.08522952e+00 6.87270701e-01
-2.68244207e-01 -7.86549449e-01 -2.60887295e-01 3.78698796e-01
-9.22170639e-01 -1.22831440e+00 -8.00257027e-01 -4.23879266e-01
7.82679141e-01 -8.80293697e-02 6.53912961e-01 -5.83636798e-02
-6.34032547e-01 1.93680257e-01 -3.55567336e-01 -7.72605300e-01
2.85289679e-02 -7.52796710e-01 3.58866215e-01 1.95165336e-01
6.05394185e-01 -1.00258517e+00 -6.78566217e-01 1.36471704e-01
-7.24242747e-01 -2.33352199e-01 4.72877681e-01 4.35616344e-01
2.88098603e-01 1.16583288e-01 8.01226914e-01 -2.42001906e-01
1.18600714e+00 -3.68890554e-01 -5.91908813e-01 -2.47385412e-01
-5.66033363e-01 -3.99275988e-01 4.80222493e-01 -2.93587506e-01
-3.61859471e-01 -4.05532792e-02 -1.52888328e-01 -1.28239706e-01
-4.29110527e-01 6.74383715e-02 8.86315033e-02 -4.75198805e-01
7.99772978e-01 3.61102670e-01 -2.04554990e-01 -5.85844636e-01
-5.92718422e-01 7.03753591e-01 6.01386368e-01 1.85956240e-01
4.20189291e-01 2.22855553e-01 -1.60439029e-01 -1.01501632e+00
2.27886096e-01 -7.90444613e-01 -6.84505522e-01 -1.75342336e-01
7.92088091e-01 -7.87527442e-01 -2.39369899e-01 7.47216165e-01
-1.24208975e+00 -1.12869866e-01 -9.18608978e-02 1.01268029e+00
-3.12774181e-01 2.99155504e-01 -1.24769121e-01 -9.67913806e-01
-8.79963398e-01 -9.94200468e-01 6.38966084e-01 -1.04328983e-01
-7.57769704e-01 -5.55673718e-01 2.60374218e-01 -3.80610943e-01
7.04334438e-01 3.70946616e-01 4.37791109e-01 -6.56120956e-01
1.66867197e-01 -4.44725454e-01 9.86781940e-02 2.50623018e-01
6.52570784e-01 -3.29576671e-01 -1.08246756e+00 -4.48992699e-01
4.55350280e-01 5.42282164e-01 2.72622019e-01 7.16470838e-01
1.07987046e+00 -8.40849895e-03 -2.08480224e-01 5.90635836e-01
1.33684647e+00 7.12142229e-01 9.62313592e-01 3.62902075e-01
-8.31318721e-02 3.29469144e-01 1.62963822e-01 6.59559250e-01
-3.96988571e-01 5.52213669e-01 9.59721655e-02 -9.07026082e-02
-2.64631733e-02 4.98652637e-01 2.42935300e-01 6.34677708e-01
-4.14983958e-01 2.47003391e-01 -9.94550645e-01 8.70292962e-01
-1.39927804e+00 -7.57927477e-01 -7.02343464e-01 2.65643311e+00
5.26847124e-01 1.41006365e-01 1.26682803e-01 7.60403991e-01
7.71991253e-01 -7.21757114e-01 -2.34270766e-01 -2.53894001e-01
-9.50756148e-02 8.21466506e-01 5.11484683e-01 2.30613515e-01
-6.46149457e-01 -1.31804822e-02 6.61098719e+00 2.65065253e-01
-1.04444158e+00 4.48789060e-01 -2.06081182e-01 -4.44264740e-01
3.09251696e-01 -5.09200454e-01 -2.10089043e-01 1.09168327e+00
1.23779655e+00 -3.07917625e-01 6.65548205e-01 1.16943374e-01
6.82832301e-01 -3.59061569e-01 -9.51950490e-01 1.63054574e+00
2.08733305e-01 -1.15298021e+00 -5.41777015e-01 -1.83974355e-01
4.11799073e-01 5.43990033e-03 -4.79768187e-01 -3.35237652e-01
-8.63303185e-01 -9.01035488e-01 7.92412937e-01 8.35665941e-01
1.06926286e+00 -8.10670495e-01 9.49364185e-01 9.72669274e-02
-9.84297335e-01 4.26547527e-02 6.11956380e-02 -3.22070628e-01
5.37356324e-02 9.50463414e-01 -4.42337632e-01 4.91822928e-01
7.79053390e-01 6.58731639e-01 -6.41077340e-01 1.75042415e+00
-2.69663721e-01 7.13535011e-01 -4.52386975e-01 1.67629704e-01
-3.74016762e-01 -8.95002410e-02 8.15711021e-01 1.31173742e+00
7.41256833e-01 3.36734027e-01 -8.21052909e-01 7.68493295e-01
4.47808415e-01 1.66749015e-01 -4.53783602e-01 3.52639109e-01
4.32995230e-01 9.99041021e-01 -7.59850681e-01 -1.01705246e-01
-3.21040899e-01 1.02070487e+00 -3.03090513e-01 3.47478986e-01
-7.13032663e-01 -1.07517064e+00 3.55979383e-01 2.84789622e-01
-3.08923721e-01 5.54319173e-02 -7.08413303e-01 -1.12029469e+00
5.19941628e-01 -8.47118735e-01 4.48502935e-02 -9.43108976e-01
-8.48669708e-01 9.22662556e-01 -1.04108743e-01 -1.41964328e+00
-8.24830458e-02 -3.93590540e-01 -7.83906221e-01 1.19914925e+00
-1.11782634e+00 -2.27362201e-01 -4.64628249e-01 9.46163416e-01
4.23805743e-01 -2.67156154e-01 1.03888261e+00 4.57146704e-01
-1.76182285e-01 4.04869348e-01 1.32283464e-01 -3.01022679e-01
8.31760943e-01 -1.03889990e+00 8.24524388e-02 1.10870898e+00
-5.79305626e-02 6.50116920e-01 8.94427657e-01 -6.52171731e-01
-9.92293894e-01 -7.92063534e-01 1.07925618e+00 -2.94381857e-01
7.15183020e-01 -3.40739787e-01 -9.84140098e-01 2.11280927e-01
4.35334712e-01 -2.60075986e-01 8.32640529e-01 -3.95199001e-01
4.66037989e-01 -9.36648995e-02 -1.22371113e+00 1.75756857e-01
6.55356526e-01 -3.67704540e-01 -1.04112983e+00 3.70342106e-01
-2.44226113e-01 -2.34093890e-01 -9.02388394e-01 2.78286099e-01
7.47636855e-01 -1.06367087e+00 5.33802330e-01 5.63409142e-02
-1.12966537e-01 -1.57530516e-01 5.77873707e-01 -1.56513000e+00
-5.10461509e-01 -1.27356410e+00 3.11063994e-02 9.10714030e-01
2.17369750e-01 -9.76498425e-01 1.80299357e-01 5.54327428e-01
-3.38035196e-01 -3.11762959e-01 -1.21741807e+00 -6.60839379e-01
-3.80537361e-01 -6.80786788e-01 5.68041563e-01 6.83186173e-01
5.57600856e-01 -1.95252195e-01 -1.51569948e-01 3.00273418e-01
6.02932036e-01 -2.86802709e-01 1.18150236e-02 -1.22409117e+00
3.04827001e-02 -4.69707876e-01 -7.26635396e-01 1.82734713e-01
-2.90487647e-01 -7.27896690e-01 -9.50531363e-02 -1.52270269e+00
1.21792499e-03 6.90668598e-02 -7.75101721e-01 5.76661348e-01
-9.10822898e-02 5.57363629e-01 -2.33234718e-01 1.37834027e-01
2.50978649e-01 -2.64635719e-02 3.35732847e-01 2.02641532e-01
-5.87319970e-01 -8.03581066e-03 -3.52047741e-01 6.94976330e-01
8.58592808e-01 -7.21164167e-01 -1.60782740e-01 -2.07055464e-01
6.63227960e-02 -4.29811440e-02 3.67280453e-01 -1.59022021e+00
3.21064949e-01 3.69175673e-01 6.11088276e-01 -5.73779106e-01
4.71860468e-02 -1.03351748e+00 6.57391429e-01 6.42381132e-01
1.24725692e-01 3.06673229e-01 7.93402493e-01 1.99769393e-01
-1.77951410e-01 -1.01972766e-01 7.07320452e-01 2.17602864e-01
-3.07413787e-01 -2.46034130e-01 -9.63744998e-01 -2.91748643e-01
1.16788864e+00 -6.32666528e-01 -3.34142596e-02 -1.19947970e-01
-9.35753644e-01 -2.65005976e-01 1.90098330e-01 3.26632112e-01
9.25957680e-01 -1.04185522e+00 -7.91986465e-01 8.07952166e-01
1.12218037e-01 -7.49157012e-01 2.90479571e-01 1.66380119e+00
-5.89686811e-01 5.56334615e-01 -7.74603784e-01 -3.55929315e-01
-1.47648990e+00 2.52276838e-01 4.22918439e-01 2.63085365e-01
-8.89162600e-01 9.11730111e-01 -2.11197317e-01 7.65736222e-01
4.41784263e-01 -4.79858577e-01 -3.39947790e-01 2.18461469e-01
1.03195763e+00 6.47958219e-01 8.76487255e-01 -3.98097664e-01
-8.41194034e-01 6.91461384e-01 5.07894635e-01 -1.70207515e-01
1.52454746e+00 -1.25008002e-01 -5.21798134e-01 5.34498751e-01
8.59794676e-01 2.25336909e-01 -6.41384959e-01 6.51617587e-01
-1.63765252e-01 -4.57807571e-01 2.47991160e-01 -1.23134506e+00
-7.91328847e-01 8.63588333e-01 1.19119704e+00 2.83237606e-01
1.67625225e+00 -6.92882955e-01 4.28908408e-01 -1.56286761e-01
9.05142426e-01 -9.92696762e-01 -6.29112661e-01 -6.58026338e-02
1.15952444e+00 -5.16776979e-01 1.05237365e-02 4.54174094e-02
-2.29048237e-01 1.31985438e+00 -2.59360760e-01 -3.48189533e-01
7.81938016e-01 5.36374032e-01 -1.75149545e-01 -2.31141463e-01
-3.69381160e-01 2.25630507e-01 4.68634874e-01 7.04477489e-01
5.46103597e-01 1.74090490e-02 -1.07378721e+00 1.01679099e+00
-1.58417106e-01 4.32281286e-01 6.60347700e-01 9.49786663e-01
-1.67916194e-01 -6.36734307e-01 -8.26394618e-01 8.01221788e-01
-6.42187595e-01 -3.91918689e-01 -3.40018123e-01 5.49086869e-01
3.50334853e-01 1.31226528e+00 -1.12459563e-01 -9.21105817e-02
5.37158072e-01 3.69670331e-01 5.92196703e-01 -3.49868596e-01
-1.20475996e+00 3.24718177e-01 -5.04994020e-02 -7.30772853e-01
-3.26454133e-01 -6.63723290e-01 -1.17435050e+00 3.94214272e-01
-4.93980557e-01 4.36454862e-01 9.63462591e-01 8.50921631e-01
6.13345683e-01 9.46262538e-01 2.78245449e-01 -7.58316576e-01
1.70723312e-02 -1.33601546e+00 -1.04837561e+00 2.75692850e-01
6.57938838e-01 -7.81178832e-01 -7.78315961e-01 3.61517429e-01] | [13.23723316192627, 3.344759225845337] |
db0dcbfa-c2ad-49f3-96e5-876c80fc524c | multi-label-zero-shot-human-action | 1709.05107 | null | http://arxiv.org/abs/1709.05107v3 | http://arxiv.org/pdf/1709.05107v3.pdf | Multi-Label Zero-Shot Human Action Recognition via Joint Latent Ranking Embedding | Human action recognition refers to automatic recognizing human actions from a
video clip. In reality, there often exist multiple human actions in a video
stream. Such a video stream is often weakly-annotated with a set of relevant
human action labels at a global level rather than assigning each label to a
specific video episode corresponding to a single action, which leads to a
multi-label learning problem. Furthermore, there are many meaningful human
actions in reality but it would be extremely difficult to collect/annotate
video clips regarding all of various human actions, which leads to a zero-shot
learning scenario. To the best of our knowledge, there is no work that has
addressed all the above issues together in human action recognition. In this
paper, we formulate a real-world human action recognition task as a multi-label
zero-shot learning problem and propose a framework to tackle this problem in a
holistic way. Our framework holistically tackles the issue of unknown temporal
boundaries between different actions for multi-label learning and exploits the
side information regarding the semantic relationship between different human
actions for knowledge transfer. Consequently, our framework leads to a joint
latent ranking embedding for multi-label zero-shot human action recognition. A
novel neural architecture of two component models and an alternate learning
algorithm are proposed to carry out the joint latent ranking embedding
learning. Thus, multi-label zero-shot recognition is done by measuring
relatedness scores of action labels to a test video clip in the joint latent
visual and semantic embedding spaces. We evaluate our framework with different
settings, including a novel data split scheme designed especially for
evaluating multi-label zero-shot learning, on two datasets: Breakfast and
Charades. The experimental results demonstrate the effectiveness of our
framework. | ['Ke Chen', 'Qian Wang'] | 2017-09-15 | null | null | null | null | ['multi-label-zero-shot-learning'] | ['computer-vision'] | [ 6.31479502e-01 -2.24338502e-01 -4.71116513e-01 -2.71300137e-01
-8.06324005e-01 -2.25147083e-01 5.53255856e-01 3.84177640e-02
-4.78941172e-01 4.62096751e-01 4.32080120e-01 2.87842542e-01
-1.87735721e-01 -5.28085053e-01 -4.39546734e-01 -8.82866740e-01
1.57759577e-01 1.56282246e-01 3.33975434e-01 1.63410172e-01
7.49206990e-02 1.67748392e-01 -1.78780711e+00 5.33509910e-01
1.90615252e-01 9.00783658e-01 3.83005440e-02 4.54265952e-01
1.11229189e-01 1.33471966e+00 -4.55907732e-01 -1.84785932e-01
2.39963561e-01 -7.12670803e-01 -9.34018135e-01 6.14316821e-01
4.37589496e-01 -4.73209232e-01 -3.08661759e-01 1.05267954e+00
3.46904933e-01 5.53903103e-01 7.41888165e-01 -1.62601089e+00
-3.88259113e-01 2.91010827e-01 -4.89330918e-01 2.63748914e-01
5.66788554e-01 1.53952306e-02 1.17947090e+00 -6.71575010e-01
6.40992761e-01 1.09013605e+00 4.30070043e-01 5.61045468e-01
-8.08052838e-01 -5.89788020e-01 2.45069176e-01 7.14859307e-01
-1.18815494e+00 -2.43311554e-01 8.36490691e-01 -6.90291584e-01
6.18542194e-01 9.67891514e-02 6.98115766e-01 1.43585563e+00
-9.93364975e-02 9.81446207e-01 9.37860310e-01 -3.15836519e-01
3.00814331e-01 -5.45917228e-02 2.20705628e-01 5.21140516e-01
-1.49392650e-01 -1.46417797e-01 -6.96227133e-01 -4.12900709e-02
6.07280850e-01 6.02280378e-01 -2.50679582e-01 -6.72830343e-01
-1.46715117e+00 6.53408825e-01 8.71566236e-02 6.02597296e-01
-4.01471227e-01 1.15565926e-01 8.42756093e-01 2.73944408e-01
3.06914032e-01 9.80560333e-02 -2.18874827e-01 -2.52318263e-01
-6.83276296e-01 3.64523306e-02 6.31492794e-01 8.60750377e-01
8.99346232e-01 -1.19431928e-01 -4.59569752e-01 7.42315769e-01
2.00060606e-01 2.60412991e-01 6.92925274e-01 -1.14603639e+00
4.07708943e-01 5.98033667e-01 2.49246150e-01 -1.32395053e+00
-3.29076111e-01 1.57437682e-01 -7.74357319e-01 1.19927503e-01
4.00242597e-01 1.33377880e-01 -5.92860162e-01 1.80085337e+00
4.86413538e-01 8.30352783e-01 1.24122642e-01 1.10519958e+00
6.12464428e-01 6.51745200e-01 4.17452008e-01 -5.72264791e-01
1.47347462e+00 -1.38419569e+00 -9.00389612e-01 1.07273804e-02
9.14415658e-01 -4.35792804e-01 8.86636138e-01 2.03358397e-01
-4.98042345e-01 -7.52970755e-01 -8.71139765e-01 7.03472048e-02
-3.88512939e-01 2.69999206e-02 3.91655266e-01 2.76697725e-01
-5.37883759e-01 5.07938981e-01 -7.91166306e-01 -6.23204827e-01
2.81167895e-01 9.03061032e-02 -6.59272492e-01 -3.20904195e-01
-1.32047665e+00 7.60447919e-01 6.51169896e-01 -1.14448488e-01
-9.59485650e-01 -2.68697709e-01 -9.47792768e-01 -8.57422128e-02
9.18214679e-01 -4.31672573e-01 1.11747944e+00 -1.36642873e+00
-1.39703345e+00 8.51404071e-01 -1.41104862e-01 -2.79854298e-01
2.72563756e-01 -2.39689037e-01 -5.96584618e-01 4.46786731e-01
3.26878965e-01 3.58226418e-01 8.72588038e-01 -1.03811407e+00
-8.60571980e-01 -3.37397009e-01 3.96754742e-01 3.99388373e-01
-6.20950520e-01 1.92905635e-01 -3.03066254e-01 -7.21529722e-01
-1.73745021e-01 -9.07945454e-01 4.20242362e-02 -1.67266965e-01
1.30193397e-01 -3.29587102e-01 7.99684584e-01 -4.97792810e-01
1.27725732e+00 -2.24699855e+00 3.77201706e-01 -1.81251004e-01
1.12614840e-01 2.72679538e-01 -2.71899641e-01 4.17569518e-01
-1.20313361e-01 -2.56113052e-01 -1.88467100e-01 -1.80936188e-01
6.99428171e-02 2.72074282e-01 -2.05089137e-01 5.78950644e-01
-7.72787035e-02 7.63181627e-01 -1.36843073e+00 -7.83646405e-01
3.93277079e-01 4.83499438e-01 -1.69677824e-01 4.06519115e-01
1.33128121e-01 5.04472494e-01 -5.65340877e-01 7.48574972e-01
1.28691927e-01 -3.70587826e-01 2.52667367e-01 -2.78627247e-01
7.84564838e-02 -4.83634114e-01 -1.31240296e+00 1.91685200e+00
-3.10132653e-01 2.70682812e-01 -3.70309204e-01 -1.36402583e+00
5.87237120e-01 7.66806126e-01 1.06317818e+00 -4.60416019e-01
9.54457745e-02 -2.00952515e-02 -3.47389966e-01 -8.55817497e-01
2.46019244e-01 -4.26601022e-01 -2.41025105e-01 7.20563352e-01
3.84796083e-01 5.59007227e-01 3.54496092e-01 -2.96441521e-02
1.31992924e+00 3.69839162e-01 7.49402940e-01 3.01410824e-01
7.02947736e-01 -2.68874228e-01 7.70047247e-01 6.38053238e-01
-7.73282290e-01 4.50352728e-01 3.80113006e-01 -6.68225944e-01
-7.70447493e-01 -8.15414488e-01 2.15100154e-01 1.41264021e+00
4.43859011e-01 -5.95170081e-01 -5.41585982e-01 -9.44245219e-01
-2.73562491e-01 4.07771945e-01 -7.39748120e-01 -2.68677205e-01
-4.38977808e-01 -3.74387532e-01 3.97508562e-01 4.55488771e-01
4.68948126e-01 -1.45938277e+00 -9.67746258e-01 2.24486321e-01
-4.55046564e-01 -1.20370924e+00 -5.47822237e-01 1.58597957e-02
-5.76660872e-01 -1.33775663e+00 -7.09538877e-01 -7.59331286e-01
3.45857620e-01 5.53399444e-01 6.98030710e-01 -2.42074400e-01
-3.34736228e-01 6.79817379e-01 -8.81532311e-01 1.60860240e-01
-1.80340186e-01 -4.03363079e-01 2.55648404e-01 6.66824937e-01
8.28539968e-01 -4.60693091e-01 -4.35933590e-01 6.20515943e-01
-1.07879150e+00 1.95383094e-02 6.49477005e-01 8.82464707e-01
6.56856537e-01 1.74791440e-01 5.64553618e-01 -6.06012166e-01
1.78920865e-01 -6.97850168e-01 -1.72454044e-02 6.36949480e-01
-1.12805478e-01 -2.49227151e-01 6.47733867e-01 -6.39863491e-01
-8.89528871e-01 3.30591828e-01 2.90822357e-01 -8.08911622e-01
-6.57126784e-01 2.99040496e-01 -3.63250077e-01 2.27119088e-01
3.00447971e-01 2.27579266e-01 -2.46019006e-01 -2.43949100e-01
5.26629210e-01 7.42840171e-01 3.24797153e-01 -2.96922833e-01
4.98171538e-01 5.09962261e-01 -7.14195007e-03 -8.41082335e-01
-1.22025573e+00 -1.13313568e+00 -9.79423821e-01 -6.72033608e-01
1.38053238e+00 -9.99852777e-01 -6.59024477e-01 5.44508159e-01
-9.66245830e-01 -1.09944664e-01 -3.25710118e-01 7.14438677e-01
-9.75542247e-01 7.61780083e-01 -4.77175057e-01 -6.68826342e-01
4.57193814e-02 -1.09287560e+00 1.19008064e+00 -1.37081370e-02
-3.10280114e-01 -9.32850063e-01 3.03658873e-01 6.42408729e-01
-1.82898358e-01 3.75951946e-01 5.19267082e-01 -6.82292044e-01
-3.22320223e-01 -1.51345059e-01 -1.93005875e-01 3.71723831e-01
4.10989523e-01 -4.88364220e-01 -8.08358192e-01 -3.31662446e-01
1.53491050e-01 -7.76874185e-01 8.20407569e-01 1.50790170e-01
8.44438195e-01 -1.53642043e-01 -2.71066546e-01 1.91907957e-01
1.30298495e+00 2.78744489e-01 4.42904621e-01 2.70201206e-01
1.05541718e+00 6.94987714e-01 1.05833697e+00 7.04863548e-01
1.86443791e-01 9.84035432e-01 1.67272761e-01 2.57896036e-01
9.15201828e-02 -1.57174796e-01 5.93981683e-01 9.66958940e-01
-2.28161514e-01 -2.97156155e-01 -7.29610145e-01 4.02129591e-01
-2.34873128e+00 -1.42356431e+00 2.14218631e-01 2.04799056e+00
5.17469227e-01 -2.29868308e-01 2.16570348e-01 1.59447461e-01
9.00432527e-01 4.78219837e-01 -5.56048691e-01 6.20871186e-02
1.65939644e-01 -2.19214767e-01 1.80839911e-01 1.46782696e-01
-1.67712903e+00 7.90575922e-01 5.03637886e+00 8.47921133e-01
-8.18676353e-01 4.67848688e-01 2.54555315e-01 -9.19829905e-02
2.10600063e-01 -2.49721836e-02 -6.43052161e-01 4.08124089e-01
8.90909612e-01 -1.25207946e-01 1.55747816e-01 7.83847690e-01
1.42326966e-01 -8.34607426e-03 -1.17292798e+00 1.30637801e+00
5.91355920e-01 -7.40309119e-01 1.90577582e-01 -6.20769002e-02
6.40408218e-01 -3.63472402e-01 -2.94249952e-01 4.73733485e-01
2.20054016e-02 -6.66522264e-01 5.64265728e-01 6.64383352e-01
5.92348993e-01 -6.15595281e-01 6.76154613e-01 4.42334384e-01
-1.36320961e+00 -4.04169708e-01 -2.63888448e-01 -1.81185901e-01
4.55965728e-01 1.62951156e-01 -2.58227229e-01 7.04612613e-01
5.47366500e-01 1.38862586e+00 -4.82677013e-01 8.75040531e-01
-1.36035845e-01 2.56515473e-01 1.98744625e-01 2.78695822e-01
3.10850799e-01 -2.11903900e-01 3.54520440e-01 9.48978662e-01
3.43813539e-01 3.83050233e-01 6.49200559e-01 2.68333077e-01
7.88669214e-02 4.03890878e-01 -8.81055355e-01 -2.10636973e-01
4.24959362e-02 1.21465683e+00 -8.28849554e-01 -3.93196315e-01
-9.40792561e-01 1.32637620e+00 3.75324428e-01 3.83327365e-01
-9.56147969e-01 -1.44091114e-01 4.71855551e-01 -2.33244196e-01
2.35431299e-01 9.86684337e-02 4.02826756e-01 -1.47843385e+00
3.01583135e-03 -8.29340339e-01 7.64344811e-01 -5.62970996e-01
-1.24581265e+00 2.97177166e-01 1.02260396e-01 -1.79231513e+00
-2.92076558e-01 -4.28178221e-01 -2.09787950e-01 1.98846951e-01
-1.23412037e+00 -1.23724055e+00 -3.77355218e-01 7.21792102e-01
7.93877125e-01 -2.32681140e-01 1.02793658e+00 5.68938136e-01
-5.01451850e-01 4.00656253e-01 4.57472429e-02 1.66023195e-01
9.50420320e-01 -1.01308119e+00 -3.24514627e-01 7.95717955e-01
4.56038922e-01 2.21933991e-01 3.54620337e-01 -6.23626828e-01
-1.08598447e+00 -1.18566525e+00 7.75088966e-01 -5.18352866e-01
8.04841101e-01 -7.30449334e-02 -9.39164162e-01 8.10370386e-01
5.26129752e-02 1.55040458e-01 1.16830063e+00 -2.76273163e-03
-3.03009719e-01 2.51044668e-02 -6.65359199e-01 3.48516285e-01
1.22748375e+00 -7.37689257e-01 -8.45171332e-01 4.08206701e-01
6.36919081e-01 2.28471324e-01 -9.98280108e-01 5.25516927e-01
6.44567966e-01 -8.76033664e-01 9.50693429e-01 -7.60279298e-01
4.94430691e-01 -4.97602254e-01 -3.61627966e-01 -1.08606946e+00
-3.91557753e-01 -4.29280475e-02 -3.21456969e-01 1.09699190e+00
-3.09090525e-01 -1.29850611e-01 6.03236377e-01 3.21968406e-01
-1.11526465e-02 -5.52584589e-01 -1.02171481e+00 -7.80996740e-01
-4.20711219e-01 -3.15714389e-01 1.64223164e-01 1.19082320e+00
1.48448646e-01 3.38420659e-01 -9.36102569e-01 7.42139742e-02
6.83125854e-01 2.03981996e-01 6.96930945e-01 -1.09725142e+00
-4.30672526e-01 -2.19254345e-01 -8.84078324e-01 -8.20566595e-01
5.73327541e-01 -7.97289014e-01 1.13831878e-01 -1.38880932e+00
5.06329834e-01 1.56540170e-01 -8.43956113e-01 5.63572407e-01
-1.95710421e-01 4.36179608e-01 2.64181137e-01 4.76156652e-01
-1.37274802e+00 6.66099370e-01 9.96438920e-01 -3.17475975e-01
2.95774769e-02 -1.61456704e-01 -1.83080077e-01 8.29486191e-01
4.78304952e-01 -4.70631272e-01 -6.51250243e-01 -1.36987492e-01
-8.57481211e-02 3.41060609e-01 4.88635689e-01 -1.27675736e+00
1.64480239e-01 -4.42692637e-01 -4.56966013e-02 -2.31364593e-01
3.74661207e-01 -1.11181545e+00 2.17688903e-01 2.31984422e-01
-4.72743332e-01 -4.90998834e-01 -3.25959593e-01 9.90939796e-01
-5.93391836e-01 -2.61146665e-01 6.65558100e-01 -2.55433738e-01
-1.50079656e+00 4.45651412e-01 -3.56384218e-01 8.89968872e-03
1.59545183e+00 -3.31685752e-01 3.12371831e-03 -2.72257149e-01
-1.05520153e+00 1.14821516e-01 3.31914842e-01 6.87458098e-01
5.69324911e-01 -1.74840140e+00 -4.27963704e-01 1.00323834e-01
7.29914904e-01 -5.56369543e-01 7.21577406e-01 9.12559867e-01
-1.79655179e-02 2.20009685e-01 -5.52233100e-01 -3.49161357e-01
-1.45348585e+00 1.02096319e+00 -4.35220227e-02 -3.95902127e-01
-7.04746127e-01 4.96786088e-01 3.03542048e-01 -2.26583287e-01
2.87482649e-01 9.24034938e-02 -6.77060485e-01 6.25986516e-01
7.86263287e-01 5.51124811e-01 -4.71925408e-01 -1.40994263e+00
-2.95389384e-01 8.38750958e-01 1.70641586e-01 1.35826185e-01
1.03858411e+00 -2.14996561e-01 -5.06786536e-03 1.09732378e+00
1.48256099e+00 -5.85699916e-01 -1.24716413e+00 -4.44732249e-01
6.06126711e-02 -6.11657381e-01 -2.74869025e-01 -3.75245720e-01
-8.78461301e-01 8.04199576e-01 7.11346924e-01 -1.43232301e-01
1.11117220e+00 6.84838276e-03 9.07681346e-01 5.80478549e-01
5.57660401e-01 -1.32907021e+00 5.39346695e-01 2.45920911e-01
4.95075047e-01 -1.54839027e+00 -1.00244135e-01 -1.51445687e-01
-8.50706637e-01 1.07243443e+00 6.52140558e-01 1.81935772e-01
6.05565429e-01 -3.45378786e-01 9.27297622e-02 -2.23537549e-01
-5.03822029e-01 -4.26532030e-01 2.17130199e-01 2.70031959e-01
5.90239055e-02 1.12620510e-01 -3.77588540e-01 4.46382135e-01
5.92379510e-01 3.24598163e-01 3.55288953e-01 1.07351422e+00
-5.31449556e-01 -1.12679994e+00 -1.29186630e-01 2.31660724e-01
-2.68485010e-01 3.64743292e-01 -4.78330366e-02 5.00242591e-01
3.71273667e-01 9.52268362e-01 -8.29265639e-02 -5.88734984e-01
2.93771654e-01 5.21089256e-01 2.94451594e-01 -6.77150190e-01
-1.35850504e-01 -3.11299153e-02 -2.51387715e-01 -8.56399477e-01
-1.18772852e+00 -8.45099390e-01 -1.16820979e+00 1.96419597e-01
-1.00549847e-01 -2.00903025e-02 -3.81637365e-03 1.11524534e+00
4.21534143e-02 4.50749964e-01 6.09322548e-01 -7.83243060e-01
-4.24508542e-01 -8.55403602e-01 -9.54842389e-01 1.17433035e+00
5.13770021e-02 -1.01527619e+00 -4.57023412e-01 4.44790244e-01] | [8.530386924743652, 0.7880268096923828] |
bd6016f9-2316-454f-bc00-1fce6b2409d6 | qu-brats-miccai-brats-2020-challenge-on | 2112.10074 | null | https://arxiv.org/abs/2112.10074v2 | https://arxiv.org/pdf/2112.10074v2.pdf | QU-BraTS: MICCAI BraTS 2020 Challenge on Quantifying Uncertainty in Brain Tumor Segmentation - Analysis of Ranking Scores and Benchmarking Results | Deep learning (DL) models have provided state-of-the-art performance in various medical imaging benchmarking challenges, including the Brain Tumor Segmentation (BraTS) challenges. However, the task of focal pathology multi-compartment segmentation (e.g., tumor and lesion sub-regions) is particularly challenging, and potential errors hinder translating DL models into clinical workflows. Quantifying the reliability of DL model predictions in the form of uncertainties could enable clinical review of the most uncertain regions, thereby building trust and paving the way toward clinical translation. Several uncertainty estimation methods have recently been introduced for DL medical image segmentation tasks. Developing scores to evaluate and compare the performance of uncertainty measures will assist the end-user in making more informed decisions. In this study, we explore and evaluate a score developed during the BraTS 2019 and BraTS 2020 task on uncertainty quantification (QU-BraTS) and designed to assess and rank uncertainty estimates for brain tumor multi-compartment segmentation. This score (1) rewards uncertainty estimates that produce high confidence in correct assertions and those that assign low confidence levels at incorrect assertions, and (2) penalizes uncertainty measures that lead to a higher percentage of under-confident correct assertions. We further benchmark the segmentation uncertainties generated by 14 independent participating teams of QU-BraTS 2020, all of which also participated in the main BraTS segmentation task. Overall, our findings confirm the importance and complementary value that uncertainty estimates provide to segmentation algorithms, highlighting the need for uncertainty quantification in medical image analyses. Finally, in favor of transparency and reproducibility, our evaluation code is made publicly available at: https://github.com/RagMeh11/QU-BraTS. | ['Mikhail Milchenko1', 'Marc-Andre Weber', 'Tommy Lofstedt', 'Ilyess Zemmoura', 'Sarahi Rosas-Gonzalez', 'Lin-min Pei', 'Yuan-han Mo', 'Hadrien Reynaud', 'Pablo Arbelaez', 'Catalina Gomez', 'Katrin Datwyler', 'Tal Arbel', 'Yarin Gal', 'Spyridon Bakas', 'Bjoern Menze', 'Christos Davatzikos', 'Justin Kirby', 'John Freymann', 'Rivka Colen', 'Aikaterini Kotrotsou', 'Daniel Marcus', 'Arash Nazeri', 'Hassan M. Fathallah-Shaykh', 'Roland Wiest', 'Andras Jakab', 'Abhishek Mahajan', 'Sanjay Talbar', 'Nisarg A. Shah', 'Craig Meyer', 'Nicholas Tustison', 'Quan Dou', 'Xue Feng', 'Ali Hatamizadeh', 'Andriy Myronenko', 'Mohammadreza Soltaninejad', 'Mehdi Amian', 'Sergi Valverde', 'Liliana Valencia', 'Arnau Oliver', 'Xavier Llado', 'Mariano Cabezas', 'Jun Ma', 'Guillaume Tochon', 'Elodie Puybareau', 'Joseph Chazalon', 'Khan M. Iftekharuddin', 'Md Monibor Rahman', 'Lasitha Vidyaratne', 'Alexis Huard', 'Nicolas Boutry', 'Jayashree Kalpathy-Cramer', 'Elizabeth R. Gerstner', 'Praveer Singh', 'Mehak Aggarwal', 'Sharut Gupta', 'Nishanth Arun', 'Mishka Gidwani', 'Katharina Hoebel', 'Ken Chang', 'Jay Patel', 'Alan Wang', 'Gonzalo Maso Talou', 'Hugh McHugh', 'Veronica Vilaplana', 'Laura Mora Ballestar', 'Tufve Nyholm', 'Minh H. Vu', 'Clovis Tauber', 'Murat AK', 'Subhashis Banerjee', 'Wenjia Bai', 'Yike Guo', 'Elsa Angelini', 'Shuo Wang', 'Chengliang Dai', 'Laura Daza', 'Joseph A. Maldjian', 'Ananth J. Madhuranthakam', 'Baowei Fei', 'Fang F. Yu', 'Ben Wagner', 'Chandan Ganesh', 'Sahil Nalawade', 'Gowtham Krishnan Murugesan', 'Piotr Radojewski', 'Raphael Meier', 'Michael Rebsamen', 'Richard McKinley', 'Chiharu Sako', 'Ujjwal Baid', 'Angelos Filos', 'Raghav Mehta'] | 2021-12-19 | null | null | null | null | ['brain-tumor-segmentation'] | ['medical'] | [-1.93603233e-01 6.97518826e-01 -3.14518183e-01 -5.76365769e-01
-1.51985955e+00 -5.73310614e-01 4.41871405e-01 7.36707389e-01
-4.21334267e-01 1.03804958e+00 6.13863707e-01 -7.11723566e-01
-3.85168761e-01 -5.18581450e-01 -6.62250578e-01 -5.48020005e-01
1.06816985e-01 8.68913114e-01 2.42453367e-02 5.44783354e-01
-8.18565935e-02 3.88560563e-01 -4.75772917e-01 4.56547678e-01
9.99003232e-01 1.21195006e+00 -1.59271330e-01 4.22695786e-01
-5.10124005e-02 9.81879592e-01 -5.32392919e-01 -6.27549410e-01
4.94755544e-02 -1.82442531e-01 -1.06473601e+00 -3.30927730e-01
1.27772123e-01 -4.26520348e-01 5.34492210e-02 1.10166621e+00
5.02274513e-01 -3.74715596e-01 8.85743976e-01 -1.29855764e+00
-1.60048574e-01 1.16513765e+00 -2.98045576e-01 4.43946064e-01
1.67163461e-02 4.71497029e-01 6.09574735e-01 -5.86389840e-01
7.09093213e-01 6.66647136e-01 1.07008290e+00 4.41012025e-01
-1.05544698e+00 -6.68227255e-01 -1.02407962e-01 2.63763629e-02
-1.26097178e+00 -3.11524332e-01 -5.65657169e-02 -9.12162960e-01
7.98438847e-01 3.44313383e-01 5.48944414e-01 9.11716580e-01
8.60329747e-01 4.75802332e-01 1.33147943e+00 1.39106795e-01
4.80000615e-01 1.18129455e-01 1.43910632e-01 4.98711467e-01
4.54472393e-01 1.25710174e-01 -1.10992275e-01 -3.48342061e-01
4.28596079e-01 -3.39983255e-01 -3.73427361e-01 4.03165855e-02
-1.36326826e+00 6.88862801e-01 6.72328293e-01 2.13713348e-01
-3.05266768e-01 3.54380608e-01 6.42442703e-01 -2.72914976e-01
5.46494365e-01 3.08172166e-01 -3.24821800e-01 -3.63667846e-01
-1.34834886e+00 9.96897519e-02 6.61223769e-01 6.89249277e-01
-1.18357688e-01 -3.51777822e-01 -6.07895732e-01 5.36930680e-01
2.93071866e-01 2.98542440e-01 3.17031264e-01 -1.07076347e+00
1.96390495e-01 1.87208787e-01 5.84636703e-02 -4.34049368e-01
-8.03766608e-01 -5.99198222e-01 -9.08807397e-01 2.29468763e-01
5.28402448e-01 -2.78657287e-01 -1.17903054e+00 1.63790262e+00
1.18159615e-01 -7.28984997e-02 -3.32295224e-02 7.82800436e-01
1.03743184e+00 9.17422995e-02 3.80893767e-01 -1.05906092e-01
1.41683102e+00 -5.91014266e-01 -6.21723831e-01 -1.20160934e-02
8.09374034e-01 -6.43978357e-01 6.47136450e-01 2.42219746e-01
-1.14727414e+00 2.94671971e-02 -7.58468032e-01 2.19847694e-01
-9.03185010e-02 -3.45745347e-02 3.80710125e-01 6.66447461e-01
-1.00214195e+00 6.08748317e-01 -1.20849276e+00 7.69943818e-02
8.68941784e-01 1.99337214e-01 -3.70636910e-01 6.80583566e-02
-1.27867055e+00 1.33933651e+00 3.87661338e-01 2.15664834e-01
-1.16442192e+00 -1.37040079e+00 -7.47002184e-01 -1.80815533e-01
3.07211816e-01 -6.08735442e-01 1.60395586e+00 -3.64019006e-01
-9.84884381e-01 7.63975084e-01 1.68067738e-01 -7.45573640e-01
1.21872842e+00 1.41604364e-01 -2.74870396e-01 1.03953935e-01
3.07023615e-01 8.66838396e-01 2.48138294e-01 -1.01536596e+00
-4.01429206e-01 -2.00469658e-01 -3.52787077e-01 -9.03272256e-02
4.32170630e-01 -1.31129166e-02 -4.91040610e-02 -4.76785421e-01
-3.79060842e-02 -8.79236579e-01 -4.20892924e-01 1.62370607e-01
-5.34954965e-01 5.11705726e-02 1.20596617e-01 -9.74730372e-01
1.02323592e+00 -1.78141332e+00 -1.37722060e-01 3.29287440e-01
6.53152525e-01 -1.01537891e-01 4.76465702e-01 -1.73565641e-01
-1.86277822e-01 5.71474016e-01 -6.43622935e-01 -1.97013661e-01
-2.41223499e-02 8.87267217e-02 1.18412718e-01 4.21867222e-01
3.19180250e-01 9.95686591e-01 -9.09491420e-01 -7.49266267e-01
2.57243127e-01 2.30581924e-01 -2.98798174e-01 -1.45894751e-01
-2.36591756e-01 7.36358404e-01 -2.10034966e-01 7.49833643e-01
6.27841413e-01 -5.02119124e-01 -1.14027718e-02 -3.84362310e-01
-2.99452269e-03 2.35891491e-01 -7.30928898e-01 1.37085736e+00
-3.49983901e-01 4.20516342e-01 -3.42192943e-03 -3.86312693e-01
4.11549658e-01 4.77979839e-01 7.69028664e-01 -2.39822850e-01
3.12392026e-01 4.52208906e-01 3.24179173e-01 -2.15718776e-01
1.28967270e-01 -3.98870796e-01 -1.47964075e-01 2.96838909e-01
-2.20197029e-02 -7.68955588e-01 -1.32886097e-01 4.13472921e-01
1.10448849e+00 -1.70005515e-01 4.43279386e-01 -5.10821819e-01
1.88658595e-01 2.11686909e-01 5.90510905e-01 6.71519935e-01
-8.87806177e-01 7.46675432e-01 1.09122765e+00 -8.24396238e-02
-8.15976501e-01 -1.34869266e+00 -9.33194697e-01 1.40667886e-01
-3.97737831e-01 -2.49068096e-01 -8.01127255e-01 -9.13291216e-01
2.02176914e-01 1.02399862e+00 -9.47994053e-01 -5.10388128e-02
7.10562766e-02 -9.27057981e-01 8.14093411e-01 8.23735833e-01
2.32969299e-01 -6.81679726e-01 -8.12640727e-01 8.39085132e-02
-4.26707268e-01 -1.15646601e+00 -3.45923781e-01 2.20285222e-01
-7.95225441e-01 -1.22400486e+00 -9.51725662e-01 6.30288869e-02
5.59647739e-01 -7.46765077e-01 1.12761033e+00 -3.67934704e-01
-2.76944906e-01 5.37485778e-01 -7.88570046e-02 -7.31659532e-01
-8.01373959e-01 -1.84714064e-01 -9.76673290e-02 -7.94829726e-01
1.43625081e-01 -3.01290434e-02 -6.13357127e-01 3.40784788e-01
-9.37671721e-01 1.85551628e-01 5.25871098e-01 8.23858738e-01
8.46045852e-01 -2.24340484e-01 4.74882603e-01 -9.54892635e-01
8.29366505e-01 -7.11531043e-01 -5.44242799e-01 3.03152323e-01
-1.02327907e+00 -4.74660099e-02 3.96200502e-03 1.18444234e-01
-9.29421723e-01 -4.47699279e-01 -2.75201589e-01 -2.86259681e-01
-1.40681202e-02 9.86763000e-01 5.03899872e-01 -2.83864364e-02
9.04841661e-01 -4.55528140e-01 1.88619986e-01 1.39821738e-01
1.96866959e-01 6.09838843e-01 4.29200202e-01 -6.43944323e-01
6.46131337e-02 3.00393850e-01 4.21695672e-02 -7.24935383e-02
-7.45812833e-01 -7.31566623e-02 -2.82362014e-01 -5.19871950e-01
9.21343684e-01 -8.39209795e-01 -3.47271711e-01 3.55676025e-01
-9.00516450e-01 -5.79430938e-01 -5.57262182e-01 5.88685751e-01
-5.49059987e-01 3.83816212e-01 -6.15374386e-01 -4.69257772e-01
-6.32897139e-01 -1.92387331e+00 9.26219881e-01 1.65579587e-01
-6.52199507e-01 -8.98267031e-01 -1.80777442e-03 3.96500289e-01
6.36813521e-01 5.90339363e-01 9.13932979e-01 -7.59366930e-01
-2.86513180e-01 -1.66555822e-01 -4.79948580e-01 3.68609786e-01
-1.42051056e-02 8.20358992e-02 -8.15147817e-01 4.05759290e-02
-7.78837577e-02 -3.56948316e-01 6.89331651e-01 1.08203351e+00
1.39986432e+00 2.12970033e-01 -3.70000511e-01 3.73972446e-01
1.20402813e+00 5.85968010e-02 4.88894373e-01 1.99538425e-01
2.07439095e-01 4.56839651e-01 4.32969481e-01 4.34508115e-01
4.40467060e-01 2.19822362e-01 5.74986994e-01 1.31901115e-01
-1.80092707e-01 1.36034831e-01 -9.31451991e-02 5.38637042e-01
1.70052454e-01 6.98107779e-02 -1.72642696e+00 5.54104626e-01
-1.66384256e+00 -3.71896029e-01 -1.42036691e-01 2.07461953e+00
1.10820532e+00 5.78244925e-01 -2.48291343e-01 -2.65288323e-01
5.50385952e-01 -8.88424888e-02 -7.15896487e-01 -3.70387584e-01
2.43888095e-01 1.57375231e-01 6.26896620e-01 5.86164832e-01
-8.00410032e-01 5.73822856e-01 6.49349928e+00 8.35272193e-01
-9.80269015e-01 3.65501016e-01 1.30372512e+00 -1.61687016e-01
-5.43964446e-01 -7.78566524e-02 -3.65343064e-01 5.33480883e-01
1.16641140e+00 -2.99864113e-01 -1.04535520e-01 6.84654593e-01
2.43924245e-01 -6.46166444e-01 -1.11714363e+00 5.80710828e-01
-3.72037470e-01 -1.63973176e+00 -3.24029803e-01 2.43501198e-02
8.45705688e-01 4.58685070e-01 1.72026560e-01 1.05560839e-01
5.20619333e-01 -1.37454498e+00 7.84938276e-01 9.35491979e-01
1.10255671e+00 -5.16713738e-01 1.36147225e+00 -9.77585092e-02
-5.58171272e-01 3.82289439e-01 -1.76059023e-01 6.02187514e-01
3.20171267e-01 1.24265862e+00 -1.37326491e+00 7.27051497e-01
8.05516124e-01 3.37762415e-01 -4.46879119e-01 1.24491107e+00
-1.85536116e-01 6.82293236e-01 -3.74509394e-01 3.18095833e-01
3.15230966e-01 2.18478605e-01 4.78731751e-01 1.36725652e+00
2.03992173e-01 1.55157924e-01 -2.79897898e-01 1.34818673e+00
-8.17695633e-02 -1.00521691e-01 -1.01563446e-01 -2.19277278e-01
4.12292033e-01 1.24513519e+00 -8.98877740e-01 -3.93024594e-01
-4.59618121e-02 2.48777807e-01 -3.92345302e-02 -9.78821069e-02
-1.12085676e+00 8.91458020e-02 4.20211673e-01 4.63531166e-02
-4.22562003e-01 1.14365824e-01 -8.93939495e-01 -7.57931471e-01
-2.89684385e-01 -7.88033247e-01 6.26217484e-01 -1.00191844e+00
-1.33725035e+00 8.30275297e-01 3.38341594e-01 -9.98200595e-01
-3.14798981e-01 -6.96838558e-01 -3.50680172e-01 1.07824135e+00
-1.22858179e+00 -9.22637880e-01 -3.74131441e-01 2.33116373e-01
1.20749868e-01 1.49041817e-01 7.70043015e-01 -1.32312402e-01
-4.46351528e-01 5.24389446e-01 -2.29608989e-03 6.90999478e-02
9.03164864e-01 -1.40690923e+00 5.98326810e-02 6.12589240e-01
-5.37928760e-01 4.24481302e-01 5.94041407e-01 -1.05048692e+00
-3.68591279e-01 -1.02942157e+00 4.39210534e-01 -7.86672890e-01
8.02995086e-01 3.67370814e-01 -7.24119425e-01 6.84966087e-01
-1.13107651e-01 1.62908733e-01 8.95619631e-01 -1.31393880e-01
-4.32551652e-02 2.52541870e-01 -1.64406598e+00 3.41030300e-01
4.13664043e-01 -4.56117451e-01 -5.16895413e-01 4.30122405e-01
6.59914315e-01 -9.50751841e-01 -1.54054248e+00 8.24375987e-01
4.73399431e-01 -9.05115962e-01 5.51825345e-01 -5.06805897e-01
8.30170751e-01 -1.13953970e-01 -1.13056585e-01 -1.45439088e+00
-4.86590229e-02 4.79859374e-02 1.19564049e-01 6.49363816e-01
8.35997283e-01 -6.38846636e-01 6.15798831e-01 1.28399682e+00
-5.25559306e-01 -1.03048062e+00 -1.29949403e+00 -3.91006351e-01
5.87511718e-01 -9.08986628e-01 4.78922546e-01 9.01581585e-01
1.63645506e-01 -6.41716719e-01 3.44078153e-01 2.36074731e-01
6.69543028e-01 -4.55301255e-01 -9.04319156e-03 -9.76087570e-01
-4.97888774e-02 -7.96506643e-01 -4.75067854e-01 2.47927874e-01
-1.18627690e-01 -1.07262814e+00 7.11931363e-02 -1.94821322e+00
4.18361962e-01 -6.28839791e-01 -4.26471889e-01 6.46676004e-01
-1.27648577e-01 -2.65729316e-02 6.74445778e-02 2.40205169e-01
-2.39203364e-01 1.31290942e-01 1.13637674e+00 -2.45101213e-01
3.52033764e-01 6.57958090e-02 -7.46332884e-01 6.08583212e-01
8.29715908e-01 -5.39281011e-01 -1.68863192e-01 -2.58480847e-01
1.44094437e-01 3.14289838e-01 5.43898821e-01 -1.14869344e+00
2.68704683e-01 -1.93562761e-01 5.82122207e-01 -6.13398671e-01
-1.09913081e-01 -7.45022595e-01 4.77879882e-01 7.34049559e-01
-3.76349121e-01 1.41280945e-02 7.01690316e-01 2.80880183e-01
-1.26151234e-01 -4.53764528e-01 8.97915483e-01 -2.91906774e-01
-3.20290953e-01 2.73514897e-01 -3.77888978e-01 2.33406261e-01
1.08554053e+00 3.35327163e-02 -4.76682067e-01 -3.39023471e-01
-1.22320735e+00 4.97310668e-01 1.69290766e-01 -6.95210174e-02
4.59677607e-01 -9.86930668e-01 -9.35217738e-01 -3.91586840e-01
2.09032193e-01 3.65441889e-01 7.02171266e-01 1.47841203e+00
-7.99038708e-01 6.41232312e-01 -2.06024349e-01 -9.36524749e-01
-7.75957465e-01 1.13137625e-01 9.07940328e-01 -6.55765116e-01
-4.25759733e-01 1.03204536e+00 -3.98643538e-02 -4.50255871e-01
1.85608044e-01 -9.13330674e-01 2.56480932e-01 -9.14979652e-02
3.85869116e-01 4.89450216e-01 5.34867287e-01 -3.04650933e-01
-5.31982124e-01 -5.48549592e-02 -3.06939125e-01 -4.75213498e-01
1.09845901e+00 1.42454937e-01 -3.11853200e-01 2.84493953e-01
6.49021506e-01 -1.04973659e-01 -1.16000414e+00 -5.91887347e-03
6.32258132e-02 -6.66409284e-02 5.53910255e-01 -1.71894944e+00
-9.16019559e-01 7.19641268e-01 7.69087076e-01 -2.14463264e-01
7.97587335e-01 5.77840656e-02 3.70447457e-01 -2.80366570e-01
3.12199265e-01 -1.04060388e+00 -2.70621419e-01 1.61900252e-01
1.02746499e+00 -1.37753773e+00 1.84109151e-01 -2.45755255e-01
-1.05061471e+00 1.05224144e+00 5.61184645e-01 4.16643411e-01
8.98571134e-01 7.16833889e-01 1.96188480e-01 -2.44488552e-01
-5.48851728e-01 3.08773041e-01 3.81045520e-01 4.24240857e-01
5.92964888e-01 6.92957282e-01 -4.28674012e-01 9.21261787e-01
-1.52778000e-01 3.92860115e-01 5.69051921e-01 6.93471134e-01
-6.41194806e-02 -7.03660011e-01 -3.88122112e-01 9.92357969e-01
-5.37524879e-01 -1.85138732e-01 -8.72835815e-02 7.35881746e-01
1.40227944e-01 6.13119304e-01 -1.39675781e-01 -1.78383768e-01
1.28742725e-01 1.29994690e-01 3.64377618e-01 -4.05409753e-01
-6.27097607e-01 -8.06115717e-02 2.62834251e-01 -6.61227942e-01
-4.68194708e-02 -7.64672101e-01 -1.39763713e+00 -1.80913717e-01
-2.45036915e-01 2.24269539e-01 8.66767883e-01 9.02283072e-01
2.41535440e-01 1.07694197e+00 -2.69488066e-01 -3.23240817e-01
-8.46811175e-01 -1.02026033e+00 -1.54698402e-01 -1.96141258e-01
1.01433866e-01 -7.61162698e-01 -2.34165609e-01 -3.89507145e-01] | [14.429834365844727, -2.0568931102752686] |
99be6800-628a-41f0-b0e8-e3bb176e0331 | an-intrinsic-entropy-model-for-exchange | 2205.01386 | null | https://arxiv.org/abs/2205.01386v1 | https://arxiv.org/pdf/2205.01386v1.pdf | An Intrinsic Entropy Model for Exchange-Traded Securities | This article introduces an intrinsic entropy model that can be used as an indicator to gauge investor interest in a given exchange-traded security, along with the state of the general market corroborated by individual security trade data. Although the syntagma of intrinsic entropy might sound somehow pleonastic, since entropy itself characterizes the fundamentals of a system, we would like to make a clear distinction between entropy models based on the values that a random variable may take and the model that we propose, which employs actual stock exchange trading data. The model we propose for intrinsic entropy does not include any exogenous factor that could influence the level of entropy. The intrinsic entropy signals whether the market is inclined to buy the security or rather to sell it. We further explore the usage of the intrinsic entropy model for algorithmic trading, in order to demonstrate the value of our model in assisting investors in the selection of the intraday stock portfolio, along with timely generated signals to support the buy / sell decision making process. The test results provide empirical evidence that the proposed intrinsic entropy model can be used as an indicator to evaluate the direction and intensity of intraday trading activity of an exchange-traded security. The data used for the test consisted of historical intraday transactions executed on The Bucharest Stock Exchange (BVB). | ['Marcel Ausloos', 'Titus-Felix Furtuna', 'Ion Smeureanu', 'Claudiu Vinte'] | 2022-05-03 | null | null | null | null | ['algorithmic-trading'] | ['time-series'] | [-4.08283085e-01 1.36853188e-01 -3.08377296e-01 -1.47190854e-01
2.63599306e-02 -7.58490860e-01 8.41437995e-01 2.80485749e-01
-5.01084626e-01 6.48391008e-01 7.07568675e-02 -6.79949105e-01
-3.89524072e-01 -1.08613193e+00 -2.79332697e-01 -7.37194419e-01
-1.45432376e-03 3.93241912e-01 2.73386892e-02 -4.29179639e-01
6.06484711e-01 6.29808664e-01 -1.08710194e+00 -1.21100545e-01
1.97447613e-01 1.43003821e+00 -1.63607478e-01 2.25903779e-01
-2.68860944e-02 6.44809604e-01 -6.97400808e-01 -5.02272367e-01
7.22258687e-01 -5.08981705e-01 -3.57154071e-01 -5.36942966e-02
-6.32483661e-01 -2.82896101e-01 2.68841714e-01 1.10111058e+00
5.13481442e-03 -1.91672117e-01 5.46103656e-01 -8.72065187e-01
-1.36464864e-01 9.56614316e-01 -1.95760243e-02 5.48364103e-01
2.02531405e-02 1.59066185e-01 1.52360404e+00 -4.23608065e-01
4.73030925e-01 5.34259975e-01 1.89495459e-01 -1.46306336e-01
-1.00406957e+00 -5.42577803e-01 1.53776035e-02 -2.72425741e-01
-7.67145216e-01 1.00871034e-01 8.43652189e-01 -6.08808160e-01
4.59555387e-01 3.69436324e-01 1.18120992e+00 6.62286878e-01
4.74250972e-01 3.11363608e-01 1.37019026e+00 -2.39863917e-01
4.77574110e-01 6.74257398e-01 1.14286676e-01 -1.28726929e-01
8.29600632e-01 5.87916315e-01 -3.71628165e-01 -3.25237662e-01
7.28197157e-01 -3.25932175e-01 -1.97618425e-01 -1.50798619e-01
-1.01526272e+00 1.02506554e+00 1.35327473e-01 7.28031754e-01
-7.14145064e-01 -4.01389122e-01 2.81671226e-01 7.88856864e-01
5.37523866e-01 7.77420580e-01 -5.26757121e-01 -3.86940390e-01
-9.76386607e-01 3.86844248e-01 1.17518032e+00 7.72590935e-02
2.69200236e-01 1.19711943e-01 1.34307042e-01 8.71773809e-02
5.83399296e-01 4.25249785e-01 6.69932306e-01 -5.84248245e-01
3.58565301e-01 6.08332813e-01 2.04069242e-01 -7.64557719e-01
-2.65105635e-01 -5.32717943e-01 -2.66445279e-01 5.68984866e-01
6.52419746e-01 -3.12803596e-01 7.06577748e-02 1.30599618e+00
3.34659102e-03 -3.66022915e-01 2.11672530e-01 6.21328652e-01
-1.00930043e-01 5.10267258e-01 -5.30228615e-01 -5.43957293e-01
1.19356728e+00 -3.07255715e-01 -6.39414966e-01 4.67595279e-01
2.43643254e-01 -6.35396659e-01 6.33099079e-01 7.58075178e-01
-1.12617493e+00 -1.26573220e-01 -1.20711613e+00 8.54474664e-01
-4.20268476e-01 -4.03591245e-01 3.25616896e-01 8.65543425e-01
-6.33761048e-01 7.86379933e-01 -6.26888275e-01 3.06257367e-01
-2.99307376e-01 2.81279951e-01 8.05601105e-02 1.29937422e+00
-1.37600768e+00 1.10288346e+00 6.98539913e-01 2.49374866e-01
-2.01142326e-01 -2.63773412e-01 -3.69565338e-01 4.37332332e-01
2.14546636e-01 -1.80967022e-02 1.03247213e+00 -9.00917530e-01
-1.53272784e+00 4.78345305e-01 7.27872729e-01 -1.11015451e+00
1.04183173e+00 1.49066180e-01 -4.49520856e-01 4.13052924e-02
-2.94821858e-01 -1.82074308e-01 5.36101937e-01 -6.51031971e-01
-5.78228951e-01 -2.52372384e-01 3.11881443e-03 -6.12093583e-02
-5.15074953e-02 -6.54968694e-02 6.41712189e-01 -1.03478396e+00
2.21463487e-01 -9.33109820e-01 1.42403096e-01 -5.18324137e-01
-1.07823677e-01 1.44110158e-01 6.08924687e-01 -3.92182291e-01
1.28320265e+00 -1.72905409e+00 -5.26554108e-01 8.69534612e-01
-2.45214462e-01 -8.67947266e-02 6.03006542e-01 6.89408064e-01
-3.19803178e-01 2.84650236e-01 -9.20292735e-02 4.67741191e-01
2.84227282e-01 -2.43320033e-01 -7.15406418e-01 3.95629764e-01
3.98691520e-02 6.36917233e-01 -4.42712784e-01 1.69284433e-01
1.94896430e-01 1.45098075e-01 -3.39421540e-01 5.38678579e-02
-2.04466462e-01 4.11169171e-01 -5.73901534e-01 3.26897889e-01
3.91238898e-01 -2.24188372e-01 2.21393451e-01 3.57332602e-02
-5.39108396e-01 6.96129560e-01 -1.42318988e+00 4.81069416e-01
-7.08046705e-02 6.07013106e-01 -3.32109720e-01 -6.73235536e-01
1.31750953e+00 5.27568161e-01 5.72446942e-01 -5.87946236e-01
4.36818928e-01 5.90562284e-01 3.96900028e-01 -1.34291099e-02
6.56204879e-01 -5.36766708e-01 -9.57631767e-02 9.64226544e-01
-3.53230655e-01 -1.60769865e-01 2.82304645e-01 -3.18799794e-01
5.11542976e-01 -1.96446642e-01 7.13540137e-01 -5.33626020e-01
5.73216617e-01 -4.76638168e-01 2.68894345e-01 3.19104671e-01
-1.47056067e-02 1.54840117e-02 1.15583980e+00 -5.31929195e-01
-9.03724194e-01 -7.79184282e-01 -4.91370559e-01 2.57286698e-01
6.12434112e-02 1.94166705e-01 -4.48171109e-01 -4.93267834e-01
1.32809862e-01 8.08413744e-01 -6.16795719e-01 1.39677688e-01
-3.76106128e-02 -1.04712713e+00 9.01576206e-02 -1.37914315e-01
7.18370616e-01 -1.21615756e+00 -1.30995250e+00 3.64557534e-01
2.32241467e-01 -6.23184204e-01 -1.66570514e-01 4.10055131e-01
-9.54196036e-01 -9.75427628e-01 -4.33256805e-01 2.12668166e-01
3.17214906e-01 -4.14954036e-01 8.59164596e-01 1.13642309e-03
4.64286804e-01 1.91323996e-01 -4.24129665e-01 -6.55177891e-01
-8.02641511e-01 2.70829927e-02 -8.64760652e-02 4.52265531e-01
3.33567351e-01 -3.41817975e-01 -5.82142174e-01 5.21116972e-01
-9.87728119e-01 -4.53019232e-01 2.99693018e-01 6.51419401e-01
1.69612542e-01 3.08197439e-01 7.29053497e-01 -8.50760639e-01
9.32708383e-01 -3.99845928e-01 -1.34501314e+00 2.11891860e-01
-1.25911081e+00 4.25312698e-01 6.23263940e-02 -1.27947703e-01
-1.04017031e+00 -5.04988551e-01 -7.83953741e-02 1.91527203e-01
3.31602395e-01 7.23597586e-01 1.63052946e-01 -3.14464490e-03
4.88712415e-02 1.79931760e-01 1.74294800e-01 -4.88989413e-01
-3.52431118e-01 4.52142119e-01 4.80701402e-02 -2.71239966e-01
6.80268645e-01 2.77439862e-01 1.16512842e-01 -4.25184041e-01
-3.63233417e-01 -2.08674908e-01 -4.26051170e-01 -1.94490626e-01
5.41901946e-01 -5.24038792e-01 -1.20370841e+00 4.97652143e-01
-7.74479806e-01 -7.85876811e-02 -6.22014284e-01 7.31756270e-01
-5.56591392e-01 -2.58905757e-02 -6.00339055e-01 -1.32941270e+00
-1.73154220e-01 -1.09582043e+00 3.00112873e-01 7.20332144e-03
-4.20558512e-01 -1.31254125e+00 3.55147362e-01 2.07307592e-01
4.82320637e-01 4.45422500e-01 5.74889123e-01 -1.42355442e+00
-8.19640040e-01 -4.70791250e-01 2.11696431e-01 5.79916775e-01
2.60433882e-01 3.10763747e-01 -7.50206828e-01 -1.27626851e-01
9.21764314e-01 7.86478743e-02 4.63589072e-01 2.73897201e-01
2.04734668e-01 -5.45966148e-01 5.76743662e-01 5.93420751e-02
1.46771884e+00 5.84299684e-01 4.26881522e-01 9.75912809e-01
-4.89137501e-01 7.78845072e-01 8.75050962e-01 1.00417495e+00
-8.42117332e-03 6.02707982e-01 4.65991110e-01 3.61993819e-01
8.64165843e-01 -1.56873345e-01 5.70114613e-01 8.51000845e-01
-1.65691137e-01 5.06707095e-02 -5.41294098e-01 4.02894989e-02
-1.26820695e+00 -1.11702847e+00 1.76598117e-01 2.36420918e+00
5.19480467e-01 9.59201336e-01 4.87387478e-01 5.63708901e-01
4.52204138e-01 3.92950058e-01 -1.57215551e-01 -5.97268343e-01
-1.83415219e-01 9.67428908e-02 7.41327047e-01 5.89872122e-01
-7.32983828e-01 7.00748563e-02 6.16327381e+00 4.11025792e-01
-1.43065178e+00 -3.83447051e-01 8.04456830e-01 6.87839761e-02
-5.43802679e-01 1.25375807e-01 -8.16376984e-01 9.22339737e-01
1.12888825e+00 -6.78319335e-01 1.82071865e-01 6.33356273e-01
4.38697338e-01 -3.51180851e-01 -8.64390790e-01 3.66830826e-01
-4.74224627e-01 -1.37538838e+00 4.89255972e-02 7.24321842e-01
4.85247016e-01 -3.37511629e-01 3.54216367e-01 -1.79022148e-01
-1.20210782e-01 -4.62589622e-01 1.04826975e+00 6.36028230e-01
-1.93871900e-01 -7.21230209e-01 1.20736241e+00 2.68131167e-01
-1.02902102e+00 -1.07168229e-02 8.66580158e-02 -2.40188718e-01
2.28903875e-01 5.98938465e-01 -9.85190094e-01 4.33838665e-01
2.06679970e-01 1.12924799e-01 -1.44944444e-01 6.40096009e-01
-3.35571282e-02 4.83432829e-01 -3.69004190e-01 -3.54259700e-01
3.68949175e-01 -8.29759836e-01 6.13716662e-01 5.79041660e-01
5.86019278e-01 -7.53950402e-02 -5.38108528e-01 1.22857046e+00
1.85405344e-01 2.26281241e-01 -5.05795300e-01 -3.34090233e-01
2.85803556e-01 8.52994680e-01 -1.06184042e+00 -1.84043437e-01
-3.46687824e-01 4.14406657e-01 -6.73494339e-01 1.11590229e-01
-6.72095954e-01 -9.17427316e-02 3.31003934e-01 2.92927891e-01
4.55948532e-01 -2.43852269e-02 -6.99218094e-01 -1.02132678e+00
2.49440864e-01 -8.04158390e-01 2.74978220e-01 -1.78791076e-01
-8.89159679e-01 4.90098566e-01 -1.06676938e-02 -1.40556526e+00
-6.06640816e-01 -6.31169498e-01 -6.86329186e-01 8.82764935e-01
-1.37577879e+00 -1.25935405e-01 3.48288953e-01 2.61981666e-01
7.74199367e-02 -5.55541217e-01 2.78977096e-01 -2.51881093e-01
-9.40898731e-02 2.03001022e-01 2.20286295e-01 2.32452095e-01
3.26859877e-02 -1.35551560e+00 3.57740194e-01 6.16463900e-01
3.84650737e-01 5.69230318e-01 1.00742257e+00 -7.66974330e-01
-5.86624980e-01 -2.78367430e-01 8.86829615e-01 -3.23720425e-01
1.21266234e+00 5.13813049e-02 -6.86429381e-01 4.68664140e-01
1.64649695e-01 -5.19542396e-01 5.83432496e-01 -2.31189549e-01
2.21943781e-02 -2.97495574e-01 -1.16468275e+00 2.54784971e-01
-1.21023171e-01 -4.08868939e-01 -9.37018752e-01 -1.32711485e-01
4.70799416e-01 8.90599415e-02 -1.25689769e+00 2.51005709e-01
7.61084676e-01 -1.60579991e+00 4.48679626e-01 2.61346977e-02
-6.48183301e-02 -2.03150585e-02 1.75295875e-03 -9.67348218e-01
2.77870476e-01 -7.60067225e-01 1.97556525e-01 1.00771129e+00
6.88279450e-01 -1.40747809e+00 8.01242232e-01 1.01634121e+00
7.01461434e-01 -6.39529943e-01 -1.07709026e+00 -9.61187959e-01
-5.46437800e-02 -2.27793321e-01 8.62456441e-01 8.30183327e-01
2.54421473e-01 -3.76021594e-01 -9.94497463e-02 -2.39193618e-01
5.05262196e-01 3.43361616e-01 3.94534439e-01 -1.26200557e+00
-5.91102660e-01 -8.28095675e-01 -3.38279337e-01 -5.04584968e-01
-3.05315435e-01 -4.96117532e-01 -6.19839191e-01 -3.58137488e-01
-3.58606040e-01 -3.36236775e-01 -7.80831814e-01 -2.26618320e-01
4.71206337e-01 -1.20412044e-01 5.44639647e-01 3.96753043e-01
2.80315965e-01 3.60650480e-01 9.06238317e-01 1.12713307e-01
-3.38003397e-01 7.46709704e-01 -5.88471651e-01 5.89381933e-01
7.68332183e-01 -3.38256001e-01 -2.37497553e-01 6.61192775e-01
6.80761874e-01 3.83355737e-01 3.37614238e-01 -6.41485691e-01
9.15580541e-02 -2.08681479e-01 5.03124855e-03 -5.09815693e-01
-1.06844422e-03 -1.10572922e+00 5.95885277e-01 9.07509148e-01
-5.95336914e-01 3.24311346e-01 -2.48062044e-01 1.81123525e-01
-7.33155847e-01 -5.64103365e-01 6.32233381e-01 -1.51687235e-01
-4.92554605e-02 -2.37198398e-02 -3.98955733e-01 -1.53461337e-01
1.18022752e+00 -3.26598644e-01 -2.46439502e-01 -5.86288750e-01
-7.88301110e-01 -9.49530602e-02 5.35987556e-01 2.33643994e-01
1.78343982e-01 -9.98966992e-01 -6.16095960e-01 3.27464908e-01
-3.45638782e-01 -7.60386288e-01 -1.82119593e-01 9.15740609e-01
-8.48745465e-01 7.25703895e-01 -2.99477696e-01 -1.17039524e-01
-7.12845266e-01 3.82049322e-01 4.58221138e-01 -6.39776468e-01
-3.80684197e-01 2.17579812e-01 -9.95194614e-02 2.43157759e-01
5.40117128e-03 -6.01710856e-01 -1.17018156e-01 7.00018346e-01
4.03181136e-01 3.75486463e-01 1.45883933e-01 -3.22717667e-01
-2.34532058e-02 1.70441672e-01 1.90308854e-01 -6.49778485e-01
1.26478004e+00 -1.43053547e-01 -1.55879945e-01 1.06378996e+00
7.76569843e-01 2.80250907e-01 -1.14970720e+00 1.29441366e-01
6.94673896e-01 -4.06577259e-01 -2.02411711e-01 -6.63954914e-01
-1.04345536e+00 4.75872427e-01 3.28384399e-01 1.06739461e+00
8.93763423e-01 -3.73297691e-01 4.14314270e-01 3.46661419e-01
5.53823411e-01 -1.29176271e+00 -3.19714367e-01 6.19847067e-02
8.98688257e-01 -9.15748298e-01 7.18209594e-02 2.01784953e-01
-7.41845310e-01 1.34178543e+00 -2.17192695e-01 -3.82280380e-01
1.15739143e+00 3.03722769e-01 5.52341528e-02 -1.52994394e-01
-8.36453199e-01 1.49789378e-01 9.23988596e-02 -2.90810317e-01
3.32450718e-01 1.92230403e-01 -7.94898570e-01 5.48312783e-01
-6.14601433e-01 -2.33956780e-02 9.12165105e-01 6.51035130e-01
-5.01001179e-01 -1.26545298e+00 -3.72701645e-01 3.81877512e-01
-7.54937708e-01 -4.33422951e-03 -4.90739822e-01 1.22869682e+00
-1.26686633e-01 4.64553833e-01 2.90094107e-01 -1.31866306e-01
2.25202426e-01 1.18884459e-01 -8.89771059e-02 -7.17841610e-02
-1.05341077e+00 2.67321408e-01 7.10900202e-02 -1.86192513e-01
-5.43998599e-01 -1.02525687e+00 -8.00609946e-01 -3.27251971e-01
-4.69268382e-01 6.87537551e-01 7.31028795e-01 8.57312262e-01
-3.08011979e-01 5.01932837e-02 1.04966748e+00 -4.16997969e-01
-1.09181941e+00 -7.73151517e-01 -1.25091255e+00 9.22768340e-02
4.65957493e-01 -4.81805265e-01 -8.15029204e-01 -3.13435227e-01] | [4.619071006774902, 4.120868682861328] |
e62ec370-ea74-4da8-aadb-a715025a2544 | mtrnet-a-generic-scene-text-eraser | 1903.04092 | null | https://arxiv.org/abs/1903.04092v3 | https://arxiv.org/pdf/1903.04092v3.pdf | MTRNet: A Generic Scene Text Eraser | Text removal algorithms have been proposed for uni-lingual scripts with regular shapes and layouts. However, to the best of our knowledge, a generic text removal method which is able to remove all or user-specified text regions regardless of font, script, language or shape is not available. Developing such a generic text eraser for real scenes is a challenging task, since it inherits all the challenges of multi-lingual and curved text detection and inpainting. To fill this gap, we propose a mask-based text removal network (MTRNet). MTRNet is a conditional adversarial generative network (cGAN) with an auxiliary mask. The introduced auxiliary mask not only makes the cGAN a generic text eraser, but also enables stable training and early convergence on a challenging large-scale synthetic dataset, initially proposed for text detection in real scenes. What's more, MTRNet achieves state-of-the-art results on several real-world datasets including ICDAR 2013, ICDAR 2017 MLT, and CTW1500, without being explicitly trained on this data, outperforming previous state-of-the-art methods trained directly on these datasets. | ['Sridha Sridharan', 'Sabesan Sivapalan', 'Rui Zeng', 'Clinton Fookes', 'Simon Denman', 'Osman Tursun'] | 2019-03-11 | null | null | null | null | ['curved-text-detection'] | ['computer-vision'] | [ 5.89915812e-01 -3.33477974e-01 6.07098043e-01 -4.91498560e-02
-6.75889194e-01 -6.95971668e-01 6.56709015e-01 -2.46806860e-01
-2.62284786e-01 3.41533482e-01 -5.56091666e-02 -2.72922546e-01
4.82500196e-01 -5.19428015e-01 -8.90682995e-01 -5.46074033e-01
4.15186971e-01 6.20209634e-01 2.86587805e-01 -3.38014573e-01
-7.22493231e-02 2.90942073e-01 -1.30212402e+00 4.90872979e-01
1.17046368e+00 6.65355861e-01 4.44714069e-01 6.50431335e-01
-2.90756166e-01 5.59363723e-01 -7.37854362e-01 -9.97720659e-01
5.07203341e-01 -2.52413690e-01 -3.90152663e-01 3.12303722e-01
9.63612080e-01 -2.29530692e-01 -5.64292789e-01 8.12079966e-01
7.61548102e-01 -2.41567314e-01 6.58706546e-01 -8.62897396e-01
-6.86016738e-01 6.91866398e-01 -7.21896470e-01 -3.58076662e-01
7.94555396e-02 2.58625507e-01 6.92542195e-01 -1.06996667e+00
8.22312176e-01 1.11905921e+00 1.00507820e+00 6.01871192e-01
-1.22977102e+00 -6.86978817e-01 1.78338334e-01 -1.85701743e-01
-1.37464786e+00 -2.20818758e-01 7.51706898e-01 -4.72050607e-01
6.86801016e-01 3.72726828e-01 3.37291688e-01 1.55436575e+00
9.25353542e-02 1.25221634e+00 9.99308169e-01 -5.89212358e-01
-2.66836341e-02 4.15720455e-02 -4.31982487e-01 4.99608070e-01
3.08976620e-01 -4.06251967e-01 -5.19293785e-01 2.42865458e-01
3.78375590e-01 -1.44829713e-02 -3.31606328e-01 -2.71845907e-01
-1.13897860e+00 7.22774267e-01 -1.54704258e-01 1.08305506e-01
1.04558811e-01 -8.53824168e-02 7.84169972e-01 3.68505299e-01
7.76433647e-01 2.64152646e-01 -3.91974092e-01 1.51950186e-02
-1.40334487e+00 1.57325149e-01 6.40115917e-01 1.23282301e+00
3.69932204e-01 4.47968215e-01 -3.48412633e-01 1.24684632e+00
-1.24723360e-01 1.04005170e+00 4.76496965e-01 -4.83605266e-02
1.06580472e+00 4.24096465e-01 -1.57462135e-01 -8.01127136e-01
-2.17274189e-01 -4.41406578e-01 -1.18772602e+00 1.65859148e-01
3.07137758e-01 -3.59139979e-01 -1.19906080e+00 1.36820245e+00
1.27120689e-01 -3.27878535e-01 -1.67732984e-01 6.37705386e-01
7.22933173e-01 6.82780683e-01 -3.17776442e-01 6.31932840e-02
1.23353338e+00 -1.14762628e+00 -9.44979310e-01 -7.82435656e-01
5.10030925e-01 -1.37279189e+00 1.27463543e+00 8.35174143e-01
-8.25927615e-01 -3.95714432e-01 -1.05686319e+00 -3.99389088e-01
-6.56389594e-01 6.99694395e-01 3.51672918e-01 8.04710567e-01
-7.95781672e-01 3.50634396e-01 -7.27717638e-01 -7.11311340e-01
4.00037915e-01 1.64552368e-02 -2.08673835e-01 -1.67078063e-01
-9.77541387e-01 8.25364828e-01 2.64859408e-01 2.53297895e-01
-6.66733980e-01 -4.87610430e-01 -8.04490626e-01 -1.43920392e-01
6.98928654e-01 -4.38122779e-01 9.34341669e-01 -1.00821257e+00
-1.61942267e+00 9.48706150e-01 -2.42818743e-02 -5.26611865e-01
1.28263819e+00 -6.29397690e-01 -5.10900378e-01 -4.37917262e-02
1.05947606e-01 4.61934000e-01 1.57733345e+00 -1.29890299e+00
-2.19313502e-01 -1.52198896e-01 -6.41807497e-01 7.92393833e-02
-6.25086010e-01 1.18537605e-01 -9.68071163e-01 -1.56022680e+00
-1.22764774e-01 -8.84353340e-01 1.51343167e-01 2.46359080e-01
-8.80923033e-01 2.80127972e-01 1.45828378e+00 -1.13171315e+00
1.22652614e+00 -2.07229471e+00 1.80488661e-01 -1.44776911e-01
-1.10921331e-01 5.45896530e-01 -2.12091029e-01 7.42861152e-01
-1.01919323e-01 1.80944234e-01 -5.62627494e-01 -9.47238326e-01
2.37360150e-01 6.85382485e-02 -6.08119547e-01 5.13327837e-01
1.66205660e-01 8.04106891e-01 -3.20657223e-01 -5.84087729e-01
5.73515892e-01 5.22213995e-01 -5.10018170e-01 -6.13096096e-02
-4.76802737e-01 1.44537523e-01 -1.63695775e-02 7.32090831e-01
9.66327190e-01 1.37884721e-01 1.01104759e-01 -1.35025218e-01
-4.97842804e-02 -2.20285773e-01 -1.18353307e+00 1.73274219e+00
-3.26559335e-01 1.09844720e+00 3.12281221e-01 -6.76163018e-01
8.71099710e-01 9.75500643e-02 2.35181972e-01 -7.38319695e-01
1.48839474e-01 1.72717959e-01 -4.70573872e-01 -4.66503799e-01
8.28250229e-01 1.41142249e-01 -7.35830963e-02 2.96085507e-01
-1.38101548e-01 -4.64020520e-01 4.87758338e-01 3.45707417e-01
1.12502575e+00 3.30877751e-01 -2.88642228e-01 3.05448808e-02
3.37231338e-01 -2.25074470e-01 2.69221544e-01 8.75771523e-01
2.15723559e-01 1.19927323e+00 3.28431934e-01 -3.56982291e-01
-1.29721117e+00 -9.20214176e-01 -2.01897800e-01 9.65095997e-01
-2.48461232e-01 -5.96746027e-01 -9.81718957e-01 -6.90133572e-01
-7.38052204e-02 7.52785683e-01 -6.41943455e-01 1.85368732e-01
-6.67511880e-01 -8.82405162e-01 8.43743443e-01 2.64666289e-01
6.04117095e-01 -1.07270193e+00 -9.80804414e-02 2.40636747e-02
-4.08825964e-01 -1.59577167e+00 -9.18201029e-01 2.23910093e-01
-5.93011320e-01 -7.91922331e-01 -7.54445255e-01 -6.62255347e-01
7.77001441e-01 1.61140859e-01 9.67453420e-01 -1.53814897e-01
-5.21768093e-01 7.19806030e-02 -6.78847432e-01 -5.82273364e-01
-6.50528610e-01 1.75281808e-01 -2.69573301e-01 2.30070040e-01
-1.56512797e-01 -2.03775734e-01 -4.17803794e-01 4.50830370e-01
-1.34793746e+00 4.08157170e-01 6.57777429e-01 8.26213777e-01
4.17944610e-01 3.24121356e-01 4.46666032e-02 -1.12332761e+00
4.25925463e-01 1.14994109e-01 -5.51149487e-01 3.39452833e-01
-3.66810799e-01 -2.39869490e-01 1.19468832e+00 -5.75603962e-01
-1.14622438e+00 1.02870204e-01 -2.74882048e-01 -4.64117765e-01
-7.91489612e-03 1.79822713e-01 -3.85049582e-01 2.17261717e-01
6.52129948e-01 6.70187891e-01 -3.96733046e-01 -7.26968408e-01
3.00229788e-01 7.08319426e-01 6.12709045e-01 -3.90807360e-01
1.28290832e+00 7.34803259e-01 -2.59617150e-01 -1.01350677e+00
-7.68436611e-01 -2.49811247e-01 -8.56605411e-01 1.83166221e-01
8.55729520e-01 -8.62664700e-01 -6.31730035e-02 1.16199911e+00
-1.03861439e+00 -7.33918667e-01 -1.16052054e-01 -1.48865879e-01
-3.42882723e-01 7.81805754e-01 -5.19691408e-01 -5.95397174e-01
-7.27191746e-01 -9.02214229e-01 1.40355659e+00 -3.42062652e-01
3.44127491e-02 -8.12365949e-01 -2.24757493e-01 4.70334947e-01
4.75033849e-01 2.12480515e-01 7.34481812e-01 -4.15994465e-01
-3.46318215e-01 -4.02734458e-01 -1.93614513e-01 4.90759552e-01
1.82839453e-01 2.58706093e-01 -1.01419854e+00 -5.80656469e-01
-4.76846844e-02 -1.62471518e-01 1.09021437e+00 1.71270937e-01
1.34612095e+00 -3.44269276e-01 -1.46872759e-01 9.16970670e-01
1.29840505e+00 -1.55075774e-01 8.68485272e-01 4.28644836e-01
1.18380690e+00 2.64810234e-01 4.89077926e-01 4.65378761e-01
7.80154765e-02 8.00864100e-01 3.50943327e-01 -5.77597201e-01
-3.61014754e-01 -2.49871582e-01 4.95185733e-01 7.49301016e-01
4.74286526e-01 -7.77865589e-01 -8.10738385e-01 4.63920712e-01
-1.70302844e+00 -8.76771271e-01 -4.33279812e-01 2.14698839e+00
1.01019597e+00 2.30100170e-01 -1.79792643e-01 2.01791942e-01
8.32000136e-01 3.51688504e-01 -6.04559362e-01 -3.84852529e-01
-5.76064706e-01 3.43948990e-01 6.70858085e-01 5.05069010e-02
-1.51393032e+00 1.29767978e+00 5.38170767e+00 1.37142503e+00
-1.29807997e+00 1.70901999e-01 3.72188717e-01 -1.37043983e-01
9.51761007e-02 -2.50677049e-01 -7.36892581e-01 6.25545382e-01
3.60561788e-01 3.67855072e-01 6.51228726e-01 6.47772908e-01
3.31553131e-01 3.73739749e-02 -7.75799811e-01 9.73689914e-01
5.23974895e-01 -9.78050113e-01 2.36432821e-01 -3.12121421e-01
8.37873042e-01 9.76729095e-02 1.17421679e-01 3.37441325e-01
2.44509622e-01 -1.02766573e+00 1.07463050e+00 2.32749820e-01
1.13054240e+00 -5.68458974e-01 3.99625212e-01 2.34147251e-01
-1.06876445e+00 1.80421278e-01 -4.04260039e-01 5.08348584e-01
1.15335882e-01 9.31154132e-01 -7.23638952e-01 7.05626965e-01
7.93744922e-01 7.44515598e-01 -1.01882851e+00 8.54954064e-01
-3.55222851e-01 7.85301864e-01 -4.77736980e-01 3.19205403e-01
-3.83262639e-03 -8.77497941e-02 6.03088737e-01 1.75621831e+00
3.46746266e-01 -5.93950868e-01 7.48933554e-02 7.73474574e-01
-3.54934782e-01 4.34842110e-01 -4.88638759e-01 -1.94582775e-01
1.18414782e-01 1.33566952e+00 -9.79887962e-01 -2.11103782e-01
-4.05656785e-01 1.53927553e+00 5.70288394e-04 3.82005870e-01
-9.42099452e-01 -6.50161743e-01 2.02791095e-01 8.86971354e-02
7.75106490e-01 -2.39334077e-01 -5.28387249e-01 -1.49869990e+00
3.82909596e-01 -1.28222430e+00 6.74326643e-02 -9.02782977e-01
-1.36052728e+00 6.41370654e-01 -4.92832422e-01 -1.48942602e+00
3.26979131e-01 -7.06903219e-01 -6.37268543e-01 8.60693395e-01
-1.33137524e+00 -1.74077010e+00 -3.56891781e-01 5.41724086e-01
1.03377330e+00 -3.05982023e-01 6.18638754e-01 4.67418879e-01
-9.35868800e-01 9.10688579e-01 6.38416588e-01 5.25688171e-01
1.28660500e+00 -1.30123568e+00 1.06351495e+00 1.36670995e+00
-1.45066138e-02 2.97323942e-01 9.61110532e-01 -9.64370370e-01
-1.66201723e+00 -1.58625019e+00 3.65599990e-01 -6.07850015e-01
5.97794950e-01 -1.15269732e+00 -8.74974132e-01 7.92721808e-01
4.51287776e-01 -2.16516048e-01 8.59677196e-02 -2.49524519e-01
-4.28350866e-01 -8.60419795e-02 -9.56339419e-01 9.21141684e-01
9.62802470e-01 -3.54669720e-01 -2.89516121e-01 7.18746722e-01
5.18144011e-01 -8.34047914e-01 -4.89292175e-01 3.18123847e-01
3.72258097e-01 -9.80340481e-01 9.00984824e-01 1.18457243e-01
6.81082249e-01 -3.04326653e-01 7.04358146e-02 -1.17474210e+00
3.07033420e-01 -9.16926622e-01 1.47417873e-01 1.52812791e+00
3.08296531e-01 -5.24553478e-01 6.59295976e-01 1.17130220e-01
-3.71685505e-01 -3.09701651e-01 -8.14039469e-01 -7.68987417e-01
2.44383141e-01 -7.55744278e-01 3.19390535e-01 9.91900086e-01
-6.07683241e-01 1.29759789e-01 -1.01860940e+00 -2.62584537e-02
5.21981835e-01 -3.70375300e-03 1.10358226e+00 -8.94204199e-01
-2.48310745e-01 -3.57575625e-01 4.36709151e-02 -9.91425753e-01
-3.19832973e-02 -8.06921840e-01 1.77778080e-01 -1.56381059e+00
1.14476360e-01 -1.92587733e-01 5.85235953e-01 6.70771182e-01
-2.05709249e-01 4.01422262e-01 3.88793916e-01 7.95007646e-02
-4.89518076e-01 7.44741797e-01 1.31649184e+00 -4.52735424e-01
3.29932123e-02 -1.85817882e-01 -4.44963753e-01 7.26748526e-01
6.00191474e-01 -5.20790219e-01 -2.59523571e-01 -6.54790640e-01
3.06191027e-01 -4.09210950e-01 2.94853449e-01 -1.01650512e+00
7.17181712e-02 1.09848753e-01 5.22104084e-01 -1.15639985e+00
2.73030341e-01 -8.26407552e-01 2.06775665e-02 9.89029184e-02
-8.38078111e-02 -7.30875060e-02 6.16651952e-01 3.50675553e-01
4.86117080e-02 -2.14565933e-01 7.31767058e-01 1.77912861e-01
-3.34666610e-01 1.71334445e-01 -5.52348614e-01 3.07858050e-01
6.24615490e-01 -1.75154790e-01 -6.31721079e-01 -2.42133260e-01
-2.65911162e-01 1.47239342e-01 8.58053565e-01 6.52266800e-01
4.63187009e-01 -9.10709739e-01 -9.17308092e-01 3.17734241e-01
5.10786213e-02 2.72552311e-01 3.32417846e-01 7.21583486e-01
-7.54133999e-01 2.67671257e-01 1.28091872e-01 -5.62529862e-01
-1.28033948e+00 7.04501152e-01 2.68750876e-01 -6.16237521e-01
-1.00136995e+00 4.13656741e-01 4.80809867e-01 -6.70967400e-01
2.11169466e-01 -3.38862598e-01 2.30971098e-01 -1.53520638e-02
2.85388082e-01 1.98236451e-01 5.07513523e-01 -4.16451931e-01
1.01088688e-01 7.99928665e-01 -3.35533530e-01 3.14484388e-02
1.23665416e+00 -1.69215709e-01 4.50650193e-02 7.49449208e-02
8.10199499e-01 5.08080840e-01 -1.26217246e+00 -2.67423868e-01
-3.17713529e-01 -4.51229632e-01 -1.24837801e-01 -1.10458219e+00
-1.28131008e+00 9.43382084e-01 6.06773078e-01 7.42164701e-02
1.23730373e+00 -5.65693498e-01 1.04295957e+00 3.95590782e-01
4.44052070e-02 -1.41548073e+00 1.82436317e-01 6.85507417e-01
1.33805382e+00 -1.18112004e+00 1.75689131e-01 -6.07603252e-01
-7.74195135e-01 1.11117709e+00 5.19212484e-01 2.43354887e-01
1.50443837e-01 6.50199175e-01 1.99514762e-01 2.53366865e-02
-5.25156200e-01 -8.36044997e-02 2.05605060e-01 5.30361533e-01
2.53996760e-01 -1.79937750e-01 -1.13324501e-01 3.30461979e-01
-3.87625307e-01 -3.21536869e-01 5.75795472e-01 9.37316418e-01
-5.92231099e-03 -1.18746078e+00 -6.88638687e-01 5.04416823e-01
-5.81035256e-01 -5.77265143e-01 -7.05411553e-01 9.85883534e-01
7.31656998e-02 7.36733735e-01 -2.30738655e-01 -1.93085983e-01
4.37716126e-01 7.86224380e-02 2.39023924e-01 -4.99135941e-01
-6.66830301e-01 4.83265340e-01 -6.32020459e-02 -6.10459559e-02
1.10584624e-01 -8.73476624e-01 -8.70606363e-01 -3.87770921e-01
-4.26055968e-01 -4.91074800e-01 8.21474671e-01 8.65451157e-01
2.20646843e-01 6.06786847e-01 4.20818925e-01 -8.61837566e-01
-3.26654851e-01 -1.19423068e+00 -4.97188270e-01 5.98819494e-01
1.14485420e-01 -2.52238989e-01 -2.71909863e-01 4.77279186e-01] | [11.901041984558105, 2.0926928520202637] |
d46a446c-659c-42fa-b7c8-7a59e3889267 | breaking-down-the-ontology-alignment-task | 1805.12402 | null | http://arxiv.org/abs/1805.12402v1 | http://arxiv.org/pdf/1805.12402v1.pdf | Breaking-down the Ontology Alignment Task with a Lexical Index and Neural Embeddings | Large ontologies still pose serious challenges to state-of-the-art ontology
alignment systems. In the paper we present an approach that combines a lexical
index, a neural embedding model and locality modules to effectively divide an
input ontology matching task into smaller and more tractable matching
(sub)tasks. We have conducted a comprehensive evaluation using the datasets of
the Ontology Alignment Evaluation Initiative. The results are encouraging and
suggest that the proposed methods are adequate in practice and can be
integrated within the workflow of state-of-the-art systems. | ['Valerie Cross', 'Ernesto Jimenez-Ruiz', 'Matthias Samwald', 'Asan Agibetov'] | 2018-05-31 | null | null | null | null | ['ontology-matching'] | ['knowledge-base'] | [ 1.40757084e-01 3.77733022e-01 -3.74935508e-01 -4.60483134e-01
-3.60622048e-01 -1.51860401e-01 6.70919657e-01 6.72472417e-01
-6.62830353e-01 3.22290540e-01 6.22039676e-01 -1.75828904e-01
-6.29116118e-01 -7.33087480e-01 -2.77587384e-01 2.14282960e-01
-1.23122953e-01 1.16759646e+00 3.17962497e-01 -8.23800564e-01
2.78261423e-01 3.80463392e-01 -2.12374544e+00 5.20537436e-01
5.22654593e-01 7.74715185e-01 -1.59979463e-01 1.16618425e-01
-5.46568751e-01 9.43749785e-01 -2.48037487e-01 -6.28042698e-01
3.85502577e-01 8.39905143e-02 -1.47414386e+00 -7.45942414e-01
9.76935506e-01 1.19985804e-01 -4.35924679e-01 1.24817121e+00
6.70784712e-01 8.06846768e-02 3.75521064e-01 -1.50639224e+00
-8.42219651e-01 5.73191166e-01 1.68970376e-01 3.67127150e-01
8.31532180e-01 -2.71440536e-01 1.42960143e+00 -7.27328002e-01
8.63296807e-01 1.24234295e+00 1.01616526e+00 3.75547826e-01
-9.84743416e-01 -6.55117095e-01 -2.15197906e-01 7.19372094e-01
-1.46811998e+00 -7.94694126e-01 4.72766459e-01 -6.85699582e-01
1.89803565e+00 3.61415058e-01 5.73779464e-01 6.06747270e-01
2.48600841e-01 2.79514909e-01 4.92732346e-01 -9.12327588e-01
-6.04005270e-02 2.40102515e-01 3.40090424e-01 7.63635218e-01
4.45957482e-01 1.65663451e-01 -4.84807789e-01 -4.48574632e-01
1.57475710e-01 -2.02339172e-01 2.57522553e-01 -6.58995152e-01
-1.08114469e+00 8.24985087e-01 4.36054528e-01 7.81448662e-01
-4.28030550e-01 -6.39381260e-02 5.35369396e-01 6.91112041e-01
4.79951292e-01 7.71976173e-01 -3.06778669e-01 -6.52177632e-02
-9.73664045e-01 6.18838012e-01 1.18533993e+00 1.24327207e+00
8.25133502e-01 -5.70010483e-01 1.20282695e-01 8.59817207e-01
4.21977937e-01 -2.40739390e-01 6.15966439e-01 -9.74031389e-01
5.96058905e-01 1.16519141e+00 1.13988325e-01 -9.33304429e-01
-5.51366329e-01 6.59726709e-02 -1.09600879e-01 1.03481598e-01
9.92799997e-02 5.32456398e-01 -2.26497799e-01 1.48137569e+00
1.53140917e-01 1.26178801e-01 3.61293644e-01 5.32108963e-01
1.03631592e+00 1.13044165e-01 2.98728734e-01 4.28889424e-01
1.71736968e+00 -1.36284709e+00 -8.24377954e-01 -3.80246252e-01
8.69681895e-01 -1.04840839e+00 9.45009172e-01 -2.28948966e-01
-1.30796695e+00 -7.62865603e-01 -1.35267842e+00 -3.77360821e-01
-1.02625012e+00 -3.19888502e-01 7.71253347e-01 5.53737044e-01
-1.31776571e+00 7.07284153e-01 -4.67203259e-01 -1.18040097e+00
2.81058758e-01 5.41767538e-01 -8.32922518e-01 -5.23061305e-02
-1.54902983e+00 1.64171851e+00 9.32478070e-01 -3.31393212e-01
-2.08887854e-04 -6.97169602e-01 -8.85810733e-01 1.55616477e-01
1.00857161e-01 -1.01027048e+00 1.34498084e+00 -5.07833898e-01
-1.22880638e+00 1.32966292e+00 -1.18350208e-01 -4.94211584e-01
8.68811682e-02 -3.53898972e-01 -7.95525730e-01 -1.49320573e-01
1.13273658e-01 6.62586153e-01 -2.18970664e-02 -6.93593383e-01
-8.82547796e-01 -2.74099827e-01 2.73593098e-01 1.28330693e-01
-7.19704986e-01 6.49029911e-01 -3.36926430e-01 -3.48652929e-01
-1.61578774e-01 -7.38544047e-01 -1.43076062e-01 6.22486183e-03
3.63099754e-01 -5.82804978e-01 3.04330587e-01 -6.76189005e-01
1.79003155e+00 -1.75924706e+00 4.50581342e-01 5.95452599e-02
9.61578265e-02 1.03004418e-01 -4.26553845e-01 9.92020369e-01
-4.37929094e-01 -6.25781417e-02 -5.80345802e-02 -4.32297975e-01
5.78644097e-01 2.89057106e-01 -9.96098444e-02 2.71530002e-01
2.45612469e-02 9.84240890e-01 -9.11047518e-01 -5.71913362e-01
2.85497963e-01 1.90127179e-01 -6.75504506e-01 3.31258595e-01
-4.00862619e-02 -1.08347788e-01 -1.03889287e-01 6.56326413e-01
1.98573351e-01 7.58477896e-02 2.74746865e-01 -4.19039130e-01
-1.77825168e-01 7.33695865e-01 -1.12496853e+00 2.13503313e+00
-5.38797855e-01 4.96489733e-01 -4.51868027e-01 -1.16158426e+00
8.58234167e-01 6.12238050e-01 7.88924038e-01 -8.78552914e-01
-1.06236339e-01 4.92274493e-01 2.81669255e-02 -7.30948091e-01
7.93743193e-01 -7.70716753e-04 -1.07028149e-02 4.67182100e-01
5.09239435e-01 2.52544343e-01 5.41124463e-01 -2.97108710e-01
1.20898044e+00 3.63093436e-01 8.47225070e-01 -5.60912848e-01
8.91680002e-01 2.07459077e-01 1.22372404e-01 5.31336606e-01
-1.85675561e-01 1.07768765e-02 2.18047202e-02 -1.03357351e+00
-1.33924723e+00 -6.48518384e-01 -2.97328472e-01 1.35361886e+00
3.25970873e-02 -8.85156572e-01 -6.08372509e-01 -4.71263677e-01
2.22988844e-01 5.64030826e-01 -3.91438812e-01 -2.50359267e-01
-6.57813668e-01 -3.57640147e-01 9.65257168e-01 5.87305665e-01
8.90802406e-03 -1.28223574e+00 -6.51785672e-01 6.85689390e-01
-2.42258072e-01 -1.42030454e+00 -1.27447501e-01 -6.37808144e-02
-5.85121632e-01 -9.50309873e-01 -7.24194720e-02 -9.79989111e-01
2.07575053e-01 -1.15644470e-01 1.72251737e+00 1.42196015e-01
-3.25377613e-01 1.67512268e-01 -3.97159964e-01 -5.19461513e-01
-5.86113274e-01 6.91242516e-01 1.75401494e-01 -6.00753188e-01
1.30930841e+00 -8.93024445e-01 -1.60450131e-01 3.74150544e-01
-9.59838092e-01 -3.28842580e-01 2.99405247e-01 5.23696959e-01
2.36518607e-01 -3.42983333e-03 3.74986500e-01 -6.42202079e-01
7.89474785e-01 -6.14324272e-01 -6.99996829e-01 4.06638175e-01
-9.05761302e-01 1.95551261e-01 3.06511283e-01 -9.86140072e-02
-4.84588623e-01 -9.19072479e-02 -3.43400151e-01 -2.00518057e-01
-2.05338851e-01 8.56086314e-01 -1.15738235e-01 -4.22391683e-01
7.29977787e-01 -3.01394522e-01 -8.94794613e-02 -7.28433132e-01
3.38486463e-01 1.04912198e+00 5.67961156e-01 -7.05481827e-01
6.81431174e-01 3.74025404e-01 -1.86828971e-01 -4.42859650e-01
-7.07872570e-01 -9.89445508e-01 -8.86584282e-01 1.00085132e-01
7.32638597e-01 -9.50954795e-01 -3.77538294e-01 -6.71388358e-02
-1.28508973e+00 7.20727965e-02 -4.31790024e-01 1.81795955e-01
-6.71796322e-01 3.39217275e-01 -2.85239279e-01 -2.92850345e-01
-5.85994840e-01 -1.00786984e+00 1.07033372e+00 1.33835047e-01
-7.23078370e-01 -1.24054277e+00 6.30749822e-01 6.49602354e-01
8.12100470e-01 -1.75542459e-01 9.63937283e-01 -1.19333553e+00
-2.88121879e-01 -5.62870145e-01 -6.46813512e-02 4.33033183e-02
1.61718264e-01 -3.17193478e-01 -9.25232947e-01 -3.26445729e-01
-5.03681064e-01 -1.75789684e-01 4.44017291e-01 -4.56722438e-01
6.29833698e-01 -2.57268250e-01 -4.05054212e-01 5.40695608e-01
1.57561076e+00 -1.37085199e-01 7.22950220e-01 1.17625415e+00
4.85297948e-01 7.30111480e-01 5.69996715e-01 1.62456363e-01
8.14978838e-01 1.41379142e+00 1.89512849e-01 5.68865624e-04
-1.79626033e-01 -3.39305997e-02 1.74400046e-01 1.17202461e+00
-1.43499166e-01 -9.21796784e-02 -1.40942061e+00 5.81106305e-01
-2.38073063e+00 -1.18206739e+00 3.12384695e-01 1.95219445e+00
7.23987818e-01 -4.49191891e-02 4.10050303e-02 -4.54742694e-03
4.01319742e-01 1.11363485e-01 1.66915193e-01 -6.82802439e-01
-6.98138848e-02 7.10497022e-01 2.90973425e-01 5.94006062e-01
-1.19479144e+00 1.22367394e+00 7.13474607e+00 3.43977809e-01
-5.64016044e-01 3.47824156e-01 -5.51608443e-01 2.57566333e-01
-1.56671017e-01 3.22932631e-01 -8.44041049e-01 2.64094323e-01
1.45128155e+00 -4.20273691e-01 5.26832461e-01 7.17825770e-01
-4.84608322e-01 4.26277101e-01 -1.70371759e+00 7.17887163e-01
8.35824832e-02 -1.48757076e+00 3.30777019e-01 1.80788547e-01
3.87342423e-01 4.16867882e-01 -4.82684404e-01 5.63665211e-01
3.82826895e-01 -1.24973691e+00 5.30854344e-01 7.68638492e-01
4.79442894e-01 -3.77070814e-01 1.35346031e+00 -5.81823923e-02
-1.42591488e+00 -2.91934997e-01 -7.27126956e-01 -3.37051153e-01
1.73714921e-01 -6.73490986e-02 -3.24398547e-01 1.28852844e+00
7.87385881e-01 7.94072509e-01 -7.78570652e-01 1.10339236e+00
1.14596687e-01 -3.17201912e-01 -3.02871943e-01 2.63044924e-01
2.57018209e-01 -2.63574749e-01 3.62387747e-01 1.35473120e+00
1.76012397e-01 -4.57191169e-01 3.02864134e-01 5.48171461e-01
-5.18724918e-02 6.19377434e-01 -7.29998291e-01 -9.71475989e-02
7.86119401e-01 8.99385512e-01 3.57223488e-02 -4.23838377e-01
-9.21701491e-01 7.50150859e-01 7.83185720e-01 -1.03213690e-01
-5.59121192e-01 -4.62738276e-01 1.07663393e+00 -1.88862160e-01
1.31032214e-01 2.54677802e-01 -3.72022651e-02 -1.17992079e+00
3.19650114e-01 -1.22784925e+00 8.64147186e-01 -6.75214291e-01
-1.41908431e+00 6.23230398e-01 2.01689884e-01 -1.19796801e+00
-3.34511667e-01 -5.75453758e-01 -2.43541315e-01 1.00254953e+00
-1.80032229e+00 -1.36425567e+00 -4.79159296e-01 2.14519843e-01
2.37319082e-01 -7.62712955e-01 1.69608641e+00 1.04153311e+00
-3.07206988e-01 7.17548847e-01 -1.74606778e-02 -1.44969821e-02
7.10837662e-01 -1.03729200e+00 7.23908305e-01 5.18621981e-01
5.08417070e-01 8.73503864e-01 5.75102508e-01 -2.71016151e-01
-1.06839907e+00 -7.73022473e-01 1.73218882e+00 -7.12167263e-01
1.20972288e+00 -7.79444352e-02 -9.27298069e-01 7.63558149e-01
7.09243834e-01 -5.29474281e-02 1.02382362e+00 2.35067293e-01
-4.50803220e-01 -1.82186633e-01 -9.56874013e-01 4.41576004e-01
1.38894141e+00 -8.01668227e-01 -1.36833513e+00 2.95915335e-01
5.18187165e-01 -3.54511976e-01 -1.60354638e+00 7.49376833e-01
9.71858263e-01 -7.48773277e-01 1.39864683e+00 -1.30931807e+00
3.10488492e-01 -2.80283809e-01 -4.90057707e-01 -9.35484111e-01
-4.89664555e-01 -6.03533566e-01 -2.03239873e-01 1.19987011e+00
3.59619200e-01 -6.98297679e-01 3.72814119e-01 3.77949566e-01
-3.65959033e-02 -4.34928477e-01 -1.26897418e+00 -8.66412878e-01
2.67212510e-01 -1.58087745e-01 1.09637439e+00 1.32615495e+00
5.81365466e-01 3.06023002e-01 3.32568139e-02 3.25724751e-01
3.38959962e-01 -9.28541198e-02 8.02160144e-01 -1.79390621e+00
1.18192889e-01 -7.90545046e-01 -1.18263686e+00 -3.61654341e-01
6.66944444e-01 -1.19463611e+00 -5.83374858e-01 -1.75278831e+00
1.73724547e-01 -3.83160681e-01 -8.50533247e-01 6.81526661e-01
-1.34914201e-02 3.01596057e-02 3.19739282e-02 4.54642385e-01
-5.14247119e-01 3.60956013e-01 2.82136112e-01 -3.23771238e-01
2.49999285e-01 -6.19836152e-01 -6.62728131e-01 3.40980738e-01
8.18732381e-01 -7.97589660e-01 -2.53452957e-01 -5.61392128e-01
4.84779507e-01 -5.49971819e-01 -1.84824653e-02 -1.33321738e+00
5.90707123e-01 1.99356794e-01 -4.27374214e-01 -7.18689933e-02
8.74530971e-02 -1.16765642e+00 4.50184882e-01 2.37727761e-01
-3.94572705e-01 6.01321518e-01 4.52672422e-01 -1.07592437e-02
-4.89150703e-01 -4.85719353e-01 5.37571549e-01 -1.25494868e-01
-8.83297324e-01 4.36668657e-02 -6.37032464e-02 3.50794308e-02
7.59032786e-01 -3.53141159e-01 -3.31773877e-01 1.03341654e-01
-4.71009105e-01 2.69593149e-01 6.62309587e-01 1.05006540e+00
1.55511320e-01 -1.60043907e+00 -7.55175471e-01 1.78955168e-01
9.02998149e-01 -6.07217312e-01 -3.54185432e-01 6.17356658e-01
-7.64450431e-01 8.27211678e-01 -7.77079046e-01 -2.12671310e-01
-1.17738259e+00 5.87344289e-01 5.27919829e-01 -6.44120932e-01
-6.64409637e-01 4.22128707e-01 -4.92076993e-01 -1.24555993e+00
3.40329379e-01 -2.09357515e-02 -6.69984341e-01 1.47925645e-01
5.14556587e-01 3.42510521e-01 4.56660658e-01 -7.93022275e-01
-5.74524343e-01 6.33391857e-01 -1.15094587e-01 1.54650528e-02
1.46423113e+00 -2.53839716e-02 -5.56053996e-01 3.27087104e-01
1.02141416e+00 -2.65983909e-01 -6.20452240e-02 -5.88554144e-01
7.93176591e-01 -2.26373002e-01 -2.23752838e-02 -5.97050309e-01
-6.15132749e-01 5.73508203e-01 9.46384490e-01 2.13709638e-01
7.48906970e-01 1.10924162e-01 6.32067382e-01 7.30041087e-01
3.39248598e-01 -1.09063792e+00 -3.64007026e-01 7.02653706e-01
6.62900150e-01 -9.31839347e-01 2.22342223e-01 -3.24639201e-01
1.64947417e-02 1.37671018e+00 6.49558961e-01 5.16246222e-02
5.69556415e-01 3.23768467e-01 3.54794204e-01 -5.97705722e-01
-7.48540759e-01 -4.55951124e-01 5.36401629e-01 4.26013976e-01
8.15219820e-01 -4.48040217e-01 -5.49435973e-01 4.04928327e-01
-4.09495145e-01 2.28854820e-01 5.90675548e-02 1.16409624e+00
-3.07190835e-01 -1.71152544e+00 5.10838069e-02 8.69336724e-02
-3.66917133e-01 -4.19756055e-01 -4.34394300e-01 1.00329280e+00
2.80312508e-01 5.39840281e-01 2.07653314e-01 -4.07336533e-01
9.32094157e-01 5.93255103e-01 5.28697371e-01 -6.44976079e-01
-8.26410472e-01 -7.32194126e-01 6.12265170e-01 -1.03783047e+00
-8.37301195e-01 -6.43699884e-01 -8.39754641e-01 -2.68312782e-01
-2.81647563e-01 1.29587412e-01 4.43132222e-01 9.73236084e-01
6.67685747e-01 4.57467407e-01 -1.48893759e-01 -6.58440351e-01
-4.90348011e-01 -1.23692584e+00 8.85294601e-02 6.07218444e-01
-2.09177732e-01 -7.28270769e-01 2.16769706e-02 7.51537755e-02] | [9.181114196777344, 8.171034812927246] |
4fbaf602-923f-4349-b703-d5d34fdb3df6 | effect-of-adaptive-and-fixed-shared-steering | 2106.03364 | null | https://arxiv.org/abs/2106.03364v1 | https://arxiv.org/pdf/2106.03364v1.pdf | Effect of Adaptive and Fixed Shared Steering Control on Distracted Driver Behavior | Driver distraction is a well-known cause for traffic collisions worldwide. Studies have indicated that shared steering control, which actively provides haptic guidance torque on the steering wheel, effectively improves the performance of distracted drivers. Recently, adaptive shared steering control based on the physiological status of the driver has been developed, although its effect on distracted driver behavior remains unclear. To this end, a high-fidelity driving simulator experiment was conducted involving 18 participants performing double lane changes. The experimental conditions comprised two driver states: attentive and distracted. Under each condition, evaluations were performed on three types of haptic guidance: none (manual), fixed authority, and adaptive authority based on feedback from the forearm surface electromyography of the driver. Evaluation results indicated that, for both attentive and distracted drivers, haptic guidance with adaptive authority yielded lower driver workload and reduced lane departure risk than manual driving and fixed authority. Moreover, there was a tendency for distracted drivers to reduce grip strength on the steering wheel to follow the haptic guidance with fixed authority, resulting in a relatively shorter double lane change duration. | ['Kimihiko Nakano', 'Bo Yang', 'Edric John Cruz Nacpil', 'Satoshi Suga', 'Zheng Wang'] | 2021-06-07 | null | null | null | null | ['steering-control'] | ['computer-vision'] | [-3.11388016e-01 2.78811753e-01 -2.18232602e-01 -1.66800439e-01
-1.10516101e-01 -4.45148349e-01 1.72573090e-01 -1.39117450e-01
-8.85456502e-01 5.12191117e-01 2.20775396e-01 -7.33958185e-01
-3.93606663e-01 -1.84373707e-01 -3.22121769e-01 -5.86322308e-01
3.01684082e-01 -4.09751505e-01 2.85883427e-01 -6.14121616e-01
6.71722531e-01 6.08762801e-01 -2.13661194e+00 -2.93926895e-01
9.51110601e-01 5.42752743e-01 7.64562905e-01 4.08644825e-01
2.58547693e-01 2.18905717e-01 -2.33646661e-01 -1.16260476e-01
1.68856815e-01 -4.05094139e-02 -6.93205222e-02 -3.53204817e-01
2.12274641e-01 -5.67256868e-01 -4.53716576e-01 5.33477128e-01
9.37743843e-01 5.59520245e-01 2.49713629e-01 -1.55716097e+00
-5.55091798e-02 -7.05050528e-02 -1.03878178e-01 3.41121137e-01
3.99434656e-01 5.81355333e-01 6.03224672e-02 -4.47124511e-01
4.91840065e-01 1.07145798e+00 2.16545388e-01 6.10370398e-01
-1.21428657e+00 -1.10817063e+00 2.52897471e-01 4.03550595e-01
-1.13745618e+00 -7.82955348e-01 7.48008430e-01 -6.38472736e-01
7.73220420e-01 7.06592649e-02 7.94201255e-01 1.05253983e+00
9.42537248e-01 1.78426787e-01 1.23619819e+00 1.78641707e-01
7.28385299e-02 5.69903016e-01 1.61795691e-01 3.51200581e-01
7.15474188e-01 7.08688080e-01 -5.70365250e-01 -1.80040412e-02
2.37378418e-01 -4.09939319e-01 -5.95174313e-01 -4.80145693e-01
-8.52031112e-01 5.13354182e-01 -6.27528578e-02 -2.35294566e-01
-8.10186625e-01 -1.09886155e-01 7.47127831e-01 2.50496835e-01
2.37582140e-02 5.07594526e-01 -1.47407129e-01 -1.00828731e+00
2.78261691e-01 4.41498846e-01 5.97554445e-01 4.93745536e-01
2.16752261e-01 2.33019844e-01 -3.15931588e-01 5.33788979e-01
3.09741162e-02 9.26522374e-01 2.94734142e-03 -9.81759906e-01
2.52678961e-01 1.83335796e-01 5.08701324e-01 -1.22085476e+00
-7.33526826e-01 -2.82623053e-01 8.67518932e-02 8.41848016e-01
7.42072240e-02 -4.33412910e-01 -4.39416021e-01 1.74892306e+00
1.95010453e-01 -4.29196209e-01 -2.19674319e-01 1.27651060e+00
3.33744884e-01 -1.92814935e-02 4.23399478e-01 -2.73733795e-01
1.26024151e+00 -1.56280026e-01 -1.46946990e+00 -2.36253679e-01
8.00251484e-01 -6.36956990e-01 1.42636728e+00 4.52317536e-01
-8.75183821e-01 -1.02463591e+00 -1.14815176e+00 4.37196910e-01
-1.96869910e-01 -6.92157715e-04 2.75736332e-01 9.35690165e-01
-5.33327579e-01 2.24710554e-01 -7.63440967e-01 -7.58370161e-02
-1.78660601e-01 1.38138711e-01 -3.82901043e-01 1.03128485e-01
-1.45076120e+00 1.65012097e+00 -1.56837404e-01 3.31490338e-01
-2.88018107e-01 -7.16433883e-01 -8.62927973e-01 -3.38404655e-01
6.19429827e-01 -6.97867692e-01 1.05437505e+00 -5.71114272e-02
-1.69995105e+00 5.02807379e-01 -1.86123610e-01 -1.31716445e-01
5.48921943e-01 -7.91611791e-01 -8.02584469e-01 -1.87344477e-02
3.78621966e-01 5.24600625e-01 9.77590442e-01 -1.11755383e+00
-3.55531752e-01 -3.62102240e-01 -2.62153745e-01 5.50089359e-01
4.40362662e-01 -2.53314883e-01 3.07239860e-01 -1.51381508e-04
-2.01529503e-01 -1.36392963e+00 2.00025246e-01 2.98388693e-02
-2.85907611e-02 -5.80969691e-01 8.01046789e-01 -2.70997524e-01
1.20610571e+00 -2.73956943e+00 -2.11201996e-01 2.52378076e-01
4.56625789e-01 2.51702219e-01 9.79897827e-02 2.08750397e-01
8.09634402e-02 -5.74792773e-02 4.81294334e-01 3.54714066e-01
1.37137711e-01 1.56654567e-01 -5.00989892e-02 5.54223120e-01
6.63409755e-02 5.34786820e-01 -8.11080158e-01 -1.73238590e-02
4.71157372e-01 1.71435058e-01 -2.58897036e-01 2.12056175e-01
6.02994144e-01 1.63038224e-01 -1.66289017e-01 2.22689718e-01
9.51934874e-01 7.63710499e-01 -1.64634615e-01 -1.17766596e-01
-9.74951684e-01 4.23206568e-01 -7.70614266e-01 9.69720304e-01
-3.26605141e-01 9.98608708e-01 4.40104872e-01 -3.58584300e-02
1.08036637e+00 2.40532830e-01 1.07805051e-01 -1.32039332e+00
4.07244772e-01 2.61430532e-01 7.86722362e-01 -1.15225506e+00
7.22619772e-01 -9.99260619e-02 -2.62441672e-02 4.71289843e-01
-8.12809348e-01 -2.59881258e-01 -3.12442742e-02 8.88912380e-02
7.24545360e-01 1.03892468e-01 -1.53602570e-01 -3.74365956e-01
9.74091664e-02 -2.81209588e-01 2.87972838e-01 7.72774637e-01
-8.59251916e-01 -3.46770048e-01 3.49788189e-01 8.78480673e-02
-2.99692065e-01 -1.12002051e+00 6.41760752e-02 1.12768900e+00
6.62299275e-01 -2.62550384e-01 -5.69604456e-01 1.65273711e-01
5.87131858e-01 1.29169798e+00 -3.66834015e-01 -9.54452276e-01
-2.75114715e-01 3.97665709e-01 2.52370238e-01 5.03430784e-01
3.77670914e-01 -8.80205691e-01 -1.22606242e+00 4.15417701e-02
-3.55115265e-01 -7.99314976e-01 -7.02045500e-01 -1.06445909e-01
-2.03548267e-01 -9.21703577e-01 -2.63661947e-02 -2.93736368e-01
2.89457738e-01 9.79391754e-01 1.83567271e-01 -5.92868542e-03
1.00191068e-02 6.20863259e-01 1.01513686e-02 -1.03717184e+00
-1.90439299e-01 -2.77494252e-01 5.24730742e-01 -2.15683803e-01
3.85692447e-01 1.46030128e-01 -7.03270018e-01 8.12025666e-01
-3.10781002e-01 1.09180026e-01 7.25606203e-01 3.86786580e-01
2.09029585e-01 -3.60048324e-01 1.06794322e+00 -1.35712400e-01
1.48623705e+00 -1.42486870e-01 -6.01286553e-02 -5.02523482e-01
-7.92370081e-01 -1.54099390e-01 6.97496394e-03 -4.23977792e-01
-1.36924171e+00 -4.29680914e-01 -3.89918350e-02 -9.19012353e-02
-3.94951463e-01 4.61947531e-01 -1.38599366e-01 -1.73340231e-01
6.70856237e-01 -2.42218509e-01 1.15235329e+00 1.24847710e-01
1.94369793e-01 8.96840513e-01 6.74607992e-01 -2.45852947e-01
3.28445375e-01 2.13054821e-01 3.06079350e-03 -8.60884905e-01
-1.82522476e-01 -4.94963199e-01 -4.28522527e-01 -1.11342633e+00
8.11035395e-01 -6.11057103e-01 -1.25101233e+00 3.14533502e-01
-6.85961068e-01 -4.33369011e-01 2.56844878e-01 1.27954781e+00
-6.10346913e-01 5.45572378e-02 1.49215102e-01 -9.44156468e-01
1.50925308e-01 -1.21004283e+00 5.90690076e-01 7.40616918e-02
-7.85086513e-01 -4.59883302e-01 -3.11638534e-01 4.17447448e-01
8.03445935e-01 1.83664799e-01 9.23398972e-01 7.37213111e-03
1.88163240e-02 -2.86947221e-01 6.42492101e-02 7.39324167e-02
5.67458928e-01 -3.33072305e-01 -7.08240986e-01 -3.73258561e-01
-8.72735754e-02 -1.38432488e-01 2.46359725e-02 5.23847163e-01
6.17746711e-01 3.22024435e-01 -5.65240622e-01 -5.36511019e-02
6.24893308e-01 7.94974923e-01 9.44360554e-01 1.87474579e-01
1.32046819e-01 9.70955074e-01 1.36592698e+00 -4.70680185e-02
7.86787122e-02 7.33916581e-01 3.88608038e-01 -2.39268124e-01
-2.38011833e-02 -5.12022013e-03 2.84580559e-01 3.29650462e-01
-2.18479589e-01 2.91261803e-02 -4.20347929e-01 2.63025463e-01
-1.10744798e+00 -8.70290995e-01 -6.08446062e-01 2.57192326e+00
3.09200436e-01 6.11482322e-01 1.62400827e-01 2.37937674e-01
6.35958493e-01 -3.61917973e-01 -8.67013633e-01 -1.02126145e+00
4.05484587e-01 -1.16020627e-01 7.30113924e-01 5.09880424e-01
-2.67093480e-01 3.19358081e-01 6.10545301e+00 1.39148250e-01
-1.34321642e+00 -2.31877938e-01 -3.89430553e-01 -3.92146111e-01
-2.80025125e-01 -3.06049407e-01 -6.15310311e-01 6.14681661e-01
1.24493718e+00 -6.34139895e-01 6.94146147e-03 6.16635680e-01
1.12653983e+00 -8.33259940e-01 -6.34210110e-01 5.82461178e-01
-1.75331473e-01 -5.95334828e-01 -7.51190126e-01 3.92124861e-01
-2.71219581e-01 -3.30138177e-01 1.51546687e-01 4.61553514e-01
-6.38274014e-01 -5.38182735e-01 7.30740190e-01 1.00501251e+00
8.02633464e-01 -8.79793882e-01 8.18392694e-01 4.66314882e-01
-8.32223594e-01 -3.68053466e-02 1.88086838e-01 -5.42178035e-01
6.56433523e-01 2.06974950e-02 -8.69630098e-01 1.02442555e-01
2.44165897e-01 7.46251270e-02 -1.88508913e-01 6.52198851e-01
-5.12981378e-02 3.55452985e-01 4.14523371e-02 -3.87707412e-01
-4.94065955e-02 -1.67622119e-01 9.19171512e-01 5.89938879e-01
-1.32269427e-01 1.36799337e-02 -2.01406449e-01 6.46298587e-01
8.57216537e-01 -1.42837733e-01 -1.15944290e+00 1.10338025e-01
5.72520494e-01 8.21728051e-01 -2.97848940e-01 8.23900104e-02
-2.56930411e-01 2.77343214e-01 -4.09345925e-01 4.85758692e-01
-8.45816076e-01 -1.09474158e+00 1.20362759e+00 5.68725705e-01
-2.65718222e-01 -5.12253821e-01 -3.53102088e-01 1.33612573e-01
1.55434892e-01 -1.87811360e-01 -2.80568361e-01 -1.09844863e+00
-5.41105926e-01 1.12069920e-01 4.03460801e-01 -1.14055765e+00
-1.42071322e-01 -6.55258477e-01 -1.60268456e-01 1.57733572e+00
-1.24996495e+00 1.69405118e-02 -6.12492323e-01 1.27727851e-01
4.99360174e-01 2.40574673e-01 4.81635720e-01 2.10545763e-01
-4.94218111e-01 5.43876708e-01 -6.40408397e-01 -8.51545215e-01
1.37242723e+00 -5.83049059e-01 -1.89314291e-01 3.67298156e-01
-1.55164778e+00 1.20893371e+00 1.01748896e+00 -9.04192805e-01
-1.67222714e+00 -5.83548069e-01 5.89592516e-01 -5.85226621e-03
3.09914738e-01 -8.06727111e-02 -1.07036698e+00 1.41610011e-01
2.42268026e-01 -5.51465690e-01 8.13803136e-01 -4.16048020e-02
3.63937587e-01 8.96122027e-03 -9.59287703e-01 9.76771533e-01
1.14805615e+00 -6.83060527e-01 -5.40484667e-01 -2.86451459e-01
5.49417019e-01 -4.81573820e-01 -6.49365485e-01 3.44784081e-01
1.17808890e+00 -6.18786991e-01 3.16027164e-01 -4.01718736e-01
-2.68328696e-01 -5.36325455e-01 5.26525140e-01 -1.68541157e+00
-4.84372050e-01 -7.05914378e-01 3.97712886e-01 3.56659681e-01
6.24976940e-02 -1.32192898e+00 6.74685389e-02 7.46780455e-01
-8.91231298e-01 -6.48671925e-01 -8.40889156e-01 -6.93882883e-01
-5.24082422e-01 -5.47494054e-01 1.29878610e-01 1.20346069e-01
5.22759318e-01 1.43710971e-01 -1.06210463e-01 1.29790306e-01
-4.24143299e-02 -4.68493342e-01 1.01483154e+00 -1.19052768e+00
5.23558259e-01 -3.51717621e-01 -4.81110781e-01 -8.84484947e-01
2.53756791e-01 -5.43827534e-01 3.55656862e-01 -1.60631454e+00
-5.31055391e-01 -2.60100007e-01 2.09389441e-02 1.68688878e-01
-3.26138794e-01 -3.50145966e-01 -1.83384761e-01 -3.07891130e-01
3.13397378e-01 3.76358151e-01 1.73826742e+00 3.37978452e-01
-7.54035711e-01 3.28165948e-01 -1.08125806e+00 -4.78516519e-02
1.02946675e+00 -1.09739490e-01 -9.96230781e-01 1.71644554e-01
1.90685600e-01 2.60103941e-01 3.56291085e-01 -5.72793126e-01
7.27591738e-02 -3.86795670e-01 -3.92088354e-01 -8.51678789e-01
4.04842317e-01 -5.08048534e-01 3.61666590e-01 8.19313169e-01
-3.41118783e-01 2.16027111e-01 9.24169183e-01 4.55007315e-01
3.31379682e-01 1.93609193e-01 4.91568655e-01 8.05094242e-01
-7.23043323e-01 -4.62362468e-01 -1.38143969e+00 -5.32747447e-01
1.13380432e+00 -6.76562190e-01 -4.20083821e-01 -4.60696816e-01
-7.82770813e-01 2.99624860e-01 1.87403068e-01 8.71620238e-01
8.36785316e-01 -1.05105340e+00 -2.34232068e-01 4.84681010e-01
3.30405384e-01 -8.55789840e-01 8.70790899e-01 1.68793881e+00
9.13348943e-02 9.23137724e-01 -6.72730744e-01 -6.05764627e-01
-1.34902191e+00 3.37530166e-01 2.40974128e-01 1.01560545e+00
-6.57617688e-01 1.96114525e-01 8.26224238e-02 1.25755936e-01
-3.55337299e-02 -2.54487067e-01 -2.40158573e-01 -2.16162086e-01
3.57441485e-01 9.03875411e-01 4.07864898e-01 -5.81005692e-01
-3.06236297e-01 3.35637271e-01 1.63162332e-02 6.77575031e-03
1.14773437e-01 -4.55038399e-01 5.21348953e-01 6.83856368e-01
5.82014740e-01 3.65118742e-01 -1.31349826e+00 4.58368480e-01
-5.02126753e-01 -8.61519992e-01 1.18421264e-01 -9.54724014e-01
-2.72650242e-01 5.63672662e-01 1.25499988e+00 6.46256283e-02
6.80162370e-01 -1.90808892e-01 3.76475185e-01 2.07116127e-01
5.02122045e-01 -1.54575861e+00 -2.30561122e-01 9.68080088e-02
1.23694015e+00 -7.26193428e-01 -3.61886293e-01 -7.68669009e-01
-8.59583080e-01 4.93178755e-01 8.90428185e-01 4.23802525e-01
5.91666877e-01 2.45531097e-01 2.64646262e-01 -4.08791959e-01
-8.11529040e-01 -3.80820066e-01 4.42342669e-01 1.00377381e+00
3.48399490e-01 4.62334394e-01 -9.95643854e-01 3.66719604e-01
-2.97329992e-01 3.05969596e-01 7.07136452e-01 1.14874637e+00
-7.38073409e-01 -4.27118182e-01 -3.13404977e-01 8.37366402e-01
4.21563953e-01 6.13010764e-01 -1.58470243e-01 1.28212500e+00
2.19762489e-01 1.45043719e+00 3.04604739e-01 -6.48971558e-01
1.24539649e+00 3.07768941e-01 2.47096792e-01 -5.22864342e-01
-8.65952596e-02 -4.17611264e-02 5.56301951e-01 -7.73743391e-01
-8.20807815e-02 -9.65799689e-01 -1.42593074e+00 -3.80894810e-01
-5.30452371e-01 1.66151866e-01 9.68203425e-01 8.36044073e-01
7.27909803e-01 7.20084906e-01 5.15734196e-01 -7.69845665e-01
-5.09027004e-01 -1.18072808e+00 -8.26750875e-01 5.73375560e-02
2.67298877e-01 -1.70892310e+00 -5.33530354e-01 -6.33244693e-01] | [5.754054069519043, 1.083006739616394] |
4033c675-1c91-44ae-9a7f-3adbfaf5b734 | swem-towards-real-time-video-object-1 | 2208.10128 | null | https://arxiv.org/abs/2208.10128v1 | https://arxiv.org/pdf/2208.10128v1.pdf | SWEM: Towards Real-Time Video Object Segmentation with Sequential Weighted Expectation-Maximization | Matching-based methods, especially those based on space-time memory, are significantly ahead of other solutions in semi-supervised video object segmentation (VOS). However, continuously growing and redundant template features lead to an inefficient inference. To alleviate this, we propose a novel Sequential Weighted Expectation-Maximization (SWEM) network to greatly reduce the redundancy of memory features. Different from the previous methods which only detect feature redundancy between frames, SWEM merges both intra-frame and inter-frame similar features by leveraging the sequential weighted EM algorithm. Further, adaptive weights for frame features endow SWEM with the flexibility to represent hard samples, improving the discrimination of templates. Besides, the proposed method maintains a fixed number of template features in memory, which ensures the stable inference complexity of the VOS system. Extensive experiments on commonly used DAVIS and YouTube-VOS datasets verify the high efficiency (36 FPS) and high performance (84.3\% $\mathcal{J}\&\mathcal{F}$ on DAVIS 2017 validation dataset) of SWEM. Code is available at: https://github.com/lmm077/SWEM. | ['Wei Liu', 'Wenhao Jiang', 'Chun Yuan', 'Ziyu Wang', 'Maomao Li', 'Tianyu Yang', 'Zhihui Lin'] | 2022-08-22 | swem-towards-real-time-video-object | http://openaccess.thecvf.com//content/CVPR2022/html/Lin_SWEM_Towards_Real-Time_Video_Object_Segmentation_With_Sequential_Weighted_Expectation-Maximization_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Lin_SWEM_Towards_Real-Time_Video_Object_Segmentation_With_Sequential_Weighted_Expectation-Maximization_CVPR_2022_paper.pdf | cvpr-2022-1 | ['semi-supervised-video-object-segmentation'] | ['computer-vision'] | [ 1.05715036e-01 -2.93785155e-01 -2.07362890e-01 -5.21952212e-01
-4.92256194e-01 -1.67900115e-01 -5.22728376e-02 -2.58259952e-01
-5.67889690e-01 4.76286829e-01 -3.66424620e-01 -1.38464421e-01
-1.74482763e-01 -7.27974653e-01 -6.83272958e-01 -6.66134894e-01
2.35818446e-01 1.00258537e-01 6.50567770e-01 3.99099350e-01
2.29691878e-01 3.05361301e-01 -1.84531903e+00 1.84757799e-01
8.68424296e-01 1.35444462e+00 6.99446559e-01 4.27972764e-01
-2.77960509e-01 6.44201398e-01 -3.95021230e-01 -2.67799139e-01
2.88900644e-01 -1.84577212e-01 -6.36154115e-01 2.83476532e-01
4.98587191e-01 -3.59788179e-01 -6.35057151e-01 1.03668857e+00
4.65168953e-01 3.13391328e-01 3.00976902e-01 -1.46504760e+00
-2.21279502e-01 9.75873321e-02 -6.72955573e-01 4.49191272e-01
2.22799599e-01 2.21550941e-01 6.53370857e-01 -1.03300631e+00
6.61699653e-01 1.03739166e+00 5.24339139e-01 5.49628258e-01
-6.94616318e-01 -8.00589383e-01 1.46067843e-01 5.28266549e-01
-1.65318584e+00 -5.63144147e-01 6.47383809e-01 -3.02645564e-01
7.89524317e-01 4.63934422e-01 7.10779190e-01 6.19806468e-01
-1.52427346e-01 1.29600763e+00 7.70667613e-01 -1.41331092e-01
5.53748310e-02 -3.75775211e-02 1.60739020e-01 9.54496324e-01
1.86640963e-01 -2.34711468e-01 -6.58921659e-01 5.54609857e-02
9.14378822e-01 2.27033779e-01 -3.27682585e-01 -3.07620287e-01
-1.04626167e+00 6.27125859e-01 -2.21132375e-02 2.10605524e-02
-1.60434067e-01 7.82968029e-02 3.50995243e-01 7.04937205e-02
1.82625607e-01 -2.41563022e-01 -4.50642675e-01 -3.33532214e-01
-1.30803657e+00 1.50355756e-01 4.62849945e-01 1.22528780e+00
8.20731521e-01 9.83370245e-02 -1.88950792e-01 7.82371581e-01
5.02257288e-01 5.68994582e-01 5.56442320e-01 -1.34481621e+00
4.44830865e-01 7.13274181e-01 -2.04058122e-02 -1.21491468e+00
-3.33027631e-01 -1.16893217e-01 -8.47238481e-01 -1.43100083e-01
3.16416979e-01 -4.91416678e-02 -8.09113681e-01 1.50444317e+00
6.19863510e-01 5.24843633e-01 -3.90177935e-01 9.98331308e-01
1.01726711e+00 6.66088700e-01 -7.61984810e-02 -4.12490606e-01
1.08887911e+00 -9.89713430e-01 -7.21297741e-01 -1.13893233e-01
3.15912724e-01 -8.44315410e-01 8.79496098e-01 2.14281142e-01
-1.23263490e+00 -7.88086414e-01 -9.98404801e-01 -3.13598365e-02
-1.10593580e-01 1.87701017e-01 7.12570190e-01 4.51612920e-01
-8.27566803e-01 6.47533596e-01 -9.52151716e-01 2.72314670e-03
7.95894921e-01 5.58609068e-01 -4.73337285e-02 -4.17344049e-02
-9.68045115e-01 3.64441216e-01 3.74420792e-01 3.11218500e-01
-5.97148299e-01 -5.13105571e-01 -8.95668268e-01 -4.64328602e-02
5.32675803e-01 -3.30588907e-01 1.04854560e+00 -8.25225174e-01
-1.43513012e+00 6.48658097e-01 -5.39990842e-01 -1.84320658e-01
5.26405156e-01 -1.50391579e-01 -2.66805440e-01 4.21149194e-01
7.70777790e-03 1.08455145e+00 1.03350997e+00 -7.70133495e-01
-7.76246011e-01 -1.61610171e-01 -1.87488735e-01 3.04685486e-03
-5.33101797e-01 1.25231773e-01 -8.64737988e-01 -5.43101311e-01
3.09904128e-01 -7.81071126e-01 -1.00947514e-01 3.45802903e-01
-2.85714795e-03 -3.53235900e-01 9.41633165e-01 -5.61957061e-01
1.53567064e+00 -2.29433894e+00 -6.51706606e-02 1.53646290e-01
1.91749066e-01 5.89454472e-01 -1.15535725e-02 -8.54762346e-02
1.96520314e-01 -1.45771995e-01 -1.36092380e-01 -3.12488139e-01
-9.60606486e-02 3.27653825e-01 1.81954414e-01 5.33265293e-01
1.51865214e-01 8.90836477e-01 -7.13655591e-01 -1.11419559e+00
4.70834881e-01 4.16102409e-01 -3.23103458e-01 6.86562508e-02
-5.30923307e-02 1.75357074e-01 -4.46182698e-01 8.26302528e-01
8.56380582e-01 -4.62239057e-01 1.09988958e-01 -1.56434581e-01
-9.77949277e-02 -3.78791839e-02 -1.72095871e+00 1.75869322e+00
7.05007836e-02 5.03548801e-01 7.57048503e-02 -1.00615120e+00
8.76335740e-01 1.95488751e-01 7.59766757e-01 -7.12022066e-01
3.75943452e-01 4.82185557e-02 -2.34145418e-01 -6.02549970e-01
4.70239758e-01 5.91029763e-01 2.72278726e-01 3.15929860e-01
1.50500655e-01 1.67741820e-01 5.18612504e-01 1.23546720e-01
7.90538967e-01 4.56285417e-01 2.39007361e-02 -1.53887600e-01
6.04602277e-01 -2.89353371e-01 1.21947956e+00 5.42069435e-01
-4.75580305e-01 6.12452567e-01 1.77624281e-02 -4.34163272e-01
-8.16643298e-01 -8.53809416e-01 -2.27018937e-01 8.09538543e-01
5.38794816e-01 -6.62996829e-01 -9.25594628e-01 -4.87098217e-01
-2.28292570e-02 3.02263439e-01 -1.67182386e-01 -3.83625180e-02
-5.82933784e-01 -5.88223159e-01 3.63209754e-01 7.09308565e-01
7.62954891e-01 -1.06577337e+00 -9.67139006e-01 1.94933727e-01
-3.80377173e-01 -1.12189913e+00 -6.00541234e-01 -2.31837064e-01
-9.12983477e-01 -1.02373397e+00 -6.31397188e-01 -7.67166913e-01
7.45925188e-01 5.36237419e-01 8.09148967e-01 3.54414493e-01
-5.45042157e-01 3.16454500e-01 -2.50130832e-01 -2.65795499e-01
1.28824353e-01 -9.54916179e-02 1.34265661e-01 4.15653065e-02
6.98844731e-01 -4.76480305e-01 -8.06965232e-01 5.40245354e-01
-8.11954796e-01 3.74158591e-01 4.03723627e-01 7.66480863e-01
9.03645694e-01 -2.17102189e-03 2.96684861e-01 -4.52602684e-01
-9.89889279e-02 -2.89423972e-01 -8.11165631e-01 2.20530421e-01
-6.06095195e-01 -2.76920259e-01 2.73937762e-01 -6.42518282e-01
-9.25516069e-01 2.46935099e-01 -3.17933853e-03 -7.89218247e-01
-1.32975623e-01 8.41210261e-02 -2.57158995e-01 -1.78496495e-01
3.18044424e-02 5.06780326e-01 1.57153368e-01 -4.87475187e-01
-7.95025229e-02 7.63953626e-01 3.35378945e-01 -4.76679057e-01
5.97357810e-01 4.95234966e-01 -1.27307758e-01 -6.52058840e-01
-6.31819189e-01 -4.44051325e-01 -5.85355401e-01 -5.67648113e-01
7.71237314e-01 -8.86603832e-01 -1.07978511e+00 6.08265877e-01
-9.19568837e-01 -1.74081862e-01 -1.95156291e-01 5.20989537e-01
-6.11167133e-01 7.32411623e-01 -7.52401888e-01 -8.49158943e-01
-2.97320157e-01 -1.18444133e+00 7.90239155e-01 6.26393914e-01
-2.45004356e-01 -5.04938364e-01 -5.47792912e-01 6.00941598e-01
2.66606033e-01 -1.06513104e-03 4.04899359e-01 -1.32864386e-01
-8.02363992e-01 -2.98919678e-02 -4.30583209e-01 3.83080542e-01
-3.26409228e-02 1.72748029e-01 -7.28347003e-01 -2.58877158e-01
8.78703520e-02 -1.70553312e-01 8.88852298e-01 5.17215133e-01
1.48118317e+00 1.18298717e-01 -2.70899862e-01 6.16808057e-01
1.21209586e+00 3.75605106e-01 6.23365343e-01 3.20882827e-01
7.05110669e-01 2.07224742e-01 1.13143635e+00 7.65306532e-01
4.12571102e-01 5.18487215e-01 3.23935360e-01 4.90303710e-02
-1.93787999e-02 -3.70541029e-02 3.21816146e-01 1.05429471e+00
-8.36602077e-02 -1.99532479e-01 -6.07667804e-01 5.23510218e-01
-2.10081220e+00 -8.88148546e-01 -1.99547097e-01 2.18447804e+00
8.49327683e-01 2.13427588e-01 -1.17479416e-03 2.50429869e-01
1.01316333e+00 2.15068772e-01 -7.02556789e-01 -1.07716173e-01
-2.17870608e-01 1.83552250e-01 4.66096938e-01 1.54625878e-01
-1.19192040e+00 8.38667512e-01 5.29839420e+00 1.26476860e+00
-8.71169090e-01 1.80846512e-01 5.54077208e-01 -3.97226870e-01
5.20603172e-02 -8.65425840e-02 -9.40993845e-01 8.87938559e-01
7.33350217e-01 4.53842022e-02 3.59761119e-01 8.16133738e-01
1.46855533e-01 -3.74382675e-01 -7.93556035e-01 1.34085333e+00
3.26603763e-02 -1.42976975e+00 -1.33914754e-01 -2.07145989e-01
5.41532934e-01 -6.35559782e-02 -3.69158834e-02 8.07850063e-02
-1.81169599e-01 -6.52906895e-01 7.63183475e-01 6.45861387e-01
7.48806655e-01 -6.31410182e-01 5.98957479e-01 2.60474473e-01
-1.48829782e+00 6.32554293e-02 -6.25772536e-01 4.54716496e-02
2.81307817e-01 6.65879548e-01 -2.07166016e-01 5.74376643e-01
9.97415066e-01 6.82196379e-01 -5.40315211e-01 1.02922177e+00
6.70751259e-02 5.16071498e-01 -5.40594816e-01 5.36966547e-02
9.29054245e-03 -1.23447262e-01 3.95362139e-01 1.23137319e+00
2.59019911e-01 2.25612789e-01 2.54083365e-01 5.92889607e-01
-8.07509571e-03 9.88721102e-02 -4.76419590e-02 2.05362830e-02
6.09576344e-01 1.28139091e+00 -9.02053773e-01 -4.85006213e-01
-6.74835682e-01 8.55750263e-01 1.32209033e-01 1.29389361e-01
-1.18860245e+00 -2.98901379e-01 4.51785117e-01 3.17549556e-02
6.12930596e-01 -4.18818384e-01 -1.04868010e-01 -1.13815296e+00
3.38835657e-01 -6.16310775e-01 3.97917241e-01 -6.34526908e-01
-8.88853252e-01 2.77508110e-01 -4.22380716e-02 -1.51991141e+00
1.54129965e-02 -4.12265360e-01 -3.14096630e-01 4.48403984e-01
-1.41514504e+00 -8.74078512e-01 -4.28927600e-01 7.45164573e-01
8.16929460e-01 -1.67497650e-01 4.60377038e-01 6.69346154e-01
-8.71187568e-01 6.56946182e-01 1.63986348e-02 2.14333773e-01
5.31207204e-01 -7.68685639e-01 2.06609711e-01 7.41296053e-01
8.58559161e-02 4.42777842e-01 2.82198906e-01 -6.81143761e-01
-1.54524791e+00 -9.97678220e-01 6.28681839e-01 -5.37155271e-02
2.49824181e-01 -1.71280637e-01 -9.70341444e-01 3.59539062e-01
-5.65737225e-02 3.58724564e-01 6.42405272e-01 -5.12914121e-01
-4.15427871e-02 -1.48987323e-01 -1.16296482e+00 5.05110323e-01
1.38436866e+00 -1.63879305e-01 -2.76627570e-01 2.28438869e-01
4.63071972e-01 -5.22597313e-01 -1.04470289e+00 5.77131391e-01
6.63571000e-01 -9.54986632e-01 8.30176115e-01 -3.23965669e-01
1.28490239e-01 -6.00667119e-01 -2.71090299e-01 -2.98076987e-01
-3.51579726e-01 -4.87458140e-01 -5.38354158e-01 1.38264585e+00
-1.20208254e-02 -4.51785922e-01 9.64604318e-01 7.82083333e-01
5.30632548e-02 -1.01959395e+00 -1.07735801e+00 -8.16826165e-01
-4.87100244e-01 -5.74101627e-01 4.79180098e-01 7.24716127e-01
-3.48511726e-01 -2.01318085e-01 -3.62692237e-01 6.83561666e-03
6.11412644e-01 3.25325042e-01 5.99980533e-01 -1.07703495e+00
-2.71781862e-01 -2.89832026e-01 -6.53169632e-01 -1.24564755e+00
1.05162978e-01 -6.83892727e-01 -1.82119861e-01 -1.22108221e+00
4.04560119e-01 -4.34222370e-01 -4.59301084e-01 4.88539934e-01
-1.61665812e-01 3.46802801e-01 2.97217399e-01 2.80172944e-01
-1.21511924e+00 4.69249666e-01 1.19698548e+00 6.55950606e-02
-7.98323471e-03 -3.74366716e-02 -1.84385955e-01 8.92450511e-01
8.02712619e-01 -4.84838814e-01 -3.24299216e-01 -5.49941421e-01
-5.29781729e-02 3.68276499e-02 2.71623135e-01 -1.04071414e+00
5.22604465e-01 -3.63606274e-01 6.51326060e-01 -9.03582752e-01
5.03496885e-01 -7.39312589e-01 2.27367699e-01 4.45892155e-01
2.10570440e-01 6.96600750e-02 1.32057935e-01 3.38189632e-01
-2.57289320e-01 -4.11911756e-01 7.31549263e-01 -1.72157452e-01
-1.14549947e+00 5.56528509e-01 -3.61380905e-01 -1.49856582e-02
1.17576826e+00 -7.02015996e-01 8.59483927e-02 -1.75675929e-01
-4.92303312e-01 5.34151077e-01 3.94573808e-01 3.94555688e-01
8.83918226e-01 -1.20823371e+00 -3.81635815e-01 2.15724453e-01
-2.52036214e-01 3.02996427e-01 6.77237868e-01 9.95621026e-01
-3.76266420e-01 1.47222176e-01 -1.70910463e-01 -8.73766243e-01
-1.43339753e+00 4.65966642e-01 -5.71607240e-03 2.27801025e-01
-4.77569938e-01 8.89742613e-01 -1.60907164e-01 1.55758858e-02
2.49842227e-01 -1.97919328e-02 1.19549502e-03 9.73845869e-02
5.65295279e-01 7.37653673e-01 -1.89001963e-01 -7.01296091e-01
-4.96304363e-01 6.13603413e-01 -1.12873219e-01 2.03305200e-01
1.32512319e+00 -4.51230288e-01 -3.69594321e-02 3.44515324e-01
1.04142082e+00 -2.39599019e-01 -1.57084692e+00 -3.52751195e-01
-9.98978168e-02 -7.48045385e-01 -8.36931318e-02 -1.76795468e-01
-1.43377185e+00 8.19140077e-01 8.09331596e-01 -2.48417601e-01
1.23439121e+00 -3.05157363e-01 1.27804101e+00 1.95272893e-01
4.21948910e-01 -1.44553030e+00 -1.27095506e-01 3.05694997e-01
2.99828231e-01 -1.27792680e+00 2.03912914e-01 -5.37729204e-01
-4.95665908e-01 1.12599623e+00 7.64352262e-01 -5.36825657e-02
6.21378303e-01 3.38670522e-01 -8.97385776e-02 4.08345722e-02
-4.75803852e-01 -9.22550783e-02 1.27357990e-01 3.38570446e-01
1.77135259e-01 -1.53693870e-01 -4.92038816e-01 5.52592158e-01
7.42413476e-02 2.29523465e-01 1.91652179e-01 1.18682182e+00
-2.78450996e-01 -9.11864042e-01 -1.79661721e-01 5.68194330e-01
-4.00082588e-01 6.60076886e-02 1.54267415e-01 6.37541294e-01
3.31151515e-01 9.88620341e-01 1.86277315e-01 -4.18473184e-01
-2.25174408e-02 1.18582509e-01 5.24556339e-01 -1.42125368e-01
-1.91233560e-01 1.71170712e-01 -3.43603700e-01 -8.07316720e-01
-6.49817646e-01 -9.11705792e-01 -1.54830587e+00 -4.07798648e-01
-4.85749483e-01 6.75975205e-03 5.35459697e-01 8.39545131e-01
6.16539180e-01 3.26670051e-01 5.71517706e-01 -8.79692256e-01
-2.66395688e-01 -7.01979995e-01 -4.15837824e-01 4.10178602e-01
-1.32266819e-01 -8.94655049e-01 -1.56429596e-02 8.21558386e-02] | [9.166088104248047, -0.12747687101364136] |
d9ceca1c-4504-4e8c-9f54-91cd0a7b53b0 | instance-based-counterfactual-explanations | 2009.13211 | null | https://arxiv.org/abs/2009.13211v2 | https://arxiv.org/pdf/2009.13211v2.pdf | Instance-based Counterfactual Explanations for Time Series Classification | In recent years, there has been a rapidly expanding focus on explaining the predictions made by black-box AI systems that handle image and tabular data. However, considerably less attention has been paid to explaining the predictions of opaque AI systems handling time series data. In this paper, we advance a novel model-agnostic, case-based technique -- Native Guide -- that generates counterfactual explanations for time series classifiers. Given a query time series, $T_{q}$, for which a black-box classification system predicts class, $c$, a counterfactual time series explanation shows how $T_{q}$ could change, such that the system predicts an alternative class, $c'$. The proposed instance-based technique adapts existing counterfactual instances in the case-base by highlighting and modifying discriminative areas of the time series that underlie the classification. Quantitative and qualitative results from two comparative experiments indicate that Native Guide generates plausible, proximal, sparse and diverse explanations that are better than those produced by key benchmark counterfactual methods. | ['Eoin Delaney', 'Mark T. Keane', 'Derek Greene'] | 2020-09-28 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 4.96570915e-01 6.50913239e-01 -4.16150481e-01 -6.51398003e-01
-4.93427336e-01 -5.53061306e-01 9.17280912e-01 9.84876081e-02
2.72502333e-01 1.09492123e+00 3.31954837e-01 -8.00573945e-01
-5.45591533e-01 -8.69830012e-01 -6.74734294e-01 -3.34361136e-01
-4.10283864e-01 4.00112629e-01 -1.56617269e-01 -2.28966236e-01
6.55125439e-01 2.16268986e-01 -2.04451680e+00 8.03065896e-01
8.42379689e-01 1.05447710e+00 -2.65936404e-01 3.97573650e-01
-4.00838971e-01 7.36601353e-01 -8.91615510e-01 -9.05425191e-01
3.65058064e-01 -5.19757628e-01 -7.28360593e-01 -2.99675554e-01
3.04049820e-01 1.11965895e-01 -1.15243971e-01 6.75862372e-01
-9.55948830e-02 4.21154611e-02 8.40920448e-01 -2.12335443e+00
-8.42587590e-01 1.11376274e+00 -2.52679676e-01 3.78653616e-01
5.65186143e-01 2.79895425e-01 9.23853934e-01 -3.53725195e-01
6.93803430e-01 1.45801032e+00 5.39560974e-01 7.08301187e-01
-1.48909914e+00 -1.17537737e+00 6.29201412e-01 7.51952112e-01
-7.66672313e-01 1.35603517e-01 1.22802818e+00 -2.69489318e-01
1.35988748e+00 9.52510715e-01 8.51219237e-01 1.16289973e+00
7.22071886e-01 6.95656180e-01 1.35334742e+00 -3.24638009e-01
5.43527186e-01 2.55452637e-02 -2.53084987e-01 1.80929333e-01
9.18458402e-02 6.96980119e-01 -6.94928229e-01 -3.06422442e-01
2.41827101e-01 1.78394288e-01 -1.55506790e-01 -5.44864595e-01
-1.56792748e+00 9.64490473e-01 5.55707753e-01 1.02275215e-01
-4.55465853e-01 1.78419366e-01 5.74454248e-01 4.92940396e-01
8.24291825e-01 1.00492299e+00 -9.89131212e-01 -2.61683077e-01
-8.71540070e-01 7.74180889e-01 7.59978116e-01 9.07586098e-01
3.36627156e-01 3.39213371e-01 -1.24286309e-01 3.31731349e-01
-1.19775414e-01 4.15745109e-01 8.75829756e-01 -9.24596846e-01
3.96334052e-01 8.47825587e-01 1.80012718e-01 -9.11217332e-01
-2.77671456e-01 -4.11041409e-01 -5.98076046e-01 1.24618441e-01
3.16823483e-01 1.95031419e-01 -8.70775759e-01 1.85749531e+00
3.41369778e-01 1.27620816e-01 2.95110047e-01 8.09033930e-01
3.06518912e-01 6.72374189e-01 2.75407970e-01 -6.53357863e-01
9.49791849e-01 -5.26670218e-01 -7.27129519e-01 -2.73622543e-01
6.26734316e-01 -6.13713861e-01 1.11017358e+00 3.25124025e-01
-7.71939158e-01 -4.79336351e-01 -9.87539768e-01 5.31019628e-01
-8.76193285e-01 -5.11966527e-01 1.22074676e+00 5.72734654e-01
-4.34354365e-01 8.15083265e-01 -3.34563285e-01 -1.16721720e-01
3.23564529e-01 3.31402481e-01 -1.47552475e-01 1.77904546e-01
-1.39878976e+00 9.54313695e-01 5.62456906e-01 -2.60270298e-01
-7.65715837e-01 -1.20287299e+00 -8.53474140e-01 1.31526351e-01
4.18489724e-01 -7.54439414e-01 1.36563909e+00 -1.30959392e+00
-1.19266045e+00 6.07446671e-01 -1.10977769e-01 -1.12598372e+00
6.98021054e-01 1.61838651e-01 -9.62229371e-01 -3.97217423e-01
3.74104291e-01 8.73015285e-01 9.73993123e-01 -1.18702602e+00
-5.59197903e-01 -4.64272857e-01 1.60944730e-01 -1.53058857e-01
3.74871433e-01 -2.58630872e-01 5.62691808e-01 -1.00730097e+00
1.12147227e-01 -8.01335573e-01 -3.11245203e-01 -1.39244691e-01
-5.59125364e-01 -3.10145795e-01 9.01871920e-01 -1.03909597e-01
1.35361469e+00 -1.87454939e+00 -3.93421650e-01 1.23189040e-01
-2.13598832e-02 -2.58148372e-01 1.13138437e-01 3.99478406e-01
-9.83243883e-01 4.27264988e-01 -4.91279423e-01 2.83379853e-01
1.87711507e-01 2.45701224e-01 -1.29917526e+00 1.22204730e-02
1.15747690e-01 6.82856560e-01 -8.71619701e-01 -2.69908607e-01
4.79498625e-01 -2.21805468e-01 -6.90033734e-01 7.96663761e-02
-7.50551999e-01 6.97243027e-03 -1.41645864e-01 6.36219621e-01
4.95109439e-01 -3.90725732e-02 7.98022896e-02 1.39191046e-01
-3.63143861e-01 2.94857562e-01 -8.15443635e-01 1.46865761e+00
-3.19917172e-01 8.12305212e-01 -9.97279644e-01 -1.22431099e+00
1.18969834e+00 2.63831586e-01 3.25792015e-01 -7.72249579e-01
-4.44177538e-02 4.60461140e-01 3.82398963e-01 -3.64850521e-01
3.80834907e-01 -6.16854310e-01 -3.81604940e-01 4.76681560e-01
-4.95674878e-01 -6.57707036e-01 3.65814269e-02 -1.39060281e-02
9.93382215e-01 5.41843116e-01 7.60914683e-01 -1.74451843e-01
2.56844074e-01 6.98339522e-01 5.65830827e-01 8.21854055e-01
-2.29880914e-01 5.09299397e-01 6.44882679e-01 -1.20828998e+00
-1.20162332e+00 -9.76876557e-01 -1.18108638e-01 7.87605822e-01
-1.08053796e-01 -2.76816100e-01 -5.55229068e-01 -9.18421447e-01
3.94772410e-01 1.66425335e+00 -1.13482320e+00 -5.26094794e-01
-4.14251715e-01 -5.42936206e-01 6.50438517e-02 4.40363228e-01
4.57998633e-01 -1.30646288e+00 -1.16010129e+00 3.63309085e-01
-1.88008890e-01 -3.51466388e-01 -1.89048871e-01 2.18609259e-01
-1.31890810e+00 -1.19152057e+00 -2.78097838e-01 1.85954720e-01
3.51388603e-01 -5.54167330e-02 1.27107966e+00 -2.75886476e-01
-4.18496817e-01 5.90410382e-02 -1.64088443e-01 -1.22339511e+00
-2.93000042e-01 -4.33906317e-01 3.47097777e-02 -7.50531405e-02
6.80676699e-01 -5.81023633e-01 -6.38637424e-01 2.02118203e-01
-7.91808546e-01 4.42333043e-01 4.23759133e-01 7.32401073e-01
5.07480264e-01 -1.23648949e-01 8.22711408e-01 -1.17535079e+00
5.10321558e-01 -7.48147666e-01 -4.51059252e-01 3.21798593e-01
-1.06238866e+00 3.95027429e-01 1.10243118e+00 -8.67288828e-01
-1.07157171e+00 -5.67135401e-02 5.07067978e-01 -4.82058376e-01
-3.15736890e-01 4.68654513e-01 4.84121479e-02 8.15732539e-01
9.59870815e-01 2.69894838e-01 -2.11006537e-01 -1.58588096e-01
6.52502954e-01 4.26827461e-01 7.09917486e-01 -3.83441359e-01
7.37624288e-01 6.71563625e-01 -7.04961419e-02 -1.42734736e-01
-9.03055549e-01 1.54480189e-01 -3.66481513e-01 -3.25714707e-01
4.42901015e-01 -3.31645817e-01 -7.93147624e-01 -3.24381590e-01
-1.22259736e+00 -7.92627260e-02 -7.95008481e-01 4.48361993e-01
-1.09541976e+00 -4.03124124e-01 4.87784117e-01 -1.01309800e+00
-6.05874509e-02 -7.24817574e-01 8.75672698e-01 -1.28092226e-02
-1.02023625e+00 -8.41327131e-01 7.81742558e-02 2.57706434e-01
3.79081070e-01 6.26387358e-01 1.38462555e+00 -1.04011118e+00
-4.96075749e-01 -2.60036439e-01 -1.20173134e-01 -4.27103698e-01
2.84227589e-03 7.40344524e-02 -1.06213534e+00 1.54968753e-01
-2.23222896e-02 1.61648318e-01 4.19450969e-01 5.95074773e-01
1.55340171e+00 -8.92532349e-01 -5.46820521e-01 3.24215621e-01
1.07777357e+00 8.80100787e-01 4.74136621e-01 6.03823483e-01
-7.02808723e-02 8.05653811e-01 1.09357572e+00 6.23006225e-01
3.48055840e-01 5.17241120e-01 7.58823216e-01 1.10498235e-01
3.67916405e-01 -4.73741651e-01 1.65041283e-01 -1.50839329e-01
4.61912192e-02 -3.82907875e-02 -8.89313757e-01 7.82154322e-01
-2.01745701e+00 -1.52111650e+00 8.35295650e-04 2.01064801e+00
6.51771486e-01 3.60132217e-01 6.76386952e-02 5.25675178e-01
5.67649603e-01 2.50359714e-01 -1.06282377e+00 -9.69566107e-01
1.02250762e-01 7.31116086e-02 8.54368806e-02 1.67875409e-01
-8.09510469e-01 6.27517641e-01 6.42300844e+00 5.14040828e-01
-1.12199724e+00 -3.37042391e-01 8.06163430e-01 -2.95493960e-01
-9.19749796e-01 2.97317356e-01 -2.72216350e-02 4.44630921e-01
1.33486366e+00 -1.07429373e+00 1.65315107e-01 1.20441520e+00
4.79377240e-01 1.04339421e-01 -1.70536160e+00 8.69459629e-01
-6.06405549e-04 -1.88605130e+00 6.26733959e-01 -1.01930469e-01
6.48609340e-01 -7.01214790e-01 3.43414813e-01 4.78878528e-01
3.39694709e-01 -1.11003721e+00 1.15821028e+00 5.82750022e-01
6.84766650e-01 -7.43009150e-01 7.07414031e-01 5.13229966e-01
-9.02939320e-01 -5.83316624e-01 -3.02824676e-01 -6.31564379e-01
-2.82552809e-01 3.80046576e-01 -1.09994030e+00 5.64694941e-01
7.38281965e-01 5.19819379e-01 -3.60543698e-01 7.40148425e-01
-5.07293753e-02 5.17283142e-01 6.54533207e-02 -3.06610733e-01
3.75157669e-02 2.46085137e-01 4.72653925e-01 9.54366386e-01
4.43580449e-01 4.84630227e-01 -3.09212953e-01 1.27460039e+00
3.20463330e-01 -9.00831074e-02 -1.25594878e+00 7.98684061e-02
5.47054052e-01 5.12875915e-01 -4.86246139e-01 -6.58889532e-01
-2.10213974e-01 6.11734390e-01 -1.70333646e-02 3.10920358e-01
-1.07479763e+00 -9.48613957e-02 6.12219751e-01 1.13212258e-01
-1.25811458e-01 4.50716138e-01 -6.67156816e-01 -1.07950473e+00
-1.40571803e-01 -1.07449162e+00 9.05176103e-01 -1.33205962e+00
-1.22967362e+00 7.24305451e-01 5.71142316e-01 -1.61838996e+00
-9.84239280e-01 -4.15789425e-01 -7.09182501e-01 7.38858283e-01
-1.05036724e+00 -8.78476024e-01 1.42396092e-01 4.64452028e-01
9.57697988e-01 -2.96724796e-01 1.08406663e+00 -5.68406999e-01
-2.74099648e-01 1.43225655e-01 -2.76446164e-01 -4.86283004e-01
3.02153051e-01 -1.39427805e+00 4.49541450e-01 6.99187100e-01
3.55724275e-01 7.66522169e-01 1.47420788e+00 -6.83174491e-01
-8.62772882e-01 -1.02379429e+00 1.26263762e+00 -4.67589706e-01
7.52349436e-01 -4.45070677e-02 -8.04511786e-01 7.84687102e-01
6.96322694e-02 -2.40130201e-01 8.08667839e-01 7.38816708e-02
-7.14144289e-01 -3.27947408e-01 -1.38100016e+00 9.66599166e-01
8.84872377e-01 -4.97781783e-01 -1.24426687e+00 2.85867214e-01
8.28879952e-01 -2.18772590e-01 -4.18818235e-01 3.46611172e-01
6.99940622e-01 -1.18064010e+00 8.80969942e-01 -1.20966923e+00
8.57598960e-01 -1.23872943e-01 -1.91246420e-01 -1.62875044e+00
-4.51960750e-02 -7.20582247e-01 -5.09918742e-02 6.58453286e-01
6.87801957e-01 -7.51530230e-01 8.82089555e-01 1.23028398e+00
-1.57642066e-01 -8.08650970e-01 -1.18012178e+00 -7.34800339e-01
6.03716411e-02 -9.46231484e-01 1.43244767e+00 1.06537640e+00
3.74680728e-01 -1.68406382e-01 -1.94337461e-02 -5.00811376e-02
5.25192618e-01 1.00101757e+00 7.14919150e-01 -1.17435253e+00
8.00546184e-02 -4.63163197e-01 -3.09859663e-01 -3.45156878e-01
3.17315012e-01 -7.94719636e-01 -2.97695816e-01 -1.07071733e+00
1.63301811e-01 -1.78498328e-01 -2.04579368e-01 5.73907435e-01
-1.09319374e-01 -9.43083689e-02 3.50017786e-01 2.43717119e-01
-1.91698089e-01 5.69409549e-01 1.03638494e+00 -1.66387349e-01
-1.07263409e-01 1.95264891e-02 -7.87048697e-01 8.79215062e-01
7.26646900e-01 -6.95247233e-01 -6.98354065e-01 1.63991284e-02
1.83903635e-01 5.97544312e-01 6.20032132e-01 -6.64130867e-01
-1.93519562e-01 -7.28897929e-01 3.83841962e-01 -5.10481060e-01
2.03719139e-01 -1.09288359e+00 6.82933450e-01 8.53820562e-01
-8.84378612e-01 5.92289567e-01 4.50390667e-01 6.73206389e-01
-3.52841228e-01 2.04771712e-01 3.66995454e-01 -2.40034163e-01
-6.71076953e-01 8.31183791e-02 -1.54107898e-01 -2.19438434e-01
1.26641834e+00 -6.25447750e-01 -6.37570024e-01 -4.80114490e-01
-6.96405768e-01 -5.18795736e-02 5.08957691e-02 8.15504670e-01
7.26929963e-01 -1.36111510e+00 -5.30232668e-01 1.38910010e-01
5.37452936e-01 -4.51302588e-01 2.89991587e-01 3.50681812e-01
-1.01666376e-01 7.83398807e-01 -4.15434659e-01 -2.69234240e-01
-9.70298111e-01 1.00711536e+00 3.88323814e-01 -1.32857665e-01
-4.62331891e-01 5.70830822e-01 3.89768898e-01 -4.96771425e-01
-1.53917447e-01 -7.16561794e-01 -1.35468289e-01 -8.32685232e-02
3.23527187e-01 2.87636131e-01 -2.43732080e-01 -1.36863247e-01
-3.54755044e-01 -4.80031036e-02 1.07383765e-01 -2.68512666e-01
1.42538977e+00 1.13312505e-01 2.37645298e-01 9.55770195e-01
7.32594430e-01 -5.11933267e-01 -1.14065766e+00 1.72339410e-01
3.00781816e-01 -9.49505925e-01 -5.17541468e-01 -1.32512438e+00
-5.10764420e-01 7.60432363e-01 4.55974162e-01 8.17560792e-01
1.18515325e+00 2.49705464e-02 1.25259891e-01 1.23010702e-01
4.43434507e-01 -7.58726001e-01 -2.14090005e-01 6.85434788e-02
1.48114347e+00 -1.12570632e+00 -8.22926387e-02 -3.11013699e-01
-6.64500713e-01 1.02430940e+00 7.36165285e-01 4.30071577e-02
2.52583385e-01 -9.29648206e-02 8.07994157e-02 -2.59189695e-01
-1.59791982e+00 4.53333855e-01 4.27795172e-01 6.48563325e-01
6.00822628e-01 4.87682909e-01 -3.03947836e-01 8.53302360e-01
-1.11510730e+00 -1.54581010e-01 4.70759004e-01 4.75508004e-01
7.64412060e-02 -7.30600715e-01 -6.52765632e-01 7.87763774e-01
-2.16364637e-01 -1.52474746e-01 -6.22378767e-01 1.35197413e+00
2.02447876e-01 9.24722970e-01 3.19877893e-01 -2.83372611e-01
3.87063652e-01 4.63031411e-01 1.66899234e-01 -5.35758138e-01
-7.17319489e-01 -3.89290959e-01 1.43130869e-01 -8.54664981e-01
-3.83209944e-01 -8.94719541e-01 -1.38722336e+00 -4.34386343e-01
-9.61187482e-02 4.07258570e-01 6.02315187e-01 1.06497371e+00
2.31014118e-01 5.79676211e-01 5.32991111e-01 -4.50825632e-01
-5.48046529e-01 -8.76496851e-01 -2.80515313e-01 7.23900020e-01
3.54065537e-01 -7.70033777e-01 -5.79608560e-01 1.42553836e-01] | [8.742711067199707, 5.6408891677856445] |
83cebba7-90b4-48b9-93ea-deefee83495c | glot500-scaling-multilingual-corpora-and | 2305.12182 | null | https://arxiv.org/abs/2305.12182v2 | https://arxiv.org/pdf/2305.12182v2.pdf | Glot500: Scaling Multilingual Corpora and Language Models to 500 Languages | The NLP community has mainly focused on scaling Large Language Models (LLMs) vertically, i.e., making them better for about 100 languages. We instead scale LLMs horizontally: we create, through continued pretraining, Glot500-m, an LLM that covers 511 predominantly low-resource languages. An important part of this effort is to collect and clean Glot500-c, a corpus that covers these 511 languages and allows us to train Glot500-m. We evaluate Glot500-m on five diverse tasks across these languages. We observe large improvements for both high-resource and low-resource languages compared to an XLM-R baseline. Our analysis shows that no single factor explains the quality of multilingual LLM representations. Rather, a combination of factors determines quality including corpus size, script, "help" from related languages and the total capacity of the model. Our work addresses an important goal of NLP research: we should not limit NLP to a small fraction of the world's languages and instead strive to support as many languages as possible to bring the benefits of NLP technology to all languages and cultures. Code, data and models are available at https://github.com/cisnlp/Glot500. | ['Ayyoob Imani', 'Hinrich Schütze', 'François Yvon', 'André F. T. Martins', 'Helmut Schmid', 'Chunlan Ma', 'Nora Kassner', 'Masoud Jalili Sabet', 'Silvia Severini', 'Amir Hossein Kargaran', 'Peiqin Lin'] | 2023-05-20 | null | null | null | null | ['xlm-r'] | ['natural-language-processing'] | [-4.49194938e-01 3.18424441e-02 -5.52648723e-01 -2.20535189e-01
-1.27219367e+00 -8.85131478e-01 5.97325861e-01 1.02409698e-01
-6.83355570e-01 8.01146209e-01 7.34944522e-01 -7.79222250e-01
2.89596617e-01 -6.07910275e-01 -7.15152979e-01 1.16744516e-02
3.99823785e-01 7.83673108e-01 -3.33610386e-01 -3.93304139e-01
-3.10073774e-02 3.03427756e-01 -6.87856436e-01 3.98193121e-01
1.11270845e+00 -1.94230881e-02 4.84363377e-01 5.35104632e-01
-4.50200111e-01 7.73940325e-01 -5.74542105e-01 -4.86487418e-01
1.65207490e-01 -3.01472187e-01 -9.67868447e-01 -3.24617326e-01
4.50075775e-01 6.81709424e-02 -9.68477800e-02 8.01448047e-01
3.15018386e-01 -1.18335851e-01 5.01934946e-01 -9.36448157e-01
-9.37855840e-01 1.34568787e+00 -4.45380062e-01 1.90774322e-01
3.34580332e-01 2.56401360e-01 1.21091413e+00 -9.64228094e-01
8.61281931e-01 1.44003236e+00 6.99018836e-01 5.52278101e-01
-1.35642636e+00 -7.98894584e-01 2.03097373e-01 -4.23432946e-01
-1.42646801e+00 -7.42923379e-01 2.98259616e-01 -4.59159732e-01
1.45435023e+00 -4.16473858e-02 3.25415701e-01 1.13957334e+00
2.98755933e-02 9.38396335e-01 1.01882124e+00 -7.76698112e-01
-3.60804290e-01 3.71052206e-01 1.16713896e-01 6.08323395e-01
3.82481277e-01 -2.89543599e-01 -5.34825325e-01 -6.76579699e-02
6.28312767e-01 -4.32411373e-01 -2.85443395e-01 2.72930712e-01
-1.44350481e+00 9.04692113e-01 1.30320579e-01 8.89393508e-01
-1.47312194e-01 4.84145805e-02 3.63068134e-01 4.78004336e-01
5.47786117e-01 8.45083892e-01 -9.98311341e-01 -3.03331733e-01
-9.85635340e-01 -7.43878856e-02 9.79103565e-01 1.13216126e+00
6.39338493e-01 2.39242569e-01 1.70068190e-01 1.17980766e+00
2.77930498e-01 7.03422785e-01 6.41752660e-01 -8.46050024e-01
1.14792705e+00 5.55601001e-01 1.50594153e-02 -3.86544049e-01
-4.03634697e-01 -4.28314269e-01 -5.37156999e-01 -4.11103576e-01
7.65514910e-01 -3.90813351e-01 -7.72440553e-01 2.00449562e+00
-2.74362743e-01 -6.28764510e-01 2.84825087e-01 4.91641521e-01
7.13857293e-01 1.08607829e+00 2.32538104e-01 1.49168432e-01
1.26207328e+00 -1.00280738e+00 -4.07120317e-01 -7.88031280e-01
1.19623494e+00 -1.11440885e+00 1.59070897e+00 2.54220754e-01
-1.29795837e+00 -4.53134775e-01 -7.73264647e-01 -3.94831419e-01
-6.02380931e-01 4.97286588e-01 7.27971971e-01 4.78288651e-01
-1.22470903e+00 3.62730086e-01 -7.53805518e-01 -6.59936309e-01
6.47373125e-02 1.10769691e-02 -5.33869982e-01 -4.39074486e-01
-1.17699993e+00 1.09991920e+00 5.43577850e-01 -4.23570573e-02
-7.04127252e-01 -5.54900169e-01 -9.39435005e-01 -6.60132244e-02
9.49163139e-02 -4.56768095e-01 1.17468750e+00 -7.98184514e-01
-1.12379420e+00 9.13964570e-01 -4.23381805e-01 -1.87505007e-01
1.08157516e-01 -4.20442402e-01 -4.68942523e-01 -3.56877476e-01
2.46732503e-01 6.89655960e-01 6.64500194e-03 -1.03825414e+00
-5.89550376e-01 -6.54538870e-02 -1.89145550e-01 2.56643772e-01
-3.71900648e-01 5.71771026e-01 -6.88261569e-01 -4.87844169e-01
-1.75867140e-01 -9.63580489e-01 -2.14339986e-01 -8.35237503e-01
-2.76009649e-01 -2.76519865e-01 -4.82853539e-02 -1.08173442e+00
1.22656572e+00 -1.91085362e+00 5.42262197e-02 1.37891928e-02
-1.74077705e-01 3.41692299e-01 -7.62606561e-01 8.40355396e-01
1.71616584e-01 6.98976040e-01 1.52435899e-03 -6.00913286e-01
-2.13399250e-02 3.86577845e-01 -4.52547252e-01 2.20082656e-01
2.33414575e-01 1.11855888e+00 -7.83565104e-01 -2.92689413e-01
2.80619543e-02 5.25863409e-01 -6.03679657e-01 -2.31018439e-02
-3.13181102e-01 5.01352131e-01 -2.03456298e-01 6.76272213e-01
3.28414530e-01 -2.59040534e-01 3.70417058e-01 4.08051461e-01
-4.60860968e-01 9.51801300e-01 -7.37600744e-01 1.90028131e+00
-9.38782930e-01 5.20626068e-01 1.49694547e-01 -4.24708962e-01
8.67622972e-01 3.78401667e-01 1.68953866e-01 -6.37245595e-01
2.03586388e-02 8.23856831e-01 2.58298337e-01 -5.80964945e-02
7.69269764e-01 -5.79697005e-02 -2.49442354e-01 4.70343620e-01
2.52270252e-01 -2.88799763e-01 3.94973844e-01 3.07031870e-01
8.87904167e-01 1.80539683e-01 4.41435337e-01 -4.82523590e-01
3.20594102e-01 2.82908589e-01 6.34507477e-01 7.57943213e-01
2.49465723e-02 3.26755524e-01 3.55157167e-01 -1.38226137e-01
-1.31613719e+00 -9.22200382e-01 -1.49311647e-01 1.30044758e+00
-7.21661448e-01 -5.91559172e-01 -4.22978908e-01 -4.87555414e-01
3.80454659e-02 9.93094027e-01 -1.60614476e-01 2.21128315e-01
-1.01602387e+00 -8.93955588e-01 8.42241228e-01 3.73171628e-01
9.06888098e-02 -1.17142856e+00 3.60042006e-01 3.00183058e-01
-5.22326410e-01 -1.22207344e+00 -5.95726132e-01 1.88078672e-01
-8.31447184e-01 -7.13359237e-01 -5.85986137e-01 -9.49780762e-01
3.82381499e-01 -1.20597929e-02 1.55179286e+00 -7.52278790e-02
5.30965216e-02 1.95212752e-01 -3.42592269e-01 -3.48154068e-01
-8.14045429e-01 6.29237652e-01 -1.12435371e-02 -5.78103065e-01
7.05428720e-01 -2.76154786e-01 1.60657197e-01 -1.89748496e-01
-5.00575423e-01 -2.60641761e-02 8.14296484e-01 5.58657587e-01
4.67911690e-01 -2.80517399e-01 4.58375573e-01 -1.03817904e+00
6.66290998e-01 -6.58895969e-01 -5.74890554e-01 4.20573145e-01
-4.49802220e-01 -7.16486573e-03 8.54233384e-01 -3.12088519e-01
-8.71033072e-01 -2.74967194e-01 -4.85268056e-01 7.03443810e-02
2.18077209e-02 7.06395924e-01 -3.09499711e-01 3.28057021e-01
5.51824689e-01 2.24302635e-01 -4.73950773e-01 -9.45423961e-01
6.41393542e-01 6.91526234e-01 2.41741881e-01 -9.08517420e-01
6.59713268e-01 -1.84465237e-02 -6.47004187e-01 -1.12393999e+00
-8.49581599e-01 -5.34125924e-01 -4.86650467e-01 4.05334055e-01
6.07923448e-01 -1.29595935e+00 -2.26852357e-01 2.71305799e-01
-1.26482582e+00 -7.80943811e-01 -2.01238886e-01 7.42328167e-01
-1.37704313e-01 1.93769857e-02 -1.14446890e+00 -5.93650937e-01
-5.40466070e-01 -9.89614546e-01 9.96443987e-01 -4.91072610e-02
-4.82552141e-01 -1.32119012e+00 3.58546287e-01 5.23742676e-01
4.50809419e-01 -2.32076064e-01 1.14529347e+00 -7.58602083e-01
-3.86868149e-01 5.76436557e-02 -1.86904073e-01 4.53172833e-01
4.36760066e-03 -5.68521209e-02 -7.16940105e-01 -5.65343678e-01
-2.91769296e-01 -6.68677807e-01 7.83805370e-01 3.75481069e-01
5.16158581e-01 -5.22700429e-01 -2.08680496e-01 7.48024821e-01
1.64329827e+00 3.50703783e-02 3.66007030e-01 4.38862026e-01
7.86820412e-01 5.70065916e-01 3.05478066e-01 1.39904236e-02
6.90811813e-01 2.65807629e-01 -4.20363724e-01 -1.43569276e-01
-2.32166603e-01 -7.92011857e-01 7.17952788e-01 1.56590223e+00
1.24840476e-01 -4.99163657e-01 -1.51193750e+00 8.79199028e-01
-1.43240845e+00 -5.22028029e-01 -8.98051262e-02 2.04049301e+00
1.31969202e+00 2.94952244e-02 -2.20611379e-01 -4.70059484e-01
3.30011219e-01 9.47445482e-02 -1.40889868e-01 -5.71517766e-01
-4.78031844e-01 3.15419018e-01 5.27314544e-01 1.05940068e+00
-6.53100550e-01 1.64536321e+00 6.55481291e+00 8.08030128e-01
-1.15468919e+00 3.28213364e-01 4.94235069e-01 -2.08066061e-01
-7.02291071e-01 2.29276881e-01 -1.46445453e+00 3.26428205e-01
1.35741055e+00 -3.22587162e-01 6.75069749e-01 8.09214294e-01
2.52579629e-01 1.93202049e-01 -1.01959157e+00 6.31337821e-01
4.61529158e-02 -1.07773459e+00 2.09051520e-01 2.65219539e-01
8.33451271e-01 9.89187539e-01 -1.76043719e-01 6.85811222e-01
1.03808916e+00 -1.21030998e+00 7.69916534e-01 2.57688671e-01
9.04218137e-01 -6.63751066e-01 5.85477054e-01 5.92038751e-01
-1.12185001e+00 2.06002623e-01 -6.25759542e-01 -4.00291085e-02
3.69958878e-01 7.09174871e-01 -8.42073739e-01 2.50941664e-01
5.77445447e-01 5.94718456e-01 -7.00437963e-01 5.02860427e-01
-4.46216285e-01 1.01599252e+00 -4.16788012e-01 7.39042610e-02
3.97491097e-01 -1.93630099e-01 4.37228054e-01 1.67599857e+00
3.44014645e-01 -2.98291028e-01 4.09046173e-01 6.73783302e-01
-3.50227863e-01 6.95173860e-01 -8.00337136e-01 -6.62928700e-01
7.57261992e-01 1.16070414e+00 -3.12842458e-01 -4.10799026e-01
-6.97982728e-01 6.52121007e-01 8.60867560e-01 4.13007796e-01
-3.93165827e-01 -3.45364124e-01 6.27678931e-01 5.97671680e-02
-2.50690803e-02 -7.10742235e-01 -1.68374375e-01 -1.55744970e+00
-1.98513374e-01 -1.38802838e+00 3.59564245e-01 -6.73797846e-01
-1.33024299e+00 8.50985229e-01 -2.96740443e-01 -5.24844885e-01
-3.77995521e-01 -7.16385126e-01 -4.28977087e-02 1.41390848e+00
-1.67558265e+00 -1.56387997e+00 3.67439985e-01 4.41258669e-01
4.79360372e-01 -3.72914940e-01 8.57459366e-01 2.74024695e-01
-5.68218768e-01 6.39806390e-01 3.01278859e-01 4.72820848e-01
8.64127159e-01 -1.17344737e+00 8.85762691e-01 9.95324135e-01
6.38356864e-01 1.16475451e+00 3.84239435e-01 -7.80425310e-01
-1.24320376e+00 -1.06307995e+00 1.95110857e+00 -8.88602734e-01
9.96174335e-01 -7.31292963e-01 -8.70487392e-01 1.36301935e+00
4.77535933e-01 -4.63650912e-01 6.84059978e-01 6.56598091e-01
-5.26335955e-01 4.38061357e-02 -7.05829620e-01 6.81808889e-01
8.65239143e-01 -6.65119708e-01 -6.46701217e-01 3.95348340e-01
9.51837718e-01 -1.88790932e-01 -1.11482584e+00 5.67599796e-02
3.26585025e-01 -3.88280928e-01 6.26190126e-01 -6.91902041e-01
3.09558421e-01 -1.62569887e-03 -3.35884631e-01 -1.61791503e+00
-4.03222561e-01 -5.88793457e-01 2.88613707e-01 1.52067029e+00
9.81822014e-01 -9.24668193e-01 4.63342518e-01 4.00855035e-01
-1.96193248e-01 -4.69784707e-01 -6.10028207e-01 -9.68523204e-01
9.08411086e-01 -6.59231484e-01 6.26977861e-01 1.24884355e+00
-8.54831040e-02 6.92281604e-01 -2.79370517e-01 4.68585640e-03
3.71743560e-01 3.18415649e-02 8.11412573e-01 -8.74448359e-01
-5.40860891e-01 -4.74293441e-01 4.00490940e-01 -1.01752460e+00
2.41833851e-01 -1.34171271e+00 -3.69361080e-02 -1.76171482e+00
2.19716102e-01 -8.17013800e-01 -9.74315405e-02 9.70611811e-01
-3.04209534e-02 1.18763879e-01 3.38039547e-01 4.34542924e-01
-1.38595119e-01 2.16651917e-01 1.05020046e+00 1.59204587e-01
-4.29250628e-01 -4.99339700e-01 -1.10605049e+00 7.37119734e-01
1.03666997e+00 -4.53498900e-01 -1.43224716e-01 -1.19750822e+00
4.43712711e-01 -1.73727557e-01 -2.46687055e-01 -7.95016885e-01
-3.28260288e-02 -3.22057128e-01 3.70571017e-01 -4.34061825e-01
1.51151478e-01 -3.90290618e-01 -8.15205127e-02 3.11366349e-01
-2.83639491e-01 2.96465278e-01 5.16076088e-01 -3.34520280e-01
-2.55227953e-01 -3.29922080e-01 6.46832407e-01 -5.40548563e-01
-4.41619515e-01 1.38222143e-01 -3.07728797e-01 6.78453684e-01
2.99979478e-01 4.26630884e-01 -5.46175599e-01 -2.13173941e-01
-2.42544293e-01 2.98426121e-01 7.02879190e-01 6.09770358e-01
-8.59213062e-03 -1.10843778e+00 -1.12292266e+00 1.23254761e-01
1.40604600e-01 -2.36715078e-01 -1.71777815e-01 5.06697118e-01
-5.69145441e-01 1.06422913e+00 1.90611601e-01 -6.74752593e-02
-8.37206304e-01 4.28829819e-01 1.49640977e-01 -6.37568951e-01
-5.13794005e-01 8.79118621e-01 1.23866975e-01 -9.73180354e-01
9.93041322e-03 -4.50519979e-01 -1.58684194e-01 -6.66179061e-02
5.06802022e-01 7.21126869e-02 -1.04893103e-01 -7.65350997e-01
-3.54791284e-01 3.77728641e-01 -1.96721971e-01 -4.86042142e-01
1.48282826e+00 -9.48688686e-02 -2.20419854e-01 5.89069486e-01
1.10549760e+00 7.87698090e-01 -6.53931022e-01 -2.85819232e-01
2.07778499e-01 -1.92794561e-01 -1.23386987e-01 -1.06535757e+00
-7.71442533e-01 9.05718207e-01 -1.95074379e-01 -3.01889122e-01
6.37903094e-01 2.23134995e-01 9.76522028e-01 4.00201380e-01
5.26379585e-01 -1.20339322e+00 -2.47269779e-01 1.04144561e+00
8.96484137e-01 -1.16015399e+00 -2.12305874e-01 1.09412111e-02
-7.21332073e-01 6.16424680e-01 4.78121400e-01 6.44408315e-02
4.29419339e-01 4.01761502e-01 2.40007311e-01 2.70629060e-02
-9.56740260e-01 -1.61550641e-01 1.79363802e-01 3.27944785e-01
1.14308190e+00 3.00475210e-01 -3.48136336e-01 6.96609974e-01
-6.91057146e-01 -3.72308753e-02 3.52653831e-01 6.77743554e-01
-4.79499221e-01 -1.55901659e+00 -2.73286819e-01 2.57305086e-01
-5.65885723e-01 -7.80370235e-01 -4.03496712e-01 1.06667948e+00
1.89420119e-01 9.53081071e-01 -4.44136523e-02 -2.53030844e-02
2.39232898e-01 4.19429958e-01 3.48679096e-01 -1.04845858e+00
-6.12809896e-01 2.12357745e-01 4.88137782e-01 -3.94246131e-01
6.06969111e-02 -7.38719046e-01 -1.03950202e+00 -6.23630524e-01
2.14664210e-02 3.06875229e-01 7.85673738e-01 6.04782045e-01
2.58480996e-01 6.40239865e-02 1.08484760e-01 -3.81349266e-01
-4.03469592e-01 -1.34785640e+00 -5.02846956e-01 1.26860812e-01
-1.75971054e-02 7.30457809e-03 -3.85582685e-01 -2.01534897e-01] | [10.874897003173828, 9.83910083770752] |
b3275bd1-b61a-4682-9800-469c116c92ad | hifi-wavegan-generative-adversarial-network | 2210.12740 | null | https://arxiv.org/abs/2210.12740v2 | https://arxiv.org/pdf/2210.12740v2.pdf | HiFi-WaveGAN: Generative Adversarial Network with Auxiliary Spectrogram-Phase Loss for High-Fidelity Singing Voice Generation | Entertainment-oriented singing voice synthesis (SVS) requires a vocoder to generate high-fidelity (e.g. 48kHz) audio. However, most text-to-speech (TTS) vocoders cannot work well in this scenario even if the neural vocoder for TTS has achieved significant progress. In this paper, we propose HiFi-WaveGAN which is designed for synthesizing the 48kHz high-quality singing voices from the full-band mel-spectrogram in real-time. Specifically, it consists of a generator improved from WaveNet, a multi-period discriminator same to HiFiGAN, and a multi-resolution spectrogram discriminator borrowed from UnivNet. To better reconstruct the high-frequency part from the full-band mel-spectrogram, we design a novel auxiliary spectrogram-phase loss to train the neural network, which can also accelerate the training process. The experimental result shows that our proposed HiFi-WaveGAN significantly outperforms other neural vocoders such as Parallel WaveGAN (PWG) and HiFiGAN in the mean opinion score (MOS) metric for the 48kHz SVS task. And a comparative study of HiFi-WaveGAN with/without phase loss term proves that phase loss indeed improves the training speed. Besides, we also compare the spectrogram generated by our HiFi-WaveGAN and PWG, which shows our HiFi-WaveGAN has a more powerful ability to model the high-frequency parts. | ['Xing He', 'Chang Zeng', 'Chunhui Wang'] | 2022-10-23 | null | null | null | null | ['singing-voice-synthesis'] | ['speech'] | [ 1.02193013e-01 -1.79242175e-02 2.12958530e-01 1.05572775e-01
-1.05564737e+00 -3.41623932e-01 1.63227007e-01 -7.92067289e-01
1.90391019e-02 6.69053972e-01 4.24842536e-01 -3.17366540e-01
1.37230039e-01 -6.55088246e-01 -7.84926176e-01 -9.32721972e-01
1.37757689e-01 -1.52340919e-01 -7.24327099e-03 -4.51495051e-01
-4.61303711e-01 -6.57389686e-02 -1.75592148e+00 1.58320859e-01
8.40342522e-01 8.46142709e-01 4.85468239e-01 1.03022480e+00
3.74850005e-01 5.85493803e-01 -1.02740860e+00 -1.74050570e-01
3.71011883e-01 -1.06831336e+00 -2.32259974e-01 -3.95669967e-01
4.06697750e-01 -3.64871264e-01 -6.64259136e-01 8.94434392e-01
1.14047980e+00 2.85256594e-01 5.04591107e-01 -1.02714932e+00
-2.31945902e-01 8.23546767e-01 -4.85735945e-02 2.21197377e-03
1.37041613e-01 4.77988243e-01 1.16174257e+00 -7.56853461e-01
3.50817472e-01 1.20517576e+00 7.73011208e-01 7.06628859e-01
-8.38187754e-01 -1.11059344e+00 -5.23601353e-01 3.22469592e-01
-1.41524506e+00 -5.26115119e-01 1.17734659e+00 -1.36082172e-01
7.20825553e-01 6.64179206e-01 7.70036519e-01 1.21611488e+00
3.46581377e-02 6.21026516e-01 9.49511945e-01 -4.82184440e-01
-1.20842069e-01 -8.35767090e-02 -5.55462420e-01 4.82359111e-01
-3.19398791e-01 5.50803304e-01 -7.80536890e-01 1.96657807e-01
8.47633779e-01 -5.43760121e-01 -8.07848871e-01 6.12234890e-01
-1.12045062e+00 5.91243982e-01 1.86460882e-01 5.65751553e-01
-3.68272603e-01 3.63409907e-01 3.14373106e-01 5.61505795e-01
3.12917471e-01 6.19017124e-01 -1.57754391e-01 -2.77894497e-01
-1.27413452e+00 3.39397520e-01 7.30759084e-01 6.53918326e-01
3.54405910e-01 1.21872807e+00 -5.68082452e-01 9.65760708e-01
2.42536932e-01 8.34017217e-01 9.47160959e-01 -9.78613079e-01
4.26937759e-01 -2.81027108e-01 -2.82016128e-01 -8.56594324e-01
-1.28349125e-01 -7.66555607e-01 -1.15470707e+00 -9.88536794e-03
2.20813602e-01 -7.01263964e-01 -7.09820032e-01 1.77277267e+00
2.55565315e-01 5.01919568e-01 1.66237265e-01 1.10840023e+00
1.09226608e+00 1.38315690e+00 -5.45958459e-01 -4.40180570e-01
1.12163472e+00 -1.22192407e+00 -1.14851058e+00 2.09569365e-01
2.17961371e-01 -8.50839615e-01 1.25977075e+00 5.59718132e-01
-1.17728961e+00 -1.15773988e+00 -1.22245991e+00 -4.08436656e-02
1.15273006e-01 3.36504430e-01 -9.27416831e-02 7.29848325e-01
-1.02514756e+00 8.96554947e-01 -3.02589476e-01 4.18894529e-01
-3.12477052e-01 1.86277822e-01 -1.71154469e-01 5.58704138e-01
-1.63032365e+00 2.67207325e-01 5.27492762e-01 6.09725565e-02
-1.11517167e+00 -9.95469987e-01 -8.00147533e-01 2.53625274e-01
1.92513734e-01 -5.35141647e-01 1.36268556e+00 -7.77505219e-01
-2.28016186e+00 -7.12609068e-02 1.28562339e-02 -5.32065511e-01
3.12852949e-01 6.26515076e-02 -8.17759693e-01 1.94094777e-01
-3.10501665e-01 3.90794665e-01 1.39347017e+00 -8.86579871e-01
-6.20638549e-01 2.08568707e-01 -4.08859909e-01 2.70054013e-01
-5.54061055e-01 -3.18486422e-01 -4.60287146e-02 -1.11217678e+00
-2.68592238e-01 -7.33052194e-01 1.43274218e-01 -6.61993742e-01
-5.89396238e-01 -1.45778939e-01 1.02149582e+00 -1.07989383e+00
1.62646651e+00 -2.40708280e+00 9.03524552e-03 3.22169736e-02
-2.16752607e-02 7.70711601e-01 -2.71975666e-01 5.49108505e-01
-1.86924115e-01 1.40329935e-02 -1.66057125e-01 -4.16964173e-01
8.47402066e-02 2.54411194e-02 -3.76760244e-01 1.77977696e-01
1.21885680e-01 6.57134056e-01 -7.72601545e-01 -4.04803634e-01
1.39981061e-01 8.79838467e-01 -6.91422760e-01 6.32039905e-01
-7.31904665e-03 6.46126568e-01 3.20470370e-02 1.63756609e-01
3.61931980e-01 3.27050805e-01 6.52574282e-03 -3.24421525e-01
-1.66417435e-01 6.05329156e-01 -1.14464748e+00 1.53350365e+00
-7.50258565e-01 7.26140380e-01 2.62668520e-01 -6.43370211e-01
1.16579127e+00 9.40194786e-01 2.52171725e-01 -6.18473411e-01
1.50908351e-01 5.18320978e-01 3.16799521e-01 -5.20803630e-01
4.61362928e-01 -5.25998116e-01 3.04658741e-01 -1.79134216e-02
3.97992402e-01 -5.98303914e-01 1.86504461e-02 -3.45934302e-01
8.14864695e-01 7.43572228e-03 -9.17027220e-02 7.33957887e-02
7.25333929e-01 -5.40662825e-01 7.17546105e-01 3.49711359e-01
2.24410221e-01 9.33541894e-01 1.37498215e-01 2.02806443e-01
-1.22861636e+00 -6.54826939e-01 1.17718592e-01 7.74140596e-01
-3.04998666e-01 -6.49573445e-01 -1.03487062e+00 -2.99305409e-01
-2.89001644e-01 6.71266973e-01 -3.20152752e-02 -2.41903275e-01
-8.58262539e-01 -2.62903661e-01 1.03181422e+00 2.90590048e-01
6.04006827e-01 -1.15479648e+00 -1.92761377e-01 4.63428140e-01
-3.13211381e-01 -8.77824068e-01 -1.02069747e+00 3.80293392e-02
-4.42788899e-01 -5.32029808e-01 -1.09509385e+00 -9.00370181e-01
-1.74843833e-01 -1.39395921e-02 5.81441522e-01 -2.59236366e-01
2.76320159e-01 -2.11332291e-01 -4.44568366e-01 -4.36446339e-01
-7.06604183e-01 1.91984028e-01 2.43836567e-01 2.39082381e-01
-3.78210574e-01 -9.04020190e-01 -5.06969213e-01 1.61772385e-01
-8.76876831e-01 1.46070182e-01 4.86771643e-01 1.05549407e+00
5.66532254e-01 5.77999473e-01 7.58256912e-01 -5.17814040e-01
8.00070167e-01 -1.47718161e-01 -4.24157798e-01 -3.48922640e-01
-3.65622640e-01 -2.32462749e-01 1.31229734e+00 -7.19746292e-01
-9.13369417e-01 -2.28671983e-01 -9.36690450e-01 -7.71879554e-01
1.99885309e-01 3.85397315e-01 -3.04798931e-01 1.84561059e-01
5.34435272e-01 4.20757979e-01 -4.88345288e-02 -6.92954719e-01
1.51030332e-01 1.16149938e+00 8.28052759e-01 -2.55724967e-01
1.18532884e+00 -1.36219889e-01 -1.23680227e-01 -1.32086325e+00
-4.20107722e-01 -2.48583809e-01 2.51178086e-01 -1.19568415e-01
7.24366188e-01 -1.02124894e+00 -8.30223441e-01 7.12953925e-01
-1.21914005e+00 -5.89483619e-01 -6.05182707e-01 1.01831782e+00
-6.05156839e-01 1.85141727e-01 -8.48205090e-01 -9.46861982e-01
-7.23525465e-01 -1.10774815e+00 8.35064888e-01 3.92264843e-01
6.27667680e-02 -6.43109858e-01 8.48835483e-02 2.58512616e-01
5.75587869e-01 3.28736156e-01 6.37555599e-01 -3.32717299e-01
-1.50841221e-01 1.12809218e-01 3.71662587e-01 9.11279380e-01
-1.29074864e-02 -9.42146406e-02 -1.33341408e+00 -3.05627197e-01
3.22742045e-01 -1.09553605e-01 6.69606566e-01 4.28986520e-01
1.15329456e+00 -7.84717023e-01 4.03217226e-01 9.44465518e-01
1.11131394e+00 5.55270731e-01 7.64007688e-01 -3.62962246e-01
8.95280123e-01 4.62308258e-01 4.71499115e-01 3.01525891e-01
1.47895604e-01 7.19980001e-01 2.69991398e-01 -2.17507347e-01
-7.38641858e-01 -6.43545508e-01 7.03298748e-01 1.87944376e+00
-2.67578483e-01 -4.89604682e-01 -2.23490164e-01 4.16078627e-01
-1.30680454e+00 -8.91100883e-01 -3.19740057e-01 1.97934997e+00
1.22662616e+00 -1.54398218e-01 2.60896087e-01 8.48928809e-01
7.31376112e-01 5.83485901e-01 -3.58201653e-01 -4.59403723e-01
-1.60452232e-01 7.73131669e-01 2.81034797e-01 5.86164117e-01
-6.45150244e-01 6.71086848e-01 5.04564762e+00 1.57479894e+00
-1.51267326e+00 3.15109402e-01 3.63718271e-01 -3.16863768e-02
-3.95743161e-01 -4.23473656e-01 -6.75757110e-01 5.83494246e-01
1.42119932e+00 -1.81434870e-01 8.45589817e-01 6.62329435e-01
3.56058896e-01 6.87705576e-01 -7.02176511e-01 1.05841446e+00
-5.11688292e-02 -1.09324753e+00 -8.99914578e-02 -2.73890980e-02
7.15733349e-01 -4.32634890e-01 3.10269713e-01 4.71244603e-01
-2.48449653e-01 -1.15003705e+00 1.00711298e+00 2.11401954e-01
1.49577308e+00 -9.24701869e-01 6.76422775e-01 4.33706880e-01
-1.32821620e+00 8.87012780e-02 -1.53509840e-01 4.61151265e-02
1.22290991e-01 5.92882574e-01 -1.01506495e+00 7.74498522e-01
4.46561992e-01 4.73653555e-01 2.54969239e-01 9.15293574e-01
-4.71858770e-01 1.25700688e+00 -1.37200296e-01 -1.54717481e-02
2.47341514e-01 2.41736168e-04 8.70238125e-01 1.18754148e+00
7.58364916e-01 1.11296624e-02 -3.17887366e-01 6.63968682e-01
-2.41164759e-01 1.30690336e-01 -3.35426480e-01 -3.61086130e-01
4.45038199e-01 1.04415095e+00 1.82100073e-01 -1.77158624e-01
1.48784220e-02 7.78207064e-01 -5.13141394e-01 4.44024444e-01
-1.01064265e+00 -1.06363297e+00 6.90111816e-01 1.89292982e-01
3.55517507e-01 -2.21018046e-01 2.96767980e-01 -8.71171474e-01
-8.22429508e-02 -1.25473714e+00 -1.79262191e-01 -8.65189493e-01
-8.33331645e-01 9.80431497e-01 -4.32483166e-01 -1.41254330e+00
-7.33221352e-01 -3.67010862e-01 -7.36901522e-01 1.10461998e+00
-1.79717910e+00 -9.95557129e-01 -1.32169753e-01 6.67571425e-01
7.13692546e-01 -2.20488057e-01 7.77760625e-01 5.81877649e-01
-4.24603373e-01 7.42203653e-01 1.17427126e-01 1.12921828e-02
7.35918283e-01 -1.15519869e+00 4.34019268e-01 7.72770405e-01
1.00915194e-01 2.43703708e-01 8.30020428e-01 -3.38086993e-01
-1.49275970e+00 -1.39116383e+00 9.03283477e-01 3.47208321e-01
2.50865132e-01 -4.13039893e-01 -8.42078984e-01 1.00990489e-01
4.72030699e-01 -1.51520550e-01 3.93776000e-01 -6.05663836e-01
-1.12933658e-01 -4.49921608e-01 -9.09726083e-01 4.99838918e-01
8.53177369e-01 -7.16516972e-01 -4.35668528e-01 1.18653961e-01
1.16906738e+00 -6.89587712e-01 -9.35513198e-01 4.34274048e-01
5.19269228e-01 -9.78280663e-01 8.03596497e-01 -1.15969621e-01
6.47114694e-01 -4.92891043e-01 -6.99509680e-02 -1.77144206e+00
-1.58236697e-01 -1.46786809e+00 -1.85359299e-01 1.52830100e+00
3.34548682e-01 -6.63883984e-01 4.87648278e-01 -5.24572968e-01
-6.55244589e-01 -6.31547272e-01 -9.37584341e-01 -1.07336140e+00
-5.35752885e-02 -4.03173268e-01 9.16485012e-01 6.83062375e-01
-2.12385833e-01 3.96013588e-01 -9.61067736e-01 1.99623927e-01
2.60167301e-01 -1.95415482e-01 7.80304015e-01 -8.32465351e-01
-7.65574753e-01 -2.54662126e-01 1.36638228e-02 -1.04758143e+00
-2.48057041e-02 -5.77434301e-01 3.85661066e-01 -1.14068341e+00
-6.82301402e-01 -1.98143691e-01 -1.73564374e-01 1.09079719e-01
-2.11472586e-01 4.76910710e-01 5.09025931e-01 -2.14523420e-01
2.70372868e-01 8.81756485e-01 1.87296796e+00 -1.43317640e-01
-3.64158690e-01 1.08227335e-01 -2.47540280e-01 6.37545645e-01
7.50785947e-01 -4.15113777e-01 -4.56300557e-01 -8.59781802e-02
-2.39460394e-01 6.36096597e-01 7.47295916e-02 -1.50251174e+00
1.58108234e-01 1.06825419e-01 -5.10151424e-02 -4.96825576e-01
7.47391760e-01 -6.28876746e-01 4.96266186e-01 5.61382115e-01
-1.82843447e-01 -3.51505935e-01 -2.89178379e-02 1.99058816e-01
-6.11764610e-01 -9.78927836e-02 8.13752830e-01 7.09218830e-02
-1.76600263e-01 2.40196079e-01 -2.71343797e-01 2.46559963e-01
4.03959125e-01 -2.74226703e-02 -2.69012392e-01 -7.67331362e-01
-3.71703446e-01 -2.43458509e-01 -1.66239604e-01 1.31349191e-01
6.05196416e-01 -1.58720827e+00 -9.44555044e-01 5.60928524e-01
-3.79259378e-01 5.87792844e-02 5.28231323e-01 6.55587792e-01
-4.47791338e-01 5.03695667e-01 5.56403538e-03 -2.24869922e-01
-1.10279953e+00 1.04001641e-01 4.12424356e-01 -1.76952764e-01
-5.28340161e-01 9.39208686e-01 2.27556556e-01 -2.87487298e-01
4.37861592e-01 -4.10642684e-01 -2.95648843e-01 -8.11036527e-02
4.88358855e-01 5.85349441e-01 3.92130688e-02 -7.04729438e-01
7.56084323e-02 5.50496519e-01 6.32918656e-01 -4.24785048e-01
1.27306187e+00 1.93991154e-01 1.70845807e-01 4.81970221e-01
1.44390988e+00 5.89840829e-01 -1.12900269e+00 1.09100036e-01
-7.71290183e-01 -2.74367064e-01 3.21383774e-01 -5.07844687e-01
-1.32336926e+00 1.08207595e+00 3.50725144e-01 3.93847793e-01
1.44639468e+00 -5.95245063e-01 1.68405330e+00 -1.39189124e-01
1.37556940e-01 -9.75524783e-01 1.95489317e-01 5.19247353e-01
1.04559004e+00 -4.04377043e-01 -6.59590364e-01 -2.43209198e-01
-6.53076053e-01 1.10462379e+00 3.86122376e-01 -2.59213150e-01
4.71017748e-01 2.76936531e-01 1.75422952e-01 4.01528060e-01
-5.35065293e-01 -1.53083295e-01 5.13322890e-01 4.44632173e-01
3.42812777e-01 6.02510832e-02 -4.26831096e-01 7.37578213e-01
-1.06877208e+00 -3.59610647e-01 5.04674911e-01 2.82811578e-02
-3.12674135e-01 -1.07391679e+00 -5.25750637e-01 2.13768616e-01
-6.71006799e-01 -2.39892304e-01 -4.71200943e-02 4.29465830e-01
3.46181393e-01 1.19839895e+00 -8.41925144e-02 -1.07663620e+00
5.48383713e-01 1.99060217e-01 1.94577917e-01 -3.98123801e-01
-1.04760289e+00 6.70445979e-01 1.76344797e-01 -3.19156498e-01
-6.11406453e-02 -2.44566068e-01 -1.14827490e+00 -3.28972548e-01
-6.01521432e-01 5.07268548e-01 8.14131379e-01 5.12891829e-01
3.46737504e-02 1.27758610e+00 1.09523904e+00 -7.40684152e-01
-6.09652698e-01 -1.16628981e+00 -9.92383420e-01 8.88520777e-02
8.50700974e-01 -2.48755477e-02 -6.16694093e-01 -1.35543281e-02] | [15.486860275268555, 6.195957660675049] |
6d09dc18-3af7-4df8-85b1-086545a75ca9 | mfqe-20-a-new-approach-for-multi-frame | 1902.09707 | null | https://arxiv.org/abs/1902.09707v6 | https://arxiv.org/pdf/1902.09707v6.pdf | MFQE 2.0: A New Approach for Multi-frame Quality Enhancement on Compressed Video | The past few years have witnessed great success in applying deep learning to enhance the quality of compressed image/video. The existing approaches mainly focus on enhancing the quality of a single frame, not considering the similarity between consecutive frames. Since heavy fluctuation exists across compressed video frames as investigated in this paper, frame similarity can be utilized for quality enhancement of low-quality frames given their neighboring high-quality frames. This task is Multi-Frame Quality Enhancement (MFQE). Accordingly, this paper proposes an MFQE approach for compressed video, as the first attempt in this direction. In our approach, we firstly develop a Bidirectional Long Short-Term Memory (BiLSTM) based detector to locate Peak Quality Frames (PQFs) in compressed video. Then, a novel Multi-Frame Convolutional Neural Network (MF-CNN) is designed to enhance the quality of compressed video, in which the non-PQF and its nearest two PQFs are the input. In MF-CNN, motion between the non-PQF and PQFs is compensated by a motion compensation subnet. Subsequently, a quality enhancement subnet fuses the non-PQF and compensated PQFs, and then reduces the compression artifacts of the non-PQF. Also, PQF quality is enhanced in the same way. Finally, experiments validate the effectiveness and generalization ability of our MFQE approach in advancing the state-of-the-art quality enhancement of compressed video. The code is available at https://github.com/RyanXingQL/MFQEv2.0.git. | ['Mai Xu', 'Zulin Wang', 'Tie Liu', 'Ren Yang', 'Qunliang Xing', 'Zhenyu Guan'] | 2019-02-26 | null | null | null | null | ['video-enhancement', 'video-restoration'] | ['computer-vision', 'computer-vision'] | [ 1.97031796e-01 -3.96215469e-01 -1.50386453e-01 -9.52126384e-02
-7.00917006e-01 1.79040655e-02 1.95782632e-01 1.80264805e-02
-4.72641051e-01 5.28245866e-01 3.46027792e-01 -1.09143786e-01
-8.09589997e-02 -8.69886100e-01 -8.16669583e-01 -6.61419809e-01
-1.23760328e-01 -6.05034053e-01 5.44388354e-01 -1.67264462e-01
3.90239686e-01 2.61829734e-01 -1.31690943e+00 5.62070966e-01
6.22576535e-01 1.31228948e+00 6.24183953e-01 8.42794180e-01
1.95900232e-01 7.40476668e-01 -4.97941643e-01 -2.88061708e-01
1.61385491e-01 -5.31612337e-01 -7.56233275e-01 -2.40375623e-02
2.39982337e-01 -9.76855099e-01 -7.02407777e-01 1.24422169e+00
6.12423539e-01 1.40573353e-01 6.76455125e-02 -1.12456453e+00
-7.03968644e-01 4.10099864e-01 -5.49488783e-01 7.37550139e-01
3.34381759e-01 1.99618399e-01 7.18050480e-01 -9.88826931e-01
3.11736703e-01 1.18716037e+00 5.23802638e-01 3.80677164e-01
-5.48318982e-01 -6.79391146e-01 -4.92103957e-02 6.26006544e-01
-1.23141336e+00 -5.10340393e-01 7.74710834e-01 -1.06285214e-01
1.05511117e+00 -1.65252298e-01 8.42439055e-01 5.96869469e-01
6.06250405e-01 7.67858326e-01 5.63982069e-01 -2.10603997e-01
2.92095169e-02 -5.79363525e-01 -3.36491525e-01 5.44079065e-01
-2.80010998e-01 3.88106972e-01 -4.82833922e-01 3.50262046e-01
8.98824155e-01 2.44533449e-01 -5.83955944e-01 2.23041877e-01
-1.23413110e+00 5.17056167e-01 6.69591486e-01 5.48315883e-01
-6.51257873e-01 2.76320010e-01 4.58697587e-01 2.52153337e-01
4.02691662e-01 -3.84867638e-02 -3.32578838e-01 -3.76240075e-01
-1.31865120e+00 1.28445610e-01 1.92459151e-01 6.00912690e-01
6.07904553e-01 6.46344721e-02 -2.56093919e-01 7.23016679e-01
3.30905139e-01 3.70271772e-01 5.14124513e-01 -1.27260435e+00
4.56748515e-01 2.25095898e-01 3.71462591e-02 -1.27193773e+00
-1.04501329e-01 -2.64613748e-01 -1.00821483e+00 2.05331683e-01
-1.32929785e-02 -2.40080692e-02 -5.85641444e-01 1.53744531e+00
1.73859119e-01 4.03469473e-01 -1.62933357e-02 1.15562522e+00
7.92045593e-01 1.19805193e+00 5.97835667e-02 -5.53043902e-01
1.07932329e+00 -1.02093017e+00 -8.48993719e-01 9.09723267e-02
2.44919255e-01 -9.50519025e-01 8.01083028e-01 3.36074948e-01
-1.48481596e+00 -1.06041121e+00 -1.09115481e+00 -9.82281789e-02
9.89682749e-02 -1.72385439e-01 -5.69736809e-02 7.89704025e-02
-1.15913808e+00 1.03423584e+00 -8.11470568e-01 5.09814322e-02
5.39793611e-01 3.22641671e-01 -1.21310890e-01 -1.77589014e-01
-1.50976562e+00 5.83758175e-01 5.55496335e-01 8.59108865e-02
-1.09129906e+00 -7.47635543e-01 -7.57626653e-01 2.43632361e-01
3.60648632e-01 -6.23084426e-01 1.32987165e+00 -1.27812111e+00
-1.31054044e+00 4.42138612e-01 -2.52192229e-01 -3.56194079e-01
1.82650253e-01 -2.88260341e-01 -6.84144497e-01 5.18504083e-01
5.80474995e-02 8.66772413e-01 9.69939530e-01 -9.23035026e-01
-9.78799999e-01 -5.14367372e-02 1.67107865e-01 2.37708315e-01
-3.03252637e-01 2.39907265e-01 -6.75325394e-01 -7.47518361e-01
5.49006183e-03 -3.17100972e-01 1.67022467e-01 1.77482322e-01
-2.48653088e-02 -2.27768943e-01 1.00609207e+00 -9.34299052e-01
1.59833431e+00 -2.34762740e+00 5.42404316e-02 -2.68069595e-01
2.99492180e-01 7.11242557e-01 -4.33238655e-01 1.39940307e-01
-6.48965091e-02 2.11493805e-01 -1.95214644e-01 -8.03945027e-03
-3.06816220e-01 -5.80857359e-02 -5.06679751e-02 2.53513366e-01
4.01478887e-01 9.00237918e-01 -1.02986765e+00 -5.82933009e-01
3.84167552e-01 7.92840719e-01 -4.94737774e-01 2.78559268e-01
4.83753234e-02 3.42352122e-01 -2.63792306e-01 5.83597958e-01
8.32590640e-01 -2.41663307e-01 -1.65176153e-01 -6.71871483e-01
-3.45208913e-01 2.57392138e-01 -9.90666389e-01 1.74928427e+00
-2.66462892e-01 4.79070932e-01 8.63901898e-02 -8.15011859e-01
6.65219963e-01 4.60574150e-01 7.32689023e-01 -1.12825656e+00
2.29086518e-01 2.48258069e-01 9.66319591e-02 -7.01286018e-01
6.14226937e-01 -1.95801169e-01 5.11660397e-01 1.78844750e-01
3.40988114e-02 1.62687063e-01 3.23884279e-01 2.70640701e-02
9.24290955e-01 3.27684313e-01 3.35294724e-01 9.48866159e-02
7.51562834e-01 -7.47282863e-01 7.51086533e-01 2.69467086e-01
-7.20078588e-01 7.93901980e-01 1.66993901e-01 -1.63078979e-01
-1.39720392e+00 -1.05270422e+00 8.98215733e-03 8.68906915e-01
4.75401491e-01 -4.59679872e-01 -8.11254859e-01 -4.21840400e-01
-3.06150913e-01 2.88089246e-01 -2.10602850e-01 -3.81399006e-01
-7.31680214e-01 -2.79238135e-01 3.85693103e-01 4.87312734e-01
1.09734654e+00 -1.28978908e+00 -8.00911963e-01 5.32781065e-01
-5.39329946e-01 -9.73632038e-01 -7.20782042e-01 -3.35025728e-01
-9.17131543e-01 -9.41529393e-01 -1.10618794e+00 -8.53520751e-01
-5.39744981e-02 5.27515173e-01 1.10153031e+00 4.51053619e-01
2.32022628e-01 -5.43092266e-02 -4.81990397e-01 1.59027576e-01
-3.10927868e-01 -3.74607474e-01 -2.00357556e-01 -3.75949405e-03
2.24313512e-01 -5.64923823e-01 -1.09168196e+00 1.56211600e-01
-1.15859234e+00 1.42730683e-01 6.05818629e-01 6.67384923e-01
8.46203089e-01 3.02745879e-01 6.28065467e-01 -3.13568227e-02
4.78107631e-01 -4.38915253e-01 -3.02979738e-01 1.78806819e-02
-5.25772214e-01 -3.68883520e-01 8.10197949e-01 -2.96813369e-01
-9.23939943e-01 -4.46255326e-01 -5.77578843e-01 -7.01944590e-01
-9.96085927e-02 4.68639612e-01 -3.14197838e-01 4.74841259e-02
7.88080916e-02 3.77709657e-01 -1.57089129e-01 -2.65767217e-01
4.50682044e-02 6.64228320e-01 6.72069490e-01 2.87757698e-03
3.69464338e-01 3.22361946e-01 -1.03402965e-01 -4.98472780e-01
-5.73090434e-01 -2.62960315e-01 -3.10256630e-01 -5.52256286e-01
9.65037107e-01 -9.03040230e-01 -7.14060724e-01 6.96036339e-01
-1.28450859e+00 -2.24617615e-01 -1.36357740e-01 6.58015251e-01
-5.37585080e-01 7.94394970e-01 -1.11808538e+00 -5.08961618e-01
-5.95075727e-01 -1.45294785e+00 1.09544659e+00 5.21590769e-01
1.24881089e-01 -8.35693538e-01 -1.05016239e-01 1.84002340e-01
5.06707430e-01 -4.02003452e-02 6.83972895e-01 1.24174906e-02
-5.97246468e-01 1.26013890e-01 -4.99902397e-01 7.36410141e-01
1.94521040e-01 6.38633221e-02 -6.51382327e-01 -4.39847380e-01
1.55870408e-01 -9.74294245e-02 8.72017860e-01 8.52539361e-01
1.14180291e+00 -2.11945906e-01 7.53222927e-02 8.91558945e-01
1.50601172e+00 5.75377107e-01 1.03887880e+00 5.39171040e-01
4.64034408e-01 1.04703556e-03 8.19580734e-01 5.32874525e-01
2.86020339e-01 5.87726831e-01 5.93800306e-01 -2.34464273e-01
-4.45960313e-01 -1.38366297e-01 5.90332448e-01 8.74041200e-01
-1.12664171e-01 -4.33198780e-01 -4.99581695e-01 4.16930497e-01
-1.60010600e+00 -1.20428896e+00 -3.18106227e-02 1.94137263e+00
7.75880873e-01 1.47451341e-01 -1.55866742e-01 3.98057789e-01
9.14437652e-01 3.12243134e-01 -6.52750432e-01 -2.23215982e-01
-2.39255697e-01 2.94461399e-01 8.50840658e-02 4.38935280e-01
-1.17290401e+00 6.30864203e-01 5.04486752e+00 1.05562389e+00
-1.24597907e+00 2.01976985e-01 7.64234126e-01 -1.76588014e-01
-7.86251798e-02 -1.22027934e-01 -4.06993836e-01 7.86325395e-01
9.22684908e-01 -9.13522691e-02 5.52319229e-01 4.96826082e-01
8.39002192e-01 -2.31640279e-01 -8.31530035e-01 1.09062099e+00
1.67586729e-02 -1.28732991e+00 5.28443232e-02 -5.04026972e-02
6.43318772e-01 -1.28566593e-01 2.55018681e-01 1.62911385e-01
-5.37441432e-01 -7.10076988e-01 8.74308348e-01 7.26302505e-01
8.48909736e-01 -9.59729552e-01 9.15994823e-01 1.37331739e-01
-1.45677686e+00 -1.30982652e-01 -4.51540321e-01 5.07942983e-04
4.67746526e-01 7.46988654e-01 7.11405650e-02 6.22142375e-01
1.07612181e+00 1.05101550e+00 -3.62622529e-01 1.07132185e+00
-1.30537391e-01 4.93834019e-01 1.20922290e-01 5.39511323e-01
2.35543936e-01 -1.23804346e-01 5.56950510e-01 1.15974581e+00
7.18137443e-01 2.71749496e-01 2.14697551e-02 6.87386036e-01
-1.28716245e-01 -5.12670949e-02 -1.05405614e-01 8.97575244e-02
2.21930295e-01 1.21073902e+00 -4.75998342e-01 -7.60303617e-01
-6.04628682e-01 1.12821245e+00 -6.70789182e-02 3.27778399e-01
-1.08619416e+00 -4.00045246e-01 4.90447342e-01 -5.23491809e-03
4.19770896e-01 9.71435905e-02 1.79685310e-01 -1.19066751e+00
1.61889106e-01 -1.08283138e+00 2.61012793e-01 -1.12581432e+00
-9.47307169e-01 6.36449397e-01 -3.26235890e-01 -1.55789840e+00
5.25270309e-03 -1.32726699e-01 -4.76657867e-01 9.00707603e-01
-1.89692211e+00 -8.00882101e-01 -4.33107853e-01 7.19391704e-01
7.35805392e-01 1.01498634e-01 2.76767254e-01 8.11290979e-01
-4.83479530e-01 3.79632831e-01 3.75757441e-02 5.49481176e-02
6.63537621e-01 -6.96183324e-01 1.31456867e-01 1.34601378e+00
-3.95423323e-01 3.42877150e-01 4.23879087e-01 -8.59142363e-01
-1.10235989e+00 -1.27115917e+00 8.19266319e-01 4.60664928e-01
3.89501840e-01 3.70122850e-01 -1.10466766e+00 1.31932706e-01
4.02274072e-01 1.74205139e-01 3.14110339e-01 -7.44013846e-01
-7.66476318e-02 -2.49385893e-01 -1.05556643e+00 3.43231469e-01
8.09958875e-01 -5.53691924e-01 -3.32674682e-01 -2.12322082e-02
8.83677840e-01 -2.07556635e-01 -1.07554078e+00 7.39424706e-01
4.81168658e-01 -1.35335743e+00 9.22146320e-01 -6.15379848e-02
1.08354628e+00 -5.97564638e-01 -3.36595863e-01 -1.11917377e+00
-5.71431518e-01 -3.86012852e-01 -3.70130122e-01 1.12293279e+00
-1.48147672e-01 -4.37366441e-02 5.50798357e-01 4.84574921e-02
-4.63459760e-01 -8.93159568e-01 -7.64585435e-01 -5.24268508e-01
-9.96859670e-02 -4.45228308e-01 5.53540945e-01 5.51840723e-01
-2.97422647e-01 9.67050716e-02 -6.41577125e-01 2.09477499e-01
4.41868097e-01 1.14289351e-01 1.67496413e-01 -4.30293381e-01
-4.35997546e-01 -5.31016886e-01 -2.79136300e-01 -1.27955341e+00
-1.58245385e-01 -7.27623284e-01 5.18104322e-02 -1.65161419e+00
3.83381248e-01 1.16900004e-01 -6.71971798e-01 2.00406030e-01
-5.05816102e-01 4.02967423e-01 6.03316247e-01 1.98007137e-01
-8.74791324e-01 7.68254519e-01 1.66189814e+00 -1.48575678e-01
-6.92814514e-02 -2.04169706e-01 -2.25194603e-01 6.78208411e-01
7.96049953e-01 -3.48445624e-01 -1.57071829e-01 -5.36862075e-01
8.42717588e-02 5.19084632e-01 5.17856181e-01 -1.33794725e+00
1.49939448e-01 4.32794653e-02 7.03033149e-01 -7.39518583e-01
2.22999305e-01 -7.43797004e-01 9.49693099e-02 7.48994470e-01
-2.48488814e-01 2.93342143e-01 6.98256791e-02 2.88224071e-01
-6.67040825e-01 -2.67381400e-01 9.98400033e-01 -2.21286103e-01
-9.51053441e-01 6.15472972e-01 -4.14473504e-01 -1.65900961e-01
8.14183295e-01 -2.54860282e-01 -2.12229952e-01 -4.98791188e-01
-6.28155053e-01 4.67514247e-02 3.60111326e-01 2.99013883e-01
1.24012256e+00 -1.46183705e+00 -6.68464780e-01 1.78821176e-01
-3.20562631e-01 -2.97355115e-01 7.49259591e-01 8.67091835e-01
-5.15864670e-01 1.98190689e-01 -4.03330863e-01 -4.52836186e-01
-1.26281786e+00 8.87239456e-01 5.21843255e-01 -2.68644661e-01
-4.83089119e-01 6.65292799e-01 1.01296693e-01 3.11569810e-01
5.36267124e-02 -2.38291726e-01 -2.67115325e-01 -1.51963338e-01
8.65269840e-01 5.37270069e-01 -3.41053531e-02 -8.35391819e-01
-2.16036022e-01 6.72011316e-01 1.06491186e-01 -6.55354757e-04
1.20612752e+00 -4.72137600e-01 -1.25752196e-01 1.79693177e-01
1.56921697e+00 -2.32565239e-01 -1.57529366e+00 -1.83819935e-01
-3.31562936e-01 -6.25554442e-01 2.86261171e-01 -5.48061252e-01
-1.57470071e+00 9.84582245e-01 1.02771807e+00 1.60800014e-02
1.90072656e+00 -3.48681122e-01 1.44049907e+00 -2.50478536e-01
6.27778918e-02 -9.29281533e-01 3.79997432e-01 5.06514549e-01
7.11805284e-01 -1.15378535e+00 -5.92694096e-02 -7.04595000e-02
-2.83066750e-01 1.36790812e+00 5.44391572e-01 -1.50415346e-01
5.78558445e-01 1.48934543e-01 -1.33453012e-01 7.24790469e-02
-6.37863278e-01 -1.06764890e-01 2.13738963e-01 3.43215674e-01
5.76514125e-01 -2.82863915e-01 -3.84933770e-01 2.92738199e-01
2.45170910e-02 5.33001125e-01 4.41483825e-01 9.13308680e-01
-6.16742015e-01 -8.49116921e-01 -2.33566999e-01 3.40486765e-01
-7.42056370e-01 -4.36514348e-01 3.38335782e-01 3.43930185e-01
5.61470687e-01 1.32158649e+00 1.80216074e-01 -6.70810878e-01
2.73793221e-01 -3.05155098e-01 2.73424596e-01 -8.41494426e-02
-7.02264309e-01 6.43803418e-01 -4.08138484e-01 -9.85989213e-01
-7.00923085e-01 -3.66266072e-01 -1.43403506e+00 -4.35314864e-01
-6.61540329e-02 -7.41678849e-02 3.10347140e-01 7.77465641e-01
3.51418555e-01 7.34932244e-01 7.77957916e-01 -8.42411399e-01
-2.87605166e-01 -8.13656926e-01 -5.55221021e-01 6.48249388e-01
6.62557542e-01 -3.39378297e-01 -2.17828333e-01 2.92140782e-01] | [11.311899185180664, -1.7516025304794312] |
76e1676b-7a61-48fe-9c63-de49384b56a5 | a-category-theory-framework-for-sense-systems | null | null | https://aclanthology.org/2022.gwll-1.7 | https://aclanthology.org/2022.gwll-1.7.pdf | A Category Theory Framework for Sense Systems | Sense repositories are a key component of many NLP applications that require the identification of word senses. Many sense repositories exist: a large proportion is based on lexicographic resources such as WordNet and various dictionaries, but there are others which are the product of clustering algorithms and other automatic techniques. Over the years these repositories have been mapped to each other. However, there have been no attempts (until now) to provide any theoretical grounding for such mappings, causing inconsistencies and unintuitive results. The present paper draws on category theory to formalise assumptions about mapped repositories that are often left implicit, providing formal grounding for this type of language resource. The paper first gives an overview of the word sense disambiguation literature and four types of sense representations: dictionary definitions, clusters of senses, domain labels, and embedding vectors. These different sense representations make different assumptions about the relations and mappings between word senses. We then introduce notation to represent the mappings and repositories as a category, which we call a “sense system”. We represent a sense system as a small category S, where the object set of S, denoted by Ob(S), is a set of sense repositories; and the homomorphism set or hom-set of S, denoted by Hom(S), is a set of mappings between these repositories. On the basis of the sense system description, we propose, formalise, and motivate four basic and two guiding criteria for such sense systems. The four basic criteria are: 1) Correctness preservation: Mappings should preserve the correctness of sense labels in all contexts. Intuitively, if the correct sense for a word token is mapped to another sense, this sense should also be correct for that token. This criterion is endorsed by virtually all existing mappings, but the formalism presented in the paper makes this assumption explicit and allows us to distinguish it from other criteria. 2) Candidacy preservation: Mappings should preserve what we call “the lexical candidacy” of sense labels. Assume that a sense s is mapped to another sense s’ in a different repository. Candidacy preservation then requires that if s is a sense associated with word type w, then so is s’. This criterion is trivially fulfilled by clustering-based approaches, but is not typically explicitly stated for repositories, and we demonstrate how a violation might occur. Our formalisation allows us to specify the difference of this criterion to correctness preservation. As we argue, candidacy preservation allows us to straightforwardly and consistently compare granularity levels by counting the number of senses for each word type. 3) Uniqueness criterion: There should be at most one mapping from one repository to another. This criterion is also fulfilled by clustering-based approaches, but is often violated by repositories that use domain labels. We argue that adhering to the uniqueness criterion provides several benefits, including: a) being able to consistently convert between sets of labels and evaluation metrics, allowing researchers to work with data and models that use different sets of labels; b) ensuring that sense repositories would form a partial preorder, which would roughly correspond to the notion of granularity; and c) ensuring transitivity of mapped senses. 4) Connectivity: A sense system should be a connected category. The connectivity criterion on its own is not very informative, but it enables other criteria by extending their benefits to the rest of the sense system, such as allowing cross-checking between multiple repositories, allowing comparison of grain level, and label conversion. As we argue, connectivity should be considered a formal requirement helping to describe sense repositories and how they relate. We also offer two guiding criteria, which we consider aspirational rather than requirements that have to be strictly fulfilled for all purposes: 1) Non-contradiction: Mappings cannot exist between senses that semantically contradict each other. The non-contradiction criterion forbids mappings between senses whose (strict) implications contradict each other. We demonstrate how such a contradiction might occur, but acknowledge the difficulty in identifying such contradictions. As we argue, the reason to consider this a guiding rather than a strict criterion is that many sense repositories lack the semantic specificity that would allow researchers to identify these contradictions. 2) Inter-annotator agreement: Mappings should correspond to a partial preorder of inter-annotator agreement levels. It has been observed that, when annotating corpora with senses from a given sense repository, inter-annotator agreement tends to drop when the repository is more fine-grained. Therefore, if one repository is coarser-grained than another, one can expect agreement levels to be higher when annotating corpora with senses from the first repository. While this criterion will necessarily be subject to empirical variability (and does apply to sense repositories using non-interpretable representations such as embeddings), we argue that strong violations suggest that the sense distinctions of the coarse-grained sense repository are unnatural, i.e. not in accordance with human linguistic intuitions. Our list is by no means exhaustive, as there are other properties that may be desirable depending on the downstream application. Our category-theory based formalism will serve as the basis for describing any such further properties. However, we also envision that the criteria we have proposed will serve as guidelines for future sense repositories and mappings, in order to avoid the inconsistencies and counterintuitive results derived from existing mappings. | ['Gladys Tyen', 'David Strohmaier'] | null | null | null | null | gwll-lrec-2022-6 | ['word-sense-disambiguation'] | ['natural-language-processing'] | [ 2.07158074e-01 8.34632069e-02 -2.39657640e-01 -1.10728987e-01
-2.47974709e-01 -1.11103034e+00 8.41234982e-01 5.94682455e-01
-5.21071851e-01 5.72461903e-01 4.57382023e-01 -3.61149728e-01
-4.36392784e-01 -9.76621866e-01 -6.36098012e-02 -4.56527770e-01
2.26542741e-01 3.24005693e-01 4.31806356e-01 -5.94202638e-01
3.61470222e-01 2.98589051e-01 -1.77925074e+00 -5.39781246e-03
6.79526567e-01 5.77781558e-01 4.29153264e-01 6.00036681e-02
-7.37450898e-01 5.23438990e-01 -6.25275373e-01 -2.54636109e-01
1.78023681e-01 -5.03185034e-01 -1.33578503e+00 -4.05533165e-02
-9.42914784e-02 6.76734686e-01 3.89924049e-01 1.43734467e+00
4.47747298e-02 -2.25646477e-02 5.73690176e-01 -1.38946342e+00
-8.31159234e-01 7.95506835e-01 -8.27225894e-02 -1.49148274e-02
7.00040400e-01 -4.20549452e-01 1.41998792e+00 -5.59262574e-01
9.27482605e-01 9.35068369e-01 6.68850780e-01 4.50097829e-01
-1.21366429e+00 -3.82250190e-01 2.81899720e-02 -1.03326272e-02
-1.65657222e+00 -1.92493424e-01 4.67881203e-01 -5.59847534e-01
9.30618048e-01 6.80582285e-01 8.85493934e-01 8.02207470e-01
-3.63468975e-02 2.05399483e-01 1.07416201e+00 -9.32244658e-01
4.59528208e-01 4.50406551e-01 6.67515874e-01 1.14110306e-01
8.49812984e-01 -1.04343101e-01 -4.68867838e-01 -4.05092835e-01
5.65896630e-01 -3.13453317e-01 -4.11755383e-01 -4.89636898e-01
-1.22187781e+00 5.97597241e-01 9.40357149e-02 1.21332359e+00
-9.91274491e-02 -3.57849300e-01 4.54286218e-01 5.59522033e-01
2.21432447e-01 9.52602267e-01 -3.41806501e-01 2.83556767e-02
-7.90338874e-01 2.04455554e-01 1.07689023e+00 1.10724545e+00
8.25307727e-01 -3.69359136e-01 4.04053003e-01 8.95966053e-01
4.92343217e-01 1.05520487e-01 8.20441842e-01 -7.17666686e-01
-2.01050729e-01 9.39505398e-01 3.81343395e-01 -1.14633107e+00
-4.46550786e-01 -3.31223369e-01 -6.27838850e-01 -1.78452492e-01
2.68760264e-01 4.53458369e-01 -4.26234573e-01 2.05077791e+00
1.82361051e-01 3.63924392e-02 2.79258013e-01 6.86686277e-01
5.61134696e-01 2.34598577e-01 1.84518158e-01 -4.96982694e-01
1.60534763e+00 -4.25036624e-02 -6.57544017e-01 -1.64060354e-01
7.94854343e-01 -8.76061678e-01 1.19185162e+00 1.36122584e-01
-8.34102273e-01 -1.14629418e-01 -1.26049888e+00 1.00461751e-01
-9.38972056e-01 -4.47664857e-01 6.33393466e-01 7.48875499e-01
-1.31829500e+00 3.63061309e-01 -4.14014190e-01 -9.76864696e-01
-4.24239814e-01 8.63800347e-02 -3.59704554e-01 3.17798316e-01
-1.66295230e+00 1.17679667e+00 6.54973030e-01 -3.03660125e-01
-8.53896737e-02 -4.83296633e-01 -8.58306289e-01 3.12210303e-02
3.97673070e-01 -7.46621728e-01 9.64851022e-01 -1.25202882e+00
-6.23270154e-01 1.41518021e+00 -1.42247528e-01 -1.71656221e-01
-1.17851654e-02 2.56500095e-01 -1.13106990e+00 -1.80902362e-01
7.97033787e-01 2.64437318e-01 1.48435950e-01 -1.52231085e+00
-8.59886527e-01 -2.59919405e-01 4.49732304e-01 4.94454661e-03
-4.63777304e-01 4.20066595e-01 -2.74152517e-01 -7.04310775e-01
4.14208829e-01 -7.01886356e-01 1.52181135e-03 -4.21503305e-01
-3.34571958e-01 -5.41738868e-01 2.41809160e-01 1.26734942e-01
1.71017885e+00 -2.14170218e+00 -1.92049906e-01 5.28654218e-01
2.59952307e-01 -3.92787308e-02 -3.58682573e-02 8.04667115e-01
-6.29259109e-01 6.46399081e-01 -4.45450723e-01 1.00198492e-01
4.66177732e-01 5.28623521e-01 -6.12251699e-01 4.47272420e-01
-1.93999320e-01 3.16832632e-01 -1.22615862e+00 -4.20264006e-01
1.49128944e-01 3.86107326e-01 -3.74799550e-01 -2.64295101e-01
-5.71318604e-02 -1.76962927e-01 -4.04866040e-01 2.97939688e-01
3.56455505e-01 -2.92876065e-01 7.61465788e-01 -1.81343541e-01
-5.12778759e-01 6.35298252e-01 -1.69188249e+00 1.42448401e+00
-3.18789959e-01 1.25085205e-01 -1.66903809e-02 -8.05764675e-01
9.32730794e-01 4.37047511e-01 7.25339234e-01 -3.16018105e-01
-1.09182119e-01 5.48699737e-01 -1.21565536e-05 -2.01730475e-01
9.40319300e-01 -5.63110828e-01 -4.64067638e-01 6.71245277e-01
-3.59856039e-02 -3.81073542e-03 3.76516908e-01 3.80845547e-01
1.08434892e+00 -1.80577457e-01 9.64103997e-01 -9.14833188e-01
5.87307215e-01 3.43276680e-01 8.60014796e-01 7.35046566e-01
-9.34214890e-02 2.37476885e-01 3.02740723e-01 -2.33739838e-01
-1.03185987e+00 -1.16227162e+00 -4.92021173e-01 9.98211503e-01
5.94487667e-01 -1.05729949e+00 -5.68187714e-01 -3.02824199e-01
-7.24508464e-02 7.01187968e-01 -5.62893689e-01 1.38902724e-01
-1.48742616e-01 -4.19336051e-01 6.47887766e-01 3.15641880e-01
1.33291453e-01 -1.06687450e+00 -8.22966516e-01 2.41214424e-01
-8.12068284e-02 -6.34950161e-01 -2.78944314e-01 1.45435736e-01
-2.32269555e-01 -1.36761236e+00 -8.18105787e-02 -6.32312655e-01
4.86326694e-01 2.15729803e-01 1.44983315e+00 3.18102747e-01
1.40798464e-01 6.23118281e-01 -7.68956721e-01 -2.65518129e-01
-4.75518286e-01 -9.41646993e-02 6.01889014e-01 -2.16052935e-01
7.60967493e-01 -6.13240182e-01 -1.46702439e-01 3.03415298e-01
-1.44242620e+00 -4.70489055e-01 1.61213987e-02 5.46422899e-01
7.79408455e-01 1.20619543e-01 5.04359901e-01 -1.03192353e+00
8.89300525e-01 -5.66856742e-01 -5.05342901e-01 5.38050711e-01
-9.17816997e-01 1.37063384e-01 3.29075038e-01 -1.65468574e-01
-5.77661514e-01 -2.72479743e-01 -1.16713606e-01 1.42492801e-01
-3.13344479e-01 8.78030300e-01 -5.27222157e-01 3.09622794e-01
6.89937770e-01 1.37573659e-01 -3.39864999e-01 -4.22518551e-01
5.83660901e-01 8.62864852e-01 5.38068950e-01 -9.29064751e-01
5.38420439e-01 4.54756856e-01 -3.88087630e-01 -8.02123606e-01
-7.67699480e-01 -8.57828438e-01 -6.43300831e-01 2.07937602e-03
8.58440042e-01 -5.58343410e-01 -4.57090020e-01 -2.39256382e-01
-1.01032948e+00 1.94261760e-01 -5.12819648e-01 1.51918635e-01
-5.89619517e-01 5.53967297e-01 -4.03472707e-02 -7.97881961e-01
-6.37674406e-02 -8.66858125e-01 6.23968840e-01 -2.02794909e-01
-1.14194524e+00 -1.30457091e+00 1.65318072e-01 -2.76207387e-01
3.28348905e-01 1.68236822e-01 1.15596569e+00 -1.07415509e+00
1.70176327e-01 7.45169520e-02 2.00866312e-01 -8.70952196e-03
5.24196804e-01 -1.95739195e-01 -6.41302824e-01 -2.37004489e-01
2.54381806e-01 2.80623734e-01 4.11112249e-01 -8.52512568e-02
4.63967800e-01 -4.66479748e-01 -6.50142491e-01 5.80837913e-02
1.81123066e+00 2.36523315e-01 4.12149668e-01 7.70016432e-01
4.10792083e-01 7.54922986e-01 4.77056086e-01 9.72479880e-02
4.25968319e-01 7.19418705e-01 -4.39468920e-02 1.12455338e-01
1.04781732e-01 -3.45816724e-02 1.91926628e-01 1.13014543e+00
1.61052719e-01 -5.46832755e-02 -1.23459053e+00 1.07095730e+00
-1.76833797e+00 -1.05281174e+00 -3.00103307e-01 2.38294411e+00
1.10149217e+00 3.73035073e-02 2.33737707e-01 2.16937169e-01
9.19133484e-01 -1.17377341e-02 -1.31095489e-02 -4.05962646e-01
-2.60656625e-01 2.45245397e-01 3.87625128e-01 8.80287886e-01
-7.03724742e-01 8.69600713e-01 6.55255604e+00 5.38860261e-01
-8.36893678e-01 3.42527986e-01 -2.84683228e-01 2.52332598e-01
-9.72238183e-01 4.60517049e-01 -4.99296278e-01 5.11026859e-01
5.42684317e-01 -6.77838445e-01 1.78381026e-01 6.47001624e-01
-1.36186153e-01 5.67101836e-02 -1.27936959e+00 8.34036112e-01
-3.69025767e-03 -1.31193638e+00 1.44868568e-01 1.60204530e-01
3.79009187e-01 -5.79776764e-02 -4.21168447e-01 -4.86134477e-02
7.42524862e-01 -8.26993525e-01 1.02162123e+00 4.17486906e-01
9.39194083e-01 -7.29958355e-01 5.69607496e-01 1.53640896e-01
-1.39643669e+00 2.20058739e-01 -4.08647388e-01 -2.04847261e-01
8.96741450e-02 7.37050831e-01 -1.52042627e-01 9.10258234e-01
6.20682299e-01 6.16887450e-01 -1.23381004e-01 8.13724756e-01
-1.54742897e-01 3.71762127e-01 -2.41771832e-01 7.25302398e-02
1.48515850e-01 -3.13841641e-01 8.43695283e-01 1.33259392e+00
3.44766945e-01 2.58215338e-01 3.16543877e-01 1.09732723e+00
3.27831119e-01 1.44834161e-01 -7.77895093e-01 -1.84485689e-05
1.14406204e+00 9.25268650e-01 -8.94323945e-01 -4.53793347e-01
-4.06916976e-01 5.13082862e-01 -1.83678016e-01 2.50804305e-01
-3.32524300e-01 -4.45981115e-01 1.23122180e+00 1.90803915e-01
-1.24794647e-01 -7.51736909e-02 -4.72277522e-01 -1.22140098e+00
1.20270349e-01 -7.17992187e-01 6.49802327e-01 -5.83850384e-01
-1.65533829e+00 6.18055046e-01 3.17280352e-01 -1.33491588e+00
-1.79519415e-01 -5.94654083e-01 -3.13714474e-01 9.58933830e-01
-1.14972341e+00 -8.30400586e-01 1.80211812e-01 6.45659983e-01
-7.57932663e-02 -2.59262137e-03 1.26696157e+00 4.85750549e-02
-4.81997058e-02 4.28039223e-01 -2.13853180e-01 1.44806936e-01
7.12748826e-01 -1.43132198e+00 1.05135381e-01 9.50890303e-01
3.03978920e-01 1.46934891e+00 1.02877140e+00 -7.30163693e-01
-1.02511430e+00 -9.42258596e-01 1.62177050e+00 -6.63940549e-01
1.10624111e+00 -1.64302349e-01 -1.00870883e+00 7.88046300e-01
3.87124300e-01 -4.83226925e-01 1.11016214e+00 4.96524632e-01
-7.09932208e-01 -3.16867754e-02 -1.05034637e+00 5.98302007e-01
1.16372371e+00 -7.50407219e-01 -1.31061912e+00 2.32680380e-01
6.03812099e-01 1.49982244e-01 -1.17632580e+00 1.49096131e-01
3.31665009e-01 -9.13288653e-01 8.10034037e-01 -5.31838894e-01
-8.08116496e-02 -9.03091431e-01 -5.20086169e-01 -1.23774374e+00
-5.10095417e-01 -4.49600756e-01 5.61315715e-01 1.62716317e+00
4.30252016e-01 -1.00421333e+00 7.07850233e-02 6.21787548e-01
-1.34483874e-01 -1.86348766e-01 -1.01926422e+00 -1.15001476e+00
4.55308497e-01 -7.98203170e-01 1.19907272e+00 1.86067879e+00
9.20183778e-01 4.94398117e-01 2.41902769e-01 3.11990604e-02
4.57785517e-01 1.18925586e-01 1.16912812e-01 -1.59613419e+00
-6.24960549e-02 -8.33924949e-01 -5.59799790e-01 -4.69781816e-01
3.43883902e-01 -1.30634105e+00 -1.14928037e-01 -1.63994539e+00
3.47958505e-02 -8.49612236e-01 -4.93843853e-01 8.40257347e-01
1.24446258e-01 3.20800506e-02 1.08760580e-01 6.34593606e-01
-3.70271862e-01 -1.84388816e-01 4.80341733e-01 -7.12727476e-03
-8.41400325e-02 -6.14733815e-01 -1.34410763e+00 8.07323098e-01
6.73807085e-01 -5.34760535e-01 -4.22990680e-01 -2.31611967e-01
1.01154137e+00 -3.72426927e-01 3.38124931e-01 -6.61177576e-01
2.73229361e-01 -3.72773826e-01 -2.43217751e-01 9.37269852e-02
-1.98580980e-01 -9.40146685e-01 6.90237880e-01 2.26872340e-01
-2.94377238e-01 1.78060874e-01 3.27421427e-02 1.68913841e-01
-3.32040727e-01 -3.51686269e-01 5.47063410e-01 -2.92813659e-01
-1.14952660e+00 -3.46281201e-01 -4.11037713e-01 2.04639323e-02
8.40707898e-01 -5.56581438e-01 -1.13938048e-01 3.58432829e-02
-7.59302735e-01 6.08113175e-03 1.17387986e+00 5.67455411e-01
2.90833354e-01 -1.64569497e+00 -4.68742043e-01 -5.48663735e-03
6.73545361e-01 -4.53570306e-01 -4.64855045e-01 5.79153955e-01
1.12637626e-02 2.62798160e-01 -2.75217649e-03 -3.88805091e-01
-1.02036071e+00 6.66594267e-01 2.93960273e-01 2.35409196e-02
-6.06719136e-01 4.92722571e-01 1.24408200e-01 -4.37046021e-01
-2.96235353e-01 -4.34667587e-01 -3.00741971e-01 1.10151440e-01
6.09439850e-01 -6.07887059e-02 1.34858102e-01 -1.00603712e+00
-8.50080609e-01 4.97382641e-01 2.70077944e-01 -3.23976457e-01
1.10135603e+00 -1.61485538e-01 -8.44658256e-01 7.43325233e-01
7.34003067e-01 3.09502721e-01 -1.05665669e-01 -7.25730434e-02
5.24515927e-01 -2.20956147e-01 -4.31487799e-01 -8.60558510e-01
-5.19991815e-01 8.99343193e-02 2.87453890e-01 8.94905210e-01
1.09247983e+00 3.61537695e-01 3.07100952e-01 -1.46575905e-02
7.29668856e-01 -1.05833280e+00 -6.97499752e-01 8.33649635e-01
8.34724307e-01 -3.92906427e-01 -3.47068459e-01 -5.33053815e-01
-4.76662636e-01 8.02709818e-01 1.00982994e-01 7.96848908e-03
6.47845447e-01 3.98131669e-01 2.04177052e-01 -4.11281049e-01
-4.20368880e-01 -5.17306089e-01 1.43341064e-01 6.93063378e-01
6.98716521e-01 2.97151893e-01 -9.93510127e-01 8.98468256e-01
-6.79219723e-01 -9.28333476e-02 5.23545146e-01 8.05715442e-01
-6.73510313e-01 -1.58687294e+00 -5.57623386e-01 1.68860167e-01
-3.17989647e-01 -3.80216002e-01 -8.43760908e-01 9.49421167e-01
5.61268806e-01 1.23271239e+00 2.63816804e-01 -5.34889638e-01
3.11706364e-01 1.07498653e-01 3.43361557e-01 -1.05131269e+00
-6.66068137e-01 -1.74717873e-01 1.69307828e-01 -3.48127723e-01
-8.18814874e-01 -5.86141288e-01 -1.45962143e+00 -3.51482570e-01
-3.55590284e-01 6.84496760e-01 1.62432224e-01 1.26803780e+00
2.78567106e-01 1.05264679e-01 4.40951735e-02 -6.93532079e-03
-1.89777941e-01 -7.70320058e-01 -9.92540896e-01 7.11304605e-01
4.36572321e-02 -8.47895682e-01 -2.73003876e-01 2.19003394e-01] | [10.231735229492188, 9.141035079956055] |
18671bf7-86ed-419a-9e66-755065d9d4bb | low-rank-matrix-completion-via-robust | 2302.11068 | null | https://arxiv.org/abs/2302.11068v1 | https://arxiv.org/pdf/2302.11068v1.pdf | Low Rank Matrix Completion via Robust Alternating Minimization in Nearly Linear Time | Given a matrix $M\in \mathbb{R}^{m\times n}$, the low rank matrix completion problem asks us to find a rank-$k$ approximation of $M$ as $UV^\top$ for $U\in \mathbb{R}^{m\times k}$ and $V\in \mathbb{R}^{n\times k}$ by only observing a few entries masked by a binary matrix $P_{\Omega}\in \{0, 1 \}^{m\times n}$. As a particular instance of the weighted low rank approximation problem, solving low rank matrix completion is known to be computationally hard even to find an approximate solution [RSW16]. However, due to its practical importance, many heuristics have been proposed for this problem. In the seminal work of Jain, Netrapalli, and Sanghavi [JNS13], they show that the alternating minimization framework provides provable guarantees for low rank matrix completion problem whenever $M$ admits an incoherent low rank factorization. Unfortunately, their algorithm requires solving two exact multiple response regressions per iteration and their analysis is non-robust as they exploit the structure of the exact solution. In this paper, we take a major step towards a more efficient and robust alternating minimization framework for low rank matrix completion. Our main result is a robust alternating minimization algorithm that can tolerate moderate errors even though the regressions are solved approximately. Consequently, we also significantly improve the running time of [JNS13] from $\widetilde{O}(mnk^2 )$ to $\widetilde{O}(mnk )$ which is nearly linear in the problem size, as verifying the low rank approximation takes $O(mnk)$ time. Our core algorithmic building block is a high accuracy regression solver that solves the regression in nearly linear time per iteration. | ['Lichen Zhang', 'Junze Yin', 'Zhao Song', 'Yuzhou Gu'] | 2023-02-21 | null | null | null | null | ['low-rank-matrix-completion', 'matrix-completion'] | ['methodology', 'methodology'] | [ 3.51565868e-01 1.92244679e-01 -9.40214545e-02 -4.41728681e-02
-1.32685566e+00 -7.79464364e-01 -2.48264670e-01 5.02159446e-02
-5.00689745e-01 6.96831286e-01 -6.05193302e-02 -6.76322699e-01
-7.94728398e-01 -8.09100866e-01 -9.30612504e-01 -7.77669311e-01
-6.16101444e-01 5.32863200e-01 -2.42652491e-01 -5.33379257e-01
1.72060430e-01 1.86054125e-01 -1.17822957e+00 2.89623201e-01
6.35258853e-01 1.14482975e+00 -1.34231314e-01 9.64877307e-01
1.37013286e-01 6.68608904e-01 -9.07931179e-02 -4.44805056e-01
1.00929189e+00 -3.24205607e-01 -1.16446531e+00 8.36577192e-02
3.66213948e-01 -1.20418213e-01 -6.27564728e-01 1.27666938e+00
2.03583091e-01 3.87717456e-01 2.83711523e-01 -1.10570073e+00
-1.15592986e-01 9.13087964e-01 -1.14626873e+00 -1.67787477e-01
5.55818260e-01 -1.48150280e-01 1.43148243e+00 -1.05135262e+00
4.39458311e-01 9.01445389e-01 4.79768932e-01 7.55959898e-02
-1.63729572e+00 -1.09511423e+00 1.38684794e-01 -9.73018706e-02
-1.73864996e+00 -4.27697986e-01 4.66270536e-01 -2.65174657e-01
5.97590923e-01 8.88743103e-01 1.60677448e-01 2.05333337e-01
-5.28958380e-01 5.29732645e-01 1.61199129e+00 -4.86544371e-01
3.57773080e-02 -2.13371918e-01 5.39210439e-01 8.00467968e-01
6.11562431e-01 -5.60965240e-02 -6.21143043e-01 -4.90820765e-01
6.15579128e-01 -1.64741259e-02 -3.33424389e-01 -8.83340836e-03
-9.93993700e-01 9.55658257e-01 2.90090233e-01 1.37454316e-01
-5.12627847e-02 4.71674204e-01 1.14032507e-01 7.03519702e-01
2.88508795e-02 3.58005702e-01 -5.24622738e-01 -4.33805920e-02
-1.06887817e+00 4.10849005e-01 8.85339677e-01 9.46736097e-01
1.13195777e+00 -3.32953073e-02 1.35270402e-01 6.04569435e-01
-2.56926063e-02 8.05616081e-01 -3.70432585e-01 -1.42859721e+00
8.81331384e-01 3.74918193e-01 2.59220302e-01 -1.30988216e+00
-4.35982376e-01 -3.33304167e-01 -1.23995721e+00 1.22339830e-01
7.41209567e-01 -5.21085374e-02 -3.76740128e-01 1.91869509e+00
1.27689287e-01 -1.09561041e-01 -1.12513587e-01 9.09006476e-01
1.82706505e-01 7.90647388e-01 -3.52601647e-01 -7.58622646e-01
1.26469886e+00 -6.31506801e-01 -3.37585211e-01 -4.09002155e-01
8.57494652e-01 -9.25998867e-01 9.41430748e-01 8.05545449e-01
-1.55416000e+00 -4.63829283e-03 -9.26360309e-01 2.46830732e-02
5.66719770e-02 -7.34997541e-02 1.13102520e+00 8.63904297e-01
-1.07351661e+00 4.54860866e-01 -3.21083218e-01 3.51171166e-01
4.42991890e-02 1.04078627e+00 -6.52750432e-01 -6.99195564e-01
-9.70969677e-01 2.94322908e-01 -1.85334727e-01 2.41798520e-01
-5.45635700e-01 -5.85078061e-01 -6.70177042e-01 -2.58535624e-01
8.63565266e-01 -2.43982702e-01 8.08504879e-01 -5.72992504e-01
-8.30786288e-01 9.47949886e-01 -4.27388906e-01 -2.13889167e-01
3.54639202e-01 6.10647425e-02 -8.32374319e-02 1.09243080e-01
1.05460174e-01 -2.40796626e-01 7.28901923e-01 -1.09748924e+00
-5.18539608e-01 -7.77064085e-01 4.51783687e-01 -1.11238033e-01
-2.05849290e-01 1.89559549e-01 -4.82363403e-01 -5.28171182e-01
8.62505436e-01 -1.10035276e+00 -9.54943359e-01 -5.01078486e-01
-2.53533810e-01 1.37127146e-01 -1.45416245e-01 -6.57837510e-01
1.61055028e+00 -2.12281990e+00 5.08386552e-01 1.08548534e+00
7.12391853e-01 -1.60498768e-01 -1.83750510e-01 5.88185310e-01
-3.13104451e-01 2.24029168e-01 -3.12403142e-01 -1.79848090e-01
4.38734740e-02 2.72705555e-01 -4.32419449e-01 8.18550348e-01
-7.15666890e-01 3.74815196e-01 -6.44970953e-01 4.69628675e-03
-2.27402970e-01 7.99311846e-02 -8.23669136e-01 -9.06570032e-02
7.23284855e-02 2.54206136e-02 -1.74984187e-01 4.75946754e-01
1.18405819e+00 -4.45543587e-01 5.95095932e-01 -2.45387942e-01
2.49284282e-02 -1.37580074e-02 -2.13360357e+00 1.44350851e+00
-3.65566462e-01 9.07228887e-02 1.02217531e+00 -1.34580493e+00
3.85989040e-01 1.86229289e-01 9.05364454e-01 -6.42263770e-01
1.00124530e-01 5.66604972e-01 -2.43271053e-01 4.09793817e-02
6.95272803e-01 -3.09165478e-01 -4.75552469e-01 7.29866743e-01
-5.77455401e-01 6.77821860e-02 4.37099904e-01 7.73131073e-01
1.62983000e+00 -5.81618845e-01 -1.93058774e-01 -2.34837309e-01
5.37872553e-01 -7.26740211e-02 7.10238159e-01 1.15288210e+00
2.49501809e-01 3.35453868e-01 8.75696719e-01 -2.27668390e-01
-8.28421354e-01 -9.23632562e-01 4.41844352e-02 1.37972403e+00
-1.91164464e-02 -9.16999161e-01 -5.50953209e-01 -2.88078725e-01
-2.09755704e-01 9.58596170e-02 -5.81832945e-01 1.54264584e-01
-6.89566076e-01 -1.00018632e+00 5.08423865e-01 3.66794169e-01
1.63551718e-01 -2.17324674e-01 1.00604884e-01 1.91038072e-01
-3.23654860e-01 -1.13304484e+00 -6.31726742e-01 5.65768063e-01
-9.62789774e-01 -1.22550619e+00 -3.71294379e-01 -4.69058424e-01
1.15194666e+00 4.55337912e-01 8.39892864e-01 2.48757571e-01
-3.58629018e-01 3.24042827e-01 -2.59656221e-01 1.81303710e-01
9.41584259e-02 -2.59098709e-01 4.23715085e-01 -3.04098967e-02
1.90211739e-02 -7.56818414e-01 -7.07141757e-01 3.89103293e-01
-1.20349383e+00 -2.61881799e-01 6.11288249e-01 7.99553514e-01
9.01689053e-01 3.81787121e-01 6.41768472e-03 -1.29328918e+00
3.73629808e-01 -2.78145909e-01 -9.64760363e-01 -6.59915432e-02
-7.67513454e-01 2.29166090e-01 8.85660350e-01 -2.46924177e-01
-2.43540138e-01 3.89407784e-01 2.40157824e-02 -3.14263582e-01
6.25238717e-01 7.03213513e-01 7.78769469e-03 -1.82917684e-01
8.32348704e-01 2.70594448e-01 -2.01008901e-01 -3.89909625e-01
6.01801991e-01 3.03747028e-01 4.69263554e-01 -1.06253374e+00
1.31579590e+00 6.36565626e-01 5.66514194e-01 -5.26937783e-01
-9.31935310e-01 -5.25646508e-01 -4.79344279e-02 2.17967674e-01
7.98247233e-02 -1.03709257e+00 -1.42740297e+00 -1.15804560e-01
-4.46828425e-01 -4.26003724e-01 -2.19426915e-01 3.35896850e-01
-6.51291788e-01 7.18231916e-01 -5.48884094e-01 -9.75757122e-01
-1.31771296e-01 -1.02397776e+00 5.80476165e-01 -5.07072866e-01
-1.58612832e-01 -3.93901169e-01 -1.11663528e-01 8.81976962e-01
2.22820655e-01 6.67186603e-02 9.54132140e-01 -2.09418491e-01
-8.44477296e-01 -5.81365824e-01 -5.99896610e-01 2.82726198e-01
-2.54499674e-01 -6.43057525e-01 -4.64935035e-01 -6.72263324e-01
4.24938053e-02 -3.33881229e-01 7.82675266e-01 1.77177250e-01
1.22564065e+00 -9.01092768e-01 5.91887021e-03 6.74472392e-01
1.62696791e+00 -5.27205527e-01 6.94809437e-01 -1.06213773e-02
5.45033574e-01 3.09848607e-01 6.41169965e-01 7.97878146e-01
1.99902460e-01 4.72736359e-01 4.06531811e-01 7.93527439e-02
5.13779938e-01 2.67706811e-01 4.63164747e-01 9.27401543e-01
-3.80788445e-01 3.97702605e-01 -5.14702082e-01 4.55078423e-01
-1.84473860e+00 -8.96866024e-01 -5.92106521e-01 2.73989081e+00
1.01845837e+00 3.03023960e-02 1.89379498e-01 5.29105604e-01
3.69380444e-01 4.49776463e-02 4.08518650e-02 -5.94999373e-01
-3.34600449e-01 1.02560329e+00 1.13622439e+00 8.64578724e-01
-5.94959259e-01 8.14074099e-01 5.36550808e+00 1.08869851e+00
-4.87578720e-01 7.97738954e-02 5.38672209e-01 -6.08625889e-01
-4.45355326e-01 4.94967908e-01 -6.29114866e-01 1.98829934e-01
8.65868807e-01 5.87490425e-02 1.17811251e+00 8.72332156e-01
1.70730054e-01 -2.98250705e-01 -1.17891371e+00 1.37533343e+00
8.80376610e-04 -1.15874374e+00 -3.91707331e-01 5.60974360e-01
9.85656917e-01 -4.03005958e-01 3.21522146e-01 2.63890624e-01
7.16274142e-01 -1.38607252e+00 2.55762249e-01 -6.25651032e-02
1.06830966e+00 -9.87237453e-01 2.64648676e-01 3.71123493e-01
-1.52936018e+00 -3.68284017e-01 -5.71986794e-01 -4.43884850e-01
2.85390280e-02 9.13579524e-01 -1.01104878e-01 5.82409620e-01
7.78927267e-01 -9.93693471e-02 -4.18809392e-02 4.50901657e-01
1.15797035e-01 4.91404861e-01 -7.64703691e-01 4.95780081e-01
2.18774974e-01 -4.35056210e-01 3.69459629e-01 9.35704648e-01
2.54409432e-01 9.02083874e-01 6.82617664e-01 2.04667985e-01
-3.85991663e-01 3.83568019e-01 -4.27382946e-01 -4.32627229e-03
1.37671918e-01 1.17798078e+00 -6.68708801e-01 -5.77463545e-02
-2.58200228e-01 8.10536981e-01 3.97650391e-01 3.67117047e-01
-7.51083016e-01 -3.37080240e-01 8.32581162e-01 3.69540840e-01
3.67981881e-01 -6.48671269e-01 -5.95448792e-01 -1.03591621e+00
2.82961279e-01 -1.28236687e+00 8.13835979e-01 -1.75827950e-01
-1.00729418e+00 2.02107757e-01 -3.32556814e-01 -9.12787080e-01
7.89303705e-02 -7.05842316e-01 2.33635709e-01 1.03724122e+00
-1.00074863e+00 -6.00753367e-01 8.33325163e-02 1.01033115e+00
-1.42015561e-01 2.62150258e-01 8.01992476e-01 4.50358242e-01
-4.30015773e-01 9.01141107e-01 3.11707973e-01 3.58002484e-02
6.06124461e-01 -1.03043449e+00 -4.65437651e-01 9.78102267e-01
-1.62403919e-02 1.02181685e+00 8.58966529e-01 -4.10794497e-01
-2.29900408e+00 -6.45079613e-01 8.46053123e-01 -3.53886694e-01
7.24277139e-01 -2.63794988e-01 -4.61099893e-01 6.94230080e-01
-3.36368144e-01 1.12916730e-01 9.75560546e-01 2.28208095e-01
-7.69134581e-01 -4.72569942e-01 -9.16843414e-01 7.93245316e-01
1.15429723e+00 -7.28741586e-01 9.50554088e-02 5.94245017e-01
4.08906758e-01 -5.72729588e-01 -1.09150612e+00 4.65988278e-01
2.88340747e-01 -7.45990396e-01 1.16288292e+00 -5.39436460e-01
2.63169557e-01 -4.32452708e-01 -8.87113988e-01 -4.07447994e-01
-3.40442002e-01 -1.22388434e+00 -3.46676290e-01 6.02812171e-01
5.70399404e-01 -3.04152906e-01 9.91332829e-01 9.89201188e-01
2.74027795e-01 -7.02169478e-01 -1.09966397e+00 -5.50324738e-01
6.94033429e-02 -1.02459061e+00 1.07785769e-01 8.13387156e-01
1.21788315e-01 9.93243083e-02 -7.81246066e-01 4.09651041e-01
7.92945862e-01 2.57044882e-01 1.02948463e+00 -9.93702590e-01
-9.37557936e-01 -2.63376206e-01 -3.47054144e-03 -1.13685703e+00
-8.18870068e-02 -1.07427084e+00 -2.90935904e-01 -1.25871634e+00
4.27260220e-01 -9.96463656e-01 -3.54109794e-01 5.83961546e-01
7.09443316e-02 6.13913417e-01 2.38949820e-01 4.29269038e-02
-8.33766937e-01 -2.26077080e-01 8.84925008e-01 -2.43896157e-01
-1.52578682e-01 7.26032630e-02 -1.22649133e+00 5.27374506e-01
2.69157976e-01 -5.18035471e-01 -1.74559832e-01 -3.77549082e-01
1.36117530e+00 7.11719215e-01 6.48919195e-02 -5.86481154e-01
3.78656775e-01 -2.56876975e-01 -8.40984508e-02 -3.92933637e-01
5.55451572e-01 -8.66530359e-01 3.05833489e-01 5.06958723e-01
-2.03352883e-01 2.07190886e-01 -4.57371399e-02 5.97514272e-01
2.23500490e-01 -2.95372576e-01 6.56828701e-01 -2.44483352e-01
-1.74092531e-01 5.27303636e-01 -2.78796107e-01 2.30308712e-01
7.59857714e-01 -2.15099707e-01 -1.81000367e-01 -6.53192341e-01
-8.65320444e-01 2.11746067e-01 3.67155820e-01 -3.89340103e-01
4.73533988e-01 -9.86146569e-01 -7.09889114e-01 -8.12428072e-02
-1.18038744e-01 1.40844166e-01 6.10176444e-01 1.16845667e+00
-4.09949809e-01 1.49505183e-01 3.92427921e-01 -2.80700415e-01
-1.22001600e+00 6.18664086e-01 -2.00767130e-01 -7.57263958e-01
-1.52283534e-02 1.11323857e+00 -1.33726880e-01 -1.63653523e-01
1.94618732e-01 1.11907646e-02 6.02556109e-01 -1.73016623e-01
7.16533780e-01 7.64285266e-01 1.52887441e-02 -5.55520833e-01
-2.34053403e-01 5.87815940e-01 -2.17874333e-01 -4.68152076e-01
1.36891127e+00 -2.10198820e-01 -9.03586626e-01 -7.35999867e-02
1.32247484e+00 6.82717323e-01 -6.29305482e-01 -3.52899790e-01
-1.97441503e-01 -6.32956386e-01 -3.85426700e-01 -3.21545243e-01
-1.18602276e+00 5.06911397e-01 3.00669253e-01 8.28227550e-02
1.38161957e+00 -9.66478363e-02 7.16296256e-01 8.20039451e-01
7.84031808e-01 -1.30079746e+00 8.30226857e-03 5.97066343e-01
5.43034732e-01 -9.57861483e-01 4.53042984e-01 -6.30553722e-01
-2.11657777e-01 8.83754969e-01 2.16972366e-01 -2.48042509e-01
5.80655158e-01 2.34057888e-01 -3.48626733e-01 -1.68962643e-01
-4.94667768e-01 -1.46441713e-01 -6.66192174e-02 3.75024565e-02
2.91255713e-01 1.89771265e-01 -6.90199792e-01 8.94769669e-01
-3.58940423e-01 -2.92399496e-01 7.36604393e-01 7.98618913e-01
-4.40409899e-01 -1.74750030e+00 -7.05770075e-01 7.15746224e-01
-7.57907987e-01 -2.33473152e-01 -3.43695208e-02 3.41923118e-01
-1.07033044e-01 1.26356328e+00 -6.37510478e-01 -4.32674199e-01
1.85914576e-01 -7.72727653e-02 5.90088785e-01 -4.46486771e-01
-6.56377494e-01 2.04144448e-01 -2.72677932e-02 -8.00909579e-01
2.68489346e-02 -5.09628892e-01 -1.51251149e+00 -8.77287507e-01
-1.25366256e-01 3.57455701e-01 5.09917617e-01 8.64901066e-01
8.59895255e-03 -6.84200227e-02 1.10231614e+00 -2.62169719e-01
-7.18018115e-01 -4.25554633e-01 -9.39493895e-01 3.42209786e-01
1.49632590e-02 -8.68707374e-02 -5.30042529e-01 -2.86479414e-01] | [6.617696762084961, 4.718226909637451] |
abe0a40e-7ff8-4fe9-b7b6-65b0fbe952b4 | systematic-generalization-what-is-required | 1811.12889 | null | http://arxiv.org/abs/1811.12889v3 | http://arxiv.org/pdf/1811.12889v3.pdf | Systematic Generalization: What Is Required and Can It Be Learned? | Numerous models for grounded language understanding have been recently
proposed, including (i) generic models that can be easily adapted to any given
task and (ii) intuitively appealing modular models that require background
knowledge to be instantiated. We compare both types of models in how much they
lend themselves to a particular form of systematic generalization. Using a
synthetic VQA test, we evaluate which models are capable of reasoning about all
possible object pairs after training on only a small subset of them. Our
findings show that the generalization of modular models is much more systematic
and that it is highly sensitive to the module layout, i.e. to how exactly the
modules are connected. We furthermore investigate if modular models that
generalize well could be made more end-to-end by learning their layout and
parametrization. We find that end-to-end methods from prior work often learn
inappropriate layouts or parametrizations that do not facilitate systematic
generalization. Our results suggest that, in addition to modularity, systematic
generalization in language understanding may require explicit regularizers or
priors. | ['Michael Noukhovitch', 'Thien Huu Nguyen', 'Harm de Vries', 'Shikhar Murty', 'Dzmitry Bahdanau', 'Aaron Courville'] | 2018-11-30 | systematic-generalization-what-is-required-1 | https://openreview.net/forum?id=HkezXnA9YX | https://openreview.net/pdf?id=HkezXnA9YX | iclr-2019-5 | ['systematic-generalization'] | ['reasoning'] | [ 6.42347187e-02 6.13645554e-01 3.66368741e-02 -4.85616833e-01
-4.67504859e-01 -1.00209630e+00 6.31204486e-01 2.42213264e-01
-1.95233822e-01 4.64698762e-01 3.03924918e-01 -4.56375182e-01
-1.94476202e-01 -8.90745461e-01 -9.55915868e-01 -1.78541183e-01
-1.22681327e-01 8.85290265e-01 2.30065659e-01 -5.32178283e-01
3.45225155e-01 2.28514418e-01 -1.49186110e+00 6.59015357e-01
1.21930110e+00 3.95487428e-01 3.47600609e-01 5.83944440e-01
-3.22972625e-01 5.12285471e-01 -6.20353699e-01 -4.94831622e-01
2.33723342e-01 -4.96105790e-01 -1.24823952e+00 1.34082019e-01
6.36214614e-01 -1.86220318e-01 1.09355852e-01 7.98404574e-01
1.07394733e-01 -9.84576494e-02 8.30986857e-01 -1.14950287e+00
-7.52092063e-01 9.40625727e-01 -2.47366726e-01 2.22019702e-01
3.37403417e-01 5.02312899e-01 1.15486622e+00 -6.63894534e-01
6.69313729e-01 1.69380808e+00 7.70646036e-01 6.12490654e-01
-1.75935256e+00 -3.61319870e-01 6.54338598e-01 -9.85683277e-02
-1.43217826e+00 -3.66948545e-01 5.08419991e-01 -6.46764636e-01
1.08419561e+00 -1.67556982e-02 6.14057124e-01 7.55145371e-01
4.60795090e-02 6.16492510e-01 1.06009650e+00 -2.73821205e-01
4.50045317e-01 2.14449883e-01 5.23392437e-03 7.63480246e-01
6.32065713e-01 -3.47555578e-01 -4.75432456e-01 -3.82778049e-01
9.36653852e-01 -3.57861519e-01 -1.23530991e-01 -9.67627585e-01
-1.17416012e+00 8.62237871e-01 4.91280437e-01 1.57438651e-01
-1.09143890e-01 3.60345244e-01 3.66524667e-01 3.64041865e-01
1.30261313e-02 1.12825787e+00 -6.59496725e-01 2.97085345e-01
-6.36543155e-01 3.91834766e-01 9.48427737e-01 1.04839051e+00
1.08981276e+00 -1.89882040e-01 8.38406235e-02 5.30671120e-01
4.19614345e-01 -1.18122483e-02 3.51632118e-01 -9.84637618e-01
6.58336043e-01 9.19228256e-01 3.54757421e-02 -6.61529481e-01
-5.83903611e-01 -5.15470088e-01 -3.53017598e-01 1.41429171e-01
4.61710632e-01 -1.79885238e-01 -8.37592065e-01 2.18353748e+00
-7.56378099e-02 -3.03599715e-01 2.40627021e-01 5.49246550e-01
6.89794898e-01 2.92854607e-01 3.51571769e-01 4.47188407e-01
1.31645918e+00 -7.57569432e-01 -1.34821147e-01 -8.10500085e-01
9.83044028e-01 -4.23170984e-01 1.46264470e+00 2.34944373e-01
-1.37493885e+00 -4.87858713e-01 -1.16003036e+00 -1.38750687e-01
-3.80336732e-01 -2.08197698e-01 9.37625885e-01 8.96837533e-01
-1.44881809e+00 6.46633565e-01 -7.45900214e-01 -7.00115085e-01
5.23509383e-01 4.96587276e-01 -1.84965819e-01 1.88403443e-01
-9.81038451e-01 1.14197135e+00 7.17577994e-01 -2.74059832e-01
-1.14276493e+00 -7.56534696e-01 -8.59377325e-01 3.28061759e-01
3.61061305e-01 -1.55499852e+00 1.39287972e+00 -1.39029038e+00
-1.21535671e+00 8.95466089e-01 -1.22429095e-01 -3.67913604e-01
1.11055501e-01 5.27339429e-02 1.85311615e-01 2.48531431e-01
5.15426844e-02 1.17036068e+00 5.72590172e-01 -1.65036976e+00
-2.47649968e-01 -2.28359118e-01 9.15156603e-01 4.46766049e-01
-2.19500035e-01 -3.44892710e-01 -1.21794350e-01 -5.27861893e-01
1.99877813e-01 -7.66300201e-01 -3.10370594e-01 1.66538641e-01
-2.04405382e-01 -1.13967255e-01 7.63924792e-02 -3.76939476e-01
1.02103508e+00 -1.72158813e+00 4.82787013e-01 5.71690239e-02
1.74294412e-01 -1.84980333e-01 -3.37203234e-01 7.72267699e-01
-4.33685444e-02 6.47759199e-01 -2.65627027e-01 -2.18458652e-01
1.93826228e-01 3.30393106e-01 -3.07729512e-01 1.34409994e-01
4.35923725e-01 9.97705758e-01 -9.84362364e-01 -1.49487510e-01
-1.63317710e-01 7.91008621e-02 -1.11285305e+00 4.56631333e-02
-6.32807076e-01 3.37340504e-01 -1.50401056e-01 2.79185474e-01
6.01145506e-01 -4.64179248e-01 3.69652897e-01 -1.07065484e-01
2.17630327e-01 7.55734980e-01 -1.29082119e+00 1.93316591e+00
-5.82941413e-01 4.42362607e-01 -5.15558757e-02 -9.43139732e-01
6.46250844e-01 2.41789997e-01 -2.37502500e-01 -2.17153385e-01
-2.43127853e-01 1.26281725e-02 4.60318297e-01 -4.07998443e-01
3.95876288e-01 -5.05439758e-01 3.40108387e-02 5.87064028e-01
3.60278815e-01 -3.66647065e-01 2.39354789e-01 5.82513630e-01
9.28591788e-01 2.09369481e-01 3.96032542e-01 -6.89583838e-01
2.25593969e-01 6.46259636e-02 4.45996732e-01 9.72308099e-01
2.32627377e-01 5.03701091e-01 6.65009081e-01 -2.70856410e-01
-1.12157416e+00 -1.34262991e+00 -7.55128078e-03 1.20104301e+00
2.10125774e-01 -7.26895392e-01 -7.72738814e-01 -5.00494421e-01
-1.26113623e-01 9.19718683e-01 -6.14443719e-01 -2.79709101e-01
-2.74233580e-01 -4.71072853e-01 6.58929825e-01 8.26794744e-01
3.10181439e-01 -7.35569715e-01 -8.00833642e-01 2.03718971e-02
-1.27644524e-01 -8.12344909e-01 -1.88210025e-01 2.69343555e-01
-1.43906236e+00 -9.99874473e-01 -2.93455750e-01 -6.84580147e-01
1.11827409e+00 4.00674224e-01 1.67815840e+00 4.55117524e-01
1.11083508e-01 8.26290846e-01 -1.23080164e-01 -3.43340933e-01
-5.75337470e-01 3.00362676e-01 -2.86924928e-01 -4.62910056e-01
8.58891085e-02 -6.46578729e-01 -5.96656859e-01 4.50409740e-01
-1.10117674e+00 2.53336251e-01 6.73240602e-01 5.29243827e-01
6.08462021e-02 -1.70981750e-01 6.29679918e-01 -1.18107283e+00
7.83165216e-01 -5.88034570e-01 -3.11431438e-01 4.61815387e-01
-4.51597482e-01 5.12700796e-01 5.44956565e-01 -3.80415499e-01
-1.06971097e+00 -2.45292354e-02 1.48291454e-01 1.29324228e-01
-2.75077969e-01 6.45532310e-01 -4.24705565e-01 2.77479384e-02
1.06935489e+00 -5.89700369e-03 -1.45543575e-01 -4.27149892e-01
7.18501389e-01 7.95940980e-02 1.34873539e-01 -1.23495507e+00
8.52091074e-01 5.67932665e-01 -2.40391940e-01 -7.78070033e-01
-4.58178163e-01 -2.28310227e-02 -7.55458951e-01 4.23690796e-01
6.59536958e-01 -1.13578117e+00 -3.02512705e-01 -3.32017019e-02
-1.19198418e+00 -5.61520815e-01 -2.59486407e-01 4.98917922e-02
-6.82527542e-01 4.44958687e-01 -4.23528731e-01 -5.13310492e-01
-4.83062342e-02 -1.17648840e+00 1.03704059e+00 1.77809238e-01
-7.23148704e-01 -1.30651772e+00 -4.00189199e-02 2.07194775e-01
5.30120611e-01 -2.16267988e-01 1.58785832e+00 -6.38416409e-01
-9.23820376e-01 3.23521197e-01 6.45195171e-02 1.29737049e-01
-8.28477293e-02 -1.19511567e-01 -8.84732664e-01 -4.34094369e-01
-2.61068285e-01 -3.66869569e-01 8.82434785e-01 1.34448960e-01
9.52466369e-01 -4.72000062e-01 -4.23351794e-01 6.94637299e-01
1.35646069e+00 -9.52752680e-02 7.69738615e-01 3.27848881e-01
4.46230888e-01 7.14175463e-01 5.90620702e-03 -3.52935642e-02
6.83434188e-01 5.62381029e-01 4.63360041e-01 4.67207395e-02
1.06109224e-01 -5.08478105e-01 3.64811301e-01 2.06546277e-01
1.08808130e-01 -5.41021638e-02 -8.95814478e-01 6.87153697e-01
-1.78019738e+00 -8.17676425e-01 3.45675677e-01 2.22219729e+00
9.92957771e-01 2.48068184e-01 2.69055992e-01 -3.40856671e-01
4.47876781e-01 -6.27717599e-02 -4.87971663e-01 -5.83265245e-01
-2.13105932e-01 2.65257984e-01 1.30606696e-01 6.12838805e-01
-7.44002998e-01 1.15333724e+00 7.23415279e+00 4.51142937e-01
-7.56625533e-01 5.44481352e-02 5.25304794e-01 1.22197419e-01
-1.01120114e+00 5.53471982e-01 -6.45639360e-01 -2.06278381e-03
6.15276098e-01 -1.08278938e-01 4.59407777e-01 5.93504250e-01
-2.11841077e-01 -1.42846107e-01 -1.64293551e+00 4.28917736e-01
-1.61144733e-01 -1.27649820e+00 5.12827337e-01 -2.07782418e-01
7.58843064e-01 -2.40314960e-01 -8.64782780e-02 4.80663329e-01
4.63505417e-01 -1.40044785e+00 7.70649254e-01 3.86520237e-01
1.05916426e-01 -5.29467165e-01 3.21616709e-01 5.54447293e-01
-9.63111281e-01 -7.64121637e-02 -5.77278435e-01 -2.76451409e-01
-5.40882349e-02 -8.70051458e-02 -9.68110919e-01 3.20682675e-01
4.49729860e-01 1.45294026e-01 -1.15033042e+00 1.15874529e+00
-6.23142123e-01 5.42020082e-01 -3.71539265e-01 5.00840954e-02
8.87283534e-02 1.30061090e-01 3.72660160e-01 1.25258470e+00
1.37937158e-01 -1.51319787e-01 4.11739759e-02 1.33714235e+00
1.28800496e-01 -4.23857607e-02 -6.03951097e-01 -3.77199650e-02
5.89023352e-01 1.04490900e+00 -8.34304452e-01 -2.31661022e-01
-4.16654229e-01 7.66093850e-01 5.86363554e-01 5.96650958e-01
-4.86055017e-01 -5.69512174e-02 6.72565877e-01 3.17662418e-01
5.46233773e-01 -3.72140825e-01 -5.36088288e-01 -1.36354399e+00
-1.14540815e-01 -9.39183652e-01 4.97493118e-01 -1.06834304e+00
-1.30707812e+00 3.46325845e-01 4.73346740e-01 -6.76066339e-01
-2.35624209e-01 -7.40359068e-01 -8.28309059e-01 9.58234906e-01
-1.02733970e+00 -1.20443130e+00 -7.72621483e-02 3.38450938e-01
3.20159435e-01 -3.72043997e-03 8.69554639e-01 -2.78370529e-01
-1.30912393e-01 5.81609190e-01 -4.69445378e-01 -5.49725480e-02
4.55933303e-01 -1.35311389e+00 6.74851477e-01 7.48381495e-01
2.70795017e-01 1.35763609e+00 9.01350319e-01 -4.57983226e-01
-1.26533902e+00 -6.41860127e-01 5.62983513e-01 -8.77545774e-01
4.85613346e-01 -5.43255270e-01 -1.08788323e+00 1.06245208e+00
3.86172116e-01 -5.20648301e-01 6.98705792e-01 7.02854514e-01
-8.15182149e-01 9.84887704e-02 -1.05449808e+00 8.64306808e-01
1.36901486e+00 -6.23048365e-01 -1.05499804e+00 2.30521500e-01
7.69128203e-01 -2.01889351e-01 -5.79932809e-01 3.18429530e-01
3.10187876e-01 -1.10544157e+00 1.09946167e+00 -1.00769734e+00
4.11256045e-01 -4.93928134e-01 -9.07036513e-02 -1.43296814e+00
-5.30086458e-01 -3.98922265e-01 5.35578467e-02 1.15958929e+00
6.83415711e-01 -7.11451292e-01 5.96948206e-01 7.17202544e-01
-1.42019674e-01 -7.57742941e-01 -6.21894538e-01 -7.91009188e-01
6.29637659e-01 -2.45604336e-01 6.35828614e-01 8.60331774e-01
2.64429510e-01 8.22746992e-01 3.52355838e-01 3.74963164e-01
3.86547267e-01 1.60035670e-01 9.34204519e-01 -1.25658798e+00
-6.79738283e-01 -7.37264931e-01 -4.53607202e-01 -1.47203302e+00
1.13091856e-01 -1.18569589e+00 -2.29833677e-01 -1.71234751e+00
4.25587207e-01 -4.19611633e-01 1.60931032e-02 4.74937290e-01
-3.10387880e-01 -2.85015583e-01 2.40152732e-01 1.18741974e-01
-6.21436834e-01 3.56110632e-01 1.12106001e+00 -2.27999061e-01
-2.01115116e-01 -1.26269966e-01 -1.49789596e+00 9.36246574e-01
8.70022535e-01 -1.65790066e-01 -8.61941636e-01 -9.24486578e-01
8.52045953e-01 -3.42686981e-01 6.88467026e-01 -7.74921417e-01
2.19328657e-01 -1.90548912e-01 3.86591405e-01 -2.46732786e-01
2.61802256e-01 -5.95605791e-01 9.31075141e-02 3.68609220e-01
-4.74492490e-01 2.48729691e-01 5.85921586e-01 4.09074128e-01
1.53868258e-01 -6.39959991e-01 5.84167719e-01 -5.68302035e-01
-7.74063945e-01 -1.35448530e-01 -5.47945082e-01 4.24813569e-01
6.76899791e-01 -4.59664077e-01 -4.81844842e-01 -5.98020196e-01
-7.24690080e-01 3.46782684e-01 8.23806226e-01 3.11809838e-01
3.57098520e-01 -1.06516981e+00 -5.42934835e-01 -9.92256552e-02
3.15823615e-01 1.63403109e-01 1.23868324e-02 4.70008790e-01
-5.06153286e-01 3.79844636e-01 -2.54968684e-02 -6.86645865e-01
-8.21320832e-01 6.78285360e-01 6.41181052e-01 2.29661502e-02
-3.52041036e-01 9.64731514e-01 8.39079022e-01 -7.16058671e-01
1.17385015e-01 -7.12421298e-01 2.35263884e-01 -1.20563589e-01
2.88326919e-01 -1.73321832e-02 -1.83492884e-01 -3.50631863e-01
-4.00609523e-01 5.69143236e-01 -1.50050461e-01 -2.22851485e-01
1.07047474e+00 -1.80933103e-01 -2.09053352e-01 3.77606839e-01
5.50930798e-01 -1.16544470e-01 -1.25331259e+00 -1.65169984e-01
1.79187745e-01 -2.31233701e-01 -4.54645127e-01 -8.89111876e-01
-7.52897799e-01 1.15840316e+00 1.56029135e-01 1.74178213e-01
9.27227736e-01 3.47782433e-01 1.08674802e-01 7.98638105e-01
6.12737775e-01 -8.68676901e-01 1.55182391e-01 5.56077421e-01
9.85268354e-01 -9.51106191e-01 2.36682415e-01 -4.04156655e-01
-7.62788057e-01 1.17408431e+00 7.29649901e-01 -1.22386821e-01
3.74613702e-01 6.21001273e-02 -1.00772781e-02 -2.65019476e-01
-1.01047778e+00 -2.29993835e-01 2.44872540e-01 9.15556312e-01
6.23821676e-01 -7.46555403e-02 -9.79764387e-02 6.79673254e-01
-3.64699513e-01 -3.06109190e-01 4.28166449e-01 7.76019216e-01
-6.51681960e-01 -9.73758340e-01 -2.67461658e-01 9.93255302e-02
1.96903553e-02 -2.59020329e-01 -7.84015536e-01 8.41982961e-01
2.40860015e-01 8.50971818e-01 -4.04164828e-02 -8.76774862e-02
2.76548892e-01 3.74694884e-01 1.04143512e+00 -1.20028150e+00
-6.83985293e-01 -3.45441192e-01 1.79512247e-01 -2.95769036e-01
-2.13617831e-01 -6.15461051e-01 -1.43786979e+00 -9.34781358e-02
-2.62495220e-01 -5.47040626e-02 3.48320663e-01 1.13769448e+00
5.67368209e-01 4.95621085e-01 -8.98289233e-02 -8.99844527e-01
-7.66125619e-01 -9.25173819e-01 -2.19687536e-01 5.49929023e-01
7.02518923e-03 -5.74090064e-01 -2.23026976e-01 4.87109683e-02] | [9.563258171081543, 6.980295181274414] |
ceb689ab-adf7-4183-ae9a-d017743585b8 | deep-image-harmonization-via-domain | 1911.13239 | null | https://arxiv.org/abs/1911.13239v3 | https://arxiv.org/pdf/1911.13239v3.pdf | DoveNet: Deep Image Harmonization via Domain Verification | Image composition is an important operation in image processing, but the inconsistency between foreground and background significantly degrades the quality of composite image. Image harmonization, aiming to make the foreground compatible with the background, is a promising yet challenging task. However, the lack of high-quality publicly available dataset for image harmonization greatly hinders the development of image harmonization techniques. In this work, we contribute an image harmonization dataset iHarmony4 by generating synthesized composite images based on COCO (resp., Adobe5k, Flickr, day2night) dataset, leading to our HCOCO (resp., HAdobe5k, HFlickr, Hday2night) sub-dataset. Moreover, we propose a new deep image harmonization method DoveNet using a novel domain verification discriminator, with the insight that the foreground needs to be translated to the same domain as background. Extensive experiments on our constructed dataset demonstrate the effectiveness of our proposed method. Our dataset and code are available at https://github.com/bcmi/Image_Harmonization_Datasets. | ['Wenyan Cong', 'Zhixin Ling', 'Weiyuan Li', 'Li Niu', 'Jianfu Zhang', 'Liu Liu', 'Liqing Zhang'] | 2019-11-27 | dovenet-deep-image-harmonization-via-domain | http://openaccess.thecvf.com/content_CVPR_2020/html/Cong_DoveNet_Deep_Image_Harmonization_via_Domain_Verification_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Cong_DoveNet_Deep_Image_Harmonization_via_Domain_Verification_CVPR_2020_paper.pdf | cvpr-2020-6 | ['image-harmonization'] | ['computer-vision'] | [ 3.49090874e-01 -3.93114924e-01 5.80312610e-02 -1.23912826e-01
-5.97977579e-01 -7.34068751e-01 6.33583963e-01 -1.38914749e-01
-2.46471599e-01 7.05116451e-01 -3.68110910e-02 -9.57181603e-02
1.33558229e-01 -6.48844123e-01 -7.05798805e-01 -8.88346076e-01
6.84678197e-01 -7.20083416e-02 2.55280048e-01 -1.96835533e-01
-4.37276950e-03 3.02667230e-01 -1.56967485e+00 2.81073332e-01
9.86464918e-01 1.04924214e+00 2.86823422e-01 6.16448045e-01
1.70289993e-01 5.92857420e-01 -5.48499703e-01 -6.56118691e-01
6.42254293e-01 -6.69839382e-01 -5.34041822e-01 3.71238887e-01
7.39240408e-01 -2.73287483e-02 -3.41518223e-01 1.36309469e+00
6.91330552e-01 3.77871916e-02 2.57247180e-01 -1.73919964e+00
-6.98161662e-01 3.12966883e-01 -6.62948668e-01 4.53427806e-02
-1.37750998e-01 4.03080165e-01 8.01748335e-01 -7.92022824e-01
8.62415910e-01 8.35962772e-01 4.43007797e-01 3.39171320e-01
-1.26473975e+00 -8.42098415e-01 -1.76304415e-01 4.13074046e-01
-1.57735562e+00 -6.12462282e-01 8.46529722e-01 -2.69278347e-01
8.81858915e-02 4.88942266e-01 7.20577836e-01 9.86070693e-01
-1.51720196e-01 5.14653206e-01 1.32566273e+00 -2.97878504e-01
-2.20410507e-02 1.95373401e-01 -4.25543219e-01 2.35288084e-01
2.36218363e-01 -1.44399598e-01 -4.60630715e-01 2.92351484e-01
4.99159247e-01 -2.24717051e-01 -6.64541960e-01 -3.01166385e-01
-1.21187031e+00 3.79828990e-01 5.62674403e-01 3.33455086e-01
-1.21891797e-01 -1.03670157e-01 1.96978942e-01 2.20418841e-01
3.79247755e-01 1.71913773e-01 -1.01697735e-01 2.23137572e-01
-9.55784082e-01 3.48600745e-01 3.25306743e-01 1.11415923e+00
7.33662248e-01 -4.90184985e-02 -1.69805259e-01 8.85849118e-01
-6.88872710e-02 6.29263103e-01 2.03717545e-01 -1.02066970e+00
4.37217861e-01 5.74484229e-01 2.84254044e-01 -1.30935562e+00
-1.97758749e-01 -3.22838873e-01 -1.10987711e+00 1.24165781e-01
3.67942065e-01 -1.35775451e-02 -5.70520818e-01 1.53717494e+00
6.08538926e-01 3.11181217e-01 1.17310792e-01 1.06697512e+00
1.03982127e+00 7.24583805e-01 -1.95879519e-01 -2.31219247e-01
1.35109317e+00 -1.16993093e+00 -6.47640467e-01 -4.20089811e-03
-4.85009812e-02 -1.19103837e+00 1.06570733e+00 4.76988763e-01
-1.05307877e+00 -9.17906761e-01 -9.57143128e-01 -2.04987124e-01
-3.04405898e-01 1.91674039e-01 2.80439645e-01 5.42018533e-01
-1.10080564e+00 1.35276929e-01 -1.97330788e-01 -4.20279741e-01
3.64863873e-01 -1.56548068e-01 -4.33975220e-01 -9.81459990e-02
-1.03355169e+00 5.84410429e-01 6.87649906e-01 2.16989756e-01
-6.80129290e-01 -6.97657943e-01 -6.60414815e-01 -3.06398600e-01
4.11093414e-01 -5.49668014e-01 9.38176811e-01 -1.16898453e+00
-1.21543801e+00 1.15122736e+00 8.49814042e-02 -4.05618578e-01
8.79241943e-01 1.43531421e-02 -6.93603337e-01 1.47375852e-01
1.76069304e-01 8.77338111e-01 9.53991055e-01 -1.54200137e+00
-9.04991746e-01 -2.03771338e-01 -5.42409644e-02 1.64331287e-01
-2.56209791e-01 -1.04143851e-01 -9.84944046e-01 -8.36623430e-01
-2.72183657e-01 -9.83471036e-01 4.64695841e-02 4.25336584e-02
-6.00376725e-01 2.33797789e-01 8.17592323e-01 -9.43844914e-01
1.17155635e+00 -2.42824960e+00 4.98646609e-02 1.22717306e-01
2.22746015e-01 2.84727842e-01 -3.54128897e-01 2.54698336e-01
-2.02925473e-01 -5.07193692e-02 -4.52466875e-01 -1.70221180e-01
-2.79130470e-02 -8.89536813e-02 -2.40660369e-01 6.27424479e-01
2.37301528e-01 8.11714947e-01 -7.29460895e-01 -6.78768396e-01
3.38825136e-01 5.39950490e-01 -4.22291845e-01 2.09293604e-01
-2.12559730e-01 7.28508174e-01 9.68566388e-02 6.86898530e-01
1.14083743e+00 -1.45603254e-01 2.50372589e-01 -7.41023242e-01
-2.76537031e-01 -4.43745017e-01 -1.17857599e+00 1.63670671e+00
-3.32214445e-01 8.82135093e-01 9.03427005e-02 -7.20310509e-01
9.96737659e-01 1.04630768e-01 5.96591532e-01 -1.05714488e+00
1.62265286e-01 3.41726363e-01 -9.08397064e-02 -1.88724831e-01
7.35736072e-01 2.38320321e-01 -5.42283691e-02 1.66291758e-01
-2.58200765e-01 -4.21022147e-01 5.31791151e-01 4.94955182e-02
5.88599384e-01 7.81740844e-02 2.51894057e-01 -9.91275534e-02
6.57693267e-01 1.95654497e-01 6.58392131e-01 4.37810600e-01
-3.90180171e-01 1.03892767e+00 3.98249716e-01 -3.37398231e-01
-1.30330539e+00 -1.09264565e+00 -1.34055302e-01 6.88604116e-01
6.78538084e-01 -3.03861409e-01 -9.55208778e-01 -1.73227072e-01
-1.49284407e-01 4.36776280e-01 -4.85245496e-01 7.64074251e-02
-4.83351976e-01 -7.53902256e-01 5.92955709e-01 6.96594939e-02
1.16471577e+00 -9.81443107e-01 -4.24142778e-01 -2.46262439e-02
-7.03258455e-01 -1.40293801e+00 -8.06080878e-01 -3.31414998e-01
-1.67750552e-01 -1.31793630e+00 -8.05228412e-01 -8.77833962e-01
5.36551535e-01 6.23213470e-01 1.13007903e+00 1.16596207e-01
-4.92334187e-01 1.28553316e-01 -4.26187813e-01 -1.72143295e-01
-4.97343928e-01 -1.67642072e-01 -2.18032971e-01 5.32960236e-01
2.96270810e-02 -3.29809457e-01 -9.64341223e-01 5.61710715e-01
-1.34789538e+00 5.00381589e-01 3.24085236e-01 7.06796467e-01
8.53661001e-01 5.05618751e-01 2.55571872e-01 -6.29048645e-01
4.43174452e-01 -1.36076972e-01 -9.42621708e-01 4.70318139e-01
-3.02034318e-01 -5.48667490e-01 4.89828169e-01 -3.83135259e-01
-1.20242274e+00 1.41571447e-01 1.51958153e-01 -4.70683813e-01
-4.84023727e-02 6.18063882e-02 -6.31811678e-01 -1.16430074e-01
5.54880559e-01 3.26956749e-01 -6.93371594e-02 -3.05076122e-01
5.61239958e-01 6.19045675e-01 1.12450802e+00 -2.44677126e-01
9.40780520e-01 7.02564955e-01 -2.62906075e-01 -8.12067270e-01
-5.20206153e-01 -3.31265152e-01 -4.57671732e-01 -4.44783628e-01
1.02489424e+00 -1.11305261e+00 -3.68395895e-01 8.16192210e-01
-1.04257715e+00 -5.70094228e-01 -1.13697551e-01 -3.49437445e-02
-3.69054466e-01 3.75194550e-01 -2.13344097e-01 -3.36762369e-01
-3.66867870e-01 -1.10665786e+00 8.78870070e-01 4.84631032e-01
5.53604849e-02 -5.01842618e-01 -9.37028416e-03 6.15892410e-01
3.72384936e-01 6.14846647e-01 4.21002239e-01 -3.13144252e-02
-5.82762718e-01 1.32552549e-01 -6.49332345e-01 4.14055169e-01
3.16622347e-01 4.12665844e-01 -8.62217546e-01 -1.70627147e-01
-1.63469642e-01 -2.36700457e-02 8.06826234e-01 2.39611819e-01
1.24275994e+00 -2.31758088e-01 4.03566882e-02 7.56893396e-01
1.43152428e+00 1.49810657e-01 1.01074004e+00 5.65432191e-01
7.68014312e-01 5.48889518e-01 8.31493735e-01 6.14238024e-01
5.34785092e-01 1.00080121e+00 3.95875782e-01 -5.86865723e-01
-6.15804315e-01 -4.63702567e-02 2.05599278e-01 6.23180449e-01
1.00054644e-01 -4.81042296e-01 -8.02844048e-01 5.07168114e-01
-1.84651387e+00 -8.75224769e-01 -3.76829684e-01 2.06367707e+00
9.67976928e-01 -2.16870129e-01 1.18220679e-01 9.06917527e-02
1.15080941e+00 1.86741456e-01 -5.54551363e-01 1.98755309e-01
-7.30174839e-01 1.08479433e-01 4.15861249e-01 4.14779276e-01
-1.36173308e+00 9.74711120e-01 4.32201433e+00 1.01506293e+00
-1.34929848e+00 2.11409867e-01 8.46527934e-01 -9.71931890e-02
-2.25891083e-01 -1.74501434e-01 -4.32200313e-01 8.50991905e-01
3.01253736e-01 -2.41793796e-01 5.44207871e-01 5.06733775e-01
2.76590079e-01 -1.86788201e-01 -6.29545748e-01 1.36776841e+00
2.40576431e-01 -1.38128650e+00 -5.63318394e-02 -1.70998707e-01
1.11530912e+00 -2.91738391e-01 2.55796969e-01 -1.90265074e-01
7.99951628e-02 -7.61370540e-01 9.68108237e-01 3.52124870e-01
9.02043104e-01 -7.53989816e-01 4.96268511e-01 -1.28530219e-01
-1.34341681e+00 1.85979620e-01 -2.54955739e-01 5.48692703e-01
1.31966501e-01 6.28860176e-01 -6.08390331e-01 7.45531082e-01
9.73668814e-01 7.86445022e-01 -9.07459795e-01 1.03511024e+00
-2.47861296e-01 1.05163760e-01 3.19087058e-02 5.02711177e-01
-1.94312572e-01 -3.72158557e-01 5.10866582e-01 1.00542617e+00
3.75067115e-01 -1.26076683e-01 5.25293015e-02 8.46484125e-01
-2.18005151e-01 2.35750839e-01 -4.13647830e-01 -1.09859258e-01
4.58287209e-01 1.49513090e+00 -7.81870544e-01 -2.76273787e-01
-2.37197056e-01 1.17801046e+00 -1.78503707e-01 3.44634950e-01
-1.33266878e+00 -3.18848610e-01 7.67030180e-01 -1.67086683e-02
2.26081267e-01 9.79499221e-02 -1.57480150e-01 -1.28521359e+00
2.39556625e-01 -1.35960865e+00 3.22943747e-01 -1.02540505e+00
-1.16709566e+00 7.19091117e-01 -2.01839864e-01 -1.53062153e+00
3.53932112e-01 -2.72695839e-01 -5.25690675e-01 6.18190467e-01
-1.48054600e+00 -1.41551936e+00 -9.84702647e-01 8.67055535e-01
3.42769712e-01 -1.03914224e-01 2.83484340e-01 7.01861203e-01
-7.56576598e-01 5.81842184e-01 7.50249773e-02 6.54434338e-02
1.11186218e+00 -8.67081761e-01 4.25799876e-01 1.17609489e+00
9.09960717e-02 1.20248161e-01 7.50184774e-01 -5.00138044e-01
-1.25266385e+00 -1.48975146e+00 5.39751470e-01 -1.17624767e-01
5.51472008e-01 -2.50207543e-01 -8.67525518e-01 2.04879686e-01
5.16820192e-01 8.84487033e-02 4.25822020e-01 -7.76571572e-01
-2.91259021e-01 -5.11875808e-01 -1.07224298e+00 8.26332271e-01
9.30367768e-01 -3.07812661e-01 6.60116374e-02 2.94899166e-01
8.58595431e-01 -5.27784228e-01 -8.47035170e-01 4.73486871e-01
5.34726679e-01 -1.12822711e+00 9.82027769e-01 3.96338739e-02
5.84336579e-01 -8.99333417e-01 -2.30972648e-01 -1.16193128e+00
-5.84413931e-02 -6.40343189e-01 3.38679612e-01 1.54554355e+00
1.50934756e-01 -5.58165550e-01 4.26178604e-01 2.25639656e-01
2.65676808e-02 -9.71057788e-02 -5.99210680e-01 -6.46679044e-01
-2.51366377e-01 -1.90178096e-01 8.02091599e-01 1.06492889e+00
-6.34254932e-01 -6.87080547e-02 -6.22679591e-01 1.74403936e-01
6.06740594e-01 5.20318151e-01 1.37616956e+00 -8.03429127e-01
-2.73923874e-01 -4.18111175e-01 -3.02075922e-01 -4.72115666e-01
-6.28617704e-02 -7.54802048e-01 2.52673440e-02 -1.30421007e+00
2.55625367e-01 -4.20757025e-01 -1.24317326e-01 3.83698940e-01
-2.55642563e-01 1.00383794e+00 5.03910720e-01 4.80580255e-02
-6.91406250e-01 4.70217168e-01 1.32899868e+00 -3.47874492e-01
-1.45479232e-01 -4.32856560e-01 -7.85875499e-01 4.71875727e-01
1.16972005e+00 -2.40817368e-01 -2.51883149e-01 -3.44770789e-01
8.33857134e-02 -2.17159837e-01 6.21694386e-01 -1.06062973e+00
9.26846191e-02 -2.51548588e-01 5.12147069e-01 -5.97416699e-01
3.26280206e-01 -8.33397806e-01 9.06722307e-01 4.90933329e-01
-5.88291474e-02 1.55774266e-01 3.37813586e-01 3.68249357e-01
-5.29113650e-01 1.28617898e-01 1.09310758e+00 3.07895355e-02
-1.08705890e+00 3.31743807e-01 1.39697596e-01 2.14637846e-01
1.19518244e+00 -1.35281771e-01 -5.45275867e-01 -3.30831110e-01
-3.83931607e-01 1.86334908e-01 8.97190869e-01 5.16642392e-01
5.45104206e-01 -1.45382833e+00 -8.98219228e-01 1.62811160e-01
4.09327149e-01 -1.90699063e-02 6.23894334e-01 8.28581333e-01
-8.63867819e-01 -1.55383512e-01 -4.86986935e-01 -5.94285011e-01
-1.59264243e+00 5.02712488e-01 2.16734856e-01 2.54011378e-02
-5.70495784e-01 6.05739653e-01 4.21692610e-01 -3.04983646e-01
9.77008417e-02 -8.01548213e-02 4.27069701e-02 2.40420714e-01
6.53144658e-01 3.36950153e-01 8.75586122e-02 -8.80221128e-01
-3.18081081e-01 4.26557511e-01 2.26932824e-01 -7.18116760e-03
1.17504072e+00 -3.71791661e-01 -4.41479981e-01 -7.34288990e-02
1.15130138e+00 7.08574504e-02 -1.09419978e+00 -1.76442802e-01
-2.13683769e-01 -8.07603002e-01 -2.19907060e-01 -5.80549955e-01
-1.39258087e+00 5.47270477e-01 8.68976951e-01 1.64300248e-01
1.66427040e+00 -2.98286021e-01 8.53496909e-01 -1.36194989e-01
1.23030201e-01 -1.10615730e+00 2.36171618e-01 1.10988066e-01
1.09238207e+00 -1.38509524e+00 -2.51886509e-02 -4.20004100e-01
-8.60744596e-01 7.70033062e-01 7.22384870e-01 9.83623937e-02
3.88619334e-01 6.43908605e-02 4.09526080e-01 1.05216399e-01
-2.75651574e-01 -3.37202430e-01 2.67015845e-01 5.25209367e-01
1.98531657e-01 1.70867607e-01 -1.52030289e-01 2.39116296e-01
-3.36259305e-01 -8.84126723e-02 5.51133573e-01 5.15532851e-01
-1.65821522e-01 -1.04118180e+00 -6.02951467e-01 -1.16611719e-01
-2.86058396e-01 -1.50994405e-01 -4.92410988e-01 7.66328514e-01
5.02126336e-01 1.11303091e+00 8.54542553e-02 -2.26812586e-01
2.00804755e-01 -4.80656505e-01 4.07765299e-01 -1.29721239e-01
-4.82780874e-01 2.42981270e-01 -1.54903620e-01 -6.18166268e-01
-6.64702833e-01 -3.72559756e-01 -8.18161130e-01 -5.50166488e-01
1.29212635e-02 -1.58536822e-01 5.76908827e-01 3.64890397e-01
3.71059328e-01 4.35430020e-01 7.21175730e-01 -8.17515790e-01
2.00147092e-01 -5.11200845e-01 -5.65129995e-01 7.04105973e-01
1.72632322e-01 -5.08424103e-01 -4.63090166e-02 6.08873129e-01] | [11.247176170349121, -1.181750774383545] |
869a74b7-dbcf-4153-a9cd-e2850e503b36 | what-do-neural-machine-translation-models | 1704.03471 | null | http://arxiv.org/abs/1704.03471v3 | http://arxiv.org/pdf/1704.03471v3.pdf | What do Neural Machine Translation Models Learn about Morphology? | Neural machine translation (MT) models obtain state-of-the-art performance
while maintaining a simple, end-to-end architecture. However, little is known
about what these models learn about source and target languages during the
training process. In this work, we analyze the representations learned by
neural MT models at various levels of granularity and empirically evaluate the
quality of the representations for learning morphology through extrinsic
part-of-speech and morphological tagging tasks. We conduct a thorough
investigation along several parameters: word-based vs. character-based
representations, depth of the encoding layer, the identity of the target
language, and encoder vs. decoder representations. Our data-driven,
quantitative evaluation sheds light on important aspects in the neural MT
system and its ability to capture word structure. | ['Hassan Sajjad', 'Nadir Durrani', 'James Glass', 'Yonatan Belinkov', 'Fahim Dalvi'] | 2017-04-11 | what-do-neural-machine-translation-models-1 | https://aclanthology.org/P17-1080 | https://aclanthology.org/P17-1080.pdf | acl-2017-7 | ['morphological-tagging'] | ['natural-language-processing'] | [ 4.69776899e-01 3.20162803e-01 -5.65603018e-01 -4.90413725e-01
-9.97383118e-01 -8.66597295e-01 7.90771306e-01 1.89155281e-01
-4.49611723e-01 5.63527942e-01 5.76442540e-01 -7.80222356e-01
3.95435601e-01 -6.51408315e-01 -9.08973694e-01 -3.59592497e-01
2.44266555e-01 6.36115074e-01 -1.35431483e-01 -2.52616823e-01
3.36102098e-02 1.88593641e-01 -9.73522484e-01 6.94025159e-01
9.90833402e-01 5.56255043e-01 3.54395539e-01 5.61884880e-01
-2.55159914e-01 6.58129930e-01 -4.58903939e-01 -8.49770963e-01
8.82012676e-03 -6.20881915e-01 -7.52355039e-01 -2.40814850e-01
4.89122003e-01 -1.14667356e-01 -5.80011487e-01 9.88716722e-01
3.00256640e-01 -2.13774994e-01 9.55767214e-01 -5.96506953e-01
-1.31130624e+00 1.05750000e+00 -2.15042770e-01 6.98224008e-01
-7.25157708e-02 3.22212726e-01 1.35313082e+00 -9.92502868e-01
8.16960752e-01 1.16529858e+00 5.11663496e-01 6.71957910e-01
-1.49919343e+00 -3.81714880e-01 1.66654423e-01 9.70151871e-02
-1.08317041e+00 -8.51456583e-01 5.32842577e-01 -4.08722609e-01
1.30317116e+00 -2.16944128e-01 2.84981251e-01 1.29347742e+00
5.81091881e-01 8.20702791e-01 1.02478421e+00 -6.41848862e-01
-9.06428173e-02 -3.51487217e-03 2.13618740e-01 8.70187223e-01
4.31365371e-01 1.95577264e-01 -7.80541480e-01 -6.84660152e-02
8.09893608e-01 -6.52667999e-01 -1.54736355e-01 -2.61626005e-01
-1.17767560e+00 9.26503897e-01 2.16480091e-01 3.88334751e-01
-2.48120680e-01 3.37879509e-01 4.45472926e-01 3.93516332e-01
4.81706679e-01 3.69441301e-01 -8.19922924e-01 -2.23866686e-01
-8.58166397e-01 -2.44197547e-01 5.91127634e-01 9.01945472e-01
7.61151612e-01 4.10576165e-01 -2.92221040e-01 9.33061719e-01
1.89497158e-01 4.32783753e-01 7.64832914e-01 -5.76531053e-01
9.62057233e-01 3.77361149e-01 -4.40326095e-01 -3.82494777e-01
-3.77686433e-02 -6.35506332e-01 -3.73004287e-01 -2.51463145e-01
4.95331556e-01 -3.41354430e-01 -9.27184880e-01 2.13066626e+00
-2.87536651e-01 -3.30652326e-01 2.92652488e-01 5.58363080e-01
5.44280350e-01 7.57246017e-01 2.01120540e-01 -4.46580574e-02
1.30929780e+00 -1.04652166e+00 -5.73780656e-01 -7.84021318e-01
8.48096013e-01 -7.21543968e-01 1.13035500e+00 -9.72504467e-02
-1.46512198e+00 -5.54034412e-01 -9.63294327e-01 -4.58364159e-01
-2.87850350e-01 5.64152300e-01 5.22739768e-01 6.42103016e-01
-1.08690917e+00 6.31884277e-01 -9.30826128e-01 -4.05424953e-01
1.81665123e-01 3.41147065e-01 -3.95069271e-01 -6.13613911e-02
-1.04135931e+00 1.40914917e+00 2.94132441e-01 -1.46146536e-01
-1.07623374e+00 -6.16842628e-01 -9.54939544e-01 1.46483302e-01
-2.54539549e-01 -7.93003857e-01 1.58539581e+00 -1.15591621e+00
-1.36582589e+00 1.06488264e+00 -4.75983560e-01 -4.14450139e-01
-7.02603906e-02 1.08531322e-02 -3.33547831e-01 7.04101250e-02
-9.05662850e-02 6.49907649e-01 5.44130743e-01 -1.02791667e+00
-3.88110459e-01 -4.34049070e-01 -2.06365049e-01 2.13832885e-01
-2.87835002e-01 3.10265332e-01 -1.53000951e-01 -9.15009916e-01
-4.22800258e-02 -8.80279839e-01 1.38482880e-02 -2.62530029e-01
-9.09259990e-02 -8.27738717e-02 -1.15410369e-02 -1.04535961e+00
1.13233840e+00 -2.09445310e+00 3.95307302e-01 -3.04087698e-01
-2.23439246e-01 2.88809478e-01 -5.99201620e-01 8.01195860e-01
1.60800957e-03 2.63101190e-01 -3.09889764e-01 -3.71802688e-01
-4.06639697e-03 3.42281193e-01 -3.58332306e-01 2.54067987e-01
6.39808536e-01 1.27901137e+00 -7.30621219e-01 -3.39983135e-01
-2.64314651e-01 5.56071997e-01 -4.25083846e-01 2.34958187e-01
-3.49213570e-01 3.06006908e-01 -3.22996020e-01 6.59298837e-01
1.87860414e-01 -4.80938097e-03 5.22405982e-01 3.72204036e-02
-2.10352931e-02 1.22389328e+00 -2.36159056e-01 1.87593937e+00
-6.16467118e-01 8.74056697e-01 -7.77300959e-03 -7.55755842e-01
8.22804689e-01 4.57695782e-01 -1.66221380e-01 -8.47633362e-01
1.21056952e-01 5.49259961e-01 6.10183597e-01 2.64970213e-02
5.65572381e-01 -3.20854604e-01 -3.02694831e-02 6.19496405e-01
5.68573117e-01 2.87927657e-01 9.25540403e-02 -1.22048378e-01
1.02696490e+00 3.06897223e-01 3.18992227e-01 -3.37574422e-01
1.07233122e-01 2.47623742e-01 4.55935121e-01 3.45159084e-01
6.47666603e-02 6.93307102e-01 4.31855410e-01 -2.81000644e-01
-1.21514714e+00 -1.13510489e+00 -1.46102518e-01 1.41867256e+00
-4.12586987e-01 -3.97273481e-01 -9.35733140e-01 -6.77749515e-01
-1.50868788e-01 9.45221901e-01 -7.31628180e-01 -7.10585713e-01
-1.02605712e+00 -7.46430933e-01 1.03365612e+00 7.23827720e-01
-4.68041748e-02 -1.08431804e+00 -3.47398669e-01 4.21857297e-01
-2.95506656e-01 -1.06803417e+00 -5.20507216e-01 5.64010918e-01
-1.34257305e+00 -6.95247531e-01 -5.07576168e-01 -9.58190262e-01
7.73979306e-01 5.74768744e-02 1.39031708e+00 -1.35230407e-01
3.48493189e-01 -7.75041804e-02 -2.51317054e-01 -1.72317415e-01
-8.33510160e-01 5.59020519e-01 -6.66688234e-02 -2.67284364e-01
3.74683440e-01 -5.40842831e-01 -2.22661138e-01 1.32201836e-01
-6.95169032e-01 -6.06378466e-02 1.01583993e+00 7.73239553e-01
5.05030572e-01 -5.51808357e-01 2.64574379e-01 -9.34598684e-01
7.36568570e-01 -4.85332102e-01 -2.32358098e-01 4.30007339e-01
-5.08522451e-01 5.34135938e-01 5.39980590e-01 -5.53346097e-01
-9.98893917e-01 -1.53142691e-01 -2.77102917e-01 -4.69960012e-02
8.99473429e-02 7.05996871e-01 -2.41434246e-01 1.65842086e-01
6.98546946e-01 4.46550816e-01 -1.44187689e-01 -7.83670604e-01
6.34159505e-01 5.91406524e-01 5.78603208e-01 -9.21133399e-01
6.26654744e-01 -6.01565540e-02 -5.41858971e-01 -4.88974512e-01
-8.25186312e-01 1.72158517e-02 -9.45898771e-01 5.18727005e-01
7.07534075e-01 -1.01526463e+00 1.71934813e-01 2.06701130e-01
-1.35499001e+00 -6.75246894e-01 -2.63730347e-01 3.15850079e-01
-5.97472191e-01 7.72092640e-02 -1.09627724e+00 -4.25401628e-01
-5.49397171e-01 -1.14768803e+00 9.03279006e-01 -1.52589396e-01
-3.51654977e-01 -1.04218173e+00 2.89058715e-01 2.49202684e-01
4.40294206e-01 -3.06842923e-01 1.74844241e+00 -7.53498971e-01
-4.38155353e-01 2.58021325e-01 -1.40685976e-01 2.83081293e-01
4.43489663e-02 -9.04180408e-02 -9.27402616e-01 -2.77291447e-01
-1.37788475e-01 -1.77043721e-01 1.06197238e+00 2.89236605e-01
4.48653162e-01 -6.03622854e-01 -1.70349807e-01 8.08362484e-01
1.31776071e+00 -4.68241759e-02 5.02198815e-01 3.20109844e-01
4.90528017e-01 6.56653881e-01 2.42974505e-01 -1.82353482e-01
4.11804795e-01 5.61477423e-01 2.89438248e-01 1.09955363e-01
-6.18199885e-01 -6.82138860e-01 8.65354002e-01 1.20897317e+00
2.29489490e-01 -4.30489212e-01 -1.06805944e+00 7.26073623e-01
-1.51614320e+00 -5.71469486e-01 6.08535931e-02 2.10928059e+00
1.25692356e+00 2.09902987e-01 -5.13933264e-02 -2.35679731e-01
7.27523327e-01 1.85935363e-01 -5.10800719e-01 -8.76061082e-01
-2.66285390e-01 3.63953620e-01 5.29106796e-01 4.88504708e-01
-6.59904838e-01 1.41485286e+00 7.34427834e+00 6.43245578e-01
-1.19255543e+00 3.57584178e-01 5.96689761e-01 1.70606926e-01
-5.31764984e-01 5.56305014e-02 -8.99530053e-01 2.04397038e-01
1.55421293e+00 -1.02468908e-01 6.12983108e-01 6.13226950e-01
-1.15323775e-01 4.05326217e-01 -1.47003353e+00 3.64506513e-01
7.33517855e-02 -1.42628777e+00 3.79238665e-01 4.83674079e-01
6.41065717e-01 7.77594745e-01 2.10308656e-01 4.34879005e-01
5.45753658e-01 -1.10493827e+00 1.16067517e+00 2.14245528e-01
1.10978806e+00 -4.94166434e-01 5.67524254e-01 5.08498788e-01
-8.63363981e-01 -6.71708807e-02 -5.97678304e-01 -5.35760596e-02
5.74492998e-02 2.73965985e-01 -8.81959200e-01 2.94333220e-01
-5.52169606e-02 3.82305443e-01 -8.01725268e-01 5.49180329e-01
-5.40834129e-01 1.06149375e+00 1.72864068e-02 7.21954256e-02
3.67734641e-01 1.21196797e-02 3.03771377e-01 1.61520207e+00
3.58971775e-01 -1.42810509e-01 -1.39376163e-01 9.53204989e-01
-4.74859983e-01 1.60955265e-01 -5.43176353e-01 -4.74551380e-01
5.41342735e-01 8.41656625e-01 -3.26619923e-01 -2.15382725e-01
-3.77476245e-01 8.04761589e-01 8.63710046e-01 3.81503969e-01
-4.67636108e-01 3.12976651e-02 9.49023783e-01 1.61241934e-01
3.83188516e-01 -6.93670511e-01 -4.98114169e-01 -1.28197634e+00
9.65402368e-03 -1.17504287e+00 2.29135811e-01 -6.11190617e-01
-1.14899874e+00 7.39341617e-01 -1.90283641e-01 -6.75915956e-01
-6.34181559e-01 -9.27137017e-01 -6.63537621e-01 1.05371130e+00
-1.49725497e+00 -1.30156589e+00 6.11542940e-01 2.35041246e-01
7.16623664e-01 -3.84696811e-01 1.06119418e+00 4.75777648e-02
-5.53778231e-01 8.56854260e-01 4.76930290e-01 6.26340508e-01
5.27767599e-01 -1.04120338e+00 1.08250344e+00 9.06368375e-01
7.10533917e-01 8.62662971e-01 3.27549577e-01 -6.57699049e-01
-1.70145476e+00 -1.05295789e+00 1.44279206e+00 -7.22768545e-01
6.88772857e-01 -4.72207397e-01 -8.87396872e-01 9.50164318e-01
5.76790333e-01 -1.30287677e-01 7.49026060e-01 4.55862731e-01
-8.31546605e-01 2.12492719e-01 -8.66003752e-01 6.07883751e-01
1.03490698e+00 -9.19466078e-01 -9.74532962e-01 -5.11727147e-02
8.96576881e-01 -1.80427819e-01 -6.34742141e-01 1.49310291e-01
5.65486312e-01 -5.81245422e-01 7.37716019e-01 -9.61496234e-01
7.03490555e-01 1.69025362e-01 -4.50136781e-01 -1.61046994e+00
-6.53200269e-01 -3.41182709e-01 -3.22023153e-01 1.30650544e+00
1.09860384e+00 -3.97109509e-01 5.42071640e-01 1.88113898e-01
-5.12660861e-01 -9.41828430e-01 -9.23920453e-01 -7.12085307e-01
7.96109498e-01 -2.97645777e-01 4.45433408e-01 8.75569642e-01
1.35234341e-01 8.51300359e-01 9.09411982e-02 3.21405521e-03
2.70496249e-01 9.38585214e-03 2.45110810e-01 -1.02827144e+00
-4.11887527e-01 -6.86046124e-01 -1.03444993e-01 -1.02069938e+00
4.20716107e-01 -1.41615593e+00 -1.12218738e-01 -1.69692469e+00
2.22825006e-01 -1.64248124e-01 -4.33333963e-01 6.30722702e-01
-1.00339524e-01 1.40894815e-01 1.46501958e-01 4.00727600e-01
-3.41196805e-02 4.23457801e-01 8.68036032e-01 -1.97655529e-01
2.51715630e-01 -2.73400784e-01 -8.68191421e-01 5.43171585e-01
1.05203116e+00 -6.80870950e-01 -2.68904895e-01 -1.55664361e+00
2.19894871e-01 1.69537775e-02 -1.60944641e-01 -5.11035323e-01
-1.79155860e-02 -1.41733140e-01 2.65046060e-01 -2.71911442e-01
3.95479441e-01 -2.66929179e-01 -3.38272423e-01 3.26665670e-01
-7.25320518e-01 4.98842835e-01 2.02012703e-01 2.67402977e-01
-1.67342693e-01 -4.80195254e-01 8.64535630e-01 -3.86663169e-01
-2.75668323e-01 1.45945996e-01 -6.42573953e-01 4.34302002e-01
2.16555446e-01 -1.52819127e-01 -3.52703601e-01 -1.54519454e-01
-5.12158394e-01 -3.10222417e-01 6.78928077e-01 5.20863712e-01
3.03583175e-01 -1.23358262e+00 -1.00431538e+00 2.45761424e-01
2.11595774e-01 -6.37997746e-01 -5.97900689e-01 5.45402706e-01
-2.82629222e-01 5.76621890e-01 -3.08987141e-01 -2.55641401e-01
-8.63511682e-01 3.80638748e-01 3.19332451e-01 -4.88136142e-01
-3.07633698e-01 7.55021930e-01 2.75114208e-01 -6.25384629e-01
6.34766975e-03 -3.62292856e-01 6.06155768e-02 1.06433500e-02
2.17580497e-01 6.14634790e-02 3.02996278e-01 -9.23236668e-01
-2.15526044e-01 4.15223449e-01 -2.72702843e-01 -5.40889204e-01
1.25881994e+00 -5.93888527e-03 -2.25333273e-02 4.69713449e-01
1.04645324e+00 2.22379602e-02 -1.11966980e+00 -3.87613505e-01
3.22759449e-01 -5.68648987e-02 2.90176664e-02 -1.04559112e+00
-7.68945932e-01 1.54445636e+00 1.38607711e-01 -4.11932051e-01
6.79318845e-01 1.32458940e-01 9.90201592e-01 5.08794129e-01
3.89766097e-01 -9.97190833e-01 -2.17248753e-01 1.07152236e+00
6.53434634e-01 -8.80018353e-01 -4.73709196e-01 -2.23436773e-01
-6.08944535e-01 9.41919863e-01 4.30321336e-01 -2.29466669e-02
1.69580892e-01 2.31181309e-01 1.45402953e-01 1.87599227e-01
-1.27279603e+00 -2.82268316e-01 3.87570053e-01 6.66523397e-01
1.07881296e+00 1.37844279e-01 -2.89815962e-01 6.96457207e-01
-4.63737339e-01 -5.46813548e-01 3.64343613e-01 7.84307420e-01
-6.16578996e-01 -1.41525018e+00 -2.28482876e-02 1.47048235e-01
-7.98992574e-01 -7.37557411e-01 -9.08818960e-01 5.80430150e-01
-5.20770159e-03 9.11614656e-01 2.58819684e-02 -2.51410425e-01
3.13995510e-01 5.91467023e-01 7.86950886e-01 -1.10418332e+00
-9.77491736e-01 -4.40215282e-02 4.60538030e-01 -1.76616445e-01
2.14586794e-01 -8.08388412e-01 -1.24175334e+00 -6.37706891e-02
-2.09078938e-01 1.65934905e-01 8.46546650e-01 9.33539927e-01
5.47492683e-01 2.56304950e-01 3.26078832e-01 -5.38877726e-01
-8.89330328e-01 -1.24254620e+00 4.25434373e-02 1.11730993e-01
2.18914643e-01 -2.06761286e-01 1.71541367e-02 1.13371752e-01] | [11.311836242675781, 9.833226203918457] |
1a0a8937-0555-477b-ad65-d174d6c94302 | learning-conditional-attributes-for-1 | 2305.17940 | null | https://arxiv.org/abs/2305.17940v2 | https://arxiv.org/pdf/2305.17940v2.pdf | Learning Conditional Attributes for Compositional Zero-Shot Learning | Compositional Zero-Shot Learning (CZSL) aims to train models to recognize novel compositional concepts based on learned concepts such as attribute-object combinations. One of the challenges is to model attributes interacted with different objects, e.g., the attribute ``wet" in ``wet apple" and ``wet cat" is different. As a solution, we provide analysis and argue that attributes are conditioned on the recognized object and input image and explore learning conditional attribute embeddings by a proposed attribute learning framework containing an attribute hyper learner and an attribute base learner. By encoding conditional attributes, our model enables to generate flexible attribute embeddings for generalization from seen to unseen compositions. Experiments on CZSL benchmarks, including the more challenging C-GQA dataset, demonstrate better performances compared with other state-of-the-art approaches and validate the importance of learning conditional attributes. Code is available at https://github.com/wqshmzh/CANet-CZSL | ['Chunhua Shen', 'Peng Wang', 'Guoqiang Liang', 'Hao Chen', 'Chenchen Jing', 'Lingqiao Liu', 'Qingsheng Wang'] | 2023-05-29 | learning-conditional-attributes-for | http://openaccess.thecvf.com//content/CVPR2023/html/Wang_Learning_Conditional_Attributes_for_Compositional_Zero-Shot_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wang_Learning_Conditional_Attributes_for_Compositional_Zero-Shot_Learning_CVPR_2023_paper.pdf | cvpr-2023-1 | ['compositional-zero-shot-learning'] | ['computer-vision'] | [ 5.66829681e-01 2.17061788e-01 -1.81130707e-01 -7.63519883e-01
-7.58337915e-01 -6.39111578e-01 7.04988956e-01 2.52179980e-01
-1.32744983e-01 4.59417909e-01 2.12567806e-01 9.37267095e-02
2.80718468e-02 -1.00478685e+00 -9.19134259e-01 -8.97555709e-01
1.01368897e-01 8.59804332e-01 -1.17019847e-01 -1.67011797e-01
-1.83936030e-01 -3.95516604e-02 -2.00190258e+00 8.34894836e-01
6.01258278e-01 1.18445325e+00 8.98715332e-02 4.30562735e-01
-4.93103534e-01 7.52083421e-01 -1.84275791e-01 -7.13308334e-01
1.71259061e-01 -3.80685568e-01 -8.18987191e-01 -4.70351204e-02
6.47234917e-01 -9.03812051e-02 9.41185802e-02 9.94663894e-01
2.86130816e-01 2.06740290e-01 1.03295040e+00 -1.60627437e+00
-1.31519783e+00 1.07429969e+00 6.12669848e-02 -3.36406589e-01
3.70107323e-01 3.52117509e-01 1.58938074e+00 -1.14442718e+00
6.57647550e-01 1.24863112e+00 4.87383097e-01 1.05779755e+00
-1.70705128e+00 -8.23483706e-01 4.27741766e-01 6.51615322e-01
-1.32045341e+00 -2.87959009e-01 7.85463333e-01 -5.87078035e-01
6.02929413e-01 2.00572625e-01 5.52540302e-01 1.73526406e+00
-3.66974115e-01 1.20109773e+00 1.06353641e+00 -4.14744675e-01
4.54193950e-01 1.91565260e-01 1.39985338e-01 4.55503762e-01
-8.99646580e-02 1.05785131e-01 -6.92607164e-01 -1.37671843e-01
-2.49161609e-02 1.40498459e-01 3.42579000e-03 -8.53311121e-01
-1.48011112e+00 8.95078123e-01 4.22166198e-01 -6.34085163e-02
-1.80339396e-01 3.12499791e-01 5.07787824e-01 3.93149227e-01
4.08523947e-01 6.16598427e-01 -5.71357906e-01 1.29164636e-01
-1.92453399e-01 3.34559292e-01 5.96633673e-01 1.53799593e+00
9.21642005e-01 1.16543815e-01 -1.27909541e-01 8.36316347e-01
2.72238553e-02 4.54120278e-01 3.18430901e-01 -9.23614800e-01
8.52719545e-02 6.84706926e-01 -3.30915242e-01 2.25607511e-02
-2.87178289e-02 -2.75182486e-01 -6.54832006e-01 2.96183139e-01
2.93746978e-01 1.19310804e-01 -9.89546120e-01 1.98447740e+00
3.25185597e-01 2.76589125e-01 2.33174294e-01 6.84313536e-01
1.00135207e+00 5.57882249e-01 6.24948263e-01 3.59033763e-01
1.49474728e+00 -8.78217578e-01 -5.94654679e-01 -1.54839531e-01
6.17122352e-01 -5.68771660e-01 1.51579082e+00 2.10545316e-01
-7.87619710e-01 -7.22520292e-01 -1.16652513e+00 -1.99063376e-01
-9.43520129e-01 -1.33773342e-01 7.59478331e-01 4.04597491e-01
-4.95593369e-01 4.93859440e-01 -4.35930192e-01 -3.82124454e-01
6.92732930e-01 2.10655630e-01 -4.25019652e-01 -2.10043207e-01
-1.15386903e+00 5.70404112e-01 6.74641848e-01 -5.87443054e-01
-1.32034898e+00 -1.22714770e+00 -1.14612436e+00 2.66134948e-01
5.52919507e-01 -8.40185940e-01 1.24695778e+00 -1.21424353e+00
-1.15000319e+00 1.03611505e+00 4.43664789e-02 -3.21442544e-01
2.32666850e-01 -2.79673755e-01 -5.13576031e-01 -2.98951343e-02
1.38390243e-01 8.92165959e-01 8.72284293e-01 -1.46998107e+00
-6.43200934e-01 -3.66530448e-01 7.80404136e-02 -1.10331038e-02
-2.96010375e-01 7.17596151e-03 3.59297365e-01 -7.60977626e-01
-8.53737891e-02 -9.30409372e-01 7.25739598e-02 1.95682526e-01
-2.72219270e-01 -4.32926148e-01 7.03165531e-01 -6.29809424e-02
5.92945397e-01 -2.54946518e+00 2.27693543e-01 -1.14513427e-01
7.40912482e-02 -3.42364460e-02 -2.60021001e-01 4.17687476e-01
-4.53164160e-01 -5.82295252e-05 -5.21426022e-01 -2.82456577e-01
5.00090539e-01 3.14523071e-01 -5.36829650e-01 1.31866053e-01
5.66096067e-01 9.42684114e-01 -1.14644945e+00 -3.56685758e-01
1.66588992e-01 5.14028370e-01 -4.40467745e-01 4.13163155e-01
-8.17940295e-01 4.24442530e-01 -1.73802942e-01 9.00108397e-01
2.96453804e-01 -1.47105187e-01 1.52134344e-01 -3.31227243e-01
2.77217001e-01 2.26279974e-01 -1.09304059e+00 1.88308680e+00
-3.39713126e-01 2.72727668e-01 -3.90532613e-01 -7.18197227e-01
9.43359673e-01 4.58978206e-01 2.70892084e-01 -4.39149261e-01
1.94349930e-01 1.94265619e-01 -1.36505561e-02 -3.54569405e-01
1.44404322e-01 -4.19124633e-01 -3.59707832e-01 5.59825778e-01
6.95705473e-01 -9.13466066e-02 4.32693176e-02 2.17538193e-01
7.18132555e-01 5.71996152e-01 3.82713854e-01 -4.08934146e-01
5.60927033e-01 -1.47695452e-01 7.38923788e-01 4.37989503e-01
-9.84535590e-02 6.48125291e-01 3.88520986e-01 -7.81335056e-01
-1.27534866e+00 -1.66403723e+00 -9.19541046e-02 1.82035613e+00
4.98371609e-02 -5.64390063e-01 -4.34788853e-01 -6.81343436e-01
2.28427693e-01 1.25276172e+00 -1.02565897e+00 -4.00770634e-01
-4.80701208e-01 -4.49371874e-01 2.44793028e-01 9.79620397e-01
9.43438336e-02 -1.12976563e+00 -5.12654722e-01 1.38316333e-01
-6.27399459e-02 -9.27762806e-01 -3.49141121e-01 4.74717200e-01
-4.02882576e-01 -9.61807549e-01 -1.93383798e-01 -8.77694666e-01
6.01117015e-01 -2.63291448e-01 1.48581147e+00 -1.16574585e-01
-4.48109627e-01 4.20211047e-01 -6.72317684e-01 -6.84397995e-01
-6.05967164e-01 6.60799667e-02 1.06341124e-01 3.88404757e-01
6.90532625e-01 -7.72280157e-01 -5.39766073e-01 1.77639544e-01
-9.27164197e-01 2.04952285e-01 5.69573402e-01 9.70819712e-01
7.77835786e-01 -5.94134986e-01 6.97534442e-01 -1.40507472e+00
2.53319472e-01 -5.32001853e-01 -1.68579414e-01 4.19875175e-01
-6.42617762e-01 3.91149312e-01 7.33879149e-01 -8.16826224e-01
-1.03458846e+00 1.93616733e-01 -7.86222145e-02 -4.58398193e-01
-4.80009228e-01 -1.91757545e-01 -7.82431841e-01 5.49421847e-01
6.03824079e-01 2.73529798e-01 -2.93025970e-02 -4.38534677e-01
1.00135851e+00 3.86317939e-01 5.90514183e-01 -1.30799282e+00
6.51419580e-01 5.00728667e-01 -1.52177066e-02 -2.53315300e-01
-8.16348553e-01 -2.90765047e-01 -8.03800046e-01 -5.82240000e-02
1.02669966e+00 -9.40673053e-01 -7.19471335e-01 2.80584484e-01
-7.92187572e-01 -4.38306898e-01 -8.61958206e-01 2.51033485e-01
-1.06182241e+00 -2.48160705e-01 -5.79860568e-01 -4.56915081e-01
-3.24139863e-01 -1.00017715e+00 9.49799657e-01 -8.46229121e-02
-6.09562993e-01 -7.56295383e-01 -6.55824915e-02 1.33343458e-01
3.95198554e-01 3.62893224e-01 1.72149277e+00 -1.20743287e+00
-5.86398482e-01 1.68228865e-01 1.00256372e-02 3.48268449e-01
2.84949869e-01 1.00633293e-01 -1.26601219e+00 -1.96747705e-01
-3.32123965e-01 -7.03001857e-01 8.24867308e-01 -2.50786781e-01
1.32390773e+00 -4.22978848e-01 -1.38037920e-01 6.80450737e-01
1.43081725e+00 1.58178851e-01 4.57740635e-01 2.06257373e-01
7.14474618e-01 6.83380127e-01 6.30020261e-01 3.52687031e-01
4.09651875e-01 4.86746520e-01 5.00681818e-01 2.60128856e-01
-4.05065387e-01 -4.83399123e-01 1.29667982e-01 5.10105431e-01
1.56011462e-01 1.30078331e-01 -7.70554721e-01 7.57813811e-01
-1.69213808e+00 -1.05418932e+00 1.59379810e-01 2.06174874e+00
1.10327852e+00 -1.46657333e-01 1.00076884e-01 2.64796824e-03
6.19474769e-01 1.13266192e-01 -6.15180612e-01 -3.84984344e-01
-2.67711937e-01 4.58879411e-01 2.37501264e-02 2.02468500e-01
-1.21610689e+00 9.10683990e-01 5.24725103e+00 6.12645447e-01
-6.58262014e-01 2.96799183e-01 5.37027121e-01 -2.19938219e-01
-6.07028604e-01 9.35215801e-02 -6.11656368e-01 5.15733659e-01
7.88845360e-01 -1.92462608e-01 3.74652088e-01 1.08566451e+00
-7.80539453e-01 4.16527957e-01 -1.83184707e+00 6.94448531e-01
2.67124355e-01 -1.25464594e+00 5.20370126e-01 -2.72407353e-01
6.23811722e-01 -3.22796285e-01 3.75609815e-01 7.27242529e-01
7.42906332e-01 -9.49800909e-01 8.14134002e-01 4.59407151e-01
9.27695870e-01 -6.49714589e-01 4.00425345e-01 -1.19490616e-01
-1.29524434e+00 -2.18891814e-01 -3.26065540e-01 1.39566362e-01
-2.23353878e-01 1.31590990e-02 -7.60693967e-01 5.20217180e-01
7.95162141e-01 6.73206985e-01 -6.08361959e-01 5.77080607e-01
-5.81228554e-01 4.66673791e-01 2.79345006e-01 5.86374067e-02
1.40698105e-02 -1.92473158e-01 3.57069314e-01 1.19104958e+00
2.65384316e-01 2.00611725e-01 5.79308160e-02 1.21857154e+00
-2.41541058e-01 9.66896936e-02 -5.92675805e-01 -4.83906157e-02
5.87534070e-01 1.09587896e+00 -2.31209055e-01 -6.33322954e-01
-5.61609268e-01 8.21910262e-01 4.34751362e-01 3.70504290e-01
-7.07600296e-01 -1.79381788e-01 1.18472183e+00 4.06894088e-02
4.45341527e-01 1.80608481e-01 -3.40086609e-01 -9.75192070e-01
-2.04877347e-01 -1.05202687e+00 8.64267826e-01 -8.34414780e-01
-1.73464119e+00 4.94626135e-01 5.01885936e-02 -1.36817396e+00
-1.92411512e-01 -7.62278140e-01 -5.68084955e-01 6.71532214e-01
-1.20385098e+00 -1.69288862e+00 -2.72889644e-01 6.49745345e-01
8.22180986e-01 -4.21251476e-01 1.30147636e+00 1.48348600e-01
-5.53095974e-02 6.72422647e-01 -6.64640069e-02 1.51837632e-01
9.59761441e-01 -1.58531344e+00 5.38445354e-01 4.70874965e-01
3.14342976e-01 5.84933877e-01 8.43794465e-01 -3.62105012e-01
-1.31083083e+00 -1.27214193e+00 7.67219543e-01 -7.36348152e-01
7.32118487e-01 -8.33384216e-01 -1.16865468e+00 1.03028870e+00
4.29210901e-01 1.97270364e-01 1.40180135e+00 2.10798651e-01
-1.25853789e+00 -2.39693537e-01 -9.35976684e-01 5.16343653e-01
1.14459550e+00 -6.70394599e-01 -7.71842360e-01 1.79228857e-01
1.07972205e+00 6.89588711e-02 -1.04767418e+00 5.53798318e-01
6.67425156e-01 -6.48220181e-01 1.15430737e+00 -1.12010980e+00
5.34856081e-01 -3.26733738e-01 -5.43757021e-01 -1.55825901e+00
-5.85045397e-01 6.76399504e-04 -2.12022334e-01 1.47341585e+00
6.09334171e-01 -5.65562844e-01 5.30602813e-01 7.12875187e-01
-2.71178275e-01 -7.24871218e-01 -7.96375632e-01 -1.01521194e+00
3.09088409e-01 -2.98242807e-01 1.23417723e+00 1.19160652e+00
-1.17734812e-01 4.21385676e-01 -1.20125942e-01 -1.98977403e-02
8.83676112e-01 4.59586799e-01 6.72332942e-01 -1.21516621e+00
-3.63750279e-01 -4.65309709e-01 -5.72695315e-01 -3.62968206e-01
5.66804647e-01 -1.28901446e+00 -8.92611817e-02 -1.24156821e+00
3.69550616e-01 -5.19244313e-01 -4.67374861e-01 7.23317027e-01
-3.95364940e-01 1.08003616e-01 2.80104697e-01 -9.37115029e-02
-5.78320444e-01 8.78906548e-01 9.77676570e-01 -6.52314901e-01
4.43908274e-01 -2.82564819e-01 -7.34223425e-01 4.02759403e-01
9.34366405e-01 -6.24688864e-01 -5.93154907e-01 -3.81210029e-01
2.56130636e-01 -5.38550973e-01 4.95005339e-01 -8.45829666e-01
2.73420978e-02 -2.08280295e-01 4.58813936e-01 -2.92068779e-01
7.23964393e-01 -9.68366325e-01 3.52438450e-01 2.52828896e-01
-7.09264517e-01 7.77416304e-02 -8.76906812e-02 5.44216216e-01
-7.31665492e-02 -1.55820802e-01 7.25230515e-01 -2.35468462e-01
-1.07733786e+00 3.41175884e-01 -2.98799369e-02 1.97163746e-01
1.15907955e+00 -4.03797589e-02 -6.47488654e-01 -7.69382790e-02
-9.04612601e-01 3.00156415e-01 6.40709877e-01 6.28498554e-01
7.33518839e-01 -1.75307810e+00 -8.80141735e-01 3.53408724e-01
9.78367805e-01 -1.51535496e-01 1.43476233e-01 2.60287702e-01
-5.13015613e-02 -1.04322590e-01 -5.68660140e-01 -4.95106161e-01
-1.19021511e+00 1.10128856e+00 2.04250626e-02 4.06370223e-01
-7.23420739e-01 1.03761387e+00 4.87903237e-01 -7.06744134e-01
2.19049498e-01 -1.97636336e-02 1.41429648e-01 7.80490711e-02
5.00543356e-01 1.43291354e-01 -2.17690900e-01 -5.58421135e-01
-2.64047861e-01 3.54833275e-01 -2.02177137e-01 1.30984947e-01
1.30748689e+00 1.82386488e-01 -7.34896809e-02 8.44270051e-01
1.20584202e+00 -4.66349214e-01 -1.34717107e+00 -7.09799170e-01
1.37725651e-01 -5.38925707e-01 -7.83547759e-01 -8.53392005e-01
-7.80663133e-01 1.07452703e+00 6.73144281e-01 -2.71842331e-01
8.83591294e-01 5.77753723e-01 4.27599013e-01 4.16827500e-01
3.22963774e-01 -1.03344822e+00 2.67530710e-01 1.57323465e-01
9.03217435e-01 -1.23544538e+00 -2.23016798e-01 -4.55171764e-01
-9.22871649e-01 9.46823120e-01 8.74083698e-01 -1.37107223e-02
3.42870086e-01 7.48559386e-02 2.11828634e-01 -1.21579543e-01
-1.11298490e+00 -3.44529629e-01 1.49997711e-01 7.69075811e-01
4.97790545e-01 4.80516374e-01 8.32229406e-02 7.75921166e-01
-2.30612710e-01 -4.97763097e-01 3.11445028e-01 7.22006738e-01
-1.53715372e-01 -1.41461182e+00 3.34510356e-02 4.75553453e-01
-4.06568460e-02 -2.32275799e-01 -4.57205951e-01 6.92810535e-01
4.43379998e-01 6.44589722e-01 1.99575245e-01 -3.18835914e-01
2.82490462e-01 8.01367104e-01 5.53116858e-01 -8.99970472e-01
-5.23985207e-01 -4.60075289e-01 1.16773546e-01 -4.42499638e-01
-3.47991794e-01 -7.58562982e-01 -1.21870649e+00 -7.56059960e-02
2.23486423e-01 6.56791329e-02 4.17118609e-01 6.32991076e-01
2.23822385e-01 5.77572703e-01 4.13955837e-01 -3.47693622e-01
-6.30474031e-01 -8.62561464e-01 -3.98782283e-01 1.19496512e+00
1.61453992e-01 -7.94366956e-01 -2.99223840e-01 3.62928063e-01] | [10.15123176574707, 2.2692601680755615] |
7b75bdda-a2d7-4ff7-8f26-92af9bff058f | impact-of-spatiotemporal-heterogeneity-in | 2204.00353 | null | https://arxiv.org/abs/2204.00353v1 | https://arxiv.org/pdf/2204.00353v1.pdf | Impact of spatiotemporal heterogeneity in heat pump loads on generation and storage requirements | This paper investigates how spatiotemporal heterogeneity in inflexible residential heat pump loads affects the need for storage and generation in the electricity system under business-as-usual and low-carbon emissions budgets. Homogeneous and heterogeneous heat pump loads are generated using population-weighted average and local temperature, respectively, assuming complete residential heat pump penetration. The results of a storage and generation optimal expansion model with network effects for spatiotemporally homogeneous and heterogeneous load profiles are compared. A case study is performed using a 3-bus network of London, Manchester, and Glasgow in Britain for load and weather data for representative weeks. Using heterogeneous heating demand data changes storage sizing: under a business-as-usual budget, 26% more total storage is built on an energy and power basis, and this storage is distributed among all of the buses in the heterogeneous case. Under a low-carbon budget, total energy storage at all buses increases 2 times on an energy basis and 40% on a power basis. The energy to power ratio of storage at each bus also increases when accounting for heterogeneity; this change suggests that storage will be needed to provide energy support in addition to power support for electric heating in high-renewable power systems. Accounting for heterogeneity also increases modeled systems costs, particularly capital costs, because of the need for higher generation capacity in the largest load center and coincidence of local peak demand at different buses. These results show the importance of accounting for heat pump load heterogeneity in power system planning. | ['Malcolm D. McCulloch', 'Filiberto Fele', 'Claire E. Halloran'] | 2022-04-01 | null | null | null | null | ['total-energy'] | ['miscellaneous'] | [-5.48388362e-01 1.49255484e-01 -3.42402995e-01 6.17302358e-02
-2.76691377e-01 -5.83946109e-01 6.96692586e-01 4.04085129e-01
7.89959431e-02 1.18685079e+00 4.43898082e-01 -5.62114358e-01
-2.99342275e-01 -1.43158460e+00 -3.15465361e-01 -1.00311077e+00
-1.29780442e-01 6.41699493e-01 -1.16829425e-01 -1.75348997e-01
-2.21676871e-01 7.14949489e-01 -1.36695147e+00 -3.83294523e-01
8.82931113e-01 5.32890916e-01 5.70539415e-01 8.04842561e-02
4.05430347e-01 1.39424698e-02 -4.17824060e-01 6.38638198e-01
3.75243396e-01 -7.03785181e-01 -5.72333038e-01 -2.69013852e-01
-8.99367034e-01 -4.81987000e-01 -2.22548954e-02 5.51353216e-01
7.83970654e-01 6.44650578e-01 8.13312232e-01 -1.30350375e+00
-1.42330036e-01 9.09313440e-01 -2.58804590e-01 2.50120670e-01
-9.52652693e-02 3.50523919e-01 9.11746621e-01 -9.01675001e-02
-1.17217392e-01 7.23670304e-01 3.63574363e-02 1.23044290e-02
-1.37661111e+00 -4.16431367e-01 -3.51828218e-01 -7.30736479e-02
-1.31516337e+00 -1.75910935e-01 5.10085225e-01 -1.54381916e-01
1.90308249e+00 1.08650231e+00 1.53150356e+00 -2.65601695e-01
3.44070554e-01 3.08647603e-01 1.16220975e+00 -5.36160648e-01
5.60740590e-01 2.79690713e-01 -2.37652034e-01 -3.95826727e-01
5.60639322e-01 6.93962574e-02 3.30457926e-01 -3.78692299e-02
1.79315642e-01 -2.75812268e-01 -1.71010584e-01 7.16502145e-02
-8.00813258e-01 4.65420514e-01 6.46925032e-01 8.10433269e-01
-3.69517952e-01 1.34224936e-01 1.83974877e-01 -2.56626189e-01
4.16320533e-01 5.81967831e-01 -3.38225424e-01 6.11593947e-02
-1.50101340e+00 2.49820426e-01 7.20169306e-01 9.85891759e-01
7.77491271e-01 3.51321876e-01 -3.60622955e-03 4.25303519e-01
1.23025022e-01 1.08668971e+00 6.07627153e-01 -8.82255018e-01
3.67597520e-01 3.67602319e-01 5.60389042e-01 -2.20319197e-01
-7.61430323e-01 -3.62238228e-01 -9.20711040e-01 -1.36375308e-01
-3.06458399e-02 -5.12404025e-01 -6.01265073e-01 1.40039253e+00
1.27185881e-01 -5.51306546e-01 -1.11693554e-01 4.29682106e-01
-4.21236783e-01 1.44359660e+00 1.90666690e-01 -7.48333216e-01
1.30751109e+00 -6.37956977e-01 -6.49102747e-01 2.55272478e-01
8.19573283e-01 -1.75680339e-01 4.45095003e-01 -3.08563560e-01
-1.84555662e+00 -5.32301739e-02 -8.46872091e-01 1.98443994e-01
-9.80428219e-01 -8.19116011e-02 -1.29507363e-01 6.27504528e-01
-1.15422976e+00 6.86837733e-01 -1.22482765e+00 -5.71450830e-01
-1.12388469e-01 1.79802217e-02 4.92564261e-01 4.26791131e-01
-1.39009452e+00 1.65066123e+00 5.97048581e-01 5.55718720e-01
-3.75745177e-01 -8.05060267e-01 -7.13477015e-01 8.40781808e-01
-3.36692929e-01 -6.00095093e-01 1.06713724e+00 2.04176784e-01
-1.02548873e+00 -3.67507994e-01 4.35928702e-02 -4.78910267e-01
3.64664406e-01 6.58272147e-01 -5.65076113e-01 -1.32701442e-01
3.23966965e-02 1.80437714e-01 -3.79075974e-01 -1.02765787e+00
-7.07272768e-01 -2.53861934e-01 -4.66310382e-01 5.88899910e-01
-1.96128935e-01 -1.74562693e-01 7.90481269e-01 -1.44057781e-01
-4.19764936e-01 -1.03090262e+00 -3.03697139e-01 -1.01977932e+00
-5.19936562e-01 -6.18052781e-01 8.53455126e-01 -9.89671171e-01
1.32232893e+00 -1.64253056e+00 9.24259201e-02 9.28245842e-01
-8.18652213e-01 -4.22540277e-01 5.93318701e-01 1.06567824e+00
-4.09133583e-01 3.92319918e-01 -4.81984347e-01 -7.76564411e-04
7.44556248e-01 5.82692802e-01 -5.92619553e-02 3.23141068e-01
-1.59090966e-01 6.73701584e-01 -9.96729374e-01 1.22628927e-01
7.37619162e-01 1.13129809e-01 6.75531775e-02 -1.85862735e-01
-4.32998203e-02 2.36236881e-02 2.75486112e-02 3.88337016e-01
7.24196672e-01 1.44335255e-01 3.81562203e-01 7.73334801e-02
-8.75824511e-01 2.60740936e-01 -1.03660524e+00 1.17782414e+00
-9.55056787e-01 4.10881907e-01 1.40304312e-01 -1.19931841e+00
3.37413102e-01 1.79857820e-01 6.51420712e-01 -1.13835073e+00
-5.56318462e-01 7.56466389e-01 1.53261527e-01 -1.48565248e-01
1.02118278e+00 -4.71419543e-01 1.71636507e-01 7.10737944e-01
-3.82387370e-01 -5.77146232e-01 6.28266752e-01 4.39115837e-02
6.44585490e-01 -3.07923168e-01 -4.50419277e-01 -1.23541903e+00
1.09785408e-01 -2.47128345e-02 3.91375184e-01 -2.70085335e-01
3.17685962e-01 -3.20669711e-01 2.87831545e-01 2.96925813e-01
-1.52426171e+00 -9.96249080e-01 -5.42223454e-01 5.59535325e-01
1.10216595e-01 3.32155153e-02 -4.53009933e-01 2.25459114e-01
2.89816409e-01 1.96573591e+00 -1.96599409e-01 2.83947904e-02
-7.62948215e-01 -1.48918819e+00 -2.70460159e-01 5.29342234e-01
5.20133853e-01 -5.23211777e-01 -1.06236684e+00 1.41576082e-01
-8.26205760e-02 -3.04761350e-01 -1.34511217e-01 4.87716883e-01
-5.26194632e-01 -6.33884609e-01 -1.20024645e+00 -9.91088822e-02
1.07137406e+00 -1.85712427e-01 9.19268906e-01 1.00334935e-01
-2.43663341e-01 2.70657450e-01 1.91264942e-01 -4.98891443e-01
-1.50563911e-01 3.24151456e-01 -3.92079726e-02 -1.16083062e+00
-4.91799600e-02 -3.05391014e-01 -1.14584959e+00 5.09293139e-01
-6.68100357e-01 1.20043144e-01 -4.83791158e-02 5.93269110e-01
3.14062685e-01 1.24627113e+00 8.74204338e-01 -1.22295707e-01
6.44973338e-01 -1.00577903e+00 -6.87495589e-01 3.97007525e-01
-1.21654439e+00 -2.48308420e-01 2.92259336e-01 5.25532663e-01
-1.36722636e+00 -3.29040796e-01 3.52982044e-01 5.06569624e-01
1.75742462e-01 6.39411271e-01 -1.67055666e-01 5.81257164e-01
-1.60502225e-01 2.10332662e-01 -1.75761342e-01 -5.36497653e-01
7.74010196e-02 2.84378827e-01 1.05419874e-01 -6.94246769e-01
9.69239533e-01 -9.12156999e-02 4.68180001e-01 -7.55929172e-01
6.38072848e-01 -1.87143534e-01 -3.74171674e-01 -2.34667733e-01
7.56517112e-01 -1.14951396e+00 -6.07813835e-01 3.63821536e-01
-4.27582443e-01 -1.05450666e+00 -8.49030077e-01 5.03112793e-01
-5.00858128e-01 -1.33205885e-02 -7.97626302e-02 -1.32572544e+00
-2.23723471e-01 -1.25014949e+00 3.74772221e-01 2.76194990e-01
-6.83842674e-02 -1.05390966e+00 1.40720397e-01 -1.16692923e-01
8.69905889e-01 4.86494184e-01 1.20614755e+00 4.73963805e-02
-5.49609244e-01 2.51481622e-01 3.95940989e-02 4.18271303e-01
4.16482538e-01 1.72307268e-01 -6.16462469e-01 -1.03288257e+00
-3.93047661e-01 1.67526379e-01 3.31928790e-01 3.75722706e-01
9.40827966e-01 -5.72760224e-01 -6.30281091e-01 -2.29472846e-01
1.98096168e+00 5.87396979e-01 5.67915618e-01 3.72593969e-01
-6.43521324e-02 8.84314775e-01 2.57731408e-01 6.95108473e-01
6.20457351e-01 4.78103846e-01 3.59024793e-01 -2.66079426e-01
3.48286033e-01 2.72456035e-02 1.50500476e-01 8.17303717e-01
-2.72619903e-01 -3.29795629e-01 -1.05776894e+00 1.28600216e+00
-1.42479134e+00 -1.17771137e+00 -1.52856961e-01 2.44397140e+00
9.04617548e-01 -2.45399028e-01 4.84497190e-01 1.93165034e-01
4.92785454e-01 -2.28261389e-02 -4.39291537e-01 -9.58147109e-01
-1.93413422e-02 1.31043687e-01 1.22797096e+00 5.08876979e-01
1.99268684e-01 -4.38846320e-01 5.78283930e+00 7.06003428e-01
-6.60409689e-01 1.39379293e-01 1.17277372e+00 -6.60558164e-01
-8.16695929e-01 1.99678004e-01 -5.35828829e-01 1.28710985e+00
1.89716291e+00 -1.10905325e+00 6.21645093e-01 2.33155027e-01
9.87033665e-01 -1.03146613e+00 -9.37513173e-01 -7.36709461e-02
-7.91272163e-01 -1.12524724e+00 -6.39171302e-01 5.63670993e-01
1.01197994e+00 3.18008363e-01 -4.66775924e-01 1.23668097e-01
4.62900072e-01 -5.46711922e-01 6.54266357e-01 6.01643682e-01
5.97523332e-01 -1.35227251e+00 8.61102045e-01 5.07316291e-01
-1.63807106e+00 -2.98348427e-01 -2.44441360e-01 -4.00602147e-02
8.97644281e-01 7.47398496e-01 -5.16550481e-01 9.89270031e-01
1.03538978e+00 4.69409823e-02 -1.97592586e-01 7.32454181e-01
1.43303558e-01 3.89428556e-01 -1.17533600e+00 -8.72940570e-02
2.64822274e-01 -6.47964239e-01 -2.09656149e-01 1.06055999e+00
8.68342221e-01 3.30871403e-01 -3.52002382e-01 9.54372048e-01
3.52276176e-01 -2.99217552e-01 -4.92068350e-01 2.37369806e-01
9.95322347e-01 1.10563016e+00 -7.87175000e-01 -6.20549381e-01
-4.62113507e-02 2.70584047e-01 -8.23314190e-01 5.28628707e-01
-5.61797142e-01 -7.81497061e-01 2.92261422e-01 5.35155416e-01
-1.31803960e-01 -2.03225419e-01 -3.98911148e-01 -6.93651021e-01
-7.86554366e-02 4.69294101e-01 6.43289015e-02 -9.16440785e-01
-1.00112689e+00 -2.32768729e-01 1.07030463e+00 -6.21084571e-01
-7.55311430e-01 6.07596561e-02 -1.11566925e+00 1.93162453e+00
-1.88082886e+00 -7.78255701e-01 1.45684198e-01 2.37617910e-01
4.19737160e-01 3.56598228e-01 7.02955246e-01 4.28954810e-01
-1.00407851e+00 -2.94841416e-02 1.11330485e+00 -4.63011742e-01
-1.82390794e-01 -1.64386988e+00 2.14651808e-01 9.48026478e-01
-1.12171805e+00 3.59879076e-01 5.24620771e-01 -6.92762196e-01
-1.35674250e+00 -9.42104936e-01 8.85076463e-01 1.41142756e-01
7.06776977e-01 -1.43670440e-01 -7.35079408e-01 6.40652955e-01
1.53858912e+00 -6.52930081e-01 6.79807484e-01 -2.75946140e-01
7.23353863e-01 -1.00591242e-01 -1.64309275e+00 2.50644445e-01
2.49410525e-01 -2.45361194e-01 -1.76088884e-01 6.37773514e-01
1.68460887e-02 -7.82019943e-02 -1.58585501e+00 1.87868595e-01
4.47280765e-01 -3.01909924e-01 4.17380184e-01 2.60834575e-01
-2.16704592e-01 -1.88109517e-01 -3.84187177e-02 -2.13436604e+00
-5.66738188e-01 -5.65284073e-01 3.29079218e-02 1.40286326e+00
6.63242459e-01 -1.07390547e+00 3.11108589e-01 1.48191249e+00
-5.73484540e-01 -6.22245133e-01 -1.52112532e+00 -9.62157667e-01
8.04974616e-01 3.68497133e-01 1.46405327e+00 1.15147781e+00
7.28229225e-01 -1.44608706e-01 3.79629314e-01 1.33997068e-01
6.38395071e-01 -1.79607179e-02 -1.06105149e-01 -7.35166073e-01
3.11403781e-01 -7.77725756e-01 4.56201673e-01 -1.12845458e-01
-6.78040460e-02 -7.26957560e-01 -2.45333374e-01 -2.25869131e+00
2.31991298e-02 -7.89417028e-01 -6.85758814e-02 7.62785733e-01
5.81704915e-01 -3.32896084e-01 4.52666879e-01 2.14706421e-01
9.01630461e-01 9.97592151e-01 6.83688283e-01 -3.21449488e-01
-2.05227688e-01 1.46735013e-02 -3.88387451e-03 1.65060118e-01
1.14129424e+00 2.78730839e-01 -7.03259945e-01 -3.08418810e-01
3.73555392e-01 -4.33657542e-02 4.16213013e-02 -9.65723515e-01
1.01083465e-01 -5.40561318e-01 4.34597641e-01 -1.11649656e+00
-1.02812655e-01 -1.45321858e+00 1.35654020e+00 9.52239156e-01
8.36094245e-02 4.45112199e-01 4.28316921e-01 2.28844628e-01
3.67334992e-01 -1.51787981e-01 5.55188239e-01 -3.26513559e-01
-2.48460084e-01 -4.71680433e-01 -5.78898609e-01 -8.99159074e-01
1.57330632e+00 -2.36688405e-01 -1.01130784e+00 5.88833280e-02
-8.64059269e-01 1.14256823e+00 7.16283679e-01 -3.83619070e-02
-1.93049803e-01 -1.48018968e+00 -4.07952338e-01 3.37855928e-02
-5.12201309e-01 1.00818790e-01 2.65182197e-01 7.74690092e-01
-5.63857079e-01 6.59943581e-01 -3.62041891e-01 -1.67782322e-01
-4.37142521e-01 3.00731122e-01 8.40531051e-01 -1.42763659e-01
-3.57643664e-01 6.59568086e-02 -3.66617322e-01 1.35890022e-01
-4.88054574e-01 -9.23599124e-01 4.43427920e-01 6.84619963e-01
9.35614184e-02 1.20888734e+00 3.94174874e-01 -4.96241480e-01
-1.84334770e-01 1.59716502e-01 6.03202581e-01 -1.24598593e-01
1.26213074e+00 -3.70989710e-01 -4.96516556e-01 5.69136024e-01
9.92535233e-01 -5.33785701e-01 -7.22488523e-01 5.28110266e-01
-4.03372288e-01 -1.11730918e-01 2.92466789e-01 -1.23531330e+00
-1.14024878e+00 4.35535997e-01 4.78951246e-01 1.23057473e+00
1.50744677e+00 -3.25007021e-01 5.16350627e-01 -3.32196057e-02
4.66425240e-01 -1.80830133e+00 -1.13936830e+00 -1.78505048e-01
9.72074091e-01 -3.76194328e-01 6.05310321e-01 3.78293723e-01
2.03302763e-02 7.22797513e-01 2.50790000e-01 1.88556567e-01
9.78747427e-01 9.61696133e-02 -1.21946740e+00 2.05171719e-01
-5.17683148e-01 9.76882577e-02 -6.39871180e-01 1.10819124e-01
1.07289195e-01 7.70912230e-01 -7.79924035e-01 -1.37092203e-01
-1.21443294e-01 -3.27400386e-01 5.64652741e-01 1.23837459e+00
-4.25816476e-01 -7.05058634e-01 -2.66161144e-01 8.28804076e-01
1.11513689e-01 7.30939135e-02 4.71732825e-01 9.08843875e-01
2.29621828e-01 7.59452045e-01 9.52372074e-01 3.46887678e-01
7.09028780e-01 1.09250523e-01 5.40048331e-02 -3.02160144e-01
-5.48798561e-01 3.07861269e-01 4.12127048e-01 -1.08671254e-02
-4.63570267e-01 -1.06776679e+00 -1.63531148e+00 -8.37141335e-01
-8.44436586e-01 9.24249947e-01 1.22237873e+00 6.27236545e-01
3.60245913e-01 5.47927856e-01 1.24898434e+00 -1.40485919e+00
-5.49119115e-01 -1.12495756e+00 -1.43947244e+00 -5.58313787e-01
-1.87608146e-03 -5.44257879e-01 -8.35150719e-01 -4.77972984e-01] | [5.72064733505249, 2.486905574798584] |
7922f0e7-d521-4fd8-81da-3598446f2487 | multi-task-collaborative-pre-training-and | 2306.11378 | null | https://arxiv.org/abs/2306.11378v1 | https://arxiv.org/pdf/2306.11378v1.pdf | Multi-task Collaborative Pre-training and Individual-adaptive-tokens Fine-tuning: A Unified Framework for Brain Representation Learning | Structural magnetic resonance imaging (sMRI) provides accurate estimates of the brain's structural organization and learning invariant brain representations from sMRI is an enduring issue in neuroscience. Previous deep representation learning models ignore the fact that the brain, as the core of human cognitive activity, is distinct from other organs whose primary attribute is anatomy. Therefore, capturing the semantic structure that dominates interindividual cognitive variability is key to accurately representing the brain. Given that this high-level semantic information is subtle, distributed, and interdependently latent in the brain structure, sMRI-based models need to capture fine-grained details and understand how they relate to the overall global structure. However, existing models are optimized by simple objectives, making features collapse into homogeneity and worsening simultaneous representation of fine-grained information and holistic semantics, causing a lack of biological plausibility and interpretation of cognition. Here, we propose MCIAT, a unified framework that combines Multi-task Collaborative pre-training and Individual-Adaptive-Tokens fine-tuning. Specifically, we first synthesize restorative learning, age prediction auxiliary learning and adversarial learning as a joint proxy task for deep semantic representation learning. Then, a mutual-attention-based token selection method is proposed to highlight discriminative features. The proposed MCIAT achieves state-of-the-art diagnosis performance on the ADHD-200 dataset compared with several sMRI-based approaches and shows superior generalization on the MCIC and OASIS datasets. Moreover, we studied 12 behavioral tasks and found significant associations between cognitive functions and MCIAT-established representations, which verifies the interpretability of our proposed framework. | ['Tianyi Yan', 'Gongshu Wang', 'Ning Jiang'] | 2023-06-20 | null | null | null | null | ['auxiliary-learning', 'anatomy'] | ['methodology', 'miscellaneous'] | [ 2.48356432e-01 -9.79767218e-02 -2.07510382e-01 -4.28649098e-01
-4.67651278e-01 -3.68565708e-01 5.81642687e-01 2.36061245e-01
-4.78834450e-01 6.97635412e-01 4.55303311e-01 1.30530357e-01
-7.83967793e-01 -6.38504148e-01 -5.14868200e-01 -6.80983484e-01
-1.43075645e-01 6.32125497e-01 3.69936526e-02 -1.11366391e-01
1.08552970e-01 3.32302511e-01 -1.45753062e+00 3.75053257e-01
1.39242172e+00 1.03370154e+00 4.92047906e-01 7.79565647e-02
-6.24336638e-02 6.97607934e-01 -5.42091012e-01 -2.76626021e-01
-1.69113189e-01 -2.92412192e-01 -8.02048385e-01 -3.23090345e-01
4.28031445e-01 -1.51710704e-01 -3.05823892e-01 1.09691060e+00
6.58581674e-01 1.63688943e-01 9.36672628e-01 -9.81324255e-01
-1.10790086e+00 5.72209299e-01 -4.28966552e-01 6.60201669e-01
-1.53042926e-02 9.54078883e-02 9.55776453e-01 -6.59723103e-01
3.60076070e-01 1.31020474e+00 4.78272051e-01 5.27045488e-01
-1.27726257e+00 -9.04537618e-01 3.95745546e-01 5.90477407e-01
-1.28753734e+00 -1.38055786e-01 6.69332027e-01 -5.80557525e-01
6.06305718e-01 7.11372937e-04 8.66397619e-01 1.33099174e+00
4.20281649e-01 5.36255896e-01 1.50705028e+00 8.87929425e-02
2.73377448e-01 -3.24328989e-01 3.89043659e-01 4.96780783e-01
5.04510522e-01 6.26822785e-02 -3.03999186e-01 -6.35695755e-02
7.61441886e-01 4.19915617e-01 -7.28383660e-02 -1.61002859e-01
-1.29619169e+00 5.01102686e-01 7.15750813e-01 5.76016784e-01
-4.88058448e-01 1.34901226e-01 5.40954709e-01 1.39823435e-02
2.64698595e-01 2.73824096e-01 -3.57579768e-01 2.83098847e-01
-1.06705582e+00 2.33806610e-01 -2.99179815e-02 2.71835238e-01
4.08564121e-01 3.62441212e-01 -5.39746583e-01 1.08951104e+00
1.03269354e-01 6.06665075e-01 1.16102779e+00 -8.71627212e-01
1.74024910e-01 7.69820809e-01 -5.35899878e-01 -8.69620204e-01
-7.40064442e-01 -1.16585588e+00 -8.75195265e-01 6.15456551e-02
5.86779341e-02 3.10474992e-01 -9.87367988e-01 2.21694088e+00
7.34397396e-02 3.53500135e-02 -3.89540464e-01 1.06614470e+00
8.82958055e-01 -9.33656543e-02 6.04704738e-01 -2.55763512e-02
1.65182626e+00 -6.68593884e-01 -4.21783119e-01 -4.44788247e-01
1.91666663e-01 -1.07719764e-01 9.33999002e-01 1.74371645e-01
-9.81160998e-01 -6.82484567e-01 -9.73886430e-01 -1.74955159e-01
-3.42940897e-01 -3.97214256e-02 8.61895442e-01 4.88759428e-01
-9.14976060e-01 3.43028337e-01 -7.42139697e-01 -4.29309867e-02
8.40613902e-01 3.17495704e-01 -4.36894864e-01 -2.14437664e-01
-1.35068715e+00 9.37133908e-01 4.36195940e-01 -1.10971928e-01
-1.11579764e+00 -1.04768991e+00 -5.55992365e-01 4.03919578e-01
2.66698360e-01 -1.18134165e+00 7.39598691e-01 -9.93819535e-01
-1.10378528e+00 8.86872470e-01 -1.21699244e-01 -3.11473221e-01
1.22679032e-01 -8.22686777e-02 -4.89992619e-01 5.03111660e-01
2.08906874e-01 6.63332462e-01 1.04243672e+00 -9.83070731e-01
-6.52139932e-02 -9.65423763e-01 1.27351657e-02 1.71010062e-01
-4.16085184e-01 -1.92822799e-01 1.88565522e-01 -9.29467738e-01
2.76939005e-01 -6.95180893e-01 -1.35331964e-02 4.63737035e-03
-1.80653661e-01 -1.57368079e-01 3.43559176e-01 -8.52663398e-01
9.48950648e-01 -2.11713648e+00 2.74181068e-01 1.37784898e-01
7.38290071e-01 1.73784643e-01 -2.68462569e-01 -2.25037903e-01
-5.44497192e-01 1.60139918e-01 -2.54488438e-01 -5.19683138e-02
3.46545242e-02 -8.54312330e-02 -4.27154042e-02 5.23059607e-01
9.45929289e-02 1.25555539e+00 -9.80261028e-01 -3.56382936e-01
8.02415088e-02 4.82106596e-01 -5.84218025e-01 -9.78036821e-02
8.48215297e-02 6.46324933e-01 -6.28465831e-01 7.35165596e-01
5.64676344e-01 -2.46113226e-01 1.62487909e-01 -3.95031959e-01
2.32155800e-01 2.29145929e-01 -5.99769115e-01 1.86178493e+00
-2.45701432e-01 1.05051547e-01 -4.96097803e-02 -1.55587959e+00
8.11214268e-01 1.94058597e-01 5.14345706e-01 -1.21335125e+00
2.63151467e-01 1.68726325e-01 5.06982386e-01 -2.80488849e-01
-1.59885302e-01 -2.80002445e-01 5.02825864e-02 4.58636731e-01
3.24089110e-01 3.48625660e-01 -1.30031392e-01 8.11201558e-02
1.07514477e+00 4.47659334e-03 4.71591204e-01 -3.61918181e-01
5.09902596e-01 -5.17631352e-01 7.55840778e-01 4.86093938e-01
-3.98040354e-01 6.48425341e-01 3.61618549e-01 -1.97149500e-01
-7.73907065e-01 -1.50972009e+00 -4.21658337e-01 1.31557977e+00
-1.59665853e-01 -3.43398266e-02 -6.97834551e-01 -3.60034585e-01
1.04995370e-01 6.23656154e-01 -1.05468357e+00 -7.44627178e-01
-2.30707631e-01 -8.97194445e-01 5.47814131e-01 7.60005474e-01
4.94433671e-01 -1.07981718e+00 -4.92145836e-01 -3.61327059e-03
-2.03579426e-01 -9.39393818e-01 -3.46890837e-01 8.40863734e-02
-9.98665929e-01 -1.15048087e+00 -8.10913086e-01 -4.70140487e-01
6.52435958e-01 3.06852251e-01 1.09847963e+00 6.52731135e-02
-6.51234806e-01 3.35003674e-01 -1.14347816e-01 -9.35762599e-02
1.34254977e-01 1.99805498e-01 2.73851696e-02 -5.03560230e-02
8.95248652e-02 -1.06277657e+00 -9.99052465e-01 9.25107151e-02
-9.95556772e-01 -9.88367349e-02 1.01518333e+00 9.34321702e-01
7.61703014e-01 -1.71643734e-01 1.04831827e+00 -7.69467652e-01
5.05461812e-01 -7.63954759e-01 1.46500230e-01 3.94833148e-01
-6.08734429e-01 1.21780641e-01 5.91706753e-01 -4.43755925e-01
-1.17661524e+00 -4.25934017e-01 -1.93050746e-02 -3.72112960e-01
-3.57294023e-01 6.20262206e-01 -3.51332247e-01 3.61559391e-02
4.75785583e-01 5.63994765e-01 3.91217805e-02 -3.97009701e-01
3.05320233e-01 1.86340481e-01 6.63453698e-01 -9.11542833e-01
5.45774639e-01 4.95467663e-01 -5.26595078e-02 -6.26991808e-01
-9.83988047e-01 -1.85772061e-01 -7.11415231e-01 1.19646795e-01
8.96937370e-01 -1.00193346e+00 -7.68870115e-01 3.88989568e-01
-8.08981597e-01 -6.82240501e-02 -2.31757358e-01 5.11627793e-01
-4.95738626e-01 2.75964379e-01 -4.70276177e-01 -4.41767901e-01
-7.31640816e-01 -1.07430327e+00 8.88566077e-01 3.01684868e-02
-2.82293677e-01 -9.41341221e-01 4.98966202e-02 7.12365091e-01
5.64508021e-01 1.49633542e-01 1.45974994e+00 -7.85071731e-01
-2.02822268e-01 1.41582116e-01 -5.46332777e-01 2.61164844e-01
5.21113724e-02 -5.86467564e-01 -1.00061655e+00 -1.67637020e-01
6.63769022e-02 -4.60802823e-01 1.16447496e+00 5.96158504e-01
1.62714100e+00 4.48802672e-02 -7.67886564e-02 6.77916110e-01
1.21232855e+00 7.84621201e-03 7.21371055e-01 2.33833790e-01
7.10138738e-01 6.20041668e-01 -1.86404921e-02 2.90997803e-01
6.48342967e-01 3.84722263e-01 3.85669649e-01 6.18468933e-02
-3.11699569e-01 -2.25210190e-03 2.50442863e-01 7.12419271e-01
-2.95660108e-01 3.46128523e-01 -7.18876779e-01 4.44998145e-01
-1.57624614e+00 -1.17857575e+00 2.34035239e-01 2.11386347e+00
6.79034591e-01 6.91557601e-02 5.65875284e-02 3.07743941e-02
7.68244565e-01 1.73255309e-01 -1.08396447e+00 -1.19283520e-01
-2.20846802e-01 5.23914874e-01 1.41117245e-01 -1.91365868e-01
-6.68128192e-01 7.06087887e-01 6.15700865e+00 8.91596258e-01
-1.20846164e+00 6.50150836e-01 6.29671752e-01 -3.17313731e-01
-5.31361043e-01 -5.49512088e-01 -2.85277903e-01 5.11657357e-01
8.65380049e-01 -2.46767297e-01 6.54686570e-01 6.40806496e-01
2.09754616e-01 2.53051311e-01 -8.45537663e-01 8.18286777e-01
1.81826696e-01 -9.90774214e-01 2.52942473e-01 -8.46957881e-03
5.47079980e-01 3.48336250e-02 5.23259521e-01 4.60287780e-01
-1.09718561e-01 -1.36732817e+00 8.17297995e-01 8.87911618e-01
9.93196368e-01 -7.17328727e-01 4.38583940e-01 2.61328351e-02
-1.18108678e+00 -3.43812168e-01 -3.69519502e-01 -6.94627222e-03
-1.64154738e-01 7.73195744e-01 -2.64037997e-01 5.04202962e-01
6.15015090e-01 6.49136245e-01 -7.68892765e-01 9.62862551e-01
-2.47008994e-01 4.44216907e-01 1.92053065e-01 4.45027083e-01
-1.12498879e-01 3.44780758e-02 3.37670505e-01 8.75452459e-01
2.71591008e-01 1.09710790e-01 1.38146132e-01 1.33231497e+00
-1.36696592e-01 5.61986417e-02 -2.23193049e-01 -2.02507779e-01
4.95612383e-01 1.34198105e+00 -7.83116698e-01 -3.20583284e-01
-3.37420672e-01 6.85524762e-01 6.16838634e-01 3.87346208e-01
-7.66941309e-01 -3.21001187e-02 7.58248389e-01 5.70774451e-02
2.19744712e-01 -2.14849204e-01 -4.62026834e-01 -1.11573601e+00
-1.31233782e-01 -8.66044700e-01 4.38043475e-01 -6.96586013e-01
-1.42313707e+00 4.62215394e-01 -1.39684051e-01 -7.70000875e-01
-5.90451881e-02 -4.90306109e-01 -6.43469930e-01 8.88200819e-01
-1.63305855e+00 -1.40908206e+00 -3.90686691e-01 7.79700220e-01
3.88230890e-01 -1.28856987e-01 8.85723889e-01 4.46615368e-01
-7.33400822e-01 6.59482300e-01 1.72837172e-02 4.51502353e-02
6.03772581e-01 -1.11014366e+00 -1.64604485e-01 5.44264436e-01
-1.65361091e-01 9.75645244e-01 1.77434266e-01 -6.90587401e-01
-1.02998173e+00 -1.13021338e+00 3.75287682e-01 -2.56574363e-01
7.11736202e-01 -3.09892088e-01 -1.28201914e+00 4.98588383e-01
-2.59167314e-01 1.39413401e-01 8.86884630e-01 2.87799180e-01
-6.89602375e-01 -2.31334090e-01 -1.17530382e+00 2.92508513e-01
1.44392061e+00 -7.93404996e-01 -1.07099938e+00 9.45151225e-02
7.43011117e-01 8.99489522e-02 -1.06550300e+00 3.47402394e-01
8.36208940e-01 -8.17826271e-01 1.46363485e+00 -6.09343052e-01
5.13102651e-01 -3.98419937e-03 -7.04991212e-03 -1.46494532e+00
-8.05356324e-01 2.99086332e-01 -1.12858437e-01 1.06387877e+00
-1.45431319e-02 -7.10047066e-01 3.85627449e-01 4.91088569e-01
-4.41712528e-01 -6.82214320e-01 -8.54575098e-01 -7.54934549e-01
3.64417791e-01 -2.83483744e-01 7.43511558e-01 1.07936728e+00
-2.06234097e-01 2.44134694e-01 9.66720690e-04 -2.58623506e-03
8.25738192e-01 1.63759932e-01 -8.79977793e-02 -1.61653388e+00
-9.69614163e-02 -7.33708560e-01 -4.61891115e-01 -1.45661354e-01
6.58889592e-01 -1.42853689e+00 -5.39974988e-01 -1.53198075e+00
7.50230253e-01 -2.06931353e-01 -9.68577623e-01 5.15371382e-01
-3.21906030e-01 1.23228922e-01 6.78029098e-03 1.62648097e-01
-6.34573162e-01 8.12569916e-01 1.48874283e+00 -3.07917923e-01
2.02447787e-01 -4.80779707e-01 -1.11842549e+00 5.97391427e-01
7.73001611e-01 -2.85762340e-01 -7.17683494e-01 -3.60690653e-01
6.01142691e-03 -2.45907038e-01 7.73069024e-01 -1.13127518e+00
-1.28612639e-02 -2.01197341e-01 8.66718292e-01 -1.81595922e-01
2.31526628e-01 -4.99922484e-01 -2.48517022e-02 6.78763390e-01
-4.29141968e-01 -1.07701890e-01 -1.11060105e-02 4.02343184e-01
-6.80878535e-02 -1.43626267e-02 8.19972634e-01 -3.03197622e-01
-7.71228194e-01 6.45629108e-01 -2.63178349e-01 4.39334273e-01
6.01869047e-01 -2.29016230e-01 -5.43282568e-01 1.41736686e-01
-8.03406775e-01 -1.33392513e-01 1.11591361e-01 4.53401655e-01
6.42913222e-01 -1.26294744e+00 -6.62071466e-01 5.26084267e-02
-3.26923281e-02 -3.18879157e-01 8.85536075e-01 1.04657650e+00
1.08090416e-01 5.26942730e-01 -8.28026891e-01 -4.10671532e-01
-6.61782205e-01 5.33972859e-01 2.14017600e-01 -2.34318897e-01
-6.19611144e-01 5.07409036e-01 8.71238589e-01 -3.01252425e-01
-1.65509045e-01 -3.25106949e-01 -5.49529791e-01 2.87794143e-01
6.64969683e-01 4.64298785e-01 1.68677807e-01 -5.23045421e-01
-4.74799365e-01 3.71505797e-01 -1.90834537e-01 3.35043609e-01
1.52495158e+00 -1.65688097e-01 -2.65746117e-01 4.29927558e-01
9.27441120e-01 -3.86072248e-01 -9.81952906e-01 -2.88476646e-01
-1.30977258e-01 -1.00708775e-01 2.07062453e-01 -9.76612151e-01
-1.40127397e+00 1.07912683e+00 9.01994884e-01 -4.77421075e-01
1.09034896e+00 1.68670624e-01 7.29364455e-01 1.27900019e-01
4.58029598e-01 -1.07862651e+00 4.08958197e-01 4.08049494e-01
9.08367455e-01 -9.74708498e-01 5.04196063e-03 -1.14059947e-01
-4.97864902e-01 7.88607478e-01 8.50309551e-01 -1.83692813e-01
3.84886771e-01 -2.36231357e-01 -4.46213245e-01 -3.31101030e-01
-5.79620957e-01 -2.88753539e-01 6.11906826e-01 7.55029202e-01
5.16894817e-01 3.36063504e-01 -4.10682052e-01 1.41776419e+00
-2.19749928e-01 -8.35546777e-02 -9.67157260e-02 3.90992671e-01
-6.34797335e-01 -7.84569204e-01 -3.45040113e-01 8.18522036e-01
-5.26198983e-01 -2.68606782e-01 -5.34680486e-02 5.72944999e-01
3.36542696e-01 3.06605548e-01 5.96806258e-02 -1.38453856e-01
7.80089647e-02 3.68064582e-01 6.61185741e-01 -6.96891904e-01
-4.06032503e-01 -2.55386710e-01 -3.49946052e-01 -6.45728946e-01
-2.89195418e-01 -6.41158819e-01 -1.52936101e+00 -1.61435768e-01
2.82013714e-01 -3.18519890e-01 4.46854740e-01 1.36345387e+00
5.06148994e-01 1.12681663e+00 2.73727357e-01 -6.72777295e-01
-4.76162374e-01 -9.17098343e-01 -7.39468336e-01 6.23886347e-01
9.57420766e-02 -1.32499695e+00 -7.37150237e-02 -1.52808771e-01] | [12.501019477844238, 3.3158321380615234] |
d8c1fa86-b585-44aa-821e-0b87f0f58175 | label-dependencies-aware-set-prediction | 2304.07022 | null | https://arxiv.org/abs/2304.07022v1 | https://arxiv.org/pdf/2304.07022v1.pdf | Label Dependencies-aware Set Prediction Networks for Multi-label Text Classification | Multi-label text classification aims to extract all the related labels from a sentence, which can be viewed as a sequence generation problem. However, the labels in training dataset are unordered. We propose to treat it as a direct set prediction problem and don't need to consider the order of labels. Besides, in order to model the correlation between labels, the adjacency matrix is constructed through the statistical relations between labels and GCN is employed to learn the label information. Based on the learned label information, the set prediction networks can both utilize the sentence information and label information for multi-label text classification simultaneously. Furthermore, the Bhattacharyya distance is imposed on the output probability distributions of the set prediction networks to increase the recall ability. Experimental results on four multi-label datasets show the effectiveness of the proposed method and it outperforms previous method a substantial margin. | ['Lv Chao', 'Sun Yalin', 'Du Xinkai', 'Han Quanjie'] | 2023-04-14 | null | null | null | null | ['multi-label-text-classification', 'multi-label-text-classification'] | ['methodology', 'natural-language-processing'] | [ 6.84360385e-01 1.93957798e-02 -3.14995855e-01 -7.61615694e-01
-3.70042562e-01 -6.88531220e-01 2.83817351e-01 9.92452130e-02
-2.76105791e-01 7.50539660e-01 1.52963459e-01 -7.87086189e-02
-3.19612026e-01 -8.71831954e-01 -2.05526203e-01 -8.20207119e-01
5.17704904e-01 4.78082508e-01 4.26276959e-02 -1.96195975e-01
4.73565549e-01 -4.52888682e-02 -1.43011725e+00 5.51162183e-01
8.23718727e-01 9.67862070e-01 3.08292478e-01 2.16431394e-01
-3.97878677e-01 9.84453022e-01 -3.91980916e-01 -1.27686203e-01
3.03058382e-02 -6.88579619e-01 -1.02927184e+00 3.70192438e-01
8.93202126e-02 4.06689085e-02 -1.69326365e-01 1.29722941e+00
2.23368019e-01 2.74114043e-01 1.04755318e+00 -1.35515010e+00
-3.83430034e-01 9.26092029e-01 -7.29504526e-01 -1.28645971e-01
1.54235512e-01 -3.51835370e-01 1.29292226e+00 -6.63813651e-01
5.21517873e-01 1.32559288e+00 4.98771757e-01 5.06504059e-01
-9.27121401e-01 -7.54607975e-01 2.58872956e-01 3.30320597e-01
-1.30470395e+00 6.38648942e-02 9.62833643e-01 -3.50275159e-01
3.41722041e-01 3.53399992e-01 2.57146984e-01 6.41028106e-01
-1.49472669e-01 8.64243090e-01 1.22894359e+00 -6.88752294e-01
-1.49264812e-01 2.37479419e-01 7.26844311e-01 9.70261037e-01
-1.42582312e-01 -3.71601582e-01 -3.25666875e-01 -1.74168944e-01
2.76237667e-01 2.23597512e-01 -2.83456355e-01 -1.10916838e-01
-1.19022322e+00 8.19976628e-01 3.33403081e-01 3.39337677e-01
1.01248018e-01 -7.69894272e-02 6.70546710e-01 2.41589159e-01
5.55383742e-01 6.98227584e-02 -4.99809474e-01 3.88537526e-01
-5.67446351e-01 -3.24297458e-01 6.77151740e-01 1.30205083e+00
1.12753069e+00 -3.64706010e-01 -2.58161992e-01 1.17191195e+00
6.38782203e-01 2.66166210e-01 5.34528732e-01 -5.08684099e-01
8.25490832e-01 8.82790387e-01 -2.85415620e-01 -1.14511693e+00
-6.57489181e-01 -3.05240780e-01 -1.24691296e+00 -4.99687374e-01
2.67639130e-01 -3.10486019e-01 -7.44799137e-01 1.57277322e+00
2.91091084e-01 4.58699971e-01 2.31864844e-02 7.05940843e-01
7.45942414e-01 8.71700883e-01 -4.25708713e-04 -7.50232458e-01
1.26940727e+00 -1.07890952e+00 -6.83636844e-01 -7.96492323e-02
1.16606987e+00 -5.19210696e-01 8.52295816e-01 1.36198416e-01
-4.22433704e-01 -6.51530743e-01 -9.27064061e-01 1.65282771e-01
-1.55042946e-01 4.09200430e-01 4.73584741e-01 4.51495290e-01
-6.81457818e-01 4.64131445e-01 -2.75364131e-01 -1.56093687e-01
2.17533097e-01 5.92793226e-01 -8.25543851e-02 -2.63616145e-01
-1.45500457e+00 5.59045434e-01 1.22128928e+00 1.73992172e-01
-2.63796061e-01 -1.72335893e-01 -6.47461832e-01 1.84334502e-01
5.39420187e-01 -3.51854295e-01 7.95326829e-01 -8.92927825e-01
-1.18063009e+00 7.50411630e-01 -2.27769613e-01 7.75713995e-02
1.25979602e-01 4.76968914e-01 -4.81242746e-01 1.89508677e-01
1.19365104e-01 5.00202358e-01 3.39055806e-01 -1.36122322e+00
-1.02035010e+00 -1.63294643e-01 -1.16087958e-01 5.81606448e-01
-6.33490384e-01 2.74040643e-03 -2.65152454e-01 -4.72021163e-01
4.80209649e-01 -1.11089969e+00 -2.00535566e-01 -6.43630207e-01
-7.95648098e-01 -5.01558781e-01 7.39578426e-01 -4.33725238e-01
1.32098913e+00 -2.01874423e+00 -1.34332422e-02 6.50732040e-01
1.46170720e-01 -7.94586316e-02 -1.55061528e-01 4.77860719e-01
-5.43938465e-02 3.06315094e-01 -2.97653377e-01 -9.54771489e-02
-9.04566050e-02 1.44722924e-01 -1.57939166e-01 2.54137427e-01
-1.35331556e-01 5.99975824e-01 -9.08619821e-01 -1.09394097e+00
2.85290722e-02 -5.73997982e-02 -1.09035686e-01 2.51060277e-02
-3.18914801e-01 4.86190677e-01 -7.92464435e-01 1.15289316e-01
6.12836361e-01 -6.36835039e-01 5.68493128e-01 -3.06774408e-01
3.28617096e-01 9.88641307e-02 -1.22141051e+00 1.51538837e+00
-4.06835049e-01 8.41164440e-02 -5.50717235e-01 -1.10049355e+00
1.15661407e+00 2.99377948e-01 6.01019800e-01 -2.43495002e-01
3.13244492e-01 -2.22970340e-02 8.85349587e-02 -5.45511901e-01
1.59411341e-01 -2.99853534e-01 -2.74799615e-01 8.74823272e-01
-5.37619554e-02 1.99239716e-01 3.91249239e-01 2.13363662e-01
7.04636037e-01 -1.15349077e-01 2.62712330e-01 -2.83352852e-01
8.72085392e-01 -9.10908133e-02 8.09970737e-01 4.85347360e-01
2.92486727e-01 3.14342499e-01 4.42584872e-01 -2.35848069e-01
-7.52827942e-01 -4.10494983e-01 4.05142829e-02 1.29947281e+00
4.16414768e-01 -3.74234438e-01 -5.39609969e-01 -1.29935408e+00
-2.64159620e-01 6.85764909e-01 -4.80073810e-01 -2.43502691e-01
-4.36723918e-01 -9.61128473e-01 4.09982532e-01 2.50499070e-01
6.69310212e-01 -9.93202984e-01 3.12305897e-01 1.68535799e-01
-4.58392620e-01 -7.73992777e-01 -8.34549010e-01 1.56587571e-01
-6.64450467e-01 -1.07205760e+00 -2.21827284e-01 -1.31483006e+00
9.01018679e-01 2.55191594e-01 7.14658320e-01 4.71815795e-01
1.85702965e-01 -1.88888192e-01 -5.29992223e-01 -7.54295588e-02
-6.12982154e-01 3.42876464e-01 -5.06032296e-02 2.98297584e-01
4.68288451e-01 -5.83905816e-01 -1.32942498e-01 4.89247322e-01
-9.07256544e-01 4.28568184e-01 4.39831883e-01 1.15711129e+00
4.77275968e-01 6.11398041e-01 1.06673372e+00 -1.36591685e+00
6.32435501e-01 -4.99530971e-01 -4.54613179e-01 8.31394613e-01
-8.33306253e-01 2.49045894e-01 8.23654711e-01 -3.60191554e-01
-1.22595131e+00 3.30638260e-01 3.27950791e-02 5.05571589e-02
-1.35687262e-01 6.70777142e-01 -3.97309870e-01 1.50328428e-01
9.44203362e-02 4.49855685e-01 -1.00381576e-01 -3.74716222e-01
3.60401273e-01 1.16985369e+00 -6.80587888e-02 -2.99701154e-01
4.54122096e-01 8.02423134e-02 4.55502838e-01 -2.33544618e-01
-1.46954381e+00 -6.25337005e-01 -1.10712922e+00 -3.42025101e-01
7.96539724e-01 -6.00961506e-01 -1.06052351e+00 3.96321356e-01
-1.06479514e+00 1.12614721e-01 3.22651416e-01 4.25697953e-01
-3.12693954e-01 6.36638403e-01 -9.61276233e-01 -7.45390356e-01
-2.62980640e-01 -9.73772407e-01 7.90095210e-01 1.76249191e-01
9.77339596e-02 -1.41009939e+00 -3.45029771e-01 3.65543902e-01
-4.07299876e-01 2.46085171e-02 1.46579039e+00 -1.08697736e+00
-4.46088254e-01 -5.84508467e-04 -3.30685318e-01 3.83869737e-01
4.19929415e-01 -3.37622792e-01 -7.27900505e-01 -2.93308139e-01
1.75423607e-01 -4.57149059e-01 9.03448522e-01 -1.38185352e-01
1.20010054e+00 -5.12031198e-01 -5.97043633e-01 1.10862240e-01
1.47314262e+00 4.61323082e-01 1.96433514e-02 -3.40943187e-02
1.35496080e+00 8.48369122e-01 8.70051086e-01 3.44161808e-01
3.82541776e-01 3.83645684e-01 1.31470516e-01 1.85023189e-01
1.54418916e-01 -3.97058189e-01 -1.07701957e-01 1.38457310e+00
4.75408345e-01 -5.67018509e-01 -9.21221852e-01 -1.29814237e-01
-2.03834438e+00 -8.04404438e-01 -5.75699627e-01 1.90153944e+00
9.84868884e-01 6.45785406e-02 -2.62175463e-02 3.11100006e-01
1.21235740e+00 -8.10951814e-02 -4.98892963e-01 -2.05599349e-02
-1.26255348e-01 -4.47629690e-01 4.08961117e-01 4.46515709e-01
-1.05982363e+00 8.87079000e-01 6.02530718e+00 1.20333445e+00
-6.92169905e-01 1.84744839e-02 9.22764242e-01 2.10706472e-01
-3.85683358e-01 1.24517992e-01 -1.13197339e+00 7.36337006e-01
5.22504985e-01 -7.95621797e-02 2.10918471e-01 5.16286790e-01
7.77339796e-04 -8.73842612e-02 -1.00905490e+00 8.81812811e-01
2.69773513e-01 -1.01371801e+00 2.07678571e-01 4.37300876e-02
8.57304215e-01 -4.42759007e-01 -2.49271169e-01 2.50317693e-01
5.30805469e-01 -8.12613189e-01 2.05710888e-01 6.34174347e-01
7.68785000e-01 -1.10246241e+00 8.95438910e-01 9.49526370e-01
-1.26279509e+00 -2.46758774e-01 -5.12303591e-01 -1.77403331e-01
1.73007846e-01 7.65816987e-01 -1.00518262e+00 8.29620540e-01
-4.73193228e-02 1.05676901e+00 -6.05098069e-01 6.91099763e-01
-2.83782214e-01 7.81684875e-01 -1.27925634e-01 -4.63230878e-01
3.28256279e-01 -6.75507128e-01 -9.55033600e-02 9.79028344e-01
2.48397365e-01 1.98622912e-01 8.60955536e-01 2.43601784e-01
-2.48261765e-01 6.09344602e-01 -4.62671340e-01 3.21881145e-01
6.20617211e-01 1.33159626e+00 -9.82827187e-01 -4.50261354e-01
-4.00315106e-01 9.40059602e-01 4.19766009e-01 2.10047796e-01
-6.48772597e-01 -3.53162885e-01 -2.41942555e-01 -5.72787464e-01
-1.53602198e-01 1.98892698e-01 -3.82200837e-01 -9.89574730e-01
-1.29495665e-01 -5.59357226e-01 5.80520272e-01 -7.20282912e-01
-1.54676032e+00 3.63931865e-01 -9.09179673e-02 -1.49765861e+00
-1.88805386e-01 -3.53756160e-01 -2.62000382e-01 9.64675725e-01
-1.12674093e+00 -1.39277124e+00 -2.73687318e-02 3.46307129e-01
4.26327497e-01 -2.82400995e-01 7.79710948e-01 1.49580315e-01
-7.50468016e-01 6.12762988e-01 5.72177231e-01 3.23109865e-01
5.93168259e-01 -1.24520421e+00 -2.23044828e-01 4.61750001e-01
1.80482820e-01 2.29240641e-01 2.25307778e-01 -7.50251710e-01
-7.45379746e-01 -1.33327305e+00 9.22272801e-01 -4.13512625e-02
5.34227192e-01 -4.19831932e-01 -7.68388391e-01 8.36527884e-01
1.87204301e-01 -2.44502753e-01 1.10400903e+00 1.89915359e-01
-2.23447293e-01 -1.56264722e-01 -9.33363378e-01 4.53535676e-01
1.19223475e+00 -4.17456478e-01 -4.80027944e-01 8.62795591e-01
9.00453508e-01 -5.34032136e-02 -9.18659270e-01 1.30929291e-01
1.67839825e-01 -5.54030716e-01 5.15572131e-01 -4.20300633e-01
4.44223344e-01 -3.04140806e-01 3.26273069e-02 -1.51256728e+00
-6.54894471e-01 8.09854195e-02 4.50890988e-01 1.70703435e+00
6.51881516e-01 -6.07690930e-01 8.08272064e-01 3.71001452e-01
-3.12979035e-02 -7.57313848e-01 -5.01981735e-01 -5.75927913e-01
-1.46695629e-01 4.75499257e-02 6.61614239e-01 1.31187880e+00
2.62061864e-01 1.14057636e+00 -6.41582072e-01 3.59910317e-02
5.89310229e-01 3.68875831e-01 1.15892984e-01 -1.47805405e+00
-2.39038363e-01 -1.63523883e-01 -7.13056000e-03 -1.27524471e+00
8.98618162e-01 -1.59750438e+00 1.61551028e-01 -1.61753178e+00
4.84330505e-01 -9.59597468e-01 -5.57515919e-01 7.26904094e-01
-4.73209262e-01 9.99951735e-02 1.58790499e-01 4.91214097e-01
-8.17541361e-01 5.55245399e-01 1.45873570e+00 -2.20889762e-01
-1.00722566e-01 2.61222839e-01 -6.49276495e-01 6.59345269e-01
9.11109984e-01 -7.68276870e-01 -9.30100977e-01 -1.61873847e-01
2.89465845e-01 4.70779598e-01 -2.49550477e-01 -8.12508047e-01
4.31655109e-01 -5.31982780e-01 2.42221177e-01 -6.83866978e-01
2.26619825e-01 -8.71203542e-01 5.64173870e-02 3.14059585e-01
-1.01563311e+00 -2.44324788e-01 -4.20101166e-01 6.77513301e-01
-1.22605532e-01 -7.50547111e-01 5.92976511e-01 -8.65927041e-02
-6.21447861e-01 3.47740173e-01 -2.02523947e-01 -3.31610776e-02
1.05439174e+00 6.72654137e-02 -2.15665534e-01 -4.05268446e-02
-6.64973497e-01 6.47208035e-01 1.24397166e-01 3.02463233e-01
5.21054208e-01 -1.59469295e+00 -6.70191467e-01 -5.66983223e-02
2.36897364e-01 5.23271300e-02 4.61270899e-01 3.94497842e-01
-8.90311897e-02 2.84029186e-01 7.95073733e-02 -3.41017812e-01
-1.61006081e+00 7.93108702e-01 1.00761019e-01 -5.96842706e-01
-2.74193108e-01 6.48911476e-01 4.66916524e-02 -8.19926620e-01
-5.90437790e-03 2.01166585e-01 -8.45739007e-01 2.62995631e-01
3.05648685e-01 2.14028850e-01 -2.85909444e-01 -7.29060888e-01
-1.74175218e-01 6.52226746e-01 -3.38141531e-01 5.41177019e-02
9.87228751e-01 -4.68408287e-01 -6.09579921e-01 8.71578336e-01
1.54556274e+00 -3.50452185e-01 -6.93220317e-01 -5.61890960e-01
2.90386707e-01 -1.45017684e-01 -2.33058229e-01 -7.32561052e-01
-1.02219737e+00 7.90922225e-01 2.77381599e-01 3.39329183e-01
1.14737463e+00 -2.74771154e-01 5.44420719e-01 6.04709864e-01
3.46241266e-01 -1.15532982e+00 1.98393553e-01 7.06953704e-01
3.24709505e-01 -1.16736543e+00 -1.52981490e-01 -1.13160121e+00
-7.48926461e-01 1.18196797e+00 7.79666960e-01 1.42097533e-01
8.88892233e-01 -6.51062131e-02 5.80234528e-02 1.42475486e-01
-6.64585650e-01 -7.90912732e-02 2.15047747e-01 3.90609831e-01
4.10575479e-01 -2.98914220e-02 -5.50320268e-01 3.33372980e-01
-7.01264665e-02 -1.53021559e-01 5.38487077e-01 5.68516016e-01
-6.74220264e-01 -1.40028036e+00 -1.58902153e-01 1.01110077e+00
-1.41502112e-01 -2.96651483e-01 -1.74244925e-01 2.15096116e-01
3.28678995e-01 1.25691235e+00 -1.67667374e-01 -7.97517955e-01
-3.75540592e-02 3.54946762e-01 6.86824694e-02 -9.44742739e-01
-5.93933165e-01 -7.74432719e-02 2.20444337e-01 2.82499343e-01
-3.75077605e-01 -5.29833913e-01 -1.61371517e+00 -7.70584960e-03
-7.47256815e-01 4.88980621e-01 3.53836149e-01 1.24850130e+00
1.37610540e-01 5.98763645e-01 1.18775475e+00 -6.65301234e-02
-6.98353291e-01 -1.35002613e+00 -1.03745556e+00 7.14774311e-01
-2.46715501e-01 -3.44938904e-01 -2.82746315e-01 1.44178405e-01] | [9.692999839782715, 4.191701889038086] |
ea5aa0ab-6513-4717-8dff-6a075c7dce84 | semantic-role-labeling-meets-definition | 2212.01094 | null | https://arxiv.org/abs/2212.01094v1 | https://arxiv.org/pdf/2212.01094v1.pdf | Semantic Role Labeling Meets Definition Modeling: Using Natural Language to Describe Predicate-Argument Structures | One of the common traits of past and present approaches for Semantic Role Labeling (SRL) is that they rely upon discrete labels drawn from a predefined linguistic inventory to classify predicate senses and their arguments. However, we argue this need not be the case. In this paper, we present an approach that leverages Definition Modeling to introduce a generalized formulation of SRL as the task of describing predicate-argument structures using natural language definitions instead of discrete labels. Our novel formulation takes a first step towards placing interpretability and flexibility foremost, and yet our experiments and analyses on PropBank-style and FrameNet-style, dependency-based and span-based SRL also demonstrate that a flexible model with an interpretable output does not necessarily come at the expense of performance. We release our software for research purposes at https://github.com/SapienzaNLP/dsrl. | ['Roberto Navigli', 'Alessandro Scirè', 'Edoardo Barba', 'Simone Conia'] | 2022-12-02 | null | null | null | null | ['semantic-role-labeling'] | ['natural-language-processing'] | [ 2.61196256e-01 5.95646322e-01 -4.96589631e-01 -7.42164850e-01
-4.18299794e-01 -1.04739559e+00 8.75775576e-01 3.82600218e-01
-3.46205443e-01 9.84545827e-01 5.72758079e-01 -6.66979134e-01
-3.26115429e-01 -6.68201029e-01 -1.81383103e-01 -2.16630131e-01
3.10028553e-01 5.13107955e-01 4.64084357e-01 -6.08203053e-01
1.35407373e-01 4.15982664e-01 -1.36475885e+00 5.04486978e-01
5.40160894e-01 6.15246713e-01 1.44682661e-01 2.01131120e-01
-4.22322392e-01 1.20209587e+00 -4.72987264e-01 -4.70063388e-01
-1.04566962e-02 -4.30243433e-01 -1.36262763e+00 4.59835678e-03
-5.92823029e-02 2.57557333e-01 1.09593578e-01 8.01581860e-01
1.19562887e-01 1.00690901e-01 2.31860965e-01 -1.23582208e+00
-3.98409843e-01 1.26211107e+00 -5.07992171e-02 1.94999188e-01
6.28445864e-01 -1.56207144e-01 1.67585170e+00 -5.55684626e-01
9.61825013e-01 1.37911987e+00 6.15174532e-01 6.17810726e-01
-1.37273920e+00 -1.28345087e-01 5.06090522e-01 1.42812040e-02
-1.08679044e+00 -5.41567326e-01 9.24132228e-01 -3.12858403e-01
9.46460724e-01 5.00986993e-01 4.58779514e-01 8.93475533e-01
-3.85957956e-01 6.09097242e-01 1.31701910e+00 -7.54498065e-01
4.87972274e-02 1.51257157e-01 7.38274693e-01 5.45885265e-01
4.73924220e-01 -3.89189422e-01 -4.29989815e-01 -2.25308180e-01
6.73753202e-01 -5.04874289e-01 -2.62080461e-01 -1.91029802e-01
-1.19235003e+00 8.35282326e-01 -9.37528685e-02 9.12013292e-01
-6.45946562e-02 1.61417514e-01 6.55934095e-01 4.26967651e-01
4.68958437e-01 5.89936018e-01 -6.69125617e-01 -2.10907966e-01
-5.54328144e-01 3.74922514e-01 9.83378351e-01 8.15845311e-01
4.82725948e-01 -2.56485760e-01 1.78900108e-01 7.77298033e-01
4.52111095e-01 -4.45248708e-02 3.19172889e-01 -1.33331001e+00
1.69666946e-01 7.43527234e-01 4.88046139e-01 -7.17634141e-01
-4.92548704e-01 -1.24857605e-01 1.60301924e-01 7.82300532e-03
8.04301322e-01 6.37404695e-02 -4.31004554e-01 1.88020861e+00
2.97872871e-01 -3.86235535e-01 2.44035169e-01 7.60523319e-01
7.49023914e-01 3.58888924e-01 5.37122488e-01 -2.31242388e-01
1.56827509e+00 -6.55124366e-01 -7.85888672e-01 -1.49605289e-01
9.16464269e-01 -4.71814394e-01 1.43633938e+00 -5.26809655e-02
-1.09660149e+00 -1.94034055e-01 -9.02888596e-01 -3.61131281e-01
-3.85880262e-01 -4.22451459e-03 8.67404819e-01 6.20874703e-01
-7.43861258e-01 3.55903029e-01 -8.02117467e-01 -3.49393547e-01
-4.01238389e-02 7.99428895e-02 -2.66337514e-01 2.35807106e-01
-1.37157500e+00 9.88365054e-01 6.86488092e-01 -1.28944457e-01
-2.20033884e-01 -4.30953443e-01 -1.01844847e+00 -3.35570164e-02
7.42703080e-01 -4.35248852e-01 1.61903203e+00 -1.11445022e+00
-1.33084428e+00 1.19360101e+00 -2.70632237e-01 -5.94913840e-01
3.94695044e-01 -1.45872787e-01 -3.03985238e-01 6.00487180e-02
4.02161241e-01 4.00794178e-01 1.39601901e-01 -1.35988450e+00
-7.35925078e-01 -1.83026582e-01 9.50521410e-01 1.86021313e-01
1.59841910e-01 6.03827059e-01 1.12194285e-01 -5.04712999e-01
2.47384235e-01 -6.82065725e-01 4.62165922e-02 -2.11296037e-01
-7.14201555e-02 -7.80675709e-01 3.17326784e-01 -4.02381927e-01
1.44814575e+00 -1.96374011e+00 6.39513582e-02 -6.46066517e-02
1.03508353e-01 1.61358967e-01 1.97379544e-01 7.32424080e-01
-3.69771451e-01 5.34115434e-01 -4.58574533e-01 -1.49501279e-01
4.16286647e-01 6.41206861e-01 -4.43152905e-01 2.96673179e-01
2.16109380e-01 8.81916046e-01 -1.07534289e+00 -4.75852460e-01
3.11746210e-01 1.71354607e-01 -2.67654866e-01 -2.53298610e-01
-7.08894491e-01 4.49193954e-01 -5.42109549e-01 4.71572965e-01
1.95530027e-01 -2.85030693e-01 9.50215042e-01 -7.14687183e-02
-4.74286735e-01 1.24833202e+00 -1.38578987e+00 1.66174817e+00
-4.30597812e-01 2.45709285e-01 2.16774270e-02 -1.06566525e+00
6.14449680e-01 5.77952266e-01 3.22525054e-01 -4.54861134e-01
3.03910021e-02 5.62532365e-01 2.83223391e-01 -2.49355733e-01
2.91963726e-01 -4.46555078e-01 -4.39669937e-01 6.39478207e-01
-2.36348689e-01 -7.83833861e-02 5.67215025e-01 -2.45476607e-02
8.49282384e-01 5.55827200e-01 7.82896101e-01 -5.90826809e-01
9.08233404e-01 4.57676828e-01 9.02642071e-01 4.96443063e-01
-3.86435613e-02 3.67462635e-01 8.73254418e-01 -4.99717087e-01
-9.55230534e-01 -9.69036639e-01 -3.83829057e-01 1.21265411e+00
-8.09358880e-02 -8.74693990e-01 -4.79362696e-01 -9.62955058e-01
-2.55096346e-01 1.15227878e+00 -1.78853258e-01 5.98467171e-01
-1.05289173e+00 -4.19045508e-01 7.50032902e-01 3.58912230e-01
1.49684668e-01 -1.18218327e+00 -1.09984374e+00 2.75293827e-01
-2.96701286e-02 -1.24187601e+00 1.22630455e-01 2.00712070e-01
-4.18854564e-01 -1.24350870e+00 1.36668272e-02 -6.86526716e-01
4.43012923e-01 -9.50769037e-02 1.43075037e+00 3.47654700e-01
3.64992589e-01 3.00865054e-01 -7.81018436e-01 -3.94510835e-01
-5.62348664e-01 -2.88833119e-02 -1.49816096e-01 -3.05096537e-01
3.77715856e-01 -6.15142405e-01 -9.44904387e-02 8.61455500e-02
-9.98818815e-01 1.41259909e-01 -2.19112396e-01 6.71947360e-01
4.05560374e-01 -1.57006178e-02 6.33846462e-01 -1.48318970e+00
6.67794764e-01 -3.21088433e-01 -6.37721121e-01 3.92506331e-01
-5.66706717e-01 2.09912732e-01 5.33663392e-01 3.79448347e-02
-1.19287825e+00 -1.06731497e-01 -4.99890208e-01 2.99293816e-01
-2.30176941e-01 7.08258033e-01 -3.72863263e-01 3.70365977e-01
6.69185638e-01 -4.88486066e-02 -9.56704840e-03 -5.15261471e-01
5.93757272e-01 4.18928653e-01 1.28993213e-01 -1.10131538e+00
6.39468193e-01 4.55033392e-01 -1.33262262e-01 -5.83444357e-01
-1.15001822e+00 -3.20450962e-01 -7.07917929e-01 1.81815088e-01
6.54708862e-01 -7.39342093e-01 -5.19227028e-01 -8.99283513e-02
-1.23505294e+00 -5.82500339e-01 -6.76622927e-01 9.07866135e-02
-6.91013575e-01 5.03015280e-01 -6.81326807e-01 -7.82832503e-01
-1.00509241e-01 -8.19180310e-01 8.09694231e-01 -6.60293922e-02
-8.21307600e-01 -1.28563178e+00 -2.94524670e-01 3.97547424e-01
2.31223762e-01 3.33426744e-01 1.08608317e+00 -9.81231928e-01
-1.20864667e-01 1.42227411e-01 -7.49333501e-02 2.54426658e-01
4.78191704e-01 -2.69821137e-01 -6.43785655e-01 8.22639540e-02
2.71870732e-01 -2.24339440e-01 2.25310907e-01 6.07532151e-02
6.33574426e-01 -4.17233050e-01 7.90262688e-03 7.51714855e-02
1.49027514e+00 2.04685315e-01 5.02317071e-01 5.60680985e-01
4.11847115e-01 8.29425275e-01 9.60138321e-01 1.05716914e-01
6.47613943e-01 7.02089190e-01 8.22014734e-03 1.54493749e-01
-2.93370068e-01 -2.68118620e-01 1.64604470e-01 2.67512381e-01
-3.09417639e-02 -2.11321875e-01 -1.20815623e+00 6.01609230e-01
-2.07934952e+00 -9.88206029e-01 -3.83557320e-01 1.77047718e+00
1.10573316e+00 3.68873298e-01 3.79511029e-01 4.66866255e-01
5.80427110e-01 4.23823357e-01 5.07350899e-02 -3.98873419e-01
-1.91150919e-01 3.04404885e-01 2.34156534e-01 9.25614357e-01
-8.48268092e-01 1.34797204e+00 5.79809427e+00 5.11046290e-01
-1.00954390e+00 4.32217479e-01 2.58901715e-01 3.14207792e-01
-7.04210162e-01 4.25762713e-01 -8.80609393e-01 3.49012762e-01
9.12868917e-01 -3.01241040e-01 1.57469109e-01 5.80278456e-01
2.77071357e-01 1.59937087e-02 -1.06604826e+00 5.94171286e-01
-2.76418835e-01 -1.47713578e+00 -5.97482771e-02 -3.44226748e-01
5.97615028e-03 -8.02998319e-02 -5.14694333e-01 -5.61177991e-02
3.99838269e-01 -8.38304579e-01 1.21627450e+00 5.02199717e-02
5.87322056e-01 -2.58638412e-01 5.47978461e-01 1.86712399e-01
-1.07850957e+00 -9.73712374e-03 -1.02799915e-01 -5.98113537e-01
5.13218522e-01 4.12413359e-01 -6.76490843e-01 6.76316619e-01
1.63717180e-01 6.76662385e-01 -2.24626660e-01 4.12450671e-01
-9.28492606e-01 8.22624445e-01 -3.36866856e-01 4.45102267e-02
1.17930457e-01 -8.37031230e-02 7.27438390e-01 1.32104790e+00
-2.63433009e-01 2.86443919e-01 3.61389995e-01 8.50192070e-01
2.62725949e-01 8.52252617e-02 -5.34004986e-01 -1.29040182e-01
7.81668484e-01 1.10434735e+00 -1.03338945e+00 -2.35014766e-01
-5.28891265e-01 4.63163763e-01 1.85486093e-01 1.78079531e-02
-8.65174174e-01 4.76123579e-02 5.30946553e-01 4.93402630e-01
-4.29596193e-02 -5.84860682e-01 -5.87749839e-01 -1.18722975e+00
2.00249389e-01 -9.01992381e-01 7.41037428e-01 -5.98675013e-01
-1.11379290e+00 4.97989088e-01 5.28152049e-01 -7.39513099e-01
-4.49115485e-01 -7.52149880e-01 -2.32123911e-01 8.27375889e-01
-1.50363576e+00 -1.40092599e+00 2.33603001e-01 1.88317508e-01
5.67652225e-01 4.26664710e-01 9.26164985e-01 2.31843919e-01
-2.61271894e-01 2.47993365e-01 -7.85715401e-01 2.19601631e-01
3.31989676e-01 -1.36642039e+00 4.63160306e-01 8.82057965e-01
3.06613863e-01 9.25876021e-01 1.05047178e+00 -4.13663447e-01
-9.41406727e-01 -6.41030312e-01 1.48449469e+00 -8.63836110e-01
1.05299437e+00 -2.91028559e-01 -8.90345812e-01 1.09205806e+00
2.57413417e-01 -1.68347567e-01 6.33252501e-01 3.44527960e-01
-4.10474360e-01 1.96052790e-01 -1.18605459e+00 4.35218781e-01
1.49900556e+00 -6.13148451e-01 -1.34512115e+00 2.95086890e-01
9.84630287e-01 -3.63163233e-01 -9.72358644e-01 4.21054572e-01
3.53857219e-01 -7.98507452e-01 7.66507030e-01 -5.84029078e-01
1.49066210e-01 -7.62457550e-01 -3.80150467e-01 -7.42083788e-01
-2.20540300e-01 -5.07040858e-01 3.66821617e-01 1.49815679e+00
7.30897009e-01 -9.06833708e-01 4.84648049e-01 9.24134314e-01
-2.95299172e-01 -4.34409469e-01 -9.33564842e-01 -7.28530943e-01
1.31571501e-01 -7.82262743e-01 5.76208532e-01 1.03835464e+00
4.11666542e-01 6.80784881e-01 2.13046923e-01 -6.95369840e-02
3.28985214e-01 3.42912167e-01 3.70506078e-01 -1.37766469e+00
-4.12117034e-01 -2.72198379e-01 -1.60078421e-01 -6.62728310e-01
6.02931976e-01 -1.01257455e+00 -2.82273948e-01 -1.78424549e+00
-3.22063416e-01 -9.73912954e-01 -1.45572618e-01 1.06163204e+00
2.89390385e-01 5.06694652e-02 2.86139578e-01 2.33106196e-01
-5.62739372e-01 3.71872587e-03 7.74570942e-01 2.69195229e-01
-2.79011279e-01 -1.27942905e-01 -1.04292345e+00 8.54199171e-01
1.09885263e+00 -5.00992417e-01 -6.13894939e-01 -5.50377846e-01
5.39517045e-01 -8.52124989e-02 3.42708200e-01 -5.98593414e-01
-8.28700960e-02 -5.55932760e-01 -4.43739146e-01 5.54527454e-02
2.06393123e-01 -6.30303741e-01 4.31839913e-01 3.96734238e-01
-4.68686521e-01 5.25339358e-02 6.67239130e-02 6.16094144e-03
-2.08451837e-01 -6.49177849e-01 5.97560346e-01 -5.70090771e-01
-9.83122766e-01 -3.63860250e-01 -1.87907144e-01 1.66472867e-01
7.63368905e-01 -1.40090078e-01 -3.77053916e-01 -5.08822016e-02
-9.10947502e-01 2.66438052e-02 7.48530447e-01 3.94851476e-01
1.58763334e-01 -8.27733994e-01 -6.84271872e-01 -1.75586894e-01
1.17342852e-01 -7.72386491e-02 -3.79216939e-01 5.76238871e-01
-5.54426312e-01 6.46941900e-01 6.16199225e-02 -1.13270007e-01
-1.24680305e+00 2.99753755e-01 3.73481452e-01 -3.61362219e-01
-7.62500226e-01 6.36546671e-01 2.97387317e-02 -4.50208545e-01
-1.81108147e-01 -3.70417088e-01 -3.40336084e-01 4.93353792e-02
3.72699350e-01 -9.08236876e-02 -1.61619172e-01 -8.10993493e-01
-6.10965312e-01 2.00282604e-01 2.19644979e-01 -4.97370094e-01
1.43377650e+00 -3.32576990e-01 -4.91376013e-01 7.70173788e-01
4.83939141e-01 4.74028230e-01 -7.39188015e-01 -2.49982566e-01
7.42817044e-01 -2.08725438e-01 -2.91188180e-01 -9.44682240e-01
-3.82662326e-01 3.22527587e-01 -1.28702849e-01 6.10490382e-01
9.02813315e-01 4.47222412e-01 5.25571108e-01 3.15334350e-02
7.16568232e-01 -9.64268863e-01 -5.37545621e-01 6.25788331e-01
9.72980917e-01 -7.63283134e-01 -7.78277591e-02 -9.21962738e-01
-5.60383856e-01 1.01621258e+00 2.83016235e-01 -1.53214857e-01
3.79868090e-01 2.19694331e-01 8.23266357e-02 -4.59564835e-01
-7.28506207e-01 -6.78675532e-01 -2.14453578e-01 4.04654443e-01
1.03353179e+00 2.09004968e-01 -1.25478697e+00 6.91960394e-01
-3.81913990e-01 1.12950364e-02 5.71833432e-01 1.24951935e+00
-3.31367254e-01 -1.94067180e+00 -1.71739310e-01 -4.12395447e-02
-8.75245333e-01 -2.68834919e-01 -4.31509674e-01 1.13298404e+00
1.62505254e-01 1.03983450e+00 -4.70912978e-02 8.65578949e-02
2.79827148e-01 5.19603312e-01 5.63893974e-01 -1.20017540e+00
-6.58406198e-01 -1.82960302e-01 1.05062532e+00 -2.17214644e-01
-9.55063641e-01 -7.74892330e-01 -1.83630002e+00 -8.93508121e-02
-4.78485078e-02 5.50155163e-01 4.15411294e-01 1.00890994e+00
1.24297507e-01 2.10473463e-01 1.52325943e-01 -2.26605564e-01
-4.94672060e-01 -6.74381018e-01 -3.32386404e-01 5.51477909e-01
4.23056521e-02 -7.86745012e-01 -4.15312529e-01 8.88100490e-02] | [10.26990032196045, 9.264399528503418] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.