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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1af8ae53-17e2-4c3b-b626-72904fe8ca31 | learning-fragment-self-attention-embeddings | null | null | https://dl.acm.org/doi/10.1145/3343031.3350940 | https://dl.acm.org/doi/pdf/10.1145/3343031.3350940 | Learning fragment self-attention embeddings for image-text matching | In image-text matching task, the key to good matching quality is to capture the rich contextual dependencies between fragments of image and text. However, previous works either simply aggregate the similarity of all possible pairs of image regions and words, or take multi-step cross attention to attend to image regions and words with each other as context, which requires exhaustive similarity computation between all image region and word pairs. In this paper, we propose Self-Attention Embeddings (SAEM) to exploit fragment relations in images or texts by self-attention mechanism, and aggregate fragment information into visual and textual embeddings. Specifically, SAEM extracts salient image regions based on bottom-up attention, and takes WordPiece tokens as sentence fragments. The self-attention layers are built to model subtle and fine-grained fragment relation in image and text respectively, which consists of multi-head self-attention sub-layer and position-wise feed-forward network sub-layer. Consequently, the fragment self-attention mechanism can discover the fragment relations and identify the semantically salient regions in images or words in sentences, and capture their interaction more accurately. By simultaneously exploiting the fine-grained fragment relation in both visual and textual modalities, our method produces more semantically consistent embeddings for representing images and texts, and demonstrates promising image-text matching accuracy and high efficiency on Flickr30K and MSCOCO datasets. | ['Qingming Huang', 'Guoli Song', 'Shuhui Wang', 'Yiling Wu'] | 2019-10-01 | null | null | null | acmmm-2019-10 | ['text-matching'] | ['natural-language-processing'] | [ 2.29086354e-01 -3.09302777e-01 -2.97843009e-01 -4.39354897e-01
-6.98136508e-01 -3.00637811e-01 8.46779287e-01 3.26099604e-01
-4.60971028e-01 1.06815752e-02 6.83576167e-01 9.74572152e-02
2.11397186e-02 -6.53704107e-01 -8.28775883e-01 -3.88240695e-01
3.44018817e-01 2.72204038e-02 2.67104208e-01 -2.10044235e-01
1.46890000e-01 2.78319806e-01 -1.63581836e+00 7.16012120e-01
6.46907270e-01 1.17889583e+00 5.61484218e-01 6.65323257e-01
-5.12289107e-01 7.23921418e-01 -2.07453564e-01 -4.50722367e-01
-8.71005505e-02 -3.45989436e-01 -9.16200936e-01 4.32259828e-01
7.90888727e-01 -3.67571592e-01 -7.84770310e-01 1.14590323e+00
4.92189050e-01 1.86657056e-01 6.27198160e-01 -1.22882879e+00
-1.47850430e+00 5.88703096e-01 -8.58231127e-01 4.62427884e-01
4.99386042e-01 2.29263276e-01 1.21320915e+00 -1.32327330e+00
2.80694455e-01 1.57592773e+00 3.52155983e-01 1.68635964e-01
-1.01257122e+00 -6.62003994e-01 4.22063798e-01 5.01389503e-01
-1.58860457e+00 -4.44692224e-01 9.14284527e-01 -3.51246506e-01
1.18061757e+00 3.39659691e-01 7.72464514e-01 1.03521800e+00
2.50722766e-01 1.14958215e+00 6.78614914e-01 -3.53349030e-01
-4.71085608e-01 1.92350492e-01 1.05312921e-01 5.98592401e-01
-9.64524373e-02 -7.10951686e-02 -4.19396460e-01 1.85463235e-01
8.39317799e-01 6.81711495e-01 -3.30593914e-01 -8.88140649e-02
-1.45570695e+00 6.88900590e-01 7.50080645e-01 7.53399611e-01
-5.01126707e-01 3.19288850e-01 4.41833764e-01 8.21423829e-02
4.13406968e-01 2.42036551e-01 -6.25866279e-02 2.20693573e-01
-8.75769615e-01 -1.57089904e-02 -1.65137798e-02 1.19829035e+00
1.00724041e+00 -2.95728356e-01 -6.56889856e-01 9.77173984e-01
4.70428467e-01 6.95656359e-01 7.47730434e-01 -3.74076635e-01
6.56438291e-01 9.60366130e-01 -2.66914696e-01 -1.57435644e+00
-8.54957923e-02 -2.90983051e-01 -9.31306601e-01 -5.28820634e-01
-2.16533214e-01 5.26083887e-01 -9.58950639e-01 1.56744289e+00
3.19910198e-01 1.32942259e-01 -6.25737756e-02 1.05598128e+00
1.27921760e+00 7.33613729e-01 2.62975723e-01 -1.35282930e-02
1.79403841e+00 -1.23845387e+00 -7.61726439e-01 -5.31920075e-01
4.30704355e-01 -8.68339598e-01 1.47661638e+00 -4.25125092e-01
-1.22630501e+00 -1.06695294e+00 -7.50537038e-01 -7.73511350e-01
-6.56109750e-01 4.39700633e-02 9.98200402e-02 -1.23915203e-01
-7.49483228e-01 3.45739424e-01 -1.83539018e-01 -4.47332501e-01
4.91892010e-01 2.23876849e-01 -4.20018762e-01 -3.99058074e-01
-1.48823726e+00 6.40251935e-01 3.74457836e-01 2.09989861e-01
-4.97475117e-01 -6.60629570e-01 -1.27862501e+00 2.77328998e-01
1.75598890e-01 -6.66362822e-01 8.63528252e-01 -1.35000014e+00
-8.37132990e-01 1.13727140e+00 -3.19596976e-01 -1.38293713e-01
1.13238163e-01 -1.59458473e-01 -7.07908154e-01 3.66450816e-01
2.27841437e-01 1.07880402e+00 1.14775515e+00 -1.21777260e+00
-6.17196202e-01 -4.72060204e-01 1.13146245e-01 4.59178954e-01
-7.67582238e-01 1.66235149e-01 -1.03322899e+00 -8.28348815e-01
-2.75525004e-02 -3.67857963e-01 1.68944038e-02 2.50050109e-02
-2.81311363e-01 -3.19671065e-01 1.00410879e+00 -7.22328305e-01
1.40474427e+00 -2.32660913e+00 7.79693350e-02 -1.94753036e-01
2.94745058e-01 1.07972726e-01 -5.84949911e-01 4.54072714e-01
-2.84859478e-01 1.56339243e-01 2.46215481e-02 -3.58568370e-01
1.10328048e-02 2.88818806e-01 -4.97229815e-01 2.96418726e-01
4.76249903e-01 1.72129762e+00 -1.03167474e+00 -1.03445005e+00
4.97046143e-01 6.12324417e-01 -3.31072778e-01 2.52241552e-01
-1.04064234e-01 -5.56457713e-02 -6.07906401e-01 6.02336049e-01
5.82865536e-01 -4.01360631e-01 -3.91512066e-01 -8.65900397e-01
9.55541581e-02 -7.21936822e-02 -7.65864015e-01 1.81064701e+00
-7.29498327e-01 7.53688037e-01 -1.92908958e-01 -8.89719665e-01
7.05041051e-01 1.10528469e-01 4.11994159e-01 -1.19175112e+00
2.17327207e-01 -1.51087597e-01 -4.98618364e-01 -9.57338452e-01
6.80575490e-01 3.37853911e-03 -2.11605966e-01 3.97986203e-01
1.31208777e-01 7.10635483e-02 -1.51648402e-01 3.39786023e-01
5.24202585e-01 -1.42526224e-01 3.09095234e-01 -6.05907775e-02
8.06712985e-01 -2.89201140e-01 2.58361492e-02 4.87706989e-01
-1.77208871e-01 7.40403175e-01 1.37080953e-01 -4.85242635e-01
-9.76278961e-01 -7.80342519e-01 3.93895730e-02 1.48348618e+00
7.51884162e-01 -4.33935791e-01 -3.81293476e-01 -6.91668749e-01
2.05844656e-01 3.57870370e-01 -1.04687250e+00 -4.55986142e-01
-3.69339019e-01 -2.16383636e-01 1.92844838e-01 7.47706532e-01
6.15370452e-01 -1.37938297e+00 -2.06925526e-01 4.31608558e-02
-2.89071858e-01 -1.23612940e+00 -1.29700351e+00 -1.85505018e-01
-3.71776432e-01 -1.02735984e+00 -8.05795729e-01 -1.18034852e+00
7.02244282e-01 8.08645904e-01 1.17914951e+00 3.52804333e-01
-5.55557013e-01 5.82535088e-01 -5.37127435e-01 6.09585568e-02
1.44075863e-02 -2.10410967e-01 -3.72800529e-01 4.43403482e-01
4.73151207e-01 -2.95758158e-01 -8.77068400e-01 3.35999489e-01
-1.26759946e+00 2.31529072e-01 6.95564508e-01 9.98527467e-01
7.81846762e-01 -6.65220097e-02 8.77231508e-02 -5.12205184e-01
5.94066620e-01 -5.21860301e-01 -5.03311343e-02 5.29258251e-01
-2.26697132e-01 -1.54044572e-02 6.07167542e-01 -5.53736985e-01
-6.88016772e-01 8.72095376e-02 -4.10346165e-02 -1.09017670e+00
-1.30958617e-01 3.63911062e-01 -4.68133956e-01 1.43263772e-01
2.67262816e-01 7.48287559e-01 -1.90673515e-01 -1.49489865e-01
6.71222091e-01 8.64071667e-01 6.82997048e-01 -4.42584872e-01
6.97088420e-01 4.74706888e-01 -4.08988386e-01 -7.83645213e-01
-1.03059304e+00 -7.93499529e-01 -7.25133121e-01 -1.85978517e-01
1.24774027e+00 -9.88522828e-01 -4.49533880e-01 2.90735781e-01
-1.14168966e+00 4.39372379e-03 -3.87232721e-01 2.68024892e-01
-3.18967551e-01 6.49600089e-01 -4.28505391e-01 -4.39316422e-01
-5.06093025e-01 -1.21732843e+00 1.72436273e+00 3.41090858e-01
8.53851289e-02 -9.81787145e-01 -3.23435694e-01 2.93093443e-01
2.14001074e-01 -5.16025648e-02 7.52751946e-01 -5.36724091e-01
-4.53038782e-01 -3.34565461e-01 -9.19098914e-01 1.90898582e-01
2.44661987e-01 -5.47261350e-02 -8.51733744e-01 -3.51473004e-01
-2.57987767e-01 -3.78868550e-01 1.09009230e+00 2.85669357e-01
1.49524641e+00 -4.48677182e-01 -4.11377400e-01 5.74600577e-01
1.43104255e+00 -2.10249990e-01 5.51126122e-01 1.37254730e-01
1.15405095e+00 5.92330277e-01 5.43826759e-01 2.06747860e-01
6.64845407e-01 7.27524161e-01 5.02327502e-01 -5.37140310e-01
-2.49135852e-01 -4.93991822e-01 2.15237722e-01 8.15517664e-01
5.11622488e-01 -1.60421327e-01 -6.76858842e-01 8.99770141e-01
-1.95106137e+00 -1.23167300e+00 -2.14260742e-02 1.84768188e+00
7.07429290e-01 -1.18061066e-01 4.09832336e-02 -2.07743123e-01
9.88356292e-01 4.52503949e-01 -5.91260672e-01 -2.84393996e-01
-2.24397868e-01 -1.02409795e-01 2.38237202e-01 4.26285565e-01
-1.14128721e+00 9.72443044e-01 5.21190691e+00 1.15797174e+00
-1.07051849e+00 1.53286517e-01 5.85697234e-01 -2.10280254e-01
-6.28656089e-01 -2.08692923e-01 -5.60211182e-01 5.79036951e-01
3.94926935e-01 -1.43845513e-01 2.26181656e-01 6.79982007e-01
-7.58454204e-03 2.78307229e-01 -1.12355494e+00 1.29254758e+00
4.91707504e-01 -1.40541756e+00 5.54691851e-01 -5.64477816e-02
6.70169055e-01 -1.93148345e-01 1.67453825e-01 2.95121431e-01
1.09424070e-02 -1.22287703e+00 7.38135338e-01 5.61991274e-01
1.03781378e+00 -8.11092079e-01 7.06123650e-01 -6.11916836e-03
-1.84787464e+00 -2.81532288e-01 -5.27012765e-01 2.34526590e-01
-3.44527178e-02 3.27762634e-01 -2.89370358e-01 6.52916133e-01
8.36857796e-01 1.18272364e+00 -7.39903331e-01 6.16129100e-01
1.51639301e-02 5.22884466e-02 4.99934368e-02 -1.26560992e-02
5.71673512e-01 1.14940323e-01 1.71296865e-01 1.55411434e+00
1.06313787e-02 1.57988057e-01 2.99743980e-01 9.81601715e-01
-1.99267805e-01 2.57825047e-01 -5.06234944e-01 -2.01434925e-01
5.51448405e-01 1.43208420e+00 -4.75825638e-01 -5.73362827e-01
-6.63869679e-01 1.44636416e+00 4.27368402e-01 3.67374986e-01
-9.48606610e-01 -4.91993040e-01 7.25123107e-01 1.16600558e-01
4.84930605e-01 8.04018900e-02 -5.10813482e-02 -1.03336394e+00
5.60322478e-02 -7.25592256e-01 5.59300303e-01 -1.19095755e+00
-1.44410253e+00 6.82004571e-01 -2.86902845e-01 -1.28454340e+00
2.68039942e-01 -4.02276158e-01 -7.43222713e-01 9.23154294e-01
-1.61738980e+00 -1.67277181e+00 -6.18128359e-01 1.02080965e+00
9.10388947e-01 2.09098130e-01 5.32985806e-01 3.29890966e-01
-6.28748059e-01 9.47320223e-01 -4.98801172e-02 3.75058651e-01
6.15812004e-01 -9.70510244e-01 4.73456562e-01 6.60696566e-01
5.00813127e-01 6.41140819e-01 1.77539647e-01 -4.86096919e-01
-1.43034112e+00 -1.51625490e+00 1.01527381e+00 -3.37559670e-01
6.35111690e-01 -3.93307090e-01 -1.09110510e+00 4.71133262e-01
4.44967717e-01 3.33143800e-01 2.85011351e-01 -2.07433328e-01
-6.47764087e-01 -7.42778853e-02 -6.86447024e-01 7.22723424e-01
1.04966676e+00 -1.22976995e+00 -9.03727531e-01 3.68323177e-01
1.12045515e+00 -2.53562301e-01 -8.31527054e-01 2.90483057e-01
5.17269373e-01 -7.15926707e-01 1.33238089e+00 -6.67397976e-01
8.27691495e-01 -3.23648453e-01 -3.36651385e-01 -9.89860356e-01
-6.86258674e-01 -2.89802432e-01 -4.94546629e-02 1.46728766e+00
1.54184625e-01 -1.56725690e-01 3.46816510e-01 2.26890922e-01
-1.70113608e-01 -8.64631236e-01 -6.75857306e-01 -4.33992982e-01
-2.82023847e-02 -4.21801597e-01 8.31028044e-01 1.16804171e+00
-1.49659310e-02 5.81708789e-01 -3.70201826e-01 1.83282554e-01
1.91434935e-01 6.15064263e-01 6.49497747e-01 -6.92344546e-01
-9.13080052e-02 -7.72078812e-01 -4.41233903e-01 -1.28888297e+00
3.37572098e-01 -1.01489043e+00 2.19110455e-02 -1.61614966e+00
6.44545913e-01 -1.13409478e-02 -5.61481535e-01 3.88535917e-01
-7.37726748e-01 5.38666546e-01 2.03188986e-01 2.40297198e-01
-9.61751759e-01 7.95596182e-01 1.56267285e+00 -5.28626621e-01
9.40668359e-02 -7.59254932e-01 -7.60687590e-01 5.17446756e-01
2.87682861e-01 -1.12348579e-01 -3.70080024e-01 -7.49597847e-01
1.11353815e-01 1.00835576e-03 8.38296711e-01 -6.50786579e-01
3.54038715e-01 -2.19499886e-01 8.33997071e-01 -8.57915640e-01
3.29617113e-01 -1.09116709e+00 -2.29756877e-01 2.43442848e-01
-7.06857443e-01 1.95942029e-01 1.19669534e-01 7.63926148e-01
-5.62353671e-01 7.60093182e-02 6.41717434e-01 -1.92346960e-01
-1.01227081e+00 7.79002130e-01 8.79774988e-02 2.32005343e-01
1.00392044e+00 -5.56652308e-01 -6.74155951e-02 -3.24469507e-01
-4.94198531e-01 5.36970377e-01 5.57932079e-01 1.01978636e+00
9.98463333e-01 -1.78174651e+00 -6.14903748e-01 5.57825267e-01
6.43118560e-01 2.85013523e-02 8.17274094e-01 6.47405148e-01
-2.72649508e-02 3.89600724e-01 -1.24470159e-01 -7.76553333e-01
-1.30473590e+00 1.08236098e+00 3.64985734e-01 -7.14916969e-03
-6.62261605e-01 1.14762020e+00 7.51132190e-01 -9.40349177e-02
1.42521605e-01 -4.27499712e-01 -2.81756371e-01 2.28120044e-01
7.86890149e-01 -2.29803950e-01 -2.88411856e-01 -1.22855735e+00
-4.46883142e-01 1.08235931e+00 -2.79463649e-01 2.84134537e-01
9.41203892e-01 -6.43616855e-01 -1.40933646e-02 1.76804468e-01
1.78751886e+00 -2.30180532e-01 -1.14453423e+00 -7.99064219e-01
-3.55533719e-01 -7.67298877e-01 1.88171491e-01 -3.70160729e-01
-1.26831400e+00 1.14910817e+00 3.54355484e-01 4.47691679e-02
1.29520452e+00 4.07861978e-01 1.09419644e+00 -7.02198073e-02
-4.13306117e-01 -9.30156946e-01 6.06012046e-01 3.40049416e-01
1.15102386e+00 -1.33725190e+00 -1.58470929e-01 -1.71002015e-01
-7.66399860e-01 9.38243687e-01 8.91330481e-01 -8.79198909e-02
4.95617777e-01 -2.21012861e-01 -1.91689700e-01 -3.23667884e-01
-7.80428469e-01 -5.19187212e-01 9.00319576e-01 5.67047238e-01
3.40047121e-01 -1.58801526e-01 3.56983915e-02 8.06020260e-01
1.99572489e-01 -3.70885342e-01 -1.91799060e-01 6.09594643e-01
-5.45110762e-01 -6.57352388e-01 -3.50229681e-01 4.35120076e-01
-1.68002322e-01 -4.92844969e-01 -5.51290393e-01 6.62984610e-01
2.95788050e-01 7.54310191e-01 4.90614623e-01 -5.40864050e-01
4.63132352e-01 -1.09067477e-01 2.85356402e-01 -4.11603451e-01
-7.05512166e-01 6.79432675e-02 -3.53792191e-01 -6.50518358e-01
-3.15342098e-01 -3.59244138e-01 -1.24468529e+00 -1.64618507e-01
-3.96253258e-01 -9.01886150e-02 2.89271325e-01 1.02570581e+00
6.03192389e-01 6.58758342e-01 8.96142304e-01 -8.07336986e-01
-1.01584859e-01 -8.34770858e-01 -4.06630456e-01 9.20796990e-01
3.92524421e-01 -5.17212987e-01 -6.96053952e-02 1.37570649e-01] | [10.702897071838379, 1.3359330892562866] |
0ec1eb8b-73d7-40ad-92c8-57e7c6f54d5f | ludb-a-new-open-access-validation-tool-for | 1809.03393 | null | https://arxiv.org/abs/1809.03393v4 | https://arxiv.org/pdf/1809.03393v4.pdf | LUDB: a new open-access validation tool for electrocardiogram delineation algorithms | We report Lobachevsky University Database (LUDB) of ECG signals, an open tool for validating ECG delineation algorithms, that is superior to the existing publicly available data bases in several aspects. LUDB contains 200 recordings of 10-second 12-lead electrocardiograms (ECG) from different subjects, representative of a variety of signal morphologies. The boundaries and peaks of QRS complexes and P and T waves are manually annotated by cardiologists for all recordings and independently for each lead, and all records received an expert classification by abnormalities. We present a case study for the recently proposed wavelet-based algorithm and the broadly used ecg-kit tool, and demonstrate the advantage of multi-lead ECG data analysis. LUDB contributes to the diversity of public databases employed in developing and validating novel ECG analysis algorithms, including the most advanced based on deep learning neural networks. | ['Mikhail V. Ivanchenko', 'Nikolai Yu. Zolotykh', 'Grigory V. Osipov', 'Konstantin A. Kosonogov', 'Alexander V. Nikolskiy', 'Victor A. Moskalenko', 'Igor I. Yusipov', 'Alena I. Kalyakulina'] | 2018-08-30 | null | null | null | null | ['ecg-wave-delineation'] | ['medical'] | [ 1.04394503e-01 -3.94963950e-01 2.49419302e-01 -2.22325832e-01
-1.16631305e+00 -8.81901562e-01 -5.45328379e-01 4.80839729e-01
-1.98026672e-01 8.89764309e-01 -2.11341709e-01 -6.76616311e-01
-6.34058833e-01 -5.23204148e-01 -1.14628062e-01 -6.95016205e-01
-8.70128274e-01 6.89196646e-01 -2.74510026e-01 -7.60807097e-02
-1.43445581e-01 8.05871665e-01 -6.98559046e-01 4.27802116e-01
5.12149692e-01 1.24846756e+00 -3.72947395e-01 1.15742064e+00
5.02626240e-01 2.33118758e-01 -9.78020191e-01 -1.52372092e-01
9.12108645e-02 -8.00975621e-01 -5.22062004e-01 -3.90983582e-01
1.58294335e-01 -6.17844090e-02 6.61808997e-02 3.21412206e-01
1.45538771e+00 -6.71357214e-01 5.59899151e-01 -8.17752361e-01
-1.32687494e-01 8.26607585e-01 -4.24310982e-01 9.25399542e-01
1.37877479e-01 1.56872913e-01 7.58762777e-01 -4.26158130e-01
6.29202187e-01 1.79423898e-01 1.70048356e+00 1.23853497e-01
-1.28447688e+00 -4.59195167e-01 -9.53765452e-01 4.26961988e-01
-1.55906916e+00 -2.14607874e-03 9.35892463e-01 -5.09969473e-01
8.42718661e-01 5.32881081e-01 1.19009745e+00 8.54649723e-01
3.67119849e-01 1.28602192e-01 1.08465290e+00 -4.21838135e-01
2.38661887e-03 -3.19412708e-01 3.02062511e-01 3.75365585e-01
4.39685345e-01 -5.33796474e-03 -2.61153460e-01 -7.59250522e-01
7.43145823e-01 -2.26742789e-01 -3.59174341e-01 -1.20423749e-01
-1.32365358e+00 4.51988339e-01 -2.63091922e-01 4.93771136e-01
-6.31930649e-01 -7.24551221e-03 8.59055519e-01 5.81753731e-01
1.76227108e-01 4.87687081e-01 -8.18634808e-01 -4.57265347e-01
-1.18624616e+00 2.01762050e-01 6.54670417e-01 5.46476662e-01
2.81490713e-01 3.18548590e-01 -3.15173924e-01 8.25387597e-01
5.03689386e-02 3.31609815e-01 5.35717010e-01 -9.91349757e-01
6.49048612e-02 2.82191843e-01 6.67541614e-03 -9.52356577e-01
-9.25884247e-01 -7.39618003e-01 -1.09812284e+00 -1.11427635e-01
4.60452884e-01 -5.52602887e-01 -2.60530025e-01 1.14109957e+00
-3.00361924e-02 1.11587130e-01 8.64071585e-03 8.34164500e-01
1.13867819e+00 -5.54802418e-02 -1.80057541e-01 -5.17009854e-01
1.27405977e+00 -4.53410596e-02 -8.25863779e-01 4.21442628e-01
7.13239968e-01 -3.84437382e-01 4.93357480e-01 1.02070856e+00
-1.11704576e+00 -6.25174284e-01 -1.24648130e+00 4.37152356e-01
5.46440817e-02 4.18581218e-01 3.20338339e-01 1.20263302e+00
-1.07096779e+00 1.13438284e+00 -6.14190876e-01 -1.82056621e-01
7.90981650e-01 2.45753869e-01 -2.65406787e-01 6.33330941e-01
-1.52599072e+00 8.21085870e-01 1.61499292e-01 2.89447367e-01
-7.12032974e-01 -8.51987541e-01 -5.94514132e-01 -1.18215375e-01
-3.28653872e-01 -4.75963175e-01 9.40287173e-01 -7.11000919e-01
-9.23094630e-01 1.34488571e+00 2.82467335e-01 -8.66632164e-01
6.46841347e-01 4.89721335e-02 -6.64087355e-01 3.07158142e-01
-1.28102347e-01 -3.98266643e-01 7.71415472e-01 -7.56063461e-01
-3.45736921e-01 -5.78189015e-01 -4.33189958e-01 -3.94326150e-01
1.68316811e-01 5.95736578e-02 7.30301673e-03 -8.26256812e-01
4.16804999e-02 -4.42106158e-01 -2.42265075e-01 -3.24115127e-01
-3.55301827e-01 4.10291590e-02 3.48282218e-01 -1.08241558e+00
1.60928857e+00 -2.20738220e+00 -7.77355283e-02 5.66004455e-01
7.18418360e-01 2.87399024e-01 1.69712141e-01 5.18537521e-01
-4.91944432e-01 3.68934572e-01 -3.82370800e-01 1.30156085e-01
-2.90441722e-01 2.41787285e-01 -1.75292820e-01 9.19484138e-01
-8.85454044e-02 1.00515962e+00 -3.80774260e-01 -4.49841261e-01
1.88217655e-01 2.58562684e-01 -4.93834633e-03 -4.74458225e-02
5.92886508e-01 8.83865416e-01 -7.59922341e-02 5.26984930e-01
5.84500372e-01 -1.16468042e-01 4.85355616e-01 -4.31671530e-01
1.03809662e-01 -4.02047597e-02 -1.15175414e+00 1.82119799e+00
5.25232963e-03 6.75510108e-01 -3.75072956e-01 -1.14121699e+00
1.05663693e+00 9.33818996e-01 9.10690248e-01 -1.49676681e-01
4.34462696e-01 2.90686846e-01 5.57642519e-01 -7.50574827e-01
-5.96888363e-01 -1.13208272e-01 1.31514296e-01 6.47482455e-01
1.55788422e-01 4.69886372e-03 1.40607715e-01 -1.32167786e-01
1.08011961e+00 5.73307313e-02 8.60454381e-01 -3.93195540e-01
5.29193997e-01 -1.52148142e-01 8.39661717e-01 9.77901757e-01
-5.70185781e-01 1.11801267e+00 7.27845311e-01 -1.17936492e+00
-8.76924038e-01 -1.03160775e+00 -7.77815282e-01 3.98645401e-01
-3.74585360e-01 -3.14819962e-01 -7.33072937e-01 -6.36921674e-02
-1.27052087e-02 1.73948053e-02 -6.40982032e-01 1.00345358e-01
-7.36482322e-01 -8.28014374e-01 1.50047982e+00 7.57234693e-01
1.10605415e-02 -1.16266084e+00 -1.38680542e+00 6.35209799e-01
-3.20083350e-01 -7.70408690e-01 -1.74683630e-02 3.61033589e-01
-1.16161203e+00 -1.52317679e+00 -8.94771338e-01 -6.63225889e-01
-6.68595135e-02 -9.43342686e-01 1.49078107e+00 2.59592801e-01
-1.11614835e+00 1.78730756e-01 -3.11178088e-01 -8.40517581e-01
-5.25115788e-01 -1.34492978e-01 9.82370973e-02 -1.25506299e-03
2.31075138e-01 -8.32257926e-01 -6.21034682e-01 5.01800887e-02
-2.02302590e-01 -4.01393831e-01 3.42671931e-01 7.94956863e-01
5.79826832e-01 -2.07884997e-01 1.19107687e+00 -8.43055308e-01
9.07883465e-01 -2.56384254e-01 -2.66578853e-01 1.52998507e-01
-7.20582306e-01 -8.48490894e-01 4.87188339e-01 -5.07217422e-02
-1.74755424e-01 -2.36430213e-01 -6.37617588e-01 -2.64105529e-01
-6.08376563e-01 4.24088717e-01 2.54027784e-01 -3.33218090e-02
9.03623700e-01 2.47446492e-01 -5.94756790e-02 -2.75835961e-01
-1.51268408e-01 6.82165921e-01 7.93422639e-01 -4.77849454e-01
2.56602079e-01 3.20309490e-01 5.37746400e-02 -7.26660550e-01
-1.27346992e-01 -4.87481296e-01 -1.12628901e+00 -2.15446800e-01
9.25224185e-01 -5.11622250e-01 -7.70285487e-01 5.81579030e-01
-1.30409992e+00 -8.46799556e-03 -5.21899521e-01 3.77956182e-01
-7.18435407e-01 5.41827500e-01 -6.84890389e-01 -7.51945078e-01
-1.04914737e+00 -7.84739614e-01 6.79655492e-01 -2.57245332e-01
-7.66934097e-01 -9.54532802e-01 4.11518306e-01 -3.88411209e-02
3.00761044e-01 1.03193104e+00 1.18139017e+00 -1.03829885e+00
2.67143518e-01 -4.29080933e-01 2.32633293e-01 5.19586146e-01
2.30362296e-01 1.19852126e-02 -1.08471644e+00 -9.47181731e-02
1.78913713e-01 -1.27299339e-01 2.92883188e-01 8.16638052e-01
1.26016963e+00 3.27116609e-01 -1.72032103e-01 9.11396623e-01
1.20762181e+00 6.76120698e-01 6.60029531e-01 3.18620428e-02
2.64837593e-01 6.24946021e-02 -7.63392001e-02 7.79083788e-01
-5.18951081e-02 1.23873636e-01 -1.41616330e-01 -6.98965013e-01
2.15145096e-01 5.50054550e-01 -3.22262913e-01 6.89263761e-01
-7.30102777e-01 2.48395056e-01 -1.27680063e+00 6.00033462e-01
-1.47355258e+00 -7.11000562e-01 -7.17657268e-01 1.91348314e+00
9.11865652e-01 8.80852044e-02 4.97706890e-01 9.45915103e-01
5.94990432e-01 -2.41162688e-01 -5.88995457e-01 -7.10288048e-01
-4.79523987e-01 8.03403974e-01 8.30988064e-02 -2.05858663e-01
-1.07110631e+00 -6.03796057e-02 7.85192537e+00 3.44421357e-01
-1.27819467e+00 3.59813541e-01 7.32530594e-01 1.25166759e-01
3.74707133e-01 -5.73246181e-01 -1.22335210e-01 1.77083015e-01
1.27123070e+00 -2.61552930e-01 4.93466035e-02 5.81100047e-01
4.01714623e-01 2.58604228e-01 -1.27084839e+00 1.54989445e+00
1.14453651e-01 -1.87459362e+00 -6.01415634e-01 -4.02353972e-01
2.36045286e-01 1.83363870e-01 -3.41025740e-01 -1.73893049e-02
-8.59618604e-01 -1.10437226e+00 6.08905375e-01 6.01389289e-01
1.43767977e+00 -7.27770090e-01 1.17738461e+00 5.04536927e-02
-1.08866310e+00 -1.11821502e-01 9.29067060e-02 -7.88680613e-02
6.37309104e-02 6.51428938e-01 -4.35313791e-01 7.49831259e-01
9.52051222e-01 9.37634230e-01 -6.52677476e-01 1.14973497e+00
8.31275284e-02 1.13632822e+00 -7.56407753e-02 4.64627534e-01
-3.42647433e-01 -6.47107512e-02 5.96511662e-01 1.26215923e+00
9.41847637e-02 1.54505581e-01 1.91760268e-02 8.58410180e-01
2.37705037e-01 2.74577916e-01 -6.19299710e-01 1.60658747e-01
1.66129321e-01 1.41050017e+00 -8.02582383e-01 -4.76662040e-01
-2.97269493e-01 3.02205682e-01 -2.56342918e-01 2.80944467e-01
-1.01041532e+00 -9.60143983e-01 7.56360516e-02 3.11180264e-01
-7.86729231e-02 2.32006356e-01 -1.00423670e+00 -6.35421515e-01
1.03680007e-01 -1.23780847e+00 7.79726744e-01 -5.90418339e-01
-1.36585569e+00 9.44054604e-01 -2.88663000e-01 -1.25318456e+00
-2.21063465e-01 -3.81577075e-01 -7.92630851e-01 1.28772330e+00
-1.03608954e+00 -7.74397194e-01 -6.75763860e-02 6.15525246e-01
2.28065670e-01 -2.22288296e-01 1.46082580e+00 7.40716040e-01
-2.52397209e-01 5.11189878e-01 -1.62274525e-01 4.86700326e-01
5.81446111e-01 -1.37576437e+00 3.10342342e-01 3.66065562e-01
1.71469525e-01 6.36777997e-01 4.56766456e-01 -4.12063599e-01
-1.03959310e+00 -8.53763640e-01 7.97434747e-01 -5.64009011e-01
2.65547395e-01 7.06213266e-02 -8.26663852e-01 4.85543311e-01
3.04928005e-01 2.12101772e-01 1.22646129e+00 1.29710704e-01
2.34324306e-01 -3.95078570e-01 -1.04935765e+00 7.39902034e-02
6.20378256e-01 -4.00790006e-01 -1.00457442e+00 1.32279500e-01
-3.12859029e-01 -5.62002003e-01 -1.39372301e+00 6.97146773e-01
1.06510878e+00 -1.10990536e+00 8.48919034e-01 -7.73730397e-01
1.01176500e-02 -1.78283025e-02 3.53445411e-01 -1.13936162e+00
-3.98405343e-01 -1.19982255e+00 -9.65689030e-03 7.25566626e-01
3.79176855e-01 -6.93674982e-01 4.01963413e-01 -3.55243832e-02
-2.40502045e-01 -1.16010213e+00 -1.12563896e+00 -4.55915332e-01
8.39159340e-02 -7.98845172e-01 5.61047494e-01 8.98807228e-01
-2.64198165e-02 -5.86553402e-02 -3.09581935e-01 -7.57121742e-02
6.41904175e-01 3.14076364e-01 2.36966699e-01 -1.54536009e+00
-3.87812972e-01 -4.22210395e-01 -7.11954176e-01 1.36145890e-01
-3.64330471e-01 -9.39949155e-01 -2.93045163e-01 -1.37698317e+00
-5.04060276e-02 -4.67807740e-01 -7.71013916e-01 5.90520263e-01
1.33584775e-02 8.59078109e-01 -2.73035586e-01 4.83295918e-02
-7.45787397e-02 -4.96539354e-01 8.17394435e-01 -1.86150912e-02
-2.24500403e-01 1.89161271e-01 -6.16840303e-01 1.09804916e+00
7.67016113e-01 -6.54632926e-01 -2.03476518e-01 -2.19285842e-02
2.14389488e-01 4.96717989e-01 3.38052094e-01 -1.27215028e+00
-2.26172253e-01 4.32705998e-01 8.39373171e-01 -1.00803816e+00
-1.48910955e-01 -4.87232357e-01 5.78381300e-01 8.24082732e-01
-2.96987176e-01 4.72187340e-01 3.85022372e-01 5.31137437e-02
-1.09574653e-01 1.13028310e-01 9.15156364e-01 -2.40017071e-01
-2.43971065e-01 4.57128257e-01 -6.97862148e-01 3.82701367e-01
1.03921795e+00 -5.21553338e-01 2.17641130e-01 -5.00947423e-02
-1.45124114e+00 -1.86471224e-01 -2.81598389e-01 1.08994558e-01
8.50134969e-01 -1.13673985e+00 -1.11872649e+00 4.64202613e-01
2.30871350e-01 -3.78173918e-01 4.46770608e-01 1.51301336e+00
-1.08958340e+00 2.38493413e-01 -6.92381024e-01 -9.44770038e-01
-1.35248160e+00 3.59063335e-02 7.75491238e-01 -1.02915771e-01
-1.32026541e+00 6.00686729e-01 -5.48861206e-01 6.70794547e-02
3.89462411e-01 -4.32621032e-01 -4.58503246e-01 2.33410984e-01
5.51797867e-01 4.78652239e-01 5.56984544e-01 -2.91544199e-01
-6.26166821e-01 5.87329090e-01 6.14686131e-01 2.23301217e-01
1.24850416e+00 2.62239762e-02 -3.25738639e-01 8.67016912e-01
6.73298776e-01 -2.54335582e-01 -2.69488990e-01 3.48222464e-01
-4.28637862e-02 -4.21587415e-02 -2.47650594e-01 -1.06556082e+00
-1.07685804e+00 1.07798135e+00 1.15425169e+00 3.70201021e-01
1.34949231e+00 -4.27845776e-01 8.67649436e-01 2.62141585e-01
5.75611234e-01 -8.17778826e-01 -4.14224535e-01 4.28712405e-02
7.70640492e-01 -6.41269028e-01 -4.06405888e-02 1.54843539e-01
-6.08881712e-01 1.38069439e+00 -1.36700362e-01 -3.13478261e-01
8.40458512e-01 4.40740377e-01 7.74193108e-01 -2.53718555e-01
-4.00036573e-01 4.02710527e-01 1.96548522e-01 1.12171400e+00
7.70650327e-01 4.17619087e-02 -7.82952428e-01 1.33461487e+00
-4.08960842e-02 4.62278515e-01 4.46158618e-01 7.57942140e-01
1.76385120e-01 -1.08962870e+00 -4.00774956e-01 7.53356695e-01
-9.56422687e-01 -2.03122988e-01 -2.29335979e-01 5.14688671e-01
6.54626548e-01 7.33377159e-01 -2.07146928e-01 -1.05530657e-01
1.76199645e-01 4.17204320e-01 6.90114439e-01 -4.00707781e-01
-1.22911835e+00 3.36218894e-01 2.68030971e-01 -2.93754041e-01
-1.85526282e-01 -7.09410727e-01 -1.20563102e+00 3.67918849e-01
-4.18191142e-02 3.99744451e-01 4.03326035e-01 7.98230410e-01
3.97799820e-01 1.01356304e+00 4.34571683e-01 -4.77154940e-01
-4.82047290e-01 -1.14483190e+00 -1.30264711e+00 4.41950440e-01
6.18746698e-01 -2.26623222e-01 -7.16748536e-02 5.89267850e-01] | [14.323134422302246, 3.2847437858581543] |
61375d87-96d8-48ed-9977-b256c11a6b8c | self-supervised-masked-convolutional | 2209.12148 | null | https://arxiv.org/abs/2209.12148v1 | https://arxiv.org/pdf/2209.12148v1.pdf | Self-Supervised Masked Convolutional Transformer Block for Anomaly Detection | Anomaly detection has recently gained increasing attention in the field of computer vision, likely due to its broad set of applications ranging from product fault detection on industrial production lines and impending event detection in video surveillance to finding lesions in medical scans. Regardless of the domain, anomaly detection is typically framed as a one-class classification task, where the learning is conducted on normal examples only. An entire family of successful anomaly detection methods is based on learning to reconstruct masked normal inputs (e.g. patches, future frames, etc.) and exerting the magnitude of the reconstruction error as an indicator for the abnormality level. Unlike other reconstruction-based methods, we present a novel self-supervised masked convolutional transformer block (SSMCTB) that comprises the reconstruction-based functionality at a core architectural level. The proposed self-supervised block is extremely flexible, enabling information masking at any layer of a neural network and being compatible with a wide range of neural architectures. In this work, we extend our previous self-supervised predictive convolutional attentive block (SSPCAB) with a 3D masked convolutional layer, as well as a transformer for channel-wise attention. Furthermore, we show that our block is applicable to a wider variety of tasks, adding anomaly detection in medical images and thermal videos to the previously considered tasks based on RGB images and surveillance videos. We exhibit the generality and flexibility of SSMCTB by integrating it into multiple state-of-the-art neural models for anomaly detection, bringing forth empirical results that confirm considerable performance improvements on five benchmarks: MVTec AD, BRATS, Avenue, ShanghaiTech, and Thermal Rare Event. We release our code and data as open source at https://github.com/ristea/ssmctb. | ['Mubarak Shah', 'Thomas B. Moeslund', 'Fahad Shahbaz Khan', 'Kamal Nasrollahi', 'Radu Tudor Ionescu', 'Nicolae-Catalin Ristea', 'Neelu Madan'] | 2022-09-25 | null | null | null | null | ['one-class-classification', 'fault-detection'] | ['miscellaneous', 'miscellaneous'] | [ 5.89602888e-01 2.07965020e-02 -1.65276174e-02 -2.25040793e-01
-3.17362398e-01 -3.08190405e-01 5.30548871e-01 1.79618984e-01
3.72627825e-02 1.50446907e-01 -3.67667109e-01 -5.05739391e-01
1.29798362e-02 -6.36316061e-01 -1.03152680e+00 -8.28737259e-01
-2.42318720e-01 5.82569614e-02 4.20436710e-01 -2.40939096e-01
1.37554973e-01 7.78979897e-01 -1.74618900e+00 6.39498413e-01
5.36585927e-01 1.73511016e+00 -1.42408028e-01 6.01868808e-01
1.22785345e-01 9.94410038e-01 -6.29251599e-01 -3.34958136e-01
1.83764428e-01 -2.87995189e-01 -6.86202884e-01 4.97722268e-01
4.71300244e-01 -3.47959667e-01 -2.24842086e-01 8.98018360e-01
1.10523835e-01 -1.57620192e-01 4.48515862e-01 -1.46400857e+00
-4.60479259e-01 2.31360465e-01 -4.99981403e-01 4.16439593e-01
3.63189608e-01 5.75530827e-01 7.67143428e-01 -9.11838353e-01
3.87014121e-01 8.64118695e-01 6.83727920e-01 4.52600628e-01
-1.04196036e+00 -3.92641038e-01 4.53529119e-01 6.23125553e-01
-1.00946677e+00 -1.69499829e-01 8.98504913e-01 -6.02549136e-01
1.16519582e+00 3.03070873e-01 6.14330709e-01 1.47712576e+00
5.14716446e-01 8.94874096e-01 7.73097754e-01 -2.60662645e-01
4.49722588e-01 -2.30962440e-01 3.26772034e-02 7.46919036e-01
2.67297655e-01 1.66925006e-02 -4.63398248e-01 -1.68976448e-02
6.09267354e-01 4.22231317e-01 -2.07522452e-01 -3.81605566e-01
-1.09817302e+00 5.69284618e-01 4.76737320e-01 2.68639356e-01
-6.64034367e-01 2.70278037e-01 6.51157856e-01 4.78331357e-01
5.14967382e-01 3.65766406e-01 -4.93226290e-01 1.03192687e-01
-6.88506722e-01 -5.44104241e-02 5.46266675e-01 5.62389672e-01
5.74799120e-01 5.80665708e-01 -2.75176734e-01 4.90009665e-01
1.44406661e-01 2.17444941e-01 5.40673971e-01 -7.02945650e-01
5.65104969e-02 8.45508754e-01 -1.71055481e-01 -8.18787634e-01
-2.96911955e-01 -5.20414293e-01 -8.83160293e-01 7.20956147e-01
2.50823081e-01 1.95438609e-01 -1.25123811e+00 1.32935357e+00
2.05414727e-01 4.24343914e-01 -1.81238283e-03 8.44836950e-01
4.16583031e-01 5.22736907e-01 -1.71609715e-01 -1.34152800e-01
1.21792781e+00 -8.76732528e-01 -6.61445796e-01 -3.08662862e-01
6.00912750e-01 -6.34958506e-01 8.94551337e-01 1.08998203e+00
-8.96966398e-01 -6.14792705e-01 -1.25575876e+00 2.98914164e-01
-6.08510554e-01 6.94838539e-02 5.37242353e-01 5.00943661e-01
-9.13379967e-01 7.54748821e-01 -1.07021677e+00 -3.26211333e-01
6.54175639e-01 3.38266969e-01 -3.46268982e-01 -8.97767693e-02
-9.41937685e-01 7.02794552e-01 2.64380664e-01 4.34834063e-01
-1.22984958e+00 -4.78072733e-01 -9.89266038e-01 -8.99076983e-02
5.31849086e-01 -3.65394950e-01 1.14538074e+00 -1.43871689e+00
-1.26066804e+00 7.91649878e-01 9.68622789e-02 -8.85127187e-01
4.52691466e-01 -4.64183927e-01 -7.56289899e-01 4.65360768e-02
-1.07811682e-01 4.25308108e-01 1.29968655e+00 -1.00794399e+00
-5.86366951e-01 -3.80186111e-01 -2.44208276e-02 -2.84660876e-01
-3.49680930e-01 -9.43922624e-02 -2.26596236e-01 -9.35060680e-01
1.31618932e-01 -7.22510517e-01 -3.22601348e-01 2.48181984e-01
-6.16947830e-01 -6.83823600e-02 1.31590438e+00 -6.69698775e-01
1.15511334e+00 -2.26062608e+00 1.27724394e-01 2.13556916e-01
3.35064754e-02 3.59427094e-01 4.25144695e-02 9.06480402e-02
-6.47096634e-01 -3.15188468e-01 -7.68358707e-01 -2.26202354e-01
-1.98400602e-01 2.86157697e-01 -3.70931953e-01 6.30967081e-01
8.00089419e-01 7.66543448e-01 -6.64435923e-01 -9.64809442e-05
5.85491002e-01 2.75739223e-01 -3.34307253e-01 1.15133196e-01
-4.42587316e-01 5.22341430e-01 -2.75121391e-01 1.27272558e+00
5.35922170e-01 -1.66970804e-01 -3.11817825e-01 -1.53998688e-01
3.60327214e-02 -3.95678431e-02 -1.01476264e+00 1.70421970e+00
-3.03767741e-01 5.45494974e-01 1.04006134e-01 -1.26261604e+00
8.20860088e-01 3.49895298e-01 6.63351655e-01 -5.64362586e-01
7.85597339e-02 2.21873179e-01 9.55122635e-02 -5.74566424e-01
3.37909073e-01 2.52638012e-01 -4.79072407e-02 1.52874917e-01
1.18957587e-01 1.65385202e-01 1.32356293e-03 -9.36835036e-02
1.53821886e+00 2.31845573e-01 2.97419846e-01 -2.52921972e-02
7.41524041e-01 9.56359208e-02 4.49458331e-01 4.97653931e-01
-2.37480655e-01 7.46325314e-01 7.77123511e-01 -6.86075628e-01
-9.32772458e-01 -1.16495574e+00 -2.76811838e-01 6.81397319e-01
-1.70180501e-04 -1.45815402e-01 -6.18603885e-01 -1.02696109e+00
5.88272326e-02 6.75709128e-01 -7.09883153e-01 -4.76569772e-01
-5.97081423e-01 -7.77338147e-01 3.94617140e-01 7.16696620e-01
4.68813449e-01 -1.28777349e+00 -6.86176836e-01 1.06134184e-01
2.53989816e-01 -1.16498840e+00 5.68265505e-02 5.25518477e-01
-8.63309264e-01 -1.25360537e+00 -3.47458392e-01 -6.05954289e-01
5.11276722e-01 -2.07854643e-01 1.16434669e+00 5.77085465e-02
-7.48158813e-01 5.85987926e-01 -2.87839055e-01 -4.75202799e-01
-6.00134850e-01 -3.17818522e-01 1.08015418e-01 4.94989485e-01
2.84758657e-01 -4.88294303e-01 -6.59047842e-01 4.22596246e-01
-1.13235557e+00 -3.62643033e-01 7.34583914e-01 8.67289007e-01
7.65057445e-01 -4.18538712e-02 3.61387879e-01 -7.79969871e-01
8.23535770e-02 -7.57548094e-01 -6.15844965e-01 -5.28720692e-02
-5.13463438e-01 -3.00093383e-01 6.85973287e-01 -3.83555740e-01
-7.56119668e-01 3.42071503e-01 -5.03014207e-01 -8.69990587e-01
-7.13334322e-01 1.36524215e-01 -2.29052782e-01 7.51962066e-02
6.76910758e-01 2.83912688e-01 2.21183494e-01 -2.85082549e-01
-2.21455973e-02 3.59192014e-01 7.44728446e-01 -1.47150993e-01
6.54421628e-01 5.95845163e-01 1.70095459e-01 -8.15228939e-01
-4.76583123e-01 -4.45495009e-01 -5.44030428e-01 -4.16762143e-01
8.88375878e-01 -6.96842849e-01 -6.02302372e-01 8.95898342e-01
-1.02549994e+00 -3.11980307e-01 -6.99916124e-01 2.19587520e-01
-5.33152580e-01 3.91165286e-01 -5.73954523e-01 -7.13613868e-01
-2.13427395e-01 -1.28883421e+00 1.16981196e+00 -1.68409899e-01
-1.37985244e-01 -8.14288259e-01 -3.09801221e-01 9.99739617e-02
3.61146212e-01 7.31321335e-01 8.53639662e-01 -9.37445223e-01
-7.66121686e-01 -4.59024131e-01 1.73154399e-01 8.45456481e-01
2.52126336e-01 -1.28281891e-01 -1.20396841e+00 -2.23248765e-01
2.52115488e-01 -2.11512804e-01 1.05337071e+00 4.38052058e-01
1.61331844e+00 -1.56896636e-01 -3.65328908e-01 5.83611548e-01
1.27302778e+00 2.35372886e-01 9.81113493e-01 4.70305026e-01
9.11136925e-01 3.49988431e-01 6.67111278e-01 3.76351893e-01
-3.37876171e-01 7.12268472e-01 1.34808087e+00 -2.74002731e-01
1.03399986e-02 4.19095695e-01 7.58978844e-01 1.70844778e-01
-1.52212158e-02 -4.03971314e-01 -7.46965110e-01 5.30295253e-01
-1.83226252e+00 -8.45806599e-01 -2.65883595e-01 2.16487646e+00
2.77482510e-01 3.26095700e-01 4.89088669e-02 6.18745863e-01
7.26153851e-01 1.51364550e-01 -9.17659283e-01 -6.23645842e-01
-8.95614177e-02 2.33625561e-01 2.77830213e-01 -3.51926917e-03
-1.40081406e+00 4.94330615e-01 5.38555431e+00 6.81902349e-01
-1.21311522e+00 5.45405895e-02 1.00523961e+00 -1.03588030e-01
2.05517516e-01 -6.08633101e-01 -4.35583711e-01 5.32034159e-01
9.08274293e-01 6.77650273e-01 1.38601094e-01 1.02596724e+00
-3.74198072e-02 -1.14949167e-01 -1.40374923e+00 7.36524463e-01
1.65480703e-01 -1.16869485e+00 3.88198569e-02 -8.73391107e-02
5.04310369e-01 4.00622463e-04 2.70708144e-01 1.03957810e-01
-1.85038388e-01 -1.06484509e+00 7.53167629e-01 3.38757157e-01
6.58593476e-01 -6.29226983e-01 8.97237420e-01 -1.02385081e-01
-1.02013731e+00 -4.61530000e-01 7.36476928e-02 8.23139325e-02
4.46764231e-02 7.05275953e-01 -7.02443182e-01 5.79006135e-01
1.00133860e+00 9.09943759e-01 -5.83603919e-01 8.08876932e-01
-1.27631724e-01 7.33990490e-01 -2.03600898e-01 4.62245971e-01
3.10261756e-01 1.82547063e-01 7.81021416e-01 1.19615459e+00
4.23308074e-01 -4.49991196e-01 3.25277895e-02 8.71810853e-01
2.57698238e-01 -2.48062849e-01 -7.16374993e-01 2.44580939e-01
-1.70549199e-01 1.18766820e+00 -8.22462976e-01 -1.13155343e-01
-6.06484771e-01 1.36758804e+00 -1.46434531e-01 1.63671836e-01
-1.08873677e+00 -2.09290028e-01 9.56812799e-01 3.15678805e-01
7.09164798e-01 1.51174411e-01 -3.57986271e-01 -8.56504261e-01
4.08826679e-01 -8.89378309e-01 6.09713793e-01 -6.53556049e-01
-1.32126153e+00 7.43409991e-01 -1.88561752e-01 -1.59470117e+00
-3.82938772e-01 -1.26663935e+00 -7.18163908e-01 4.87957895e-01
-1.38191581e+00 -1.07630205e+00 -4.61698622e-01 7.35854089e-01
8.40409398e-01 -3.58003050e-01 6.94432974e-01 2.30467886e-01
-8.04295421e-01 4.74207997e-01 -2.04239577e-01 1.11705214e-01
4.88413453e-01 -1.38289189e+00 6.07719004e-01 1.33395171e+00
-6.86162934e-02 6.79388642e-02 5.05297422e-01 -6.16421759e-01
-1.44039094e+00 -1.43420219e+00 -9.53012519e-03 -5.40549874e-01
7.29665518e-01 -3.54137391e-01 -1.17234528e+00 7.83814847e-01
1.55657336e-01 4.90170598e-01 2.07001939e-01 -3.43677610e-01
-2.96354830e-01 -2.53675520e-01 -1.25954878e+00 3.47418010e-01
9.78370667e-01 -2.20105365e-01 -1.89515486e-01 4.44847822e-01
6.35038674e-01 -6.26522064e-01 -6.63192213e-01 7.31545866e-01
1.06828161e-01 -1.20613658e+00 9.92890716e-01 -4.24519897e-01
7.05337167e-01 -5.59446037e-01 -1.44656301e-01 -1.08635736e+00
-1.22491077e-01 -4.38362867e-01 -7.08450317e-01 8.60739946e-01
4.24047768e-01 -7.94206977e-01 7.51891553e-01 1.06412143e-01
-7.91475177e-01 -1.14915311e+00 -1.02874470e+00 -7.55973697e-01
-4.48164135e-01 -8.37685406e-01 4.08464760e-01 7.84217000e-01
-4.03776497e-01 -1.49151355e-01 -3.31957936e-01 5.75880826e-01
2.58642972e-01 -1.35048434e-01 3.41082990e-01 -1.05685234e+00
-4.60717589e-01 -3.89200479e-01 -1.00690043e+00 -6.62891924e-01
-1.00020386e-01 -6.95020914e-01 1.57198086e-01 -1.14724374e+00
-3.32478732e-01 -7.57578015e-02 -6.09879017e-01 7.47926891e-01
9.04079154e-02 5.95295846e-01 -3.18852931e-01 5.93401790e-02
-5.21985650e-01 2.52394676e-01 8.09335828e-01 -2.84978062e-01
2.76582632e-02 2.54532754e-01 -2.45987087e-01 8.54300141e-01
7.33178198e-01 -1.30817965e-01 -2.39255339e-01 -1.35216206e-01
-3.23525779e-02 -3.29334021e-01 9.68863308e-01 -1.40051281e+00
-4.95325355e-03 3.08870673e-01 6.19989514e-01 -5.32075465e-01
4.41201299e-01 -1.12890029e+00 9.45478119e-03 8.03833187e-01
7.97250250e-04 3.56664300e-01 4.83023196e-01 7.96719551e-01
-2.84625620e-01 -7.10043758e-02 6.51994705e-01 -6.57753870e-02
-1.18625891e+00 2.67227054e-01 -5.34441650e-01 -4.37093914e-01
1.41272497e+00 -5.42086124e-01 -2.29563087e-01 -8.37740079e-02
-8.81875157e-01 1.01574203e-02 4.16018039e-01 5.53606331e-01
8.19129109e-01 -1.17383075e+00 -5.11948943e-01 5.65256238e-01
3.60578507e-01 1.77921385e-01 3.46804798e-01 9.83388782e-01
-5.62418938e-01 1.79095864e-01 -3.61005157e-01 -1.20345032e+00
-1.08109426e+00 8.38714540e-01 5.45196354e-01 -1.42289922e-01
-7.86343455e-01 5.67418575e-01 1.59617662e-01 7.58670047e-02
4.58908647e-01 -6.18706942e-01 8.20044428e-03 -2.20180139e-01
4.95151490e-01 3.86739969e-01 6.72704279e-01 -4.31580186e-01
-4.19577360e-01 2.76246697e-01 -1.50588363e-01 4.97765511e-01
1.18395245e+00 2.23795027e-01 -9.33785141e-02 5.25573671e-01
8.63706946e-01 -4.25401270e-01 -1.38500381e+00 1.69686861e-02
1.04028592e-02 -1.15275592e-01 1.61364600e-02 -8.14637244e-01
-1.46194315e+00 6.51463151e-01 1.06905913e+00 4.17913556e-01
1.64258397e+00 5.26211895e-02 6.79684758e-01 3.73484612e-01
1.50045082e-01 -7.59446979e-01 4.54367250e-01 2.53404140e-01
9.25534129e-01 -1.24639940e+00 -7.61861876e-02 -4.86942381e-01
-6.74354792e-01 1.18595231e+00 8.77493501e-01 -1.86213806e-01
6.08060837e-01 4.93670642e-01 -3.78152803e-02 -3.60103428e-01
-5.84247231e-01 -4.82422374e-02 3.39898080e-01 5.79743624e-01
2.64997542e-01 -2.30156004e-01 3.53193879e-01 4.29736316e-01
3.38447601e-01 -2.28048280e-01 4.62132424e-01 1.08290064e+00
-1.59396335e-01 -7.94677556e-01 -5.44208288e-01 6.12376511e-01
-6.80564344e-01 7.96966404e-02 -1.67922378e-01 7.56280839e-01
4.46681142e-01 7.23759949e-01 5.19592524e-01 -4.05945033e-01
4.13108021e-01 2.64131844e-01 1.80275396e-01 -4.57533121e-01
-6.45415902e-01 -1.58253580e-01 -4.12137955e-02 -1.10597718e+00
-3.07477802e-01 -7.41038144e-01 -9.76170421e-01 4.40527759e-02
-2.13192314e-01 -4.06986475e-01 5.82725406e-01 9.73231196e-01
3.49806696e-01 1.07459867e+00 5.19573390e-01 -8.81103873e-01
-1.42922506e-01 -8.56335282e-01 -4.52751666e-01 4.57536668e-01
6.95475757e-01 -6.40213907e-01 -2.97915936e-01 9.46893319e-02] | [7.668502330780029, 1.9458937644958496] |
107f7eac-22ae-45b2-b1d3-df93ee342bb5 | towards-ethical-content-based-detection-of | 1908.11030 | null | https://arxiv.org/abs/1908.11030v1 | https://arxiv.org/pdf/1908.11030v1.pdf | Towards Ethical Content-Based Detection of Online Influence Campaigns | The detection of clandestine efforts to influence users in online communities is a challenging problem with significant active development. We demonstrate that features derived from the text of user comments are useful for identifying suspect activity, but lead to increased erroneous identifications when keywords over-represented in past influence campaigns are present. Drawing on research in native language identification (NLI), we use "named entity masking" (NEM) to create sentence features robust to this shortcoming, while maintaining comparable classification accuracy. We demonstrate that while NEM consistently reduces false positives when key named entities are mentioned, both masked and unmasked models exhibit increased false positive rates on English sentences by Russian native speakers, raising ethical considerations that should be addressed in future research. | ['Herna Viktor', 'Evan Crothers', 'Nathalie Japkowicz'] | 2019-08-29 | null | null | null | null | ['native-language-identification'] | ['natural-language-processing'] | [ 4.76527512e-01 7.13672936e-02 -4.58607227e-01 -2.39493966e-01
-9.90053952e-01 -7.66547084e-01 1.05263042e+00 3.05350631e-01
-6.88816547e-01 7.55143046e-01 5.93207657e-01 -8.47316146e-01
7.77116120e-02 -6.61587536e-01 -3.27027500e-01 -1.79603100e-02
1.38270900e-01 -1.38295749e-02 7.73705542e-02 1.59051642e-01
7.49639690e-01 3.66417676e-01 -9.62904930e-01 4.43241298e-01
9.87194061e-01 -1.79028809e-01 -3.56800139e-01 6.96052372e-01
-1.30099967e-01 1.18107784e+00 -1.03873003e+00 -8.36632371e-01
4.24636230e-02 -2.33444676e-01 -9.56716716e-01 5.30126989e-02
8.72605741e-01 -2.67332882e-01 -5.02496600e-01 1.08851111e+00
2.86609173e-01 -3.57906997e-01 6.01241291e-01 -9.19565618e-01
-7.90587366e-01 1.07343268e+00 -7.36688316e-01 8.56525242e-01
6.98365688e-01 1.23768918e-01 1.16240287e+00 -8.84345531e-01
8.86042774e-01 1.21596253e+00 9.93698239e-01 2.96127945e-01
-1.59404957e+00 -9.13186193e-01 5.29747643e-02 -1.44371182e-01
-1.51210356e+00 -8.12467158e-01 2.65472859e-01 -6.43431187e-01
9.05086160e-01 4.18787092e-01 3.32611561e-01 1.42660487e+00
-1.74901605e-01 5.95310748e-01 1.18765879e+00 -2.60366201e-01
-3.60393018e-01 6.38011336e-01 5.24162531e-01 4.65080321e-01
7.75231004e-01 -2.29115486e-01 -7.07705140e-01 -1.09183311e+00
2.71529686e-02 -2.63979167e-01 -1.40013710e-01 4.90265399e-01
-1.03901875e+00 1.09512973e+00 -2.41392329e-01 7.74671018e-01
-1.69303223e-01 -2.95600802e-01 2.66330361e-01 4.10143793e-01
7.40220964e-01 8.84266973e-01 -3.75599027e-01 -5.10121346e-01
-1.27134860e+00 1.88923717e-01 1.22560775e+00 5.76985121e-01
6.22665882e-01 -2.34470442e-02 -2.11814508e-01 9.38484967e-01
1.77861243e-01 7.53366530e-01 4.33374763e-01 -4.40646231e-01
6.07109666e-01 7.69854248e-01 2.35000327e-01 -1.45113575e+00
-4.41202551e-01 -6.69127643e-01 -2.96543002e-01 -3.20867836e-01
5.15363634e-01 -5.33786833e-01 -3.69332165e-01 1.54281807e+00
9.97563824e-02 -4.32735942e-02 -2.69497782e-01 2.02613279e-01
3.23624611e-01 2.82049835e-01 4.14449155e-01 -2.38555178e-01
1.38032150e+00 -2.28563324e-01 -7.51765370e-01 -6.42183602e-01
8.60560596e-01 -8.67861927e-01 5.72044849e-01 -7.14865774e-02
-7.53991783e-01 -3.20195556e-02 -8.22458804e-01 2.16879860e-01
-3.69275033e-01 2.76765507e-02 4.81145054e-01 1.44769490e+00
-8.87008965e-01 3.95004451e-01 -4.05338168e-01 -5.20622075e-01
5.83130062e-01 1.83866978e-01 -4.06480342e-01 2.97791421e-01
-1.26998723e+00 9.91803408e-01 -2.93577909e-01 -1.24649800e-01
-5.77523768e-01 -7.82793641e-01 -4.90926117e-01 -1.95438385e-01
4.30501640e-01 1.97447747e-01 1.17497861e+00 -8.47521365e-01
-4.15887117e-01 1.16442335e+00 -5.82241595e-01 -4.59287524e-01
3.43647599e-01 -1.40980586e-01 -1.01364303e+00 2.10338891e-01
6.49415612e-01 -2.03177467e-01 1.00779128e+00 -7.67878771e-01
-6.75445855e-01 -4.98674780e-01 -1.58084616e-01 -1.24458380e-01
-7.49562323e-01 9.10640776e-01 3.18303049e-01 -6.80264831e-01
-1.34471685e-01 -8.86109650e-01 -1.17308244e-01 -5.66973448e-01
-6.34270072e-01 -2.23635957e-01 8.14362347e-01 -1.23511803e+00
1.81253898e+00 -1.80212295e+00 -6.58712864e-01 5.16077578e-01
6.38155758e-01 5.42928100e-01 1.00191243e-01 4.54736918e-01
1.26638606e-01 1.17241871e+00 1.03091411e-01 -1.45770404e-02
-2.44902059e-01 -5.31475663e-01 -4.13075715e-01 7.68312097e-01
2.88217604e-01 8.65870059e-01 -8.04560006e-01 -3.98341179e-01
-4.45006132e-01 1.88496396e-01 -3.06590378e-01 -4.44773316e-01
3.24707806e-01 1.16171930e-02 -4.72006232e-01 7.79636383e-01
5.55421472e-01 -5.01539588e-01 4.30647463e-01 3.76642585e-01
-3.37919891e-01 9.49443460e-01 -9.42507148e-01 3.75446826e-01
-2.82195151e-01 1.25190711e+00 6.54703259e-01 -3.94122392e-01
6.58543050e-01 2.45706335e-01 -2.62049790e-02 -1.27997220e-01
9.26205665e-02 2.41851345e-01 2.83977211e-01 -5.30525446e-01
7.44081795e-01 2.05649704e-01 7.47798532e-02 8.48406851e-01
-5.00043154e-01 5.32223344e-01 4.58735257e-01 7.10650623e-01
1.43842530e+00 -8.80873501e-01 3.03997934e-01 -2.93290079e-01
5.74445665e-01 3.04838449e-01 4.50360566e-01 1.37941432e+00
-4.15315688e-01 9.42312777e-02 6.15952969e-01 2.04997212e-01
-9.91822720e-01 -6.56121075e-01 -3.92892003e-01 1.17600548e+00
-4.42786902e-01 -7.91893244e-01 -6.73432887e-01 -9.70015764e-01
2.19088748e-01 7.56051004e-01 -4.63177830e-01 2.61779483e-02
-5.28085411e-01 -1.02117157e+00 1.10004652e+00 3.41427810e-02
9.81025845e-02 -6.16781235e-01 -1.54372156e-02 3.38768363e-01
-1.90249562e-01 -1.19223166e+00 -4.92823899e-01 -3.16897452e-01
-4.07414049e-01 -1.44996262e+00 -5.73998153e-01 -5.07587850e-01
7.57707834e-01 4.84741151e-01 9.83045638e-01 4.90150183e-01
-4.03308332e-01 5.45786798e-01 -2.75249720e-01 -3.67544144e-01
-1.04282820e+00 4.23648030e-01 5.35578988e-02 9.42498147e-02
1.18441045e+00 -3.29252452e-01 5.05263358e-02 1.62719443e-01
-6.53824329e-01 -5.12587249e-01 5.09944320e-01 5.00993907e-01
-5.49864054e-01 -2.17749953e-01 7.01874793e-01 -1.53113151e+00
1.13324523e+00 -7.19114542e-01 -1.39208242e-01 2.78390050e-01
-7.10065484e-01 -2.86592454e-01 2.40466058e-01 -7.37832308e-01
-1.13539004e+00 -5.44413686e-01 9.80390795e-03 3.84770721e-01
-1.44528925e-01 3.84802580e-01 2.64746100e-01 -3.26761246e-01
1.17016101e+00 9.03935060e-02 1.89303860e-01 -5.81101000e-01
-1.60392135e-01 1.22290361e+00 -4.02031913e-02 -1.71984375e-01
1.13090396e+00 4.07038391e-01 -7.05893874e-01 -1.35303342e+00
-8.75774145e-01 -6.73669457e-01 -4.93751228e-01 -8.19962546e-02
5.19258976e-01 -1.00826514e+00 -5.66236258e-01 4.36516285e-01
-1.10813344e+00 3.45863938e-01 4.81193334e-01 4.24682736e-01
4.21158254e-01 7.01057255e-01 -8.37767422e-01 -1.10613084e+00
-2.62836397e-01 -7.15127826e-01 5.63516319e-01 1.09999962e-01
-9.92440701e-01 -8.94049823e-01 9.24629867e-02 8.17449033e-01
4.82792139e-01 -8.54758024e-02 6.63244665e-01 -1.55543280e+00
-1.94872156e-01 -6.31231725e-01 -1.87884673e-01 4.78559323e-02
1.85037494e-01 3.94470125e-01 -8.25451195e-01 -2.57002175e-01
-6.09676577e-02 -7.79697374e-02 6.41424239e-01 -1.54451221e-01
3.94598871e-01 -8.08184206e-01 -6.92288339e-01 -1.65307745e-01
7.62609184e-01 -2.19741818e-02 2.33777046e-01 2.88240284e-01
5.91734827e-01 6.43569767e-01 1.40001122e-02 5.20306110e-01
1.01849489e-01 2.37520054e-01 -3.41829479e-01 1.61875054e-01
2.49268234e-01 -5.49723387e-01 5.12479424e-01 3.78313899e-01
3.32787007e-01 -2.90368851e-02 -1.20898879e+00 7.72137702e-01
-1.05007184e+00 -1.38755333e+00 -6.90297306e-01 2.00731635e+00
1.03387165e+00 3.76158923e-01 2.66725600e-01 1.20954894e-01
1.39165390e+00 3.22898448e-01 -2.21674412e-01 -3.30779135e-01
-4.22808707e-01 2.29298890e-01 9.28438485e-01 7.07047164e-01
-1.18990278e+00 6.52955115e-01 7.16893196e+00 6.37004852e-01
-7.60172188e-01 3.07645679e-01 5.10701239e-01 -1.32109433e-01
-3.95047188e-01 2.70504355e-01 -1.31741989e+00 6.18722320e-01
1.24231660e+00 -5.73121965e-01 2.72964180e-01 6.73272908e-01
3.17496836e-01 -9.38993543e-02 -6.01547956e-01 4.35908765e-01
3.75694305e-01 -1.39212155e+00 -1.10605069e-01 4.67420608e-01
9.57094371e-01 1.48942441e-01 2.68518608e-02 2.18769521e-01
4.67108369e-01 -8.18568230e-01 4.89048243e-01 5.53445742e-02
4.96104032e-01 -4.63391423e-01 3.46505910e-01 5.87533057e-01
-6.19064927e-01 -3.32483470e-01 -1.17874071e-01 -2.41620749e-01
1.04287922e-01 8.54323387e-01 -1.26537025e+00 -1.63103402e-01
2.18215808e-01 5.64926565e-01 -1.09460175e+00 8.13860834e-01
-2.76501149e-01 1.25047266e+00 -3.33288133e-01 -4.60610330e-01
1.93833008e-01 2.02065110e-01 9.34694648e-01 1.33407307e+00
-1.58557519e-01 -3.22195739e-01 -2.07032889e-01 8.48277688e-01
-3.38556617e-01 2.27062955e-01 -8.80219460e-01 -8.89780581e-01
6.62646949e-01 1.33006549e+00 -6.48157895e-01 -4.40461874e-01
-6.46496713e-01 7.08297372e-01 3.80769789e-01 3.30931842e-01
-3.14372927e-01 -4.57178324e-01 6.54710650e-01 7.42307067e-01
-2.32118014e-02 -1.74919724e-01 -3.70702446e-01 -1.21251094e+00
-1.07582025e-01 -1.13307142e+00 2.85528988e-01 -2.53977835e-01
-1.40738833e+00 1.60960376e-01 -2.28821471e-01 -6.36055589e-01
-2.81603903e-01 -3.22992116e-01 -6.33028746e-01 9.45761979e-01
-7.79127836e-01 -8.06621194e-01 2.49564141e-01 -1.53668849e-02
1.86001167e-01 -1.73261896e-01 4.86289352e-01 4.66199338e-01
-7.71087050e-01 8.43081832e-01 7.48108476e-02 6.94668651e-01
8.40661407e-01 -8.72713387e-01 6.99872196e-01 9.80007172e-01
4.11089599e-01 1.41380751e+00 5.99987805e-01 -1.28857040e+00
-9.75982845e-01 -8.16806555e-01 1.81668484e+00 -8.84516001e-01
1.22460115e+00 -4.51529205e-01 -8.74177873e-01 8.26823413e-01
1.14585213e-01 -8.06893408e-01 1.06549346e+00 5.45829952e-01
-6.28258705e-01 6.63923383e-01 -1.16466320e+00 7.73798823e-01
1.10834849e+00 -1.20634043e+00 -7.73244441e-01 6.20211542e-01
2.84597665e-01 2.26684883e-01 -7.28170991e-01 7.24925399e-02
3.76712114e-01 -6.84462011e-01 5.98467886e-01 -9.76829290e-01
1.81453750e-01 1.49494827e-01 5.22936583e-01 -8.23109627e-01
-4.34246153e-01 -8.97076428e-01 6.53525442e-02 1.67164898e+00
1.02004170e+00 -6.82904899e-01 8.04072082e-01 1.04633462e+00
5.62947273e-01 -1.75322428e-01 -8.32023084e-01 -3.96987826e-01
2.40822285e-02 -2.23337889e-01 2.40304004e-02 1.36001492e+00
3.71684462e-01 5.60206592e-01 -3.10252070e-01 1.93872899e-01
4.71636504e-01 -5.12969971e-01 6.12027824e-01 -1.35433424e+00
1.33571413e-03 -4.00484473e-01 -1.99390694e-01 -5.99783540e-01
5.02591372e-01 -8.30576122e-01 -5.03977239e-01 -1.04732978e+00
5.47353268e-01 -1.42073587e-01 3.81714553e-01 3.92692506e-01
-3.60655218e-01 4.13544595e-01 8.46334547e-02 3.58469069e-01
-2.78777599e-01 -1.11770451e-01 5.87542176e-01 -3.94415081e-01
-2.55435407e-01 3.29644561e-01 -1.34474111e+00 9.12371993e-01
7.26767480e-01 -8.79885435e-01 1.74591824e-01 3.41615379e-02
4.60647643e-01 -4.76416767e-01 2.23738894e-01 -7.82792926e-01
2.50496805e-01 -9.54598412e-02 3.73602033e-01 -4.79688793e-01
-1.17295749e-01 -4.00297880e-01 -1.21648803e-01 4.98125046e-01
-6.61402285e-01 1.98747069e-01 -1.20637137e-02 5.50218463e-01
8.35058913e-02 -7.41809428e-01 5.53422272e-01 -2.67223686e-01
-1.67902172e-01 -2.24171847e-01 -1.30430996e+00 2.77055025e-01
6.34860575e-01 -1.58215240e-01 -6.75972998e-01 -5.22646129e-01
-5.05340099e-01 -9.40752104e-02 5.20505190e-01 3.75276893e-01
1.08416960e-01 -8.04482579e-01 -9.34848785e-01 1.02089666e-01
2.53639035e-02 -1.23321676e+00 7.51898661e-02 9.01538432e-01
-1.73454791e-01 4.33547705e-01 4.61525023e-01 1.18993238e-01
-1.63442171e+00 1.84644341e-01 5.85650019e-02 -1.58529162e-01
-1.16721965e-01 6.85760796e-01 -4.57365125e-01 -3.53287160e-01
2.64178719e-02 4.11309272e-01 -2.63488352e-01 4.18123722e-01
9.88117337e-01 7.78243005e-01 -2.54621636e-02 -1.02522159e+00
-5.08920074e-01 -3.19714725e-01 -8.23158443e-01 -1.74154639e-01
8.95262241e-01 -1.32178605e-01 -3.81915718e-01 1.15451425e-01
1.21539092e+00 1.05573773e+00 -2.09003195e-01 -2.87067860e-01
5.81351876e-01 -6.27744317e-01 -8.63551423e-02 -5.76350749e-01
-4.44597781e-01 3.16834629e-01 1.29827410e-01 5.73349118e-01
8.23416784e-02 -2.18806081e-02 4.20834184e-01 4.92298990e-01
-1.15700187e-02 -1.13055539e+00 -3.65394801e-01 5.77184677e-01
4.02875096e-01 -1.13245630e+00 3.94434780e-01 -4.94486272e-01
-4.70731974e-01 7.32937992e-01 6.82840168e-01 2.66522765e-01
6.94055080e-01 2.86504120e-01 6.28797384e-03 -2.07682967e-01
-5.26965201e-01 -9.21485871e-02 1.32626757e-01 5.04980445e-01
5.96116185e-01 -1.34112060e-01 -1.02257204e+00 5.88283598e-01
-1.55174613e-01 -5.70168555e-01 1.02751994e+00 9.71564770e-01
-4.41261441e-01 -1.07643366e+00 -7.40747035e-01 1.22745144e+00
-1.35316253e+00 -5.40845573e-01 -1.23253620e+00 6.60896897e-01
-2.06491187e-01 1.40919626e+00 7.43074194e-02 -4.18032706e-01
-5.64102456e-02 2.92429924e-01 -1.37376383e-01 -8.44794631e-01
-1.41002297e+00 -2.72713214e-01 9.20911908e-01 -2.71461043e-03
-4.61649686e-01 -9.29855049e-01 -5.98783970e-01 -6.12256467e-01
-5.77163458e-01 4.71622437e-01 5.34419894e-01 9.41916347e-01
5.70322692e-01 -2.67496914e-01 6.09617591e-01 4.75065671e-02
-8.58079016e-01 -1.38347590e+00 -4.94849056e-01 4.72787291e-01
1.61016390e-01 -3.55170012e-01 -9.49726522e-01 -2.01046512e-01] | [8.583666801452637, 10.344955444335938] |
6ecba356-4680-44b5-82f0-c9118bf56c8f | neural-machine-translation-with-phrase-level | 2203.10299 | null | https://arxiv.org/abs/2203.10299v1 | https://arxiv.org/pdf/2203.10299v1.pdf | Neural Machine Translation with Phrase-Level Universal Visual Representations | Multimodal machine translation (MMT) aims to improve neural machine translation (NMT) with additional visual information, but most existing MMT methods require paired input of source sentence and image, which makes them suffer from shortage of sentence-image pairs. In this paper, we propose a phrase-level retrieval-based method for MMT to get visual information for the source input from existing sentence-image data sets so that MMT can break the limitation of paired sentence-image input. Our method performs retrieval at the phrase level and hence learns visual information from pairs of source phrase and grounded region, which can mitigate data sparsity. Furthermore, our method employs the conditional variational auto-encoder to learn visual representations which can filter redundant visual information and only retain visual information related to the phrase. Experiments show that the proposed method significantly outperforms strong baselines on multiple MMT datasets, especially when the textual context is limited. | ['Yang Feng', 'Qingkai Fang'] | 2022-03-19 | null | https://aclanthology.org/2022.acl-long.390 | https://aclanthology.org/2022.acl-long.390.pdf | acl-2022-5 | ['multimodal-machine-translation'] | ['natural-language-processing'] | [ 4.22865391e-01 -1.52618706e-01 -4.62234586e-01 4.87058908e-02
-1.25294733e+00 -5.60909748e-01 4.30535316e-01 -2.82839358e-01
-3.36980999e-01 5.96209168e-01 2.95305699e-01 -3.58134925e-01
5.02072990e-01 -5.88045239e-01 -1.06996787e+00 -7.94881821e-01
7.59035528e-01 1.71958268e-01 2.73972340e-02 -2.04963133e-01
1.69302210e-01 2.99800988e-02 -9.36587512e-01 8.60374451e-01
8.88603270e-01 7.61271238e-01 9.13905025e-01 2.67251402e-01
-4.22428250e-01 4.45541501e-01 -5.20355046e-01 -3.37254167e-01
3.29024762e-01 -6.10598743e-01 -4.96307135e-01 3.02653044e-01
8.61179054e-01 -5.39144814e-01 -5.64286649e-01 9.83548522e-01
5.94590724e-01 7.22153857e-02 7.31207550e-01 -1.11863708e+00
-1.32900250e+00 2.68798888e-01 -1.14658248e+00 1.74937412e-01
2.71878362e-01 3.51163656e-01 9.48613584e-01 -1.55545545e+00
9.71418262e-01 1.59155059e+00 5.98832630e-02 6.80709600e-01
-1.33472764e+00 -6.69429719e-01 3.26735348e-01 2.05552697e-01
-1.23874187e+00 -3.69125694e-01 8.75519693e-01 -1.14193678e-01
9.03641403e-01 9.53409076e-02 6.06946349e-01 1.34308136e+00
2.56982267e-01 1.31605136e+00 1.26805675e+00 -4.81898934e-01
-1.70114830e-01 2.93015659e-01 -5.19654393e-01 7.84779966e-01
-1.61311731e-01 -7.60160536e-02 -8.62996876e-01 3.27718765e-01
1.06684053e+00 2.78885812e-01 -4.49142694e-01 -3.95301104e-01
-1.59293413e+00 8.25774908e-01 6.68922484e-01 2.84331620e-01
-1.72366306e-01 3.21115330e-02 1.79273412e-01 3.43815893e-01
4.38220978e-01 4.46779542e-02 -2.24888638e-01 3.79773647e-01
-1.09620762e+00 -1.91672400e-01 -9.46825966e-02 1.19396758e+00
7.98874676e-01 1.09668583e-01 -6.82805479e-01 8.86701167e-01
4.67888832e-01 1.23272169e+00 3.51078331e-01 -6.98771536e-01
1.23720205e+00 5.26827991e-01 -6.93384185e-02 -1.02737045e+00
1.05309047e-01 -3.57801914e-01 -7.77543902e-01 -1.36512309e-01
1.41486809e-01 7.38470182e-02 -1.20862961e+00 1.73445714e+00
1.64248303e-01 -2.18010888e-01 2.43170038e-01 1.29387844e+00
1.11894858e+00 1.10641301e+00 -1.02982409e-01 -5.78028500e-01
1.24367511e+00 -1.35600579e+00 -9.04248059e-01 -5.43057263e-01
3.93212587e-01 -9.55641687e-01 1.46425688e+00 -9.46797505e-02
-1.20757759e+00 -5.08586228e-01 -8.79614711e-01 -5.51888585e-01
-1.46769851e-01 2.89956570e-01 2.60452241e-01 1.05065577e-01
-9.59001005e-01 -1.20036744e-01 -5.32181501e-01 -2.36677021e-01
6.25267923e-01 1.02835327e-01 -4.39673781e-01 -6.95512950e-01
-1.09586275e+00 8.67863774e-01 5.61262891e-02 2.07314506e-01
-1.04534221e+00 -4.24152404e-01 -9.08859909e-01 -7.15879649e-02
2.41432175e-01 -8.73511493e-01 9.39601779e-01 -1.11913788e+00
-1.16618502e+00 8.56176674e-01 -7.41706491e-01 4.39672768e-02
4.15605694e-01 2.71013938e-02 4.03782316e-02 6.94978654e-01
4.30763274e-01 1.23120046e+00 1.22404301e+00 -1.56302917e+00
-2.85766572e-01 -4.06312197e-01 -1.43774182e-01 5.33329904e-01
-4.38498020e-01 -1.79586425e-01 -9.47828293e-01 -7.91770697e-01
2.03582034e-01 -6.97915494e-01 -6.00989088e-02 9.98523384e-02
-2.31048375e-01 -3.98941562e-02 9.42951083e-01 -7.23585367e-01
7.10419536e-01 -2.18544388e+00 5.69724679e-01 -2.52184361e-01
3.58266458e-02 -5.59165925e-02 -6.45121336e-01 4.13798898e-01
1.18081875e-01 -7.09901284e-03 -6.74686879e-02 -6.10022604e-01
-2.30178565e-01 3.89250368e-01 -5.73940039e-01 2.35377789e-01
3.37984949e-01 1.60784781e+00 -8.88827801e-01 -1.18024433e+00
3.39232028e-01 5.92333555e-01 -2.28366852e-01 1.90228686e-01
-3.17244440e-01 7.74556816e-01 -5.93398273e-01 7.86202371e-01
8.34904611e-01 -2.84471214e-01 -8.80030096e-02 -4.60858345e-01
2.66747754e-02 -1.60087362e-01 -3.28369528e-01 2.26368070e+00
-5.38415909e-01 7.63833523e-01 1.92735773e-02 -8.63247693e-01
6.48397744e-01 2.57765055e-01 3.71446550e-01 -1.33153772e+00
1.11757278e-01 8.06934834e-02 -4.69449639e-01 -6.38953388e-01
2.43994281e-01 -2.44367331e-01 5.68902567e-02 3.43823075e-01
8.36215913e-02 -6.73392713e-02 -6.54078647e-02 4.22532469e-01
4.77685541e-01 3.12092811e-01 -3.52739453e-01 1.80570349e-01
2.61221319e-01 2.55696893e-01 5.46946585e-01 5.63091040e-01
-1.51724830e-01 9.63108957e-01 2.49846727e-01 -5.63748628e-02
-1.14148498e+00 -1.35734236e+00 2.31068730e-02 9.00682390e-01
5.06632209e-01 -1.52127728e-01 -5.77438593e-01 -7.74744868e-01
-3.81821692e-01 4.77067828e-01 -4.68804419e-01 -1.70111999e-01
-6.34605825e-01 -4.14216340e-01 3.06738280e-02 4.22725469e-01
6.42202437e-01 -9.65069592e-01 -2.17667133e-01 -5.60831279e-02
-8.37318420e-01 -1.30159199e+00 -1.02406311e+00 -1.73925459e-01
-1.08814931e+00 -6.19245768e-01 -1.25290990e+00 -1.18604493e+00
1.01226509e+00 1.05320811e+00 1.01783061e+00 4.08266904e-03
-1.60944700e-01 3.13971370e-01 -2.97586054e-01 -7.64949173e-02
7.15792179e-03 -2.06921294e-01 -2.30263889e-01 1.40117984e-02
1.92800626e-01 -3.57207149e-01 -9.20069456e-01 3.17108840e-01
-9.53706324e-01 6.22742951e-01 1.12708044e+00 1.11417174e+00
8.44404340e-01 -4.28797543e-01 3.32551599e-01 -3.17633241e-01
3.97297800e-01 -2.37939268e-01 -2.23963156e-01 6.15906179e-01
-3.91060323e-01 -1.01828195e-01 2.83838034e-01 -6.48929298e-01
-1.05515158e+00 -2.52088401e-02 3.39485615e-01 -1.09158993e+00
2.82427192e-01 3.74545544e-01 -3.67111892e-01 -4.56621237e-02
2.69844681e-01 7.61683345e-01 -1.04398750e-01 -3.04029733e-01
5.38245320e-01 5.27814746e-01 3.74794006e-01 -4.51812327e-01
8.01311374e-01 4.52525049e-01 -5.26318140e-02 -4.46108133e-01
-7.76724517e-01 -3.50314856e-01 -6.63766623e-01 -2.40461573e-01
1.11097181e+00 -1.20799351e+00 -3.69354248e-01 1.55815864e-02
-1.31314802e+00 -8.02283958e-02 2.17816412e-01 4.78083014e-01
-3.96236509e-01 4.80633318e-01 -7.21948147e-01 -7.43608177e-01
-5.14212489e-01 -1.36520028e+00 1.64990270e+00 1.67737097e-01
4.31888878e-01 -6.71788037e-01 -3.55434000e-01 9.45822775e-01
1.39975563e-01 -1.20104358e-01 9.29636538e-01 1.41228691e-01
-1.15781927e+00 2.05586314e-01 -7.22371459e-01 1.49468765e-01
-8.76575708e-02 -4.72059906e-01 -7.71423936e-01 -6.27669513e-01
3.01876906e-02 -3.88927072e-01 1.22983646e+00 4.27538872e-01
7.23566771e-01 -3.25356513e-01 -3.80365610e-01 5.40011048e-01
1.45322764e+00 2.44322140e-02 6.28629386e-01 1.28043488e-01
1.15825951e+00 5.83579838e-01 8.25398862e-01 -2.77298898e-01
3.56072396e-01 8.07344556e-01 3.94983977e-01 -5.81005394e-01
-4.44867134e-01 -5.57084918e-01 7.44451821e-01 9.09486055e-01
1.83670670e-01 -2.84316033e-01 -4.80796367e-01 7.21383333e-01
-2.10067821e+00 -8.23119104e-01 -7.72098079e-02 1.90773761e+00
8.62744987e-01 -2.15699583e-01 -1.63548276e-01 -4.07732040e-01
6.94704354e-01 1.00358881e-01 -5.47854364e-01 -1.77650079e-01
-4.24877197e-01 -1.68195635e-01 4.32392567e-01 3.45450014e-01
-6.51901424e-01 1.16534698e+00 6.25733757e+00 1.06001639e+00
-1.07414365e+00 5.62894046e-01 4.82370645e-01 -4.34900731e-01
-5.94511151e-01 -2.18668878e-01 -4.42345768e-01 4.59072083e-01
3.66450459e-01 2.76977152e-01 3.00551206e-01 4.32925045e-01
4.25893426e-01 -1.54448047e-01 -8.99017990e-01 1.30410945e+00
3.35629523e-01 -1.31789660e+00 5.51765382e-01 2.71934628e-01
1.03127515e+00 -7.44468579e-03 5.71252525e-01 2.78008670e-01
-2.68792927e-01 -9.84163582e-01 6.45489395e-01 3.99082720e-01
9.65609848e-01 -6.88944697e-01 5.80599666e-01 2.70207763e-01
-1.07891786e+00 1.91143118e-02 -5.08711696e-01 3.15385431e-01
2.02936232e-01 2.58891463e-01 -6.93377316e-01 6.73448741e-01
6.99301243e-01 9.26241517e-01 -5.75205147e-01 5.95262706e-01
-3.11764091e-01 3.42232376e-01 2.17797551e-02 1.18907057e-01
3.58427525e-01 -2.60913879e-01 5.40200949e-01 1.00802386e+00
3.27043772e-01 -6.47832230e-02 4.00483519e-01 1.12478840e+00
-1.40176892e-01 3.15124840e-01 -7.71703005e-01 -6.29622564e-02
2.74563074e-01 1.10792983e+00 -5.14987409e-01 -2.90642798e-01
-7.06802726e-01 1.46536720e+00 3.07924956e-01 8.44685078e-01
-6.77231312e-01 1.05904020e-01 1.61960706e-01 -9.75137949e-02
4.59179461e-01 -3.03076327e-01 -3.79872173e-01 -1.56758666e+00
3.59344721e-01 -7.43686914e-01 1.52105644e-01 -1.38269496e+00
-1.16581786e+00 3.80897850e-01 -4.74576987e-02 -1.53084052e+00
1.36065215e-01 -5.63725412e-01 -4.25349027e-01 9.68824446e-01
-1.48165226e+00 -1.81299770e+00 -4.15453613e-02 9.29227650e-01
9.00976837e-01 -3.69829796e-02 4.36107337e-01 2.26788133e-01
-5.19227684e-01 7.04857767e-01 1.12571828e-01 1.49117306e-01
9.97692406e-01 -9.54999566e-01 -3.37518305e-02 8.65723252e-01
3.73487800e-01 6.38503850e-01 4.39669967e-01 -7.58427560e-01
-1.76499116e+00 -1.11951041e+00 7.08394408e-01 -6.37365639e-01
2.71434069e-01 -4.79462087e-01 -7.87990212e-01 5.13890743e-01
6.85841680e-01 -2.36998364e-01 4.06757236e-01 -2.63357133e-01
-4.84410673e-01 -1.00916162e-01 -8.72569799e-01 7.90499687e-01
8.76328528e-01 -7.12718725e-01 -7.89484203e-01 5.53308666e-01
9.80734110e-01 -3.93087626e-01 -6.21431530e-01 1.69572353e-01
3.79342288e-01 -5.26303053e-01 1.20020688e+00 -2.98131526e-01
8.87705982e-01 -5.71284354e-01 -2.68393785e-01 -1.25318789e+00
-3.06691397e-02 -3.34463030e-01 2.42082751e-03 1.14991200e+00
5.84629297e-01 -1.79821357e-01 4.50185955e-01 1.78096518e-01
-1.72853932e-01 -7.42655218e-01 -9.90593374e-01 -5.09394109e-01
1.72471985e-01 -1.50701348e-02 2.41821036e-01 8.28323543e-01
-2.93507464e-02 8.81283283e-01 -8.53348255e-01 3.66522297e-02
7.92987525e-01 6.11121297e-01 6.17796183e-01 -4.63940352e-01
-2.20769987e-01 -3.06146860e-01 -2.40410548e-02 -1.34233332e+00
1.37148928e-02 -1.03708518e+00 1.38075590e-01 -1.93495011e+00
8.31718922e-01 2.37749457e-01 -3.17475080e-01 6.08125627e-01
-2.59486049e-01 6.97751820e-01 3.61721575e-01 5.10096073e-01
-7.03148365e-01 8.49285841e-01 2.32355499e+00 -6.32790685e-01
2.68410090e-02 -5.80787957e-01 -5.57646692e-01 6.98800161e-02
3.75816941e-01 -4.38345343e-01 -5.70418537e-01 -9.79467869e-01
1.59564570e-01 3.49732161e-01 5.21953762e-01 -3.46992105e-01
1.34810731e-01 -4.34710562e-01 9.88844812e-01 -1.00006819e+00
4.69853073e-01 -7.52269387e-01 -2.78180748e-01 2.47519419e-01
-2.96151102e-01 1.11213349e-01 9.74979401e-02 8.91590834e-01
-3.24387908e-01 4.94098887e-02 5.89512229e-01 -1.99842721e-01
-5.52298605e-01 4.34663028e-01 -1.71694309e-01 -1.46314159e-01
7.06398606e-01 -2.21631527e-01 -5.03462732e-01 -3.24326694e-01
-4.23680097e-01 6.07038915e-01 6.86949253e-01 7.25869060e-01
1.06990039e+00 -1.71578479e+00 -8.09688687e-01 1.07788360e-02
3.65534931e-01 5.37944138e-02 5.12885988e-01 1.26363742e+00
-9.59158689e-02 5.17443240e-01 -3.26701909e-01 -9.34547007e-01
-1.34892964e+00 1.00606239e+00 -1.25206299e-02 5.86234666e-02
-8.28527749e-01 6.88867867e-01 7.47032106e-01 -3.33176143e-02
6.10037372e-02 -3.82441736e-04 -1.12832479e-01 -1.06870748e-01
5.00864983e-01 -3.85129809e-01 -2.75041938e-01 -8.15333366e-01
-3.09995502e-01 9.25476551e-01 -3.02535176e-01 -5.47461808e-01
9.76875305e-01 -6.55759394e-01 -9.87293720e-02 3.48124444e-01
1.45764077e+00 -1.47380263e-01 -1.24317372e+00 -5.59566021e-01
-6.49155140e-01 -7.85519183e-01 2.45139852e-01 -7.48552680e-01
-1.25302196e+00 1.41124582e+00 6.76298499e-01 -6.10813498e-01
1.29068792e+00 1.91100001e-01 1.13307202e+00 2.79677302e-01
2.99183279e-01 -9.60824907e-01 5.61043680e-01 2.27932528e-01
1.09211099e+00 -1.61957932e+00 -7.60648176e-02 -2.79832214e-01
-8.25593233e-01 8.89987290e-01 7.98106194e-01 1.73969224e-01
-1.45395786e-01 -3.92146438e-01 3.11768681e-01 -6.82384074e-02
-8.36888433e-01 -2.29546979e-01 6.89134955e-01 6.82902336e-01
2.35319927e-01 -6.81013465e-02 -3.23477387e-01 3.95262718e-01
3.81902248e-01 -4.23466146e-01 6.90910667e-02 8.48701417e-01
-4.24979478e-01 -1.07011712e+00 -5.45996845e-01 7.00943470e-02
-2.25710228e-01 -5.14903665e-01 -6.55701458e-01 4.92557138e-01
3.21161300e-02 1.06964695e+00 -1.92186713e-01 -3.13835919e-01
1.31145373e-01 -1.46455780e-01 8.62107515e-01 -5.65072298e-01
-2.02987149e-01 8.58715534e-01 -3.39623272e-01 -5.26274741e-01
-5.78496099e-01 -5.61950207e-01 -1.13375866e+00 4.99014296e-02
-2.35381618e-01 5.97470887e-02 5.90943873e-01 9.83780980e-01
3.45341444e-01 3.86762887e-01 6.68525040e-01 -9.38212991e-01
7.21723959e-03 -9.67218578e-01 -2.36370504e-01 4.36063319e-01
5.73301733e-01 -7.15988934e-01 -2.07592010e-01 1.93935737e-01] | [11.455172538757324, 1.4782785177230835] |
f8900f90-f14e-44b3-91cd-45be3e481cd7 | usage-of-multiple-rtl-features-for-earthquake | 1905.10805 | null | https://arxiv.org/abs/1905.10805v1 | https://arxiv.org/pdf/1905.10805v1.pdf | Usage of multiple RTL features for Earthquake prediction | We construct a classification model that predicts if an earthquake with the magnitude above a threshold will take place at a given location in a time range 30-180 days from a given moment of time. A common approach is to use expert forecasts based on features like Region-Time-Length (RTL) characteristics. The proposed approach uses machine learning on top of multiple RTL features to take into account effects at various scales and to improve prediction accuracy. For historical data about Japan earthquakes 1992-2005 and predictions at locations given in this database the best model has precision up to ~ 0.95 and recall up to ~ 0.98. | ['I. Braslavsky', 'E. Egorov', 'P. Proskura', 'E. Burnaev', 'A. Zaytsev'] | 2019-05-26 | null | null | null | null | ['earthquake-prediction'] | ['computer-vision'] | [-4.51909840e-01 -2.39774659e-01 -9.95876268e-02 -5.40270686e-01
-7.90421009e-01 -5.55976987e-01 5.65124989e-01 9.93916273e-01
-3.96780252e-01 9.34562206e-01 3.22515845e-01 -6.26757562e-01
-3.23354185e-01 -1.22464919e+00 -4.37764049e-01 -5.55962205e-01
-6.51704013e-01 2.58670509e-01 4.06877309e-01 -3.40613961e-01
6.85482144e-01 9.41339493e-01 -1.09294200e+00 3.83251131e-01
6.58529639e-01 1.08311796e+00 1.86233252e-01 6.31016493e-01
1.75427541e-01 7.73625195e-01 -8.71711910e-01 1.88214600e-01
2.48204023e-01 1.70218885e-01 -5.56137681e-01 -5.75616837e-01
-4.51601624e-01 -2.18084008e-01 -1.89571396e-01 3.78746569e-01
1.90906450e-01 1.36508167e-01 1.10762596e+00 -9.01455045e-01
-1.81491360e-01 6.06370151e-01 -5.43041408e-01 8.98728848e-01
4.40806806e-01 -2.75287569e-01 7.34701991e-01 -7.43018508e-01
2.59918153e-01 7.33130932e-01 9.79861081e-01 -5.23514628e-01
-9.83124852e-01 -4.85441506e-01 -1.44486755e-01 2.19387025e-01
-1.58608603e+00 2.59370744e-01 6.39072418e-01 -8.49956214e-01
1.37269533e+00 4.34256941e-02 5.48728287e-01 2.96964169e-01
1.08821356e+00 4.89277542e-02 9.73653734e-01 -2.97522962e-01
2.36540869e-01 -2.12681487e-01 -9.19917077e-02 -4.27244380e-02
2.34770104e-01 -1.46244811e-02 -1.88008606e-01 -4.43404615e-01
4.87261623e-01 2.51946330e-01 1.63206846e-01 7.23295867e-01
-1.11198735e+00 1.19275153e+00 5.65257907e-01 6.56641603e-01
-8.08402002e-01 -1.76755726e-01 6.16943836e-01 3.43089938e-01
8.79373848e-01 6.98608696e-01 -1.00799823e+00 3.59079316e-02
-1.22966170e+00 3.65839899e-01 4.73648846e-01 2.10403770e-01
9.65685248e-01 1.55505091e-01 1.66536480e-01 5.10244310e-01
-1.61132053e-01 6.88749433e-01 4.23124701e-01 -3.81904423e-01
6.78852618e-01 4.47775662e-01 3.66526455e-01 -1.27286792e+00
-9.59197044e-01 -4.83675420e-01 -8.05248439e-01 5.87291643e-02
3.06261778e-01 -5.50420284e-01 -7.09761858e-01 1.35906041e+00
4.21083383e-02 3.71637493e-02 2.45370954e-01 4.84472752e-01
2.38820523e-01 1.54688263e+00 1.75204612e-02 -3.23633194e-01
1.04431772e+00 7.00180978e-02 -3.18537593e-01 -1.46865726e-01
1.27741253e+00 -6.53664529e-01 3.02594811e-01 3.97353947e-01
-4.13322806e-01 -4.58796710e-01 -8.03382039e-01 7.48284876e-01
-7.71466851e-01 1.71470940e-01 3.67716908e-01 1.00290753e-01
-7.74323285e-01 7.87725925e-01 -6.54540241e-01 -2.55108982e-01
-2.72432923e-01 1.48194224e-01 -3.06442410e-01 3.50391030e-01
-1.61787128e+00 1.21742952e+00 4.95321333e-01 1.13977544e-01
-4.54636902e-01 -5.32489300e-01 -6.44995689e-01 1.35144107e-02
-1.77941859e-01 2.26645451e-02 8.69505882e-01 -5.30098319e-01
-7.49110162e-01 2.39920542e-01 9.42855179e-02 -7.46617317e-01
1.78809166e-01 -2.10752785e-01 -7.86421061e-01 5.06304670e-04
4.10813779e-01 -2.98652090e-02 7.63961375e-02 -7.78797388e-01
-8.98230076e-01 -3.63613307e-01 -5.39318398e-02 1.33903742e-01
-1.12447359e-01 2.61974752e-01 3.13775897e-01 -9.11323547e-01
3.41972768e-01 -8.31980944e-01 -6.34779572e-01 -1.14937401e+00
-2.39082456e-01 -3.44613433e-01 5.91543257e-01 -1.25762796e+00
1.63584685e+00 -1.58388758e+00 -3.78704906e-01 4.96980280e-01
-3.91366601e-01 -7.69187603e-03 4.25138295e-01 1.13026726e+00
-2.66026527e-01 1.15089551e-01 -4.24444564e-02 5.36543250e-01
-6.99405849e-01 9.61805880e-02 -7.05378234e-01 4.93619531e-01
6.93886429e-02 1.32715940e-01 -6.80697978e-01 -9.67432782e-02
2.93271214e-01 5.79739511e-02 -1.76547736e-01 2.92571425e-01
2.37318963e-01 4.07191783e-01 -6.35848284e-01 5.56932092e-01
6.40061855e-01 1.03260651e-01 -9.12815556e-02 2.48728365e-01
-6.45914793e-01 2.08896935e-01 -9.89637613e-01 7.16493964e-01
-6.41824782e-01 7.18129754e-01 -8.16778004e-01 -1.15223062e+00
1.53204215e+00 8.16898406e-01 7.55336285e-01 -4.01038587e-01
-3.50437462e-01 4.18668985e-01 -3.21109220e-03 -4.52450693e-01
4.17010576e-01 9.12860259e-02 -5.00753999e-01 3.47324371e-01
-4.31166649e-01 -7.70852529e-03 6.90804794e-02 -1.83751822e-01
1.17336011e+00 -5.61317921e-01 7.88750410e-01 -5.47882736e-01
2.90985644e-01 6.76899329e-02 7.31639028e-01 6.45839393e-01
2.32186273e-01 4.09464747e-01 6.14767313e-01 -1.00603962e+00
-1.19212711e+00 -5.14236093e-01 -3.33599448e-01 9.32608843e-01
-4.66016978e-01 5.81266731e-02 -2.22375706e-01 -2.86430329e-01
1.28364682e-01 8.47321689e-01 -7.71699727e-01 1.24587089e-01
-6.28296435e-01 -1.05910349e+00 4.02068317e-01 6.14029348e-01
2.17495665e-01 -1.03014398e+00 -8.44922066e-01 4.31735486e-01
-1.61310971e-01 -6.80409133e-01 6.81726933e-02 5.88539541e-01
-1.09471607e+00 -5.01476228e-01 -8.07224631e-01 -5.90363622e-01
4.81854320e-01 -2.61366993e-01 9.64189351e-01 -1.15340069e-01
-9.56505165e-03 -4.53845650e-01 -5.13532817e-01 -4.96432126e-01
-2.23265260e-01 3.76774430e-01 9.54830647e-02 -1.00791737e-01
-4.71281409e-02 -6.57087386e-01 -6.47929430e-01 3.02260995e-01
-6.32297575e-01 -1.98608682e-01 5.52626014e-01 5.44840693e-01
1.47703409e-01 3.51041168e-01 1.28068328e+00 -4.47354645e-01
3.42710823e-01 -1.17434430e+00 -6.60073876e-01 2.29034528e-01
-5.73350072e-01 -1.11932255e-01 8.36675942e-01 -3.89249414e-01
-8.49042535e-01 1.87495038e-01 -8.86143222e-02 2.13057175e-01
-1.65524796e-01 1.21576715e+00 3.53975654e-01 3.64268869e-01
7.00660408e-01 2.81437218e-01 -7.09449887e-01 -5.01550198e-01
-2.03329980e-01 6.45776391e-01 3.72240841e-01 -2.71007955e-01
4.72196907e-01 1.75614655e-01 1.50283128e-01 -6.46670282e-01
-7.25265443e-01 -6.17531061e-01 -8.23854208e-01 -4.65537369e-01
5.03107429e-01 -1.07052803e+00 -2.36432880e-01 4.77275848e-01
-1.16604900e+00 -1.16224423e-01 3.76638740e-01 9.19404268e-01
-2.54113585e-01 -1.75609335e-01 -5.68312526e-01 -1.04181147e+00
-3.16412866e-01 -6.63848996e-01 4.89530265e-01 2.99434692e-01
-2.85022795e-01 -1.28159618e+00 4.61816072e-01 -4.06798124e-01
4.51919496e-01 6.45995319e-01 7.20041752e-01 -1.04780865e+00
5.47069833e-02 -8.15039456e-01 2.39703283e-02 -1.06099710e-01
2.79145420e-01 3.21589828e-01 -6.45502090e-01 -2.04732865e-01
-1.37116298e-01 1.22123718e-01 6.63885117e-01 8.34011614e-01
8.89095128e-01 -4.43607956e-01 -4.65057194e-01 2.85596311e-01
1.64084160e+00 9.26600695e-01 5.38757145e-01 6.91888928e-01
5.97484261e-02 6.00793004e-01 1.02331531e+00 9.86712039e-01
3.05747271e-01 6.82719111e-01 1.55462772e-01 -7.36515671e-02
8.07749152e-01 6.31315410e-02 8.71981457e-02 6.48205400e-01
-2.29393661e-01 -2.08564848e-01 -1.66103613e+00 9.54778612e-01
-1.67347014e+00 -1.14152479e+00 -2.97768861e-01 2.38381791e+00
5.58263123e-01 2.83467233e-01 4.08994704e-02 3.95152628e-01
7.36810803e-01 4.17910278e-01 -1.16629966e-01 -7.29310930e-01
-1.66864961e-01 -3.35760295e-01 1.05363107e+00 4.91239429e-01
-1.14660621e+00 4.38570470e-01 7.00032806e+00 5.42876899e-01
-1.44193745e+00 -2.17085600e-01 1.04631090e+00 5.70226908e-01
4.22435105e-02 2.37148374e-01 -8.34027886e-01 5.20151019e-01
1.39391685e+00 -4.14509565e-01 -1.34028420e-02 8.17276120e-01
5.90930700e-01 -5.35885572e-01 -4.79015380e-01 3.50896269e-01
-2.34364271e-01 -1.48325491e+00 -1.77639887e-01 1.49927989e-01
8.95587087e-01 1.25864983e-01 -1.80755422e-01 1.89769432e-01
8.74749199e-02 -8.07329595e-01 6.22133851e-01 8.78843784e-01
4.60271120e-01 -1.14305043e+00 1.21964037e+00 6.30497813e-01
-1.40618634e+00 -3.74412447e-01 -4.46524203e-01 -5.06051540e-01
2.51533985e-01 1.16718698e+00 -1.41241205e+00 6.13737166e-01
7.51079857e-01 6.65088892e-01 -2.33992070e-01 1.21954489e+00
9.28379744e-02 9.74204540e-01 -6.03946567e-01 1.50537202e-02
3.44363332e-01 1.56703293e-01 2.23021731e-01 1.33100700e+00
1.03491712e+00 3.78731191e-01 1.86868429e-01 -4.58441637e-02
5.79706848e-01 3.05841863e-01 -9.92362916e-01 7.00572968e-01
7.22583294e-01 8.31836581e-01 -8.03198040e-01 -3.95340502e-01
-2.57808894e-01 3.45145762e-01 6.22230098e-02 2.07509369e-01
-8.04770708e-01 -7.24047601e-01 1.00755617e-01 3.39078039e-01
2.25787327e-01 -3.04564416e-01 -6.48037732e-01 -6.47476196e-01
-2.20539510e-01 -5.82959354e-02 5.90446770e-01 -8.65329146e-01
-1.04429388e+00 7.73134887e-01 2.41925642e-01 -1.53247464e+00
-5.71173310e-01 -3.40158105e-01 -1.18791938e+00 1.01313090e+00
-1.14084578e+00 -6.78164840e-01 3.39924514e-01 1.84190482e-01
4.94253695e-01 -2.06359386e-01 8.33672941e-01 8.38405117e-02
-2.84514904e-01 3.76461335e-02 6.08935058e-01 2.82081515e-01
3.93581331e-01 -1.16898859e+00 4.09454495e-01 6.22342706e-01
-2.94398725e-01 2.03296691e-01 9.10411000e-01 -1.02992785e+00
-4.21411395e-01 -1.10440660e+00 1.55944061e+00 -2.14863509e-01
8.23455215e-01 2.09245995e-01 -1.00164700e+00 5.96487164e-01
-8.26524943e-02 -3.31294537e-03 5.24288535e-01 1.76820636e-01
-1.02797508e-01 -4.18447137e-01 -1.12489760e+00 6.32916093e-02
-1.12275556e-01 -3.84078979e-01 -7.00560451e-01 4.84925836e-01
2.28851989e-01 -1.79624230e-01 -1.48368883e+00 6.17686152e-01
4.31979775e-01 -7.19853997e-01 7.33426273e-01 -2.66265810e-01
1.75231516e-01 -1.80150241e-01 -2.36440808e-01 -1.73428750e+00
-4.93881702e-01 -2.98740327e-01 5.59795201e-01 9.59595859e-01
8.18409622e-01 -8.25839877e-01 3.32051516e-01 2.62106806e-01
-4.24852222e-02 -8.52065265e-01 -1.26655626e+00 -7.87312448e-01
2.75042653e-01 -4.24349397e-01 8.55805635e-01 8.78727794e-01
3.30870867e-01 -1.60768002e-01 -4.38887179e-01 7.47796476e-01
2.28336290e-01 2.13387981e-01 3.15370739e-01 -1.31752753e+00
3.71843845e-01 1.27718691e-02 -4.27104503e-01 -4.26852077e-01
-2.11839527e-01 -4.10333574e-01 -1.52392626e-01 -1.46741307e+00
2.41820421e-02 -6.80638433e-01 -5.57716370e-01 4.84181166e-01
1.44417444e-02 9.23527405e-04 -1.98312968e-01 4.24576551e-01
9.59821641e-02 9.49777961e-02 5.49724102e-01 3.36127132e-01
-1.80396363e-01 5.00453949e-01 8.75640810e-02 8.08052242e-01
1.27605438e+00 -7.63437450e-01 -1.10074520e-01 -2.11923480e-01
4.57482159e-01 9.88693953e-01 -1.40590787e-01 -1.20427334e+00
-5.25526851e-02 -7.33766258e-01 5.26272893e-01 -1.17059791e+00
2.35202033e-02 -6.66623533e-01 5.09011090e-01 6.53090060e-01
-3.82110298e-01 7.72388458e-01 1.81606412e-01 4.27886546e-01
-4.49730277e-01 -2.37215519e-01 4.35810119e-01 1.26902759e-01
-5.65096498e-01 1.66373961e-02 -1.02533126e+00 -5.27882040e-01
1.17101955e+00 1.13174647e-01 -3.53396051e-02 -6.07253730e-01
-7.48552680e-01 5.85792586e-03 1.66392803e-01 1.76352799e-01
3.60460699e-01 -1.19295394e+00 -9.37469363e-01 -1.53627828e-01
1.97650254e-01 -6.32658958e-01 2.06733853e-01 6.55536175e-01
-8.72337818e-01 7.97203839e-01 -2.87974507e-01 -3.63429785e-01
-7.16321707e-01 3.41229707e-01 2.87669480e-01 -7.14004159e-01
-3.89382601e-01 3.88281196e-01 -1.02303386e-01 -3.08216929e-01
-3.92549932e-01 -6.40184402e-01 -6.70608819e-01 2.48433173e-01
5.63177943e-01 4.70536292e-01 2.33855307e-01 -8.14744473e-01
-3.60073417e-01 6.44548893e-01 3.69817317e-01 -3.62169802e-01
1.43747616e+00 1.00959297e-02 1.39828682e-01 9.34697628e-01
1.10293806e+00 5.90143725e-04 -1.12126517e+00 -1.72537640e-02
4.53494906e-01 -3.90370876e-01 5.72584271e-02 -8.29383135e-01
-8.89924765e-01 6.40357435e-01 3.27902496e-01 4.38929141e-01
1.01825416e+00 -1.49211586e-01 6.24621153e-01 4.15890872e-01
4.45078731e-01 -1.19433689e+00 -4.02704775e-01 8.01530182e-01
1.09085441e+00 -1.03150117e+00 1.60428822e-01 9.92148221e-02
-6.96210504e-01 1.36925459e+00 8.98912549e-04 -3.95948827e-01
1.18265700e+00 3.06266010e-01 -5.73680364e-02 1.68826729e-01
-9.05826449e-01 1.70103595e-01 1.84091687e-01 3.88208598e-01
5.06553292e-01 5.52724838e-01 -5.99425197e-01 4.76165950e-01
-1.52180314e-01 -2.36194789e-01 5.93234479e-01 9.06542778e-01
-9.59104657e-01 -5.34238100e-01 -8.13556075e-01 7.40204751e-01
-7.52341688e-01 -4.01658192e-02 3.11399519e-01 5.46208799e-01
8.55696350e-02 1.13675642e+00 4.50866878e-01 -4.90624189e-01
-6.48063347e-02 -1.08723812e-01 -4.41338360e-01 -3.88060451e-01
-6.11960828e-01 4.77056056e-02 1.33305103e-01 -2.31648535e-01
-8.53376687e-02 -1.15602779e+00 -1.54680240e+00 -2.90217161e-01
-3.87317777e-01 4.65455085e-01 6.71366572e-01 1.09342861e+00
1.55725973e-02 -5.59621211e-03 1.49507427e+00 -9.36857045e-01
-2.39848703e-01 -1.31520486e+00 -8.76034617e-01 1.35101661e-01
3.61485988e-01 -4.86405760e-01 -6.75162971e-01 -2.69761868e-03] | [6.513430595397949, 3.039111614227295] |
9b720c7a-2704-4811-893d-39b8306b77ae | self-mutual-distillation-learning-for | null | null | http://openaccess.thecvf.com//content/ICCV2021/html/Hao_Self-Mutual_Distillation_Learning_for_Continuous_Sign_Language_Recognition_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Hao_Self-Mutual_Distillation_Learning_for_Continuous_Sign_Language_Recognition_ICCV_2021_paper.pdf | Self-Mutual Distillation Learning for Continuous Sign Language Recognition | In recent years, deep learning moves video-based Continuous Sign Language Recognition (CSLR) significantly forward. Currently, a typical network combination for CSLR includes a visual module, which focuses on spatial and short-temporal information, followed by a contextual module, which focuses on long-temporal information, and the Connectionist Temporal Classification (CTC) loss is adopted to train the network. However, due to the limitation of chain rules in back-propagation, the visual module is hard to adjust for seeking optimized visual features. As a result, it enforces that the contextual module focuses on contextual information optimization only rather than balancing efficient visual and contextual information. In this paper, we propose a Self-Mutual Knowledge Distillation (SMKD) method, which enforces the visual and contextual modules to focus on short-term and long-term information and enhances the discriminative power of both modules simultaneously. Specifically, the visual and contextual modules share the weights of their corresponding classifiers, and train with CTC loss simultaneously. Moreover, the spike phenomenon widely exists with CTC loss. Although it can help us choose a few of the key frames of a gloss, it does drop other frames in a gloss and makes the visual feature saturation in the early stage. A gloss segmentation is developed to relieve the spike phenomenon and decrease saturation in the visual module. We conduct experiments on two CSLR benchmarks: PHOENIX14 and PHOENIX14-T. Experimental results demonstrate the effectiveness of the SMKD. | ['Xilin Chen', 'Yuecong Min', 'Aiming Hao'] | 2021-01-01 | null | null | null | iccv-2021-1 | ['sign-language-recognition'] | ['computer-vision'] | [-1.30101219e-01 -4.10448998e-01 -2.15697393e-01 -4.03096288e-01
-3.96432847e-01 -3.90283205e-02 2.28023812e-01 -3.30330193e-01
-8.13578188e-01 4.20550823e-01 1.28369421e-01 3.79809774e-02
-1.66707009e-01 -4.50558752e-01 -5.28468907e-01 -1.06113851e+00
2.58702755e-01 -7.50626922e-02 5.73720753e-01 1.37912491e-02
8.89817160e-03 2.04223379e-01 -1.25536287e+00 3.50752443e-01
1.02235889e+00 1.07381034e+00 6.55655444e-01 1.84887186e-01
-3.00880581e-01 1.02253711e+00 -3.88771623e-01 -1.22032702e-01
7.64955133e-02 -4.04354572e-01 -2.86239475e-01 -4.18037623e-02
2.35055968e-01 -4.09885824e-01 -9.32687640e-01 1.11593235e+00
7.31442571e-01 2.62392223e-01 2.81947196e-01 -1.27648807e+00
-6.74948514e-01 3.65914583e-01 -6.91669822e-01 1.10913016e-01
-5.42137362e-02 7.21943080e-01 1.03525662e+00 -7.99217999e-01
4.84263062e-01 1.16905713e+00 4.27232206e-01 5.54820001e-01
-8.01820815e-01 -7.81987786e-01 7.84563780e-01 5.52277863e-01
-1.44387066e+00 -2.19765782e-01 9.41937566e-01 -2.28471309e-01
6.62591338e-01 6.03327192e-02 1.13379180e+00 8.72542441e-01
-4.76137884e-02 1.52383256e+00 9.00470912e-01 -1.72916442e-01
-2.66293958e-02 -9.62006301e-02 2.56389171e-01 7.44405389e-01
-1.63532034e-01 2.95364648e-01 -5.76694965e-01 3.25161368e-01
7.71561205e-01 4.52420294e-01 -5.18040121e-01 -2.29688019e-01
-9.67691541e-01 5.65814972e-01 8.66466224e-01 3.86530668e-01
-2.33515233e-01 2.47396126e-01 3.40656549e-01 3.15981656e-01
-3.98943089e-02 -2.17052847e-01 -2.70508468e-01 -8.32712278e-02
-8.61936510e-01 -1.18506029e-01 2.10752726e-01 8.16239655e-01
7.76992798e-01 -1.05390819e-02 -5.22023141e-01 1.06447446e+00
5.99807024e-01 4.27798182e-01 5.27790546e-01 -6.61199450e-01
5.84233344e-01 7.38739610e-01 -3.53366524e-01 -7.93257058e-01
-2.53934443e-01 -4.63156998e-01 -8.15804422e-01 2.41823956e-01
2.82071948e-01 6.16883906e-03 -1.34803581e+00 1.82334828e+00
-5.91028072e-02 3.43542993e-01 -2.83288777e-01 1.37438846e+00
8.39728236e-01 6.42881930e-01 1.50610879e-01 -3.14394504e-01
8.87243271e-01 -1.08049655e+00 -7.17335641e-01 -2.47106925e-01
3.09712261e-01 -5.15222013e-01 1.25414515e+00 2.13116884e-01
-9.57234442e-01 -4.78746861e-01 -9.23956215e-01 -1.71565413e-01
-5.83785661e-02 3.00581247e-01 5.09152234e-01 -1.47714198e-01
-6.93460226e-01 2.87196845e-01 -9.25646484e-01 -6.42732084e-02
5.27711570e-01 1.80576429e-01 -1.09305963e-01 -3.04062903e-01
-1.19064152e+00 5.90719640e-01 2.79740572e-01 5.22346675e-01
-5.68329453e-01 -3.23845059e-01 -7.16091871e-01 -1.05062351e-01
2.10809708e-01 -5.08357823e-01 8.53657484e-01 -1.17235732e+00
-1.56536102e+00 6.52965903e-01 -1.71612859e-01 -6.47627786e-02
8.22791159e-01 -2.09507585e-01 -4.87159073e-01 8.47026855e-02
-8.71044174e-02 8.01913261e-01 8.66408169e-01 -1.12742877e+00
-9.01957452e-01 -2.56744772e-01 1.74359590e-01 4.22217250e-01
-3.79570991e-01 -1.59172058e-01 -1.16449738e+00 -7.55823195e-01
3.36051166e-01 -8.12506497e-01 -1.29441515e-01 2.69186348e-01
-2.41522431e-01 -2.88846791e-01 8.19628835e-01 -6.53014600e-01
1.36593938e+00 -2.34789348e+00 2.76603371e-01 3.26542586e-01
7.54993856e-02 4.14877325e-01 -2.23953709e-01 1.12832487e-02
5.38751930e-02 -1.72870457e-01 -3.99106622e-01 -3.48202735e-01
-2.39524528e-01 4.80784357e-01 -1.51969284e-01 4.53092545e-01
1.31080940e-01 1.00193477e+00 -9.56001043e-01 -6.40856624e-01
4.09050614e-01 6.08566165e-01 -5.86724102e-01 8.31364770e-04
-9.57393497e-02 4.67876673e-01 -5.29739499e-01 6.59963310e-01
7.37963676e-01 -2.11687863e-01 -5.31115867e-02 -3.97243172e-01
-2.64131725e-01 8.21130350e-02 -9.73006368e-01 1.89679205e+00
-3.64495873e-01 6.20947897e-01 -6.76259771e-02 -9.12206173e-01
6.53275669e-01 -6.20113201e-02 4.85092610e-01 -1.15177810e+00
9.80646759e-02 2.02921510e-01 -7.19605312e-02 -6.85468197e-01
-2.42971070e-02 1.21513838e-02 2.31044129e-01 3.92128527e-02
-1.44941673e-01 3.02294165e-01 4.76936512e-02 9.45472047e-02
9.85439301e-01 2.53965795e-01 -2.78645188e-01 2.31107160e-01
6.12732708e-01 -3.77539903e-01 1.02980053e+00 4.74803329e-01
-4.45836633e-01 7.13972032e-01 1.70639902e-01 -1.74199015e-01
-5.77783227e-01 -1.05088973e+00 1.11934289e-01 8.98768663e-01
8.06332350e-01 -2.24503607e-01 -1.94070235e-01 -9.68179822e-01
-2.07540039e-02 4.88885164e-01 -3.65338624e-01 -4.72740531e-01
-8.17472100e-01 -4.78420705e-01 3.16804051e-01 5.69548845e-01
8.33243489e-01 -1.33222413e+00 -3.78726214e-01 1.80694684e-01
-1.98205397e-01 -8.92758250e-01 -8.77251565e-01 -5.01583749e-03
-6.41342700e-01 -9.18033659e-01 -8.76692712e-01 -1.17595863e+00
6.42138064e-01 4.05659676e-01 4.30420935e-01 3.41954082e-01
-1.13408826e-01 1.16475403e-01 -4.13048565e-01 -5.28023578e-02
2.85664588e-01 -2.64210552e-01 -2.94772625e-01 2.58261859e-01
4.03687000e-01 -6.47084296e-01 -8.49623024e-01 3.02424371e-01
-8.16699982e-01 1.43783540e-01 8.05899084e-01 1.10366714e+00
6.23326182e-01 -3.08392216e-02 1.65876821e-01 -2.47917488e-01
2.78712630e-01 -1.33424550e-01 -4.66756225e-01 5.18943965e-01
-5.37422836e-01 5.91267124e-02 3.94497424e-01 -7.31186569e-01
-9.37985718e-01 6.56198934e-02 -2.68618226e-01 -8.42438042e-01
2.36608565e-01 5.47210276e-01 -3.95258754e-01 -1.52322084e-01
-1.04177207e-01 5.93745470e-01 -5.30255698e-02 -4.39492702e-01
1.61367789e-01 5.95786452e-01 4.10087764e-01 -1.90764666e-01
6.55243754e-01 5.41939318e-01 -4.89876509e-01 -6.56713784e-01
-7.72182047e-01 -5.06618500e-01 -2.88711280e-01 -4.53266025e-01
8.81139755e-01 -9.26671147e-01 -7.01325238e-01 8.76180053e-01
-9.89288867e-01 -4.16382879e-01 -3.56952548e-01 8.44124734e-01
-2.56954640e-01 4.69942629e-01 -7.04196990e-01 -6.80560350e-01
-1.69985041e-01 -1.21581984e+00 1.02677989e+00 5.88002801e-01
3.40598434e-01 -6.61264837e-01 -7.23891258e-02 1.88727498e-01
1.12932153e-01 -2.59686202e-01 6.53088808e-01 5.22268079e-02
-8.07518482e-01 -6.23838278e-03 -5.51795304e-01 6.00239217e-01
3.58547294e-03 7.88771585e-02 -9.25470114e-01 -3.30138505e-01
-2.22173288e-01 -3.14218402e-01 1.55473363e+00 5.56031704e-01
1.19087481e+00 -2.42108107e-01 -2.46832550e-01 8.52599204e-01
1.29065704e+00 4.04911637e-01 8.88395607e-01 2.64905155e-01
9.06119525e-01 4.20205146e-01 6.58698678e-01 7.87472725e-02
6.32328391e-01 7.65922606e-01 2.82004237e-01 -1.58193648e-01
-3.91198486e-01 -4.34297860e-01 5.77880025e-01 1.07935476e+00
-5.29996678e-02 -2.14072578e-02 -5.90041876e-01 3.74306738e-01
-2.13198590e+00 -9.05622780e-01 2.73348987e-01 2.09213853e+00
1.05683076e+00 2.43129760e-01 -1.66881561e-01 9.02322382e-02
7.51172245e-01 3.50890458e-01 -7.49357045e-01 1.49161756e-01
-3.87512147e-01 -1.53157160e-01 4.01297778e-01 4.77839112e-01
-9.82737780e-01 9.73197997e-01 4.62491417e+00 1.17793429e+00
-1.52705371e+00 1.01494551e-01 2.89271533e-01 -3.30083311e-01
-2.63961852e-01 1.43347397e-01 -7.27691829e-01 8.53286147e-01
4.75034155e-02 3.04030806e-01 4.41531390e-01 7.44337142e-01
2.54889697e-01 -3.22114304e-02 -7.77484298e-01 1.33534026e+00
-9.64840725e-02 -1.04488885e+00 7.50079975e-02 -6.99978694e-02
4.91479754e-01 2.80160636e-01 -4.31055669e-03 5.21819055e-01
9.12186280e-02 -7.22302318e-01 8.16430211e-01 9.35771942e-01
7.87485659e-01 -8.54381502e-01 5.48855543e-01 4.66801882e-01
-1.47998595e+00 -2.85203159e-01 -2.84444720e-01 2.80343235e-01
1.80189297e-01 4.89988178e-01 5.06527871e-02 2.43502215e-01
8.16704452e-01 1.07763505e+00 -2.72169024e-01 1.32009566e+00
-5.94602942e-01 4.53830093e-01 -3.67556781e-01 -1.53529763e-04
4.62102711e-01 -3.80841464e-01 6.92376196e-01 1.10769308e+00
1.96324121e-02 2.03077629e-01 3.74991834e-01 6.55444026e-01
2.28266995e-02 -5.99485487e-02 -2.45822415e-01 2.18857065e-01
3.08832705e-01 9.37189400e-01 -3.50103915e-01 -2.27486119e-01
-4.72842127e-01 1.20894933e+00 3.98194939e-01 8.02003920e-01
-7.87000835e-01 -5.42709053e-01 5.58554471e-01 -1.23827055e-01
3.97349566e-01 -2.10267693e-01 -2.42532685e-01 -1.44457603e+00
4.23660070e-01 -6.32103145e-01 4.97452348e-01 -6.62043273e-01
-1.23565125e+00 3.23167533e-01 -4.76553947e-01 -1.48906648e+00
2.57608116e-01 -3.18444103e-01 -7.61491477e-01 7.79898703e-01
-1.90245724e+00 -1.21106958e+00 -4.32382584e-01 8.92640769e-01
6.14072263e-01 -1.28801046e-02 1.48175776e-01 5.42558312e-01
-8.36072981e-01 8.34501565e-01 1.38376638e-01 2.91344047e-01
6.41447961e-01 -9.28232193e-01 -2.20800713e-01 8.31709564e-01
9.46643651e-02 4.85367894e-01 2.11657137e-01 -5.95359206e-01
-1.07169366e+00 -1.09583616e+00 7.45279968e-01 1.62323773e-01
4.94057566e-01 -6.35021031e-02 -1.05077863e+00 3.54907453e-01
-1.81910083e-01 1.51544482e-01 2.09113717e-01 -1.77597672e-01
-4.18521911e-01 -4.68830943e-01 -7.00544953e-01 6.98789656e-01
1.22697377e+00 -7.90288270e-01 -5.34933805e-01 2.73897886e-01
6.63503289e-01 -2.35809371e-01 -5.20246983e-01 5.93303859e-01
7.59458899e-01 -8.53749156e-01 8.62458587e-01 -3.81431431e-01
1.49630681e-01 -5.43288350e-01 9.67941806e-02 -1.10967839e+00
-2.68262953e-01 -4.05877173e-01 -2.33377755e-01 1.18115556e+00
2.34579816e-01 -5.56427896e-01 6.31759107e-01 3.90404195e-01
-2.11869746e-01 -9.39043164e-01 -1.00662339e+00 -7.85198748e-01
-1.95470542e-01 -5.64535439e-01 2.58325666e-01 7.86579847e-01
-1.96281239e-01 2.90386826e-01 -5.41889906e-01 6.38529882e-02
4.56918001e-01 4.18014526e-01 4.12753105e-01 -8.80285382e-01
-2.94544131e-01 -7.00394213e-01 -3.55200708e-01 -1.59752774e+00
-4.20128629e-02 -7.89937556e-01 2.60064065e-01 -1.56687200e+00
2.75983423e-01 -6.53875232e-01 -7.98380136e-01 6.52426600e-01
-4.06920195e-01 -6.24812581e-02 2.99165130e-01 5.02767682e-01
-7.23275483e-01 9.95886505e-01 1.52686703e+00 -3.34132165e-01
-4.23707843e-01 -3.63585502e-02 -1.99084774e-01 8.93421173e-01
5.31845808e-01 -3.42328787e-01 -4.48243409e-01 -6.90030754e-01
-1.95306446e-02 -1.04587950e-01 6.10728562e-01 -9.27349329e-01
6.86251163e-01 -1.55700892e-01 3.81626219e-01 -7.34564424e-01
3.37278336e-01 -7.77586639e-01 -3.31550151e-01 5.95549047e-01
-2.16003522e-01 -4.89638478e-01 -9.37042013e-03 7.22708523e-01
-4.54168141e-01 1.96739241e-01 9.81997073e-01 -1.07016815e-02
-1.07345212e+00 7.87341654e-01 -3.60158421e-02 7.86197037e-02
7.71283150e-01 -3.66963059e-01 -1.89013883e-01 -2.78376102e-01
-6.66823745e-01 8.48890960e-01 3.23564112e-01 5.95909595e-01
1.01111889e+00 -1.46465743e+00 -5.15912175e-01 3.81662667e-01
1.90006688e-01 -1.43469200e-02 5.89672804e-01 1.15277302e+00
-2.40477160e-01 1.95739344e-01 -1.08907454e-01 -7.66050398e-01
-1.17003369e+00 1.99589983e-01 4.96654123e-01 -1.77939042e-01
-9.27060544e-01 1.14057708e+00 5.02256155e-01 -1.70264482e-01
6.29319489e-01 -2.47433245e-01 -4.33391094e-01 4.66230102e-02
4.79593635e-01 1.08901903e-01 -3.22735965e-01 -6.15063906e-01
-5.08659840e-01 8.54214072e-01 -1.02201723e-01 -1.16218299e-01
1.13106680e+00 -2.94759393e-01 8.59548226e-02 5.44669390e-01
1.31149924e+00 -1.34893700e-01 -1.62056923e+00 -5.34822881e-01
-2.75756538e-01 -3.73137623e-01 3.11162919e-01 -6.89922988e-01
-1.57435334e+00 1.12929749e+00 7.25832522e-01 -4.32436943e-01
1.41281784e+00 -2.91272044e-01 1.12936139e+00 2.98431247e-01
2.45718986e-01 -1.32923758e+00 2.08681509e-01 6.40153229e-01
8.40263188e-01 -1.23087406e+00 -3.15366745e-01 -1.21220119e-01
-8.19469452e-01 8.78046870e-01 7.67779589e-01 -1.18887290e-01
6.89288735e-01 9.54974890e-02 5.55926375e-02 -9.29794535e-02
-5.64100206e-01 -4.61392283e-01 3.99863511e-01 2.88611025e-01
-6.98832422e-02 -1.10982552e-01 -4.82459038e-01 5.97411931e-01
3.65791261e-01 1.21178158e-01 -1.62498608e-01 9.50194061e-01
-3.78256202e-01 -8.81031334e-01 1.68529168e-01 3.75237346e-01
-1.14129506e-01 -2.83391356e-01 -2.51725435e-01 7.46741712e-01
4.75721270e-01 7.25571156e-01 -4.05713543e-02 -6.58869326e-01
3.59437197e-01 -2.86339194e-01 2.32896388e-01 -2.08386689e-01
-3.83528531e-01 4.05079186e-01 -1.71288103e-01 -7.11051941e-01
-3.33594710e-01 -6.07671022e-01 -1.60825133e+00 5.11143844e-05
-5.37201107e-01 5.34144975e-03 3.32409441e-01 1.07931292e+00
1.03554666e-01 5.23162246e-01 5.85050523e-01 -4.63058919e-01
-2.77769864e-01 -8.02480102e-01 -5.72430491e-01 3.01636517e-01
4.61219013e-01 -7.09774256e-01 -4.48600352e-01 -1.66491106e-01] | [9.21223258972168, -6.47495698928833] |
edc8b19f-9b54-4eb1-9d5d-346305b341a6 | unifying-motion-deblurring-and-frame | 2203.12178 | null | https://arxiv.org/abs/2203.12178v2 | https://arxiv.org/pdf/2203.12178v2.pdf | Unifying Motion Deblurring and Frame Interpolation with Events | Slow shutter speed and long exposure time of frame-based cameras often cause visual blur and loss of inter-frame information, degenerating the overall quality of captured videos. To this end, we present a unified framework of event-based motion deblurring and frame interpolation for blurry video enhancement, where the extremely low latency of events is leveraged to alleviate motion blur and facilitate intermediate frame prediction. Specifically, the mapping relation between blurry frames and sharp latent images is first predicted by a learnable double integral network, and a fusion network is then proposed to refine the coarse results via utilizing the information from consecutive blurry inputs and the concurrent events. By exploring the mutual constraints among blurry frames, latent images, and event streams, we further propose a self-supervised learning framework to enable network training with real-world blurry videos and events. Extensive experiments demonstrate that our method compares favorably against the state-of-the-art approaches and achieves remarkable performance on both synthetic and real-world datasets. | ['Lei Yu', 'Xiang Zhang'] | 2022-03-23 | unifying-motion-deblurring-and-frame-1 | https://openaccess.thecvf.com/content/CVPR2022/html/Zhang_Unifying_Motion_Deblurring_and_Frame_Interpolation_With_Events_CVPR_2022_paper.html | https://openaccess.thecvf.com/content/CVPR2022/papers/Zhang_Unifying_Motion_Deblurring_and_Frame_Interpolation_With_Events_CVPR_2022_paper.pdf | cvpr-2022-6 | ['video-enhancement'] | ['computer-vision'] | [ 3.52104396e-01 -6.42305791e-01 -2.41324112e-01 -3.87349784e-01
-4.45807248e-01 -3.39676917e-01 4.67128336e-01 -5.44987500e-01
-1.96281120e-01 8.36840987e-01 5.73761165e-01 -2.64813043e-02
-1.83835059e-01 -1.50428504e-01 -8.03555250e-01 -5.33150434e-01
-2.01970890e-01 -6.10984504e-01 2.78442055e-01 4.40066725e-01
1.86203301e-01 2.48568237e-01 -1.14721727e+00 5.73224485e-01
8.77981484e-01 1.02065969e+00 4.32599992e-01 9.23002541e-01
3.57989788e-01 1.77250993e+00 -4.25778747e-01 -6.75428733e-02
2.45029390e-01 -4.14063573e-01 -5.40806592e-01 4.65868950e-01
5.08570611e-01 -1.12939572e+00 -1.21583116e+00 9.19237673e-01
2.53080875e-01 1.65357903e-01 1.02858365e-01 -9.98005331e-01
-9.35119212e-01 3.01227510e-01 -7.04160213e-01 6.51599407e-01
3.84116530e-01 5.92778146e-01 3.82642955e-01 -9.78574574e-01
5.21379769e-01 1.11742771e+00 6.92248523e-01 4.28067416e-01
-1.09220469e+00 -6.28713012e-01 2.01476678e-01 6.32388055e-01
-1.08642471e+00 -8.33565831e-01 8.19122612e-01 -3.36183399e-01
6.40437007e-01 1.23897716e-01 5.47196448e-01 1.15021300e+00
3.65863174e-01 7.85480678e-01 8.43983114e-01 -5.29581755e-02
1.34971097e-01 -3.97552937e-01 -1.39146954e-01 6.16694331e-01
2.57279463e-02 4.20968741e-01 -8.82995069e-01 -3.14071067e-02
1.20069897e+00 5.13281822e-01 -9.63078022e-01 -1.32236436e-01
-1.40932250e+00 2.30068147e-01 2.88788199e-01 -2.42318343e-02
-5.00851572e-01 3.99183333e-01 3.32359493e-01 1.93548769e-01
6.45086944e-01 1.96075544e-01 -3.77493948e-01 -2.85274595e-01
-1.34766603e+00 6.42687976e-02 4.48718488e-01 7.91176021e-01
5.86722255e-01 1.77986696e-01 -5.38593292e-01 6.75379157e-01
3.15663487e-01 3.04559946e-01 2.29398251e-01 -1.51558805e+00
4.24389392e-01 5.50516471e-02 5.83921671e-01 -8.18099022e-01
1.98402807e-01 -5.83805740e-02 -1.07568789e+00 6.39039874e-02
4.12866443e-01 -2.14758456e-01 -8.27185750e-01 1.57393348e+00
-2.01193001e-02 1.13898909e+00 -1.37322217e-01 1.27457130e+00
3.93665671e-01 8.54460180e-01 -7.68189952e-02 -7.70147383e-01
1.07287240e+00 -1.14100194e+00 -1.10911977e+00 -2.64207244e-01
-2.92412817e-01 -9.05223131e-01 6.87789559e-01 2.53016412e-01
-1.54278028e+00 -8.96446705e-01 -9.43503320e-01 -1.55350804e-01
3.08286875e-01 8.80352315e-03 4.65320110e-01 6.41722530e-02
-9.71923530e-01 6.44683838e-01 -1.04130757e+00 2.23652571e-01
6.34330988e-01 1.01527743e-01 -4.23813500e-02 -5.29301524e-01
-1.21760368e+00 6.51037753e-01 2.45231330e-01 2.76723921e-01
-1.25515485e+00 -1.06207395e+00 -9.85604048e-01 1.13175742e-01
3.20233941e-01 -9.44754839e-01 1.28673303e+00 -9.80723977e-01
-1.45682764e+00 3.30226243e-01 -4.33262438e-01 -6.62904561e-01
6.74868941e-01 -6.92872345e-01 -4.08354700e-01 4.55918312e-01
-1.43196613e-01 6.20407760e-01 1.46896183e+00 -1.06822121e+00
-7.56712258e-01 1.84683904e-01 1.88424550e-02 2.55204111e-01
-1.81531444e-01 1.93962336e-01 -5.71469605e-01 -9.44202125e-01
-2.69057900e-01 -5.34797609e-01 -7.88260102e-02 2.47602358e-01
3.70635511e-03 3.11464608e-01 1.19562590e+00 -9.45596695e-01
1.40422320e+00 -2.29158568e+00 1.14776827e-01 -5.75400174e-01
4.61172551e-01 2.92831153e-01 -2.22424185e-03 -2.16644444e-03
-1.77391455e-01 -4.40419942e-01 -7.85294846e-02 -4.49746996e-01
-3.56760085e-01 2.03422263e-01 -6.99253440e-01 5.51901758e-01
2.64859200e-01 9.12900209e-01 -1.20420814e+00 -3.14150333e-01
8.19053829e-01 7.70760894e-01 -4.07455653e-01 6.42081678e-01
-4.23399061e-02 7.39022672e-01 -1.44151062e-01 5.51110268e-01
8.15527856e-01 -6.00814104e-01 -1.23362675e-01 -6.02858782e-01
-1.26584396e-01 6.69151619e-02 -9.47550654e-01 1.81738079e+00
-3.61065239e-01 1.12010741e+00 1.85275003e-02 -3.90647948e-01
3.22164387e-01 4.51212674e-01 5.63576281e-01 -4.39592093e-01
-6.70806244e-02 -9.88081247e-02 -5.09292901e-01 -8.46973836e-01
6.69146359e-01 1.83067918e-01 3.94677669e-01 3.44778806e-01
-9.32865217e-03 4.96630259e-02 3.10145300e-02 1.47893786e-01
1.09750652e+00 3.85985583e-01 1.76693425e-01 1.37236610e-01
5.45683205e-01 -6.16141915e-01 6.64372444e-01 7.40143418e-01
-4.73765165e-01 8.81071568e-01 8.68327841e-02 -7.11205781e-01
-1.25781667e+00 -1.24636257e+00 1.30448341e-01 7.62114882e-01
7.59780467e-01 -1.89971849e-01 -7.57853270e-01 -4.13501501e-01
-2.95737594e-01 5.84521472e-01 -3.23864192e-01 -2.61172682e-01
-7.31430411e-01 -6.86810732e-01 2.85963714e-01 4.84069377e-01
6.66570723e-01 -7.30808020e-01 -6.20251000e-01 2.93987483e-01
-5.28044164e-01 -1.50312197e+00 -1.06757402e+00 -2.34057650e-01
-8.16572249e-01 -1.09917438e+00 -7.90102780e-01 -5.93325257e-01
5.02244890e-01 8.33911896e-01 9.92189109e-01 -1.12457685e-01
-1.41364142e-01 2.73375869e-01 1.87990330e-02 2.67767608e-01
-3.89429599e-01 -5.56610644e-01 3.34111825e-02 3.24453652e-01
-2.11865697e-02 -6.18239105e-01 -1.13773835e+00 2.03743204e-01
-1.27143347e+00 5.77602506e-01 4.82851177e-01 9.87258971e-01
1.86085969e-01 8.76832604e-02 2.73889780e-01 -2.40125328e-01
3.34011465e-01 -4.33662534e-01 -6.38425291e-01 2.77329177e-01
-4.02310193e-01 -8.09670687e-02 5.08146167e-01 -7.73591578e-01
-1.68616998e+00 -8.93220678e-02 3.83943737e-01 -1.05025125e+00
-1.29466861e-01 -1.01571549e-02 1.02960803e-01 -6.27042279e-02
3.81304175e-01 3.68862242e-01 -2.04397291e-01 -3.09138298e-01
4.01648849e-01 6.25081420e-01 1.05014765e+00 -1.29931256e-01
5.84553957e-01 7.91914344e-01 -4.16280985e-01 -4.03939843e-01
-1.04437482e+00 -3.67891967e-01 -3.55647624e-01 -4.17205155e-01
7.71923363e-01 -1.51383948e+00 -5.47362030e-01 9.50670958e-01
-1.60976863e+00 -3.79236132e-01 -2.28238925e-01 6.63735867e-01
-6.00496411e-01 7.01002121e-01 -1.38627934e+00 -6.02467597e-01
-1.90920755e-01 -1.23214519e+00 1.02717674e+00 5.05688846e-01
2.00968105e-02 -9.71514821e-01 -8.75500739e-02 4.18770313e-01
4.42674249e-01 -1.03742257e-01 3.03064346e-01 2.89386392e-01
-1.29607177e+00 1.95711404e-01 -7.83010244e-01 5.34931779e-01
6.14745438e-01 -1.11659952e-01 -1.10574257e+00 -3.40727359e-01
5.55685222e-01 1.63197014e-02 9.79671717e-01 8.18012357e-01
1.37014461e+00 -4.14991885e-01 -1.54657394e-01 1.02115846e+00
1.36831415e+00 1.91407099e-01 8.20739031e-01 1.18555769e-01
9.08932745e-01 -1.42040721e-03 5.07368922e-01 6.72070742e-01
3.64283264e-01 4.26638931e-01 3.16203386e-01 -1.13718525e-01
-4.60323691e-01 -1.65295258e-01 6.28863096e-01 6.87262654e-01
-7.18365014e-02 -1.82977632e-01 -3.21464211e-01 6.16151392e-01
-2.04977703e+00 -1.38550663e+00 1.67062074e-01 2.02757740e+00
1.15731704e+00 1.15677062e-02 -3.47558081e-01 -2.93281376e-01
9.73408461e-01 7.59286404e-01 -6.53519690e-01 3.54513913e-01
-2.21962422e-01 -1.98353335e-01 4.86355156e-01 7.73130119e-01
-1.16193008e+00 6.39870822e-01 6.78084612e+00 7.05183685e-01
-1.16340470e+00 1.99901581e-01 9.92864788e-01 -3.52025181e-01
1.89345796e-02 4.55132164e-02 -3.33957195e-01 1.04101992e+00
7.85814881e-01 -1.28242016e-01 9.92879212e-01 3.74267429e-01
8.94874215e-01 -1.72807828e-01 -1.10594583e+00 1.24469554e+00
1.11362524e-01 -1.66212070e+00 -1.90606534e-01 -2.92176187e-01
1.08412266e+00 4.62255329e-02 3.02042756e-02 -2.64151275e-01
2.08859771e-01 -7.62313426e-01 7.74369836e-01 9.96229529e-01
8.87458384e-01 -2.95437634e-01 4.38467920e-01 1.59044743e-01
-1.09043992e+00 -2.18517318e-01 -2.00317994e-01 -2.90224224e-01
5.79953849e-01 8.49282622e-01 -3.38231444e-01 4.94620800e-01
8.18387806e-01 1.33930838e+00 -4.06232893e-01 1.08545363e+00
-2.18352377e-01 5.90611577e-01 1.41175866e-01 7.71792471e-01
-3.89252044e-02 -1.50651922e-02 6.64601862e-01 1.31256759e+00
3.04454982e-01 7.28916898e-02 -1.59701973e-01 8.64948690e-01
-7.59008899e-02 -9.03494537e-01 -2.86966920e-01 3.37847143e-01
3.83767217e-01 1.12747335e+00 -1.55966461e-01 -6.26255810e-01
-7.27546275e-01 1.38957441e+00 1.90995838e-02 7.42017567e-01
-1.37779486e+00 -4.29070741e-02 9.56463754e-01 -2.79977232e-01
5.03070831e-01 -2.27951899e-01 -1.82473496e-01 -1.87070882e+00
7.01723173e-02 -7.20942378e-01 1.58474475e-01 -1.34373367e+00
-1.38994527e+00 3.93249154e-01 -1.06474891e-01 -1.40621603e+00
-1.84832409e-01 -2.96819210e-01 -5.63947856e-01 8.74290228e-01
-2.00108528e+00 -8.59902740e-01 -6.66692674e-01 6.58734560e-01
1.08409822e+00 1.21607676e-01 4.05445434e-02 4.72989082e-01
-5.33823133e-01 6.71432447e-03 1.78342685e-01 -2.63009928e-02
1.11281645e+00 -8.73346984e-01 3.75106007e-01 1.39986122e+00
-1.22423261e-01 2.56283611e-01 7.45742679e-01 -7.27989554e-01
-1.37319720e+00 -1.38749349e+00 6.01474881e-01 -2.87868351e-01
7.43218899e-01 8.19680914e-02 -1.08449781e+00 5.80371261e-01
5.76373458e-01 4.85019922e-01 8.03412646e-02 -7.01343358e-01
-2.62765765e-01 -3.24893594e-01 -6.96251690e-01 5.39844394e-01
1.02669442e+00 -8.18109035e-01 -5.55374146e-01 1.49936259e-01
1.00986719e+00 -7.74962485e-01 -7.30517566e-01 3.72192055e-01
4.90379155e-01 -1.20728576e+00 1.22856390e+00 -1.92970589e-01
8.96524549e-01 -5.87219834e-01 2.28021502e-01 -1.12295675e+00
-4.98571783e-01 -1.14570034e+00 -9.73422468e-01 1.06559360e+00
-3.37494105e-01 -1.07992589e-01 5.50325930e-01 5.73144734e-01
-6.65434003e-02 -4.37158912e-01 -6.34145021e-01 -3.93879145e-01
-8.28330636e-01 -3.81376773e-01 3.99004132e-01 9.52446759e-01
-2.34367698e-01 1.09377369e-01 -1.19669211e+00 4.32939321e-01
9.25221205e-01 9.40396562e-02 3.69475424e-01 -3.71870756e-01
-5.14156938e-01 1.86933037e-02 -6.02608509e-02 -1.67059875e+00
-9.61134210e-02 1.23384176e-03 2.01489449e-01 -9.93440866e-01
4.64422017e-01 5.00990376e-02 -5.22112548e-01 6.68880530e-03
-7.23760366e-01 3.14916551e-01 2.62952685e-01 5.55651307e-01
-1.03203881e+00 4.85745311e-01 1.42013371e+00 -3.24921198e-02
-1.76595032e-01 -2.21072987e-01 -3.65371734e-01 8.07583690e-01
2.64942229e-01 -1.93154871e-01 -4.15824771e-01 -9.36713815e-01
-9.68180373e-02 6.32405043e-01 8.12776983e-01 -1.01023054e+00
6.10901237e-01 -4.29255366e-01 6.39756083e-01 -4.89926904e-01
3.56569916e-01 -7.62217462e-01 4.13105547e-01 2.55549908e-01
-4.53690648e-01 -1.12317398e-01 8.96856710e-02 8.91428113e-01
-3.86320978e-01 1.54355675e-01 8.94034326e-01 7.44540170e-02
-8.26806962e-01 5.24981260e-01 -2.32524142e-01 -7.12906793e-02
8.76227260e-01 -1.91057831e-01 -5.00018001e-01 -6.38663828e-01
-4.20561045e-01 4.04198840e-02 6.65225387e-01 4.52841401e-01
8.83445323e-01 -1.25326562e+00 -5.54241061e-01 2.68362045e-01
-3.08436692e-01 -7.58166565e-03 6.93919241e-01 8.78807127e-01
-5.45288742e-01 1.83530331e-01 -2.43445292e-01 -6.97702765e-01
-1.12549591e+00 7.43733168e-01 3.38801771e-01 -1.16488673e-01
-5.12337565e-01 6.43891573e-01 5.32859325e-01 6.51524782e-01
2.36397550e-01 -5.64855039e-01 2.13419095e-01 -4.78650272e-01
1.02450693e+00 5.55857122e-01 -4.60520297e-01 -3.73046935e-01
6.17785007e-02 1.61341220e-01 -2.27057889e-01 1.58578753e-01
1.34131110e+00 -8.97299409e-01 -6.96897805e-02 2.38600656e-01
1.06437123e+00 -9.37096179e-02 -2.47514224e+00 -4.41563874e-01
-3.72494817e-01 -9.77171063e-01 1.31744087e-01 -5.53389311e-01
-1.11032915e+00 6.17753029e-01 6.08062983e-01 5.48025630e-02
1.73817587e+00 -1.98527709e-01 1.14344978e+00 -8.59661028e-02
6.68395460e-02 -7.47345865e-01 3.27250272e-01 3.55592757e-01
5.67876816e-01 -1.31344926e+00 1.29525229e-01 -4.48726416e-01
-4.65082020e-01 1.17822742e+00 4.30786163e-01 -2.39178523e-01
2.82940120e-01 4.45960075e-01 -8.06427896e-02 3.17474574e-01
-1.08241439e+00 3.53443205e-01 3.05356324e-01 3.45170885e-01
1.03971526e-01 -4.61187661e-01 1.43427253e-01 2.91857451e-01
6.40983462e-01 7.86874890e-01 5.64134240e-01 6.88736320e-01
-2.15146989e-01 -6.00884020e-01 -5.32768369e-01 1.40801072e-01
-5.78053474e-01 -3.42021883e-01 3.22607189e-01 2.54343122e-01
2.47480303e-01 8.34503889e-01 2.01160803e-01 -7.04459250e-02
-1.43830627e-01 -3.18899006e-01 5.50515532e-01 -7.40156844e-02
-2.09978387e-01 2.84606189e-01 -2.40334690e-01 -9.19771731e-01
-8.13263059e-01 -5.82208633e-01 -6.63502991e-01 -4.44493502e-01
-2.34230816e-01 -3.19612980e-01 5.03281765e-02 1.06221282e+00
6.87425494e-01 7.38366187e-01 8.20742726e-01 -1.17813408e+00
-3.77775609e-01 -7.45085061e-01 -2.72371590e-01 5.74391901e-01
1.06011963e+00 -2.96312362e-01 -4.77512002e-01 9.84975398e-01] | [11.191200256347656, -2.291553497314453] |
68131d85-7e04-464e-b92b-26af14572cbf | language-to-rewards-for-robotic-skill | 2306.08647 | null | https://arxiv.org/abs/2306.08647v2 | https://arxiv.org/pdf/2306.08647v2.pdf | Language to Rewards for Robotic Skill Synthesis | Large language models (LLMs) have demonstrated exciting progress in acquiring diverse new capabilities through in-context learning, ranging from logical reasoning to code-writing. Robotics researchers have also explored using LLMs to advance the capabilities of robotic control. However, since low-level robot actions are hardware-dependent and underrepresented in LLM training corpora, existing efforts in applying LLMs to robotics have largely treated LLMs as semantic planners or relied on human-engineered control primitives to interface with the robot. On the other hand, reward functions are shown to be flexible representations that can be optimized for control policies to achieve diverse tasks, while their semantic richness makes them suitable to be specified by LLMs. In this work, we introduce a new paradigm that harnesses this realization by utilizing LLMs to define reward parameters that can be optimized and accomplish variety of robotic tasks. Using reward as the intermediate interface generated by LLMs, we can effectively bridge the gap between high-level language instructions or corrections to low-level robot actions. Meanwhile, combining this with a real-time optimizer, MuJoCo MPC, empowers an interactive behavior creation experience where users can immediately observe the results and provide feedback to the system. To systematically evaluate the performance of our proposed method, we designed a total of 17 tasks for a simulated quadruped robot and a dexterous manipulator robot. We demonstrate that our proposed method reliably tackles 90% of the designed tasks, while a baseline using primitive skills as the interface with Code-as-policies achieves 50% of the tasks. We further validated our method on a real robot arm where complex manipulation skills such as non-prehensile pushing emerge through our interactive system. | ['Fei Xia', 'Yuval Tassa', 'Jie Tan', 'Dorsa Sadigh', 'Nicolas Heess', 'Tingnan Zhang', 'Andy Zeng', 'Peng Xu', 'Ted Xiao', 'Brian Ichter', 'Jan Humplik', 'Leonard Hasenclever', 'Tom Erez', 'Hao-Tien Lewis Chiang', 'Montse Gonzalez Arenas', 'Kuang-Huei Lee', 'Sean Kirmani', 'Chuyuan Fu', 'Nimrod Gileadi', 'Wenhao Yu'] | 2023-06-14 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [-3.27491648e-02 4.64603156e-01 -3.80175114e-01 -1.27207339e-01
-5.12861550e-01 -6.61306918e-01 5.54241657e-01 -2.45262951e-01
-3.30749691e-01 4.80855465e-01 3.60384164e-03 -4.45131510e-01
-2.17049360e-01 -5.03958821e-01 -1.12910557e+00 -3.26819211e-01
-3.47416908e-01 5.99581361e-01 1.85683459e-01 -7.04181015e-01
5.09944737e-01 5.22330165e-01 -1.83414555e+00 2.00023085e-01
1.13583827e+00 5.75379789e-01 1.01464462e+00 6.74355984e-01
-5.98991401e-02 9.97749448e-01 -4.88647640e-01 3.12092513e-01
3.38945985e-01 1.63980052e-01 -8.37345898e-01 -2.39033937e-01
-4.73093688e-02 -3.07662398e-01 -1.39523432e-01 8.27491462e-01
2.76949972e-01 1.70132846e-01 4.53062713e-01 -1.51306224e+00
-4.80185568e-01 1.09476602e+00 -9.66800749e-03 -7.46601999e-01
5.85907340e-01 7.26196945e-01 8.99414361e-01 -4.87774849e-01
6.82059646e-01 1.58994484e+00 3.02241832e-01 8.79999161e-01
-1.09583914e+00 -4.76368815e-01 1.00431815e-01 1.18005108e-02
-1.04787171e+00 -6.51143715e-02 5.76665640e-01 -6.21761143e-01
1.27557266e+00 -1.41622365e-01 5.60613096e-01 1.33342755e+00
7.01085091e-01 9.37686622e-01 8.27601373e-01 -3.40582043e-01
3.19181859e-01 -1.54163002e-03 -1.73018113e-01 8.46188247e-01
3.04066762e-02 4.31868881e-01 -3.59451592e-01 1.26079068e-01
8.72149050e-01 5.55944145e-02 8.46236646e-02 -7.29845464e-01
-1.46562934e+00 5.09835541e-01 6.34468853e-01 1.62962139e-01
-2.34777108e-01 6.52830958e-01 3.54924023e-01 3.60952795e-01
-5.17443299e-01 1.27304816e+00 -8.13803256e-01 -2.78354764e-01
-1.91995695e-01 5.25074542e-01 1.09923756e+00 1.47966671e+00
6.69493973e-01 1.08519971e-01 -3.29973161e-01 6.97799265e-01
2.23259911e-01 4.71234769e-01 5.25318921e-01 -1.40185690e+00
6.29505098e-01 8.65735769e-01 3.38873059e-01 -6.45772159e-01
-7.71017551e-01 -1.88459054e-01 -1.45742342e-01 6.94855452e-01
1.32415295e-01 -8.04568604e-02 -8.84645700e-01 1.77389848e+00
1.57439113e-01 -2.78597891e-01 3.49644244e-01 1.06989813e+00
1.94459304e-01 6.34746850e-01 2.65109569e-01 3.78706276e-01
1.18281281e+00 -1.31070185e+00 -3.26506257e-01 -3.74467462e-01
9.52166736e-01 -2.02609390e-01 1.82968891e+00 5.58445573e-01
-9.15935636e-01 -6.38395011e-01 -1.25969970e+00 2.12819129e-02
-2.03266710e-01 2.50676006e-01 9.06137645e-01 1.79556068e-02
-9.67045128e-01 1.06894374e+00 -1.11207640e+00 -3.03643614e-01
1.83303416e-01 7.27913797e-01 -2.56465793e-01 4.72511351e-02
-9.78972316e-01 1.25630641e+00 7.12766290e-01 -1.81953222e-01
-1.53453064e+00 -4.60434049e-01 -1.06022012e+00 1.47106245e-01
8.31470728e-01 -7.32303023e-01 1.66583180e+00 -6.07773304e-01
-2.24528313e+00 5.43733299e-01 5.59547901e-01 -3.59668583e-01
2.11398631e-01 -5.07992208e-01 -1.90758538e-02 -6.50272965e-02
1.80317506e-01 9.18425262e-01 7.46673286e-01 -1.26987982e+00
-5.51382244e-01 8.32908303e-02 6.57711864e-01 2.05846652e-01
-1.59984156e-01 -2.68012583e-01 -2.37779185e-01 -4.18168813e-01
-1.45305008e-01 -1.42853427e+00 -4.70604569e-01 -1.39365032e-01
-2.55969852e-01 -2.32420236e-01 5.67271113e-01 -2.46107712e-01
7.57591724e-01 -2.00128841e+00 7.50104904e-01 9.11958516e-03
-1.13867454e-01 3.77089009e-02 -4.03682202e-01 5.69417715e-01
2.54954219e-01 6.85342550e-02 -1.20407462e-01 -8.73840675e-02
5.29861927e-01 6.53992891e-01 -2.97800839e-01 1.03367353e-03
3.52012575e-01 9.06907022e-01 -1.20512438e+00 -4.16964442e-01
2.68850833e-01 -6.56711385e-02 -1.02489805e+00 4.66484070e-01
-8.15233648e-01 7.86597192e-01 -6.95015132e-01 7.21778154e-01
-4.02058326e-02 -8.53659660e-02 3.99103642e-01 3.12471628e-01
-7.43647963e-02 3.10716867e-01 -8.74417365e-01 2.27375150e+00
-1.03816807e+00 1.23240627e-01 3.40157807e-01 -8.05467963e-01
8.87410760e-01 2.30482519e-02 3.25756580e-01 -5.34359813e-01
1.46806091e-01 4.44276690e-01 1.34454116e-01 -7.58632541e-01
6.35046184e-01 1.19875111e-01 -6.78239286e-01 1.62468120e-01
2.33805493e-01 -9.24861312e-01 7.56833777e-02 -1.66585609e-01
1.27783298e+00 1.08277535e+00 2.88322896e-01 -1.88524947e-01
4.30036932e-01 4.61762100e-01 2.92300463e-01 9.76381183e-01
-6.16750971e-04 1.17047958e-01 4.39845949e-01 -1.26105845e-01
-1.01959205e+00 -1.05570924e+00 2.89456278e-01 1.48243546e+00
2.37580553e-01 -3.05659115e-01 -8.19511950e-01 -4.58268791e-01
9.88589153e-02 1.05241776e+00 -1.98307976e-01 -3.60909462e-01
-8.38698983e-01 6.16361499e-02 5.26023686e-01 3.78392398e-01
1.73263982e-01 -1.61313510e+00 -1.12350130e+00 2.32065544e-01
1.46725953e-01 -1.20341623e+00 -2.42011532e-01 4.27588075e-01
-1.00006783e+00 -9.33796763e-01 -2.30467141e-01 -9.53635335e-01
5.76366186e-01 -4.09850441e-02 9.54851210e-01 2.39661783e-02
-3.06069583e-01 4.50454444e-01 -4.23980087e-01 -2.09780172e-01
-9.17625725e-01 2.80935377e-01 2.94008404e-01 -8.22221279e-01
-4.62629706e-01 -6.79024458e-01 -1.96705177e-01 1.96772859e-01
-7.61007488e-01 3.36267173e-01 1.13120103e+00 8.67468417e-01
2.17128679e-01 -2.24335745e-01 5.19116998e-01 -5.19881845e-01
8.12045634e-01 -3.70774269e-01 -7.87012637e-01 2.24222153e-01
-5.84498167e-01 6.34921014e-01 9.19217646e-01 -8.05871129e-01
-9.14667487e-01 3.93990189e-01 6.48829155e-03 -3.68734598e-01
-1.60538927e-01 7.32422829e-01 -1.03466034e-01 -4.36006114e-02
7.45361447e-01 7.23793507e-02 1.65779382e-01 -2.24021018e-01
7.91247308e-01 5.34844100e-01 6.57727122e-01 -1.51536953e+00
5.60516238e-01 -1.51839286e-01 9.32334960e-02 -3.53284806e-01
-3.46874893e-01 -1.13795862e-01 -3.53605658e-01 -6.45857230e-02
6.22749150e-01 -7.69344926e-01 -1.38448298e+00 2.85644472e-01
-1.07092047e+00 -9.92123485e-01 -1.94467142e-01 3.96034151e-01
-1.29528069e+00 3.63610778e-03 -6.70219243e-01 -7.85905540e-01
-1.07368063e-02 -1.74978173e+00 1.30248487e+00 1.46602437e-01
-5.08680642e-01 -4.62293118e-01 -3.49045902e-01 7.49337152e-02
4.21790719e-01 2.14223653e-01 1.17603683e+00 -4.95429307e-01
-6.88040614e-01 -9.14186388e-02 2.62078285e-01 7.93904886e-02
3.69991059e-03 -3.26796174e-01 -5.43771088e-01 -2.20156655e-01
-1.56250715e-01 -7.86472619e-01 5.46711534e-02 -6.25193864e-02
1.31001341e+00 -3.55316579e-01 -4.63339955e-01 4.77892399e-01
1.10613120e+00 2.01525524e-01 5.61422825e-01 6.50471807e-01
6.59634769e-01 6.43642306e-01 1.06125653e+00 3.11514586e-01
4.59968388e-01 9.69583690e-01 7.51620471e-01 5.00356972e-01
3.82527593e-03 -5.57157636e-01 9.25762594e-01 6.22987568e-01
-1.14042036e-01 2.16048315e-01 -1.05182409e+00 2.11760506e-01
-2.10198641e+00 -5.47392726e-01 2.68628746e-01 1.88036621e+00
1.08640957e+00 2.53778428e-01 -2.12065175e-01 -4.18747574e-01
3.39825690e-01 -2.97510207e-01 -8.06456804e-01 -4.11105692e-01
5.91801882e-01 3.02264363e-01 4.90401775e-01 4.12567377e-01
-7.89626896e-01 1.49035013e+00 5.76153040e+00 7.58370638e-01
-1.21264410e+00 -1.68060258e-01 -2.67717600e-01 6.32218421e-02
8.64490941e-02 1.36865407e-01 -6.40767872e-01 1.42126471e-01
9.07272160e-01 4.48502488e-02 8.85780811e-01 1.58414125e+00
2.91998088e-01 -8.85002613e-02 -1.55846536e+00 7.64979303e-01
-4.59828109e-01 -1.08733034e+00 -1.27030117e-02 -2.52436459e-01
5.35111070e-01 -2.88139507e-02 1.67385861e-02 1.09357393e+00
7.92167962e-01 -1.15684140e+00 1.23902309e+00 4.87243384e-01
7.40048528e-01 -4.64010179e-01 3.69875878e-01 8.95210743e-01
-9.41195846e-01 -6.08157337e-01 -1.40943915e-01 -4.70275432e-01
8.77837911e-02 -1.60300046e-01 -1.30224442e+00 5.56031227e-01
4.47455406e-01 5.24772346e-01 -2.27483407e-01 5.54867029e-01
-6.52319491e-01 -1.08644748e-02 -7.35118836e-02 -5.12848079e-01
1.75190806e-01 4.67735603e-02 5.36107898e-01 8.36612999e-01
3.39282364e-01 -2.23893598e-01 6.01423740e-01 1.22635448e+00
2.47348830e-01 -2.42399380e-01 -7.98493326e-01 -4.25149262e-01
3.28943938e-01 1.12407649e+00 -3.68589342e-01 -1.46259055e-01
-2.48687584e-02 8.47185910e-01 6.04105949e-01 3.14778447e-01
-1.01623988e+00 -5.00635564e-01 8.54335129e-01 -2.71672785e-01
-2.38639154e-02 -8.81044805e-01 -3.15528929e-01 -1.05480933e+00
-5.59775420e-02 -1.13143003e+00 -2.76190579e-01 -1.08331525e+00
-8.20162773e-01 2.69254923e-01 2.37056330e-01 -1.28306580e+00
-5.54347336e-01 -1.03718817e+00 -2.32365549e-01 6.25676930e-01
-1.26789117e+00 -1.01321578e+00 -2.08576277e-01 2.44683608e-01
7.66958177e-01 -3.08584988e-01 8.98148298e-01 -1.15003563e-01
-2.35457346e-01 2.26861030e-01 -4.80483204e-01 -2.88217694e-01
4.46364373e-01 -1.04045331e+00 3.03598046e-01 3.60275865e-01
-4.37418401e-01 1.01684058e+00 7.19724000e-01 -6.12894952e-01
-2.05599356e+00 -8.34259927e-01 2.45184109e-01 -5.27354717e-01
8.09199452e-01 -5.41871905e-01 -4.83656675e-01 7.08781660e-01
-2.02038080e-01 -2.88508028e-01 -6.16709329e-02 -1.00052329e-02
-1.89490765e-02 1.23811327e-01 -9.77464199e-01 1.13868797e+00
1.42211390e+00 -3.62436980e-01 -9.26293671e-01 3.56402487e-01
1.39172328e+00 -8.85547757e-01 -8.43220770e-01 5.18581510e-01
6.19112730e-01 -4.41505194e-01 1.03289115e+00 -7.82685220e-01
8.21451724e-01 -5.36862791e-01 -2.12276801e-01 -1.40764105e+00
-1.32217184e-01 -7.19160199e-01 -9.58784670e-02 7.21459985e-01
4.02706236e-01 -6.11784756e-01 4.60608274e-01 5.94917297e-01
-8.76604497e-01 -8.30622256e-01 -4.00402486e-01 -9.83923137e-01
-6.83885962e-02 -6.97976649e-01 6.57836795e-01 5.30802488e-01
5.61071873e-01 1.70174405e-01 -2.19980270e-01 2.95707166e-01
6.99640810e-02 1.33063689e-01 1.12037683e+00 -8.83747697e-01
-7.07205474e-01 -6.67339444e-01 6.72582090e-02 -1.30463326e+00
7.02503026e-01 -1.23827481e+00 7.57288158e-01 -1.37086666e+00
-1.56856507e-01 -9.27378595e-01 1.99844465e-02 7.84718871e-01
2.27279842e-01 -7.85873234e-01 4.69059944e-01 3.27432960e-01
-5.28669000e-01 6.41146719e-01 1.64897025e+00 -2.33330891e-01
-4.99732822e-01 -4.93110828e-02 -5.17416418e-01 7.35990465e-01
7.32005477e-01 -3.48240346e-01 -5.30660987e-01 -5.16177535e-01
3.29879403e-01 4.84253258e-01 2.43067890e-01 -1.23705447e+00
2.58742839e-01 -6.84922636e-01 -2.37107083e-01 1.12555727e-01
2.22338423e-01 -1.03675508e+00 -4.45498377e-02 8.52626681e-01
-5.99902332e-01 -4.81693894e-02 2.85397530e-01 4.01906550e-01
6.85575232e-02 -3.79635274e-01 3.06359380e-01 -2.87764102e-01
-1.19584012e+00 -7.20778629e-02 -4.50543761e-01 -3.58847529e-01
1.17507231e+00 1.14638902e-01 -3.92211407e-01 -9.86046866e-02
-7.46583402e-01 6.42208755e-01 6.25449181e-01 8.08425546e-01
5.05778909e-01 -1.04424465e+00 -5.59476055e-02 1.65082775e-02
3.01557779e-01 1.81710303e-01 -2.69528508e-01 6.73175216e-01
-6.67990923e-01 4.98460919e-01 -5.84590673e-01 -6.83188379e-01
-5.53263068e-01 7.32712328e-01 1.43173352e-01 -3.13695639e-01
-6.47675097e-01 4.73018587e-01 1.62868008e-01 -1.18054771e+00
2.98648447e-01 -9.18173850e-01 -4.14536148e-02 -6.93181038e-01
-5.55132367e-02 -9.36201215e-03 -2.80250549e-01 7.69023746e-02
-3.15386385e-01 4.57619876e-01 3.42767209e-01 -1.41277105e-01
1.44591045e+00 2.62390196e-01 -8.47620368e-02 4.70554680e-01
6.14795268e-01 -2.75071979e-01 -1.58627570e+00 2.99654961e-01
3.38757515e-01 -5.09819426e-02 -5.61567843e-01 -9.48040485e-01
-4.35918331e-01 6.75305605e-01 1.43844128e-01 -2.93236375e-01
6.98344946e-01 4.83524539e-02 5.81519246e-01 1.11566365e+00
1.31669199e+00 -1.27538431e+00 6.88124239e-01 9.77522373e-01
1.19371426e+00 -1.15813136e+00 -2.99376845e-01 -2.81191528e-01
-8.87474716e-01 1.23955691e+00 9.76597667e-01 -3.12533259e-01
-2.87059676e-02 4.20666426e-01 -2.38335997e-01 5.92884682e-02
-8.11112106e-01 -1.52695581e-01 -3.70461307e-02 7.47323215e-01
1.43325165e-01 2.18816549e-01 -1.40108541e-01 8.27171803e-01
-3.72592986e-01 2.51043290e-01 6.00269914e-01 1.33258176e+00
-5.59280932e-01 -1.25232410e+00 -2.58640528e-01 3.65354270e-02
8.92113149e-02 1.92498103e-01 4.71883230e-02 1.00682187e+00
2.06010626e-03 6.96696222e-01 -4.06806320e-01 -6.50785565e-01
5.62300384e-01 7.63736740e-02 7.55506635e-01 -1.16389084e+00
-5.80224752e-01 -5.13487875e-01 1.54038161e-01 -1.07057714e+00
4.60040532e-02 -3.68274242e-01 -2.03459525e+00 2.15447903e-01
8.73332918e-02 -1.76964983e-01 1.03131235e+00 9.62734818e-01
5.25711238e-01 8.73019397e-01 2.44771957e-01 -1.47993159e+00
-1.15331638e+00 -8.22687685e-01 -2.05048084e-01 3.27030092e-01
1.74579889e-01 -1.04612529e+00 1.49386991e-02 -2.03516241e-02] | [4.490677833557129, 0.8815512657165527] |
c6a6847d-3e68-4f85-abd1-06b71f8c0396 | reinterpreting-causal-discovery-as-the-task | 2305.06894 | null | https://arxiv.org/abs/2305.06894v1 | https://arxiv.org/pdf/2305.06894v1.pdf | Reinterpreting causal discovery as the task of predicting unobserved joint statistics | If $X,Y,Z$ denote sets of random variables, two different data sources may contain samples from $P_{X,Y}$ and $P_{Y,Z}$, respectively. We argue that causal discovery can help inferring properties of the `unobserved joint distributions' $P_{X,Y,Z}$ or $P_{X,Z}$. The properties may be conditional independences (as in `integrative causal inference') or also quantitative statements about dependences. More generally, we define a learning scenario where the input is a subset of variables and the label is some statistical property of that subset. Sets of jointly observed variables define the training points, while unobserved sets are possible test points. To solve this learning task, we infer, as an intermediate step, a causal model from the observations that then entails properties of unobserved sets. Accordingly, we can define the VC dimension of a class of causal models and derive generalization bounds for the predictions. Here, causal discovery becomes more modest and better accessible to empirical tests than usual: rather than trying to find a causal hypothesis that is `true' a causal hypothesis is {\it useful} whenever it correctly predicts statistical properties of unobserved joint distributions. This way, a sparse causal graph that omits weak influences may be more useful than a dense one (despite being less accurate) because it is able to reconstruct the full joint distribution from marginal distributions of smaller subsets. Within such a `pragmatic' application of causal discovery, some popular heuristic approaches become justified in retrospect. It is, for instance, allowed to infer DAGs from partial correlations instead of conditional independences if the DAGs are only used to predict partial correlations. | ['Leena Chennuru Vankadara', 'Philipp M. Faller', 'Dominik Janzing'] | 2023-05-11 | null | null | null | null | ['causal-inference', 'causal-discovery', 'causal-inference'] | ['knowledge-base', 'knowledge-base', 'miscellaneous'] | [ 2.03477532e-01 5.51364779e-01 -5.65577626e-01 -4.98280942e-01
-5.64273894e-01 -6.25065625e-01 7.14266837e-01 2.75230080e-01
1.48372352e-01 1.40861881e+00 2.41035610e-01 -5.54232359e-01
-7.95122206e-01 -1.29702413e+00 -8.80171299e-01 -1.01113951e+00
-6.63467884e-01 7.06105709e-01 1.75272912e-01 2.02713758e-01
3.51341963e-02 2.96504438e-01 -1.62569177e+00 1.11575939e-01
6.68420017e-01 6.72514260e-01 4.44193035e-02 4.68044043e-01
-2.55444676e-01 1.01091719e+00 -3.95301849e-01 -5.18368959e-01
-1.33893251e-01 -7.50199795e-01 -9.57459748e-01 -2.26329267e-01
-5.75134577e-03 1.52125701e-01 1.01095373e-02 1.19449139e+00
-2.19677925e-01 -7.06566125e-02 9.96658802e-01 -1.51438701e+00
-6.39893353e-01 1.22029090e+00 -6.47969067e-01 1.32019594e-01
4.58110541e-01 2.91143567e-03 1.65031672e+00 -4.33170140e-01
5.73112726e-01 1.29837489e+00 3.02347332e-01 1.77437156e-01
-1.73183119e+00 -9.16492939e-01 2.95640498e-01 1.60032377e-01
-1.27474034e+00 -2.23576173e-01 6.18257225e-01 -5.54274142e-01
4.50489283e-01 5.36067307e-01 3.55917573e-01 1.36159372e+00
-6.80648535e-02 6.53024971e-01 1.12126565e+00 -5.36889315e-01
5.23393571e-01 2.14518458e-01 3.36596042e-01 7.21632361e-01
6.74558103e-01 4.19939995e-01 -7.65499473e-01 -4.34480906e-01
7.85444617e-01 -1.94000185e-01 -4.14812893e-01 -3.39147866e-01
-1.03117514e+00 1.04647291e+00 1.30435959e-01 3.04790735e-01
-3.13695371e-01 3.80992830e-01 -5.30907586e-02 3.74569774e-01
2.39064857e-01 5.56503356e-01 -9.56278801e-01 5.94757423e-02
-6.88047647e-01 2.93095231e-01 9.81086731e-01 9.17881608e-01
1.02048922e+00 -3.45581353e-01 5.22912703e-02 2.28101477e-01
3.92981708e-01 3.96322250e-01 -2.07524851e-01 -8.36569130e-01
2.58435965e-01 5.02807796e-01 3.29763502e-01 -7.59784281e-01
-2.56500334e-01 -8.78056809e-02 -1.00782621e+00 4.56350967e-02
9.74041343e-01 -2.09614843e-01 -7.70161808e-01 2.21681523e+00
1.24650553e-01 3.40460598e-01 -1.64455160e-01 7.68774092e-01
4.21384037e-01 4.70767766e-01 2.38706470e-01 -7.39666522e-01
1.06619132e+00 3.76878045e-02 -5.21606386e-01 1.88912228e-02
4.02891219e-01 -4.35212582e-01 7.83831954e-01 4.49227184e-01
-8.45112503e-01 -2.19331235e-01 -5.70856631e-01 3.60165328e-01
-3.05250794e-01 -2.57186353e-01 1.10615838e+00 5.27063668e-01
-7.69082904e-01 7.81556249e-01 -6.69135809e-01 -1.41210526e-01
3.53709996e-01 3.54092389e-01 -3.82999927e-01 -8.71806890e-02
-1.36219215e+00 5.11172175e-01 2.84928232e-01 -1.80587664e-01
-9.53475654e-01 -6.97509170e-01 -4.96140271e-01 3.53662014e-01
6.81970775e-01 -6.49314523e-01 7.11258948e-01 -8.21381867e-01
-7.08413661e-01 6.71154082e-01 -3.44673783e-01 -1.40352666e-01
3.32290769e-01 1.37079999e-01 -3.65725249e-01 -1.33233055e-01
4.18204010e-01 1.91036910e-01 6.35936439e-01 -1.37780809e+00
-7.71155894e-01 -4.76222068e-01 2.36244887e-01 -4.44997996e-01
2.14588001e-01 -1.36106417e-01 -1.13884807e-01 -9.44624245e-02
3.93355757e-01 -5.92927694e-01 -2.76089191e-01 -3.05831283e-01
-1.21067572e+00 -3.38722020e-01 3.56579006e-01 -8.68974999e-02
1.04446256e+00 -2.03662753e+00 2.28375018e-01 7.21127629e-01
4.81321484e-01 -5.21064878e-01 2.72826016e-01 3.98928165e-01
-5.76291859e-01 5.97892046e-01 -4.22406554e-01 9.30617005e-03
-1.89480316e-02 4.49432909e-01 -4.76960361e-01 5.92806399e-01
3.77431214e-01 5.20318151e-01 -9.15241599e-01 -3.94818187e-01
1.11190952e-01 5.67552373e-02 -5.29434741e-01 1.59938753e-01
-6.40022039e-01 7.58766294e-01 -6.79932714e-01 3.53849113e-01
4.41334397e-01 -5.97258627e-01 4.94609565e-01 1.62410215e-01
-1.94541648e-01 5.20179868e-01 -1.37195468e+00 1.05681288e+00
-8.73235017e-02 3.72768074e-01 -2.27172941e-01 -1.55979156e+00
8.11189055e-01 4.19268250e-01 5.75165212e-01 -2.58066058e-01
1.19118311e-01 5.09842113e-02 1.03391491e-01 -6.29754305e-01
-1.04216121e-01 -6.08435392e-01 -2.15578482e-01 4.38597411e-01
1.02640994e-01 1.89124554e-01 1.40780330e-01 1.93452165e-01
1.42305911e+00 6.37128577e-03 4.77495432e-01 -3.88836145e-01
2.16342941e-01 1.36652023e-01 6.69408202e-01 1.03305197e+00
2.71706671e-01 3.50216389e-01 1.38046741e+00 -1.98033094e-01
-9.08904672e-01 -1.60791099e+00 -3.14366817e-01 8.84267211e-01
1.38680369e-01 -3.98343474e-01 -1.12624593e-01 -7.72260308e-01
2.93037239e-02 1.05431831e+00 -1.01119280e+00 2.67607998e-02
-2.97272861e-01 -9.87193167e-01 2.12706327e-01 2.73684412e-01
5.30899167e-02 -8.50578725e-01 -2.90065527e-01 -4.47507650e-02
-1.83255255e-01 -5.35126925e-01 3.42688560e-01 7.59433985e-01
-7.90138602e-01 -1.53797841e+00 -2.07483042e-02 -1.34223521e-01
7.01214015e-01 -2.00519398e-01 1.29668987e+00 -3.63359600e-02
-8.96784440e-02 -6.66467622e-02 -3.01324069e-01 -2.97964156e-01
-2.29326323e-01 -2.91731387e-01 1.12327352e-01 -9.12543163e-02
5.06498992e-01 -1.12188339e+00 -2.61532009e-01 8.26795846e-02
-6.42958522e-01 -2.00688481e-01 6.54966831e-01 7.54442930e-01
5.06785572e-01 5.59776902e-01 4.73267406e-01 -1.37431514e+00
1.60216585e-01 -9.86783504e-01 -5.23307860e-01 2.31632397e-01
-5.15791476e-01 4.33771938e-01 7.50906229e-01 -4.17519033e-01
-1.05889177e+00 -1.79112166e-01 1.91784427e-01 -2.58738846e-01
-4.43113029e-01 8.21541488e-01 -5.33464134e-01 9.34980690e-01
8.11487556e-01 -1.51073381e-01 -3.97650719e-01 -6.22564793e-01
3.98394316e-01 1.29796475e-01 3.76859099e-01 -9.79113638e-01
8.06433439e-01 5.40755749e-01 3.41480821e-01 -3.65439028e-01
-9.05532897e-01 -1.70915559e-01 -7.16265023e-01 3.46640646e-02
7.45030642e-01 -5.98647058e-01 -1.04701567e+00 -5.19679248e-01
-1.02718031e+00 -1.75993323e-01 -6.24297082e-01 7.76849866e-01
-5.18569410e-01 -3.50699395e-01 -6.27219751e-02 -1.10263407e+00
6.54100418e-01 -9.27954972e-01 8.14474404e-01 -2.90386262e-03
-6.41536295e-01 -1.23712564e+00 -1.16370860e-02 -5.25564142e-02
-2.66176462e-01 2.15451553e-01 1.60076690e+00 -7.50250280e-01
-8.52066696e-01 -2.50909328e-01 -4.46839094e-01 -1.65340230e-01
2.70756096e-01 2.80524343e-01 -8.47572148e-01 8.10730606e-02
-2.43980646e-01 -1.93455070e-02 8.25141668e-01 6.68269157e-01
1.28266728e+00 -6.06496036e-01 -8.60063374e-01 1.06587172e-01
1.42617929e+00 9.30568799e-02 6.12508893e-01 -2.27351069e-01
4.96907949e-01 8.95290136e-01 3.53028059e-01 4.47817504e-01
1.04700424e-01 3.23316753e-01 5.13489544e-01 1.11042492e-01
2.96390206e-01 -5.04444242e-01 2.05495521e-01 1.85084403e-01
-4.09369260e-01 -2.82432020e-01 -8.17943215e-01 5.32326102e-01
-1.64538610e+00 -1.32676804e+00 -6.37074649e-01 2.44921160e+00
1.08782279e+00 1.43877804e-01 1.14429906e-01 1.57046780e-01
7.88434386e-01 -9.00024772e-02 -4.36259151e-01 1.27756959e-02
-1.31934196e-01 5.36320210e-01 3.90325934e-01 6.66737378e-01
-7.00843453e-01 5.46345651e-01 6.25398302e+00 6.41438603e-01
-8.16406488e-01 8.84811282e-02 7.40194380e-01 -1.07895792e-01
-9.94897485e-01 6.18934631e-01 -7.54063070e-01 6.43781722e-01
8.25637400e-01 -1.60907522e-01 2.44963199e-01 6.44044578e-01
3.60263318e-01 -4.39653069e-01 -1.60779691e+00 4.95750040e-01
-6.08159840e-01 -1.30295336e+00 -1.15064517e-01 4.53850091e-01
6.49760127e-01 -4.45061952e-01 -9.54984222e-03 1.12686090e-01
1.27238619e+00 -1.36258721e+00 6.06554270e-01 6.77419007e-01
8.79667521e-01 -7.03922153e-01 5.25836825e-01 6.94723010e-01
-8.08390617e-01 -8.91847834e-02 -2.95197845e-01 -5.61183095e-01
2.09136307e-02 1.26694107e+00 -6.28751755e-01 5.04101038e-01
6.87551081e-01 4.80512738e-01 -8.70185569e-02 5.54266095e-01
-6.59254253e-01 1.01552165e+00 -2.82233298e-01 -3.45571376e-02
-1.39017805e-01 -2.96042264e-01 4.51261133e-01 9.16888237e-01
2.88903505e-01 4.69104826e-01 -7.63627887e-02 1.38554049e+00
2.73174215e-02 -1.58901110e-01 -8.20011377e-01 5.78186475e-02
4.64549839e-01 8.79620194e-01 -8.28026831e-01 -2.54634798e-01
-3.26208979e-01 3.39296758e-01 4.78180647e-01 6.48997724e-01
-9.28737879e-01 3.19530606e-01 7.89228618e-01 1.67612582e-01
1.09898604e-01 1.58062316e-02 -5.21022320e-01 -1.19305313e+00
-2.39087507e-01 -4.34112251e-01 4.87835705e-01 -6.31044507e-01
-1.52763808e+00 -2.37998972e-03 2.72956043e-01 -9.35671151e-01
-3.96610796e-01 -4.24029619e-01 -5.85242867e-01 1.13853073e+00
-1.01312804e+00 -7.00398207e-01 3.29210788e-01 6.07717812e-01
2.14392617e-02 1.03289403e-01 9.17486250e-01 -1.66109428e-01
-4.51837778e-01 1.49112046e-01 -4.96889316e-02 4.49911840e-02
3.15927744e-01 -1.46126723e+00 -2.99016684e-01 7.27242768e-01
3.14255357e-01 8.47759843e-01 1.18364966e+00 -7.69920349e-01
-1.13133061e+00 -8.81603241e-01 1.14243829e+00 -5.84550977e-01
9.45012152e-01 -3.41221273e-01 -8.48834276e-01 1.06777787e+00
-1.08495302e-01 -6.91739246e-02 8.16442311e-01 9.55791712e-01
-4.58836347e-01 6.40407950e-02 -9.73338008e-01 5.13041794e-01
1.13202310e+00 -3.03122789e-01 -4.89030123e-01 3.46916080e-01
5.78158081e-01 2.62788415e-01 -9.07680035e-01 2.67642677e-01
4.31211799e-01 -1.30637765e+00 8.67601633e-01 -9.82561171e-01
8.98325622e-01 -1.73021883e-01 -3.89269680e-01 -1.26173675e+00
-3.48557651e-01 -2.57784873e-01 1.30916864e-01 1.41128993e+00
8.55921686e-01 -4.50371653e-01 8.37420821e-01 8.16711962e-01
3.27407718e-01 -4.65832353e-01 -1.06916571e+00 -5.45695961e-01
1.80842161e-01 -8.79725933e-01 5.92546463e-01 1.37326992e+00
2.40075514e-01 5.41588008e-01 -2.50028640e-01 5.18512785e-01
8.09260726e-01 5.44266224e-01 5.18765867e-01 -1.66613925e+00
-7.47894347e-01 -3.36145520e-01 -1.67433336e-01 -7.49711514e-01
2.65756339e-01 -8.23305130e-01 2.79789767e-03 -1.22083509e+00
4.98884916e-01 -1.04717124e+00 -1.21986367e-01 6.24787092e-01
-1.16247579e-01 -2.75104314e-01 -1.89889699e-01 2.89462537e-01
-1.49785399e-01 9.66158137e-02 1.25156653e+00 -6.94226995e-02
-1.89628407e-01 2.95930415e-01 -8.89189899e-01 8.54560494e-01
5.30253589e-01 -6.17579103e-01 -5.32483637e-01 5.76441213e-02
6.45942748e-01 7.88073719e-01 7.22377956e-01 -2.24210709e-01
1.22562736e-01 -6.65893316e-01 4.28167790e-01 -3.18421900e-01
8.25756341e-02 -6.47431314e-01 5.27352095e-01 2.20116600e-01
-5.81407964e-01 -5.89109480e-01 -2.97199488e-01 6.15294337e-01
-7.35582709e-02 -1.89209715e-01 4.82886225e-01 -3.22554618e-01
-2.92037517e-01 1.56894431e-01 -2.46986762e-01 -1.61523342e-01
7.77651489e-01 5.12547791e-02 -2.58232683e-01 -3.84026796e-01
-1.09764743e+00 -7.33022438e-03 1.34953737e-01 -9.35264826e-02
4.91277069e-01 -1.00542057e+00 -6.55553639e-01 8.45868886e-03
-1.13998950e-02 2.42219210e-01 -1.79316159e-02 1.00448775e+00
3.10061604e-01 3.10410470e-01 1.40127480e-01 -4.94038433e-01
-7.98516154e-01 7.18866110e-01 5.08388784e-03 -2.57958561e-01
-4.01143461e-01 9.81667221e-01 6.12933755e-01 -1.49397284e-01
-1.72490314e-01 -2.97835797e-01 -2.01378971e-01 2.20427185e-01
2.94552445e-01 2.62979835e-01 -4.42460567e-01 -2.01286554e-01
-3.43760371e-01 -5.03167091e-03 2.01595455e-01 -1.94022804e-01
1.37469184e+00 -1.80429459e-01 -4.82078791e-01 9.25475717e-01
1.09744644e+00 2.93391168e-01 -1.04523039e+00 -6.22101203e-02
1.50318310e-01 -6.00812316e-01 -3.66107851e-01 -9.20975208e-01
-1.07397783e+00 7.81327903e-01 -1.09330360e-02 8.32685232e-01
9.65274513e-01 7.94772029e-01 -2.02540323e-01 -2.04590216e-01
5.72740734e-01 -3.81844282e-01 -2.65915573e-01 1.38926730e-01
6.43466234e-01 -1.20412827e+00 -5.92995249e-02 -7.32830763e-01
-1.51290849e-01 9.46502268e-01 2.00706527e-01 1.49011817e-02
6.89686954e-01 4.02565449e-01 -4.93964434e-01 -5.95581532e-01
-1.07860124e+00 -3.64557445e-01 8.42585638e-02 4.97477233e-01
5.74231327e-01 4.64549839e-01 -1.95966870e-01 7.23042488e-01
-5.21273792e-01 -1.56367168e-01 5.06608069e-01 2.64976829e-01
-3.05497438e-01 -1.05664361e+00 -4.88350719e-01 8.59668493e-01
-3.59494984e-01 -1.54393628e-01 -3.36768866e-01 1.07710874e+00
5.04065335e-01 1.00646687e+00 2.18334094e-01 -3.43060583e-01
3.51708755e-02 9.82624292e-03 4.87924844e-01 -7.74576426e-01
2.11196542e-01 1.45439342e-01 2.18457177e-01 -4.50166076e-01
-6.07755899e-01 -1.05416560e+00 -1.27215528e+00 -6.23964965e-01
-3.11871886e-01 3.91988188e-01 2.81866252e-01 1.38671339e+00
-2.60327637e-01 4.39762175e-01 5.20528853e-01 -1.75838172e-01
-1.66564494e-01 -7.61298478e-01 -1.21834528e+00 3.79102111e-01
2.41436958e-01 -9.84523475e-01 -6.39069259e-01 2.66687632e-01] | [7.874304294586182, 5.359012603759766] |
071081d8-3002-45d7-8062-4d04f2325d2d | the-bdcam-oes-collection-of-portuguese | null | null | https://aclanthology.org/2020.lrec-1.106 | https://aclanthology.org/2020.lrec-1.106.pdf | The BDCam\~oes Collection of Portuguese Literary Documents: a Research Resource for Digital Humanities and Language Technology | This paper presents the BDCam{\~o}es Collection of Portuguese Literary Documents, a new corpus of literary texts written in Portuguese that in its inaugural version includes close to 4 million words from over 200 complete documents from 83 authors in 14 genres, covering a time span from the 16th to the 21st century, and adhering to different orthographic conventions. Many of the texts in the corpus have also been automatically parsed with state-of-the-art language processing tools, forming the BDCam{\~o}es Treebank subcorpus. This set of characteristics makes of BDCam{\~o}es an invaluable resource for research in language technology (e.g. authorship detection, genre classification, etc.) and in language science and digital humanities (e.g. comparative literature, diachronic linguistics, etc.). | ["Ant{\\'o}nio Branco", 'Jo{\\~a}o Silva', "M{\\'a}rcia Bolrinha", 'Sara Grilo', 'Rui Vaz'] | 2020-05-01 | null | null | null | lrec-2020-5 | ['genre-classification'] | ['computer-vision'] | [-1.41734064e-01 -1.10372998e-01 -2.79195577e-01 2.23857015e-01
-4.36144710e-01 -1.05593598e+00 9.70352113e-01 3.72191191e-01
-8.09863746e-01 1.01945758e+00 3.71518373e-01 -5.81537664e-01
-2.88248360e-01 -5.28726161e-01 -1.45232275e-01 -2.63702363e-01
4.00693178e-01 7.68298566e-01 2.76416123e-01 -6.11190379e-01
9.17836547e-01 5.84578514e-01 -1.47045016e+00 -2.03384478e-02
6.28602505e-01 4.16324496e-01 5.35730243e-01 8.52356017e-01
-8.00981700e-01 3.98186535e-01 -1.02042818e+00 -1.15101099e+00
-2.39214540e-01 -3.03241551e-01 -1.12412155e+00 -2.92951047e-01
1.95812330e-01 1.82198122e-01 -4.23159689e-01 1.20547330e+00
3.02923769e-01 -6.16550967e-02 6.06406212e-01 -6.07965291e-01
-5.49967110e-01 1.03878427e+00 -2.63686389e-01 7.36581862e-01
4.86194074e-01 -3.67099136e-01 1.10634553e+00 -9.81162310e-01
1.27223647e+00 1.32714081e+00 4.28963751e-01 3.31832379e-01
-8.66650462e-01 -5.43664753e-01 -4.40665901e-01 8.66729766e-02
-1.49537277e+00 -3.45202863e-01 7.43078232e-01 -6.77470028e-01
8.69827271e-01 1.90901995e-01 7.52711952e-01 1.10580742e+00
2.55809873e-01 5.81767201e-01 1.15864861e+00 -9.86892462e-01
7.31752440e-02 3.04285586e-01 2.16126516e-01 3.85704517e-01
4.15992856e-01 -6.67617023e-01 -6.56595767e-01 -1.46345809e-01
7.09155321e-01 -4.73507196e-01 4.76753190e-02 6.39148593e-01
-1.23567986e+00 6.84326947e-01 -6.74175024e-01 1.17861891e+00
2.69274898e-02 -2.77952403e-01 1.08430469e+00 3.47451001e-01
1.51642159e-01 5.45897961e-01 -4.23875093e-01 -7.15726554e-01
-8.47924829e-01 1.06775336e-01 9.11905944e-01 1.24124002e+00
1.96456298e-01 -1.57591388e-01 4.00821030e-01 1.14691448e+00
3.11179727e-01 4.94122803e-01 7.86990762e-01 -7.97417760e-01
8.17215562e-01 5.35703361e-01 -1.14711253e-02 -8.96135509e-01
-2.16940299e-01 -3.67781259e-02 -4.56770152e-01 -4.90805864e-01
5.66100717e-01 -6.38157800e-02 -2.37361670e-01 1.22993898e+00
7.29056522e-02 -8.42876971e-01 -3.20293196e-02 4.11706865e-01
1.11328614e+00 8.09452951e-01 1.99067429e-01 -6.10186875e-01
1.74928153e+00 -4.60268676e-01 -9.31817353e-01 1.69217363e-01
2.97117442e-01 -1.43923128e+00 1.17808199e+00 6.83344066e-01
-1.39821565e+00 -3.57942164e-01 -1.10229254e+00 -4.31028068e-01
-6.98075175e-01 2.07902491e-01 2.25589558e-01 8.46739709e-01
-6.99222445e-01 4.78489310e-01 -5.73860109e-01 -6.13954902e-01
2.90332526e-01 -1.66265666e-02 -1.25860944e-01 4.13657665e-01
-9.43377256e-01 1.00042069e+00 6.24094784e-01 -2.45443031e-01
-1.64130032e-01 -3.04079354e-01 -3.72752249e-01 -2.57511616e-01
5.36894321e-01 -1.08710781e-01 1.16995597e+00 -7.27623343e-01
-1.56403804e+00 1.51496613e+00 -1.48256756e-02 -7.66375437e-02
5.41562438e-01 -3.40957075e-01 -8.84171903e-01 3.29154134e-01
3.19377214e-01 -3.10871825e-02 4.14277226e-01 -7.43651330e-01
-8.42646241e-01 -3.95344883e-01 -2.17913672e-01 1.74422320e-02
-5.98790944e-01 9.72800553e-01 -5.73637962e-01 -1.31796467e+00
-4.64651659e-02 -5.57604134e-01 3.01656365e-01 -3.82605970e-01
-2.77597368e-01 -6.14958107e-01 6.03635073e-01 -9.90172565e-01
1.77988636e+00 -1.89050198e+00 4.10775274e-01 1.69032708e-01
2.67408848e-01 3.45744908e-01 4.06217754e-01 9.88064110e-01
4.42888051e-01 4.77356881e-01 -5.61386757e-02 -2.56862372e-01
8.93370509e-02 6.69044435e-01 2.61852751e-03 4.42441702e-01
-2.79470623e-01 5.77861965e-01 -9.88859594e-01 -8.91985416e-01
6.13103509e-02 1.53049961e-01 2.38300428e-01 -3.63741547e-01
-1.84520856e-02 7.52409399e-02 -6.40939891e-01 5.58698177e-01
7.30984509e-02 2.58576632e-01 3.49162012e-01 4.72357780e-01
-8.44581723e-01 7.22256541e-01 -9.18273747e-01 1.79518926e+00
-4.94266242e-01 1.04855025e+00 5.39610051e-02 -6.63576007e-01
8.81468475e-01 4.95503306e-01 3.43147486e-01 -5.37818849e-01
7.53097117e-01 7.08995223e-01 1.68775484e-01 -5.06495774e-01
1.11970079e+00 -8.26555863e-03 -3.16073477e-01 3.86699557e-01
3.04249853e-01 -1.91342056e-01 1.08218813e+00 1.82301179e-01
6.76816404e-01 2.32784525e-02 8.16231728e-01 -8.42781782e-01
8.63854825e-01 3.16473782e-01 3.58871788e-01 2.66311258e-01
4.16353457e-02 3.34228545e-01 4.27702516e-01 -3.80233765e-01
-1.58110225e+00 -8.34955454e-01 -6.98042393e-01 9.97592986e-01
-2.47441441e-01 -8.05406094e-01 -8.13600302e-01 -1.17169902e-01
-3.79261911e-01 5.80051243e-01 -2.87970781e-01 4.10097241e-01
-9.74542260e-01 -3.65050554e-01 1.11840510e+00 2.03458622e-01
4.35533136e-01 -1.43576980e+00 -9.56455290e-01 4.52761561e-01
-1.26242086e-01 -1.09434748e+00 -2.42545962e-01 7.35790431e-02
-4.66979802e-01 -1.01079071e+00 -6.73864365e-01 -8.99305642e-01
7.73003250e-02 -3.77731830e-01 1.29466093e+00 1.86303452e-01
-3.49882841e-01 9.54507887e-02 -7.32096791e-01 -7.07240641e-01
-9.91877615e-01 4.12880838e-01 3.52059752e-02 -5.98548472e-01
4.76959348e-01 -3.88963819e-01 3.45374376e-01 -9.48980972e-02
-6.96061313e-01 -3.41205478e-01 2.29460567e-01 6.46236718e-01
4.84536707e-01 -7.81559721e-02 4.55561429e-01 -9.44708407e-01
8.16037834e-01 -2.81779498e-01 -6.63586915e-01 2.17128903e-01
-3.02082002e-01 -2.43502155e-01 9.91687596e-01 -3.59572560e-01
-9.76998329e-01 -7.12771237e-01 -5.89070201e-01 2.65327126e-01
-6.09799698e-02 5.66009998e-01 -3.38221043e-01 4.54154640e-01
5.26131332e-01 1.56633809e-01 -3.61335635e-01 -8.84629667e-01
2.61464208e-01 1.05090368e+00 1.05060852e+00 -7.05104828e-01
5.74977756e-01 2.22049002e-02 5.07570198e-03 -1.44745362e+00
-4.86189455e-01 -7.21396625e-01 -1.05844522e+00 -3.11336845e-01
9.17052388e-01 -4.00058717e-01 -6.09930813e-01 2.82996505e-01
-1.28532255e+00 1.81202680e-01 -3.83375138e-01 3.59414786e-01
-4.06354487e-01 3.97244692e-01 -8.90876889e-01 -7.87532508e-01
-3.49134594e-01 -8.08765471e-01 8.73057306e-01 2.76701629e-01
-7.34458447e-01 -1.19827759e+00 2.18961641e-01 9.29610431e-02
-2.55173296e-01 2.80302584e-01 1.09051979e+00 -5.44597387e-01
3.55120242e-01 2.00577304e-02 1.30920798e-01 1.14756413e-01
-1.36864437e-02 4.51540709e-01 -4.84861583e-01 8.65329579e-02
1.04407951e-01 3.37006897e-02 3.42719495e-01 -1.27212167e-01
5.48090219e-01 -3.38413239e-01 -6.59071282e-02 1.04819290e-01
1.29496050e+00 4.53727931e-01 7.61991560e-01 7.25503743e-01
4.12859201e-01 5.76880038e-01 4.23660368e-01 5.31444430e-01
2.37621829e-01 5.59433758e-01 -2.13731691e-01 7.25889087e-01
-2.36020118e-01 -9.14341956e-02 4.07017440e-01 1.62061524e+00
-3.50153297e-01 -2.66333312e-01 -1.13527429e+00 8.29989195e-01
-1.43190396e+00 -9.77655888e-01 -6.31600559e-01 2.01414466e+00
9.27534819e-01 2.94001400e-01 3.10511410e-01 5.56110799e-01
7.59812474e-01 4.25517447e-02 1.71535954e-01 -1.03258419e+00
-4.03637975e-01 6.36872590e-01 4.34126109e-01 2.76455909e-01
-7.10464537e-01 1.13180506e+00 5.96274614e+00 1.16872764e+00
-9.23665106e-01 1.31218076e-01 -1.40555203e-01 1.92293033e-01
-2.19632015e-01 -9.13712308e-02 -8.75514269e-01 6.76602125e-01
9.86751080e-01 -5.98047197e-01 2.47677833e-01 5.84044218e-01
1.18405350e-01 -2.96789050e-01 -5.74101806e-01 1.03095698e+00
2.29009792e-01 -1.54246366e+00 -3.05633973e-02 2.79171765e-01
6.56481147e-01 1.00196183e-01 -3.70293051e-01 -5.95062375e-02
5.27788028e-02 -8.03079844e-01 1.40310192e+00 2.54483163e-01
1.19720757e+00 -7.78172970e-01 7.49325097e-01 3.89908522e-01
-1.16125560e+00 2.13016883e-01 -4.53009158e-01 -1.25627548e-01
3.31379086e-01 3.47052187e-01 -3.21567267e-01 7.12027729e-01
7.20102191e-01 7.01502323e-01 -6.32099390e-01 6.84952199e-01
-3.76236945e-01 9.27376688e-01 -2.83321351e-01 -8.15515161e-01
3.51481110e-01 -4.67419714e-01 9.75018740e-01 1.55681634e+00
1.25826582e-01 1.86068252e-01 -1.13218464e-01 3.87800902e-01
-3.34941328e-01 7.27339447e-01 -2.61139542e-01 -4.29069817e-01
7.91644216e-01 1.01304269e+00 -1.41933012e+00 -4.73075062e-01
-2.37322181e-01 6.87605560e-01 7.91048855e-02 -2.24788725e-01
-3.05450976e-01 -8.19905221e-01 1.35992587e-01 3.34861189e-01
2.10026026e-01 -4.83421028e-01 -6.02490604e-01 -9.56380427e-01
1.08415477e-01 -7.27191150e-01 4.50893253e-01 -4.49923575e-01
-1.37786078e+00 6.66700482e-01 1.08576067e-01 -8.58491063e-01
-2.59735107e-01 -9.97337997e-01 -3.41378987e-01 7.23238111e-01
-6.85775995e-01 -8.83465767e-01 4.19602454e-01 3.67527187e-01
6.43008888e-01 -5.60238123e-01 9.19912696e-01 4.64889765e-01
-3.67352307e-01 2.67679870e-01 6.35698199e-01 3.42608333e-01
4.81123626e-01 -1.38019323e+00 1.79523453e-01 6.56896770e-01
2.60006845e-01 6.48446262e-01 7.73089528e-01 -6.37010813e-01
-1.34143412e+00 -4.98626798e-01 1.68845999e+00 -4.47258234e-01
1.28397894e+00 -3.98905814e-01 -6.01561427e-01 3.80312830e-01
5.96182108e-01 -7.29161322e-01 6.75264776e-01 -1.42202750e-01
-1.08582102e-01 2.46750191e-01 -1.00740159e+00 9.38729048e-01
1.14539599e+00 -5.51112235e-01 -1.22953057e+00 2.26010650e-01
3.63839686e-01 -2.18525395e-01 -1.27710605e+00 -3.53918284e-01
7.55455375e-01 -5.57859063e-01 3.22006047e-01 -3.65310133e-01
6.18454337e-01 -3.34732197e-02 -5.43865468e-03 -8.05441737e-01
3.41399498e-02 -1.17530274e+00 3.80187899e-01 1.49343061e+00
4.39059764e-01 -5.37919641e-01 2.47292846e-01 -1.56933144e-01
-6.40116036e-01 -2.97602564e-01 -1.44363964e+00 -6.21336460e-01
4.29851383e-01 -6.40371263e-01 3.65046203e-01 1.02912605e+00
3.66739273e-01 4.81237411e-01 4.27634716e-02 -7.41237402e-01
2.09728763e-01 -2.46077478e-01 3.94134849e-01 -1.56819570e+00
-4.20775265e-02 -1.14213717e+00 -4.00780559e-01 -7.95311332e-01
1.25742584e-01 -7.61853635e-01 -2.56297350e-01 -1.46390128e+00
-1.01144299e-01 -2.86729336e-01 4.93999034e-01 1.43159121e-01
2.72899151e-01 4.29303318e-01 1.85942546e-01 5.32683671e-01
-3.25545073e-01 1.16200238e-01 1.12280321e+00 2.06629440e-01
-1.85040116e-01 -4.48252290e-01 -4.11712736e-01 1.01631880e+00
7.21045375e-01 -5.97214818e-01 1.24651283e-01 -3.46405268e-01
7.15913057e-01 -8.73336270e-02 -2.73357928e-01 -4.45094377e-01
1.55034214e-01 -9.77933779e-02 2.44026501e-02 -6.59140289e-01
6.35501444e-02 -4.91620153e-01 6.45582005e-02 4.99256700e-01
8.94700587e-02 7.62557566e-01 3.31716716e-01 -2.33262539e-01
-3.27307940e-01 -9.26867723e-01 5.61994612e-01 -3.03392947e-01
-5.93936324e-01 -2.98009932e-01 -1.08063245e+00 5.49663663e-01
8.49955857e-01 -3.93701196e-01 -3.55007112e-01 3.08393091e-01
-3.86092395e-01 -1.18880026e-01 5.87220013e-01 3.93293649e-01
1.65847346e-01 -1.00914204e+00 -7.22559750e-01 -1.68968111e-01
3.20301317e-02 -2.68859565e-01 -3.68613958e-01 4.62052017e-01
-1.21621752e+00 4.08076882e-01 -4.06099528e-01 -1.36880711e-01
-1.23740435e+00 3.49049151e-01 -2.56574452e-01 -3.47037554e-01
-6.93635464e-01 3.87179792e-01 -7.70935237e-01 -7.36626759e-02
3.19281258e-02 2.44133398e-02 -6.21220469e-01 4.39934999e-01
5.14558375e-01 8.17712784e-01 8.96092430e-02 -1.37653399e+00
-2.25518405e-01 5.98201871e-01 2.09166318e-01 -6.52628541e-01
1.16503966e+00 -3.41410965e-01 -6.56564057e-01 1.04977310e+00
9.49692726e-01 8.04217517e-01 -9.86954048e-02 3.06552231e-01
5.78797579e-01 -3.10692817e-01 -2.42231965e-01 -8.11890841e-01
-3.75232428e-01 7.45274186e-01 -2.40819231e-01 4.32499468e-01
8.66459966e-01 9.21953097e-02 6.77272201e-01 3.37683499e-01
3.75770390e-01 -1.72221076e+00 -4.49399590e-01 9.58327174e-01
7.84588635e-01 -4.82419670e-01 1.60610080e-01 -3.13962936e-01
-4.18744892e-01 1.55818915e+00 -2.25387827e-01 -5.47025390e-02
7.52110660e-01 3.93798918e-01 -3.77803817e-02 -2.76132077e-01
-2.35434979e-01 -8.83777998e-03 2.83602774e-01 3.12920243e-01
7.42077470e-01 9.19581875e-02 -1.68079424e+00 6.80606663e-01
-1.01921988e+00 -3.47475499e-01 8.41664195e-01 1.05133450e+00
-3.54621351e-01 -1.61122119e+00 -6.70986831e-01 1.76352218e-01
-1.04346442e+00 -6.96659163e-02 -7.49725819e-01 1.34742117e+00
4.39387888e-01 8.96901906e-01 2.06548139e-01 5.22309132e-02
2.05954179e-01 4.78044339e-02 7.30294645e-01 -5.04392922e-01
-1.14134133e+00 4.73198146e-01 4.81811762e-01 2.49795526e-01
-6.19630694e-01 -1.02257073e+00 -1.40520906e+00 -6.80963099e-01
-1.60595089e-01 4.99380618e-01 7.73615062e-01 1.10716438e+00
-3.88577253e-01 3.83630276e-01 1.53511852e-01 -6.52004004e-01
-1.67115018e-01 -1.14586258e+00 -7.82657743e-01 6.20537391e-03
-3.61347556e-01 -5.08702517e-01 -1.44877434e-01 2.09768742e-01] | [10.332056045532227, 10.24225902557373] |
7542fb51-84a5-424d-a99a-3198f5be9e11 | deep-learning-based-on-chip-rapid-spectral | 2301.06321 | null | https://arxiv.org/abs/2301.06321v1 | https://arxiv.org/pdf/2301.06321v1.pdf | Deep-learning-based on-chip rapid spectral imaging with high spatial resolution | Spectral imaging extends the concept of traditional color cameras to capture images across multiple spectral channels and has broad application prospects. Conventional spectral cameras based on scanning methods suffer from low acquisition speed and large volume. On-chip computational spectral imaging based on metasurface filters provides a promising scheme for portable applications, but endures long computation time for point-by-point iterative spectral reconstruction and mosaic effect in the reconstructed spectral images. In this study, we demonstrated on-chip rapid spectral imaging eliminating the mosaic effect in the spectral image by deep-learning-based spectral data cube reconstruction. We experimentally achieved four orders of magnitude speed improvement than iterative spectral reconstruction and high fidelity of spectral reconstruction over 99% for a standard color board. In particular, we demonstrated video-rate spectral imaging for moving objects and outdoor driving scenes with good performance for recognizing metamerism, where the concolorous sky and white cars can be distinguished via their spectra, showing great potential for autonomous driving and other practical applications in the field of intelligent perception. | ['Fang Liu', 'Xue Feng', 'Wei zhang', 'Yidong Huang', 'Kaiyu Cui', 'Jiawei Yang'] | 2023-01-16 | null | null | null | null | ['spectral-reconstruction', 'metamerism'] | ['computer-vision', 'computer-vision'] | [ 8.92246366e-01 -6.21162355e-01 3.05890620e-01 -2.66809314e-01
-7.36045122e-01 -4.28789914e-01 1.04670428e-01 -5.47989964e-01
-4.33000892e-01 4.39295441e-01 -4.20239985e-01 -7.06839189e-02
-1.07213147e-01 -7.94146359e-01 -7.41236925e-01 -1.09825146e+00
4.17717844e-01 1.05862387e-01 4.04776275e-01 -1.43780380e-01
2.52743721e-01 8.56778696e-02 -1.95355630e+00 3.00068766e-01
1.07605827e+00 1.04575646e+00 6.62719548e-01 3.37478876e-01
-9.36640427e-02 3.70825887e-01 -2.27306455e-01 1.44880325e-01
4.38382626e-01 -3.31566572e-01 1.06654629e-01 3.26478809e-01
8.09892058e-01 -5.82778156e-01 -4.46463436e-01 1.47287583e+00
1.63707182e-01 -1.28115758e-01 3.26849073e-01 -7.32886791e-01
-7.59852946e-01 2.52819657e-01 -1.02685213e+00 -5.38892746e-01
2.26817071e-01 4.69118893e-01 5.93203068e-01 -9.35305238e-01
1.39153734e-01 6.56582832e-01 6.85127914e-01 3.37336361e-01
-1.23518634e+00 -8.94593894e-01 -5.76802492e-01 3.95726472e-01
-1.34785700e+00 -3.62871617e-01 9.87317502e-01 -3.46321344e-01
7.61234999e-01 1.48000970e-01 9.15107250e-01 5.80017686e-01
4.39708345e-02 9.39572453e-02 1.49742651e+00 -2.29560241e-01
8.33663568e-02 -1.10248052e-01 -1.29276320e-01 8.41292143e-01
3.94840300e-01 5.04903257e-01 -8.20881486e-01 1.85131326e-01
7.87869513e-01 2.23411992e-01 -6.90409064e-01 -1.02979638e-01
-1.34301162e+00 2.34622061e-01 3.36981088e-01 -1.90257937e-01
7.64430389e-02 3.30905467e-01 -1.15268424e-01 1.23489134e-01
2.20549777e-01 4.88571048e-01 -1.13256071e-02 2.26294562e-01
-9.54105496e-01 -9.48210955e-02 3.85849811e-02 4.88877863e-01
9.49259281e-01 4.55231011e-01 6.73658669e-01 9.83562589e-01
2.24115521e-01 1.33054304e+00 2.96695471e-01 -1.36428511e+00
-1.30747065e-01 4.81345206e-01 2.46183708e-01 -2.36856833e-01
-4.32273686e-01 -3.72282982e-01 -7.98244238e-01 6.00940466e-01
2.52836049e-01 -1.87702812e-02 -8.98456156e-01 1.29848015e+00
1.10819824e-01 3.04681569e-01 1.32008120e-02 1.46228695e+00
6.49443686e-01 1.02630675e+00 -6.41787112e-01 -4.11590964e-01
1.17171478e+00 -6.16556406e-01 -4.43259537e-01 -2.96127439e-01
1.56956077e-01 -9.55269575e-01 1.26165652e+00 8.79900575e-01
-1.02846181e+00 -5.90385139e-01 -1.24062538e+00 1.52082462e-03
1.58866495e-01 4.57999170e-01 6.30498469e-01 7.12016404e-01
-7.12126493e-01 3.91783744e-01 -5.67506313e-01 -6.95667276e-03
2.18541026e-01 6.80901110e-02 -1.43040597e-01 -3.49265635e-01
-6.67939425e-01 3.41613024e-01 1.60859555e-01 -1.71627179e-01
-5.57678878e-01 -1.12241066e+00 -2.89178193e-01 -1.64886981e-01
5.70386708e-01 -5.78045070e-01 9.63871717e-01 -1.07654250e+00
-2.06824136e+00 9.72548425e-01 -3.87219042e-02 -3.48244280e-01
7.98108578e-02 -2.00470537e-01 -7.79088557e-01 3.08540046e-01
-1.44837424e-01 1.23469196e-01 1.07202983e+00 -1.01551580e+00
-3.06957871e-01 -4.86057460e-01 -5.14071405e-01 1.04761682e-01
-3.28494906e-01 -1.30675629e-01 -1.76378146e-01 1.47768438e-01
4.24926251e-01 -8.95174801e-01 5.68838641e-02 2.94464588e-01
-1.67827964e-01 4.35896188e-01 1.05561960e+00 -1.24706380e-01
5.13265729e-01 -2.36456800e+00 -2.63449758e-01 -1.90396205e-01
1.28807008e-01 4.07433510e-01 -2.26385921e-01 3.69398743e-01
1.53992191e-01 -5.74466586e-01 -5.51451862e-01 4.60788846e-01
-4.61592942e-01 -4.45703447e-01 -1.50645196e-01 7.57328868e-01
-1.06282704e-01 5.01660228e-01 -5.66364050e-01 -1.96928643e-02
3.58423769e-01 5.20528555e-01 -3.88823241e-01 -5.70275523e-02
-5.63088775e-01 3.03529501e-01 -8.51146802e-02 6.55749559e-01
1.19990695e+00 -3.22788447e-01 3.46941471e-01 -6.72238469e-01
-7.45314538e-01 2.44116932e-02 -9.60741043e-01 1.92499459e+00
-3.90528738e-01 9.30798650e-01 5.75136662e-01 -7.14816630e-01
9.75810766e-01 -2.07248017e-01 6.45043612e-01 -1.20960736e+00
-1.30445108e-01 5.65550089e-01 -1.76573917e-01 -4.75603163e-01
7.08666146e-01 -4.56209481e-01 1.42397564e-02 3.04271042e-01
-2.40919366e-01 -6.30851150e-01 8.09325487e-04 -1.87833264e-01
5.83148062e-01 1.21690415e-01 -1.37456089e-01 -5.26569128e-01
4.05150622e-01 1.08169407e-01 3.05702776e-01 4.61689949e-01
2.69457638e-01 5.24529397e-01 -2.03290045e-01 -3.34747583e-01
-1.32635069e+00 -1.41393876e+00 -1.99218661e-01 7.52200603e-01
7.28801012e-01 -1.97022781e-01 -4.44356799e-01 4.87442404e-01
1.17053166e-01 5.14502406e-01 1.62308201e-01 2.02065352e-02
-3.52055550e-01 -8.75287056e-01 5.15174046e-02 7.96163380e-02
9.22946632e-01 -3.38112563e-01 -1.04428184e+00 6.87958822e-02
4.04790370e-03 -1.15692687e+00 -1.40259534e-01 -2.46264145e-01
-6.52358174e-01 -1.26236534e+00 -3.60260755e-01 -4.98488635e-01
2.77614355e-01 1.25520861e+00 7.86204994e-01 -3.32281083e-01
-9.23340499e-01 4.82057244e-01 1.16189703e-01 -3.91742170e-01
-1.42365247e-01 -4.45867896e-01 8.79129469e-02 2.08159417e-01
9.15258303e-02 -5.66139877e-01 -8.81097496e-01 4.94376838e-01
-6.96617782e-01 7.17079520e-01 6.55602455e-01 7.78037608e-01
6.65424407e-01 8.63937810e-02 3.38863045e-01 -8.71413648e-01
1.24216313e-02 1.37831986e-01 -1.48242426e+00 9.96352881e-02
-5.41502118e-01 -1.59309402e-01 7.42525220e-01 -1.30644381e-01
-1.61180162e+00 2.18000308e-01 3.12530473e-02 -6.30778968e-02
-1.56119633e-02 2.92186558e-01 1.65645584e-01 -4.35111493e-01
1.00275970e+00 4.91709471e-01 3.60931098e-01 -1.72009706e-01
3.02793235e-01 7.01031685e-01 8.46717656e-01 -2.82385916e-01
7.22623467e-01 1.09072447e+00 1.26824856e-01 -1.62749910e+00
-5.87472916e-01 -6.73860312e-01 -1.91771716e-01 -6.76968634e-01
8.63533914e-01 -1.42754006e+00 -9.88431990e-01 6.86355650e-01
-7.86803603e-01 -5.02999187e-01 3.05802785e-02 6.97591901e-01
-3.84755135e-01 4.19227719e-01 -4.48596925e-01 -7.15110183e-01
-1.14548698e-01 -9.50860798e-01 1.12401903e+00 4.18045044e-01
4.91453111e-01 -3.40344131e-01 -2.34343007e-01 6.50830328e-01
4.82697964e-01 -8.98872465e-02 8.35685074e-01 9.69894171e-01
-1.48346412e+00 3.00102755e-02 -6.50809646e-01 1.60420705e-02
-2.87003610e-02 7.92592242e-02 -1.06970382e+00 -1.46259353e-01
1.12021975e-01 -4.49451059e-01 1.32901406e+00 4.27247584e-01
1.20636165e+00 4.63259041e-01 -1.93167597e-01 1.04966819e+00
2.09511018e+00 2.57065684e-01 5.78105211e-01 2.15491354e-02
7.80267954e-01 4.47219759e-01 5.68176448e-01 3.45781684e-01
-2.83075392e-01 6.37456059e-01 5.34143448e-01 -1.57688424e-01
-3.64035636e-01 2.92346776e-01 6.23378098e-01 6.92461610e-01
1.77786071e-02 -1.35515900e-02 -7.08785951e-01 3.56906503e-01
-1.51831377e+00 -1.01321709e+00 -9.89688396e-01 2.32414079e+00
4.99451399e-01 -3.21918726e-01 -1.43463120e-01 -3.05018246e-01
5.68876088e-01 1.92170382e-01 -7.88945317e-01 1.93547517e-01
-4.49726701e-01 3.69856566e-01 9.22092557e-01 3.73695135e-01
-6.38737798e-01 7.71411777e-01 6.03540754e+00 6.51642382e-01
-1.38833511e+00 1.46807641e-01 6.51284605e-02 -5.54760396e-01
-5.65728843e-01 -1.65043429e-01 -5.25006235e-01 3.22369337e-01
7.80926287e-01 -3.01386602e-02 7.20969558e-01 3.77012372e-01
1.51006311e-01 -6.75428569e-01 -4.34659958e-01 1.51335561e+00
-1.12414248e-01 -1.65673673e+00 -1.13890067e-01 6.02906086e-02
9.67964888e-01 4.36157495e-01 4.47103620e-01 -7.09997594e-01
-8.18151981e-02 -6.44726813e-01 6.89005852e-01 5.61657846e-01
1.43457317e+00 -6.86725974e-01 -8.55833665e-02 4.44865286e-01
-1.00879848e+00 -2.59800702e-01 -7.82259583e-01 -3.91834341e-02
5.92378490e-02 9.90249395e-01 -3.25421125e-01 2.86989003e-01
7.70093143e-01 7.13916063e-01 -7.31603801e-02 7.34925985e-01
1.53472111e-01 7.72641301e-01 -3.53536725e-01 4.62202951e-02
5.33491895e-02 -9.62110996e-01 4.10150021e-01 8.42569888e-01
7.77314067e-01 2.18910262e-01 3.26340720e-02 1.28661263e+00
-4.39969786e-02 -3.15078825e-01 -5.73613286e-01 -1.42695680e-01
1.59950942e-01 1.38466501e+00 -4.76304650e-01 7.10723773e-02
-9.29447830e-01 8.94212723e-01 -2.75314480e-01 3.66366982e-01
-8.90212774e-01 -3.59641522e-01 8.34096432e-01 4.41200912e-01
1.23468772e-01 -7.91029155e-01 -3.67747366e-01 -1.22071064e+00
-6.01665154e-02 -2.74629384e-01 -1.37237266e-01 -1.31113291e+00
-8.71621251e-01 -6.48928136e-02 -5.32875419e-01 -1.27016389e+00
5.52616358e-01 -9.19414103e-01 -1.98365316e-01 8.95928681e-01
-1.61065674e+00 -9.70558286e-01 -7.15253711e-01 7.26895452e-01
3.72404903e-01 -1.26983270e-01 6.68087780e-01 1.10683799e-01
-2.04929307e-01 -1.42060310e-01 7.79739439e-01 -4.11450714e-01
9.82523143e-01 -7.79129684e-01 9.34462845e-02 8.71886134e-01
1.38176322e-01 2.12988332e-01 4.85588938e-01 -4.78460044e-01
-2.08565855e+00 -1.12789321e+00 -3.02819282e-01 3.39765131e-01
5.41357756e-01 -2.31144920e-01 -7.67139018e-01 -7.58414268e-02
4.17586684e-01 -2.68810660e-01 7.17632055e-01 -4.33672011e-01
-5.85595191e-01 -6.36678874e-01 -8.49888921e-01 2.94467807e-01
1.21266580e+00 -8.58432770e-01 8.59708041e-02 5.17136395e-01
5.60106635e-01 -2.55733192e-01 -3.36696863e-01 3.84259552e-01
8.25419784e-01 -1.47985113e+00 9.68404412e-01 3.77336830e-01
4.63434637e-01 -6.73149586e-01 -1.23437345e-02 -1.04638958e+00
-5.20162940e-01 -5.26312172e-01 6.03711665e-01 6.91910028e-01
3.39591324e-01 -7.93491960e-01 8.93127084e-01 -3.80657725e-02
-5.21637142e-01 1.27894089e-01 -5.99619985e-01 -7.59797573e-01
-3.42965335e-01 -4.64040637e-01 3.63724500e-01 8.90425503e-01
-4.07635659e-01 3.09527904e-01 -3.64298820e-01 5.22359133e-01
1.33856833e+00 9.02628303e-01 4.56973642e-01 -1.30197251e+00
-3.89642775e-01 -3.35744232e-01 -2.93584839e-02 -8.89422953e-01
7.76535720e-02 -8.95311356e-01 1.09329700e-01 -1.32778919e+00
3.62849742e-01 3.11859902e-02 6.45087585e-02 -2.15080634e-01
2.41924763e-01 6.91420674e-01 1.16417162e-01 2.76916534e-01
-5.39984226e-01 3.12558591e-01 1.19538832e+00 -3.07239324e-01
-1.96802951e-02 -2.96559304e-01 -3.13076317e-01 4.55593348e-01
7.19027221e-01 3.83042805e-02 -6.19008839e-01 -7.30700612e-01
7.07488835e-01 2.80959636e-01 6.64151013e-01 -1.41881979e+00
4.49494570e-01 -4.68366474e-01 3.10735643e-01 -4.12992328e-01
5.90807617e-01 -7.32493520e-01 5.25692225e-01 4.11663979e-01
1.90332532e-02 -8.97620201e-01 9.25412476e-02 7.19977856e-01
-1.79632336e-01 1.91687625e-02 1.19083023e+00 -4.03714836e-01
-1.30751336e+00 1.11602686e-01 -6.39260352e-01 -2.91780055e-01
9.03890610e-01 -4.51367736e-01 -7.98532844e-01 -1.43974572e-01
-6.41685203e-02 -2.10166246e-01 9.70560849e-01 -2.53430784e-01
9.29947972e-01 -1.04016399e+00 -4.76648957e-01 4.86442715e-01
2.73957163e-01 -2.61367470e-01 9.04661119e-01 7.14306653e-01
-9.61765409e-01 2.61963308e-01 -4.08206761e-01 -9.75196064e-01
-1.30504608e+00 1.35336637e-01 5.60703695e-01 7.43465900e-01
-8.30112219e-01 9.27103817e-01 3.21573824e-01 -1.60908520e-01
-3.87044638e-01 -1.12539366e-01 5.97179294e-01 -3.73479635e-01
4.54360425e-01 4.40357000e-01 1.36494100e-01 -3.49720359e-01
-1.34734601e-01 1.00856888e+00 4.81171936e-01 -7.96096101e-02
1.41971374e+00 -1.75112545e-01 -3.25346768e-01 5.76925159e-01
1.04552507e+00 1.89132899e-01 -1.67348921e+00 -3.70758295e-01
-6.05822206e-01 -7.57502377e-01 3.71238679e-01 -7.96423674e-01
-1.08968544e+00 1.10237384e+00 8.10766637e-01 1.52136832e-01
1.39216650e+00 -1.70929492e-01 8.23510826e-01 4.30897117e-01
6.69492781e-01 -1.38014352e+00 9.11533982e-02 2.06898317e-01
2.91791975e-01 -1.22051215e+00 2.59082526e-01 -7.60749459e-01
-4.05772269e-01 1.47745895e+00 3.49340379e-01 2.94700311e-03
3.41150224e-01 2.99317271e-01 5.45590632e-02 -2.31700271e-01
-5.94133496e-01 -1.80107176e-01 1.66352093e-01 7.74337471e-01
3.76980335e-01 4.76518780e-01 -3.46314982e-02 -5.46712242e-02
1.63043737e-02 -1.05743781e-01 8.86131644e-01 2.20029801e-01
-8.87902081e-01 -6.85048103e-01 -2.87062883e-01 4.57455397e-01
7.77436718e-02 -5.26504405e-02 -1.65905684e-01 3.15311521e-01
9.30309370e-02 7.68762708e-01 1.52873084e-01 -4.47055012e-01
9.62429121e-02 1.73453256e-01 9.33680952e-01 -3.96036476e-01
3.71990241e-02 3.89256269e-01 -3.07480115e-02 -8.84637773e-01
-6.20103955e-01 -7.45690107e-01 -1.25666559e+00 -8.55599195e-02
-2.87915647e-01 -3.17812294e-01 9.03414905e-01 4.90965247e-01
3.03152412e-01 1.96509525e-01 5.43495297e-01 -6.27328336e-01
-1.27014592e-01 -3.58107448e-01 -1.10397410e+00 1.59442395e-01
2.89557546e-01 -4.45356876e-01 -1.29859701e-01 1.47624671e-01] | [10.242741584777832, -2.513437509536743] |
13988cae-470c-4e5d-988d-0f14cc86fba0 | the-gem-benchmark-natural-language-generation | 2102.01672 | null | https://arxiv.org/abs/2102.01672v3 | https://arxiv.org/pdf/2102.01672v3.pdf | The GEM Benchmark: Natural Language Generation, its Evaluation and Metrics | We introduce GEM, a living benchmark for natural language Generation (NLG), its Evaluation, and Metrics. Measuring progress in NLG relies on a constantly evolving ecosystem of automated metrics, datasets, and human evaluation standards. Due to this moving target, new models often still evaluate on divergent anglo-centric corpora with well-established, but flawed, metrics. This disconnect makes it challenging to identify the limitations of current models and opportunities for progress. Addressing this limitation, GEM provides an environment in which models can easily be applied to a wide set of tasks and in which evaluation strategies can be tested. Regular updates to the benchmark will help NLG research become more multilingual and evolve the challenge alongside models. This paper serves as the description of the data for which we are organizing a shared task at our ACL 2021 Workshop and to which we invite the entire NLG community to participate. | ['Simon Mille', 'Angelina McMillan-Major', 'Dhruv Kumar', 'Mihir Kale', 'Jiawei Zhou', 'Akhila Yerukola', 'Diyi Yang', 'Wei Xu', 'Nishant Subramani', 'Hendrik Strobelt', 'Marco Antonio Sobrevilla Cabezudo', 'Anastasia Shimorina', 'Samira Shaikh', 'Thibault Sellam', 'João Sedoc', 'Sashank Santhanam', 'Juan Diego Rodriguez', 'Vikas Raunak', 'Niranjan Ramesh Rao', 'Laura Perez-Beltrachini', 'Ankur Parikh', 'Salomey Osei', 'Rubungo Andre Niyongabo', 'Vitaly Nikolaev', 'Shashi Narayan', 'Moin Nadeem', 'Emiel van Miltenburg', 'Pedro Henrique Martins', 'Bodhisattwa Prasad Majumder', 'Saad Mahamood', 'Khyati Mahajan', 'Mounica Maddela', 'Aman Madaan', 'Faisal Ladhak', 'Shailza Jolly', 'Yangfeng Ji', 'Harsh Jhamtani', 'Yacine Jernite', 'Yufang Hou', 'Tatsunori Hashimoto', 'Cristina Garbacea', 'Varun Gangal', 'Chris Emezue', 'Ondřej Dušek', 'Esin Durmus', 'Wanyu Du', 'Kaustubh D. Dhole', 'Dipanjan Das', 'Miruna Clinciu', 'Khyathi Raghavi Chandu', 'Antoine Bosselut', 'Aremu Anuoluwapo', 'Pawan Sasanka Ammanamanchi', 'Karmanya Aggarwal', 'Tosin Adewumi', 'Sebastian Gehrmann'] | 2021-02-02 | null | https://aclanthology.org/2021.gem-1.10 | https://aclanthology.org/2021.gem-1.10.pdf | acl-gem-2021-8 | ['extreme-summarization'] | ['natural-language-processing'] | [-1.33274933e-02 1.59581244e-01 -2.90350974e-01 -3.34921211e-01
-1.21107781e+00 -8.18849027e-01 1.14036465e+00 5.71397282e-02
-5.00768542e-01 1.21351361e+00 7.72106946e-01 -2.56921977e-01
1.47972479e-01 -4.76391971e-01 -3.01534534e-01 -3.86815481e-02
1.58683620e-02 8.83506298e-01 -6.09514788e-02 -5.42899191e-01
1.72590464e-01 1.15367576e-01 -1.45130968e+00 4.24512923e-01
8.26356828e-01 2.73992449e-01 4.88682576e-02 6.85727358e-01
-1.71226382e-01 6.15168035e-01 -9.32924688e-01 -5.90549827e-01
1.52084664e-01 -6.37776196e-01 -1.23241770e+00 -4.27848637e-01
6.53116107e-01 1.00391932e-01 2.32358783e-01 7.94868469e-01
8.71285737e-01 1.44221216e-01 4.60064471e-01 -1.28946757e+00
-6.07418060e-01 1.04098237e+00 9.70131382e-02 2.01676428e-01
7.70460367e-01 2.81850994e-01 1.30751729e+00 -7.77020812e-01
1.35790849e+00 1.35744774e+00 6.60865784e-01 8.00192893e-01
-1.20247066e+00 -4.24564153e-01 6.62063137e-02 -2.03949977e-02
-9.69282866e-01 -7.99030602e-01 2.78238773e-01 -5.40969133e-01
1.13080657e+00 3.76540720e-01 6.50212824e-01 1.57318401e+00
-8.85864422e-02 7.35581279e-01 1.18646193e+00 -4.87046838e-01
-1.08680785e-01 4.53941114e-02 -2.74518933e-02 4.31810707e-01
3.39212388e-01 1.90660492e-01 -7.26202011e-01 -4.77889031e-02
2.59848088e-01 -9.01630878e-01 -3.88276994e-01 -8.99577811e-02
-1.74196303e+00 7.16069818e-01 -1.27442524e-01 7.19244540e-01
6.14922643e-02 9.42208618e-02 4.17675316e-01 5.49816310e-01
4.68373448e-01 9.79789793e-01 -5.03982306e-01 -7.03702927e-01
-9.24353302e-01 9.11243737e-01 1.13207901e+00 7.81661808e-01
4.37938243e-01 -1.39515400e-01 -3.08755994e-01 9.49802756e-01
3.27971250e-01 2.89088339e-01 5.54738283e-01 -1.14268517e+00
6.83091879e-01 4.94569451e-01 2.29543120e-01 -6.69781506e-01
-3.91209036e-01 -4.43100572e-01 -5.24771035e-01 2.10071355e-01
7.48933792e-01 -2.93886304e-01 -5.65964937e-01 1.95893443e+00
-3.61682288e-03 -4.16396499e-01 9.58962291e-02 7.74636745e-01
8.27232838e-01 4.24229890e-01 7.01540187e-02 -3.59672569e-02
9.79075491e-01 -6.85388744e-01 -4.40987796e-01 -1.57599106e-01
1.09130335e+00 -1.19672096e+00 1.34985471e+00 4.89967734e-01
-1.39108682e+00 -5.49190283e-01 -1.13239646e+00 -2.09413543e-01
-5.01925051e-01 -2.46912196e-01 6.99337065e-01 6.18832946e-01
-1.38607764e+00 5.20462573e-01 -7.45406866e-01 -6.85857892e-01
-1.35012493e-01 -1.67998243e-02 -4.04082537e-01 -1.43732697e-01
-1.46695220e+00 1.27075851e+00 5.17949939e-01 -7.99743086e-02
-6.14006937e-01 -6.76411390e-01 -7.30521381e-01 -5.88235080e-01
7.97983408e-02 -7.36409783e-01 1.57862043e+00 -5.79474449e-01
-1.25968754e+00 1.12715304e+00 1.87089920e-01 -3.90172064e-01
8.55859220e-01 -2.27913171e-01 -7.37266600e-01 -5.14277697e-01
6.28738642e-01 8.37810457e-01 -9.76476222e-02 -1.00329590e+00
-7.48818219e-01 -1.17135957e-01 3.59547995e-02 4.52894896e-01
7.64765739e-02 2.49417692e-01 -1.58581674e-01 -5.93942761e-01
-2.53981948e-01 -7.68582523e-01 -1.64108455e-01 -7.30671287e-01
-2.15138659e-01 -4.63961095e-01 3.54330301e-01 -5.38401246e-01
1.36342812e+00 -1.75270832e+00 1.75375432e-01 -3.37941796e-01
1.53190613e-01 1.22760832e-01 -3.36679637e-01 8.67896140e-01
2.62214631e-01 5.90520442e-01 -3.86496224e-02 -4.05723810e-01
3.92199367e-01 -1.76593944e-01 -1.56029314e-01 1.04173504e-01
8.73393044e-02 1.13391864e+00 -1.32544398e+00 -4.96787697e-01
-5.46758063e-03 2.04469994e-01 -3.95070672e-01 -3.85966264e-02
-4.50125128e-01 3.85285348e-01 -2.19953075e-01 4.67694283e-01
1.80231750e-01 -1.46752134e-01 2.11816683e-01 1.78095937e-01
-5.54664612e-01 7.36042976e-01 -1.11237144e+00 2.04778457e+00
-3.40322465e-01 7.71698356e-01 -1.19912244e-01 -4.74313676e-01
9.22204077e-01 4.81402040e-01 9.43551734e-02 -5.88076890e-01
-1.05758518e-01 6.42867565e-01 2.76692152e-01 -1.73181683e-01
7.68513203e-01 -3.40038761e-02 -4.13358569e-01 7.30916619e-01
3.34075481e-01 -4.40011173e-01 8.58689964e-01 3.25231910e-01
1.05217421e+00 3.32614571e-01 2.44760200e-01 -5.37749588e-01
4.14491177e-01 3.52101594e-01 4.50256109e-01 7.87306726e-01
-3.71207297e-01 4.54487175e-01 3.03911269e-01 -5.93495965e-01
-1.09446383e+00 -1.02104640e+00 -1.14022903e-01 1.09011030e+00
-4.48314577e-01 -7.34843671e-01 -6.83942497e-01 -6.18653834e-01
-2.58740425e-01 7.95560718e-01 -5.19881606e-01 1.86937690e-01
-6.16141260e-01 -7.22845912e-01 9.21134353e-01 1.99154913e-01
3.89980286e-01 -1.26340699e+00 -3.97265792e-01 4.22516286e-01
-5.59496760e-01 -9.76142406e-01 -2.46457189e-01 -2.88930565e-01
-5.92515171e-01 -1.04275191e+00 -7.23470151e-01 -8.58742952e-01
-1.24930814e-01 -1.75092742e-01 1.66893423e+00 -1.14005834e-01
4.20845463e-04 2.47783720e-01 -3.93942207e-01 -6.36329889e-01
-9.81657147e-01 6.58293486e-01 -9.89620313e-02 -6.80845141e-01
3.76072973e-01 -2.56150097e-01 -3.70194405e-01 2.49270678e-01
-6.21495485e-01 1.58892497e-01 4.30653512e-01 6.54837966e-01
2.52523541e-01 -6.28998041e-01 1.04643786e+00 -1.02170467e+00
1.14942992e+00 -4.22185034e-01 -4.38897341e-01 4.28075433e-01
-9.28552628e-01 -1.71810426e-02 3.34237367e-01 5.41201904e-02
-1.01393366e+00 -5.21744847e-01 -1.18191607e-01 4.76826012e-01
-1.47495672e-01 7.25655496e-01 -2.47003153e-01 2.85969675e-01
9.91894841e-01 -1.74471274e-01 -2.31788799e-01 -4.92455631e-01
7.21524656e-01 5.82935333e-01 5.61990619e-01 -7.91696429e-01
5.86730063e-01 -4.44474779e-02 -3.57263088e-01 -5.96495569e-01
-7.19543457e-01 -1.77046403e-01 -4.78521824e-01 -2.35789329e-01
7.50079989e-01 -9.19673443e-01 -4.23140498e-03 3.58319283e-01
-1.29797494e+00 -6.80717707e-01 -3.61496806e-01 5.72050214e-01
-4.46416616e-01 2.71620542e-01 -6.35813057e-01 -3.50820780e-01
-4.02926952e-01 -1.04310739e+00 9.26989377e-01 -5.52423075e-02
-8.32659602e-01 -1.36064672e+00 7.14707255e-01 6.85737967e-01
5.17340422e-01 4.15352821e-01 6.65200651e-01 -6.33170307e-01
-3.73288542e-01 -6.88625947e-02 1.50735918e-02 3.63087803e-01
-1.33683547e-01 1.59246415e-01 -7.10494220e-01 -2.92881519e-01
-2.66742498e-01 -7.51089215e-01 6.23736143e-01 -1.21180629e-02
3.97229046e-01 -1.04337789e-01 -1.96513087e-01 3.82178575e-01
1.06589639e+00 -2.63589080e-02 5.38093567e-01 5.84708691e-01
4.19692487e-01 6.93855047e-01 5.09258389e-01 8.21300521e-02
7.96801627e-01 7.17534304e-01 -1.13926016e-01 1.01224124e-01
-3.91834050e-01 -2.61466235e-01 3.98257256e-01 1.29510772e+00
-2.82002598e-01 -6.13542497e-01 -1.36712039e+00 6.01700723e-01
-1.76070762e+00 -1.13655496e+00 -3.58102560e-01 2.28308296e+00
1.07974184e+00 2.00186044e-01 1.59484997e-01 -1.67105585e-01
5.30103803e-01 3.25078517e-01 7.77641162e-02 -6.13316953e-01
-4.36202765e-01 2.23590896e-01 4.08073748e-03 7.49970853e-01
-8.06708455e-01 1.19387770e+00 7.53688955e+00 5.83831310e-01
-9.11795795e-01 1.36571720e-01 5.53638697e-01 -9.40210000e-02
-5.56385159e-01 1.25086531e-01 -8.84154260e-01 3.07769746e-01
1.21870327e+00 -8.20146978e-01 4.54997391e-01 4.09091949e-01
2.06805676e-01 1.98875234e-01 -1.30490696e+00 6.38605595e-01
1.92572787e-01 -1.39557087e+00 -3.68479826e-02 2.21014276e-01
9.45767641e-01 7.11644590e-01 -1.71400905e-01 5.71627498e-01
8.58761609e-01 -1.16774940e+00 7.29645848e-01 3.82166386e-01
9.06916857e-01 -3.58506054e-01 6.32566035e-01 1.17877729e-01
-9.64698374e-01 3.80411685e-01 -1.22962534e-01 -3.13868791e-01
4.65300977e-01 2.41681829e-01 -8.17682981e-01 6.05833113e-01
3.60568762e-01 5.48236609e-01 -6.65770113e-01 1.00006342e+00
-4.22647089e-01 5.86623073e-01 -3.56160589e-02 -1.74255624e-01
3.38132471e-01 -6.76450208e-02 6.97854578e-01 1.49830413e+00
3.20951492e-01 -4.56645012e-01 2.87569910e-01 7.02900827e-01
-1.62601009e-01 4.45248038e-01 -8.28664601e-01 -3.96553934e-01
5.27549088e-01 1.30219710e+00 -4.60311592e-01 -2.61573821e-01
-4.01066422e-01 7.35094368e-01 5.32681942e-01 2.33711496e-01
-5.49245715e-01 -3.59968036e-01 5.68637073e-01 5.96731193e-02
-4.95307177e-01 -3.25622618e-01 -1.12253085e-01 -1.16907215e+00
-8.66118520e-02 -1.36191821e+00 5.23778796e-01 -6.35257542e-01
-1.22608006e+00 8.40389609e-01 2.08870292e-01 -8.93166363e-01
-8.26270282e-01 -4.32968885e-01 -3.98836672e-01 9.84585583e-01
-1.17834401e+00 -1.12535954e+00 -2.40739599e-01 2.20495686e-01
4.00221020e-01 -3.57840568e-01 9.89368796e-01 2.54339933e-01
-2.57377416e-01 6.21918082e-01 -1.78734094e-01 1.05495444e-02
1.12543190e+00 -1.46473706e+00 9.67692792e-01 7.25285351e-01
5.85461855e-01 5.36675036e-01 6.57904744e-01 -7.17462957e-01
-8.47847879e-01 -9.13970590e-01 1.64228082e+00 -9.18297052e-01
9.80998456e-01 -4.87004966e-01 -3.91163766e-01 7.98867345e-01
4.12850946e-01 -5.95369041e-01 7.28973567e-01 4.14180517e-01
-2.59521246e-01 2.21572012e-01 -7.27586746e-01 7.22460926e-01
1.40813792e+00 -4.01329219e-01 -7.12402403e-01 5.09160161e-01
7.27235258e-01 -5.53578794e-01 -9.69931126e-01 3.59937310e-01
6.43213689e-01 -9.71953452e-01 4.34794068e-01 -7.22817838e-01
4.30986851e-01 -1.57100156e-01 -1.68467090e-01 -1.60390568e+00
-1.80910870e-01 -1.04677975e+00 4.30795193e-01 1.50554347e+00
9.79932189e-01 -7.34632969e-01 6.49056375e-01 3.96885723e-01
-5.12521744e-01 -5.36417365e-01 -8.60377133e-01 -8.09324503e-01
4.89754885e-01 -7.07612514e-01 7.68684745e-01 1.11333907e+00
2.06121087e-01 6.75927401e-01 -5.09976409e-02 -5.56656241e-01
5.16796589e-01 -2.10583419e-01 1.03181863e+00 -1.39012527e+00
-2.36429065e-01 -8.23579311e-01 -4.84220743e-01 -5.37031710e-01
-6.74223676e-02 -1.31327581e+00 -1.61662310e-01 -1.75604951e+00
8.47309157e-02 -3.77751201e-01 -1.82270885e-01 3.33719999e-01
-8.78608525e-02 3.46272975e-01 2.74510324e-01 2.01598495e-01
-7.12859869e-01 1.92918047e-01 1.20299518e+00 1.37318661e-02
-1.70055121e-01 -2.93773830e-01 -9.50082302e-01 5.27728319e-01
1.05272388e+00 -1.66114777e-01 -4.58062440e-01 -7.90347934e-01
5.76217055e-01 -3.60233933e-01 7.89090544e-02 -1.26004553e+00
3.25109474e-02 -1.53028741e-01 1.62561670e-01 -2.84224510e-01
7.11834654e-02 8.77183676e-02 4.19251591e-01 1.56974509e-01
-4.81649041e-01 4.55943227e-01 2.08840668e-01 -4.88639288e-02
-3.35537463e-01 -1.55537101e-02 3.21503103e-01 -4.43134934e-01
-5.96712470e-01 2.80795306e-01 -1.61018044e-01 9.32257533e-01
7.17847407e-01 -1.27783880e-01 -7.28401423e-01 -4.08392936e-01
-4.99428421e-01 3.05276245e-01 7.57778943e-01 6.82960272e-01
9.87304747e-02 -1.44347513e+00 -1.38360739e+00 -1.26307085e-01
3.45397651e-01 -3.10727537e-01 -4.24843207e-02 6.83271050e-01
-6.20808721e-01 6.29814029e-01 -1.85138032e-01 -1.87385440e-01
-8.88875246e-01 -1.02350432e-02 2.65088588e-01 -7.45777369e-01
-2.59827763e-01 7.44866848e-01 -1.71752587e-01 -7.08853662e-01
-6.22123480e-02 -1.31067067e-01 -7.71087930e-02 1.98775426e-01
5.04779220e-01 5.08843243e-01 7.90837929e-02 -6.29532099e-01
-1.94925830e-01 5.19454032e-02 -2.88275406e-02 -8.27024102e-01
1.13446164e+00 6.79820850e-02 -2.76396964e-02 8.69668663e-01
8.25749993e-01 2.72856116e-01 -7.81982243e-01 1.52943417e-01
2.85847634e-01 -7.59654343e-02 -1.49563462e-01 -1.35020089e+00
-5.56345403e-01 7.09595203e-01 2.67045408e-01 2.75008261e-01
5.11696517e-01 5.71880117e-02 6.64456904e-01 2.63407320e-01
4.89615798e-01 -1.30571878e+00 -1.34319901e-01 8.00141990e-01
1.26631081e+00 -9.51655447e-01 -1.68847620e-01 -9.54429284e-02
-5.50327718e-01 9.45049047e-01 6.26466691e-01 2.51738787e-01
2.03634426e-01 1.97350487e-01 3.91479254e-01 -1.80180565e-01
-1.06249905e+00 -1.60147682e-01 2.85609871e-01 8.50557506e-01
1.17800283e+00 1.73993871e-01 -1.01927352e+00 2.77851075e-01
-8.87947917e-01 4.52568717e-02 4.08590376e-01 5.70041776e-01
-2.77682900e-01 -1.85066509e+00 5.64332902e-02 3.66101027e-01
-4.77717012e-01 -3.37475181e-01 -7.90311456e-01 1.27274001e+00
2.15770602e-01 1.03580940e+00 -2.36660928e-01 -2.87700057e-01
3.97687465e-01 3.03737432e-01 6.59985006e-01 -8.71123075e-01
-7.42860138e-01 -2.30818629e-01 7.61534572e-01 -5.91243267e-01
-2.64878601e-01 -9.54562604e-01 -8.00134540e-01 -4.22359079e-01
-1.02572873e-01 4.93539989e-01 7.52241135e-01 6.57568097e-01
3.58651668e-01 8.01223367e-02 1.16126209e-01 -5.35050273e-01
-5.87760687e-01 -1.31493759e+00 -3.32141578e-01 4.52388465e-01
-2.27481723e-01 -3.90980691e-01 -3.29444587e-01 -9.73009989e-02] | [11.326005935668945, 9.576085090637207] |
b914dd24-7090-421d-a756-237350cfb2cf | learning-to-prompt-for-open-vocabulary-object | 2203.14940 | null | https://arxiv.org/abs/2203.14940v1 | https://arxiv.org/pdf/2203.14940v1.pdf | Learning to Prompt for Open-Vocabulary Object Detection with Vision-Language Model | Recently, vision-language pre-training shows great potential in open-vocabulary object detection, where detectors trained on base classes are devised for detecting new classes. The class text embedding is firstly generated by feeding prompts to the text encoder of a pre-trained vision-language model. It is then used as the region classifier to supervise the training of a detector. The key element that leads to the success of this model is the proper prompt, which requires careful words tuning and ingenious design. To avoid laborious prompt engineering, there are some prompt representation learning methods being proposed for the image classification task, which however can only be sub-optimal solutions when applied to the detection task. In this paper, we introduce a novel method, detection prompt (DetPro), to learn continuous prompt representations for open-vocabulary object detection based on the pre-trained vision-language model. Different from the previous classification-oriented methods, DetPro has two highlights: 1) a background interpretation scheme to include the proposals in image background into the prompt training; 2) a context grading scheme to separate proposals in image foreground for tailored prompt training. We assemble DetPro with ViLD, a recent state-of-the-art open-world object detector, and conduct experiments on the LVIS as well as transfer learning on the Pascal VOC, COCO, Objects365 datasets. Experimental results show that our DetPro outperforms the baseline ViLD in all settings, e.g., +3.4 APbox and +3.0 APmask improvements on the novel classes of LVIS. Code and models are available at https://github.com/dyabel/detpro. | ['Guoqi Li', 'Yue Gao', 'Miaojing Shi', 'Zihe Zhang', 'Fangyun Wei', 'Yu Du'] | 2022-03-28 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Du_Learning_To_Prompt_for_Open-Vocabulary_Object_Detection_With_Vision-Language_Model_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Du_Learning_To_Prompt_for_Open-Vocabulary_Object_Detection_With_Vision-Language_Model_CVPR_2022_paper.pdf | cvpr-2022-1 | ['open-vocabulary-object-detection'] | ['computer-vision'] | [ 1.68134049e-01 -3.64061743e-02 -1.99876517e-01 -3.11210275e-01
-8.03402007e-01 -5.35069942e-01 7.55470634e-01 -9.00987834e-02
-5.51815748e-01 1.67345896e-01 -1.39086500e-01 -3.83447677e-01
6.60941660e-01 -6.12761021e-01 -8.40031981e-01 -6.78050101e-01
3.99390817e-01 4.75724638e-01 8.81608844e-01 -1.10168653e-02
9.72796902e-02 3.25638324e-01 -1.76862085e+00 5.32513261e-01
5.47602832e-01 1.09828067e+00 6.60631180e-01 7.95707166e-01
-4.65936124e-01 8.37987244e-01 -4.84575242e-01 -3.91078740e-01
2.56768525e-01 -1.53324351e-01 -5.63546479e-01 1.42456457e-01
8.84091735e-01 -4.81770545e-01 -2.85833657e-01 1.09585738e+00
6.00383341e-01 -2.50854611e-01 9.05680358e-01 -1.11046827e+00
-8.84235203e-01 3.50840569e-01 -5.75499177e-01 4.94337559e-01
1.14927001e-01 4.84015584e-01 1.11471522e+00 -1.48624825e+00
5.12602329e-01 1.20554304e+00 3.82986456e-01 8.96404207e-01
-1.16387713e+00 -6.46761537e-01 2.59494483e-01 3.87720913e-01
-1.34559274e+00 -3.31991553e-01 6.43806875e-01 -8.06821048e-01
9.64995623e-01 1.61133274e-01 4.51168865e-01 1.27095342e+00
-1.43262735e-02 1.19626665e+00 9.10031796e-01 -6.59386992e-01
1.20432854e-01 5.48633218e-01 5.03479064e-01 8.44427586e-01
2.26547197e-01 3.05623859e-01 -2.27183387e-01 1.05156697e-01
5.93266904e-01 -7.86493793e-02 -2.08200678e-01 -4.82291162e-01
-1.03016698e+00 9.26149607e-01 5.48914254e-01 6.64644316e-02
-2.13471264e-01 5.63087091e-02 5.35724044e-01 -1.03428243e-02
4.56202179e-01 4.33342308e-02 -4.70366180e-01 4.05051380e-01
-6.88630402e-01 1.93834439e-01 5.78839779e-01 1.07342517e+00
7.05234468e-01 5.90249784e-02 -7.12182939e-01 8.52065146e-01
6.53529346e-01 5.42296052e-01 3.49084824e-01 -4.13461417e-01
3.40957910e-01 8.05786729e-01 -1.59196258e-01 -4.78511631e-01
-1.63825393e-01 -4.63380247e-01 -4.37732607e-01 4.09509689e-01
3.99385244e-01 1.20601654e-01 -1.29146934e+00 1.26353645e+00
4.80348945e-01 1.84366867e-01 1.03749298e-01 8.21514368e-01
1.48129153e+00 8.32244635e-01 3.49616706e-01 1.10546909e-01
1.83925581e+00 -1.25389409e+00 -5.13995051e-01 -5.83201945e-01
6.43359482e-01 -8.91947627e-01 1.21875083e+00 2.31708899e-01
-6.80927515e-01 -1.00683498e+00 -1.00976264e+00 -1.90044597e-01
-4.73619640e-01 6.58334672e-01 4.45581406e-01 4.97061878e-01
-1.01890016e+00 -1.82781026e-01 -5.57628334e-01 -5.87555349e-01
6.35481358e-01 2.68546361e-02 -5.45620406e-03 -8.07036981e-02
-8.83477509e-01 8.82027388e-01 6.39095962e-01 -2.74615288e-02
-1.35002387e+00 -6.05869055e-01 -8.83693516e-01 -7.38708824e-02
6.58623874e-01 -5.77632368e-01 1.42850792e+00 -1.06216979e+00
-1.22618866e+00 1.36747456e+00 3.31273079e-02 -4.97776836e-01
5.27806282e-01 -9.12992656e-02 -2.27003515e-01 1.62343845e-01
2.90166646e-01 1.11476600e+00 1.29570436e+00 -1.24902391e+00
-1.05414939e+00 -6.22228198e-02 2.08714619e-01 3.27550583e-02
-2.42137730e-01 4.76093739e-01 -8.74733150e-01 -6.19017780e-01
-2.62990087e-01 -7.40504444e-01 -1.94189936e-01 3.48057181e-01
-2.36049190e-01 -8.65843832e-01 9.72218633e-01 -5.49421728e-01
9.96058285e-01 -2.32475233e+00 -2.13446379e-01 -3.62143457e-01
2.71448255e-01 7.33590126e-01 -2.65246868e-01 -1.08864188e-01
5.11149131e-02 -3.63077611e-01 9.41346288e-02 -2.88241565e-01
-4.01231162e-02 2.60015935e-01 -5.28504491e-01 3.14615369e-01
5.62797785e-01 9.83916521e-01 -7.98759460e-01 -7.67458379e-01
4.68351394e-01 3.62296402e-01 -4.64417070e-01 5.25825441e-01
-5.54295838e-01 -5.88051155e-02 -3.83654833e-01 7.83876479e-01
4.91902739e-01 -3.66783142e-01 -3.46935898e-01 -2.71665365e-01
-2.01981857e-01 1.38770923e-01 -1.15819669e+00 1.30425560e+00
-2.28157133e-01 6.95405185e-01 5.18813953e-02 -8.98703635e-01
1.01221085e+00 1.00563042e-01 -1.26537666e-01 -4.46282417e-01
2.99229175e-01 1.56286180e-01 -8.28194544e-02 -6.60633266e-01
3.80305886e-01 3.25018376e-01 1.18407957e-01 -8.97976309e-02
4.64887619e-01 -3.49426270e-02 4.72423553e-01 3.39664251e-01
8.06644261e-01 1.75699264e-01 2.80838430e-01 -2.41825283e-01
9.30099308e-01 7.27031231e-02 4.60634112e-01 8.29044342e-01
-5.24882138e-01 6.47150159e-01 2.03483358e-01 -5.12795329e-01
-7.92083561e-01 -9.24080431e-01 -3.45713645e-01 1.49455702e+00
4.01505604e-02 -4.16961312e-01 -7.29300141e-01 -1.05935979e+00
-4.22892310e-02 7.00233459e-01 -5.17753243e-01 3.11289057e-02
-3.65032226e-01 -5.48765182e-01 3.91663134e-01 8.57680917e-01
3.39602947e-01 -1.20791984e+00 -8.04490268e-01 6.00551106e-02
5.77120893e-02 -1.33695030e+00 -5.49909174e-01 4.95358497e-01
-4.18045402e-01 -1.03569019e+00 -7.41942823e-01 -1.33509672e+00
7.19148159e-01 4.77987826e-01 9.61471498e-01 3.24738771e-02
-6.23736024e-01 7.41794348e-01 -3.71497810e-01 -7.65378714e-01
-6.28340721e-01 -1.69724107e-01 6.60077087e-04 2.22212374e-01
7.44674325e-01 3.00486207e-01 -5.96445501e-01 2.29048833e-01
-6.34752214e-01 1.65062442e-01 6.35171235e-01 9.11721885e-01
7.60852754e-01 -4.89320397e-01 2.69210696e-01 -5.05763531e-01
2.05483764e-01 -6.00952208e-02 -1.08512914e+00 3.12080383e-01
-4.65991497e-01 -1.46529093e-01 3.99512619e-01 -6.44872248e-01
-9.76849139e-01 4.49939430e-01 -1.96634695e-01 -7.24021852e-01
-3.73005003e-01 -3.62646729e-02 -1.71276078e-01 -1.27282783e-01
8.51485312e-01 3.06863874e-01 -1.95923075e-01 -4.03358549e-01
6.95525944e-01 9.07427490e-01 5.97264290e-01 -4.23772991e-01
7.71951556e-01 3.59230161e-01 -5.03291011e-01 -9.76711333e-01
-9.04877961e-01 -8.11334491e-01 -5.27859688e-01 -3.09401125e-01
1.08962762e+00 -1.20410275e+00 -2.39696145e-01 3.75146031e-01
-1.38494945e+00 -4.86126751e-01 -3.83427739e-01 1.41296789e-01
-3.95409614e-01 2.94205010e-01 -3.84858876e-01 -7.08282709e-01
-4.61645246e-01 -1.22685671e+00 1.22126341e+00 4.01254535e-01
5.42473532e-02 -6.94984376e-01 -9.87394899e-02 4.62595731e-01
1.16481468e-01 -2.63575673e-01 4.72997010e-01 -8.26185882e-01
-7.66348898e-01 -2.28294730e-01 -6.07755542e-01 7.80677795e-01
-2.20742553e-01 1.27584562e-01 -1.34196866e+00 -1.87684476e-01
-1.56088963e-01 -6.41486406e-01 1.14239240e+00 3.09228927e-01
1.10419428e+00 -5.01199029e-02 -5.46795070e-01 6.51777208e-01
1.25520146e+00 1.24723032e-01 4.16022092e-01 4.05090302e-01
8.08443308e-01 5.28367043e-01 8.17645311e-01 1.45248830e-01
3.74358416e-01 6.93985879e-01 4.60284472e-01 -3.17270011e-01
-6.09562695e-01 -2.00112626e-01 6.42639995e-01 3.64607781e-01
1.99112520e-01 -3.67294694e-03 -9.72445726e-01 5.95980525e-01
-1.70349872e+00 -7.74392426e-01 -3.57677758e-01 1.92416573e+00
8.77756715e-01 4.09421623e-01 -1.51104080e-02 -7.68292919e-02
6.14762783e-01 1.55463889e-01 -4.21826869e-01 -2.87132651e-01
-4.96531315e-02 9.78561416e-02 4.45386350e-01 3.30468595e-01
-1.56331205e+00 1.36071849e+00 5.10432959e+00 9.46281612e-01
-1.21243799e+00 3.98589373e-01 4.27119017e-01 1.73858970e-01
2.71514803e-01 -4.62493338e-02 -1.57386136e+00 3.01042378e-01
5.25011122e-01 2.26613447e-01 1.75319856e-03 1.31011009e+00
-3.49222608e-02 -5.04779220e-02 -1.28420162e+00 1.20583296e+00
3.66598636e-01 -1.38027740e+00 1.77249193e-01 -3.01545560e-01
4.26918060e-01 3.87721628e-01 -8.83493200e-02 7.43075371e-01
1.24831542e-01 -5.17370343e-01 9.01490688e-01 3.25742476e-02
8.90814483e-01 -1.92026585e-01 4.29558247e-01 3.59662533e-01
-1.27404678e+00 -1.44149035e-01 -6.04148507e-01 1.68834120e-01
-7.22434893e-02 2.12548763e-01 -1.17231870e+00 1.01841658e-01
8.46766531e-01 5.34259021e-01 -9.13809478e-01 9.93566453e-01
-7.42071509e-01 8.57518196e-01 -1.65205836e-01 -1.44157380e-01
2.36258924e-01 1.74052343e-01 5.52147567e-01 1.48992491e+00
-2.55992770e-01 1.34064620e-02 5.60027659e-01 9.64783788e-01
1.22154325e-01 1.96345434e-01 -3.96979988e-01 7.24319890e-02
1.58894867e-01 1.54187512e+00 -7.32966244e-01 -6.11293554e-01
-8.13075244e-01 9.36959565e-01 2.94853449e-01 3.36649150e-01
-9.89033997e-01 -2.81258076e-01 4.03613567e-01 1.57234296e-01
6.84128404e-01 2.50573969e-03 -2.94213574e-02 -1.05877912e+00
-1.55491829e-02 -8.17364633e-01 6.35616362e-01 -7.55064845e-01
-1.27388597e+00 7.07761467e-01 -8.71877149e-02 -1.12076604e+00
-9.06432047e-03 -1.20479000e+00 -6.66827917e-01 7.06530690e-01
-1.78456831e+00 -1.42772353e+00 -3.22015017e-01 6.30041420e-01
1.11026466e+00 -3.01054835e-01 5.94328701e-01 3.44502151e-01
-7.55403221e-01 6.39643192e-01 -2.66308218e-01 4.11463171e-01
7.16437519e-01 -1.22987998e+00 3.38050872e-01 1.12208056e+00
3.88594776e-01 2.56985903e-01 5.62907398e-01 -4.91121322e-01
-1.23661518e+00 -1.46915507e+00 8.95565987e-01 -7.69351542e-01
7.66999722e-01 -6.75675035e-01 -9.90063488e-01 6.85715795e-01
1.92691982e-01 5.40223420e-01 3.37755680e-01 -1.67985424e-01
-4.62105602e-01 -1.53803840e-01 -6.63399816e-01 5.75431585e-01
9.03778434e-01 -4.71369535e-01 -9.99044001e-01 5.97103059e-01
8.40082943e-01 -5.50121188e-01 -2.16726214e-01 3.91327351e-01
2.67125696e-01 -5.85767269e-01 1.08266592e+00 -4.57350761e-01
8.84944424e-02 -5.79437196e-01 -1.25047922e-01 -6.39798462e-01
-4.46704149e-01 -2.55039871e-01 -2.52325863e-01 1.48824501e+00
3.04585844e-01 -4.73931015e-01 4.19368833e-01 1.96473107e-01
-1.65514350e-01 -7.49442518e-01 -7.04200208e-01 -8.29180360e-01
-2.31379226e-01 -6.03087604e-01 2.85865311e-02 5.57629764e-01
-6.69409811e-01 6.95111930e-01 3.40204537e-02 3.74484748e-01
6.00957811e-01 -5.58827631e-02 1.10848415e+00 -1.09865689e+00
-3.79895091e-01 -4.24355090e-01 -4.92226720e-01 -1.27621722e+00
8.29976127e-02 -1.03879917e+00 2.32426584e-01 -1.57595396e+00
3.56695354e-01 -3.64557385e-01 -3.79103363e-01 6.98433697e-01
-4.14467037e-01 2.99767137e-01 3.08943778e-01 1.43235222e-01
-7.78673649e-01 4.93819118e-01 9.50651228e-01 -4.76802021e-01
-2.24704832e-01 6.93321526e-02 -5.00705957e-01 9.18796897e-01
5.53782701e-01 -4.75535542e-01 -3.17588091e-01 -2.86992043e-01
-8.88325199e-02 -4.94495273e-01 8.43971968e-01 -9.96199489e-01
1.54970065e-01 1.22041196e-01 3.33646983e-01 -9.81778502e-01
1.03662759e-01 -6.37261569e-01 -5.12831211e-01 4.85660523e-01
-1.54658720e-01 -4.64989811e-01 3.69366378e-01 5.93375385e-01
1.57104712e-02 -3.78612816e-01 9.61784959e-01 2.79794186e-02
-1.39053464e+00 3.22113574e-01 -3.06909949e-01 1.29007071e-01
1.19058287e+00 -2.99791992e-01 -3.82682264e-01 2.19184428e-01
-6.18743718e-01 5.46507299e-01 1.80843636e-01 8.11156869e-01
7.28999376e-01 -1.04857230e+00 -8.50159943e-01 2.85293758e-01
7.25493073e-01 1.95074812e-01 1.13344127e-02 7.68183887e-01
-3.82954925e-01 3.92097861e-01 7.93335810e-02 -1.19111860e+00
-1.46922565e+00 9.96571720e-01 3.83449167e-01 -3.02618388e-02
-7.92653918e-01 1.24026442e+00 9.21356380e-01 -1.64176270e-01
7.31147528e-01 -5.77982605e-01 -3.46198767e-01 1.39681011e-01
8.93425703e-01 -8.96561220e-02 -8.29750206e-03 -6.17605269e-01
-4.87752795e-01 5.04207373e-01 -4.74078506e-01 2.81000942e-01
9.89726245e-01 2.41468754e-03 6.29751608e-02 3.84095669e-01
8.59674156e-01 -2.86437333e-01 -1.50631964e+00 -5.95034420e-01
-4.73670699e-02 -2.07088143e-01 1.61445275e-01 -7.59700596e-01
-8.12477469e-01 1.04966056e+00 1.04714739e+00 -4.72772717e-02
9.70757127e-01 4.20275539e-01 3.12801719e-01 2.14480117e-01
2.90918469e-01 -1.01825845e+00 3.16995412e-01 6.02864325e-01
9.76434946e-01 -1.53598213e+00 -3.30352187e-01 -3.39353770e-01
-6.40480280e-01 1.14206016e+00 1.06446457e+00 -7.68180564e-02
4.67957079e-01 1.77131683e-01 3.63131076e-01 -1.55104458e-01
-6.76274419e-01 -6.81390643e-01 4.97342587e-01 6.74699485e-01
2.66857207e-01 -2.16064844e-02 -9.97886434e-02 6.82705998e-01
3.07667971e-01 -1.63401619e-01 2.75204331e-01 7.76281834e-01
-7.84561694e-01 -8.58774483e-01 -6.19777262e-01 2.93711454e-01
-1.73788667e-01 -2.04819500e-01 -3.64213079e-01 5.32899797e-01
4.53538328e-01 8.04071307e-01 -8.61055553e-02 -7.30161220e-02
3.55172515e-01 2.11572886e-01 4.31636423e-01 -1.00568199e+00
-3.86051983e-01 6.22297674e-02 -4.81069349e-02 -5.78750014e-01
-1.66157275e-01 -6.37479722e-01 -1.18882811e+00 5.74814260e-01
-5.97077847e-01 -3.10416967e-01 4.62859541e-01 7.39756227e-01
2.01657146e-01 7.30349064e-01 3.03390950e-01 -8.81547451e-01
-6.92368686e-01 -9.62162971e-01 -2.68391501e-02 2.09208503e-01
4.26260799e-01 -8.23171735e-01 -2.51746774e-01 2.28534296e-01] | [9.638030052185059, 1.4514102935791016] |
ae33c588-5153-4b9b-a9d6-14b20622741a | zero-shot-object-detection-learning-to | 1803.06049 | null | http://arxiv.org/abs/1803.06049v1 | http://arxiv.org/pdf/1803.06049v1.pdf | Zero-Shot Object Detection: Learning to Simultaneously Recognize and Localize Novel Concepts | Current Zero-Shot Learning (ZSL) approaches are restricted to recognition of
a single dominant unseen object category in a test image. We hypothesize that
this setting is ill-suited for real-world applications where unseen objects
appear only as a part of a complex scene, warranting both the `recognition' and
`localization' of an unseen category. To address this limitation, we introduce
a new \emph{`Zero-Shot Detection'} (ZSD) problem setting, which aims at
simultaneously recognizing and locating object instances belonging to novel
categories without any training examples. We also propose a new experimental
protocol for ZSD based on the highly challenging ILSVRC dataset, adhering to
practical issues, e.g., the rarity of unseen objects. To the best of our
knowledge, this is the first end-to-end deep network for ZSD that jointly
models the interplay between visual and semantic domain information. To
overcome the noise in the automatically derived semantic descriptions, we
utilize the concept of meta-classes to design an original loss function that
achieves synergy between max-margin class separation and semantic space
clustering. Furthermore, we present a baseline approach extended from
recognition to detection setting. Our extensive experiments show significant
performance boost over the baseline on the imperative yet difficult ZSD
problem. | ['Shafin Rahman', 'Salman Khan', 'Fatih Porikli'] | 2018-03-16 | null | null | null | null | ['zero-shot-object-detection', 'novel-concepts'] | ['computer-vision', 'reasoning'] | [ 3.58145863e-01 9.08485800e-02 -1.26267588e-02 -4.48472798e-01
-9.16241169e-01 -4.58805412e-01 5.84073007e-01 1.85344949e-01
-4.48334962e-01 2.47735545e-01 -1.56703055e-01 3.18720862e-02
-6.60098046e-02 -5.42610288e-01 -6.07654631e-01 -6.46571457e-01
1.56499937e-01 4.12950277e-01 3.21981966e-01 2.98417755e-03
4.08534892e-02 2.40472764e-01 -1.92899406e+00 3.90734196e-01
6.25524580e-01 1.26443911e+00 4.76104677e-01 2.46186227e-01
6.42118603e-02 5.29960096e-01 -5.18321514e-01 -2.34309301e-01
3.90846282e-01 -1.91660479e-01 -6.76034868e-01 5.11464775e-01
6.87674582e-01 -3.99588674e-01 -2.28151858e-01 1.08206379e+00
5.77696800e-01 5.47197282e-01 6.67006016e-01 -1.40726793e+00
-7.36198425e-01 1.16916776e-01 -4.79137897e-01 2.48953462e-01
1.15045786e-01 1.96522608e-01 1.19686854e+00 -1.29720175e+00
6.24651790e-01 1.14994180e+00 3.26626062e-01 7.25409806e-01
-1.33066010e+00 -3.87427419e-01 5.26941359e-01 4.98328626e-01
-1.63475978e+00 -5.86871505e-01 7.82888114e-01 -4.87657309e-01
8.65125895e-01 1.85297012e-01 1.03205144e-01 1.41707838e+00
-5.77992678e-01 1.20238507e+00 7.22361743e-01 -4.40053552e-01
6.67878687e-01 3.06362957e-01 2.96201408e-01 3.81882071e-01
2.36235723e-01 -1.49552040e-02 -3.89818609e-01 6.87866658e-02
2.64217526e-01 2.12718099e-01 -2.19274297e-01 -1.04834521e+00
-1.14317298e+00 7.84524560e-01 5.85849226e-01 2.59743989e-01
-5.49728982e-02 -2.56463021e-01 3.44310254e-01 1.00241669e-01
3.52388501e-01 3.50809783e-01 -4.53052849e-01 1.35936782e-01
-7.52843559e-01 8.07541534e-02 5.52794039e-01 1.12139702e+00
7.00060487e-01 -1.13958150e-01 -2.47209549e-01 9.78346229e-01
1.53399333e-01 2.37125620e-01 4.71208870e-01 -5.77241063e-01
3.98047984e-01 6.67399466e-01 2.16902554e-01 -8.68098736e-01
-2.10017458e-01 -6.76817298e-01 -6.51889443e-01 2.22141802e-01
2.28903934e-01 2.05325782e-01 -1.26717710e+00 2.00482130e+00
6.00507140e-01 4.53721523e-01 2.26170823e-01 1.30987418e+00
9.70503390e-01 4.22025412e-01 5.88523485e-02 4.46120948e-02
1.35406792e+00 -9.13226545e-01 -4.76851523e-01 -5.48580527e-01
5.42703927e-01 -3.19699228e-01 1.22421837e+00 1.71833023e-01
-5.44041812e-01 -5.55709600e-01 -1.08721268e+00 -8.99993628e-02
-7.17407942e-01 1.58914149e-01 4.89867538e-01 3.10087144e-01
-6.22507930e-01 2.96434164e-01 -6.47494614e-01 -6.50537789e-01
7.40870714e-01 1.55450881e-01 -4.68680441e-01 -2.94191599e-01
-9.67413127e-01 6.92453265e-01 7.06842005e-01 1.62289411e-01
-1.14856136e+00 -5.26664019e-01 -1.08710325e+00 3.73252183e-02
9.45145726e-01 -4.36627001e-01 1.18966198e+00 -1.01152682e+00
-9.03255582e-01 1.21569741e+00 -9.54908356e-02 -3.49330753e-01
4.09163088e-01 -1.71025321e-01 -4.29578871e-01 1.11365795e-01
4.19879407e-01 5.44049442e-01 8.44743192e-01 -1.54970860e+00
-7.59026110e-01 -4.98961836e-01 8.40361118e-02 2.74524242e-01
-4.50068653e-01 -3.49168897e-01 -5.97853422e-01 -6.33024693e-01
2.15148732e-01 -6.88298523e-01 -1.05158679e-01 2.20927626e-01
-4.50581253e-01 -4.28000391e-01 9.83689427e-01 -1.49402633e-01
8.99669588e-01 -2.52583456e+00 1.41276553e-01 -6.71812743e-02
3.01308513e-01 4.75100905e-01 -2.52545774e-01 1.68052495e-01
1.17177861e-02 -1.77853703e-01 -3.70620430e-01 -5.03785014e-01
2.86160916e-01 3.43989104e-01 -4.41744983e-01 4.68746990e-01
3.67016673e-01 9.29142475e-01 -1.09104848e+00 -3.60352635e-01
3.85185212e-01 2.57263720e-01 -3.24853808e-01 2.96249658e-01
-1.31901592e-01 1.84443071e-01 -2.89065003e-01 8.95837128e-01
6.64508402e-01 -3.78430963e-01 -3.99166085e-02 -2.01127902e-01
1.92570075e-01 -1.17648743e-01 -1.31283057e+00 1.95926380e+00
-1.96552128e-01 3.66206825e-01 -7.52054080e-02 -1.25314200e+00
7.60628939e-01 1.08038783e-01 2.79210597e-01 -6.17710292e-01
1.67938322e-01 2.23205447e-01 -2.20720291e-01 -4.85472649e-01
2.12741435e-01 -3.12403947e-01 -2.34517053e-01 1.61382094e-01
5.63737094e-01 2.53111869e-01 1.28394783e-01 2.43677557e-01
1.00761855e+00 -1.13838501e-02 3.94236386e-01 -3.51353846e-02
1.80272505e-01 3.96790057e-02 6.30123377e-01 1.02556551e+00
-7.14612007e-01 8.54844511e-01 2.65627027e-01 -3.70975494e-01
-8.39250863e-01 -1.37122583e+00 -2.44935215e-01 1.23657691e+00
6.62623167e-01 -1.54952452e-01 -4.31250513e-01 -9.68083978e-01
1.21747486e-01 8.86552334e-01 -7.72132576e-01 -4.72605318e-01
-4.09126952e-02 -6.76343262e-01 1.28329143e-01 4.14617002e-01
2.73255676e-01 -8.13159406e-01 -7.70615399e-01 9.42365080e-02
-8.16484466e-02 -1.30301332e+00 -2.60487765e-01 3.71659309e-01
-4.23070312e-01 -1.15271604e+00 -5.61441600e-01 -9.83643651e-01
6.50411487e-01 7.31849909e-01 8.97140622e-01 -2.34666929e-01
-5.37151754e-01 5.16546905e-01 -5.46556413e-01 -3.25109780e-01
1.96694452e-02 -2.21907079e-01 8.07841346e-02 4.98986930e-01
7.59430885e-01 -3.41975689e-01 -7.85936832e-01 3.96830857e-01
-8.57933819e-01 2.22534314e-02 5.81547022e-01 7.85307765e-01
6.56633317e-01 1.25398174e-01 6.00745976e-01 -5.53006232e-01
1.99471042e-01 -7.22229242e-01 -2.76971728e-01 2.89928228e-01
-3.34481925e-01 -2.27314293e-01 5.15834093e-01 -6.18019223e-01
-8.63118827e-01 1.59975633e-01 2.37830296e-01 -9.23955798e-01
-5.50636768e-01 1.58455804e-01 -4.54593897e-01 1.86435640e-01
5.28784335e-01 2.79790312e-01 -9.75441933e-02 -5.84455729e-01
4.62271333e-01 7.81576633e-01 7.79096603e-01 -3.08775395e-01
8.42741311e-01 7.20007956e-01 -2.71965355e-01 -8.09002578e-01
-1.27417016e+00 -1.02415621e+00 -7.08470583e-01 -8.89937356e-02
7.88117766e-01 -1.17071581e+00 -3.78414601e-01 2.01969340e-01
-9.79808629e-01 -8.58840123e-02 -5.47065198e-01 3.72056186e-01
-5.86539924e-01 3.04070294e-01 -1.89261168e-01 -8.20309877e-01
-9.72723588e-02 -9.01053309e-01 1.49581301e+00 -5.75446002e-02
-1.50505900e-01 -6.43333733e-01 -2.00568542e-01 2.64942646e-01
2.63121147e-02 4.32493508e-01 6.71591997e-01 -1.08901894e+00
-5.00441968e-01 -3.97041589e-01 -4.30707335e-01 4.53576326e-01
1.42357960e-01 -5.36934853e-01 -1.30694139e+00 -5.61742842e-01
1.38297632e-01 -6.20654643e-01 1.10569668e+00 -3.82627547e-02
1.19271350e+00 -8.67652223e-02 -3.46128583e-01 4.98262942e-01
1.54735172e+00 1.34587020e-01 2.58073300e-01 1.96436018e-01
8.59237850e-01 7.52256095e-01 8.16675901e-01 5.04345357e-01
2.66866803e-01 7.59067953e-01 6.39512420e-01 -1.74579605e-01
-1.06540948e-01 -3.39177638e-01 7.68073946e-02 2.91129470e-01
4.09536779e-01 -4.29472715e-01 -9.02004242e-01 8.08579445e-01
-2.06641650e+00 -8.85046303e-01 2.15380087e-01 2.29733539e+00
4.10186589e-01 9.19368491e-02 -4.62852381e-02 1.17931262e-01
8.80929589e-01 1.98171064e-01 -8.33948314e-01 1.49867713e-01
-7.99303949e-02 -6.46370798e-02 1.87779397e-01 1.11239634e-01
-1.60855711e+00 9.95122313e-01 5.00765610e+00 1.04842186e+00
-9.04883087e-01 1.90285698e-01 3.92137080e-01 -3.21559131e-01
1.92078054e-02 -2.06475005e-01 -8.66632342e-01 5.04694998e-01
4.79668319e-01 -5.53279147e-02 1.07180320e-01 1.07689250e+00
-1.80644114e-02 -1.45804472e-02 -1.34540606e+00 1.24522090e+00
3.85007381e-01 -8.55060577e-01 1.51652247e-01 -9.49624628e-02
5.62578440e-01 -6.71310863e-03 1.61168784e-01 5.16389549e-01
1.05628721e-01 -7.89803624e-01 9.05064404e-01 1.54500991e-01
7.86277235e-01 -5.32502294e-01 5.03331184e-01 4.86907721e-01
-1.15355301e+00 -3.63871455e-01 -5.11250556e-01 9.97309983e-02
-1.12227853e-02 4.35959965e-01 -8.42057645e-01 3.59497100e-01
6.82454705e-01 9.39148009e-01 -6.27401054e-01 1.02372360e+00
-9.73300710e-02 2.18790114e-01 -2.16946855e-01 1.85631096e-01
4.41028595e-01 1.01834245e-01 6.87948942e-01 1.03012514e+00
2.44262457e-01 1.90069884e-01 3.29260439e-01 1.11700833e+00
2.95637976e-02 -7.04495683e-02 -7.63398051e-01 9.27350596e-02
3.88692170e-01 1.12497115e+00 -8.25036824e-01 -3.67655814e-01
-4.55289036e-01 1.24555743e+00 5.00121236e-01 5.27972877e-01
-7.14012027e-01 -4.77168381e-01 7.39790201e-01 -1.02953641e-02
6.98327839e-01 -3.52372825e-02 -4.86367941e-02 -1.49231529e+00
2.30051309e-01 -5.77595651e-01 5.85445523e-01 -6.02276266e-01
-1.55042171e+00 3.53858173e-01 -1.02002971e-01 -1.54008043e+00
1.05624236e-01 -7.12319493e-01 -3.66413027e-01 5.09184957e-01
-1.58365130e+00 -1.17057669e+00 -4.22906697e-01 5.74972808e-01
8.83216977e-01 -1.41933545e-01 7.36308038e-01 3.25983137e-01
-6.05626822e-01 5.73470414e-01 2.29758173e-01 1.48934454e-01
6.53900266e-01 -1.37970734e+00 2.18560070e-01 1.03145611e+00
1.56008899e-01 5.08641601e-01 6.23436749e-01 -4.07332331e-01
-1.33465874e+00 -1.38956213e+00 7.62038291e-01 -6.26012444e-01
5.93170226e-01 -9.08513784e-01 -1.06287277e+00 5.48395216e-01
-4.41994727e-01 3.53331298e-01 6.65227652e-01 1.57467052e-02
-5.16794503e-01 -1.57939166e-01 -1.06762636e+00 5.05787671e-01
1.29570329e+00 -6.44832730e-01 -8.26321959e-01 3.55262250e-01
8.18210244e-01 -1.94627553e-01 -4.51821953e-01 6.76564336e-01
2.11783394e-01 -7.81669259e-01 1.13039756e+00 -6.94624543e-01
1.43125385e-01 -5.57751894e-01 -5.25678873e-01 -1.09371042e+00
-2.28537410e-01 -1.65839538e-01 -3.32439005e-01 1.25080431e+00
9.72058065e-03 -2.40923345e-01 7.38826931e-01 5.76263309e-01
-2.40725204e-01 -5.82110405e-01 -1.10754442e+00 -1.01992500e+00
-2.96647519e-01 -5.88458300e-01 3.15254420e-01 1.00193119e+00
-2.13929236e-01 3.97740722e-01 -3.87151033e-01 4.82009858e-01
9.46533322e-01 3.04332465e-01 6.73635364e-01 -1.23831713e+00
-2.58716613e-01 -2.61167765e-01 -8.46028209e-01 -9.75668073e-01
1.66637495e-01 -9.41182733e-01 4.30384874e-01 -1.39478397e+00
3.95581603e-01 -3.38824660e-01 -7.09848940e-01 4.72636878e-01
-1.95488185e-01 2.97331721e-01 2.35505372e-01 1.87904447e-01
-1.25547957e+00 7.42677510e-01 9.36818779e-01 -2.76877224e-01
-9.22097936e-02 -8.67305994e-02 -8.62320364e-01 5.26753247e-01
5.31811297e-01 -3.09799373e-01 -6.72685266e-01 -4.79907721e-01
-3.12948316e-01 -3.74544948e-01 7.91707993e-01 -8.71648312e-01
1.79495409e-01 -4.61163931e-02 2.37288982e-01 -5.60184240e-01
4.50159043e-01 -1.03070104e+00 -2.95834661e-01 1.64381996e-01
-3.42530966e-01 -7.01915562e-01 6.79226667e-02 1.04011416e+00
-2.67430961e-01 -6.06376082e-02 9.14008975e-01 -8.95313025e-02
-1.44848609e+00 4.24264073e-01 3.01741473e-02 1.55809730e-01
1.48656404e+00 -5.08137405e-01 -3.09044272e-01 2.80537233e-02
-8.90882492e-01 4.25214618e-01 4.01461631e-01 8.26331019e-01
8.59428287e-01 -1.34350419e+00 -4.68950421e-01 3.87775779e-01
7.87017822e-01 2.75662810e-01 3.86192799e-01 6.32254958e-01
3.94459143e-02 3.10884356e-01 9.22009423e-02 -7.96660423e-01
-1.10141230e+00 1.08068657e+00 2.99635857e-01 2.23484457e-01
-7.87257373e-01 9.32112932e-01 8.49495649e-01 -4.49721664e-01
5.82180560e-01 7.87264016e-03 -1.10200964e-01 1.89446270e-01
6.36129498e-01 2.81215250e-01 1.74158230e-01 -7.46233702e-01
-5.85770726e-01 3.95698041e-01 -1.50392711e-01 1.88028783e-01
1.29300165e+00 -4.08918947e-01 2.97330976e-01 7.30518043e-01
1.45312381e+00 -7.14859545e-01 -1.50553882e+00 -6.46124184e-01
2.09849924e-01 -5.85252821e-01 5.13322204e-02 -8.03641558e-01
-8.19444180e-01 9.05560672e-01 8.70822906e-01 -1.04838975e-01
9.69010055e-01 3.55317563e-01 6.20807588e-01 4.17410970e-01
4.39915866e-01 -1.19147348e+00 2.71410167e-01 2.85824448e-01
6.73051775e-01 -1.69415867e+00 -3.77470464e-01 -2.85400808e-01
-6.96480155e-01 7.93537915e-01 8.25226426e-01 3.82681563e-02
4.22491610e-01 -2.30836302e-01 -9.93847251e-02 -2.80070841e-01
-6.92092299e-01 -7.35400021e-01 4.49310243e-01 6.14356995e-01
-2.79565901e-01 2.45624222e-02 9.82877389e-02 7.25331604e-01
3.25918913e-01 -1.92266196e-01 1.71328366e-01 1.09948587e+00
-7.66296923e-01 -5.53582311e-01 -1.98248431e-01 3.88623387e-01
-3.05536073e-02 2.33324729e-02 -4.47942048e-01 7.48695135e-01
4.28889573e-01 1.07811296e+00 2.16437250e-01 -2.47261599e-01
5.19927621e-01 1.11403361e-01 2.20594287e-01 -9.44869578e-01
1.53541833e-01 5.48642427e-02 -2.59688854e-01 -4.43980575e-01
-2.80735493e-01 -4.98948246e-01 -1.03220415e+00 1.73083231e-01
-2.26465806e-01 -6.16727173e-02 4.06898767e-01 8.77810419e-01
4.22381312e-01 3.81132752e-01 7.35377550e-01 -8.37939382e-01
-6.56552136e-01 -6.45851076e-01 -9.02137697e-01 9.66248512e-01
5.49556494e-01 -9.73412573e-01 -6.60205424e-01 -1.38165250e-01] | [9.765237808227539, 2.094698667526245] |
bc2ce675-68d9-47e4-a85c-66ccbf42a675 | breaking-the-lower-bound-with-little | 2302.06763 | null | https://arxiv.org/abs/2302.06763v1 | https://arxiv.org/pdf/2302.06763v1.pdf | Breaking the Lower Bound with (Little) Structure: Acceleration in Non-Convex Stochastic Optimization with Heavy-Tailed Noise | We consider the stochastic optimization problem with smooth but not necessarily convex objectives in the heavy-tailed noise regime, where the stochastic gradient's noise is assumed to have bounded $p$th moment ($p\in(1,2]$). Zhang et al. (2020) is the first to prove the $\Omega(T^{\frac{1-p}{3p-2}})$ lower bound for convergence (in expectation) and provides a simple clipping algorithm that matches this optimal rate. Cutkosky and Mehta (2021) proposes another algorithm, which is shown to achieve the nearly optimal high-probability convergence guarantee $O(\log(T/\delta)T^{\frac{1-p}{3p-2}})$, where $\delta$ is the probability of failure. However, this desirable guarantee is only established under the additional assumption that the stochastic gradient itself is bounded in $p$th moment, which fails to hold even for quadratic objectives and centered Gaussian noise. In this work, we first improve the analysis of the algorithm in Cutkosky and Mehta (2021) to obtain the same nearly optimal high-probability convergence rate $O(\log(T/\delta)T^{\frac{1-p}{3p-2}})$, without the above-mentioned restrictive assumption. Next, and curiously, we show that one can achieve a faster rate than that dictated by the lower bound $\Omega(T^{\frac{1-p}{3p-2}})$ with only a tiny bit of structure, i.e., when the objective function $F(x)$ is assumed to be in the form of $\mathbb{E}_{\Xi\sim\mathcal{D}}[f(x,\Xi)]$, arguably the most widely applicable class of stochastic optimization problems. For this class of problems, we propose the first variance-reduced accelerated algorithm and establish that it guarantees a high-probability convergence rate of $O(\log(T/\delta)T^{\frac{1-p}{2p-1}})$ under a mild condition, which is faster than $\Omega(T^{\frac{1-p}{3p-2}})$. Notably, even when specialized to the finite-variance case, our result yields the (near-)optimal high-probability rate $O(\log(T/\delta)T^{-1/3})$. | ['Zhengyuan Zhou', 'Jiawei Zhang', 'Zijian Liu'] | 2023-02-14 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [ 1.74871117e-01 1.11197338e-01 1.22799836e-01 5.06883338e-02
-1.30432940e+00 -6.20040655e-01 -1.60210222e-01 1.07515991e-01
-8.43957186e-01 9.70732808e-01 -6.29985571e-01 -7.16575325e-01
-7.08660543e-01 -8.11537743e-01 -8.72916102e-01 -1.26067531e+00
-5.51010907e-01 2.04171002e-01 1.81327879e-01 -1.42963201e-01
2.00652838e-01 2.44831026e-01 -1.72988117e+00 -6.27216697e-01
1.03781962e+00 1.61300671e+00 -4.57624719e-02 9.61715460e-01
-2.21785512e-02 1.95211127e-01 -4.99289513e-01 -5.17634034e-01
5.02527833e-01 -6.20180309e-01 -4.21448082e-01 -3.01510721e-01
1.25621676e-01 1.17831707e-01 2.58831140e-02 1.85784340e+00
5.81268549e-01 2.61147618e-01 6.82872653e-01 -1.03268933e+00
-1.36217788e-01 4.55039978e-01 -1.03105247e+00 3.17287743e-01
-4.45254296e-02 1.19547591e-01 8.57955694e-01 -5.28668404e-01
5.38824871e-02 6.91005588e-01 8.33533049e-01 4.24416959e-01
-9.05521333e-01 -8.89058948e-01 2.24952668e-01 -3.06934774e-01
-1.74001241e+00 1.13251954e-01 4.05159414e-01 -3.80034208e-01
4.03327525e-01 5.53534389e-01 2.58722842e-01 4.39674020e-01
1.91071242e-01 5.31968832e-01 1.14250970e+00 -4.73958343e-01
3.73025954e-01 7.44879693e-02 7.88697526e-02 7.49518096e-01
2.76265979e-01 2.30391249e-02 -1.37630224e-01 -2.09193349e-01
6.51842952e-01 -3.11932951e-01 -4.89072800e-01 3.65157008e-01
-7.19219685e-01 9.02684510e-01 -1.01358458e-01 5.26890755e-01
-3.23689878e-01 4.71074522e-01 1.01906210e-01 2.29569644e-01
5.49843013e-01 4.93835360e-02 -6.31723285e-01 -5.68337739e-01
-1.03472352e+00 3.58087450e-01 7.36870408e-01 1.22100616e+00
4.22890782e-01 2.37987772e-01 -3.20802331e-01 6.08808160e-01
3.19709629e-01 1.19316339e+00 -1.34898618e-01 -1.26715946e+00
7.26161242e-01 -1.08314991e-01 4.73213315e-01 -8.26428115e-01
-1.47770286e-01 -7.70769000e-01 -1.00690103e+00 2.34265551e-01
1.01708698e+00 -5.11324167e-01 -3.53585958e-01 2.10226703e+00
3.79547253e-02 -3.34743589e-01 -3.11100513e-01 7.56049097e-01
-1.61263019e-01 8.02054226e-01 -3.12116623e-01 -1.00069022e+00
1.17665744e+00 -1.66996986e-01 -4.16996360e-01 1.82051510e-01
3.86280417e-01 -7.54071474e-01 1.39052665e+00 5.98100781e-01
-1.46816564e+00 -1.46285996e-01 -7.81835318e-01 6.86105847e-01
9.22556967e-02 -4.25854594e-01 2.09074646e-01 1.21539760e+00
-9.87081349e-01 6.33221149e-01 -5.36516130e-01 2.63538182e-01
2.80817419e-01 3.78450692e-01 1.70457382e-02 1.02288030e-01
-8.75290871e-01 2.40897194e-01 -2.70255268e-01 1.42185152e-01
-5.85643411e-01 -8.27667296e-01 -4.45781499e-01 1.35072827e-01
5.74164987e-01 -2.62609631e-01 8.95089388e-01 -4.97065246e-01
-1.48403573e+00 5.64775467e-01 -2.27024347e-01 -2.25895718e-01
9.20661211e-01 1.01039156e-01 2.76048034e-02 1.63301259e-01
1.92761630e-01 -2.81683028e-01 6.71968460e-01 -9.66267407e-01
-8.32485318e-01 -8.07431459e-01 -1.04830123e-01 -1.79854006e-01
-2.58658469e-01 4.40707564e-01 -2.97956526e-01 -5.33661485e-01
9.69751030e-02 -8.31749976e-01 -3.78375709e-01 -1.55107945e-01
-2.72672355e-01 -1.39447063e-01 2.30607301e-01 -3.90389383e-01
1.40821552e+00 -2.12378454e+00 -1.30582929e-01 6.16278231e-01
3.81856039e-02 1.71860039e-01 5.74524522e-01 1.34534761e-01
9.35206190e-04 5.43508410e-01 -7.20487952e-01 -5.14980078e-01
2.27328897e-01 -1.18321188e-01 -9.66348276e-02 8.59127998e-01
-2.70225644e-01 1.88937411e-01 -8.29338014e-01 -1.19089425e-01
-2.25718737e-01 4.93070513e-01 -3.50478023e-01 -3.55298251e-01
-1.23374142e-01 3.78687441e-01 -6.20044231e-01 3.74684572e-01
1.01921439e+00 -2.09530324e-01 -1.72588646e-01 3.00924540e-01
-4.54439133e-01 -3.78334880e-01 -1.75307977e+00 1.06594539e+00
-3.35965037e-01 3.38499546e-01 8.05409431e-01 -1.43348217e+00
7.16722012e-01 1.58809990e-01 7.32047379e-01 -4.73590910e-01
4.08047706e-01 4.31631774e-01 -2.43097261e-01 -2.15046257e-01
6.78118365e-03 -9.69930053e-01 -2.00244367e-01 4.72829014e-01
-2.67693639e-01 -2.71872878e-01 2.58890212e-01 -1.39352605e-01
1.16431904e+00 -2.77679801e-01 -1.60058871e-01 -7.48255074e-01
5.82359910e-01 -4.88988668e-01 6.96633577e-01 1.22671914e+00
-5.63674927e-01 4.30992633e-01 1.02864432e+00 1.55844569e-01
-9.09597278e-01 -1.14183521e+00 -4.95439440e-01 1.02010202e+00
3.66564542e-02 -4.20594327e-02 -8.31009269e-01 -3.99957120e-01
-1.63997412e-01 9.47402954e-01 -6.97023451e-01 1.67348802e-01
-2.94648796e-01 -1.14609468e+00 6.69757009e-01 3.34171295e-01
4.94014591e-01 -6.19491518e-01 -4.80634093e-01 2.37327013e-02
9.44573730e-02 -7.64091432e-01 -7.36225128e-01 5.84888935e-01
-6.75679803e-01 -7.97911584e-01 -1.06775141e+00 -2.09470302e-01
7.46606588e-01 -1.95421517e-01 6.42210126e-01 -1.77225709e-01
-6.76697940e-02 5.05756438e-01 -1.63957372e-01 -7.19280481e-01
-2.75095887e-02 -5.17050505e-01 2.32299864e-01 5.39557412e-02
-2.20125690e-02 -6.73241973e-01 -5.98553181e-01 2.29762033e-01
-9.12789345e-01 -7.04364657e-01 9.84593183e-02 8.08403134e-01
9.10181403e-01 7.62330234e-01 2.83048868e-01 -4.86102283e-01
5.29754400e-01 -2.06471562e-01 -1.35358787e+00 -1.69314239e-02
-7.64234424e-01 1.78599507e-02 1.04293871e+00 -4.01367098e-01
-6.49406433e-01 -2.72441506e-01 -3.11614990e-01 -4.73243117e-01
1.83883488e-01 4.52142239e-01 2.98479162e-02 -6.09617606e-02
5.58842480e-01 4.79367197e-01 -1.07101828e-01 -6.06129169e-01
1.38077661e-01 4.57301021e-01 6.74486995e-01 -8.82083595e-01
8.96652818e-01 5.40497482e-01 4.64654952e-01 -6.23319268e-01
-9.00002122e-01 -9.09298658e-02 4.15984467e-02 -6.25012070e-02
4.81682956e-01 -4.04884905e-01 -1.52408385e+00 1.31082445e-01
-5.59406579e-01 -2.16246620e-01 -6.10071421e-01 6.60120606e-01
-9.61719990e-01 4.71764624e-01 -4.87823963e-01 -1.79749370e+00
-3.08841556e-01 -1.01446688e+00 5.32874882e-01 3.38534772e-01
2.62251973e-01 -8.14261913e-01 -2.86942840e-01 1.92248508e-01
6.59618974e-01 2.94791818e-01 7.32148528e-01 -3.27561051e-01
-2.24523902e-01 -5.76449871e-01 -3.67323190e-01 8.32063794e-01
-1.86317354e-01 3.48912901e-03 -4.64139134e-01 -2.85844505e-01
6.44020677e-01 2.76665002e-01 5.77340066e-01 7.48060703e-01
1.45095789e+00 -6.80778503e-01 -2.54580844e-02 5.42567313e-01
1.59784079e+00 2.53252983e-01 5.36928117e-01 6.31926432e-02
-8.26896131e-02 3.63119125e-01 3.90211761e-01 8.48778486e-01
3.09733022e-02 4.17082250e-01 5.18601537e-01 4.23347861e-01
4.04140770e-01 1.03328899e-01 4.43685025e-01 5.28469920e-01
-2.37288654e-01 -2.26196066e-01 -6.53939128e-01 6.32859588e-01
-1.45742917e+00 -9.70996916e-01 -3.54199141e-01 2.88101459e+00
1.07303822e+00 3.95838439e-01 2.27686435e-01 3.44115317e-01
7.75732398e-01 -1.64104804e-01 -2.43564263e-01 -5.22907019e-01
-3.56470525e-01 8.95747185e-01 8.77378702e-01 6.77941084e-01
-7.01566994e-01 3.02551687e-01 4.45587111e+00 1.48075485e+00
-8.32289934e-01 2.43961409e-01 7.53297031e-01 -3.83631170e-01
-3.14519107e-01 4.89866585e-02 -7.38844931e-01 1.25663018e+00
9.85731363e-01 -5.41303635e-01 3.25031936e-01 9.14008498e-01
2.22244889e-01 -5.04506230e-01 -6.37138188e-01 1.06438637e+00
-1.63507834e-01 -8.35393310e-01 -7.10646689e-01 4.03910041e-01
6.86284602e-01 -3.38732779e-01 4.47745621e-01 1.33574605e-01
4.20195103e-01 -1.08891344e+00 5.60637593e-01 2.27808580e-01
8.21645260e-01 -1.02163363e+00 8.28044295e-01 5.21148682e-01
-1.19800901e+00 -3.08545291e-01 -2.79549778e-01 -4.89951894e-02
4.41571951e-01 1.22521710e+00 -8.63809697e-03 6.76313400e-01
1.11226368e+00 -2.19230220e-01 5.00439525e-01 1.06718171e+00
-1.42424628e-02 6.88074350e-01 -1.01212931e+00 -2.03635797e-01
2.36442044e-01 -3.35278392e-01 9.02430952e-01 1.11504245e+00
8.15591097e-01 4.19832170e-01 1.22302976e-02 6.58776343e-01
-9.41758081e-02 2.64193714e-01 -1.68077111e-01 2.15162009e-01
3.32205623e-01 7.28768826e-01 -8.16392779e-01 -9.89424810e-02
-1.70991898e-01 5.53819776e-01 -1.38353825e-01 4.25011307e-01
-9.64382410e-01 -9.40185308e-01 8.12758267e-01 2.51504660e-01
6.81479871e-01 -3.66270423e-01 -6.00877404e-01 -7.20160663e-01
5.54560065e-01 -3.80498916e-01 4.36709285e-01 -9.32915583e-02
-1.34019303e+00 1.82807088e-01 5.84974512e-02 -9.24994409e-01
-2.59685423e-03 -6.61566377e-01 -3.41996431e-01 1.23343730e+00
-1.08985686e+00 -3.05418760e-01 2.93803245e-01 6.84462428e-01
-1.04086943e-01 -6.35805214e-03 5.79639256e-01 3.67295861e-01
-4.63355303e-01 9.92893457e-01 7.14679360e-01 -2.15079505e-02
2.11608559e-01 -1.07567716e+00 -4.05533969e-01 9.17647719e-01
-4.00908887e-01 4.24049616e-01 1.11451888e+00 -3.79264712e-01
-1.36358953e+00 -6.50042057e-01 7.83977270e-01 -3.25831473e-01
8.34437490e-01 -1.70770973e-01 -6.62125468e-01 2.41789177e-01
-3.15819532e-01 2.12668151e-01 6.52928829e-01 -2.35172391e-01
-1.40169293e-01 -2.94869274e-01 -1.55124795e+00 5.63087225e-01
8.76424670e-01 -2.96824276e-01 5.93123883e-02 3.43785465e-01
4.75667447e-01 -2.75641143e-01 -1.15395558e+00 5.78740001e-01
4.10418510e-01 -1.07780337e+00 6.80426598e-01 -3.99508983e-01
4.56545502e-02 -3.18125278e-01 -6.82543993e-01 -6.41107738e-01
1.73203439e-01 -1.14945245e+00 1.01882983e-02 9.79205132e-01
6.50629640e-01 -8.17904234e-01 7.43134618e-01 7.17266917e-01
6.34914543e-03 -9.88861799e-01 -1.76564550e+00 -1.08627796e+00
7.47824132e-01 -9.14574564e-01 -1.22907132e-01 4.95947957e-01
-2.52520815e-02 -4.92175132e-01 -2.73810148e-01 1.57834768e-01
8.45773816e-01 -3.51532161e-01 3.33871484e-01 -9.43289161e-01
-7.67658532e-01 -7.67600954e-01 -1.69976264e-01 -1.05160975e+00
-2.57801056e-01 -4.09471154e-01 2.95563161e-01 -9.75436687e-01
1.65204555e-01 -8.22013021e-01 -5.88729858e-01 1.02262780e-01
-1.99277103e-01 3.76734063e-02 1.92772746e-01 -1.55968368e-01
-5.09632766e-01 5.53538442e-01 1.05424988e+00 7.08422363e-02
-8.11221302e-02 5.18376887e-01 -8.48611355e-01 7.28107214e-01
4.70474839e-01 -5.70564926e-01 -2.64705330e-01 -2.60360539e-01
5.16673863e-01 4.13888782e-01 5.08044064e-02 -8.93795967e-01
3.29360753e-01 -2.41604030e-01 -1.70378134e-01 -3.55603576e-01
2.66264886e-01 -4.59115773e-01 3.32233310e-03 3.78605753e-01
-3.62821221e-02 -1.46535560e-02 1.36745632e-01 6.29657269e-01
1.94210440e-01 -5.53100467e-01 1.02262855e+00 -8.90987962e-02
4.03778791e-01 3.08625787e-01 -4.53315169e-01 3.30838263e-01
1.04700291e+00 -1.29626505e-02 -2.47916266e-01 -7.60455132e-01
-6.86957836e-01 2.47701257e-01 2.03014433e-01 -3.70980442e-01
1.07388198e-01 -8.48547161e-01 -8.28007281e-01 -1.16753921e-01
-5.03841102e-01 3.34634781e-01 5.59427500e-01 1.32969880e+00
-3.16723645e-01 9.58481878e-02 4.71565127e-01 -6.12396359e-01
-6.63143933e-01 5.16450524e-01 3.71105075e-01 -4.33975786e-01
-2.08411306e-01 1.40708864e+00 -1.91185325e-01 1.15134604e-01
3.48392278e-01 -1.94160938e-01 6.03778481e-01 -2.92885244e-01
6.12767816e-01 6.59653842e-01 -4.29063737e-02 -2.34697878e-01
-4.12639230e-01 6.53959870e-01 1.94781005e-01 -6.45152450e-01
1.16125834e+00 -2.39564985e-01 -2.53704786e-01 4.59072262e-01
1.34504032e+00 4.21970159e-01 -1.31950569e+00 -2.93612964e-02
-1.63438186e-01 -4.54237282e-01 -3.37361097e-01 -5.71099281e-01
-1.03845191e+00 9.79494691e-01 7.34974861e-01 5.10188937e-01
1.24478745e+00 -7.22185615e-03 5.96255124e-01 -1.60849720e-01
6.56914115e-01 -1.45104957e+00 -1.68676317e-01 6.73018634e-01
5.22645235e-01 -9.68108594e-01 -1.40316367e-01 -2.59759218e-01
-2.96487302e-01 9.45751190e-01 9.66797993e-02 -6.31982535e-02
1.10199952e+00 3.66803527e-01 -3.46102059e-01 9.85423028e-02
-1.59565777e-01 -1.84427679e-01 -2.41547544e-02 3.02134275e-01
4.03495550e-01 1.12974316e-01 -7.87547231e-01 1.10632193e+00
-4.31539446e-01 -1.17753796e-01 3.31738263e-01 8.57675493e-01
-5.46089232e-01 -9.59814787e-01 -4.27907646e-01 5.18202484e-01
-1.10407186e+00 -8.28582048e-02 4.60352570e-01 5.73433936e-01
1.08126074e-01 1.19526279e+00 -1.64675564e-01 5.04731759e-03
2.23928645e-01 8.01570043e-02 4.16147739e-01 -1.24217160e-01
-2.83059150e-01 2.84413874e-01 -2.48853981e-01 -4.04456198e-01
-2.67880596e-02 -5.98616123e-01 -1.21811867e+00 -6.10465407e-01
-1.41683817e-01 5.61639190e-01 8.76400471e-01 1.01629233e+00
-3.31995785e-02 1.90981627e-01 7.85270929e-01 -3.22137356e-01
-8.78765404e-01 -7.63039231e-01 -1.10991704e+00 1.05111100e-01
1.18589498e-01 -5.97012341e-01 -9.04356301e-01 -2.92144865e-01] | [6.463344097137451, 4.528484344482422] |
b9ccc75c-6f8b-4d3c-ba4d-d5df64d1a1d4 | 3d-action-recognition-from-novel-viewpoints | null | null | http://openaccess.thecvf.com/content_cvpr_2016/html/Rahmani_3D_Action_Recognition_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/Rahmani_3D_Action_Recognition_CVPR_2016_paper.pdf | 3D Action Recognition From Novel Viewpoints | We propose a human pose representation model that transfers human poses acquired from different unknown views to a view-invariant high-level space. The model is a deep convolutional neural network and requires a large corpus of multiview training data which is very expensive to acquire. Therefore, we propose a method to generate this data by fitting synthetic 3D human models to real motion capture data and rendering the human poses from numerous viewpoints. While learning the CNN model, we do not use action labels but only the pose labels after clustering all training poses into k clusters. The proposed model is able to generalize to real depth images of unseen poses without the need for re-training or fine-tuning. Real depth videos are passed through the model frame-wise to extract view-invariant features. For spatio-temporal representation, we propose group sparse Fourier Temporal Pyramid which robustly encodes the action specific most discriminative output features of the proposed human pose model. Experiments on two multiview and three single-view benchmark datasets show that the proposed method dramatically outperforms existing state-of-the-art in action recognition. | ['Ajmal Mian', 'Hossein Rahmani'] | 2016-06-01 | null | null | null | cvpr-2016-6 | ['3d-human-action-recognition'] | ['computer-vision'] | [ 2.43258998e-01 -1.42888948e-01 2.23230906e-02 -4.59809065e-01
-6.77271247e-01 -4.20520067e-01 5.57466686e-01 -6.52381182e-01
-2.74032921e-01 5.10169625e-01 3.42577398e-01 5.05789042e-01
1.66596249e-01 -5.88504970e-01 -1.01570427e+00 -4.81778741e-01
1.00320041e-01 6.55514896e-01 3.95399004e-01 -1.31279737e-01
4.61053774e-02 4.93042976e-01 -1.86415291e+00 6.00752115e-01
1.83853745e-01 9.00967777e-01 1.73574328e-01 6.22175217e-01
6.02626801e-01 9.43580747e-01 -4.33467358e-01 -3.47119421e-02
6.26755536e-01 -4.15752232e-01 -6.06752336e-01 8.26641262e-01
8.91288638e-01 -7.93118119e-01 -6.49259567e-01 7.43614078e-01
5.38082957e-01 3.86756659e-01 4.87057835e-01 -1.27912843e+00
-3.23945433e-01 -1.56194746e-01 -5.03973663e-01 5.32636829e-02
8.53212059e-01 2.04428688e-01 7.04063177e-01 -1.10148716e+00
1.13146794e+00 1.44306743e+00 3.56837958e-01 6.96078241e-01
-8.86690557e-01 -3.16531301e-01 2.16064453e-01 5.12970626e-01
-1.25262511e+00 -2.93633223e-01 9.20342922e-01 -6.98534310e-01
1.02248526e+00 7.53721222e-02 1.15947318e+00 1.38773549e+00
2.82947987e-01 9.81436431e-01 8.90890718e-01 -2.55000323e-01
1.31604046e-01 -4.58814502e-01 -3.43715042e-01 9.22429264e-01
2.88067106e-02 -2.48499271e-02 -7.97688007e-01 -3.18760984e-02
1.18250191e+00 6.02144301e-01 -4.79311675e-01 -1.10554934e+00
-1.58442652e+00 6.50448799e-01 2.80723840e-01 -1.05845980e-01
-6.18099928e-01 2.62741655e-01 4.13049579e-01 -4.52587977e-02
1.75318435e-01 -1.32613778e-01 -5.17585993e-01 1.14814728e-01
-6.18511081e-01 3.82876217e-01 1.82306737e-01 8.91867340e-01
7.18904495e-01 1.50797099e-01 -2.82020643e-02 4.33874875e-01
2.11551964e-01 4.85164225e-01 6.23772204e-01 -1.38439739e+00
5.99009275e-01 7.79665768e-01 3.03114653e-01 -1.14014506e+00
-4.42370087e-01 -6.12327605e-02 -8.28907251e-01 2.26630084e-02
3.01866621e-01 1.67612299e-01 -9.37576771e-01 1.44718587e+00
5.82506239e-01 1.68017685e-01 -1.95000082e-01 1.09282565e+00
6.62777364e-01 5.21427214e-01 -3.17772925e-01 -1.98121622e-01
1.28479338e+00 -1.06058097e+00 -5.94679475e-01 -2.15731144e-01
3.13198507e-01 -2.90683120e-01 7.34460354e-01 6.69949412e-01
-1.19452751e+00 -1.00013793e+00 -1.04772079e+00 -1.73226148e-01
-5.28722852e-02 3.45452666e-01 4.93137687e-01 2.46322751e-01
-8.15909386e-01 6.65196300e-01 -1.13693810e+00 -2.40885153e-01
2.44819045e-01 3.02883893e-01 -1.01976609e+00 -2.83521295e-01
-1.01773179e+00 6.01159871e-01 3.67211640e-01 2.26132408e-01
-1.42240775e+00 -1.76619396e-01 -1.27970719e+00 -3.18744153e-01
4.19504553e-01 -1.08982158e+00 9.67265129e-01 -9.31196213e-01
-1.46023059e+00 8.00685763e-01 -2.36376449e-01 -1.22239515e-01
7.88411498e-01 -5.04050493e-01 -9.47024077e-02 5.76722682e-01
1.85753807e-01 6.58275843e-01 1.15594959e+00 -1.32152617e+00
-5.39251208e-01 -7.68681407e-01 1.68766886e-01 5.41487336e-01
-1.00784510e-01 -3.45564455e-01 -6.09966576e-01 -5.31177163e-01
6.07288837e-01 -1.10374987e+00 -2.05330253e-01 4.68192399e-02
-1.69430494e-01 -5.63035533e-02 7.85034597e-01 -7.94353306e-01
6.10003352e-01 -1.77697802e+00 8.00220132e-01 -4.30503078e-02
1.05744444e-01 3.70289683e-02 5.73385209e-02 3.65264058e-01
-2.02186272e-01 -5.43019950e-01 1.00849465e-01 -4.05850142e-01
-2.08053097e-01 4.70153600e-01 5.23452125e-02 9.11635578e-01
-6.73090070e-02 8.53006244e-01 -8.42170954e-01 -4.80892479e-01
7.33641863e-01 6.76478088e-01 -7.85019577e-01 5.50073504e-01
-1.03109658e-01 9.67736304e-01 -3.69451433e-01 5.28565943e-01
3.96721333e-01 -4.81254607e-01 1.80311844e-01 -5.69801867e-01
2.36254051e-01 -1.15487106e-01 -1.42223001e+00 2.37784314e+00
-1.65878400e-01 1.12497084e-01 -2.72715211e-01 -1.21085310e+00
5.65468848e-01 4.13413137e-01 7.25183666e-01 -3.26848179e-01
2.08387807e-01 -1.32274002e-01 -5.40161014e-01 -7.89638340e-01
2.79139727e-01 -6.06220253e-02 -1.62317336e-01 1.99261427e-01
3.93570453e-01 1.41821235e-01 -9.01757628e-02 -4.10386361e-02
8.78696918e-01 7.45655060e-01 5.40355325e-01 3.28956932e-01
8.65016818e-01 -2.41428480e-01 8.31836641e-01 1.03614256e-01
-2.36881688e-01 8.37814867e-01 1.02430768e-01 -9.26706016e-01
-9.78517056e-01 -1.28332806e+00 4.24145192e-01 9.50408518e-01
6.44925758e-02 -2.81817436e-01 -6.50669873e-01 -7.09014595e-01
-2.34223485e-01 4.17070165e-02 -8.31808209e-01 2.42595337e-02
-9.30084765e-01 -7.58925900e-02 1.40607223e-01 6.65400565e-01
6.61955714e-01 -1.03181827e+00 -1.17179680e+00 -2.47456953e-02
-7.27812827e-01 -1.34351730e+00 -4.43037391e-01 -2.50763446e-01
-8.86471212e-01 -1.22025681e+00 -9.25333500e-01 -7.04589546e-01
7.74835289e-01 2.33870372e-01 9.42276478e-01 -3.44594568e-01
-4.42363828e-01 8.67804945e-01 -4.36689645e-01 1.34216234e-01
-3.76244523e-02 -5.28158426e-01 4.14881200e-01 4.79943454e-01
1.73741460e-01 -6.63727701e-01 -8.87423754e-01 4.12896603e-01
-7.00008929e-01 8.78804103e-02 4.59732175e-01 9.45320487e-01
8.13737631e-01 -3.20065588e-01 2.58962154e-01 -5.34627438e-01
-1.87964380e-01 -1.59726124e-02 -3.03556830e-01 1.85420856e-01
3.18034858e-01 -1.40173584e-01 5.23304164e-01 -4.36501145e-01
-8.96877468e-01 8.09641063e-01 1.86172977e-01 -1.06909049e+00
-4.67990220e-01 2.81688012e-02 -4.55202788e-01 1.17850132e-01
6.32340372e-01 4.14152265e-01 -7.95066077e-03 -3.36581618e-01
4.60058808e-01 1.96109965e-01 7.29710758e-01 -3.90663385e-01
7.35169232e-01 9.38227296e-01 1.44497171e-01 -7.60103166e-01
-8.55146587e-01 -5.74766994e-01 -1.38313317e+00 -6.20384097e-01
1.35390127e+00 -1.32945228e+00 -6.41322076e-01 4.33684856e-01
-1.24067473e+00 -1.16522852e-02 -1.33534327e-01 6.07254386e-01
-1.28193438e+00 7.50973523e-01 -5.24097919e-01 -5.17009795e-01
-7.25272745e-02 -1.18493795e+00 1.61252820e+00 -4.25121814e-01
-1.82100952e-01 -6.59744740e-01 5.08559644e-02 8.33005726e-01
-3.32462788e-01 8.73165309e-01 3.80585641e-01 -2.34665886e-01
-7.82343090e-01 -2.85028070e-01 3.51471007e-01 5.44433057e-01
1.28972456e-01 -3.73269945e-01 -8.35008323e-01 -4.40418571e-01
2.14981467e-01 -7.70097256e-01 6.88804984e-01 4.85511601e-01
1.15323865e+00 -2.14025438e-01 -2.33549550e-01 6.76535070e-01
1.25576293e+00 -5.87392002e-02 6.91634238e-01 1.48738146e-01
1.12813032e+00 6.99985504e-01 7.75603592e-01 7.73663402e-01
4.35775608e-01 8.61582935e-01 5.88521838e-01 1.87825605e-01
2.24624142e-01 -3.78296018e-01 6.12770915e-01 8.93119514e-01
-6.25401556e-01 2.02756509e-01 -6.80939257e-01 3.71059448e-01
-1.94516742e+00 -1.29580247e+00 1.50728285e-01 2.19770813e+00
3.46355140e-01 -8.61734673e-02 2.17018813e-01 3.41607481e-01
4.61688340e-01 2.40466312e-01 -5.22131443e-01 2.85758436e-01
3.21125299e-01 7.77770355e-02 1.00882642e-01 3.44908804e-01
-1.26615381e+00 8.19333732e-01 5.09570217e+00 3.01549733e-01
-9.72671866e-01 -1.31716970e-02 1.48472458e-01 -2.33334869e-01
1.27096310e-01 -3.16932231e-01 -4.27085549e-01 -9.59743708e-02
4.82032239e-01 1.88195512e-01 1.22158207e-01 9.77424145e-01
2.73997813e-01 8.13166127e-02 -1.51970279e+00 1.52879310e+00
5.85236490e-01 -1.09716892e+00 4.45829600e-01 2.03355253e-02
8.35574329e-01 -2.03904644e-01 -7.27233589e-02 1.55997485e-01
-8.56983941e-03 -9.30851758e-01 8.37623000e-01 8.36904705e-01
5.93984306e-01 -7.62914181e-01 3.67200881e-01 6.87695205e-01
-1.34490263e+00 -1.73658028e-01 -5.59597969e-01 -1.64159209e-01
4.62387085e-01 5.24807423e-02 -4.20193076e-01 7.97815561e-01
8.12205315e-01 1.24876356e+00 -5.86078644e-01 4.81451958e-01
-3.07374522e-02 -2.68253863e-01 6.60475492e-02 5.51407158e-01
-1.59286261e-02 -2.37521715e-02 3.44731778e-01 5.45594871e-01
3.86415750e-01 2.19622418e-01 5.15678227e-01 3.92682612e-01
3.58999223e-01 -1.66609734e-01 -9.17156100e-01 2.95161098e-01
-1.12107597e-01 1.11650658e+00 -5.06702006e-01 -5.04034877e-01
-4.76216912e-01 1.60940564e+00 2.91227728e-01 3.77631903e-01
-8.93584669e-01 4.03717846e-01 4.54207629e-01 2.33056486e-01
5.88803411e-01 -4.79261905e-01 4.47700471e-01 -1.62224138e+00
1.70414731e-01 -1.11539936e+00 3.56490046e-01 -8.79353762e-01
-1.11179471e+00 4.99245912e-01 2.82654464e-01 -1.87151563e+00
-7.50694752e-01 -7.62370050e-01 2.14567911e-02 4.94904190e-01
-8.78012359e-01 -1.38086891e+00 -6.96127594e-01 1.24257982e+00
9.43382144e-01 -1.90525994e-01 7.86130488e-01 1.63170770e-01
-1.98237151e-02 1.74612895e-01 -4.15570736e-01 2.27748677e-01
8.00538242e-01 -9.08269525e-01 3.56372088e-01 5.23385763e-01
1.93685010e-01 5.15933335e-01 4.93140846e-01 -5.54117680e-01
-1.60763896e+00 -1.18382013e+00 4.91917402e-01 -9.58540559e-01
5.36387525e-02 -3.27544659e-01 -4.48469460e-01 1.01211238e+00
-4.86129373e-02 4.81942981e-01 6.47354901e-01 -5.19991815e-01
-1.89841732e-01 7.77615681e-02 -9.96206522e-01 3.92445445e-01
1.36425209e+00 -4.10282552e-01 -7.54018307e-01 3.29899371e-01
4.21777427e-01 -5.17658710e-01 -8.93936992e-01 5.66429496e-01
8.23172152e-01 -1.34454596e+00 1.33250141e+00 -7.75761604e-01
5.14504850e-01 -4.19409096e-01 -5.90777636e-01 -1.07598424e+00
-3.39990437e-01 -1.34154662e-01 -2.88944066e-01 4.58472848e-01
-2.81391293e-01 -2.57390767e-01 9.50078845e-01 1.76675469e-01
2.11567730e-02 -6.70921803e-01 -1.00285244e+00 -7.16585219e-01
-3.16941023e-01 -2.61720270e-01 1.04112297e-01 9.14209187e-01
-2.88408369e-01 2.83436924e-01 -7.82196462e-01 3.69723529e-01
9.71080840e-01 3.25840823e-02 1.28714395e+00 -1.22276354e+00
-4.83639985e-01 4.07880276e-01 -9.16695297e-01 -1.17910147e+00
2.44051918e-01 -5.33470511e-01 -1.38016269e-01 -1.58995116e+00
1.64788097e-01 5.10891676e-01 9.24820229e-02 2.67216802e-01
3.78197990e-02 4.20932621e-01 1.25497788e-01 1.99782804e-01
-7.80319273e-01 8.51727664e-01 1.63498175e+00 -1.32569456e-02
4.30264696e-02 -1.22140296e-01 1.56390101e-01 1.13820362e+00
3.94256443e-01 -1.40344009e-01 -7.59228945e-01 -2.57065535e-01
3.91323119e-02 5.86682498e-01 7.05524087e-01 -1.33479190e+00
4.33472544e-02 -3.32584321e-01 9.57878709e-01 -9.57553327e-01
9.24101651e-01 -9.58880305e-01 4.39864695e-01 5.51972926e-01
-1.18327014e-01 3.15318316e-01 -2.19322905e-01 1.00263083e+00
-1.81184515e-01 2.87402511e-01 7.17724442e-01 -7.50069022e-01
-1.00591278e+00 6.11862242e-01 -1.35473549e-01 8.20549801e-02
1.20352435e+00 -3.17547649e-01 1.91383436e-01 -5.93698978e-01
-1.23738348e+00 -2.57245377e-02 6.43569171e-01 5.79750419e-01
1.03591430e+00 -1.65009046e+00 -4.85442996e-01 6.02244198e-01
2.49092311e-01 1.59451321e-01 6.83242381e-01 5.70399523e-01
-6.46915674e-01 3.51599067e-01 -7.02686012e-01 -1.08106732e+00
-1.35126841e+00 7.91930258e-01 4.08894241e-01 -3.30943614e-02
-9.20004308e-01 5.83041489e-01 5.39867997e-01 -6.16144300e-01
3.04763436e-01 -2.82055318e-01 -3.15505564e-01 -3.10806446e-02
5.89963734e-01 4.21674103e-01 -2.28817105e-01 -1.26843905e+00
-3.62315595e-01 8.98914516e-01 2.39562884e-01 -2.77947307e-01
1.30228686e+00 -6.75670505e-02 1.78424604e-02 6.38619900e-01
1.36946321e+00 -4.86776471e-01 -1.58058381e+00 -1.33240148e-01
-5.38592517e-01 -7.82011926e-01 -5.24042904e-01 -5.90391196e-02
-1.06446970e+00 1.05413032e+00 6.02114856e-01 -5.28878510e-01
1.05498540e+00 -1.11289524e-01 6.23529255e-01 6.33569539e-01
9.19101357e-01 -1.19454265e+00 8.49805295e-01 3.65238816e-01
1.29201508e+00 -1.12044561e+00 2.78889745e-01 -3.47709209e-01
-7.60882974e-01 1.05154705e+00 9.46962059e-01 -5.07677853e-01
6.11239374e-01 -3.52265537e-01 2.25248211e-03 -2.82507241e-01
-6.80516243e-01 -4.46758932e-03 4.04826164e-01 7.48059511e-01
1.04319103e-01 -7.44253844e-02 -4.10926491e-02 3.03771466e-01
-5.69886528e-02 1.17501177e-01 5.10758519e-01 9.99954343e-01
-3.97811055e-01 -9.13262904e-01 -4.02012557e-01 9.30031613e-02
-2.78851092e-01 4.92846251e-01 -3.05178672e-01 7.60235608e-01
3.08314532e-01 5.59271574e-01 -4.19084392e-02 -5.92966616e-01
6.45690084e-01 3.19435924e-01 9.38382208e-01 -7.27884591e-01
-1.82035983e-01 2.09071293e-01 -1.46506414e-01 -1.16040468e+00
-9.86313105e-01 -8.81078482e-01 -1.09216487e+00 1.15208499e-01
1.27951279e-01 -4.29830581e-01 1.28755510e-01 9.08705890e-01
8.07150751e-02 5.03655374e-01 4.89890873e-01 -1.74756896e+00
-5.88461161e-01 -7.95524359e-01 -5.13033986e-01 8.97382319e-01
4.24363226e-01 -1.21464860e+00 -2.27334037e-01 4.56935406e-01] | [7.1799750328063965, -0.6504600644111633] |
a842b6aa-8a55-4a03-af88-fdbea3c01191 | analyzing-intentional-behavior-in-autonomous | 2307.01532 | null | https://arxiv.org/abs/2307.01532v1 | https://arxiv.org/pdf/2307.01532v1.pdf | Analyzing Intentional Behavior in Autonomous Agents under Uncertainty | Principled accountability for autonomous decision-making in uncertain environments requires distinguishing intentional outcomes from negligent designs from actual accidents. We propose analyzing the behavior of autonomous agents through a quantitative measure of the evidence of intentional behavior. We model an uncertain environment as a Markov Decision Process (MDP). For a given scenario, we rely on probabilistic model checking to compute the ability of the agent to influence reaching a certain event. We call this the scope of agency. We say that there is evidence of intentional behavior if the scope of agency is high and the decisions of the agent are close to being optimal for reaching the event. Our method applies counterfactual reasoning to automatically generate relevant scenarios that can be analyzed to increase the confidence of our assessment. In a case study, we show how our method can distinguish between 'intentional' and 'accidental' traffic collisions. | ['Bettina Könighofer', 'Ruzica Piskac', 'Scott J. Shapiro', 'Nicholas Shoemaker', 'Katrine Bjørner', 'Timos Antonopoulos', 'Samuel Judson', 'Filip Cano Córdoba'] | 2023-07-04 | null | null | null | null | ['decision-making'] | ['reasoning'] | [ 3.75104994e-01 8.84975672e-01 -4.18032706e-02 -3.36010695e-01
-7.07830846e-01 -5.59332490e-01 8.99302900e-01 3.35533947e-01
-5.97309649e-01 9.99710858e-01 5.31235874e-01 -8.20181489e-01
-2.75487959e-01 -8.33566070e-01 -9.55397666e-01 -4.30222660e-01
-1.12648569e-01 5.32094419e-01 2.39687741e-01 1.86245754e-01
1.99359432e-01 4.53014851e-01 -1.18641245e+00 2.51594007e-01
7.18909264e-01 5.63516676e-01 -2.82620132e-01 6.95598781e-01
3.52484375e-01 1.40626836e+00 -5.28313935e-01 -4.26438898e-01
3.38990629e-01 -2.90234715e-01 -7.71216691e-01 -9.17244405e-02
-4.75889683e-01 -7.30842292e-01 -8.11371058e-02 1.23780203e+00
6.37922958e-02 -5.27349152e-02 8.64423811e-01 -1.69070482e+00
2.13862602e-02 1.10152936e+00 9.53377634e-02 -5.47419190e-02
6.99555933e-01 8.12151372e-01 6.20230079e-01 1.36830941e-01
5.90362132e-01 1.64983940e+00 3.00796062e-01 6.09986484e-01
-1.13671589e+00 -4.30829674e-01 1.97273657e-01 1.10129453e-01
-9.90071893e-01 -4.36562151e-01 5.68868160e-01 -7.04781473e-01
7.55730987e-01 2.09499910e-01 5.14886260e-01 1.01746523e+00
8.23963821e-01 4.71170843e-01 1.51854670e+00 -3.87712270e-01
8.69229436e-01 -9.26555693e-02 8.53977725e-02 2.06970885e-01
9.41334128e-01 8.84507358e-01 -1.40930489e-01 -6.96360111e-01
1.14923522e-01 -1.31287113e-01 1.97264686e-01 -2.04644993e-01
-1.16276753e+00 5.17166734e-01 -1.95340559e-01 -2.62709379e-01
-9.90595043e-01 6.52958274e-01 2.84340203e-01 1.38977513e-01
-2.63000935e-01 5.24318397e-01 -2.31599197e-01 -2.01992229e-01
-1.66841835e-01 8.41517925e-01 7.91551709e-01 5.96117258e-01
1.63611785e-01 -2.18668014e-01 -4.07667518e-01 -4.08250779e-01
6.22347057e-01 7.68269479e-01 -6.66896328e-02 -1.68262863e+00
4.07507867e-01 4.87425476e-01 1.00547528e+00 -4.53339785e-01
-1.56642899e-01 1.88722000e-01 4.22630049e-02 9.73761559e-01
5.33841968e-01 -4.71450061e-01 -5.84312320e-01 1.90568352e+00
1.72303051e-01 -6.42210841e-02 5.34678400e-01 4.17466819e-01
-1.65982321e-01 3.80090982e-01 5.11494458e-01 -4.34416890e-01
1.07435310e+00 1.14572257e-01 -8.20402384e-01 -1.08298875e-01
6.94954157e-01 -3.13887522e-02 6.34507477e-01 2.89163858e-01
-1.09750187e+00 3.60143900e-01 -9.55831051e-01 6.00552320e-01
3.25500757e-01 -9.35778320e-01 5.17064154e-01 5.73183596e-01
-4.19938207e-01 4.13283855e-01 -8.04520369e-01 9.95571986e-02
5.19795895e-01 -1.22941574e-02 -1.98976353e-01 1.61766037e-02
-1.33102643e+00 1.15048552e+00 3.37138295e-01 -2.63350666e-01
-1.72283399e+00 -2.12151334e-01 -8.03218126e-01 9.91757736e-02
7.38491833e-01 -5.42251408e-01 1.50263202e+00 -6.16368771e-01
-8.68409514e-01 6.10781193e-01 1.66648868e-02 -1.03768194e+00
1.05426323e+00 1.21681750e-01 -4.99732405e-01 2.43836287e-02
4.82176572e-01 6.28750250e-02 4.69943583e-01 -1.36566150e+00
-6.74752116e-01 -4.05176073e-01 6.44362450e-01 6.85261488e-02
5.96875429e-01 4.48455542e-01 7.18546212e-01 8.41342881e-02
-1.96267307e-01 -1.22020197e+00 -7.74681926e-01 -2.74514765e-01
-6.29670322e-01 -1.76506937e-01 1.43508717e-01 -2.89006922e-02
1.00260973e+00 -1.98414445e+00 -8.56283069e-01 4.23641056e-01
1.52811468e-01 -3.01222295e-01 4.19281185e-01 2.32767388e-01
2.10610792e-01 3.99886549e-01 -3.60943943e-01 2.87959278e-01
4.15404379e-01 2.82787740e-01 -3.71703535e-01 6.71046853e-01
8.53187963e-02 6.08126998e-01 -7.76093781e-01 -4.97788191e-01
1.29419148e-01 -1.78449437e-01 -3.95150751e-01 2.49366030e-01
-3.03950250e-01 2.26829842e-01 -7.36665130e-01 3.41844916e-01
2.18095705e-01 4.33945358e-01 2.65547782e-01 5.79020143e-01
-2.95058131e-01 5.67318141e-01 -1.30032289e+00 6.04484677e-01
-2.12774083e-01 2.34266713e-01 -1.78924143e-01 -2.74930775e-01
2.65038311e-01 5.67182720e-01 -3.48655991e-02 -4.10974145e-01
2.65946984e-01 2.17771471e-01 3.65235209e-01 -6.41934514e-01
4.88182716e-02 -6.54668689e-01 -6.69411182e-01 1.00720370e+00
-7.21301138e-01 -1.70177832e-01 -1.70781627e-01 2.15728298e-01
1.56792212e+00 -1.90296080e-02 5.91896117e-01 -3.16971809e-01
4.99301925e-02 3.33240122e-01 9.47592676e-01 1.12182248e+00
-7.02793717e-01 -7.56111667e-02 1.17219007e+00 -4.24204141e-01
-9.63864267e-01 -1.09865940e+00 5.09820692e-02 3.50380152e-01
2.81948894e-01 1.85142979e-01 -9.23544526e-01 -7.32876956e-01
9.07027572e-02 1.78562832e+00 -7.83109605e-01 -4.02408034e-01
-2.39334047e-01 -4.45534378e-01 4.93862301e-01 3.75766426e-01
3.33848417e-01 -1.16502261e+00 -1.52654672e+00 1.62586153e-01
1.58456329e-03 -9.25815880e-01 -1.37743458e-01 -7.38255829e-02
-5.21585584e-01 -1.36516786e+00 2.99884737e-01 3.40967000e-01
7.42048562e-01 -1.66169077e-01 8.18563342e-01 -9.38146710e-02
3.91788423e-01 4.10149962e-01 -3.21846418e-02 -1.05085635e+00
-1.21868658e+00 -1.02333796e+00 4.83809173e-01 -3.03608775e-01
4.08485621e-01 -2.52613276e-01 -3.96084875e-01 2.40010574e-01
-8.80424917e-01 -2.95881808e-01 3.32372308e-01 1.85049534e-01
4.06417131e-01 5.39767504e-01 4.57068920e-01 -8.80530834e-01
1.05000257e+00 -6.30870283e-01 -7.88901329e-01 2.84119099e-01
-8.23836923e-01 3.19755822e-01 3.11987907e-01 -2.47010350e-01
-1.40819061e+00 3.79483611e-03 5.13168156e-01 9.45462212e-02
-6.28687561e-01 4.37159806e-01 -7.43951201e-01 6.29978299e-01
7.36243248e-01 -3.75533879e-01 -1.50394365e-01 1.20343983e-01
2.31636047e-01 5.63237488e-01 3.46470803e-01 -1.09726620e+00
5.25010943e-01 6.89127266e-01 3.12355310e-01 7.59174675e-02
-5.27106106e-01 2.69478381e-01 -4.44743745e-02 -6.49544060e-01
8.48672628e-01 -5.99629521e-01 -1.21689069e+00 1.22516476e-01
-1.06875837e+00 -5.16800880e-01 -6.42571867e-01 7.90560186e-01
-9.88890648e-01 4.33439724e-02 5.52777648e-02 -1.62839293e+00
3.36435378e-01 -1.41339684e+00 4.64995772e-01 -1.96745872e-01
-5.16542077e-01 -5.20135701e-01 1.89288095e-01 3.58589917e-01
2.99583399e-03 6.66224003e-01 6.56181872e-01 -1.00850773e+00
-6.42165184e-01 -6.01300478e-01 3.06405455e-01 6.21156022e-02
-2.42172554e-01 2.06484482e-01 -8.62154424e-01 2.69265056e-01
2.52299309e-01 -3.85539718e-02 7.28601515e-02 4.38581288e-01
5.23467720e-01 -1.06140280e+00 -3.93389821e-01 -4.26293164e-01
1.09555101e+00 7.78570175e-01 8.58480334e-01 7.23099828e-01
7.19772652e-02 1.10268462e+00 9.74441290e-01 4.84349847e-01
4.85970169e-01 6.78750813e-01 6.91059530e-01 6.09695494e-01
4.62936938e-01 -3.36198956e-01 6.37377083e-01 -6.09763324e-01
-1.94558352e-01 -1.02543034e-01 -1.11339712e+00 7.20140696e-01
-1.96781182e+00 -1.53804421e+00 -2.32175305e-01 2.50891352e+00
6.37497723e-01 7.38332093e-01 2.20457226e-01 5.28281480e-02
9.28403080e-01 -2.77867198e-01 -5.66873848e-01 -9.83911633e-01
2.61735260e-01 -6.32551432e-01 8.17871988e-01 8.01846147e-01
-4.51875210e-01 3.88571769e-01 6.61977577e+00 3.37436527e-01
-1.29094973e-01 2.26804450e-01 7.27857709e-01 -2.01212302e-01
-6.82426870e-01 4.73417044e-01 -5.11198819e-01 5.64697266e-01
1.41193271e+00 -6.74752533e-01 1.06681334e-02 8.01099300e-01
7.68603146e-01 -3.54460686e-01 -1.29533482e+00 -1.77470781e-02
-6.93093002e-01 -1.00735176e+00 -1.26348749e-01 3.53864044e-01
5.15541613e-01 -2.70929068e-01 -2.80071825e-01 1.59518465e-01
1.37876463e+00 -1.06654429e+00 1.58429134e+00 8.46526027e-01
5.06308496e-01 -1.02203488e+00 9.57684815e-01 6.59643412e-01
-3.48697513e-01 -4.56518382e-01 1.97849214e-01 -5.01636446e-01
4.81608152e-01 6.76812589e-01 -1.03049958e+00 -2.98017319e-02
2.51087338e-01 -3.99357229e-01 6.70883581e-02 7.28229523e-01
-6.46649122e-01 6.53475106e-01 -3.69304746e-01 -1.11580873e-03
1.84013724e-01 1.04625933e-02 9.62067008e-01 7.16537476e-01
-4.77480739e-02 3.32306296e-01 -9.12751779e-02 1.05039597e+00
2.98754990e-01 -7.33774722e-01 -1.07475293e+00 1.80580541e-01
7.56110668e-01 4.58570063e-01 -4.35348421e-01 -5.24305582e-01
7.47695416e-02 2.70805005e-02 -3.13750952e-01 1.65370390e-01
-7.03594089e-01 9.85597298e-02 6.73730195e-01 1.93197262e-02
-3.74818534e-01 1.30163088e-01 -5.27805150e-01 -5.87994576e-01
1.51843131e-01 -8.30663741e-01 2.58250594e-01 -1.00501323e+00
-6.95704281e-01 2.15582550e-01 4.33595628e-01 -1.15393579e+00
-5.85555136e-01 -2.69106269e-01 -8.49591017e-01 7.08963811e-01
-9.68655705e-01 -6.17193222e-01 3.18917155e-01 2.21620902e-01
-1.23929279e-02 1.97844476e-01 4.94543284e-01 -6.51041389e-01
-2.80518085e-01 -3.89041118e-02 -5.18895209e-01 -2.06632823e-01
1.89802721e-01 -1.27259672e+00 3.53185564e-01 1.25579357e+00
-6.60610080e-01 5.44535697e-01 1.37855875e+00 -1.09690094e+00
-1.08758175e+00 -1.11539781e+00 8.56057882e-01 -8.82529438e-01
8.03311884e-01 1.82518885e-01 -5.36775827e-01 1.01223862e+00
-1.75855175e-01 -3.19941282e-01 3.94601643e-01 -2.14825317e-01
-3.72278661e-01 1.92355305e-01 -1.82201922e+00 1.01707447e+00
1.02596438e+00 -4.48531121e-01 -1.29404688e+00 1.61452085e-01
8.50875616e-01 1.97067201e-01 -6.74401283e-01 3.78889620e-01
6.99708402e-01 -9.57981229e-01 4.55301642e-01 -9.91441071e-01
2.78720707e-01 -6.41731739e-01 -3.79461557e-01 -1.06006730e+00
-1.35984058e-02 -7.05821693e-01 1.74153373e-01 8.49529624e-01
5.88181973e-01 -8.87958109e-01 2.00274795e-01 1.73198283e+00
1.57328770e-02 -2.41097882e-01 -1.22108614e+00 -8.16403210e-01
1.43237114e-02 -1.01736891e+00 1.09260726e+00 7.30071127e-01
4.07021761e-01 -3.56474817e-01 -4.36857715e-03 6.02948844e-01
1.15373266e+00 -5.00591360e-02 5.13955235e-01 -1.02616501e+00
-9.27262083e-02 -1.29644245e-01 -5.03794134e-01 8.37444291e-02
2.50388801e-01 -2.65850723e-01 5.69157183e-01 -1.53553331e+00
4.47183222e-01 -3.51690143e-01 -1.14178091e-01 5.23225904e-01
-1.51611656e-01 -5.48768103e-01 2.19511807e-01 2.37202823e-01
-6.58997238e-01 9.88985896e-02 9.43940699e-01 -6.94046766e-02
5.23879640e-02 3.81071746e-01 -9.38849390e-01 1.01575351e+00
8.73447895e-01 -7.92411923e-01 -2.09203869e-01 1.66142866e-01
6.16800964e-01 6.02443993e-01 6.89180553e-01 -6.89138591e-01
8.05431232e-03 -9.45992827e-01 -3.33606005e-01 -4.58476692e-02
-2.62160063e-01 -1.06989276e+00 6.92265749e-01 9.68034148e-01
-1.01636016e+00 -1.73102483e-01 -9.74088982e-02 8.87144744e-01
1.88240543e-01 -4.08786625e-01 4.51217115e-01 -1.78470030e-01
-3.26613456e-01 -4.97842878e-02 -1.13446164e+00 -2.48505041e-01
1.41971004e+00 7.71203116e-02 -4.87484127e-01 -6.39646053e-01
-7.33141601e-01 2.88207471e-01 8.35712075e-01 -2.18519270e-01
4.67385590e-01 -1.03089666e+00 -7.84694731e-01 -5.05568862e-01
1.58244282e-01 -2.31081963e-01 1.70878489e-02 7.89350986e-01
-1.49506256e-01 3.28476012e-01 -1.84388161e-01 -6.82715923e-02
-8.67446840e-01 6.87851489e-01 5.49095929e-01 -1.30370736e-01
-3.37057859e-01 4.36777249e-02 3.35751802e-01 -1.04099043e-01
-1.27069890e-01 -3.29705894e-01 -1.09304853e-01 -5.08174062e-01
7.37432957e-01 7.12484241e-01 -2.46872589e-01 -6.81923330e-01
-5.65738142e-01 -2.41755307e-01 2.23660111e-01 -7.56521463e-01
9.82860565e-01 -1.41286612e-01 -1.24476627e-01 5.79675317e-01
2.72889405e-01 2.22354695e-01 -1.66858947e+00 3.46660227e-01
1.57192767e-01 -4.61128145e-01 -1.38271317e-01 -1.05549860e+00
-2.47872218e-01 3.20853621e-01 3.25276494e-01 6.45684481e-01
5.98308742e-01 -9.96877812e-03 3.18732187e-02 4.54741806e-01
9.84502077e-01 -1.09592545e+00 -7.13627398e-01 -3.79148349e-02
8.23023140e-01 -9.46601629e-01 1.28244758e-01 -1.92968231e-02
-1.02623963e+00 4.61124450e-01 1.20114110e-01 -4.52770591e-02
2.17788339e-01 5.15097082e-01 -8.73320829e-03 -2.63644546e-01
-1.04371309e+00 -1.08560048e-01 -2.34795690e-01 6.14267826e-01
-3.56607646e-01 8.98024559e-01 -4.24898207e-01 6.18525922e-01
-1.43611163e-01 2.96946257e-01 1.27643168e+00 9.80848730e-01
-6.14735544e-01 -5.46406627e-01 -7.71061957e-01 5.13067663e-01
-7.22788334e-01 2.26244122e-01 -5.02066255e-01 6.75016284e-01
1.08614616e-01 1.41302669e+00 -6.73648566e-02 -1.16508104e-01
5.90675235e-01 2.02456140e-03 6.62169233e-02 -5.42572260e-01
-1.47198364e-02 -4.56558794e-01 8.29800069e-01 -9.61134017e-01
-5.52962840e-01 -1.14582813e+00 -1.48104715e+00 -2.23805159e-01
3.33780497e-02 2.68239021e-01 5.82236052e-01 1.41536260e+00
1.16821155e-01 4.39834774e-01 6.28886044e-01 -2.15456516e-01
-9.20651793e-01 -5.22052765e-01 -4.75623101e-01 3.35811973e-01
4.13409442e-01 -7.29803681e-01 -8.63844037e-01 -1.36043236e-01] | [8.309354782104492, 5.85219669342041] |
91586084-5c3d-4c85-90ea-4754a85de6fa | region-of-interest-focused-mri-to-synthetic | 2203.16288 | null | https://arxiv.org/abs/2203.16288v2 | https://arxiv.org/pdf/2203.16288v2.pdf | Region of Interest focused MRI to Synthetic CT Translation using Regression and Classification Multi-task Network | In this work, we present a method for synthetic CT (sCT) generation from zero-echo-time (ZTE) MRI aimed at structural and quantitative accuracies of the image, with a particular focus on the accurate bone density value prediction. We propose a loss function that favors a spatially sparse region in the image. We harness the ability of a multi-task network to produce correlated outputs as a framework to enable localisation of region of interest (RoI) via classification, emphasize regression of values within RoI and still retain the overall accuracy via global regression. The network is optimized by a composite loss function that combines a dedicated loss from each task. We demonstrate how the multi-task network with RoI focused loss offers an advantage over other configurations of the network to achieve higher accuracy of performance. This is relevant to sCT where failure to accurately estimate high Hounsfield Unit values of bone could lead to impaired accuracy in clinical applications. We compare the dose calculation maps from the proposed sCT and the real CT in a radiation therapy treatment planning setup. | ['Bjoern Menze', 'Florian Wiesinger', 'Tufve Nyholm', 'Joakim Jonsson', 'Kesavadas Chandrasekharan', 'Vikas Chauhan', 'Bhairav Mehta', 'Carolin Pirkl', 'Marion I Menzel', 'Jonathan J Wyatt', 'Steven F Petit', 'Dattesh Shanbhag', 'Cristina Cozzini', 'Mikael Bylund', 'Sandeep Kaushik'] | 2022-03-30 | null | null | null | null | ['value-prediction'] | ['computer-code'] | [ 6.48914218e-01 4.69768554e-01 1.78024635e-01 -5.18983305e-01
-1.22154665e+00 1.27699459e-02 4.41437155e-01 3.39471221e-01
-5.96662343e-01 8.63759398e-01 2.94114441e-01 -1.26424268e-01
-6.25766456e-01 -7.29760289e-01 -6.12498939e-01 -9.78591859e-01
-4.06832725e-01 7.86833286e-01 4.18923944e-01 -4.09344956e-02
5.64567558e-02 7.60518730e-01 -9.98632252e-01 7.19076216e-01
6.21872902e-01 1.11563492e+00 5.16096473e-01 4.85430896e-01
2.45425254e-01 1.09217846e+00 -4.66890812e-01 -3.79499234e-03
3.00455242e-01 -3.97576004e-01 -6.16176605e-01 -3.51719737e-01
3.91711801e-01 -1.93062261e-01 -1.51447132e-01 7.06690431e-01
8.24394584e-01 -1.27789482e-01 1.02197945e+00 -9.12758350e-01
2.53348261e-01 7.50205219e-01 -4.53227162e-01 3.15940082e-01
-5.07382452e-02 1.51207596e-01 5.95439613e-01 -8.43011439e-01
5.56052685e-01 7.85638571e-01 1.02793682e+00 3.15180004e-01
-1.42723393e+00 -3.27621430e-01 -5.43966830e-01 -1.40787270e-02
-1.29403090e+00 -1.78788096e-01 5.57294130e-01 -5.60431123e-01
8.03759336e-01 4.12912190e-01 4.30170298e-01 9.06681597e-01
8.64433527e-01 3.62675726e-01 1.44193423e+00 -4.86331254e-01
1.24599166e-01 9.56257507e-02 -3.37260932e-01 7.91849315e-01
2.17149585e-01 1.82771161e-01 -3.78814489e-01 7.76145905e-02
1.02667224e+00 -2.28920087e-01 -3.65340561e-01 -8.02643239e-01
-1.33421433e+00 8.29979002e-01 9.21013355e-01 7.05760837e-01
-6.39872015e-01 4.91988331e-01 4.15819228e-01 2.14566458e-02
5.99094093e-01 4.22218949e-01 -7.93989524e-02 2.94168234e-01
-1.46467304e+00 2.25885302e-01 4.94708300e-01 2.74773270e-01
3.70688081e-01 5.92200235e-02 -9.50535357e-01 9.78402615e-01
2.34377161e-01 3.17178458e-01 6.59842610e-01 -8.25131655e-01
2.93520868e-01 1.46651015e-01 -2.34056905e-01 -6.98909700e-01
-7.44355261e-01 -8.92332613e-01 -8.59224916e-01 6.49988592e-01
7.13787198e-01 -3.09062079e-02 -1.11963952e+00 1.54102707e+00
1.99270621e-01 -2.19296273e-02 -3.08385789e-01 1.05912864e+00
6.56975567e-01 2.03059077e-01 1.51020333e-01 -2.47268319e-01
1.11007738e+00 -7.37218440e-01 -4.60987598e-01 -1.87090710e-01
6.56817853e-01 -5.15260339e-01 9.22550023e-01 3.52431297e-01
-1.29292572e+00 -3.51899713e-01 -1.12048411e+00 4.01977986e-01
4.55258302e-02 1.18048146e-01 2.89258480e-01 8.28158081e-01
-1.14820802e+00 1.06355524e+00 -8.59065711e-01 -6.80335909e-02
6.45288944e-01 6.11205697e-01 -1.30703837e-01 -7.27717653e-02
-1.05764842e+00 1.43348861e+00 2.61382073e-01 1.05081692e-01
-1.05923998e+00 -1.18125319e+00 -7.03526080e-01 1.03772162e-02
1.14907190e-01 -6.54604614e-01 1.28309941e+00 -9.08352673e-01
-1.23151481e+00 7.46698141e-01 3.24076325e-01 -7.64306188e-01
1.02844930e+00 2.86172152e-01 9.42557305e-02 4.25898015e-01
2.68670589e-01 6.70358479e-01 8.01782072e-01 -1.29517949e+00
-1.11321554e-01 -2.94857621e-01 -6.18206024e-01 2.22415645e-02
2.20947899e-02 -6.72152713e-02 2.46648431e-01 -7.17757940e-01
3.28865707e-01 -6.04283094e-01 -5.48441887e-01 1.97422206e-01
-3.70652765e-01 3.61576051e-01 3.98193955e-01 -8.13366055e-01
8.78117859e-01 -1.50508583e+00 -7.95094669e-02 4.99896079e-01
4.08439279e-01 -2.05616891e-01 1.64059982e-01 1.99577976e-02
-4.76594299e-01 -1.21428132e-01 -6.08836293e-01 -1.21632911e-01
-4.68348294e-01 6.86862916e-02 3.32444817e-01 6.67174697e-01
1.70366824e-01 1.02675605e+00 -8.15664172e-01 -7.51785934e-01
3.73128802e-01 5.63370347e-01 -4.29300129e-01 2.06090976e-02
-5.55711351e-02 7.90842474e-01 -5.68253577e-01 3.76960069e-01
6.31544054e-01 -1.66824669e-01 1.55929744e-01 -5.99438012e-01
-8.93059522e-02 -5.25814667e-03 -7.31326580e-01 1.59406960e+00
-6.90537870e-01 3.77570271e-01 1.39582545e-01 -1.02847040e+00
8.95767868e-01 3.42277795e-01 1.23888814e+00 -1.08023632e+00
2.23450586e-01 5.50604284e-01 3.54943931e-01 -3.40273470e-01
9.26182643e-02 -8.70264649e-01 1.80945456e-01 4.73325074e-01
1.86713681e-01 -5.00399113e-01 -2.42675439e-01 -7.67990351e-02
1.21260881e+00 1.53278366e-01 4.44235727e-02 -5.74533284e-01
4.88717794e-01 4.05377448e-02 7.91471750e-02 8.62155974e-01
-1.91302776e-01 1.09581637e+00 5.51068783e-01 -3.83908808e-01
-1.52329850e+00 -1.08359933e+00 -5.28042972e-01 5.50909162e-01
-4.11357850e-01 2.07732826e-01 -6.08099818e-01 -8.09191227e-01
-7.49302581e-02 8.30278099e-01 -7.96124756e-01 -1.62605211e-01
-9.66644406e-01 -1.16448581e+00 5.00695646e-01 2.94856787e-01
1.29874259e-01 -1.13571143e+00 -9.35670018e-01 4.02704626e-01
-2.79304504e-01 -8.72194290e-01 -1.07822567e-01 7.55216599e-01
-1.08101785e+00 -7.65426636e-01 -1.31550956e+00 -4.20230776e-01
5.44660509e-01 -4.28481758e-01 1.36458921e+00 -1.11517362e-01
-3.97115022e-01 3.18878770e-01 -2.30928123e-01 -2.04812616e-01
-9.46779251e-01 2.17936903e-01 -4.47248459e-01 -1.60064355e-01
-3.46236527e-01 -5.76883078e-01 -8.10685873e-01 4.07820642e-01
-9.34685528e-01 2.43404686e-01 5.07614195e-01 9.21422064e-01
6.39504194e-01 -1.72028780e-01 5.16471684e-01 -8.95935774e-01
4.74336475e-01 -5.03186047e-01 -2.53336847e-01 3.02963227e-01
-7.20429778e-01 3.11624408e-01 2.82894731e-01 -3.68096344e-02
-1.09720576e+00 2.02973619e-01 -3.19290698e-01 -2.37583101e-01
2.93621659e-01 3.39239240e-01 5.41156828e-01 -6.35801613e-01
1.09757972e+00 3.11199427e-01 4.07724917e-01 -1.86225459e-01
4.78767306e-02 1.90840438e-01 2.97291517e-01 -5.30782163e-01
3.09382170e-01 4.55390960e-01 6.28927112e-01 -3.38931590e-01
-7.50140250e-01 -2.58062869e-01 -8.41844201e-01 -7.24165142e-01
6.65421307e-01 -4.51003969e-01 -4.83054340e-01 1.73434868e-01
-8.33391964e-01 -5.48332334e-01 -7.63198495e-01 6.52231216e-01
-9.20871139e-01 2.54019499e-01 -5.38136244e-01 -6.23553991e-01
-6.24114394e-01 -1.42475510e+00 1.02131891e+00 -1.68806612e-02
-1.16015300e-02 -1.05073667e+00 5.31503260e-02 1.02783456e-01
9.10116255e-01 6.32787943e-01 9.00180876e-01 -4.90100682e-01
-4.73770857e-01 1.31237581e-01 -3.00538510e-01 3.51428270e-01
-4.41733263e-02 -3.51115674e-01 -1.05401683e+00 -2.34993950e-01
3.39757115e-01 -5.06327152e-01 1.07078266e+00 1.19913840e+00
1.44920385e+00 2.14905769e-01 -3.39017600e-01 6.55262947e-01
1.69990122e+00 8.04339871e-02 6.90409303e-01 3.14721316e-01
4.05810535e-01 7.36320972e-01 5.01227736e-01 3.98608416e-01
-4.05896753e-02 8.44402134e-01 6.96818233e-01 -4.65878278e-01
-5.37963748e-01 6.04244471e-02 -7.00403601e-02 3.51346165e-01
-1.50537968e-01 1.25220018e-02 -1.22592413e+00 4.27804977e-01
-1.39853883e+00 -7.18868077e-01 -3.14357728e-01 2.36938334e+00
6.43909752e-01 2.14261815e-01 8.77194181e-02 4.05352712e-02
6.78575575e-01 3.01070940e-02 -4.51153696e-01 -2.73688406e-01
-2.43316293e-02 6.48700535e-01 1.08392000e+00 5.52623928e-01
-9.10759866e-01 2.31504083e-01 6.94005489e+00 1.08578801e+00
-1.32072198e+00 6.75543904e-01 1.17423344e+00 -2.29172409e-01
-3.22876126e-01 -3.97203863e-01 -2.20010787e-01 2.43957743e-01
1.07879257e+00 4.54646498e-02 -1.86218973e-02 3.74203414e-01
4.02463049e-01 -4.13722128e-01 -8.50847483e-01 5.56216419e-01
1.29146323e-01 -1.31472933e+00 -3.11772972e-01 1.16517119e-01
5.11892021e-01 2.85741866e-01 -1.27660658e-03 9.26690549e-02
-1.61114205e-02 -1.34111297e+00 8.53197396e-01 6.92823708e-01
1.04836953e+00 -4.63082999e-01 6.91497862e-01 2.79978663e-01
-8.29495668e-01 -4.58538644e-02 -1.38517350e-01 5.71517289e-01
3.61642659e-01 7.78483868e-01 -1.51430905e+00 6.24692678e-01
4.20226246e-01 1.78630114e-01 -6.17143452e-01 1.45682907e+00
2.57383853e-01 2.57009834e-01 -4.71253127e-01 1.29481807e-01
2.93380022e-01 -3.44992466e-02 4.12569135e-01 1.39762044e+00
6.09162390e-01 -2.62920469e-01 -3.89986113e-02 9.16229486e-01
1.97909117e-01 2.31894538e-01 -4.30119008e-01 6.09435081e-01
-1.12200081e-01 1.37841916e+00 -1.12793434e+00 -2.01099560e-01
3.58413495e-02 6.73892677e-01 1.32439241e-01 -1.20321244e-01
-8.98441315e-01 1.90886363e-01 -1.49517089e-01 6.95591390e-01
1.32668778e-01 6.63487762e-02 -5.41466296e-01 -5.85261941e-01
-2.33396575e-01 -3.85810465e-01 3.94366056e-01 -9.16453242e-01
-1.15973616e+00 8.96448731e-01 2.46005669e-01 -1.14283073e+00
-4.40429240e-01 -6.56345546e-01 -4.96614158e-01 1.07212806e+00
-1.46133733e+00 -1.15095699e+00 -4.26087230e-01 3.61622363e-01
2.96736896e-01 9.71337855e-02 5.53169668e-01 5.27732730e-01
4.98537757e-02 4.14146125e-01 3.17362957e-02 -2.35092252e-01
6.19780481e-01 -1.47256076e+00 -1.26445174e-01 3.01252991e-01
-2.47414291e-01 -1.36707559e-01 7.87456334e-01 -7.51035988e-01
-6.65825903e-01 -1.03005934e+00 3.82625937e-01 -3.36291015e-01
3.91628712e-01 -1.54067487e-01 -7.03230560e-01 3.89121532e-01
-4.34010141e-02 2.61443824e-01 1.80973202e-01 -4.68134135e-01
2.74888724e-01 -2.84120232e-01 -1.57860279e+00 3.76881883e-02
4.49054986e-01 -1.37841493e-01 -1.66327000e-01 6.12975717e-01
2.61860818e-01 -6.22355640e-01 -1.11892700e+00 6.39370501e-01
6.02856994e-01 -1.31432927e+00 1.30116785e+00 -1.43441513e-01
7.18883336e-01 5.24518415e-02 4.52040359e-02 -1.42518055e+00
-2.82029629e-01 1.36883259e-01 3.38723153e-01 6.04842484e-01
5.57934821e-01 -3.90891105e-01 1.15067232e+00 3.83107454e-01
-3.52232605e-01 -8.67991567e-01 -1.53498602e+00 -4.65156674e-01
4.78137374e-01 -5.02860606e-01 7.69646391e-02 6.25832736e-01
-4.35283870e-01 -2.01923281e-01 -3.80202740e-01 -2.17814028e-01
9.02236283e-01 -3.23357314e-01 2.85912175e-02 -9.95431244e-01
-2.73066849e-01 -4.43417639e-01 -1.94285378e-01 -3.45825881e-01
-3.19830060e-01 -1.25858760e+00 1.84342891e-01 -1.51194668e+00
3.26321989e-01 -9.02101576e-01 -5.89394510e-01 2.24801645e-01
3.77663672e-01 5.66971302e-01 9.78785902e-02 1.82966501e-01
-5.19970655e-02 5.06595492e-01 1.85467148e+00 -7.89854303e-02
1.71830788e-01 1.09480947e-01 -2.31507897e-01 3.92014056e-01
5.71922183e-01 -1.00475287e+00 -2.31566146e-01 -1.57042816e-01
-9.87419933e-02 6.13275111e-01 7.32212901e-01 -1.35797000e+00
1.65121928e-01 1.91335037e-01 8.02197158e-01 -4.97324318e-01
5.04241884e-01 -9.34008002e-01 3.65303159e-01 9.36335206e-01
-4.16730523e-01 -2.43686259e-01 2.89457273e-02 7.61598721e-02
-1.58720210e-01 -6.08384907e-01 1.31763172e+00 -5.27636647e-01
-6.64264783e-02 1.65872931e-01 -1.54374257e-01 -3.10437202e-01
9.44416046e-01 -4.32963550e-01 8.89430791e-02 -3.09518397e-01
-1.07500494e+00 -8.42834711e-02 4.74067219e-02 -1.59711540e-01
5.75333059e-01 -1.30910993e+00 -1.00912285e+00 1.06422625e-01
-1.45667791e-01 -1.67825237e-01 4.41689551e-01 1.36089182e+00
-8.94947112e-01 3.66863012e-01 -5.49665630e-01 -9.72970784e-01
-1.00875473e+00 7.22669661e-02 1.31569469e+00 -9.06047821e-01
-6.84413195e-01 7.41606772e-01 1.57381997e-01 -5.07680237e-01
-2.28291422e-01 -4.79057014e-01 -2.90556163e-01 -6.59868196e-02
1.39325887e-01 1.69518262e-01 6.63986802e-01 -6.60396457e-01
-2.36221239e-01 4.61521775e-01 4.39739786e-02 -2.04198420e-01
1.51613164e+00 4.37378399e-02 1.51848391e-01 2.43237615e-01
1.02092540e+00 -2.75692552e-01 -1.34972239e+00 -1.73107848e-01
-1.55268721e-02 -4.66261715e-01 6.28321767e-01 -1.26372516e+00
-1.25218129e+00 7.48296261e-01 1.15270352e+00 4.25554402e-02
1.04147995e+00 -6.39575766e-03 3.88650626e-01 -2.92883158e-01
3.98628950e-01 -8.40914667e-01 2.40903407e-01 1.37548953e-01
1.10124052e+00 -1.14869285e+00 2.03876600e-01 -4.47069168e-01
-6.78205490e-01 1.32967854e+00 4.93620098e-01 -2.10777581e-01
5.64290166e-01 5.45364082e-01 -1.22041203e-01 -4.75691497e-01
-3.53547007e-01 -2.06385348e-02 3.38319927e-01 6.94649637e-01
6.87962294e-01 -6.83906153e-02 -3.85143727e-01 2.40104824e-01
-1.30553106e-02 1.27620265e-01 5.13024449e-01 7.70821869e-01
-4.46677446e-01 -1.13007998e+00 -6.84107780e-01 8.72947752e-01
-7.06709445e-01 -1.37402058e-01 2.52701312e-01 8.45861137e-01
1.05596229e-01 1.96680844e-01 2.52459683e-02 1.42682567e-01
3.67472380e-01 -1.18952312e-01 8.70772362e-01 -6.01747572e-01
-1.07019937e+00 2.36336902e-01 5.20926379e-02 -5.03103971e-01
-3.19408447e-01 -5.49083412e-01 -1.01293874e+00 1.22951806e-01
-1.09289624e-01 -8.37180689e-02 9.47288990e-01 8.14897358e-01
-3.33690703e-01 1.02739239e+00 6.54914916e-01 -1.02873766e+00
-4.71573740e-01 -1.08467698e+00 -8.83439660e-01 2.76526123e-01
2.30905280e-01 -7.91121542e-01 -1.12987548e-01 -3.05955589e-01] | [13.884561538696289, -2.486630916595459] |
d5dffc64-c402-4433-a13a-a2aca920415e | a-principled-approach-to-data-valuation-for | 2009.06192 | null | https://arxiv.org/abs/2009.06192v1 | https://arxiv.org/pdf/2009.06192v1.pdf | A Principled Approach to Data Valuation for Federated Learning | Federated learning (FL) is a popular technique to train machine learning (ML) models on decentralized data sources. In order to sustain long-term participation of data owners, it is important to fairly appraise each data source and compensate data owners for their contribution to the training process. The Shapley value (SV) defines a unique payoff scheme that satisfies many desiderata for a data value notion. It has been increasingly used for valuing training data in centralized learning. However, computing the SV requires exhaustively evaluating the model performance on every subset of data sources, which incurs prohibitive communication cost in the federated setting. Besides, the canonical SV ignores the order of data sources during training, which conflicts with the sequential nature of FL. This paper proposes a variant of the SV amenable to FL, which we call the federated Shapley value. The federated SV preserves the desirable properties of the canonical SV while it can be calculated without incurring extra communication cost and is also able to capture the effect of participation order on data value. We conduct a thorough empirical study of the federated SV on a range of tasks, including noisy label detection, adversarial participant detection, and data summarization on different benchmark datasets, and demonstrate that it can reflect the real utility of data sources for FL and has the potential to enhance system robustness, security, and efficiency. We also report and analyze "failure cases" and hope to stimulate future research. | ['Tianhao Wang', 'Johannes Rausch', 'Ce Zhang', 'Dawn Song', 'Ruoxi Jia'] | 2020-09-14 | null | null | null | null | ['data-summarization'] | ['miscellaneous'] | [ 4.12141308e-02 1.95492327e-01 -5.08222938e-01 -2.84467250e-01
-8.26041341e-01 -8.12719643e-01 3.88998389e-01 3.70681256e-01
-5.81612349e-01 8.90225053e-01 7.65288994e-02 -4.76583123e-01
-4.27176625e-01 -7.33588636e-01 -7.97110796e-01 -8.34450364e-01
-3.74048322e-01 1.91228360e-01 -2.48871267e-01 2.03052275e-02
-1.86392978e-01 4.45888907e-01 -1.32508564e+00 1.35710061e-01
6.05006635e-01 1.36112261e+00 -2.70338356e-01 2.77614832e-01
1.54837891e-01 1.03475606e+00 -8.68731737e-01 -9.34434474e-01
8.18329275e-01 -3.08439195e-01 -7.45378911e-01 -3.03961635e-01
2.09095567e-01 -3.87183458e-01 -1.80431768e-01 1.11705649e+00
6.74936950e-01 -5.73535487e-02 4.13552761e-01 -2.00483441e+00
-4.72277641e-01 1.03299081e+00 -4.03765410e-01 -2.98255552e-02
-1.12814270e-01 1.18957795e-01 1.23694718e+00 -3.13166708e-01
5.81029534e-01 7.78391063e-01 6.46893322e-01 7.03227878e-01
-1.14876330e+00 -7.87617028e-01 -8.71209130e-02 -1.60048548e-02
-8.38990390e-01 -4.89843369e-01 7.82608986e-01 -1.98693261e-01
3.10877144e-01 5.43182075e-01 3.74319226e-01 9.77836430e-01
-3.53999920e-02 9.60279226e-01 1.08411658e+00 -2.29684249e-01
6.88638687e-01 2.57429838e-01 6.45468831e-02 4.41235542e-01
5.20925581e-01 3.45373154e-01 -7.56206810e-01 -7.51111031e-01
1.58943817e-01 1.02747850e-01 -3.00889790e-01 -8.10582936e-01
-1.01741755e+00 9.15185928e-01 4.92921948e-01 -4.74308468e-02
-3.31950516e-01 2.29549631e-01 6.83572769e-01 8.17554593e-01
4.12889302e-01 2.38247737e-01 -5.49067378e-01 1.24107428e-01
-6.73286855e-01 2.24418253e-01 1.00132728e+00 6.79010391e-01
5.51760733e-01 -6.61348104e-02 -4.14459348e-01 2.98193365e-01
-3.23047303e-02 4.13374603e-01 4.36473221e-01 -1.42899024e+00
5.96974730e-01 6.31580651e-01 1.69472858e-01 -6.51196122e-01
-1.15514763e-01 -5.42176366e-01 -1.05600715e+00 2.10781634e-01
4.74007875e-01 -4.50510502e-01 8.33866373e-02 2.11508846e+00
5.32073557e-01 -5.05986698e-02 2.01533794e-01 8.15427303e-01
4.13870215e-01 1.76429257e-01 -8.23892467e-03 -4.23596323e-01
9.20367360e-01 -6.39094949e-01 -5.74602604e-01 3.12414110e-01
9.02112305e-01 -6.98999241e-02 8.08146179e-01 3.19457531e-01
-1.03770006e+00 1.77847192e-01 -9.05452669e-01 2.21006706e-01
-2.06987247e-01 -2.63096690e-01 8.21268082e-01 1.06765318e+00
-9.79265571e-01 6.50165319e-01 -5.43188393e-01 1.67192370e-01
8.56763005e-01 5.65250933e-01 -4.90802288e-01 1.33921489e-01
-1.26449966e+00 5.06070197e-01 -8.10700282e-03 -3.63112658e-01
-1.08728838e+00 -1.00366056e+00 -5.91154873e-01 3.55374753e-01
5.50482333e-01 -7.47615576e-01 1.41004789e+00 -1.01577628e+00
-1.06301641e+00 6.37899756e-01 3.36551487e-01 -8.57897043e-01
1.09759796e+00 3.55196476e-01 -1.15661882e-02 -2.46623941e-02
-4.28296626e-02 2.49210626e-01 6.27569377e-01 -1.16392732e+00
-7.70207703e-01 -4.72495437e-01 2.50120968e-01 6.82775378e-02
-1.01594412e+00 -3.01734246e-02 2.15604514e-01 -6.40551507e-01
-4.93194163e-01 -6.39232337e-01 -1.28612384e-01 2.92591125e-01
-1.46259353e-01 -4.13751245e-01 9.07502174e-01 -2.61936158e-01
1.02090919e+00 -2.14652157e+00 -1.91268295e-01 2.95145094e-01
4.60007221e-01 1.35230824e-01 -2.36603096e-01 3.64525765e-01
2.65909106e-01 3.23586345e-01 -3.30125719e-01 -4.68591124e-01
2.46804476e-01 1.85988098e-02 -2.63849974e-01 7.91815937e-01
-3.10484588e-01 9.69327629e-01 -8.07771146e-01 -3.80659819e-01
-3.36138874e-01 1.42778546e-01 -6.32105172e-01 1.03945263e-01
-1.64871633e-01 8.41926336e-02 -4.44350064e-01 5.83177149e-01
5.87431550e-01 -3.10953140e-01 3.46088529e-01 1.16969869e-01
3.09886396e-01 -1.95214469e-02 -1.36658335e+00 1.23760784e+00
-2.81609297e-01 2.91571468e-01 5.31436861e-01 -9.77854669e-01
6.35186732e-01 4.36411023e-01 8.56582046e-01 -6.66619837e-01
3.22218806e-01 2.24158719e-01 -4.22588557e-01 -2.32314929e-01
2.64007211e-01 -2.14838907e-02 -3.43964666e-01 1.16257477e+00
-1.22649260e-01 5.73847830e-01 -1.79870933e-01 6.17702007e-01
1.23493218e+00 -5.30370653e-01 3.53692412e-01 -2.92628825e-01
3.90618503e-01 -2.67905474e-01 8.32727432e-01 8.03577602e-01
-6.22166753e-01 2.77977511e-02 7.16001451e-01 -3.66107821e-01
-7.56866276e-01 -9.92920518e-01 1.32355958e-01 1.24204040e+00
1.50489405e-01 -1.59946308e-01 -7.64424205e-01 -1.18052375e+00
5.86848736e-01 4.07277137e-01 -5.62108994e-01 -3.80115509e-01
-4.74002361e-02 -5.19527137e-01 8.18616927e-01 2.77836561e-01
5.93582034e-01 -8.35264087e-01 -8.04094434e-01 -6.69936240e-02
-3.58544737e-01 -6.68664992e-01 -7.20950544e-01 1.73938006e-01
-6.17205441e-01 -1.40950048e+00 -4.54788059e-01 -4.15357828e-01
6.05600536e-01 4.88768816e-01 8.16757679e-01 6.28353804e-02
7.60704502e-02 5.98796010e-01 -1.65489897e-01 -6.93936527e-01
-5.35232604e-01 -2.85965391e-02 3.01423967e-01 3.17193329e-01
1.99351579e-01 -2.81824112e-01 -5.47612607e-01 2.81637728e-01
-1.09779608e+00 -3.49497467e-01 2.27336422e-01 9.28456604e-01
3.99381578e-01 1.32208303e-01 1.26276290e+00 -1.19463921e+00
8.30744386e-01 -7.36760139e-01 -6.11307323e-01 5.83039463e-01
-9.21210289e-01 1.44806996e-01 8.85592639e-01 -4.20873225e-01
-9.00696516e-01 9.67188105e-02 4.90673900e-01 -4.65161622e-01
3.24612260e-01 3.22511941e-01 -4.06009525e-01 -3.45880181e-01
7.31144130e-01 -1.65188327e-01 3.80970865e-01 -4.10563231e-01
4.54047769e-01 9.06273961e-01 4.56789702e-01 -7.02169955e-01
7.13612020e-01 4.59394634e-01 1.45858839e-01 -1.27397403e-01
-5.53177059e-01 -2.84328461e-01 9.51474719e-03 -9.56589654e-02
1.30251035e-01 -7.90187478e-01 -1.33153296e+00 3.76220644e-01
-8.88062656e-01 -1.71782091e-01 -9.01680708e-01 2.46130913e-01
-5.67656279e-01 3.68816912e-01 -4.51389313e-01 -8.90309215e-01
-6.84302866e-01 -7.70132482e-01 2.92096019e-01 -1.42919049e-01
2.37921551e-01 -9.71190155e-01 -8.86233989e-03 6.03050470e-01
5.38743973e-01 5.40533960e-01 8.88199091e-01 -8.81296337e-01
-4.35564905e-01 -3.92838925e-01 3.68308909e-02 4.49773848e-01
-1.02388458e-02 -4.93416011e-01 -1.13623941e+00 -7.90660143e-01
2.85201132e-01 -6.97873950e-01 6.82673097e-01 -1.10598505e-02
1.32650471e+00 -1.01419103e+00 4.68985736e-02 4.32303220e-01
1.40370178e+00 -2.77790666e-01 -5.89646921e-02 1.69059336e-01
2.40837201e-01 8.93918633e-01 3.32721680e-01 8.39349091e-01
5.80729127e-01 2.25996330e-01 7.07945526e-01 1.73929427e-02
2.46550217e-01 -3.09983015e-01 5.07196665e-01 2.55081445e-01
3.19933921e-01 -1.38696641e-01 -5.58332503e-01 4.39425617e-01
-1.83878267e+00 -9.79514182e-01 2.25827217e-01 2.53228664e+00
9.15620863e-01 -2.97540933e-01 4.42685604e-01 4.54409510e-01
5.96845448e-01 6.05072230e-02 -1.00941467e+00 -4.11130875e-01
-3.33985716e-01 -1.59672871e-01 9.45037127e-01 2.43279058e-02
-7.72483110e-01 2.70010412e-01 6.23457336e+00 7.55423784e-01
-9.25993741e-01 6.23791099e-01 8.08369756e-01 -6.81770802e-01
-4.87298340e-01 -1.69227988e-01 -3.91590565e-01 5.45910299e-01
1.04534996e+00 -9.68063176e-01 7.09877849e-01 1.00633490e+00
3.98925804e-02 1.90894634e-01 -1.27625167e+00 8.47936511e-01
-1.78700238e-01 -1.40876389e+00 -1.28288120e-01 3.06174338e-01
9.88361239e-01 2.74420470e-01 1.56172693e-01 1.55577436e-01
6.80131972e-01 -7.19709277e-01 8.64859641e-01 1.76060602e-01
8.98211598e-01 -9.69129741e-01 6.89817667e-01 8.19667995e-01
-8.60790014e-01 -8.17443490e-01 -2.23795816e-01 1.52913079e-01
-3.11142534e-01 5.46231747e-01 -4.10217285e-01 6.07760191e-01
6.04034603e-01 2.33802438e-01 -5.16210437e-01 8.56989861e-01
2.48116374e-01 5.48713386e-01 -3.07256162e-01 1.55157506e-01
-5.61540313e-02 1.48320332e-01 4.31834042e-01 6.81220353e-01
1.18371621e-01 -1.14568442e-01 -2.81812549e-02 6.68427885e-01
-1.01062083e+00 1.78865865e-01 -6.91504776e-01 4.14569266e-02
8.64718139e-01 1.21802795e+00 -7.83276632e-02 -1.12090364e-01
-2.58774966e-01 4.56571311e-01 4.12735999e-01 1.98691279e-01
-5.20149231e-01 -8.85654241e-02 6.99427783e-01 -1.74154453e-02
-1.44785836e-01 2.32459947e-01 -5.25008082e-01 -1.06942248e+00
1.71398461e-01 -1.07724679e+00 9.21464443e-01 1.29552567e-02
-1.56228042e+00 2.14401811e-01 -1.63242877e-01 -1.21696150e+00
-1.48609594e-01 -6.92952052e-03 -6.09016836e-01 5.88392973e-01
-1.58928025e+00 -1.00293756e+00 -9.00654495e-02 1.00150335e+00
-1.92444652e-01 -3.43872339e-01 6.83802783e-01 2.36581564e-01
-5.80027938e-01 1.19562662e+00 5.46875119e-01 -3.87062542e-02
6.97641134e-01 -1.05613816e+00 -1.83180661e-03 7.25671053e-01
7.49456361e-02 3.59925628e-01 3.25574338e-01 -3.85736763e-01
-1.50641394e+00 -1.31464732e+00 9.63438809e-01 -4.63230699e-01
3.63310754e-01 -2.95030951e-01 -5.81692338e-01 4.88630265e-01
-1.25872076e-01 3.76675695e-01 9.11446333e-01 -1.67315662e-01
-5.90224445e-01 -7.02195942e-01 -1.87256753e+00 1.91784054e-01
9.61908937e-01 -4.88638490e-01 -5.53328097e-02 3.26309144e-01
8.42136443e-01 6.24694228e-02 -8.54971409e-01 2.36393437e-01
3.52603912e-01 -1.00363374e+00 6.47398055e-01 -7.13746905e-01
-1.81168690e-02 1.64497793e-01 -1.35958061e-01 -1.01668131e+00
-1.87248383e-02 -1.00642729e+00 -6.01123869e-01 1.42649925e+00
1.81774572e-01 -1.00612438e+00 1.07361197e+00 1.15824568e+00
5.09672880e-01 -7.69505203e-01 -1.37331283e+00 -1.06701851e+00
2.74444640e-01 -2.23861203e-01 1.14576471e+00 1.21502244e+00
1.13052249e-01 -3.09315801e-01 -4.33006018e-01 -1.41528606e-01
1.23528302e+00 -3.95100983e-03 6.35207117e-01 -1.33633840e+00
-2.56654024e-01 -3.86193961e-01 -1.10722356e-01 -3.78906429e-01
4.58856434e-01 -1.25611830e+00 -5.19973814e-01 -1.11079192e+00
3.51544946e-01 -7.91658282e-01 -6.41517520e-01 9.07930672e-01
1.83966160e-01 1.11357391e-01 5.74901998e-01 5.21872163e-01
-7.37487257e-01 5.14702797e-01 8.07593048e-01 -2.66370386e-01
-1.39020279e-01 2.82346606e-01 -1.19163430e+00 1.62309170e-01
9.57214177e-01 -7.45759189e-01 -5.26322722e-01 -2.15646371e-01
2.71580756e-01 2.05597788e-01 5.50296426e-01 -6.60360396e-01
6.23140812e-01 -3.47841442e-01 -7.15939552e-02 -3.74631062e-02
-1.55903637e-01 -1.21794677e+00 3.00213933e-01 6.63222194e-01
-7.64448047e-01 -1.31224543e-01 -4.59056139e-01 7.61248171e-01
8.03362578e-02 -1.65941387e-01 7.03706205e-01 4.16258089e-02
-2.21917108e-02 5.27054012e-01 1.14321008e-01 2.40065828e-01
1.48893201e+00 5.32449894e-02 -5.18528819e-01 -3.99948388e-01
-9.41337198e-02 6.70639217e-01 5.55816114e-01 2.96555907e-01
3.32614690e-01 -1.37765312e+00 -7.13239551e-01 3.66859138e-01
1.51861459e-02 -7.04604462e-02 1.16194375e-01 6.53967559e-01
2.95505784e-02 2.64145464e-01 -1.16090223e-01 -6.57595992e-02
-1.26888442e+00 7.63428211e-01 4.57738787e-01 -2.82666653e-01
-4.11131799e-01 4.74904537e-01 -2.55629331e-01 -4.25407022e-01
7.82564461e-01 2.16371536e-01 1.16370164e-01 2.12710962e-01
4.42984998e-01 8.03715587e-01 1.09883666e-01 -3.70474607e-01
-1.82307303e-01 -2.83403456e-01 1.60197437e-01 -3.37522924e-02
1.29827571e+00 -1.34326488e-01 -3.95231582e-02 9.75623876e-02
1.25949800e+00 -9.97700617e-02 -1.43330216e+00 -6.46868885e-01
1.17775276e-02 -3.99698108e-01 1.11532919e-01 -9.12163317e-01
-1.48172700e+00 5.20546377e-01 4.73263949e-01 2.64074475e-01
1.21208847e+00 -1.55313671e-01 8.64770710e-01 4.46025282e-01
8.43560278e-01 -1.05253685e+00 -1.34603634e-01 -7.33700618e-02
5.14082551e-01 -1.20024943e+00 -9.26558003e-02 1.48690313e-01
-8.25139880e-01 6.11933291e-01 4.08942282e-01 3.16064924e-01
6.43387496e-01 3.60980809e-01 -4.78369594e-02 2.06296057e-01
-1.05404580e+00 4.14076865e-01 -1.98577732e-01 5.76707065e-01
-2.74959147e-01 2.89737374e-01 -4.77239400e-01 1.15313864e+00
6.03872864e-03 1.51858538e-01 5.37436724e-01 1.01661253e+00
-1.37337029e-01 -1.18358147e+00 -2.86290258e-01 6.60355747e-01
-7.36604154e-01 3.00841272e-01 -6.00137472e-01 1.33356094e-01
1.79342613e-01 1.19248724e+00 -2.44189009e-01 -3.38216722e-01
1.95245892e-01 -2.28520799e-02 -6.86750114e-02 -1.48731589e-01
-1.25423515e+00 -5.22396564e-01 -2.45620236e-01 -6.08539999e-01
-4.91014123e-01 -4.63341057e-01 -1.17547250e+00 -8.88450444e-01
-4.23199534e-01 5.18004358e-01 8.37025642e-01 4.61638838e-01
5.53500533e-01 -7.30740651e-02 1.39100742e+00 -1.98028296e-01
-1.61662459e+00 -4.24360096e-01 -8.31616223e-01 6.32617891e-01
5.00049353e-01 -1.67702496e-01 -7.35128224e-01 -2.15637967e-01] | [5.873630046844482, 6.467391014099121] |
728988f0-326a-4979-b0ea-d7247c53be2c | adversarial-training-for-multi-channel-sign | 2008.12405 | null | https://arxiv.org/abs/2008.12405v1 | https://arxiv.org/pdf/2008.12405v1.pdf | Adversarial Training for Multi-Channel Sign Language Production | Sign Languages are rich multi-channel languages, requiring articulation of both manual (hands) and non-manual (face and body) features in a precise, intricate manner. Sign Language Production (SLP), the automatic translation from spoken to sign languages, must embody this full sign morphology to be truly understandable by the Deaf community. Previous work has mainly focused on manual feature production, with an under-articulated output caused by regression to the mean. In this paper, we propose an Adversarial Multi-Channel approach to SLP. We frame sign production as a minimax game between a transformer-based Generator and a conditional Discriminator. Our adversarial discriminator evaluates the realism of sign production conditioned on the source text, pushing the generator towards a realistic and articulate output. Additionally, we fully encapsulate sign articulators with the inclusion of non-manual features, producing facial features and mouthing patterns. We evaluate on the challenging RWTH-PHOENIX-Weather-2014T (PHOENIX14T) dataset, and report state-of-the art SLP back-translation performance for manual production. We set new benchmarks for the production of multi-channel sign to underpin future research into realistic SLP. | ['Ben Saunders', 'Richard Bowden', 'Necati Cihan Camgoz'] | 2020-08-27 | null | null | null | null | ['sign-language-production'] | ['natural-language-processing'] | [ 5.84747791e-01 2.77963459e-01 1.18790060e-01 -3.98474008e-01
-1.17064393e+00 -1.01414049e+00 9.52606559e-01 -1.08268893e+00
-1.83547065e-01 5.66638052e-01 5.42253673e-01 -2.24144801e-01
3.18237007e-01 -4.03379023e-01 -9.25756216e-01 -6.48313224e-01
5.91089129e-02 5.23892701e-01 -4.74953562e-01 -3.52276832e-01
-2.95281500e-01 3.71498406e-01 -1.27919734e+00 5.40313959e-01
5.67435086e-01 8.20421219e-01 -2.84077883e-01 1.27394581e+00
2.03921184e-01 8.56908917e-01 -8.46944630e-01 -9.01975691e-01
6.80804551e-01 -8.60152006e-01 -4.88512456e-01 -1.69324249e-01
9.49101806e-01 -5.49529195e-01 -2.31456026e-01 7.85471737e-01
9.31971252e-01 -4.63932812e-01 9.67763543e-01 -1.52477193e+00
-9.61651146e-01 6.24941230e-01 -2.49281064e-01 -8.16799462e-01
4.80969191e-01 7.61252165e-01 1.20728660e+00 -7.67069876e-01
1.09129047e+00 1.54680312e+00 6.26178563e-01 1.00835693e+00
-1.16507304e+00 -1.02502203e+00 -2.02616364e-01 -3.09856296e-01
-1.22377396e+00 -7.49354303e-01 5.24801016e-01 -4.54119921e-01
5.96337020e-01 3.90789926e-01 7.42676079e-01 1.80773854e+00
-1.15744963e-01 1.00169206e+00 1.57713139e+00 -4.88147169e-01
-1.69665933e-01 -2.87127972e-01 -7.94283688e-01 6.78306282e-01
-3.40621352e-01 7.83123672e-01 -8.94078791e-01 2.83801332e-02
7.17815816e-01 -7.80568421e-01 -3.12072992e-01 -4.56644408e-02
-1.23725343e+00 5.81896007e-01 1.64924264e-01 -8.78987461e-02
-3.45353007e-01 7.07192004e-01 1.23386376e-01 7.23755836e-01
-1.38254330e-01 2.40452439e-01 -4.33410704e-01 -4.08932984e-01
-1.06790841e+00 5.69195330e-01 9.06192899e-01 1.17295110e+00
-1.38817698e-01 2.99686760e-01 -6.67772651e-01 6.44375741e-01
6.67768002e-01 1.46217549e+00 2.50309199e-01 -1.06814802e+00
6.12967134e-01 -7.39099085e-02 -1.13259673e-01 -3.07932645e-01
-2.91384999e-02 -5.37710041e-02 -7.39403784e-01 9.18357670e-01
8.20548713e-01 -5.19540727e-01 -1.47250128e+00 2.02256846e+00
-8.29408765e-02 -1.38224870e-01 2.06661150e-01 1.02168798e+00
7.99589455e-01 4.26555812e-01 9.40370709e-02 2.62329638e-01
1.15254521e+00 -8.24335337e-01 -6.44339859e-01 7.77062625e-02
2.17229679e-01 -1.09290361e+00 1.05578291e+00 3.73358399e-01
-1.27511680e+00 -2.55432218e-01 -6.83345199e-01 -1.45287156e-01
-5.24284542e-02 2.88520813e-01 3.63438666e-01 8.01088929e-01
-1.26455748e+00 4.05689001e-01 -5.98907828e-01 -1.25223950e-01
4.19082701e-01 3.43693852e-01 -6.01310253e-01 6.69473186e-02
-1.14242399e+00 9.96389508e-01 -2.52361238e-01 1.94896296e-01
-7.66177356e-01 -7.37978160e-01 -8.90931964e-01 -5.83202004e-01
-3.69058579e-01 -8.34544182e-01 1.56204104e+00 -1.17410946e+00
-2.23683548e+00 1.02753937e+00 -8.09646025e-02 -3.17906380e-01
1.43139529e+00 -1.08174838e-01 -3.97984594e-01 -3.96341458e-02
-2.11207360e-01 1.11558461e+00 1.40978599e+00 -1.33249295e+00
-2.99525917e-01 6.71991035e-02 -3.71262819e-01 1.13430634e-01
5.47195673e-01 3.07694167e-01 -1.69068277e-01 -1.09494102e+00
-1.70470580e-01 -1.28493309e+00 3.72622430e-01 6.29368782e-01
-6.55946910e-01 3.85169297e-01 8.06489348e-01 -1.23347604e+00
5.11660039e-01 -2.01925254e+00 3.22627425e-01 4.12912041e-01
-8.67860466e-02 1.59634292e-01 -6.31858826e-01 3.67497981e-01
8.07638764e-02 4.06181104e-02 -5.36429703e-01 -6.13516033e-01
5.12582660e-01 4.37252998e-01 -6.77211821e-01 5.07744074e-01
3.92851561e-01 1.48591006e+00 -7.72242010e-01 -4.38713491e-01
9.31059346e-02 8.41075778e-01 -6.32771194e-01 6.99686930e-02
-2.81283855e-01 9.77072120e-01 1.14962541e-01 1.05357444e+00
5.75319886e-01 4.32586074e-01 1.48943774e-02 6.13436960e-02
4.97506894e-02 -7.35432059e-02 -8.90648067e-01 1.58784485e+00
-8.36196125e-01 8.38919461e-01 3.54425579e-01 -1.24779493e-01
7.85333991e-01 4.08580929e-01 1.22095212e-01 -7.44408309e-01
2.80781657e-01 6.65157974e-01 3.32577080e-01 -2.63374031e-01
9.14963260e-02 -6.06956482e-01 -3.14735800e-01 4.90412593e-01
2.83904523e-01 -9.12274718e-01 -6.94050267e-02 -1.25824720e-01
9.43698585e-01 4.62971032e-01 -1.42975703e-01 2.42018715e-01
1.13908149e-01 -4.63195831e-01 4.74239364e-02 6.52196765e-01
-1.05887510e-01 1.11383915e+00 2.94777393e-01 2.22956449e-01
-9.72209096e-01 -1.52563012e+00 1.48848265e-01 7.51466930e-01
-4.63141203e-01 6.87072724e-02 -7.51995564e-01 -5.81625879e-01
2.78163880e-01 8.31293583e-01 -6.24458969e-01 5.79073764e-02
-7.91220844e-01 -1.21255659e-01 1.63071728e+00 4.44275111e-01
4.13647264e-01 -1.49162972e+00 -4.88791078e-01 6.53470382e-02
-2.36554056e-01 -1.21558130e+00 -9.16269064e-01 -2.05826461e-01
-1.74888879e-01 -8.63041580e-01 -1.39645970e+00 -7.68177092e-01
4.69428241e-01 -8.24120343e-01 8.48327160e-01 -1.96399629e-01
-2.83624738e-01 3.96051854e-01 -4.15340573e-01 -6.17290795e-01
-1.21854901e+00 -4.80749220e-01 4.87580411e-02 2.42187366e-01
-2.57304251e-01 -5.90014875e-01 -2.68749803e-01 8.13324898e-02
-7.45130122e-01 3.43609303e-01 8.91248107e-01 1.02793419e+00
2.81065732e-01 -9.18260872e-01 1.31830230e-01 -2.43274510e-01
6.62942886e-01 1.46547273e-01 -4.23566401e-01 2.30097905e-01
-7.93147907e-02 4.54580605e-01 5.72601974e-01 -6.96310401e-01
-9.76364970e-01 2.14041382e-01 -4.24178094e-01 -4.13556218e-01
5.56781925e-02 -1.12753578e-01 -3.13487858e-01 -4.13608909e-01
6.13578737e-01 5.67588568e-01 2.29068696e-01 -3.12762320e-01
1.05493343e+00 9.00881052e-01 9.27175164e-01 -7.68521547e-01
1.26420903e+00 4.37378794e-01 1.05262160e-01 -7.30329335e-01
-1.21482126e-01 2.54477739e-01 -6.40209615e-01 -3.14824283e-01
6.47727609e-01 -7.46777773e-01 -8.18975687e-01 1.19546723e+00
-1.40763474e+00 -8.27247322e-01 -6.26642466e-01 2.90434837e-01
-1.00026989e+00 9.88964587e-02 -4.66746062e-01 -6.97142780e-01
-5.20437598e-01 -1.04513037e+00 1.62339807e+00 -1.69214666e-01
-6.42586946e-01 -5.45669854e-01 2.00216830e-01 2.58096963e-01
6.16593599e-01 6.31889403e-01 5.29981911e-01 -7.16802059e-03
-5.33801317e-01 -1.83059216e-01 -1.93499506e-01 7.31805265e-01
2.76019387e-02 2.39039972e-01 -1.19528317e+00 -1.48148775e-01
-7.00102806e-01 -5.39422154e-01 7.49717295e-01 2.00595677e-01
2.78414726e-01 -6.33796751e-01 3.52091163e-01 9.70702231e-01
9.18198586e-01 -1.93155780e-01 6.54947162e-01 -3.09529066e-01
5.85065305e-01 5.73621392e-01 4.46844846e-02 3.21044177e-01
3.75603467e-01 7.74352193e-01 1.17144272e-01 -1.11285433e-01
-1.04871190e+00 -8.11802149e-01 8.56326699e-01 6.70288205e-01
-4.25998330e-01 -3.08994263e-01 -7.49977946e-01 4.65989679e-01
-1.24039483e+00 -1.03408670e+00 6.59854989e-03 1.71116292e+00
1.25963402e+00 -2.57575005e-01 5.43406000e-03 2.92179525e-01
3.45970184e-01 1.43123627e-01 -4.00007606e-01 -6.91036463e-01
-6.15324318e-01 8.93744290e-01 6.35416389e-01 7.93527782e-01
-9.46333826e-01 1.21123195e+00 6.25850916e+00 4.99293387e-01
-1.56091356e+00 1.35055095e-01 -1.64595619e-01 -2.48770118e-01
-3.58832836e-01 -2.74980307e-01 -4.64471430e-01 4.98429418e-01
6.18777215e-01 2.42926665e-02 7.54058599e-01 2.24768162e-01
1.50863096e-01 4.49705929e-01 -1.07660973e+00 1.04350388e+00
1.02148347e-01 -1.00463223e+00 3.02885622e-01 1.33999690e-01
8.17866445e-01 1.91599384e-01 4.16306347e-01 3.06893498e-01
6.88574433e-01 -1.43826091e+00 1.59981883e+00 6.61642909e-01
1.68090212e+00 -2.08206579e-01 3.00335884e-01 7.52846757e-03
-1.00461483e+00 5.88829406e-02 7.61826515e-01 6.18430488e-02
7.35335231e-01 -2.79864743e-02 -8.39810014e-01 2.69236863e-01
1.27931952e-01 2.24135429e-01 -1.62037998e-01 7.06344962e-01
-9.33950126e-01 5.99825263e-01 -6.01691306e-01 4.32649553e-02
1.36569798e-01 1.67955726e-01 7.75799155e-01 1.43809688e+00
4.66200650e-01 -2.48667017e-01 -2.57953435e-01 9.69965696e-01
-2.98318803e-01 8.24260041e-02 -6.55877471e-01 -1.89813778e-01
1.44803062e-01 6.45664513e-01 2.34476868e-02 -7.36216307e-02
-1.45013452e-01 1.56972969e+00 -2.26945311e-01 4.27336574e-01
-7.71822453e-01 -2.54177898e-01 1.02497411e+00 -5.43037839e-02
2.34434813e-01 -3.40897590e-01 -3.93900096e-01 -1.09863567e+00
3.50033820e-01 -1.18770540e+00 -1.51068971e-01 -9.38520372e-01
-1.20921195e+00 4.17751223e-01 -3.74782711e-01 -1.32572162e+00
-7.26203382e-01 -6.78647220e-01 -4.88456756e-01 1.16029787e+00
-1.47853804e+00 -2.09650445e+00 -9.23805460e-02 6.31655514e-01
1.42248601e-01 -9.87936854e-02 1.10700154e+00 3.46791089e-01
2.87643999e-01 1.30891275e+00 -1.13609187e-01 6.09526217e-01
7.76156843e-01 -1.04807842e+00 9.09113288e-01 8.28431785e-01
4.02993947e-01 2.00889543e-01 7.03992009e-01 -5.76910913e-01
-1.28515351e+00 -1.06513488e+00 1.11347425e+00 -6.78506076e-01
5.69980443e-01 -5.22517323e-01 -9.03029665e-02 5.80173731e-01
-2.78583057e-02 8.69492367e-02 2.24357426e-01 -6.67800307e-01
-7.49122858e-01 3.18456680e-01 -1.38992369e+00 8.32285523e-01
1.37607086e+00 -8.85479450e-01 -4.46828902e-01 9.76400375e-02
4.41886961e-01 -7.03386605e-01 -6.40351176e-01 3.11878920e-01
1.35842168e+00 -5.11016607e-01 8.55795264e-01 -5.05301058e-01
5.63065410e-01 -2.71438301e-01 -3.43895078e-01 -1.49253416e+00
3.80645305e-01 -1.27975500e+00 -1.78636968e-01 1.11936057e+00
5.30220270e-01 -5.33680618e-01 5.07247984e-01 3.55706334e-01
6.46707937e-02 -3.73165667e-01 -1.19026685e+00 -1.10465407e+00
4.17894304e-01 -7.15314150e-01 5.34216702e-01 6.26205087e-01
-3.78203958e-01 -1.21903688e-01 -9.30055082e-01 -1.54403364e-02
7.60969996e-01 -5.50250150e-02 1.06428742e+00 -8.34924221e-01
-6.00992441e-01 -6.83203101e-01 -4.13073838e-01 -9.19965327e-01
2.79050499e-01 -1.28062367e+00 3.24665636e-01 -1.23020017e+00
-4.82847124e-01 -1.98394954e-01 3.81962329e-01 7.85439968e-01
4.54281390e-01 4.44505721e-01 7.81759143e-01 2.47603245e-02
2.79983938e-01 5.14323533e-01 1.81458819e+00 -3.59702110e-01
-9.08418149e-02 1.29978970e-01 -4.26468819e-01 5.21404147e-01
4.48616356e-01 -4.12802011e-01 1.46946236e-01 -5.73781550e-01
-2.02734739e-01 2.93987785e-02 7.67922223e-01 -7.59956598e-01
1.43247973e-02 7.97761157e-02 1.67797327e-01 2.97557283e-02
6.23256624e-01 -5.47052145e-01 1.65577173e-01 6.60109282e-01
-3.21987718e-01 -2.63606697e-01 3.69292237e-02 1.68610606e-02
-2.42181078e-01 3.33080798e-01 9.47204292e-01 4.67935726e-02
-3.58673930e-01 2.18139157e-01 -2.01785952e-01 4.03255075e-01
6.31202161e-01 -1.01836167e-01 -1.82740614e-01 -8.35341811e-01
-7.09787726e-01 2.98591563e-03 3.93147230e-01 6.88223481e-01
6.49240136e-01 -1.40970314e+00 -1.30948615e+00 6.80405438e-01
4.87354025e-02 -3.81698132e-01 -5.59399426e-02 6.33810818e-01
-8.28410327e-01 1.85725927e-01 -3.81209224e-01 -3.24086040e-01
-1.41679001e+00 -2.71085083e-01 5.10786772e-01 4.13752568e-04
-6.52656078e-01 1.06621480e+00 -3.24337453e-01 -8.35544169e-01
3.59766603e-01 -5.58534443e-01 5.53842187e-01 -3.94905172e-03
3.36775094e-01 5.59645370e-02 -2.92677939e-01 -1.04754329e+00
-2.88994312e-01 8.95637929e-01 6.86434388e-01 -9.85277951e-01
1.00488448e+00 4.34697062e-01 1.93950251e-01 7.32094096e-03
1.24454892e+00 3.87913227e-01 -1.22441697e+00 2.73222458e-02
-4.71751422e-01 -2.75441557e-01 -3.00481737e-01 -1.49100149e+00
-9.53452766e-01 9.81988728e-01 5.46084583e-01 -6.62441611e-01
8.72366667e-01 -1.30598098e-01 1.04451168e+00 1.91147864e-01
5.17354965e-01 -8.97275150e-01 -1.76936671e-01 7.09382594e-01
1.77704859e+00 -9.72523391e-01 -5.86745560e-01 -1.41710401e-01
-9.29639041e-01 9.28381503e-01 3.27489786e-02 2.97788624e-02
5.92247128e-01 7.64313102e-01 7.91514397e-01 2.28925079e-01
-2.03718722e-01 -2.52849609e-01 4.83730793e-01 8.74119163e-01
1.41220003e-01 4.58797067e-01 -3.79160233e-02 5.08319557e-01
-9.64519799e-01 1.73203155e-01 1.62840694e-01 9.08677340e-01
4.08563256e-01 -1.36516964e+00 -4.98457432e-01 1.24793120e-01
-2.05031067e-01 -4.41013187e-01 -8.43571007e-01 8.22085261e-01
3.96194160e-01 7.19419062e-01 -1.94729641e-01 -3.83614540e-01
6.35305882e-01 4.44825143e-01 1.01293075e+00 -2.09917709e-01
-6.60275996e-01 -3.33602488e-01 2.60309726e-01 -4.89691764e-01
-1.65127352e-01 -9.27346170e-01 -1.17845595e+00 -2.33527914e-01
1.29260466e-01 -6.53497398e-01 9.43801522e-01 8.86871517e-01
6.51386231e-02 2.31066018e-01 2.31544256e-01 -1.03215969e+00
-8.92956972e-01 -1.05820417e+00 -4.57764059e-01 6.23767734e-01
6.45473778e-01 -3.04178178e-01 -4.29249138e-01 2.72956073e-01] | [9.216592788696289, -6.540419101715088] |
c733eaa4-108a-4cd6-b1ff-805f6f916c16 | unrealtext-synthesizing-realistic-scene-text | 2003.10608 | null | https://arxiv.org/abs/2003.10608v6 | https://arxiv.org/pdf/2003.10608v6.pdf | UnrealText: Synthesizing Realistic Scene Text Images from the Unreal World | Synthetic data has been a critical tool for training scene text detection and recognition models. On the one hand, synthetic word images have proven to be a successful substitute for real images in training scene text recognizers. On the other hand, however, scene text detectors still heavily rely on a large amount of manually annotated real-world images, which are expensive. In this paper, we introduce UnrealText, an efficient image synthesis method that renders realistic images via a 3D graphics engine. 3D synthetic engine provides realistic appearance by rendering scene and text as a whole, and allows for better text region proposals with access to precise scene information, e.g. normal and even object meshes. The comprehensive experiments verify its effectiveness on both scene text detection and recognition. We also generate a multilingual version for future research into multilingual scene text detection and recognition. Additionally, we re-annotate scene text recognition datasets in a case-sensitive way and include punctuation marks for more comprehensive evaluations. The code and the generated datasets are released at: https://github.com/Jyouhou/UnrealText/ . | ['Shangbang Long', 'Cong Yao'] | 2020-03-24 | unrealtext-synthesizing-realistic-scene-text-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Long_UnrealText_Synthesizing_Realistic_Scene_Text_Images_From_the_Unreal_World_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Long_UnrealText_Synthesizing_Realistic_Scene_Text_Images_From_the_Unreal_World_CVPR_2020_paper.pdf | cvpr-2020-6 | ['scene-text-detection'] | ['computer-vision'] | [ 2.78331190e-01 -4.09584612e-01 1.55171901e-01 -5.03141642e-01
-7.86171019e-01 -6.25758469e-01 8.64771724e-01 -4.15878817e-02
-3.93084913e-01 2.18585819e-01 -2.68266685e-02 -2.93260545e-01
7.69509017e-01 -7.46516109e-01 -7.97543049e-01 -5.21830916e-01
8.19543004e-01 6.32906556e-01 4.73888099e-01 -1.98993653e-01
4.83304501e-01 4.29475009e-01 -1.65236259e+00 5.88123560e-01
1.11518526e+00 5.37274837e-01 5.88480771e-01 7.53297687e-01
-6.67507350e-01 5.07269263e-01 -7.44193196e-01 -5.66428542e-01
1.25262335e-01 -2.31023759e-01 -3.71024102e-01 5.63004136e-01
7.73861229e-01 -4.24957126e-01 -3.21689039e-01 1.17212343e+00
5.13922513e-01 -2.10635260e-01 7.42031574e-01 -1.00164878e+00
-6.04505897e-01 5.09106398e-01 -6.89897418e-01 -2.49812856e-01
5.81999660e-01 3.35508078e-01 4.87224460e-01 -1.37222409e+00
5.93194246e-01 1.20037866e+00 3.56208265e-01 3.68189812e-01
-8.51242244e-01 -7.13855982e-01 2.50584520e-02 6.40888810e-02
-1.56859946e+00 -6.31574035e-01 7.48326421e-01 -4.84752476e-01
7.37704217e-01 6.55577123e-01 4.43844408e-01 1.31536746e+00
-7.22680837e-02 1.14411807e+00 1.02735496e+00 -7.99249887e-01
-1.37144819e-01 6.03787541e-01 -1.05213150e-01 8.37533951e-01
2.03044266e-01 -4.06048954e-01 -3.73005331e-01 1.95580617e-01
7.64628470e-01 -2.27391481e-01 -4.20976967e-01 -2.41506889e-01
-1.29650962e+00 6.45143032e-01 7.54809938e-04 2.97794044e-01
1.10859044e-01 -2.14072317e-01 5.47258079e-01 6.11410625e-02
4.97062862e-01 1.20835260e-01 -8.65028985e-03 -4.78825942e-02
-9.42904115e-01 3.09767313e-02 4.88770097e-01 1.11643708e+00
5.16946971e-01 3.91743809e-01 -8.42169523e-02 1.15647042e+00
3.30507159e-01 1.04174900e+00 5.69658697e-01 -2.30240479e-01
8.04804265e-01 8.28422904e-01 2.22118851e-02 -1.10170794e+00
-6.94122957e-03 -5.18674143e-02 -6.80069447e-01 4.67545912e-02
3.84564042e-01 2.37308607e-01 -9.44892466e-01 8.89084876e-01
4.12153721e-01 -1.20405173e-02 -1.29697889e-01 1.03350747e+00
1.00853109e+00 8.58453929e-01 -2.42291942e-01 2.80859828e-01
1.41042614e+00 -1.12184310e+00 -7.60571241e-01 -1.71133772e-01
8.82372797e-01 -1.52010751e+00 1.67933381e+00 4.02594566e-01
-8.12931478e-01 -4.68451202e-01 -9.29347038e-01 -1.96869105e-01
-6.32734179e-01 7.17869222e-01 3.50413620e-01 1.01685667e+00
-8.04830909e-01 -8.64260718e-02 -7.74760664e-01 -5.71076095e-01
2.48388231e-01 -8.96386206e-02 -2.05095589e-01 -2.23580092e-01
-9.94822681e-01 7.84738183e-01 3.60245198e-01 7.10981935e-02
-6.76306248e-01 -1.86726809e-01 -8.90818596e-01 -3.95955265e-01
2.09896415e-01 -2.17606649e-01 1.18614578e+00 -1.06100452e+00
-1.49524415e+00 1.34003603e+00 -6.28014654e-02 1.76132452e-02
1.00530100e+00 -5.19235693e-02 -6.30608439e-01 1.63123116e-01
1.27765507e-01 5.65735042e-01 1.01343381e+00 -1.30119908e+00
-4.76393282e-01 -1.83800712e-01 -3.08680356e-01 5.48561931e-01
-3.88866454e-01 3.25067192e-01 -9.53426600e-01 -9.44102526e-01
7.91203603e-03 -6.91303849e-01 1.95546225e-01 2.89538741e-01
-7.08166778e-01 2.06949171e-02 1.07100487e+00 -9.70389187e-01
9.82384086e-01 -2.09573627e+00 -5.08127570e-01 1.92015674e-02
-6.74177483e-02 4.36749935e-01 -6.55426681e-02 4.97938782e-01
6.11021966e-02 1.68880612e-01 -9.21309814e-02 -5.53988934e-01
9.72407013e-02 -8.06920081e-02 -4.27503198e-01 6.01065814e-01
-5.15333526e-02 6.59848094e-01 -4.73270833e-01 -1.00556707e+00
8.10392797e-01 5.53881466e-01 -1.52024940e-01 1.14172120e-02
-4.67296541e-01 2.81385720e-01 -5.92663825e-01 7.79366612e-01
9.09649551e-01 -9.22148973e-02 -1.56097904e-01 -1.57259166e-01
-2.67330915e-01 6.35630498e-03 -1.49578643e+00 1.67819083e+00
-3.70680869e-01 1.02461553e+00 -5.98581806e-02 -5.55610597e-01
8.85256827e-01 1.98762879e-01 3.96952406e-02 -9.08532560e-01
3.08169752e-01 2.13775799e-01 -6.86997235e-01 -6.06589496e-01
1.09219909e+00 4.19862002e-01 -1.57858618e-02 4.01290894e-01
-5.72910011e-01 -5.87478042e-01 3.09165001e-01 3.84758323e-01
3.37907314e-01 2.37492934e-01 -1.28231365e-02 1.43033557e-03
6.22655451e-01 4.70680416e-01 1.04909480e-01 6.28998220e-01
1.06277585e-01 9.26025271e-01 1.37764156e-01 -3.15251648e-01
-1.42148030e+00 -8.49009156e-01 -3.45461011e-01 8.45624566e-01
2.94209659e-01 -3.45675170e-01 -1.07191217e+00 -4.47157919e-01
-2.61884153e-01 9.14069295e-01 -2.73078203e-01 4.03985739e-01
-4.84235376e-01 -6.05436325e-01 8.12189221e-01 2.64353275e-01
7.89845467e-01 -8.95331264e-01 -3.27927947e-01 -2.79266417e-01
-3.52299750e-01 -1.27816737e+00 -8.22989106e-01 -4.14571375e-01
-5.36553204e-01 -1.02757740e+00 -8.69681656e-01 -1.01187611e+00
8.19443166e-01 5.23064435e-01 9.23415005e-01 2.39198089e-01
-6.51264846e-01 2.25295767e-01 -5.65319061e-01 -3.24121088e-01
-7.69735336e-01 -2.06741378e-01 -1.14823952e-01 1.17432676e-01
8.09215382e-02 1.84305131e-01 -4.83424664e-01 7.09422052e-01
-1.06684768e+00 8.21788192e-01 3.76780063e-01 5.36059558e-01
6.07334673e-01 5.10745309e-02 -1.52468113e-02 -1.03716338e+00
5.38942516e-01 2.38935038e-01 -8.82141531e-01 4.32381958e-01
-1.20000206e-01 -3.49528790e-01 5.87799072e-01 -4.23887610e-01
-1.26462030e+00 2.06922993e-01 -2.55713016e-01 -3.49496990e-01
-4.09866273e-01 1.58635631e-01 -2.34277129e-01 -1.22125953e-01
8.28105748e-01 7.21326888e-01 -3.42363089e-01 -3.90128285e-01
3.99685591e-01 1.26312137e+00 3.99863541e-01 -4.64263946e-01
8.88562143e-01 4.19630378e-01 -6.98522627e-01 -1.32623589e+00
-4.40296650e-01 -5.13012290e-01 -6.30427003e-01 -4.23826396e-01
7.61716485e-01 -9.95673418e-01 -3.64905477e-01 8.63688827e-01
-1.08346546e+00 -4.77646858e-01 1.49410054e-01 3.33704591e-01
-2.21166998e-01 7.00563073e-01 -4.92331624e-01 -7.94143200e-01
-3.35457981e-01 -1.26284397e+00 1.68413699e+00 9.57830101e-02
2.13480532e-01 -7.72233069e-01 -2.00853452e-01 6.15747213e-01
1.10598303e-01 1.77130058e-01 5.99790454e-01 -3.48543674e-01
-6.41815007e-01 -4.48829263e-01 -5.63459396e-01 1.08955450e-01
-1.30317435e-02 4.98602331e-01 -1.15226018e+00 -1.17530279e-01
-2.35951528e-01 -2.96060801e-01 4.87575412e-01 2.73695923e-02
1.04588127e+00 -2.44657360e-02 -3.15219402e-01 4.50770795e-01
1.40192592e+00 5.63673638e-02 7.09764004e-01 3.31997365e-01
1.05141020e+00 4.98508632e-01 7.53003061e-01 4.16170120e-01
3.55560601e-01 8.34640443e-01 1.71803594e-01 -4.26567107e-01
-4.46686864e-01 -3.69883090e-01 2.60583788e-01 1.03598499e+00
4.40087616e-01 -5.14330387e-01 -1.17330539e+00 2.21105307e-01
-1.47417235e+00 -7.42654681e-01 -7.20271111e-01 2.17854428e+00
7.92736709e-01 -1.32331839e-02 -7.70601556e-02 2.46292353e-01
1.12120807e+00 1.42533937e-02 -3.67742270e-01 -2.34515727e-01
-4.80269998e-01 -1.62111551e-01 5.47549784e-01 3.90926510e-01
-1.05955577e+00 1.53357935e+00 5.28827572e+00 1.22471380e+00
-1.39571071e+00 5.97511046e-02 4.46068645e-01 2.89935589e-01
-1.56419486e-01 -1.92574397e-01 -7.91121542e-01 5.23016036e-01
3.74223083e-01 -2.47143097e-02 2.64826715e-01 8.63345563e-01
4.60848153e-01 -2.08487406e-01 -7.03227699e-01 1.41285348e+00
3.51781130e-01 -1.32553363e+00 4.42197055e-01 -1.75868392e-01
7.05586731e-01 1.42803848e-01 -6.01280518e-02 6.10765070e-02
5.09686619e-02 -8.97507012e-01 1.11267054e+00 3.17039192e-01
1.22224510e+00 -3.96988451e-01 3.74250263e-01 4.84010667e-01
-1.23498535e+00 6.13534391e-01 -5.00623763e-01 3.86767179e-01
-2.43565403e-02 4.81215239e-01 -1.26262939e+00 4.58708763e-01
4.05555546e-01 5.01191854e-01 -1.10678113e+00 1.04512668e+00
-2.52751201e-01 6.01764977e-01 -3.08131278e-01 -3.61130655e-01
5.06429300e-02 -3.06943238e-01 2.61610597e-01 1.54583442e+00
1.90038383e-01 -3.63287300e-01 3.23140472e-01 8.18070889e-01
3.86578999e-02 7.87818670e-01 -6.91723347e-01 -2.59600431e-01
3.41193765e-01 1.15860975e+00 -1.16087973e+00 -3.57727051e-01
-4.12847936e-01 1.24218667e+00 -6.08219318e-02 3.50920618e-01
-9.94938314e-01 -5.72277546e-01 -5.45687266e-02 1.02382511e-01
-1.04430310e-01 -3.25062484e-01 -4.33256149e-01 -1.64433742e+00
1.36032030e-01 -1.06783855e+00 5.48474006e-02 -1.27703524e+00
-8.23965847e-01 6.26126826e-01 -4.16207790e-01 -1.35271811e+00
8.42324793e-02 -7.32601225e-01 -4.14263725e-01 8.92204702e-01
-1.24177217e+00 -1.47054636e+00 -6.12848997e-01 7.41650522e-01
1.11571324e+00 -8.83729234e-02 8.46141875e-01 4.23520386e-01
-8.92998338e-01 7.13688672e-01 3.95807832e-01 4.88570273e-01
8.53826523e-01 -9.71836746e-01 7.90627122e-01 9.36629713e-01
4.01589364e-01 2.08020493e-01 7.84829974e-01 -7.64607489e-01
-1.63723004e+00 -1.09444511e+00 5.16386092e-01 -5.86747408e-01
4.24957126e-01 -8.36858571e-01 -9.10020888e-01 6.01626098e-01
1.93356499e-01 -5.21872580e-01 1.46969825e-01 -4.05006379e-01
-2.41818428e-01 2.55693525e-01 -1.01199770e+00 1.08710873e+00
6.19914114e-01 -4.98074412e-01 -1.92878440e-01 8.17210436e-01
4.14961725e-01 -6.81975007e-01 -5.00057280e-01 8.78734738e-02
4.72417951e-01 -9.88185644e-01 6.05826616e-01 2.45465696e-01
2.67739981e-01 -5.60612857e-01 -1.80922106e-01 -8.37401509e-01
3.80724490e-01 -2.50239015e-01 5.61163902e-01 1.35300922e+00
5.04593849e-01 -6.76476181e-01 9.11455750e-01 2.93091804e-01
-1.98155209e-01 -2.88878709e-01 -4.56239611e-01 -4.68731493e-01
-1.50117144e-01 -8.08307767e-01 3.41152549e-01 1.25212181e+00
-2.54915744e-01 2.14782596e-01 -3.51658732e-01 2.10431099e-01
4.66685712e-01 1.79663002e-01 1.11970603e+00 -7.12410033e-01
5.78093901e-03 -5.67784369e-01 -3.71396810e-01 -1.05640888e+00
6.39819950e-02 -1.01280797e+00 6.57403320e-02 -1.60215807e+00
2.36877233e-01 -5.61120808e-01 6.49168968e-01 8.39374065e-02
-6.17639907e-02 4.50195223e-01 7.78961554e-02 3.57670605e-01
-5.10847807e-01 4.71729845e-01 1.48022974e+00 -2.30019867e-01
1.21076241e-01 -1.24972045e-01 -2.18419671e-01 7.64774323e-01
9.89334941e-01 -2.91424483e-01 -2.87929565e-01 -7.38581479e-01
4.06508967e-02 -3.93535160e-02 2.05073044e-01 -7.74798691e-01
2.58530498e-01 -1.82049572e-01 5.73941827e-01 -1.00478423e+00
4.10396159e-01 -5.76367795e-01 -4.61112261e-02 2.32236296e-01
-1.53782636e-01 2.82054693e-02 2.83897996e-01 2.63140857e-01
-4.71757092e-02 -3.68499756e-01 8.14274251e-01 -2.61897780e-02
-7.67519414e-01 8.70123804e-02 -3.32211226e-01 -1.95286386e-02
9.06575263e-01 -5.72470427e-01 -4.93661165e-01 -3.53228718e-01
-1.55715108e-01 1.98444486e-01 1.11564946e+00 5.26875436e-01
7.81916380e-01 -9.75931466e-01 -7.96459258e-01 2.95584381e-01
4.45824414e-01 5.86131215e-02 2.84643888e-01 5.06861746e-01
-1.34277105e+00 4.96888429e-01 8.49281326e-02 -9.55013990e-01
-1.53366613e+00 4.70036924e-01 2.82076746e-01 1.34575620e-01
-8.32409561e-01 5.10729134e-01 4.42199856e-01 -6.45735681e-01
3.54055583e-01 -2.54075289e-01 7.05644190e-02 -2.84671217e-01
6.10671043e-01 8.26272070e-02 2.77804643e-01 -8.96137655e-01
-1.04658492e-01 7.28825629e-01 -6.57321140e-02 -2.60611981e-01
6.70330405e-01 -2.61096925e-01 1.80848524e-01 5.70704937e-01
7.60980844e-01 2.72274077e-01 -8.89001608e-01 -1.62849024e-01
-2.40852103e-01 -8.22670221e-01 1.01660732e-02 -8.34299326e-01
-8.65061820e-01 8.83515775e-01 7.31538236e-01 5.04495837e-02
9.36287284e-01 -2.74977565e-01 5.17994642e-01 3.45888168e-01
3.29585582e-01 -1.16555285e+00 9.01124775e-02 3.88444871e-01
9.54389095e-01 -1.27516782e+00 5.49827516e-02 -7.59016752e-01
-7.86462545e-01 1.05340278e+00 6.73343122e-01 2.18507051e-01
2.02468876e-02 3.73925000e-01 4.85447615e-01 2.58542900e-03
-2.89439231e-01 -1.00598425e-01 1.77679166e-01 4.95004952e-01
6.57379150e-01 1.85934961e-01 -5.24865882e-03 -1.59546807e-01
-3.16575527e-01 -5.22532940e-01 6.27766132e-01 7.94040203e-01
-3.69316339e-01 -1.08652687e+00 -8.58848274e-01 4.23788548e-01
-2.68169641e-01 -3.57451230e-01 -5.99554300e-01 9.83509004e-01
-2.78362125e-01 8.61485302e-01 -9.24134403e-02 -1.98281720e-01
3.96131128e-01 8.87032002e-02 4.57204014e-01 -6.76198065e-01
-4.36738193e-01 2.83692867e-01 1.93022355e-01 -4.94170748e-02
-3.74441668e-02 -5.75427532e-01 -1.21777451e+00 -4.29398179e-01
-6.70711398e-01 -1.05317660e-01 1.28043151e+00 5.87686896e-01
1.03824519e-01 3.45970809e-01 5.36386013e-01 -8.63512516e-01
-1.97455242e-01 -1.12630081e+00 -5.88129699e-01 4.07641798e-01
-2.13156998e-01 -3.37011933e-01 -1.08146816e-01 4.71928746e-01] | [11.993362426757812, 2.207533836364746] |
d079598b-323f-4063-a76d-c54b76b51634 | can-chatgpt-defend-the-truth-automatic | 2305.13160 | null | https://arxiv.org/abs/2305.13160v1 | https://arxiv.org/pdf/2305.13160v1.pdf | Can ChatGPT Defend the Truth? Automatic Dialectical Evaluation Elicits LLMs' Deficiencies in Reasoning | We explore testing the reasoning ability of large language models (LLMs), such as ChatGPT, by engaging with them in a debate-like conversation that probes deeper into their understanding of the subject. Specifically, we formulate a new task where given a question, the LLM can generate a correct solution while the user believes in a wrong solution in the beginning, and they need to discuss to make the correct decision through dialogue. Such a setting requires the LLM to not only achieve the correct answer on its own (which could be done by shallow memorization), but also be able to defend the truth instead of blindly believing or getting misled by the user's (invalid) arguments and critiques, thus testing in greater depth whether the LLM grasps the essence of the reasoning required to solve the problem. To automate this evaluation framework and save human labor, we simulate the user using another LLM conditioned on a synthesized wrong solution. Across a range of complex reasoning benchmarks spanning math, commonsense, logic and tasks from BIG-Bench, we find that despite being able to generate correct step-by-step solutions in the beginning, ChatGPT cannot maintain its belief in truth for a significant portion of examples when challenged by often-time absurdly invalid arguments. Our work reveals LLMs' weaknesses not captured by conventional benchmarking, and also points to danger zones of aligning models with human feedback. | ['Huan Sun', 'Xiang Yue', 'Boshi Wang'] | 2023-05-22 | null | null | null | null | ['memorization'] | ['natural-language-processing'] | [ 2.28817359e-01 1.01347005e+00 3.33903432e-01 -2.82810211e-01
-1.06389511e+00 -1.02639854e+00 6.93899035e-01 4.10077423e-01
-1.14171892e-01 7.18805373e-01 2.57973671e-01 -1.19781971e+00
1.03895381e-01 -9.56834376e-01 -7.06425846e-01 -4.25346456e-02
4.98969615e-01 7.32848048e-01 1.71593487e-01 -5.12690842e-01
4.83013064e-01 -1.30275607e-01 -9.32262123e-01 7.27166116e-01
9.01295185e-01 3.53736341e-01 -1.30229875e-01 9.61980999e-01
-1.83806539e-01 1.69348884e+00 -7.66750276e-01 -9.91909504e-01
1.85927242e-01 -7.27725863e-01 -1.66857255e+00 -1.89169094e-01
6.40736103e-01 -3.82321328e-01 2.09202096e-01 1.26245666e+00
1.49312139e-01 -2.05577925e-01 4.26977485e-01 -1.25236535e+00
-2.88227320e-01 1.20626450e+00 -7.37762377e-02 -1.72497615e-01
8.95687103e-01 8.98632944e-01 1.33607352e+00 -3.70104522e-01
6.84498489e-01 1.63801801e+00 6.63497865e-01 8.11860800e-01
-1.40506840e+00 -3.63718271e-01 1.32606745e-01 5.99692911e-02
-6.36514306e-01 -5.05303383e-01 3.65903258e-01 -4.32332098e-01
8.26614439e-01 6.85894549e-01 3.92282039e-01 1.17155504e+00
1.88489795e-01 4.27506089e-01 1.37077844e+00 -3.59170318e-01
3.74236554e-01 3.52295667e-01 3.14988643e-01 7.41304696e-01
2.97560245e-01 -2.24044457e-01 -4.53431845e-01 -4.50668335e-01
2.19259650e-01 -7.67529905e-01 -3.41723680e-01 7.50945434e-02
-1.16960788e+00 7.46646762e-01 2.40088940e-01 2.23155603e-01
-2.07653984e-01 -8.82540867e-02 2.58087218e-01 8.60562444e-01
-1.87788755e-01 1.23053062e+00 -5.32285869e-01 -4.64838713e-01
-8.29584420e-01 7.77463436e-01 1.62006903e+00 4.39714164e-01
3.12671065e-01 -3.67074251e-01 -1.88135445e-01 1.14929251e-01
2.50621796e-01 2.08113104e-01 2.85702378e-01 -1.60053039e+00
6.41275883e-01 9.41565990e-01 6.10885262e-01 -1.05532706e+00
-1.98232591e-01 -3.16849500e-01 -2.85572290e-01 6.20944858e-01
1.18113303e+00 -3.32696080e-01 -9.11774486e-02 2.02090764e+00
2.96768755e-01 -1.25523880e-01 4.47893023e-01 9.56927299e-01
5.70388854e-01 5.32289505e-01 -9.29232389e-02 -3.17255408e-02
1.30991340e+00 -7.35809505e-01 -2.93025732e-01 -5.32031536e-01
1.06029141e+00 -5.27721584e-01 1.54838097e+00 6.49144590e-01
-1.68126583e+00 -2.28179067e-01 -9.96583700e-01 -2.90950328e-01
2.80402422e-01 -6.41638935e-01 3.67990941e-01 3.69347960e-01
-1.01257193e+00 6.55306637e-01 -5.30270457e-01 -1.73353732e-01
1.86141625e-01 -2.01240972e-01 -2.42971435e-01 -2.33408272e-01
-1.40319955e+00 1.21686888e+00 1.78678900e-01 1.93418697e-01
-8.80479991e-01 -7.24614382e-01 -7.02164054e-01 3.02829862e-01
6.40151858e-01 -7.29168534e-01 1.82508123e+00 -8.81780505e-01
-1.45245624e+00 1.01653385e+00 -4.19327728e-02 -6.36417508e-01
1.29553497e+00 -1.38280243e-01 9.96343568e-02 -3.05765867e-02
2.11181611e-01 4.87199724e-01 2.63741434e-01 -1.13674581e+00
-2.79941887e-01 -1.90078169e-01 1.07280576e+00 1.32401779e-01
4.26236480e-01 -1.05550578e-02 2.94746101e-01 4.53280751e-03
2.74487734e-01 -8.56935620e-01 -4.20759432e-02 3.84159274e-02
-7.79347837e-01 -1.67028785e-01 4.40334618e-01 -6.61361277e-01
1.01270604e+00 -1.84007466e+00 6.51756525e-02 1.90190569e-01
6.06879890e-01 2.55176246e-01 -1.27626583e-01 6.29591763e-01
1.06000239e-02 6.14169300e-01 -8.76263008e-02 -1.69002473e-01
3.75680476e-01 1.60703193e-02 -6.53208435e-01 1.86003566e-01
1.08350635e-01 1.09627223e+00 -1.00503314e+00 -3.83267015e-01
-4.13904041e-02 -1.95189238e-01 -9.24558043e-01 4.21728313e-01
-8.04167449e-01 4.50639069e-01 -1.89319819e-01 2.17225447e-01
4.49069530e-01 -7.65182674e-01 4.98854369e-01 1.56339988e-01
2.72742867e-01 1.06728494e+00 -1.36646676e+00 1.08899260e+00
-4.63384002e-01 6.24563813e-01 2.50488698e-01 -7.38455951e-01
4.34922546e-01 2.30663195e-01 -7.00305045e-01 -5.85280538e-01
1.38807535e-01 2.98076838e-01 5.79426467e-01 -5.30051589e-01
1.31764501e-01 -4.52864766e-01 -5.37499972e-02 9.84795094e-01
-4.19284672e-01 -5.44279516e-01 7.10607693e-02 7.73659825e-01
1.24867606e+00 -1.85003608e-01 1.08372718e-01 -2.65346020e-01
8.03544164e-01 3.07126671e-01 3.25591832e-01 1.26326847e+00
7.64490739e-02 2.71055102e-01 1.33186567e+00 -4.40891027e-01
-8.23251367e-01 -9.11305845e-01 3.94112468e-01 9.03731346e-01
-1.09612700e-02 -5.63796222e-01 -9.14367437e-01 -7.01265454e-01
-1.14374280e-01 1.40362740e+00 -4.96896684e-01 -2.25962833e-01
-6.46335781e-01 -2.87003219e-02 8.83468211e-01 1.20904274e-01
7.51846075e-01 -1.04315364e+00 -1.07588077e+00 1.79942921e-01
-7.18880177e-01 -9.92009878e-01 8.26897025e-02 -1.59092948e-01
-6.61833107e-01 -1.53117347e+00 1.88296556e-01 -1.58873349e-01
6.41659737e-01 -1.45428494e-01 1.36730933e+00 9.02946293e-01
2.23288432e-01 2.53252238e-01 -1.04479507e-01 -2.50146121e-01
-1.30332935e+00 -5.39906472e-02 -4.28727865e-01 -4.81679469e-01
6.46509826e-02 -5.11625707e-01 -3.03770185e-01 4.27397698e-01
-5.98501384e-01 4.39358622e-01 1.81870013e-01 6.91918314e-01
-3.90610844e-01 -5.56919500e-02 2.36820251e-01 -1.29162991e+00
1.02859545e+00 -3.13911647e-01 -6.79016709e-01 4.39417332e-01
-4.21530902e-01 2.90520042e-01 7.63745427e-01 -2.75801182e-01
-8.80466163e-01 -8.74533057e-01 2.29576081e-02 2.77150512e-01
1.21880390e-01 6.24992669e-01 -2.70537473e-02 1.38535380e-01
1.13370967e+00 1.00838244e-01 2.31876373e-01 -1.36492491e-01
3.69536340e-01 1.73407272e-01 6.15411043e-01 -1.42200470e+00
1.01503038e+00 8.35921150e-03 -2.54002601e-01 -1.27746388e-01
-1.20931602e+00 3.74213845e-01 -1.18053379e-02 -3.20553519e-02
4.43681926e-01 -7.30151713e-01 -1.53594482e+00 3.99418205e-01
-1.45517135e+00 -8.78383219e-01 -1.83455735e-01 -4.01343107e-02
-4.33182329e-01 3.59807283e-01 -7.67751038e-01 -8.00983012e-01
-1.79989338e-01 -1.24746871e+00 4.70385015e-01 9.65880305e-02
-1.11737990e+00 -8.27926040e-01 -2.03319624e-01 7.93526471e-01
5.66811919e-01 6.09714761e-02 1.51523602e+00 -1.06079328e+00
-5.88385046e-01 -1.85617879e-01 -4.86043021e-02 4.04912710e-01
-1.25133008e-01 -8.50998536e-02 -8.95208716e-01 -5.95882423e-02
3.51511925e-01 -8.62262964e-01 1.04349546e-01 -3.92386287e-01
7.21387088e-01 -7.74831295e-01 1.76542208e-01 6.41963333e-02
8.27310801e-01 -3.17466229e-01 7.17613697e-01 3.06006610e-01
1.27481282e-01 7.79293776e-01 2.50146061e-01 -1.95473600e-02
6.86662197e-01 1.56803191e-01 2.81373292e-01 2.67638236e-01
2.34470293e-01 -4.95989144e-01 2.30955139e-01 6.80680871e-02
2.60740042e-01 -1.66865870e-01 -1.08820236e+00 1.80426553e-01
-1.91218221e+00 -9.75123763e-01 -3.24269980e-01 2.06593752e+00
1.46537960e+00 7.60221004e-01 -2.11667851e-01 3.83099198e-01
2.60055333e-01 1.97663620e-01 -5.06443024e-01 -7.65117466e-01
1.01257652e-01 1.80174485e-01 -3.05039942e-01 1.23604465e+00
-2.53498971e-01 9.59180295e-01 5.84840107e+00 3.46286058e-01
-1.10321021e+00 -8.19212720e-02 7.64166772e-01 1.10918403e-01
-7.89136350e-01 6.34299636e-01 -2.75566518e-01 2.57981122e-01
8.82091343e-01 -3.32090348e-01 6.23760462e-01 5.42072296e-01
1.10963516e-01 -5.59675992e-01 -1.65705943e+00 2.56129891e-01
-2.37735361e-01 -1.54256833e+00 3.79124284e-02 -2.47080460e-01
3.08122605e-01 -4.41706687e-01 -1.80116698e-01 5.60400367e-01
8.03076506e-01 -1.38133824e+00 1.02149665e+00 2.72580922e-01
2.11897045e-01 -3.04420203e-01 7.64266729e-01 1.14506269e+00
-3.41668010e-01 1.97068807e-02 5.31998761e-02 -9.76362228e-01
1.99183896e-02 2.38222420e-01 -1.06905043e+00 6.11470714e-02
3.64522040e-02 -8.73067528e-02 -7.28148162e-01 2.97315836e-01
-8.07560444e-01 7.14582324e-01 -2.09045574e-01 -1.69515461e-01
3.29621613e-01 -3.57889757e-02 4.50453281e-01 8.24747324e-01
-2.25281134e-01 5.48622191e-01 1.02987647e-01 1.48615956e+00
-1.41234651e-01 -3.40778887e-01 -1.11591078e-01 -1.80836350e-01
4.66133446e-01 1.12544465e+00 -3.30689490e-01 -6.09036982e-01
1.82946473e-02 6.38142586e-01 4.10384834e-01 2.69605994e-01
-5.93177199e-01 8.52596853e-03 4.71679270e-01 3.18896055e-01
-4.46195304e-01 -8.71094540e-02 -5.53174436e-01 -1.07153070e+00
2.17234820e-01 -1.71512067e+00 1.98930368e-01 -1.31293905e+00
-1.04047954e+00 3.14669371e-01 -2.35689312e-01 -3.66772801e-01
-4.39286709e-01 -4.76208061e-01 -1.06701791e+00 1.22957432e+00
-1.21762693e+00 -7.05887735e-01 -2.85363734e-01 3.41219902e-01
3.62629265e-01 5.55233002e-01 7.29130924e-01 -3.08722466e-01
-1.16175935e-01 4.13071990e-01 -1.08124626e+00 2.85553873e-01
6.37780964e-01 -1.30355608e+00 6.80900991e-01 7.05541134e-01
-3.42827439e-02 1.18404055e+00 1.31978416e+00 -6.45081699e-01
-1.33641005e+00 -4.31503981e-01 1.33850145e+00 -9.62247550e-01
9.32096720e-01 -2.69643813e-01 -1.31044412e+00 9.27138329e-01
1.63944975e-01 -3.72379273e-01 2.72290051e-01 8.77612829e-02
-7.94623196e-01 3.87027085e-01 -1.36360276e+00 9.11843777e-01
8.18017960e-01 -8.51181686e-01 -1.30542064e+00 4.55834538e-01
6.53168380e-01 -7.68762887e-01 -2.57928520e-01 7.67117739e-02
4.71878529e-01 -1.21647406e+00 5.57509542e-01 -1.24096107e+00
9.48524415e-01 -3.10330361e-01 -1.54546484e-01 -1.10945368e+00
1.76903471e-01 -8.40238690e-01 1.93208367e-01 1.04512882e+00
6.50595963e-01 -8.37356746e-01 5.08122206e-01 1.43859982e+00
1.79188624e-01 -6.37415886e-01 -6.82322681e-01 -1.50453061e-01
3.59200656e-01 -7.01657593e-01 4.62130636e-01 9.91242528e-01
5.61792016e-01 4.07730669e-01 4.27475721e-02 3.46231908e-01
3.00627530e-01 1.79874077e-01 1.06042659e+00 -1.17346120e+00
-6.12457037e-01 -5.20550668e-01 1.72009125e-01 -9.09583032e-01
2.09196761e-01 -9.00524497e-01 9.28551480e-02 -1.66139686e+00
1.45173877e-01 -3.33962083e-01 3.14721465e-01 6.63776815e-01
-2.51477778e-01 -2.57565945e-01 5.15443027e-01 7.37117007e-02
-5.93541265e-01 -2.10938007e-01 1.19596732e+00 -3.17343734e-02
1.36872232e-01 -4.25868370e-02 -1.30546832e+00 1.18858242e+00
4.93308395e-01 -3.21996093e-01 -4.19308513e-01 -3.19491118e-01
1.04788268e+00 5.03591895e-01 9.45382357e-01 -8.06127310e-01
5.29148161e-01 -4.38305557e-01 1.02079322e-03 1.14450872e-01
-1.22160889e-01 -4.38746363e-01 7.98243433e-02 7.90889144e-01
-9.17863190e-01 2.56432474e-01 1.59939677e-01 -1.70900039e-02
4.64658476e-02 -2.84803778e-01 6.96211338e-01 -5.32941639e-01
-1.75800502e-01 -5.83561420e-01 -4.08076555e-01 6.71885073e-01
5.08826971e-01 -2.08679736e-02 -7.61176705e-01 -6.87332869e-01
-7.29708493e-01 5.74827194e-01 7.10679531e-01 2.99206860e-02
3.11397046e-01 -5.83131313e-01 -8.81157756e-01 -5.29141258e-03
-1.85248345e-01 1.02383077e-01 8.73964354e-02 9.03998435e-01
-6.66327000e-01 3.60776819e-02 1.52333677e-01 -4.28287804e-01
-1.08961475e+00 1.73037544e-01 7.71865845e-01 -4.85838801e-01
-5.53154588e-01 9.23955023e-01 1.80650383e-01 -7.02163458e-01
4.48978916e-02 -5.64845145e-01 1.92318752e-01 -1.69365063e-01
5.56506515e-01 -4.34530303e-02 8.13097656e-02 1.13215029e-01
-3.45640779e-01 1.51819959e-01 -3.15486267e-02 -4.16034788e-01
9.76481497e-01 4.96837944e-02 -3.40435594e-01 4.50641245e-01
5.39650679e-01 4.16788548e-01 -8.79592597e-01 -2.75967777e-01
9.23405588e-03 -4.78307486e-01 -5.57807446e-01 -1.53533757e+00
-3.40147167e-01 8.92400265e-01 -4.62831378e-01 4.23459828e-01
3.52439106e-01 -3.87462564e-02 5.22873580e-01 9.92040992e-01
5.37602484e-01 -7.08062232e-01 2.67784923e-01 6.05566502e-01
1.07256532e+00 -1.19791615e+00 -1.17840707e-01 -1.82749420e-01
-8.38323832e-01 1.08186877e+00 8.43887329e-01 6.36446057e-05
-1.28680840e-01 2.72202313e-01 3.78706038e-01 -2.89706409e-01
-1.27422583e+00 5.80915809e-01 -2.35269427e-01 -6.04241453e-02
4.33002591e-01 3.30820754e-02 -1.30432919e-01 4.80873048e-01
-7.90774941e-01 -1.35032386e-01 9.34137404e-01 8.22296858e-01
-4.18081462e-01 -9.03591096e-01 -4.24101770e-01 9.48423371e-02
-3.97650898e-01 -2.68717557e-01 -7.43294418e-01 7.87283540e-01
-1.97475985e-01 1.21700561e+00 -2.77853191e-01 -4.66161743e-02
2.46640861e-01 3.76135319e-01 4.73161042e-01 -7.86448061e-01
-9.54434752e-01 -5.08211076e-01 5.96275449e-01 -7.40024328e-01
8.13344195e-02 -3.39297414e-01 -1.39326108e+00 -8.43136847e-01
-1.57144815e-02 4.18324143e-01 2.31990978e-01 1.45127070e+00
1.72145531e-01 2.32981920e-01 1.97979569e-01 -1.93831369e-01
-1.16442621e+00 -8.77723873e-01 1.70639455e-01 5.41906297e-01
4.50699598e-01 -3.11157876e-03 -7.19983220e-01 -2.73440510e-01] | [9.415384292602539, 7.213832855224609] |
366373a5-60b5-4c60-930c-c8474c4ad415 | masc-multi-scale-affinity-with-sparse | 1902.04478 | null | http://arxiv.org/abs/1902.04478v1 | http://arxiv.org/pdf/1902.04478v1.pdf | MASC: Multi-scale Affinity with Sparse Convolution for 3D Instance Segmentation | We propose a new approach for 3D instance segmentation based on sparse
convolution and point affinity prediction, which indicates the likelihood of
two points belonging to the same instance. The proposed network, built upon
submanifold sparse convolution [3], processes a voxelized point cloud and
predicts semantic scores for each occupied voxel as well as the affinity
between neighboring voxels at different scales. A simple yet effective
clustering algorithm segments points into instances based on the predicted
affinity and the mesh topology. The semantic for each instance is determined by
the semantic prediction. Experiments show that our method outperforms the
state-of-the-art instance segmentation methods by a large margin on the widely
used ScanNet benchmark [2]. We share our code publicly at
https://github.com/art-programmer/MASC. | ['Yasutaka Furukawa', 'Chen Liu'] | 2019-02-12 | null | null | null | null | ['3d-instance-segmentation-1'] | ['computer-vision'] | [-3.86174135e-02 2.81598389e-01 -1.45253539e-01 -5.19406736e-01
-6.74268544e-01 -3.18511784e-01 4.03276235e-01 1.14492029e-01
-2.09329173e-01 3.92307103e-01 -1.42977700e-01 1.25570118e-01
-1.65954828e-01 -1.03036129e+00 -1.11113429e+00 -4.92350757e-01
-2.99378335e-01 1.27391481e+00 4.60272282e-01 3.16458225e-01
4.66542393e-01 8.98936212e-01 -1.59073234e+00 3.11666191e-01
7.62823582e-01 1.32081175e+00 3.05520922e-01 2.91988879e-01
-5.47655940e-01 6.78058118e-02 -2.24132948e-02 -3.80259752e-02
4.13651139e-01 6.14093505e-02 -1.19622540e+00 6.09888174e-02
5.19003451e-01 1.28467381e-01 -1.77856490e-01 1.11460221e+00
-8.96655992e-02 2.66181856e-01 7.97717750e-01 -9.72457588e-01
-4.01330411e-01 6.12532616e-01 -6.22813046e-01 4.21319678e-02
-2.04196289e-01 7.02759484e-03 9.14668858e-01 -1.11806250e+00
7.47301102e-01 9.88228679e-01 8.02760422e-01 3.00628424e-01
-1.21328330e+00 -5.31288207e-01 8.98206756e-02 2.10392311e-01
-1.50433731e+00 1.73151836e-01 8.54357123e-01 -6.11879230e-01
8.30077291e-01 3.07767183e-01 9.15419281e-01 4.96449143e-01
-6.06662072e-02 6.30559385e-01 1.03239787e+00 1.70919105e-01
5.57412446e-01 -1.55908883e-01 2.67312318e-01 8.32988381e-01
-9.02795941e-02 -2.31504336e-01 -1.66330427e-01 -2.69069403e-01
1.17672586e+00 2.07623377e-01 -8.28085989e-02 -6.22773111e-01
-1.37786043e+00 7.18498886e-01 8.72490942e-01 2.85766840e-01
-5.44711173e-01 4.77271885e-01 5.38019575e-02 -2.12340117e-01
6.80437624e-01 2.72959441e-01 -6.28803253e-01 1.14317477e-01
-1.17208409e+00 1.96422562e-01 7.96456635e-01 7.32157648e-01
1.21552050e+00 -5.18579304e-01 -1.59194451e-02 7.95724034e-01
4.67913985e-01 1.58266723e-01 -4.16024551e-02 -1.38647413e+00
4.10333350e-02 9.42983687e-01 -1.04657479e-01 -6.75564528e-01
-3.89434159e-01 -4.19676125e-01 -7.80706346e-01 3.38105917e-01
4.29915100e-01 2.31778026e-01 -1.34055650e+00 1.22188425e+00
6.65060580e-01 9.40522552e-01 -4.29958463e-01 1.10950923e+00
8.51029873e-01 5.79830766e-01 1.24170734e-02 4.12884772e-01
1.19795442e+00 -9.55753386e-01 3.92126665e-02 -8.90076905e-02
2.47284144e-01 -5.90748727e-01 8.08990598e-01 1.54491171e-01
-1.21224630e+00 -4.36218351e-01 -7.26343036e-01 7.31887147e-02
-1.89759552e-01 -1.37004063e-01 6.74173295e-01 1.85584635e-01
-1.06917942e+00 1.17551458e+00 -1.32044959e+00 -3.52523804e-01
9.26553011e-01 6.31393254e-01 -1.16986744e-01 1.30669683e-01
-5.84849179e-01 4.63160455e-01 1.33153170e-01 -7.58330747e-02
-6.13280654e-01 -1.21450365e+00 -6.79413855e-01 1.21415496e-01
-8.74643177e-02 -1.01495516e+00 1.01280558e+00 -7.97903478e-01
-1.19932055e+00 1.08572781e+00 -2.15956405e-01 -5.07804394e-01
5.87705553e-01 3.56128700e-02 2.26535752e-01 3.30809563e-01
2.67574281e-01 1.08371484e+00 6.61894917e-01 -1.47268510e+00
-5.70241988e-01 -6.44297838e-01 -1.20299272e-01 2.27016881e-01
2.19966114e-01 -4.40942526e-01 -7.34566689e-01 -3.04401338e-01
7.11456120e-01 -8.25297475e-01 -6.76740468e-01 1.74033493e-01
-6.55640662e-01 -3.88244957e-01 7.22637653e-01 -4.49021369e-01
3.80374044e-01 -2.07518315e+00 3.15415293e-01 7.60011554e-01
4.25573736e-01 -2.48308077e-01 1.33358777e-01 5.96675975e-03
-4.02493067e-02 1.99493408e-01 -7.42802501e-01 -4.39259320e-01
1.89532954e-02 3.92482020e-02 1.01716869e-01 7.12118864e-01
4.54942137e-02 8.89571130e-01 -6.77213490e-01 -5.42034864e-01
5.46390235e-01 7.73707390e-01 -5.07926166e-01 4.64148298e-02
-4.38242614e-01 7.81560600e-01 -5.50207555e-01 6.00364089e-01
7.71097898e-01 -7.02639818e-01 -1.88395247e-01 -2.51539856e-01
-1.85256749e-01 1.92405343e-01 -1.24020326e+00 2.02174330e+00
-1.65220574e-01 1.74312919e-01 1.57321364e-01 -9.35767293e-01
8.35307062e-01 5.53330779e-02 9.37243164e-01 -1.70711085e-01
1.26885220e-01 3.27262670e-01 -3.04226428e-01 -1.02751218e-01
1.66911870e-01 -4.59584258e-02 2.31564283e-01 1.74235940e-01
3.19047756e-02 -3.20802093e-01 -9.78834033e-02 1.32161051e-01
1.15163863e+00 9.78246927e-02 -2.50514477e-01 -4.74001616e-01
3.96776229e-01 4.08501893e-01 3.68913442e-01 5.40789425e-01
1.77872941e-01 8.98092210e-01 2.55495578e-01 -5.70570886e-01
-1.04508162e+00 -1.33846402e+00 -5.05161762e-01 5.49931705e-01
6.63851261e-01 -3.24580409e-02 -1.06414509e+00 -5.62597334e-01
2.79786378e-01 4.66662019e-01 -8.24801803e-01 3.47536772e-01
-5.59211016e-01 -4.68417615e-01 7.84829631e-02 6.49840236e-01
3.61568898e-01 -1.22793472e+00 -3.51013929e-01 -1.15779703e-02
-3.74851027e-03 -8.60546529e-01 -2.59996891e-01 1.97897270e-01
-1.27191973e+00 -1.01895916e+00 -5.57585180e-01 -1.00259256e+00
1.05419075e+00 3.67179215e-02 1.13804626e+00 4.27786022e-01
-3.24097872e-01 4.23399687e-01 1.74163394e-02 2.25176848e-02
-1.40623869e-02 3.52470249e-01 -2.35021323e-01 -1.73877254e-02
2.53666818e-01 -8.80131960e-01 -1.03931010e+00 2.55890220e-01
-5.19513786e-01 2.77936459e-01 4.39175814e-01 3.97681385e-01
1.58559799e+00 -1.40925348e-01 -5.14363162e-02 -1.05460835e+00
-1.41361386e-01 -7.61055887e-01 -5.69993198e-01 -1.63246125e-01
-3.44892740e-01 -6.66936189e-02 1.57929838e-01 -1.04923651e-01
-7.35044658e-01 4.54877287e-01 -2.68186659e-01 -7.32305944e-01
-8.30376267e-01 3.31093043e-01 1.83833256e-01 -2.63760656e-01
2.77917892e-01 3.72473449e-02 -1.11435153e-01 -6.20910466e-01
4.73574668e-01 2.01096267e-01 4.75540638e-01 -7.36062348e-01
8.35223734e-01 9.51101720e-01 2.08030954e-01 -6.09217644e-01
-6.61361098e-01 -7.18443990e-01 -1.06335711e+00 -3.22936386e-01
1.17982674e+00 -7.50157654e-01 -5.92361987e-01 2.52676964e-01
-1.21570134e+00 -5.87152541e-01 -4.88167316e-01 4.37832743e-01
-7.98167825e-01 1.15566790e-01 -7.82569051e-01 -2.67445356e-01
-3.04306269e-01 -1.25633347e+00 1.30634987e+00 1.73124671e-01
-1.83249712e-01 -8.39137673e-01 5.45128882e-02 6.18830681e-01
1.02043152e-01 2.44820684e-01 9.51175451e-01 -5.56860209e-01
-1.17903447e+00 -7.77384788e-02 -3.71745527e-01 1.13671295e-01
-3.02864730e-01 -2.02257764e-02 -6.67306900e-01 3.31569165e-02
-1.33238629e-01 1.45889968e-01 9.07038152e-01 9.25706685e-01
1.75141740e+00 1.96241606e-02 -6.74597859e-01 9.55366433e-01
1.63105333e+00 -2.11809576e-01 6.88497424e-01 1.51000544e-01
1.15343142e+00 3.80618185e-01 3.66207242e-01 1.39335841e-01
3.34389865e-01 4.18496847e-01 7.16016293e-01 -3.24315161e-01
-9.43683833e-02 5.85709698e-02 -3.32112610e-01 5.70286274e-01
-3.06478024e-01 2.13671625e-01 -1.14826524e+00 6.43087924e-01
-1.96205544e+00 -7.08122075e-01 -6.17534161e-01 1.99500537e+00
4.98608708e-01 1.74046218e-01 1.56793408e-02 -2.94286877e-01
9.24004138e-01 -1.27532125e-01 -6.44966781e-01 -7.72331953e-02
3.21650296e-01 7.97604680e-01 6.64650977e-01 5.69702029e-01
-1.09014070e+00 9.34241414e-01 5.57172298e+00 8.61696064e-01
-9.22758222e-01 3.19286525e-01 1.06260049e+00 -1.58627048e-01
-2.84960717e-01 -1.54207563e-02 -5.12925982e-01 4.43404257e-01
6.17211223e-01 1.82827637e-01 4.62110758e-01 1.02240348e+00
2.70825978e-02 -4.21231799e-02 -1.00260210e+00 6.87821448e-01
-1.86761424e-01 -1.91044319e+00 -5.14918659e-03 2.35491782e-01
9.56498206e-01 6.96954012e-01 -2.80382354e-02 -2.47615978e-01
7.48774931e-02 -1.05650091e+00 7.61520565e-01 6.44725740e-01
6.43014014e-01 -8.09428692e-01 4.59146112e-01 1.66664436e-01
-1.30507696e+00 3.40098083e-01 -3.21854651e-01 8.48714784e-02
1.47043198e-01 7.68485188e-01 -6.80520892e-01 2.10427418e-01
1.00023651e+00 6.62875354e-01 -2.24090949e-01 1.30761909e+00
-1.21359117e-01 7.62560368e-01 -4.77687716e-01 2.84655780e-01
4.00624812e-01 -5.03597021e-01 5.86242974e-01 9.78392422e-01
3.61330450e-01 2.76110947e-01 3.78002942e-01 1.46025920e+00
-2.14919686e-01 2.01169789e-01 -2.67654508e-01 4.83233660e-01
5.19954324e-01 1.47284389e+00 -1.40984583e+00 -3.31974655e-01
-3.19508016e-01 9.66451705e-01 4.52509016e-01 1.98191181e-01
-7.56996214e-01 -4.16621082e-02 1.00702429e+00 4.44900274e-01
5.77572942e-01 -2.84540385e-01 -9.68369186e-01 -6.79552913e-01
-1.30300391e-02 1.08428905e-02 9.67845693e-02 -6.76823318e-01
-1.47913778e+00 5.10396063e-01 -2.46243924e-01 -9.91535425e-01
2.32038796e-01 -5.72708428e-01 -8.45298231e-01 8.30613196e-01
-1.29378629e+00 -9.99919057e-01 -3.81451964e-01 5.68133831e-01
5.60061336e-01 2.97408968e-01 5.20365953e-01 7.95906112e-02
-3.05886179e-01 -1.19960144e-01 3.89885763e-03 1.59113675e-01
-1.21902777e-02 -1.45717835e+00 6.10064149e-01 2.34634832e-01
1.60637349e-01 4.42104161e-01 4.24690336e-01 -9.25090909e-01
-1.05549324e+00 -1.28812730e+00 5.37194014e-01 -5.17050147e-01
5.99671900e-01 -2.93342769e-01 -1.19316542e+00 6.13213599e-01
-2.08759055e-01 2.40160689e-01 5.40564299e-01 7.19574019e-02
6.37945719e-03 2.65039563e-01 -1.37040865e+00 3.43777239e-01
1.32166517e+00 -2.05175385e-01 -4.80664968e-01 6.05808079e-01
7.68191576e-01 -6.27031982e-01 -1.25860250e+00 5.33544302e-01
1.78353041e-01 -9.57123160e-01 1.29723227e+00 -2.93861866e-01
7.02943504e-01 -3.43778491e-01 -1.16968580e-01 -1.03908336e+00
-6.27887964e-01 7.14071691e-02 -6.49608970e-02 7.57499814e-01
4.93429422e-01 -3.20110530e-01 1.34832537e+00 6.76846623e-01
-4.82119352e-01 -1.31472480e+00 -1.22153866e+00 -6.30411863e-01
1.40597492e-01 -6.90050006e-01 8.50673676e-01 9.30893958e-01
-3.52728277e-01 -9.96174291e-02 3.51287425e-01 7.04527974e-01
1.02402639e+00 5.32403767e-01 4.06171679e-01 -1.69759691e+00
-3.45677622e-02 -6.79885745e-01 -6.28821790e-01 -7.76928186e-01
3.69429559e-01 -1.39657605e+00 6.07399717e-02 -1.89943171e+00
1.84921831e-01 -7.26114213e-01 -3.04288507e-01 3.88193965e-01
6.44862503e-02 7.49439538e-01 -9.20714512e-02 4.78408068e-01
-6.44845247e-01 3.50331753e-01 1.15003407e+00 -1.68186739e-01
-2.94738084e-01 -9.01360586e-02 -1.93086073e-01 8.96573663e-01
9.93116856e-01 -4.31479782e-01 4.26209718e-03 -4.37057108e-01
-1.87855974e-01 -5.45495935e-03 7.78390944e-01 -1.38787997e+00
2.15664834e-01 -1.07207052e-01 6.67341352e-01 -8.35278988e-01
4.89843190e-01 -9.20016468e-01 5.43005824e-01 4.20943409e-01
-2.70480990e-01 -3.83083463e-01 3.90489586e-02 5.37102997e-01
4.92020734e-02 -1.02547154e-01 8.28421831e-01 -4.43127692e-01
-7.23393261e-01 9.14392769e-01 1.63641155e-01 -1.21525832e-01
1.02990258e+00 -3.74887317e-01 -1.05113842e-01 1.38147533e-01
-1.06356108e+00 2.08305836e-01 8.38248014e-01 1.22404642e-01
7.59990811e-01 -1.30996716e+00 -6.04446352e-01 1.74969390e-01
-1.54411629e-01 4.78381276e-01 4.25876319e-01 9.02064323e-01
-6.00389004e-01 -5.45044476e-03 -1.43293813e-01 -9.66425538e-01
-1.00327861e+00 1.09442979e-01 5.25667846e-01 1.46473482e-01
-1.06730282e+00 1.00451851e+00 3.20922673e-01 -6.20536923e-01
9.75459144e-02 -2.35823542e-01 9.61134657e-02 -4.67157602e-01
4.46897335e-02 4.98864889e-01 5.38334213e-02 -8.25868011e-01
-4.86243784e-01 7.85584211e-01 1.51955977e-01 2.45423391e-02
1.60925293e+00 2.03959018e-01 -4.92781043e-01 5.02451658e-01
1.16173720e+00 -4.92197841e-01 -1.35975027e+00 -8.79010111e-02
-1.73636958e-01 -5.65978110e-01 2.68505782e-01 -6.76160991e-01
-1.54361951e+00 6.39806330e-01 6.85236692e-01 7.14245513e-02
5.45340002e-01 7.78386831e-01 1.08523118e+00 3.39426212e-02
5.64750552e-01 -8.73790741e-01 -1.96134493e-01 4.18467939e-01
6.37654483e-01 -1.01904190e+00 -1.49067730e-01 -8.94577324e-01
-3.24307024e-01 8.21644366e-01 7.63586104e-01 -7.07837343e-01
1.02079821e+00 3.17921266e-02 -1.35236606e-01 -8.14570069e-01
-1.46845862e-01 -1.78525746e-02 5.44743538e-01 4.58960503e-01
2.57142484e-01 3.88237476e-01 -8.81924480e-02 4.79612082e-01
-2.57253557e-01 -8.35130215e-02 -2.18046028e-02 5.08757710e-01
-5.33891022e-01 -8.75914574e-01 -5.72301745e-01 9.40944195e-01
-2.16582730e-01 2.67337002e-02 -2.28492022e-01 4.37887758e-01
3.16539317e-01 1.95596606e-01 6.28801167e-01 -3.23760241e-01
1.03725128e-01 1.66228693e-02 4.55671787e-01 -7.71092951e-01
-5.93829930e-01 -4.96991305e-03 -3.91078979e-01 -9.54900205e-01
-3.73264134e-01 -9.07650888e-01 -2.09029555e+00 -2.54557729e-01
1.19808018e-01 1.24591149e-01 8.77936244e-01 8.52167189e-01
5.38814783e-01 3.82627487e-01 4.09094870e-01 -1.43861783e+00
1.52969034e-02 -6.03543401e-01 -7.47813642e-01 5.31797826e-01
-2.09028333e-01 -6.67624831e-01 -3.20451289e-01 -1.85365364e-01] | [8.167673110961914, -3.256761312484741] |
8628b0a0-e50e-4c69-afc1-ef949be80591 | subgroup-discovery-in-unstructured-data | 2207.07781 | null | https://arxiv.org/abs/2207.07781v1 | https://arxiv.org/pdf/2207.07781v1.pdf | Subgroup Discovery in Unstructured Data | Subgroup discovery is a descriptive and exploratory data mining technique to identify subgroups in a population that exhibit interesting behavior with respect to a variable of interest. Subgroup discovery has numerous applications in knowledge discovery and hypothesis generation, yet it remains inapplicable for unstructured, high-dimensional data such as images. This is because subgroup discovery algorithms rely on defining descriptive rules based on (attribute, value) pairs, however, in unstructured data, an attribute is not well defined. Even in cases where the notion of attribute intuitively exists in the data, such as a pixel in an image, due to the high dimensionality of the data, these attributes are not informative enough to be used in a rule. In this paper, we introduce the subgroup-aware variational autoencoder, a novel variational autoencoder that learns a representation of unstructured data which leads to subgroups with higher quality. Our experimental results demonstrate the effectiveness of the method at learning subgroups with high quality while supporting the interpretability of the concepts. | ['Martin Ester', 'Jialin Lu', 'Dev Arora', 'Ali Arab'] | 2022-07-15 | null | null | null | null | ['subgroup-discovery'] | ['methodology'] | [ 2.22240537e-01 4.32464510e-01 -3.74886483e-01 -4.50938106e-01
-1.16828874e-01 -2.09064350e-01 3.93193394e-01 6.53264701e-01
-9.47439298e-02 8.51995945e-01 3.39493573e-01 -1.28199056e-01
-7.42198944e-01 -1.15382397e+00 -6.95582271e-01 -1.06151843e+00
-4.31792259e-01 8.97984147e-01 -6.97314367e-02 -3.80651057e-02
-9.62933153e-02 2.30456010e-01 -2.44545794e+00 6.36212289e-01
9.43196893e-01 1.10156918e+00 7.08761960e-02 8.55398327e-02
-3.00751656e-01 5.19647419e-01 -3.63672554e-01 -2.15242565e-01
8.07437971e-02 -4.34949547e-01 -6.97132528e-01 4.16619390e-01
1.37390912e-01 -5.61243519e-02 2.43600890e-01 1.07135630e+00
1.01971485e-01 5.11088014e-01 9.83161211e-01 -1.25687242e+00
-4.71976727e-01 8.29317749e-01 -3.07942897e-01 -1.09669805e-01
2.36453176e-01 -3.12456787e-01 1.49374866e+00 -6.39701486e-01
8.82037878e-01 1.11995268e+00 3.34670126e-01 5.17347932e-01
-1.39726043e+00 -5.13446748e-01 3.28361779e-01 2.93900669e-01
-1.42723095e+00 -9.91783887e-02 9.28261280e-01 -5.51101029e-01
4.26534742e-01 3.25290203e-01 8.76429975e-01 1.19245851e+00
2.44421768e-03 7.36420393e-01 9.21735048e-01 -2.28187576e-01
7.27662444e-01 4.99962479e-01 3.11748654e-01 6.13474011e-01
4.74478781e-01 9.82170701e-02 -6.14750564e-01 -4.32140440e-01
2.81621456e-01 3.56300205e-01 -1.37447700e-01 -5.75788736e-01
-1.07338500e+00 1.28703475e+00 2.39619642e-01 3.92353803e-01
-8.67185771e-01 -2.57954806e-01 2.58813500e-01 3.51112783e-01
4.12154585e-01 4.25414115e-01 -3.42608392e-01 7.67135322e-02
-9.76897120e-01 2.79937565e-01 7.89686203e-01 5.75026512e-01
1.02774405e+00 -2.04783797e-01 -1.44531861e-01 5.03116369e-01
3.11016411e-01 4.70778719e-02 6.26963913e-01 -9.14984822e-01
-4.82459627e-02 1.12331700e+00 -1.77441597e-01 -9.88028109e-01
-4.15783882e-01 -5.16740382e-01 -1.07953870e+00 4.86328527e-02
2.89853793e-02 6.26328290e-02 -6.68223977e-01 1.71169317e+00
5.80258250e-01 7.16267675e-02 2.69773990e-01 7.41860032e-01
8.93311143e-01 4.95295256e-01 5.87698556e-02 -5.96278369e-01
1.38285995e+00 -1.05466254e-01 -7.84108281e-01 2.92508543e-01
4.25905347e-01 6.61310330e-02 8.88541222e-01 4.57773417e-01
-4.82768267e-01 -5.21252334e-01 -8.95543516e-01 2.91853994e-01
-4.43199247e-01 -3.13825577e-01 9.13844705e-01 4.07409519e-01
-6.02015555e-01 6.87915802e-01 -6.23757839e-01 -3.26671302e-01
6.20534956e-01 6.54099345e-01 -4.80682909e-01 1.58526525e-01
-1.21150768e+00 1.31068707e-01 5.55516541e-01 -2.99421698e-01
-7.06329525e-01 -8.24000776e-01 -8.09209704e-01 2.34845713e-01
6.44093931e-01 -8.04032862e-01 6.99852824e-01 -1.09976292e+00
-8.26211870e-01 7.15972304e-01 -3.02981168e-01 -7.53574729e-01
3.37094694e-01 2.37284973e-01 -5.33389032e-01 6.28760010e-02
2.83164203e-01 4.33237344e-01 1.00785887e+00 -1.54347205e+00
-9.06808019e-01 -7.32680142e-01 -6.98395967e-02 -7.68469423e-02
-4.96023357e-01 -5.80903172e-01 -1.23631908e-02 -3.57310832e-01
2.09435463e-01 -7.46047616e-01 -4.57877368e-01 -2.87606478e-01
-6.21207595e-01 -6.62529051e-01 1.00756347e+00 -1.87116250e-01
1.20836890e+00 -2.26747799e+00 1.37940407e-01 7.75203466e-01
7.09575832e-01 -2.81733543e-01 6.06702685e-01 2.02270821e-01
1.47320315e-01 2.87056535e-01 -5.09624422e-01 -7.83877373e-02
-2.00748052e-02 5.82446158e-01 -2.25619361e-01 3.08958620e-01
1.71102174e-02 5.15899301e-01 -7.65979767e-01 -7.76131928e-01
3.06676608e-03 2.17076018e-01 -8.06249857e-01 1.85745001e-01
-5.53863525e-01 3.45712662e-01 -9.17687118e-01 6.32323384e-01
3.33681911e-01 -3.86070281e-01 2.13126212e-01 -1.88818708e-01
-5.02334163e-02 -2.68359333e-01 -1.44289541e+00 1.06719089e+00
-4.26026061e-02 4.46426392e-01 -2.20036402e-01 -1.37155151e+00
9.39634204e-01 2.75816947e-01 1.05202043e+00 -2.20793784e-01
-3.94280367e-02 3.76356058e-02 7.47296363e-02 -5.87789416e-01
4.99110430e-01 -3.71797442e-01 4.18934412e-02 5.17163396e-01
-1.26369208e-01 4.39402461e-01 4.41276699e-01 -3.68729457e-02
8.78241301e-01 -3.60569209e-01 3.34931761e-01 -4.38241333e-01
4.74346519e-01 1.39148220e-01 1.05099130e+00 7.79500365e-01
2.01667324e-01 4.04440671e-01 6.86879277e-01 -6.15193963e-01
-9.22201097e-01 -1.08511484e+00 -4.64290708e-01 7.24411726e-01
-5.62600568e-02 -5.24716914e-01 -5.08922517e-01 -5.82507312e-01
1.67652115e-01 6.60705268e-01 -1.12402928e+00 -1.81093529e-01
-2.01652441e-02 -7.78262138e-01 -3.75852212e-02 2.45089144e-01
5.10459542e-01 -1.06150746e+00 -7.39272833e-01 1.88351050e-01
-9.69737843e-02 -8.93494427e-01 5.32973222e-02 2.98243016e-01
-1.02300155e+00 -1.36844718e+00 -6.63658455e-02 -7.47347474e-01
8.22904766e-01 -2.16450572e-01 1.22847044e+00 7.55570680e-02
-1.31298915e-01 2.53818363e-01 -4.97279525e-01 -5.60254872e-01
-4.28755254e-01 -1.15839534e-01 3.29164118e-01 6.04175329e-01
6.88758552e-01 -6.65954053e-01 -5.10294855e-01 2.43717153e-03
-9.93488729e-01 -2.13581592e-01 5.40670037e-01 9.52335358e-01
1.13924694e+00 6.81647301e-01 5.04717767e-01 -1.27387035e+00
6.48473680e-01 -7.43306816e-01 -3.69520932e-01 1.03550032e-01
-8.72761905e-01 4.33043361e-01 6.79199815e-01 -4.94414926e-01
-8.29883218e-01 1.28916189e-01 2.11975217e-01 -4.20587271e-01
-4.09720719e-01 7.29496658e-01 -2.78363943e-01 5.92333257e-01
6.21525288e-01 3.13445747e-01 3.05307657e-01 -4.96599704e-01
8.64332765e-02 5.78389764e-01 1.36847898e-01 -5.62985778e-01
4.50819790e-01 7.71280766e-01 2.80380428e-01 -1.12746358e+00
-6.88320518e-01 -4.41290766e-01 -6.05387092e-01 5.31203719e-03
8.32629204e-01 -6.66528940e-01 -1.01798785e+00 -3.64956468e-01
-4.72339153e-01 2.36683607e-01 -7.82002926e-01 6.22232020e-01
-7.21541047e-01 1.97654769e-01 1.41448811e-01 -7.69126773e-01
-2.01718137e-01 -1.21118391e+00 7.75516868e-01 7.83999115e-02
-4.76004601e-01 -8.74314964e-01 -6.63908869e-02 2.20240951e-01
-2.53154814e-01 5.58679402e-01 1.48766255e+00 -1.20706594e+00
-5.43193698e-01 1.84296034e-02 3.72839749e-01 5.69002517e-02
3.85660768e-01 -2.89197080e-03 -7.54321158e-01 -5.28308749e-02
-1.74144045e-01 -3.31356302e-02 8.40091705e-01 6.52114511e-01
1.59978366e+00 -5.75184941e-01 -4.84396994e-01 4.98685449e-01
1.10803056e+00 2.92416573e-01 1.80307031e-01 1.12560473e-01
5.37298203e-01 1.14207160e+00 4.27137822e-01 7.51311541e-01
3.20417523e-01 5.07152140e-01 3.71501863e-01 8.99000391e-02
4.54122484e-01 -1.90430537e-01 3.87551589e-03 1.45597383e-01
-2.19968528e-01 -9.25092474e-02 -6.23798609e-01 6.21320426e-01
-1.91356492e+00 -1.31142533e+00 -8.45109001e-02 2.22324204e+00
7.89059222e-01 -1.06795646e-01 2.75551051e-01 4.53930140e-01
6.12990618e-01 -7.57469609e-02 -6.86618149e-01 -3.64769727e-01
-3.30097347e-01 -7.31027648e-02 -3.18225585e-02 2.43928865e-01
-9.09521639e-01 5.39554894e-01 5.40366936e+00 4.24364984e-01
-7.84906626e-01 -3.04111451e-01 6.77042544e-01 1.72135442e-01
-8.78354907e-01 -1.48725614e-01 -6.52137339e-01 5.40297389e-01
7.25774765e-01 -2.58512467e-01 2.08652124e-01 8.66273046e-01
2.66754180e-01 4.02146317e-02 -1.42304492e+00 9.25411224e-01
-1.57913536e-01 -1.51993406e+00 4.63062137e-01 6.28263295e-01
7.38323212e-01 -5.45117617e-01 3.24394137e-01 1.41250461e-01
2.23087698e-01 -1.16435695e+00 3.08523715e-01 4.96261299e-01
2.32538328e-01 -1.12666249e+00 5.90187192e-01 4.29343820e-01
-6.99597180e-01 -3.52003962e-01 -4.23311949e-01 1.72931239e-01
-2.78054982e-01 1.01833689e+00 -1.06691468e+00 2.62041390e-01
9.06917036e-01 7.54967988e-01 -3.64775658e-01 5.41935980e-01
1.36889964e-01 5.59056103e-01 -3.00681055e-01 -1.39390081e-01
8.68131742e-02 -3.99709910e-01 6.40727460e-01 5.34319818e-01
3.47014129e-01 3.08365226e-01 3.11559647e-01 1.03608263e+00
-6.25464618e-02 2.55853087e-01 -9.60079312e-01 -1.52483463e-01
3.47905427e-01 8.81211042e-01 -6.10400021e-01 -1.90912664e-01
-2.79202104e-01 7.08705604e-01 -3.77546698e-02 3.84886175e-01
-2.70167351e-01 1.01842597e-01 1.05995119e+00 4.28893745e-01
4.77941692e-01 1.50571674e-01 -1.94358096e-01 -9.49821293e-01
9.04645678e-03 -9.67249036e-01 8.68040860e-01 -1.29578263e-01
-1.21853173e+00 5.29650450e-01 1.83929443e-01 -1.34706128e+00
-8.42022121e-01 -5.03445089e-01 -2.74413526e-01 3.16128582e-01
-9.09123778e-01 -9.36904371e-01 -1.83167592e-01 9.75741148e-01
3.81493509e-01 -5.72834909e-01 9.17642534e-01 -1.50271952e-01
-2.98335999e-01 5.02048314e-01 3.00102204e-01 3.31241302e-02
2.03710526e-01 -1.46093500e+00 -4.37061667e-01 4.65208262e-01
4.99059796e-01 7.56493747e-01 1.08168244e+00 -6.68913245e-01
-1.41160083e+00 -1.13082826e+00 8.12360764e-01 -2.00320527e-01
3.14007044e-01 -3.10916692e-01 -9.52930272e-01 5.14465213e-01
-1.71275109e-01 9.37708393e-02 1.31948245e+00 5.68042219e-01
-1.30046725e-01 -2.33793810e-01 -1.28910458e+00 6.18315160e-01
8.48305643e-01 -2.89102703e-01 -5.64967513e-01 1.02984548e-01
7.67865002e-01 2.55230963e-01 -1.17728114e+00 3.53480339e-01
4.44672316e-01 -1.09771609e+00 9.17268336e-01 -9.57140625e-01
4.28418010e-01 -3.87113780e-01 -3.86742204e-01 -1.30904114e+00
-3.18735898e-01 -4.18295920e-01 -5.48793912e-01 1.12268591e+00
5.04949093e-01 -3.42292041e-01 1.13108075e+00 5.75756669e-01
1.91608682e-01 -9.32759404e-01 -1.03113508e+00 -4.98686284e-01
-2.61184514e-01 -3.92114013e-01 1.09880018e+00 9.53156590e-01
-1.91795349e-01 3.23012263e-01 -1.74611241e-01 2.37461805e-01
1.04321575e+00 7.56531596e-01 5.03543139e-01 -2.00403762e+00
-2.01834425e-01 -2.31444865e-01 -7.15960205e-01 -2.77807087e-01
4.21059668e-01 -8.91793907e-01 -2.07391083e-01 -1.36782181e+00
1.92724437e-01 -5.05076766e-01 -2.60949850e-01 2.67429918e-01
8.33461806e-02 -1.48510963e-01 -2.40572318e-01 1.09294821e-02
-4.32102561e-01 7.12919295e-01 9.80244279e-01 -4.92803514e-01
-5.83907247e-01 3.19923550e-01 -1.01856625e+00 6.16950572e-01
7.24698544e-01 -4.38805968e-01 -7.53876090e-01 3.32778692e-01
1.12912640e-01 -1.88297432e-04 2.79497951e-01 -7.80445337e-01
2.57716421e-02 -3.71846884e-01 4.69099164e-01 -6.06367230e-01
1.89220309e-01 -1.08513260e+00 4.21095520e-01 4.41556901e-01
-3.66963953e-01 -1.84045091e-01 -3.01585674e-01 7.94451118e-01
-4.14015412e-01 -6.52842373e-02 4.49654520e-01 -1.27363488e-01
-1.10952842e+00 6.44825399e-01 -3.32777768e-01 2.88918527e-04
1.06612086e+00 -5.25572002e-01 1.52460262e-01 -3.96479934e-01
-9.34645891e-01 2.40817547e-01 3.51509213e-01 3.99623573e-01
9.22971129e-01 -1.35061443e+00 -8.13291371e-01 4.75621492e-01
5.81944287e-01 2.78381854e-01 3.39232653e-01 6.31026924e-01
-5.94381383e-03 5.55352345e-02 -1.18078202e-01 -7.60576427e-01
-1.11576140e+00 9.37157035e-01 7.84275830e-02 1.74325947e-02
-9.82001901e-01 4.78466332e-01 4.03407604e-01 -4.38687280e-02
2.44208634e-01 -3.16716194e-01 -7.10301340e-01 6.01878345e-01
5.42789817e-01 1.25284657e-01 -2.26328559e-02 -6.79990590e-01
-3.52766782e-01 3.29654545e-01 6.26611942e-03 7.48938844e-02
1.54789853e+00 3.37264761e-02 -1.53341264e-01 7.36520708e-01
1.12242270e+00 -1.24134898e-01 -9.33300674e-01 -1.77654594e-01
7.90689215e-02 -2.69428074e-01 1.44431740e-01 -2.54376918e-01
-1.01866961e+00 5.41457772e-01 5.84812045e-01 3.77875090e-01
1.12981462e+00 3.75921428e-01 3.34336489e-01 7.35574961e-01
1.16332948e-01 -1.26186478e+00 -1.44817606e-01 6.94973245e-02
6.46255970e-01 -1.44756722e+00 -7.54285455e-02 -2.45248199e-01
-8.28185260e-01 1.01312327e+00 3.53353679e-01 1.82758808e-01
9.91948307e-01 6.15444630e-02 -2.14476526e-01 -4.56234992e-01
-8.86721134e-01 -3.72973531e-01 4.41045821e-01 9.60151017e-01
2.02426642e-01 2.63390601e-01 -1.58741161e-01 8.22956383e-01
-4.05930907e-01 -3.48418176e-01 3.56875509e-01 6.17013931e-01
-4.63344246e-01 -9.77105618e-01 -3.06343347e-01 1.09898043e+00
-3.47830057e-01 1.67479992e-01 -3.60134751e-01 6.44577861e-01
4.31421280e-01 9.37811077e-01 4.75613177e-01 -5.20729661e-01
2.03430653e-01 -3.76408249e-02 5.65083213e-02 -6.04190826e-01
-2.80834049e-01 -1.20634146e-01 -1.17596291e-01 -5.11780858e-01
-4.96312678e-01 -1.01864457e+00 -1.37282872e+00 -1.67994678e-01
2.64655888e-01 6.92177236e-01 2.95910984e-01 1.05523562e+00
3.67146730e-01 4.14895386e-01 7.23316491e-01 -3.70170623e-02
-3.57985824e-01 -5.33751905e-01 -1.03380251e+00 7.37691641e-01
7.43321240e-01 -9.13499236e-01 -2.48849943e-01 2.37397596e-01] | [7.849210262298584, 4.777099609375] |
56992c95-0743-4e69-ac66-a544e8a11ee8 | mac-po-multi-agent-experience-replay-via | 2302.10418 | null | https://arxiv.org/abs/2302.10418v2 | https://arxiv.org/pdf/2302.10418v2.pdf | MAC-PO: Multi-Agent Experience Replay via Collective Priority Optimization | Experience replay is crucial for off-policy reinforcement learning (RL) methods. By remembering and reusing the experiences from past different policies, experience replay significantly improves the training efficiency and stability of RL algorithms. Many decision-making problems in practice naturally involve multiple agents and require multi-agent reinforcement learning (MARL) under centralized training decentralized execution paradigm. Nevertheless, existing MARL algorithms often adopt standard experience replay where the transitions are uniformly sampled regardless of their importance. Finding prioritized sampling weights that are optimized for MARL experience replay has yet to be explored. To this end, we propose MAC-PO, which formulates optimal prioritized experience replay for multi-agent problems as a regret minimization over the sampling weights of transitions. Such optimization is relaxed and solved using the Lagrangian multiplier approach to obtain the close-form optimal sampling weights. By minimizing the resulting policy regret, we can narrow the gap between the current policy and a nominal optimal policy, thus acquiring an improved prioritization scheme for multi-agent tasks. Our experimental results on Predator-Prey and StarCraft Multi-Agent Challenge environments demonstrate the effectiveness of our method, having a better ability to replay important transitions and outperforming other state-of-the-art baselines. | ['Peng Wei', 'Guru Venkataramani', 'Tian Lan', 'Hanhan Zhou', 'Yongsheng Mei'] | 2023-02-21 | null | null | null | null | ['starcraft'] | ['playing-games'] | [-2.33047366e-01 -1.66954219e-01 -6.94143534e-01 1.10610142e-01
-8.14679921e-01 -4.83544946e-01 6.00968778e-01 3.02670926e-01
-1.01708555e+00 1.25357842e+00 1.93196297e-01 -2.89575934e-01
-2.43639767e-01 -5.44849098e-01 -7.56852090e-01 -8.92340660e-01
-5.78946173e-01 5.89239478e-01 4.34317254e-02 -4.05441761e-01
2.80466199e-01 1.72344923e-01 -1.46322072e+00 -7.17032179e-02
9.80357587e-01 7.51370549e-01 3.77436370e-01 7.32063711e-01
2.79769540e-01 1.11332798e+00 -6.41340196e-01 1.38296202e-01
3.57990563e-01 -5.26149631e-01 -6.78658843e-01 1.28087088e-01
-1.63722098e-01 -7.10105300e-01 -1.61578983e-01 8.86381507e-01
6.79227889e-01 8.51829648e-01 2.12063268e-01 -1.39886260e+00
-2.63098538e-01 8.73970509e-01 -7.57899821e-01 3.43287706e-01
3.39922458e-01 4.30620641e-01 9.21801627e-01 -2.37154320e-01
4.77590501e-01 1.34648442e+00 2.70551354e-01 7.07673192e-01
-1.02815735e+00 -2.75355518e-01 6.52555585e-01 3.62573713e-01
-7.39736676e-01 -1.49303436e-01 3.65049928e-01 9.55131724e-02
1.21329820e+00 -1.13623112e-01 9.82309222e-01 1.19143438e+00
3.65287423e-01 1.32912481e+00 1.21994770e+00 -3.38739663e-01
9.06858265e-01 -1.82644248e-01 -4.18156981e-01 7.86929548e-01
2.82609493e-01 3.53104293e-01 -5.99811673e-01 -3.11525851e-01
6.51765227e-01 2.26426676e-01 -1.31857589e-01 -5.52296638e-01
-1.21498919e+00 7.90351510e-01 2.65089989e-01 -1.74434319e-01
-8.90217304e-01 4.25385445e-01 5.75222373e-01 8.01819086e-01
1.51389420e-01 6.18075013e-01 -6.68470204e-01 -4.12858635e-01
-6.27677977e-01 6.78100049e-01 7.66605616e-01 6.08335972e-01
7.81465471e-01 2.70217240e-01 -2.98700988e-01 6.37065649e-01
1.31790757e-01 6.06703937e-01 6.92915261e-01 -1.25982285e+00
4.42844927e-01 2.30949089e-01 6.63195252e-01 -1.90395579e-01
-2.41044804e-01 -3.70721072e-01 -4.25351828e-01 6.40885711e-01
3.46207917e-01 -8.64394903e-01 -5.31950176e-01 1.77518511e+00
5.63064516e-01 3.96089524e-01 5.53232193e-01 9.93282497e-01
-1.83965296e-01 6.99514508e-01 7.47412145e-02 -5.23689747e-01
1.09380162e+00 -1.18788838e+00 -5.11728644e-01 -4.90571111e-01
4.40372825e-01 -2.83487588e-01 1.03368509e+00 4.89711434e-01
-1.05071652e+00 -9.72789451e-02 -1.21587157e+00 5.78106642e-01
1.57184392e-01 -5.72413206e-02 6.96366251e-01 2.16765091e-01
-7.07912028e-01 9.56400752e-01 -1.11754894e+00 -5.95105141e-02
2.83802897e-01 2.99784124e-01 8.99937749e-02 1.24287240e-01
-9.39656615e-01 1.01292396e+00 5.10343969e-01 -1.71484664e-01
-1.46671605e+00 -7.08734572e-01 -6.99242830e-01 -2.97098979e-02
1.10944092e+00 -6.82987034e-01 2.01826715e+00 -1.01745546e+00
-2.21460438e+00 1.65700130e-02 2.73452997e-01 -8.54987919e-01
5.57842970e-01 -5.10353625e-01 -1.46929339e-01 2.26956367e-01
1.07306622e-01 4.23091620e-01 9.22535002e-01 -1.19913757e+00
-1.15408874e+00 1.39054298e-01 4.60991770e-01 8.97129953e-01
-1.07297748e-01 -5.38945973e-01 3.51890139e-02 -2.24188909e-01
-6.77166283e-01 -1.10605764e+00 -7.04740942e-01 -4.29332018e-01
3.23935360e-01 -2.46911511e-01 6.16956592e-01 -1.09174125e-01
9.02337492e-01 -1.97584307e+00 3.22798193e-01 -5.38413227e-03
-1.34176075e-01 2.83233494e-01 -4.51124877e-01 6.54950976e-01
6.80496931e-01 -5.61839819e-01 7.73505718e-02 -3.53051752e-01
1.81876957e-01 5.59534848e-01 -4.21111584e-01 4.99074996e-01
-1.32691026e-01 6.85158372e-01 -1.63547444e+00 -2.14507312e-01
3.04935992e-01 -1.01103351e-01 -7.98199058e-01 4.30006146e-01
-7.46909738e-01 4.26430106e-01 -6.59013331e-01 4.85060543e-01
2.52671540e-01 -9.09814015e-02 6.02622032e-01 5.50058305e-01
-2.32426345e-01 2.41973117e-01 -1.23579860e+00 1.75505602e+00
-7.64554620e-01 1.89629510e-01 6.57720789e-02 -7.65241086e-01
5.49442589e-01 2.52923220e-01 7.85148799e-01 -9.30496514e-01
6.80496218e-03 6.40517101e-02 -5.95582277e-02 -3.52275193e-01
7.43107498e-01 2.14627199e-02 -2.04904631e-01 7.42217064e-01
-2.00612321e-02 -6.81069307e-03 4.41960990e-01 1.60294726e-01
1.06793296e+00 5.57575166e-01 7.02763796e-01 -8.56191888e-02
1.32873625e-01 1.32612854e-01 8.23686302e-01 1.24531496e+00
-4.05606449e-01 -3.53057534e-01 3.92278671e-01 -3.10723752e-01
-7.67373979e-01 -9.09048736e-01 5.65622807e-01 1.40125716e+00
3.48455220e-01 -1.94593683e-01 -4.76778001e-01 -7.68530905e-01
1.40551239e-01 7.97963381e-01 -5.36736608e-01 -1.72398537e-01
-6.51517332e-01 -6.59974337e-01 1.76059231e-01 2.73341596e-01
6.86067045e-01 -1.59594798e+00 -1.62084305e+00 6.43265605e-01
-8.48518610e-02 -7.89868176e-01 -6.44736767e-01 3.38124573e-01
-7.21958399e-01 -9.74709511e-01 -7.46101856e-01 -4.69262898e-01
3.67347747e-01 4.38575596e-01 9.91948426e-01 2.16244329e-02
7.03855753e-02 7.10917950e-01 -4.76036727e-01 -4.46693569e-01
-2.66965568e-01 5.79920597e-02 4.52430576e-01 -2.82215178e-01
-3.52385826e-02 -6.43972814e-01 -8.54804456e-01 1.70746110e-02
-7.98613787e-01 -1.02087863e-01 9.07275558e-01 1.06218803e+00
5.78002036e-01 4.49186936e-02 8.71738970e-01 -6.35726213e-01
1.06412315e+00 -6.72434628e-01 -7.61328757e-01 3.28300506e-01
-6.83977306e-01 5.46824574e-01 9.51015413e-01 -9.19270813e-01
-1.22165346e+00 -1.99909329e-01 3.39280874e-01 -2.54049987e-01
5.00874072e-02 4.95565563e-01 2.94394761e-01 2.08285689e-01
6.75791740e-01 3.61258090e-01 1.19807154e-01 -2.85925413e-03
5.11534572e-01 3.25770319e-01 2.67634749e-01 -1.07369328e+00
2.85588115e-01 2.34773040e-01 -1.12794340e-01 -5.58883846e-01
-8.46536875e-01 -4.58871305e-01 1.86086759e-01 -4.09413397e-01
2.75384486e-01 -9.86757636e-01 -1.38713515e+00 3.10596645e-01
-5.57924449e-01 -1.07970738e+00 -5.52307487e-01 6.03595734e-01
-1.00507903e+00 3.95202428e-01 -6.39858365e-01 -1.12473786e+00
-4.19320166e-01 -1.12307632e+00 6.43894494e-01 5.33974051e-01
4.27468307e-02 -9.03981626e-01 5.94618142e-01 -2.24095702e-01
3.24989647e-01 1.97978601e-01 5.47175348e-01 -2.10559323e-01
-4.06812936e-01 3.56510580e-01 3.99171770e-01 -1.00482158e-01
-9.65849310e-02 -4.94321913e-01 -4.30162668e-01 -8.82506132e-01
-2.00895816e-01 -8.27949464e-01 7.70467639e-01 3.96034271e-01
5.77043831e-01 -6.74824774e-01 -1.38582125e-01 1.63850874e-01
1.53800845e+00 3.46852064e-01 2.61221856e-01 8.43826592e-01
5.70607334e-02 1.04563177e-01 1.20180881e+00 1.23790395e+00
4.98691171e-01 4.27528530e-01 6.08445823e-01 4.43129420e-01
3.21814746e-01 -4.53303337e-01 1.02512217e+00 4.82792497e-01
-3.33514065e-02 -9.42383856e-02 -4.52352703e-01 5.77998877e-01
-2.55709457e+00 -1.22012746e+00 7.22170889e-01 2.30484366e+00
1.02674043e+00 1.77346036e-01 5.28540790e-01 -3.59280497e-01
2.54068792e-01 1.88594475e-01 -1.22955596e+00 -5.25996149e-01
2.88197458e-01 4.63503934e-02 6.56980872e-01 5.90426624e-01
-9.05509651e-01 1.04641104e+00 6.08603191e+00 7.54822254e-01
-8.84582877e-01 1.19700963e-02 7.06672817e-02 -5.42953789e-01
-1.63020450e-03 1.28033251e-01 -7.30875254e-01 3.66638631e-01
8.79777372e-01 -4.40266788e-01 1.27931762e+00 1.06750178e+00
4.04533982e-01 -5.81939578e-01 -1.03059959e+00 7.57146120e-01
-3.19905519e-01 -1.13523102e+00 -1.61838129e-01 -6.16453700e-02
9.81336057e-01 2.45857835e-01 5.47245778e-02 1.01062620e+00
1.15150249e+00 -4.48703259e-01 8.94142747e-01 3.09888333e-01
3.04083854e-01 -1.07535231e+00 3.97001356e-01 6.15401685e-01
-1.09575284e+00 -4.66577709e-01 -4.44180757e-01 -5.12515604e-01
1.18730299e-01 1.32111162e-01 -8.90218019e-01 5.51453888e-01
4.19192821e-01 5.74813187e-01 3.22844498e-02 1.07608104e+00
-3.97567630e-01 3.91279131e-01 -2.33819380e-01 -6.12038970e-01
7.68920541e-01 -2.86417425e-01 6.50998592e-01 8.25659335e-01
1.76569372e-01 -1.21847382e-02 8.11828673e-01 2.76234478e-01
3.65015194e-02 -6.66792318e-02 -4.10609126e-01 3.94461788e-02
6.76951289e-01 1.18348444e+00 -5.97589433e-01 -3.56328815e-01
-2.05851555e-01 1.01046216e+00 7.29560435e-01 4.76424813e-01
-8.97510171e-01 -1.95660904e-01 1.02972686e+00 -6.18391395e-01
4.85178709e-01 -4.06950027e-01 3.88802320e-01 -1.15398061e+00
-1.99435025e-01 -1.21277046e+00 5.01978397e-01 -2.24531218e-01
-1.08945847e+00 3.39170009e-01 4.65591289e-02 -1.34714043e+00
-6.63254440e-01 -2.03476548e-01 -6.40952349e-01 1.26543999e-01
-1.75179172e+00 -5.13793945e-01 2.99443603e-01 5.71084857e-01
8.63185942e-01 -3.54820579e-01 7.19421029e-01 -9.09137204e-02
-5.60743272e-01 4.71042097e-01 4.50065672e-01 -4.64388490e-01
6.02348566e-01 -1.46467805e+00 3.77612263e-02 7.35811830e-01
-1.47826895e-01 2.56727517e-01 9.12168503e-01 -5.83786607e-01
-1.72174537e+00 -9.41928208e-01 -6.10569678e-02 1.56011835e-01
6.98143840e-01 1.36764005e-01 -3.89498830e-01 6.10311449e-01
4.82618839e-01 -2.94571161e-01 5.09154797e-01 1.13649391e-01
8.95569026e-02 -4.07045707e-02 -1.00479400e+00 1.16793406e+00
6.84704959e-01 -8.79781097e-02 -5.21264672e-01 1.63727373e-01
5.75384498e-01 -5.42903543e-01 -6.56309247e-01 -1.47645161e-01
4.38254416e-01 -6.73669517e-01 7.80162632e-01 -7.18672693e-01
1.91715926e-01 -2.43920460e-01 6.41892850e-02 -2.06077576e+00
-1.02835737e-01 -1.10815692e+00 -6.18020773e-01 6.32199824e-01
5.08638211e-02 -5.37809730e-01 6.89120829e-01 2.44517699e-01
-8.29825550e-02 -8.43168855e-01 -8.04440856e-01 -9.67987239e-01
-2.00177431e-01 3.98503877e-02 5.31075001e-01 5.32460093e-01
4.05860633e-01 3.24338645e-01 -8.17418396e-01 1.68285325e-01
9.57826793e-01 2.62457818e-01 7.85375595e-01 -5.87252259e-01
-8.07772160e-01 -4.26899910e-01 4.66835231e-01 -1.32860827e+00
3.66613626e-01 -5.67521989e-01 3.82564247e-01 -1.41460514e+00
4.29301038e-02 -4.55423355e-01 -4.17892814e-01 5.55789530e-01
-3.12803060e-01 -5.14574170e-01 4.42211568e-01 4.93184887e-02
-1.44514704e+00 1.07098019e+00 1.48093116e+00 -5.72439581e-02
-7.00549960e-01 7.57666528e-02 -5.30278802e-01 6.35317028e-01
1.12941015e+00 -5.49417615e-01 -6.08496308e-01 -2.70440698e-01
2.09817484e-01 6.11804008e-01 -2.55177449e-02 -1.01477551e+00
2.42632419e-01 -9.28343296e-01 2.51198150e-02 -2.71668136e-01
2.89540410e-01 -6.21925950e-01 -2.58184314e-01 8.46645296e-01
-5.03385961e-01 3.25420529e-01 1.58267856e-01 1.01276791e+00
1.90858871e-01 -4.06264961e-01 5.89163542e-01 -4.33970004e-01
-1.10790098e+00 4.08661872e-01 -9.31911170e-01 2.64140040e-01
1.27882838e+00 3.21656674e-01 -3.09520125e-01 -3.80822152e-01
-4.48245734e-01 9.24114883e-01 2.71382183e-01 2.35117525e-01
6.34103477e-01 -1.18942070e+00 -4.98183131e-01 -2.09861487e-01
-4.38064970e-02 -2.20495269e-01 2.09227026e-01 4.91561592e-01
-2.21944198e-01 -2.22688019e-02 -6.54976368e-01 -9.74583104e-02
-9.32425976e-01 5.37816465e-01 3.66136402e-01 -7.51763105e-01
-6.72122180e-01 2.92794079e-01 -5.93236208e-01 -2.80215055e-01
3.94385964e-01 -2.52137154e-01 -7.28823543e-02 -1.08878307e-01
6.12075984e-01 7.12551773e-01 -3.36500376e-01 2.56355733e-01
-2.00548202e-01 -1.32641226e-01 -3.65444988e-01 -6.01802528e-01
1.43135214e+00 -1.86580196e-01 3.60662222e-01 3.93477499e-01
5.15698195e-01 -2.69738555e-01 -2.10715318e+00 -4.37723249e-01
-1.70020647e-02 -4.09854025e-01 1.21974558e-01 -8.40856314e-01
-6.32598102e-01 2.60264933e-01 3.28837276e-01 7.38736317e-02
8.90471458e-01 -4.51531380e-01 8.57611954e-01 9.73508596e-01
9.03141618e-01 -1.53197563e+00 5.10156095e-01 6.73084974e-01
5.70322990e-01 -1.11382890e+00 9.22718197e-02 5.44629872e-01
-1.26127315e+00 9.08125818e-01 8.26797187e-01 -4.58857477e-01
1.21036984e-01 1.66362509e-01 1.32459784e-02 2.21312419e-01
-1.23697567e+00 -4.53444660e-01 -4.03752804e-01 4.79015619e-01
-2.43084282e-01 3.87081921e-01 -2.56025821e-01 2.20944881e-01
1.75013185e-01 5.39462417e-02 6.25220835e-01 1.30266261e+00
-6.85871601e-01 -1.25102770e+00 -2.12735742e-01 3.11869055e-01
-3.25845957e-01 2.63159722e-01 2.55153537e-01 5.19146442e-01
-3.01367164e-01 1.01641941e+00 -6.57025650e-02 3.92394289e-02
2.13946626e-01 -3.22705299e-01 8.29169929e-01 -3.83128673e-01
-7.35357821e-01 -4.28757481e-02 1.08772214e-03 -6.18997872e-01
-3.97168487e-01 -7.69518137e-01 -1.58588409e+00 -4.85834330e-01
-6.62384629e-02 4.36523110e-01 2.64532626e-01 1.01949334e+00
3.19784075e-01 6.51137292e-01 7.36577809e-01 -1.04586422e+00
-1.37470710e+00 -6.78169549e-01 -5.82439005e-01 2.60708719e-01
6.49588168e-01 -8.59079778e-01 -2.60417052e-02 -5.52414358e-01] | [3.902498483657837, 2.0250329971313477] |
18aeb6b4-ff44-4d2d-905a-ec491f2893e3 | real-time-super-resolution-system-of-4k-video | 2107.05307 | null | https://arxiv.org/abs/2107.05307v2 | https://arxiv.org/pdf/2107.05307v2.pdf | Real-Time Super-Resolution System of 4K-Video Based on Deep Learning | Video super-resolution (VSR) technology excels in reconstructing low-quality video, avoiding unpleasant blur effect caused by interpolation-based algorithms. However, vast computation complexity and memory occupation hampers the edge of deplorability and the runtime inference in real-life applications, especially for large-scale VSR task. This paper explores the possibility of real-time VSR system and designs an efficient and generic VSR network, termed EGVSR. The proposed EGVSR is based on spatio-temporal adversarial learning for temporal coherence. In order to pursue faster VSR processing ability up to 4K resolution, this paper tries to choose lightweight network structure and efficient upsampling method to reduce the computation required by EGVSR network under the guarantee of high visual quality. Besides, we implement the batch normalization computation fusion, convolutional acceleration algorithm and other neural network acceleration techniques on the actual hardware platform to optimize the inference process of EGVSR network. Finally, our EGVSR achieves the real-time processing capacity of 4K@29.61FPS. Compared with TecoGAN, the most advanced VSR network at present, we achieve 85.04% reduction of computation density and 7.92x performance speedups. In terms of visual quality, the proposed EGVSR tops the list of most metrics (such as LPIPS, tOF, tLP, etc.) on the public test dataset Vid4 and surpasses other state-of-the-art methods in overall performance score. The source code of this project can be found on https://github.com/Thmen/EGVSR. | ['He Li', 'Yongming Tang', 'Changjun Song', 'Chengcheng Wang', 'Yanpeng Cao'] | 2021-07-12 | null | null | null | null | ['video-super-resolution'] | ['computer-vision'] | [ 5.36593199e-02 -4.32257235e-01 -5.35107777e-02 -1.19337045e-01
-5.73786616e-01 -3.89542818e-01 2.10385069e-01 -5.68498611e-01
-4.63705391e-01 5.74214160e-01 -6.90137371e-02 -5.43667078e-01
-1.32783726e-01 -8.34382176e-01 -7.74310589e-01 -5.58381200e-01
-2.82865670e-02 -3.07962179e-01 4.12824631e-01 -3.79008532e-01
1.30887508e-01 7.94924080e-01 -1.57335520e+00 3.74308676e-01
7.31386065e-01 1.15581489e+00 2.18582019e-01 9.15579140e-01
1.24460131e-01 9.64240491e-01 -5.97347677e-01 -5.13888121e-01
4.97462690e-01 -9.09531489e-02 -3.88438076e-01 -6.52178824e-01
6.54148221e-01 -7.51863301e-01 -9.13283765e-01 1.14110553e+00
8.25339198e-01 6.55289367e-02 1.09963074e-01 -1.13243163e+00
-7.07084358e-01 5.53355336e-01 -7.87241638e-01 6.75241113e-01
1.73041418e-01 3.47397864e-01 4.96294707e-01 -7.84480333e-01
4.83956873e-01 1.15705943e+00 6.77820206e-01 6.32012486e-01
-9.34203744e-01 -1.07250714e+00 -2.19453409e-01 4.98534709e-01
-1.72785223e+00 -6.66291714e-01 5.80004573e-01 8.20284188e-02
9.15197074e-01 4.82501894e-01 4.06350762e-01 1.05838192e+00
2.99207240e-01 4.56527174e-01 1.08705425e+00 -2.92737596e-02
5.42053394e-02 -1.27534181e-01 -6.83135167e-02 5.54740906e-01
2.29331613e-01 1.29415736e-01 -7.25515723e-01 1.46697059e-01
1.39557934e+00 -1.45215109e-01 -3.71877789e-01 4.44316775e-01
-1.05348313e+00 2.47387439e-01 3.43950033e-01 2.14872137e-01
-2.08643004e-01 4.02923077e-01 4.98619348e-01 3.52862209e-01
2.60588288e-01 -2.63553150e-02 -3.27397883e-01 -2.35514298e-01
-1.22686863e+00 1.54637307e-01 2.21845224e-01 9.54578876e-01
3.20801049e-01 6.37721241e-01 -2.97535986e-01 6.16320431e-01
-1.88412033e-02 6.14414334e-01 3.51139277e-01 -1.46471345e+00
5.17464995e-01 1.08457178e-01 -5.28452639e-03 -1.05290008e+00
-2.08895713e-01 -4.18081045e-01 -1.34120977e+00 5.48638225e-01
3.23042780e-01 4.08852939e-03 -7.47037888e-01 1.35066676e+00
2.02206567e-01 5.46038508e-01 2.20079973e-01 1.21749818e+00
1.07990670e+00 8.38680387e-01 -2.07938284e-01 -1.96763545e-01
1.48319721e+00 -8.01773727e-01 -8.42637837e-01 1.53894231e-01
5.36852665e-02 -8.59799862e-01 1.05831361e+00 6.92061543e-01
-1.36422408e+00 -8.34823251e-01 -1.31036890e+00 -2.13313833e-01
9.57496390e-02 2.71004081e-01 7.24292696e-01 5.86768985e-01
-1.30560303e+00 9.00436044e-01 -8.25667441e-01 2.57705674e-02
6.37249887e-01 4.34776127e-01 -1.84899122e-01 -1.09575473e-01
-1.27474368e+00 4.95699018e-01 2.86091298e-01 1.83665410e-01
-8.36124063e-01 -1.05927789e+00 -3.81940037e-01 7.19147623e-02
3.87335092e-01 -6.02120876e-01 9.57189560e-01 -9.08738434e-01
-1.63166296e+00 5.88235736e-01 2.30042171e-02 -6.45669460e-01
7.03346789e-01 -2.96170503e-01 -7.95680881e-01 3.05865198e-01
-4.73781526e-01 2.23138243e-01 1.04774821e+00 -7.98574567e-01
-6.51376605e-01 -1.68805599e-01 8.76480713e-02 1.27387226e-01
-1.97835848e-01 3.19633156e-01 -7.20324516e-01 -8.27177048e-01
-1.40424281e-01 -5.94499886e-01 -1.31930292e-01 1.12790458e-01
4.15863004e-03 4.83643785e-02 8.66464317e-01 -1.00705695e+00
1.31295502e+00 -2.35902333e+00 -2.98615813e-01 -2.05977187e-01
4.55010325e-01 7.17844903e-01 8.85274783e-02 -1.97103098e-01
-1.87591523e-01 6.00590520e-02 1.24396071e-01 1.99710671e-02
-2.59041697e-01 -1.33872807e-01 -3.33407402e-01 6.64204657e-01
-7.78900646e-03 7.45520294e-01 -5.02352893e-01 -4.86160100e-01
3.64985138e-01 9.72174704e-01 -4.99600589e-01 1.38538912e-01
2.48479486e-01 2.95598626e-01 -3.09415847e-01 7.77707338e-01
9.96805668e-01 -1.63209409e-01 -5.23253269e-02 -7.20131159e-01
-2.90158451e-01 -4.27358923e-03 -1.49862838e+00 1.73979354e+00
-4.30248529e-01 7.92746663e-01 2.71018267e-01 -6.78676128e-01
9.54202533e-01 4.34184849e-01 2.84376383e-01 -9.27595258e-01
2.66648352e-01 1.42180860e-01 -3.71032000e-01 -3.95115942e-01
8.39197993e-01 2.64478445e-01 2.31650710e-01 -1.39533207e-01
-1.75308120e-02 5.53893268e-01 -1.46999091e-01 3.17348063e-01
1.22572160e+00 3.81561100e-01 2.52834052e-01 -1.59675896e-01
6.40525758e-01 -3.47649783e-01 6.77556932e-01 5.49442112e-01
-2.88424253e-01 5.25077760e-01 2.07853094e-01 -4.62100834e-01
-1.34588659e+00 -1.20934820e+00 -7.30922446e-02 7.79882312e-01
2.82171190e-01 -2.35537574e-01 -6.99980855e-01 1.61536504e-02
-3.82745415e-01 4.51108992e-01 -8.80440399e-02 1.23512939e-01
-9.62788880e-01 -6.66292012e-01 9.66493726e-01 4.95679170e-01
1.01503813e+00 -7.70035505e-01 -5.87360561e-01 2.62668300e-02
-1.71382889e-01 -1.45684946e+00 -3.62274349e-01 -4.75534201e-01
-8.81949663e-01 -7.14698613e-01 -5.92257798e-01 -4.97204721e-01
1.55848354e-01 3.53941411e-01 1.10244477e+00 6.27819896e-02
-5.03525674e-01 -9.97455791e-03 -2.10494637e-01 1.25182420e-01
-2.73752928e-01 -2.21697688e-01 1.39972210e-01 -1.94076672e-01
1.14501067e-01 -9.31221545e-01 -1.00602865e+00 1.95880532e-01
-8.35026443e-01 1.56584471e-01 5.49717665e-01 4.17566299e-01
7.44069099e-01 4.75944251e-01 4.65499163e-01 -5.03989160e-01
1.80160105e-01 -1.70846328e-01 -1.02092803e+00 -7.19345734e-03
-6.03088439e-01 -1.25047565e-01 9.56292570e-01 -5.09298265e-01
-1.24209535e+00 -2.36239746e-01 -3.20779651e-01 -8.88587773e-01
5.32185808e-02 -2.21778005e-01 -1.55223593e-01 -2.43676230e-01
5.09878635e-01 3.25270593e-01 -1.15882076e-01 -3.97586197e-01
3.08845282e-01 7.27375507e-01 9.86408234e-01 -3.33027095e-01
9.96599495e-01 6.66415095e-01 1.66943818e-01 -7.76939213e-01
-3.05778563e-01 -1.14045322e-01 3.29445116e-02 -3.15542400e-01
8.36283088e-01 -1.36618936e+00 -1.20850015e+00 5.55942714e-01
-9.98523176e-01 -2.54904300e-01 -8.79041329e-02 5.07079422e-01
-3.34707499e-01 3.65043521e-01 -9.87778544e-01 -8.07438910e-01
-9.16983664e-01 -1.05671585e+00 7.13744640e-01 3.48822385e-01
3.41649830e-01 -4.00235116e-01 -4.52431321e-01 4.29564625e-01
7.55844295e-01 2.97544718e-01 2.38200411e-01 -1.21444501e-01
-1.13576066e+00 1.25781804e-01 -6.35627329e-01 5.26845932e-01
-3.03724915e-01 2.21862480e-01 -9.24308777e-01 -5.40604889e-01
1.02965765e-01 2.84590721e-02 7.00217009e-01 4.57431376e-01
1.44150198e+00 -3.89019161e-01 6.98148906e-02 1.23333991e+00
2.03843880e+00 2.49793351e-01 1.10740137e+00 3.22444201e-01
8.17556858e-01 2.19020024e-02 6.85218930e-01 5.50211608e-01
1.32961348e-01 7.22535074e-01 5.57931602e-01 -3.72916162e-02
-4.12341833e-01 2.53127255e-02 5.76179087e-01 8.82422924e-01
-5.96733034e-01 -2.90211350e-01 -4.70106393e-01 2.08056882e-01
-1.47471833e+00 -1.10486472e+00 -2.70273149e-01 2.23446107e+00
5.98556280e-01 2.44379967e-01 -1.49860382e-01 1.79839000e-01
8.67168427e-01 4.35077935e-01 -4.68251824e-01 -4.85210329e-01
-2.04508170e-01 4.63638395e-01 1.02236223e+00 2.50933379e-01
-8.39385629e-01 8.95242393e-01 5.01084757e+00 1.36579800e+00
-1.16198874e+00 3.22156727e-01 6.16767764e-01 -4.41430509e-01
1.51080981e-01 -2.93025881e-01 -1.00363219e+00 4.53927636e-01
1.23424661e+00 1.62258856e-02 9.97084141e-01 7.30946064e-01
4.30789590e-01 2.67155707e-01 -6.17070317e-01 1.39541113e+00
-3.71336266e-02 -1.55124307e+00 1.20310653e-02 -1.62922516e-01
3.17399085e-01 5.16937748e-02 1.93116784e-01 2.08129317e-01
1.26348615e-01 -1.09414160e+00 5.73492348e-01 4.15220410e-01
1.50890851e+00 -1.11950898e+00 6.54888511e-01 -8.56426656e-02
-1.49155295e+00 6.69045821e-02 -4.56386060e-01 8.43221098e-02
1.30916953e-01 4.06618476e-01 -2.53249943e-01 7.68480003e-01
1.16355491e+00 4.01463181e-01 -4.61465120e-01 4.88016754e-01
1.05840266e-01 6.33325100e-01 -1.95205942e-01 3.95216316e-01
-2.17658840e-02 -1.18806995e-01 6.58879042e-01 1.15879893e+00
4.72429335e-01 4.42604780e-01 -2.99235731e-01 5.82759023e-01
-2.28395775e-01 -1.93893611e-02 -2.00841725e-01 2.49736011e-01
7.19621658e-01 1.47733581e+00 -6.36208415e-01 -4.42049891e-01
-4.48614836e-01 1.10027337e+00 2.56372597e-02 2.68484771e-01
-1.46929145e+00 -5.12969494e-01 7.99660802e-01 3.19027662e-01
2.82982647e-01 -1.70185909e-01 -1.85796469e-01 -1.10704041e+00
1.49838224e-01 -1.10029697e+00 1.12379536e-01 -8.87187123e-01
-7.62647331e-01 7.74620056e-01 -3.15890580e-01 -1.33616614e+00
2.98957765e-01 -4.84955281e-01 -3.90869558e-01 7.88929164e-01
-1.66444743e+00 -7.83772051e-01 -6.83078647e-01 9.20525312e-01
5.46352923e-01 -2.48956844e-01 3.89654160e-01 8.64789844e-01
-5.31539440e-01 9.07072365e-01 1.08537607e-01 2.19796956e-01
7.48421967e-01 -8.47028077e-01 6.14800572e-01 1.19709373e+00
-2.89416641e-01 4.46141481e-01 8.00002337e-01 -3.86999398e-01
-1.75272167e+00 -1.33725631e+00 1.39127687e-01 -2.96831466e-02
5.89157462e-01 -1.75768673e-01 -1.03159821e+00 4.91746664e-01
1.50441498e-01 4.75597054e-01 1.47831962e-01 -5.18823326e-01
-4.93014246e-01 -5.68994164e-01 -1.35263336e+00 8.60581458e-01
1.14770675e+00 -4.26275164e-01 -6.45036399e-02 4.76830341e-02
1.30394340e+00 -7.67054617e-01 -1.20268655e+00 4.53326762e-01
5.97843766e-01 -1.15708947e+00 1.30366886e+00 -8.38222131e-02
6.42521441e-01 -5.51185966e-01 -4.29767430e-01 -4.72084194e-01
-3.46139073e-01 -8.66181672e-01 -3.65297794e-01 1.45587993e+00
-8.68668854e-02 -6.34669781e-01 8.47283125e-01 3.08500171e-01
-1.17369451e-01 -7.40882039e-01 -1.05571008e+00 -8.08580101e-01
-3.85134757e-01 -5.17502487e-01 6.28065944e-01 8.41593981e-01
-7.35416114e-01 6.18192134e-03 -6.29363894e-01 5.53908527e-01
1.01292384e+00 -1.86567098e-01 6.32179856e-01 -5.74726403e-01
-5.24658442e-01 -1.58598393e-01 -5.75656235e-01 -8.14925075e-01
-2.67323494e-01 -7.38113105e-01 -4.43772554e-01 -1.27268279e+00
-9.45494026e-02 -3.06701928e-01 -3.81805658e-01 9.48888436e-02
8.13675672e-02 5.77287734e-01 4.16020542e-01 -6.58033555e-03
-5.94563186e-01 1.46544158e-01 1.37272501e+00 1.95082709e-01
-1.75669923e-01 -3.40507239e-01 -4.82752472e-01 7.74917305e-01
1.04253912e+00 -2.19370767e-01 -3.61721456e-01 -4.47384596e-01
1.02063790e-01 5.76015353e-01 6.00816548e-01 -1.22545004e+00
3.02045017e-01 2.40277909e-02 5.45642495e-01 -6.12085819e-01
4.10456419e-01 -7.35179186e-01 6.67155325e-01 4.10135895e-01
4.70723351e-03 2.02029198e-01 2.07167625e-01 2.61150956e-01
2.73931511e-02 1.27583459e-01 1.12131989e+00 -4.45230678e-02
-9.47292328e-01 5.61535776e-01 -1.13872811e-01 9.02623236e-02
7.77412057e-01 -3.10396254e-01 -6.68290913e-01 -3.89933288e-02
-3.31863075e-01 -2.24475458e-01 4.60708976e-01 3.17532390e-01
8.20555031e-01 -1.25901616e+00 -8.97707343e-01 2.76534129e-02
-5.24924219e-01 5.53444214e-02 9.02873635e-01 5.93101978e-01
-9.76042628e-01 1.24113217e-01 -3.81791174e-01 -3.47488552e-01
-1.66437709e+00 7.38400757e-01 2.87946492e-01 -2.60552406e-01
-1.11153030e+00 7.24980950e-01 -9.99207720e-02 2.30762616e-01
1.40243515e-01 -3.65663692e-02 -1.41514346e-01 -4.02989715e-01
9.63541567e-01 9.13338602e-01 2.40613502e-02 -4.84864295e-01
-3.17402661e-01 5.16541898e-01 4.66051511e-02 1.85419068e-01
1.21946740e+00 -1.94883734e-01 -1.66383952e-01 -9.13655609e-02
1.19605696e+00 1.39422119e-01 -1.43231654e+00 -2.02229321e-01
-6.02961123e-01 -6.59378290e-01 3.23701143e-01 -4.99486685e-01
-1.49938381e+00 4.88392919e-01 1.05584693e+00 -2.64513850e-01
1.64775252e+00 -3.48217994e-01 1.10511184e+00 5.28946035e-02
5.14558434e-01 -8.46830010e-01 -1.45044595e-01 9.54192728e-02
6.45823240e-01 -9.60056126e-01 3.72910291e-01 -3.70338440e-01
-4.19122547e-01 1.00241530e+00 6.57003880e-01 -6.22882724e-01
3.44597757e-01 7.56046534e-01 -9.92664844e-02 1.28255635e-01
-7.83093512e-01 3.69525850e-01 -4.55587283e-02 5.03419578e-01
2.30097368e-01 1.65323243e-02 -2.40949705e-01 4.74039525e-01
-3.12973738e-01 2.30217457e-01 6.55705214e-01 5.17928720e-01
-1.89858198e-01 -5.16182840e-01 -4.86540735e-01 3.75303596e-01
-1.13678527e+00 -3.15876156e-01 6.19487941e-01 7.32616067e-01
1.75012723e-01 6.76914752e-01 1.47359982e-01 -5.83510876e-01
3.13980550e-01 -5.84541202e-01 5.04722238e-01 2.38947913e-01
-6.11424923e-01 -1.14619911e-01 2.00715456e-02 -1.12842464e+00
-4.36032981e-01 -2.29860529e-01 -1.19317138e+00 -1.01984751e+00
-6.99119195e-02 -3.38586360e-01 7.81450450e-01 4.38946366e-01
3.34380567e-01 1.09310937e+00 4.84543651e-01 -6.31091177e-01
-4.42082584e-01 -5.68117261e-01 -5.99605083e-01 -3.04694064e-02
3.26934367e-01 -1.65291920e-01 -3.43223333e-01 1.03040539e-01] | [11.004667282104492, -1.8960858583450317] |
b88d2e3d-9437-405c-bacc-f6e6b5f6f5dc | real-time-ground-fault-detection-for-inverter | 2304.12445 | null | https://arxiv.org/abs/2304.12445v1 | https://arxiv.org/pdf/2304.12445v1.pdf | Real-Time Ground Fault Detection for Inverter-Based Microgrid Systems | Ground fault detection in inverter-based microgrid systems is challenging, particularly in a real-time setting, as the fault current deviates slightly from the nominal value. This difficulty is reinforced when natural disturbances exhibit similar output patterns as a faulty setting does. The conventional solution of installing more relays to obtain additional measurements is costly and also increases the complexity of the system. In this paper, we propose diagnosis schemes based on optimization-based fault detection filters with the output current as the only measurement. Modeling the microgrid dynamics and the diagnosis filter, we formulate the filter design as a linear programming (LP) problem that accounts for decoupling a class of disturbances and ensuring fault sensitivity simultaneously. Next, we robustify the filter to disturbances that cannot be fully decoupled. To this end, we leverage tools from the existing literature and extend the optimization program to a quadratic programming (QP) problem in which the filter is trained for this class of disturbances. To ease the computational effort, we also provide an approximate but analytical solution to this QP. Additionally, we use classical statistical results to provide a thresholding mechanism that enjoys probabilistic false-alarm guarantees. Finally, we verify the effectiveness of the proposed methods through several numerical simulations. | ['Peyman Mohajerin Esfahani', 'Yucheng Liao', 'Jingwei Dong'] | 2023-04-24 | null | null | null | null | ['fault-detection'] | ['miscellaneous'] | [-1.92052443e-02 4.56635356e-02 8.00149515e-02 5.50281033e-02
-7.74424374e-01 -7.78446496e-01 1.02326035e-01 1.87167838e-01
2.04963520e-01 9.45829928e-01 -3.21662843e-01 -3.35316569e-01
-4.68339741e-01 -6.62631869e-01 -7.32075334e-01 -1.10658324e+00
-2.03794867e-01 -9.00340732e-03 -1.38653040e-01 -6.30617887e-02
-7.00335056e-02 4.53759074e-01 -1.13446999e+00 -4.32852566e-01
1.33050370e+00 1.22669840e+00 -9.46979504e-03 3.49444956e-01
9.47974503e-01 4.72624332e-01 -1.15255213e+00 1.82363704e-01
3.56866330e-01 -3.86649340e-01 -2.74965107e-01 7.35970140e-01
-1.90304443e-01 -5.52758813e-01 -5.40870488e-01 1.53999412e+00
4.84724492e-01 2.08499044e-01 5.46510696e-01 -1.86419964e+00
-2.43364513e-01 3.44793439e-01 -3.66436839e-01 1.29412621e-01
3.83049041e-01 1.37779742e-01 5.74807584e-01 -8.59593391e-01
-3.05061229e-02 9.27669644e-01 3.60873729e-01 -2.87223514e-03
-1.22102392e+00 -3.90550256e-01 4.40718859e-01 2.56051719e-01
-1.45508254e+00 -1.48869947e-01 6.59640610e-01 -4.19221431e-01
7.17459440e-01 1.98101655e-01 4.76699203e-01 7.22514153e-01
2.75262266e-01 5.88066041e-01 1.01698804e+00 -1.43360466e-01
5.46544492e-01 -6.65164813e-02 2.97645718e-01 3.48806590e-01
6.84102297e-01 -4.44828831e-02 -2.24979669e-01 -2.08285660e-01
3.78876239e-01 5.28609715e-02 -1.05996704e+00 -2.36207172e-01
-7.63672352e-01 7.61547387e-01 1.96957573e-01 -6.88781515e-02
-4.18900192e-01 -1.60670757e-01 1.62046075e-01 6.01190686e-01
3.24300438e-01 1.87951803e-01 -3.64065081e-01 2.58260667e-01
-6.71956480e-01 1.79569185e-01 1.17042184e+00 8.69849205e-01
4.28901106e-01 5.73024511e-01 2.13439986e-01 2.62988031e-01
4.20941502e-01 5.69695115e-01 9.81675461e-02 -5.28557718e-01
3.33888471e-01 3.98702770e-01 5.31347632e-01 -8.18759263e-01
-2.59623170e-01 -9.22156334e-01 -8.91571522e-01 2.98265219e-01
5.02696753e-01 -5.30629039e-01 -6.24306977e-01 1.53975940e+00
2.46221110e-01 1.80263862e-01 -2.93442444e-03 9.68192756e-01
-3.43181312e-01 6.49922669e-01 -6.07241273e-01 -7.54747748e-01
1.20599282e+00 -5.20516574e-01 -1.04543996e+00 -9.50747505e-02
3.22994143e-01 -4.38165724e-01 6.23372018e-01 7.98805058e-01
-8.52267325e-01 1.97633151e-02 -1.58622742e+00 5.93373775e-01
-2.03006625e-01 3.30457896e-01 -1.22969396e-01 5.82530022e-01
-9.38955426e-01 5.00410259e-01 -7.82508910e-01 -2.15347737e-01
-1.83716193e-01 4.37538207e-01 -9.78098661e-02 5.34951650e-02
-1.26110661e+00 1.08511889e+00 3.47191602e-01 4.77917939e-01
-9.07052517e-01 -6.71595335e-01 -8.63756299e-01 8.92046317e-02
8.58014345e-01 -3.39759201e-01 1.19577610e+00 -5.31295598e-01
-1.56208968e+00 -1.40510142e-01 -3.28443982e-02 -5.72421372e-01
5.36764622e-01 -3.09707940e-01 -3.70496154e-01 3.84593457e-01
-3.14505622e-02 -6.12399042e-01 9.64834154e-01 -1.21622586e+00
-7.14261532e-01 -2.10228518e-01 1.61090493e-01 -1.23437773e-02
-8.11389610e-02 -2.34665081e-01 5.99051639e-02 -6.98983490e-01
1.12987615e-01 -4.74002987e-01 -2.26752251e-01 -1.97194800e-01
-7.68934131e-01 6.26292750e-02 8.41699898e-01 -8.32914889e-01
1.18499780e+00 -2.15067339e+00 9.05706435e-02 5.66668510e-01
-1.52244523e-01 -5.24290577e-02 5.52784085e-01 7.84931302e-01
3.96637842e-02 -3.27274173e-01 -3.76944453e-01 -1.77421317e-01
2.80904680e-01 4.31787461e-01 -5.09396493e-01 1.09299421e+00
4.79340911e-01 1.81155130e-01 -8.74708712e-01 2.74852782e-01
2.89358705e-01 1.48293123e-01 -2.90588051e-01 2.77782232e-01
4.98868823e-02 4.14348245e-01 -3.38367820e-01 6.84399188e-01
6.92037761e-01 -1.19015396e-01 3.34447950e-01 -4.87873226e-01
-5.60715534e-02 1.61478415e-01 -1.79303992e+00 1.05560553e+00
-3.42523485e-01 2.47228011e-01 7.71691799e-01 -1.66760910e+00
6.00436449e-01 4.02359366e-01 3.93991321e-01 -3.03392589e-01
2.42679596e-01 2.37371489e-01 -4.67757545e-02 -3.07489753e-01
-3.91037576e-02 -3.62026654e-02 -1.11969426e-01 3.11204672e-01
2.59304732e-01 -2.80969858e-01 3.97830606e-01 -1.43940626e-02
1.05490506e+00 -4.42607135e-01 4.23599571e-01 -7.62880921e-01
5.74480236e-01 -1.82865232e-01 1.00372016e+00 5.06987751e-01
-6.51575206e-03 7.05343485e-02 7.98035085e-01 3.62230748e-01
-4.47886109e-01 -1.09797800e+00 -3.16264987e-01 7.30932727e-02
2.36477315e-01 -1.42864153e-01 -6.63605213e-01 -5.83984613e-01
3.82657349e-01 7.57419825e-01 -3.57559085e-01 -4.31620777e-01
-2.46013373e-01 -1.23578918e+00 1.17621265e-01 3.89378875e-01
1.74241379e-01 1.80801690e-01 -4.70606476e-01 4.25840139e-01
8.71880278e-02 -9.86087918e-01 -4.31444228e-01 8.23232114e-01
-4.40120220e-01 -1.29506814e+00 -4.29356217e-01 -7.15951443e-01
1.04761469e+00 3.78641896e-02 7.40367174e-01 1.10608362e-01
-1.94300279e-01 4.58375722e-01 -2.13352501e-01 -3.31897557e-01
-1.02885321e-01 -4.43227202e-01 5.93225658e-01 2.13322848e-01
-3.44521075e-01 -6.06228650e-01 -3.35039943e-01 4.32142615e-01
-9.44712400e-01 -5.49765229e-01 2.94301033e-01 1.19796038e+00
3.61086458e-01 9.16640043e-01 1.13501430e+00 -3.54713917e-01
7.25631058e-01 -7.26761341e-01 -1.12774229e+00 9.57124233e-02
-6.41492844e-01 -2.18087241e-01 1.11858058e+00 -5.58021426e-01
-6.87628865e-01 8.47151503e-02 1.52291149e-01 -3.08051378e-01
1.85981482e-01 7.08916545e-01 -8.30158114e-01 -3.33922982e-01
7.59836435e-02 2.02423722e-01 1.27079397e-01 -4.17714596e-01
1.53499335e-01 6.91161692e-01 7.47053266e-01 -5.21368265e-01
1.26289225e+00 2.64974535e-01 6.48358762e-02 -8.11387300e-01
-7.23653495e-01 -3.09318542e-01 -2.59203553e-01 -2.89520830e-01
7.73787200e-02 -9.88229036e-01 -1.20479012e+00 6.48121357e-01
-7.70013392e-01 -1.69257268e-01 -1.18554138e-01 4.50461656e-01
-3.01581144e-01 4.92080569e-01 -6.38077736e-01 -1.15856540e+00
3.62118110e-02 -1.18959057e+00 8.03341746e-01 2.62308538e-01
2.98243314e-01 -1.04771292e+00 -3.33210230e-01 -4.05647397e-01
1.54702693e-01 6.19412124e-01 6.28856301e-01 -4.42076385e-01
-5.30421615e-01 -4.23575342e-01 1.84920684e-01 6.26034439e-01
4.10129756e-01 -1.22435674e-01 -6.01180911e-01 -8.21829915e-01
7.55615354e-01 -5.47858216e-02 2.41987005e-01 2.17794672e-01
7.33984888e-01 -5.52676976e-01 -2.10943267e-01 4.74713027e-01
1.78504694e+00 3.37504238e-01 -8.01549256e-02 1.77566409e-01
3.55188161e-01 4.78018373e-01 6.20406806e-01 7.51194656e-01
4.75631595e-01 5.07029176e-01 5.85531592e-01 -5.04263416e-02
6.50344014e-01 1.44777536e-01 5.59005678e-01 7.49759614e-01
7.51333237e-01 -5.37083983e-01 -5.10960221e-01 5.95969677e-01
-1.85209513e+00 -3.89767677e-01 7.19734468e-04 2.44997191e+00
8.01296949e-01 1.01185746e-01 4.43334579e-02 8.38723540e-01
6.38558030e-01 -4.36325073e-01 -6.76098049e-01 -3.36710699e-02
-3.25606346e-01 -7.95278549e-02 7.01404929e-01 5.97527981e-01
-1.02156115e+00 -2.93902643e-02 5.80291080e+00 4.19928610e-01
-1.17151964e+00 -7.52063543e-02 4.01476055e-01 4.33751903e-02
2.44518623e-01 -6.50859028e-02 -6.10344172e-01 8.27666104e-01
1.07012260e+00 -7.29508340e-01 5.43961644e-01 5.70470870e-01
7.30204821e-01 -4.89871055e-01 -1.18598270e+00 6.59359992e-01
5.77090010e-02 -4.97060388e-01 -4.12581325e-01 1.09324887e-01
6.82904303e-01 -4.27853048e-01 -1.53719380e-01 -2.13807356e-02
2.16635108e-01 -6.29703224e-01 9.75350201e-01 3.66482168e-01
1.77965492e-01 -9.65486705e-01 7.71429420e-01 3.65223706e-01
-1.10687757e+00 -5.39566696e-01 -1.86254799e-01 -4.70690459e-01
6.60460770e-01 1.21509612e+00 -4.62298512e-01 1.13702750e+00
4.64768171e-01 6.22530878e-01 -2.53609598e-01 1.41719139e+00
-6.37924254e-01 7.24789202e-01 -6.95921838e-01 3.28476787e-01
-1.40242085e-01 -3.28369260e-01 5.98736465e-01 7.33723462e-01
3.09129953e-01 6.15638234e-02 7.90253282e-01 6.57626510e-01
1.49239570e-01 -3.65952015e-01 -3.40062678e-01 3.24866205e-01
7.89232671e-01 1.05354273e+00 -6.05512500e-01 -3.40916336e-01
-5.82205415e-01 9.04778481e-01 -1.00414373e-01 6.65619910e-01
-7.25751758e-01 -5.83769321e-01 9.14406657e-01 -4.08468425e-01
5.93305789e-02 -2.30126396e-01 -1.98028564e-01 -1.31741107e+00
5.00129938e-01 -9.62774336e-01 2.70699948e-01 -2.84405202e-01
-1.51381135e+00 2.20987752e-01 1.59442872e-02 -1.38321149e+00
-5.07041276e-01 -4.91345167e-01 -7.82245457e-01 1.06443703e+00
-1.57158899e+00 -4.78952438e-01 -5.53247444e-02 5.78791440e-01
2.31869549e-01 5.47821939e-01 4.66772914e-01 5.93991518e-01
-1.18612933e+00 4.82313722e-01 4.43604916e-01 -7.95783550e-02
4.68028367e-01 -1.60753500e+00 -1.37738034e-01 1.61237967e+00
-6.21205926e-01 4.28030699e-01 1.21822917e+00 -6.33526146e-01
-1.97675920e+00 -1.15520155e+00 3.50861460e-01 2.14007482e-01
1.18017447e+00 -4.97546554e-01 -1.00330150e+00 5.26040614e-01
1.96053937e-01 1.14387594e-01 1.61396027e-01 -7.20354199e-01
3.21571440e-01 -2.06459969e-01 -1.31899977e+00 2.25130349e-01
3.50784928e-01 -5.81618309e-01 -4.80541319e-01 4.88404006e-01
3.75904232e-01 -4.85540837e-01 -1.19197047e+00 4.06591237e-01
-5.52109443e-03 -1.99637860e-01 7.69538164e-01 -6.32866770e-02
-4.74691600e-01 -9.24114406e-01 -3.16934288e-01 -1.91192424e+00
8.56689587e-02 -1.00354719e+00 -4.97154385e-01 1.23618829e+00
9.31494236e-02 -1.16014111e+00 2.22236633e-01 2.82805771e-01
-2.20383763e-01 -6.58030152e-01 -9.72032666e-01 -1.19041049e+00
-6.68207332e-02 -1.15291961e-01 3.59896064e-01 8.56765389e-01
6.07632995e-01 -4.58336771e-02 -2.92777270e-01 1.33051860e+00
8.98006201e-01 -3.88775095e-02 3.09605032e-01 -9.00662482e-01
-5.56686759e-01 -5.73411025e-02 -3.30017477e-01 -9.06603634e-01
1.22895405e-01 -3.43201637e-01 6.01135492e-01 -1.31226504e+00
-5.04324436e-01 3.03643700e-02 -1.54656023e-01 2.96185791e-01
-4.56567943e-01 -3.74488942e-02 -4.06734049e-02 -1.72591344e-01
-4.04229537e-02 6.90845788e-01 5.19214809e-01 -2.63012677e-01
9.19014029e-03 1.44161925e-01 -4.27772224e-01 5.77828467e-01
7.64539659e-01 -9.78176445e-02 -5.80251515e-01 -2.01670527e-01
-1.38408750e-01 3.76477420e-01 2.58314013e-01 -1.02546799e+00
3.91728520e-01 8.09234604e-02 1.51191115e-01 -4.30118412e-01
1.42436147e-01 -1.04360473e+00 1.01315312e-01 7.47707427e-01
3.21392059e-01 2.66020715e-01 -5.00865653e-02 6.13505185e-01
-4.03963208e-01 -1.58773884e-01 7.05944777e-01 4.73478019e-01
-1.82861209e-01 2.00350843e-02 -9.16702211e-01 -1.06254615e-01
1.05059540e+00 2.07823277e-01 -4.47131753e-01 -4.87378955e-01
-7.94218481e-01 8.62087548e-01 4.69910413e-01 1.33288473e-01
3.46309453e-01 -9.17115569e-01 -4.09963846e-01 1.97488070e-01
-1.54680178e-01 -1.83925390e-01 5.73950857e-02 1.31313396e+00
8.40258449e-02 3.67267102e-01 3.06267351e-01 -4.52287287e-01
-6.84347332e-01 5.82368970e-01 6.09370351e-01 4.26755846e-02
-4.23656493e-01 2.84854680e-01 -1.09557472e-01 1.18906118e-01
4.82608318e-01 -8.05650592e-01 1.19645469e-01 1.88283399e-01
5.82152665e-01 5.79184830e-01 4.84786481e-01 -2.68736720e-01
-3.03825557e-01 2.33958751e-01 3.78216177e-01 7.91069493e-02
1.07508564e+00 -5.76974511e-01 -8.63050371e-02 4.44988579e-01
8.68402123e-01 -2.16262475e-01 -1.44638979e+00 3.04630194e-02
-3.76941822e-02 -1.85164765e-01 1.50450945e-01 -7.43855000e-01
-1.21849048e+00 3.42641205e-01 3.59750718e-01 6.86752200e-01
1.47150385e+00 -5.27765512e-01 3.83418083e-01 2.95729458e-01
6.38848484e-01 -8.46990466e-01 -4.16890085e-01 3.11429292e-01
6.25588894e-01 -6.97875500e-01 -9.88290980e-02 -6.49465799e-01
9.44893137e-02 1.02876329e+00 2.98434705e-01 -5.62925875e-01
8.98982227e-01 6.76813543e-01 -2.40464300e-01 2.20341995e-01
-7.88553596e-01 -1.39178680e-02 -3.30838598e-02 4.27482426e-01
-1.21276490e-01 1.59952402e-01 -5.61641872e-01 7.66228378e-01
6.86478987e-02 -2.44223773e-01 9.85320568e-01 1.13303089e+00
-2.55647600e-01 -9.14074004e-01 -7.21007168e-01 5.24852216e-01
-6.93146527e-01 2.39640638e-01 1.50666356e-01 6.23107135e-01
-1.51167095e-01 1.57508039e+00 -8.20507482e-02 1.30291870e-02
6.82169378e-01 -7.73117989e-02 2.21067190e-01 -4.36976969e-01
-2.73715764e-01 3.02706480e-01 -4.23889756e-02 -5.43902159e-01
-1.01148494e-01 -6.92912161e-01 -1.16173410e+00 4.81952466e-02
-8.34849894e-01 4.52250510e-01 4.99719173e-01 1.24622273e+00
-3.71990986e-02 9.31587934e-01 9.60905850e-01 -8.22998106e-01
-1.15727496e+00 -8.49316835e-01 -1.06727982e+00 -1.17954463e-01
6.31892204e-01 -9.22734141e-01 -9.65404034e-01 -1.16302304e-01] | [5.621012210845947, 2.577685594558716] |
a16f9245-a1ba-4960-9668-35feacc245cd | an-investigation-on-selecting-audio-pre | 2208.06127 | null | https://arxiv.org/abs/2208.06127v1 | https://arxiv.org/pdf/2208.06127v1.pdf | An investigation on selecting audio pre-trained models for audio captioning | Audio captioning is a task that generates description of audio based on content. Pre-trained models are widely used in audio captioning due to high complexity. Unless a comprehensive system is re-trained, it is hard to determine how well pre-trained models contribute to audio captioning system. To prevent the time consuming and energy consuming process of retraining, it is necessary to propose a preditor of performance for the pre-trained model in audio captioning. In this paper, a series of pre-trained models are investigated for the correlation between extracted audio features and the performance of audio captioning. A couple of predictor is proposed based on the experiment results.The result demonstrates that the kurtosis and skewness of audio features extracted may act as an indicator of the performance of audio captioning systems for pre-trained audio due to the high correlation between kurtosis and skewness of audio features and the performance of audio captioning systems. | ['Shengchen Li', 'Peiran Yan'] | 2022-08-12 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 2.34968677e-01 9.15898606e-02 2.71975875e-01 -3.03022921e-01
-1.06428981e+00 -5.62339127e-01 2.11501881e-01 1.77855775e-01
-1.06750511e-01 6.52422190e-01 6.00538552e-01 1.59734592e-01
1.74746998e-02 -2.23678112e-01 -7.43935108e-01 -3.93904537e-01
-2.17508927e-01 3.57446253e-01 6.45076707e-02 -1.20346583e-01
2.70144761e-01 1.19021960e-01 -1.94156516e+00 6.74096465e-01
4.96450275e-01 9.87508535e-01 3.15250069e-01 1.15383780e+00
-8.71835947e-02 5.19607842e-01 -9.40375388e-01 -3.04673407e-02
-5.08299023e-02 -5.87646127e-01 -7.10052013e-01 -1.00819498e-01
1.21790290e-01 -5.29308915e-02 -3.67557816e-02 8.82099926e-01
6.70202672e-01 -5.08347228e-02 5.63548148e-01 -1.56704402e+00
-3.77330512e-01 8.77987504e-01 7.45458528e-02 3.79568666e-01
7.94147909e-01 -1.43621489e-01 9.38630402e-01 -7.86939800e-01
2.98216511e-02 9.85965073e-01 5.45405805e-01 4.32353765e-01
-7.33141780e-01 -8.91899526e-01 -2.29617745e-01 2.53139764e-01
-1.46264398e+00 -5.10813355e-01 9.62319374e-01 -7.01623976e-01
4.29854721e-01 4.55777526e-01 5.67181945e-01 8.82359147e-01
-6.12902045e-02 4.74843711e-01 8.08417559e-01 -5.93015194e-01
2.06324860e-01 5.07010818e-01 -6.36009350e-02 5.07222712e-02
5.19741476e-02 -1.45640150e-01 -8.33382905e-01 -2.77954310e-01
3.97333354e-01 -6.65458977e-01 -3.64550292e-01 7.03353658e-02
-9.86949742e-01 6.53312624e-01 -3.05565689e-02 4.84088749e-01
-4.07886535e-01 6.56580254e-02 6.33047521e-01 7.47559592e-02
2.46974379e-01 7.48312712e-01 -2.30428994e-01 -6.27386212e-01
-1.14760041e+00 9.82986689e-02 4.67953682e-01 6.92576289e-01
4.92471039e-01 1.23138860e-01 -2.52093136e-01 7.77948201e-01
3.33446383e-01 4.56611484e-01 8.57959807e-01 -5.52973390e-01
4.65614170e-01 4.30731624e-01 2.52773762e-01 -7.25971043e-01
-1.71236664e-01 -2.66450316e-01 -2.65900850e-01 -4.09210056e-01
1.57703042e-01 -3.49671990e-01 -8.27534556e-01 1.45009375e+00
-8.16394091e-02 4.36896652e-01 3.49693656e-01 1.05139625e+00
9.96800780e-01 1.32832778e+00 2.45513171e-01 -3.01125288e-01
1.35832942e+00 -6.86676025e-01 -9.34825778e-01 -1.30281642e-01
2.90012032e-01 -1.26342738e+00 1.32049227e+00 2.41301432e-01
-9.07238543e-01 -8.63784313e-01 -9.86419857e-01 5.24662554e-01
-2.77696270e-02 2.93676525e-01 3.39138001e-01 7.23020971e-01
-9.08824027e-01 2.06492320e-01 -3.33300650e-01 -2.97669381e-01
-2.50111580e-01 4.52973127e-01 -2.07964361e-01 4.69921350e-01
-1.33034182e+00 5.28183401e-01 6.20570362e-01 8.47504437e-02
-9.11642194e-01 -5.64880073e-01 -4.58749503e-01 3.49856138e-01
-2.29882151e-01 -2.57073909e-01 1.58610964e+00 -1.09832013e+00
-1.42523253e+00 2.23636329e-01 1.64828017e-01 -3.55600059e-01
2.33106181e-01 -3.04864556e-01 -7.88753867e-01 3.68387550e-01
-3.93575095e-02 7.83172786e-01 9.63783801e-01 -1.27511942e+00
-6.33159637e-01 1.67118266e-01 -2.96447337e-01 4.28457737e-01
-5.30647159e-01 2.15760618e-01 -1.70064732e-01 -4.58827376e-01
-1.49739772e-01 -9.52345073e-01 3.37080538e-01 -5.56357443e-01
-2.63231725e-01 -1.73192590e-01 9.98869419e-01 -7.03189492e-01
1.51707518e+00 -2.42836833e+00 -5.38835287e-01 2.31511816e-01
-4.96670038e-01 4.99595016e-01 -1.51503339e-01 4.41235244e-01
-3.39315236e-01 2.70644188e-01 7.64586255e-02 1.98294103e-01
-4.36810553e-02 -1.72952339e-01 -4.26254898e-01 -3.11098080e-02
4.31065738e-01 2.81671673e-01 -8.91531110e-01 -6.89616859e-01
7.13702738e-02 6.47244155e-01 -3.27919215e-01 6.42945290e-01
-1.82378963e-01 3.41406673e-01 -4.90878612e-01 4.38432842e-01
4.19538200e-01 7.42814839e-02 -2.55343020e-01 -2.30847627e-01
3.37430164e-02 4.12152827e-01 -1.20984888e+00 1.21599209e+00
-2.53499120e-01 8.98796141e-01 -8.19328278e-02 -7.64247596e-01
1.22098565e+00 1.00771880e+00 5.34835219e-01 -4.51845139e-01
3.55893940e-01 3.49525005e-01 2.49991752e-02 -9.78160739e-01
6.77599013e-01 -1.62791073e-01 -1.63214102e-01 5.23478806e-01
4.94083855e-03 -1.50647402e-01 1.27814531e-01 -6.51966929e-02
6.37486041e-01 -1.64405927e-01 -9.22526345e-02 1.26914261e-02
6.80823565e-01 -1.12303674e-01 5.96171804e-02 3.41528028e-01
-3.17713916e-01 8.96759033e-01 1.72297180e-01 -1.00843832e-02
-1.17053521e+00 -7.30972111e-01 -2.02175565e-02 1.30814576e+00
-2.25758389e-01 -2.84971774e-01 -8.80485356e-01 -2.12673113e-01
-4.12674338e-01 5.23945034e-01 -3.74809980e-01 -2.96387941e-01
-3.56800973e-01 -4.64246690e-01 8.56955767e-01 3.49166781e-01
3.23959023e-01 -1.15044451e+00 -3.87619108e-01 3.22636187e-01
-5.89373469e-01 -1.11345947e+00 -5.08484602e-01 5.77420322e-03
-7.15160429e-01 -7.04922259e-01 -9.46733952e-01 -1.04408658e+00
4.29262221e-01 1.20734252e-01 7.42630005e-01 5.09523265e-02
3.00628811e-01 4.62620527e-01 -7.48590946e-01 -9.35974360e-01
-9.46122706e-01 2.97581375e-01 1.64324909e-01 1.65130183e-01
2.23452657e-01 -4.54873711e-01 -4.04859960e-01 4.09564883e-01
-9.51276779e-01 -6.72584772e-02 7.16983020e-01 3.79394978e-01
3.74591619e-01 9.16771218e-02 9.40453231e-01 -2.47538373e-01
1.15257919e+00 -4.65832889e-01 -3.09835523e-01 3.31982493e-01
-3.68718565e-01 1.17345043e-01 6.59701407e-01 -7.78216481e-01
-7.88398981e-01 2.93633074e-01 -7.35871214e-03 -3.64920139e-01
-2.20522180e-01 6.04415357e-01 -8.94009396e-02 1.39759481e-01
6.18502736e-01 2.04985276e-01 3.74723561e-02 -4.92685467e-01
-2.12225705e-01 1.26045227e+00 6.50074542e-01 -3.52772772e-01
8.00585270e-01 -1.76415041e-01 -3.37636501e-01 -9.88905072e-01
-7.68987894e-01 -6.25539601e-01 -2.25491554e-01 -7.43539393e-01
8.09156895e-01 -1.04848337e+00 -4.67067480e-01 2.30157837e-01
-1.15344548e+00 2.95912683e-01 -1.01112269e-01 8.98418784e-01
-4.36102927e-01 2.21224800e-01 -2.48606522e-02 -1.11748040e+00
-5.60253859e-01 -1.00403607e+00 9.07639027e-01 4.43283051e-01
-5.09328604e-01 -5.74599147e-01 3.40333760e-01 5.73841631e-01
3.38823408e-01 -6.80118380e-03 6.34028554e-01 -9.57558870e-01
-1.48149788e-01 -5.08702338e-01 -2.97869891e-02 5.62074542e-01
-5.32508530e-02 8.72028396e-02 -1.43410182e+00 -7.39602000e-02
-1.43922433e-01 -2.85103172e-01 4.54535522e-02 3.24571013e-01
9.38296854e-01 -3.68037909e-01 1.95800707e-01 -5.83236627e-02
1.06843483e+00 4.71311033e-01 6.44256592e-01 3.94031733e-01
2.38053039e-01 5.09542346e-01 8.59797359e-01 3.97279441e-01
-1.61435187e-01 5.22819161e-01 4.02815670e-01 7.37725496e-02
3.23027931e-02 -6.62901282e-01 6.98986709e-01 1.16688645e+00
7.04447553e-02 -3.17944378e-01 -9.13718641e-01 6.38247848e-01
-1.42707491e+00 -1.00498486e+00 -1.05537653e-01 2.14841080e+00
7.31048763e-01 2.16829970e-01 4.33508128e-01 7.64976263e-01
1.19410706e+00 -2.87300885e-01 -1.48897082e-01 -7.65923083e-01
1.92632556e-01 1.89990867e-02 1.31422833e-01 2.94042498e-01
-8.93016040e-01 4.48357850e-01 7.19044113e+00 5.08222044e-01
-1.36155319e+00 -2.10054621e-01 3.96792531e-01 2.53934078e-02
-1.42861471e-01 -1.41789854e-01 -6.10239863e-01 8.17156553e-01
1.54565632e+00 -5.32883584e-01 1.30238205e-01 7.74707377e-01
5.23569345e-01 -5.46758696e-02 -1.16510439e+00 1.13856030e+00
1.37317687e-01 -9.09683645e-01 3.31066102e-01 -1.84701353e-01
4.90352273e-01 -2.68656492e-01 1.06566519e-01 3.38287115e-01
-5.24058342e-01 -8.74360800e-01 8.05448711e-01 2.69149810e-01
5.69329739e-01 -7.20301270e-01 1.25983512e+00 8.18316936e-02
-1.18167341e+00 -7.69329593e-02 -1.69907644e-01 -1.51667371e-01
3.74746054e-01 2.57827878e-01 -1.81999123e+00 2.99573123e-01
5.63689888e-01 -1.48425758e-01 -6.72010362e-01 1.75008941e+00
1.24904908e-01 1.21109879e+00 -4.18279201e-01 -2.42203265e-01
1.92791790e-01 1.88356146e-01 5.97971678e-01 1.31665540e+00
8.20417345e-01 -1.47717193e-01 -1.52600348e-01 3.62087280e-01
7.15743229e-02 5.25975049e-01 -3.48978102e-01 -6.21568024e-01
6.44346476e-01 8.73943210e-01 -5.53685784e-01 -1.14399262e-01
-2.12866999e-03 3.19174856e-01 -5.31351924e-01 1.18402012e-01
-1.06147945e+00 -3.75538230e-01 1.62398264e-01 5.01542866e-01
8.38189647e-02 9.61039066e-02 -1.30742684e-01 -5.52589238e-01
1.12233348e-01 -7.73790002e-01 3.26262176e-01 -1.37977934e+00
-9.54754412e-01 9.35203969e-01 1.91144571e-01 -1.69620013e+00
-4.94750440e-01 -5.71102276e-02 -6.58256829e-01 6.75716579e-01
-1.21530807e+00 -8.24111402e-01 -3.67374927e-01 4.20658141e-01
4.73622590e-01 -3.05496275e-01 7.27494061e-01 5.44013679e-01
-3.24687004e-01 6.19700134e-01 -1.68939233e-01 9.19139683e-02
8.08102965e-01 -7.46916652e-01 -2.01972842e-01 6.97900772e-01
1.85705811e-01 1.13280445e-01 1.45585430e+00 -3.99307251e-01
-9.60290968e-01 -1.02450323e+00 1.03958571e+00 -1.73108891e-01
5.25389671e-01 -2.17079744e-02 -1.09356463e+00 1.94455802e-01
1.34367675e-01 -4.49696422e-01 1.05646145e+00 -7.38888383e-02
2.45955121e-02 -3.24617356e-01 -8.26131761e-01 9.75415260e-02
1.54022723e-01 -7.34650910e-01 -9.44793642e-01 1.98989615e-01
5.80640256e-01 -1.52761802e-01 -7.59100556e-01 1.69938222e-01
5.78139067e-01 -5.49926698e-01 6.01975799e-01 -3.46732944e-01
2.14443862e-01 -4.34994966e-01 3.77873443e-02 -1.05102754e+00
-1.16988882e-01 -7.95552671e-01 3.74547064e-01 1.66369843e+00
8.79207909e-01 -4.76449393e-02 8.62209976e-01 4.94035661e-01
-2.13721305e-01 -8.43892768e-02 -8.50639045e-01 -8.31377149e-01
-4.57673967e-01 -3.99475306e-01 6.15170658e-01 6.95692062e-01
2.00726271e-01 6.44814074e-01 -4.78039384e-01 3.89652848e-01
4.63121617e-03 -3.58236492e-01 7.03500628e-01 -1.42090881e+00
8.08172598e-02 -6.46791309e-02 -6.20988667e-01 -4.08180714e-01
-9.32430699e-02 -6.20430052e-01 3.21220130e-01 -1.48694003e+00
1.13145806e-01 -1.96748108e-01 -3.44600081e-01 4.17120993e-01
-4.82260920e-02 3.85452420e-01 2.47102752e-01 4.01433885e-01
-4.56790030e-01 3.58525276e-01 1.03438878e+00 -8.44723508e-02
-6.25778556e-01 2.98954457e-01 -5.92104077e-01 5.02224863e-01
1.04769289e+00 -7.40070760e-01 -6.76846445e-01 -1.51162863e-01
3.30553263e-01 9.57911983e-02 1.04781650e-01 -1.46185434e+00
6.84148371e-02 -1.79374591e-03 1.16209455e-01 -6.77231610e-01
4.62440878e-01 -8.95838499e-01 4.17534858e-01 2.23625556e-01
-5.37543714e-01 2.37899557e-01 4.00256455e-01 2.69334316e-01
-7.02672005e-01 -6.07534289e-01 6.21667683e-01 1.16231978e-01
-5.01303911e-01 -1.91839173e-01 -7.76344121e-01 -1.12130247e-01
9.41056848e-01 -5.41475415e-01 5.34178652e-02 -9.54834878e-01
-7.75734663e-01 -7.43738711e-02 4.06145789e-02 6.35266304e-01
4.53015566e-01 -1.32867777e+00 -7.57577896e-01 1.46509215e-01
2.44620249e-01 -2.99806356e-01 2.27992788e-01 3.47511649e-01
-6.05958104e-01 6.40463114e-01 -6.76441491e-01 -5.94303548e-01
-1.57737887e+00 2.77293265e-01 8.29997137e-02 4.72741462e-02
-5.48271043e-03 5.45158684e-01 -2.11686984e-01 3.50199997e-01
4.60731417e-01 -2.39216775e-01 -4.91468579e-01 1.17097214e-01
5.55815399e-01 2.64295727e-01 1.76312596e-01 -8.46396506e-01
-1.81165710e-01 3.10291678e-01 2.05568597e-02 -4.31547135e-01
1.09355545e+00 -1.57364979e-01 2.18374446e-01 5.69759727e-01
1.08168948e+00 -1.69893563e-01 -8.22121203e-01 3.61486614e-01
-3.92146930e-02 -2.45343789e-01 1.40980572e-01 -6.89065993e-01
-5.95382214e-01 9.01445031e-01 8.07059348e-01 4.91707325e-01
1.20946181e+00 -1.45597294e-01 7.36231208e-01 1.89871311e-01
-2.89240349e-02 -1.30654991e+00 2.28448108e-01 3.99132490e-01
9.80461538e-01 -9.37877834e-01 -2.25958407e-01 -2.92677462e-01
-7.58978724e-01 1.17388713e+00 6.66306019e-01 1.96035445e-01
5.36571145e-01 1.13892026e-01 3.05918455e-01 1.71292007e-01
-5.33673525e-01 8.61677676e-02 5.82302928e-01 7.34064639e-01
5.88582933e-01 -2.80603133e-02 -2.37511680e-01 8.73338997e-01
-8.20083916e-01 5.73023483e-02 7.60849237e-01 7.33925521e-01
-7.28846312e-01 -9.91815567e-01 -9.42779839e-01 1.87828630e-01
-6.15157545e-01 1.95150360e-01 -5.98467469e-01 5.48230827e-01
8.29079002e-02 1.18525147e+00 1.05113335e-01 -6.32499516e-01
2.69860893e-01 4.64730203e-01 -7.94480648e-03 -5.42802572e-01
-6.50741518e-01 1.72579870e-01 1.91085339e-01 1.21302597e-01
-5.12214482e-01 -3.29318464e-01 -1.15939438e+00 1.56481549e-01
-6.88041687e-01 1.08837390e+00 9.60926950e-01 6.40333056e-01
4.46654946e-01 3.43153805e-01 7.45517373e-01 -6.74588203e-01
-3.47233772e-01 -1.34016681e+00 -3.47953141e-01 5.54324150e-01
1.96978778e-01 -4.40043747e-01 -5.94991863e-01 3.82614523e-01] | [15.288854598999023, 4.887353897094727] |
4fc7163d-7ab2-4e96-a6e5-bf1718a50089 | how-different-are-pre-trained-transformers | 2204.07233 | null | https://arxiv.org/abs/2204.07233v1 | https://arxiv.org/pdf/2204.07233v1.pdf | How Different are Pre-trained Transformers for Text Ranking? | In recent years, large pre-trained transformers have led to substantial gains in performance over traditional retrieval models and feedback approaches. However, these results are primarily based on the MS Marco/TREC Deep Learning Track setup, with its very particular setup, and our understanding of why and how these models work better is fragmented at best. We analyze effective BERT-based cross-encoders versus traditional BM25 ranking for the passage retrieval task where the largest gains have been observed, and investigate two main questions. On the one hand, what is similar? To what extent does the neural ranker already encompass the capacity of traditional rankers? Is the gain in performance due to a better ranking of the same documents (prioritizing precision)? On the other hand, what is different? Can it retrieve effectively documents missed by traditional systems (prioritizing recall)? We discover substantial differences in the notion of relevance identifying strengths and weaknesses of BERT that may inspire research for future improvement. Our results contribute to our understanding of (black-box) neural rankers relative to (well-understood) traditional rankers, help understand the particular experimental setting of MS-Marco-based test collections. | ['Jaap Kamps', 'David Rau'] | 2022-04-05 | null | null | null | null | ['passage-retrieval'] | ['natural-language-processing'] | [ 4.10852283e-02 -7.85228238e-02 -1.70723230e-01 -2.53934413e-01
-1.18546903e+00 -6.14684165e-01 8.80060375e-01 5.50735474e-01
-8.23814929e-01 6.91806734e-01 8.70584846e-01 -4.52426046e-01
-8.84007156e-01 -5.97391307e-01 -6.41937613e-01 -5.32918036e-01
-4.45709825e-01 8.03084314e-01 4.03962582e-01 -7.38116980e-01
5.87567866e-01 1.49926022e-01 -1.59314704e+00 5.89164257e-01
4.55665946e-01 8.26880574e-01 -2.30674855e-02 8.02702844e-01
1.49162561e-01 8.63575459e-01 -7.50460267e-01 -4.19556051e-01
2.26550773e-01 -1.84533656e-01 -1.13489044e+00 -6.10183835e-01
6.73395276e-01 -6.81657553e-01 -5.62480509e-01 7.12163329e-01
5.58253169e-01 2.75162876e-01 7.44506180e-01 -5.84197879e-01
-9.66297925e-01 8.69631588e-01 -1.80110008e-01 8.24235499e-01
8.54454711e-02 4.08339389e-02 1.53471053e+00 -5.19013882e-01
4.99949664e-01 1.21241343e+00 4.41973746e-01 3.74866337e-01
-1.02645993e+00 -2.87831903e-01 -6.28454238e-03 2.58467704e-01
-1.18898201e+00 -5.24729311e-01 1.03302860e-04 -2.18110383e-01
1.05380750e+00 2.57596344e-01 3.71092290e-01 7.80009925e-01
1.62699044e-01 1.02337766e+00 8.66767287e-01 -3.61034721e-01
2.71839518e-02 -4.33903150e-02 4.55839783e-01 2.72137672e-01
4.23513174e-01 3.38078529e-01 -6.50874376e-01 -9.40660611e-02
4.69691992e-01 -2.06662312e-01 -5.53720355e-01 -1.16650335e-01
-1.15448070e+00 7.77713418e-01 6.15462244e-01 5.43738246e-01
-2.30593726e-01 2.63658345e-01 5.75847566e-01 8.05370033e-01
3.83471131e-01 1.10646081e+00 -5.90268075e-01 -4.35204059e-01
-1.24031305e+00 6.05174601e-01 7.50428557e-01 5.46653926e-01
3.92386973e-01 -1.66217431e-01 -4.64912891e-01 6.88066125e-01
1.21797517e-01 4.01438504e-01 6.19708419e-01 -7.40947068e-01
4.86374289e-01 2.11751387e-01 2.83362478e-01 -7.61522949e-01
-4.07453209e-01 -8.99573445e-01 -1.90326527e-01 3.45452391e-02
5.48998475e-01 -5.40837529e-04 -7.42066205e-01 1.66805863e+00
-5.50988019e-01 -3.15902531e-01 -1.40840173e-01 1.07889330e+00
5.74900389e-01 4.52055663e-01 4.01109569e-02 2.07375243e-01
1.27633297e+00 -6.78937316e-01 -3.82496715e-01 -5.07498920e-01
8.49154532e-01 -7.76313663e-01 1.09390116e+00 1.38362974e-01
-1.16154194e+00 -3.82041454e-01 -1.28318346e+00 -2.40798980e-01
-3.53156298e-01 -9.97626036e-02 7.69533753e-01 4.49935347e-01
-1.43505466e+00 9.44878995e-01 -6.37554526e-01 -3.99460524e-01
-7.50765726e-02 4.41340864e-01 -2.17568129e-02 -1.59653619e-01
-1.71605814e+00 1.11877000e+00 2.02957615e-01 2.55960915e-02
-9.27028477e-01 -6.05891526e-01 -3.76789033e-01 5.22796035e-01
3.32488358e-01 -5.98498583e-01 1.56961346e+00 -8.26921582e-01
-9.09517705e-01 5.50310135e-01 4.25965227e-02 -6.57341957e-01
3.17798436e-01 -5.06995201e-01 -1.95587784e-01 1.79072663e-01
5.70207741e-03 5.10158241e-01 1.64175350e-02 -8.90013099e-01
-6.79966509e-01 -3.53560388e-01 4.32557076e-01 5.76449096e-01
-4.01057422e-01 2.29276896e-01 -4.47978437e-01 -4.81229335e-01
-1.00666337e-01 -7.47274816e-01 1.66432232e-01 -4.50432807e-01
2.56933749e-01 -3.95072281e-01 1.63854197e-01 -5.44427574e-01
1.38318491e+00 -1.94572687e+00 -2.09646404e-01 6.83021322e-02
3.75468522e-01 4.33921784e-01 -5.12125492e-01 8.68254244e-01
1.53027192e-01 3.91781777e-01 3.76379877e-01 2.79100724e-02
7.48551488e-02 -2.66082376e-01 -5.63337862e-01 2.30982646e-01
9.50864553e-02 1.01742625e+00 -1.08057249e+00 -3.34857881e-01
-2.10874766e-01 4.56737667e-01 -3.01697046e-01 -1.28699183e-01
-2.21460581e-01 -1.46799579e-01 -5.53347945e-01 2.96117246e-01
1.97550669e-01 -4.78687555e-01 -4.57232408e-02 1.42551705e-01
4.47518341e-02 1.09930146e+00 -7.40727782e-01 1.28723323e+00
-7.53021538e-02 1.11375570e+00 -1.37464672e-01 -8.74444127e-01
5.04240692e-01 3.77880782e-01 2.66572922e-01 -1.02630317e+00
-2.68935524e-02 3.72135222e-01 3.49281251e-01 -2.08402678e-01
1.09230006e+00 -5.39835766e-02 4.75747764e-01 7.93893874e-01
-5.23523055e-02 2.26012424e-01 3.77267510e-01 5.05488753e-01
1.48581421e+00 -1.26649275e-01 -2.13778839e-01 -3.27201694e-01
-8.13248232e-02 9.61591899e-02 1.24474980e-01 1.31620204e+00
-3.11104715e-01 7.96325803e-01 5.02610207e-01 -1.20720752e-01
-9.31888342e-01 -9.26905274e-01 -2.65402883e-01 1.49433088e+00
4.94283624e-02 -4.45479512e-01 -3.31133157e-01 -4.41695452e-01
9.69772488e-02 5.00229478e-01 -8.07626009e-01 -2.68938273e-01
-5.11292160e-01 -6.84694648e-01 7.17493117e-01 7.81204462e-01
5.67486882e-02 -1.06285751e+00 -4.69039738e-01 2.08888039e-01
-1.92887321e-01 -6.89242661e-01 -4.07954127e-01 3.66465598e-01
-1.12023103e+00 -9.11139965e-01 -1.09390235e+00 -4.12499309e-01
1.65857837e-01 6.78245068e-01 1.65962601e+00 5.20178258e-01
2.45143622e-01 4.24413025e-01 -5.30288160e-01 -4.02695566e-01
-1.87212110e-01 4.51810628e-01 -8.46644342e-02 -6.10007405e-01
7.34478295e-01 -1.80002511e-01 -1.11499071e+00 3.13833296e-01
-1.04641962e+00 -5.14743745e-01 1.08708477e+00 8.13582242e-01
-2.77556665e-02 -3.04545872e-02 7.59358943e-01 -8.97103190e-01
1.22813582e+00 -4.70545352e-01 -3.79929930e-01 3.45289588e-01
-1.07419550e+00 3.84329408e-01 8.76213238e-02 -1.73483133e-01
-7.92304397e-01 -1.00316489e+00 -1.76842473e-02 3.76836807e-02
3.49213362e-01 9.80066001e-01 5.31254053e-01 2.85939693e-01
1.00799382e+00 -4.68193181e-02 -1.20136440e-01 -5.43241382e-01
2.02551156e-01 6.74287081e-01 2.40236789e-01 -8.64070237e-01
3.87501329e-01 1.38145700e-01 -4.17079985e-01 -4.09321338e-01
-8.36447954e-01 -8.43546093e-01 -1.14056908e-01 9.01141763e-02
5.62129557e-01 -9.92871940e-01 -5.16388834e-01 -5.19514978e-02
-8.03547859e-01 -4.63377655e-01 -2.80993819e-01 5.74230790e-01
-2.60734379e-01 3.01870912e-01 -1.06690288e+00 -7.43339360e-01
-5.48931420e-01 -1.10451281e+00 1.08493841e+00 1.99247733e-01
-2.86058456e-01 -9.65885341e-01 2.42923051e-01 4.52427626e-01
9.99628782e-01 -4.54476058e-01 1.12376416e+00 -8.92798781e-01
-7.28052557e-01 -4.58654583e-01 -5.29037535e-01 3.13181996e-01
-3.63860905e-01 -2.09893107e-01 -9.23101544e-01 -6.02523923e-01
-3.94266009e-01 -6.17005169e-01 1.23294401e+00 3.45134676e-01
5.73866904e-01 7.49257281e-02 -1.96661487e-01 -8.68148059e-02
1.26688826e+00 5.77821173e-02 6.14180565e-01 7.29745924e-01
5.99817783e-02 5.84830523e-01 4.69541281e-01 1.31841851e-02
3.02631497e-01 7.03886926e-01 1.71009913e-01 2.76180446e-01
-2.07054257e-01 -1.70304671e-01 3.56207132e-01 7.73691893e-01
-1.98148236e-01 -4.71356124e-01 -8.68706346e-01 8.05858970e-01
-1.79759407e+00 -1.02245486e+00 2.11724937e-01 2.54471159e+00
9.51181829e-01 5.40736437e-01 -3.57855158e-03 8.12760368e-02
3.74745756e-01 3.79132837e-01 -3.09659421e-01 -2.98002094e-01
-1.71561554e-01 3.36969554e-01 5.18649995e-01 3.81632626e-01
-8.37175548e-01 5.60871422e-01 7.04618359e+00 7.49148726e-01
-1.06682205e+00 -3.40432972e-02 5.55051029e-01 -4.66612607e-01
-3.56347471e-01 6.80365488e-02 -8.38300884e-01 9.47250128e-02
1.26990211e+00 -2.69537300e-01 4.57509458e-01 5.35696805e-01
-7.02694952e-02 1.14971280e-01 -1.43956804e+00 4.04036462e-01
2.24257056e-02 -1.19661784e+00 -6.05854951e-03 2.90118515e-01
5.56971610e-01 5.81449330e-01 2.74186850e-01 9.18813050e-01
5.28609455e-01 -1.10821557e+00 6.71609819e-01 4.93059486e-01
4.41362977e-01 -3.68731230e-01 1.09761524e+00 1.54425502e-01
-6.19794905e-01 1.45247057e-02 -6.16426587e-01 -1.44894153e-01
-2.04165637e-01 5.00911117e-01 -8.33879828e-01 4.56666976e-01
5.71417451e-01 2.26869345e-01 -9.34181690e-01 1.38673234e+00
-1.70752462e-02 8.90474677e-01 -2.36379743e-01 -3.73002559e-01
5.66432416e-01 3.97895455e-01 4.83058333e-01 1.24243140e+00
-9.88551136e-03 -2.56180286e-01 -1.86866149e-01 4.43776697e-01
-2.55547374e-01 1.03677444e-01 -1.93305075e-01 -3.20595175e-01
4.15325105e-01 9.03236330e-01 -4.72948641e-01 -3.36687744e-01
-4.05333191e-01 4.24682885e-01 4.67649072e-01 5.06690741e-01
-3.22900176e-01 -4.65721369e-01 3.65730613e-01 4.12796378e-01
2.86541581e-02 -6.58375025e-02 -1.22364208e-01 -1.08072805e+00
1.00670243e-02 -1.03450382e+00 5.35533071e-01 -7.89725602e-01
-1.13680577e+00 4.52995747e-01 -2.34431378e-03 -7.62375474e-01
-3.70343298e-01 -5.60911536e-01 -2.77618676e-01 9.67164636e-01
-1.57186151e+00 -5.22105992e-01 1.03977479e-01 7.41036534e-02
4.96883959e-01 -5.22409007e-02 6.76476002e-01 5.32011092e-01
6.81215599e-02 7.48292744e-01 6.67548120e-01 2.26372585e-01
1.05748963e+00 -1.50734365e+00 3.65449667e-01 5.56158602e-01
3.68174583e-01 1.07489896e+00 6.65014744e-01 -4.08266306e-01
-1.22392666e+00 -5.48843503e-01 1.00740576e+00 -8.44937861e-01
6.93753719e-01 1.27288222e-01 -9.68643844e-01 6.12596393e-01
2.46443808e-01 -5.60060620e-01 5.55023491e-01 9.34766352e-01
-4.70453799e-01 -1.90785658e-02 -5.64707398e-01 6.80280924e-01
6.88107789e-01 -7.30270684e-01 -7.87150919e-01 3.18920493e-01
7.60511875e-01 -1.57934874e-01 -6.75763309e-01 3.74211937e-01
8.14642727e-01 -8.99463236e-01 9.62183475e-01 -8.13134253e-01
7.59044588e-01 -2.97206286e-02 -1.56736091e-01 -1.23685873e+00
-5.24414539e-01 -1.47660494e-01 2.24583689e-03 9.46669161e-01
6.24984622e-01 -2.90529966e-01 7.96779513e-01 7.87475526e-01
-1.73197895e-01 -9.97432768e-01 -6.52767241e-01 -6.37830436e-01
6.38732255e-01 -3.29998970e-01 4.90275532e-01 6.88819289e-01
4.81837913e-02 5.45371830e-01 -3.61371376e-02 -3.05067599e-01
2.53660560e-01 -2.16626719e-01 4.10512596e-01 -1.26267731e+00
-5.31240404e-01 -8.57331097e-01 -2.90932208e-01 -1.34860623e+00
-3.27570438e-01 -8.42580855e-01 -2.02850997e-02 -1.60768414e+00
4.07358915e-01 -3.90290827e-01 -9.50254560e-01 3.51978004e-01
-4.32930678e-01 2.03915283e-01 2.72729754e-01 5.00445187e-01
-9.81018186e-01 2.30376050e-01 1.04646301e+00 -2.29189575e-01
-3.25961001e-02 2.27380078e-02 -1.26652050e+00 1.59182832e-01
4.18888330e-01 -3.16177636e-01 -5.69178045e-01 -9.88292456e-01
1.03323030e+00 1.35000274e-01 3.74003388e-02 -9.08718586e-01
3.52153480e-01 1.75173178e-01 2.70012170e-01 -3.68517309e-01
1.78628594e-01 -3.10420513e-01 -4.78468716e-01 2.70642579e-01
-1.03965199e+00 2.96968341e-01 2.76056617e-01 6.38191402e-01
-3.18657488e-01 -5.56754112e-01 3.79889339e-01 -2.45918259e-01
-5.64820170e-01 -1.07042659e-02 -4.52532798e-01 3.29193234e-01
2.21212413e-02 1.44724160e-01 -7.37701595e-01 -8.89300585e-01
-3.27497095e-01 3.15169543e-01 2.38777384e-01 5.89152217e-01
2.79850990e-01 -1.12344170e+00 -9.92564440e-01 -4.65218246e-01
3.53650600e-01 -3.53140920e-01 -2.61073671e-02 6.56054199e-01
-3.94748777e-01 1.13547730e+00 1.38196424e-01 -2.02405587e-01
-9.98284876e-01 1.89889222e-01 3.07717592e-01 -8.25665116e-01
-3.87502462e-01 8.48983765e-01 2.50624180e-01 -2.10573170e-02
4.75891054e-01 -1.21902771e-01 -2.81960905e-01 1.39560729e-01
7.14396536e-01 3.41476172e-01 5.59250891e-01 -2.42566273e-01
-6.94423467e-02 2.06518099e-02 -7.28873372e-01 -4.19375628e-01
1.29848766e+00 -1.24137178e-02 1.53337687e-01 3.24830055e-01
1.14367628e+00 2.96874382e-02 -6.83801591e-01 -4.65650469e-01
2.49452919e-01 -2.76473910e-01 3.52395564e-01 -1.29315078e+00
-8.01668823e-01 1.01005948e+00 6.82453811e-01 2.22417310e-01
8.27680707e-01 -1.58278584e-01 6.61035359e-01 7.69314766e-01
3.59189957e-01 -1.12831831e+00 1.71787925e-02 6.87669277e-01
8.22350562e-01 -1.13362360e+00 1.82414457e-01 5.40131629e-01
-5.15545428e-01 9.93807733e-01 2.86254823e-01 -3.64155136e-03
3.35474759e-01 -3.33968610e-01 5.54744080e-02 -5.47268271e-01
-1.07520282e+00 -3.76712888e-01 4.35723245e-01 1.73221603e-01
1.02332211e+00 -2.73253731e-02 -5.54412723e-01 2.58989245e-01
-4.10809934e-01 5.90801314e-02 3.86783600e-01 8.40049684e-01
-7.75782406e-01 -9.13025141e-01 -2.41310060e-01 7.31940389e-01
-8.67553592e-01 -4.96006042e-01 -4.45646316e-01 8.06151450e-01
-4.21258777e-01 1.17793512e+00 6.54023215e-02 -4.54043210e-01
1.47452414e-01 2.51438290e-01 3.16680640e-01 -7.03147173e-01
-8.53730917e-01 -3.86811942e-02 3.15611959e-01 -2.62527525e-01
-4.73064780e-02 -6.07230842e-01 -8.16871762e-01 -4.51368481e-01
-5.84031165e-01 6.07947409e-01 6.38796866e-01 9.41746891e-01
3.96599203e-01 4.84975994e-01 1.03440389e-01 -4.19299722e-01
-1.14145553e+00 -1.18406355e+00 -5.13522446e-01 4.45742339e-01
4.36988503e-01 -4.74902958e-01 -5.29019296e-01 -5.91322184e-01] | [11.455028533935547, 7.614954948425293] |
a24abc04-5f7a-4593-82d4-37d688cb1f81 | monitoring-machine-learning-ml-based-risk | 2211.09781 | null | https://arxiv.org/abs/2211.09781v2 | https://arxiv.org/pdf/2211.09781v2.pdf | Monitoring machine learning (ML)-based risk prediction algorithms in the presence of confounding medical interventions | Performance monitoring of machine learning (ML)-based risk prediction models in healthcare is complicated by the issue of confounding medical interventions (CMI): when an algorithm predicts a patient to be at high risk for an adverse event, clinicians are more likely to administer prophylactic treatment and alter the very target that the algorithm aims to predict. A simple approach is to ignore CMI and monitor only the untreated patients, whose outcomes remain unaltered. In general, ignoring CMI may inflate Type I error because (i) untreated patients disproportionally represent those with low predicted risk and (ii) evolution in both the model and clinician trust in the model can induce complex dependencies that violate standard assumptions. Nevertheless, we show that valid inference is still possible if one monitors conditional performance and if either conditional exchangeability or time-constant selection bias hold. Specifically, we develop a new score-based cumulative sum (CUSUM) monitoring procedure with dynamic control limits. Through simulations, we demonstrate the benefits of combining model updating with monitoring and investigate how over-trust in a prediction model may delay detection of performance deterioration. Finally, we illustrate how these monitoring methods can be used to detect calibration decay of an ML-based risk calculator for postoperative nausea and vomiting during the COVID-19 pandemic. | ['Romain Pirracchio', 'Berkman Sahiner', 'Nicholas Petrick', 'Gene Pennello', 'Alexej Gossmann', 'Jean Feng'] | 2022-11-17 | null | null | null | null | ['selection-bias'] | ['natural-language-processing'] | [ 4.46742326e-01 4.09131289e-01 -5.81804335e-01 -1.68817177e-01
-4.18595433e-01 -4.60998356e-01 1.49299875e-01 9.85268414e-01
-5.54200709e-01 7.67047524e-01 1.69349581e-01 -9.05567467e-01
-4.91259038e-01 -6.56387568e-01 -7.56075382e-01 -7.35263050e-01
-4.97704834e-01 6.92505598e-01 -1.80757985e-01 4.26448345e-01
-7.62755722e-02 3.30210090e-01 -1.01544952e+00 2.02096075e-01
9.90715981e-01 6.87499762e-01 -3.96436870e-01 7.70020545e-01
6.57720685e-01 1.19329166e+00 -6.19782209e-01 -3.23410511e-01
4.12635535e-01 -5.66051304e-01 -2.27977976e-01 -4.05171633e-01
-1.96810484e-01 -4.57541317e-01 2.40404278e-01 8.17941189e-01
3.62040609e-01 -3.01912904e-01 8.89098406e-01 -1.34170425e+00
2.47861043e-01 7.90994346e-01 -2.68283129e-01 3.50085669e-03
-1.96199194e-02 1.74330384e-01 4.33888972e-01 3.30974497e-02
1.04962461e-01 9.81003225e-01 9.96869504e-01 5.24280727e-01
-1.28837454e+00 -6.61215007e-01 3.38473171e-01 -3.19787413e-01
-9.96439517e-01 -1.52924508e-01 1.09241068e-01 -6.40815973e-01
5.22645891e-01 6.38313115e-01 7.58520007e-01 7.01669931e-01
9.92920637e-01 2.18887359e-01 9.83736515e-01 -3.27265114e-01
3.00123513e-01 3.43105227e-01 1.83000028e-01 3.70479703e-01
9.62124586e-01 8.58303607e-01 -6.92602396e-02 -1.03358269e+00
6.21495545e-01 4.86086488e-01 -2.21215874e-01 -4.07935947e-01
-1.13709891e+00 1.01099586e+00 -5.02842702e-02 -1.71047851e-01
-7.32779682e-01 1.06047206e-01 4.03783560e-01 3.76542032e-01
3.56636912e-01 4.04766202e-01 -7.54575789e-01 7.76699185e-03
-7.80959785e-01 2.41894707e-01 8.89642417e-01 3.97742897e-01
-5.71369193e-02 -2.05967322e-01 -4.61843580e-01 1.63678274e-01
2.30849296e-01 5.31343937e-01 3.73608947e-01 -1.16103351e+00
1.00889429e-01 3.89507860e-01 6.08700156e-01 -7.18460023e-01
-8.14315856e-01 -8.16804290e-01 -9.04052377e-01 2.82446712e-01
4.39587206e-01 -6.64394498e-01 -6.01241648e-01 1.89609015e+00
1.95837498e-01 -6.61436841e-02 1.40104786e-01 3.79806668e-01
-2.59451419e-01 3.15991074e-01 4.01845038e-01 -1.08277822e+00
1.00595486e+00 -3.45209867e-01 -6.24340475e-01 -3.78256850e-02
1.16120744e+00 -2.03157857e-01 3.58555645e-01 5.43326437e-01
-1.41228700e+00 5.38838021e-02 -8.03580523e-01 9.06011939e-01
2.68910348e-01 -4.79052603e-01 4.29964304e-01 7.76505351e-01
-6.24501765e-01 9.43217218e-01 -1.10535121e+00 -1.97430328e-01
2.57581323e-01 4.74955320e-01 1.91440970e-01 1.44856378e-01
-1.21353781e+00 1.09325874e+00 1.20124519e-01 -1.37415722e-01
-8.61438870e-01 -1.22735345e+00 -6.27327263e-01 2.06878901e-01
3.59091133e-01 -1.05299675e+00 1.12494326e+00 -9.89400327e-01
-7.33405709e-01 4.73900050e-01 5.62748080e-03 -8.73344779e-01
9.03771460e-01 -1.72438741e-01 -1.67004243e-01 -2.23738179e-01
-1.42299727e-01 -1.71447992e-01 5.18924236e-01 -1.14029300e+00
-7.94100404e-01 -5.78003883e-01 -3.26210886e-01 2.51362771e-01
-1.33457825e-01 1.95262790e-01 3.61199588e-01 -6.24068618e-01
-1.76305607e-01 -9.73034799e-01 -8.24969649e-01 -3.97813797e-01
-2.89214164e-01 4.09525841e-01 -1.49109349e-01 -5.06079435e-01
1.60561514e+00 -1.95856273e+00 -3.49425733e-01 5.07250071e-01
2.91930109e-01 -2.13810913e-02 3.21347713e-01 4.34759289e-01
-2.08854884e-01 1.87061533e-01 -4.85323370e-01 1.55874178e-01
-3.61649126e-01 1.21283561e-01 1.95024547e-03 6.63287222e-01
-1.55656755e-01 5.02234519e-01 -7.16855884e-01 -3.15016955e-01
1.56524450e-01 1.86135992e-01 -9.68518198e-01 3.11803669e-01
4.05143052e-02 5.35357237e-01 -1.69568434e-01 1.90650538e-01
4.40702379e-01 -4.31747913e-01 6.53953910e-01 2.81754702e-01
-7.69136474e-02 1.83708087e-01 -1.10867941e+00 3.23176175e-01
-3.15451413e-01 -1.44214079e-01 3.62628400e-02 -1.10984731e+00
2.11237341e-01 5.06785870e-01 7.35903621e-01 -1.64098218e-01
5.49658649e-02 4.12364155e-02 5.23679912e-01 -4.24167693e-01
-4.24575537e-01 -7.57895529e-01 7.19073042e-02 6.43992901e-01
-7.87872672e-01 1.69085890e-01 -5.12869775e-01 9.62494314e-03
1.19782686e+00 -4.69271243e-01 7.98327267e-01 -4.56668496e-01
4.56298366e-02 3.97197157e-02 1.04969144e+00 1.05551159e+00
-2.40761146e-01 2.00551525e-01 8.04333031e-01 -2.54009604e-01
-6.52273118e-01 -1.18239439e+00 -4.82947499e-01 6.44224644e-01
-4.09149647e-01 3.23789157e-02 -4.32227135e-01 -5.64572752e-01
3.92899692e-01 9.52121735e-01 -1.03169763e+00 -5.30922413e-01
-3.99418682e-01 -1.40925145e+00 2.10317403e-01 5.39774418e-01
-3.48850459e-01 -5.10086596e-01 -1.23355138e+00 5.38651347e-01
3.23565006e-01 -1.91388145e-01 -2.85833150e-01 3.48571777e-01
-1.11568391e+00 -1.51539111e+00 -6.30010366e-01 -1.58074945e-01
7.13224769e-01 -1.63051188e-01 1.10705113e+00 3.57287943e-01
1.17519692e-01 4.59758490e-01 6.90325573e-02 -9.86053824e-01
-8.49141657e-01 -4.86683339e-01 2.99223900e-01 -2.28518516e-01
2.13278472e-01 -1.46533102e-01 -9.82484281e-01 1.82881191e-01
-7.04469383e-01 -8.00946131e-02 3.31331432e-01 9.19827759e-01
3.61378103e-01 -5.49476482e-02 8.02512825e-01 -1.47169173e+00
5.73852777e-01 -6.77681148e-01 -7.62229621e-01 5.16453683e-01
-1.55691850e+00 -2.00017706e-01 2.31155932e-01 -6.40585959e-01
-8.44709039e-01 -1.19275134e-02 5.24345577e-01 -2.03071699e-01
2.40318984e-01 8.14818025e-01 1.43387616e-01 4.64673072e-01
5.47728419e-01 -1.29355907e-01 4.36838627e-01 -1.63313150e-01
-2.85977960e-01 5.14280558e-01 1.60306290e-01 -3.33468378e-01
5.31006694e-01 3.56158048e-01 2.57250249e-01 -2.27641910e-01
-7.84613371e-01 -1.03906214e-01 -1.06643908e-01 -1.26793250e-01
5.92598379e-01 -1.13935041e+00 -1.05174613e+00 1.75800264e-01
-6.42108977e-01 -6.35424197e-01 -2.83244461e-01 8.52423191e-01
-4.89998162e-01 8.62340108e-02 -4.94981855e-01 -1.27632082e+00
-2.58053690e-01 -1.08673537e+00 2.25204572e-01 -4.03444730e-02
-6.06565237e-01 -1.32910609e+00 2.93999553e-01 -9.33010727e-02
4.97379929e-01 6.11862838e-01 1.20924461e+00 -9.61194873e-01
-8.48260745e-02 -5.41177690e-01 2.50258952e-01 1.83791205e-01
3.77590954e-01 1.40769213e-01 -5.49228549e-01 -5.71479082e-01
4.52532202e-01 2.93695360e-01 4.25628781e-01 9.46922362e-01
1.02768195e+00 -9.51496184e-01 -5.90905428e-01 4.91338968e-01
1.16647673e+00 6.80501521e-01 1.08598329e-01 2.09418982e-01
2.69661158e-01 8.99324477e-01 6.32524252e-01 8.83659482e-01
3.67253870e-01 4.61799532e-01 4.11374152e-01 -3.09597969e-01
8.10634613e-01 -9.03117284e-02 2.32274935e-01 3.01978856e-01
4.11403514e-02 5.13569787e-02 -1.05525935e+00 3.54007095e-01
-1.77087259e+00 -8.86903882e-01 -3.62401515e-01 3.03929043e+00
1.02906418e+00 3.92336011e-01 4.11818624e-01 -2.90495995e-02
5.48998594e-01 -5.62826335e-01 -8.65855098e-01 -6.44365191e-01
1.63627177e-01 -2.10067242e-01 8.94348741e-01 6.61771119e-01
-6.39851272e-01 -1.08861700e-02 7.08387995e+00 -4.03584391e-02
-8.06019962e-01 2.40298018e-01 1.24165118e+00 -3.34715307e-01
-3.00419658e-01 2.51141004e-02 -3.68071586e-01 4.59557384e-01
1.47908342e+00 -5.36951661e-01 -7.84915462e-02 5.48985124e-01
6.09661222e-01 -1.99631855e-01 -1.38573170e+00 2.58391440e-01
-2.32333794e-01 -9.97632205e-01 -3.14730048e-01 1.83857337e-01
6.39037728e-01 -1.66116595e-01 6.21432066e-02 2.13336483e-01
6.47702038e-01 -1.05040026e+00 4.61960256e-01 6.76402450e-01
7.18319416e-01 -9.44933474e-01 9.51732934e-01 4.96286303e-01
-4.40955788e-01 -4.13644999e-01 -3.79448682e-02 -3.00629884e-01
9.79504436e-02 9.21101093e-01 -8.60794783e-01 2.15113416e-01
3.03313822e-01 2.07602859e-01 -1.19946897e-01 1.02746940e+00
2.15136841e-01 9.38259244e-01 -2.10865408e-01 3.47955972e-01
-2.32122183e-01 -6.24358095e-02 3.64794344e-01 9.11385596e-01
8.54706392e-02 3.48937511e-01 -9.00774673e-02 4.11750555e-01
4.48592722e-01 3.40428613e-02 -5.66650987e-01 1.44948706e-01
3.37320596e-01 5.23998559e-01 -3.42251152e-01 -5.82129419e-01
-4.62291092e-01 4.25744087e-01 -2.97161669e-01 2.02810898e-01
-7.08039403e-01 2.65837371e-01 8.03025663e-01 5.13959944e-01
-3.34137976e-01 4.06023085e-01 -7.49737561e-01 -9.20539796e-01
-3.25831324e-01 -8.72782171e-01 1.02134466e+00 -1.14591695e-01
-1.00974166e+00 3.70642804e-02 1.39067829e-01 -1.20526505e+00
-5.66847265e-01 -2.51589000e-01 -5.08582771e-01 9.45600986e-01
-1.37115204e+00 -2.18175203e-01 2.73436755e-01 3.25931072e-01
-4.96523716e-02 4.86403108e-01 8.06519687e-01 -9.10964012e-02
-6.94993079e-01 7.63418019e-01 2.59025723e-01 -2.07309619e-01
7.76420593e-01 -1.14349306e+00 -2.92165726e-01 5.79089999e-01
-6.94116652e-01 7.05692172e-01 8.03706884e-01 -1.05778670e+00
-9.48592007e-01 -1.24661970e+00 9.09689188e-01 -7.80101955e-01
3.87409091e-01 1.46736592e-01 -9.89129126e-01 8.43456388e-01
-3.78170133e-01 -3.06018829e-01 9.24484551e-01 -5.02707288e-02
2.14655534e-03 -6.76504970e-02 -1.33921468e+00 5.20277441e-01
5.69696009e-01 7.23461881e-02 -3.49270999e-01 5.25629878e-01
7.24589109e-01 -1.14704356e-01 -1.21821249e+00 9.46135342e-01
7.96405017e-01 -1.01520753e+00 8.33450854e-01 -1.17207468e+00
7.33501241e-02 2.06890583e-01 3.19727734e-02 -1.14827037e+00
-4.20388341e-01 -7.36312509e-01 -1.30632460e-01 6.75152838e-01
7.66252339e-01 -1.02441120e+00 5.47945142e-01 1.10624659e+00
3.19660366e-01 -7.28812993e-01 -7.59545267e-01 -7.28273451e-01
5.24989009e-01 -3.08605492e-01 5.42836845e-01 1.08091760e+00
3.85777384e-01 -2.78961867e-01 -4.32244956e-01 5.45492649e-01
8.42411399e-01 -1.78619567e-02 2.68384576e-01 -1.30705559e+00
-5.88624954e-01 -2.81387091e-01 6.59199804e-02 -1.00618348e-01
-3.94070625e-01 -3.81533831e-01 -1.14120707e-01 -1.27650452e+00
6.91570044e-01 -5.86217642e-01 -8.34335625e-01 4.21896905e-01
-6.34388983e-01 -5.23382068e-01 1.45078942e-01 2.68360496e-01
-2.82905605e-02 1.33199170e-01 8.49547982e-01 2.32981741e-01
-5.08029938e-01 7.33421445e-01 -6.59480214e-01 7.45689392e-01
7.35278904e-01 -9.18104470e-01 -4.03558075e-01 1.52028248e-01
4.20054287e-01 8.31603527e-01 3.03805947e-01 -6.63511038e-01
-1.87296923e-02 -7.07413733e-01 3.80403936e-01 -3.43297482e-01
-3.96773756e-01 -1.04801571e+00 7.08928466e-01 1.66604114e+00
-6.65877283e-01 3.35729569e-01 2.08884552e-01 6.20874286e-01
3.55930030e-01 3.87929305e-02 9.10875380e-01 -5.31225950e-02
5.94875753e-01 1.79283902e-01 -7.79387712e-01 6.59049079e-02
1.08375430e+00 2.49862015e-01 -1.11460470e-01 -4.87069845e-01
-7.85271525e-01 3.48804832e-01 5.05750299e-01 -8.42896663e-03
1.87627718e-01 -8.56969595e-01 -8.24894726e-01 4.27609161e-02
1.12554260e-01 -1.54414043e-01 2.23966882e-01 1.36212349e+00
-2.53656983e-01 3.30958247e-01 3.88081402e-01 -3.52181882e-01
-1.10663271e+00 1.01136923e+00 7.33830214e-01 -4.39280361e-01
-3.94681692e-01 5.08754492e-01 6.08685076e-01 1.58205424e-02
3.14409673e-01 -3.06243122e-01 1.38409480e-01 -2.31132954e-02
7.17670202e-01 3.95557523e-01 4.86896001e-02 -1.39643326e-01
-5.32955348e-01 3.16379406e-02 -6.33173883e-02 -4.08569677e-03
1.18228900e+00 -1.78639024e-01 -6.66934252e-02 5.93556166e-01
7.48975217e-01 -1.08376205e-01 -1.37732816e+00 8.14842153e-03
5.10223992e-02 -1.13593345e-03 -4.30962518e-02 -1.12370324e+00
-6.98443234e-01 5.71736336e-01 8.96439373e-01 2.20031649e-01
1.10819030e+00 -3.42951953e-01 1.14959784e-01 -6.70196265e-02
1.04382955e-01 -7.89630175e-01 -6.98608518e-01 -2.73093671e-01
5.78415513e-01 -1.15254390e+00 1.80646181e-01 3.63758206e-03
-6.43317401e-01 5.31370878e-01 2.39211589e-01 2.17530638e-01
1.03347862e+00 3.25128704e-01 1.72545314e-01 -3.76324840e-02
-1.21120512e+00 4.45789039e-01 -7.13610947e-02 4.48398203e-01
2.34607056e-01 6.06809139e-01 -6.30441248e-01 8.59165192e-01
1.35443121e-01 2.79593438e-01 8.75083685e-01 9.61360276e-01
-1.51994556e-01 -8.24493051e-01 -5.52508533e-01 1.07074857e+00
-7.83941209e-01 -2.14319572e-01 1.20331205e-01 7.53323436e-01
-1.19377449e-01 9.39872861e-01 1.92654982e-01 -3.03259250e-02
4.70425993e-01 2.54176587e-01 1.47646248e-01 -5.85830510e-01
-6.96350694e-01 1.75906107e-01 4.74416278e-02 -4.91031677e-01
-3.48165035e-01 -1.01877439e+00 -1.14068794e+00 -1.63276225e-01
-1.43616408e-01 3.33782881e-01 4.12246585e-01 8.61934781e-01
3.60960871e-01 5.18433332e-01 9.67525005e-01 -7.16012567e-02
-1.13795376e+00 -6.37513995e-01 -5.15066087e-01 9.70280245e-02
6.93348706e-01 -5.31661570e-01 -9.53582764e-01 -1.94187388e-01] | [8.027194023132324, 5.379754066467285] |
b394871a-c6f2-4b31-b258-622a679b0d3e | mcmia-model-compression-against-membership | 2008.13578 | null | https://arxiv.org/abs/2008.13578v4 | https://arxiv.org/pdf/2008.13578v4.pdf | Against Membership Inference Attack: Pruning is All You Need | The large model size, high computational operations, and vulnerability against membership inference attack (MIA) have impeded deep learning or deep neural networks (DNNs) popularity, especially on mobile devices. To address the challenge, we envision that the weight pruning technique will help DNNs against MIA while reducing model storage and computational operation. In this work, we propose a pruning algorithm, and we show that the proposed algorithm can find a subnetwork that can prevent privacy leakage from MIA and achieves competitive accuracy with the original DNNs. We also verify our theoretical insights with experiments. Our experimental results illustrate that the attack accuracy using model compression is up to 13.6% and 10% lower than that of the baseline and Min-Max game, accordingly. | ['Caiwen Ding', 'Jinbo Bi', 'Hang Liu', 'Yijue Wang', 'Zigeng Wang', 'Shanglin Zhou', 'Sanguthevar Rajasekaran', 'Chenghong Wang'] | 2020-08-28 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [ 9.03886929e-02 3.92555207e-01 -4.58977610e-01 -1.51759967e-01
-1.35345981e-01 -5.67667186e-01 6.27200752e-02 -1.44293025e-01
-7.23154426e-01 7.47915506e-01 -3.65716070e-01 -1.10285497e+00
-1.96652070e-01 -1.14411139e+00 -9.48238969e-01 -3.94920468e-01
-1.61155552e-01 -1.74823701e-02 3.60352337e-01 1.18160598e-01
-1.26595512e-01 5.79313934e-01 -1.12951946e+00 -1.27414033e-01
7.86202848e-01 1.40372086e+00 -4.92082864e-01 1.66341826e-01
2.61473328e-01 7.25984156e-01 -8.70772421e-01 -1.18671155e+00
5.88039041e-01 1.24928392e-01 -6.37438416e-01 -7.76003838e-01
5.50575316e-01 -9.38440382e-01 -1.07098246e+00 1.53361404e+00
4.23331261e-01 -2.63099253e-01 2.63779193e-01 -1.77361596e+00
-1.38965607e-01 1.13743591e+00 -4.25750315e-01 3.01451143e-02
-4.35403079e-01 -2.37477735e-01 6.98063135e-01 -1.96132362e-01
2.94314176e-01 1.10006881e+00 9.59443748e-01 1.05200481e+00
-8.80561590e-01 -1.67147303e+00 1.74466327e-01 2.85496175e-01
-1.63208950e+00 -5.96908867e-01 4.99834925e-01 1.19708709e-01
9.13192272e-01 4.71933633e-01 6.19100630e-01 1.27898824e+00
-1.54556319e-01 7.60840118e-01 5.15686154e-01 -1.39646968e-02
4.37369376e-01 1.57712817e-01 3.95132810e-01 9.05394495e-01
1.14100635e+00 3.41712348e-02 -3.23449910e-01 -6.04096353e-01
5.88906169e-01 1.84836790e-01 -1.61724344e-01 -1.36111304e-01
-1.69324949e-01 8.31736267e-01 1.62114546e-01 -2.98819065e-01
6.55857101e-02 6.11078918e-01 6.18829310e-01 4.13369387e-01
-2.30046082e-02 -3.29623595e-02 -7.16736317e-01 -1.86267346e-01
-8.27445447e-01 2.86355466e-01 1.16343975e+00 1.16424131e+00
2.57449389e-01 1.73200846e-01 4.00745124e-01 1.83857903e-01
4.99100536e-01 4.13938522e-01 1.70183226e-01 -1.01986933e+00
6.39717340e-01 3.98459315e-01 -3.93264562e-01 -1.07721031e+00
-2.18948990e-01 -5.83430588e-01 -1.38429737e+00 2.43051443e-02
4.89972740e-01 -7.11841822e-01 -4.99829620e-01 2.04072142e+00
2.16043785e-01 7.55258024e-01 3.21207225e-01 2.82879949e-01
5.62029004e-01 2.99995214e-01 1.99783117e-01 -1.43414304e-01
1.16725075e+00 -5.77234387e-01 -5.44764459e-01 -7.63113797e-02
1.09245169e+00 -5.78278787e-02 5.25839567e-01 4.19146597e-01
-1.13580179e+00 2.67410241e-02 -1.42731643e+00 2.45649368e-02
-2.75763810e-01 -6.56429604e-02 8.85650635e-01 1.47548592e+00
-8.03354144e-01 3.77136648e-01 -1.02765596e+00 4.89531495e-02
1.42831612e+00 7.87289858e-01 -1.45524904e-01 1.41432151e-01
-1.50352931e+00 3.94661874e-01 5.53148866e-01 -2.53267109e-01
-9.09697294e-01 -1.06451285e+00 -6.67895973e-01 4.71985966e-01
3.12119246e-01 -6.09932125e-01 1.04592991e+00 -1.48201734e-01
-1.27434516e+00 6.05078161e-01 1.54870138e-01 -1.62525368e+00
6.10927105e-01 -1.90872088e-01 -3.75115305e-01 1.46608770e-01
-7.96253681e-01 4.94955391e-01 4.03287947e-01 -8.74526083e-01
-7.82521904e-01 -4.77710247e-01 4.99816328e-01 -4.63344544e-01
-1.03429806e+00 -6.73773438e-02 -1.66939482e-01 -5.79212546e-01
-8.21150914e-02 -8.38113248e-01 -4.14130837e-01 4.15433586e-01
-5.51456392e-01 -6.38658181e-02 1.31570554e+00 -3.91778439e-01
1.65033853e+00 -2.08185720e+00 -6.43599868e-01 7.70584643e-01
6.67375624e-01 8.99454057e-01 1.66058496e-01 -1.62653118e-01
3.17204058e-01 4.86597925e-01 1.73241626e-02 -3.28417152e-01
1.23874642e-01 2.61485070e-01 -6.21420741e-01 4.30110037e-01
-5.02257705e-01 7.74134755e-01 -3.66568863e-01 -3.39003861e-01
-2.70432740e-01 4.77081388e-01 -9.91492152e-01 -3.12930465e-01
-3.06882858e-02 -3.69914472e-01 -3.65320444e-01 5.95901072e-01
1.27973843e+00 -1.27132714e-01 5.46158433e-01 -2.03872457e-01
6.19526148e-01 3.94890040e-01 -9.37497854e-01 1.07297528e+00
-8.40285644e-02 6.56269431e-01 1.43834427e-01 -1.02483428e+00
8.80264759e-01 2.16217592e-01 2.22491995e-01 -5.24868011e-01
4.82282937e-01 9.64378119e-02 9.48223472e-02 -6.18231557e-02
3.22400848e-03 3.39257300e-01 5.83129525e-02 4.05105114e-01
-1.24126524e-01 7.54744947e-01 -4.02434468e-01 2.28127763e-01
1.16118288e+00 -8.44212294e-01 3.76613140e-02 -2.41795138e-01
3.79246235e-01 -6.00048184e-01 8.97261858e-01 1.07166123e+00
-6.46483243e-01 -1.61860064e-01 9.22008574e-01 -4.39403206e-01
-6.80449367e-01 -9.73934650e-01 -2.48191714e-01 6.66454434e-01
2.83627898e-01 -8.12008023e-01 -1.22976542e+00 -8.45238507e-01
1.29800260e-01 4.98802841e-01 -2.88180411e-01 -5.58898449e-01
-4.85323459e-01 -7.12286413e-01 1.60910594e+00 5.79395711e-01
1.14153874e+00 -2.52492040e-01 -5.11293650e-01 -1.59110680e-01
8.75954628e-02 -1.33069158e+00 -7.63230622e-02 4.06055152e-02
-1.05576634e+00 -1.08695972e+00 -2.29058877e-01 -8.18536103e-01
5.69071889e-01 1.05692372e-01 5.43374956e-01 2.70762652e-01
1.45733461e-01 -2.99751103e-01 3.15054476e-01 -7.17131972e-01
-3.06757331e-01 4.65526462e-01 4.77081716e-01 -3.13693434e-01
7.09537506e-01 -1.05172122e+00 -5.31584084e-01 9.67939049e-02
-8.22920263e-01 -1.47666842e-01 2.46576190e-01 5.03781676e-01
3.47701371e-01 4.98938650e-01 6.52549207e-01 -1.16226137e+00
5.53772390e-01 -3.88739794e-01 -1.01640797e+00 2.97232091e-01
-8.16041648e-01 -2.08959118e-01 7.05555797e-01 -7.35059500e-01
-4.13790256e-01 3.92799340e-02 -3.63546252e-01 -8.11044097e-01
9.36316848e-02 3.18406940e-01 -7.43854761e-01 -6.48062706e-01
4.85990316e-01 -3.01927514e-02 3.12583223e-02 -3.92212331e-01
9.80239138e-02 6.25931859e-01 5.17451227e-01 -3.97242874e-01
9.41581666e-01 4.14268792e-01 3.75328273e-01 -5.94651103e-01
-4.89771068e-01 3.24146301e-01 -5.41136786e-02 1.80742756e-01
3.37099642e-01 -1.13200057e+00 -1.26492596e+00 7.18852997e-01
-1.04486918e+00 -2.25254491e-01 6.15458377e-02 3.10834676e-01
2.97763105e-02 4.89821583e-01 -7.81185448e-01 -8.29161167e-01
-8.25257361e-01 -9.30345833e-01 3.61351445e-02 4.81789798e-01
4.46880683e-02 -9.02028859e-01 -6.62250042e-01 4.37641561e-01
3.63853753e-01 2.20368192e-01 1.03678095e+00 -1.09451616e+00
-6.40094101e-01 -3.48178029e-01 -3.45175892e-01 2.99765885e-01
-3.33841890e-01 -2.26655722e-01 -1.22067630e+00 -4.44135040e-01
6.34004623e-02 -2.89012641e-01 8.13268125e-01 2.55752206e-01
2.03454804e+00 -9.94314611e-01 -4.63766038e-01 1.19813848e+00
1.19289196e+00 2.79383063e-01 8.81006122e-01 2.98617572e-01
7.91450441e-01 -1.36824809e-02 1.36776164e-01 5.36438406e-01
2.45266184e-01 2.25286603e-01 6.84623659e-01 1.12036511e-01
3.66342485e-01 -5.83425164e-01 1.63324624e-01 4.86543477e-01
2.24902123e-01 -2.93490529e-01 -6.60343289e-01 2.59570658e-01
-1.75006342e+00 -8.78177881e-01 1.72569707e-01 2.05155420e+00
9.84681368e-01 6.90749228e-01 9.39327925e-02 5.11692703e-01
2.78892428e-01 3.17793004e-02 -7.77548552e-01 -4.41455573e-01
-1.02155544e-01 3.10653955e-01 1.24138546e+00 4.25837219e-01
-1.15938306e+00 9.77244020e-01 6.64518261e+00 1.23664057e+00
-8.69662583e-01 8.68699476e-02 1.13508642e+00 -4.37372714e-01
-1.78081915e-01 -3.71039480e-01 -1.06737447e+00 3.92178148e-01
1.33126068e+00 -5.38827598e-01 3.72585714e-01 1.19384742e+00
-2.98495561e-01 6.11322284e-01 -9.12049651e-01 1.04948199e+00
-2.86269426e-01 -1.64879143e+00 2.26989672e-01 5.08008301e-01
5.37464797e-01 -7.12022558e-03 3.50782543e-01 2.97216505e-01
3.59275907e-01 -1.06700766e+00 4.31327313e-01 1.40833497e-01
8.69613349e-01 -1.65491247e+00 7.30690598e-01 5.42625070e-01
-8.91056895e-01 -4.11765963e-01 -8.35533023e-01 -7.88133740e-02
-2.91836649e-01 3.45336348e-01 -5.74648082e-01 1.15883641e-01
6.72400475e-01 3.30598533e-01 -4.63245213e-01 9.51490402e-01
6.42741844e-02 9.72009838e-01 -7.30087757e-01 -2.18211010e-01
2.34834496e-02 7.41537847e-03 4.33890045e-01 9.49068069e-01
2.18788475e-01 2.16377303e-01 -3.24119717e-01 7.93358982e-01
-8.87919843e-01 -3.86693031e-01 -6.85511529e-01 2.42398903e-02
1.17824519e+00 7.92290211e-01 -2.34900668e-01 -3.65204453e-01
-1.65308252e-01 6.11907780e-01 2.46557444e-01 2.68803537e-01
-1.18277538e+00 -7.26959705e-01 1.36175215e+00 1.02247475e-02
1.69081017e-01 1.46230860e-02 -7.19833374e-01 -1.05067337e+00
9.38035995e-02 -9.15668130e-01 4.02778685e-01 6.31000251e-02
-9.61695969e-01 3.58760327e-01 -1.68906465e-01 -8.31620395e-01
2.42591202e-01 -5.13541102e-01 -6.36289656e-01 4.07141894e-01
-1.22254956e+00 -8.42989445e-01 2.69700617e-01 6.85119212e-01
-3.08427662e-01 -5.12489021e-01 8.47725093e-01 7.26640105e-01
-1.00367522e+00 2.09816647e+00 2.74364561e-01 6.67698324e-01
-4.39312346e-02 -5.85849881e-01 6.33199215e-01 1.07481170e+00
5.97262159e-02 8.92922401e-01 3.61780733e-01 -4.03509080e-01
-1.34764576e+00 -1.38707018e+00 8.10695231e-01 4.77781855e-02
5.62880814e-01 -7.16303051e-01 -8.22038352e-01 7.78582215e-01
-2.28028864e-01 8.52546766e-02 9.72705543e-01 -4.83970456e-02
-6.36370122e-01 -4.42980170e-01 -1.79016149e+00 9.70878541e-01
1.31096864e+00 -6.67683244e-01 1.24535352e-01 -3.24991792e-02
1.09731936e+00 -4.63713527e-01 -7.80927122e-01 4.19511557e-01
7.56405354e-01 -8.01434457e-01 1.23261368e+00 -9.05236661e-01
-2.06075132e-01 2.22840920e-01 -2.16644660e-01 -4.26482350e-01
7.10870102e-02 -1.06411099e+00 -1.23378384e+00 1.20776534e+00
5.45275867e-01 -9.55231130e-01 1.53994942e+00 1.01923513e+00
4.76061881e-01 -9.58891988e-01 -1.31490827e+00 -8.34858477e-01
3.08350265e-01 -6.90343618e-01 1.12769687e+00 9.90682006e-01
-1.39491156e-01 -1.29288003e-01 -3.90875190e-01 5.62427640e-01
8.44664633e-01 -4.45379704e-01 6.55110419e-01 -1.42690325e+00
-1.01433553e-01 -5.18601477e-01 -6.63048029e-01 -1.19505644e+00
3.22848707e-01 -7.74788678e-01 -8.55179548e-01 -7.51735747e-01
-5.55534177e-02 -4.52683806e-01 -6.43457651e-01 7.66832888e-01
3.82448047e-01 2.00394794e-01 2.32869059e-01 -2.10345536e-01
-5.76169014e-01 2.77155101e-01 6.86564744e-01 -2.13368803e-01
1.87398195e-02 3.35578352e-01 -1.02814972e+00 9.58300889e-01
1.20629239e+00 -6.71620846e-01 -7.74721682e-01 -2.73056179e-01
1.31466150e-01 -2.09295660e-01 3.24197531e-01 -1.28977859e+00
3.45923483e-01 5.22904433e-02 2.72246480e-01 -4.75209028e-01
1.54835373e-01 -1.23241556e+00 2.29911786e-02 1.01156294e+00
-4.18331742e-01 -3.22450191e-01 3.78623456e-01 6.91456795e-01
3.38119596e-01 -6.41239733e-02 8.77705753e-01 5.10671556e-01
-1.37256190e-01 7.74498045e-01 -4.59374219e-01 -8.10888037e-02
1.06745899e+00 -1.98037431e-01 -5.23805559e-01 -3.05059522e-01
-5.23693740e-01 3.39537472e-01 1.76698044e-01 9.99026671e-02
7.79169440e-01 -1.35110819e+00 -1.83280975e-01 4.91148800e-01
-3.56402516e-01 6.77749291e-02 8.69120583e-02 5.03971457e-01
-7.26725101e-01 5.91948450e-01 -1.66218415e-01 1.29462793e-01
-1.78887796e+00 4.11614388e-01 5.73975503e-01 -1.66828960e-01
-6.18450940e-01 1.08640730e+00 -4.64374274e-02 -1.12708464e-01
1.35658014e+00 -5.17472744e-01 1.75125673e-02 -4.54368085e-01
8.13277364e-01 6.59686089e-01 -2.26883233e-01 -1.49125382e-01
-2.77361870e-01 -1.27711087e-01 -3.74497116e-01 2.40785107e-01
1.03999376e+00 -1.58383511e-02 1.84927076e-01 -5.10756969e-01
1.50353193e+00 -1.10052668e-01 -1.16119635e+00 -4.01884228e-01
-2.27669522e-01 -4.21783209e-01 3.62295806e-01 -5.55637598e-01
-1.62024295e+00 9.06189978e-01 7.35746562e-01 1.53820887e-01
1.27799487e+00 -4.05812711e-01 1.55662155e+00 9.95100498e-01
5.64738333e-01 -8.58610034e-01 -5.07729888e-01 4.05416936e-01
-8.48300308e-02 -6.88385189e-01 -1.07635267e-01 -4.29388344e-01
-1.16609409e-01 8.95231128e-01 9.45752501e-01 1.52343791e-02
1.27567625e+00 5.47335327e-01 -1.79832399e-01 8.69908780e-02
-9.17470157e-01 4.42695260e-01 -3.05215865e-01 8.00410569e-01
-3.80886912e-01 -1.87359452e-02 -7.31155500e-02 1.44268751e+00
-5.21755576e-01 9.79590788e-02 3.90766084e-01 8.56464088e-01
-4.41210717e-01 -1.07828820e+00 1.91759050e-01 5.65828145e-01
-9.93990242e-01 2.16413923e-02 -4.60703462e-01 5.36960840e-01
1.67287424e-01 7.73246288e-01 1.16164580e-01 -1.02584577e+00
8.55371654e-02 -1.79905742e-01 1.67750299e-01 1.72983423e-01
-4.48425502e-01 -3.28942418e-01 2.35871524e-01 -5.95558941e-01
3.59011114e-01 -7.43263438e-02 -1.30329442e+00 -1.15890145e+00
-4.54759419e-01 -7.70072173e-03 3.67040426e-01 5.58621109e-01
4.88752902e-01 4.97886725e-02 6.72414243e-01 2.22021341e-01
-7.01109350e-01 -5.04008651e-01 -6.77228987e-01 -1.27234370e-01
2.85590827e-01 -2.50500232e-01 -3.86743963e-01 -5.12991786e-01] | [5.878729343414307, 7.054177761077881] |
a87a0b5c-db9e-4b64-94f9-f1d042899c46 | distantly-supervised-road-segmentation | 1708.06118 | null | http://arxiv.org/abs/1708.06118v1 | http://arxiv.org/pdf/1708.06118v1.pdf | Distantly Supervised Road Segmentation | We present an approach for road segmentation that only requires image-level
annotations at training time. We leverage distant supervision, which allows us
to train our model using images that are different from the target domain.
Using large publicly available image databases as distant supervisors, we
develop a simple method to automatically generate weak pixel-wise road masks.
These are used to iteratively train a fully convolutional neural network, which
produces our final segmentation model. We evaluate our method on the Cityscapes
dataset, where we compare it with a fully supervised approach. Further, we
discuss the trade-off between annotation cost and performance. Overall, our
distantly supervised approach achieves 93.8% of the performance of the fully
supervised approach, while using orders of magnitude less annotation work. | ['Satoshi Tsutsui', 'Tommi Kerola', 'Shunta Saito'] | 2017-08-21 | null | null | null | null | ['road-segementation'] | ['computer-vision'] | [ 4.31646883e-01 5.94502449e-01 -1.26710802e-01 -5.82016408e-01
-1.12794769e+00 -8.91083598e-01 5.65238416e-01 -1.11372940e-01
-6.93285108e-01 5.52803874e-01 -9.84992310e-02 -5.06429553e-01
4.42866355e-01 -8.84155869e-01 -1.09427047e+00 -2.77797759e-01
1.49906531e-01 5.31111479e-01 7.48403788e-01 2.01093704e-02
-5.15390113e-02 4.78311300e-01 -1.23060977e+00 2.08929479e-01
1.08564961e+00 7.35949159e-01 -1.19109806e-02 8.23046684e-01
3.54237147e-02 6.74304962e-01 -4.69262421e-01 -3.88182402e-01
5.44419408e-01 -9.70129445e-02 -1.26089442e+00 5.25844514e-01
8.13634455e-01 -4.77488279e-01 -2.69397587e-01 6.89210713e-01
1.32774249e-01 4.02751453e-02 7.88757563e-01 -8.55201662e-01
-4.35395390e-01 5.76113045e-01 -6.24231458e-01 6.64544627e-02
-2.01821133e-01 1.57662883e-01 9.63094473e-01 -7.23707378e-01
5.74053049e-01 8.74907553e-01 8.57797980e-01 2.55554408e-01
-1.39597762e+00 -3.82101953e-01 3.72826457e-01 -2.54010439e-01
-1.48314524e+00 -4.06616211e-01 5.85081160e-01 -5.55180550e-01
7.82046258e-01 -1.10755593e-01 2.63166338e-01 6.50983691e-01
-3.67520809e-01 8.93246889e-01 1.03603566e+00 -4.61159199e-01
2.98183574e-03 1.19378231e-01 8.30116794e-02 9.20057714e-01
2.94661615e-02 -1.28999993e-01 1.05192825e-01 2.21317112e-01
9.63654995e-01 -3.41609985e-01 -1.40577946e-02 -3.01056355e-01
-9.43671823e-01 7.50990689e-01 6.58761859e-01 5.35024814e-02
-3.86666544e-02 3.17342490e-01 4.03429940e-02 2.47853938e-02
6.57669365e-01 2.44055241e-01 -5.71527064e-01 1.66880280e-01
-1.08895302e+00 -9.41276476e-02 7.02931404e-01 9.47192490e-01
1.18760312e+00 -2.42942959e-01 7.26361200e-02 8.53359997e-01
5.86251281e-02 3.91749054e-01 -2.22925499e-01 -1.36461723e+00
5.93973339e-01 5.57177424e-01 2.05835328e-01 -5.77133000e-01
-3.37346643e-01 -4.02015060e-01 -3.54714841e-01 3.40682924e-01
7.75652289e-01 -5.14982045e-01 -1.36712372e+00 1.38622320e+00
2.54124612e-01 2.85879254e-01 -3.22237751e-03 5.12591481e-01
5.01710057e-01 4.80295509e-01 6.71528950e-02 5.00518918e-01
9.49670017e-01 -1.53842068e+00 -1.61733732e-01 -5.40342450e-01
9.24818635e-01 -5.69915295e-01 1.22863209e+00 2.29395434e-01
-9.77118373e-01 -5.83051980e-01 -8.97381127e-01 -1.97134748e-01
-5.02253175e-01 4.66998070e-01 4.11520183e-01 5.91355503e-01
-1.34075320e+00 5.25033414e-01 -8.85179758e-01 -2.42652714e-01
7.54651546e-01 2.01191723e-01 -2.65510470e-01 8.32302868e-03
-8.09011400e-01 6.98467612e-01 4.32127893e-01 -3.41758281e-02
-9.20582771e-01 -4.78069276e-01 -1.07460678e+00 -2.28450149e-02
3.97114396e-01 -3.27609599e-01 1.63770783e+00 -1.06151760e+00
-1.34265780e+00 1.15851319e+00 -1.99923679e-01 -5.03413737e-01
6.58089936e-01 -2.54972905e-01 -4.90013659e-02 3.71680051e-01
4.34939653e-01 1.28121352e+00 4.95525390e-01 -1.41187680e+00
-9.90678310e-01 -5.64396847e-03 2.08531916e-01 1.91925719e-01
4.03221659e-02 -2.04303965e-01 -1.12916875e+00 -4.54426855e-01
-8.64059851e-02 -1.12462676e+00 -6.43736899e-01 1.82905763e-01
-5.44333041e-01 -8.53443518e-02 9.24344301e-01 -6.75820291e-01
8.72162044e-01 -2.05847764e+00 -3.85758221e-01 3.99947822e-01
1.17820390e-01 2.60093868e-01 -2.59750724e-01 1.96844302e-02
1.00363120e-01 3.62157613e-01 -7.63400018e-01 -5.46447158e-01
-2.87004381e-01 3.28603685e-01 -1.32818550e-01 1.76454961e-01
6.22963428e-01 1.14057446e+00 -7.91701138e-01 -6.15608335e-01
2.58080751e-01 3.27507734e-01 -4.27308440e-01 2.28139684e-02
-2.98161775e-01 4.74335372e-01 -2.92275876e-01 5.19527316e-01
6.23958886e-01 -3.76474231e-01 -5.88400327e-02 2.43161116e-02
-1.39689922e-01 1.48793936e-01 -7.97509372e-01 1.63622880e+00
-6.22592688e-01 9.06428337e-01 5.99853508e-02 -9.39582288e-01
8.64378870e-01 7.28852749e-02 2.23065630e-01 -6.71784937e-01
-6.37590885e-02 2.45729208e-01 -3.62320572e-01 -2.45034963e-01
2.90664196e-01 3.09266150e-01 -6.06147759e-02 3.29738081e-01
-5.63557222e-02 -2.56786257e-01 1.47838026e-01 1.49377510e-01
1.14528036e+00 4.09184426e-01 -1.18741266e-01 -1.37126967e-01
3.48924637e-01 5.45508981e-01 2.91754425e-01 7.11750686e-01
-2.25267828e-01 9.80628729e-01 6.46731973e-01 -5.20851731e-01
-1.11280727e+00 -1.00904000e+00 -1.72701657e-01 8.91167223e-01
9.69702750e-02 3.84041248e-03 -1.23996842e+00 -1.26638246e+00
-1.69067219e-01 4.34670657e-01 -7.09811151e-01 2.86298692e-01
-7.54696131e-01 -4.63806599e-01 8.99416208e-01 8.39201748e-01
8.31836343e-01 -7.82146811e-01 -4.19886768e-01 -3.17782797e-02
-5.15196957e-02 -1.56561172e+00 -3.95499170e-01 1.74941331e-01
-8.11296284e-01 -1.05103040e+00 -7.88706422e-01 -9.89707768e-01
9.67115104e-01 3.35363895e-01 1.32728958e+00 1.54492661e-01
-1.24402709e-01 1.96027368e-01 -2.20471561e-01 -2.99478859e-01
-2.54684865e-01 5.71837544e-01 -7.94967353e-01 -2.30628267e-01
2.59586275e-01 -2.39784136e-01 -5.67359269e-01 4.20262605e-01
-7.71845758e-01 1.50529921e-01 7.26003647e-01 5.73792696e-01
7.80636132e-01 8.01297277e-02 4.67778802e-01 -1.35740733e+00
2.04893306e-01 -3.31164688e-01 -9.09658670e-01 9.48995724e-02
-4.53856051e-01 3.25553901e-02 3.84842366e-01 -5.20764738e-02
-1.20126486e+00 8.45448256e-01 -4.14428651e-01 -1.19450070e-01
-6.15991950e-01 3.38493407e-01 -1.57033220e-01 -1.77517772e-01
7.29588032e-01 -1.29558787e-01 -2.00770259e-01 -5.43274343e-01
6.47266448e-01 8.27520788e-01 8.33972752e-01 -5.16021907e-01
9.29348767e-01 7.03565478e-01 -3.87261689e-01 -6.32576704e-01
-1.40995443e+00 -5.41491270e-01 -1.15528643e+00 -6.18135836e-03
1.17999697e+00 -1.05611670e+00 5.22962622e-02 5.11909306e-01
-1.12045348e+00 -1.09664035e+00 -2.68134564e-01 1.22753032e-01
-5.17460704e-01 3.71868536e-02 -5.77057898e-01 -3.87711495e-01
-7.56804347e-02 -1.04682529e+00 1.36165357e+00 1.44994959e-01
2.01705378e-02 -1.20211124e+00 -4.21165489e-02 6.34127080e-01
1.33800194e-01 4.71678585e-01 4.36777979e-01 -6.28623128e-01
-7.09212363e-01 -6.06694929e-02 -6.05659902e-01 3.99835408e-01
1.58852205e-01 1.16269901e-01 -1.26178062e+00 1.14224188e-01
-7.21325040e-01 -3.76071423e-01 1.13190329e+00 3.09264839e-01
1.35289538e+00 -3.19746621e-02 -5.05922556e-01 6.90853179e-01
1.32473993e+00 -1.02688618e-01 7.43381143e-01 5.06499350e-01
9.98426855e-01 7.44681776e-01 7.54713476e-01 -1.62930012e-01
6.99805975e-01 5.61677396e-01 2.81469047e-01 -8.86709273e-01
-2.41813362e-01 -4.56981778e-01 9.06074867e-02 2.34189525e-01
7.47163668e-02 -1.58899248e-01 -1.30614996e+00 9.78663981e-01
-1.95429361e+00 -6.38053775e-01 -4.00631964e-01 1.93060040e+00
8.87328804e-01 4.66459662e-01 2.27736264e-01 -2.78285071e-02
6.19315207e-01 -5.37345335e-02 -4.92069781e-01 -3.77726674e-01
3.07156704e-02 3.81743640e-01 1.08840764e+00 8.22256863e-01
-1.66942191e+00 1.52029979e+00 7.62366915e+00 7.21038997e-01
-9.94932234e-01 -8.16002954e-03 1.12511444e+00 2.40160450e-01
-2.14516565e-01 5.06835952e-02 -6.93179429e-01 1.45315275e-01
1.04069483e+00 4.48850960e-01 9.52148214e-02 9.91394460e-01
3.78485292e-01 -1.55100644e-01 -9.23533082e-01 4.25652117e-01
-2.02819511e-01 -1.42917621e+00 -1.97210267e-01 1.42220438e-01
1.26497769e+00 5.33841252e-01 -1.18693233e-01 1.23919554e-01
9.46393669e-01 -1.24863625e+00 5.48417270e-01 1.61922127e-01
9.06465471e-01 -8.74458075e-01 7.20767856e-01 3.73569071e-01
-1.29992080e+00 1.95178628e-01 -1.09553762e-01 -3.80740054e-02
2.69325674e-01 7.29734123e-01 -9.03278053e-01 2.65636265e-01
8.38986695e-01 7.83584416e-01 -7.69654572e-01 8.59350026e-01
-7.13407516e-01 8.84988725e-01 -4.87911254e-01 7.47092128e-01
5.91266572e-01 -1.57244951e-01 -1.61718786e-01 1.46090138e+00
-8.57396051e-02 -6.48538694e-02 4.90166575e-01 9.04799402e-01
-3.28042805e-01 -1.71662137e-01 -8.23350966e-01 3.51915479e-01
4.43563402e-01 1.31942594e+00 -1.07740784e+00 -5.61662614e-01
-5.62111557e-01 1.15851796e+00 6.64348900e-01 5.11302412e-01
-9.02449071e-01 -5.65615416e-01 3.67716044e-01 1.07817240e-01
6.50761783e-01 -3.39501649e-01 -4.58178490e-01 -8.73478770e-01
1.20622478e-02 -5.85772097e-01 2.07625613e-01 -7.77614892e-01
-9.98306096e-01 4.42810267e-01 3.10134701e-02 -9.15964603e-01
-7.11694360e-02 -5.13093531e-01 -8.22188735e-01 8.42053115e-01
-1.99548340e+00 -1.40039921e+00 -5.05324841e-01 4.04341042e-01
6.79493606e-01 4.05626357e-01 4.70895946e-01 3.55897695e-01
-6.65482700e-01 4.62000251e-01 -1.09377578e-01 7.00936615e-01
6.21737957e-01 -1.45695639e+00 9.52592671e-01 1.02066994e+00
2.85288215e-01 2.30067581e-01 1.90139994e-01 -5.74873149e-01
-3.26657027e-01 -1.61833549e+00 8.19443643e-01 -6.04655445e-01
6.39168143e-01 -2.93277800e-01 -8.28166962e-01 1.05233967e+00
1.67025656e-01 2.56191701e-01 3.76848578e-01 -3.54253016e-02
-2.96755046e-01 9.83900428e-02 -1.13249099e+00 5.29436946e-01
1.15934324e+00 -4.92748678e-01 -4.35460776e-01 2.19934121e-01
9.27030146e-01 -5.05663157e-01 -7.05493450e-01 2.77987897e-01
2.96301991e-01 -8.29632282e-01 8.69501352e-01 -2.02303827e-01
4.96124357e-01 -4.50764537e-01 2.30513573e-01 -1.27249670e+00
-1.12747801e-02 -3.08839977e-01 3.28323305e-01 1.24982738e+00
9.97974336e-01 -5.49896657e-01 1.11947203e+00 6.55039847e-01
-2.34417498e-01 -6.49700165e-01 -5.05223751e-01 -7.47391403e-01
4.47861522e-01 -4.27726507e-01 4.40103382e-01 8.72559607e-01
-4.73383367e-01 2.63658643e-01 2.62241773e-02 3.98196518e-01
4.78865862e-01 4.19602208e-02 8.77217174e-01 -1.19549966e+00
-9.07679573e-02 -2.43905425e-01 -1.77841455e-01 -1.39824736e+00
2.80070543e-01 -7.30811477e-01 3.22077721e-01 -1.85722864e+00
-5.27430838e-03 -6.53886080e-01 -4.99655306e-02 9.92510557e-01
-1.47298440e-01 7.42504179e-01 -1.35598168e-01 1.96187958e-01
-6.78881705e-01 7.71645978e-02 1.13559723e+00 -2.62333900e-01
-3.16405952e-01 5.48546910e-02 -8.14580560e-01 1.01592529e+00
1.12754738e+00 -4.25347954e-01 -5.99912763e-01 -8.87186110e-01
-1.42745554e-01 -5.67225873e-01 4.04989749e-01 -9.23250735e-01
1.29585573e-02 -1.36485323e-01 2.98158437e-01 -5.98320842e-01
2.41078157e-02 -9.11051750e-01 -3.02743763e-01 2.04282090e-01
-4.05912399e-01 -1.43306449e-01 2.68792003e-01 4.62092608e-01
-2.79989183e-01 -1.82522506e-01 8.48328769e-01 -1.43386468e-01
-7.95513511e-01 2.42618918e-01 -3.52279603e-01 1.99418738e-01
1.12544703e+00 -3.14669698e-01 -3.08762193e-01 -3.45633090e-01
-7.59418249e-01 4.41181809e-01 7.50932276e-01 1.70637622e-01
2.74540782e-01 -9.64444816e-01 -6.83159053e-01 -4.80863899e-02
1.75286755e-01 6.58004224e-01 -2.43703678e-01 5.82417190e-01
-8.02128613e-01 3.00903291e-01 1.20741747e-01 -7.95478344e-01
-9.97941375e-01 2.01621965e-01 4.69647855e-01 -3.48186046e-01
-6.95925891e-01 6.90736175e-01 2.04582334e-01 -9.02562439e-01
-1.73959155e-02 -3.96157295e-01 -1.94453135e-01 -2.96444207e-01
2.76893526e-01 8.28108862e-02 1.20802455e-01 -5.86635709e-01
-2.61364222e-01 6.61672294e-01 2.17743181e-02 -3.70229483e-01
1.30066383e+00 -2.05878913e-01 2.81586081e-01 1.56987950e-01
1.23274851e+00 9.24377218e-02 -1.87383997e+00 -2.61119097e-01
2.13056728e-01 -3.65702361e-01 1.89721614e-01 -8.81881595e-01
-1.24553430e+00 9.11651433e-01 3.61159474e-01 3.33126783e-02
9.41564381e-01 2.63042122e-01 8.27522755e-01 5.18540263e-01
1.97730646e-01 -1.30685890e+00 -3.35540660e-02 6.50942385e-01
4.16907847e-01 -1.52791595e+00 -2.30620220e-01 -8.02412093e-01
-6.46663070e-01 7.86542296e-01 7.29654372e-01 -2.53224403e-01
5.49781024e-01 5.69263995e-01 4.55579847e-01 -2.12544620e-01
-3.27428043e-01 -7.55994260e-01 2.66544402e-01 8.14707160e-01
3.76195967e-01 -5.67525849e-02 7.04154149e-02 1.69489115e-01
-1.46925941e-01 9.17980745e-02 4.55885738e-01 9.13513780e-01
-6.23874009e-01 -1.20365167e+00 -1.85721040e-01 2.93919295e-01
-3.61333400e-01 -1.31523713e-01 -7.55617976e-01 8.36300850e-01
1.76993579e-01 1.06636071e+00 3.54243964e-01 -1.14204302e-01
3.70450974e-01 -8.93232822e-02 5.16530424e-02 -8.26470256e-01
-3.02137464e-01 8.58082399e-02 5.37131071e-01 -5.90635896e-01
-4.86217171e-01 -5.17270803e-01 -1.53676462e+00 -8.88273194e-02
-2.17281803e-01 -4.33812737e-02 5.46752453e-01 1.03424394e+00
4.54201758e-01 4.32944506e-01 7.15501308e-01 -1.00701272e+00
-3.76745971e-04 -5.31876445e-01 -7.58829340e-02 3.66648048e-01
3.13845396e-01 -3.52378488e-01 -1.95890158e-01 4.83502984e-01] | [9.5100736618042, 0.4623180031776428] |
bf794136-7787-4073-905d-ced95f64a8b3 | a-hybrid-morphological-disambiguation-system | null | null | https://aclanthology.org/I13-1175 | https://aclanthology.org/I13-1175.pdf | A Hybrid Morphological Disambiguation System for Turkish | null | ['Ilyas Cicekli', 'Mucahid Kutlu'] | 2013-10-01 | a-hybrid-morphological-disambiguation-system-1 | https://aclanthology.org/I13-1175 | https://aclanthology.org/I13-1175.pdf | ijcnlp-2013-10 | ['morphological-disambiguation'] | ['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] | [-7.343428611755371, 3.710782289505005] |
c8af8824-1015-4c02-a55b-dce6531a4d2a | large-scale-autonomous-flight-with-real-time | 2109.06479 | null | https://arxiv.org/abs/2109.06479v4 | https://arxiv.org/pdf/2109.06479v4.pdf | Large-scale Autonomous Flight with Real-time Semantic SLAM under Dense Forest Canopy | Semantic maps represent the environment using a set of semantically meaningful objects. This representation is storage-efficient, less ambiguous, and more informative, thus facilitating large-scale autonomy and the acquisition of actionable information in highly unstructured, GPS-denied environments. In this letter, we propose an integrated system that can perform large-scale autonomous flights and real-time semantic mapping in challenging under-canopy environments. We detect and model tree trunks and ground planes from LiDAR data, which are associated across scans and used to constrain robot poses as well as tree trunk models. The autonomous navigation module utilizes a multi-level planning and mapping framework and computes dynamically feasible trajectories that lead the UAV to build a semantic map of the user-defined region of interest in a computationally and storage efficient manner. A drift-compensation mechanism is designed to minimize the odometry drift using semantic SLAM outputs in real time, while maintaining planner optimality and controller stability. This leads the UAV to execute its mission accurately and safely at scale. Code is released at: https://github.com/KumarRobotics/kr_autonomous_flight and https://github.com/KumarRobotics/sloam. | ['Vijay Kumar', 'Camillo J. Taylor', 'Roseli A. F. Romero', 'Steven W. Chen', 'Chao Qu', 'Thomas Donnelly', 'Alex Zhou', 'Yuezhan Tao', 'Fernando Cladera Ojeda', 'Guilherme V. Nardari', 'Xu Liu'] | 2021-09-14 | null | null | null | null | ['semantic-slam'] | ['computer-vision'] | [-2.09492266e-01 -4.43194695e-02 -4.15790975e-02 -2.98962355e-01
-9.68282893e-02 -1.01642597e+00 4.69233207e-02 3.26010883e-01
-1.37828052e-01 9.85472143e-01 -5.74192286e-01 -9.48678851e-02
-5.98255932e-01 -1.26824868e+00 -6.44845188e-01 -3.74066383e-01
-3.77189845e-01 1.05321038e+00 6.34514809e-01 -5.05188525e-01
2.46328846e-01 9.78807390e-01 -1.92192256e+00 -6.99920475e-01
1.12444639e+00 1.10546017e+00 1.05078685e+00 6.37558401e-01
3.00804406e-01 1.64052814e-01 -8.06402937e-02 5.01970470e-01
5.34985006e-01 9.92880687e-02 -6.36516213e-01 1.59348458e-01
6.73147663e-03 -2.61481762e-01 -1.45451585e-02 1.08928871e+00
2.70564884e-01 3.34410518e-01 4.86705676e-02 -1.31066406e+00
5.61737679e-02 -1.58064961e-01 -1.18253537e-01 -2.98985422e-01
2.69511253e-01 2.24780768e-01 4.03053761e-01 -6.69904053e-01
8.57466280e-01 9.12629664e-01 5.34560502e-01 -1.14873283e-01
-8.64408135e-01 -5.64820945e-01 1.24383293e-01 8.00196454e-02
-1.97882915e+00 -3.45401645e-01 1.52625516e-01 -2.25615382e-01
7.58296788e-01 3.68117958e-01 1.00670791e+00 1.96503028e-01
5.78769028e-01 -9.25617889e-02 4.19142544e-01 -1.32810384e-01
3.88766974e-01 -2.32706532e-01 -3.21546584e-01 1.10470021e+00
9.01149631e-01 1.65277272e-01 -6.50297165e-01 7.41486102e-02
8.25912058e-01 1.18485093e-01 -2.51191556e-01 -1.07850802e+00
-1.32597160e+00 6.69739246e-01 7.51581788e-01 -2.64488190e-01
-6.27244115e-01 1.66377276e-01 -2.42097881e-02 5.06537929e-02
1.38684466e-01 5.55904210e-01 -4.28647906e-01 -7.86221698e-02
-5.89468539e-01 4.83799189e-01 6.36764467e-01 1.62905192e+00
1.03345346e+00 2.83176303e-01 5.41524112e-01 3.67538124e-01
1.60630658e-01 1.10654759e+00 1.41362295e-01 -1.47663188e+00
3.08487117e-01 8.15282047e-01 5.92159510e-01 -1.02143800e+00
-7.47236729e-01 -2.17742264e-01 -5.10871053e-01 2.65983641e-01
-2.08885506e-01 -7.78752193e-02 -8.53722453e-01 1.42583907e+00
8.25735867e-01 -1.77813530e-01 1.56681165e-01 1.12729394e+00
4.00034413e-02 4.98717725e-01 -3.70099217e-01 -1.39982089e-01
1.21618950e+00 -6.89977527e-01 -6.69695377e-01 -6.46949947e-01
6.71203971e-01 -6.25109375e-01 6.71616316e-01 1.84748903e-01
-6.53384626e-01 -4.02015865e-01 -1.23725462e+00 1.23540824e-02
-5.56533933e-01 3.75741899e-01 6.51847780e-01 1.77024771e-02
-1.26669550e+00 4.62335855e-01 -1.00573528e+00 -8.64390314e-01
9.09475833e-02 4.62701708e-01 -2.41766125e-01 -1.66335940e-01
-9.86794174e-01 1.28807259e+00 8.71653259e-01 3.23447376e-01
-1.02179444e+00 -3.34639877e-01 -9.37931061e-01 -3.27950567e-01
7.36888111e-01 -7.76150823e-01 1.06281948e+00 -2.20603094e-01
-1.39594245e+00 5.73309839e-01 -1.69222638e-01 -4.56230879e-01
1.15011148e-01 -3.87716979e-01 -1.43849924e-01 2.74072476e-02
5.19473076e-01 8.86258483e-01 4.63519484e-01 -1.16250205e+00
-1.07138193e+00 -8.63836348e-01 -5.64688928e-02 8.13991904e-01
1.85444742e-01 -6.43286645e-01 -3.00612062e-01 8.02271962e-02
9.46460664e-01 -1.48165834e+00 -4.91265953e-01 3.45682204e-01
-1.72241166e-01 4.29054469e-01 9.96756732e-01 -3.68965060e-01
8.89238894e-01 -1.74280763e+00 1.79849848e-01 7.91313499e-02
-1.55744389e-01 -1.50243312e-01 1.12367101e-01 6.83598757e-01
7.33566821e-01 -1.28196031e-01 -4.38445359e-01 1.01758286e-01
-3.41714799e-01 7.75095880e-01 -3.51286322e-01 4.41495717e-01
-2.33908013e-01 5.54004192e-01 -7.82631159e-01 -2.50620842e-01
3.94789279e-01 7.89442956e-02 -2.84411907e-01 -2.35424303e-02
-3.63645464e-01 5.87939382e-01 -8.67629051e-01 1.14919758e+00
8.95657122e-01 2.28497505e-01 4.57402885e-01 2.38349259e-01
-8.91932368e-01 2.89887726e-01 -1.35886860e+00 2.14903784e+00
-3.73553365e-01 3.81557792e-01 6.92163289e-01 -5.09113789e-01
1.18349576e+00 -3.06182832e-01 4.26888645e-01 -5.23595691e-01
5.09744659e-02 4.83257443e-01 -5.13463378e-01 -2.56527439e-02
1.15848947e+00 3.54206741e-01 -2.62284160e-01 -1.83429644e-01
-4.41151440e-01 -1.07064736e+00 5.61534017e-02 -1.65504336e-01
7.93876469e-01 5.28692067e-01 6.68210626e-01 -5.50327778e-01
4.30086523e-01 9.86604631e-01 7.06043243e-01 4.00369495e-01
-9.28346440e-02 6.67036464e-03 -3.80469173e-01 -4.32974488e-01
-8.11372459e-01 -1.16358960e+00 -1.04694806e-01 6.09086454e-01
1.07636881e+00 -3.21941733e-01 -2.99953997e-01 1.98341966e-01
5.31215727e-01 6.22897565e-01 3.23546790e-02 -9.92673635e-02
-5.17302752e-01 -2.23617747e-01 9.69697312e-02 1.96570128e-01
5.59745133e-01 -6.49723649e-01 -1.54893649e+00 3.85381758e-01
-2.93285608e-01 -1.07359815e+00 1.76525921e-01 4.55272764e-01
-1.04296410e+00 -1.21014750e+00 2.35289872e-01 -5.59890747e-01
7.83214152e-01 7.83983648e-01 5.56134045e-01 -1.25920191e-01
-5.02444565e-01 2.26930246e-01 -3.64295542e-01 -5.46702266e-01
1.83403522e-01 9.13824071e-04 5.80040455e-01 -6.36205494e-01
-2.26828277e-01 -4.96425360e-01 -3.73409122e-01 5.79177916e-01
-3.09156001e-01 2.25648612e-01 2.20662937e-01 3.63563895e-01
1.28087854e+00 2.32694238e-01 -2.16668770e-02 -1.54583722e-01
1.59436956e-01 -3.58437181e-01 -1.44276083e+00 1.68045796e-02
-7.39418387e-01 -1.79611176e-01 4.97176558e-01 1.16304077e-01
-6.89414918e-01 6.01490378e-01 5.84561586e-01 -4.31099951e-01
-5.99584319e-02 4.87392068e-01 -1.63678184e-01 -4.37426984e-01
4.40996617e-01 2.85254389e-01 1.73702821e-01 -2.61757731e-01
4.65512991e-01 5.08179903e-01 7.14922428e-01 -5.20043433e-01
1.01440239e+00 7.52475441e-01 6.70178175e-01 -8.65534663e-01
-5.22334218e-01 -4.20150936e-01 -1.20508754e+00 -2.58212417e-01
6.01668477e-01 -1.08150828e+00 -4.49755311e-01 1.69670627e-01
-9.47664678e-01 -4.58297908e-01 -5.50826192e-01 4.46929306e-01
-8.41258168e-01 5.12428582e-02 2.46584147e-01 -8.88030112e-01
-2.42990673e-01 -1.03458309e+00 1.02209222e+00 3.49676937e-01
1.76368952e-02 -4.47215557e-01 6.01409562e-02 1.26530165e-02
2.50638247e-01 7.35998213e-01 1.75347418e-01 1.06923364e-01
-1.38736165e+00 -2.00026467e-01 8.46907943e-02 -4.57499325e-01
1.15125038e-01 -6.28971756e-02 -3.53511661e-01 -6.67808056e-01
-2.55966157e-01 -1.07098885e-01 2.39128932e-01 3.19668204e-01
6.85241401e-01 -2.35934541e-01 -8.23792875e-01 8.87017965e-01
1.57174194e+00 4.73964870e-01 1.24614634e-01 6.99653506e-01
2.94476628e-01 7.61534095e-01 1.76941025e+00 6.70022964e-01
7.51993060e-01 8.71820509e-01 1.24287152e+00 3.96333098e-01
4.15449798e-01 -4.47783858e-01 2.35246494e-01 5.19840240e-01
1.63965166e-01 -3.15807402e-01 -1.14188266e+00 5.45362830e-01
-2.12923622e+00 -4.14005995e-01 -3.35700154e-01 2.48113275e+00
8.98333788e-02 -2.03413829e-01 -4.33229119e-01 -4.27448988e-01
7.78652072e-01 1.00437306e-01 -8.24347079e-01 -2.04655156e-01
1.04920328e-01 -2.14150339e-01 1.35700798e+00 7.33446836e-01
-8.81130517e-01 1.54117811e+00 4.69901133e+00 2.11687997e-01
-1.10563076e+00 1.75182551e-01 -5.35980523e-01 -1.77895874e-01
1.64275482e-01 5.87977886e-01 -8.98282886e-01 2.06773832e-01
1.16878760e+00 -3.83698434e-01 6.14970744e-01 1.13580453e+00
4.87952650e-01 -6.69274628e-01 -3.94824922e-01 7.21722126e-01
-4.21905428e-01 -1.48245347e+00 -3.10258687e-01 2.56537884e-01
4.61949915e-01 3.06501359e-01 -2.42428467e-01 -1.91254631e-01
5.90686500e-01 -3.61849695e-01 1.28682482e+00 5.10375321e-01
9.95650053e-01 -7.77841032e-01 3.83881152e-01 8.23723137e-01
-1.82130992e+00 -4.18562710e-01 -8.07124376e-01 -3.77256930e-01
4.61739868e-01 3.59080613e-01 -1.18288636e+00 1.04023659e+00
8.14743578e-01 5.50913632e-01 -3.73259634e-01 1.04737139e+00
-3.03502351e-01 -2.04526544e-01 -8.35200191e-01 4.24023233e-02
3.15577090e-01 -6.32915258e-01 8.15872252e-01 2.25247309e-01
8.07974994e-01 2.64832973e-01 7.31178880e-01 6.69941306e-01
4.12328392e-01 -1.22648738e-01 -9.86748576e-01 6.86001778e-02
1.11374974e+00 1.18410110e+00 -7.70794332e-01 -8.37044269e-02
4.27347988e-01 7.99810469e-01 1.57729566e-01 -9.13089812e-02
-7.24499404e-01 -3.61016303e-01 9.94944990e-01 2.66500980e-01
-2.81560719e-02 -9.40342009e-01 -1.79032996e-01 -6.98667467e-01
1.38473794e-01 -8.68613422e-02 1.16864569e-01 -1.19915748e+00
4.43868525e-02 4.47565347e-01 -2.55924612e-02 -1.49221277e+00
-3.60216737e-01 -2.25276619e-01 -1.04089737e-01 8.31427634e-01
-1.60426402e+00 -1.22254896e+00 -1.01223648e+00 3.41040909e-01
5.91492355e-01 -7.23935589e-02 9.39219356e-01 -3.40162814e-01
-1.94304422e-01 -5.49703419e-01 2.15897843e-01 -7.89658070e-01
4.80513394e-01 -8.36074889e-01 3.33390683e-01 1.18643332e+00
-3.04069966e-01 3.58139992e-01 7.59155095e-01 -1.29686165e+00
-1.79606903e+00 -1.45600224e+00 5.10733545e-01 -2.21940473e-01
4.26379025e-01 -2.88226247e-01 -4.73388880e-01 7.54793406e-01
-4.75540727e-01 -1.38215289e-01 -1.72933921e-01 -3.68552953e-01
4.50089514e-01 -4.18768793e-01 -1.21801782e+00 4.62721765e-01
1.39745045e+00 -9.07019600e-02 -1.19318277e-01 6.73891842e-01
9.74919021e-01 -9.78813529e-01 -7.29969382e-01 6.94845736e-01
4.96249795e-01 -7.17400670e-01 8.22836816e-01 -7.05021843e-02
-5.55275142e-01 -8.40742648e-01 -5.44664979e-01 -1.16022742e+00
-3.99309576e-01 -5.40877223e-01 1.09124087e-01 7.51954019e-01
2.26553623e-02 -1.04425323e+00 6.83228195e-01 2.06346482e-01
-6.43144667e-01 -3.10534805e-01 -1.16240227e+00 -1.05088198e+00
-7.05942810e-01 -9.85493213e-02 7.25289524e-01 5.08524418e-01
-3.77034634e-01 -1.63256422e-01 -1.69253692e-01 1.16499436e+00
6.50349677e-01 4.13665891e-01 9.57268178e-01 -1.44984710e+00
6.78538382e-01 1.89186424e-01 -4.87033784e-01 -7.59228945e-01
2.05278590e-01 -6.86098337e-01 4.67456162e-01 -1.76960588e+00
-7.52506793e-01 -1.02420354e+00 4.17119294e-01 5.72377384e-01
6.19624019e-01 -2.99108267e-01 8.62905681e-02 7.24394143e-01
-4.98610944e-01 7.07346320e-01 1.10490489e+00 2.98466057e-01
-3.69452715e-01 6.29973859e-02 -1.60760418e-01 5.52893341e-01
1.07185280e+00 -4.43871319e-01 -6.89299047e-01 -7.42970765e-01
-1.61160782e-01 5.03445148e-01 3.68357331e-01 -1.48213911e+00
5.83744228e-01 -7.46414125e-01 7.70522282e-02 -1.11516201e+00
8.67431223e-01 -1.11538076e+00 5.43690443e-01 9.25717175e-01
3.76752377e-01 3.89927745e-01 3.47783655e-01 6.93949282e-01
-2.73836516e-02 -1.90816179e-01 6.95061326e-01 -3.34886968e-01
-1.26659679e+00 5.42865455e-01 -3.08801502e-01 -3.82784128e-01
1.50409794e+00 -3.61728251e-01 -3.49210143e-01 -8.07993859e-02
-4.81285572e-01 8.59713972e-01 1.10333169e+00 5.18819392e-01
7.48807251e-01 -9.14545357e-01 -1.68879971e-01 4.71412241e-01
2.23020509e-01 6.60170734e-01 2.38589570e-01 6.21953607e-01
-1.35550869e+00 8.70643437e-01 -6.44784331e-01 -8.18928003e-01
-9.99783397e-01 3.03476006e-01 2.04656556e-01 2.61394769e-01
-5.90835392e-01 5.26364684e-01 -2.35972494e-01 -7.35410571e-01
-3.36727768e-01 -4.11789775e-01 2.92438328e-01 -1.90485805e-01
1.87378094e-01 4.51655567e-01 5.91205433e-02 -8.20685983e-01
-8.13257575e-01 8.55654418e-01 7.70500243e-01 -1.36758611e-01
1.12914860e+00 -8.77362430e-01 -2.90218085e-01 9.16293040e-02
2.84606338e-01 -1.56504899e-01 -1.29201734e+00 1.92126408e-01
1.78814799e-01 -7.83683658e-01 2.16299742e-01 -5.51570714e-01
-5.11213124e-01 4.62514579e-01 6.99024558e-01 -3.72903608e-02
9.52280819e-01 -3.23333740e-01 5.85705280e-01 6.98569238e-01
1.37181759e+00 -1.18863165e+00 -4.20084625e-01 7.65140653e-01
9.87937152e-01 -8.91230226e-01 2.34652475e-01 -6.09201312e-01
-7.22746849e-01 1.04328966e+00 8.23776782e-01 -8.23710561e-02
2.72999167e-01 3.70615810e-01 -1.07674256e-01 -2.71000624e-01
-6.62461638e-01 -5.41113853e-01 -4.19247240e-01 1.06838918e+00
-6.35567546e-01 4.82817203e-01 -1.37697505e-02 8.89623165e-02
-4.59794343e-01 -1.31623536e-01 6.04074001e-01 1.45071018e+00
-1.31212175e+00 -7.58865535e-01 -6.51877880e-01 9.40389484e-02
7.34944999e-01 2.79491305e-01 -2.30346605e-01 1.01565325e+00
2.22647116e-01 7.17004478e-01 2.27758542e-01 -3.33171129e-01
5.27558267e-01 -6.46345168e-02 2.12227687e-01 -8.36637139e-01
1.35374531e-01 -2.32344836e-01 3.35693419e-01 -1.08693862e+00
1.80438563e-01 -6.65874124e-01 -1.71108770e+00 -2.12875798e-01
-2.99222082e-01 2.30623484e-01 1.17210233e+00 3.78077030e-01
8.74979258e-01 2.36122921e-01 5.70479691e-01 -1.05209053e+00
-5.08848727e-01 -5.69052100e-01 -6.39083028e-01 -5.89844584e-01
9.87555236e-02 -1.26256442e+00 1.47797093e-01 -2.80909240e-01] | [7.2968549728393555, -1.997840404510498] |
48c3468e-9d91-4511-9b72-ec631b121027 | introduction-to-protein-folding | 2307.02174 | null | https://arxiv.org/abs/2307.02174v2 | https://arxiv.org/pdf/2307.02174v2.pdf | Introduction to Protein Folding | While many good textbooks are available on Protein Structure, Molecular Simulations, Thermodynamics and Bioinformatics methods in general, there is no good introductory level book for the field of Structural Bioinformatics. This book aims to give an introduction into Structural Bioinformatics, which is where the previous topics meet to explore three dimensional protein structures through computational analysis. We provide an overview of existing computational techniques, to validate, simulate, predict and analyse protein structures. More importantly, it will aim to provide practical knowledge about how and when to use such techniques. We will consider proteins from three major vantage points: Protein structure quantification, Protein structure prediction, and Protein simulation & dynamics. In this chapter we explore basic physical and chemical concepts required to understand protein folding. We introduce major (de)stabilising factors of folded protein structures such as the hydrophobic effect and backbone entropy. In addition, we consider different states along the folding pathway, as well as natively disordered proteins and aggregated protein states. In this chapter, an intuitive understanding is provided about the protein folding process, to prepare for the next chapter on the thermodynamics of protein folding. In particular, it is emphasized that protein folding is a stochastic process and that proteins unfold and refold in a dynamic equilibrium. The effect of temperature on the stability of the folded and unfolded states is also explained. | ['Sanne Abeln', 'K. Anton Feenstra', 'Isabel Houtkamp', 'Annika Jacobsen', 'Jochem Bijlard', 'Halima Mouhib', 'Ali May', 'Erik van Dijk', 'Juami H. M. van Gils'] | 2023-07-05 | null | null | null | null | ['protein-structure-prediction', 'protein-folding'] | ['miscellaneous', 'natural-language-processing'] | [ 2.13790908e-01 -1.39869690e-01 -2.38785103e-01 -3.99577707e-01
-1.17974356e-03 -6.31459951e-01 -2.07364336e-01 4.16248888e-01
-1.36325642e-01 1.22804368e+00 -6.95173889e-02 -7.55953610e-01
2.77331620e-01 -3.36157918e-01 -7.51769066e-01 -1.29342639e+00
-3.10430467e-01 2.96352357e-01 1.32837920e-02 -5.29344201e-01
1.42896295e-01 7.13222682e-01 -1.38899314e+00 1.02850378e-01
1.22379255e+00 2.05584496e-01 5.89458704e-01 8.37610126e-01
-2.51978308e-01 5.99301040e-01 -2.70743936e-01 -3.42459261e-01
-2.90073901e-01 -9.01087821e-01 -1.05680084e+00 -5.46325594e-02
-1.74331680e-01 1.63430631e-01 1.97636455e-01 5.71961582e-01
5.14587462e-01 -3.85091305e-02 6.19588912e-01 -3.63978326e-01
-7.23174214e-01 3.46761122e-02 -6.21460527e-02 2.05064386e-01
6.36087060e-01 5.27440131e-01 6.36828899e-01 -7.54785895e-01
8.16598833e-01 9.94664013e-01 5.01583517e-01 8.07817757e-01
-1.53967774e+00 -4.11645994e-02 -6.28566369e-02 2.92048991e-01
-8.23340833e-01 -3.24859321e-01 3.98820549e-01 -5.81788182e-01
1.64271784e+00 3.51259917e-01 8.32352340e-01 6.54069602e-01
1.01258934e+00 3.00619066e-01 1.31669044e+00 -5.35897672e-01
3.00736696e-01 -1.43255889e-01 7.88944066e-01 5.37714660e-01
2.53999114e-01 2.74215400e-01 -4.87586558e-01 -4.56240237e-01
-1.77761186e-02 6.37932122e-02 -2.40740195e-01 -7.43888021e-01
-7.50143707e-01 5.86959898e-01 2.77955793e-02 3.18922430e-01
-5.61767161e-01 -2.37959787e-01 3.05909008e-01 7.75118694e-02
7.60684684e-02 2.14232177e-01 -8.84452522e-01 -1.63938835e-01
-6.02566183e-01 5.06178319e-01 9.89636660e-01 3.90377939e-01
7.71714032e-01 -4.07557517e-01 4.37489718e-01 6.17383540e-01
5.06378770e-01 5.08824706e-01 2.42070258e-01 -5.09709179e-01
-1.58348441e-01 4.47873354e-01 3.07666659e-01 -3.67369145e-01
-4.95416403e-01 5.15182734e-01 -6.08373702e-01 3.01514655e-01
5.27690947e-01 -3.54544334e-02 -9.52665627e-01 1.64624941e+00
3.70263308e-01 -6.59118176e-01 2.47103408e-01 8.84580314e-01
8.37898433e-01 5.65211058e-01 6.34762228e-01 -1.07057416e+00
1.63775611e+00 -5.16635895e-01 -8.34973037e-01 1.98329702e-01
5.11929333e-01 -9.01246846e-01 5.95078409e-01 1.58902228e-01
-1.20112813e+00 -2.67328829e-01 -1.00891984e+00 -3.85183603e-01
-4.94194984e-01 -1.94380865e-01 5.04000485e-01 8.00602317e-01
-7.29014575e-01 1.04433441e+00 -1.33132625e+00 -9.20989215e-01
-2.27178007e-01 3.90316635e-01 -3.25430036e-01 2.97496259e-01
-1.28362894e+00 1.60604846e+00 4.68679309e-01 -1.37112767e-01
-3.46458048e-01 -5.87098241e-01 -4.46024239e-01 -4.54552203e-01
-3.41727525e-01 -6.75031900e-01 1.07239377e+00 -5.55574000e-01
-1.60653865e+00 1.23743749e+00 -1.06890261e+00 -2.42095053e-01
-2.75035232e-01 6.32106587e-02 -1.35029435e-01 8.30178857e-02
-4.58415478e-01 1.85963765e-01 -1.75723538e-01 -1.03676736e+00
-1.69171952e-02 -6.46579683e-01 -2.38238439e-01 3.17941725e-01
6.67613626e-01 4.18970972e-01 5.15559077e-01 -2.73628861e-01
3.44658107e-01 -8.42703283e-01 -5.97912610e-01 -3.74780893e-01
-1.33057892e-01 -1.98687017e-01 5.22458971e-01 -7.35448897e-01
9.43461478e-01 -1.57248640e+00 5.11329532e-01 1.76144660e-01
3.43301415e-01 2.35508785e-01 4.74800318e-01 1.28797817e+00
-6.19686663e-01 -4.87280525e-02 -4.03902143e-01 5.53827941e-01
-3.27145189e-01 2.71315187e-01 -2.00299948e-01 6.80867136e-01
-8.92972574e-02 1.03772533e+00 -7.32411742e-01 -1.74633658e-03
4.17488664e-01 6.43629372e-01 -1.39635727e-01 1.98794186e-01
-1.45021543e-01 5.91647506e-01 -1.97021261e-01 7.52740085e-01
1.05454540e+00 -2.28116125e-01 9.85380530e-01 -2.02668637e-01
-4.99949485e-01 6.24107182e-01 -3.67680937e-01 9.85958159e-01
4.40499961e-01 -4.76651229e-02 -2.23584683e-03 -7.98110187e-01
1.11565185e+00 3.46336752e-01 6.46395206e-01 -3.46605927e-01
-1.88585266e-01 3.74911636e-01 3.42848480e-01 -6.13894105e-01
6.07414693e-02 -7.80645549e-01 4.67982739e-01 4.03562725e-01
-1.55142143e-01 2.43842691e-01 3.59768271e-01 1.15028173e-01
7.27686882e-01 4.79650855e-01 7.32749939e-01 -7.70196378e-01
7.98517704e-01 3.58761996e-01 5.94028413e-01 -6.94818422e-02
-5.20980358e-01 -1.07862139e-02 6.26991272e-01 -9.21397924e-01
-1.34246707e+00 -8.48090768e-01 -4.03993428e-01 1.23395050e+00
3.31378847e-01 -8.01968515e-01 -1.20085192e+00 6.35155961e-02
-1.32590175e-01 2.82808483e-01 -2.71392703e-01 -1.40008971e-01
-8.26978564e-01 -1.70073271e+00 2.70102825e-02 6.45574629e-02
-1.29100397e-01 -1.34608793e+00 -7.56757617e-01 3.99876803e-01
-2.63482690e-01 -4.46565151e-01 -7.53205568e-02 8.64910722e-01
-1.22310984e+00 -1.53595233e+00 -3.80704641e-01 -7.84190059e-01
5.67128479e-01 1.68186724e-01 1.16417718e+00 3.93188655e-01
-4.21542883e-01 -2.00121015e-01 -2.73106426e-01 -3.01448554e-01
-6.93285346e-01 -1.60083964e-01 2.96677679e-01 -7.06895411e-01
1.25114894e+00 -8.00763965e-01 -8.88840973e-01 3.14761251e-01
-6.17898881e-01 5.66159338e-02 4.06751961e-01 7.81014144e-01
8.47535491e-01 -5.60941577e-01 3.60422693e-02 -7.00071156e-01
6.25116587e-01 -1.40485540e-01 -1.33987248e-01 5.43595910e-01
-9.02451217e-01 3.88999343e-01 4.97753918e-01 1.02368407e-01
-9.51257288e-01 4.55697834e-01 -6.68169916e-01 7.08593369e-01
-3.99956822e-01 1.91688418e-01 -4.63174701e-01 -1.78656802e-01
7.70933867e-01 5.95295191e-01 5.83581924e-01 -6.85901284e-01
4.91003245e-02 3.47470284e-01 -4.51502465e-02 -6.88367844e-01
2.60033011e-01 2.06547886e-01 9.73948315e-02 -9.05280292e-01
-3.20719421e-01 -3.61205876e-01 -9.54603255e-01 -7.07504153e-02
8.82482886e-01 -2.35824853e-01 -1.41648805e+00 4.01475757e-01
-9.45873737e-01 -2.68938333e-01 2.00142935e-01 4.35068250e-01
-9.60197389e-01 1.03919518e+00 -8.47817481e-01 -7.65636861e-01
-6.10061824e-01 -1.58189702e+00 7.08549500e-01 1.66955888e-01
-3.22166055e-01 -1.00891423e+00 5.85905254e-01 3.42356026e-01
1.44962713e-01 6.35866702e-01 1.26377082e+00 -1.07289530e-01
-2.16537088e-01 5.37319005e-01 4.10625428e-01 -4.71599251e-02
2.03964725e-01 4.81298447e-01 -4.90455836e-01 -3.73643726e-01
9.10853222e-02 -2.55352169e-01 9.44997370e-01 6.07136548e-01
5.26918650e-01 -1.51877448e-01 -6.21226966e-01 4.65574712e-01
1.44044781e+00 5.54350197e-01 8.45344841e-01 3.84078264e-01
2.14496955e-01 9.96698380e-01 7.39735365e-01 -9.37107578e-03
-1.69182941e-01 4.38167572e-01 2.23783359e-01 1.19446041e-02
3.00252825e-01 -2.29820721e-02 5.81778824e-01 8.47717285e-01
-8.05518031e-01 1.06027849e-01 -1.02610517e+00 -1.15350164e-01
-1.57107234e+00 -1.05858111e+00 -6.02384567e-01 2.10422087e+00
1.43186831e+00 -3.04299951e-01 5.04084527e-01 -1.55275598e-01
5.62722862e-01 -3.02094042e-01 -8.63225400e-01 -9.29342985e-01
-4.11010712e-01 3.75292242e-01 6.29807889e-01 8.31009388e-01
-7.65711963e-01 8.79037321e-01 8.61388779e+00 2.57824004e-01
-1.07084250e+00 -2.11045384e-01 4.62029189e-01 1.85068935e-01
-1.25358790e-01 4.51721191e-01 -7.54519463e-01 4.95663702e-01
1.34133852e+00 -3.37662064e-02 2.63635695e-01 6.38502955e-01
6.41918778e-01 -3.03587914e-01 -7.46474922e-01 4.89080489e-01
-5.73793054e-01 -1.30614984e+00 -2.68386781e-01 3.58672827e-01
2.39252314e-01 -1.25732630e-01 -3.66102368e-01 -2.19531640e-01
1.44588932e-01 -1.19910765e+00 6.58348575e-02 6.50352657e-01
4.91239399e-01 -7.89762735e-01 6.64067030e-01 4.45266724e-01
-1.02883410e+00 3.67893785e-01 -8.87946069e-01 -3.52540940e-01
3.90443236e-01 8.02777052e-01 -3.85485768e-01 2.64028221e-01
6.56558931e-01 4.62116838e-01 -1.55639917e-01 4.86725420e-01
-6.73473030e-02 3.16689730e-01 -1.26061842e-01 -4.37760174e-01
-1.40784949e-01 -8.66932452e-01 1.65161625e-01 1.17556489e+00
-3.58036727e-01 8.77260029e-01 -9.03692171e-02 6.85723066e-01
4.17762637e-01 4.20548469e-01 -9.13993865e-02 -1.41068801e-01
1.79644693e-02 9.62422132e-01 -6.37997031e-01 -2.56772190e-01
-3.34037036e-01 7.79151797e-01 2.22752064e-01 4.19671685e-01
-5.57440639e-01 -2.96066284e-01 1.26964152e+00 2.43059069e-01
6.07601665e-02 -3.76334101e-01 8.20386633e-02 -9.90876973e-01
-1.05989940e-01 -8.50584328e-01 -1.12491377e-01 -2.84628808e-01
-1.09591126e+00 7.79702365e-02 -2.46358797e-01 -1.81046709e-01
2.51921833e-01 -9.45151925e-01 -2.67296553e-01 1.24421191e+00
-1.18349063e+00 -6.24286056e-01 1.47801563e-01 -1.32321725e-02
1.23616360e-01 3.05110246e-01 9.86259818e-01 -1.77111581e-01
-5.72113812e-01 2.44340196e-01 9.30581033e-01 -5.07407367e-01
7.89988339e-01 -1.44602871e+00 6.37358904e-01 3.59247684e-01
-9.26361918e-01 1.42716253e+00 1.33628178e+00 -1.00518048e+00
-1.62831354e+00 -6.21027231e-01 1.17978096e+00 -6.05827212e-01
1.01832390e-01 -3.52438986e-01 -1.13977945e+00 4.05788273e-01
2.29256555e-01 -7.19743848e-01 1.32193339e+00 1.85368862e-02
1.26430213e-01 4.41249967e-01 -1.21109498e+00 3.64057809e-01
8.43159080e-01 -2.45428175e-01 -6.77768767e-01 6.24474823e-01
3.36052656e-01 -4.53291476e-01 -1.30562615e+00 3.47589314e-01
9.35912311e-01 -1.32718074e+00 9.81332481e-01 -1.12448597e+00
7.86716267e-02 -4.05652314e-01 1.10349663e-01 -6.25617683e-01
-6.62602067e-01 -7.40687251e-01 -3.97432357e-01 3.75392616e-01
4.28824723e-01 -5.79905450e-01 9.19110298e-01 9.85319078e-01
1.61471535e-02 -1.18932998e+00 -5.45590937e-01 -5.73930323e-01
5.99709213e-01 5.85145354e-01 4.21891630e-01 5.37888587e-01
8.72337639e-01 4.05242115e-01 -2.19870299e-01 -5.00704050e-01
5.68048954e-01 3.71675700e-01 4.22825634e-01 -9.32267487e-01
1.32252231e-01 -4.17988673e-02 -3.68332833e-01 -1.14991450e+00
-1.20588191e-01 -6.17193937e-01 -1.96759656e-01 -1.43631399e+00
7.10901916e-01 2.48301223e-01 -5.91848465e-03 9.21662152e-02
-2.99187928e-01 -2.15630934e-01 -1.82650343e-01 3.93368423e-01
-1.08819030e-01 2.83199966e-01 1.20902157e+00 3.71783674e-01
-3.10966671e-01 -2.01390147e-01 -6.36368930e-01 2.65912533e-01
1.28124714e+00 -3.53351116e-01 -1.55383706e-01 6.30783200e-01
3.60682875e-01 -2.33340308e-01 -7.20684007e-02 -2.29382649e-01
-3.32831442e-01 -6.10752821e-01 3.16363096e-01 -9.53665555e-01
1.25787959e-01 -5.61643481e-01 4.77583170e-01 1.06845760e+00
-2.76016723e-02 1.81164563e-01 -4.78480756e-02 4.26353306e-01
1.45253152e-01 -2.03881487e-01 1.34645379e+00 -4.44542170e-01
-3.46234530e-01 1.31136747e-02 -1.05509877e+00 -4.22015727e-01
1.40075946e+00 -7.19438016e-01 -2.30484232e-01 2.62735337e-01
-1.37553215e+00 1.10091083e-01 1.20219660e+00 -2.63596594e-01
2.92868972e-01 -8.68344009e-01 -2.16386959e-01 -2.56379619e-02
-8.05099458e-02 -6.75826252e-01 4.79255229e-01 1.12740433e+00
-1.44965136e+00 9.76273715e-01 -3.58899713e-01 -5.54975271e-01
-1.73706150e+00 1.12284684e+00 8.33787560e-01 -3.99992578e-02
-2.00184301e-01 2.91799396e-01 1.48288980e-01 -6.28963530e-01
-2.48540565e-01 -2.63569325e-01 -1.50095508e-01 -5.03208995e-01
5.85657001e-01 3.53676409e-01 2.28975654e-01 -7.67186284e-01
-7.20677733e-01 7.79091120e-01 -1.33160293e-01 7.18506932e-01
1.08822155e+00 -5.27892351e-01 -9.29255605e-01 3.08151960e-01
8.58013153e-01 -3.91600400e-01 -8.74531388e-01 1.38573766e-01
1.78325996e-01 9.78072062e-02 -7.24485993e-01 -7.94841111e-01
-2.07901120e-01 7.90167272e-01 7.12852180e-01 2.13553631e-05
9.85815525e-01 1.67450249e-01 9.66615081e-01 5.57183981e-01
1.02657504e-01 -1.08261633e+00 -6.26725733e-01 5.92068017e-01
5.56478143e-01 -9.50207591e-01 1.75297260e-01 -5.70570111e-01
-4.01335895e-01 1.37482536e+00 3.97909760e-01 1.18183248e-01
6.36854410e-01 2.47987211e-01 1.87381044e-01 -3.63272548e-01
-1.00965989e+00 -5.55565208e-02 -2.40525395e-01 7.12533534e-01
1.34775162e+00 1.67007357e-01 -1.28429055e+00 2.81025976e-01
-3.07813823e-01 -2.51838356e-01 3.22079390e-01 1.24281514e+00
-1.04690301e+00 -1.76764429e+00 -5.25418341e-01 7.33800754e-02
-6.69677675e-01 -6.02049194e-02 -1.15223610e+00 3.30544144e-01
2.19081379e-02 8.68188560e-01 -5.62893927e-01 -3.85796539e-02
1.28216177e-01 6.96779490e-01 7.67274737e-01 -2.22770020e-01
-5.33614337e-01 -1.08606905e-01 -1.24384630e-02 -5.13801694e-01
-9.31377590e-01 -6.24344647e-01 -1.59295881e+00 -1.03065503e+00
-4.11042243e-01 8.43635142e-01 7.04625189e-01 7.76162803e-01
5.46100616e-01 1.06601074e-01 2.02474579e-01 -5.46568215e-01
-2.77015299e-01 -7.09395707e-01 -7.92483628e-01 2.52047539e-01
3.31737369e-01 -5.53387046e-01 -1.27960190e-01 3.44430923e-01] | [4.734662055969238, 5.285408973693848] |
545eef46-42c6-45a8-935d-a9654089bae3 | faceformer-speech-driven-3d-facial-animation | 2112.05329 | null | https://arxiv.org/abs/2112.05329v4 | https://arxiv.org/pdf/2112.05329v4.pdf | FaceFormer: Speech-Driven 3D Facial Animation with Transformers | Speech-driven 3D facial animation is challenging due to the complex geometry of human faces and the limited availability of 3D audio-visual data. Prior works typically focus on learning phoneme-level features of short audio windows with limited context, occasionally resulting in inaccurate lip movements. To tackle this limitation, we propose a Transformer-based autoregressive model, FaceFormer, which encodes the long-term audio context and autoregressively predicts a sequence of animated 3D face meshes. To cope with the data scarcity issue, we integrate the self-supervised pre-trained speech representations. Also, we devise two biased attention mechanisms well suited to this specific task, including the biased cross-modal multi-head (MH) attention and the biased causal MH self-attention with a periodic positional encoding strategy. The former effectively aligns the audio-motion modalities, whereas the latter offers abilities to generalize to longer audio sequences. Extensive experiments and a perceptual user study show that our approach outperforms the existing state-of-the-arts. The code will be made available. | ['Taku Komura', 'Wenping Wang', 'Jun Saito', 'Zhaojiang Lin', 'Yingruo Fan'] | 2021-12-10 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Fan_FaceFormer_Speech-Driven_3D_Facial_Animation_With_Transformers_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Fan_FaceFormer_Speech-Driven_3D_Facial_Animation_With_Transformers_CVPR_2022_paper.pdf | cvpr-2022-1 | ['3d-face-animation'] | ['computer-vision'] | [ 2.08523959e-01 2.40418807e-01 -3.10944077e-02 -3.47359955e-01
-8.78030539e-01 -2.07983050e-02 5.74815571e-01 -5.77320755e-01
7.98028484e-02 4.15859073e-01 6.69694304e-01 1.09103270e-01
7.89964795e-02 -3.96674573e-01 -6.87368631e-01 -6.70971930e-01
-2.21935734e-01 3.17121476e-01 -7.50942379e-02 -1.31763980e-01
1.12280749e-01 5.69234967e-01 -2.13560581e+00 4.03769344e-01
4.90345627e-01 1.21884882e+00 2.39639744e-01 5.99462569e-01
-1.90819472e-01 7.21464574e-01 -1.60330698e-01 -2.47275278e-01
4.53917459e-02 -5.12666166e-01 -4.71314192e-01 3.28425914e-01
6.12922072e-01 -3.02923560e-01 -2.66785681e-01 7.64883578e-01
8.51726413e-01 1.48855090e-01 7.10851908e-01 -1.40332580e+00
-5.46840370e-01 2.05635533e-01 -6.30120575e-01 1.37828365e-01
6.37821913e-01 1.18671291e-01 9.16525483e-01 -1.34186506e+00
6.12593949e-01 1.65412676e+00 5.96980572e-01 8.32318246e-01
-1.03091717e+00 -7.78822064e-01 4.36204225e-01 3.43692660e-01
-1.62110269e+00 -9.75103021e-01 1.35186338e+00 -5.85508585e-01
8.27340305e-01 2.24178553e-01 7.71409571e-01 1.56353676e+00
1.29899681e-02 6.46933973e-01 1.00231993e+00 -3.32207859e-01
1.65236592e-01 2.30261572e-02 -6.04207337e-01 6.41911328e-01
-6.27409697e-01 2.38991708e-01 -9.51810181e-01 -1.93260923e-01
8.94220293e-01 -3.22637171e-01 -1.49246246e-01 -5.10951042e-01
-8.16082001e-01 6.07564092e-01 9.32860374e-02 2.61446446e-01
-5.62982202e-01 8.31542015e-02 3.76094401e-01 1.20358624e-01
9.24130857e-01 -9.71050188e-02 -2.95929104e-01 -1.63066164e-01
-1.19327164e+00 3.41801882e-01 3.85464698e-01 8.74536514e-01
4.90930617e-01 5.11472523e-01 -6.98850630e-03 9.54262912e-01
6.19236648e-01 4.70589936e-01 4.34552789e-01 -1.03429139e+00
1.69904307e-01 4.22030240e-02 -2.56853163e-01 -1.06373298e+00
-2.47812778e-01 -2.61914045e-01 -8.24799061e-01 2.59684712e-01
1.25504225e-01 -5.44480002e-03 -7.19013989e-01 2.05250478e+00
6.32074177e-01 6.37162328e-01 -2.77099043e-01 1.03376067e+00
9.65627670e-01 6.26776159e-01 2.09348843e-01 -5.92125714e-01
1.25453925e+00 -6.40176833e-01 -9.96155500e-01 -1.15939595e-01
-1.50745623e-02 -7.88210571e-01 1.14912617e+00 2.06961095e-01
-1.39112747e+00 -7.46292472e-01 -7.46268272e-01 8.40858091e-04
-1.67308912e-01 -1.46340907e-01 2.30810419e-01 4.65704918e-01
-1.16494155e+00 4.05804843e-01 -7.13460088e-01 -1.84780717e-01
4.94005710e-01 1.86803252e-01 -3.44349593e-01 1.59343600e-01
-1.14410889e+00 5.44247389e-01 -3.12342435e-01 2.02863172e-01
-9.80180383e-01 -8.08548033e-01 -1.12746119e+00 1.22001739e-02
2.66982883e-01 -6.71245158e-01 1.14768255e+00 -1.32091796e+00
-1.99087679e+00 8.77124786e-01 -4.15690243e-01 3.71987838e-03
5.31593800e-01 -3.91846299e-01 -2.44288310e-01 2.57405907e-01
-1.97162211e-01 8.61011326e-01 1.49479198e+00 -1.39850020e+00
-2.47619152e-01 -5.15953660e-01 -2.07712993e-01 3.85084271e-01
-4.48694795e-01 1.47416860e-01 -3.68587643e-01 -9.83144462e-01
-9.02488083e-02 -7.53993273e-01 -2.07790546e-02 2.23249152e-01
-1.41856804e-01 -1.55015975e-01 1.05229938e+00 -5.13176560e-01
1.14972329e+00 -2.29199505e+00 5.07374823e-01 -1.06680058e-01
-1.25321060e-01 3.41172852e-02 -2.48040497e-01 2.80817330e-01
-3.58148575e-01 -4.30458784e-02 -2.86054797e-02 -8.32142174e-01
-1.46972924e-01 3.48156616e-02 -3.76511514e-01 4.86490160e-01
5.11854589e-01 6.01659715e-01 -7.24166751e-01 -8.06390047e-01
1.88046470e-01 1.00205505e+00 -8.02747607e-01 5.79557836e-01
-2.91551441e-01 7.87467122e-01 -2.99552590e-01 7.46481001e-01
6.08217776e-01 -6.01053275e-02 -1.10242315e-01 -2.31594965e-01
-1.11141481e-01 1.90637216e-01 -1.15847588e+00 1.99377656e+00
-6.17674828e-01 5.48809052e-01 5.09008408e-01 -8.27327728e-01
8.14861596e-01 7.27185249e-01 6.07509792e-01 -5.76053262e-01
1.97723955e-01 9.71800834e-03 -3.03125024e-01 -8.22022498e-01
3.29463691e-01 -3.39896232e-01 2.58727759e-01 1.36559844e-01
1.27836198e-01 -2.07554102e-01 -4.53304738e-01 -1.62145391e-01
5.70980608e-01 6.43480480e-01 1.48078978e-01 -8.96472335e-02
7.46056914e-01 -6.57596648e-01 5.90326011e-01 1.02198698e-01
-1.90663949e-01 8.96113992e-01 4.03307855e-01 -1.99132517e-01
-8.56669724e-01 -8.39129388e-01 -1.78287383e-02 1.41507483e+00
-1.93281561e-01 -4.69159186e-01 -7.61198461e-01 -5.24762571e-01
-1.42628089e-01 4.67624575e-01 -7.73433924e-01 -8.16668123e-02
-6.32243395e-01 -5.60719073e-02 2.46807799e-01 5.17632425e-01
9.50952470e-02 -1.18660712e+00 -5.49935579e-01 8.29468668e-02
-1.10534631e-01 -1.02749789e+00 -6.08123899e-01 -4.35713857e-01
-5.49685478e-01 -7.51438618e-01 -9.74301457e-01 -6.71760678e-01
2.99841523e-01 3.14409770e-02 9.28169072e-01 -2.26740003e-01
-2.82716066e-01 6.74010873e-01 -3.15286636e-01 -3.12144458e-01
-2.43636549e-01 -1.87165841e-01 2.96767622e-01 4.71004218e-01
1.53107988e-02 -1.04828501e+00 -5.34629226e-01 2.69252360e-01
-6.69768095e-01 6.22727349e-02 4.01079923e-01 8.53362560e-01
6.62880182e-01 -4.57267582e-01 8.08099031e-01 -5.90293050e-01
3.78305793e-01 -6.30113602e-01 -2.83664912e-01 -1.25794858e-01
-1.63533524e-01 -2.29051009e-01 3.97472948e-01 -8.84847820e-01
-1.16544425e+00 2.42710739e-01 -5.27014673e-01 -1.07817543e+00
-2.87649691e-01 2.85172999e-01 -5.37630141e-01 3.42525542e-02
2.94823110e-01 9.72625092e-02 2.68405497e-01 -5.65571725e-01
3.59396070e-01 6.60007238e-01 5.00875235e-01 -4.65149939e-01
5.39998174e-01 4.52348530e-01 7.27308989e-02 -1.24849105e+00
-4.63195294e-01 -4.02157381e-02 -5.56573451e-01 -6.21263385e-01
7.80125141e-01 -1.01079249e+00 -7.17077255e-01 3.82336080e-01
-1.16472805e+00 -3.60680312e-01 -2.62267262e-01 3.04031581e-01
-1.06918132e+00 3.69636208e-01 -4.99303192e-01 -1.27822316e+00
-2.23640084e-01 -1.15024698e+00 1.49083638e+00 -8.75397250e-02
-4.03034002e-01 -6.67970061e-01 7.90863931e-02 1.01171479e-01
3.12632591e-01 3.93171728e-01 7.47982264e-01 -3.74738812e-01
-3.37463796e-01 1.13684461e-01 1.17774092e-01 1.21613555e-01
-6.72966540e-02 7.39548057e-02 -1.46307230e+00 -2.09358245e-01
-4.48900238e-02 -4.76680487e-01 5.07462740e-01 5.99780023e-01
1.32248044e+00 -3.98433000e-01 -8.99658725e-02 5.49609900e-01
7.94064462e-01 -9.41022560e-02 4.28910166e-01 -2.56363660e-01
6.42833591e-01 1.07200027e+00 7.77666926e-01 7.64030635e-01
4.22305673e-01 9.39748108e-01 7.16271460e-01 -3.13876979e-02
-3.16884965e-01 -5.56446373e-01 3.18373919e-01 9.69117045e-01
-1.33120090e-01 -9.96684209e-02 -6.42896891e-01 6.25440776e-01
-1.59504080e+00 -9.80715454e-01 4.17669296e-01 2.00066471e+00
6.99324906e-01 -1.17076389e-01 2.86026508e-01 3.00197393e-01
5.90729415e-01 3.66242886e-01 -4.83579427e-01 -2.84906209e-01
2.38943510e-02 1.05476730e-01 -2.74650842e-01 5.95180869e-01
-9.84638751e-01 9.09982026e-01 5.95290041e+00 8.50480080e-01
-1.26226938e+00 2.02810943e-01 5.11967659e-01 -3.06334674e-01
-3.71674448e-01 -3.23478669e-01 -5.27446866e-01 4.97816205e-01
1.01779079e+00 1.90741315e-01 2.01661631e-01 7.31735945e-01
4.43464339e-01 3.58486921e-01 -1.17387307e+00 1.18595374e+00
2.22190976e-01 -1.06710625e+00 6.72798082e-02 2.50615478e-01
3.55107039e-01 -4.24572945e-01 4.42001939e-01 1.91776156e-01
-2.67563701e-01 -1.13649368e+00 1.04114032e+00 8.29719484e-01
1.22808206e+00 -8.18653226e-01 1.82803839e-01 3.29713643e-01
-1.42913043e+00 -8.96161571e-02 3.07239238e-02 4.41242270e-02
3.15953791e-01 1.85039967e-01 -4.47901398e-01 3.52785110e-01
8.45893323e-01 8.26848567e-01 -2.68855616e-02 6.99576676e-01
1.92006826e-02 4.40119714e-01 -1.75530255e-01 2.97210515e-01
-6.44911081e-02 1.09055839e-01 9.27312493e-01 1.10619760e+00
5.01290679e-01 2.28751898e-01 1.06546935e-02 7.52756178e-01
6.19674213e-02 2.92145193e-01 -8.09394956e-01 -8.73033330e-03
4.11993802e-01 1.01898396e+00 -1.63384721e-01 -7.61854574e-02
-3.99612725e-01 9.40643311e-01 1.02245688e-01 4.22931463e-01
-6.34863079e-01 3.04010734e-02 9.78422225e-01 4.09871697e-01
4.02071148e-01 -2.05211550e-01 -5.36010647e-03 -8.81766319e-01
-1.51329651e-01 -9.21549737e-01 2.26483047e-01 -9.58631933e-01
-1.29140306e+00 8.52184594e-01 -3.55158709e-02 -1.31101286e+00
-6.90820992e-01 -2.69853771e-01 -4.24383283e-01 7.31323719e-01
-1.55518913e+00 -1.45690262e+00 -1.73500121e-01 6.51388943e-01
1.07383049e+00 -1.43598437e-01 9.05929625e-01 4.74704146e-01
-3.13426763e-01 7.72785664e-01 -4.83018547e-01 -3.34913433e-01
8.61885726e-01 -7.07863986e-01 1.49612248e-01 3.34927410e-01
-3.94263193e-02 3.08819443e-01 9.39639449e-01 -5.37737727e-01
-1.50100935e+00 -9.83952343e-01 9.57178473e-01 -1.35758847e-01
4.63755995e-01 -5.62353969e-01 -1.18401432e+00 4.67604131e-01
1.73617885e-01 2.68151253e-01 7.91240990e-01 -5.74590936e-02
-3.54521573e-01 -6.86472952e-02 -1.05154431e+00 5.77805281e-01
1.29498780e+00 -6.83220506e-01 -4.10758942e-01 5.10489941e-03
6.66211247e-01 -5.01748085e-01 -8.48466635e-01 4.95381951e-01
8.09094548e-01 -1.01717734e+00 1.06758845e+00 -5.97136021e-01
5.16231775e-01 -2.79144421e-02 -2.49130994e-01 -1.14440536e+00
-2.29186803e-01 -8.64503443e-01 -3.72357517e-01 1.54672766e+00
-3.42496037e-02 -1.73379824e-01 7.50289977e-01 2.09528372e-01
-7.08365142e-02 -8.50598574e-01 -1.29375863e+00 -3.44037265e-01
5.77932829e-03 -6.49197936e-01 6.71506226e-01 9.13554430e-01
-2.39884816e-02 3.16395998e-01 -7.89221108e-01 2.67893195e-01
4.62673604e-01 -4.94542532e-02 6.78031683e-01 -1.29740107e+00
-2.07628310e-01 -4.40122604e-01 -3.94207567e-01 -1.07688713e+00
6.99068666e-01 -5.38013279e-01 1.68672696e-01 -8.05915952e-01
-1.21553063e-01 -1.17881380e-01 6.81899264e-02 1.94216385e-01
6.42359555e-02 2.49646828e-01 1.50791228e-01 -7.95874149e-02
-3.32926035e-01 1.13764727e+00 1.21312737e+00 1.28475249e-01
-2.72856951e-01 1.22337915e-01 -3.15894395e-01 8.78751874e-01
3.02226186e-01 -1.62851572e-01 -6.01464152e-01 -2.26487994e-01
-2.83003636e-02 3.96004915e-01 3.55994582e-01 -8.62303734e-01
1.48541182e-01 -1.24442779e-01 3.40218335e-01 -6.42912745e-01
1.01257730e+00 -9.33923542e-01 1.16849601e-01 -7.67155886e-02
-5.77324390e-01 -5.73495850e-02 2.21146405e-01 7.09492862e-01
-2.48981044e-01 2.67170429e-01 8.33853006e-01 -3.98914590e-02
-3.20968270e-01 5.61156631e-01 -4.92614537e-01 1.17088016e-02
7.25021303e-01 -2.87520230e-01 4.05353159e-01 -7.68981874e-01
-9.13131952e-01 7.66990483e-02 4.12681103e-01 6.67847395e-01
9.21406329e-01 -1.64794469e+00 -7.50155449e-01 6.34640217e-01
-6.15022797e-03 -1.16794661e-01 6.18480146e-01 7.94242620e-01
2.06674322e-01 1.94079295e-01 -2.65851915e-01 -7.73011088e-01
-1.53635192e+00 6.10717118e-01 2.95975864e-01 2.11292788e-01
-6.25274181e-01 8.77334833e-01 3.42369050e-01 -1.22149803e-01
5.91709197e-01 6.10213317e-02 -4.89276588e-01 2.74751842e-01
6.50694847e-01 2.78009385e-01 -1.87735036e-01 -1.28943038e+00
-3.39084417e-01 9.43792403e-01 2.84293354e-01 -2.13615447e-01
1.35155904e+00 -4.36853439e-01 2.01015666e-01 7.25832522e-01
1.24870312e+00 1.84645019e-02 -1.73017848e+00 -3.05491000e-01
-4.49764550e-01 -5.76254427e-01 6.03925660e-02 -1.91264465e-01
-1.10110819e+00 1.34114528e+00 7.00213492e-01 4.09711860e-02
1.32742727e+00 -1.49535611e-02 5.71683705e-01 -3.90721351e-01
7.50009865e-02 -8.28522205e-01 3.23432535e-01 4.10154611e-01
1.31939137e+00 -1.00802386e+00 -1.95033967e-01 -6.34910405e-01
-7.64822721e-01 9.47339118e-01 6.08207285e-01 5.31553403e-02
9.01019692e-01 2.79887199e-01 -2.90219914e-02 -8.87918845e-02
-9.78082955e-01 -2.99739510e-01 5.08007348e-01 8.49442184e-01
5.78121662e-01 -5.06396294e-01 6.86183423e-02 7.05799520e-01
-1.87016800e-01 3.42610814e-02 8.36109743e-02 6.80811048e-01
-1.49211824e-01 -8.07133734e-01 -4.34129655e-01 -1.05101429e-01
-5.29616237e-01 -1.33153982e-03 -2.77105212e-01 6.01386249e-01
6.99402392e-02 8.27485681e-01 3.13715190e-01 -3.20495039e-01
4.27104980e-01 4.05525386e-01 4.89816308e-01 -3.65558773e-01
-2.62523443e-01 7.02618718e-01 -9.10828710e-02 -8.07934523e-01
-5.86143255e-01 -8.11262906e-01 -8.81066740e-01 -2.26712301e-02
-2.98825018e-02 -2.22481519e-01 6.49660707e-01 6.85788631e-01
6.67054534e-01 5.83427131e-01 8.17071319e-01 -1.61005819e+00
-3.59487444e-01 -1.05085051e+00 -5.46305478e-01 3.81659389e-01
7.48310506e-01 -1.07148075e+00 -4.33725566e-01 1.18762553e-01] | [13.202736854553223, -0.40648630261421204] |
2268a0e0-6af8-4d4b-97c3-19ab9d325de1 | 190503445 | 1905.03445 | null | https://arxiv.org/abs/1905.03445v1 | https://arxiv.org/pdf/1905.03445v1.pdf | Two-Stage Convolutional Neural Network Architecture for Lung Nodule Detection | Early detection of lung cancer is an effective way to improve the survival rate of patients. It is a critical step to have accurate detection of lung nodules in computed tomography (CT) images for the diagnosis of lung cancer. However, due to the heterogeneity of the lung nodules and the complexity of the surrounding environment, robust nodule detection has been a challenging task. In this study, we propose a two-stage convolutional neural network (TSCNN) architecture for lung nodule detection. The CNN architecture in the first stage is based on the improved UNet segmentation network to establish an initial detection of lung nodules. Simultaneously, in order to obtain a high recall rate without introducing excessive false positive nodules, we propose a novel sampling strategy, and use the offline hard mining idea for training and prediction according to the proposed cascaded prediction method. The CNN architecture in the second stage is based on the proposed dual pooling structure, which is built into three 3D CNN classification networks for false positive reduction. Since the network training requires a significant amount of training data, we adopt a data augmentation method based on random mask. Furthermore, we have improved the generalization ability of the false positive reduction model by means of ensemble learning. The proposed method has been experimentally verified on the LUNA dataset. Experimental results show that the proposed TSCNN architecture can obtain competitive detection performance. | ['Chih-Cheng Hung', 'Enmin Song', 'Xiangyang Xu', 'Tengying Liu', 'Renchao Jin', 'Hong Liu', 'Haichao Cao', 'Guangzhi Ma'] | 2019-05-09 | null | null | null | null | ['unet-segmentation', 'lung-nodule-detection'] | ['computer-vision', 'medical'] | [ 2.32721061e-01 3.19720618e-02 -8.89762193e-02 6.65318817e-02
-3.16399604e-01 1.25018835e-01 2.84054816e-01 -1.42127812e-01
-5.22585571e-01 2.92619407e-01 -2.66084105e-01 -3.42727691e-01
-1.18771471e-01 -1.08945465e+00 -2.95872182e-01 -8.87700021e-01
1.28268734e-01 4.21091527e-01 7.93585598e-01 3.05496380e-02
-1.07650585e-01 6.31791115e-01 -1.23846042e+00 3.56784672e-01
9.08680201e-01 1.18971431e+00 5.58199525e-01 3.22249442e-01
-2.01388732e-01 7.31875300e-01 -1.33027464e-01 7.97600895e-02
4.39863324e-01 -6.02900922e-01 -4.89899248e-01 2.26059780e-01
-3.09512585e-01 -4.42350566e-01 -2.88646966e-01 9.93850350e-01
4.60576236e-01 -1.12940960e-01 5.76959252e-01 -7.73691118e-01
6.23781607e-02 5.18394589e-01 -4.41970795e-01 2.67592996e-01
-5.59601426e-01 8.81110430e-02 6.68226182e-01 -1.03997970e+00
1.19421361e-02 7.78274119e-01 6.81383431e-01 4.73635674e-01
-7.20918357e-01 -8.70647371e-01 -2.10251123e-01 3.70443702e-01
-1.49797142e+00 -7.92665221e-03 4.38450217e-01 -3.77135843e-01
2.84495890e-01 2.60341018e-01 9.46484983e-01 5.72200954e-01
1.77585840e-01 4.90552217e-01 7.82574892e-01 -2.73977727e-01
-8.15528035e-02 3.98906618e-01 -7.85623714e-02 9.77092326e-01
6.89875901e-01 3.06896809e-02 4.67922240e-01 5.70888110e-02
1.11779332e+00 5.65760612e-01 -1.96925834e-01 -1.85391709e-01
-1.05836368e+00 8.24214995e-01 1.17768347e+00 8.27772200e-01
-3.22246969e-01 -9.44980093e-06 2.37492412e-01 -3.98960084e-01
2.98675656e-01 2.85533190e-01 -9.04200152e-02 7.05536544e-01
-9.55391109e-01 -1.27657399e-01 6.09396219e-01 3.55599463e-01
5.76988280e-01 -1.62563175e-01 -5.62961519e-01 7.13954449e-01
3.11857820e-01 3.13514590e-01 9.28052425e-01 -2.15625629e-01
3.22379261e-01 9.68379200e-01 -2.10432038e-01 -9.09374177e-01
-5.08087397e-01 -1.01799476e+00 -1.40792727e+00 8.09313804e-02
1.63214579e-01 -8.22105333e-02 -1.11552215e+00 1.10343719e+00
2.69989341e-01 3.31979662e-01 -1.83829471e-01 9.22625244e-01
8.52212310e-01 4.55451012e-01 4.28032428e-02 -3.26753765e-01
1.45572615e+00 -1.09429479e+00 -4.68131840e-01 -7.22556412e-02
6.17903769e-01 -4.41421926e-01 5.85689485e-01 4.86590788e-02
-5.96194863e-01 -7.45876729e-01 -1.01552916e+00 2.38704041e-01
1.48977069e-02 9.48638678e-01 3.54965746e-01 5.12034595e-01
-6.20235682e-01 3.91627818e-01 -1.04177511e+00 -3.65560114e-01
6.70475423e-01 5.77127576e-01 -6.56485185e-02 -3.04608196e-02
-9.68797505e-01 7.40377247e-01 8.98382187e-01 4.51957196e-01
-7.26613820e-01 -3.61224115e-01 -3.74832898e-01 3.97217929e-01
6.40964687e-01 -7.17648089e-01 1.16985166e+00 -8.49007547e-01
-1.20073104e+00 4.03389603e-01 5.25337793e-02 -5.65685987e-01
7.23932624e-01 2.62517422e-01 -9.70563218e-02 1.63960949e-01
-8.61907005e-03 3.48742932e-01 6.59951329e-01 -7.80449092e-01
-7.75869012e-01 -2.24084526e-01 -4.11713570e-01 2.69288898e-01
-5.10937572e-01 -3.68241757e-01 -5.01907349e-01 -5.55479705e-01
4.61802840e-01 -1.02176833e+00 -7.31032550e-01 3.51997651e-02
-3.82966220e-01 -2.13429838e-01 9.80244696e-01 -5.69465458e-01
1.38259768e+00 -2.09392357e+00 -1.45780697e-01 5.58378160e-01
5.00711977e-01 4.17675883e-01 3.97962362e-01 -3.45581740e-01
-9.63322520e-02 3.48720580e-01 -3.20611894e-01 5.64908832e-02
-6.37612164e-01 2.21168362e-02 3.78700554e-01 6.54783994e-02
3.34854007e-01 8.19757879e-01 -4.70232219e-01 -1.00775242e+00
2.82546878e-01 2.41142556e-01 -3.99866372e-01 4.02940452e-01
-1.85903251e-01 5.52771568e-01 -7.76666939e-01 4.58216339e-01
6.41572773e-01 -6.06045008e-01 1.34490773e-01 -4.17220712e-01
-1.93013474e-01 -1.01959758e-01 -1.03104997e+00 1.03928828e+00
-3.40370089e-01 3.90598029e-02 -1.11320674e-01 -8.82875621e-01
9.31915522e-01 5.19672692e-01 5.39626241e-01 -4.85221505e-01
4.16279316e-01 5.57573020e-01 5.00661433e-01 -6.90812349e-01
-1.63525745e-01 -2.86517948e-01 3.33181947e-01 1.56796873e-01
-2.61883408e-01 -2.87064933e-03 1.87301055e-01 -1.94365799e-01
1.12748778e+00 -4.53282535e-01 4.43419993e-01 -8.57273042e-02
9.23192978e-01 1.11964434e-01 6.09664619e-01 4.52929944e-01
-1.59887411e-02 4.87566203e-01 2.54920006e-01 -3.59738231e-01
-8.87925148e-01 -5.52120864e-01 -2.59131998e-01 3.84209067e-01
3.06968465e-02 1.78308442e-01 -5.79086304e-01 -8.98995280e-01
-2.94207007e-01 1.80754915e-01 -6.82182074e-01 -1.89728737e-01
-7.43746161e-01 -1.15702879e+00 3.67919415e-01 5.27355373e-01
1.28811395e+00 -9.81805205e-01 -4.92110759e-01 2.24094674e-01
-2.46076420e-01 -8.03480983e-01 -1.37107551e-01 2.98636258e-01
-1.09686601e+00 -1.25254869e+00 -7.61219978e-01 -9.38914299e-01
7.36711383e-01 4.31065351e-01 5.47257483e-01 6.36664867e-01
-4.45191205e-01 -2.86603153e-01 -2.63160229e-01 -3.77093554e-01
-4.22111750e-01 4.34627086e-01 -3.51593316e-01 1.62337571e-01
9.89182815e-02 -2.55851716e-01 -7.21520364e-01 3.92373621e-01
-8.87068510e-01 1.94257155e-01 1.53864253e+00 8.74713659e-01
7.30315387e-01 6.42051518e-01 4.39289987e-01 -9.75306451e-01
1.94395602e-01 -6.01910472e-01 -5.17518818e-01 1.50742665e-01
-4.46445853e-01 -8.91406536e-02 5.94604969e-01 -3.55730385e-01
-1.11840665e+00 3.49204093e-01 -2.34028786e-01 -4.41187829e-01
5.17782830e-02 4.39746469e-01 -3.25742476e-02 -2.28537962e-01
5.77452123e-01 2.16538087e-01 2.33972549e-01 -3.41619849e-01
-3.10340434e-01 4.97032911e-01 1.77205518e-01 2.80717403e-01
9.89376664e-01 3.29105765e-01 3.12478960e-01 -6.79479182e-01
-8.35713685e-01 -5.73757529e-01 -6.20615125e-01 -2.02416807e-01
1.14029598e+00 -9.04672921e-01 -3.76215696e-01 2.43923277e-01
-8.90876591e-01 3.05040833e-02 -6.96949810e-02 9.28138912e-01
1.99084431e-02 2.63743818e-01 -5.39875269e-01 -5.63091755e-01
-5.83027482e-01 -1.08936179e+00 5.19996226e-01 4.68375027e-01
3.20982218e-01 -6.67688489e-01 -1.65944085e-01 1.32104784e-01
6.26677275e-01 9.79771167e-02 9.31206107e-01 -1.09001124e+00
-9.02828872e-01 -4.87353861e-01 -5.13074517e-01 5.12010455e-01
2.55420834e-01 -8.60684216e-02 -6.67870700e-01 -1.94195524e-01
4.06912416e-01 -3.52023281e-02 1.15126491e+00 3.94144595e-01
1.62399387e+00 -2.14702174e-01 -7.89493620e-01 4.55602169e-01
1.49444485e+00 3.00541878e-01 5.54770350e-01 2.35150516e-01
7.59128511e-01 1.42820433e-01 5.39116919e-01 2.75826365e-01
-2.36829966e-01 2.64220685e-01 6.94642127e-01 -3.26167226e-01
-2.33241752e-01 9.01186839e-02 -1.45156235e-01 7.58634090e-01
-3.83743346e-01 -4.15578559e-02 -8.24195087e-01 2.88856328e-01
-1.66035664e+00 -9.62866962e-01 -5.05583704e-01 2.13727355e+00
5.23984730e-01 3.38965356e-01 -1.08792596e-01 2.84770012e-01
8.05472672e-01 -2.29879588e-01 -2.89168894e-01 3.43869537e-01
3.66761118e-01 3.65723997e-01 4.82316315e-01 -8.07025656e-02
-1.31082916e+00 3.03745657e-01 4.83248281e+00 1.10167360e+00
-1.37061691e+00 1.73046947e-01 7.74697185e-01 1.44550383e-01
1.36978135e-01 -3.43099505e-01 -8.70399535e-01 3.87575775e-01
3.42621416e-01 5.63685447e-02 -1.06549554e-01 8.80571663e-01
2.72100687e-01 -9.26469788e-02 -7.57205963e-01 6.57209218e-01
-1.34894833e-01 -1.15055382e+00 3.58347557e-02 6.08150810e-02
5.53521156e-01 -9.11971461e-03 -7.87407011e-02 3.21816236e-01
-2.82485247e-01 -1.03382802e+00 1.17389977e-01 5.52550912e-01
6.31751657e-01 -7.00042903e-01 1.35542953e+00 7.24634588e-01
-1.33797586e+00 -3.70330244e-01 -4.49896663e-01 3.11486334e-01
-3.69242460e-01 9.10264432e-01 -1.51874745e+00 6.50395334e-01
7.08558977e-01 5.47634065e-01 -9.17568803e-01 1.52691126e+00
-1.52612291e-02 6.93291247e-01 -4.70838964e-01 -3.39036077e-01
1.01328343e-01 -1.47647616e-02 3.75346750e-01 9.07930613e-01
6.92917883e-01 2.37301201e-01 3.68775994e-01 9.49153602e-01
-2.16887861e-01 3.55752140e-01 -4.19570178e-01 1.53777331e-01
3.02094072e-01 1.67340910e+00 -1.08690333e+00 -3.49186659e-01
-2.05021515e-01 6.52535021e-01 2.74578426e-02 -1.43770605e-01
-9.95777667e-01 -2.64488697e-01 -4.45794076e-01 3.52667719e-01
4.50139582e-01 2.23420545e-01 -2.61846453e-01 -9.58301306e-01
-2.94058956e-02 -4.95939791e-01 2.82532394e-01 -2.13321820e-01
-9.32627022e-01 7.12687194e-01 -3.01000327e-01 -1.54450798e+00
5.11879660e-02 -5.55444360e-01 -1.00120556e+00 7.57870317e-01
-1.19574428e+00 -1.06658757e+00 -9.25289989e-01 4.39815640e-01
4.72092241e-01 -8.29117298e-02 6.61592782e-01 3.47791404e-01
-9.86932278e-01 2.68328279e-01 -1.79205939e-01 2.93002069e-01
4.23056632e-01 -1.00425696e+00 -3.53823721e-01 7.38977551e-01
-3.83658856e-01 -1.57279503e-02 5.15838340e-02 -7.55537927e-01
-6.52462840e-01 -1.66689694e+00 4.48015779e-01 8.98849443e-02
2.22586870e-01 4.11937349e-02 -1.05139041e+00 5.22807479e-01
-2.49080285e-01 3.01671892e-01 4.66584206e-01 -5.01257896e-01
3.28874081e-01 -3.13393325e-01 -1.19872844e+00 3.75588208e-01
4.49307561e-01 1.16696261e-01 -3.49142045e-01 4.47354525e-01
5.75467348e-01 -1.99648231e-01 -6.47025526e-01 9.48281646e-01
4.23834771e-01 -9.75211442e-01 8.30317318e-01 8.91628116e-02
3.11103195e-01 -4.63473320e-01 3.94946277e-01 -7.78976679e-01
-8.09709728e-01 4.35437202e-01 2.61116713e-01 7.83471942e-01
5.29165268e-01 -4.13295835e-01 1.15534556e+00 2.44081542e-01
-1.23487316e-01 -1.10562587e+00 -7.47598588e-01 -4.79214728e-01
-3.08616877e-01 -4.63698879e-02 2.18788028e-01 4.63577151e-01
-5.95557630e-01 2.92273849e-01 1.07523359e-01 2.14226812e-01
4.82372671e-01 -1.09003097e-01 4.57367212e-01 -1.43429482e+00
-4.84376132e-01 -5.17062664e-01 -2.20308706e-01 -7.15102017e-01
-5.36025763e-01 -9.77205992e-01 1.48978293e-01 -1.60236585e+00
5.90948582e-01 -5.56370854e-01 -5.68479359e-01 3.20610434e-01
-4.32995796e-01 4.20524359e-01 3.45767965e-03 5.02993584e-01
-4.01609510e-01 4.28586096e-01 1.51200628e+00 -1.17429033e-01
-2.16823488e-01 7.68748939e-01 -3.54329616e-01 8.06957364e-01
9.16173458e-01 -4.03937697e-01 -2.79335529e-01 -1.43170372e-01
-1.60562843e-01 1.45550638e-01 6.08686328e-01 -1.42220283e+00
3.07639718e-01 2.55507082e-02 8.68229091e-01 -8.37326825e-01
1.59955561e-01 -1.15782917e+00 8.66161138e-02 1.36844814e+00
-1.76721383e-02 -4.53128785e-01 1.42431244e-01 6.70232058e-01
-2.34960660e-01 -4.64312226e-01 9.56275403e-01 -5.13813674e-01
-4.49099928e-01 5.40950894e-01 -4.16970938e-01 -5.38041949e-01
1.32055187e+00 -1.16466843e-01 1.78423017e-01 2.03880355e-01
-7.35898852e-01 3.04879785e-01 -1.03033982e-01 -1.27982363e-01
5.79887211e-01 -1.42126596e+00 -6.94844127e-01 2.64823884e-01
-1.00664003e-02 5.29377878e-01 2.40711808e-01 1.22672856e+00
-7.42463648e-01 4.03180629e-01 2.04792581e-02 -6.97402656e-01
-1.30909598e+00 3.53129625e-01 7.91604638e-01 -8.39465261e-01
-3.91084433e-01 8.06554019e-01 3.31887513e-01 -1.36691645e-01
9.90206748e-02 -5.32832682e-01 -6.53990805e-01 -1.98850170e-01
2.75097817e-01 3.42728019e-01 2.86561817e-01 -6.16544634e-02
-1.88810214e-01 3.57034981e-01 -2.42613778e-01 2.69144654e-01
1.06509006e+00 3.04545820e-01 -3.68510962e-01 1.33975267e-01
9.17423904e-01 -1.23177923e-01 -6.50252759e-01 -9.81712714e-02
-5.81467897e-02 -6.62727058e-02 1.62304118e-01 -6.68026447e-01
-1.19098473e+00 8.36973965e-01 9.21792388e-01 2.01578721e-01
1.22512019e+00 -1.89979166e-01 7.83934534e-01 5.24287403e-01
-6.72554448e-02 -6.31869495e-01 3.54685903e-01 2.33767882e-01
6.43402219e-01 -1.43194759e+00 5.99746890e-02 -7.28782237e-01
-1.75435811e-01 1.34153712e+00 1.03424394e+00 -2.38622710e-01
8.22472394e-01 -7.53151178e-02 -3.27923477e-01 -6.00793883e-02
-5.15119076e-01 -3.94696504e-01 3.49060059e-01 -5.12703061e-02
2.62690246e-01 7.86371008e-02 -4.97609645e-01 9.95358109e-01
1.87664196e-01 1.71849027e-01 2.14668661e-01 6.44802332e-01
-9.60656643e-01 -8.66633117e-01 -4.61588591e-01 9.41064835e-01
-5.11978447e-01 1.19001549e-02 -1.67139888e-01 9.93546546e-01
5.95105708e-01 5.58575630e-01 1.28585249e-01 -4.50498044e-01
2.72967875e-01 -4.11682203e-03 2.65995175e-01 -8.94218802e-01
-6.39106870e-01 3.32791716e-01 -3.01436037e-01 -8.22251141e-02
-3.16634268e-01 -2.76916444e-01 -1.39735973e+00 2.46714890e-01
-7.54813075e-01 2.60942429e-01 3.99733394e-01 9.76236463e-01
-5.42285591e-02 9.28549290e-01 8.97962630e-01 -5.90903819e-01
-6.63105011e-01 -1.19559062e+00 -4.88122702e-01 4.38142531e-02
-9.22659878e-03 -4.49447602e-01 -3.42512578e-01 -2.97986299e-01] | [15.353765487670898, -2.1170554161071777] |
6e51cbe5-cb12-4d95-bf45-b79852ff5096 | multi-mention-learning-for-reading | 1711.00894 | null | http://arxiv.org/abs/1711.00894v2 | http://arxiv.org/pdf/1711.00894v2.pdf | Multi-Mention Learning for Reading Comprehension with Neural Cascades | Reading comprehension is a challenging task, especially when executed across
longer or across multiple evidence documents, where the answer is likely to
reoccur. Existing neural architectures typically do not scale to the entire
evidence, and hence, resort to selecting a single passage in the document
(either via truncation or other means), and carefully searching for the answer
within that passage. However, in some cases, this strategy can be suboptimal,
since by focusing on a specific passage, it becomes difficult to leverage
multiple mentions of the same answer throughout the document. In this work, we
take a different approach by constructing lightweight models that are combined
in a cascade to find the answer. Each submodel consists only of feed-forward
networks equipped with an attention mechanism, making it trivially
parallelizable. We show that our approach can scale to approximately an order
of magnitude larger evidence documents and can aggregate information at the
representation level from multiple mentions of each answer candidate across the
document. Empirically, our approach achieves state-of-the-art performance on
both the Wikipedia and web domains of the TriviaQA dataset, outperforming more
complex, recurrent architectures. | ['Tom Kwiatkowski', 'Swabha Swayamdipta', 'Ankur P. Parikh'] | 2017-11-02 | multi-mention-learning-for-reading-1 | https://openreview.net/forum?id=HyRnez-RW | https://openreview.net/pdf?id=HyRnez-RW | iclr-2018-1 | ['triviaqa'] | ['miscellaneous'] | [ 1.11086309e-01 8.78067762e-02 -3.23502533e-02 -1.03401177e-01
-1.30734742e+00 -8.87567759e-01 5.00720978e-01 7.51937270e-01
-6.18298352e-01 5.85979879e-01 5.92206240e-01 -6.65715516e-01
-2.47196078e-01 -7.85560369e-01 -9.89276648e-01 -2.00161859e-01
2.58679479e-01 5.19428253e-01 4.00657952e-01 -2.29184642e-01
5.08743227e-01 5.42081706e-02 -1.34330630e+00 4.71652359e-01
9.47533727e-01 7.06666827e-01 2.63246804e-01 8.22924137e-01
-3.10856044e-01 8.69571209e-01 -6.89364970e-01 -7.71047294e-01
-9.29459855e-02 -4.55052614e-01 -1.10917711e+00 -2.79301643e-01
9.44980860e-01 -5.59630692e-01 -2.03641862e-01 8.11658144e-01
3.27647269e-01 2.40010932e-01 5.07541180e-01 -3.72109532e-01
-7.66382158e-01 1.03100967e+00 -4.40860927e-01 5.82233191e-01
6.36947215e-01 -1.73537508e-01 1.45741129e+00 -1.06846404e+00
4.27305669e-01 1.06355584e+00 5.58211148e-01 2.78245538e-01
-1.11841428e+00 -1.93864420e-01 6.54191434e-01 3.79892617e-01
-7.85533309e-01 -3.51096392e-01 4.04546380e-01 -1.32956579e-01
1.36318088e+00 1.28545672e-01 5.02670705e-01 9.26505148e-01
1.29838198e-01 1.02721047e+00 6.27738595e-01 -5.23418009e-01
3.84990200e-02 -3.29425126e-01 6.56555235e-01 8.48190367e-01
2.86800861e-01 -5.60934424e-01 -7.89767146e-01 -1.54517889e-01
1.14879228e-01 -4.72098403e-02 -3.67908001e-01 5.88266402e-02
-1.18557930e+00 8.34273696e-01 5.01883566e-01 1.65766701e-01
-6.32547081e-01 1.80755407e-01 3.22515517e-01 5.60276926e-01
2.01498270e-01 6.36091232e-01 -5.38147688e-01 -3.13097000e-01
-1.15007782e+00 5.21622539e-01 1.03819633e+00 5.44540048e-01
5.43400466e-01 -4.91276205e-01 -2.68397748e-01 7.72877038e-01
8.99083763e-02 1.33229196e-01 4.38109636e-01 -9.75388110e-01
8.23182523e-01 6.85397387e-01 3.20166975e-01 -9.10406649e-01
-4.65532929e-01 -5.44009924e-01 -3.32428694e-01 -1.69928730e-01
7.07893908e-01 -2.67914295e-01 -8.41021955e-01 1.86571348e+00
1.52687192e-01 -2.69396901e-01 4.78338730e-03 9.00826871e-01
7.99670458e-01 8.40013385e-01 -5.26361465e-02 -1.19397882e-02
1.40362489e+00 -1.30796063e+00 -5.00109851e-01 -5.39459169e-01
5.02265990e-01 -7.03459382e-01 1.16580653e+00 4.62595522e-01
-1.52898419e+00 -3.34242702e-01 -1.08177602e+00 -4.79537487e-01
-3.38059753e-01 1.04475198e-02 3.46084863e-01 8.78727436e-02
-9.34587240e-01 4.56979036e-01 -5.98454416e-01 -3.49932462e-01
3.77870798e-02 7.31369853e-02 -9.38821211e-02 -4.22806323e-01
-1.27200878e+00 1.21768701e+00 1.48653492e-01 2.68552870e-01
-6.24266386e-01 -6.65459514e-01 -6.64141536e-01 5.21792650e-01
4.27845538e-01 -9.12529051e-01 1.53973377e+00 -8.05052638e-01
-1.27570605e+00 5.85945308e-01 -5.91643155e-01 -5.29440999e-01
2.18737349e-01 -7.21380413e-01 -1.49415448e-01 1.43049315e-01
4.76229936e-02 3.06453884e-01 6.67532086e-01 -7.07937241e-01
-4.78034198e-01 -2.83695072e-01 6.23899102e-01 3.09450597e-01
-2.71583408e-01 1.88587278e-01 -5.88742137e-01 -3.20683986e-01
2.01248497e-01 -7.29855835e-01 -2.95385551e-02 -3.29242319e-01
-3.64970773e-01 -5.79716265e-01 2.08518744e-01 -8.99326503e-01
1.51793349e+00 -1.91663396e+00 4.91666883e-01 1.22910783e-01
2.32240096e-01 1.81266993e-01 -2.34856606e-01 7.23096490e-01
3.13643575e-01 7.52882436e-02 -1.81297034e-01 -2.40879253e-01
8.28811750e-02 -9.72755626e-03 -5.31787574e-01 4.07273583e-02
2.18310818e-01 9.18737948e-01 -9.31511164e-01 -2.69554585e-01
-4.92734164e-01 1.80721432e-01 -6.66754127e-01 1.68036520e-02
-5.44986308e-01 1.57476636e-03 -5.25973916e-01 4.14008260e-01
1.90905258e-01 -7.40112424e-01 6.84721321e-02 3.17332484e-02
-3.75641026e-02 1.07554603e+00 -9.39770937e-01 1.76686227e+00
-5.98528922e-01 6.11860037e-01 -5.19474745e-02 -8.16006720e-01
5.17883420e-01 2.16354579e-01 -5.24477921e-02 -6.45551145e-01
-1.49300963e-01 5.55179596e-01 2.71044880e-01 -5.31391799e-01
6.70745850e-01 1.87508985e-01 -6.25726432e-02 8.95260394e-01
1.58002406e-01 2.25686878e-01 6.75032020e-01 5.49291611e-01
1.48793972e+00 -1.08783707e-01 1.58728465e-01 1.38221458e-01
4.62112218e-01 -9.16789919e-02 6.36352226e-02 1.17387331e+00
1.73800826e-01 6.22981369e-01 6.62549615e-01 -3.43596607e-01
-9.23768878e-01 -1.02844441e+00 7.92137831e-02 1.34591246e+00
-8.45377967e-02 -5.29622138e-01 -6.82070613e-01 -7.61678338e-01
9.84279364e-02 7.09766150e-01 -5.87796807e-01 -1.90756083e-01
-9.34120536e-01 -4.62417156e-01 4.14468408e-01 6.34512603e-01
4.34496015e-01 -1.06210756e+00 -9.10633802e-01 5.49916387e-01
-3.79287064e-01 -6.79733634e-01 -4.83436018e-01 3.44463199e-01
-7.95198321e-01 -9.13852394e-01 -7.65105128e-01 -6.22057974e-01
4.24602777e-01 1.92939147e-01 1.46460581e+00 4.85935807e-01
2.16639698e-01 3.46297532e-01 -2.54521251e-01 -1.88529968e-01
-1.16560720e-01 5.78876972e-01 -3.99656802e-01 -1.70449525e-01
4.76377368e-01 -2.76170522e-01 -7.93802202e-01 -7.23828077e-02
-6.16020918e-01 -3.01290751e-01 6.25658751e-01 1.05844080e+00
3.30603033e-01 -3.60574335e-01 8.81409824e-01 -8.36069286e-01
1.12146902e+00 -7.49996364e-01 -3.86084318e-01 5.90033889e-01
-4.37452227e-01 1.49102271e-01 6.62537992e-01 -3.36011618e-01
-9.27484155e-01 -2.67690539e-01 -2.42623568e-01 -4.51602042e-02
1.55734569e-01 1.14745140e+00 3.93702388e-01 3.51875305e-01
6.34462297e-01 2.66712278e-01 -2.51142055e-01 -5.64960897e-01
5.15966713e-01 4.09911871e-01 4.41840857e-01 -5.86699069e-01
4.22374845e-01 1.08947083e-02 -4.31995451e-01 -4.18601602e-01
-1.40346324e+00 -3.50115895e-01 -4.48895276e-01 -1.35514528e-01
6.03458941e-01 -7.05130279e-01 -5.81366777e-01 1.57181367e-01
-1.41134274e+00 -1.77890927e-01 8.50118101e-02 3.74542505e-01
-5.35518527e-02 1.02496535e-01 -9.82360542e-01 -4.69586790e-01
-4.68281865e-01 -1.01653981e+00 8.43764961e-01 4.48432893e-01
-5.08331597e-01 -7.22264826e-01 7.79648870e-02 3.84203017e-01
5.79710841e-01 -4.21053380e-01 1.33793402e+00 -9.16086137e-01
-6.95353448e-01 -4.82979476e-01 -1.48374885e-01 1.62091926e-02
-2.28722632e-01 4.09523547e-02 -6.69031143e-01 -1.16508991e-01
-5.00968844e-02 -8.02148044e-01 1.41303372e+00 9.25430804e-02
1.07655573e+00 -4.59288478e-01 -1.14258446e-01 -3.34221683e-02
1.02063620e+00 -2.71199942e-01 3.39433312e-01 4.98355180e-01
3.84714276e-01 6.36024714e-01 3.04140121e-01 1.57543674e-01
6.69008791e-01 3.92048538e-01 3.15404654e-01 3.47649455e-01
-1.02415495e-01 -2.96140194e-01 3.61391515e-01 9.48273540e-01
4.28014360e-02 -5.55248618e-01 -9.24113214e-01 7.62677431e-01
-1.69300270e+00 -1.04765725e+00 -1.47023931e-01 2.13732791e+00
9.55986619e-01 1.88169003e-01 -4.97756787e-02 5.32392114e-02
3.39627683e-01 2.22377896e-01 -7.18796372e-01 -6.25724435e-01
-1.62449013e-02 2.44329512e-01 3.41158919e-02 6.03773594e-01
-7.55520880e-01 6.47874475e-01 6.90851259e+00 4.33624685e-01
-1.06788945e+00 2.07769386e-02 4.78917241e-01 -4.38747376e-01
-5.48049212e-01 3.04223169e-02 -8.74016643e-01 3.56315166e-01
1.08570993e+00 5.17856255e-02 5.15378237e-01 3.45101446e-01
-1.78640813e-01 -4.73594278e-01 -1.13535404e+00 3.04746777e-01
3.01512659e-01 -1.23292828e+00 1.68715343e-01 -2.97012717e-01
6.57226205e-01 3.00429136e-01 5.02530038e-02 3.72943193e-01
2.90689349e-01 -1.06132758e+00 7.96433389e-01 4.66262490e-01
1.18648343e-01 -6.34637296e-01 5.71647108e-01 5.27300775e-01
-7.79941261e-01 -2.91663796e-01 -3.51096660e-01 -6.55783340e-02
2.45817482e-01 4.54676867e-01 -5.03105819e-01 3.32990855e-01
6.95138812e-01 3.49370152e-01 -7.47317076e-01 1.20982003e+00
-5.74558198e-01 8.19011509e-01 -4.24874663e-01 -3.93259078e-01
4.86062616e-01 1.75020352e-01 4.09117848e-01 1.12271571e+00
4.13169950e-01 7.89145902e-02 -7.08830133e-02 7.02752054e-01
-4.88801271e-01 1.65852547e-01 -2.80276358e-01 -5.20779286e-03
6.26808941e-01 1.03594279e+00 -3.46792459e-01 -3.88978332e-01
-8.19097877e-01 8.49258065e-01 1.04828393e+00 4.76786613e-01
-6.86563492e-01 -5.12526333e-01 1.46990307e-02 -1.87320799e-01
6.23171926e-01 -2.11726665e-01 -1.55668184e-01 -1.19475615e+00
5.03868759e-01 -1.12720311e+00 7.57276177e-01 -7.35313654e-01
-1.20473421e+00 5.03409028e-01 -3.30220878e-01 -7.42925882e-01
-3.68261456e-01 -5.05150318e-01 -8.12976599e-01 1.00699818e+00
-1.58186388e+00 -7.59766817e-01 -6.09859219e-03 2.92045861e-01
6.10924363e-01 2.15444162e-01 7.43683398e-01 3.11380714e-01
-4.84852642e-01 5.69875598e-01 6.34251609e-02 -7.23182186e-02
6.35029316e-01 -1.25753224e+00 2.05787286e-01 9.76272523e-01
4.45200920e-01 1.21599913e+00 6.03343070e-01 -4.57826883e-01
-1.45965648e+00 -5.20209551e-01 1.37598443e+00 -4.89017397e-01
8.65687788e-01 -1.41609728e-01 -1.38168812e+00 7.78856337e-01
5.48491180e-01 -3.37195575e-01 5.25813758e-01 6.75832510e-01
-6.50961637e-01 7.13711828e-02 -5.86751521e-01 7.70057023e-01
6.93903387e-01 -7.77487636e-01 -1.02537191e+00 2.66798288e-01
7.08661556e-01 -5.17663121e-01 -7.22540081e-01 1.68687060e-01
7.20692813e-01 -8.22295308e-01 6.98971748e-01 -8.22380424e-01
9.13528264e-01 -2.34834135e-01 -2.34027840e-02 -1.38641739e+00
-2.40256637e-01 -4.72114861e-01 -7.33822405e-01 9.58428860e-01
1.16053402e+00 -4.43872690e-01 4.89937335e-01 6.41201437e-01
-2.32199669e-01 -1.05915308e+00 -1.00536430e+00 -3.01199943e-01
3.66176307e-01 -1.81694746e-01 5.80480933e-01 4.27800089e-01
4.08829898e-01 6.62687719e-01 -1.36877000e-01 1.71078905e-01
1.79668337e-01 4.50991571e-01 4.01639909e-01 -9.59104061e-01
-4.73736376e-01 -7.08284080e-01 3.01154345e-01 -1.67277598e+00
5.93631901e-02 -9.21441913e-01 2.27520451e-01 -1.90362239e+00
2.29291275e-01 -1.47470802e-01 -5.10692656e-01 4.22079116e-01
-6.17832839e-01 -1.03536718e-01 7.33942315e-02 1.78827941e-01
-8.21859658e-01 2.96777636e-01 1.16848755e+00 -2.44309157e-01
9.95074436e-02 -1.48945078e-01 -8.71819913e-01 6.44247890e-01
6.50436580e-01 -3.56965393e-01 -2.85172462e-01 -1.11212540e+00
8.38312566e-01 1.49131536e-01 3.57898295e-01 -7.15097070e-01
6.18116260e-01 3.14034760e-01 4.97941136e-01 -6.49757028e-01
2.73123503e-01 -3.97373170e-01 -4.45955902e-01 3.22122365e-01
-8.38826418e-01 3.58881027e-01 2.13705540e-01 6.37096882e-01
-2.68083662e-01 -6.19112730e-01 3.13557446e-01 -1.75104231e-01
-2.80101955e-01 -8.82397965e-02 -5.02389252e-01 1.37264907e-01
3.70139927e-01 2.12029114e-01 -7.07648277e-01 -4.67361867e-01
-4.70566183e-01 3.95416707e-01 1.66843802e-01 3.93535733e-01
5.77935934e-01 -9.06381309e-01 -7.61380196e-01 -1.40762761e-01
6.40642792e-02 2.90396512e-02 1.56341597e-01 8.49868596e-01
-1.47549331e-01 5.79144299e-01 2.70262241e-01 -2.94714004e-01
-9.33038354e-01 3.87442231e-01 3.71929973e-01 -5.58681250e-01
-6.84970796e-01 1.00115705e+00 -1.12035252e-01 -2.56378412e-01
3.71228695e-01 -2.94579744e-01 -4.89717811e-01 3.27656806e-01
7.46780276e-01 1.73798636e-01 3.78802776e-01 -1.45589024e-01
-2.19560355e-01 5.65812349e-01 -5.40065289e-01 -1.98301405e-01
1.22748113e+00 -2.13231847e-01 -3.12323689e-01 5.67281544e-01
9.21416342e-01 2.57288367e-01 -1.02390039e+00 -3.65138859e-01
1.85048878e-01 -9.66108516e-02 -1.35988712e-01 -1.13582563e+00
-7.53467739e-01 8.34091306e-01 -1.61463946e-01 3.35590661e-01
9.01485324e-01 1.76258311e-01 9.09633994e-01 9.62172270e-01
1.70498788e-01 -8.76090348e-01 1.89791694e-01 8.75064373e-01
1.05899453e+00 -1.15555561e+00 -5.11869267e-02 1.31515369e-01
-3.34263504e-01 1.32943988e+00 6.19033337e-01 -1.42493099e-01
3.08419853e-01 -1.06957525e-01 -8.95266533e-02 -4.39430684e-01
-1.28929591e+00 1.16728500e-01 4.57994223e-01 -2.73790538e-01
8.25174034e-01 -2.57812619e-01 -4.45999950e-01 5.41366220e-01
-5.47721535e-02 -2.90112495e-01 5.69990456e-01 1.01302421e+00
-6.85891271e-01 -9.31667089e-01 -1.68902650e-01 8.42174053e-01
-7.82290995e-01 -5.37607908e-01 -3.38882864e-01 4.15111899e-01
-4.22561109e-01 1.13570666e+00 9.48653743e-03 2.34474093e-02
3.67230564e-01 4.80983824e-01 5.34692526e-01 -5.50145090e-01
-8.43652666e-01 -1.67214274e-01 3.60539109e-01 -5.06524980e-01
-1.26220003e-01 -7.12975383e-01 -1.26050019e+00 -3.19281220e-01
-2.51111746e-01 1.11430101e-01 4.87906456e-01 1.14363229e+00
4.70949560e-01 5.40403962e-01 1.28908515e-01 -3.50155085e-01
-8.02227497e-01 -1.05576634e+00 -4.16667573e-02 1.20049030e-01
5.98080635e-01 -4.71855581e-01 -3.06115508e-01 -2.25566268e-01] | [11.200199127197266, 8.03711986541748] |
ece4c0b9-0f71-4acc-8bb0-efc0bd004c2e | comparative-studies-of-detecting-abusive | 1808.10245 | null | http://arxiv.org/abs/1808.10245v1 | http://arxiv.org/pdf/1808.10245v1.pdf | Comparative Studies of Detecting Abusive Language on Twitter | The context-dependent nature of online aggression makes annotating large
collections of data extremely difficult. Previously studied datasets in abusive
language detection have been insufficient in size to efficiently train deep
learning models. Recently, Hate and Abusive Speech on Twitter, a dataset much
greater in size and reliability, has been released. However, this dataset has
not been comprehensively studied to its potential. In this paper, we conduct
the first comparative study of various learning models on Hate and Abusive
Speech on Twitter, and discuss the possibility of using additional features and
context data for improvements. Experimental results show that bidirectional GRU
networks trained on word-level features, with Latent Topic Clustering modules,
is the most accurate model scoring 0.805 F1. | ['Seunghyun Yoon', 'Younghun Lee', 'Kyomin Jung'] | 2018-08-30 | comparative-studies-of-detecting-abusive-1 | https://aclanthology.org/W18-5113 | https://aclanthology.org/W18-5113.pdf | ws-2018-10 | ['twitter-sentiment-analysis', 'abuse-detection'] | ['natural-language-processing', 'natural-language-processing'] | [-3.64216775e-01 -1.24377951e-01 -3.22955519e-01 -3.71901333e-01
-5.02880156e-01 -3.47613424e-01 6.93450987e-01 3.48792017e-01
-6.12256467e-01 7.86069930e-01 3.91027153e-01 -8.57135355e-02
-4.22138274e-02 -4.83487129e-01 -1.17890969e-01 -5.25991917e-01
-9.74146873e-02 4.64956373e-01 1.09245300e-01 -4.71123099e-01
3.29497635e-01 3.47420722e-01 -1.29637182e+00 4.30627733e-01
7.44922996e-01 5.01295805e-01 -3.19315374e-01 5.15347540e-01
-3.64110053e-01 1.10454023e+00 -1.15925598e+00 -9.54927444e-01
-2.43167222e-01 -4.08893138e-01 -7.75721431e-01 -2.93426543e-01
6.78668559e-01 -7.99689412e-01 -3.99135590e-01 8.90859067e-01
6.73722327e-01 1.32797480e-01 7.24914491e-01 -1.24825573e+00
-9.65746641e-01 7.71771967e-01 -5.79116642e-01 8.15221786e-01
3.43806148e-01 -3.40551808e-02 8.90530586e-01 -5.70048273e-01
6.01850390e-01 1.35416794e+00 7.14987516e-01 1.03910220e+00
-8.99855256e-01 -1.18582380e+00 -1.60360977e-01 2.21116304e-01
-1.12489986e+00 -2.66118348e-01 9.55962956e-01 -6.39784455e-01
1.10790384e+00 1.62177637e-01 6.49208248e-01 1.98940754e+00
-5.26681207e-02 7.06659377e-01 9.67174351e-01 -1.18650027e-01
-8.91956165e-02 4.20741975e-01 5.10799646e-01 6.30059302e-01
4.20630634e-01 -2.34585315e-01 -8.47360134e-01 -4.72266436e-01
1.65077895e-01 -9.35451537e-02 2.36054838e-01 2.49763995e-01
-5.83158880e-02 1.59807181e+00 3.31216127e-01 7.53073454e-01
9.22027975e-02 6.35885224e-02 9.45409238e-01 2.70319998e-01
1.10651553e+00 5.66046417e-01 1.11599974e-01 -7.20740139e-01
-9.19768155e-01 1.03497483e-01 7.35415637e-01 4.49995428e-01
1.81692660e-01 2.77034044e-01 1.16139702e-01 1.48613214e+00
1.06474515e-02 4.20538396e-01 9.20473099e-01 -4.72216785e-01
4.41659659e-01 5.72085977e-01 -4.44270283e-01 -1.37765217e+00
-6.60380542e-01 -2.16358393e-01 -4.76585656e-01 -2.18723983e-01
3.07194889e-01 -3.98422539e-01 -5.01197517e-01 1.60058773e+00
1.98191375e-01 2.00654522e-01 -3.87830228e-01 7.11410463e-01
1.01051962e+00 5.40574729e-01 5.70516586e-01 -2.59967595e-01
1.00588131e+00 -7.64925838e-01 -9.77433264e-01 -2.75152177e-01
1.12445676e+00 -6.12078071e-01 9.02322471e-01 5.56835711e-01
-9.70626831e-01 1.08228053e-03 -8.23113441e-01 -3.94603193e-01
-9.21006680e-01 -3.22095364e-01 8.20461690e-01 1.31103086e+00
-8.51594210e-01 3.16245049e-01 -5.26779175e-01 -6.10789776e-01
6.63945794e-01 2.53930748e-01 -4.82993156e-01 2.57962257e-01
-1.47358739e+00 1.19016349e+00 6.62119985e-02 -3.62147123e-01
-9.01485205e-01 -5.90504646e-01 -6.82055891e-01 -1.82527468e-01
9.66102704e-02 -9.21961814e-02 1.13481665e+00 -8.61945271e-01
-1.29642308e+00 1.03698659e+00 1.17989272e-01 -6.15896702e-01
2.60050356e-01 -3.32781672e-01 -5.37248850e-01 8.57630894e-02
1.29775908e-02 3.71499747e-01 8.66726577e-01 -9.97430742e-01
-5.26080906e-01 -7.35303223e-01 1.30689621e-01 2.78869318e-03
-1.45282149e+00 6.56245828e-01 2.36454919e-01 -6.00936353e-01
-6.88215554e-01 -8.43254387e-01 2.99212098e-01 -4.27909851e-01
-1.91562086e-01 -6.16490722e-01 1.20066226e+00 -9.47268188e-01
1.79998207e+00 -2.06403470e+00 6.27629086e-02 -1.22297667e-01
4.24061090e-01 7.07915843e-01 3.11569363e-01 5.58775663e-01
1.17604673e-01 5.74144721e-01 -1.76427573e-01 -8.04492772e-01
-3.37403491e-02 3.92333925e-01 -2.65300274e-01 6.44062281e-01
2.11102515e-02 5.62588453e-01 -9.81547415e-01 -5.89110374e-01
5.30520454e-02 4.79309380e-01 -7.18980730e-01 2.83096462e-01
3.57294679e-01 1.44844130e-01 -2.11919397e-01 4.71491098e-01
4.19281095e-01 3.66006494e-01 -8.27257112e-02 6.38943374e-01
-1.08694583e-01 5.46522558e-01 -3.31409216e-01 1.25299096e+00
-6.95169806e-01 1.04611170e+00 1.48034140e-01 -6.78075254e-01
9.50525999e-01 1.32652998e-01 4.73392636e-01 -6.41010463e-01
4.32907254e-01 1.68223217e-01 2.52906084e-01 -8.49137306e-01
6.88245952e-01 -3.88687670e-01 -3.72503370e-01 4.60116416e-01
3.24888825e-01 -7.33677624e-03 2.58803159e-01 4.08839136e-01
1.20768392e+00 -5.50035477e-01 2.29647070e-01 -2.07829736e-02
3.77110898e-01 -1.77197322e-01 1.76830143e-01 7.04444170e-01
-6.74923778e-01 4.18718070e-01 7.85059392e-01 -2.98287779e-01
-1.05079353e+00 -7.16202676e-01 -4.37442482e-01 1.66658819e+00
-4.34945285e-01 -6.95812762e-01 -9.68172014e-01 -9.34848428e-01
2.05873121e-02 9.10667956e-01 -9.25983608e-01 -4.67747539e-01
-5.10030806e-01 -9.57515895e-01 1.19241655e+00 3.89125228e-01
2.53333330e-01 -1.00088322e+00 -2.60996521e-01 -2.94646136e-02
-4.21358682e-02 -9.17768359e-01 -1.04230195e-01 -1.76990200e-02
-5.85699081e-01 -8.69135320e-01 -3.61430824e-01 -3.63825023e-01
1.54058680e-01 2.67535686e-01 5.95302105e-01 4.75884408e-01
-3.06151688e-01 7.29199126e-02 -6.93419278e-01 -4.96861368e-01
-6.15635693e-01 3.90700251e-01 3.44756186e-01 -1.03420183e-01
8.95949125e-01 -5.21734536e-01 -7.16441199e-02 -4.33330089e-02
-9.79232907e-01 -5.23521304e-01 1.57949124e-02 8.70769322e-01
-5.95447600e-01 -3.29140425e-01 6.86369419e-01 -1.21370780e+00
1.14515138e+00 -1.10351217e+00 1.47050813e-01 -4.57082570e-01
-1.86878040e-01 -6.83340490e-01 6.55466318e-01 -4.66274917e-01
-1.08030105e+00 -5.94316483e-01 -5.93134344e-01 -3.50897223e-01
-5.35966158e-01 2.75503427e-01 4.18195456e-01 9.90507230e-02
8.60970259e-01 -5.62634990e-02 -9.45678260e-03 -5.97847164e-01
1.48030119e-02 1.25655258e+00 -6.56179115e-02 -1.87890530e-01
6.23238802e-01 4.12800759e-01 -4.11286592e-01 -1.42158711e+00
-1.00110173e+00 -8.65211785e-01 -7.93641090e-01 -3.16425264e-01
1.02257550e+00 -5.67866504e-01 -6.01946592e-01 5.40732205e-01
-1.11489880e+00 -1.52916864e-01 2.93366253e-01 2.20079169e-01
-2.36462861e-01 4.59553421e-01 -9.94227886e-01 -1.19405508e+00
-3.35357398e-01 -7.90446341e-01 9.10251856e-01 -1.46697491e-01
-6.22698367e-01 -1.12812579e+00 4.51883852e-01 8.57127249e-01
5.19984245e-01 1.43386811e-01 6.30807936e-01 -1.43237329e+00
5.88706851e-01 -3.89117777e-01 -1.15888258e-02 5.09431481e-01
1.03851080e-01 3.02139223e-01 -1.19761419e+00 -2.08681405e-01
8.54834765e-02 -8.51507246e-01 8.06319177e-01 -5.70128821e-02
1.12554550e+00 -3.62247378e-01 -3.27541769e-01 2.03869477e-01
8.12493503e-01 2.68366992e-01 3.85909677e-01 3.77474874e-01
7.78251469e-01 7.75728166e-01 2.70842969e-01 7.19726384e-01
5.27065359e-02 6.79017961e-01 2.96278447e-01 3.12652647e-01
-2.22104087e-01 -2.95691669e-01 5.70020020e-01 9.58018005e-01
1.26561195e-01 -2.43044227e-01 -1.13325799e+00 5.80617964e-01
-1.48739469e+00 -1.29544055e+00 -3.89996856e-01 1.66996598e+00
6.77003801e-01 6.14664219e-02 6.18937790e-01 6.04368448e-02
6.66606963e-01 3.42344522e-01 -1.98750973e-01 -1.09488869e+00
4.57207225e-02 -1.61780138e-02 5.30384369e-02 6.27535224e-01
-1.22730362e+00 1.13577580e+00 6.73720551e+00 1.03597176e+00
-1.06180048e+00 6.74813867e-01 5.88452339e-01 -5.93008637e-01
8.73821154e-02 -6.16943359e-01 -1.00917864e+00 9.26928997e-01
1.48584557e+00 -1.53392795e-02 6.46311864e-02 1.04023921e+00
1.33684218e-01 9.98016372e-02 -6.21475935e-01 7.17540205e-01
7.74196386e-01 -8.42238367e-01 -1.26390047e-02 2.26798192e-01
4.87413079e-01 4.34449362e-03 4.50627148e-01 6.42467022e-01
4.36281681e-01 -1.05565834e+00 4.00718391e-01 1.30508780e-01
3.42203438e-01 -9.98911321e-01 9.43960965e-01 5.61614692e-01
-1.18072309e-01 -5.86731017e-01 -5.72579324e-01 -3.86060059e-01
1.34668434e-02 1.29184574e-01 -1.07715154e+00 -1.36351109e-01
9.62387681e-01 7.67639458e-01 -7.33566821e-01 6.67311490e-01
-1.63462341e-01 1.18836689e+00 -3.37292731e-01 -5.93300998e-01
5.94425917e-01 1.34892799e-02 5.84789395e-01 1.57575846e+00
2.11675428e-02 -7.06889480e-03 -1.62826583e-01 4.24937606e-01
-1.86092511e-01 5.05502105e-01 -1.21650994e+00 -2.26763174e-01
3.51787657e-01 1.26630223e+00 -2.24968612e-01 -1.62850723e-01
-5.75014532e-01 7.01088488e-01 1.11263502e+00 -3.21308851e-01
-1.02831018e+00 -2.70266496e-02 8.99035037e-01 3.19780260e-01
-2.91194320e-01 -3.45682740e-01 -3.80205631e-01 -9.16078508e-01
-5.84186971e-01 -7.49493599e-01 8.94706130e-01 -5.01397371e-01
-1.82452309e+00 5.38949192e-01 1.61956280e-01 -6.91250026e-01
-3.25644761e-01 -6.62274241e-01 -4.98271585e-01 3.39170843e-01
-9.84453917e-01 -1.20081115e+00 -5.57668600e-03 4.43660587e-01
8.15158367e-01 -4.60522503e-01 6.64061189e-01 4.60705370e-01
-1.13309717e+00 8.26258957e-01 1.66586682e-01 3.74257654e-01
7.73944139e-01 -1.12295866e+00 -1.27877846e-01 5.05848289e-01
-3.25933448e-03 7.27402866e-01 7.01188028e-01 -7.71047056e-01
-6.95901215e-01 -8.02900374e-01 7.88903177e-01 -1.04646277e+00
1.38806331e+00 -8.38966906e-01 -1.11845016e+00 7.49419153e-01
3.80349129e-01 -5.12446404e-01 1.36754453e+00 4.32021886e-01
-5.68209410e-01 1.93691149e-01 -1.25387132e+00 5.35991073e-01
1.12985182e+00 -5.06103456e-01 -8.11997354e-01 6.48011446e-01
5.95690668e-01 -1.23292439e-01 -8.46254051e-01 -1.78930789e-01
2.94733077e-01 -1.01951575e+00 6.89004719e-01 -1.18189275e+00
8.36978674e-01 7.37447441e-01 9.43322927e-02 -1.28897583e+00
-4.09819096e-01 -2.65503138e-01 -1.41593710e-01 1.59719372e+00
3.23652655e-01 -4.77468729e-01 6.79978311e-01 5.74322701e-01
-1.67053625e-01 -4.72674370e-01 -1.36956966e+00 -6.50702238e-01
7.37841725e-01 -4.93721157e-01 1.40450895e-01 1.45443869e+00
5.55093765e-01 5.03597617e-01 -8.26666772e-01 -2.16631144e-01
2.19523385e-01 -4.49050903e-01 8.54696274e-01 -1.17595994e+00
2.49407068e-01 -7.50156164e-01 -5.25188386e-01 -2.77658015e-01
7.68888533e-01 -8.56561363e-01 -4.40922052e-01 -1.04189825e+00
4.60960984e-01 -1.47985205e-01 9.89282727e-02 4.91864115e-01
-2.00036138e-01 6.56669080e-01 1.91527277e-01 9.94653255e-02
-5.35117269e-01 6.24610782e-01 7.00585604e-01 -1.92170218e-01
-6.54097414e-03 -3.24573606e-01 -6.11471295e-01 8.65876079e-01
1.04392064e+00 -6.67214513e-01 -2.74309695e-01 -3.27288002e-01
-5.31923212e-02 -4.50856328e-01 -3.60436216e-02 -9.26093340e-01
1.00477025e-01 -6.64155781e-02 2.33561248e-01 -1.54163361e-01
8.73088181e-01 -6.44497395e-01 -5.41550100e-01 3.43117565e-01
-5.87231636e-01 2.64466628e-02 3.26128632e-01 4.66174513e-01
-1.22232847e-01 -7.94110894e-01 7.46888340e-01 -1.03776585e-02
-4.33316648e-01 2.99025536e-01 -9.35862422e-01 1.17533058e-01
1.13386083e+00 1.57320667e-02 -7.01743484e-01 -6.59437656e-01
-6.66518867e-01 -1.30712360e-01 1.62318617e-01 8.84158909e-01
4.80212808e-01 -9.71412838e-01 -7.30049670e-01 -1.48733243e-01
2.38840077e-02 -8.00706565e-01 4.81018275e-01 7.42854834e-01
-3.56715560e-01 5.33469796e-01 -3.20521712e-01 -2.46813610e-01
-1.58968079e+00 8.14906895e-01 1.99018508e-01 -2.13203043e-01
-2.07099289e-01 8.40043485e-01 -2.03160822e-01 -4.51197296e-01
4.10616249e-02 5.16498148e-01 -7.25862920e-01 5.35147727e-01
7.76288271e-01 6.00034893e-01 -8.94269422e-02 -1.29610205e+00
-1.73747256e-01 -4.58208583e-02 -4.16198492e-01 9.81646255e-02
1.41515255e+00 1.00220442e-02 -3.17937255e-01 6.18986905e-01
1.48048019e+00 2.56591231e-01 -3.87610793e-01 1.81327164e-01
1.30504429e-01 -7.93480575e-01 9.31275636e-02 -6.30735636e-01
-9.15365219e-01 1.28276193e+00 2.80971855e-01 8.12039375e-01
5.63389242e-01 -2.17858795e-03 9.23001289e-01 2.26995185e-01
1.42481536e-01 -1.27424932e+00 5.82605600e-01 7.28561997e-01
9.50806499e-01 -1.22186804e+00 -8.30928832e-02 -3.32546145e-01
-7.52686858e-01 9.44216311e-01 1.30035520e+00 -9.12261307e-02
6.28400087e-01 2.55076647e-01 1.11573920e-01 -4.46808368e-01
-7.29810894e-01 -6.00906238e-02 -1.74265932e-02 7.67313838e-01
6.98305488e-01 -3.07912733e-02 -7.26588964e-01 6.99716926e-01
-6.14109635e-01 -5.70552707e-01 7.79206514e-01 6.68779016e-01
-5.57828963e-01 -8.97791147e-01 -3.07685405e-01 5.85609138e-01
-9.48304832e-01 -1.08741187e-02 -1.30047834e+00 9.90923464e-01
6.26490265e-02 1.22423983e+00 2.39423275e-01 -5.78933418e-01
1.12752631e-01 4.04421985e-01 -6.34620637e-02 -8.16822886e-01
-1.06497598e+00 -3.71285796e-01 6.38967156e-01 -2.25399062e-01
-1.47969410e-01 -6.85222089e-01 -8.99028242e-01 -8.03267300e-01
-3.63025874e-01 1.09804712e-01 6.91891909e-01 8.99576843e-01
1.57868087e-01 8.01780373e-02 4.73259687e-01 -4.63173628e-01
-3.79140556e-01 -1.52988100e+00 -6.16222441e-01 8.19007754e-01
1.95318416e-01 -9.74312425e-01 -5.76577842e-01 -5.04882991e-01] | [8.788386344909668, 10.545029640197754] |
a584c4d9-2f3a-48a5-b168-fefd05eeae38 | distributed-stochastic-optimization-under-a | 2301.12677 | null | https://arxiv.org/abs/2301.12677v2 | https://arxiv.org/pdf/2301.12677v2.pdf | Distributed Stochastic Optimization under a General Variance Condition | Distributed stochastic optimization has drawn great attention recently due to its effectiveness in solving large-scale machine learning problems. Though numerous algorithms have been proposed and successfully applied to general practical problems, their theoretical guarantees mainly rely on certain boundedness conditions on the stochastic gradients, varying from uniform boundedness to the relaxed growth condition. In addition, how to characterize the data heterogeneity among the agents and its impacts on the algorithmic performance remains challenging. In light of such motivations, we revisit the classical Federated Averaging (FedAvg) algorithm for solving the distributed stochastic optimization problem and establish the convergence results under only a mild variance condition on the stochastic gradients for smooth nonconvex objective functions. Almost sure convergence to a stationary point is also established under the condition. Moreover, we discuss a more informative measurement for data heterogeneity as well as its implications. | ['Shi Pu', 'Xiao Li', 'Kun Huang'] | 2023-01-30 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [-1.22772448e-01 -1.89093798e-01 -1.68051153e-01 -2.44145960e-01
-1.02228785e+00 -4.41222221e-01 1.36553869e-01 3.04508865e-01
-5.53671002e-01 1.01872873e+00 8.27537850e-02 8.19570869e-02
-5.88095486e-01 -4.07338887e-01 -7.22487748e-01 -1.28615165e+00
-1.85418099e-01 3.07649285e-01 -1.25392184e-01 7.88629800e-02
1.37631238e-01 2.78758049e-01 -1.18358469e+00 -5.17193019e-01
1.32078862e+00 1.08530998e+00 3.50131631e-01 4.61857140e-01
5.78482822e-03 6.19661391e-01 -6.27781808e-01 -2.65453845e-01
4.18702662e-01 -4.19086993e-01 -2.88789958e-01 2.99826294e-01
1.04341529e-01 -1.89200729e-01 8.57716203e-02 1.55456948e+00
7.85151064e-01 2.99883723e-01 5.57492554e-01 -1.37079155e+00
-3.11307847e-01 7.40921021e-01 -7.94788241e-01 1.73331469e-01
-2.49084756e-02 -6.39413744e-02 1.19938636e+00 -8.28381181e-01
4.12625849e-01 8.37788880e-01 6.53754771e-01 3.14602464e-01
-1.02815580e+00 -3.01290691e-01 4.43406999e-01 2.56366074e-01
-1.46114409e+00 -2.85030842e-01 6.74919903e-01 -4.18236017e-01
3.05374563e-01 3.21959585e-01 2.83618480e-01 6.98032796e-01
6.62066713e-02 9.71714258e-01 9.51136351e-01 -3.35770160e-01
6.63210750e-01 1.73963726e-01 2.16659874e-01 6.10250294e-01
6.81148946e-01 -3.88429314e-01 -2.94505894e-01 -3.49544138e-01
4.48166370e-01 -1.20495051e-01 -5.44010162e-01 -4.87467170e-01
-9.91830885e-01 1.07906663e+00 2.89007604e-01 2.74136961e-01
-8.89666498e-01 1.96725726e-02 3.29209596e-01 2.53627717e-01
7.62648582e-01 1.88246459e-01 -4.44783241e-01 -1.12138875e-01
-6.38097525e-01 4.78923500e-01 9.03876364e-01 8.66873980e-01
6.87977731e-01 2.22455084e-01 -2.69371718e-01 7.47944057e-01
6.56540617e-02 6.93911254e-01 2.02118620e-01 -1.04202497e+00
7.85956383e-01 3.28650028e-01 4.35041517e-01 -1.21618259e+00
-3.27682167e-01 -8.23246598e-01 -9.45368290e-01 -1.38345271e-01
7.74036825e-01 -7.45342553e-01 -9.53020751e-02 1.87446737e+00
5.71633518e-01 1.33934155e-01 -1.84881419e-01 1.26325333e+00
5.67640252e-02 5.29901981e-01 -2.02815294e-01 -6.67174578e-01
7.11194396e-01 -7.98268378e-01 -8.64618540e-01 1.53659359e-02
7.38819480e-01 -3.21727723e-01 1.05722451e+00 1.28219232e-01
-1.04833126e+00 2.44128138e-01 -7.01756239e-01 3.39709610e-01
6.57056421e-02 5.80045134e-02 5.39182186e-01 8.41298223e-01
-8.98720384e-01 3.48750949e-01 -8.51294935e-01 -2.68001199e-01
4.09289658e-01 3.70660722e-01 1.70805112e-01 2.35987944e-03
-7.98969269e-01 4.32317734e-01 -5.36173396e-02 4.67820376e-01
-8.37954044e-01 -8.39814186e-01 -5.23722410e-01 8.46626312e-02
6.84005678e-01 -8.54694605e-01 1.10502398e+00 -9.30937588e-01
-1.66673255e+00 4.81877834e-01 -1.92025259e-01 -5.24344504e-01
1.07713318e+00 -1.91836312e-01 1.41748071e-01 2.38987766e-02
7.61775821e-02 -4.67300177e-01 5.92634022e-01 -1.06201267e+00
-7.02102482e-01 -6.81890190e-01 3.13970894e-01 4.36937630e-01
-7.56426394e-01 1.39411956e-01 -1.62482336e-01 -6.96227372e-01
-3.38404000e-01 -9.00368273e-01 -5.69822907e-01 -1.91152632e-01
-3.89103323e-01 -1.68820530e-01 4.69251961e-01 -3.85704279e-01
1.27716196e+00 -1.97354150e+00 4.39137787e-01 1.86287731e-01
2.83978730e-02 -7.52792284e-02 -3.58161889e-02 4.26640004e-01
4.46991235e-01 -6.65130913e-02 -3.51878822e-01 -4.96862978e-01
2.70411879e-01 -8.85432065e-02 -2.62326032e-01 1.07660580e+00
-2.96011209e-01 5.00410676e-01 -8.93774211e-01 -1.36406869e-01
-3.99699202e-03 2.27895707e-01 -7.21822739e-01 3.96037735e-02
-1.68231010e-01 5.32379091e-01 -1.04146743e+00 3.89206260e-01
7.51082659e-01 -4.24960554e-01 1.01385221e-01 7.94828385e-02
-1.69786856e-01 -3.36198807e-01 -1.31376195e+00 1.34813845e+00
-6.41590118e-01 3.01520377e-01 9.96924937e-01 -1.45527959e+00
4.79759634e-01 1.51361063e-01 8.35469842e-01 -1.53485402e-01
2.99865127e-01 4.27379817e-01 -1.69087127e-01 -4.27802801e-01
4.57889020e-01 -3.54987085e-01 1.11869508e-02 5.68204939e-01
-4.88557070e-01 1.29196957e-01 2.25311831e-01 1.43483207e-01
8.10868859e-01 -4.26446438e-01 1.66398734e-01 -7.06367731e-01
5.90642869e-01 -8.03488269e-02 6.77898824e-01 9.77246463e-01
-3.25680554e-01 4.02474850e-01 4.27783668e-01 -1.92148406e-02
-7.41875231e-01 -5.96039653e-01 2.29495578e-02 1.25434077e+00
3.78876865e-01 -9.61816162e-02 -9.03531730e-01 -4.93059337e-01
2.47881502e-01 6.02488875e-01 -6.92728281e-01 7.54022673e-02
-3.44145238e-01 -1.34886801e+00 -1.38768442e-02 8.94446298e-02
5.13351262e-01 -4.61233228e-01 -4.36454177e-01 1.39302164e-01
-5.85603081e-02 -1.06403363e+00 -7.03247547e-01 -2.11288154e-01
-6.94286644e-01 -9.41500783e-01 -1.14551854e+00 -3.47564310e-01
7.22451508e-01 5.35736799e-01 7.31501818e-01 -1.95983633e-01
2.89182141e-02 6.48184776e-01 -1.60776168e-01 -4.99284029e-01
-3.05160806e-02 2.38569126e-01 3.11892778e-01 5.82895339e-01
-9.84000489e-02 -2.71885723e-01 -6.28060699e-01 4.93103534e-01
-8.86246204e-01 -2.71431923e-01 3.19781691e-01 5.92153728e-01
4.98003423e-01 3.22112471e-01 8.27485085e-01 -7.47336745e-01
9.66040969e-01 -7.16250718e-01 -7.80506134e-01 3.36835653e-01
-3.87558490e-01 1.21854723e-03 9.24418509e-01 -3.67178023e-01
-1.13814485e+00 -2.32701614e-01 2.93777823e-01 -5.21400869e-01
3.87770444e-01 8.09699595e-01 -2.72631198e-01 2.48735435e-02
3.32830638e-01 1.96186364e-01 1.94308102e-01 -3.65076482e-01
2.74570197e-01 4.96408254e-01 3.39344814e-02 -8.92648041e-01
7.70682096e-01 8.09261918e-01 3.37094292e-02 -1.04391086e+00
-1.01036441e+00 -2.94984639e-01 -1.34099469e-01 -2.19575822e-01
4.79529172e-01 -7.83307135e-01 -8.48664343e-01 5.06993771e-01
-8.38364124e-01 -4.60961163e-01 -2.89768308e-01 6.29028141e-01
-7.04790711e-01 3.22169602e-01 -5.71362019e-01 -1.26814568e+00
-1.80403575e-01 -1.13566494e+00 5.46428800e-01 2.47047842e-01
3.28793645e-01 -1.42615962e+00 5.21537885e-02 1.85447901e-01
7.90088058e-01 2.13336781e-01 4.94242996e-01 -5.59448659e-01
-5.32442033e-01 -2.31248572e-01 -2.37045661e-01 1.59870282e-01
1.38900876e-01 -1.70081869e-01 -4.34340239e-01 -5.18840909e-01
3.58287305e-01 -5.12233451e-02 7.45563328e-01 9.17346179e-01
1.06373048e+00 -6.13311946e-01 -2.42384762e-01 6.11322880e-01
1.55247641e+00 -1.83601514e-01 -1.84109762e-01 4.14024115e-01
5.13236761e-01 6.62079692e-01 5.54930508e-01 1.15300310e+00
2.75130063e-01 5.68183541e-01 4.95330036e-01 2.23720387e-01
4.25174236e-01 1.47496074e-01 4.21650201e-01 8.84226084e-01
-1.09712198e-01 -2.59901613e-01 -5.92158258e-01 4.83710140e-01
-2.10581636e+00 -9.05543447e-01 -1.86250135e-01 2.55077291e+00
6.24653995e-01 -4.42644417e-01 3.52389902e-01 -1.65238336e-01
1.19635892e+00 3.04467767e-01 -9.11379993e-01 -1.23423308e-01
-5.72417259e-01 -4.88177866e-01 8.54886115e-01 5.91066003e-01
-8.45295310e-01 5.09263277e-01 6.23899746e+00 1.07568550e+00
-1.10012758e+00 3.13967496e-01 8.06020856e-01 -3.90181005e-01
-3.88697594e-01 -3.57898504e-01 -6.50153399e-01 6.53139055e-01
5.66948414e-01 -7.43893623e-01 6.28984749e-01 8.49100828e-01
8.45623136e-01 -1.34897321e-01 -6.55045867e-01 1.05083013e+00
-7.63949677e-02 -1.01783037e+00 -2.68940449e-01 3.28150094e-01
1.26726592e+00 1.60716772e-01 3.35294276e-01 -4.01414111e-02
5.09399176e-01 -6.23809457e-01 6.77876949e-01 3.07395935e-01
4.26284015e-01 -9.20706809e-01 7.23260880e-01 4.92656827e-01
-9.75218058e-01 -4.71262485e-01 -4.97196823e-01 -2.87484467e-01
3.65495652e-01 1.12331057e+00 -1.55307814e-01 7.36459136e-01
3.14126134e-01 5.91782272e-01 -1.19863905e-01 1.20017791e+00
2.24142689e-02 6.11786246e-01 -6.52848125e-01 -3.54288906e-01
3.05295467e-01 -6.75634563e-01 8.27397227e-01 8.53052676e-01
4.95761842e-01 -2.04751939e-01 2.45805457e-01 5.94081461e-01
-1.05548948e-01 6.11322045e-01 -2.20777526e-01 -3.51844057e-02
3.96995366e-01 1.09028649e+00 -5.41865408e-01 -1.38095528e-01
-6.37992740e-01 7.80300617e-01 4.58970457e-01 7.95481265e-01
-9.14648831e-01 -2.11628392e-01 8.29801977e-01 -3.15357819e-02
2.88593799e-01 -2.70186514e-01 -4.84686553e-01 -1.38952863e+00
5.01294196e-01 -5.55873275e-01 5.05535126e-01 9.03026834e-02
-1.45987332e+00 2.72864282e-01 -2.08705053e-01 -9.65823114e-01
-2.35365406e-02 -2.20115453e-01 -6.93815470e-01 6.88805163e-01
-1.44543624e+00 -6.77160025e-01 -1.02354996e-01 7.04805613e-01
3.96556050e-01 -2.38339230e-01 3.00254822e-01 4.69808817e-01
-1.23673010e+00 4.87310648e-01 9.48392928e-01 -2.83762068e-01
4.19229358e-01 -1.16730475e+00 -5.47790885e-01 8.65057766e-01
-2.53628820e-01 5.08691013e-01 1.11935592e+00 -3.21300298e-01
-1.83704293e+00 -1.10754859e+00 5.11764228e-01 3.18378098e-02
1.05887985e+00 -1.15471393e-01 -6.14479601e-01 3.78970921e-01
5.83587289e-02 2.18091354e-01 4.18359190e-01 1.28617018e-01
1.38957351e-01 -3.01224679e-01 -1.13249564e+00 5.05197763e-01
8.17826688e-01 -1.32223487e-01 2.34941557e-01 6.15229845e-01
2.70777255e-01 -1.88330278e-01 -6.55204892e-01 1.55186638e-01
3.21105212e-01 -8.31573188e-01 6.08917952e-01 -6.76265538e-01
2.22577557e-01 -8.73752683e-02 -4.08500314e-01 -1.39864004e+00
-3.07164975e-02 -1.08668423e+00 -7.33571798e-02 1.14412320e+00
3.29439253e-01 -1.03596270e+00 9.34317827e-01 8.58652353e-01
2.18460113e-01 -9.42080200e-01 -1.17181027e+00 -1.09463227e+00
2.63636976e-01 -3.04675013e-01 3.06461692e-01 9.11201239e-01
-2.23426502e-02 -6.54065982e-03 -5.70350587e-01 3.45559418e-01
8.98501575e-01 2.62257576e-01 8.75912309e-01 -8.43979001e-01
-3.84272218e-01 -6.99441969e-01 -5.02317250e-02 -1.30189323e+00
3.49272072e-01 -7.54096746e-01 -2.72733159e-02 -1.31838512e+00
4.09054071e-01 -5.56862116e-01 -4.17620212e-01 -3.87674063e-01
-4.70530003e-01 1.68550760e-02 3.02744031e-01 3.56417120e-01
-9.87250566e-01 9.74431157e-01 1.05287421e+00 4.15346847e-04
-2.31248930e-01 5.12755692e-01 -6.61612391e-01 5.90124607e-01
8.42369795e-01 -2.92950571e-01 -4.58468437e-01 -6.37718618e-01
3.75599056e-01 1.54615641e-01 -5.54907843e-02 -6.33016109e-01
3.18702728e-01 -4.35943604e-01 -5.15674531e-01 -1.69736847e-01
7.75251538e-02 -8.17521870e-01 -3.84462588e-02 3.79623681e-01
-4.60028023e-01 -3.55821736e-02 -3.11505228e-01 1.00572264e+00
-1.96180582e-01 -4.27382350e-01 8.02638412e-01 1.21756732e-01
-7.16172606e-02 6.10084236e-01 -3.56372863e-01 3.70957702e-01
1.31681514e+00 1.09206602e-01 1.34779047e-02 -7.88336575e-01
-5.73162556e-01 4.89965916e-01 4.22799975e-01 -8.62413347e-02
2.02495024e-01 -1.15805030e+00 -1.01323044e+00 -1.88433468e-01
-3.21597278e-01 -1.48740992e-01 4.09082204e-01 1.40689623e+00
-2.29446083e-01 3.75997186e-01 5.12763858e-01 -4.44350362e-01
-9.18912947e-01 6.54198825e-01 3.75813097e-01 -2.41694868e-01
-3.15213710e-01 7.36200571e-01 3.55232567e-01 -1.00056969e-01
3.40392232e-01 -1.35584280e-01 1.54724464e-01 1.53492823e-01
4.56667870e-01 7.94817746e-01 -5.04116667e-03 -4.51397419e-01
-3.57628435e-01 4.39794451e-01 1.96627408e-01 -1.11211248e-01
1.39172792e+00 -6.77633047e-01 -6.86103180e-02 5.12315869e-01
1.35559189e+00 3.40514779e-01 -1.44637108e+00 -4.88393933e-01
-1.12340022e-02 -4.33478713e-01 1.20350402e-02 -2.15092376e-01
-1.48404241e+00 5.10622799e-01 1.74918398e-01 5.34121037e-01
9.07928646e-01 -1.11277007e-01 5.96654832e-01 2.73992211e-01
7.08712697e-01 -1.22869551e+00 -2.34804586e-01 4.42533910e-01
7.69151747e-01 -1.20235825e+00 1.05694219e-01 -3.93487275e-01
-5.62618315e-01 8.59432161e-01 4.77789491e-02 -3.55246216e-01
7.55950868e-01 1.71819821e-01 -9.96329114e-02 1.02574319e-01
-5.78357279e-01 -1.23706497e-01 -2.05113471e-01 2.89358616e-01
2.60390580e-01 1.00120775e-01 -8.34782660e-01 5.38617969e-01
6.97628036e-02 -1.27038643e-01 4.76566017e-01 5.29162109e-01
-3.39463890e-01 -6.93850875e-01 -4.24987555e-01 3.73866022e-01
-8.70823622e-01 2.46541366e-01 -8.28871951e-02 4.80188042e-01
-5.02836764e-01 1.03791416e+00 -1.41573802e-01 2.11469620e-01
4.35389169e-02 -4.51686472e-01 3.15105975e-01 -3.55140537e-01
-4.17482078e-01 1.22226186e-01 -1.27834111e-01 -3.86574745e-01
-6.45463347e-01 -9.43126142e-01 -8.70470107e-01 -4.23533678e-01
-6.16104841e-01 5.72913110e-01 7.08471119e-01 9.49131191e-01
4.31361854e-01 2.36662313e-01 9.61176813e-01 -8.03435624e-01
-1.17812800e+00 -8.43899608e-01 -9.69086647e-01 2.92081326e-01
2.86005229e-01 -5.41587591e-01 -8.73204172e-01 -3.98895741e-01] | [6.3261399269104, 4.817956447601318] |
e4ecc74b-3ab0-4b56-a565-1f343a07024a | composer-creative-and-controllable-image | 2302.09778 | null | https://arxiv.org/abs/2302.09778v2 | https://arxiv.org/pdf/2302.09778v2.pdf | Composer: Creative and Controllable Image Synthesis with Composable Conditions | Recent large-scale generative models learned on big data are capable of synthesizing incredible images yet suffer from limited controllability. This work offers a new generation paradigm that allows flexible control of the output image, such as spatial layout and palette, while maintaining the synthesis quality and model creativity. With compositionality as the core idea, we first decompose an image into representative factors, and then train a diffusion model with all these factors as the conditions to recompose the input. At the inference stage, the rich intermediate representations work as composable elements, leading to a huge design space (i.e., exponentially proportional to the number of decomposed factors) for customizable content creation. It is noteworthy that our approach, which we call Composer, supports various levels of conditions, such as text description as the global information, depth map and sketch as the local guidance, color histogram for low-level details, etc. Besides improving controllability, we confirm that Composer serves as a general framework and facilitates a wide range of classical generative tasks without retraining. Code and models will be made available. | ['Jingren Zhou', 'Deli Zhao', 'Yujun Shen', 'Yu Liu', 'Di Chen', 'Lianghua Huang'] | 2023-02-20 | null | null | null | null | ['pose-transfer', 'virtual-try-on'] | ['computer-vision', 'computer-vision'] | [ 8.00305977e-02 6.96306750e-02 -8.82931948e-02 -3.16723399e-02
-2.12677374e-01 -7.95702755e-01 7.98913777e-01 -3.04557323e-01
1.24836817e-01 4.54358846e-01 1.69355765e-01 2.29076780e-02
-9.68310013e-02 -1.23344707e+00 -8.36531997e-01 -8.36331189e-01
3.28464061e-01 3.55471522e-01 2.23549291e-01 -4.36516762e-01
1.92212164e-01 6.43522501e-01 -1.72702003e+00 3.84818316e-01
9.90639448e-01 1.04005361e+00 4.62872416e-01 5.04132032e-01
-2.63794184e-01 3.82623404e-01 -6.28890395e-01 -5.58848560e-01
2.56565303e-01 -5.64886272e-01 -1.70417026e-01 4.14806753e-01
4.84206885e-01 -3.18601817e-01 -3.03663164e-01 8.19597602e-01
4.55296040e-01 2.41417438e-02 6.16696715e-01 -1.19410372e+00
-1.37490427e+00 6.76451564e-01 -1.96648613e-01 -4.36668426e-01
1.82216421e-01 5.45367181e-01 9.58275437e-01 -8.54448617e-01
8.64113927e-01 1.20549214e+00 1.16253346e-01 6.53270125e-01
-1.37261772e+00 -6.10232472e-01 4.11903232e-01 8.57463777e-02
-1.28578866e+00 -4.34697568e-01 9.03788745e-01 -3.10433686e-01
4.94791001e-01 5.49095929e-01 1.11037278e+00 1.37842619e+00
1.41884282e-01 7.30218709e-01 9.34653759e-01 -1.83009923e-01
4.35673147e-01 4.71440405e-02 -5.83541572e-01 7.21050382e-01
1.07851833e-01 -2.74611730e-02 -4.21080440e-01 1.53722689e-01
1.35449302e+00 3.60091925e-01 -3.19440454e-01 -4.86220956e-01
-1.44930208e+00 6.52960002e-01 5.90540409e-01 2.27303371e-01
-3.14471602e-01 2.39612550e-01 -4.70804386e-02 3.53195041e-01
7.41490945e-02 7.63059378e-01 -1.23393767e-01 1.07207619e-01
-8.54010761e-01 3.14251602e-01 6.01160705e-01 1.32792950e+00
8.87915194e-01 2.50058800e-01 -4.69152361e-01 7.68398404e-01
-3.24544013e-02 6.36345685e-01 4.55314934e-01 -9.63343322e-01
3.39826018e-01 8.08001280e-01 7.69754201e-02 -1.21008742e+00
-4.04680967e-02 -6.22193813e-01 -1.24712336e+00 3.48444790e-01
1.48059145e-01 8.93536806e-02 -1.04613078e+00 1.85174286e+00
2.25087970e-01 -3.04424316e-01 -3.72208118e-01 8.88700902e-01
5.74863970e-01 7.85431743e-01 -2.25484058e-01 -1.20261922e-01
1.36073101e+00 -1.11980176e+00 -7.93138802e-01 -1.79721594e-01
-9.33594406e-02 -6.97400689e-01 1.45282316e+00 5.54483294e-01
-1.35565186e+00 -6.53588057e-01 -1.03449559e+00 -1.97592825e-01
-3.68804038e-01 2.31215864e-01 9.45010066e-01 3.36980790e-01
-1.20135522e+00 5.99701405e-01 -5.69264233e-01 -5.01500554e-02
3.06605220e-01 8.28230828e-02 -2.56685108e-01 -1.01142399e-01
-9.19666350e-01 4.75186974e-01 1.64478794e-01 1.93412781e-01
-1.00562811e+00 -6.31546438e-01 -5.84002614e-01 2.61509567e-01
4.55276757e-01 -1.06134474e+00 7.15693116e-01 -9.11973059e-01
-1.85448241e+00 5.91205359e-01 2.39411116e-01 2.43226260e-01
7.57873058e-01 3.80176604e-02 -3.55930507e-01 3.18206921e-02
-1.86879545e-01 7.83135056e-01 1.36340940e+00 -1.46281362e+00
-4.00241107e-01 -2.40091681e-01 3.47796857e-01 2.53725760e-02
-5.81637800e-01 -4.41583842e-01 -6.85413003e-01 -9.32831407e-01
1.11467242e-01 -8.67472887e-01 -3.36620241e-01 3.41625035e-01
-4.16845113e-01 5.41136786e-02 6.16981149e-01 -3.06365371e-01
1.54931617e+00 -2.20991468e+00 7.44174659e-01 2.93951571e-01
4.73222375e-01 -6.76200613e-02 -3.39680701e-01 5.96435726e-01
4.71605174e-02 3.67524534e-01 -1.18037745e-01 -1.82752863e-01
2.12191939e-01 9.63635221e-02 -3.50077391e-01 -6.47323532e-03
2.99889028e-01 1.26395404e+00 -7.06350148e-01 -2.55651414e-01
2.53931642e-01 5.51382124e-01 -6.17900491e-01 2.80441433e-01
-5.73336363e-01 4.16324317e-01 -6.39997244e-01 4.27946359e-01
6.66859269e-01 -4.69660789e-01 9.00070928e-03 -4.01820987e-01
-2.38145605e-01 -1.96176112e-01 -1.36613309e+00 1.99758649e+00
-6.89796984e-01 2.77115107e-01 2.01063618e-01 -3.79355192e-01
1.25591409e+00 3.39658149e-02 6.68525919e-02 -6.00627720e-01
-2.29252130e-02 1.20292924e-01 -1.37003452e-01 -4.13848072e-01
5.04355669e-01 1.08138099e-02 -1.97576061e-01 5.34688711e-01
-1.51983932e-01 -4.69650120e-01 2.49624997e-01 1.84457898e-01
8.58818352e-01 2.65556067e-01 3.53228711e-02 -1.57275632e-01
2.25223899e-01 -3.11436057e-01 2.76368111e-01 4.46308345e-01
4.57059145e-01 7.25868404e-01 6.32728577e-01 -5.87571144e-01
-1.41932309e+00 -1.05544245e+00 1.01864494e-01 9.81391490e-01
1.46808431e-01 -6.88038468e-01 -8.25322032e-01 -7.44121373e-02
-1.56767458e-01 5.15562534e-01 -7.67941713e-01 -1.83120698e-01
-6.43126070e-01 -4.47005242e-01 8.24262202e-02 4.77674335e-01
4.09604877e-01 -1.18217623e+00 -5.20933151e-01 7.78161222e-03
1.02134809e-01 -6.41370356e-01 -6.58251047e-01 4.83937711e-02
-8.16321969e-01 -5.88714898e-01 -8.80229652e-01 -8.65817964e-01
9.15047526e-01 1.00786887e-01 1.14629757e+00 3.64844382e-01
-4.28465813e-01 -9.63972136e-02 -2.00168714e-01 2.55951315e-01
-4.46491838e-01 2.19432730e-02 -3.90185326e-01 1.20765917e-01
-5.14354467e-01 -1.12598038e+00 -1.03970540e+00 3.64311904e-01
-1.23292327e+00 7.23373711e-01 7.58905351e-01 8.04098785e-01
7.15328991e-01 -4.77516977e-03 3.80475700e-01 -8.06044281e-01
8.75080168e-01 -2.05969319e-01 -5.16204536e-01 3.84868175e-01
-5.68666160e-01 2.87197202e-01 1.13232028e+00 -7.26172209e-01
-1.00162196e+00 -1.25840008e-01 2.42879800e-02 -4.14373308e-01
-6.35324866e-02 9.14383978e-02 -5.99618673e-01 1.25011548e-01
5.74830055e-01 3.59935582e-01 -9.83050838e-02 -5.00434935e-01
8.63341212e-01 3.49971116e-01 3.49151462e-01 -7.07612574e-01
9.71587300e-01 3.30073804e-01 -1.25691749e-03 -5.02504051e-01
-2.48383194e-01 4.26736653e-01 -6.67648435e-01 -8.28666538e-02
7.25272775e-01 -6.61432564e-01 -6.85148358e-01 3.63053888e-01
-1.00813687e+00 -4.07977194e-01 -5.13147950e-01 -1.14437342e-01
-5.62424302e-01 -9.04289186e-02 -6.87070370e-01 -3.95490974e-01
-1.40003577e-01 -1.29204226e+00 1.14191186e+00 2.90001720e-01
-4.38719839e-02 -7.93523967e-01 -3.36140394e-01 3.20957974e-02
6.95993006e-01 4.52258408e-01 1.28105903e+00 1.00794509e-01
-1.20116866e+00 -1.58394292e-01 -3.26797180e-02 3.25980037e-02
2.21184194e-01 3.77356976e-01 -6.72656238e-01 -2.73049116e-01
-1.97861582e-01 -1.66368291e-01 7.32028782e-01 7.52774905e-03
1.43853724e+00 -6.23652577e-01 -2.08004624e-01 8.75319481e-01
1.34475565e+00 1.41021520e-01 8.36829185e-01 1.46773919e-01
9.78690922e-01 3.50398570e-01 2.89753266e-03 5.66282213e-01
2.43253246e-01 5.79233646e-01 4.47767317e-01 -1.90064907e-01
-2.91396797e-01 -7.23380029e-01 8.29678774e-02 9.02246058e-01
-3.80221158e-01 -4.72008556e-01 -4.01333451e-01 1.41486004e-01
-1.76794541e+00 -1.06711614e+00 3.15758914e-01 2.04971266e+00
7.46683061e-01 -1.42733473e-03 6.95580542e-02 -4.31665480e-02
7.21683383e-01 4.14563805e-01 -6.24574900e-01 -2.45204300e-01
-1.80061385e-01 5.54173701e-02 2.46213023e-02 2.30251178e-01
-5.24862707e-01 1.01926625e+00 6.30559731e+00 1.10418200e+00
-1.31663609e+00 -1.24051519e-01 6.39926136e-01 -2.43841633e-01
-1.13807869e+00 2.25707125e-02 -4.25913155e-01 5.64422786e-01
3.27289313e-01 -5.09217381e-02 9.79117334e-01 7.14551926e-01
1.55499205e-01 2.80943453e-01 -9.89020288e-01 9.41195369e-01
-4.53243479e-02 -1.72910249e+00 6.67295873e-01 7.61741325e-02
8.37798953e-01 -8.08240294e-01 3.67529780e-01 1.04456097e-01
2.26978078e-01 -1.07287610e+00 1.07023025e+00 6.15190089e-01
1.24104393e+00 -8.64758492e-01 -3.26837748e-02 4.22396302e-01
-1.16965103e+00 -7.41617903e-02 -4.07636613e-01 6.30657151e-02
2.17888847e-01 2.74541199e-01 -1.50599882e-01 3.92290741e-01
3.40706527e-01 5.79557657e-01 -7.21503496e-01 5.62787950e-01
-4.92049456e-01 7.03327805e-02 -1.16088316e-01 -1.63114697e-01
3.47277261e-02 -4.54009742e-01 3.26903045e-01 8.78677905e-01
5.66819906e-01 7.80861650e-04 2.00863361e-01 1.34446108e+00
-1.73138320e-01 1.08221292e-01 -4.55176353e-01 -2.54418850e-01
5.29487669e-01 1.46754467e+00 -8.79550993e-01 -2.80644298e-01
-7.36836046e-02 1.11001480e+00 3.71289790e-01 4.76461679e-01
-8.30062926e-01 -2.66322911e-01 4.15903270e-01 3.26521039e-01
3.32451671e-01 -3.31912309e-01 -3.51541936e-01 -1.33675230e+00
1.12729380e-02 -1.09373033e+00 -1.76460594e-01 -9.98098552e-01
-1.00687635e+00 8.41137469e-01 -1.60872281e-01 -1.20201540e+00
-1.10256650e-01 -4.61089879e-01 -7.26342857e-01 7.39924610e-01
-1.07944906e+00 -1.48375416e+00 -5.78043461e-01 6.57378852e-01
5.97037435e-01 -1.16868354e-01 8.02300513e-01 2.00537130e-01
-6.58762217e-01 6.01370096e-01 2.10902411e-02 -2.18588546e-01
5.92071116e-01 -1.09933627e+00 5.31954706e-01 6.55916631e-01
2.31520757e-01 8.83720756e-01 5.61387658e-01 -5.53133667e-01
-1.82869077e+00 -8.21844280e-01 3.45820874e-01 -2.51505494e-01
5.97386122e-01 -7.59583056e-01 -5.88703752e-01 4.26888764e-01
2.01691151e-01 -2.17422634e-01 3.51539850e-01 2.99194288e-02
-4.08924282e-01 -3.20970595e-01 -8.03999782e-01 1.11281550e+00
1.47595251e+00 -2.62198001e-01 -1.10001005e-01 1.25554100e-01
1.07134020e+00 -3.62074554e-01 -8.92623603e-01 1.55681238e-01
6.47453189e-01 -1.17041349e+00 8.97272885e-01 -3.50024223e-01
7.48870432e-01 -4.61843848e-01 2.78420672e-02 -1.33446968e+00
-8.85817647e-01 -9.89093721e-01 -1.11307040e-01 1.28273523e+00
6.15217865e-01 -4.54196960e-01 5.45625508e-01 6.68542445e-01
-3.03311676e-01 -8.42063308e-01 -4.43409652e-01 -4.75481659e-01
-1.31841883e-01 -2.39206329e-01 1.17817938e+00 6.11102283e-01
-2.38669544e-01 3.74661207e-01 -5.51663995e-01 -2.03643113e-01
3.59880179e-01 7.72267699e-01 8.59085798e-01 -1.01119494e+00
-5.32596946e-01 -6.88865840e-01 -1.46847218e-01 -1.22388804e+00
-1.98428854e-01 -7.35328734e-01 -2.73595303e-01 -1.52933669e+00
1.64968044e-01 -6.48507118e-01 -4.13627885e-02 3.83056492e-01
-3.21080768e-03 1.76622540e-01 5.61862409e-01 4.04819638e-01
-3.63500983e-01 6.97573841e-01 2.29613137e+00 -1.27071217e-01
-1.58121362e-01 -1.30325317e-01 -1.05369091e+00 4.18354899e-01
5.01049101e-01 -3.02245077e-02 -6.32431149e-01 -6.09799385e-01
5.28526723e-01 1.06232762e-01 3.47511142e-01 -8.15928340e-01
1.03346609e-01 -5.97613335e-01 7.34870136e-01 -1.96222797e-01
3.45304400e-01 -8.00035834e-01 7.63208866e-01 3.02110821e-01
-4.99827147e-01 1.46660760e-01 -1.05291329e-01 5.15702784e-01
-1.04068756e-01 1.01114273e-01 6.46071792e-01 -3.81349236e-01
-5.21495700e-01 6.43016398e-01 8.67966339e-02 -1.80754602e-01
9.20220733e-01 -3.27069879e-01 -5.28993845e-01 -3.52301866e-01
-6.02590501e-01 -5.34154102e-03 1.04564905e+00 6.79753006e-01
6.03779197e-01 -1.60797179e+00 -4.42205340e-01 6.55274808e-01
-4.18257574e-03 2.28202716e-01 5.15046537e-01 3.13531280e-01
-4.85573798e-01 5.54586202e-02 -5.35495400e-01 -3.73370022e-01
-5.88216066e-01 9.89319623e-01 -1.11865103e-01 -6.45443499e-02
-8.38541329e-01 6.72752380e-01 5.79277933e-01 3.82356085e-02
5.55405132e-02 -4.57316071e-01 5.82142398e-02 1.30095109e-01
6.24565542e-01 1.66755542e-01 -2.60405719e-01 -7.79639781e-02
1.75528433e-02 8.06036174e-01 1.26269028e-01 -6.88120350e-02
1.47427106e+00 -1.25240922e-01 -2.25895271e-01 3.38656127e-01
9.11523521e-01 2.17921644e-01 -1.77842295e+00 1.13856673e-01
-5.80888212e-01 -5.93470275e-01 -1.08318664e-01 -8.20793331e-01
-1.14356732e+00 9.46369588e-01 2.60129571e-01 3.68313134e-01
1.28341460e+00 -4.97359969e-02 6.76451921e-01 7.07819909e-02
5.48059523e-01 -9.66651618e-01 4.72696871e-01 1.68324843e-01
1.35956550e+00 -7.62388229e-01 -2.20502213e-01 -2.92327851e-01
-7.21896768e-01 1.17552662e+00 6.71657383e-01 -2.71193326e-01
1.55341670e-01 5.19093752e-01 -2.80055195e-01 -9.31576341e-02
-9.21594143e-01 1.18065983e-01 3.87924582e-01 5.83215773e-01
3.05147111e-01 1.13172635e-01 -1.12317190e-01 6.08985007e-01
-4.32832450e-01 -1.69150934e-01 2.30103061e-01 5.00120819e-01
-5.06620884e-01 -1.14640892e+00 -1.91523299e-01 2.19233498e-01
1.23938672e-01 -8.75247419e-02 -2.93682754e-01 7.39414752e-01
3.92093748e-01 4.57008988e-01 3.89908142e-02 -3.80126059e-01
4.15932238e-01 -1.81607872e-01 4.97192204e-01 -5.33615947e-01
-5.10618329e-01 2.75927842e-01 -3.41879547e-01 -7.22918272e-01
3.69042829e-02 -1.49054766e-01 -9.17198241e-01 -5.07018328e-01
-6.19356409e-02 -8.23055357e-02 5.18533289e-01 5.00703633e-01
5.90345144e-01 7.02822685e-01 6.13574743e-01 -9.66004014e-01
-3.32470536e-01 -7.61101484e-01 -6.90497458e-01 4.75234270e-01
3.00380677e-01 -4.94030207e-01 -3.18052649e-01 2.69928962e-01] | [11.429956436157227, -0.3955236077308655] |
0158a38d-3cd7-4145-b29e-15850dd7ed78 | dlolis-a-description-logic-based-text | 1303.5929 | null | http://arxiv.org/abs/1303.5929v1 | http://arxiv.org/pdf/1303.5929v1.pdf | DLOLIS-A: Description Logic based Text Ontology Learning | Ontology Learning has been the subject of intensive study for the past
decade. Researchers in this field have been motivated by the possibility of
automatically building a knowledge base on top of text documents so as to
support reasoning based knowledge extraction. While most works in this field
have been primarily statistical (known as light-weight Ontology Learning) not
much attempt has been made in axiomatic Ontology Learning (called heavy-weight
Ontology Learning) from Natural Language text documents. Heavy-weight Ontology
Learning supports more precise formal logic-based reasoning when compared to
statistical ontology learning. In this paper we have proposed a sound Ontology
Learning tool DLOL_(IS-A) that maps English language IS-A sentences into their
equivalent Description Logic (DL) expressions in order to automatically
generate a consistent pair of T-box and A-box thereby forming both regular
(definitional form) and generalized (axiomatic form) DL ontology. The current
scope of the paper is strictly limited to IS-A sentences that exclude the
possible structures of: (i) implicative IS-A sentences, and (ii) "Wh" IS-A
questions. Other linguistic nuances that arise out of pragmatics and epistemic
of IS-A sentences are beyond the scope of this present work. We have adopted
Gold Standard based Ontology Learning evaluation on chosen IS-A rich Wikipedia
documents. | ['Rupali KaPatel', 'Ankur Padia', 'Sourish Dasgupta', 'Prasenjit Majumder', 'Kushal Shah'] | 2013-03-24 | null | null | null | null | ['formal-logic'] | ['reasoning'] | [ 1.44849360e-01 8.66983771e-01 -2.33868450e-01 -6.43048942e-01
-2.82352120e-01 -3.48200232e-01 8.48599672e-01 5.09749651e-01
-4.71409112e-01 8.78512800e-01 4.86571968e-01 -8.63183379e-01
-9.01967406e-01 -1.00865805e+00 -5.41584253e-01 6.75792340e-03
3.86172719e-02 8.39004457e-01 5.58986366e-01 -6.16068244e-01
1.69046164e-01 1.79334387e-01 -1.88257194e+00 6.30038083e-01
7.32489049e-01 5.92671454e-01 2.45394170e-01 2.24102795e-01
-9.02146697e-01 1.19679928e+00 -3.45375836e-01 -5.11113644e-01
-1.27976060e-01 -2.85298795e-01 -1.53085697e+00 -1.31249666e-01
2.87712723e-01 2.19406456e-01 2.73153007e-01 1.16683424e+00
1.73806652e-01 4.64997739e-02 6.86143756e-01 -1.22254765e+00
-5.40839732e-01 1.10323727e+00 1.94799170e-01 2.40042135e-01
8.63562346e-01 -4.23743993e-01 1.30106986e+00 -3.98629844e-01
8.04309905e-01 1.80268860e+00 4.77546990e-01 5.45206428e-01
-9.97979939e-01 -3.78084272e-01 -2.51470692e-02 6.07286215e-01
-1.23164117e+00 -1.65906101e-01 7.52929628e-01 -4.67356682e-01
1.36450720e+00 3.01357329e-01 5.43281078e-01 5.20471156e-01
3.47774982e-01 6.85699701e-01 1.34180629e+00 -1.33407807e+00
3.45001906e-01 6.33700430e-01 7.01919138e-01 5.37371397e-01
4.83501792e-01 -2.71805376e-01 -3.59258503e-01 1.81709994e-02
4.66449782e-02 -4.92068708e-01 2.05207944e-01 -9.85836312e-02
-6.06411099e-01 9.98602569e-01 -1.55172408e-01 8.94159615e-01
-2.53293276e-01 -1.46535829e-01 4.91456717e-01 4.26967233e-01
-1.36254594e-01 4.32010412e-01 -7.77886987e-01 7.95825571e-02
-3.96917284e-01 6.74953461e-01 1.22476935e+00 1.03701460e+00
7.15404928e-01 -3.52042437e-01 3.63556355e-01 6.10236287e-01
7.28953302e-01 1.65974408e-01 7.00462937e-01 -9.54846025e-01
2.63575166e-01 1.05034447e+00 1.57155558e-01 -6.66724682e-01
-4.90319222e-01 1.54333591e-01 -3.01230587e-02 1.38960063e-01
4.26332653e-01 -1.82830900e-01 -2.30455086e-01 1.51667893e+00
3.87726843e-01 -3.46233785e-01 7.18262494e-01 2.09089860e-01
1.18909073e+00 4.61543590e-01 3.80316019e-01 -2.87162930e-01
1.86700678e+00 -2.55006015e-01 -1.19268525e+00 -1.80623904e-01
1.01335239e+00 -6.88691735e-01 1.23997104e+00 3.76664013e-01
-7.90432930e-01 -1.04828931e-01 -9.43262160e-01 -2.20284358e-01
-9.71705079e-01 -4.48897213e-01 9.29449975e-01 8.18053305e-01
-5.85595369e-01 1.90968618e-01 -2.30767235e-01 -9.09361839e-01
1.09127842e-01 3.50000173e-01 -4.00316417e-01 8.09312165e-02
-1.66100717e+00 1.35138214e+00 1.35934663e+00 -4.39251930e-01
-4.53548208e-02 -2.57501930e-01 -1.09554458e+00 -1.43909812e-01
8.34934711e-01 -3.89541119e-01 1.22896552e+00 -8.64423454e-01
-1.40889513e+00 1.23469746e+00 -5.22568375e-02 -8.02406311e-01
1.49429083e-01 -9.46960077e-02 -9.49648380e-01 -5.21729402e-02
2.74890482e-01 2.60244071e-01 1.11575231e-01 -1.19904840e+00
-9.47176635e-01 -5.18905699e-01 6.66468441e-01 7.60451471e-03
-2.49324650e-01 6.03601635e-01 2.70860344e-01 -2.80784637e-01
1.96498573e-01 -7.02136636e-01 3.44175041e-01 -4.63344306e-01
-2.22569600e-01 -9.60572183e-01 6.85859323e-01 -3.14411253e-01
1.67008460e+00 -1.59883296e+00 -4.08433586e-01 2.15661660e-01
-3.13089043e-01 2.24196836e-01 4.33013886e-01 6.72335625e-01
-2.32997149e-01 3.33325058e-01 6.58944622e-02 5.25024891e-01
6.09075129e-01 9.23232198e-01 -4.50807512e-01 -1.23871133e-01
-1.79478258e-01 4.39717919e-01 -8.32040191e-01 -1.08685672e+00
4.81676280e-01 -1.40530780e-01 -7.42957056e-01 -1.56288400e-01
-6.30445421e-01 -1.43834338e-01 -6.06708705e-01 4.06324446e-01
2.60809720e-01 2.01881453e-01 6.31711125e-01 -2.78241307e-01
-3.15067887e-01 5.96777022e-01 -1.42352605e+00 1.28671324e+00
-7.16886461e-01 5.15414894e-01 -5.65610230e-01 -1.13786709e+00
8.25451374e-01 7.56687880e-01 2.34102562e-01 -5.76419532e-01
2.74688274e-01 5.17552197e-01 3.99337679e-01 -1.11125100e+00
1.77949473e-01 -9.69537079e-01 -6.45553395e-02 2.44538069e-01
2.26806223e-01 -2.59106636e-01 6.97576821e-01 -2.10004061e-01
5.98016083e-01 4.37196612e-01 1.07437253e+00 -7.34395921e-01
1.10143209e+00 1.56371981e-01 5.24023533e-01 5.41435719e-01
6.20668121e-02 -3.44615877e-01 7.01094747e-01 -6.42557919e-01
-8.59884381e-01 -8.42275143e-01 -8.64224494e-01 9.38566983e-01
-7.33642504e-02 -8.03616464e-01 -5.56946635e-01 -5.31236649e-01
-1.44220188e-01 1.53477335e+00 -3.84307414e-01 3.52064967e-01
-4.74302441e-01 -4.72700864e-01 5.51988363e-01 7.25024343e-02
4.48498338e-01 -1.40917659e+00 -6.86292589e-01 4.55329776e-01
-1.75840959e-01 -1.16832137e+00 6.35771215e-01 3.77679884e-01
-8.58233511e-01 -1.40666664e+00 4.33162421e-01 -8.15387607e-01
2.41394266e-01 -6.07757449e-01 9.54909325e-01 7.67790806e-03
3.60492840e-02 2.01451749e-01 -7.18357146e-01 -1.36327302e+00
-8.09431195e-01 -3.47623765e-01 1.13541253e-01 -1.96078941e-01
1.23560596e+00 -5.29098928e-01 3.54629338e-01 -1.44383814e-02
-1.19556201e+00 -3.00698638e-01 3.17625970e-01 3.58200341e-01
3.23909819e-01 7.76099682e-01 4.81121063e-01 -1.20490789e+00
7.44341433e-01 -4.51236993e-01 -5.49672306e-01 4.38744217e-01
-6.13871098e-01 5.16393602e-01 5.22781193e-01 -1.04297400e-01
-1.30485857e+00 -3.78797024e-01 -3.25895399e-01 4.97439712e-01
-4.92328167e-01 9.28429246e-01 -5.39268315e-01 3.09694022e-01
8.52415979e-01 -1.19419617e-03 -9.26729813e-02 -4.39566404e-01
1.49505004e-01 1.00516510e+00 2.23831877e-01 -1.19601810e+00
6.79235518e-01 3.56434554e-01 2.17168018e-01 -1.22230792e+00
-1.15213835e+00 -4.82703239e-01 -6.56568587e-01 -1.35715425e-01
9.44945514e-01 -5.27985036e-01 -7.09220946e-01 -2.24259838e-01
-1.06543899e+00 -4.29138727e-02 -4.74812537e-01 5.57537735e-01
-7.75524020e-01 3.74371737e-01 3.77741002e-04 -1.18460739e+00
-4.83784601e-02 -6.94129109e-01 4.68859762e-01 8.39436240e-03
-7.20586360e-01 -1.31275070e+00 2.23746356e-02 5.09383619e-01
-1.39183074e-01 1.62821561e-01 1.50596333e+00 -1.15405118e+00
3.74658331e-02 -3.26808512e-01 1.99138835e-01 3.38927209e-01
1.98874623e-01 -1.29069224e-01 -7.21880257e-01 3.53391320e-01
1.00560121e-01 -3.61657619e-01 3.25391404e-02 1.29303932e-01
5.74896634e-01 -6.52339399e-01 -7.29299858e-02 -1.00993425e-01
1.79199529e+00 3.89208853e-01 1.02203119e+00 8.97025704e-01
-1.62107050e-01 8.80411804e-01 6.90197647e-01 2.54911005e-01
5.56399405e-01 8.36803675e-01 -2.45003123e-02 7.40866899e-01
3.65027115e-02 -1.77166507e-01 2.47070000e-01 5.79395533e-01
-1.54142022e-01 -4.13032323e-02 -1.20192111e+00 6.45731151e-01
-1.89052987e+00 -1.40303695e+00 -4.64851320e-01 1.88996291e+00
1.24806058e+00 5.75114667e-01 -9.02489480e-03 4.45056558e-01
4.59222257e-01 -1.55414537e-01 4.35954183e-01 -7.86519945e-01
-2.87208825e-01 3.93818587e-01 -4.79083471e-02 1.05523372e+00
-9.40565526e-01 1.00157773e+00 5.01030111e+00 7.09240675e-01
-5.40289819e-01 7.23752901e-02 -4.84934598e-01 5.08760870e-01
-5.57699323e-01 6.40999019e-01 -1.00415289e+00 3.41338694e-01
8.84917438e-01 -4.53332961e-01 -6.15786016e-03 9.61131275e-01
3.17372203e-01 -3.13928366e-01 -1.03903699e+00 5.77376068e-01
6.25320673e-02 -1.25251842e+00 4.84565586e-01 4.35612537e-02
4.23273206e-01 -3.69139403e-01 -7.72441447e-01 4.47160333e-01
3.36578369e-01 -7.80353189e-01 8.99621904e-01 5.01729548e-01
2.93313801e-01 -5.79371810e-01 1.27685726e+00 5.68366885e-01
-8.74588311e-01 -2.83322811e-01 -5.12931824e-01 -4.99143004e-01
1.26827911e-01 4.12217826e-01 -6.88036025e-01 7.48837709e-01
7.50791132e-01 2.69505620e-01 -3.53490502e-01 7.40886271e-01
-4.34363276e-01 4.79794174e-01 -4.90829021e-01 -3.84748280e-01
2.54535794e-01 -2.14990735e-01 6.49461985e-01 1.24798751e+00
6.81361705e-02 3.54191959e-01 5.42758778e-02 6.30224884e-01
3.96295756e-01 4.81285006e-01 -7.26035476e-01 2.07750425e-02
3.32817614e-01 6.97993755e-01 -5.66362083e-01 -4.81262654e-01
-6.65593684e-01 9.96116847e-02 -6.41707554e-02 -1.01916768e-01
-2.42886424e-01 -5.58351398e-01 3.15217495e-01 3.90643269e-01
-4.59209606e-02 2.43822768e-01 9.59147885e-02 -9.27541077e-01
1.55344279e-02 -7.66869545e-01 7.53232360e-01 -7.23161697e-01
-1.23010635e+00 2.56806374e-01 6.19079828e-01 -9.60507333e-01
-4.39384520e-01 -1.08939767e+00 -4.28228170e-01 6.43740296e-01
-1.47729778e+00 -1.27117825e+00 -1.20770941e-02 4.67205882e-01
3.64583492e-01 -3.17203522e-01 1.27287436e+00 1.49684489e-01
2.81510372e-02 -8.34647790e-02 -3.48547995e-01 3.97940800e-02
3.45466852e-01 -1.47630835e+00 -5.99655330e-01 6.35113478e-01
2.60895863e-02 9.48953629e-01 1.39013314e+00 -5.54346144e-01
-9.47288752e-01 -5.22760093e-01 1.99388719e+00 -6.13561273e-01
1.01562047e+00 1.30388290e-01 -8.10720026e-01 1.07011890e+00
3.14753652e-01 -2.25745007e-01 1.01975691e+00 3.12990069e-01
-2.39653409e-01 -1.30444035e-01 -1.15496588e+00 7.06182063e-01
5.92362106e-01 -5.81518471e-01 -1.78070831e+00 4.26320761e-01
4.44247872e-01 5.35200052e-02 -8.77895296e-01 3.35577399e-01
4.70148385e-01 -1.05495584e+00 7.42273092e-01 -1.02640104e+00
1.25136077e-01 -6.14374876e-01 -4.21628952e-01 -5.04537284e-01
9.67071727e-02 -5.09718180e-01 2.11529750e-02 1.14439166e+00
4.43019509e-01 -5.91120422e-01 3.98263872e-01 6.38122678e-01
-1.42174050e-01 -3.32482219e-01 -9.57671463e-01 -9.07294393e-01
2.94554740e-01 -9.69269931e-01 4.81049776e-01 1.04537404e+00
6.39911711e-01 5.91991663e-01 2.19913423e-01 5.64008392e-02
6.07388079e-01 7.57035427e-03 5.97533107e-01 -1.76406717e+00
-1.05404213e-01 -4.59643632e-01 -6.93946362e-01 -4.51880872e-01
5.74628770e-01 -1.14366782e+00 -2.61446357e-01 -1.86284602e+00
-5.95038868e-02 -2.73831218e-01 1.09376840e-01 6.31116092e-01
3.16850752e-01 -3.10389310e-01 -2.10038797e-05 -1.76504910e-01
-4.44749534e-01 1.76014096e-01 9.53566313e-01 1.69569045e-01
-6.08517826e-02 -1.55346870e-01 -7.09283471e-01 1.40034628e+00
7.01929033e-01 -6.01001382e-01 -6.59731686e-01 1.51218250e-01
8.86950433e-01 -1.37822986e-01 2.49426767e-01 -9.57828999e-01
3.34746502e-02 -5.82875788e-01 -3.88877958e-01 -1.34789631e-01
-2.07564548e-01 -1.28922307e+00 1.18951999e-01 4.30400401e-01
-3.87446105e-01 -1.84208468e-01 2.68801659e-01 1.42144546e-01
-3.39164376e-01 -1.23766792e+00 6.94129705e-01 -3.26523691e-01
-1.12470555e+00 -5.08902013e-01 -5.21203458e-01 4.33532298e-01
9.84785020e-01 -3.99641871e-01 5.25656529e-02 -1.20842889e-01
-9.31195080e-01 -8.37360546e-02 8.85726511e-02 3.42042536e-01
1.89813122e-01 -9.97788906e-01 -5.71723998e-01 -3.91464025e-01
4.78072554e-01 -1.35563046e-01 -3.82080488e-02 5.41507185e-01
-8.68365347e-01 7.95472205e-01 -2.79005706e-01 8.92656595e-02
-1.17493951e+00 4.32000190e-01 3.46824288e-01 -2.83372223e-01
-6.66844666e-01 3.98120284e-01 -3.57936829e-01 -8.30502570e-01
2.29785919e-01 -2.71985471e-01 -8.18770051e-01 1.55717328e-01
4.70416903e-01 7.69188926e-02 -4.03392874e-02 -7.38812506e-01
-5.59323251e-01 3.62533033e-01 1.55006886e-01 -1.59456342e-01
1.38981605e+00 -4.62783985e-02 -5.48175097e-01 7.70078897e-01
6.70420170e-01 2.64889657e-01 -1.41950950e-01 -3.54006469e-01
8.89381528e-01 -9.00046751e-02 -1.01450510e-01 -8.60849798e-01
2.35574562e-02 5.59292495e-01 3.55233967e-01 6.74512446e-01
5.88345885e-01 4.25734609e-01 3.36273521e-01 9.18913662e-01
5.06984890e-01 -1.58843017e+00 -2.64602691e-01 7.72475779e-01
9.49087918e-01 -9.63324249e-01 3.17367375e-01 -6.23255610e-01
-3.80595773e-01 1.44036770e+00 4.32832450e-01 1.93044990e-01
7.50681043e-01 2.94696510e-01 1.30138323e-01 -6.08658433e-01
-5.90356648e-01 -5.63304126e-01 2.12931246e-01 6.83391929e-01
7.97977328e-01 4.95997220e-02 -1.08597314e+00 8.18662643e-01
-6.81815684e-01 3.09893310e-01 5.98651886e-01 1.12291694e+00
-7.68277466e-01 -1.40109825e+00 -2.99628615e-01 3.88059288e-01
-6.67492509e-01 2.67872144e-03 -1.90097392e-01 1.33594882e+00
5.99224508e-01 1.04891634e+00 -1.10675544e-01 9.32772532e-02
4.62100267e-01 5.24335921e-01 6.54786825e-01 -9.90883291e-01
-2.56781876e-01 -4.35787380e-01 7.37299800e-01 -2.74111200e-02
-9.90163565e-01 -6.42137706e-01 -1.80952680e+00 -1.30887195e-01
-3.41661930e-01 4.31176782e-01 6.03392184e-01 1.74559402e+00
-4.23252106e-01 3.03520918e-01 -3.12007397e-01 -1.37604289e-02
-5.40106893e-01 -9.12964523e-01 -6.92589641e-01 6.01254523e-01
-1.65006056e-01 -7.10451901e-01 -4.48058575e-01 1.38586178e-01] | [9.77171802520752, 8.726953506469727] |
aed133b8-1a2d-47dd-b22f-9234428624dd | a-conglomerate-of-multiple-ocr-table | 2010.08591 | null | https://arxiv.org/abs/2010.08591v1 | https://arxiv.org/pdf/2010.08591v1.pdf | A Conglomerate of Multiple OCR Table Detection and Extraction | Information representation as tables are compact and concise method that eases searching, indexing, and storage requirements. Extracting and cloning tables from parsable documents is easier and widely used, however industry still faces challenge in detecting and extracting tables from OCR documents or images. This paper proposes an algorithm that detects and extracts multiple tables from OCR document. The algorithm uses a combination of image processing techniques, text recognition and procedural coding to identify distinct tables in same image and map the text to appropriate corresponding cell in dataframe which can be stored as Comma-separated values, Database, Excel and multiple other usable formats. | ['Sumit Kumar', 'Raj Ratn Pranesh', 'Smita Pallavi'] | 2020-10-16 | null | null | null | null | ['table-detection'] | ['miscellaneous'] | [ 5.18975854e-01 -5.24938881e-01 -1.43326238e-01 -1.31848127e-01
-4.08517540e-01 -1.20770359e+00 1.37985483e-01 8.26394916e-01
-1.83998123e-01 6.68160081e-01 1.79270163e-01 -4.74637091e-01
-3.18641096e-01 -8.59499753e-01 -3.51222306e-01 8.02988857e-02
3.43436688e-01 5.38333356e-01 4.24328357e-01 1.50210429e-02
9.28066373e-01 1.16910374e+00 -1.56273901e+00 6.79021358e-01
5.22225261e-01 9.08273935e-01 2.29149997e-01 1.11990726e+00
-8.22088540e-01 9.40732598e-01 -9.41249907e-01 -3.92612278e-01
4.23692018e-01 -7.64940754e-02 -6.49223626e-01 3.12433958e-01
6.34934157e-02 -2.97450453e-01 -3.07774365e-01 9.59491014e-01
-2.85028312e-02 -3.46306652e-01 7.84251094e-01 -1.04856849e+00
-8.79523635e-01 4.16654527e-01 -9.31024790e-01 2.31044129e-01
7.57992208e-01 -6.77371025e-01 1.73641235e-01 -9.78472948e-01
8.04434717e-01 1.35313702e+00 4.80446756e-01 -3.64837587e-01
-6.40049756e-01 -6.46987677e-01 -5.53255796e-01 -2.03117624e-01
-1.61124277e+00 -1.84281141e-01 3.50072503e-01 -4.84995037e-01
1.17992139e+00 9.02430654e-01 5.57304621e-01 -1.05026446e-01
6.55580342e-01 5.05681753e-01 1.11733782e+00 -8.91950250e-01
-1.77471742e-01 4.24257994e-01 3.21136713e-01 5.08167565e-01
8.61663103e-01 -6.46234632e-01 -4.88623589e-01 2.89138764e-01
8.51916015e-01 4.08576936e-01 1.37070164e-01 2.70470213e-02
-1.24092078e+00 1.98377982e-01 -6.60004690e-02 3.28453571e-01
-1.54184192e-01 -2.73605168e-01 3.65826905e-01 2.04913378e-01
-2.86245435e-01 2.60697424e-01 1.01895798e-02 -2.80562788e-01
-9.64594007e-01 -4.51323427e-02 4.68902886e-01 1.45611870e+00
4.76085514e-01 -2.95407832e-01 6.23874739e-02 7.30450630e-01
3.02331954e-01 6.66640282e-01 5.54573119e-01 -4.29026753e-01
8.26804221e-01 1.35275888e+00 -2.16644316e-04 -1.45907819e+00
-1.42580211e-01 4.88289177e-01 -5.42587042e-01 2.31913552e-01
6.64248541e-02 3.60700905e-01 -1.24988163e+00 1.49624839e-01
3.20652500e-02 -1.10700655e+00 1.14730954e-01 2.88861752e-01
8.24421346e-01 1.15470958e+00 -2.97617763e-01 6.74945563e-02
1.87930250e+00 -4.06912625e-01 -1.17370081e+00 -5.58687039e-02
1.61016613e-01 -1.61964405e+00 6.26584530e-01 6.16972923e-01
-1.10401392e+00 -4.05214369e-01 -1.39111626e+00 -6.65666997e-01
-1.19735539e+00 6.31118357e-01 3.45659792e-01 5.66507161e-01
-6.48100078e-01 5.99466451e-02 -3.87587577e-01 -5.76244533e-01
3.27316731e-01 5.30927122e-01 -6.91865385e-01 1.56628653e-01
-3.26141059e-01 7.01041460e-01 6.33770585e-01 -7.02296346e-02
3.63125533e-01 -3.77977714e-02 -6.77086651e-01 -3.54891196e-02
2.66447812e-01 4.61017191e-02 7.11213529e-01 -4.25813138e-01
-1.02830493e+00 1.20454288e+00 3.79440710e-02 -2.08015621e-01
2.73785293e-01 -1.11138374e-01 -6.99992120e-01 3.41872543e-01
1.47459446e-04 4.42090333e-01 7.81611323e-01 -9.39779580e-01
-8.81433189e-01 -6.64376140e-01 -5.04799247e-01 1.64928153e-01
2.98458040e-02 6.03491724e-01 -6.60802305e-01 -9.30313468e-01
7.51203358e-01 -5.33683479e-01 3.86295736e-01 3.53514180e-02
-7.38892317e-01 1.38615012e-01 1.14931488e+00 -1.06908298e+00
1.59975970e+00 -2.08205462e+00 -5.39231479e-01 6.76527441e-01
-4.34401929e-02 1.81500509e-01 6.04546070e-01 8.05472732e-01
1.58019438e-02 4.49254453e-01 1.44645572e-01 5.18660247e-01
-1.24385878e-01 -3.04193478e-02 -3.11616987e-01 2.01325282e-01
-2.35234834e-02 3.56715411e-01 8.04326590e-03 -1.05652499e+00
3.25342536e-01 3.87868464e-01 1.15511931e-01 -1.54548019e-01
4.20135200e-01 -5.65347254e-01 -2.60583460e-01 1.21919191e+00
8.69086444e-01 1.54213801e-01 1.90042049e-01 -4.21728104e-01
-6.90938830e-01 2.34849155e-01 -2.02158332e+00 6.33734405e-01
2.77436562e-02 1.04232633e+00 -1.88748181e-01 -7.83246279e-01
1.47161317e+00 1.55620694e-01 -3.41533348e-02 -7.73275673e-01
2.69904286e-01 2.80503631e-01 -5.42196870e-01 -5.70801675e-01
1.08260489e+00 5.27976215e-01 -2.68415865e-02 4.48063552e-01
-4.41821188e-01 -2.52706110e-01 7.82981873e-01 3.85285795e-01
6.38743281e-01 -1.57908738e-01 8.40169966e-01 -1.87547401e-01
7.04325378e-01 3.40524763e-01 5.83932400e-02 5.90290248e-01
1.30901456e-01 5.31798124e-01 3.70494306e-01 -5.85566461e-01
-1.16001821e+00 -1.01367605e+00 -3.86005521e-01 6.15426481e-01
1.78130507e-01 -4.07032758e-01 -4.97176558e-01 -4.53254655e-02
2.64939547e-01 1.94905043e-01 -2.84062505e-01 4.26778376e-01
-7.26052284e-01 -1.30278513e-01 6.71993345e-02 5.81578374e-01
3.56220514e-01 -1.06202173e+00 -8.36819530e-01 2.32484967e-01
2.76827753e-01 -8.58613074e-01 -3.84130061e-01 3.27923298e-01
-9.54490066e-01 -8.55615199e-01 -4.18874323e-01 -1.19844675e+00
1.00583208e+00 5.10287344e-01 7.93086290e-01 2.29417697e-01
-9.95972097e-01 1.93921193e-01 -2.31118158e-01 -1.03449917e+00
-3.60816538e-01 -2.47990843e-02 -3.10546309e-01 -4.27183181e-01
7.60248303e-01 1.25217780e-01 -2.78867573e-01 1.70039326e-01
-1.13406670e+00 6.36385828e-02 5.08476913e-01 9.55826119e-02
8.52772951e-01 4.16694462e-01 -1.03074849e-01 -7.29739904e-01
7.43368089e-01 2.00308442e-01 -8.25646698e-01 6.36361957e-01
-5.50818563e-01 2.05033660e-01 4.39772874e-01 1.72480181e-01
-5.77949524e-01 2.96203583e-01 4.85259593e-01 3.71320456e-01
-1.27924949e-01 2.56519407e-01 -6.29644915e-02 3.98081616e-02
4.65374440e-01 4.50604886e-01 4.23851125e-02 -5.75502098e-01
1.15261845e-01 1.56498492e+00 7.71663368e-01 4.02009785e-02
1.00789428e+00 3.55783761e-01 -2.29780138e-01 -9.34501886e-01
1.40197814e-01 -7.42901623e-01 -9.97808278e-01 -3.65508527e-01
8.04048002e-01 -7.18538940e-01 -4.88159239e-01 2.71226734e-01
-9.31355238e-01 7.66166270e-01 6.11557029e-02 3.64644974e-01
6.00578561e-02 2.40307420e-01 -5.54450870e-01 -8.15258741e-01
-4.95570332e-01 -8.42700481e-01 8.74822199e-01 5.51299930e-01
-3.24985147e-01 -1.78813532e-01 -2.13703781e-01 2.62548059e-01
1.54853567e-01 2.60570168e-01 9.90027010e-01 -3.37429106e-01
-9.54653263e-01 -8.02026272e-01 -5.03315330e-01 -3.23250651e-01
5.05258381e-01 8.96364152e-01 -2.88071185e-01 2.12459311e-01
-3.51685613e-01 5.37080392e-02 5.34761071e-01 -8.74582827e-02
1.09430861e+00 -4.31350291e-01 -7.03704357e-01 3.31333667e-01
1.87617695e+00 1.34676445e+00 1.23000538e+00 9.86448944e-01
4.33763146e-01 3.87783945e-01 5.47817290e-01 3.63656402e-01
-9.53969429e-04 2.05681995e-01 -3.60042930e-01 5.47631122e-02
2.42490798e-01 -7.45678917e-02 -9.50388238e-02 7.11587131e-01
7.61936083e-02 -3.06868017e-01 -9.81203616e-01 2.17491210e-01
-1.09768260e+00 -9.42436934e-01 -1.35554761e-01 2.07543373e+00
9.15100455e-01 3.68848503e-01 1.22703463e-01 8.79783928e-01
1.04664016e+00 -3.63480538e-01 -8.35719635e-04 -1.07909203e+00
-3.44923921e-02 3.43223631e-01 9.77799952e-01 2.66458362e-01
-8.97388041e-01 5.86783648e-01 7.25540543e+00 4.67840165e-01
-1.12812281e+00 -7.26127267e-01 6.88916445e-01 1.65129855e-01
1.04349345e-01 -2.27815092e-01 -1.05561352e+00 4.26224619e-01
6.27967238e-01 -4.59789664e-01 3.04943204e-01 8.32029462e-01
-2.18645409e-01 -6.13793194e-01 -7.77030230e-01 1.34043705e+00
1.02134436e-01 -1.40994775e+00 6.43150210e-01 5.41633070e-02
2.09246919e-01 -9.32074606e-01 8.50195736e-02 -3.94675404e-01
-9.47091952e-02 -1.07685280e+00 1.06484985e+00 4.27968472e-01
7.87320673e-01 -8.40563655e-01 5.46318889e-01 -4.76170063e-01
-1.52044058e+00 3.75463627e-02 -5.99895298e-01 1.05504252e-01
-2.85439610e-01 2.57811658e-02 -8.21615040e-01 1.49724633e-01
9.51684237e-01 1.39403865e-01 -1.13720381e+00 1.13331401e+00
4.68131214e-01 -1.54749975e-01 -5.39167345e-01 -4.48624462e-01
-6.81740269e-02 -4.61956829e-01 6.45434782e-02 1.57638288e+00
6.99681044e-01 8.22519213e-02 -3.01137477e-01 3.04633439e-01
-8.74330178e-02 6.96652532e-01 -6.39827967e-01 -6.65077329e-01
8.90718400e-01 1.11013269e+00 -1.80183744e+00 -5.69626153e-01
-3.07717651e-01 6.65460408e-01 -3.18169385e-01 -1.61617562e-01
-3.96170259e-01 -1.56942105e+00 -1.46812513e-01 3.91328901e-01
4.31077361e-01 -4.04407948e-01 -8.03266287e-01 -4.09057558e-01
2.20185176e-01 -1.11151385e+00 4.41951185e-01 -9.85917628e-01
-2.45146886e-01 4.94102925e-01 -4.10610177e-02 -1.35594344e+00
6.09773248e-02 -9.86645579e-01 -2.62546480e-01 6.23398960e-01
-5.21400452e-01 -6.35769844e-01 -4.54631627e-01 5.03400564e-01
3.99930328e-01 -3.71955365e-01 5.34592390e-01 2.44618341e-01
-8.48897576e-01 4.74276632e-01 6.49377048e-01 5.66100657e-01
6.92750871e-01 -1.48111463e+00 2.37293124e-01 9.87715244e-01
2.68602669e-01 9.31084275e-01 5.01899958e-01 -6.89092398e-01
-1.64413524e+00 -4.75725383e-01 1.20159316e+00 -2.47835010e-01
2.60194421e-01 -6.63564622e-01 -7.65398204e-01 7.39732563e-01
4.11271870e-01 -4.09270883e-01 5.32152712e-01 -7.22360194e-01
-2.15876624e-01 -6.80869043e-01 -1.24261057e+00 5.32762706e-01
2.77252346e-01 -6.33174062e-01 -7.46540368e-01 3.14480036e-01
6.66112825e-02 -6.35569572e-01 -6.82756901e-01 -3.75174075e-01
8.53550315e-01 -7.09288418e-01 9.66789722e-01 -1.14689842e-01
3.52860868e-01 -7.28254020e-01 -1.52106002e-01 -2.26133123e-01
-5.70483692e-02 -6.91110015e-01 4.04096007e-01 1.44366336e+00
5.58050811e-01 -4.27888989e-01 6.15955532e-01 6.90147281e-01
3.90456289e-01 -2.45496541e-01 -4.92102504e-01 -5.22512734e-01
-5.94873667e-01 3.18988293e-01 7.34076440e-01 5.75650096e-01
2.58722305e-01 2.45684832e-01 -6.87353238e-02 -1.16766252e-01
3.15565377e-01 2.63396651e-01 5.60042202e-01 -1.31444812e+00
5.25339603e-01 -1.66356593e-01 -7.68578827e-01 -4.67285961e-01
-9.43719029e-01 -6.93669915e-01 -2.83699155e-01 -2.00197458e+00
4.86094505e-02 -3.03127676e-01 9.94235948e-02 1.83186397e-01
3.70868802e-01 3.89519930e-01 3.97594452e-01 4.79152590e-01
-4.89180654e-01 -5.28645337e-01 7.96742618e-01 -8.69387463e-02
-2.86479890e-01 -6.21017575e-01 -6.54222310e-01 4.31559235e-01
7.11029053e-01 -7.26369560e-01 -1.37069494e-01 -1.21119127e-01
4.88479435e-01 -8.14288855e-02 -3.84965718e-01 -1.27694762e+00
4.37164724e-01 -1.82930291e-01 1.25179935e+00 -1.61826789e+00
-2.73271918e-01 -1.01787210e+00 5.59118807e-01 4.72249150e-01
-1.87633425e-01 9.67945635e-01 3.81721020e-01 2.70990074e-01
-2.99419314e-01 -7.09821582e-01 7.24988341e-01 -3.50805432e-01
-6.17453039e-01 -3.07182252e-01 -9.33485389e-01 -4.51164812e-01
1.18731344e+00 -1.15258193e+00 -4.06674445e-01 2.30192356e-02
-2.92521745e-01 -2.22999230e-01 6.12051189e-01 3.87272447e-01
7.31290221e-01 -1.32471144e+00 -9.87912416e-02 4.79196489e-01
9.69578773e-02 -1.76866323e-01 -2.96609670e-01 -1.32982852e-04
-2.02126169e+00 9.15058851e-01 -7.61201322e-01 -3.27907354e-01
-1.87315023e+00 7.43758619e-01 -2.55686045e-01 3.62132043e-02
-6.07387781e-01 3.94854754e-01 -6.64004207e-01 1.97280496e-01
2.86390185e-01 -7.47639418e-01 -6.54023886e-01 2.55835384e-01
1.03931475e+00 7.25363553e-01 3.43487889e-01 -5.50277948e-01
-6.26691699e-01 1.01690686e+00 -5.14855981e-01 -6.71037510e-02
8.54023397e-01 -2.98662931e-01 -4.74397779e-01 4.21321839e-01
1.20958674e+00 4.84693676e-01 -4.49691653e-01 2.52924800e-01
2.63889581e-01 -7.83849120e-01 -3.29695046e-01 -6.70665681e-01
-6.00881457e-01 4.41998094e-01 8.37630332e-01 5.76645195e-01
9.91919219e-01 -3.05132896e-01 5.96468449e-01 5.75091481e-01
8.29575211e-02 -1.56446862e+00 -6.05175011e-02 1.53837204e-01
9.75461006e-01 -8.36044550e-01 5.68716705e-01 -6.56593323e-01
-2.96963841e-01 1.87169540e+00 3.71276379e-01 -7.02186301e-02
7.11648405e-01 9.12856162e-01 2.56080925e-01 -1.38896942e-01
-3.22232544e-01 2.44170442e-01 1.65539607e-01 6.83413506e-01
7.98196077e-01 -5.54253906e-02 -4.92167741e-01 2.17012525e-01
-5.94393551e-01 1.23499975e-01 8.87021661e-01 1.76812172e+00
-8.85969043e-01 -9.18324351e-01 -1.19252527e+00 7.64648974e-01
-9.37434077e-01 1.43274218e-01 -8.12861204e-01 1.02094030e+00
-1.99745581e-01 7.06056476e-01 5.29751182e-01 -1.43130288e-01
3.76111090e-01 2.68466473e-02 4.46000576e-01 -2.05207646e-01
-5.40761530e-01 3.43746059e-02 -2.77153224e-01 -3.17725427e-02
-1.77314028e-01 -5.61535776e-01 -1.33275259e+00 -5.21690845e-01
-6.48831502e-02 8.43928680e-02 9.88126993e-01 4.49268401e-01
3.65449131e-01 3.30429435e-01 1.83130443e-01 -9.29451287e-02
2.76978891e-02 -5.77795446e-01 -6.31429911e-01 1.56970784e-01
2.18488753e-01 -2.97661692e-01 2.38919035e-02 5.88103533e-01] | [11.755926132202148, 2.832470655441284] |
fc7058a7-15c2-4167-8d6a-0e3d14bf4903 | multi-label-video-classification-for | 2305.17338 | null | https://arxiv.org/abs/2305.17338v1 | https://arxiv.org/pdf/2305.17338v1.pdf | Multi-label Video Classification for Underwater Ship Inspection | Today ship hull inspection including the examination of the external coating, detection of defects, and other types of external degradation such as corrosion and marine growth is conducted underwater by means of Remotely Operated Vehicles (ROVs). The inspection process consists of a manual video analysis which is a time-consuming and labor-intensive process. To address this, we propose an automatic video analysis system using deep learning and computer vision to improve upon existing methods that only consider spatial information on individual frames in underwater ship hull video inspection. By exploring the benefits of adding temporal information and analyzing frame-based classifiers, we propose a multi-label video classification model that exploits the self-attention mechanism of transformers to capture spatiotemporal attention in consecutive video frames. Our proposed method has demonstrated promising results and can serve as a benchmark for future research and development in underwater video inspection applications. | ['Martin Ludvigsen', 'Brian Elvesæter', 'Maryna Waszak', 'Ahmed Mohammed', 'Md Abulkalam Azad'] | 2023-05-27 | null | null | null | null | ['video-classification'] | ['computer-vision'] | [ 1.89930394e-01 -3.88792902e-01 6.38927996e-01 -3.29818219e-01
-6.44329548e-01 -4.17851120e-01 -4.32487056e-02 1.80615723e-01
-6.15849376e-01 2.77646095e-01 -5.34296259e-02 -2.02998862e-01
-6.81053698e-02 -6.86323285e-01 -7.05091834e-01 -1.06105781e+00
-2.38531560e-01 -1.18087828e-01 5.86267114e-01 -2.92661369e-01
4.02192265e-01 4.21270430e-01 -1.73663366e+00 3.92039716e-01
4.98864770e-01 9.96474504e-01 5.09933591e-01 7.44751513e-01
2.93901801e-01 1.00662827e+00 -4.76502091e-01 -9.06910598e-02
1.19571507e-01 -3.69583756e-01 -6.58642888e-01 2.66440719e-01
2.42387176e-01 -6.37642205e-01 -3.49400669e-01 1.21178722e+00
3.87396097e-01 4.82683808e-01 3.91087472e-01 -7.59181440e-01
-5.27347088e-01 1.43950850e-01 -5.31606078e-01 7.83826649e-01
2.37537384e-01 -6.80462494e-02 9.18171823e-01 -6.92407727e-01
2.48096377e-01 9.74042714e-01 7.10345149e-01 3.91231447e-01
-3.81571054e-01 -3.15837443e-01 3.19911510e-01 7.55699515e-01
-8.68995011e-01 -2.93874145e-01 8.01753104e-01 -4.67511863e-01
6.72417581e-01 5.97863533e-02 8.44222844e-01 4.75748003e-01
4.85431492e-01 8.78350377e-01 6.89687729e-01 -3.98472935e-01
2.70481735e-01 -4.23969120e-01 -1.42222615e-02 7.92433619e-01
-6.55735061e-02 -1.03855021e-01 -1.51291072e-01 2.63159484e-01
5.20533621e-01 3.33326876e-01 -5.18318474e-01 2.09229872e-01
-5.78614712e-01 7.23095417e-01 3.02248687e-01 1.62036464e-01
-3.29971492e-01 2.20257446e-01 6.55541658e-01 3.85972708e-01
6.03857398e-01 2.80357420e-01 -5.00547528e-01 -5.35716042e-02
-8.04873943e-01 -3.27371567e-01 3.55439961e-01 5.32772183e-01
7.75321066e-01 4.17656392e-01 2.64982074e-01 8.07442605e-01
5.61808884e-01 3.63805324e-01 5.37316799e-01 -1.05300117e+00
5.18128313e-02 2.94626087e-01 -1.62455775e-02 -8.45327735e-01
-6.13999963e-01 -4.50063385e-02 -2.63505757e-01 3.53790939e-01
8.77226219e-02 -3.21588516e-01 -7.97280073e-01 9.01342869e-01
2.17386797e-01 4.53493118e-01 2.93925911e-01 1.13477397e+00
9.95784938e-01 7.47483552e-01 7.82319978e-02 -2.88488477e-01
1.46333241e+00 -1.16035116e+00 -7.81231999e-01 -5.95194846e-02
6.47574842e-01 -4.43614155e-01 7.39957392e-01 4.90981162e-01
-1.05017567e+00 -4.76608455e-01 -1.14876354e+00 4.91899885e-02
-1.36119410e-01 2.14949965e-01 4.77096826e-01 3.81841660e-01
-9.24270511e-01 5.31811774e-01 -1.19285476e+00 -1.93966590e-02
4.04867798e-01 2.79438019e-01 -5.53268611e-01 -5.40965199e-01
-9.89488900e-01 5.55004776e-01 -1.66436017e-01 6.84619188e-01
-1.53387630e+00 -2.68363535e-01 -1.34238648e+00 2.38636620e-02
2.79044956e-01 1.02509342e-01 1.19933558e+00 -9.84881341e-01
-1.39831853e+00 5.11200309e-01 -1.22935876e-01 -1.27856508e-01
1.40728697e-01 -4.86170352e-01 -1.89561054e-01 6.45472646e-01
-2.32885871e-02 3.34702544e-02 6.77444518e-01 -1.10280442e+00
-9.11512911e-01 -2.95266002e-01 3.62865120e-01 2.40566418e-01
-5.42848706e-01 3.17101151e-01 -4.83615160e-01 -4.38169718e-01
2.61480600e-01 -9.33271885e-01 -9.03099701e-02 -1.84532311e-02
1.64930090e-01 -1.81162462e-01 1.17158234e+00 -1.00746667e+00
6.62320793e-01 -2.25591922e+00 1.17951192e-01 -3.20049912e-01
-4.94867973e-02 2.99259722e-01 -1.14457257e-01 2.84618169e-01
1.40364140e-01 -1.47190750e-01 -1.72004119e-01 -4.75474030e-01
-5.87817311e-01 2.24666595e-01 1.33881122e-01 1.00469792e+00
1.42880037e-01 4.05618221e-01 -1.19985735e+00 -5.26759088e-01
3.64703536e-01 2.67538011e-01 -4.61563826e-01 2.96121627e-01
8.34399760e-02 5.78965187e-01 -3.37803334e-01 1.19009578e+00
5.43519676e-01 4.35101949e-02 2.00094078e-02 -5.12877226e-01
-3.98418933e-01 -2.37853050e-01 -5.94168842e-01 1.63209224e+00
-4.43720132e-01 1.05734003e+00 2.93642908e-01 -1.37277770e+00
7.45861709e-01 4.41069186e-01 6.40910149e-01 -5.63431680e-01
3.06914091e-01 -6.63592592e-02 -2.52078146e-01 -1.37518978e+00
4.28476155e-01 -3.54308397e-01 2.29864627e-01 8.20193291e-02
-8.53647292e-03 1.90914601e-01 1.90131322e-01 -2.44859666e-01
1.19871223e+00 2.48747230e-01 -2.13038236e-01 -1.72865346e-01
4.41307485e-01 -6.83156354e-03 8.86550605e-01 4.35324252e-01
-2.80198574e-01 5.53351521e-01 6.17832579e-02 -7.46690810e-01
-8.46533477e-01 -6.26778841e-01 -1.18179128e-01 1.07760549e+00
6.00144506e-01 1.63541138e-01 -4.45934713e-01 -4.01722461e-01
-2.28770033e-01 3.18867266e-02 -6.77176833e-01 -1.78072736e-01
-6.32578373e-01 -9.33990121e-01 3.50083441e-01 7.66336143e-01
3.53154898e-01 -1.13847494e+00 -8.42499077e-01 2.81598717e-01
-2.02259377e-01 -1.34181929e+00 -1.22211389e-01 2.27295831e-01
-9.42063272e-01 -1.23183131e+00 -5.67450643e-01 -1.33677912e+00
6.82830215e-01 6.21975243e-01 6.77378416e-01 4.87117320e-01
-4.76295322e-01 4.00953382e-01 -8.80434692e-01 -3.61177891e-01
-9.24434140e-03 -5.49687326e-01 -7.28081260e-03 2.15776280e-01
6.62052259e-02 -2.88268805e-01 -7.62963533e-01 4.22200143e-01
-1.06558478e+00 -4.79966849e-01 4.60877746e-01 9.25392985e-01
2.27994740e-01 5.03976166e-01 4.17479157e-01 -4.07927245e-01
1.04020469e-01 -6.37935817e-01 -7.47043312e-01 2.01402545e-01
1.43082649e-01 -3.92309457e-01 4.60829318e-01 -1.17150679e-01
-1.05891740e+00 -4.68296222e-02 -3.63891214e-01 -6.00350499e-01
-7.85423890e-02 9.18287575e-01 9.81388092e-02 -3.88197124e-01
-1.23702854e-01 2.98370540e-01 2.82825708e-01 -5.33682227e-01
-5.51109791e-01 6.63659513e-01 5.18246710e-01 -1.82855964e-01
3.76537472e-01 9.09286320e-01 -2.63957940e-02 -1.22470701e+00
-8.98780286e-01 -7.33053446e-01 -5.61549306e-01 -9.22883511e-01
1.38503397e+00 -1.25120938e+00 -1.04381239e+00 7.55795002e-01
-1.03676200e+00 -2.94829726e-01 3.15565199e-01 7.26509929e-01
-3.74532957e-03 9.80663478e-01 -1.16675925e+00 -1.05253661e+00
-2.28370935e-01 -1.46934438e+00 1.12246346e+00 3.90595287e-01
6.26892924e-01 -1.09559882e+00 -8.60477313e-02 6.88788295e-01
1.85175627e-01 1.59138069e-01 3.87035787e-01 -2.66369164e-01
-6.30143762e-01 -4.24281180e-01 -4.18459736e-02 6.85839057e-01
2.77356952e-01 2.16648236e-01 -8.07625651e-01 -4.73306388e-01
2.43435174e-01 -3.87007296e-01 9.40771222e-01 6.59931660e-01
9.96838093e-01 1.97015584e-01 -1.36404052e-01 6.54953837e-01
1.49925232e+00 6.21664345e-01 7.38004565e-01 6.34998441e-01
5.36234617e-01 8.91566515e-01 9.30277169e-01 6.80699468e-01
5.83266914e-01 4.10548657e-01 1.07408798e+00 -2.72077471e-01
3.76056135e-01 5.58422267e-01 5.39315820e-01 9.66156185e-01
-5.33466399e-01 -4.33992475e-01 -6.57208741e-01 9.10861492e-01
-1.54912460e+00 -1.21743667e+00 -2.40128919e-01 1.69741130e+00
2.06769988e-01 -2.36202583e-01 -4.28667247e-01 1.71148121e-01
7.52004743e-01 -1.26643851e-01 -3.52803677e-01 -6.20988682e-02
-1.60024464e-01 -8.84136111e-02 4.17218477e-01 1.88020021e-01
-1.43943596e+00 5.19502342e-01 6.38317442e+00 1.45462379e-01
-1.04970682e+00 9.91716385e-02 2.95422196e-01 -2.69295350e-02
3.50341052e-02 -8.07777494e-02 -5.91272354e-01 4.90555167e-01
6.48895741e-01 6.95741415e-01 -2.77461242e-02 7.61101067e-01
4.54148471e-01 -2.00273469e-01 -5.60559869e-01 6.10355198e-01
4.64196354e-01 -1.26180053e+00 -2.95682818e-01 -6.32313415e-02
5.87148011e-01 2.17095494e-01 -3.34612161e-01 6.65982545e-04
-3.98070216e-02 -3.96384299e-01 7.50328720e-01 5.18507183e-01
3.31476241e-01 -3.47407132e-01 1.26283848e+00 -1.99967831e-01
-1.38983214e+00 -6.43440902e-01 -3.97058398e-01 -2.89583921e-01
5.82567096e-01 2.05541998e-01 -1.69835672e-01 4.45692956e-01
1.21385968e+00 1.15160537e+00 -1.90417349e-01 1.44058967e+00
-6.60681501e-02 5.95936656e-01 7.02600107e-02 1.45914555e-01
4.30586666e-01 -3.34729910e-01 4.57136244e-01 9.67489958e-01
5.03226757e-01 2.86596149e-01 3.42256218e-01 -7.19299465e-02
1.60425797e-01 -1.50949433e-01 -4.51119274e-01 2.96690881e-01
-2.44038962e-02 1.22933018e+00 -8.82310987e-01 -1.67503715e-01
-1.02561212e+00 8.11156034e-01 -7.99473748e-02 1.88347787e-01
-9.05378997e-01 -5.55837691e-01 7.01594651e-01 -2.04462603e-01
7.20430076e-01 -5.38855612e-01 2.87720144e-01 -1.13840890e+00
-5.05804233e-02 -3.72627199e-01 4.95386541e-01 -9.72475886e-01
-9.33358967e-01 6.84838653e-01 -2.19332471e-01 -1.69224513e+00
5.13803780e-01 -7.47470498e-01 -8.43224823e-01 1.32871225e-01
-2.17005491e+00 -8.99379730e-01 -8.43798876e-01 2.45559394e-01
1.15908122e+00 -1.87759444e-01 2.99682021e-01 6.26303196e-01
-7.66616464e-01 2.51861244e-01 2.91526854e-01 3.82008493e-01
3.44863743e-01 -9.51117754e-01 -2.17978954e-01 9.95630920e-01
-3.55862647e-01 9.64185596e-02 7.65659690e-01 -5.73727667e-01
-1.58296478e+00 -1.03009164e+00 1.02552086e-01 7.64023885e-02
7.32141614e-01 1.80789545e-01 -1.03101099e+00 6.77302420e-01
3.19032311e-01 2.04005286e-01 9.21330035e-01 -4.57937598e-01
3.55114698e-01 -6.07037768e-02 -7.74363458e-01 5.36616482e-02
5.58065295e-01 -2.55195469e-01 -4.21478897e-01 5.43522954e-01
5.04171908e-01 -2.95106202e-01 -1.09477782e+00 6.06795490e-01
5.33939004e-01 -6.38058901e-01 8.02923203e-01 -2.57281303e-01
5.13792515e-01 -6.64953589e-01 -2.25393265e-01 -9.78289843e-01
-1.40275821e-01 -1.63731486e-01 4.74847794e-01 8.69130790e-01
9.34310108e-02 -2.17463464e-01 5.87086976e-01 3.91646236e-01
-1.16021407e+00 -3.84815156e-01 -9.68899846e-01 -6.08330309e-01
-5.14939606e-01 -3.22492182e-01 -5.46697043e-02 7.60227323e-01
-2.36499868e-03 -1.88214079e-01 -6.60320997e-01 8.43885124e-01
6.75858557e-01 4.15408723e-02 1.90356046e-01 -1.31504035e+00
-1.09672435e-01 5.49878776e-02 -7.72110283e-01 -9.12514091e-01
2.57415920e-01 -3.06463778e-01 6.31775558e-01 -1.57465434e+00
-6.42029643e-02 -1.09120493e-03 -4.53015000e-01 4.72333461e-01
-6.92318827e-02 8.39207232e-01 -1.75099805e-01 2.89986193e-01
-8.87617648e-01 7.06178069e-01 1.15647578e+00 -3.98145050e-01
1.53074563e-01 -6.96261600e-02 -1.08056478e-01 6.30690753e-01
4.82727766e-01 -3.73196036e-01 -5.89601435e-02 -8.70682299e-01
1.67080671e-01 3.38925302e-01 1.99245617e-01 -1.15684819e+00
4.51077610e-01 -1.48121659e-02 1.80220716e-02 -3.72578412e-01
4.95910048e-01 -9.19197381e-01 -1.69657066e-01 7.08001018e-01
1.70155033e-01 3.61218959e-01 3.00290406e-01 9.12864327e-01
-7.27218509e-01 -7.16286719e-01 8.27990234e-01 -4.80445862e-01
-1.22206092e+00 9.08415988e-02 -1.17558801e+00 -4.63585705e-01
1.29813325e+00 -2.97768503e-01 -4.12225723e-01 -1.35225192e-01
-7.14727998e-01 4.47621495e-01 3.36446375e-01 2.88732618e-01
1.09446847e+00 -6.39744818e-01 -7.21255004e-01 2.53941447e-01
6.26873970e-02 -6.99404180e-02 8.14520895e-01 1.04464102e+00
-1.08997214e+00 -1.26729921e-01 -2.38710478e-01 -6.94924474e-01
-1.46822751e+00 2.59242386e-01 4.34833527e-01 2.43926838e-01
-9.19041693e-01 9.24623013e-01 8.67298245e-02 2.67362237e-01
6.82350025e-02 -3.73591900e-01 -7.92028904e-01 2.26203084e-01
7.16024578e-01 4.36354786e-01 1.06395401e-01 -6.12597823e-01
-3.05642843e-01 8.38836312e-01 5.58820330e-02 3.88808936e-01
1.74589813e+00 -4.45560873e-01 -3.74317348e-01 3.03895056e-01
1.18482161e+00 -4.38032568e-01 -1.72233474e+00 1.68599844e-01
-1.60140753e-01 -4.40282732e-01 4.44801241e-01 -9.53369513e-02
-1.59182942e+00 8.72871280e-01 5.15442967e-01 3.10038507e-01
1.26735067e+00 8.07887986e-02 8.98043692e-01 4.45563078e-01
3.15757722e-01 -1.16337705e+00 3.79102260e-01 3.97321016e-01
4.76934344e-01 -1.56824744e+00 9.44001824e-02 -5.01565151e-02
-6.95770621e-01 1.45412660e+00 6.76630020e-01 -1.61045879e-01
7.70255923e-01 3.41475219e-01 4.48423654e-01 -4.83408242e-01
-6.95653677e-01 -1.57823205e-01 -4.94038276e-02 3.75151545e-01
1.94543242e-01 -6.04043603e-01 -4.85427380e-02 2.05567911e-01
7.27993906e-01 -2.78002590e-01 9.41326141e-01 1.39646244e+00
-7.04901814e-01 -5.95826924e-01 -2.29038328e-01 3.69843513e-01
-8.38732958e-01 -8.58260393e-02 6.37574315e-01 4.38343048e-01
-6.13194481e-02 1.22103357e+00 2.07355842e-01 -4.36058670e-01
2.35053629e-01 -3.56896281e-01 1.46257058e-01 -4.77859825e-01
-4.04778063e-01 7.64213204e-02 -2.89821904e-03 -3.00814688e-01
-1.12147832e+00 -1.02330828e+00 -1.24994481e+00 4.69119668e-01
-6.29762650e-01 4.29977864e-01 8.94679666e-01 1.20427358e+00
-9.77122635e-02 9.88690853e-01 8.83269429e-01 -1.38728404e+00
1.15434602e-01 -9.95399594e-01 -8.39248955e-01 4.12689209e-01
8.03538918e-01 -1.06473100e+00 -7.52368391e-01 4.91895497e-01] | [10.676136016845703, -3.50374698638916] |
614e0352-2f08-483c-b7d4-976af2ffdf2b | learning-a-representation-for-cover-song | 1911.00334 | null | https://arxiv.org/abs/1911.00334v1 | https://arxiv.org/pdf/1911.00334v1.pdf | Learning a Representation for Cover Song Identification Using Convolutional Neural Network | Cover song identification represents a challenging task in the field of Music Information Retrieval (MIR) due to complex musical variations between query tracks and cover versions. Previous works typically utilize hand-crafted features and alignment algorithms for the task. More recently, further breakthroughs are achieved employing neural network approaches. In this paper, we propose a novel Convolutional Neural Network (CNN) architecture based on the characteristics of the cover song task. We first train the network through classification strategies; the network is then used to extract music representation for cover song identification. A scheme is designed to train robust models against tempo changes. Experimental results show that our approach outperforms state-of-the-art methods on all public datasets, improving the performance especially on the large dataset. | ['Xiaoou Chen', 'Xiaoshuo Xu', 'Zhesong Yu', 'Deshun Yang'] | 2019-11-01 | learning-a-representation-for-cover-song-1 | null | null | arxiv-2019-11 | ['cover-song-identification'] | ['music'] | [ 4.87583369e-01 -7.02670515e-01 -3.79326791e-01 7.23320991e-02
-1.01503932e+00 -6.71615422e-01 2.65232652e-01 -2.93781370e-01
-3.63749325e-01 3.13747495e-01 1.09927215e-01 2.20836535e-01
-3.02874863e-01 -6.14316583e-01 -6.39557838e-01 -5.50515890e-01
-9.08079222e-02 2.51006275e-01 -2.60396630e-01 -3.06987166e-01
3.86260927e-01 2.78234571e-01 -1.71598232e+00 5.15215933e-01
3.59433323e-01 1.29956877e+00 2.36988012e-02 5.05273461e-01
-7.95547068e-02 6.88350439e-01 -7.79235303e-01 -2.33177841e-01
4.78256881e-01 -5.66604435e-01 -7.04627156e-01 -1.85502529e-01
4.34339404e-01 -2.30438802e-02 -5.39359689e-01 8.63557220e-01
8.54780257e-01 6.74818233e-02 4.24189299e-01 -8.59615862e-01
-2.63148546e-01 1.18997765e+00 -4.33012396e-01 2.32127428e-01
1.65591806e-01 -1.24017119e-01 1.40653491e+00 -7.70937324e-01
3.17633569e-01 7.64415264e-01 9.68188167e-01 3.36268008e-01
-1.01448047e+00 -1.10361373e+00 -1.49206728e-01 6.68060005e-01
-1.81251144e+00 -3.83486748e-01 1.19974339e+00 -4.06122088e-01
6.60001814e-01 5.63705623e-01 7.32176840e-01 1.11623538e+00
-3.01411688e-01 1.04673457e+00 5.69179237e-01 -3.87343198e-01
-1.34641826e-01 -3.26794952e-01 4.70063603e-03 7.92898685e-02
-9.86630917e-02 1.88220050e-02 -8.58081281e-01 -1.34824449e-02
6.29601419e-01 -1.61309883e-01 -2.64416873e-01 3.38238478e-02
-1.37699914e+00 5.91891587e-01 4.17897940e-01 7.37748086e-01
-3.33707422e-01 2.99771219e-01 7.17299581e-01 6.02274299e-01
3.62401217e-01 6.89011514e-01 -2.37386927e-01 -2.69379914e-01
-1.45552814e+00 5.50699592e-01 6.87834918e-01 6.62958682e-01
2.48286441e-01 2.61253208e-01 -3.98370504e-01 9.26555276e-01
-1.32322341e-01 1.97253272e-01 6.80921614e-01 -5.34072340e-01
4.13310409e-01 3.67245674e-01 -1.39490902e-01 -1.12652647e+00
-3.86215001e-01 -1.30213261e+00 -1.00294697e+00 -2.77442157e-01
3.58891904e-01 2.89206237e-01 -4.96124178e-01 1.51419616e+00
-1.14685133e-01 3.92356724e-01 -2.13876754e-01 8.52925956e-01
8.94433498e-01 3.28857154e-01 -5.15422344e-01 2.48537380e-02
1.08328581e+00 -1.02977061e+00 -7.09200740e-01 1.33572116e-01
2.28061035e-01 -9.02500212e-01 9.42413867e-01 7.33313084e-01
-8.11262250e-01 -9.64757085e-01 -1.41693437e+00 2.37647146e-01
-1.97997048e-01 5.98518372e-01 3.96065533e-01 5.37600756e-01
-5.87000608e-01 9.98421550e-01 -3.55718851e-01 1.35586441e-01
4.65356857e-01 5.57034552e-01 1.89746786e-02 3.73949558e-01
-1.29869342e+00 2.64729381e-01 7.79316843e-01 3.17814082e-01
-8.26717079e-01 -4.64071035e-01 -3.71928513e-01 2.07700402e-01
2.84019321e-01 -3.92751098e-01 1.27425480e+00 -1.17895854e+00
-1.52712286e+00 7.45331645e-01 3.92520398e-01 -6.96370900e-01
5.43173909e-01 -3.50280344e-01 -5.45421064e-01 -1.77823693e-01
-1.09500058e-01 2.30402127e-01 1.10199356e+00 -8.57317984e-01
-5.05238652e-01 -9.96284261e-02 -1.21548742e-01 7.52302930e-02
-6.94104135e-01 1.87945768e-01 -7.11654365e-01 -1.16384912e+00
1.48074999e-01 -1.15814924e+00 7.72516280e-02 -5.82072318e-01
-7.34457374e-01 -1.98913097e-01 5.17470181e-01 -6.01235449e-01
1.76429725e+00 -2.37812304e+00 3.46082628e-01 2.78294891e-01
-1.41901672e-01 4.03113425e-01 -2.53515512e-01 6.15161777e-01
-1.23286337e-01 -1.74005210e-01 -4.39290702e-02 -2.19169512e-01
2.21542180e-01 -3.24623376e-01 -5.56133330e-01 3.30757678e-01
-1.00679152e-01 7.92554319e-01 -4.81352091e-01 -1.16694629e-01
-1.22062914e-01 4.29909766e-01 -4.09216881e-01 1.96608186e-01
-1.79867223e-01 6.90508425e-01 -2.85506397e-01 7.10394561e-01
3.45179796e-01 -1.23158433e-01 1.84244215e-01 -2.25263745e-01
4.15226184e-02 5.06106853e-01 -1.45663381e+00 2.16503453e+00
-2.77968943e-01 7.96488404e-01 -2.68012941e-01 -9.06728625e-01
1.09592414e+00 4.91721690e-01 6.68250203e-01 -4.59954023e-01
3.57717425e-01 5.08345366e-01 1.45908892e-01 -2.40542650e-01
6.09386206e-01 1.22170158e-01 -2.22402036e-01 5.53283215e-01
-1.18836261e-01 3.03733528e-01 1.00239754e-01 -4.23386574e-01
8.85663450e-01 1.57700852e-01 1.46986663e-01 2.05412924e-01
7.50494957e-01 -1.23006500e-01 6.05697513e-01 8.85092854e-01
1.51658326e-01 7.46728420e-01 -2.84074023e-02 -7.41789758e-01
-8.22970092e-01 -5.53635240e-01 -2.79170513e-01 1.33930182e+00
-2.55651195e-02 -6.37264073e-01 -6.69600308e-01 -4.11655605e-01
-1.65344514e-02 1.06111199e-01 -4.83773589e-01 -1.20899133e-01
-8.62919629e-01 -5.35822511e-01 1.13987279e+00 6.05214357e-01
5.73060691e-01 -1.45635116e+00 -4.47605640e-01 4.68016058e-01
-2.91888088e-01 -8.66679490e-01 -5.47616601e-01 -5.44879027e-02
-6.06162727e-01 -8.97024155e-01 -7.68286109e-01 -7.96172023e-01
-1.90163568e-01 2.52079964e-01 1.10919154e+00 1.67994410e-01
-3.50826859e-01 -2.51523200e-02 -4.85050768e-01 -6.42816544e-01
-2.01625213e-01 7.98637688e-01 1.57955900e-01 3.74840677e-01
4.09754992e-01 -8.05770755e-01 -5.64321756e-01 2.21425727e-01
-1.01828122e+00 -9.77334678e-02 6.68042183e-01 8.80992353e-01
7.50267267e-01 2.02767506e-01 7.63972759e-01 -6.48700178e-01
6.47743940e-01 -1.67034224e-01 -3.89230818e-01 4.51038443e-02
-3.72196913e-01 -1.67264506e-01 5.67658901e-01 -7.28004158e-01
-2.46271491e-01 1.34028912e-01 -1.26362637e-01 -7.11533964e-01
-4.92716581e-02 7.28461325e-01 -1.49054304e-01 -2.15798244e-01
5.14246702e-01 3.67262363e-01 -4.35046285e-01 -1.05184078e+00
1.23067096e-01 1.10799980e+00 7.97836483e-01 -4.50238198e-01
9.84291375e-01 3.16793263e-01 -1.42015383e-01 -5.21545589e-01
-1.16209698e+00 -5.75699389e-01 -7.79447436e-01 -4.29302126e-01
4.45561022e-01 -1.02774370e+00 -7.07028687e-01 4.62077498e-01
-9.08875942e-01 1.07555620e-01 -2.25440249e-01 5.87888718e-01
-6.62046313e-01 1.58375904e-01 -5.73795855e-01 -7.78603554e-01
-8.01013470e-01 -8.99335921e-01 1.10951412e+00 4.58984450e-02
-3.60623986e-01 -4.23993349e-01 4.24773008e-01 3.40353280e-01
5.15797615e-01 3.31928462e-01 6.02325976e-01 -8.54388416e-01
-6.54976428e-01 -5.36967576e-01 1.13855861e-01 2.91807622e-01
2.08009854e-02 -3.29866141e-01 -1.39328003e+00 -4.60145861e-01
-1.80090249e-01 -3.04531991e-01 1.25389111e+00 1.57152861e-01
1.59566402e+00 -1.77290246e-01 -8.98906291e-02 9.22079444e-01
1.22205591e+00 8.45149811e-03 5.04253626e-01 8.07871640e-01
7.93772578e-01 2.39819691e-01 5.73604047e-01 4.95943040e-01
-1.11149371e-01 1.12541246e+00 3.47090453e-01 7.62282684e-02
-1.01533987e-01 -2.51627922e-01 3.07010770e-01 1.11236262e+00
-2.04823732e-01 3.08210682e-02 -7.62444317e-01 5.00669718e-01
-1.93724811e+00 -1.10443425e+00 2.31460959e-01 2.18457961e+00
8.81522238e-01 3.42196673e-02 3.75916094e-01 8.80380332e-01
6.36259854e-01 3.21046978e-01 -6.06331468e-01 7.16941804e-02
-2.31473520e-01 4.18489218e-01 2.19655484e-01 -2.24467158e-01
-1.64437318e+00 8.19984853e-01 6.15582800e+00 1.31212974e+00
-1.27244878e+00 -4.31798473e-02 1.40483066e-01 -3.98501754e-01
1.28601238e-01 -5.01191497e-01 -5.55581748e-01 2.86365300e-01
6.63816094e-01 7.41586685e-02 7.16255486e-01 5.59129238e-01
-1.76209822e-01 6.99883342e-01 -1.04107082e+00 1.39488626e+00
1.96677461e-01 -1.35734475e+00 4.73804809e-02 -2.64440142e-02
7.25649118e-01 1.35616392e-01 2.56478935e-01 3.70403379e-01
-4.24299836e-01 -1.15074790e+00 9.24301863e-01 5.86767077e-01
9.72064793e-01 -1.07402730e+00 8.44319165e-01 2.86283582e-01
-1.49135828e+00 -1.14164636e-01 -2.55489200e-01 -1.03654005e-01
-2.78567970e-01 2.35469177e-01 -5.97933173e-01 8.28242719e-01
7.38591135e-01 1.07432866e+00 -5.03837585e-01 1.36778271e+00
4.79661971e-02 7.45772481e-01 -9.33514312e-02 1.28614426e-01
3.91537808e-02 -5.32846432e-04 9.02802706e-01 1.24599361e+00
2.52297461e-01 -6.65912986e-01 2.45117515e-01 7.69930303e-01
-3.97501767e-01 4.12450165e-01 -4.35789883e-01 -3.99351716e-01
2.42154941e-01 1.28292584e+00 -5.33749521e-01 1.25803845e-02
-4.76091541e-02 8.41447115e-01 9.19949114e-02 1.04308717e-01
-6.68849885e-01 -5.19910872e-01 5.80195606e-01 -1.57538146e-01
6.79674685e-01 -1.06815621e-01 -6.67580143e-02 -1.16492856e+00
1.65848434e-01 -1.25554752e+00 4.71676439e-01 -3.24954718e-01
-1.28082287e+00 7.36266911e-01 -4.83076215e-01 -1.92971003e+00
-3.92854512e-01 -3.81433219e-01 -5.84190488e-01 6.23521149e-01
-1.48670876e+00 -1.23784959e+00 -1.46107048e-01 5.16752124e-01
5.44446588e-01 -7.11258352e-01 1.13668895e+00 6.56094253e-01
-5.04210234e-01 9.68843579e-01 4.86080110e-01 6.72888517e-01
7.00900853e-01 -9.65672612e-01 3.27451915e-01 4.58109081e-01
1.02264023e+00 3.57724667e-01 4.04817402e-01 -1.69199824e-01
-1.44845295e+00 -1.11746359e+00 6.30547941e-01 -2.06787601e-01
5.22053957e-01 -3.20055962e-01 -6.86965644e-01 3.94148350e-01
1.37457043e-01 -3.83079171e-01 1.19766319e+00 3.62601072e-01
-6.77860141e-01 -3.47321659e-01 -4.51372981e-01 2.51349479e-01
9.40965772e-01 -8.44060719e-01 -3.53763521e-01 1.08025566e-01
5.10091841e-01 -5.42942464e-01 -8.93686235e-01 5.73438644e-01
1.17329597e+00 -6.24274015e-01 1.04929471e+00 -6.89249516e-01
2.19053879e-01 -5.67831039e-01 -2.53748715e-01 -9.76313055e-01
-2.60780811e-01 -7.68921256e-01 -3.02779406e-01 1.16065109e+00
2.32434928e-01 1.32359937e-01 6.73063278e-01 -5.03539622e-01
1.26444265e-01 -6.40380383e-01 -9.16920662e-01 -8.69621158e-01
-1.95937037e-01 -7.90053070e-01 8.23938310e-01 9.83984172e-01
-1.98899537e-01 4.52327460e-01 -8.32787395e-01 -1.53244406e-01
4.19480592e-01 8.16022158e-01 1.02574301e+00 -1.47084939e+00
-7.84218967e-01 -7.53713369e-01 -7.35614896e-01 -7.45567203e-01
1.56290233e-01 -1.10550427e+00 -1.55557275e-01 -8.79696250e-01
2.71056116e-01 -9.49807391e-02 -1.04060674e+00 4.39927787e-01
-9.88494605e-03 8.97142529e-01 4.73691255e-01 6.09197438e-01
-8.75323892e-01 4.80259210e-01 9.62372839e-01 -5.61024845e-01
-3.44284475e-01 4.59094197e-01 -5.89762092e-01 5.63576102e-01
1.07936811e+00 -5.45278609e-01 2.23444216e-02 -3.86312246e-01
5.31195998e-01 -1.98918819e-01 2.40254879e-01 -1.59469950e+00
6.83245435e-02 3.31656277e-01 4.20779616e-01 -1.02577186e+00
3.28265846e-01 -6.62483811e-01 3.80040079e-01 3.10899556e-01
-6.90194070e-01 -2.38816813e-01 2.71922886e-01 6.57403886e-01
-5.69046378e-01 -1.91646039e-01 3.45623374e-01 1.39233574e-01
-2.71457881e-01 3.44212472e-01 -7.04697613e-03 -2.34165117e-01
3.33206087e-01 6.69232756e-02 2.94884235e-01 -4.48568821e-01
-7.46704936e-01 -2.95986325e-01 -2.06100475e-02 6.84495449e-01
3.91439587e-01 -1.74046957e+00 -9.09487784e-01 1.86476663e-01
3.67956072e-01 -3.33657533e-01 1.08854830e-01 7.81258225e-01
-2.54258960e-01 4.89479214e-01 -2.07584351e-01 -5.41860342e-01
-1.46822464e+00 3.49918514e-01 3.99907053e-01 -3.83250594e-01
-6.36142552e-01 7.94500470e-01 -2.24064887e-01 -3.03880811e-01
7.62911499e-01 1.98409706e-03 -5.42391658e-01 2.26822153e-01
8.62276435e-01 2.03922927e-01 1.49405569e-01 -7.77652919e-01
-6.11894093e-02 7.62122989e-01 -2.16797980e-05 2.94031426e-02
1.45944560e+00 2.05749452e-01 -9.15190671e-03 7.12974608e-01
1.09878719e+00 9.82046798e-02 -6.66698813e-01 -6.73659325e-01
1.60940990e-01 -4.55071509e-01 4.64087985e-02 -6.20716512e-01
-1.15220881e+00 8.49218726e-01 7.82898426e-01 2.69881427e-01
1.37412190e+00 -2.61186332e-01 1.00517905e+00 6.45096242e-01
3.60718369e-01 -1.06902170e+00 -8.43786001e-02 5.75656712e-01
9.01120186e-01 -8.23429167e-01 -1.43569559e-01 8.86293501e-02
-1.22038908e-01 1.24229527e+00 1.61454424e-01 -2.87707478e-01
5.54384589e-01 -1.64701026e-02 1.18039802e-01 -1.13105707e-01
-3.41577023e-01 -4.31198567e-01 9.10643339e-01 9.49477404e-02
5.58432817e-01 9.51858610e-02 -2.31806159e-01 1.13816524e+00
-6.38136923e-01 -9.28461552e-03 -3.82911824e-02 6.02474451e-01
-3.33226264e-01 -1.35825598e+00 -4.33325350e-01 3.15335184e-01
-8.59213591e-01 -1.37014896e-01 -8.38539124e-01 5.68021536e-01
2.94893175e-01 7.48870492e-01 -1.04820400e-01 -8.85950327e-01
2.67521858e-01 4.02477644e-02 4.49463397e-01 -4.16962922e-01
-1.18169618e+00 5.09782195e-01 -3.45611759e-02 -4.07855988e-01
-6.43108726e-01 -4.41147447e-01 -7.75451124e-01 5.11449948e-02
-4.94518220e-01 1.77162036e-01 6.47560418e-01 7.80322373e-01
3.12799335e-01 8.39410782e-01 9.03200090e-01 -1.03104985e+00
-6.14338398e-01 -1.26803446e+00 -7.64399350e-01 4.48344111e-01
2.53313839e-01 -5.12522280e-01 3.39590316e-03 -9.40913484e-02] | [15.77473258972168, 5.214552402496338] |
40ce52ac-4b2d-46c4-b570-f4815f4e3c1e | lightdefectnet-a-highly-compact-deep-anti | 2204.11765 | null | https://arxiv.org/abs/2204.11765v1 | https://arxiv.org/pdf/2204.11765v1.pdf | LightDefectNet: A Highly Compact Deep Anti-Aliased Attention Condenser Neural Network Architecture for Light Guide Plate Surface Defect Detection | Light guide plates are essential optical components widely used in a diverse range of applications ranging from medical lighting fixtures to back-lit TV displays. An essential step in the manufacturing of light guide plates is the quality inspection of defects such as scratches, bright/dark spots, and impurities. This is mainly done in industry through manual visual inspection for plate pattern irregularities, which is time-consuming and prone to human error and thus act as a significant barrier to high-throughput production. Advances in deep learning-driven computer vision has led to the exploration of automated visual quality inspection of light guide plates to improve inspection consistency, accuracy, and efficiency. However, given the cost constraints in visual inspection scenarios, the widespread adoption of deep learning-driven computer vision methods for inspecting light guide plates has been greatly limited due to high computational requirements. In this study, we explore the utilization of machine-driven design exploration with computational and "best-practices" constraints as well as L$_1$ paired classification discrepancy loss to create LightDefectNet, a highly compact deep anti-aliased attention condenser neural network architecture tailored specifically for light guide plate surface defect detection in resource-constrained scenarios. Experiments show that LightDetectNet achieves a detection accuracy of $\sim$98.2% on the LGPSDD benchmark while having just 770K parameters ($\sim$33$\times$ and $\sim$6.9$\times$ lower than ResNet-50 and EfficientNet-B0, respectively) and $\sim$93M FLOPs ($\sim$88$\times$ and $\sim$8.4$\times$ lower than ResNet-50 and EfficientNet-B0, respectively) and $\sim$8.8$\times$ faster inference speed than EfficientNet-B0 on an embedded ARM processor. | ['Alexander Wong', 'Mohammad Javad Shafiee', 'Gautam Bathla', 'Mahmoud Famouri', 'Carol Xu'] | 2022-04-25 | null | null | null | null | ['defect-detection'] | ['computer-vision'] | [ 4.04781073e-01 -1.58324063e-01 3.07424754e-01 2.41997223e-02
-4.52146947e-01 -8.98097977e-02 -2.46312559e-01 3.28486226e-02
-2.98875690e-01 2.03498036e-01 -8.55055749e-01 -5.77662349e-01
-3.02181751e-01 -8.32438052e-01 -6.18092537e-01 -7.59095430e-01
2.31881827e-01 2.90943235e-01 2.50390708e-01 -3.35269243e-01
5.71151435e-01 4.84218806e-01 -1.83369172e+00 1.72796085e-01
6.38360143e-01 1.40406621e+00 3.95171285e-01 5.50643027e-01
3.07293952e-01 5.20529389e-01 -4.43324149e-01 -4.37746853e-01
4.41231102e-01 -1.02540933e-01 -4.40972298e-01 1.07864588e-01
4.38450605e-01 -4.93093193e-01 -2.16186211e-01 1.23690045e+00
5.31916976e-01 1.59027085e-01 5.87291360e-01 -9.56443965e-01
-6.64568186e-01 -1.33359537e-01 -9.25195158e-01 3.12142223e-01
5.87487407e-02 6.65652871e-01 1.01786041e+00 -8.97838652e-01
8.16390142e-02 6.90532446e-01 6.87285900e-01 3.73151600e-01
-8.44793260e-01 -7.36751139e-01 -3.73900324e-01 2.32559785e-01
-1.19842017e+00 -3.67990077e-01 6.99236512e-01 -3.72056663e-01
1.68728268e+00 3.46393399e-02 5.85983038e-01 1.63482442e-01
4.24553543e-01 4.93411243e-01 7.79508233e-01 -6.97745144e-01
2.95751423e-01 -5.46592884e-02 1.81045100e-01 1.39859438e+00
4.32676703e-01 2.62355268e-01 -3.41391444e-01 3.71990502e-01
7.89267004e-01 -6.65324777e-02 -1.45089328e-01 2.68804640e-01
-3.33999246e-01 7.73888290e-01 4.70461935e-01 -2.44090278e-02
-2.25558981e-01 2.15471044e-01 1.98664501e-01 1.10329412e-01
3.66060019e-01 9.87393379e-01 -2.59374678e-01 -1.97970927e-01
-7.20889807e-01 4.61208373e-02 4.20502901e-01 9.37670171e-01
9.32183385e-01 5.90505838e-01 2.45967492e-01 1.27439916e+00
2.19048709e-01 8.39570105e-01 6.47826493e-02 -9.16680932e-01
2.89008409e-01 9.38973963e-01 -6.24464452e-03 -1.14907885e+00
-5.82156599e-01 -3.05279285e-01 -1.00968754e+00 6.45879149e-01
1.28715470e-01 -4.25120723e-03 -1.13084388e+00 1.10772908e+00
-1.73156206e-02 -5.08759201e-01 -4.49462533e-01 7.86377430e-01
6.24786079e-01 6.71553850e-01 -5.25160015e-01 -4.89288801e-03
1.51640952e+00 -8.48162889e-01 -2.53932089e-01 -4.96332288e-01
3.75033081e-01 -1.02821088e+00 1.28596950e+00 8.62480283e-01
-1.47956669e+00 -8.02606046e-01 -1.58575535e+00 -2.67712206e-01
-1.83933973e-01 3.61839622e-01 5.04028022e-01 7.12252557e-01
-9.14185286e-01 3.70719314e-01 -8.32883656e-01 1.24660410e-01
5.74510276e-01 8.69794607e-01 1.60025179e-01 -3.88423651e-01
-5.51320195e-01 5.99581003e-01 -2.92084575e-01 1.38503015e-01
-7.57925332e-01 -7.49514043e-01 -7.59206533e-01 1.55103788e-01
4.69577879e-01 -2.40608409e-01 1.06023407e+00 -2.83869267e-01
-1.50511861e+00 7.63055444e-01 -4.79474198e-03 -3.61949861e-01
-2.03917414e-01 -1.47074059e-01 -3.53628337e-01 9.63079706e-02
-2.15524107e-01 4.31651503e-01 7.76056111e-01 -8.92458916e-01
-1.04798210e+00 -4.43941683e-01 1.73947558e-01 -9.56333131e-02
-3.35466385e-01 2.72652488e-02 -6.94595873e-01 -2.27746934e-01
1.99637935e-01 -1.08416760e+00 -1.39097422e-01 2.28230879e-01
-4.89257187e-01 -3.44697416e-01 7.96667933e-01 -4.75510389e-01
8.96658540e-01 -2.06796885e+00 -6.65336192e-01 2.60142267e-01
3.34094763e-01 7.21717060e-01 1.16564013e-01 2.82528694e-03
1.31383523e-01 -2.61689782e-01 -1.65594026e-01 -2.60286212e-01
-1.68933973e-01 -1.52542293e-01 2.34311715e-01 6.19312525e-01
2.82406420e-01 4.43748236e-01 -4.66963202e-01 1.14229612e-01
4.98895228e-01 2.76754886e-01 -7.71606922e-01 2.06850439e-01
-1.84977069e-01 -3.39119554e-01 -1.06568322e-01 1.13039660e+00
6.70216501e-01 -4.69887495e-01 -2.92471051e-01 -4.85751092e-01
-2.12141901e-01 1.34976715e-01 -9.79302585e-01 1.41388285e+00
-7.17365503e-01 9.59969699e-01 3.75663161e-01 -1.05556262e+00
1.25283754e+00 -5.24151251e-02 3.58673602e-01 -1.20209861e+00
4.14933205e-01 2.77440548e-01 5.46155609e-02 -6.31610751e-01
6.13841772e-01 -2.65130967e-01 9.27325860e-02 4.04628813e-01
-2.92835325e-01 -5.90393901e-01 1.77858248e-01 -3.48484844e-01
1.37001288e+00 -2.54064202e-01 -2.51558006e-01 -4.53622401e-01
2.92646557e-01 6.74120933e-02 5.71822703e-01 6.17833257e-01
-9.90671664e-03 5.31911612e-01 8.04929137e-02 -7.52119124e-01
-1.13339341e+00 -9.21365619e-01 -1.35664985e-01 8.25810909e-01
2.68449634e-01 2.54828006e-01 -6.98597431e-01 -1.33970052e-01
-1.16321802e-01 6.31182432e-01 -9.84781981e-03 -3.82541031e-01
-6.35379136e-01 -8.95923018e-01 4.21949297e-01 5.74947417e-01
6.36386812e-01 -9.48980510e-01 -1.17569780e+00 2.09119469e-01
4.22694147e-01 -9.96201336e-01 -6.59483448e-02 3.92480820e-01
-4.36296284e-01 -1.33289278e+00 -4.18430179e-01 -1.08209383e+00
8.37037504e-01 2.09114790e-01 1.12458038e+00 3.52295071e-01
-1.18913388e+00 1.55937076e-01 -2.00121142e-02 -7.72569776e-01
-3.97939146e-01 -1.60680056e-01 2.23789796e-01 -4.31072146e-01
6.40614152e-01 -2.18264058e-01 -9.94127631e-01 3.53383005e-01
-7.32153594e-01 5.53000672e-03 6.00393414e-01 1.07362938e+00
4.95384723e-01 4.48569655e-01 3.76279026e-01 -6.58642352e-01
4.77490395e-01 1.02429256e-01 -1.26301789e+00 -4.99501713e-02
-1.06850898e+00 -2.00458825e-01 9.31199491e-01 -1.26305865e-02
-8.97052705e-01 -2.98209876e-01 -4.50418293e-01 -5.44297636e-01
-6.41918778e-02 2.19374418e-01 1.56579137e-01 -2.62746245e-01
7.82927513e-01 1.10239470e-02 1.17045388e-01 -2.37264663e-01
-3.72796059e-01 6.79182470e-01 5.72538137e-01 -2.02738702e-01
7.00057685e-01 5.22042215e-01 2.66058147e-01 -1.19165468e+00
-3.12709719e-01 -3.16318840e-01 1.98892251e-01 -2.11227611e-01
7.32226014e-01 -6.81591153e-01 -1.23349106e+00 7.66220629e-01
-9.16132927e-01 -3.12662601e-01 -1.11369364e-01 2.74121404e-01
-2.51760960e-01 2.75392890e-01 -8.34510505e-01 -7.71370769e-01
-8.61799479e-01 -1.70062435e+00 8.13460112e-01 3.70216399e-01
-5.05179018e-02 -5.52843451e-01 -5.81389427e-01 7.51834810e-01
3.39670599e-01 -1.37321260e-02 1.42255473e+00 -1.08777128e-01
-7.47609854e-01 -5.22165596e-01 -6.18894398e-01 9.91968930e-01
4.33593750e-01 2.20245048e-02 -9.86909211e-01 -4.59249198e-01
1.28776237e-01 -3.98739606e-01 5.41381180e-01 5.28885365e-01
1.22705698e+00 8.30839649e-02 -3.03130031e-01 5.55281341e-01
1.76024985e+00 9.40436482e-01 6.66990578e-01 5.49882464e-02
5.66001356e-01 4.14457381e-01 5.19300461e-01 5.45311689e-01
-1.52995318e-01 4.26379293e-01 6.96614623e-01 -3.32791388e-01
-8.56960490e-02 3.19948792e-01 -1.26040876e-01 7.93855369e-01
1.27646688e-03 -4.89285082e-01 -1.00133419e+00 6.63520873e-01
-8.59651566e-01 -4.49617833e-01 -8.39317031e-03 2.18665552e+00
6.11038268e-01 5.82316518e-01 -4.04242873e-01 4.10646290e-01
6.20859265e-01 -3.58608633e-01 -7.41105497e-01 -7.88686693e-01
3.40796053e-01 7.21655071e-01 6.51616216e-01 1.58029661e-01
-6.38073742e-01 6.11439824e-01 4.50176573e+00 1.09038973e+00
-1.29094744e+00 -2.28076920e-01 7.83433199e-01 -3.38396966e-01
1.56372622e-01 -5.55082738e-01 -9.84410763e-01 5.34484208e-01
5.32403767e-01 5.61091661e-01 4.53320622e-01 7.91315019e-01
1.52957410e-01 -4.92107898e-01 -1.20585513e+00 1.28295374e+00
7.62307970e-03 -1.62139738e+00 -4.54941362e-01 7.68472627e-02
7.72620857e-01 1.62860990e-01 5.13736844e-01 -1.07623577e-01
3.23029935e-01 -7.64836192e-01 3.61519903e-01 -1.61822781e-01
1.31471801e+00 -8.90427589e-01 7.43437052e-01 7.72458734e-04
-8.11405838e-01 -3.20387274e-01 -5.84443152e-01 -1.42905846e-01
4.22500893e-02 8.31550002e-01 -1.01759839e+00 7.17131943e-02
1.23452985e+00 1.49375632e-01 7.80642927e-02 7.64691234e-01
-2.28222297e-03 6.29520297e-01 -5.42846382e-01 -4.18138683e-01
3.40722442e-01 -8.53553116e-02 3.35694373e-01 9.14101660e-01
3.63321751e-01 1.43469900e-01 -6.81166500e-02 9.52907562e-01
-1.44863293e-01 -4.95520353e-01 -3.91938865e-01 2.68709868e-01
3.03299099e-01 1.05397880e+00 -8.77684653e-01 1.11860007e-01
-4.76513475e-01 6.61522090e-01 -1.91344336e-01 3.08728945e-02
-8.05874884e-01 -9.67686415e-01 9.05892909e-01 4.53399062e-01
2.66349882e-01 -1.63666949e-01 -5.53122222e-01 -3.17028970e-01
-3.52542065e-02 -7.98745573e-01 1.21143973e-02 -8.66999030e-01
-1.02751493e+00 6.05203450e-01 -4.05152231e-01 -1.11772847e+00
1.98767602e-01 -1.20716238e+00 -2.90042639e-01 8.71864200e-01
-1.45753205e+00 -7.03005970e-01 -4.71857131e-01 3.58513564e-01
9.42622423e-01 -4.10258621e-01 5.35063505e-01 3.98460060e-01
-7.99100757e-01 8.23897600e-01 2.74064302e-01 8.52470174e-02
2.42890820e-01 -8.10118020e-01 3.99053842e-01 8.29387248e-01
-1.93899706e-01 4.56403285e-01 6.09722435e-01 -3.25910121e-01
-1.70447779e+00 -8.28502059e-01 1.44684300e-01 1.52495936e-01
2.15187266e-01 -1.87365383e-01 -7.51926541e-01 -1.04402281e-01
2.10030258e-01 -1.20614395e-01 4.27225411e-01 -2.09691346e-01
7.44652152e-02 -5.29718995e-01 -1.44540167e+00 6.20752811e-01
7.49314249e-01 -4.45857495e-01 -7.95008317e-02 3.21834624e-01
5.06203234e-01 -4.81928647e-01 -6.19031549e-01 4.85839158e-01
2.64120489e-01 -1.10588181e+00 8.54358792e-01 2.84784853e-01
6.48537278e-01 -7.14505911e-02 -4.63509299e-02 -9.27584767e-01
-4.43995178e-01 -4.94995475e-01 4.21572030e-01 8.47919881e-01
5.00732601e-01 -6.68965340e-01 1.20308995e+00 8.10330510e-01
-8.73398066e-01 -1.07269347e+00 -1.08170509e+00 -4.79911923e-01
-2.11586401e-01 -6.17626786e-01 2.55031466e-01 4.24165189e-01
-3.06425989e-01 1.64543748e-01 5.21070063e-02 3.33072454e-01
4.67403561e-01 -1.84157804e-01 2.76181757e-01 -1.02623284e+00
-3.74589145e-01 -5.33004820e-01 -4.22703415e-01 -9.79780912e-01
-3.27025205e-01 -2.71785885e-01 3.89257550e-01 -1.36846673e+00
-1.36519387e-01 -8.25494349e-01 -2.53596246e-01 5.97181141e-01
2.32903808e-01 6.46313727e-01 -1.58619553e-01 -1.25560895e-01
-2.15471193e-01 2.29115605e-01 9.68340755e-01 -6.38136923e-01
-6.80053681e-02 8.99744779e-03 -4.49974895e-01 8.55290174e-01
6.47437990e-01 -1.65995181e-01 -5.03953457e-01 -8.82037103e-01
6.93292558e-01 -7.29146600e-02 2.21785188e-01 -1.47531092e+00
4.08859521e-01 1.24381959e-01 2.40428030e-01 -5.27564883e-01
6.62258208e-01 -8.02862704e-01 -4.31086928e-01 6.74367070e-01
1.59132212e-01 1.02279961e-01 5.53321421e-01 2.48654753e-01
-2.49356646e-02 -4.46558684e-01 1.16501403e+00 -2.56321520e-01
-7.31811047e-01 6.33071288e-02 -5.60298204e-01 -1.43817469e-01
8.24356139e-01 -6.72833443e-01 -4.85008836e-01 2.58832723e-02
-1.20078579e-01 -4.22349013e-02 4.00522053e-01 2.11960465e-01
1.08642054e+00 -5.21308362e-01 -3.78978938e-01 6.89650357e-01
-9.69813466e-02 4.28302795e-01 6.19728625e-01 4.04169381e-01
-1.12547779e+00 4.86699373e-01 -7.84708112e-02 -6.86603546e-01
-1.21054852e+00 3.29380810e-01 3.63992810e-01 -7.61221126e-02
-3.82747918e-01 1.44836676e+00 9.81884226e-02 5.88261858e-02
-1.40225098e-01 -5.01383424e-01 2.59894431e-01 -4.66197282e-01
3.33920807e-01 1.08570099e+00 7.04211414e-01 5.72785772e-02
-1.50574908e-01 6.94341362e-01 -2.35864729e-01 3.43738586e-01
1.37204957e+00 1.17968656e-01 -1.49577454e-01 -1.86800718e-01
1.28478193e+00 -3.29878002e-01 -1.26123607e+00 -7.72290155e-02
-2.78661668e-01 -2.59077042e-01 4.65362310e-01 -8.49690199e-01
-1.48312199e+00 1.06440973e+00 9.82332587e-01 1.52853057e-01
1.41077113e+00 -1.95081145e-01 1.03922200e+00 5.69492936e-01
6.07778490e-01 -1.53073037e+00 3.07617158e-01 2.93567479e-01
5.45030653e-01 -1.23956895e+00 1.80780023e-01 -2.82898754e-01
-2.99014375e-02 9.53361928e-01 1.08805215e+00 -2.03077465e-01
5.81709027e-01 6.14293456e-01 -8.69726837e-02 -6.67553902e-01
-5.40598631e-01 3.04924160e-01 1.28626302e-01 4.63778645e-01
1.57173976e-01 -2.69296259e-01 2.24915311e-01 2.22513407e-01
2.76107434e-02 4.50258106e-02 5.10484636e-01 1.10349619e+00
-4.58869964e-01 -5.54864228e-01 -2.57378906e-01 9.23879087e-01
-6.42831266e-01 -2.34283701e-01 4.18133467e-01 6.52186453e-01
4.21731919e-01 1.31329715e+00 5.84653795e-01 -4.82287556e-01
3.83311570e-01 -4.24838006e-01 3.10686082e-01 -5.89150965e-01
-5.07391810e-01 -3.33158672e-02 -3.24879028e-02 -4.08278048e-01
1.53989019e-02 -1.85552984e-01 -1.37639594e+00 -3.05592865e-01
-5.82346678e-01 -2.57260323e-01 8.99079502e-01 7.53816307e-01
3.52923542e-01 6.35481656e-01 6.28048956e-01 -6.09425604e-01
-4.29036319e-01 -6.79853678e-01 -6.58840001e-01 4.55446122e-03
1.82427287e-01 -6.60059333e-01 -2.76607066e-01 -1.32389471e-01] | [7.425858974456787, 1.8738305568695068] |
d0940a9b-256a-4afb-96ca-63e290a7c48e | shapepu-a-new-pu-learning-framework | 2206.02118 | null | https://arxiv.org/abs/2206.02118v1 | https://arxiv.org/pdf/2206.02118v1.pdf | ShapePU: A New PU Learning Framework Regularized by Global Consistency for Scribble Supervised Cardiac Segmentation | Cardiac segmentation is an essential step for the diagnosis of cardiovascular diseases. However, pixel-wise dense labeling is both costly and time-consuming. Scribble, as a form of sparse annotation, is more accessible than full annotations. However, it's particularly challenging to train a segmentation network with weak supervision from scribbles. To tackle this problem, we propose a new scribble-guided method for cardiac segmentation, based on the Positive-Unlabeled (PU) learning framework and global consistency regularization, and termed as ShapePU. To leverage unlabeled pixels via PU learning, we first present an Expectation-Maximization (EM) algorithm to estimate the proportion of each class in the unlabeled pixels. Given the estimated ratios, we then introduce the marginal probability maximization to identify the classes of unlabeled pixels. To exploit shape knowledge, we apply cutout operations to training images, and penalize the inconsistent segmentation results. Evaluated on two open datasets, i.e, ACDC and MSCMRseg, our scribble-supervised ShapePU surpassed the fully supervised approach respectively by 1.4% and 9.8% in average Dice, and outperformed the state-of-the-art weakly supervised and PU learning methods by large margins. Our code is available at https://github.com/BWGZK/ShapePU. | ['Xiahai Zhuang', 'Ke Zhang'] | 2022-06-05 | null | null | null | null | ['cardiac-segmentation'] | ['medical'] | [ 7.48322085e-02 1.58949912e-01 -4.03443456e-01 -3.98991138e-01
-1.10009551e+00 -6.18781924e-01 8.09599757e-02 2.48922154e-01
-3.19264174e-01 8.90147924e-01 -1.05408020e-01 -2.27913484e-01
2.08807498e-01 -4.45683867e-01 -7.24827230e-01 -1.02119172e+00
3.46645802e-01 5.94554961e-01 2.40645081e-01 5.20959198e-01
1.27885770e-02 2.90849358e-01 -7.09069252e-01 -3.58484387e-02
1.07750463e+00 9.63697851e-01 9.56012011e-02 4.07081425e-01
-1.76632613e-01 7.01171875e-01 -5.49559854e-02 -3.23418260e-01
2.12561503e-01 -6.25583768e-01 -6.85112000e-01 4.78457779e-01
4.18669045e-01 -1.97726011e-01 5.03796302e-02 1.13075662e+00
4.85818952e-01 -3.92716005e-02 7.16824889e-01 -8.95178378e-01
-5.03000021e-01 6.22259557e-01 -9.87753272e-01 3.40689480e-01
-1.74598187e-01 1.42037645e-01 8.02701592e-01 -9.71070230e-01
6.43472672e-01 8.08702528e-01 6.91274583e-01 5.12026489e-01
-1.24676263e+00 -5.42572498e-01 1.42171122e-02 -2.01908663e-01
-1.42368340e+00 -2.61236787e-01 7.26716995e-01 -6.32878661e-01
8.95281211e-02 2.26015106e-01 5.48391461e-01 6.21266067e-01
-7.00939745e-02 1.06596315e+00 1.27265406e+00 -3.69593114e-01
2.02506796e-01 1.60191581e-01 3.28801125e-01 1.03249228e+00
2.27231801e-01 -1.05177671e-01 -3.67361754e-02 -2.06005469e-01
1.00279236e+00 8.46057236e-02 -4.29926217e-01 -6.09364212e-01
-1.25252593e+00 6.69092774e-01 4.91407782e-01 4.80701849e-02
-2.67747998e-01 4.05180939e-02 3.17519248e-01 -2.74663955e-01
6.12621605e-01 -2.68061310e-02 -4.63123947e-01 8.55206326e-02
-1.17044783e+00 -1.58664882e-01 7.62152016e-01 7.07995713e-01
7.47274518e-01 -2.14129657e-01 -2.63131440e-01 9.84636605e-01
5.40671647e-01 6.02415740e-01 2.35010952e-01 -1.06514287e+00
2.04306245e-01 5.72935760e-01 -1.26902536e-01 -7.55894840e-01
-3.11924040e-01 -5.49732566e-01 -1.15865850e+00 1.06001846e-01
8.26727211e-01 -4.15779859e-01 -1.21719682e+00 1.43986142e+00
6.08548880e-01 5.27137041e-01 -2.66319573e-01 1.08982468e+00
9.05706882e-01 4.43397731e-01 1.67181566e-01 -4.24136758e-01
1.20216656e+00 -1.17699230e+00 -6.31983578e-01 -4.34019491e-02
7.23927796e-01 -6.70860291e-01 8.57830822e-01 2.07523659e-01
-1.06551683e+00 -2.59132624e-01 -7.85326183e-01 1.39895633e-01
2.11904526e-01 4.58455980e-01 3.54999244e-01 7.10586727e-01
-7.68557131e-01 5.20380914e-01 -1.27189028e+00 -1.00570370e-03
1.06281769e+00 1.41488656e-01 -2.05878660e-01 -1.44182935e-01
-6.72910154e-01 4.84986126e-01 2.93594629e-01 1.00395806e-01
-9.07840610e-01 -8.10820162e-01 -8.97357345e-01 -1.71361282e-01
6.13719761e-01 -5.10074675e-01 8.90367389e-01 -8.19173515e-01
-1.38834286e+00 1.13667452e+00 -1.16200529e-01 -4.84641284e-01
7.87162244e-01 -1.27689958e-01 2.06644192e-01 4.22794163e-01
2.19158694e-01 7.51674414e-01 8.09721410e-01 -1.39093983e+00
-3.10580313e-01 -3.61721098e-01 -4.57614928e-01 5.52978665e-02
1.47892358e-02 -1.43905655e-01 -7.26161480e-01 -7.74861753e-01
3.33573461e-01 -1.07781935e+00 -4.50233102e-01 4.49456334e-01
-5.89978635e-01 -3.30729187e-02 6.31261945e-01 -8.68983209e-01
1.11930251e+00 -2.28163218e+00 -7.06136748e-02 3.45045805e-01
4.83430922e-01 3.54030162e-01 3.44504207e-01 -4.00051802e-01
1.67837262e-01 8.68954957e-02 -1.00455570e+00 -4.85810220e-01
-3.29493761e-01 2.56215513e-01 1.69227853e-01 6.40729070e-01
1.71907358e-02 8.61253142e-01 -1.09241426e+00 -9.56161320e-01
3.46047372e-01 3.81743461e-01 -5.48809290e-01 1.27404243e-01
-1.07855514e-01 9.80911076e-01 -4.29610819e-01 8.64818335e-01
7.03978062e-01 -6.45607054e-01 2.32468799e-01 -1.22526146e-01
3.24848667e-02 -3.39600205e-01 -1.13356161e+00 1.77555478e+00
-1.13855101e-01 2.54681587e-01 1.49972513e-01 -1.12226260e+00
8.65443647e-01 4.44768429e-01 7.54578888e-01 -1.46735251e-01
2.82356083e-01 5.27455926e-01 -1.59201533e-01 -4.48305905e-01
-3.04021209e-01 -2.06888586e-01 2.14246020e-01 3.79956990e-01
-3.76945138e-02 -1.29204720e-01 2.17286885e-01 2.15355486e-01
7.88580298e-01 3.52631390e-01 3.09703678e-01 -3.70828658e-01
5.60587168e-01 -1.19464383e-01 1.00193191e+00 5.77346623e-01
-5.43308377e-01 1.09860623e+00 5.59908867e-01 -2.75432795e-01
-7.18118906e-01 -1.09968448e+00 -3.53762537e-01 5.21122217e-01
1.88004017e-01 -1.81936711e-01 -8.90324295e-01 -1.06388938e+00
-1.45913959e-01 1.38508558e-01 -4.33761120e-01 2.36084998e-01
-5.16725481e-01 -1.04050136e+00 3.38197917e-01 5.77840626e-01
5.63252747e-01 -8.98805976e-01 -5.23935378e-01 9.00822580e-02
-2.86041021e-01 -1.10919487e+00 -7.20589101e-01 7.24784732e-02
-1.05912054e+00 -1.20969427e+00 -1.25719047e+00 -8.09092402e-01
1.06472075e+00 -9.89192054e-02 1.16399217e+00 2.06100658e-01
-3.08336109e-01 9.84242782e-02 -3.06499660e-01 -1.97486132e-01
-1.64950758e-01 -4.19207625e-02 -3.37779552e-01 1.53833017e-01
-1.26297921e-02 -4.71814007e-01 -8.99939835e-01 2.35965654e-01
-6.80693746e-01 1.14358932e-01 7.01515257e-01 1.03392172e+00
1.20306182e+00 -2.71457165e-01 4.76526946e-01 -1.50944161e+00
4.76269908e-02 -3.24045479e-01 -6.07062578e-01 2.58361250e-01
-6.13367081e-01 -1.37086794e-01 2.80129522e-01 -2.34666899e-01
-1.03456330e+00 5.33531666e-01 -1.79426447e-01 -6.46485507e-01
-2.05747887e-01 4.82020885e-01 4.36375439e-02 -5.60441911e-02
4.94934410e-01 -3.22852731e-02 8.78529549e-02 -4.45638090e-01
4.00119662e-01 5.43734312e-01 6.33084059e-01 -6.35749519e-01
4.53159273e-01 7.21054018e-01 -9.40144211e-02 -5.76595485e-01
-1.15408278e+00 -6.39817834e-01 -8.08440804e-01 -1.58597857e-01
1.20636499e+00 -7.96294034e-01 -4.99189124e-02 4.70667899e-01
-8.78207862e-01 -5.62220693e-01 -3.51872623e-01 5.65041125e-01
-4.20917064e-01 7.74434566e-01 -7.52439320e-01 -5.00831008e-01
-5.15591562e-01 -1.24208844e+00 9.18132544e-01 3.66600722e-01
-7.87969157e-02 -1.12984300e+00 7.58128986e-02 6.58374488e-01
3.13360877e-02 4.72985953e-01 6.91637099e-01 -6.49516463e-01
-4.99841481e-01 -1.70421720e-01 -5.56871712e-01 6.79268718e-01
1.75040334e-01 -7.63691217e-02 -7.84135580e-01 -1.38123363e-01
2.38055717e-02 -3.15278113e-01 9.14643407e-01 9.05266404e-01
1.38540900e+00 3.00553143e-02 -3.71283293e-01 7.62565732e-01
1.31069672e+00 -3.13880853e-02 3.69659573e-01 -8.69997293e-02
9.99011338e-01 3.81851435e-01 6.12248063e-01 5.29736996e-01
3.09602827e-01 3.08391213e-01 2.33229175e-01 -4.42522913e-01
-3.68908882e-01 -8.89685228e-02 -8.35710987e-02 9.47829902e-01
-2.60617230e-02 3.52283455e-02 -1.06183648e+00 4.66756225e-01
-1.98341501e+00 -3.63094121e-01 -3.18273306e-01 2.18308330e+00
1.12518227e+00 1.93944737e-01 2.90943682e-02 -7.71023408e-02
9.42312479e-01 1.22957968e-03 -5.28444827e-01 1.88883454e-01
7.77392015e-02 1.94512397e-01 6.00221992e-01 5.67018509e-01
-1.34456158e+00 8.28157306e-01 5.37133980e+00 8.08592975e-01
-9.56457138e-01 4.15009528e-01 1.33143330e+00 1.36061966e-01
-7.53183849e-03 1.70704201e-02 -5.55398464e-01 6.91751838e-01
5.20657122e-01 3.26712638e-01 1.38199423e-02 6.54420674e-01
4.03210372e-01 -2.24014640e-01 -8.54314089e-01 1.08374834e+00
1.16913073e-01 -1.26206040e+00 -2.84615487e-01 -1.38691157e-01
1.05219352e+00 1.26951650e-01 -2.54390895e-01 3.53038423e-02
2.53955901e-01 -7.76753426e-01 2.84214467e-01 4.55004394e-01
7.70787716e-01 -3.71469647e-01 9.31233287e-01 3.34436536e-01
-1.06013393e+00 3.71686280e-01 -2.08030835e-01 3.46582204e-01
4.35895532e-01 1.18701684e+00 -4.52173412e-01 3.52109015e-01
6.18578076e-01 1.07544100e+00 -3.01268131e-01 1.27920902e+00
-5.15387237e-01 9.07072544e-01 -2.75133252e-01 4.94243473e-01
5.96493669e-02 -6.03754163e-01 4.26436216e-01 1.12679398e+00
1.40183285e-01 1.12904526e-01 5.38367748e-01 8.88101459e-01
-3.88883770e-01 2.05776155e-01 -7.68161342e-02 2.26153076e-01
3.11786413e-01 1.52447712e+00 -1.20849979e+00 -6.65727854e-01
-4.06570673e-01 9.30549204e-01 1.03975728e-01 2.84603953e-01
-8.87820601e-01 3.39694545e-02 -2.15175286e-01 7.40175098e-02
1.93742156e-01 5.52724637e-02 -4.96824503e-01 -1.23988378e+00
-1.42734740e-02 -5.93854249e-01 6.00834966e-01 -5.39070725e-01
-1.31579769e+00 3.60685796e-01 -9.28846896e-02 -1.20486510e+00
2.61644274e-01 -3.35168302e-01 -6.50861144e-01 7.21603334e-01
-1.72247946e+00 -9.70754564e-01 -4.68687385e-01 3.23231280e-01
4.94800448e-01 2.10689828e-01 5.29437363e-01 5.45757830e-01
-8.15058768e-01 3.26696366e-01 1.22161388e-01 4.39947248e-01
7.92444170e-01 -1.43586862e+00 -1.91159084e-01 7.28360593e-01
1.57205358e-01 2.94657320e-01 2.83201218e-01 -6.58964515e-01
-7.49651849e-01 -1.18882978e+00 5.37146926e-01 -2.49534801e-01
5.01099288e-01 9.60279331e-02 -8.91340077e-01 6.99688792e-01
-1.13083199e-02 7.81479657e-01 8.04111242e-01 -3.37661177e-01
5.21596670e-02 8.74631405e-02 -1.17446065e+00 3.55032146e-01
7.31204093e-01 -9.96428430e-02 -3.93942416e-01 4.59770173e-01
4.24260974e-01 -6.67579651e-01 -9.82230842e-01 4.73987162e-01
2.94283360e-01 -7.16452718e-01 8.75615954e-01 -1.19001925e-01
4.11523044e-01 -4.15613353e-01 1.47692546e-01 -1.03429174e+00
-1.79414395e-02 -4.17681664e-01 -1.59832656e-01 1.19847810e+00
5.47836244e-01 -5.09231627e-01 1.12812936e+00 6.51272595e-01
-2.58717835e-01 -1.03855479e+00 -8.03576171e-01 -4.71833795e-01
2.82523513e-01 -1.70124456e-01 -1.29598141e-01 9.52784061e-01
-1.82635441e-01 1.53090224e-01 -3.13054174e-01 7.04989657e-02
9.46790516e-01 1.42563328e-01 3.87830466e-01 -1.16718805e+00
-3.48892033e-01 -3.41322988e-01 -1.22684188e-01 -1.07021379e+00
8.35085139e-02 -1.09906030e+00 2.26877898e-01 -1.51394939e+00
6.00011349e-01 -7.79578269e-01 -4.13582206e-01 5.76909959e-01
-5.69170058e-01 6.38890624e-01 5.85192908e-03 4.76011276e-01
-6.98635995e-01 3.45335901e-01 1.57728553e+00 -1.42079845e-01
-2.35006258e-01 2.01706856e-01 -5.59082985e-01 9.45842922e-01
1.02325881e+00 -5.34981906e-01 -3.68495941e-01 -2.82362282e-01
-1.18208513e-01 1.89369321e-01 3.64689767e-01 -8.61350358e-01
4.73266356e-02 1.10160343e-01 2.99273461e-01 -5.26781142e-01
-9.20374691e-02 -7.14578569e-01 -2.03692064e-01 4.97928172e-01
-3.44873786e-01 -4.28786248e-01 -9.13160667e-02 5.58556199e-01
-2.07205728e-01 -3.33357602e-01 1.12203419e+00 -3.48550409e-01
-2.74401546e-01 6.63296223e-01 -3.83632220e-02 5.02058864e-01
1.08208489e+00 4.90418039e-02 -3.19308676e-02 -2.78298222e-02
-1.08364189e+00 4.37254727e-01 2.23462030e-01 -3.22021276e-01
6.25391901e-01 -1.13382757e+00 -7.84513891e-01 -1.70969181e-02
-1.61045536e-01 5.85336387e-01 2.81130910e-01 1.51085019e+00
-7.74273992e-01 6.91073537e-02 8.23642015e-02 -9.77262974e-01
-1.25907350e+00 3.24507833e-01 4.39847559e-01 -3.76122594e-01
-8.59424651e-01 9.26063061e-01 2.76776016e-01 -4.91837263e-01
1.82672173e-01 -3.51775885e-01 -8.33082944e-02 -1.60550147e-01
2.54590839e-01 3.14889252e-01 -5.57904765e-02 -4.99000967e-01
-2.83550322e-01 6.86769485e-01 -3.33227888e-02 9.15827081e-02
1.08002782e+00 -2.54027426e-01 -1.85092360e-01 3.15496325e-01
1.09827721e+00 -1.58653200e-01 -1.63287413e+00 -3.92228305e-01
6.16923384e-02 -3.53570938e-01 1.60594195e-01 -7.40067840e-01
-1.55899549e+00 9.04198587e-01 6.56909525e-01 -1.57489344e-01
9.12097692e-01 6.93406910e-02 9.81070042e-01 -1.34912565e-01
1.17093613e-02 -8.86560142e-01 -7.04100356e-02 3.48474830e-02
3.48331541e-01 -1.68522787e+00 1.59730062e-01 -9.42328930e-01
-7.90418446e-01 8.89936149e-01 4.01403695e-01 -1.56020105e-01
9.14082050e-01 2.33640224e-01 4.42992657e-01 -1.86702132e-01
-2.14608327e-01 -1.52782276e-02 3.19919080e-01 2.91938454e-01
5.73452473e-01 2.45166183e-01 -4.58893508e-01 4.44766045e-01
3.79923612e-01 2.61858255e-01 2.31707036e-01 8.71307850e-01
-4.05379415e-01 -1.02318597e+00 -3.08421165e-01 6.92485452e-01
-8.10105085e-01 -1.36111230e-01 -7.27737322e-04 4.29128647e-01
1.46109745e-01 6.39210641e-01 -1.72880545e-01 2.20261857e-01
-1.54865190e-01 -6.37450069e-02 3.16906393e-01 -8.25581670e-01
-2.46838659e-01 4.95734304e-01 -2.67547339e-01 -3.27399373e-01
-7.07151592e-01 -7.51768231e-01 -1.59911859e+00 2.06911415e-01
-4.30730253e-01 8.33392739e-02 2.49236330e-01 1.02147448e+00
4.68482003e-02 4.98078078e-01 5.06673515e-01 -6.22060716e-01
-4.25496966e-01 -7.63257563e-01 -6.04470193e-01 5.74170709e-01
1.23644300e-01 -4.64966446e-01 -3.74761671e-01 4.49481457e-01] | [14.659512519836426, -2.2042324542999268] |
0346b308-eb87-4b58-a382-e7c21a20ddb4 | submodular-maximization-under-fading-model | 1901.07708 | null | https://arxiv.org/abs/1901.07708v4 | https://arxiv.org/pdf/1901.07708v4.pdf | Cascade Submodular Maximization: Question Selection and Sequencing in Online Personality Quiz | Personality quiz is a powerful tool that enables costumer segmentation by actively asking them questions, and marketers are using it as an effective method of generating leads and increasing e-commerce sales. In this paper, we study the problem of how to select and sequence a group of quiz questions so as to optimize the quality of customer segmentation. We assume that the customer will sequentially scan the list of questions. After reading a question, the customer makes two, possibly correlated, random decisions: 1) she first decides whether to answer this question or not, and then 2) decides whether to continue reading the next question or not. We further assume that the utility of questions that have been answered can be captured by a monotone and submodular function. In general, our problem falls into the category of non-adaptive active learning based customer profiling. Note that under the our model, the probability of a question being answered depends on the location of that question, as well as the set of other questions placed ahead of that question, this makes our problem fundamentally different from existing studies on submodular optimization. We develop a series of question selection and sequencing strategies with provable performance bound. Although we focus on the application of quiz design in this paper, our results apply to a broad range of applications, including assortment optimization with position bias effect. | ['Shaojie Tang', 'Jing Yuan'] | 2019-01-23 | null | null | null | null | ['question-selection'] | ['natural-language-processing'] | [ 2.46343553e-01 2.60144919e-01 -7.71060705e-01 -5.69098890e-01
-1.00043106e+00 -9.03269947e-01 -3.68952721e-01 4.24718022e-01
-4.83224839e-01 7.08851397e-01 -2.85937302e-02 -3.37377340e-01
-7.21556127e-01 -9.29282665e-01 -6.87327087e-01 -6.42787814e-01
5.50941378e-02 1.15213335e+00 1.20813213e-02 -2.22603768e-01
4.61847693e-01 1.77041188e-01 -1.14081717e+00 -8.08603987e-02
1.06178284e+00 1.21336854e+00 4.50518668e-01 7.53693938e-01
-3.77582848e-01 7.40536630e-01 -5.97312331e-01 -7.30074942e-01
5.61963141e-01 -5.14685988e-01 -1.10911143e+00 6.08458936e-01
3.36561515e-03 -3.21746260e-01 2.76889771e-01 1.00226915e+00
3.50159109e-01 3.92943412e-01 3.09649587e-01 -1.36823726e+00
-4.75932777e-01 8.85623693e-01 -6.01142645e-01 3.55246216e-01
1.57850385e-01 -3.14093791e-02 1.76966655e+00 -2.82444865e-01
3.84789795e-01 7.66498983e-01 6.97022751e-02 4.08550084e-01
-1.20111692e+00 -4.85262215e-01 6.99067533e-01 2.81343699e-01
-7.79654086e-01 -1.30029291e-01 8.96355569e-01 -2.40557224e-01
2.16375530e-01 7.54011214e-01 8.33727002e-01 3.17379028e-01
-2.57125068e-02 1.21634805e+00 8.97675216e-01 -3.80216181e-01
7.58780241e-01 5.01852334e-01 5.83418846e-01 2.32350960e-01
2.51537472e-01 -1.83479786e-01 -5.77052593e-01 -2.73367018e-01
2.14233384e-01 3.42965089e-02 8.32212120e-02 -4.84727085e-01
-3.69808078e-01 1.50464821e+00 6.35279566e-02 -1.71154663e-01
-7.07351983e-01 -2.61296511e-01 -1.87619403e-01 4.42968369e-01
2.45155960e-01 9.55533504e-01 -6.34027898e-01 -7.88281485e-02
-8.00929666e-01 5.56732297e-01 1.14041078e+00 9.39081311e-01
7.86850035e-01 -4.44434822e-01 -3.17489445e-01 8.03675652e-01
2.41121024e-01 7.19387531e-02 4.01819162e-02 -1.12171805e+00
6.29418612e-01 7.09327281e-01 3.50220650e-01 -6.90649390e-01
-7.23642185e-02 -4.58963931e-01 -1.62458807e-01 -2.20595613e-01
4.90439564e-01 -5.10873139e-01 -5.03481448e-01 1.50147378e+00
3.28178108e-01 -4.70694065e-01 -5.07648230e-01 9.07458365e-01
3.21416616e-01 5.43823540e-01 -3.56520355e-01 -6.40081108e-01
1.19526184e+00 -9.54262793e-01 -7.47542918e-01 -3.60494822e-01
3.01608533e-01 -6.83424056e-01 9.50987279e-01 6.39466047e-01
-1.22639751e+00 -7.69419074e-02 -6.71511054e-01 1.56894356e-01
-5.91003746e-02 -3.68642479e-01 1.00613141e+00 9.56929564e-01
-7.16825247e-01 8.97940919e-02 -2.61533290e-01 -2.91607203e-03
5.42818666e-01 7.58501768e-01 5.00286698e-01 -8.02021697e-02
-9.32975113e-01 2.34966174e-01 -7.62655912e-03 2.49005351e-02
-6.72837079e-01 -5.65695226e-01 -4.73190933e-01 3.71386647e-01
1.13065970e+00 -5.43167114e-01 1.66133273e+00 -1.03902519e+00
-1.34050834e+00 6.93619549e-01 -4.40491110e-01 -5.11164963e-01
5.05132556e-01 -1.46219060e-01 9.77335572e-02 -2.14540824e-01
3.77329469e-01 5.20502865e-01 4.62635875e-01 -1.10702574e+00
-1.24675608e+00 -8.01625073e-01 7.89464533e-01 5.72775781e-01
-1.23369113e-01 -1.13685578e-01 -5.21991551e-01 -6.64884597e-02
1.79638386e-01 -7.85211980e-01 -6.51008606e-01 -4.98033375e-01
-7.06989467e-01 -4.65380460e-01 5.03877103e-01 -4.11075056e-01
1.20118380e+00 -1.88026571e+00 -9.18532610e-02 3.30787063e-01
2.39721298e-01 -3.44000995e-01 3.88732515e-02 4.47250634e-01
4.27781165e-01 2.84660786e-01 -1.26751333e-01 -1.08579591e-01
2.01479509e-01 2.30495781e-01 -3.72009426e-01 2.10692406e-01
-1.92435488e-01 9.31983113e-01 -6.34515882e-01 -2.24976987e-01
-3.02318960e-01 -6.52844548e-01 -8.44326198e-01 2.84855664e-01
-5.19568622e-01 2.31541708e-01 -6.18071377e-01 7.84164608e-01
7.64468253e-01 -3.09605390e-01 8.25290605e-02 4.93929207e-01
1.07532628e-01 4.06237721e-01 -1.29383588e+00 7.57515788e-01
-2.41864145e-01 3.70893717e-01 3.43227655e-01 -1.24713790e+00
6.72040284e-01 -2.31511757e-01 5.43403089e-01 -8.24484289e-01
1.94501370e-01 1.06813252e-01 2.23754749e-01 -5.79198837e-01
6.81385159e-01 -8.16614926e-02 -2.96272367e-01 4.84121054e-01
-2.44605854e-01 9.36389416e-02 5.78401804e-01 1.47938952e-01
6.50786817e-01 -5.40775001e-01 1.76114351e-01 -2.15332091e-01
1.59974948e-01 3.37727338e-01 6.63545072e-01 1.16989756e+00
-2.22397983e-01 3.01788270e-01 9.94760871e-01 1.79890290e-01
-7.30377734e-01 -1.06520283e+00 2.86586001e-03 1.39731932e+00
5.96612573e-01 2.03727275e-01 -5.82196116e-01 -7.87916005e-01
2.29351282e-01 1.05612850e+00 -5.77557385e-01 2.78224230e-01
-3.95434737e-01 -7.68842757e-01 -5.03947198e-01 2.53082007e-01
4.25956666e-01 -9.15039480e-01 -3.41972888e-01 1.94685534e-01
-9.19843838e-02 -7.05396354e-01 -1.01181901e+00 4.18562889e-01
-8.32066357e-01 -1.19829535e+00 -5.87696910e-01 -9.14797723e-01
4.86087143e-01 3.15700740e-01 1.14015901e+00 -5.48303604e-01
5.94368763e-02 7.00176120e-01 -3.55727196e-01 -5.87996781e-01
2.33897775e-01 4.53466594e-01 -6.20499790e-01 2.68530667e-01
5.60018361e-01 -5.38950115e-02 -8.16781223e-01 4.53936964e-01
-7.47181237e-01 -3.65020663e-01 3.06430995e-01 5.50908148e-01
8.15160036e-01 3.07617366e-01 9.18033421e-01 -1.43360031e+00
9.14955735e-01 -7.29042172e-01 -8.86102974e-01 2.47672439e-01
-6.94831073e-01 -3.31783652e-01 3.48197818e-01 -5.41725099e-01
-8.87081563e-01 5.76842763e-02 -2.05326289e-01 2.49833107e-01
2.06415012e-01 5.47662556e-01 -4.94758189e-01 3.12093407e-01
2.27766767e-01 1.25442492e-02 -2.03927323e-01 -3.62536728e-01
2.62588233e-01 6.53060198e-01 -9.40192863e-02 -2.25915045e-01
5.16400933e-01 1.34760469e-01 -1.85059920e-01 -6.80854559e-01
-9.83069837e-01 -8.37250412e-01 -1.05434500e-01 -1.03256397e-01
8.16459060e-01 -3.43387961e-01 -1.30278885e+00 5.58574013e-02
-6.34211779e-01 -2.54145771e-01 -6.48654819e-01 1.87639624e-01
-5.01055360e-01 -8.17045346e-02 -3.06815535e-01 -1.37790072e+00
3.97238582e-02 -1.13985407e+00 6.01370156e-01 6.31614089e-01
-1.37606695e-01 -8.90166461e-01 -5.43656908e-02 1.13971400e+00
1.43425167e-01 -2.95851111e-01 1.19987690e+00 -1.07751763e+00
-1.35873628e+00 -2.61094540e-01 3.03581029e-01 8.32050964e-02
1.27464635e-02 -6.64161563e-01 -4.72886771e-01 -9.06313509e-02
3.54211003e-01 -1.32041976e-01 6.45222425e-01 8.08833182e-01
1.15501678e+00 -7.67547727e-01 -1.08286321e-01 4.14034009e-01
1.21920300e+00 8.69421840e-01 3.58391613e-01 3.30276877e-01
3.23168188e-01 1.05785930e+00 7.99733758e-01 4.87228721e-01
8.06396604e-01 5.23516357e-01 6.50464773e-01 1.48546964e-01
8.68376195e-01 -3.15660775e-01 2.02866062e-03 3.01490426e-01
5.76881945e-01 -7.02441573e-01 -3.61995071e-01 6.32176042e-01
-1.92565203e+00 -7.33559310e-01 -1.21547095e-01 2.40374517e+00
6.75391316e-01 9.61241797e-02 6.17999017e-01 1.75761998e-01
8.19534540e-01 -3.85635048e-02 -1.01289284e+00 -6.94376767e-01
-1.60033926e-01 7.37925693e-02 9.17098880e-01 6.28582716e-01
-8.42127502e-01 6.92842126e-01 5.89024496e+00 7.98949659e-01
-8.05213392e-01 2.15170234e-01 1.31927741e+00 -3.26929808e-01
-7.85115480e-01 1.45911440e-01 -1.15659738e+00 6.38898432e-01
3.36987436e-01 -4.16436315e-01 6.50735974e-01 8.45908940e-01
1.54468864e-01 -5.78340173e-01 -1.20445967e+00 7.55025387e-01
-1.13659631e-02 -1.10213649e+00 -2.51785100e-01 4.22908157e-01
1.10257375e+00 -7.05274940e-01 3.08597177e-01 4.11146551e-01
4.87666965e-01 -9.30283487e-01 6.79852962e-01 2.28322238e-01
3.22946981e-02 -1.01883531e+00 6.26491308e-01 8.00724447e-01
-6.47323906e-01 -6.01644099e-01 -6.98768795e-02 -2.00209096e-01
4.23573107e-01 5.51494241e-01 -7.46338427e-01 2.14497745e-01
5.68425834e-01 -2.37375885e-01 -2.83438206e-01 1.46243429e+00
-4.39144075e-02 8.90332520e-01 -2.37396121e-01 -4.42073435e-01
2.46302813e-01 -4.42955464e-01 3.70298564e-01 4.20288861e-01
1.16704032e-01 2.84226835e-01 3.19262415e-01 8.07496488e-01
-2.06180707e-01 5.20731807e-01 2.57903878e-02 -3.52742106e-01
3.87370706e-01 9.92677569e-01 -8.78637135e-01 1.79442447e-02
-2.70609379e-01 7.96122253e-01 -5.93325198e-02 4.43857163e-01
-5.06262839e-01 -3.36367011e-01 3.53182644e-01 4.93421167e-01
4.29006696e-01 2.25540131e-01 -8.31871867e-01 -6.34388268e-01
4.38479483e-02 -7.17039406e-01 6.84601903e-01 -2.43123859e-01
-1.25286758e+00 -4.55659330e-02 -1.24485508e-01 -4.76842999e-01
-2.28893146e-01 -2.26781517e-01 -6.90836787e-01 7.56625473e-01
-1.30694997e+00 -3.55768293e-01 1.09806471e-01 1.71935469e-01
8.40700150e-01 6.89538196e-02 1.37448430e-01 2.68376410e-01
-4.46340084e-01 6.56862736e-01 1.06870420e-01 -1.49870738e-01
1.17812082e-01 -1.38899148e+00 -1.22737743e-01 3.79291594e-01
-1.98721397e-03 3.54537070e-01 5.43734014e-01 -5.89245439e-01
-1.38258243e+00 -6.69019580e-01 1.11574066e+00 -2.22217679e-01
1.83453411e-01 -4.87238258e-01 -4.55361575e-01 7.01448023e-01
2.29955073e-02 -7.16573775e-01 8.75141621e-01 4.22479600e-01
4.72086430e-01 -4.01259452e-01 -1.38400185e+00 5.92547297e-01
8.18641186e-01 6.85574207e-03 -2.22101614e-01 3.33698839e-01
7.09008574e-01 -1.81288600e-01 -1.98517144e-01 1.48612097e-01
2.76569664e-01 -1.07844615e+00 5.49925864e-01 -5.75380862e-01
3.58612835e-01 3.30732465e-01 -2.37288356e-01 -9.06722426e-01
-4.31386977e-01 -8.34961832e-01 5.85338436e-02 1.32045984e+00
8.71607482e-01 -8.53515148e-01 1.49716902e+00 7.93099642e-01
2.02113733e-01 -1.39643013e+00 -8.80076647e-01 -5.24720609e-01
7.67434910e-02 -2.30925664e-01 7.01120496e-01 5.59034288e-01
-1.24202542e-01 5.88275075e-01 -2.27767050e-01 -8.52664784e-02
5.61296880e-01 5.45063913e-01 5.31702340e-01 -1.12142193e+00
-6.25676811e-01 -6.19036257e-01 2.27637976e-01 -1.80771756e+00
-2.98785388e-01 -6.30810261e-01 1.76902577e-01 -1.53207743e+00
4.59206939e-01 -5.78030527e-01 -1.15792572e-01 -2.73693740e-01
-2.91419506e-01 -2.46079668e-01 3.65499616e-01 6.34970516e-03
-8.03780675e-01 2.08908349e-01 1.46973801e+00 -1.44267380e-01
-6.63179338e-01 9.66548681e-01 -1.51172030e+00 4.19578075e-01
8.20882857e-01 -3.37782979e-01 -7.25065470e-01 -3.43230106e-02
6.84455514e-01 6.13659501e-01 -7.07278252e-02 8.08377936e-02
3.73658985e-01 -6.32819831e-01 6.19673654e-02 -1.00366724e+00
3.50670576e-01 -6.99937463e-01 -2.14282945e-01 1.90636635e-01
-7.10239232e-01 2.82287207e-02 -4.43972319e-01 5.36571324e-01
8.50190991e-04 -9.14177954e-01 4.87606376e-01 -2.65916169e-01
-2.35223472e-01 6.25505805e-01 -5.10078847e-01 5.79842746e-01
1.13615751e+00 -4.62180287e-01 2.22934872e-01 -9.91263866e-01
-6.64996505e-01 9.09967601e-01 1.52309403e-01 1.46652341e-01
2.98806727e-01 -8.68936360e-01 -2.93361366e-01 -1.62951991e-01
-1.40407547e-01 1.28190562e-01 2.69641608e-01 6.55968130e-01
-1.13763530e-02 6.18183672e-01 3.95302296e-01 -4.00092453e-01
-9.27071869e-01 4.26906735e-01 8.40858519e-02 -4.62523133e-01
2.67678738e-01 1.20174229e+00 4.27568376e-01 -3.32779825e-01
4.91395295e-01 6.66632950e-02 -4.95434523e-01 5.87473691e-01
-9.89316702e-02 7.05788076e-01 -1.27256975e-01 -1.80606991e-01
-1.06824130e-01 1.93695962e-01 -3.55791152e-01 -3.05831909e-01
1.19295335e+00 -3.89122576e-01 7.12989047e-02 5.13043761e-01
9.13692653e-01 2.58571565e-01 -1.10086942e+00 -1.30313084e-01
1.68199837e-01 -7.17832685e-01 -1.04781017e-01 -8.71353567e-01
-1.15996325e+00 2.79725701e-01 3.78532618e-01 9.60786581e-01
1.15513253e+00 3.17250162e-01 1.03259063e+00 2.12426901e-01
3.17275614e-01 -1.25381267e+00 2.69360363e-01 2.70715475e-01
3.85427058e-01 -1.23838639e+00 -3.15487415e-01 -6.75079405e-01
-8.66803169e-01 4.45279837e-01 5.72120845e-01 -1.50681108e-01
5.06694078e-01 1.06132915e-02 -1.08827703e-01 -1.78708062e-01
-6.30367637e-01 -4.36469138e-01 1.25534683e-01 3.59308243e-01
2.62596346e-02 2.15632349e-01 -6.33036673e-01 9.82926726e-01
-5.86656272e-01 -1.46600708e-01 5.27079105e-01 6.78188324e-01
-7.19819725e-01 -1.23738956e+00 -4.21432644e-01 9.91099656e-01
-6.53204918e-01 5.45504577e-02 -5.42534411e-01 3.28645438e-01
2.74927586e-01 1.37143767e+00 2.33901292e-01 -9.40648466e-02
4.18550044e-01 -1.31805480e-01 4.26895469e-01 -9.09690857e-01
-6.26056314e-01 8.24909434e-02 9.20898244e-02 -1.85829552e-03
1.25230834e-01 -8.94092321e-01 -8.35593581e-01 -2.67586231e-01
-7.46552527e-01 4.34698194e-01 4.34116364e-01 1.03314030e+00
5.98812029e-02 2.58637577e-01 1.12754130e+00 -8.60048383e-02
-8.76707852e-01 -4.82576162e-01 -9.38397825e-01 7.58840293e-02
1.98582843e-01 -3.65893036e-01 -4.50165570e-01 -4.16022569e-01] | [4.652952671051025, 3.345635175704956] |
da686673-96e7-48f0-899e-5e125d692488 | mtab-matching-tabular-data-to-knowledge-graph | 1910.00246 | null | https://arxiv.org/abs/1910.00246v2 | https://arxiv.org/pdf/1910.00246v2.pdf | MTab: Matching Tabular Data to Knowledge Graph using Probability Models | This paper presents the design of our system, namely MTab, for Semantic Web Challenge on Tabular Data to Knowledge Graph Matching (SemTab 2019). MTab combines the voting algorithm and the probability models to solve critical problems of the matching tasks. Results on SemTab 2019 show that MTab obtains promising performance for the three matching tasks. | ['Phuc Nguyen', 'Ryutaro Ichise', 'Natthawut Kertkeidkachorn', 'Hideaki Takeda'] | 2019-10-01 | null | null | null | null | ['table-annotation', 'table-annotation'] | ['knowledge-base', 'natural-language-processing'] | [-3.31676602e-02 3.36606622e-01 -5.57134807e-01 -4.10850585e-01
-9.18580055e-01 -6.63440228e-01 8.29050779e-01 2.75330693e-01
-6.47520497e-02 8.32176924e-01 6.64631545e-04 -4.78553981e-01
-7.52550423e-01 -1.11661744e+00 -6.33888841e-01 1.96332321e-01
2.67786443e-01 1.39788818e+00 7.35731840e-01 -4.81239319e-01
1.79074109e-01 -1.01086251e-01 -1.66560030e+00 8.93140793e-01
7.64508009e-01 1.46908498e+00 -8.95460322e-02 2.62463033e-01
-7.68817723e-01 1.21791589e+00 -5.34827076e-03 -1.28379357e+00
4.04172719e-01 1.99032918e-01 -1.79140532e+00 -6.32334650e-01
1.22174394e+00 5.12850046e-01 -5.42405665e-01 1.35626876e+00
1.05591804e-01 1.29420921e-01 6.18697166e-01 -1.82047296e+00
-3.31341952e-01 1.19047296e+00 -2.80528277e-01 4.63653058e-02
9.76627231e-01 -7.10260391e-01 1.70334208e+00 -6.31386042e-01
1.15776527e+00 1.56562591e+00 9.85282600e-01 2.95133859e-01
-9.57677186e-01 -5.46687007e-01 -4.05292183e-01 1.24414802e+00
-1.50872660e+00 -4.19882298e-01 1.28804266e-01 -2.48992607e-01
1.29936016e+00 6.52391970e-01 3.14660072e-01 5.79298735e-01
3.04871559e-01 5.75793982e-01 1.10159087e+00 -4.46518064e-01
4.47633602e-02 1.24882154e-01 4.54601914e-01 1.06543899e+00
5.28588951e-01 1.65469170e-01 -9.13246155e-01 -7.13228464e-01
1.24020219e-01 -7.68565536e-01 2.53349632e-01 -6.62436724e-01
-6.75462604e-01 6.46705091e-01 3.02413255e-01 -3.46448831e-02
-4.44786437e-02 1.56816646e-01 4.75093842e-01 6.78667486e-01
8.51830244e-02 2.71259993e-01 -6.93682790e-01 2.98321515e-01
-6.40915215e-01 6.21139348e-01 1.19906831e+00 1.68366003e+00
5.46597064e-01 -5.86661100e-01 -3.87044579e-01 8.73292387e-01
6.72939181e-01 6.58713639e-01 4.01574373e-03 -9.95452166e-01
8.43758047e-01 8.82949412e-01 -5.88493608e-02 -1.10814381e+00
-2.96575665e-01 1.77711740e-01 -8.88856128e-02 -2.69613385e-01
5.11209726e-01 5.69601178e-01 -8.47108841e-01 1.28175974e+00
3.06283832e-01 -2.88841784e-01 1.26369536e-01 6.65463448e-01
1.38795602e+00 9.74672958e-02 4.77474570e-01 1.12526573e-01
1.80987382e+00 -1.02801394e+00 -7.79118955e-01 -2.95198321e-01
6.27215683e-01 -9.98723984e-01 4.81251985e-01 7.95391798e-02
-1.09984875e+00 -2.48601377e-01 -8.44608366e-01 -1.94818884e-01
-9.20227051e-01 -7.66173899e-01 8.43247831e-01 8.17848504e-01
-1.12313759e+00 3.98378521e-01 -1.52689591e-01 -7.73210645e-01
2.96697915e-01 2.20895708e-01 -5.04098296e-01 -5.21906853e-01
-1.84709609e+00 1.29547226e+00 8.04823339e-01 -5.91193438e-01
-2.01284423e-01 -6.40933573e-01 -1.02353418e+00 1.56152546e-01
7.18076289e-01 -1.16124427e+00 1.36914766e+00 -2.11278677e-01
-8.23323011e-01 1.12454474e+00 -1.18595257e-01 -6.18970156e-01
4.41297233e-01 3.67020220e-01 -7.77442813e-01 -2.75645256e-02
2.44005859e-01 4.96351987e-01 2.84265995e-01 -8.51324260e-01
-1.00898695e+00 -4.58737105e-01 -1.19024981e-02 -2.80778836e-02
1.94702968e-02 1.54394418e-01 -7.48273730e-01 -1.99531883e-01
1.89532816e-01 -6.07134998e-01 7.02248737e-02 -4.09230053e-01
-2.53652453e-01 -7.58364141e-01 5.59376836e-01 -7.02384651e-01
1.07572651e+00 -1.20509636e+00 4.45480756e-02 7.43543327e-01
9.03652906e-02 -1.39775783e-01 -1.86027691e-01 6.27211273e-01
-2.15124674e-02 -1.34501263e-01 1.20372929e-01 2.30964318e-01
5.90411782e-01 3.12088132e-01 -3.57906550e-01 -1.33243963e-01
-6.18900537e-01 1.45480049e+00 -8.33222628e-01 -1.29198945e+00
2.72367597e-02 -4.51614499e-01 -3.09186339e-01 -3.86661552e-02
-6.22518718e-01 -7.67387152e-01 -3.08286399e-01 1.06614983e+00
7.35151589e-01 -3.83908838e-01 9.85917330e-01 -6.19677126e-01
3.96574736e-01 5.12232900e-01 -1.11314917e+00 1.71955705e+00
8.88668448e-02 1.47374332e-01 -2.61410177e-01 -7.17920065e-01
7.66539097e-01 8.38169754e-02 6.32986963e-01 -1.28825366e+00
-9.23244804e-02 4.78940427e-01 -5.08889258e-01 -4.56745416e-01
7.09559560e-01 -1.91264786e-02 -4.31620181e-01 3.84243160e-01
2.94487953e-01 -3.88326913e-01 6.02462351e-01 5.92076719e-01
1.30233335e+00 1.94975987e-01 5.32577634e-01 -7.12754130e-01
5.19920588e-01 7.51340449e-01 3.10507655e-01 1.01196778e+00
1.70893401e-01 -2.50442419e-02 2.44859755e-01 -7.64419436e-01
-6.82456791e-01 -1.12401938e+00 -1.52086258e-01 9.88622248e-01
5.14439702e-01 -7.95437515e-01 -5.43360829e-01 -1.39403248e+00
7.34059691e-01 6.73223615e-01 -3.91083121e-01 -9.03977156e-02
-2.06568778e-01 -4.69609857e-01 5.52002907e-01 3.38026047e-01
5.42385936e-01 -1.17502785e+00 1.20863952e-01 1.33138523e-01
-5.92402756e-01 -1.50447392e+00 -3.20861191e-01 -7.67445415e-02
-4.94037509e-01 -1.97380912e+00 2.85961628e-01 -8.60776901e-01
2.62074947e-01 2.75808960e-01 1.69505119e+00 4.43504900e-02
-3.58981967e-01 5.87253213e-01 -2.76114315e-01 -2.52325714e-01
-2.52640069e-01 2.55202115e-01 -3.31328064e-01 -6.76354825e-01
6.48607671e-01 -1.93423271e-01 -1.27629146e-01 6.70874000e-01
-5.50720751e-01 -3.17665772e-03 7.71186352e-02 6.38769984e-01
4.53495353e-01 2.62829095e-01 2.31851682e-01 -1.13948441e+00
4.46530700e-01 -5.37140429e-01 -1.16175401e+00 9.85791743e-01
-1.46315432e+00 2.34477967e-01 1.53888732e-01 2.47166738e-01
-9.02931929e-01 -1.48966283e-01 1.24642268e-01 9.31766555e-02
5.48854709e-01 8.49152327e-01 -1.62036344e-01 -4.38334495e-01
6.05114222e-01 -1.97520837e-01 -2.15971157e-01 -5.42232752e-01
4.04337734e-01 5.84959388e-01 6.47974551e-01 -1.01025403e+00
9.17600632e-01 3.03215891e-01 2.52468854e-01 3.67457241e-01
-1.06511962e+00 -5.94733775e-01 -2.37776488e-01 -3.92870784e-01
3.67356658e-01 -5.39671123e-01 -9.67773318e-01 2.31214657e-01
-8.97344708e-01 -5.89083843e-02 -9.67799872e-02 -1.65892482e-01
-7.45281637e-01 3.54134023e-01 -3.45904768e-01 -4.79016185e-01
-7.91440248e-01 -4.98157769e-01 8.59663904e-01 4.66583148e-02
-2.15880632e-01 -1.28538859e+00 2.67394841e-01 1.14066267e+00
2.49955922e-01 -2.13533804e-01 1.40027034e+00 -1.03361225e+00
-6.28626406e-01 -2.73997098e-01 -7.05558360e-01 -4.83594358e-01
-6.96489364e-02 -3.40137601e-01 -5.61602831e-01 -2.70866930e-01
-8.68817031e-01 -6.39465928e-01 8.41986001e-01 -4.03424650e-02
1.10563350e+00 -2.25385264e-01 -6.39414549e-01 3.17815959e-01
1.54770970e+00 1.51258156e-01 8.38575840e-01 7.88580298e-01
5.53686142e-01 7.02291250e-01 9.05932546e-01 9.71868336e-02
1.30626845e+00 1.20695281e+00 4.01650548e-01 2.74843007e-01
-4.04941440e-01 -5.47015846e-01 -6.99669197e-02 7.90222406e-01
2.46955186e-01 -3.57314259e-01 -1.28498495e+00 5.36856472e-01
-2.36850405e+00 -9.13533509e-01 -4.87345904e-01 1.86640298e+00
7.65478253e-01 7.07386211e-02 -4.37017567e-02 -4.84564304e-02
8.63126040e-01 8.60363431e-03 -1.79294795e-01 -3.56020510e-01
-3.99832606e-01 5.19304991e-01 7.79783607e-01 6.54432297e-01
-8.09915602e-01 1.16806686e+00 7.72897005e+00 1.37418747e+00
-3.94189507e-02 1.17062345e-01 -1.08426571e-01 6.58867598e-01
-6.32831156e-01 4.39451128e-01 -8.28720331e-01 1.93561241e-01
8.02612484e-01 -8.21937382e-01 8.40144992e-01 6.18082166e-01
-9.18135345e-01 1.69889648e-02 -1.02684534e+00 7.17806935e-01
5.53018898e-02 -1.75822878e+00 7.32784212e-01 -1.94809675e-01
6.38349473e-01 -7.67195448e-02 -3.39587420e-01 5.94125867e-01
1.06523705e+00 -1.13078392e+00 7.79908895e-01 5.88140249e-01
6.04291677e-01 -4.66940135e-01 9.62593913e-01 -3.16639572e-01
-1.46061158e+00 1.34696096e-01 -4.36640650e-01 4.90569472e-01
-8.78103524e-02 1.80868521e-01 -7.91229486e-01 1.58933318e+00
9.73170161e-01 6.69064283e-01 -6.03788733e-01 9.92368400e-01
-1.27270855e-02 -6.99609146e-02 -1.08007036e-01 4.92864624e-02
-1.48374155e-01 1.67636707e-01 4.16419625e-01 8.66869986e-01
1.83648884e-01 -1.04870871e-01 3.74892913e-02 6.44195855e-01
-3.06568652e-01 1.45048589e-01 -3.72277021e-01 1.35923564e-01
9.52190936e-01 1.19485736e+00 -1.71903312e-01 -4.07184720e-01
-3.80200028e-01 3.02243829e-01 6.39428675e-01 -8.17741901e-02
-6.11521959e-01 -3.59234005e-01 2.82669783e-01 1.36666134e-01
2.49220744e-01 5.50927997e-01 -3.42799604e-01 -9.38855290e-01
6.88196719e-02 -1.08570671e+00 1.76198256e+00 -1.14804411e+00
-1.65904653e+00 4.44736272e-01 4.31323797e-01 -5.53066909e-01
-1.97943926e-01 -7.51707256e-01 -1.02883019e-01 6.57568216e-01
-1.65410066e+00 -1.27930081e+00 -4.50819641e-01 9.67505336e-01
5.05738929e-02 -5.59163511e-01 8.82149458e-01 6.07069850e-01
-1.09327033e-01 7.83721089e-01 -2.75725842e-01 -2.07208052e-01
6.46271884e-01 -1.42839956e+00 7.53433406e-01 3.55434537e-01
3.78736645e-01 1.95620522e-01 6.33879960e-01 -9.76459265e-01
-1.53283548e+00 -9.41480875e-01 1.52642417e+00 -6.80993736e-01
1.05815959e+00 -9.59245116e-02 -5.52204788e-01 7.07019508e-01
1.16900288e-01 -1.07955471e-01 2.08699346e-01 2.68146306e-01
-9.56666708e-01 -3.38797450e-01 -1.38825464e+00 3.14612359e-01
1.52576768e+00 -5.91904342e-01 -9.44672883e-01 4.81135905e-01
3.66255164e-01 -4.50300574e-01 -1.26978159e+00 8.90886605e-01
9.42355573e-01 -4.36126709e-01 1.13611555e+00 -1.09125853e+00
1.49122462e-01 -2.20355287e-01 -5.41683853e-01 -9.96717215e-01
-3.89573008e-01 -5.20419896e-01 -3.45834851e-01 1.07052481e+00
6.42306328e-01 -6.37458563e-01 1.12606990e+00 8.27105224e-01
3.25112402e-01 5.71403839e-02 -1.48191369e+00 -1.02086711e+00
-2.00931966e-01 -2.52341777e-01 1.03678679e+00 1.31063569e+00
5.06067038e-01 1.72910035e-01 -9.04963687e-02 -7.62089342e-02
1.18429101e+00 5.92751324e-01 5.89539468e-01 -1.61258316e+00
9.72585455e-02 -3.81671160e-01 -5.27780771e-01 -4.68547225e-01
4.12808090e-01 -1.62780046e+00 -4.33871359e-01 -1.82636619e+00
8.61431956e-01 -6.00228727e-01 -5.60599089e-01 8.08620989e-01
-5.20635881e-02 6.83917999e-02 -1.03972159e-01 9.81584489e-02
-1.20354080e+00 2.64701903e-01 8.71391833e-01 -6.82817161e-01
7.08894074e-01 -4.16496426e-01 -5.69050252e-01 1.90563947e-01
5.69678605e-01 -8.77547026e-01 -3.22466433e-01 -2.47458383e-01
7.33008265e-01 3.32395554e-01 2.03432873e-01 -5.70665061e-01
7.65817046e-01 -4.76979315e-01 -2.01079533e-01 -7.65072525e-01
-5.42073511e-02 -1.22550499e+00 8.14628065e-01 5.88359594e-01
-2.55102158e-01 3.66402447e-01 3.14148098e-01 3.64187360e-01
-2.12091208e-01 -2.45120898e-01 3.61454606e-01 8.48163962e-02
-1.07749736e+00 2.85649151e-01 -4.86992672e-02 5.20819664e-01
7.20247269e-01 -5.19616250e-03 -1.12922394e+00 1.62995961e-02
-5.77689052e-01 8.56228650e-01 1.62318170e-01 7.46690869e-01
4.39955354e-01 -1.76811743e+00 -7.34138787e-01 -1.86931744e-01
7.24168658e-01 -7.77725160e-01 -1.36268303e-01 6.54832244e-01
-4.72718984e-01 8.62427831e-01 -4.95791584e-01 -1.16885304e-02
-1.60607278e+00 4.85908628e-01 4.89776731e-01 -7.88046002e-01
-2.39807338e-01 3.61828744e-01 -6.27275348e-01 -8.68820846e-01
2.19469652e-01 4.95134681e-01 -3.30040812e-01 8.43461156e-02
8.95804316e-02 7.43009031e-01 5.62637746e-01 -1.87781438e-01
-8.64931107e-01 3.42471808e-01 -2.55633086e-01 9.83467251e-02
1.06104743e+00 -8.73083398e-02 -7.06118286e-01 -1.81567132e-01
5.42987823e-01 -3.12873334e-01 5.49883842e-02 -8.16381991e-01
8.87221038e-01 -5.25578678e-01 1.13084525e-01 -1.45297766e+00
-1.04767847e+00 -1.68706015e-01 2.28061169e-01 3.68073791e-01
5.26192367e-01 1.63820550e-01 1.07579851e+00 6.46389902e-01
1.05426550e+00 -1.21890152e+00 -5.97664833e-01 6.22618020e-01
7.27328062e-01 -1.10110641e+00 2.01927796e-01 -8.70290279e-01
-2.34739885e-01 9.98611867e-01 6.99926615e-01 4.60710526e-01
6.03792548e-01 2.16192305e-01 3.87578122e-02 -9.23187017e-01
-1.21814835e+00 -3.90745014e-01 8.88964415e-01 5.91103137e-01
-1.33496463e-01 1.93329770e-02 -4.90458906e-01 5.43830156e-01
-3.99875402e-01 -8.54221582e-02 -2.05488101e-01 8.61063659e-01
-2.21459687e-01 -1.48599541e+00 -2.30837405e-01 5.67096531e-01
-2.19480455e-01 -3.40475559e-01 -9.89555359e-01 1.06745124e+00
-3.54420066e-01 1.11668289e+00 -1.40510499e-01 -8.13066542e-01
6.53921604e-01 2.16870934e-01 8.63645852e-01 -1.07506186e-01
-7.85076559e-01 -8.43350232e-01 1.07117093e+00 -9.31273162e-01
-3.79679412e-01 -6.58378184e-01 -9.20458496e-01 -1.03796196e+00
-2.85884380e-01 7.12176800e-01 3.85661423e-01 7.77545154e-01
1.52568892e-01 1.92038372e-01 3.00001204e-01 4.95598733e-01
-3.79303724e-01 -6.10795140e-01 -7.07283854e-01 5.40526390e-01
-5.39759815e-01 -7.55055726e-01 -7.32227266e-02 -1.81244582e-01] | [9.260332107543945, 8.021889686584473] |
bb9b73d2-4a97-41f7-af7a-1345c2173918 | splitting-numerical-integration-for-matrix | 2202.06482 | null | https://arxiv.org/abs/2202.06482v1 | https://arxiv.org/pdf/2202.06482v1.pdf | Splitting numerical integration for matrix completion | Low rank matrix approximation is a popular topic in machine learning. In this paper, we propose a new algorithm for this topic by minimizing the least-squares estimation over the Riemannian manifold of fixed-rank matrices. The algorithm is an adaptation of classical gradient descent within the framework of optimization on manifolds. In particular, we reformulate an unconstrained optimization problem on a low-rank manifold into a differential dynamic system. We develop a splitting numerical integration method by applying a splitting integration scheme to the dynamic system. We conduct the convergence analysis of our splitting numerical integration algorithm. It can be guaranteed that the error between the recovered matrix and true result is monotonically decreasing in the Frobenius norm. Moreover, our splitting numerical integration can be adapted into matrix completion scenarios. Experimental results show that our approach has good scalability for large-scale problems with satisfactory accuracy | ['Qianqian Song'] | 2022-02-14 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [-2.01974064e-01 -2.94526387e-02 1.39240980e-01 -6.43656924e-02
-8.55006039e-01 -3.96312833e-01 2.17601120e-01 -5.62865376e-01
-4.23890024e-01 5.79788744e-01 -3.27796750e-02 -1.98624849e-01
-2.69856840e-01 -3.08677733e-01 -7.01915145e-01 -8.33794534e-01
-1.02398708e-01 3.86335254e-02 -1.68058544e-01 -4.17969793e-01
1.16038568e-01 5.62786281e-01 -7.08886147e-01 -3.37577552e-01
1.00642860e+00 7.22123265e-01 3.95016745e-04 6.12845242e-01
3.08180302e-01 6.63249850e-01 3.83786224e-02 -2.56500661e-01
7.11589456e-01 -3.76546383e-01 -8.61948907e-01 5.82887590e-01
6.29187584e-01 -3.21751326e-01 -5.14557481e-01 1.42617500e+00
3.43620688e-01 3.82847726e-01 5.14519870e-01 -1.11363840e+00
-2.48661771e-01 6.26706108e-02 -7.04550028e-01 -1.26829520e-02
2.52595127e-01 -3.52938592e-01 1.03378236e+00 -1.65131187e+00
5.30631781e-01 1.29136992e+00 7.67533243e-01 1.73769876e-01
-1.47803891e+00 -9.45985317e-02 2.81004421e-02 5.81532195e-02
-1.84285378e+00 -3.11487973e-01 8.07172596e-01 -7.27743745e-01
2.83943683e-01 4.84471858e-01 4.20908958e-01 1.07820638e-01
5.32657243e-02 7.31860995e-01 8.51138711e-01 -2.63739526e-01
1.37488723e-01 6.21922985e-02 1.33715600e-01 9.55373228e-01
3.29270929e-01 -3.48334730e-01 -8.75561610e-02 -2.87593901e-01
8.52667809e-01 -9.14031174e-04 -3.89052600e-01 -7.13083982e-01
-1.41671944e+00 1.11516333e+00 4.29826021e-01 3.11198980e-01
-2.10308194e-01 7.53029585e-02 -3.13196927e-02 3.68283540e-01
5.89427352e-01 1.71404034e-01 2.06373949e-02 1.27593368e-01
-7.36477673e-01 1.53798223e-01 1.13712728e+00 8.33492279e-01
1.03821325e+00 3.18624943e-01 2.74042338e-01 8.73945653e-01
4.67115313e-01 6.80868626e-01 2.26913884e-01 -1.22686696e+00
4.21987206e-01 4.60993230e-01 1.95146024e-01 -1.54863584e+00
-4.12992835e-01 -3.83972168e-01 -1.03491068e+00 1.56206697e-01
3.93704176e-01 -4.22487319e-01 4.13869172e-02 1.58460605e+00
7.16117263e-01 3.53104919e-01 -4.04772013e-02 1.18317950e+00
1.76928967e-01 7.31155932e-01 -7.75249183e-01 -5.90756238e-01
8.50285530e-01 -1.06377804e+00 -8.62591326e-01 1.61276519e-01
9.26142514e-01 -7.47415423e-01 7.80191362e-01 4.13139164e-01
-9.47541535e-01 -2.53131032e-01 -1.23885036e+00 -9.04725045e-02
1.70231864e-01 6.27376378e-01 4.10353601e-01 4.58837718e-01
-1.08061302e+00 9.34202611e-01 -8.12343776e-01 -1.44306749e-01
-2.94422865e-01 4.34663236e-01 -5.93430817e-01 2.15942264e-01
-9.02398229e-01 6.42670572e-01 7.46249482e-02 5.94579935e-01
-4.34353203e-01 -5.57080090e-01 -9.70871329e-01 -4.90089566e-01
3.11897099e-01 -3.89184356e-01 9.75632727e-01 -8.20470035e-01
-1.63939607e+00 7.21286714e-01 -1.53897628e-01 -5.73242195e-02
6.63899958e-01 -4.83911842e-01 -1.69154987e-01 1.11969724e-01
1.25144467e-01 -1.08456559e-01 1.16652369e+00 -9.90961015e-01
-2.34143555e-01 -3.07148218e-01 9.44175720e-02 3.93559724e-01
-5.13810873e-01 -2.36567110e-01 -4.83443975e-01 -6.41741395e-01
4.91755128e-01 -1.30841923e+00 -5.75215340e-01 1.22500472e-01
-3.49318624e-01 1.01109535e-01 6.39224589e-01 -8.15795541e-01
1.45028687e+00 -2.27494025e+00 7.01242387e-01 2.71488249e-01
2.28184402e-01 -1.23379612e-02 6.73958194e-03 5.46848297e-01
-1.56024769e-01 -1.76733866e-01 -5.67874551e-01 -2.61481136e-01
-1.37496695e-01 -4.37176600e-02 -3.50515276e-01 1.06762874e+00
-2.99157351e-02 4.61365372e-01 -9.54029858e-01 -4.05745715e-01
2.86410679e-03 4.12722290e-01 -6.24937713e-01 1.69360876e-01
4.10776913e-01 3.65582973e-01 -5.71625412e-01 2.13842019e-01
9.34503973e-01 -2.70389020e-01 2.67784476e-01 -6.78943098e-01
-2.48023003e-01 -2.47476637e-01 -1.87241089e+00 1.79247093e+00
-4.19001639e-01 4.39229071e-01 8.20344746e-01 -1.22008586e+00
6.43872023e-01 8.50330144e-02 7.83737183e-01 2.38752082e-01
1.11446790e-01 4.99143183e-01 -8.21752995e-02 -3.82038414e-01
5.09761810e-01 -1.42189264e-01 2.10981250e-01 3.19581091e-01
-2.53880948e-01 -1.42362982e-01 4.44259703e-01 3.71593624e-01
7.20427990e-01 -3.27865407e-02 2.02869296e-01 -7.08741784e-01
1.12347674e+00 -1.59799039e-01 5.75596988e-01 1.26688510e-01
1.22914940e-01 5.71054161e-01 5.10757923e-01 -2.24122196e-01
-9.73105252e-01 -7.89424479e-01 -4.36230242e-01 6.92632556e-01
1.41718864e-01 -4.32448804e-01 -1.05398571e+00 -5.60437024e-01
6.85140938e-02 3.52117904e-02 -3.80447447e-01 -3.90296653e-02
-6.97226822e-01 -1.02777588e+00 1.31582201e-01 1.55015633e-01
6.14232600e-01 -1.09580122e-01 2.33705491e-01 1.54272854e-01
-1.03042282e-01 -1.14037681e+00 -1.08261442e+00 -5.62541366e-01
-1.15183425e+00 -1.13599074e+00 -7.93233812e-01 -8.95736814e-01
1.03296995e+00 6.50398850e-01 5.12673497e-01 5.99549785e-02
-1.88331127e-01 6.29343390e-01 1.42089918e-01 4.48978454e-01
-1.74492434e-01 -1.49089351e-01 5.60723543e-01 8.32167566e-01
-1.93879724e-01 -4.53846812e-01 -5.94515085e-01 6.37997389e-01
-1.03217959e+00 -4.89434786e-03 3.14584941e-01 8.05738807e-01
7.69487500e-01 -9.52709913e-02 2.94065714e-01 -6.31637156e-01
7.99701154e-01 -4.43141222e-01 -9.67909932e-01 5.54309934e-02
-6.10140860e-01 2.96916515e-01 5.84548593e-01 -4.06020105e-01
-7.40699887e-01 4.76604164e-01 1.94778755e-01 -5.11288702e-01
8.30477059e-01 7.15154469e-01 -2.03603640e-01 -6.41547024e-01
4.87520695e-01 1.25096366e-01 4.13419247e-01 -7.62465537e-01
6.99131131e-01 5.86289048e-01 3.89186114e-01 -6.69927895e-01
1.20830882e+00 8.70478630e-01 4.43644464e-01 -1.11377060e+00
-8.41196775e-01 -7.16276169e-01 -8.48873138e-01 -2.78235018e-01
6.00234985e-01 -9.78335619e-01 -6.86926544e-01 4.36323076e-01
-9.14900720e-01 -2.90867567e-01 1.44403540e-02 8.83434474e-01
-7.11993337e-01 9.39756930e-01 -7.85782695e-01 -7.37815976e-01
-1.76941678e-01 -1.02776611e+00 8.85395885e-01 -1.56608656e-01
1.78321853e-01 -1.38197434e+00 5.43178141e-01 1.16064072e-01
1.96685847e-02 2.47804016e-01 2.52845526e-01 -2.35314399e-01
-5.20415604e-01 -4.09835398e-01 -2.69964606e-01 6.72996998e-01
7.72019178e-02 -8.92795920e-02 -4.00716811e-01 -6.00181997e-01
5.62294900e-01 4.05067950e-03 4.94645774e-01 3.04232333e-02
6.72865570e-01 -5.37894785e-01 -1.07091494e-01 9.02951360e-01
1.54217350e+00 -4.40983802e-01 2.95054406e-01 1.34391770e-01
1.11623657e+00 6.09511971e-01 8.38466465e-01 3.54723662e-01
2.17705041e-01 8.27764988e-01 1.82161346e-01 -3.01147848e-02
2.80909240e-01 -1.90916732e-02 7.21457720e-01 1.50511134e+00
-1.42790545e-02 6.74019158e-01 -7.12016582e-01 4.39767480e-01
-2.28033876e+00 -6.59335911e-01 -5.08204639e-01 2.57497430e+00
8.55438352e-01 -4.20923382e-01 1.79625556e-01 1.98834520e-02
8.84150624e-01 4.71349806e-02 -4.62721050e-01 -2.51121908e-01
-3.87796126e-02 -3.97885859e-01 6.18311703e-01 1.07114160e+00
-1.17532599e+00 7.68425465e-01 6.44619322e+00 8.69694889e-01
-1.02473092e+00 9.00460109e-02 1.33382678e-01 9.33970883e-02
3.51036340e-02 1.01203665e-01 -7.51442611e-01 1.70586795e-01
6.32024109e-01 -4.93256629e-01 7.05795765e-01 1.01726580e+00
4.60421294e-01 1.70277014e-01 -9.65556324e-01 1.16825855e+00
1.19612552e-01 -1.01873600e+00 -1.66392505e-01 3.04812640e-01
1.02103281e+00 -1.57428265e-01 7.72755593e-03 -8.62441733e-02
3.04813534e-02 -6.80009186e-01 3.08786362e-01 4.79472995e-01
6.82564795e-01 -5.96305013e-01 3.15643817e-01 2.93590873e-01
-1.32958221e+00 2.06808671e-01 -5.31634152e-01 -1.26759052e-01
3.07033777e-01 9.05396760e-01 -4.06795472e-01 5.19153714e-01
2.17592135e-01 1.09164178e+00 -3.56867909e-01 9.96725738e-01
-6.88061416e-02 3.19886684e-01 -4.62536186e-01 2.00704217e-01
1.89634398e-01 -1.51305807e+00 9.98744667e-01 1.05542409e+00
3.02881092e-01 2.04547629e-01 3.48408669e-01 5.81473649e-01
-6.01952597e-02 7.51751363e-01 -7.32441008e-01 6.08020723e-02
-1.53126374e-01 1.76536918e+00 -3.99165273e-01 -1.21971130e-01
-6.01575434e-01 1.07201028e+00 4.62362766e-01 4.92248774e-01
-5.43281853e-01 -4.00156617e-01 6.61443174e-01 -7.45291635e-02
2.86493152e-01 -6.87611639e-01 -6.49208799e-02 -1.78177452e+00
3.65450352e-01 -8.02469611e-01 2.06617668e-01 -5.00158608e-01
-1.09502554e+00 3.64215225e-01 -1.59964800e-01 -1.42934132e+00
-1.16447404e-01 -8.81632209e-01 -5.95295131e-01 5.90768337e-01
-9.81585860e-01 -7.07828164e-01 -8.04867130e-03 8.18697810e-01
1.55210763e-01 -3.12620103e-02 4.07573551e-01 5.32531977e-01
-9.00916398e-01 3.53857160e-01 7.01650560e-01 2.65804499e-01
4.98788893e-01 -1.48334670e+00 -2.02964574e-01 1.10248876e+00
6.18555583e-04 8.19959819e-01 5.91628373e-01 -5.40447772e-01
-1.95229113e+00 -1.16099262e+00 4.94675219e-01 -1.61666363e-01
1.26217163e+00 -2.53194600e-01 -9.90611374e-01 8.32072735e-01
-5.71868345e-02 7.29200318e-02 5.50474107e-01 -6.17308058e-02
-1.62026241e-01 -3.07974696e-01 -9.90286529e-01 7.37617254e-01
8.57499540e-01 -5.88794291e-01 -2.78614610e-01 8.39495063e-01
4.66773242e-01 -4.49460417e-01 -1.28588557e+00 3.96675348e-01
2.62994170e-01 -2.90067017e-01 8.59666586e-01 -6.71920061e-01
-3.80674824e-02 -7.43932903e-01 -2.75361419e-01 -1.19873500e+00
-2.70395160e-01 -1.28337550e+00 -3.13294500e-01 8.63316536e-01
3.57987642e-01 -7.00578570e-01 5.65477550e-01 6.61164999e-01
3.41492854e-02 -8.54564011e-01 -8.65968287e-01 -1.08921850e+00
1.32199332e-01 -2.82561570e-01 -1.55242667e-01 8.80848408e-01
3.84497643e-01 4.89352852e-01 -7.44880378e-01 2.66925573e-01
9.95233953e-01 1.02620281e-01 9.54718769e-01 -9.19744730e-01
-3.33322078e-01 -1.61726072e-01 -3.70414108e-01 -1.37912726e+00
3.47972572e-01 -1.06630671e+00 3.94001864e-02 -1.24750412e+00
3.13936844e-02 -1.90111641e-02 -3.52220610e-02 -2.28593752e-01
-1.34477571e-01 2.39118114e-01 1.77560180e-01 4.83975440e-01
-6.02173924e-01 7.13646591e-01 1.21822429e+00 6.37998730e-02
-2.56739438e-01 5.15288748e-02 -2.78726488e-01 9.17622983e-01
5.98419011e-01 -2.57770300e-01 -2.67738581e-01 -1.98571399e-01
4.47548151e-01 -3.22950482e-02 4.09810729e-02 -8.67507637e-01
2.08423495e-01 -1.08389907e-01 -4.00365502e-01 -3.06515574e-01
3.34001601e-01 -7.09193766e-01 1.89946547e-01 5.25975764e-01
-5.16133606e-02 6.63489178e-02 -1.69057563e-01 6.23511970e-01
-1.82925835e-01 -4.20191377e-01 8.51449907e-01 2.79477924e-01
-2.53718704e-01 5.45542538e-01 -3.66804838e-01 1.59291252e-01
7.33797073e-01 1.57059312e-01 2.48373792e-01 -5.32604456e-01
-9.09602940e-01 2.27046162e-01 5.31160355e-01 9.62996557e-02
5.91946721e-01 -1.81918371e+00 -7.22421229e-01 -5.27578581e-04
-2.62059957e-01 -2.86066115e-01 -5.34535013e-02 1.51183426e+00
-7.89556086e-01 3.15051317e-01 4.34293747e-01 -6.32636786e-01
-1.06958139e+00 6.12535298e-01 6.79691792e-01 -1.96963415e-01
-3.09224308e-01 3.80920172e-01 2.98040390e-01 -5.92770457e-01
-8.42821524e-02 -1.85006838e-02 -3.53952800e-03 6.19847886e-03
6.37379825e-01 9.04864073e-01 -7.97443613e-02 -1.07212698e+00
-1.89814851e-01 1.01480556e+00 9.64567214e-02 -4.48901892e-01
1.14784837e+00 -4.06449109e-01 -7.49132574e-01 5.87041140e-01
1.97189355e+00 3.58097017e-01 -1.25212026e+00 -4.60896879e-01
3.89509127e-02 -3.81486028e-01 1.51547894e-01 3.93133551e-01
-1.25501347e+00 5.04960179e-01 4.02349234e-01 2.83470452e-01
7.66076207e-01 -4.43993360e-01 6.86390877e-01 8.52195263e-01
1.91636041e-01 -1.24839032e+00 3.79178561e-02 5.30307949e-01
1.19536829e+00 -1.17948067e+00 3.39198112e-01 -7.32970715e-01
-2.74970084e-01 1.17223990e+00 4.25168604e-01 -7.04939604e-01
1.05019963e+00 -3.41807991e-01 -3.09789088e-02 -4.68661562e-02
-3.35687339e-01 -1.84456203e-02 6.78116918e-01 1.79011337e-02
4.46946293e-01 -3.46674211e-02 -9.64576960e-01 5.99398911e-02
-6.16124831e-02 -2.65166402e-01 7.28145063e-01 5.61177671e-01
-3.46449494e-01 -1.11676240e+00 -5.48795044e-01 6.59794286e-02
-4.04320210e-01 6.70524985e-02 -3.13635945e-01 6.03194892e-01
-5.34001410e-01 9.55841064e-01 -3.67447704e-01 -2.50176936e-01
3.01597744e-01 -1.27630889e-01 4.79803115e-01 -5.77588499e-01
2.14679576e-02 3.36752802e-01 -5.05071841e-02 -7.54433990e-01
-3.91784757e-01 -8.61616373e-01 -1.22875452e+00 -2.42062852e-01
-3.54159981e-01 4.81140047e-01 7.48786569e-01 8.15512300e-01
3.32927287e-01 -1.13239355e-01 1.17895639e+00 -6.72251642e-01
-1.03162694e+00 -7.96242714e-01 -9.76609468e-01 4.61706758e-01
3.54939222e-01 -5.60076296e-01 -8.10203850e-01 5.75144067e-02] | [7.434655666351318, 4.338878631591797] |
4691ff4a-a2ee-430f-902c-89ef41ab04a2 | face-photo-sketch-recognition-using | 2108.09898 | null | https://arxiv.org/abs/2108.09898v1 | https://arxiv.org/pdf/2108.09898v1.pdf | Face Photo-Sketch Recognition Using Bidirectional Collaborative Synthesis Network | This research features a deep-learning based framework to address the problem of matching a given face sketch image against a face photo database. The problem of photo-sketch matching is challenging because 1) there is large modality gap between photo and sketch, and 2) the number of paired training samples is insufficient to train deep learning based networks. To circumvent the problem of large modality gap, our approach is to use an intermediate latent space between the two modalities. We effectively align the distributions of the two modalities in this latent space by employing a bidirectional (photo -> sketch and sketch -> photo) collaborative synthesis network. A StyleGAN-like architecture is utilized to make the intermediate latent space be equipped with rich representation power. To resolve the problem of insufficient training samples, we introduce a three-step training scheme. Extensive evaluation on public composite face sketch database confirms superior performance of our method compared to existing state-of-the-art methods. The proposed methodology can be employed in matching other modality pairs. | ['Juneho Yi', 'Hyunkyu Park', 'Nizam Ud Din', 'Seho Bae'] | 2021-08-23 | null | null | null | null | ['sketch-recognition'] | ['computer-vision'] | [ 4.09786224e-01 1.21612169e-01 -2.32987866e-01 -4.26647007e-01
-8.57614756e-01 -5.59875965e-01 1.06283772e+00 -7.47466743e-01
5.38647622e-02 4.33047891e-01 2.57894099e-01 -1.05860069e-01
1.63163200e-01 -7.25813746e-01 -7.48666704e-01 -6.97084129e-01
6.22012198e-01 1.96693331e-01 -2.06054077e-01 1.81862023e-02
1.96152508e-01 5.22451162e-01 -1.52511775e+00 5.36423087e-01
3.96164328e-01 1.03515625e+00 4.69978489e-02 3.53772730e-01
-5.64918041e-01 4.44187284e-01 -2.10389525e-01 -7.00918555e-01
8.12240779e-01 -7.85862446e-01 -6.44588470e-01 2.37781316e-01
1.03480780e+00 -6.78884447e-01 -5.96879184e-01 1.08329129e+00
6.68013334e-01 -1.20348416e-01 6.70426309e-01 -1.68923581e+00
-8.86642456e-01 2.69845575e-01 -6.81548834e-01 -4.55772787e-01
5.29315412e-01 2.62571461e-02 6.57387316e-01 -1.05771494e+00
6.18154943e-01 1.51712584e+00 7.00139999e-01 8.75439584e-01
-1.01576471e+00 -9.69416142e-01 -1.81474835e-01 -5.64706288e-02
-1.40867400e+00 -9.29439604e-01 1.24269748e+00 -3.98406744e-01
2.80215174e-01 6.42622449e-03 3.68370414e-01 1.29931676e+00
-2.06913203e-01 6.39069915e-01 1.14801478e+00 -5.53892076e-01
-1.63362086e-01 2.36150667e-01 -6.21758401e-01 8.82815540e-01
-3.43849361e-01 1.55194804e-01 -7.59632409e-01 -3.52483153e-01
1.15406954e+00 3.38141859e-01 8.30823630e-02 -4.16946709e-01
-1.10841155e+00 5.40062428e-01 3.27458113e-01 2.98550218e-01
-3.02061737e-01 2.23720163e-01 1.60990909e-01 4.90388453e-01
2.21897423e-01 -1.34627577e-02 -2.68921517e-02 2.29041710e-01
-1.23650050e+00 2.96516400e-02 5.75069666e-01 1.10519922e+00
7.87327409e-01 8.49852562e-02 -1.52661383e-01 9.98903632e-01
7.44220138e-01 5.59590399e-01 2.44153291e-01 -1.03176785e+00
7.94372261e-01 5.69148302e-01 -8.45988691e-02 -1.15011382e+00
4.16553140e-01 1.85692981e-01 -1.05675817e+00 3.25126946e-01
5.38231611e-01 1.13009281e-01 -9.16255176e-01 1.86636782e+00
3.06433439e-01 4.04800266e-01 5.54261841e-02 6.99859560e-01
8.71528029e-01 5.70570469e-01 5.75693138e-02 -4.13067499e-03
1.30139065e+00 -8.30581844e-01 -6.98791265e-01 4.56646569e-02
-2.33718585e-02 -1.19348943e+00 8.89758825e-01 -1.74242482e-01
-1.38850403e+00 -8.70719135e-01 -1.16112530e+00 -2.44952694e-01
-4.30064470e-01 4.78820264e-01 4.03889507e-01 7.50268698e-01
-1.15411866e+00 4.70052183e-01 -3.42884600e-01 -5.30481458e-01
5.78146338e-01 4.21453327e-01 -8.44689310e-01 -3.19946587e-01
-8.89363527e-01 5.69006503e-01 1.55339837e-01 3.54661256e-01
-7.95226157e-01 -7.17152536e-01 -7.30500817e-01 1.79972593e-02
-4.58679199e-02 -7.26154327e-01 8.34905982e-01 -1.51683223e+00
-1.91199923e+00 1.09038734e+00 -1.52789667e-01 1.90541357e-01
8.09837639e-01 1.57906383e-01 -3.82616460e-01 1.83050886e-01
-7.32881799e-02 9.65564907e-01 1.38323307e+00 -1.39151871e+00
-3.98357123e-01 -3.89635235e-01 3.04181222e-02 1.44815743e-01
-3.85121167e-01 -1.61517863e-04 -7.55068064e-01 -6.31122530e-01
3.78611743e-01 -1.02894306e+00 2.97634691e-01 4.43035007e-01
-2.10988507e-01 -2.79453456e-01 9.50326622e-01 -7.18248010e-01
6.89507723e-01 -2.05022264e+00 -5.57701029e-02 2.87355453e-01
9.34779048e-02 3.92599612e-01 -4.91710812e-01 7.64231026e-01
-1.99244991e-01 -1.06828421e-01 -1.49815232e-01 -7.05370486e-01
1.45076394e-01 7.82054737e-02 -5.20828784e-01 5.63184977e-01
1.94301963e-01 8.21788192e-01 -6.14746451e-01 -6.11104190e-01
1.24964848e-01 6.78675592e-01 -4.06786770e-01 6.76724255e-01
-5.76300435e-02 5.52073956e-01 -2.15061322e-01 9.54353154e-01
1.25916457e+00 -1.20935209e-01 4.37567741e-01 -6.18188560e-01
1.60541788e-01 -1.38221085e-01 -1.23278797e+00 2.03244758e+00
-4.79701966e-01 5.97754836e-01 2.63645276e-02 -7.07187474e-01
1.09250128e+00 6.42768800e-01 3.79703015e-01 -5.43631136e-01
1.59272730e-01 2.35846251e-01 -3.80360097e-01 -6.14011049e-01
2.50503927e-01 -3.55324388e-01 2.54865289e-01 5.89207053e-01
3.31582725e-01 9.08526257e-02 -3.08869153e-01 -6.30626902e-02
6.73667312e-01 4.26328391e-01 -9.25084203e-02 2.79438496e-02
8.70124817e-01 -6.89260840e-01 4.32683438e-01 6.03396356e-01
-1.10502616e-01 7.16546237e-01 4.86750096e-01 -3.80452156e-01
-1.42563868e+00 -1.01783931e+00 2.16662973e-01 8.95799279e-01
-1.60991140e-02 -3.24964561e-02 -6.68437481e-01 -7.28516459e-01
9.94897559e-02 -1.17294349e-01 -6.93256676e-01 1.77953392e-01
-6.22565985e-01 -9.75185912e-03 8.90668869e-01 5.26632488e-01
8.20544481e-01 -9.99997914e-01 1.47652686e-01 -1.22848794e-01
-3.21881287e-02 -1.20807028e+00 -6.50020897e-01 -8.04169297e-01
-8.19602966e-01 -1.09795272e+00 -9.25477624e-01 -1.09649873e+00
1.03273118e+00 3.12177181e-01 8.68922353e-01 4.09270465e-01
5.45694306e-02 5.67669988e-01 4.39893501e-03 2.85428576e-02
-5.35802782e-01 -1.10851079e-01 2.74017043e-02 4.47540104e-01
3.19710702e-01 -7.62355030e-01 -8.34042192e-01 2.60329157e-01
-1.07467234e+00 2.08241165e-01 6.60540342e-01 1.03769612e+00
1.55838490e-01 -3.70697945e-01 5.98187447e-01 -6.77612603e-01
5.04390895e-01 -3.67788672e-01 -6.96138918e-01 7.42228687e-01
-3.44803005e-01 1.06111929e-01 4.19113994e-01 -5.18399477e-01
-1.29083288e+00 4.20851141e-01 -1.19319230e-01 -6.05219543e-01
-7.59734064e-02 1.22207016e-01 -5.28508663e-01 -3.32244635e-01
8.74446705e-02 1.54203132e-01 3.04690212e-01 -5.13178587e-01
5.06584704e-01 9.00849283e-01 6.81313276e-01 -6.68383658e-01
1.10711408e+00 5.98534763e-01 2.90141463e-01 -6.18660450e-01
-1.77546248e-01 -1.29637375e-01 -8.18941951e-01 -1.60512269e-01
6.84849799e-01 -9.64080334e-01 -9.93744791e-01 6.11905396e-01
-1.28027797e+00 7.50689954e-02 2.05635488e-01 2.66184211e-01
-5.22961795e-01 5.93314409e-01 -4.68619078e-01 -8.19035351e-01
-3.73061538e-01 -1.15183496e+00 1.22718740e+00 3.66141349e-01
3.24518204e-01 -9.23015118e-01 1.24473631e-01 4.57246959e-01
4.10166860e-01 1.77040860e-01 7.81233668e-01 -3.66142035e-01
-8.34086180e-01 -4.21223760e-01 -5.94881535e-01 2.57545501e-01
3.06793213e-01 1.73177496e-01 -1.20416164e+00 -3.93953234e-01
-2.19943449e-01 -5.24766088e-01 4.99589771e-01 -1.85493886e-01
9.78521705e-01 -2.32839793e-01 -2.45966494e-01 6.97760403e-01
1.44393528e+00 -3.26901637e-02 8.39637756e-01 -1.07330993e-01
8.59414697e-01 7.08158433e-01 2.30064467e-01 1.56748801e-01
4.31701899e-01 7.02221513e-01 4.50519621e-02 -2.55814612e-01
-3.58266860e-01 -8.94353092e-01 2.27822185e-01 7.05727160e-01
1.29984513e-01 -5.10183014e-02 -5.94878972e-01 4.22495723e-01
-1.69705844e+00 -1.05390239e+00 4.59102720e-01 2.10383368e+00
7.03103185e-01 -4.15414214e-01 7.28129502e-03 4.72621284e-02
8.04845095e-01 1.27358481e-01 -2.78957933e-01 -3.88598554e-02
-7.92865679e-02 2.15475028e-03 2.48499200e-01 4.69214290e-01
-9.06761646e-01 8.82921576e-01 6.05782604e+00 7.45121300e-01
-1.26821339e+00 1.36963084e-01 4.18012351e-01 2.88314223e-01
-3.90846521e-01 2.13988230e-01 -5.71224570e-01 5.48689246e-01
4.90958333e-01 4.11112398e-01 6.03546619e-01 4.51549917e-01
-2.51166523e-01 2.24622592e-01 -1.47825658e+00 1.39256394e+00
3.03464711e-01 -1.55102479e+00 2.12289199e-01 9.23415199e-02
7.91109025e-01 -4.45709825e-01 3.07579160e-01 7.53236413e-02
-1.03593066e-01 -9.52357471e-01 5.27374446e-01 7.21285105e-01
1.25615728e+00 -4.58671331e-01 4.46768939e-01 5.34716472e-02
-1.24404263e+00 1.53343789e-02 -4.78391886e-01 2.89997905e-01
-1.48156524e-01 -7.09592598e-03 -6.95364475e-01 6.21459424e-01
3.67734551e-01 4.35565144e-01 -3.42818528e-01 6.78553939e-01
-7.41604418e-02 4.13987003e-02 -2.36089304e-01 5.77125788e-01
3.86438146e-02 -3.60330731e-01 2.18694031e-01 8.68787527e-01
5.28952658e-01 5.86547740e-02 6.50836453e-02 1.00268865e+00
-5.57557285e-01 -7.06185251e-02 -1.02335560e+00 -3.31166267e-01
7.60393918e-01 1.21851456e+00 -3.49049866e-01 -3.67844045e-01
-5.16196430e-01 1.24703991e+00 1.64133981e-01 3.54966879e-01
-5.20896316e-01 -1.05494760e-01 4.06282425e-01 -9.48617831e-02
1.37213349e-01 -9.60358083e-02 -1.11460470e-01 -1.12251508e+00
8.33400264e-02 -8.93253565e-01 2.12503582e-01 -8.57252300e-01
-1.58084691e+00 5.72816849e-01 -1.13250233e-01 -1.42078233e+00
-2.86660194e-01 -3.63349527e-01 -6.61352515e-01 1.25348616e+00
-1.56850374e+00 -2.02552056e+00 -4.27013099e-01 8.56583118e-01
3.13469291e-01 -7.52878487e-01 7.82986879e-01 6.76568747e-01
-3.22356105e-01 1.08706379e+00 -8.65796357e-02 3.19097728e-01
9.55472112e-01 -8.11450481e-01 2.94416636e-01 7.52712429e-01
3.95963080e-02 7.58301079e-01 2.75252849e-01 -6.94645107e-01
-1.80685270e+00 -6.71045065e-01 9.58750308e-01 -4.09019321e-01
2.65958190e-01 -4.38706398e-01 -6.49228036e-01 4.83804554e-01
5.34626842e-01 1.99606508e-01 8.02337646e-01 -2.56617785e-01
-7.39588618e-01 -3.53077531e-01 -1.41034591e+00 5.89531958e-01
9.54292357e-01 -1.14194942e+00 -4.91760969e-01 1.35024831e-01
2.73898900e-01 -2.48710409e-01 -9.66371775e-01 3.41467589e-01
1.23369551e+00 -8.98569047e-01 9.89670634e-01 -4.90006685e-01
5.19968033e-01 -2.70479202e-01 -4.00265276e-01 -7.71434009e-01
3.07311565e-02 -8.47124279e-01 -2.05886178e-02 1.74472082e+00
-2.63055321e-02 -5.06410003e-01 1.04367495e+00 7.37985015e-01
4.36547011e-01 -4.43205237e-01 -7.56829977e-01 -4.12066042e-01
-4.11833636e-02 2.30845660e-01 1.13235259e+00 1.09932220e+00
-2.82803357e-01 1.82194695e-01 -7.93794453e-01 1.37292236e-01
7.28464186e-01 1.91641346e-01 1.06741118e+00 -1.00290489e+00
-1.84249252e-01 -3.03176492e-01 -2.48148188e-01 -9.87933218e-01
4.80037689e-01 -7.51721799e-01 -1.96754530e-01 -1.13554776e+00
2.06724852e-01 -6.14829421e-01 -2.17486605e-01 5.31179190e-01
3.63666788e-02 5.99512339e-01 5.08529067e-01 4.19348359e-01
-5.67861972e-03 5.96923053e-01 1.19969141e+00 -9.56047028e-02
2.04896219e-02 -1.49097577e-01 -3.89096409e-01 4.25038755e-01
4.92889434e-01 -3.75008076e-01 -5.33369303e-01 -5.46379685e-01
3.76782287e-03 4.52945173e-01 4.09903944e-01 -7.52967834e-01
5.45183122e-01 -7.05339015e-02 7.09741950e-01 -5.92145741e-01
6.17510080e-01 -1.11733580e+00 4.39616919e-01 8.67912844e-02
-4.45638895e-01 1.57915920e-01 -9.51090157e-02 4.45804089e-01
-3.12108845e-01 2.79063266e-03 8.88100564e-01 -1.43845767e-01
-3.45793873e-01 5.93871176e-01 2.58066565e-01 -4.96082962e-01
6.56863391e-01 -4.00285959e-01 -2.97590882e-01 -3.75609815e-01
-3.76944959e-01 -1.42136931e-01 6.11348748e-01 6.31497204e-01
7.06861138e-01 -2.00005817e+00 -5.45788169e-01 7.51734078e-01
1.78756282e-01 -6.61227107e-01 4.30390000e-01 6.39443636e-01
-4.64086145e-01 2.77306110e-01 -6.69039309e-01 -4.52574164e-01
-1.38365817e+00 4.16101009e-01 3.70550454e-01 6.42765686e-02
-3.66799206e-01 7.75690615e-01 4.89094220e-02 -6.24410331e-01
3.94747853e-01 4.27928925e-01 8.20576102e-02 2.50113104e-02
4.67382342e-01 1.56061739e-01 -2.73277462e-01 -9.19759154e-01
-3.10638458e-01 7.61970222e-01 1.74576268e-01 -3.53799641e-01
1.21456051e+00 -1.69984609e-01 -4.13745582e-01 2.49879416e-02
1.60941374e+00 -1.44397214e-01 -1.35054684e+00 -4.32061225e-01
-3.59493494e-01 -9.54971373e-01 -2.37637401e-01 -5.26129663e-01
-1.18907619e+00 1.05462193e+00 1.04821122e+00 -5.59500158e-02
1.05465925e+00 -3.56142968e-01 8.74872744e-01 1.80659175e-01
-6.69428036e-02 -9.94556427e-01 2.26713568e-01 7.33078942e-02
8.77446234e-01 -1.46068966e+00 1.18716797e-02 -2.10011005e-01
-3.43735814e-01 1.33636749e+00 5.89396060e-01 -1.59363523e-01
6.68103993e-01 -1.48397267e-01 1.78107128e-01 -1.53400511e-01
-4.49560553e-01 1.06087230e-01 5.75529337e-01 4.26364630e-01
4.39815968e-01 -2.44376749e-01 -6.51176870e-02 -2.12861300e-02
2.04229236e-01 2.47028738e-01 6.45196214e-02 9.84627843e-01
1.24390900e-01 -1.79028499e+00 -4.87760127e-01 4.13032062e-02
-3.74632031e-01 1.28150523e-01 -3.51753771e-01 5.26980340e-01
2.66060859e-01 7.77105510e-01 -7.44278915e-03 -3.28543544e-01
2.80323643e-02 2.38627926e-01 9.68978107e-01 -1.27010807e-01
-4.19340819e-01 1.46995082e-01 -2.65956998e-01 -5.23751736e-01
-7.66780376e-01 -3.75205308e-01 -5.16634703e-01 -5.58032453e-01
-1.25441268e-01 -2.27007791e-01 9.10568595e-01 1.00303340e+00
4.18219149e-01 -6.53211176e-02 9.33808565e-01 -1.01258683e+00
-5.86891532e-01 -7.83604681e-01 -3.96526635e-01 5.22631764e-01
4.31026459e-01 -5.62257588e-01 3.59247215e-02 2.34336555e-01] | [12.456687927246094, -0.00370511831715703] |
07bbf193-24b6-4922-bd95-8534e1c529e9 | end-to-end-dense-video-captioning-with | 2108.07781 | null | https://arxiv.org/abs/2108.07781v2 | https://arxiv.org/pdf/2108.07781v2.pdf | End-to-End Dense Video Captioning with Parallel Decoding | Dense video captioning aims to generate multiple associated captions with their temporal locations from the video. Previous methods follow a sophisticated "localize-then-describe" scheme, which heavily relies on numerous hand-crafted components. In this paper, we proposed a simple yet effective framework for end-to-end dense video captioning with parallel decoding (PDVC), by formulating the dense caption generation as a set prediction task. In practice, through stacking a newly proposed event counter on the top of a transformer decoder, the PDVC precisely segments the video into a number of event pieces under the holistic understanding of the video content, which effectively increases the coherence and readability of predicted captions. Compared with prior arts, the PDVC has several appealing advantages: (1) Without relying on heuristic non-maximum suppression or a recurrent event sequence selection network to remove redundancy, PDVC directly produces an event set with an appropriate size; (2) In contrast to adopting the two-stage scheme, we feed the enhanced representations of event queries into the localization head and caption head in parallel, making these two sub-tasks deeply interrelated and mutually promoted through the optimization; (3) Without bells and whistles, extensive experiments on ActivityNet Captions and YouCook2 show that PDVC is capable of producing high-quality captioning results, surpassing the state-of-the-art two-stage methods when its localization accuracy is on par with them. Code is available at https://github.com/ttengwang/PDVC. | ['Ping Luo', 'Ran Cheng', 'Feng Zheng', 'Zhichao Lu', 'Ruimao Zhang', 'Teng Wang'] | 2021-08-17 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Wang_End-to-End_Dense_Video_Captioning_With_Parallel_Decoding_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Wang_End-to-End_Dense_Video_Captioning_With_Parallel_Decoding_ICCV_2021_paper.pdf | iccv-2021-1 | ['dense-video-captioning'] | ['computer-vision'] | [ 3.09606463e-01 1.09460235e-01 -1.88076630e-01 -2.70199150e-01
-1.09247959e+00 -4.75768536e-01 6.14538133e-01 -7.82444030e-02
-1.58987433e-01 8.34625125e-01 6.61664307e-01 -9.41602290e-02
3.46195847e-01 -3.26682895e-01 -1.11316288e+00 -5.13553381e-01
1.88021332e-01 4.83042926e-01 2.92043984e-01 -2.58394033e-02
6.52216747e-03 -1.03832096e-01 -1.57470512e+00 5.20564437e-01
6.97862506e-01 1.03793025e+00 7.72680104e-01 6.13729715e-01
-1.22187689e-01 1.08230305e+00 -3.32048386e-01 -6.08240306e-01
-9.50712115e-02 -7.93898463e-01 -5.46454072e-01 1.38705537e-01
1.92152962e-01 -3.82345259e-01 -5.80404699e-01 8.13047826e-01
4.85078692e-01 -2.96615474e-02 2.92467505e-01 -1.24955583e+00
-7.33253002e-01 8.47909391e-01 -6.06192946e-01 2.81203032e-01
5.58375895e-01 1.91365093e-01 1.16164529e+00 -1.01103568e+00
4.83670473e-01 8.55069697e-01 5.16098142e-01 6.21824861e-01
-9.45846438e-01 -7.78520525e-01 3.18939567e-01 4.69796956e-01
-1.46530485e+00 -6.21533036e-01 6.49052918e-01 -3.33966643e-01
8.15915525e-01 2.97555774e-01 6.39970422e-01 1.36815751e+00
-1.45438671e-01 9.74588931e-01 5.01232266e-01 -1.34350419e-01
1.08247302e-01 -9.90928411e-02 -9.01788771e-02 4.93884176e-01
-4.24210280e-02 -6.33711889e-02 -7.30322242e-01 -1.30049109e-01
7.52548993e-01 -4.72119451e-02 -6.61154985e-01 -6.34456649e-02
-1.57996261e+00 6.09548688e-01 2.53415346e-01 2.50303119e-01
-7.11790323e-01 4.88664210e-01 4.90368783e-01 -2.95879662e-01
2.96353281e-01 7.45094568e-02 -2.70150721e-01 -4.37589824e-01
-1.20846832e+00 1.30909488e-01 5.63208282e-01 1.27601731e+00
6.70671761e-01 -1.60237074e-01 -6.32611573e-01 6.27151430e-01
3.64558071e-01 4.61263835e-01 4.72610146e-01 -6.66896403e-01
8.66373360e-01 1.27751678e-01 3.12356234e-01 -8.39658737e-01
-5.18973581e-02 -4.84543502e-01 -7.57719755e-01 -5.56035042e-01
7.12154880e-02 -2.63663322e-01 -9.93367612e-01 1.90889287e+00
8.67038220e-02 8.10859919e-01 -1.61113515e-02 1.24147475e+00
7.04307675e-01 1.23781025e+00 3.48792762e-01 -3.67454708e-01
1.64399600e+00 -1.31095636e+00 -9.62956905e-01 -5.78772545e-01
2.09544539e-01 -6.93584025e-01 9.14521873e-01 1.30709037e-01
-1.15891111e+00 -6.03429258e-01 -9.18087244e-01 -1.44254193e-01
2.17730388e-01 2.58226454e-01 6.21548116e-01 1.18551046e-01
-1.12192869e+00 1.47281989e-01 -7.32328355e-01 -2.20585957e-01
3.97616863e-01 1.53595015e-01 -2.13306487e-01 -4.46223468e-02
-1.23391640e+00 6.08516395e-01 6.50893927e-01 1.65706173e-01
-1.07837045e+00 -5.98425567e-01 -9.45003927e-01 3.02027792e-01
4.10584480e-01 -8.26671064e-01 1.50268948e+00 -1.13812947e+00
-1.46677244e+00 4.94105130e-01 -6.72988355e-01 -6.02499902e-01
3.84652764e-01 -4.16199774e-01 -3.94563824e-01 4.50297743e-01
2.81259716e-01 1.01742804e+00 7.70263553e-01 -1.09410453e+00
-7.05778956e-01 2.21432358e-01 -6.30461499e-02 4.48770225e-01
-1.28794670e-01 2.67936885e-01 -1.18084550e+00 -7.86164463e-01
-1.10914141e-01 -8.49493027e-01 -1.22819141e-01 -1.71166852e-01
-4.39698011e-01 4.80925106e-03 7.01813996e-01 -8.78889441e-01
1.42663217e+00 -2.22416997e+00 1.98843151e-01 -2.30171010e-01
1.60195455e-01 2.27076456e-01 -1.45116106e-01 5.21758795e-01
-1.50578380e-01 -2.49082246e-03 -3.14511597e-01 -6.20319188e-01
-3.50927711e-02 -6.38409983e-03 -5.79183996e-01 2.61973739e-01
4.17887717e-01 1.06054711e+00 -1.16708398e+00 -6.27999246e-01
1.78625643e-01 6.20412111e-01 -5.21492183e-01 4.57109183e-01
-5.49546480e-01 3.37784648e-01 -4.40394938e-01 5.08034647e-01
3.97750705e-01 -8.05675626e-01 2.94301003e-01 -3.61160994e-01
8.67355429e-03 4.34889853e-01 -9.05179322e-01 1.87532568e+00
-3.11157495e-01 6.66816771e-01 -1.37473151e-01 -7.23222375e-01
6.32187963e-01 8.42769325e-01 3.34191084e-01 -6.11264348e-01
6.37942031e-02 1.18428618e-01 -5.45374215e-01 -4.19772565e-01
5.76303899e-01 3.62995937e-02 -2.35575050e-01 2.47287884e-01
2.39978120e-01 5.18621385e-01 3.04502785e-01 3.65685612e-01
1.06117296e+00 3.90807867e-01 2.65556276e-01 1.58663958e-01
3.90065998e-01 -1.31493500e-02 6.04513705e-01 5.59353292e-01
-1.22837827e-01 1.07158780e+00 5.41548491e-01 -3.53393480e-02
-1.08647418e+00 -1.07791495e+00 2.44590506e-01 9.14455771e-01
3.62079084e-01 -5.93351781e-01 -8.19749355e-01 -5.11365414e-01
-4.53450322e-01 7.92095661e-01 -4.29732859e-01 3.39595340e-02
-6.66492105e-01 -6.03625000e-01 5.29610693e-01 5.86959004e-01
5.68331361e-01 -1.09942734e+00 -3.19060147e-01 4.83919948e-01
-9.45572019e-01 -1.49789906e+00 -1.00679994e+00 3.82277109e-02
-5.03934681e-01 -7.36274421e-01 -9.78527427e-01 -9.71943378e-01
5.48838198e-01 3.59902143e-01 1.18481827e+00 -9.62690786e-02
3.57622027e-01 7.51805604e-02 -5.77880979e-01 -6.55345842e-02
-3.01069319e-01 -1.02614509e-02 -2.43729547e-01 2.50490218e-01
2.73507565e-01 -6.62140667e-01 -5.80409229e-01 2.46965989e-01
-8.11841011e-01 8.17657948e-01 8.06339979e-01 7.17658997e-01
7.34239578e-01 -4.86324102e-01 6.56212270e-01 -4.72186744e-01
1.76474586e-01 -8.35970342e-01 -4.77044880e-01 1.90049484e-01
-2.09236756e-01 -4.38391529e-02 7.55808473e-01 -4.61603075e-01
-1.00745749e+00 3.57419044e-01 -1.99368477e-01 -7.29742825e-01
-3.43312547e-02 3.59202415e-01 -2.76985794e-01 5.27333260e-01
2.57436752e-01 6.44273102e-01 -9.62837189e-02 -3.96512449e-01
4.50118512e-01 6.24268234e-01 8.45463514e-01 -3.23187739e-01
6.34287655e-01 2.23208383e-01 -5.78750968e-01 -3.71095717e-01
-8.32870245e-01 -4.21809226e-01 -1.43190801e-01 -3.21276993e-01
1.08175874e+00 -1.42231262e+00 -6.18146598e-01 3.46638829e-01
-1.52318513e+00 -1.66709855e-01 -1.05685800e-01 4.94993299e-01
-6.83331251e-01 4.35989439e-01 -8.41993868e-01 -5.79023957e-01
-3.88251066e-01 -1.10538554e+00 1.26437175e+00 2.95827717e-01
-1.12795010e-01 -5.52585900e-01 -8.96247104e-02 4.80161041e-01
2.20201880e-01 7.26542398e-02 5.10240138e-01 -5.39687574e-01
-8.51739883e-01 1.41137606e-02 -4.17292386e-01 1.96511731e-01
-1.50260121e-01 -3.73751819e-01 -9.79481757e-01 -3.12559903e-02
-2.16474801e-01 -1.75347000e-01 6.90801978e-01 3.20273787e-01
1.15890682e+00 -4.67545688e-01 -3.49055618e-01 5.54884493e-01
1.45350409e+00 2.72885680e-01 8.88679504e-01 1.23891696e-01
7.73350060e-01 2.55608112e-01 5.62987626e-01 5.81845164e-01
5.81196189e-01 8.66971910e-01 5.81377447e-01 -9.05902963e-03
-2.95244128e-01 -7.25143671e-01 6.54409170e-01 8.31146777e-01
1.65681571e-01 -8.34210873e-01 -6.79238796e-01 7.24282205e-01
-2.24680471e+00 -1.14087200e+00 -2.76191556e-03 1.89171290e+00
9.28630471e-01 -1.94532387e-02 1.96273047e-02 -3.07823300e-01
1.23928905e+00 4.06912237e-01 -3.76641572e-01 -3.06369923e-02
-1.32008687e-01 -4.04333435e-02 3.70734066e-01 2.48395532e-01
-9.95348394e-01 9.77102041e-01 5.36081076e+00 9.18461144e-01
-9.34642732e-01 3.23340416e-01 6.73495531e-01 -2.55481243e-01
-3.56891900e-01 7.08698481e-02 -9.77225423e-01 9.67036128e-01
1.30512512e+00 -1.95222229e-01 3.48430514e-01 6.16267443e-01
4.82892603e-01 9.42065045e-02 -1.16794944e+00 1.20632410e+00
1.35356382e-01 -1.66343844e+00 1.64345115e-01 -2.28893921e-01
4.56200421e-01 9.22937021e-02 -2.95938611e-01 3.46757829e-01
1.20388819e-02 -8.24559629e-01 1.27817595e+00 3.67994875e-01
8.67579818e-01 -4.96215343e-01 6.52714431e-01 1.88424394e-01
-1.34218192e+00 2.85303649e-02 -6.67883977e-02 6.05181642e-02
9.00484622e-01 6.11834824e-01 -7.49214947e-01 5.96608698e-01
5.33422947e-01 7.28478253e-01 -1.35223582e-01 1.14184988e+00
-4.25572872e-01 9.13979888e-01 -3.22917998e-01 5.15423752e-02
4.37529117e-01 1.07401006e-01 5.39808512e-01 1.35150182e+00
5.17182708e-01 1.40408456e-01 6.36706650e-02 8.26295853e-01
-3.24167341e-01 -6.40189424e-02 -1.32305697e-01 -1.45136923e-01
6.41468883e-01 1.06623888e+00 -5.76002777e-01 -5.25533795e-01
-4.81679440e-01 1.32673132e+00 1.16707027e-01 3.46490324e-01
-1.56214690e+00 -2.09217817e-01 5.35901845e-01 1.04737431e-01
7.56544650e-01 8.82021263e-02 -3.32080126e-02 -1.32902300e+00
2.97168523e-01 -8.34172845e-01 2.14130908e-01 -1.20835567e+00
-9.43167508e-01 8.79910409e-01 -5.72368912e-02 -1.40537977e+00
-4.13036078e-01 -2.53639463e-02 -4.98871356e-01 7.69996345e-01
-1.49765456e+00 -1.14324617e+00 -3.53014022e-01 5.01217365e-01
8.48120093e-01 2.34246120e-01 6.69781804e-01 6.64833724e-01
-6.96051002e-01 4.36642349e-01 -1.76347092e-01 1.03143923e-01
6.54395163e-01 -8.40674579e-01 4.93652731e-01 1.21995902e+00
1.36994645e-01 3.93191539e-02 7.75505006e-01 -6.67587519e-01
-1.21937311e+00 -1.30516350e+00 1.26170301e+00 -1.44793138e-01
5.37620723e-01 -5.21855772e-01 -8.05092990e-01 8.10480773e-01
4.47559148e-01 -4.86194268e-02 4.30692047e-01 -3.38309854e-01
-2.39507303e-01 1.69209223e-02 -6.36227846e-01 6.55579269e-01
1.13481104e+00 -4.33640599e-01 -6.14583492e-01 5.52275777e-01
1.11283004e+00 -4.81753856e-01 -4.27065343e-01 1.92662746e-01
2.52034068e-01 -8.40092778e-01 7.90164888e-01 -2.40613937e-01
7.74792552e-01 -5.81943870e-01 -1.35328203e-01 -8.45498145e-01
-4.43546802e-01 -9.55861509e-01 -5.02760291e-01 1.45764494e+00
5.26963055e-01 -1.93439096e-01 6.71689093e-01 2.83371359e-01
-5.79180717e-01 -7.31626511e-01 -7.93568671e-01 -5.10870755e-01
-7.29907632e-01 -5.14823973e-01 5.05944967e-01 5.66363037e-01
-9.25777107e-02 5.85815728e-01 -8.52058828e-01 5.32636344e-01
4.05862898e-01 1.03750773e-01 3.79906654e-01 -6.15354657e-01
-5.10271013e-01 -1.15433045e-01 -1.47573411e-01 -1.57197547e+00
8.32881406e-02 -6.97509468e-01 5.11643589e-01 -1.67631912e+00
3.54612559e-01 -1.66836753e-01 -2.64883786e-01 6.81528986e-01
-3.09089631e-01 3.75941247e-01 3.49375963e-01 4.30283308e-01
-1.01248729e+00 6.95410073e-01 1.00217700e+00 1.66662902e-01
-1.52888447e-01 -2.54343420e-01 -9.06224251e-01 4.90576535e-01
5.90332925e-01 -4.42578495e-01 -5.33114016e-01 -6.20802462e-01
9.32791978e-02 4.82540131e-01 4.75720257e-01 -1.17517734e+00
3.16392392e-01 2.13083276e-03 1.59200892e-01 -6.88368142e-01
5.90923667e-01 -6.11313403e-01 5.62775373e-01 9.40337628e-02
-2.71310657e-01 1.72308654e-01 3.14357638e-01 7.47621953e-01
-4.42734569e-01 -8.34146142e-02 5.35144210e-01 -4.66385111e-02
-1.07693219e+00 4.19785827e-01 -4.45841581e-01 3.38347144e-02
1.21564496e+00 -1.06548006e-02 -2.78847903e-01 -6.14102304e-01
-4.57271367e-01 3.78668815e-01 3.64220738e-01 5.39460540e-01
6.16000414e-01 -1.28769839e+00 -8.31508815e-01 -8.33762586e-02
1.35533549e-02 7.54764630e-03 4.68044817e-01 8.28031003e-01
-3.88979644e-01 4.81227458e-01 8.15671459e-02 -5.70863724e-01
-9.80167687e-01 6.17233813e-01 5.42727159e-03 -3.62740219e-01
-6.79226398e-01 9.02047038e-01 4.63897228e-01 3.13663334e-01
2.52596945e-01 -2.14349255e-02 -1.02058344e-01 4.09468636e-02
7.57700682e-01 -9.60331485e-02 -1.06159225e-01 -6.86867177e-01
-3.96170944e-01 2.08599761e-01 -1.95193484e-01 -1.30695805e-01
1.25761473e+00 -3.32663745e-01 1.77072614e-01 2.41336729e-02
1.09925747e+00 -1.49138674e-01 -1.58964920e+00 -7.61968419e-02
-2.03059182e-01 -1.75948024e-01 -3.65028046e-02 -8.16814721e-01
-1.01997316e+00 5.30246556e-01 2.18381166e-01 -1.27290323e-01
1.36368954e+00 2.07866192e-01 1.27651870e+00 -1.26289546e-01
2.97231764e-01 -6.76739693e-01 5.67285083e-02 3.68853301e-01
8.21883261e-01 -9.06251490e-01 -3.47751528e-01 -6.26605511e-01
-9.76449609e-01 7.83568680e-01 5.92484891e-01 6.13809787e-02
1.60700783e-01 1.93852663e-01 -1.44843951e-01 5.44359945e-02
-1.02313197e+00 -2.08333686e-01 1.23111382e-01 2.90780962e-01
2.46525288e-01 -1.63344264e-01 -3.13218653e-01 9.35873449e-01
3.71253490e-02 3.22604835e-01 4.78023499e-01 6.30892634e-01
-3.29355747e-01 -8.99550378e-01 -2.66680300e-01 2.04728067e-01
-6.00307882e-01 -5.17867923e-01 1.29975840e-01 5.49176991e-01
1.29672736e-01 9.03131008e-01 2.63023078e-01 -4.52922553e-01
1.42022625e-01 8.13225731e-02 1.55359711e-02 -5.62991321e-01
-4.84036475e-01 1.50891930e-01 2.80073881e-01 -7.82253206e-01
-3.34008306e-01 -5.41194737e-01 -1.34008098e+00 -5.36253005e-02
-2.25436211e-01 4.21321750e-01 4.25683022e-01 8.73458326e-01
8.65747988e-01 4.94682699e-01 4.64283168e-01 -1.08831716e+00
-2.72458017e-01 -8.84390116e-01 -1.74471185e-01 3.27713758e-01
3.21219534e-01 -4.04150337e-01 -3.27461630e-01 4.20615226e-01] | [10.404016494750977, 0.6768828630447388] |
a86a119e-02b8-457e-8d81-758797502e2c | effects-of-real-life-traffic-sign-alteration | 2305.05499 | null | https://arxiv.org/abs/2305.05499v1 | https://arxiv.org/pdf/2305.05499v1.pdf | Effects of Real-Life Traffic Sign Alteration on YOLOv7- an Object Recognition Model | The advancement of Image Processing has led to the widespread use of Object Recognition (OR) models in various applications, such as airport security and mail sorting. These models have become essential in signifying the capabilities of AI and supporting vital services like national postal operations. However, the performance of OR models can be impeded by real-life scenarios, such as traffic sign alteration. Therefore, this research investigates the effects of altered traffic signs on the accuracy and performance of object recognition models. To this end, a publicly available dataset was used to create different types of traffic sign alterations, including changes to size, shape, color, visibility, and angles. The impact of these alterations on the YOLOv7 (You Only Look Once) model's detection and classification abilities were analyzed. It reveals that the accuracy of object detection models decreases significantly when exposed to modified traffic signs under unlikely conditions. This study highlights the significance of enhancing the robustness of object detection models in real-life scenarios and the need for further investigation in this area to improve their accuracy and reliability. | ['Edmon Begoli', 'Edward Michaud', 'Md Saif Hassan Onim', 'Shahinul Hoque', 'Farhin Farhad Riya'] | 2023-05-09 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [ 3.36448461e-01 -4.99590963e-01 2.47167766e-01 -2.37818837e-01
6.29831925e-02 -5.50658584e-01 5.19466639e-01 -1.05539791e-01
-4.78846639e-01 3.45675856e-01 -4.47655797e-01 -6.52187645e-01
-2.74091512e-01 -6.49314582e-01 -5.88340819e-01 -5.35188735e-01
4.62637432e-02 -1.58221796e-01 6.50317490e-01 -2.16153979e-01
7.01583564e-01 1.15279877e+00 -2.28091931e+00 1.71009257e-01
5.11790812e-01 1.01254582e+00 6.44222423e-02 7.92426467e-01
-1.00816645e-01 2.52898008e-01 -9.29352820e-01 -3.42289895e-01
4.52910870e-01 -1.56359121e-01 9.73513629e-03 2.98064739e-01
6.08727217e-01 -4.28855896e-01 -2.51133382e-01 9.64556694e-01
5.08430719e-01 2.11004958e-01 4.15464729e-01 -1.39542723e+00
-7.32555151e-01 -1.99954093e-01 -2.78998703e-01 4.96735275e-01
3.79912704e-01 5.30445874e-01 3.44114870e-01 -5.39521754e-01
3.37898493e-01 1.08973265e+00 4.30187017e-01 6.33012429e-02
-1.05851030e+00 -9.66816425e-01 9.43068415e-03 6.17804050e-01
-1.12855613e+00 -4.01773453e-01 6.14258051e-01 -4.03188646e-01
7.29223490e-01 3.85697067e-01 6.71301365e-01 6.79813027e-01
2.91326016e-01 2.77365357e-01 1.33781409e+00 -7.06966639e-01
-3.18480507e-02 4.24269587e-01 2.65456319e-01 5.91902912e-01
1.05980694e+00 4.90307003e-01 -3.00289094e-01 2.61781901e-01
5.44229567e-01 -2.11612806e-01 5.93474358e-02 -1.15045197e-01
-8.99083972e-01 3.11425954e-01 2.50033587e-01 3.87389272e-01
-2.48632431e-01 -5.34695275e-02 1.33555010e-01 3.27945948e-01
-8.03764611e-02 6.44240558e-01 -4.16182950e-02 -2.98465848e-01
-4.68374580e-01 1.46692730e-02 3.48363489e-01 5.28360784e-01
2.63868570e-01 2.47124061e-01 -8.46019164e-02 6.69877529e-01
1.63955867e-01 8.32311809e-01 2.62642056e-01 -6.64433479e-01
2.47486442e-01 7.80479193e-01 2.81151205e-01 -1.51864278e+00
-4.90865022e-01 -3.87480468e-01 -1.92650273e-01 6.63557351e-01
6.62576616e-01 2.73416340e-01 -1.08257997e+00 9.80112731e-01
3.67524773e-01 3.85053605e-02 -3.40383768e-01 9.48545754e-01
7.96627045e-01 4.02557880e-01 1.92930683e-01 1.92571357e-01
1.32769918e+00 -3.43065947e-01 -8.30859363e-01 -9.72127393e-02
5.99648952e-01 -1.11072242e+00 1.06727719e+00 4.31689858e-01
-4.98156697e-01 -8.49660099e-01 -1.06964815e+00 3.85138363e-01
-7.93372512e-01 3.66842628e-01 5.93553603e-01 1.35828757e+00
-4.45508122e-01 1.47994757e-01 -4.26868647e-01 -5.05809307e-01
1.32281527e-01 3.02828252e-01 -3.70453566e-01 -2.97494411e-01
-9.75902200e-01 1.25411093e+00 1.34736657e-01 6.55432522e-01
-2.64173120e-01 -3.71514291e-01 -6.48670912e-01 -2.19603732e-01
1.76720068e-01 -5.75923659e-02 7.79505312e-01 -8.97845447e-01
-1.12440419e+00 7.53088951e-01 -6.05144985e-02 -2.83742726e-01
6.22545719e-01 -1.18809305e-01 -8.96673977e-01 -3.62397507e-02
-1.80409923e-01 2.87182540e-01 7.26543009e-01 -1.35014415e+00
-6.14437044e-01 -2.27029160e-01 5.82277328e-02 -1.37598574e-01
3.88938002e-02 3.43953669e-01 -1.62929744e-01 -4.20116216e-01
1.15179196e-01 -1.00001299e+00 1.70126066e-01 1.88194096e-01
-1.20178409e-01 1.74403459e-01 1.19077373e+00 -8.56587172e-01
1.15348947e+00 -2.03287554e+00 -8.82966101e-01 5.60566187e-01
-2.91723013e-01 8.62386167e-01 -2.23032176e-01 3.35406244e-01
-4.21423018e-02 7.39154443e-02 6.88491166e-02 3.76766026e-01
-8.41518492e-02 1.33121178e-01 -9.65601578e-02 4.79877979e-01
1.82873145e-01 5.80876827e-01 -4.36287045e-01 -2.10147172e-01
7.07712829e-01 3.68633330e-01 -1.30006626e-01 -5.61098531e-02
3.58794332e-01 2.65303373e-01 -1.41152695e-01 9.07779634e-01
7.15190947e-01 4.52972561e-01 -2.80155987e-01 -2.92830348e-01
-3.11031669e-01 -4.32114229e-02 -1.16492784e+00 3.32463741e-01
-2.28699878e-01 1.23240602e+00 -2.45424092e-01 -8.39989960e-01
1.10573947e+00 1.05523497e-01 1.80212423e-01 -1.33346701e+00
2.54287571e-01 2.11900219e-01 5.04614949e-01 -9.15875971e-01
7.55420744e-01 1.21112689e-01 3.51311743e-01 1.56607255e-01
-8.24754238e-01 -6.10324219e-02 4.18950379e-01 -1.75025105e-01
5.51222980e-01 -2.33816922e-01 -7.42284581e-02 2.18173444e-01
5.29815555e-01 -1.33616114e-02 1.04572572e-01 7.92042553e-01
-4.59659904e-01 2.65870839e-01 2.55137116e-01 -3.83525550e-01
-9.39870298e-01 -7.60293543e-01 -3.54618102e-01 6.55660808e-01
2.53215313e-01 3.09262544e-01 -3.93291533e-01 -2.92913675e-01
2.40676299e-01 9.96264338e-01 -5.14441073e-01 -5.48261404e-01
-4.54061329e-01 -7.78064430e-01 5.93969703e-01 4.17248070e-01
7.15226531e-01 -1.15564156e+00 -9.80769336e-01 -2.66132616e-02
1.37233093e-01 -1.25111103e+00 1.56392768e-01 -4.24607605e-01
-7.01487422e-01 -1.47117126e+00 -5.07574320e-01 -4.81278658e-01
1.00277758e+00 5.77415705e-01 2.99731821e-01 5.57788432e-01
-6.76263392e-01 6.28242612e-01 -5.22159994e-01 -5.37982106e-01
-4.85261768e-01 -3.26878011e-01 -4.90541980e-02 3.34959269e-01
2.84668773e-01 1.56929195e-01 -3.32819790e-01 9.66401577e-01
-1.04986489e+00 -2.47642234e-01 5.99737167e-01 3.37661326e-01
9.68695432e-02 4.28959936e-01 4.38968003e-01 -5.58065832e-01
7.89145052e-01 7.63774887e-02 -8.13861191e-01 3.87002259e-01
-7.13011324e-01 -3.42902720e-01 1.39133096e-01 -4.68630999e-01
-1.14690530e+00 -4.06643629e-01 2.87825853e-01 4.40810174e-02
-5.48126340e-01 2.07174763e-01 -1.67420372e-01 -5.99673212e-01
4.67231035e-01 2.44157091e-02 1.45831794e-01 -2.94991970e-01
-1.75462142e-01 8.39821458e-01 2.99381018e-01 -2.71984637e-01
1.04311848e+00 3.32832754e-01 4.24864143e-01 -1.31041586e+00
-2.48198628e-01 -3.81777614e-01 -5.20345867e-01 -8.84828091e-01
5.36713004e-01 -1.68123096e-01 -6.61863863e-01 8.10587525e-01
-8.95193636e-01 -3.96302454e-02 2.07948327e-01 7.44744956e-01
1.09336041e-01 5.35774469e-01 -1.79229509e-02 -1.08678925e+00
2.88153678e-01 -1.10273468e+00 5.84388137e-01 4.45274919e-01
-7.78985322e-02 -5.04234970e-01 -4.23462927e-01 8.14111769e-01
5.69003224e-01 1.99918345e-01 8.71583164e-01 -2.19360530e-01
-6.49819672e-01 -8.61613870e-01 -4.05125588e-01 4.34786588e-01
1.30428419e-01 5.24990141e-01 -7.70482540e-01 4.69572060e-02
-4.64134067e-01 1.83909923e-01 5.75587094e-01 1.50295869e-01
8.77386570e-01 -6.47478411e-03 -6.29822984e-02 2.84727424e-01
9.80989397e-01 9.32809234e-01 9.51934099e-01 7.21821904e-01
2.88495868e-01 9.02605653e-01 9.17838931e-01 5.50740724e-03
-1.62919492e-01 7.75272965e-01 4.78596151e-01 -1.31549329e-01
-3.79855275e-01 8.29648599e-02 2.25090981e-01 1.71101198e-01
-2.92396605e-01 -1.66355416e-01 -8.92266452e-01 2.52487689e-01
-1.28816140e+00 -1.17116678e+00 -4.80605721e-01 2.27542281e+00
-9.45167392e-02 2.66249180e-01 9.47504677e-03 6.64309978e-01
8.56937170e-01 -1.72499776e-01 -4.00600612e-01 -6.35682940e-01
-1.82378411e-01 -2.33370438e-01 6.25396609e-01 1.64703339e-01
-7.42138982e-01 6.15507960e-01 6.70195389e+00 3.59597176e-01
-1.32189870e+00 -4.71930802e-01 5.50401986e-01 1.44356623e-01
3.07775894e-03 -2.81524092e-01 -6.75945520e-01 6.23884678e-01
5.50745904e-01 2.21762151e-01 2.54853576e-01 5.89767039e-01
7.05391407e-01 -6.55182898e-01 -3.54391336e-01 6.11307025e-01
3.53832066e-01 -7.62328804e-01 8.34602714e-02 -2.48583034e-02
2.50801563e-01 -7.22970307e-01 3.07563603e-01 2.76157588e-01
-2.93195963e-01 -8.04601967e-01 6.98637962e-01 6.15273833e-01
6.30500376e-01 -4.91733134e-01 1.02340400e+00 -9.73953232e-02
-8.94196391e-01 -1.23068109e-01 -1.14128247e-01 -2.04683155e-01
2.27594882e-01 1.33627519e-01 -1.06380928e+00 6.00225069e-02
5.49265683e-01 9.83111002e-03 -1.15222001e+00 1.57840478e+00
-1.63955793e-01 8.13274622e-01 -4.78658199e-01 -2.52274722e-01
-3.40344384e-02 -2.46373668e-01 5.45383513e-01 9.67177570e-01
3.26596797e-01 -5.52205974e-03 -2.94448048e-01 4.33348089e-01
4.83920187e-01 -9.86954644e-02 -6.03292286e-01 -3.03308666e-01
5.54576516e-01 7.47948706e-01 -1.07658815e+00 -1.56138302e-03
-3.59734267e-01 5.44177532e-01 -4.89520401e-01 6.21255398e-01
-9.70220685e-01 -4.98273700e-01 4.42642778e-01 6.45015240e-01
2.16455176e-01 -4.85795140e-01 -5.08322418e-01 -4.13173020e-01
1.50011882e-01 -7.08914220e-01 1.02469936e-01 -1.02309096e+00
-7.09806859e-01 1.57223195e-01 2.93136775e-01 -1.19762969e+00
3.68399918e-01 -1.02803969e+00 -6.84335113e-01 4.62253958e-01
-1.32337821e+00 -1.16892529e+00 -5.03241837e-01 2.03152567e-01
3.52005005e-01 -2.96613216e-01 1.97165847e-01 4.49906260e-01
-7.46722221e-01 6.01931572e-01 -9.01138410e-02 2.31744453e-01
5.37338734e-01 -3.59868377e-01 6.58014938e-02 1.02950668e+00
-5.22266813e-02 4.14433599e-01 7.73009419e-01 -6.67268813e-01
-1.12180686e+00 -6.14044189e-01 7.58915901e-01 -2.39962682e-01
3.76242191e-01 2.98739821e-02 -5.50114512e-01 3.51061106e-01
-4.41291064e-01 -2.71883428e-01 5.77512801e-01 -1.20154060e-01
-1.94679409e-01 -4.45828855e-01 -1.17699122e+00 7.83086240e-01
7.30072260e-01 -2.63129860e-01 -4.57402349e-01 -1.64667398e-01
4.89537977e-02 -5.20631447e-02 -2.95948029e-01 3.42787951e-01
1.11025119e+00 -1.00642872e+00 1.12976682e+00 -5.75872540e-01
-2.04021603e-01 -5.63204885e-01 -1.79322511e-02 -7.67318428e-01
-2.15344518e-01 -3.46451811e-02 3.45275640e-01 1.03493989e+00
4.09232527e-01 -1.04345024e+00 4.29608583e-01 1.04780936e+00
-7.83606395e-02 -2.39891559e-01 -9.83157039e-01 -8.92052948e-01
-7.86737561e-01 -7.83034563e-01 4.49231774e-01 5.47954679e-01
-4.46348220e-01 -3.75524491e-01 -1.99191809e-01 4.22770351e-01
2.85298616e-01 -1.19647510e-01 1.02704751e+00 -1.07119012e+00
2.44581386e-01 -5.70407093e-01 -1.06049943e+00 -2.42942289e-01
-3.87610763e-01 -2.71649063e-01 -2.98493624e-01 -1.35659993e+00
-3.77161562e-01 -5.43352008e-01 -2.46942922e-01 3.09157580e-01
-1.51383772e-01 4.35956270e-01 7.20890403e-01 2.37981826e-01
-1.23058841e-01 1.81522012e-01 1.27858543e+00 3.26378196e-02
-4.34750170e-02 2.97061175e-01 -2.85390198e-01 5.21927834e-01
8.72892082e-01 -4.32216644e-01 -1.71476543e-01 -1.65224001e-01
2.25485891e-01 -4.53607261e-01 6.53155327e-01 -1.26999509e+00
1.59545138e-01 -2.24391758e-01 5.16810119e-01 -5.32990217e-01
3.83294225e-01 -1.19487333e+00 1.21179111e-01 6.83164358e-01
-4.22863439e-02 3.11300606e-01 5.41031599e-01 4.05749768e-01
-3.79627459e-02 -3.85037035e-01 5.89558065e-01 3.25727701e-01
-1.05577385e+00 -3.41837615e-01 -6.88713074e-01 -5.76919317e-01
1.47074068e+00 -1.03314745e+00 -3.45487833e-01 -3.70356500e-01
-2.36197203e-01 -8.81789848e-02 3.44325304e-01 6.82849228e-01
5.79308271e-01 -8.94665837e-01 -1.88380301e-01 6.75849617e-01
1.29133120e-01 -5.54459214e-01 3.07212442e-01 1.01930332e+00
-9.39400077e-01 6.34293854e-01 -6.20464265e-01 -3.78531784e-01
-1.65629900e+00 3.36044252e-01 2.84126759e-01 3.15605521e-01
-2.34541863e-01 3.69372398e-01 -2.54555941e-01 -3.13496999e-02
2.24711180e-01 -3.74808282e-01 -3.80080014e-01 -3.01649538e-03
3.76123935e-01 8.85754585e-01 9.11747757e-03 -6.05807483e-01
-2.93942153e-01 6.91384435e-01 4.97475676e-02 6.83256015e-02
6.13406897e-01 -6.99070245e-02 3.31777453e-01 2.89568633e-01
5.31624973e-01 1.00706391e-01 -9.55306768e-01 2.58385569e-01
-6.79100528e-02 -9.13241684e-01 -9.72474292e-02 -1.02128685e+00
-8.65399778e-01 6.55583739e-01 9.10746276e-01 2.13698059e-01
8.91286612e-01 -4.59150732e-01 5.08185387e-01 6.25633121e-01
2.04225570e-01 -1.16992331e+00 -6.97178170e-02 1.78990349e-01
7.19745874e-01 -1.18141127e+00 3.28424685e-02 -4.38312232e-01
-6.14119172e-01 1.08628380e+00 6.43297791e-01 2.06478581e-01
4.17408973e-01 -2.12628394e-02 3.94680023e-01 -1.69512197e-01
-1.34042799e-01 -4.16694701e-01 5.41213334e-01 9.12446499e-01
2.56347358e-01 -4.30058548e-03 -4.90599364e-01 2.61234716e-02
-1.59247637e-01 4.03677933e-02 3.78999889e-01 9.94695961e-01
-5.85883081e-01 -8.16443801e-01 -1.04785442e+00 5.56886911e-01
-4.12149504e-02 3.08698803e-01 -5.67714036e-01 1.25188839e+00
3.42758387e-01 1.06581879e+00 6.33191243e-02 -4.06912953e-01
6.57389343e-01 9.63485911e-02 3.07175905e-01 -1.63844168e-01
-3.71216357e-01 -6.03061676e-01 4.16534960e-01 -2.46253759e-01
-3.32901508e-01 -7.18847752e-01 -7.99019575e-01 -4.24225718e-01
-6.21619165e-01 -2.16591492e-01 1.24333596e+00 9.00241196e-01
2.03218132e-01 4.82076883e-01 1.92619056e-01 -5.50534785e-01
-2.44934008e-01 -7.73746133e-01 -3.76652479e-01 6.02936924e-01
-3.60357091e-02 -1.07630241e+00 -3.63620967e-01 -3.29172015e-02] | [8.000892639160156, -0.8840978741645813] |
b0d3b7d1-7bf4-4f71-83f9-b0436db17db0 | global-consistent-point-cloud-registration | 2208.07103 | null | https://arxiv.org/abs/2208.07103v1 | https://arxiv.org/pdf/2208.07103v1.pdf | Global Consistent Point Cloud Registration Based on Lie-algebraic Cohomology | We present a novel, effective method for global point cloud registration problems by geometric topology. Based on many point cloud pairwise registration methods (e.g ICP), we focus on the problem of accumulated error for the composition of transformations along any loops. The major technical contribution of this paper is a linear method for the elimination of errors, using only solving a Poisson equation. We demonstrate the consistency of our method from Hodge-Helmhotz decomposition theorem and experiments on multiple RGBD datasets of real-world scenes. The experimental results also demonstrate that our global registration method runs quickly and provides accurate reconstructions. | ['Xianfeng David Gu', 'Na lei', 'Wei Chen', 'Baowei Jiang', 'Yuxue Ren'] | 2022-08-15 | null | null | null | null | ['point-cloud-registration'] | ['computer-vision'] | [-5.25988676e-02 -2.84189999e-01 3.15873682e-01 -2.25329369e-01
-7.54828513e-01 -3.90116543e-01 4.46040034e-01 -4.59932089e-02
-3.30641955e-01 6.68531358e-01 -4.16684061e-01 -2.37649884e-02
-2.19164118e-01 -9.38578963e-01 -8.10554802e-01 -7.39796102e-01
2.08906904e-02 9.27143335e-01 3.56645763e-01 -4.45270747e-01
4.63128656e-01 9.51592803e-01 -1.26905477e+00 -6.30546272e-01
8.71027529e-01 8.80557716e-01 4.54959413e-03 5.35477459e-01
-1.60959959e-01 2.08075240e-01 -2.18315378e-01 -1.41302660e-01
6.85841322e-01 -2.80185044e-01 -1.02129674e+00 1.12454861e-01
6.97310627e-01 -1.40290573e-01 -1.01362437e-01 9.46904063e-01
4.69264865e-01 1.69496775e-01 6.35243297e-01 -1.51403630e+00
-2.07878008e-01 -2.22654864e-01 -8.26840222e-01 -3.66084635e-01
6.53303087e-01 -3.46475154e-01 5.31253159e-01 -1.12115741e+00
8.34006608e-01 1.12570345e+00 1.37480628e+00 3.52977626e-02
-1.30543697e+00 -4.69101042e-01 -3.73212010e-01 -8.76056179e-02
-1.90263104e+00 -2.93581992e-01 6.80428505e-01 -2.49641985e-01
9.94871318e-01 6.39767885e-01 7.35698938e-01 1.24816582e-01
4.81734127e-01 -1.31830484e-01 1.16002882e+00 -5.22214770e-01
1.40127629e-01 -3.65919143e-01 1.33432314e-01 6.35792196e-01
5.00140846e-01 -5.30311055e-02 -2.75254041e-01 -6.79884136e-01
1.33532727e+00 3.43507081e-02 -1.19105995e-01 -7.82925367e-01
-1.22399950e+00 7.22305179e-01 4.53282684e-01 2.21176520e-01
-3.06698769e-01 5.53084373e-01 -1.32069483e-01 7.40856081e-02
5.36407292e-01 -3.00159380e-02 -3.45810592e-01 5.66620417e-02
-7.18040287e-01 4.81201738e-01 9.19264495e-01 1.44869983e+00
1.29469430e+00 -4.82314467e-01 7.83783734e-01 4.92304653e-01
5.90516508e-01 1.10581362e+00 -3.41714680e-01 -1.19219363e+00
3.34998131e-01 4.70088363e-01 1.03589334e-01 -1.41994631e+00
-3.59778345e-01 1.94997013e-01 -8.66712689e-01 3.09974104e-01
-1.13600008e-02 3.34050387e-01 -6.02801919e-01 1.12834990e+00
6.79330587e-01 4.28082556e-01 -1.33307323e-01 5.63777864e-01
5.49229980e-01 5.82681060e-01 -3.30948114e-01 -4.22600120e-01
9.05970693e-01 -4.67910647e-01 -7.91569233e-01 3.54215413e-01
3.29860032e-01 -1.07273901e+00 4.01610315e-01 1.27050430e-01
-1.14274967e+00 -2.28495255e-01 -7.52055526e-01 -2.40163296e-01
6.89351633e-02 -3.28476161e-01 4.89519477e-01 2.39874870e-01
-1.42301691e+00 8.45436335e-01 -1.02902782e+00 -8.71990860e-01
-4.64677960e-02 8.01702023e-01 -5.47412336e-01 2.36902997e-01
-3.69164914e-01 1.04251111e+00 -2.78749410e-02 2.47516468e-01
-8.75543803e-02 -5.79755783e-01 -6.64442062e-01 -5.74153125e-01
-9.27116275e-02 -9.62452233e-01 9.26840782e-01 -5.00263155e-01
-1.34982479e+00 1.10999012e+00 -4.81058985e-01 -1.23535627e-02
6.22102618e-01 4.60048467e-02 2.52822757e-01 7.51477107e-02
1.73408613e-01 3.53331625e-01 3.53022307e-01 -1.82481468e+00
-2.74919420e-01 -5.78330517e-01 -4.95944262e-01 2.31281593e-01
4.36949432e-01 1.50269672e-01 -2.74771184e-01 -3.26644212e-01
1.18306589e+00 -1.23937953e+00 -6.10506773e-01 2.95257807e-01
-1.37851432e-01 -1.37478873e-01 8.48917007e-01 -6.85866177e-01
3.80081296e-01 -1.96863973e+00 2.46714890e-01 7.83233941e-01
1.54353648e-01 -4.51419771e-01 1.51512399e-01 5.33147514e-01
4.89868447e-02 2.45632663e-01 -7.24401951e-01 -7.19395161e-01
-2.72862554e-01 5.51784694e-01 -4.30552244e-01 1.12698221e+00
-2.22405180e-01 5.41074455e-01 -7.28052080e-01 -7.69924581e-01
4.56624180e-01 6.90705597e-01 -1.58444896e-01 6.55099079e-02
3.83408606e-01 7.67296672e-01 -2.37480730e-01 5.85850716e-01
1.34190857e+00 2.04547301e-01 -2.37389207e-01 -4.52605158e-01
-5.07463753e-01 1.58797905e-01 -1.65143657e+00 1.82519174e+00
-2.60239273e-01 1.87875167e-01 3.42591822e-01 -6.08720005e-01
1.12318957e+00 2.44291261e-01 1.11461043e+00 -2.93213129e-01
1.40687943e-01 5.42610466e-01 -7.02730417e-01 4.42363620e-02
7.02833056e-01 -3.36701900e-01 1.20138545e-02 3.79387945e-01
-9.16185752e-02 -9.24388051e-01 -4.45567310e-01 1.05018780e-01
9.10171926e-01 3.81857514e-01 4.97996867e-01 -7.03877747e-01
4.05721724e-01 5.49423695e-01 5.82601011e-01 4.26196009e-01
-1.24824969e-02 1.01487660e+00 -2.50130773e-01 -5.05511761e-01
-1.17013752e+00 -1.18678856e+00 -5.45371950e-01 5.98728172e-02
7.55930185e-01 -4.79752153e-01 -5.53744256e-01 8.09073001e-02
1.52653754e-01 1.43723726e-01 -2.35490218e-01 4.43590403e-01
-9.16429281e-01 -7.87546456e-01 3.02502155e-01 2.86959171e-01
6.29192650e-01 -4.61671174e-01 -3.97136778e-01 1.21358618e-01
-3.64091516e-01 -1.10502589e+00 -1.36324093e-01 -1.72613531e-01
-1.55870271e+00 -1.16275740e+00 -2.26451680e-01 -7.41988719e-01
1.03579330e+00 7.14235067e-01 1.13676596e+00 4.22898591e-01
-3.73203486e-01 9.32203412e-01 -1.10589890e-02 -1.54151276e-01
-1.91654906e-01 -3.28885376e-01 2.55205542e-01 -2.97752500e-01
1.42479718e-01 -8.98581624e-01 -2.72673666e-01 7.00897098e-01
-4.60275948e-01 -1.90577835e-01 -9.22728851e-02 2.22484469e-01
1.34795058e+00 -9.38012749e-02 -5.41913092e-01 -4.61529315e-01
2.25301951e-01 -1.14832550e-01 -9.68218982e-01 2.25242302e-01
-3.59495431e-01 -1.70354307e-01 -2.20537648e-01 5.27002625e-02
-7.49292314e-01 5.46154141e-01 5.29743405e-03 -4.45199341e-01
1.10128127e-01 -2.56358329e-02 -4.21751328e-02 -1.17591441e+00
5.51460147e-01 1.55158073e-01 1.31971255e-01 -5.06445527e-01
2.67923534e-01 3.93546820e-01 6.75924182e-01 -8.88478100e-01
1.20708919e+00 1.13496280e+00 5.86358905e-01 -1.22418475e+00
1.53638855e-01 -9.38819647e-01 -1.19813943e+00 -3.07848632e-01
7.86943793e-01 -7.66956925e-01 -8.64122748e-01 5.38213074e-01
-1.69317269e+00 4.54615150e-03 -3.29442620e-01 3.35023373e-01
-9.64900196e-01 5.70902050e-01 -3.79588038e-01 -7.80105412e-01
-2.26332977e-01 -1.19687772e+00 1.37318993e+00 -1.69125229e-01
-5.38959913e-02 -1.14692628e+00 8.69460285e-01 -1.81046668e-02
2.94189006e-01 7.83472598e-01 8.70592818e-02 1.49204507e-01
-1.18085766e+00 1.65304225e-02 -4.07214910e-02 -1.83488950e-01
2.56180555e-01 5.01627445e-01 -7.38848865e-01 -1.38551474e-01
6.62416041e-01 4.09001619e-01 1.95774227e-01 1.70732051e-01
6.71322167e-01 -1.52232423e-01 -5.95977783e-01 1.12695229e+00
2.14933133e+00 -1.43598676e-01 1.02660370e+00 4.97164398e-01
1.02506983e+00 3.93748999e-01 6.36193752e-01 1.52654409e-01
5.87961078e-01 8.90921772e-01 4.36077893e-01 -1.59473345e-01
1.35400504e-01 8.07717815e-02 -1.29228503e-01 1.35866857e+00
-9.29347396e-01 5.21273315e-01 -1.26919782e+00 3.44144762e-01
-2.03405428e+00 -6.96407974e-01 -1.10304701e+00 2.40094256e+00
5.54366052e-01 -7.37414300e-01 -1.15092635e-01 -4.40105684e-02
7.65534222e-01 -3.32531005e-01 6.06404282e-02 -2.46788561e-01
-2.51102656e-01 3.95642459e-01 1.25580859e+00 9.22370613e-01
-9.72993076e-01 7.94671178e-01 7.55385447e+00 2.99057186e-01
-6.93205416e-01 4.60520655e-01 -1.85678229e-01 5.32666922e-01
-3.37111354e-01 3.95019203e-01 -7.71072745e-01 -2.31195837e-02
5.46051145e-01 -2.02805385e-01 4.08148885e-01 5.31078458e-01
5.32882772e-02 -3.81873548e-01 -7.27225780e-01 1.29715097e+00
1.82969436e-01 -1.20572376e+00 -1.35640770e-01 3.75153691e-01
9.40606177e-01 3.44397992e-01 -5.54296911e-01 -5.67742229e-01
4.10647869e-01 -6.52684510e-01 6.39706254e-01 5.50510287e-01
5.90478063e-01 -4.80382949e-01 5.55751026e-01 1.72731519e-01
-1.61397696e+00 7.82680154e-01 -7.84509063e-01 -8.32383931e-02
3.56244951e-01 7.01645911e-01 -3.95475090e-01 9.61046159e-01
6.91194296e-01 7.35996366e-01 -3.80485028e-01 1.32694972e+00
1.04961708e-01 -1.13586940e-01 -1.00271070e+00 4.88651186e-01
-3.08637887e-01 -1.06808281e+00 6.93391562e-01 8.34781587e-01
6.56121254e-01 7.49665380e-01 6.22166693e-02 7.78509140e-01
3.26409817e-01 3.32975537e-01 -9.55040157e-01 8.82350147e-01
3.31010789e-01 1.15103114e+00 -9.05634999e-01 8.85168314e-02
-1.94379136e-01 9.07712221e-01 1.36172799e-02 5.63263707e-02
-7.28259385e-01 -4.12273929e-02 7.00172722e-01 1.33953601e-01
-4.89870310e-02 -8.96486282e-01 -8.28586400e-01 -1.09327722e+00
1.31131783e-01 -2.26634637e-01 -7.32394159e-02 -9.20901120e-01
-1.23681188e+00 5.16348302e-01 3.18836361e-01 -1.48985219e+00
3.13286006e-01 -4.27510172e-01 -6.19066775e-01 9.96803105e-01
-1.13129997e+00 -1.19788754e+00 -6.48774207e-01 1.02285123e+00
-2.63972491e-01 5.21935284e-01 7.69863546e-01 2.00183794e-01
7.75692612e-02 -6.23032600e-02 1.43124759e-01 -1.96586430e-01
7.11836874e-01 -1.04346216e+00 5.01186073e-01 7.18345046e-01
-1.87806636e-01 7.64194727e-01 8.03519487e-01 -1.02678311e+00
-1.79892004e+00 -8.20594668e-01 9.83307838e-01 -7.40348399e-01
5.82138956e-01 -2.54180014e-01 -8.62585247e-01 1.10201645e+00
4.48625498e-02 1.89658463e-01 2.90915191e-01 -1.42064318e-01
-1.44253112e-02 -3.22511524e-01 -1.50231373e+00 1.24458522e-01
1.18637669e+00 -4.28583920e-01 -7.10140109e-01 6.07362986e-01
5.35193801e-01 -9.17952776e-01 -1.06933391e+00 5.60967207e-01
1.38823003e-01 -7.69851923e-01 1.21855152e+00 1.95741415e-01
-1.21594310e-01 -6.79895222e-01 -4.23879325e-01 -8.25487375e-01
-1.37897074e-01 -6.92248404e-01 5.11908770e-01 1.17065358e+00
-2.52939820e-01 -9.71431613e-01 4.27490324e-01 8.36805344e-01
-3.92560884e-02 -1.18889861e-01 -1.36037827e+00 -9.23642576e-01
-7.82845020e-02 -3.47507983e-01 6.45941973e-01 1.15774512e+00
-2.66657919e-01 -2.31892958e-01 -1.34039193e-01 5.01818180e-01
1.39359665e+00 -5.69807319e-03 1.07203865e+00 -1.53762734e+00
3.33257973e-01 6.66882545e-02 -8.59521866e-01 -6.87192738e-01
-9.69446674e-02 -7.06690133e-01 3.79312932e-01 -1.72956002e+00
8.55466798e-02 -1.03611302e+00 4.21725780e-01 2.58496732e-01
2.79151529e-01 4.98274654e-01 1.53968871e-01 8.61835301e-01
-3.44166905e-01 4.01945561e-01 1.00585139e+00 1.87882513e-01
-1.51260689e-01 -1.59471288e-01 -4.04731333e-02 6.91988885e-01
7.19046056e-01 -7.07185328e-01 1.58912361e-01 -5.36198378e-01
3.95761102e-01 -1.01725936e-01 5.97718298e-01 -1.08489001e+00
4.70027834e-01 -3.57168704e-01 -1.35119200e-01 -8.41896653e-01
4.32485521e-01 -1.36010337e+00 1.17303312e+00 4.31740135e-01
5.88828027e-01 7.56273746e-01 1.30572006e-01 5.34248173e-01
-1.26651451e-01 -1.99892342e-01 7.18591094e-01 -2.18676969e-01
-3.34409148e-01 4.42138165e-01 2.38041759e-01 -4.02301997e-01
1.16620052e+00 -3.99880409e-01 -1.60065860e-01 -4.78956625e-02
-6.15058959e-01 -6.37736619e-02 1.22556221e+00 -1.63960695e-01
6.05303824e-01 -1.73573041e+00 -6.97295070e-01 2.36946568e-01
-9.44987386e-02 4.22605067e-01 -2.54587531e-01 1.05009460e+00
-1.47189760e+00 2.10046768e-01 -2.09822223e-01 -1.31933963e+00
-1.35791540e+00 1.35973185e-01 5.72301805e-01 2.33947426e-01
-6.88208580e-01 5.74538708e-01 -2.13294074e-01 -8.95799398e-01
-9.74157751e-02 -3.26835573e-01 4.98464167e-01 -4.11055982e-01
3.70227955e-02 6.42156124e-01 3.48052025e-01 -1.21655953e+00
-7.42429137e-01 1.64871621e+00 7.20547497e-01 -3.70334655e-01
1.49393749e+00 -2.87919581e-01 -1.17068839e+00 4.76025581e-01
1.15900970e+00 2.04765648e-01 -7.35462129e-01 9.86519177e-03
-3.76243025e-01 -7.49165356e-01 -1.48884729e-01 1.13345414e-01
-9.46034253e-01 5.73279083e-01 6.07728004e-01 3.76916118e-03
9.21281576e-01 5.53930067e-02 4.90499616e-01 3.09899271e-01
1.21011817e+00 -6.82883382e-01 -6.51944757e-01 5.17913043e-01
1.05687702e+00 -9.85645592e-01 5.10650277e-01 -1.03548479e+00
5.07200919e-02 1.16645145e+00 2.46829808e-01 -6.49236500e-01
8.68491948e-01 3.42953324e-01 7.03808479e-03 -3.10392708e-01
1.19590260e-01 7.95422941e-02 -1.80111468e-01 6.26317561e-01
8.48252699e-02 2.60288604e-02 -5.95497549e-01 -3.33047837e-01
-2.95153022e-01 -8.93968642e-02 4.89808768e-01 1.14348662e+00
-4.06892061e-01 -1.33964312e+00 -1.13695991e+00 -5.73056228e-02
9.73836184e-02 1.78840980e-01 -3.95001233e-01 9.17256653e-01
1.83196384e-02 6.21285081e-01 4.29964930e-01 -1.47554308e-01
4.72828448e-01 -2.34616905e-01 9.74855006e-01 -1.89727619e-01
-4.87570822e-01 1.30906671e-01 -4.02413607e-01 -9.26622629e-01
-1.18699563e+00 -1.04213405e+00 -1.42911410e+00 -8.08159351e-01
-4.50379610e-01 2.74128467e-01 1.15516043e+00 7.10158050e-01
1.57010481e-01 -8.50292891e-02 7.14859664e-01 -1.25332820e+00
-2.10556969e-01 -4.67294723e-01 -6.79907203e-01 3.62433881e-01
1.24867715e-01 -7.25193501e-01 -6.02876961e-01 1.64556608e-01] | [7.777622222900391, -2.8320934772491455] |
f25851e1-6302-4c54-a495-4db2b78763a9 | knowing-depth-quality-in-advance-a-depth | 2008.04157 | null | https://arxiv.org/abs/2008.04157v1 | https://arxiv.org/pdf/2008.04157v1.pdf | Knowing Depth Quality In Advance: A Depth Quality Assessment Method For RGB-D Salient Object Detection | Previous RGB-D salient object detection (SOD) methods have widely adopted deep learning tools to automatically strike a trade-off between RGB and D (depth), whose key rationale is to take full advantage of their complementary nature, aiming for a much-improved SOD performance than that of using either of them solely. However, such fully automatic fusions may not always be helpful for the SOD task because the D quality itself usually varies from scene to scene. It may easily lead to a suboptimal fusion result if the D quality is not considered beforehand. Moreover, as an objective factor, the D quality has long been overlooked by previous work. As a result, it is becoming a clear performance bottleneck. Thus, we propose a simple yet effective scheme to measure D quality in advance, the key idea of which is to devise a series of features in accordance with the common attributes of high-quality D regions. To be more concrete, we conduct D quality assessments for each image region, following a multi-scale methodology that includes low-level edge consistency, mid-level regional uncertainty and high-level model variance. All these components will be computed independently and then be assembled with RGB and D features, applied as implicit indicators, to guide the selective fusion. Compared with the state-of-the-art fusion schemes, our method can achieve a more reasonable fusion status between RGB and D. Specifically, the proposed D quality measurement method achieves steady performance improvements for almost 2.0\% in general. | ['Chenglizhao Chen', 'Xuehao Wang', 'Shuai Li', 'Hong Qin', 'Aimin Hao'] | 2020-08-07 | null | null | null | null | ['rgb-d-salient-object-detection'] | ['computer-vision'] | [-2.98489314e-02 -2.66923517e-01 3.96700986e-02 -2.93746293e-01
-9.78736877e-01 -2.10940987e-01 5.43571174e-01 3.08409274e-01
-4.31523561e-01 6.08652830e-01 -2.21368410e-02 1.48913264e-02
-2.82594353e-01 -9.65253532e-01 -2.78348118e-01 -9.54523742e-01
3.01327974e-01 1.03215367e-01 4.88740116e-01 -3.15159053e-01
2.53629565e-01 7.32167065e-01 -1.77407122e+00 -1.94972068e-01
1.22800565e+00 1.53274977e+00 2.79478878e-01 1.47185594e-01
-3.22711587e-01 4.42718297e-01 -4.08194393e-01 -3.57324690e-01
2.88901687e-01 -3.63512009e-01 -3.87773216e-01 1.93591014e-01
-2.06687618e-02 -4.16031331e-01 -1.27026320e-01 1.30474532e+00
5.56776166e-01 -9.21160430e-02 7.56312072e-01 -1.04897761e+00
-2.25372344e-01 1.86490387e-01 -8.94082427e-01 1.61279187e-01
2.87036717e-01 3.23080689e-01 9.88736093e-01 -7.70719767e-01
2.58051872e-01 9.76591170e-01 5.00268698e-01 -3.46466387e-03
-9.95751798e-01 -5.39821208e-01 1.82396963e-01 1.64776519e-01
-1.43848431e+00 -2.32270449e-01 1.28113484e+00 -2.90836155e-01
4.25765544e-01 2.09168598e-01 8.96253228e-01 5.20605862e-01
3.14116180e-01 7.66007662e-01 1.24350226e+00 -3.38695526e-01
2.96475798e-01 1.40282199e-01 -1.33060575e-01 5.27642369e-01
4.05582577e-01 2.21151970e-02 -3.68503600e-01 2.45577842e-01
6.37183249e-01 -1.24511287e-01 -3.57559264e-01 -5.45214891e-01
-1.08330512e+00 4.92550761e-01 8.05153787e-01 5.23319304e-01
-4.69851255e-01 -1.28057405e-01 2.17155844e-01 -4.83620204e-02
4.72819477e-01 1.52980447e-01 -2.07501262e-01 -6.86917752e-02
-9.18537319e-01 2.06348717e-01 2.29516372e-01 4.63181615e-01
1.01768386e+00 -1.07414700e-01 -8.27775896e-02 5.89049280e-01
5.09323299e-01 3.46729755e-01 3.04117441e-01 -8.24542761e-01
3.13427418e-01 1.00126851e+00 2.96461195e-01 -1.33626759e+00
-5.53993821e-01 -4.90209550e-01 -1.02246499e+00 5.77673078e-01
5.84502637e-01 9.56821218e-02 -7.38090873e-01 1.50984216e+00
5.57236791e-01 -1.51239902e-01 -5.38465846e-03 1.21838236e+00
8.28220785e-01 2.72140354e-01 3.43163945e-02 -2.89065480e-01
1.31583118e+00 -4.57042694e-01 -8.46951663e-01 -1.00986823e-01
3.39401305e-01 -8.25873077e-01 9.60965097e-01 4.49760616e-01
-9.84057784e-01 -7.77356327e-01 -1.33927524e+00 1.90564003e-02
-4.63594049e-01 1.86917886e-01 7.03133523e-01 6.81560457e-01
-9.75278020e-01 4.40182090e-01 -6.32380545e-01 -1.47930741e-01
2.10101336e-01 1.23512164e-01 -2.42417291e-01 1.55763775e-01
-1.29040420e+00 9.93439436e-01 3.76736403e-01 4.67150927e-01
-5.60513437e-01 -4.01290268e-01 -6.26577616e-01 -1.38718635e-01
5.31964898e-01 -6.01546586e-01 1.05065310e+00 -7.75495172e-01
-1.45858383e+00 8.70684028e-01 1.04832783e-01 -1.34362668e-01
8.18762898e-01 -3.28191042e-01 -3.93480629e-01 2.12309986e-01
8.56303349e-02 3.94785881e-01 7.70286560e-01 -1.68336952e+00
-8.73491347e-01 -5.59274137e-01 3.04235578e-01 4.51094210e-01
-3.46904308e-01 -2.39387512e-01 -6.91998780e-01 -5.67837059e-01
5.18727899e-01 -4.18845505e-01 -1.98123977e-01 3.49148721e-01
-2.93996423e-01 -1.09600484e-01 4.92145568e-01 -4.62270379e-01
1.25035167e+00 -2.15314341e+00 1.59968048e-01 1.24710150e-01
3.58902395e-01 2.67741382e-01 4.10875291e-01 -7.69408196e-02
2.78071970e-01 -1.09165207e-01 -4.14825380e-01 -3.48851085e-01
-1.46732092e-01 -7.02428520e-02 1.84999794e-01 6.53261483e-01
3.84325057e-01 6.84102714e-01 -9.77580547e-01 -8.12315404e-01
6.91810369e-01 4.85547543e-01 -2.26611018e-01 1.68305635e-01
-1.24221653e-01 3.55384409e-01 -7.19217837e-01 8.26433957e-01
9.62779701e-01 -1.19312219e-01 -6.34257644e-02 -5.24103284e-01
-4.21480089e-01 -2.98380130e-03 -1.56905186e+00 1.64086664e+00
-3.21802288e-01 2.31231645e-01 2.10196361e-01 -9.97279048e-01
1.31366313e+00 9.02689844e-02 8.08211446e-01 -8.89604867e-01
3.64372134e-01 5.11407197e-01 -8.06408972e-02 -4.06510562e-01
5.24983943e-01 -2.40537629e-01 -5.96284755e-02 1.35271445e-01
-3.28600317e-01 -3.88178825e-01 -1.13012522e-01 -1.51946217e-01
7.94856608e-01 2.14319959e-01 5.50244391e-01 -1.92062974e-01
8.08031201e-01 -8.57398510e-02 6.04229510e-01 5.14721155e-01
-5.06409705e-01 7.35445321e-01 4.64053124e-01 -1.81115136e-01
-8.10434520e-01 -1.11545432e+00 -2.79889852e-01 2.63530135e-01
7.35778213e-01 -6.19687606e-03 -5.00186861e-01 -6.17711544e-01
6.50217310e-02 3.54870349e-01 -5.91885567e-01 -3.16403806e-01
-4.17575836e-01 -8.49427998e-01 2.14607209e-01 3.85712624e-01
1.08391356e+00 -8.42362344e-01 -9.21908677e-01 3.25733632e-01
-1.91687644e-02 -9.47925270e-01 2.08998099e-02 3.74801904e-01
-8.20995569e-01 -1.04455709e+00 -8.73439431e-01 -3.24387878e-01
3.65371495e-01 5.04178762e-01 9.93918717e-01 9.58230942e-02
-5.19349091e-02 -7.72326812e-02 -4.60018665e-01 -2.58885622e-01
-1.60551995e-01 -3.29950266e-02 4.97377478e-02 1.77357644e-01
1.82607606e-01 -4.29035366e-01 -8.94431055e-01 3.21349531e-01
-1.00353074e+00 1.06773883e-01 8.28892708e-01 6.67585194e-01
5.97207606e-01 2.66315341e-01 5.53573787e-01 -2.47346491e-01
2.91092217e-01 -2.73529947e-01 -5.78637242e-01 3.30506384e-01
-6.62953198e-01 -5.75972907e-02 4.87227619e-01 -4.38088812e-02
-1.23143506e+00 5.45507483e-02 -3.97387624e-01 -3.79349202e-01
-2.15725541e-01 4.09560889e-01 -6.31282985e-01 -2.06107106e-02
3.52897257e-01 2.87099868e-01 5.43822302e-03 -4.19829488e-01
2.93221503e-01 7.13714182e-01 4.23452318e-01 -4.42380846e-01
7.47784197e-01 5.88753343e-01 8.20213035e-02 -5.72344005e-01
-7.92282760e-01 -4.86753643e-01 -7.72376478e-01 -5.95234215e-01
8.70649576e-01 -8.26938748e-01 -5.55715084e-01 7.32334554e-01
-1.06872880e+00 2.21831724e-02 -1.57386422e-01 4.29736018e-01
-4.51481611e-01 5.73651016e-01 -2.44927436e-01 -1.00821900e+00
-5.87406084e-02 -1.40638101e+00 1.27859390e+00 4.56970274e-01
2.19675109e-01 -7.79427946e-01 -1.90122768e-01 2.19652250e-01
3.08213234e-01 4.08218235e-01 6.65460646e-01 -4.70860535e-03
-5.82475364e-01 -8.19935128e-02 -5.94526649e-01 3.55437785e-01
4.65284079e-01 1.73087418e-01 -1.05273604e+00 6.31961375e-02
2.94053316e-01 -8.76431987e-02 8.03566039e-01 4.65616673e-01
1.00995958e+00 3.40402544e-01 -2.04825744e-01 6.44540489e-01
1.69960117e+00 2.94498622e-01 6.60927653e-01 6.38392866e-01
5.63733697e-01 4.71772820e-01 1.04231453e+00 6.45281076e-01
3.80213112e-01 9.58443999e-01 7.58279562e-01 -4.38896209e-01
-1.77807420e-01 1.66967046e-02 3.22949030e-02 6.00852072e-01
-1.26155972e-01 -1.28754452e-01 -7.93832481e-01 3.99977982e-01
-1.80594313e+00 -6.82550490e-01 -7.63700902e-02 2.18859863e+00
7.75525093e-01 5.93234003e-01 1.50199085e-01 5.78675568e-01
6.35979116e-01 3.14276755e-01 -5.53282559e-01 1.01485804e-01
-3.16014141e-01 -2.23673373e-01 4.54863489e-01 1.67266458e-01
-1.06348264e+00 4.07023996e-01 5.39770842e+00 1.00853395e+00
-1.26265538e+00 -1.51347488e-01 7.71082699e-01 1.78820416e-01
-4.34130698e-01 -5.41289523e-02 -6.83931470e-01 5.91464698e-01
2.52363533e-01 -2.22758837e-02 -3.87587696e-02 7.42002368e-01
2.93163806e-01 -5.80487072e-01 -6.92651808e-01 1.17451143e+00
-1.75533921e-01 -1.02531910e+00 -1.90218017e-01 6.18805923e-02
5.72243810e-01 -3.04976821e-01 1.00654429e-02 3.84077765e-02
-5.48865609e-02 -6.24320984e-01 9.98922169e-01 6.06290162e-01
5.99538445e-01 -8.58604670e-01 9.60403681e-01 3.39037448e-01
-1.43279707e+00 -3.36122839e-03 -2.83694893e-01 4.18767966e-02
3.20893168e-01 9.99104738e-01 -1.31351933e-01 1.24394619e+00
8.63717794e-01 6.22503817e-01 -5.79421580e-01 1.22253263e+00
-1.68657899e-01 -1.78709421e-02 -4.52082783e-01 -5.08655906e-02
2.68671244e-01 -3.00306052e-01 5.18093646e-01 9.08399403e-01
4.39307630e-01 1.97213009e-01 4.96903025e-02 7.34054208e-01
1.39461488e-01 2.49519378e-01 -5.28431237e-01 5.17246664e-01
3.09404403e-01 1.39589930e+00 -1.07636511e+00 -3.30083817e-01
-5.40351868e-01 7.38069654e-01 2.17544168e-01 1.14867635e-01
-9.74030554e-01 -2.06311554e-01 6.76837921e-01 1.19579636e-01
2.99558252e-01 -1.24003947e-01 -6.42379284e-01 -1.05970168e+00
1.80828869e-01 -5.31996131e-01 2.09830269e-01 -7.24741817e-01
-1.30003822e+00 6.98552430e-01 -2.51031644e-03 -1.81090760e+00
2.11232882e-02 -5.71231604e-01 -3.93612593e-01 7.91859090e-01
-1.76110923e+00 -1.00210941e+00 -6.68585598e-01 4.12476391e-01
2.86197692e-01 1.74991846e-01 4.24733877e-01 5.39499402e-01
-6.27652109e-01 3.74435335e-01 6.61881566e-02 -1.47355050e-01
5.58580339e-01 -1.12926149e+00 -1.54954776e-01 1.13524234e+00
-1.39917016e-01 2.13499412e-01 7.45179355e-01 -5.85946202e-01
-1.23273349e+00 -7.02313125e-01 3.92641634e-01 -1.93979785e-01
6.00737333e-01 6.92545772e-02 -8.71577084e-01 -1.41840935e-01
-2.40778014e-01 5.87809756e-02 5.34444153e-02 -1.66782007e-01
3.60122994e-02 -5.47097743e-01 -1.14778948e+00 5.57261229e-01
9.56927657e-01 -4.74056304e-01 -5.70087135e-01 -2.05809996e-01
5.93839645e-01 -3.38619798e-01 -9.09795105e-01 8.08021605e-01
4.66930211e-01 -1.57617879e+00 1.00172698e+00 3.38524789e-01
4.09136862e-01 -7.37511516e-01 -1.90828234e-01 -1.04731333e+00
-9.55942571e-02 -1.76424041e-01 -3.83236445e-02 1.47099268e+00
-3.18805054e-02 -5.34855306e-01 6.00559056e-01 3.83154035e-01
-2.70743042e-01 -9.04152274e-01 -1.12275279e+00 -6.50699198e-01
-2.37226069e-01 -6.17149293e-01 8.09027195e-01 7.23257899e-01
-3.59834582e-01 4.15808782e-02 -1.93033218e-01 1.05303504e-01
6.71956718e-01 3.28321844e-01 7.25073516e-01 -1.25631249e+00
1.06420554e-03 -9.17909920e-01 -4.53610241e-01 -1.11284459e+00
-3.44508946e-01 -3.93340558e-01 2.54748166e-01 -1.62547433e+00
-4.12730388e-02 -9.06175196e-01 -6.77738309e-01 1.16465777e-01
-4.86491263e-01 4.57912296e-01 -4.29590158e-02 3.29550922e-01
-4.52076793e-01 8.54164302e-01 1.55063438e+00 -7.53514990e-02
-2.12056860e-01 -2.57034991e-02 -7.57752299e-01 7.69314170e-01
6.49974048e-01 -1.35692134e-01 -2.23170966e-01 -2.17216924e-01
1.17322266e-01 7.50687867e-02 3.77751648e-01 -1.37256813e+00
1.70869499e-01 -1.89658046e-01 5.27476430e-01 -7.67454982e-01
3.80310357e-01 -1.00778937e+00 1.75862059e-01 3.79489541e-01
2.02383742e-01 -1.97746933e-01 8.28425884e-02 4.24183398e-01
-5.47775567e-01 3.01534962e-02 9.60838556e-01 -7.68002570e-02
-1.04840124e+00 2.51704067e-01 5.53224944e-02 -4.03258830e-01
1.08953977e+00 -5.50766051e-01 -7.46856406e-02 -2.26290151e-01
-3.69559854e-01 5.10585569e-02 7.77127862e-01 2.29716361e-01
5.81511140e-01 -1.38547432e+00 -4.51115042e-01 1.52108043e-01
3.74163568e-01 1.99742824e-01 4.68867093e-01 1.04325080e+00
-4.53075320e-01 5.41505292e-02 -3.21218669e-01 -8.59188735e-01
-7.76313245e-01 5.82224905e-01 3.38129044e-01 -2.14330375e-01
-5.46048045e-01 5.87637663e-01 1.15763946e-02 4.80601713e-02
1.66243583e-01 -5.10301650e-01 -3.09810191e-01 4.22689587e-01
3.72937977e-01 1.46637067e-01 3.57066959e-01 -7.61162937e-01
-5.48471093e-01 9.43322480e-01 3.21874887e-01 3.37382555e-02
1.18619049e+00 -5.04298806e-01 9.04195756e-02 4.80139762e-01
9.59459484e-01 -1.18529135e-02 -1.55033708e+00 -1.93163678e-01
-1.44856021e-01 -6.88191235e-01 3.17321986e-01 -5.00494897e-01
-1.22545886e+00 8.24644208e-01 8.22452068e-01 4.81986612e-01
1.59303844e+00 5.67377359e-02 5.37694931e-01 -6.95803240e-02
5.54323375e-01 -1.10070956e+00 6.28723800e-02 -1.16497077e-01
6.74430549e-01 -1.59318995e+00 3.36018175e-01 -4.29695874e-01
-4.57142323e-01 1.06563997e+00 6.90317988e-01 5.30097447e-02
6.80200219e-01 2.80374914e-01 1.24783471e-01 -1.49359882e-01
-1.34061262e-01 -6.17424548e-01 4.26348150e-01 3.99257720e-01
2.27502868e-01 -1.15072437e-01 -3.53659332e-01 4.13402766e-01
1.14309117e-01 -1.71811897e-02 5.07354736e-04 7.30666339e-01
-5.56504011e-01 -8.89805973e-01 -6.39460027e-01 4.03042406e-01
-2.35924855e-01 3.74966174e-01 4.89017405e-02 1.00579906e+00
4.35905695e-01 8.34350467e-01 -3.19341794e-02 -6.19931698e-01
4.93386567e-01 -3.33800912e-01 3.62209082e-01 -1.07785933e-01
-4.15688962e-01 1.34252250e-01 -1.93172395e-01 -6.47088766e-01
-8.48188937e-01 -6.79686785e-01 -1.13607275e+00 -2.51599282e-01
-5.20744979e-01 -2.08897769e-01 6.32890105e-01 1.08041048e+00
-8.94675553e-02 5.95927298e-01 7.80033648e-01 -1.09509516e+00
-3.00898403e-01 -7.22127676e-01 -6.09243214e-01 4.28367436e-01
3.04510832e-01 -1.17271733e+00 -5.23186266e-01 -3.64560932e-01] | [9.649033546447754, -0.9450000524520874] |
824abfb8-4d0a-4e2b-9013-ab78303e68a1 | improving-aspect-level-sentiment-analysis | 2005.06607 | null | https://arxiv.org/abs/2005.06607v1 | https://arxiv.org/pdf/2005.06607v1.pdf | Improving Aspect-Level Sentiment Analysis with Aspect Extraction | Aspect-based sentiment analysis (ABSA), a popular research area in NLP has two distinct parts -- aspect extraction (AE) and labeling the aspects with sentiment polarity (ALSA). Although distinct, these two tasks are highly correlated. The work primarily hypothesize that transferring knowledge from a pre-trained AE model can benefit the performance of ALSA models. Based on this hypothesis, word embeddings are obtained during AE and subsequently, feed that to the ALSA model. Empirically, this work show that the added information significantly improves the performance of three different baseline ALSA models on two distinct domains. This improvement also translates well across domains between AE and ALSA tasks. | ['Louis-Philippe Morency', 'Alexander Gelbukh', 'Rishabh Bhardwaj', 'Soujanya Poria', 'Amir Hussain', 'Navonil Majumder', 'Amir Zadeh'] | 2020-05-03 | null | null | null | null | ['aspect-extraction'] | ['natural-language-processing'] | [ 2.0937046e-01 3.9925322e-01 -4.3973339e-01 -7.3242217e-01
-8.2617468e-01 -7.7790403e-01 9.0871698e-01 2.5145060e-01
-3.1573468e-01 4.7411504e-01 6.3300079e-01 -3.0155677e-01
2.5914177e-01 -8.9600968e-01 -5.2720350e-01 -4.2062277e-01
4.5017070e-01 5.5602211e-01 -1.7483285e-01 -3.6838380e-01
2.6154920e-01 1.9749591e-01 -1.1295063e+00 5.7886147e-01
6.0498255e-01 1.0156986e+00 -5.7548046e-01 5.4031485e-01
-7.2494227e-01 6.4738137e-01 -5.6823254e-01 -8.7413365e-01
1.8609634e-01 -1.0839704e-01 -9.3835539e-01 -1.3408080e-01
3.3142740e-01 2.0785590e-01 2.9972219e-01 7.6412529e-01
1.5268834e-01 -1.6290255e-01 8.6428136e-01 -1.3443781e+00
-5.4441714e-01 7.6694155e-01 -6.4281940e-01 -5.4386489e-02
4.2100376e-01 -4.9360882e-02 1.7241013e+00 -1.0033888e+00
6.8484163e-01 1.2764853e+00 1.0017067e+00 3.5702929e-01
-1.1570176e+00 -4.3886781e-01 5.1978940e-01 -1.7534208e-01
-4.7542101e-01 -2.7908370e-01 9.5916212e-01 -3.2713518e-01
1.4841266e+00 -4.7760669e-02 1.1158429e+00 1.1694472e+00
4.2050242e-01 1.1876829e+00 1.5744405e+00 -4.1862610e-01
3.6647499e-01 4.7123465e-01 7.0730501e-01 2.3467886e-01
7.5088549e-01 -3.1794053e-01 -9.1742814e-01 -2.7530006e-01
-8.3548918e-02 -2.6373672e-01 9.1209531e-02 -2.4836607e-01
-8.8640302e-01 1.0371594e+00 -3.6316570e-02 3.0827644e-01
-7.6401806e-01 -1.9447957e-01 3.7603116e-01 3.4669566e-01
9.0677214e-01 1.0871475e+00 -9.0773213e-01 -3.6338970e-01
-8.7749028e-01 3.6926869e-01 1.3730842e+00 7.4110973e-01
7.5460231e-01 5.8375288e-02 -9.7692922e-02 5.6930137e-01
5.8365303e-01 7.2185141e-01 4.9590981e-01 -4.5906946e-01
5.2584028e-01 1.1813604e+00 -9.5225796e-02 -9.5746666e-01
-4.3709567e-01 -3.3492494e-01 2.7127346e-04 2.1856898e-01
2.7303499e-01 -3.7864760e-01 -9.7503930e-01 1.5294352e+00
2.6333961e-01 -3.7573814e-01 3.4433025e-01 5.1845223e-01
6.2977320e-01 7.2685879e-01 3.3136445e-01 1.3147570e-01
1.7065040e+00 -1.2491711e+00 -8.5628712e-01 -1.0761949e+00
6.3812780e-01 -1.1341411e+00 1.2738711e+00 4.5100948e-01
-1.1070045e+00 -1.3911609e-01 -1.3712364e+00 -2.7852672e-01
-9.1334254e-01 1.9036888e-01 1.0714424e+00 7.7690101e-01
-1.0942036e+00 2.5719035e-01 -7.4218750e-01 -5.1521033e-01
7.1332216e-01 2.2443028e-01 -4.3047014e-01 -6.5150887e-02
-9.6572953e-01 9.8442394e-01 -1.3421120e-01 -3.3297214e-01
-3.7505171e-01 -8.8277715e-01 -9.8764306e-01 1.7133459e-02
1.7561026e-01 -8.1284851e-01 1.2999552e+00 -1.2749950e+00
-1.5063305e+00 1.0174906e+00 -6.7923844e-01 -5.3690583e-01
-9.2192002e-02 -8.7412834e-01 -2.1581131e-01 -6.5934531e-02
3.8001943e-01 5.1651508e-01 8.4273493e-01 -1.2558403e+00
-7.3187560e-01 -5.0808120e-01 3.8581103e-01 4.2158380e-01
-5.3674465e-01 5.9001978e-02 -3.3304960e-01 -6.3048744e-01
-3.3929050e-01 -9.1228575e-01 -1.6431381e-01 -4.3150067e-01
-2.5277454e-01 -4.3412185e-01 7.7349216e-01 -6.0488194e-01
1.1704012e+00 -1.9448878e+00 -6.4551353e-02 2.7939588e-01
1.2785081e-02 1.9626993e-01 -5.1010942e-01 4.3085417e-01
-2.1671022e-01 1.6202275e-01 -3.4714362e-01 -4.1963738e-01
8.8972092e-02 8.2111627e-02 -5.5384117e-01 -7.2768532e-02
6.6522074e-01 1.1944957e+00 -6.1668962e-01 -1.6564944e-01
-1.1653543e-01 4.3654048e-01 -6.9178706e-01 3.0734512e-01
-4.0762660e-01 -1.2462411e-01 -5.0185108e-01 8.4725332e-01
6.0422224e-01 -1.6653058e-01 4.3869546e-01 -2.9400611e-01
-5.1063649e-05 9.4758821e-01 -5.7181907e-01 1.4568669e+00
-8.3345127e-01 7.7723759e-01 -1.9465694e-01 -8.3363771e-01
8.3174229e-01 3.3437356e-01 5.0157320e-01 -6.9771856e-01
3.2912979e-01 1.4926636e-01 1.2684116e-01 -2.0695317e-01
7.4687207e-01 -3.4124002e-01 -2.1201973e-01 8.9388770e-01
3.2550195e-01 -4.5516366e-01 4.7275531e-01 1.6037478e-01
9.8201931e-01 3.5600555e-01 7.2240525e-01 -4.3584621e-01
4.5517460e-01 6.6444403e-01 5.7980585e-01 2.2158743e-01
-2.7993035e-01 3.7351868e-01 8.3121735e-01 -4.4161677e-01
-1.0385246e+00 -7.9553729e-01 -1.4075265e-02 1.0467343e+00
-2.0537551e-01 -7.5378990e-01 -7.4183130e-01 -1.3148916e+00
4.0297091e-02 8.7620342e-01 -1.0354248e+00 3.0944817e-02
-6.1830401e-01 -9.5543987e-01 1.5263866e-01 8.1277603e-01
1.3793816e-01 -1.1564019e+00 -3.1230012e-01 7.1894355e-02
4.0726632e-02 -1.3811406e+00 -9.3759514e-02 3.5697833e-01
-8.1217706e-01 -1.1486099e+00 -2.8747201e-01 -7.3144120e-01
5.3799099e-01 2.3458415e-01 1.6898366e+00 -4.6799827e-01
1.2151365e-01 5.6904453e-01 -5.4056752e-01 -1.1845350e+00
-2.1651372e-01 4.3434423e-01 -1.0131830e-01 -2.2256272e-02
1.3486285e+00 -5.8927894e-01 -4.2748582e-01 -2.1051300e-01
-7.2287178e-01 -2.1665366e-01 1.0018814e+00 3.6889547e-01
7.4515969e-01 -3.2897747e-01 4.1721141e-01 -1.5778481e+00
8.5653454e-01 -4.4207329e-01 -3.1363282e-01 9.1288211e-03
-1.1355944e+00 3.5244068e-03 6.0444474e-01 -3.2151312e-02
-1.1643879e+00 -2.1686234e-01 -2.5571415e-01 1.3367167e-01
-3.6959445e-01 7.7025527e-01 -2.8993142e-01 3.9136529e-01
3.2288328e-01 -2.2164616e-01 -2.2138478e-01 -1.2734394e-01
4.9363866e-01 5.6413573e-01 -2.0247176e-01 -2.9352915e-01
8.5346985e-01 7.4943018e-01 -2.2598188e-01 -7.2801518e-01
-1.5657499e+00 -7.0838362e-01 -4.1241243e-01 1.5579905e-01
8.2987833e-01 -1.1852913e+00 -4.9484435e-02 3.3013955e-01
-1.0945864e+00 -1.0257426e-01 -7.1772456e-01 3.6738637e-01
-2.9599121e-01 -4.6341166e-02 -3.8301182e-01 -5.9767807e-01
-7.5429511e-01 -1.0052015e+00 1.0588765e+00 4.6249071e-01
-9.2594057e-01 -1.3541855e+00 6.0580546e-01 5.8791417e-01
5.2875096e-01 -5.7388838e-02 1.0839742e+00 -1.2083178e+00
-8.2113475e-02 -3.2956651e-01 -3.4488019e-02 4.4589564e-01
2.8371540e-01 -7.8679360e-02 -1.4604315e+00 7.1797878e-03
1.9168870e-01 -3.0753893e-01 6.9344229e-01 5.3241080e-01
2.5751385e-01 -2.3142898e-01 -2.1743110e-01 3.8943768e-01
1.3570139e+00 1.8713275e-01 4.5891091e-01 8.0095404e-01
6.2770802e-01 7.4891257e-01 9.1883850e-01 8.0807013e-03
6.3723922e-01 2.5369325e-01 9.4548345e-02 3.4692079e-02
-1.6500147e-01 -9.7220466e-02 1.0072176e+00 1.4178749e+00
4.4330455e-02 -2.0703964e-01 -9.5814651e-01 1.0015627e+00
-1.5336632e+00 -4.7255683e-01 -1.1632852e-01 1.5225322e+00
1.0233370e+00 2.3966299e-01 3.1985074e-02 3.7025504e-02
-5.0853666e-02 6.9065332e-01 -3.5310277e-01 -1.1997526e+00
-3.7679112e-01 7.3753119e-01 2.1890786e-01 4.2968383e-01
-1.2692981e+00 1.0536629e+00 6.5167179e+00 4.8169968e-01
-9.7083819e-01 1.7493133e-01 4.9343064e-01 -6.9054626e-03
-7.6819658e-01 2.8663749e-01 -8.6883634e-01 -4.0310804e-02
1.1758304e+00 -2.9459080e-01 -7.4546278e-02 1.2996227e+00
-1.1773586e-01 -3.4851581e-02 -1.1511464e+00 3.7692603e-01
4.7478378e-01 -1.0778399e+00 6.4309108e-01 1.0421004e-01
9.8862904e-01 1.3138522e-01 1.7200209e-01 5.0655532e-01
4.6106935e-01 -9.1113138e-01 3.6997676e-01 1.4685565e-01
3.1363145e-01 -8.0452144e-01 1.1913979e+00 -3.9078158e-01
-1.1223115e+00 3.1277728e-01 -1.8981764e-01 -1.4410083e-01
2.0424628e-01 7.9815429e-01 -8.2775313e-01 6.0993141e-01
5.3198195e-01 1.0611159e+00 -5.0290114e-01 4.2380372e-01
-7.9317462e-01 1.0482466e+00 -8.4800404e-03 -1.1615482e-01
4.6664569e-01 -5.5787390e-01 7.1849906e-01 1.2730383e+00
4.1518927e-02 -3.2878765e-01 -1.9671306e-01 6.2397158e-01
-2.2147961e-01 4.7685185e-01 -7.3257118e-01 -8.1098896e-01
1.1019139e-02 1.5286168e+00 -4.6599820e-01 -4.1912037e-01
-9.4976479e-01 7.0509022e-01 2.5233141e-01 2.4522321e-01
-5.7549369e-01 -4.9901599e-01 1.0979563e+00 1.4083302e-01
5.3669089e-01 1.2656263e-01 -9.1152805e-01 -1.3630266e+00
1.3273735e-01 -1.2431974e+00 3.5397613e-01 -5.8129513e-01
-1.5631481e+00 7.4026603e-01 -3.8784409e-01 -1.0896071e+00
-2.4620377e-01 -1.0440378e+00 -8.0294901e-01 8.9553946e-01
-2.1203723e+00 -1.5697647e+00 5.5899505e-02 1.7741388e-01
7.6038879e-01 -2.7176869e-01 9.4022876e-01 -7.0604663e-03
-3.9934939e-01 3.8676530e-01 -3.8413221e-01 2.6711228e-01
8.3270580e-01 -1.7010258e+00 9.1712594e-01 9.6275419e-01
4.0564403e-01 8.3129209e-01 5.9795773e-01 -6.6758978e-01
-1.4130424e+00 -9.8735070e-01 1.4812171e+00 -8.7378883e-01
9.7466391e-01 -2.5670737e-01 -7.0214289e-01 1.0465864e+00
8.5480928e-01 -5.0329119e-01 1.1254655e+00 8.1783724e-01
-7.3714310e-01 5.6591049e-02 -6.9858921e-01 5.7098413e-01
4.1543266e-01 -6.4006150e-01 -1.1957579e+00 -6.1660204e-02
6.2454617e-01 -5.7068501e-02 -8.1973982e-01 4.0103087e-01
6.0161942e-01 -7.8963870e-01 7.5534654e-01 -8.4089494e-01
1.0118961e+00 -4.4383958e-02 -1.7298487e-01 -1.6175991e+00
6.7141354e-02 -1.9886668e-01 -1.6181152e-01 1.2650782e+00
9.8623824e-01 -7.7042967e-01 7.9557812e-01 6.1117607e-01
1.5351953e-01 -9.3756533e-01 -3.0626592e-01 -4.4262961e-01
4.0831044e-01 -6.4176399e-01 6.1085826e-01 1.0359186e+00
4.2270806e-02 1.0887902e+00 2.1725895e-01 2.4905518e-02
1.9452973e-01 6.4153636e-01 8.6465496e-01 -1.2627884e+00
-1.9721568e-01 -5.1571441e-01 -7.0422471e-02 -6.7350960e-01
1.4951058e-01 -8.7208784e-01 -1.3023011e-01 -1.8223433e+00
2.8011584e-01 -2.0094255e-01 -3.7163544e-01 6.5887940e-01
-5.2048403e-01 1.5974031e-01 1.2915245e-01 -7.5652681e-02
-3.8485780e-01 5.6632310e-01 8.3381397e-01 -1.9484389e-01
-2.1926230e-01 1.0234378e-02 -1.1961647e+00 1.2559067e+00
8.6844242e-01 -7.5664115e-01 -4.4836512e-01 -4.2538947e-01
8.4496552e-01 -8.4255904e-01 -3.9392826e-01 -6.0463572e-01
-2.1432932e-01 8.6626999e-02 9.2803061e-02 -4.9734816e-01
3.8023591e-01 -7.3934829e-01 -7.8770298e-01 8.6320259e-02
-1.9770165e-01 3.5445061e-01 5.2868044e-01 4.2579496e-01
-6.8122393e-01 -1.8061352e-01 3.2126090e-01 6.7988709e-02
-5.3313714e-01 1.9804720e-02 -3.8701367e-01 4.9995962e-01
7.3422003e-01 -1.3118520e-01 -2.9541504e-01 -2.9763812e-01
-4.3994942e-01 -9.1780104e-02 3.4341693e-01 4.6621597e-01
3.0055583e-01 -1.0431415e+00 -4.2826468e-01 1.6656239e-01
2.3512733e-01 -4.7891282e-02 -1.4998096e-01 9.7064185e-01
-2.3096883e-01 6.2127411e-01 -1.2792823e-01 -1.4562307e-01
-1.0600240e+00 1.0003045e-01 -1.4939852e-01 -9.4064492e-01
-1.0941588e-01 7.8358299e-01 1.8036726e-01 -8.3292866e-01
-2.1448462e-01 -3.0161470e-01 -6.7166275e-01 6.3968045e-01
3.8228467e-01 -4.2590078e-02 1.1973572e-01 -6.1574709e-01
-4.0104514e-01 7.4385715e-01 -4.0600389e-01 -3.2743907e-01
1.6531763e+00 -3.2545026e-02 -2.0636949e-01 7.0480448e-01
1.0457840e+00 6.7549098e-01 -7.2130948e-01 -2.2036339e-01
2.7842686e-01 -1.7871980e-01 -3.8590945e-02 -1.0181886e+00
-9.9550742e-01 1.0007904e+00 -1.0776690e-02 1.2198691e-01
9.7358131e-01 -1.2537991e-01 1.0110884e+00 3.5141590e-01
-5.9769014e-03 -1.2580770e+00 -8.0927778e-03 8.3338165e-01
5.9329915e-01 -1.3704486e+00 2.9942626e-01 -3.7532699e-01
-1.1346099e+00 9.3708760e-01 5.6757915e-01 -2.4919751e-01
7.4900222e-01 5.7944947e-01 6.9867939e-01 -4.1772953e-01
-1.0062288e+00 -2.9897398e-01 4.9347261e-01 5.6186849e-01
7.6591635e-01 -1.6030202e-02 -5.4774201e-01 1.2092944e+00
-4.9674562e-01 -2.3030971e-01 4.3566114e-01 1.1660342e+00
-6.8681352e-02 -1.3160270e+00 3.5919100e-02 5.0260317e-01
-7.9667908e-01 -5.2233797e-01 -9.0932035e-01 1.0467769e+00
-2.5502419e-01 8.9891636e-01 2.9707782e-02 -2.2860429e-01
4.9668980e-01 5.0332886e-01 -2.2889366e-03 -8.6487073e-01
-1.0220622e+00 -3.5383165e-02 6.2588567e-01 -7.2408128e-01
-8.3391237e-01 -8.0281013e-01 -9.5522231e-01 4.4562970e-03
-1.8251550e-01 3.1580043e-01 8.6762452e-01 1.2071636e+00
5.2290422e-01 4.4898582e-01 4.3846092e-01 -1.5114084e-01
-1.0152953e-01 -1.0386602e+00 -3.0283561e-01 5.5067039e-01
-1.6811613e-02 -2.0986018e-01 -5.0345063e-01 1.2979758e-02] | [11.424331665039062, 6.715104103088379] |
c708f98c-d8e5-4d67-9c1f-30124bc3dd2e | nautilus-a-versatile-voice-cloning-system | 2005.11004 | null | https://arxiv.org/abs/2005.11004v2 | https://arxiv.org/pdf/2005.11004v2.pdf | NAUTILUS: a Versatile Voice Cloning System | We introduce a novel speech synthesis system, called NAUTILUS, that can generate speech with a target voice either from a text input or a reference utterance of an arbitrary source speaker. By using a multi-speaker speech corpus to train all requisite encoders and decoders in the initial training stage, our system can clone unseen voices using untranscribed speech of target speakers on the basis of the backpropagation algorithm. Moreover, depending on the data circumstance of the target speaker, the cloning strategy can be adjusted to take advantage of additional data and modify the behaviors of text-to-speech (TTS) and/or voice conversion (VC) systems to accommodate the situation. We test the performance of the proposed framework by using deep convolution layers to model the encoders, decoders and WaveNet vocoder. Evaluations show that it achieves comparable quality with state-of-the-art TTS and VC systems when cloning with just five minutes of untranscribed speech. Moreover, it is demonstrated that the proposed framework has the ability to switch between TTS and VC with high speaker consistency, which will be useful for many applications. | ['Hieu-Thi Luong', 'Junichi Yamagishi'] | 2020-05-22 | null | null | null | null | ['voice-cloning'] | ['speech'] | [ 2.00234458e-01 2.21898034e-01 2.91926503e-01 -3.01190257e-01
-5.29193997e-01 -4.84605551e-01 5.53656995e-01 -6.54978871e-01
-2.20055610e-01 5.97908199e-01 5.96496947e-02 -4.70645517e-01
5.27434528e-01 -4.53010827e-01 -7.10440338e-01 -8.44205916e-01
3.91014129e-01 3.94308209e-01 2.32406214e-01 -3.69309276e-01
-1.71302348e-01 5.88929534e-01 -1.77549672e+00 2.72669226e-01
6.49560392e-01 8.52121115e-01 7.13760257e-01 1.02803349e+00
-2.16749787e-01 3.96389574e-01 -1.18760824e+00 -1.19511694e-01
2.67245829e-01 -5.96611857e-01 -4.66567516e-01 1.73439994e-01
2.64320672e-01 -2.44581178e-01 -4.92561609e-01 8.71551633e-01
9.89561975e-01 1.83089033e-01 4.26279485e-01 -8.75532448e-01
-6.49086356e-01 8.26587856e-01 2.72111207e-01 1.36567503e-01
1.72645569e-01 2.24953666e-01 5.49711943e-01 -8.15697551e-01
4.08886045e-01 1.27146745e+00 1.83070108e-01 9.08695936e-01
-1.02753794e+00 -9.38418031e-01 4.40366492e-02 6.43727705e-02
-1.43107033e+00 -1.30821586e+00 7.62970984e-01 -2.94392824e-01
1.03072345e+00 5.01218915e-01 4.27570134e-01 1.35836160e+00
1.58903003e-01 4.82125789e-01 6.13263190e-01 -5.66181600e-01
2.22671539e-01 6.35554731e-01 -4.72769886e-01 4.38831270e-01
-5.19724190e-01 3.65925133e-01 -5.47317266e-01 1.80314168e-01
6.03604913e-01 -6.21505976e-01 -7.04904974e-01 1.85270488e-01
-1.16548181e+00 4.80575681e-01 1.90470591e-01 4.11072701e-01
-1.86142698e-01 -4.96074674e-04 3.43674421e-01 4.85832721e-01
2.24080265e-01 2.82559097e-01 -3.63710731e-01 -6.85669556e-02
-1.03904307e+00 -1.80665150e-01 8.29473197e-01 1.42889011e+00
2.80036032e-01 9.42162693e-01 -2.08489195e-01 1.12019432e+00
2.00951859e-01 5.07906854e-01 1.16825080e+00 -5.60307801e-01
6.37857199e-01 -1.13972433e-01 -2.61322670e-02 -2.18802989e-01
-7.34212995e-02 -6.96277678e-01 -7.37946987e-01 1.06570296e-01
-1.60120577e-01 -5.02124250e-01 -1.15144062e+00 1.63412261e+00
3.33153069e-01 3.43244463e-01 4.59582150e-01 8.28688264e-01
9.41281021e-01 1.04923654e+00 -4.71673310e-01 -3.91296715e-01
9.88872170e-01 -1.20968068e+00 -9.87436235e-01 -3.09978783e-01
1.53359890e-01 -1.01802468e+00 1.04380083e+00 1.49711877e-01
-1.01810372e+00 -1.10985529e+00 -1.23647523e+00 2.27311805e-01
-3.41112196e-01 5.18474936e-01 -1.58067062e-01 9.28005040e-01
-1.15971708e+00 2.99497932e-01 -5.64561725e-01 -2.61733949e-01
-4.00394440e-01 4.02291387e-01 -4.32355218e-02 5.16869664e-01
-1.48960853e+00 7.17976928e-01 4.60374564e-01 3.61331344e-01
-1.16713703e+00 -5.12300193e-01 -6.95461690e-01 2.62197256e-01
-1.39163081e-02 -5.61770320e-01 1.57811940e+00 -1.17056131e+00
-2.59557271e+00 2.62877911e-01 -2.19378963e-01 -3.56726319e-01
5.77884853e-01 2.08778799e-01 -1.06259048e+00 -8.59894964e-04
-1.75635606e-01 5.53669333e-01 1.37169218e+00 -8.49431157e-01
-6.81032538e-01 1.48055881e-01 -4.66629833e-01 3.09000224e-01
-2.21106648e-01 1.76206359e-03 -3.72513741e-01 -6.49766743e-01
-5.91773912e-02 -9.60709929e-01 3.23776841e-01 -3.74707371e-01
-7.44404674e-01 7.31470361e-02 9.67188776e-01 -7.33327806e-01
9.93455410e-01 -2.50166965e+00 2.66794562e-01 -1.41937256e-01
-4.04849619e-01 4.47331369e-01 -2.52165496e-01 4.47981715e-01
-1.24241158e-01 8.31188576e-04 -2.12680977e-02 -5.37208438e-01
-1.32248357e-01 7.88512528e-02 -5.22682846e-01 2.52708822e-01
1.50301158e-01 3.70827764e-01 -3.99906337e-01 -1.81788281e-01
3.30864817e-01 5.68160057e-01 -3.45168650e-01 6.29504025e-01
-1.57451093e-01 6.87582254e-01 1.13711528e-01 1.38911590e-01
3.31104517e-01 3.18611681e-01 1.14077643e-01 2.51824018e-02
-3.06363910e-01 9.08639073e-01 -1.24003482e+00 1.27989745e+00
-8.34794998e-01 1.05705655e+00 4.70444977e-01 -5.82015395e-01
1.12556827e+00 1.00600410e+00 -6.22317344e-02 -3.85599762e-01
7.69012123e-02 3.27990711e-01 3.76186728e-01 -4.02154595e-01
5.02410650e-01 -2.84120321e-01 2.10791767e-01 1.88476816e-01
3.30673575e-01 -4.77695704e-01 -1.95419416e-01 -2.80054241e-01
5.93221784e-01 -2.62535304e-01 -1.35083959e-01 -6.59384206e-02
7.95269072e-01 -4.36310142e-01 3.17893118e-01 4.00826901e-01
-7.04313349e-03 6.37011707e-01 -8.43994170e-02 1.45205125e-01
-1.02574182e+00 -1.25201368e+00 -1.90791890e-01 1.15393889e+00
-1.87570930e-01 -4.25547324e-02 -8.58222961e-01 -2.33641993e-02
-3.44665945e-01 1.14319718e+00 -1.15902565e-01 -1.40102282e-01
-5.41811287e-01 -2.87701219e-01 8.54249597e-01 1.90440044e-01
5.53060830e-01 -1.22962284e+00 -7.03700706e-02 5.19472897e-01
-2.01013014e-01 -1.20157707e+00 -8.60110283e-01 3.50379020e-01
-5.45309246e-01 -3.69508624e-01 -8.33900928e-01 -1.19374633e+00
3.86369675e-01 4.63046432e-02 4.31948751e-01 -3.65315616e-01
2.16684923e-01 9.46824849e-02 -1.48759499e-01 -3.03058624e-01
-1.38625348e+00 4.03119057e-01 3.95603210e-01 2.72729188e-01
-1.96233317e-01 -4.95976061e-01 -1.66177645e-01 3.71684194e-01
-8.06637228e-01 2.55681485e-01 4.34810936e-01 7.73265302e-01
2.54541755e-01 3.27372223e-01 8.72350514e-01 -4.45188314e-01
7.71364689e-01 -2.30126128e-01 -8.06851506e-01 1.73261106e-01
-3.88675421e-01 5.70091084e-02 1.12066209e+00 -7.30874062e-01
-1.25159764e+00 -2.12662760e-02 -4.56123888e-01 -5.29501617e-01
-1.94157064e-01 1.79177403e-01 -4.84796345e-01 1.72864616e-01
4.99192238e-01 7.88183212e-01 -9.31125656e-02 -5.18560469e-01
4.01149571e-01 1.62432706e+00 7.33797610e-01 -8.41948017e-02
7.78868020e-01 -1.48483559e-01 -7.19551742e-01 -1.28699017e+00
-4.52636369e-02 -2.40191624e-01 -7.06479251e-01 -2.20382094e-01
5.99923968e-01 -1.02852833e+00 -4.84501570e-01 7.37922251e-01
-1.45009470e+00 -3.24919939e-01 -2.09309570e-02 6.82890892e-01
-4.59871978e-01 7.00151101e-02 -7.03134716e-01 -7.49646783e-01
-4.31003809e-01 -1.49501777e+00 9.85427737e-01 1.10364616e-01
7.97634125e-02 -6.88528001e-01 -1.93561852e-01 3.07923406e-01
6.50777578e-01 -4.42546487e-01 8.67870271e-01 -7.03847289e-01
-6.86100543e-01 3.08304210e-03 6.08280122e-01 8.04464579e-01
5.56383073e-01 2.32680380e-01 -1.36280513e+00 -3.96973997e-01
1.45275995e-01 5.51736839e-02 4.86163318e-01 1.16623141e-01
7.35906422e-01 -5.88176548e-01 -7.58572817e-02 6.10866129e-01
8.83268297e-01 8.02957714e-01 6.23771191e-01 -1.42848492e-01
4.94903564e-01 4.02807921e-01 -3.98121737e-02 1.91733539e-01
-8.86749185e-04 7.66619503e-01 1.11188605e-01 5.36553934e-02
-3.82041723e-01 -2.50594825e-01 7.96968639e-01 1.52263713e+00
3.52650732e-01 -5.68199217e-01 -4.11268204e-01 4.17762965e-01
-1.08014166e+00 -1.04026377e+00 1.84231341e-01 2.33559966e+00
1.06576848e+00 1.79136187e-01 -1.74591988e-01 2.40672931e-01
1.19797862e+00 2.57131249e-01 -6.71648085e-01 -9.41266716e-01
1.61235183e-01 2.84630686e-01 3.08478773e-01 7.50687361e-01
-7.37758994e-01 1.11831868e+00 6.42797852e+00 7.04859674e-01
-1.93239748e+00 -2.70514507e-02 1.92518637e-01 -2.66895205e-01
-1.50820225e-01 -5.65372169e-01 -8.37462723e-01 5.71242869e-01
1.50230443e+00 -2.49743283e-01 1.05524302e+00 7.85804868e-01
4.35109675e-01 6.47152126e-01 -1.27599764e+00 8.27666700e-01
-4.25521359e-02 -1.11866164e+00 2.17270795e-02 -2.66178101e-01
4.65851426e-01 1.05422445e-01 2.47091830e-01 5.00905514e-01
-5.50658777e-02 -8.79993498e-01 1.15875244e+00 9.11145732e-02
1.10432911e+00 -5.18803716e-01 4.73923445e-01 4.41916496e-01
-1.21538460e+00 4.69563678e-02 -3.31297904e-01 1.70916021e-01
4.42413166e-02 2.24325925e-01 -1.63780868e+00 3.16010416e-01
2.90777355e-01 5.58571927e-02 -1.71225652e-01 7.12947130e-01
-2.97756433e-01 8.50107074e-01 -2.12202981e-01 -1.97556034e-01
1.60166934e-01 1.35140106e-01 6.98623657e-01 1.21079230e+00
6.15647435e-01 -2.53670156e-01 -2.20511556e-01 8.36636066e-01
-2.73284882e-01 2.18736559e-01 -3.17160398e-01 -1.83510825e-01
8.69843245e-01 8.70431781e-01 -3.80287468e-02 -3.10121894e-01
-2.42264256e-01 1.21670282e+00 3.86994183e-02 7.41354644e-01
-8.97053182e-01 -6.54230535e-01 8.18386555e-01 -5.79553358e-02
5.87054908e-01 -1.99534535e-01 1.01285852e-01 -8.83099258e-01
6.22112341e-02 -9.44864810e-01 -3.62778902e-01 -1.01519072e+00
-7.80509770e-01 1.26157510e+00 -4.70590413e-01 -1.24688375e+00
-6.67859018e-01 -4.04804140e-01 -7.08325326e-01 1.38295615e+00
-1.39454293e+00 -6.67027831e-01 4.22413051e-02 5.40791333e-01
8.97293806e-01 -5.07782400e-01 8.36481690e-01 2.13020995e-01
-6.88416481e-01 6.93728268e-01 4.31809634e-01 8.66340175e-02
7.24131286e-01 -9.47360516e-01 7.15661347e-01 8.35272491e-01
1.81727633e-01 4.00833964e-01 7.40709007e-01 -3.70146275e-01
-1.12843633e+00 -1.31980431e+00 8.18269014e-01 3.22095722e-01
2.81151533e-01 -7.83476949e-01 -8.27072680e-01 7.03475356e-01
4.64860141e-01 -3.43565643e-01 6.43324375e-01 -2.99913228e-01
2.00261772e-02 -4.11988288e-01 -8.79537106e-01 7.43746579e-01
6.80640817e-01 -8.22625339e-01 -7.21467495e-01 1.91805273e-01
1.27288043e+00 -7.23315358e-01 -5.26374340e-01 -2.19255798e-02
4.64338213e-01 -7.97808766e-01 6.15898252e-01 -3.32908958e-01
9.89012495e-02 -3.52279931e-01 -1.58933461e-01 -1.78794527e+00
-3.38899583e-01 -9.85798061e-01 3.24482381e-01 1.36679542e+00
7.46959805e-01 -8.40328991e-01 3.35935235e-01 3.18109602e-01
-6.78313017e-01 -1.38514444e-01 -1.30601609e+00 -9.55057025e-01
-6.30355328e-02 -4.34591949e-01 1.00206387e+00 6.46578610e-01
-6.84703141e-02 5.90187252e-01 -4.10251707e-01 7.43124723e-01
1.22085614e-02 -5.84600195e-02 7.87821591e-01 -6.70998633e-01
-5.08217692e-01 -2.27548197e-01 -1.42666459e-01 -1.23641884e+00
2.13303313e-01 -9.63598549e-01 4.71607536e-01 -1.18918407e+00
-8.39118481e-01 -2.41269901e-01 -1.16150767e-01 1.34158224e-01
-3.17589231e-02 -3.04906964e-01 2.86288172e-01 6.86189458e-02
3.35695475e-01 9.36566710e-01 1.33229101e+00 -2.73127437e-01
-6.45044684e-01 3.56750727e-01 -2.65959203e-01 2.34967515e-01
9.23693478e-01 -3.27891529e-01 -5.98692119e-01 -5.26363194e-01
-6.89359844e-01 4.81241375e-01 8.23462829e-02 -1.23756444e+00
3.14504355e-01 3.47355604e-02 1.35307074e-01 -3.90394628e-01
5.66471398e-01 -8.18555355e-01 4.91657466e-01 3.94936979e-01
-5.58737934e-01 -1.85087904e-01 4.10084337e-01 4.02490497e-01
-2.80294210e-01 -4.05077308e-01 8.67070794e-01 9.66774896e-02
-2.99260288e-01 -3.17993499e-02 -9.23173904e-01 -2.76799172e-01
6.85366809e-01 -1.52895957e-01 -2.34852776e-01 -6.40907526e-01
-7.39657462e-01 -1.65322736e-01 4.42852080e-02 6.23025239e-01
6.15493655e-01 -1.30359387e+00 -6.79272473e-01 7.71301389e-01
-2.60012168e-02 -2.44064301e-01 1.42219007e-01 1.78301170e-01
-4.27621782e-01 7.21950054e-01 1.46555714e-02 -5.35381258e-01
-1.13088322e+00 5.96554101e-01 7.78892875e-01 4.64745253e-01
-4.99616057e-01 8.02411675e-01 3.03528309e-01 -6.53954029e-01
5.24514616e-01 -6.49694681e-01 7.60685429e-02 -3.51825386e-01
4.23467100e-01 8.72920081e-02 3.06399912e-01 -6.57870352e-01
-1.93458244e-01 2.48832740e-02 8.78002346e-02 -5.00800133e-01
8.32847118e-01 -2.06336543e-01 3.50705177e-01 6.05176449e-01
1.18941724e+00 1.93218589e-01 -1.12227571e+00 -3.36551927e-02
-5.54559946e-01 -1.36494040e-01 1.81636781e-01 -8.52101207e-01
-1.19351649e+00 9.92788374e-01 5.42635918e-01 2.22566307e-01
1.03030169e+00 -2.79788792e-01 9.05438840e-01 3.72184396e-01
1.91191435e-01 -1.13607645e+00 -2.30688050e-01 3.82250309e-01
1.19084942e+00 -7.18554378e-01 -7.52646387e-01 -3.64594698e-01
-6.43885255e-01 1.30523396e+00 4.56499815e-01 2.64821202e-01
6.05638266e-01 4.52035457e-01 4.64436561e-01 5.17262936e-01
-1.03274632e+00 4.67344634e-02 1.90650031e-01 6.08982861e-01
3.95862222e-01 2.24939138e-01 2.18812004e-01 3.53658050e-01
-8.37332547e-01 -2.29114667e-01 7.05933154e-01 3.47777605e-01
-5.30348361e-01 -9.80810404e-01 -6.12925649e-01 1.31814837e-01
-2.30797902e-01 -3.11998397e-01 -3.05901229e-01 5.46133637e-01
3.88825722e-02 1.26584709e+00 2.09038079e-01 -6.13220572e-01
2.85635203e-01 4.96405274e-01 2.23559469e-01 -7.99760818e-01
-7.73367703e-01 3.75639975e-01 1.61091715e-01 9.55808982e-02
4.04586382e-02 -4.56272185e-01 -1.20406783e+00 -1.19195499e-01
-5.64742446e-01 2.49004662e-01 9.35137272e-01 8.61886024e-01
3.98919672e-01 1.00327122e+00 1.01449251e+00 -6.47924781e-01
-6.84172511e-01 -1.33061945e+00 -7.36841142e-01 -1.62161171e-01
7.04202652e-01 -2.69377679e-01 -5.49323380e-01 3.66080910e-01] | [14.931519508361816, 6.626309871673584] |
0935b6aa-a2a2-4366-a6d8-e20ea8702997 | a-novel-clustering-based-algorithm-for | 2110.06996 | null | https://arxiv.org/abs/2110.06996v2 | https://arxiv.org/pdf/2110.06996v2.pdf | A Novel Clustering-Based Algorithm for Continuous and Non-invasive Cuff-Less Blood Pressure Estimation | Extensive research has been performed on continuous, non-invasive, cuffless blood pressure (BP) measurement using artificial intelligence algorithms. This approach involves extracting certain features from physiological signals like ECG, PPG, ICG, BCG, etc. as independent variables and extracting features from Arterial Blood Pressure (ABP) signals as dependent variables, and then using machine learning algorithms to develop a blood pressure estimation model based on these data. The greatest challenge of this field is the insufficient accuracy of estimation models. This paper proposes a novel blood pressure estimation method with a clustering step for accuracy improvement. The proposed method involves extracting Pulse Transit Time (PTT), PPG Intensity Ratio (PIR), and Heart Rate (HR) features from Electrocardiogram (ECG) and Photoplethysmogram (PPG) signals as the inputs of clustering and regression, extracting Systolic Blood Pressure (SBP) and Diastolic Blood Pressure (DBP) features from ABP signals as dependent variables, and finally developing regression models by applying Gradient Boosting Regression (GBR), Random Forest Regression (RFR), and Multilayer Perceptron Regression (MLP) on each cluster. The method was implemented using the MIMICII dataset with the silhouette criterion used to determine the optimal number of clusters. The results showed that because of the inconsistency, high dispersion, and multi-trend behavior of the extracted features vectors, the accuracy can be significantly improved by running a clustering algorithm and then developing a regression model on each cluster, and finally weighted averaging of the results based on the error of each cluster. When implemented with 5 clusters and GBR, this approach yielded an MAE of 2.56 for SBP estimates and 2.23 for DBP estimates, which were significantly better than the best results without clustering (DBP: 6.27, SBP: 6.36). | ['Elham Akhondzadeh Noughabi', 'Reza Baradaran Kazemzadeh', 'Ali Farki'] | 2021-10-13 | null | null | null | null | ['blood-pressure-estimation'] | ['medical'] | [-3.19310762e-02 -2.54413456e-01 1.02027640e-01 -5.22578061e-01
-8.90618339e-02 -1.05228402e-01 -5.26489206e-02 2.57881850e-01
-3.53462696e-01 9.96923983e-01 6.25932887e-02 -5.66268861e-01
-4.44968164e-01 -7.65917301e-01 -1.08930632e-01 -7.81431019e-01
-4.99908775e-01 3.42159897e-01 -6.80561513e-02 2.14688480e-01
4.05450612e-01 5.86189687e-01 -1.30683005e+00 -1.50555998e-01
1.25725436e+00 1.09384882e+00 -4.03214633e-01 8.66571546e-01
-2.40723696e-02 6.74643099e-01 -6.90191746e-01 1.25928953e-01
3.15365195e-01 -6.90742791e-01 -2.55487829e-01 -2.31919177e-02
4.79374975e-02 -7.84194767e-02 1.10531054e-01 4.96213347e-01
9.03602421e-01 1.08773313e-01 5.58291733e-01 -9.26822543e-01
-2.49520212e-01 5.07475138e-01 -6.71354055e-01 2.43632257e-01
-2.12774854e-02 6.38633296e-02 1.88120931e-01 -3.95196170e-01
-4.72709611e-02 8.52912068e-01 8.38536441e-01 6.69528097e-02
-1.55693388e+00 -5.65158188e-01 -3.66798133e-01 2.70538568e-01
-1.51653194e+00 -2.71891594e-01 6.83620751e-01 -5.86830556e-01
5.13159275e-01 6.99535191e-01 1.05845439e+00 7.93870240e-02
3.26807767e-01 -1.19634695e-01 1.67834723e+00 -7.46432543e-01
3.73826146e-01 7.00130463e-01 6.60367012e-01 4.04157162e-01
4.25955653e-01 1.03937216e-01 2.27338016e-01 -3.86849225e-01
6.66580439e-01 3.62905152e-02 -2.75805086e-01 1.44168526e-01
-6.20564103e-01 7.25835323e-01 1.47818416e-01 4.50473726e-01
-6.37222350e-01 -5.56050301e-01 2.25326642e-01 3.65954220e-01
3.19861591e-01 5.09023726e-01 -4.73169863e-01 -5.42360544e-02
-1.10271180e+00 1.21439599e-01 1.12360740e+00 3.13267857e-01
5.86312592e-01 -5.32756336e-02 -2.38795996e-01 9.78014112e-01
4.14828956e-01 5.98386109e-01 4.63820547e-01 -8.94797385e-01
1.12479366e-01 6.81910098e-01 2.63562679e-01 -1.41347337e+00
-7.10827410e-01 -3.08660299e-01 -9.86640513e-01 2.18517527e-01
6.49461210e-01 -6.88098431e-01 -6.52811468e-01 1.17791426e+00
2.52438366e-01 -5.53605892e-02 -2.07723692e-01 9.41478848e-01
6.88059866e-01 3.67263258e-01 4.63796049e-01 -7.15616643e-01
1.31912863e+00 -4.03524816e-01 -6.89160883e-01 2.26849943e-01
2.73546875e-01 -2.83784777e-01 3.86466593e-01 3.57209414e-01
-6.42088592e-01 -1.02955508e+00 -8.12592328e-01 5.52052975e-01
-3.11574996e-01 2.24391386e-01 3.60598475e-01 9.55382049e-01
-7.18224704e-01 9.63713825e-01 -7.48315871e-01 -2.28651673e-01
9.66680422e-02 3.53680789e-01 -9.67579782e-02 2.39494964e-01
-1.16129565e+00 1.22078574e+00 3.65222283e-02 4.28776920e-01
3.67762327e-01 -7.46145666e-01 -4.82453436e-01 -1.57624111e-01
-4.28586870e-01 -5.09466410e-01 6.70033693e-02 -5.10904670e-01
-1.55889606e+00 4.60688382e-01 -1.17239252e-01 -3.19168866e-01
4.58903700e-01 -2.88200583e-02 -7.02201545e-01 1.96346551e-01
-1.89998239e-01 4.88568656e-02 6.44187391e-01 -8.19670200e-01
-2.99232036e-01 -9.81038272e-01 -7.67223775e-01 -5.01259267e-02
1.36087134e-01 2.42888063e-01 2.32071251e-01 1.76831722e-01
4.97532010e-01 -6.76402330e-01 -6.59915328e-01 -4.25524205e-01
-1.39514893e-01 -2.99667828e-02 3.19505244e-01 -1.40785158e+00
1.53563237e+00 -1.99393928e+00 1.38361789e-02 8.86321127e-01
3.39092851e-01 3.47046435e-01 5.79593003e-01 2.94741504e-02
-1.95955798e-01 9.16446820e-02 -1.70667797e-01 3.87207538e-01
-2.40688384e-01 1.10062003e-01 1.93115026e-01 4.51022595e-01
8.65438059e-02 5.46866059e-01 -5.11687994e-01 -6.64659083e-01
7.85390079e-01 4.43753809e-01 -1.97113752e-01 2.15228066e-01
4.64102060e-01 7.97896802e-01 -9.36230794e-02 3.02788585e-01
7.95653045e-01 1.85223222e-01 2.95585096e-01 -3.45842719e-01
-4.62516427e-01 -2.50737965e-01 -1.33926952e+00 6.34681940e-01
-9.39417481e-02 5.61009407e-01 -1.06690854e-01 -1.21068537e+00
1.88644385e+00 5.10229111e-01 7.66069233e-01 -5.03943145e-01
1.94817230e-01 1.43014625e-01 5.73566496e-01 -8.99813712e-01
-2.74814039e-01 -2.54591465e-01 4.12191898e-01 1.84869558e-01
-2.70666718e-01 1.80293024e-01 1.86057687e-01 -3.16972911e-01
6.92380846e-01 -1.25691637e-01 4.21655387e-01 -3.49855393e-01
8.58372867e-01 -9.90409851e-02 4.41404492e-01 7.78499901e-01
-5.87590456e-01 3.27702135e-01 4.66528058e-01 -7.81579256e-01
-9.51010942e-01 -8.63787234e-01 -7.28505671e-01 3.37413043e-01
-2.11576238e-01 4.57201910e-04 -2.59010285e-01 8.28397796e-02
4.17207360e-01 5.34596920e-01 -3.74033839e-01 8.46945271e-02
-6.03438139e-01 -1.34125555e+00 2.55607158e-01 3.30631375e-01
5.55933595e-01 -8.74891818e-01 -6.27420545e-01 5.47160625e-01
3.45522985e-02 -6.93134069e-01 6.56869709e-01 6.48254231e-02
-1.27774036e+00 -9.90490556e-01 -5.73092580e-01 -2.01240599e-01
4.58608687e-01 -5.42667031e-01 8.38702500e-01 -3.86973657e-02
-6.32063210e-01 -8.61934647e-02 -1.26825154e-01 -4.80273783e-01
-2.94366688e-01 -3.07138860e-01 7.18682259e-02 1.29334226e-01
7.61886537e-01 -9.76639271e-01 -7.60014951e-01 2.91968346e-01
1.05368730e-03 -2.95796365e-01 3.47713798e-01 3.10790360e-01
3.82585704e-01 -1.38538554e-01 8.38468015e-01 -6.49084866e-01
7.43147254e-01 -5.22829294e-01 -4.85381752e-01 2.41862200e-02
-9.30829167e-01 -5.37575006e-01 6.90057218e-01 -2.85625547e-01
-5.78603923e-01 -1.94881096e-01 -1.11391384e-03 -1.16615243e-01
-4.14402962e-01 2.84819514e-01 2.68719494e-01 4.05205563e-02
1.10516131e+00 3.67904544e-01 6.06375515e-01 -4.33153391e-01
-3.77880558e-02 1.10831487e+00 3.22262853e-01 -4.83304501e-01
2.67301768e-01 2.98759025e-02 2.72395402e-01 -9.51390028e-01
-3.56524616e-01 -3.33691448e-01 -9.93809342e-01 -4.65385258e-01
8.29934716e-01 -6.73004925e-01 -1.05524838e+00 3.30906391e-01
-3.35369110e-01 2.11835548e-01 -7.27323219e-02 1.05145633e+00
-1.94339618e-01 5.15175104e-01 -3.80358011e-01 -1.38039482e+00
-7.20507503e-01 -5.58902442e-01 -9.56556424e-02 4.87292618e-01
-4.24368650e-01 -8.26229453e-01 3.46760824e-02 4.81844336e-01
8.61653268e-01 8.67624223e-01 8.27797592e-01 -5.63423157e-01
1.97008610e-01 -5.87811351e-01 -2.69550025e-01 6.60167038e-01
3.42347950e-01 3.02900374e-01 -5.97532451e-01 1.22426927e-01
2.93620020e-01 4.51662719e-01 3.32586080e-01 8.48461092e-01
9.44335103e-01 -1.87986672e-01 -2.25544795e-01 4.04733956e-01
1.52900708e+00 6.77209258e-01 9.17175531e-01 1.71785355e-01
4.05372083e-01 3.93994898e-01 1.56620830e-01 6.64907813e-01
3.61065418e-01 3.68653417e-01 -1.98315218e-01 -2.57287502e-01
2.38442108e-01 4.73170280e-01 -5.39184995e-02 7.47896135e-01
-8.20771098e-01 8.04799616e-01 -8.28712940e-01 7.00141415e-02
-1.66624606e+00 -1.06159818e+00 -7.13264883e-01 2.27888894e+00
8.27203393e-01 -7.41390735e-02 4.65345293e-01 5.07900774e-01
8.78420413e-01 -3.54492009e-01 -3.12604010e-01 -8.87801647e-01
2.04965457e-01 4.28510100e-01 5.17704546e-01 3.40665162e-01
-1.00465775e+00 3.75179648e-02 6.17297125e+00 -3.41254860e-01
-1.26475430e+00 -3.75461847e-01 8.20944369e-01 2.17497751e-01
5.83321929e-01 4.46448922e-02 -5.50550759e-01 1.03146088e+00
1.34198785e+00 -1.55048490e-01 5.02444386e-01 7.17536151e-01
6.69921339e-01 -4.37874585e-01 -5.92038512e-01 8.77444565e-01
-2.18035206e-01 -7.69199848e-01 -7.38509655e-01 -2.60761321e-01
1.93330944e-01 -2.13959679e-01 -6.85019135e-01 3.38572234e-01
-2.57358104e-01 -8.60836923e-01 -2.68681616e-01 1.19706476e+00
4.17402536e-01 -3.57430816e-01 9.41212833e-01 2.42599949e-01
-6.95839167e-01 -2.46599764e-01 -4.72842336e-01 -7.18852997e-01
-1.12194330e-01 1.07337224e+00 -6.81008399e-01 5.48079133e-01
6.44911528e-01 4.17419851e-01 -5.55660188e-01 1.37486076e+00
1.11597955e-01 8.36533964e-01 -6.11815214e-01 -2.87126690e-01
-3.75133276e-01 -7.93989182e-01 2.07004830e-01 1.02169311e+00
1.00554615e-01 4.40693229e-01 4.59421538e-02 1.15095544e+00
8.38172197e-01 4.95163828e-01 -1.67386472e-01 5.67996085e-01
6.00692809e-01 1.22343588e+00 -5.28242528e-01 -6.12958074e-01
-2.90393353e-01 1.19007658e-02 -1.35059461e-01 2.60362774e-01
-8.23810577e-01 -7.60939002e-01 1.16305165e-01 3.38580877e-01
-1.89551428e-01 9.17012393e-02 -9.78278697e-01 -7.74848342e-01
5.29435501e-02 -4.65018839e-01 4.34204102e-01 -3.43334407e-01
-1.15359342e+00 3.34076732e-01 5.97423539e-02 -8.96077871e-01
-1.41819015e-01 -3.25838774e-01 -7.19522417e-01 1.70390940e+00
-1.11529303e+00 -1.85348928e-01 -4.98777032e-01 4.36023533e-01
-3.40422630e-01 -4.52007800e-02 9.75809515e-01 4.97795254e-01
-7.52007484e-01 1.85740620e-01 -1.35285139e-01 2.69254029e-01
4.80675071e-01 -1.28337491e+00 -5.67465007e-01 4.45333779e-01
-7.28196084e-01 5.86668730e-01 5.93160391e-01 -4.17749107e-01
-9.65784848e-01 -9.46882010e-01 1.07381737e+00 -2.69537926e-01
1.69761136e-01 2.15921327e-01 -7.53665030e-01 1.65177852e-01
-3.76901031e-01 1.78496599e-01 1.01667786e+00 1.39186487e-01
1.83811262e-01 -4.44281578e-01 -1.50901651e+00 3.53005946e-01
6.20741323e-02 1.27531558e-01 -6.59255087e-01 1.36466473e-01
-4.19258140e-02 -2.45121136e-01 -1.85167205e+00 6.23775601e-01
1.02472568e+00 -1.04303002e+00 9.25174415e-01 -4.92834270e-01
2.02143282e-01 -5.07386088e-01 1.59942642e-01 -1.00202322e+00
-6.71774089e-01 -1.93916798e-01 6.33070152e-03 1.08944428e+00
4.53181773e-01 -9.21361983e-01 3.96897078e-01 1.35238290e+00
1.12505846e-01 -7.05400944e-01 -7.08986461e-01 -3.62329811e-01
-1.42365575e-01 7.30232671e-02 2.93379247e-01 9.38943028e-01
4.28252846e-01 3.22146535e-01 -4.52678442e-01 2.81008761e-02
8.18667054e-01 2.88415849e-01 7.61949360e-01 -1.55567372e+00
-2.51313955e-01 -1.41504094e-01 -5.78123868e-01 -2.62170434e-02
-5.72317719e-01 -5.57888687e-01 -3.17844123e-01 -1.54505432e+00
-8.64375308e-02 -6.34556711e-01 -5.90423584e-01 4.14671838e-01
-4.25869375e-01 1.29678771e-01 1.88082516e-01 1.71617821e-01
2.38840684e-01 -1.33689508e-01 1.07007670e+00 3.35351050e-01
-1.09446335e+00 5.90578437e-01 -5.66288590e-01 3.59293222e-01
1.13633966e+00 -3.81938010e-01 2.22563129e-02 4.50785846e-01
-4.29645687e-01 5.11939824e-01 2.01364413e-01 -1.24156845e+00
5.76614635e-03 1.85506046e-02 1.31419814e+00 -5.60638130e-01
-5.97559242e-03 -9.21351910e-01 4.96710718e-01 8.13374698e-01
-6.76855370e-02 -2.74799198e-01 7.10581988e-02 -7.99169987e-02
-1.00712210e-01 -9.22867879e-02 8.62483084e-01 -3.26053828e-01
-1.57700121e-01 -1.75439492e-01 -4.76961643e-01 -3.84701997e-01
9.29691195e-01 -5.87943912e-01 -1.43761784e-01 -9.60788652e-02
-1.37103820e+00 1.42822936e-01 -2.71551311e-01 2.59069987e-02
3.04931819e-01 -1.08776593e+00 -1.13315392e+00 4.94394213e-01
-2.81357020e-01 -4.44904536e-01 3.74977082e-01 1.48510194e+00
-5.55034637e-01 2.25866705e-01 -5.31058669e-01 -7.91804671e-01
-1.36060190e+00 2.87765712e-01 5.87225914e-01 1.12370268e-04
-6.95230484e-01 1.97935253e-01 -9.47834253e-01 -1.79745898e-01
2.03433752e-01 -1.14381723e-01 -7.20362306e-01 5.59107997e-02
5.59788525e-01 8.94685030e-01 -1.75848231e-02 -1.32085949e-01
-3.90564710e-01 7.01938391e-01 4.03666526e-01 2.40453854e-01
1.18454158e+00 -2.20825553e-01 -5.97372830e-01 3.91012281e-01
8.93905401e-01 -1.19641498e-01 -4.48990583e-01 -2.33392324e-02
3.25305872e-02 -4.40667987e-01 -1.75717622e-01 -1.10128498e+00
-9.00400758e-01 6.12753272e-01 9.31336939e-01 3.91085595e-01
1.22314954e+00 -5.27521193e-01 1.17452890e-01 3.73175852e-02
-3.88090196e-03 -1.11902714e+00 -8.52412999e-01 -3.75890406e-03
6.45789504e-01 -8.10892165e-01 3.74098927e-01 -2.31294811e-01
-5.42805195e-01 1.33395135e+00 4.12049592e-01 -4.09085631e-01
1.19112301e+00 4.00735587e-02 2.93240488e-01 1.14481010e-01
-4.38858002e-01 1.07361183e-01 6.86067119e-02 6.28688514e-01
5.78685462e-01 5.30863345e-01 -1.06380904e+00 6.06892526e-01
-2.17336237e-01 5.00963092e-01 2.67345339e-01 5.69065869e-01
-8.54274273e-01 -6.63209498e-01 -6.39282823e-01 9.79167700e-01
-4.65007365e-01 1.09359510e-01 1.38134761e-02 6.47024870e-01
2.00095192e-01 1.24308026e+00 8.95842016e-02 -4.54873234e-01
4.75453883e-01 5.67779064e-01 3.33473772e-01 -8.43233168e-02
-6.98133409e-01 2.04172768e-02 3.06480769e-02 -1.52609423e-01
-4.52067673e-01 -3.78013790e-01 -1.01663160e+00 -2.30974659e-01
-3.51076812e-01 5.19897223e-01 9.98859406e-01 7.25616395e-01
3.78605962e-01 3.64486396e-01 1.14496195e+00 -4.51455384e-01
-6.30886555e-01 -1.41448963e+00 -8.20894241e-01 3.43811780e-01
4.79228375e-03 -3.80166054e-01 -5.99907696e-01 -6.78862408e-02] | [14.061084747314453, 2.9777305126190186] |
17f6fed3-1b61-41cf-ad42-ecc025e1ecee | a-novel-enhanced-convolution-neural-network | 2208.02953 | null | https://arxiv.org/abs/2208.02953v1 | https://arxiv.org/pdf/2208.02953v1.pdf | A Novel Enhanced Convolution Neural Network with Extreme Learning Machine: Facial Emotional Recognition in Psychology Practices | Facial emotional recognition is one of the essential tools used by recognition psychology to diagnose patients. Face and facial emotional recognition are areas where machine learning is excelling. Facial Emotion Recognition in an unconstrained environment is an open challenge for digital image processing due to different environments, such as lighting conditions, pose variation, yaw motion, and occlusions. Deep learning approaches have shown significant improvements in image recognition. However, accuracy and time still need improvements. This research aims to improve facial emotion recognition accuracy during the training session and reduce processing time using a modified Convolution Neural Network Enhanced with Extreme Learning Machine (CNNEELM). The system entails (CNNEELM) improving the accuracy in image registration during the training session. Furthermore, the system recognizes six facial emotions happy, sad, disgust, fear, surprise, and neutral with the proposed CNNEELM model. The study shows that the overall facial emotion recognition accuracy is improved by 2% than the state of art solutions with a modified Stochastic Gradient Descent (SGD) technique. With the Extreme Learning Machine (ELM) classifier, the processing time is brought down to 65ms from 113ms, which can smoothly classify each frame from a video clip at 20fps. With the pre-trained InceptionV3 model, the proposed CNNEELM model is trained with JAFFE, CK+, and FER2013 expression datasets. The simulation results show significant improvements in accuracy and processing time, making the model suitable for the video analysis process. Besides, the study solves the issue of the large processing time required to process the facial images. | ['Omar Hisham Alsadoon', 'Tarik A. Rashid', 'Ahmed Dawoud', 'P. W. C. Prasad', 'Abeer Alsadoon', 'Nitesh Banskota'] | 2022-08-05 | null | null | null | null | ['facial-emotion-recognition'] | ['computer-vision'] | [-3.23399156e-01 -3.09556544e-01 1.54514477e-01 -6.38238609e-01
1.57632619e-01 1.46382272e-01 1.20895309e-03 -6.50027215e-01
-6.34977221e-01 3.64141196e-01 -3.34925473e-01 2.26688728e-01
1.36660948e-01 -2.37238139e-01 -1.12229884e-01 -8.89321864e-01
6.50191903e-02 -2.30192363e-01 -6.55190945e-01 -4.16556209e-01
6.25731945e-02 8.81826937e-01 -1.96527719e+00 3.60281020e-01
6.51165545e-01 1.60530245e+00 -2.35156417e-01 5.94905674e-01
-4.02930528e-01 8.67690265e-01 -5.31454384e-01 -5.05880952e-01
1.57289758e-01 -1.32982478e-01 -4.04652327e-01 3.54373813e-01
1.96280062e-01 -2.52642065e-01 -3.62734981e-02 1.16519153e+00
9.38651621e-01 1.90043703e-01 3.19691986e-01 -1.75075233e+00
-4.96804953e-01 -3.68302763e-01 -9.11113024e-01 8.32261331e-03
3.19533646e-01 5.49515672e-02 -6.29082918e-02 -1.33630729e+00
3.39235425e-01 1.23928034e+00 6.36877358e-01 9.66701865e-01
-5.47538042e-01 -9.56739306e-01 -1.04011573e-01 7.12110221e-01
-1.56289792e+00 -4.59789157e-01 8.74509871e-01 -2.96539385e-02
1.06965518e+00 2.39470333e-01 6.03567064e-01 7.29405165e-01
3.98028791e-01 5.53905249e-01 1.29387760e+00 -4.26981181e-01
2.91874498e-01 2.00705186e-01 1.45979840e-02 8.31003428e-01
-3.54158312e-01 -6.63876534e-02 -4.53567743e-01 3.39452401e-02
3.30990136e-01 3.51609677e-01 -2.06183925e-01 4.78682846e-01
-2.46313125e-01 5.97211063e-01 2.04590395e-01 3.07375550e-01
-6.94621205e-01 -1.11910179e-01 7.30112791e-01 5.28041601e-01
6.52988434e-01 -3.52092475e-01 -4.10223722e-01 -4.16458130e-01
-9.18862879e-01 -5.29238880e-01 7.12435961e-01 5.67848086e-01
5.11791229e-01 4.63750541e-01 3.01564604e-01 1.05681181e+00
4.44850892e-01 5.26344359e-01 7.62793481e-01 -9.50643837e-01
-1.58761084e-01 5.61185241e-01 -2.31469303e-01 -1.39686310e+00
-3.92389476e-01 -6.24670237e-02 -1.11853302e+00 7.82070398e-01
-7.44794384e-02 -4.43037778e-01 -8.46449375e-01 1.41707361e+00
5.57245970e-01 2.03796074e-01 2.87299067e-01 9.67432737e-01
1.11695027e+00 8.79890025e-01 2.28051960e-01 -6.93637013e-01
1.35531390e+00 -9.23503935e-01 -1.22068274e+00 6.25662282e-02
5.77081442e-01 -8.33185732e-01 8.03325355e-01 6.34006977e-01
-9.01127100e-01 -6.53274179e-01 -6.87925160e-01 2.46719360e-01
-2.65823841e-01 3.63949388e-01 8.53400946e-01 1.02200282e+00
-1.07743371e+00 3.00699741e-01 -6.22538447e-01 -2.59953827e-01
5.86841702e-01 5.91664135e-01 -5.77537775e-01 1.25978356e-02
-8.99166822e-01 7.31961548e-01 1.37867391e-01 7.28764296e-01
-2.13355705e-01 -3.69619846e-01 -7.74591386e-01 8.12528580e-02
-1.83639690e-01 1.86802335e-02 7.45523810e-01 -1.93180835e+00
-1.95704794e+00 1.05333781e+00 -3.61256689e-01 -6.20048819e-03
1.67576611e-01 -9.06002223e-02 -8.44432712e-01 2.72143692e-01
-6.26556158e-01 6.27449393e-01 8.98196816e-01 -5.76506436e-01
-1.87593475e-01 -6.18969202e-01 -5.72030783e-01 1.84238613e-01
-6.17661655e-01 8.13862145e-01 -1.98781684e-01 -2.47494057e-01
-7.61459842e-02 -9.04787481e-01 -2.74644401e-02 3.42225224e-01
3.42487901e-01 -2.69491285e-01 1.32616818e+00 -7.34422565e-01
8.14957500e-01 -2.65230846e+00 -3.36257637e-01 1.85633361e-01
-1.09854862e-01 6.70977056e-01 3.73816825e-02 -2.39346266e-01
-5.68817139e-01 -1.39839455e-01 2.69062549e-01 -2.32885823e-01
-1.66331887e-01 1.28442779e-01 1.41582832e-01 5.68065345e-01
2.08386928e-01 7.90736377e-01 -2.38209754e-01 -6.74162686e-01
4.05656874e-01 1.05696392e+00 -3.88041466e-01 3.32874089e-01
2.97599733e-01 2.34574854e-01 -7.08400309e-02 1.02383006e+00
1.09507906e+00 1.67163461e-01 -3.08072418e-01 -2.74433017e-01
5.49834631e-02 -1.07976878e+00 -1.37624526e+00 1.16577709e+00
-5.58411121e-01 7.10501850e-01 2.85839915e-01 -1.01030326e+00
1.38970292e+00 6.61818206e-01 5.45293093e-01 -6.91119313e-01
9.19333398e-01 1.16645642e-01 -2.35731214e-01 -1.13394630e+00
8.58431458e-02 -3.36952776e-01 4.42023605e-01 2.13416234e-01
2.94473022e-01 4.43497419e-01 -2.42984295e-01 -3.45394254e-01
3.84613574e-01 2.32815500e-02 2.42252499e-02 1.26783624e-01
8.15949261e-01 -4.55136746e-01 7.70232379e-01 -2.37729266e-01
-7.46528566e-01 1.53138772e-01 2.20016390e-01 -8.88491094e-01
-5.35743773e-01 -3.60217273e-01 -1.18519381e-01 9.66198444e-01
-1.24688730e-01 2.85634343e-02 -9.13423538e-01 -2.02602223e-01
-4.98000085e-01 3.65765035e-01 -4.35721159e-01 -5.03564835e-01
-9.69271362e-02 -1.11345708e+00 4.16196436e-01 3.92196774e-01
9.76130545e-01 -1.23587763e+00 -7.84393668e-01 -2.41007194e-01
-3.71235237e-02 -1.07400835e+00 -1.84320603e-02 -2.53538340e-01
-8.09795618e-01 -6.61542416e-01 -7.12627411e-01 -1.00943935e+00
7.65369654e-01 -3.21002662e-01 7.04716623e-01 2.56962508e-01
-7.07177222e-01 3.86778653e-01 -2.99971908e-01 -5.64972103e-01
1.03260018e-01 -7.98644185e-01 3.25054497e-01 5.78988791e-01
8.68230104e-01 -3.75520468e-01 -4.83741075e-01 9.46795493e-02
-6.12910807e-01 -2.60433555e-01 2.47931853e-01 8.04969072e-01
4.17635471e-01 2.51338214e-01 5.36278665e-01 -3.59438181e-01
5.60144603e-01 -4.05299067e-01 -2.99390286e-01 1.29191205e-01
-6.03141129e-01 -5.14603794e-01 6.44250572e-01 -7.14967251e-01
-1.41139150e+00 8.57798830e-02 -4.60552961e-01 -7.52913058e-01
-4.58765566e-01 2.56531537e-01 -2.03360677e-01 -6.39971375e-01
3.67530137e-01 6.68524951e-02 4.80625510e-01 6.54539466e-02
-1.65470347e-01 1.17671025e+00 2.97575027e-01 -7.51006529e-02
-8.13646913e-02 5.24497449e-01 -1.82008073e-02 -1.06739545e+00
-3.28150958e-01 -1.05843358e-01 -2.48819605e-01 -8.17079663e-01
7.63532162e-01 -8.50834131e-01 -1.46541166e+00 1.03918958e+00
-9.22559559e-01 1.66873157e-01 7.99327195e-02 7.97651589e-01
-4.47477907e-01 1.96023732e-01 -8.25494349e-01 -1.20037568e+00
-9.23211038e-01 -1.03717065e+00 6.62758350e-01 8.38296592e-01
-1.29476368e-01 -8.25588226e-01 -4.02410656e-01 2.87018329e-01
5.56027651e-01 5.11398017e-01 5.94646156e-01 -2.39218801e-01
1.91145957e-01 -5.48114300e-01 1.61067713e-02 6.43003762e-01
-3.06804050e-02 5.99605858e-01 -1.06267762e+00 -4.02380042e-02
5.10134518e-01 -5.94006121e-01 3.78394067e-01 4.06351775e-01
1.33933997e+00 -4.38625999e-02 1.57099113e-01 9.81217504e-01
1.43279517e+00 5.29321849e-01 1.03356934e+00 2.01014265e-01
2.14166224e-01 6.93973780e-01 6.51325703e-01 5.85381925e-01
2.31740400e-02 8.11828896e-02 6.05343819e-01 -5.87957978e-01
2.50133634e-01 5.91764152e-01 4.91605699e-01 7.04936564e-01
-2.14926615e-01 1.98989391e-01 -6.86484933e-01 4.84104864e-02
-1.43792832e+00 -1.27479851e+00 -1.30007207e-01 1.81799161e+00
4.66448128e-01 -4.38380599e-01 -2.58596987e-01 2.11750433e-01
8.54778528e-01 -5.20805195e-02 -4.96740282e-01 -1.30445206e+00
-2.03466132e-01 5.01021683e-01 -2.12862194e-01 1.91031396e-01
-8.10897648e-01 8.71308982e-01 5.60534000e+00 7.17625380e-01
-1.91682708e+00 -1.22773610e-01 1.07839870e+00 -5.29977858e-01
5.83430767e-01 -5.58256924e-01 -6.53897345e-01 6.87303066e-01
1.20275092e+00 1.34767130e-01 3.88261467e-01 1.20426309e+00
4.50029284e-01 -6.25518411e-02 -5.67837715e-01 1.79077423e+00
5.46780109e-01 -6.73134565e-01 -1.75238416e-01 -2.85570443e-01
4.10405666e-01 -2.99601614e-01 3.06608319e-01 5.59583068e-01
-5.29992163e-01 -1.29586816e+00 6.10651337e-02 7.80788422e-01
8.46115708e-01 -1.29821551e+00 1.09572005e+00 2.45611757e-01
-8.34335983e-01 -1.86432555e-01 -4.73867357e-01 -1.28620729e-01
-4.03114408e-02 4.90179688e-01 -3.65611523e-01 6.73487112e-02
1.04358697e+00 3.83226991e-01 -1.93141654e-01 5.20550728e-01
2.35869214e-02 4.32182312e-01 -3.45150411e-01 -2.83774316e-01
6.13120347e-02 -3.35620493e-01 9.60531533e-02 9.91245389e-01
3.75262529e-01 5.20655513e-01 -8.00257772e-02 5.16717792e-01
6.14925958e-02 4.99396563e-01 -3.41153234e-01 -5.21560088e-02
-8.82414728e-02 1.77006638e+00 -5.93943954e-01 -1.97743446e-01
-4.05035049e-01 1.33957696e+00 2.29192805e-02 3.12949032e-01
-9.95202005e-01 -6.37900174e-01 8.13768566e-01 -5.20163357e-01
-6.69823661e-02 3.33517417e-02 -7.35329464e-02 -9.46254492e-01
1.11650452e-01 -8.84937406e-01 3.27615649e-01 -1.09224677e+00
-6.95178151e-01 9.41426396e-01 -5.99990070e-01 -9.94117677e-01
-2.13768527e-01 -9.00003016e-01 -7.21784115e-01 8.25231016e-01
-1.50510573e+00 -8.29963923e-01 -6.34376585e-01 9.93673563e-01
4.59851921e-01 -2.47996151e-01 1.04419851e+00 4.76975530e-01
-9.21513617e-01 7.77946889e-01 -1.08244337e-01 2.62483716e-01
5.91730595e-01 -5.42267442e-01 -5.57549298e-01 5.00114441e-01
-3.49341601e-01 2.86143631e-01 4.58781987e-01 -2.03608915e-01
-1.43984711e+00 -8.96243334e-01 9.09079492e-01 3.25170547e-01
-4.18095216e-02 -7.84994103e-03 -7.28639781e-01 2.89714336e-01
2.52187163e-01 6.77892625e-01 1.22667372e+00 -1.52793244e-01
2.18578130e-02 -3.08646739e-01 -1.68712759e+00 5.35170555e-01
4.49464053e-01 -3.43550831e-01 -2.04771385e-01 5.91074862e-02
3.52975577e-02 -3.17550242e-01 -9.58313107e-01 4.26742017e-01
8.25774312e-01 -9.59989548e-01 4.72571999e-01 -6.14343405e-01
4.76831973e-01 -5.14690727e-02 -1.64379328e-01 -1.11921084e+00
-6.70391077e-04 -3.92744064e-01 -5.15559688e-02 1.25780785e+00
-1.41834076e-02 -7.12469876e-01 8.84557068e-01 1.18965507e+00
3.39900225e-01 -1.09530044e+00 -9.31379497e-01 -4.85081226e-01
-3.37282926e-01 -4.21119124e-01 3.89610410e-01 1.01127207e+00
2.14898780e-01 -5.15022762e-02 -2.88057268e-01 -6.41876040e-03
2.94138551e-01 -7.65245967e-03 5.07513165e-01 -9.77227926e-01
1.84664473e-01 -1.28131405e-01 -9.79049981e-01 -6.65383935e-02
5.36514759e-01 -7.00504720e-01 -4.72736657e-01 -9.92748737e-01
1.38032511e-01 1.27257630e-01 -4.14229184e-01 6.27125382e-01
1.27563521e-01 3.92208904e-01 1.48058563e-01 -1.69125780e-01
-4.46069002e-01 9.02957737e-01 8.80448222e-01 9.01598558e-02
-1.27189532e-01 -1.80771425e-01 -4.15393323e-01 1.02650869e+00
8.40851963e-01 -2.16473565e-01 -2.39187345e-01 -2.78413594e-01
7.25343674e-02 2.71208249e-02 2.38349095e-01 -1.00602853e+00
3.91276330e-01 -1.34302927e-02 1.08459163e+00 -2.91264266e-01
6.06857598e-01 -1.23527110e+00 2.52822489e-01 5.76411903e-01
1.40477598e-01 9.22326371e-02 4.35025185e-01 -2.79482603e-02
-5.68876326e-01 -1.78849012e-01 1.01009226e+00 -8.07321072e-02
-9.73965645e-01 4.22365099e-01 -4.65250373e-01 -3.49520564e-01
1.53347158e+00 -6.92369163e-01 9.62455571e-02 -5.07622302e-01
-1.00779891e+00 -3.30076441e-02 9.51298922e-02 2.18745291e-01
1.00432336e+00 -1.44812369e+00 -6.30121171e-01 5.40808260e-01
-2.59389371e-01 -4.53993797e-01 7.25545526e-01 1.06743920e+00
-5.47607541e-01 -1.85603395e-01 -6.05634868e-01 -5.94597518e-01
-2.01271248e+00 3.61514866e-01 6.70489550e-01 4.75710988e-01
-3.06325376e-01 1.01217663e+00 -3.53435099e-01 -2.85927743e-01
2.64807522e-01 2.76719779e-01 -5.61996460e-01 1.02475227e-03
9.11518812e-01 5.41960716e-01 2.15923101e-01 -9.93310690e-01
-4.83656555e-01 7.10841537e-01 -1.23674944e-01 2.85729766e-01
1.34479570e+00 5.49804159e-02 -4.00834262e-01 3.40841599e-02
1.55005634e+00 -4.41792220e-01 -9.59607482e-01 2.27861688e-01
-2.81953722e-01 -4.15247351e-01 3.10457110e-01 -7.28117943e-01
-1.57284117e+00 9.22718287e-01 1.36786306e+00 -5.09391189e-01
1.83907235e+00 -4.92009759e-01 7.10736394e-01 3.44887435e-01
1.72325626e-01 -1.46119416e+00 1.96975786e-02 5.36864102e-01
8.10055494e-01 -1.27388263e+00 -3.55253786e-01 -2.17978656e-01
-7.62415588e-01 1.44308412e+00 8.28356445e-01 -1.86868384e-01
9.92236555e-01 6.85081065e-01 3.56184572e-01 -2.23069191e-01
-7.92169392e-01 2.30784386e-01 -8.81334841e-02 3.39680940e-01
4.96988922e-01 -1.11024722e-01 -3.21066260e-01 7.19701886e-01
-4.74823117e-02 4.14959759e-01 1.74122438e-01 9.11502123e-01
-2.52326459e-01 -4.37757760e-01 -5.38843513e-01 2.83086389e-01
-9.21277881e-01 4.20405120e-02 -3.19323272e-01 3.73706132e-01
1.96544737e-01 1.02045739e+00 2.28368983e-01 -4.77643281e-01
1.36554137e-01 3.57300729e-01 3.24207187e-01 6.86613992e-02
-5.43156385e-01 -1.42134517e-01 -3.63793343e-01 -6.91333055e-01
-4.98639911e-01 -2.08277270e-01 -1.65433896e+00 -4.95782733e-01
-3.69465560e-01 1.60430118e-01 1.10152483e+00 8.23265612e-01
7.11929440e-01 1.83911726e-01 1.01994729e+00 -6.60586059e-01
-1.93104997e-01 -7.57202983e-01 -6.99351728e-01 4.28138793e-01
6.81041181e-03 -5.08535147e-01 -4.53466982e-01 1.62167117e-01] | [13.541625022888184, 1.7801958322525024] |
0c2bdddd-be82-4471-b151-51fa68d418ee | efficient-structured-inference-for-transition | null | null | https://aclanthology.org/Q16-1014 | https://aclanthology.org/Q16-1014.pdf | Efficient Structured Inference for Transition-Based Parsing with Neural Networks and Error States | Transition-based approaches based on local classification are attractive for dependency parsing due to their simplicity and speed, despite producing results slightly below the state-of-the-art. In this paper, we propose a new approach for approximate structured inference for transition-based parsing that produces scores suitable for global scoring using local models. This is accomplished with the introduction of error states in local training, which add information about incorrect derivation paths typically left out completely in locally-trained models. Using neural networks for our local classifiers, our approach achieves 93.61{\%} accuracy for transition-based dependency parsing in English. | ['Ashish Vaswani', 'Kenji Sagae'] | 2016-01-01 | null | null | null | tacl-2016-1 | ['transition-based-dependency-parsing'] | ['natural-language-processing'] | [-9.10054669e-02 3.35923433e-01 -4.58227634e-01 -1.16605818e+00
-1.74236953e+00 -4.40800935e-01 9.98872891e-02 5.37620962e-01
-4.62248057e-01 9.63312387e-01 2.62227565e-01 -6.99295521e-01
2.40466923e-01 -7.65924215e-01 -7.53519058e-01 -2.94942468e-01
-5.34421066e-04 7.23713696e-01 5.24142325e-01 -6.90403506e-02
1.54319461e-02 3.33461732e-01 -8.33786905e-01 8.44979644e-01
9.63347495e-01 7.35776067e-01 1.02367386e-01 9.74547088e-01
-5.75165629e-01 1.07163405e+00 -9.40190315e-01 -9.30007577e-01
-4.33749586e-01 -7.15271950e-01 -1.11286759e+00 -8.64128768e-01
7.19922125e-01 -4.82702821e-01 1.66860908e-01 1.04256189e+00
1.47545695e-01 -1.00228257e-01 4.37522531e-01 -4.44573462e-01
-4.86180991e-01 1.32571375e+00 1.40131801e-01 3.01030606e-01
6.08206213e-01 -4.13890570e-01 1.45954490e+00 -8.98755908e-01
6.96674705e-01 1.30718350e+00 8.70092571e-01 6.67743444e-01
-1.15440679e+00 -3.46996844e-01 3.55353624e-01 1.52329788e-01
-8.40987265e-01 -5.26915491e-01 6.10082805e-01 8.16506967e-02
1.83259523e+00 5.15127480e-02 2.51999885e-01 8.31669092e-01
6.89812899e-01 7.87452519e-01 1.15513122e+00 -7.63292789e-01
3.80507588e-01 -2.28927895e-01 6.94628775e-01 1.00932169e+00
1.41108736e-01 1.95333865e-02 -7.16781259e-01 -2.47684732e-01
3.87791395e-01 -6.40245378e-01 4.29741532e-01 1.49477184e-01
-6.56397998e-01 9.69700515e-01 5.01564085e-01 3.97505909e-01
-6.58104941e-02 1.32672876e-01 5.17044723e-01 3.05415332e-01
7.53599226e-01 6.31702319e-02 -1.01818371e+00 -3.68197232e-01
-1.15880704e+00 1.55128375e-01 9.61782575e-01 9.93345082e-01
7.04133213e-01 1.14597254e-01 -9.92100388e-02 7.98853636e-01
2.47221842e-01 2.41002530e-01 1.82261899e-01 -1.10857856e+00
1.06079972e+00 4.50924695e-01 -6.81912675e-02 -1.90410584e-01
-6.37014389e-01 -2.10774615e-01 -3.59069824e-01 2.21255347e-01
8.30808818e-01 -1.48614183e-01 -9.76049304e-01 1.68404186e+00
2.37173378e-01 -3.06313664e-01 1.93309143e-01 3.47751498e-01
4.99700159e-01 7.19283819e-01 4.84160304e-01 -4.06145662e-01
1.14974153e+00 -1.06100643e+00 -8.40738952e-01 -4.71334368e-01
1.38655031e+00 -5.81330240e-01 8.11723888e-01 5.05537510e-01
-1.46294427e+00 -4.48875844e-01 -8.44623923e-01 -3.12352300e-01
-1.84684083e-01 1.31230712e-01 6.95955873e-01 8.45021367e-01
-1.23449123e+00 1.08601201e+00 -1.29286361e+00 -8.81886557e-02
1.58027530e-01 3.96770924e-01 -4.69463617e-01 -8.12043697e-02
-1.10049200e+00 1.39900911e+00 4.21616435e-01 2.49803081e-01
-2.09183514e-01 -1.62654057e-01 -1.07121539e+00 7.42227659e-02
-1.15466163e-01 -1.01129018e-01 1.77772713e+00 -5.43518543e-01
-1.78790724e+00 6.00429118e-01 -6.24995589e-01 -4.43739206e-01
4.06571440e-02 -3.42780262e-01 -1.86163157e-01 -1.86675545e-02
2.53609240e-01 3.11242998e-01 3.96639884e-01 -7.69982517e-01
-7.60349691e-01 -4.67528760e-01 -1.47243381e-01 -5.34592615e-03
9.38499570e-02 6.83005810e-01 -1.99678555e-01 -4.04314309e-01
5.89010715e-01 -6.63200438e-01 -2.95562893e-01 -5.91323256e-01
-2.06385642e-01 -6.33725882e-01 3.20384204e-01 -1.26314461e+00
1.48762417e+00 -1.55625331e+00 1.19956367e-01 3.77580635e-02
-4.76143688e-01 3.51364583e-01 -2.39097103e-02 2.31986836e-01
4.06216905e-02 3.19361001e-01 -4.55929339e-01 -6.85760498e-01
-5.78607768e-02 5.89278579e-01 4.45530936e-02 4.95252609e-02
5.51633000e-01 9.35362518e-01 -9.04125571e-01 -8.08197856e-01
1.76865265e-01 -7.97227863e-03 -4.83362019e-01 2.02167913e-01
-1.15941875e-01 1.28535315e-01 -4.79462266e-01 1.01800168e+00
6.72933042e-01 4.96577382e-01 7.23317802e-01 3.14896613e-01
-1.53453171e-01 1.25324631e+00 -8.37148368e-01 1.88853049e+00
-6.70627892e-01 1.56875730e-01 1.64002970e-01 -9.60162997e-01
8.46331239e-01 4.19668615e-01 -3.56195718e-01 -6.12260222e-01
-9.47909150e-03 5.25038302e-01 -4.31829281e-02 6.16547540e-02
4.09013450e-01 -2.30672881e-01 -7.17694223e-01 3.17472368e-01
5.38094044e-01 -1.47739440e-01 1.44555449e-01 1.67134814e-02
1.46503878e+00 7.42852569e-01 3.92523527e-01 -2.74656206e-01
3.14866722e-01 3.57515663e-01 9.03064489e-01 1.01333570e+00
-3.75145763e-01 3.77906442e-01 5.28160751e-01 -6.36334658e-01
-8.34089756e-01 -1.16116560e+00 -4.59205747e-01 1.35456789e+00
-3.76208693e-01 -5.20757675e-01 -1.04570687e+00 -1.20413375e+00
-5.58799267e-01 9.43298459e-01 -2.47579888e-01 3.63297835e-02
-1.41890931e+00 -6.54023230e-01 8.66658926e-01 1.22935295e+00
3.87267470e-02 -1.34851670e+00 -2.63655543e-01 7.50368178e-01
-4.64761823e-01 -1.22050440e+00 8.84645209e-02 1.08013546e+00
-1.65608227e+00 -4.91105914e-01 -1.57496110e-01 -1.09067345e+00
3.70928973e-01 -6.60212159e-01 1.35026360e+00 8.09078142e-02
1.30949780e-01 -5.30052662e-01 -6.13005459e-01 2.41125207e-02
-9.25979733e-01 4.28002656e-01 -1.30496696e-01 -1.00977731e+00
5.48155785e-01 -1.80004835e-01 1.31306037e-01 -1.74283564e-01
-9.28783491e-02 -3.37416351e-01 7.93002367e-01 1.13489783e+00
3.67372006e-01 -5.31178653e-01 3.30179125e-01 -1.24431705e+00
3.22306305e-01 -2.25890815e-01 -3.49101931e-01 2.93681353e-01
-5.52094996e-01 4.95101988e-01 9.59665298e-01 7.08158985e-02
-1.61961854e+00 1.73519760e-01 -8.15527022e-01 4.32615250e-01
-4.75106478e-01 5.76783359e-01 -1.41723573e-01 1.16178378e-01
6.75526202e-01 -2.20621064e-01 -7.61680067e-01 -8.65319788e-01
1.06342629e-01 5.23168445e-01 4.76404846e-01 -8.85968626e-01
2.85341889e-01 -2.30031475e-01 -3.75586636e-02 -3.44607383e-01
-1.13230288e+00 -1.82911500e-01 -1.17234862e+00 2.38466993e-01
9.01072443e-01 -7.20473230e-01 1.09583475e-01 4.70163018e-01
-1.49858880e+00 -4.18054402e-01 -6.57214001e-02 4.51250583e-01
-4.19382453e-01 4.69252676e-01 -1.36364043e+00 -8.95281672e-01
-3.12263638e-01 -1.00547183e+00 1.02480388e+00 1.67318657e-02
-3.10092598e-01 -1.09626865e+00 2.12698832e-01 4.26484048e-01
2.21706793e-01 -1.50332630e-01 1.18444729e+00 -8.41994226e-01
-2.69605786e-01 -3.94829035e-01 8.49258155e-02 5.07621884e-01
-1.44972101e-01 6.11322224e-02 -9.72904086e-01 1.03020661e-01
-2.50752866e-01 -4.86850411e-01 8.26363444e-01 5.10881305e-01
6.74120665e-01 -1.69326931e-01 -2.66549259e-01 4.38794017e-01
1.29344463e+00 8.98834765e-02 2.38736525e-01 2.61687160e-01
4.60420877e-01 7.26742268e-01 9.66227829e-01 1.05255112e-01
4.56728876e-01 6.68679953e-01 1.98438197e-01 3.06669950e-01
-1.32105753e-01 -2.76764780e-01 7.57084072e-01 1.06494462e+00
-1.54621348e-01 -1.48895115e-01 -9.08442914e-01 6.65678501e-01
-1.83988202e+00 -7.00580895e-01 -5.42386889e-01 2.04508591e+00
1.08120239e+00 7.12233782e-01 -2.61498123e-01 3.89614180e-02
7.57807314e-01 8.52567330e-02 9.01996996e-03 -1.62861824e+00
5.36127947e-02 1.20179391e+00 3.54260564e-01 8.63495469e-01
-1.09882677e+00 1.54890347e+00 7.33429289e+00 5.90504110e-01
-5.08790791e-01 4.76181835e-01 6.24739707e-01 3.79272938e-01
-4.73137088e-02 4.37492192e-01 -1.36480713e+00 1.86586276e-01
1.78922546e+00 7.94935167e-01 -2.56045498e-02 1.02232242e+00
-2.24987268e-01 -2.91119397e-01 -1.12188220e+00 3.01025286e-02
-1.36927187e-01 -1.11303484e+00 -1.36555687e-01 -5.40005028e-01
4.81363684e-01 1.74346954e-01 -5.06291270e-01 6.31907463e-01
6.28612220e-01 -8.55556250e-01 7.89798796e-01 7.08415583e-02
6.49297476e-01 -8.46170306e-01 1.16478443e+00 5.19833446e-01
-1.19509327e+00 -2.27205548e-02 -6.74948514e-01 -4.45589751e-01
2.84037083e-01 5.51242769e-01 -5.71111619e-01 5.33648431e-01
8.04345131e-01 8.67196247e-02 -3.40368748e-01 3.60967487e-01
-8.87278914e-01 1.15777946e+00 -5.65768838e-01 -2.54740179e-01
3.74176323e-01 -1.39383182e-01 1.23691328e-01 1.71393991e+00
1.30015239e-01 -3.46906073e-02 1.63648073e-02 5.22337139e-01
3.86411995e-02 1.82268918e-01 -3.59723419e-01 4.96253163e-01
5.57034492e-01 9.87492919e-01 -5.30949533e-01 -3.13030511e-01
-4.98375267e-01 1.08012009e+00 1.19965100e+00 -2.10354403e-01
-6.29120886e-01 -4.25569743e-01 2.79253215e-01 -4.73654270e-01
4.16407466e-01 -4.06601131e-01 -5.52603602e-01 -1.12544239e+00
3.09424043e-01 -4.49483603e-01 5.49257457e-01 -3.65228891e-01
-9.87969398e-01 7.72099793e-01 1.37055069e-01 -5.46896219e-01
-8.54247630e-01 -1.09996045e+00 -8.47645104e-01 8.71112347e-01
-1.43468249e+00 -1.32221055e+00 3.08496624e-01 8.48129690e-02
6.21156335e-01 2.73822993e-01 1.45508957e+00 -1.21901527e-01
-4.69554693e-01 9.76321995e-01 1.23296924e-01 4.37801838e-01
7.45983839e-01 -1.64982617e+00 9.37425435e-01 1.23161125e+00
4.21996593e-01 6.04890406e-01 4.42626387e-01 -5.85576117e-01
-9.57218051e-01 -7.56363213e-01 1.89446354e+00 -7.70082295e-01
5.49849927e-01 -3.06763738e-01 -8.93149316e-01 8.87872696e-01
-1.19463228e-01 2.24695846e-01 4.96020108e-01 7.29040027e-01
-5.21775782e-01 6.51158839e-02 -1.18572974e+00 1.74985379e-01
1.14316392e+00 -5.40237546e-01 -1.03763819e+00 1.39717758e-01
4.26599443e-01 -5.93718946e-01 -1.07344139e+00 2.93437868e-01
5.13931751e-01 -1.00647473e+00 5.98199964e-01 -7.03962505e-01
3.38188112e-01 2.85537213e-01 -2.98917472e-01 -1.37086487e+00
-4.53503102e-01 -3.25547159e-01 -2.26423800e-01 1.25941467e+00
1.04016650e+00 -4.79783893e-01 9.66870248e-01 7.09777474e-01
-6.03387117e-01 -6.22373104e-01 -1.52774847e+00 -6.48159742e-01
7.46205330e-01 -4.84244347e-01 2.33985037e-01 5.15625179e-01
4.63363826e-01 3.37118745e-01 -7.66463354e-02 2.62273610e-01
5.77175260e-01 8.92308801e-02 9.05366316e-02 -1.03992903e+00
-3.24255198e-01 -1.24632381e-01 -2.43853167e-01 -9.77670908e-01
5.24664581e-01 -7.17340052e-01 5.34120500e-01 -1.57071519e+00
8.74493830e-03 -6.46487594e-01 -3.46521020e-01 8.05788517e-01
-3.83086503e-01 2.50935614e-01 -1.43114629e-03 -1.66746248e-02
-4.29941416e-01 -4.81214449e-02 6.69008851e-01 2.93680906e-01
-1.31463632e-01 4.29515541e-02 -2.87806123e-01 9.14909601e-01
9.84445810e-01 -1.02091455e+00 3.31607819e-01 -7.55349874e-01
2.24229738e-01 5.54615974e-01 -2.24590033e-01 -1.09526777e+00
1.29616871e-01 -7.10540265e-02 3.80629569e-01 -7.53198802e-01
4.57378805e-01 -3.79097044e-01 -5.50032914e-01 5.39623380e-01
-3.57514411e-01 2.80265450e-01 -1.51174143e-03 3.35682750e-01
-4.32157725e-01 -8.97427082e-01 7.48749137e-01 -2.93889642e-01
-5.00610232e-01 -3.17783207e-01 -5.79133630e-01 1.20248072e-01
4.85228330e-01 4.03609015e-02 -3.10883343e-01 5.59773855e-03
-9.95745361e-01 -3.74129534e-01 1.31234959e-01 -6.15005940e-02
3.56756181e-01 -8.17451775e-01 -7.56495595e-01 2.02014536e-01
-3.23343240e-02 2.48556081e-02 -1.65646955e-01 5.00149190e-01
-8.60545695e-01 5.50433397e-01 -1.12474144e-01 -4.05524462e-01
-1.21758711e+00 -8.06067884e-02 1.77706368e-02 -8.00519347e-01
-4.78494108e-01 1.13580096e+00 -6.16048276e-01 -1.07764065e+00
-4.60483283e-02 -1.01701474e+00 6.66339323e-02 -2.84343898e-01
1.83960512e-01 1.61251023e-01 6.27655149e-01 -5.13078511e-01
-7.14835584e-01 2.76524931e-01 -2.58152157e-01 -1.69459149e-01
1.07371581e+00 2.92672902e-01 -1.45789504e-01 4.90845650e-01
9.35691416e-01 1.06243864e-01 -9.19692099e-01 -3.36132832e-02
5.67027092e-01 -1.04120754e-01 1.16372362e-01 -1.20000696e+00
-2.82283187e-01 1.15505493e+00 2.09446579e-01 -9.94241312e-02
6.70199275e-01 2.42010772e-01 9.02757943e-01 8.20026696e-01
6.19240165e-01 -1.33880639e+00 -4.53610271e-01 1.11472058e+00
8.30541700e-02 -1.23421824e+00 -3.53176780e-02 -3.97913396e-01
-4.13659632e-01 1.31433249e+00 5.82379699e-01 -3.70333344e-01
3.52589190e-01 6.50613606e-01 2.31523827e-01 3.75401020e-01
-9.06793475e-01 1.10399552e-01 -1.40807137e-01 5.94768763e-01
8.73077273e-01 3.65424186e-01 -5.49631357e-01 9.10311580e-01
-1.52368292e-01 -5.58209419e-01 3.51053655e-01 1.35336876e+00
-8.01794171e-01 -1.73989153e+00 -2.07929224e-01 3.24933618e-01
-9.80983973e-01 -5.50423741e-01 -3.87502342e-01 7.09253430e-01
-1.22730859e-01 1.06221998e+00 -3.32916453e-02 -8.62205327e-02
1.26791313e-01 7.86945462e-01 8.03877532e-01 -9.40810740e-01
-1.05555964e+00 -2.15552449e-01 9.55132961e-01 -7.21528769e-01
-2.42284358e-01 -9.53816891e-01 -1.54375112e+00 -4.22705263e-02
-7.73857296e-01 2.84884572e-01 9.67677176e-01 1.21946084e+00
-4.32796106e-02 2.36753225e-01 4.55941767e-01 -7.47212350e-01
-9.86302137e-01 -1.28028405e+00 -2.99341053e-01 -8.54728520e-02
1.61622018e-02 -2.38394454e-01 -3.78240287e-01 -2.00841695e-01] | [10.354787826538086, 9.741865158081055] |
1082fbaf-a6a9-4d77-9342-404e9f6ade81 | looking-through-the-glass-neural-surface | 2304.08706 | null | https://arxiv.org/abs/2304.08706v1 | https://arxiv.org/pdf/2304.08706v1.pdf | Looking Through the Glass: Neural Surface Reconstruction Against High Specular Reflections | Neural implicit methods have achieved high-quality 3D object surfaces under slight specular highlights. However, high specular reflections (HSR) often appear in front of target objects when we capture them through glasses. The complex ambiguity in these scenes violates the multi-view consistency, then makes it challenging for recent methods to reconstruct target objects correctly. To remedy this issue, we present a novel surface reconstruction framework, NeuS-HSR, based on implicit neural rendering. In NeuS-HSR, the object surface is parameterized as an implicit signed distance function (SDF). To reduce the interference of HSR, we propose decomposing the rendered image into two appearances: the target object and the auxiliary plane. We design a novel auxiliary plane module by combining physical assumptions and neural networks to generate the auxiliary plane appearance. Extensive experiments on synthetic and real-world datasets demonstrate that NeuS-HSR outperforms state-of-the-art approaches for accurate and robust target surface reconstruction against HSR. Code is available at https://github.com/JiaxiongQ/NeuS-HSR. | ['Bo Ren', 'Ming-Ming Cheng', 'Ze-Xin Yin', 'Yifan Zhu', 'Peng-Tao Jiang', 'Jiaxiong Qiu'] | 2023-04-18 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Qiu_Looking_Through_the_Glass_Neural_Surface_Reconstruction_Against_High_Specular_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Qiu_Looking_Through_the_Glass_Neural_Surface_Reconstruction_Against_High_Specular_CVPR_2023_paper.pdf | cvpr-2023-1 | ['neural-rendering'] | ['computer-vision'] | [ 5.57209432e-01 3.55548114e-02 4.71842200e-01 -2.82006174e-01
-6.40815616e-01 -2.25227892e-01 3.13349634e-01 -7.58620560e-01
4.40520883e-01 4.31953967e-01 -4.28701751e-02 3.12383659e-03
1.61172032e-01 -8.38438630e-01 -8.02870393e-01 -7.54336715e-01
3.93923193e-01 2.92269886e-01 4.92877513e-01 -2.32136205e-01
1.74418360e-01 7.14391053e-01 -1.76900554e+00 6.24260604e-01
9.44302320e-01 1.12907255e+00 2.27625743e-01 3.01659882e-01
2.13773884e-02 4.39363956e-01 -4.58168685e-01 -3.27385604e-01
6.31159544e-01 -6.20936379e-02 -9.36013013e-02 5.11119291e-02
7.59196341e-01 -6.29682660e-01 -2.98737615e-01 1.11357570e+00
3.36216986e-01 -1.11751701e-03 8.03867936e-01 -1.04143727e+00
-8.21614921e-01 -2.99737185e-01 -9.89047289e-01 -6.03902876e-01
4.83377665e-01 1.07100442e-01 5.67952335e-01 -1.18900156e+00
5.35433948e-01 1.42226124e+00 7.55052924e-01 7.17359364e-01
-1.15014184e+00 -7.98820496e-01 2.74049342e-01 -2.44858071e-01
-1.32889843e+00 -5.02966166e-01 1.28331566e+00 -3.72282505e-01
3.77149761e-01 6.27658367e-01 5.87356031e-01 1.17386508e+00
1.18328124e-01 6.70186877e-01 1.23862362e+00 -1.86277017e-01
-5.27528897e-02 4.30819243e-02 -2.13307198e-02 6.06265843e-01
4.94278967e-01 3.61113757e-01 -6.43828154e-01 -5.04260778e-01
1.31920075e+00 1.07815608e-01 -8.19823265e-01 -5.53154886e-01
-1.02517331e+00 3.56574953e-01 3.66410166e-01 -2.59121984e-01
-3.72181952e-01 5.95244877e-02 -1.43575341e-01 7.58610442e-02
8.77537906e-01 1.95671111e-01 -1.87822163e-01 5.39069951e-01
-4.31159675e-01 3.90469521e-01 6.83609784e-01 9.95483577e-01
7.16659784e-01 3.19582075e-01 1.12801306e-01 1.08294380e+00
5.88163197e-01 9.38851118e-01 -2.07824975e-01 -1.31639564e+00
2.25698471e-01 5.04461884e-01 6.32155061e-01 -1.15009058e+00
1.10119641e-01 -4.77612734e-01 -8.64956617e-01 7.18494773e-01
4.70537730e-02 2.20733076e-01 -9.06855643e-01 1.19831240e+00
8.46837759e-01 3.98573458e-01 -6.21274896e-02 1.34726000e+00
1.08739734e+00 8.26480806e-01 -6.64235413e-01 -2.86562610e-02
9.53455210e-01 -8.73507261e-01 -6.02491796e-01 4.03805003e-02
4.31255996e-02 -9.67625678e-01 1.20395744e+00 6.86752737e-01
-1.14725244e+00 -2.64363170e-01 -1.00881720e+00 -6.82422966e-02
2.06150055e-01 1.05091214e-01 5.36874771e-01 2.74885148e-01
-8.51111829e-01 4.42228675e-01 -6.45712793e-01 2.73845285e-01
2.88919240e-01 3.69516090e-02 -1.83816969e-01 -3.42624068e-01
-7.75316179e-01 3.47613484e-01 -2.78802127e-01 5.14122903e-01
-7.89943099e-01 -1.02769673e+00 -7.59180665e-01 -2.89387584e-01
3.99747968e-01 -7.78068542e-01 1.22948980e+00 -1.27309430e+00
-1.92914438e+00 9.90265131e-01 -2.15007514e-01 4.06417519e-01
5.12355268e-01 -4.94647175e-01 -2.67930597e-01 4.87578958e-02
-2.41138294e-01 4.01343266e-03 1.04619455e+00 -2.30109119e+00
1.91548448e-02 -4.10430133e-01 -6.15211427e-02 3.41268718e-01
1.82665735e-01 -8.83769691e-02 -3.79285097e-01 -6.38398528e-01
5.23980737e-01 -7.25739062e-01 -7.88552463e-02 5.81214249e-01
-5.39604843e-01 1.43459484e-01 1.04531062e+00 -6.34626985e-01
6.46533191e-01 -2.22480249e+00 -1.30807996e-01 9.92684588e-02
3.48595262e-01 3.98625761e-01 -4.14105862e-01 1.64598405e-01
-2.32841889e-03 -3.82617086e-01 -3.68852079e-01 -5.11047721e-01
-1.54410884e-01 -4.23012823e-02 -5.89853764e-01 6.16125405e-01
4.62406129e-02 5.87805629e-01 -6.56834662e-01 -7.31917918e-02
1.18289277e-01 1.01876795e+00 -4.10036176e-01 5.22989571e-01
-4.24958915e-01 5.82399189e-01 -5.73602796e-01 1.08731556e+00
1.31467772e+00 -2.93371022e-01 -5.26381359e-02 -4.66422260e-01
-3.05176646e-01 2.25452289e-01 -1.21351302e+00 1.27479470e+00
-2.79895037e-01 3.55878383e-01 5.39278150e-01 -5.16274095e-01
1.20837688e+00 2.16977775e-01 4.24714148e-01 -8.60643506e-01
-1.25154823e-01 4.49600726e-01 -3.68649334e-01 -2.83105493e-01
3.29167426e-01 -2.46018291e-01 5.67778766e-01 3.07554659e-02
-7.64710605e-01 -3.69148523e-01 -6.52745366e-01 -9.29179490e-02
8.11948001e-01 6.15509510e-01 -7.17354119e-02 -1.25535607e-01
4.91710752e-01 -1.26378566e-01 8.28651309e-01 3.35118324e-01
2.61136353e-01 1.26332986e+00 1.42390072e-01 -6.48320794e-01
-9.52158868e-01 -1.43736470e+00 -1.02690995e-01 4.99762505e-01
6.47499859e-01 -4.30568494e-03 -5.99238098e-01 -2.25392759e-01
1.58034980e-01 6.98836684e-01 -6.00782454e-01 1.94116856e-03
-8.06084692e-01 -4.31216925e-01 4.57174033e-02 1.95762902e-01
4.49299932e-01 -9.44758773e-01 -5.49633980e-01 -2.39295289e-02
-2.77590573e-01 -9.79380012e-01 -3.91098797e-01 -5.40112019e-01
-7.96902597e-01 -1.16877306e+00 -1.00435996e+00 -3.75559062e-01
8.05324435e-01 8.32009614e-01 1.34664702e+00 2.18102157e-01
-3.17037344e-01 5.81340194e-01 -2.11563110e-01 -4.36807901e-01
-2.39057690e-01 -7.10374534e-01 -1.00873888e-01 4.38202918e-01
-8.79735053e-02 -7.76532114e-01 -7.44764447e-01 7.74542570e-01
-8.27004373e-01 6.86616123e-01 3.49270403e-01 5.66837728e-01
8.17438364e-01 -3.96597266e-01 -1.65216543e-03 -8.83625209e-01
6.41774014e-03 -2.69525349e-01 -9.79150414e-01 2.08997075e-02
-1.25500873e-01 -4.45866108e-01 3.15166146e-01 -6.19772553e-01
-1.45155442e+00 -1.74326405e-01 4.80923653e-02 -9.63592649e-01
-1.37963176e-01 -5.11850172e-04 -3.19117367e-01 -3.03831577e-01
4.02063727e-01 2.44941607e-01 -4.84670214e-02 -7.35215664e-01
-3.32718730e-01 5.17425656e-01 2.70587832e-01 -6.07757092e-01
1.13991451e+00 9.63793695e-01 1.70178264e-01 -1.30182540e+00
-1.08452713e+00 -3.28480363e-01 -1.97568893e-01 -4.66989279e-01
4.90184546e-01 -8.47637773e-01 -8.22992146e-01 9.53068137e-01
-1.46482098e+00 -7.70616353e-01 -1.10371992e-01 2.03917652e-01
-5.25337338e-01 2.60180056e-01 -4.76256549e-01 -9.64016855e-01
-3.60007256e-01 -8.99776578e-01 1.44918668e+00 9.56631228e-02
1.48256391e-01 -6.45222843e-01 6.40990511e-02 6.21523261e-01
4.63816196e-01 5.21197975e-01 6.65587485e-01 3.55484933e-01
-1.27631330e+00 1.87792048e-01 -5.44407248e-01 3.86069059e-01
2.92106718e-01 1.41690120e-01 -1.18885505e+00 -3.08167964e-01
3.92021984e-01 -1.35968894e-01 5.86873353e-01 4.52001482e-01
1.25513303e+00 -3.42963636e-01 -2.81127185e-01 1.22649479e+00
1.61863625e+00 1.43629298e-01 9.11890805e-01 2.79606640e-01
1.08635080e+00 6.71050668e-01 8.42080414e-01 3.94085050e-01
2.48160914e-01 8.75885904e-01 8.29112232e-01 -3.46829623e-01
-4.08048809e-01 2.96659507e-02 5.01142025e-01 6.79158032e-01
-3.96220416e-01 -3.81253868e-01 -6.49474978e-01 5.48413023e-03
-1.50350368e+00 -5.29168248e-01 -5.48646569e-01 2.25062656e+00
5.28097332e-01 -2.51426607e-01 -4.07233387e-01 -2.24829286e-01
6.11488998e-01 2.80712217e-01 -7.09734738e-01 7.42277801e-02
-2.38274798e-01 -5.27944900e-02 1.92553267e-01 7.05075979e-01
-5.95547855e-01 7.21007228e-01 5.14091396e+00 5.09971738e-01
-1.50556171e+00 6.62204670e-03 3.72278184e-01 -9.06153247e-02
-8.07519197e-01 -3.01856160e-01 -8.03675354e-01 2.36073315e-01
2.57697552e-01 3.74137610e-01 4.14561242e-01 5.61518371e-01
9.86569822e-02 2.60365475e-02 -7.23214865e-01 1.14343154e+00
1.98522702e-01 -1.17622721e+00 3.14628303e-01 -6.84649348e-02
7.77205110e-01 7.40878507e-02 2.95723438e-01 -2.08974242e-01
1.17135987e-01 -8.28223109e-01 7.35844374e-01 8.44858110e-01
1.05922818e+00 -2.72405535e-01 2.93940216e-01 1.38311341e-01
-1.13338757e+00 5.19711792e-01 -3.37853491e-01 2.30253786e-01
1.14546210e-01 6.66637778e-01 -3.50617558e-01 6.69751883e-01
7.37117708e-01 8.25914800e-01 -5.61084859e-02 9.43387508e-01
-1.77924290e-01 3.40181321e-01 -3.56864601e-01 3.50161999e-01
-6.17619567e-02 -5.10419607e-01 1.06848049e+00 7.84516454e-01
4.38450128e-01 6.05901361e-01 3.74111570e-02 1.30692601e+00
1.51293606e-01 -1.25344574e-01 -6.33833110e-01 2.92169750e-01
8.81680176e-02 1.07778263e+00 -3.66887748e-01 -6.83377832e-02
-4.14539933e-01 1.06412244e+00 1.53242633e-01 8.69416595e-01
-7.92909324e-01 -1.37271127e-02 8.65970135e-01 7.57800341e-01
3.80086526e-02 -2.08980367e-01 -2.49251008e-01 -1.39183390e+00
3.71017665e-01 -8.92342389e-01 -2.10314631e-01 -1.30311882e+00
-1.37654591e+00 4.95880187e-01 -1.34240478e-01 -1.50847316e+00
5.37974417e-01 -8.44960749e-01 -4.78177160e-01 1.06015277e+00
-1.79883492e+00 -1.26699317e+00 -6.86428308e-01 5.03936827e-01
4.74070996e-01 1.68594584e-01 6.81373119e-01 2.13244691e-01
-2.54508048e-01 1.01570100e-01 1.15494058e-01 -2.99229801e-01
6.60125494e-01 -7.22246051e-01 3.38402599e-01 5.18182278e-01
-4.60501492e-01 5.13317883e-01 7.04687297e-01 -8.06062877e-01
-1.73663080e+00 -1.02323818e+00 -1.11618198e-01 -2.33358875e-01
5.68966381e-02 -4.78328258e-01 -1.41125131e+00 5.81945181e-01
-6.03664070e-02 2.41004273e-01 3.41405362e-01 -2.79592186e-01
-6.99653029e-01 -2.01830432e-01 -1.13752007e+00 7.38982379e-01
1.22405112e+00 -1.94072098e-01 -2.33448178e-01 1.72029316e-01
8.28451335e-01 -8.50704014e-01 -4.65003639e-01 9.16447937e-01
7.13431597e-01 -1.46118557e+00 1.13173902e+00 2.56425664e-02
5.09877920e-01 -4.84512001e-01 -4.48347926e-01 -1.15069044e+00
-2.91594595e-01 -7.08721876e-01 -3.18415403e-01 8.01566601e-01
-5.23672551e-02 -1.10698783e+00 8.93347085e-01 4.90552038e-01
-6.03227794e-01 -1.08741128e+00 -6.03043675e-01 -7.24201798e-01
-3.81606251e-01 -5.58596775e-02 5.05065382e-01 9.98995304e-01
-1.08126223e+00 -9.57958326e-02 -3.10164809e-01 8.08194876e-01
1.33316910e+00 5.79499304e-01 1.01919007e+00 -1.34873915e+00
-3.14316392e-01 -8.00860226e-02 1.10840246e-01 -1.28399742e+00
5.43409213e-02 -4.56218064e-01 2.48739690e-01 -1.57764256e+00
8.49368051e-03 -8.76949489e-01 2.19611893e-03 1.38412237e-01
6.88120946e-02 3.77301335e-01 -1.53723331e-02 3.16326141e-01
-1.81982711e-01 8.80467355e-01 1.83340120e+00 1.29158080e-01
-2.32933179e-01 1.05209745e-01 -4.54109937e-01 1.15212214e+00
6.67023838e-01 -3.12264413e-01 -1.58585370e-01 -7.89585710e-01
2.23877952e-01 7.60571659e-02 6.85598731e-01 -7.79429257e-01
-2.63275653e-01 -4.77718711e-01 3.53149742e-01 -9.90093291e-01
1.01515973e+00 -1.03356385e+00 6.41536534e-01 2.71639377e-01
1.37402624e-01 -4.94082063e-01 5.49992472e-02 4.35500830e-01
1.28482678e-03 5.13689965e-02 1.12120640e+00 -1.33139715e-01
-1.48825288e-01 4.96804893e-01 1.72552466e-01 -1.98611379e-01
6.93430781e-01 -4.09468442e-01 -3.60910505e-01 -1.51131943e-01
-9.65876430e-02 -1.06909953e-01 9.45597887e-01 2.81477988e-01
1.16827905e+00 -1.24331903e+00 -6.86029851e-01 5.22618055e-01
-6.36311471e-02 6.03341639e-01 3.56990457e-01 6.45306766e-01
-7.80501664e-01 -2.61495933e-02 1.02658279e-01 -6.50848269e-01
-1.44123089e+00 1.37574628e-01 6.38961792e-01 1.60467520e-01
-1.04227173e+00 8.34248126e-01 1.14451635e+00 -6.53517246e-01
1.23096265e-01 -2.21724421e-01 2.54567385e-01 -5.50757885e-01
5.58270156e-01 5.40551782e-01 -2.26264060e-01 -5.83053648e-01
2.30801590e-02 9.66325343e-01 2.77796872e-02 1.26083061e-01
1.49522793e+00 1.57130603e-02 -2.96604544e-01 5.68136871e-01
9.45447266e-01 3.61021221e-01 -1.61000407e+00 -1.65090740e-01
-7.24366844e-01 -9.87143278e-01 -3.84110063e-02 -5.96448183e-01
-1.11444986e+00 8.18237662e-01 3.05677950e-01 -1.64011031e-01
1.07129359e+00 -2.05617726e-01 1.05207586e+00 1.86934188e-01
6.16707623e-01 -3.93345445e-01 3.11402977e-01 5.28650343e-01
1.50303149e+00 -1.06897378e+00 5.75544201e-02 -1.22118080e+00
-2.27325708e-01 9.67749476e-01 9.25117910e-01 -4.27343279e-01
6.39655411e-01 4.24956203e-01 2.86047459e-01 -4.48788494e-01
-4.78360474e-01 4.33738440e-01 4.80115592e-01 6.12534702e-01
2.26511776e-01 -6.15784042e-02 3.53833377e-01 2.30349556e-01
1.43645555e-01 -1.45104647e-01 4.18789387e-01 7.66044736e-01
-2.82515228e-01 -7.76836216e-01 -7.07622170e-01 2.19656155e-01
-2.23875850e-01 -1.06750540e-02 -1.95684969e-01 6.27016306e-01
-1.60436198e-01 2.95679003e-01 -1.81181394e-02 1.72970928e-02
5.32673120e-01 -4.68995064e-01 7.64040887e-01 -7.96759546e-01
-2.06143573e-01 2.41146177e-01 2.11416166e-02 -8.26774776e-01
-3.55254948e-01 -5.35902798e-01 -1.10880399e+00 -4.87544909e-02
-3.65916580e-01 -2.39804849e-01 5.49132824e-01 2.92989433e-01
3.46354634e-01 3.76629531e-01 7.47825086e-01 -1.45612788e+00
-3.79819334e-01 -5.78392446e-01 -7.17502534e-01 3.85553181e-01
6.60967529e-01 -9.75953937e-01 -7.03223407e-01 -5.63842356e-02] | [9.729333877563477, -3.069593667984009] |
88b59335-d4ff-4c1a-9fdc-cdc5714a0446 | the-big-data-myth-using-diffusion-models-for | 2306.09762 | null | https://arxiv.org/abs/2306.09762v1 | https://arxiv.org/pdf/2306.09762v1.pdf | The Big Data Myth: Using Diffusion Models for Dataset Generation to Train Deep Detection Models | Despite the notable accomplishments of deep object detection models, a major challenge that persists is the requirement for extensive amounts of training data. The process of procuring such real-world data is a laborious undertaking, which has prompted researchers to explore new avenues of research, such as synthetic data generation techniques. This study presents a framework for the generation of synthetic datasets by fine-tuning pretrained stable diffusion models. The synthetic datasets are then manually annotated and employed for training various object detection models. These detectors are evaluated on a real-world test set of 331 images and compared against a baseline model that was trained on real-world images. The results of this study reveal that the object detection models trained on synthetic data perform similarly to the baseline model. In the context of apple detection in orchards, the average precision deviation with the baseline ranges from 0.09 to 0.12. This study illustrates the potential of synthetic data generation techniques as a viable alternative to the collection of extensive training data for the training of deep models. | ['Klaas Dijkstra', 'Maya Aghaei', 'Roy Voetman'] | 2023-06-16 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 5.36970854e-01 2.72140533e-01 1.20833300e-01 -2.20485300e-01
-5.29335439e-01 -7.79377282e-01 9.74721909e-01 2.39316598e-01
-4.72241461e-01 4.63627249e-01 -4.46555465e-01 -1.44383565e-01
2.66093403e-01 -1.01653671e+00 -6.29529715e-01 -7.24964261e-01
8.60734284e-02 4.12132472e-01 6.27667487e-01 -5.41711412e-02
4.19200450e-01 6.96606159e-01 -1.75311708e+00 3.30263197e-01
6.87433481e-01 7.52510607e-01 5.91698229e-01 8.31084609e-01
-6.32270649e-02 4.21561450e-01 -1.09094787e+00 -4.73530143e-01
4.05340075e-01 -2.90845007e-01 -3.56421977e-01 5.00633657e-01
5.13905108e-01 -5.73629797e-01 -3.21918987e-02 1.09838593e+00
5.10765433e-01 -1.43193245e-01 7.28667557e-01 -1.34827805e+00
-8.39872599e-01 4.12630171e-01 -7.02847123e-01 3.02880079e-01
-1.10300422e-01 5.80977261e-01 6.66049957e-01 -8.54093015e-01
7.62495935e-01 1.07317114e+00 6.25460863e-01 5.37814736e-01
-1.58414209e+00 -6.18473351e-01 -9.60353166e-02 -2.85066843e-01
-1.24624336e+00 -3.06146920e-01 4.99486953e-01 -5.63958764e-01
5.73287070e-01 4.67242636e-02 5.17009735e-01 1.29751337e+00
7.14098141e-02 9.30702090e-01 1.10048330e+00 -6.49284303e-01
4.00354445e-01 6.68278992e-01 7.06002638e-02 5.62186539e-01
8.45659018e-01 4.26505208e-01 -1.51366115e-01 1.66078750e-03
8.45728099e-01 -3.00272226e-01 1.85160115e-01 -5.90591371e-01
-1.10225892e+00 1.12370253e+00 6.62612259e-01 2.33318329e-01
-5.92061102e-01 -2.12329775e-01 3.63490999e-01 -1.14400320e-01
6.24814510e-01 7.62748837e-01 -9.80441347e-02 3.03382903e-01
-1.20389807e+00 4.85823035e-01 7.44057477e-01 9.49787974e-01
3.12297493e-01 3.10159057e-01 -3.49705696e-01 7.04128981e-01
3.71991843e-01 4.43753719e-01 4.36951935e-01 -7.82527149e-01
1.73529893e-01 9.06955540e-01 4.56316531e-01 -9.77249563e-01
-8.82608369e-02 -3.59130561e-01 -4.65485364e-01 6.42699599e-01
6.69969499e-01 -2.59734333e-01 -1.26798272e+00 1.36421859e+00
4.08011645e-01 -2.36396551e-01 1.56099260e-01 5.73466182e-01
6.46069050e-01 6.91741109e-01 3.13417584e-01 1.57220185e-01
1.08649313e+00 -9.11509395e-01 -3.34602982e-01 -2.33068377e-01
4.46625650e-01 -1.04165876e+00 9.60609257e-01 2.66434789e-01
-8.07252467e-01 -7.31481433e-01 -1.22714186e+00 3.83579433e-01
-6.70430124e-01 5.90796232e-01 6.55309319e-01 9.79954898e-01
-9.74953771e-01 3.56927931e-01 -8.44189644e-01 -7.54926920e-01
7.68828750e-01 2.27244616e-01 -2.90376833e-03 -1.42919600e-01
-6.27841294e-01 8.68855715e-01 7.97383189e-01 4.34108786e-02
-1.40085006e+00 -5.20745516e-01 -4.21036243e-01 -1.28570274e-01
2.02948228e-01 -1.78857759e-01 1.32197392e+00 -8.66958976e-01
-9.84031320e-01 1.06898856e+00 3.75943899e-01 -7.54488766e-01
9.05938625e-01 -1.25798032e-01 -2.06813797e-01 -5.33538610e-02
4.43873852e-01 1.08748484e+00 9.22461748e-01 -1.68597996e+00
-8.56426358e-01 -3.05028647e-01 3.55751067e-02 -1.92714483e-01
-2.48342231e-01 9.87066478e-02 -1.41742095e-01 -8.01004887e-01
-2.34839946e-01 -1.11858249e+00 -4.08745289e-01 4.46992844e-01
-6.68543398e-01 -9.34396461e-02 1.06996930e+00 -3.71001244e-01
7.96072543e-01 -2.12645698e+00 -4.54648107e-01 -1.97447930e-02
1.92576572e-01 7.09949791e-01 -3.99331570e-01 2.75655031e-01
-7.69768804e-02 1.64932072e-01 -3.48227710e-01 1.49323745e-02
-2.07835302e-01 -6.38565570e-02 -2.63869971e-01 1.89031735e-01
6.79960608e-01 7.86180139e-01 -1.03634751e+00 -4.39190447e-01
2.99472988e-01 3.87706429e-01 -8.34056363e-02 3.14393640e-01
-4.89477396e-01 5.69274202e-02 -4.46690470e-01 8.03902984e-01
8.43286037e-01 -2.39437670e-01 -8.98275301e-02 6.40345439e-02
-2.08980769e-01 -2.11522236e-01 -1.18309653e+00 1.19289100e+00
1.90836377e-02 9.22441721e-01 -1.43517219e-02 -6.81782603e-01
1.11765528e+00 6.64506778e-02 2.99249977e-01 -4.41954523e-01
3.95428240e-02 1.30028754e-01 2.80708492e-01 -3.77648383e-01
6.45191729e-01 2.57356346e-01 2.25888535e-01 6.97140157e-01
2.46338844e-02 -4.89214212e-01 7.24660218e-01 3.19646239e-01
1.12589049e+00 3.09446901e-01 -1.28227562e-01 -2.04238653e-01
1.81092486e-01 6.71810567e-01 8.01552609e-02 8.22286606e-01
-2.47609168e-01 5.66697180e-01 4.06981677e-01 -5.06882012e-01
-1.36944902e+00 -1.20755172e+00 -1.84604079e-01 7.85568595e-01
4.31295745e-02 -4.16431203e-02 -1.05419564e+00 -7.96600342e-01
1.22504458e-01 1.02862442e+00 -8.69537354e-01 -2.19550386e-01
-2.03864485e-01 -1.38228321e+00 7.30432868e-01 5.23089647e-01
7.03127682e-01 -1.41916704e+00 -1.14984739e+00 3.48670751e-01
2.82446474e-01 -9.81739819e-01 1.02953978e-01 1.33095786e-01
-8.81633937e-01 -1.13459504e+00 -9.31785107e-01 -6.80665016e-01
8.58361602e-01 6.45302713e-01 1.09958160e+00 1.86593402e-02
-8.68626237e-01 5.05198678e-03 -4.47966129e-01 -1.03600597e+00
-1.00549984e+00 1.80579767e-01 -1.80386424e-01 -1.88157454e-01
6.27770126e-01 1.14631683e-01 -6.39982820e-01 2.96385527e-01
-1.17283225e+00 -1.04022615e-01 8.69633794e-01 8.77060294e-01
5.11132300e-01 7.13721216e-02 6.76459849e-01 -9.13309872e-01
9.57159638e-01 -4.11456168e-01 -1.03966832e+00 3.64497781e-01
-6.34428024e-01 -9.16834027e-02 1.31266043e-01 -7.70738065e-01
-1.24727166e+00 5.14245808e-01 3.75305295e-01 -1.10336065e-01
-3.63174796e-01 -1.03309333e-01 1.90901935e-01 -4.52344082e-02
1.20762300e+00 1.11134201e-02 6.70700371e-02 -2.88140208e-01
5.29875576e-01 5.21772921e-01 2.93132633e-01 -2.35908300e-01
9.04358327e-01 4.55932319e-01 -3.60822588e-01 -1.04710042e+00
-6.95108235e-01 -1.21289670e-01 -9.36805248e-01 -2.26555422e-01
7.42240846e-01 -5.73249280e-01 1.46336272e-01 6.83160305e-01
-1.25070083e+00 -4.21687961e-01 -6.34391665e-01 2.52530754e-01
-1.84770301e-01 1.41116753e-01 -2.91670144e-01 -8.56036365e-01
-2.47533962e-01 -1.07146227e+00 1.11514735e+00 2.78272778e-01
-3.25477630e-01 -6.59241736e-01 2.76563372e-02 1.59587339e-01
3.36492419e-01 5.49447775e-01 7.21916437e-01 -4.90697891e-01
-7.85932302e-01 -7.46417642e-01 -7.11282253e-01 4.71528172e-01
2.64571190e-01 5.78926802e-01 -1.19855309e+00 -1.85488954e-01
-9.13242400e-02 -5.58972418e-01 7.31945693e-01 4.04036552e-01
9.22242224e-01 1.52365297e-01 -6.66989505e-01 1.17972381e-01
1.40146947e+00 4.67825651e-01 4.68331844e-01 4.79946047e-01
4.09217149e-01 7.74192631e-01 8.72840047e-01 3.49959373e-01
-4.32117097e-02 3.62622112e-01 5.32265663e-01 -3.67899418e-01
-3.22380811e-01 -1.89111277e-01 -5.28047457e-02 -6.64741695e-02
6.04470931e-02 -4.63639498e-01 -8.28617990e-01 8.21074724e-01
-1.43534458e+00 -1.05254996e+00 -2.98792511e-01 2.09186077e+00
5.72289288e-01 3.47360134e-01 2.36957729e-01 2.05830082e-01
8.27037513e-01 1.02815196e-01 -8.10132742e-01 -3.29213530e-01
4.44452325e-03 2.17477694e-01 5.47819853e-01 -4.92845476e-02
-1.36205828e+00 1.08242321e+00 7.06436825e+00 5.13118625e-01
-1.19718468e+00 -1.40673831e-01 6.53018236e-01 4.92866077e-02
3.47037435e-01 -2.89359540e-01 -8.65256608e-01 4.37182039e-01
1.01695919e+00 -2.19679564e-01 -1.51779931e-02 9.98614848e-01
1.94141328e-01 -5.29853821e-01 -9.38770115e-01 4.08932149e-01
-6.59999177e-02 -1.28388894e+00 1.29209533e-01 1.78039372e-01
9.82124448e-01 1.23373814e-01 1.61359608e-01 1.81004554e-01
8.42240930e-01 -8.32346678e-01 7.06134796e-01 1.15919657e-01
5.30378580e-01 -3.32491815e-01 5.70311964e-01 4.35590327e-01
-7.79776454e-01 -4.62344177e-02 -5.31879425e-01 2.85362035e-01
-8.90449509e-02 4.10644203e-01 -1.50796843e+00 -5.45253567e-02
5.90081632e-01 3.14496905e-01 -1.10600770e+00 1.19222581e+00
-2.94593796e-02 5.69472015e-01 -4.32749242e-01 -2.73424059e-01
3.29414785e-01 8.83917958e-02 2.04966471e-01 1.15097940e+00
2.32261911e-01 -4.88209218e-01 -8.10128972e-02 1.30383813e+00
1.71706565e-02 -9.92460325e-02 -1.08622968e+00 -4.84136254e-01
4.27521825e-01 1.31290340e+00 -1.30783379e+00 -3.33017498e-01
-2.61122227e-01 9.18268442e-01 6.30833283e-02 2.05470979e-01
-7.73168027e-01 -3.76904517e-01 6.34888634e-02 1.99775532e-01
3.74578863e-01 -1.04561463e-01 -4.97636646e-01 -6.28502309e-01
2.56053731e-03 -8.94533634e-01 4.81111556e-02 -7.34098136e-01
-1.18648148e+00 7.94890046e-01 5.69750555e-02 -1.13480985e+00
-2.52206266e-01 -7.01188982e-01 -6.77949250e-01 8.58882725e-01
-1.13458586e+00 -1.25815451e+00 -7.73420274e-01 1.53667545e-02
8.77891660e-01 -3.86220634e-01 1.01916921e+00 4.33917381e-02
-5.85383654e-01 3.22557092e-01 2.79540479e-01 4.99747507e-02
4.21548605e-01 -1.09609592e+00 9.22379911e-01 9.06242907e-01
3.58516723e-01 4.40932900e-01 6.98064387e-01 -7.43676305e-01
-8.75514805e-01 -1.28560114e+00 3.16253603e-01 -7.14074790e-01
6.12618625e-01 -3.01258028e-01 -8.22233021e-01 3.36364001e-01
2.35724926e-01 -1.16010211e-01 4.82572168e-01 -4.92641658e-01
-7.52708763e-02 8.86096954e-02 -1.35436010e+00 4.47934359e-01
5.93488872e-01 -1.12185061e-01 -3.80797982e-01 4.47860658e-01
3.98603767e-01 -8.43754262e-02 -4.85936940e-01 3.51635844e-01
5.73878407e-01 -8.22507858e-01 9.37709868e-01 -4.41024154e-01
5.47075212e-01 -1.69206262e-01 2.79803704e-02 -1.37198770e+00
-1.15154728e-01 -1.89588383e-01 5.94031662e-02 1.27013540e+00
6.77228034e-01 -2.11461529e-01 1.20986104e+00 6.24390662e-01
3.24368507e-01 -4.04796869e-01 -3.01697075e-01 -7.44374752e-01
-1.57353505e-02 -8.96886587e-02 2.91389018e-01 8.21164072e-01
-8.79846275e-01 1.74782097e-01 1.67280331e-01 -2.70151608e-02
8.22801948e-01 3.46193905e-03 1.05723917e+00 -1.43539965e+00
1.03276428e-02 -3.59590948e-01 -4.35425460e-01 -5.97180128e-01
-2.12026179e-01 -5.82383037e-01 1.73846215e-01 -1.70207882e+00
6.49708882e-02 -6.15454495e-01 3.03394850e-02 1.19647019e-01
-1.01309635e-01 4.24544334e-01 2.15240628e-01 8.67423266e-02
1.11137014e-02 2.84705430e-01 1.36715567e+00 -3.13147068e-01
-3.46623361e-01 2.42518961e-01 -6.17192626e-01 7.02493429e-01
1.05414867e+00 -6.46542609e-01 -6.29834235e-01 -4.02966768e-01
-3.64360869e-01 -4.81314659e-01 5.02777219e-01 -1.09615695e+00
-1.19544953e-01 -6.19092733e-02 7.18636334e-01 -6.91630960e-01
3.84392679e-01 -6.56784415e-01 -1.84806973e-01 6.95064962e-01
-2.82309860e-01 1.40859187e-01 4.55811888e-01 6.42327070e-01
8.29594731e-02 -5.32228351e-01 9.74714935e-01 -3.96967977e-01
-1.00444901e+00 -1.05397321e-01 -3.81502122e-01 -7.42079243e-02
1.72700047e+00 -5.23722768e-01 -3.90231580e-01 4.42325845e-02
-6.68289125e-01 -1.00527540e-01 4.69065428e-01 6.64521158e-01
5.53564668e-01 -9.28402543e-01 -7.61058152e-01 3.57909232e-01
2.04776898e-01 1.14444494e-01 -3.12576920e-01 1.54939294e-01
-8.49725544e-01 1.92457303e-01 -6.03712618e-01 -7.18091607e-01
-1.23611534e+00 5.94945908e-01 1.96699828e-01 -1.24433950e-01
-5.86563647e-01 8.19456398e-01 2.62047499e-01 -1.57153055e-01
2.56420106e-01 -3.53079498e-01 -3.28836404e-03 2.21284047e-01
4.64527011e-01 4.02789742e-01 -1.48114666e-01 -5.44621885e-01
6.33732695e-03 -8.34146701e-03 -4.37721848e-01 -2.02915326e-01
1.29998755e+00 4.79272604e-01 3.31078559e-01 2.25926057e-01
6.14396513e-01 -4.78206396e-01 -1.40233266e+00 8.56856257e-02
3.89189541e-01 -6.20375216e-01 -1.17630344e-02 -8.93083811e-01
-9.49047863e-01 9.06132579e-01 1.07553554e+00 5.82811415e-01
8.01179647e-01 -1.67997077e-01 2.19071582e-01 3.56980175e-01
2.96274751e-01 -1.07463145e+00 4.14914638e-01 -3.92321236e-02
9.13381398e-01 -1.50195169e+00 2.89347004e-02 -4.03187186e-01
-6.69930995e-01 8.47393215e-01 9.27242994e-01 -3.56543034e-01
4.75160837e-01 4.12432432e-01 2.50434786e-01 -3.44889015e-01
-4.38195199e-01 -1.10087737e-01 -1.11226305e-01 9.66793060e-01
4.98833388e-01 7.01870769e-02 -1.93629384e-01 -1.48610309e-01
1.50586545e-01 2.04621077e-01 6.47919893e-01 1.18846560e+00
-3.71444106e-01 -1.12060273e+00 -4.14987236e-01 7.04393268e-01
-3.59236330e-01 2.01487884e-01 -8.36653829e-01 1.20663679e+00
2.17346828e-02 9.58232760e-01 1.05568748e-02 -2.78438814e-02
4.28111047e-01 -3.36711332e-02 3.29718560e-01 -9.21789348e-01
-5.55959225e-01 -2.03570336e-01 1.54036107e-02 -1.04655415e-01
-5.12736380e-01 -6.62096798e-01 -8.06391180e-01 -4.96714227e-02
-7.09335208e-01 -2.29203373e-01 9.04548407e-01 3.90431017e-01
2.08077028e-01 6.26399994e-01 3.30531836e-01 -1.09697378e+00
-8.51639211e-01 -1.15439332e+00 -5.69323063e-01 4.30694461e-01
8.71142820e-02 -7.33917356e-01 -2.39614956e-02 3.77511263e-01] | [8.659360885620117, -0.9733677506446838] |
866c7535-57bc-442f-a1cb-fd50d69cc256 | designing-optimal-behavioral-experiments | 2305.07721 | null | https://arxiv.org/abs/2305.07721v1 | https://arxiv.org/pdf/2305.07721v1.pdf | Designing Optimal Behavioral Experiments Using Machine Learning | Computational models are powerful tools for understanding human cognition and behavior. They let us express our theories clearly and precisely, and offer predictions that can be subtle and often counter-intuitive. However, this same richness and ability to surprise means our scientific intuitions and traditional tools are ill-suited to designing experiments to test and compare these models. To avoid these pitfalls and realize the full potential of computational modeling, we require tools to design experiments that provide clear answers about what models explain human behavior and the auxiliary assumptions those models must make. Bayesian optimal experimental design (BOED) formalizes the search for optimal experimental designs by identifying experiments that are expected to yield informative data. In this work, we provide a tutorial on leveraging recent advances in BOED and machine learning to find optimal experiments for any kind of model that we can simulate data from, and show how by-products of this procedure allow for quick and straightforward evaluation of models and their parameters against real experimental data. As a case study, we consider theories of how people balance exploration and exploitation in multi-armed bandit decision-making tasks. We validate the presented approach using simulations and a real-world experiment. As compared to experimental designs commonly used in the literature, we show that our optimal designs more efficiently determine which of a set of models best account for individual human behavior, and more efficiently characterize behavior given a preferred model. We provide code to replicate all analyses as well as tutorial notebooks and pointers to adapt the methodology to other experimental settings. | ['Christopher G. Lucas', 'Michael U. Gutmann', 'Peggy Seriès', 'Neil R. Bramley', 'Steven Kleinegesse', 'Simon Valentin'] | 2023-05-12 | null | null | null | null | ['experimental-design'] | ['methodology'] | [-1.42889783e-01 -3.15335393e-01 -5.34811199e-01 -3.92513841e-01
-1.33739397e-01 -5.82936108e-01 5.12815833e-01 3.16895209e-02
-7.34590173e-01 7.04148293e-01 -6.77967668e-02 -8.00260603e-01
-5.80705523e-01 -4.00246054e-01 -5.69495499e-01 -4.44989890e-01
-1.21554077e-01 5.89713037e-01 -2.57911414e-01 -1.03312902e-01
6.11997068e-01 3.44840020e-01 -1.78537202e+00 1.07869841e-01
8.25434566e-01 7.60794997e-01 2.78196305e-01 7.18399286e-01
1.60096183e-01 4.89105523e-01 -4.02053744e-01 -5.63281894e-01
3.24552476e-01 -3.84108484e-01 -7.35435426e-01 1.18118942e-01
9.62880254e-02 -4.17160630e-01 1.47115337e-02 8.42258692e-01
3.49249691e-01 1.36535957e-01 6.59865558e-01 -1.21212316e+00
-7.96407998e-01 6.17113411e-01 -3.63686025e-01 3.88207912e-01
3.93718064e-01 4.47157383e-01 1.01235092e+00 -1.70582727e-01
1.50556773e-01 1.50159383e+00 6.59441888e-01 4.76770043e-01
-1.66919434e+00 -5.37580311e-01 4.14469987e-01 2.81987846e-01
-9.37742114e-01 -4.13847417e-01 5.90606272e-01 -5.23913920e-01
5.70398748e-01 4.00193423e-01 9.82000291e-01 1.49185419e+00
1.14820123e-01 9.35854077e-01 1.54331279e+00 -6.47871733e-01
5.14138281e-01 4.48578596e-01 6.92240119e-01 3.98349702e-01
6.83300257e-01 7.39576280e-01 -3.58776718e-01 -3.29141647e-01
6.31006300e-01 1.25882447e-01 -1.17654592e-01 -4.27452862e-01
-9.09661353e-01 1.03094614e+00 2.22094968e-01 1.20469168e-01
-5.82106709e-01 1.22787856e-01 5.56118228e-02 3.62150013e-01
1.32636074e-02 8.88423681e-01 -5.54881752e-01 -1.58740446e-01
-7.98937917e-01 7.59938121e-01 9.34044302e-01 5.75795174e-01
3.71811420e-01 -2.03130797e-01 -2.67316341e-01 7.38976717e-01
2.86911279e-01 3.91309351e-01 7.62335718e-01 -1.17165363e+00
4.84655872e-02 3.10061932e-01 8.35817635e-01 -8.23836565e-01
-6.32069290e-01 -5.27168930e-01 -3.22727799e-01 1.13216847e-01
8.22871745e-01 -6.11776561e-02 -5.80887794e-01 1.79890418e+00
6.81080073e-02 -2.97608793e-01 -2.76588053e-01 8.06232333e-01
1.56313609e-02 2.99064875e-01 1.66565046e-01 -4.78158236e-01
1.41116416e+00 -5.21584332e-01 -6.45244539e-01 -6.67060256e-01
7.48930514e-01 -5.22027731e-01 1.58860517e+00 5.78390956e-01
-1.09053504e+00 -3.77978265e-01 -8.31852615e-01 3.22207987e-01
-2.53867179e-01 2.73508765e-02 1.09910703e+00 1.24933481e+00
-7.38899350e-01 7.78067589e-01 -9.05484140e-01 -3.21857363e-01
2.10955918e-01 2.54154772e-01 2.90772587e-01 1.16295025e-01
-1.02640021e+00 1.11502707e+00 4.14363623e-01 2.14897543e-01
-7.45666683e-01 -5.93038380e-01 -5.32621384e-01 1.91251710e-01
6.33606851e-01 -7.14262903e-01 1.84999907e+00 -1.10855722e+00
-1.48705852e+00 6.35323405e-01 -1.94994062e-01 -6.99718535e-01
5.37733197e-01 -2.43814111e-01 -5.57451360e-02 -2.61520535e-01
-1.93160743e-01 2.86672890e-01 3.91698271e-01 -1.04710793e+00
-3.84122729e-01 -4.36565220e-01 1.10304408e-01 3.18184346e-02
-2.09376037e-01 2.01843739e-01 2.58515123e-02 -5.13709903e-01
-1.98724315e-01 -1.03605151e+00 -4.74070489e-01 -2.57904500e-01
-3.49549413e-01 -1.72825903e-02 -1.93142388e-02 -3.00543606e-01
1.46947193e+00 -1.77948022e+00 -2.01431707e-01 2.87966341e-01
9.38929543e-02 6.81262761e-02 -5.42896688e-02 3.36200804e-01
-8.74643102e-02 4.01908249e-01 1.33237407e-01 -3.50852787e-01
5.25184393e-01 1.66982397e-01 -1.72173560e-01 4.96532410e-01
-3.60754728e-01 8.43921065e-01 -6.00149512e-01 -2.57136077e-01
2.84324646e-01 -2.96133105e-02 -9.66778219e-01 2.29055822e-01
-2.48911962e-01 2.62367666e-01 -6.19861245e-01 3.24582070e-01
4.31908399e-01 -3.69796664e-01 3.34267974e-01 3.42000365e-01
-2.17475757e-01 4.40482914e-01 -1.27240980e+00 8.03512037e-01
-4.57815319e-01 3.65487069e-01 -4.10448201e-02 -1.28691173e+00
5.40997624e-01 -2.00546592e-01 1.65930241e-02 -7.80595720e-01
4.87503111e-01 7.55338296e-02 3.79619867e-01 -6.10603452e-01
1.41351193e-01 -2.79216528e-01 -7.98233002e-02 7.46412814e-01
-2.37744555e-01 5.03474325e-02 2.06550807e-01 -2.86283582e-01
6.76268637e-01 -1.34076729e-01 7.15137243e-01 -6.00355387e-01
1.75824270e-01 -1.25024006e-01 3.79446864e-01 1.52164817e+00
-2.86596119e-01 -1.33083969e-01 6.62533700e-01 -6.57056153e-01
-1.01178348e+00 -7.59574294e-01 -1.79288521e-01 1.32486010e+00
-7.54541680e-02 -8.62155333e-02 -7.38239586e-01 -1.21660456e-01
1.43230677e-01 1.41783917e+00 -9.86819744e-01 -1.86329275e-01
-1.73037261e-01 -1.06990182e+00 1.07923083e-01 3.15827757e-01
1.73346594e-01 -1.05038071e+00 -1.03081763e+00 -1.24449559e-01
-5.29544428e-02 -7.03599870e-01 -2.74303854e-01 1.03154331e-01
-9.05027688e-01 -1.11903989e+00 -3.88440102e-01 -2.13875398e-01
4.02052253e-01 3.36366445e-01 1.15168869e+00 2.65194237e-01
-3.34602743e-01 4.48375195e-01 -2.33674720e-01 -7.14876831e-01
-4.60559070e-01 -2.42338642e-01 3.16321045e-01 -2.07861766e-01
7.15693831e-01 -4.32470888e-01 -7.70975053e-01 5.02943099e-01
-5.83858430e-01 1.17996559e-01 7.45302737e-01 9.43192482e-01
2.24170331e-02 -4.39555794e-02 4.15616393e-01 -8.61521482e-01
9.99901533e-01 -5.01774669e-01 -1.17738438e+00 4.06017810e-01
-1.05280507e+00 3.44751656e-01 4.65932816e-01 -6.80915833e-01
-1.08870745e+00 -2.52026170e-01 2.22098321e-01 -1.79324657e-01
-3.62947643e-01 6.92467868e-01 2.68527687e-01 1.79674804e-01
8.60517919e-01 1.00519061e-01 5.51780611e-02 -5.64592838e-01
3.71031791e-01 6.37025535e-01 2.23663878e-02 -1.03074288e+00
1.80772394e-01 8.52613896e-02 -1.60949022e-01 -5.97479939e-01
-9.99876201e-01 -4.92515452e-02 -3.45623225e-01 -1.13979124e-01
4.69331563e-01 -3.97828966e-01 -1.49210167e+00 4.79795560e-02
-6.91130996e-01 -5.72789907e-01 -3.47187072e-02 7.98663855e-01
-9.99865055e-01 2.41983701e-02 -3.07908177e-01 -1.54338384e+00
2.13341117e-01 -1.25433457e+00 6.29316688e-01 3.89044434e-01
-5.01370370e-01 -1.04452300e+00 9.14992113e-03 4.55560923e-01
3.14190775e-01 -2.70435631e-01 9.79270756e-01 -9.46152091e-01
-4.81864452e-01 -1.36049643e-01 2.02368096e-01 2.89324927e-03
-1.44398808e-01 -4.44110595e-02 -7.29694426e-01 -3.92103493e-01
2.37541452e-01 -3.64451289e-01 7.09130347e-01 9.70032275e-01
1.42726517e+00 -6.72936976e-01 -4.06185478e-01 4.49590743e-01
1.03171015e+00 3.56395125e-01 2.86904514e-01 6.14383638e-01
-7.39187747e-02 9.23999846e-01 5.97945511e-01 5.79087257e-01
3.18780184e-01 8.27139556e-01 1.62735984e-01 3.10690939e-01
6.79273009e-01 -3.62054676e-01 3.81148189e-01 -2.04506833e-02
-3.85221429e-02 -1.98128209e-01 -7.62528300e-01 4.03294474e-01
-1.82009268e+00 -1.13626647e+00 2.67161690e-02 2.80096054e+00
7.43645847e-01 2.87974507e-01 7.16764748e-01 -4.01388705e-02
6.13285244e-01 -3.39734823e-01 -4.83889848e-01 -5.83927333e-01
2.72656679e-01 -7.54476935e-02 5.37378073e-01 5.80229223e-01
-8.15968871e-01 6.92913830e-01 8.06379414e+00 5.73817670e-01
-6.69984400e-01 -1.78403452e-01 8.63889992e-01 -4.27154630e-01
-1.57498017e-01 7.03094080e-02 -7.90968180e-01 5.07897198e-01
1.36850858e+00 -4.67366636e-01 8.73005509e-01 8.48239958e-01
6.96444213e-01 -4.03949380e-01 -1.49079788e+00 8.26221466e-01
-3.68432701e-01 -1.04761648e+00 -2.29685485e-01 2.25859836e-01
4.43964452e-01 -4.82277393e-01 3.71529013e-01 3.95422757e-01
8.55832398e-01 -1.15376079e+00 9.39394534e-01 4.66275901e-01
-8.41143280e-02 -4.63054508e-01 6.92991614e-01 8.08779716e-01
-2.50223190e-01 -7.64651358e-01 -4.41840500e-01 -6.89238191e-01
-1.10033333e-01 2.95143753e-01 -8.04442883e-01 5.98668754e-02
7.35055625e-01 1.61085278e-01 -4.97827977e-01 1.28128910e+00
-1.20933428e-01 6.97158158e-01 -3.88710767e-01 -5.33661604e-01
2.61808187e-01 -1.91471562e-01 2.83376157e-01 9.97254252e-01
3.12008530e-01 1.26160428e-01 2.51697302e-02 1.28082287e+00
4.16635275e-01 1.08983859e-01 -5.15644073e-01 -1.96759701e-01
6.88077569e-01 8.00765812e-01 -7.93803394e-01 -1.88093558e-01
-4.04579431e-01 3.98401648e-01 4.06915307e-01 4.17613953e-01
-6.93909883e-01 2.27229625e-01 7.46304512e-01 1.19927730e-02
2.37381697e-01 -2.13273823e-01 -5.53798318e-01 -1.07484972e+00
-2.16483533e-01 -1.14628410e+00 5.19608676e-01 -6.13111854e-01
-1.41703033e+00 -3.31537277e-02 6.52957439e-01 -6.69164002e-01
-4.60615456e-01 -9.51011658e-01 -3.72151613e-01 7.94411361e-01
-9.79643166e-01 -6.14418209e-01 -7.03261932e-03 2.47716174e-01
4.55248564e-01 1.49580002e-01 5.06071270e-01 -1.58292428e-01
-8.09328198e-01 5.73642492e-01 2.26816297e-01 -3.31295550e-01
4.26047295e-01 -1.17688107e+00 2.36204326e-01 4.92393732e-01
9.28205326e-02 1.20325637e+00 1.29414034e+00 -4.44020629e-01
-1.19642997e+00 -2.68273920e-01 3.57835442e-01 -5.81823170e-01
8.18608105e-01 -4.02023047e-01 -5.28083265e-01 1.00381458e+00
-5.86900748e-02 -5.56305528e-01 9.07807291e-01 7.53251791e-01
-1.56528935e-01 1.07803397e-01 -8.83445740e-01 9.78338301e-01
9.87532496e-01 3.77605297e-03 -8.36269975e-01 3.73176306e-01
3.92371148e-01 -6.58230335e-02 -3.83903742e-01 -3.29854488e-02
9.61505175e-01 -1.44017398e+00 1.02245450e+00 -1.13487494e+00
1.42313600e-01 1.58285663e-01 -1.11483119e-01 -1.46033001e+00
-4.29581106e-01 -8.14633071e-01 1.92069188e-01 7.41358280e-01
4.22836959e-01 -7.70649016e-01 6.16591454e-01 1.04143190e+00
2.42259637e-01 -6.23977005e-01 -4.87646312e-01 -8.50302041e-01
2.33147308e-01 -7.04851985e-01 5.66953301e-01 7.55997658e-01
1.49410084e-01 -6.56928075e-03 -4.69140947e-01 -2.17236485e-02
8.18409264e-01 3.07033032e-01 8.06811988e-01 -1.36689234e+00
-6.83855832e-01 -9.11508560e-01 1.31892338e-02 -1.26148665e+00
1.62224874e-01 -3.65328908e-01 -1.09854206e-01 -9.21995342e-01
5.00175714e-01 -3.69502008e-01 -2.10150480e-01 2.51025140e-01
-2.96432704e-01 -2.33413652e-01 2.04599693e-01 1.35877788e-01
-3.79270703e-01 4.24580663e-01 9.42342043e-01 2.33383223e-01
-4.72260982e-01 3.84321511e-01 -1.32904530e+00 9.48828816e-01
7.33652234e-01 -4.42514420e-01 -5.37854373e-01 -5.85668208e-03
3.61147702e-01 1.61179364e-01 6.91773355e-01 -5.19642353e-01
1.18710436e-01 -8.09365094e-01 5.19740939e-01 -1.75481886e-01
1.79344520e-01 -7.11550713e-01 1.74701475e-02 5.68625212e-01
-7.72868097e-01 6.05210587e-02 3.06518257e-01 5.58673143e-01
3.82718623e-01 -5.15676498e-01 7.13428199e-01 -3.27389896e-01
-3.74743402e-01 -1.24227151e-01 -6.49868488e-01 -2.36512423e-01
1.09777117e+00 -1.35789424e-01 -3.62750620e-01 -8.72558415e-01
-9.87353981e-01 4.42673653e-01 4.64916229e-01 2.20100328e-01
1.72532186e-01 -9.31653619e-01 -4.21605855e-01 1.79827616e-01
-6.65289396e-03 -8.84827793e-01 3.18521082e-01 9.98512328e-01
-2.96038896e-01 7.61789441e-01 -2.84981221e-01 -2.00481474e-01
-1.10352981e+00 1.06185961e+00 4.98542666e-01 -3.22043091e-01
-1.32112846e-01 5.32163620e-01 4.22390193e-01 -1.87735319e-01
9.80742350e-02 -5.14112532e-01 -1.06732212e-01 -1.54767156e-01
7.17927158e-01 4.44673300e-01 -2.05438882e-01 -2.50137039e-02
-2.53535986e-01 9.13440064e-02 -2.11511806e-01 -2.39076808e-01
1.27863884e+00 -3.57025921e-01 1.98060304e-01 6.10921621e-01
5.69805026e-01 -2.08971024e-01 -1.20269048e+00 -4.54306193e-02
1.49259016e-01 -7.32110202e-01 1.50577530e-01 -9.37974513e-01
-3.97933900e-01 6.90181732e-01 4.63985205e-01 6.50057852e-01
1.02376294e+00 -3.94790955e-02 -1.58009142e-01 7.43592501e-01
3.59119892e-01 -9.63312805e-01 -6.27933666e-02 2.05727257e-02
7.61824727e-01 -1.17093039e+00 2.12141544e-01 -1.04607446e-02
-6.13077939e-01 8.59146893e-01 5.32191455e-01 -8.56711194e-02
5.17304838e-01 2.00300086e-02 -2.21635580e-01 -1.30497783e-01
-1.07251561e+00 -1.65262297e-01 1.48072705e-01 4.69043821e-01
3.84734839e-01 2.27226734e-01 -6.33514583e-01 9.58728194e-01
-6.69557571e-01 1.56068504e-01 6.12322628e-01 7.26482153e-01
-5.50486684e-01 -1.04485452e+00 -7.46652603e-01 5.98044157e-01
-5.14817953e-01 4.05192152e-02 -3.69273931e-01 1.10452795e+00
-3.79240423e-01 1.10022867e+00 1.27500296e-01 -1.64385810e-01
2.45673120e-01 1.06824726e-01 6.50651693e-01 -2.62286872e-01
-2.82896012e-01 3.71810561e-03 9.34742838e-02 -7.05462575e-01
-2.24395394e-01 -9.01328027e-01 -4.73021805e-01 -6.97886288e-01
-2.82850385e-01 2.43236527e-01 4.35829133e-01 1.06100857e+00
1.96009338e-01 2.29452118e-01 3.63648236e-01 -9.04461741e-01
-1.15888584e+00 -1.03675473e+00 -6.24471784e-01 2.81780183e-01
3.61331403e-01 -1.15235412e+00 -6.21292174e-01 -1.92792222e-01] | [4.648215293884277, 3.071427345275879] |
c7380bb9-5f43-458a-a3bd-70063deb3fa2 | diabetic-retinopathy-detection-using-ensemble | 2106.12545 | null | https://arxiv.org/abs/2106.12545v1 | https://arxiv.org/pdf/2106.12545v1.pdf | Diabetic Retinopathy Detection using Ensemble Machine Learning | Diabetic Retinopathy (DR) is among the worlds leading vision loss causes in diabetic patients. DR is a microvascular disease that affects the eye retina, which causes vessel blockage and therefore cuts the main source of nutrition for the retina tissues. Treatment for this visual disorder is most effective when it is detected in its earliest stages, as severe DR can result in irreversible blindness. Nonetheless, DR identification requires the expertise of Ophthalmologists which is often expensive and time-consuming. Therefore, automatic detection systems were introduced aiming to facilitate the identification process, making it available globally in a time and cost-efficient manner. However, due to the limited reliable datasets and medical records for this particular eye disease, the obtained predictions accuracies were relatively unsatisfying for eye specialists to rely on them as diagnostic systems. Thus, we explored an ensemble-based learning strategy, merging a substantial selection of well-known classification algorithms in one sophisticated diagnostic model. The proposed framework achieved the highest accuracy rates among all other common classification algorithms in the area. 4 subdatasets were generated to contain the top 5 and top 10 features of the Messidor dataset, selected by InfoGainEval. and WrapperSubsetEval., accuracies of 70.7% and 75.1% were achieved on the InfoGainEval. top 5 and original dataset respectively. The results imply the impressive performance of the subdataset, which significantly conduces to a less complex classification process | ['Mohammad Alauthman', 'Mouhammd Alkasassbeh', 'Israa Odeh'] | 2021-06-22 | null | null | null | null | ['diabetic-retinopathy-detection'] | ['medical'] | [-8.68250907e-04 -1.48237962e-02 -3.21032666e-02 -1.58872098e-01
-3.90663505e-01 -2.20209867e-01 3.64200354e-01 2.02006236e-01
-3.23179960e-01 1.02991021e+00 5.38563579e-02 -2.49135792e-01
-4.81352806e-01 -7.25146770e-01 -4.71292548e-02 -1.02692771e+00
1.20932683e-01 2.65062660e-01 -1.66279897e-02 1.83938235e-01
5.08106947e-01 7.29177058e-01 -2.25223088e+00 1.68900728e-01
1.73013365e+00 1.18085945e+00 2.22864807e-01 3.36373806e-01
-2.09765598e-01 8.10140073e-01 -2.53582269e-01 -4.62675840e-01
2.65952766e-01 -3.73573840e-01 -2.80728906e-01 7.31274784e-02
3.03965092e-01 -4.48216498e-01 -3.11588570e-02 8.88306201e-01
7.74379969e-01 -3.00402790e-01 9.17315245e-01 -5.88206172e-01
-4.64112967e-01 3.72911990e-02 -6.65544271e-01 1.93246469e-01
2.06382573e-01 1.80975989e-01 5.19144535e-01 -8.28814268e-01
4.12402749e-01 9.06835794e-01 2.97511220e-01 3.21346462e-01
-1.01258087e+00 -4.49091911e-01 -3.90725732e-01 4.75938201e-01
-1.18317807e+00 -4.83530939e-01 3.32188934e-01 -8.64804029e-01
6.08880281e-01 3.39994818e-01 8.45844328e-01 4.56183553e-01
3.36455673e-01 4.29122001e-01 1.61094618e+00 -3.46663147e-01
1.49246886e-01 3.23724985e-01 3.73020500e-01 5.55427432e-01
6.00151360e-01 9.75228846e-02 -7.88305104e-02 6.06026920e-03
4.83124584e-01 -7.47539196e-03 -3.44501227e-01 -1.07268065e-01
-7.71849751e-01 5.30088902e-01 3.08806986e-01 2.72896048e-02
-6.80204630e-01 -6.61410272e-01 2.42310822e-01 2.28561461e-01
2.40208521e-01 2.10312545e-01 -2.35782146e-01 1.23285362e-02
-5.12469709e-01 -6.53903335e-02 3.55640203e-01 3.59388411e-01
4.42728400e-01 -1.16651386e-01 -2.21947357e-01 9.84354317e-01
2.30204210e-01 5.26005089e-01 2.37999067e-01 -4.88003731e-01
1.00029245e-01 1.25372624e+00 2.12259009e-01 -1.01716948e+00
-3.53289366e-01 -7.80749738e-01 -9.70965505e-01 5.15963316e-01
3.75142664e-01 -2.46102542e-01 -1.17745960e+00 9.65921640e-01
3.37942660e-01 -8.33201781e-02 1.91773549e-01 9.75204229e-01
9.06443715e-01 3.87373596e-01 1.78679183e-01 -4.86969113e-01
1.44008064e+00 -4.79591578e-01 -5.66370964e-01 1.30728170e-01
3.69239181e-01 -1.00426376e+00 5.82870185e-01 7.40314603e-01
-7.96252251e-01 -4.71824229e-01 -8.56361628e-01 4.41215783e-02
-1.60726041e-01 7.66096711e-01 7.94807196e-01 4.52089995e-01
-9.29123044e-01 1.78201124e-01 -4.46383476e-01 -6.75726414e-01
7.00262666e-01 2.84652174e-01 -4.74001557e-01 -4.26523030e-01
-5.92440307e-01 1.16856575e+00 6.92896917e-02 4.47926462e-01
-4.34383392e-01 -4.16133404e-01 -2.37296954e-01 -4.49291110e-01
9.63323861e-02 -8.47452700e-01 5.82127869e-01 -8.04029226e-01
-1.13633919e+00 9.81259346e-01 -3.25142413e-01 -5.04947543e-01
6.93866313e-01 -2.39926681e-01 -6.00161314e-01 2.10532889e-01
-1.86374515e-01 2.10424792e-02 7.47868896e-01 -1.03895140e+00
-1.02690363e+00 -7.38265038e-01 -2.40239084e-01 8.06420818e-02
-1.32328406e-01 6.92458302e-02 -6.50403351e-02 -1.58524543e-01
3.31153631e-01 -5.69252968e-01 -1.23382010e-01 -1.07822366e-01
-6.25897408e-01 -3.15687925e-01 4.40561980e-01 -8.89936209e-01
1.34033716e+00 -1.94216371e+00 3.89456600e-02 1.73164696e-01
4.41825032e-01 7.10088670e-01 3.82012457e-01 3.20892215e-01
-1.06489509e-01 3.25783044e-02 5.54094724e-02 2.16890842e-01
-6.01982415e-01 -1.10668555e-01 1.85674489e-01 4.67702329e-01
1.95245445e-01 1.42618001e-01 -4.01345998e-01 -4.35118943e-01
3.35600644e-01 5.74593961e-01 -2.87930399e-01 3.41707706e-01
5.30299991e-02 4.97924805e-01 -7.05209613e-01 9.94936526e-01
6.28899813e-01 -7.66268522e-02 -6.33573011e-02 -4.02298838e-01
-6.03555381e-01 -9.20535773e-02 -1.07719660e+00 9.53329325e-01
-1.08277775e-01 3.47201943e-01 -2.26395816e-01 -9.86478508e-01
1.27370334e+00 2.30804026e-01 4.18382853e-01 -5.91832697e-01
3.07855964e-01 4.88591999e-01 1.73126996e-01 -1.08600187e+00
-8.12073946e-02 2.01009825e-01 5.21916091e-01 -2.33852834e-01
-6.02468252e-01 5.80817938e-01 2.41024807e-01 -1.34848505e-01
8.17686379e-01 1.01981990e-01 6.86999440e-01 1.04314387e-02
9.06060100e-01 2.39125431e-01 5.93416929e-01 2.52339691e-01
-1.04789458e-01 4.11136478e-01 4.56302524e-01 -6.78739250e-01
-8.15455854e-01 -7.87703395e-01 -6.29529893e-01 6.68528378e-02
-6.03370219e-02 2.25301266e-01 -4.00366515e-01 -3.22635055e-01
2.05630779e-01 4.31720078e-01 -3.90382916e-01 5.82489818e-02
-7.76349567e-03 -9.11981821e-01 2.14139476e-01 4.07767482e-03
7.87902594e-01 -5.59918344e-01 -5.41963041e-01 1.11565605e-01
3.05675864e-01 -5.97901464e-01 3.19742948e-01 -2.46370494e-01
-1.06742203e+00 -1.55554950e+00 -8.95789564e-01 -6.32984340e-01
7.57635415e-01 2.40399301e-01 6.46524847e-01 1.25689745e-01
-7.94652283e-01 -2.30999261e-01 -2.72156298e-01 -4.19747472e-01
-3.33270848e-01 -2.79749721e-01 -9.18704122e-02 4.26950723e-01
7.44092822e-01 -5.64667940e-01 -1.02831912e+00 1.03170358e-01
-1.56829610e-01 -2.35548839e-02 1.15770113e+00 6.15018547e-01
5.69720507e-01 7.44308159e-02 9.03719902e-01 -8.27729464e-01
2.18872607e-01 -4.27006334e-01 -8.68203998e-01 1.77738175e-01
-8.26273561e-01 -2.20947668e-01 4.23592687e-01 1.25412315e-01
-1.03735209e+00 1.37493670e-01 1.17948428e-01 -1.54997306e-02
-3.66747558e-01 5.12890577e-01 -1.34754419e-01 -6.72636330e-02
5.95117092e-01 2.75019974e-01 3.17417741e-01 -7.24037766e-01
-2.74293534e-02 1.18075645e+00 3.73819083e-01 -1.76045988e-02
4.75137502e-01 1.68655127e-01 1.75151095e-01 -9.86038864e-01
-5.25502920e-01 -5.60362160e-01 -2.35203132e-01 -3.46863389e-01
1.03278863e+00 -1.04660308e+00 -8.03662956e-01 7.98551440e-01
-8.95501971e-01 6.47818208e-01 4.16507214e-01 8.36250424e-01
-1.57797396e-01 2.79101551e-01 -7.58165196e-02 -1.09813535e+00
-4.47174340e-01 -1.02234662e+00 3.71950716e-01 5.71233213e-01
1.53969914e-01 -4.66811538e-01 -1.75170347e-01 7.27782667e-01
3.55696112e-01 5.37139177e-01 1.35816932e+00 -4.74339604e-01
-8.38278413e-01 -3.38288903e-01 -6.40494227e-01 6.25315905e-01
4.12741840e-01 5.15975952e-01 -1.07771158e+00 3.35221216e-02
-3.95689815e-01 -1.20603189e-01 9.34309959e-01 5.59879005e-01
9.75968778e-01 -8.82424489e-02 -4.07449961e-01 3.97453904e-01
1.87751329e+00 6.83178067e-01 9.79462385e-01 4.45270479e-01
3.12257677e-01 5.63204765e-01 6.88405454e-01 6.11185968e-01
3.39373112e-01 4.25649583e-01 6.71144724e-01 -3.11100006e-01
-2.78452188e-01 1.12984911e-01 -7.89934248e-02 5.31866789e-01
-7.77892590e-01 -1.84960857e-01 -8.50229979e-01 6.36357069e-01
-1.44725776e+00 -9.37061965e-01 -6.38724566e-01 2.54045057e+00
7.62193739e-01 -7.00686350e-02 1.14438482e-01 1.74209341e-01
8.35496247e-01 -4.18721616e-01 -5.67871094e-01 -2.07846880e-01
-2.07167208e-01 1.95572928e-01 4.26385552e-01 -1.99893117e-02
-9.29324567e-01 3.81673217e-01 5.32244253e+00 3.22447568e-01
-1.10222197e+00 -4.69098836e-01 5.42464018e-01 -4.10775281e-02
6.10067509e-02 -1.07680053e-01 -8.91014755e-01 6.65766418e-01
6.38706923e-01 -5.15579060e-02 1.19531736e-01 7.09205985e-01
8.81628931e-01 -5.04444301e-01 -7.61826217e-01 1.00653660e+00
-2.54875153e-01 -1.08212805e+00 2.32105523e-01 1.71726406e-01
6.64248705e-01 -1.57846943e-01 2.44855881e-04 -1.42511532e-01
-7.26420060e-02 -1.06251085e+00 -1.50551081e-01 1.13888705e+00
8.09681892e-01 -9.10874486e-01 1.20114982e+00 1.36635885e-01
-6.21655762e-01 -3.21987361e-01 -4.26538527e-01 -6.18291972e-03
4.42831049e-04 1.08078003e+00 -9.05124545e-01 6.55464232e-01
6.06416464e-01 8.59986722e-01 -6.03215575e-01 1.83128059e+00
6.31997287e-02 4.18252110e-01 3.49783897e-02 1.23941340e-01
-2.84007400e-01 -5.46806037e-01 6.61121786e-01 4.73945618e-01
7.70457506e-01 2.39100799e-01 -1.45479366e-01 4.66443270e-01
2.01998740e-01 5.91106653e-01 -4.95933861e-01 4.44103926e-02
4.42646652e-01 1.13273752e+00 -1.30168095e-01 -1.87059585e-02
-4.90200579e-01 3.91973168e-01 3.28983068e-02 1.71303824e-01
-5.71707964e-01 -3.43431622e-01 6.80618823e-01 4.34050530e-01
1.58984214e-01 3.78448427e-01 -4.00134534e-01 -8.71446490e-01
-3.24700326e-02 -9.63625371e-01 3.64392489e-01 -6.01998925e-01
-1.25264776e+00 6.50426328e-01 -4.46047246e-01 -1.46589077e+00
8.35178718e-02 -5.76731980e-01 -3.20551932e-01 1.17084610e+00
-1.77147150e+00 -9.91060674e-01 -6.85021639e-01 4.72757667e-01
3.03720355e-01 -5.16750395e-01 7.87217319e-01 4.43065882e-01
-1.10637581e+00 2.81262130e-01 3.38158190e-01 -2.87023962e-01
8.75755072e-01 -1.03689194e+00 -6.52714491e-01 7.31438994e-01
-7.06308484e-01 5.82511902e-01 6.05331957e-01 -6.82978570e-01
-1.30013311e+00 -8.64484191e-01 9.11305189e-01 6.48736767e-03
2.53309727e-01 5.65121949e-01 -7.52171218e-01 8.83697942e-02
7.63470232e-02 -2.10728034e-01 6.06784523e-01 -9.33179930e-02
9.06586573e-02 -5.48481405e-01 -1.24008322e+00 4.57253247e-01
6.98006213e-01 -1.43294549e-02 -5.25178909e-01 2.21354052e-01
-4.10308689e-02 -9.64909792e-02 -1.06564689e+00 4.82870102e-01
6.88307703e-01 -1.45439887e+00 6.92792892e-01 -4.92155224e-01
6.34673357e-01 -5.41183710e-01 8.79185926e-03 -8.95812154e-01
2.12795641e-02 -1.93736374e-01 1.37503326e-01 1.32354593e+00
5.37341774e-01 -9.83342052e-01 4.94453013e-01 5.97567677e-01
-7.91736543e-02 -8.39086890e-01 -5.69827318e-01 -3.16544890e-01
-4.27219182e-01 2.25023925e-01 2.20622003e-01 6.61097765e-01
-4.91841912e-01 1.38180047e-01 -2.31002390e-01 3.33404422e-01
8.55159879e-01 2.00724136e-02 7.50582695e-01 -1.77855599e+00
-8.38594045e-03 -2.27457687e-01 -6.92946970e-01 -4.60666776e-01
-5.54990947e-01 -6.52498782e-01 -5.38030922e-01 -1.94281936e+00
1.96063548e-01 -5.47200918e-01 -3.39231074e-01 3.82962644e-01
-1.03449076e-01 1.73599467e-01 -3.33311498e-01 3.73021454e-01
1.68588743e-01 1.30168587e-01 1.42212629e+00 7.17125386e-02
-4.37589169e-01 4.82282430e-01 -8.04026127e-01 7.06458211e-01
8.00735354e-01 -1.09152347e-01 -6.52173221e-01 -2.05440864e-01
5.36170416e-02 5.01565151e-02 3.24946195e-01 -9.32956934e-01
6.40436560e-02 -1.56348288e-01 4.69751298e-01 -7.62401283e-01
1.59760654e-01 -7.58494079e-01 3.97985965e-01 4.03788686e-01
7.30253160e-02 -4.10474807e-01 2.52847131e-02 5.33512950e-01
-2.83922076e-01 6.79734498e-02 1.08400404e+00 1.31888622e-02
-8.31917465e-01 8.35347995e-02 -3.32157731e-01 -3.99611354e-01
1.43107808e+00 -6.22362316e-01 -5.42830527e-01 -2.19148435e-02
-8.22241306e-01 7.90070295e-02 3.15318078e-01 1.61302816e-02
6.13188326e-01 -8.08089674e-01 -9.81570959e-01 1.36394233e-01
3.07987899e-01 -1.58545393e-02 4.96067882e-01 1.27203381e+00
-7.49415517e-01 3.90127033e-01 -4.04868275e-01 -6.68739974e-01
-1.61255741e+00 2.40684956e-01 3.29387605e-01 1.09467231e-01
-6.87323451e-01 6.00999534e-01 -3.24054897e-01 3.52983534e-01
2.35293761e-01 -2.65643477e-01 -1.08002377e+00 3.70294213e-01
7.62437940e-01 8.17223847e-01 1.85319290e-01 -4.20228750e-01
-3.52441430e-01 9.18132484e-01 -3.46028000e-01 7.37028778e-01
1.26982439e+00 -3.12004149e-01 -5.14810920e-01 -5.66843490e-04
6.71077728e-01 1.68292627e-01 -7.88106441e-01 1.81698259e-02
-4.62367162e-02 -6.00808680e-01 3.52808475e-01 -1.22718966e+00
-9.76719558e-01 6.00378335e-01 9.03582335e-01 1.16605334e-01
1.49200416e+00 -3.26996833e-01 4.66623187e-01 8.62580165e-02
5.40517390e-01 -7.20277846e-01 -6.37131691e-01 -2.52056807e-01
7.00259149e-01 -1.33050406e+00 2.85654426e-01 -6.74930274e-01
-5.36777735e-01 1.19266248e+00 4.29987699e-01 1.57061722e-02
4.07461792e-01 -2.93620139e-01 3.09913009e-01 -7.34833330e-02
-7.49264061e-01 -3.70883733e-01 3.61686975e-01 6.60840213e-01
5.02157211e-01 -3.98038980e-03 -1.09825027e+00 4.84211117e-01
1.47855818e-01 4.48918283e-01 5.44153988e-01 4.98848855e-01
-7.43605256e-01 -1.04119062e+00 -2.71300286e-01 9.74430740e-01
-5.75912833e-01 4.92978096e-02 -2.06938803e-01 9.13519740e-01
4.03057784e-01 1.06828988e+00 -1.82304144e-01 -2.77332395e-01
3.88357222e-01 8.45279952e-04 1.89582422e-01 -3.39591652e-01
-2.14261949e-01 -2.92761791e-02 5.22547305e-01 -4.27217960e-01
-6.65900469e-01 -6.67226672e-01 -9.97958899e-01 -2.71168292e-01
-1.92040354e-01 -5.91418892e-02 6.39362931e-01 9.15397227e-01
4.41170037e-01 3.03046077e-01 9.33632553e-01 2.26030219e-02
-5.81959963e-01 -9.19222176e-01 -6.31522954e-01 -5.95318526e-02
4.15473968e-01 -8.56162250e-01 -4.35437500e-01 -3.32597941e-02] | [15.839580535888672, -3.995589017868042] |
3bd96c4d-a842-4850-bd08-2da6390fcc06 | t-sciq-teaching-multimodal-chain-of-thought | 2305.03453 | null | https://arxiv.org/abs/2305.03453v2 | https://arxiv.org/pdf/2305.03453v2.pdf | T-SciQ: Teaching Multimodal Chain-of-Thought Reasoning via Large Language Model Signals for Science Question Answering | Large Language Models (LLMs) have recently demonstrated exceptional performance in various Natural Language Processing (NLP) tasks. They have also shown the ability to perform chain-of-thought (CoT) reasoning to solve complex problems. Recent studies have explored CoT reasoning in complex multimodal scenarios, such as the science question answering task, by fine-tuning multimodal models with high-quality human-annotated CoT rationales. However, collecting high-quality COT rationales is usually time-consuming and costly. Besides, the annotated rationales are hardly accurate due to the redundant information involved or the essential information missed. To address these issues, we propose a novel method termed \emph{T-SciQ} that aims at teaching science question answering with LLM signals. The T-SciQ approach generates high-quality CoT rationales as teaching signals and is advanced to train much smaller models to perform CoT reasoning in complex modalities. Additionally, we introduce a novel data mixing strategy to produce more effective teaching data samples for simple and complex science question answer problems. Extensive experimental results show that our T-SciQ method achieves a new state-of-the-art performance on the ScienceQA benchmark, with an accuracy of 96.18%. Moreover, our approach outperforms the most powerful fine-tuned baseline by 4.5%. | ['Heng Tao Shen', 'Hui Liu', 'Ning Liu', 'Xing Xu', 'Jiabang He', 'Yi Hu', 'Lei Wang'] | 2023-05-05 | null | null | null | null | ['science-question-answering'] | ['miscellaneous'] | [ 2.40183666e-01 1.47268698e-01 1.37295321e-01 -5.25040925e-01
-1.53187752e+00 -6.91535473e-01 4.94052529e-01 4.99819905e-01
-4.00943965e-01 3.75631034e-01 7.91727677e-02 -4.34317857e-01
-2.31125906e-01 -5.81402659e-01 -1.08261526e+00 -4.67197299e-01
5.85996807e-01 7.02531755e-01 1.34135380e-01 -4.12590623e-01
1.91434681e-01 -4.85819019e-02 -1.50686431e+00 7.69251406e-01
1.48841321e+00 7.81454504e-01 2.51854807e-01 8.10680866e-01
-7.67659903e-01 1.15152812e+00 -8.73206735e-01 -9.16770577e-01
-3.77627909e-01 -4.15912420e-01 -9.71101820e-01 -4.00092185e-01
7.58813024e-01 -2.91284546e-02 8.01261812e-02 9.21577334e-01
5.38986921e-01 2.77899444e-01 5.03432870e-01 -1.21364689e+00
-6.41982377e-01 9.52471733e-01 -6.12049639e-01 5.52183948e-03
8.96238387e-01 5.17717190e-02 1.12845731e+00 -9.73121285e-01
2.84661263e-01 1.57632244e+00 2.14864612e-01 7.99147427e-01
-1.22821140e+00 -7.48575389e-01 2.43460804e-01 5.04266441e-01
-9.21124220e-01 -5.94490841e-02 8.52010548e-01 -2.42722720e-01
5.82278192e-01 5.13640642e-01 3.28400582e-01 1.27209806e+00
-2.45106351e-02 1.27403116e+00 1.09238720e+00 -5.03421307e-01
1.16291404e-01 3.33449543e-02 3.43404979e-01 8.86737704e-01
-1.97098792e-01 -6.33714497e-01 -9.23038006e-01 -3.00301779e-02
3.00140202e-01 -2.22626761e-01 -3.47677052e-01 4.22881097e-02
-1.69595242e+00 8.66291761e-01 4.00436461e-01 1.87534407e-01
-3.22430164e-01 1.92146897e-01 1.12880513e-01 5.13777316e-01
4.33935188e-02 7.39194810e-01 -6.07577026e-01 -3.39009553e-01
-5.84069431e-01 2.14701220e-01 7.83610880e-01 1.03007138e+00
3.03595871e-01 -2.30261892e-01 -5.34715950e-01 1.06040347e+00
5.06222844e-01 7.51982629e-01 2.57010430e-01 -1.11639965e+00
9.70522344e-01 9.16494608e-01 -1.73683599e-01 -7.51083851e-01
-2.92445153e-01 -3.51219654e-01 -8.96089852e-01 -4.19481546e-01
7.02573240e-01 -4.42353263e-02 -7.29913652e-01 1.78779829e+00
4.29336727e-01 -2.51272246e-02 2.46100575e-01 9.45059597e-01
1.53632128e+00 9.09209847e-01 3.32603753e-01 1.24250978e-01
1.80386722e+00 -1.15069389e+00 -8.06673229e-01 -3.58938910e-02
4.76763666e-01 -9.23258126e-01 1.65055680e+00 6.82003975e-01
-1.22960949e+00 -5.86871147e-01 -5.96000552e-01 -3.22453827e-01
-1.45639271e-01 2.22091183e-01 5.60666680e-01 4.55452144e-01
-6.59756124e-01 3.24840322e-02 -3.22911173e-01 -5.41296788e-02
3.70881319e-01 1.21772289e-01 -1.55661225e-01 -5.66610873e-01
-1.18843353e+00 6.79572761e-01 1.21788979e-01 1.72561631e-01
-9.04514313e-01 -1.10402906e+00 -8.01045179e-01 1.38063788e-01
6.00424886e-01 -9.58131850e-01 1.51694763e+00 -4.89935219e-01
-1.84896886e+00 6.39821827e-01 -1.34567440e-01 1.26659662e-01
4.87498879e-01 -3.96836996e-01 -3.61011088e-01 5.40263176e-01
2.31667757e-02 9.66097951e-01 7.12306917e-01 -1.18578649e+00
-2.48348877e-01 -1.56593964e-01 2.74037153e-01 1.52979895e-01
-2.88401008e-01 -1.82919741e-01 -6.43490314e-01 -5.87908864e-01
1.80544272e-01 -6.56190932e-01 -4.83133346e-02 -2.32728645e-01
-2.97000110e-01 -7.91479051e-01 1.88272625e-01 -4.11247581e-01
1.02217841e+00 -1.82320261e+00 4.48633313e-01 8.13651830e-03
3.34173024e-01 1.77199483e-01 -6.29440427e-01 5.32372117e-01
2.36160740e-01 -2.22243425e-02 -1.95107609e-01 -1.79492667e-01
3.37594509e-01 3.13982338e-01 -6.35019243e-01 -3.07501823e-01
3.40296835e-01 1.19092786e+00 -1.15591240e+00 -7.43327141e-01
2.43709646e-02 3.07871848e-01 -5.44031084e-01 5.65517128e-01
-7.45914102e-01 5.90199471e-01 -6.07777357e-01 8.16626489e-01
4.65831071e-01 -5.34865499e-01 3.00879008e-03 -3.44594836e-01
2.37116203e-01 1.51563987e-01 -9.00978804e-01 2.06886888e+00
-5.47615051e-01 4.56923008e-01 -5.59628718e-02 -7.47533560e-01
8.13070655e-01 3.40395540e-01 1.02030843e-01 -9.46544409e-01
9.94835347e-02 2.46934086e-01 -5.20764254e-02 -1.07750416e+00
2.90960878e-01 -1.20849945e-01 -1.67147949e-01 3.55111599e-01
2.67493457e-01 -6.79216087e-01 4.13090676e-01 5.70705891e-01
1.12003088e+00 6.19408451e-02 -3.50124925e-01 6.53001294e-02
1.04709053e+00 3.34276780e-02 2.88579792e-01 8.08019400e-01
-1.14643678e-01 4.33638543e-01 4.54933643e-01 -3.27568576e-02
-3.81203622e-01 -1.03842211e+00 2.07874373e-01 1.43971980e+00
1.01703502e-01 -5.59277833e-01 -7.01255143e-01 -7.51965702e-01
-1.80377543e-01 1.00205815e+00 -3.23784113e-01 -5.96680008e-02
-4.94437873e-01 -6.38570368e-01 7.28874326e-01 3.84714514e-01
5.48158824e-01 -1.03072071e+00 -3.27931941e-01 1.10132977e-01
-7.95817971e-01 -1.35452557e+00 -3.05472344e-01 -1.55265659e-01
-7.19012976e-01 -1.19654059e+00 -7.36071944e-01 -7.40279317e-01
7.72169173e-01 2.85215795e-01 1.46105945e+00 2.06946611e-01
3.17041799e-02 7.99593091e-01 -4.86898780e-01 -4.94357020e-01
-4.27858472e-01 -8.36204886e-02 -3.95372987e-01 5.11206454e-04
4.43558425e-01 -1.63677588e-01 -4.92500663e-01 9.32438821e-02
-1.12112617e+00 2.39036009e-01 8.94881725e-01 8.10762942e-01
5.25699615e-01 -1.72062457e-01 8.15684438e-01 -7.73910403e-01
1.04168451e+00 -3.32629561e-01 -3.46933007e-01 8.18210006e-01
-1.63248375e-01 2.98739702e-01 7.48300016e-01 -6.54213607e-01
-1.41209769e+00 -2.89151251e-01 -3.86334717e-01 -2.71454960e-01
-1.94632545e-01 7.06612885e-01 -2.46777594e-01 -3.06390449e-02
4.36614513e-01 1.44431576e-01 -4.06199843e-01 -5.76046109e-01
7.50042558e-01 4.22553658e-01 5.15706897e-01 -1.28755581e+00
7.91408002e-01 -2.77934354e-02 -2.77652144e-02 -8.27727616e-01
-1.29221082e+00 -3.68492305e-01 -2.75548130e-01 -3.56236488e-01
9.10552621e-01 -8.98097098e-01 -1.35573387e+00 2.76356846e-01
-1.33980954e+00 -1.51683643e-01 -7.06150085e-02 6.27252042e-01
-2.81548947e-01 5.81313252e-01 -7.92516530e-01 -7.60048866e-01
-4.80809540e-01 -1.20711839e+00 1.16116166e+00 5.75116336e-01
-2.58856654e-01 -8.45855474e-01 6.54819384e-02 1.46476090e+00
1.61703393e-01 -1.34011149e-01 1.49092889e+00 -6.53854728e-01
-7.20652819e-01 6.84821457e-02 -2.00486794e-01 -4.76874150e-02
-4.17190164e-01 -1.16120046e-02 -1.15327382e+00 8.34498927e-02
-9.07201469e-02 -8.45206857e-01 8.50184441e-01 -6.61444217e-02
1.38264716e+00 -6.72792122e-02 1.62795529e-01 6.18190244e-02
9.61842656e-01 -3.46002914e-02 1.19811893e-01 -3.81942727e-02
7.44795024e-01 9.31528389e-01 6.62014484e-01 1.32890083e-02
9.33491647e-01 3.53948236e-01 3.07588100e-01 1.39190838e-01
-6.56549707e-02 -4.42091018e-01 4.62314546e-01 1.57527304e+00
2.21096635e-01 -2.43447393e-01 -1.07108760e+00 4.87573326e-01
-1.84374869e+00 -8.39874268e-01 -6.42392576e-01 1.71881497e+00
1.26788354e+00 -1.02555670e-01 -2.51365423e-01 6.58237711e-02
2.43271053e-01 -2.79906660e-01 -4.35445756e-01 -2.76790142e-01
-7.81297237e-02 2.66777515e-01 -3.33374500e-01 4.38953489e-01
-6.83757067e-01 7.40205288e-01 5.51609421e+00 1.08478379e+00
-5.44942975e-01 1.00584373e-01 2.60780543e-01 1.18483700e-01
-1.00930309e+00 -3.25585097e-01 -5.17696798e-01 2.24908337e-01
8.88258755e-01 2.65578300e-01 2.68340230e-01 2.59094238e-01
-1.58091977e-01 -3.03456843e-01 -1.11892724e+00 1.20547879e+00
2.93822318e-01 -1.13919902e+00 4.18121964e-01 -5.57747066e-01
7.75763154e-01 -2.96770394e-01 -7.70327449e-02 7.39720285e-01
3.21109653e-01 -9.65993822e-01 6.34348333e-01 6.07211173e-01
3.70731354e-01 -6.21085882e-01 6.52883828e-01 7.09995508e-01
-9.83891785e-01 -1.08454183e-01 -1.81116253e-01 1.67981789e-01
1.24877505e-02 7.42487788e-01 -6.90922141e-01 8.10286880e-01
6.61783040e-01 2.98491687e-01 -9.47399080e-01 7.87309170e-01
-6.98107421e-01 8.19494843e-01 -2.39539891e-01 -5.09912789e-01
2.11937562e-01 -1.53208241e-01 3.72361928e-01 1.08006561e+00
1.05498649e-01 3.56340557e-01 -9.75530297e-02 1.13321531e+00
-3.54238153e-01 1.67078272e-01 -7.92449415e-02 -4.41603541e-01
3.28324765e-01 1.36605561e+00 -5.08322120e-01 -3.48327637e-01
-4.33014780e-01 7.19026864e-01 3.12179148e-01 4.39539045e-01
-7.61143029e-01 -3.66030216e-01 1.37955666e-01 -5.53449214e-01
-7.36243352e-02 -1.21678352e-01 -1.05508879e-01 -1.54759157e+00
3.81952245e-03 -1.32362759e+00 6.97705328e-01 -1.13090789e+00
-1.67472136e+00 4.69282418e-01 -6.10951334e-03 -9.38739657e-01
7.63436556e-02 -6.02837563e-01 -3.70450199e-01 6.16418779e-01
-1.65558553e+00 -1.24265981e+00 -4.58914161e-01 6.63017988e-01
8.16140354e-01 -2.30129153e-01 8.56644511e-01 3.41027021e-01
-7.11029291e-01 4.76210833e-01 -2.54958540e-01 -1.31531721e-02
8.79907250e-01 -1.49537146e+00 -5.58577590e-02 5.53091466e-01
4.03840572e-01 8.46091866e-01 6.58788621e-01 -3.98184448e-01
-2.11395645e+00 -6.05240881e-01 9.67970669e-01 -8.98634911e-01
6.25786543e-01 -3.10398519e-01 -1.17021704e+00 3.07856262e-01
4.49490130e-01 -4.85508889e-01 1.05061495e+00 1.24248877e-01
-6.30263567e-01 -1.42067730e-01 -1.00382972e+00 7.10784435e-01
5.17199755e-01 -5.97172141e-01 -1.18439162e+00 3.54698241e-01
9.63608086e-01 -4.02942151e-01 -9.72740293e-01 4.36528116e-01
4.18514073e-01 -5.78455627e-01 9.76505160e-01 -7.13007033e-01
7.59221911e-01 -3.50423455e-01 -1.54241636e-01 -1.18129992e+00
2.16754586e-01 -4.52077061e-01 -2.42090121e-01 1.47196841e+00
4.15708780e-01 -2.02654347e-01 4.78991747e-01 6.66814029e-01
-1.03561266e-03 -7.64912903e-01 -7.41860330e-01 -2.60667682e-01
3.38021517e-02 -7.50233173e-01 5.58278918e-01 1.03723669e+00
1.56537086e-01 9.08059776e-01 -3.07943323e-04 1.05441861e-01
7.02402055e-01 5.56793809e-01 7.66499698e-01 -1.20940876e+00
-1.63742363e-01 -5.42252064e-01 3.62229764e-01 -1.36669087e+00
3.07248265e-01 -9.29372907e-01 9.00930166e-02 -1.80010033e+00
2.31055602e-01 -1.91709250e-01 -2.17509702e-01 3.84406984e-01
-6.34208083e-01 -2.04872396e-02 3.61287266e-01 -1.12109788e-01
-1.06058788e+00 8.35436165e-01 1.78622353e+00 -4.77970481e-01
1.22399561e-01 -1.63063500e-02 -6.27391219e-01 6.56881273e-01
5.21149814e-01 -3.06721210e-01 -6.84432685e-01 -5.63738525e-01
8.91691208e-01 2.91771233e-01 4.29494351e-01 -5.58290005e-01
4.51925308e-01 -3.47539276e-01 2.87562102e-01 -9.16064143e-01
4.66055661e-01 -8.81487727e-01 -3.60150099e-01 2.61132151e-01
-6.76936448e-01 4.97660711e-02 2.68186092e-01 5.19332111e-01
-4.20875341e-01 -3.94676685e-01 2.18250439e-01 -6.69244677e-02
-5.76708376e-01 -1.62386045e-01 -2.63112396e-01 4.52180862e-01
6.40396595e-01 4.50696528e-01 -6.98246002e-01 -5.04746139e-01
-3.48045141e-01 8.90665948e-01 -3.60781193e-01 6.21292472e-01
8.10151100e-01 -1.27152061e+00 -7.62893319e-01 -2.98751984e-02
3.57965350e-01 1.65804341e-01 5.41110814e-01 9.76131797e-01
-3.78949076e-01 6.61689103e-01 1.25319496e-01 -8.18399251e-01
-1.23766220e+00 3.35767627e-01 6.54844344e-02 -2.21564263e-01
-3.81202936e-01 1.19274306e+00 4.90233302e-02 -9.54415858e-01
4.78159815e-01 -5.80928087e-01 -4.76923406e-01 2.88031787e-01
5.75485587e-01 2.53097892e-01 -5.39151542e-02 -1.59042910e-01
-2.12120548e-01 7.96267688e-01 2.38273308e-01 -1.46953180e-01
1.22119045e+00 7.03119338e-02 -1.73581123e-01 5.73600411e-01
7.36522198e-01 2.26307631e-01 -7.45148659e-01 -2.63541013e-01
1.49812013e-01 -2.06118375e-01 -3.08258206e-01 -1.28584099e+00
-6.95731282e-01 1.34916890e+00 5.72664849e-02 1.03965431e-01
1.07559764e+00 1.15914762e-01 9.19974923e-01 8.46697271e-01
6.83663860e-02 -8.56509984e-01 5.73143899e-01 5.34242034e-01
1.07770014e+00 -1.49713778e+00 -3.03640693e-01 -3.82857293e-01
-8.27579081e-01 1.23452294e+00 8.95532370e-01 6.51206553e-01
1.00120388e-01 -1.40139565e-01 4.05957133e-01 -4.28055227e-01
-1.16317940e+00 -3.46574076e-02 6.97153151e-01 1.40677914e-01
6.23035848e-01 8.64460096e-02 -1.00457363e-01 8.77567887e-01
-1.45441025e-01 -1.28929511e-01 2.95241266e-01 6.69863939e-01
-3.31150800e-01 -1.13375926e+00 -6.55023873e-01 1.41349226e-01
-2.16802344e-01 -8.57464299e-02 -6.12211645e-01 4.65667933e-01
-8.22556913e-02 1.44352102e+00 -4.53158468e-01 -2.48884901e-01
5.99928856e-01 3.65868717e-01 8.12419415e-01 -4.66529042e-01
-1.00520599e+00 -6.82327971e-02 2.89221928e-02 -4.50115561e-01
-6.73375547e-01 -2.04363510e-01 -1.44960880e+00 -2.39594877e-01
-1.14308394e-01 4.71914202e-01 5.98627448e-01 1.13470852e+00
2.21661329e-01 9.17407274e-01 7.03580752e-02 -7.02473521e-02
-7.16671109e-01 -1.00708878e+00 1.67620167e-01 4.81878966e-01
3.65575731e-01 -3.46131712e-01 -4.00941148e-02 -6.14491217e-02] | [11.085721015930176, 7.971562385559082] |
170f3778-d997-428e-9148-39736b7b7ef4 | probing-what-different-nlp-tasks-teach | 1904.11544 | null | https://arxiv.org/abs/1904.11544v2 | https://arxiv.org/pdf/1904.11544v2.pdf | Probing What Different NLP Tasks Teach Machines about Function Word Comprehension | We introduce a set of nine challenge tasks that test for the understanding of function words. These tasks are created by structurally mutating sentences from existing datasets to target the comprehension of specific types of function words (e.g., prepositions, wh-words). Using these probing tasks, we explore the effects of various pretraining objectives for sentence encoders (e.g., language modeling, CCG supertagging and natural language inference (NLI)) on the learned representations. Our results show that pretraining on language modeling performs the best on average across our probing tasks, supporting its widespread use for pretraining state-of-the-art NLP models, and CCG supertagging and NLI pretraining perform comparably. Overall, no pretraining objective dominates across the board, and our function word probing tasks highlight several intuitive differences between pretraining objectives, e.g., that NLI helps the comprehension of negation. | ['Ellie Pavlick', 'Samuel R. Bowman', 'Patrick Xia', 'Alex Wang', 'Roma Patel', 'Tal Linzen', 'R. Thomas McCoy', 'Najoung Kim', 'Benjamin Van Durme', 'Adam Poliak', 'Ian Tenney', 'Alexis Ross'] | 2019-04-25 | probing-what-different-nlp-tasks-teach-1 | https://aclanthology.org/S19-1026 | https://aclanthology.org/S19-1026.pdf | semeval-2019-6 | ['ccg-supertagging'] | ['natural-language-processing'] | [ 3.37115020e-01 7.53117681e-01 -2.39178762e-01 -7.59035349e-01
-6.86218917e-01 -6.19096637e-01 7.75619209e-01 4.74265605e-01
-6.60091817e-01 6.76322937e-01 1.04169333e+00 -7.05316186e-01
2.29784518e-01 -9.40757036e-01 -1.14877844e+00 -1.62320599e-01
-6.88786134e-02 3.96475673e-01 1.10739149e-01 -5.30416191e-01
1.57475129e-01 -7.84222037e-03 -1.18680120e+00 7.81116426e-01
6.68528318e-01 6.69845104e-01 3.32116127e-01 4.64955360e-01
-3.19026470e-01 1.03414822e+00 -6.50799215e-01 -5.97871780e-01
-1.31336257e-01 -2.24330395e-01 -1.16258597e+00 -5.65616548e-01
4.25262690e-01 -5.16720951e-01 -1.47916198e-01 7.81298757e-01
3.10906053e-01 1.97638974e-01 5.74099243e-01 -7.72623301e-01
-1.03230333e+00 1.52804959e+00 -1.16282158e-01 3.59871298e-01
6.21131957e-01 2.89519608e-01 1.79476178e+00 -8.65982831e-01
6.45064950e-01 1.62678885e+00 7.28188217e-01 7.22042143e-01
-1.54744172e+00 -5.56015193e-01 3.58177781e-01 1.04626373e-03
-9.04564619e-01 -5.29388607e-01 4.10042137e-01 -1.78279802e-01
1.70662701e+00 2.28794608e-02 2.00382397e-01 1.33764005e+00
4.84342754e-01 9.32518244e-01 8.96403491e-01 -5.13331771e-01
-1.21662483e-01 -7.46293291e-02 5.20154834e-01 8.33963513e-01
3.63587946e-01 1.52143255e-01 -8.74380231e-01 -7.60980397e-02
2.89138466e-01 -4.54656005e-01 -4.21077758e-01 1.98647335e-01
-1.18547463e+00 1.09416938e+00 4.20239359e-01 2.95452833e-01
-1.13469854e-01 4.63598520e-01 5.82839489e-01 4.29475605e-01
2.88886547e-01 1.03650141e+00 -1.06772304e+00 -1.99991941e-01
-4.31739777e-01 2.61860728e-01 9.87642109e-01 8.19349229e-01
7.79933512e-01 -1.21619366e-02 -5.29308617e-01 7.68189251e-01
4.34559077e-01 1.88719228e-01 5.71002662e-01 -7.91623354e-01
7.43837595e-01 4.68261272e-01 -9.53110084e-02 -5.06012380e-01
-6.11698210e-01 -5.18184900e-01 -7.01070670e-03 -4.49957222e-01
6.35906339e-01 -4.55403507e-01 -9.64034557e-01 2.38929510e+00
-3.11331362e-01 -1.52227655e-01 2.37949744e-01 3.89176995e-01
1.23180330e+00 9.23337579e-01 6.06194794e-01 1.03558354e-01
1.52683091e+00 -8.57546449e-01 -6.82983160e-01 -1.08485425e+00
1.20375311e+00 -4.13899660e-01 1.62990189e+00 1.64591834e-01
-1.38071775e+00 -5.89792550e-01 -9.93178070e-01 -5.34195900e-01
-5.28520226e-01 -3.78764451e-01 7.21832335e-01 3.33477288e-01
-1.10442019e+00 5.70511997e-01 -8.11659276e-01 -3.79086554e-01
1.70447916e-01 2.51291096e-01 -4.04163450e-01 -1.99886486e-01
-1.61573613e+00 1.40913725e+00 6.65342808e-01 -2.04122320e-01
-9.03503835e-01 -1.17548954e+00 -1.29081202e+00 4.72979784e-01
1.73033148e-01 -6.96642041e-01 1.69035697e+00 -7.72569239e-01
-1.18510222e+00 1.26026654e+00 -2.12428004e-01 -7.40675926e-01
-9.13685486e-02 -5.54946840e-01 -1.75172806e-01 -2.23675519e-01
1.96865678e-01 1.10203540e+00 4.57552969e-01 -1.02076197e+00
-4.64315593e-01 -1.22615010e-01 3.54137152e-01 2.48100787e-01
-3.16136181e-01 1.33201152e-01 1.51176497e-01 -4.58325088e-01
-1.22178935e-01 -3.24749947e-01 3.46404552e-01 -4.04125720e-01
-5.64827144e-01 -5.96415818e-01 2.83513546e-01 -7.75279701e-01
1.11068237e+00 -2.12154746e+00 1.89413816e-01 -1.36499003e-01
1.56273499e-01 -1.48939088e-01 -4.11670238e-01 5.71492434e-01
-3.40458870e-01 4.68470216e-01 -1.09854393e-01 -4.82137471e-01
2.91656494e-01 5.41395307e-01 -4.37401414e-01 6.71329573e-02
4.61738348e-01 1.42022133e+00 -7.72012949e-01 -2.08960131e-01
-4.21284437e-02 1.18759885e-01 -8.45638812e-01 2.86463797e-01
-7.54604936e-01 1.23819254e-01 3.95485573e-02 2.57904679e-01
1.69164330e-01 -1.70922428e-01 2.05396429e-01 -2.47146770e-01
7.48867691e-02 1.39640081e+00 -3.24658751e-01 1.55412483e+00
-6.76032841e-01 4.35610026e-01 -9.70942713e-03 -9.06491101e-01
5.21358430e-01 2.17403233e-01 -2.11518154e-01 -7.35370219e-01
2.06037566e-01 2.73639590e-01 7.69305646e-01 -3.47144812e-01
3.84998292e-01 -5.25798976e-01 -2.99099922e-01 4.24189210e-01
4.54867244e-01 -1.32721271e-02 1.18826702e-01 1.62723526e-01
1.29257858e+00 3.96940298e-02 4.42667931e-01 -6.19875669e-01
2.67506123e-01 -1.00885235e-01 3.84010941e-01 8.64377677e-01
4.55692224e-02 3.15485954e-01 8.15593421e-01 -1.23781696e-01
-6.91065192e-01 -1.17936599e+00 3.25956359e-03 1.86589897e+00
-2.61860430e-01 -6.05478168e-01 -6.27754152e-01 -6.82097912e-01
1.81274220e-01 1.50298786e+00 -8.75122368e-01 -6.07952416e-01
-8.73247743e-01 -5.31657517e-01 7.43977785e-01 9.26481366e-01
3.39603275e-01 -1.48081088e+00 -5.01305282e-01 2.86654383e-01
-1.76535532e-01 -1.10994601e+00 -2.90200055e-01 5.62588632e-01
-8.03791881e-01 -7.01487184e-01 -1.59981042e-01 -1.07208633e+00
3.91118616e-01 -1.21001959e-01 1.85145378e+00 3.56092244e-01
2.31683359e-01 3.35028112e-01 -4.41273510e-01 -3.98240775e-01
-6.13600194e-01 5.00587046e-01 -4.26510841e-01 -5.90261281e-01
4.74910706e-01 -4.59153831e-01 -3.55276436e-01 -2.82314420e-03
-8.18079829e-01 2.67615728e-02 6.45822406e-01 9.56907272e-01
1.22359186e-01 -3.65734696e-01 6.12052917e-01 -1.46926200e+00
9.85177875e-01 -5.61566651e-01 -1.42077252e-01 3.21787655e-01
-2.36965671e-01 3.72700870e-01 6.02246940e-01 -2.41714284e-01
-1.01989567e+00 -6.19421422e-01 -6.03059113e-01 3.47631872e-01
-4.89778072e-02 1.04204297e+00 -2.71156967e-01 2.84203798e-01
8.35734427e-01 1.66805923e-01 -4.36174832e-02 -3.75494272e-01
3.84660512e-01 -1.58345867e-02 3.47063661e-01 -9.36217546e-01
6.27646625e-01 -8.91723707e-02 -5.97945809e-01 -7.85583258e-01
-1.48085165e+00 -7.82804713e-02 -1.41126752e-01 5.55345237e-01
1.22245646e+00 -9.35911596e-01 -6.50097787e-01 2.79159993e-01
-1.31943965e+00 -8.62295628e-01 -3.93632233e-01 1.36254221e-01
-4.51551706e-01 1.26110569e-01 -7.34296560e-01 -2.99631447e-01
-4.06101108e-01 -9.82648015e-01 1.00719094e+00 -2.18169257e-01
-7.80124485e-01 -1.28832352e+00 -3.35881531e-01 2.67816812e-01
5.94328880e-01 -1.05342530e-01 1.71952462e+00 -1.05554867e+00
-2.48361364e-01 2.55844444e-01 -1.83354646e-01 4.72189337e-01
-1.38588285e-03 -4.63382870e-01 -9.20234144e-01 -1.72818378e-01
1.03257768e-01 -8.46935213e-01 1.09811783e+00 3.31009060e-01
1.07388055e+00 -6.16478026e-01 -2.20705152e-01 4.71515566e-01
1.08395481e+00 -1.09598376e-01 6.69754684e-01 1.86917633e-01
4.70944047e-01 5.61685562e-01 3.86438549e-01 5.32931313e-02
6.27363324e-01 2.05157012e-01 3.08855742e-01 2.33231455e-01
-7.04067424e-02 -6.35427237e-01 6.68218017e-01 4.32725608e-01
5.65908194e-01 -5.91418386e-01 -1.08038509e+00 5.76132715e-01
-1.43355191e+00 -4.58500892e-01 2.25033402e-01 1.91330564e+00
1.26015902e+00 6.23473227e-01 -3.32854867e-01 -1.14641920e-01
1.99920177e-01 5.41513085e-01 -4.01489407e-01 -6.52845562e-01
-1.74634904e-02 8.19157720e-01 3.85419518e-01 1.03788817e+00
-7.53784359e-01 1.36079168e+00 6.37491131e+00 3.22370976e-01
-9.42575693e-01 2.62141466e-01 5.11343956e-01 2.83425272e-01
-7.94840217e-01 1.62801798e-02 -1.10716820e+00 2.65979677e-01
1.08097386e+00 -1.05300173e-02 3.64687234e-01 4.81233507e-01
-9.49702039e-02 -6.53530732e-02 -1.68876076e+00 3.35487992e-01
1.13393422e-02 -1.26103318e+00 2.96794415e-01 -2.50505894e-01
2.77548641e-01 3.85592341e-01 -1.06344283e-01 7.93278933e-01
4.93509829e-01 -1.30093682e+00 9.15918827e-01 1.36565804e-01
4.26978588e-01 -4.45366263e-01 8.21925282e-01 3.67458999e-01
-6.06484950e-01 -8.38302374e-02 -2.52484202e-01 -4.68928516e-01
3.35853726e-01 2.82575339e-01 -8.37869525e-01 -7.24194422e-02
3.79718989e-01 4.95946288e-01 -6.93102181e-01 4.06285524e-01
-9.63533640e-01 9.98043716e-01 -2.74340153e-01 -3.07416886e-01
3.57800275e-01 3.33413810e-01 2.68307686e-01 1.31877148e+00
-8.86502638e-02 8.68528932e-02 1.41789973e-01 9.23561811e-01
-4.37532216e-01 -3.11264372e-03 -4.86892670e-01 -3.21061641e-01
5.85584283e-01 7.97362983e-01 -2.65516341e-01 -3.33248228e-01
-6.35016620e-01 6.25535488e-01 7.29549766e-01 2.49098942e-01
-8.37863326e-01 -2.03522459e-01 8.13015163e-01 4.13609177e-01
2.82067090e-01 -3.47465873e-01 -5.54430962e-01 -1.13441336e+00
-1.28425449e-01 -9.95813727e-01 5.50736845e-01 -9.16318119e-01
-1.28551960e+00 2.39800051e-01 1.40852302e-01 -1.48781538e-01
-5.00181556e-01 -9.49736536e-01 -9.18794632e-01 9.51242507e-01
-1.62192202e+00 -1.08636999e+00 1.42228380e-02 3.23223203e-01
5.05314112e-01 2.35250860e-01 1.01922619e+00 -1.25118971e-01
-2.90978730e-01 8.11391473e-01 -5.12346566e-01 3.97789240e-01
5.74024081e-01 -1.40192175e+00 8.71334434e-01 7.12603927e-01
7.15609714e-02 1.26792824e+00 7.95543134e-01 -6.46571517e-01
-1.36325765e+00 -9.10411417e-01 1.33922839e+00 -5.43198049e-01
9.03421283e-01 -9.21324611e-01 -1.12972045e+00 1.34903133e+00
5.44377625e-01 -3.47828776e-01 7.88352191e-01 7.41607010e-01
-5.98635733e-01 1.51448414e-01 -1.00079942e+00 5.61181009e-01
1.16723573e+00 -5.93943238e-01 -1.24057961e+00 4.86486286e-01
1.26791370e+00 -4.59436685e-01 -7.09621727e-01 2.94359595e-01
3.00802052e-01 -6.73602104e-01 8.86097372e-01 -1.10371423e+00
9.00486946e-01 2.75444388e-01 -3.05995882e-01 -1.53247654e+00
-4.98885334e-01 -2.21159518e-01 -1.25900656e-01 1.18522930e+00
1.13492155e+00 -9.83845353e-01 7.08919704e-01 6.84555352e-01
-5.25501788e-01 -7.01647758e-01 -7.70768523e-01 -4.79811549e-01
5.32516003e-01 -3.81312817e-01 2.35385507e-01 7.65434802e-01
2.04706788e-01 1.08148766e+00 1.51202098e-01 -4.86825220e-02
1.05550975e-01 -2.05454603e-01 3.82126331e-01 -1.05360663e+00
-6.70143068e-01 -3.85761380e-01 5.48034050e-02 -1.23134792e+00
6.64292395e-01 -1.24740684e+00 2.22411841e-01 -1.83796680e+00
-3.61377150e-02 -4.00636159e-02 -5.99811960e-04 1.02043402e+00
-3.51632476e-01 -2.96428084e-01 9.96160507e-02 -4.10347372e-01
-1.81380957e-01 5.45288801e-01 1.05457282e+00 -9.10972729e-02
9.73328277e-02 -2.34284908e-01 -1.31657326e+00 7.92435169e-01
7.82496452e-01 -2.93558598e-01 -5.61191976e-01 -1.17518020e+00
5.49339294e-01 -1.67631045e-01 1.79747820e-01 -6.19388640e-01
-5.50866872e-02 1.69785902e-01 2.23948374e-01 -1.50014460e-01
3.87345105e-01 -3.03506136e-01 -7.46358633e-01 5.30283034e-01
-7.52613842e-01 3.72705042e-01 5.89677215e-01 1.80363417e-01
-2.50788480e-01 -4.76427168e-01 8.21204424e-01 -4.76712734e-01
-6.48712337e-01 -3.12035143e-01 -4.04794276e-01 8.73266637e-01
3.35606694e-01 -1.34366333e-01 -5.87877452e-01 -4.66177642e-01
-7.27733731e-01 4.27951813e-01 -2.09069476e-02 4.72794443e-01
5.90707898e-01 -7.76540995e-01 -7.67249882e-01 2.74119765e-01
1.99936032e-01 2.56159622e-02 -2.84218520e-01 5.97750306e-01
-2.98261672e-01 4.94983077e-01 2.62759496e-02 -2.25003824e-01
-8.15215886e-01 2.20964849e-01 3.15646738e-01 -4.49445128e-01
-4.18582052e-01 1.46745896e+00 6.49564505e-01 -8.30271900e-01
2.45248601e-01 -1.08317852e+00 -5.26508456e-03 8.36149044e-03
3.32194626e-01 -9.33511108e-02 2.12865606e-01 -9.68354195e-02
-3.74230683e-01 -2.80683637e-02 -5.25053203e-01 -6.27028272e-02
1.36352038e+00 2.98274845e-01 -2.73438752e-01 5.08244991e-01
1.27446270e+00 -9.64785516e-02 -6.51592135e-01 -2.16401875e-01
3.53159964e-01 1.00044399e-01 -1.18917182e-01 -1.18631971e+00
-5.68145812e-01 9.02311444e-01 -2.07800031e-01 1.00717522e-01
7.44315863e-01 2.98146307e-01 8.32788229e-01 6.90166533e-01
2.16526389e-01 -7.61585355e-01 -3.43676284e-02 1.16835105e+00
1.02865291e+00 -9.73518133e-01 -2.60330588e-01 -5.65191746e-01
-4.09018934e-01 7.40131497e-01 9.84254479e-01 -1.82343498e-01
5.96387267e-01 2.69825369e-01 -1.53328225e-01 -3.57996583e-01
-1.08449841e+00 -6.26714677e-02 8.56601726e-03 2.13543132e-01
1.08801699e+00 3.55476253e-02 -5.72160780e-01 6.74575329e-01
-8.21594059e-01 -4.92063999e-01 3.77996624e-01 8.67021561e-01
-6.26318812e-01 -9.31592941e-01 6.74853772e-02 8.52745354e-01
-5.70987403e-01 -8.27847362e-01 -5.74758053e-01 9.67401624e-01
-5.98630309e-02 7.26688802e-01 2.31342673e-01 -1.21665485e-01
5.28017342e-01 5.45350015e-01 5.43377817e-01 -1.37196875e+00
-1.02060080e+00 -5.92506349e-01 6.96386278e-01 -5.26184559e-01
1.74626544e-01 -6.14504457e-01 -1.44265938e+00 -1.48892254e-01
-2.85491079e-01 2.39647463e-01 2.61830330e-01 1.21697760e+00
1.61124766e-01 6.04927957e-01 -1.49568975e-01 -3.28292191e-01
-8.53656590e-01 -1.48379910e+00 -9.48283002e-02 5.05964637e-01
2.50782520e-01 -5.88484228e-01 -3.52436274e-01 -4.12186027e-01] | [10.627246856689453, 8.898904800415039] |
b51b4703-1ddf-4f84-bc5b-48b2c1dcef38 | zambezi-voice-a-multilingual-speech-corpus | 2306.04428 | null | https://arxiv.org/abs/2306.04428v2 | https://arxiv.org/pdf/2306.04428v2.pdf | Zambezi Voice: A Multilingual Speech Corpus for Zambian Languages | This work introduces Zambezi Voice, an open-source multilingual speech resource for Zambian languages. It contains two collections of datasets: unlabelled audio recordings of radio news and talk shows programs (160 hours) and labelled data (over 80 hours) consisting of read speech recorded from text sourced from publicly available literature books. The dataset is created for speech recognition but can be extended to multilingual speech processing research for both supervised and unsupervised learning approaches. To our knowledge, this is the first multilingual speech dataset created for Zambian languages. We exploit pretraining and cross-lingual transfer learning by finetuning the Wav2Vec2.0 large-scale multilingual pre-trained model to build end-to-end (E2E) speech recognition models for our baseline models. The dataset is released publicly under a Creative Commons BY-NC-ND 4.0 license and can be accessed via https://github.com/unza-speech-lab/zambezi-voice . | ['Antonios Anastasopoulos', 'Mayumbo Nyirenda', 'David Zulu', 'Martin Phiri', 'Mofya Phiri', 'Bangiwe Zulu', 'Stanly Mwape', 'Kalinda Siaminwe', 'Claytone Sikasote'] | 2023-06-07 | null | null | null | null | ['cross-lingual-transfer'] | ['natural-language-processing'] | [-1.56809077e-01 1.64734274e-01 -3.18525970e-01 -5.39126396e-01
-1.31031311e+00 -7.62962639e-01 7.01065660e-01 -1.71882957e-01
-4.85082358e-01 5.82938313e-01 7.12515414e-01 -1.04079270e+00
3.94633502e-01 -4.21544045e-01 -6.46481216e-01 -4.59036380e-01
1.02523908e-01 8.27096701e-01 -1.33302465e-01 -2.39066079e-01
-2.13079572e-01 2.05326155e-01 -1.12067175e+00 5.25272310e-01
4.22505289e-01 3.86212885e-01 4.94687617e-01 1.05709624e+00
3.82207558e-02 8.44700754e-01 -4.22811151e-01 -4.64824736e-01
-5.49763888e-02 -1.99577749e-01 -1.28663743e+00 -1.90301016e-02
3.31811756e-01 -1.31161660e-01 -6.30765319e-01 7.30216324e-01
8.61534894e-01 1.18875444e-01 2.13142425e-01 -8.12254786e-01
-8.37973535e-01 1.19210792e+00 6.29055128e-02 6.30531251e-01
2.29180276e-01 5.44610806e-02 1.17197454e+00 -1.03778672e+00
6.76989198e-01 1.49639606e+00 2.95284927e-01 5.53203046e-01
-9.36826229e-01 -8.32081676e-01 -2.43406624e-01 2.19065532e-01
-1.23007870e+00 -1.32561111e+00 6.22110546e-01 -3.37251753e-01
1.35894132e+00 2.34794974e-01 7.76949301e-02 1.61625421e+00
-3.04198384e-01 1.08644795e+00 1.15795040e+00 -8.46205175e-01
-2.31278360e-01 4.13833261e-01 -1.33596972e-01 5.16254663e-01
-5.58045566e-01 1.40910611e-01 -5.72424948e-01 -3.24031413e-02
5.43493688e-01 -5.26199400e-01 -2.62222886e-01 2.20551178e-01
-1.35954201e+00 8.45245123e-01 2.43203733e-02 5.71524560e-01
-2.22945556e-01 -2.98825622e-01 6.71996236e-01 7.11308300e-01
7.85063624e-01 6.57850802e-02 -9.97468174e-01 -4.55432326e-01
-8.07328761e-01 -1.17800273e-01 7.05870509e-01 8.44968200e-01
6.26809716e-01 3.86209190e-01 2.21665308e-01 1.78032506e+00
3.90214562e-01 8.91819656e-01 8.80587637e-01 -3.84684563e-01
7.76894867e-01 -1.68732524e-01 -5.35941958e-01 -3.66601288e-01
-1.39633343e-01 -1.05014659e-01 -2.71531433e-01 -4.01246607e-01
2.02030092e-01 -4.19936031e-01 -8.30132723e-01 1.43812525e+00
2.09789068e-01 3.45010869e-02 4.55132127e-01 5.57601869e-01
1.34152675e+00 1.14419007e+00 9.84099228e-03 -3.10250539e-02
1.30711484e+00 -1.49694335e+00 -8.29960644e-01 -1.05603658e-01
6.23281121e-01 -1.34065616e+00 1.29088199e+00 3.87067854e-01
-1.08961475e+00 -4.36786771e-01 -6.85899734e-01 -3.04655492e-01
-7.68555045e-01 2.53533840e-01 3.49804014e-01 7.37448812e-01
-1.15647817e+00 -3.44478875e-01 -7.97288537e-01 -6.07070506e-01
2.48610198e-01 2.41160346e-03 -5.14647245e-01 -1.07051276e-01
-1.27759778e+00 7.75996149e-01 2.63462126e-01 -2.09106296e-01
-1.31510127e+00 -4.76188958e-01 -1.00591409e+00 -3.54001284e-01
4.26144034e-01 2.63719916e-01 1.81683290e+00 -9.17175710e-01
-1.81470239e+00 1.20793271e+00 -2.95366555e-01 -4.33470398e-01
2.01645255e-01 -5.33041470e-02 -8.39635789e-01 -1.35488719e-01
9.03142616e-02 5.57410061e-01 6.06800497e-01 -1.00275588e+00
-7.00144589e-01 -3.43323678e-01 -2.44468078e-01 1.14135437e-01
-1.58830702e-01 6.55990243e-01 -5.87032199e-01 -8.82517338e-01
-2.97137320e-01 -1.11031199e+00 1.78671762e-01 -1.05563414e+00
-3.23183298e-01 -2.51362592e-01 8.29013586e-01 -1.38056314e+00
1.16462088e+00 -2.01853895e+00 -1.78738371e-01 -1.72512874e-01
-4.21233118e-01 4.93018538e-01 -4.98626679e-01 7.35398412e-01
-8.06355849e-02 7.04040602e-02 6.84319735e-02 -5.74602962e-01
-1.15905538e-01 2.67456979e-01 -3.99266839e-01 3.97476256e-01
4.60014530e-02 8.28810990e-01 -1.00187874e+00 4.37184609e-03
3.28817189e-01 7.38650620e-01 -2.17921555e-01 1.88297153e-01
2.03247488e-01 7.22960770e-01 3.64639759e-02 6.28368497e-01
3.77336890e-01 5.95627844e-01 2.32937917e-01 1.82980061e-01
-3.97198290e-01 1.15156758e+00 -5.51985919e-01 1.64464259e+00
-1.02677453e+00 9.05123591e-01 4.76180077e-01 -8.91670644e-01
8.67713451e-01 1.08960485e+00 3.21917050e-02 -7.92815447e-01
3.13799590e-01 5.97398341e-01 -2.21858546e-02 -5.00434101e-01
5.83629906e-01 2.95445118e-02 -1.84642166e-01 2.88094103e-01
7.76437104e-01 -4.80820060e-01 4.15297374e-02 -5.97047955e-02
7.33612478e-01 -2.93881804e-01 2.57712424e-01 -2.93642938e-01
6.88006639e-01 -9.19991285e-02 3.49409282e-01 3.82595092e-01
-2.41760507e-01 4.68384445e-01 1.60017163e-02 -2.74241656e-01
-1.03649950e+00 -1.00229287e+00 -3.01650137e-01 1.70182419e+00
-7.95953214e-01 -4.08960074e-01 -6.33786798e-01 -6.56562626e-01
-2.43542939e-01 8.71995926e-01 -1.74290478e-01 5.33633292e-01
-5.19105852e-01 -1.60334319e-01 1.09425676e+00 2.01919734e-01
9.13567245e-02 -1.34617054e+00 2.87389457e-01 2.46170789e-01
-4.82599258e-01 -1.21299040e+00 -5.98578274e-01 3.33299309e-01
-3.13020915e-01 -7.33938992e-01 -8.17474663e-01 -1.23753679e+00
8.66286084e-02 1.14730142e-01 1.10591507e+00 -4.19967353e-01
1.84058733e-02 4.60796654e-01 -5.98217189e-01 -6.99743807e-01
-8.77557456e-01 3.34921181e-01 2.84821868e-01 -1.63964033e-01
6.45357668e-01 -3.60547036e-01 2.86894739e-01 1.65166736e-01
-6.01818740e-01 -6.46444783e-02 3.50002229e-01 8.34988892e-01
6.22825027e-01 -3.99440788e-02 8.68391752e-01 -6.94650233e-01
7.47050822e-01 -6.60957277e-01 -5.86700857e-01 1.54184476e-01
-1.63281001e-02 -3.10239434e-01 3.70708615e-01 -2.07875535e-01
-1.27545142e+00 -1.85402438e-01 -7.20615029e-01 -3.53113383e-01
-6.18835449e-01 6.02852821e-01 -3.90252650e-01 3.93085092e-01
2.18393847e-01 1.82080567e-01 -3.81335765e-01 -6.88071251e-01
8.55053246e-01 1.62920916e+00 7.64635921e-01 -5.01165926e-01
3.94810915e-01 1.82690881e-02 -1.03903246e+00 -1.39523983e+00
-4.89798188e-01 -8.26615214e-01 -5.86714745e-01 -2.21134812e-01
8.53833258e-01 -1.17819977e+00 -2.18831778e-01 6.65488183e-01
-1.06536579e+00 -7.15232253e-01 7.13435784e-02 5.57950497e-01
-4.51070011e-01 -1.08146861e-01 -9.08948421e-01 -8.54774356e-01
-3.79132181e-01 -1.37789214e+00 9.62483585e-01 -2.51013935e-01
-3.06295037e-01 -1.22734582e+00 4.69531357e-01 7.77220845e-01
5.30202687e-01 -5.67804515e-01 5.20825326e-01 -1.03200328e+00
8.48441347e-02 4.66947518e-02 2.62451023e-01 7.36602366e-01
3.41768593e-01 1.55539718e-02 -1.41060066e+00 -4.16820109e-01
-3.04175615e-01 -7.29248643e-01 6.28686190e-01 3.42885911e-01
5.60003996e-01 -6.17828310e-01 7.46354982e-02 3.10082018e-01
8.33411992e-01 4.88225073e-01 4.52121764e-01 3.01919222e-01
7.02058196e-01 7.09161818e-01 1.34804741e-01 1.09327629e-01
5.80782473e-01 6.10328913e-01 -1.92360625e-01 8.03968124e-03
-4.04094905e-01 -3.93346161e-01 6.62549794e-01 1.83910787e+00
2.22666129e-01 -3.04819703e-01 -1.39782703e+00 1.22271097e+00
-1.39534450e+00 -9.02477145e-01 7.80294687e-02 2.25393724e+00
1.30883837e+00 -2.74316192e-01 2.23463178e-01 8.16572309e-02
6.37444317e-01 2.84166962e-01 4.73372713e-02 -7.81100988e-01
-2.25034222e-01 4.99325126e-01 4.36387062e-01 1.10988891e+00
-1.33727229e+00 1.50151920e+00 5.06474447e+00 1.22630978e+00
-1.56300128e+00 7.50360429e-01 5.52099288e-01 -1.70505434e-01
-2.32982874e-01 -1.88492939e-01 -7.72787809e-01 2.94694126e-01
1.88905418e+00 -1.15006603e-01 6.89079583e-01 8.08625102e-01
3.10668796e-01 3.75822872e-01 -6.63417339e-01 1.07343197e+00
1.15037128e-01 -1.24256146e+00 -2.41763264e-01 5.49072474e-02
7.17159629e-01 1.08241320e+00 1.63058102e-01 7.09010601e-01
5.73943377e-01 -1.00040650e+00 8.32000077e-01 -2.55873740e-01
9.08530772e-01 -9.94094074e-01 6.19810820e-01 2.91860700e-01
-1.09640777e+00 2.34998196e-01 -1.48811281e-01 2.50682950e-01
2.80934036e-01 3.02204967e-01 -1.32586873e+00 4.21057254e-01
7.85980523e-01 6.20093048e-01 -3.39678138e-01 5.82300365e-01
-3.37903947e-01 1.54022956e+00 -1.43753022e-01 1.92330375e-01
5.59324384e-01 -4.07662690e-02 6.88206375e-01 1.71627247e+00
2.82952845e-01 -2.87556916e-01 1.78616360e-01 1.13183230e-01
-2.71839023e-01 4.74351436e-01 -8.30292523e-01 -4.30201083e-01
7.52310932e-01 1.08430862e+00 -4.05842155e-01 -3.25749785e-01
-6.91110790e-01 8.36955726e-01 2.67574370e-01 3.55016083e-01
-4.21475559e-01 -4.32585359e-01 5.25269091e-01 -7.44737759e-02
2.30034530e-01 -3.21652114e-01 1.72699943e-01 -1.19491339e+00
-3.23357314e-01 -1.39276147e+00 2.81389534e-01 -7.42138624e-01
-1.24709797e+00 1.12535369e+00 -2.06601605e-01 -7.57871032e-01
-4.99791473e-01 -7.24236548e-01 -3.77283871e-01 9.94922042e-01
-1.44861126e+00 -1.43188071e+00 3.22047234e-01 5.54578185e-01
1.24419177e+00 -7.91621089e-01 1.16227996e+00 6.77020311e-01
-4.86539662e-01 6.32919014e-01 3.92633498e-01 5.78909934e-01
8.45787346e-01 -1.00041759e+00 7.51373589e-01 6.54510558e-01
5.06669760e-01 4.35121268e-01 3.40374023e-01 -4.15440589e-01
-1.11452532e+00 -1.27207410e+00 1.52991414e+00 -6.22188687e-01
1.20491290e+00 -6.02129459e-01 -8.36555004e-01 1.17982233e+00
8.81710172e-01 -3.18568349e-01 9.60190654e-01 3.29395801e-01
-4.13553387e-01 1.18370794e-01 -7.34819293e-01 4.18634057e-01
5.09286940e-01 -1.12026751e+00 -6.88733339e-01 3.47142190e-01
7.45957613e-01 -2.85432041e-01 -7.19879866e-01 1.12492591e-01
4.30238485e-01 -4.11943108e-01 6.95998490e-01 -5.05828321e-01
1.17797151e-01 -4.83022891e-02 -6.77617192e-01 -1.76640904e+00
1.08298279e-01 -7.46932328e-01 4.06702042e-01 1.67971241e+00
8.63049865e-01 -7.28175342e-01 -2.90610716e-02 -3.90545398e-01
-4.65022892e-01 -4.10071641e-01 -1.20054746e+00 -9.38839197e-01
4.46580589e-01 -1.09076452e+00 6.43457055e-01 1.37972391e+00
6.26039058e-02 5.58350742e-01 -3.43699127e-01 3.17732900e-01
1.63430691e-01 -5.78641236e-01 8.02840948e-01 -6.76256895e-01
-2.70965844e-01 -4.83723313e-01 -2.30404094e-01 -8.32843304e-01
5.96462846e-01 -1.24971139e+00 2.05135167e-01 -1.21778083e+00
-2.63423175e-01 -4.14181143e-01 -6.96680397e-02 8.62250388e-01
2.15872064e-01 1.83016926e-01 -6.29905090e-02 8.02905113e-02
-1.22082628e-01 4.75983202e-01 7.24653661e-01 -4.35380429e-01
-3.97976547e-01 -1.70286149e-01 -2.34680414e-01 4.13904130e-01
1.01183438e+00 -4.38709497e-01 -4.43449467e-01 -8.81849110e-01
-4.73041058e-01 8.85105953e-02 -1.42784819e-01 -8.00263405e-01
-1.20216444e-01 4.53137010e-02 -4.53362286e-01 -6.08788311e-01
3.11426580e-01 -3.53276849e-01 -1.36418313e-01 -1.79579616e-01
-4.99933481e-01 2.48561576e-02 5.57506204e-01 -5.52357212e-02
-5.65199733e-01 -4.62092040e-03 6.63857460e-01 -5.81387654e-02
-7.15072870e-01 1.72471985e-01 -8.65383744e-01 9.63213369e-02
4.55005199e-01 4.11298424e-01 -5.61931252e-01 -6.58358991e-01
-6.78464413e-01 1.03218779e-01 4.19288203e-02 1.04628241e+00
2.49342278e-01 -1.35936379e+00 -1.11254835e+00 5.28572381e-01
3.36507112e-01 -4.50460881e-01 1.43913895e-01 7.11861610e-01
-3.68226171e-01 8.99133503e-01 4.81145084e-02 -3.27030689e-01
-1.30836511e+00 2.02829778e-01 6.88413531e-02 8.29353184e-02
-3.30847740e-01 1.08317578e+00 -7.42304046e-03 -1.41284335e+00
3.74848992e-01 -1.53677747e-01 -2.54343394e-02 5.27298637e-03
6.94952726e-01 2.62738407e-01 3.57194841e-01 -1.43174660e+00
-4.04524893e-01 -1.22873083e-01 -1.61321506e-01 -7.23798871e-01
1.25937247e+00 -4.24087942e-01 -1.53016299e-01 8.64305854e-01
1.47496378e+00 7.20972836e-01 -4.62770164e-01 -3.16801608e-01
-1.16915517e-01 -3.55398022e-02 5.07745922e-01 -8.76576662e-01
-9.42540407e-01 1.07889104e+00 4.69634235e-01 -1.55218933e-02
7.54196405e-01 3.19994718e-01 9.76457477e-01 4.77019578e-01
2.85708845e-01 -1.18044913e+00 -5.02760112e-01 1.17079043e+00
9.88675594e-01 -1.32166326e+00 -7.62716353e-01 -7.65240565e-02
-9.84809279e-01 8.61354172e-01 2.18019709e-01 2.87435561e-01
8.88323009e-01 3.24781954e-01 9.14544761e-01 4.85046282e-02
-7.65597701e-01 -2.99175143e-01 2.37280801e-01 7.84201801e-01
9.89790142e-01 6.00125909e-01 7.69258812e-02 7.18779206e-01
-8.71059537e-01 -3.01625222e-01 3.91675174e-01 7.88801551e-01
-1.89985424e-01 -1.34926724e+00 -3.38910967e-01 -1.39106745e-02
-7.64117301e-01 -5.42547405e-01 -4.16837603e-01 5.68938732e-01
-1.38497025e-01 1.51128697e+00 1.49565041e-01 -3.02033514e-01
2.42974013e-01 3.16470355e-01 1.30829573e-01 -7.66650558e-01
-3.64738941e-01 4.99620110e-01 6.01564050e-01 -1.84731647e-01
-3.87940526e-01 -5.82116663e-01 -9.22681510e-01 -1.66817844e-01
-2.58083731e-01 4.35805053e-01 9.70391810e-01 7.19576001e-01
1.72518522e-01 3.64739209e-01 6.24146938e-01 -9.90556657e-01
-1.07067721e-02 -1.28614688e+00 -4.50980902e-01 5.57986693e-03
4.16992337e-01 -2.87618548e-01 -3.12998801e-01 3.70025277e-01] | [14.360841751098633, 6.938343048095703] |
fb580a5b-6ef4-4c84-ae4c-5a402a8a049f | fine-grained-video-text-retrieval-with | 2003.00392 | null | https://arxiv.org/abs/2003.00392v1 | https://arxiv.org/pdf/2003.00392v1.pdf | Fine-grained Video-Text Retrieval with Hierarchical Graph Reasoning | Cross-modal retrieval between videos and texts has attracted growing attentions due to the rapid emergence of videos on the web. The current dominant approach for this problem is to learn a joint embedding space to measure cross-modal similarities. However, simple joint embeddings are insufficient to represent complicated visual and textual details, such as scenes, objects, actions and their compositions. To improve fine-grained video-text retrieval, we propose a Hierarchical Graph Reasoning (HGR) model, which decomposes video-text matching into global-to-local levels. To be specific, the model disentangles texts into hierarchical semantic graph including three levels of events, actions, entities and relationships across levels. Attention-based graph reasoning is utilized to generate hierarchical textual embeddings, which can guide the learning of diverse and hierarchical video representations. The HGR model aggregates matchings from different video-text levels to capture both global and local details. Experimental results on three video-text datasets demonstrate the advantages of our model. Such hierarchical decomposition also enables better generalization across datasets and improves the ability to distinguish fine-grained semantic differences. | ['Shizhe Chen', 'Qi Wu', 'Yida Zhao', 'Qin Jin'] | 2020-03-01 | fine-grained-video-text-retrieval-with-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Chen_Fine-Grained_Video-Text_Retrieval_With_Hierarchical_Graph_Reasoning_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Chen_Fine-Grained_Video-Text_Retrieval_With_Hierarchical_Graph_Reasoning_CVPR_2020_paper.pdf | cvpr-2020-6 | ['video-text-retrieval'] | ['computer-vision'] | [-2.19856098e-01 -4.90424693e-01 -2.97641158e-01 -1.78445339e-01
-8.77201498e-01 -6.68544471e-01 8.60105932e-01 4.73563969e-01
-1.20198347e-01 5.74768381e-03 8.44599545e-01 2.18478963e-01
-2.28889734e-01 -5.55911124e-01 -4.12446856e-01 -5.72215915e-01
9.57841948e-02 1.81008235e-01 3.05447966e-01 7.23247752e-02
1.98923215e-01 2.39713356e-01 -1.70645773e+00 8.36356699e-01
6.30008042e-01 9.98297632e-01 1.26672998e-01 5.88329494e-01
-5.82060814e-01 8.35935712e-01 -3.43403965e-01 -3.74757558e-01
2.84282845e-02 -3.79050285e-01 -5.13579547e-01 3.48825008e-01
8.81645501e-01 -4.51496184e-01 -9.64762211e-01 1.10117376e+00
2.82323986e-01 3.11955482e-01 7.90549517e-01 -1.48469567e+00
-1.21797073e+00 4.99053955e-01 -6.25874817e-01 1.20314851e-01
7.71676898e-01 9.62617323e-02 1.61847389e+00 -8.84760737e-01
5.95988989e-01 1.70157754e+00 4.17558789e-01 3.21841329e-01
-1.12181675e+00 -5.51177204e-01 4.69355166e-01 6.64783537e-01
-1.49500287e+00 -8.90192613e-02 7.98640013e-01 -6.63240433e-01
7.98721373e-01 2.36866772e-01 7.67714560e-01 1.13514400e+00
7.63133317e-02 1.22668338e+00 7.17731714e-01 -6.19401298e-02
-1.15335487e-01 -7.96610564e-02 1.39043599e-01 9.35581267e-01
3.14639866e-01 -5.52918017e-01 -7.58613706e-01 -1.81661561e-01
8.20393920e-01 6.16598904e-01 -5.05029857e-01 -7.60668337e-01
-1.42102814e+00 8.78925979e-01 5.14808357e-01 6.20729864e-01
-3.60084236e-01 1.50224850e-01 7.31720924e-01 1.13489948e-01
2.92468518e-01 2.48685867e-01 7.93117192e-03 7.54868537e-02
-8.77178133e-01 4.75947969e-02 4.93083924e-01 1.02635503e+00
7.84530997e-01 -1.37649089e-01 -5.02450347e-01 1.04177713e+00
4.44161117e-01 3.12781960e-01 7.27973580e-01 -7.47509301e-01
8.13331485e-01 1.10078382e+00 -2.59403229e-01 -1.65672219e+00
2.46807132e-02 1.84301749e-01 -7.76246130e-01 -3.51566404e-01
7.21041858e-02 5.39977610e-01 -7.79500842e-01 1.45188618e+00
2.39078224e-01 9.92161110e-02 -1.63592041e-01 9.93710637e-01
1.17859280e+00 8.65659833e-01 1.94776207e-01 1.01141252e-01
1.70949960e+00 -1.05775023e+00 -8.80513430e-01 2.33482104e-02
5.62678933e-01 -4.89320546e-01 1.06930101e+00 -1.81810901e-01
-1.01861048e+00 -6.46351755e-01 -8.29186141e-01 -5.94010711e-01
-7.47887909e-01 -1.16762459e-01 2.29721636e-01 3.23400199e-02
-9.83231604e-01 2.35385299e-01 -5.54776311e-01 -5.70328176e-01
2.88605630e-01 -4.23728786e-02 -6.16157830e-01 -4.68730032e-01
-1.38296151e+00 3.74312669e-01 5.87965548e-01 -1.60437390e-01
-6.78376198e-01 -5.68483412e-01 -1.21984673e+00 3.29743594e-01
3.74956310e-01 -8.37984204e-01 6.46306753e-01 -8.67821097e-01
-9.45636153e-01 7.97335625e-01 -1.80780932e-01 -2.58234479e-02
2.52162099e-01 2.18247087e-03 -4.03527379e-01 6.71894133e-01
2.88720936e-01 7.16294765e-01 1.21008003e+00 -1.22752094e+00
-7.24548697e-01 -5.66284239e-01 3.35557342e-01 5.33583164e-01
-8.56982291e-01 -1.15747601e-01 -9.90799606e-01 -9.66007054e-01
9.65950042e-02 -6.57858551e-01 2.64084697e-01 1.92361087e-01
-1.41181692e-01 -6.31043434e-01 1.02992606e+00 -8.47755075e-01
1.54747772e+00 -2.26340604e+00 6.25588477e-01 -1.90068260e-01
5.42309523e-01 -1.46496162e-01 -4.25755590e-01 7.91011810e-01
-1.10996723e-01 1.63501382e-01 3.10827792e-01 -3.40205044e-01
3.65885049e-01 1.94169775e-01 -1.79167360e-01 3.68886292e-01
7.15817958e-02 1.24287176e+00 -9.61543739e-01 -9.26086485e-01
3.21568877e-01 6.81527853e-01 -5.34195185e-01 3.15814108e-01
-1.38373330e-01 -4.40435000e-02 -8.27417135e-01 7.10878670e-01
3.98666590e-01 -5.41726589e-01 1.68273121e-01 -8.61773908e-01
1.76443741e-01 -1.26013741e-01 -1.00964451e+00 1.80383086e+00
-3.21930408e-01 8.04870725e-01 -5.03198337e-03 -1.13527310e+00
4.53506917e-01 2.54839987e-01 5.38160563e-01 -7.45341957e-01
6.14035167e-02 -2.63453543e-01 -4.92644787e-01 -8.07695150e-01
6.95980191e-01 1.19997494e-01 -3.99375975e-01 3.92638266e-01
1.90794155e-01 5.92401959e-02 3.55715483e-01 6.75447226e-01
9.24737930e-01 5.52972034e-02 3.16491634e-01 -7.49151111e-02
6.26928091e-01 -1.28277391e-01 2.74727404e-01 4.67336208e-01
-2.06801549e-01 5.96313775e-01 4.87012714e-01 -3.84308398e-01
-8.32243204e-01 -1.00283170e+00 1.93887144e-01 1.47907543e+00
5.51082432e-01 -8.91530335e-01 -5.10346770e-01 -7.41641045e-01
2.85966188e-01 1.84288502e-01 -8.11757445e-01 -1.65035605e-01
-3.82092327e-01 -2.66429126e-01 1.81057990e-01 5.70526958e-01
4.01107490e-01 -9.13612604e-01 -5.41167669e-02 -1.15586780e-01
-6.33867145e-01 -1.41276765e+00 -9.88636017e-01 -4.60960567e-01
-6.61403298e-01 -1.16195774e+00 -7.85877168e-01 -9.93909836e-01
5.96557736e-01 8.63905489e-01 1.08674228e+00 1.45792082e-01
-5.52660465e-01 1.06907248e+00 -7.25996673e-01 3.73619676e-01
-1.38142601e-01 -3.28096986e-01 -1.03391901e-01 3.99355263e-01
4.31045681e-01 -2.10019633e-01 -7.47768581e-01 2.54094422e-01
-1.36679089e+00 8.54575410e-02 4.86828953e-01 7.25271821e-01
5.63904166e-01 2.32699007e-01 -9.72587895e-03 -4.38947469e-01
6.38372242e-01 -6.05581403e-01 -1.80734605e-01 6.51536763e-01
-6.65640607e-02 1.32920057e-01 6.57363951e-01 -5.72125673e-01
-8.05169463e-01 -4.11236405e-01 5.41163862e-01 -9.50666547e-01
-1.71304271e-01 5.14126897e-01 -2.05980793e-01 1.60277292e-01
6.76417351e-02 5.46779454e-01 -3.18214148e-01 -3.17903727e-01
6.41795754e-01 6.48552299e-01 1.76135764e-01 -7.57106900e-01
7.73702621e-01 5.95312357e-01 -2.06311792e-01 -1.09819520e+00
-8.07806492e-01 -8.92835498e-01 -6.49764836e-01 -2.89809555e-01
1.37167978e+00 -1.14741337e+00 -6.04745865e-01 2.35463470e-01
-9.75153208e-01 -4.93270867e-02 4.45853424e-04 2.86592454e-01
-4.53038782e-01 9.41327989e-01 -7.98624337e-01 -2.52983093e-01
-8.15134794e-02 -1.08198464e+00 1.53398585e+00 -5.18566072e-02
5.50744385e-02 -1.35216260e+00 -1.03522398e-01 6.50116503e-01
-2.29080785e-02 8.81982446e-02 1.15041959e+00 -4.46256757e-01
-7.58224666e-01 -2.88278282e-01 -6.46140218e-01 8.43484700e-03
2.55094737e-01 1.79758564e-01 -4.41066861e-01 -4.94410872e-01
-3.36810172e-01 -4.86577332e-01 8.91801894e-01 1.20083779e-01
1.30538976e+00 -4.01431769e-01 -4.52780604e-01 3.36008251e-01
1.37124646e+00 -1.23719752e-01 4.76149440e-01 3.07608783e-01
1.31379378e+00 7.25702465e-01 4.62461561e-01 3.68149132e-01
6.94847941e-01 7.07206607e-01 3.72163713e-01 1.50594756e-01
-3.63686502e-01 -5.23917735e-01 5.00599146e-01 1.25920093e+00
2.81354934e-01 -2.85361409e-01 -6.90493941e-01 5.59209824e-01
-2.01797915e+00 -1.25331199e+00 1.70802604e-02 1.79431236e+00
6.08750820e-01 -2.66293943e-01 1.39371855e-02 -1.73222154e-01
1.04763317e+00 6.32952631e-01 -3.87527138e-01 1.51054701e-02
3.38319018e-02 -4.62494165e-01 -8.51856098e-02 2.78587282e-01
-9.83654916e-01 8.84932160e-01 5.34929943e+00 1.12018096e+00
-8.60379457e-01 -2.67867837e-02 7.45869353e-02 1.39589468e-02
-5.24235308e-01 -2.20536649e-01 -6.24440014e-01 4.12878066e-01
3.66766810e-01 -1.97079182e-01 2.71757007e-01 5.44525087e-01
-3.82992625e-02 3.26719791e-01 -1.21228278e+00 1.35817099e+00
6.08922720e-01 -1.31651235e+00 1.02343941e+00 -1.89672500e-01
7.52768040e-01 -4.09984916e-01 5.81713300e-03 5.55756390e-01
2.53991276e-01 -6.62380874e-01 6.12756550e-01 3.70395601e-01
6.96162224e-01 -5.10941088e-01 2.89670169e-01 3.91267687e-02
-1.90483570e+00 -1.38192594e-01 -3.20139647e-01 4.30584401e-01
1.92707386e-02 3.16788889e-02 -2.71424890e-01 8.23072553e-01
9.01000202e-01 1.35513544e+00 -7.61269152e-01 6.80248201e-01
1.90301035e-02 -9.99660566e-02 2.16510817e-01 -1.04798116e-01
4.73092109e-01 -3.89418155e-01 4.32607889e-01 1.42224872e+00
8.12970474e-02 1.15205057e-01 4.44853336e-01 5.18086791e-01
-2.40369245e-01 2.35836446e-01 -6.13349974e-01 -5.32429159e-01
3.72021109e-01 1.28259957e+00 -6.43959582e-01 -5.85171819e-01
-8.25258374e-01 1.17356265e+00 5.60049772e-01 6.39882565e-01
-9.09553885e-01 -4.21271086e-01 8.04389179e-01 -2.76747514e-02
4.84859854e-01 -2.28297606e-01 3.75662684e-01 -1.71394086e+00
1.95386494e-03 -9.25114870e-01 8.54489148e-01 -1.03920972e+00
-1.55387032e+00 3.27680260e-01 9.82669294e-02 -1.37700939e+00
1.20028563e-01 -6.82110012e-01 -1.81300491e-01 4.00903970e-01
-1.25751746e+00 -1.37660539e+00 -6.71309948e-01 1.03069723e+00
9.92708325e-01 9.93628204e-02 6.13450646e-01 3.61672401e-01
-5.58612525e-01 5.95396757e-01 1.58376679e-01 4.67985392e-01
7.04852581e-01 -1.14197993e+00 -6.75180852e-02 5.76143384e-01
4.43196446e-01 6.37246907e-01 4.14703310e-01 -5.22931576e-01
-1.82363784e+00 -1.17752135e+00 6.35198772e-01 -4.78063613e-01
1.23203695e+00 -4.47508067e-01 -1.09508479e+00 6.20840311e-01
2.90750861e-01 -3.43800671e-02 7.20011771e-01 -2.56007388e-02
-1.00138485e+00 -7.14611933e-02 -6.37514234e-01 9.72503603e-01
9.31245327e-01 -1.21044886e+00 -8.83903205e-01 2.53826916e-01
8.31273019e-01 2.70137452e-02 -1.12074125e+00 2.29799360e-01
6.29942656e-01 -9.32814002e-01 1.00500143e+00 -6.83370054e-01
5.36079526e-01 -2.37799123e-01 -4.27104831e-01 -1.06814098e+00
-5.50029576e-01 -2.21251011e-01 -2.61652440e-01 1.29024994e+00
-1.55506179e-01 -3.72883290e-01 5.29469788e-01 3.50923508e-01
1.09303802e-01 -4.67406482e-01 -6.67802453e-01 -5.56749284e-01
-2.16164052e-01 -2.35105336e-01 2.68291086e-01 1.41968811e+00
2.52805710e-01 5.01134455e-01 -1.84142962e-01 1.94764137e-01
6.59779668e-01 6.84158921e-01 6.95394337e-01 -1.15108800e+00
-1.95502162e-01 -7.49930978e-01 -8.55849266e-01 -1.25908077e+00
3.97096902e-01 -1.02047598e+00 -1.37012929e-01 -1.87930822e+00
7.46371150e-01 1.53174132e-01 -3.75889093e-01 2.54726738e-01
-3.69477153e-01 3.27434599e-01 3.09969574e-01 3.90393585e-01
-1.22938108e+00 7.46522069e-01 1.39915681e+00 -4.94810551e-01
2.01683402e-01 -8.42813671e-01 -4.37793285e-01 5.70989668e-01
3.49591732e-01 -1.99966624e-01 -4.30429131e-01 -6.94028974e-01
1.75958440e-01 2.51241535e-01 4.09142673e-01 -7.57278979e-01
3.06352079e-01 -7.50846714e-02 3.15611482e-01 -6.13441169e-01
3.75885487e-01 -1.02704716e+00 5.72760515e-02 2.31954738e-01
-4.89958882e-01 1.34376958e-01 2.51664664e-03 1.04359543e+00
-6.69055939e-01 1.52795121e-01 3.93574744e-01 -9.59757790e-02
-9.54538524e-01 7.51843274e-01 -2.55887508e-01 2.14508504e-01
9.62352753e-01 -4.50294882e-01 -3.18968505e-01 -4.49052691e-01
-7.30109274e-01 5.45534492e-01 5.01624286e-01 1.03667343e+00
7.99495757e-01 -1.80217457e+00 -4.72027093e-01 -2.98481830e-03
7.08850563e-01 -2.89427340e-01 6.14609241e-01 6.69950426e-01
-4.18298393e-01 5.77305973e-01 -1.42761588e-01 -7.54409134e-01
-1.52283156e+00 9.20884371e-01 3.83397862e-02 -1.32613435e-01
-6.94092870e-01 5.46246111e-01 9.38534439e-01 -1.09956041e-01
3.47694576e-01 -2.26259917e-01 -3.67498547e-01 5.54019988e-01
6.34446919e-01 3.33504558e-01 -4.10006106e-01 -1.00082004e+00
-1.87303931e-01 1.07553315e+00 -2.07170874e-01 2.33215839e-01
9.89814103e-01 -5.25027454e-01 -2.21105218e-01 6.07161582e-01
1.74541438e+00 -2.50396490e-01 -1.12543356e+00 -5.06123364e-01
-1.11448571e-01 -7.30638027e-01 2.17033029e-02 -4.95708250e-02
-9.64365602e-01 1.10909200e+00 1.74820259e-01 5.99499643e-01
9.74013031e-01 2.73197979e-01 8.64558399e-01 3.07796121e-01
1.49508074e-01 -9.25915778e-01 8.00783038e-01 3.17889601e-01
9.22098637e-01 -1.09149885e+00 -6.48753196e-02 -2.01922119e-01
-6.40372932e-01 1.21944356e+00 5.07868350e-01 -8.91065225e-02
4.31734949e-01 -3.86614799e-01 -1.97572932e-01 -4.10935670e-01
-6.90089762e-01 -4.13312674e-01 8.09523702e-01 3.44636440e-01
3.74218762e-01 -5.87567389e-02 -3.00793126e-02 1.70337230e-01
4.55221534e-01 -5.08067131e-01 9.95820761e-02 7.33421564e-01
-5.37354171e-01 -8.54650259e-01 -2.43528083e-01 3.06879371e-01
-4.44372207e-01 -9.50957462e-02 -4.41673189e-01 9.38345194e-01
-2.44389474e-01 7.48656154e-01 2.80986220e-01 -2.61915237e-01
2.74860680e-01 7.98761770e-02 4.87767994e-01 -5.63322425e-01
-1.76129445e-01 1.46126479e-01 -3.22935134e-01 -8.20765913e-01
-5.72265804e-01 -6.45550370e-01 -1.13223135e+00 -2.24286437e-01
-1.09859362e-01 2.86182642e-01 2.78204113e-01 7.63148487e-01
4.68102872e-01 5.45557737e-01 5.54838121e-01 -9.21163976e-01
-2.12972611e-01 -5.84487557e-01 -6.85827553e-01 1.11008132e+00
3.53167742e-01 -6.83768511e-01 -6.33729994e-01 2.08178848e-01] | [10.423310279846191, 1.0558546781539917] |
45099667-8816-4852-8eb8-af6e1c751059 | dm_nlp-at-semeval-2018-task-12-a-pipeline | null | null | https://aclanthology.org/S19-2156 | https://aclanthology.org/S19-2156.pdf | DM\_NLP at SemEval-2018 Task 12: A Pipeline System for Toponym Resolution | This paper describes DM-NLP{'}s system for toponym resolution task at Semeval 2019. Our system was developed for toponym detection, disambiguation and end-to-end resolution which is a pipeline of the former two. For toponym detection, we utilized the state-of-the-art sequence labeling model, namely, BiLSTM-CRF model as backbone. A lot of strategies are adopted for further improvement, such as pre-training, model ensemble, model averaging and data augment. For toponym disambiguation, we adopted the widely used searching and ranking framework. For ranking, we proposed several effective features for measuring the consistency between the detected toponym and toponyms in GeoNames. Eventually, our system achieved the best performance among all the submitted results in each sub task. | ['Chunping Ma', 'Pengjun Xie', 'Chu Liu', 'Luo Si', 'Xiaobin Wang', 'Linlin Li', 'Huafei Zheng'] | 2019-06-01 | null | null | null | semeval-2019-6 | ['toponym-resolution'] | ['natural-language-processing'] | [-1.28287673e-01 -2.19118685e-01 -5.93273975e-02 -5.14276862e-01
-9.82115209e-01 -4.44736093e-01 6.84538662e-01 4.42959785e-01
-9.97644603e-01 8.84907782e-01 1.75947919e-01 1.43806273e-02
-7.02802613e-02 -6.79221869e-01 -3.50008637e-01 -1.67780727e-01
1.23272426e-01 1.25704467e+00 2.38147184e-01 -7.32358158e-01
5.38154364e-01 2.82198668e-01 -1.43107724e+00 5.17468989e-01
1.21855521e+00 3.77111107e-01 3.55867207e-01 2.82516569e-01
-5.87792754e-01 2.54017383e-01 -6.42924547e-01 -6.79723322e-01
-1.94639906e-01 -1.07032964e-02 -8.81557226e-01 -9.52067614e-01
6.05986536e-01 7.77737200e-02 -8.97198990e-02 1.11543429e+00
8.14938724e-01 1.47204831e-01 5.69813728e-01 -7.16235042e-01
-5.22546291e-01 1.28731859e+00 -3.42026442e-01 2.29039833e-01
7.94996977e-01 -6.33217156e-01 1.12959909e+00 -1.30955589e+00
7.25445390e-01 1.32031190e+00 6.42096341e-01 5.02552211e-01
-6.18614614e-01 -8.84132326e-01 -1.20849699e-01 3.71916533e-01
-1.92839730e+00 -2.56582141e-01 -6.12351857e-02 -3.92855108e-01
1.59309304e+00 1.18884668e-01 2.81227708e-01 9.18581665e-01
-1.18419603e-01 4.18733180e-01 7.77370870e-01 -8.63090396e-01
-1.26943514e-01 -4.50865440e-02 4.54742998e-01 6.02455199e-01
2.61612087e-01 -2.57818043e-01 -6.63944960e-01 -5.22091091e-01
5.47835052e-01 -3.73365790e-01 1.35943368e-01 1.99981049e-01
-1.06367958e+00 3.56358767e-01 2.56752759e-01 5.56266487e-01
-3.90635580e-01 -2.15952471e-02 3.38642478e-01 1.68494090e-01
2.46089235e-01 8.20554972e-01 -8.18344593e-01 -2.09388882e-02
-1.08343196e+00 3.95999938e-01 6.24932349e-01 1.12799048e+00
6.83527410e-01 -5.53740203e-01 -2.61076033e-01 1.22838652e+00
5.46633124e-01 3.85092676e-01 9.34348941e-01 -6.58780098e-01
8.22869241e-01 6.12566292e-01 4.54363644e-01 -5.83454728e-01
-9.09093142e-01 -4.45379347e-01 -3.55246991e-01 -5.69807112e-01
1.57365277e-01 -8.14003963e-03 -1.04311228e+00 1.92085993e+00
2.56625377e-02 5.33951759e-01 2.02096179e-02 6.11949861e-01
1.10559762e+00 3.39480579e-01 5.78549385e-01 -9.01995376e-02
1.70564413e+00 -7.09970355e-01 -6.89054430e-01 -6.13821894e-02
1.05565524e+00 -1.19579268e+00 7.34709978e-01 2.57829100e-01
-9.05219615e-01 -7.84425139e-01 -9.72261727e-01 -9.85669419e-02
-8.03100586e-01 2.66423941e-01 4.23756182e-01 6.21387661e-01
-9.38657284e-01 8.96231055e-01 -4.14058149e-01 -6.87719107e-01
-3.15653741e-01 4.59004909e-01 -4.09952790e-01 -4.05381527e-03
-2.14491487e+00 1.28898966e+00 9.34163809e-01 -1.07428925e-02
-3.15359414e-01 -5.30253470e-01 -6.19246304e-01 1.32255051e-02
-1.78616673e-01 -1.06174445e+00 1.14117837e+00 -2.84310639e-01
-1.11210084e+00 1.28371966e+00 -1.12533964e-01 -4.71723437e-01
1.17499024e-01 -7.80710876e-01 -1.10639071e+00 -4.40834880e-01
4.99834150e-01 7.49728739e-01 1.89326808e-01 -6.61310434e-01
-8.44536722e-01 -2.91029632e-01 -4.15587813e-01 5.59035182e-01
8.24968738e-04 5.75002432e-01 -6.75623357e-01 -4.30595130e-01
1.30585387e-01 -9.10948157e-01 -4.82644364e-02 -1.07226908e+00
-2.19900876e-01 -7.48884737e-01 -3.17461714e-02 -9.80849683e-01
1.81576371e+00 -1.94987714e+00 -2.67756641e-01 2.91296482e-01
7.56792948e-02 7.37327814e-01 -1.96031421e-01 9.18833792e-01
-1.64826185e-01 1.79320589e-01 7.34455884e-02 -4.79944348e-01
9.31335688e-02 5.83197400e-02 -2.73709357e-01 -4.26440462e-02
-1.36172190e-01 6.21043622e-01 -1.09821057e+00 -7.95563757e-01
2.56061047e-01 4.92755860e-01 -1.96977377e-01 3.05748284e-02
-2.51338303e-01 1.30378053e-01 -4.39166814e-01 4.57560837e-01
8.07151198e-01 5.03530502e-02 3.94079685e-01 1.72672436e-01
-1.21911965e-01 6.41212761e-01 -1.36646199e+00 2.28831792e+00
-5.65489113e-01 -7.11882561e-02 -5.52268505e-01 -2.49701604e-01
1.19528127e+00 5.00743151e-01 2.93669254e-01 -7.31227994e-01
-1.86956733e-01 1.00368488e+00 -2.02492759e-01 -5.90976417e-01
1.05721378e+00 2.30373934e-01 -4.05036777e-01 -1.87364578e-01
1.56612560e-01 3.73755366e-01 3.93956959e-01 2.46491268e-01
8.04155171e-01 5.86108327e-01 6.00833058e-01 -1.54473633e-01
6.64193392e-01 1.74707800e-01 8.01290810e-01 9.41394985e-01
1.36493240e-02 3.76036257e-01 -1.47970051e-01 -5.22818863e-01
-7.86121547e-01 -8.84278476e-01 -3.08923721e-01 1.31518650e+00
1.80724114e-01 -7.81813741e-01 -6.58828914e-01 -3.91632497e-01
-1.22151300e-01 8.21521640e-01 -5.18883802e-02 1.63846180e-01
-9.66874719e-01 -8.59856606e-01 1.12851346e+00 1.98673204e-01
5.42454183e-01 -1.09782219e+00 -1.54197976e-01 5.79882562e-01
-5.34763575e-01 -1.26156759e+00 -3.46964538e-01 2.81003773e-01
-5.16128659e-01 -9.35328841e-01 -5.91648281e-01 -9.89261687e-01
1.07849918e-01 -8.41482449e-03 1.49095380e+00 3.68779674e-02
-1.83764160e-01 -3.92428786e-01 -4.34198558e-01 -1.28302485e-01
-9.45678502e-02 3.74631345e-01 3.10203522e-01 -5.04099131e-01
7.70117402e-01 -5.56172907e-01 -3.60849500e-01 9.18635651e-02
-2.88700104e-01 -3.50182027e-01 6.57301128e-01 7.64541328e-01
4.75266695e-01 -4.12564784e-01 6.86366498e-01 -1.24905741e+00
6.08721972e-01 -2.84739733e-01 -3.72256875e-01 5.99997044e-01
-8.17612529e-01 3.31025779e-01 6.32197917e-01 1.39843285e-01
-1.23788929e+00 2.38154978e-01 -4.97630954e-01 1.64004892e-01
-2.36188024e-01 6.69563949e-01 -2.79086202e-01 2.14995414e-01
8.07175159e-01 7.32386261e-02 -1.07189202e+00 -9.84698296e-01
8.91548872e-01 1.01258302e+00 7.27734387e-01 -6.04687154e-01
2.36595854e-01 -3.06029588e-01 -3.48709375e-01 -5.50232232e-01
-8.21258187e-01 -8.65926743e-01 -8.65066409e-01 3.58893991e-01
7.50577331e-01 -1.42151392e+00 -4.41330045e-01 4.53286409e-01
-1.60521722e+00 3.79595637e-01 3.54024291e-01 6.08712792e-01
-7.73237720e-02 5.76247692e-01 -5.57865024e-01 -4.42993820e-01
-7.64716327e-01 -6.32173121e-01 1.05322838e+00 3.86112571e-01
-5.05746663e-01 -6.94834948e-01 7.45133281e-01 3.21823567e-01
4.00696158e-01 -1.72594160e-01 1.01729512e+00 -1.18282676e+00
-7.62887001e-02 -1.67364106e-01 -1.61539957e-01 -3.20090324e-01
-1.78894803e-01 7.86795169e-02 -1.01178443e+00 2.53397357e-02
-8.41238678e-01 -1.44747466e-01 1.16131151e+00 1.33914724e-01
6.43164456e-01 2.35883534e-01 -6.51828766e-01 5.84133267e-01
1.42263651e+00 1.41347364e-01 5.05240619e-01 5.48012316e-01
7.14810431e-01 3.02730024e-01 9.91775036e-01 5.10594726e-01
5.80047011e-01 1.26794088e+00 1.27120927e-01 1.89177603e-01
1.12150490e-01 -5.04784286e-01 3.76346149e-02 8.33913743e-01
4.40426208e-02 -3.41244936e-01 -1.15110195e+00 4.94709074e-01
-2.09962296e+00 -8.90103221e-01 -3.34694177e-01 2.33507776e+00
8.12581956e-01 3.99731211e-02 3.50409225e-02 -3.31989557e-01
1.26386011e+00 -9.29153711e-02 6.32094070e-02 -6.01551235e-01
-2.29048893e-01 3.17437917e-01 4.67384011e-01 6.72883570e-01
-1.28560519e+00 1.88123369e+00 6.19963169e+00 1.03637087e+00
-6.96606755e-01 1.25577236e-02 -1.20575160e-01 3.59934539e-01
-1.88254565e-01 2.76920795e-01 -1.39151645e+00 6.32451415e-01
9.80929017e-01 1.27832085e-01 1.95731983e-01 8.01915586e-01
-1.45552278e-01 1.39534786e-01 -7.71285951e-01 1.07151246e+00
-6.85259327e-02 -9.09308255e-01 2.09740743e-01 -4.46127921e-01
4.29771006e-01 3.78455490e-01 -4.75306004e-01 6.25119328e-01
5.93494415e-01 -8.87327075e-01 2.86093563e-01 5.15229225e-01
7.48687327e-01 -8.79828513e-01 9.29671228e-01 3.03925991e-01
-1.30660129e+00 1.70057893e-01 -6.00164175e-01 1.97786968e-02
2.65598238e-01 8.95658016e-01 -9.43169892e-01 1.00413477e+00
4.08528924e-01 6.85669124e-01 -6.34814620e-01 1.31376135e+00
-5.21443069e-01 1.13106176e-01 -3.97065490e-01 1.37438685e-01
-2.26026773e-02 -6.24124706e-02 2.99779594e-01 1.76564682e+00
4.47809905e-01 -2.61145294e-01 3.07308137e-01 2.45674595e-01
-1.49549678e-01 4.57865864e-01 -3.02800626e-01 2.77191728e-01
1.47566986e+00 1.36768675e+00 -4.31833357e-01 -6.40263319e-01
-2.14672044e-01 8.89351010e-01 5.78953028e-01 4.93304841e-02
-8.46501648e-01 -8.17098141e-01 3.95105958e-01 -4.16250467e-01
-6.03278987e-02 -3.16007659e-02 -1.15316324e-01 -1.18073821e+00
-1.47888184e-01 -5.81144512e-01 1.04775512e+00 -6.69175506e-01
-1.09337032e+00 7.12868690e-01 3.46080102e-02 -1.07077122e+00
-3.85785103e-01 -4.03781235e-01 -3.93007547e-01 1.05556774e+00
-1.48606896e+00 -1.26695728e+00 -1.69262528e-01 4.00666624e-01
2.20528811e-01 -2.22769469e-01 1.21187842e+00 1.00402296e+00
-6.40374541e-01 7.56738484e-01 6.85287938e-02 2.24284038e-01
1.42223394e+00 -1.43543386e+00 9.34522390e-01 8.78546536e-01
4.19777691e-01 1.04013288e+00 5.36483049e-01 -9.91810918e-01
-6.33971274e-01 -9.10333872e-01 1.95576620e+00 -4.75884497e-01
5.64005971e-01 -1.34916618e-01 -6.20710671e-01 6.05192065e-01
3.21255587e-02 -4.52672035e-01 6.71663821e-01 3.91878903e-01
-5.72969913e-01 8.68115872e-02 -8.13120365e-01 2.95802206e-01
1.26501524e+00 -5.00538111e-01 -1.07623518e+00 2.37494081e-01
7.73766696e-01 -5.37152529e-01 -1.05283225e+00 3.10386211e-01
5.66436350e-01 -3.64819795e-01 1.21868062e+00 -5.64200044e-01
8.98057744e-02 -4.38233018e-01 -3.07095408e-01 -1.28659558e+00
-6.18028820e-01 -4.96865094e-01 8.28280225e-02 1.64291215e+00
7.42886841e-01 -4.78827655e-01 2.28421301e-01 2.25007728e-01
-4.24094558e-01 -2.05342144e-01 -1.03192234e+00 -6.75494730e-01
-5.56856431e-02 -3.48017849e-02 8.27063739e-01 1.18953502e+00
1.63492560e-01 6.52939558e-01 -5.70185900e-01 2.71134138e-01
3.47218692e-01 5.47234751e-02 2.24736348e-01 -1.37198663e+00
-1.43611118e-01 -2.22610340e-01 -3.47596735e-01 -9.08926308e-01
1.86423406e-01 -1.15881205e+00 -5.77246472e-02 -1.59005177e+00
1.50111064e-01 -4.50758487e-01 -8.94114912e-01 4.82822776e-01
-5.80574989e-01 6.25416487e-02 3.36383171e-02 3.51841867e-01
-8.80840063e-01 1.51094049e-01 6.06206179e-01 4.02701557e-01
-3.14608157e-01 -1.32647291e-01 -6.20146155e-01 5.75777531e-01
6.02006972e-01 -8.64444613e-01 1.21973164e-01 -7.62817979e-01
8.14047873e-01 1.01892278e-01 -1.08195335e-01 -9.30595279e-01
3.44165534e-01 1.86345145e-01 3.71790558e-01 -9.52250063e-01
3.36008161e-01 -3.20642889e-01 9.11974460e-02 5.93283892e-01
-3.48514795e-01 5.74744284e-01 -1.30456269e-01 2.27012813e-01
-3.45767766e-01 -3.00795913e-01 6.93538308e-01 -3.17538559e-01
-1.16665924e+00 -4.99481056e-03 -1.83447689e-01 3.29067022e-01
3.76757681e-01 3.51845652e-01 -7.17953682e-01 2.53904372e-01
-7.76279807e-01 4.24526364e-01 2.78147489e-01 3.99575263e-01
3.49326879e-01 -1.34204137e+00 -1.07217669e+00 -2.24020779e-01
4.58094746e-01 -4.53507751e-01 1.76379874e-01 4.33444202e-01
-6.16721928e-01 7.03121781e-01 -3.17288667e-01 -1.88563421e-01
-1.33228099e+00 3.21591407e-01 2.84709215e-01 -7.69046962e-01
-3.20249945e-01 1.11777687e+00 -3.81740034e-01 -8.17932487e-01
1.88730106e-01 -7.37814680e-02 -8.46027911e-01 2.68313885e-01
4.92485881e-01 4.71627980e-01 3.31199378e-01 -6.03810489e-01
-9.67267931e-01 8.55527282e-01 -4.76878911e-01 -1.98113084e-01
1.05090916e+00 -2.18709767e-01 -2.52488643e-01 5.74716739e-02
6.54056609e-01 -3.90293822e-02 -1.21335790e-01 -4.72166389e-01
7.90562212e-01 -1.78368360e-01 -2.71221191e-01 -1.10525084e+00
-2.09227592e-01 7.17888832e-01 7.28576958e-01 -2.40163520e-01
7.98394203e-01 -2.94917583e-01 9.21581089e-01 7.35273302e-01
5.59221029e-01 -1.45699430e+00 -5.70032954e-01 1.18256462e+00
2.46135116e-01 -1.03878379e+00 5.01841540e-03 -3.17089111e-01
-5.12101769e-01 9.25164998e-01 6.40897930e-01 -2.39748042e-02
4.44582671e-01 -1.94621071e-01 2.51491934e-01 -1.34344369e-01
-6.71150684e-01 -3.34784359e-01 3.45840633e-01 3.09665889e-01
9.56155300e-01 2.75370806e-01 -9.76598561e-01 8.44667017e-01
-3.85416895e-01 4.58853468e-02 2.75175393e-01 3.66732240e-01
-8.00439179e-01 -1.57070279e+00 -1.19463235e-01 1.15580551e-01
-5.91453433e-01 -8.42851877e-01 -4.88407284e-01 6.87848628e-01
4.51768488e-01 8.37259769e-01 -3.64198178e-01 -5.53578258e-01
3.74642789e-01 4.76618260e-01 4.08804029e-01 -9.33231235e-01
-1.00210881e+00 -6.56913817e-02 5.69095016e-01 -6.72044575e-01
-3.27963471e-01 -5.18202007e-01 -1.51737344e+00 4.42344928e-03
-4.18513447e-01 4.69068855e-01 7.10718215e-01 1.19571078e+00
4.87467647e-01 2.62283057e-01 4.34086740e-01 -3.23058248e-01
-4.04350221e-01 -1.22055447e+00 -4.65891987e-01 3.67741078e-01
-3.05226833e-01 -5.30898750e-01 2.09225133e-01 -4.36757773e-01] | [9.505732536315918, 9.154616355895996] |
a1b13a33-66ce-4a88-86ad-f10e7bdc4d25 | sac-gan-structure-aware-image-to-image | 2112.06596 | null | https://arxiv.org/abs/2112.06596v5 | https://arxiv.org/pdf/2112.06596v5.pdf | SAC-GAN: Structure-Aware Image Composition | We introduce an end-to-end learning framework for image-to-image composition, aiming to plausibly compose an object represented as a cropped patch from an object image into a background scene image. As our approach emphasizes more on semantic and structural coherence of the composed images, rather than their pixel-level RGB accuracies, we tailor the input and output of our network with structure-aware features and design our network losses accordingly, with ground truth established in a self-supervised setting through the object cropping. Specifically, our network takes the semantic layout features from the input scene image, features encoded from the edges and silhouette in the input object patch, as well as a latent code as inputs, and generates a 2D spatial affine transform defining the translation and scaling of the object patch. The learned parameters are further fed into a differentiable spatial transformer network to transform the object patch into the target image, where our model is trained adversarially using an affine transform discriminator and a layout discriminator. We evaluate our network, coined SAC-GAN, for various image composition scenarios in terms of quality, composability, and generalizability of the composite images. Comparisons are made to state-of-the-art alternatives, including Instance Insertion, ST-GAN, CompGAN and PlaceNet, confirming superiority of our method. | ['Ling-Xiao Zhang', 'Ali Mahdavi-Amiri', 'Lin Gao', 'Rui Ma', 'Hao Zhang', 'Hang Zhou'] | 2021-12-13 | null | null | null | null | ['image-augmentation'] | ['computer-vision'] | [ 9.12291765e-01 3.77431393e-01 2.03874320e-01 -2.93464750e-01
-6.95233226e-01 -9.60983336e-01 7.25489736e-01 -2.80260146e-01
-7.41120204e-02 4.13516760e-01 5.65956868e-02 -4.45940159e-02
7.24637583e-02 -1.06247663e+00 -1.37946844e+00 -8.30052912e-01
3.15975726e-01 2.83313304e-01 1.81334857e-02 -1.16346262e-01
1.11570589e-01 5.83145738e-01 -1.31603742e+00 3.55751544e-01
8.50837588e-01 1.36211801e+00 2.18995959e-01 8.40862393e-01
8.42561200e-02 7.39151835e-01 -5.11608839e-01 -7.71022916e-01
7.02143013e-01 -7.56319404e-01 -5.50556719e-01 5.10316312e-01
9.12203968e-01 -2.50880808e-01 -3.87356937e-01 9.36768830e-01
1.70924217e-01 -1.71253234e-01 6.32911861e-01 -1.33292341e+00
-1.09810138e+00 5.56208253e-01 -4.24818426e-01 -5.74147522e-01
1.93335772e-01 6.03601635e-01 1.10736310e+00 -6.55739963e-01
8.42133045e-01 1.06903374e+00 5.96200347e-01 3.64368945e-01
-1.54162419e+00 -5.64220846e-01 8.82169232e-02 -1.27127707e-01
-1.04429173e+00 -4.13691461e-01 9.98344541e-01 -4.73891526e-01
4.80057895e-01 3.12894523e-01 8.35595012e-01 1.15883338e+00
9.03759748e-02 4.51939225e-01 1.05948436e+00 -4.92251545e-01
2.69585133e-01 1.63238123e-01 -7.65978515e-01 6.84372365e-01
1.30163161e-02 3.06950957e-01 -3.22973818e-01 2.70360887e-01
1.07974112e+00 4.43997085e-02 -2.80690581e-01 -7.80673265e-01
-1.30267155e+00 5.81761181e-01 1.00505888e+00 6.19641840e-02
-5.23470044e-01 6.20430112e-01 -1.50386930e-01 1.35002151e-01
3.40447247e-01 6.10636353e-01 -2.54768461e-01 2.96178818e-01
-9.04630661e-01 4.06826973e-01 4.90855038e-01 1.19115984e+00
9.67756212e-01 2.52869874e-01 -3.82320166e-01 5.02363920e-01
1.99131179e-03 7.27705717e-01 -2.83171260e-03 -1.08566296e+00
3.95665348e-01 6.89706206e-01 -7.35716149e-02 -8.83613348e-01
1.27091452e-01 -5.85759282e-01 -9.04018521e-01 4.52043504e-01
2.49439910e-01 1.21107008e-02 -1.06242192e+00 1.76108134e+00
2.26544157e-01 9.81948674e-02 -6.04620576e-02 9.30396080e-01
4.54358935e-01 5.02269387e-01 3.12184785e-02 3.11794043e-01
1.04954231e+00 -1.25359976e+00 -3.14416766e-01 -2.97582239e-01
7.30003566e-02 -7.16148078e-01 1.12408435e+00 2.22117960e-01
-1.43920374e+00 -8.04706097e-01 -9.56776202e-01 -2.41504535e-01
-2.17531174e-01 3.44275385e-01 2.73334980e-01 4.38723832e-01
-1.12161279e+00 5.91474891e-01 -3.92951071e-01 -4.08593938e-02
6.09459043e-01 2.23941848e-01 -3.90692949e-01 1.69703718e-02
-6.81432664e-01 6.64918363e-01 3.90563160e-01 1.20220771e-02
-1.29228210e+00 -1.02988899e+00 -9.13522124e-01 9.68159512e-02
1.57737002e-01 -1.10027981e+00 9.31017399e-01 -1.63625252e+00
-1.53188062e+00 9.82118964e-01 2.58218348e-01 -3.38367432e-01
7.61544943e-01 2.63595194e-01 -1.24958366e-01 4.43221964e-02
1.48486793e-01 8.94316137e-01 1.28485453e+00 -1.74011552e+00
-7.64964402e-01 -9.14648473e-02 3.23996663e-01 1.14157669e-01
5.79915829e-02 -4.55457211e-01 -4.55304921e-01 -9.36798692e-01
-8.73391051e-03 -8.74468267e-01 -3.81850405e-03 4.78379875e-01
-7.21855760e-01 6.48165762e-01 9.59490597e-01 -8.27342212e-01
6.59417570e-01 -2.49433446e+00 5.27610242e-01 3.37860614e-01
1.88052893e-01 -1.24400713e-01 -5.76342344e-01 3.53556871e-01
-2.27914616e-01 2.62634153e-03 -6.60021782e-01 -4.82591718e-01
7.15050399e-02 7.02033192e-02 -4.72885698e-01 5.12841523e-01
5.79021811e-01 1.46865523e+00 -9.56476510e-01 -2.29526330e-02
4.57412511e-01 6.29022658e-01 -8.09365094e-01 5.79520524e-01
-6.07144594e-01 7.43131101e-01 -9.74898338e-02 4.56903219e-01
6.02856457e-01 -1.03004768e-01 2.34850273e-02 -4.72527444e-01
1.31649509e-01 6.70007989e-02 -8.10516715e-01 1.92064881e+00
-6.69979215e-01 5.76319039e-01 -4.36327048e-02 -9.62444067e-01
9.10030186e-01 -4.60384637e-02 2.41838396e-01 -8.79915535e-01
-4.52298597e-02 9.99370813e-02 -1.48589149e-01 -3.37694257e-01
4.01591241e-01 -7.88652822e-02 -1.21105857e-01 2.18782544e-01
1.56134143e-01 -5.85527182e-01 -2.42983967e-01 5.81058078e-02
8.59789431e-01 5.36355913e-01 -5.93890138e-02 -1.05696552e-01
3.92761409e-01 -4.39617299e-02 4.46890201e-03 4.54094768e-01
4.66239840e-01 1.13174450e+00 6.29934430e-01 -3.11154157e-01
-1.71828818e+00 -1.44738185e+00 1.69473261e-01 8.42051208e-01
2.59896368e-01 2.38666311e-03 -1.11717308e+00 -7.03967452e-01
-1.09372092e-02 7.76528418e-01 -8.47730458e-01 -2.71657467e-01
-6.51732743e-01 5.13910055e-02 4.77624536e-01 4.07911927e-01
8.48527610e-01 -1.07348907e+00 -5.33571243e-01 1.03124321e-01
-8.14834684e-02 -1.22741747e+00 -7.38578677e-01 1.46017056e-02
-5.36473393e-01 -1.03566718e+00 -6.22934878e-01 -8.52628767e-01
1.05980742e+00 -2.03248523e-02 1.11983812e+00 7.41522834e-02
-4.16386217e-01 2.05504343e-01 -2.07627922e-01 3.86227667e-02
-5.17602921e-01 -1.01596221e-01 -3.99693042e-01 6.18889570e-01
-6.02616489e-01 -7.86399543e-01 -8.70021462e-01 9.78677124e-02
-1.33576345e+00 6.47711515e-01 8.10025930e-01 8.85919392e-01
6.14607573e-01 1.64680257e-01 -9.76800025e-02 -9.69487965e-01
8.46534520e-02 -1.91128194e-01 -6.17031872e-01 3.47373813e-01
-3.09154510e-01 1.01029247e-01 7.64863193e-01 -4.22363609e-01
-9.61666286e-01 2.16600418e-01 -2.44256984e-02 -5.93527734e-01
-8.27707648e-02 3.53742242e-02 -6.26362205e-01 -2.32017666e-01
6.92271411e-01 5.03111541e-01 -1.46208897e-01 -1.81040943e-01
9.58015323e-01 2.27235660e-01 1.08476400e+00 -6.78197384e-01
1.23060632e+00 6.16914451e-01 2.37722639e-02 -3.46101493e-01
-6.84875548e-01 1.97312370e-01 -7.17085421e-01 -5.93031645e-02
9.91429865e-01 -8.69262338e-01 -4.97912437e-01 6.82237923e-01
-1.19997215e+00 -6.64577305e-01 -7.95659304e-01 -2.16755465e-01
-8.25748801e-01 -1.98507115e-01 -3.19509804e-01 -3.66531849e-01
-2.08687261e-01 -1.20656812e+00 1.50748575e+00 -2.60373652e-02
9.01425704e-02 -9.75041926e-01 -3.97751719e-01 3.71747613e-01
5.97936153e-01 7.60964811e-01 1.25061083e+00 2.38848981e-02
-1.21011364e+00 -5.34986742e-02 -4.32883918e-01 5.45958579e-01
1.56135872e-01 1.35605231e-01 -1.01853788e+00 -2.16987669e-01
-2.17154294e-01 -1.89442322e-01 5.60176790e-01 2.40405202e-01
1.28041303e+00 -8.14256608e-01 -3.77738848e-02 1.15224683e+00
1.72121370e+00 -1.58175416e-02 8.73152852e-01 1.91929907e-01
1.04621196e+00 5.99326551e-01 -8.58215801e-03 1.02755822e-01
2.41434067e-01 7.45187640e-01 7.23739862e-01 -4.47563201e-01
-5.92972636e-01 -8.09064746e-01 2.60400593e-01 2.35330895e-01
1.20062932e-01 -3.24883312e-01 -4.17719156e-01 3.58971983e-01
-1.47905242e+00 -7.54650712e-01 2.28400677e-01 1.93867421e+00
7.55798221e-01 -6.06766865e-02 -6.30712807e-02 4.48119864e-02
6.01163507e-01 1.58260420e-01 -6.59126401e-01 -1.04881205e-01
-3.63220513e-01 4.87779588e-01 6.33366764e-01 4.65821475e-01
-8.37557077e-01 9.64781821e-01 5.62322378e+00 7.37100005e-01
-1.64906812e+00 1.06624961e-01 9.53018010e-01 8.77976939e-02
-7.81193256e-01 1.32654443e-01 -1.53192297e-01 4.07035649e-01
4.63769406e-01 9.04818550e-02 9.39106524e-01 6.62205935e-01
-1.00674540e-01 1.73187405e-01 -1.26343596e+00 9.09543991e-01
1.53629854e-01 -1.65447199e+00 3.49339604e-01 2.72107571e-02
1.12599814e+00 -3.59481394e-01 4.69932050e-01 -1.16676033e-01
2.31838897e-01 -1.21501207e+00 1.44658804e+00 4.90644902e-01
1.41629291e+00 -4.46914256e-01 1.82867274e-01 -3.61295678e-02
-9.93685722e-01 -4.72636037e-02 1.11106120e-01 1.11131363e-01
4.86447997e-02 3.97007257e-01 -5.73669076e-01 5.68669915e-01
4.52828556e-01 6.21676981e-01 -5.91226578e-01 6.43878162e-01
-5.44015408e-01 2.89444149e-01 -1.11393044e-02 5.09901702e-01
2.30580568e-01 -2.87676603e-01 3.68399650e-01 9.07802761e-01
3.24336410e-01 -2.07804948e-01 -1.55520841e-01 1.72524548e+00
-4.74878520e-01 -2.63195664e-01 -5.58493435e-01 -9.75079089e-02
3.08449447e-01 1.25901341e+00 -5.89029074e-01 -2.88807750e-01
-1.35966474e-02 1.45184755e+00 1.94697067e-01 5.55248797e-01
-9.84737217e-01 -2.79069692e-01 5.22513449e-01 3.40230525e-01
6.34183288e-01 4.12703380e-02 -6.37781560e-01 -1.02396321e+00
3.64483625e-01 -9.68378484e-01 -3.45070720e-01 -1.12178802e+00
-1.16357768e+00 7.61860907e-01 -1.83564842e-01 -1.13391364e+00
-1.21979952e-01 -4.66986328e-01 -7.06752777e-01 9.58493590e-01
-1.29817820e+00 -1.80777287e+00 -5.75190842e-01 3.75171542e-01
3.09027284e-01 -2.17487946e-01 6.90254867e-01 2.29104951e-01
-1.07664593e-01 7.18922734e-01 -8.10773969e-02 3.36863726e-01
3.24399978e-01 -1.15734386e+00 7.85977662e-01 9.02241766e-01
1.11232080e-01 2.79843360e-01 4.59294587e-01 -3.79064262e-01
-1.65624070e+00 -1.61030507e+00 3.18788111e-01 -4.68525141e-01
4.28314328e-01 -7.94803560e-01 -4.38010812e-01 6.93131745e-01
2.85816967e-01 2.83111066e-01 -8.44512507e-02 -7.45164514e-01
-6.47743225e-01 -2.56041765e-01 -1.24801779e+00 6.13874257e-01
1.30136085e+00 -6.50210798e-01 -1.42461449e-01 1.00166835e-01
1.08979857e+00 -7.00441778e-01 -7.12461472e-01 3.78202528e-01
6.45975590e-01 -9.93513048e-01 1.06164193e+00 -3.52804184e-01
1.05177569e+00 -4.82649803e-01 -2.94665426e-01 -1.41552484e+00
-3.71021092e-01 -4.95992094e-01 3.79270375e-01 1.09587717e+00
4.72883672e-01 -3.85492653e-01 6.48962438e-01 4.10611689e-01
-1.96112350e-01 -6.37475908e-01 -6.94258809e-01 -4.23771858e-01
5.75618707e-02 -2.54546583e-01 1.02220619e+00 8.06443453e-01
-8.30649436e-01 2.50188559e-01 -3.13332140e-01 1.67366341e-01
7.02232182e-01 2.67363250e-01 9.95737433e-01 -5.61746299e-01
-5.91801405e-01 -6.11373186e-01 -4.58254218e-01 -9.47795928e-01
7.24583492e-02 -8.01939189e-01 -6.20162161e-03 -1.21568251e+00
-9.36371181e-03 -7.17075765e-01 2.04247490e-01 4.65385228e-01
2.26196684e-02 7.60391831e-01 4.48849022e-01 5.05401231e-02
-1.81949973e-01 6.18909955e-01 1.69723880e+00 -4.32045341e-01
7.14373291e-02 -3.62560391e-01 -7.59844005e-01 2.76619703e-01
4.08216953e-01 -1.97257325e-01 -3.00426215e-01 -6.81704223e-01
1.52627066e-01 1.66873895e-02 8.59203815e-01 -1.04795086e+00
-1.12168631e-02 -2.36948371e-01 6.56060934e-01 3.10359579e-02
3.09134036e-01 -1.04946637e+00 7.43591964e-01 2.95346916e-01
-6.45556152e-01 -2.20742002e-01 -2.18178332e-02 3.11069489e-01
-6.71370775e-02 7.88394138e-02 7.82106519e-01 2.12007575e-02
-4.31950837e-01 4.72356290e-01 5.96935570e-01 -1.07817547e-02
1.05897868e+00 -4.14871633e-01 -3.61105204e-01 -3.50365579e-01
-4.94808108e-01 -3.42990547e-01 1.09284413e+00 5.15038013e-01
5.68530798e-01 -1.43578672e+00 -7.72085905e-01 7.61354268e-01
2.34100789e-01 3.93323094e-01 2.21744895e-01 4.39788431e-01
-8.47137094e-01 1.58618227e-01 -4.84236717e-01 -7.17114627e-01
-6.42647088e-01 6.10514402e-01 5.08783519e-01 1.19681813e-01
-5.74818909e-01 6.65754676e-01 7.50942945e-01 -6.02376521e-01
1.83399945e-01 -4.09592301e-01 6.15128160e-01 -5.33372104e-01
1.00474179e-01 -2.22996593e-01 5.99022664e-04 -6.35911703e-01
1.49479792e-01 6.54446542e-01 4.82443988e-01 -2.27472454e-01
1.28729630e+00 4.47317995e-02 -3.69517416e-01 9.11939889e-02
1.42410290e+00 4.58960086e-02 -1.72985125e+00 -2.09060371e-01
-4.67357606e-01 -6.24922574e-01 -9.41446945e-02 -9.49373364e-01
-1.51772559e+00 6.24849796e-01 4.81116414e-01 -5.20129092e-02
1.24933326e+00 5.03047332e-02 6.89640462e-01 -1.87204570e-01
3.33650738e-01 -4.97717202e-01 3.45004618e-01 8.28971267e-02
1.27249706e+00 -8.94903004e-01 -3.47388327e-01 -4.03225422e-01
-5.85782468e-01 8.56778979e-01 4.81078923e-01 -5.56104720e-01
3.05049300e-01 3.07829887e-01 -7.59571493e-02 5.86902862e-03
-4.29407924e-01 2.89688334e-02 6.02495790e-01 7.09269941e-01
1.05087325e-01 1.97484404e-01 2.88484573e-01 1.72149509e-01
-6.19583905e-01 -1.93745643e-01 1.19034268e-01 5.05323589e-01
1.42635047e-01 -9.87768471e-01 -2.86475718e-01 1.46450877e-01
-1.93163961e-01 -1.78043231e-01 -3.30763459e-01 5.13442039e-01
6.49131477e-01 3.07421654e-01 3.89872164e-01 -4.41984922e-01
4.93228495e-01 -3.25263917e-01 7.39443958e-01 -4.37646955e-01
-7.05613136e-01 -1.33451045e-01 -1.96826801e-01 -5.96277058e-01
-2.71680355e-01 -2.43618563e-01 -9.25019205e-01 -2.04225138e-01
6.28858805e-02 -3.44686925e-01 9.86046672e-01 6.88616812e-01
3.58433723e-01 8.22941065e-01 8.45606923e-01 -1.08543241e+00
-3.29297155e-01 -7.20292091e-01 -3.75540316e-01 8.91335428e-01
5.45151293e-01 -2.87351459e-01 -3.21716011e-01 5.19434571e-01] | [11.59553050994873, -0.4958001673221588] |
dc487793-677b-4973-808c-9bb56808ac89 | ow-detr-open-world-detection-transformer | 2112.01513 | null | https://arxiv.org/abs/2112.01513v3 | https://arxiv.org/pdf/2112.01513v3.pdf | OW-DETR: Open-world Detection Transformer | Open-world object detection (OWOD) is a challenging computer vision problem, where the task is to detect a known set of object categories while simultaneously identifying unknown objects. Additionally, the model must incrementally learn new classes that become known in the next training episodes. Distinct from standard object detection, the OWOD setting poses significant challenges for generating quality candidate proposals on potentially unknown objects, separating the unknown objects from the background and detecting diverse unknown objects. Here, we introduce a novel end-to-end transformer-based framework, OW-DETR, for open-world object detection. The proposed OW-DETR comprises three dedicated components namely, attention-driven pseudo-labeling, novelty classification and objectness scoring to explicitly address the aforementioned OWOD challenges. Our OW-DETR explicitly encodes multi-scale contextual information, possesses less inductive bias, enables knowledge transfer from known classes to the unknown class and can better discriminate between unknown objects and background. Comprehensive experiments are performed on two benchmarks: MS-COCO and PASCAL VOC. The extensive ablations reveal the merits of our proposed contributions. Further, our model outperforms the recently introduced OWOD approach, ORE, with absolute gains ranging from 1.8% to 3.3% in terms of unknown recall on MS-COCO. In the case of incremental object detection, OW-DETR outperforms the state-of-the-art for all settings on PASCAL VOC. Our code is available at https://github.com/akshitac8/OW-DETR. | ['Mubarak Shah', 'Fahad Shahbaz Khan', 'Salman Khan', 'K J Joseph', 'Sanath Narayan', 'Akshita Gupta'] | 2021-12-02 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Gupta_OW-DETR_Open-World_Detection_Transformer_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Gupta_OW-DETR_Open-World_Detection_Transformer_CVPR_2022_paper.pdf | cvpr-2022-1 | ['open-world-object-detection'] | ['computer-vision'] | [ 9.53155458e-02 -1.08558431e-01 -2.05278218e-01 -2.11435571e-01
-1.11220753e+00 -6.56411409e-01 5.41899085e-01 -6.36664703e-02
-4.34047699e-01 5.48182130e-01 -1.96478933e-01 8.49270672e-02
1.78488299e-01 -3.78801048e-01 -8.68744075e-01 -5.26162505e-01
-6.76155090e-04 5.39196849e-01 7.33420968e-01 2.85003215e-01
-9.39528197e-02 4.02695209e-01 -1.65960217e+00 3.37800503e-01
7.71525681e-01 1.33996010e+00 4.00779784e-01 7.66699076e-01
-4.40853909e-02 7.31968582e-01 -3.18433285e-01 -5.34910798e-01
4.88938063e-01 -8.28088522e-02 -6.09389365e-01 1.85215473e-01
1.10951078e+00 -5.22956014e-01 -4.28831130e-01 1.02223027e+00
6.24890327e-01 9.62444916e-02 4.68738168e-01 -1.40201426e+00
-9.04866397e-01 3.24244261e-01 -6.23599231e-01 6.59546077e-01
-1.72364891e-01 5.81572711e-01 1.09969616e+00 -1.47075522e+00
5.14881194e-01 1.22269642e+00 6.53802037e-01 6.44977689e-01
-1.35043275e+00 -8.38790476e-01 6.35749936e-01 4.96063828e-01
-1.39125061e+00 -4.32350844e-01 4.82879043e-01 -5.34565330e-01
6.27434969e-01 2.31109723e-01 3.71189982e-01 1.20485640e+00
-1.86201945e-01 1.29838276e+00 9.51286793e-01 -1.20050371e-01
2.06781209e-01 2.48085350e-01 4.27338034e-01 6.50550723e-01
5.32555044e-01 1.84109658e-01 -4.38406646e-01 -2.94471309e-02
3.29516768e-01 2.39693671e-01 -2.74025261e-01 -6.44068182e-01
-1.17147684e+00 5.21391153e-01 7.35507786e-01 -1.16619520e-01
-1.62220076e-01 1.72619998e-01 4.19748366e-01 2.41151191e-02
4.00382996e-01 3.67366999e-01 -6.76891565e-01 8.74301642e-02
-5.99812388e-01 2.86653191e-01 4.90400523e-01 1.07923567e+00
6.78378165e-01 9.65451240e-04 -5.36832690e-01 7.87593722e-01
2.15438694e-01 6.41136646e-01 4.12061483e-01 -6.71805024e-01
4.46770310e-01 5.36956966e-01 2.64365941e-01 -5.69655359e-01
-8.80028903e-02 -9.60271835e-01 -2.64267981e-01 1.28156036e-01
4.51764047e-01 2.79820524e-02 -1.27801716e+00 1.70650017e+00
6.76142812e-01 6.04247868e-01 -6.84863701e-03 9.80738461e-01
1.05108082e+00 6.33445442e-01 1.49845690e-01 7.72911385e-02
1.36570537e+00 -1.36382377e+00 -4.66891795e-01 -6.84928834e-01
2.51225144e-01 -5.93249619e-01 1.06341839e+00 2.45562240e-01
-7.70296991e-01 -6.48208499e-01 -8.30068409e-01 -1.11254156e-02
-2.65836090e-01 4.89577413e-01 5.04600823e-01 4.23174560e-01
-6.18367076e-01 1.64222613e-01 -7.06203997e-01 -2.93864161e-01
1.10823417e+00 9.92468521e-02 -1.64527401e-01 -4.08335954e-01
-6.86507642e-01 5.09756804e-01 6.04689538e-01 1.23061761e-01
-1.60459936e+00 -8.52090001e-01 -7.18944669e-01 9.31185782e-02
9.88445520e-01 -7.08690286e-01 1.51694465e+00 -9.57199872e-01
-9.14807975e-01 7.88911521e-01 -1.10165447e-01 -6.16048336e-01
7.12986767e-01 -5.53445816e-01 -2.72076845e-01 2.08178675e-03
3.65038723e-01 8.28352273e-01 9.84251022e-01 -1.24752271e+00
-1.02332282e+00 -4.05503988e-01 3.52241769e-02 9.63147879e-02
-3.43544692e-01 -1.31368294e-01 -7.37866163e-01 -7.49092817e-01
9.15687457e-02 -8.94511759e-01 -1.47177160e-01 4.86002326e-01
-3.33486378e-01 -3.59543979e-01 1.01020372e+00 -2.40874603e-01
8.44054401e-01 -2.28666162e+00 -1.48005247e-01 -3.93337727e-01
4.55239236e-01 4.82053101e-01 -3.70526165e-01 -1.41974837e-01
1.41482145e-01 -2.88331181e-01 -2.40191281e-01 -3.47062737e-01
1.04094446e-01 2.41429657e-02 -4.95185107e-01 3.46626163e-01
6.02858245e-01 1.12074494e+00 -9.92094636e-01 -2.15564907e-01
7.95308053e-02 2.79235035e-01 -3.40174586e-01 2.33944923e-01
-4.86856341e-01 5.88619448e-02 -3.07433695e-01 9.97466445e-01
7.00261652e-01 -3.81414533e-01 -2.83662319e-01 -4.51154970e-02
1.49583817e-01 -4.39714082e-02 -1.39319623e+00 1.39141905e+00
-2.02755600e-01 7.46604979e-01 -8.55952948e-02 -5.74393153e-01
6.76472127e-01 -7.79543668e-02 -2.44148187e-02 -5.38777232e-01
1.19252071e-01 3.26869071e-01 5.06463684e-02 -3.68070573e-01
4.13282633e-01 1.43550649e-01 1.51570693e-01 1.47277653e-01
4.51745033e-01 3.24788243e-01 2.59914756e-01 3.51620823e-01
1.06987083e+00 7.30128437e-02 2.61260182e-01 -1.68533549e-01
3.05821300e-01 -6.13817498e-02 9.21384752e-01 1.09825218e+00
-5.11025369e-01 5.49937129e-01 2.09280938e-01 -4.89528894e-01
-6.28796577e-01 -1.36247647e+00 -3.01339597e-01 1.21940053e+00
3.66512746e-01 -5.20678684e-02 -3.57868284e-01 -1.09806621e+00
3.33091766e-01 5.60130358e-01 -5.88682413e-01 -1.57983527e-01
-2.93002546e-01 -6.20582581e-01 2.78048933e-01 5.64509630e-01
3.22938859e-01 -9.85017061e-01 -5.84601283e-01 2.03033507e-01
-1.56416729e-01 -1.37900722e+00 -7.68817067e-01 1.08938567e-01
-7.40470529e-01 -1.16597927e+00 -7.16419280e-01 -7.94879198e-01
5.79247296e-01 6.40830815e-01 1.03159404e+00 -3.72928977e-01
-8.36603343e-01 5.09330034e-01 -2.83160359e-01 -7.90316224e-01
-2.60850787e-02 -1.75183073e-01 1.77924111e-02 5.56935728e-01
4.22556996e-01 -1.61982104e-01 -6.88148081e-01 5.24867952e-01
-6.78279102e-01 -1.62404627e-01 8.23432207e-01 9.03973043e-01
8.11239779e-01 -2.08499119e-01 6.58441126e-01 -8.65632176e-01
-1.17238432e-01 -6.52310431e-01 -7.43609369e-01 1.55507699e-01
-3.79182845e-01 -2.17004851e-01 3.90020907e-01 -8.30203116e-01
-9.92163420e-01 1.46549866e-01 2.21072361e-01 -8.64360392e-01
-2.14853734e-01 -7.26237446e-02 -3.20577592e-01 -3.47328968e-02
7.94740379e-01 2.75631636e-01 -4.83286381e-01 -5.24348021e-01
4.11098003e-01 5.81827223e-01 8.17892671e-01 -4.45560157e-01
9.49682832e-01 7.17048228e-01 -5.59202492e-01 -5.97466350e-01
-1.27976489e+00 -7.68962443e-01 -3.66632551e-01 -1.37539536e-01
5.20183623e-01 -1.31861734e+00 -2.81253070e-01 6.78453743e-01
-1.02266431e+00 -3.29739392e-01 -6.97213709e-01 3.29226553e-01
-2.55519331e-01 1.94225430e-01 -3.54968667e-01 -9.05556798e-01
-4.33395088e-01 -1.00075877e+00 1.18828332e+00 3.34340066e-01
2.61030972e-01 -5.24601817e-01 -2.73241550e-01 5.32154143e-01
2.02893078e-01 1.14390321e-01 4.20003116e-01 -8.75481129e-01
-1.07431602e+00 -3.01823795e-01 -5.48789144e-01 3.69870871e-01
-8.23250189e-02 -2.75053054e-01 -1.27473354e+00 -5.07341623e-01
-2.43991897e-01 -6.30552649e-01 1.39724314e+00 2.01167539e-01
1.25620151e+00 -2.64470696e-01 -4.77837414e-01 5.57934165e-01
1.31309891e+00 3.08117066e-02 2.55151600e-01 1.85940430e-01
6.86345518e-01 3.04934800e-01 9.18589592e-01 3.97676259e-01
4.40519899e-01 6.75051689e-01 6.52404904e-01 -7.87381753e-02
-3.85354787e-01 -1.37877762e-01 5.07619321e-01 1.51544577e-02
3.96644861e-01 -3.14913899e-01 -8.89338732e-01 1.01812589e+00
-1.81639183e+00 -8.86810303e-01 -1.27891637e-02 2.19876766e+00
5.61474085e-01 5.31431496e-01 2.49811396e-01 -1.77846432e-01
8.35635185e-01 -1.50132021e-02 -1.03811812e+00 3.44951421e-01
-9.34747383e-02 -4.00060751e-02 3.37748647e-01 2.06542805e-01
-1.32698321e+00 9.38955724e-01 4.89842939e+00 8.71466517e-01
-9.96055782e-01 5.29749691e-01 6.65761054e-01 -5.04503131e-01
1.40439764e-01 -1.13788143e-01 -1.28382361e+00 3.88059825e-01
4.27589417e-01 -4.09086458e-02 4.00202442e-03 1.15750301e+00
-1.57906130e-01 -8.60445276e-02 -1.29554009e+00 1.06443691e+00
1.84981644e-01 -1.23081946e+00 9.34713483e-02 -2.51556426e-01
8.78696203e-01 5.33298135e-01 1.97779566e-01 6.99036062e-01
2.60177404e-01 -5.14542699e-01 1.09098184e+00 2.29735434e-01
6.74140036e-01 -3.45110238e-01 5.86891711e-01 3.34681362e-01
-1.25716746e+00 -5.60783982e-01 -3.89464080e-01 1.80257082e-01
-9.31754187e-02 6.22899532e-01 -8.72587621e-01 2.81156301e-01
9.69487250e-01 7.42963374e-01 -8.04704309e-01 1.48290229e+00
-2.05987230e-01 8.27891290e-01 -2.64849603e-01 1.60425186e-01
3.63635570e-01 3.52369368e-01 9.68190074e-01 1.23341072e+00
2.25436538e-02 2.43350957e-02 4.77503181e-01 1.08407807e+00
-2.97897935e-01 -1.33440644e-01 -2.14067131e-01 6.50169477e-02
4.30763632e-01 1.34824145e+00 -6.82445109e-01 -4.40755874e-01
-4.39578623e-01 9.63162303e-01 4.28984910e-01 3.19258451e-01
-1.02825046e+00 -4.48262990e-01 6.41122878e-01 1.19617626e-01
9.07533646e-01 1.74926832e-01 1.57477230e-01 -1.34051323e+00
4.80336815e-01 -8.35411072e-01 7.32129276e-01 -6.02646708e-01
-1.52914870e+00 6.00744486e-01 -2.51312643e-01 -1.19033432e+00
4.62274134e-01 -7.55474448e-01 -6.11472905e-01 5.78482449e-01
-1.83587563e+00 -1.28361475e+00 -5.48791051e-01 3.11246663e-01
1.08238423e+00 -1.64454311e-01 3.44342768e-01 3.94732922e-01
-7.87138104e-01 6.46315575e-01 1.94327861e-01 1.41808972e-01
6.42387509e-01 -1.23666358e+00 5.77553689e-01 1.15962434e+00
4.58390921e-01 2.23923668e-01 4.03880179e-01 -5.70302069e-01
-1.21833777e+00 -1.66086864e+00 4.73558605e-01 -7.46815681e-01
6.63101196e-01 -7.98586488e-01 -1.10882509e+00 6.63509369e-01
-4.49075580e-01 9.09969747e-01 3.61363351e-01 4.05237125e-03
-7.56939769e-01 -3.75431329e-01 -9.95235324e-01 3.97065252e-01
1.31704235e+00 -3.59176189e-01 -5.23115516e-01 3.92764390e-01
9.82635379e-01 -5.38127184e-01 -3.75170469e-01 4.54375178e-01
4.46842581e-01 -8.49556863e-01 1.04103112e+00 -7.92405605e-01
1.23345241e-01 -5.57468653e-01 -2.22593680e-01 -9.42051291e-01
-3.76720935e-01 -5.47190309e-01 -6.00495279e-01 1.18176532e+00
4.05562907e-01 -7.45619714e-01 7.09662318e-01 2.95885652e-01
-1.59326583e-01 -8.48427117e-01 -9.82531071e-01 -1.18302011e+00
-3.16903144e-01 -5.95060050e-01 2.85013258e-01 7.74584115e-01
-6.93195403e-01 3.78728658e-01 -2.36471996e-01 4.90490794e-01
8.55499089e-01 3.69850069e-01 9.22047615e-01 -1.20133829e+00
-5.66279531e-01 -2.51624852e-01 -6.25432849e-01 -1.25087595e+00
-1.15418136e-01 -9.41289246e-01 1.86925292e-01 -1.37789321e+00
5.61423898e-01 -5.54271579e-01 -6.32779479e-01 5.94613194e-01
-6.09311879e-01 4.56119478e-01 4.95239109e-01 2.11507276e-01
-1.18664575e+00 7.63404191e-01 1.01644754e+00 -3.95765513e-01
-1.94123715e-01 1.67815536e-01 -8.27323318e-01 7.36527622e-01
5.24445355e-01 -7.33771443e-01 -1.85979754e-01 -5.01796007e-01
-2.05441296e-01 -3.36415082e-01 8.97654772e-01 -1.14719260e+00
-1.59216058e-02 -1.84528418e-02 4.54402715e-01 -7.29042232e-01
3.25189531e-01 -7.90413737e-01 -1.76754281e-01 6.34597301e-01
-1.81264773e-01 -4.91663188e-01 5.46646059e-01 1.14475691e+00
-1.38784386e-03 -7.28206411e-02 1.01512229e+00 1.11104369e-01
-1.20808887e+00 6.02925777e-01 9.89790075e-04 4.97869432e-01
1.48487604e+00 -1.78116992e-01 -4.70536858e-01 8.99121314e-02
-7.50399232e-01 5.49381375e-01 9.56594199e-02 7.62055457e-01
6.76800013e-01 -1.14065301e+00 -9.69572604e-01 4.77205254e-02
6.87304914e-01 2.29462415e-01 3.48739028e-01 7.04547644e-01
-1.26836836e-01 3.18770409e-01 5.85453436e-02 -8.93489897e-01
-1.29685032e+00 6.56553090e-01 4.99850094e-01 -9.08909366e-02
-6.03872538e-01 1.22978687e+00 7.61228919e-01 -4.67350751e-01
5.35316586e-01 -3.33132058e-01 1.86972886e-01 -6.98002055e-02
7.05531418e-01 5.84269941e-01 -6.44886633e-04 -3.76848340e-01
-3.72629523e-01 2.41030157e-01 -5.38589120e-01 4.37588781e-01
1.22501028e+00 -7.58373439e-02 2.64599383e-01 4.94945884e-01
9.36363697e-01 -3.48024040e-01 -1.70551944e+00 -8.43963146e-01
4.10317890e-02 -7.47547746e-01 3.43275033e-02 -1.00745344e+00
-8.95854235e-01 7.48627901e-01 1.00175643e+00 -3.43367428e-01
9.84916389e-01 4.10529792e-01 7.37755418e-01 5.41801035e-01
3.05264503e-01 -7.02121198e-01 4.38684821e-01 3.90365690e-01
8.65319014e-01 -1.66943777e+00 -1.71312869e-01 -4.53414172e-01
-6.53505981e-01 6.42836034e-01 1.15144360e+00 1.05512969e-01
2.66148865e-01 9.54852626e-03 -9.50833224e-03 -8.70000273e-02
-8.98437679e-01 -4.37405467e-01 5.45629442e-01 4.28080827e-01
-2.23946750e-01 -1.79772936e-02 3.84271532e-01 6.48272455e-01
4.55226898e-01 -7.19996095e-02 3.44185263e-01 7.96630919e-01
-6.01712525e-01 -5.35660505e-01 -4.88570362e-01 7.01951802e-01
-3.66391391e-01 -9.60255116e-02 -3.55191618e-01 5.54824829e-01
4.49331999e-01 6.32123590e-01 2.09259894e-02 7.34808072e-02
5.39910018e-01 -4.58708070e-02 2.47778565e-01 -8.10243487e-01
-3.07402164e-01 -9.19021145e-02 -2.43614599e-01 -5.55946469e-01
6.08917046e-03 -7.49886274e-01 -9.81452167e-01 3.48389030e-01
-8.09229851e-01 -9.60128680e-02 3.17454368e-01 5.47791123e-01
5.66880345e-01 5.73461413e-01 4.60878253e-01 -8.22179079e-01
-7.67339349e-01 -7.48422921e-01 -3.03335607e-01 3.15330923e-01
6.87375426e-01 -8.27355146e-01 -2.31251806e-01 -7.20107183e-02] | [9.391746520996094, 1.421911597251892] |
a7ee5ca8-a03d-4da3-b508-e2f20eceeb25 | a-novel-motion-detection-method-resistant-to | 1612.03382 | null | http://arxiv.org/abs/1612.03382v6 | http://arxiv.org/pdf/1612.03382v6.pdf | A Novel Motion Detection Method Resistant to Severe Illumination Changes | Recently, there has been a considerable attention given to the motion
detection problem due to the explosive growth of its applications in video
analysis and surveillance systems. While the previous approaches can produce
good results, an accurate detection of motion remains a challenging task due to
the difficulties raised by illumination variations, occlusion, camouflage,
burst physical motion, dynamic texture, and environmental changes such as those
on climate changes, sunlight changes during a day, etc. In this paper, we
propose a novel per-pixel motion descriptor for both motion detection and
dynamic texture segmentation which outperforms the current methods in the
literature particularly in severe scenarios. The proposed descriptor is based
on two complementary three-dimensional-discrete wavelet transform (3D-DWT) and
three-dimensional wavelet leader. In this approach, a feature vector is
extracted for each pixel by applying a novel three dimensional wavelet-based
motion descriptor. Then, the extracted features are clustered by a clustering
method such as well-known k-means algorithm or Gaussian Mixture Model (GMM).
The experimental results demonstrate the effectiveness of our proposed method
compared to the other motion detection approaches from the literature. The
application of the proposed method and additional experimental results for the
different datasets are available at
(http://dspl.ce.sharif.edu/motiondetector.html). | ['Marius Staring', 'Jeremy Lin', 'Sahar Yousefi', 'M. T. Manzuri Shalmani'] | 2016-12-11 | null | null | null | null | ['motion-detection'] | ['computer-vision'] | [ 2.61213422e-01 -9.71397817e-01 -5.78269400e-02 1.75518021e-01
-3.49195898e-01 -4.68960971e-01 4.32498366e-01 -1.37485251e-01
-4.39499080e-01 3.46957088e-01 2.86957715e-02 2.38020718e-02
-1.13649212e-01 -6.53052807e-01 -4.82537746e-02 -1.23001552e+00
5.31211644e-02 -2.44657010e-01 9.11488235e-01 9.92352888e-02
3.54888529e-01 5.52639365e-01 -1.63349509e+00 -1.69786453e-01
5.62282920e-01 9.95056391e-01 4.28820133e-01 7.80350089e-01
6.73963353e-02 4.45901245e-01 -5.98714769e-01 2.79727608e-01
3.03145975e-01 -4.34665710e-01 -2.65671968e-01 6.66016400e-01
-3.67742851e-02 -4.05267000e-01 -2.27542415e-01 1.02303016e+00
4.35566038e-01 4.79800493e-01 5.99539399e-01 -1.14286280e+00
-2.44290918e-01 -4.16196734e-01 -1.13195753e+00 5.76969624e-01
3.62946779e-01 1.44044638e-01 9.89077240e-02 -7.57686079e-01
6.67598009e-01 1.39398444e+00 4.08928066e-01 3.37465733e-01
-6.84598923e-01 -4.08264279e-01 -2.30492666e-01 6.03465080e-01
-1.42452800e+00 -3.37883890e-01 1.02439296e+00 -6.24580681e-01
5.83512306e-01 3.01940113e-01 5.55439413e-01 7.65715659e-01
3.65612656e-01 6.43900096e-01 1.12704384e+00 -4.09633368e-01
1.97714016e-01 -6.21617027e-02 -2.52504386e-02 6.65882349e-01
3.24714482e-01 -4.71286988e-03 4.76102754e-02 -2.46011317e-01
5.91748118e-01 1.87151894e-01 -2.94568986e-01 -2.78472275e-01
-1.01260173e+00 5.58209598e-01 -1.71170920e-01 4.64625537e-01
-4.12950635e-01 -7.48745203e-02 4.44799274e-01 -1.26453191e-01
5.12329578e-01 -4.66681749e-01 -7.49865547e-02 -3.73129696e-01
-1.19749999e+00 1.83357239e-01 4.72606152e-01 4.97654229e-01
4.48721468e-01 2.10009068e-01 1.10699564e-01 7.45303631e-01
4.52311069e-01 8.05663466e-01 5.74879467e-01 -8.96118522e-01
1.75434738e-01 3.56295466e-01 3.06201458e-01 -1.83361340e+00
-2.77729362e-01 9.54645202e-02 -1.01941991e+00 1.19474649e-01
2.70691544e-01 -1.43432707e-01 -6.74366832e-01 1.16131818e+00
7.64504671e-01 4.49819565e-01 1.78077325e-01 9.83294249e-01
7.45639563e-01 1.00338650e+00 -1.05040267e-01 -6.15492284e-01
1.36037982e+00 -7.10725963e-01 -1.03929424e+00 1.34311974e-01
2.31640354e-01 -1.21423721e+00 4.78418350e-01 5.70514679e-01
-6.15237176e-01 -7.10933685e-01 -9.31856632e-01 6.15103126e-01
-2.17541739e-01 2.74520516e-01 1.89785063e-01 7.53597140e-01
-7.76396751e-01 1.38955206e-01 -1.01772416e+00 -8.25925052e-01
1.65377378e-01 1.30720824e-01 -2.52247363e-01 -2.00059384e-01
-9.26156402e-01 6.64101303e-01 3.70187551e-01 2.80295461e-01
-6.42781079e-01 2.48004287e-01 -7.43836462e-01 -4.05177474e-01
4.36679006e-01 -2.74068624e-01 7.81721890e-01 -8.67975473e-01
-1.42409289e+00 6.31443381e-01 -4.89893585e-01 8.01257696e-03
4.43289727e-01 -7.15937391e-02 -7.04233527e-01 5.34853876e-01
3.46880443e-02 1.51601523e-01 1.00972474e+00 -9.92151260e-01
-8.12863350e-01 -3.09895396e-01 -3.46102178e-01 1.33273855e-01
-3.21543574e-01 3.99010807e-01 -5.73328137e-01 -7.09024966e-01
1.16992898e-01 -9.63330925e-01 -1.17879421e-01 -9.34185609e-02
-1.98491171e-01 -1.75638556e-01 1.56283569e+00 -8.34053338e-01
1.43551195e+00 -2.30731559e+00 6.79234490e-02 7.80456811e-02
-2.06271991e-01 7.22772062e-01 2.04244629e-01 4.87306982e-01
3.13609630e-01 -1.22832619e-01 -3.14363569e-01 9.34436917e-02
-4.50766772e-01 1.92481026e-01 1.73754871e-01 8.70534778e-01
1.28724828e-01 4.44230326e-02 -8.20047259e-01 -7.14878082e-01
6.97807431e-01 5.05406678e-01 4.20131683e-02 8.27022642e-02
2.54677445e-01 6.04607701e-01 -6.95183575e-01 8.92293036e-01
1.10624838e+00 2.21003070e-01 -1.18223593e-01 -2.02811182e-01
-4.02624011e-01 -5.97451746e-01 -1.55004120e+00 1.08773649e+00
5.77501804e-02 7.21841156e-01 7.97902122e-02 -1.09139943e+00
1.02419460e+00 4.73179281e-01 6.56192720e-01 -3.50955427e-01
3.64064664e-01 2.46888533e-01 -7.38712847e-02 -1.20822215e+00
5.50783753e-01 1.60336837e-01 2.11144254e-01 1.03621505e-01
-4.53143001e-01 5.71864396e-02 4.78443414e-01 -1.36939839e-01
8.38922143e-01 2.26797894e-01 4.50804532e-01 -2.04650596e-01
9.36393380e-01 1.98098257e-01 9.28429186e-01 2.28366017e-01
-6.00586057e-01 3.33044022e-01 -4.25679348e-02 -4.27087903e-01
-9.26671863e-01 -5.95669448e-01 -6.64536655e-02 4.18941140e-01
5.27002573e-01 1.48310876e-02 -7.38792479e-01 -2.13013411e-01
-1.81670576e-01 1.25833392e-01 -2.77921587e-01 -7.15440437e-02
-6.40087843e-01 -1.06021762e+00 2.90924937e-01 1.96710378e-01
9.95713532e-01 -9.34214413e-01 -1.13353360e+00 2.77436376e-01
-1.85977563e-01 -1.38611495e+00 -3.19269449e-01 -6.35477602e-01
-9.00316775e-01 -1.09590924e+00 -1.01395416e+00 -8.38050902e-01
5.38767993e-01 8.97662759e-01 2.89677858e-01 3.48253846e-02
-6.40441895e-01 4.74547297e-01 -6.73833191e-01 -1.71822146e-01
-2.82000780e-01 -4.31265771e-01 1.04414530e-01 3.84496152e-01
5.58388591e-01 -1.49255171e-01 -9.58795547e-01 5.17422318e-01
-1.20747781e+00 -1.97391540e-01 4.27540034e-01 4.55607653e-01
3.90266061e-01 7.56938636e-01 1.24789245e-01 -4.35707748e-01
4.91675943e-01 -5.65331757e-01 -7.31541753e-01 3.21145728e-02
-1.16942853e-01 -4.64318663e-01 2.84320295e-01 -6.50177896e-01
-1.15744829e+00 1.06748827e-01 2.21964449e-01 -3.41803789e-01
-5.94137192e-01 3.42587143e-01 -1.31863981e-01 -1.12844557e-01
2.56102860e-01 5.42662323e-01 6.91016540e-02 -3.68249536e-01
3.95599827e-02 9.10798848e-01 5.82651913e-01 -1.95643112e-01
9.33370709e-01 9.10910726e-01 1.47712767e-01 -1.38393772e+00
-2.82544374e-01 -8.20946515e-01 -5.34705698e-01 -6.00703418e-01
1.19188190e+00 -5.82313001e-01 -4.82659876e-01 9.94134426e-01
-1.23286855e+00 2.84351438e-01 5.27591109e-01 7.92580843e-01
-2.75924623e-01 1.03719187e+00 -4.31594610e-01 -1.31604004e+00
-3.12192798e-01 -1.25651264e+00 7.77665675e-01 6.39035821e-01
-8.59114230e-02 -9.71006036e-01 3.80359143e-02 3.05910558e-01
5.43242693e-01 7.92695642e-01 5.14712632e-01 -1.10881448e-01
-4.55389202e-01 -2.69800901e-01 6.93586469e-03 3.58164877e-01
5.46439230e-01 4.49100971e-01 -6.53667450e-01 -1.81257278e-01
2.47640535e-01 3.02953780e-01 6.90420270e-01 6.91335797e-01
6.65402234e-01 -3.74649316e-02 -4.28956568e-01 4.45486069e-01
1.51202142e+00 7.99454272e-01 5.75800955e-01 5.50310493e-01
5.58637023e-01 6.70441329e-01 1.06178212e+00 8.09369564e-01
6.18719533e-02 6.94747925e-01 2.90782183e-01 -2.09461361e-01
1.28717989e-01 3.71138304e-01 4.01887119e-01 8.18005860e-01
-2.78674603e-01 -2.49152854e-01 -8.23894083e-01 6.24761462e-01
-1.90368366e+00 -1.26139677e+00 -4.98229742e-01 2.17727232e+00
2.50731230e-01 -8.58934894e-02 1.99348047e-01 4.34786260e-01
1.00443256e+00 3.00129890e-01 -2.93099135e-01 -2.69866914e-01
-2.95798182e-01 -2.47817650e-01 5.66686749e-01 2.65456498e-01
-1.46939242e+00 7.33053386e-01 5.17934036e+00 9.30186033e-01
-1.15482688e+00 5.92241101e-02 4.11080450e-01 3.10790539e-01
4.29227054e-01 -4.11777869e-02 -6.27604485e-01 7.74767101e-01
4.48554695e-01 -6.06897846e-02 7.36490730e-03 4.96471792e-01
9.06569839e-01 -8.21284592e-01 -1.30147204e-01 1.20820928e+00
1.61024660e-01 -7.46020794e-01 -1.40417576e-01 6.62760511e-02
6.74516618e-01 -5.02773285e-01 9.90537480e-02 -3.34179163e-01
-3.26013148e-01 -5.46770930e-01 3.38367254e-01 6.17467761e-01
3.58541071e-01 -8.46967399e-01 8.54898572e-01 4.09121126e-01
-1.46968555e+00 3.77499647e-02 -5.27350962e-01 -2.91682482e-02
3.11448544e-01 7.75428593e-01 -3.94737184e-01 5.21589935e-01
6.75697267e-01 7.70326853e-01 -2.56647646e-01 1.30284572e+00
1.66048616e-01 6.15616083e-01 -3.13291401e-01 -8.36992934e-02
4.03895110e-01 -5.13187408e-01 8.56435001e-01 1.30746758e+00
6.36174381e-01 3.22425455e-01 2.98327595e-01 3.35308462e-01
6.95354342e-01 1.86721429e-01 -6.18314803e-01 -2.78753452e-02
3.66808504e-01 1.39283645e+00 -1.06945646e+00 -5.59132397e-01
-4.65721190e-01 1.03525555e+00 -5.92485487e-01 4.93850321e-01
-7.29863703e-01 -4.10760820e-01 4.54892665e-01 5.54583520e-02
4.68408376e-01 -6.40471101e-01 4.30322528e-01 -9.61409628e-01
8.33052918e-02 -6.45861149e-01 3.95296365e-01 -3.72531742e-01
-8.79739046e-01 3.66879404e-01 2.57143736e-01 -1.53796232e+00
-1.34329498e-01 -5.11644602e-01 -8.13379407e-01 5.94714284e-01
-1.37294328e+00 -7.78162658e-01 -4.86553669e-01 6.21877611e-01
8.35489810e-01 -2.57411212e-01 4.71052974e-01 4.87943321e-01
-8.59022141e-01 5.26112691e-02 5.00302374e-01 6.09681420e-02
5.78374207e-01 -6.39822841e-01 -9.31432471e-02 1.19324148e+00
-4.88090068e-01 1.85863838e-01 8.75589907e-01 -7.08897412e-01
-1.51281095e+00 -8.40513766e-01 5.64337254e-01 1.05305590e-01
5.69096744e-01 6.45350590e-02 -7.96151161e-01 7.25056231e-02
2.01895237e-01 1.11669637e-01 6.12022102e-01 -9.13225472e-01
2.55251020e-01 -1.81679986e-02 -1.25335538e+00 2.66124666e-01
6.14095032e-01 1.42081067e-01 -3.24367315e-01 6.25819862e-02
2.07521226e-02 -2.09265724e-01 -7.80643702e-01 5.03539681e-01
6.02587104e-01 -1.01365149e+00 6.82309031e-01 -9.51338112e-02
-3.73582647e-04 -8.20461154e-01 -2.34667331e-01 -9.27267015e-01
-2.38240436e-01 -5.59808016e-01 6.44370317e-02 1.25492477e+00
-3.48587781e-01 -5.42911947e-01 5.60992479e-01 1.74208641e-01
3.29033017e-01 -5.07450104e-01 -8.98067594e-01 -7.61499822e-01
-5.86379826e-01 -9.47294831e-02 1.61494501e-02 9.46572423e-01
-2.42199719e-01 -1.86537489e-01 -6.88961446e-01 3.33874732e-01
7.79962659e-01 1.05665125e-01 7.49331832e-01 -1.03229523e+00
-2.27375656e-01 -3.18432659e-01 -9.22734082e-01 -8.06127012e-01
-2.21647352e-01 -2.60166168e-01 -3.56192701e-02 -1.48971832e+00
2.81460702e-01 1.66806616e-02 -5.65995760e-02 -1.90997347e-01
-3.36339474e-01 2.56874621e-01 1.90528244e-01 4.00460839e-01
-5.65554976e-01 3.96700233e-01 1.09720528e+00 -4.11898419e-02
-2.43512377e-01 3.42543684e-02 1.04188092e-01 8.65788102e-01
9.77718294e-01 -4.08382416e-01 -3.06785136e-01 -3.19016069e-01
-4.35745358e-01 2.09445775e-01 2.82102376e-01 -1.20187378e+00
2.72736281e-01 -3.12072694e-01 4.34726387e-01 -7.67868400e-01
3.60118836e-01 -8.88506293e-01 2.41904899e-01 7.68058121e-01
4.71128196e-01 2.85248190e-01 2.30734691e-01 6.94562912e-01
-4.73419011e-01 -1.08777888e-01 8.79550815e-01 -1.35698929e-01
-1.22386956e+00 7.38954321e-02 -9.76246774e-01 -3.14713508e-01
1.33946991e+00 -7.43626177e-01 -2.52854884e-01 -3.38920474e-01
-6.39680088e-01 1.43460487e-03 4.58737761e-01 3.59364510e-01
8.27691555e-01 -1.35620773e+00 -6.78645909e-01 4.88807783e-02
1.21444324e-03 -3.21629226e-01 5.32003582e-01 1.29568958e+00
-8.37221086e-01 2.14119360e-01 -3.91884685e-01 -7.98510075e-01
-1.73705781e+00 4.79404360e-01 -1.74359989e-03 1.36837915e-01
-4.53924477e-01 2.56845236e-01 -1.03201596e-02 4.08620447e-01
3.63360532e-02 -1.16726741e-01 -4.80730027e-01 -8.17422345e-02
7.11041451e-01 8.35523546e-01 -3.71651053e-01 -1.10233855e+00
-4.86772418e-01 1.22396743e+00 2.90346324e-01 -9.02582109e-02
9.97675955e-01 -4.03509110e-01 -1.31144449e-01 4.61635053e-01
1.17147255e+00 -5.66719100e-02 -1.04280818e+00 -1.06795095e-01
1.13705553e-01 -7.87755430e-01 -1.46471635e-01 -6.77180439e-02
-9.99803841e-01 8.62070560e-01 9.64987159e-01 2.54003286e-01
1.46177769e+00 -3.82010370e-01 9.91156876e-01 -4.25878493e-03
3.03741097e-01 -1.22438502e+00 -6.09280132e-02 1.56288803e-01
3.67612153e-01 -1.11836827e+00 2.06941217e-02 -5.69453180e-01
-5.20500243e-01 1.17721462e+00 3.30222726e-01 -2.09955618e-01
6.70403183e-01 1.90180719e-01 3.28525156e-01 2.16668516e-01
-3.53187114e-01 -2.85495728e-01 1.56382799e-01 6.63755715e-01
3.47459286e-01 -8.84694383e-02 -6.86807334e-01 -8.10818374e-02
3.79240185e-01 3.67733277e-02 5.86992979e-01 1.20103490e+00
-8.76852453e-01 -8.82802308e-01 -1.05967236e+00 1.58573180e-01
-6.11689270e-01 4.64499831e-01 -2.03775883e-01 7.70348966e-01
2.60790706e-01 1.38690042e+00 9.52833891e-03 -4.08826292e-01
1.41282931e-01 -2.21183404e-01 3.16010267e-01 -1.01991169e-01
3.47844735e-02 5.71943223e-01 -1.68046728e-01 -4.68470335e-01
-1.06965578e+00 -7.42814362e-01 -1.18317509e+00 -2.49109134e-01
-1.98138624e-01 7.46360123e-02 7.15819597e-01 7.22757459e-01
1.56744301e-01 2.09830433e-01 7.14215040e-01 -9.45476770e-01
-1.28198802e-01 -9.90779698e-01 -7.21725464e-01 6.44271255e-01
1.97575942e-01 -9.05261636e-01 -4.29109633e-01 3.23143274e-01] | [8.98155689239502, -0.9060744047164917] |
2644832a-9211-4dba-84f0-5c61be115332 | self-attention-networks-can-process-bounded | 2105.11115 | null | https://arxiv.org/abs/2105.11115v3 | https://arxiv.org/pdf/2105.11115v3.pdf | Self-Attention Networks Can Process Bounded Hierarchical Languages | Despite their impressive performance in NLP, self-attention networks were recently proved to be limited for processing formal languages with hierarchical structure, such as $\mathsf{Dyck}_k$, the language consisting of well-nested parentheses of $k$ types. This suggested that natural language can be approximated well with models that are too weak for formal languages, or that the role of hierarchy and recursion in natural language might be limited. We qualify this implication by proving that self-attention networks can process $\mathsf{Dyck}_{k, D}$, the subset of $\mathsf{Dyck}_{k}$ with depth bounded by $D$, which arguably better captures the bounded hierarchical structure of natural language. Specifically, we construct a hard-attention network with $D+1$ layers and $O(\log k)$ memory size (per token per layer) that recognizes $\mathsf{Dyck}_{k, D}$, and a soft-attention network with two layers and $O(\log k)$ memory size that generates $\mathsf{Dyck}_{k, D}$. Experiments show that self-attention networks trained on $\mathsf{Dyck}_{k, D}$ generalize to longer inputs with near-perfect accuracy, and also verify the theoretical memory advantage of self-attention networks over recurrent networks. | ['Karthik Narasimhan', 'Christos Papadimitriou', 'Binghui Peng', 'Shunyu Yao'] | 2021-05-24 | null | https://aclanthology.org/2021.acl-long.292 | https://aclanthology.org/2021.acl-long.292.pdf | acl-2021-5 | ['hard-attention'] | ['methodology'] | [-1.41617134e-01 6.13647282e-01 1.68251025e-03 -1.77194089e-01
-7.25731730e-01 -7.82466292e-01 1.92115046e-02 1.06915183e-01
-4.73282427e-01 6.78184986e-01 -1.80734284e-02 -1.10427618e+00
-1.77830637e-01 -1.26810610e+00 -1.19212949e+00 -3.63688827e-01
-5.88945329e-01 5.66590965e-01 1.31065529e-02 -1.05211049e-01
5.04617020e-02 4.40782964e-01 -1.63858020e+00 2.87706047e-01
6.33594453e-01 1.21860337e+00 1.18112296e-01 9.63385642e-01
-6.34115219e-01 1.05199850e+00 -6.53715968e-01 -5.32315969e-01
1.91725165e-01 -1.85875461e-01 -1.44766510e+00 -4.89583105e-01
4.23123091e-01 -4.24543589e-01 -6.06438279e-01 1.19733679e+00
-1.77020296e-01 -3.18218805e-02 4.04725641e-01 -7.94058204e-01
-1.17632377e+00 1.40082109e+00 -3.87760699e-01 3.23703706e-01
1.41342938e-01 1.34360492e-01 1.52176499e+00 -6.59411907e-01
1.66983068e-01 1.28518522e+00 6.12923026e-01 7.11210489e-01
-1.19326150e+00 -9.98721242e-01 3.65100920e-01 -2.37329677e-01
-1.52697873e+00 -2.15318859e-01 2.38082737e-01 -2.16131926e-01
1.69433570e+00 3.56697738e-01 3.49992156e-01 3.52513701e-01
7.30463862e-02 7.70696521e-01 8.41786325e-01 -6.27374768e-01
5.20613976e-02 -1.69000477e-01 6.86201274e-01 1.17437792e+00
3.18200231e-01 -2.36052796e-01 -2.25482464e-01 -1.11861899e-01
9.86675501e-01 5.00160195e-02 3.75923477e-02 4.73710656e-01
-8.52284133e-01 9.48361754e-01 4.52555209e-01 6.35681987e-01
1.13637507e-01 7.12338626e-01 4.53686476e-01 4.60714579e-01
1.29357651e-01 6.63154304e-01 -7.59698927e-01 4.67044376e-02
-9.19546962e-01 1.51844293e-01 7.40683079e-01 1.42610776e+00
9.93346214e-01 4.58007544e-01 1.61588058e-01 5.39632380e-01
-1.44077493e-02 6.48033202e-01 4.77903008e-01 -1.26050198e+00
5.53881645e-01 8.59965086e-01 3.84214660e-03 -5.79419255e-01
-4.65826839e-01 -5.86821027e-02 -1.27779102e+00 -4.76523489e-02
5.61737359e-01 -1.57779545e-01 -9.56346810e-01 2.15986156e+00
-4.51812178e-01 -4.55611408e-01 -2.39666216e-02 4.66956496e-01
5.23851037e-01 1.11719716e+00 1.25449106e-01 -3.54196072e-01
1.32810652e+00 -6.80191994e-01 -2.27529034e-01 -3.63773823e-01
1.14277685e+00 -2.56807715e-01 1.37813866e+00 3.37869972e-01
-1.92319417e+00 -4.75489259e-01 -8.19570184e-01 -4.75514144e-01
-5.32058418e-01 -2.25248441e-01 1.07124591e+00 6.07582092e-01
-1.46988428e+00 4.98475313e-01 -5.73929787e-01 2.09069923e-01
6.77537546e-02 7.77343094e-01 -2.27856651e-01 4.23885398e-02
-1.37139773e+00 4.86867368e-01 3.83363485e-01 2.60704577e-01
-8.78022194e-01 -6.72630370e-01 -1.02422476e+00 4.42693621e-01
3.31064671e-01 -3.93447727e-01 1.28025246e+00 -9.81735587e-01
-1.08731508e+00 1.13450718e+00 -3.52198273e-01 -1.01167929e+00
-9.85053405e-02 -4.15883446e-03 -1.83889344e-01 -5.20231016e-02
2.23071445e-02 5.03689647e-01 4.81068343e-01 -9.73152995e-01
-5.71212471e-01 -5.75080276e-01 6.24307632e-01 -2.89695024e-01
-1.27022162e-01 3.23985785e-01 -1.63489223e-01 -3.12663466e-01
3.47588360e-01 -7.24134326e-01 -1.80177629e-01 -4.67609465e-01
-3.70055884e-01 -6.72238648e-01 1.31351411e-01 -3.07772309e-01
1.75104320e+00 -1.95274806e+00 -5.39254025e-02 4.33041364e-01
7.05236435e-01 4.47039992e-01 1.84189156e-01 1.08604640e-01
-1.52999058e-01 7.41778255e-01 -4.42766130e-01 -2.18862787e-01
3.21981668e-01 4.37936038e-01 -5.35432696e-01 1.49225727e-01
1.46253586e-01 8.99712622e-01 -6.23533845e-01 -3.39068025e-01
-2.95259893e-01 -6.39766082e-02 -7.79654503e-01 2.86858767e-01
-8.12300622e-01 -5.79778254e-01 -2.73057878e-01 5.75999796e-01
5.84440768e-01 -7.28437722e-01 3.14379483e-01 3.91323417e-01
-1.71697274e-01 5.58392286e-01 -1.16159487e+00 1.24218190e+00
-6.17101312e-01 3.22226167e-01 5.93943000e-01 -1.04809570e+00
6.63990676e-01 8.32974985e-02 1.32854700e-01 -9.12165642e-01
1.06992729e-01 3.13815266e-01 1.76685378e-01 -9.75010768e-02
6.14303708e-01 -5.12774348e-01 -5.01404226e-01 9.54166532e-01
1.35795146e-01 -4.90463302e-02 2.89749116e-01 5.70525527e-01
1.31026840e+00 -5.15119553e-01 -2.84233272e-01 -3.91382098e-01
5.48140943e-01 -1.54441193e-01 5.50692201e-01 1.14135575e+00
-3.90010444e-03 1.42921418e-01 8.52816284e-01 -5.84441543e-01
-1.27330410e+00 -1.10114479e+00 1.44289192e-02 1.77440476e+00
-1.71522155e-01 -5.10305583e-01 -9.96786654e-01 -4.74933386e-01
-9.59726349e-02 6.66216671e-01 -7.91860580e-01 -5.57212941e-02
-9.24229920e-01 -4.02545780e-01 1.22769201e+00 1.07756054e+00
4.97714609e-01 -1.41338277e+00 -6.65763855e-01 3.28785837e-01
1.72950789e-01 -5.86466789e-01 -4.53750253e-01 8.94483328e-01
-7.05586135e-01 -9.17272866e-01 -6.15056336e-01 -9.96736288e-01
7.46069610e-01 -2.93347001e-01 1.30667281e+00 5.23288310e-01
-1.34567574e-01 3.06416750e-01 4.24182490e-02 -2.59864390e-01
-3.00795704e-01 3.32441509e-01 -2.09403448e-02 -5.68985760e-01
5.28997302e-01 -5.04565418e-01 -3.81703198e-01 -5.29512167e-02
-1.30445230e+00 -2.18086526e-01 4.69413221e-01 9.52266812e-01
4.11965340e-01 8.77844840e-02 2.22895294e-01 -1.14672124e+00
3.20056528e-01 -2.64501065e-01 -7.70639300e-01 1.27162889e-01
-4.72873181e-01 4.30747539e-01 1.41482186e+00 -2.21373633e-01
-6.26286626e-01 -2.96311796e-01 -3.36774349e-01 -5.41019022e-01
-8.26945454e-02 6.60643518e-01 9.15816203e-02 2.67315209e-01
5.88621259e-01 5.65629780e-01 -2.22215921e-01 -3.20474982e-01
3.89659852e-01 3.88613373e-01 6.19477749e-01 -1.14129651e+00
5.24828970e-01 2.39787787e-01 -2.49560490e-01 -6.48390830e-01
-8.18333447e-01 4.71160375e-02 -2.86486357e-01 5.91098666e-01
5.29948175e-01 -7.61334360e-01 -1.41395187e+00 2.33054623e-01
-1.22223318e+00 -9.17379498e-01 -4.09021199e-01 -5.95173836e-02
-5.60195982e-01 3.51708889e-01 -1.26993537e+00 -1.23611081e+00
-6.52957976e-01 -8.44454050e-01 8.00659001e-01 -7.60441944e-02
-3.78582895e-01 -8.70848060e-01 -5.69632709e-01 4.47626784e-02
3.55142832e-01 -4.37589586e-01 1.71384108e+00 -5.71142435e-01
-8.23034465e-01 -1.84857488e-01 -6.52527690e-01 6.00468755e-01
-4.43131059e-01 -1.36034280e-01 -6.01957679e-01 -6.41404558e-03
-1.41111657e-01 -6.35752201e-01 8.62588882e-01 2.72012591e-01
1.65480042e+00 -9.00528133e-01 -6.23972341e-02 4.62929845e-01
1.38194668e+00 2.03176156e-01 7.86425948e-01 -4.66325544e-02
4.20073003e-01 1.77494481e-01 -3.06290805e-01 3.69803160e-01
5.36886215e-01 -2.07703993e-01 3.19081157e-01 5.25728688e-02
2.47909516e-01 -1.08044267e-01 3.08059931e-01 8.65175724e-01
5.64230457e-02 -2.57928014e-01 -1.23625469e+00 8.61974955e-01
-1.24710751e+00 -9.36768115e-01 7.10920021e-02 2.11463714e+00
1.16400874e+00 5.29223800e-01 -1.92875769e-02 3.31182837e-01
4.92766291e-01 9.42887068e-02 -4.19626474e-01 -1.39740789e+00
-1.90369904e-01 1.19097853e+00 6.98881745e-01 7.59656191e-01
-6.28306270e-01 1.26197267e+00 5.98083496e+00 6.96440279e-01
-9.82615530e-01 -1.59924284e-01 7.82935381e-01 -1.21717215e-01
-5.65035641e-01 -1.48096949e-01 -9.38808203e-01 5.34151316e-01
1.32718134e+00 -1.23798050e-01 7.05886841e-01 9.48365271e-01
-3.53700310e-01 -1.09501973e-01 -1.39835489e+00 6.18591905e-01
-1.88161552e-01 -1.46672380e+00 6.15292378e-02 1.27636284e-01
4.72111076e-01 -1.48824394e-01 2.98052758e-01 6.42244101e-01
1.07440090e+00 -1.56223869e+00 6.73606455e-01 4.89956439e-02
1.12898171e+00 -8.36795092e-01 4.86344874e-01 5.61893582e-01
-1.27645552e+00 -4.52343434e-01 -5.46920300e-01 -4.16263729e-01
-2.36502796e-01 3.66070777e-01 -4.70169425e-01 8.23335275e-02
9.97932673e-01 -1.65879682e-01 -1.40707046e-02 1.00699462e-01
4.50392440e-02 7.11317003e-01 -6.83832645e-01 -2.65648365e-01
7.88383543e-01 1.32208362e-01 -1.91556066e-01 1.45902240e+00
3.60413015e-01 7.91704834e-01 -6.04362153e-02 1.15619540e+00
-4.54980731e-01 -1.47395805e-01 -4.67394501e-01 -2.75985867e-01
4.42466706e-01 5.80485404e-01 -5.31508744e-01 -7.59311438e-01
-1.79690495e-01 5.65639377e-01 7.14850962e-01 2.58438319e-01
-7.50854313e-01 -7.28144467e-01 6.41672552e-01 4.36257780e-01
4.18673605e-01 -5.34773648e-01 -5.50253272e-01 -9.43959713e-01
-4.44034971e-02 -8.10619712e-01 7.35330820e-01 -7.51819909e-01
-9.82719004e-01 6.41015589e-01 -2.78959721e-01 -2.89130896e-01
-2.25336805e-01 -8.19345295e-01 -1.76573902e-01 1.17767107e+00
-1.07745397e+00 -8.65853906e-01 3.20703954e-01 7.61847377e-01
1.62926987e-01 2.02781409e-01 1.19905424e+00 -5.12433387e-02
-3.23023707e-01 1.14784122e+00 1.15700178e-02 6.47462726e-01
-8.74779299e-02 -1.41120732e+00 2.83208311e-01 6.07902944e-01
-1.11449316e-01 1.43903267e+00 3.78364086e-01 -2.75571704e-01
-1.79350090e+00 -9.62150753e-01 1.26246524e+00 -1.78922668e-01
8.11936021e-01 -4.61308509e-01 -1.13020706e+00 1.26705611e+00
1.49802387e-01 6.87869266e-03 5.30851185e-01 2.32367247e-01
-9.88132000e-01 -1.32366776e-01 -1.04773986e+00 6.44104898e-01
1.26040649e+00 -8.96468461e-01 -7.00269938e-01 2.43634693e-02
9.94506240e-01 -5.11304915e-01 -9.85451102e-01 1.68482348e-01
4.55984443e-01 -8.79403412e-01 6.48328662e-01 -7.35527992e-01
5.90013266e-01 7.27796257e-02 -2.40498319e-01 -6.96025133e-01
-5.07345617e-01 -7.62094259e-01 -3.32152933e-01 9.38260436e-01
7.29880512e-01 -6.72829509e-01 1.06316590e+00 9.97889221e-01
-2.00228587e-01 -4.58975941e-01 -1.05423057e+00 -5.50910652e-01
1.09046733e+00 -7.10185051e-01 6.16287053e-01 6.28472984e-01
2.83521682e-01 2.52440870e-01 -7.03659430e-02 1.17967822e-01
1.58401072e-01 3.09963346e-01 4.03124779e-01 -9.67861354e-01
-3.65172207e-01 -7.13711202e-01 -3.56280394e-02 -1.41437364e+00
5.22559702e-01 -9.31499720e-01 1.08417884e-01 -1.23027027e+00
7.31130783e-03 -9.26323712e-01 -3.55594456e-01 7.44318366e-01
4.20342505e-01 -8.49308670e-02 -2.53988653e-02 -1.77286521e-01
-6.53455555e-01 -3.77168767e-02 9.24355328e-01 -3.76478046e-01
1.15477659e-01 -4.13281173e-01 -1.02328503e+00 7.28677928e-01
4.75036621e-01 -1.70311898e-01 -6.40427247e-02 -9.51386631e-01
7.86455870e-01 3.44405442e-01 6.84788823e-02 -7.97078252e-01
4.84023064e-01 -9.09112021e-02 1.93161547e-01 -4.73666936e-01
1.49159193e-01 -5.63067496e-01 -5.11056900e-01 3.48342687e-01
-7.21646309e-01 3.36552829e-01 3.70261878e-01 1.16492927e-01
-7.11016953e-02 -1.76985994e-01 7.71386743e-01 -8.40789258e-01
-3.01872641e-01 4.27278012e-01 -5.44586480e-01 3.74804586e-01
5.78652382e-01 -2.68525094e-01 -3.13174546e-01 -2.30173349e-01
-8.08532178e-01 3.16397548e-01 2.85861880e-01 -2.36423403e-01
4.66734797e-01 -7.71974325e-01 -2.78201550e-01 5.19044340e-01
-1.85519144e-01 9.19656396e-01 4.00140136e-01 2.57884562e-01
-6.83304787e-01 7.97760844e-01 1.60246044e-01 -8.44272673e-02
-7.74285376e-01 7.45130360e-01 4.47137386e-01 -6.33244932e-01
-2.49984875e-01 1.44399869e+00 3.63563001e-01 -4.41241354e-01
3.77506971e-01 -1.11306906e+00 5.73033750e-01 -3.45040739e-01
6.35382891e-01 8.60442892e-02 -6.19246326e-02 -3.26457292e-01
-3.99921477e-01 2.88709223e-01 -2.85414904e-01 1.63280498e-02
1.16058636e+00 1.36785626e-01 -7.42654860e-01 3.80159855e-01
1.28442097e+00 -1.81404397e-01 -6.92149818e-01 -3.42060685e-01
2.15633065e-02 2.52194941e-01 -4.57229257e-01 -6.14208162e-01
-8.43253672e-01 1.27654219e+00 1.08688638e-01 8.39922667e-01
1.02905726e+00 2.42347538e-01 1.07496047e+00 5.11410594e-01
5.64294040e-01 -1.14816415e+00 -4.30703051e-02 1.27977300e+00
6.04310095e-01 -6.47683561e-01 -5.22726357e-01 1.59631804e-01
2.21722070e-02 1.02551067e+00 6.59412742e-01 -3.35259914e-01
6.38450503e-01 7.26773679e-01 -2.88430333e-01 -8.12390149e-02
-9.88254786e-01 -1.68995827e-01 -2.54097074e-01 1.01404496e-01
6.60662115e-01 1.72068506e-01 1.23730712e-01 1.09291852e+00
-5.63757896e-01 4.35010456e-02 3.82287502e-01 1.01265824e+00
-7.34078646e-01 -9.14083004e-01 -3.27856898e-01 6.60686553e-01
-8.74666393e-01 -6.01224303e-01 -2.13137358e-01 7.39745736e-01
2.72161037e-01 6.61007464e-01 5.08742571e-01 -5.02167642e-02
-1.33465007e-02 4.28436339e-01 5.97693384e-01 -8.00096214e-01
-9.72510874e-01 -1.42691582e-01 -1.27503529e-01 -2.49613404e-01
2.95078754e-01 -1.53410405e-01 -1.83815956e+00 -9.62787271e-01
-9.25651044e-02 3.64800185e-01 6.99881464e-02 8.69385183e-01
2.63832480e-01 2.48834744e-01 4.05625790e-01 -1.47835612e-01
-6.12828672e-01 -7.84736216e-01 -9.17240262e-01 1.33425415e-01
3.67020935e-01 -1.20410034e-02 -4.31148678e-01 -6.27951026e-02] | [9.515735626220703, 7.1205363273620605] |
ed98a428-bee4-4476-847c-7364f573bed8 | relocnet-continuous-metric-learning | null | null | http://openaccess.thecvf.com/content_ECCV_2018/html/Vassileios_Balntas_RelocNet_Continous_Metric_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Vassileios_Balntas_RelocNet_Continous_Metric_ECCV_2018_paper.pdf | RelocNet: Continuous Metric Learning Relocalisation using Neural Nets | We propose a method of learning suitable convolutional representations for camera pose retrieval based on nearest neighbour matching and continuous metric learning-based feature descriptors. We introduce information from camera frusta overlaps between pairs of images to optimise our feature embedding network. Thus, the final camera pose descriptor differences represent camera pose changes. In addition, we build a pose regressor that is trained with a geometric loss to infer finer relative poses between a query and nearest neighbour images. Experiments show that our method is able to generalise in a meaningful way, and outperforms related methods across several experiments. | ['Shuda Li', 'Vassileios Balntas', 'Victor Prisacariu'] | 2018-09-01 | null | null | null | eccv-2018-9 | ['pose-retrieval'] | ['computer-vision'] | [ 9.47270822e-03 -2.15781882e-01 -2.05699369e-01 -6.52111709e-01
-7.74080575e-01 -8.85988533e-01 9.85043228e-01 1.35440394e-01
-6.20346308e-01 2.10911874e-02 2.53942370e-01 3.02446634e-01
-5.15670717e-01 -6.45195723e-01 -8.36305678e-01 -4.03124124e-01
-1.76060259e-01 1.71462670e-01 2.97987968e-01 -2.26209313e-01
5.80662131e-01 1.13254321e+00 -1.41721451e+00 2.32234951e-02
-5.65548725e-02 1.18525970e+00 -1.60864457e-01 6.00078940e-01
4.58145678e-01 4.57806319e-01 -3.06742549e-01 -4.81037170e-01
7.84951031e-01 -1.07401401e-01 -7.93841302e-01 1.43969938e-01
9.85585153e-01 -3.97559702e-01 -1.00648701e+00 7.47696996e-01
4.96664703e-01 3.61965358e-01 7.51201332e-01 -1.09833074e+00
-5.68655610e-01 -4.09277305e-02 -4.01671946e-01 2.75398612e-01
1.02995968e+00 9.58457068e-02 1.04624140e+00 -1.05157948e+00
6.36913717e-01 1.36280131e+00 9.79406893e-01 2.06781372e-01
-1.07648468e+00 -3.00265372e-01 3.28451917e-02 3.14047515e-01
-1.77191985e+00 -3.64838511e-01 1.01357126e+00 -3.81129384e-01
9.64517772e-01 2.52158195e-01 8.20688784e-01 9.67820287e-01
3.19679976e-01 4.39752787e-01 5.75483263e-01 -3.66147548e-01
-3.64951491e-02 -6.95420578e-02 -3.77313673e-01 7.44158924e-01
-4.25776765e-02 4.71470833e-01 -4.55496579e-01 -4.17320523e-03
9.01258707e-01 3.69247705e-01 -3.03048015e-01 -1.17972064e+00
-1.36189854e+00 9.89957333e-01 1.21111500e+00 5.67917526e-02
-1.45679012e-01 5.36204338e-01 1.42877772e-01 3.18602711e-01
9.99368727e-02 8.39950502e-01 -1.97994471e-01 2.00346648e-03
-7.76105821e-01 1.97088867e-01 6.77917719e-01 1.10732722e+00
9.66356218e-01 -5.39609134e-01 -8.59521404e-02 5.85593820e-01
3.56868327e-01 4.88880843e-01 4.21839833e-01 -1.10446799e+00
2.51010597e-01 4.75681782e-01 -6.26548454e-02 -1.58532476e+00
-3.96298736e-01 1.33165404e-01 -3.64820600e-01 1.13444485e-01
5.54385856e-02 4.10484225e-01 -6.39754176e-01 1.24159515e+00
1.12321056e-01 1.71208858e-01 -1.41513839e-01 1.21048677e+00
7.05160499e-01 3.73935431e-01 -3.19542855e-01 3.65529776e-01
9.62568641e-01 -6.46368980e-01 -3.17898065e-01 -3.64967436e-01
5.36597371e-01 -8.87712181e-01 5.81773460e-01 2.10456878e-01
-8.84711027e-01 -5.18722355e-01 -1.16783071e+00 -2.67967701e-01
-5.82364738e-01 1.03973605e-01 4.89322454e-01 2.81480044e-01
-1.06566203e+00 7.53008664e-01 -6.38628602e-01 -5.18886209e-01
1.75585687e-01 6.41533494e-01 -6.80318773e-01 -1.79198250e-01
-9.86819088e-01 1.10321915e+00 4.30141211e-01 2.51783192e-01
-5.23817539e-01 -3.42163056e-01 -1.39964163e+00 -1.72058702e-01
-6.86187856e-03 -7.18692601e-01 9.98714805e-01 -6.23742163e-01
-1.48095489e+00 1.04516363e+00 3.21813643e-01 -1.69750541e-01
4.03167695e-01 -1.95072412e-01 -2.69141614e-01 3.19464475e-01
-1.61381558e-01 1.04757941e+00 1.07741237e+00 -9.82662976e-01
-3.79951864e-01 -4.92080092e-01 2.25562483e-01 5.28167605e-01
-1.99046701e-01 -5.63169494e-02 -6.82039320e-01 -5.09819388e-01
4.00392950e-01 -1.07075465e+00 -2.99252331e-01 4.71485764e-01
-7.55481943e-02 -8.60836133e-02 8.51189673e-01 -2.43493095e-01
9.38600183e-01 -2.20914221e+00 3.87163520e-01 4.61164922e-01
2.98395663e-01 1.99596491e-02 -3.75901848e-01 4.23650771e-01
-1.61583349e-01 -1.88900620e-01 1.41076505e-01 -2.66452208e-02
-4.86735143e-02 1.80066720e-01 -1.36517197e-01 9.56063032e-01
4.75179344e-01 1.03952444e+00 -8.38649809e-01 -2.77294010e-01
7.83495426e-01 8.11131895e-01 -4.61497694e-01 3.83603513e-01
3.40957373e-01 2.14748587e-02 -3.10260147e-01 4.10464376e-01
6.20072424e-01 2.43471980e-01 -2.83501565e-01 -5.50351441e-01
1.19952291e-01 7.26757059e-03 -1.12091553e+00 2.07066298e+00
-5.80869436e-01 9.99153137e-01 -5.01799524e-01 -6.89150095e-01
1.31715834e+00 -2.22614467e-01 3.63581032e-01 -5.61313808e-01
3.00167054e-01 -3.74501273e-02 -4.92909372e-01 -3.27147216e-01
5.99652469e-01 3.30925703e-01 -3.48347485e-01 -2.03622878e-02
1.54363587e-01 -7.25506067e-01 -2.47018382e-01 -1.04186721e-01
1.09131944e+00 1.69893548e-01 4.21969265e-01 7.83840567e-02
7.98023880e-01 -3.70551527e-01 1.02986597e-01 3.82880330e-01
-2.68832326e-01 1.02051985e+00 2.11554512e-01 -1.16429698e+00
-1.11254382e+00 -1.12682080e+00 -1.34938747e-01 7.55954087e-01
7.07727611e-01 -5.81285954e-01 -6.00773931e-01 -8.28333974e-01
2.60039598e-01 1.01008900e-01 -6.97081208e-01 -6.06609702e-01
-6.11445963e-01 8.77541751e-02 4.12995100e-01 7.45392621e-01
2.88773000e-01 -6.07173622e-01 -6.33996010e-01 -1.92406580e-01
2.68768877e-01 -1.17335117e+00 -8.97666574e-01 2.16471270e-01
-7.93840408e-01 -1.29899180e+00 -5.26224911e-01 -8.32322478e-01
5.65085232e-01 4.91185725e-01 9.39705551e-01 -2.42266059e-01
-6.20880127e-01 8.78100455e-01 -4.39744771e-01 7.62737095e-02
6.11756817e-02 1.00830965e-01 7.17618912e-02 -9.93682295e-02
7.85090327e-01 -1.51177749e-01 -7.21442580e-01 5.09405971e-01
-9.32870924e-01 -7.20514953e-01 7.48802900e-01 7.20618784e-01
5.69702923e-01 -3.87384564e-01 -3.64916265e-01 -2.08267078e-01
5.08083224e-01 -1.99214548e-01 -6.91187501e-01 2.13096589e-01
-4.07906979e-01 3.88745129e-01 2.19084635e-01 -3.28612298e-01
-2.24594742e-01 6.93201721e-01 1.86013892e-01 -9.43119586e-01
-2.63131976e-01 1.75105900e-01 -1.59270421e-01 -6.21155858e-01
7.64066637e-01 -1.21147716e-02 5.99645749e-02 -2.09926412e-01
7.61033118e-01 5.68745077e-01 7.39980042e-01 -2.13197127e-01
1.15831983e+00 4.49518323e-01 1.22495569e-01 -7.99302697e-01
-6.56564176e-01 -9.35888946e-01 -1.28364313e+00 -1.99808940e-01
8.37975621e-01 -1.00946403e+00 -6.34816051e-01 2.22250625e-01
-1.30036306e+00 1.91276982e-01 -2.32802883e-01 7.37891793e-01
-8.94903719e-01 3.19522828e-01 -2.92282701e-01 -2.39061445e-01
-2.82458635e-03 -1.19841170e+00 1.74061465e+00 5.68515323e-02
-2.97678471e-01 -1.18269897e+00 4.62733388e-01 -2.73849238e-02
4.96594310e-02 3.82368743e-01 1.08514138e-01 -4.76802021e-01
-7.14543641e-01 -7.94455647e-01 -2.70411313e-01 1.58450648e-01
2.91573219e-02 8.81121010e-02 -8.99150908e-01 -6.74284816e-01
-1.88361004e-01 -2.46004716e-01 7.84361243e-01 1.59917727e-01
1.20674086e+00 -2.22670287e-01 -4.31833684e-01 1.29416442e+00
1.40392613e+00 -1.67856529e-01 8.52642298e-01 5.31376243e-01
7.88460255e-01 4.66801792e-01 5.93523622e-01 2.16048554e-01
2.18369201e-01 9.95981991e-01 6.53126240e-01 -5.09770513e-02
-1.16067538e-02 -4.04106259e-01 3.32486898e-01 5.33548534e-01
1.06253333e-01 1.31921917e-02 -9.07819867e-01 3.14610004e-01
-1.66987300e+00 -8.14077377e-01 2.00078622e-01 2.25664377e+00
1.56412184e-01 -1.95994437e-01 -4.41406928e-02 -7.34308138e-02
6.79310024e-01 3.93863857e-01 -4.65194702e-01 -5.63369572e-01
4.55228053e-02 7.70251453e-02 9.00859118e-01 4.53291863e-01
-1.65510237e+00 9.74555492e-01 7.34517574e+00 2.96066314e-01
-9.67260480e-01 -3.68550181e-01 9.78863090e-02 -8.02125633e-02
-9.12650153e-02 -7.19992816e-02 -6.40608966e-01 1.19695544e-01
6.23529613e-01 1.00414127e-01 4.23661053e-01 8.82797897e-01
-3.49567384e-01 2.16493562e-01 -1.50202560e+00 1.54214084e+00
7.27242768e-01 -1.15942180e+00 1.02976806e-01 8.71336758e-02
5.34410834e-01 1.16971873e-01 2.29118079e-01 8.50409418e-02
7.01695457e-02 -1.14932561e+00 5.82323313e-01 7.89348304e-01
9.12549198e-01 -8.82424533e-01 8.06169212e-01 -6.45339563e-02
-1.32553124e+00 -3.08671445e-01 -5.84033310e-01 -7.16233477e-02
-3.17691833e-01 -4.37406898e-02 -9.27057743e-01 3.78272384e-01
7.25109756e-01 1.24525332e+00 -1.04562175e+00 1.30501735e+00
-1.79495409e-01 -4.22552198e-01 -3.07005912e-01 1.24069609e-01
4.42817807e-01 -7.74653256e-02 4.17786807e-01 1.19017756e+00
2.38653928e-01 -3.41251791e-01 2.41475642e-01 6.84189379e-01
-2.84039918e-02 -1.18123002e-01 -1.20086622e+00 4.26748425e-01
5.12227714e-01 1.26239574e+00 -5.52982986e-01 3.22924703e-01
-4.51778769e-01 1.42993510e+00 3.99141222e-01 1.00157395e-01
-7.50021636e-01 -6.76074803e-01 8.59951913e-01 1.02851786e-01
5.12294590e-01 -2.51837671e-01 5.07559717e-01 -1.07975912e+00
1.38624519e-01 -5.52532673e-01 2.64585465e-01 -9.55521584e-01
-1.41864264e+00 5.40959477e-01 1.50477231e-01 -1.66496849e+00
-4.90415931e-01 -9.82759774e-01 -4.93085831e-01 6.58232272e-01
-1.41582215e+00 -1.41079843e+00 -6.03476644e-01 6.85252726e-01
3.40969384e-01 -1.30893826e-01 7.68246830e-01 7.82177672e-02
-2.28570670e-01 8.14837158e-01 9.27483961e-02 4.67172593e-01
9.14913356e-01 -1.22158194e+00 5.62983632e-01 5.32118917e-01
4.46746618e-01 5.22091568e-01 2.50695109e-01 -1.29405130e-02
-1.59700584e+00 -1.09809136e+00 5.79591751e-01 -1.06170881e+00
5.41832626e-01 -4.15621817e-01 -4.31703717e-01 7.61028647e-01
-9.61325169e-02 5.44468164e-01 4.27911043e-01 -1.13860473e-01
-7.64865160e-01 -3.41954172e-01 -9.51282620e-01 2.17694715e-01
9.12863851e-01 -9.78743076e-01 -7.46637523e-01 1.66484073e-01
5.90461254e-01 -6.69426143e-01 -1.09882283e+00 2.91195959e-01
7.47883558e-01 -9.64382410e-01 1.45254552e+00 -3.79456431e-01
8.12731683e-02 -2.45295048e-01 -3.18310440e-01 -1.22821426e+00
-5.68270981e-01 -5.10624349e-01 4.12288010e-02 6.48430705e-01
5.88042755e-03 -2.15605646e-01 5.63436627e-01 4.17207479e-01
8.05982500e-02 -7.06447721e-01 -9.20251191e-01 -1.03452539e+00
2.05308288e-01 -2.34167382e-01 7.16730356e-01 6.34083927e-01
-6.35804310e-02 1.73296735e-01 -4.79558371e-02 3.21080148e-01
3.42002869e-01 -1.27965152e-01 8.70393515e-01 -1.26376259e+00
5.78285977e-02 -6.01689339e-01 -1.56665123e+00 -1.09760821e+00
4.59122270e-01 -8.60982358e-01 9.38057899e-03 -1.09377801e+00
1.93586260e-01 -2.07704175e-02 -3.46765995e-01 3.07456944e-02
8.84706303e-02 5.73696673e-01 2.38086402e-01 2.28829846e-01
-8.68514597e-01 5.79328477e-01 9.81469989e-01 -1.79069385e-01
-1.38871958e-02 -1.29678965e-01 -2.48608306e-01 7.10113645e-01
5.06594539e-01 -3.43337804e-01 -2.45502904e-01 -5.49918354e-01
3.94474268e-01 -1.90923810e-01 8.99973512e-01 -1.17267644e+00
2.92538077e-01 1.43351778e-01 8.43785167e-01 -4.76695150e-01
5.10726631e-01 -1.15332663e+00 -3.22183669e-02 2.67770201e-01
-4.71560329e-01 3.65130305e-01 8.62696767e-03 7.54856765e-01
-3.80930573e-01 -1.45287454e-01 6.23448968e-01 9.49537531e-02
-1.05094635e+00 5.57411134e-01 6.45743385e-02 -9.23411027e-02
1.12420094e+00 -6.84153497e-01 2.07692325e-01 -4.92239833e-01
-3.99918407e-01 -5.73117398e-02 8.90230775e-01 7.89640665e-01
9.30025578e-01 -1.85398996e+00 -3.03997338e-01 5.28790414e-01
7.28216946e-01 -1.80322766e-01 -3.06147575e-01 4.66072977e-01
-9.05188441e-01 5.92100978e-01 -1.96627080e-01 -9.52829778e-01
-1.08564556e+00 8.24331522e-01 5.67765236e-01 1.33807242e-01
-4.27046955e-01 8.39745820e-01 -1.77715927e-01 -8.39865923e-01
3.45435619e-01 -4.54197973e-01 -1.06902696e-01 -2.75597125e-02
4.00520444e-01 3.62457156e-01 1.66507810e-01 -1.02626908e+00
-6.01624906e-01 1.24259293e+00 9.61139798e-02 8.42975751e-02
1.20061457e+00 -3.44173431e-01 9.44638774e-02 3.02492648e-01
2.08313656e+00 -3.09189230e-01 -1.55487347e+00 -2.74696410e-01
6.42567873e-02 -9.98442173e-01 2.70342439e-01 -2.58500665e-01
-9.67250407e-01 8.22185695e-01 9.56813276e-01 -2.68899798e-01
9.69030380e-01 2.21977726e-01 5.47227085e-01 8.80196452e-01
2.85822660e-01 -9.23579991e-01 3.12877148e-01 5.80853462e-01
1.11360025e+00 -1.55936253e+00 3.41636389e-01 -9.73080937e-03
-3.23971957e-01 1.50456512e+00 3.72183561e-01 -5.33343732e-01
6.30492926e-01 -1.50657162e-01 6.69085309e-02 -4.23437089e-01
-2.45336413e-01 -1.93043247e-01 8.78968894e-01 8.88191700e-01
3.49532157e-01 -2.01259345e-01 3.40853363e-01 -1.95511773e-01
-3.06008935e-01 -5.13686121e-01 2.22754583e-01 6.15270555e-01
-2.50701517e-01 -1.03240156e+00 -4.22464788e-01 2.49565169e-02
1.76247343e-01 1.23762809e-01 -8.21465909e-01 1.01315761e+00
-6.18668012e-02 5.52526057e-01 4.78358328e-01 -6.64820194e-01
5.66623211e-01 -1.80501848e-01 8.26564729e-01 -5.14266610e-01
-3.67155164e-01 -1.59242675e-01 -5.35970867e-01 -1.15921795e+00
-4.30945277e-01 -6.29877627e-01 -5.42476892e-01 -3.00295681e-01
-5.04827976e-01 -2.37029836e-01 6.91289663e-01 5.31324208e-01
3.37895095e-01 -2.25219484e-02 1.15267944e+00 -1.38361895e+00
-6.19614184e-01 -5.92038810e-01 -3.87520313e-01 6.92693233e-01
5.64724445e-01 -7.78499961e-01 -4.66160655e-01 1.53092481e-02] | [7.906341075897217, -2.180464029312134] |
90327f8e-cb6e-46be-87c3-3981ee8d1db9 | iitk-at-the-finsim-task-hypernym-detection-in | 2007.11201 | null | https://arxiv.org/abs/2007.11201v1 | https://arxiv.org/pdf/2007.11201v1.pdf | IITK at the FinSim Task: Hypernym Detection in Financial Domain via Context-Free and Contextualized Word Embeddings | In this paper, we present our approaches for the FinSim 2020 shared task on "Learning Semantic Representations for the Financial Domain". The goal of this task is to classify financial terms into the most relevant hypernym (or top-level) concept in an external ontology. We leverage both context-dependent and context-independent word embeddings in our analysis. Our systems deploy Word2vec embeddings trained from scratch on the corpus (Financial Prospectus in English) along with pre-trained BERT embeddings. We divide the test dataset into two subsets based on a domain rule. For one subset, we use unsupervised distance measures to classify the term. For the second subset, we use simple supervised classifiers like Naive Bayes, on top of the embeddings, to arrive at a final prediction. Finally, we combine both the results. Our system ranks 1st based on both the metrics, i.e., mean rank and accuracy. | ['Ashutosh Modi', 'Vishal Keswani', 'Sakshi Singh'] | 2020-07-22 | null | https://aclanthology.org/2020.finnlp-1.14 | https://aclanthology.org/2020.finnlp-1.14.pdf | finnlp-coling-2020-1 | ['learning-semantic-representations'] | ['methodology'] | [-4.29067552e-01 4.42086719e-02 -2.12297261e-01 -5.84869325e-01
-3.81707460e-01 -7.58390784e-01 7.18789458e-01 5.26203573e-01
-7.86289334e-01 4.13769871e-01 5.90320289e-01 -4.27813828e-01
-3.03969651e-01 -1.08549535e+00 -2.22315609e-01 -2.86297143e-01
2.64701582e-02 6.77367270e-01 2.23955825e-01 -2.84301728e-01
4.88407791e-01 2.55046964e-01 -1.54488087e+00 2.58996785e-01
7.07848668e-01 1.41608226e+00 2.39369534e-02 1.74527243e-01
-6.85419142e-01 9.62834597e-01 -1.22578509e-01 -7.55198002e-01
2.78675258e-01 9.78981778e-02 -1.04824972e+00 -4.00904179e-01
-8.35807621e-02 -5.73462658e-02 -3.74956191e-01 1.18648601e+00
1.85556397e-01 3.40013891e-01 9.36217427e-01 -8.68827820e-01
-7.70237982e-01 7.90223539e-01 3.10919080e-02 1.52201205e-01
2.30378389e-01 -4.54382122e-01 1.71369493e+00 -1.22771227e+00
7.62731493e-01 1.08296514e+00 4.70362514e-01 3.44932348e-01
-9.59247708e-01 -6.25848293e-01 -2.93356013e-02 5.15014350e-01
-1.18514216e+00 -1.56414315e-01 5.87816358e-01 -7.96897471e-01
1.10691071e+00 -1.66263089e-01 4.22847062e-01 8.39582860e-01
8.20873976e-02 3.00613433e-01 7.74639487e-01 -5.41480958e-01
4.84453171e-01 4.62682873e-01 4.93809909e-01 4.27396804e-01
4.75309074e-01 -8.47169608e-02 -3.63772511e-01 -1.45360574e-01
3.33423987e-02 8.84747226e-03 -2.26258067e-03 -4.01677787e-01
-6.46953464e-01 1.22600198e+00 3.74309450e-01 6.35299802e-01
-3.29616785e-01 -8.50723460e-02 4.10441667e-01 3.01506460e-01
6.57228053e-01 8.75987172e-01 -6.97503269e-01 -5.16771190e-02
-8.20221364e-01 3.35848749e-01 9.25829470e-01 4.98287410e-01
9.33763325e-01 -5.56739569e-01 5.76778539e-02 1.04081714e+00
6.32147312e-01 -9.93907899e-02 1.05594707e+00 -5.42866290e-01
4.05013621e-01 8.24661911e-01 2.19953477e-01 -9.85154688e-01
-3.55988860e-01 -3.01288694e-01 8.98691639e-03 7.56667405e-02
1.26469165e-01 -1.22803152e-01 -8.26308608e-01 1.48493421e+00
1.89941689e-01 1.41530022e-01 4.47224528e-01 8.21918607e-01
7.88881779e-01 4.54570800e-01 1.69678509e-01 2.66423345e-01
1.56130421e+00 -9.57651258e-01 -5.46160817e-01 -1.16025493e-01
7.90458202e-01 -7.09832549e-01 7.16661215e-01 1.63204938e-01
-4.81804848e-01 -3.20666850e-01 -1.29526651e+00 -1.37316480e-01
-1.13079107e+00 -2.83366084e-01 5.78944147e-01 5.06506145e-01
-6.92972660e-01 8.70153606e-01 -4.55277711e-01 -3.57393503e-01
1.02759108e-01 -4.35892269e-02 -2.82456309e-01 -2.00734496e-01
-1.67322171e+00 1.06827605e+00 7.74785221e-01 -7.55273283e-01
-5.39415181e-01 -6.70981526e-01 -1.09830797e+00 2.46774852e-01
1.36522025e-01 -4.38521922e-01 1.04703450e+00 -5.20008862e-01
-1.06243336e+00 1.10547566e+00 5.07985093e-02 -6.95719421e-01
1.70596629e-01 -2.92344362e-01 -7.90987909e-01 1.29267186e-01
4.16360587e-01 3.65580171e-01 2.90964544e-01 -1.06929874e+00
-8.19618821e-01 -2.77666807e-01 2.64819592e-01 1.74531981e-01
-9.68751788e-01 1.28887549e-01 -4.61383998e-01 -5.64769685e-01
3.02175265e-02 -5.74700952e-01 -1.59469187e-01 -4.30959880e-01
-2.08111525e-01 -7.15410471e-01 4.22344834e-01 -6.69780612e-01
1.48885858e+00 -2.29733109e+00 -1.22629374e-01 3.96090925e-01
2.78987706e-01 1.51855769e-02 -1.63138285e-02 3.82252425e-01
-2.90555954e-01 2.89422810e-01 5.12474515e-02 -3.43493193e-01
2.00517967e-01 1.74112871e-01 -3.84958208e-01 -8.10335129e-02
2.77781099e-01 6.18123055e-01 -1.01316774e+00 -4.22772229e-01
8.03811923e-02 -1.25201968e-02 -6.93098485e-01 2.18995333e-01
-2.15057924e-01 -1.88913077e-01 -4.95415896e-01 4.03100729e-01
4.81730729e-01 -8.38914886e-02 5.00049293e-01 -1.57584369e-01
-1.31426036e-01 6.98856831e-01 -1.09268594e+00 1.61195982e+00
-6.82618558e-01 3.56878966e-01 -6.27657950e-01 -1.26516080e+00
1.21799743e+00 2.67671406e-01 6.05274022e-01 -6.03364289e-01
2.45679021e-01 5.03392100e-01 -1.11205667e-01 -5.53291142e-01
6.75703287e-01 -2.96475202e-01 -3.69893759e-01 3.66335511e-01
4.81527239e-01 2.69682109e-01 4.29685295e-01 1.24373592e-01
1.23248971e+00 2.40242019e-01 4.93805677e-01 -4.96493638e-01
4.34921920e-01 6.13774545e-02 5.70694685e-01 2.98684508e-01
7.52293542e-02 5.36285996e-01 6.75039768e-01 -6.91115201e-01
-8.45632553e-01 -1.06729639e+00 -4.18795198e-01 1.17009830e+00
9.87218022e-02 -7.07096040e-01 -4.17641461e-01 -8.70112598e-01
4.61621761e-01 9.77431774e-01 -6.39436841e-01 -1.33735657e-01
-1.75993353e-01 -5.15292943e-01 3.57679605e-01 5.47518432e-01
8.11616331e-02 -8.55856597e-01 -4.93848532e-01 3.69004965e-01
1.25565425e-01 -1.12166762e+00 -3.52137275e-02 4.77483392e-01
-5.84683061e-01 -1.15094244e+00 -5.72612941e-01 -8.08224499e-01
2.03588709e-01 -2.07445413e-01 1.11068380e+00 -2.33376920e-01
-1.31297484e-01 1.91193491e-01 -6.98387206e-01 -2.00129926e-01
3.62575278e-02 2.31682090e-03 3.15084755e-01 1.50637835e-01
1.03376567e+00 -4.50419158e-01 -4.84833598e-01 -5.11608692e-03
-8.20109367e-01 -6.10943556e-01 2.23492593e-01 6.81420386e-01
3.89263839e-01 2.22222447e-01 4.80657190e-01 -8.15970659e-01
7.80455530e-01 -8.28988016e-01 -4.32810843e-01 3.22022498e-01
-9.83404994e-01 3.22765112e-01 6.07165575e-01 -1.55768573e-01
-7.37695754e-01 -2.69374877e-01 -5.91839012e-03 -4.36877102e-01
1.91812560e-01 1.08333516e+00 -1.49841294e-01 3.53686005e-01
6.13199472e-01 -1.48626804e-01 -3.99544835e-01 -7.65089631e-01
4.97571021e-01 9.23478186e-01 3.58510166e-01 -6.82368636e-01
6.78849638e-01 1.73136555e-02 -4.58565533e-01 -4.36761588e-01
-1.10541606e+00 -7.46113420e-01 -5.21904349e-01 1.10020235e-01
1.13079548e+00 -7.90147126e-01 -2.41834491e-01 -7.18477443e-02
-1.10147297e+00 2.11666226e-01 -3.66906494e-01 8.43039751e-01
-2.70874858e-01 1.50693521e-01 -3.58843416e-01 -7.03720272e-01
-1.71628803e-01 -7.74337113e-01 5.16881466e-01 1.76923245e-01
-3.68975520e-01 -1.26896012e+00 4.36940670e-01 2.36967027e-01
2.68583566e-01 7.75622651e-02 1.22045684e+00 -1.57335556e+00
2.88004819e-02 -4.45432842e-01 -3.67964387e-01 6.05707288e-01
5.90403602e-02 -3.38948578e-01 -1.01799572e+00 1.15447119e-01
-1.51655838e-01 -2.84367591e-01 1.23148596e+00 -8.13892186e-02
1.14318180e+00 1.55034466e-02 -3.09068590e-01 3.64662170e-01
1.72630060e+00 2.58885771e-01 4.99405563e-01 8.01352561e-01
4.66832310e-01 7.01751888e-01 7.33482957e-01 6.72632515e-01
6.29222155e-01 5.28489649e-01 1.74675643e-01 5.40035844e-01
2.46014789e-01 -4.30445194e-01 3.74061227e-01 7.70166874e-01
8.27285722e-02 -1.13553070e-01 -1.32465470e+00 7.48162210e-01
-1.73359668e+00 -7.38980830e-01 2.69096285e-01 2.16196346e+00
7.00425744e-01 3.34330738e-01 -2.25975513e-01 1.40965164e-01
5.49082994e-01 1.18683740e-01 -2.48816133e-01 -5.20056546e-01
1.28869429e-01 6.44188583e-01 4.96582210e-01 3.85953993e-01
-1.26358569e+00 9.46743369e-01 5.56090832e+00 6.98130190e-01
-9.00241911e-01 2.00045377e-01 2.92470455e-01 -1.84521284e-02
-6.41017318e-01 2.99993664e-01 -8.12515318e-01 6.41692877e-01
1.07253635e+00 -5.23852885e-01 2.01710165e-01 1.02712607e+00
-3.60513330e-01 1.88984022e-01 -1.13889134e+00 6.49734020e-01
1.35714278e-01 -1.20701218e+00 1.66540697e-01 6.86337426e-02
5.15234649e-01 1.68327484e-02 -1.86754242e-01 7.72451401e-01
5.52890122e-01 -1.11271942e+00 5.91917932e-01 5.76057673e-01
6.38546348e-01 -8.01768303e-01 1.16079938e+00 7.55159631e-02
-1.32896042e+00 -2.74897963e-01 -4.59979266e-01 -5.55064064e-04
8.76953080e-02 8.65629137e-01 -4.72154647e-01 9.25571740e-01
7.37174928e-01 9.24366117e-01 -3.82335305e-01 8.05359364e-01
-2.42436975e-01 2.66470253e-01 -1.15904994e-01 -4.76881824e-02
3.81606877e-01 -2.83106476e-01 2.00354651e-01 1.10381424e+00
4.47327435e-01 -6.25918480e-03 2.77459741e-01 7.50205576e-01
-3.03052843e-01 4.13444161e-01 -5.45888126e-01 -2.88627982e-01
5.90692103e-01 1.19218743e+00 -3.75490755e-01 -6.07103169e-01
-6.44995391e-01 6.96260691e-01 4.99313563e-01 9.83324796e-02
-6.13714218e-01 -8.70358706e-01 9.41999793e-01 -2.28745509e-02
4.75703657e-01 -5.17408215e-02 -2.78331190e-01 -1.40828609e+00
4.61414307e-02 -2.52845734e-01 7.54400909e-01 -5.45881689e-01
-1.57773554e+00 5.43419540e-01 -1.08197004e-01 -1.17659390e+00
-3.91001880e-01 -1.15398347e+00 -6.33135378e-01 8.61438870e-01
-1.75496876e+00 -6.90488040e-01 5.14181191e-03 3.24800700e-01
1.19854651e-01 -5.39828241e-01 1.09598231e+00 5.69803655e-01
-3.64352226e-01 5.21025896e-01 2.07463548e-01 4.56398875e-01
8.62249792e-01 -1.48296118e+00 1.84866875e-01 4.34855133e-01
1.85577720e-01 7.85925269e-01 4.97627437e-01 -5.80589294e-01
-5.93363047e-01 -1.05754423e+00 1.65610754e+00 -4.21710640e-01
1.12905979e+00 -1.70505613e-01 -7.91454375e-01 5.00782311e-01
-3.83943580e-02 8.68866146e-02 1.11976004e+00 5.13747811e-01
-7.99010992e-01 -9.57502425e-02 -1.17351055e+00 2.91726381e-01
8.57488632e-01 -6.00610852e-01 -1.26144063e+00 6.66260496e-02
8.92298877e-01 2.32916191e-01 -1.19906127e+00 1.97158307e-01
5.99976182e-01 -6.20460153e-01 8.28166187e-01 -1.06143105e+00
9.42440867e-01 -1.19104594e-01 -6.69095218e-01 -1.30395901e+00
-4.12784487e-01 1.64304867e-01 1.18784092e-01 1.31145179e+00
6.72401428e-01 -7.69606411e-01 5.21725655e-01 4.19206977e-01
1.80881042e-02 -7.66317964e-01 -9.87008035e-01 -9.56451476e-01
4.05771732e-01 -6.11597240e-01 7.83795595e-01 1.23822129e+00
3.53381902e-01 3.25679570e-01 8.58870894e-03 -6.17197417e-02
3.82382780e-01 2.86732823e-01 1.80975959e-01 -1.60816836e+00
-1.38708919e-01 -4.57592964e-01 -7.74023294e-01 -7.49829710e-01
3.53574604e-01 -1.25201058e+00 -1.71771586e-01 -1.52110159e+00
1.54032409e-01 -4.76298183e-01 -1.00270414e+00 3.97974908e-01
-5.43647306e-03 2.22594403e-02 1.06394388e-01 1.96135610e-01
-4.96269733e-01 4.64770079e-01 4.17543441e-01 -1.36953175e-01
1.26460850e-01 -4.70913947e-01 -8.97870958e-01 7.48951912e-01
8.51849139e-01 -5.93372643e-01 -1.01294264e-01 -4.17773902e-01
2.98885435e-01 -3.31110209e-01 1.21916041e-01 -9.37999249e-01
4.77677174e-02 -1.59613356e-01 1.46581858e-01 -1.25560910e-01
1.04212239e-01 -8.20211172e-01 -4.65994716e-01 3.22147280e-01
-5.59920907e-01 -1.49655551e-01 -1.71964675e-01 3.73231620e-01
-4.55729008e-01 -8.21290195e-01 6.42285526e-01 -2.21533217e-02
-1.07458961e+00 1.65821224e-01 -8.68401304e-02 2.63066739e-01
8.85322154e-01 2.04486027e-01 -6.78508133e-02 -2.08183099e-02
-9.22139823e-01 2.81039327e-01 2.96176672e-01 7.63023973e-01
4.04463291e-01 -1.66308320e+00 -5.87460577e-01 1.06158387e-02
6.43957138e-01 -3.71310234e-01 -2.67852813e-01 3.20974618e-01
-3.83108169e-01 5.05509794e-01 -7.30344728e-02 6.94646984e-02
-8.01697254e-01 5.72298348e-01 2.29638502e-01 -6.03573620e-01
-2.89695263e-01 7.26403177e-01 1.87679064e-02 -5.16614914e-01
1.33743331e-01 -3.21955532e-01 -8.21761191e-01 6.25726223e-01
5.49166441e-01 3.16556506e-02 1.57429144e-01 -4.12992328e-01
-5.89757860e-01 4.97882724e-01 -1.35686129e-01 -3.52622032e-01
1.56584096e+00 2.33899787e-01 -7.53403008e-02 5.19795060e-01
1.52765906e+00 -7.72804320e-02 -3.83867562e-01 -4.61425543e-01
7.45308101e-01 -4.50651139e-01 9.23635513e-02 -6.12444520e-01
-8.45522165e-01 8.10039222e-01 5.74311137e-01 3.60847443e-01
6.71944261e-01 7.88109601e-02 7.49533892e-01 4.38530058e-01
3.41757894e-01 -1.51873112e+00 -4.61599901e-02 8.61366868e-01
5.67866385e-01 -1.01564801e+00 -1.13402128e-01 -5.14628403e-02
-6.43099308e-01 1.38179839e+00 3.08562487e-01 -4.02969927e-01
9.76499796e-01 -5.40828928e-02 1.66833818e-01 -3.27024639e-01
-6.98087454e-01 -4.45932984e-01 3.42866927e-01 3.33826751e-01
6.30785644e-01 9.75178853e-02 -6.65832460e-01 1.19925964e+00
-5.15818954e-01 -2.61072069e-03 1.93665549e-01 6.75296605e-01
-7.58080363e-01 -1.27457988e+00 2.34415103e-02 6.14091635e-01
-4.83246356e-01 -3.18951637e-01 -1.67551830e-01 4.02667105e-01
3.43241841e-01 9.25290406e-01 3.31165493e-01 -7.41587460e-01
5.05836070e-01 6.30435586e-01 -1.31667659e-01 -8.56467783e-01
-3.51977944e-01 -4.37419593e-01 4.19260591e-01 -4.34110403e-01
-1.66484509e-02 -6.27745926e-01 -1.24049342e+00 -1.41442761e-01
-3.87950763e-02 3.92796576e-01 6.92486525e-01 1.11861718e+00
2.78981537e-01 4.54251170e-01 7.79423714e-01 -3.72987807e-01
-8.49453151e-01 -9.55990553e-01 -7.34650791e-01 7.64724553e-01
-1.37618765e-01 -9.24092770e-01 -5.78612208e-01 -2.57000685e-01] | [10.320415496826172, 8.804228782653809] |
83113458-39e4-48bf-95ae-f0ae8aa1936b | dispositionet-disentangled-pose-and-identity | 2211.05499 | null | https://arxiv.org/abs/2211.05499v1 | https://arxiv.org/pdf/2211.05499v1.pdf | DisPositioNet: Disentangled Pose and Identity in Semantic Image Manipulation | Graph representation of objects and their relations in a scene, known as a scene graph, provides a precise and discernible interface to manipulate a scene by modifying the nodes or the edges in the graph. Although existing works have shown promising results in modifying the placement and pose of objects, scene manipulation often leads to losing some visual characteristics like the appearance or identity of objects. In this work, we propose DisPositioNet, a model that learns a disentangled representation for each object for the task of image manipulation using scene graphs in a self-supervised manner. Our framework enables the disentanglement of the variational latent embeddings as well as the feature representation in the graph. In addition to producing more realistic images due to the decomposition of features like pose and identity, our method takes advantage of the probabilistic sampling in the intermediate features to generate more diverse images in object replacement or addition tasks. The results of our experiments show that disentangling the feature representations in the latent manifold of the model outperforms the previous works qualitatively and quantitatively on two public benchmarks. Project Page: https://scenegenie.github.io/DispositioNet/ | ['Nassir Navab', 'Federico Tombari', 'Helisa Dhamo', 'Yousef Yeganeh', 'Azade Farshad'] | 2022-11-10 | null | null | null | null | ['image-manipulation'] | ['computer-vision'] | [ 1.13363959e-01 7.31278956e-02 -3.16159315e-02 -2.67920375e-01
6.46616158e-04 -9.04352725e-01 8.04521918e-01 -7.16159195e-02
7.23650753e-02 2.86582291e-01 3.82890403e-01 3.20054024e-01
-1.95277646e-01 -7.83789992e-01 -9.55085158e-01 -8.91582370e-01
2.13700667e-01 4.94017959e-01 1.30458206e-01 -1.43427998e-01
2.05937549e-01 6.16534889e-01 -1.78292036e+00 1.68579608e-01
6.55910969e-01 6.49519324e-01 4.87229437e-01 6.38141334e-01
6.17363416e-02 5.30633390e-01 -2.90209949e-01 -3.62426639e-01
5.54567754e-01 -1.76610440e-01 -4.55507904e-01 5.52387893e-01
8.32714021e-01 -3.97955805e-01 -7.63045251e-01 1.15470624e+00
2.37289160e-01 2.43305653e-01 8.94167304e-01 -1.66940618e+00
-1.15635848e+00 5.94804943e-01 -7.20564246e-01 -2.23178953e-01
4.07145530e-01 4.44566607e-01 1.04624462e+00 -7.95266271e-01
7.40847170e-01 1.60741532e+00 -8.74698628e-03 4.86934304e-01
-1.59854174e+00 -5.42445362e-01 3.49615425e-01 2.72730529e-01
-1.39064288e+00 -4.82584178e-01 1.09282804e+00 -7.25886226e-01
4.16245103e-01 4.55387354e-01 7.12272525e-01 1.18693888e+00
1.61472261e-01 7.28547990e-01 8.56055379e-01 -2.60261387e-01
2.82681938e-02 2.16628373e-01 3.02451104e-02 8.20420563e-01
5.09763241e-01 2.91857449e-03 -5.01171708e-01 -2.49145910e-01
1.23805678e+00 2.90472507e-01 -3.64118338e-01 -1.39194095e+00
-1.47194219e+00 7.35751510e-01 7.41023302e-01 -1.49764284e-01
-3.18080455e-01 4.02954668e-01 1.43937513e-01 -1.51148468e-01
2.48486146e-01 8.11026692e-01 -1.98440462e-01 2.21600696e-01
-2.62492836e-01 3.88794601e-01 5.20326614e-01 1.38534856e+00
9.40017223e-01 -8.24029222e-02 -6.99162185e-01 4.88565564e-01
2.88120449e-01 4.31365848e-01 -3.53200138e-02 -1.18426991e+00
4.59174603e-01 8.28847647e-01 2.36898661e-01 -1.34954500e+00
-2.29178835e-02 -7.47394338e-02 -7.60771513e-01 3.07480037e-01
3.79108757e-01 2.77088374e-01 -1.13874495e+00 1.86910808e+00
4.57532883e-01 2.02906758e-01 -3.21222097e-01 8.22754562e-01
8.49193215e-01 3.64093274e-01 1.41907018e-02 2.56565690e-01
1.40172195e+00 -1.05629432e+00 -7.72153139e-01 -7.84990117e-02
1.93165943e-01 -6.85089827e-01 1.17006516e+00 6.76368326e-02
-1.05655479e+00 -5.20405650e-01 -8.63671541e-01 -3.11379045e-01
-3.76328170e-01 7.71590248e-02 8.09053302e-01 3.15532357e-01
-9.03792977e-01 4.12301004e-01 -8.48162472e-01 -1.72331616e-01
5.97074807e-01 2.87598580e-01 -6.44553125e-01 -3.55679154e-01
-7.04521537e-01 6.54213607e-01 2.69907922e-01 1.57304987e-01
-1.09065151e+00 -8.31575572e-01 -1.30712759e+00 1.32628277e-01
4.98021722e-01 -1.04625201e+00 6.93412006e-01 -5.61478019e-01
-1.30474830e+00 8.40542376e-01 -1.03377588e-02 1.71509326e-01
5.28220713e-01 -4.39008363e-02 3.39771241e-01 1.67061966e-02
7.19503984e-02 9.20866072e-01 1.12068057e+00 -1.55812573e+00
-1.08337782e-01 -5.67198694e-01 5.48476219e-01 3.68829459e-01
-1.18534602e-01 -4.13407385e-01 -6.13349020e-01 -6.22406542e-01
1.44988820e-01 -1.07080829e+00 -2.96226054e-01 3.81651729e-01
-8.07380736e-01 -1.34095907e-01 9.52227950e-01 -3.45304608e-01
7.23413944e-01 -2.37382984e+00 7.79430568e-01 -1.61883026e-01
7.12953806e-01 -1.68511719e-01 -2.96688020e-01 4.76398706e-01
-2.39064008e-01 2.53403187e-01 5.48856184e-02 -5.78079939e-01
8.59338716e-02 3.68335605e-01 -1.41298369e-01 7.61751413e-01
1.33279964e-01 1.02228189e+00 -1.17254162e+00 -4.21006441e-01
5.88115990e-01 5.39314866e-01 -8.85681927e-01 3.69027346e-01
-3.48336577e-01 7.02332914e-01 -6.42643154e-01 3.30729634e-01
7.95165956e-01 -4.12420541e-01 -5.66829108e-02 -5.97587109e-01
2.77529806e-01 1.36324808e-01 -1.25991559e+00 1.91824138e+00
-1.77497968e-01 4.99024510e-01 -7.48839453e-02 -6.67136908e-01
5.74816465e-01 9.92754921e-02 4.25299913e-01 -1.98624119e-01
8.98139328e-02 -4.57985252e-01 -1.84889078e-01 -6.23599768e-01
5.93636572e-01 4.54461843e-01 -1.66582897e-01 2.47503489e-01
3.83225158e-02 -6.19832039e-01 1.96826488e-01 6.61812067e-01
9.51494753e-01 2.88331270e-01 2.09392309e-01 -1.99657097e-01
1.94336958e-02 -1.68337300e-01 3.13514829e-01 6.89225793e-01
2.54011229e-02 7.07055748e-01 4.19004261e-01 -2.46093273e-01
-1.04365790e+00 -1.50260055e+00 -8.26944113e-02 7.43232310e-01
5.51681936e-01 -4.87116486e-01 -4.33384269e-01 -4.61487859e-01
2.04301953e-01 8.72731686e-01 -9.07020688e-01 -5.08007109e-01
-2.98990309e-01 -3.34790617e-01 1.77699625e-02 3.26356500e-01
3.49425793e-01 -9.84409809e-01 -3.91253173e-01 -3.48587513e-01
-1.16508387e-01 -1.16385436e+00 -6.65583849e-01 -6.93144053e-02
-5.81541419e-01 -1.16449845e+00 -2.67418385e-01 -6.49424553e-01
1.28022122e+00 6.48326635e-01 9.99889851e-01 -1.12828270e-01
-5.47447860e-01 7.11453140e-01 -4.41720217e-01 -2.01043814e-01
-1.64342284e-01 -2.25161090e-01 -1.43773751e-02 1.96888939e-01
-2.23923326e-01 -8.14091086e-01 -8.05303395e-01 2.65870780e-01
-1.07813489e+00 5.15261471e-01 2.26325125e-01 6.98404789e-01
4.71372992e-01 1.02983071e-02 -8.88439864e-02 -9.86431003e-01
3.28792483e-01 -4.59868342e-01 -6.14418685e-01 2.31139570e-01
-2.11249337e-01 4.17975754e-01 1.71917021e-01 -6.72530234e-01
-8.75209868e-01 3.44431788e-01 7.10339606e-01 -8.33616674e-01
-8.82108435e-02 2.97281099e-03 -5.42809606e-01 -1.75342187e-02
7.15982080e-01 4.32916731e-03 -1.48707060e-02 -3.78147721e-01
9.97960269e-01 5.02295345e-02 2.60693252e-01 -7.46781051e-01
1.25151694e+00 6.71491623e-01 4.95057888e-02 -5.83529472e-01
-7.86864638e-01 -3.40797633e-01 -8.22039723e-01 -1.00584291e-01
1.01416218e+00 -8.83889139e-01 -6.52971148e-01 3.80389273e-01
-1.19260287e+00 -1.95375964e-01 -5.36830187e-01 2.35114068e-01
-6.52141273e-01 3.11996937e-01 -3.35051864e-01 -3.72535259e-01
3.10527176e-01 -1.44746041e+00 1.50896502e+00 1.31237209e-01
-1.25860870e-01 -8.27124953e-01 -2.23915711e-01 4.02868360e-01
1.24867387e-01 6.12918615e-01 9.93287146e-01 -1.19852960e-01
-1.21113753e+00 -1.99482158e-01 -2.08290681e-01 1.06374107e-01
6.06777132e-01 1.54261410e-01 -8.16861868e-01 -2.72867441e-01
-1.88836902e-01 -1.76413953e-01 7.71229029e-01 3.40290785e-01
1.42491174e+00 -4.30190653e-01 -3.92551661e-01 8.47686768e-01
1.18444633e+00 -2.28905022e-01 7.20895231e-01 -1.95283927e-02
1.33963883e+00 6.69783890e-01 4.14063603e-01 4.23387825e-01
4.65469807e-01 8.46447945e-01 7.73280978e-01 5.06007671e-02
-3.42861980e-01 -5.21875441e-01 2.71514624e-01 4.88837898e-01
9.11500528e-02 -3.75225514e-01 -5.66834569e-01 4.88387942e-01
-1.87187457e+00 -8.15019488e-01 -1.65509909e-01 2.06478286e+00
7.81552017e-01 -3.07307392e-01 -2.90061712e-01 -2.54587293e-01
7.80430198e-01 4.56199884e-01 -6.08226657e-01 -1.09336991e-02
7.26253241e-02 -1.95662230e-01 5.15746117e-01 3.93599004e-01
-8.71789932e-01 9.66069937e-01 5.77211332e+00 6.35050654e-01
-8.11307013e-01 -1.30455136e-01 3.05236161e-01 -2.67659426e-01
-6.84804440e-01 1.41608730e-01 -6.18044913e-01 1.94105759e-01
-4.83824015e-02 -2.63552099e-01 7.99852014e-01 7.65312076e-01
1.33068621e-01 4.09412943e-02 -1.50936401e+00 1.11389816e+00
2.62410164e-01 -1.34832692e+00 6.00810230e-01 7.92018026e-02
8.02581668e-01 -2.89289951e-01 3.84433061e-01 2.13471241e-02
5.75585783e-01 -8.54620159e-01 1.01703465e+00 5.79053640e-01
8.34144413e-01 -2.72256732e-01 -1.53194992e-02 2.74237990e-01
-1.03155220e+00 2.40206257e-01 -3.41823876e-01 -2.54684184e-02
6.00619093e-02 3.57159704e-01 -8.26847553e-01 4.68972206e-01
6.10323966e-01 7.49384582e-01 -8.29903364e-01 8.93618226e-01
-5.39067030e-01 1.72791556e-01 -1.43555954e-01 2.06901670e-01
-2.83143431e-01 -4.53670800e-01 8.99856925e-01 6.12765193e-01
-3.61944586e-02 5.59653528e-02 1.95901245e-01 1.47387159e+00
-1.54712930e-01 -2.11894006e-01 -9.59927082e-01 -2.64703482e-01
4.75988835e-01 1.30094302e+00 -5.78248680e-01 -2.50040710e-01
1.34313703e-02 1.06943166e+00 5.17460406e-01 6.46300077e-01
-9.63322461e-01 -6.26189075e-03 8.78144443e-01 3.26042682e-01
4.00447518e-01 -5.35669625e-01 -2.19593138e-01 -1.39889383e+00
2.00244874e-01 -7.16985822e-01 -7.09434003e-02 -1.10606062e+00
-1.34333622e+00 2.76428401e-01 6.36631012e-01 -1.10267234e+00
1.36453016e-02 -6.11639023e-01 -4.05974209e-01 8.20422590e-01
-9.64074254e-01 -1.39966953e+00 -7.24595368e-01 6.50474966e-01
4.93054509e-01 1.43436834e-01 7.10531414e-01 1.36482781e-02
-5.85696936e-01 4.36217934e-01 -2.48776793e-01 7.94711113e-02
5.69719672e-01 -1.23282111e+00 3.04658979e-01 6.60485685e-01
3.15427423e-01 8.84305239e-01 8.88218105e-01 -5.72340608e-01
-1.73201954e+00 -1.07433522e+00 1.81510493e-01 -8.91847193e-01
5.02205253e-01 -9.59556818e-01 -6.49789810e-01 9.84258175e-01
1.37840122e-01 3.24791014e-01 3.23819518e-01 4.49726805e-02
-6.47644699e-01 1.09811798e-01 -1.01000905e+00 1.04666889e+00
1.44306242e+00 -5.82711995e-01 -3.61135036e-01 5.13223231e-01
1.15642989e+00 -5.95531464e-01 -5.95109224e-01 3.25099498e-01
4.89255250e-01 -7.71829247e-01 1.21401107e+00 -6.50694668e-01
5.93764126e-01 -4.59747434e-01 -1.83870077e-01 -1.43602943e+00
-6.64252937e-01 -4.89209682e-01 9.11718328e-03 1.23490620e+00
1.68319643e-01 -6.50708735e-01 6.40047610e-01 8.63592565e-01
2.20177509e-02 -5.46539664e-01 -5.97738981e-01 -5.72231114e-01
-3.50146532e-01 -2.58713156e-01 7.14057863e-01 9.20761824e-01
-5.17478347e-01 3.30514789e-01 -2.97751069e-01 4.42389816e-01
8.25124681e-01 2.45898366e-01 1.15745556e+00 -9.70710158e-01
-3.35225582e-01 -2.78677106e-01 -8.19651604e-01 -1.12947202e+00
1.68412119e-01 -1.00026691e+00 -6.46924153e-02 -1.56158757e+00
6.26319826e-01 -4.83472526e-01 -5.31035522e-03 4.27217484e-01
-3.60321164e-01 -9.38273072e-02 3.89192849e-01 9.11736637e-02
-6.37689173e-01 9.82632697e-01 1.88966000e+00 -3.92801076e-01
-1.13768682e-01 -3.18333983e-01 -7.54253268e-01 6.26541197e-01
5.50295293e-01 -5.32832861e-01 -5.61021984e-01 -7.90360093e-01
1.83731914e-01 -1.25116721e-01 7.78282762e-01 -3.93150121e-01
-1.46951117e-02 -4.34502006e-01 2.61813015e-01 -3.01400244e-01
6.64186239e-01 -9.25118506e-01 3.97363454e-01 4.75715660e-02
-3.27964038e-01 -4.43732515e-02 1.25304550e-01 8.88467252e-01
2.09729120e-01 -5.62730283e-02 5.59893191e-01 -1.05442315e-01
-6.00051522e-01 5.91570079e-01 6.10015914e-03 -1.18361607e-01
1.19383550e+00 -3.54062855e-01 -4.12330598e-01 -4.12734479e-01
-6.50305092e-01 2.73340613e-01 9.44291532e-01 9.22177494e-01
6.25309825e-01 -1.51649666e+00 -4.58923340e-01 2.89150834e-01
4.77940768e-01 4.55171734e-01 2.59180248e-01 5.27951062e-01
-4.65038985e-01 -1.80906102e-01 -3.96581799e-01 -8.13426852e-01
-1.27684617e+00 7.87561297e-01 2.46201962e-01 7.90310800e-02
-5.41634560e-01 8.78205001e-01 9.82255816e-01 -4.36919332e-01
1.91539586e-01 -4.65702176e-01 8.86912365e-03 -2.67570734e-01
7.10489005e-02 1.42672777e-01 -6.06918275e-01 -6.35051250e-01
-1.08144909e-01 4.11866009e-01 -5.51182665e-02 8.54930952e-02
1.31649137e+00 -2.18327731e-01 -4.26235259e-01 5.86781204e-01
1.13473010e+00 1.62649125e-01 -1.47389460e+00 -1.36920676e-01
-5.35655618e-01 -9.89093721e-01 1.12316953e-02 -3.84609640e-01
-1.09322071e+00 6.31290913e-01 3.83769065e-01 1.65425241e-02
8.20595026e-01 2.76046664e-01 2.60226607e-01 2.28058666e-01
5.30808628e-01 -4.07496035e-01 4.79773998e-01 1.87549919e-01
1.38786006e+00 -1.27461898e+00 2.42193192e-01 -9.28844333e-01
-6.24294102e-01 6.80923998e-01 7.14826226e-01 -3.81509870e-01
7.00796247e-01 2.01611482e-02 -2.63884991e-01 -6.26774311e-01
-4.98245150e-01 6.43673632e-03 5.63586056e-01 6.97169006e-01
9.35240611e-02 4.10719484e-01 2.31991187e-01 1.37773260e-01
-2.40804121e-01 -6.11967385e-01 3.88578624e-01 7.36377776e-01
3.58396322e-02 -1.11503696e+00 -1.43142074e-01 3.99092376e-01
1.60935074e-01 1.42634949e-02 -4.42208678e-01 7.52368331e-01
1.97130814e-01 7.54499435e-01 9.76458564e-03 -3.02764416e-01
4.84622002e-01 -4.09646511e-01 9.75743234e-01 -1.19272745e+00
-1.12381235e-01 -2.50875860e-01 -1.27463266e-01 -8.59975576e-01
-2.71432787e-01 -6.46855533e-01 -9.76102829e-01 -5.29862270e-02
-3.97641838e-01 -1.69697925e-01 5.22150755e-01 6.49355471e-01
5.36139131e-01 7.05207407e-01 5.97433686e-01 -1.37051952e+00
-5.18011332e-01 -8.07025373e-01 -7.59351552e-01 1.01493716e+00
3.15662444e-01 -1.24447858e+00 -4.71289098e-01 2.96107531e-01] | [9.07448673248291, -3.1237783432006836] |
6016a9c2-73ae-4763-8826-5d9b735e0137 | smatch-standardized-and-extended-evaluation | 2305.06993 | null | https://arxiv.org/abs/2305.06993v1 | https://arxiv.org/pdf/2305.06993v1.pdf | SMATCH++: Standardized and Extended Evaluation of Semantic Graphs | The Smatch metric is a popular method for evaluating graph distances, as is necessary, for instance, to assess the performance of semantic graph parsing systems. However, we observe some issues in the metric that jeopardize meaningful evaluation. E.g., opaque pre-processing choices can affect results, and current graph-alignment solvers do not provide us with upper-bounds. Without upper-bounds, however, fair evaluation is not guaranteed. Furthermore, adaptions of Smatch for extended tasks (e.g., fine-grained semantic similarity) are spread out, and lack a unifying framework. For better inspection, we divide the metric into three modules: pre-processing, alignment, and scoring. Examining each module, we specify its goals and diagnose potential issues, for which we discuss and test mitigation strategies. For pre-processing, we show how to fully conform to annotation guidelines that allow structurally deviating but valid graphs. For safer and enhanced alignment, we show the feasibility of optimal alignment in a standard evaluation setup, and develop a lossless graph compression method that shrinks the search space and significantly increases efficiency. For improved scoring, we propose standardized and extended metric calculation of fine-grained sub-graph meaning aspects. Our code is available at https://github.com/flipz357/smatchpp | ['Juri Opitz'] | 2023-05-11 | null | null | null | null | ['semantic-textual-similarity', 'semantic-similarity'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.53248322e-01 2.68305272e-01 -1.22801133e-01 -6.18577242e-01
-7.75023878e-01 -9.22367811e-01 1.90300673e-01 6.28642678e-01
-4.07084465e-01 4.41517800e-01 2.49658436e-01 -6.03951275e-01
-4.57819939e-01 -8.07996571e-01 -6.17945313e-01 -2.66331673e-01
-6.50484115e-02 4.73473370e-01 4.87590492e-01 -2.95085665e-02
4.39508349e-01 2.00459778e-01 -1.52224958e+00 2.46381417e-01
9.27070260e-01 6.77163959e-01 1.50562376e-01 7.01326907e-01
-2.94911683e-01 3.04689318e-01 -7.15262234e-01 -9.64436471e-01
2.31920555e-01 -4.30042118e-01 -1.35220134e+00 -2.94319481e-01
6.39792264e-01 -2.53621697e-01 2.75074691e-02 1.23577809e+00
3.77773196e-01 -2.42824927e-02 3.02563846e-01 -1.49938953e+00
-3.41582537e-01 1.11940765e+00 -3.24895889e-01 1.69955060e-01
4.90833640e-01 1.27128080e-01 1.51936030e+00 -3.45528573e-01
8.48021090e-01 1.13256049e+00 7.13793874e-01 3.02218735e-01
-9.89487767e-01 -3.36273104e-01 1.51753426e-01 2.63707012e-01
-1.31677997e+00 -3.50692183e-01 4.49435115e-01 -1.60178840e-01
1.12413979e+00 7.49487460e-01 4.20074135e-01 8.07882786e-01
-1.25852019e-01 5.21459401e-01 7.54330516e-01 -4.72417444e-01
2.38732338e-01 -2.71458536e-01 5.34532964e-01 6.62743211e-01
6.63743794e-01 -1.32247597e-01 -4.70164388e-01 -3.16935569e-01
2.95746028e-01 -5.56800306e-01 -2.57794946e-01 -6.36496246e-01
-1.07037210e+00 8.02198529e-01 1.75252318e-01 3.89591575e-01
1.86444730e-01 -9.37464554e-03 7.22948492e-01 4.89120036e-01
2.08316267e-01 6.74918115e-01 -4.76865053e-01 -3.58988076e-01
-9.16692376e-01 3.30297083e-01 7.85489202e-01 1.16172564e+00
7.27255523e-01 -5.13149321e-01 -1.15067609e-01 7.72641122e-01
1.01561856e-03 -2.94502694e-02 2.28928953e-01 -1.18748689e+00
6.53632164e-01 6.22396648e-01 -3.72922897e-01 -1.04263127e+00
-5.27568460e-01 -3.06888819e-01 -4.02054906e-01 7.30925798e-02
7.47121572e-01 1.70048445e-01 -5.72873354e-01 1.97543669e+00
2.98011839e-01 -9.12134126e-02 -7.30266348e-02 8.83227527e-01
6.63437128e-01 9.69020128e-02 9.35084000e-02 -5.32270186e-02
1.62950778e+00 -8.72444391e-01 -4.76797432e-01 -4.38927412e-01
1.33909202e+00 -7.69882441e-01 1.42877269e+00 2.32822448e-01
-1.19601464e+00 -7.48065636e-02 -1.16416264e+00 -3.98519456e-01
-3.14111561e-01 -4.02516216e-01 7.67954409e-01 8.61369073e-01
-1.10054374e+00 9.50885117e-01 -8.96735728e-01 -6.37120903e-01
1.93941295e-02 2.37718761e-01 -4.96126622e-01 -9.31243896e-02
-1.11512327e+00 8.94548833e-01 5.65372467e-01 -3.21877539e-01
-4.74123843e-02 -5.72357833e-01 -1.01136935e+00 2.27323785e-01
6.54511154e-01 -7.65443027e-01 1.24403608e+00 -5.68376184e-01
-1.07828808e+00 1.03351688e+00 4.30369936e-02 -4.33989167e-01
2.92155921e-01 1.95591222e-03 -1.87467501e-01 1.32358953e-01
1.26454473e-01 5.64138591e-01 2.98909426e-01 -7.90926099e-01
-4.80932325e-01 -3.74185205e-01 4.53135371e-01 3.65116835e-01
-4.45747912e-01 2.29985625e-01 -6.89715683e-01 -5.14891982e-01
1.75996214e-01 -9.06444967e-01 -7.79596791e-02 -2.60809690e-01
-4.69666511e-01 5.49844885e-03 3.16853255e-01 -6.79833353e-01
1.57166851e+00 -2.19194126e+00 -8.71552620e-03 2.93837368e-01
3.99231493e-01 1.53106198e-01 -3.15730959e-01 4.40448910e-01
1.67251043e-02 5.58957458e-01 -5.31233668e-01 -2.72390008e-01
2.84601718e-01 2.34989852e-01 1.34091079e-01 2.66304314e-01
5.30663840e-02 8.77962947e-01 -9.13018882e-01 -6.78404212e-01
-2.42601358e-03 1.56451598e-01 -9.29458380e-01 1.20372344e-02
-1.76930770e-01 -5.02777100e-02 -2.18329534e-01 6.06096566e-01
6.78794920e-01 -4.08977121e-01 4.89803731e-01 -3.04875374e-01
1.87671453e-01 9.54405010e-01 -1.32519221e+00 1.91282952e+00
-1.70886144e-01 3.70506614e-01 4.53411341e-02 -9.21949446e-01
6.66751206e-01 -2.81699270e-01 3.03069204e-01 -6.63902879e-01
-3.23021300e-02 3.29789370e-01 1.16241239e-01 -1.72344685e-01
8.91470134e-01 2.68267661e-01 -1.68609619e-01 5.72782338e-01
-6.20829351e-02 -1.46868676e-01 6.07274890e-01 4.05163944e-01
1.77242279e+00 1.20203428e-01 4.38262969e-01 -3.43425065e-01
1.93687990e-01 1.27490237e-01 4.57074493e-01 6.21965051e-01
-1.90098271e-01 8.06965590e-01 7.71079302e-01 -1.14852257e-01
-9.76054668e-01 -7.83968508e-01 2.28795931e-02 1.23985231e+00
2.62973309e-01 -1.13485587e+00 -9.99055505e-01 -9.37068999e-01
-1.36423260e-01 7.52156317e-01 -2.89636552e-01 -2.56092489e-01
-6.66348815e-01 -6.99342966e-01 9.03322816e-01 4.90150839e-01
3.92495990e-02 -7.69046068e-01 -8.05400729e-01 1.03067353e-01
-3.77384990e-01 -1.26388943e+00 -5.61550498e-01 1.11978665e-01
-8.68166268e-01 -1.31953335e+00 -1.28731549e-01 -6.79749370e-01
5.51701605e-01 3.79553169e-01 1.55467558e+00 6.40598416e-01
-1.66609883e-01 3.32004964e-01 -5.63297749e-01 1.05799817e-01
-6.35924160e-01 2.66077518e-01 -4.34683532e-01 -8.05292904e-01
3.47482949e-01 -5.95543087e-01 -5.35311818e-01 3.80584657e-01
-8.84476244e-01 1.40811071e-01 4.62844640e-01 4.96172249e-01
6.11672461e-01 -4.31017354e-02 1.25043884e-01 -1.17251933e+00
7.44095385e-01 -2.15731278e-01 -7.14669883e-01 4.20325756e-01
-1.02830911e+00 4.15401608e-01 4.52284843e-01 1.01570182e-01
-4.31523353e-01 -2.51505494e-01 -3.32962573e-01 -1.65650416e-02
-1.39677048e-01 5.72666764e-01 -2.24905625e-01 -1.14838623e-01
6.72463596e-01 -1.98872417e-01 1.83831379e-02 -4.07294929e-01
5.59829175e-01 4.00382727e-01 6.17512584e-01 -9.44189072e-01
6.14794075e-01 1.36485159e-01 -1.49538398e-01 -2.37160251e-01
-5.75530052e-01 -5.23872733e-01 -3.43792558e-01 2.41671726e-01
6.62809730e-01 -5.33297837e-01 -5.68539977e-01 3.52918021e-02
-1.11324871e+00 -3.60672444e-01 -2.52323031e-01 3.04269493e-01
-6.25689685e-01 8.68077099e-01 -7.06788182e-01 -2.95955598e-01
-4.39403236e-01 -1.27122176e+00 1.05431426e+00 -2.04157054e-01
-6.90241098e-01 -9.17407572e-01 -2.60993671e-02 5.30350804e-01
3.51582170e-01 1.43107787e-01 1.18938744e+00 -1.04376841e+00
-4.12130564e-01 1.74362406e-01 -3.08472872e-01 2.01953799e-01
-1.33974001e-01 2.72204038e-02 -7.16735721e-01 -3.31227541e-01
-2.49440312e-01 -1.64262444e-01 5.70586205e-01 4.34367210e-02
1.20940161e+00 -3.31364840e-01 -2.79115200e-01 7.14043438e-01
1.27599335e+00 -1.57936186e-01 5.84039807e-01 6.69720531e-01
6.80547178e-01 7.33922422e-01 7.35366821e-01 3.46050769e-01
6.46809757e-01 6.99616671e-01 5.02168477e-01 1.37410656e-01
-1.93211138e-01 -1.19906515e-01 2.27017060e-01 9.47774649e-01
-3.05593405e-02 -4.23374623e-01 -9.28184986e-01 3.77020717e-01
-1.62994015e+00 -7.27055073e-01 -1.85291320e-01 2.44926834e+00
6.85881019e-01 4.77083832e-01 2.40944132e-01 3.67344201e-01
7.81133950e-01 1.23998985e-01 -3.56996983e-01 -7.11634815e-01
-1.20417558e-01 3.98461312e-01 7.26263463e-01 7.57490218e-01
-5.85424423e-01 1.02156806e+00 5.93485785e+00 6.94757700e-01
-7.38546193e-01 6.97920248e-02 3.93659264e-01 1.16327628e-02
-7.57346094e-01 3.90558749e-01 -5.81649363e-01 2.44831666e-01
9.33020294e-01 -4.07365471e-01 5.94530821e-01 8.22041690e-01
-2.26184934e-01 2.13294804e-01 -1.31748259e+00 7.47035086e-01
-7.94873014e-02 -1.15572500e+00 -1.51223123e-01 5.23988977e-02
1.13960542e-02 1.41687542e-01 -2.98716843e-01 2.88002610e-01
3.66622567e-01 -8.09978604e-01 7.57986248e-01 -2.14478195e-01
6.82876825e-01 -5.75605869e-01 6.52298152e-01 -1.23125911e-01
-1.24704444e+00 2.30713442e-01 -4.31265801e-01 3.04119196e-02
6.33337945e-02 6.39999926e-01 -7.18882680e-01 8.52348447e-01
5.77106833e-01 3.88273925e-01 -8.87080431e-01 9.61600959e-01
-2.29328796e-01 5.22888541e-01 -3.89850438e-01 1.08912610e-01
1.01891391e-01 -2.01736093e-01 5.86372077e-01 1.33478224e+00
5.21017969e-01 -7.36658946e-02 4.87745889e-02 6.34788156e-01
-8.02410021e-02 3.61461759e-01 -5.00177264e-01 -1.30316064e-01
9.20155585e-01 1.24788761e+00 -9.72690701e-01 -1.75390556e-01
-5.22368729e-01 9.43627894e-01 7.14094639e-01 2.71364637e-02
-9.87650931e-01 -4.00322080e-01 7.89729416e-01 2.24693358e-01
1.52873382e-01 -1.44232035e-01 -6.13937795e-01 -1.06705773e+00
3.28879148e-01 -1.25357330e+00 9.70725536e-01 -5.73820174e-01
-1.00229919e+00 5.84570110e-01 2.29932666e-01 -8.03958595e-01
-4.67956662e-01 -5.79728186e-01 -3.37602854e-01 4.35248584e-01
-1.24167907e+00 -7.70299911e-01 -4.25495803e-01 3.65665317e-01
1.65478215e-01 3.29523265e-01 7.51293898e-01 3.31082940e-01
-2.57226020e-01 1.10426307e+00 -4.96460617e-01 -7.59585649e-02
7.86073565e-01 -1.33634984e+00 9.98077154e-01 1.12512469e+00
2.84299135e-01 6.26499832e-01 9.56624627e-01 -4.38609302e-01
-1.43820107e+00 -8.83261204e-01 9.66481388e-01 -2.79371262e-01
7.54630208e-01 -3.77941757e-01 -1.11878479e+00 6.59995079e-01
-1.72766849e-01 -1.25194758e-01 5.38039148e-01 5.12860119e-01
-7.11928666e-01 1.07871324e-01 -1.04976761e+00 7.63782740e-01
1.67793560e+00 -4.87220585e-01 -4.85301018e-01 3.54548395e-01
1.02774239e+00 -7.11782098e-01 -1.15737927e+00 5.41024506e-01
4.05244857e-01 -1.24042118e+00 8.40005457e-01 -4.28063422e-01
1.74940094e-01 -1.42152548e-01 -2.65657723e-01 -1.07255495e+00
-4.17533755e-01 -7.22183287e-01 8.15587714e-02 1.28234220e+00
6.05560899e-01 -7.77264476e-01 9.35702980e-01 8.50735784e-01
-4.54172999e-01 -6.46327078e-01 -8.09610963e-01 -1.05185628e+00
9.14086029e-02 -7.10636616e-01 1.00031233e+00 1.12707734e+00
4.60143119e-01 3.42823297e-01 6.68812245e-02 3.30310315e-01
5.20434558e-01 1.48850441e-01 6.87050343e-01 -9.02053833e-01
-5.01118302e-01 -8.36866975e-01 -6.42898321e-01 -7.27667809e-01
9.63633582e-02 -1.14513493e+00 -2.85701931e-01 -1.79539526e+00
2.27832228e-01 -6.70124292e-01 -3.47434610e-01 7.43543327e-01
-1.98642716e-01 7.60077983e-02 2.29733199e-01 3.45767140e-02
-7.81116962e-01 -9.53592584e-02 8.71881783e-01 1.50481313e-01
2.19726674e-02 -4.22077537e-01 -1.00776851e+00 4.56830412e-01
1.18093348e+00 -5.75479388e-01 -5.71130574e-01 -7.01787293e-01
5.30212045e-01 -1.40245602e-01 2.72132188e-01 -1.08685505e+00
1.62486777e-01 -4.70990576e-02 -4.63950723e-01 -7.51415119e-02
-4.22957391e-02 -4.61003572e-01 2.58564413e-01 4.33368713e-01
-4.52536762e-01 4.32484984e-01 5.90189174e-02 5.65358587e-02
-1.45265177e-01 -5.04251957e-01 4.89813060e-01 3.04869302e-02
-6.78720355e-01 2.55014718e-01 2.80932099e-01 4.71479297e-01
6.34974480e-01 -2.55618244e-01 -7.78511643e-01 -1.62677795e-01
-4.10427272e-01 2.14607164e-01 1.00970626e+00 4.25791085e-01
2.17921466e-01 -1.09998047e+00 -6.44354343e-01 -7.97647536e-02
4.00154144e-01 -6.66913539e-02 9.44159776e-02 7.44481266e-01
-5.92900097e-01 1.65850833e-01 -1.34508863e-01 -3.04194123e-01
-1.53613424e+00 6.52826369e-01 5.35674673e-03 -4.87075865e-01
-7.22506940e-01 5.66894770e-01 1.27780765e-01 -4.66741174e-01
2.07551524e-01 -7.09454954e-01 -2.22349167e-02 -2.13345781e-01
3.52275282e-01 3.55816960e-01 6.77007735e-01 -4.09427285e-01
-6.29987895e-01 4.11970794e-01 -3.72259617e-02 -5.85645996e-02
1.13646185e+00 -2.03757927e-01 -2.74825305e-01 -7.80612454e-02
1.08232391e+00 1.41016349e-01 -7.29942918e-01 -5.88438995e-02
3.64118159e-01 -3.13903809e-01 -2.63694167e-01 -7.34300375e-01
-8.97375286e-01 5.42204857e-01 1.96484670e-01 4.66703057e-01
1.28463697e+00 5.19971773e-02 9.10893738e-01 4.39099908e-01
4.16616261e-01 -1.01045966e+00 -4.77349877e-01 4.62489009e-01
5.25612235e-01 -9.16202545e-01 1.21744893e-01 -9.65369463e-01
-4.13516790e-01 9.75129724e-01 6.04518175e-01 3.26720417e-01
1.62604049e-01 5.12917042e-01 -1.06787473e-01 -3.28656107e-01
-8.07824671e-01 -2.67891198e-01 1.48991033e-01 4.66605157e-01
6.21006846e-01 1.54312536e-01 -7.76157498e-01 4.74273354e-01
-8.70917261e-01 -3.81915599e-01 5.39709210e-01 9.44479167e-01
-3.81451309e-01 -1.42329836e+00 -1.67636648e-01 3.72049689e-01
-5.32465219e-01 -3.85985494e-01 -4.56805289e-01 9.28306222e-01
-2.06329927e-01 9.65953469e-01 4.44234610e-02 -5.67706227e-01
3.33185583e-01 1.17368670e-02 7.03182876e-01 -5.38226128e-01
-4.94098157e-01 -4.03773040e-01 7.72281229e-01 -8.85767102e-01
-1.39314368e-01 -5.37425280e-01 -1.43266273e+00 -6.12434208e-01
-2.72985905e-01 2.70815849e-01 6.88715100e-01 8.53067398e-01
6.98461354e-01 2.96300113e-01 4.07902012e-03 -1.94429681e-01
-5.02157331e-01 -7.11229086e-01 -1.91774607e-01 6.55769289e-01
-2.78873295e-01 -3.84035468e-01 -5.39428055e-01 -1.90444887e-01] | [9.489320755004883, 8.140239715576172] |
ca9c52fc-a341-47ca-8f50-0b1c20d62b1c | timebankpt-a-timeml-annotated-corpus-of | null | null | https://aclanthology.org/L12-1096 | https://aclanthology.org/L12-1096.pdf | TimeBankPT: A TimeML Annotated Corpus of Portuguese | In this paper, we introduce TimeBankPT, a TimeML annotated corpus of Portuguese. It has been produced by adapting an existing resource for English, namely the data used in the first TempEval challenge. TimeBankPT is the first corpus of Portuguese with rich temporal annotations (i.e. it includes annotations not only of temporal expressions but also about events and temporal relations). In addition, it was subjected to an automated error mining procedure that checks the consistency of the annotated temporal relations based on their logical properties. This procedure allowed for the detection of some errors in the annotations, that also affect the original English corpus. The Portuguese language is currently undergoing a spelling reform, and several countries where Portuguese is official are in a transitional period where old and new orthographies are valid. TimeBankPT adopts the recent spelling reform. This decision is to preserve its future usefulness. TimeBankPT is freely available for download. | ["Ant{\\'o}nio Branco", 'Francisco Costa'] | 2012-05-01 | null | null | null | lrec-2012-5 | ['temporal-information-extraction'] | ['natural-language-processing'] | [-2.94290334e-02 4.76242632e-01 -2.88003802e-01 -1.17890127e-01
-3.68900746e-01 -7.83311725e-01 9.24615502e-01 6.56902194e-01
-7.72277653e-01 1.10946786e+00 5.01729369e-01 -3.38469267e-01
-2.10582450e-01 -6.01408422e-01 -2.96416581e-01 -1.50166601e-01
-2.80193895e-01 7.46262789e-01 6.76578403e-01 -2.22498074e-01
2.58734345e-01 4.75077420e-01 -1.42040169e+00 4.65939969e-01
6.99904442e-01 5.88361025e-01 1.75911009e-01 4.81624395e-01
-1.14633583e-01 1.00587547e+00 -7.13517010e-01 -4.61300790e-01
-8.71707499e-03 -3.91584933e-01 -1.42217338e+00 -3.32046092e-01
-4.65347409e-01 1.71885714e-01 -2.12698370e-01 7.06226408e-01
1.10438518e-01 2.73331702e-01 1.31236240e-01 -1.22103751e+00
-1.30393565e-01 8.19061279e-01 -3.02384770e-03 3.14085066e-01
6.71242177e-01 -1.12977579e-01 1.00472999e+00 -5.95389724e-01
1.29082549e+00 9.59595025e-01 5.69727004e-01 1.83316767e-01
-7.71396816e-01 -2.40652964e-01 -1.02327742e-01 4.89647120e-01
-1.19961512e+00 -4.17048246e-01 3.58668804e-01 -5.12478471e-01
1.57215774e+00 2.23234758e-01 8.18083823e-01 9.57642078e-01
3.49022388e-01 2.44877547e-01 1.14074731e+00 -9.79563355e-01
3.82150739e-01 -1.71635643e-01 -1.93702467e-02 2.74166137e-01
-6.41167723e-03 1.74394511e-02 -6.69924378e-01 -1.46128893e-01
3.24153215e-01 -5.95310867e-01 -6.10892661e-02 2.11205348e-01
-1.20944071e+00 3.02600771e-01 -4.30889398e-01 1.07612038e+00
-1.71614036e-01 -5.55567630e-02 9.35605407e-01 3.48837733e-01
3.43591720e-01 3.05450201e-01 -7.53586888e-01 -8.59187007e-01
-7.91742027e-01 1.94205508e-01 1.01849079e+00 8.58024657e-01
2.62693912e-01 -2.94644266e-01 1.15214631e-01 6.11700833e-01
3.65217119e-01 3.69115286e-02 6.33185506e-01 -1.06026137e+00
5.75170159e-01 8.06738615e-01 5.11889756e-01 -5.55879712e-01
-5.28675735e-01 1.63617432e-01 -2.01563388e-01 -2.00123921e-01
6.89110041e-01 2.36978885e-02 -5.57084382e-01 1.52113450e+00
3.08984816e-01 -9.05556530e-02 2.42334291e-01 3.23950291e-01
4.76120234e-01 5.56289852e-01 1.10296451e-01 -8.83935452e-01
1.33309209e+00 -5.34526765e-01 -1.35254192e+00 -1.41906496e-02
8.30040038e-01 -8.03760827e-01 6.99054778e-01 5.75752616e-01
-1.15815949e+00 -8.60551149e-02 -9.60436046e-01 -6.68130443e-02
-7.32014060e-01 -1.13379039e-01 4.05037045e-01 4.72939700e-01
-7.45566666e-01 7.99776435e-01 -1.22000277e+00 -8.73589277e-01
-1.33047491e-01 1.17474005e-01 -5.30514657e-01 3.12964141e-01
-1.42085254e+00 1.22850323e+00 9.26063478e-01 1.30048186e-01
-3.67566407e-01 -2.81920731e-01 -9.07953262e-01 -3.34657967e-01
6.32190347e-01 2.73903847e-01 1.60818160e+00 -7.35286772e-01
-1.35111260e+00 1.00196481e+00 -2.14627594e-01 -7.88433611e-01
6.99963152e-01 -1.20849527e-01 -1.27107894e+00 1.22853480e-01
5.19420683e-01 6.90883547e-02 1.40598103e-01 -5.36694527e-01
-9.00732279e-01 -3.90417844e-01 8.32910687e-02 -2.87145853e-01
-2.59221587e-02 8.16514850e-01 -6.42760456e-01 -7.82613337e-01
-3.11679989e-02 -9.13464844e-01 -1.19047128e-01 -4.84184355e-01
-1.26747459e-01 -4.94734138e-01 5.35055459e-01 -9.55565453e-01
1.95443130e+00 -2.09494090e+00 -7.31854960e-02 6.17433861e-02
-4.48705584e-01 1.21513829e-01 3.40352118e-01 1.00009310e+00
-2.14127153e-01 4.45906907e-01 -4.91045982e-01 -8.52027312e-02
1.32943660e-01 8.74646068e-01 -2.34510124e-01 5.18871605e-01
1.07915759e-01 5.00284553e-01 -1.06272805e+00 -8.23446631e-01
1.74470257e-03 -2.05630921e-02 -1.57441959e-01 -1.55303448e-01
-5.28051674e-01 5.25554478e-01 -3.19684356e-01 5.27739346e-01
2.49862328e-01 2.73010343e-01 7.23627687e-01 1.59074634e-01
-1.05566585e+00 8.72486293e-01 -1.12431157e+00 1.97525847e+00
-3.40119362e-01 4.59566981e-01 -4.70319897e-01 -4.71807271e-01
8.49052727e-01 9.23148215e-01 9.32147324e-01 -9.31775630e-01
7.21335411e-02 6.43853784e-01 -5.98595478e-02 -7.36917436e-01
8.40095520e-01 6.57498986e-02 -3.92302901e-01 5.47549844e-01
-1.15511090e-01 -2.59725183e-01 1.05527043e+00 1.51798084e-01
1.03564811e+00 8.17708850e-01 9.21403110e-01 -3.10617715e-01
7.40626633e-01 3.51138800e-01 1.07073295e+00 -5.21656796e-02
-2.05686063e-01 2.03489706e-01 7.18264282e-01 -5.60987592e-01
-1.24883997e+00 -6.19625211e-01 -3.36790085e-01 4.96659487e-01
-3.45660895e-01 -1.11117911e+00 -2.55940676e-01 -6.80903316e-01
-6.61961854e-01 8.68172467e-01 -5.13540268e-01 3.51740956e-01
-9.81182516e-01 -3.11696708e-01 8.45354795e-01 3.70861739e-01
3.00729245e-01 -1.34083271e+00 -1.06011319e+00 5.69560647e-01
-4.91444319e-01 -1.35816479e+00 -1.71834722e-01 3.88510138e-01
-6.44298673e-01 -1.32135642e+00 6.12987727e-02 -3.63307029e-01
1.24173164e-01 -8.10955107e-01 9.85100865e-01 -3.95536348e-02
2.66068757e-01 8.35666955e-02 -7.67886519e-01 -7.59085715e-01
-6.14900351e-01 3.39276046e-02 -9.10609588e-02 -3.67797166e-01
2.62637734e-01 -4.95478511e-01 1.46525994e-01 1.73696324e-01
-1.00771856e+00 -2.55530059e-01 -2.04374820e-01 4.60079402e-01
4.49112892e-01 3.51618707e-01 4.43030924e-01 -7.75198817e-01
2.57472903e-01 -4.89667267e-01 -7.33027756e-01 3.65376413e-01
-6.08689070e-01 -3.50467078e-02 3.72836143e-01 1.62287410e-02
-1.39217854e+00 -1.18007101e-01 -2.08851427e-01 3.86534005e-01
-7.91243538e-02 8.57831836e-01 -1.11313621e-02 5.86390555e-01
4.60839599e-01 -2.39270359e-01 -6.64182127e-01 -5.12641490e-01
3.08814552e-02 4.14162457e-01 7.24693835e-01 -7.84224749e-01
3.47113192e-01 4.60203081e-01 -1.57819912e-02 -6.92954361e-01
-4.35909152e-01 -4.74971682e-01 -1.24527800e+00 -3.44460994e-01
8.52583051e-01 -5.88367760e-01 -2.96306640e-01 3.21646869e-01
-1.36076117e+00 -5.01067579e-01 -5.70101619e-01 5.80175102e-01
-5.33026218e-01 4.44633603e-01 -6.25243843e-01 -8.99040163e-01
1.68141067e-01 -5.90344369e-01 5.71717203e-01 -7.45385736e-02
-8.59022677e-01 -1.03455150e+00 3.96930009e-01 -1.90044001e-01
-2.90641516e-01 5.11713505e-01 6.94509625e-01 -7.38604069e-01
8.76514316e-02 6.01647943e-02 3.05577278e-01 -1.59526080e-01
1.13342106e-01 5.94371140e-01 -5.44762790e-01 4.14027907e-02
-9.58787184e-03 -3.30637768e-02 1.74737751e-01 -2.15379700e-01
3.33998263e-01 -4.13311690e-01 -2.84060091e-01 -1.50412649e-01
1.34667838e+00 9.40708995e-01 8.19984913e-01 9.20330584e-01
-4.18756604e-02 8.31710279e-01 1.19184518e+00 7.52040088e-01
5.10683894e-01 9.98397171e-01 1.40506938e-01 5.11533618e-01
5.04508354e-02 -1.89840138e-01 6.23376250e-01 1.09991789e+00
-2.26663306e-01 -1.45809785e-01 -1.32506967e+00 1.12173378e+00
-2.03295183e+00 -8.32395673e-01 -7.83668280e-01 2.22036910e+00
1.06656432e+00 3.01118404e-01 -7.60685792e-03 5.38163781e-01
4.16742712e-01 8.86450931e-02 2.33038411e-01 -8.82557452e-01
-1.56703249e-01 5.78985929e-01 3.46555531e-01 5.57622671e-01
-9.17626977e-01 9.43782151e-01 5.90404892e+00 3.86445224e-01
-7.55777001e-01 3.87888551e-01 -1.79511219e-01 1.72201902e-01
-2.27650434e-01 6.51713133e-01 -4.78490651e-01 4.22167033e-01
1.28619301e+00 -4.28922325e-01 1.43076047e-01 2.12369174e-01
7.39331126e-01 -5.06959438e-01 -8.90937865e-01 4.04368997e-01
-2.57121503e-01 -1.08546126e+00 -5.65726697e-01 -1.02519721e-01
6.51526570e-01 6.99888542e-02 -7.21916914e-01 1.15629174e-01
3.34976912e-01 -5.04902780e-01 1.55845547e+00 5.91969609e-01
7.45413125e-01 -7.40546942e-01 8.52887750e-01 2.48585865e-01
-1.38782704e+00 -3.37217972e-02 2.19295695e-01 -3.48641366e-01
6.28216326e-01 5.45982897e-01 -7.76788831e-01 1.07695985e+00
9.22252834e-01 9.54813421e-01 -4.12101299e-01 9.06823277e-01
-6.75114870e-01 4.36274648e-01 -3.41503054e-01 3.11179996e-01
1.28780127e-01 -1.98334560e-01 7.41177499e-01 1.32476258e+00
5.76338172e-01 -2.25450471e-02 1.23955704e-01 3.19419831e-01
3.45951170e-01 8.91145393e-02 -4.43551958e-01 -2.22479850e-01
7.32919693e-01 6.48751378e-01 -9.56604660e-01 -3.19952786e-01
-3.34113657e-01 8.20437133e-01 -9.72018018e-02 1.41845778e-01
-8.68828356e-01 -3.01936239e-01 7.10073635e-02 1.57481596e-01
2.28751808e-01 -5.18994153e-01 6.06116205e-02 -1.12185884e+00
3.47081065e-01 -6.69954181e-01 8.35566282e-01 -9.21030879e-01
-5.45966446e-01 6.76665783e-01 3.22292000e-01 -1.16044950e+00
-4.61452752e-01 -3.72761846e-01 -3.26390058e-01 5.24404109e-01
-1.15221524e+00 -1.13281608e+00 4.00580287e-01 4.49755371e-01
3.53966147e-01 3.50566983e-01 9.65215802e-01 5.17144382e-01
-3.78932744e-01 -1.04735523e-01 -4.08609360e-01 -1.75653677e-02
9.47928011e-01 -1.24809945e+00 2.69114286e-01 1.18422759e+00
-2.93774940e-02 4.78930593e-01 1.10105562e+00 -1.02297843e+00
-7.85743177e-01 -1.00533843e+00 2.22363162e+00 -3.45912218e-01
1.13717377e+00 -2.25762017e-02 -7.66259134e-01 1.21306574e+00
6.08658552e-01 -2.85325736e-01 2.92243391e-01 -2.99341440e-01
-1.55964658e-01 1.11430518e-01 -1.01043463e+00 4.84076262e-01
9.27338600e-01 -7.09265709e-01 -1.03414690e+00 3.01815212e-01
6.70765102e-01 -5.09792745e-01 -1.28342891e+00 4.38176721e-01
5.20082355e-01 -7.66741276e-01 1.19933762e-01 -3.65036726e-01
6.32278845e-02 -7.45484293e-01 -1.14696555e-01 -9.62735713e-01
1.19108498e-01 -1.13382065e+00 -1.38203343e-02 1.79306138e+00
5.78727186e-01 -7.03356147e-01 1.76459059e-01 4.69186157e-01
-5.48971832e-01 -1.04940556e-01 -1.46762407e+00 -9.62827027e-01
-2.15503216e-01 -9.43621159e-01 6.63673282e-01 1.27662194e+00
6.63915753e-01 -8.14352334e-02 -1.96599662e-01 -8.69004205e-02
-9.34506431e-02 -1.87120140e-01 4.12349105e-01 -1.32199144e+00
-3.38161252e-02 -3.01501513e-01 -7.30476901e-02 -1.10835366e-01
1.42128184e-01 -7.87807047e-01 -8.83238912e-02 -1.54430032e+00
-4.19779420e-01 -3.04316133e-01 2.24967435e-01 9.20326293e-01
5.49814582e-01 -1.20228134e-01 1.10142082e-01 2.55560637e-01
-4.90435988e-01 1.77030861e-01 8.37331176e-01 8.41884837e-02
-3.46194685e-01 -2.06554770e-01 6.57433867e-02 8.51676881e-01
8.15764308e-01 -8.14308345e-01 -8.16252977e-02 -2.47094810e-01
8.30338955e-01 2.44955406e-01 -3.71021889e-02 -7.89359510e-01
3.78830850e-01 -3.92910719e-01 -4.64778990e-01 -7.63728678e-01
2.99611147e-02 -1.10115385e+00 9.28715050e-01 7.60179877e-01
-7.30295107e-02 6.01673484e-01 4.87730116e-01 1.32212043e-01
-5.37081361e-01 -5.66580713e-01 5.22083402e-01 -2.77628124e-01
-1.02633059e+00 -1.12426497e-01 -8.31037641e-01 -1.64271761e-02
1.24144602e+00 -1.73397630e-01 -3.77156794e-01 2.28257552e-01
-9.60465431e-01 1.42766535e-02 4.84637946e-01 3.42653692e-01
1.09658241e-02 -1.23062861e+00 -3.87670368e-01 -4.04111445e-01
2.58039445e-01 -2.31329620e-01 -7.78078660e-02 1.09816468e+00
-7.46126354e-01 4.50416952e-01 -2.69164532e-01 -4.99218740e-02
-1.19341695e+00 6.33260727e-01 1.81461141e-01 -7.62926280e-01
-6.61286712e-01 -4.63479348e-02 -9.34569001e-01 -2.78982997e-01
-7.90529177e-02 -7.71783531e-01 -6.34708285e-01 4.39946979e-01
3.88040811e-01 4.71734494e-01 4.42904174e-01 -8.37331772e-01
-6.50752604e-01 3.07200074e-01 5.09407282e-01 -7.29091704e-01
1.49801958e+00 -2.37235636e-01 -8.17853570e-01 1.00985324e+00
6.20141625e-01 6.87752187e-01 -5.37965715e-01 3.24363522e-02
9.27791715e-01 -1.97629333e-01 -4.91819441e-01 -9.01256502e-01
-3.32274824e-01 2.43143201e-01 6.36129305e-02 5.16809165e-01
1.08513355e+00 -1.19236104e-01 5.55062771e-01 2.15065237e-02
6.14210725e-01 -1.39933276e+00 -4.42913502e-01 1.11112630e+00
9.07245457e-01 -5.65153360e-01 1.31820753e-01 -5.07347584e-01
-3.14174980e-01 1.25747931e+00 2.96532929e-01 4.57146823e-01
4.45674866e-01 3.97862554e-01 -2.37957478e-01 -2.75041670e-01
-1.00835860e+00 -3.12184870e-01 -4.24273685e-02 4.27067995e-01
6.41600788e-01 1.07538596e-01 -1.51532078e+00 7.99451172e-01
-3.27889949e-01 3.91865253e-01 7.86004007e-01 1.29705787e+00
1.25609770e-01 -1.67992008e+00 -4.85792696e-01 -3.23204279e-01
-8.07092130e-01 -5.13505936e-02 -6.81235075e-01 1.32092690e+00
6.34550691e-01 1.11924660e+00 1.90880015e-01 -5.17275743e-02
5.72405756e-01 1.90155774e-01 4.93926108e-01 -5.29248834e-01
-9.01770353e-01 3.71532626e-02 9.04323757e-01 -6.07917130e-01
-8.25319469e-01 -1.06654036e+00 -1.87244260e+00 -6.36258051e-02
1.54584870e-02 5.98241985e-01 5.69552958e-01 1.09930730e+00
-1.23836078e-01 5.68231285e-01 1.57039657e-01 -2.62784660e-01
2.53171444e-01 -9.46853578e-01 -4.16073054e-01 1.64505914e-01
-1.59073338e-01 -5.33605695e-01 -2.49883700e-02 3.64119172e-01] | [9.256660461425781, 9.265345573425293] |
c2a94690-1639-4feb-9fef-aae325020da8 | towards-a-better-understanding-of-the-4 | 2305.06773 | null | https://arxiv.org/abs/2305.06773v1 | https://arxiv.org/pdf/2305.06773v1.pdf | Towards a Better Understanding of the Computer Vision Research Community in Africa | Computer vision is a broad field of study that encompasses different tasks (e.g., object detection, semantic segmentation, 3D reconstruction). Although computer vision is relevant to the African communities in various applications, yet computer vision research is under-explored in the continent and constructs only 0.06% of top-tier publications in the last 10 years. In this paper, our goal is to have a better understanding of the computer vision research conducted in Africa and provide pointers on whether there is equity in research or not. We do this through an empirical analysis of the African computer vision publications that are Scopus indexed. We first study the opportunities available for African institutions to publish in top-tier computer vision venues. We show that African publishing trends in top-tier venues over the years do not exhibit consistent growth. We also devise a novel way to retrieve African authors through their affiliation history to have a better understanding of their contributions in top-tier venues. Moreover, we study all computer vision publications beyond top-tier venues in different African regions to find that mainly Northern and Southern Africa are publishing in computer vision with more than 85% of African publications. Finally, we present the most recurring keywords in computer vision publications. In summary, our analysis reveals that African researchers are key contributors to African research, yet there exists multiple barriers to publish in top-tier venues and the current trend of topics published in the continent might not necessarily reflect the communities' needs. This work is part of a community based effort that is focused on improving computer vision research in Africa. | ['Mennatullah Siam', "Ro'ya-CV4Africa", 'Karim Gamal', 'Yvan Pimi', 'Abigail Oppong', 'Idriss Tondji', 'Mahmod Abdien', 'Houcemeddine Turki', 'Ismaila Lukman', 'Zainab Akinjobi', 'Gbetondji Dovonon', 'Eman Ehab', 'Mai Gamal', 'Abdul-Hakeem Omotayo'] | 2023-05-11 | null | null | null | null | ['3d-reconstruction'] | ['computer-vision'] | [ 1.17816208e-02 -1.34225145e-01 -2.94913054e-01 2.47969508e-01
-6.23556316e-01 -7.28632092e-01 5.50806701e-01 1.52966335e-01
-6.99073434e-01 1.82823062e-01 2.02316761e-01 -8.82616401e-01
-1.91234469e-01 -5.73497295e-01 -6.90378070e-01 -4.89802271e-01
2.63612509e-01 3.66275042e-01 1.41607001e-01 1.01579450e-01
8.99820447e-01 9.08630371e-01 -1.30366111e+00 -3.31792742e-01
1.09831905e+00 2.13188767e-01 3.56199592e-01 7.36836255e-01
-3.42301548e-01 3.28158468e-01 -7.09952712e-01 -4.93690312e-01
2.30122954e-01 -3.67659837e-01 -7.18163669e-01 3.49454507e-02
8.62489820e-01 -1.94817726e-02 -8.50493535e-02 1.07530630e+00
1.68777525e-01 -4.95788097e-01 7.03293502e-01 -1.18726480e+00
-8.87249947e-01 1.64632052e-01 -1.06204283e+00 7.81280041e-01
2.60416597e-01 7.93509483e-02 8.06434989e-01 -7.76676297e-01
1.05986905e+00 1.40645933e+00 6.65321469e-01 5.48066273e-02
-8.95976424e-01 -9.17246997e-01 1.71822712e-01 1.74509436e-01
-1.09454215e+00 -2.43726328e-01 7.37093449e-01 -9.84953284e-01
6.86108589e-01 1.63470477e-01 6.78231001e-01 7.15286374e-01
3.81604105e-01 1.17321368e-02 1.53888106e+00 -6.61514997e-01
-2.70784658e-04 3.01256239e-01 2.34612040e-02 8.33506763e-01
9.79433715e-01 -4.32941347e-01 -4.79673207e-01 -1.90297693e-01
1.00080514e+00 1.32663220e-01 2.01410614e-02 1.90939218e-01
-1.56223512e+00 8.53160679e-01 4.09872293e-01 4.47208703e-01
-4.00117725e-01 -1.09651610e-01 7.44392499e-02 2.88769543e-01
3.79986376e-01 4.33610290e-01 -2.78478563e-01 -8.50012451e-02
-1.36456835e+00 1.65844947e-01 1.00326359e+00 6.20927155e-01
5.00856340e-01 -3.13905887e-02 3.38264614e-01 8.66268158e-01
6.19360685e-01 5.69225907e-01 -5.93785802e-03 -1.33499599e+00
2.36974388e-01 6.21131837e-01 -1.64518684e-01 -1.31114602e+00
2.20491797e-01 -1.59958899e-01 -3.70923460e-01 3.59667599e-01
5.77199757e-01 -1.29464075e-01 -1.13641191e+00 1.16364503e+00
-7.74035463e-03 -1.77163631e-01 -2.40889713e-01 7.62636006e-01
9.35557187e-01 6.35779023e-01 1.76482752e-01 -2.68019021e-01
1.72769809e+00 -8.11078370e-01 -3.13322693e-01 -1.53078049e-01
-3.73567253e-01 -1.33585906e+00 5.37255883e-01 4.49311256e-01
-1.12482166e+00 -3.57668847e-01 -7.41766095e-01 -1.74523350e-02
-5.74910223e-01 -9.07503143e-02 3.76648813e-01 6.59978807e-01
-1.28158998e+00 1.03072181e-01 -6.29712760e-01 -1.24470067e+00
7.52462506e-01 8.74093398e-02 -1.98816434e-01 -2.31877059e-01
-4.52810198e-01 1.37610734e+00 -4.31883410e-02 -3.47084105e-01
-6.56900287e-01 -8.71929765e-01 -2.38932848e-01 -5.70929587e-01
4.91155207e-01 -6.58675253e-01 9.65109944e-01 -8.72378349e-01
-6.13416076e-01 1.45892930e+00 -4.99355078e-01 -3.04961771e-01
5.77758551e-01 7.69297779e-02 -5.36811113e-01 5.73399246e-01
6.97724223e-01 7.15938210e-01 7.52354562e-01 -1.33756268e+00
-1.26487648e+00 -5.48561871e-01 -1.53918594e-01 -1.11596510e-01
-3.30920935e-01 7.96352148e-01 -7.81015038e-01 -9.07740414e-01
3.42548043e-01 -8.71850789e-01 -3.77136260e-01 2.03652948e-01
2.29936838e-01 -2.36143395e-01 8.94939065e-01 -9.01381314e-01
9.85982060e-01 -1.63632524e+00 -1.78428724e-01 6.48656338e-02
3.82292241e-01 -1.30661473e-01 4.64771748e-01 4.87293333e-01
4.40734863e-01 6.08438671e-01 -2.64667302e-01 2.98567444e-01
-3.81895959e-01 2.43211836e-01 -1.04478456e-01 6.75831139e-01
1.09207511e-01 5.83208740e-01 -8.63199830e-01 -8.44778657e-01
3.85245159e-02 4.79741007e-01 -7.32304379e-02 -4.94751871e-01
2.12097168e-01 2.56632537e-01 -5.93264937e-01 1.26139462e+00
9.61005867e-01 -1.30524993e-01 6.25789315e-02 2.42803127e-01
-8.52351189e-01 -2.08163887e-01 -8.44328105e-01 1.18890524e+00
-1.50821850e-01 1.23153627e+00 5.38864255e-01 -9.74405468e-01
6.99608445e-01 9.68442410e-02 5.07642627e-01 -3.91771078e-01
-1.43759519e-01 6.24082208e-01 1.92514583e-01 -2.76697725e-01
6.20384693e-01 5.20635955e-02 3.19776475e-01 3.18674982e-01
-2.27520660e-01 -4.62102406e-02 1.90866187e-01 4.12253708e-01
7.04028904e-01 1.22769676e-01 2.60574996e-01 -4.63355720e-01
2.53629863e-01 7.02075064e-01 5.45937657e-01 9.81216133e-01
-4.64685768e-01 5.69601655e-01 3.68169188e-01 -2.22788304e-01
-1.19866562e+00 -1.01888335e+00 -5.82134068e-01 9.26855266e-01
-2.22390026e-01 -1.63521290e-01 -9.27527726e-01 -4.58345227e-02
-1.15834735e-01 1.39286801e-01 -4.01484579e-01 6.54695868e-01
-4.87311602e-01 -7.52751529e-01 4.93377745e-01 2.04483226e-01
8.28598619e-01 -1.18835127e+00 -1.37445951e+00 5.59506826e-02
1.79875061e-01 -1.17626870e+00 -1.58356547e-01 -1.71728760e-01
-1.16591370e+00 -1.06193888e+00 -1.50366402e+00 -1.03658772e+00
8.45241547e-01 4.76850986e-01 1.16737974e+00 8.67333077e-03
-6.43831730e-01 8.00408244e-01 -1.06470801e-01 -9.32347894e-01
-2.47184843e-01 -4.35805554e-03 -2.85007447e-01 -5.72422862e-01
6.63718820e-01 -1.57119870e-01 -3.38611275e-01 -4.78581712e-02
-5.91563344e-01 -3.55566561e-01 7.75793076e-01 3.40850174e-01
3.53419900e-01 -8.13977942e-02 3.87538731e-01 -8.09460342e-01
5.44304371e-01 -7.37452626e-01 -6.48417354e-01 1.97693631e-01
-9.69572902e-01 -6.65510476e-01 -1.36472017e-01 -1.75174281e-01
-1.04432261e+00 -4.15886879e-01 3.73960495e-01 -2.57614821e-01
-4.50796515e-01 7.29217172e-01 2.75483191e-01 -2.31585547e-01
3.68758589e-01 -1.62201002e-01 9.64321867e-02 -5.55334389e-01
-8.41767527e-03 7.66763568e-01 4.78556246e-01 -5.20227611e-01
7.48576641e-01 7.42117345e-01 1.38414934e-01 -1.46779656e+00
-2.02497408e-01 -7.60019481e-01 -4.83257949e-01 -3.75363082e-01
1.03421807e+00 -8.28349471e-01 -4.75197494e-01 4.10208553e-01
-1.15069747e+00 3.06258857e-01 2.39018649e-01 4.79136407e-01
4.35912088e-02 2.75290489e-01 -4.24554348e-01 -9.98308063e-01
-1.33079559e-01 -1.36581230e+00 7.03539968e-01 6.80623651e-01
-2.42158934e-01 -9.60718930e-01 -1.07212082e-01 8.05910408e-01
2.97338992e-01 4.12741572e-01 6.46272063e-01 -2.37222522e-01
-7.17801213e-01 3.69941384e-01 -6.05524123e-01 -8.33924711e-02
3.83660495e-02 5.73022664e-01 -7.90730596e-01 -1.28012598e-01
-7.71676302e-02 5.02726257e-01 1.26779807e+00 8.70246112e-01
6.71458840e-01 -2.41521269e-01 -6.52032793e-01 2.95210540e-01
1.66503501e+00 6.88744187e-01 3.98269445e-01 1.11435473e+00
7.62510300e-01 1.00671780e+00 4.67418432e-01 4.19284180e-02
5.50623953e-01 1.76028877e-01 8.89106169e-02 -1.85063891e-02
1.70497335e-02 -2.84180753e-02 3.19745153e-01 7.02404022e-01
-8.28165531e-01 3.70336324e-01 -1.58169436e+00 1.13088393e+00
-1.54269159e+00 -9.46464181e-01 -6.11070991e-01 2.02146983e+00
5.02925873e-01 2.18444631e-01 4.13640171e-01 -3.93293649e-01
7.99321651e-01 -5.18093891e-02 -4.03755367e-01 -8.17440152e-01
8.78367014e-03 4.63827431e-01 9.63256657e-01 2.08281800e-01
-9.14766490e-01 9.84164119e-01 6.83562708e+00 5.09135246e-01
-1.03064299e+00 1.29894689e-01 6.68132067e-01 4.26061265e-02
-2.42158383e-01 3.52298975e-01 -5.96617103e-01 2.28571892e-01
8.90024781e-01 -1.58097297e-01 1.33018643e-01 7.17772663e-01
3.11016262e-01 -3.25595081e-01 -2.57549107e-01 8.56338024e-01
1.32398933e-01 -1.53169358e+00 -6.09327480e-02 7.19803452e-01
9.61597621e-01 3.25306714e-01 2.58870542e-01 -3.92767519e-01
3.43133122e-01 -1.21075130e+00 6.52392685e-01 4.62135166e-01
5.59002995e-01 -9.22199428e-01 3.50435674e-01 2.46651769e-02
-8.79891276e-01 1.77614577e-02 -4.05629456e-01 -2.08864644e-01
-1.23558089e-01 5.18373489e-01 -6.54672980e-01 5.56269825e-01
1.38755250e+00 6.89950109e-01 -5.28670907e-01 1.20768678e+00
1.20394327e-01 9.45454121e-01 -9.44557935e-02 6.75158277e-02
6.27740204e-01 -5.97200334e-01 8.29418063e-01 1.50515771e+00
2.71468163e-01 1.84696034e-01 1.13092903e-02 7.40122676e-01
-1.12962745e-01 2.78296582e-02 -7.33449936e-01 -4.73549277e-01
5.60894430e-01 1.13083196e+00 -1.38818526e+00 -7.16588795e-02
-8.48523736e-01 8.94046187e-01 -2.84031481e-01 5.78218818e-01
-1.74959451e-01 -5.37507474e-01 5.77735305e-01 2.12376118e-01
3.10344070e-01 -3.28542262e-01 -4.92263496e-01 -8.04124832e-01
-1.42299384e-01 -9.79244232e-01 3.49329174e-01 -5.93017101e-01
-1.16692615e+00 1.89631104e-01 6.86071301e-03 -5.68085730e-01
3.20335865e-01 -6.75942779e-01 -4.55554545e-01 1.17799592e+00
-1.36000121e+00 -1.28863800e+00 -1.06707014e-01 2.61652410e-01
6.86222792e-01 -4.39117283e-01 3.30155671e-01 -2.85470877e-02
-2.85313755e-01 1.37770683e-01 2.28170902e-01 1.81995466e-01
5.99292397e-01 -1.03777862e+00 2.93498367e-01 8.48264337e-01
1.36441201e-01 1.03995204e+00 5.35020232e-01 -1.05577838e+00
-1.41736019e+00 -4.95030701e-01 1.21927142e+00 -7.57855177e-01
9.27300274e-01 3.73944819e-01 -6.43702686e-01 6.30556762e-01
7.10442662e-01 -5.64656734e-01 4.31335837e-01 5.07047698e-02
-1.51230127e-01 1.01649001e-01 -1.29454350e+00 7.01902032e-01
8.35169971e-01 -3.99682373e-01 -1.05392587e+00 2.02200785e-02
2.67084539e-01 5.18802330e-02 -7.97279239e-01 -8.57780725e-02
9.07613575e-01 -6.27808809e-01 1.16951203e+00 -3.12462807e-01
7.52222359e-01 -1.65505216e-01 6.37540361e-03 -7.72436976e-01
-1.42151892e-01 -5.31662285e-01 3.86524677e-01 1.38953722e+00
1.94473505e-01 -9.22094345e-01 7.68655777e-01 3.39596719e-01
-4.10225429e-02 -6.52507722e-01 -9.09585357e-01 -4.28350359e-01
6.82606101e-01 -1.99145600e-01 -1.29971460e-01 1.20109391e+00
-7.30178475e-01 -8.08589086e-02 2.03091070e-01 -1.01980723e-01
1.06128943e+00 7.94967338e-02 4.84362602e-01 -1.49291134e+00
3.88021797e-01 -9.55720127e-01 -4.74459976e-01 -2.35238075e-01
-2.36849219e-01 -6.47885561e-01 -4.06911463e-01 -2.17774820e+00
6.16631150e-01 -2.93984592e-01 -7.35332966e-02 2.57009804e-01
9.81064793e-03 3.33177507e-01 3.61239254e-01 7.67431378e-01
1.15311220e-01 -7.31161892e-01 1.24086595e+00 -1.03464529e-01
2.98910029e-02 -2.71533996e-01 -1.32549548e+00 1.10881865e+00
6.43732071e-01 -2.78864473e-01 2.16962006e-02 -3.79256368e-01
3.19491848e-02 -5.06384373e-01 6.52677536e-01 -7.71627128e-01
3.86401087e-01 -6.29461408e-01 4.61859703e-01 -8.26125324e-01
-2.27473542e-01 -9.19733226e-01 1.61628723e-02 5.17023206e-01
1.39916569e-01 4.31209475e-01 2.00451329e-01 3.24038833e-01
-1.92414567e-01 -4.58606631e-01 6.58386528e-01 -7.32840598e-01
-9.09005344e-01 2.27809530e-02 -6.73194051e-01 2.56917745e-01
1.05350661e+00 -7.08985031e-01 -1.93240061e-01 4.54720072e-02
-4.81969297e-01 -1.99739933e-02 9.16496277e-01 3.01925659e-01
5.11079669e-01 -6.55172706e-01 -9.18241143e-01 -5.63743889e-01
-9.10153706e-03 -3.52708727e-01 -2.96587706e-01 8.50259602e-01
-1.11172843e+00 5.99008143e-01 -4.94481236e-01 -4.58408266e-01
-1.30449486e+00 1.55714631e-01 -2.04263926e-01 2.44665056e-01
-8.06724191e-01 7.08180249e-01 -1.73597321e-01 -9.86638740e-02
3.64271313e-01 1.47670507e-01 -4.57538903e-01 6.07984841e-01
2.99702525e-01 8.73908043e-01 -2.96871603e-01 -9.72898006e-01
-7.69788563e-01 8.94840598e-01 -1.26245946e-01 -7.14581609e-01
1.48317456e+00 1.78629216e-02 -4.85668570e-01 5.54166436e-01
8.14503491e-01 1.64278194e-01 -7.09682167e-01 2.97477901e-01
3.12885344e-01 -6.18899226e-01 3.12179506e-01 -9.24397290e-01
-1.27959502e+00 8.07353020e-01 6.54023111e-01 -4.50412892e-02
1.01122737e+00 2.54873484e-01 3.31798881e-01 -7.61544034e-02
3.80809367e-01 -1.39252365e+00 -2.02992246e-01 2.07674742e-01
8.22198570e-01 -1.15162539e+00 3.84452194e-01 -2.34845921e-01
-6.91138864e-01 1.11115718e+00 2.75875628e-01 4.00329493e-02
7.33194530e-01 -1.06199700e-02 3.75575155e-01 -4.01489675e-01
-1.78919464e-01 -1.89496979e-01 1.58781409e-01 7.69864142e-01
7.01557755e-01 5.38680851e-02 -8.80528450e-01 1.91527419e-02
-1.99264497e-01 2.06313461e-01 6.85646832e-01 1.14441669e+00
-5.04033923e-01 -8.31551433e-01 -1.09550607e+00 8.05956244e-01
-1.24428630e+00 -2.22307354e-01 -8.81162345e-01 9.36015725e-01
2.87258565e-01 1.06984866e+00 2.26644188e-01 2.55510569e-01
4.57862839e-02 -2.44535834e-01 6.01328731e-01 -2.88300186e-01
-8.12706828e-01 4.35799450e-01 -3.99327166e-02 2.20776826e-01
-9.18751121e-01 -1.10258198e+00 -1.10280252e+00 -6.00413442e-01
-1.33823799e-02 2.35055566e-01 1.21790242e+00 5.65984368e-01
1.20344445e-01 3.75268131e-01 4.22384813e-02 -4.82595295e-01
1.91676766e-01 -6.80024981e-01 -4.35896963e-01 -2.61382878e-01
2.45907262e-01 -5.55420280e-01 -3.22288752e-01 3.33781391e-01] | [9.559334754943848, 8.237627983093262] |
98bbfd53-6901-47d7-a973-88609f5d8139 | gesture-based-bootstrapping-for-egocentric | 1612.02889 | null | http://arxiv.org/abs/1612.02889v2 | http://arxiv.org/pdf/1612.02889v2.pdf | Gesture-based Bootstrapping for Egocentric Hand Segmentation | Accurately identifying hands in images is a key sub-task for human activity
understanding with wearable first-person point-of-view cameras. Traditional
hand segmentation approaches rely on a large corpus of manually labeled data to
generate robust hand detectors. However, these approaches still face challenges
as the appearance of the hand varies greatly across users, tasks, environments
or illumination conditions. A key observation in the case of many wearable
applications and interfaces is that, it is only necessary to accurately detect
the user's hands in a specific situational context. Based on this observation,
we introduce an interactive approach to learn a person-specific hand
segmentation model that does not require any manually labeled training data.
Our approach proceeds in two steps, an interactive bootstrapping step for
identifying moving hand regions, followed by learning a personalized user
specific hand appearance model. Concretely, our approach uses two convolutional
neural networks: (1) a gesture network that uses pre-defined motion information
to detect the hand region; and (2) an appearance network that learns a person
specific model of the hand region based on the output of the gesture network.
During training, to make the appearance network robust to errors in the gesture
network, the loss function of the former network incorporates the confidence of
the gesture network while learning. Experiments demonstrate the robustness of
our approach with an F1 score over 0.8 on all challenging datasets across a
wide range of illumination and hand appearance variations, improving over a
baseline approach by over 10%. | ['Kris M. Kitani', 'Yubo Zhang', 'Vishnu Naresh Boddeti'] | 2016-12-09 | null | null | null | null | ['hand-segmentation'] | ['computer-vision'] | [ 3.05610299e-01 -2.55464941e-01 -2.02693209e-01 -2.59859055e-01
-3.61938655e-01 -8.45595658e-01 2.86039740e-01 -2.84736007e-01
-5.01091897e-01 2.34922990e-01 -6.98335096e-02 -2.62076836e-02
2.90484011e-01 -3.15980136e-01 -6.73775017e-01 -6.98408604e-01
3.52540195e-01 4.33888197e-01 5.29826999e-01 2.43422419e-01
1.91391632e-02 6.35905445e-01 -1.32335401e+00 2.67135762e-02
4.76993591e-01 1.05163729e+00 1.83921918e-01 1.01905382e+00
2.85739571e-01 3.59785527e-01 -5.60520828e-01 -1.93719819e-01
4.20876652e-01 -3.87726128e-01 -8.34363163e-01 4.59735483e-01
4.75630671e-01 -6.60960734e-01 -2.61794537e-01 8.86508048e-01
7.41839647e-01 2.69439638e-01 4.33888376e-01 -9.73237514e-01
-6.87142611e-02 1.69903070e-01 -7.25331247e-01 2.04007141e-02
5.94709396e-01 6.26185060e-01 6.98245704e-01 -8.08211684e-01
6.19077563e-01 1.05736399e+00 6.00191474e-01 6.74428642e-01
-1.11303067e+00 -5.47384143e-01 5.70415795e-01 -1.16119564e-01
-1.55863965e+00 -2.94895291e-01 8.01501215e-01 -7.23212600e-01
6.24283969e-01 1.45389408e-01 8.71885777e-01 1.35954714e+00
-1.76612124e-01 8.42833877e-01 9.99329150e-01 -5.46637297e-01
2.70360559e-01 2.01017290e-01 2.05755204e-01 8.27563345e-01
7.85914287e-02 -2.16274410e-01 -2.48983636e-01 -6.58134595e-02
1.19247746e+00 2.46889129e-01 -2.96790779e-01 -5.19348264e-01
-9.88198221e-01 2.62588292e-01 3.77276659e-01 2.76532114e-01
-6.25571012e-01 1.49358392e-01 1.79358646e-01 -1.85211331e-01
1.38479918e-01 -4.69617657e-02 -4.11465079e-01 -2.56136954e-02
-1.06345057e+00 2.49199226e-01 8.00957203e-01 8.03564131e-01
3.34956616e-01 -3.64219874e-01 -4.10786629e-01 6.35743618e-01
3.26728314e-01 5.25961339e-01 1.18542448e-01 -6.83932066e-01
4.58564967e-01 6.85595453e-01 3.27568591e-01 -6.22631907e-01
-4.33248192e-01 -2.90277153e-01 -3.91754508e-01 3.52060735e-01
1.04259729e+00 -4.09850568e-01 -1.16826856e+00 1.62264287e+00
6.42543435e-01 -4.76627015e-02 -4.59941626e-01 1.27189434e+00
4.76117253e-01 2.42380366e-01 2.59652883e-01 1.19189620e-01
1.53348219e+00 -1.01874566e+00 -4.49826181e-01 -4.08484787e-01
-3.11935190e-02 -5.63841879e-01 1.21640909e+00 4.48436618e-01
-9.57341850e-01 -6.76153183e-01 -8.55240583e-01 2.44669363e-01
-9.98321995e-02 4.25864190e-01 4.44989771e-01 8.54461968e-01
-8.05737436e-01 4.14299279e-01 -9.42600489e-01 -6.66749775e-01
4.19752598e-01 6.04174554e-01 -2.49310300e-01 1.68431457e-02
-5.62674165e-01 5.57489872e-01 3.30731183e-01 3.23922187e-01
-7.67690361e-01 -3.29750150e-01 -5.23457885e-01 -8.74756649e-02
5.70373833e-01 -8.31060171e-01 1.24441409e+00 -1.37072802e+00
-1.61683619e+00 8.59940767e-01 -2.49342054e-01 1.43054545e-01
1.11676335e+00 -4.08962190e-01 -4.62207459e-02 2.89163440e-01
-2.23946780e-01 4.76426631e-01 1.20589602e+00 -1.44421422e+00
-7.50538349e-01 -7.00597882e-01 1.52261361e-01 1.90402746e-01
-9.99133959e-02 1.50971189e-01 -1.07139885e+00 -5.34703076e-01
-8.29578415e-02 -1.15345156e+00 -2.82188468e-02 1.68860182e-01
-4.93071049e-01 -3.39462787e-01 8.86479199e-01 -1.05414343e+00
9.74208415e-01 -1.80671573e+00 2.41878539e-01 5.95783770e-01
2.20807761e-01 3.91184032e-01 4.68610153e-02 -1.33259103e-01
1.44709498e-01 -2.59334326e-01 -4.03199866e-02 -3.49505156e-01
-2.08337024e-01 -1.65395245e-01 6.39107078e-02 4.56319839e-01
6.41830906e-04 9.16733205e-01 -9.58482087e-01 -3.80800813e-01
4.24916834e-01 7.90830553e-01 -2.96036065e-01 5.88282526e-01
-2.21537888e-01 8.30779731e-01 -3.95712107e-01 7.76870668e-01
2.24657699e-01 -3.29704791e-01 4.27998692e-01 -2.31824920e-01
1.01366639e-01 -7.11791664e-02 -1.53480875e+00 1.68813908e+00
-3.40899527e-01 3.62820983e-01 1.63648054e-01 -5.39615870e-01
4.97125417e-01 5.44869661e-01 4.26545709e-01 -1.12738475e-01
4.29223448e-01 -3.06173414e-03 -2.22299639e-02 -7.28696108e-01
-3.26502584e-02 1.92605615e-01 3.10589433e-01 6.78068638e-01
9.85429063e-02 3.96012247e-01 5.11315977e-03 -1.81014284e-01
8.27241182e-01 5.09914994e-01 2.10104361e-02 1.82364941e-01
3.68880719e-01 -4.05889899e-01 3.64141166e-01 7.24283397e-01
-3.98251116e-01 8.89567912e-01 1.33921430e-01 -5.84866285e-01
-8.92806470e-01 -9.42452371e-01 2.14429274e-01 1.32424390e+00
1.72186583e-01 5.49697457e-03 -1.20621264e+00 -1.04625154e+00
-1.74395159e-01 2.63492107e-01 -6.85720026e-01 2.90868044e-01
-7.74871707e-01 -6.14158034e-01 3.21900576e-01 1.09210038e+00
5.80546498e-01 -1.20099437e+00 -9.82738972e-01 -6.86457381e-02
-9.16483179e-02 -1.19258213e+00 -8.99341404e-01 -1.25375092e-01
-6.45683885e-01 -1.24672961e+00 -1.04132974e+00 -7.35266328e-01
1.01105821e+00 1.57524556e-01 7.16196775e-01 1.42529413e-01
-4.99223679e-01 8.34905505e-01 -1.77191630e-01 -2.74706781e-01
-7.16218576e-02 2.44853929e-01 8.95403698e-02 4.04868782e-01
3.43853831e-01 -4.53512490e-01 -1.02999270e+00 3.87456149e-01
-3.08441937e-01 -1.13582790e-01 5.30870080e-01 4.71559852e-01
3.50205064e-01 -3.48193377e-01 9.69802141e-02 -5.84185004e-01
5.20210862e-01 2.09790673e-02 -4.58584577e-01 5.79976618e-01
-3.13064992e-01 -2.35771328e-01 3.56347322e-01 -8.07595968e-01
-1.09279811e+00 8.59542966e-01 -8.15379396e-02 -4.56882149e-01
-4.05821145e-01 -2.98074454e-01 -3.51663888e-01 -7.48298913e-02
6.52531147e-01 1.63120940e-01 -1.89328223e-01 -3.58964205e-01
3.85663927e-01 8.30073476e-01 6.77251279e-01 -6.38129115e-01
6.92106128e-01 5.73974729e-01 -4.76233721e-01 -9.64265585e-01
-6.91621244e-01 -6.83228970e-01 -1.33509612e+00 -6.88390195e-01
1.12107253e+00 -6.57989144e-01 -9.40246880e-01 6.96914375e-01
-1.22972560e+00 -6.76384151e-01 -5.48186526e-02 3.84787112e-01
-3.15858662e-01 2.95404434e-01 -3.77490729e-01 -1.21242726e+00
-5.54123580e-01 -1.20050418e+00 1.24203694e+00 4.57361132e-01
-5.53641379e-01 -8.88707459e-01 -2.66525626e-01 4.45843458e-01
9.46914256e-02 4.00734276e-01 4.72177774e-01 -4.25369084e-01
-5.22943139e-01 -5.06876230e-01 -1.04907863e-01 2.62594253e-01
3.57947052e-01 9.16401669e-03 -1.38896990e+00 -3.86443675e-01
-2.67095804e-01 -1.64431393e-01 5.21671116e-01 4.95887280e-01
1.16579115e+00 -9.05845240e-02 -4.22736287e-01 4.79080826e-01
9.70384777e-01 2.17345133e-01 3.26911271e-01 9.42730252e-03
1.05130875e+00 6.44650936e-01 2.03505740e-01 3.26939911e-01
2.02244475e-01 7.47399092e-01 1.44856751e-01 -2.51225799e-01
-2.03319058e-01 -2.09297046e-01 2.85959125e-01 2.00570957e-03
-7.48415411e-01 3.54479486e-03 -8.22128892e-01 3.63662243e-01
-1.75724697e+00 -8.04185867e-01 1.72090158e-01 2.50629616e+00
6.08047783e-01 5.00322767e-02 9.45915401e-01 1.82624564e-01
7.91814089e-01 -1.28130540e-01 -8.46475124e-01 1.06444634e-01
3.66337150e-01 1.64436117e-01 3.94015640e-01 3.40843171e-01
-1.32595468e+00 9.31657135e-01 5.88336468e+00 4.83977571e-02
-1.32440758e+00 4.40952033e-02 3.73854011e-01 -1.46133333e-01
3.60165477e-01 -2.81297892e-01 -7.19912171e-01 4.33348566e-01
2.68813103e-01 5.24753392e-01 5.97482562e-01 7.94805825e-01
3.30542326e-01 -5.83040155e-02 -1.19178271e+00 1.07224751e+00
1.69302598e-01 -5.74342310e-01 -2.03509226e-01 -1.03638597e-01
4.96133417e-01 -3.92190725e-01 -9.73048732e-02 -1.23578221e-01
3.13284323e-02 -9.71205533e-01 8.43251228e-01 4.33047354e-01
8.09696853e-01 -4.88247812e-01 4.34437335e-01 4.30929750e-01
-1.36001706e+00 -9.44002643e-02 2.53425211e-01 4.19374555e-02
1.41066328e-01 6.45553917e-02 -9.44532871e-01 -3.26254754e-04
8.29229057e-01 1.57902524e-01 -4.90120709e-01 9.37912822e-01
-5.44375598e-01 5.30571878e-01 -3.40758532e-01 1.48368031e-01
-4.55454178e-02 9.11046937e-02 4.57481742e-01 1.41573203e+00
-1.13144748e-01 2.49040976e-01 4.41763848e-01 7.91559637e-01
4.14385907e-02 -6.35853931e-02 -2.34670967e-01 1.58268392e-01
3.82714927e-01 1.24360216e+00 -1.06675792e+00 -3.02871436e-01
-3.16731006e-01 1.59188676e+00 9.34671536e-02 6.18921459e-01
-6.29012465e-01 -3.10734510e-01 4.01969999e-01 4.14198756e-01
3.17786545e-01 -2.44223043e-01 -2.42529169e-01 -9.99416232e-01
4.33999687e-01 -7.78390884e-01 3.68321002e-01 -5.12890995e-01
-9.94209111e-01 5.30855060e-01 -1.52393326e-01 -8.32394660e-01
-3.88880223e-01 -6.89617157e-01 -6.99219942e-01 9.53402996e-01
-9.46273267e-01 -1.59709597e+00 -8.55798244e-01 6.75756216e-01
7.51960576e-01 1.10104777e-01 6.29404128e-01 1.22103117e-01
-8.03069890e-01 7.66247809e-01 -5.15947402e-01 6.60174549e-01
6.52174115e-01 -1.34345317e+00 4.65052605e-01 9.03915286e-01
-2.57347561e-02 8.51930499e-01 4.89554524e-01 -6.45095408e-01
-1.31538904e+00 -9.00625646e-01 4.56422716e-01 -8.72762024e-01
1.11876965e-01 -5.22025287e-01 -7.47287452e-01 9.16974127e-01
-6.16325391e-03 1.95681229e-01 4.95921910e-01 1.33375421e-01
-3.54084879e-01 -1.63318701e-02 -1.12195742e+00 5.56249917e-01
1.23877931e+00 -3.70593280e-01 -4.38077301e-01 2.53839672e-01
-8.26799590e-03 -6.74377561e-01 -6.44805312e-01 1.02920443e-01
1.31044805e+00 -7.58924901e-01 1.01405323e+00 -6.34399593e-01
-9.75848734e-02 -4.55635846e-01 1.98310971e-01 -8.58432353e-01
-1.66051432e-01 -6.40402377e-01 -3.14849585e-01 1.02238965e+00
8.98566991e-02 -2.39202589e-01 9.74450171e-01 1.10168350e+00
5.89195132e-01 -6.35701418e-01 -5.35650373e-01 -6.09143913e-01
-3.31415594e-01 -4.27808225e-01 2.08152831e-01 5.85326970e-01
-1.28773108e-01 4.27203864e-01 -2.94436067e-01 3.03586900e-01
7.02142835e-01 3.24714109e-02 1.09049952e+00 -1.25683570e+00
-3.81821394e-01 -4.28246379e-01 -1.78634033e-01 -1.22785830e+00
-8.22178870e-02 -4.53145564e-01 1.72294721e-01 -1.47405469e+00
4.77825552e-01 -3.58198941e-01 -1.07229494e-01 5.97753942e-01
-4.04352576e-01 3.68167639e-01 3.95043254e-01 2.85443068e-01
-5.95849812e-01 -1.13820575e-01 1.10594845e+00 -1.01593755e-01
-8.09752107e-01 5.61451197e-01 -3.48949075e-01 1.01855135e+00
5.70881188e-01 -1.22813374e-01 -3.05062890e-01 -3.28280091e-01
-2.50423044e-01 -1.86216235e-01 7.34156787e-01 -9.54040229e-01
2.33059287e-01 -1.65294990e-01 8.14478338e-01 -3.26814860e-01
2.12332457e-01 -8.74129653e-01 5.56416810e-02 5.52718103e-01
-8.72366279e-02 -1.49496585e-01 2.50907894e-02 4.71721202e-01
4.16758955e-01 -7.13110790e-02 8.15274417e-01 -3.08180451e-01
-7.26485074e-01 2.57274091e-01 -1.73539683e-01 -1.26912221e-01
1.03145778e+00 -5.83118916e-01 2.70306289e-01 -3.58493209e-01
-9.64330375e-01 1.40583366e-01 4.60656315e-01 6.10630810e-01
4.30740863e-01 -9.51079905e-01 -3.95581692e-01 3.50604922e-01
-4.78661396e-02 1.69297799e-01 5.55652492e-02 8.34919930e-01
-4.24697578e-01 1.25900805e-01 -3.83569822e-02 -8.73987436e-01
-1.53783500e+00 3.20977628e-01 6.82096720e-01 5.08808345e-03
-7.77439892e-01 7.77907312e-01 1.23857327e-01 -1.98478684e-01
7.14095056e-01 -3.97821158e-01 -1.22319767e-02 -8.46809968e-02
8.64853740e-01 5.10974824e-01 -1.23712479e-03 -7.72967279e-01
-4.65423495e-01 9.33689654e-01 -1.71952753e-03 -2.24646643e-01
9.98830557e-01 1.66999679e-02 2.85145849e-01 3.80434155e-01
8.06524873e-01 -1.08032167e-01 -1.83270288e+00 -2.28914693e-01
-2.29244903e-01 -4.57115442e-01 -1.77613392e-01 -1.04949367e+00
-1.10398436e+00 9.26135600e-01 9.84358788e-01 -2.55842656e-01
1.07477653e+00 2.71370728e-02 6.77258611e-01 2.74977088e-01
3.83892357e-01 -1.27991903e+00 2.91145056e-01 2.62489021e-01
7.80906618e-01 -1.21154165e+00 -2.22833782e-01 -4.17916685e-01
-6.52568460e-01 1.18543983e+00 7.16354549e-01 1.78855121e-01
4.59247082e-01 1.49630666e-01 3.05506825e-01 -8.31141993e-02
1.09306596e-01 -3.76056582e-01 5.92547119e-01 6.01236820e-01
2.86306381e-01 8.56563449e-02 -1.74816549e-02 6.43293977e-01
9.58370715e-02 2.38204911e-01 -2.03527421e-01 1.02186775e+00
-2.92216599e-01 -9.55901086e-01 -6.27380192e-01 3.31920117e-01
-4.47107524e-01 4.59273875e-01 -7.13081837e-01 7.29896784e-01
3.07859302e-01 7.64483809e-01 -3.61131094e-02 -3.79564136e-01
6.32304430e-01 4.11114126e-01 8.24206769e-01 -6.25768423e-01
-7.17824876e-01 1.70798227e-01 -4.66919243e-01 -5.26982725e-01
-3.92019868e-01 -6.85226858e-01 -1.21855927e+00 -1.27038043e-02
-2.14707583e-01 -4.34954703e-01 6.08252466e-01 1.16527784e+00
6.28519291e-03 4.04304504e-01 3.11458886e-01 -1.41957188e+00
-4.24235672e-01 -9.46687460e-01 -4.13336515e-01 5.82110345e-01
2.50468701e-01 -6.93037271e-01 5.09619936e-02 3.85721624e-01] | [6.65946626663208, -0.5908035039901733] |
069f2729-f866-4c0a-91bc-ed7c1baa8bd6 | decoupling-pragmatics-discriminative-decoding | null | null | https://aclanthology.org/2021.reinact-1.7 | https://aclanthology.org/2021.reinact-1.7.pdf | Decoupling Pragmatics: Discriminative Decoding for Referring Expression Generation | The shift to neural models in Referring Expression Generation (REG) has enabled more natural set-ups, but at the cost of interpretability. We argue that integrating pragmatic reasoning into the inference of context-agnostic generation models could reconcile traits of traditional and neural REG, as this offers a separation between context-independent, literal information and pragmatic adaptation to context. With this in mind, we apply existing decoding strategies from discriminative image captioning to REG and evaluate them in terms of pragmatic informativity, likelihood to ground-truth annotations and linguistic diversity. Our results show general effectiveness, but a relatively small gain in informativity, raising important questions for REG in general. | ['Sina Zarrieß', 'Simeon Schüz'] | null | null | null | null | reinact-2021-10 | ['referring-expression-generation'] | ['computer-vision'] | [ 3.80425453e-01 8.98773551e-01 -1.54494882e-01 -7.21105933e-01
-1.21083117e+00 -6.16307914e-01 7.75016725e-01 -7.33107179e-02
-3.27367008e-01 8.33895981e-01 1.26621473e+00 -3.03954333e-01
-7.14996010e-02 -5.28835952e-01 -5.51770687e-01 -3.49371076e-01
2.63400882e-01 5.06612897e-01 -3.17583591e-01 -4.58188504e-01
5.44945784e-02 1.64512768e-01 -9.37281430e-01 8.50947142e-01
6.38217688e-01 7.97381878e-01 1.16469003e-01 3.35139155e-01
9.10953879e-02 1.07125688e+00 -5.23396671e-01 -7.88043737e-01
-6.98952228e-02 -7.62584448e-01 -1.20361817e+00 -1.84646666e-01
2.65616536e-01 -2.91927159e-01 -1.28999844e-01 7.38421738e-01
5.38548887e-01 1.61166534e-01 4.79406238e-01 -6.85844004e-01
-1.48762155e+00 1.10482776e+00 -8.46308693e-02 2.80729681e-01
4.61371064e-01 5.80123782e-01 1.50092649e+00 -5.05943775e-01
1.10176694e+00 1.67085755e+00 6.44868374e-01 9.39453781e-01
-1.46064556e+00 9.76359844e-03 3.88456970e-01 -7.69579858e-02
-9.73635375e-01 -8.18325400e-01 6.57817364e-01 -3.28053415e-01
1.02246201e+00 3.41105461e-01 7.66314805e-01 1.54001713e+00
-2.36815587e-01 8.20339322e-01 1.43970633e+00 -4.80394691e-01
2.12538809e-01 -1.29587293e-01 -5.13514653e-02 4.38844651e-01
7.94961005e-02 2.27172375e-01 -6.65150046e-01 -3.45898345e-02
5.50568283e-01 -8.09492648e-01 -4.64574009e-01 2.00001031e-01
-1.28002787e+00 9.13177371e-01 7.09442496e-01 3.54374766e-01
-3.52967769e-01 5.29749751e-01 4.81520385e-01 1.05403043e-01
1.95399895e-01 9.91284788e-01 -4.17459965e-01 -6.21652044e-02
-5.11676371e-01 3.52688938e-01 6.57704830e-01 8.94756973e-01
3.87782127e-01 4.31748256e-02 -5.52019417e-01 7.77832329e-01
4.05418128e-01 2.31233820e-01 4.69852328e-01 -1.47942400e+00
3.81024420e-01 3.48945558e-01 -6.65949583e-02 -8.83244276e-01
-4.45163727e-01 -1.60395786e-01 -1.12980969e-01 -2.76277423e-01
3.51168603e-01 -2.25346759e-01 -5.11994779e-01 2.53286600e+00
-3.83581407e-02 -6.16723657e-01 2.55311191e-01 1.15259171e+00
6.65307522e-01 4.80806917e-01 6.67863846e-01 -5.23920804e-02
1.59689772e+00 -6.94103479e-01 -5.77387512e-01 -8.01715493e-01
7.45755851e-01 -4.46075261e-01 1.61765730e+00 -7.76184127e-02
-1.16363430e+00 -8.60884227e-03 -7.84427762e-01 -6.06986880e-01
-4.23526689e-02 -7.46962652e-02 1.02920425e+00 4.40787673e-01
-1.26134670e+00 2.46001840e-01 -4.79221284e-01 -5.90679824e-01
5.49769044e-01 7.08568171e-02 -2.32048020e-01 1.41488031e-01
-1.20619857e+00 1.20936358e+00 2.72055387e-01 2.99613684e-01
-4.58033890e-01 -2.25841954e-01 -9.35807228e-01 -6.64722621e-02
2.83200741e-01 -1.19456327e+00 1.52230704e+00 -1.42382228e+00
-1.52122009e+00 1.23373818e+00 -4.12714891e-02 -5.01173973e-01
2.64165998e-01 -9.50044319e-02 -2.32782587e-01 1.00257680e-01
1.71678960e-01 1.14161706e+00 2.16362223e-01 -1.09137905e+00
-9.90042388e-02 -4.57112491e-01 5.22486269e-01 6.66380525e-01
2.04352275e-01 2.42988199e-01 2.65208930e-02 -5.96830904e-01
9.03650448e-02 -8.85576904e-01 -4.93712015e-02 4.34838049e-02
-2.67423362e-01 -6.08791970e-02 2.33948529e-01 -7.65117764e-01
8.12401354e-01 -2.16927147e+00 3.69159549e-01 -3.47340614e-01
-6.61903098e-02 -2.71447182e-01 -2.30069786e-01 1.70639217e-01
1.60967767e-01 4.65013176e-01 -2.68984437e-01 -3.02867889e-01
2.78076082e-01 4.67698872e-01 -4.52105016e-01 2.03058302e-01
5.84547281e-01 1.44221437e+00 -1.02842426e+00 -6.07545078e-01
-1.44848615e-01 3.41858029e-01 -9.80512500e-01 -1.78986043e-01
-6.65036738e-01 5.98694623e-01 -4.14551765e-01 6.89236403e-01
-3.59042585e-02 -2.13402450e-01 6.78884506e-01 -1.96433425e-01
4.42970917e-02 9.54980314e-01 -3.11853409e-01 1.87850356e+00
-5.42957187e-01 8.16496134e-01 3.34449053e-01 -7.95975566e-01
5.39744198e-01 3.23374957e-01 -2.71633655e-01 -8.14442158e-01
2.57890075e-01 3.90835434e-01 4.06992942e-01 -4.62761402e-01
5.48854589e-01 -7.69375920e-01 -4.58252132e-01 5.02486408e-01
5.57467863e-02 -3.99544090e-01 -1.28169760e-01 8.86598229e-02
7.35705793e-01 6.12358749e-01 3.13458294e-01 -4.98466223e-01
1.72560487e-03 2.25647360e-01 5.13964713e-01 6.15638852e-01
-2.61662096e-01 6.85451508e-01 4.42488939e-01 -2.95235693e-01
-8.74071658e-01 -1.07658124e+00 -2.53501028e-01 9.90744770e-01
-7.94638991e-02 -2.96114713e-01 -8.41282904e-01 -5.80078602e-01
-3.47318321e-01 1.26445699e+00 -5.97199440e-01 -2.28140578e-01
-8.59979868e-01 -7.67690539e-01 7.04900503e-01 4.93837297e-01
2.76552916e-01 -1.45001650e+00 -8.42212200e-01 1.14522740e-01
-4.07286763e-01 -1.16335452e+00 -9.33798850e-02 6.94175735e-02
-7.88355350e-01 -6.36247277e-01 -3.75047326e-01 -6.07486963e-01
5.06878376e-01 -2.50711650e-01 1.34432435e+00 2.26055691e-03
1.87810555e-01 4.88496333e-01 -2.12422624e-01 -2.23331183e-01
-6.75253391e-01 -3.17061134e-02 -3.16864163e-01 -5.04065990e-01
2.01508686e-01 -5.58879733e-01 -5.92593849e-01 -1.51641697e-01
-7.13260531e-01 3.41233462e-01 5.52195013e-01 9.16634977e-01
4.22917157e-01 -1.03958297e+00 8.21440756e-01 -9.09721196e-01
1.04431844e+00 -4.25727993e-01 -1.49544880e-01 1.54709086e-01
-4.47553068e-01 2.27254614e-01 4.03688490e-01 -1.93437561e-01
-1.51618207e+00 -3.89495045e-01 -2.34823197e-01 2.29149982e-01
-1.24815881e-01 6.36214077e-01 -2.39133015e-01 3.12639445e-01
1.04834580e+00 -2.02499300e-01 -4.60254699e-02 -3.06112349e-01
1.12575197e+00 5.14341950e-01 6.20277107e-01 -1.16039431e+00
9.77056175e-02 4.55519795e-01 -3.17204863e-01 -5.53252101e-01
-1.33151317e+00 3.59911352e-01 -3.47974509e-01 1.32807389e-01
9.98390675e-01 -9.93861675e-01 -3.65946770e-01 -2.01801270e-01
-1.22764790e+00 -5.69121540e-01 -9.37200427e-01 2.79359460e-01
-1.02442753e+00 2.72656471e-01 -7.69363284e-01 -5.41166008e-01
-2.84965754e-01 -1.13918126e+00 1.23598814e+00 -4.29392084e-02
-9.92202401e-01 -9.97909665e-01 -1.42785266e-01 6.86720431e-01
4.91092682e-01 3.80274117e-01 1.21961904e+00 -6.52379751e-01
-7.31436908e-01 1.32197335e-01 -3.30996454e-01 2.42874965e-01
-7.63769969e-02 -4.01450664e-01 -1.20646501e+00 3.37752134e-01
2.86626756e-01 -6.30317926e-01 6.37055337e-01 3.15630257e-01
6.86961412e-01 -8.51332963e-01 -1.39127523e-01 5.14814854e-01
1.23230577e+00 4.68858890e-02 6.01910949e-01 2.56654173e-01
4.19940293e-01 7.65061677e-01 3.61222059e-01 -7.26343021e-02
7.21754551e-01 7.51734078e-01 5.41654117e-02 1.21259645e-01
-5.25454223e-01 -4.55385268e-01 4.67344046e-01 5.07232547e-01
1.64045975e-01 -3.05821061e-01 -7.98531115e-01 8.47521305e-01
-1.91202450e+00 -1.04814243e+00 1.95494622e-01 1.68438816e+00
1.41276813e+00 -1.55639928e-02 -1.12300187e-01 -4.79662508e-01
4.42068279e-01 4.21437025e-01 -1.21799886e-01 -8.81802559e-01
-5.32873034e-01 8.85856077e-02 5.82229579e-04 7.90390551e-01
-5.19652069e-01 1.32053888e+00 7.41156244e+00 4.26293463e-01
-1.07100749e+00 3.60821605e-01 9.37058926e-01 -2.27977470e-01
-8.83341312e-01 4.55042154e-01 -3.57092679e-01 2.38606066e-01
9.92306411e-01 1.19517297e-01 6.71990335e-01 7.13274002e-01
1.37911484e-01 -1.34942815e-01 -1.40987027e+00 5.35548389e-01
1.72991261e-01 -1.47682643e+00 1.88180387e-01 -1.10348970e-01
6.67029977e-01 1.26323581e-01 -6.29417598e-02 2.32809454e-01
4.35004562e-01 -1.02868557e+00 1.17683899e+00 3.59137535e-01
6.77827835e-01 -1.18087724e-01 5.52464426e-01 1.38992012e-01
-3.07617962e-01 -8.42963159e-02 6.25671521e-02 -4.03660387e-01
6.27251148e-01 1.33741498e-01 -7.33750582e-01 2.97236978e-03
2.69211531e-01 1.67281643e-01 -3.38614762e-01 2.96653599e-01
-6.75559878e-01 4.95340914e-01 -3.57005507e-01 -1.45303696e-01
2.98063993e-01 -9.20356214e-02 6.87066138e-01 1.28284502e+00
-3.12542580e-02 2.01124415e-01 -1.93885460e-01 1.40655112e+00
-6.19148575e-02 8.45704079e-02 -7.97853053e-01 -1.43798456e-01
4.40109938e-01 9.37664211e-01 -5.21087885e-01 -2.09905580e-01
-2.41259709e-01 9.96180356e-01 6.04322255e-01 3.28890562e-01
-6.59447193e-01 3.14887226e-01 4.79853809e-01 -8.75060447e-03
-4.29283418e-02 -1.94075078e-01 -6.06873989e-01 -1.07318199e+00
1.49744347e-01 -8.25578034e-01 2.85617322e-01 -1.07242095e+00
-1.15530217e+00 4.03933585e-01 2.21192226e-01 -2.73464084e-01
-5.44183969e-01 -7.85139859e-01 -2.69748509e-01 7.98723578e-01
-1.28232503e+00 -1.45964861e+00 9.47765708e-02 -2.76058558e-02
4.78108019e-01 3.57695132e-01 1.03523159e+00 -1.66038781e-01
-2.90530592e-01 4.18031603e-01 -5.10753989e-01 1.00974189e-02
4.03707206e-01 -1.21082532e+00 5.02383590e-01 7.03782022e-01
2.97037363e-01 9.61753845e-01 7.19429731e-01 -3.57062459e-01
-1.26291382e+00 -6.46853149e-01 1.12279069e+00 -9.60127890e-01
5.51227808e-01 -3.32649797e-01 -5.51729739e-01 1.08375430e+00
3.69570494e-01 -2.39481255e-01 6.86597884e-01 4.65456009e-01
-3.87456775e-01 3.63173366e-01 -1.23818183e+00 1.18120432e+00
1.38221133e+00 -7.94524372e-01 -1.02681446e+00 3.71553957e-01
1.03246880e+00 -3.49910170e-01 -7.37115204e-01 2.04240158e-01
3.99952948e-01 -8.28842282e-01 7.05528140e-01 -6.55392706e-01
5.92905641e-01 -2.07991078e-01 -5.57377338e-01 -9.82984900e-01
-2.99248636e-01 -7.86520422e-01 2.35741392e-01 1.32646465e+00
6.90951645e-01 -5.04420459e-01 2.79185236e-01 8.48906219e-01
-4.17176425e-01 -7.66427636e-01 -1.13181353e+00 -3.18299532e-01
2.64581531e-01 -5.28331280e-01 5.97280562e-01 9.25096393e-01
5.41818976e-01 1.13521016e+00 -9.37160179e-02 -3.03163350e-01
-1.09670475e-01 1.83770165e-01 3.84702444e-01 -9.23652530e-01
-5.74741840e-01 -6.32110238e-01 -9.31263566e-02 -8.82417440e-01
4.38740343e-01 -1.10213649e+00 2.54401594e-01 -1.85021389e+00
9.00454074e-02 -4.98035192e-01 1.83936283e-01 6.10052228e-01
-4.38357890e-02 2.35601395e-01 3.40980440e-01 1.90135121e-01
-5.64080536e-01 5.41870952e-01 1.22510731e+00 1.23903245e-01
-6.82662949e-02 -6.19684517e-01 -1.64372993e+00 9.28229630e-01
6.81705296e-01 -1.54834166e-01 -4.71472114e-01 -9.19570386e-01
6.51729465e-01 -5.97956330e-02 5.00235140e-01 -4.50405687e-01
-4.42140698e-01 -2.96030879e-01 1.70688760e-02 3.54228735e-01
5.59824407e-01 -3.75061929e-01 -1.69594772e-02 1.19994385e-02
-7.74294794e-01 -2.68277116e-02 3.76058906e-01 3.57129991e-01
5.53017184e-02 -2.50850439e-01 6.16996527e-01 -5.99331498e-01
-7.32457757e-01 -4.86784875e-01 -2.62343973e-01 5.45276880e-01
4.85183984e-01 -3.24923456e-01 -6.95125759e-01 -4.53706175e-01
-9.06912625e-01 -2.86919028e-02 6.82783127e-01 3.24894279e-01
2.17494041e-01 -1.27523220e+00 -5.79340398e-01 -3.31277609e-01
2.22936362e-01 9.62934736e-03 -2.91059650e-02 8.26723337e-01
-2.63124079e-01 4.77700353e-01 -1.24869002e-02 -3.47409308e-01
-4.96937990e-01 4.09496576e-01 4.87367868e-01 7.37064481e-02
-6.59621537e-01 9.54423904e-01 4.06474918e-01 -4.65828866e-01
-2.26535231e-01 -5.23321867e-01 1.27710029e-01 -7.76020885e-02
7.45045841e-02 -3.66735756e-01 -2.64415830e-01 -8.60739410e-01
-3.06774586e-01 2.18854710e-01 1.96780086e-01 -6.23632014e-01
1.06878996e+00 -1.44076571e-01 -1.17105700e-01 4.69972014e-01
7.41346836e-01 3.41498822e-01 -1.20228982e+00 7.52974302e-02
2.40010902e-01 -2.32387096e-01 -1.20892778e-01 -1.31142461e+00
-5.89836001e-01 6.10382199e-01 1.13837093e-01 -2.25072175e-01
9.12660122e-01 5.57641745e-01 4.70541179e-01 4.17414069e-01
2.77621299e-01 -1.11436999e+00 -1.32014886e-01 4.38253164e-01
1.15208888e+00 -9.69233930e-01 -1.31134331e-01 -2.15780959e-01
-9.44903851e-01 5.68304002e-01 6.42139435e-01 -3.90598141e-02
-6.67207688e-02 9.93524566e-02 1.32519722e-01 -4.33665991e-01
-8.70124578e-01 -1.57098889e-01 3.20123392e-04 8.23638439e-01
8.56619179e-01 2.64345437e-01 -7.50085354e-01 7.65494287e-01
-6.33707166e-01 -1.41572341e-01 2.66399562e-01 6.98182166e-01
-1.67966560e-01 -9.68197405e-01 -8.45614895e-02 1.94309596e-02
-5.78538358e-01 -6.42276883e-01 -7.14056611e-01 1.03711855e+00
1.05500510e-02 6.70705616e-01 1.33340240e-01 1.46088362e-01
2.29595110e-01 3.47249418e-01 6.52375698e-01 -8.18984449e-01
-6.75623834e-01 -2.59597376e-02 9.11798358e-01 -4.87253368e-01
-4.64669049e-01 -8.84949505e-01 -1.28680444e+00 9.91710201e-02
-3.06670815e-01 -4.57220562e-02 3.76624823e-01 1.10190272e+00
7.66254902e-01 3.36364120e-01 -1.93491489e-01 -4.39806104e-01
-6.19685233e-01 -8.91500235e-01 6.59147650e-02 5.08575976e-01
2.43016630e-01 -1.89305902e-01 -2.18827561e-01 2.08314061e-02] | [10.71831226348877, 9.082273483276367] |
9d393049-8037-4da7-90a1-43503a0b485b | improving-end-to-end-slu-performance-with | 2305.08067 | null | https://arxiv.org/abs/2305.08067v1 | https://arxiv.org/pdf/2305.08067v1.pdf | Improving End-to-End SLU performance with Prosodic Attention and Distillation | Most End-to-End SLU methods depend on the pretrained ASR or language model features for intent prediction. However, other essential information in speech, such as prosody, is often ignored. Recent research has shown improved results in classifying dialogue acts by incorporating prosodic information. The margins of improvement in these methods are minimal as the neural models ignore prosodic features. In this work, we propose prosody-attention, which uses the prosodic features differently to generate attention maps across time frames of the utterance. Then we propose prosody-distillation to explicitly learn the prosodic information in the acoustic encoder rather than concatenating the implicit prosodic features. Both the proposed methods improve the baseline results, and the prosody-distillation method gives an intent classification accuracy improvement of 8\% and 2\% on SLURP and STOP datasets over the prosody baseline. | ['Shangeth Rajaa'] | 2023-05-14 | null | null | null | null | ['intent-classification'] | ['natural-language-processing'] | [ 1.85667947e-01 5.96678615e-01 -3.04753363e-01 -7.15557277e-01
-8.65809321e-01 -3.20352197e-01 3.63072783e-01 -2.19378695e-02
-4.29961145e-01 8.61112952e-01 1.13372481e+00 2.79298872e-02
6.12623453e-01 -3.63345981e-01 -2.25309148e-01 -5.11975288e-01
2.25549534e-01 2.43454546e-01 1.32486895e-01 -3.58204514e-01
2.67369688e-01 -2.15331137e-01 -9.08590257e-01 5.28217435e-01
7.12167799e-01 1.01937795e+00 3.69273514e-01 9.00098681e-01
-3.64381224e-01 1.07586396e+00 -7.53387630e-01 -6.83002025e-02
-1.23186700e-01 -7.52253771e-01 -8.07428837e-01 -2.00133815e-01
-8.44787713e-03 -4.20648545e-01 -2.54706770e-01 7.11748481e-01
5.89602232e-01 2.84490585e-01 5.26935935e-01 -6.22324347e-01
-6.60023034e-01 1.19064999e+00 -4.51420605e-01 2.81185091e-01
3.84856701e-01 2.18639169e-02 1.49212313e+00 -7.77922690e-01
2.51411885e-01 1.32909405e+00 5.27892888e-01 8.51110518e-01
-1.26115656e+00 -5.66362739e-01 1.01042598e-01 2.14474902e-01
-8.18331897e-01 -7.47206688e-01 1.22480893e+00 -3.19980323e-01
1.20299697e+00 2.27350920e-01 2.16936782e-01 1.06047690e+00
2.24602684e-01 1.19016600e+00 8.21753442e-01 -5.87180018e-01
2.42298730e-02 1.09861113e-01 7.81256080e-01 2.47804895e-01
-7.76437998e-01 -1.72601268e-02 -5.72873473e-01 1.22502625e-01
4.44781095e-01 -3.61853510e-01 -3.44105035e-01 4.92062986e-01
-8.90841663e-01 1.01559722e+00 2.74116307e-01 2.64802784e-01
-3.63670737e-01 -1.07522003e-01 8.63960505e-01 1.88281879e-01
6.92395210e-01 6.35032773e-01 -5.66761374e-01 -6.86192214e-01
-7.68897533e-01 -4.97953705e-02 8.03792715e-01 7.37938762e-01
4.00633007e-01 3.51131469e-01 -4.34699446e-01 1.39018595e+00
2.74938226e-01 9.67606753e-02 7.63246834e-01 -1.10668933e+00
6.05511010e-01 2.01329723e-01 -9.97325033e-02 -5.61850131e-01
-3.63133311e-01 5.69130816e-02 -5.36523223e-01 -2.27157250e-01
1.22200318e-01 -6.48398876e-01 -8.58488917e-01 1.98814905e+00
-1.55676678e-01 -6.88939989e-02 4.49040890e-01 7.81633377e-01
9.90419924e-01 1.27916527e+00 2.84696460e-01 -5.82407951e-01
1.18967962e+00 -1.39211535e+00 -1.34866905e+00 -3.12604070e-01
5.53103745e-01 -7.64514089e-01 1.43382728e+00 3.19091588e-01
-1.28126895e+00 -8.23797405e-01 -1.03442585e+00 -3.07232589e-01
8.17089826e-02 2.00082615e-01 2.72017300e-01 5.46346068e-01
-8.96823645e-01 2.63477862e-01 -7.73324072e-01 3.54221463e-02
-4.00460325e-02 3.47775787e-01 -5.63481264e-02 4.42455053e-01
-1.31317711e+00 6.51148260e-01 7.26699889e-01 -1.55075341e-01
-3.94562453e-01 -7.67548144e-01 -1.17575336e+00 3.10443878e-01
1.44116849e-01 -1.73971020e-02 1.65706539e+00 -8.77911806e-01
-2.24298096e+00 4.05627877e-01 -2.95560867e-01 -7.10073769e-01
-3.52969505e-02 -5.20380437e-01 -4.65972573e-02 1.95715204e-02
-2.18442947e-01 1.01800346e+00 2.85413742e-01 -7.54858136e-01
-6.95673227e-01 -1.18061751e-02 -1.70570109e-02 5.66270292e-01
-2.33577207e-01 3.54393542e-01 -2.01906309e-01 -7.44354606e-01
-2.37332284e-01 -7.97342598e-01 -1.88364405e-02 -6.61723375e-01
-4.20908809e-01 -4.96157318e-01 1.03994966e+00 -1.10229933e+00
1.47164834e+00 -2.30838180e+00 1.87802866e-01 -5.74058950e-01
-3.56496960e-01 2.01852411e-01 -5.91263399e-02 2.67983764e-01
5.35275321e-03 1.96076617e-01 -2.97987640e-01 -6.83164477e-01
2.80041154e-02 3.78317535e-01 -6.56537592e-01 -1.66278973e-01
3.90466481e-01 7.11756945e-01 -5.71232557e-01 -1.71854645e-01
2.70382613e-01 4.36988443e-01 -9.00680006e-01 4.88469362e-01
-4.24219221e-01 7.77814925e-01 3.19980495e-02 1.41292289e-01
1.34210527e-01 4.42628354e-01 1.12927943e-01 1.08501568e-01
-1.39857411e-01 1.40847206e+00 -5.22833347e-01 1.75206017e+00
-6.38424397e-01 6.82492495e-01 1.25598580e-01 -7.35322118e-01
1.16143250e+00 8.86007071e-01 2.30108440e-01 -4.34128463e-01
1.28550619e-01 -2.45507974e-02 5.34544051e-01 -1.77025333e-01
6.06743276e-01 -4.92054790e-01 -4.29935485e-01 2.76190102e-01
2.12201193e-01 -2.06355050e-01 6.45346381e-03 -2.00137988e-01
8.62799764e-01 -1.47486739e-02 4.63484496e-01 -2.72007525e-01
7.03123271e-01 -2.37126589e-01 9.70105529e-01 3.37442845e-01
-3.00999075e-01 7.29490519e-01 6.53447926e-01 -1.41321257e-01
-9.47980881e-01 -8.13425124e-01 -1.38059169e-01 1.51637900e+00
-3.12592238e-01 -5.37952363e-01 -8.73417199e-01 -7.84633875e-01
-5.14067233e-01 1.25425506e+00 -2.44161680e-01 -3.06310877e-02
-1.08145583e+00 -6.77456737e-01 6.62036121e-01 8.05763960e-01
5.27793765e-01 -1.63515902e+00 -1.11162327e-01 5.09672403e-01
-4.18583214e-01 -1.22267723e+00 -8.92953575e-01 3.13852280e-01
-6.55354917e-01 -4.52683449e-01 -5.31713843e-01 -9.28168535e-01
8.70367140e-02 -3.68428439e-01 6.91790342e-01 -3.71859908e-01
2.75278121e-01 -1.50693730e-01 -6.44860387e-01 -3.95094603e-01
-7.21772134e-01 1.52707189e-01 1.18911371e-01 -1.50485471e-01
4.56258029e-01 -4.07542378e-01 -1.93666995e-01 6.13357648e-02
-2.75763869e-01 2.15560243e-01 2.36904845e-01 1.13800156e+00
3.66205096e-01 -2.15176910e-01 1.22883546e+00 -8.80750716e-01
8.49912107e-01 -2.62148470e-01 2.97061559e-02 -3.18791449e-01
-1.90619379e-01 1.17820628e-01 7.65548944e-01 -2.59372830e-01
-1.51528978e+00 8.68042000e-03 -8.11067581e-01 -2.26872057e-01
-4.88171041e-01 5.54999828e-01 -5.28049588e-01 7.02616513e-01
2.51592398e-01 1.98849663e-01 4.82652076e-02 -6.93682134e-01
2.56967127e-01 1.04812884e+00 3.61389339e-01 -5.03886521e-01
-6.60936162e-02 -3.04071009e-01 -7.04953611e-01 -1.19227588e+00
-1.05453432e+00 -4.66158360e-01 -6.68936431e-01 2.20085800e-01
1.23175061e+00 -7.28733063e-01 -4.20811951e-01 3.47861350e-01
-1.51243496e+00 -5.13244987e-01 -3.05708975e-01 6.73843443e-01
-9.14572418e-01 5.15892267e-01 -1.28827775e+00 -1.01931000e+00
-5.73391020e-01 -1.41913724e+00 8.29687238e-01 2.37276822e-01
-6.64363205e-01 -8.95551562e-01 1.16490513e-01 5.41904271e-01
2.45629355e-01 -2.19407827e-01 1.01658583e+00 -1.06640840e+00
4.67785522e-02 2.12466151e-01 -2.87161954e-03 7.98769176e-01
4.50438499e-01 -3.80395114e-01 -1.25424910e+00 1.24511197e-01
3.59984905e-01 -4.02065247e-01 7.63337314e-01 5.30446410e-01
8.62071335e-01 -7.62046754e-01 6.26775399e-02 1.95654303e-01
7.12323129e-01 8.24750960e-01 5.42109728e-01 -3.57786000e-01
5.65793574e-01 9.00645852e-01 7.10416973e-01 1.00223735e-01
3.40154469e-01 7.56944418e-01 -5.46635278e-02 2.11761847e-01
-3.66393119e-01 -2.57940412e-01 9.07060027e-01 1.52389741e+00
4.67990488e-02 -2.23929241e-01 -8.44920576e-01 6.72839344e-01
-1.82638562e+00 -8.38321209e-01 8.20387602e-02 1.96519029e+00
1.50210953e+00 3.36436450e-01 2.03305170e-01 1.52271181e-01
6.96928918e-01 4.29972619e-01 -1.02185093e-01 -9.31375742e-01
1.91318333e-01 2.03655094e-01 -5.27264401e-02 1.01077950e+00
-1.15790153e+00 1.29458988e+00 6.49034166e+00 7.08726645e-01
-1.15991735e+00 1.97726548e-01 7.33492672e-01 6.26044022e-03
-4.17764392e-03 -2.10939199e-01 -1.07114887e+00 4.74995911e-01
1.36602330e+00 -4.90074866e-02 1.40706852e-01 9.55506444e-01
5.79100013e-01 6.85500354e-02 -1.30895495e+00 5.93612909e-01
1.96017027e-01 -1.03193390e+00 -1.90258533e-01 -2.40696277e-02
3.86444002e-01 -9.63161364e-02 -2.04377577e-01 8.49283516e-01
2.16536075e-01 -9.18742001e-01 5.08372962e-01 2.45074183e-01
4.52679873e-01 -9.04273272e-01 9.64222908e-01 5.52816093e-01
-1.03186381e+00 8.89961720e-02 -3.03702772e-01 -4.34963673e-01
7.26332903e-01 1.11652305e-02 -1.39607680e+00 1.55890226e-01
2.03655705e-01 6.70432270e-01 -9.47764516e-03 6.69292927e-01
-4.96911615e-01 1.32104874e+00 -1.91370800e-01 -7.27649033e-02
3.51757884e-01 -1.32718235e-01 7.65022516e-01 1.56876111e+00
-2.93333735e-02 2.76342034e-01 3.77676964e-01 7.80310392e-01
-6.23584492e-04 3.64809602e-01 -5.55968702e-01 -1.54383063e-01
2.33208969e-01 8.10773551e-01 -2.56913900e-01 -1.96442589e-01
-4.93741721e-01 8.75417054e-01 1.44635662e-01 1.28936574e-01
-8.21385741e-01 -4.49847937e-01 8.10178399e-01 -1.62184030e-01
2.44675115e-01 -3.03497195e-01 -5.14939487e-01 -7.86058187e-01
-2.44964942e-01 -5.90499103e-01 2.38519415e-01 -5.73800206e-01
-1.15670049e+00 6.11025810e-01 -5.10381535e-02 -6.81859076e-01
-7.96915531e-01 -5.61827719e-01 -1.01177359e+00 1.05500650e+00
-1.35498846e+00 -9.54343081e-01 3.27249348e-01 -1.05623126e-01
1.64969063e+00 -1.68931171e-01 9.39851761e-01 1.27277011e-02
-5.50852895e-01 5.62736094e-01 -4.15371746e-01 5.53592980e-01
7.67582476e-01 -1.55033827e+00 2.44269416e-01 5.82080543e-01
-7.36168772e-03 4.64897543e-01 6.71541989e-01 -4.76040393e-01
-6.10521674e-01 -8.47919583e-01 1.08571815e+00 -3.28830957e-01
5.77681780e-01 -4.77931947e-01 -1.18458879e+00 8.39271307e-01
9.63471055e-01 -4.77784872e-01 1.11887431e+00 4.22543645e-01
1.43729467e-02 1.78159058e-01 -8.79323840e-01 6.26400352e-01
5.86742878e-01 -4.59801555e-01 -1.43735087e+00 -1.84101656e-01
1.39616561e+00 -2.71250218e-01 -8.60448718e-01 4.80028003e-01
3.25437635e-01 -7.36802101e-01 5.57797790e-01 -5.28266072e-01
5.39120138e-01 1.09385133e-01 -3.72082442e-01 -1.35449886e+00
-2.12477505e-01 -5.94039679e-01 -8.85105431e-02 1.59609425e+00
6.74214005e-01 -3.18262577e-01 6.13427818e-01 5.59242308e-01
-7.76481211e-01 -4.71828103e-01 -9.22715127e-01 -4.46694434e-01
2.20676273e-01 -3.57436091e-01 7.15680718e-02 7.80613601e-01
7.00015485e-01 1.18180311e+00 -5.87906063e-01 -1.48118660e-01
1.51877496e-02 -2.55783051e-01 4.54996914e-01 -9.52113748e-01
-6.18831873e-01 -5.36483645e-01 -1.37696505e-01 -1.46051776e+00
5.32086611e-01 -7.09348857e-01 6.03480816e-01 -1.27032542e+00
-8.77147317e-02 -1.85174689e-01 -2.79724896e-01 7.38638818e-01
-3.79601628e-01 -1.13284260e-01 4.15144771e-01 3.26523334e-02
-2.16547087e-01 9.80807602e-01 1.04249513e+00 -2.75803626e-01
-7.37189949e-01 1.22590333e-01 -4.57233012e-01 8.96781206e-01
1.08248234e+00 -1.97085246e-01 -5.34938991e-01 -1.60177976e-01
-6.03792608e-01 5.50066590e-01 -3.74564856e-01 -9.17415619e-01
1.03611099e-02 -1.25950933e-01 -8.84502307e-02 -8.08144093e-01
9.07675803e-01 -4.12465394e-01 -5.76319993e-01 2.39846393e-01
-7.78674126e-01 -2.86758959e-01 3.98271710e-01 4.36552048e-01
-6.35669470e-01 -5.18332064e-01 7.82123148e-01 -1.60825416e-01
-5.93721688e-01 -2.73105919e-01 -6.04867578e-01 5.49560674e-02
7.34795630e-01 2.93518882e-03 -2.35488102e-01 -5.76019704e-01
-1.09858108e+00 1.91487163e-01 -1.14546314e-01 5.71657240e-01
3.81473094e-01 -1.15697742e+00 -7.24469662e-01 3.24839413e-01
-2.96210945e-01 1.01812333e-01 1.77347496e-01 8.20622921e-01
-1.80977270e-01 7.52177894e-01 -1.26352772e-01 -4.83769178e-01
-1.40445018e+00 2.99293339e-01 1.65433168e-01 -4.03206885e-01
-5.44974625e-01 8.21631014e-01 7.66340494e-01 -5.99082708e-01
4.97095942e-01 -5.36328495e-01 -5.75172186e-01 1.17425889e-01
6.00671649e-01 2.56792773e-02 -2.35864475e-01 -6.73453510e-01
-9.85011980e-02 1.54278353e-01 -4.73536372e-01 -4.87884164e-01
1.23391354e+00 -2.32465014e-01 1.62090912e-01 9.38436210e-01
9.54623759e-01 4.56706285e-01 -1.42961407e+00 -1.32441837e-02
2.17096269e-01 1.57655463e-01 5.30147888e-02 -9.38398063e-01
-5.51954329e-01 1.32068443e+00 -2.58014183e-02 4.05763626e-01
9.69728410e-01 -3.77881229e-02 1.03639305e+00 2.08877891e-01
1.74543932e-02 -1.07587433e+00 1.29868671e-01 1.17166483e+00
1.17970943e+00 -1.13655782e+00 -6.89024925e-01 -5.83723068e-01
-1.25544882e+00 1.02484250e+00 8.97424757e-01 -1.89927503e-01
6.09180868e-01 3.33809763e-01 1.66522473e-01 3.54550540e-01
-9.51099694e-01 -7.29144141e-02 2.08988562e-01 4.06462222e-01
1.00372648e+00 3.88331115e-02 -7.00361848e-01 1.23713052e+00
-5.01926482e-01 -3.84125620e-01 5.82674801e-01 5.39081514e-01
-7.96476901e-01 -1.30293095e+00 -8.82682353e-02 1.53598800e-01
-7.05980599e-01 -4.54919487e-01 -5.36754549e-01 5.60484529e-01
-2.79695302e-01 9.53251541e-01 4.14487034e-01 -4.87053275e-01
1.86200306e-01 7.97360659e-01 -7.39150867e-02 -1.28328347e+00
-7.08578289e-01 5.87776601e-01 6.20564759e-01 -2.13958323e-01
-1.48723319e-01 -6.13218427e-01 -1.66586769e+00 3.23723286e-01
-2.35009894e-01 4.50543821e-01 3.25102448e-01 9.69546318e-01
3.01770717e-01 7.11319208e-01 6.67042613e-01 -8.73892665e-01
-5.39942443e-01 -1.64802229e+00 -2.75487244e-01 1.23524413e-01
4.04188663e-01 -4.68177676e-01 -3.89577150e-01 1.17416248e-01] | [14.417622566223145, 6.989426136016846] |
88ad6cc4-8fb0-4afb-b3f7-8b39eb95036b | automated-crystal-orientation-mapping-by | 2102.09711 | null | https://arxiv.org/abs/2102.09711v2 | https://arxiv.org/pdf/2102.09711v2.pdf | Automated crystal orientation mapping by precession electron diffraction assisted four-dimensional scanning transmission electron microscopy (4D-STEM) using a scintillator based CMOS detector | The recent development of electron sensitive and pixelated detectors has attracted the use of four-dimensional scanning transmission electron microscopy (4D-STEM). Here, we present a precession electron diffraction assisted 4D-STEM technique for automated orientation mapping using diffraction spot patterns directly captured by an in-column scintillator based complementary metal-oxide-semiconductor (CMOS) detector. We compare the results to a conventional approach, which utilizes a fluorescent screen filmed by an external CCD camera. The high dynamic range and signal-to-noise characteristics of the detector largely improve the image quality of the diffraction patterns, especially the visibility of diffraction spots at high scattering angles. In the orientation maps reconstructed via the template matching process, the CMOS data yields a significant reduction of false indexing and higher reliability compared to the conventional approach. The angular resolution of misorientation measurement could also be improved by masking reflections close to the direct beam. This is because the orientation sensitive, weak and small diffraction spots at high scattering angle are more significant. The results show that fine details such as nanograins, nanotwins and sub-grain boundaries can be resolved with a sub-degree angular resolution which is comparable to orientation mapping using Kikuchi diffraction patterns. | ['Christian H. Liebscher', 'Gerhard Dehm', 'Niels Cautaerts', 'Jiwon Jeong'] | 2021-02-19 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [ 7.17245579e-01 -1.82036668e-01 5.56230009e-01 -2.01077104e-01
-5.83737731e-01 -7.16632307e-02 4.61758018e-01 8.57317522e-02
-9.65555787e-01 6.36278629e-01 -2.95775503e-01 1.81097314e-01
8.06020871e-02 -7.40767300e-01 -4.09881234e-01 -1.09843314e+00
4.08551663e-01 1.04454744e+00 9.66412842e-01 -5.83976880e-02
5.07120609e-01 6.74412012e-01 -1.64422572e+00 1.88773096e-01
2.30915561e-01 9.55670476e-01 9.83353734e-01 3.29860210e-01
5.69344200e-02 2.47482017e-01 -2.98193663e-01 1.22371629e-01
-2.26602659e-01 -2.74470329e-01 -1.95505381e-01 1.60895467e-01
-1.28830105e-01 -3.66480738e-01 3.29066157e-01 1.11427009e+00
1.92545578e-01 -3.43535155e-01 5.66453159e-01 -3.52904588e-01
-5.65152824e-01 8.74327198e-02 -5.83121717e-01 2.41712034e-01
6.88566327e-01 1.37379870e-01 3.67508560e-01 -1.31138122e+00
1.16729772e+00 1.01065457e+00 3.93353045e-01 5.40498674e-01
-1.27944815e+00 -4.63713348e-01 -7.96313703e-01 3.92918557e-01
-1.16739023e+00 -4.64552969e-01 7.29173601e-01 -4.69616413e-01
1.17147231e+00 4.44756418e-01 3.98007274e-01 6.97597861e-01
6.10856652e-01 -2.71311074e-01 1.88465202e+00 -8.25359285e-01
3.34445238e-01 2.19659761e-01 3.78561318e-01 4.68581468e-01
6.96778953e-01 2.13567317e-01 -1.63622767e-01 3.22183132e-01
6.16650820e-01 3.52695227e-01 -4.70005661e-01 -2.90004641e-01
-1.15651202e+00 1.34310946e-01 1.09155610e-01 8.52357686e-01
-4.25624907e-01 -3.59357387e-01 -8.53298455e-02 -5.86763099e-02
2.97511369e-01 5.48878193e-01 1.67149246e-01 -6.20338432e-02
-8.33923280e-01 2.69057080e-02 4.26671028e-01 5.35073996e-01
5.91760635e-01 -1.99084565e-01 1.63243100e-01 2.48416647e-01
4.64039803e-01 8.30825210e-01 3.24259907e-01 -6.06204867e-01
-5.57073206e-02 5.84168017e-01 2.61395395e-01 -4.82848942e-01
-1.66025400e-01 3.02800566e-01 -5.39387286e-01 7.11262524e-01
6.11164212e-01 3.47238362e-01 -9.89872992e-01 8.80324185e-01
4.33564276e-01 -4.36531305e-01 -7.64001682e-02 1.32131791e+00
3.75943065e-01 4.16361958e-01 -5.86090207e-01 -5.80629408e-01
1.61146080e+00 -4.36095864e-01 -9.83520210e-01 -1.06378883e-01
1.24781728e-01 -9.22543228e-01 9.65540588e-01 1.96271643e-01
-9.88774836e-01 -2.79789288e-02 -1.45513642e+00 9.03723016e-02
-2.27649614e-01 -1.08450510e-01 -1.62275150e-01 4.80892956e-01
-5.04059255e-01 6.77602887e-01 -1.17261434e+00 -4.73134309e-01
1.59722149e-01 5.13744473e-01 -5.09314954e-01 -1.22071333e-01
-5.51604092e-01 7.28924990e-01 3.48774411e-05 -4.16162878e-01
-1.86290555e-02 -3.39936733e-01 -1.11795411e-01 -1.33873165e-01
1.89783767e-01 -1.36067346e-01 8.73283803e-01 -2.50134200e-01
-1.82977927e+00 1.41068375e+00 -4.39169466e-01 -3.19522321e-01
6.82820156e-02 6.78427219e-02 -5.32982051e-01 6.38330460e-01
4.14043754e-01 -2.99787134e-01 5.18053949e-01 -1.06535769e+00
-2.97187358e-01 -1.00621712e+00 -6.85753942e-01 -3.77324484e-02
1.24907859e-01 4.48282957e-01 -5.43708205e-02 8.57967734e-02
6.01869285e-01 -6.01444900e-01 -1.42351732e-01 -3.52499306e-01
-2.22866312e-01 1.07963160e-01 1.37019467e+00 -3.13707292e-01
8.78338754e-01 -2.19751692e+00 -2.86553681e-01 1.78372696e-01
2.28743702e-01 1.51663348e-01 6.95049107e-01 7.36039221e-01
1.11644924e-01 -4.97437507e-01 -1.16288476e-01 1.43638079e-03
-2.89312214e-01 -2.04884753e-01 2.50492513e-01 8.23437691e-01
-2.03629166e-01 4.34647650e-01 -5.04235983e-01 -2.85160840e-01
3.88420850e-01 4.75945085e-01 8.40408131e-02 -5.62942177e-02
2.06631690e-01 5.24653971e-01 -4.92018908e-01 7.59065330e-01
8.67763400e-01 -5.02325058e-01 2.87210196e-01 -1.92707777e-01
-8.97072792e-01 1.37665927e-01 -9.86506760e-01 1.02741468e+00
4.91354018e-01 2.20493719e-01 5.24398446e-01 -5.53900480e-01
1.10504246e+00 3.54623616e-01 2.44358465e-01 -1.53597951e+00
-6.76482767e-02 7.45813310e-01 -2.65306860e-01 -4.84945118e-01
3.24799001e-01 -8.35785031e-01 4.71739382e-01 4.17169183e-01
-3.95295560e-01 -6.29722849e-02 3.87451500e-02 -3.26225549e-01
8.49263012e-01 -2.30355680e-01 2.59252876e-01 -8.36032391e-01
5.14550924e-01 1.09172277e-01 1.51398316e-01 4.50922340e-01
2.38993078e-01 8.69890392e-01 -5.25321402e-02 -4.95207280e-01
-1.71587014e+00 -6.57690525e-01 -5.83591521e-01 8.92391726e-02
3.68298441e-01 -6.13374077e-02 -1.04326880e+00 3.74976069e-01
-2.54209489e-01 2.97796100e-01 -1.58860132e-01 1.76078275e-01
-5.45539796e-01 -1.05384195e+00 -3.10709268e-01 3.42024565e-01
3.15906048e-01 -9.70649660e-01 -1.19669616e+00 4.17107493e-01
4.80494082e-01 -1.14751780e+00 2.15510130e-01 2.92673439e-01
-8.93083751e-01 -1.03818774e+00 -6.51483774e-01 -4.87350315e-01
7.81257272e-01 3.16278338e-01 8.08516204e-01 -2.84151554e-01
-4.57973063e-01 3.38321686e-01 -3.90108347e-01 -2.04412565e-01
-2.39313141e-01 -4.17212993e-01 2.88489103e-01 -1.57477096e-01
9.84660745e-01 -6.23227000e-01 -7.46220887e-01 4.07350093e-01
-6.04358613e-01 7.86709562e-02 5.02107441e-01 6.57586336e-01
1.12772691e+00 -6.36867881e-02 5.03373407e-02 -9.60527599e-01
3.30682695e-01 2.77099431e-01 -1.03840435e+00 -2.29789853e-01
-1.11499119e+00 -1.60758030e-02 6.19797409e-01 3.69793065e-02
-1.02005863e+00 -2.64198035e-01 -1.69323862e-01 -1.07690051e-01
-4.66520458e-01 8.13114718e-02 -1.44409329e-01 -1.35901719e-01
6.47864103e-01 2.38972515e-01 3.16296577e-01 -5.51562667e-01
-8.01773429e-01 7.50868142e-01 4.86028850e-01 -1.50022358e-01
5.76712668e-01 9.19330418e-01 4.07070816e-01 -1.25283039e+00
-3.94214958e-01 -7.13944018e-01 -3.11144352e-01 -2.99859583e-01
1.15504944e+00 -5.75933933e-01 -7.18748868e-01 7.39896238e-01
-9.77847219e-01 1.07332222e-01 -3.00642550e-01 9.30353940e-01
-1.46961421e-01 4.61478382e-01 -9.13444221e-01 -8.95267129e-01
-4.24193293e-01 -1.36296940e+00 1.09483945e+00 2.57649243e-01
-1.57921925e-01 -6.19709194e-01 2.34187562e-02 4.18990225e-01
5.12125731e-01 4.35012877e-01 8.91569316e-01 -1.07047230e-01
-8.09626400e-01 -4.58808839e-01 -6.22312501e-02 -8.53917003e-02
5.12873866e-02 -1.26026407e-01 -8.48592937e-01 -1.28805876e-01
8.01080704e-01 6.98096529e-02 5.38506866e-01 6.35124028e-01
4.36854541e-01 3.73304486e-01 -7.36858428e-01 4.09381241e-01
1.84885359e+00 7.29635417e-01 1.06168592e+00 8.89894426e-01
6.95065618e-01 3.64977419e-01 8.72569144e-01 2.11427718e-01
-2.47642845e-01 9.12381411e-01 1.87273264e-01 -6.92644343e-02
1.41781449e-01 3.60571504e-01 1.16915539e-01 7.22920358e-01
-6.75354958e-01 -1.02904074e-01 -8.36043715e-01 1.12590238e-01
-1.20638108e+00 -1.17721343e+00 -9.57491040e-01 2.45958519e+00
5.25970936e-01 1.80349305e-01 -1.47560060e-01 5.29754937e-01
8.47701132e-01 -2.88533449e-01 -3.06556433e-01 -2.15849802e-01
-4.04270828e-01 6.97644889e-01 7.41165042e-01 8.37002754e-01
-4.83135045e-01 4.57830667e-01 6.54115868e+00 4.83654857e-01
-1.38989735e+00 -1.66391919e-03 9.79331285e-02 9.51645225e-02
-2.67038286e-01 6.51391968e-02 -9.67476070e-01 8.53831768e-01
6.45928144e-01 2.18499526e-01 1.79379910e-01 5.54947138e-01
3.07470590e-01 -7.08522201e-01 -8.05550575e-01 9.92700577e-01
-3.03359032e-01 -1.40494990e+00 -2.57222563e-01 5.71306348e-01
3.21960300e-01 -8.42683092e-02 -6.84908703e-02 -9.29304361e-01
-4.84242141e-01 -7.38872647e-01 5.98901987e-01 5.79136550e-01
1.10981441e+00 -6.84857845e-01 6.90777481e-01 1.13888472e-01
-8.24521840e-01 3.57127726e-01 -3.91026169e-01 -3.06622475e-01
4.55404162e-01 1.12922430e+00 -9.81937945e-01 1.26877993e-01
1.00567937e+00 2.49050736e-01 -1.81299254e-01 4.90840822e-01
3.38901430e-01 3.70065629e-01 -6.65669382e-01 -2.21668929e-01
4.80025373e-02 -1.03423131e+00 8.58405232e-01 8.98498297e-01
5.93102753e-01 3.76181990e-01 -4.70915377e-01 9.10720468e-01
3.99568498e-01 -3.10605049e-01 -5.60893834e-01 -9.82143879e-02
3.90964001e-01 1.15562439e+00 -1.35346043e+00 -2.83360213e-01
-5.88121116e-01 8.98287237e-01 -2.01327249e-01 9.81464535e-02
-1.15896225e-01 -4.25334305e-01 9.99484677e-03 1.05500937e+00
6.11255884e-01 -2.44262189e-01 -2.97336608e-01 -9.92322206e-01
1.58988446e-01 -4.96833712e-01 -2.40868449e-01 -7.51686037e-01
-8.71610045e-01 4.63417292e-01 -2.09769756e-01 -7.57410347e-01
1.07742801e-01 -1.00248134e+00 -3.68421853e-01 8.94181550e-01
-9.24835920e-01 -5.25297165e-01 -1.28261819e-01 3.94551218e-01
-1.29855471e-02 1.09747343e-01 9.80889261e-01 2.00891122e-01
-2.34486550e-01 -3.07375222e-01 7.53707290e-01 -3.81557435e-01
6.22263193e-01 -1.31598520e+00 -2.29177661e-02 8.68397236e-01
-5.59475124e-01 3.17957133e-01 1.09815228e+00 -6.70837700e-01
-1.64795423e+00 -3.45106602e-01 1.04275739e+00 -3.94578278e-01
2.52196938e-01 -3.79066199e-01 -8.98943245e-01 3.05158794e-01
2.23896369e-01 -3.38909924e-01 6.25321507e-01 -5.23627758e-01
4.35409307e-01 -4.03328724e-02 -1.45318604e+00 2.12607458e-01
6.39165103e-01 -5.21165490e-01 -5.84784627e-01 1.95844442e-01
-9.99170355e-03 -2.14469224e-01 -6.84724271e-01 4.58320558e-01
8.09753716e-01 -1.54892635e+00 6.13785684e-01 2.09556863e-01
3.28662336e-01 -4.71464068e-01 -2.33573139e-01 -5.47991395e-01
-6.11865401e-01 -6.03083484e-02 5.38950741e-01 1.01212764e+00
2.27783874e-01 -8.66562903e-01 1.05413759e+00 4.91649032e-01
-3.37258190e-01 -4.46607411e-01 -1.01446450e+00 -6.29269004e-01
-4.73349094e-01 3.54052216e-01 4.52048451e-01 6.54712558e-01
1.99061915e-01 6.50281310e-01 1.14193447e-01 7.80921578e-02
1.09044349e+00 3.13593805e-01 -2.75484100e-02 -1.48301232e+00
-2.13392586e-01 2.39735499e-01 -6.07437015e-01 -5.52036762e-01
-5.57557881e-01 -4.80699360e-01 -2.31414422e-01 -1.22866273e+00
2.11314484e-01 -8.13970342e-02 1.80654258e-01 -2.99139947e-01
3.07511985e-01 4.26154315e-01 -2.77669638e-01 6.30224049e-01
-4.81040955e-01 1.91082373e-01 1.41890454e+00 2.71454841e-01
-6.19361084e-03 -2.64761984e-01 -7.54283369e-02 5.85699439e-01
4.04425889e-01 -7.56110549e-01 7.59286806e-02 -1.82147175e-01
3.31257761e-01 5.86674847e-02 2.05540225e-01 -1.30695987e+00
4.16914970e-01 2.16036946e-01 5.00413656e-01 -7.22915292e-01
2.97007769e-01 -1.18575573e+00 8.98714542e-01 7.57190108e-01
4.46647555e-01 1.32603785e-02 -3.61183882e-01 3.27519715e-01
-3.56242597e-01 -4.32584167e-01 1.04848003e+00 -5.93825817e-01
-5.40920615e-01 -2.68483877e-01 -9.82490540e-01 -6.43433511e-01
1.12956583e+00 -1.01337767e+00 -3.90026271e-01 2.26895288e-01
-7.57280946e-01 -4.52483505e-01 1.37403333e+00 -6.35148823e-01
7.78209448e-01 -1.22371912e+00 1.65674016e-02 4.17074174e-01
8.28753505e-03 1.42622724e-01 3.60852659e-01 1.20274317e+00
-1.07867658e+00 5.59486806e-01 -4.47570235e-01 -8.91079545e-01
-1.36143219e+00 3.58193159e-01 3.34721684e-01 -2.94204414e-01
-9.37592804e-01 5.07694185e-01 3.06576286e-02 1.96130395e-01
-5.82006037e-01 -1.04037739e-01 -2.08508208e-01 -3.07348490e-01
8.23026478e-01 5.61816752e-01 4.47950780e-01 -5.63650608e-01
-6.30091369e-01 9.21331286e-01 -1.60900623e-01 -1.81819826e-01
1.60626614e+00 -3.31984848e-01 -2.14659378e-01 6.15772605e-01
9.35001254e-01 1.63843140e-01 -1.09674942e+00 1.02426343e-01
8.16747826e-03 -2.60264188e-01 7.58741349e-02 -4.76771325e-01
-5.74195743e-01 6.75490081e-01 8.09051514e-01 2.66514361e-01
9.71119583e-01 1.06763221e-01 5.38538814e-01 1.44524544e-01
5.51147044e-01 -1.38807511e+00 -3.40715438e-01 1.44623592e-01
3.09103727e-01 -9.42720950e-01 5.70202842e-02 -6.06793761e-01
-1.26613274e-01 1.46896195e+00 2.99613103e-02 -2.82808244e-01
3.74424011e-01 9.32767987e-01 -1.15858451e-01 -9.11553979e-01
-3.91255498e-01 2.61825733e-02 -2.39702597e-01 5.75356126e-01
3.63141507e-01 2.54432820e-02 -5.19485116e-01 2.84806281e-01
3.82241420e-02 2.20500156e-01 5.57300866e-01 1.20534277e+00
-7.82343686e-01 -1.04639220e+00 -8.80038917e-01 4.38291579e-01
-6.64430201e-01 1.45177484e-01 -3.62208307e-01 6.59696817e-01
-3.27143341e-01 6.35572374e-01 3.57317179e-01 -1.32123739e-01
5.14707327e-01 2.79174466e-02 8.16454589e-01 -4.61051822e-01
1.16243243e-01 4.80334222e-01 5.42407967e-02 -5.04241765e-01
-9.41443443e-01 -7.26921380e-01 -1.40242410e+00 -1.85409486e-01
-4.58249688e-01 3.10781687e-01 7.15115070e-01 1.07234919e+00
2.71701962e-01 1.32786006e-01 2.85780340e-01 -8.40247035e-01
-1.59131929e-01 -8.02396774e-01 -1.34450579e+00 3.04432631e-01
2.21381918e-01 -6.07984960e-01 -6.37896895e-01 -7.41921067e-02] | [12.909872055053711, -2.7917346954345703] |
73da5a1a-9b1f-443b-a002-ac5bb6f6c7e6 | on-linear-structure-from-motion-for-light | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Johannsen_On_Linear_Structure_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Johannsen_On_Linear_Structure_ICCV_2015_paper.pdf | On Linear Structure From Motion for Light Field Cameras | We present a novel approach to relative pose estimation which is tailored to 4D light field cameras. From the relationships between scene geometry and light field structure and an analysis of the light field projection in terms of Pluecker ray coordinates, we deduce a set of linear constraints on ray space correspondences between a light field camera pair. These can be applied to infer relative pose of the light field cameras and thus obtain a point cloud reconstruction of the scene. While the proposed method has interesting relationships to pose estimation for generalized cameras based on ray-to-ray correspondence, our experiments demonstrate that our approach is both more accurate and computationally more efficient. It also compares favourably to direct linear pose estimation based on aligning the 3D point clouds obtained by reconstructing depth for each individual light field. To further validate the method, we employ the pose estimates to merge light fields captured with hand-held consumer light field cameras into refocusable panoramas. | ['Bastian Goldluecke', 'Antonin Sulc', 'Ole Johannsen'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['point-cloud-reconstruction'] | ['computer-vision'] | [ 5.06331623e-01 -2.35729203e-01 2.50316888e-01 -4.92995501e-01
-3.72296870e-01 -8.38251531e-01 7.00117111e-01 -1.81743145e-01
-5.51496685e-01 5.42129457e-01 -1.39770731e-01 -2.48884223e-02
-2.16991186e-01 -6.70776725e-01 -7.22200513e-01 -4.09832358e-01
5.57135761e-01 1.01291668e+00 6.04666948e-01 2.40323618e-02
6.34313583e-01 1.02142727e+00 -1.47270417e+00 -7.44482949e-02
1.85752794e-01 5.91115713e-01 5.47835529e-01 9.66887712e-01
6.47971034e-02 4.91716117e-01 -7.35604614e-02 -3.39914471e-01
4.93839711e-01 -1.39326781e-01 -7.42815614e-01 5.53890586e-01
9.92542803e-01 -6.25754595e-01 -1.89060755e-02 1.01200604e+00
1.10048085e-01 -5.48957810e-02 3.49342942e-01 -8.60379100e-01
2.09351674e-01 -2.63943315e-01 -7.54519761e-01 7.06870034e-02
1.13559353e+00 -2.36471280e-01 7.53676891e-01 -9.83850241e-01
9.78082776e-01 9.80936468e-01 5.35914660e-01 1.49496377e-01
-1.21589816e+00 -3.42097908e-01 -2.79807210e-01 5.98816760e-02
-1.14167607e+00 -5.40111899e-01 7.96988487e-01 -6.06949925e-01
7.56976962e-01 3.51812422e-01 8.60981882e-01 4.55615550e-01
3.66288602e-01 4.95592132e-03 1.39507997e+00 -9.40881968e-01
1.19061477e-01 2.73299843e-01 1.66647300e-01 8.08169961e-01
3.56592596e-01 2.56405950e-01 -5.67968726e-01 -3.19679528e-01
1.21441019e+00 1.49387687e-01 -5.80893934e-01 -7.86590278e-01
-1.29020870e+00 4.96452004e-01 4.86470535e-02 1.24944709e-01
-1.22655287e-01 6.90757036e-02 -3.80696744e-01 -1.68942988e-01
3.38250726e-01 4.84156966e-01 -1.77645907e-01 1.52633443e-01
-5.16400576e-01 2.94696540e-01 7.68729746e-01 1.17007446e+00
1.20071328e+00 -4.37618136e-01 4.97450173e-01 2.58538365e-01
6.89285934e-01 6.81512296e-01 -3.37173939e-01 -1.32346511e+00
1.88335374e-01 5.08886874e-01 2.62776136e-01 -8.82843316e-01
-3.01604450e-01 6.31298348e-02 -1.21543156e-02 5.53484082e-01
7.32604027e-01 2.01956913e-01 -5.71400762e-01 1.10148466e+00
4.56403702e-01 2.53503859e-01 -5.94871864e-02 8.83980513e-01
3.98371071e-01 3.47144544e-01 -8.10300946e-01 -5.11411607e-01
1.39923501e+00 -2.56644815e-01 -5.37057936e-01 -7.63642192e-02
1.20694190e-01 -1.36548817e+00 4.21178997e-01 6.13262951e-01
-1.61876011e+00 -2.20499739e-01 -8.10904264e-01 -1.84966967e-01
2.03355491e-01 1.92068249e-03 3.58544081e-01 4.60847408e-01
-8.77988636e-01 2.04625383e-01 -8.72694314e-01 -6.77102029e-01
-2.57364791e-02 5.28959870e-01 -3.83101225e-01 -1.58833280e-01
-2.62909710e-01 1.00951815e+00 1.99903727e-01 -1.63093925e-01
-5.19660592e-01 -4.95522261e-01 -5.34438610e-01 -2.61726230e-01
3.07366073e-01 -9.23384249e-01 1.24376845e+00 -3.22957188e-01
-1.72354031e+00 1.18624473e+00 -2.52733201e-01 1.18373707e-01
3.32956702e-01 -1.08135447e-01 -3.53908278e-02 5.28364062e-01
-1.15199238e-01 2.65433937e-01 5.79054475e-01 -1.55318058e+00
-9.43885326e-01 -6.88706338e-01 2.84381539e-01 3.95253003e-01
2.16613084e-01 2.13643238e-01 -6.07860625e-01 2.37152502e-01
7.37848878e-01 -1.13380933e+00 -1.30110949e-01 1.53026596e-01
-3.70578736e-01 1.83989078e-01 9.73566473e-01 -2.06008807e-01
5.31753838e-01 -1.64113033e+00 3.24891806e-01 3.33412647e-01
2.67125219e-01 -2.05660686e-01 3.10532987e-01 5.14949441e-01
-1.09153800e-01 -7.20406950e-01 -5.74086234e-02 -2.07284391e-01
-5.62053382e-01 3.19496244e-01 -2.07463369e-01 9.35733378e-01
-5.24306595e-01 3.17956716e-01 -8.09780538e-01 -3.81551713e-01
6.54054224e-01 5.53193092e-01 -6.52480006e-01 3.01645070e-01
-2.38621563e-01 7.27941155e-01 -4.15046781e-01 3.75629544e-01
9.66113865e-01 2.77632270e-02 1.95057571e-01 -4.33947355e-01
-4.67170984e-01 -4.68759127e-02 -1.28143060e+00 1.81904340e+00
-4.29604709e-01 7.58680344e-01 2.66856223e-01 -3.00474226e-01
8.23704779e-01 2.78474122e-01 6.37236774e-01 -1.26279548e-01
2.81520814e-01 1.56695873e-01 -5.18425524e-01 -4.56161499e-01
5.93906581e-01 -4.51031864e-01 4.24706757e-01 5.03144503e-01
7.63891712e-02 -8.51466954e-01 1.39546573e-01 -2.17167698e-02
6.86370552e-01 4.42385375e-01 4.88996059e-01 -4.19656456e-01
7.50125110e-01 -2.13842660e-01 1.33008823e-01 1.72549203e-01
5.07622480e-01 6.25801563e-01 -3.09733599e-02 -5.96391797e-01
-1.06598878e+00 -1.18013942e+00 -5.62504232e-01 4.13115621e-01
4.39112931e-01 -3.39640886e-01 -8.02010953e-01 -1.69365197e-01
-2.34732792e-01 5.50586462e-01 -1.13129996e-01 6.48443282e-01
-6.18318498e-01 -5.61690032e-01 -1.80023715e-01 -3.83307263e-02
1.29171818e-01 -2.63537407e-01 -1.25466621e+00 -1.40299261e-01
-1.12614900e-01 -1.32031024e+00 -1.52565077e-01 -5.03520258e-02
-9.07797694e-01 -1.56073010e+00 -4.44804877e-01 -5.72406232e-01
9.32256460e-01 6.65638745e-01 1.09382176e+00 -2.10785046e-01
-1.39814392e-01 1.06783271e+00 -8.70831832e-02 -4.29169357e-01
-3.48025918e-01 -5.95317721e-01 2.70686716e-01 9.67442542e-02
3.92236680e-01 -6.85918212e-01 -6.82327271e-01 4.91250366e-01
-9.04438317e-01 1.27901673e-01 1.04113542e-01 2.02668265e-01
6.55470252e-01 -9.17013548e-03 -6.79928362e-01 -8.57632458e-01
9.99770835e-02 6.52811825e-02 -1.40061569e+00 9.28770974e-02
-3.38594764e-01 -1.80523768e-01 2.72812665e-01 1.06378205e-01
-1.26811624e+00 5.98062694e-01 1.70538336e-01 -5.67863286e-01
-3.68756056e-01 -3.00203025e-01 9.06446427e-02 -6.13254189e-01
6.41586900e-01 -1.12850711e-01 -1.87691540e-01 -6.00340247e-01
3.71930182e-01 3.79523635e-01 7.07943380e-01 -4.65571404e-01
9.29896057e-01 1.20695508e+00 5.63214004e-01 -1.02694738e+00
-6.09901190e-01 -6.88400030e-01 -1.05396533e+00 -4.61464763e-01
1.23889303e+00 -6.58541858e-01 -1.16079628e+00 2.08679259e-01
-1.59925377e+00 4.03367698e-01 -1.44411504e-01 1.10183036e+00
-8.76664817e-01 6.28703594e-01 -3.77947688e-01 -6.72222674e-01
3.00514489e-01 -1.30956209e+00 1.47996390e+00 1.87217575e-02
-2.78060362e-02 -1.18306291e+00 5.09905756e-01 5.46946228e-01
-3.23107958e-01 1.39154539e-01 6.97079301e-01 5.32878280e-01
-1.24852991e+00 -1.33679360e-01 5.49436957e-02 -3.51878777e-02
2.79467087e-02 2.71323025e-01 -1.21758103e+00 -3.07953537e-01
5.24009049e-01 2.11553231e-01 4.14370954e-01 5.86284816e-01
5.24952829e-01 3.50131318e-02 -5.77120900e-01 9.48596537e-01
2.11511159e+00 3.20239067e-01 5.99421084e-01 3.05878282e-01
7.90437818e-01 8.07895362e-01 4.90883350e-01 4.03290361e-01
2.20085278e-01 1.09581745e+00 5.91401696e-01 1.24450354e-02
-9.07826275e-02 -4.66499366e-02 1.92780063e-01 7.57443368e-01
-5.28432012e-01 -7.10853040e-02 -8.64174783e-01 9.87433363e-03
-1.34553623e+00 -8.80754650e-01 -7.85591185e-01 2.50585294e+00
3.94745797e-01 -1.42852798e-01 -4.21239585e-02 5.46626225e-02
6.07812703e-01 -2.75123984e-01 -1.56234289e-02 -2.08823532e-01
2.16650635e-01 8.86609703e-02 8.13144386e-01 1.33987498e+00
-4.15537953e-01 6.50323510e-01 7.07077503e+00 1.49133518e-01
-9.32627439e-01 -7.84754753e-04 -2.86862731e-01 -4.05009128e-02
-5.46129823e-01 5.45587063e-01 -1.02838612e+00 2.45367572e-01
1.80069193e-01 2.60379985e-02 6.20397449e-01 3.28882843e-01
4.24080640e-02 -7.92664707e-01 -1.57293093e+00 1.26078451e+00
1.81206837e-01 -1.19388413e+00 3.76284122e-04 3.62013012e-01
6.73224688e-01 6.99078068e-02 -1.94299936e-01 -8.41851175e-01
1.95231661e-01 -3.63333166e-01 5.32367885e-01 6.55530691e-01
7.19369650e-01 -3.71001273e-01 2.36004964e-01 5.85606992e-01
-1.00073481e+00 2.12978169e-01 -4.82912153e-01 -3.52889806e-01
4.86388355e-01 5.71320534e-01 -1.20258522e+00 5.57480693e-01
5.47716737e-01 5.19822717e-01 -2.72177130e-01 7.90186048e-01
1.76911559e-02 -1.50473163e-01 -4.77985919e-01 4.49905992e-01
-1.01447649e-01 -8.86426568e-01 8.23233843e-01 8.04884315e-01
4.44594622e-01 3.59843761e-01 -1.15471251e-01 9.83297706e-01
4.44153845e-01 -2.51615614e-01 -9.03180003e-01 7.12400913e-01
-4.22360487e-02 1.34518051e+00 -9.23591077e-01 -4.29508053e-02
-7.51569510e-01 7.68625021e-01 -1.84668899e-01 3.94907326e-01
-4.59050715e-01 -2.93546226e-02 3.92570823e-01 4.43931729e-01
1.77505482e-02 -3.10308307e-01 -1.36571396e-02 -1.34413600e+00
4.79411520e-02 -2.58202076e-01 -5.93920164e-02 -1.39388669e+00
-8.99779320e-01 6.09807372e-01 6.08116865e-01 -1.45070887e+00
-3.24173719e-01 -8.41344237e-01 -3.47400934e-01 9.99419272e-01
-1.67247319e+00 -9.89615679e-01 -4.08399254e-01 9.34823990e-01
3.07394236e-01 1.58917129e-01 7.12581694e-01 -4.56079543e-02
2.05749393e-01 -2.78114676e-01 1.42185837e-01 -3.42691392e-01
3.84578735e-01 -1.13379169e+00 -8.09205472e-02 9.17519033e-01
2.63759971e-01 5.74643850e-01 7.52392113e-01 -3.45497459e-01
-1.65754712e+00 -4.40792531e-01 7.54614174e-01 -1.15941405e+00
1.86350629e-01 -4.09604371e-01 -4.16751981e-01 9.90556419e-01
3.46980274e-01 -1.62120432e-01 4.50291127e-01 -1.38888672e-01
8.39022174e-02 -2.08956271e-01 -9.98091877e-01 1.34756386e-01
8.88311923e-01 -6.12461507e-01 -6.96348965e-01 7.03827679e-01
1.02426752e-01 -8.04528236e-01 -6.73144460e-01 2.26131603e-01
7.05818713e-01 -1.38408494e+00 1.09699392e+00 7.12214485e-02
7.21304938e-02 -5.20465910e-01 -2.88422346e-01 -9.67506170e-01
-2.13763312e-01 -6.43485487e-01 4.83769834e-01 8.28430653e-01
-2.71803200e-01 -6.77778363e-01 8.77802372e-01 2.31371149e-01
2.78762989e-02 -2.57892579e-01 -7.81411290e-01 -4.46991116e-01
-4.50540721e-01 -3.12947512e-01 1.41515836e-01 9.43703413e-01
-2.98336577e-02 6.25047863e-01 -7.08156973e-02 5.07412374e-01
1.08818781e+00 3.65475476e-01 8.82514060e-01 -1.62846518e+00
-4.72841859e-01 -7.20967576e-02 -4.28673446e-01 -1.28496921e+00
2.64150742e-02 -8.41044307e-01 -1.57296792e-01 -1.36608469e+00
2.77749866e-01 -3.40440452e-01 4.48497653e-01 -1.73750401e-01
3.96346152e-01 4.42561209e-01 1.63056388e-01 4.53063279e-01
-2.27399811e-01 -1.41513079e-01 1.32607090e+00 6.56659842e-01
-1.11618698e-01 2.10369557e-01 -5.27598746e-02 1.24715090e+00
1.55763000e-01 -3.44550699e-01 -4.73894447e-01 -7.88017571e-01
6.69235051e-01 5.48979461e-01 3.44494998e-01 -9.58908498e-01
3.41677606e-01 -2.37905368e-01 2.82394379e-01 -8.87600064e-01
7.39462852e-01 -1.50471711e+00 5.34667969e-01 3.28904748e-01
3.14990208e-02 1.18879683e-01 -1.66391090e-01 5.20120084e-01
1.05709054e-01 -5.29917657e-01 9.33547437e-01 -5.30058265e-01
-4.92403656e-01 1.93281069e-01 -2.31577381e-01 -3.87944281e-01
1.25578213e+00 -6.09323442e-01 -1.40750319e-01 -3.74172926e-01
-5.26026487e-01 -3.46115023e-01 1.18044448e+00 -1.78918898e-01
7.96128750e-01 -1.07328510e+00 -4.73410636e-01 6.67817771e-01
2.12755591e-01 2.91985571e-01 -2.78165072e-01 7.49383628e-01
-1.31149161e+00 7.15372443e-01 -3.32822412e-01 -1.22583222e+00
-1.64085746e+00 4.64151531e-01 4.29270059e-01 2.64902055e-01
-5.83414674e-01 6.97526097e-01 5.65893769e-01 -3.00761729e-01
-7.77363405e-02 -5.80117643e-01 -7.37600215e-03 -3.68839115e-01
3.16654116e-01 6.92722380e-01 1.10883690e-01 -9.05748844e-01
-4.55540776e-01 1.57539093e+00 2.02547520e-01 -4.34783399e-01
1.28981400e+00 -5.02455354e-01 -3.73594433e-01 2.48692989e-01
1.15606606e+00 6.13349020e-01 -1.27706230e+00 -3.24228078e-01
-4.38250661e-01 -1.10945904e+00 7.74023160e-02 -2.79779375e-01
-8.44179273e-01 8.80021214e-01 4.06587690e-01 1.09032653e-02
1.23617876e+00 1.61002800e-01 3.64382625e-01 4.25430775e-01
6.06405616e-01 -7.10191905e-01 -2.09299549e-01 2.02735037e-01
5.50867021e-01 -8.90138865e-01 4.95005369e-01 -8.46676946e-01
3.82595621e-02 1.46721709e+00 1.96121141e-01 -1.59260362e-01
5.24048507e-01 3.85080278e-01 -4.46005203e-02 -4.42365795e-01
-4.15459573e-01 2.34869756e-02 1.71702653e-01 5.88522732e-01
3.67940575e-01 -2.60250926e-01 -1.93420216e-01 -7.73913085e-01
-1.19978271e-01 1.30962104e-01 1.05390894e+00 7.42335021e-01
-5.88738501e-01 -1.28404403e+00 -8.70750606e-01 -1.00264601e-01
-1.71731681e-01 1.62267849e-01 -2.09754586e-01 6.98861599e-01
1.88169867e-01 8.42111111e-01 4.16058898e-01 3.47622000e-02
3.75333756e-01 -3.73848498e-01 1.33035612e+00 -7.30432451e-01
2.30783168e-02 4.54023570e-01 -1.55673385e-01 -6.10438049e-01
-1.06641221e+00 -7.37971723e-01 -7.90912330e-01 -1.18237056e-01
-6.19835198e-01 7.16530085e-02 9.22211766e-01 8.06112826e-01
-2.48340219e-01 -1.91460907e-01 7.72904098e-01 -1.14754605e+00
-1.62535906e-01 -4.37330782e-01 -9.23473299e-01 2.99676120e-01
5.86948991e-01 -6.04520977e-01 -4.35180992e-01 4.06857133e-01] | [9.456084251403809, -2.688197374343872] |
d0413599-0a03-4e11-8969-025aad31281d | creating-personalized-synthetic-voices-from | 2305.17436 | null | https://arxiv.org/abs/2305.17436v1 | https://arxiv.org/pdf/2305.17436v1.pdf | Creating Personalized Synthetic Voices from Post-Glossectomy Speech with Guided Diffusion Models | This paper is about developing personalized speech synthesis systems with recordings of mildly impaired speech. In particular, we consider consonant and vowel alterations resulted from partial glossectomy, the surgical removal of part of the tongue. The aim is to restore articulation in the synthesized speech and maximally preserve the target speaker's individuality. We propose to tackle the problem with guided diffusion models. Specifically, a diffusion-based speech synthesis model is trained on original recordings, to capture and preserve the target speaker's original articulation style. When using the model for inference, a separately trained phone classifier will guide the synthesis process towards proper articulation. Objective and subjective evaluation results show that the proposed method substantially improves articulation in the synthesized speech over original recordings, and preserves more of the target speaker's individuality than a voice conversion baseline. | ['Tan Lee', 'Guangyan Zhang', 'Yusheng Tian'] | 2023-05-27 | null | null | null | null | ['voice-conversion', 'voice-conversion', 'speech-synthesis'] | ['audio', 'speech', 'speech'] | [ 1.69713542e-01 7.69586682e-01 -3.86461079e-01 7.31049627e-02
-1.00701737e+00 -2.58238614e-01 2.67158598e-01 -4.26021606e-01
-9.24141854e-02 5.98852873e-01 1.20292258e+00 -1.27990171e-01
1.37970701e-01 -3.02820563e-01 -1.36272103e-01 -9.00927603e-01
4.08780873e-01 2.88167953e-01 -1.23975269e-01 -1.72397614e-01
-1.43344909e-01 4.61501330e-01 -1.64333522e+00 3.53193253e-01
1.16530466e+00 5.59584975e-01 8.06075037e-01 8.65650177e-01
-3.77598181e-02 5.17697155e-01 -9.56591606e-01 -1.92260996e-01
-1.74749121e-01 -6.75103307e-01 -6.61776125e-01 5.59323609e-01
9.10833627e-02 -4.05683100e-01 -4.39308345e-01 1.39358473e+00
7.86732733e-01 2.83552885e-01 6.69158161e-01 -6.62753582e-01
-9.06610727e-01 8.14351141e-01 4.42554832e-01 3.22835445e-02
2.85758734e-01 1.86920017e-01 8.58867288e-01 -6.88865483e-01
4.17204469e-01 1.38299239e+00 1.87414795e-01 1.23439252e+00
-1.11322927e+00 -1.68573156e-01 2.24099100e-01 -9.81758609e-02
-1.15492725e+00 -1.20472574e+00 8.98190975e-01 -2.01410636e-01
9.90834892e-01 4.65210706e-01 7.52833366e-01 1.35684240e+00
-4.75828238e-02 9.13171411e-01 6.18071258e-01 -6.32978559e-01
3.13738853e-01 1.69880807e-01 -2.28964522e-01 3.85036856e-01
-2.75906026e-01 2.90226191e-01 -4.25416738e-01 1.39407426e-01
2.58420527e-01 -4.17213678e-01 -6.81282938e-01 2.95286834e-01
-1.17524266e+00 4.70925570e-01 -1.56869754e-01 6.19683385e-01
-7.48780549e-01 -2.32095003e-01 2.97857970e-01 3.13549876e-01
7.30077565e-01 8.55102539e-02 -1.43936723e-01 -2.99795300e-01
-8.83859217e-01 -5.94216539e-03 6.89422131e-01 1.01343715e+00
-2.15313151e-01 6.54203296e-01 -4.52659369e-01 1.59268665e+00
4.73423451e-01 8.14900160e-01 1.00050330e+00 -1.08763933e+00
3.40299606e-01 2.62583464e-01 -3.04436445e-01 -2.46845961e-01
-5.51956072e-02 -7.28484750e-01 -6.66943491e-01 2.27198586e-01
4.31937650e-02 -3.77537370e-01 -1.17654550e+00 1.84589243e+00
1.01132967e-01 -3.12634021e-01 5.26365221e-01 5.48306942e-01
6.81124747e-01 8.45164776e-01 -1.33722037e-01 -8.45313966e-01
1.01749420e+00 -1.39601326e+00 -1.37371993e+00 -5.44604540e-01
4.72514004e-01 -9.76244271e-01 1.16223955e+00 3.29589337e-01
-1.52596283e+00 -5.07256806e-01 -7.54503608e-01 -2.84380605e-03
2.55139828e-01 3.75253767e-01 -3.01600784e-01 8.76149058e-01
-1.27188313e+00 3.90019804e-01 -6.66103065e-01 -2.50171065e-01
7.17697889e-02 3.20800781e-01 -1.95716619e-01 -4.78863791e-02
-9.69337583e-01 9.08200681e-01 -1.73128247e-02 -1.09068736e-01
-7.64972270e-01 -5.57627976e-01 -6.75049186e-01 1.13867342e-01
-8.27532783e-02 -9.59908426e-01 1.57467461e+00 -8.08370948e-01
-2.22468209e+00 7.42953956e-01 -4.61899042e-01 -2.10599199e-01
5.20656049e-01 2.94133592e-02 -7.89798439e-01 1.19474247e-01
-1.02895878e-01 2.58373320e-01 9.73156869e-01 -1.27666378e+00
-8.05280685e-01 -5.01759827e-01 -5.05544484e-01 5.26794791e-01
-4.41996902e-01 -1.18877612e-01 -4.64058012e-01 -9.87705410e-01
1.02926586e-02 -1.06104994e+00 1.06207781e-01 -1.59899205e-01
-7.47183025e-01 -1.38311088e-01 7.41934717e-01 -1.47600865e+00
1.33811653e+00 -2.45029831e+00 5.81279278e-01 -8.16588998e-02
-1.32647306e-01 6.39355958e-01 -9.72890407e-02 3.22024196e-01
2.89976865e-01 -1.37410775e-01 -4.95604813e-01 -9.90476429e-01
-2.97526065e-02 4.92586374e-01 -1.97125643e-01 2.92286664e-01
-3.89795274e-01 3.95750433e-01 -6.05771482e-01 -2.07982704e-01
2.21890137e-01 6.59852326e-01 -9.06872034e-01 3.82760882e-01
-3.79471719e-01 5.58688879e-01 1.02377325e-01 6.86417282e-01
2.54823297e-01 6.33439720e-01 2.52209216e-01 5.34103476e-02
-6.84836581e-02 5.57853818e-01 -6.55507147e-01 1.68237233e+00
-6.79591358e-01 4.46647972e-01 5.32398283e-01 -6.39572620e-01
9.42874670e-01 8.39675426e-01 2.55644470e-01 -5.45737207e-01
-3.10390536e-02 7.40190685e-01 5.66386223e-01 -7.83133090e-01
1.28521845e-01 -4.67996836e-01 4.05417264e-01 -2.65552104e-02
-4.63610254e-02 -5.26511610e-01 -3.14780772e-01 -3.48075360e-01
6.64959013e-01 -4.35252070e-01 2.05131203e-01 -2.34282255e-01
6.50229037e-01 -5.30456483e-01 3.10283244e-01 1.35354027e-01
-3.32401454e-01 5.42822123e-01 -1.10766135e-01 4.83291954e-01
-9.38829780e-01 -1.31661892e+00 5.94085790e-02 7.44984150e-01
-5.03180087e-01 1.24388278e-01 -1.15312123e+00 -2.97644705e-01
-1.84183493e-01 1.35538638e+00 -2.03091308e-01 -5.60438752e-01
-6.50584042e-01 -1.46490231e-01 5.12864947e-01 1.97320387e-01
3.07112843e-01 -1.16464365e+00 3.76106977e-01 3.85824710e-01
-6.33806109e-01 -8.86511385e-01 -1.03064406e+00 -1.43043563e-01
-8.24998677e-01 -3.22773367e-01 -1.23049521e+00 -1.44802463e+00
4.63804930e-01 -2.51881033e-02 2.05179751e-01 -9.15067345e-02
1.72534794e-01 1.59871474e-01 -1.82078362e-01 -2.81985372e-01
-1.31482232e+00 1.15386799e-01 4.69276607e-01 4.08172570e-02
-2.92590261e-01 -8.07495475e-01 -4.20343101e-01 4.52582464e-02
-6.37683451e-01 -3.10126245e-01 3.83724838e-01 7.99265623e-01
5.70460558e-01 2.38822743e-01 9.06037509e-01 -3.18705678e-01
1.19796455e+00 -2.73608238e-01 4.51420136e-02 2.61405502e-02
-6.49038076e-01 -1.09548695e-01 7.08584964e-01 -6.15518868e-01
-1.37275076e+00 -2.22019345e-01 -8.79654229e-01 -4.56952214e-01
-7.09699914e-02 2.02310219e-01 -7.25714803e-01 4.93141592e-01
5.20416737e-01 6.83576465e-01 3.92120183e-01 -7.78622568e-01
4.18407470e-01 1.42393684e+00 9.18500841e-01 -2.49895200e-01
2.78623581e-01 1.25645772e-01 -7.98309147e-01 -1.31477892e+00
-2.98899978e-01 -2.59317875e-01 -3.60733211e-01 -1.00394867e-01
6.31490409e-01 -5.85792601e-01 -4.73052919e-01 7.73759067e-01
-1.21782827e+00 -4.85692084e-01 -7.00231493e-01 8.75761926e-01
-9.42750156e-01 1.75596818e-01 -4.38416094e-01 -1.06502748e+00
-5.63269556e-01 -1.52923167e+00 9.32927489e-01 7.51332492e-02
-4.51014251e-01 -1.16377378e+00 1.93991452e-01 6.16505623e-01
3.83197367e-01 -4.48492438e-01 1.24172306e+00 -6.03416085e-01
-7.29146898e-02 -1.45550385e-01 6.24761343e-01 1.22748125e+00
1.02102685e+00 -9.07748938e-02 -1.11614370e+00 -1.83713257e-01
4.16141391e-01 4.53506470e-01 4.97035354e-01 6.73146665e-01
6.78891182e-01 -8.07239294e-01 -1.51405528e-01 4.80853528e-01
5.27784765e-01 8.04283738e-01 7.12501526e-01 -2.41705328e-01
3.07801515e-01 8.05213094e-01 4.42996621e-02 9.74924937e-02
2.23316342e-01 7.89270699e-01 7.95257092e-02 1.64627016e-01
-1.44170702e+00 -4.09587443e-01 6.97995484e-01 1.68537188e+00
3.41663603e-03 -6.68540537e-01 -4.52936858e-01 1.03349543e+00
-1.26668179e+00 -9.18254793e-01 1.66982681e-01 2.09463143e+00
9.97759044e-01 -7.02530146e-02 1.55762076e-01 4.17228192e-01
8.99537623e-01 1.99007779e-01 -5.21876812e-01 -4.86833334e-01
-1.35937855e-01 5.56153879e-02 4.82161976e-02 1.22869813e+00
-1.80796057e-01 1.03853416e+00 6.32992887e+00 9.59127784e-01
-1.19058645e+00 9.82503369e-02 1.26276771e-02 -3.76283735e-01
-4.72402781e-01 -4.93545860e-01 -5.26744366e-01 5.23273110e-01
1.16258729e+00 -4.67799872e-01 9.28773880e-01 5.85953712e-01
6.90700948e-01 4.47708130e-01 -8.75491321e-01 7.12865829e-01
1.17341392e-01 -1.05094254e+00 8.35240707e-02 2.77674645e-01
6.61715031e-01 -1.56305566e-01 3.91959846e-01 7.39977732e-02
-5.54555375e-03 -6.52439356e-01 9.51855421e-01 7.68406689e-01
9.43887770e-01 -7.34071553e-01 2.48736247e-01 4.75814879e-01
-6.91533327e-01 -1.89165324e-01 6.88607618e-02 3.68487149e-01
3.03264558e-01 4.69254106e-01 -1.17644167e+00 6.23304397e-02
7.07639456e-02 1.78136736e-01 2.73591071e-01 9.71397638e-01
-6.30052626e-01 7.81138897e-01 7.19416216e-02 4.86452132e-02
-1.76365361e-01 1.09927244e-01 1.20066178e+00 9.74699795e-01
7.17826545e-01 -3.77984978e-02 -8.09645057e-02 5.55812359e-01
-2.38936052e-01 3.43595803e-01 -6.64880574e-01 -1.53469250e-01
6.82950139e-01 3.26288164e-01 -6.70597702e-03 -2.88243115e-01
-1.23730324e-01 1.39301658e+00 -3.60405743e-02 4.94908780e-01
-1.44208074e-01 -2.72071600e-01 1.16250026e+00 1.73987541e-02
3.50346789e-02 -9.62495059e-02 -3.22076589e-01 -8.19817185e-01
1.49889782e-01 -1.24198985e+00 -1.56243458e-01 -7.70326197e-01
-7.55444705e-01 8.47226441e-01 -4.04027134e-01 -8.26524615e-01
-6.04251146e-01 -3.14198077e-01 -6.33131325e-01 1.19696355e+00
-1.14804208e+00 -1.00235617e+00 3.04270595e-01 6.53135777e-01
1.24291599e+00 -5.05911946e-01 1.04895055e+00 4.12191391e-01
-5.83426058e-01 7.30319440e-01 3.54246587e-01 -1.94957420e-01
4.49268222e-01 -9.91521776e-01 2.80960977e-01 9.56189752e-01
-8.17688853e-02 6.32512391e-01 9.89568591e-01 -8.61022472e-01
-9.00704443e-01 -1.06441593e+00 1.41739082e+00 1.32864252e-01
3.10923368e-01 4.55777645e-02 -8.72499764e-01 5.73854148e-01
3.15924793e-01 -4.82964277e-01 5.37789345e-01 -3.94802958e-01
2.27172941e-01 -1.25300869e-01 -1.42636728e+00 9.06915486e-01
1.27475369e+00 -9.45626557e-01 -7.57908881e-01 6.08405352e-01
1.16931033e+00 -2.81994969e-01 -8.04375947e-01 2.17731693e-04
3.93629134e-01 -3.65324706e-01 9.40031469e-01 -2.75904030e-01
1.39814299e-02 3.72254960e-02 -6.40099168e-01 -2.17956591e+00
-1.27906978e-01 -9.59514558e-01 -8.87457654e-02 1.37993574e+00
5.19707978e-01 -6.89417660e-01 4.96898592e-01 3.43091697e-01
-1.03423595e+00 -1.59857810e-01 -1.11623478e+00 -9.68231201e-01
3.25190514e-01 -3.34563971e-01 4.42088097e-01 5.03414631e-01
3.16345751e-01 1.54073775e-01 -4.96358842e-01 5.19518554e-01
4.29112881e-01 -4.42868143e-01 3.11955094e-01 -8.49974453e-01
-2.38826394e-01 -6.04309320e-01 -6.92753447e-03 -8.94990504e-01
4.18039262e-01 -1.12834251e+00 3.33700001e-01 -1.96240425e+00
-6.19974375e-01 9.54918712e-02 2.94814616e-01 2.63945282e-01
7.41549581e-03 -3.43257755e-01 1.96869820e-01 2.39393741e-01
7.29972482e-01 9.55043435e-01 1.80808854e+00 -4.00417149e-01
-7.01295733e-01 8.04107189e-01 -7.87282825e-01 7.14855790e-01
8.22427511e-01 -5.31360686e-01 -8.74010384e-01 -5.24374247e-01
-8.22686434e-01 6.05666399e-01 -1.31579369e-01 -8.22530925e-01
3.63203257e-01 -9.86120328e-02 -5.00391781e-01 -1.53220460e-01
6.72589600e-01 -6.02234542e-01 1.78005639e-02 8.90558183e-01
-6.37764752e-01 -4.55979437e-01 1.42279699e-01 2.99473703e-01
-2.67586648e-01 -3.61798912e-01 1.19211018e+00 6.30036145e-02
-2.23303432e-04 5.43906875e-02 -7.90185392e-01 -1.44538835e-01
6.40784144e-01 9.79165658e-02 -2.74632037e-01 -4.72394109e-01
-1.52758074e+00 -4.05257672e-01 1.26158684e-01 6.17365181e-01
6.91225111e-01 -1.34685087e+00 -8.92363727e-01 3.90913993e-01
-1.26063749e-01 -3.22360605e-01 2.46941775e-01 7.71052659e-01
-2.42635623e-01 4.53685492e-01 1.22603312e-01 -1.28865987e-01
-1.35789216e+00 4.67562288e-01 7.25947022e-01 2.60925263e-01
-8.95337045e-01 8.75019729e-01 9.33566689e-02 -4.42541033e-01
5.41621685e-01 -7.39596605e-01 1.52430506e-02 -5.45275509e-02
5.77243686e-01 6.06891632e-01 3.53358120e-01 -6.64261162e-01
-1.47072658e-01 2.11169243e-01 -7.66049838e-03 -6.88695610e-01
1.06232834e+00 -6.49266899e-01 1.65265173e-01 2.91572243e-01
1.26796913e+00 7.57954776e-01 -9.67498779e-01 -3.78189832e-01
-3.94317806e-01 -2.61567622e-01 4.67708141e-01 -1.18980634e+00
-1.13239145e+00 8.86858463e-01 6.27652645e-01 1.53600767e-01
1.09890139e+00 -5.59835322e-02 1.01377940e+00 3.49553600e-02
9.58246365e-03 -1.16379964e+00 2.12728325e-02 3.13587636e-01
1.43561244e+00 -4.63002652e-01 -8.53937566e-01 -4.13805753e-01
-7.03998804e-01 9.03388143e-01 6.46272376e-02 2.00428456e-01
6.69920087e-01 1.47119135e-01 3.12109351e-01 3.15518409e-01
-5.31004369e-01 -3.78351957e-01 4.12753642e-01 1.13349843e+00
2.60328025e-01 4.18118387e-01 -2.71185666e-01 5.89611590e-01
-8.49717975e-01 -2.73389220e-01 4.22259778e-01 3.46976340e-01
-8.44588757e-01 -1.18838251e+00 -4.61083353e-01 4.93686199e-02
-3.66052926e-01 -2.74056226e-01 -4.07765210e-01 3.23535949e-01
1.13619596e-01 1.53476310e+00 -1.42363027e-01 -2.77412504e-01
8.28913033e-01 5.64196289e-01 3.03434521e-01 -8.10281038e-01
-2.76434988e-01 6.10426247e-01 4.39861178e-01 -6.91912025e-02
-4.33735326e-02 -9.46638763e-01 -1.42169023e+00 -1.42454982e-01
3.34707997e-03 2.83588618e-02 9.96163130e-01 1.15709043e+00
1.95560735e-02 1.12991607e+00 6.79772973e-01 -2.78978527e-01
-7.89427698e-01 -1.29618120e+00 -8.39543104e-01 2.15212516e-02
1.03368247e+00 -2.10763022e-01 -6.92141354e-01 1.62854806e-01] | [14.867825508117676, 6.484362602233887] |
84f671bf-d3eb-4fa8-a1d2-102432338b96 | achieving-fairness-in-multi-agent-markov | 2306.00324 | null | https://arxiv.org/abs/2306.00324v1 | https://arxiv.org/pdf/2306.00324v1.pdf | Achieving Fairness in Multi-Agent Markov Decision Processes Using Reinforcement Learning | Fairness plays a crucial role in various multi-agent systems (e.g., communication networks, financial markets, etc.). Many multi-agent dynamical interactions can be cast as Markov Decision Processes (MDPs). While existing research has focused on studying fairness in known environments, the exploration of fairness in such systems for unknown environments remains open. In this paper, we propose a Reinforcement Learning (RL) approach to achieve fairness in multi-agent finite-horizon episodic MDPs. Instead of maximizing the sum of individual agents' value functions, we introduce a fairness function that ensures equitable rewards across agents. Since the classical Bellman's equation does not hold when the sum of individual value functions is not maximized, we cannot use traditional approaches. Instead, in order to explore, we maintain a confidence bound of the unknown environment and then propose an online convex optimization based approach to obtain a policy constrained to this confidence region. We show that such an approach achieves sub-linear regret in terms of the number of episodes. Additionally, we provide a probably approximately correct (PAC) guarantee based on the obtained regret bound. We also propose an offline RL algorithm and bound the optimality gap with respect to the optimal fair solution. To mitigate computational complexity, we introduce a policy-gradient type method for the fair objective. Simulation experiments also demonstrate the efficacy of our approach. | ['Ness B. Shroff', 'Arnob Ghosh', 'Peizhong Ju'] | 2023-06-01 | null | null | null | null | ['offline-rl'] | ['playing-games'] | [-3.86738628e-01 1.87799647e-01 -1.35278970e-01 -2.23983437e-01
-5.27474582e-01 -5.40161848e-01 2.14275926e-01 3.59787047e-01
-8.43842089e-01 1.42344630e+00 -2.27999404e-01 -2.95861751e-01
-5.52037895e-01 -8.40352356e-01 -5.43832779e-01 -8.21020126e-01
-2.55576730e-01 6.13001168e-01 -1.39272541e-01 -1.31943181e-01
3.96761447e-01 3.04441333e-01 -1.00719583e+00 -5.45052826e-01
1.18324256e+00 1.05662704e+00 1.38427168e-01 7.59223521e-01
7.63846040e-02 9.45831776e-01 -7.10274220e-01 -2.26363659e-01
5.35621822e-01 -6.30339146e-01 -7.35622287e-01 1.84097797e-01
-4.99382675e-01 -8.54158819e-01 -9.36651826e-02 1.36472011e+00
5.08392870e-01 6.43561840e-01 4.45036232e-01 -1.66999435e+00
-3.51107627e-01 6.64024591e-01 -7.28046179e-01 5.07307425e-02
9.18696746e-02 4.85200286e-02 8.60255957e-01 9.94959194e-03
3.74764472e-01 1.29483962e+00 2.70539671e-01 7.28904843e-01
-9.01629150e-01 -4.70994562e-01 5.53989530e-01 3.26972567e-02
-9.70730424e-01 -1.87700585e-01 2.07714871e-01 2.11603530e-02
5.30416191e-01 3.28493506e-01 6.86609745e-01 3.37955743e-01
2.87321389e-01 8.29433918e-01 1.17648935e+00 -3.42703134e-01
7.70805299e-01 9.48041826e-02 -1.50155723e-01 4.46795791e-01
4.85721022e-01 1.92293435e-01 -5.42190745e-02 -3.34669679e-01
7.64383852e-01 4.60969210e-02 -1.30291834e-01 -3.08439821e-01
-7.59259522e-01 9.86883700e-01 2.40647092e-01 -1.38509512e-01
-6.25794649e-01 2.33269513e-01 2.09401667e-01 5.26242852e-01
3.79027158e-01 3.93845767e-01 -1.79982498e-01 -2.52834022e-01
-6.80678129e-01 5.41517854e-01 1.38959837e+00 7.44456649e-01
3.49160343e-01 3.70029919e-03 -3.12058419e-01 4.65540856e-01
2.51029313e-01 4.35537755e-01 -1.98859554e-02 -1.40024018e+00
4.88455176e-01 4.84372117e-03 9.37882543e-01 -7.95944750e-01
-1.99468136e-01 -3.11675251e-01 -7.78479993e-01 5.54083169e-01
7.47828841e-01 -7.48184025e-01 -2.89924920e-01 1.82694173e+00
5.45159996e-01 -6.87050596e-02 2.68511713e-01 1.15412986e+00
-3.28186512e-01 7.69842744e-01 -1.08617768e-01 -1.04806483e+00
8.75530541e-01 -9.28051591e-01 -7.64628053e-01 1.11918204e-01
7.78376833e-02 -4.20507431e-01 4.85911489e-01 3.06554228e-01
-1.28996587e+00 3.30858082e-01 -7.36728311e-01 5.75058758e-01
1.41123563e-01 -5.29065967e-01 4.16788399e-01 7.70234823e-01
-7.77149796e-01 7.50654757e-01 -8.88693810e-01 -2.16931850e-01
1.80908099e-01 4.55851376e-01 2.49104336e-01 1.97719142e-01
-9.32940006e-01 8.57396007e-01 2.95176685e-01 2.38908485e-01
-1.09052718e+00 -4.50358003e-01 -4.03367847e-01 2.38070950e-01
1.07659173e+00 -6.80892229e-01 1.56426191e+00 -1.05030060e+00
-1.94193816e+00 1.76576391e-01 1.39165357e-01 -6.32361650e-01
1.03265345e+00 -1.14267714e-01 1.40370905e-01 2.49459460e-01
3.32301445e-02 -7.64904767e-02 6.11558378e-01 -1.24416900e+00
-9.17802691e-01 -1.28442645e-01 7.90247798e-01 6.76430106e-01
-2.55821317e-01 8.60627592e-02 2.91320890e-01 -2.03097165e-01
-3.08012992e-01 -9.23857510e-01 -8.60042870e-01 9.38151553e-02
-3.37153345e-01 2.09370982e-02 2.10007161e-01 -3.74801964e-01
9.63239372e-01 -1.72238016e+00 -6.43979236e-02 3.20445955e-01
1.38851404e-01 -1.98213523e-03 8.74616355e-02 5.63462675e-01
6.30060971e-01 1.66622013e-01 -2.65252233e-01 -3.31333548e-01
5.59818685e-01 2.90111244e-01 -2.42054582e-01 6.75444126e-01
-3.15099925e-01 4.59961653e-01 -9.81529236e-01 -2.92201489e-01
1.09436847e-01 -1.94869950e-01 -6.93024039e-01 4.68290955e-01
-3.54706645e-01 3.64366978e-01 -8.61602783e-01 3.21118593e-01
7.71041691e-01 2.21922677e-02 4.67266858e-01 4.48345721e-01
-4.38374221e-01 -3.08236837e-01 -1.48539090e+00 9.85219121e-01
-5.02223194e-01 -3.05873118e-02 5.81617892e-01 -1.17574775e+00
5.32031357e-01 2.52240777e-01 6.93130255e-01 -5.82502127e-01
3.26585382e-01 2.67397493e-01 5.37343435e-02 -2.48910651e-01
6.51529193e-01 -5.60667813e-01 -1.40303999e-01 8.91135275e-01
-5.67742109e-01 1.40074208e-01 2.34829769e-01 1.41128197e-01
8.54818463e-01 -1.06432088e-01 6.17304087e-01 -3.92400831e-01
3.77811402e-01 -1.26149222e-01 7.74207234e-01 1.02033186e+00
-6.89261258e-01 -1.45883616e-02 8.92765701e-01 -1.72624335e-01
-1.02451718e+00 -7.14579701e-01 1.94995135e-01 8.82717788e-01
6.10108435e-01 1.33728951e-01 -7.55566597e-01 -4.74252999e-01
1.79936767e-01 6.75938547e-01 -5.82544982e-01 1.10685632e-01
-2.72099763e-01 -1.02119517e+00 6.09166212e-02 -2.11694743e-02
6.89064622e-01 -8.72294843e-01 -1.08065498e+00 4.83538270e-01
2.67730299e-02 -8.57506335e-01 -5.96042693e-01 -8.33042711e-02
-4.59796876e-01 -9.78149652e-01 -6.93764925e-01 -1.91356316e-01
6.11051798e-01 9.42370072e-02 6.78400278e-01 -4.07356098e-02
1.25071287e-01 5.72786033e-01 -1.47839859e-01 -4.35824215e-01
-2.65975714e-01 -2.25244954e-01 2.24205360e-01 6.46256283e-02
-2.52532482e-01 -3.74555320e-01 -7.93062091e-01 1.42966807e-01
-8.34359288e-01 -1.90541044e-01 2.82796711e-01 8.24597478e-01
5.69446564e-01 3.79657209e-01 9.72883165e-01 -6.27513587e-01
1.09008515e+00 -5.69088995e-01 -1.12632692e+00 4.82479900e-01
-6.12986386e-01 2.30685458e-01 8.80854428e-01 -3.85091156e-01
-1.22561252e+00 -2.37154692e-01 3.11505139e-01 -1.94225550e-01
2.23280475e-01 4.05903459e-01 -1.01420067e-01 -1.32902354e-01
-2.47331546e-03 3.62632684e-02 2.24647745e-01 6.42704545e-03
2.20942125e-01 5.48910797e-01 1.26947895e-01 -1.07114327e+00
5.05831599e-01 3.00441921e-01 2.44655088e-01 -3.78206074e-01
-8.50247800e-01 -5.79595054e-03 1.24066502e-01 -4.91257399e-01
3.97253782e-01 -5.04565775e-01 -1.75561070e+00 2.63312072e-01
-8.86127174e-01 -3.48656297e-01 -2.92735904e-01 5.93998492e-01
-9.95382786e-01 4.11904573e-01 -6.90123975e-01 -1.84919035e+00
-3.25732082e-01 -8.15331995e-01 3.65359426e-01 6.31974101e-01
2.07495734e-01 -9.27873671e-01 2.78943896e-01 -9.65299532e-02
5.12928724e-01 3.93297791e-01 3.07536423e-01 -4.30016547e-01
-6.90466464e-01 7.28096887e-02 -1.60865024e-01 -3.13299634e-02
4.69221473e-02 -1.90524340e-01 -3.31431925e-01 -5.59284449e-01
2.36902207e-01 -4.25501943e-01 4.12460476e-01 5.03097832e-01
9.42959428e-01 -1.01364493e+00 -5.46905659e-02 2.04788953e-01
1.67639387e+00 5.39005995e-01 2.32986122e-01 6.65281057e-01
1.47397891e-01 6.67114794e-01 9.64475214e-01 1.38607514e+00
6.26265764e-01 3.69112700e-01 7.55372524e-01 4.00129765e-01
8.55773926e-01 4.89942394e-02 4.20416087e-01 2.51527667e-01
-3.23809832e-01 -3.45246315e-01 -5.29121876e-01 3.80699039e-01
-2.29409814e+00 -1.25842357e+00 3.66309583e-01 2.61762929e+00
8.70231986e-01 -5.36397249e-02 4.68352765e-01 -1.65926144e-01
8.45539272e-01 -2.19742626e-01 -1.00722456e+00 -7.52772987e-01
-2.82025244e-02 -1.83511868e-01 7.64988244e-01 8.07746291e-01
-8.41440797e-01 6.50348663e-01 5.55199814e+00 6.98821247e-01
-8.37590694e-01 4.76360135e-02 8.25490355e-01 -4.54410166e-01
-1.75857857e-01 4.53903824e-02 -4.67361569e-01 5.74338675e-01
8.70078504e-01 -9.73523021e-01 1.00453699e+00 7.55028188e-01
8.09188426e-01 -5.19573689e-01 -7.65632033e-01 6.90657556e-01
-5.05034387e-01 -8.51856291e-01 -6.21978760e-01 3.26577693e-01
8.97473156e-01 -6.00639820e-01 -1.07986383e-01 2.75118649e-01
7.33726501e-01 -7.97469854e-01 8.80940676e-01 6.02438986e-01
2.41547018e-01 -1.38789022e+00 7.23143637e-01 6.35883033e-01
-1.03849590e+00 -3.39444607e-01 -4.97378409e-01 -5.03063679e-01
4.70748842e-01 5.63395202e-01 -3.96332622e-01 7.16578305e-01
3.19746584e-01 1.74897000e-01 3.74148458e-01 1.33441269e+00
3.51521112e-02 8.61935988e-02 -4.77347136e-01 -3.88438284e-01
4.27210391e-01 -6.04333460e-01 5.93240082e-01 6.39807224e-01
2.92845607e-01 4.47225451e-01 6.13881767e-01 7.87056804e-01
-2.56230738e-02 2.18649656e-01 -1.30077779e-01 -1.63940296e-01
5.55152059e-01 1.17177308e+00 -9.39785123e-01 -3.59321088e-01
-3.79757360e-02 1.10792208e+00 4.49186891e-01 3.83928359e-01
-9.60461497e-01 -2.81933576e-01 9.28770900e-01 -5.60855389e-01
1.00678258e-01 -6.39014766e-02 -1.91827953e-01 -1.08824289e+00
1.39743611e-01 -7.31427431e-01 4.44318384e-01 -4.64530028e-02
-1.38622200e+00 3.73652041e-01 -2.24372685e-01 -1.04549193e+00
-4.50322151e-01 1.09861763e-02 -6.76802993e-01 6.93636060e-01
-1.68110394e+00 -5.00711262e-01 2.12765589e-01 5.45478880e-01
1.55921355e-01 1.78557739e-01 5.43671668e-01 1.25724316e-01
-8.08604598e-01 3.61859798e-01 7.36167908e-01 -3.69291604e-01
4.30946320e-01 -1.39540589e+00 -3.95229340e-01 7.59266973e-01
-6.98797107e-01 2.70871103e-01 1.10223305e+00 -5.91123104e-01
-1.38935959e+00 -9.20984983e-01 3.87642086e-01 2.16345355e-01
8.42155635e-01 1.49324298e-01 -5.48846900e-01 3.60751987e-01
2.38397375e-01 -6.33783191e-02 3.11435461e-01 -1.58681676e-01
3.35181952e-01 -1.67781591e-01 -1.43280816e+00 6.90903604e-01
7.59381235e-01 2.98553929e-02 -8.84878561e-02 2.68551081e-01
5.87799191e-01 -4.36054558e-01 -7.60723829e-01 4.28235754e-02
6.83994830e-01 -7.77987897e-01 4.72482860e-01 -8.36829126e-01
3.41858357e-01 -2.00283065e-01 -1.74399212e-01 -1.61561024e+00
3.02091949e-02 -9.31664407e-01 -2.14689761e-01 1.05156684e+00
1.03926435e-01 -8.74258757e-01 7.92671084e-01 1.15679526e+00
1.28458470e-01 -7.67963231e-01 -1.01473927e+00 -1.13776708e+00
2.45129928e-01 -5.70640294e-03 6.20760202e-01 8.42083871e-01
3.25815976e-01 -2.30721205e-01 -8.55248690e-01 3.46912742e-01
1.11130154e+00 4.94029969e-01 4.69562024e-01 -6.87876523e-01
-6.39013946e-01 -5.73016882e-01 2.78672725e-01 -8.39012206e-01
2.90641248e-01 -2.73631066e-01 3.44512910e-01 -1.49896348e+00
4.63937908e-01 -7.54593194e-01 -2.87074506e-01 1.09281339e-01
-1.86835140e-01 -4.54054087e-01 6.73044264e-01 9.24802348e-02
-1.13904214e+00 9.30687308e-01 1.35998142e+00 7.24064857e-02
-3.13316584e-01 3.01519662e-01 -6.14046693e-01 3.48645270e-01
1.08790028e+00 -5.55125475e-01 -3.21659416e-01 -1.45512372e-01
1.65404268e-02 7.58873820e-01 3.73711623e-02 -5.31421900e-01
3.19841385e-01 -9.99588728e-01 -2.29698062e-01 -1.40369251e-01
1.36215255e-01 -7.77472556e-01 2.17502534e-01 8.74383986e-01
-4.34807807e-01 -1.07603200e-01 -2.20570609e-01 7.78477907e-01
3.56361605e-02 -6.12454236e-01 8.13597023e-01 -2.00668871e-01
-2.28497371e-01 5.00644207e-01 -5.23627520e-01 1.09627314e-01
1.36113524e+00 2.40749434e-01 -3.49457026e-01 -8.39902401e-01
-6.22411430e-01 1.03392291e+00 3.75831008e-01 -1.39693514e-01
3.61378193e-01 -1.02759600e+00 -5.78379452e-01 -5.54556549e-01
-5.29287398e-01 -4.00165886e-01 4.56719488e-01 7.73482919e-01
-3.47573519e-01 3.44009280e-01 -3.83648276e-01 1.48593992e-01
-9.23666954e-01 5.08911729e-01 7.03890026e-01 -4.97413486e-01
-1.11909412e-01 3.23811024e-01 7.21332133e-02 -1.10840559e-01
1.87427655e-01 2.73288153e-02 6.34925365e-02 -3.15416940e-02
6.92322433e-01 7.05976248e-01 -4.60282683e-01 -1.21315464e-01
-2.26387560e-01 7.03717917e-02 4.18974794e-02 -5.76727569e-01
1.38420689e+00 -5.73149920e-01 -1.42011002e-01 1.76280290e-01
6.57696962e-01 -1.47742763e-01 -1.61792493e+00 -1.06826201e-01
-3.59760262e-02 -6.58158660e-01 -4.10913974e-02 -6.65522754e-01
-1.04065216e+00 4.10211861e-01 3.31315607e-01 7.03257442e-01
1.12803555e+00 -4.98678237e-01 5.64476967e-01 4.26829636e-01
7.49283910e-01 -1.36874568e+00 -2.18514293e-01 4.48356807e-01
6.41919732e-01 -1.18050063e+00 -6.14594817e-02 -1.35804176e-01
-8.15966249e-01 1.17301619e+00 5.56456745e-01 -2.22909003e-01
4.26927358e-01 2.39433542e-01 -3.32500905e-01 3.26939791e-01
-9.17154908e-01 -4.37657177e-01 -4.74866420e-01 2.23424658e-01
1.21677751e-02 5.42050004e-01 -8.12689841e-01 6.54962420e-01
-7.43240491e-02 5.01743369e-02 9.47843194e-01 1.15058625e+00
-7.27693796e-01 -1.07177663e+00 -4.80912417e-01 4.54113126e-01
-6.61494970e-01 2.08701521e-01 -3.10817035e-04 4.36467886e-01
-3.65338027e-01 1.19712472e+00 -3.95663083e-02 3.17759454e-01
2.45385557e-01 -4.52238262e-01 5.56248963e-01 -2.39454776e-01
-4.07404512e-01 1.56332344e-01 1.83152348e-01 -4.67330039e-01
-4.98976588e-01 -6.78375781e-01 -1.23869359e+00 -1.06410503e+00
-2.16221064e-01 5.23793876e-01 4.25241560e-01 1.14201331e+00
-1.61466151e-01 4.26974386e-01 9.96534526e-01 -5.25561571e-01
-1.16481221e+00 -5.36344826e-01 -9.33343947e-01 -5.65603077e-02
3.98575097e-01 -5.07604718e-01 -4.26733077e-01 -5.27605295e-01] | [4.2667155265808105, 2.702383041381836] |
71d1eac5-42c6-4cfc-968d-99851702e0c4 | single-stage-broad-multi-instance-multi-label | 2209.02625 | null | https://arxiv.org/abs/2209.02625v2 | https://arxiv.org/pdf/2209.02625v2.pdf | Single-Stage Broad Multi-Instance Multi-Label Learning (BMIML) with Diverse Inter-Correlations and its application to medical image classification | described by multiple instances (e.g., image patches) and simultaneously associated with multiple labels. Existing MIML methods are useful in many applications but most of which suffer from relatively low accuracy and training efficiency due to several issues: i) the inter-label correlations(i.e., the probabilistic correlations between the multiple labels corresponding to an object) are neglected; ii) the inter-instance correlations (i.e., the probabilistic correlations of different instances in predicting the object label) cannot be learned directly (or jointly) with other types of correlations due to the missing instance labels; iii) diverse inter-correlations (e.g., inter-label correlations, inter-instance correlations) can only be learned in multiple stages. To resolve these issues, a new single-stage framework called broad multi-instance multi-label learning (BMIML) is proposed. In BMIML, there are three innovative modules: i) an auto-weighted label enhancement learning (AWLEL) based on broad learning system (BLS) is designed, which simultaneously and efficiently captures the inter-label correlations while traditional BLS cannot; ii) A specific MIML neural network called scalable multi-instance probabilistic regression (SMIPR) is constructed to effectively estimate the inter-instance correlations using the object label only, which can provide additional probabilistic information for learning; iii) Finally, an interactive decision optimization (IDO) is designed to combine and optimize the results from AWLEL and SMIPR and form a single-stage framework. Experiments show that BMIML is highly competitive to (or even better than) existing methods in accuracy and much faster than most MIML methods even for large medical image data sets (> 90K images). | ['DeShuang Huang', 'Chi-Man Vong', 'Yanfen Gan', 'Jianhang Zhou', 'Qi Lai'] | 2022-09-06 | null | null | null | null | ['multi-label-learning'] | ['methodology'] | [ 2.62826055e-01 -2.04940096e-01 -5.21076500e-01 -5.11131883e-01
-1.14672625e+00 2.85091288e-02 2.73444384e-01 1.58887282e-01
-2.56867200e-01 6.19128585e-01 -2.22736835e-01 2.70393968e-01
-5.38069785e-01 -4.54936862e-01 -4.77630168e-01 -1.09525251e+00
9.59490091e-02 7.89621592e-01 2.79747546e-01 4.21288788e-01
1.22601958e-02 1.88611060e-01 -1.64474022e+00 3.79698128e-01
8.00297320e-01 1.21751916e+00 3.24771762e-01 1.73187688e-01
-4.47039634e-01 9.12010312e-01 -4.51472938e-01 -5.37769087e-02
-2.07295626e-01 -2.83885032e-01 -7.56373644e-01 3.54265302e-01
2.80458689e-01 6.41022399e-02 1.98173821e-01 1.09701085e+00
4.28799450e-01 -1.35183185e-01 8.52709234e-01 -1.35571170e+00
-3.59332830e-01 3.92376453e-01 -1.08030105e+00 -2.78306812e-01
-1.58945367e-01 -2.66117066e-01 1.04343784e+00 -1.04698086e+00
3.54693443e-01 1.32283390e+00 7.18572557e-01 2.48157278e-01
-1.25592172e+00 -8.89507711e-01 3.88689309e-01 2.89192677e-01
-1.64588416e+00 9.65033844e-02 8.79009306e-01 -3.98928821e-01
3.86701405e-01 1.70267612e-01 3.27079326e-01 9.04682815e-01
1.50674090e-01 1.15663970e+00 1.43554986e+00 -2.61306703e-01
4.06570286e-02 4.56610531e-01 3.79444718e-01 7.65380859e-01
-1.37738541e-01 -3.09806764e-02 -4.35616791e-01 -3.65329891e-01
4.99964744e-01 2.63359994e-01 -3.48640159e-02 -2.12523043e-01
-1.19855106e+00 7.25186527e-01 4.02912557e-01 3.26212347e-01
-2.52222627e-01 -2.05364451e-02 3.90024424e-01 -4.24699634e-02
6.11858606e-01 1.81260213e-01 -7.76748300e-01 3.93586814e-01
-8.61080647e-01 8.71849582e-02 5.37699699e-01 9.46280837e-01
1.19717777e+00 -4.55211192e-01 -3.51245791e-01 1.35955536e+00
4.91470933e-01 4.14233714e-01 4.79612917e-01 -6.90449417e-01
3.23092043e-01 6.47923291e-01 -1.04627140e-01 -9.46755826e-01
-8.07111323e-01 -6.40413582e-01 -1.04275179e+00 -4.04841378e-02
1.60460949e-01 -2.47443072e-03 -7.17944920e-01 1.68957376e+00
5.60486376e-01 5.80625176e-01 -2.05515444e-01 5.99038184e-01
1.21019077e+00 8.02937686e-01 4.58838582e-01 -4.83556688e-01
1.57227969e+00 -1.13497210e+00 -8.26617062e-01 -3.73258233e-01
6.07945621e-01 -7.73271680e-01 9.20008361e-01 3.16645384e-01
-6.87452853e-01 -5.56332886e-01 -8.21476817e-01 1.80377930e-01
-2.51209855e-01 4.65021014e-01 6.24197841e-01 1.44886479e-01
-5.55984318e-01 2.72495508e-01 -5.54379344e-01 1.28041491e-01
5.01196921e-01 5.06845295e-01 -2.13163584e-01 -4.88737017e-01
-1.15009749e+00 6.65886581e-01 5.31682193e-01 5.93228787e-02
-6.97937548e-01 -7.15417027e-01 -7.45341778e-01 -1.40880302e-01
7.62245893e-01 -3.72879386e-01 8.24634075e-01 -1.06873000e+00
-1.15766692e+00 7.89837062e-01 -2.63139606e-01 4.15693879e-01
1.50829330e-01 4.26978357e-02 -3.80060434e-01 6.46492699e-03
2.49723271e-01 7.96244085e-01 6.01325095e-01 -1.64409471e+00
-8.38386416e-01 -3.69354844e-01 -2.68801123e-01 3.21050316e-01
-3.19245249e-01 2.30419397e-01 -6.24775052e-01 -7.91954160e-01
3.76498163e-01 -1.06818116e+00 -1.28641531e-01 -9.32921469e-02
-5.68269551e-01 -5.67324221e-01 9.57640886e-01 -5.57296991e-01
1.14574909e+00 -2.10553527e+00 4.05940637e-02 8.63893479e-02
2.50771493e-01 3.48885149e-01 -2.96689570e-01 1.16779022e-01
-2.07153901e-01 -7.65177235e-02 -3.47358972e-01 -5.94757020e-01
-2.18948409e-01 4.19680953e-01 4.61336747e-02 3.02076995e-01
3.84052634e-01 8.17537844e-01 -1.01090431e+00 -9.92780149e-01
2.35264778e-01 5.41388750e-01 -9.46815358e-04 3.47873360e-01
-2.81060696e-01 7.18157947e-01 -6.54283285e-01 9.47104156e-01
8.14205587e-01 -7.32582510e-01 1.04853623e-01 -4.61803496e-01
2.18488008e-01 -1.69499516e-01 -1.39771175e+00 1.32194757e+00
-6.31208181e-01 -1.19871259e-01 -2.43034407e-01 -1.09724271e+00
8.15810919e-01 5.44408500e-01 9.03467655e-01 -2.62170672e-01
6.93809288e-03 2.98641473e-01 -4.24924850e-01 -5.90566576e-01
-1.68019757e-01 -4.56703275e-01 5.74645847e-02 5.26277125e-01
2.26204306e-01 3.14929366e-01 3.13226059e-02 -1.20270602e-01
7.63333499e-01 4.88792434e-02 3.94217312e-01 -1.01805091e-01
6.16343558e-01 -3.13255519e-01 1.16010737e+00 5.94296455e-01
-1.98934048e-01 6.78590119e-01 4.35644180e-01 -4.66972619e-01
-6.69277191e-01 -9.87962663e-01 -4.91805732e-01 9.51757967e-01
5.50469816e-01 -3.07364941e-01 -3.97026807e-01 -9.87974346e-01
-1.38876289e-02 3.79850626e-01 -6.22400343e-01 1.17409676e-02
-2.94189006e-01 -1.38164926e+00 2.07041249e-01 5.64603388e-01
5.18547118e-01 -1.18692529e+00 1.99724227e-01 3.15910786e-01
-4.47258532e-01 -1.13730085e+00 -4.29532200e-01 4.14081901e-01
-8.63588452e-01 -1.10889351e+00 -5.30346096e-01 -7.67218351e-01
7.86349237e-01 2.61324078e-01 1.16869354e+00 8.70454386e-02
-3.50851834e-01 2.29278710e-02 -3.15888047e-01 -2.60520190e-01
-1.74838677e-01 -1.21026635e-01 -1.19781680e-01 2.90103227e-01
4.06794876e-01 -4.76442248e-01 -3.68677050e-01 7.45761096e-01
-8.12392771e-01 2.53874719e-01 9.49284673e-01 1.34436333e+00
1.18111324e+00 5.04655957e-01 8.60835850e-01 -1.22381938e+00
5.98026440e-02 -7.17949986e-01 -5.69591641e-01 7.46384025e-01
-9.13161457e-01 -1.58679202e-01 5.21284342e-01 -7.69410849e-01
-1.14124525e+00 1.32048264e-01 -6.02618903e-02 -3.94463986e-01
-2.58666128e-01 5.37393212e-01 -4.24270540e-01 -3.10303550e-02
2.11241886e-01 9.49411467e-02 -1.07359543e-01 -5.47130466e-01
7.62118772e-02 7.33570397e-01 1.04081491e-03 -6.39731765e-01
4.48367774e-01 2.64662236e-01 1.73394442e-01 -2.82867432e-01
-1.51006866e+00 -8.01225245e-01 -6.57865226e-01 -2.39369899e-01
8.32935274e-01 -1.18229723e+00 -5.46423495e-01 8.35702062e-01
-8.74851406e-01 -6.92154244e-02 -1.04431836e-02 7.19723046e-01
-3.13168198e-01 3.03182989e-01 -8.44084382e-01 -6.62773728e-01
-1.25999168e-01 -1.49520338e+00 1.29554927e+00 4.60262150e-01
5.71413413e-02 -1.18204081e+00 -1.27255678e-01 7.32019722e-01
7.32375681e-02 1.07674524e-01 1.05359137e+00 -5.06645799e-01
-5.18170714e-01 -2.39543721e-01 -6.26366556e-01 5.20492971e-01
2.48687938e-01 -1.10511757e-01 -9.39928472e-01 -1.98250696e-01
9.91606386e-04 -6.70676947e-01 7.10633218e-01 5.76852441e-01
1.35304606e+00 -1.27456233e-01 -5.65767109e-01 3.69786799e-01
1.48477614e+00 3.24479602e-02 2.69232750e-01 7.43591711e-02
9.46072280e-01 7.44357586e-01 9.71048295e-01 4.30451930e-01
4.78423804e-01 8.03293824e-01 3.18275422e-01 -2.51760602e-01
-1.96950644e-01 3.83002407e-05 2.09688336e-01 1.11659479e+00
1.66749939e-01 2.67190649e-03 -6.46027744e-01 2.11452588e-01
-2.13327169e+00 -5.05540133e-01 -3.26095015e-01 2.02695990e+00
9.82477486e-01 -1.04294404e-01 -8.22490081e-02 -1.53584555e-01
9.47832406e-01 1.22268766e-01 -8.01141858e-01 2.88878262e-01
-1.38036758e-01 9.60056391e-03 3.28169376e-01 3.12459826e-01
-1.32506752e+00 6.74148619e-01 5.44190693e+00 1.33168733e+00
-7.43041456e-01 4.91743386e-01 9.32360590e-01 6.49722144e-02
-6.79052100e-02 -5.92477731e-02 -9.76791978e-01 6.07335985e-01
5.07694542e-01 4.45493788e-01 2.09789991e-01 9.65463519e-01
-3.04595262e-01 -2.15931654e-01 -8.07683825e-01 1.08593285e+00
1.02068610e-01 -8.79677057e-01 -1.97675347e-01 6.07935973e-02
9.31995869e-01 -3.19045112e-02 -1.79205313e-02 4.46003318e-01
3.47852945e-01 -8.55561018e-01 4.35419500e-01 5.60156226e-01
6.92319572e-01 -7.79680669e-01 1.11809862e+00 7.14728296e-01
-1.27780294e+00 -2.12299570e-01 -4.86234486e-01 4.46525544e-01
6.05138279e-02 1.26674688e+00 -4.42266315e-01 8.21326792e-01
7.02494264e-01 8.01172733e-01 -5.25448024e-01 9.81231868e-01
-4.50742006e-01 5.46120107e-01 -3.03202480e-01 2.21413240e-01
1.68971673e-01 -2.10711733e-01 2.04520479e-01 1.05996943e+00
1.88116014e-01 -1.34814292e-01 5.98629057e-01 7.64118850e-01
4.28574272e-02 1.78545460e-01 3.92640457e-02 3.67351979e-01
3.55506450e-01 1.63644648e+00 -6.79322839e-01 -2.33433843e-01
-6.54204726e-01 5.58955491e-01 4.52748030e-01 3.12077194e-01
-8.21343482e-01 5.87028302e-02 3.70136321e-01 -2.65324712e-01
1.03343673e-01 3.59052986e-01 -3.16290319e-01 -1.02180898e+00
-3.81080806e-02 -5.92848599e-01 5.74620664e-01 -7.46846139e-01
-1.84121406e+00 5.42772412e-01 -4.58761491e-02 -1.30053926e+00
1.82136357e-01 -3.96433413e-01 -2.64259309e-01 9.91418362e-01
-1.79978061e+00 -1.45535910e+00 -3.10938507e-01 5.52659690e-01
3.17458063e-01 -1.30755380e-01 1.00938272e+00 6.20547175e-01
-8.41701508e-01 6.52762353e-01 1.56604782e-01 -1.10269859e-01
9.95110750e-01 -1.22050941e+00 -5.06708443e-01 2.63120681e-01
1.85532078e-01 2.30650872e-01 2.60514840e-02 -5.81926942e-01
-7.39514053e-01 -1.42061806e+00 9.51559782e-01 -2.67123461e-01
4.10875171e-01 -1.46239042e-01 -1.07432282e+00 5.32559037e-01
-3.82938057e-01 4.57480341e-01 1.20954037e+00 5.05180776e-01
-4.85688031e-01 -3.64317834e-01 -1.10901713e+00 2.41349176e-01
5.27812243e-01 -4.15158480e-01 -1.84341252e-01 7.14662492e-01
5.03066838e-01 -2.40844503e-01 -1.15602553e+00 8.68978024e-01
3.81658405e-01 -7.82794535e-01 9.55911517e-01 -1.56789154e-01
5.03985167e-01 -6.58933818e-01 -2.13310003e-01 -1.06545341e+00
-5.08424342e-01 2.35032246e-01 -1.63038135e-01 1.52626991e+00
3.97261471e-01 -5.59862912e-01 5.00434399e-01 5.06064713e-01
2.82958057e-02 -1.28638709e+00 -9.25527275e-01 -6.77506208e-01
-3.16936165e-01 -4.42977458e-01 5.02607405e-01 1.15330660e+00
-2.62294024e-01 2.73922592e-01 -6.19845748e-01 4.05839235e-01
8.52693021e-01 2.56224096e-01 2.10126519e-01 -1.34718966e+00
-3.92419040e-01 -1.76352262e-01 -1.94136202e-01 -7.78999984e-01
3.66190523e-01 -9.12339747e-01 2.73233324e-01 -1.47997582e+00
8.81747186e-01 -1.09046066e+00 -9.43773389e-01 7.67729044e-01
-8.24911475e-01 2.97460973e-01 -4.01245132e-02 6.40180647e-01
-9.57315683e-01 5.28657377e-01 1.33150637e+00 -1.36216372e-01
1.49456441e-01 3.02517146e-01 -7.13206172e-01 7.86668181e-01
4.53255624e-01 -9.57153320e-01 -3.62235844e-01 1.00871146e-01
1.16224453e-01 2.04652667e-01 2.48528436e-01 -9.31488454e-01
1.54724136e-01 -2.39651233e-01 3.04265052e-01 -8.17999065e-01
2.17828929e-01 -8.69869411e-01 2.20440432e-01 1.57295600e-01
-3.09058964e-01 -5.43408394e-01 -2.89712310e-01 6.70157313e-01
-4.86358196e-01 -4.52847809e-01 9.72172499e-01 -1.07465059e-01
-4.61447418e-01 6.27612948e-01 4.60005738e-02 -1.12870529e-01
1.06992531e+00 1.07411064e-01 -2.01058269e-01 3.51996459e-02
-7.39402175e-01 5.45049250e-01 1.19461760e-01 2.57920086e-01
3.46392930e-01 -1.58412802e+00 -6.44435227e-01 -1.49818093e-01
4.45442170e-01 4.59198177e-01 6.49474561e-01 9.90319669e-01
1.86507747e-01 1.08865999e-01 2.60018945e-01 -7.67157674e-01
-1.33121884e+00 6.42070830e-01 1.71656981e-01 -9.64962482e-01
-3.66260380e-01 9.42737103e-01 6.61798358e-01 -9.30821180e-01
1.46039352e-01 1.51159644e-01 -3.61542761e-01 2.56318361e-01
6.17689371e-01 1.85787961e-01 -6.20168038e-02 -7.00964868e-01
-3.76041949e-01 9.13888752e-01 -3.04397464e-01 4.76176351e-01
1.41292989e+00 -6.78163692e-02 -4.19631273e-01 8.38163555e-01
1.42429447e+00 -3.37627113e-01 -1.43441212e+00 -7.26263821e-01
-2.19572242e-02 -2.13028714e-01 6.64854646e-02 -9.66450214e-01
-1.28230953e+00 7.25435555e-01 6.00523472e-01 -2.83606797e-01
1.22702086e+00 2.51944244e-01 7.83670247e-01 3.32722478e-02
3.26753050e-01 -1.27088511e+00 3.16186786e-01 2.39228994e-01
5.09073198e-01 -1.57294571e+00 2.22592875e-01 -7.58614182e-01
-8.23400080e-01 1.06734276e+00 6.25678003e-01 1.46883368e-01
1.02616262e+00 2.24717498e-01 2.12445199e-01 -2.37069353e-01
-6.28420293e-01 -6.10949658e-03 4.67135876e-01 4.18788105e-01
2.87053794e-01 1.78581938e-01 -2.32969433e-01 7.81940222e-01
5.91812849e-01 -1.63992203e-03 -1.77944601e-01 5.52574635e-01
-1.89801112e-01 -1.30025113e+00 -4.55139995e-01 6.31833613e-01
-3.95891994e-01 4.06886041e-02 2.11543620e-01 5.18741190e-01
7.26796806e-01 9.60722744e-01 -2.72527367e-01 -5.02517462e-01
2.20700856e-02 1.15545206e-02 2.28714138e-01 -6.83463812e-01
-4.74356592e-01 2.97840089e-01 -1.66544542e-01 -4.48533237e-01
-8.06982338e-01 -7.14562356e-01 -1.20208597e+00 3.00872654e-01
-6.97500288e-01 -1.60503760e-02 6.54787600e-01 1.26460540e+00
5.95195703e-02 7.14763045e-01 8.64465952e-01 -6.15438640e-01
-3.59601319e-01 -1.06400275e+00 -1.01978791e+00 4.14483875e-01
-1.49777001e-02 -9.34206307e-01 -4.68760103e-01 -1.07282184e-01] | [9.571893692016602, 4.0221476554870605] |
8f576c64-eff6-452e-8445-027f5e3a646c | a-hierarchical-transformer-with-speaker | 2012.14781 | null | https://arxiv.org/abs/2012.14781v1 | https://arxiv.org/pdf/2012.14781v1.pdf | A Hierarchical Transformer with Speaker Modeling for Emotion Recognition in Conversation | Emotion Recognition in Conversation (ERC) is a more challenging task than conventional text emotion recognition. It can be regarded as a personalized and interactive emotion recognition task, which is supposed to consider not only the semantic information of text but also the influences from speakers. The current method models speakers' interactions by building a relation between every two speakers. However, this fine-grained but complicated modeling is computationally expensive, hard to extend, and can only consider local context. To address this problem, we simplify the complicated modeling to a binary version: Intra-Speaker and Inter-Speaker dependencies, without identifying every unique speaker for the targeted speaker. To better achieve the simplified interaction modeling of speakers in Transformer, which shows excellent ability to settle long-distance dependency, we design three types of masks and respectively utilize them in three independent Transformer blocks. The designed masks respectively model the conventional context modeling, Intra-Speaker dependency, and Inter-Speaker dependency. Furthermore, different speaker-aware information extracted by Transformer blocks diversely contributes to the prediction, and therefore we utilize the attention mechanism to automatically weight them. Experiments on two ERC datasets indicate that our model is efficacious to achieve better performance. | ['Weiping Wang', 'Qingyi Si', 'Peng Fu', 'Zheng Lin', 'Jiangnan Li'] | 2020-12-29 | null | null | null | null | ['emotion-recognition-in-conversation'] | ['natural-language-processing'] | [ 5.21340445e-02 -1.86648428e-01 6.17890395e-02 -8.39312375e-01
-5.30639708e-01 -2.10117385e-01 4.62911963e-01 -2.25052685e-01
-2.78547049e-01 2.09223390e-01 7.42068291e-01 -8.78316164e-02
5.81115559e-02 -3.18710119e-01 -2.24518478e-01 -7.57528543e-01
1.32268205e-01 1.95272699e-01 4.11924571e-02 -4.51175958e-01
1.53275728e-01 2.16733798e-01 -1.68749607e+00 5.63315392e-01
1.11970437e+00 9.54243243e-01 2.61620641e-01 4.27862972e-01
-7.84383595e-01 7.17851639e-01 -7.94913828e-01 -2.86415577e-01
-6.43429637e-01 -5.45693099e-01 -5.62528610e-01 1.20919414e-01
-5.26530743e-01 5.75555116e-02 9.75511745e-02 9.24192250e-01
6.21890068e-01 8.56770948e-02 6.58549786e-01 -1.25443006e+00
-4.17676508e-01 8.80358279e-01 -4.88073021e-01 1.19914003e-01
4.17524368e-01 -3.61979395e-01 8.65357339e-01 -8.48416090e-01
1.28971403e-02 1.39133453e+00 4.47470844e-01 6.22252405e-01
-4.62556988e-01 -9.65546370e-01 9.13271308e-01 4.87138867e-01
-1.38589573e+00 -5.65725088e-01 1.18425095e+00 -2.31190398e-01
7.75403500e-01 5.47005773e-01 5.35747111e-01 1.29437673e+00
-1.26995072e-01 9.40426707e-01 9.55074251e-01 -3.04325759e-01
-2.45451853e-02 6.39191091e-01 3.70476067e-01 2.44577572e-01
-5.69935083e-01 -4.27234530e-01 -6.24427438e-01 -7.45790005e-02
1.43925145e-01 1.23134129e-01 -5.29661059e-01 2.33762339e-01
-8.90985489e-01 6.32130742e-01 1.33028746e-01 5.86874127e-01
-1.82934314e-01 -3.44117314e-01 5.25323570e-01 7.87590444e-02
5.93945324e-01 -2.65302032e-01 -6.22846663e-01 -4.48124915e-01
-4.17295754e-01 -3.82854372e-01 8.49306285e-01 9.10412312e-01
6.94027066e-01 5.66729158e-03 -8.54580030e-02 1.22396851e+00
6.15564048e-01 4.18908894e-01 8.66045713e-01 -2.19563648e-01
5.28055727e-01 8.04324389e-01 1.64958264e-03 -1.22024775e+00
-5.01808226e-01 -4.82993424e-01 -9.45937157e-01 -5.18370688e-01
-1.97888106e-01 -3.37508410e-01 -5.64618945e-01 1.81323349e+00
4.98311877e-01 2.37516522e-01 1.42207429e-01 8.63291621e-01
1.05370140e+00 9.46798503e-01 3.69774818e-01 -5.71173549e-01
1.51899731e+00 -1.06952894e+00 -1.24620736e+00 -3.23683262e-01
4.51907992e-01 -8.14349771e-01 1.15858102e+00 3.06109756e-01
-4.81847793e-01 -3.75344366e-01 -9.08047080e-01 -4.45577828e-03
-5.05297422e-01 2.97023743e-01 5.71621418e-01 7.94149697e-01
-6.50181115e-01 -8.05273354e-02 -3.71994436e-01 -1.45058990e-01
-3.89002673e-02 3.14785480e-01 -4.75197472e-02 3.21491480e-01
-1.66618896e+00 7.69635797e-01 -4.68410477e-02 6.67949796e-01
-3.41636002e-01 -1.65394247e-01 -8.05681050e-01 3.26572120e-01
3.51807892e-01 3.16320779e-03 1.01419699e+00 -1.38409722e+00
-1.95593417e+00 3.98268968e-01 -8.26466441e-01 3.24113131e-01
2.17366219e-01 -1.87243208e-01 -1.01323271e+00 -2.75988013e-01
-3.56511414e-01 1.29401058e-01 7.99489021e-01 -1.36976349e+00
-6.77327514e-01 -4.55788493e-01 -2.56238222e-01 6.52434349e-01
-7.48315513e-01 4.98804301e-01 -8.39660883e-01 -5.71771085e-01
1.50278330e-01 -5.94487906e-01 -5.82284331e-02 -6.05291545e-01
-3.54575276e-01 -4.54573780e-01 1.12260246e+00 -6.41240656e-01
1.62546563e+00 -2.39665747e+00 2.83888057e-02 2.68385500e-01
-1.23743750e-01 1.60299778e-01 1.83618627e-03 3.67265373e-01
-2.52089761e-02 1.35195285e-01 -7.63793439e-02 -5.66691577e-01
2.51940370e-01 3.26630950e-01 -3.47126544e-01 4.26404104e-02
4.49545905e-02 6.58076942e-01 -4.13904667e-01 -6.51903093e-01
1.59137875e-01 7.85812557e-01 -3.95543814e-01 2.53939241e-01
3.93670313e-02 4.29584712e-01 -7.77281523e-01 5.10704339e-01
7.38016844e-01 1.59205310e-02 3.17848057e-01 -2.71912932e-01
-6.82568625e-02 2.83811331e-01 -1.33620214e+00 1.20890152e+00
-7.30473042e-01 2.51222759e-01 4.00535583e-01 -1.13743412e+00
1.22722864e+00 4.90181535e-01 2.97041297e-01 -5.46754479e-01
4.27314252e-01 -6.84438124e-02 -1.45569101e-01 -8.55502486e-01
3.50615740e-01 -3.10983002e-01 -3.68642271e-01 4.29374039e-01
-1.15072407e-01 3.45874876e-01 -3.93618852e-01 7.84860179e-02
5.32668412e-01 -2.93382049e-01 -1.01056937e-02 -1.99354261e-01
9.77592170e-01 -6.81455851e-01 9.16972697e-01 2.33602583e-01
-4.94288206e-01 1.38400793e-01 4.91484940e-01 -1.59012228e-01
-2.07224235e-01 -4.38348204e-01 -2.92433472e-03 1.55751753e+00
5.31167626e-01 -5.06395519e-01 -7.89165139e-01 -8.56090665e-01
-3.91363174e-01 6.83587253e-01 -6.88070416e-01 -3.44453454e-01
-2.93320745e-01 -7.99666524e-01 3.26878130e-01 4.67979580e-01
5.41958749e-01 -1.28089249e+00 -4.02156189e-02 2.71381229e-01
-6.45725906e-01 -9.50251937e-01 -8.73780727e-01 2.06276044e-01
-2.54412562e-01 -6.87348425e-01 -4.53416824e-01 -9.72887576e-01
3.88267100e-01 1.27323434e-01 8.83230150e-01 7.55185559e-02
8.30926076e-02 3.57333362e-01 -7.05141783e-01 -6.03974164e-01
-3.77020217e-03 1.09150179e-01 3.09755560e-03 7.13236988e-01
7.31225729e-01 -5.60108304e-01 -3.73766094e-01 6.58457756e-01
-7.47407019e-01 1.88147761e-02 4.34578419e-01 7.47287154e-01
2.37316668e-01 2.88422585e-01 8.70809853e-01 -7.25271404e-01
7.22656131e-01 -4.47419673e-01 6.56378493e-02 6.02089584e-01
-2.31735617e-01 -1.85928315e-01 6.48438990e-01 -5.66791654e-01
-1.64076889e+00 7.59972706e-02 -4.72818166e-01 -8.28765109e-02
-2.71497220e-01 5.86050332e-01 -9.43631172e-01 2.22335264e-01
-7.37797990e-02 4.55942154e-01 -2.60807335e-01 -4.93198246e-01
1.64855406e-01 1.28701460e+00 6.87487274e-02 -6.40778184e-01
1.98542565e-01 5.39980382e-02 -6.43878877e-01 -9.00664747e-01
-7.38588274e-01 -4.39563602e-01 -3.09077352e-01 -5.37412584e-01
8.88773680e-01 -8.28598619e-01 -1.05084598e+00 6.22895896e-01
-1.20171654e+00 -1.99370813e-02 3.32946569e-01 5.29634595e-01
1.38849467e-01 2.60111272e-01 -6.11916065e-01 -1.23743713e+00
-3.33935231e-01 -1.32181799e+00 1.09961903e+00 5.02368748e-01
-3.22463363e-02 -1.05494058e+00 -3.23839933e-01 3.77144605e-01
5.04548848e-01 -2.90348798e-01 7.67223418e-01 -7.54373908e-01
-3.74261960e-02 -7.67243505e-02 -5.50681949e-02 4.11468834e-01
3.55899125e-01 1.68766394e-01 -1.16875029e+00 2.07013294e-01
2.95585543e-01 -1.35561585e-01 6.43782437e-01 -1.40908331e-01
1.43739724e+00 -2.41289914e-01 -3.70552093e-01 3.08922023e-01
8.78332138e-01 5.55315793e-01 5.43423474e-01 -7.16549307e-02
7.23883629e-01 9.83861387e-01 5.33341706e-01 5.56008875e-01
8.39383364e-01 5.74671149e-01 9.47103798e-02 -4.32680398e-02
5.91965020e-01 -1.10784687e-01 5.35920203e-01 1.33755505e+00
1.09715380e-01 -4.19979274e-01 -5.94748378e-01 1.63099900e-01
-1.82245016e+00 -9.51231778e-01 -1.41737327e-01 1.74305475e+00
8.54224801e-01 -3.63927111e-02 -1.90623179e-01 5.90206347e-02
9.53126431e-01 2.68242151e-01 -4.84212995e-01 -5.75556397e-01
-1.26909629e-01 -2.17563584e-01 -1.74955949e-01 5.56620181e-01
-1.01076305e+00 8.63337457e-01 5.93243217e+00 1.00786650e+00
-1.28217518e+00 1.52070392e-02 8.63936007e-01 -6.77845627e-03
-6.47996724e-01 -2.68997431e-01 -9.42108154e-01 8.42748642e-01
7.20618427e-01 -5.24774753e-02 3.71629745e-01 7.38006711e-01
2.51879245e-01 1.57237247e-01 -8.28862488e-01 1.23037267e+00
3.88323963e-01 -5.62973142e-01 7.32505098e-02 -2.14141846e-01
2.37632796e-01 -5.05463541e-01 -1.05888173e-01 7.68032551e-01
4.41844538e-02 -9.25111055e-01 5.50543845e-01 5.26593924e-01
3.91224325e-01 -9.28903401e-01 7.69580841e-01 5.59318602e-01
-1.57727075e+00 -7.93954432e-02 -3.79933268e-02 6.14077970e-02
2.52408981e-01 6.07831597e-01 -1.56880260e-01 5.89365840e-01
9.18591440e-01 6.57962084e-01 -2.45449349e-01 3.73082340e-01
-2.11389348e-01 6.93632185e-01 -3.86909366e-01 -4.45473582e-01
4.74542901e-02 -2.59316027e-01 1.92141458e-01 1.38324380e+00
3.31748337e-01 4.09908414e-01 2.60568827e-01 3.58375698e-01
-1.06382407e-02 5.94218135e-01 -3.88168022e-02 1.88135639e-01
6.09660685e-01 1.36355340e+00 -3.62707198e-01 -4.76284295e-01
-3.81700039e-01 1.12715077e+00 4.06269968e-01 4.74102527e-01
-1.12323618e+00 -6.24721944e-01 5.90023339e-01 -3.84474546e-01
2.86107093e-01 1.65174797e-01 -1.20408840e-01 -1.32973242e+00
2.06067383e-01 -8.97067964e-01 3.77807289e-01 -7.60728657e-01
-1.27821672e+00 8.43485355e-01 -4.54112113e-01 -9.94802415e-01
9.30045620e-02 -4.14586782e-01 -1.04512250e+00 1.10983038e+00
-1.35320055e+00 -1.04497910e+00 -4.50749487e-01 8.27249587e-01
5.77116072e-01 1.60418540e-01 9.01155472e-01 5.19365251e-01
-1.14979661e+00 8.48716497e-01 -2.36248914e-02 -8.20972696e-02
9.63099897e-01 -9.78474379e-01 -3.10012311e-01 4.92824525e-01
-2.21974283e-01 6.63433313e-01 4.52075511e-01 -3.65766883e-01
-1.23335874e+00 -8.97615016e-01 1.27685988e+00 -1.42069444e-01
4.77246225e-01 -6.05631828e-01 -1.12514055e+00 7.24913716e-01
2.41366506e-01 -3.47818851e-01 1.04320288e+00 6.17008567e-01
-2.86128104e-01 -4.11437392e-01 -8.58918905e-01 5.46619415e-01
8.08429241e-01 -6.64457142e-01 -6.17466688e-01 -1.52267292e-02
7.95632482e-01 -1.04930006e-01 -6.28004730e-01 4.11444366e-01
5.05831897e-01 -8.99748027e-01 5.88493049e-01 -2.68925130e-01
1.93694234e-01 -2.68960446e-01 -9.31482017e-02 -1.33491886e+00
-1.57090425e-01 -4.73877460e-01 2.37985983e-01 1.98234499e+00
4.29978162e-01 -9.25803363e-01 5.95417619e-01 8.67323875e-01
-3.25896055e-01 -8.90238523e-01 -7.97687411e-01 -2.25805089e-01
-2.14409649e-01 -5.73824823e-01 1.03093255e+00 1.28244722e+00
5.64640343e-01 6.18157208e-01 -5.98097384e-01 2.32156068e-01
6.33180439e-02 6.57370612e-02 5.11023283e-01 -1.08847404e+00
-1.28572971e-01 -5.98816156e-01 4.13737968e-02 -1.37161887e+00
3.46659482e-01 -4.38296080e-01 3.57994050e-01 -1.34277391e+00
8.79649892e-02 -5.72636127e-01 -4.60020244e-01 3.76082450e-01
-5.45808315e-01 -2.11122274e-01 -2.33798608e-01 2.50050128e-02
-7.66308188e-01 1.13004804e+00 1.11372721e+00 -9.77606326e-02
-3.24261934e-01 2.30055507e-02 -9.37177896e-01 8.29000831e-01
8.49332809e-01 -1.04874320e-01 -5.11975944e-01 -3.36639732e-01
-5.55809066e-02 1.37720019e-01 -6.01125173e-02 -5.60233533e-01
4.23875332e-01 -2.24785030e-01 3.62836331e-01 -6.98264778e-01
5.13971806e-01 -9.74407077e-01 3.66287604e-02 -7.85012618e-02
-3.32810491e-01 -2.83777237e-01 1.13733709e-01 4.98438418e-01
-5.52551031e-01 -5.63861150e-03 3.36006731e-01 8.93048570e-02
-7.17281520e-01 2.81601399e-01 -5.64336896e-01 -1.86843038e-01
8.69573951e-01 8.23558569e-02 -1.06917210e-01 -5.61350107e-01
-9.19652462e-01 6.10286951e-01 -1.64412960e-01 6.13621235e-01
4.43422616e-01 -1.33287585e+00 -4.44856137e-01 2.66724706e-01
2.33937889e-01 -1.12182543e-01 9.53244388e-01 6.69135869e-01
2.37060845e-01 1.35857135e-01 3.27169478e-01 -3.80926371e-01
-1.55986428e+00 5.04225552e-01 4.83747423e-01 -2.69251466e-01
-1.11536108e-01 9.01593149e-01 4.74064410e-01 -6.36905074e-01
5.83186388e-01 -1.41457915e-01 -8.32132757e-01 2.80701488e-01
6.47450149e-01 1.86623801e-02 -1.23462819e-01 -9.43325758e-01
-6.09691203e-01 6.95746899e-01 -2.01611593e-01 6.99645951e-02
9.84153688e-01 -6.73442543e-01 -1.90723404e-01 6.79065466e-01
1.22931433e+00 2.00309709e-01 -9.27479565e-01 -3.88709158e-01
-2.31788471e-01 -9.42400694e-02 2.47507598e-02 -9.27398026e-01
-1.04542828e+00 1.05237913e+00 3.01138073e-01 3.13465804e-01
1.37831688e+00 1.08181899e-02 6.34511471e-01 2.24932224e-01
1.12940960e-01 -1.33382976e+00 -3.65049019e-02 7.53357470e-01
9.88660812e-01 -1.06035113e+00 -4.51919109e-01 -7.19756424e-01
-1.02600777e+00 9.17152584e-01 9.74403501e-01 4.65339690e-01
1.05140984e+00 1.86534896e-01 4.38792974e-01 -4.69323061e-02
-8.25358570e-01 -4.56443205e-02 3.00106734e-01 2.90941834e-01
5.48480332e-01 1.06199861e-01 -2.89361984e-01 1.55974185e+00
-1.68088496e-01 -3.49394619e-01 1.03806391e-01 6.64282441e-01
-4.59665954e-01 -1.14877176e+00 -3.80185068e-01 2.31283993e-01
-4.27784503e-01 -1.39328778e-01 -4.51197684e-01 3.94871503e-01
2.55936623e-01 1.36294806e+00 -1.99476001e-03 -7.52451122e-01
4.63901401e-01 3.20687711e-01 -2.92937517e-01 -2.28563443e-01
-6.76103294e-01 3.99362177e-01 1.63208976e-01 -3.72546256e-01
-5.47357440e-01 -4.69971865e-01 -1.32031298e+00 -2.33802900e-01
-4.83626932e-01 6.19800448e-01 7.16624260e-01 1.19641566e+00
4.95281041e-01 8.54273319e-01 1.12937939e+00 -7.60071337e-01
-2.23639637e-01 -1.22530031e+00 -7.75474191e-01 3.51380438e-01
1.05014086e-01 -6.39341295e-01 -7.63819933e-01 -7.14200288e-02] | [13.034701347351074, 6.067753791809082] |
df57570b-d312-4ea0-a00e-d1ee5daa8af1 | decision-making-using-rough-set-based | 2107.12477 | null | https://arxiv.org/abs/2107.12477v1 | https://arxiv.org/pdf/2107.12477v1.pdf | Decision Making Using Rough Set based Spanning Sets for a Decision System | Rough Set based concepts of Span and Spanning Sets were recently proposed to deal with uncertainties in data. Here, this paper, presents novel concepts for generic decision-making process using Rough Set based span for a decision table. Majority of problems in Artificial Intelligence deal with decision making. This paper provides real life applications of proposed Rough Set based span for decision tables. Here, novel concept of span for a decision table is proposed, illustrated with real life example of flood relief and rescue team assignment. Its uses, applications and properties are explored. The key contribution of paper is primarily to study decision making using Rough Set based Span for a decision tables, as against an information system in prior works. Here, the main contribution is that decision classes are automatically learned by the technique of Rough Set based span, for a particular problem, hence automating the decision-making process. These decision-making tools based on span can guide an expert in taking decisions in tough and time-bound situations. | ['Nidhika Yadav'] | 2021-07-21 | null | null | null | null | ['novel-concepts'] | ['reasoning'] | [ 2.27592409e-01 4.05870676e-01 2.35533297e-01 -8.99151325e-01
-4.16839570e-01 -3.36522877e-01 1.00241683e-01 9.23807442e-01
5.58782220e-02 1.15650332e+00 2.05271780e-01 -2.65716702e-01
-1.17302561e+00 -1.25199723e+00 -2.08708644e-02 -1.90268010e-01
-7.00615525e-01 1.17076302e+00 -1.64137036e-01 -7.70385683e-01
7.02786863e-01 9.27914262e-01 -1.91251349e+00 5.07121921e-01
1.00760686e+00 1.27096128e+00 -1.77854523e-01 2.76789099e-01
-3.06630284e-01 8.23766768e-01 -7.19196498e-01 -1.60604373e-01
7.23674536e-01 -2.35390645e-02 -1.08350968e+00 1.32580651e-02
-3.89449447e-01 -1.72258556e-01 3.12384993e-01 6.99768066e-01
3.55969936e-01 6.59883618e-01 1.18845916e+00 -1.35917795e+00
-8.49769190e-02 1.23166227e+00 -3.85375440e-01 3.13573182e-01
4.06927109e-01 -3.78954411e-01 5.24780691e-01 -4.70742315e-01
1.26647145e-01 1.75604284e+00 6.01542294e-01 -1.36240080e-01
-7.04473794e-01 -3.85926783e-01 1.03321418e-01 2.84335375e-01
-1.32600379e+00 -5.24518527e-02 4.70338166e-01 -3.92524838e-01
9.27208662e-01 3.06724817e-01 5.94440341e-01 -4.20217633e-01
8.52202892e-01 1.99457631e-01 1.47281730e+00 -5.93115687e-01
9.25544620e-01 -1.20237187e-01 7.36204147e-01 3.30952406e-01
9.77659464e-01 4.19186912e-02 -1.85819343e-01 -6.04825877e-02
-4.08378467e-02 1.77820280e-01 7.66178071e-02 1.26261145e-01
-5.15173674e-01 9.01362658e-01 2.59392291e-01 3.63273233e-01
-7.22941935e-01 -2.67351449e-01 5.89784682e-01 7.55021155e-01
2.98366398e-01 4.70588714e-01 -4.51038182e-01 2.63393283e-01
-7.89460659e-01 5.83249092e-01 9.47034657e-01 1.05357897e+00
7.16039181e-01 1.86115876e-01 -3.77310097e-01 1.87459052e-01
5.67226410e-01 3.84067327e-01 1.56030387e-01 -5.90792775e-01
4.52045113e-01 7.97429323e-01 2.23311201e-01 -1.04391003e+00
-9.77917492e-01 -8.37407410e-02 -5.20715892e-01 6.81079388e-01
-8.32291991e-02 -3.41087312e-01 -1.33064985e+00 9.97136831e-01
5.62007010e-01 -5.76929986e-01 6.00540221e-01 3.36077482e-01
6.74354613e-01 5.66386282e-01 2.56792396e-01 -5.57240963e-01
1.70487487e+00 2.85940617e-01 -1.17398369e+00 2.03721583e-01
5.30137837e-01 -4.64455307e-01 3.58624339e-01 9.10488367e-01
-6.47234738e-01 -3.18030089e-01 -1.55033803e+00 4.91022855e-01
-7.18105614e-01 -4.73234594e-01 8.99766088e-01 7.37811208e-01
-6.11417949e-01 7.73851156e-01 -2.40918949e-01 -4.04869825e-01
3.78082544e-01 6.52270675e-01 -1.15436576e-02 -3.96047235e-01
-1.62292039e+00 1.61097479e+00 8.71737599e-01 1.61533087e-01
-7.31727540e-01 -4.31248367e-01 -1.16269028e+00 -3.54115814e-02
4.87005383e-01 -5.36944926e-01 1.06763387e+00 1.92252249e-01
-1.10125673e+00 4.67057437e-01 5.61179936e-01 -1.01665103e+00
2.90566534e-01 -1.63282260e-01 -8.44563961e-01 5.64569235e-02
2.81897485e-01 -2.34810878e-02 4.35893118e-01 -1.17462015e+00
-8.92096579e-01 -7.56695688e-01 1.21615104e-01 2.00419813e-01
6.26408160e-01 -1.08717792e-02 9.39537346e-01 -3.46608400e-01
3.53995442e-01 -3.44585091e-01 -6.28936529e-01 -7.45919883e-01
-2.27142781e-01 -4.09739494e-01 4.91393566e-01 -2.98005998e-01
1.54803526e+00 -1.49083805e+00 -6.48359835e-01 7.61758983e-01
-3.04468125e-01 -5.53864613e-02 6.81756198e-01 1.04525220e+00
-2.47410149e-03 2.71976054e-01 -8.12528968e-01 4.13687766e-01
-3.96221839e-02 4.91953969e-01 -3.71305525e-01 3.30183476e-01
3.23251426e-01 -1.68187916e-01 -5.17437041e-01 -6.82970703e-01
5.03884256e-01 -1.13791242e-01 3.05697739e-01 -3.12400043e-01
1.88474897e-02 -2.53673255e-01 -8.45546007e-01 1.03194404e+00
1.06609917e+00 7.57204354e-01 6.44052550e-02 -2.49663010e-01
-2.78890640e-01 -6.87929094e-02 -1.63295603e+00 1.32592773e+00
-4.90867019e-01 -8.04925570e-04 -2.66161263e-01 -1.31716609e+00
1.55714858e+00 3.60045165e-01 5.49558043e-01 -3.07747245e-01
6.16480410e-01 1.45733640e-01 1.21120118e-01 -5.05882204e-01
4.81858969e-01 -7.44190335e-01 -5.12367785e-01 6.48638427e-01
-2.37416521e-01 -7.24325478e-01 6.32507145e-01 -3.13949920e-02
1.05817235e+00 4.10290845e-02 1.00995076e+00 -8.30643773e-01
6.21987879e-01 5.34937561e-01 7.38058627e-01 2.90978998e-01
-1.13211453e-01 3.40944648e-01 3.98948938e-01 -1.05439603e+00
-6.03861094e-01 -6.01281643e-01 -6.94328666e-01 5.20378530e-01
1.73108339e-01 8.50715265e-02 -5.79687119e-01 -5.22953570e-01
4.45135385e-01 1.03943765e+00 -8.39534342e-01 7.50278076e-03
-2.15198502e-01 -9.47710395e-01 2.86538631e-01 1.20867796e-01
5.72550356e-01 -1.03732657e+00 -1.24518859e+00 8.12921524e-01
4.69246328e-01 -7.14973807e-01 5.78226388e-01 3.90944809e-01
-1.24824333e+00 -1.17148483e+00 1.57166407e-01 -1.93379432e-01
5.56174040e-01 2.32748520e-02 1.26914716e+00 3.41667533e-02
-5.02751410e-01 1.23330830e-02 -8.89228404e-01 -1.49185193e+00
-1.11000121e-01 -5.63653708e-01 5.34984648e-01 -2.01262981e-01
5.42265654e-01 -5.89196205e-01 -3.87088329e-01 3.75422239e-01
-1.05052912e+00 -5.55939317e-01 6.22236609e-01 3.55406135e-01
5.91034591e-01 9.61474597e-01 1.28051603e+00 -1.07083344e+00
9.50483978e-01 -6.98292017e-01 -5.18409669e-01 4.81817573e-01
-9.84698176e-01 4.67839599e-01 2.91082919e-01 5.34001589e-01
-1.20221615e+00 -1.18176527e-01 1.10828891e-01 2.21675381e-01
-1.84499726e-01 6.83990240e-01 -4.30088609e-01 2.22412303e-01
9.95782495e-01 -5.13141870e-01 -6.38399199e-02 -3.16418529e-01
-3.96277867e-02 8.25293362e-01 2.44123340e-02 -7.67334700e-01
5.77139914e-01 5.01677692e-01 6.95063829e-01 -1.98273316e-01
-8.32765639e-01 -3.30481827e-01 -6.94055974e-01 -3.29917252e-01
4.00947213e-01 -6.39412284e-01 -5.01447141e-01 -2.69898742e-01
-5.57125092e-01 3.02789688e-01 -5.87798417e-01 4.00632888e-01
-9.24753726e-01 -2.30524927e-01 7.49450028e-02 -1.39512074e+00
-7.70955026e-01 -7.24412978e-01 3.57618213e-01 3.09127569e-01
-2.94734418e-01 -7.40430295e-01 -1.67788103e-01 6.42753616e-02
1.22754768e-01 1.26724744e+00 8.84887695e-01 -1.00411332e+00
-4.90769139e-03 -5.02091110e-01 2.67397404e-01 4.36351635e-02
4.21319962e-01 5.96072823e-02 -6.41508996e-01 8.80706385e-02
3.94468546e-01 -1.76087856e-01 4.60458428e-01 3.23296010e-01
6.33598626e-01 -2.60649383e-01 -1.91077366e-01 -2.14700936e-03
1.86573350e+00 8.24646950e-01 7.35497117e-01 5.35122693e-01
-1.77524552e-01 1.20948076e+00 1.96502423e+00 1.06612194e+00
4.09880608e-01 -1.25292882e-01 2.85782456e-01 6.27029657e-01
4.46407229e-01 3.81057829e-01 -1.76011816e-01 1.70299903e-01
-2.89728522e-01 -1.57484174e-01 -1.18983841e+00 5.53921163e-01
-1.84052527e+00 -8.83299768e-01 -4.58893403e-02 2.13926101e+00
2.59884775e-01 4.19407994e-01 2.00392120e-02 1.11944771e+00
8.01872075e-01 -4.44834560e-01 -3.88463974e-01 -1.20308435e+00
1.00318886e-01 2.36720905e-01 6.89558744e-01 5.99160075e-01
-8.78133893e-01 4.28017646e-01 5.50390625e+00 5.79272568e-01
-5.58221757e-01 -4.54177082e-01 5.45948088e-01 2.29540825e-01
-3.38468671e-01 2.04159468e-01 -8.68670046e-01 2.19417319e-01
9.44961488e-01 -6.55196488e-01 -8.00102428e-02 6.17550790e-01
3.33250821e-01 -6.50487661e-01 -1.15963852e+00 7.26305366e-01
-2.33623013e-01 -1.15465152e+00 2.68589586e-01 -1.81428745e-01
5.65447927e-01 -6.65046096e-01 -3.49321872e-01 3.74794632e-01
9.12894607e-01 -1.08307898e+00 5.59256971e-01 8.01381767e-01
5.49312353e-01 -1.66158605e+00 1.43626988e+00 8.23204070e-02
-1.17263579e+00 -4.57248330e-01 -6.75139666e-01 -6.63855195e-01
1.92706063e-01 1.07680678e+00 -1.11619282e+00 1.38585293e+00
7.74610519e-01 3.58604014e-01 -1.12931952e-01 1.05092669e+00
1.40389442e-01 1.44090593e-01 -5.37351727e-01 9.47870612e-02
2.01185212e-01 -5.65699711e-02 3.30317974e-01 9.17748570e-01
3.56631577e-01 8.95751297e-01 2.16089338e-01 3.97369653e-01
5.17430484e-01 2.77146846e-01 -1.01618826e+00 4.75873739e-01
9.01287079e-01 8.81137967e-01 -1.08783877e+00 -4.39570844e-01
3.93313468e-01 -2.84050524e-01 -1.53580382e-01 -2.50557572e-01
-1.47990897e-01 -1.02377570e+00 3.81214678e-01 2.92239040e-01
-7.03286752e-02 4.67176847e-02 -1.22826397e+00 -2.07418114e-01
-9.52919573e-02 -5.27252972e-01 1.10295320e+00 -3.29857200e-01
-1.23378932e+00 7.96658576e-01 9.36334610e-01 -1.73324633e+00
-2.73802608e-01 -4.50957477e-01 -8.67261767e-01 7.35657752e-01
-1.12268603e+00 -7.33753800e-01 -2.29894146e-01 2.43693620e-01
5.85219920e-01 -2.91317910e-01 6.60337627e-01 -3.27823192e-01
-1.29925147e-01 1.87330544e-02 3.19334865e-02 -4.56942469e-01
3.03332299e-01 -1.22994244e+00 -6.55371835e-03 8.27830613e-01
-7.53605008e-01 4.32939738e-01 1.26608419e+00 -1.02871573e+00
-1.14042675e+00 -9.68441248e-01 4.54424649e-01 -3.69375527e-01
1.97277993e-01 1.03870362e-01 -7.37379670e-01 3.14165324e-01
-7.79636279e-02 -1.63062692e-01 7.67538190e-01 -3.26226689e-02
3.38874936e-01 -8.12684178e-01 -2.39784765e+00 -7.43487552e-02
6.40881598e-01 3.71972263e-01 -1.17516351e+00 1.43633991e-01
4.46253777e-01 -1.63095519e-01 -1.36717153e+00 8.03878844e-01
3.76086861e-01 -1.00272846e+00 6.00670278e-01 -3.73783171e-01
-9.68545377e-02 -7.37200856e-01 -3.80898565e-01 -1.36128175e+00
-1.23068221e-01 -6.15358353e-01 9.71812680e-02 1.33002865e+00
3.31913471e-01 -4.75375652e-01 3.69014084e-01 9.86281395e-01
-3.20554107e-01 -9.99770582e-01 -1.08241796e+00 -5.53828061e-01
-1.62337214e-01 -3.91969025e-01 1.37173200e+00 6.20338380e-01
2.73720343e-02 1.00427449e-01 1.54784052e-02 2.10046321e-01
9.93905425e-01 8.71492997e-02 2.85731256e-01 -1.82206154e+00
6.39747202e-01 2.69227833e-01 -9.58247244e-01 4.66572344e-01
-5.32250166e-01 -2.49460697e-01 6.76812083e-02 -2.07348132e+00
-6.34572685e-01 -1.13927555e+00 -4.97257084e-01 3.83011490e-01
1.21357754e-01 -2.30149552e-01 3.63371551e-01 1.18365064e-01
-1.51107714e-01 2.76731879e-01 9.85685289e-01 -2.96539851e-02
-2.64053673e-01 2.99540371e-01 -1.16182029e+00 8.81505787e-01
1.16323888e+00 -8.43935609e-01 -8.19341242e-01 1.99916407e-01
5.21715879e-01 4.63604122e-01 -7.62888134e-01 -1.51072800e+00
6.35531023e-02 -6.12714112e-01 4.31993186e-01 -1.17362738e+00
-7.25637227e-02 -1.07918644e+00 4.21378106e-01 6.86385870e-01
-6.68465272e-02 3.32869321e-01 2.44412765e-01 5.15528381e-01
-2.98891574e-01 -7.03253150e-01 7.51221657e-01 -3.19637775e-01
-9.14156973e-01 1.71977922e-01 -5.42773902e-01 -6.40853271e-02
1.69246233e+00 -6.47979081e-01 2.00462028e-01 -1.15481704e-01
-9.18993533e-01 7.20625341e-01 -3.71005833e-02 -1.45187080e-01
7.02872574e-01 -1.13678968e+00 -1.03775215e+00 -3.21484834e-01
1.25173833e-02 7.22161472e-01 4.96956229e-01 2.21701115e-01
-9.28804040e-01 2.81995565e-01 -9.79539454e-01 -6.04513884e-02
-8.86473835e-01 7.31984138e-01 9.90866423e-02 -4.84624922e-01
-5.05882740e-01 5.14865100e-01 -4.42351162e-01 -9.79512408e-02
-9.64013934e-02 -8.55104625e-01 -9.95457292e-01 6.05921030e-01
8.39440286e-01 8.22323203e-01 5.86265028e-01 -3.53400975e-01
-4.70644325e-01 8.33502591e-01 -3.51878814e-02 -2.55724281e-01
1.39926839e+00 -4.15532589e-01 -1.01027265e-01 6.96816683e-01
2.74865568e-01 -4.52472478e-01 -6.90942943e-01 -8.31977054e-02
4.71132010e-01 -5.30409932e-01 -1.19635753e-01 -1.11123693e+00
-3.00237954e-01 5.27841926e-01 7.41283536e-01 3.34604472e-01
1.59049857e+00 -5.04041910e-01 3.02648842e-01 8.30051184e-01
1.05368555e+00 -1.42065775e+00 -7.67100930e-01 4.95580107e-01
1.38890290e+00 -1.27274704e+00 7.64850795e-01 -5.06592035e-01
-9.05505240e-01 1.32845771e+00 2.50358433e-01 -4.15183812e-01
1.06535351e+00 5.02671838e-01 -6.65054703e-03 -4.49626267e-01
-7.53150403e-01 -3.75597179e-01 -1.32254839e-01 1.06309152e+00
1.68652441e-02 4.41166461e-01 -7.65943348e-01 7.91287124e-01
-6.01506293e-01 1.84415653e-01 9.08055484e-01 1.74510133e+00
-1.14488637e+00 -8.04723203e-01 -9.64081764e-01 8.88102353e-01
-6.61775112e-01 1.57638267e-01 -6.01561069e-02 9.26400840e-01
4.99509215e-01 1.40609992e+00 1.05314881e-01 -5.61413884e-01
6.61156297e-01 -2.40316227e-01 2.64792979e-01 -8.51470947e-01
-7.47013986e-01 -1.00425887e+00 4.86929804e-01 -4.08091675e-03
-2.86116153e-01 -7.46859729e-01 -1.86210465e+00 -2.51408607e-01
-7.72579983e-02 8.80539119e-01 7.85883486e-01 1.04570067e+00
4.81365547e-02 5.67220509e-01 8.07974935e-01 -5.31235397e-01
-7.33743846e-01 -1.26355565e+00 -1.01743650e+00 1.03602685e-01
8.81483778e-02 -1.21290755e+00 -6.00365214e-02 -2.41702959e-01] | [8.271103858947754, 4.776122570037842] |
dd2cae67-6153-4efe-9964-0b0d81530f6b | audio-dequantization-for-high-fidelity-audio | 2008.06867 | null | https://arxiv.org/abs/2008.06867v1 | https://arxiv.org/pdf/2008.06867v1.pdf | Audio Dequantization for High Fidelity Audio Generation in Flow-based Neural Vocoder | In recent works, a flow-based neural vocoder has shown significant improvement in real-time speech generation task. The sequence of invertible flow operations allows the model to convert samples from simple distribution to audio samples. However, training a continuous density model on discrete audio data can degrade model performance due to the topological difference between latent and actual distribution. To resolve this problem, we propose audio dequantization methods in flow-based neural vocoder for high fidelity audio generation. Data dequantization is a well-known method in image generation but has not yet been studied in the audio domain. For this reason, we implement various audio dequantization methods in flow-based neural vocoder and investigate the effect on the generated audio. We conduct various objective performance assessments and subjective evaluation to show that audio dequantization can improve audio generation quality. From our experiments, using audio dequantization produces waveform audio with better harmonic structure and fewer digital artifacts. | ['Seong-Whan Lee', 'Hyeong-Rae Noh', 'Sang-Hoon Lee', 'Hyun-Wook Yoon'] | 2020-08-16 | null | null | null | null | ['audio-generation', 'audio-dequantization'] | ['audio', 'audio'] | [ 6.62551150e-02 -1.70021847e-01 4.53524850e-02 -9.02123526e-02
-6.57377005e-01 -2.70360559e-01 4.64021087e-01 -1.71867549e-01
-9.53650451e-04 1.09984767e+00 6.69081450e-01 -2.17855535e-03
1.59396827e-01 -8.15255284e-01 -6.29539430e-01 -6.42181993e-01
-9.80724841e-02 1.84496827e-02 1.38039201e-01 -5.59198996e-03
1.17234178e-01 2.54112124e-01 -1.76472998e+00 3.40598732e-01
8.70438695e-01 9.39712286e-01 1.94561392e-01 1.03565896e+00
-1.36022538e-01 6.29428208e-01 -1.28511524e+00 -9.28351358e-02
2.32692495e-01 -8.65131915e-01 -3.54091614e-01 -1.76764458e-01
3.36999923e-01 -5.92461884e-01 -5.55052221e-01 1.24981928e+00
8.74863863e-01 2.95987308e-01 1.07540667e+00 -1.42449176e+00
-8.44509006e-01 8.08395326e-01 -3.40895146e-01 2.94387817e-01
2.50773996e-01 8.61324734e-05 7.61637807e-01 -7.50225186e-01
4.17123973e-01 1.51662886e+00 4.23792511e-01 6.68889105e-01
-1.38396990e+00 -1.02786684e+00 -4.92405176e-01 3.42300862e-01
-1.54125202e+00 -5.40805161e-01 1.05009425e+00 -3.43188971e-01
7.04801381e-01 8.84847715e-02 7.37510920e-01 1.00940514e+00
3.00787061e-01 7.01379061e-01 8.16434503e-01 -3.39087397e-01
4.02071595e-01 8.53893459e-02 -4.19697493e-01 2.36853570e-01
-5.84306605e-02 5.27379096e-01 -8.26545477e-01 3.71957719e-02
1.02373934e+00 -6.10426724e-01 -6.83543324e-01 1.18940838e-01
-8.03356111e-01 7.31770396e-01 1.72595859e-01 1.00283995e-01
-2.63701677e-01 5.58763266e-01 5.85988462e-01 4.77565914e-01
3.51033151e-01 3.64031911e-01 4.72205952e-02 -5.66867411e-01
-1.35160100e+00 4.39242005e-01 5.28105915e-01 6.29523933e-01
4.40059125e-01 1.00151217e+00 -1.72272325e-01 9.92249846e-01
3.20380211e-01 4.86698836e-01 8.53479803e-01 -1.05917609e+00
3.27477813e-01 -2.03322798e-01 -1.33192584e-01 -9.96328712e-01
2.86479950e-01 -4.32846069e-01 -1.15799749e+00 5.02470315e-01
2.62807757e-01 -3.74614745e-01 -8.25378120e-01 1.72395277e+00
4.66825217e-02 4.09479439e-01 1.24924570e-01 9.00770187e-01
6.16565883e-01 1.34076381e+00 -1.45342723e-01 -3.83436412e-01
8.79919112e-01 -6.65042222e-01 -1.33301687e+00 6.57946765e-01
1.29900903e-01 -1.01823437e+00 1.12429357e+00 8.11572373e-01
-1.13481963e+00 -9.49575663e-01 -1.35324562e+00 -6.85213413e-03
-6.74585477e-02 1.52979434e-01 1.90089777e-01 9.25762534e-01
-1.13026440e+00 6.66221082e-01 -5.70584774e-01 2.77648360e-01
2.05179632e-01 1.73939556e-01 -1.17933393e-01 5.90070128e-01
-1.48889673e+00 3.17147017e-01 4.71075445e-01 -2.70998597e-01
-1.35187697e+00 -8.91878784e-01 -7.39811599e-01 2.84535289e-01
-8.88835192e-02 -5.22542179e-01 1.39483464e+00 -9.05450940e-01
-1.89283228e+00 -6.88256919e-02 -4.58032526e-02 -8.08898747e-01
3.05103749e-01 -1.45472154e-01 -5.83603084e-01 2.99919784e-01
-2.75080562e-01 1.13491976e+00 1.27769160e+00 -9.90510166e-01
-6.39314055e-01 3.69152755e-01 -2.65836120e-01 2.68228538e-02
-4.39809501e-01 -3.87048572e-01 3.59124273e-01 -1.17149770e+00
-2.09453776e-01 -4.10018772e-01 1.36278167e-01 9.36182290e-02
-2.57874489e-01 -1.38334289e-01 1.01325119e+00 -6.35369658e-01
1.46369851e+00 -2.29418659e+00 -6.95259795e-02 -1.83131531e-01
-6.27836511e-02 3.12861770e-01 -5.36703579e-02 4.30779546e-01
-2.10323691e-01 3.60497296e-01 8.19916055e-02 -2.87469268e-01
-7.39210248e-02 -2.58290134e-02 -7.59325206e-01 3.97588387e-02
2.84372151e-01 3.32446218e-01 -6.12465620e-01 -6.01535976e-01
1.70078665e-01 8.83723378e-01 -1.02653933e+00 1.76775604e-01
-2.66828388e-01 2.39573792e-01 7.24604726e-02 3.36468399e-01
7.75149047e-01 5.00177324e-01 -4.12859559e-01 -3.91587377e-01
-5.76779619e-02 3.22730571e-01 -1.24399245e+00 1.49260545e+00
-5.15260279e-01 1.17143071e+00 2.68230420e-02 -7.67390013e-01
1.14834881e+00 8.33070934e-01 2.22982138e-01 -4.54432368e-01
1.46804452e-01 2.03181103e-01 1.40493035e-01 -3.67658287e-01
8.08745801e-01 -5.37595332e-01 4.22418684e-01 3.00770253e-01
3.44790339e-01 -5.70715964e-01 3.57780010e-01 3.08635994e-03
5.15605569e-01 -1.26464888e-01 4.67924692e-05 -1.93327859e-01
5.34133911e-01 -4.90883440e-01 4.67918187e-01 3.78665239e-01
-2.05706626e-01 8.10107350e-01 4.19430107e-01 1.20825797e-01
-1.28532541e+00 -1.38523769e+00 -4.28463668e-02 6.34283841e-01
-2.74383277e-01 -4.34103549e-01 -9.53963101e-01 1.54210553e-02
-3.64471823e-01 6.39211774e-01 -1.61158532e-01 -3.25056106e-01
-4.07439172e-01 -3.62254828e-01 9.57903087e-01 3.73686731e-01
6.16141617e-01 -1.10986662e+00 -1.77061588e-01 4.97542858e-01
-2.26461768e-01 -7.27782428e-01 -8.12558651e-01 -1.46506205e-01
-9.64045882e-01 -4.44591969e-01 -1.19771290e+00 -7.87564635e-01
2.04019383e-01 -1.75298110e-01 7.19546497e-01 -4.14834350e-01
-2.47301161e-01 -1.08250871e-01 -3.66937459e-01 -5.04789233e-01
-7.63543129e-01 -9.62965414e-02 3.41312617e-01 -1.38570210e-02
-2.33033255e-01 -1.02724302e+00 -5.00124156e-01 9.50742885e-02
-1.23033357e+00 -1.49135487e-02 2.06772119e-01 8.30330610e-01
4.31184173e-01 6.41414106e-01 1.09412670e+00 -2.86074448e-03
1.36518013e+00 -2.45099604e-01 -6.53286457e-01 -3.07810366e-01
-2.25011095e-01 4.48487699e-01 9.74442661e-01 -7.74358869e-01
-1.16215146e+00 -2.45036155e-01 -3.84684205e-01 -6.21201634e-01
5.05453125e-02 4.79117632e-01 -1.46434382e-01 3.48691672e-01
7.84785688e-01 3.54288995e-01 6.52193278e-02 -4.48200613e-01
2.46524528e-01 1.06018651e+00 7.43293941e-01 -5.72363257e-01
7.45260477e-01 1.63253009e-01 -2.74830274e-02 -1.17730534e+00
-9.19375792e-02 3.72333862e-02 5.26997484e-02 -3.22209597e-01
8.27744305e-01 -8.52632701e-01 -6.09815717e-01 4.83397394e-01
-1.32293189e+00 -2.93147802e-01 -4.27496940e-01 8.14164817e-01
-6.87629640e-01 2.29272515e-01 -6.93463445e-01 -1.08152556e+00
-4.28962737e-01 -1.23377049e+00 7.29415119e-01 3.70616466e-01
-1.28560096e-01 -8.09191227e-01 3.09938401e-01 -2.52232552e-01
5.07459700e-01 1.71759397e-01 9.01303232e-01 1.39485747e-01
-6.30112827e-01 2.18494251e-01 4.80529070e-02 8.16292048e-01
2.64914513e-01 3.28986853e-01 -1.04694891e+00 -4.18014117e-02
9.55375284e-02 -2.86524445e-01 7.91846156e-01 6.10734701e-01
1.17192137e+00 -3.21026683e-01 1.74023762e-01 4.56149697e-01
1.18952072e+00 6.47305071e-01 8.50781262e-01 -3.38641971e-01
1.75791606e-01 2.84364045e-01 3.33743840e-01 5.06013989e-01
-3.97968888e-01 5.26430964e-01 1.41117945e-01 7.23954663e-02
-6.96512341e-01 -6.59703612e-01 6.15909874e-01 1.14303327e+00
1.29452914e-01 -5.37984669e-01 -3.53217185e-01 6.33979857e-01
-1.25435793e+00 -1.11557364e+00 1.29517332e-01 2.07395983e+00
1.17715943e+00 1.94524661e-01 2.47099265e-01 8.13847423e-01
9.01044846e-01 -1.88734196e-02 -2.09480241e-01 -6.30822420e-01
8.49458873e-02 4.88378346e-01 1.43202186e-01 6.45246446e-01
-5.67659795e-01 5.74018776e-01 6.85982847e+00 1.36294937e+00
-1.49067283e+00 2.98216064e-02 4.21752602e-01 -2.11596996e-01
-3.81209850e-01 -3.49064678e-01 -5.92144847e-01 5.40349066e-01
9.99885142e-01 -6.91055954e-01 4.14614379e-01 6.86950386e-01
6.53324783e-01 5.13474755e-02 -1.00672162e+00 1.21396804e+00
-1.75178766e-01 -1.42021906e+00 6.13689244e-01 7.20518529e-02
7.25029528e-01 -5.64720452e-01 3.67027670e-01 2.49296010e-01
-1.06384866e-01 -1.26060450e+00 9.37063575e-01 4.21553850e-01
1.03204620e+00 -1.11553693e+00 4.44384038e-01 1.63632259e-01
-1.42968047e+00 1.37972504e-01 -5.07301688e-01 -1.28785729e-01
5.34663141e-01 6.76516950e-01 -1.14232767e+00 1.77296042e-01
2.68051118e-01 3.39572936e-01 -1.63848829e-02 1.36971247e+00
-3.86195518e-02 9.40962613e-01 -1.37555838e-01 -8.37753415e-02
6.05297461e-03 1.38326185e-02 7.41259217e-01 9.52619553e-01
7.06656218e-01 -1.67749807e-01 -3.06386918e-01 1.08637369e+00
-9.10767093e-02 8.78173951e-03 -6.60258353e-01 -4.33571190e-01
5.42832315e-01 6.53809011e-01 -4.53367531e-01 -2.67212510e-01
1.70887455e-01 6.15725577e-01 -4.34775263e-01 5.00641942e-01
-9.78686512e-01 -9.23137009e-01 7.56330490e-01 3.41841698e-01
7.30376877e-03 -3.73733848e-01 1.10242605e-01 -7.32707441e-01
-2.60244489e-01 -8.04946721e-01 -1.95533693e-01 -1.02465141e+00
-9.22466218e-01 6.20418370e-01 9.06623155e-02 -1.37261415e+00
-6.05406165e-01 -3.21505308e-01 -6.02701902e-01 8.20890069e-01
-1.15513051e+00 -3.29330176e-01 4.91181873e-02 4.31968212e-01
7.97378838e-01 -4.01326299e-01 6.86539471e-01 7.16219068e-01
-6.33265376e-02 7.79387474e-01 2.56460835e-03 5.81002794e-03
6.74998581e-01 -1.05439246e+00 2.73080677e-01 7.75680721e-01
1.82265118e-01 4.27366465e-01 1.00601721e+00 -5.80080926e-01
-8.41847360e-01 -1.08507276e+00 3.95575523e-01 2.78665394e-01
3.26437533e-01 -2.95006812e-01 -9.42465067e-01 5.49612790e-02
5.59614837e-01 -4.38025534e-01 5.09834468e-01 -5.36179662e-01
-6.50465069e-03 -3.26580256e-01 -1.08384943e+00 5.85748911e-01
6.25811636e-01 -6.26458108e-01 -4.68869299e-01 -2.64362186e-01
9.31681514e-01 -1.10852510e-01 -7.13487327e-01 2.95929790e-01
4.64859784e-01 -9.54953790e-01 8.31778288e-01 6.45920485e-02
7.12176502e-01 -5.89726388e-01 -9.96232107e-02 -1.58781528e+00
1.77512020e-02 -1.06852245e+00 -4.91448343e-02 1.49307954e+00
3.89401942e-01 -4.00338739e-01 7.32965410e-01 -3.45728874e-01
-1.14627749e-01 -3.42421412e-01 -8.79729688e-01 -1.02130949e+00
9.23026055e-02 -3.94978583e-01 6.65780544e-01 4.06179398e-01
7.51099959e-02 3.06596905e-01 -7.16656983e-01 -4.78849448e-02
4.28842574e-01 -4.27216381e-01 7.07233071e-01 -8.86399627e-01
-4.52389598e-01 -3.56460094e-01 -4.98897910e-01 -1.26131344e+00
-2.18489468e-02 -7.55297601e-01 2.20820874e-01 -1.30537653e+00
-3.11620116e-01 -1.14637427e-01 1.04114413e-01 -2.32508451e-01
2.09289134e-01 3.64766181e-01 2.63179004e-01 -2.06445679e-01
2.27077842e-01 1.02696180e+00 1.52188003e+00 -1.37226641e-01
-3.59357655e-01 -8.68984610e-02 -2.01528773e-01 3.99964035e-01
9.55550194e-01 -5.68726778e-01 -1.07494426e+00 -1.59374297e-01
4.66564782e-02 4.45946872e-01 1.13145925e-01 -1.59225237e+00
1.02592692e-01 -1.20613426e-02 2.56652564e-01 -7.87244916e-01
6.25671625e-01 -3.88925344e-01 3.46062183e-01 5.13368785e-01
-4.31707531e-01 2.16974458e-03 1.94215208e-01 4.60236549e-01
-7.54717350e-01 -3.84284645e-01 8.98734629e-01 1.61653221e-01
-2.01351926e-01 1.69460356e-01 -8.82085562e-01 1.74299940e-01
4.83088672e-01 -2.92625397e-01 1.89170223e-02 -9.67392504e-01
-6.02935433e-01 -4.15669262e-01 -1.69664733e-02 2.62299418e-01
8.13246489e-01 -1.63699186e+00 -6.81642711e-01 4.48216885e-01
-5.35268784e-01 -1.74292699e-01 1.39429510e-01 2.54365861e-01
-6.59650385e-01 4.39796776e-01 -4.71157044e-01 -5.47015429e-01
-1.27543974e+00 2.49692589e-01 4.70777094e-01 1.89853370e-01
-2.28190646e-01 7.89259255e-01 2.52093017e-01 1.68511420e-01
5.40205598e-01 -7.31543601e-01 -1.37058556e-01 5.37869968e-02
6.17947876e-01 5.30185878e-01 -1.96773008e-01 -2.88302511e-01
1.89521223e-01 3.29426587e-01 2.27916136e-01 -9.25761640e-01
8.02449167e-01 -1.65251121e-02 2.46015027e-01 4.36930954e-01
1.37009823e+00 2.33149342e-03 -1.16616666e+00 3.48484397e-01
-6.42639577e-01 -5.78418911e-01 2.15840638e-01 -3.43229949e-01
-1.17042708e+00 1.31154990e+00 7.26264954e-01 4.04562593e-01
1.27616882e+00 -5.93156040e-01 1.01733923e+00 1.18307054e-01
1.57530189e-01 -1.16866553e+00 6.37267768e-01 4.00564730e-01
1.08636487e+00 -4.40767884e-01 -3.97098899e-01 -2.46364549e-01
-3.16884726e-01 1.06814182e+00 5.64596653e-01 -2.60180950e-01
8.62260044e-01 5.54403424e-01 4.08925042e-02 4.03330237e-01
-7.83018708e-01 1.95155874e-01 5.49332052e-02 8.77196729e-01
5.56596279e-01 -8.23917091e-02 -3.71245146e-01 2.94582099e-01
-9.34141159e-01 2.61915743e-01 9.46400702e-01 4.27439451e-01
-6.47406995e-01 -1.11577129e+00 -6.52343571e-01 1.76341161e-01
-5.75660169e-01 -1.50235742e-01 -4.00160104e-02 4.34654236e-01
1.05020523e-01 1.04529083e+00 2.91151643e-01 -3.63893956e-01
5.83650135e-02 1.05059661e-01 3.76951993e-01 -3.18365812e-01
-1.83245048e-01 4.94825184e-01 -4.56717350e-02 -3.14958058e-02
-2.01120973e-01 -2.05443010e-01 -1.43973434e+00 -2.45237231e-01
-5.20096719e-01 5.46318948e-01 7.64193237e-01 4.35890555e-01
1.51008889e-01 9.76620138e-01 5.46688378e-01 -7.86831260e-01
-5.58116913e-01 -1.16874063e+00 -7.72994518e-01 2.03589752e-01
6.59215808e-01 -5.65216362e-01 -5.31805336e-01 4.80352610e-01] | [15.511391639709473, 5.9861602783203125] |
4d66af46-bc39-43d9-804d-43186b9b4a48 | learning-to-restore-3d-face-from-in-the-wild | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zhang_Learning_To_Restore_3D_Face_From_In-the-Wild_Degraded_Images_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zhang_Learning_To_Restore_3D_Face_From_In-the-Wild_Degraded_Images_CVPR_2022_paper.pdf | Learning To Restore 3D Face From In-the-Wild Degraded Images | In-the-wild 3D face modelling is a challenging problem as the predicted facial geometry and texture suffer from a lack of reliable clues or priors, when the input images are degraded. To address such a problem, in this paper we propose a novel Learning to Restore (L2R) 3D face framework for unsupervised high-quality face reconstruction from low-resolution images. Rather than directly refining 2D image appearance, L2R learns to recover fine-grained 3D details on the proxy against degradation via extracting generative facial priors. Concretely, L2R proposes a novel albedo restoration network to model high-quality 3D facial texture, in which the diverse guidance from the pre-trained Generative Adversarial Networks (GANs) is leveraged to complement the lack of input facial clues. With the finer details of the restored 3D texture, L2R then learns displacement maps from scratch to enhance the significant facial structure and geometry. Both of the procedures are mutually optimized with a novel 3D-aware adversarial loss, which further improves the modelling performance and suppresses the potential uncertainty. Extensive experiments on benchmarks show that L2R outperforms state-of-the-art methods under the condition of low-quality inputs, and obtains superior performances than 2D pre-processed modelling approaches with limited 3D proxy. | ['Zhifeng Xie', 'Dongjin Huang', 'Hao Tang', 'Chengjie Wang', 'Xiaoming Huang', 'Ying Tai', 'Yanhao Ge', 'Zhenyu Zhang'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['3d-face-modeling', 'face-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 3.67506355e-01 4.92407829e-01 1.41320467e-01 -3.49545777e-01
-1.00063682e+00 -1.53940514e-01 4.52068150e-01 -9.96363640e-01
3.40626448e-01 5.00380635e-01 3.14292908e-01 2.72590309e-01
-6.22267313e-02 -8.16541433e-01 -9.48503673e-01 -9.79988873e-01
3.75989944e-01 5.19744396e-01 -2.14058980e-01 -3.59232217e-01
-2.25628674e-01 9.43390548e-01 -1.74701202e+00 1.46099359e-01
7.82779276e-01 1.25544012e+00 -1.33424580e-01 2.83530384e-01
1.57515362e-01 5.91391623e-01 -2.21706584e-01 -5.10579407e-01
6.44909084e-01 -4.12281275e-01 -2.88494080e-01 5.60977101e-01
7.37568676e-01 -6.97156131e-01 -5.01738846e-01 1.02427876e+00
5.30563951e-01 -1.23399988e-01 7.50044465e-01 -9.13149774e-01
-7.33720362e-01 -3.20614725e-01 -8.62518072e-01 -4.22971755e-01
3.74488711e-01 4.14844692e-01 3.95731628e-01 -1.23254991e+00
5.79778552e-01 1.73211539e+00 6.90616667e-01 8.90873909e-01
-1.36258221e+00 -7.74542630e-01 8.66265818e-02 -3.46110970e-01
-1.48134005e+00 -7.84339845e-01 1.30418241e+00 -3.30088615e-01
3.32705766e-01 8.71210918e-02 3.44857991e-01 1.29454947e+00
3.65280323e-02 3.96034926e-01 1.36470211e+00 -2.39320591e-01
-2.64994968e-02 -2.43811220e-01 -8.48122895e-01 9.25294340e-01
-1.36391684e-01 5.66240370e-01 -6.97354019e-01 -1.47448331e-01
1.24489892e+00 1.09146565e-01 -4.25257266e-01 -2.05010504e-01
-5.35467029e-01 4.18719739e-01 3.74996394e-01 -3.05110514e-01
-5.07020891e-01 4.34826724e-02 -2.87471712e-01 6.47764206e-02
1.14882588e+00 7.64689175e-03 -4.06348199e-01 3.31752360e-01
-9.17509615e-01 2.44661704e-01 2.46480659e-01 7.29494691e-01
9.83360887e-01 4.21647340e-01 -9.32212770e-02 9.29655612e-01
7.46196687e-01 1.02057791e+00 -1.61619067e-01 -1.28123748e+00
2.79982388e-01 4.79467481e-01 1.01928584e-01 -1.06434083e+00
7.64392018e-02 -4.08298999e-01 -1.13194525e+00 6.20497286e-01
2.98372418e-01 1.59058243e-01 -1.37537849e+00 1.76364756e+00
7.53208935e-01 3.26533884e-01 -1.52582139e-01 1.04984486e+00
8.74663055e-01 5.98751485e-01 -3.13503504e-01 -2.68423170e-01
9.18245792e-01 -5.80692530e-01 -6.08672082e-01 -3.14699948e-01
-4.28775668e-01 -9.19663608e-01 9.30453122e-01 3.08531582e-01
-1.33538616e+00 -6.08709216e-01 -6.70359135e-01 -1.35364071e-01
3.97199541e-01 -7.10178688e-02 3.31876427e-01 6.51423991e-01
-1.08955634e+00 5.99343121e-01 -8.24038804e-01 1.60137817e-01
1.04096758e+00 1.78639024e-01 -5.78147233e-01 -6.09896541e-01
-8.64330113e-01 4.87568080e-01 -3.87614191e-01 5.32324433e-01
-1.38145399e+00 -1.05683649e+00 -9.76006150e-01 -2.41450474e-01
3.39343160e-01 -7.77121365e-01 7.33222187e-01 -9.48129356e-01
-1.72963297e+00 1.11317492e+00 -2.03641728e-01 3.10621530e-01
8.40138078e-01 -2.02135280e-01 -1.23519994e-01 3.04518849e-01
1.12080099e-02 6.49039567e-01 1.56593299e+00 -1.82461476e+00
3.02172136e-02 -8.32885265e-01 -2.19490543e-01 6.04124218e-02
8.60666409e-02 -1.85681179e-01 -6.89596236e-01 -9.72227633e-01
2.94701546e-01 -6.04840577e-01 -1.56836390e-01 5.00278592e-01
-2.51578480e-01 2.80047238e-01 7.57227302e-01 -1.02428627e+00
4.59549993e-01 -1.92679882e+00 2.88089365e-01 2.34597996e-01
1.90217420e-01 4.84310538e-02 -3.26961190e-01 -2.30685487e-01
-6.70506060e-02 1.43720686e-01 -5.28247118e-01 -7.14467227e-01
-9.41086337e-02 2.83881307e-01 -4.13069546e-01 5.96443415e-01
6.47077560e-01 8.66810739e-01 -5.72116137e-01 -3.54756206e-01
2.02373222e-01 1.26713789e+00 -6.78062916e-01 5.78719616e-01
-4.55955952e-01 1.02768409e+00 -5.58126330e-01 1.09217501e+00
1.21627057e+00 -8.05440471e-02 -3.92830335e-02 -3.97869170e-01
5.25685012e-01 -1.41445950e-01 -9.67971027e-01 1.66011596e+00
-5.36796093e-01 7.56867006e-02 4.25911486e-01 -6.72874987e-01
1.21649539e+00 2.54886508e-01 5.44788837e-01 -8.32955837e-01
-4.19686474e-02 1.24846816e-01 -5.67012131e-01 -3.66322756e-01
-8.76783282e-02 -5.25101125e-01 3.00237834e-01 2.13913590e-01
1.75621118e-02 -6.14585042e-01 -8.26368093e-01 -1.53631881e-01
8.51210773e-01 7.55438805e-01 -2.58169711e-01 9.59450081e-02
4.90542799e-01 -6.70729578e-01 8.86359155e-01 2.25834459e-01
1.29298255e-01 1.23600090e+00 3.30864847e-01 -3.47080380e-01
-1.11213863e+00 -1.23748863e+00 -1.33085191e-01 5.72501242e-01
-8.18056017e-02 -8.27409104e-02 -8.33399951e-01 -7.58262277e-01
3.91727351e-02 2.29533523e-01 -9.35550630e-01 -1.31585240e-01
-5.98767161e-01 -7.50124037e-01 3.44387650e-01 2.57747561e-01
5.68118393e-01 -8.40762258e-01 3.34938616e-03 -1.12068862e-01
-1.55624822e-01 -1.13562953e+00 -5.22703886e-01 -3.02926868e-01
-8.46989751e-01 -1.01400101e+00 -6.92293525e-01 -5.03744006e-01
1.04874730e+00 9.78598297e-02 1.31186318e+00 3.24663371e-01
-3.26378495e-01 4.23169911e-01 -1.07839979e-01 -1.07944153e-01
-4.83961314e-01 -6.59112930e-01 2.62840781e-02 5.27636290e-01
-1.95235714e-01 -1.08137763e+00 -6.48633122e-01 5.20294070e-01
-9.62033391e-01 1.60118103e-01 5.34904003e-01 1.06123447e+00
9.99887705e-01 2.83216238e-01 3.28469068e-01 -7.85115480e-01
-1.20346412e-01 -3.02760065e-01 -6.05402827e-01 1.26870915e-01
-5.26736736e-01 1.13433778e-01 4.69643056e-01 -3.97321612e-01
-1.60136640e+00 1.74255550e-01 -4.33797807e-01 -9.68224823e-01
-2.05802694e-01 -1.01974629e-01 -1.00690866e+00 -3.28929126e-01
4.86983508e-01 2.85802662e-01 2.61327833e-01 -8.20231557e-01
1.94532171e-01 2.51794189e-01 5.63248456e-01 -9.00199056e-01
1.34488475e+00 8.12299967e-01 4.48101521e-01 -6.02368057e-01
-1.13819337e+00 2.19938219e-01 -5.51054537e-01 -3.74076307e-01
5.96230924e-01 -1.16933012e+00 -5.33729672e-01 8.06273222e-01
-1.03374588e+00 -5.77667356e-01 -2.97165453e-01 1.45387258e-02
-7.23396480e-01 3.03599060e-01 -5.83763778e-01 -8.80341172e-01
-3.04558814e-01 -1.22184145e+00 1.56105268e+00 1.94858342e-01
4.86804336e-01 -6.56333327e-01 -3.00703496e-01 9.02755082e-01
2.14747399e-01 6.77921355e-01 8.11033487e-01 3.43257099e-01
-8.65304053e-01 8.55403394e-02 -1.50794923e-01 7.68341660e-01
3.84549975e-01 -8.11203867e-02 -1.35016298e+00 -1.70998946e-01
2.64679879e-01 -4.05298531e-01 7.66545177e-01 3.65033954e-01
1.41673148e+00 -3.77340674e-01 8.87661427e-02 1.05936038e+00
1.22626674e+00 -3.51456374e-01 9.42918062e-01 -2.46026367e-01
1.07179284e+00 8.54663610e-01 6.54329777e-01 5.48080027e-01
2.80359715e-01 7.57467449e-01 8.24117780e-01 -1.41167015e-01
-6.17434561e-01 -6.43948495e-01 4.40605223e-01 4.12239671e-01
-4.28302675e-01 8.43442306e-02 -5.94341278e-01 1.96098387e-01
-1.46253908e+00 -7.07722723e-01 3.18262905e-01 2.15151763e+00
1.13163877e+00 -1.37092531e-01 -2.94285685e-01 -1.25091910e-01
5.53460062e-01 3.98096085e-01 -8.16346169e-01 2.92415410e-01
-4.20137733e-01 4.84182775e-01 1.02428876e-01 6.94626868e-01
-6.55107141e-01 1.04445958e+00 5.38162470e+00 1.15974021e+00
-1.05072689e+00 1.26113161e-01 1.20999265e+00 -2.05876395e-01
-7.24195957e-01 -1.63133740e-01 -7.90912211e-01 4.08836842e-01
4.29801881e-01 6.09990954e-01 7.32588649e-01 5.30947268e-01
3.03558081e-01 2.34306023e-01 -7.85616994e-01 1.10905075e+00
2.53795475e-01 -1.21437263e+00 2.12722480e-01 3.25051308e-01
1.07996261e+00 -3.36658180e-01 3.99718910e-01 -9.80232880e-02
3.19870383e-01 -1.54023564e+00 8.31007779e-01 1.10641778e+00
1.31788695e+00 -9.93695855e-01 3.34255278e-01 1.88038677e-01
-8.84556651e-01 2.11864784e-01 -3.69456857e-01 3.03493828e-01
1.35628521e-01 8.82062316e-01 -2.31663093e-01 6.65657580e-01
9.22577679e-01 7.04197288e-01 -4.05929595e-01 3.37716013e-01
-7.14205801e-01 5.82720757e-01 -3.39809567e-01 9.71279681e-01
-2.63956666e-01 -2.00319320e-01 6.00828588e-01 4.48564053e-01
3.79531771e-01 3.57309043e-01 2.56436132e-02 1.14355171e+00
-4.75330293e-01 -2.78175861e-01 -4.17715311e-01 3.05282921e-01
2.95708656e-01 1.19917715e+00 -2.72182226e-01 1.00441776e-01
-1.93918630e-01 9.31974947e-01 2.06673890e-01 4.72180665e-01
-6.67756379e-01 5.06040215e-01 1.01375163e+00 7.03737259e-01
2.77301818e-01 -1.57504659e-02 -1.75155580e-01 -1.08624387e+00
2.60946482e-01 -1.12215316e+00 -3.76252569e-02 -1.11464787e+00
-1.57843876e+00 6.60343885e-01 -2.61251539e-01 -1.02125680e+00
-2.18697473e-01 -5.97759426e-01 -3.69682312e-01 1.12650180e+00
-1.94911671e+00 -1.61850321e+00 -4.71551836e-01 7.63322592e-01
3.88764977e-01 -1.32228836e-01 6.36718512e-01 1.15201823e-01
-4.78610933e-01 6.18856311e-01 -1.23406738e-01 -6.77171256e-03
8.82881045e-01 -8.83493304e-01 2.35772625e-01 8.24963927e-01
-1.79685786e-01 2.54201561e-01 4.94006872e-01 -8.43631089e-01
-1.63032913e+00 -1.38898432e+00 3.33580136e-01 -4.93838072e-01
1.16676446e-02 -3.04837972e-01 -1.00717151e+00 4.35370654e-01
-2.24871367e-01 6.99534178e-01 3.87623727e-01 -3.93120736e-01
-6.72965527e-01 -3.84653747e-01 -1.65029681e+00 4.54134107e-01
1.46384263e+00 -6.82411611e-01 -2.86407322e-01 1.96794152e-01
6.24458551e-01 -6.15523517e-01 -1.15888715e+00 7.57888019e-01
4.58554327e-01 -1.07637548e+00 1.30496109e+00 -3.02563071e-01
6.99180722e-01 -3.89480919e-01 -4.24937874e-01 -1.19230235e+00
7.72339106e-02 -1.02261055e+00 -2.55598009e-01 1.40981746e+00
-1.28253222e-01 -4.10673618e-01 9.23073769e-01 6.33684993e-01
-1.25688642e-01 -8.60865414e-01 -8.98763180e-01 -4.88120794e-01
8.22621807e-02 -4.05601591e-01 9.82735038e-01 7.54177153e-01
-1.13347411e+00 -6.14377260e-02 -6.80106521e-01 4.48964417e-01
1.13950956e+00 2.32333317e-01 8.59877765e-01 -1.25616550e+00
-2.44902655e-01 -4.90539894e-02 -2.08986446e-01 -1.13080609e+00
5.55554092e-01 -6.50264978e-01 1.34782389e-01 -1.01210260e+00
2.09727101e-02 -5.01435459e-01 1.85129255e-01 5.12767673e-01
-1.87386557e-01 7.17179298e-01 -9.73918512e-02 1.26467690e-01
-1.05675481e-01 1.08435810e+00 1.76070738e+00 -1.43307716e-01
5.36473207e-02 -9.05297846e-02 -8.64110351e-01 9.18967187e-01
2.36153364e-01 -4.14763212e-01 -3.45603913e-01 -4.97173488e-01
1.27026409e-01 1.01782896e-01 7.97000945e-01 -6.77537918e-01
-2.40962207e-01 -3.69233787e-01 1.05661190e+00 -4.11429673e-01
7.29391873e-01 -8.47566307e-01 3.59736979e-01 -1.97721422e-01
1.25130638e-01 -6.26035571e-01 1.58554524e-01 6.15566611e-01
-1.36402994e-01 3.94340128e-01 1.21766162e+00 -1.44759908e-01
-2.45310858e-01 9.58977878e-01 3.67044657e-01 1.41300708e-01
4.46865648e-01 -2.37018272e-01 -3.98099311e-02 -4.78381157e-01
-7.15031207e-01 -1.80292696e-01 9.62007642e-01 3.79958093e-01
8.98036182e-01 -1.50331020e+00 -8.78054619e-01 6.39624298e-01
-1.14270568e-01 6.74132645e-01 6.35025501e-01 3.90082896e-01
-2.67723709e-01 -1.99292600e-01 -1.95612893e-01 -6.44717336e-01
-1.06244624e+00 3.19741964e-01 6.83053017e-01 -4.79425453e-02
-7.49441326e-01 9.34118271e-01 6.56411707e-01 -7.04371750e-01
8.09741840e-02 1.64353922e-01 1.90226421e-01 -2.36460641e-01
5.73004901e-01 4.46596928e-02 1.01663381e-01 -1.04346788e+00
-2.16419727e-01 1.11464262e+00 1.89199194e-01 5.65176941e-02
1.83657241e+00 -3.04661751e-01 -2.04575151e-01 -1.51841298e-01
1.01840019e+00 2.48833492e-01 -2.21285820e+00 -4.27362174e-01
-6.26175523e-01 -9.93951619e-01 2.99170434e-01 -7.97660470e-01
-1.76716185e+00 8.56664300e-01 5.25907576e-01 -5.82932830e-01
1.56153095e+00 1.30674355e-02 6.75476313e-01 -2.65951663e-01
4.39021111e-01 -6.20448470e-01 5.02703249e-01 1.89955130e-01
1.25081789e+00 -1.15685666e+00 1.29540518e-01 -6.65832698e-01
-3.93210232e-01 1.00170350e+00 7.28484154e-01 -1.69177696e-01
8.77884328e-01 2.59457767e-01 -1.18214125e-02 -2.87307918e-01
-5.19721448e-01 -3.04069407e-02 5.17810643e-01 8.99661541e-01
1.23583591e-02 -2.92343706e-01 5.62067449e-01 8.27691734e-01
-7.41809979e-02 -1.78873807e-01 1.48243785e-01 4.64679718e-01
8.85002092e-02 -1.06429648e+00 -8.09398115e-01 1.23858951e-01
-3.57674211e-01 -3.78327221e-02 -2.43386745e-01 5.31260729e-01
3.12899590e-01 6.99634850e-01 -1.31429374e-01 -2.52916276e-01
2.66050607e-01 -1.41446352e-01 8.81332159e-01 -4.16061461e-01
-4.01939936e-02 4.77203757e-01 -2.56837487e-01 -9.11884487e-01
-3.91483814e-01 -5.16682982e-01 -8.94232333e-01 -3.95764768e-01
2.95090880e-02 -3.26732934e-01 4.05948341e-01 8.08096230e-01
5.57084382e-01 3.59728962e-01 1.06675589e+00 -1.26911497e+00
-3.59410048e-01 -8.18291366e-01 -6.82335675e-01 4.15123701e-01
5.74762225e-01 -1.02050376e+00 -5.72514534e-01 1.21796526e-01] | [12.794160842895508, -0.1456453949213028] |
d90dbc03-8963-4639-94d5-ee94edb6d08b | adversarial-sparse-transformer-for-time | null | null | http://proceedings.neurips.cc/paper/2020/hash/c6b8c8d762da15fa8dbbdfb6baf9e260-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/c6b8c8d762da15fa8dbbdfb6baf9e260-Paper.pdf | Adversarial Sparse Transformer for Time Series Forecasting | Many approaches have been proposed for time series forecasting, in light of its significance in wide applications including business demand prediction.
However, the existing methods suffer from two key limitations. Firstly, most point prediction models only predict an exact value of each time step without flexibility, which can hardly capture the stochasticity of data. Even probabilistic prediction using the likelihood estimation suffers these problems in the same way. Besides, most of them use the auto-regressive generative mode, where ground-truth is provided during training and replaced by the network’s own one-step ahead output during inference, causing the error accumulation in inference. Thus they may fail to forecast time series for long time horizon due to the error accumulation. To solve these issues, in this paper, we propose a new time series forecasting model -- Adversarial Sparse Transformer (AST), based on Generated Adversarial Networks (GANs). Specifically, AST adopts a Sparse Transformer as the generator to learn a sparse attention map for time series forecasting, and uses a discriminator to improve the prediction performance from sequence level. Extensive experiments on several real-world datasets show the effectiveness and efficiency of our method. | ['Junzhou Huang', 'Ying WEI', 'Peilin Zhao', 'Qianggang Ding', 'Xi Xiao', 'Sifan Wu'] | 2020-12-01 | null | null | null | neurips-2020-12 | ['probabilistic-time-series-forecasting'] | ['time-series'] | [ 7.55509436e-02 -1.31462410e-01 -8.46644938e-02 -4.26591486e-01
-5.42021334e-01 -2.98101127e-01 5.51825762e-01 -5.13912022e-01
3.22255492e-01 7.40863025e-01 2.34270200e-01 -2.16313660e-01
4.68726158e-02 -9.99917626e-01 -7.63025165e-01 -7.97944665e-01
1.70339286e-01 2.66435742e-01 3.47821377e-02 -2.30962917e-01
-1.48156984e-02 -3.25324461e-02 -1.17008388e+00 9.83964354e-02
1.17334032e+00 1.40520072e+00 2.70006299e-01 1.09962121e-01
-3.70575488e-01 1.11438203e+00 -7.33145893e-01 -5.29370308e-01
1.78283527e-01 -5.63244045e-01 -2.13506725e-03 -2.08033308e-01
-4.80737299e-01 -5.75948656e-01 -6.85623050e-01 9.68977511e-01
3.45160186e-01 1.13378055e-01 5.73624969e-01 -1.64054072e+00
-8.00663233e-01 7.96805859e-01 -3.98781270e-01 1.40011728e-01
-4.87541743e-02 1.76654682e-01 5.76389849e-01 -5.96987903e-01
-1.16369370e-02 1.08507764e+00 8.47897053e-01 5.09473264e-01
-9.12773073e-01 -9.70662296e-01 3.30938160e-01 2.68188417e-01
-1.17535436e+00 -2.15157524e-01 1.33587515e+00 -4.40071166e-01
6.93560362e-01 1.29281968e-01 5.92060804e-01 1.27269578e+00
4.23505396e-01 8.04443717e-01 1.07730329e+00 1.13925152e-01
2.67716110e-01 -1.11531928e-01 -4.42631543e-01 1.40385658e-01
-3.81062299e-01 2.69270688e-01 -1.00639373e-01 -1.07385457e-01
8.36203873e-01 7.39727497e-01 -1.87812224e-01 2.26717785e-01
-1.25533628e+00 7.59284079e-01 3.77672940e-01 2.65787899e-01
-7.00747609e-01 9.55292955e-02 5.51673710e-01 3.64266843e-01
8.51668656e-01 -8.60260800e-02 -5.01164317e-01 -3.19465071e-01
-9.96476531e-01 8.67581069e-02 6.47492826e-01 8.96419942e-01
3.76285344e-01 7.74954200e-01 -2.43298858e-01 6.06540442e-01
1.80407181e-01 6.07256413e-01 1.04096723e+00 -4.76282984e-01
7.27631390e-01 3.63552809e-01 7.49253109e-02 -1.46871579e+00
-1.04553275e-01 -7.41442263e-01 -1.37035549e+00 -2.21046954e-01
3.00956257e-02 -4.84027982e-01 -9.37209010e-01 1.81342340e+00
1.97679643e-02 8.51887524e-01 8.66663530e-02 7.73947597e-01
5.62615097e-01 1.38469899e+00 8.03526714e-02 -5.45480609e-01
7.14709401e-01 -9.36965048e-01 -8.92506540e-01 -5.79985082e-02
3.00045133e-01 -5.44457793e-01 8.39037776e-01 3.64890754e-01
-8.49223197e-01 -7.36909747e-01 -6.82928264e-01 3.99709165e-01
-1.21987738e-01 1.38782859e-01 5.03405750e-01 4.35943723e-01
-5.94802439e-01 5.38231015e-01 -9.63549912e-01 3.14507067e-01
2.59712249e-01 5.46245985e-02 1.77597567e-01 1.40232176e-01
-1.53201962e+00 6.05669498e-01 3.90628546e-01 4.24328417e-01
-8.73013973e-01 -7.71201313e-01 -6.20586276e-01 3.13101202e-01
1.52880251e-01 -4.64293212e-01 1.29301345e+00 -1.24069822e+00
-1.77188265e+00 -1.08187445e-01 -3.19340467e-01 -6.75579011e-01
6.52898073e-01 -4.61455174e-02 -9.98019159e-01 -5.20058334e-01
-9.94664878e-02 1.14501774e-01 9.93743241e-01 -7.99754620e-01
-4.66002882e-01 -2.35502139e-01 -1.98505864e-01 -1.07669272e-01
-2.84465998e-01 -2.54208058e-01 4.35498729e-02 -1.19351745e+00
9.78745967e-02 -6.81764364e-01 -4.58230644e-01 -6.58779800e-01
-3.60192478e-01 -3.14731330e-01 1.07367516e+00 -1.10042298e+00
1.54955161e+00 -2.15077138e+00 -1.05695464e-01 1.39105856e-01
4.95373309e-02 9.78584737e-02 5.95319197e-02 6.76240742e-01
-1.44419536e-01 2.12701447e-02 -2.10550562e-01 -1.46866143e-01
-6.57614740e-03 3.32360774e-01 -1.26798213e+00 1.37369484e-01
1.38228804e-01 1.11571741e+00 -8.43289673e-01 -1.97207168e-01
3.54504257e-01 6.08652234e-01 -1.62285775e-01 3.45575660e-01
-5.59126079e-01 7.23267317e-01 -7.07624555e-01 3.32477897e-01
6.63371444e-01 -3.54863346e-01 -1.34601876e-01 -1.13052316e-01
-7.24752024e-02 2.52415866e-01 -9.07268524e-01 1.37500978e+00
-6.18100345e-01 2.55720973e-01 -7.29419231e-01 -1.30321455e+00
1.25517392e+00 6.26358569e-01 5.83744287e-01 -6.70052648e-01
-5.42458668e-02 3.70495766e-01 -8.40113610e-02 -3.02615792e-01
2.85869896e-01 -3.22950751e-01 1.11831054e-01 3.37288797e-01
-3.18256766e-01 2.97307633e-02 -4.17214036e-01 -1.31880999e-01
7.51242995e-01 2.17260391e-01 -6.21712282e-02 3.92491311e-01
4.79389727e-01 -2.28538543e-01 1.05098999e+00 2.36003831e-01
2.25123107e-01 6.36610448e-01 5.19686222e-01 -6.99174941e-01
-1.19036078e+00 -6.99631572e-01 1.67021483e-01 6.75486982e-01
-1.06409654e-01 -1.68017238e-01 -4.78625476e-01 -6.32620454e-01
-2.08783060e-01 1.10615849e+00 -5.25372267e-01 -3.03401947e-01
-5.66449285e-01 -7.62827098e-01 2.00831458e-01 6.71699405e-01
4.99286145e-01 -1.24280393e+00 -2.07592010e-01 6.19887114e-01
-2.98138708e-01 -8.34462702e-01 -4.86554056e-01 -2.75689095e-01
-9.97817993e-01 -3.92798007e-01 -1.03123307e+00 -4.49429065e-01
5.10063827e-01 1.37417957e-01 9.30051625e-01 -3.10887158e-01
4.50636029e-01 -1.48659930e-01 -3.35220248e-01 -5.49916565e-01
-4.45810676e-01 -3.24315988e-02 -1.48520902e-01 3.02004993e-01
3.14782113e-01 -9.56115782e-01 -7.22763717e-01 4.57679868e-01
-8.43983114e-01 2.43655026e-01 3.71462584e-01 1.06065488e+00
7.14881539e-01 4.68219221e-01 1.14727330e+00 -7.45622575e-01
6.13443017e-01 -1.07885051e+00 -6.78584814e-01 2.61907339e-01
-7.95355201e-01 -2.21616477e-01 1.33727789e+00 -6.66055262e-01
-1.01336884e+00 -1.63683668e-01 -2.13851199e-01 -9.40505385e-01
2.02648059e-01 9.08440411e-01 -9.01295170e-02 4.20645177e-01
8.16429332e-02 9.91733730e-01 -4.62145284e-02 -4.85362411e-01
1.18884537e-02 5.27387083e-01 4.37383920e-01 -2.86985159e-01
7.48140454e-01 2.10690796e-01 -8.12553391e-02 -6.47286847e-02
-8.67694855e-01 1.45632088e-01 3.27342004e-02 -1.69235885e-01
3.98598224e-01 -1.07582438e+00 -5.92485487e-01 6.89206362e-01
-1.12569904e+00 -2.24726051e-01 -3.78676713e-01 6.06618166e-01
-6.03925109e-01 1.36138111e-01 -4.70856488e-01 -1.10542822e+00
-3.92823249e-01 -9.34417307e-01 8.33429396e-01 3.42155069e-01
2.35102177e-01 -1.05139422e+00 -5.68419918e-02 -9.59027112e-02
7.67708421e-01 4.83463854e-01 8.15384448e-01 -6.35175884e-01
-5.36239564e-01 -4.94941205e-01 -5.53257726e-02 5.31020939e-01
1.17288835e-01 -9.34163854e-02 -7.92490065e-01 -9.97649357e-02
4.35151279e-01 -2.39217319e-02 5.02622545e-01 4.32943672e-01
1.78950846e+00 -7.35689580e-01 -2.28486881e-01 6.56384051e-01
1.31496644e+00 7.55537748e-01 8.29735577e-01 -2.20514368e-02
7.78326392e-01 2.45639861e-01 5.64128399e-01 7.39946306e-01
4.92983818e-01 4.37266439e-01 4.61375922e-01 1.91897988e-01
3.84649068e-01 -9.15040910e-01 4.18471754e-01 1.42398357e+00
-2.00862125e-01 -5.18086195e-01 -8.12649846e-01 3.69073749e-01
-1.97433102e+00 -1.40486658e+00 5.06912582e-02 2.20768046e+00
6.55321479e-01 1.95836827e-01 7.07592890e-02 2.42063314e-01
6.00056052e-01 3.23516816e-01 -1.01394367e+00 9.43823531e-02
-9.68145281e-02 -1.00513652e-01 3.40763927e-01 -2.82221995e-02
-6.32034838e-01 5.27433217e-01 5.61704206e+00 1.13318765e+00
-1.57573700e+00 2.18161747e-01 1.19100404e+00 1.93185166e-01
-7.15817928e-01 -7.80083984e-02 -5.37646770e-01 1.44160616e+00
1.11037648e+00 -4.85744476e-01 6.30764425e-01 1.03506064e+00
3.80103350e-01 6.45124972e-01 -6.71453834e-01 1.04651058e+00
-8.41441080e-02 -1.14739311e+00 4.51074354e-02 -2.21125588e-01
9.14539099e-01 1.17090009e-02 3.83179456e-01 6.41476274e-01
2.58094072e-01 -1.12829411e+00 7.02629864e-01 9.71547425e-01
7.58376420e-01 -9.41197217e-01 8.70981812e-01 7.97252536e-01
-1.19107294e+00 -2.19798580e-01 -3.50304186e-01 -1.76763877e-01
4.22251672e-01 9.08745885e-01 -6.53948247e-01 6.41952395e-01
2.93081999e-01 9.56768453e-01 1.89903509e-02 8.16900730e-01
-7.03150630e-02 1.14442098e+00 -3.95392597e-01 -1.53497402e-02
2.03142822e-01 -3.99660349e-01 4.30767983e-01 6.35671854e-01
9.28424835e-01 2.25641713e-01 3.09630036e-01 9.08906698e-01
7.02491775e-02 5.18349931e-02 -3.88566911e-01 -2.77355790e-01
5.91704547e-01 8.07190478e-01 -3.31286341e-01 -3.38190854e-01
-5.34303665e-01 8.39942515e-01 -1.05521090e-01 5.76757491e-01
-1.21368837e+00 -4.03712362e-01 1.66890681e-01 -5.25760315e-02
3.78680050e-01 -1.89797834e-01 -2.44175777e-01 -1.41781223e+00
3.63655329e-01 -9.23192799e-01 1.83456466e-01 -9.09404695e-01
-1.53615499e+00 7.41390646e-01 -3.49650979e-01 -1.66687500e+00
-6.79025292e-01 -1.41053053e-03 -8.69879305e-01 1.17691731e+00
-1.44753563e+00 -1.13258624e+00 -1.75077975e-01 5.58663189e-01
7.95549154e-01 -3.68474692e-01 6.35263562e-01 3.02264303e-01
-6.05487347e-01 4.89386469e-01 5.43496311e-01 2.50873491e-02
1.92344502e-01 -9.62610841e-01 7.00941801e-01 7.53547490e-01
-7.29980469e-02 3.31701666e-01 6.09697342e-01 -6.22571826e-01
-1.40037394e+00 -1.38471878e+00 8.12523663e-01 -1.96132034e-01
7.33613253e-01 -7.95654729e-02 -1.17209804e+00 8.33430529e-01
-2.69345120e-02 6.41353875e-02 3.36208582e-01 -2.22306639e-01
-2.20460296e-01 -4.22989339e-01 -1.04953289e+00 3.51959616e-01
7.14056730e-01 -4.22292918e-01 -4.96252060e-01 4.25089628e-01
1.07868695e+00 -4.81957465e-01 -9.08449769e-01 4.59889919e-01
4.01279569e-01 -8.32255900e-01 7.61514723e-01 -3.59825134e-01
8.87550175e-01 -2.27237374e-01 -9.51751769e-02 -1.53669977e+00
-3.84473592e-01 -7.60779023e-01 -5.94135702e-01 1.52522659e+00
1.81516901e-01 -1.16778660e+00 5.78372061e-01 5.18125832e-01
-2.11175352e-01 -9.23005462e-01 -9.36699033e-01 -8.80843699e-01
-1.54391220e-02 -5.37701070e-01 1.46904933e+00 1.12213850e+00
-3.21921140e-01 1.08776852e-01 -8.79482687e-01 4.53814492e-02
2.83663303e-01 4.46040869e-01 7.20532358e-01 -9.99139190e-01
-3.26236904e-01 -3.69859397e-01 -3.67260188e-01 -1.19682813e+00
1.60220116e-01 -4.65513051e-01 -7.86360875e-02 -1.18475604e+00
-1.73615798e-01 -6.18978977e-01 -5.82194149e-01 2.55850852e-01
-1.67307332e-01 -2.75044173e-01 3.29357311e-02 4.98758525e-01
-1.74780518e-01 9.43427444e-01 1.28605700e+00 -1.46177933e-02
-4.05590050e-02 3.96966398e-01 -4.81653333e-01 5.40097475e-01
8.86160910e-01 -3.60221654e-01 -8.15134525e-01 -4.56520587e-01
1.87320679e-01 6.86699688e-01 4.74369824e-01 -1.00759280e+00
1.20435640e-01 -5.30690074e-01 3.96096051e-01 -9.14919078e-01
2.89644063e-01 -1.02220714e+00 7.47754693e-01 3.83525640e-01
-2.16490492e-01 3.43903154e-01 -1.07483290e-01 8.79431903e-01
-5.78285754e-01 1.38763443e-01 2.32712358e-01 3.20939510e-03
-6.15468979e-01 8.19669247e-01 1.32303655e-01 -2.99220048e-02
9.06188548e-01 5.68096712e-02 -1.31916165e-01 -7.28029013e-01
-3.82555068e-01 2.36891657e-01 1.18699722e-01 4.83944863e-01
4.10896420e-01 -1.64397407e+00 -7.72558689e-01 3.42407107e-01
-2.31702238e-01 8.94694328e-02 6.97759390e-01 7.72310317e-01
-6.45992607e-02 3.77312720e-01 7.29861259e-02 -4.49834138e-01
-2.24292874e-01 9.20675099e-01 2.90017158e-01 -3.74976575e-01
-7.07479298e-01 4.01959330e-01 2.67418325e-01 -2.40291536e-01
-2.49030106e-02 -3.40554208e-01 -1.83240533e-01 -2.27141738e-01
5.33922613e-01 2.62439311e-01 -1.37060240e-01 -5.52169204e-01
1.11432269e-01 3.47377896e-01 7.13131130e-02 3.08314357e-02
1.49523008e+00 -1.43558189e-01 9.46382210e-02 9.18003201e-01
1.03646123e+00 -2.04925537e-01 -1.37565827e+00 -4.32096153e-01
-3.35839808e-01 -5.21275163e-01 1.12028839e-02 -7.34705150e-01
-1.47671485e+00 9.22556341e-01 2.75460750e-01 7.53885448e-01
1.37094200e+00 -5.68490386e-01 1.44023204e+00 -1.68190440e-04
4.15229350e-01 -6.68590605e-01 -5.09135187e-01 3.77273172e-01
9.04787660e-01 -1.08189261e+00 -4.70490813e-01 -1.01047270e-01
-8.10446620e-01 1.08197665e+00 4.50928450e-01 -2.91829318e-01
7.56956041e-01 7.32482895e-02 -1.38015851e-01 3.36423546e-01
-9.95717406e-01 3.98291618e-01 2.46293485e-01 4.24450517e-01
3.56827825e-01 2.40288168e-01 -2.09043816e-01 1.11100423e+00
-3.51544738e-01 4.38212961e-01 1.23117276e-01 5.18978357e-01
3.79896723e-02 -8.44789147e-01 -1.51549667e-01 6.86131597e-01
-5.16433060e-01 -2.33521741e-02 3.81556749e-01 2.78160751e-01
-8.49797949e-02 8.15036893e-01 1.05115004e-01 -6.43905520e-01
1.63447872e-01 1.33379951e-01 -1.98436737e-01 -6.67524561e-02
-3.31499606e-01 1.19757168e-01 -4.55393255e-01 -3.36283714e-01
-2.42687821e-01 -7.66943991e-01 -9.94365692e-01 -4.82602119e-01
-2.68378139e-01 3.23466659e-01 5.22419512e-01 9.23328936e-01
5.99379063e-01 6.78893626e-01 1.29447567e+00 -5.38743794e-01
-9.91417646e-01 -1.05461586e+00 -5.62080562e-01 3.29364240e-01
3.50142986e-01 -4.75949585e-01 -3.61046404e-01 1.70994565e-01] | [6.925973892211914, 3.078063726425171] |
462939fa-dddf-4937-beef-25f47f004091 | in-out-diverse-image-outpainting-via-gan | 2104.00675 | null | https://arxiv.org/abs/2104.00675v1 | https://arxiv.org/pdf/2104.00675v1.pdf | In&Out : Diverse Image Outpainting via GAN Inversion | Image outpainting seeks for a semantically consistent extension of the input image beyond its available content. Compared to inpainting -- filling in missing pixels in a way coherent with the neighboring pixels -- outpainting can be achieved in more diverse ways since the problem is less constrained by the surrounding pixels. Existing image outpainting methods pose the problem as a conditional image-to-image translation task, often generating repetitive structures and textures by replicating the content available in the input image. In this work, we formulate the problem from the perspective of inverting generative adversarial networks. Our generator renders micro-patches conditioned on their joint latent code as well as their individual positions in the image. To outpaint an image, we seek for multiple latent codes not only recovering available patches but also synthesizing diverse outpainting by patch-based generation. This leads to richer structure and content in the outpainted regions. Furthermore, our formulation allows for outpainting conditioned on the categorical input, thereby enabling flexible user controls. Extensive experimental results demonstrate the proposed method performs favorably against existing in- and outpainting methods, featuring higher visual quality and diversity. | ['Ming-Hsuan Yang', 'Sergey Tulyakov', 'Jian Ren', 'Hsin-Ying Lee', 'Chieh Hubert Lin', 'Yen-Chi Cheng'] | 2021-04-01 | null | null | null | null | ['image-outpainting'] | ['computer-vision'] | [ 8.55867922e-01 4.25379544e-01 -1.76588863e-01 4.07695770e-02
-8.23564231e-01 -6.77388608e-01 3.36737752e-01 -4.55997080e-01
2.67533455e-02 1.13370919e+00 1.51733339e-01 1.55215889e-01
3.02196771e-01 -9.35433745e-01 -1.07448137e+00 -8.92416716e-01
4.95242864e-01 1.80493221e-01 -2.35097244e-01 -1.73489109e-01
1.58899620e-01 5.66300750e-01 -1.35070276e+00 4.01652247e-01
1.09190631e+00 5.76015413e-01 5.12717545e-01 5.75675488e-01
-9.03590024e-02 8.91116679e-01 -6.95332348e-01 -2.61415690e-01
6.97181702e-01 -1.02214611e+00 -3.42299312e-01 6.94876969e-01
7.72968352e-01 -3.69996965e-01 -4.33993965e-01 1.19660795e+00
3.71312909e-02 -2.38237344e-02 4.80785638e-01 -1.18010676e+00
-1.19510972e+00 3.30051690e-01 -7.95673311e-01 -4.63413835e-01
4.28176224e-01 4.22352076e-01 6.76298082e-01 -6.85249090e-01
9.64111626e-01 1.17772460e+00 5.29942930e-01 7.00850904e-01
-1.88517082e+00 -5.15866697e-01 -1.71334893e-01 -2.88037211e-01
-1.46760285e+00 -3.82672131e-01 1.25451910e+00 -2.12832496e-01
4.09018189e-01 6.63543463e-01 5.89947820e-01 9.86956298e-01
4.95627493e-01 6.18749082e-01 1.45357573e+00 -5.75047195e-01
2.69632250e-01 1.06404282e-01 -9.01484609e-01 6.09954953e-01
-5.45140065e-04 1.14412934e-01 -3.99469227e-01 -1.11008301e-01
1.38759673e+00 3.21756005e-01 -4.97564405e-01 -2.72169024e-01
-1.23367715e+00 7.37461805e-01 3.82448345e-01 1.62680000e-01
-6.00461066e-01 2.48135865e-01 -7.31771290e-02 4.00662065e-01
5.58251977e-01 5.21666527e-01 -6.53229579e-02 4.23855424e-01
-1.55545664e+00 5.18926203e-01 4.44206923e-01 1.09221792e+00
1.24595523e+00 3.52860540e-01 -3.72266561e-01 7.12633669e-01
-7.84557760e-02 4.85822380e-01 3.03551376e-01 -1.33007789e+00
5.28441191e-01 3.24009389e-01 1.53549522e-01 -1.08612466e+00
5.29915690e-01 -2.20746025e-01 -1.19119084e+00 7.50134408e-01
1.29922122e-01 -8.23161229e-02 -1.15368783e+00 1.97610188e+00
1.63297683e-01 6.18114322e-02 -1.03395708e-01 6.20395422e-01
2.60720044e-01 1.00258541e+00 -1.45622090e-01 -4.89787072e-01
9.87470925e-01 -1.06621599e+00 -9.31042194e-01 -4.12249506e-01
-4.73964587e-02 -9.33303177e-01 9.67734098e-01 1.86965391e-01
-1.49173820e+00 -7.96700060e-01 -1.01453996e+00 -2.13055149e-01
1.31527409e-01 -1.29175738e-01 2.36518215e-02 4.28807527e-01
-1.38148618e+00 4.87844437e-01 -4.64944571e-01 1.21755444e-01
3.14328313e-01 1.16260834e-01 -6.34587944e-01 -2.10082248e-01
-8.47568572e-01 5.72432637e-01 4.64863420e-01 -2.74575174e-01
-8.12292993e-01 -8.53375137e-01 -9.73636806e-01 -2.91238278e-02
1.22523293e-01 -9.34763253e-01 7.23333299e-01 -1.61363554e+00
-1.27266514e+00 8.55798006e-01 -2.64662176e-01 -3.16029847e-01
8.64376009e-01 1.96964592e-01 -8.21697190e-02 2.35157743e-01
4.29948926e-01 1.23321044e+00 1.51448214e+00 -1.72686732e+00
-3.10522437e-01 -2.55678836e-02 -2.12294057e-01 3.09191495e-01
1.00770935e-01 -5.18480241e-01 -5.18814087e-01 -1.39450419e+00
1.21994488e-01 -7.83569574e-01 -4.29215133e-01 5.06638229e-01
-5.40255368e-01 6.13752782e-01 1.24096596e+00 -1.02119350e+00
9.58967686e-01 -2.28373170e+00 4.23674405e-01 2.30862990e-01
3.33564907e-01 -7.89396167e-02 -3.49326164e-01 6.25924468e-01
-2.76420921e-01 9.91295800e-02 -8.04955244e-01 -7.11017668e-01
-3.64993453e-01 6.50141597e-01 -6.90707743e-01 3.66834342e-01
4.69838291e-01 1.12683249e+00 -7.50364065e-01 -6.47795260e-01
2.08513230e-01 6.09396219e-01 -7.71732926e-01 3.40403527e-01
-4.94072080e-01 9.71577764e-01 -2.91681975e-01 6.67115927e-01
9.40643311e-01 6.61151484e-02 -1.31696343e-01 -9.73001681e-03
6.50237650e-02 -4.88551587e-01 -1.00172400e+00 1.95472980e+00
-4.96367961e-01 6.67137802e-01 3.25562805e-01 -8.09001982e-01
1.08926034e+00 1.84749767e-01 3.85222703e-01 -7.04248369e-01
-2.75368065e-01 2.14513078e-01 -3.68877530e-01 -3.18670243e-01
7.38592088e-01 -4.19118851e-01 4.28745225e-02 4.21072751e-01
-1.10464178e-01 -5.14424980e-01 1.45699695e-01 3.11718971e-01
7.70614207e-01 3.18298787e-01 8.92958343e-02 -1.01155460e-01
2.11793318e-01 -1.89272225e-01 4.52217698e-01 7.53955662e-01
2.67232716e-01 1.14539993e+00 3.29875231e-01 -2.08910584e-01
-1.60639393e+00 -1.09103763e+00 3.45414169e-02 2.34180272e-01
1.62635580e-01 -2.96235085e-04 -1.13620698e+00 -3.91846746e-01
-2.01805010e-01 6.69886291e-01 -8.92191529e-01 -4.83676838e-03
-6.89887404e-01 -9.06394124e-02 3.14851403e-01 2.26674512e-01
7.04375148e-01 -1.39024222e+00 -4.50661987e-01 3.96240294e-01
-5.05915582e-01 -8.46462488e-01 -9.43733215e-01 -2.66356990e-02
-1.03003705e+00 -5.89650869e-01 -1.35357797e+00 -1.09228659e+00
1.17409813e+00 2.24301651e-01 1.13590872e+00 1.17008612e-01
-4.10260528e-01 1.10646524e-01 -1.09440237e-01 1.60442844e-01
-9.64692473e-01 -2.83833593e-01 -2.93924361e-01 2.89012074e-01
-5.75302541e-01 -1.04894161e+00 -6.99136257e-01 9.11881104e-02
-1.71651316e+00 5.74519455e-01 7.72562265e-01 1.07402956e+00
9.54297066e-01 1.98406771e-01 3.33482951e-01 -1.00295961e+00
4.43286091e-01 -3.81131351e-01 -1.71730652e-01 2.50355721e-01
-4.47784096e-01 1.38605431e-01 8.46208334e-01 -6.65186465e-01
-1.17154753e+00 1.29649684e-01 -9.61574167e-02 -8.47154438e-01
-2.25840136e-01 1.70449689e-02 -4.33362424e-01 -1.24013960e-01
6.84337437e-01 8.21281433e-01 2.95401335e-01 -2.42413893e-01
7.61238337e-01 2.19314456e-01 8.90429735e-01 -7.25449502e-01
1.14335680e+00 7.15246141e-01 -2.37352289e-02 -6.09442353e-01
-3.30062062e-01 1.18795171e-01 -3.85091245e-01 -1.24252051e-01
8.24276447e-01 -9.49667394e-01 1.21741347e-01 3.78368050e-01
-1.10413837e+00 -2.27776930e-01 -9.70425487e-01 -2.19479188e-01
-8.96556914e-01 6.01585507e-01 -5.97366333e-01 -5.58747530e-01
-1.58405110e-01 -1.26186740e+00 1.14467323e+00 1.35887340e-01
-2.53439367e-01 -9.35380101e-01 3.80857512e-02 1.83400095e-01
3.29703748e-01 7.86060631e-01 8.11393559e-01 3.17565858e-01
-9.60684836e-01 -4.41404991e-02 -6.14330582e-02 5.61034918e-01
4.59779680e-01 -2.09382176e-01 -6.87418699e-01 -3.59492153e-01
1.85928598e-01 -9.32448134e-02 8.41689467e-01 4.35996294e-01
1.13419819e+00 -8.25924337e-01 -1.53079137e-01 7.43041039e-01
1.77183855e+00 2.19603300e-01 1.17433524e+00 1.57944635e-01
7.96168745e-01 6.51721001e-01 3.54930431e-01 4.13209170e-01
-3.53677988e-01 4.59029317e-01 4.76331234e-01 -5.54195881e-01
-4.01055157e-01 -7.65574157e-01 2.27664009e-01 3.70542675e-01
1.54934376e-01 -3.29977930e-01 -2.17876956e-01 6.77467644e-01
-1.78359711e+00 -1.22148108e+00 1.51574016e-01 1.92699468e+00
1.35359037e+00 -2.99553812e-01 -1.44407377e-01 4.96547297e-02
8.84584308e-01 2.91754663e-01 -5.88790894e-01 -2.74993509e-01
-3.45778644e-01 5.05581141e-01 4.18145418e-01 7.73926318e-01
-5.38935423e-01 7.74581671e-01 6.21732950e+00 1.30840826e+00
-1.04383528e+00 5.20214699e-02 1.18387246e+00 -7.80226439e-02
-9.10792351e-01 2.41422221e-01 -2.83805609e-01 7.03785717e-01
2.42079854e-01 1.30975202e-01 6.63647234e-01 6.27446592e-01
2.24563465e-01 -1.83591098e-01 -8.45708609e-01 1.00564897e+00
1.60277784e-01 -1.69263053e+00 4.94760364e-01 1.02601059e-01
1.38660157e+00 -7.46837318e-01 3.97850931e-01 -2.06146494e-01
-6.34626150e-02 -9.79486227e-01 1.20427048e+00 7.09656298e-01
1.36035264e+00 -9.12963390e-01 1.01958066e-01 5.41019499e-01
-8.04891288e-01 2.01034084e-01 -1.97299272e-01 3.22086513e-02
2.55074710e-01 6.83141768e-01 -3.57684702e-01 3.98768723e-01
4.35043842e-01 5.20717859e-01 -2.18216389e-01 5.49604893e-01
-3.27493191e-01 2.74204016e-01 -6.44548833e-02 5.93846619e-01
2.91094165e-02 -6.27548933e-01 6.65385127e-01 9.80563641e-01
5.32156765e-01 -2.91692209e-03 1.52144194e-01 1.58153927e+00
-2.14879021e-01 -4.94272821e-02 -8.88554096e-01 -2.13607475e-02
3.00048321e-01 1.02410209e+00 -7.46067584e-01 -4.08280075e-01
-4.01799642e-02 1.68979144e+00 7.32644424e-02 6.69346154e-01
-8.65428746e-01 -4.85038102e-01 4.91192758e-01 3.90441447e-01
2.10962236e-01 -9.90358442e-02 -5.18449664e-01 -1.11007047e+00
2.79059201e-01 -1.09589326e+00 -2.36652970e-01 -1.03054476e+00
-1.12445533e+00 6.80686057e-01 -1.27332166e-01 -1.54495382e+00
-3.11800718e-01 1.15939721e-01 -6.41918600e-01 1.40289283e+00
-1.28561091e+00 -1.44870257e+00 -1.59401625e-01 7.33184159e-01
8.78555954e-01 -9.06704590e-02 6.41293764e-01 3.40553857e-02
-4.07633781e-02 5.10744750e-01 2.88756877e-01 -1.56355694e-01
6.70105159e-01 -8.80057871e-01 3.84087294e-01 1.04951394e+00
1.44765228e-01 4.74333525e-01 8.76140118e-01 -8.74973893e-01
-1.02018476e+00 -1.22754622e+00 6.46187782e-01 -9.55232531e-02
1.25365123e-01 -1.66832894e-01 -9.43546474e-01 6.65763617e-01
6.76905513e-01 -1.00852214e-01 2.11875573e-01 -8.75823438e-01
-2.99395442e-01 1.42384648e-01 -1.62122822e+00 7.78872848e-01
8.08341205e-01 -5.74184239e-01 -3.09865415e-01 1.69545606e-01
8.25621843e-01 -4.54336703e-01 -4.33365911e-01 8.04505032e-03
2.10392639e-01 -1.04665542e+00 1.09093773e+00 -1.05527505e-01
1.04834878e+00 -5.92340648e-01 -1.44254163e-01 -1.26016319e+00
-4.49553937e-01 -1.00271857e+00 4.55434211e-02 1.27975309e+00
9.54444259e-02 -3.36809516e-01 9.70193744e-01 4.93003517e-01
-8.28795061e-02 -6.34066463e-01 -8.27512443e-01 -6.38775051e-01
3.71202901e-02 -2.02510040e-02 4.39212799e-01 9.79804218e-01
-5.56498230e-01 -1.82047337e-01 -8.62384915e-01 6.85525611e-02
8.91896129e-01 2.47823387e-01 6.43876135e-01 -4.72047955e-01
-7.73697138e-01 -1.95731923e-01 -1.15929730e-01 -1.11105561e+00
1.37983173e-01 -6.47343099e-01 2.49199867e-01 -1.35195720e+00
2.70474046e-01 -3.98263812e-01 2.40756497e-01 5.78179419e-01
-1.84213042e-01 8.57822299e-01 2.86936104e-01 5.47828496e-01
6.39803484e-02 5.65682352e-01 1.95024145e+00 -4.25995260e-01
-1.66492015e-01 -3.78030300e-01 -8.20706904e-01 3.85594517e-01
7.37447083e-01 -5.03076315e-01 -5.75005472e-01 -3.40791851e-01
-6.21889085e-02 5.60859203e-01 5.78835189e-01 -9.56956506e-01
-1.88723370e-01 -4.16416466e-01 6.75515890e-01 -3.73723149e-01
4.20388967e-01 -7.33261883e-01 8.46271992e-01 3.61340046e-01
-3.10352117e-01 4.00095657e-02 1.75359309e-01 6.47003412e-01
-4.47382629e-01 -3.00955832e-01 1.01494706e+00 -4.71150607e-01
-5.71664691e-01 2.84118056e-01 -2.47524008e-01 -4.66491990e-02
1.10542345e+00 -7.50382364e-01 4.19343524e-02 -6.03644788e-01
-7.94811070e-01 -3.73907447e-01 9.81603980e-01 2.41533682e-01
8.65912855e-01 -1.48160815e+00 -7.71764219e-01 6.67485893e-01
-2.23603517e-01 1.48224324e-01 5.28324723e-01 3.09875757e-01
-8.36761057e-01 -9.06628594e-02 -7.17267632e-01 -4.59618747e-01
-1.00421011e+00 6.65068567e-01 6.27401695e-02 -2.53262848e-01
-6.82978094e-01 7.45287597e-01 7.43123591e-01 -4.75023836e-02
-1.00874454e-01 -1.37896910e-01 4.42210287e-01 -3.16339463e-01
4.12014216e-01 -1.27500758e-01 -3.79055679e-01 -5.75415254e-01
3.40123653e-01 4.81445849e-01 -1.86960828e-02 -3.67414534e-01
1.13309491e+00 -2.70453364e-01 -3.77199590e-01 2.44591734e-03
1.17194843e+00 4.73542154e-01 -1.74483705e+00 -7.30178133e-02
-7.04786837e-01 -8.88837636e-01 -5.08755669e-02 -5.55300832e-01
-1.26764429e+00 4.82082993e-01 3.57893676e-01 6.13935515e-02
1.41296387e+00 -2.43932858e-01 9.86672878e-01 -2.69498497e-01
3.95805031e-01 -7.02023268e-01 2.52292961e-01 -1.36559740e-01
1.30115795e+00 -8.87564957e-01 -2.67468952e-02 -4.31359887e-01
-4.99655247e-01 8.90834570e-01 4.26210791e-01 -5.93817115e-01
2.56627560e-01 3.87155235e-01 7.66126737e-02 3.02159458e-01
-3.61559331e-01 2.91885436e-01 1.26796201e-01 7.67327189e-01
1.47901461e-01 7.39227375e-03 -2.12021858e-01 -4.70299534e-02
2.09692735e-02 -1.21359281e-01 4.97179389e-01 9.71311688e-01
-2.40680635e-01 -1.54985332e+00 -8.76875639e-01 1.65776700e-01
-3.36416602e-01 -2.49115348e-01 -2.98283458e-01 5.92903316e-01
4.55936700e-01 6.82858050e-01 6.45475462e-02 -5.01611866e-02
-8.78974516e-03 -2.00202927e-01 5.71364820e-01 -5.76032341e-01
-2.40872070e-01 4.22724545e-01 -4.83283758e-01 -5.39528310e-01
-3.59435767e-01 -5.01130760e-01 -8.52702558e-01 -1.81876168e-01
-2.72716526e-02 1.20122116e-02 3.01207393e-01 5.55268347e-01
4.49514925e-01 6.35752678e-01 6.80159986e-01 -1.20640922e+00
-3.04940045e-01 -6.46161973e-01 -8.40192735e-01 6.08258069e-01
5.32089114e-01 -1.24738820e-01 -2.37118199e-01 7.06685364e-01] | [11.634339332580566, -0.8556600213050842] |
a16de766-4970-4606-bd79-80565cff7aff | time-contrastive-learning-based-deep | 1905.04554 | null | https://arxiv.org/abs/1905.04554v1 | https://arxiv.org/pdf/1905.04554v1.pdf | Time-Contrastive Learning Based Deep Bottleneck Features for Text-Dependent Speaker Verification | There are a number of studies about extraction of bottleneck (BN) features from deep neural networks (DNNs)trained to discriminate speakers, pass-phrases and triphone states for improving the performance of text-dependent speaker verification (TD-SV). However, a moderate success has been achieved. A recent study [1] presented a time contrastive learning (TCL) concept to explore the non-stationarity of brain signals for classification of brain states. Speech signals have similar non-stationarity property, and TCL further has the advantage of having no need for labeled data. We therefore present a TCL based BN feature extraction method. The method uniformly partitions each speech utterance in a training dataset into a predefined number of multi-frame segments. Each segment in an utterance corresponds to one class, and class labels are shared across utterances. DNNs are then trained to discriminate all speech frames among the classes to exploit the temporal structure of speech. In addition, we propose a segment-based unsupervised clustering algorithm to re-assign class labels to the segments. TD-SV experiments were conducted on the RedDots challenge database. The TCL-DNNs were trained using speech data of fixed pass-phrases that were excluded from the TD-SV evaluation set, so the learned features can be considered phrase-independent. We compare the performance of the proposed TCL bottleneck (BN) feature with those of short-time cepstral features and BN features extracted from DNNs discriminating speakers, pass-phrases, speaker+pass-phrase, as well as monophones whose labels and boundaries are generated by three different automatic speech recognition (ASR) systems. Experimental results show that the proposed TCL-BN outperforms cepstral features and speaker+pass-phrase discriminant BN features, and its performance is on par with those of ASR derived BN features. Moreover,.... | ['Suwon Shon', 'Zheng-Hua Tan', 'Achintya kr. Sarkar', 'James Glass', 'Hao Tang'] | 2019-05-11 | null | null | null | null | ['text-dependent-speaker-verification'] | ['speech'] | [ 9.49354023e-02 -3.33916426e-01 -2.90704221e-01 -7.69114852e-01
-8.94891918e-01 -5.18237233e-01 7.87297964e-01 -5.11931479e-02
-7.31988966e-01 4.65911895e-01 2.76240677e-01 -3.48267823e-01
-2.89715510e-02 -1.20924547e-01 -1.01942778e-01 -1.18729854e+00
-2.22069472e-01 4.19423819e-01 1.49192378e-01 -6.53060749e-02
2.87032966e-02 6.24798119e-01 -1.52619815e+00 4.27338541e-01
3.32322687e-01 9.77892220e-01 2.25095138e-01 7.25455821e-01
-1.25860482e-01 4.57846820e-01 -8.90706480e-01 1.50707453e-01
-3.54650035e-03 -7.25080073e-01 -7.12910712e-01 -1.68591812e-01
1.98132664e-01 8.84678587e-02 -4.29998457e-01 9.62238610e-01
8.29750836e-01 2.37409145e-01 8.04718196e-01 -1.27633834e+00
-6.75181821e-02 9.30234253e-01 -2.83098787e-01 9.12957668e-01
1.68890640e-01 -1.24892995e-01 8.39818716e-01 -8.30859721e-01
2.39564419e-01 1.25822759e+00 4.55421716e-01 9.31523561e-01
-1.36064219e+00 -9.75669265e-01 7.45225102e-02 4.42741752e-01
-1.44569659e+00 -1.07842112e+00 8.30815732e-01 -6.30797863e-01
1.16824162e+00 3.57716262e-01 5.99437654e-01 1.44058597e+00
2.17308775e-01 7.55903602e-01 1.03273475e+00 -5.15722752e-01
3.40856999e-01 2.17588723e-01 6.77843273e-01 2.97786444e-01
-6.05475426e-01 3.56738955e-01 -1.01034594e+00 -8.33701938e-02
4.02875513e-01 -4.60624903e-01 -4.42578882e-01 5.29278331e-02
-1.33490229e+00 7.46449471e-01 -1.31480336e-01 9.18298006e-01
-2.55827576e-01 -3.19360822e-01 8.25395346e-01 4.82434958e-01
5.28822899e-01 -2.32510120e-01 -6.51398122e-01 -2.42461085e-01
-1.17606533e+00 -2.46137068e-01 6.62848353e-01 7.10268140e-01
4.24409062e-01 3.73046845e-01 -3.36148649e-01 1.11256242e+00
3.33877444e-01 6.98904991e-01 1.17021120e+00 -4.60727900e-01
2.69680083e-01 -1.26398742e-01 -5.32262325e-01 -6.37343287e-01
-7.20893562e-01 -3.13609242e-01 -1.06603217e+00 -3.09484303e-01
2.26914421e-01 -1.01513252e-01 -1.01262093e+00 2.13617373e+00
1.85700879e-01 8.74405950e-02 2.77240783e-01 6.45554006e-01
1.00591397e+00 9.21319246e-01 5.46763279e-02 -6.79871559e-01
1.31624329e+00 -5.94247401e-01 -8.65787089e-01 1.48472011e-01
5.07018089e-01 -5.45043647e-01 6.27057850e-01 4.93688017e-01
-7.18879879e-01 -8.06095898e-01 -8.22790384e-01 5.75608730e-01
-2.65686601e-01 1.72494620e-01 1.50704041e-01 1.05772614e+00
-1.20819628e+00 2.83585638e-01 -1.10261202e+00 -2.91075438e-01
7.70916939e-02 5.36219358e-01 -4.83032703e-01 4.23983216e-01
-1.37346959e+00 7.45992482e-01 6.63779020e-01 1.85745761e-01
-1.39449155e+00 -4.84523296e-01 -9.40779448e-01 -3.86733375e-02
-2.00661376e-01 -1.95832662e-02 1.24355519e+00 -1.14403260e+00
-1.72685444e+00 8.30769479e-01 -5.20736694e-01 -6.95627213e-01
-4.23902161e-02 4.98997450e-01 -9.27895367e-01 3.91394168e-01
-4.77667600e-02 6.59730673e-01 1.17914772e+00 -7.70043433e-01
-6.38403594e-01 -3.72753352e-01 -4.95035797e-01 4.17676494e-02
-4.59977329e-01 4.66621637e-01 7.35679641e-02 -6.66647792e-01
4.04695332e-01 -9.11026657e-01 3.44867080e-01 -7.64261544e-01
-6.26131833e-01 -6.85264766e-01 8.40656519e-01 -7.85831749e-01
1.13041878e+00 -2.50022221e+00 6.64754137e-02 2.39338711e-01
-6.65050671e-02 2.90739685e-01 -2.82760859e-01 8.76432955e-02
-5.24852157e-01 3.18388641e-02 -2.08803162e-01 -4.29041207e-01
5.99814616e-02 2.34998718e-01 -2.16138601e-01 6.88294113e-01
1.32588848e-01 3.32709253e-01 -5.16315281e-01 -5.05238175e-01
1.96416155e-01 5.51546633e-01 -1.49387807e-01 -3.51632014e-03
2.21052021e-01 7.87617326e-01 -3.02117392e-02 2.59826869e-01
5.48284471e-01 5.22837996e-01 -3.39538530e-02 -1.35078505e-01
-2.01030612e-01 8.56923163e-01 -8.30892265e-01 1.39057457e+00
-2.97268450e-01 1.14288902e+00 -5.26612177e-02 -1.48556697e+00
9.48800087e-01 1.02028680e+00 4.77448732e-01 -7.01338530e-01
1.99959010e-01 8.69076848e-02 4.93614346e-01 -3.89769405e-01
-1.09705307e-01 -4.41862226e-01 -6.08445816e-02 2.97881663e-01
6.33073270e-01 1.04934059e-01 1.87375546e-01 -1.10726170e-01
8.95289004e-01 -3.43606979e-01 4.97645922e-02 -5.75372756e-01
1.12234187e+00 -4.40123707e-01 7.87141860e-01 5.17088473e-01
-7.27126777e-01 3.76429796e-01 4.14350331e-01 -1.30225033e-01
-6.59851313e-01 -1.16326153e+00 -6.26579225e-01 1.30114388e+00
-3.00983399e-01 -2.70309567e-01 -8.45778286e-01 -5.66419065e-01
-4.90540266e-01 8.36244345e-01 -4.64054286e-01 -2.00933263e-01
-6.13718510e-01 -8.01713347e-01 9.51592207e-01 4.07808244e-01
3.53706241e-01 -1.30112362e+00 -3.73032600e-01 2.05620304e-01
-4.62425023e-01 -1.18047953e+00 -5.15626967e-01 7.00856924e-01
-7.16228664e-01 -5.99976242e-01 -7.48426914e-01 -1.11150360e+00
3.76872122e-01 -7.71324411e-02 6.18123055e-01 -6.10154450e-01
6.64223284e-02 4.68405008e-01 -4.08750176e-01 -2.84568757e-01
-7.33796716e-01 -2.94537414e-02 6.90898895e-01 2.86017895e-01
7.25941837e-01 -5.03469169e-01 -1.82663023e-01 6.53421760e-01
-6.73718810e-01 -3.80445451e-01 3.11554700e-01 9.45390582e-01
4.85361248e-01 4.08451915e-01 9.83100593e-01 -1.78691283e-01
4.51469541e-01 -1.79609343e-01 -3.96315157e-01 -4.27837074e-02
-7.70480782e-02 -2.15652157e-02 5.41972220e-01 -7.78331280e-01
-1.00437880e+00 1.13815829e-01 -3.95380855e-01 -4.25235331e-01
-6.23079896e-01 6.02328479e-01 -2.07601160e-01 1.99348003e-01
5.35945892e-01 7.25937843e-01 -1.11528728e-02 -3.53528351e-01
-5.71737345e-03 1.08162415e+00 4.77912933e-01 -4.31431592e-01
4.82420266e-01 7.58749321e-02 -4.86123741e-01 -1.51349103e+00
-5.17228484e-01 -8.64427030e-01 -8.63905191e-01 -2.19452709e-01
9.92472351e-01 -8.78177047e-01 -5.66381931e-01 6.62678242e-01
-1.18439484e+00 -2.02464789e-01 -8.71328339e-02 1.08443022e+00
-5.57564735e-01 4.70774263e-01 -6.52244985e-01 -1.13890111e+00
-4.00956243e-01 -1.18431258e+00 7.60614276e-01 1.02561355e-01
-2.61670917e-01 -7.07343519e-01 1.57014132e-01 2.89739996e-01
1.12586468e-01 -3.12393993e-01 8.61325860e-01 -1.40450811e+00
2.16307446e-01 -4.84306142e-02 3.13786298e-01 8.50679338e-01
1.35359198e-01 -1.29205287e-01 -1.46208394e+00 -4.94499654e-01
5.83243847e-01 -1.00731887e-01 8.64488661e-01 7.35684276e-01
7.58054197e-01 -1.44228086e-01 -2.30573103e-01 2.85542935e-01
7.64793694e-01 8.66378605e-01 3.01541686e-01 -6.40816689e-02
2.95084625e-01 6.44585252e-01 2.58915156e-01 2.05921620e-01
1.64598841e-02 5.89396060e-01 -1.39372304e-01 2.00508475e-01
-6.47917986e-02 2.78057516e-01 1.00988197e+00 1.47363150e+00
8.03068280e-02 -6.26944453e-02 -1.06188202e+00 7.26668179e-01
-1.39307153e+00 -1.14891326e+00 -5.37301525e-02 2.41439772e+00
8.85158658e-01 2.84908146e-01 5.01838744e-01 6.48822784e-01
1.02085614e+00 8.73647630e-02 -4.02752966e-01 -4.21943307e-01
-1.99507684e-01 3.48006636e-01 7.59277418e-02 2.51300097e-01
-1.32878721e+00 7.08714247e-01 5.39734030e+00 1.09570348e+00
-1.43931329e+00 3.21380883e-01 5.54214537e-01 3.82431820e-02
3.46769720e-01 -4.18670446e-01 -1.00728321e+00 3.87860537e-01
1.69832039e+00 -1.19006343e-01 2.63433456e-01 5.12907505e-01
4.52184677e-01 -3.61669548e-02 -1.47590613e+00 1.17889583e+00
3.03609252e-01 -6.82943702e-01 -2.72902876e-01 -8.87513161e-02
3.62119377e-01 3.30585599e-01 1.71124533e-01 3.98693621e-01
-2.42084742e-01 -7.24383831e-01 9.68358040e-01 8.60730708e-02
7.63459086e-01 -8.78135562e-01 8.25805664e-01 5.94538867e-01
-1.33767438e+00 -5.94034344e-02 -3.32961440e-01 3.00985754e-01
-7.14249015e-02 4.26266700e-01 -1.12153125e+00 4.33018208e-01
7.31243610e-01 5.96643984e-01 -2.96414196e-01 9.86009300e-01
-7.08224550e-02 1.07052195e+00 -2.49126941e-01 -1.27180412e-01
1.82598233e-01 8.19954053e-02 7.76903570e-01 1.49812472e+00
1.74822778e-01 9.99919939e-05 -2.80840732e-02 6.04030371e-01
3.85438263e-01 1.42697632e-01 -4.33506310e-01 -3.01482737e-01
3.97772402e-01 1.10165179e+00 -9.20838356e-01 -4.02342290e-01
-2.71266967e-01 6.55715942e-01 -2.00462356e-01 3.14592600e-01
-7.76276350e-01 -4.01632518e-01 5.53866446e-01 -3.89275998e-01
3.65197122e-01 -3.64805281e-01 1.88925728e-01 -1.12774837e+00
-2.11558223e-01 -8.26204181e-01 3.79463315e-01 -3.54007691e-01
-1.43868256e+00 1.17380965e+00 1.41159788e-01 -1.26585019e+00
-4.11131412e-01 -5.21764219e-01 -3.92295301e-01 1.02877903e+00
-1.39810359e+00 -8.71389747e-01 3.73531044e-01 9.67252672e-01
9.13957536e-01 -4.92779285e-01 1.00848508e+00 2.83177733e-01
-8.08914602e-01 6.21708453e-01 1.99713394e-01 5.88711023e-01
6.01778686e-01 -9.81972575e-01 1.16489559e-01 8.59908164e-01
4.37992960e-01 6.07468724e-01 4.51690197e-01 -2.22306505e-01
-7.24887490e-01 -8.66484404e-01 1.19034600e+00 -1.08483896e-01
4.33407545e-01 -7.54176140e-01 -7.30613530e-01 5.68846583e-01
2.32246041e-01 -2.35762700e-01 1.03523898e+00 -2.73150532e-03
-2.74294734e-01 -3.01591426e-01 -9.74118292e-01 1.37002587e-01
5.33922136e-01 -9.75694537e-01 -1.08302617e+00 3.42527777e-01
5.88912666e-01 8.20582509e-02 -6.09214723e-01 2.54854977e-01
3.82924438e-01 -9.38992202e-01 8.30103219e-01 -3.92091513e-01
-1.60636574e-01 -2.04092324e-01 -3.42276037e-01 -1.39974153e+00
-2.67326146e-01 -5.08683503e-01 3.68669003e-01 1.45245528e+00
5.50580800e-01 -7.29316831e-01 5.06197453e-01 2.68691033e-01
-2.81992406e-01 -2.89520584e-02 -1.49717820e+00 -1.04516423e+00
-1.22144008e-02 -7.81042814e-01 2.64622569e-01 9.48815465e-01
2.91666269e-01 6.61356628e-01 2.30331682e-02 3.06194425e-01
4.54782277e-01 -2.20856726e-01 1.72321260e-01 -1.42067981e+00
-1.09107316e-01 -5.03637671e-01 -6.30891204e-01 -8.65889609e-01
6.22641563e-01 -1.14962661e+00 3.40337515e-01 -8.68246257e-01
9.58594978e-02 -3.42935055e-01 -7.59369075e-01 5.94783306e-01
3.75218332e-01 8.24297871e-03 7.52997473e-02 1.90507799e-01
-1.52505189e-01 5.58048248e-01 7.01460719e-01 -4.60401028e-01
-4.61339474e-01 4.15704161e-01 1.30719826e-01 6.77142024e-01
7.76667535e-01 -6.51251972e-01 -5.08812249e-01 1.41356170e-01
-6.35184705e-01 4.13202435e-01 1.27565950e-01 -1.15810323e+00
3.22335213e-01 3.37039649e-01 3.55706483e-01 -9.37330067e-01
5.13038099e-01 -6.78301930e-01 -8.34040567e-02 5.32883167e-01
-4.82622772e-01 -3.01006973e-01 2.87148297e-01 4.42294717e-01
-4.29102331e-01 -3.92789602e-01 1.03323686e+00 1.18594557e-01
-5.02128065e-01 6.00614585e-02 -9.91037011e-01 -2.32394338e-01
7.21376538e-01 -2.13724047e-01 6.31625652e-02 -3.87081236e-01
-1.08438373e+00 -2.88879484e-01 -4.99226451e-01 3.51640642e-01
6.43861890e-01 -1.23458493e+00 -6.69667780e-01 6.20552897e-01
-3.49835232e-02 -2.79810518e-01 3.76539558e-01 1.15722239e+00
-1.59944501e-02 9.24582243e-01 -2.77986884e-01 -1.16068268e+00
-1.49529302e+00 5.28963745e-01 3.40190679e-01 5.94781935e-02
-4.25380647e-01 9.60686505e-01 4.14435208e-01 -4.60672945e-01
5.65704286e-01 -5.64883947e-01 -5.85200727e-01 4.31581974e-01
4.84942853e-01 6.56635165e-02 3.71134758e-01 -1.21943188e+00
-6.48398817e-01 3.08061302e-01 -1.52193457e-01 -5.39836645e-01
1.39345455e+00 -2.00682744e-01 -7.62480423e-02 9.17598903e-01
1.35457420e+00 -3.21116030e-01 -6.07218444e-01 -3.51912439e-01
2.35775217e-01 2.16229439e-01 2.58490741e-01 -6.50302768e-01
-1.08837104e+00 1.02978706e+00 8.83503020e-01 4.64569122e-01
1.28837109e+00 1.90154314e-02 6.43967390e-01 2.64092147e-01
3.49193275e-01 -9.91285920e-01 -7.44499564e-02 6.31244898e-01
7.04312861e-01 -9.86161470e-01 -7.60183156e-01 -1.37435654e-02
-4.97306049e-01 1.36220229e+00 3.05646688e-01 2.56091118e-01
9.29454803e-01 1.42342910e-01 1.39983907e-01 6.71429560e-02
-6.35668516e-01 -8.43372568e-02 2.80593604e-01 7.48403370e-01
4.18125927e-01 1.04545213e-01 -2.52235681e-01 8.32232952e-01
-3.85871530e-01 -4.18477178e-01 2.96129584e-01 6.59604847e-01
-4.14798945e-01 -9.46751535e-01 -5.28531730e-01 2.17771843e-01
-4.47766632e-01 -1.99304685e-01 -5.43127507e-02 5.55564344e-01
9.82175544e-02 1.28142118e+00 1.18866585e-01 -5.88381827e-01
9.25860330e-02 6.65976465e-01 2.31657907e-01 -6.43058717e-01
-5.96637249e-01 5.78193247e-01 1.08803660e-01 -2.05613837e-01
-7.52539575e-01 -8.48109424e-01 -1.22671533e+00 1.05986960e-01
-5.09352088e-01 3.71537864e-01 8.90201330e-01 1.07081044e+00
-4.53373119e-02 5.46029210e-01 9.54036891e-01 -9.53224301e-01
-5.16358554e-01 -1.25334096e+00 -8.44294369e-01 1.51187137e-01
7.96172261e-01 -6.15592122e-01 -7.21457183e-01 4.49376762e-01] | [14.454752922058105, 6.197702407836914] |
8c48b853-1aa1-4e6a-9889-8b6ddd950195 | locality-aware-attention-network-with | 2208.05636 | null | https://arxiv.org/abs/2208.05636v1 | https://arxiv.org/pdf/2208.05636v1.pdf | Locality-aware Attention Network with Discriminative Dynamics Learning for Weakly Supervised Anomaly Detection | Video anomaly detection is recently formulated as a multiple instance learning task under weak supervision, in which each video is treated as a bag of snippets to be determined whether contains anomalies. Previous efforts mainly focus on the discrimination of the snippet itself without modeling the temporal dynamics, which refers to the variation of adjacent snippets. Therefore, we propose a Discriminative Dynamics Learning (DDL) method with two objective functions, i.e., dynamics ranking loss and dynamics alignment loss. The former aims to enlarge the score dynamics gap between positive and negative bags while the latter performs temporal alignment of the feature dynamics and score dynamics within the bag. Moreover, a Locality-aware Attention Network (LA-Net) is constructed to capture global correlations and re-calibrate the location preference across snippets, followed by a multilayer perceptron with causal convolution to obtain anomaly scores. Experimental results show that our method achieves significant improvements on two challenging benchmarks, i.e., UCF-Crime and XD-Violence. | ['Xiaoyu Wu', 'Yujiang Pu'] | 2022-08-11 | null | null | null | null | ['supervised-anomaly-detection'] | ['computer-vision'] | [-6.56105429e-02 -5.99071443e-01 -2.96163242e-02 -3.84612560e-01
-2.51773566e-01 -1.54511079e-01 4.80131686e-01 4.92827654e-01
-3.79635036e-01 2.05409646e-01 1.78537026e-01 8.69695395e-02
-1.49279490e-01 -6.06116474e-01 -8.92991841e-01 -9.30658579e-01
-5.11058688e-01 3.24422926e-01 4.36441064e-01 7.05230981e-02
1.95003867e-01 2.98691690e-01 -1.24682438e+00 2.58980066e-01
8.10321331e-01 1.22814536e+00 -2.12021351e-01 4.54384208e-01
6.18869402e-02 1.06989837e+00 -4.13321495e-01 -2.66498029e-01
9.18743312e-02 -3.16819906e-01 -4.11096066e-01 1.47971898e-01
6.23515427e-01 -5.25396585e-01 -8.84234309e-01 1.18884957e+00
2.62169480e-01 3.51575553e-01 5.45114338e-01 -1.49893117e+00
-4.17779654e-01 2.41676778e-01 -9.50251162e-01 1.12383628e+00
1.06875956e-01 5.10983944e-01 1.04948437e+00 -6.95251882e-01
1.24765769e-01 1.31201482e+00 4.98857826e-01 2.94249266e-01
-1.17707324e+00 -7.47184634e-01 8.74690294e-01 8.58697057e-01
-1.20696604e+00 -2.33247593e-01 9.43036795e-01 -9.08239305e-01
7.60451019e-01 1.52643159e-01 6.44198060e-01 1.24464703e+00
-5.21055330e-03 1.05182123e+00 3.63347203e-01 2.17765749e-01
8.96121711e-02 -3.13270688e-01 3.02834094e-01 6.59632564e-01
-7.15857372e-03 -1.77741453e-01 -5.77791989e-01 -6.24849126e-02
7.61724353e-01 4.75972384e-01 -1.90995842e-01 -4.36829060e-01
-9.76772726e-01 6.79366827e-01 4.28063631e-01 2.43409444e-03
-4.23568249e-01 2.02826053e-01 8.69906962e-01 1.63590848e-01
6.48652554e-01 -1.82972133e-01 -3.51129681e-01 -3.16936105e-01
-5.83921134e-01 3.82328987e-01 2.57345438e-01 4.53811288e-01
6.06529355e-01 -1.73810005e-01 -4.10059273e-01 7.89201558e-01
2.99802721e-01 2.09771350e-01 5.72755098e-01 -3.90466243e-01
9.23594713e-01 7.38506854e-01 -6.60106726e-03 -1.58615828e+00
-2.62877762e-01 -1.87297165e-01 -8.59032631e-01 -1.96007177e-01
4.35457021e-01 -6.28861561e-02 -7.87401617e-01 1.81925201e+00
3.80339026e-01 9.81979728e-01 -4.51925427e-01 9.93279040e-01
3.85283381e-01 7.97682166e-01 2.24541232e-01 -2.36698136e-01
7.50454366e-01 -1.14937699e+00 -6.86078548e-01 -8.76691341e-02
6.40439808e-01 -7.49936253e-02 1.02045989e+00 2.73565441e-01
-1.06348133e+00 -3.87466133e-01 -7.71428823e-01 3.14503133e-01
-1.84484154e-01 -1.94902629e-01 2.94531077e-01 6.85801962e-03
-5.87117434e-01 7.87716627e-01 -1.32315075e+00 -1.64611393e-03
7.91952908e-01 2.30600774e-01 -1.42055362e-01 9.06375200e-02
-1.02079451e+00 3.64815205e-01 4.79313880e-01 2.11252987e-01
-9.53618765e-01 -8.02102268e-01 -9.32209730e-01 1.53129175e-01
3.98542076e-01 -1.86238468e-01 7.64757991e-01 -1.18664694e+00
-1.07974315e+00 7.49695241e-01 -2.18562111e-01 -6.63098872e-01
6.79082334e-01 -4.23697084e-01 -6.78476036e-01 -6.86102733e-02
2.24924877e-01 1.15583614e-02 8.49752307e-01 -7.94721484e-01
-9.65014279e-01 -6.23942673e-01 -9.85228941e-02 1.96555018e-01
-7.19455600e-01 7.65231177e-02 -9.03026879e-01 -9.29221690e-01
7.93255791e-02 -6.06680870e-01 -7.43025094e-02 -1.35846689e-01
-3.54510278e-01 -2.96248078e-01 1.10250175e+00 -1.03444254e+00
1.97554648e+00 -2.34247637e+00 4.53391075e-01 4.54505891e-01
3.30846936e-01 2.32276842e-01 -3.54470871e-02 -6.23112470e-02
-2.68427849e-01 -1.86141521e-01 -2.52256274e-01 -2.52296418e-01
-2.41196156e-01 2.39639223e-01 -3.85275692e-01 6.29827261e-01
4.15417254e-01 5.63781321e-01 -1.01639581e+00 -2.47161627e-01
8.22568834e-02 9.48061142e-03 -8.63692880e-01 2.29432300e-01
-2.20106512e-01 5.78449249e-01 -3.99580240e-01 6.22117937e-01
6.67397082e-01 -3.25445294e-01 -2.19866350e-01 7.21843615e-02
3.05234250e-02 9.65497196e-02 -1.12050235e+00 1.51245296e+00
-9.11865979e-02 4.25767094e-01 -2.59176672e-01 -1.26932490e+00
6.04912341e-01 1.87669359e-02 6.99621379e-01 -7.59465218e-01
-3.47368568e-01 2.35219393e-02 4.32147905e-02 -8.79083693e-01
1.23893723e-01 3.47705513e-01 -3.30088809e-02 1.84628889e-01
-1.48258880e-01 8.37790430e-01 2.91559517e-01 1.93762287e-01
1.37993979e+00 -6.20170915e-03 -2.01454341e-01 5.18957563e-02
7.09430456e-01 -3.65158647e-01 1.03144825e+00 5.79836905e-01
-4.75501537e-01 5.29385984e-01 9.38717604e-01 -7.79638231e-01
-9.30742979e-01 -1.27451992e+00 -6.92473492e-03 1.06182253e+00
6.35794282e-01 -3.24121326e-01 -4.72712457e-01 -1.10055077e+00
2.45837733e-01 3.82512003e-01 -8.23921740e-01 -4.63033050e-01
-8.79129767e-01 -1.08360100e+00 1.62227765e-01 6.80727422e-01
4.39200491e-01 -9.92012858e-01 -2.82597780e-01 2.74174303e-01
-2.47817039e-01 -9.83258069e-01 -8.61507058e-01 -2.98353642e-01
-7.60810077e-01 -1.13607478e+00 -2.47089967e-01 -4.39160079e-01
5.34294009e-01 -2.99813449e-02 8.30835342e-01 6.99859634e-02
-1.52282119e-01 2.55146831e-01 -2.17777908e-01 -1.32842287e-01
1.95607290e-01 -1.91110671e-01 1.99805394e-01 6.69382095e-01
6.54628396e-01 -9.10898507e-01 -8.65651369e-01 3.88425827e-01
-8.61067593e-01 -2.80150354e-01 3.14214647e-01 9.32895422e-01
6.67161286e-01 1.00634485e-01 4.21005845e-01 -6.22842431e-01
2.90716946e-01 -1.04861081e+00 -5.06834924e-01 6.40681013e-02
-2.99168795e-01 -2.20741570e-01 5.42635500e-01 -5.08344591e-01
-7.33080029e-01 -1.40878126e-01 2.14260086e-01 -9.44319546e-01
-9.00022909e-02 4.53266025e-01 -3.12924981e-01 4.45540488e-01
2.26524040e-01 3.18624586e-01 -1.58212394e-01 -3.72977287e-01
-2.08073139e-01 1.95294991e-01 7.34273612e-01 -3.24774295e-01
6.75873518e-01 5.98623753e-01 -2.84460723e-01 -6.31682038e-01
-8.84419978e-01 -8.32714379e-01 -5.63253760e-01 -3.17655236e-01
9.15225148e-01 -1.03488231e+00 -5.03871202e-01 8.41486931e-01
-9.59575653e-01 -2.36933380e-01 -1.37491763e-01 3.17239761e-01
-3.89377803e-01 3.39858562e-01 -5.83672881e-01 -6.87893450e-01
4.74933609e-02 -1.00357735e+00 1.03046370e+00 2.26846129e-01
6.37429282e-02 -1.09868646e+00 2.71306515e-01 3.08189571e-01
-4.46669385e-02 5.22407949e-01 9.26026285e-01 -7.65326798e-01
-6.04425728e-01 -4.70246613e-01 -3.54111910e-01 3.80836159e-01
3.86941656e-02 4.50911783e-02 -7.60336578e-01 -4.19524282e-01
-3.32600266e-01 -5.51024973e-02 1.16149592e+00 4.03315634e-01
1.86295497e+00 -5.42632461e-01 -4.23571557e-01 7.53357947e-01
1.02300882e+00 4.17200148e-01 6.11282289e-01 3.81949514e-01
1.19066191e+00 3.47151190e-01 5.08589149e-01 7.22659349e-01
4.31612492e-01 8.76913369e-01 6.79677010e-01 2.64868438e-01
3.60943466e-01 -2.74576426e-01 5.16559422e-01 7.40412712e-01
-1.59422252e-02 -1.67400762e-01 -1.02684331e+00 7.21648216e-01
-2.44843745e+00 -1.29875922e+00 -8.08879826e-03 2.23124576e+00
3.31405938e-01 3.35157305e-01 4.85223234e-01 -6.35042116e-02
8.72900367e-01 4.04723376e-01 -9.94064391e-01 3.20855603e-02
-2.05403697e-02 -2.81549156e-01 1.50609538e-01 8.25396031e-02
-1.58977139e+00 6.61478400e-01 4.45417833e+00 8.55140805e-01
-1.11216223e+00 1.62911877e-01 9.35555220e-01 -6.22370720e-01
1.07524693e-01 -3.46796483e-01 -6.22425497e-01 1.21924603e+00
6.74149692e-01 -8.23517051e-03 2.90162414e-01 6.07685626e-01
5.77498138e-01 1.85643896e-01 -1.03923237e+00 1.02149868e+00
1.58675984e-02 -1.06956232e+00 1.93001643e-01 -1.35032788e-01
5.73571563e-01 9.13190395e-02 1.79164529e-01 3.21496516e-01
9.66710895e-02 -8.94845784e-01 7.34346569e-01 8.22020590e-01
2.64298588e-01 -8.54610384e-01 5.65877676e-01 2.71927893e-01
-1.19951010e+00 -5.24671018e-01 -1.69187501e-01 1.06566116e-01
1.05792895e-01 4.80844259e-01 -1.69253245e-01 2.84918487e-01
1.11727977e+00 1.36373770e+00 -4.96953458e-01 1.26542139e+00
5.97176254e-02 7.39620447e-01 -1.87154934e-01 3.34054053e-01
3.56454551e-01 -4.80602056e-01 9.35241997e-01 1.21117413e+00
1.93012744e-01 -8.37066174e-02 4.80669558e-01 6.93659663e-01
1.48214892e-01 2.52105558e-04 -2.98759669e-01 1.18297249e-01
2.17645153e-01 1.01337278e+00 -2.92770237e-01 -1.16037473e-01
-6.54368758e-01 1.13057566e+00 5.04094064e-01 3.57450843e-01
-1.04226398e+00 -1.86015323e-01 1.08912146e+00 3.37516218e-01
4.10487503e-01 1.88104976e-02 -2.56375484e-02 -1.53092670e+00
4.06480283e-01 -5.39376676e-01 7.82298744e-01 -3.37141454e-01
-1.55626953e+00 2.90669560e-01 -7.15279132e-02 -1.39724600e+00
-1.96155198e-02 -3.91087443e-01 -1.24122179e+00 6.98320627e-01
-1.19383371e+00 -9.30905819e-01 -4.99987334e-01 7.58222938e-01
6.20841026e-01 -1.60874933e-01 1.85873821e-01 6.97103679e-01
-1.46692336e+00 7.34462142e-01 2.20423549e-01 4.55604732e-01
5.40499449e-01 -1.28773594e+00 3.36005211e-01 1.13356328e+00
-1.26859978e-01 2.04890609e-01 5.45462906e-01 -7.23193049e-01
-8.11842144e-01 -1.59651232e+00 3.98383766e-01 -4.42258060e-01
1.11829937e+00 -2.62296647e-01 -1.32368255e+00 7.71856546e-01
-4.03011888e-01 5.54887414e-01 4.77629960e-01 9.78638697e-03
-2.96650141e-01 -3.05727363e-01 -7.38952577e-01 4.93292093e-01
1.33905888e+00 -4.22142357e-01 -1.31272778e-01 4.16234761e-01
4.96798366e-01 -5.30180633e-01 -4.40031648e-01 5.80720067e-01
2.91379243e-01 -9.43332732e-01 8.29249680e-01 -1.07630050e+00
7.42082834e-01 -2.52100796e-01 -2.80493945e-02 -1.32430315e+00
-4.23480898e-01 -4.33257729e-01 -7.91664541e-01 1.17517269e+00
1.91717088e-01 -5.09790421e-01 7.72811651e-01 5.16815841e-01
-2.19142258e-01 -1.03327298e+00 -9.62556779e-01 -7.75760055e-01
-2.78303087e-01 -4.16805059e-01 5.56420088e-01 1.04704416e+00
-7.15087950e-02 9.14164484e-02 -4.88734365e-01 6.13397658e-01
4.91496384e-01 -1.67022720e-01 5.61695755e-01 -1.09616804e+00
-4.69758481e-01 -6.58338606e-01 -9.25747156e-01 -1.08300781e+00
1.74562722e-01 -7.57940233e-01 -1.55720115e-01 -8.48800421e-01
3.98463964e-01 -3.13995510e-01 -8.39075446e-01 1.40686572e-01
-5.21536052e-01 -1.56835526e-01 -1.74811691e-01 3.86567265e-01
-1.12356806e+00 7.65414000e-01 8.67829084e-01 -2.37591773e-01
-3.63230288e-01 2.41593301e-01 -2.86151856e-01 1.01249242e+00
5.71232617e-01 -3.89031708e-01 -3.40977669e-01 -4.84480143e-01
8.96838084e-02 -1.80112779e-01 6.62853956e-01 -1.08743918e+00
1.36245221e-01 -2.22986102e-01 3.99374992e-01 -6.48929358e-01
-5.32459803e-02 -6.73308432e-01 -2.11330920e-01 2.91316301e-01
-3.20409745e-01 4.24937636e-01 1.94957897e-01 1.27097940e+00
-4.82955575e-01 1.70515448e-01 6.71300948e-01 1.57660455e-01
-8.33146095e-01 1.17180967e+00 -6.07804246e-02 2.72895694e-01
1.26050782e+00 -5.77296093e-02 -7.24567771e-02 -1.71448052e-01
-9.25688684e-01 8.48482430e-01 1.95595548e-01 6.64083779e-01
6.30955696e-01 -1.74047148e+00 -6.19728506e-01 2.56338120e-01
3.07961315e-01 2.32191399e-01 7.88535237e-01 9.61021304e-01
-3.86431664e-01 4.17408645e-02 -6.32162467e-02 -8.41355920e-01
-1.28961658e+00 5.69708407e-01 5.46414614e-01 -5.37480414e-01
-7.08416581e-01 1.04037392e+00 4.13950175e-01 -7.23996907e-02
6.42146707e-01 -7.85716325e-02 -4.77553666e-01 1.48921058e-01
6.93455219e-01 4.59345937e-01 -8.94734114e-02 -6.22426748e-01
-4.76372391e-01 2.67695516e-01 -4.22964692e-01 3.31446975e-01
1.40757430e+00 -1.06917806e-01 -1.05154671e-01 6.07437670e-01
1.21348679e+00 -4.11222816e-01 -1.79133177e+00 -4.89922136e-01
2.19179854e-01 -6.65081918e-01 2.60148533e-02 -3.31686854e-01
-1.28528988e+00 7.43862092e-01 7.14826405e-01 1.68794066e-01
1.30604124e+00 -1.39840603e-01 1.00131059e+00 1.38717905e-01
-1.61113605e-01 -1.13276601e+00 5.82869947e-01 5.91393828e-01
8.02975714e-01 -1.37403309e+00 -4.20601249e-01 -6.05785698e-02
-5.79976261e-01 8.16832483e-01 1.20459282e+00 -2.72452652e-01
7.84444392e-01 -2.22074792e-01 -4.00400192e-01 -3.17882359e-01
-7.01379836e-01 1.49227113e-01 4.49204236e-01 1.42400265e-01
-5.77679239e-02 -6.97405636e-02 1.10269018e-01 7.41800308e-01
3.37191403e-01 -3.81236047e-01 3.32310162e-02 5.85130870e-01
-2.01534808e-01 -4.69126046e-01 -4.49947082e-02 8.13867927e-01
-5.18589556e-01 3.07875797e-02 -6.73596263e-02 4.52924043e-01
3.47947419e-01 4.37001735e-01 8.11480939e-01 -5.47185004e-01
5.17080307e-01 3.61086577e-02 1.32322475e-01 -4.02827799e-01
-3.60947460e-01 -8.45764503e-02 -4.03744102e-01 -8.19073319e-01
-1.40640691e-01 -1.10366976e+00 -1.03632879e+00 -2.11330995e-01
-1.47632763e-01 -4.14559916e-02 6.74053505e-02 1.03285480e+00
3.39513034e-01 6.60479128e-01 9.28617656e-01 -7.05702424e-01
-3.87724817e-01 -8.42645168e-01 -5.24002135e-01 1.01648676e+00
7.41478026e-01 -7.07171202e-01 -4.69452024e-01 2.10802611e-02] | [7.8381218910217285, 1.5920089483261108] |
c62d30d8-b3a4-4497-b622-d432ee564b55 | generative-probabilistic-novelty-detection | 1807.02588 | null | http://arxiv.org/abs/1807.02588v2 | http://arxiv.org/pdf/1807.02588v2.pdf | Generative Probabilistic Novelty Detection with Adversarial Autoencoders | Novelty detection is the problem of identifying whether a new data point is
considered to be an inlier or an outlier. We assume that training data is
available to describe only the inlier distribution. Recent approaches primarily
leverage deep encoder-decoder network architectures to compute a reconstruction
error that is used to either compute a novelty score or to train a one-class
classifier. While we too leverage a novel network of that kind, we take a
probabilistic approach and effectively compute how likely is that a sample was
generated by the inlier distribution. We achieve this with two main
contributions. First, we make the computation of the novelty probability
feasible because we linearize the parameterized manifold capturing the
underlying structure of the inlier distribution, and show how the probability
factorizes and can be computed with respect to local coordinates of the
manifold tangent space. Second, we improved the training of the autoencoder
network. An extensive set of results show that the approach achieves
state-of-the-art results on several benchmark datasets. | ['Donald A. Adjeroh', 'Ranya Almohsen', 'Stanislav Pidhorskyi', 'Gianfranco Doretto'] | 2018-07-06 | generative-probabilistic-novelty-detection-1 | http://papers.nips.cc/paper/7915-generative-probabilistic-novelty-detection-with-adversarial-autoencoders | http://papers.nips.cc/paper/7915-generative-probabilistic-novelty-detection-with-adversarial-autoencoders.pdf | neurips-2018-12 | ['one-class-classifier'] | ['methodology'] | [-1.76134542e-01 2.83791428e-03 2.18494167e-03 -3.50605309e-01
-9.68238771e-01 -4.67720419e-01 6.60121441e-01 2.56106555e-01
-4.02428448e-01 4.93853867e-01 2.27209374e-01 5.33341654e-02
8.82034078e-02 -6.24485254e-01 -1.31469345e+00 -7.42675066e-01
-2.01195925e-01 5.22218466e-01 -1.89800039e-02 2.22244143e-01
4.07998741e-01 6.62638545e-01 -1.52175319e+00 9.84234735e-02
3.69626403e-01 1.11911726e+00 -3.53721201e-01 8.16060483e-01
2.93551683e-01 5.74906886e-01 -6.55710161e-01 -2.48064250e-01
5.39845705e-01 -2.46429488e-01 -5.37202597e-01 -2.77715102e-02
7.28717089e-01 -7.12229908e-01 -4.27785754e-01 1.08778262e+00
3.44700992e-01 2.40520284e-01 8.55274737e-01 -1.29500544e+00
-2.96932131e-01 1.39633879e-01 -1.26253933e-01 1.37442306e-01
3.35866004e-01 6.00640103e-02 1.00773716e+00 -1.34123158e+00
7.92929530e-01 7.13365316e-01 1.02529967e+00 2.97658164e-02
-1.22214234e+00 -1.54513016e-01 -6.72326460e-02 1.73024222e-01
-1.45288754e+00 -4.55636293e-01 9.12061632e-01 -6.76645219e-01
8.47214282e-01 -8.42832550e-02 7.58525133e-01 1.09011996e+00
2.47076780e-01 8.65580440e-01 5.39303422e-01 -3.64585102e-01
4.87053365e-01 6.23943135e-02 -6.94122612e-02 6.53142989e-01
2.86207974e-01 1.92971796e-01 -4.09653395e-01 -3.02836180e-01
6.65060103e-01 1.97827548e-01 -2.73474306e-01 -7.10893810e-01
-1.19569981e+00 1.00383937e+00 5.70041299e-01 2.33974680e-01
-5.14131248e-01 3.59573334e-01 2.07865179e-01 2.68334270e-01
4.13576514e-01 4.45515990e-01 -2.85420179e-01 -3.76088232e-01
-1.12569153e+00 2.68544585e-01 9.95178699e-01 7.69867122e-01
8.71223748e-01 4.89231497e-02 2.54896075e-01 4.18782890e-01
1.35678887e-01 3.54072034e-01 5.52563727e-01 -1.07900059e+00
7.57275075e-02 4.05550450e-01 1.22211322e-01 -1.05355990e+00
-3.86173159e-01 -4.21930224e-01 -4.42099541e-01 4.46516722e-01
5.76345086e-01 -6.36548102e-02 -5.75541079e-01 1.66257048e+00
3.75437468e-01 5.37842333e-01 2.13924404e-02 6.86477184e-01
2.47716248e-01 6.56899333e-01 -6.80079579e-01 1.64837152e-01
8.79530132e-01 -7.54722059e-01 -2.72875130e-01 8.12102854e-02
7.18668997e-01 -6.85384691e-01 7.19675779e-01 4.97407138e-01
-8.94128740e-01 -3.28518152e-01 -1.26552999e+00 -2.75044162e-02
-2.40682527e-01 3.18366498e-01 2.38933295e-01 1.65420115e-01
-7.87990272e-01 1.07734108e+00 -1.13138771e+00 -3.08162838e-01
3.36255431e-01 5.29182665e-02 -4.88448948e-01 9.12657231e-02
-7.46930838e-01 8.76875401e-01 4.38308120e-01 -1.76282674e-01
-1.00062358e+00 -7.34862804e-01 -8.66154194e-01 8.89033526e-02
6.59367740e-02 -7.10719526e-01 1.33164370e+00 -9.06655908e-01
-1.12927020e+00 6.02438867e-01 -8.64886194e-02 -7.52102971e-01
5.86517155e-01 -3.55749369e-01 -2.05882028e-01 4.30171579e-01
2.22854335e-02 5.02873480e-01 1.12216985e+00 -1.01095700e+00
-5.71146607e-01 -1.35198534e-01 -1.69888705e-01 -6.90671662e-03
-1.97750300e-01 -3.54686618e-01 -2.06706122e-01 -6.82693660e-01
2.74965316e-01 -1.03611958e+00 1.30638167e-01 4.03980017e-01
-5.46925306e-01 -9.69537795e-02 7.43923247e-01 -6.46748066e-01
9.00170386e-01 -2.22505474e+00 -7.46129677e-02 3.12904775e-01
2.28702545e-01 -5.18611213e-03 1.80324852e-01 5.43649793e-01
-1.99408039e-01 -1.02968581e-01 -2.68067807e-01 -6.01101279e-01
1.28104568e-01 1.09373569e-01 -7.21728504e-01 8.71294379e-01
3.39203715e-01 5.47877967e-01 -7.82097757e-01 3.75352614e-02
2.13594615e-01 5.46652377e-01 -6.31227314e-01 3.36846471e-01
-1.95404515e-01 1.67603150e-01 1.45922659e-03 3.99001837e-01
5.50923288e-01 -6.36860654e-02 -3.87813747e-01 -3.58735055e-01
8.99060816e-02 4.97961223e-01 -1.44901216e+00 1.76252532e+00
-2.51108706e-01 7.89260864e-01 -4.24700618e-01 -7.50048399e-01
7.64880538e-01 1.44420296e-01 6.15163445e-01 -3.53234299e-02
1.78887829e-01 4.84289318e-01 -2.57727921e-01 -3.16055268e-01
4.92073029e-01 -1.03042357e-01 2.71362960e-01 4.60281223e-01
4.07450795e-01 2.08837718e-01 -9.72134247e-03 -6.19387999e-02
1.22173643e+00 2.49151841e-01 1.89008921e-01 -4.40319106e-02
9.46723521e-02 -1.31900012e-01 6.02504134e-01 6.86585665e-01
-4.39912416e-02 8.55971515e-01 6.40944958e-01 -5.81196129e-01
-1.27690876e+00 -1.33595884e+00 -2.42844924e-01 3.63283753e-01
-8.80737081e-02 -4.30809885e-01 -5.92678249e-01 -7.45838523e-01
7.16334879e-02 9.33783591e-01 -5.93379617e-01 -4.15894806e-01
-4.28865135e-01 -3.08642507e-01 4.22018826e-01 5.47658920e-01
2.36246258e-01 -5.94786465e-01 -6.61402345e-01 1.33798182e-01
3.39079015e-02 -9.14801359e-01 -3.87514889e-01 1.75991178e-01
-9.37852800e-01 -1.16676867e+00 -4.94614035e-01 -5.52299142e-01
7.44523227e-01 -8.30216408e-02 9.20232058e-01 -2.16073915e-01
-2.16018975e-01 4.30303693e-01 -1.81853145e-01 -3.41528237e-01
-3.35020483e-01 -9.93487760e-02 2.65175968e-01 1.30370736e-01
5.29611826e-01 -7.56447673e-01 -5.91899693e-01 4.05013282e-03
-8.43225241e-01 -4.59866703e-01 3.39871973e-01 8.73146832e-01
7.54782081e-01 3.23945135e-01 2.77423412e-01 -5.55866301e-01
2.77391940e-01 -6.28402948e-01 -5.87945342e-01 -1.30430236e-01
-3.91042948e-01 3.32879722e-01 7.58562088e-01 -4.58450913e-01
-4.30616379e-01 1.94268361e-01 -3.23160619e-01 -8.70594919e-01
-2.36490875e-01 5.12750566e-01 4.96495105e-02 -3.88432033e-02
6.84241056e-01 1.19116403e-01 -6.69662580e-02 -6.10651016e-01
2.16032937e-01 3.38323027e-01 7.03635395e-01 -5.69673419e-01
8.50129545e-01 8.32619607e-01 8.04519877e-02 -6.62943602e-01
-7.96623766e-01 -5.93763113e-01 -7.43253171e-01 -2.17875242e-02
5.12733102e-01 -9.71212387e-01 -5.48175097e-01 3.29845399e-01
-1.17857969e+00 -3.95912640e-02 -6.72195137e-01 8.25580776e-01
-7.79348612e-01 2.32770696e-01 -3.77033949e-01 -6.88141882e-01
-1.38014570e-01 -1.03213024e+00 1.12556636e+00 -1.25937417e-01
-2.26522401e-01 -9.30880606e-01 4.05022025e-01 -3.59028786e-01
1.83399737e-01 5.15777230e-01 6.62650585e-01 -9.52896297e-01
-6.29291177e-01 -8.10117483e-01 1.97626576e-01 5.27853191e-01
-1.78512305e-01 2.80365169e-01 -1.18083990e+00 -4.56424713e-01
3.66078645e-01 -1.09799594e-01 8.26526046e-01 1.93111494e-01
1.10598826e+00 -4.82941121e-01 -1.74197420e-01 8.30416501e-01
1.48630786e+00 -2.17544153e-01 4.73683000e-01 3.41737896e-01
6.19118273e-01 2.82999873e-01 3.47168088e-01 5.88703752e-01
3.30983728e-01 7.11578906e-01 4.26260531e-01 2.42014661e-01
4.40333560e-02 -6.02432847e-01 5.21537960e-01 7.10723698e-01
3.31412435e-01 6.64841235e-02 -8.46743524e-01 6.85097754e-01
-1.87318051e+00 -9.87995446e-01 1.93612829e-01 2.52234340e+00
4.76129383e-01 2.31375024e-01 2.18730316e-01 1.87349677e-01
4.47896302e-01 -1.78130612e-01 -6.99639380e-01 -1.93780974e-01
-9.18583386e-03 3.16338526e-04 4.92027193e-01 6.08365953e-01
-1.10345590e+00 6.26418889e-01 6.49876595e+00 3.40441346e-01
-1.29193461e+00 -1.18039772e-01 2.48156324e-01 -2.70961106e-01
-1.58529982e-01 2.01661110e-01 -7.52639413e-01 4.74716038e-01
1.05648065e+00 -7.38368928e-03 3.73090893e-01 1.20646489e+00
-1.42589927e-01 9.58792958e-03 -1.56469893e+00 1.05304444e+00
3.79509121e-01 -1.23878181e+00 -6.03387058e-02 1.19965076e-01
5.79912126e-01 4.07097310e-01 1.20794624e-01 2.07620189e-01
8.25742334e-02 -6.99222386e-01 9.16110814e-01 9.42951500e-01
4.98602986e-01 -7.06079423e-01 6.38884962e-01 4.14545894e-01
-6.96484506e-01 9.29651558e-02 -5.57569861e-01 -8.07137191e-02
6.09623780e-03 9.01939631e-01 -1.30562103e+00 2.92611986e-01
6.85797095e-01 9.22710359e-01 -5.69905519e-01 1.36019015e+00
-3.43505561e-01 6.14578009e-01 -8.13840628e-01 2.69436717e-01
6.15236610e-02 -4.10242081e-02 1.05825841e+00 7.96453238e-01
7.12227941e-01 -5.80236256e-01 2.59904087e-01 9.88867104e-01
-1.64942056e-01 -2.42048010e-01 -7.96839654e-01 -1.04094362e-02
4.51538920e-01 9.98263001e-01 -3.08296472e-01 -1.69422269e-01
-3.10563862e-01 9.61922765e-01 4.25884813e-01 3.14316481e-01
-5.21312058e-01 -6.25476182e-01 7.29353011e-01 1.14799574e-01
6.89394593e-01 -3.77787083e-01 -1.77272603e-01 -1.55371428e+00
5.42666316e-01 -5.68210840e-01 2.38170579e-01 -9.31904435e-01
-1.24547434e+00 5.70780039e-01 -1.23778589e-01 -1.48127651e+00
-7.04069912e-01 -7.26341844e-01 -7.54600346e-01 5.42655885e-01
-1.29710817e+00 -8.37861121e-01 -1.86046794e-01 2.41181150e-01
1.15537681e-01 -2.07307965e-01 8.97591829e-01 2.45470434e-01
-3.98337722e-01 5.89800775e-01 5.10913432e-01 2.41494954e-01
7.38094091e-01 -1.37907958e+00 3.74933332e-01 1.10737848e+00
4.24395055e-01 6.75879657e-01 8.57722700e-01 -4.22811478e-01
-1.22718835e+00 -1.02690732e+00 8.19098353e-01 -7.09706485e-01
8.58491123e-01 -3.12827885e-01 -9.12666023e-01 9.77841973e-01
-4.12454903e-01 3.74474704e-01 4.86539662e-01 -5.42111248e-02
-5.58620274e-01 -1.40024960e-01 -1.19304645e+00 5.68145335e-01
6.05069876e-01 -6.75817847e-01 -6.37269258e-01 2.39998206e-01
6.02130830e-01 -4.50316966e-01 -9.68190610e-01 2.42651135e-01
4.60352242e-01 -1.14021099e+00 8.43900144e-01 -5.30474246e-01
4.08928275e-01 -5.26269972e-01 -4.68517005e-01 -1.32411253e+00
-8.32226221e-03 -6.28126323e-01 -5.57394981e-01 9.20528769e-01
2.37241670e-01 -5.98090231e-01 8.58146489e-01 4.81625050e-01
-3.43757182e-01 -1.02498734e+00 -1.23112917e+00 -8.52210462e-01
1.34141058e-01 -5.98363400e-01 6.94007158e-01 7.53403842e-01
-2.62541533e-01 1.19651295e-01 -2.25125104e-01 3.39122653e-01
5.91555417e-01 3.35598225e-03 9.29785609e-01 -1.27453327e+00
-4.76130843e-01 -2.21157089e-01 -9.70994890e-01 -1.21107137e+00
1.68038696e-01 -9.66736615e-01 3.17448378e-03 -1.19130480e+00
-1.91212937e-01 -2.00581506e-01 -3.03942710e-01 3.80816102e-01
1.88851550e-01 5.10343090e-02 -2.92814616e-02 4.79146302e-01
-5.35523593e-01 7.02341378e-01 4.79314834e-01 1.09942943e-01
-1.02721170e-01 8.02980959e-02 -3.92712623e-01 9.63947296e-01
6.62133574e-01 -6.41875029e-01 -8.58063400e-02 -4.20182079e-01
3.44528079e-01 -2.89888501e-01 7.15481281e-01 -1.35996258e+00
3.04893941e-01 3.21630865e-01 5.78584135e-01 -6.98100448e-01
4.44434196e-01 -9.65363681e-01 -9.42907110e-02 2.91624248e-01
-2.93266326e-01 -4.28785458e-02 4.56169620e-02 8.47408891e-01
-1.74538553e-01 -4.57651585e-01 7.87553847e-01 2.90120393e-01
-3.66026461e-01 3.49358201e-01 -9.78307128e-02 -9.59872268e-03
9.11154389e-01 -1.49847120e-01 -2.96686441e-02 -5.24053276e-01
-4.79735136e-01 -1.73576549e-01 8.83348405e-01 2.99983263e-01
6.87808990e-01 -1.56477427e+00 -6.54374361e-01 5.00408709e-01
3.13689411e-01 1.17837690e-01 -1.78040072e-01 8.58962119e-01
-6.32905245e-01 8.04035738e-02 1.41923755e-01 -8.34114134e-01
-7.22396672e-01 5.90719581e-01 4.38407987e-01 4.16751578e-02
-8.88866365e-01 5.96819818e-01 -1.68706700e-01 -5.01499474e-01
4.31808203e-01 -4.80383068e-01 3.57252479e-01 -1.64131999e-01
5.62093258e-01 3.85196924e-01 3.36040676e-01 -7.59141982e-01
-3.14098090e-01 3.30385357e-01 -2.06235293e-02 -1.03860475e-01
1.37267637e+00 2.09372178e-01 4.71313186e-02 8.54689837e-01
1.79016757e+00 1.05303386e-02 -1.50797880e+00 -3.34012002e-01
-4.18877974e-02 -5.46181917e-01 6.13258732e-03 -4.58176345e-01
-7.41948545e-01 8.55605602e-01 5.69156826e-01 -4.72158156e-02
8.75039756e-01 -1.71890602e-01 9.82002735e-01 7.89273083e-01
1.87889054e-01 -1.08060443e+00 4.26011272e-02 6.00363076e-01
8.90329838e-01 -1.23770094e+00 -4.41698171e-02 1.15261532e-01
-3.43993843e-01 1.44050992e+00 1.84112743e-01 -7.44158506e-01
8.98052394e-01 7.68381655e-02 -1.01837672e-01 -1.10713117e-01
-5.57871222e-01 8.75446126e-02 4.23353493e-01 4.16348040e-01
1.81743056e-01 -2.33604223e-01 2.46099547e-01 3.55424017e-01
-5.71260691e-01 -1.07479906e-02 6.55336916e-01 8.02196503e-01
-5.49446583e-01 -7.18115866e-01 -3.41552377e-01 4.62848216e-01
-4.53170687e-01 1.28483726e-02 -2.14616224e-01 6.40060604e-01
-7.02269152e-02 4.87784356e-01 2.34341577e-01 -4.63807136e-01
1.99607089e-01 2.02425614e-01 3.29443663e-01 -5.01304448e-01
-1.23567171e-01 8.53902176e-02 -1.87617183e-01 -6.21662140e-01
2.23450568e-02 -6.97930157e-01 -1.26868665e+00 -1.56230181e-01
-2.63319790e-01 1.10569552e-01 9.27212119e-01 8.51594567e-01
5.20125926e-01 1.98643550e-01 7.12299645e-01 -9.31763768e-01
-8.45993400e-01 -7.03648508e-01 -5.94433308e-01 4.89810705e-01
7.64551938e-01 -6.53685629e-01 -8.31670702e-01 5.35068102e-03] | [7.694599628448486, 2.348661422729492] |
5c0200f4-79d9-447a-abfd-0f53a267bf20 | cross-lingual-transfer-with-target-language | 2306.02767 | null | https://arxiv.org/abs/2306.02767v1 | https://arxiv.org/pdf/2306.02767v1.pdf | Cross-Lingual Transfer with Target Language-Ready Task Adapters | Adapters have emerged as a modular and parameter-efficient approach to (zero-shot) cross-lingual transfer. The established MAD-X framework employs separate language and task adapters which can be arbitrarily combined to perform the transfer of any task to any target language. Subsequently, BAD-X, an extension of the MAD-X framework, achieves improved transfer at the cost of MAD-X's modularity by creating "bilingual" adapters specific to the source-target language pair. In this work, we aim to take the best of both worlds by (i) fine-tuning task adapters adapted to the target language(s) (so-called "target language-ready" (TLR) adapters) to maintain high transfer performance, but (ii) without sacrificing the highly modular design of MAD-X. The main idea of "target language-ready" adapters is to resolve the training-vs-inference discrepancy of MAD-X: the task adapter "sees" the target language adapter for the very first time during inference, and thus might not be fully compatible with it. We address this mismatch by exposing the task adapter to the target language adapter during training, and empirically validate several variants of the idea: in the simplest form, we alternate between using the source and target language adapters during task adapter training, which can be generalized to cycling over any set of language adapters. We evaluate different TLR-based transfer configurations with varying degrees of generality across a suite of standard cross-lingual benchmarks, and find that the most general (and thus most modular) configuration consistently outperforms MAD-X and BAD-X on most tasks and languages. | ['Anna Korhonen', 'Ivan Vulić', 'Alan Ansell', 'Marinela Parović'] | 2023-06-05 | null | null | null | null | ['zero-shot-cross-lingual-transfer', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing'] | [ 4.76214960e-02 1.07488453e-01 -1.63642094e-01 -2.74938792e-01
-1.09508538e+00 -8.73101473e-01 7.98986852e-01 -2.22134963e-01
-5.41603446e-01 7.93171883e-01 2.41756335e-01 -7.31140673e-01
-1.02747999e-01 -6.77376568e-01 -7.73103237e-01 -5.03301263e-01
1.79158077e-01 7.19404519e-01 1.35575041e-01 -6.25992000e-01
-1.69066221e-01 2.77056396e-01 -1.40096080e+00 4.12202120e-01
9.54964399e-01 5.25142491e-01 3.36508602e-01 3.79757315e-01
-3.10568303e-01 4.39161211e-01 -4.40701574e-01 -7.76511610e-01
1.55775800e-01 -4.16632414e-01 -1.24516284e+00 -3.43002677e-01
3.72126132e-01 1.26585871e-01 6.86093643e-02 7.58161366e-01
4.59881693e-01 -8.33575055e-02 6.32851839e-01 -1.12230825e+00
-5.26513517e-01 8.23488474e-01 -1.17699131e-01 -1.47818411e-02
4.06074136e-01 2.33426467e-01 1.09654689e+00 -7.50863791e-01
6.24195635e-01 1.20999694e+00 8.45809519e-01 8.88861835e-01
-1.71640038e+00 -6.74281299e-01 1.36723667e-01 -2.45630279e-01
-1.44636250e+00 -7.52894342e-01 4.03463155e-01 -5.35906196e-01
1.24049377e+00 3.27354997e-01 1.90988198e-01 1.36794603e+00
1.92462474e-01 6.80356979e-01 1.34066415e+00 -5.58609247e-01
-8.93505812e-02 5.05472183e-01 2.35393450e-01 6.44319177e-01
-9.77074355e-02 1.95585668e-01 -5.12372196e-01 -1.57385737e-01
3.71135056e-01 -4.50358301e-01 -4.92868841e-01 -2.68366545e-01
-1.52025843e+00 7.05513179e-01 3.21456075e-01 5.75802505e-01
-3.09190410e-03 -6.12147413e-02 6.91756248e-01 1.01884747e+00
4.96743530e-01 7.01394558e-01 -8.13489974e-01 -1.78953737e-01
-5.47224343e-01 1.11734338e-01 9.31902885e-01 1.06313837e+00
1.03351951e+00 -1.00227423e-01 -2.81442881e-01 8.56836319e-01
-3.88920680e-02 1.76823705e-01 7.75826216e-01 -5.34792185e-01
7.51145601e-01 5.40353119e-01 -1.04222126e-01 -1.62060514e-01
-2.90718973e-01 -4.65710789e-01 -6.79698646e-01 1.92558065e-01
5.99341571e-01 -2.46210977e-01 -5.33020973e-01 2.29917836e+00
1.59403235e-01 1.39381168e-02 4.05798137e-01 6.26861691e-01
6.07905746e-01 4.53574508e-01 1.75256938e-01 1.15943521e-01
1.45969152e+00 -8.76539826e-01 -2.04350501e-01 -4.15065318e-01
1.23029470e+00 -8.00457418e-01 1.80502033e+00 -4.39092740e-02
-1.03719091e+00 -7.12681353e-01 -9.05457079e-01 -1.75367311e-01
-6.69358015e-01 -3.71203497e-02 5.34253240e-01 5.95877647e-01
-1.23748934e+00 5.33521354e-01 -4.15699393e-01 -6.06109440e-01
-1.26530677e-01 4.45809752e-01 -7.40052700e-01 -7.98122399e-03
-1.31654882e+00 1.16108441e+00 4.87602293e-01 -3.54364693e-01
-8.78498375e-01 -9.22870517e-01 -7.31377304e-01 1.25061572e-01
1.94476441e-01 -1.13186491e+00 1.35032654e+00 -1.33533144e+00
-1.65510631e+00 1.25947344e+00 -1.10441461e-01 -2.47957274e-01
4.37494010e-01 -2.33977526e-01 -2.89519191e-01 -3.71974170e-01
7.43007287e-02 7.11194873e-01 6.39712393e-01 -1.17027867e+00
-5.07012129e-01 -2.75070429e-01 4.72204477e-01 1.74047932e-01
-4.41066265e-01 2.23604336e-01 -3.76214445e-01 -6.44950688e-01
-4.50452954e-01 -1.07737231e+00 2.07089752e-01 -2.59071410e-01
-1.87839344e-01 -3.12014937e-01 4.38549727e-01 -3.57217729e-01
1.10170698e+00 -2.29034758e+00 5.27667046e-01 -1.87316924e-01
-4.40621153e-02 4.49604392e-01 -4.75608170e-01 5.23730099e-01
-2.29601800e-01 -1.00341409e-01 -2.89230525e-01 -8.02650332e-01
1.16687141e-01 5.31083465e-01 -4.02839243e-01 1.58353746e-01
2.53694087e-01 1.09687984e+00 -9.14994419e-01 -4.05479997e-01
-4.50710543e-02 4.02470261e-01 -7.63652742e-01 5.14466763e-01
-3.10745448e-01 6.31514847e-01 -1.94180906e-01 9.58920792e-02
3.10969323e-01 -5.44270016e-02 1.90212712e-01 -1.75633773e-01
-1.02083892e-01 8.01918447e-01 -8.96773160e-01 1.95203555e+00
-1.07000422e+00 3.84841233e-01 2.08690286e-01 -8.68984163e-01
7.61439264e-01 5.67873180e-01 3.82962339e-02 -6.41742170e-01
-9.70385522e-02 3.81571233e-01 4.72810864e-02 -3.68625998e-01
2.57467419e-01 -5.96874297e-01 -2.89255291e-01 6.73547804e-01
4.22414631e-01 1.77360311e-01 -4.38131317e-02 3.28226462e-02
9.33026314e-01 4.48794276e-01 3.88735086e-01 -7.20884979e-01
8.22303295e-01 -2.43954599e-01 2.58142263e-01 6.02438271e-01
1.57945737e-01 2.82012135e-01 3.75595331e-01 -2.79099911e-01
-7.56999135e-01 -1.34708786e+00 -3.83927627e-03 1.75379384e+00
-3.65354717e-01 -3.85671675e-01 -8.98623526e-01 -8.77182245e-01
5.61334305e-02 9.63510752e-01 -7.05015004e-01 -3.29438269e-01
-7.41937935e-01 -2.84942120e-01 9.88938332e-01 4.43883419e-01
2.50090599e-01 -1.04254293e+00 -3.34981173e-01 1.93325534e-01
-3.41930360e-01 -8.97537649e-01 -6.29444242e-01 3.64814043e-01
-5.89111209e-01 -8.25639248e-01 -4.81715500e-01 -7.31615901e-01
5.65066397e-01 1.14767373e-01 1.52885222e+00 7.33764172e-02
3.30812961e-01 4.09086287e-01 -2.60560632e-01 -1.51639983e-01
-8.32837582e-01 7.37313032e-01 7.65060857e-02 1.24430783e-01
2.93502897e-01 -6.65587008e-01 -1.01570219e-01 4.22675282e-01
-8.29037607e-01 7.65860826e-02 6.68773413e-01 8.03490520e-01
9.52037051e-02 -5.32589674e-01 6.80691540e-01 -1.04015803e+00
7.05494940e-01 -5.36587715e-01 -4.65588838e-01 2.90587395e-01
-4.18720812e-01 3.66323113e-01 9.85755444e-01 -5.59613228e-01
-1.04921985e+00 -3.81292969e-01 -4.31897432e-01 -2.44994849e-01
-2.25973949e-01 4.32078481e-01 -5.62499464e-01 -1.99712869e-02
6.93514943e-01 3.18318844e-01 1.39315978e-01 -9.24029410e-01
5.28880477e-01 7.42743134e-01 4.78817999e-01 -1.22579575e+00
6.05363429e-01 1.03658728e-01 -3.63870084e-01 -5.21239877e-01
-8.05047452e-01 -2.18758628e-01 -6.13934457e-01 2.10534245e-01
7.13308334e-01 -8.17676187e-01 -6.33616209e-01 3.75212997e-01
-1.27383244e+00 -8.16813767e-01 -4.69517410e-01 3.94726306e-01
-6.79088354e-01 5.97863570e-02 -6.09566927e-01 -1.47308916e-01
-2.35458910e-01 -1.41950083e+00 1.26884079e+00 -4.05720979e-01
-4.62555915e-01 -1.36894310e+00 2.56306201e-01 1.00628704e-01
6.99446261e-01 -3.70101392e-01 1.26835775e+00 -7.37625778e-01
-3.07425171e-01 1.56409487e-01 -2.12716058e-01 5.99205673e-01
1.80951372e-01 -4.18708652e-01 -1.08524382e+00 -5.83316505e-01
9.44757313e-02 -2.32158497e-01 5.05947292e-01 -2.47325137e-01
6.39169693e-01 -3.60936701e-01 -3.76265168e-01 7.50110865e-01
1.42848337e+00 -3.10836643e-01 5.55860519e-01 3.64338577e-01
6.03221118e-01 6.34278595e-01 1.45059049e-01 -3.36655885e-01
6.11361027e-01 1.25597978e+00 -4.72068414e-02 -6.62786374e-03
-2.49039188e-01 -3.26002896e-01 8.32384765e-01 1.12497628e+00
9.09687430e-02 -9.39936005e-03 -9.17891622e-01 5.95092654e-01
-1.51561177e+00 -6.51634037e-01 2.06112638e-01 2.44625092e+00
1.27638113e+00 4.94114980e-02 2.75115013e-01 -6.10233732e-02
3.52368057e-01 -1.62139535e-01 -4.25842963e-02 -6.24898374e-01
2.01090164e-02 3.50765258e-01 7.51615241e-02 8.65347564e-01
-7.20173895e-01 1.07584131e+00 6.24571466e+00 8.26939046e-01
-1.33545649e+00 6.37282670e-01 7.73365647e-02 7.98771679e-02
-4.48687464e-01 2.55471319e-01 -1.11074245e+00 5.34740865e-01
1.00859749e+00 -3.02589029e-01 5.91679156e-01 5.63713372e-01
-3.29316765e-01 4.39444155e-01 -1.75812006e+00 4.16902393e-01
-1.56523228e-01 -1.07232046e+00 2.36631468e-01 6.42349187e-04
3.57214093e-01 3.62857759e-01 -1.14081576e-01 7.17223287e-01
3.89511377e-01 -9.65910316e-01 8.10326278e-01 3.29097569e-01
1.04480720e+00 -4.96158540e-01 5.40506244e-01 4.89576399e-01
-1.09240305e+00 1.29370503e-02 -7.21033961e-02 -1.91527590e-01
9.61199477e-02 2.40535975e-01 -5.42700768e-01 7.55862594e-01
4.53870744e-01 3.51877391e-01 -5.16880274e-01 4.22166049e-01
-1.45226434e-01 5.50350666e-01 -2.01061308e-01 3.71169686e-01
3.96175712e-01 -2.40696847e-01 5.85760176e-01 1.64532495e+00
2.91517437e-01 -4.28904444e-01 1.59171894e-01 8.26374531e-01
-1.19088553e-01 3.33387792e-01 -1.00027966e+00 3.71320009e-01
4.14182484e-01 1.00433743e+00 -2.53927819e-02 -4.15689796e-01
-7.37093806e-01 9.09893990e-01 8.37778509e-01 3.02197516e-01
-7.36370921e-01 -2.08806664e-01 1.03761518e+00 3.26282948e-01
1.96978197e-01 -2.32118651e-01 -6.81891888e-02 -1.18644392e+00
4.03045826e-02 -1.12652147e+00 4.60175633e-01 -6.40884519e-01
-1.39504015e+00 8.59201312e-01 2.10163221e-01 -1.09151483e+00
-3.10709089e-01 -5.97640693e-01 -5.26434958e-01 1.12374687e+00
-1.59294164e+00 -1.57275283e+00 2.19085142e-02 8.93386185e-01
3.91295254e-01 -3.41733992e-01 1.31217217e+00 6.45291626e-01
-2.35973462e-01 1.03506649e+00 -2.66722497e-02 5.30944839e-02
9.18018639e-01 -1.23461032e+00 5.42371392e-01 5.85278153e-01
8.76372457e-02 9.44348454e-01 5.10156333e-01 -1.62261650e-01
-1.34335411e+00 -1.32516718e+00 1.25476205e+00 -7.83824861e-01
8.15462828e-01 -6.66314304e-01 -1.14621079e+00 1.28780031e+00
3.54678720e-01 -3.03521901e-01 7.47623324e-01 5.84004998e-01
-8.25972915e-01 -3.68458182e-01 -8.33148599e-01 7.59504974e-01
1.16160250e+00 -9.08155322e-01 -8.31895232e-01 2.62094140e-01
8.78637135e-01 -1.48612723e-01 -1.08516800e+00 2.36570224e-01
4.07519817e-01 -1.05061221e+00 9.05859172e-01 -6.99691892e-01
1.66306227e-01 -2.47960627e-01 -3.00262660e-01 -1.52729011e+00
-1.89746827e-01 -8.94255042e-01 2.95639485e-01 1.56809640e+00
5.41183174e-01 -1.17810261e+00 2.60998785e-01 -3.73864211e-02
-4.44398791e-01 -6.15673482e-01 -1.25174332e+00 -1.10330760e+00
6.66628063e-01 -4.48969632e-01 8.87249231e-01 1.03207433e+00
6.43905252e-02 9.08725381e-01 -2.37202585e-01 -1.25002146e-01
2.40107238e-01 4.67425622e-02 1.19161248e+00 -1.12637496e+00
-8.13129008e-01 -6.17658496e-01 -4.81735431e-02 -1.17331100e+00
5.31047165e-01 -1.50041306e+00 -5.77504523e-02 -8.35624874e-01
-2.32080910e-02 -8.68244469e-01 -2.96686679e-01 7.43946016e-01
-2.63931543e-01 2.59065747e-01 1.55682534e-01 3.47949028e-01
-1.84569061e-01 3.86112154e-01 9.99432921e-01 5.31256981e-02
-1.06220290e-01 2.08534915e-02 -1.00376475e+00 4.19981360e-01
5.83589554e-01 -4.03548867e-01 -5.37629485e-01 -9.68871713e-01
2.97836155e-01 -1.67283401e-01 3.01414400e-01 -9.28847492e-01
-3.77995819e-02 1.51423365e-01 -4.03169811e-01 2.09801987e-01
2.73451716e-01 -7.62495160e-01 3.18794191e-01 3.12175989e-01
-3.37034494e-01 2.75513887e-01 3.73593509e-01 1.23859372e-03
-1.45799726e-01 -2.21108228e-01 9.39111829e-01 -9.27149504e-02
-4.00437444e-01 -2.85947099e-02 -1.85289890e-01 2.68643111e-01
8.61657023e-01 -5.45262061e-02 -4.09845173e-01 -2.80112773e-02
-5.22955656e-01 -8.53299275e-02 7.09521115e-01 6.82479739e-01
-3.89345251e-02 -1.26424837e+00 -9.10384119e-01 5.27788103e-01
5.53246200e-01 -2.08039075e-01 -1.52212486e-01 1.05788088e+00
4.50958908e-02 4.92129266e-01 -2.59970665e-01 -5.21323144e-01
-1.10875750e+00 7.46109545e-01 4.86029685e-01 -5.41335940e-01
-6.09589696e-01 7.34467924e-01 5.72151423e-01 -9.20598149e-01
-1.85148232e-02 -1.84945047e-01 1.84066743e-01 -9.57006589e-02
3.31575483e-01 -5.99091090e-02 2.94541985e-01 -8.04790378e-01
-3.26016784e-01 6.95901334e-01 -2.66938895e-01 -6.28786981e-02
1.01837134e+00 -1.14875883e-01 -2.41362095e-01 6.17641568e-01
1.26446700e+00 1.73765123e-01 -7.50605106e-01 -5.41780949e-01
6.34399354e-02 -1.30703628e-01 -3.89869928e-01 -7.59455025e-01
-8.07071030e-01 1.06210494e+00 1.84050590e-01 6.16156571e-02
1.01185322e+00 1.76505685e-01 6.76073670e-01 2.11109757e-01
5.68518102e-01 -6.35477424e-01 -2.64065772e-01 5.80872476e-01
8.39964747e-01 -8.49591613e-01 -3.76973003e-01 -1.88340366e-01
-5.71869969e-01 6.70709908e-01 5.83026111e-01 3.21680419e-02
4.88549113e-01 4.13321078e-01 3.19914490e-01 -6.81737140e-02
-9.59228396e-01 -2.65607029e-01 2.75579840e-01 7.17083752e-01
6.93027914e-01 1.44596606e-01 -1.05076529e-01 5.33174932e-01
-3.85549396e-01 2.70039085e-02 -1.43814161e-01 7.92366266e-01
2.16098893e-02 -1.55221474e+00 -1.98237509e-01 1.15674935e-01
-3.04817408e-01 -2.58845627e-01 -2.21550807e-01 1.05520272e+00
2.88188756e-01 5.70400417e-01 3.29089761e-02 -2.52381265e-01
5.45076549e-01 4.66352165e-01 6.60268784e-01 -7.12981522e-01
-1.04284847e+00 -3.79238665e-01 2.74592221e-01 -4.58643347e-01
-2.16594771e-01 -4.31030810e-01 -7.59193122e-01 -3.88810843e-01
-2.61629403e-01 2.83195198e-01 4.96666521e-01 1.23470521e+00
6.24271572e-01 2.95670778e-01 4.33710486e-01 -8.17845583e-01
-7.81152666e-01 -1.00584304e+00 -1.26714796e-01 6.88113093e-01
3.01040232e-01 -6.43630385e-01 -2.19486639e-01 7.36440765e-03] | [11.099739074707031, 9.889382362365723] |
ead1b84a-72a1-448b-8bd6-6b9bb53f79aa | graph-based-multi-ode-neural-networks-for | 2305.18687 | null | https://arxiv.org/abs/2305.18687v2 | https://arxiv.org/pdf/2305.18687v2.pdf | Graph-based Multi-ODE Neural Networks for Spatio-Temporal Traffic Forecasting | There is a recent surge in the development of spatio-temporal forecasting models in the transportation domain. Long-range traffic forecasting, however, remains a challenging task due to the intricate and extensive spatio-temporal correlations observed in traffic networks. Current works primarily rely on road networks with graph structures and learn representations using graph neural networks (GNNs), but this approach suffers from over-smoothing problem in deep architectures. To tackle this problem, recent methods introduced the combination of GNNs with residual connections or neural ordinary differential equations (ODE). However, current graph ODE models face two key limitations in feature extraction: (1) they lean towards global temporal patterns, overlooking local patterns that are important for unexpected events; and (2) they lack dynamic semantic edges in their architectural design. In this paper, we propose a novel architecture called Graph-based Multi-ODE Neural Networks (GRAM-ODE) which is designed with multiple connective ODE-GNN modules to learn better representations by capturing different views of complex local and global dynamic spatio-temporal dependencies. We also add some techniques like shared weights and divergence constraints into the intermediate layers of distinct ODE-GNN modules to further improve their communication towards the forecasting task. Our extensive set of experiments conducted on six real-world datasets demonstrate the superior performance of GRAM-ODE compared with state-of-the-art baselines as well as the contribution of different components to the overall performance. The code is available at https://github.com/zbliu98/GRAM-ODE | ['Chandan K Reddy', 'Parshin Shojaee', 'Zibo Liu'] | 2023-05-30 | null | null | null | null | ['traffic-prediction', 'spatio-temporal-forecasting'] | ['time-series', 'time-series'] | [-4.05200720e-01 -2.91448444e-01 -2.68726647e-01 -4.60621804e-01
2.37557366e-01 -2.74280280e-01 8.44997168e-01 -2.39016965e-01
2.06029013e-01 5.30256927e-01 3.88211489e-01 -6.39575183e-01
-4.82844055e-01 -9.74077404e-01 -5.70384085e-01 -4.03194070e-01
-4.65241969e-01 4.11385447e-01 6.12965345e-01 -7.77727306e-01
-6.45026043e-02 8.53013337e-01 -1.38605213e+00 2.41905019e-01
9.10295248e-01 9.47593451e-01 -1.75026402e-01 5.66285849e-01
-3.10071588e-01 1.25540662e+00 -7.16356784e-02 -5.02310216e-01
3.04670036e-01 -2.28253916e-01 -4.61056411e-01 -1.96221069e-01
3.19828093e-01 -9.74307135e-02 -1.20006418e+00 6.51228845e-01
2.33017504e-01 6.47733927e-01 6.28694177e-01 -1.70151627e+00
-8.78028631e-01 3.71120125e-01 -3.63376051e-01 5.44932485e-01
-1.24236517e-01 3.85055542e-01 9.50809598e-01 -6.57881677e-01
4.88653183e-01 1.33979118e+00 9.17709231e-01 2.94997722e-01
-1.05009997e+00 -6.96759701e-01 6.35005414e-01 5.63295126e-01
-1.28290284e+00 -3.48119140e-01 9.63839054e-01 -5.43100774e-01
1.27170670e+00 1.21583782e-01 5.34025908e-01 1.27074826e+00
4.92337078e-01 5.67356348e-01 6.08601809e-01 3.82610977e-01
-7.84380585e-02 -2.65648276e-01 3.30658793e-01 7.32324243e-01
8.13884884e-02 2.23775759e-01 -3.62348944e-01 1.51640415e-01
7.89449811e-01 3.94629449e-01 1.43893480e-01 -2.42798597e-01
-7.90525496e-01 8.43756020e-01 9.98216391e-01 4.59050417e-01
-4.12325561e-01 4.70127612e-01 5.08062422e-01 3.32636833e-01
7.68564761e-01 -6.13090098e-02 -2.47066125e-01 -4.31953669e-02
-6.87809110e-01 2.40033746e-01 7.95179605e-01 7.59169817e-01
7.94092953e-01 3.50872964e-01 -6.69587776e-02 7.18677342e-01
2.53459692e-01 2.14317694e-01 5.59203178e-02 -4.94061440e-01
8.10837865e-01 9.72660959e-01 -4.50172186e-01 -1.70901239e+00
-7.35866010e-01 -6.79758847e-01 -1.34365046e+00 9.79337841e-02
2.71729052e-01 -1.86249331e-01 -9.86343503e-01 1.64502561e+00
1.34879246e-01 7.02800512e-01 -2.60643244e-01 8.05621624e-01
1.01857316e+00 8.62995267e-01 1.59500912e-01 4.41183478e-01
8.22635293e-01 -1.29042006e+00 -6.57625973e-01 -3.76093566e-01
6.80584729e-01 -2.39505842e-01 6.14085853e-01 -1.98596552e-01
-7.09758520e-01 -7.05124319e-01 -7.51429081e-01 -9.03315917e-02
-1.01550519e+00 -4.82085943e-02 8.60269547e-01 1.88323945e-01
-1.23144984e+00 6.30909979e-01 -8.06205451e-01 -3.41730416e-01
4.29519355e-01 2.63657331e-01 -2.42849410e-01 -4.04071249e-02
-1.62007642e+00 8.95409584e-01 6.89393356e-02 4.88865763e-01
-4.97467667e-01 -7.89101303e-01 -9.42928195e-01 1.61581293e-01
4.48947132e-01 -6.26376808e-01 6.22419477e-01 -8.24193299e-01
-1.31102204e+00 2.51206756e-01 -1.70157433e-01 -4.87623245e-01
5.70852876e-01 -2.34659780e-02 -1.00919390e+00 -2.23950297e-01
-7.61373788e-02 3.02296966e-01 5.48975885e-01 -8.59974504e-01
-4.40924376e-01 -6.57994226e-02 2.20405757e-01 -1.57191381e-01
-1.14249885e-01 -1.36326224e-01 -4.56031799e-01 -8.20640326e-01
-8.52550790e-02 -9.66016233e-01 -4.91965413e-01 -8.38845149e-02
-3.10458481e-01 -5.13980210e-01 1.06845224e+00 -5.64746797e-01
1.73848403e+00 -1.94068265e+00 -1.17024221e-03 3.79492640e-01
5.30680776e-01 6.02712452e-01 -3.34023595e-01 8.74630034e-01
-2.10506544e-01 5.44165596e-02 -1.30021617e-01 -3.11746091e-01
8.69321898e-02 4.84075636e-01 -3.41367275e-01 3.04620892e-01
4.40461099e-01 1.34990573e+00 -8.94883096e-01 -7.46547580e-02
4.31497097e-01 7.51718760e-01 -2.91544884e-01 -6.26147613e-02
-2.37894639e-01 4.76845026e-01 -4.78650212e-01 4.07483190e-01
7.03336656e-01 -5.00696599e-01 -4.81820107e-02 -8.46271589e-02
-1.56976849e-01 4.48105752e-01 -1.16324282e+00 1.30013549e+00
-4.15856391e-01 7.98332930e-01 -1.87016770e-01 -1.29045999e+00
9.27180946e-01 1.49241716e-01 6.57442808e-01 -9.56061602e-01
4.36833091e-02 4.43027541e-02 8.34734887e-02 -6.10012472e-01
4.31037903e-01 2.78205872e-01 1.50857717e-01 3.14748615e-01
-8.76027718e-02 4.90348190e-01 4.26214933e-01 3.61905783e-01
1.30210102e+00 2.31917780e-02 -2.06827715e-01 -1.87214151e-01
6.79790378e-01 -7.85811022e-02 6.37255847e-01 3.73941779e-01
-8.05138946e-02 3.83808643e-01 6.96965694e-01 -9.83164132e-01
-7.27382004e-01 -8.53763223e-01 7.18832240e-02 1.05549300e+00
-7.23083550e-03 -5.37327349e-01 -4.28223275e-02 -7.45522559e-01
2.85061061e-01 5.97459733e-01 -6.71244979e-01 -1.11007497e-01
-9.39848304e-01 -7.19940960e-01 6.09990120e-01 8.05907905e-01
5.39687812e-01 -8.82326961e-01 -1.18224479e-01 4.23654854e-01
-2.38303114e-02 -1.21283269e+00 -3.84301573e-01 -2.49327138e-01
-7.56365120e-01 -9.80474055e-01 -4.90595907e-01 -4.28671032e-01
5.11025548e-01 5.11905909e-01 1.19295752e+00 2.09590092e-01
-5.65768033e-02 1.07107997e-01 -2.45218486e-01 -8.15514103e-02
-1.12671159e-01 3.83777797e-01 -1.46144554e-01 3.68402600e-01
4.43872273e-01 -9.80380356e-01 -6.39297009e-01 5.73402584e-01
-7.68520296e-01 1.87309682e-01 5.16330659e-01 5.41610479e-01
6.14890419e-02 8.58069360e-02 7.05204844e-01 -8.22198808e-01
7.24034309e-01 -1.05800033e+00 -4.52452540e-01 7.45946914e-02
-5.48046291e-01 2.19936687e-02 8.12826753e-01 -2.71554887e-01
-9.20052528e-01 -3.63748193e-01 -1.77201629e-01 -5.60875595e-01
-2.58654445e-01 8.51135731e-01 2.23806784e-01 -5.44489212e-02
4.81269479e-01 -3.62818711e-03 5.27985170e-02 -4.21809435e-01
3.57762784e-01 2.03489840e-01 1.92867890e-01 -3.50637704e-01
8.20304215e-01 4.55212891e-01 3.93309861e-01 -7.54315794e-01
-6.28052771e-01 -3.65490466e-01 -5.77830732e-01 -4.69942361e-01
6.81219280e-01 -9.22922432e-01 -6.47833765e-01 4.87364650e-01
-9.33752239e-01 -6.60152912e-01 3.91988829e-02 3.28743666e-01
-9.39219892e-02 2.57273018e-01 -6.67005777e-01 -6.83511078e-01
1.24315321e-02 -8.55119884e-01 7.10202694e-01 1.12166986e-01
-3.39144133e-02 -1.53418338e+00 6.78666010e-02 1.01678371e-01
1.06231451e+00 5.48975945e-01 8.97111773e-01 -4.76098746e-01
-6.98043108e-01 -2.53515482e-01 -5.31148016e-01 1.43873677e-01
9.61957723e-02 1.50644273e-01 -6.08956337e-01 1.92110799e-02
-6.50885344e-01 1.72266245e-01 1.11389661e+00 5.12121081e-01
1.15507519e+00 -3.30642939e-01 -4.40506130e-01 6.62057161e-01
1.24042571e+00 -2.68377792e-02 7.03885555e-01 1.14332631e-01
1.09480965e+00 8.01012814e-01 6.77423924e-02 1.81798607e-01
9.52939332e-01 6.19760215e-01 5.49778044e-01 -2.82133192e-01
-4.14060146e-01 -2.01734379e-01 4.22253966e-01 9.85135972e-01
-4.33344781e-01 -6.72627866e-01 -1.23988819e+00 6.61784291e-01
-2.45601797e+00 -1.19303155e+00 -7.89309323e-01 1.57060313e+00
5.61008938e-02 2.47894287e-01 3.37781578e-01 -1.33794606e-01
6.28521621e-01 6.42834783e-01 -6.61097288e-01 -5.50645411e-01
-1.77157238e-01 -4.36374806e-02 5.50635159e-01 4.22343582e-01
-1.04230523e+00 9.93551850e-01 5.52019310e+00 5.99306583e-01
-1.26606071e+00 -1.12261325e-02 5.23768365e-01 8.27617757e-03
-2.28504404e-01 -9.12743658e-02 -7.20871568e-01 5.37352264e-01
1.30070388e+00 2.88812723e-03 4.70341891e-01 5.57071507e-01
3.35245848e-01 3.36536437e-01 -7.62593031e-01 7.73078203e-01
-1.95550859e-01 -1.61342621e+00 1.62102312e-01 1.77220285e-01
8.58487010e-01 5.95883965e-01 -5.71351498e-03 4.92398947e-01
6.39224946e-01 -1.16476893e+00 3.43739718e-01 9.42458212e-01
4.31088895e-01 -7.20743895e-01 6.33143842e-01 2.01213241e-01
-1.85555816e+00 -1.14913844e-01 -2.92288691e-01 -4.08025682e-01
4.81318355e-01 7.20136106e-01 -2.52813458e-01 9.77296293e-01
5.85393965e-01 1.48961806e+00 -6.43553674e-01 8.76650691e-01
-2.00110599e-01 7.16100574e-01 -2.56131530e-01 4.50808816e-02
7.95183063e-01 -5.36142647e-01 5.42301118e-01 1.36144578e+00
2.24536180e-01 -1.59604564e-01 1.81697369e-01 9.10857379e-01
7.43344007e-03 -1.91104919e-01 -9.91261542e-01 -3.63760069e-02
1.25778615e-01 1.20252895e+00 -5.19165397e-01 -2.86285907e-01
-8.28655005e-01 5.43648005e-01 4.88056004e-01 7.08530009e-01
-1.13772678e+00 -3.10408562e-01 1.01213384e+00 3.58142257e-01
3.55035782e-01 -6.55052125e-01 -1.87742248e-01 -1.13155758e+00
4.21862565e-02 -4.60164726e-01 4.86952066e-01 -4.88041729e-01
-1.71853542e+00 7.69556761e-01 6.37442665e-03 -1.23286414e+00
-2.08758712e-01 -6.48055553e-01 -9.51949596e-01 8.20082605e-01
-1.67996120e+00 -1.37447095e+00 -3.80797744e-01 7.32505798e-01
4.50634480e-01 -2.07723781e-01 3.72832626e-01 7.25063980e-01
-8.49179149e-01 3.93708974e-01 1.63703989e-02 3.04948866e-01
2.93783456e-01 -8.97734821e-01 1.05833042e+00 9.56793010e-01
-9.13479552e-02 4.18960184e-01 2.77228475e-01 -7.18838990e-01
-1.30706537e+00 -1.47004497e+00 1.22016430e+00 -3.97548616e-01
1.17086792e+00 -5.32801390e-01 -1.09168494e+00 9.48197126e-01
1.13819048e-01 4.75386888e-01 3.48547965e-01 2.56255895e-01
-3.99041355e-01 -2.97018915e-01 -6.58359706e-01 6.55085504e-01
1.58668268e+00 -4.99708593e-01 -1.01412028e-01 2.44739845e-01
4.33604121e-01 -2.51589537e-01 -7.52195716e-01 4.11943495e-01
4.92094070e-01 -9.67240632e-01 9.26515222e-01 -9.05888319e-01
4.14690226e-01 -3.41728508e-01 1.76080391e-01 -1.34491396e+00
-7.81440675e-01 -4.60785121e-01 -5.05935192e-01 1.19100833e+00
4.41947430e-01 -1.09141850e+00 5.94739974e-01 6.33180022e-01
-3.16098809e-01 -9.82896984e-01 -8.43660414e-01 -9.01584864e-01
9.64481011e-02 -6.41058922e-01 8.83469880e-01 1.09237587e+00
-3.91701758e-01 4.24688756e-01 -6.02854669e-01 1.11919269e-01
2.54897714e-01 -6.32027313e-02 8.87178659e-01 -1.50290453e+00
5.46051078e-02 -7.14951813e-01 -7.58060455e-01 -1.14654648e+00
2.92353481e-01 -1.05290878e+00 -4.28967685e-01 -1.77908027e+00
-3.90805900e-01 -5.19074202e-01 -4.54392701e-01 4.59891915e-01
7.72749409e-02 -9.62847769e-02 1.32519409e-01 1.21983826e-01
-6.31012559e-01 7.07291543e-01 1.13481653e+00 -1.49858534e-01
-2.14761674e-01 1.66799903e-01 -4.29710537e-01 5.46108246e-01
8.90907526e-01 -3.47918123e-01 -6.00661635e-01 -7.58791566e-01
4.75575000e-01 -1.36464983e-01 6.30520463e-01 -1.18131316e+00
5.56375682e-01 -1.44731432e-01 1.18910611e-01 -7.05176890e-01
1.24789588e-01 -8.67933512e-01 4.67436969e-01 2.85974145e-01
-7.80991465e-02 5.75588226e-01 2.68308133e-01 7.75809526e-01
-3.03272843e-01 6.02357984e-01 2.90586054e-01 1.24599459e-02
-1.05636168e+00 7.65341878e-01 -6.13885045e-01 -5.56603447e-02
9.40576971e-01 -1.57960758e-01 -5.45560539e-01 -5.52203834e-01
-6.02208316e-01 6.41256392e-01 -7.30736274e-03 9.00957167e-01
3.99602860e-01 -1.56354308e+00 -6.92918003e-01 2.53912359e-01
-1.56552959e-02 -1.43905059e-01 4.69702452e-01 1.15006936e+00
-4.80383933e-01 5.87875247e-01 -1.99416935e-01 -4.61326838e-01
-7.07172513e-01 3.94932061e-01 4.38874274e-01 -6.37329698e-01
-9.07120228e-01 5.93995810e-01 1.01494573e-01 -6.39144897e-01
8.89617130e-02 -4.49493468e-01 -2.31397539e-01 7.28021637e-02
1.62529588e-01 7.67090261e-01 6.08140640e-02 -1.00309324e+00
-4.81373250e-01 5.25993466e-01 1.53982192e-01 3.67552519e-01
1.67245817e+00 -1.80858150e-01 -2.30130572e-02 4.23537016e-01
1.25853670e+00 -4.81564760e-01 -1.31945550e+00 -5.28815567e-01
9.21663791e-02 -2.77728230e-01 7.50282258e-02 -5.17784834e-01
-1.64623570e+00 1.04587948e+00 2.76192486e-01 4.40583467e-01
8.80634129e-01 -1.17015101e-01 1.05699468e+00 3.63395989e-01
-8.33658595e-03 -9.44029748e-01 -1.21505700e-01 9.53294694e-01
9.72553372e-01 -1.18375361e+00 -1.77481905e-01 -3.28269303e-01
-5.63184381e-01 1.28262055e+00 6.95429444e-01 -4.54502523e-01
1.20650578e+00 3.20791416e-02 -9.46706086e-02 -6.08708739e-01
-1.00947142e+00 -5.03518462e-01 7.16259837e-01 3.18491489e-01
3.03595126e-01 2.79901698e-02 -6.37412295e-02 3.31144899e-01
9.26301852e-02 -8.78065825e-03 7.68617243e-02 7.47812986e-01
1.10942740e-02 -8.85264337e-01 2.95118064e-01 5.19218385e-01
-1.12047710e-01 5.49952425e-02 -3.41476440e-01 9.63529527e-01
6.38171881e-02 1.02725077e+00 2.27660254e-01 -6.86795115e-01
4.88625318e-01 -2.27744319e-02 -1.78048134e-01 -1.98675841e-01
-6.59564853e-01 -3.56355041e-01 2.30870247e-01 -9.00293767e-01
-4.64423895e-01 -4.95575637e-01 -1.08538270e+00 -8.76528442e-01
1.50550716e-02 -1.64982110e-01 4.34015125e-01 9.00534332e-01
9.22289073e-01 9.93374944e-01 4.81910765e-01 -9.14985299e-01
3.58389765e-02 -8.82553995e-01 -3.71163040e-01 4.92857397e-01
4.66375500e-01 -8.99282515e-01 -2.96079904e-01 -3.26585084e-01] | [6.498448371887207, 2.10390567779541] |
060e4e65-259d-4961-bec3-9fcf38351f1b | task-oriented-conversational-modelling-with | 2303.17695 | null | https://arxiv.org/abs/2303.17695v1 | https://arxiv.org/pdf/2303.17695v1.pdf | Task Oriented Conversational Modelling With Subjective Knowledge | Existing conversational models are handled by a database(DB) and API based systems. However, very often users' questions require information that cannot be handled by such systems. Nonetheless, answers to these questions are available in the form of customer reviews and FAQs. DSTC-11 proposes a three stage pipeline consisting of knowledge seeking turn detection, knowledge selection and response generation to create a conversational model grounded on this subjective knowledge. In this paper, we focus on improving the knowledge selection module to enhance the overall system performance. In particular, we propose entity retrieval methods which result in an accurate and faster knowledge search. Our proposed Named Entity Recognition (NER) based entity retrieval method results in 7X faster search compared to the baseline model. Additionally, we also explore a potential keyword extraction method which can improve the accuracy of knowledge selection. Preliminary results show a 4 \% improvement in exact match score on knowledge selection task. The code is available https://github.com/raja-kumar/knowledge-grounded-TODS | ['Raja Kumar'] | 2023-03-30 | null | null | null | null | ['response-generation', 'keyword-extraction'] | ['natural-language-processing', 'natural-language-processing'] | [-2.08570510e-01 2.64698416e-01 -8.98722857e-02 -4.36284274e-01
-1.26967418e+00 -8.50182354e-01 6.37394071e-01 2.71457642e-01
-4.87218440e-01 9.18291807e-01 5.71870625e-01 -2.62755632e-01
-1.39830187e-01 -8.94532621e-01 -2.22390115e-01 -5.61492331e-02
5.90362966e-01 7.12358534e-01 4.09813523e-01 -4.75462914e-01
6.57371342e-01 2.11438820e-01 -1.31576729e+00 7.41687536e-01
1.02068758e+00 6.20565891e-01 2.56087661e-01 8.31332564e-01
-6.60333157e-01 8.28658104e-01 -7.13219941e-01 -5.91649115e-01
-5.51378317e-02 -2.14416146e-01 -1.46034670e+00 -2.26587832e-01
-4.50686961e-02 -3.08919966e-01 4.50839438e-02 5.32761395e-01
6.82305932e-01 3.95233959e-01 3.25695783e-01 -1.01563323e+00
-5.08890033e-01 7.95791864e-01 -4.86991275e-03 1.31350800e-01
8.66508722e-01 -1.83894083e-01 1.03145909e+00 -1.09113216e+00
7.52765477e-01 1.03210294e+00 2.72728741e-01 4.89760429e-01
-5.82940698e-01 -5.44233501e-01 -1.54700011e-01 4.81813848e-01
-1.50255811e+00 -6.73418701e-01 4.75768447e-01 -1.03018411e-01
1.44059145e+00 5.53174734e-01 1.93504632e-01 6.55921400e-01
-3.27331901e-01 9.53247845e-01 9.24774289e-01 -8.00107360e-01
6.52866364e-02 7.31182754e-01 4.70353961e-01 3.28332067e-01
2.45739520e-03 -4.29067224e-01 -8.43622208e-01 -3.59123290e-01
2.39426941e-01 -3.83013844e-01 -2.62113392e-01 1.57533020e-01
-9.44936156e-01 6.48227990e-01 1.18444107e-01 3.64617199e-01
-8.85911763e-01 -4.58213300e-01 3.19906324e-01 3.18147719e-01
1.89405993e-01 9.08768356e-01 -6.50392234e-01 -6.09922767e-01
-5.69060028e-01 3.22098374e-01 1.50808275e+00 1.09091628e+00
6.96565449e-01 -5.48566759e-01 -3.11120838e-01 1.10543370e+00
2.79987514e-01 3.69947881e-01 2.37535387e-01 -9.84913230e-01
6.11955762e-01 9.97706532e-01 6.04177237e-01 -8.85375261e-01
-2.24546313e-01 -1.91442043e-01 1.69496212e-04 -6.16725087e-01
1.90857515e-01 -3.30006510e-01 -4.21045780e-01 1.21481299e+00
5.06324649e-01 -2.55192339e-01 5.11145532e-01 6.88301980e-01
1.21185899e+00 6.14610612e-01 5.95195480e-02 -2.01633275e-01
1.52874696e+00 -1.02624321e+00 -9.42558050e-01 1.09394275e-01
8.16306651e-01 -1.35743356e+00 7.97164023e-01 1.48427650e-01
-1.08004940e+00 -2.06060544e-01 -6.26802802e-01 -6.13672845e-02
-4.86114621e-01 2.01309100e-01 6.73784554e-01 7.29255497e-01
-9.96184647e-01 -1.63356155e-01 -4.75698918e-01 -8.11048746e-01
-2.30195627e-01 3.37680846e-01 -5.33630513e-02 -2.40233690e-02
-1.49714458e+00 1.19355619e+00 2.98220485e-01 1.30690783e-02
-2.07024857e-01 -4.25370723e-01 -5.65304339e-01 3.46853621e-02
4.58995789e-01 -6.61936760e-01 1.63686204e+00 -5.59718549e-01
-1.62917101e+00 4.29673404e-01 -5.82909942e-01 -2.20360100e-01
7.41881430e-02 -2.95376539e-01 -4.95779097e-01 2.94776052e-01
1.51544705e-01 6.50875092e-01 3.00724041e-02 -1.03225446e+00
-1.03814518e+00 -6.23897370e-03 4.40561026e-01 6.18574679e-01
-2.34734282e-01 5.23832798e-01 -8.27910125e-01 -1.57042757e-01
-2.94334590e-01 -9.48759139e-01 4.54205312e-02 -7.46300638e-01
-3.38032991e-01 -6.12645328e-01 4.24454004e-01 -9.96493399e-01
1.69552326e+00 -1.66634762e+00 -3.88930559e-01 3.20822537e-01
-2.40638837e-01 6.23152733e-01 -3.87113094e-02 1.05877388e+00
3.80576104e-01 3.34369570e-01 3.26294899e-01 8.24700668e-02
3.82647775e-02 -8.90837982e-02 -1.10647701e-01 -4.12121564e-01
2.33331978e-01 9.32943046e-01 -7.05274343e-01 -5.56489468e-01
-6.34500384e-02 5.00401676e-01 -3.46364796e-01 3.04431498e-01
-2.21960112e-01 9.48471054e-02 -7.16888011e-01 7.46765018e-01
3.64066422e-01 -2.28918225e-01 3.52506578e-01 -4.58332486e-02
-1.22211240e-01 8.22822034e-01 -1.18312371e+00 1.21657526e+00
-7.05531418e-01 3.05607080e-01 9.91335139e-03 -4.12230283e-01
8.88520002e-01 7.10558891e-01 1.86174452e-01 -6.99213207e-01
8.62662494e-02 2.53610462e-01 -1.15779400e-01 -5.89947760e-01
8.79146516e-01 3.52481902e-01 -5.50381690e-02 6.20151818e-01
-2.74691194e-01 8.91112909e-02 2.67103821e-01 5.54611564e-01
1.25587046e+00 -1.21309549e-01 3.43467683e-01 9.00558606e-02
6.45574391e-01 5.94024777e-01 3.72974575e-01 7.62776434e-01
-7.26618394e-02 9.56186056e-02 8.34589154e-02 6.72353059e-02
-6.90783560e-01 -4.88304973e-01 5.47327846e-02 1.07038510e+00
3.93945798e-02 -5.86285710e-01 -7.70406187e-01 -6.36020839e-01
-1.26331002e-01 1.12500870e+00 -6.35714903e-02 1.51256556e-02
-5.07400572e-01 -4.32643652e-01 4.87703711e-01 4.70119774e-01
6.11986458e-01 -1.40862131e+00 -1.63870484e-01 4.20480341e-01
-7.32115686e-01 -1.08910549e+00 -4.95823413e-01 -1.03939950e-01
-4.38743711e-01 -1.03340030e+00 -4.11765069e-01 -7.87204742e-01
3.27379793e-01 3.89813721e-01 1.02509379e+00 1.02445506e-01
-4.03673537e-02 7.44921505e-01 -9.51473117e-01 -5.07867575e-01
-3.94980848e-01 4.79782492e-01 -2.61064947e-01 -2.51876235e-01
9.80187953e-01 -2.01697603e-01 -7.77086020e-01 5.21654785e-01
-5.88281631e-01 -1.33110672e-01 6.43529892e-01 5.83510637e-01
7.79245272e-02 -1.13389738e-01 1.13092017e+00 -8.08087647e-01
1.18445694e+00 -6.31245196e-01 -2.82454610e-01 6.37959778e-01
-9.16056931e-01 5.43467589e-02 5.22065908e-02 -1.78043276e-01
-1.55936980e+00 -1.55837135e-02 -2.09257677e-01 3.83885920e-01
-2.45545700e-01 8.30511510e-01 -8.79879668e-03 3.89255174e-02
7.01005101e-01 1.15083501e-01 -1.42526746e-01 -6.06498182e-01
3.44293952e-01 1.27717924e+00 -4.80120778e-02 -4.76589978e-01
2.94433713e-01 -2.82588247e-02 -8.39254260e-01 -6.90776467e-01
-5.84613860e-01 -1.15288508e+00 -3.54587466e-01 -2.34189942e-01
6.46826327e-01 -9.46132839e-01 -9.68897641e-01 1.74207911e-01
-1.21273792e+00 -6.23017550e-02 9.81775969e-02 5.10403633e-01
-2.21377045e-01 2.03085363e-01 -6.39627278e-01 -1.08958781e+00
-9.02783394e-01 -7.64714003e-01 7.75906622e-01 5.69410503e-01
-6.72578812e-01 -7.82694697e-01 1.19538762e-01 1.10269189e+00
6.49469137e-01 -4.90429521e-01 8.31181884e-01 -1.27455342e+00
-6.58731103e-01 -3.26732844e-01 -1.30996332e-01 1.62736416e-01
3.66314757e-03 -8.36131275e-02 -6.11016870e-01 6.39350489e-02
-4.05866444e-01 -2.59397417e-01 4.14817840e-01 -5.51170856e-02
2.93382943e-01 -3.85481387e-01 -4.20780003e-01 -3.08541209e-01
1.14870572e+00 6.37331247e-01 5.65326929e-01 5.27265310e-01
3.83242667e-01 9.62965369e-01 9.96770978e-01 4.07169044e-01
1.03948641e+00 8.47177327e-01 -2.09663257e-01 2.93179989e-01
-1.18122950e-01 -7.04478472e-02 2.50080138e-01 1.00600934e+00
1.94045141e-01 -3.43732506e-01 -1.04061580e+00 9.23894644e-01
-1.96797550e+00 -9.50861573e-01 -2.01666489e-01 2.08751440e+00
1.19275737e+00 -1.01832263e-01 1.27828509e-01 -2.90817589e-01
7.25282073e-01 -4.32152420e-01 -3.68935287e-01 -7.20858335e-01
5.68150319e-02 3.70645702e-01 2.64024496e-01 8.63978088e-01
-5.63147485e-01 1.20040464e+00 5.55603504e+00 5.63382983e-01
-7.74947464e-01 8.20229799e-02 2.36802846e-01 -1.61265191e-02
-4.22952354e-01 1.93614215e-01 -1.33181298e+00 2.23800808e-01
1.16246855e+00 -6.48693025e-01 2.76381433e-01 8.12184095e-01
3.18829119e-01 -4.05428201e-01 -6.56161606e-01 6.90811872e-01
-2.19911840e-02 -1.37321651e+00 6.91204816e-02 -1.46175055e-02
5.16253769e-01 -1.31448388e-01 -4.81411278e-01 7.00445890e-01
4.36341047e-01 -5.57701230e-01 8.73139650e-02 8.01389992e-01
7.09465370e-02 -9.86076832e-01 9.80091453e-01 4.21544164e-01
-1.03760111e+00 1.24642693e-01 -5.64731993e-02 2.48104244e-01
3.11679482e-01 3.71874928e-01 -1.67177606e+00 6.04463696e-01
6.08838677e-01 -3.83348949e-02 -3.86960983e-01 1.17016029e+00
-1.78599387e-01 6.59540057e-01 -3.40834558e-01 -4.63562667e-01
-4.42953873e-03 1.00453608e-01 4.31957334e-01 1.50512576e+00
1.85092866e-01 5.68727732e-01 2.49367997e-01 4.23924327e-01
-1.67590290e-01 6.72222078e-01 -2.88778722e-01 -1.46357372e-01
1.02380347e+00 1.27379286e+00 -4.10302669e-01 -4.59507555e-01
-4.04838443e-01 8.34146678e-01 2.26051882e-01 4.02715743e-01
-3.76620144e-01 -7.67857611e-01 4.33204114e-01 7.47000566e-03
4.33214128e-01 -4.24603447e-02 3.09533980e-02 -1.00560820e+00
2.81982869e-01 -1.13941979e+00 3.59818906e-01 -6.31323934e-01
-9.94399965e-01 6.39983058e-01 9.83441714e-03 -7.60600209e-01
-7.27202177e-01 -8.15032497e-02 -3.93625498e-01 9.83554482e-01
-1.37409961e+00 -1.04914892e+00 -2.24691108e-01 4.80022639e-01
6.33519948e-01 8.59905481e-02 1.04412663e+00 5.44289112e-01
-4.39131171e-01 6.70305908e-01 -4.44493368e-02 1.79951608e-01
1.03269041e+00 -1.13375640e+00 2.62595892e-01 5.42596400e-01
-4.89673167e-02 9.87485409e-01 5.22860646e-01 -8.60211074e-01
-1.42667544e+00 -7.92032242e-01 1.56597042e+00 -7.40440309e-01
4.62293774e-01 5.22484072e-03 -1.01511610e+00 3.58570606e-01
5.79627812e-01 -8.40536475e-01 1.08400285e+00 4.41265047e-01
-1.45059317e-01 -5.00186272e-02 -1.10757816e+00 4.73328888e-01
5.02777219e-01 -6.80319607e-01 -7.06286848e-01 3.85827839e-01
5.57323277e-01 -2.18004942e-01 -1.03469932e+00 2.39642620e-01
5.77731311e-01 -5.05930424e-01 6.16594791e-01 -5.04097879e-01
-5.98821715e-02 -3.74809086e-01 -8.82924646e-02 -1.02395439e+00
-1.20584918e-02 -6.79817438e-01 -1.06329113e-01 1.63349509e+00
1.04629755e+00 -5.63861370e-01 5.67977369e-01 1.24743629e+00
6.55821264e-02 -5.79321563e-01 -5.50595522e-01 -4.32427913e-01
-3.81155312e-01 -2.93590873e-01 4.63042527e-01 9.87113416e-01
6.32263780e-01 7.11849749e-01 -9.51586440e-02 2.53465831e-01
4.45534103e-02 1.53049141e-01 7.62223065e-01 -7.28883564e-01
5.45855351e-02 -1.35943890e-01 3.59405242e-02 -1.00217104e+00
-3.25796306e-02 -6.67408764e-01 1.18221238e-01 -1.95498276e+00
3.46692532e-01 -3.81863534e-01 -1.24194488e-01 5.87246716e-01
-2.60332465e-01 -2.77662426e-01 5.66908009e-02 1.38219312e-01
-9.52507913e-01 2.88207382e-01 7.87147343e-01 3.22198391e-01
-6.14385962e-01 1.27381966e-01 -9.65579033e-01 1.81142285e-01
1.09908354e+00 -4.08229768e-01 -3.68554801e-01 -1.92959875e-01
3.07758301e-01 2.38703787e-01 -1.63374394e-01 -5.74119925e-01
8.27814758e-01 -1.55974627e-01 -8.18731338e-02 -6.22092366e-01
4.33861941e-01 -4.19950217e-01 1.09207079e-01 -2.82491185e-02
-4.44250077e-01 1.40689328e-01 1.99376971e-01 2.25212425e-01
-4.36877668e-01 -3.67991239e-01 1.90711752e-01 -2.13617831e-01
-6.80156410e-01 -2.65719920e-01 -6.65165961e-01 -1.31468520e-01
7.67076135e-01 -6.62702462e-03 -5.69181204e-01 -8.30888510e-01
-5.97335577e-01 5.46178401e-01 1.39293224e-01 6.94339156e-01
4.99423236e-01 -9.89182293e-01 -7.12161839e-01 -4.13298845e-01
3.14773023e-01 -4.76688594e-01 1.19137503e-01 7.53383577e-01
-3.47486615e-01 1.03242791e+00 -4.22203727e-03 3.46219204e-02
-1.44017684e+00 1.89449772e-01 2.30713096e-02 -2.72468418e-01
9.22370702e-02 8.49552691e-01 -4.90595549e-01 -7.34132051e-01
2.70539939e-01 8.76944438e-02 -6.87470555e-01 2.77622610e-01
7.46982932e-01 3.98234308e-01 2.77669996e-01 -5.12140751e-01
-6.67463243e-01 2.05474421e-02 -5.42892516e-01 -3.68691832e-01
1.01772833e+00 -4.01337743e-01 -1.21549942e-01 1.88850220e-02
9.82328594e-01 3.59543294e-01 -2.59162277e-01 -5.24837911e-01
5.07757306e-01 -2.75994509e-01 -4.75892350e-02 -1.39805901e+00
-5.25618494e-01 3.39305520e-01 3.86663109e-01 2.72139490e-01
1.01175523e+00 7.20052719e-02 9.42572355e-01 8.93901348e-01
3.81190032e-01 -1.31252015e+00 -2.42659494e-01 6.97809994e-01
8.86344016e-01 -1.40995276e+00 -2.42624268e-01 -4.87970054e-01
-9.27507222e-01 8.52953434e-01 7.44974017e-01 5.11626899e-01
5.69579363e-01 -1.08562179e-01 5.50914407e-01 -2.64229625e-01
-1.20886028e+00 -5.38249373e-01 4.09336686e-01 2.18921334e-01
6.04684114e-01 -1.13165537e-02 -8.04937363e-01 6.21005714e-01
-1.77426353e-01 -3.15249041e-02 3.79376948e-01 1.20643687e+00
-5.26428282e-01 -1.51773059e+00 -9.92427990e-02 4.15264487e-01
-6.59353614e-01 -5.32422125e-01 -8.63895893e-01 4.79301572e-01
-4.93161470e-01 1.65107250e+00 -3.85127515e-01 -5.42594075e-01
6.85189366e-01 5.03669858e-01 -1.02005497e-01 -9.87538338e-01
-1.14302444e+00 6.72701001e-03 1.07359731e+00 -3.70388418e-01
-4.69313413e-01 -6.13810360e-01 -1.21351838e+00 -4.10403043e-01
-8.39811087e-01 8.33864510e-01 9.03930545e-01 7.75607586e-01
9.93004560e-01 -4.96031567e-02 5.65992355e-01 5.84884323e-02
-3.04559410e-01 -1.19082129e+00 4.25882749e-02 1.52471796e-01
-3.48644763e-01 -2.84712166e-01 -1.44705400e-01 9.54076424e-02] | [12.233174324035645, 7.982776641845703] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.