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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
36510044-1e43-4445-82f5-9ae82065e708 | exploiting-temporal-context-for-3d-human-pose | 1905.04266 | null | https://arxiv.org/abs/1905.04266v1 | https://arxiv.org/pdf/1905.04266v1.pdf | Exploiting temporal context for 3D human pose estimation in the wild | We present a bundle-adjustment-based algorithm for recovering accurate 3D human pose and meshes from monocular videos. Unlike previous algorithms which operate on single frames, we show that reconstructing a person over an entire sequence gives extra constraints that can resolve ambiguities. This is because videos often give multiple views of a person, yet the overall body shape does not change and 3D positions vary slowly. Our method improves not only on standard mocap-based datasets like Human 3.6M -- where we show quantitative improvements -- but also on challenging in-the-wild datasets such as Kinetics. Building upon our algorithm, we present a new dataset of more than 3 million frames of YouTube videos from Kinetics with automatically generated 3D poses and meshes. We show that retraining a single-frame 3D pose estimator on this data improves accuracy on both real-world and mocap data by evaluating on the 3DPW and HumanEVA datasets. | ['Carl Doersch', 'Anurag Arnab', 'Andrew Zisserman'] | 2019-05-10 | exploiting-temporal-context-for-3d-human-pose-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Arnab_Exploiting_Temporal_Context_for_3D_Human_Pose_Estimation_in_the_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Arnab_Exploiting_Temporal_Context_for_3D_Human_Pose_Estimation_in_the_CVPR_2019_paper.pdf | cvpr-2019-6 | ['monocular-3d-human-pose-estimation'] | ['computer-vision'] | [-2.00406373e-01 -1.90316401e-02 2.20716849e-01 -3.98944676e-01
-7.21031070e-01 -7.42524862e-01 3.74207020e-01 -4.04840052e-01
-6.67984426e-01 6.40949965e-01 6.12074852e-01 6.41915143e-01
4.18854713e-01 -1.69468388e-01 -9.60814297e-01 -1.11274123e-01
-2.52473623e-01 8.80348742e-01 3.47853124e-01 -2.53257930e-01
-1.59030303e-01 3.04453552e-01 -1.54761016e+00 1.13790974e-01
6.69347644e-02 6.22099161e-01 -2.89829761e-01 1.07812512e+00
5.61827123e-01 3.18161845e-01 -3.98917228e-01 -6.03158891e-01
7.94004738e-01 -3.54924560e-01 -6.43675387e-01 6.66713238e-01
1.37777042e+00 -7.39090502e-01 -8.25649440e-01 4.35736686e-01
7.56378233e-01 1.84334427e-01 3.21024060e-01 -1.18869126e+00
1.74577028e-01 -1.66024029e-01 -4.81318623e-01 1.21780254e-01
1.29000795e+00 2.95013249e-01 7.67510056e-01 -8.47807705e-01
1.31497097e+00 1.48954642e+00 1.29244745e+00 7.01851785e-01
-1.21520662e+00 -2.83954352e-01 -2.08424833e-02 1.54760733e-01
-1.44116461e+00 -6.18939400e-01 4.82416779e-01 -6.99435472e-01
8.60276997e-01 1.38943210e-01 1.27334964e+00 1.49439740e+00
6.49129301e-02 8.40471804e-01 7.90044427e-01 -6.75813556e-02
-1.98946640e-01 -5.89891851e-01 -3.10915291e-01 8.67447913e-01
2.05091760e-01 5.58982752e-02 -9.71500993e-01 -6.62804917e-02
1.29479301e+00 6.23709038e-02 -2.65332431e-01 -9.17092741e-01
-1.59260595e+00 4.29651678e-01 9.71857980e-02 -4.18937534e-01
-2.35889301e-01 3.88331234e-01 2.16913998e-01 2.24829525e-01
1.91074431e-01 5.19500561e-02 -5.16562164e-01 -6.17314398e-01
-8.20103943e-01 9.76523280e-01 7.65017986e-01 1.08141220e+00
4.80341375e-01 -2.20357731e-01 1.94310755e-01 4.26734060e-01
1.23806067e-01 7.81538665e-01 2.00155899e-01 -1.70624375e+00
5.78411818e-01 3.98745477e-01 5.21502614e-01 -1.12309086e+00
-7.57847607e-01 -3.78030166e-02 -2.59764433e-01 1.52552918e-01
8.89547944e-01 -2.51121402e-01 -7.71208704e-01 1.58307934e+00
7.76687860e-01 1.33316755e-01 -4.99189794e-01 1.42808020e+00
8.38812292e-01 7.45116696e-02 -3.76282066e-01 1.43046260e-01
1.25189686e+00 -8.44409704e-01 -5.78370571e-01 -3.61503631e-01
3.04891855e-01 -5.56547105e-01 7.34744191e-01 5.60499489e-01
-1.25880444e+00 -6.51513278e-01 -8.14516664e-01 -3.39778513e-01
2.29995608e-01 -2.39536062e-01 4.39365834e-01 4.70569938e-01
-8.44131231e-01 6.38907611e-01 -1.07695842e+00 -8.33484411e-01
1.24177881e-01 2.90968299e-01 -1.13931298e+00 -1.74547695e-02
-7.52348423e-01 8.07178259e-01 -1.20598055e-01 3.28182615e-02
-8.64231884e-01 -5.43386102e-01 -1.11024833e+00 -6.05016053e-01
4.97338027e-01 -1.03720617e+00 1.45583129e+00 -6.25845373e-01
-1.40624917e+00 1.18355942e+00 -2.22986683e-01 -3.91609073e-01
1.25072098e+00 -8.88513923e-01 -3.62180620e-02 5.94880939e-01
1.86254047e-02 9.22626376e-01 7.31684685e-01 -1.05804098e+00
-6.15117073e-01 -5.94451845e-01 1.73420638e-01 5.53841829e-01
2.43856698e-01 -2.32588097e-01 -9.79139030e-01 -5.15813589e-01
2.86463767e-01 -1.20610225e+00 -4.19540294e-02 6.01810932e-01
-1.83596492e-01 2.03923419e-01 4.73703563e-01 -9.28221047e-01
7.11216927e-01 -1.81551313e+00 7.94682145e-01 -6.86377287e-02
3.34210098e-01 -1.55638009e-01 1.27114996e-01 1.55725524e-01
1.85865298e-01 -3.59105945e-01 5.16429953e-02 -7.65221477e-01
3.10514029e-02 5.01986504e-01 3.56183171e-01 9.81688678e-01
-2.61366814e-01 9.50519800e-01 -8.20217848e-01 -6.20272636e-01
3.94119352e-01 6.56805873e-01 -8.47015083e-01 2.46408656e-01
1.06701124e-02 7.46812284e-01 -9.40044224e-03 7.57956088e-01
3.86328906e-01 -8.16557780e-02 1.04727946e-01 -4.67916168e-02
1.72537923e-01 -1.79250494e-01 -1.49660730e+00 2.56985903e+00
1.65632620e-01 5.98392248e-01 2.62378246e-01 -4.61092740e-01
4.04436946e-01 4.18938041e-01 8.41358304e-01 -2.75923491e-01
1.07670791e-01 -9.48213637e-02 -3.37232709e-01 -6.11744404e-01
3.35465819e-01 -1.56760827e-01 -2.23053873e-01 9.72455740e-02
2.46488124e-01 -1.22574344e-01 2.27357388e-01 1.80927038e-01
1.25121891e+00 7.90330648e-01 1.47892892e-01 -5.37792183e-02
1.28497392e-01 6.94392100e-02 7.49312162e-01 4.50963616e-01
-5.19473195e-01 1.32431710e+00 2.86719024e-01 -8.70994449e-01
-1.46293139e+00 -1.34958458e+00 1.02736257e-01 7.31274784e-01
1.25880733e-01 -7.83516526e-01 -8.37054312e-01 -5.35055995e-01
4.20074701e-01 -8.78976285e-02 -7.42349982e-01 5.18136203e-01
-7.35827982e-01 -1.17491744e-01 5.03794789e-01 5.81449032e-01
4.01426166e-01 -5.61391771e-01 -1.01374996e+00 1.27435893e-01
-5.68924844e-01 -1.56193268e+00 -7.75803506e-01 -4.95530874e-01
-8.18844318e-01 -1.54161143e+00 -8.15807462e-01 -4.13230181e-01
5.83698750e-01 2.18965277e-01 1.22763252e+00 -4.41973098e-02
-5.08826911e-01 8.86588871e-01 -2.78233349e-01 4.39681262e-02
1.48893893e-01 -1.89545467e-01 5.57209611e-01 -1.14853218e-01
3.01098734e-01 -6.69989705e-01 -5.57316720e-01 5.20035565e-01
-1.54443279e-01 1.41368091e-01 -2.07597464e-02 4.66723025e-01
6.24413252e-01 -6.21841908e-01 -1.60716087e-01 -6.29484713e-01
-1.57063767e-01 -8.78737867e-03 -3.81784499e-01 -2.34117359e-01
1.62048444e-01 -3.01219046e-01 9.68848988e-02 -2.92336643e-01
-6.90411568e-01 6.56134546e-01 -3.34283113e-02 -7.65642881e-01
-3.16130489e-01 -1.66488409e-01 -8.76407139e-03 2.64814403e-02
9.03600037e-01 -8.19204524e-02 2.60196418e-01 -5.64003825e-01
2.26386368e-01 2.43175760e-01 1.23911405e+00 -6.05085135e-01
9.55753565e-01 9.63377893e-01 7.81055540e-02 -8.18558455e-01
-8.20634067e-01 -6.72302246e-01 -1.41197824e+00 -5.17308235e-01
1.05091429e+00 -1.45638764e+00 -9.80616093e-01 4.71961766e-01
-1.06390393e+00 -3.42444241e-01 -1.48684502e-01 5.91615736e-01
-9.84908164e-01 5.44338703e-01 -7.25641608e-01 -6.03163123e-01
9.73319188e-02 -8.84239078e-01 1.48148811e+00 -1.66772977e-01
-7.77632475e-01 -7.89890826e-01 3.05936664e-01 6.17433310e-01
-3.52738231e-01 9.23378170e-01 -2.28976741e-01 -1.93405449e-02
-3.22715968e-01 -3.39385599e-01 3.63238513e-01 -6.35854602e-02
-1.86544925e-01 -2.53685713e-01 -6.15765691e-01 -4.28417265e-01
-3.64828825e-01 -3.71955127e-01 4.56226319e-01 3.12899470e-01
5.16072333e-01 -4.03658673e-02 -1.56391129e-01 7.53412247e-01
1.05825281e+00 -5.33849716e-01 5.55963159e-01 3.72960001e-01
9.95209992e-01 6.23728752e-01 5.13405263e-01 5.55062711e-01
7.65702307e-01 1.03131533e+00 2.73150355e-01 2.52531141e-01
-2.73808837e-01 -5.22749245e-01 4.50794995e-01 5.97574055e-01
-6.94354296e-01 2.37383544e-01 -8.38739991e-01 4.48850870e-01
-2.05445957e+00 -1.12179673e+00 -2.93737024e-01 2.29170537e+00
6.82568729e-01 1.33740187e-01 8.01225424e-01 4.17752154e-02
5.31960130e-01 2.52287369e-03 -4.64429975e-01 3.11257333e-01
-1.54146060e-01 -2.90147513e-02 4.98521328e-01 7.86674678e-01
-1.19178152e+00 7.13139355e-01 7.12185526e+00 -7.36705288e-02
-5.24906218e-01 4.83208969e-02 -6.99718669e-02 -9.45955992e-01
8.61893818e-02 -2.47782737e-01 -7.55783975e-01 1.85135841e-01
5.19778371e-01 1.34770662e-01 3.82697195e-01 7.16727555e-01
2.46434480e-01 -1.81053758e-01 -1.45317876e+00 1.57059646e+00
4.55795854e-01 -9.96829033e-01 -3.51246297e-01 2.96196252e-01
7.32495248e-01 1.96433235e-02 -4.27558273e-01 3.76358889e-02
1.07815206e-01 -8.97536755e-01 1.09646285e+00 7.39456892e-01
6.93170547e-01 -6.84598327e-01 4.78379637e-01 4.50330585e-01
-1.13103819e+00 2.76263982e-01 -2.55166590e-01 -4.23444092e-01
4.29514468e-01 2.69769013e-01 -4.34431255e-01 3.66895497e-01
1.04893744e+00 1.05132580e+00 -6.77167594e-01 9.84551013e-01
-2.62665033e-01 5.42377234e-02 -6.52339220e-01 4.45827335e-01
-2.79637814e-01 2.12225206e-02 7.20251381e-01 1.05792522e+00
2.01511949e-01 3.36341292e-01 4.57373589e-01 1.82519093e-01
9.61781070e-02 -2.84323215e-01 -5.53369343e-01 4.11962271e-01
1.48447096e-01 1.16580021e+00 -4.10553604e-01 -3.32897931e-01
-3.63071769e-01 1.30895698e+00 2.93680191e-01 8.05072263e-02
-9.02577937e-01 7.94256628e-02 9.62749183e-01 6.26919210e-01
5.06422460e-01 -8.75311792e-01 4.71056402e-02 -1.59479094e+00
2.68390685e-01 -9.60403681e-01 4.55951363e-01 -9.07613099e-01
-1.04718888e+00 1.62185445e-01 1.34022266e-01 -1.32846296e+00
-5.41190326e-01 -5.13892353e-01 4.50955518e-02 3.59079778e-01
-7.94656932e-01 -1.01211357e+00 -5.54221570e-01 8.20821047e-01
6.94797039e-01 1.02712780e-01 5.53501189e-01 2.73320764e-01
-2.41331354e-01 3.50832254e-01 -3.55597466e-01 2.75487661e-01
9.80838776e-01 -1.25367033e+00 7.90071487e-01 6.90714598e-01
3.74721944e-01 4.24069852e-01 1.01398182e+00 -7.76865423e-01
-1.78990161e+00 -5.86706042e-01 7.50239253e-01 -1.47310567e+00
2.50418752e-01 -7.13204503e-01 -3.66676241e-01 1.18657374e+00
-1.24248162e-01 3.29251260e-01 4.18455362e-01 1.74191490e-01
-3.81884873e-01 1.72578245e-01 -1.17635071e+00 5.20467103e-01
1.96438658e+00 -3.50626588e-01 -8.74776185e-01 3.14667314e-01
3.17273080e-01 -1.16461658e+00 -9.71290886e-01 1.27462476e-01
1.16470373e+00 -1.22585666e+00 1.26976728e+00 -6.81846201e-01
4.00564075e-01 -3.86836231e-01 -3.79998863e-01 -1.19058168e+00
-2.96280295e-01 -8.65105450e-01 -4.44245756e-01 5.30821979e-01
-2.34328926e-01 -3.29676806e-03 1.17926800e+00 7.54396498e-01
2.77783513e-01 -2.87168801e-01 -1.07846296e+00 -7.10990012e-01
-3.62724662e-01 -4.73466784e-01 1.88090593e-01 5.66821873e-01
3.85228097e-02 1.75371990e-01 -8.01670372e-01 7.09450096e-02
1.09676445e+00 -2.89533138e-01 1.66949904e+00 -1.33265650e+00
-4.24308181e-01 2.21266717e-01 -1.01652730e+00 -1.28136671e+00
2.35017054e-02 -3.64852995e-01 -3.13864462e-02 -1.25405288e+00
2.28330821e-01 3.27182144e-01 4.75191593e-01 4.00410473e-01
-9.65880528e-02 6.37550175e-01 3.46675187e-01 2.02604070e-01
-9.21350181e-01 2.99024850e-01 1.17264283e+00 3.37653190e-01
2.16344729e-01 -2.73042679e-01 -1.33418635e-01 1.06256044e+00
1.92205623e-01 -3.71793151e-01 -1.28550932e-01 -6.04445994e-01
1.90028310e-01 3.30608219e-01 8.84146273e-01 -1.38984108e+00
4.87086736e-02 5.44536784e-02 9.72335517e-01 -6.51635587e-01
7.84088314e-01 -6.91475809e-01 6.12557352e-01 4.48162705e-01
4.17784378e-02 3.60453486e-01 1.85275357e-02 7.85589635e-01
3.07475924e-01 3.04909259e-01 3.99308711e-01 -6.70404553e-01
-8.98735404e-01 4.27199095e-01 -6.62172884e-02 4.67229098e-01
8.78907502e-01 -5.14913917e-01 1.14984661e-01 -7.84542561e-01
-1.17594147e+00 3.76070529e-01 1.02157915e+00 5.72500229e-01
5.16171873e-01 -1.55636322e+00 -9.49053526e-01 1.54728532e-01
3.83682474e-02 2.38052800e-01 1.37204304e-01 6.76535785e-01
-1.03155625e+00 1.72001421e-01 -5.04467547e-01 -1.10209131e+00
-1.58209527e+00 1.82431355e-01 3.70296896e-01 2.92628139e-01
-9.93712962e-01 8.20755541e-01 -3.67395058e-02 -6.57589316e-01
2.55497932e-01 -5.81583343e-02 2.96625078e-01 -1.41299352e-01
5.54355204e-01 6.04277790e-01 -1.52892634e-01 -1.20150530e+00
-4.79908675e-01 9.21547592e-01 4.07057106e-01 -5.52448988e-01
1.31596053e+00 -4.78724748e-01 3.49441260e-01 5.08094490e-01
1.11411119e+00 2.12217808e-01 -2.00151014e+00 -1.67909995e-01
-4.41417068e-01 -1.01256883e+00 -5.17875969e-01 -3.98234427e-01
-7.81805336e-01 6.40092909e-01 4.01231676e-01 -3.68219048e-01
3.99128884e-01 -8.76866805e-04 9.25748527e-01 4.71693277e-01
8.49954605e-01 -1.32009232e+00 1.81246027e-01 3.71696591e-01
9.82573807e-01 -1.18219471e+00 4.28607136e-01 -3.87664348e-01
-5.46519458e-01 9.71819282e-01 6.26235306e-01 -3.46441716e-01
2.49049425e-01 1.31119668e-01 1.02901310e-01 -1.89134806e-01
-4.60629553e-01 -5.72395138e-02 3.25227976e-01 8.09825838e-01
2.59535491e-01 2.51543168e-02 -6.34521544e-02 2.33842477e-01
-6.90193117e-01 1.02167219e-01 6.41104281e-01 1.20398962e+00
-3.00868064e-01 -9.65208590e-01 -7.83109665e-01 4.31765094e-02
-2.77710885e-01 6.69048905e-01 -4.75045383e-01 1.05934179e+00
1.88086510e-01 8.16666543e-01 1.96770921e-01 -4.90924090e-01
7.78367579e-01 1.77145809e-01 1.24332035e+00 -5.32366335e-01
-3.71445179e-01 2.59911835e-01 3.76602173e-01 -1.17365813e+00
-6.63110137e-01 -1.21333790e+00 -1.16312230e+00 -6.96044445e-01
3.00154895e-01 -2.42723808e-01 4.26451445e-01 8.58295560e-01
3.26347440e-01 9.47388560e-02 8.51532146e-02 -1.54214203e+00
-3.01865101e-01 -8.02256048e-01 -6.03361309e-01 9.99409974e-01
4.72288370e-01 -9.28779423e-01 -1.21759653e-01 5.32300711e-01] | [7.090700149536133, -0.8899604082107544] |
998141e6-66c2-48d4-b547-eeef614bbe41 | we-never-go-out-of-style-motion | 2306.00559 | null | https://arxiv.org/abs/2306.00559v1 | https://arxiv.org/pdf/2306.00559v1.pdf | We never go out of Style: Motion Disentanglement by Subspace Decomposition of Latent Space | Real-world objects perform complex motions that involve multiple independent motion components. For example, while talking, a person continuously changes their expressions, head, and body pose. In this work, we propose a novel method to decompose motion in videos by using a pretrained image GAN model. We discover disentangled motion subspaces in the latent space of widely used style-based GAN models that are semantically meaningful and control a single explainable motion component. The proposed method uses only a few $(\approx10)$ ground truth video sequences to obtain such subspaces. We extensively evaluate the disentanglement properties of motion subspaces on face and car datasets, quantitatively and qualitatively. Further, we present results for multiple downstream tasks such as motion editing, and selective motion transfer, e.g. transferring only facial expressions without training for it. | ['R. Venkatesh Babu', 'Piyush Tiwari', 'Raghav Magazine', 'Rishubh Parihar'] | 2023-06-01 | null | null | null | null | ['motion-disentanglement', 'disentanglement'] | ['computer-vision', 'methodology'] | [ 6.93865195e-02 1.54918060e-01 -1.93605751e-01 -3.45010847e-01
-5.83151519e-01 -8.36436749e-01 6.33886039e-01 -1.11908233e+00
-1.14281476e-01 8.08600485e-01 4.88894522e-01 2.31911182e-01
3.61466318e-01 -3.77954870e-01 -9.98776197e-01 -9.47360218e-01
4.38837618e-01 4.09120262e-01 -5.15613675e-01 -2.81136613e-02
-2.81555951e-01 4.46259677e-01 -1.16463864e+00 2.01352239e-01
4.23023432e-01 4.66151983e-01 -1.35606542e-01 8.62515390e-01
2.69355416e-01 9.70877826e-01 -4.34484929e-01 -4.05168086e-01
2.38093853e-01 -8.77726495e-01 -7.47833133e-01 5.85464120e-01
6.07228160e-01 -5.86994171e-01 -8.10241342e-01 8.88077497e-01
1.00633487e-01 3.35247338e-01 7.63393462e-01 -1.38929570e+00
-6.55490577e-01 2.99704373e-01 -7.22458184e-01 -2.02040628e-01
5.13980389e-01 5.21344841e-01 6.97572529e-01 -7.13650942e-01
1.06801867e+00 1.36024642e+00 1.81627274e-01 1.36025023e+00
-1.52625740e+00 -7.55569637e-01 2.15122655e-01 8.20495933e-02
-1.17724705e+00 -7.78147161e-01 1.14323688e+00 -6.08357131e-01
5.25557220e-01 4.68425542e-01 7.17132092e-01 1.63903332e+00
6.63703382e-02 8.73819470e-01 7.18484461e-01 -9.76305529e-02
9.16998163e-02 -1.38670161e-01 -4.11515743e-01 7.77778506e-01
6.25842214e-02 -7.88742229e-02 -6.96521759e-01 6.09307997e-02
1.07104945e+00 -3.12862620e-02 -4.36152399e-01 -6.28382802e-01
-1.45588577e+00 7.63263524e-01 4.44302522e-03 2.60453504e-02
-2.13292882e-01 6.77746236e-01 1.54124469e-01 1.09805420e-01
3.58531505e-01 4.05255407e-02 -1.65561706e-01 -1.74056754e-01
-7.54892886e-01 2.74156749e-01 5.37850082e-01 1.21520793e+00
7.57427275e-01 4.89898533e-01 -2.35731393e-01 3.20046306e-01
1.82771698e-01 6.71783924e-01 3.87232929e-01 -1.59273291e+00
4.28269804e-01 6.18007183e-02 3.43936622e-01 -1.02400029e+00
-2.49986537e-02 1.77427277e-01 -1.12882483e+00 1.70505401e-02
4.13309872e-01 -3.29910517e-01 -8.04368079e-01 2.37436509e+00
3.29652071e-01 6.01563156e-01 2.79120319e-02 9.05921996e-01
5.58111787e-01 6.14114523e-01 7.44265318e-02 -4.08971816e-01
1.05504394e+00 -1.06620693e+00 -1.01995635e+00 -1.39187142e-01
4.98488337e-01 -5.42404413e-01 1.14393294e+00 3.29040796e-01
-1.42422390e+00 -5.72002411e-01 -7.49369562e-01 -1.79938465e-01
4.07348007e-01 2.36251146e-01 6.84722245e-01 3.62929970e-01
-1.01189768e+00 4.96717066e-01 -1.19675589e+00 -2.24194720e-01
3.74750763e-01 2.69426256e-01 -7.19276965e-01 -1.53319895e-01
-8.32050920e-01 5.30777156e-01 -6.04643784e-02 1.43623650e-01
-1.32321417e+00 -4.95998621e-01 -1.03706741e+00 -1.73834562e-01
1.03853650e-01 -1.32635641e+00 1.10772336e+00 -1.23344195e+00
-1.93443966e+00 9.39847648e-01 -5.44688225e-01 -6.89627230e-02
7.29896069e-01 -4.60839927e-01 -1.59300819e-01 3.17721069e-01
-1.23563148e-01 1.06851125e+00 1.26997757e+00 -1.19027340e+00
-2.54964866e-02 -3.81820023e-01 -5.93084469e-02 2.94489473e-01
-1.94706261e-01 -1.02729023e-01 -6.30520403e-01 -1.01424277e+00
-2.02569470e-01 -1.39373112e+00 -1.33916482e-01 1.93236753e-01
-4.54008132e-01 2.98189938e-01 9.95713949e-01 -8.79016638e-01
9.01134014e-01 -2.16676879e+00 1.14781380e+00 -2.25590557e-01
3.07012856e-01 -1.78293794e-01 -2.09601700e-01 1.86832976e-02
-1.24331467e-01 2.54304390e-02 -2.68929064e-01 -6.44606590e-01
-1.43170953e-01 4.11253333e-01 -2.87546068e-01 6.04560077e-01
1.11688696e-01 1.15350592e+00 -7.24654853e-01 -1.75291836e-01
1.40447676e-01 7.14822114e-01 -8.80155981e-01 4.19966906e-01
-2.05734670e-01 1.45766556e+00 -2.96322882e-01 2.78460562e-01
4.63492036e-01 -9.69443619e-02 2.23965943e-01 -3.22842568e-01
3.94106299e-01 -2.48501301e-01 -8.03432047e-01 2.33391500e+00
-4.99970347e-01 8.64319324e-01 1.71447501e-01 -8.85415792e-01
4.11182165e-01 5.59574485e-01 7.15424895e-01 -2.68452093e-02
2.36914188e-01 -1.70431599e-01 -1.43554688e-01 -5.74031711e-01
1.72668591e-01 -4.04437900e-01 -3.97990011e-02 3.30886304e-01
2.77601153e-01 -1.13402158e-01 -1.73480138e-01 1.35208830e-01
9.71081197e-01 4.10031259e-01 2.01527804e-01 -8.94454345e-02
4.31253254e-01 -3.87030423e-01 5.58440030e-01 6.71862438e-02
-2.07925573e-01 8.19568753e-01 3.90012830e-01 -3.86708379e-01
-1.11364722e+00 -1.18894720e+00 4.60734397e-01 7.23567784e-01
1.18430957e-01 -1.12120509e-01 -1.02924407e+00 -6.27944231e-01
-2.71656036e-01 8.60448956e-01 -7.25642204e-01 -4.55416471e-01
-9.24992442e-01 -5.21229804e-01 5.36868274e-01 7.60010660e-01
4.32159513e-01 -9.60691810e-01 -4.64724869e-01 -4.35182899e-02
-6.94801211e-01 -1.30956447e+00 -9.52582359e-01 -6.42588615e-01
-8.78241420e-01 -6.70647025e-01 -8.84465039e-01 -5.21088958e-01
8.50783825e-01 3.55146199e-01 1.04741669e+00 -2.43190452e-01
-3.15571398e-01 7.59193897e-01 1.62665937e-02 1.24380626e-01
-2.79138088e-01 -3.61052871e-01 2.20697030e-01 3.97385865e-01
3.12605291e-04 -7.36918688e-01 -7.17248440e-01 3.50401103e-01
-8.69300961e-01 3.60737681e-01 2.91012824e-01 8.45455408e-01
4.84386325e-01 -4.27971572e-01 1.64505959e-01 -6.41380608e-01
1.16921633e-01 -3.58800918e-01 -2.88669914e-01 3.92037854e-02
1.26412690e-01 1.51195303e-01 3.90582383e-01 -6.20685279e-01
-1.35907102e+00 3.75151277e-01 5.42591922e-02 -1.05171728e+00
-3.84614140e-01 -1.57005057e-01 -7.61964321e-01 8.42613876e-02
4.09866780e-01 2.68053710e-01 3.38469446e-02 -2.02323139e-01
9.32134211e-01 -1.39305638e-02 7.43896186e-01 -6.76257610e-01
1.03601277e+00 1.08851290e+00 1.59397703e-02 -1.00299692e+00
-4.66993332e-01 -9.65133682e-02 -7.73032486e-01 -3.28675836e-01
1.40144718e+00 -1.15174913e+00 -8.05105090e-01 4.27078873e-01
-1.32212222e+00 -4.14906293e-01 -3.64735454e-01 5.79732418e-01
-1.14015388e+00 3.35235059e-01 -7.17760265e-01 -3.80891114e-01
-6.03486933e-02 -1.13712275e+00 1.24938619e+00 -4.44311798e-02
-6.40141487e-01 -8.76202226e-01 2.52638310e-01 5.56325853e-01
-1.17525533e-02 6.15985811e-01 6.80147886e-01 9.33969244e-02
-8.53944361e-01 1.06556237e-01 3.50499630e-01 2.53537357e-01
3.19676459e-01 4.32766899e-02 -8.63009691e-01 -5.18749654e-01
1.13158531e-01 -2.29814455e-01 7.32948363e-01 4.90533978e-01
1.25207627e+00 -7.28267074e-01 -2.52210349e-01 1.13643646e+00
9.66268957e-01 3.17485183e-01 7.60723114e-01 -3.29854071e-01
1.23928523e+00 8.08327854e-01 2.20279694e-01 1.98421001e-01
2.13819221e-01 8.65626454e-01 1.91376194e-01 7.88846761e-02
-1.37272611e-01 -3.84172648e-01 8.28665257e-01 1.04785573e+00
-6.01845860e-01 -3.07926476e-01 -4.19710368e-01 3.17399591e-01
-1.77505553e+00 -1.19235051e+00 2.27385521e-01 1.88153458e+00
5.91739595e-01 -3.39687139e-01 4.99903336e-02 -2.68209636e-01
5.07985830e-01 2.50702679e-01 -7.64501154e-01 1.41178006e-02
-1.57736748e-01 9.72090736e-02 1.47524297e-01 6.74814641e-01
-8.97847772e-01 1.13032103e+00 6.17191458e+00 3.78948838e-01
-1.07404315e+00 2.21542478e-01 6.32665515e-01 -5.67894638e-01
-7.04367518e-01 -6.62718639e-02 -3.94001305e-01 1.65048212e-01
6.28865421e-01 -1.28734529e-01 5.79556763e-01 6.42183006e-01
3.69597703e-01 4.06897128e-01 -1.45267761e+00 1.30490685e+00
2.35997051e-01 -1.42498720e+00 4.70636576e-01 2.11868528e-02
9.33110595e-01 -6.28362000e-01 3.37913752e-01 2.38052364e-02
2.61131734e-01 -1.18450952e+00 8.67260516e-01 6.49849415e-01
1.00180531e+00 -6.36829853e-01 7.92065114e-02 2.67419428e-01
-9.84560728e-01 2.80342251e-01 1.50183711e-04 1.20390907e-01
4.82994020e-01 2.15588994e-02 -5.04660420e-02 4.83783394e-01
4.33023661e-01 8.65193963e-01 -4.74657416e-02 8.54835510e-02
-4.48327988e-01 4.37965333e-01 1.20594256e-01 5.70068300e-01
-8.68322253e-02 -5.41128397e-01 8.64651978e-01 8.26713443e-01
4.67413247e-01 5.08352399e-01 -4.01151955e-01 1.05676103e+00
-2.89680570e-01 -3.11893791e-01 -9.10414457e-01 7.43520353e-03
5.23596331e-02 1.08950222e+00 -4.31807280e-01 -3.11563879e-01
-4.53334868e-01 1.72616756e+00 1.43391937e-01 7.97124267e-01
-1.14163280e+00 3.15502912e-01 1.26556873e+00 -9.03661400e-02
1.10696390e-01 -6.84068441e-01 -1.80615392e-02 -1.95838463e+00
-2.08603553e-02 -8.37347269e-01 9.74245428e-04 -9.08106744e-01
-9.45714891e-01 5.07073522e-01 1.76643893e-01 -1.23134983e+00
-7.86508322e-01 -3.65887344e-01 -6.10283732e-01 6.43383443e-01
-6.85859323e-01 -1.38693166e+00 -5.41525960e-01 1.00078356e+00
8.30653548e-01 -1.38272911e-01 6.69104040e-01 2.01076597e-01
-6.99814737e-01 6.08334482e-01 -1.31331488e-01 1.70833483e-01
6.23364925e-01 -8.44108820e-01 4.50418293e-01 8.43102634e-01
4.21788752e-01 6.63411975e-01 8.58056724e-01 -4.81427133e-01
-1.68093181e+00 -1.22646892e+00 2.76266634e-01 -6.87545478e-01
3.13832343e-01 -5.58043778e-01 -7.02305734e-01 1.25971794e+00
4.10366893e-01 -4.82613593e-03 7.49709964e-01 -4.79902655e-01
-2.52442449e-01 6.96842596e-02 -1.05054331e+00 8.57655048e-01
1.53873062e+00 -5.66495180e-01 -3.30054820e-01 8.35271669e-04
8.18714619e-01 -4.08201575e-01 -5.09867072e-01 3.51663679e-01
7.77546942e-01 -7.97318995e-01 1.10705626e+00 -1.14561641e+00
5.68485856e-01 -2.05129564e-01 -2.90538073e-01 -1.36651516e+00
-2.88783610e-01 -1.08566904e+00 -5.22435546e-01 1.00645936e+00
-7.36807510e-02 -2.30146825e-01 1.07372642e+00 7.83515692e-01
1.80949554e-01 -3.07345033e-01 -8.01134109e-01 -6.97064877e-01
1.90276206e-01 -1.78542718e-01 3.36456537e-01 1.18942511e+00
-3.72699082e-01 2.73743719e-01 -9.83221114e-01 -9.71361715e-03
7.35481620e-01 3.52728404e-02 1.23009360e+00 -5.24178028e-01
-6.18058264e-01 -1.38289437e-01 -4.95931536e-01 -1.29183853e+00
6.94027126e-01 -6.88165605e-01 -1.25821501e-01 -9.36190128e-01
4.01066810e-01 3.14632863e-01 2.42869988e-01 1.00874379e-01
-7.01391697e-02 2.27463111e-01 3.86888236e-01 3.39126647e-01
-2.64683694e-01 1.04853356e+00 1.58932734e+00 -1.66031674e-01
-6.34043217e-02 -1.71405405e-01 -5.21755397e-01 1.05673897e+00
5.78973711e-01 -2.57142216e-01 -7.10385561e-01 -7.55251527e-01
-9.65410993e-02 5.46731532e-01 6.53790176e-01 -7.19117880e-01
-1.63692743e-01 -5.00877678e-01 4.47851866e-01 -1.56653404e-01
6.35282874e-01 -7.47734427e-01 6.47535741e-01 4.38305289e-01
-2.88512290e-01 1.58677220e-01 2.41957933e-01 8.54906380e-01
-2.59788573e-01 4.98135000e-01 8.95310462e-01 -1.63576692e-01
-5.52502096e-01 5.78996241e-01 -4.16081041e-01 5.44976741e-02
1.26755917e+00 -5.43380938e-02 1.69171989e-02 -9.55761671e-01
-1.26461494e+00 -1.35227114e-01 5.96562684e-01 5.19185662e-01
6.05966926e-01 -1.69547820e+00 -5.96024573e-01 2.14303315e-01
-1.64700046e-01 6.50475323e-02 5.65566540e-01 7.50908554e-01
-5.68000019e-01 3.12680811e-01 -5.62906623e-01 -6.65100217e-01
-1.25986207e+00 6.00766182e-01 3.68369639e-01 1.24473616e-01
-5.69386721e-01 7.86074340e-01 9.16316211e-01 -1.39912471e-01
-1.87806591e-01 -3.19000393e-01 8.92419666e-02 -2.66349792e-01
3.89153957e-01 4.70032424e-01 -5.43119133e-01 -1.04398203e+00
-1.91509679e-01 8.87396157e-01 3.68433952e-01 -5.80493510e-01
1.03496003e+00 -3.13340336e-01 -3.24968882e-02 4.18047160e-01
1.42784071e+00 9.52281877e-02 -1.64347064e+00 1.90364718e-01
-5.04007697e-01 -6.35233164e-01 -5.19790173e-01 -8.53983462e-02
-1.42056441e+00 1.03326344e+00 2.69566059e-01 -5.59476316e-01
1.25124049e+00 1.30717337e-01 6.86382294e-01 1.83809116e-01
4.68222797e-01 -6.50596917e-01 4.64290321e-01 6.30546883e-02
1.16341448e+00 -1.01816249e+00 -3.81634861e-01 -5.48649430e-01
-8.62921536e-01 8.29922616e-01 7.75241494e-01 -2.00192019e-01
5.49529850e-01 3.55702788e-02 -1.23585887e-01 4.89391275e-02
-8.19866061e-01 2.54291654e-01 4.20688748e-01 5.30369461e-01
3.24423224e-01 2.42793918e-01 -8.81559551e-02 4.83360440e-01
-1.72221676e-01 -4.69206832e-02 4.03940350e-01 5.55027425e-01
2.03990370e-01 -1.00217438e+00 -2.56889939e-01 -2.05737695e-01
-2.25486845e-01 3.47568572e-01 -4.35060650e-01 7.73951888e-01
5.68376109e-03 6.33549452e-01 8.20410997e-02 -4.65191960e-01
9.51092392e-02 1.37578666e-01 9.49713349e-01 -3.82037431e-01
1.38898894e-01 1.83142170e-01 -8.06379989e-02 -1.01034749e+00
-7.10255802e-01 -8.96116674e-01 -1.14161742e+00 -4.27540749e-01
3.79486501e-01 -1.97656050e-01 1.61081076e-01 8.47535849e-01
3.81860226e-01 4.12183464e-01 5.02738059e-01 -1.09914398e+00
-2.33700097e-01 -8.06292951e-01 -3.96326393e-01 8.94203246e-01
4.11257982e-01 -6.68894947e-01 -3.33540559e-01 8.34215879e-01] | [10.980628967285156, -0.660194993019104] |
c3add9ba-0f50-4a8a-8acd-99cec7f8f3ce | automatic-text-summarization-methods-a | 2204.01849 | null | https://arxiv.org/abs/2204.01849v1 | https://arxiv.org/pdf/2204.01849v1.pdf | Automatic Text Summarization Methods: A Comprehensive Review | One of the most pressing issues that have arisen due to the rapid growth of the Internet is known as information overloading. Simplifying the relevant information in the form of a summary will assist many people because the material on any topic is plentiful on the Internet. Manually summarising massive amounts of text is quite challenging for humans. So, it has increased the need for more complex and powerful summarizers. Researchers have been trying to improve approaches for creating summaries since the 1950s, such that the machine-generated summary matches the human-created summary. This study provides a detailed state-of-the-art analysis of text summarization concepts such as summarization approaches, techniques used, standard datasets, evaluation metrics and future scopes for research. The most commonly accepted approaches are extractive and abstractive, studied in detail in this work. Evaluating the summary and increasing the development of reusable resources and infrastructure aids in comparing and replicating findings, adding competition to improve the outcomes. Different evaluation methods of generated summaries are also discussed in this study. Finally, at the end of this study, several challenges and research opportunities related to text summarization research are mentioned that may be useful for potential researchers working in this area. | ['Arun Kumar Yadav', 'Jalpa Desai', 'Divakar Yadav'] | 2022-03-03 | null | null | null | null | ['art-analysis'] | ['computer-vision'] | [ 2.44756326e-01 2.07355604e-01 -3.07775319e-01 -2.62004565e-02
-5.85156679e-01 -5.41247487e-01 5.99027157e-01 1.05233240e+00
-3.78217757e-01 1.09470809e+00 9.78035212e-01 4.01171297e-02
-2.17712983e-01 -5.39599836e-01 -6.92827255e-02 -1.24471344e-01
2.96309233e-01 3.05309325e-01 2.44773492e-01 -4.33941901e-01
1.26384747e+00 2.12799132e-01 -1.88529885e+00 3.86004895e-01
1.46241784e+00 4.03571993e-01 5.38208246e-01 7.48077273e-01
-7.86716640e-01 4.29102570e-01 -1.42906141e+00 -4.54232454e-01
-2.19401002e-01 -6.23862624e-01 -1.02684009e+00 -1.40080497e-01
5.99383056e-01 -4.74175811e-02 -2.78219692e-02 1.02923727e+00
7.47607529e-01 2.84192383e-01 6.71939969e-01 -1.14189661e+00
-4.86642629e-01 8.77384722e-01 -4.72511113e-01 5.57338715e-01
8.93856645e-01 -3.39827806e-01 7.53944099e-01 -3.43543917e-01
7.86029935e-01 1.21012866e+00 3.43718052e-01 2.51693189e-01
-6.07448637e-01 -4.29246366e-01 -9.85460952e-02 2.62936264e-01
-8.25126946e-01 -3.52715969e-01 6.97836995e-01 -2.63978362e-01
1.17925870e+00 8.62630963e-01 7.39558756e-01 9.15464461e-01
4.10452634e-01 6.58997357e-01 6.75883353e-01 -7.11674511e-01
7.11816102e-02 2.72995979e-01 7.25158215e-01 1.31796733e-01
1.12398565e+00 -8.77654791e-01 -5.83411932e-01 -1.39931604e-01
2.35366467e-02 -5.78391328e-02 -2.47483537e-01 3.39154601e-01
-1.10663545e+00 7.43902981e-01 -1.12778828e-01 7.66542077e-01
-6.39399827e-01 -4.61817592e-01 1.05890226e+00 1.84816971e-01
7.29329348e-01 1.01445794e+00 -4.87179197e-02 -4.71323490e-01
-1.43895757e+00 5.99217236e-01 1.06248450e+00 9.79196727e-01
4.66934144e-01 -2.04053193e-01 -6.31384671e-01 6.61534667e-01
-2.81925201e-02 2.47227222e-01 8.73942196e-01 -9.08772051e-01
8.81195009e-01 1.18197942e+00 5.33661097e-02 -1.27885866e+00
-3.49478781e-01 -5.35988986e-01 -8.48989129e-01 -3.25268298e-01
-1.72309935e-01 -2.47465760e-01 -4.35340315e-01 1.01165140e+00
4.72987033e-02 -4.73916262e-01 1.68604985e-01 4.76257145e-01
1.40484500e+00 1.02379382e+00 -1.23127867e-02 -6.60618663e-01
1.53532171e+00 -1.02844334e+00 -1.27212226e+00 -2.07935423e-01
4.43363249e-01 -1.21923506e+00 9.90426064e-01 3.08474034e-01
-1.30824852e+00 -4.80054617e-01 -1.15927768e+00 -2.44740650e-01
-5.94267905e-01 1.25640407e-01 3.22328568e-01 5.61811924e-01
-9.38019812e-01 7.75210857e-01 -4.83412206e-01 -9.40203547e-01
2.62662709e-01 -6.46720231e-02 -2.45132789e-01 9.32117701e-02
-1.09699154e+00 1.13360286e+00 7.68760443e-01 -3.83149326e-01
-2.23100018e-02 -7.11291611e-01 -5.52982390e-01 1.90356568e-01
3.90907049e-01 -8.54316056e-01 1.35842431e+00 -4.46998090e-01
-1.08758569e+00 5.56619287e-01 -3.89280260e-01 -4.95688856e-01
3.71492535e-01 -5.05966246e-01 -2.51989454e-01 2.34878495e-01
5.63048184e-01 2.01755211e-01 1.51798546e-01 -9.60461795e-01
-8.41099322e-01 -3.98089319e-01 -2.54529327e-01 5.33305764e-01
-5.55554807e-01 4.24100965e-01 -2.54373163e-01 -7.57214010e-01
-2.31056854e-01 -3.55618477e-01 -7.01206475e-02 -7.77001023e-01
-5.90520024e-01 -5.97865880e-01 9.95265424e-01 -1.19554245e+00
2.09989238e+00 -1.53942776e+00 -1.29576400e-02 -3.27033311e-01
2.82815427e-01 7.42974460e-01 2.12886095e-01 1.41572917e+00
3.48498553e-01 7.26504207e-01 -4.70679142e-02 -2.49149621e-01
-7.44473264e-02 -1.79347947e-01 -3.34345311e-01 -7.11076558e-02
-3.18979353e-01 6.46551609e-01 -9.67783332e-01 -7.78959453e-01
7.60542452e-02 1.70015201e-01 6.32148609e-02 2.04053402e-01
-1.03415422e-01 9.78653356e-02 -6.17061555e-01 3.72795582e-01
5.16494513e-01 3.89356911e-03 -2.48731032e-01 -1.83208257e-01
-7.02046037e-01 5.64510524e-01 -1.07294977e+00 1.64343166e+00
-1.94689855e-01 9.73632276e-01 -3.48495543e-01 -8.64739358e-01
1.01370668e+00 3.97449195e-01 2.36190915e-01 -4.98703271e-01
4.85589802e-02 1.90974653e-01 -2.77318150e-01 -8.71277869e-01
1.25076652e+00 3.39514732e-01 -1.71422094e-01 6.03356063e-01
-2.15553522e-01 -3.74957800e-01 1.10650718e+00 8.14889014e-01
9.27740335e-01 -3.38662773e-01 8.41900110e-01 -2.50293195e-01
5.38864136e-01 3.52237612e-01 1.51396990e-01 7.72784173e-01
6.60731196e-02 3.81192535e-01 5.37478030e-01 -6.50823489e-02
-1.03981888e+00 -3.76420051e-01 1.84100702e-01 7.14397192e-01
-4.73752581e-02 -1.07368183e+00 -9.84781265e-01 -3.43279630e-01
-1.88504532e-01 1.11125708e+00 -3.66287321e-01 -5.44158742e-02
-3.68847162e-01 -2.37897292e-01 4.86516923e-01 1.61897510e-01
8.53718162e-01 -1.20085108e+00 -9.90624905e-01 2.11693719e-01
-6.30315006e-01 -7.25316226e-01 -3.22794169e-01 -4.41386223e-01
-1.19028914e+00 -8.83181810e-01 -8.25665534e-01 -5.11734962e-01
4.51292098e-01 7.52144933e-01 1.08156765e+00 3.17440927e-01
-2.09534436e-01 3.02446544e-01 -8.18254709e-01 -9.90674436e-01
-6.01473570e-01 5.95032513e-01 -3.18599254e-01 -7.43775308e-01
2.44280726e-01 -2.34179795e-01 -4.80742753e-01 -1.82347327e-01
-1.20729184e+00 1.55771613e-01 6.80370331e-01 2.63513446e-01
1.45119965e-01 2.52497822e-01 9.73935187e-01 -9.77846801e-01
1.63128686e+00 -5.11112571e-01 2.94555463e-02 5.09430826e-01
-7.70608604e-01 1.69240758e-01 5.19654930e-01 -5.14425039e-02
-1.17041087e+00 -7.22152650e-01 8.35415572e-02 4.80856299e-01
-9.46343392e-02 8.95787835e-01 1.09241314e-01 5.15373647e-01
8.64256561e-01 9.30583701e-02 -3.05394605e-02 -5.45853913e-01
1.14990152e-01 1.10396826e+00 2.33804911e-01 -2.65804589e-01
3.25884461e-01 3.37088443e-02 -2.24407271e-01 -1.34686160e+00
-8.61273944e-01 -9.22569573e-01 -3.25348347e-01 -4.78374183e-01
4.86294150e-01 -3.70011479e-01 -1.62183151e-01 2.80733645e-01
-1.40150785e+00 3.03919077e-01 -3.98758262e-01 4.23697680e-01
-5.18339835e-02 8.67342353e-01 -1.17511421e-01 -7.74759650e-01
-1.18309844e+00 -8.14218521e-01 8.02252173e-01 7.31517136e-01
-9.49951112e-01 -8.80207896e-01 2.34399870e-01 6.00801706e-01
6.56391144e-01 3.05213064e-01 6.87522829e-01 -1.00089419e+00
4.51011583e-03 -4.93322313e-01 -3.37448828e-02 2.28055239e-01
3.83888662e-01 3.49597722e-01 -4.80854243e-01 -8.02911595e-02
9.40076187e-02 7.84350261e-02 7.72465289e-01 4.50042695e-01
8.61484349e-01 -8.04713130e-01 -4.19140965e-01 -3.32213968e-01
1.16236997e+00 1.17808007e-01 6.93959534e-01 5.98570704e-01
2.34312579e-01 8.71027529e-01 9.13530350e-01 5.70734859e-01
3.39442968e-01 2.65506417e-01 1.23278186e-01 4.64227587e-01
-3.42340201e-01 -2.89997399e-01 9.74318981e-02 1.27218044e+00
-2.01045945e-01 -4.94228989e-01 -8.00706863e-01 4.36955839e-01
-1.93585432e+00 -1.21515703e+00 -4.62715626e-01 2.14673042e+00
8.07131171e-01 3.40912640e-02 1.68335527e-01 4.58232164e-01
8.39287102e-01 2.14998364e-01 -1.64821267e-01 -7.37552166e-01
-2.65667699e-02 2.29607020e-02 1.67091087e-01 3.75413299e-01
-5.77286005e-01 6.46510184e-01 6.16397095e+00 7.07498848e-01
-1.05858541e+00 -1.75877228e-01 3.15864325e-01 1.06211089e-01
-2.81539291e-01 8.72220546e-02 -9.42126453e-01 6.75271571e-01
9.94227767e-01 -1.19403160e+00 -4.24162596e-02 7.13594556e-01
5.97619712e-01 -6.29363596e-01 -6.22702718e-01 8.58654320e-01
5.65913975e-01 -1.53584254e+00 4.28257465e-01 -2.05080435e-01
8.29923689e-01 -3.56189787e-01 -3.32043588e-01 2.39363849e-01
-2.35191882e-01 -6.42141104e-01 4.66379166e-01 5.20339727e-01
2.60991454e-01 -6.30258203e-01 1.16594136e+00 4.77781266e-01
-8.36614549e-01 2.25912690e-01 -4.99932647e-01 -4.64132756e-01
1.86871514e-01 7.63093174e-01 -8.97992551e-01 9.72899199e-01
5.51461339e-01 6.70713484e-01 -8.87941957e-01 1.46404159e+00
-2.67661572e-01 4.41692054e-01 4.14802358e-02 -7.18814552e-01
-7.34511390e-02 -2.55738199e-01 8.28983247e-01 1.59568763e+00
5.19962847e-01 -6.40507508e-03 -3.89828123e-02 3.50228965e-01
-6.44385442e-03 6.53042674e-01 -7.04132080e-01 -2.46951133e-01
7.44607985e-01 1.35125303e+00 -9.85633194e-01 -7.44459510e-01
-6.41441122e-02 7.10860133e-01 -2.24518779e-04 4.02366044e-03
-3.37401718e-01 -7.86515594e-01 -1.07400000e-01 1.89189747e-01
-3.30543011e-01 -2.86925696e-02 -6.17591143e-01 -9.45616722e-01
2.29655325e-01 -9.57936049e-01 4.57654178e-01 -6.96136594e-01
-9.28706050e-01 4.69776183e-01 5.76010704e-01 -1.05816114e+00
-1.75099149e-01 -1.69218574e-02 -9.89485562e-01 7.72768974e-01
-1.11761165e+00 -6.27339542e-01 -6.56642020e-01 -8.31532106e-02
1.09432137e+00 -8.16257894e-02 6.44250274e-01 1.71758562e-01
-6.34173572e-01 1.42339379e-01 2.32399534e-02 -4.71233934e-01
8.41051519e-01 -1.09929645e+00 4.00795579e-01 9.62723792e-01
-1.75116584e-01 7.35961556e-01 1.19898832e+00 -1.07594085e+00
-1.12495959e+00 -6.98560774e-01 1.42683125e+00 -1.61392435e-01
4.76769716e-01 2.54705131e-01 -9.11227167e-01 2.00058579e-01
9.08514321e-01 -1.16687107e+00 7.22495973e-01 -1.10077634e-01
2.41753757e-01 -1.85961444e-02 -1.09395647e+00 7.28356004e-01
6.79143906e-01 -8.25951435e-03 -1.19541252e+00 3.96739900e-01
6.42283261e-01 -2.68188328e-01 -5.55948496e-01 -7.24891527e-03
2.28277728e-01 -8.18014622e-01 4.57670778e-01 -4.16221321e-01
6.47978723e-01 -6.17465703e-03 3.53994489e-01 -1.55689096e+00
6.68287128e-02 -7.69701302e-01 -8.83271173e-02 1.65973353e+00
4.35562313e-01 -5.29039025e-01 5.45907915e-01 7.40635097e-01
-4.62814450e-01 -5.39828420e-01 -5.16992509e-01 -5.53872764e-01
-3.76464352e-02 1.55479625e-01 2.91554034e-01 6.27751052e-01
5.14221728e-01 7.72942007e-01 5.99926710e-02 -4.83258367e-01
3.18138212e-01 -1.31491393e-01 9.87069726e-01 -1.58055258e+00
5.07450819e-01 -8.23968530e-01 -2.95967638e-01 -7.31788039e-01
-2.68577784e-01 -7.10950971e-01 -4.07216728e-01 -2.65734982e+00
5.70405602e-01 3.47674936e-01 3.68800670e-01 5.44616431e-02
-4.69859987e-01 -4.04280752e-01 3.50831032e-01 3.10056031e-01
-8.30067158e-01 2.16882572e-01 1.36776078e+00 -5.97915761e-02
-3.34859997e-01 6.06527850e-02 -1.28001833e+00 4.43138421e-01
1.28286874e+00 -4.66780066e-01 -5.21252692e-01 -5.64111993e-02
5.26765883e-01 -2.37937961e-02 -2.70725608e-01 -1.07014370e+00
6.62835002e-01 -9.81393605e-02 -3.30125988e-02 -1.06395400e+00
-9.63371992e-02 -3.99102420e-01 2.01746193e-03 4.37723815e-01
-5.06992280e-01 5.25713861e-01 2.71585196e-01 1.40387982e-01
-4.09633040e-01 -9.31331396e-01 3.26556116e-01 -2.86968589e-01
-3.95064741e-01 -2.86621034e-01 -3.87563735e-01 3.23045731e-01
9.73634541e-01 -4.25763875e-01 -6.77855849e-01 -6.09740973e-01
-2.82591805e-02 2.24402443e-01 2.62179494e-01 4.56156820e-01
4.69709754e-01 -6.94806278e-01 -1.04823327e+00 -4.79802847e-01
-4.99307103e-02 -4.79131788e-02 3.52269888e-01 5.61235309e-01
-8.83677959e-01 7.87173092e-01 -5.18858016e-01 -1.06866136e-01
-1.60749042e+00 4.12312180e-01 -5.29438674e-01 -6.65278196e-01
-6.40974402e-01 2.33427167e-01 -4.22151685e-01 1.04585111e-01
3.73764127e-01 -4.38652903e-01 -8.75296593e-01 6.35192871e-01
1.03777492e+00 1.32907462e+00 3.70495349e-01 -3.47348392e-01
-3.93787539e-03 2.51877159e-01 -3.42751443e-01 -1.32669225e-01
1.24215817e+00 -3.80315542e-01 -5.49254894e-01 5.67538738e-01
8.89322937e-01 2.26337343e-01 -3.43400657e-01 1.98312402e-01
3.15662026e-01 -2.77284592e-01 -5.50845116e-02 -9.49121296e-01
-4.34717119e-01 6.27338171e-01 -4.21047397e-02 9.25704896e-01
9.28197742e-01 -1.13916926e-01 6.67715430e-01 5.19884408e-01
-6.30271062e-02 -1.39015889e+00 3.93247940e-02 5.46665490e-01
1.27613759e+00 -1.10146654e+00 6.60132229e-01 -3.67779732e-01
-7.68617213e-01 1.35652006e+00 2.41671875e-01 1.09807543e-01
1.47989824e-01 -2.38271914e-02 -1.22470468e-01 -3.69286865e-01
-5.46696246e-01 1.88743129e-01 4.07973379e-01 4.34188455e-01
9.12496567e-01 -1.38763323e-01 -1.42891037e+00 6.59130752e-01
-5.68817019e-01 -1.07224276e-02 8.97406638e-01 1.04143679e+00
-8.37636828e-01 -1.07279944e+00 -5.32723665e-01 9.22130704e-01
-8.13237786e-01 1.06453918e-01 -9.03704286e-01 8.57388616e-01
-5.25488973e-01 1.45812643e+00 -1.93314031e-01 1.21656284e-02
6.01661801e-01 2.62170043e-02 1.31001741e-01 -9.11948621e-01
-8.18532526e-01 -3.83505136e-01 3.52247834e-01 1.94516554e-02
-5.57906032e-01 -7.84793258e-01 -1.18877625e+00 -6.66009009e-01
-3.67874533e-01 8.37917268e-01 1.09068155e+00 6.68303847e-01
6.15969956e-01 6.36216581e-01 2.62354702e-01 -7.83943474e-01
-4.96493280e-01 -1.49722469e+00 -1.53674975e-01 2.60567218e-01
2.43207384e-02 -2.53512919e-01 -3.78316432e-01 5.94363473e-02] | [12.435808181762695, 9.534560203552246] |
93ed24b8-472a-4a24-95de-a9caf0e97962 | infrared-and-visible-image-fusion-using-a-1 | 1811.02291 | null | https://arxiv.org/abs/1811.02291v5 | https://arxiv.org/pdf/1811.02291v5.pdf | MDLatLRR: A novel decomposition method for infrared and visible image fusion | Image decomposition is crucial for many image processing tasks, as it allows to extract salient features from source images. A good image decomposition method could lead to a better performance, especially in image fusion tasks. We propose a multi-level image decomposition method based on latent low-rank representation(LatLRR), which is called MDLatLRR. This decomposition method is applicable to many image processing fields. In this paper, we focus on the image fusion task. We develop a novel image fusion framework based on MDLatLRR, which is used to decompose source images into detail parts(salient features) and base parts. A nuclear-norm based fusion strategy is used to fuse the detail parts, and the base parts are fused by an averaging strategy. Compared with other state-of-the-art fusion methods, the proposed algorithm exhibits better fusion performance in both subjective and objective evaluation. | ['Xiao-Jun Wu', 'Josef Kittler', 'Hui Li'] | 2018-11-06 | null | null | null | null | ['infrared-and-visible-image-fusion'] | ['computer-vision'] | [ 4.07493472e-01 -6.42322540e-01 7.96415657e-02 -5.97852357e-02
-1.02322817e+00 4.67532920e-03 3.78469706e-01 8.44258443e-02
-2.83928573e-01 5.94899595e-01 6.07819378e-01 1.69343933e-01
-9.84908715e-02 -6.34001315e-01 -1.48477167e-01 -1.14338517e+00
3.63261014e-01 -4.88828748e-01 1.03199989e-01 -3.28646988e-01
9.19379517e-02 2.90091306e-01 -1.62630868e+00 2.82319844e-01
1.10813522e+00 1.03523862e+00 4.30013984e-01 2.64912724e-01
4.91403453e-02 7.52097547e-01 -4.07217622e-01 -6.34237230e-02
2.17181757e-01 -5.26485682e-01 -3.89992237e-01 5.80099940e-01
2.60272585e-02 -6.99435323e-02 -3.51172015e-02 1.58718717e+00
3.30409795e-01 3.99733394e-01 6.82105541e-01 -9.52296019e-01
-6.59947932e-01 2.25290924e-01 -1.07167113e+00 2.84935415e-01
4.32190865e-01 -2.33543858e-01 7.54267693e-01 -9.63739812e-01
1.57981738e-01 1.51787925e+00 2.96586573e-01 -4.57837395e-02
-1.27442133e+00 -4.80725259e-01 1.53193831e-01 2.97877789e-01
-1.40612590e+00 -4.19106454e-01 8.68467629e-01 -4.34093803e-01
1.79648593e-01 2.19762236e-01 2.44726166e-01 4.66077209e-01
3.83846819e-01 8.47009063e-01 1.63306761e+00 -4.47563797e-01
5.62193096e-02 -2.88868934e-01 2.23912880e-01 6.70396209e-01
5.65446973e-01 -1.51417196e-01 -3.15056711e-01 -5.75744957e-02
7.86328554e-01 4.55320597e-01 -6.86662793e-01 -3.57007720e-02
-1.46913075e+00 6.83320105e-01 5.10387599e-01 6.31743729e-01
-9.30213630e-01 -2.86500037e-01 1.87682062e-01 -5.32637276e-02
5.13093054e-01 -2.11693108e-01 8.19543079e-02 4.15592462e-01
-1.03354955e+00 1.71258703e-01 1.84337065e-01 1.98760301e-01
8.63300800e-01 8.75678062e-02 -5.68419337e-01 1.01676095e+00
6.45501673e-01 5.51083803e-01 6.67168140e-01 -1.01728725e+00
3.45291853e-01 6.90704465e-01 2.96517536e-02 -1.42392766e+00
-1.37037337e-01 -4.60835516e-01 -1.36660802e+00 4.15717781e-01
-1.61857054e-01 9.42429453e-02 -8.67215812e-01 1.47097588e+00
2.29944304e-01 2.70822525e-01 5.45048177e-01 1.06994164e+00
1.05448306e+00 1.12633026e+00 -5.54784015e-02 -7.42962778e-01
1.68407679e+00 -7.92815387e-01 -1.04414892e+00 7.32209161e-02
-7.86316693e-02 -1.11544394e+00 6.27446711e-01 7.37562299e-01
-1.10590720e+00 -9.30933118e-01 -1.18032432e+00 -6.58750609e-02
-7.29049519e-02 4.76531386e-01 3.68432313e-01 3.24468613e-01
-8.83660316e-01 8.23369771e-02 -6.62584722e-01 -1.08718976e-01
2.36569569e-02 1.21308908e-01 -6.65239513e-01 -4.27625418e-01
-1.06371403e+00 8.52385223e-01 5.26941419e-01 2.50913918e-01
-5.88803530e-01 -1.49777800e-01 -1.11877835e+00 9.20696333e-02
3.00076753e-01 -6.35139108e-01 9.61564064e-01 -6.75706744e-01
-1.22516239e+00 4.50989485e-01 -5.60578644e-01 -3.28144401e-01
3.66320158e-03 -1.51699156e-01 -4.95557189e-01 3.47780377e-01
2.88181394e-01 2.83464313e-01 1.16363537e+00 -1.56082261e+00
-6.84233844e-01 -4.10551459e-01 -1.78625390e-01 4.28126216e-01
-1.50402799e-01 -1.46658272e-02 -5.45572519e-01 -1.16591024e+00
6.17603600e-01 -4.13959056e-01 -4.76823777e-01 -5.41783512e-01
-1.71433259e-02 -1.18219234e-01 8.96430433e-01 -1.02500510e+00
1.30361950e+00 -2.27066159e+00 4.67219681e-01 1.32231176e-01
2.58569509e-01 3.05395037e-01 -9.36153978e-02 1.81964830e-01
-2.95221299e-01 -2.77396470e-01 -3.81391466e-01 -3.18695843e-01
-3.90597165e-01 -7.23108649e-02 -2.50841588e-01 5.13420820e-01
1.44699188e-02 5.16062319e-01 -6.81377947e-01 -6.31740570e-01
5.13016105e-01 6.52028501e-01 -1.10469677e-01 2.56641656e-01
2.99421847e-01 6.07053280e-01 -3.76335621e-01 7.08350480e-01
9.31968570e-01 -1.17272273e-01 -3.18525136e-01 -8.61040771e-01
-2.08666354e-01 -4.32255715e-01 -1.31159484e+00 1.59126461e+00
-4.37193304e-01 2.79102057e-01 3.80932510e-01 -1.03594542e+00
1.25199687e+00 3.78721684e-01 6.40068114e-01 -4.62464899e-01
3.30129474e-01 6.26867935e-02 -1.65282845e-01 -3.47764671e-01
4.81809765e-01 -3.27878207e-01 -3.16180885e-02 -3.29547338e-02
-9.91908684e-02 4.15018760e-02 4.39195871e-01 1.75845444e-01
6.69531465e-01 -9.45960134e-02 7.06347346e-01 -1.78832427e-01
1.31691539e+00 -4.16555047e-01 8.83039594e-01 1.93152875e-01
3.17757800e-02 6.46411538e-01 9.79101583e-02 -1.67545483e-01
-5.39861560e-01 -9.97257173e-01 8.94664750e-02 5.13506234e-01
4.81306523e-01 -4.05696154e-01 -5.90574741e-01 -2.34983385e-01
-3.15515429e-01 3.64078045e-01 -2.66670227e-01 -1.50750086e-01
-1.80110425e-01 -8.38095903e-01 -1.71451211e-01 1.72937527e-01
1.05255079e+00 -7.55552411e-01 -2.77545184e-01 2.47144789e-01
-7.07355320e-01 -1.08829772e+00 -4.81133491e-01 -3.52354556e-01
-8.53140056e-01 -8.78255308e-01 -1.14375412e+00 -6.66419089e-01
8.79174948e-01 1.03789091e+00 6.29943252e-01 4.55601886e-02
-9.16233882e-02 2.19947681e-01 -7.30791152e-01 -5.69572821e-02
-2.62940496e-01 -3.59346092e-01 8.89822841e-02 6.59196079e-01
4.77568246e-02 -3.53993922e-01 -6.12243354e-01 2.73557663e-01
-1.24338448e+00 2.86734223e-01 1.06246901e+00 8.52172852e-01
9.31641459e-01 7.75426090e-01 2.87739515e-01 -3.99310052e-01
9.28729177e-01 -2.48435840e-01 -6.79956675e-01 4.36257720e-01
-3.21241379e-01 1.90993883e-02 6.33554399e-01 4.23918897e-03
-1.38642395e+00 1.05866052e-01 -3.20574753e-02 -6.00379109e-01
-7.38184378e-02 7.08939016e-01 -3.83753657e-01 -2.01928586e-01
2.68049717e-01 4.81026500e-01 6.95768371e-02 -6.95958912e-01
2.89397597e-01 8.22682023e-01 7.75850832e-01 -5.43679118e-01
7.61772454e-01 4.51670945e-01 2.22055003e-01 -8.53266537e-01
-8.69687915e-01 -6.74597561e-01 -4.50817674e-01 -2.12389857e-01
1.18125188e+00 -1.23628974e+00 -7.53218174e-01 4.83009934e-01
-9.92206752e-01 4.41685170e-01 8.30961168e-02 8.54392469e-01
-3.39270115e-01 7.47146368e-01 -3.90829086e-01 -7.93150485e-01
-3.38784009e-01 -1.57720709e+00 1.08567405e+00 5.25642335e-01
4.32404488e-01 -7.32614160e-01 -7.10725635e-02 5.77138186e-01
1.38439357e-01 4.32520270e-01 4.53109056e-01 -6.31330013e-02
-3.76466990e-01 -1.27139119e-02 -4.81338322e-01 7.54837275e-01
5.39196670e-01 -2.81214237e-01 -5.53382456e-01 -4.20461774e-01
4.04550105e-01 -1.38508910e-02 1.28237724e+00 5.36921978e-01
1.01260138e+00 -2.71014839e-01 -1.32814109e-01 4.27381963e-01
1.56429172e+00 2.55559176e-01 7.10562885e-01 9.67815965e-02
6.88918889e-01 4.14853662e-01 9.09894109e-01 3.25642049e-01
3.38252515e-01 6.72440052e-01 8.33375230e-02 -5.83650112e-01
-1.48894146e-01 1.15112603e-01 4.37234282e-01 1.20887268e+00
-2.07445607e-01 8.23004544e-02 -6.32851601e-01 3.42064083e-01
-2.21950364e+00 -8.87422621e-01 -1.21093102e-01 2.00020194e+00
5.09729624e-01 -1.21435240e-01 -1.98907077e-01 3.27747762e-01
9.69610989e-01 3.40342999e-01 -2.13790357e-01 -6.42693713e-02
-2.27247298e-01 7.87101761e-02 2.48030216e-01 4.42542136e-01
-1.18230104e+00 4.88460660e-01 6.02491236e+00 1.26654959e+00
-7.89290428e-01 1.94055349e-01 6.64733291e-01 5.74210882e-01
-1.18688293e-01 1.44663844e-02 -5.90489984e-01 4.39402312e-01
2.72154868e-01 -1.77624390e-01 3.37073386e-01 4.23263282e-01
4.46517944e-01 -4.21072841e-01 -4.48246360e-01 1.55015516e+00
4.52216774e-01 -1.03891110e+00 3.26757669e-01 -6.67525921e-03
7.97304869e-01 -6.07977509e-01 1.40033411e-02 2.05979366e-02
1.50605544e-01 -8.13354492e-01 2.12319791e-01 7.90848732e-01
4.47498500e-01 -9.26481128e-01 1.04943180e+00 2.95776695e-01
-1.62858891e+00 6.02639932e-03 -3.47623736e-01 -1.78836063e-02
2.43300885e-01 1.01058125e+00 7.88479075e-02 1.24896312e+00
5.78719437e-01 9.67750669e-01 -4.92410749e-01 9.71239626e-01
-3.42676461e-01 2.10418954e-01 8.83913636e-02 6.15442097e-01
-3.55862044e-02 -7.04361022e-01 5.64771354e-01 9.04138505e-01
5.00637472e-01 4.97202903e-01 6.06829345e-01 4.35542434e-01
8.59652821e-04 3.11054707e-01 -3.87455940e-01 1.98765054e-01
1.02843821e-01 1.62725782e+00 -8.12105477e-01 -4.42553043e-01
-5.18738389e-01 9.37456250e-01 -2.21093372e-01 3.51704419e-01
-5.43392003e-01 -3.67195010e-01 3.94116491e-01 -4.27031696e-01
2.63371080e-01 -2.90953428e-01 -1.51981816e-01 -1.67319560e+00
-6.95279613e-02 -1.08703983e+00 5.13183713e-01 -8.30547154e-01
-1.22760141e+00 5.05791306e-01 1.97825059e-01 -1.59105086e+00
1.17559060e-01 -3.70611429e-01 -6.06100440e-01 9.73698258e-01
-1.69469929e+00 -1.14774227e+00 -5.50466001e-01 8.14172268e-01
5.30605733e-01 -3.25982213e-01 4.59699631e-01 1.81852967e-01
-5.28479397e-01 1.51005819e-01 9.12903175e-02 -8.95046070e-03
5.60175359e-01 -8.25870454e-01 -3.63608927e-01 1.42345953e+00
6.79517463e-02 6.85110450e-01 6.00647032e-01 -5.97215116e-01
-1.12204647e+00 -9.98605072e-01 3.37796539e-01 3.60610366e-01
1.82951748e-01 2.85175264e-01 -9.29237425e-01 2.87772059e-01
4.07322764e-01 4.61403951e-02 6.96398556e-01 -2.40593255e-01
-1.78566769e-01 -4.92630810e-01 -1.17122769e+00 6.02680981e-01
3.52325052e-01 -2.18897179e-01 -9.30412352e-01 -8.79399180e-02
6.55508459e-01 -1.44203588e-01 -1.10688770e+00 8.47851396e-01
2.66973525e-01 -8.47300291e-01 1.07504642e+00 2.99954474e-01
2.98350215e-01 -1.06728482e+00 -3.20517808e-01 -1.31745565e+00
-7.41946280e-01 -3.33537370e-01 -6.28904626e-02 1.47026801e+00
-2.91852325e-01 -6.09397650e-01 1.41886041e-01 1.09421320e-01
1.20169163e-01 -5.79837561e-01 -6.98534727e-01 -5.67337036e-01
-6.30806923e-01 1.23873197e-01 3.60515773e-01 5.89261949e-01
-2.38194942e-01 4.09247965e-01 -4.87948716e-01 2.66784787e-01
9.86269891e-01 3.92959476e-01 5.26772976e-01 -1.16594398e+00
-1.50091395e-01 -3.63789827e-01 -5.64597130e-01 -7.28927016e-01
1.06622644e-01 -6.03311777e-01 1.62010133e-01 -1.84843981e+00
4.80848014e-01 1.62493318e-01 -6.87460899e-01 4.46594924e-01
-5.58969259e-01 5.09086370e-01 4.77907091e-01 4.44730014e-01
-6.39909089e-01 7.68966675e-01 1.16638505e+00 -1.99559495e-01
-1.35366455e-01 -5.13820015e-02 -9.44305420e-01 8.10580969e-01
7.11506665e-01 -3.22426669e-02 -2.82608956e-01 -1.23197086e-01
-2.44696662e-01 3.30930710e-01 2.46477738e-01 -1.35241091e+00
1.32222652e-01 -1.59743562e-01 4.91890430e-01 -7.97649443e-01
3.97997677e-01 -9.16545153e-01 3.45762849e-01 4.83103842e-01
-1.34634793e-01 -9.25627127e-02 -2.68979240e-02 3.82932186e-01
-8.52950037e-01 -2.18851075e-01 9.22784328e-01 -2.94443935e-01
-8.01100731e-01 3.18716913e-01 -3.57413918e-01 -5.65545022e-01
1.07090163e+00 -6.98675439e-02 -2.20562816e-01 -3.10717911e-01
-6.54736161e-01 1.37149677e-01 2.92519182e-01 3.49802315e-01
9.60662484e-01 -1.61120260e+00 -9.23363447e-01 9.64663401e-02
1.31993070e-01 -2.13348836e-01 4.56444085e-01 9.26821887e-01
-3.43370050e-01 1.87062055e-01 -2.88553268e-01 -4.79645103e-01
-1.63627684e+00 7.27683663e-01 -2.11218551e-01 -5.34231246e-01
-3.26387316e-01 2.99808800e-01 6.22541189e-01 2.21228451e-01
-6.73413947e-02 -1.70739308e-01 -7.28149772e-01 2.00437978e-01
8.73271942e-01 4.27994490e-01 3.67600173e-02 -1.30930376e+00
-2.91645378e-01 1.03649807e+00 -4.26961966e-02 -1.75311252e-01
1.09631777e+00 -4.39491093e-01 -6.11650348e-01 3.02396715e-01
1.25602019e+00 1.03011489e-01 -9.53059793e-01 -3.78197610e-01
-3.77307624e-01 -6.69394553e-01 5.40461600e-01 -3.92031252e-01
-1.15655088e+00 8.51942837e-01 8.90534282e-01 1.05978698e-01
1.83264768e+00 -2.54617304e-01 7.52920330e-01 1.07795693e-01
4.60035533e-01 -6.17757916e-01 1.07739031e-01 1.53449222e-01
1.02576602e+00 -1.27244508e+00 4.49616939e-01 -5.31046152e-01
-5.47057092e-01 1.06062901e+00 3.18260819e-01 -1.64271772e-01
5.64901531e-01 -1.09443821e-01 -8.14136025e-03 1.90763827e-02
-4.10037935e-01 -6.82099342e-01 6.17191136e-01 3.51320028e-01
3.59739631e-01 -4.86402288e-02 -7.72159576e-01 5.74864805e-01
3.56490701e-01 1.59268185e-01 1.80072576e-01 6.49641633e-01
-7.09917307e-01 -1.16708922e+00 -1.07128012e+00 3.69016349e-01
-7.16824234e-01 1.23404860e-01 2.41882965e-01 5.94295524e-02
4.31647182e-01 1.30471313e+00 -3.08030576e-01 -4.90139931e-01
1.81823045e-01 -4.79567707e-01 2.97014564e-01 -5.00780642e-01
-2.10796073e-01 5.66298902e-01 -4.91455257e-01 -4.03143615e-01
-1.07552063e+00 -5.76966941e-01 -1.05683756e+00 8.91112238e-02
-3.66216749e-01 3.42728823e-01 4.73847061e-01 7.84750640e-01
5.46368957e-02 6.49812996e-01 8.35473955e-01 -1.11052144e+00
2.48471443e-02 -7.85287619e-01 -6.53676271e-01 5.06093264e-01
4.89004940e-01 -8.75025153e-01 -2.04189613e-01 3.77241969e-01] | [10.589327812194824, -1.904952049255371] |
35021838-071b-414b-bc10-54da7be9ab4f | dynamics-of-order-positions-and-related | 1505.04810 | null | http://arxiv.org/abs/1505.04810v2 | http://arxiv.org/pdf/1505.04810v2.pdf | Dynamics of Order Positions and Related Queues in a Limit Order Book | Order positions are key variables in algorithmic trading. This paper studies
the limiting behavior of order positions and related queues in a limit order
book. In addition to the fluid and diffusion limits for the processes,
fluctuations of order positions and related queues around their fluid limits
are analyzed. As a corollary, explicit analytical expressions for various
quantities of interests in a limit order book are derived. | [] | 2015-10-13 | null | null | null | null | ['algorithmic-trading'] | ['time-series'] | [-8.61231267e-01 -4.62266326e-01 1.27714381e-01 -1.11545287e-02
1.33236155e-01 -1.12429118e+00 7.36120343e-01 1.83530405e-01
-5.82619190e-01 1.06618369e+00 -3.47548336e-01 -3.38070571e-01
-5.61302185e-01 -7.25168884e-01 -2.89258122e-01 -6.51828945e-01
-6.26239419e-01 1.01710331e+00 5.16248345e-01 -3.58824372e-01
5.63949287e-01 9.36796963e-01 -9.93892312e-01 -1.87778950e-01
6.01019919e-01 1.46978927e+00 -3.21148336e-01 9.83147979e-01
-6.75502300e-01 1.12250280e+00 -8.84958506e-01 -4.48336095e-01
8.80032122e-01 -6.80442631e-01 4.97547388e-02 -8.48375484e-02
-5.24641812e-01 -8.03431094e-01 -5.11145115e-01 1.33848453e+00
2.36220863e-02 1.63312092e-01 9.27549839e-01 -1.41248894e+00
-6.51944041e-01 8.65178049e-01 -5.18120706e-01 1.05979466e+00
-4.71951663e-01 -1.15686871e-01 1.23318076e+00 -3.82837683e-01
1.11909524e-01 1.08871603e+00 3.03045005e-01 -2.07280321e-03
-9.85327959e-01 -7.70896018e-01 1.03697799e-01 -4.33261603e-01
-7.45393693e-01 6.51280954e-02 5.00364065e-01 -2.86261261e-01
6.26990557e-01 3.68252814e-01 1.11834085e+00 1.52432635e-01
1.49462891e+00 5.72199702e-01 1.25990570e+00 -2.18329597e-02
4.90801901e-01 1.30198881e-01 5.17224312e-01 -3.22203040e-02
1.00230134e+00 2.93362886e-01 -8.87082294e-02 -5.97095311e-01
1.44937181e+00 3.18561196e-01 2.02128112e-01 -2.29613796e-01
-6.86357915e-01 8.83031428e-01 -7.10382462e-02 1.61569551e-01
-7.08440542e-01 2.89663911e-01 1.29751995e-01 6.85692728e-01
7.02999294e-01 4.53785032e-01 -3.65520954e-01 -3.34565878e-01
-5.98553836e-01 6.20608807e-01 1.69390118e+00 1.14653528e+00
1.26832008e-01 -1.96451440e-01 -2.25191250e-01 5.88164516e-02
1.11944400e-01 8.67414355e-01 3.01945835e-01 -1.19552922e+00
3.67159337e-01 3.77072468e-02 8.32054615e-01 -4.61822003e-01
-3.94919187e-01 -7.34346449e-01 -7.30119228e-01 -7.62563245e-03
7.75236428e-01 -3.60837817e-01 -5.17799892e-02 7.91636527e-01
-1.20521367e-01 -4.02208656e-01 9.51584503e-02 6.40822291e-01
-2.25431189e-01 7.57799029e-01 -3.81854445e-01 -1.02642250e+00
1.08999968e+00 -6.81275487e-01 -1.23210752e+00 4.30366427e-01
-1.59803644e-01 -9.24140930e-01 3.54270071e-01 3.20907176e-01
-1.38372278e+00 8.12197030e-02 -5.48564970e-01 7.00359523e-01
-7.97438696e-02 -8.91221285e-01 4.52135265e-01 1.52589321e-01
-1.00813496e+00 9.99821901e-01 -5.53280532e-01 2.58401096e-01
3.63702513e-02 3.02986175e-01 9.91094470e-01 9.87514138e-01
-1.06135380e+00 4.34097707e-01 -2.75475621e-01 3.52347344e-01
-1.32087171e-01 -9.32140827e-01 1.48214489e-01 9.60952118e-02
2.37406790e-01 -2.61354893e-01 1.98218989e+00 -1.52154163e-01
-1.47900164e+00 3.18389475e-01 2.97233224e-01 -8.61352861e-01
1.22690928e+00 -4.18999791e-01 -3.15402806e-01 3.38407218e-01
9.84268710e-02 -5.69375992e-01 4.90984917e-01 -9.22361255e-01
-9.49234545e-01 -1.79142043e-01 -2.39429623e-01 2.84624279e-01
1.76213562e-01 3.57186384e-02 6.86464667e-01 -5.28767407e-01
1.71840668e-01 -7.16415942e-01 -4.76631969e-01 -3.72195572e-01
-2.16733813e-02 -2.04608068e-01 2.47331649e-01 -2.92273611e-02
1.01571977e+00 -1.79003847e+00 -6.95616722e-01 3.55424136e-01
5.07547736e-01 -4.18096274e-01 6.13143742e-01 8.44280422e-01
7.00723767e-01 3.34592760e-01 5.68712473e-01 5.20390809e-01
5.44656038e-01 2.23887071e-01 -8.76457810e-01 7.10251868e-01
-2.04517737e-01 1.02423823e+00 -8.37631106e-01 -1.03746243e-01
-3.70952748e-02 -1.37510151e-01 -3.02369475e-01 2.77317464e-01
-1.10173494e-01 1.89760819e-01 -9.28189516e-01 2.17739210e-01
9.08009648e-01 -3.18856716e-01 -3.16565596e-02 4.91478682e-01
-5.87950528e-01 1.32795557e-01 -7.45591164e-01 -2.29248300e-01
-2.86686093e-01 3.24278265e-01 5.81988096e-01 -2.14029461e-01
1.03157449e+00 2.41718609e-02 7.19922125e-01 -7.30923653e-01
4.50698644e-01 5.78719616e-01 5.74251473e-01 2.94995680e-02
6.40869856e-01 -7.88297117e-01 -1.71530738e-01 1.09733212e+00
-4.99545842e-01 -2.43055418e-01 4.04482692e-01 2.66866952e-01
9.09784555e-01 -5.15045822e-01 6.15843199e-02 -9.04362798e-01
2.49717668e-01 -2.18844280e-01 2.11022869e-01 8.11975479e-01
-5.62957048e-01 -3.76235813e-01 1.17474985e+00 -5.28821528e-01
-1.23262370e+00 -1.45307493e+00 -2.85640836e-01 8.19141269e-01
6.78299367e-01 1.09649681e-01 -5.43515027e-01 -4.66811694e-02
7.95962512e-01 5.50695777e-01 -5.62237442e-01 3.13463539e-01
-6.71325803e-01 -9.18399572e-01 -2.45104671e-01 2.20119014e-01
4.70172673e-01 -1.13892400e+00 -7.94458568e-01 5.42024076e-01
7.42429197e-01 -1.13613641e+00 -8.11525702e-01 4.41871434e-01
-1.01629579e+00 -1.03311849e+00 -9.12444055e-01 2.25772671e-02
4.98953104e-01 -1.99927285e-01 1.06669497e+00 -2.10203379e-01
3.47754687e-01 3.53742957e-01 4.45491932e-02 -5.62083483e-01
-5.73058248e-01 -4.58666719e-02 1.90602750e-01 3.58665735e-01
2.91693985e-01 -3.87753904e-01 -1.14468837e+00 5.01654029e-01
-6.94407582e-01 -7.83185422e-01 2.91697234e-01 3.24357808e-01
5.44637859e-01 2.90317863e-01 3.25285047e-01 -7.88838863e-01
1.63357282e+00 -1.89020112e-01 -1.61227584e+00 1.06723532e-02
-9.55398560e-01 2.74290174e-01 9.09378231e-01 -4.80861396e-01
-9.31902766e-01 -8.25593948e-01 8.70115578e-01 -2.87414283e-01
4.87396538e-01 -2.56511271e-01 2.42720082e-01 1.76719278e-01
-4.34971809e-01 3.27008307e-01 3.80055279e-01 -6.37259722e-01
6.73308522e-02 4.06582654e-01 2.54923195e-01 -3.74282181e-01
8.61090481e-01 5.05011380e-01 4.88272488e-01 -7.93878973e-01
-5.85630298e-01 -1.35830924e-01 -7.46639907e-01 -1.70499504e-01
1.70309395e-01 -1.11501478e-01 -1.54335892e+00 6.02433860e-01
-9.86680090e-01 -2.90562898e-01 -8.69915962e-01 5.43094277e-01
-8.49404097e-01 -1.45755172e-01 -1.61829937e+00 -1.70040882e+00
-9.29705575e-02 -5.48437417e-01 3.18296790e-01 5.21885753e-01
-1.89747483e-01 -1.35083151e+00 1.46131128e-01 -4.63068634e-01
6.94087744e-01 1.45460963e-01 7.63883710e-01 -1.26597416e+00
-9.60146368e-01 -3.37040871e-01 -1.28284946e-01 -5.75530864e-02
1.34519100e-01 -1.84435308e-01 5.30447699e-02 -2.06855476e-01
7.37401962e-01 5.95595658e-01 3.73211056e-01 8.53671789e-01
2.56968499e-03 -7.79949963e-01 -3.18397790e-01 4.13208097e-01
1.22782588e+00 1.10329199e+00 -1.76007792e-01 5.91393352e-01
-1.59105122e-01 6.00333750e-01 6.80595219e-01 9.07335103e-01
-3.13666135e-01 -8.57465491e-02 -1.83537841e-01 4.47160959e-01
7.46488273e-01 -2.99027264e-01 1.10975482e-01 1.02452326e+00
1.54123649e-01 -2.03251526e-01 -6.14394724e-01 4.74797636e-02
-1.57719159e+00 -9.46477532e-01 -3.41505885e-01 1.97321391e+00
5.70536137e-01 4.23366904e-01 6.57626092e-01 -4.22525495e-01
8.96064878e-01 2.73248125e-02 -1.07949650e+00 -5.14604390e-01
-2.65903533e-01 -1.02713600e-01 1.31162751e+00 8.06590796e-01
-4.99845892e-01 5.89010835e-01 8.20061493e+00 3.35056841e-01
-7.38145769e-01 -6.51000738e-02 7.82802761e-01 -5.51366091e-01
-3.73719573e-01 -5.66618480e-02 -1.19568729e+00 7.65182912e-01
9.81388152e-01 -9.71644282e-01 6.69924736e-01 4.17167425e-01
3.97371680e-01 -2.77770132e-01 -7.56214201e-01 8.46888185e-01
-9.26146865e-01 -1.13228714e+00 -9.45725590e-02 9.54820752e-01
5.32254457e-01 -4.78595793e-01 4.25025940e-01 -1.82440192e-01
7.30607271e-01 -5.05850852e-01 8.44456971e-01 8.57678831e-01
1.97991177e-01 -1.15336156e+00 1.07107854e+00 1.82745442e-01
-1.04280269e+00 -1.86030522e-01 -5.97216666e-01 -6.12477005e-01
6.45453513e-01 7.20232368e-01 -9.21716914e-02 -9.09243152e-02
4.78589684e-01 3.12399983e-01 -9.62517112e-02 9.28719103e-01
2.76325732e-01 3.76507849e-01 -4.65134829e-01 -9.02607441e-01
3.71376038e-01 -1.16248429e+00 4.55245256e-01 5.26278019e-01
2.25471243e-01 3.97991955e-01 -2.12592602e-01 1.01405644e+00
3.86608355e-02 2.57989243e-02 -3.12845260e-01 -5.54760635e-01
5.00504076e-01 8.74279559e-01 -1.46765876e+00 -2.56276309e-01
-6.97521716e-02 8.92256081e-01 -4.45565075e-01 5.38507342e-01
-5.47046602e-01 -6.13384068e-01 8.33218157e-01 6.99278533e-01
3.14352959e-01 -3.70310783e-01 -1.36335522e-01 -7.58952200e-01
5.04267626e-02 -5.69166578e-02 7.63362572e-02 1.10256501e-01
-1.73733723e+00 6.65998161e-01 2.73595992e-02 -1.11516523e+00
-4.11049336e-01 -5.01718342e-01 -4.89995211e-01 7.88514555e-01
-1.23008108e+00 1.94667593e-01 7.77996659e-01 2.71229237e-01
1.70410231e-01 -5.47380269e-01 -2.15838343e-01 -3.70601773e-01
-1.85299903e-01 -1.16932727e-02 1.31159520e+00 1.43633619e-01
-1.72367454e-01 -1.44513452e+00 6.45379961e-01 2.17462614e-01
-2.12333888e-01 6.05092287e-01 9.61814582e-01 -1.07946670e+00
-1.41234469e+00 -4.25893366e-01 6.70277596e-01 -2.36715272e-01
1.75538945e+00 -5.37354589e-01 -5.91664791e-01 3.98782223e-01
4.60050762e-01 -1.38850687e-02 3.56032312e-01 -4.04306024e-01
5.96075773e-01 -3.14917713e-01 -8.61537039e-01 3.93747061e-01
8.58780086e-01 -1.63077906e-01 -2.05309957e-01 2.27618709e-01
6.08460188e-01 -2.48646975e-01 -7.54353762e-01 -2.27412969e-01
8.07957470e-01 -1.12315285e+00 4.67514634e-01 -5.11413872e-01
-3.22596133e-01 3.08132887e-01 7.89607037e-03 -1.17294824e+00
-2.02680320e-01 -1.72097111e+00 -2.14424998e-01 1.22019076e+00
1.60507351e-01 -1.45046830e+00 6.79059029e-01 7.69714236e-01
8.62702608e-01 -6.67855203e-01 -1.05502498e+00 -1.46662617e+00
5.58918118e-01 4.26709682e-01 7.02501416e-01 2.02400424e-02
-8.43248889e-02 2.97409475e-01 -2.01028176e-02 -3.51592660e-01
1.15353572e+00 6.58479214e-01 1.04335407e-02 -1.25303721e+00
-1.53019026e-01 -9.65406299e-01 8.97160694e-02 -1.27409041e+00
3.25923599e-02 -4.56549883e-01 -1.50243029e-01 -1.06098068e+00
-3.38937826e-02 -5.09397149e-01 -5.64984143e-01 -1.03565454e+00
4.17769074e-01 -2.41445750e-01 6.09856248e-01 9.76820469e-01
-7.59106219e-01 6.04537189e-01 1.53437543e+00 3.34797204e-01
-3.65862072e-01 8.50752234e-01 -4.61984187e-01 4.22618240e-01
1.14187515e+00 -3.21333468e-01 -2.81594515e-01 4.80261236e-01
3.18623722e-01 4.16733712e-01 -7.19017014e-02 -4.38701630e-01
3.62076521e-01 -3.75332147e-01 2.39798322e-01 -1.19984746e+00
-5.44058718e-02 -7.59120405e-01 1.67159036e-01 9.31962371e-01
-7.25808382e-01 8.56281459e-01 -2.10165635e-01 7.47316003e-01
-4.76584118e-03 -4.53605056e-01 5.20613432e-01 -2.66184449e-01
2.26348341e-01 3.25004965e-01 -9.53401089e-01 5.39755762e-01
1.27935004e+00 -6.40292764e-02 -2.41674677e-01 -8.90219331e-01
-6.72752202e-01 5.92595398e-01 2.06683651e-01 6.26758039e-02
3.02161723e-01 -1.09679186e+00 -2.37009883e-01 1.27243936e-01
-9.40890312e-01 -5.70573032e-01 -3.28240842e-01 7.61382759e-01
-1.00066447e+00 8.10803056e-01 -2.01667488e-01 -2.59577692e-01
-5.26000440e-01 7.52162397e-01 7.59787917e-01 -5.12602150e-01
-4.93351132e-01 2.87648022e-01 2.14253291e-01 2.63967603e-01
-2.65209079e-02 -6.55422688e-01 2.91118026e-01 3.58035594e-01
7.81141877e-01 8.00623775e-01 -5.19574761e-01 -2.38073856e-01
1.36378482e-01 5.07560015e-01 -1.87544346e-01 -5.40795386e-01
1.30561256e+00 -6.42935693e-01 -6.10171080e-01 1.24139702e+00
7.85449088e-01 6.97492212e-02 -1.65769184e+00 -1.24894127e-01
5.10022283e-01 -5.78793943e-01 -5.16248703e-01 -3.34810138e-01
-9.65415061e-01 7.36246228e-01 -9.67149064e-03 1.47386277e+00
4.59257513e-01 2.22945452e-01 1.06255472e+00 -2.39031538e-02
4.94006395e-01 -1.16555858e+00 -3.26666653e-01 7.51542330e-01
7.70306110e-01 -4.36595619e-01 -2.92281330e-01 -3.23454231e-01
-3.83911669e-01 1.09895349e+00 7.00301602e-02 -6.56213880e-01
1.36864996e+00 1.09255707e+00 1.96231112e-01 -1.12759843e-01
-8.62573266e-01 1.98525473e-01 -2.85602599e-01 4.09827173e-01
3.71471405e-01 1.48491070e-01 -6.22193098e-01 7.87919998e-01
-6.66149318e-01 -1.48362041e-01 7.52840281e-01 8.30970764e-01
-6.38990760e-01 -7.25634694e-01 -4.27589685e-01 8.64344478e-01
-9.15882051e-01 -8.58922005e-02 -5.41654229e-01 1.07063520e+00
-8.27747166e-01 6.74093187e-01 9.27667439e-01 2.26466015e-01
4.21584070e-01 -1.22372754e-01 4.11903828e-01 -1.22039495e-02
-7.57631540e-01 6.16235912e-01 -4.19616520e-01 -3.49711686e-01
2.50876427e-01 -8.43252897e-01 -1.32476568e+00 -1.10112071e+00
-2.15265200e-01 9.84144270e-01 -1.43843591e-01 6.66515648e-01
-1.66255057e-01 2.56713659e-01 8.11310112e-01 -3.47580522e-01
-1.53435898e+00 -6.62420571e-01 -2.09092498e+00 7.31476620e-02
6.65894806e-01 -3.92321229e-01 -7.77264655e-01 -6.27236962e-01] | [4.8436994552612305, 3.987200975418091] |
9f88f5c3-d12b-4401-a192-2231e6797768 | spatially-resolved-gene-expression-prediction | 2306.01859 | null | https://arxiv.org/abs/2306.01859v1 | https://arxiv.org/pdf/2306.01859v1.pdf | Spatially Resolved Gene Expression Prediction from H&E Histology Images via Bi-modal Contrastive Learning | Histology imaging is an important tool in medical diagnosis and research, enabling the examination of tissue structure and composition at the microscopic level. Understanding the underlying molecular mechanisms of tissue architecture is critical in uncovering disease mechanisms and developing effective treatments. Gene expression profiling provides insight into the molecular processes underlying tissue architecture, but the process can be time-consuming and expensive. In this study, we present BLEEP (Bi-modaL Embedding for Expression Prediction), a bi-modal embedding framework capable of generating spatially resolved gene expression profiles of whole-slide Hematoxylin and eosin (H&E) stained histology images. BLEEP uses a contrastive learning framework to construct a low-dimensional joint embedding space from a reference dataset using paired image and expression profiles at micrometer resolution. With this framework, the gene expression of any query image patch can be imputed using the expression profiles from the reference dataset. We demonstrate BLEEP's effectiveness in gene expression prediction by benchmarking its performance on a human liver tissue dataset captured via the 10x Visium platform, where it achieves significant improvements over existing methods. Our results demonstrate the potential of BLEEP to provide insights into the molecular mechanisms underlying tissue architecture, with important implications in diagnosis and research of various diseases. The proposed framework can significantly reduce the time and cost associated with gene expression profiling, opening up new avenues for high-throughput analysis of histology images for both research and clinical applications. | ['Bo wang', 'Gary D. Bader', 'Kuan Pang', 'Ronald Xie'] | 2023-06-02 | null | null | null | null | ['medical-diagnosis'] | ['medical'] | [ 6.57741129e-02 -5.94946980e-01 -9.62904096e-02 -1.33777469e-01
-1.00046933e+00 -7.14767933e-01 2.14578155e-02 4.64817137e-01
-3.24999303e-01 4.52147365e-01 1.37703672e-01 -9.79607180e-02
-1.83070540e-01 -7.39536107e-01 -9.87304822e-02 -1.62650752e+00
-4.06967372e-01 2.44140461e-01 -3.63997787e-01 1.93533689e-01
-1.14520594e-01 7.55402923e-01 -1.04538405e+00 1.53544605e-01
1.54823601e-01 8.75690699e-01 2.27768756e-02 7.98265934e-01
-4.58001764e-03 4.41810757e-01 -1.49155810e-01 -1.55186772e-01
1.90263465e-01 -3.66229415e-01 -6.06230438e-01 1.48970515e-01
2.10008711e-01 -1.03985779e-01 -3.42687458e-01 1.08096623e+00
7.92034745e-01 -2.54978329e-01 7.84640312e-01 -9.11048412e-01
-4.17238712e-01 3.38766500e-02 -5.77032089e-01 3.15003633e-01
-3.01222336e-02 3.24535877e-01 9.01703954e-01 -8.69861424e-01
6.91846728e-01 6.24250531e-01 5.92302561e-01 5.34816265e-01
-1.75140750e+00 -3.76010299e-01 -7.72738516e-01 2.86145508e-01
-1.71012211e+00 -3.42369258e-01 6.04193985e-01 -8.56991053e-01
5.92674792e-01 5.70593596e-01 9.61190104e-01 5.77210307e-01
7.37964153e-01 4.87158626e-01 1.49938166e+00 -1.76890910e-01
2.63201684e-01 -1.67605475e-01 -1.27324328e-01 1.18903410e+00
-1.14983909e-01 -2.90466659e-02 -2.99127698e-01 -3.41455013e-01
8.10505092e-01 2.59081572e-01 -5.58953643e-01 -2.97231495e-01
-1.51174152e+00 7.00402915e-01 6.43469632e-01 3.15794528e-01
-5.47094941e-01 2.06660613e-01 5.62936366e-01 2.17362329e-01
3.36417824e-01 5.07470608e-01 -1.98935285e-01 1.12513326e-01
-8.99151266e-01 -3.04208308e-01 4.50329840e-01 2.24428549e-01
8.15309703e-01 -2.93540984e-01 -1.21664762e-01 8.25716078e-01
4.72136408e-01 2.36503959e-01 5.64929783e-01 -8.34235966e-01
-5.11971354e-01 7.68845439e-01 -3.17761958e-01 -1.02943027e+00
-4.24979031e-01 -2.70344496e-01 -1.12355840e+00 2.78397590e-01
3.05843383e-01 1.14587724e-01 -6.11325860e-01 1.37916648e+00
5.62296152e-01 3.34663361e-01 -1.37913615e-01 1.13621116e+00
6.96057320e-01 6.27203107e-01 2.98860550e-01 -1.96177289e-01
1.59080231e+00 -5.65696716e-01 -4.64583874e-01 3.32944334e-01
7.30555832e-01 -5.83351731e-01 9.15831983e-01 7.80421495e-02
-4.51190770e-01 1.68475002e-01 -9.18244481e-01 -1.03039250e-01
-2.34091461e-01 2.97292084e-01 6.16452992e-01 3.17961752e-01
-1.17544675e+00 3.94171506e-01 -1.19350648e+00 -5.62531114e-01
5.61680019e-01 5.44121265e-01 -9.74937797e-01 -3.59369636e-01
-6.56254411e-01 4.87102240e-01 -1.55975953e-01 3.11072916e-01
-8.71745467e-01 -1.06935179e+00 -7.90873885e-01 1.65343776e-01
-2.30275735e-01 -9.69685972e-01 7.50131726e-01 -1.81024551e-01
-1.47170770e+00 1.17427135e+00 -7.37476721e-02 2.30416693e-02
1.01152219e-01 5.24848819e-01 -7.18466332e-03 3.83844674e-01
5.07844202e-02 4.88306761e-01 2.76669919e-01 -7.30597854e-01
-4.73568700e-02 -7.22149968e-01 -5.51756382e-01 -1.11749724e-01
-1.92999244e-01 -9.89448130e-02 -3.22753042e-01 -3.63371253e-01
3.27315964e-02 -9.84994888e-01 -4.26011503e-01 6.74188673e-01
-2.66420275e-01 1.93947032e-01 6.57987118e-01 -6.20131195e-01
8.12635779e-01 -2.34614825e+00 3.55958760e-01 2.25850239e-01
4.31928635e-01 -3.76917049e-02 -2.18094558e-01 4.69730586e-01
-8.56844559e-02 2.65307218e-01 3.96956839e-02 8.39445069e-02
-1.21068351e-01 -3.31842005e-02 2.27718726e-01 1.14391398e+00
7.58369714e-02 1.22591305e+00 -9.30262446e-01 -6.35928392e-01
5.81356287e-01 8.77889395e-01 -1.85906872e-01 3.54404151e-01
1.63347244e-01 5.43861806e-01 -4.01492745e-01 8.71663749e-01
2.67648399e-01 -6.40877068e-01 4.23439115e-01 -6.69199526e-01
5.76663855e-03 -3.08592170e-01 -4.55590963e-01 1.55843389e+00
-4.81638312e-01 8.22400630e-01 4.40475971e-01 -8.42843056e-01
6.32806063e-01 2.95167238e-01 7.54544616e-01 -4.04244035e-01
4.49241221e-01 -1.18925534e-01 -9.27051231e-02 -8.13075900e-01
-4.95650768e-01 -7.32282162e-01 4.20359485e-02 4.25549477e-01
7.06667751e-02 -1.26087889e-01 -1.89772900e-02 -2.06269503e-01
1.36220324e+00 -4.48824525e-01 3.88332754e-01 -4.00152117e-01
5.37737012e-01 7.58484900e-02 3.79159898e-01 -6.00043610e-02
-5.55108964e-01 3.54730129e-01 5.18731594e-01 -6.00761414e-01
-1.20371366e+00 -1.18610585e+00 -4.42553759e-01 7.28756607e-01
-1.33351743e-01 -2.52035826e-01 -3.80013973e-01 -5.80979764e-01
7.87795410e-02 -3.07442099e-01 -1.00490475e+00 1.00057185e-01
-1.42123073e-01 -1.30058408e+00 5.74464679e-01 2.98882604e-01
2.75092907e-02 -5.51440060e-01 -3.09649020e-01 2.07194924e-01
-1.28031000e-02 -7.03687370e-01 -3.65231037e-01 1.66882738e-01
-8.56761754e-01 -1.30546987e+00 -6.10094309e-01 -9.66272235e-01
1.12187362e+00 1.50688916e-01 7.52416730e-01 2.90490866e-01
-1.30697155e+00 4.48714554e-01 2.57592183e-02 4.24635187e-02
-3.75189722e-01 -2.21944347e-01 -1.26127675e-01 -4.68225926e-02
2.07307652e-01 -6.01170182e-01 -1.15846741e+00 1.26153961e-01
-1.05864501e+00 -9.00951214e-03 8.55272412e-01 1.27007711e+00
1.00913620e+00 2.88377646e-02 3.70257497e-01 -8.51217985e-01
4.32626307e-01 -5.68981469e-01 -6.37360215e-01 2.56692350e-01
-5.04838645e-01 -5.08953780e-02 4.83211100e-01 2.27029800e-01
-6.53561115e-01 1.24578550e-01 -3.45066607e-01 -3.41698706e-01
-5.50335385e-02 8.89781177e-01 1.57302931e-01 -4.11929160e-01
4.36427087e-01 3.27542335e-01 5.75548053e-01 -2.19186887e-01
1.01882573e-02 6.10483527e-01 4.70484823e-01 -2.79029965e-01
2.79993981e-01 7.65688956e-01 5.98593533e-01 -9.38407600e-01
-4.63736653e-01 -9.40960407e-01 -6.20589912e-01 -9.70448628e-02
7.91290283e-01 -8.99534285e-01 -8.05065691e-01 2.99991816e-01
-4.57106352e-01 -4.73239511e-01 2.17013866e-01 4.40064281e-01
-4.37465459e-01 3.56091529e-01 -1.30833793e+00 -1.07957296e-01
-6.50696874e-01 -1.22655702e+00 1.23301113e+00 -1.80487167e-02
-1.80363387e-01 -1.37456167e+00 6.16562784e-01 4.12639707e-01
2.77656972e-01 5.09346366e-01 1.55650902e+00 -4.11625020e-02
-7.51256406e-01 -4.05426800e-01 -1.76533878e-01 2.40114287e-01
4.08902138e-01 2.78345853e-01 -6.57659590e-01 -4.91518289e-01
-2.72153258e-01 -2.66362190e-01 6.22906029e-01 2.90915310e-01
1.11924863e+00 -4.01845932e-01 -4.14818555e-01 8.45557392e-01
2.09557819e+00 -3.11960280e-01 6.06657624e-01 1.93733707e-01
3.91116947e-01 6.36723459e-01 4.09033418e-01 3.55846792e-01
1.44170687e-01 5.83160102e-01 4.05440420e-01 -4.48856324e-01
-2.17211723e-01 8.88351426e-02 3.01725138e-02 9.01107311e-01
2.58540303e-01 -1.64670292e-02 -1.04481626e+00 9.50668395e-01
-1.36538184e+00 -8.61840725e-01 -8.30908343e-02 1.98276377e+00
8.12010407e-01 -8.78007770e-01 -1.19076088e-01 -1.13164634e-01
5.09891391e-01 -1.60104603e-01 -3.77781242e-01 -1.30044550e-01
-2.23499946e-02 3.37873995e-01 2.43964747e-01 3.17719638e-01
-1.00981200e+00 3.88491273e-01 7.09811592e+00 4.40049410e-01
-1.75594544e+00 1.25748008e-01 6.55835807e-01 -2.17392921e-01
-4.51526307e-02 -4.42302078e-01 -1.03578277e-01 3.27817738e-01
7.88102388e-01 -5.51649332e-01 2.73203075e-01 6.37972355e-01
3.50533873e-01 1.64831169e-02 -1.16243529e+00 8.67230415e-01
-1.72203556e-01 -1.70748675e+00 -1.96170628e-01 3.81264985e-01
5.59174478e-01 3.00259292e-01 1.92859262e-01 -6.42834902e-02
-3.16179991e-02 -1.26992834e+00 -3.23669761e-01 2.45147809e-01
1.23303831e+00 -5.36924541e-01 1.25386369e+00 3.82494032e-02
-8.44874084e-01 1.80782482e-01 -2.26673767e-01 2.60585129e-01
-8.41174275e-02 7.49614358e-01 -1.60070074e+00 2.70374447e-01
5.03413379e-01 6.05275452e-01 -4.12812859e-01 1.12178135e+00
-8.01837444e-02 5.67825377e-01 -2.55141240e-02 1.10376514e-01
6.68722987e-02 -3.52420360e-01 2.89645225e-01 1.47261500e+00
3.77167404e-01 1.51354939e-01 1.03016578e-01 7.15418041e-01
-1.21727638e-01 3.54461730e-01 -3.79557818e-01 -2.85704195e-01
2.74870753e-01 1.88456190e+00 -9.12405550e-01 9.54298116e-03
-3.41222584e-01 9.72598672e-01 3.30826402e-01 1.75415859e-01
-5.32744348e-01 -8.75233859e-02 1.09478974e+00 1.55095279e-01
-2.05174973e-03 -6.28391653e-02 6.80701807e-02 -9.53845680e-01
-4.39410836e-01 -5.43727040e-01 3.33995551e-01 -3.82889122e-01
-1.52868676e+00 2.79800564e-01 -5.88402689e-01 -1.06813061e+00
-7.04742819e-02 -9.24004495e-01 -4.96936500e-01 8.14293504e-01
-1.53803158e+00 -1.20615566e+00 -5.22693992e-01 2.38205984e-01
1.18761301e-01 2.14318410e-01 1.43115604e+00 3.79195124e-01
-8.14225614e-01 2.68690497e-01 5.50905943e-01 3.04669589e-01
6.70513988e-01 -1.31818163e+00 -4.45859104e-01 3.07537466e-01
1.23391837e-01 7.41938233e-01 6.64165556e-01 -1.45838141e-01
-1.87872183e+00 -1.43190753e+00 4.10528630e-01 -1.66302174e-01
7.02064395e-01 -1.86079174e-01 -7.48439372e-01 5.92887342e-01
-5.32656088e-02 6.69947922e-01 1.79486692e+00 -1.68241322e-01
-1.89855278e-01 -1.98450848e-01 -1.40102029e+00 5.22666931e-01
1.53140843e-01 -7.26102352e-01 -9.77248624e-02 4.25763696e-01
-9.68306363e-02 -3.92335020e-02 -1.55315077e+00 2.72682816e-01
9.42504346e-01 -7.48277009e-01 8.41759562e-01 -4.06709164e-01
4.93020982e-01 -6.65270686e-01 -1.70372471e-01 -1.56955183e+00
-6.33973837e-01 -8.63369033e-02 2.40872011e-01 7.25547910e-01
1.36189818e-01 -2.97544152e-01 1.08887100e+00 4.37375486e-01
9.41944569e-02 -1.30333710e+00 -1.13035476e+00 -3.94661278e-01
-1.57227572e-02 1.77002013e-01 2.91284800e-01 9.83648002e-01
4.26226854e-01 1.73520952e-01 2.44874567e-01 2.57501692e-01
6.16752625e-01 2.73981750e-01 5.83029687e-01 -7.99262106e-01
-2.45457545e-01 -3.87506753e-01 -1.13493657e+00 -2.96587169e-01
3.08209985e-01 -1.20525193e+00 -4.39707674e-02 -1.45753384e+00
7.53625810e-01 -4.14201319e-01 -6.00270629e-01 4.67145771e-01
6.44377153e-03 9.85658705e-01 -1.62045568e-01 2.16810718e-01
-3.13325673e-01 4.22972560e-01 1.13426304e+00 -4.67226833e-01
3.21367025e-01 -6.05688334e-01 -6.20964289e-01 3.67999434e-01
6.54939890e-01 -3.28819245e-01 1.18065905e-02 -1.56468228e-01
-2.20362544e-02 7.25962147e-02 5.71178615e-01 -8.56333792e-01
2.90103912e-01 -9.75964442e-02 6.33995771e-01 -4.41434681e-02
2.52337992e-01 -9.09742355e-01 6.32971346e-01 6.58634007e-01
-2.33452410e-01 -6.34552762e-02 1.86713651e-01 5.55171192e-01
-4.25583690e-01 8.52961540e-02 8.88561308e-01 -1.67456210e-01
-6.36079013e-01 5.51798642e-01 -5.97832382e-01 -7.61223495e-01
1.37516272e+00 -8.69185850e-02 -3.99226159e-01 4.87373844e-02
-7.96925604e-01 1.60664201e-01 7.90121794e-01 -1.81436345e-01
7.77831018e-01 -1.44345737e+00 -6.86385572e-01 2.67935216e-01
4.62885022e-01 -1.81971729e-01 7.88933694e-01 1.26160944e+00
-9.24227357e-01 2.66183883e-01 -3.76690358e-01 -9.32758927e-01
-1.64477217e+00 3.26869965e-01 4.09496546e-01 -4.52357233e-01
-4.24330354e-01 8.59735131e-01 4.65568095e-01 -3.19521546e-01
-3.70626569e-01 7.98648298e-02 -3.10466555e-03 -1.10417247e-01
6.03086233e-01 3.16766232e-01 2.16197774e-01 -5.16749322e-01
-2.64838427e-01 6.25262499e-01 -3.24969478e-02 2.04506740e-01
1.56576073e+00 -2.30022654e-01 -8.13745260e-01 4.45964962e-01
1.73050225e+00 2.27841474e-02 -7.88950980e-01 1.92714080e-01
-2.90147811e-01 -6.15278006e-01 6.11820638e-01 -9.98324275e-01
-1.12791073e+00 8.71623278e-01 8.36091757e-01 -2.45798513e-01
1.20513272e+00 1.50189593e-01 6.36684239e-01 -3.75660807e-02
4.54244643e-01 -3.67311299e-01 -1.17291681e-01 -2.11558387e-01
6.91422343e-01 -1.19027245e+00 6.99775368e-02 -6.64631724e-01
-1.19321205e-01 1.25072443e+00 2.50495493e-01 1.83097944e-02
6.22744679e-01 4.68339920e-01 4.72250998e-01 -5.92269301e-01
-9.61234331e-01 1.99177578e-01 9.50970575e-02 3.42657447e-01
8.81165385e-01 3.86548579e-01 -8.39527026e-02 2.60718226e-01
1.57577738e-01 1.68547228e-01 3.54152679e-01 6.19334579e-01
-8.04146752e-02 -1.16559148e+00 -1.28054559e-01 7.37890303e-01
-8.48288953e-01 2.12810785e-02 -2.62350976e-01 5.97378016e-01
-1.42758012e-01 2.86389440e-01 1.03261948e-01 -2.51992047e-01
-8.91373158e-02 8.49860758e-02 7.32324183e-01 -6.92805946e-01
-3.29708397e-01 2.32140154e-01 -2.79086947e-01 -4.10760939e-01
-2.40869001e-01 -6.82263136e-01 -1.02844727e+00 -3.14127415e-01
-1.53011397e-01 6.99516386e-02 8.03215683e-01 4.69819963e-01
7.60989547e-01 3.75375152e-01 7.85141289e-01 -6.19758010e-01
-2.77955502e-01 -6.96324944e-01 -8.96852851e-01 3.39183003e-01
5.00343680e-01 -3.73884946e-01 -2.94823080e-01 3.90959978e-01] | [15.10315990447998, -2.9880263805389404] |
054f0f5c-409e-473c-bc0d-bf22bb3742bd | joint-graph-learning-and-model-fitting-in | 2305.02573 | null | https://arxiv.org/abs/2305.02573v1 | https://arxiv.org/pdf/2305.02573v1.pdf | Joint Graph Learning and Model Fitting in Laplacian Regularized Stratified Models | Laplacian regularized stratified models (LRSM) are models that utilize the explicit or implicit network structure of the sub-problems as defined by the categorical features called strata (e.g., age, region, time, forecast horizon, etc.), and draw upon data from neighboring strata to enhance the parameter learning of each sub-problem. They have been widely applied in machine learning and signal processing problems, including but not limited to time series forecasting, representation learning, graph clustering, max-margin classification, and general few-shot learning. Nevertheless, existing works on LRSM have either assumed a known graph or are restricted to specific applications. In this paper, we start by showing the importance and sensitivity of graph weights in LRSM, and provably show that the sensitivity can be arbitrarily large when the parameter scales and sample sizes are heavily imbalanced across nodes. We then propose a generic approach to jointly learn the graph while fitting the model parameters by solving a single optimization problem. We interpret the proposed formulation from both a graph connectivity viewpoint and an end-to-end Bayesian perspective, and propose an efficient algorithm to solve the problem. Convergence guarantees of the proposed optimization algorithm is also provided despite the lack of global strongly smoothness of the Laplacian regularization term typically required in the existing literature, which may be of independent interest. Finally, we illustrate the efficiency of our approach compared to existing methods by various real-world numerical examples. | ['Stephen Boyd', 'Akshay Agrawal', 'Junzi Zhang', 'Ziheng Cheng'] | 2023-05-04 | null | null | null | null | ['graph-clustering'] | ['graphs'] | [ 1.24707311e-01 3.63701612e-01 -4.06814575e-01 -2.54008621e-01
-6.52170420e-01 -3.99875820e-01 5.59283197e-01 3.78986180e-01
-2.32896823e-02 6.15474880e-01 1.16845869e-01 -2.91115552e-01
-8.28709126e-01 -7.33451366e-01 -5.09507298e-01 -9.17581439e-01
-5.31853199e-01 4.26198989e-01 9.53900665e-02 3.96843627e-02
1.21622033e-01 4.16668564e-01 -9.52566087e-01 -5.44825435e-01
1.05263734e+00 1.09395719e+00 -1.25630990e-01 3.66873056e-01
1.45092234e-01 6.24029934e-01 -2.36366481e-01 -3.03369492e-01
3.47283214e-01 -2.73688048e-01 -4.62644249e-01 7.79844582e-01
2.79110987e-02 2.58235902e-01 -4.52862620e-01 1.10211694e+00
3.96779835e-01 4.66451466e-01 8.31283450e-01 -1.34840846e+00
-3.95948917e-01 7.78746784e-01 -8.96448851e-01 1.37648344e-01
-6.75578194e-04 -1.71742082e-01 9.93420362e-01 -6.90122545e-01
4.75370705e-01 9.30054247e-01 7.50137568e-01 8.40532109e-02
-1.33487475e+00 -3.44840437e-01 6.05836213e-01 2.85337627e-01
-1.54414809e+00 -2.38918543e-01 1.14497817e+00 -6.90987408e-01
2.35342443e-01 2.97201902e-01 4.86014634e-01 6.87807202e-01
7.54124660e-04 3.05295497e-01 1.07582140e+00 -3.47204030e-01
6.00344658e-01 1.52241260e-01 4.72195297e-01 8.11338186e-01
3.79539490e-01 -2.27602974e-01 -2.60537356e-01 -5.73574245e-01
5.08326054e-01 7.01432005e-02 -3.80730331e-01 -7.58409381e-01
-9.08580422e-01 1.00334013e+00 5.40879309e-01 1.49720699e-01
-3.41963649e-01 -9.81463566e-02 2.77994126e-01 2.18244150e-01
1.00223923e+00 2.85684480e-03 -3.10357362e-01 4.68036562e-01
-1.03922880e+00 -8.70437920e-02 9.74874437e-01 9.56664801e-01
8.15995991e-01 1.85615480e-01 -3.30374464e-02 8.35015178e-01
4.78559971e-01 9.25415084e-02 1.46688789e-01 -6.55074120e-01
4.68475729e-01 4.94447768e-01 -2.31614476e-03 -1.28909707e+00
-5.56241870e-01 -6.63983345e-01 -1.16936874e+00 -2.34247327e-01
4.75844055e-01 -3.82688254e-01 -7.29575515e-01 1.88316464e+00
5.90873778e-01 8.19190979e-01 -1.92152530e-01 7.03433990e-01
4.51851755e-01 5.76261342e-01 -3.20373885e-02 -6.64219916e-01
1.13090611e+00 -6.81859612e-01 -6.59751594e-01 -2.37468511e-01
5.51299870e-01 -3.68639112e-01 7.25934207e-01 1.72217831e-01
-8.75065088e-01 -9.26447362e-02 -8.67572248e-01 4.40416604e-01
-3.12400192e-01 7.18760816e-03 5.82706988e-01 6.48425043e-01
-1.02509928e+00 6.02680683e-01 -7.95976758e-01 -4.48803991e-01
2.42629424e-01 2.38982916e-01 3.44620235e-02 -6.44501075e-02
-1.20952916e+00 5.86781800e-01 2.29457855e-01 2.99032837e-01
-6.29442453e-01 -6.16288722e-01 -7.35684097e-01 2.35805959e-01
7.57784426e-01 -7.65456498e-01 7.03148365e-01 -7.75485575e-01
-1.23982060e+00 5.43293238e-01 -1.72538497e-02 -5.42444825e-01
4.92133617e-01 2.68264353e-01 -3.83989483e-01 1.36003628e-01
-1.69942662e-01 -1.79809511e-01 1.07118785e+00 -1.02816927e+00
-2.98611611e-01 -5.17235279e-01 1.70837805e-01 1.46650061e-01
-3.69707704e-01 -2.28590533e-01 -4.83347207e-01 -8.14780951e-01
2.02541396e-01 -8.66898358e-01 -8.43112469e-01 -1.32153511e-01
-5.51158011e-01 -1.85850084e-01 6.11557007e-01 -7.07661569e-01
1.49324799e+00 -2.17921376e+00 4.52848524e-01 8.04053724e-01
3.04609209e-01 -2.40446135e-01 2.01496169e-01 5.42591214e-01
-7.93642998e-02 1.75143452e-03 -5.49196362e-01 -2.43581712e-01
3.99324000e-02 1.14309683e-01 -1.35689348e-01 1.09477961e+00
-1.94199055e-01 5.47443092e-01 -8.86595905e-01 -5.09046435e-01
1.34582579e-01 4.21992064e-01 -3.78513604e-01 -4.67739068e-03
-3.09723839e-02 3.82990599e-01 -6.30652189e-01 5.30255854e-01
5.31562150e-01 -6.70679092e-01 5.07417917e-01 -1.83140755e-01
1.55906111e-01 -1.81833535e-01 -1.53495383e+00 1.37061059e+00
-3.81596029e-01 3.48710746e-01 4.23846304e-01 -1.50683331e+00
6.82900906e-01 2.60057688e-01 6.73282146e-01 6.24101423e-02
3.36695276e-02 2.31001694e-02 -1.67249173e-01 -2.41271809e-01
2.04242244e-01 -8.23431611e-02 -8.96585137e-02 1.89965770e-01
-1.28861934e-01 9.52817500e-02 2.37086579e-01 3.43770862e-01
8.92617106e-01 -2.98729181e-01 6.23814821e-01 -4.86254603e-01
5.01074135e-01 -3.07308435e-01 6.44115865e-01 6.08951807e-01
-1.02932313e-02 3.67811233e-01 7.78641999e-01 6.88053444e-02
-8.24411690e-01 -9.49100077e-01 -2.81053394e-01 8.94858062e-01
8.18440467e-02 -3.39976132e-01 -5.45716107e-01 -6.37052417e-01
2.97245942e-02 5.18808842e-01 -6.05803907e-01 -7.65719712e-02
-2.31673986e-01 -1.23781490e+00 -2.90007256e-02 2.35523462e-01
3.37022990e-02 -3.43815029e-01 -4.98872101e-02 2.19917342e-01
1.56136170e-01 -1.10350895e+00 -6.61108673e-01 3.10166702e-02
-1.00752413e+00 -1.13174927e+00 -7.15754390e-01 -6.12827301e-01
9.74686444e-01 1.14797510e-01 9.06902373e-01 -1.12003982e-01
-2.03895152e-01 7.45610595e-01 -1.30053416e-01 -3.40652689e-02
5.81645221e-02 1.40421435e-01 6.44343346e-02 6.56441391e-01
-1.83166727e-01 -8.89631748e-01 -7.07802713e-01 2.77501464e-01
-8.17776442e-01 -1.09441116e-01 3.98679644e-01 7.14717031e-01
4.68124598e-01 1.96548805e-01 8.59858811e-01 -1.25676250e+00
6.35158002e-01 -9.88359272e-01 -9.05689597e-01 4.25291806e-01
-9.52136040e-01 -1.79245606e-01 3.95739675e-01 -4.93519753e-01
-9.22256708e-01 1.01043046e-01 3.79138231e-01 -4.27087873e-01
1.51925549e-01 1.09920382e+00 -1.32384405e-01 -3.02397549e-01
2.81111687e-01 2.07606442e-02 -6.83998549e-03 -4.72689867e-01
5.80790579e-01 4.62683409e-01 2.33199880e-01 -4.42325830e-01
8.46620798e-01 5.98257124e-01 4.04963553e-01 -9.48997974e-01
-9.01022971e-01 -5.87171435e-01 -5.05070865e-01 -4.20139998e-01
5.14975071e-01 -8.59683692e-01 -5.06745994e-01 2.43871942e-01
-6.92220509e-01 -2.31047764e-01 -2.54352748e-01 4.58162695e-01
-5.65639257e-01 6.68015242e-01 -4.95904654e-01 -1.11870182e+00
-2.84061562e-02 -6.81617498e-01 5.96853375e-01 5.64421341e-02
9.75338891e-02 -1.65900624e+00 -1.83183327e-02 7.77279139e-02
2.59981871e-01 5.02152085e-01 1.07159448e+00 -6.44782305e-01
-4.20961231e-01 -3.68677795e-01 -1.97310135e-01 8.00263733e-02
1.93185493e-01 -9.01534855e-02 -6.72059059e-01 -5.19325972e-01
1.12329550e-01 1.60521939e-01 8.78753901e-01 7.42213428e-01
1.33071852e+00 -6.01467133e-01 -4.95693922e-01 6.37792587e-01
1.55050874e+00 -1.78916767e-01 1.60618693e-01 -1.72400683e-01
8.29430759e-01 9.16550219e-01 4.38288182e-01 8.04708719e-01
4.64651048e-01 6.76729679e-01 4.15343463e-01 -1.57143727e-01
2.05067754e-01 -2.20420863e-02 2.46556982e-01 1.08848727e+00
-7.75087923e-02 -2.12847561e-01 -8.09144199e-01 5.62557101e-01
-2.25463438e+00 -7.33705640e-01 -1.10193707e-01 2.51764202e+00
5.48051178e-01 4.38230149e-02 1.90202460e-01 1.38098478e-01
1.01484728e+00 4.07183886e-01 -5.96936107e-01 -1.90427259e-03
-5.32865860e-02 -2.40070373e-01 8.13191533e-01 6.11303270e-01
-1.11157024e+00 4.73505139e-01 6.01485395e+00 1.01820791e+00
-8.67890894e-01 1.18337058e-01 8.21299374e-01 5.14896251e-02
-3.02345812e-01 2.96778917e-01 -4.81127053e-01 3.84233952e-01
9.26022708e-01 -6.07998610e-01 5.87933302e-01 7.47139931e-01
4.15471792e-01 -1.98919163e-03 -9.48393047e-01 8.54636788e-01
1.12913147e-01 -1.11624217e+00 -3.56015950e-01 2.40692198e-01
8.42254400e-01 -2.16773883e-01 1.77973192e-02 1.20004810e-01
1.79435655e-01 -7.76205659e-01 3.10789049e-01 6.87528491e-01
5.46939909e-01 -6.90555811e-01 3.52120668e-01 4.51741368e-01
-1.46918356e+00 -2.22408056e-01 -2.45494545e-01 -5.33881299e-02
3.83068174e-01 1.02157176e+00 -5.77327967e-01 1.00625551e+00
3.20039630e-01 8.63306522e-01 -5.48641980e-01 1.34693623e+00
-3.08797252e-03 9.46151853e-01 -4.69431102e-01 8.68549496e-02
2.19991282e-01 -6.90537930e-01 7.62409985e-01 9.22490180e-01
3.45616579e-01 3.13470401e-02 5.00007749e-01 5.31100690e-01
2.38213800e-02 2.73366094e-01 -5.60654342e-01 -2.26192232e-02
3.30745935e-01 1.41181993e+00 -9.62807178e-01 -3.00798595e-01
-7.15770245e-01 4.12147909e-01 2.95621753e-01 8.67913306e-01
-6.39451444e-01 -3.26511949e-01 3.15391243e-01 3.28601807e-01
3.86022747e-01 -2.95373738e-01 -1.87930137e-01 -1.23735881e+00
-8.14758018e-02 -5.08253574e-01 7.94239640e-01 -2.51342893e-01
-1.74236095e+00 3.25490981e-01 3.03634763e-01 -1.24709249e+00
-3.30790162e-01 -4.09070045e-01 -6.70138717e-01 6.45358980e-01
-1.37217999e+00 -1.03754187e+00 -5.66531643e-02 5.72227240e-01
3.08148086e-01 -9.22181904e-02 3.77015173e-01 3.72298986e-01
-9.17274714e-01 2.52562135e-01 3.14407974e-01 -7.89547414e-02
4.18017805e-01 -1.20050704e+00 -2.19357256e-02 9.37735379e-01
1.77683719e-02 4.81755972e-01 8.61355960e-01 -6.47858739e-01
-1.24825919e+00 -1.29031801e+00 7.08489239e-01 3.53931300e-02
1.09490514e+00 -4.58199084e-01 -8.80354106e-01 6.87589765e-01
5.89417405e-02 2.42117018e-01 5.85725367e-01 3.52789760e-01
-5.08476458e-02 -2.45729595e-01 -1.08750010e+00 5.54610968e-01
8.54009271e-01 -3.62165153e-01 -1.37803704e-01 7.18724906e-01
4.63881224e-01 -2.15691969e-01 -1.09906125e+00 3.12535644e-01
1.02964811e-01 -6.17561519e-01 9.62514400e-01 -7.45110810e-01
-1.42987579e-01 -2.53328770e-01 -3.25047486e-02 -1.34650815e+00
-3.77851516e-01 -7.97710598e-01 -4.50315297e-01 1.43890429e+00
3.72664541e-01 -7.76737690e-01 7.09262848e-01 6.03837788e-01
1.44578675e-02 -8.54787886e-01 -1.16345537e+00 -7.86001265e-01
-7.34760612e-02 -4.88038599e-01 3.49187255e-01 1.12342966e+00
1.76637590e-01 4.04141039e-01 -6.52163744e-01 5.28052449e-01
1.03942299e+00 1.95628509e-01 3.57303917e-01 -1.49254620e+00
-4.81585830e-01 -3.78386825e-01 -4.40007150e-01 -9.10740376e-01
3.42780352e-01 -9.07128096e-01 -1.03690371e-01 -1.51619291e+00
1.46375909e-01 -6.58058584e-01 -4.52973634e-01 8.68965015e-02
-2.33665302e-01 -1.17952466e-01 -3.24233435e-04 9.35073346e-02
-5.65318763e-01 5.58000982e-01 9.27784204e-01 -9.65993777e-02
-2.65429825e-01 4.98251528e-01 -6.23062670e-01 8.36138129e-01
6.71999335e-01 -3.52576107e-01 -7.28407025e-01 5.52898869e-02
2.63739914e-01 3.31242383e-01 4.47244018e-01 -6.11812770e-01
4.29430723e-01 -3.26640695e-01 -1.26856312e-01 -1.97007582e-01
2.53359765e-01 -1.01348710e+00 2.23368615e-01 2.36099824e-01
-3.31354022e-01 -1.32120997e-01 -2.70586818e-01 1.21345663e+00
2.89420877e-03 -2.64830053e-01 7.07406521e-01 1.01316303e-01
-4.63092923e-01 6.61541939e-01 -1.93725005e-01 2.21370667e-01
1.11736608e+00 8.20489228e-02 -1.12994593e-02 -7.33952403e-01
-9.49055076e-01 4.71703202e-01 2.12713644e-01 9.56137385e-03
3.84672105e-01 -1.24488246e+00 -7.23459125e-01 -1.87424660e-01
2.54888218e-02 -2.02138007e-01 4.65517282e-01 1.26224208e+00
-1.14925373e-02 6.41261116e-02 3.54351699e-01 -4.82544571e-01
-9.72792804e-01 8.57758164e-01 2.10115626e-01 -4.29296374e-01
-7.24856436e-01 4.44036692e-01 1.91454396e-01 -1.48001641e-01
5.17344296e-01 -1.87529832e-01 -3.08499426e-01 3.18709284e-01
4.72976975e-02 6.53687835e-01 -1.42079070e-01 -5.85919142e-01
-3.39935720e-01 7.62018561e-01 2.58159250e-01 -3.58199179e-02
1.32388091e+00 -5.76728046e-01 -8.83398280e-02 7.13633060e-01
9.05907035e-01 9.77471247e-02 -1.15984428e+00 -6.28592908e-01
2.91949183e-01 -1.78875878e-01 1.31257817e-01 -1.96762249e-01
-1.39549339e+00 5.70718288e-01 3.29029381e-01 7.36377001e-01
1.19528925e+00 1.45995647e-01 3.13054889e-01 1.29900515e-01
3.02186191e-01 -1.07045889e+00 -2.76665032e-01 1.59788847e-01
7.61079431e-01 -1.04344428e+00 3.16725254e-01 -7.54880607e-01
-4.31601286e-01 8.75028610e-01 1.22141831e-01 -2.47416973e-01
1.28127992e+00 -3.47610004e-02 -2.59231776e-01 -1.33364230e-01
-7.59109735e-01 -2.34320611e-01 6.11199677e-01 4.47955459e-01
2.60638475e-01 7.69812763e-02 -5.69645941e-01 6.10065699e-01
2.26163983e-01 -3.50362182e-01 5.21397114e-01 5.70612550e-01
-2.74201065e-01 -7.74306536e-01 -3.08753550e-01 7.21952617e-01
-4.54761207e-01 4.07822840e-02 -1.02300309e-01 6.18032634e-01
-2.03439727e-01 9.85142112e-01 -1.74787804e-01 -1.69165190e-02
2.14517564e-01 -2.39890758e-02 2.40452379e-01 -4.73592162e-01
-2.46570185e-01 3.73599201e-01 1.95621476e-01 -2.45016724e-01
-4.81054991e-01 -8.19510937e-01 -7.77219236e-01 -1.75573796e-01
-4.33300793e-01 3.06517571e-01 3.83852005e-01 9.29128289e-01
2.65511900e-01 5.12783051e-01 9.55944061e-01 -7.90669560e-01
-6.79050624e-01 -9.20728326e-01 -1.17561769e+00 1.47580534e-01
3.11107755e-01 -7.45966673e-01 -8.61701190e-01 -5.24849407e-02] | [7.111222743988037, 4.891982555389404] |
0453ccc1-0d80-4e1d-adf5-f1540f89654f | learning-word-embeddings-efficiently-with | null | null | http://papers.nips.cc/paper/5165-learning-word-embeddings-efficiently-with-noise-contrastive-estimation | http://papers.nips.cc/paper/5165-learning-word-embeddings-efficiently-with-noise-contrastive-estimation.pdf | Learning word embeddings efficiently with noise-contrastive estimation | Continuous-valued word embeddings learned by neural language models have recently been shown to capture semantic and syntactic information about words very well, setting performance records on several word similarity tasks. The best results are obtained by learning high-dimensional embeddings from very large quantities of data, which makes scalability of the training method a critical factor. We propose a simple and scalable new approach to learning word embeddings based on training log-bilinear models with noise-contrastive estimation. Our approach is simpler, faster, and produces better results than the current state-of-the art method of Mikolov et al. (2013a). We achieve results comparable to the best ones reported, which were obtained on a cluster, using four times less data and more than an order of magnitude less computing time. We also investigate several model types and find that the embeddings learned by the simpler models perform at least as well as those learned by the more complex ones. | ['Andriy Mnih', 'Koray Kavukcuoglu'] | 2013-12-01 | null | null | null | neurips-2013-12 | ['learning-word-embeddings'] | ['methodology'] | [-4.10449743e-01 -1.17885925e-01 -2.34752968e-01 -2.74219155e-01
-7.74539530e-01 -5.54121077e-01 8.82279992e-01 7.26465583e-01
-1.13081324e+00 5.08342385e-01 3.25563729e-01 -3.69124025e-01
-1.04167253e-01 -8.01270485e-01 -5.53121984e-01 -5.83011210e-01
-2.38332212e-01 7.79690683e-01 3.05359453e-01 -3.13063890e-01
3.34914178e-01 3.95835996e-01 -1.64781559e+00 -1.44849345e-02
5.47703028e-01 8.56650591e-01 2.43078873e-01 7.61329174e-01
-5.63398898e-01 1.68020934e-01 -1.85805157e-01 -4.21011806e-01
1.14617161e-01 7.37271011e-02 -8.24474990e-01 -4.47447866e-01
4.25144166e-01 -5.31393439e-02 -2.09021196e-01 8.31435919e-01
6.21645570e-01 2.74626106e-01 6.74174726e-01 -9.00660038e-01
-1.21764195e+00 4.40957516e-01 -2.03824297e-01 2.60415047e-01
1.48640141e-01 -2.23132357e-01 1.46364486e+00 -1.23022139e+00
5.36216438e-01 1.24884975e+00 8.22534263e-01 4.95283782e-01
-1.42730141e+00 -3.53883237e-01 -1.45990700e-02 5.04697740e-01
-1.55806828e+00 -2.61432022e-01 2.58670211e-01 -2.59454787e-01
1.45583713e+00 2.44509093e-02 4.28016782e-01 9.20208752e-01
-4.33138870e-02 4.25414920e-01 9.20277953e-01 -7.67971575e-01
3.57170790e-01 4.30330664e-01 3.72177929e-01 6.96542203e-01
6.07450664e-01 -8.96734446e-02 -1.40237138e-01 -5.10900855e-01
3.21537614e-01 1.51269734e-01 1.05545677e-01 -4.70555395e-01
-1.13984847e+00 1.30910230e+00 4.26151514e-01 8.08321774e-01
-3.81453007e-01 2.72641629e-01 5.62653482e-01 4.15036887e-01
8.93001378e-01 6.24358058e-01 -8.26621711e-01 -4.09602046e-01
-7.22408414e-01 2.21700937e-01 8.58486414e-01 5.94844043e-01
7.95520782e-01 -1.25415027e-01 1.71094522e-01 9.14824486e-01
2.69860327e-01 2.76540786e-01 9.47791278e-01 -5.58268011e-01
1.21369399e-01 1.95025519e-01 1.71887726e-01 -8.71555209e-01
-6.11780047e-01 -1.07390225e-01 -5.22156060e-01 -2.63872705e-02
6.09300554e-01 9.13925469e-02 -8.23711693e-01 1.68917394e+00
4.44636941e-02 1.83816999e-01 6.99203238e-02 5.57594776e-01
5.89979470e-01 8.34049225e-01 1.38653815e-01 5.95871918e-02
1.61959958e+00 -9.02611375e-01 -5.67307293e-01 -1.84160978e-01
1.14098382e+00 -8.71086836e-01 1.25327384e+00 3.12375098e-01
-8.72894764e-01 -5.64212382e-01 -9.95188832e-01 -3.08253706e-01
-8.89428914e-01 -1.42506540e-01 7.99110293e-01 8.58604550e-01
-1.24874818e+00 7.52143919e-01 -7.80570269e-01 -6.58302069e-01
3.00473064e-01 2.35652611e-01 -4.94173706e-01 -2.04045042e-01
-1.28873801e+00 1.17655492e+00 4.45294023e-01 -3.78559202e-01
-3.00689667e-01 -7.79523551e-01 -9.65605378e-01 1.64995074e-01
1.34168491e-01 -2.10189879e-01 1.14617777e+00 -4.73861277e-01
-1.26821673e+00 8.42162013e-01 -2.21915603e-01 -5.44646740e-01
-4.82596923e-03 -2.48795062e-01 -2.80017912e-01 -8.80175978e-02
-1.72010824e-01 5.46121299e-01 3.70814890e-01 -8.85979652e-01
-1.71766862e-01 -3.39856058e-01 5.02453037e-02 -2.64500767e-01
-1.11710751e+00 1.25963643e-01 -2.42687821e-01 -5.54319263e-01
-3.47479224e-01 -6.72314048e-01 -3.53038937e-01 3.14858168e-01
3.18091989e-01 -7.88878500e-01 3.44895869e-01 -6.05447352e-01
1.31772459e+00 -2.03705001e+00 2.21064035e-02 -4.37590815e-02
2.17234254e-01 6.81916952e-01 -5.68846703e-01 8.00390780e-01
-2.51219988e-01 3.63709062e-01 -1.06784903e-01 -6.39079511e-01
4.47820991e-01 5.28378308e-01 -1.38219416e-01 4.78731573e-01
3.69822532e-01 9.99786854e-01 -9.67665255e-01 -2.34916151e-01
1.62802979e-01 6.33007646e-01 -6.41241789e-01 1.77746370e-01
-3.72798815e-02 -3.39751214e-01 -1.19480990e-01 1.80024356e-01
5.19759834e-01 -2.41301656e-01 2.16376647e-01 1.99904051e-02
-6.16087280e-02 5.45297384e-01 -1.20649397e+00 1.74002647e+00
-9.00164545e-01 7.66446054e-01 -2.45980680e-01 -1.33353627e+00
9.87635314e-01 4.17241782e-01 2.70804018e-01 -7.47909784e-01
1.39669046e-01 4.27793890e-01 5.28556891e-02 -6.01964295e-01
5.90384901e-01 -4.35121179e-01 -3.69854160e-02 7.31806397e-01
4.48490113e-01 1.54157039e-02 3.34013790e-01 6.27196729e-02
1.02027857e+00 -3.42412531e-01 5.95672071e-01 -5.56193411e-01
3.23538125e-01 -3.08268577e-01 5.61371408e-02 6.56339347e-01
-1.65332064e-01 4.93914604e-01 4.73530293e-01 -6.12193167e-01
-1.44037211e+00 -9.24611807e-01 -3.75466824e-01 1.35699439e+00
-2.14081138e-01 -6.68911934e-01 -5.99241555e-01 -3.24779779e-01
1.38501808e-01 7.19906867e-01 -8.56472254e-01 -1.73205286e-01
-4.47110355e-01 -7.87894011e-01 5.09443402e-01 8.89933527e-01
-2.82301635e-01 -7.76112616e-01 -1.99905321e-01 2.09828749e-01
3.67279619e-01 -1.06148875e+00 -1.96096659e-01 3.83923709e-01
-8.21895421e-01 -7.01675355e-01 -7.82439053e-01 -9.08783019e-01
4.41433102e-01 1.44734204e-01 1.23181009e+00 1.69037208e-01
-5.46485782e-01 4.46604937e-03 -5.47679543e-01 -3.68604988e-01
-2.79848576e-01 2.42180809e-01 5.01399219e-01 -2.20891446e-01
8.08393002e-01 -4.67467457e-01 -2.58767664e-01 -1.58804387e-01
-1.17459559e+00 -6.66433811e-01 5.84562719e-01 1.05609000e+00
1.77078709e-01 -2.62372792e-01 6.47615314e-01 -7.35596001e-01
9.11354601e-01 -6.60456777e-01 -5.68049133e-01 -1.39907589e-02
-1.09111166e+00 6.46314085e-01 1.01722288e+00 -5.90740502e-01
-3.85855377e-01 -2.79976755e-01 -4.97138530e-01 -2.06468597e-01
-1.87289074e-01 2.42617384e-01 2.60330528e-01 5.85127957e-02
6.73621655e-01 -1.20019717e-02 -3.25604305e-02 -8.31152916e-01
8.61884594e-01 8.21679175e-01 3.28571498e-02 -5.22455573e-01
8.95500660e-01 3.19286495e-01 -3.36331874e-01 -9.16013002e-01
-5.01528680e-01 -7.72966325e-01 -7.61364758e-01 4.93613899e-01
8.96787763e-01 -5.84665477e-01 -3.05123270e-01 6.82214200e-02
-1.43711996e+00 7.52808303e-02 -3.81282896e-01 7.36927271e-01
-2.19062403e-01 4.45129961e-01 -5.26219845e-01 -7.20422208e-01
-1.42223373e-01 -7.50224233e-01 9.88318145e-01 -1.07132182e-01
-3.63557935e-01 -1.52095926e+00 5.97457588e-01 -1.20037153e-01
7.90480316e-01 -2.62294263e-01 1.19514894e+00 -1.24082100e+00
8.08738172e-02 -4.80605960e-01 -4.41034108e-01 5.71330786e-01
5.67376688e-02 -2.54673809e-01 -1.01962554e+00 -3.98774594e-01
-2.43634671e-01 -4.10448968e-01 1.03676450e+00 5.14566414e-02
1.03549325e+00 -9.05232579e-02 -1.13669075e-01 3.93914789e-01
1.70107639e+00 -1.13197297e-01 4.61554855e-01 3.27637464e-01
5.14387667e-01 5.19505203e-01 3.10697764e-01 2.99244732e-01
3.35415781e-01 6.83848321e-01 2.32886940e-01 -1.59399733e-01
8.01749751e-02 -1.12030417e-01 2.83554375e-01 1.21508205e+00
-1.28767015e-02 -2.28192747e-01 -8.92005205e-01 1.03518677e+00
-1.79389238e+00 -8.21484387e-01 -1.09497927e-01 2.17432761e+00
7.00155973e-01 1.33631825e-01 1.39030769e-01 2.31030434e-01
3.08492452e-01 3.87262195e-01 -8.65901075e-03 -9.95212793e-01
1.22796714e-01 9.78904128e-01 5.74808359e-01 6.90740108e-01
-8.80534112e-01 9.18726802e-01 7.08453465e+00 9.01976466e-01
-8.83543372e-01 5.11974394e-01 2.08753616e-01 -2.61130422e-01
-4.69793707e-01 -1.30150780e-01 -4.95185226e-01 3.22438419e-01
1.48110950e+00 -3.00931722e-01 3.94425303e-01 8.63540411e-01
-2.37264782e-02 6.16248958e-02 -1.15564442e+00 1.08540678e+00
3.66610110e-01 -1.36289191e+00 -1.06949925e-01 3.75369824e-02
6.72702730e-01 2.53030866e-01 -1.66495621e-01 3.63756597e-01
2.38799557e-01 -1.21043062e+00 2.85632282e-01 3.45557094e-01
4.66976702e-01 -7.65648723e-01 9.85652149e-01 2.19537377e-01
-1.16355717e+00 -1.48207515e-01 -8.59069824e-01 -4.04932529e-01
8.11767802e-02 7.31585443e-01 -6.37812138e-01 3.48497927e-01
6.63245201e-01 5.95191598e-01 -6.72543824e-01 7.88172126e-01
-2.90636942e-02 6.97714269e-01 -3.75267684e-01 -7.33819962e-01
5.35986304e-01 -8.62037092e-02 9.08843875e-02 1.36306429e+00
2.66441762e-01 -3.09038043e-01 -3.86171937e-02 7.21273005e-01
-1.47021845e-01 4.57331330e-01 -8.71986985e-01 -3.80545110e-01
3.58236343e-01 1.33258462e+00 -4.13169652e-01 -5.06133974e-01
-6.89271331e-01 8.94390583e-01 7.59104371e-01 -1.56186586e-02
-6.22009456e-01 -8.03350925e-01 9.61874068e-01 -3.79035296e-03
6.15141749e-01 -6.25994861e-01 -1.67126015e-01 -1.03707695e+00
1.67775199e-01 -3.15476775e-01 2.17644557e-01 -4.25767541e-01
-1.74652004e+00 7.06335247e-01 2.32867170e-02 -8.78180385e-01
-4.56522495e-01 -1.47750759e+00 -5.08220255e-01 8.96702409e-01
-1.67745292e+00 -5.77645361e-01 1.75617561e-01 7.44249299e-02
3.87037665e-01 -1.48136377e-01 1.37055445e+00 4.67896819e-01
-1.13839336e-01 5.58177531e-01 6.61798835e-01 4.92535047e-02
6.19906843e-01 -1.28420281e+00 6.00482464e-01 4.92263436e-01
6.59035742e-01 6.46372676e-01 6.69627547e-01 3.04783955e-02
-1.45613825e+00 -6.82505369e-01 1.61658466e+00 -5.81860840e-01
1.18454826e+00 -8.26439738e-01 -1.16172206e+00 3.24477941e-01
4.53436375e-01 2.65732259e-01 8.86691630e-01 4.51045841e-01
-7.66232729e-01 -1.58700906e-02 -8.46710145e-01 2.80813217e-01
8.75865579e-01 -7.77900159e-01 -1.00421023e+00 4.12446856e-01
9.15676117e-01 3.02701354e-01 -8.51209700e-01 -4.16129529e-02
6.40499175e-01 -7.80154169e-01 1.05416846e+00 -1.07082784e+00
3.43927473e-01 2.33498871e-01 -3.29547912e-01 -1.48898327e+00
-4.23073262e-01 -1.54618546e-01 2.03154255e-02 1.07780707e+00
5.18623888e-01 -1.01805890e+00 3.93380076e-01 3.11261475e-01
2.30623707e-01 -9.32222188e-01 -9.97812510e-01 -1.18058085e+00
7.15769172e-01 -5.77173233e-01 5.51836789e-01 1.00869930e+00
1.68381006e-01 4.13661271e-01 -5.78075796e-02 -2.69169450e-01
2.58409500e-01 -1.71175092e-01 3.91753614e-01 -1.50348461e+00
-2.41938263e-01 -5.24214089e-01 -7.36547768e-01 -9.09998119e-01
5.74502230e-01 -1.18941498e+00 -1.30688563e-01 -1.57247794e+00
1.09355293e-01 -3.56972545e-01 -6.74306273e-01 4.82784510e-01
-1.37317821e-01 5.37775755e-01 3.53596024e-02 -2.61593491e-01
-3.09518695e-01 4.84585822e-01 5.63247979e-01 3.43794078e-02
2.68842578e-01 -5.42488933e-01 -8.32392633e-01 6.82096899e-01
7.11983562e-01 -6.52679384e-01 -2.04318166e-01 -6.96119905e-01
3.46140116e-01 -6.11052155e-01 2.37829059e-01 -6.61780596e-01
1.40047163e-01 1.32437035e-01 1.09185889e-01 -1.68355823e-01
4.21835840e-01 -6.82477117e-01 -5.56023896e-01 5.47470808e-01
-4.25856143e-01 5.52792132e-01 2.50349015e-01 4.37311590e-01
-1.56682044e-01 -6.25800431e-01 6.57648802e-01 7.88863562e-03
-7.52334356e-01 2.22615823e-01 -4.12539572e-01 2.00376377e-01
7.23424077e-01 3.75391208e-02 -7.15652034e-02 -3.24293733e-01
-5.19292057e-01 -2.11814925e-01 2.79123396e-01 8.15958917e-01
3.79786283e-01 -1.53770173e+00 -7.23667264e-01 2.42697194e-01
5.13560355e-01 -4.71355379e-01 -2.83639759e-01 5.70949793e-01
-4.45086598e-01 6.98417962e-01 8.24780297e-03 -2.62957931e-01
-1.05311525e+00 7.82937288e-01 -6.74719363e-02 -4.16971534e-01
-4.00196999e-01 7.80352235e-01 -2.92828768e-01 -7.77195811e-01
5.33492491e-02 -5.13334155e-01 -6.85453862e-02 3.18820238e-01
7.05382347e-01 3.76973689e-01 2.47337177e-01 -5.76687753e-01
-5.41260302e-01 7.49920130e-01 -1.10036097e-01 -1.17980175e-01
1.70709181e+00 2.37087190e-01 -1.20999701e-01 6.57619894e-01
1.87329292e+00 -5.80150485e-02 -5.36833823e-01 -5.05837619e-01
1.70672908e-01 -5.21169722e-01 1.57691255e-01 -2.94540882e-01
-6.29198670e-01 1.32566977e+00 7.82223642e-01 5.50549567e-01
5.57968080e-01 3.24917138e-01 8.90076816e-01 6.05995655e-01
2.81105846e-01 -1.21312511e+00 1.32032990e-01 6.25809848e-01
5.02023160e-01 -1.21055186e+00 -3.31726670e-02 1.16080053e-01
-1.81071535e-01 1.23004127e+00 1.45708069e-01 -5.48496068e-01
7.90654242e-01 2.59507358e-01 3.04715000e-02 4.56913337e-02
-9.83469188e-01 -3.95639390e-01 3.44654888e-01 6.29536748e-01
5.11933982e-01 5.34578320e-03 -6.28605902e-01 5.53253114e-01
-3.32078673e-02 -2.50612795e-01 2.10815758e-01 6.49733961e-01
-5.44071317e-01 -1.55242598e+00 -1.20715216e-01 4.31304842e-01
-2.97989041e-01 -4.96960878e-01 -1.77026495e-01 8.35096240e-01
-1.28103942e-01 8.21144938e-01 4.11332875e-01 -3.65385622e-01
2.61194348e-01 4.85212445e-01 3.64151925e-01 -7.50713825e-01
-3.88358206e-01 -5.12033284e-01 9.10266191e-02 -5.28134823e-01
-2.61493057e-01 -3.90350431e-01 -1.26535642e+00 -4.04037714e-01
-3.67181271e-01 2.46568158e-01 9.64371562e-01 1.09221101e+00
4.16295677e-01 2.43569180e-01 6.39210641e-01 -1.03063774e+00
-1.06258917e+00 -1.12033784e+00 -5.26676238e-01 5.32851815e-01
2.12338373e-01 -6.71700776e-01 -5.64488351e-01 -3.59119505e-01] | [10.491302490234375, 8.702796936035156] |
35daddd5-da66-416b-9728-b26d24b0390a | architectural-vision-for-quantum-computing-in | 2305.05238 | null | https://arxiv.org/abs/2305.05238v1 | https://arxiv.org/pdf/2305.05238v1.pdf | Architectural Vision for Quantum Computing in the Edge-Cloud Continuum | Quantum processing units (QPUs) are currently exclusively available from cloud vendors. However, with recent advancements, hosting QPUs is soon possible everywhere. Existing work has yet to draw from research in edge computing to explore systems exploiting mobile QPUs, or how hybrid applications can benefit from distributed heterogeneous resources. Hence, this work presents an architecture for Quantum Computing in the edge-cloud continuum. We discuss the necessity, challenges, and solution approaches for extending existing work on classical edge computing to integrate QPUs. We describe how warm-starting allows defining workflows that exploit the hierarchical resources spread across the continuum. Then, we introduce a distributed inference engine with hybrid classical-quantum neural networks (QNNs) to aid system designers in accommodating applications with complex requirements that incur the highest degree of heterogeneity. We propose solutions focusing on classical layer partitioning and quantum circuit cutting to demonstrate the potential of utilizing classical and quantum computation across the continuum. To evaluate the importance and feasibility of our vision, we provide a proof of concept that exemplifies how extending a classical partition method to integrate quantum circuits can improve the solution quality. Specifically, we implement a split neural network with optional hybrid QNN predictors. Our results show that extending classical methods with QNNs is viable and promising for future work. | ['Felix Truger', 'Philipp Raith', 'Frank Leymann', 'Schahram Dustdar', 'Marvin Bechtold', 'Johanna Barzen', 'Alireza Furutanpey'] | 2023-05-09 | null | null | null | null | ['edge-computing'] | ['time-series'] | [-4.35360484e-02 -9.31433886e-02 -1.46079838e-01 -9.81535837e-02
-7.23709166e-01 -4.40085709e-01 5.76489195e-02 -1.66220576e-01
-1.96747348e-01 6.59901023e-01 -2.59238660e-01 -6.18110478e-01
-2.26686239e-01 -1.15518689e+00 -6.18540943e-01 -5.33633709e-01
-5.10022417e-02 4.86470342e-01 1.21010557e-01 -4.88205910e-01
2.84642905e-01 4.35410976e-01 -1.77356315e+00 6.88123882e-01
7.73713470e-01 1.23804533e+00 -4.98356037e-02 8.53370428e-01
4.82938848e-02 8.51485193e-01 -3.00260931e-01 -2.39086166e-01
6.24896646e-01 -4.04477060e-01 -8.34084094e-01 -7.45553195e-01
6.28213286e-02 -2.64573544e-01 -3.25772941e-01 9.28736687e-01
8.15890312e-01 4.00358811e-02 2.01015070e-01 -1.21537364e+00
-6.14620388e-01 5.97479343e-01 1.05452336e-01 2.34648407e-01
5.17632365e-02 3.00119609e-01 1.19576180e+00 -3.58447284e-01
8.62792969e-01 3.87922943e-01 9.30980742e-01 4.90646482e-01
-1.34256625e+00 -5.93542576e-01 -5.41215599e-01 7.82995224e-01
-1.79780591e+00 -6.18450880e-01 3.28068584e-01 2.17925981e-02
1.89428985e+00 3.37835014e-01 8.88981938e-01 4.43236589e-01
4.57151741e-01 5.62873483e-01 1.15124786e+00 -5.80325007e-01
6.92958236e-01 3.09772998e-01 1.95305631e-01 4.87522066e-01
1.44204944e-01 3.23520213e-01 -8.48478913e-01 -6.37016669e-02
3.31291080e-01 -1.91636220e-01 5.65569252e-02 -1.39890820e-01
-6.83828890e-01 4.56180722e-01 6.37878656e-01 2.65495449e-01
-4.91895109e-01 5.30748308e-01 5.10049522e-01 3.38943213e-01
3.30624759e-01 6.28537714e-01 -4.69846427e-01 -4.86856580e-01
-1.31238925e+00 1.13275759e-01 1.12170446e+00 1.20096970e+00
1.19310308e+00 -3.12542111e-01 -3.28322113e-01 4.41677980e-02
-1.08430274e-01 2.82558292e-01 -1.37926117e-01 -1.45305467e+00
-2.10659996e-01 3.02552313e-01 -9.70874280e-02 -2.26423636e-01
-5.36246896e-01 -4.02562946e-01 -5.04823327e-01 2.26783678e-01
-2.57815063e-01 -2.71563530e-01 -6.41872525e-01 1.14722633e+00
2.32008800e-01 5.36746800e-01 1.39809206e-01 8.53673458e-01
6.91568792e-01 4.80259746e-01 -3.95193160e-01 1.65730074e-01
1.38290417e+00 -1.29004633e+00 -4.86594349e-01 2.87866652e-01
1.05726194e+00 -5.96450269e-01 6.13525093e-01 4.04337645e-01
-1.16259527e+00 -3.18804741e-01 -1.35194468e+00 -3.18310767e-01
-5.22783458e-01 -3.29988599e-01 9.78436828e-01 1.25481379e+00
-1.70323598e+00 9.85715032e-01 -1.03020012e+00 -3.20908934e-01
2.06138149e-01 9.72717106e-01 5.44296503e-02 -5.03505990e-02
-1.37092698e+00 8.04517031e-01 6.37522519e-01 -6.46091402e-02
-3.27771842e-01 -6.45840287e-01 -3.05635691e-01 5.39818645e-01
2.26216510e-01 -1.30371690e+00 1.24688077e+00 -6.05594277e-01
-1.93607461e+00 3.47988814e-01 -1.09866098e-01 -6.79554939e-01
-9.75216553e-02 4.99425083e-01 -3.42967719e-01 3.22684675e-01
-1.01072520e-01 6.67312801e-01 2.31576130e-01 -4.43396121e-01
-7.08314478e-01 -9.28018466e-02 4.25072879e-01 2.33022779e-01
-3.75676602e-01 -1.64981976e-01 -2.68948466e-01 6.18426859e-01
5.49616516e-02 -1.16212392e+00 -2.82029539e-01 -6.06268644e-01
1.38394199e-02 -1.04431827e-02 4.48709518e-01 -2.58469768e-03
1.27965260e+00 -2.09808850e+00 -1.62519231e-01 3.29129040e-01
4.25847620e-01 -6.20972402e-02 2.17124462e-01 7.48646677e-01
4.20894295e-01 1.44335940e-01 8.14866200e-02 -2.99968153e-01
4.36822921e-01 9.82383564e-02 6.40300214e-02 4.81436178e-02
-3.38853337e-02 1.09347486e+00 -7.50752866e-01 -2.42309168e-01
-1.31536067e-01 4.53014255e-01 -1.15666068e+00 -3.49856198e-01
-2.13776529e-01 2.77844846e-01 -1.18480407e-01 8.71123075e-01
8.90724838e-01 -6.54564500e-01 2.91577667e-01 -9.73332301e-02
-4.41510022e-01 5.94181657e-01 -1.20608938e+00 1.73951936e+00
-5.02196014e-01 7.32174575e-01 4.99659032e-01 -7.03669429e-01
3.07641566e-01 3.34779441e-01 4.23551440e-01 -8.47185254e-01
-4.41226596e-03 6.67752028e-01 2.51444310e-01 -3.00071746e-01
1.13490176e+00 -4.38108146e-01 5.58296219e-02 4.70224768e-01
2.74757653e-01 -2.71745920e-01 2.88100779e-01 2.58273363e-01
1.44379222e+00 2.49181360e-01 -2.15746611e-01 -3.81890297e-01
1.36708677e-01 1.20465562e-01 4.90991741e-01 1.06130695e+00
-7.16828883e-01 3.28891158e-01 5.94640017e-01 -3.21282506e-01
-1.21234369e+00 -9.96725857e-01 -5.31030476e-01 1.17536974e+00
3.21836978e-01 -1.08692622e+00 -8.99084032e-01 -8.48524719e-02
-2.79614329e-01 5.17316461e-01 -1.29981399e-01 5.54506369e-02
4.46607582e-02 -1.13315582e+00 6.49198294e-01 4.74693477e-01
5.84107459e-01 -8.72603774e-01 -7.11504519e-01 1.00532621e-01
1.07272476e-01 -1.28236055e+00 2.54685760e-01 5.75075388e-01
-8.25539887e-01 -4.36063558e-01 5.06950833e-04 -3.56649816e-01
1.00587748e-01 3.56355608e-01 1.19685769e+00 -3.02188657e-02
-2.75686830e-01 2.70120472e-01 -2.39461154e-01 2.11339500e-02
-2.39326566e-01 4.88929123e-01 2.51982510e-01 -4.85812038e-01
6.63196981e-01 -7.53297806e-01 -9.22149360e-01 -2.31827825e-01
-7.98211455e-01 2.94816047e-01 5.13225257e-01 7.98721969e-01
3.92137438e-01 1.36561215e-01 2.02513248e-01 -9.83069539e-01
5.19200444e-01 -5.22039652e-01 -5.57646215e-01 8.75485316e-02
-1.04042590e+00 8.64110738e-02 8.00105691e-01 4.14482385e-01
-8.42162609e-01 -2.92195767e-01 -3.21174562e-01 -1.96117088e-01
-4.18730117e-02 6.22687042e-01 4.35879499e-01 -4.74503458e-01
8.94526303e-01 -1.16441868e-01 -4.51172113e-01 3.39115620e-01
5.74638188e-01 1.01879454e+00 -2.17908651e-01 -8.12757432e-01
-1.76257025e-02 8.20415795e-01 3.03454906e-01 -7.89102018e-01
-2.88237482e-01 -5.57661951e-01 -4.22530144e-01 -1.92000642e-02
9.72141981e-01 -1.02983117e+00 -1.25674903e+00 -1.64139211e-01
-9.20459867e-01 -4.73391771e-01 -2.68270701e-01 5.25589645e-01
-6.13590956e-01 2.07187161e-01 -1.16265035e+00 -9.66048419e-01
-6.31188452e-01 -1.54139853e+00 1.09672129e+00 6.15399778e-01
4.71220054e-02 -7.19325840e-01 2.88367458e-02 6.80461347e-01
9.97900546e-01 -4.36494678e-01 4.99335110e-01 -3.75169426e-01
-1.49421263e+00 -4.81587350e-02 -4.84902382e-01 1.45281628e-01
-6.33628547e-01 8.32907706e-02 -1.27291763e+00 -3.37346524e-01
-1.54186115e-01 -2.55098224e-01 7.33267128e-01 2.22038969e-01
7.69980967e-01 4.12503511e-01 -3.34552109e-01 1.07214570e+00
1.55864918e+00 -4.12646234e-02 6.32802427e-01 2.92865902e-01
5.00902295e-01 1.21445000e-01 -1.09947927e-01 2.57257134e-01
5.63255489e-01 4.27867413e-01 2.30298564e-01 2.31313214e-01
2.18551248e-01 2.78934687e-01 4.75076765e-01 1.36130261e+00
-4.13044006e-01 2.89081842e-01 -9.85687912e-01 1.44726396e-01
-1.85598505e+00 -1.07190502e+00 -2.29640067e-01 2.07871580e+00
2.67603695e-01 3.97863775e-01 -2.16221958e-01 -3.67679775e-01
4.71933156e-01 -3.39501411e-01 -4.70739216e-01 -1.09917319e+00
2.48391345e-01 9.20589268e-01 8.88995588e-01 2.75656521e-01
-7.24336982e-01 1.04015660e+00 6.52576780e+00 1.01266754e+00
-1.15393281e+00 6.67359710e-01 3.92600834e-01 -3.61239165e-01
-4.00040805e-01 7.66599298e-01 -8.20399284e-01 3.53611261e-01
1.82698476e+00 -9.73172784e-02 1.09478617e+00 9.23414528e-01
-1.65520892e-01 -1.98775113e-01 -9.61325109e-01 1.00698888e+00
-3.22855443e-01 -1.83578002e+00 -7.39855349e-01 1.81039289e-01
1.10273027e+00 9.74597156e-01 -7.61785656e-02 7.59401083e-01
1.16217732e-01 -8.74996662e-01 4.92433727e-01 4.97222245e-01
7.90326178e-01 -7.03796744e-01 9.61348712e-01 4.35096234e-01
-1.13760722e+00 -1.12623505e-01 -3.79318357e-01 -8.74870241e-01
8.75163674e-02 3.49276125e-01 -8.35774243e-01 8.87612104e-01
8.87668490e-01 3.13675344e-01 -2.36344755e-01 9.51993942e-01
8.81790221e-02 3.57545316e-01 -6.84776723e-01 -4.42838609e-01
2.48950407e-01 -6.81202710e-01 4.38953750e-02 1.12295997e+00
5.04656076e-01 -1.07458429e-02 -8.76473486e-02 1.31982303e+00
-6.23654164e-02 9.99228284e-02 -6.81999147e-01 -1.33175716e-01
5.52377760e-01 1.61929238e+00 -1.17676842e+00 -4.43015546e-01
-6.07337058e-01 1.18192732e+00 3.13257098e-01 3.06688726e-01
-6.77720010e-01 -4.90978211e-01 5.22588491e-01 -1.77937463e-01
3.61937165e-01 -3.01923752e-01 -7.03428745e-01 -1.45475721e+00
-1.45591885e-01 -5.12513578e-01 -2.91320607e-02 -7.75924981e-01
-1.10755968e+00 5.70288897e-01 -3.86925906e-01 -8.38293195e-01
5.97822033e-02 -7.64488399e-01 -3.75412583e-01 1.00080478e+00
-1.42052686e+00 -1.10259128e+00 -8.88179243e-02 -3.08448933e-02
-2.61285752e-01 1.63177431e-01 1.15812266e+00 6.42270684e-01
-5.80835044e-01 4.56599325e-01 6.20782733e-01 -3.82106245e-01
4.52364355e-01 -1.16976380e+00 4.92853284e-01 6.39665306e-01
1.46357223e-01 9.89849806e-01 3.72933865e-01 -3.77685040e-01
-1.88856244e+00 -4.14855301e-01 5.00405967e-01 -5.41450620e-01
6.83754563e-01 -6.77766562e-01 -2.98744112e-01 5.63359141e-01
4.80782628e-01 1.74901202e-01 1.14996302e+00 6.46506608e-01
2.81806923e-02 -1.47800922e-01 -8.37179482e-01 5.90395510e-01
9.39924955e-01 -1.27636313e+00 2.83043832e-01 3.97027463e-01
4.82285768e-01 -5.58108687e-01 -9.66766894e-01 2.90440410e-01
6.19382560e-01 -1.54704499e+00 5.72153866e-01 -3.11338603e-01
2.36872479e-01 -3.45576614e-01 -2.81676382e-01 -8.83375764e-01
-3.31429034e-01 -9.97700870e-01 -2.39012748e-01 5.58316588e-01
3.86772066e-01 -7.19069421e-01 1.29333472e+00 9.34633732e-01
-4.90907073e-01 -7.49771416e-01 -1.16127288e+00 -5.89675009e-01
1.07543178e-01 -8.33494246e-01 5.46065152e-01 7.95260310e-01
5.32775760e-01 6.42576218e-01 7.65301362e-02 1.21978097e-01
2.43264794e-01 3.06518793e-01 5.36113322e-01 -6.53823793e-01
-8.01885366e-01 -6.01438940e-01 -7.07646608e-01 -8.51724684e-01
-2.03149632e-01 -1.36832786e+00 -1.77925095e-01 -1.43019366e+00
3.23694974e-01 -5.23367584e-01 -5.03658414e-01 -6.00133352e-02
8.74179415e-03 4.86888438e-01 4.74694490e-01 2.11287439e-01
-1.09427440e+00 4.33411300e-01 8.44930947e-01 1.63301021e-01
-2.73810953e-01 -2.43246153e-01 -5.65121591e-01 2.76117533e-01
6.97997451e-01 -3.69080454e-01 -2.55847394e-01 -2.50251651e-01
1.04256010e+00 3.49979736e-02 6.11682013e-02 -1.51916814e+00
9.78292048e-01 3.93567026e-01 -9.23192501e-02 -2.45755911e-01
5.07063389e-01 -5.39134920e-01 1.85265064e-01 2.99984574e-01
3.82725447e-01 -5.12745418e-02 1.07606754e-01 2.61884153e-01
-2.56791022e-02 -4.00653660e-01 3.78946871e-01 -2.67289281e-01
-7.95163631e-01 1.73970178e-01 -4.13493335e-01 -2.84427136e-01
1.14434934e+00 -2.65140593e-01 -4.60276693e-01 -1.96838647e-01
-7.70382404e-01 2.99798816e-01 9.25837100e-01 -3.75574738e-01
1.90531507e-01 -8.35767567e-01 -2.03936651e-01 2.85828650e-01
-3.00544631e-02 -2.79719710e-01 6.76210523e-01 1.16839814e+00
-1.20709372e+00 6.76918685e-01 -3.19069892e-01 -6.48868561e-01
-6.34013236e-01 6.25137806e-01 5.84047079e-01 -2.07714826e-01
-3.08825225e-01 9.36454237e-01 -3.75028640e-01 -6.69622838e-01
-4.15345371e-01 -2.47307986e-01 5.53386390e-01 -2.09731877e-01
1.00575075e-01 4.59042221e-01 6.93750381e-01 -1.08639963e-01
-2.71334589e-01 -1.26791093e-02 1.22474506e-01 -3.89093369e-01
1.08081305e+00 -3.18895467e-02 -6.90407932e-01 4.43821967e-01
1.05565429e+00 -1.21966384e-01 -8.10370207e-01 1.74100950e-01
-1.20136157e-01 2.73461580e-01 3.30281883e-01 -5.31074047e-01
-4.78333294e-01 1.15765810e+00 7.80925155e-01 5.34801543e-01
1.19872320e+00 -2.66596884e-01 1.16509080e+00 6.75636470e-01
8.49317729e-01 -1.49510634e+00 -6.93318546e-01 8.71051252e-01
-5.30038238e-01 -9.53119993e-01 -5.97692616e-02 -2.65399426e-01
-1.45980835e-01 1.21583104e+00 3.44287395e-01 -3.94319855e-02
8.98216069e-01 4.64463055e-01 -1.00872755e-01 -3.72644663e-01
-1.02580237e+00 -6.52661443e-01 -2.92200387e-01 1.16576269e-01
6.35007977e-01 5.06767035e-01 -3.27020943e-01 4.48296666e-01
-2.64781326e-01 3.76183003e-01 6.83260143e-01 1.22575200e+00
-1.58047751e-01 -1.51948047e+00 -2.52586186e-01 6.33495808e-01
-4.69793409e-01 -6.67674005e-01 9.86871943e-02 3.27199429e-01
6.63556635e-01 7.05686450e-01 1.54069275e-01 -7.93970346e-01
-1.16901584e-01 3.02529812e-01 6.17118955e-01 -7.95987606e-01
-8.88638794e-01 -2.08052933e-01 1.12026326e-01 -6.02212012e-01
-1.20627746e-01 -3.65218610e-01 -1.39160132e+00 -7.38245964e-01
-5.71027756e-01 2.38866642e-01 1.00965619e+00 8.87047768e-01
1.13473737e+00 8.41641307e-01 4.09142897e-02 -1.07351446e+00
-3.87491643e-01 -5.97968876e-01 -6.92389369e-01 -1.97942376e-01
-3.30029801e-02 -3.39217991e-01 -1.04109533e-01 -4.16974306e-01] | [5.587704181671143, 4.93940544128418] |
e4ecaf73-73a3-43fd-8a96-13084f720f65 | a-survey-of-unsupervised-dependency-parsing | 2010.01535 | null | https://arxiv.org/abs/2010.01535v1 | https://arxiv.org/pdf/2010.01535v1.pdf | A Survey of Unsupervised Dependency Parsing | Syntactic dependency parsing is an important task in natural language processing. Unsupervised dependency parsing aims to learn a dependency parser from sentences that have no annotation of their correct parse trees. Despite its difficulty, unsupervised parsing is an interesting research direction because of its capability of utilizing almost unlimited unannotated text data. It also serves as the basis for other research in low-resource parsing. In this paper, we survey existing approaches to unsupervised dependency parsing, identify two major classes of approaches, and discuss recent trends. We hope that our survey can provide insights for researchers and facilitate future research on this topic. | ['Kewei Tu', 'Hwee Tou Ng', 'Yong Jiang', 'Wenjuan Han'] | 2020-10-04 | null | https://aclanthology.org/2020.coling-main.227 | https://aclanthology.org/2020.coling-main.227.pdf | coling-2020-8 | ['unsupervised-dependency-parsing'] | ['natural-language-processing'] | [-1.39758978e-02 5.34014642e-01 -5.66971838e-01 -1.04520786e+00
-7.55172551e-01 -8.01784694e-01 9.57247764e-02 3.52159381e-01
-3.19516748e-01 8.81237984e-01 5.21859050e-01 -6.78052485e-01
3.99477214e-01 -6.93165064e-01 -2.80152380e-01 -4.02402490e-01
-2.33948022e-01 3.66664469e-01 3.34125161e-01 -1.32539377e-01
1.84735030e-01 3.88354987e-01 -1.08307958e+00 1.69339567e-01
5.66003978e-01 7.71624893e-02 3.00582498e-01 6.94979310e-01
-6.20848179e-01 1.17497468e+00 -7.07800925e-01 -6.14450514e-01
-2.51034677e-01 -6.49330974e-01 -1.29619336e+00 -2.27465019e-01
-6.75141588e-02 -8.82531181e-02 1.37420937e-01 1.17267203e+00
1.14703380e-01 -7.78506026e-02 1.68434590e-01 -8.82926643e-01
-5.37732303e-01 1.29081988e+00 -1.73465163e-01 5.52072942e-01
4.43234175e-01 -5.29542863e-01 1.37277687e+00 -5.03724158e-01
7.07940996e-01 1.27453613e+00 1.82365239e-01 7.26270139e-01
-6.90746009e-01 -4.76024687e-01 5.26353419e-01 -1.24040537e-01
-6.64773881e-01 -3.95311236e-01 9.60073888e-01 -1.77802876e-01
1.26606381e+00 7.24184290e-02 1.25984818e-01 7.76575506e-01
4.84483361e-01 7.37825453e-01 1.15567172e+00 -9.18483436e-01
1.90987006e-01 -2.69623905e-01 9.89447176e-01 7.10061550e-01
4.13496822e-01 -1.83705658e-01 -5.07606626e-01 -1.42870650e-01
6.15896404e-01 -4.47735310e-01 2.03820467e-01 4.49189991e-02
-8.28650475e-01 1.09200728e+00 4.97459807e-02 5.43630600e-01
8.22648332e-02 -3.29510681e-02 4.38426077e-01 4.01118517e-01
4.96234745e-01 3.67769629e-01 -8.39435995e-01 -3.66291195e-01
-3.23612899e-01 -2.06424132e-01 1.35927570e+00 1.15356112e+00
5.93717933e-01 -3.65541615e-02 4.18612331e-01 7.11923778e-01
2.36929700e-01 2.06615120e-01 2.54252940e-01 -8.85794580e-01
7.10075855e-01 4.83189672e-01 -3.33691612e-02 -3.74566674e-01
-6.14368260e-01 3.78331274e-01 -4.80883867e-01 -6.95878938e-02
3.67594093e-01 -5.30604243e-01 -7.69455910e-01 1.59098291e+00
3.01929533e-01 -5.47312140e-01 3.61506820e-01 7.08633304e-01
8.97128046e-01 5.97730100e-01 6.33132577e-01 -5.60851157e-01
1.75735509e+00 -1.07666230e+00 -1.00195897e+00 -6.25051081e-01
7.95700610e-01 -8.69243741e-01 8.39167237e-01 3.74443717e-02
-8.43872964e-01 -3.73574138e-01 -6.77294314e-01 -2.58616775e-01
-2.29068145e-01 -9.69716907e-02 1.31493807e+00 9.29385602e-01
-9.78782713e-01 3.50269556e-01 -1.17846787e+00 -5.72290123e-01
1.31730378e-01 4.37030315e-01 -3.72177184e-01 -1.57011881e-01
-1.12653410e+00 8.79789352e-01 4.97024894e-01 4.52383049e-02
-1.65800318e-01 1.07582219e-01 -1.18893874e+00 -1.17378123e-01
5.07341266e-01 -1.86185911e-01 1.59833634e+00 -8.74619961e-01
-1.54448962e+00 8.50485563e-01 -6.42982244e-01 -2.16796473e-01
-3.08604866e-01 -4.56479430e-01 -4.45151627e-01 4.34577577e-02
3.58210325e-01 3.21357310e-01 4.41645175e-01 -6.85289323e-01
-7.77165890e-01 -4.87289310e-01 3.49102616e-01 1.67818964e-01
-1.14587508e-01 1.00829720e+00 -4.76908505e-01 -6.61566913e-01
4.16049153e-01 -8.71424198e-01 -5.85745096e-01 -8.69379997e-01
-5.09083450e-01 -6.84657991e-01 5.29830456e-01 -5.49479127e-01
1.23771477e+00 -1.92487907e+00 -4.51562293e-02 -4.34876770e-01
8.59681442e-02 -3.72071639e-02 3.04171536e-02 6.78369284e-01
-2.33745128e-01 5.23642838e-01 -5.69662273e-01 -2.38592774e-01
-3.78157973e-01 7.47917235e-01 -2.04273954e-01 1.93217456e-01
6.55272305e-01 9.48991776e-01 -1.06838036e+00 -7.06613660e-01
7.16583356e-02 9.97368023e-02 -1.64942130e-01 4.86999840e-01
-1.11731023e-01 7.64036179e-01 -9.77778971e-01 8.78845036e-01
5.14672756e-01 7.37426355e-02 8.49901557e-01 3.68997723e-01
-6.60927415e-01 8.54581296e-01 -7.27314711e-01 1.44174659e+00
-2.96297401e-01 6.29211247e-01 2.24806279e-01 -1.18717742e+00
8.53247106e-01 3.74360532e-01 1.04242757e-01 -4.35357869e-01
1.92225739e-01 8.99670050e-02 1.52804822e-01 -6.37059212e-01
3.51919621e-01 -2.62271374e-01 -9.19732332e-01 6.44472480e-01
1.16491169e-01 -2.80079581e-02 5.58578491e-01 1.52747974e-01
1.24552608e+00 -1.52400985e-01 7.14936674e-01 -5.63752055e-01
5.65784872e-01 2.96950340e-01 7.16592252e-01 6.95230842e-01
-3.73315305e-01 1.71512350e-01 7.88525879e-01 -8.73164713e-01
-7.50051141e-01 -9.10744250e-01 -1.43126518e-01 1.52435040e+00
-5.54426797e-02 -6.16791904e-01 -7.71366298e-01 -1.17605543e+00
-5.95841348e-01 4.37514365e-01 -4.54998434e-01 5.90947151e-01
-1.32066846e+00 -8.46285522e-01 4.59433854e-01 1.06629157e+00
2.40903661e-01 -1.57451415e+00 -6.46020949e-01 2.50434875e-01
-2.87651211e-01 -1.36989963e+00 1.45947054e-01 8.17988694e-01
-1.20384490e+00 -1.40559125e+00 -1.15943894e-01 -1.60455501e+00
1.01012063e+00 3.15425932e-01 1.45161605e+00 4.04218912e-01
6.29628897e-02 1.88337266e-01 -1.08633959e+00 -4.43597347e-01
-5.93054771e-01 3.74157101e-01 -7.15097561e-02 -8.07808816e-01
8.68732870e-01 -3.18957239e-01 3.05543244e-02 -9.42111015e-04
-8.12287033e-01 -2.01682761e-01 2.85015196e-01 6.81139946e-01
4.28623527e-01 5.03021851e-02 4.51304376e-01 -1.74737835e+00
8.66011143e-01 -4.68583614e-01 -6.89679086e-01 1.64573476e-01
-3.38466316e-01 2.05659941e-01 9.99045610e-01 2.03616574e-01
-1.35471690e+00 2.71728456e-01 -5.53011656e-01 7.04381466e-01
-5.35453796e-01 5.94155550e-01 -3.27903271e-01 2.20461562e-01
1.40471593e-01 -2.50979125e-01 -7.04249978e-01 -7.29126871e-01
4.02223229e-01 4.73739684e-01 4.26963031e-01 -8.07186484e-01
5.92401147e-01 2.76753247e-01 -2.19331950e-01 -8.16640019e-01
-1.08601916e+00 -4.93313402e-01 -1.30781913e+00 2.90990800e-01
1.15573454e+00 -6.24831378e-01 -1.25814840e-01 1.31169513e-01
-1.49683261e+00 -2.11204663e-01 -3.86626348e-02 4.98087198e-01
-2.39170656e-01 6.75803304e-01 -9.88647342e-01 -7.97397316e-01
-2.97341615e-01 -9.40117002e-01 7.90321290e-01 5.48991084e-01
-3.31924319e-01 -1.19695675e+00 3.03162903e-01 2.48423263e-01
-7.74731040e-02 -4.76002172e-02 1.06510115e+00 -8.97508264e-01
-2.03349829e-01 -2.06105202e-01 1.04521411e-02 3.39018255e-01
4.03914243e-01 2.11210743e-01 -7.53370225e-01 2.25948673e-02
3.11051846e-01 -3.68523747e-01 6.32896423e-01 2.66515046e-01
8.55692208e-01 -2.27865763e-02 -2.82066673e-01 2.21048504e-01
1.22641861e+00 3.60196203e-01 2.20747411e-01 2.61118919e-01
6.93882942e-01 1.04828858e+00 8.72072637e-01 8.91081840e-02
6.45406365e-01 1.06554031e-01 1.84784263e-01 6.46466389e-02
1.58346444e-01 -7.06576034e-02 3.54339659e-01 1.40808284e+00
-2.24590778e-01 -2.42771238e-01 -1.16472638e+00 4.42839175e-01
-1.71963012e+00 -5.43178082e-01 -5.22504508e-01 1.66911113e+00
5.90286911e-01 3.06603402e-01 -7.65218586e-02 4.29404527e-02
1.02059054e+00 3.34692359e-01 -1.81057066e-01 -1.07120526e+00
-6.01894297e-02 7.53373742e-01 2.60732919e-01 5.09861290e-01
-1.44285059e+00 1.66512477e+00 7.83901024e+00 -1.84600651e-01
-7.79609084e-01 1.27994016e-01 4.67893511e-01 6.60540342e-01
-1.94679033e-02 5.46146929e-01 -1.00145459e+00 1.38334155e-01
1.20155263e+00 -3.70028280e-02 -2.87155546e-02 1.05619693e+00
8.17665756e-02 -3.38203758e-01 -1.04735494e+00 2.71280497e-01
-1.86468571e-01 -9.26426232e-01 -2.70762026e-01 -3.54537159e-01
2.68688053e-01 1.93395600e-01 -5.86393237e-01 3.27529341e-01
8.34151566e-01 -9.32481587e-01 2.23284289e-01 -3.38759869e-01
4.44249392e-01 -7.60310233e-01 9.13771033e-01 5.47732413e-01
-1.46990860e+00 7.51050096e-03 -6.70432210e-01 -8.70418012e-01
5.29927850e-01 5.32388389e-01 -3.19463462e-01 4.43192035e-01
7.78702259e-01 5.39999545e-01 -3.82168323e-01 4.41139340e-01
-1.14599550e+00 1.11157286e+00 -7.28484169e-02 -1.85026795e-01
1.21351287e-01 -1.91871002e-01 2.31365770e-01 1.59261501e+00
-3.43166113e-01 6.96737468e-01 5.17666578e-01 5.48962578e-02
-8.59582499e-02 3.07425797e-01 -4.83094364e-01 -2.29214266e-01
6.57501221e-01 1.15349758e+00 -1.36356568e+00 -3.48657876e-01
-8.55300605e-01 8.29304099e-01 7.79526055e-01 -7.58604184e-02
-5.97829461e-01 -3.89152348e-01 4.84149873e-01 -4.23564017e-01
2.77153909e-01 -9.03506994e-01 -3.47758681e-01 -1.36471093e+00
1.38235176e-02 -5.58200240e-01 7.81147659e-01 -3.75971556e-01
-1.24769044e+00 9.26301360e-01 1.89518452e-01 -6.96156979e-01
-3.95631254e-01 -7.78439581e-01 -7.71669328e-01 6.46889448e-01
-1.75368285e+00 -8.73899400e-01 9.24893394e-02 2.31595591e-01
1.08242953e+00 1.54146492e-01 1.37497282e+00 -2.34469280e-01
-8.26783240e-01 3.52246583e-01 -3.47100139e-01 8.06200564e-01
5.38568854e-01 -1.22174168e+00 1.11012065e+00 1.18655348e+00
4.32607174e-01 8.52332473e-01 4.76621330e-01 -7.72287369e-01
-1.59950650e+00 -9.26824510e-01 1.49329436e+00 -7.20297277e-01
8.04806352e-01 -3.90216917e-01 -9.12108839e-01 9.45914507e-01
3.78223598e-01 -1.52890477e-02 1.06779969e+00 4.74403411e-01
-3.50056171e-01 3.54962528e-01 -8.41004431e-01 3.57013166e-01
1.06979179e+00 -2.90247560e-01 -1.06009197e+00 1.68478161e-01
8.97766888e-01 -4.69703674e-01 -8.93369496e-01 1.67648539e-01
3.86812061e-01 -9.37922776e-01 3.20211977e-01 -9.05856550e-01
4.85718817e-01 1.53233141e-01 -6.56213285e-03 -9.65740502e-01
-3.96749347e-01 -5.10244131e-01 1.01101406e-01 1.51592886e+00
6.69302166e-01 -5.85325241e-01 9.04911458e-01 1.08565879e+00
-1.39164031e-01 -4.10988629e-01 -7.01700032e-01 -6.87191367e-01
4.40336734e-01 -6.74504519e-01 3.11495930e-01 9.98165369e-01
4.07981753e-01 7.59370446e-01 -2.85056215e-02 2.77564883e-01
5.75065851e-01 3.12428981e-01 5.60502410e-01 -1.34922945e+00
1.58700775e-02 1.07128978e-01 3.51182967e-02 -1.17139196e+00
6.78724885e-01 -5.95178545e-01 3.44569594e-01 -1.54175103e+00
1.18904613e-01 -4.85544413e-01 -1.60674319e-01 6.96862042e-01
-2.94225663e-01 3.67414393e-02 1.33305460e-01 3.27923834e-01
-5.96677244e-01 -1.74202338e-01 1.00467980e+00 2.11247414e-01
-2.57142693e-01 1.53373212e-01 -1.07030189e+00 1.07517707e+00
1.29913116e+00 -1.15002751e+00 -2.49479160e-01 -8.12705636e-01
7.87681937e-02 1.89027190e-01 -5.05661786e-01 -4.69116628e-01
2.33293161e-01 -5.19735873e-01 2.20215954e-02 -5.07511616e-01
-2.08027750e-01 -6.00605249e-01 -5.32035291e-01 2.03278810e-01
-1.85286433e-01 5.45894861e-01 -1.25823030e-02 3.64081442e-01
-3.91173929e-01 -7.35004604e-01 4.76564616e-01 -5.93594313e-01
-1.01897812e+00 -1.84799805e-01 -8.66038024e-01 2.30212718e-01
8.49914193e-01 1.44053295e-01 -2.07751393e-01 2.45353207e-02
-5.92349112e-01 2.47159094e-01 4.16195154e-01 6.49329066e-01
4.21796471e-01 -6.28473103e-01 -6.06450200e-01 1.47645399e-01
-6.22736402e-02 2.30911970e-01 -3.10721159e-01 3.82542424e-02
-5.17964482e-01 6.68881834e-01 -1.65142715e-01 -2.26546377e-01
-1.43743789e+00 5.24683893e-01 -4.69989777e-01 -6.42232180e-01
-6.65887892e-01 8.76997888e-01 -1.39131192e-02 -3.99799824e-01
8.61392915e-02 -3.11784148e-01 -6.40729308e-01 -2.35565513e-01
5.90423584e-01 -1.92664444e-01 -7.80203342e-02 -7.39418089e-01
-7.40892708e-01 3.66444558e-01 -2.36458719e-01 5.39363846e-02
1.39306819e+00 -6.16699271e-02 -6.94780707e-01 5.08622289e-01
9.01664138e-01 3.22085291e-01 -7.87020385e-01 1.37567848e-01
6.79675639e-01 -1.29171297e-01 -3.57631087e-01 -4.18240577e-01
-8.04615676e-01 7.35525250e-01 -1.65739104e-01 5.80322504e-01
1.17431295e+00 5.63919544e-01 6.38635993e-01 7.14859784e-01
6.56009316e-01 -1.09636486e+00 -2.29143992e-01 1.21189356e+00
3.10349286e-01 -1.35561287e+00 7.90592656e-02 -1.15284729e+00
-5.73054492e-01 1.35353899e+00 6.33231640e-01 -3.54935229e-01
7.87394166e-01 8.24026346e-01 4.33855087e-01 -2.27288261e-01
-6.20159507e-01 -4.30261374e-01 -4.38926101e-01 8.86552930e-01
1.12046242e+00 1.65728614e-01 -8.86254191e-01 1.11350977e+00
-4.57668424e-01 -4.12557691e-01 6.47724748e-01 1.57288826e+00
-5.94156623e-01 -2.03924966e+00 -2.62752265e-01 6.33176640e-02
-8.91833186e-01 -4.39559221e-01 -7.64320195e-01 9.57827508e-01
-3.01592588e-01 1.32171738e+00 8.94415751e-03 -1.23808697e-01
2.46810555e-01 2.62436837e-01 4.69205707e-01 -1.29004860e+00
-6.46542311e-01 -1.19393460e-01 6.25381112e-01 -3.39823812e-01
-7.61357129e-01 -5.82270801e-01 -1.66500521e+00 1.32822189e-02
-3.38234395e-01 6.36664152e-01 6.43760979e-01 1.12326002e+00
-1.26159452e-02 1.70370981e-01 6.99961543e-01 -5.31912327e-01
-8.33116770e-02 -6.83320522e-01 -3.52111161e-01 2.26181410e-02
-3.15636694e-02 -2.71469563e-01 -2.43674703e-02 4.21042562e-01] | [10.368191719055176, 9.72996711730957] |
e4bc04e6-801d-4ca9-8ebb-a64890b35f9f | practical-bandits-an-industry-perspective | 2302.01223 | null | https://arxiv.org/abs/2302.01223v1 | https://arxiv.org/pdf/2302.01223v1.pdf | Practical Bandits: An Industry Perspective | The bandit paradigm provides a unified modeling framework for problems that require decision-making under uncertainty. Because many business metrics can be viewed as rewards (a.k.a. utilities) that result from actions, bandit algorithms have seen a large and growing interest from industrial applications, such as search, recommendation and advertising. Indeed, with the bandit lens comes the promise of direct optimisation for the metrics we care about. Nevertheless, the road to successfully applying bandits in production is not an easy one. Even when the action space and rewards are well-defined, practitioners still need to make decisions regarding multi-arm or contextual approaches, on- or off-policy setups, delayed or immediate feedback, myopic or long-term optimisation, etc. To make matters worse, industrial platforms typically give rise to large action spaces in which existing approaches tend to break down. The research literature on these topics is broad and vast, but this can overwhelm practitioners, whose primary aim is to solve practical problems, and therefore need to decide on a specific instantiation or approach for each project. This tutorial will take a step towards filling that gap between the theory and practice of bandits. Our goal is to present a unified overview of the field and its existing terminology, concepts and algorithms -- with a focus on problems relevant to industry. We hope our industrial perspective will help future practitioners who wish to leverage the bandit paradigm for their application. | ['Devesh Parekh', 'Zahra Nazari', 'Ben London', 'Ying Li', 'Olivier Jeunen', 'Bram van den Akker'] | 2023-02-02 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 1.65150434e-01 1.63571220e-02 -1.01861906e+00 -3.73955339e-01
-9.17850912e-01 -8.03050876e-01 4.88044620e-01 -1.19834982e-01
-1.17734514e-01 9.95513260e-01 9.12840888e-02 -7.83390760e-01
-9.30333257e-01 -6.13208532e-01 -5.71328819e-01 -7.89595366e-01
-1.20397501e-01 5.64923882e-01 -2.95657516e-01 -1.26488462e-01
4.32240695e-01 3.79125863e-01 -1.43189621e+00 1.70756713e-01
5.37951827e-01 1.46322453e+00 1.76521838e-02 6.20578229e-01
-2.74988502e-01 7.80409515e-01 -4.29114103e-01 -8.44451904e-01
5.81116915e-01 -3.41089159e-01 -8.79556775e-01 2.40165368e-02
-1.79497778e-01 -4.40373749e-01 8.01357999e-02 8.74442816e-01
3.31240803e-01 1.10617287e-01 2.88009614e-01 -1.35628510e+00
-4.75058556e-01 9.30827558e-01 -5.39424956e-01 2.07182720e-01
1.46809220e-02 2.99669862e-01 1.57458603e+00 -3.39035783e-03
3.05643082e-02 1.25162017e+00 3.10911804e-01 1.69500887e-01
-1.48748958e+00 -2.65516341e-01 6.45806968e-01 2.97034770e-01
-6.29562140e-01 -4.10522193e-01 5.83378494e-01 -3.95583451e-01
7.86791801e-01 9.08548355e-01 9.09336507e-01 8.49714041e-01
-7.08222836e-02 1.26931238e+00 1.20259798e+00 -4.35362697e-01
4.28387195e-01 3.07559788e-01 1.81920022e-01 -1.96389705e-01
3.67565691e-01 4.61499959e-01 -2.73262978e-01 -2.69769311e-01
7.47282863e-01 2.24143565e-01 -6.73077404e-02 -6.03310287e-01
-1.06460559e+00 1.10202527e+00 3.21742207e-01 1.87401533e-01
-8.00459445e-01 5.39853454e-01 3.24746162e-01 6.67009354e-01
4.77250785e-01 7.40771174e-01 -5.49626172e-01 -6.73759043e-01
-1.03834200e+00 5.34453690e-01 9.42530751e-01 7.35291600e-01
5.65236628e-01 -1.55941442e-01 -5.08098602e-01 9.13663864e-01
3.94373596e-01 1.44473046e-01 1.74269259e-01 -1.18415070e+00
5.14153540e-01 1.31940678e-01 7.13635743e-01 -7.38374352e-01
-1.95413455e-01 -5.90635538e-01 -3.28827679e-01 2.65623301e-01
5.60039878e-01 -2.87541837e-01 -7.04840004e-01 1.29299939e+00
1.32555589e-01 -2.21268237e-01 -1.76817521e-01 9.10297096e-01
-1.63942695e-01 6.99749351e-01 -2.14800626e-01 -6.56406462e-01
1.17422867e+00 -1.05926156e+00 -8.53538096e-01 -6.62721813e-01
4.90747452e-01 -7.82425106e-01 7.37986743e-01 8.44104111e-01
-1.34998262e+00 1.49339169e-01 -8.12825501e-01 5.97599506e-01
-1.98018506e-01 -4.21514839e-01 1.27559543e+00 1.25405312e+00
-6.71013594e-01 6.94890261e-01 -7.12974310e-01 -1.29066855e-01
4.53351438e-01 5.10899961e-01 3.87652904e-01 -2.09551305e-01
-8.82388711e-01 1.08275986e+00 2.63695776e-01 4.32817012e-01
-5.48140109e-01 -6.70409203e-01 -3.33205551e-01 2.56963372e-01
9.33501899e-01 -6.03526771e-01 1.90580237e+00 -1.19559944e+00
-1.74666286e+00 2.53579497e-01 2.67783165e-01 -7.50817895e-01
6.75457001e-01 -2.68426895e-01 -1.76633984e-01 -5.00652552e-01
-1.38740674e-01 6.53112978e-02 6.94001317e-01 -9.10544515e-01
-1.04629707e+00 -2.79508471e-01 6.01752162e-01 2.23905027e-01
-8.82545412e-02 3.48064423e-01 -7.32360855e-02 -5.29911280e-01
-1.14077583e-01 -1.05036056e+00 -5.86017609e-01 -5.88902831e-01
-2.55444407e-01 -1.49888650e-01 3.12170506e-01 -2.91575789e-01
1.51350617e+00 -1.78326797e+00 -1.51628554e-01 2.90160447e-01
-3.55441600e-01 -8.90888367e-03 1.59299061e-01 7.64167666e-01
-1.13469124e-01 4.01570439e-01 1.84418991e-01 7.40374848e-02
6.06432199e-01 2.12507620e-01 -4.49372023e-01 4.96650577e-01
-5.74054867e-02 8.86857092e-01 -7.60492146e-01 -7.75258895e-03
3.78801942e-01 -2.86649495e-01 -5.41902959e-01 -1.19823828e-01
-5.25470495e-01 -3.98053974e-02 -7.78828502e-01 9.04328465e-01
3.70123267e-01 -3.20158571e-01 1.11003280e-01 5.01085937e-01
-2.27405682e-01 5.57018459e-01 -1.38985693e+00 1.20443869e+00
-6.35880172e-01 3.41826886e-01 3.64687562e-01 -1.55557454e+00
5.00030935e-01 3.24404597e-01 7.84532964e-01 -5.47649264e-01
1.22596979e-01 3.95348191e-01 2.42980272e-01 -2.21583217e-01
4.30124164e-01 -2.78692126e-01 -9.02543515e-02 6.75333381e-01
-4.28242385e-01 -2.24225998e-01 3.80407333e-01 -3.34212571e-01
9.79705811e-01 -4.68080044e-02 3.17146152e-01 -1.83012515e-01
8.91887993e-02 1.87509298e-01 3.80649656e-01 1.00587225e+00
-1.03632092e-01 1.19870424e-01 8.64754438e-01 -6.06672943e-01
-6.70437753e-01 -5.79489946e-01 -1.40109807e-01 1.35222876e+00
-3.88299748e-02 5.01945168e-02 -4.39928174e-01 -6.76031530e-01
4.09797341e-01 8.91989410e-01 -3.90359402e-01 1.74980700e-01
-2.40897149e-01 -9.53002632e-01 -3.39225829e-01 2.91563481e-01
3.45064588e-02 -8.68410349e-01 -6.63464248e-01 7.48982787e-01
2.59441901e-02 -6.84962988e-01 -2.42500678e-01 3.53103578e-01
-1.00204623e+00 -7.77313471e-01 -7.92506218e-01 2.93350648e-02
4.23244350e-02 2.69101828e-01 1.19029486e+00 -4.03462291e-01
-8.28064010e-02 5.29422998e-01 -3.73494208e-01 -9.01894271e-01
-2.54917592e-02 -1.82200313e-01 -1.82031065e-01 1.96039304e-01
2.88714916e-01 -5.40544689e-01 -8.99635315e-01 6.21983051e-01
-6.75121844e-01 -3.25881064e-01 7.56039262e-01 8.84670019e-01
3.70638162e-01 -4.08039987e-03 7.06599534e-01 -8.12148213e-01
9.26617384e-01 -5.31403244e-01 -1.07814658e+00 4.32767838e-01
-9.90858912e-01 -6.95208833e-02 -1.74322091e-02 -4.20527399e-01
-8.46633255e-01 -2.47388050e-01 2.57912993e-01 -1.21266022e-01
2.40160525e-01 9.41417456e-01 1.53968155e-01 1.54714674e-01
5.70889771e-01 -1.91910982e-01 -1.43456385e-01 -3.59562725e-01
5.71029186e-01 8.46053362e-01 -3.86118367e-02 -8.15340042e-01
2.00553864e-01 2.26916254e-01 -1.66823640e-01 -3.02860796e-01
-7.79337168e-01 -5.34710646e-01 1.66169927e-01 -1.65371493e-01
3.76619369e-01 -6.23066425e-01 -8.29399645e-01 -1.37918919e-01
-8.74551058e-01 -4.33155954e-01 -5.77493191e-01 6.05366170e-01
-1.16165268e+00 -1.74579725e-01 -3.25066209e-01 -1.55101287e+00
-4.79396023e-02 -1.43653774e+00 5.43199301e-01 2.61021435e-01
-2.80852377e-01 -7.57878184e-01 -1.86439022e-01 7.45851994e-01
6.18963182e-01 1.36353999e-01 6.00880206e-01 -6.00898981e-01
-1.18251765e+00 -5.35807371e-01 -1.42339002e-02 4.39493120e-01
4.89561148e-02 -8.63042995e-02 -8.33200991e-01 -2.78429240e-01
-3.51535529e-02 -2.76677728e-01 4.82495636e-01 9.65802073e-01
1.25166190e+00 -6.13529444e-01 -2.65380442e-01 1.83819905e-01
1.25616896e+00 5.92579842e-01 5.04341841e-01 7.33075023e-01
-1.10242575e-01 7.51735032e-01 1.08580828e+00 6.88766062e-01
2.26067175e-04 9.06618536e-01 8.96905661e-01 2.85876691e-01
5.85028887e-01 1.08654313e-01 3.25312316e-01 1.68251857e-01
-3.18944782e-01 -2.56725043e-01 -6.98916674e-01 6.96349561e-01
-2.26768041e+00 -1.09914160e+00 1.39932245e-01 2.49495149e+00
7.17088878e-01 4.37651753e-01 5.44578552e-01 1.60116091e-01
6.48580015e-01 2.87574437e-02 -6.99722528e-01 -9.68084574e-01
3.45642447e-01 -2.49187157e-01 9.32104528e-01 2.33988285e-01
-9.87160981e-01 4.46516782e-01 6.71249628e+00 9.25049365e-01
-1.03636360e+00 -8.12165588e-02 1.09915292e+00 -6.00213766e-01
-2.56991118e-01 2.22562522e-01 -6.65748358e-01 5.85791171e-01
1.21806192e+00 -4.81223226e-01 9.34873819e-01 1.17690229e+00
5.07062912e-01 -2.75875241e-01 -1.35916257e+00 1.13955879e+00
-7.46578276e-01 -1.49574089e+00 -6.70329452e-01 6.27901971e-01
9.36713934e-01 8.47485811e-02 3.45328331e-01 4.43933159e-01
9.03103411e-01 -1.08093381e+00 7.60420501e-01 2.87448764e-01
2.44225129e-01 -8.32284570e-01 6.72570884e-01 4.81420100e-01
-4.32945669e-01 -6.67487264e-01 -2.49002963e-01 -4.93288845e-01
3.08240145e-01 7.65170991e-01 -6.72063649e-01 5.18252790e-01
6.70585394e-01 2.92228371e-01 4.48443532e-01 1.61031556e+00
-4.68775723e-03 6.36573792e-01 -5.23676336e-01 -4.71537054e-01
6.30639672e-01 -5.55772066e-01 3.20988297e-01 8.79038572e-01
4.66626704e-01 -9.80849639e-02 8.14863071e-02 7.77623475e-01
2.15615913e-01 7.64302388e-02 -4.07449573e-01 -4.32019085e-01
3.67982030e-01 1.05010962e+00 -6.44372404e-01 -1.03115231e-01
-5.70495605e-01 3.04582715e-01 -2.25213662e-01 4.22023982e-01
-7.34681427e-01 -8.85462463e-02 1.20088291e+00 1.33285254e-01
4.34114814e-01 1.24830857e-01 -6.57318532e-01 -5.59080720e-01
9.81647894e-02 -1.22325945e+00 3.96412551e-01 -5.27083814e-01
-1.12523580e+00 -5.47419339e-02 2.45574102e-01 -9.75562394e-01
-6.64074838e-01 -6.27406001e-01 -4.08060551e-01 7.82752872e-01
-1.55200589e+00 -7.00919032e-01 4.66247916e-01 1.08111605e-01
6.09613419e-01 1.16022483e-01 4.91020471e-01 1.20059401e-01
-4.39104199e-01 2.31204480e-01 7.01586485e-01 -4.34307158e-01
3.67268354e-01 -1.21786988e+00 1.13589846e-01 4.40022528e-01
2.87201464e-01 5.41290998e-01 9.68926132e-01 -1.42345250e-01
-1.55870497e+00 -5.52898586e-01 3.95974308e-01 -4.91352230e-01
1.06354082e+00 -8.78087729e-02 -1.21039100e-01 8.98383558e-01
8.10268670e-02 -5.49853966e-02 5.25149226e-01 6.67932570e-01
1.82942465e-01 -5.39058924e-01 -1.02346206e+00 6.08085096e-01
6.98242903e-01 -1.78099602e-01 -1.94630608e-01 5.07237196e-01
5.39097190e-01 -3.64368290e-01 -8.90710831e-01 1.80900127e-01
7.64134407e-01 -1.18041480e+00 8.68780494e-01 -7.88976610e-01
1.52701512e-01 1.97629929e-01 -1.62851915e-01 -1.40660608e+00
-2.83416599e-01 -1.32577622e+00 -1.20718136e-01 1.12294030e+00
6.81654513e-01 -8.05747867e-01 1.06937933e+00 1.06720686e+00
-7.31336232e-03 -1.00125504e+00 -1.07614028e+00 -1.06779981e+00
1.99597441e-02 -8.93878937e-01 9.68628645e-01 6.65845811e-01
2.63224602e-01 1.11493520e-01 -5.20176351e-01 -2.67117441e-01
4.35106069e-01 5.32374382e-01 7.81782568e-01 -1.17226028e+00
-4.85565186e-01 -9.41084146e-01 -1.43262357e-01 -1.43756557e+00
-5.20451248e-01 -2.92059839e-01 -1.40717840e-02 -1.55307114e+00
-6.64362758e-02 -6.96871996e-01 -7.36640632e-01 3.02486598e-01
2.28900120e-01 -1.85559586e-01 3.98776650e-01 -2.63854372e-03
-5.45598269e-01 2.30954647e-01 1.26133871e+00 -1.49174720e-01
-3.49584728e-01 9.19988632e-01 -1.25465405e+00 5.35506248e-01
8.92174304e-01 -4.64116693e-01 -5.85868537e-01 -1.98494256e-01
6.40703619e-01 4.45900321e-01 1.81837603e-01 -3.82410288e-01
4.02297676e-02 -8.95746231e-01 -2.12228760e-01 -4.87567753e-01
5.02326250e-01 -1.17307532e+00 4.05154407e-01 1.83463022e-01
-3.85004163e-01 -7.52948448e-02 3.32873035e-03 7.67232180e-01
-2.23459184e-01 -5.30831218e-01 3.80062401e-01 -3.06282312e-01
-2.52876252e-01 2.65716553e-01 -2.70973951e-01 -1.36642829e-01
1.20563591e+00 -2.97198415e-01 -1.35845780e-01 -8.57707620e-01
-7.96275675e-01 5.46729326e-01 1.11401796e-01 3.49172980e-01
-1.08961456e-01 -1.25650346e+00 -5.07431567e-01 -3.84407789e-01
-6.65480569e-02 -1.51354894e-01 7.99210966e-02 1.04979992e+00
-1.73252955e-01 9.68329847e-01 9.91460457e-02 -4.30379301e-01
-9.11114812e-01 6.77955091e-01 2.29201049e-01 -6.84374452e-01
-2.50685569e-02 8.17822218e-01 1.06905987e-02 1.36947215e-01
4.07646954e-01 -5.80712736e-01 1.37832820e-01 3.59985203e-01
4.77250099e-01 5.28672338e-01 1.68749079e-01 1.80488706e-01
-6.46803603e-02 -4.10678275e-02 -5.40966131e-02 -2.53034145e-01
1.69840479e+00 -2.63101280e-01 2.37196852e-02 5.37336588e-01
6.04641438e-01 -2.43696615e-01 -1.40074348e+00 -1.66492105e-01
3.67915004e-01 -7.62618780e-01 4.31031525e-01 -1.01872766e+00
-9.56799388e-01 6.29230738e-01 4.19168264e-01 1.29767740e+00
9.24312830e-01 -5.22846170e-02 3.11976016e-01 3.41478854e-01
5.17478168e-01 -1.46174645e+00 -3.10129315e-01 1.40225485e-01
9.14162219e-01 -1.15679491e+00 2.37592116e-01 -1.24120317e-01
-5.83515167e-01 8.57254386e-01 -7.52657279e-02 6.62952662e-02
6.40365362e-01 -9.93933808e-03 -1.38005182e-01 9.13454071e-02
-9.33479548e-01 -2.91348934e-01 1.25022918e-01 5.40857792e-01
3.71044457e-01 3.14949453e-01 -4.91167605e-01 6.48787558e-01
-3.13719153e-01 1.14021555e-01 4.29968297e-01 8.29345822e-01
-4.42509025e-01 -1.50506318e+00 -7.13129818e-01 9.42277908e-01
-8.79725456e-01 7.52052069e-02 -1.35773405e-01 7.97612488e-01
-3.26866329e-01 1.21437764e+00 -1.55883789e-01 8.43743756e-02
4.38062876e-01 -1.96150243e-02 4.62412566e-01 -5.07701457e-01
-5.08819282e-01 3.95776123e-01 5.36361158e-01 -6.89656913e-01
-3.42176884e-01 -1.09313595e+00 -3.47685635e-01 -4.94329810e-01
-7.14912713e-01 3.14463526e-01 1.01396978e+00 7.79280424e-01
3.73883694e-02 5.22487521e-01 7.05589175e-01 -8.49288881e-01
-1.29981840e+00 -8.42475832e-01 -8.69631171e-01 -1.69055074e-01
4.70392227e-01 -7.91988492e-01 -1.58430398e-01 -3.71729493e-01] | [4.526704788208008, 3.2534019947052] |
0a295534-9c7d-46bc-94e9-6ef7090f067d | odianlps-participation-in-wat2020 | null | null | https://aclanthology.org/2020.wat-1.10 | https://aclanthology.org/2020.wat-1.10.pdf | ODIANLP’s Participation in WAT2020 | This paper describes the ODIANLP submission to WAT 2020. We have participated in the English-Hindi Multimodal task and Indic task. We have used the state-of-the-art Transformer model for the translation task and InceptionResNetV2 for the Hindi Image Captioning task. Our submission tops in English->Hindi Multimodal task in its track and Odia<->English translation tasks. Also, our submissions performed well in the Indic Multilingual tasks. | ['Ondřej Bojar', 'Biranchi Narayan Nayak', 'Priyanka Pattnaik', 'Satya Prakash Biswal', 'Debasish Kumar Mallick', 'Satya Ranjan Dash', 'Amulya Ratna Dash', 'Petr Motlicek', 'Shantipriya Parida'] | null | null | null | null | aacl-wat-2020-12 | ['hindi-image-captioning'] | ['computer-vision'] | [-3.32389235e-01 1.33047014e-01 -1.22405902e-01 -4.96942073e-01
-1.62001622e+00 -7.28931189e-01 1.10884583e+00 -5.21858573e-01
-7.73435712e-01 1.21848786e+00 4.61224616e-01 -4.44293916e-01
4.82012689e-01 -2.10729420e-01 -9.60428178e-01 -1.16327748e-01
3.49402219e-01 1.32912445e+00 1.69593483e-01 -6.95574224e-01
-5.37264824e-01 -8.22782889e-02 -5.03367841e-01 9.88637269e-01
4.49158877e-01 7.48349130e-01 2.79899418e-01 8.20228159e-01
3.98637466e-02 1.24936652e+00 -4.95423764e-01 -8.68545830e-01
2.61906590e-02 -5.06494582e-01 -1.31571639e+00 -6.37051642e-01
6.73062384e-01 -3.00015867e-01 -8.86435449e-01 9.47808921e-01
7.77960479e-01 -3.95237654e-01 4.44239914e-01 -1.36873734e+00
-9.70787406e-01 1.05285776e+00 -2.78278291e-01 2.56345302e-01
4.73510712e-01 -5.98567873e-02 1.03169894e+00 -1.27445161e+00
1.25177372e+00 1.32439923e+00 3.83666635e-01 4.90941256e-01
-1.00056732e+00 -7.31960952e-01 -4.59288955e-01 7.05969810e-01
-1.40639162e+00 -5.46304464e-01 5.44991076e-01 -1.69121027e-01
1.48182404e+00 3.65763932e-01 4.10414696e-01 1.50668073e+00
2.74503529e-01 1.56063581e+00 1.29142082e+00 -2.62679935e-01
-6.33387327e-01 3.62189233e-01 -1.55598268e-01 5.29708982e-01
-7.30941057e-01 8.39611441e-02 -5.72950363e-01 4.53914315e-01
1.35345355e-01 -9.09491897e-01 2.20227063e-01 3.39463681e-01
-1.83063996e+00 8.08427215e-01 2.83138901e-01 2.81594425e-01
-1.95189252e-01 2.77866513e-01 8.61368537e-01 7.94426858e-01
4.57871735e-01 2.17327997e-01 -4.87663239e-01 -5.93845010e-01
-1.04502606e+00 1.85068786e-01 7.44359672e-01 1.46163940e+00
4.81143355e-01 2.54085865e-02 -4.76239562e-01 1.21447098e+00
2.68871754e-01 1.12045193e+00 2.89636761e-01 -9.96658623e-01
1.13650560e+00 -1.43160760e-01 -2.87424456e-02 -1.79672644e-01
-2.99822479e-01 9.28517710e-03 -8.21537971e-01 -1.54563963e-01
9.50118825e-02 -1.59730330e-01 -1.12343538e+00 1.60085237e+00
-2.77803719e-01 -6.91000938e-01 3.67463470e-01 7.83025801e-01
1.30916274e+00 1.09321415e+00 3.38779747e-01 1.42697930e-01
1.45040369e+00 -1.35881746e+00 -9.19851899e-01 -2.40218893e-01
4.70317543e-01 -1.38579702e+00 1.05000329e+00 1.10857978e-01
-1.61344767e+00 -8.83197665e-01 -9.73241091e-01 -4.73180979e-01
-5.39924979e-01 3.68155777e-01 7.24437833e-02 8.86822641e-02
-1.56644523e+00 -5.96690550e-02 -4.84818757e-01 -9.23037112e-01
-2.15690941e-01 2.20609292e-01 -6.65380478e-01 -3.79999369e-01
-1.88380110e+00 1.57517707e+00 6.37212038e-01 4.12428193e-02
-1.17013097e+00 -4.05551553e-01 -8.89728010e-01 -6.06666028e-01
-2.23291799e-01 -4.88599151e-01 1.58929849e+00 -5.56374311e-01
-1.29364419e+00 1.40596688e+00 2.66069770e-01 -8.09442520e-01
7.70622373e-01 -1.87380090e-01 -1.04469228e+00 2.42386892e-01
3.86006951e-01 1.50201559e+00 3.71196389e-01 -8.92377555e-01
-7.19406128e-01 -2.12833673e-01 1.28328666e-01 5.99338412e-01
2.43372038e-01 5.36060333e-01 -9.03447092e-01 -4.98500794e-01
-3.69406193e-01 -1.23444605e+00 2.16149941e-01 -4.41806376e-01
-6.69450760e-01 -7.82604292e-02 1.12277484e+00 -1.14979088e+00
8.10987890e-01 -2.22452307e+00 2.86762238e-01 -4.38842893e-01
-2.29580373e-01 1.68721974e-01 -5.53617537e-01 7.20167398e-01
-1.34761282e-03 -1.25731915e-01 1.96755067e-01 -3.95566434e-01
5.07225096e-01 6.53735995e-01 -4.70257342e-01 4.07806396e-01
4.01542224e-02 1.51914871e+00 -6.60847127e-01 -1.02115273e+00
3.87492925e-01 4.50313210e-01 1.31615236e-01 4.55391943e-01
-1.46058112e-01 3.30828041e-01 5.72634377e-02 4.79329735e-01
4.78961885e-01 1.72822252e-01 1.19051682e-02 -5.67176104e-01
-3.22475642e-01 2.28629336e-01 -2.15660125e-01 2.34573221e+00
-7.31585801e-01 8.56592238e-01 2.37880163e-02 -7.99381793e-01
5.20923972e-01 6.69028640e-01 3.97251666e-01 -1.20473242e+00
-2.98818536e-02 6.25508666e-01 -1.07074484e-01 -5.73241770e-01
7.68332720e-01 -1.93686754e-01 -6.62936270e-01 -9.04645175e-02
4.89788175e-01 -5.91291189e-01 6.09056711e-01 4.18100476e-01
5.11523664e-01 6.46888733e-01 5.25613204e-02 -4.21017438e-01
3.91538024e-01 3.84783566e-01 -1.70472801e-01 7.13908434e-01
-1.85791716e-01 8.75906289e-01 2.68434770e-02 -4.57396477e-01
-1.29492569e+00 -1.39483786e+00 -8.69378373e-02 1.12862039e+00
-1.43675953e-01 -2.58446097e-01 -5.89715302e-01 -6.36379182e-01
-7.01079786e-01 1.12445545e+00 -5.73817372e-01 2.50812322e-01
-7.70650566e-01 -6.98689520e-01 1.31935489e+00 5.03006339e-01
8.64527762e-01 -1.42089963e+00 -1.68270096e-01 5.88996351e-01
-1.01937175e+00 -1.71500337e+00 -6.19931459e-01 3.50393355e-01
-2.39086561e-02 -3.93317759e-01 -1.25917947e+00 -1.20948017e+00
-2.02468976e-01 -3.73607785e-01 1.58771324e+00 -8.57191741e-01
-3.00480783e-01 4.21541005e-01 -3.17572534e-01 -3.13348979e-01
-6.68974638e-01 4.91551399e-01 -4.44125324e-01 -7.71995246e-01
2.80053258e-01 -1.09545380e-01 -1.19464412e-01 2.68266171e-01
-5.69625020e-01 4.47681218e-01 5.16098857e-01 7.24971116e-01
5.01592577e-01 -6.44367576e-01 1.70009792e-01 -5.03481686e-01
4.29998815e-01 -1.69781461e-01 -4.96778548e-01 6.45986259e-01
-3.97385620e-02 -7.66348094e-02 5.83717406e-01 1.03388615e-02
-1.11155403e+00 4.70588394e-02 -7.13346779e-01 -2.53916811e-02
-7.40123317e-02 6.66234553e-01 -3.80679905e-01 4.42068249e-01
3.93394142e-01 3.56360137e-01 -3.14315945e-01 -3.19616437e-01
8.59050632e-01 8.29252839e-01 9.03248787e-01 -6.30834222e-01
5.24574995e-01 1.57921940e-01 -2.91748941e-01 -9.22847748e-01
-6.96847916e-01 -1.61750406e-01 -7.20248342e-01 -2.37782270e-01
1.57717884e+00 -1.45785403e+00 -1.27012029e-01 6.11145914e-01
-1.73239732e+00 -5.27103662e-01 -3.48346978e-01 3.94797921e-01
-8.45675349e-01 2.78979354e-02 -1.29411244e+00 -1.68953836e-01
-6.44927561e-01 -1.51186168e+00 1.19139946e+00 -3.50367546e-01
-2.43104994e-01 -9.39462006e-01 5.28482080e-01 4.18101311e-01
4.75375682e-01 -1.93440303e-01 8.71823967e-01 -4.90791529e-01
-1.76624835e-01 -1.84988096e-01 -7.36320257e-01 2.45818168e-01
-6.09982073e-01 -2.19452560e-01 -7.31749237e-01 -1.50255948e-01
-6.01741791e-01 -1.02227199e+00 7.14868248e-01 1.99438423e-01
4.24425930e-01 -1.03978552e-01 -2.60040723e-02 3.69594425e-01
1.23001337e+00 3.16396534e-01 9.81862426e-01 4.55383509e-01
6.45829260e-01 3.69610578e-01 6.29017174e-01 -1.17426284e-01
1.03989398e+00 1.34762335e+00 -3.02827954e-02 -6.44276857e-01
-7.60610521e-01 -4.06624198e-01 7.02621758e-01 1.50824237e+00
3.06941658e-01 -5.30220389e-01 -9.79367793e-01 8.11274767e-01
-2.00754476e+00 -7.66887009e-01 -3.77817214e-01 1.48543811e+00
9.97496963e-01 -1.02733143e-01 1.37024224e-01 -6.10048711e-01
4.05944675e-01 1.29281789e-01 8.81341174e-02 -7.31495440e-01
-5.81814826e-01 1.73899516e-01 3.68061841e-01 8.79960477e-01
-1.23271859e+00 1.64268410e+00 6.75800991e+00 9.68885303e-01
-8.55737746e-01 7.23931253e-01 4.44196850e-01 2.70572305e-01
-1.31623164e-01 -1.46538600e-01 -9.45337296e-01 1.49464756e-01
1.47227538e+00 -1.32515326e-01 6.43406808e-01 4.83853877e-01
-1.18924983e-01 9.69658270e-02 -8.85787904e-01 9.68663871e-01
2.52494156e-01 -1.11334157e+00 1.19340532e-01 -1.56381309e-01
4.72652912e-01 1.12916493e+00 1.33104071e-01 7.42716849e-01
6.72753274e-01 -1.01362491e+00 1.14217746e+00 2.04561144e-01
1.44933724e+00 -6.82888687e-01 8.49259317e-01 -2.27572829e-01
-1.18511891e+00 4.77592200e-01 -4.13120717e-01 4.91644502e-01
6.80693686e-01 8.99577066e-02 -6.91873848e-01 7.55377710e-01
6.77607834e-01 7.93076217e-01 -5.90726137e-01 5.04202127e-01
-5.21104217e-01 6.48357511e-01 -3.12939703e-01 1.98273391e-01
8.56991887e-01 -1.35876715e-01 4.82068807e-01 1.87599969e+00
2.92427540e-01 -4.59898740e-01 5.78444064e-01 3.12359184e-01
-2.71252304e-01 2.46379122e-01 -7.89125323e-01 -1.45488754e-01
-1.09150819e-01 1.09522319e+00 -2.65735120e-01 -6.91089571e-01
-5.34851432e-01 1.52637315e+00 8.70106220e-02 4.25343722e-01
-1.32561600e+00 -4.06988800e-01 1.24241635e-01 -8.05829465e-02
7.89376944e-02 -1.96254909e-01 1.89487368e-01 -1.18692756e+00
-3.41429740e-01 -1.04777527e+00 6.16210818e-01 -1.41454434e+00
-1.01698744e+00 1.38050520e+00 1.75501570e-01 -1.05284178e+00
-6.05565131e-01 -4.39235985e-01 -9.83538106e-02 7.40260065e-01
-1.47113216e+00 -2.08167338e+00 2.59993702e-01 9.70958710e-01
8.31775665e-01 -3.92897308e-01 1.14220369e+00 8.92216086e-01
-1.10081755e-01 4.71497327e-01 1.73352793e-01 3.39792579e-01
9.96960878e-01 -1.17187428e+00 7.82509267e-01 6.15046501e-01
4.27425176e-01 -3.98965329e-02 8.42854857e-01 -4.77027267e-01
-1.17347348e+00 -1.11945009e+00 1.65021503e+00 -3.45250607e-01
1.23741770e+00 -4.43973690e-01 -1.51736081e-01 9.71106529e-01
1.37280798e+00 -2.80848622e-01 2.44571239e-01 -2.28275999e-01
-3.91565651e-01 4.84126732e-02 -8.86638880e-01 5.26466429e-01
7.66485512e-01 -9.41973448e-01 -5.30969918e-01 7.97134101e-01
7.39534080e-01 -5.61649621e-01 -9.55990732e-01 1.42226130e-01
6.30855024e-01 -1.64477378e-01 1.05308914e+00 -7.76909411e-01
8.77533853e-01 -2.14192688e-01 -7.63771951e-01 -1.40844083e+00
-3.33235770e-01 -4.97518897e-01 3.83950889e-01 1.13301492e+00
1.07440782e+00 -5.91431521e-02 2.02536836e-01 -3.38960276e-03
-2.81122893e-01 4.77556325e-03 -1.48644078e+00 -8.72172415e-01
4.42967117e-01 -5.11586428e-01 3.14917006e-02 6.25105560e-01
3.15462574e-02 9.92006898e-01 -9.53746080e-01 -4.02510971e-01
5.72515368e-01 -1.61073059e-01 6.49262488e-01 -4.91539627e-01
-2.48656571e-01 -2.12498214e-02 -3.08376253e-01 -8.29641998e-01
3.79036546e-01 -1.34721315e+00 2.67342031e-01 -1.91455817e+00
4.98680621e-01 4.58702058e-01 -2.10878238e-01 8.66669774e-01
1.54914469e-01 9.53961074e-01 6.59802437e-01 1.59766838e-01
-1.03211725e+00 4.64487672e-01 1.10697627e+00 -7.26671934e-01
5.55480540e-01 -4.91912842e-01 -1.15715787e-01 1.23826638e-01
6.18230224e-01 -3.96696836e-01 -3.32477212e-01 -1.00428200e+00
1.74410999e-01 4.21777397e-01 1.83044434e-01 -9.51832533e-01
-1.46631181e-01 3.18813264e-01 3.01286578e-01 -1.06044388e+00
6.66766167e-01 -5.94976723e-01 3.18609774e-01 4.50867206e-01
-4.14220005e-01 7.41384089e-01 5.58447599e-01 -4.17287380e-01
-6.64967537e-01 1.19513713e-01 1.03757370e+00 -2.69431412e-01
-9.43990827e-01 1.53840408e-01 -7.66886413e-01 6.39316201e-01
5.62132657e-01 4.25282896e-01 -6.19364977e-01 -7.57737994e-01
-8.93733978e-01 2.11070389e-01 1.50915921e-01 8.03192496e-01
4.40266073e-01 -1.51816261e+00 -1.40815973e+00 -1.31509200e-01
3.44335914e-01 -1.24750721e+00 7.72006661e-02 1.08580923e+00
-6.16628647e-01 8.27984571e-01 -6.73524976e-01 -4.62928593e-01
-1.10829937e+00 4.02192116e-01 -1.73112713e-02 -6.31994367e-01
-4.45010543e-01 7.45676041e-01 1.31436903e-02 -8.32685828e-01
2.91859806e-02 3.89971375e-01 1.79397594e-02 1.66205242e-02
4.25853133e-01 1.22032933e-01 5.10147586e-02 -1.09617281e+00
-6.69611096e-01 3.02423745e-01 -3.67160290e-01 -9.75920320e-01
1.13597691e+00 -2.46486753e-01 -2.10026413e-01 6.06831849e-01
1.66527677e+00 -4.54065263e-01 -3.23699683e-01 1.47879288e-01
-1.28477678e-01 3.57239991e-01 -1.13439910e-01 -1.65834260e+00
-8.15339029e-01 1.16453815e+00 8.05409491e-01 -2.47885555e-01
7.68475771e-01 5.09865955e-02 1.15468872e+00 3.68447989e-01
5.14680088e-01 -1.44956350e+00 -4.92120951e-01 1.12401342e+00
1.09930110e+00 -1.30401313e+00 -2.93444425e-01 2.48565897e-01
-1.21292937e+00 8.95528197e-01 4.04796600e-01 4.96306479e-01
3.64230037e-01 5.61671972e-01 6.85755074e-01 -8.80409032e-02
-5.52093089e-01 -2.77126759e-01 4.25955981e-01 5.75512767e-01
7.23758996e-01 2.79068768e-01 -5.66844225e-01 1.11522421e-01
-3.17016661e-01 1.00111775e-01 1.03130087e-01 6.67696595e-01
1.21078290e-01 -1.18210340e+00 3.53496112e-02 -2.27491409e-01
-4.11296904e-01 -6.93902671e-01 -6.57234490e-01 1.17889059e+00
-1.43920481e-01 9.16084409e-01 -3.66436958e-01 -4.51138139e-01
5.60072541e-01 1.27981022e-01 6.92313850e-01 -1.53211609e-01
-7.30164468e-01 2.85414487e-01 6.75151110e-01 -5.83752334e-01
-1.37964860e-01 -5.27997255e-01 -9.51740384e-01 -1.69162959e-01
3.37761730e-01 4.69766915e-01 8.84510458e-01 8.15815270e-01
-9.96776149e-02 2.79152513e-01 9.73817892e-03 -5.42224586e-01
-1.76857814e-01 -1.15055454e+00 -3.36786300e-01 1.59218386e-01
-2.39824504e-01 8.79116580e-02 3.00598770e-01 1.54797658e-01] | [11.438986778259277, 1.5291764736175537] |
6f606846-c92d-4ccb-850f-76fb54aaab76 | efficient-belief-space-planning-in-high | 2112.14428 | null | https://arxiv.org/abs/2112.14428v1 | https://arxiv.org/pdf/2112.14428v1.pdf | Efficient Belief Space Planning in High-Dimensional State Spaces using PIVOT: Predictive Incremental Variable Ordering Tactic | In this work, we examine the problem of online decision making under uncertainty, which we formulate as planning in the belief space. Maintaining beliefs (i.e., distributions) over high-dimensional states (e.g., entire trajectories) was not only shown to significantly improve accuracy, but also allows planning with information-theoretic objectives, as required for the tasks of active SLAM and information gathering. Nonetheless, planning under this "smoothing" paradigm holds a high computational complexity, which makes it challenging for online solution. Thus, we suggest the following idea: before planning, perform a standalone state variable reordering procedure on the initial belief, and "push forwards" all the predicted loop closing variables. Since the initial variable order determines which subset of them would be affected by incoming updates, such reordering allows us to minimize the total number of affected variables, and reduce the computational complexity of candidate evaluation during planning. We call this approach PIVOT: Predictive Incremental Variable Ordering Tactic. Applying this tactic can also improve the state inference efficiency; if we maintain the PIVOT order after the planning session, then we should similarly reduce the cost of loop closures, when they actually occur. To demonstrate its effectiveness, we applied PIVOT in a realistic active SLAM simulation, where we managed to significantly reduce the computation time of both the planning and inference sessions. The approach is applicable to general distributions, and induces no loss in accuracy. | ['Vadim Indelman', 'Khen Elimelech'] | 2021-12-29 | null | null | null | null | ['decision-making-under-uncertainty', 'decision-making-under-uncertainty'] | ['medical', 'reasoning'] | [ 2.13863924e-01 4.49582219e-01 -2.24029481e-01 -1.87889993e-01
-6.22660697e-01 -5.94400227e-01 5.61644375e-01 7.82822430e-01
-6.96081102e-01 1.11592197e+00 6.91235736e-02 -3.98069322e-01
-4.16970342e-01 -1.04424894e+00 -7.74833560e-01 -6.99620485e-01
-4.68714595e-01 9.68724728e-01 4.74169284e-01 -7.89935514e-02
4.47642475e-01 6.93562865e-01 -1.61078143e+00 -4.14182991e-01
8.53419662e-01 8.64972949e-01 3.51571858e-01 5.58599472e-01
-4.52101380e-02 7.19262600e-01 -4.40249205e-01 1.89599041e-02
2.40150034e-01 -2.26102084e-01 -1.00894499e+00 1.08341850e-01
-1.58873588e-01 -4.64797765e-01 -1.01396434e-01 9.85362411e-01
2.46224031e-01 6.37031853e-01 2.64220804e-01 -1.14371610e+00
5.37543178e-01 5.66706479e-01 -2.36178830e-01 4.04081047e-02
4.78977382e-01 6.04544617e-02 8.75359118e-01 -3.15876454e-01
7.18332410e-01 1.05124915e+00 3.44887018e-01 1.02578811e-01
-1.39713430e+00 -2.16433182e-01 3.53450090e-01 1.14648044e-01
-1.38602889e+00 -5.62573373e-01 4.33784217e-01 -3.09048772e-01
7.51739085e-01 3.77375960e-01 9.20696616e-01 2.14294374e-01
5.14432430e-01 5.95765948e-01 8.34374249e-01 -3.49418014e-01
8.08831215e-01 9.35079232e-02 5.52760474e-02 8.30224216e-01
3.18698704e-01 3.02815467e-01 -7.07746029e-01 -3.00503641e-01
3.29941154e-01 -2.38907546e-01 -2.22584561e-01 -6.72452629e-01
-1.22729790e+00 8.73834312e-01 2.51984894e-01 -1.68558452e-02
-4.91011739e-01 2.79433101e-01 2.09873140e-01 3.36036563e-01
3.61749440e-01 6.24381006e-01 -2.69044787e-01 -5.01965940e-01
-1.00149405e+00 4.41273540e-01 9.07361150e-01 7.60323822e-01
1.22971058e+00 -4.01620537e-01 1.68650523e-01 2.23401040e-01
2.55402356e-01 4.02146071e-01 -1.12015389e-01 -1.09041595e+00
4.75324631e-01 4.86654550e-01 7.01639831e-01 -1.00226116e+00
-7.61684060e-01 -1.87094167e-01 -4.73953515e-01 3.57187837e-01
4.60529923e-01 -3.72782052e-01 -1.02440989e+00 1.71841526e+00
6.03163123e-01 -1.26579657e-01 8.01393166e-02 7.69892931e-01
-3.01099420e-01 6.73049688e-01 -2.80799240e-01 -6.48438871e-01
9.26649928e-01 -3.31866115e-01 -7.83749104e-01 -4.78482306e-01
7.02219546e-01 -5.68606734e-01 5.15441895e-01 3.72218579e-01
-1.03074741e+00 -1.06773354e-01 -1.08415461e+00 3.33668232e-01
-7.40780532e-02 -4.44257498e-01 8.42837632e-01 3.45483750e-01
-9.04056549e-01 7.48959541e-01 -1.40354335e+00 -2.58736193e-01
-8.28841403e-02 5.42301774e-01 -1.30204141e-01 -1.99881103e-02
-1.08451831e+00 1.31274188e+00 6.80428445e-01 1.62780032e-01
-9.43822622e-01 -2.87237495e-01 -9.14435446e-01 1.31433845e-01
1.03536820e+00 -4.16125804e-01 1.24468470e+00 -4.13402945e-01
-1.42526031e+00 6.85648099e-02 -5.79181194e-01 -6.65370762e-01
4.83288199e-01 -1.37285218e-01 1.51436374e-01 -1.56186074e-01
-3.99030326e-03 5.73593020e-01 5.23607016e-01 -1.32508457e+00
-7.95140326e-01 -3.87583017e-01 3.80900502e-01 5.86659312e-01
2.34138861e-01 -4.85532880e-01 -3.86837453e-01 3.14149022e-01
8.29053104e-01 -1.33428180e+00 -6.99483812e-01 -2.92893946e-01
-2.70193994e-01 1.16510957e-01 3.06291342e-01 -3.42741013e-01
1.21132219e+00 -2.06326699e+00 3.89232725e-01 7.71647274e-01
-2.53805853e-02 -3.53747964e-01 3.16885352e-01 7.14115322e-01
5.26534081e-01 -2.02214211e-01 -4.21537876e-01 -3.69819582e-01
-1.40791796e-02 6.18460238e-01 -4.29009259e-01 6.61704421e-01
-2.54155397e-01 4.96036500e-01 -1.02528918e+00 -5.36128521e-01
3.74177992e-01 -1.71981514e-01 -7.36811340e-01 -7.46343583e-02
-5.63502848e-01 4.44467068e-01 -5.89642942e-01 3.04830074e-01
4.42213863e-01 -4.78281043e-02 4.67551768e-01 1.75259069e-01
-5.10255933e-01 4.89220083e-01 -1.51794243e+00 1.53591406e+00
-5.72125256e-01 5.15110314e-01 1.85365006e-01 -7.01101661e-01
7.79992580e-01 -4.81715351e-02 7.46306837e-01 -4.49547917e-01
4.45473902e-02 1.92146644e-01 -1.48260314e-02 -1.58601537e-01
9.24030185e-01 -1.66271344e-01 -2.49541923e-01 2.48628557e-01
-2.98582971e-01 -5.38355947e-01 3.55949283e-01 3.16162169e-01
1.06187832e+00 1.21063180e-01 5.50015032e-01 -2.93467700e-01
2.85236895e-01 4.97793645e-01 6.69936895e-01 8.02948713e-01
-4.56201173e-02 -3.04387975e-02 6.50993526e-01 -3.12970430e-01
-6.07080817e-01 -8.18869174e-01 1.88922063e-02 6.98281348e-01
8.12206626e-01 -5.15539646e-01 -3.13001305e-01 -3.73822123e-01
1.55329704e-01 9.88413393e-01 -4.76562172e-01 -9.97867063e-03
-4.01466727e-01 -5.87785304e-01 3.48981954e-02 4.29486260e-02
3.53592932e-01 -6.46708548e-01 -1.21382821e+00 4.85172302e-01
-2.36300975e-01 -6.34020865e-01 5.70537187e-02 6.43820822e-01
-8.14119637e-01 -1.06334400e+00 -1.21394182e-02 -8.58911872e-02
7.67782390e-01 -1.78384092e-02 8.12130690e-01 5.43590300e-02
2.83677816e-01 3.29517543e-01 -3.08606327e-01 -3.53785872e-01
-2.82370538e-01 -6.78887069e-02 4.44891155e-02 -2.54741102e-01
-2.10385591e-01 -3.82587820e-01 -1.76898405e-01 1.56231403e-01
-6.16813779e-01 1.60390019e-01 4.17605996e-01 7.42803335e-01
7.18809128e-01 4.67083752e-01 6.05489202e-02 -8.63486707e-01
5.40018022e-01 -2.50660270e-01 -1.16059506e+00 1.90037355e-01
-6.98063791e-01 5.02995193e-01 3.15496147e-01 -8.07397962e-02
-1.01521432e+00 2.45286554e-01 -1.97630841e-02 -1.87172871e-02
1.74448580e-01 1.07132256e+00 2.70255636e-02 -2.15956464e-01
3.68839145e-01 2.55971074e-01 5.72017618e-02 -8.19941908e-02
1.47359669e-01 1.94211230e-01 2.78152943e-01 -5.99232733e-01
6.62651658e-01 4.97753948e-01 4.36216265e-01 -6.51680768e-01
-6.37980282e-01 -3.20046037e-01 -5.01961946e-01 -3.66500914e-01
5.03095865e-01 -5.17128229e-01 -1.11058795e+00 6.60057738e-02
-1.04909050e+00 -5.52480459e-01 -3.47980410e-01 6.06743217e-01
-8.16188633e-01 3.27038497e-01 -7.39189982e-02 -1.28411603e+00
2.61495143e-01 -1.26947677e+00 7.83956468e-01 4.45739925e-02
-3.30341578e-01 -8.79519284e-01 1.48369998e-01 -1.08699761e-01
2.55321205e-01 3.56194705e-01 7.16568649e-01 -4.62855697e-01
-1.05412447e+00 -1.28066108e-01 2.57444173e-01 -3.70746762e-01
-1.68689072e-01 -2.53049284e-01 -4.62076217e-01 -4.86769378e-01
-5.62332869e-02 -6.05763979e-02 6.21882319e-01 2.06156030e-01
6.55103505e-01 -2.00397357e-01 -6.12141609e-01 2.47063056e-01
1.35527110e+00 4.60473895e-01 4.43183005e-01 4.56463665e-01
1.41888589e-01 7.14339495e-01 1.26316798e+00 7.93564618e-01
4.57452506e-01 7.87205637e-01 6.41750813e-01 3.50144207e-01
3.28823119e-01 -3.00639927e-01 2.97671556e-01 4.42810684e-01
-5.51485009e-02 -3.57967705e-01 -1.16911960e+00 5.92557609e-01
-2.09946609e+00 -7.89473891e-01 1.88903078e-01 2.54248500e+00
6.30926490e-01 4.15907115e-01 -2.42914110e-01 1.14329003e-01
4.67234701e-01 2.27968946e-01 -5.20141840e-01 -2.77347594e-01
4.03190553e-01 -8.66885334e-02 9.99478579e-01 1.22398686e+00
-8.91391516e-01 8.53442609e-01 5.73679781e+00 4.89858329e-01
-1.04169416e+00 3.80410440e-02 5.05758859e-02 -2.07556829e-01
-3.67275417e-01 5.67407846e-01 -7.57940352e-01 4.46336091e-01
9.52728033e-01 -2.73376107e-01 7.25086927e-01 8.23895812e-01
2.28635088e-01 -1.00225008e+00 -1.15538347e+00 5.00806630e-01
-2.73388445e-01 -1.31639099e+00 -3.82484168e-01 3.20506036e-01
4.91201192e-01 2.17979364e-02 -5.10253489e-01 2.00910076e-01
5.17846763e-01 -5.74319303e-01 8.59509885e-01 3.83610219e-01
3.16607207e-01 -1.00462830e+00 8.63517284e-01 8.95461321e-01
-1.05642366e+00 -1.22912154e-01 -2.41121903e-01 -3.44138294e-01
7.13855147e-01 7.71038294e-01 -1.14622951e+00 7.52603889e-01
1.84974030e-01 1.39024064e-01 -6.69874251e-02 1.11638248e+00
-2.32761145e-01 1.91210866e-01 -8.72088253e-01 -2.87661344e-01
2.62543350e-01 -2.86163747e-01 7.39698291e-01 6.06845260e-01
2.83966631e-01 1.61986038e-01 6.05596364e-01 4.53826070e-01
5.43750405e-01 -2.71804810e-01 -4.07282323e-01 1.89659700e-01
8.36584866e-01 6.55258298e-01 -9.85362530e-01 -3.42997104e-01
5.79457134e-02 6.08097076e-01 2.48944834e-01 1.47989854e-01
-7.73170233e-01 -1.40582129e-01 4.92270380e-01 1.01930700e-01
9.62275565e-02 -6.79464936e-01 -2.97943681e-01 -8.66238117e-01
-1.66882679e-01 -5.24042249e-01 2.02862158e-01 -3.37553352e-01
-3.00935388e-01 3.68332475e-01 4.36556399e-01 -9.72424328e-01
-6.27244234e-01 -2.01095194e-02 -2.38749012e-01 6.47083580e-01
-1.28301775e+00 -5.12053668e-01 -2.08621830e-01 4.02731299e-01
3.37782353e-01 5.29782712e-01 6.97012007e-01 5.04106469e-02
-1.83537170e-01 -6.61792932e-03 5.13627157e-02 -4.95884240e-01
2.59258062e-01 -1.09870017e+00 6.57546371e-02 1.02557600e+00
-3.42570618e-02 7.50358939e-01 1.22123098e+00 -1.06999433e+00
-1.67759192e+00 -7.07356393e-01 8.12595546e-01 -1.25710368e-01
5.66570222e-01 -1.94255501e-01 -6.47874296e-01 8.17519486e-01
-2.83452690e-01 -4.80264813e-01 1.08154394e-01 5.21303236e-01
5.06288707e-01 -6.60770293e-03 -1.05556035e+00 5.98313749e-01
7.97234058e-01 -1.55901298e-01 -5.24206936e-01 3.63057315e-01
4.41828191e-01 -7.43867815e-01 -6.63242519e-01 6.27300620e-01
3.71700853e-01 -8.35535705e-01 4.88275290e-01 -7.19128698e-02
-2.24254221e-01 -6.35608494e-01 -1.95436522e-01 -1.35791957e+00
-5.31229302e-02 -7.61454105e-01 -3.35617661e-01 8.68502796e-01
5.34840465e-01 -8.50912452e-01 8.79458070e-01 8.15165162e-01
-2.61073291e-01 -8.17145109e-01 -1.30070400e+00 -6.35273874e-01
-5.61303258e-01 -4.66486007e-01 4.96371448e-01 5.16800880e-01
2.32524738e-01 1.29460052e-01 -3.23155046e-01 5.39400101e-01
5.70065916e-01 3.44344467e-01 1.03035891e+00 -1.08748794e+00
-3.70422453e-01 4.29553464e-02 -3.46351057e-01 -1.26903141e+00
4.83085820e-03 -4.43081617e-01 6.39913976e-01 -1.57157481e+00
-1.24897085e-01 -9.35266316e-01 4.13898677e-02 4.42757338e-01
1.22501887e-02 -5.56108832e-01 3.99962157e-01 3.12664807e-01
-5.90332627e-01 4.92860675e-01 1.02873576e+00 1.10739611e-01
-5.05900621e-01 3.08819443e-01 -1.13401003e-01 6.74831629e-01
5.90058506e-01 -4.61044103e-01 -5.51861525e-01 -2.78066784e-01
3.84885013e-01 7.30844438e-01 7.89097995e-02 -1.07244039e+00
5.09516776e-01 -6.37430191e-01 -8.25283453e-02 -9.16501820e-01
7.94576824e-01 -1.00296938e+00 5.12523293e-01 8.14330101e-01
-2.39996120e-01 -7.68668354e-02 1.28598109e-01 6.55174732e-01
-1.80442199e-01 -4.52132761e-01 6.67147100e-01 -2.15104558e-02
-9.11658347e-01 5.70784397e-02 -5.79855382e-01 -4.49099928e-01
1.12159514e+00 -1.49065793e-01 -4.00069468e-02 -5.37249386e-01
-8.27804983e-01 5.78998923e-01 7.24147439e-01 -1.77439615e-01
3.97328347e-01 -7.36253619e-01 -1.04753338e-01 9.35666040e-02
-1.87464148e-01 4.04196918e-01 2.14708224e-02 9.77913260e-01
-5.20246744e-01 4.31110352e-01 3.55039574e-02 -6.43243730e-01
-1.01298296e+00 5.10371208e-01 1.94946349e-01 -5.08122087e-01
-5.05138516e-01 7.29410052e-01 -2.21505210e-01 -1.83287904e-01
3.01432639e-01 -3.18635970e-01 5.76737896e-02 2.14312851e-01
2.08621964e-01 4.10946667e-01 8.09962153e-02 -8.19942728e-02
-5.54761887e-01 2.39559352e-01 8.32633674e-02 -5.58510482e-01
1.12423456e+00 -4.77815717e-01 -6.83878362e-02 5.54744899e-01
4.66567993e-01 3.79032701e-01 -1.32770312e+00 -2.91854199e-02
3.71886641e-01 -6.52258456e-01 1.96947709e-01 -5.74375987e-01
-4.83939767e-01 4.44229752e-01 2.25286543e-01 3.87014896e-01
8.48649681e-01 4.29327786e-02 4.52374995e-01 8.24743509e-01
1.26025796e+00 -1.07135391e+00 -5.05743980e-01 8.37880671e-01
4.63888496e-01 -1.00956595e+00 4.75557625e-01 -4.88497227e-01
-5.75529337e-01 8.31346810e-01 4.98618662e-01 1.17232226e-01
2.45508447e-01 3.07013929e-01 -2.77909189e-01 -1.87691122e-01
-8.99821341e-01 -4.94170666e-01 -2.49544293e-01 1.07624635e-01
-1.64852664e-01 2.81281948e-01 -4.13250744e-01 -1.77193165e-01
-4.16286975e-01 -8.94291550e-02 7.52362549e-01 1.28867817e+00
-1.05108047e+00 -9.00841892e-01 -5.48889220e-01 4.82041657e-01
3.13080996e-02 2.12173179e-01 -8.62042755e-02 9.09213901e-01
7.78399929e-02 1.00925469e+00 2.30216816e-01 -3.64444137e-01
2.43171856e-01 -5.80548458e-02 4.83764708e-01 -6.74083591e-01
-1.41429722e-01 -6.48650154e-02 5.24041355e-01 -7.40099311e-01
-2.09778607e-01 -8.44287813e-01 -1.59158373e+00 -4.68804032e-01
-6.06093884e-01 6.16709292e-01 8.60170126e-01 1.13735938e+00
2.69640833e-01 4.47411269e-01 4.45697755e-01 -9.35502470e-01
-7.24956095e-01 -7.13667512e-01 -3.72267187e-01 -7.84639716e-02
4.93713021e-01 -1.14083028e+00 -3.27463239e-01 -3.86913896e-01] | [4.804194450378418, 1.8388780355453491] |
a423c621-6c29-4ced-97a6-4a6d8964292c | covi-agentsim-an-agent-based-model-for | 2010.16004 | null | https://arxiv.org/abs/2010.16004v1 | https://arxiv.org/pdf/2010.16004v1.pdf | COVI-AgentSim: an Agent-based Model for Evaluating Methods of Digital Contact Tracing | The rapid global spread of COVID-19 has led to an unprecedented demand for effective methods to mitigate the spread of the disease, and various digital contact tracing (DCT) methods have emerged as a component of the solution. In order to make informed public health choices, there is a need for tools which allow evaluation and comparison of DCT methods. We introduce an agent-based compartmental simulator we call COVI-AgentSim, integrating detailed consideration of virology, disease progression, social contact networks, and mobility patterns, based on parameters derived from empirical research. We verify by comparing to real data that COVI-AgentSim is able to reproduce realistic COVID-19 spread dynamics, and perform a sensitivity analysis to verify that the relative performance of contact tracing methods are consistent across a range of settings. We use COVI-AgentSim to perform cost-benefit analyses comparing no DCT to: 1) standard binary contact tracing (BCT) that assigns binary recommendations based on binary test results; and 2) a rule-based method for feature-based contact tracing (FCT) that assigns a graded level of recommendation based on diverse individual features. We find all DCT methods consistently reduce the spread of the disease, and that the advantage of FCT over BCT is maintained over a wide range of adoption rates. Feature-based methods of contact tracing avert more disability-adjusted life years (DALYs) per socioeconomic cost (measured by productive hours lost). Our results suggest any DCT method can help save lives, support re-opening of economies, and prevent second-wave outbreaks, and that FCT methods are a promising direction for enriching BCT using self-reported symptoms, yielding earlier warning signals and a significantly reduced spread of the virus per socioeconomic cost. | ['Yoshua Bengio', 'Eilif B. Muller', 'Joanna Merckx', 'Christopher Pal', 'Irina Rish', 'Jian Tang', 'Bernhard Schölkopf', 'Yang Zhang', 'Joumana Ghosn', 'David Buckeridge', 'Marc-Andre Rousseau', 'Satya Ortiz-Gagné', 'Pierre Luc Carrier', 'Gaétan Marceau Caron', 'Olexa Bilaniuk', 'Meng Qu', 'Akshay Patel', 'Andrew Williams', 'Tristan Deleu', 'Pierre-Luc St. Charles', 'Victor Schmidt', 'Soren Harnois-Leblanc', 'Nanor Minoyan', 'Abhinav Sharma', 'Hannah Alsdurf', 'Nasim Rahaman', 'Martin Weiss', 'Tegan Maharaj', 'Prateek Gupta'] | 2020-10-30 | null | https://openreview.net/forum?id=07iDTU-KFK | https://openreview.net/pdf?id=07iDTU-KFK | null | ['virology'] | ['miscellaneous'] | [ 2.08499402e-01 -3.21349889e-01 -2.14134455e-01 7.49932528e-02
-4.92632762e-02 -6.21659994e-01 1.00863874e+00 7.24472582e-01
-5.32596469e-01 7.61532724e-01 1.40964925e-01 -7.47641265e-01
-7.63374031e-01 -1.07176018e+00 -2.75519252e-01 -3.45671535e-01
-5.76161504e-01 7.62686133e-01 2.78149903e-01 -2.56133944e-01
-1.00206867e-01 6.80886149e-01 -9.71009851e-01 -1.46488309e-01
9.31309879e-01 2.99201220e-01 1.41945139e-01 5.45243919e-01
2.21740380e-01 1.12829730e-01 -8.20907831e-01 -3.64224225e-01
1.68795943e-01 -4.51492578e-01 -2.38500625e-01 -5.38031757e-01
-5.67906797e-01 -7.49735892e-01 -5.02085574e-02 4.21069771e-01
3.05783659e-01 -3.93014759e-01 1.06236982e+00 -1.53596044e+00
-1.14903390e-01 6.44846261e-02 -4.14810658e-01 1.24603152e-01
5.09732723e-01 3.42089832e-01 2.64732093e-01 -9.82151851e-02
1.07428336e+00 1.31025493e+00 1.17874599e+00 2.81338006e-01
-1.64228868e+00 -5.50095260e-01 -1.46431103e-01 -2.36336052e-01
-1.17867255e+00 -8.10890570e-02 1.56366989e-01 -6.75780356e-01
1.07140672e+00 5.09321094e-01 9.29252028e-01 1.07266176e+00
5.52976429e-01 -8.45278203e-02 1.13470852e+00 -7.92612135e-03
1.90543830e-01 1.87233821e-01 1.23717720e-02 6.51290238e-01
1.06863439e+00 2.72901803e-01 -1.58456042e-02 -1.20639908e+00
7.24916220e-01 5.52203178e-01 -2.02971846e-01 -5.12727313e-02
-1.08794200e+00 1.02935672e+00 4.68691923e-02 1.75619870e-01
-5.53097308e-01 7.04440996e-02 2.39616632e-01 2.37687796e-01
7.27581799e-01 3.93887788e-01 -3.90987635e-01 1.34174228e-01
-5.64195395e-01 5.60440481e-01 8.00820470e-01 2.09838688e-01
5.63881099e-01 -2.33340427e-01 -2.94488907e-01 6.50547981e-01
2.95932293e-01 1.22415054e+00 -9.58369002e-02 -8.71971250e-01
5.42898513e-02 8.04056585e-01 3.32155377e-01 -1.17719555e+00
-7.44726598e-01 -2.98197150e-01 -7.00268269e-01 7.78663307e-02
1.58638060e-01 -7.08734572e-01 -6.28609776e-01 1.76214957e+00
5.10655701e-01 1.26187414e-01 -3.16887915e-01 2.69182384e-01
5.94462119e-02 7.94469237e-01 1.85524642e-01 -4.62336034e-01
1.08093619e+00 -4.71589752e-02 -4.28970128e-01 1.74449816e-01
7.85787344e-01 -2.80479670e-01 4.15694028e-01 -9.42175984e-02
-5.46360731e-01 5.22992127e-02 -8.46828401e-01 1.07059681e+00
-6.36979043e-01 -5.04596293e-01 6.33737803e-01 9.30834711e-01
-1.18985689e+00 5.90244174e-01 -9.85445023e-01 -9.91512239e-01
3.31329346e-01 4.13473368e-01 -1.44896105e-01 -1.22145191e-02
-1.03053725e+00 8.50675821e-01 -3.88540894e-01 -2.22069323e-01
-6.87317431e-01 -9.96725023e-01 -5.16953826e-01 -1.71945244e-01
8.28024670e-02 -9.25621331e-01 5.42241454e-01 -4.02625680e-01
-7.97115505e-01 3.49892169e-01 -2.72790566e-02 -4.86766785e-01
7.03463376e-01 2.78592676e-01 -4.36322123e-01 2.33422861e-01
2.60192573e-01 4.01891738e-01 3.29208821e-01 -1.16924310e+00
-5.95723867e-01 -4.76606965e-01 -4.01930995e-02 1.72148906e-02
-3.05970758e-01 3.33119690e-01 1.33023307e-01 -5.61731517e-01
-8.20765674e-01 -1.07263708e+00 -3.38243335e-01 -9.04151425e-02
-1.90558717e-01 -1.05039760e-01 9.87759054e-01 -7.00412989e-01
9.44012940e-01 -1.58355069e+00 -2.34214544e-01 3.89877439e-01
3.12565148e-01 6.54006779e-01 -2.37877771e-01 1.07030106e+00
7.63726354e-01 4.12625492e-01 -4.55126464e-01 3.85054536e-02
-2.79906720e-01 -6.87288027e-03 -5.52131087e-02 5.50298095e-01
3.53096366e-01 5.74997008e-01 -9.76801991e-01 -1.84210137e-01
2.30683282e-01 9.02934551e-01 -6.75791442e-01 2.48113573e-01
-2.78201520e-01 4.45302308e-01 -5.56898415e-01 3.77786815e-01
6.50607646e-01 -4.61501420e-01 6.39527380e-01 4.86235976e-01
-5.49055412e-02 -8.15986916e-02 -5.67229450e-01 4.48037028e-01
-1.30664319e-01 3.40781569e-01 2.04865843e-01 -5.14164984e-01
6.01861119e-01 1.73786804e-01 7.07667530e-01 -3.79244179e-01
-2.22188663e-02 -1.03382252e-01 1.57664120e-01 -2.72183239e-01
-2.65717041e-03 1.95526659e-01 1.16635203e-01 1.03582406e+00
-4.63496625e-01 1.37867391e-01 1.21910863e-01 3.83578539e-01
1.52836907e+00 -1.97755769e-01 1.55372709e-01 -3.69308352e-01
-1.50841892e-01 4.35774118e-01 4.70363975e-01 8.52736533e-01
-1.67486623e-01 -7.95191750e-02 5.20295799e-01 -1.85672939e-01
-6.28548145e-01 -1.23792279e+00 -2.26877898e-01 7.58377016e-01
1.10859588e-01 -1.24359615e-01 -6.83430851e-01 -4.41556543e-01
3.88829082e-01 3.62860829e-01 -6.53623641e-01 -9.46942493e-02
-4.16076481e-01 -1.42705870e+00 6.64823353e-01 1.56813767e-02
4.15315032e-01 -8.37526500e-01 -7.57703841e-01 3.93101990e-01
3.59141789e-02 -6.07169628e-01 -2.56667256e-01 -2.55046546e-01
-7.18463600e-01 -1.36386144e+00 -9.43833113e-01 -2.59513319e-01
7.82033622e-01 3.54970723e-01 6.53374672e-01 6.41842186e-01
-4.24130529e-01 7.19309688e-01 -2.71060973e-01 -3.90953034e-01
-6.90168619e-01 -1.19314127e-01 4.40898508e-01 -2.44350448e-01
1.05974995e-01 -2.47360647e-01 -1.06063235e+00 5.84326923e-01
-6.92842543e-01 -2.20494062e-01 1.94913939e-01 2.79599190e-01
4.92558181e-02 -1.70404240e-01 1.06271672e+00 -1.05366457e+00
1.06609488e+00 -9.87980902e-01 -7.10731745e-01 3.29154074e-01
-1.02228975e+00 -4.70845252e-01 2.08433330e-01 -5.28989255e-01
-1.03474593e+00 -3.89838636e-01 1.69948548e-01 4.29680318e-01
1.15026243e-01 5.23239374e-01 4.81319100e-01 3.58871892e-02
4.98570561e-01 -1.21277384e-01 4.07447577e-01 -2.69812733e-01
-9.71320719e-02 7.50341237e-01 -7.85468817e-02 -2.44669154e-01
6.80089474e-01 6.54403150e-01 3.77081782e-02 -1.12897277e+00
1.77320048e-01 -1.70044437e-01 -3.84615660e-02 -2.98895985e-01
7.81690538e-01 -6.53922081e-01 -1.04685938e+00 6.94696069e-01
-9.65463281e-01 -8.70565712e-01 3.62049043e-01 4.39449549e-01
-1.20339446e-01 7.80349076e-02 -7.10666776e-01 -9.84045982e-01
-3.27720702e-01 -9.61925149e-01 6.16416514e-01 5.92613481e-02
-6.19353354e-01 -1.29522467e+00 7.80937195e-01 1.29301727e-01
8.46326768e-01 6.95416570e-01 1.09552228e+00 -7.48954952e-01
-3.99794161e-01 -2.35642135e-01 -3.54475945e-01 -1.70629203e-01
6.33749485e-01 2.53234565e-01 -4.72694844e-01 -6.11143708e-01
-3.01779985e-01 1.20424382e-01 4.75888431e-01 6.68748260e-01
2.38330781e-01 -5.51278949e-01 -1.10488212e+00 4.17656571e-01
1.32224631e+00 5.32415450e-01 2.93749511e-01 2.50176769e-02
1.99374542e-01 9.17075276e-01 3.02891910e-01 4.06976461e-01
5.50876260e-01 7.10525334e-01 2.00589493e-01 -4.02658880e-01
1.94271263e-02 -2.23124847e-01 1.74991742e-01 3.53599340e-01
-3.84491324e-01 -3.94057244e-01 -1.01433301e+00 5.36195397e-01
-1.44897699e+00 -9.68771100e-01 -1.05568320e-01 2.42579794e+00
5.66759229e-01 6.01720065e-02 7.45842695e-01 -2.14854598e-01
6.20975912e-01 -2.14242578e-01 -4.28568929e-01 -4.26946700e-01
1.61367461e-01 -5.34651689e-02 5.46817839e-01 4.98401463e-01
-5.16374111e-01 2.70401448e-01 7.36984348e+00 4.16023403e-01
-1.01211560e+00 4.36627641e-02 7.29377925e-01 2.18072727e-01
-3.37034136e-01 -2.36925721e-01 -6.31109893e-01 5.41846156e-01
1.15546846e+00 -4.44266558e-01 5.95031559e-01 8.93295333e-02
6.03513002e-01 -3.26440573e-01 -7.52573013e-01 1.68773070e-01
-2.88819134e-01 -1.61079597e+00 4.82340977e-02 4.69877154e-01
7.42752910e-01 2.63564855e-01 -7.55621195e-02 1.15235923e-02
5.88301480e-01 -6.79693878e-01 1.02334870e-02 4.76961225e-01
1.03763366e+00 -7.28510439e-01 7.95839369e-01 2.20855817e-01
-1.08156824e+00 2.45656565e-01 2.15953961e-02 -7.02594582e-04
3.21303725e-01 3.66975337e-01 -1.21489704e+00 2.63054967e-01
5.28957963e-01 2.57350206e-01 -3.69782448e-01 9.99672174e-01
4.06156033e-01 7.74774313e-01 -5.00426233e-01 -2.75786638e-01
-1.34647172e-02 -2.29430974e-01 5.49686790e-01 1.21307325e+00
2.46052593e-01 3.60069498e-02 -1.33710325e-01 7.43416071e-01
2.83445269e-01 -2.10610837e-01 -9.80679750e-01 -2.52250552e-01
7.73857832e-01 8.27597618e-01 -9.76393461e-01 -2.52187997e-01
-1.15525804e-01 7.19618082e-01 -2.18388155e-01 5.80605507e-01
-5.01545250e-01 -5.56335449e-01 9.39139545e-01 6.44565940e-01
7.65933469e-02 -8.36343020e-02 2.90418893e-01 -5.84194601e-01
-5.58025956e-01 -7.43865252e-01 3.85313779e-01 -4.23099756e-01
-1.02208722e+00 5.43080568e-01 5.52672148e-01 -6.88493848e-01
-5.37540615e-01 -3.20971936e-01 -7.85367370e-01 7.32500136e-01
-1.12465787e+00 -7.99767017e-01 5.07729761e-02 3.12782794e-01
8.57594423e-03 2.33459860e-01 8.18692565e-01 1.02861397e-01
-4.79321629e-01 3.53888154e-01 4.72649127e-01 -1.96320727e-01
3.33357960e-01 -5.92685282e-01 3.46144795e-01 2.95730859e-01
-5.59273243e-01 9.89021122e-01 6.48648083e-01 -1.44160354e+00
-1.33171809e+00 -1.33461976e+00 7.05730975e-01 -5.34924924e-01
5.91829896e-01 -4.59747285e-01 -5.65680742e-01 6.51721716e-01
7.10924491e-02 -6.65607274e-01 4.46324348e-01 -1.97075576e-01
-6.63395748e-02 -4.21634316e-02 -1.77349007e+00 6.63283527e-01
9.42861140e-01 -3.10939550e-01 -1.08502962e-01 4.30268168e-01
9.33360696e-01 4.72043216e-01 -8.12723935e-01 5.36470413e-01
6.89472973e-01 -8.48049283e-01 1.05320537e+00 -3.25432181e-01
-1.95995212e-01 1.63920131e-02 1.09785169e-01 -1.44759452e+00
-1.38853922e-01 -7.11178422e-01 4.09685224e-01 1.14824092e+00
4.75895524e-01 -1.41582716e+00 2.33266428e-01 3.08856577e-01
5.79466522e-01 -8.18511367e-01 -7.20403552e-01 -9.96771514e-01
-1.08182795e-01 1.14161372e-01 7.70628095e-01 9.72968698e-01
-6.51151389e-02 -9.49098617e-02 -3.00505728e-01 1.22328922e-01
6.68125212e-01 -2.34905899e-01 7.07402587e-01 -1.38552582e+00
-2.41429895e-01 -2.85692930e-01 -1.93581000e-01 -1.26068592e-01
-4.11209136e-01 -5.39537311e-01 -4.63562757e-01 -1.74018943e+00
5.34262657e-01 -7.92144001e-01 5.14176227e-02 3.74597639e-01
1.33211901e-02 2.68720120e-01 -3.16838361e-02 2.17237309e-01
-4.67790775e-02 1.88638747e-01 6.50893927e-01 -1.04892954e-01
-5.51010132e-01 9.49902087e-02 -3.92817438e-01 5.46351790e-01
8.27420592e-01 -6.43239498e-01 -5.97066402e-01 -1.56896606e-01
2.83276647e-01 2.28657231e-01 5.75725257e-01 -6.57755375e-01
-1.67296752e-01 -5.85103154e-01 2.07936019e-01 -5.29639602e-01
3.07209730e-01 -6.51498199e-01 6.94144607e-01 1.41009223e+00
1.44080162e-01 1.71361640e-01 3.02686870e-01 7.62716413e-01
7.35768259e-01 2.96784103e-01 2.58698106e-01 3.54328632e-01
3.57075810e-01 2.65305936e-01 -9.94709671e-01 -4.35623713e-02
1.06139648e+00 -1.06688380e-01 -1.01742804e+00 -4.54044372e-01
-1.98236614e-01 2.40111440e-01 7.25091577e-01 1.71707094e-01
3.03094417e-01 -9.22102869e-01 -8.86229932e-01 4.65699211e-02
-9.86563191e-02 -8.84423256e-01 5.00525944e-02 8.51885140e-01
-8.01320255e-01 5.12904167e-01 -2.31630445e-01 -3.89770210e-01
-1.20578086e+00 4.97868448e-01 3.60084981e-01 -1.91910937e-01
-4.06788856e-01 1.72790751e-01 2.12685257e-01 -5.22286355e-01
6.19872957e-02 9.01369303e-02 1.36193320e-01 -2.62291729e-02
7.30007470e-01 9.13613319e-01 -1.97859764e-01 -4.82223958e-01
-7.61500716e-01 3.64931613e-01 1.97353587e-01 -8.84348527e-02
1.57247710e+00 -5.53593300e-02 -1.27700970e-01 6.22836165e-02
7.48082340e-01 1.20262057e-01 -1.14944410e+00 2.95120448e-01
-8.20461586e-02 -1.70191512e-01 -2.97588468e-01 -1.10579586e+00
-6.42120600e-01 2.13145897e-01 8.46041441e-01 6.50685847e-01
7.59998977e-01 -1.07675083e-01 6.44762754e-01 2.35044003e-01
6.68817163e-01 -4.09738153e-01 -3.69068742e-01 -2.42868051e-01
7.05639184e-01 -8.25463831e-01 1.30722582e-01 -2.92617768e-01
-1.58437774e-01 1.79635242e-01 1.05688378e-01 -5.76170348e-02
8.17421436e-01 3.22589606e-01 1.81860738e-02 -3.36443216e-01
-7.54074812e-01 3.08619142e-02 -2.15833798e-01 1.26202822e+00
-5.66538312e-02 4.81121808e-01 -7.22677708e-01 2.92381257e-01
5.27480602e-01 3.09334812e-03 4.94784355e-01 8.11315060e-01
-3.16821903e-01 -1.22141111e+00 -4.98829693e-01 1.11376894e+00
-2.66010523e-01 1.07179530e-01 -7.02539444e-01 1.01437509e+00
6.18480220e-02 1.32094705e+00 1.87532768e-01 -3.11780572e-01
3.52346674e-02 -2.44906843e-01 2.12419569e-01 -3.98594916e-01
-7.72065341e-01 -2.25255296e-01 3.74042124e-01 -2.71836221e-01
-3.85800511e-01 -6.12066746e-01 -1.13785577e+00 -9.27563965e-01
-2.08897725e-01 1.26902899e-02 7.10358262e-01 7.05006540e-01
8.01866889e-01 3.13506760e-02 6.77782178e-01 -6.89323962e-01
-2.41706952e-01 -7.71326244e-01 -4.61239427e-01 1.17408685e-01
3.70714903e-01 -7.07769334e-01 -5.10165572e-01 -4.14527088e-01] | [5.984170913696289, 4.387811660766602] |
c7e4ed0c-b3c7-43be-969d-58bb0840148b | alarm-based-prescriptive-process-monitoring | 1803.08706 | null | http://arxiv.org/abs/1803.08706v2 | http://arxiv.org/pdf/1803.08706v2.pdf | Alarm-Based Prescriptive Process Monitoring | Predictive process monitoring is concerned with the analysis of events
produced during the execution of a process in order to predict the future state
of ongoing cases thereof. Existing techniques in this field are able to
predict, at each step of a case, the likelihood that the case will end up in an
undesired outcome. These techniques, however, do not take into account what
process workers may do with the generated predictions in order to decrease the
likelihood of undesired outcomes. This paper proposes a framework for
prescriptive process monitoring, which extends predictive process monitoring
approaches with the concepts of alarms, interventions, compensations, and
mitigation effects. The framework incorporates a parameterized cost model to
assess the cost-benefit tradeoffs of applying prescriptive process monitoring
in a given setting. The paper also outlines an approach to optimize the
generation of alarms given a dataset and a set of cost model parameters. The
proposed approach is empirically evaluated using a range of real-life event
logs. | ['Fabrizio Maria Maggi', 'Niek Tax', 'Irene Teinemaa', 'Massimiliano de Leoni', 'Marlon Dumas'] | 2018-03-23 | null | null | null | null | ['predictive-process-monitoring'] | ['time-series'] | [ 9.06853735e-01 3.69671047e-01 2.64584959e-01 -3.25155556e-01
-2.05434442e-01 -2.18485191e-01 8.50453436e-01 1.21817148e+00
-9.74919349e-02 5.91601253e-01 8.88438374e-02 -4.39707190e-01
-6.07355773e-01 -1.10954571e+00 -2.87855476e-01 -3.65801871e-01
3.36424410e-02 6.18683696e-01 1.84781820e-01 5.08154094e-01
4.83258456e-01 7.08658397e-01 -1.37353122e+00 2.58544952e-01
3.59552681e-01 9.08691227e-01 2.94408239e-02 7.01521933e-01
-1.09379090e-01 1.07126391e+00 -8.31813753e-01 -1.12695480e-02
2.32973054e-01 -3.92719984e-01 -3.98892164e-01 4.60353941e-01
-4.18233573e-01 -2.39181027e-01 3.01997513e-01 5.09701133e-01
1.36505608e-02 2.37456098e-01 8.33344698e-01 -1.35451221e+00
2.53725976e-01 3.53557557e-01 -1.52261361e-01 3.54977965e-01
3.90862674e-01 6.88973725e-01 5.85983276e-01 -2.17162967e-01
5.76979900e-03 1.25730968e+00 2.39774838e-01 -8.86090100e-03
-1.30823302e+00 -2.54660755e-01 3.53208154e-01 -1.48622975e-01
-8.27003121e-01 -2.29592443e-01 5.47396839e-01 -6.17076933e-01
1.22311497e+00 1.67582169e-01 5.45118213e-01 7.02725470e-01
6.92332923e-01 8.08101669e-02 9.74739194e-01 -4.86360788e-01
7.55224347e-01 1.48778465e-02 1.42031834e-01 -6.33323789e-02
7.17249334e-01 4.18017000e-01 -3.40064734e-01 -5.30271053e-01
5.83579004e-01 4.27188337e-01 2.55292714e-01 -1.72229558e-01
-7.34012187e-01 3.42871457e-01 -1.92024857e-01 1.74414858e-01
-1.04363072e+00 1.58349976e-01 3.06443870e-01 3.72607745e-02
2.36106113e-01 6.23758793e-01 -3.23424518e-01 -6.92403838e-02
-6.23912930e-01 2.74922937e-01 9.16944385e-01 6.69784248e-01
3.54277760e-01 1.46142505e-02 -7.92056620e-01 1.95895970e-01
4.10131067e-01 2.25400776e-02 1.64890841e-01 -8.24957967e-01
4.11399871e-01 9.28651690e-01 7.35134900e-01 -4.47856426e-01
-1.76813379e-01 -1.45094246e-01 -2.28601098e-01 4.37720478e-01
1.41952157e-01 -3.41849595e-01 -7.38976717e-01 1.12660432e+00
1.62941590e-01 1.33887738e-01 -7.49641359e-02 7.83801302e-02
-4.16728020e-01 7.72448421e-01 6.32475138e-01 -8.95543993e-01
9.36119676e-01 -3.07772905e-01 -8.05695236e-01 -1.55011296e-01
5.67707531e-02 -5.44403613e-01 5.34646749e-01 5.55587769e-01
-1.02247655e+00 -5.39441645e-01 -7.18894780e-01 1.01057529e+00
9.60111711e-03 -1.81113839e-01 2.67177910e-01 5.89253843e-01
-4.68731433e-01 8.58523548e-01 -1.14559698e+00 -5.01341224e-01
-9.96412151e-03 2.39863425e-01 3.89886349e-01 2.90771961e-01
-5.91005027e-01 9.92251277e-01 6.32771492e-01 1.13530904e-01
-1.41629624e+00 -5.60210645e-01 -4.01610732e-01 6.25346482e-01
6.38708293e-01 -4.50834006e-01 1.60554647e+00 -3.52970719e-01
-9.72722769e-01 -1.37132034e-01 -1.37278274e-01 -6.60354674e-01
8.07702422e-01 -3.46941292e-01 -5.99988163e-01 -1.51558280e-01
-1.63384065e-01 -9.32518691e-02 5.91538787e-01 -1.15582812e+00
-1.05649769e+00 -3.22559774e-01 -8.92080814e-02 1.46100044e-01
1.15697207e-02 -4.43693958e-02 3.58566612e-01 -1.19027495e-01
-1.83543250e-01 -6.80572808e-01 -7.65238762e-01 -5.18215299e-01
-4.22908634e-01 -3.08285560e-03 5.67581475e-01 -1.12691060e-01
1.25300717e+00 -1.67694700e+00 -5.41619122e-01 5.47768414e-01
-3.61863583e-01 1.16321944e-01 4.34186697e-01 9.96217132e-01
-9.28499624e-02 -1.59915921e-03 -1.96987316e-01 -7.67524764e-02
-2.98843980e-01 2.46566564e-01 -4.45327550e-01 -1.10303489e-02
4.99333292e-01 -4.32663113e-02 -7.57034838e-01 -7.79410973e-02
7.91586399e-01 9.28267241e-02 -1.04284376e-01 7.18421340e-01
-4.43839610e-01 4.13427353e-01 -6.35359764e-01 4.47832227e-01
1.22473560e-01 1.06997088e-01 4.87809688e-01 3.77150208e-01
-4.42739248e-01 2.48163372e-01 -1.23446751e+00 3.33030492e-01
-5.44652045e-01 -8.94746333e-02 -1.95022762e-01 -7.68896818e-01
9.35656130e-01 7.51772225e-01 8.02740097e-01 -1.44625485e-01
2.21880153e-02 -2.38907225e-02 -8.25079083e-02 -5.52358747e-01
2.93483824e-01 -2.76364565e-01 1.50348768e-01 6.60179317e-01
-3.72636259e-01 7.03914315e-02 3.73180181e-01 -2.90283769e-01
1.54139507e+00 6.76426739e-02 7.81700313e-01 2.94146873e-02
4.99443054e-01 -8.29889998e-02 6.53492868e-01 7.16053069e-01
-1.90285116e-01 -6.21448755e-02 5.28748751e-01 -4.83260572e-01
-7.96498597e-01 -9.41265702e-01 2.85395026e-01 6.82574749e-01
1.44192129e-02 -8.88351277e-02 -3.63942534e-01 -4.31075007e-01
-8.71228799e-03 1.35930884e+00 -3.87048155e-01 -2.21899971e-01
-2.24310189e-01 -9.87370551e-01 -7.21106604e-02 3.42985779e-01
2.87248045e-02 -1.30425191e+00 -1.39514256e+00 8.76510441e-01
3.79812539e-01 -6.59223855e-01 5.09025119e-02 3.63423467e-01
-1.28751802e+00 -1.47265518e+00 2.74887174e-01 6.53823987e-02
7.65507638e-01 -1.41365469e-01 8.13558757e-01 -2.07131073e-01
-2.99126357e-01 4.59593832e-01 -1.24618009e-01 -9.79798734e-01
-9.11215484e-01 -4.41242695e-01 -1.75489798e-01 1.24653287e-01
5.65536559e-01 -3.74565005e-01 -4.00941402e-01 1.53865397e-01
-9.76722002e-01 -3.59372467e-01 7.01207697e-01 3.73830646e-01
6.38090253e-01 5.91202617e-01 7.03860164e-01 -1.06039286e+00
1.10651219e+00 -4.80913550e-01 -5.60506284e-01 5.96733749e-01
-1.14544344e+00 1.72785640e-01 4.32971448e-01 -3.39006305e-01
-1.57751560e+00 3.27633679e-01 3.59240174e-01 -2.37291548e-02
-7.56356359e-01 3.79248172e-01 -2.18838319e-01 8.11076164e-01
3.34695250e-01 1.40331998e-01 -7.27840215e-02 -3.67449522e-01
-2.73728639e-01 2.86891341e-01 2.70090491e-01 -4.35446858e-01
4.27115560e-01 3.05463105e-01 3.05567265e-01 -3.00257564e-01
-5.17218471e-01 -6.72490716e-01 -3.42839390e-01 -6.51598334e-01
4.22949761e-01 -3.44902635e-01 -8.62336874e-01 6.38360158e-02
-9.29452062e-01 -5.49215302e-02 -5.39859474e-01 2.90557802e-01
-6.79460585e-01 -1.32937819e-01 -2.55403042e-01 -1.58548057e+00
-3.17784965e-01 -8.63468885e-01 5.19388437e-01 1.35745361e-01
-6.63087130e-01 -8.51697624e-01 1.54834017e-01 1.08120255e-01
3.41329008e-01 5.52247047e-01 8.50804865e-01 -1.01723552e+00
-5.92792928e-01 -6.03359342e-01 4.69599456e-01 3.08062494e-01
6.57651007e-01 3.11207294e-01 -6.38216913e-01 -1.30797446e-01
4.10725594e-01 4.74316061e-01 8.55576620e-02 3.51921886e-01
9.01595414e-01 -6.44163966e-01 -5.31251073e-01 -3.06390762e-01
1.43282974e+00 1.00523043e+00 4.96175051e-01 4.16519523e-01
1.47567675e-01 9.88401055e-01 1.20652425e+00 9.72201228e-01
-1.83264554e-01 2.51853585e-01 7.10502744e-01 3.23537678e-01
5.49415290e-01 -2.55525410e-01 3.23020458e-01 -1.88264832e-01
-1.67129308e-01 -4.98756856e-01 -9.55869377e-01 6.75974667e-01
-1.80414426e+00 -1.34192848e+00 -3.20972085e-01 2.42739439e+00
2.47256324e-01 6.23943806e-01 6.21922910e-02 5.01739025e-01
8.83996725e-01 -1.90565243e-01 -4.83903229e-01 -7.00801790e-01
5.37010133e-01 -4.12423834e-02 7.24441051e-01 4.00273800e-01
-8.55612755e-01 2.24361587e-02 6.50994778e+00 -8.48909393e-02
-5.88653982e-01 -2.26373464e-01 7.35613465e-01 -2.05480471e-01
-7.64785782e-02 2.82627344e-01 -9.10115540e-01 4.22341079e-01
1.56525147e+00 -5.12479424e-01 2.71863993e-02 7.31882095e-01
1.05013180e+00 -4.03623968e-01 -1.57374084e+00 8.51339027e-02
-5.34002542e-01 -9.75604594e-01 1.78189412e-01 2.12039560e-01
5.16060710e-01 -5.64206183e-01 -3.40956658e-01 -1.56750515e-01
5.35338223e-01 -8.40533078e-01 6.86071157e-01 9.59050000e-01
-1.20658912e-01 -9.52800632e-01 7.11135685e-01 5.56295514e-01
-9.16597426e-01 -6.14471495e-01 1.99935988e-01 -5.01330316e-01
5.71430504e-01 6.77001357e-01 -1.66001904e+00 3.84034812e-01
3.47539276e-01 4.54607531e-02 -2.43782371e-01 1.11994183e+00
-2.18769521e-01 7.07573950e-01 -8.16833749e-02 2.33141538e-02
-1.09910645e-01 -1.94640070e-01 5.98630130e-01 1.00954592e+00
4.59091276e-01 -2.37877563e-01 2.11242154e-01 7.84860790e-01
6.66566074e-01 -5.26948422e-02 -5.67837000e-01 6.72481663e-04
7.25585639e-01 7.84452319e-01 -6.98310316e-01 -4.19052124e-01
-7.97962695e-02 3.71046990e-01 -3.31470072e-01 1.79814488e-01
-3.58170152e-01 -4.98840818e-03 4.22471374e-01 6.45572782e-01
4.90451083e-02 -4.02859636e-02 -6.89311028e-01 -7.01911896e-02
-3.18449363e-02 -5.17005622e-01 5.35686791e-01 -5.37170589e-01
-1.29365528e+00 1.03140451e-01 4.01643425e-01 -1.04071128e+00
-5.70662856e-01 -3.45508963e-01 -1.06321859e+00 9.72570837e-01
-1.06805468e+00 -4.93694335e-01 -2.94415325e-01 3.40851918e-02
6.27700984e-01 1.32692143e-01 6.94403112e-01 -2.86362916e-01
-4.92058247e-01 -4.55392092e-01 5.94384223e-03 -5.29529095e-01
4.55896944e-01 -1.12503612e+00 1.52308401e-02 9.47335482e-01
-6.78313076e-01 4.79295492e-01 1.20557809e+00 -1.19724917e+00
-7.69846141e-01 -1.41415179e+00 8.91559660e-01 -2.99584597e-01
6.19489610e-01 1.47019133e-01 -8.79373848e-01 8.05806339e-01
2.03210637e-02 -6.78002477e-01 6.63966417e-01 -6.45089149e-02
3.94104540e-01 -2.97825217e-01 -1.57163239e+00 3.16049188e-01
5.78990579e-01 -8.67077261e-02 -7.05501199e-01 1.59195572e-01
2.87158251e-01 1.40637070e-01 -9.29187357e-01 4.39669609e-01
1.42924771e-01 -1.01755869e+00 5.77459157e-01 -4.75359648e-01
2.14293599e-01 -2.96919346e-01 3.43344539e-01 -1.29627967e+00
-1.91044807e-01 -4.94644016e-01 -2.62842655e-01 1.25247657e+00
3.24897856e-01 -4.79664564e-01 4.45161760e-01 1.26939845e+00
-1.15130357e-01 -4.47664618e-01 -7.65403450e-01 -6.10987663e-01
-7.41464794e-01 -4.36535865e-01 8.43727529e-01 4.42564398e-01
-9.27542523e-02 9.36514810e-02 -1.36163637e-01 5.26187539e-01
6.92186296e-01 -7.90741295e-02 5.21369278e-01 -1.33542788e+00
-2.68706828e-01 -1.05119586e-01 -2.68790364e-01 7.75714964e-02
-5.21782458e-01 3.06646861e-02 2.11733535e-01 -1.74231315e+00
1.86798215e-01 -2.76122808e-01 -6.01646781e-01 4.74369586e-01
-3.73484373e-01 -8.02091718e-01 2.33500108e-01 3.65523934e-01
-1.68704197e-01 3.39752674e-01 5.85286438e-01 1.16068855e-01
-7.14396417e-01 5.16231656e-01 -3.72900307e-01 6.70534790e-01
1.03838921e+00 -5.56256473e-01 -6.72659993e-01 2.51852274e-01
-6.18896745e-02 3.69410694e-01 4.03702110e-01 -1.09104133e+00
1.98637366e-01 -8.43363285e-01 3.09314936e-01 -5.27683914e-01
1.21394604e-01 -1.32813764e+00 1.14690697e+00 1.00951314e+00
-6.96300745e-01 3.28042477e-01 -9.98223387e-03 1.11598682e+00
-2.50135630e-01 -4.05869335e-01 4.84360069e-01 -1.45999521e-01
-4.26280826e-01 -1.41256396e-02 -8.43173563e-01 -6.46200418e-01
1.58154488e+00 -3.21537793e-01 -1.02424622e-01 -1.51221186e-01
-9.51177180e-01 2.99993455e-01 1.27464354e-01 1.72630981e-01
2.22314700e-01 -6.81954503e-01 -3.65077436e-01 -1.76260099e-01
3.46666835e-02 -7.03014061e-02 1.00128092e-01 4.83223557e-01
-3.49128485e-01 4.94652212e-01 -2.33068556e-01 -1.88739032e-01
-1.14719987e+00 6.68893814e-01 4.00898039e-01 -8.03328633e-01
-3.72573882e-01 1.28526939e-02 1.09736379e-02 2.52428174e-01
6.94403574e-02 -4.80168372e-01 -3.09757024e-01 -5.15369140e-02
5.46840310e-01 6.88279390e-01 2.37835392e-01 -7.00004920e-02
-1.09153599e-01 -3.25206280e-01 1.67338371e-01 -2.55529612e-01
1.36235833e+00 -3.27975079e-02 6.96132779e-02 7.63191640e-01
7.80453682e-02 -1.80704594e-01 -1.84788907e+00 -2.17795745e-03
6.86946154e-01 -4.86313760e-01 -1.05078094e-01 -1.08465624e+00
-4.98232514e-01 3.88288021e-01 5.47754645e-01 6.14708006e-01
1.26662540e+00 -2.04573736e-01 2.80677509e-02 1.75173521e-01
3.72928441e-01 -1.27673900e+00 -7.42784292e-02 1.05021685e-01
7.96968997e-01 -6.95308447e-01 4.60200757e-02 -5.09107888e-01
-7.34546483e-01 9.83422816e-01 5.68344176e-01 9.48632061e-02
5.24427950e-01 2.45166942e-01 -3.34452540e-01 -2.54256755e-01
-1.11621583e+00 2.47053117e-01 -4.25596774e-01 6.59782588e-01
4.07736897e-01 5.33822179e-01 -3.21212113e-01 2.57386476e-01
3.20782363e-01 2.84721375e-01 8.64134133e-01 1.51745009e+00
-6.43812656e-01 -9.59167957e-01 -8.37480307e-01 8.37870777e-01
-4.83950675e-01 4.11137730e-01 -4.01670188e-01 5.01275539e-01
2.17888728e-01 1.33049095e+00 4.24531877e-01 4.08776738e-02
9.74036992e-01 3.98420662e-01 -2.15824768e-02 -1.08480299e+00
-6.86802804e-01 1.32997245e-01 3.99286091e-01 -3.71482521e-01
-3.67487818e-01 -1.06524074e+00 -1.23577595e+00 -1.03561208e-02
9.12509188e-02 -4.19290587e-02 6.22144043e-01 9.16518748e-01
2.97225118e-01 1.12650371e+00 6.08989298e-01 -4.86376047e-01
-8.56254756e-01 -1.11648083e+00 -4.52718556e-01 5.42053640e-01
-3.65953483e-02 -6.06531203e-01 -3.01821053e-01 2.66701549e-01] | [8.587885856628418, 5.985487461090088] |
73c20861-c707-46f4-b179-846f9e2dd1c4 | motion-aware-memory-network-for-fast-video | 2208.00946 | null | https://arxiv.org/abs/2208.00946v1 | https://arxiv.org/pdf/2208.00946v1.pdf | Motion-aware Memory Network for Fast Video Salient Object Detection | Previous methods based on 3DCNN, convLSTM, or optical flow have achieved great success in video salient object detection (VSOD). However, they still suffer from high computational costs or poor quality of the generated saliency maps. To solve these problems, we design a space-time memory (STM)-based network, which extracts useful temporal information of the current frame from adjacent frames as the temporal branch of VSOD. Furthermore, previous methods only considered single-frame prediction without temporal association. As a result, the model may not focus on the temporal information sufficiently. Thus, we initially introduce object motion prediction between inter-frame into VSOD. Our model follows standard encoder--decoder architecture. In the encoding stage, we generate high-level temporal features by using high-level features from the current and its adjacent frames. This approach is more efficient than the optical flow-based methods. In the decoding stage, we propose an effective fusion strategy for spatial and temporal branches. The semantic information of the high-level features is used to fuse the object details in the low-level features, and then the spatiotemporal features are obtained step by step to reconstruct the saliency maps. Moreover, inspired by the boundary supervision commonly used in image salient object detection (ISOD), we design a motion-aware loss for predicting object boundary motion and simultaneously perform multitask learning for VSOD and object motion prediction, which can further facilitate the model to extract spatiotemporal features accurately and maintain the object integrity. Extensive experiments on several datasets demonstrated the effectiveness of our method and can achieve state-of-the-art metrics on some datasets. The proposed model does not require optical flow or other preprocessing, and can reach a speed of nearly 100 FPS during inference. | ['Xiaofei He', 'Ronghua Liang', 'Dongdong Zhao', 'Guodao Sun', 'Peipei Li', 'Haoran Liang', 'Xing Zhao'] | 2022-08-01 | null | null | null | null | ['video-salient-object-detection'] | ['computer-vision'] | [ 1.70606777e-01 -4.01576012e-01 -5.62209725e-01 -2.02052444e-01
-3.21353674e-01 1.21355787e-01 2.46423945e-01 -2.23440438e-04
-4.32607502e-01 6.81442976e-01 3.18955719e-01 1.79150790e-01
-1.13165928e-02 -7.67683446e-01 -6.77946270e-01 -6.70703709e-01
9.85039398e-02 -3.31788123e-01 1.02275169e+00 5.31485640e-02
4.59033459e-01 1.64684057e-01 -1.69611084e+00 3.32787931e-01
9.49710548e-01 1.39259934e+00 9.31825280e-01 3.38317215e-01
-2.71416068e-01 1.03749883e+00 -2.27412283e-01 1.84714317e-01
-7.65906693e-03 -4.13590729e-01 -6.81036294e-01 9.08949152e-02
2.67898917e-01 -6.69017553e-01 -6.23152554e-01 1.07486951e+00
3.46052259e-01 2.13550463e-01 8.80340040e-02 -1.30563962e+00
-5.87929189e-01 2.29931906e-01 -6.27981305e-01 6.50983632e-01
2.11024120e-01 2.31161416e-01 8.10040474e-01 -1.01890790e+00
6.16964102e-01 1.28498554e+00 2.32259527e-01 5.39562821e-01
-8.87464941e-01 -6.06002390e-01 4.55830991e-01 6.44547105e-01
-1.08429420e+00 -4.41196322e-01 1.01384640e+00 -3.97989243e-01
6.06633186e-01 1.72855717e-03 8.33854556e-01 8.58392835e-01
3.45706582e-01 1.29509413e+00 7.49316931e-01 5.31188212e-03
-2.46941857e-02 -1.82339415e-01 -9.16332379e-02 9.03830409e-01
2.74927542e-02 1.73933536e-01 -8.17988813e-01 3.10001075e-01
1.12603843e+00 3.13518286e-01 -6.38081908e-01 -2.98643142e-01
-1.58303010e+00 6.14927173e-01 7.34825015e-01 3.56652409e-01
-4.03008521e-01 1.86458126e-01 2.66593277e-01 -1.74082145e-01
3.07092071e-01 -4.23569232e-02 -2.91990280e-01 -6.54104501e-02
-1.07522488e+00 5.44576496e-02 1.45566776e-01 8.85789275e-01
9.79700565e-01 1.71310119e-02 -6.38034999e-01 5.82352996e-01
4.78293628e-01 3.66924703e-01 5.73152840e-01 -1.09542131e+00
5.68594873e-01 4.64234799e-01 3.22881371e-01 -1.43074882e+00
-3.46635044e-01 -2.36740500e-01 -8.26076388e-01 1.03557870e-01
2.58172423e-01 1.31195918e-01 -9.34839070e-01 1.74650371e+00
3.26752663e-01 7.74057984e-01 -8.65917727e-02 1.51865232e+00
9.03020084e-01 8.42851400e-01 1.24808826e-01 -4.89811808e-01
1.14159107e+00 -1.25441074e+00 -9.35962975e-01 -2.42179543e-01
4.43614066e-01 -6.90063417e-01 9.20723617e-01 5.93794137e-02
-1.20127726e+00 -9.00539994e-01 -1.03799391e+00 -3.39937866e-01
4.95589189e-02 2.84341216e-01 6.01302207e-01 -1.21771358e-01
-8.87959421e-01 6.19643271e-01 -1.00538242e+00 8.31681192e-02
7.26893961e-01 3.11373740e-01 -8.16817135e-02 -7.30249882e-02
-1.28651595e+00 8.15528572e-01 5.53546488e-01 3.17348421e-01
-1.07216787e+00 -4.99805003e-01 -1.03078556e+00 7.24961981e-02
4.38417196e-01 -7.33989060e-01 1.00000489e+00 -1.06285238e+00
-1.26871157e+00 3.66145223e-01 -6.28799498e-01 -3.24809700e-01
2.84085333e-01 -1.82593599e-01 -3.70492518e-01 5.44366300e-01
2.42713794e-01 1.01945436e+00 9.56399262e-01 -9.73543942e-01
-1.17488647e+00 -5.16147837e-02 1.05181746e-01 3.02143216e-01
-4.79864508e-01 -7.60196894e-02 -5.15977383e-01 -7.01151669e-01
3.32446337e-01 -4.70113337e-01 -1.42791599e-01 5.18163025e-01
-1.37777552e-01 -1.55340075e-01 1.19587362e+00 -6.83862865e-01
1.26804388e+00 -2.21663332e+00 2.89148331e-01 -3.47094029e-01
3.55985969e-01 5.53635955e-01 -1.77717477e-01 -3.21419299e-01
2.92023808e-01 -1.72757700e-01 -2.16510147e-01 -2.93572098e-01
-4.92679268e-01 8.30982029e-02 -3.30149323e-01 2.37037897e-01
4.78781790e-01 1.06429207e+00 -1.23366237e+00 -9.18925762e-01
5.52087426e-01 5.68493068e-01 -4.69932079e-01 1.85525820e-01
-3.21912885e-01 4.35335726e-01 -7.75963485e-01 5.33678532e-01
6.89635456e-01 -3.76456559e-01 -4.17815208e-01 -3.79260480e-01
-3.11990649e-01 4.24536556e-01 -1.01829970e+00 1.98440993e+00
-3.05076987e-01 6.13467693e-01 -1.05444707e-01 -1.07553089e+00
8.81307900e-01 1.05011724e-01 5.76813817e-01 -8.37934554e-01
7.01247156e-02 3.03412795e-01 -1.62443504e-01 -6.37230277e-01
3.35483670e-01 7.29478821e-02 4.03834581e-01 9.57257226e-02
6.76889792e-02 3.55703741e-01 1.09611116e-01 5.50135374e-02
7.28865504e-01 4.32107896e-01 7.49675333e-02 -9.21094194e-02
9.68518436e-01 -9.80291292e-02 1.11407590e+00 2.15039819e-01
-5.48389137e-01 8.39210033e-01 3.04634541e-01 -5.75462282e-01
-7.94749379e-01 -9.45370018e-01 2.47731283e-02 6.83390319e-01
1.01265502e+00 -3.20863008e-01 -5.35260379e-01 -6.48701072e-01
-1.41848996e-01 3.84563744e-01 -4.98340845e-01 -4.69126970e-01
-7.68672109e-01 -4.45662320e-01 -2.97772605e-02 5.05400538e-01
9.60954607e-01 -1.34751582e+00 -9.50127363e-01 6.13021791e-01
-6.32062912e-01 -1.31818211e+00 -7.91531980e-01 -4.21442956e-01
-9.66150701e-01 -8.75034094e-01 -8.25373352e-01 -1.07473850e+00
4.39307719e-01 8.18808496e-01 6.36294723e-01 2.96339124e-01
-1.16238825e-01 -2.93262661e-01 -3.01887780e-01 6.98890397e-03
1.33919761e-01 -1.80202857e-01 -1.64909825e-01 3.64230812e-01
2.56292641e-01 -5.54930925e-01 -8.69579613e-01 4.47973162e-01
-9.02730167e-01 5.14334202e-01 5.86665869e-01 8.53389621e-01
5.96726656e-01 -1.28004014e-01 5.97055376e-01 -9.99334082e-02
2.21332274e-02 -3.74604523e-01 -5.36811471e-01 9.49092135e-02
-3.06303769e-01 -3.44398208e-02 5.89887440e-01 -4.22085494e-01
-1.06840694e+00 9.11199898e-02 1.10020585e-01 -9.02280509e-01
1.26878157e-01 3.79690200e-01 -1.98298648e-01 -1.75956748e-02
5.92529960e-02 6.56192601e-01 3.74760255e-02 -3.74832749e-01
-9.67163220e-02 5.35286486e-01 5.28328121e-01 -1.26704007e-01
5.35869896e-01 6.91655993e-01 8.86293966e-03 -4.98157710e-01
-1.15904415e+00 -3.21024984e-01 -6.16557539e-01 -2.87169546e-01
1.06812906e+00 -1.05028534e+00 -7.04158068e-01 6.98064506e-01
-1.38731170e+00 -6.64734840e-02 -4.91631813e-02 7.53027499e-01
-6.29295647e-01 5.07495224e-01 -6.18191957e-01 -6.07607663e-01
-2.57014930e-01 -1.32254267e+00 1.14452457e+00 6.41066194e-01
2.83836484e-01 -8.40930521e-01 -4.75859553e-01 1.90918803e-01
3.57757300e-01 1.05562232e-01 6.32510066e-01 8.11327547e-02
-1.17293906e+00 3.59405488e-01 -6.08932376e-01 2.07220376e-01
3.34956646e-01 -3.71313025e-03 -7.57440090e-01 -1.80368766e-01
8.37370604e-02 -1.40458152e-01 1.23918259e+00 6.55264616e-01
1.39134574e+00 -1.66242003e-01 -5.57432175e-01 6.78885937e-01
1.24810648e+00 1.59449577e-01 5.99779069e-01 3.71630341e-01
9.38073575e-01 5.15099406e-01 1.09774029e+00 3.71895999e-01
5.38523734e-01 7.79942632e-01 6.49076223e-01 -8.08626413e-02
-3.25657725e-01 -4.62344348e-01 5.43298364e-01 6.11943901e-01
-4.62458916e-02 -6.40247017e-02 -5.19957423e-01 7.07848728e-01
-2.16971064e+00 -1.11042237e+00 -1.65652350e-01 2.13363934e+00
7.77397275e-01 2.85382003e-01 1.56581122e-02 9.26447660e-02
9.53176856e-01 4.74115849e-01 -6.08448148e-01 1.21898867e-01
-2.72583127e-01 -2.92001516e-01 1.37999624e-01 4.43496108e-01
-1.23723996e+00 1.05222416e+00 5.05754662e+00 8.20700347e-01
-1.33089745e+00 1.63919076e-01 5.84684432e-01 -2.93237954e-01
-3.28714937e-01 8.84380788e-02 -8.46981347e-01 9.11920249e-01
3.81891131e-01 -1.64931819e-01 2.35577703e-01 6.91195607e-01
5.17517328e-01 -3.09868604e-01 -7.72973835e-01 1.05615127e+00
4.86587100e-02 -1.59383285e+00 1.31246820e-01 -1.46154270e-01
6.69099987e-01 -1.18862994e-01 -7.40016950e-03 -5.81247313e-03
-3.82794589e-01 -7.68455029e-01 8.57249737e-01 7.12457597e-01
6.03559732e-01 -6.16661727e-01 6.29790843e-01 4.72654343e-01
-1.47040057e+00 -2.98183501e-01 -6.02798283e-01 -8.31644461e-02
4.78254557e-01 7.46681631e-01 -2.20201805e-01 5.71772099e-01
7.95296431e-01 1.39554179e+00 -4.08524126e-01 1.33136916e+00
-1.51685134e-01 3.10470939e-01 -2.59481013e-01 -8.98498669e-02
4.60135937e-01 -7.27097690e-02 6.98933780e-01 8.41441810e-01
3.76320899e-01 1.47987619e-01 3.38662863e-01 8.94344509e-01
1.60225406e-01 -4.81055342e-02 -2.76722431e-01 2.72870004e-01
3.81510139e-01 1.20807564e+00 -5.66394567e-01 -5.08967519e-01
-5.63326478e-01 1.02440119e+00 2.93221444e-01 3.47163439e-01
-1.00477982e+00 -4.55095917e-01 6.36001885e-01 8.52085650e-02
6.29580677e-01 -1.52401820e-01 -1.72605872e-01 -1.37872756e+00
4.13775206e-01 -3.20996314e-01 2.55105168e-01 -8.80237579e-01
-7.36303508e-01 3.99139673e-01 -2.72988200e-01 -1.57257128e+00
1.27159640e-01 -2.25642741e-01 -6.14051223e-01 8.22587013e-01
-2.00523973e+00 -1.04963136e+00 -5.38108408e-01 6.34572744e-01
7.27665663e-01 4.73584980e-02 2.22269699e-01 3.36088657e-01
-5.87802649e-01 1.71087056e-01 -2.36922055e-01 1.79855198e-01
5.06560087e-01 -7.84747601e-01 1.32834882e-01 1.06587541e+00
-6.98895678e-02 2.39641935e-01 3.32087934e-01 -6.76240623e-01
-1.09193933e+00 -1.27934742e+00 1.00376141e+00 -2.12524738e-02
4.10731018e-01 -1.55861959e-01 -1.16886532e+00 3.94223183e-01
-1.46415293e-01 4.82356608e-01 -5.30100688e-02 -5.95655859e-01
1.15904681e-01 -2.52644300e-01 -8.85031223e-01 5.06785393e-01
1.32696152e+00 -3.37974846e-01 -7.72239566e-01 8.41367468e-02
1.02369285e+00 -3.74012887e-01 -5.07945061e-01 6.31585598e-01
3.77307922e-01 -1.14043164e+00 9.03647482e-01 -3.82242054e-01
8.06299150e-01 -7.70363927e-01 8.02316591e-02 -9.18099344e-01
-5.02376854e-01 -5.82995892e-01 -5.09058297e-01 1.21612155e+00
1.51987802e-02 -3.65241975e-01 6.78452671e-01 2.45409116e-01
-4.01510328e-01 -9.51060951e-01 -9.99913692e-01 -5.35336494e-01
-4.72453773e-01 -2.20053792e-01 3.58654588e-01 6.83536470e-01
-1.73720568e-01 1.63035870e-01 -5.11864007e-01 1.70018032e-01
5.83617210e-01 5.14396727e-01 3.29618216e-01 -1.10557640e+00
2.34120823e-02 -4.44219321e-01 -6.25921786e-01 -1.61171138e+00
2.23864347e-01 -7.10562706e-01 1.75728023e-01 -1.53052056e+00
2.69509792e-01 -3.51203233e-01 -6.59084320e-01 5.18503487e-01
-6.10683680e-01 1.38269261e-01 3.32992166e-01 3.58893454e-01
-7.76794016e-01 1.00656843e+00 1.89692187e+00 -6.66176751e-02
-2.15201482e-01 -1.77891672e-01 -3.95959616e-01 7.08976626e-01
5.80546856e-01 -3.07676703e-01 -4.32422996e-01 -5.85684657e-01
-3.65340322e-01 2.82359719e-01 7.14385390e-01 -1.11139667e+00
3.16680104e-01 -3.81345868e-01 6.51186705e-01 -7.78259993e-01
4.63027596e-01 -6.54056787e-01 -3.81487429e-01 5.90164363e-01
-1.88230544e-01 -2.34215215e-01 -2.70481110e-02 6.32816851e-01
-5.34867883e-01 2.24260669e-02 9.10275161e-01 4.89618890e-02
-1.19115937e+00 7.44727612e-01 -1.40754685e-01 -3.55901904e-02
1.12473547e+00 -2.91294426e-01 -2.58347094e-01 -2.04673052e-01
-4.66239005e-01 5.11087775e-01 4.39204454e-01 8.43846798e-01
1.06899869e+00 -1.50950933e+00 -5.65323412e-01 3.81832987e-01
-1.34443462e-01 2.79478103e-01 4.31561112e-01 1.13651240e+00
-2.94866830e-01 3.78061831e-01 -5.21317124e-01 -8.90384018e-01
-9.05431807e-01 8.06059659e-01 3.60538185e-01 1.96466789e-01
-6.71709657e-01 7.59244502e-01 4.96245027e-01 3.43843490e-01
6.49015652e-03 -3.92053396e-01 -4.21436101e-01 -1.84609238e-02
7.93934822e-01 2.01556832e-01 -4.17084187e-01 -7.93531597e-01
-4.60132092e-01 8.05277944e-01 -2.56989077e-02 6.42365590e-02
1.16631591e+00 -3.90243530e-01 -9.73965749e-02 4.66643244e-01
1.33360565e+00 -4.52577174e-01 -1.84751844e+00 -4.25257087e-01
-1.18715540e-01 -8.96979034e-01 3.56964886e-01 -2.79175550e-01
-1.35622334e+00 1.14108336e+00 4.53970730e-01 -1.67956218e-01
1.33683264e+00 -1.37626648e-01 1.31577158e+00 -8.51840675e-02
4.17945385e-01 -1.00299835e+00 2.30375543e-01 3.47873658e-01
6.57858729e-01 -1.32009256e+00 -1.30155891e-01 -6.25391066e-01
-5.40427566e-01 1.06905603e+00 9.73238170e-01 -1.32980213e-01
5.60227811e-01 -1.30222380e-01 -1.97897390e-01 2.12371364e-01
-8.29147100e-01 -3.35600168e-01 3.32876801e-01 5.25308728e-01
3.55870537e-02 -4.01293159e-01 -3.27874631e-01 3.87873679e-01
2.95888931e-01 3.33464235e-01 4.69424337e-01 7.91529775e-01
-7.84829617e-01 -7.38361716e-01 -1.44564454e-02 4.50900733e-01
-3.35823715e-01 -1.05425134e-01 2.50551701e-01 3.78898621e-01
2.96972305e-01 9.03232336e-01 1.74626619e-01 -4.49463010e-01
1.49095073e-01 -4.07978863e-01 2.51836210e-01 -3.90688390e-01
-3.66651751e-02 1.83710083e-01 -2.34762222e-01 -8.89073312e-01
-7.67175138e-01 -6.78134263e-01 -1.60578823e+00 -3.86395113e-04
-3.57024133e-01 1.12757020e-01 2.15427741e-01 1.00612688e+00
5.03719628e-01 4.88135815e-01 8.74400139e-01 -1.11483836e+00
-1.37360051e-01 -7.35652268e-01 -3.42724472e-01 4.13774371e-01
7.63866127e-01 -1.04180491e+00 -3.05198759e-01 1.19173683e-01] | [9.507818222045898, -0.33586257696151733] |
19d93a14-18a9-43f8-824f-078e711fcfb4 | medical-code-prediction-with-multi-view | 1811.01468 | null | http://arxiv.org/abs/1811.01468v1 | http://arxiv.org/pdf/1811.01468v1.pdf | Medical code prediction with multi-view convolution and description-regularized label-dependent attention | A ubiquitous task in processing electronic medical data is the assignment of
standardized codes representing diagnoses and/or procedures to free-text
documents such as medical reports. This is a difficult natural language
processing task that requires parsing long, heterogeneous documents and
selecting a set of appropriate codes from tens of thousands of
possibilities---many of which have very few positive training samples. We
present a deep learning system that advances the state of the art for the
MIMIC-III dataset, achieving a new best micro F1-measure of 55.85\%,
significantly outperforming the previous best result (Mullenbach et al. 2018).
We achieve this through a number of enhancements, including two major novel
contributions: multi-view convolutional channels, which effectively learn to
adjust kernel sizes throughout the input; and attention regularization,
mediated by natural-language code descriptions, which helps overcome sparsity
for thousands of uncommon codes. These and other modifications are selected to
address difficulties inherent to both automated coding specifically and deep
learning generally. Finally, we investigate our accuracy results in detail to
individually measure the impact of these contributions and point the way
towards future algorithmic improvements. | ['David Suendermann-Oeft', 'Greg P. Finley', 'Slava Baryshnikov', 'Vignesh Murali', 'Najmeh Sadoughi', 'Mark Miller', 'Nico Axtmann', 'James Fone', 'Maxim Korenevski'] | 2018-11-05 | null | null | null | null | ['medical-code-prediction'] | ['medical'] | [ 3.49766284e-01 2.71958202e-01 -3.03598374e-01 -5.62928379e-01
-1.46121478e+00 -5.83997726e-01 2.04109833e-01 7.45232821e-01
-4.37452883e-01 5.70686042e-01 6.02569580e-01 -5.45780122e-01
-1.39438778e-01 -4.03690904e-01 -4.80453163e-01 -2.97671705e-01
-3.93141359e-01 7.34780014e-01 -4.25977379e-01 9.21246782e-02
2.13410944e-01 1.16818361e-01 -1.13338637e+00 7.49292314e-01
8.35508823e-01 9.19308782e-01 -1.44207731e-01 8.33731592e-01
-2.19914228e-01 1.24391544e+00 -2.75380760e-01 -5.97110152e-01
-8.45464692e-02 -4.80807424e-02 -1.11094153e+00 -3.63741145e-02
4.75271672e-01 -4.27003026e-01 -1.30806074e-01 9.04093444e-01
4.83693004e-01 -2.37436950e-01 7.61368692e-01 -5.46133101e-01
-9.17247951e-01 7.32245684e-01 -5.06181180e-01 2.91119337e-01
5.22636592e-01 1.13797560e-01 1.24010098e+00 -7.69485772e-01
8.23162138e-01 9.68256295e-01 9.97694254e-01 6.97766304e-01
-1.14948058e+00 -6.21146560e-01 -2.05181643e-01 -1.96613431e-01
-1.46755660e+00 -5.55130601e-01 7.36041591e-02 -9.09182131e-01
1.39177251e+00 1.73178956e-01 2.90315956e-01 1.07148254e+00
5.30987263e-01 7.21655905e-01 6.37852848e-01 -3.16958100e-01
1.83365792e-01 5.99444807e-02 3.61283094e-01 1.15449321e+00
3.61214370e-01 -3.96243125e-01 -1.10830545e-01 -7.36136198e-01
4.16916639e-01 9.64523703e-02 -2.10052669e-01 -2.92692930e-01
-1.38374543e+00 1.02843559e+00 3.26655418e-01 2.50049114e-01
-1.69860661e-01 5.63975126e-02 8.76527727e-01 -8.15348700e-03
3.67979944e-01 7.91689336e-01 -7.64443934e-01 -1.92669079e-01
-1.00065792e+00 2.52128154e-01 8.21474850e-01 1.11351109e+00
2.97751158e-01 -4.14137930e-01 -4.81755286e-01 8.93639922e-01
8.68329108e-02 1.82075068e-01 5.23720622e-01 -6.29587173e-01
8.32371235e-01 6.50555909e-01 -1.14142478e-01 -8.92750144e-01
-9.74337697e-01 -6.71357930e-01 -1.05871058e+00 -3.69589269e-01
2.59845108e-01 -3.09428304e-01 -1.18139315e+00 1.58156776e+00
-1.96312428e-01 -3.04634959e-01 3.31252143e-02 5.68027318e-01
1.03225195e+00 4.68221724e-01 3.55599672e-01 4.02093902e-02
1.64201331e+00 -9.89521861e-01 -6.83348238e-01 -2.93061018e-01
1.22841108e+00 -6.88317239e-01 1.03138971e+00 3.75313759e-01
-1.24117064e+00 -1.78102583e-01 -6.17989898e-01 -5.51431179e-01
-3.47315818e-01 5.53447425e-01 8.96406531e-01 4.42851305e-01
-1.19515109e+00 1.51437610e-01 -9.18511987e-01 -4.21079218e-01
9.15895641e-01 4.84251410e-01 -4.51278359e-01 -2.94653147e-01
-1.09787261e+00 7.42756188e-01 1.79891467e-01 -2.68047273e-01
-5.52064478e-01 -9.56647098e-01 -1.14345336e+00 4.40180063e-01
1.69424787e-01 -9.13193226e-01 1.23792315e+00 -7.60845602e-01
-8.21259022e-01 1.50726378e+00 -1.80545032e-01 -5.59042454e-01
4.55135584e-01 2.70482264e-02 -4.04404193e-01 1.77709594e-01
3.61963868e-01 6.58720076e-01 1.70007274e-01 -6.54474974e-01
-3.04070443e-01 -9.18155685e-02 -8.88875797e-02 -7.88466930e-02
-4.26347971e-01 1.35399193e-01 -7.09198356e-01 -8.38956118e-01
-1.97236419e-01 -9.48522806e-01 -4.86556560e-01 1.50577143e-01
-5.41590214e-01 -9.53348279e-02 -4.12726820e-01 -8.21414351e-01
1.46130884e+00 -2.24642897e+00 4.19839993e-02 2.55865186e-01
5.95673680e-01 -1.04610376e-01 -1.31455392e-01 3.17893922e-01
-3.16324621e-01 3.87350380e-01 -6.19124174e-01 -4.91768152e-01
-1.84343830e-01 -5.27332425e-02 -1.94951236e-01 4.54806834e-01
5.38698018e-01 8.75011206e-01 -9.00237978e-01 -5.78268826e-01
-2.93053538e-01 2.85378098e-01 -1.01242089e+00 1.44285839e-02
1.83294863e-02 7.91345090e-02 -4.46690679e-01 7.82215416e-01
5.34046829e-01 -9.43229854e-01 1.14548065e-01 8.55996385e-02
1.39853016e-01 6.97176278e-01 -7.93793559e-01 1.89757705e+00
-4.81981933e-01 4.10112232e-01 2.26594910e-01 -8.83735597e-01
4.45160747e-01 2.91411579e-01 7.79615402e-01 4.78257947e-02
1.93533629e-01 1.84922203e-01 6.10392950e-02 -8.03564310e-01
3.72936815e-01 3.37063335e-02 -2.84375489e-01 2.33751625e-01
1.43335387e-01 8.12189002e-03 2.45365694e-01 5.28776586e-01
1.70382142e+00 -5.92736781e-01 7.89349437e-01 -4.98585701e-01
3.97784680e-01 1.21847928e-01 4.39168870e-01 9.48631406e-01
-1.73277363e-01 9.33875322e-01 9.03196633e-01 -7.56675899e-01
-8.77531469e-01 -9.14117634e-01 -4.94508177e-01 8.28294933e-01
-5.57212472e-01 -6.32863224e-01 -5.01418352e-01 -7.55387664e-01
2.39760116e-01 4.61131483e-01 -1.05078125e+00 -1.24467224e-01
-1.52070150e-01 -1.03826714e+00 5.97106099e-01 7.12574899e-01
-1.11546583e-01 -8.03906441e-01 -4.44794804e-01 2.54055709e-01
-1.31964654e-01 -1.12923229e+00 -6.49161160e-01 4.89795595e-01
-6.69952273e-01 -1.26449394e+00 -6.26982987e-01 -8.62485886e-01
9.18181181e-01 -2.54065990e-01 1.50216675e+00 4.68976170e-01
-6.91181302e-01 3.37001622e-01 -3.67858350e-01 -3.89498830e-01
-6.27761364e-01 5.47872126e-01 -2.24992633e-01 -2.59949893e-01
6.11933470e-01 1.32797927e-01 -3.70999694e-01 -4.78735507e-01
-9.23579752e-01 1.85882837e-01 7.12718487e-01 1.01413083e+00
5.00240088e-01 -5.48290074e-01 3.24589223e-01 -1.53043604e+00
7.15533376e-01 -8.31741273e-01 -3.25096309e-01 1.90504670e-01
-6.01627111e-01 1.25090584e-01 7.16447532e-01 5.34445280e-03
-4.57846433e-01 2.82181948e-01 -5.38112938e-01 -2.48565041e-02
-1.14923023e-01 7.37255514e-01 4.63998199e-01 1.97524399e-01
7.02750683e-01 3.90858315e-02 -1.11328140e-01 -1.95225686e-01
3.51433977e-02 7.46156514e-01 3.66088003e-01 -3.21956158e-01
2.87948608e-01 3.61568242e-01 -2.93126732e-01 -2.69765526e-01
-1.18650436e+00 -6.28324330e-01 -4.66648012e-01 3.97054821e-01
1.16119981e+00 -1.19595873e+00 -5.14312804e-01 2.25520227e-02
-1.15962613e+00 -1.32597119e-01 -4.60049435e-02 4.56763417e-01
-3.41999561e-01 5.23132943e-02 -1.15390658e+00 -2.27495611e-01
-4.95156348e-01 -1.35421658e+00 1.39656734e+00 -1.90210015e-01
-6.22556269e-01 -9.29035962e-01 7.76281953e-02 4.20110673e-01
4.00675476e-01 2.53097385e-01 1.15855503e+00 -8.32587898e-01
-1.09748527e-01 -1.83593616e-01 -4.27845210e-01 -9.63972807e-02
2.68408865e-01 7.90218711e-02 -9.35313821e-01 -4.12485570e-01
-3.69946957e-01 -5.31574309e-01 1.22358727e+00 5.39344490e-01
1.81404376e+00 -1.73389956e-01 -6.11063302e-01 9.75830615e-01
1.26169538e+00 -6.60362318e-02 3.41148406e-01 2.02754080e-01
7.07082212e-01 2.89627761e-01 8.68922193e-03 5.87741196e-01
5.69280148e-01 4.14727181e-01 1.64350018e-01 -4.12766606e-01
2.13893220e-01 2.24702373e-01 -9.28378999e-02 9.72035766e-01
3.60694528e-01 -6.69451524e-03 -1.49472427e+00 6.57762349e-01
-1.42772651e+00 -6.70935273e-01 -9.07624140e-02 1.85884976e+00
1.12994981e+00 2.29138598e-01 -1.72434241e-01 -3.79907638e-01
4.97869849e-01 -9.99935418e-02 -6.56166673e-01 -6.00815654e-01
1.83698558e-03 2.72170305e-01 6.12565815e-01 4.72830921e-01
-1.50200820e+00 5.46437740e-01 7.14673281e+00 4.83741343e-01
-9.71212089e-01 5.36643639e-02 1.17476165e+00 -2.14044675e-01
-3.72605592e-01 -6.51721299e-01 -5.81375122e-01 5.04716396e-01
1.14703810e+00 1.94726121e-02 3.47862303e-01 9.59734023e-01
-1.75722301e-01 2.06461206e-01 -1.28375006e+00 1.23769712e+00
2.74609357e-01 -1.70439148e+00 2.98113246e-02 3.46419774e-02
8.70596528e-01 4.86945033e-01 7.77446479e-02 3.19979906e-01
4.63946372e-01 -1.08934426e+00 5.26790738e-01 3.84054244e-01
1.25941098e+00 -5.07040560e-01 9.81829762e-01 -1.39597610e-01
-1.03921616e+00 -3.51449519e-01 -2.57343948e-01 1.51502207e-01
-1.41016975e-01 8.83414507e-01 -8.87104690e-01 2.59424567e-01
7.58529603e-01 9.58952665e-01 -7.17590332e-01 8.80572438e-01
2.31956735e-01 6.74142659e-01 1.12980073e-02 1.32391572e-01
2.62562186e-01 4.52835768e-01 3.42491046e-02 1.88674986e+00
2.64383763e-01 1.73596039e-01 2.46606171e-01 1.03938460e+00
-4.60569412e-01 1.89759552e-01 -5.32582700e-01 -9.43929777e-02
3.12252969e-01 1.30240357e+00 -7.79600620e-01 -5.80140352e-01
-6.45087719e-01 7.91586697e-01 7.81646430e-01 5.02734147e-02
-8.29740703e-01 -4.00555849e-01 7.81540990e-01 -1.27091989e-01
1.45360664e-01 -2.75576301e-02 -5.90012789e-01 -1.44960928e+00
-1.26965567e-01 -1.05123401e+00 7.86271214e-01 -5.98442256e-01
-1.57019782e+00 6.62074149e-01 -4.70670253e-01 -1.32020152e+00
-2.75892794e-01 -8.53494704e-01 5.02941012e-02 6.94087505e-01
-1.46825528e+00 -6.93637609e-01 -1.21034890e-01 3.76896441e-01
4.22444075e-01 -2.97866225e-01 1.27360666e+00 6.43141985e-01
-4.87419188e-01 9.90788817e-01 3.20088953e-01 5.32240391e-01
9.63928699e-01 -1.30705047e+00 6.11037076e-01 5.05184650e-01
-1.56085253e-01 9.25441265e-01 2.16244489e-01 -3.40776622e-01
-1.36503685e+00 -1.23084879e+00 1.14849126e+00 -6.85086668e-01
6.95871830e-01 -6.62359476e-01 -8.39603245e-01 8.04138124e-01
-8.76870975e-02 2.27561921e-01 1.13167191e+00 2.16390848e-01
-3.86191726e-01 8.54826793e-02 -1.14880371e+00 3.82431000e-01
8.37149441e-01 -5.63674331e-01 -3.86534423e-01 6.11785710e-01
6.44976258e-01 -8.02711844e-01 -8.96645188e-01 1.92071900e-01
3.47218364e-01 -7.55510390e-01 8.32510293e-01 -8.34407687e-01
1.09491301e+00 1.41783759e-01 7.96030238e-02 -1.16065764e+00
-6.93930924e-01 -3.86050105e-01 2.42985711e-01 8.28901052e-01
6.99891865e-01 -4.71803844e-01 5.15188217e-01 8.88914824e-01
-4.01802838e-01 -1.02537370e+00 -7.48720706e-01 -1.25160053e-01
2.50961542e-01 -3.51545632e-01 4.59401548e-01 1.34512246e+00
3.20170522e-01 1.12120986e-01 -1.43272042e-01 -4.00068518e-03
2.22922862e-01 -1.35068119e-01 2.49594599e-01 -1.11597991e+00
-2.07124665e-01 -4.96780723e-01 -2.58225918e-01 -6.98485255e-01
5.03752455e-02 -1.37539482e+00 -1.47227813e-02 -1.64784884e+00
6.94164038e-01 -4.47212279e-01 -3.70145947e-01 7.46246338e-01
-4.31125551e-01 6.23682477e-02 -6.19301526e-03 1.60522431e-01
-7.20980346e-01 -7.46986866e-02 9.32612300e-01 -3.85870129e-01
-4.22844291e-02 -2.53718942e-01 -1.13375878e+00 5.67555189e-01
4.92672265e-01 -6.43747628e-01 -1.00951537e-01 -7.80651689e-01
4.95635152e-01 1.02859899e-01 9.71873775e-02 -8.33468974e-01
5.54459691e-02 3.29114199e-02 5.02873063e-01 -1.50833130e-01
-2.15623960e-01 -6.98374391e-01 -2.77568609e-01 6.56834543e-01
-1.02523386e+00 3.41745794e-01 4.82245207e-01 4.36131865e-01
-2.24596933e-02 -1.61998019e-01 7.14211106e-01 -4.31469023e-01
-2.87857056e-01 2.18331262e-01 -6.56253099e-01 3.48384380e-01
7.01414824e-01 2.09626153e-01 -3.67218047e-01 -2.24667743e-01
-7.08612621e-01 4.08544779e-01 2.11921141e-01 3.48131627e-01
3.93302172e-01 -1.06799853e+00 -9.16397214e-01 3.83159906e-01
4.79749292e-01 4.36973646e-02 2.15158716e-01 7.97388732e-01
-9.20099497e-01 6.63090646e-01 2.82004736e-02 -4.97864842e-01
-8.92097533e-01 6.59259200e-01 3.41480911e-01 -7.69129932e-01
-7.45360136e-01 1.09905756e+00 1.52447075e-01 -6.61494970e-01
2.73579031e-01 -9.71821070e-01 -2.94045657e-01 -5.19810133e-02
7.35212326e-01 -9.65811461e-02 4.45544511e-01 -2.76410908e-01
-6.86982512e-01 3.84819955e-01 -4.93451476e-01 3.23147804e-01
1.47532964e+00 3.75909396e-02 -3.58401000e-01 2.32308745e-01
1.44120097e+00 4.98097837e-02 -7.57198513e-01 4.68927855e-03
6.44143298e-02 -1.53559953e-01 -3.31727713e-02 -1.09397757e+00
-1.10903978e+00 6.88098311e-01 5.40747046e-01 6.13126643e-02
1.02066612e+00 1.86248407e-01 7.76247799e-01 4.41180676e-01
1.28719406e-02 -8.92352343e-01 -2.72164732e-01 6.37828469e-01
7.17666805e-01 -1.44949281e+00 8.19147006e-02 -1.22003324e-01
-6.21927857e-01 1.30092812e+00 4.36323196e-01 6.07360415e-02
6.47954464e-01 7.49072850e-01 5.74493818e-02 -3.74187261e-01
-1.02045417e+00 1.96077690e-01 4.51956511e-01 2.58563757e-01
1.08944511e+00 1.24014601e-01 -2.59799480e-01 9.49043214e-01
-1.88848749e-01 9.39591676e-02 6.30365133e-01 7.49247193e-01
-2.21636016e-02 -7.18535483e-01 -3.04815173e-01 1.34148788e+00
-8.63332212e-01 -5.52267134e-01 -1.54689506e-01 4.16926235e-01
1.16668276e-01 7.05069482e-01 2.89289534e-01 -2.20391944e-01
1.86033428e-01 1.05471365e-01 -4.60008681e-02 -1.05945146e+00
-9.20386195e-01 -1.08098142e-01 2.26483885e-02 -5.94019771e-01
-1.71531782e-01 -7.97104299e-01 -1.49969649e+00 -6.11797832e-02
1.29603958e-02 -8.38606134e-02 2.27992088e-01 8.39434743e-01
6.47572935e-01 8.79123569e-01 9.39934887e-03 -5.13292670e-01
-6.43185973e-01 -6.86405957e-01 -4.26457405e-01 5.32435954e-01
8.50971878e-01 -4.38138008e-01 -2.51808703e-01 3.09982300e-01] | [8.028382301330566, 6.861257553100586] |
42f415fc-d174-4a8a-9f0c-6473ca07d0ae | investigations-on-the-inference-optimization | 1911.12993 | null | https://arxiv.org/abs/1911.12993v1 | https://arxiv.org/pdf/1911.12993v1.pdf | Investigations on the inference optimization techniques and their impact on multiple hardware platforms for Semantic Segmentation | In this work, the task of pixel-wise semantic segmentation in the context of self-driving with a goal to reduce the inference time is explored. Fully Convolutional Network (FCN-8s, FCN-16s, and FCN-32s) with a VGG16 encoder architecture and skip connections is trained and validated on the Cityscapes dataset. Numerical investigations are carried out for several inference optimization techniques built into TensorFlow and TensorRT to quantify their impact on the inference time and network size. Finally, the trained network is ported on to an embedded platform (Nvidia Jetson TX1) and the inference time, as well as the total energy consumed for inference across hardware platforms, are compared. | ['Sethu Hareesh Kolluru'] | 2019-11-29 | null | null | null | null | ['inference-optimization'] | ['audio'] | [ 4.99748848e-02 3.64404470e-01 -1.20516486e-01 -4.67298537e-01
-8.86670724e-02 -3.48298222e-01 3.07327062e-01 -3.06565940e-01
-8.07021081e-01 5.95797002e-01 -3.77802730e-01 -8.20771813e-01
-1.34803116e-01 -7.14727700e-01 -6.78443849e-01 -5.21867514e-01
5.87746166e-02 3.90439294e-03 2.75712967e-01 1.51125744e-01
2.46504694e-01 6.43310428e-01 -1.50737822e+00 -1.78911507e-01
7.45349467e-01 1.45279825e+00 -1.85824577e-02 9.57008839e-01
8.59748721e-02 1.01032412e+00 -6.80148184e-01 -6.87380672e-01
3.56350660e-01 1.24725729e-01 -9.37176883e-01 -1.20294623e-01
5.58236659e-01 -4.74606663e-01 -5.43650270e-01 9.93838668e-01
3.95484447e-01 2.86237538e-01 8.74885395e-02 -1.40823865e+00
1.74795061e-01 6.11211777e-01 8.08939990e-03 6.70573056e-01
-6.48955405e-01 4.96606797e-01 6.57514811e-01 -2.86604136e-01
4.67277467e-01 1.32248247e+00 5.89456618e-01 3.16204369e-01
-8.19185972e-01 -7.40984142e-01 -1.98637456e-01 4.15060639e-01
-1.39714658e+00 -2.41331890e-01 3.39215070e-01 -2.60292411e-01
1.52360523e+00 1.80405617e-01 6.43496871e-01 1.00080085e+00
3.84948224e-01 5.31237125e-01 9.41265285e-01 -1.23094037e-01
6.16034806e-01 -8.68106559e-02 3.96499932e-01 7.95654893e-01
3.61535758e-01 -1.81339979e-02 -4.35228586e-01 5.88648736e-01
5.45602262e-01 -3.92618209e-01 3.68151963e-01 5.66953182e-01
-6.99747086e-01 9.61298525e-01 7.03570724e-01 8.34025666e-02
-1.76455110e-01 6.73926950e-01 9.32439685e-01 -1.74351081e-01
6.73513532e-01 2.74772048e-01 -5.32028735e-01 -3.67618978e-01
-1.11996436e+00 1.37769550e-01 8.09117615e-01 1.10013461e+00
7.16381371e-01 4.11502212e-01 -3.13822210e-01 3.45561922e-01
2.64069110e-01 2.59223729e-01 4.00885493e-01 -1.29189503e+00
6.32905900e-01 7.64353096e-01 -3.53489667e-01 -5.86050928e-01
-5.65929472e-01 -2.51133263e-01 -7.78959751e-01 3.64752620e-01
1.72573090e-01 -8.40069830e-01 -1.12257349e+00 9.72167552e-01
2.23824427e-01 3.72123569e-01 8.67135078e-02 9.12748814e-01
9.32515442e-01 2.69848466e-01 2.83863872e-01 5.58408797e-01
1.32073367e+00 -1.59088659e+00 -6.78452790e-01 -4.04230595e-01
6.90673649e-01 -4.40161198e-01 7.43753612e-01 1.61833867e-01
-1.13020360e+00 -9.30504441e-01 -1.37945974e+00 -6.65130436e-01
-8.54018748e-01 5.97665608e-01 9.16855395e-01 9.98813212e-01
-1.17964923e+00 9.80186284e-01 -1.25593567e+00 -2.31193721e-01
9.06153262e-01 5.46681404e-01 2.94577777e-01 3.25266838e-01
-1.07934833e+00 8.38177443e-01 8.38541687e-01 6.19237006e-01
-7.68752515e-01 -7.27296650e-01 -7.40936637e-01 1.51777565e-01
3.47588025e-02 -6.84666812e-01 1.12982452e+00 -1.01681685e+00
-1.80786884e+00 5.57542026e-01 2.28841618e-01 -1.06240499e+00
5.93468487e-01 -1.86792284e-01 -5.05287468e-01 2.27347761e-01
-3.65258753e-02 1.21291912e+00 6.11963987e-01 -3.08853298e-01
-8.18590522e-01 -4.26914096e-01 2.97810227e-01 2.37202317e-01
-2.59075426e-02 -2.74874747e-01 -4.77363229e-01 -2.74429739e-01
-2.94338673e-01 -1.11349487e+00 -3.82206291e-01 -1.29582047e-01
-8.80294263e-01 -1.26485348e-01 1.10965836e+00 -8.09841573e-01
9.31574702e-01 -2.18706918e+00 -2.45960474e-01 2.97724038e-01
-4.05722335e-02 6.61210775e-01 2.18237996e-01 -3.86102170e-01
1.64531589e-01 8.99417400e-02 -1.47602886e-01 -5.90912879e-01
1.67507865e-02 5.72359383e-01 1.75170541e-01 3.27483326e-01
3.66351932e-01 9.32318211e-01 -3.80568713e-01 -5.92363119e-01
6.23225749e-01 6.63771987e-01 -5.45072496e-01 -4.65225428e-02
-1.83130026e-01 2.68812060e-01 -4.44284141e-01 5.91268063e-01
7.90698290e-01 -1.36371881e-01 7.91744068e-02 -2.48696074e-01
-3.63605738e-01 2.98354805e-01 -8.05522084e-01 1.64360738e+00
-5.76709390e-01 1.09898901e+00 1.24311775e-01 -8.64363670e-01
8.05903018e-01 -1.91895757e-02 4.20580544e-02 -8.51092279e-01
6.25903308e-01 7.56624639e-02 9.98693034e-02 -5.58834970e-01
8.54895592e-01 6.39690995e-01 3.12980562e-01 -1.74522892e-01
3.89771581e-01 1.88524097e-01 5.83465755e-01 -1.26309432e-02
9.76581991e-01 2.38928288e-01 -4.85457063e-01 -6.06678367e-01
4.30563778e-01 3.92214656e-01 1.74909025e-01 3.94027263e-01
-1.64858729e-01 7.11894408e-02 6.29989028e-01 -4.47367460e-01
-1.20212281e+00 -8.47954929e-01 -2.82549471e-01 8.45224857e-01
1.27830893e-01 -2.47916818e-01 -1.36845171e+00 -4.68030185e-01
-1.00012302e-01 9.78376508e-01 -4.10630435e-01 -3.38718779e-02
-3.30496222e-01 -7.72244692e-01 9.83297229e-01 8.80091250e-01
1.31055462e+00 -1.03653657e+00 -1.42017603e+00 1.45035997e-01
3.02221745e-01 -1.75895226e+00 1.26149103e-01 4.83941942e-01
-1.13021505e+00 -1.12856615e+00 -1.35599270e-01 -5.04389465e-01
4.89347488e-01 -4.16820973e-01 1.04101253e+00 3.39405704e-03
-6.31876290e-01 7.72275403e-02 -4.75107227e-03 -5.20395279e-01
1.13300845e-01 6.20698750e-01 -6.05455399e-01 -1.49626955e-01
3.80225301e-01 -1.09282747e-01 -8.84230852e-01 -1.59479365e-01
-6.12998068e-01 3.56193155e-01 4.03812170e-01 3.12584430e-01
4.50979114e-01 3.14015359e-01 -1.08165853e-01 -8.71835828e-01
2.54353553e-01 -2.18423232e-01 -9.23183918e-01 -1.74818411e-01
-8.54501128e-01 1.58884574e-03 5.52540362e-01 7.13762045e-02
-1.19597304e+00 1.95108965e-01 -4.31591630e-01 -5.10537386e-01
-1.03773393e-01 1.48515955e-01 -1.37329949e-02 -2.45955065e-01
2.16765806e-01 -2.98415810e-01 1.06658861e-01 -2.85623729e-01
2.39124060e-01 5.56055486e-01 5.76809764e-01 -7.88366497e-02
6.89117536e-02 6.22264445e-01 2.31300816e-01 -9.62682188e-01
-8.50487351e-01 -6.76790476e-02 -6.73990130e-01 -2.92811871e-01
1.42353833e+00 -1.12371886e+00 -9.31670964e-01 6.60662234e-01
-1.03957784e+00 -8.57172787e-01 -1.46275252e-01 4.65473354e-01
-3.09137881e-01 -3.13610882e-01 -7.41762459e-01 -5.15834987e-01
-8.55642080e-01 -1.52741277e+00 9.07366097e-01 7.38889873e-01
-4.44142409e-02 -1.25494504e+00 -6.30879402e-01 6.15966260e-01
4.60810333e-01 2.79748708e-01 6.08922243e-01 -3.17879766e-01
-6.38880610e-01 1.03061602e-01 -8.11476648e-01 7.17962742e-01
-5.79579473e-01 3.41788918e-01 -1.08040285e+00 -6.96515962e-02
-3.71479899e-01 -5.42755499e-02 7.18241990e-01 7.61294127e-01
1.63556230e+00 -2.05363274e-01 -2.98509091e-01 9.67537582e-01
1.75380826e+00 6.35204837e-02 8.36345017e-01 4.27915603e-01
9.78205860e-01 1.75164104e-01 3.83462042e-01 2.80773401e-01
4.12178516e-01 4.19471949e-01 7.58064628e-01 -3.18401724e-01
-3.56363416e-01 1.31915346e-01 -6.77571222e-02 3.89775395e-01
-1.61725700e-01 -3.43883753e-01 -7.15122879e-01 3.99542123e-01
-1.52211726e+00 -3.32814127e-01 -5.24943650e-01 1.63118017e+00
2.69380033e-01 4.82012987e-01 -1.66806623e-01 2.23412022e-01
6.71324611e-01 6.70133159e-02 -8.70201528e-01 -1.08709657e+00
1.68472290e-01 7.66081333e-01 1.32233548e+00 2.86739796e-01
-1.04310477e+00 9.96927679e-01 6.26283503e+00 8.54387522e-01
-1.52547073e+00 1.98531851e-01 1.20885789e+00 -2.37463504e-01
3.79728526e-01 -9.72645432e-02 -1.07135499e+00 7.46067882e-01
1.64155936e+00 3.73337507e-01 5.36223531e-01 1.08925939e+00
2.18559682e-01 -6.14889860e-01 -7.10634291e-01 7.63636887e-01
-4.24650759e-01 -1.54277956e+00 -3.80821645e-01 1.24068875e-02
4.82413977e-01 5.64412534e-01 -7.41208196e-02 4.70644593e-01
2.09357783e-01 -1.02454233e+00 7.43920147e-01 1.91116095e-01
8.39743376e-01 -1.25229263e+00 9.49251652e-01 -8.53545293e-02
-9.24854934e-01 -5.71046546e-02 -2.13807449e-01 -2.15936333e-01
1.01216637e-01 5.24772465e-01 -7.89534926e-01 3.17784816e-01
1.10586965e+00 5.24991989e-01 -9.03198540e-01 4.82909441e-01
-1.84231058e-01 8.42842281e-01 -2.92206883e-01 -1.49840936e-01
7.82944739e-01 -2.75590420e-01 4.78493944e-02 1.36215127e+00
1.32034048e-01 -1.03447191e-01 -3.08704287e-01 8.82571161e-01
-8.60949904e-02 -3.81309122e-01 9.51103959e-03 -7.23887281e-03
3.37912172e-01 1.76196170e+00 -1.19467008e+00 -4.81749773e-01
-2.14399874e-01 1.01097178e+00 1.82446748e-01 2.62129396e-01
-1.31348217e+00 -6.02467239e-01 9.48864996e-01 3.24019343e-02
4.22239929e-01 -1.22108482e-01 -8.15950036e-01 -5.12949407e-01
-2.08584160e-01 -1.09637983e-01 1.07249513e-01 -8.93896580e-01
-5.77139497e-01 4.70991224e-01 -1.61938071e-01 -3.05471867e-01
5.81598617e-02 -9.93242145e-01 -7.50265002e-01 9.08614159e-01
-1.66020072e+00 -8.28057289e-01 -5.64917088e-01 3.96745890e-01
5.94244599e-01 -1.63407549e-01 4.55693156e-01 7.07134247e-01
-1.21025431e+00 6.76867306e-01 -5.62398620e-02 3.20913523e-01
-3.65295172e-01 -1.11832118e+00 7.71589458e-01 8.74388933e-01
-5.49095452e-01 1.89426631e-01 4.60488141e-01 -3.68547678e-01
-1.42942405e+00 -1.76981270e+00 5.91512561e-01 1.42512649e-01
5.07431626e-01 -2.60073721e-01 -4.95561808e-01 7.48667598e-01
4.52477396e-01 3.00046533e-01 6.33147135e-02 -2.84269542e-01
2.40657315e-01 1.84674151e-02 -1.47426164e+00 6.79486930e-01
7.81632721e-01 -3.49634141e-01 3.35269153e-01 3.55200917e-01
7.37533391e-01 -9.60210145e-01 -9.49821532e-01 1.64467722e-01
3.00871015e-01 -9.58340824e-01 8.67463946e-01 -1.27500981e-01
6.31607354e-01 4.40562032e-02 1.26909837e-01 -8.43326807e-01
7.30631351e-02 -3.08895767e-01 -9.77262035e-02 9.80599761e-01
3.62488896e-01 -6.49627328e-01 1.23306084e+00 5.61510801e-01
-4.80938137e-01 -7.93952405e-01 -1.16835284e+00 -6.84024811e-01
-1.72248006e-01 -4.49896067e-01 4.95137841e-01 3.25134009e-01
-7.22760379e-01 3.54875952e-01 2.90679723e-01 1.94398433e-01
4.10902023e-01 -3.91545922e-01 4.52648252e-01 -7.61599541e-01
1.25314668e-01 -3.79213452e-01 -5.26944816e-01 -5.31628311e-01
3.82145375e-01 -7.39961505e-01 -6.88339621e-02 -1.31115270e+00
-4.01174814e-01 -3.74022663e-01 8.67914557e-02 3.82647038e-01
1.50940701e-01 4.44278836e-01 4.55755182e-02 -3.64920557e-01
-4.89359498e-01 1.09855160e-01 1.10485721e+00 -7.50912130e-02
1.13943502e-01 -1.64675444e-01 -2.42927354e-02 6.51092529e-01
9.45840716e-01 -1.27477095e-01 -4.37964052e-01 -5.59723318e-01
-2.56467145e-02 -2.22089186e-01 7.87331223e-01 -1.51956630e+00
2.28516355e-01 4.62687701e-01 4.46198821e-01 -8.76143157e-01
4.80140477e-01 -7.33143508e-01 1.30091995e-01 7.24991381e-01
-2.18548834e-01 1.05425611e-01 5.60230374e-01 1.19553939e-01
-7.56089091e-02 -4.50725019e-01 7.13078499e-01 9.59896073e-02
-1.20495534e+00 1.69148549e-01 -3.85567218e-01 2.56790053e-02
1.15741718e+00 -3.29684496e-01 -3.22137833e-01 3.15217614e-01
-6.35761678e-01 3.15976292e-01 -1.63151488e-01 2.92433381e-01
2.05701038e-01 -1.01090491e+00 -2.70767540e-01 2.21847117e-01
-3.90232056e-01 4.00621027e-01 5.56968033e-01 7.56701648e-01
-1.30089152e+00 6.68163538e-01 -5.02341628e-01 -5.92041850e-01
-1.09807730e+00 1.22746088e-01 6.09308481e-01 -1.83121204e-01
-5.98783314e-01 9.32812154e-01 -4.84053433e-01 -2.80994385e-01
2.44414970e-01 -7.19844639e-01 -6.32600859e-02 -3.13117020e-02
1.29329920e-01 1.05699813e+00 5.14700770e-01 -3.00312042e-01
-3.27593952e-01 -2.36624852e-02 2.58762896e-01 5.50951287e-02
1.00170934e+00 -1.24110892e-01 -8.06245953e-03 -6.17548116e-02
1.41045678e+00 -1.03904903e+00 -1.72897708e+00 4.13951367e-01
1.81653630e-02 1.08813561e-01 7.43975580e-01 -7.69726515e-01
-1.81738615e+00 9.01256442e-01 1.00214338e+00 -2.08395913e-01
1.03997731e+00 -5.50255239e-01 9.37952757e-01 1.04772955e-01
-3.30046602e-02 -1.71845806e+00 -3.24960142e-01 6.58383608e-01
1.08764514e-01 -1.18030369e+00 -1.10568572e-02 -3.71440798e-01
-5.36912262e-01 1.26681268e+00 7.95111656e-01 -4.02724296e-01
7.09764123e-01 6.11573279e-01 -2.39005804e-01 -4.22779322e-01
-3.72140676e-01 -1.99388474e-01 -9.12500545e-02 2.00460166e-01
2.99597323e-01 3.11736792e-01 -1.67373177e-02 1.73339069e-01
-5.61384797e-01 2.72080898e-01 4.04081732e-01 6.93661928e-01
1.88590854e-01 -2.54158646e-01 8.28524493e-03 6.56823993e-01
-7.20747232e-01 -1.14523649e-01 1.08456546e-02 8.56606603e-01
7.00585663e-01 9.23187256e-01 7.81004965e-01 -2.80172706e-01
1.46791980e-01 -1.60944194e-01 3.33804756e-01 -2.05747724e-01
-1.11486530e+00 -3.22207123e-01 3.70406210e-01 -7.68977344e-01
-2.80755490e-01 -4.17984486e-01 -1.55843568e+00 -6.62204325e-01
-2.09959820e-01 -2.82662600e-01 1.55356276e+00 1.09592760e+00
4.65206474e-01 1.33967745e+00 8.59526619e-02 -8.32137346e-01
-8.09353068e-02 -9.42425251e-01 -5.66962719e-01 -1.94671497e-01
-7.51173273e-02 -3.33515882e-01 -1.88271865e-01 -3.39839719e-02] | [9.402623176574707, -0.2659558057785034] |
58a871c3-9dca-4466-8b9b-2e555ccb2e22 | cross-lingual-unsupervised-sentiment | null | null | https://aclanthology.org/2020.acl-main.510 | https://aclanthology.org/2020.acl-main.510.pdf | Cross-Lingual Unsupervised Sentiment Classification with Multi-View Transfer Learning | Recent neural network models have achieved impressive performance on sentiment classification in English as well as other languages. Their success heavily depends on the availability of a large amount of labeled data or parallel corpus. In this paper, we investigate an extreme scenario of cross-lingual sentiment classification, in which the low-resource language does not have any labels or parallel corpus. We propose an unsupervised cross-lingual sentiment classification model named multi-view encoder-classifier (MVEC) that leverages an unsupervised machine translation (UMT) system and a language discriminator. Unlike previous language model (LM) based fine-tuning approaches that adjust parameters solely based on the classification error on training data, we employ the encoder-decoder framework of a UMT as a regularization component on the shared network parameters. In particular, the cross-lingual encoder of our model learns a shared representation, which is effective for both reconstructing input sentences of two languages and generating more representative views from the input for classification. Extensive experiments on five language pairs verify that our model significantly outperforms other models for 8/11 sentiment classification tasks. | ['Hongliang Fei', 'Ping Li'] | 2020-07-01 | null | null | null | acl-2020-6 | ['unsupervised-machine-translation'] | ['natural-language-processing'] | [ 1.68113753e-01 -1.53631881e-01 -5.39276600e-01 -9.90053654e-01
-1.13326430e+00 -7.02984035e-01 5.60108662e-01 -3.96146685e-01
-4.73300874e-01 6.47256076e-01 2.99362659e-01 -5.34070432e-01
7.73864627e-01 -4.46846277e-01 -6.83722496e-01 -6.13667369e-01
7.07393587e-01 3.80822837e-01 -4.47298318e-01 -4.36961353e-01
-1.97673067e-01 -2.32559025e-01 -8.47881615e-01 7.27838576e-01
7.37471819e-01 9.86011446e-01 2.01355919e-01 4.74216193e-01
-3.13645244e-01 7.96178937e-01 -4.33777064e-01 -1.07709229e+00
2.39357024e-01 -5.89829028e-01 -6.35504961e-01 1.65203530e-02
2.82585591e-01 9.78293829e-03 2.08180308e-01 1.01363087e+00
6.22035980e-01 -2.19806120e-01 7.96701729e-01 -1.02380991e+00
-1.04545832e+00 8.69118512e-01 -7.14235425e-01 -2.67755866e-01
-1.17116906e-02 -4.11102086e-01 1.23345780e+00 -1.18821979e+00
5.76628268e-01 1.25122714e+00 7.58321226e-01 7.77462721e-01
-1.22435200e+00 -7.86290407e-01 4.01405126e-01 -2.14078620e-01
-1.06585109e+00 -5.80919087e-01 9.14750338e-01 -3.68273437e-01
9.83565390e-01 -1.47089809e-01 2.16229677e-01 1.61608660e+00
4.41492498e-01 9.54475403e-01 1.11069763e+00 -7.60966361e-01
2.11715177e-02 9.01205420e-01 -1.30036414e-01 5.14356136e-01
-6.80339113e-02 -2.52683908e-01 -8.27235162e-01 4.45347279e-02
1.77429155e-01 -2.57901669e-01 -1.30732313e-01 -4.54350501e-01
-9.71907318e-01 1.04681838e+00 1.07458323e-01 1.64342433e-01
1.24604724e-01 -1.07282586e-01 9.14786696e-01 7.40897715e-01
1.02729130e+00 2.69622296e-01 -1.17412448e+00 2.42952362e-01
-7.41289198e-01 -5.21324873e-01 8.01709533e-01 1.04732358e+00
6.17955267e-01 3.10315043e-01 2.43882522e-01 1.20889401e+00
3.26043487e-01 6.51139855e-01 1.07470345e+00 -2.40288585e-01
8.63990784e-01 5.92734516e-01 -4.78790343e-01 -5.33040643e-01
-1.28283411e-01 -5.80790281e-01 -9.59204912e-01 -2.29555756e-01
-6.74572438e-02 -4.84413505e-01 -6.09591424e-01 2.03367472e+00
3.81740765e-03 -3.46482307e-01 8.85460675e-01 4.40412432e-01
7.01875806e-01 5.09984970e-01 -4.50381599e-02 5.50436713e-02
1.22360992e+00 -1.31016767e+00 -4.83318925e-01 -5.74136496e-01
1.30507481e+00 -8.91631246e-01 1.25054896e+00 1.58414021e-01
-7.04046011e-01 -5.99505246e-01 -1.19794548e+00 -2.08792865e-01
-6.06096327e-01 6.51904166e-01 4.23311502e-01 7.48502016e-01
-8.90910983e-01 1.93958282e-01 -6.12635195e-01 -2.76115149e-01
2.19235644e-01 2.18293905e-01 -5.92926025e-01 -1.10623658e-01
-1.19425476e+00 9.25588548e-01 1.70551613e-01 7.13883564e-02
-6.02718532e-01 -3.75470549e-01 -1.22990060e+00 -3.07682812e-01
-6.00481145e-02 -6.20890141e-01 1.22400510e+00 -1.85759211e+00
-1.64819074e+00 1.12714708e+00 -2.71813244e-01 -2.66235501e-01
4.12553757e-01 -1.81680501e-01 -5.22841632e-01 -2.63470560e-01
5.39511204e-01 5.83516002e-01 7.53715932e-01 -1.16682613e+00
-4.91883099e-01 -4.70352590e-01 -4.28712741e-02 4.44115400e-01
-6.96122885e-01 3.96159291e-02 -4.66599613e-01 -6.76575065e-01
-6.20434880e-02 -1.02345622e+00 -2.07467467e-01 -4.31553513e-01
-4.92240727e-01 -9.41452906e-02 6.40764296e-01 -4.88374323e-01
8.47271144e-01 -2.11436868e+00 2.10015312e-01 -1.87703222e-01
-3.87581855e-01 6.59756958e-02 -3.75067562e-01 3.93123209e-01
-2.57459432e-01 9.04169753e-02 -1.57766774e-01 -1.04934788e+00
-7.86499828e-02 1.98378861e-01 -5.71856201e-01 4.13665831e-01
2.34342873e-01 8.34466577e-01 -5.98788321e-01 -3.40029627e-01
-3.28539312e-01 5.11758506e-01 -6.11738145e-01 1.66533172e-01
-5.35902195e-02 6.43272579e-01 -4.78800446e-01 4.32096303e-01
3.90653789e-01 -4.04562831e-01 6.42761946e-01 -2.52299458e-01
1.07959606e-01 6.77043140e-01 -6.29807293e-01 1.96652031e+00
-1.09797144e+00 5.35620451e-01 -1.98664099e-01 -1.14778531e+00
1.11255634e+00 5.74365139e-01 7.31913298e-02 -7.99270213e-01
2.87145436e-01 5.67673206e-01 -2.24708453e-01 -3.44397038e-01
3.87456089e-01 -4.12243247e-01 -5.63265860e-01 8.18830729e-01
4.16388333e-01 -1.33493438e-01 5.20909056e-02 1.19010888e-01
4.05548900e-01 4.97991174e-01 2.20153376e-01 -3.51871967e-01
8.55413795e-01 -6.54120594e-02 7.20064819e-01 3.18635285e-01
3.53473462e-02 5.84477663e-01 5.60475290e-01 -4.73616660e-01
-1.04379308e+00 -4.66403693e-01 -2.57679671e-01 1.37170303e+00
-1.88542143e-01 -5.43113649e-01 -7.80943453e-01 -1.04098666e+00
-3.19708526e-01 7.24724293e-01 -6.81266248e-01 -2.79221237e-01
-4.10835296e-01 -9.72801924e-01 6.78308547e-01 6.39853299e-01
3.29593867e-01 -8.69401157e-01 -2.81656273e-02 -1.16548926e-01
-3.39210719e-01 -1.43461895e+00 -6.97521985e-01 7.54518569e-01
-7.24595904e-01 -7.08802938e-01 -4.31785971e-01 -1.23207426e+00
8.28747690e-01 1.96782932e-01 1.30290556e+00 -3.94834250e-01
4.06481922e-01 1.72152832e-01 -4.55786467e-01 -5.07696569e-01
-6.79522038e-01 5.25588095e-01 2.26398364e-01 2.41802320e-01
7.84517169e-01 -3.32172036e-01 -5.19895218e-02 2.18675330e-01
-6.67951524e-01 4.00918186e-01 4.64533657e-01 1.15857255e+00
6.81332648e-01 -3.24901879e-01 9.43491459e-01 -1.36422276e+00
6.23249650e-01 -4.80908841e-01 -3.82136613e-01 3.96713614e-01
-6.44836545e-01 1.99979067e-01 1.18092763e+00 -3.06596816e-01
-1.21803081e+00 -2.20359885e-03 -1.18855312e-01 -3.92504960e-01
1.15976371e-01 7.41801441e-01 -3.14187944e-01 4.02079284e-01
4.70603883e-01 4.14775699e-01 -1.77591354e-01 -4.82739896e-01
3.83429170e-01 9.67180848e-01 6.74623996e-02 -4.95270759e-01
4.18339342e-01 3.11391145e-01 -4.91324693e-01 -4.87713754e-01
-1.35044718e+00 -1.74055204e-01 -1.07858717e+00 2.96227962e-01
9.03200388e-01 -1.47055924e+00 -8.58975649e-02 4.11965311e-01
-1.11223960e+00 -2.83094853e-01 -2.27604229e-02 6.27655447e-01
-4.25606251e-01 -6.38094395e-02 -6.09663546e-01 -4.21208918e-01
-4.58143145e-01 -1.42519069e+00 1.23818302e+00 -2.17615783e-01
-1.07706435e-01 -1.55467379e+00 2.67309815e-01 5.36879957e-01
2.85690635e-01 -2.97290802e-01 9.19810772e-01 -9.35605288e-01
-3.15721467e-04 -3.13522100e-01 1.33418456e-01 1.00262761e+00
2.28236318e-01 -2.96392024e-01 -1.21642029e+00 -4.91423041e-01
2.68050581e-01 -9.19004142e-01 8.77274454e-01 1.18931837e-01
9.44803715e-01 -9.89602506e-02 -1.56210646e-01 9.57434833e-01
1.49591088e+00 1.47230709e-02 9.62292254e-02 2.85282373e-01
8.60228240e-01 8.47667754e-01 2.01993659e-01 -9.27245468e-02
8.14179420e-01 5.09603798e-01 -9.68653411e-02 -2.29389712e-01
1.91603228e-01 -4.13031459e-01 8.97609532e-01 1.69765210e+00
3.86264235e-01 -3.19382012e-01 -6.81071877e-01 4.78340209e-01
-1.62246537e+00 -5.69627643e-01 1.90610901e-01 1.98773634e+00
1.04775310e+00 6.12965152e-02 -3.68607163e-01 -3.28173608e-01
4.73125219e-01 1.99072868e-01 -7.11786449e-01 -8.02569151e-01
-3.82050812e-01 2.85981614e-02 4.13377821e-01 4.57675368e-01
-1.14036870e+00 1.26715529e+00 5.70463753e+00 6.65217042e-01
-1.40760922e+00 4.37623084e-01 9.63247597e-01 -1.63006887e-01
-4.63729292e-01 -1.85150102e-01 -1.16280270e+00 2.63528317e-01
1.06296730e+00 -2.95547582e-03 2.24593915e-02 1.08914804e+00
-9.30876955e-02 2.73886293e-01 -1.16709340e+00 7.34308541e-01
6.58167899e-01 -1.00533938e+00 2.24439606e-01 -2.02737257e-01
9.10446167e-01 4.59552109e-01 2.71545440e-01 5.09637952e-01
4.56598580e-01 -8.28595400e-01 7.69549727e-01 2.27141194e-02
1.16125810e+00 -7.38137841e-01 8.29504609e-01 4.29404020e-01
-1.02875280e+00 2.22436830e-01 -4.08291221e-01 1.80786237e-01
1.14454821e-01 2.75818110e-01 -6.09215379e-01 5.31407833e-01
7.04986215e-01 1.17210805e+00 -4.44342315e-01 -2.96125978e-01
-4.12551284e-01 5.29146016e-01 9.55511406e-02 1.04983523e-01
3.14806789e-01 -4.02952284e-01 2.29925271e-02 1.33702850e+00
5.14075197e-02 -6.20059192e-01 1.88906968e-01 4.59040791e-01
-4.08503175e-01 4.27381873e-01 -8.20547640e-01 -3.78637239e-02
6.97597042e-02 1.33484316e+00 -3.44309717e-01 -3.29080582e-01
-1.04141927e+00 1.11293137e+00 7.71353543e-01 3.45176786e-01
-5.36588132e-01 -1.30523264e-01 5.66854119e-01 -5.70640445e-01
3.25475246e-01 -6.07799329e-02 -3.86143565e-01 -1.85842133e+00
2.33144015e-01 -1.18080485e+00 1.05884619e-01 -6.08263671e-01
-1.54222631e+00 1.24552691e+00 -5.32419205e-01 -1.44272614e+00
-4.72732067e-01 -9.83593583e-01 -3.11452925e-01 1.09381783e+00
-1.84072876e+00 -1.67671824e+00 4.23226297e-01 6.75329268e-01
7.62097716e-01 -8.46627235e-01 1.22577500e+00 4.02536362e-01
-8.96172822e-01 1.00386071e+00 4.57767338e-01 4.47168589e-01
1.10745394e+00 -1.14177525e+00 3.60605061e-01 7.01726019e-01
3.05259705e-01 6.46273673e-01 1.95064589e-01 -4.01314467e-01
-1.29347765e+00 -1.24755216e+00 1.23117042e+00 -6.52128160e-01
7.78692603e-01 -8.81136239e-01 -5.38194120e-01 1.24062288e+00
5.28532147e-01 -6.38258308e-02 1.47154844e+00 3.39323133e-01
-7.72094727e-01 -1.35585934e-01 -5.01006365e-01 6.48781955e-01
6.09299242e-01 -1.00264478e+00 -4.01255935e-01 4.65094268e-01
7.64363766e-01 -2.83901066e-01 -6.46300137e-01 2.54565626e-01
6.66541874e-01 -8.54148984e-01 5.77719808e-01 -8.43991518e-01
9.29848552e-01 7.10251406e-02 -6.49536014e-01 -1.63244534e+00
8.39636177e-02 -4.62638699e-02 4.66269225e-01 1.26336074e+00
9.99625444e-01 -7.79597223e-01 5.43814003e-01 1.98012337e-01
-2.16896564e-01 -1.04252219e+00 -6.43075287e-01 -4.73655909e-01
4.65141863e-01 -6.88610137e-01 4.36446577e-01 1.23749864e+00
-2.91712284e-02 1.12181950e+00 -5.37087679e-01 -5.50171770e-02
3.64167482e-01 3.85037154e-01 8.48313689e-01 -7.25945652e-01
-4.59796429e-01 -2.28981957e-01 1.32657334e-01 -1.06684673e+00
8.12033176e-01 -1.44583201e+00 -1.31677672e-01 -1.06543636e+00
2.61264175e-01 -5.54017365e-01 -3.50633591e-01 4.74782020e-01
-8.96370783e-02 3.81555706e-01 -1.32905915e-01 2.27005392e-01
-4.74489123e-01 8.47560525e-01 1.15779924e+00 -1.05080172e-01
1.51840057e-02 -3.26489843e-02 -1.08051062e+00 1.00020635e+00
7.53281593e-01 -6.73350692e-01 -4.74852175e-01 -9.49046910e-01
5.20932436e-01 -9.72294882e-02 -2.17301816e-01 -3.43521535e-01
7.92070180e-02 6.70583844e-02 4.26619589e-01 -3.99659485e-01
4.40888107e-01 -7.70606697e-01 -4.20435756e-01 1.15425229e-01
-7.67629862e-01 3.00078988e-01 7.01798275e-02 3.76951218e-01
-5.75380802e-01 -2.59619445e-01 6.94519460e-01 -1.50563464e-01
-4.52361971e-01 2.41450399e-01 -1.87944904e-01 1.39630273e-01
6.16173089e-01 1.33675575e-01 -2.75129408e-01 -3.77906442e-01
-3.90863150e-01 1.40727863e-01 4.78939414e-01 8.39023530e-01
2.50627309e-01 -1.47084308e+00 -8.25587332e-01 6.05087817e-01
4.65829670e-01 -2.73188859e-01 -7.83956200e-02 6.88031018e-01
-2.19031125e-02 4.83720511e-01 2.02370346e-01 -5.07206619e-01
-1.01362121e+00 2.93535829e-01 6.02095723e-01 -4.98358369e-01
-2.28674755e-01 1.02889872e+00 5.98540425e-01 -1.19953227e+00
-3.75802331e-02 1.10573143e-01 -2.84959733e-01 2.61896014e-01
3.11253309e-01 -4.01296437e-01 1.46284178e-01 -1.08574188e+00
-1.97691947e-01 6.00634575e-01 -4.44361985e-01 -2.17442155e-01
1.40423822e+00 -3.55021656e-01 -1.70444071e-01 9.66378212e-01
1.72365689e+00 1.48961201e-01 -1.00866032e+00 -6.03407681e-01
-2.01215431e-01 7.05400258e-02 2.28921324e-02 -6.35709584e-01
-1.29782331e+00 1.05858958e+00 2.44043097e-01 -1.06016278e-01
1.10746372e+00 -6.04059808e-02 7.82183409e-01 4.70687866e-01
3.02742422e-01 -1.27030706e+00 7.87197649e-02 8.83008659e-01
7.11839914e-01 -1.66151536e+00 -2.54845738e-01 -7.81847164e-02
-1.24411857e+00 9.83388186e-01 7.66547441e-01 -4.57142629e-02
7.94449806e-01 3.62354130e-01 7.84082353e-01 3.02605983e-02
-1.10046077e+00 2.28777379e-01 2.78766990e-01 2.46549800e-01
1.05115187e+00 1.39538318e-01 6.73411563e-02 9.44047391e-01
-6.15316689e-01 -3.21854681e-01 3.42132568e-01 7.84675241e-01
9.94163007e-02 -1.49576306e+00 1.23596422e-01 2.58616328e-01
-7.17830718e-01 -5.31061888e-01 -2.84857005e-01 4.09295201e-01
3.28555927e-02 9.02022481e-01 5.80854081e-02 -5.28052807e-01
1.54964343e-01 4.26080525e-01 1.49531275e-01 -8.94150317e-01
-7.21575856e-01 5.81174716e-02 1.01521425e-01 -3.09684128e-01
-6.28466129e-01 -5.85381389e-01 -7.89892495e-01 1.18504010e-01
-4.60487515e-01 1.87502325e-01 9.74518359e-01 1.17183161e+00
4.51161265e-01 4.25166398e-01 9.90679383e-01 -4.54786330e-01
-6.51795745e-01 -9.60711479e-01 -6.83333755e-01 3.80134523e-01
3.93599212e-01 -2.47022033e-01 -3.12783957e-01 3.25543433e-01] | [11.198049545288086, 9.76653003692627] |
0c413ff2-2308-408a-979a-c141505193de | k-way-p-spectral-clustering-on-grassmann | 2008.13210 | null | https://arxiv.org/abs/2008.13210v2 | https://arxiv.org/pdf/2008.13210v2.pdf | Multiway $p$-spectral graph cuts on Grassmann manifolds | Nonlinear reformulations of the spectral clustering method have gained a lot of recent attention due to their increased numerical benefits and their solid mathematical background. We present a novel direct multiway spectral clustering algorithm in the $p$-norm, for $p \in (1, 2]$. The problem of computing multiple eigenvectors of the graph $p$-Laplacian, a nonlinear generalization of the standard graph Laplacian, is recasted as an unconstrained minimization problem on a Grassmann manifold. The value of $p$ is reduced in a pseudocontinuous manner, promoting sparser solution vectors that correspond to optimal graph cuts as $p$ approaches one. Monitoring the monotonic decrease of the balanced graph cuts guarantees that we obtain the best available solution from the $p$-levels considered. We demonstrate the effectiveness and accuracy of our algorithm in various artificial test-cases. Our numerical examples and comparative results with various state-of-the-art clustering methods indicate that the proposed method obtains high quality clusters both in terms of balanced graph cut metrics and in terms of the accuracy of the labelling assignment. Furthermore, we conduct studies for the classification of facial images and handwritten characters to demonstrate the applicability in real-world datasets. | ['Olaf Schenk', 'Christie Louis Alappat', 'Dimosthenis Pasadakis', 'Gerhard Wellein'] | 2020-08-30 | null | null | null | null | ['spectral-graph-clustering'] | ['graphs'] | [ 9.69480649e-02 9.95035619e-02 -1.43854216e-01 -1.98703453e-01
-4.52760011e-01 -4.00970817e-01 -3.13108088e-03 3.76223251e-02
-2.79213548e-01 3.85084867e-01 -3.49042028e-01 7.77897835e-02
-7.25614548e-01 -5.95921755e-01 -3.91697407e-01 -9.97199833e-01
-4.22388375e-01 3.78931344e-01 -9.13918018e-02 -3.00765969e-02
4.42149282e-01 6.01493061e-01 -1.47823453e+00 -2.67205387e-01
1.14541447e+00 9.41630423e-01 -1.62184969e-01 3.60840708e-01
-2.01939475e-02 1.86803892e-01 -2.47435316e-01 -4.70215172e-01
5.14551103e-01 -4.95108902e-01 -8.85225296e-01 7.62609839e-01
3.12423527e-01 7.48235822e-01 -1.95142791e-01 1.55754948e+00
2.30762407e-01 2.88599283e-01 8.31647515e-01 -1.42744839e+00
-4.64423627e-01 3.56819123e-01 -1.30638111e+00 -9.53082964e-02
1.75345078e-01 -2.95860916e-01 9.71198499e-01 -8.02274704e-01
6.65238202e-01 9.35226619e-01 6.36036217e-01 8.31238255e-02
-1.61646211e+00 -6.79000854e-01 5.14740460e-02 2.96580732e-01
-2.03307009e+00 -1.43346071e-01 1.05590129e+00 -6.09466553e-01
5.27281821e-01 3.07864368e-01 5.68998933e-01 -2.77497824e-02
-1.22575022e-01 2.29984313e-01 9.44379032e-01 -5.82305610e-01
2.53952920e-01 2.15097610e-02 1.64327785e-01 1.10424387e+00
2.52370298e-01 -3.34869385e-01 -4.07302469e-01 -1.51352361e-01
4.89937901e-01 -3.24679971e-01 -3.81385386e-01 -1.02091956e+00
-8.24493170e-01 8.64510298e-01 4.94231552e-01 6.59647286e-01
-2.53501803e-01 1.30652145e-01 -4.33541574e-02 -1.30608566e-02
4.03964281e-01 3.02395999e-01 1.50845140e-01 2.57525623e-01
-1.07462442e+00 -1.23070076e-01 6.94131672e-01 1.05737722e+00
9.25271571e-01 -1.68762002e-02 4.52104151e-01 9.81139660e-01
3.14849854e-01 4.43924606e-01 1.09888554e-01 -1.16773200e+00
3.03906232e-01 9.95719433e-01 -1.27090484e-01 -1.78033936e+00
-3.71191621e-01 -4.19308364e-01 -1.18636262e+00 8.91391858e-02
4.45710927e-01 1.29759610e-01 -4.91017461e-01 1.59885538e+00
4.05027330e-01 2.36401692e-01 -2.06159309e-01 8.88378739e-01
2.04152495e-01 6.43754125e-01 -1.64021879e-01 -8.38060200e-01
9.58498776e-01 -4.89273995e-01 -6.60215795e-01 1.15041770e-01
4.28892016e-01 -8.23360980e-01 8.82836819e-01 6.11760199e-01
-1.02259135e+00 -3.60323817e-01 -9.32786763e-01 5.55199206e-01
-1.05273932e-01 2.19465747e-01 2.82859653e-01 7.84137249e-01
-1.29345834e+00 5.87740541e-01 -7.81391084e-01 -3.65424246e-01
1.94969743e-01 5.03859103e-01 -3.96020919e-01 -2.70979255e-01
-5.55872142e-01 4.07518148e-01 5.96325435e-02 3.91369343e-01
-6.79470375e-02 -3.30598682e-01 -6.77773118e-01 -7.95391425e-02
3.39985281e-01 -1.07304677e-01 3.10608834e-01 -9.66197371e-01
-1.15525103e+00 1.09041774e+00 -9.97031853e-02 8.81976560e-02
3.64934891e-01 4.13260341e-01 -2.99824446e-01 5.42713404e-01
1.97181717e-01 3.45677167e-01 9.62218344e-01 -1.34125972e+00
-3.75862986e-01 -6.69712663e-01 -4.37796444e-01 2.22796172e-01
-4.19429481e-01 -2.38247439e-01 -4.98227984e-01 -5.22219121e-01
6.25566781e-01 -1.00560534e+00 -5.09589911e-01 -2.65140891e-01
-3.99430156e-01 -1.71630681e-01 7.71205783e-01 -4.20128733e-01
1.40421915e+00 -2.29822230e+00 6.99046135e-01 1.09973276e+00
1.83742464e-01 -5.16277961e-02 6.27919659e-02 3.82539988e-01
-3.02350700e-01 1.09828159e-01 -8.40380549e-01 2.43381746e-02
-2.21566483e-01 -3.74700129e-02 3.49823833e-01 1.06539392e+00
-2.76790053e-01 2.42630631e-01 -8.39006484e-01 -7.82893002e-01
1.83967367e-01 3.66854340e-01 -4.81749117e-01 -2.50510693e-01
3.03391337e-01 8.00591782e-02 -2.30275482e-01 3.64389747e-01
7.19262958e-01 -2.29944259e-01 5.16454995e-01 -2.49663547e-01
-1.34413794e-01 -8.59782040e-01 -1.73218548e+00 1.61916697e+00
4.57695685e-02 4.78224069e-01 7.01512635e-01 -1.56313062e+00
1.04719901e+00 5.85843064e-02 1.10375845e+00 -2.71317065e-01
1.46990448e-01 1.21289529e-01 -4.82461825e-02 -3.29248130e-01
2.71239161e-01 -2.60249645e-01 -5.68807460e-02 3.66225958e-01
-1.39485508e-01 -2.98545957e-01 6.69559181e-01 4.76517409e-01
9.80455160e-01 -4.25457567e-01 1.80317283e-01 -1.08553863e+00
8.36682856e-01 -4.28278521e-02 4.44646299e-01 1.78715825e-01
-2.54544467e-01 4.22192067e-01 6.04333401e-01 2.28185341e-01
-6.21496797e-01 -8.93883049e-01 -1.98356718e-01 6.10800624e-01
3.71491492e-01 -1.68262824e-01 -1.23516011e+00 -4.17796880e-01
-5.15818447e-02 3.68586510e-01 -5.29921174e-01 -4.14252728e-02
-5.11223674e-01 -1.07062590e+00 9.60305706e-02 5.86319342e-02
5.37670612e-01 -8.55823755e-01 -4.69408959e-01 1.51309902e-02
-3.27437639e-01 -9.30016816e-01 -6.66674972e-01 -1.64703280e-02
-8.11214626e-01 -1.63265491e+00 -5.99237621e-01 -1.38196754e+00
1.18667364e+00 3.45699102e-01 6.38276160e-01 1.47695765e-01
-6.71709776e-01 7.11727023e-01 -2.46541947e-01 4.04667258e-01
-6.89363852e-02 -9.75236669e-02 9.15140193e-03 6.06572330e-01
1.02820940e-01 -5.57880580e-01 -5.31051278e-01 6.26968443e-01
-8.85560572e-01 -5.80997348e-01 2.19919831e-01 5.76300979e-01
7.94579685e-01 6.56225204e-01 4.13687348e-01 -6.37753308e-01
6.98224902e-01 -1.80081397e-01 -9.29906070e-01 1.32550389e-01
-8.87551725e-01 2.11067237e-02 5.70898831e-01 -4.30912785e-02
-6.55663431e-01 2.64558941e-01 3.27312589e-01 -4.67721254e-01
6.43554106e-02 4.94380504e-01 -1.69717297e-01 -3.95780891e-01
6.27545238e-01 8.34512636e-02 1.74166724e-01 -1.47542760e-01
7.03201711e-01 4.26626235e-01 6.09572649e-01 -4.67332453e-01
8.94583523e-01 6.53732538e-01 3.97333384e-01 -1.25710690e+00
-2.06459671e-01 -9.10714269e-01 -8.01019967e-01 -5.78850210e-01
8.72362196e-01 -2.40106821e-01 -1.00936711e+00 4.41661537e-01
-6.42196238e-01 -6.74288794e-02 -1.99008733e-01 2.71719068e-01
-7.41305590e-01 8.76173556e-01 -4.89340752e-01 -7.72485614e-01
-3.56538035e-02 -1.12650692e+00 4.69839633e-01 4.68533859e-02
2.65117772e-02 -9.50075030e-01 3.68037000e-02 1.48636118e-01
-8.41761008e-02 4.96090859e-01 1.22454393e+00 3.18736844e-02
-2.33856320e-01 -2.19943061e-01 -1.44712165e-01 3.47919554e-01
1.72251090e-01 1.86798126e-01 -3.68723571e-01 -4.89523590e-01
1.83029518e-01 1.89232424e-01 5.45184374e-01 5.84401667e-01
1.04059732e+00 -4.24448550e-01 -4.22187895e-01 4.59066749e-01
1.73562455e+00 2.72234768e-01 3.60132456e-01 -1.92033634e-01
5.70918441e-01 8.23995054e-01 5.90384722e-01 2.88875610e-01
-5.82885891e-02 5.12143910e-01 4.58547831e-01 -3.09289396e-02
1.95363149e-01 2.79084593e-01 -2.12025009e-02 1.17983747e+00
-1.80340692e-01 4.75020818e-02 -1.06591320e+00 6.45777285e-01
-1.97402382e+00 -8.80654156e-01 -2.62014627e-01 2.34613776e+00
5.18076599e-01 -2.32881621e-01 2.72300750e-01 5.56030750e-01
1.24369133e+00 1.71436369e-02 -2.04855308e-01 -2.47343466e-01
9.63297859e-02 2.57982492e-01 4.69238728e-01 7.15515256e-01
-9.46493566e-01 6.45722330e-01 6.18504524e+00 8.58615220e-01
-8.88245940e-01 -1.14341989e-01 6.39607012e-01 -1.66394878e-02
3.62521643e-03 -2.86785364e-01 -2.37370327e-01 3.87131542e-01
5.67313910e-01 -1.52726650e-01 6.11165822e-01 7.08016813e-01
2.12307811e-01 -1.49764925e-01 -7.81163990e-01 1.37077749e+00
2.48864740e-01 -1.10934508e+00 -1.99568942e-01 3.17343861e-01
8.16583037e-01 -4.58487660e-01 1.81141227e-01 -3.64956707e-01
1.32327810e-01 -9.48824167e-01 3.79887342e-01 1.96466133e-01
8.05303395e-01 -1.29548466e+00 3.59582514e-01 1.90813780e-01
-1.54528379e+00 -3.55429947e-02 -2.54272163e-01 2.47841045e-01
1.24642149e-01 6.73093021e-01 -4.00414735e-01 6.06043398e-01
5.96485138e-01 7.10235775e-01 -3.44571441e-01 1.01144457e+00
1.04844414e-01 2.98218578e-01 -2.74860948e-01 4.84636091e-02
3.69331837e-01 -1.13493609e+00 4.41102207e-01 1.14226818e+00
2.68723577e-01 7.04794347e-01 2.48357311e-01 6.22389138e-01
-1.32900979e-02 7.71405399e-01 -5.31388044e-01 -3.93040851e-03
2.80162245e-01 1.29613233e+00 -1.33119941e+00 -8.76862463e-03
-1.31397918e-01 6.88614249e-01 2.60685176e-01 3.60638708e-01
-6.58651829e-01 -6.43909156e-01 4.83460099e-01 2.51684129e-01
2.95402169e-01 -4.28688258e-01 -2.45359123e-01 -8.70772362e-01
2.20173039e-03 -7.53055811e-01 5.62361479e-01 -3.49121660e-01
-1.11339331e+00 6.17749095e-01 2.44859643e-02 -1.05947256e+00
2.70618424e-02 -5.51648021e-01 -2.85301298e-01 4.32257235e-01
-9.07715321e-01 -4.81114686e-01 -2.26130381e-01 8.73210669e-01
1.37241617e-01 1.12340346e-01 5.00165701e-01 3.38238478e-01
-5.79655051e-01 4.94856447e-01 2.29476616e-01 1.53473377e-01
2.45528772e-01 -1.00575244e+00 -4.26154792e-01 1.08517277e+00
1.91220030e-01 4.10939217e-01 6.65595710e-01 -3.82212013e-01
-1.29981327e+00 -8.84771824e-01 4.98261392e-01 1.24828048e-01
6.23068094e-01 -1.29523039e-01 -7.03825891e-01 4.33177680e-01
1.20420687e-01 -8.97546411e-02 6.24824464e-01 -9.68589187e-02
-7.91154727e-02 -3.98083299e-01 -1.44646049e+00 4.89658803e-01
1.09468591e+00 -1.61947444e-01 -9.42887738e-02 3.40216428e-01
1.60671592e-01 1.44274965e-01 -9.95942056e-01 2.67814934e-01
2.07282692e-01 -1.07647657e+00 8.48356783e-01 -1.74784005e-01
-3.25133279e-02 -4.06927854e-01 -1.95199072e-01 -1.29257631e+00
-4.37685132e-01 -6.99769020e-01 4.85981882e-01 9.66061831e-01
2.22381085e-01 -5.63474119e-01 1.03300655e+00 2.66609728e-01
6.47492632e-02 -6.75353110e-01 -1.17859769e+00 -7.09356785e-01
-4.33767177e-02 -2.40171865e-01 6.72600372e-03 1.17166209e+00
6.52680576e-01 2.39763707e-01 3.99146639e-02 1.74313143e-01
1.23541939e+00 2.54821062e-01 3.68382365e-01 -1.19875848e+00
-1.10571533e-01 -6.11544371e-01 -6.86509371e-01 -6.37252212e-01
3.78079802e-01 -1.05040729e+00 2.56684124e-01 -1.28779674e+00
1.58317059e-01 -6.72021031e-01 -8.86925310e-03 1.55559769e-02
1.45476833e-01 4.66383636e-01 1.74091265e-01 7.82190785e-02
-4.47040409e-01 3.08462381e-01 8.71938050e-01 -1.57415017e-01
-3.25928956e-01 -2.24023774e-01 -4.46145207e-01 8.40716660e-01
5.44291139e-01 -3.19854558e-01 -4.32450682e-01 -1.02258340e-01
-1.73750073e-02 1.78650916e-01 -1.95713788e-01 -1.10128462e+00
1.44844651e-01 -1.84190154e-01 -1.00234523e-01 -3.50100696e-01
3.22799653e-01 -1.11022878e+00 2.62050748e-01 5.00364065e-01
-1.55913949e-01 1.13508791e-01 -1.94871321e-01 7.26235032e-01
-3.32324535e-01 -1.99660003e-01 1.30083227e+00 -4.47183251e-02
-4.66090471e-01 3.55087101e-01 -2.31584504e-01 1.02083437e-01
1.37481332e+00 -3.56424779e-01 6.34129494e-02 -3.48925591e-01
-9.69142079e-01 1.96772367e-01 5.64423800e-01 -1.09589718e-01
5.02639890e-01 -1.15121698e+00 -4.36823517e-01 1.33379623e-01
-1.44790709e-01 -1.70175910e-01 6.64825290e-02 1.04768968e+00
-6.88955367e-01 2.22247168e-01 -2.22916007e-02 -8.79497886e-01
-1.39626312e+00 7.37774253e-01 4.33014542e-01 -4.24757525e-02
-3.82734835e-01 7.10286736e-01 9.55143420e-04 -8.36156309e-02
2.22249448e-01 -1.63102433e-01 -1.10279024e-01 1.68860763e-01
4.79326360e-02 8.18662703e-01 3.94565091e-02 -9.61892068e-01
-5.37897468e-01 1.06187999e+00 5.11658669e-01 -2.80469209e-01
1.16598356e+00 -2.05606073e-01 -6.17761791e-01 4.41874154e-02
1.37046313e+00 8.38571638e-02 -8.61813188e-01 6.56529889e-02
1.35362104e-01 -5.53246856e-01 -1.86562091e-01 -1.59697279e-01
-1.50391710e+00 5.76173723e-01 5.72696686e-01 5.88109732e-01
1.29005039e+00 7.65411481e-02 2.14517266e-01 1.05606616e-01
5.26199579e-01 -1.34678972e+00 1.87683076e-01 -4.20701131e-02
6.83697939e-01 -6.58990264e-01 -4.05118689e-02 -9.91172850e-01
-3.94908160e-01 9.59631920e-01 3.28368485e-01 -3.95304799e-01
9.93449926e-01 -1.04241651e-02 1.34619445e-01 -4.23425496e-01
-1.96518004e-02 -3.70056555e-02 1.72449932e-01 4.93808597e-01
3.20887327e-01 1.38483256e-01 -7.80662179e-01 7.89259747e-03
-2.20844164e-01 -4.80893135e-01 5.56080699e-01 6.87253475e-01
-5.76370239e-01 -7.86644638e-01 -6.09391689e-01 3.37102175e-01
-2.74822056e-01 3.14512610e-01 -3.11706871e-01 8.31151426e-01
-3.51531082e-03 1.30947816e+00 -7.56046399e-02 -9.53664705e-02
3.87985766e-01 1.39822483e-01 6.27629399e-01 -4.14827406e-01
-1.93816647e-01 2.86186934e-01 -2.94580132e-01 -7.46881127e-01
-7.65752375e-01 -8.01867068e-01 -1.43132126e+00 -2.66598850e-01
-4.07295674e-01 8.85733724e-01 6.58612430e-01 6.36825323e-01
1.87772155e-01 1.39112324e-01 7.92660296e-01 -6.67470098e-01
-2.48310864e-01 -5.48291087e-01 -1.10388410e+00 5.36948740e-01
-1.53853834e-01 -6.66698337e-01 -5.74012995e-01 3.17151695e-01] | [7.499814510345459, 4.750634670257568] |
0a6e8e29-5cac-4f80-8ab6-070c9a26aa43 | sign-language-translation-from-instructional | 2304.06371 | null | https://arxiv.org/abs/2304.06371v2 | https://arxiv.org/pdf/2304.06371v2.pdf | Sign Language Translation from Instructional Videos | The advances in automatic sign language translation (SLT) to spoken languages have been mostly benchmarked with datasets of limited size and restricted domains. Our work advances the state of the art by providing the first baseline results on How2Sign, a large and broad dataset. We train a Transformer over I3D video features, using the reduced BLEU as a reference metric for validation, instead of the widely used BLEU score. We report a result of 8.03 on the BLEU score, and publish the first open-source implementation of its kind to promote further advances. | ['Xavier Giró-i-Nieto', 'Jordi Torres', 'Amanda Duarte', 'Gerard I. Gállego', 'Laia Tarrés'] | 2023-04-13 | null | null | null | null | ['sign-language-translation'] | ['computer-vision'] | [ 9.69796479e-02 -3.32738638e-01 -5.56341588e-01 -3.81568193e-01
-1.39745152e+00 -7.25101531e-01 9.26714540e-01 -8.19232583e-01
-6.73595726e-01 9.09394264e-01 5.91548443e-01 -1.88694358e-01
4.30705220e-01 -1.23008996e-01 -7.53397346e-01 -4.52342361e-01
1.65524065e-01 6.25676215e-01 5.46933830e-01 -1.61645904e-01
7.36334026e-02 1.29455298e-01 -1.58383095e+00 2.55856752e-01
7.39492655e-01 7.52004266e-01 -3.68462592e-01 6.32376671e-01
-6.88816532e-02 7.36629426e-01 -8.26319516e-01 -6.58208966e-01
3.64886463e-01 -8.56686532e-01 -7.37520516e-01 -2.42568865e-01
1.02400196e+00 -6.06727898e-01 -5.07677257e-01 7.12797225e-01
8.94742846e-01 -2.15031564e-01 7.50835538e-01 -1.18941128e+00
-8.31662893e-01 4.47765470e-01 -5.91788553e-02 -1.71442598e-01
5.37709415e-01 4.73597854e-01 1.05574918e+00 -9.66539800e-01
1.04657638e+00 1.09748662e+00 5.20468056e-01 1.12872231e+00
-5.04320264e-01 -8.96577895e-01 -4.50398885e-02 2.81044602e-01
-1.12642968e+00 -5.48680544e-01 2.04463929e-01 -3.67002577e-01
9.49221551e-01 6.75131381e-02 6.82577133e-01 1.58762097e+00
-2.35262230e-01 1.13810980e+00 1.49387586e+00 -6.86189711e-01
6.19829632e-03 -9.60121229e-02 -1.56185895e-01 6.86239123e-01
2.77686149e-01 1.55952975e-01 -8.76379251e-01 2.94514179e-01
6.08756542e-01 -6.38204813e-01 -2.67781079e-01 -2.26172686e-01
-1.64466298e+00 6.09975517e-01 -3.50675583e-02 1.78675339e-01
1.00049987e-01 2.83139735e-01 4.72740918e-01 9.75787997e-01
1.69384629e-01 2.31153890e-01 -5.38026214e-01 -8.28287780e-01
-8.69378090e-01 3.60088497e-01 9.51332688e-01 1.20415270e+00
9.62727070e-02 -1.27308339e-01 -3.56922805e-01 7.78067589e-01
3.39829862e-01 9.82846916e-01 5.77800691e-01 -1.02299035e+00
7.08781660e-01 3.47034931e-01 9.04947966e-02 7.99002349e-02
1.43820047e-01 8.41770843e-02 -1.32222310e-01 5.00368178e-01
9.44360852e-01 -2.63622940e-01 -1.35687280e+00 1.59253240e+00
-6.04221933e-02 -4.33809832e-02 1.71916541e-02 1.26130950e+00
1.13556564e+00 3.04231763e-01 -1.30418286e-01 1.70953810e-01
9.22905684e-01 -1.25538206e+00 -5.24336934e-01 -5.65378144e-02
7.53864110e-01 -9.94628489e-01 1.55387104e+00 5.83191991e-01
-8.19226623e-01 -1.76736675e-02 -9.91763115e-01 -1.01547681e-01
-2.86164612e-01 3.09525460e-01 6.03630960e-01 7.37706304e-01
-1.18525219e+00 3.94932210e-01 -6.93053305e-01 -9.01162803e-01
1.88338324e-01 1.08165056e-01 -5.10083914e-01 -2.65932530e-01
-1.06429482e+00 1.23196304e+00 -2.64723718e-01 -3.51751983e-01
-7.76029587e-01 -5.57703376e-01 -3.98693442e-01 -8.19869280e-01
1.20630361e-01 -3.08675468e-01 1.50780380e+00 -8.78305912e-01
-2.23248196e+00 1.41937220e+00 1.96386240e-02 -2.63710737e-01
1.07821333e+00 -3.52976918e-01 -5.63193738e-01 -6.49617910e-02
-1.43706203e-01 7.76730239e-01 5.96919179e-01 -6.69852674e-01
-7.57132530e-01 -1.59271378e-02 -9.26831141e-02 1.65332660e-01
-1.11789852e-01 6.37328506e-01 -7.10515201e-01 -5.67195773e-01
-6.81143105e-02 -1.17136586e+00 2.37410590e-01 2.78403848e-01
1.06837481e-01 -3.54234070e-01 5.66242993e-01 -8.20102036e-01
1.06742287e+00 -1.84916008e+00 1.74244925e-01 -2.16415241e-01
-2.28915542e-01 4.63240266e-01 -4.13831085e-01 4.07233834e-01
3.12730402e-01 -6.35564774e-02 -1.64557174e-01 -2.46324778e-01
2.52279013e-01 1.53386593e-01 -2.19947830e-01 4.30721641e-01
-1.47660952e-02 1.01654315e+00 -1.00841653e+00 -4.25568223e-01
1.76531404e-01 2.43564576e-01 -3.97402227e-01 7.48365074e-02
-1.16873287e-01 5.82703292e-01 -1.83608517e-01 8.32751811e-01
3.36531818e-01 6.32590652e-02 8.58464185e-03 1.91009551e-01
-3.91975522e-01 5.74453354e-01 -7.41019249e-01 2.14757729e+00
-3.14989686e-01 1.20754349e+00 -3.86358500e-01 -5.01012504e-01
8.71388912e-01 3.37671459e-01 3.24394435e-01 -8.51576328e-01
8.62878263e-02 8.74393582e-01 -1.35634720e-01 -5.31856894e-01
-1.41673284e-02 1.25357345e-01 -2.51544148e-01 3.47872317e-01
3.17288637e-01 -3.46665025e-01 3.40021282e-01 -1.46879718e-01
1.29102457e+00 6.49719536e-01 9.90001336e-02 -1.20111279e-01
4.51403379e-01 1.73367392e-02 2.47876495e-01 6.44355476e-01
-6.22543454e-01 1.00382102e+00 2.70882010e-01 -3.36377710e-01
-1.22897291e+00 -1.24552608e+00 -5.21120913e-02 1.00421894e+00
-1.06416591e-01 -3.07531118e-01 -7.46153116e-01 -1.26653767e+00
2.05874685e-02 5.19004524e-01 -3.25455129e-01 3.57279390e-01
-8.84711504e-01 -2.83188850e-01 1.15972459e+00 5.67925632e-01
5.96848965e-01 -1.00282073e+00 -3.93400997e-01 -4.60970178e-02
-2.87716806e-01 -1.17965925e+00 -8.17434967e-01 -1.93846539e-01
-6.57448947e-01 -1.00557196e+00 -1.33862352e+00 -1.06979549e+00
1.91423640e-01 -2.21814826e-01 1.15951562e+00 -3.06712538e-01
-3.94445397e-02 3.19580615e-01 -7.56406069e-01 -5.25546789e-01
-4.68194753e-01 4.13315371e-02 4.04457040e-02 -5.03754914e-01
8.14436972e-01 -1.45524263e-01 -4.96014208e-01 6.36944711e-01
-5.13546824e-01 -1.37752011e-01 5.72543144e-01 8.11767459e-01
1.95698828e-01 -1.41712832e+00 3.26801866e-01 -2.26331219e-01
4.55444932e-01 9.54289883e-02 -6.74355030e-01 4.05585855e-01
-7.05267370e-01 4.91369098e-01 2.30780631e-01 -4.94602531e-01
-7.77375042e-01 -6.81606978e-02 -2.00000241e-01 -1.65520698e-01
-1.63852662e-01 1.87411994e-01 -6.06048806e-03 -2.49967098e-01
5.96101403e-01 4.25583422e-01 -1.21791707e-02 -5.87562501e-01
3.91991407e-01 1.09552789e+00 5.97676516e-01 -3.75335962e-01
6.69103503e-01 2.02426091e-01 -2.89092690e-01 -4.79852468e-01
-5.86501658e-01 -3.77498686e-01 -6.62422121e-01 -4.33127642e-01
6.08999848e-01 -9.86056268e-01 -4.40479130e-01 8.02550614e-01
-9.24865425e-01 -6.17728472e-01 -2.78820008e-01 8.93204868e-01
-8.32874954e-01 4.09829229e-01 -3.97950977e-01 -3.35795552e-01
-3.87322992e-01 -9.52295542e-01 1.28375351e+00 -7.50750080e-02
-3.07153046e-01 -5.79968810e-01 6.77549183e-01 4.83104467e-01
5.17257571e-01 2.68403560e-01 1.09605268e-01 -3.99800628e-01
-5.66142023e-01 -1.95429057e-01 -3.52895617e-01 5.91549337e-01
4.15215231e-02 9.15961787e-02 -8.60149562e-01 -2.16294140e-01
-7.01104462e-01 -7.66231835e-01 1.06207418e+00 1.65930659e-01
5.19260943e-01 -3.86314839e-02 -3.55239608e-04 6.32508397e-01
1.15976810e+00 4.59342450e-02 6.66684091e-01 6.61254942e-01
3.90958756e-01 2.12398589e-01 9.07724857e-01 2.00073779e-01
4.99226362e-01 9.62389171e-01 -1.68053299e-01 2.29032919e-01
-9.46160614e-01 -5.53005397e-01 8.78830850e-01 1.15561986e+00
-5.66402137e-01 -3.40369046e-01 -9.98823643e-01 6.97463870e-01
-1.80529881e+00 -8.30403924e-01 3.84876542e-02 2.25174665e+00
1.07489562e+00 -1.12924889e-01 3.85639995e-01 7.78728127e-02
4.05129045e-01 1.76096439e-01 -3.30436260e-01 -4.02762711e-01
-2.97776192e-01 3.21474284e-01 8.10341299e-01 4.52666581e-01
-1.05130541e+00 1.34139204e+00 7.75295353e+00 5.59346855e-01
-1.37980127e+00 2.39467233e-01 -1.63407058e-01 -2.62347013e-01
-3.41542177e-02 -5.19996397e-02 -8.45769644e-01 4.59893435e-01
1.11269128e+00 3.82074192e-02 5.84478974e-01 7.48831391e-01
1.27717048e-01 2.58993119e-01 -1.13082707e+00 1.16255963e+00
3.74767214e-01 -9.75681543e-01 -6.82611763e-02 1.30642340e-01
8.94223988e-01 7.89301693e-01 1.51984379e-01 3.83187652e-01
5.66191077e-01 -8.31559122e-01 1.11513543e+00 4.54096675e-01
1.51558101e+00 -2.31627509e-01 6.77387536e-01 -7.17031509e-02
-9.14767563e-01 2.29151145e-01 -5.28798848e-02 -3.30314063e-03
2.16570660e-01 -7.08670914e-02 -9.04560328e-01 7.59236962e-02
6.75803423e-01 8.92862737e-01 -3.31322640e-01 1.23623025e+00
-6.64690852e-01 1.01496208e+00 -6.50248051e-01 -5.04968524e-01
3.94445509e-01 5.32660671e-02 5.55966318e-01 1.57626605e+00
5.43733954e-01 -3.31792206e-01 -2.63009131e-01 1.67091206e-01
-2.05504045e-01 2.94912905e-01 -6.68601394e-01 -3.96439992e-02
3.06774408e-01 4.87891376e-01 -5.79034500e-02 -3.90890658e-01
-7.42799819e-01 1.18375933e+00 9.11968574e-03 2.62447149e-01
-1.02464235e+00 -3.77286941e-01 9.95670438e-01 -5.36109395e-02
3.63410830e-01 -3.62222403e-01 -2.28048965e-01 -1.51063764e+00
3.80295962e-01 -1.19307745e+00 3.32651675e-01 -5.32321155e-01
-1.12000465e+00 3.93545300e-01 -4.40938100e-02 -1.93262374e+00
-5.52075982e-01 -9.59434807e-01 -2.61298060e-01 5.92396021e-01
-1.80311251e+00 -1.36828554e+00 -3.16572994e-01 5.85737050e-01
6.65120840e-01 -5.07673621e-01 9.58988786e-01 5.02918482e-01
-1.31259367e-01 1.36091387e+00 4.23561305e-01 3.37382227e-01
1.54816306e+00 -1.21586001e+00 7.34020174e-01 7.79506385e-01
2.59209663e-01 2.03250438e-01 6.44863963e-01 -4.73536164e-01
-1.43978155e+00 -6.94396973e-01 1.33856571e+00 -1.12255812e+00
8.51702869e-01 -2.40954116e-01 -1.25452906e-01 6.88688755e-01
1.29723325e-01 -1.31634567e-02 3.09643924e-01 -1.52202874e-01
-7.32841551e-01 -8.13181400e-02 -1.11752582e+00 5.87393224e-01
1.77083945e+00 -4.72840667e-01 -5.70551813e-01 1.89200416e-01
5.32688379e-01 -5.56942642e-01 -9.91490543e-01 3.52237612e-01
1.35080898e+00 -6.58237815e-01 6.16428852e-01 -6.55872047e-01
4.35246319e-01 -2.68421948e-01 -3.43038321e-01 -1.20218050e+00
1.15046725e-01 -7.28886485e-01 -5.23948148e-02 1.00453711e+00
7.49901652e-01 -5.22152722e-01 8.94540310e-01 2.65185028e-01
-3.35169464e-01 -3.67757767e-01 -1.26417768e+00 -1.20750391e+00
3.58476251e-01 -4.31570381e-01 5.80772579e-01 5.45326352e-01
1.33647576e-01 5.68954237e-02 -6.70526803e-01 -4.39154863e-01
7.26075411e-01 1.33121591e-02 1.27005088e+00 -1.08334589e+00
-6.45955354e-02 -5.36918402e-01 -6.60959899e-01 -1.30496454e+00
-1.16800942e-01 -7.40564346e-01 3.99496406e-01 -1.48527932e+00
9.57345068e-02 6.77745938e-02 -2.80514330e-01 6.08832657e-01
1.48485988e-01 6.39648855e-01 3.35700333e-01 3.31193447e-01
-8.05288017e-01 3.80801409e-01 1.53811944e+00 -3.67015064e-01
-4.23559994e-02 -7.77051076e-02 -2.02298746e-01 5.48319340e-01
5.66671252e-01 -2.72560567e-01 -8.04966688e-02 -7.90801108e-01
-1.68824509e-01 -4.76582468e-01 -3.64533812e-02 -1.13155210e+00
3.71297039e-02 -1.65883198e-01 -1.84246808e-01 -4.81295586e-01
1.18773371e-01 -5.41042387e-01 -3.04550588e-01 3.76891643e-01
-4.28579003e-01 -7.84545913e-02 -3.87191735e-02 1.25407547e-01
-3.69130701e-01 2.08410129e-01 8.21412206e-01 3.45901139e-02
-1.11580634e+00 1.82123393e-01 -1.15270786e-01 3.56796354e-01
8.03600848e-01 -4.08471078e-01 -3.92653704e-01 -4.79088157e-01
-4.57542062e-01 8.60840455e-03 6.65503144e-01 8.45097482e-01
5.73434949e-01 -1.58326662e+00 -1.23528218e+00 1.28923789e-01
6.67744458e-01 -7.33927906e-01 -5.39036989e-01 8.04466963e-01
-8.42929959e-01 6.99419200e-01 -4.26762640e-01 -5.06743670e-01
-1.50366032e+00 -2.22495154e-01 1.92607999e-01 -1.41498536e-01
-7.19932377e-01 8.66222024e-01 -5.19823492e-01 -6.99187577e-01
4.80931848e-01 -6.50450110e-01 7.86862820e-02 -2.35774800e-01
5.84807158e-01 4.51691598e-01 8.62037763e-02 -5.84212482e-01
-7.05225587e-01 9.15933251e-01 2.08350629e-01 -6.63996398e-01
1.29778659e+00 1.36598712e-02 3.69365931e-01 3.83959770e-01
1.16526115e+00 -1.70471321e-04 -1.34588909e+00 -3.66305500e-01
1.09712325e-01 -5.90336323e-01 -1.73529789e-01 -1.24618113e+00
-8.21657896e-01 5.28752685e-01 9.01192427e-01 -5.63129365e-01
9.70166028e-01 1.57282978e-01 1.02650213e+00 5.72348356e-01
7.82962441e-01 -1.30713189e+00 -1.13022208e-01 9.33828235e-01
1.12110233e+00 -1.51336563e+00 -3.04577023e-01 1.35117307e-01
-8.17894161e-01 9.03828025e-01 4.03534383e-01 -2.69741476e-01
4.13946182e-01 4.12769079e-01 8.24954331e-01 2.41932556e-01
-5.76856315e-01 -4.15416628e-01 5.39686859e-01 7.83837795e-01
7.95060396e-01 1.88760445e-01 -1.05592024e+00 2.45581821e-01
-3.87878805e-01 6.37135983e-01 2.23952234e-01 8.66879165e-01
-3.17108035e-01 -1.45579171e+00 -2.09437340e-01 2.38227278e-01
-3.13959688e-01 4.26696874e-02 -8.99478376e-01 9.86698985e-01
-9.33419615e-02 7.80860960e-01 -3.29413384e-01 -6.58066332e-01
7.20305860e-01 2.83073902e-01 9.71798182e-01 -2.84127802e-01
-3.68699372e-01 -4.08637911e-01 4.70385879e-01 -7.31638134e-01
-6.67305291e-01 -9.72678661e-01 -9.49824512e-01 -3.04165810e-01
1.41233094e-02 -2.72182763e-01 8.14152002e-01 9.10356641e-01
2.13207621e-02 -2.33304068e-01 3.25550765e-01 -6.01907730e-01
-6.54079258e-01 -1.15698504e+00 -2.62973964e-01 5.19046545e-01
3.36122096e-01 -5.06655097e-01 -3.43922228e-01 2.71437727e-02] | [9.20506763458252, -6.533027172088623] |
80669735-b00c-4fff-b134-b8bd8eadb390 | data-augmentation-and-cnn-classification-for | 2108.07148 | null | https://arxiv.org/abs/2108.07148v2 | https://arxiv.org/pdf/2108.07148v2.pdf | Data Augmentation and CNN Classification For Automatic COVID-19 Diagnosis From CT-Scan Images On Small Dataset | We present an automatic COVID1-19 diagnosis framework from lung CT images. The focus is on signal processing and classification on small datasets with efforts putting into exploring data preparation and augmentation to improve the generalization capability of the 2D CNN classification models. We propose a unique and effective data augmentation method using multiple Hounsfield Unit (HU) normalization windows. In addition, the original slice image is cropped to exclude background, and a filter is applied to filter out closed-lung images. For the classification network, we choose to use 2D Densenet and Xception with the feature pyramid network (FPN). To further improve the classification accuracy, an ensemble of multiple CNN models and HU windows is used. On the training/validation dataset, we achieve a patient classification accuracy of 93.39%. | ['Hongwei Guo', 'Weijun Tan'] | 2021-08-16 | null | null | null | null | ['covid-19-detection'] | ['medical'] | [ 3.87671500e-01 1.50112540e-01 -1.54678211e-01 -2.14294299e-01
-5.62041104e-01 6.97300062e-02 1.88425079e-01 1.83374867e-01
-7.39950061e-01 4.44305509e-01 6.80009797e-02 -4.25126106e-01
5.91031387e-02 -8.92122328e-01 -1.76603243e-01 -7.81907976e-01
-8.73961598e-02 1.74964905e-01 5.57468235e-01 1.46578953e-01
-3.90880585e-01 7.65585005e-01 -9.82394159e-01 4.69365478e-01
5.65976620e-01 1.34129167e+00 2.08146304e-01 7.80004799e-01
1.93658546e-02 1.01429617e+00 -4.42891926e-01 6.45024851e-02
4.63283718e-01 -3.55898052e-01 -6.18292391e-01 -4.41016033e-02
2.25160301e-01 -5.32689869e-01 -6.36666298e-01 9.80256259e-01
8.82380247e-01 6.52944818e-02 4.24204469e-01 -1.00441515e+00
-1.18531950e-01 5.74799418e-01 -4.79049623e-01 8.72406125e-01
-3.28509063e-01 2.97743022e-01 3.90965849e-01 -1.00119853e+00
3.23553562e-01 5.39914310e-01 9.48280692e-01 5.48110604e-01
-9.64477181e-01 -7.87744939e-01 -4.31732267e-01 1.65202588e-01
-1.49753428e+00 1.94720641e-01 5.76916933e-01 -4.05177325e-01
8.33294928e-01 3.30076545e-01 9.20431018e-01 8.99243414e-01
4.67577547e-01 5.71861923e-01 9.66121852e-01 -5.04850090e-01
-9.32018310e-02 -3.06934621e-02 1.60750464e-01 8.07707787e-01
3.48977387e-01 5.34972921e-02 -1.74426530e-02 -1.00994892e-01
1.06838882e+00 4.56323683e-01 -4.57712084e-01 -1.29823506e-01
-1.40736544e+00 8.51242781e-01 8.64277005e-01 6.18156314e-01
-6.44658625e-01 -7.70028308e-02 5.21717787e-01 -8.91563594e-02
1.86215892e-01 4.49485958e-01 -1.71380505e-01 2.82201648e-01
-8.74315500e-01 -8.69907662e-02 3.67663771e-01 7.36359000e-01
3.29929829e-01 4.13762704e-02 -7.71630585e-01 8.02552760e-01
1.63211059e-02 2.55687654e-01 9.48540688e-01 -4.56507057e-01
2.84834892e-01 7.30366588e-01 -5.79332948e-01 -5.58039248e-01
-8.70132267e-01 -8.11572671e-01 -1.41211760e+00 1.48074761e-01
3.04012418e-01 -2.36473128e-01 -1.43691278e+00 1.12549984e+00
1.24223642e-01 2.39811316e-01 -8.91714245e-02 9.19729471e-01
1.23232031e+00 1.52080834e-01 2.42829084e-01 -2.70665020e-01
1.36391509e+00 -8.85000527e-01 -6.82926059e-01 2.34795496e-01
9.54955518e-01 -3.87252480e-01 1.01068413e+00 2.90244251e-01
-9.98228133e-01 -7.29846418e-01 -1.17830229e+00 9.74188298e-02
-2.10105121e-01 4.14986402e-01 5.18253028e-01 6.96909845e-01
-7.78695524e-01 7.11857975e-01 -1.21505082e+00 -2.92711496e-01
7.45040357e-01 6.53483868e-01 -3.40391904e-01 -1.99563116e-01
-1.12659121e+00 9.51778114e-01 6.20010674e-01 -2.70805266e-02
-8.59718025e-01 -9.27817345e-01 -7.22713768e-01 3.07739656e-02
1.78697005e-01 -7.85634756e-01 1.10530221e+00 -5.33991575e-01
-1.12838650e+00 6.75084770e-01 4.16311771e-01 -7.06835508e-01
5.60809791e-01 3.63866717e-01 -4.40479636e-01 3.75362664e-01
-3.29954252e-02 5.49228370e-01 5.28044820e-01 -6.11113906e-01
-6.30464375e-01 -3.40420187e-01 -3.35824400e-01 2.17536792e-01
-5.93517423e-01 1.81268621e-02 -4.60709780e-01 -8.83798480e-01
4.74810004e-01 -7.73178220e-01 -5.19361794e-01 -1.10034170e-02
-3.93193185e-01 1.88522831e-01 8.70668709e-01 -7.74695337e-01
1.32658076e+00 -2.24616671e+00 -4.59289283e-01 6.18735611e-01
7.59160459e-01 1.48745984e-01 6.43120706e-02 -5.71600080e-01
-4.13727075e-01 8.99096727e-02 -2.81310618e-01 7.67600387e-02
-7.59354591e-01 2.41362751e-01 3.77096027e-01 3.51595610e-01
1.82943806e-01 8.89587224e-01 -4.14269149e-01 -7.74436712e-01
3.12110633e-01 4.06905979e-01 -6.81802511e-01 1.92074835e-01
4.08055753e-01 5.41559517e-01 -4.76233035e-01 6.37108982e-01
7.94482470e-01 -3.99326831e-01 -7.71254301e-02 -4.40659404e-01
2.09588185e-02 1.57800298e-02 -9.80046511e-01 1.53875840e+00
-3.17948192e-01 2.68836111e-01 4.37662713e-02 -8.58655035e-01
7.26403117e-01 5.92197120e-01 1.14390695e+00 -4.58591551e-01
5.94693601e-01 1.14584215e-01 5.48732758e-01 -6.45847440e-01
-8.56803805e-02 -3.67166281e-01 2.21061334e-01 7.08605275e-02
1.91731870e-01 -2.35329971e-01 4.53940555e-02 -6.71370998e-02
1.19700158e+00 -5.66377103e-01 3.08133334e-01 -3.58428031e-01
4.73712444e-01 2.62930036e-01 5.95007062e-01 8.09584856e-01
-4.47297573e-01 9.14788961e-01 3.85016918e-01 -5.99760950e-01
-9.18940306e-01 -9.63277102e-01 -5.97017646e-01 5.38508236e-01
-2.89131522e-01 -2.35689402e-01 -7.07884192e-01 -8.68213177e-01
-1.43135250e-01 3.30114007e-01 -9.24306929e-01 -2.46406466e-01
-6.26464427e-01 -1.24711716e+00 7.69598424e-01 1.05754757e+00
7.92592704e-01 -9.15212095e-01 -8.10178995e-01 3.61281455e-01
-1.46293223e-01 -1.04207242e+00 -3.14451903e-01 6.86299324e-01
-1.12860334e+00 -1.23410308e+00 -8.17537069e-01 -8.31828237e-01
6.52432978e-01 -5.10036834e-02 7.30136156e-01 2.70879328e-01
-6.05550289e-01 1.19256854e-01 -2.31542304e-01 -6.57773256e-01
-3.13857138e-01 1.22338571e-01 3.91579121e-02 -3.42309684e-01
3.21446687e-01 -2.73500174e-01 -5.42118788e-01 1.05061039e-01
-8.84742916e-01 1.52951851e-01 7.73843288e-01 1.05420959e+00
7.75937617e-01 1.03209808e-01 2.16749519e-01 -7.05736756e-01
5.04099071e-01 -2.10473478e-01 -1.93817496e-01 7.62245730e-02
-5.26718795e-01 -3.66117388e-01 5.39253414e-01 -5.65826297e-01
-7.62521684e-01 2.78532267e-01 -4.36032981e-01 -8.18017721e-01
-1.59146756e-01 4.33777779e-01 1.85104296e-01 -3.69613320e-01
9.40128624e-01 -5.08200703e-03 2.56172985e-01 -1.68318704e-01
-1.81120321e-01 5.00411868e-01 4.75533456e-01 -6.70973852e-04
5.77079117e-01 3.74513030e-01 2.49557197e-01 -7.93779492e-01
-7.97862053e-01 -3.92909735e-01 -1.11416006e+00 -1.96090296e-01
1.33700025e+00 -8.03518057e-01 -2.41828457e-01 3.68826807e-01
-7.48635709e-01 -2.72400886e-01 -6.48209691e-01 1.20887423e+00
-1.90835625e-01 1.16768710e-01 -9.31673765e-01 -8.51475000e-02
-6.76587284e-01 -1.33068216e+00 3.09829682e-01 2.90349275e-01
1.62836979e-03 -7.03116536e-01 -1.85571015e-02 7.03225583e-02
7.55430460e-01 3.40438157e-01 1.01007152e+00 -1.04023647e+00
-1.82061136e-01 -4.71142620e-01 -3.04215252e-01 6.11771822e-01
1.05253845e-01 -1.79408029e-01 -9.86702323e-01 -1.76386923e-01
2.00389370e-01 -1.06279336e-01 8.73466313e-01 6.97802424e-01
1.82299006e+00 1.51132211e-01 -3.36092561e-01 9.14187670e-01
1.10462213e+00 3.98312211e-01 5.48554301e-01 1.43406481e-01
7.01631665e-01 -2.95396075e-02 1.76816553e-01 3.28898489e-01
-4.83758524e-02 1.29880652e-01 3.42193455e-01 -6.35238171e-01
-4.19457883e-01 1.16057210e-01 -3.95131171e-01 7.54625618e-01
-4.11078990e-01 2.32769921e-01 -1.10926497e+00 2.06141874e-01
-1.13587248e+00 -7.02565074e-01 -2.74350613e-01 1.98327672e+00
5.38282335e-01 3.74567628e-01 -1.48126259e-01 2.94682384e-01
6.72570586e-01 -2.25689068e-01 -3.81593049e-01 2.11887866e-01
-5.07988110e-02 6.66058898e-01 7.24656522e-01 1.60964966e-01
-1.39423633e+00 4.58436787e-01 6.47722483e+00 5.96346140e-01
-1.60373116e+00 3.09307009e-01 8.51467609e-01 -9.66978893e-02
2.75402218e-01 -5.93839586e-01 -5.64439714e-01 -2.11260957e-03
6.62967861e-01 2.10129749e-02 -1.31248340e-01 9.41683710e-01
1.32863566e-01 3.38008888e-02 -6.66204214e-01 9.35245156e-01
9.51607376e-02 -1.30578756e+00 -1.61414102e-01 -4.94857654e-02
4.92621809e-01 2.77458459e-01 -1.15431160e-01 6.08951926e-01
-2.42034242e-01 -1.09204829e+00 1.21202029e-01 2.20916197e-01
1.15450597e+00 -6.06166482e-01 1.11350513e+00 2.92838633e-01
-1.17048180e+00 -3.16859454e-01 -3.21092844e-01 2.16321155e-01
-2.06250280e-01 5.14932930e-01 -1.45745420e+00 4.24325705e-01
9.74822879e-01 4.26740170e-01 -8.56445968e-01 1.30864692e+00
6.79792017e-02 6.07026875e-01 -4.83023793e-01 1.49402380e-01
1.70702115e-01 2.03255862e-01 1.92279637e-01 1.29057384e+00
4.52770799e-01 5.67819238e-01 4.51461285e-01 5.72130919e-01
-2.03147978e-01 3.79026055e-01 -4.79823232e-01 4.63000029e-01
-1.69008002e-02 1.57628357e+00 -1.03201032e+00 -6.76800847e-01
-4.49616581e-01 5.97926438e-01 -2.86709994e-01 3.15117128e-02
-9.28699136e-01 -3.19527984e-01 1.35690719e-01 3.88350934e-01
4.58538011e-02 1.12225942e-01 -6.55765831e-01 -1.07991683e+00
-3.58521104e-01 -5.37983298e-01 7.61402667e-01 -5.95124900e-01
-9.82764065e-01 7.58048713e-01 4.18990059e-03 -1.37735915e+00
1.43904105e-01 -7.57965624e-01 -9.06958401e-01 7.80503154e-01
-1.30731010e+00 -8.41798484e-01 -8.78474712e-01 9.21247065e-01
1.63445100e-01 -1.14351526e-01 8.12604666e-01 5.70065916e-01
-5.44619560e-01 5.40986896e-01 -2.50035971e-01 5.84983170e-01
6.02051377e-01 -1.13369477e+00 -1.90245301e-01 7.95735776e-01
-4.94516611e-01 1.67921349e-01 5.01450859e-02 -6.51118159e-01
-9.31670189e-01 -1.52381480e+00 4.47152287e-01 -1.91815749e-01
4.03285295e-01 -1.19740695e-01 -9.69165921e-01 6.85691714e-01
1.58587061e-02 6.96308017e-01 8.54364932e-01 -5.56806326e-01
1.96912900e-01 7.33201355e-02 -1.50909638e+00 3.35949838e-01
5.73264599e-01 -2.18608618e-01 -5.98773897e-01 3.24902654e-01
6.20428801e-01 -7.57648170e-01 -1.23758125e+00 1.03081250e+00
2.35593662e-01 -6.12847865e-01 9.69454527e-01 -5.85470080e-01
1.24058954e-01 -1.83691859e-01 4.09389436e-02 -1.05733371e+00
-7.16622710e-01 1.34074777e-01 1.84265018e-01 4.42631185e-01
4.77859169e-01 -3.79427284e-01 9.88193512e-01 4.97431040e-01
-4.08905387e-01 -9.81813014e-01 -1.01654100e+00 -3.11603159e-01
2.06523195e-01 -6.57163918e-01 2.22441420e-01 1.00985909e+00
-1.27205059e-01 6.90647736e-02 9.17595774e-02 1.95918381e-01
5.33658445e-01 -3.15557569e-01 2.85747498e-01 -1.25079107e+00
-9.01936442e-02 -4.34338450e-01 -4.53511119e-01 -4.65950847e-01
-4.54485744e-01 -1.16695166e+00 -1.97718456e-01 -1.47469056e+00
3.04439127e-01 -2.90050924e-01 -8.24831307e-01 7.54177094e-01
-3.27921622e-02 7.17207372e-01 1.88467577e-01 8.52039158e-02
-6.86189085e-02 3.31219316e-01 1.79478049e+00 -2.65040129e-01
-3.73192281e-01 3.70985955e-01 -3.92907500e-01 8.28808308e-01
1.01097679e+00 -4.79535699e-01 -1.93969488e-01 6.92867616e-05
-6.66882515e-01 2.58535713e-01 4.40872669e-01 -1.49561274e+00
2.17944220e-01 1.48166955e-01 1.13235605e+00 -9.18954730e-01
1.78826302e-01 -9.05414402e-01 -2.35730633e-01 1.01959085e+00
-2.60257661e-01 1.22593820e-01 3.83084357e-01 7.42311701e-02
-2.52954304e-01 -1.35946885e-01 1.28800893e+00 -3.05553049e-01
-3.49293977e-01 7.90249527e-01 -5.25471449e-01 -1.99036643e-01
1.19349158e+00 -2.03686178e-01 8.71967375e-02 1.77569672e-01
-1.27139807e+00 8.01353976e-02 6.30539358e-02 9.90963355e-02
7.83722878e-01 -1.43263984e+00 -6.19697690e-01 6.00612640e-01
1.21712379e-01 1.57423928e-01 4.15540278e-01 1.49354362e+00
-7.69909739e-01 3.55769068e-01 -4.95216101e-01 -8.24029148e-01
-1.32100022e+00 4.46858108e-01 9.86216128e-01 -5.94345689e-01
-9.59639788e-01 8.84582698e-01 1.77584156e-01 -3.07967573e-01
3.13993394e-01 -8.52676928e-01 -4.43657577e-01 -1.51352942e-01
5.59918106e-01 1.39393300e-01 5.24837434e-01 -2.47020692e-01
-4.17262673e-01 2.65195668e-01 -1.77545920e-01 2.17402056e-01
1.30209351e+00 4.90753680e-01 2.21102014e-02 -1.53115243e-01
1.26387715e+00 -2.75311112e-01 -9.28914309e-01 -2.39706501e-01
-2.39001676e-01 -1.71576425e-01 4.59728390e-01 -7.25061178e-01
-1.41111958e+00 1.03497088e+00 1.22391903e+00 3.15905772e-02
1.31120563e+00 -1.10741779e-01 4.93099123e-01 2.67254889e-01
-3.03596437e-01 -8.76641989e-01 8.45383257e-02 6.29576206e-01
6.99683487e-01 -1.20087433e+00 -5.93421049e-02 -3.62009406e-01
-6.17381394e-01 1.41555941e+00 9.84943867e-01 -1.59719586e-01
1.02979708e+00 7.16701865e-01 1.15544811e-01 -2.64580667e-01
-4.25169140e-01 -1.27755195e-01 2.46118337e-01 4.82091218e-01
5.50810158e-01 -7.24119321e-02 -1.04327179e-01 9.26977396e-01
-6.37161359e-02 2.19217554e-01 5.54817736e-01 8.92581761e-01
-5.05866110e-01 -5.52456379e-01 -5.07724941e-01 9.78148758e-01
-7.51532495e-01 -1.74000084e-01 1.77715033e-01 1.03930581e+00
3.80681694e-01 3.71023864e-01 1.94373727e-01 -4.86001283e-01
5.95381558e-01 1.68541342e-01 3.64332855e-01 -7.05316544e-01
-9.12680268e-01 2.44220600e-01 -3.49901438e-01 -3.70154321e-01
-3.51376347e-02 -3.13997835e-01 -1.51539934e+00 1.20536827e-01
-3.42122525e-01 -9.58625823e-02 5.57933688e-01 6.78478420e-01
-1.62057742e-01 1.21120024e+00 4.54516560e-01 -5.60950518e-01
-5.86186647e-01 -1.23771703e+00 -6.07358932e-01 2.42560104e-01
2.06785753e-01 -5.72086096e-01 -2.39803225e-01 -1.41459465e-01] | [15.20095443725586, -2.1770846843719482] |
b72bec36-1731-4260-8c31-7e469979d2ca | underwater-image-filtering-methods-datasets | 2012.12258 | null | https://arxiv.org/abs/2012.12258v1 | https://arxiv.org/pdf/2012.12258v1.pdf | Underwater image filtering: methods, datasets and evaluation | Underwater images are degraded by the selective attenuation of light that distorts colours and reduces contrast. The degradation extent depends on the water type, the distance between an object and the camera, and the depth under the water surface the object is at. Underwater image filtering aims to restore or to enhance the appearance of objects captured in an underwater image. Restoration methods compensate for the actual degradation, whereas enhancement methods improve either the perceived image quality or the performance of computer vision algorithms. The growing interest in underwater image filtering methods--including learning-based approaches used for both restoration and enhancement--and the associated challenges call for a comprehensive review of the state of the art. In this paper, we review the design principles of filtering methods and revisit the oceanology background that is fundamental to identify the degradation causes. We discuss image formation models and the results of restoration methods in various water types. Furthermore, we present task-dependent enhancement methods and categorise datasets for training neural networks and for method evaluation. Finally, we discuss evaluation strategies, including subjective tests and quality assessment measures. We complement this survey with a platform ( https://puiqe.eecs.qmul.ac.uk/ ), which hosts state-of-the-art underwater filtering methods and facilitates comparisons. | ['Andrea Cavallaro', 'Riccardo Mazzon', 'Chau Yi Li'] | 2020-12-22 | null | null | null | null | ['underwater-image-restoration'] | ['computer-vision'] | [ 5.68793476e-01 -3.89068514e-01 9.77016509e-01 -2.59436995e-01
-4.16005820e-01 -4.99820799e-01 4.32321280e-02 -1.13154380e-02
-8.43349397e-01 6.00715995e-01 4.85388339e-01 1.06767483e-01
-1.45280302e-01 -8.22405219e-01 -6.92130268e-01 -1.20172405e+00
-4.42020833e-01 -4.87898946e-01 1.47810534e-01 -3.97920132e-01
3.88619542e-01 3.23167950e-01 -2.08459139e+00 2.95109481e-01
8.84877980e-01 9.27708209e-01 6.79080546e-01 1.21847737e+00
2.28773355e-01 4.46037471e-01 -7.14211643e-01 -3.49953473e-01
4.64973480e-01 -6.09177113e-01 -3.36216241e-01 2.23245531e-01
8.95961702e-01 -6.46956027e-01 -5.47046423e-01 1.33134067e+00
8.79441857e-01 4.16391313e-01 4.91873801e-01 -3.53782028e-01
-6.34082496e-01 9.81793460e-03 6.33425638e-02 4.62248385e-01
1.84212983e-01 9.00154486e-02 4.89116848e-01 -1.03133941e+00
5.22099137e-01 1.10937715e+00 9.02036786e-01 8.33970428e-01
-8.20285559e-01 -2.47158200e-01 -2.23535702e-01 2.94650912e-01
-6.09826744e-01 -8.38203847e-01 3.34457278e-01 -2.44699076e-01
8.26215386e-01 3.65549088e-01 9.70935762e-01 2.61813223e-01
5.45711696e-01 6.98334515e-01 1.32371080e+00 -4.18704599e-01
2.18312696e-01 -2.31093466e-01 -7.90187493e-02 3.38209063e-01
2.18414038e-01 3.48284036e-01 -5.69962621e-01 2.35117912e-01
6.41918838e-01 -4.28833961e-02 -7.95469582e-01 1.10364377e-01
-4.17660534e-01 4.43436325e-01 4.47401077e-01 -4.76900302e-02
-1.85809448e-01 -2.30460912e-02 2.13613585e-01 7.60935545e-01
6.00969493e-01 5.39938867e-01 -5.50435841e-01 1.57038793e-01
-8.83012056e-01 2.35534430e-01 7.19342470e-01 3.82685810e-01
8.77212048e-01 4.04520303e-01 2.68370330e-01 1.15014899e+00
7.23364055e-01 8.26136351e-01 1.32795468e-01 -1.41956627e+00
-1.84844047e-01 -4.33719456e-02 1.47110954e-01 -5.33421159e-01
-3.19209337e-01 -1.55737326e-01 -5.46161294e-01 1.02845943e+00
1.62410349e-01 -3.14878076e-01 -1.07746994e+00 1.11177576e+00
-5.40060773e-02 -8.52223039e-02 6.59115195e-01 1.02327597e+00
1.32335913e+00 8.21694553e-01 -3.43935758e-01 -5.27421832e-01
1.12946665e+00 -8.24878693e-01 -1.04461980e+00 -4.03559744e-01
8.34305882e-02 -1.05408585e+00 5.48405230e-01 4.91008401e-01
-1.43357086e+00 -4.83414114e-01 -1.16302419e+00 4.45544580e-03
-2.37815961e-01 -1.76436365e-01 3.96643728e-02 7.40284383e-01
-1.32728648e+00 9.79426503e-01 -7.88204908e-01 -2.23143175e-01
2.05481708e-01 2.44789749e-01 -3.09151709e-01 -5.30879080e-01
-9.71595585e-01 1.12002277e+00 -5.72652407e-02 4.38965261e-01
-1.07496917e+00 -8.01236808e-01 -9.68922198e-01 -3.23790073e-01
-4.19235170e-01 -1.46825805e-01 1.36099148e+00 -9.70720410e-01
-1.50192642e+00 6.89234674e-01 -5.71741071e-03 -2.10284278e-01
4.82422084e-01 -4.50756043e-01 -4.05420154e-01 3.87789100e-01
-2.40591615e-01 4.32941347e-01 8.43927622e-01 -1.81082594e+00
-9.95491922e-01 -2.97553897e-01 1.01192892e-01 6.98194802e-01
-4.41918410e-02 1.18414626e-01 -4.43329185e-01 -6.19213700e-01
1.73236817e-01 -5.44998527e-01 -1.37802064e-01 5.88786006e-01
3.10636193e-01 4.89643067e-01 1.11919057e+00 -9.74588096e-01
7.60799885e-01 -2.47144270e+00 8.91723260e-02 -2.14094982e-01
-1.07885920e-01 5.08862257e-01 -3.99358004e-01 7.75915325e-01
1.85922056e-01 -3.79477859e-01 -6.64185643e-01 -5.77757478e-01
-4.12158251e-01 6.48575425e-01 -1.37750953e-02 1.10484874e+00
-2.76870847e-01 1.72740340e-01 -9.13421869e-01 -1.13403380e-01
4.52920973e-01 8.09246004e-01 -3.23849231e-01 5.95552802e-01
2.87051946e-01 3.48738074e-01 4.31865126e-01 7.61037111e-01
1.21205473e+00 7.85526276e-01 -8.16791728e-02 -4.53003049e-01
-7.79013455e-01 -9.47658047e-02 -1.28721476e+00 1.09970081e+00
-4.59816873e-01 1.07992864e+00 9.13077056e-01 -6.11295283e-01
8.18397939e-01 3.02869469e-01 3.72778662e-02 -6.95458710e-01
-1.87812567e-01 5.28941333e-01 3.68627347e-02 -9.92681563e-01
5.89761674e-01 -4.82114583e-01 8.33648682e-01 -4.05919291e-02
4.57814150e-02 -2.63970524e-01 4.17061210e-01 -1.69512957e-01
6.56922340e-01 1.11297585e-01 -9.96097699e-02 -5.47385395e-01
2.34147936e-01 -3.09526384e-01 4.64206904e-01 8.39770317e-01
-2.30440512e-01 7.82199085e-01 -1.91234320e-01 -5.27392507e-01
-9.30630207e-01 -8.84016752e-01 -6.23166621e-01 1.07171547e+00
7.04126656e-01 2.51690120e-01 -7.18823791e-01 2.04346240e-01
-9.73985568e-02 2.82127887e-01 -8.45896900e-01 5.42487309e-04
-4.54449922e-01 -1.12286592e+00 3.61422271e-01 2.71734476e-01
6.75841272e-01 -1.33270872e+00 -7.60062039e-01 3.26856166e-01
-1.87723979e-01 -7.43708193e-01 -2.93739121e-02 3.13233912e-01
-1.28125799e+00 -1.01634586e+00 -8.49602759e-01 -1.07871282e+00
7.33829260e-01 7.15901315e-01 9.96661186e-01 6.49748266e-01
-2.07158476e-01 5.18227637e-01 -6.85421705e-01 -4.58811462e-01
-3.99774313e-01 -1.07618916e+00 8.41677636e-02 -1.64611638e-01
-1.18895121e-01 -4.59538281e-01 -1.06667542e+00 2.05404222e-01
-1.27591002e+00 -4.55333591e-01 5.37260592e-01 8.28167558e-01
2.58062780e-01 1.52289987e-01 -1.67295501e-01 -3.37694079e-01
3.84928167e-01 2.60009076e-02 -4.26680744e-01 -5.98986913e-03
-3.62249076e-01 -3.38873744e-01 4.12092581e-02 -1.33548260e-01
-1.26824415e+00 -1.66446820e-01 -4.64672446e-01 2.53251135e-01
-2.08876058e-01 5.89147747e-01 -1.30008921e-01 -6.73727751e-01
7.69666672e-01 4.43580449e-01 2.20891148e-01 -6.49214327e-01
-2.63558507e-01 6.64682090e-01 6.60727262e-01 4.25193533e-02
6.72592521e-01 8.90393615e-01 -1.77760631e-01 -1.61209893e+00
-6.28346980e-01 -6.73311353e-01 -3.59328926e-01 -8.48974884e-01
7.58977473e-01 -1.04804850e+00 -2.75455207e-01 1.14403713e+00
-9.89563227e-01 -6.84742153e-01 -1.04606420e-01 5.09857118e-01
-4.25829999e-02 6.10700846e-01 -1.01820660e+00 -1.04776990e+00
-4.26977873e-01 -1.25632703e+00 7.53547668e-01 6.35214388e-01
6.58699393e-01 -1.08156097e+00 3.09195757e-01 1.36221096e-01
8.03340614e-01 3.73371430e-02 3.94767895e-02 3.77735943e-01
-1.97674543e-01 5.13059599e-03 -3.30658138e-01 9.22831535e-01
1.77232102e-01 1.46237731e-01 -1.21403754e+00 -6.41094148e-01
7.67038912e-02 -4.47607040e-02 1.29969513e+00 8.48311663e-01
3.95492613e-01 -1.83495373e-01 1.57577187e-01 9.53210771e-01
1.81901920e+00 7.29405582e-02 1.21537721e+00 7.85953522e-01
7.95927644e-02 9.69502091e-01 4.14849162e-01 4.13612723e-01
-1.22420944e-01 2.20028967e-01 9.48696852e-01 -4.40682381e-01
-4.80751097e-01 5.28639674e-01 5.58530807e-01 7.04920590e-01
-6.94834054e-01 -3.04597735e-01 -1.90448612e-01 6.80689633e-01
-1.16778183e+00 -9.13229704e-01 -5.99713504e-01 1.94321346e+00
8.55721653e-01 -4.79244709e-01 -3.92473906e-01 3.05506498e-01
6.02177560e-01 1.27506219e-02 -1.88590407e-01 -2.14786857e-01
-5.40973365e-01 4.82821614e-02 8.26016665e-01 1.04321504e+00
-1.23531961e+00 4.80722278e-01 6.61757231e+00 1.82371140e-01
-8.65129530e-01 -7.74153322e-02 -1.10950693e-02 2.24238843e-01
-4.21698652e-02 -3.79172832e-01 -6.08787000e-01 2.68323749e-01
5.68900883e-01 4.66430604e-01 4.32716966e-01 2.76890546e-01
6.44573510e-01 -3.73269916e-01 -4.56330538e-01 6.02930486e-01
2.20286876e-01 -1.04232049e+00 -2.24582061e-01 1.32735297e-02
9.26739335e-01 3.78600806e-01 -1.28433853e-01 -3.24778348e-01
9.20996908e-03 -6.48532271e-01 8.35604131e-01 6.82016969e-01
7.28127420e-01 -3.26100618e-01 1.25538719e+00 -1.77328125e-01
-9.34782147e-01 -2.57144392e-01 -7.96393931e-01 -5.02371430e-01
2.84967333e-01 4.72153574e-01 2.06147090e-01 1.45855814e-01
1.49002433e+00 7.27534592e-01 -3.65281582e-01 1.86757004e+00
-4.70364690e-01 5.51384747e-01 -3.03810835e-01 1.99439496e-01
1.82282850e-01 -4.46682900e-01 6.94989562e-01 1.68981481e+00
4.15646523e-01 5.54667652e-01 -2.75341362e-01 5.58287948e-02
9.00715142e-02 -1.19811065e-01 -2.14727938e-01 5.24022520e-01
1.62752286e-01 1.11175752e+00 -4.60873514e-01 -1.72379136e-01
-5.84945023e-01 9.86452579e-01 -4.84359115e-01 4.81146693e-01
-3.64291761e-03 -6.82315230e-01 1.10378623e+00 -4.91104536e-02
2.61814624e-01 -2.03379840e-01 -3.38130556e-02 -8.63247633e-01
-1.55711189e-01 -8.10818851e-01 1.58410415e-01 -9.69869375e-01
-1.32019663e+00 6.11030042e-01 -2.03947395e-01 -1.40094900e+00
6.80919170e-01 -9.65708911e-01 -7.21212327e-01 8.67594004e-01
-2.22206903e+00 -6.74275100e-01 -7.89269090e-01 1.64542079e-01
7.96198130e-01 3.01345766e-01 7.54216731e-01 4.03313994e-01
-1.22550227e-01 -4.03373465e-02 8.84076834e-01 8.05017054e-02
7.38496423e-01 -1.41477323e+00 -1.43514900e-02 1.18539977e+00
-5.75715959e-01 2.59145141e-01 1.41557586e+00 -6.27891719e-01
-1.47211516e+00 -7.77906656e-01 5.86100280e-01 1.18195795e-01
4.77596462e-01 2.12968633e-01 -1.23116553e+00 2.01458588e-01
6.70997918e-01 -3.48263495e-02 8.99136066e-01 -7.55041599e-01
2.15628371e-01 -2.12061703e-01 -1.22642756e+00 3.60167593e-01
6.58811092e-01 -4.52450737e-02 -5.16438365e-01 2.04839721e-01
9.60556418e-02 -5.73695302e-01 -8.30960095e-01 4.71436203e-01
9.01294589e-01 -1.30739200e+00 9.28226769e-01 -2.63716206e-02
6.31181479e-01 -6.43529952e-01 -4.31182742e-01 -1.52235174e+00
-3.07348371e-01 -2.15827599e-01 3.03933650e-01 6.74391210e-01
2.85301089e-01 -5.01653850e-01 4.06373978e-01 2.93259650e-01
-6.81955636e-01 -5.21183349e-02 -7.73675442e-01 -4.47644770e-01
-1.34102926e-01 -2.30283946e-01 -2.19692320e-01 3.84625763e-01
-2.47231305e-01 -2.78277874e-01 -5.57472467e-01 8.62197876e-01
1.25445580e+00 -7.32158422e-02 4.21107739e-01 -1.12363219e+00
-6.36893660e-02 -2.37307504e-01 -3.86374593e-01 -1.04041851e+00
-5.28003693e-01 4.90830131e-02 8.02342951e-01 -2.14818621e+00
-1.35787770e-01 2.85605133e-01 9.07466859e-02 1.91369101e-01
-3.27308953e-01 1.02208269e+00 5.88312298e-02 3.77992958e-01
-1.71221942e-01 4.56441432e-01 1.35386121e+00 -9.97900032e-03
-2.20850930e-01 1.78523250e-02 -3.55622351e-01 6.08440101e-01
8.78554821e-01 -2.52209276e-01 1.20386370e-01 -1.02541983e+00
2.52813458e-01 -1.08637877e-01 2.17686117e-01 -1.07420230e+00
2.61651158e-01 9.11109447e-02 4.13790256e-01 -1.66459054e-01
4.07946736e-01 -8.55677068e-01 -1.17376950e-02 8.32932174e-01
-9.33487490e-02 -2.15257168e-01 2.66604394e-01 6.69466257e-01
-4.37180638e-01 -9.05760050e-01 1.53617442e+00 -4.32579815e-01
-1.09108961e+00 -1.15361311e-01 -1.05151641e+00 -4.24689502e-01
3.84387970e-01 -5.08510709e-01 -5.44617653e-01 -4.76607084e-01
-8.44590366e-01 2.12603524e-01 6.80266440e-01 -4.47448231e-02
1.02658689e+00 -4.92942899e-01 -1.16269743e+00 1.17016308e-01
-1.08745389e-01 -2.93421775e-01 5.33516586e-01 7.38416076e-01
-1.38144422e+00 -6.88196361e-01 -3.71779025e-01 -2.25084692e-01
-1.67381465e+00 -1.53057247e-01 9.33674395e-01 3.96820813e-01
-7.37869918e-01 1.15554857e+00 1.52620792e-01 -2.65119523e-01
2.25008458e-01 9.12604630e-02 -6.23569489e-01 -2.50412319e-02
1.08097100e+00 7.93561339e-01 1.71116933e-01 -5.65403819e-01
-1.00421607e-01 8.05145979e-01 3.86294603e-01 3.17833647e-02
1.86218619e+00 -6.44624233e-01 -5.39503813e-01 9.70702618e-02
9.13069427e-01 -1.37885123e-01 -1.92421961e+00 -6.81316182e-02
-7.78734207e-01 -6.53244495e-01 6.12920046e-01 -8.82669449e-01
-1.29751706e+00 9.39056754e-01 1.05169952e+00 5.40333092e-01
1.50096273e+00 -2.90605873e-01 5.20106673e-01 2.92450488e-01
-2.01337621e-01 -1.13668156e+00 -2.13097408e-03 5.22600949e-01
1.03860605e+00 -1.24617410e+00 3.77611130e-01 -3.19943905e-01
-3.26995194e-01 1.56860721e+00 2.42865533e-01 -1.75379246e-01
8.47675204e-01 5.70514381e-01 8.63063812e-01 -2.06028804e-01
-1.03503384e-01 -4.43430692e-01 -3.23189646e-02 9.94386792e-01
2.75554895e-01 -3.74491572e-01 -3.10410023e-01 -1.67861104e-01
7.22027570e-02 -3.07057858e-01 1.00782502e+00 1.15120089e+00
-1.17331755e+00 -7.09026277e-01 -7.19947457e-01 2.40201816e-01
-6.80394948e-01 -3.02459717e-01 9.97764468e-02 2.66283035e-01
3.27660114e-01 1.33158064e+00 -1.26899481e-01 -4.69117351e-02
5.28970778e-01 -5.33696830e-01 3.78986418e-01 -2.89935917e-01
-4.45373148e-01 4.44913596e-01 8.74521881e-02 -4.09859493e-02
-1.29091823e+00 -9.12110507e-01 -9.17241633e-01 -1.01781249e-01
-4.58693653e-01 5.02988398e-01 7.92788923e-01 6.81758523e-01
-3.92206699e-01 5.06600440e-01 5.77039838e-01 -1.59898484e+00
-4.48849909e-02 -1.37614787e+00 -1.05435777e+00 6.96462616e-02
7.91442394e-01 -4.01474565e-01 -1.08809018e+00 5.13782263e-01] | [10.695423126220703, -3.495652914047241] |
e4fc17bd-32ce-4d48-8b26-c07065e18d78 | pe-gan-prior-embedding-gan-for-pxd-images-at | 2303.00693 | null | https://arxiv.org/abs/2303.00693v1 | https://arxiv.org/pdf/2303.00693v1.pdf | PE-GAN: Prior Embedding GAN for PXD images at Belle II | The pixel vertex detector (PXD) is an essential part of the Belle II detector recording particle positions. Data from the PXD and other sensors allow us to reconstruct particle tracks and decay vertices. The effect of background hits on track reconstruction is simulated by adding measured or simulated background hit patterns to the hits produced by simulated signal particles. This model requires a large set of statistically independent PXD background noise samples to avoid a systematic bias of reconstructed tracks. However, data from the fine-grained PXD requires a substantial amount of storage. As an efficient way of producing background noise, we explore the idea of an on-demand PXD background generator using conditional Generative Adversarial Networks (GANs) with contrastive learning, adapted by the number of PXD sensors in order to both increase the image fidelity and produce sensor-dependent PXD hitmaps. | ['Matej srebre', 'Martin Ritter', 'Thomas Kuhr', 'Nikolai Hartmann', 'Hosein Hashemi'] | 2023-03-01 | null | null | null | null | ['conditional-image-generation'] | ['computer-vision'] | [ 2.76364058e-01 -3.57095420e-01 2.41799876e-01 -2.80395877e-02
-1.24482012e+00 -5.68068385e-01 8.62228036e-01 -1.58384934e-01
-5.10192811e-01 7.38661230e-01 2.20895130e-02 -2.29306325e-01
6.40731752e-01 -1.43909192e+00 -1.18292451e+00 -8.56072426e-01
3.71400863e-01 7.72448123e-01 6.51004553e-01 2.89897382e-01
-3.55153412e-01 7.97650397e-01 -1.14900017e+00 2.66411036e-01
2.10060298e-01 9.43222761e-01 1.66600451e-01 9.32104468e-01
-2.47579902e-01 9.94464397e-01 -9.95979548e-01 -3.17600220e-01
6.28658354e-01 -9.07092631e-01 3.69774491e-01 -4.57881719e-01
2.64051080e-01 -4.73360687e-01 -8.06658447e-01 1.16114283e+00
6.56315267e-01 -1.83157489e-01 6.04760170e-01 -1.16109598e+00
-1.05917066e-01 7.43264735e-01 -4.71168429e-01 4.04901475e-01
8.33322033e-02 8.26307774e-01 1.20951205e-01 -3.51257592e-01
7.10402250e-01 9.48100030e-01 9.22798753e-01 7.59956717e-01
-1.30024004e+00 -8.55354249e-01 -7.86344528e-01 -3.11582446e-01
-1.13916266e+00 -2.49180481e-01 7.07062185e-01 -3.94158542e-01
5.83719075e-01 4.77359861e-01 6.52361155e-01 1.60443783e+00
2.37192526e-01 3.44715834e-01 1.26781642e+00 -3.88218552e-01
4.51634735e-01 1.59699142e-01 -1.92938372e-02 4.47269917e-01
6.60277903e-01 8.54410112e-01 -5.37908912e-01 -3.03696066e-01
1.29449213e+00 -1.10194966e-01 -3.04188907e-01 -4.61431444e-02
-9.97222722e-01 6.92839444e-01 2.66071796e-01 2.88522383e-03
-5.02962351e-01 7.30738699e-01 1.93357453e-01 9.87517610e-02
8.69820118e-02 1.76564991e-01 3.17653008e-02 -2.28840619e-01
-8.91755700e-01 3.84353846e-01 5.44089556e-01 9.09446776e-01
5.19799829e-01 2.64312565e-01 -8.81919503e-01 -3.73489037e-02
4.18026000e-01 1.43259716e+00 -1.75395444e-01 -6.15535617e-01
4.60966408e-01 1.68492094e-01 3.99677247e-01 -5.41408122e-01
-1.22176513e-01 -2.53244340e-01 -8.79884243e-01 9.58866537e-01
7.30439663e-01 -3.39224219e-01 -1.40218735e+00 1.47183514e+00
4.03962135e-01 4.17077869e-01 -1.67542160e-01 8.77477467e-01
6.56526566e-01 7.75855422e-01 -1.08505160e-01 -2.67110467e-02
1.15081525e+00 -5.24472117e-01 -8.23581576e-01 -1.19579032e-01
-1.98712368e-02 -7.09565282e-01 8.31327677e-01 1.09358341e-01
-1.31727684e+00 -6.79405808e-01 -1.30376816e+00 -2.70950440e-02
3.00649200e-02 -1.04589656e-01 3.71310890e-01 1.07883620e+00
-6.24087512e-01 4.93611485e-01 -1.14098263e+00 4.54402447e-01
4.98943895e-01 1.36032671e-01 3.27724487e-01 8.47780984e-03
-1.03431296e+00 7.46945202e-01 1.79827094e-01 -2.42752358e-01
-1.31959617e+00 -9.77775633e-01 -5.47071099e-01 -1.98224217e-01
7.13726804e-02 -8.09242845e-01 1.19430256e+00 -3.91282260e-01
-1.54052722e+00 5.34585416e-01 1.71610057e-01 -8.84040654e-01
1.01259995e+00 2.84272164e-01 -4.09039646e-01 1.08357564e-01
-1.04081430e-01 2.16793522e-01 9.91042674e-01 -1.21377969e+00
-3.30726445e-01 -3.06932274e-02 -6.30766749e-01 -1.27676293e-01
6.95432425e-01 9.75701958e-03 -6.45117462e-01 -5.35414100e-01
-2.98934221e-01 -7.17322946e-01 -3.82323831e-01 -8.87821391e-02
-4.78868037e-01 5.09185553e-01 7.55385101e-01 -5.14353454e-01
6.41007900e-01 -2.17740965e+00 -5.54456115e-01 3.10205042e-01
2.19121620e-01 2.72700697e-01 2.28725478e-01 2.21300453e-01
2.89049357e-01 -3.51311594e-01 -1.28470346e-01 -5.90456963e-01
5.66442534e-02 1.83694363e-01 -4.52025145e-01 3.91619503e-01
1.77157447e-02 1.08784187e+00 -8.30087841e-01 -9.29340944e-02
4.96548265e-01 5.38937449e-01 -4.46348712e-02 5.57554625e-02
-4.99274671e-01 8.63690853e-01 -4.01424110e-01 3.12759072e-01
1.11230969e+00 -1.39735445e-01 -4.23539728e-01 -2.74423778e-01
-1.06420755e-01 3.79732519e-01 -1.20232093e+00 1.20019984e+00
3.17634009e-02 3.52802962e-01 1.41127571e-01 -1.08943291e-01
8.24807346e-01 1.36033624e-01 1.97746009e-01 -1.07308614e+00
5.38410664e-01 1.52612209e-01 -2.70034492e-01 1.12017207e-01
5.72943211e-01 -5.00701845e-01 -2.67824769e-01 5.38406789e-01
-2.87175834e-01 -3.60281676e-01 -3.11975181e-01 2.54958689e-01
1.62977743e+00 -1.80869877e-01 -3.69692534e-01 1.07305735e-01
-2.22024858e-01 2.53538340e-01 2.43836597e-01 1.36190116e+00
6.94331154e-02 8.32101822e-01 8.06467533e-02 -2.60266036e-01
-1.68645322e+00 -1.35558856e+00 -9.84349772e-02 -1.52036026e-01
1.64696515e-01 -1.39529914e-01 -9.50430512e-01 -4.56667334e-01
1.29908146e-02 1.02976906e+00 -4.96956825e-01 -1.69793516e-01
-6.38459802e-01 -1.02934980e+00 9.46832240e-01 7.02223420e-01
5.38267434e-01 -1.14898348e+00 -5.52014172e-01 2.58579731e-01
3.47860217e-01 -9.85907972e-01 -4.12963510e-01 4.98184830e-01
-2.85622209e-01 -9.09846663e-01 -4.09676909e-01 4.25299928e-02
7.01068342e-01 -9.37618390e-02 1.33776307e+00 -4.26282078e-01
-3.82803023e-01 3.00431222e-01 -2.42845953e-01 -7.00850308e-01
-1.20567143e+00 -7.80440927e-01 -1.70405507e-02 -1.94365069e-01
3.51438552e-01 -2.77975470e-01 -4.27121222e-01 -4.03199755e-02
-8.95990849e-01 6.85484335e-02 3.71449918e-01 4.50077236e-01
1.16566980e+00 3.29133600e-01 -1.93307221e-01 -8.41406524e-01
2.01218799e-01 -1.31278262e-01 -1.37329865e+00 -3.11961740e-01
-5.83858881e-03 1.12651423e-01 7.62536049e-01 -4.20262396e-01
-1.04606318e+00 -1.05333485e-01 -3.72782379e-01 -8.08672309e-01
7.32810721e-02 -5.39177179e-01 -4.75336611e-01 -2.42322683e-01
8.82094562e-01 1.57772437e-01 -2.68969029e-01 -2.04410404e-01
4.07350987e-01 1.39373168e-01 8.94277632e-01 -3.35468769e-01
1.13930464e+00 7.45967627e-01 2.79469728e-01 -5.29157460e-01
-3.51547986e-01 5.51024266e-02 6.38610572e-02 -1.69093400e-01
1.14477539e+00 -9.73553181e-01 -8.33382368e-01 8.29811990e-01
-9.81977880e-01 -7.47626901e-01 -1.13736606e+00 5.37000179e-01
-3.73117387e-01 1.76546630e-02 -8.31685901e-01 -8.38889480e-01
-1.74741611e-01 -1.12011075e+00 1.11871076e+00 3.65971088e-01
1.41596124e-01 -4.30886030e-01 1.59926370e-01 -9.59693417e-02
2.96581447e-01 5.70434928e-01 4.45169151e-01 -1.90811008e-01
-1.37521815e+00 -5.00076234e-01 -1.52156381e-02 4.13614124e-01
-1.96435377e-01 -2.62860537e-01 -7.98631608e-01 -9.51791927e-02
5.87282121e-01 3.02866958e-02 8.64208341e-01 8.51272941e-01
8.96221757e-01 1.19796492e-01 -6.80459380e-01 6.44408524e-01
1.64789009e+00 4.28377032e-01 9.18963015e-01 -1.39841512e-01
8.47932637e-01 -3.90252858e-01 2.26233035e-01 3.43268514e-01
-1.76039070e-01 6.32611156e-01 2.04385430e-01 -1.50388241e-01
-8.19482803e-01 -5.78853488e-01 2.85500675e-01 3.13640207e-01
2.60163456e-01 -7.32887328e-01 -6.72616363e-01 -3.86089198e-02
-1.24492002e+00 -1.32163453e+00 -7.81584263e-01 2.54711103e+00
6.57042086e-01 4.32773978e-01 -4.63245884e-02 -6.89598098e-02
8.19225073e-01 -9.40751508e-02 -5.31953096e-01 -3.60696688e-02
-3.27017039e-01 6.51404738e-01 1.34935856e+00 5.08019269e-01
-7.17765033e-01 4.73986834e-01 7.20095158e+00 8.76272976e-01
-8.00076604e-01 3.07278514e-01 5.10793507e-01 -3.13893497e-01
-7.07609832e-01 -3.31132680e-01 -1.43801928e+00 1.28764391e+00
1.03510654e+00 4.59629923e-01 5.08837819e-01 4.36618626e-01
-1.17364163e-02 -3.64585757e-01 -7.80086577e-01 1.13180578e+00
-1.97452307e-01 -1.63936925e+00 -1.95270270e-01 1.85767651e-01
9.04249251e-01 3.07658494e-01 2.96811946e-02 1.48170635e-01
9.92367208e-01 -7.41533756e-01 9.00927842e-01 9.04631615e-01
7.98066258e-01 -7.40131676e-01 4.57025558e-01 5.37194848e-01
-6.92383230e-01 4.22502011e-01 -4.16844428e-01 2.43597686e-01
7.61406720e-01 1.25698698e+00 -8.28306556e-01 2.51439363e-01
6.51926100e-01 -3.51636440e-01 -1.97830334e-01 8.19636405e-01
-1.76422849e-01 9.62474942e-01 -9.11986530e-01 -9.96722057e-02
3.97373550e-02 -4.67539042e-01 8.06522965e-01 1.03117156e+00
5.19712448e-01 1.17791183e-01 -1.92616120e-01 1.20713699e+00
-3.38793188e-01 -9.26111758e-01 -7.53173232e-01 6.49183318e-02
5.60284495e-01 8.65051687e-01 -8.18227112e-01 -4.17810470e-01
-4.60514098e-01 1.04623055e+00 -2.55591869e-01 9.92730781e-02
-1.34795761e+00 2.00864241e-01 4.71636802e-01 5.24685323e-01
6.62458122e-01 -1.72065496e-01 -3.69730115e-01 -7.16969132e-01
1.26627982e-01 -8.03504705e-01 -9.23125967e-02 -6.89122856e-01
-1.54979587e+00 4.26510692e-01 -2.52064139e-01 -9.86878753e-01
7.95330703e-02 -4.43593919e-01 -5.46194613e-01 1.10355043e+00
-1.03320372e+00 -9.00339246e-01 -5.13236284e-01 5.96278548e-01
-2.79592752e-01 9.44267809e-02 6.80173516e-01 5.00629008e-01
-8.20364133e-02 3.73632401e-01 3.84954482e-01 2.37580881e-01
4.19460267e-01 -1.19429994e+00 1.19653010e+00 9.71489668e-01
2.66308695e-01 -1.47839859e-01 9.92402017e-01 -1.25073755e+00
-1.44574022e+00 -1.32425070e+00 1.45594239e-01 -1.02916038e+00
3.00537944e-01 -8.01846027e-01 -5.49189329e-01 6.94777310e-01
-3.88954580e-02 4.79554802e-01 2.29418278e-01 -7.52989173e-01
-1.85960047e-02 -2.22767517e-02 -1.50030208e+00 5.09516060e-01
1.05981076e+00 -5.96975207e-01 -1.75780937e-01 3.78520429e-01
6.04647756e-01 -9.43490803e-01 -2.54765183e-01 3.27967465e-01
-1.25186279e-01 -1.00756943e+00 1.21188664e+00 8.74211788e-02
-4.39818464e-02 -8.49977851e-01 -4.29613926e-02 -9.70400035e-01
-2.76735604e-01 -6.14908099e-01 -1.93527818e-01 1.23662353e+00
2.21036181e-01 -4.93214399e-01 1.10519624e+00 5.10554850e-01
-3.23320270e-01 1.58063099e-02 -1.03052270e+00 -8.16652119e-01
-1.98677450e-01 -6.88119709e-01 6.72470450e-01 3.45922917e-01
-9.62454200e-01 2.24024635e-02 -4.70803499e-01 5.78112662e-01
8.89051497e-01 -1.36470094e-01 8.79622161e-01 -7.02842116e-01
-7.83091426e-01 -3.43555436e-02 -3.85828823e-01 -9.83583450e-01
-6.36819661e-01 -7.21396327e-01 3.07910979e-01 -1.19767880e+00
4.56753559e-03 -7.71865904e-01 -2.43556853e-02 -7.79880434e-02
-1.43121868e-01 6.10007823e-01 1.18593037e-01 -2.53678989e-02
-2.42839381e-01 2.43745297e-01 1.22732210e+00 -6.94851577e-02
1.06031574e-01 3.10730964e-01 -2.27534883e-02 5.64871192e-01
5.70416510e-01 -9.01184678e-01 -1.81515768e-01 -2.77199805e-01
3.80177736e-01 3.25099289e-01 9.16366577e-01 -1.49142003e+00
1.66156203e-01 3.68090928e-01 1.04582596e+00 -1.06797934e+00
6.08066916e-01 -8.14014852e-01 1.08150196e+00 3.72388422e-01
2.17786402e-01 -1.67889863e-01 3.88389289e-01 7.19001353e-01
-2.87180115e-02 -4.12917346e-01 9.55215454e-01 -4.39688534e-01
-7.47068450e-02 4.03348207e-01 -3.62677664e-01 1.08295262e-01
1.10278213e+00 2.29231827e-02 -3.46210539e-01 -7.57455304e-02
-3.22583854e-01 -4.34774190e-01 8.90233517e-01 -1.38840467e-01
9.92539823e-02 -1.48378623e+00 -4.35035437e-01 4.99803394e-01
-3.66905302e-01 2.30606779e-01 3.46988052e-01 2.70234913e-01
-8.76060069e-01 1.93991601e-01 6.71712160e-02 -4.18793440e-01
-7.68571436e-01 7.59559691e-01 3.84104282e-01 -5.52926362e-01
-1.06598258e+00 1.11199367e+00 -1.09431319e-01 -3.19726020e-02
-2.47253031e-02 -4.51442420e-01 7.68042028e-01 -5.03088653e-01
8.36052477e-01 2.82245576e-01 1.29722506e-01 -9.23116207e-02
4.99947816e-02 2.25680098e-01 4.13615614e-01 -3.82263482e-01
8.48079562e-01 2.68232226e-01 5.00900805e-01 3.32030386e-01
5.81225455e-01 6.47544503e-01 -1.88036513e+00 1.60149366e-01
-5.58315873e-01 -3.43715221e-01 4.66905236e-02 -1.07130349e+00
-9.54560041e-01 5.58319330e-01 7.16839731e-01 2.13401526e-01
9.60168004e-01 1.46741480e-01 1.10061276e+00 -2.08741024e-01
8.26038420e-01 -7.23085701e-01 -2.70878702e-01 2.52694905e-01
2.39483267e-01 -9.05749261e-01 -2.82567590e-01 -2.10664362e-01
-5.03988922e-01 5.29088736e-01 2.30745420e-01 -3.91658783e-01
2.98614383e-01 1.35916436e+00 1.08267948e-01 -3.81682605e-01
-1.23020798e-01 9.40275714e-02 -1.65784985e-01 7.77896047e-01
-6.24931693e-01 2.19294518e-01 2.80702978e-01 5.08422554e-01
-8.25558007e-02 -7.75146037e-02 2.89786279e-01 7.43990481e-01
-7.29422197e-02 -1.34938800e+00 -7.47930408e-01 4.47856694e-01
-2.04938695e-01 -2.36637771e-01 -1.50872068e-02 6.99523270e-01
4.05919135e-01 4.53338504e-01 3.80712211e-01 3.24308760e-02
3.01509321e-01 -8.75740219e-03 8.95898938e-01 -3.22305143e-01
-7.67603934e-01 -1.03522189e-01 -2.89083961e-02 -5.10329783e-01
-4.43588244e-03 -7.69917727e-01 -1.25481069e+00 -4.01700526e-01
-2.68830091e-01 -1.15182474e-01 7.49164999e-01 5.05291402e-01
2.38096938e-01 9.34734583e-01 3.02544028e-01 -5.37850499e-01
-4.89866287e-01 -7.47310400e-01 -6.81648016e-01 4.02312189e-01
2.22774461e-01 -3.80424768e-01 -5.23525894e-01 -3.28088775e-02] | [9.288627624511719, -3.5184519290924072] |
a05504da-6add-4dcb-a6c4-d300ed72094e | learning-opinion-dynamics-from-social-traces | 2006.01673 | null | https://arxiv.org/abs/2006.01673v1 | https://arxiv.org/pdf/2006.01673v1.pdf | Learning Opinion Dynamics From Social Traces | Opinion dynamics - the research field dealing with how people's opinions form and evolve in a social context - traditionally uses agent-based models to validate the implications of sociological theories. These models encode the causal mechanism that drives the opinion formation process, and have the advantage of being easy to interpret. However, as they do not exploit the availability of data, their predictive power is limited. Moreover, parameter calibration and model selection are manual and difficult tasks. In this work we propose an inference mechanism for fitting a generative, agent-like model of opinion dynamics to real-world social traces. Given a set of observables (e.g., actions and interactions between agents), our model can recover the most-likely latent opinion trajectories that are compatible with the assumptions about the process dynamics. This type of model retains the benefits of agent-based ones (i.e., causal interpretation), while adding the ability to perform model selection and hypothesis testing on real data. We showcase our proposal by translating a classical agent-based model of opinion dynamics into its generative counterpart. We then design an inference algorithm based on online expectation maximization to learn the latent parameters of the model. Such algorithm can recover the latent opinion trajectories from traces generated by the classical agent-based model. In addition, it can identify the most likely set of macro parameters used to generate a data trace, thus allowing testing of sociological hypotheses. Finally, we apply our model to real-world data from Reddit to explore the long-standing question about the impact of backfire effect. Our results suggest a low prominence of the effect in Reddit's political conversation. | ['Francesco Bonchi', 'Corrado Monti', 'Gianmarco De Francisci Morales'] | 2020-06-02 | null | null | null | null | ['link-sign-prediction'] | ['graphs'] | [-8.64328668e-02 3.04387987e-01 -5.53374738e-02 -2.20700249e-01
1.39927249e-02 -6.77838624e-01 1.15921986e+00 4.81776714e-01
-3.43611121e-01 7.93117762e-01 1.90285519e-01 -6.93107426e-01
-2.66514361e-01 -1.15836322e+00 -5.82100868e-01 -7.94429958e-01
-1.09355822e-01 8.30892146e-01 7.84624219e-02 -4.32338893e-01
2.12094292e-01 7.08421096e-02 -1.61840510e+00 -1.60336286e-01
1.12584579e+00 3.56982321e-01 3.16435695e-02 7.16619790e-01
8.83345008e-02 8.09693038e-01 -7.78386831e-01 -6.75814211e-01
-1.46156270e-02 -6.99688613e-01 -4.50214893e-01 2.00921953e-01
-1.72054529e-01 -1.07959852e-01 -1.16198249e-01 9.59375262e-01
2.77534038e-01 -1.46487877e-01 8.95587087e-01 -1.51451075e+00
-4.58440065e-01 7.53287137e-01 -8.94451365e-02 1.93808720e-01
3.77929449e-01 3.47405791e-01 1.15202546e+00 -2.25333422e-01
1.01484263e+00 1.39706481e+00 3.65211427e-01 2.49109253e-01
-1.40818250e+00 -2.90829778e-01 3.85220259e-01 1.62042931e-01
-9.84599292e-01 -3.34401011e-01 6.32342160e-01 -6.66535974e-01
4.52183872e-01 2.05750421e-01 1.28225303e+00 1.43321168e+00
3.12782466e-01 6.07087553e-01 1.42419994e+00 -3.68733644e-01
6.75218344e-01 3.28922451e-01 3.62733334e-01 6.90160573e-01
4.58499640e-01 2.94575304e-01 -5.40917993e-01 -8.35836947e-01
3.71123672e-01 -1.17420666e-01 -1.09416425e-01 -1.77525327e-01
-7.92717934e-01 1.05935621e+00 5.48673719e-02 3.68586272e-01
-6.37273431e-01 1.78259462e-02 8.23291913e-02 6.44187927e-01
8.71236801e-01 2.88011461e-01 -5.66623151e-01 -2.58938223e-01
-4.71141577e-01 4.94350374e-01 1.26548839e+00 3.76826674e-01
8.98658395e-01 -3.27539355e-01 4.77746539e-02 2.73311108e-01
7.14737594e-01 8.38411689e-01 1.61507353e-01 -8.08512688e-01
-1.11958608e-01 7.98836350e-01 4.36185479e-01 -9.82031345e-01
-2.93199360e-01 -2.31621355e-01 -5.93902588e-01 1.50903508e-01
6.21798277e-01 -6.41984642e-01 -5.81668258e-01 1.88485944e+00
5.94954133e-01 3.01567376e-01 1.58752874e-01 5.83788037e-01
1.06605984e-01 6.21996582e-01 -7.60301277e-02 -7.07054794e-01
1.24220634e+00 -2.09843814e-01 -6.26535237e-01 -1.67295933e-01
6.06392622e-01 -3.80368501e-01 8.41391206e-01 1.92879662e-01
-8.26943099e-01 -9.13883522e-02 -5.91528356e-01 7.08103478e-01
-2.51055241e-01 -3.95110399e-01 8.45235348e-01 7.07529843e-01
-1.00057244e+00 5.12288213e-01 -7.63423800e-01 -5.64870715e-01
-8.88407901e-02 2.58343309e-01 8.13148990e-02 3.87862474e-01
-1.33223987e+00 7.86905527e-01 6.54307231e-02 1.09733373e-01
-6.88052058e-01 -3.69046539e-01 -3.88500065e-01 -1.32808298e-01
5.41539490e-01 -1.00460291e+00 1.25235832e+00 -1.37675250e+00
-1.73136079e+00 5.52706540e-01 -4.86148477e-01 -5.68375826e-01
6.15218937e-01 1.64104730e-01 -3.05859089e-01 -1.53653890e-01
-4.49367911e-02 -2.67848396e-03 8.77512634e-01 -1.39720380e+00
-6.56697869e-01 -2.84156412e-01 3.88697296e-01 -7.26447850e-02
-6.59450963e-02 -1.97558090e-01 -2.34358888e-02 -2.32616231e-01
-3.42417240e-01 -1.38672304e+00 -3.27139139e-01 -5.50051987e-01
-2.75385201e-01 -4.41001743e-01 4.15440708e-01 -2.90287852e-01
1.30355120e+00 -1.93676686e+00 3.93998384e-01 5.79058766e-01
2.12247983e-01 4.23799828e-02 -5.21525443e-02 9.10416603e-01
3.56977463e-01 3.57265860e-01 -2.78601378e-01 -3.97669822e-01
1.45347565e-01 3.44557405e-01 -3.80835116e-01 5.51496685e-01
-5.74747808e-02 8.33998978e-01 -9.50194240e-01 -1.52059495e-01
1.67116281e-02 2.84156531e-01 -5.52552998e-01 2.76526660e-01
-7.20061541e-01 5.43558776e-01 -5.11791885e-01 9.00055394e-02
2.45912060e-01 -2.91761577e-01 7.81888843e-01 3.03080916e-01
-1.57583356e-01 2.29088992e-01 -1.04708254e+00 7.54602730e-01
-4.40611869e-01 5.90738297e-01 9.42899659e-02 -7.30607748e-01
7.34423220e-01 2.89940834e-01 3.38386774e-01 -3.25609237e-01
3.72612149e-01 2.16881484e-01 5.25355756e-01 -7.16861188e-01
1.99329823e-01 -1.69582605e-01 -9.12621850e-04 9.78946865e-01
-2.22005606e-01 -1.96260512e-02 4.17552173e-01 2.87341654e-01
1.07445312e+00 -1.37111887e-01 4.25909609e-01 -3.79696876e-01
4.30815548e-01 -1.26610026e-01 4.90204036e-01 9.16576505e-01
1.78105354e-01 -5.13152853e-02 9.40353155e-01 -4.27148521e-01
-1.02637255e+00 -7.34097719e-01 1.45729169e-01 8.44159842e-01
5.87163679e-02 -5.64841270e-01 -8.12899172e-01 -6.52198792e-01
1.12633206e-01 9.13358510e-01 -9.57639039e-01 -2.86725461e-02
-4.60103422e-01 -1.14687061e+00 2.41087098e-02 -2.66040742e-01
4.35324833e-02 -1.19839227e+00 -5.70709884e-01 2.77266860e-01
-2.89911330e-01 -6.79455638e-01 5.59029356e-02 -2.55709171e-01
-6.02042437e-01 -1.40154874e+00 -3.02690528e-02 -3.62840183e-02
8.49730253e-01 -8.41018781e-02 1.21337020e+00 3.91466022e-01
1.86940387e-01 6.32403314e-01 -4.32553232e-01 -6.08060479e-01
-9.56821203e-01 2.28645846e-01 2.21282691e-01 4.90255594e-01
3.70112002e-01 -7.86306739e-01 -6.06095791e-01 2.82845765e-01
-9.88180161e-01 -1.63745224e-01 4.46586818e-01 5.88708699e-01
-3.70098501e-02 6.52959570e-02 5.57371140e-01 -1.20630026e+00
1.19335496e+00 -8.37855458e-01 -8.22357357e-01 3.28539699e-01
-9.08292592e-01 1.94266379e-01 4.46385711e-01 -5.96548080e-01
-9.95152891e-01 -4.99129325e-01 1.27159417e-01 2.07103357e-01
-1.87743053e-01 8.37426007e-01 1.79241002e-01 6.22949600e-01
6.34920061e-01 8.69178176e-02 1.92010388e-01 -1.78682998e-01
3.59863430e-01 6.17727160e-01 -2.09098980e-01 -5.82085013e-01
9.02290642e-01 7.74614036e-01 -1.90137580e-01 -1.02545512e+00
-7.34141886e-01 -3.88525650e-02 -4.61793393e-01 -3.41893762e-01
6.66620612e-01 -5.37813604e-01 -1.13712108e+00 5.56148291e-01
-1.15953290e+00 -5.92137396e-01 -2.80126214e-01 3.54461491e-01
-5.13328850e-01 1.42056316e-01 -3.92605126e-01 -1.44284749e+00
-4.20242660e-02 -9.11470056e-01 8.18407595e-01 1.84740543e-01
-5.25507152e-01 -1.54774153e+00 6.77597404e-01 1.20330505e-01
3.34199220e-01 3.82671691e-02 9.57813025e-01 -6.80866897e-01
-6.39094293e-01 -8.75407606e-02 3.75143617e-01 -1.54819682e-01
1.10351384e-01 6.80652797e-01 -7.32019484e-01 -3.58749211e-01
1.03502765e-01 1.93478599e-01 5.55830538e-01 5.40256262e-01
2.64526546e-01 -5.84335566e-01 -3.75759989e-01 1.01127001e-02
1.08599830e+00 1.94722667e-01 3.93037587e-01 1.47312447e-01
1.85982019e-01 9.95843589e-01 3.53350163e-01 4.60851550e-01
6.67569995e-01 6.95859671e-01 3.26026857e-01 2.03187421e-01
4.95136231e-01 -3.80816638e-01 8.43003929e-01 8.79378796e-01
-2.83957034e-01 -6.80574954e-01 -8.07787478e-01 4.82515395e-01
-2.29512596e+00 -1.16747057e+00 -2.93208778e-01 2.11948609e+00
6.21012926e-01 3.63302976e-01 5.68105996e-01 -2.47007370e-01
5.72613060e-01 1.65782616e-01 -4.28703636e-01 -2.59318739e-01
-1.24249406e-01 -8.18617195e-02 2.32773095e-01 1.05280781e+00
-5.51085949e-01 7.68248260e-01 6.10459805e+00 3.35783184e-01
-9.73136246e-01 7.34184682e-02 4.54209059e-01 1.73712298e-01
-7.80376554e-01 4.61067319e-01 -6.91525459e-01 5.68720698e-01
1.32292640e+00 -4.42888916e-01 4.94713813e-01 2.54953921e-01
7.07556725e-01 -4.65592116e-01 -1.01570833e+00 3.93032759e-01
-1.49780259e-01 -1.21584296e+00 -1.83939755e-01 6.16967559e-01
7.23897636e-01 -6.20728098e-02 -2.65946567e-01 1.00881115e-01
9.04671073e-01 -6.93345189e-01 8.26256752e-01 7.50356793e-01
5.50065131e-04 -5.16031682e-01 8.46116781e-01 7.38959789e-01
-6.62151992e-01 -2.68295377e-01 -7.69350231e-02 -5.72992384e-01
3.91664416e-01 9.72096741e-01 -9.22010839e-01 2.96986789e-01
1.41828880e-01 5.36963046e-01 -5.49772263e-01 6.21663511e-01
-6.07910156e-01 1.11438131e+00 -4.35553700e-01 -3.96769196e-01
2.47593243e-02 -8.12871218e-01 7.61785448e-01 7.78930008e-01
2.57468402e-01 5.84167521e-03 2.84157433e-02 8.45263124e-01
2.51076818e-01 2.25483556e-03 -9.47414815e-01 -2.13598013e-01
4.96066570e-01 1.12685180e+00 -7.59269476e-01 -3.24891627e-01
-1.58396095e-01 6.63472474e-01 3.09694350e-01 5.24909794e-01
-6.66039824e-01 3.91527236e-01 7.63678551e-01 2.60840476e-01
1.87476113e-01 -2.99433321e-01 2.34857813e-01 -1.27317226e+00
-9.25691202e-02 -1.23085642e+00 5.24399653e-02 -2.31427506e-01
-1.49140775e+00 3.84387165e-01 1.42811432e-01 -6.36806369e-01
-7.25038469e-01 -3.52352142e-01 -9.21200633e-01 6.04401529e-01
-1.06154907e+00 -9.64618921e-01 7.75742456e-02 2.78984219e-01
1.51938364e-01 5.14659137e-02 6.42600238e-01 -2.22073436e-01
-6.78491712e-01 -9.87102911e-02 1.05727740e-01 -1.37820646e-01
3.92950475e-01 -1.37897217e+00 4.83969629e-01 6.58735394e-01
2.81192213e-01 8.28298748e-01 1.15125883e+00 -9.22697544e-01
-1.34241092e+00 -5.69471598e-01 1.04093635e+00 -7.53180504e-01
1.14471877e+00 -5.81493855e-01 -6.82257354e-01 5.97422123e-01
3.62564594e-01 -6.91709340e-01 6.59937918e-01 5.07069886e-01
-5.38105518e-02 5.20653166e-02 -7.00184107e-01 9.83180702e-01
9.80504990e-01 -5.22805035e-01 -4.49678868e-01 3.52815986e-01
4.59373683e-01 2.39879370e-01 -6.81819558e-01 -6.05389327e-02
6.24803424e-01 -1.17601812e+00 3.84901553e-01 -6.36246920e-01
2.39622548e-01 -2.80533254e-01 1.47074804e-01 -1.66367328e+00
-5.39041311e-02 -9.94625628e-01 -1.88822612e-01 1.15467179e+00
6.38820529e-01 -1.12118936e+00 4.78029609e-01 5.20204842e-01
6.08093441e-01 -5.74688911e-01 -6.78758144e-01 -4.27405328e-01
-4.51828353e-02 -2.80103981e-01 5.52786231e-01 8.59886229e-01
-1.12795539e-01 7.34516084e-01 -3.37126970e-01 2.74959028e-01
6.30306125e-01 2.58852363e-01 1.16938996e+00 -1.70994294e+00
-5.84813058e-01 -3.55184168e-01 -5.34825101e-02 -6.87210381e-01
4.75376129e-01 -5.03053427e-01 -1.84589460e-01 -1.29243362e+00
1.06316589e-01 -4.84161913e-01 2.28949070e-01 -6.32528542e-03
-3.15758556e-01 -2.77331322e-01 1.90366149e-01 3.22636575e-01
-4.11290497e-01 4.81791854e-01 9.13502276e-01 1.65552497e-01
-4.63092953e-01 2.76440322e-01 -6.27131879e-01 9.08076048e-01
8.73921752e-01 -6.27399325e-01 -6.38918400e-01 8.54601413e-02
9.53083158e-01 1.97026357e-01 4.71582949e-01 -2.19369307e-01
2.29964599e-01 -5.86518168e-01 -3.84563655e-01 -1.17716320e-01
1.26368608e-02 -7.28863776e-01 5.22126913e-01 6.60730183e-01
-2.23326385e-01 2.49781668e-01 -3.21226895e-01 8.03455889e-01
-1.01279162e-01 -2.43094847e-01 2.62575626e-01 -1.33299261e-01
-1.12398885e-01 1.95731953e-01 -8.52746964e-01 2.65702810e-02
9.73867834e-01 2.12746650e-01 -4.23475295e-01 -8.47394824e-01
-7.81325221e-01 2.01956362e-01 7.52535760e-01 3.31513047e-01
1.49057537e-01 -7.37670302e-01 -9.70614314e-01 9.17238966e-02
-9.49494913e-02 -4.92062867e-01 8.80689081e-03 8.82721066e-01
-2.06267372e-01 -2.20485814e-02 2.48022735e-01 -4.58266616e-01
-1.05421293e+00 4.76293862e-01 5.26374936e-01 -5.07779002e-01
-3.77661705e-01 2.14561924e-01 7.10627437e-02 -5.38233995e-01
-3.54577810e-01 -1.13793522e-01 -1.66807920e-01 5.22494912e-01
2.77018905e-01 3.01740170e-01 -3.14725310e-01 -4.81180370e-01
-1.75824642e-01 2.57101268e-01 1.71184689e-01 -5.62548220e-01
1.31692469e+00 -2.75532871e-01 -4.67884809e-01 8.77232254e-01
6.78940058e-01 4.88420278e-01 -1.09484172e+00 -9.18340683e-02
1.03402309e-01 -1.17751189e-01 -2.57284760e-01 -5.75948298e-01
-6.14136279e-01 4.61241484e-01 2.15867385e-01 1.07514989e+00
5.78090131e-01 7.07296804e-02 2.58316994e-01 3.95081758e-01
3.63964468e-01 -9.39740181e-01 -1.20268330e-01 4.16314006e-01
6.18936896e-01 -1.12003541e+00 -6.31702840e-02 -2.98263311e-01
-4.11875814e-01 7.65420139e-01 2.37878278e-01 -1.28289759e-01
9.24443185e-01 9.05369297e-02 3.28137219e-01 -5.64137757e-01
-1.33384991e+00 -4.28110778e-01 -3.43564823e-02 4.92814988e-01
2.42434785e-01 3.57315987e-01 -5.56040585e-01 7.21626878e-02
-3.77180517e-01 -1.41166031e-01 7.47084141e-01 6.36979461e-01
-2.42998317e-01 -1.42608094e+00 -1.55266851e-01 3.32790852e-01
-1.50378630e-01 6.18214048e-02 -8.07791173e-01 7.16152191e-01
1.71594888e-01 1.28700125e+00 5.32541201e-02 -2.07657993e-01
3.11247379e-01 1.85202256e-01 1.67586938e-01 -5.02315700e-01
-7.43626654e-01 -1.14559503e-02 2.99286813e-01 -2.01651752e-01
-7.80268252e-01 -9.29474890e-01 -6.88766062e-01 -7.89279878e-01
-4.55687195e-01 4.74591017e-01 4.61753398e-01 1.21409488e+00
5.45102596e-01 1.77352786e-01 7.19541490e-01 -4.99165088e-01
-4.82248664e-01 -9.89273846e-01 -4.50996369e-01 6.60074830e-01
2.69209653e-01 -5.91639042e-01 -7.71587431e-01 -1.71870720e-02] | [6.725534439086914, 5.081334590911865] |
7a563a18-4670-4314-9114-3799bbe2e487 | optimizing-software-effort-estimation-models | 1903.02079 | null | http://arxiv.org/abs/1903.02079v1 | http://arxiv.org/pdf/1903.02079v1.pdf | Optimizing Software Effort Estimation Models Using Firefly Algorithm | Software development effort estimation is considered a fundamental task for
software development life cycle as well as for managing project cost, time and
quality. Therefore, accurate estimation is a substantial factor in projects
success and reducing the risks. In recent years, software effort estimation has
received a considerable amount of attention from researchers and became a
challenge for software industry. In the last two decades, many researchers and
practitioners proposed statistical and machine learning-based models for
software effort estimation. In this work, Firefly Algorithm is proposed as a
metaheuristic optimization method for optimizing the parameters of three
COCOMO-based models. These models include the basic COCOMO model and other two
models proposed in the literature as extensions of the basic COCOMO model. The
developed estimation models are evaluated using different evaluation metrics.
Experimental results show high accuracy and significant error minimization of
Firefly Algorithm over other metaheuristic optimization algorithms including
Genetic Algorithms and Particle Swarm Optimization. | ['Rizik M. H. Al-Sayyed', 'Hossam Faris', 'Ibrahim Aljarah', 'Nazeeh Ghatasheh'] | 2019-01-08 | null | null | null | null | ['metaheuristic-optimization'] | ['methodology'] | [-1.27097458e-01 -6.02071166e-01 6.28110319e-02 4.45500808e-03
-1.38144925e-01 -3.20010811e-01 -4.51492891e-03 6.13942623e-01
-4.31726038e-01 8.89172137e-01 -3.41209620e-01 1.48360461e-01
-6.71193540e-01 -9.78000700e-01 2.85264775e-02 -6.94314957e-01
4.14528877e-01 4.33535665e-01 -4.66761664e-02 -1.66023850e-01
1.02479482e+00 4.90548730e-01 -1.70407927e+00 -5.95949948e-01
1.03068209e+00 5.21647990e-01 3.35624903e-01 5.36347926e-01
-1.56162068e-01 4.11948353e-01 -7.77132511e-01 -4.04710412e-01
7.06931576e-02 -4.75204974e-01 -3.06875318e-01 -1.94306746e-02
-7.28231549e-01 4.77639884e-01 3.20395708e-01 1.14458168e+00
5.90425789e-01 1.85720205e-01 4.19136673e-01 -1.59798801e+00
-4.08717334e-01 4.89335805e-01 -8.75216842e-01 3.35394681e-01
4.17959064e-01 -1.33113921e-01 7.57935584e-01 -5.79282641e-01
2.97125071e-01 7.84622252e-01 6.75813258e-01 -7.92781860e-02
-7.32901752e-01 -5.70801854e-01 -3.75095546e-01 5.45674264e-01
-1.43499482e+00 1.54588759e-01 8.22225988e-01 -4.79485601e-01
1.04800022e+00 5.45539379e-01 1.00373125e+00 -1.33240491e-01
5.88137686e-01 4.42234814e-01 1.15191555e+00 -9.71704304e-01
5.27210116e-01 4.29844350e-01 1.42768845e-01 4.89493608e-01
7.42836654e-01 -1.91847518e-01 -2.78372735e-01 -4.03021514e-01
2.79602081e-01 1.81688249e-01 -1.18410751e-01 -9.76611376e-02
-6.42742872e-01 9.58332360e-01 -2.43673503e-01 6.72444820e-01
-7.15378881e-01 9.21725929e-02 1.75459325e-01 9.97922048e-02
5.34102917e-01 6.62044346e-01 -2.34874278e-01 -8.45797896e-01
-9.16744173e-01 3.26491416e-01 8.46188724e-01 7.61346638e-01
4.63896364e-01 2.04928577e-01 2.12637782e-01 7.45866954e-01
5.49656153e-01 2.99871415e-01 3.18279952e-01 -7.48798430e-01
1.88716233e-01 1.10635912e+00 3.64125043e-01 -1.29361618e+00
-4.86776948e-01 -7.21037865e-01 -4.77181792e-01 2.65715361e-01
1.44872759e-02 -3.12978297e-01 -2.12932855e-01 1.04865718e+00
4.16736931e-01 -3.72587480e-02 -1.97383925e-01 3.07953030e-01
2.34064370e-01 6.03496194e-01 -1.15328878e-01 -7.01451063e-01
1.17544830e+00 -7.71547914e-01 -8.49957049e-01 9.97978523e-02
2.64180273e-01 -1.38007522e+00 4.17496443e-01 7.34918356e-01
-1.25816536e+00 -3.62381488e-01 -1.11344755e+00 8.85933459e-01
-1.92026675e-01 1.09271139e-01 7.66320527e-01 1.45388913e+00
-8.06929469e-01 5.15686214e-01 -8.52537692e-01 -5.09023666e-01
2.34911233e-01 2.50339121e-01 1.89086542e-01 5.45624411e-03
-6.43865168e-01 9.06814516e-01 2.51607597e-01 5.82837779e-03
-2.12228000e-01 -2.75780946e-01 -4.00302827e-01 2.60712653e-01
1.15826286e-01 -2.98511267e-01 6.64127350e-01 -7.10136771e-01
-1.32446027e+00 4.26777273e-01 -2.04560217e-02 6.26273900e-02
3.85230035e-01 1.12371854e-01 -2.44610757e-01 -4.21969533e-01
-8.40417203e-03 -4.43428248e-01 5.61049163e-01 -8.17272544e-01
-7.83527493e-01 -5.38112521e-01 1.12878662e-02 1.84473783e-01
-4.66842651e-01 6.30676150e-01 -5.92472740e-02 -5.89147270e-01
1.41128838e-01 -9.79098499e-01 -3.05942267e-01 -5.19778907e-01
1.44857168e-01 -3.01108569e-01 1.89594269e-01 -6.33346796e-01
1.67408311e+00 -1.54684985e+00 3.04065585e-01 4.54383373e-01
-1.83726653e-01 1.66877285e-01 3.94065976e-01 7.46604979e-01
4.01453376e-02 6.97186217e-02 6.39190823e-02 9.24088508e-02
-1.03842087e-01 -2.29250625e-01 7.43361056e-01 4.82188702e-01
-7.20831603e-02 1.51921034e-01 -6.92300797e-01 -3.81539285e-01
2.79739380e-01 3.97385836e-01 -4.74552214e-01 2.88129151e-01
2.53176033e-01 -1.42227024e-01 -4.90147829e-01 1.06283820e+00
8.49995673e-01 -8.88126865e-02 2.58286119e-01 -1.82372855e-03
-6.21017456e-01 -5.87149739e-01 -1.48204541e+00 1.02128112e+00
-7.46885359e-01 5.22940814e-01 -2.02038556e-01 -1.19100106e+00
1.13366342e+00 3.48873377e-01 7.30570138e-01 -3.54739606e-01
6.81244671e-01 1.46070421e-01 2.68011659e-01 -6.12307429e-01
6.16789341e-01 -1.41678993e-02 2.18895733e-01 7.73989499e-01
-2.67067641e-01 -2.59765476e-01 6.33965075e-01 -3.98569465e-01
8.27801347e-01 1.07681826e-01 7.21657276e-01 -1.02835365e-01
9.27187562e-01 5.59829362e-02 8.13919544e-01 2.61274457e-01
-2.49617055e-01 3.14454019e-01 4.48142529e-01 -3.98136497e-01
-9.95290458e-01 -4.33629245e-01 1.46604944e-02 8.22441459e-01
-6.11775219e-02 -2.58739650e-01 -8.32585812e-01 -2.68924385e-01
-1.14156365e-01 8.90529573e-01 -4.15008336e-01 8.86065364e-02
-5.19991636e-01 -1.10913968e+00 1.28431663e-01 2.59496897e-01
4.19934362e-01 -1.00133991e+00 -9.81063068e-01 5.73648632e-01
-1.86007586e-03 -5.48544168e-01 -7.82263651e-03 -1.17388986e-01
-6.68627322e-01 -1.33136868e+00 -7.29851484e-01 -5.64075351e-01
8.06561351e-01 3.80731583e-01 1.04204416e+00 6.57873750e-01
-8.53230596e-01 2.93202609e-01 -7.23665357e-01 -8.25166583e-01
-5.51824689e-01 -1.14083812e-01 -1.99412882e-01 -2.66904712e-01
5.39507210e-01 -4.40839231e-01 -3.93424988e-01 5.42428017e-01
-4.81706679e-01 -4.22802329e-01 4.07514274e-01 6.70120180e-01
3.67976487e-01 8.70264113e-01 6.43230975e-01 -4.85859036e-01
9.85295773e-01 -5.66663444e-01 -1.15766358e+00 5.47696590e-01
-1.02029359e+00 -1.47757992e-01 4.73968089e-01 -2.75356025e-02
-1.01685846e+00 -2.79144704e-01 -4.09127437e-02 2.46463284e-01
1.88912824e-01 7.30058551e-01 -7.39573538e-02 -5.87832689e-01
2.82674849e-01 3.06030244e-01 1.41818598e-01 -3.73875856e-01
-4.80038315e-01 9.59166110e-01 -5.03522515e-01 -5.15669703e-01
5.82753658e-01 1.10575847e-01 1.93916842e-01 -7.84258783e-01
-3.08624864e-01 -5.29445708e-01 -9.88263488e-02 -6.13061070e-01
6.53364837e-01 -2.05758065e-01 -8.16617846e-01 4.83042598e-01
-1.15213561e+00 7.21570075e-01 1.58403054e-01 4.86634344e-01
-1.67904168e-01 3.26660007e-01 1.38805643e-01 -1.27708983e+00
-5.29493034e-01 -1.45714891e+00 1.41274169e-01 7.58947074e-01
-1.01693042e-01 -9.82986212e-01 2.48009086e-01 5.19627452e-01
7.48156309e-01 5.52424133e-01 7.60866523e-01 -2.54759431e-01
-3.71125370e-01 -7.22178578e-01 -3.14552709e-02 3.67070585e-01
3.18376005e-01 5.87137520e-01 -2.20378250e-01 -2.88179815e-01
4.38240647e-01 1.71644539e-01 -2.36500442e-01 6.77150428e-01
6.81285381e-01 1.65029079e-01 -2.19697013e-01 1.76618159e-01
1.88931561e+00 9.81314301e-01 5.52455962e-01 8.79153430e-01
-4.15187962e-02 5.60933769e-01 1.29149318e+00 1.11321545e+00
1.76247269e-01 4.91168857e-01 5.43972790e-01 5.48233032e-01
5.58231056e-01 6.44758642e-01 -1.89219475e-01 1.16224885e+00
-6.61520302e-01 -2.92595744e-01 -1.20002496e+00 4.90076810e-01
-1.67950714e+00 -9.95368421e-01 -3.82445306e-02 2.23285437e+00
4.87112820e-01 1.81272849e-01 1.72083259e-01 6.13783956e-01
1.04085898e+00 -1.36035070e-01 -5.36141805e-02 -6.80760026e-01
2.06446946e-01 4.51413393e-01 2.43608907e-01 1.52225286e-01
-7.32744575e-01 2.71601349e-01 5.30424643e+00 6.16198421e-01
-8.58866215e-01 2.54091978e-01 1.48332715e-01 6.91731349e-02
-1.56026140e-01 1.60174936e-01 -6.46660328e-01 7.90858805e-01
5.79396725e-01 -8.75194252e-01 4.57297236e-01 9.75000978e-01
4.20171320e-01 -4.35027689e-01 -5.35048485e-01 9.28146005e-01
8.67393464e-02 -1.24088860e+00 -6.82717383e-01 -1.98899396e-02
1.04406643e+00 -5.17498136e-01 -3.44346404e-01 -6.31365404e-02
2.20503076e-03 -8.21724057e-01 2.84743279e-01 5.53020895e-01
-1.82863716e-02 -1.30219936e+00 1.32557416e+00 2.96729088e-01
-1.14995146e+00 -1.21872351e-01 -5.64736366e-01 -2.75799245e-01
1.39576375e-01 4.42966491e-01 -4.47290927e-01 5.75719059e-01
7.14989245e-01 3.18358898e-01 -3.05179030e-01 1.81730521e+00
1.77928686e-01 5.57776451e-01 -1.97312877e-01 -9.30460215e-01
-1.86838917e-02 -5.03421426e-01 5.10798693e-01 7.61892319e-01
6.65231526e-01 -8.23653862e-02 -3.30104798e-01 7.46274889e-01
3.29460263e-01 5.25191486e-01 -1.05603606e-01 -2.77976334e-01
8.07186544e-01 1.24139464e+00 -1.02289701e+00 2.75280476e-01
-4.51533169e-01 1.57987043e-01 -3.35413843e-01 -1.61890298e-01
-1.05778229e+00 -1.07813466e+00 4.64986801e-01 -8.52759257e-02
6.13371702e-03 -5.20370863e-02 -3.62551093e-01 -5.60615778e-01
-1.06246151e-01 -8.20993483e-01 7.56148994e-02 -3.95095527e-01
-5.60047269e-01 6.55483544e-01 -8.53191689e-02 -1.57801509e+00
5.15273288e-02 -3.62961143e-01 -8.04316223e-01 9.02150691e-01
-1.22641480e+00 -7.03235030e-01 -8.11024725e-01 -1.66673243e-01
7.09193110e-01 -6.99351966e-01 5.92438459e-01 3.19428802e-01
-9.07378674e-01 3.58402640e-01 4.05984938e-01 -4.74077761e-01
2.51530737e-01 -1.01318216e+00 -1.27801582e-01 9.91748095e-01
-3.34208459e-01 7.83755839e-01 1.17904174e+00 -5.24994135e-01
-1.37677658e+00 -4.28848863e-01 6.57124877e-01 1.11841068e-01
4.54044461e-01 3.18545222e-01 -3.88491631e-01 8.17172602e-02
6.34818137e-01 -4.45979923e-01 9.18983757e-01 -1.38238937e-01
6.18507445e-01 -1.24517344e-01 -1.44034863e+00 2.26652890e-01
1.62525415e-01 3.05435002e-01 -2.06781924e-01 3.42702746e-01
1.52639329e-01 -2.04132378e-01 -1.21056044e+00 -4.38094810e-02
5.97963691e-01 -1.36831915e+00 7.09409595e-01 -2.38325387e-01
2.51514256e-01 -1.49709389e-01 4.62909602e-02 -1.23307872e+00
-3.37880313e-01 -4.69143599e-01 3.23456138e-01 1.36862028e+00
6.87198192e-02 -7.17250586e-01 1.20358038e+00 7.37852633e-01
3.05724710e-01 -9.70604658e-01 -8.74898911e-01 -6.39922857e-01
-3.55495065e-01 -1.41272843e-01 6.80118263e-01 9.91485119e-01
-9.46289003e-02 -1.87558636e-01 -8.26835558e-02 -1.14829689e-01
9.79785383e-01 1.41589522e-01 6.91858172e-01 -1.39830756e+00
-3.53990406e-01 -5.25996387e-01 -7.59996951e-01 7.51479641e-02
-2.67596632e-01 -2.66713262e-01 -2.86522150e-01 -1.61822581e+00
2.30204418e-01 -6.37697995e-01 -3.64416331e-01 5.06969541e-02
-3.15770239e-01 -4.51160036e-02 8.95836204e-02 -9.31720659e-02
-2.32954711e-01 3.04008901e-01 7.26230502e-01 2.77413242e-02
-3.03329647e-01 7.27368712e-01 -5.80902457e-01 7.84355223e-01
8.90547752e-01 -8.86107028e-01 -3.80459726e-01 -1.97313339e-01
7.94488907e-01 3.83999228e-01 -2.91792899e-01 -1.00741673e+00
2.89466619e-01 -5.71777880e-01 -1.32296100e-01 -5.89464068e-01
2.36679539e-02 -1.13925266e+00 6.67939961e-01 6.16279840e-01
1.67479590e-01 3.61591220e-01 -3.08976062e-02 2.78385520e-01
-4.02339280e-01 -1.07545650e+00 8.00750077e-01 -6.86361641e-02
-4.28467900e-01 -2.23082423e-01 -2.54167467e-01 8.77768546e-03
1.47627854e+00 -6.92497790e-01 -2.40311638e-01 -8.07443410e-02
-5.06164312e-01 1.44365892e-01 4.37161386e-01 3.30781549e-01
7.30470777e-01 -1.02170241e+00 -7.58498967e-01 -3.99495572e-01
-4.95525487e-02 -5.79097629e-01 2.14680851e-01 1.07434571e+00
-1.24735403e+00 3.52095455e-01 -6.56724215e-01 -1.57678947e-01
-1.89467406e+00 -1.47959394e-02 5.34870103e-02 -3.53356332e-01
2.91798204e-01 8.92127275e-01 -7.49135733e-01 -2.45194826e-02
-1.72125120e-02 1.29032955e-01 -5.95018387e-01 1.19215557e-02
3.19102496e-01 1.35197103e+00 5.36183827e-02 -4.30794388e-01
-6.49282634e-01 8.53391171e-01 8.67844224e-02 2.89817899e-01
1.84984469e+00 -1.48193270e-01 -6.26685441e-01 4.37439494e-02
7.72914290e-01 2.19808698e-01 -4.01615024e-01 1.47320539e-01
3.03222120e-01 -9.57701564e-01 3.62722278e-01 -7.90031791e-01
-1.02686024e+00 6.31748855e-01 7.28717446e-01 5.99798799e-01
1.38725901e+00 -7.16036916e-01 2.62464374e-01 3.88920680e-02
6.22562468e-01 -1.59286714e+00 7.99675658e-02 -9.72735975e-03
7.85613835e-01 -1.17041934e+00 6.04998469e-01 -3.91528308e-01
-4.79511946e-01 1.38726878e+00 6.69360876e-01 -3.10830735e-02
7.82192767e-01 3.91858995e-01 -2.50248730e-01 8.65698531e-02
-4.19743419e-01 2.32828200e-01 -4.10837354e-04 5.43291450e-01
7.41564155e-01 -1.93774506e-01 -1.12357259e+00 8.89492631e-01
1.57110155e-01 1.66806042e-01 7.54648685e-01 1.51822603e+00
-6.96423531e-01 -1.38484526e+00 -7.71421671e-01 4.63421315e-01
-7.06165314e-01 7.12361187e-02 9.44726840e-02 8.69159758e-01
1.83683917e-01 1.12030590e+00 -4.33315456e-01 -4.49417830e-01
2.69872218e-01 -3.62191230e-01 5.60103536e-01 -4.41054374e-01
-8.82817924e-01 -3.77048962e-02 3.40053402e-02 -1.22936040e-01
-6.18174613e-01 -8.50732028e-01 -1.00406361e+00 -4.06416923e-01
-1.03801644e+00 7.93805778e-01 1.30525994e+00 8.28190088e-01
1.77273169e-01 7.96297133e-01 1.01809740e+00 -6.34226084e-01
-4.52756673e-01 -8.28819215e-01 -6.78882003e-01 -2.85940379e-01
-5.78325808e-01 -9.23767924e-01 -3.40965241e-01 -2.92763174e-01] | [5.664723873138428, 3.4973769187927246] |
66b5101c-e641-4ce1-ba5f-ca5206a1113f | unsupervised-person-re-identification-with-2 | 2110.15610 | null | https://arxiv.org/abs/2110.15610v2 | https://arxiv.org/pdf/2110.15610v2.pdf | Unsupervised Person Re-Identification with Wireless Positioning under Weak Scene Labeling | Existing unsupervised person re-identification methods only rely on visual clues to match pedestrians under different cameras. Since visual data is essentially susceptible to occlusion, blur, clothing changes, etc., a promising solution is to introduce heterogeneous data to make up for the defect of visual data. Some works based on full-scene labeling introduce wireless positioning to assist cross-domain person re-identification, but their GPS labeling of entire monitoring scenes is laborious. To this end, we propose to explore unsupervised person re-identification with both visual data and wireless positioning trajectories under weak scene labeling, in which we only need to know the locations of the cameras. Specifically, we propose a novel unsupervised multimodal training framework (UMTF), which models the complementarity of visual data and wireless information. Our UMTF contains a multimodal data association strategy (MMDA) and a multimodal graph neural network (MMGN). MMDA explores potential data associations in unlabeled multimodal data, while MMGN propagates multimodal messages in the video graph based on the adjacency matrix learned from histogram statistics of wireless data. Thanks to the robustness of the wireless data to visual noise and the collaboration of various modules, UMTF is capable of learning a model free of the human label on data. Extensive experimental results conducted on two challenging datasets, i.e., WP-ReID and DukeMTMC-VideoReID demonstrate the effectiveness of the proposed method. | ['Houqiang Li', 'Qiaokang Xie', 'Wengang Zhou', 'Yiheng Liu'] | 2021-10-29 | null | null | null | null | ['scene-labeling', 'unsupervised-person-re-identification'] | ['computer-vision', 'computer-vision'] | [-9.91878808e-02 -2.44191036e-01 -1.56617090e-01 -3.23467463e-01
-2.55712986e-01 -5.97474337e-01 4.34121549e-01 1.88394859e-01
-5.02593875e-01 6.04414463e-01 1.45238012e-01 -5.56447217e-03
4.28640936e-03 -7.91097939e-01 -7.51934350e-01 -7.92069554e-01
1.07057728e-01 1.83126152e-01 1.54580325e-01 1.24038368e-01
-1.19793512e-01 1.21528856e-01 -1.28393340e+00 -1.08623110e-01
8.12862515e-01 6.32308066e-01 2.68401057e-01 6.04930758e-01
-3.00234184e-02 6.22471035e-01 -4.73546058e-01 -7.99730182e-01
1.92122951e-01 1.04423784e-01 -4.12462562e-01 3.64904881e-01
4.46115881e-01 -3.38758618e-01 -8.10449898e-01 1.27847636e+00
7.32670486e-01 1.95459470e-01 4.06699598e-01 -1.63861239e+00
-6.34942055e-01 5.55857420e-01 -7.34157085e-01 6.73016161e-02
7.34632194e-01 1.02030575e-01 3.22137773e-01 -6.82377815e-01
4.79671806e-01 1.07723916e+00 1.10443354e+00 5.23481846e-01
-1.05198610e+00 -6.57482684e-01 4.77758259e-01 5.03356040e-01
-1.79014277e+00 -3.49334866e-01 8.88136268e-01 -3.41070801e-01
3.37617636e-01 3.82782102e-01 8.34230721e-01 1.30782056e+00
-3.75170857e-01 6.75030649e-01 9.29416776e-01 -3.80852938e-01
-9.57227945e-02 4.20066833e-01 3.42813790e-01 7.68798530e-01
4.37620431e-01 -6.51835874e-02 -6.70430303e-01 -1.61816925e-01
4.18756813e-01 4.56967205e-01 -3.34225625e-01 -4.75736290e-01
-1.07474339e+00 2.61283427e-01 4.30788487e-01 1.34557173e-01
-7.56819248e-02 8.94632377e-03 1.63082510e-01 -4.29367833e-02
1.61520600e-01 -3.67934734e-01 5.03682420e-02 3.14377308e-01
-7.14324832e-01 -2.11304650e-01 5.70555568e-01 1.28366446e+00
9.24752891e-01 -2.56774306e-01 -5.66701442e-02 7.71040261e-01
7.70650148e-01 9.45186377e-01 2.07628548e-01 -5.28921723e-01
8.32023323e-01 6.79856598e-01 2.50510782e-01 -1.58354354e+00
-5.51861703e-01 -3.40507835e-01 -1.02415442e+00 -5.46413422e-01
4.86286819e-01 -4.39790905e-01 -6.90062642e-01 1.76355267e+00
4.99173582e-01 6.64705217e-01 4.07408774e-02 9.24942672e-01
1.16126037e+00 4.54235792e-01 3.24370325e-01 -1.62213206e-01
1.20055258e+00 -7.18755662e-01 -6.90674365e-01 -9.83266681e-02
5.79211414e-01 -3.81877989e-01 5.19948184e-01 1.01649258e-02
-5.93899071e-01 -8.28491092e-01 -9.47229028e-01 3.40099454e-01
-6.68033838e-01 2.72442281e-01 3.80334526e-01 1.10964465e+00
-1.22035658e+00 -1.73120141e-01 -4.25286978e-01 -9.58144903e-01
1.62646234e-01 6.73466623e-01 -5.53748429e-01 -4.74501461e-01
-1.19153178e+00 4.21416909e-01 4.19934839e-01 5.71266413e-01
-6.39560401e-01 -2.50831246e-01 -8.45294654e-01 -2.62590528e-01
3.54751855e-01 -8.09173346e-01 5.95035732e-01 -6.75766230e-01
-1.03784275e+00 5.98114431e-01 -2.64680088e-01 -3.02638322e-01
5.92225671e-01 1.92685649e-01 -7.75079727e-01 2.39717156e-01
2.62492180e-01 4.89010304e-01 7.27563739e-01 -1.76550710e+00
-9.42951739e-01 -4.97686982e-01 1.01592476e-02 2.60877162e-01
-7.18430400e-01 -3.86116832e-01 -1.04064143e+00 -4.51418072e-01
-2.67552547e-02 -9.07092929e-01 -1.57862768e-01 -5.53611338e-01
-7.01577187e-01 -1.32161885e-01 8.13145220e-01 -8.56511116e-01
1.11668205e+00 -2.07121968e+00 -8.87798965e-02 7.28146553e-01
3.16977471e-01 1.22469455e-01 -1.02266446e-01 2.24567965e-01
1.28543854e-01 4.83804159e-02 1.21278681e-01 -7.30476558e-01
8.57365876e-02 9.29056183e-02 1.85242310e-01 6.73970878e-01
-4.78168041e-01 7.94451833e-01 -9.79394734e-01 -8.19185376e-01
4.69089240e-01 4.44362342e-01 -1.33106917e-01 1.36245817e-01
4.48872060e-01 6.95546448e-01 -4.91593391e-01 9.32592213e-01
1.06089818e+00 -1.58496082e-01 4.84547466e-01 -6.57786787e-01
-6.07201979e-02 -7.03678548e-01 -1.39219868e+00 1.65931928e+00
1.10431172e-01 3.63924921e-01 1.23531453e-01 -1.10770965e+00
6.56587124e-01 2.88540602e-01 6.71106935e-01 -6.62445009e-01
3.13586891e-02 -1.62275240e-01 -6.87316000e-01 -7.52710044e-01
4.11796838e-01 4.99337554e-01 -9.88803878e-02 1.10864803e-01
-8.88587534e-03 9.19309139e-01 8.61330852e-02 5.12537360e-01
8.51548195e-01 -1.32698178e-01 -2.98886240e-01 1.40896395e-01
6.95581257e-01 -1.12548515e-01 5.54147243e-01 1.11836505e+00
-3.18837374e-01 2.78055340e-01 -2.05454841e-01 -3.22186172e-01
-6.86955273e-01 -1.08899057e+00 6.93812147e-02 8.87732923e-01
1.05864167e+00 -4.64015573e-01 -7.23277450e-01 -9.00189400e-01
-9.95026380e-02 1.05729975e-01 -4.02725488e-01 1.78531446e-02
-4.35060024e-01 -1.02284813e+00 6.93933487e-01 2.86517173e-01
1.02005482e+00 -2.52463251e-01 1.72252014e-01 -4.44118381e-02
-7.23691761e-01 -1.51926184e+00 -6.46663189e-01 -4.81621832e-01
-3.81257355e-01 -1.22824049e+00 -7.94827759e-01 -1.00742102e+00
1.01269341e+00 9.39420521e-01 5.08001149e-01 2.22329855e-01
-4.79287766e-02 1.31063819e+00 -3.53506565e-01 1.00844810e-02
1.93614408e-01 2.11660340e-02 4.79640841e-01 7.52247870e-01
5.46311498e-01 -4.82232720e-01 -6.75243735e-01 6.11433983e-01
-4.35002744e-01 -6.21167608e-02 2.92761296e-01 5.14351785e-01
3.93804044e-01 4.04759824e-01 3.23248535e-01 -6.04637802e-01
2.57242054e-01 -7.40370095e-01 -3.84398222e-01 6.75046563e-01
-2.97899902e-01 -5.70398927e-01 3.28876793e-01 -5.36125720e-01
-1.03100300e+00 3.97702217e-01 3.03141445e-01 -2.58366495e-01
-4.06113386e-01 3.32139939e-01 -6.80440426e-01 -4.86364305e-01
1.44953668e-01 3.11121464e-01 -3.74207765e-01 -2.70010382e-01
4.61890787e-01 7.37970769e-01 7.99654365e-01 -3.97048563e-01
1.09024131e+00 5.58057606e-01 -1.29769653e-01 -9.69823122e-01
-1.71782553e-01 -8.63117039e-01 -7.82520711e-01 -8.92921746e-01
1.10833395e+00 -1.22906208e+00 -1.08287656e+00 5.89535415e-01
-1.06018305e+00 7.79463649e-02 3.69279206e-01 6.75044477e-01
1.22873157e-01 1.04415417e+00 -3.42930734e-01 -1.02070129e+00
1.24760807e-01 -9.01881695e-01 9.03741598e-01 6.70105934e-01
2.49669343e-01 -1.20428038e+00 -2.19414428e-01 7.26167262e-01
1.46770906e-02 2.09920943e-01 4.46234494e-01 -3.92436206e-01
-7.98007011e-01 -4.75144684e-01 -4.62059230e-01 -1.91998154e-01
2.50004619e-01 -4.58191037e-01 -1.09392810e+00 -4.71875429e-01
-4.13955837e-01 7.65332356e-02 7.17773139e-01 4.15785819e-01
9.71408069e-01 -3.07345957e-01 -9.16947305e-01 8.15029681e-01
1.24696934e+00 1.15541935e-01 4.31401342e-01 3.71447057e-01
1.46171391e+00 7.40066648e-01 4.06892687e-01 4.23441648e-01
1.16773057e+00 7.45878696e-01 4.45756793e-01 -3.31690043e-01
-2.81786993e-02 -4.72505718e-01 3.62308234e-01 7.05638170e-01
-3.84584755e-01 -5.76805115e-01 -8.06086242e-01 3.86235535e-01
-2.26391506e+00 -9.86480713e-01 -4.25471723e-01 2.18645287e+00
7.01566711e-02 -2.99312234e-01 4.02148426e-01 9.16936062e-03
1.28605592e+00 -1.24967867e-03 -3.16378355e-01 5.27673304e-01
-3.50533247e-01 -6.41696274e-01 1.05619359e+00 4.11017567e-01
-1.43811488e+00 6.72030330e-01 4.95366955e+00 7.39975512e-01
-6.18627071e-01 2.60616630e-01 3.64311904e-01 2.62638420e-01
-1.21806502e-01 -2.13905692e-01 -9.49928939e-01 8.46760094e-01
5.06347001e-01 3.24263096e-01 4.57455665e-01 4.68387157e-01
2.29318753e-01 -1.37482673e-01 -8.11252534e-01 1.52480209e+00
3.80072385e-01 -1.06734872e+00 -5.86542003e-02 9.71948877e-02
5.87738991e-01 -2.92975932e-01 1.68572515e-01 5.35155684e-02
2.76921064e-01 -5.92797399e-01 5.52477181e-01 9.28806424e-01
5.50239921e-01 -7.44249403e-01 9.22191620e-01 3.10713708e-01
-1.76003969e+00 -2.71048099e-01 -3.23244959e-01 4.45709258e-01
2.43876517e-01 2.06051543e-01 -5.94706655e-01 1.06849158e+00
9.35582817e-01 1.07889724e+00 -9.56424057e-01 1.24018919e+00
1.33205384e-01 3.82760525e-01 -3.70575100e-01 1.49752095e-01
-2.13605732e-01 -1.46667376e-01 5.27317703e-01 1.33427739e+00
4.79968816e-01 1.57490969e-01 6.69324994e-01 3.61355275e-01
-3.62177752e-02 2.52231490e-03 -7.43276119e-01 3.26944232e-01
6.28957391e-01 1.19189084e+00 -6.21971011e-01 -1.75885394e-01
-6.42290294e-01 1.11556947e+00 1.42680272e-01 9.24914777e-01
-9.50340033e-01 3.49872783e-02 1.84273094e-01 4.79448102e-02
1.28035340e-02 -4.40023661e-01 3.09623107e-02 -1.36168969e+00
4.77293916e-02 -4.40164089e-01 6.93976820e-01 -5.92838943e-01
-1.48668647e+00 1.32210776e-01 1.15382141e-02 -1.33121264e+00
2.39713237e-01 -3.06306124e-01 -3.58082294e-01 6.39486670e-01
-1.46629870e+00 -1.67514551e+00 -7.89302826e-01 1.21281457e+00
5.01925200e-02 -4.21523511e-01 3.96330714e-01 8.21302116e-01
-1.01034117e+00 1.01511586e+00 3.67810167e-02 4.81504589e-01
8.08156490e-01 -9.03619826e-01 -1.26257867e-01 1.15049672e+00
5.92248514e-02 5.57142079e-01 3.81126285e-01 -1.00916970e+00
-1.77406883e+00 -1.35909057e+00 7.38624096e-01 -4.05648738e-01
3.81150007e-01 -4.22557682e-01 -4.89190012e-01 6.01658106e-01
1.42251417e-01 -8.61546248e-02 8.67306173e-01 1.01015402e-03
-1.19372137e-01 -4.18999583e-01 -1.08594930e+00 4.36951786e-01
1.27574134e+00 -6.60388231e-01 -1.17377959e-01 4.64396566e-01
3.66743565e-01 -1.25569910e-01 -6.94201350e-01 3.66533160e-01
5.25919974e-01 -6.93365455e-01 1.36686814e+00 -1.54700920e-01
-6.31030977e-01 -7.40044713e-01 -2.51905829e-01 -9.48204458e-01
-3.11674058e-01 -3.85041386e-01 -5.61458729e-02 1.78474891e+00
5.52774668e-02 -6.83777928e-01 7.77654707e-01 9.40851808e-01
2.32425258e-01 2.65531927e-01 -9.01518703e-01 -5.40117741e-01
-9.57859695e-01 -5.23619115e-01 7.49333858e-01 1.12452221e+00
-1.04302049e-01 2.96586845e-02 -9.75000858e-01 1.11251426e+00
1.19265831e+00 -3.59577328e-01 1.04175019e+00 -1.11729801e+00
-1.34501204e-01 3.37202698e-02 -6.39001787e-01 -1.17071164e+00
1.35737345e-01 -7.95268357e-01 -1.20824091e-01 -1.48252225e+00
5.34596026e-01 -5.49799800e-01 -2.76554763e-01 2.90162593e-01
-1.89821109e-01 4.60184425e-01 2.32704848e-01 3.25930744e-01
-1.12651634e+00 3.88804138e-01 6.89621806e-01 -7.59454012e-01
-2.82878637e-01 1.47664696e-01 -3.92283022e-01 6.97030187e-01
6.11377776e-01 -2.58656174e-01 -4.39919621e-01 -7.99238324e-01
9.97598376e-03 -1.15649089e-01 9.66973007e-01 -1.22244632e+00
9.54648256e-01 1.32357985e-01 7.30597436e-01 -7.23351419e-01
3.64501476e-01 -1.20446646e+00 3.76399994e-01 1.24337867e-01
4.75850441e-02 5.66916391e-02 -2.03339872e-03 1.10574460e+00
6.94840178e-02 -4.94512683e-03 2.91639477e-01 4.39011045e-02
-1.13370323e+00 4.50323373e-01 -4.55406994e-01 -4.57112193e-01
1.00641084e+00 -5.36884844e-01 -5.29708982e-01 -5.30429721e-01
-9.19921458e-01 6.88624442e-01 3.43363941e-01 2.75808871e-01
6.82878137e-01 -1.54781532e+00 -3.37992102e-01 1.40780643e-01
2.33323991e-01 -3.67436975e-01 7.09842622e-01 8.75537217e-01
-9.61328223e-02 2.48148784e-01 1.06998637e-01 -8.55647385e-01
-1.52362859e+00 6.58591032e-01 2.58861691e-01 1.34363562e-01
-3.17645222e-01 6.62039578e-01 8.58616307e-02 -5.06866574e-01
5.87499857e-01 1.76804662e-01 -5.49112380e-01 1.51948676e-01
4.45488185e-01 5.14772058e-01 -4.20722634e-01 -1.21681261e+00
-5.27804852e-01 9.79226053e-01 2.53188610e-01 9.12088528e-02
7.99966633e-01 -9.79848266e-01 7.31353909e-02 1.19611114e-01
8.67708862e-01 9.74226817e-02 -1.00502443e+00 -3.85454416e-01
-1.54187113e-01 -4.40955192e-01 -3.24554920e-01 -4.02199268e-01
-1.04824793e+00 5.21130323e-01 1.14532697e+00 3.65329951e-01
1.03464770e+00 -1.97109878e-01 7.62804389e-01 3.95883262e-01
7.10442483e-01 -1.22836518e+00 3.65342088e-02 5.08650988e-02
1.91210415e-02 -1.44130802e+00 -7.68836811e-02 -5.91895342e-01
-4.41380709e-01 8.71072233e-01 5.21238029e-01 4.54294384e-01
7.17889488e-01 -2.00221851e-01 2.94372421e-02 9.83601138e-02
2.53646702e-01 -5.36300540e-01 2.16785431e-01 1.26309705e+00
-4.00877714e-01 6.80072829e-02 2.54252583e-01 7.32265353e-01
6.38725907e-02 -2.30852425e-01 4.78002764e-02 6.50482476e-01
-1.86243117e-01 -1.03863704e+00 -6.67888701e-01 2.19388828e-02
1.61973163e-02 1.74012288e-01 -3.87895346e-01 6.40042245e-01
7.60839820e-01 1.59340978e+00 -2.86712110e-01 -8.99901390e-01
2.31794879e-01 -2.77581483e-01 2.79546291e-01 -6.27635717e-02
-2.28664234e-01 1.85431931e-02 4.26700860e-02 -3.45116168e-01
-1.13461208e+00 -5.21212578e-01 -8.12324047e-01 -4.75447178e-01
-2.07081392e-01 1.23413779e-01 4.61086124e-01 1.09682894e+00
7.59179667e-02 2.83873975e-01 7.27825463e-01 -8.80866706e-01
3.53820950e-01 -4.85110432e-01 -5.98926544e-01 5.06476820e-01
2.76976407e-01 -6.82090402e-01 -3.12640369e-02 2.76436895e-01] | [14.801668167114258, 1.0316684246063232] |
af42efab-375e-48eb-bc80-1c7e41cb954f | improving-implicit-feedback-based | 2305.05585 | null | https://arxiv.org/abs/2305.05585v1 | https://arxiv.org/pdf/2305.05585v1.pdf | Improving Implicit Feedback-Based Recommendation through Multi-Behavior Alignment | Recommender systems that learn from implicit feedback often use large volumes of a single type of implicit user feedback, such as clicks, to enhance the prediction of sparse target behavior such as purchases. Using multiple types of implicit user feedback for such target behavior prediction purposes is still an open question. Existing studies that attempted to learn from multiple types of user behavior often fail to: (i) learn universal and accurate user preferences from different behavioral data distributions, and (ii) overcome the noise and bias in observed implicit user feedback. To address the above problems, we propose multi-behavior alignment (MBA), a novel recommendation framework that learns from implicit feedback by using multiple types of behavioral data. We conjecture that multiple types of behavior from the same user (e.g., clicks and purchases) should reflect similar preferences of that user. To this end, we regard the underlying universal user preferences as a latent variable. The variable is inferred by maximizing the likelihood of multiple observed behavioral data distributions and, at the same time, minimizing the Kullback-Leibler divergence (KL-divergence) between user models learned from auxiliary behavior (such as clicks or views) and the target behavior separately. MBA infers universal user preferences from multi-behavior data and performs data denoising to enable effective knowledge transfer. We conduct experiments on three datasets, including a dataset collected from an operational e-commerce platform. Empirical results demonstrate the effectiveness of our proposed method in utilizing multiple types of behavioral data to enhance the prediction of the target behavior. | ['Zhaochun Ren', 'Maarten de Rijke', 'Joemon Jose', 'Hengliang Luo', 'Xinlei Shi', 'Jiahuan Lei', 'Zhumin Chen', 'Pengjie Ren', 'Hanbing Wang', 'Xiangyuan Liu', 'Xin Xin'] | 2023-05-09 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 1.97317332e-01 -2.62862027e-01 -5.22413135e-01 -9.53360021e-01
-3.76884580e-01 -4.75497186e-01 3.74028653e-01 3.04221660e-02
-2.03333437e-01 4.78795201e-01 5.50770104e-01 -2.28900015e-01
-1.72444835e-01 -7.76067436e-01 -6.76502347e-01 -5.70046186e-01
1.35635752e-02 4.23436403e-01 -2.29819223e-01 -2.06381679e-01
3.89312088e-01 -1.55377358e-01 -1.56309068e+00 6.17014885e-01
1.13972354e+00 1.13134205e+00 3.63800108e-01 4.10676241e-01
-4.97078188e-02 7.60352850e-01 -1.62389964e-01 -5.13898790e-01
4.84839171e-01 -2.95448363e-01 -2.28781253e-01 1.67968288e-01
5.54784477e-01 -8.25370193e-01 -2.31068149e-01 1.04081511e+00
7.51497746e-02 3.89747530e-01 6.81622028e-01 -1.21760654e+00
-1.32174110e+00 5.39766490e-01 -2.87731022e-01 -3.80564434e-03
2.53805429e-01 3.39198820e-02 1.46153462e+00 -1.00781596e+00
2.87314981e-01 1.14916885e+00 6.07033074e-01 4.75994468e-01
-1.69104087e+00 -6.89362228e-01 5.76157689e-01 1.47665337e-01
-9.86148238e-01 -1.88719496e-01 8.68471622e-01 -3.98239136e-01
2.11332172e-01 5.69873929e-01 4.77241099e-01 1.37186122e+00
-3.26042712e-01 1.18816161e+00 1.15925479e+00 -5.24610057e-02
2.89412141e-01 9.52968895e-01 6.91664696e-01 4.50421751e-01
2.90902495e-01 1.56188294e-01 -4.83571947e-01 -5.38604558e-01
6.48133814e-01 7.88622320e-01 -2.31752947e-01 -3.96041781e-01
-7.26195991e-01 1.10316157e+00 3.31139177e-01 5.47153801e-02
-5.19330323e-01 -3.34679484e-01 -1.69950068e-01 7.88468719e-01
4.25075740e-01 3.34760070e-01 -9.18587983e-01 2.54403204e-02
-6.24685347e-01 2.18380287e-01 1.08361673e+00 1.01573575e+00
1.35896862e+00 2.63878424e-02 -1.00385562e-01 8.85094106e-01
6.67565703e-01 6.27661645e-01 6.22257233e-01 -7.31515229e-01
4.29008007e-01 7.78723419e-01 5.01739502e-01 -1.32172751e+00
1.97210297e-01 -4.62261677e-01 -6.48248672e-01 -1.77733585e-01
4.57019836e-01 -1.85071200e-01 -8.72623742e-01 1.97099090e+00
2.46405751e-01 3.05952519e-01 -3.15728724e-01 9.75592136e-01
3.86487991e-01 5.58135390e-01 -9.52652004e-03 -4.26822543e-01
7.71377385e-01 -7.79605865e-01 -6.38215840e-01 -6.27229735e-02
7.67019451e-01 -5.09533226e-01 1.48330784e+00 3.99149269e-01
-5.84586859e-01 -6.06315017e-01 -5.29566526e-01 1.65049568e-01
-2.85149515e-01 1.51708409e-01 7.57318020e-01 6.41215622e-01
-4.39283818e-01 4.05492991e-01 -7.41536617e-01 -1.49997324e-01
-5.93273304e-02 4.64836150e-01 -4.28242348e-02 8.33986849e-02
-1.06906581e+00 3.87224048e-01 -2.21559390e-01 -2.26020217e-02
-6.13786936e-01 -1.21538901e+00 -4.89215344e-01 1.37039140e-01
6.17064714e-01 -5.99011004e-01 1.34329438e+00 -1.82460856e+00
-1.42992628e+00 1.89776391e-01 -3.00334990e-01 -4.34480011e-01
1.83929086e-01 -7.14796245e-01 -5.98976016e-01 -5.87852120e-01
2.17231065e-02 -7.89777115e-02 7.40994334e-01 -1.39064062e+00
-8.30364466e-01 -7.95404315e-01 2.54613422e-02 2.02462420e-01
-8.51545274e-01 -5.26417494e-01 -2.98379481e-01 -6.87869668e-01
-9.72484052e-02 -1.00223243e+00 -3.06916326e-01 -3.17374140e-01
-6.05298532e-03 -1.82193115e-01 8.53025317e-01 -8.75826299e-01
1.37462616e+00 -2.06464076e+00 -1.16872636e-03 5.15016675e-01
1.73950389e-01 8.44243690e-02 -2.53358752e-01 1.59820825e-01
3.71116698e-01 2.96846479e-02 1.92136690e-01 -2.87110239e-01
1.84077874e-01 4.27411586e-01 -6.31676376e-01 2.43519112e-01
-4.23085183e-01 7.37920284e-01 -1.00563121e+00 8.96408632e-02
1.47522777e-01 1.36895806e-01 -1.06373167e+00 6.16224825e-01
-3.37369591e-01 3.75378817e-01 -7.27323115e-01 4.78592724e-01
5.58966458e-01 -4.33408290e-01 6.26101434e-01 -4.82403219e-01
1.88542642e-02 5.95317543e-01 -1.16967690e+00 1.30973101e+00
-5.44421792e-01 5.19603044e-02 8.09294060e-02 -1.10277677e+00
9.24351633e-01 2.51732439e-01 7.84370422e-01 -1.13648820e+00
-1.67806998e-01 7.44258538e-02 5.02933562e-02 -5.20490766e-01
3.87905389e-01 2.12114770e-02 1.30276487e-03 7.60606647e-01
1.20280847e-01 7.34012961e-01 -3.15664448e-02 1.91858441e-01
8.51978540e-01 -1.34173691e-01 3.26552421e-01 -5.89502417e-02
3.76806289e-01 -1.69170201e-01 9.42388833e-01 1.07795036e+00
9.37981457e-02 2.02350557e-01 -6.79288805e-02 -4.67378706e-01
-9.74478483e-01 -1.17237151e+00 2.13620692e-01 1.56169069e+00
1.24222524e-01 -4.29235190e-01 -1.19361229e-01 -9.58063424e-01
3.85384619e-01 1.04135740e+00 -9.12443161e-01 -1.79404989e-01
-3.11421216e-01 -5.76281250e-01 -8.55110064e-02 3.99737835e-01
2.19439894e-01 -8.64045680e-01 1.72251821e-01 1.95157692e-01
-1.60038650e-01 -5.47869265e-01 -9.28754568e-01 -5.89653030e-02
-1.13537455e+00 -1.02577651e+00 -2.72881538e-01 -3.88639182e-01
7.45136201e-01 6.39385164e-01 1.20121276e+00 7.30799837e-03
1.33934990e-01 6.80250645e-01 -4.45512861e-01 2.89887283e-02
-1.45808443e-01 -1.13419682e-01 4.65501517e-01 6.86409712e-01
9.59429145e-01 -8.45111191e-01 -7.95665801e-01 7.04466343e-01
-8.36621583e-01 -9.49042961e-02 8.19358766e-01 8.76066923e-01
5.65105796e-01 -3.88015419e-01 6.43278420e-01 -1.56624520e+00
7.26262212e-01 -1.10873628e+00 -4.63510782e-01 4.52085286e-01
-1.24784994e+00 1.77747346e-02 7.19379246e-01 -9.46822762e-01
-1.42450333e+00 -1.41758814e-01 8.78764614e-02 -3.91108721e-01
-4.13984984e-01 7.27377295e-01 -6.13058247e-02 4.78117734e-01
6.21762455e-01 4.47775155e-01 -8.36210996e-02 -9.99511421e-01
7.07148671e-01 9.19221997e-01 1.63473010e-01 -4.44788694e-01
7.07498193e-01 3.37113112e-01 -6.36486530e-01 -6.67797685e-01
-1.15574348e+00 -9.14742410e-01 -4.77953553e-01 5.03719226e-02
4.31597263e-01 -6.70761883e-01 -6.66515529e-01 2.99931358e-04
-5.57357788e-01 -2.44158611e-01 -1.39496550e-01 8.20506632e-01
-4.12671626e-01 2.97645807e-01 -5.56018054e-01 -9.67256963e-01
-3.47348660e-01 -7.01926589e-01 5.69978595e-01 2.31345862e-01
-4.85708654e-01 -1.35489035e+00 4.21760738e-01 4.67322767e-01
5.01913428e-01 -4.02480066e-01 8.35791409e-01 -1.00482237e+00
-5.93359053e-01 -2.99160719e-01 1.27979005e-02 5.41577876e-01
7.79555500e-01 -4.42645729e-01 -5.30758440e-01 -3.47039014e-01
2.16481701e-01 -3.86795282e-01 4.37739342e-01 3.19510698e-01
1.13356292e+00 -9.82386053e-01 -5.52210957e-02 4.96544212e-01
1.23201191e+00 2.01438278e-01 4.07552898e-01 7.75319105e-03
6.15583360e-01 6.02460504e-01 7.89461553e-01 5.91834188e-01
3.93395483e-01 6.98708773e-01 3.20400566e-01 1.58550203e-01
3.61641645e-01 -6.74354017e-01 7.21308589e-01 1.14268482e+00
-1.96482867e-01 1.88464880e-01 -6.96747378e-02 2.92455107e-01
-2.16950846e+00 -1.18117177e+00 -9.04404521e-02 2.54789758e+00
8.72925222e-01 -2.62358367e-01 1.98813036e-01 -4.20344114e-01
5.29108882e-01 -2.12165907e-01 -9.92741227e-01 -1.92398190e-01
2.09811911e-01 -1.66749954e-01 3.07494879e-01 4.01396096e-01
-6.05294228e-01 4.75434065e-01 5.46465588e+00 3.94833446e-01
-1.14120853e+00 1.82692736e-01 5.16413033e-01 -2.55086273e-01
-8.98398697e-01 -5.14922850e-02 -7.81034350e-01 7.14056313e-01
8.30181360e-01 -2.94623464e-01 7.88011193e-01 1.09653437e+00
4.42511380e-01 3.57163787e-01 -1.37336159e+00 9.08363640e-01
-6.47646934e-03 -1.05602062e+00 2.03756168e-01 4.12109256e-01
1.08136010e+00 7.63914222e-03 3.70906144e-01 8.19709659e-01
8.05488229e-01 -6.46859407e-01 4.83476877e-01 7.26626813e-01
2.77752489e-01 -1.97125450e-01 3.35741401e-01 7.14531243e-01
-9.65589404e-01 -3.95330846e-01 -3.31841230e-01 -1.13159716e-01
-4.83777896e-02 5.32159209e-01 -8.24241340e-01 2.71629661e-01
7.18856633e-01 1.03037071e+00 -3.61916572e-01 8.45008135e-01
1.76914930e-01 1.07744980e+00 -2.88189232e-01 -1.81270507e-03
9.77244303e-02 -8.07223439e-01 3.47563386e-01 8.72526586e-01
4.96336490e-01 9.49229151e-02 3.75514984e-01 1.13371134e+00
-2.40676284e-01 3.29685956e-01 -6.07889235e-01 -2.43882880e-01
3.02010715e-01 1.30385685e+00 1.48798928e-01 -3.78470659e-01
-1.00090277e+00 8.31761122e-01 4.51397300e-01 7.08829641e-01
-6.77788019e-01 3.65297705e-01 1.03739524e+00 2.61338294e-01
4.16317582e-01 -7.55536035e-02 -2.86884934e-01 -1.54077506e+00
3.42109352e-02 -9.15079713e-01 5.73249876e-01 -3.71072799e-01
-1.82808197e+00 -8.93137082e-02 -3.24457437e-01 -1.26782417e+00
-1.89563841e-01 -5.03749609e-01 -5.06047606e-01 9.04233634e-01
-1.03687489e+00 -7.74570465e-01 -2.71117151e-01 9.43193197e-01
7.32533813e-01 -2.10199580e-01 8.08712900e-01 5.38798869e-01
-1.63405374e-01 6.90578759e-01 4.04435515e-01 6.73327073e-02
8.12636673e-01 -1.33555162e+00 -1.13448597e-01 2.92823225e-01
4.25672889e-01 1.39519870e+00 7.07128882e-01 -8.34998548e-01
-1.80083966e+00 -1.15849614e+00 3.37031215e-01 -5.35094380e-01
7.36109078e-01 -1.69403121e-01 -1.04210389e+00 1.13732958e+00
-1.78008392e-01 -2.88752079e-01 1.27797127e+00 8.77877653e-01
-4.87795174e-01 -3.37347448e-01 -9.32487309e-01 6.24194741e-01
7.88663983e-01 -5.97941697e-01 -7.31675446e-01 3.35491560e-02
5.27429163e-01 3.46079946e-01 -8.98855865e-01 3.20595115e-01
8.60387206e-01 -9.26095009e-01 9.00751233e-01 -1.08211648e+00
2.43595004e-01 1.56292412e-02 -5.77754557e-01 -1.45444775e+00
-6.61920071e-01 -1.95598722e-01 -6.31962717e-01 1.20096850e+00
4.65986162e-01 -5.83324969e-01 1.02592015e+00 1.12085629e+00
5.77547587e-02 -6.47003651e-01 -9.92980748e-02 -4.19496745e-01
-4.09677237e-01 -2.69830674e-01 4.91077721e-01 1.12616694e+00
-7.94204623e-02 5.14861226e-01 -1.08042765e+00 1.34229109e-01
7.42908776e-01 6.14060998e-01 1.15373039e+00 -1.34178472e+00
-6.42866790e-01 -9.98669420e-04 1.81879550e-01 -1.78578746e+00
-1.81881562e-01 -7.47562766e-01 -1.97499752e-01 -1.03891003e+00
3.13035458e-01 -5.68100214e-01 -8.15496445e-01 1.76163778e-01
-1.27827331e-01 -2.26445179e-02 3.81023474e-02 4.32538837e-01
-6.67832494e-01 7.09418654e-01 1.19173503e+00 -1.63994908e-01
-5.13416409e-01 5.73619723e-01 -1.20863426e+00 7.58864164e-01
6.27641439e-01 -4.08404380e-01 -7.83822060e-01 -1.76659971e-01
2.63696700e-01 1.70816794e-01 8.92183185e-02 -2.15973049e-01
6.66793063e-02 -6.88304663e-01 2.89705366e-01 -5.09909987e-01
3.73651326e-01 -1.13828492e+00 6.18846193e-02 2.90809005e-01
-7.78761685e-01 -1.64290980e-01 -4.55335766e-01 1.18711531e+00
-2.05095291e-01 4.90779104e-03 1.92248151e-01 -2.77337104e-01
-8.30026567e-01 6.85317218e-01 -2.89390236e-01 -1.43563136e-01
6.29811049e-01 -8.49442258e-02 -1.36527523e-01 -6.29392862e-01
-1.00485384e+00 1.72130600e-01 2.33042285e-01 8.14694941e-01
5.63934445e-01 -1.54686475e+00 -4.26526487e-01 3.11090112e-01
4.31373641e-02 -6.66311264e-01 1.39450848e-01 8.97374988e-01
4.68136936e-01 3.61187845e-01 -9.16877165e-02 -3.85825664e-01
-1.41939116e+00 5.42357624e-01 6.04114234e-02 -2.80714154e-01
-2.47417063e-01 3.52578461e-01 5.66784441e-01 -9.86277878e-01
1.44005477e-01 -1.24626175e-01 -1.89667612e-01 -4.67820466e-02
6.10652328e-01 3.58682513e-01 -1.94996417e-01 -4.31102812e-01
1.35930941e-01 1.73935629e-02 -4.87996489e-01 7.76017830e-02
1.51926076e+00 -5.31021893e-01 1.35583252e-01 7.60729074e-01
1.36778307e+00 2.14797348e-01 -1.31684315e+00 -8.42654943e-01
4.45632311e-03 -1.05937302e+00 -1.17607981e-01 -8.86967301e-01
-1.02645969e+00 5.33553541e-01 5.23511887e-01 5.66169739e-01
8.53915572e-01 -2.93251783e-01 7.81933486e-01 6.32481039e-01
2.78886288e-01 -1.29912984e+00 3.48452926e-01 2.39862293e-01
5.35174608e-01 -1.50415206e+00 -2.17157081e-01 -8.27142820e-02
-9.18984056e-01 8.55097413e-01 9.41773772e-01 -1.09322824e-01
9.30725574e-01 -2.28844017e-01 1.69415906e-01 5.58958529e-03
-1.08078468e+00 6.83386102e-02 7.62592137e-01 3.26798648e-01
6.42196894e-01 3.28342050e-01 -1.25940114e-01 9.69395697e-01
3.72781456e-02 1.64876819e-01 2.92189457e-02 6.47539556e-01
-5.56188703e-01 -1.16217816e+00 -3.42343077e-02 1.03924656e+00
-1.60862461e-01 -7.22694770e-02 -2.45096222e-01 2.78679579e-01
-5.73006682e-02 8.61879766e-01 -9.80859101e-02 -7.00725019e-01
1.45170674e-01 2.01164335e-02 1.55342937e-01 -7.99272895e-01
-3.88318479e-01 -8.35483894e-02 -1.73347592e-01 -7.58370697e-01
-3.85343194e-01 -6.46357894e-01 -6.67079449e-01 -4.09300804e-01
-3.46093118e-01 4.02346969e-01 4.70724791e-01 9.98187661e-01
4.09266800e-01 8.36919025e-02 9.98597205e-01 -4.91162360e-01
-1.36302125e+00 -1.00038683e+00 -7.86329031e-01 1.31017542e+00
2.27338195e-01 -6.42342269e-01 -4.15797472e-01 2.87288249e-01] | [10.017494201660156, 5.567943572998047] |
6e3f994b-ff2a-4e4c-a9a7-106d3d58f1a8 | fully-digital-second-order-level-crossing | 2211.09848 | null | https://arxiv.org/abs/2211.09848v2 | https://arxiv.org/pdf/2211.09848v2.pdf | Fully Digital Second-order Level-crossing Sampling ADC for Data Saving in Sensing Sparse Signals | This paper presents a fully integrated second-order level-crossing sampling data converter for real-time data compression and feature extraction. Compared with level-sampling ADCs which sample at fixed voltage levels, the proposed circuits updates tracking thresholds using linear extrapolation, which forms a second-order level-crossing sampling ADC that has sloped sampling levels. The computing is done digitally and is implemented by modifying the digital control logic of a conventional SAR ADC. The system selects only the turning points in the input waveform for quantization. The output of the proposed data converter consists of both the digital value of the selected sampling points and the timestamp between the selected sampling points. The main advantages are data savings and power savings for the data converter and the following digital signal processing or communication circuits, which are ideal for low-power sensors. The test chip was fabricated using a 180nm CMOS process. The proposed ADC saves 30% compared to a conventional SAR ADC and achieves a compression factor of 6.17 for tracking ECG signals. | ['Wei Tang', 'Jaime Ramirez-Angulo', 'Xiaochen Tang', 'Mario Renteria-Pinon'] | 2022-11-17 | null | null | null | null | ['data-compression'] | ['time-series'] | [ 7.81729579e-01 -2.11388320e-01 -3.52342516e-01 -4.45175916e-01
-5.00824630e-01 -4.73253310e-01 5.63767105e-02 9.28754926e-01
-8.26860309e-01 5.90549946e-01 -1.69535950e-01 -2.22915024e-01
1.31506085e-01 -7.42400050e-01 -1.98768124e-01 -3.70046824e-01
-1.13329142e-01 -1.15259755e-02 6.11971855e-01 1.69940680e-01
4.22556251e-01 6.08825207e-01 -1.07309353e+00 3.93249363e-01
8.07273507e-01 1.48752964e+00 -7.40951896e-02 8.95496428e-01
2.60836244e-01 4.08343852e-01 -8.56616318e-01 1.08730063e-01
3.03510666e-01 -6.54988229e-01 -1.53659984e-01 -5.72337985e-01
3.25790256e-01 -8.78847182e-01 -4.45740640e-01 1.10108399e+00
4.88766164e-01 -4.02764052e-01 5.74376702e-01 -1.02006137e+00
4.40788865e-02 7.75530219e-01 -4.84703064e-01 4.31909531e-01
4.05718297e-01 -9.74497348e-02 1.85533196e-01 -3.86666477e-01
3.43276888e-01 8.82309854e-01 9.32586730e-01 -8.24728757e-02
-1.24173093e+00 -9.17366803e-01 -6.98617339e-01 2.54312992e-01
-1.59911776e+00 -2.33310536e-01 8.14973652e-01 -2.98446208e-01
8.45504701e-01 3.84783387e-01 9.59830582e-01 2.01424837e-01
9.26179767e-01 -2.97825933e-02 1.28737342e+00 -3.26937109e-01
6.35455072e-01 1.30693112e-02 4.02975678e-01 3.05761576e-01
7.64132082e-01 -2.73519903e-01 -3.56902748e-01 -4.60471213e-01
1.19118154e+00 2.74217218e-01 -4.28682983e-01 2.16735274e-01
-1.00571799e+00 3.58710289e-01 5.34793198e-01 3.75320435e-01
-5.72531700e-01 5.64605415e-01 4.27261680e-01 2.18365446e-01
-5.07225573e-01 2.84067065e-01 -6.95720837e-02 -2.76463270e-01
-1.53970599e+00 -1.70477912e-01 6.18014693e-01 7.93557346e-01
8.45104977e-02 4.99888659e-01 5.87037094e-02 -1.43189102e-01
3.56770128e-01 6.16790533e-01 8.89781713e-01 -6.76523328e-01
1.39645591e-01 7.25493491e-01 2.27220342e-01 -7.55733073e-01
-4.67631340e-01 -1.34459332e-01 -9.37029898e-01 2.54519641e-01
4.46936429e-01 -2.93079704e-01 -7.60832667e-01 1.03615892e+00
5.23222424e-02 -2.26035506e-01 1.13079414e-01 7.22455442e-01
3.50245386e-01 8.14762533e-01 -9.31219757e-02 -7.41501153e-01
1.77700233e+00 -7.13622421e-02 -1.42872727e+00 -8.29580426e-02
-2.69304901e-01 -6.03740513e-01 7.55834877e-01 6.64067805e-01
-1.26881135e+00 -9.10843849e-01 -2.04593897e+00 -3.76084417e-01
1.00586899e-01 3.94808412e-01 -3.21023422e-03 5.66180110e-01
-6.66420758e-01 7.04869509e-01 -9.91778910e-01 3.41319107e-02
2.23295942e-01 4.49613601e-01 3.24328423e-01 8.01791966e-01
-1.11605620e+00 5.16755760e-01 6.31956458e-02 -8.41489360e-02
-2.27757499e-01 -9.54974115e-01 -5.19006252e-01 1.77922651e-01
-4.62706476e-01 -8.23193640e-02 1.13670588e+00 -9.25528824e-01
-1.37358367e+00 3.45254868e-01 -1.55616432e-01 -1.15053248e+00
5.16096115e-01 -9.07358602e-02 -7.32366681e-01 6.95859373e-01
-4.52174544e-01 9.44130570e-02 1.06371605e+00 -2.87382871e-01
-4.90468919e-01 -3.72624695e-01 -7.06875563e-01 -2.31894925e-01
-1.67371169e-01 -4.87805367e-01 1.38070047e-01 -5.72416902e-01
3.32531124e-01 -3.95198286e-01 -3.66377890e-01 4.52614099e-01
1.41271338e-01 5.81248164e-01 9.37216759e-01 -4.53511953e-01
1.82201993e+00 -2.38368464e+00 -6.97930872e-01 3.70995820e-01
-1.40865371e-01 4.44273084e-01 8.15499008e-01 3.74657273e-01
3.12000394e-01 -4.59446520e-01 -3.44985962e-01 4.56527621e-01
-5.27292550e-01 -4.79899317e-01 -3.11522156e-01 6.94794774e-01
2.99105734e-01 5.38372755e-01 -5.80045164e-01 -2.31064335e-01
1.16968807e-02 8.58170629e-01 -1.48664042e-01 -2.52227873e-01
3.74844164e-01 4.59536426e-02 -2.03591600e-01 4.17082906e-01
6.47174358e-01 2.62306929e-01 6.72604859e-01 -1.00434947e+00
-3.49630058e-01 7.83954740e-01 -1.41462421e+00 1.28155339e+00
-5.53046204e-02 8.46526027e-01 2.14974806e-02 -2.58487999e-01
1.56697071e+00 2.79977828e-01 5.70866503e-02 -7.06442177e-01
5.73281467e-01 4.44028407e-01 1.15299568e-01 -2.08156660e-01
5.83506107e-01 -1.86554253e-01 -3.49754333e-01 -9.84520689e-02
-3.66370320e-01 -4.71582323e-01 -1.69918820e-01 6.62348792e-02
7.90320992e-01 -1.94438890e-01 1.02261686e+00 -5.06123126e-01
5.06731689e-01 5.22735864e-02 6.08385503e-01 5.75979315e-02
-3.83294672e-01 4.83435765e-02 3.60647649e-01 -2.96477109e-01
-1.01417530e+00 -1.08131123e+00 -3.93045217e-01 1.18574882e-02
1.83968857e-01 -3.72013897e-01 -7.19959855e-01 1.97262809e-01
5.04709482e-02 7.20992684e-01 1.08838424e-01 -2.55414486e-01
-7.00935960e-01 -3.42739001e-02 8.67395163e-01 6.96196437e-01
6.92216635e-01 -3.17295879e-01 -1.95500231e+00 6.90867305e-01
1.85646147e-01 -1.02485335e+00 -6.63981497e-01 4.81416494e-01
-1.62720299e+00 -7.05625474e-01 -3.82925756e-02 -9.00005221e-01
5.63543856e-01 -4.30292487e-01 3.10846984e-01 -2.86283940e-01
-5.65597713e-01 -1.46837071e-01 1.12242348e-01 -7.79404581e-01
-1.93463981e-01 -4.93188471e-01 9.34356302e-02 -9.65894759e-02
6.60404384e-01 -5.94442010e-01 -1.12964141e+00 -1.51078284e-01
-5.62886178e-01 -3.43244433e-01 7.36092269e-01 4.74991530e-01
8.30534220e-01 4.33011092e-02 9.88172770e-01 -5.53815544e-01
5.48981547e-01 3.18226337e-01 -1.05029023e+00 -3.79894823e-01
-6.27434373e-01 -1.56237632e-02 1.25295985e+00 -5.00199735e-01
-5.63147306e-01 6.31350994e-01 1.43744973e-02 3.25838700e-02
2.64641404e-01 -4.14697193e-02 8.57645937e-04 -4.39610258e-02
7.65577555e-01 1.99151635e-01 5.17405212e-01 -2.17173129e-01
-1.25350252e-01 1.13070714e+00 1.07188165e+00 2.98121661e-01
1.92020297e-01 3.59373778e-01 3.16949964e-01 -9.25503552e-01
5.51844418e-01 -1.79266945e-01 -5.46802819e-01 -7.12729096e-02
6.11467063e-01 -1.13700223e+00 -8.15544069e-01 4.17398423e-01
-8.48524630e-01 3.00169826e-01 -5.59905708e-01 6.95993423e-01
-3.60296726e-01 2.09841549e-01 -9.82164443e-01 -9.55682278e-01
-1.16263950e+00 -9.51595664e-01 4.94491726e-01 5.83394468e-01
-6.34579360e-01 -2.90339172e-01 -4.43605214e-01 -4.04403150e-01
4.26920861e-01 6.17464602e-01 8.53762984e-01 -2.02747911e-01
-3.22722971e-01 -5.84815800e-01 4.08835888e-01 9.41314474e-02
1.22886479e-01 1.00730825e-02 -7.55351603e-01 -4.15981025e-01
6.50342762e-01 4.95165512e-02 3.88462156e-01 4.25498128e-01
6.50671542e-01 -3.95514905e-01 -5.93743205e-01 4.33019459e-01
1.95532417e+00 9.52462196e-01 9.00114059e-01 -3.23210418e-01
-4.38765064e-02 -1.30792141e-01 7.54307866e-01 5.88509381e-01
-5.29741589e-03 1.94942698e-01 1.16845220e-02 -1.01948380e-01
1.35380030e-01 -3.09884250e-01 4.02717352e-01 8.73095095e-01
1.61607042e-01 3.82152081e-01 -5.65247953e-01 4.83904630e-01
-9.25530314e-01 -8.36516261e-01 -6.14212751e-01 2.67014694e+00
1.28230846e+00 3.11884582e-01 -1.15393415e-01 8.80885661e-01
6.19338334e-01 -3.08926284e-01 -7.87715971e-01 -1.10855174e+00
4.44747835e-01 9.18658555e-01 9.78970230e-01 3.94095898e-01
-6.02569401e-01 -9.51641947e-02 5.59597635e+00 2.89319336e-01
-1.53453112e+00 -2.05892473e-01 2.23219991e-01 -1.88619167e-01
2.71010816e-01 -3.77814136e-02 -8.32396030e-01 4.36445415e-01
1.27300417e+00 -3.62892449e-01 -1.54905900e-01 7.26566255e-01
4.41620409e-01 -4.82188225e-01 -1.20561647e+00 9.27608430e-01
-3.95866662e-01 -1.04444897e+00 1.16733454e-01 -1.98936477e-01
2.37982363e-01 -7.93080568e-01 -1.54976040e-01 -2.70005286e-01
-7.20593572e-01 -5.77669740e-01 7.65727341e-01 4.81175125e-01
1.53555417e+00 -9.70379770e-01 5.02491057e-01 1.20143238e-02
-1.48853958e+00 -3.43212306e-01 -3.33148986e-01 -4.24281746e-01
2.25033075e-01 7.27188408e-01 -1.17720985e+00 8.87966156e-02
4.04835075e-01 -5.91051839e-02 -2.24488556e-01 8.14883292e-01
2.14616701e-01 7.59545028e-01 -4.44365859e-01 -3.85682285e-01
-2.22124219e-01 -1.22336611e-01 3.15117568e-01 1.35962403e+00
3.85361701e-01 5.54123700e-01 -4.03106004e-01 4.32234049e-01
1.95259809e-01 -2.03851059e-01 -5.09236678e-02 -6.48578480e-02
1.16630137e+00 9.81635511e-01 -5.20793855e-01 -7.64849663e-01
-1.04082219e-01 7.60114253e-01 -5.43906212e-01 -3.71857256e-01
-8.77167642e-01 -1.66641307e+00 2.27172628e-01 6.35954142e-01
2.34088615e-01 -3.78202468e-01 -6.24368548e-01 -1.16751678e-01
1.51475258e-02 -6.88934028e-01 3.46253246e-01 -2.30277568e-01
-4.00498569e-01 4.83472377e-01 -2.64419466e-01 -1.56598055e+00
-2.99464732e-01 -1.31527856e-01 -4.59096164e-01 8.25369656e-01
-1.07412148e+00 -1.22945033e-01 -6.78764582e-01 5.58086991e-01
3.48329842e-01 4.52375531e-01 8.38419378e-01 1.85097709e-01
4.15772647e-01 6.84121132e-01 1.86733920e-02 1.62732869e-01
6.80583000e-01 -8.83751750e-01 3.28703016e-01 8.74989629e-01
-6.98937356e-01 7.50306070e-01 6.14665210e-01 -9.12558854e-01
-1.76941705e+00 -1.04538929e+00 8.52425516e-01 6.88370883e-01
3.00558269e-01 -4.51615125e-01 -9.49645638e-01 2.59081393e-01
1.78308785e-01 -9.22480002e-02 4.88766313e-01 -1.40188718e+00
5.89063130e-02 -8.75114501e-01 -1.91144466e+00 3.56010348e-01
1.76901788e-01 -3.39866638e-01 -7.66798854e-01 -5.36201537e-01
4.66101229e-01 -4.46242750e-01 -1.15982044e+00 2.70269722e-01
9.58994329e-01 -5.12539864e-01 5.59174061e-01 4.15935993e-01
-5.60746295e-03 -8.49341989e-01 1.16263945e-02 -7.97012806e-01
-2.83651263e-01 -7.46918201e-01 1.82598114e-01 1.01152277e+00
2.74816573e-01 -6.95027053e-01 3.65684837e-01 3.41939300e-01
3.35855126e-01 -6.67163134e-01 -1.03854704e+00 -3.67339551e-01
-5.20568490e-01 3.21533352e-01 3.77297968e-01 2.87601113e-01
7.07837999e-01 2.45912120e-01 5.58020696e-02 3.87811333e-01
7.82580078e-01 -2.41271198e-01 -1.08505175e-01 -1.00734234e+00
-2.34082773e-01 -3.15356292e-02 -1.04450369e+00 -9.32394862e-01
-1.25397360e+00 -3.80431235e-01 -1.27066582e-01 -1.23071480e+00
-5.50701380e-01 1.15578147e-02 -1.26517519e-01 1.40858769e-01
1.48193285e-01 4.11397338e-01 -6.07094355e-02 -7.29115854e-04
2.97631025e-01 -2.07992062e-01 8.12240601e-01 -3.92645411e-03
-7.32756793e-01 -9.76906940e-02 -8.47821534e-02 7.07537651e-01
8.46323609e-01 -1.04175329e-01 -6.65927649e-01 2.11243436e-01
-2.04577670e-01 5.93741000e-01 1.47282615e-01 -1.65652847e+00
5.60446560e-01 3.84911954e-01 1.15073085e+00 -1.04530334e+00
3.65712821e-01 -1.27573991e+00 6.60051763e-01 1.45952106e+00
-2.73491502e-01 5.57244003e-01 1.98741764e-01 2.56804883e-01
-8.16888139e-02 -1.42209321e-01 1.37516797e+00 2.45097011e-01
-3.86566401e-01 -5.60972214e-01 -7.29850233e-01 -2.16557413e-01
1.21925521e+00 -6.64989591e-01 -2.23152205e-01 -3.20552066e-02
-5.01099229e-01 -1.18186258e-01 4.32642072e-01 -1.92500338e-01
1.03617394e+00 -1.21172750e+00 -1.24499820e-01 5.64615667e-01
-4.86435354e-01 -3.63613814e-02 8.55735168e-02 7.53333688e-01
-9.44469571e-01 3.43284845e-01 -7.44816184e-01 -6.44431114e-01
-1.43502510e+00 1.68008074e-01 3.91526014e-01 2.00792164e-01
-6.76847696e-01 2.00380698e-01 -1.18045688e+00 1.18069077e+00
8.32808688e-02 -1.11885023e+00 1.41190300e-02 3.97644877e-01
1.05359399e+00 8.01513314e-01 7.33161941e-02 4.20442298e-02
-6.23363256e-01 6.79080188e-01 1.30201668e-01 -2.87082493e-01
9.92179930e-01 1.02366045e-01 -5.30339368e-02 9.16060030e-01
1.15306449e+00 7.58354440e-02 -1.00911355e+00 1.87762722e-01
-1.05466835e-01 -1.98278368e-01 -5.56741431e-02 -7.89783716e-01
-6.64420784e-01 7.18961656e-01 9.84977543e-01 -7.77882487e-02
1.83460867e+00 -9.32117820e-01 9.26346660e-01 -9.46738198e-02
7.51186609e-01 -1.15196466e+00 -3.04953575e-01 -1.51062459e-01
6.20622277e-01 -1.31472712e-02 6.55103266e-01 -1.28908902e-01
-4.18166727e-01 1.62739289e+00 3.06243658e-01 -7.01237738e-01
4.15458918e-01 1.01447964e+00 3.89565155e-02 4.62403715e-01
-4.57102418e-01 6.30293429e-01 -9.93125737e-02 6.29327655e-01
4.89544153e-01 1.56894252e-01 -1.29143476e+00 8.03581536e-01
-2.74407953e-01 5.75342774e-01 1.18052709e+00 1.09777737e+00
-7.98058391e-01 -4.30406511e-01 -5.57621479e-01 7.31635273e-01
-7.38110840e-01 1.86157495e-01 -1.52681470e-01 4.15111303e-01
-4.97085741e-03 8.26639891e-01 6.15163743e-01 -2.74049938e-01
6.24053717e-01 1.09358259e-01 5.49366891e-01 1.73166290e-01
-9.72890675e-01 2.76998311e-01 -2.07855418e-01 -7.01816082e-01
1.48044989e-01 -4.15458381e-01 -2.12556839e+00 2.22098663e-01
-1.65762350e-01 1.33443609e-01 1.02409482e+00 1.24369331e-01
3.87476206e-01 8.47919822e-01 4.96815592e-01 -9.09193754e-02
-9.22947884e-01 -7.57973552e-01 -5.23607552e-01 -2.63568193e-01
7.61979461e-01 5.73401749e-02 -2.01861501e-01 2.48948261e-01] | [13.94454288482666, 3.1611549854278564] |
20e6bf2e-a9ac-4e5a-a9fa-c436e1291559 | tutoring-instruction-grounded-conversational | 2302.12623 | null | https://arxiv.org/abs/2302.12623v1 | https://arxiv.org/pdf/2302.12623v1.pdf | TUTORING: Instruction-Grounded Conversational Agent for Language Learners | In this paper, we propose Tutoring bot, a generative chatbot trained on a large scale of tutor-student conversations for English-language learning. To mimic a human tutor's behavior in language education, the tutor bot leverages diverse educational instructions and grounds to each instruction as additional input context for the tutor response generation. As a single instruction generally involves multiple dialogue turns to give the student sufficient speaking practice, the tutor bot is required to monitor and capture when the current instruction should be kept or switched to the next instruction. For that, the tutor bot is learned to not only generate responses but also infer its teaching action and progress on the current conversation simultaneously by a multi-task learning scheme. Our Tutoring bot is deployed under a non-commercial use license at https://tutoringai.com. | ['Jinyoung Yeo', 'Junmyung Lee', 'Hyejoong Kim', 'Wonseok Jeong', 'Chaehyeong Kim', 'Minjin Kim', 'Hyungjoo Chae'] | 2023-02-24 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [ 3.21996808e-01 6.25031292e-01 -2.78646588e-01 -2.71546483e-01
-6.48947120e-01 -9.78612363e-01 5.11644721e-01 -1.46589383e-01
-2.69611012e-02 9.33221221e-01 -2.48106811e-02 -9.03598487e-01
5.59532881e-01 -7.82928467e-01 -4.16891843e-01 -4.12621886e-01
4.87642229e-01 7.04029500e-01 5.01118600e-01 -4.73460615e-01
4.34816889e-02 3.19492072e-01 -1.28101611e+00 2.80951113e-01
1.14696705e+00 2.04537272e-01 3.22675347e-01 1.30346167e+00
-2.03485891e-01 1.71366525e+00 -1.16299653e+00 -2.56338835e-01
-1.86191380e-01 -9.53961372e-01 -1.17880559e+00 2.41779648e-02
2.43265927e-01 -8.83834004e-01 -3.62447113e-01 7.92022884e-01
3.54162186e-01 3.00300777e-01 5.43323457e-01 -1.72044122e+00
-3.56569141e-01 1.00105679e+00 1.95783898e-01 2.11437196e-01
6.41677916e-01 6.81116283e-01 8.42112124e-01 -2.02972069e-01
5.16149879e-01 1.15930974e+00 2.40716361e-03 1.01037109e+00
-8.08181942e-01 -8.77305567e-01 6.25468642e-02 -1.41452821e-02
-3.24724317e-01 -4.09174085e-01 7.37665832e-01 -4.76147830e-01
4.55064714e-01 1.01640113e-01 8.16165626e-01 1.66726887e+00
3.49351674e-01 1.16686571e+00 1.16765285e+00 -3.96149695e-01
3.80061306e-02 3.40774506e-01 1.83819398e-01 8.39793086e-01
-2.43817300e-01 -2.16814265e-01 -4.38651681e-01 -1.95061579e-01
8.07873249e-01 -9.36645195e-02 -1.19011968e-01 1.45814657e-01
-8.34011018e-01 9.76373315e-01 -3.29582877e-02 2.47327536e-01
-3.84377927e-01 2.34452173e-01 2.63874590e-01 8.54223847e-01
-4.43242956e-03 6.93856299e-01 -3.81131709e-01 -9.25817668e-01
-2.98215717e-01 4.20361131e-01 1.46499574e+00 1.15053868e+00
8.09628367e-01 1.57169834e-01 -4.96755064e-01 5.55907369e-01
2.22389966e-01 4.23612893e-01 4.81071979e-01 -1.32189202e+00
4.12306249e-01 9.48387027e-01 -3.75221036e-02 -2.64819473e-01
1.14997171e-01 1.17243253e-01 -1.80878311e-01 2.19502226e-02
5.59449613e-01 -8.64212513e-01 -4.68092680e-01 1.38104558e+00
5.45604110e-01 4.52140898e-01 7.50695169e-02 4.28153336e-01
1.04005659e+00 4.29575592e-01 4.88126017e-02 -1.49800461e-02
1.35196459e+00 -1.17487180e+00 -6.59544706e-01 -2.84002811e-01
8.98611844e-01 -8.88299763e-01 9.90794420e-01 2.87900537e-01
-1.06896579e+00 -2.35387146e-01 -5.49875140e-01 1.29889861e-01
-3.47629301e-02 -2.12094530e-01 3.69188517e-01 4.09329236e-01
-8.77272189e-01 2.98233122e-01 -7.22265005e-01 -6.41488433e-02
1.23966306e-01 1.34502172e-01 1.43189356e-01 5.94718717e-02
-1.05621743e+00 6.68331623e-01 -2.97848344e-01 -4.63481545e-01
-1.41254556e+00 -6.46407902e-01 -8.26859236e-01 2.61283278e-01
5.10617554e-01 -2.46161461e-01 2.46297717e+00 -8.72567177e-01
-2.43033671e+00 6.36628509e-01 3.32884714e-02 -1.69128552e-01
6.93796396e-01 -6.64614961e-02 3.17815363e-01 1.98935121e-01
2.39791229e-01 4.17012721e-01 7.04056144e-01 -9.93028462e-01
-6.83184147e-01 5.45369163e-02 4.15147781e-01 3.60409588e-01
4.03760634e-02 8.57200474e-02 1.56854242e-01 -7.34791681e-02
-5.59626162e-01 -1.14083922e+00 5.08691333e-02 -7.60541677e-01
-3.49724174e-01 -7.15944529e-01 1.17003155e+00 -1.56102806e-01
8.96711409e-01 -1.64939463e+00 -4.31125686e-02 -1.26704574e-01
4.22033876e-01 4.07505482e-01 -3.38224530e-01 6.60148263e-01
2.94378191e-01 2.56436318e-02 2.70288348e-01 -8.03254396e-02
3.82979773e-02 2.93701321e-01 -3.00273985e-01 7.46234804e-02
1.00047022e-01 8.59736979e-01 -1.24953651e+00 -4.60333347e-01
6.66754320e-02 1.07709870e-01 -5.56855202e-01 1.18571758e+00
-5.52955270e-01 9.11563933e-01 -9.99177158e-01 4.73468572e-01
-6.11312455e-03 -2.67273813e-01 1.94969878e-01 8.69546950e-01
-3.70248146e-02 9.74616408e-01 -6.40931129e-01 1.00505209e+00
-8.53743792e-01 6.39204860e-01 5.95798850e-01 -8.05552185e-01
8.42365503e-01 7.29457796e-01 2.77943879e-01 -2.10619748e-01
3.28621060e-01 2.95650475e-02 5.28186202e-01 -6.33968771e-01
3.42349976e-01 3.84890378e-01 -1.55151218e-01 1.32262325e+00
2.92871475e-01 -3.65860254e-01 1.16174277e-02 4.98708278e-01
1.53428996e+00 2.98823416e-01 2.10619003e-01 1.70757845e-01
4.06312883e-01 5.53747416e-02 3.25009346e-01 8.71464312e-01
-4.35818017e-01 -1.91041842e-01 8.63529444e-01 -6.86166212e-02
-5.12369990e-01 -5.86426795e-01 5.94775558e-01 1.88452542e+00
-3.57063025e-01 -1.80712536e-01 -1.05304098e+00 -1.22170961e+00
-4.22266126e-02 1.00398934e+00 -2.47455478e-01 -2.34996751e-01
-8.59549522e-01 4.15068090e-01 7.54086494e-01 1.05799742e-01
4.12724674e-01 -1.71376085e+00 -1.03844750e+00 4.75451231e-01
-3.55356634e-01 -9.34985340e-01 -1.09131980e+00 3.58726114e-01
-4.16108876e-01 -1.33585072e+00 -3.10362488e-01 -8.46002817e-01
5.70316494e-01 3.01728398e-01 1.15762901e+00 6.20931983e-01
8.17576796e-02 8.44458401e-01 -3.18534195e-01 -4.22669649e-01
-1.23658502e+00 4.72439528e-01 -2.57780820e-01 -3.64074349e-01
3.65017056e-01 -7.05128610e-01 -2.86209464e-01 3.59950602e-01
-4.00261432e-01 2.68968433e-01 3.92283410e-01 8.66556704e-01
-3.06600153e-01 -5.39634824e-01 9.39849317e-01 -1.18743920e+00
1.15783608e+00 -6.07467055e-01 -7.87696898e-01 2.35645056e-01
-4.19389904e-01 -7.99580198e-03 1.02119029e+00 -6.92819476e-01
-1.14129746e+00 -9.23216194e-02 4.97008413e-02 -4.57873523e-01
-6.89077854e-01 1.01913415e-01 1.63009819e-02 4.98992614e-02
5.00516117e-01 5.10489881e-01 1.21946566e-01 -1.28488243e-01
3.33027504e-02 1.04855001e+00 2.94484079e-01 -1.06050706e+00
6.47405267e-01 -6.43153787e-01 -4.71664160e-01 -4.90939111e-01
-7.48165131e-01 -1.97512954e-01 -5.43287337e-01 -7.57979393e-01
6.55880213e-01 -8.13410640e-01 -1.61898530e+00 4.54941362e-01
-1.26876068e+00 -1.39580250e+00 2.02958025e-02 -9.22922790e-03
-5.61739206e-01 -3.60492766e-01 -9.13113773e-01 -8.67270648e-01
-1.68265387e-01 -1.73908067e+00 6.61407769e-01 6.25972927e-01
-4.84922647e-01 -1.16844940e+00 1.36417046e-01 7.22129881e-01
4.29854155e-01 -2.53659785e-01 6.32135272e-01 -1.22041881e+00
-7.46698022e-01 1.04959846e-01 4.50634390e-01 -1.86817860e-03
3.23904008e-01 3.22516888e-01 -8.69952202e-01 -3.63897458e-02
-1.62086841e-02 -9.14533198e-01 -1.32982522e-01 -1.32007524e-01
5.84724128e-01 -8.83490384e-01 -1.83148533e-02 1.02078699e-01
3.85756344e-01 6.00830376e-01 5.57124726e-02 2.48112306e-01
7.54195809e-01 6.65098965e-01 4.63652402e-01 9.46457461e-02
6.74236000e-01 4.87709910e-01 2.24379331e-01 5.24693012e-01
-7.21758930e-03 -6.81842804e-01 8.87578249e-01 9.99402404e-01
3.24245274e-01 -1.56662658e-01 -1.02324533e+00 4.13603812e-01
-1.64843416e+00 -9.05275106e-01 -8.48720670e-02 1.78391004e+00
1.51317322e+00 -8.60457122e-02 2.91740179e-01 -5.41751266e-01
4.75638002e-01 6.20352253e-02 -6.57714486e-01 -8.13695014e-01
7.64379561e-01 3.02640051e-01 2.73735851e-01 7.70966291e-01
-3.85580480e-01 1.22704983e+00 5.85313129e+00 3.99603933e-01
-1.27859950e+00 1.23604713e-02 3.55446935e-01 2.35190377e-01
-2.02248126e-01 2.93948986e-02 -1.06188667e+00 4.51567829e-01
1.40832329e+00 -4.89059269e-01 6.65197909e-01 9.11539614e-01
2.18817532e-01 -1.78220466e-01 -1.18505752e+00 1.72957301e-01
-2.31862336e-01 -1.07734561e+00 -2.87505120e-01 2.14658290e-01
5.18589616e-01 -1.90937012e-01 -1.38739824e-01 7.67508209e-01
1.47108352e+00 -1.02388191e+00 2.63280421e-01 -3.35247293e-02
3.83686364e-01 -6.61896050e-01 2.84450501e-01 9.73424554e-01
-8.19165409e-01 -7.84291606e-03 1.77879721e-01 -2.76154488e-01
-2.98433900e-01 -4.78231579e-01 -1.73852146e+00 -2.52711810e-02
1.16390504e-01 3.44592154e-01 -3.80394250e-01 4.06200290e-01
-9.42330003e-01 1.43766892e+00 1.59585848e-01 -4.45886284e-01
3.44508350e-01 -4.11839396e-01 4.40816462e-01 9.38079834e-01
9.12916437e-02 5.10639727e-01 7.21490741e-01 8.22540998e-01
-4.06886518e-01 -2.21882075e-01 -8.63469899e-01 -1.74407125e-01
8.13637316e-01 1.42522967e+00 -3.17814976e-01 -5.28898060e-01
-2.94315666e-01 9.65334594e-01 4.34294790e-01 3.62082183e-01
-5.75220585e-01 -3.61913055e-01 7.15936959e-01 1.36238337e-01
1.37591446e-02 7.35486671e-02 5.85175082e-02 -8.44853520e-01
-6.02488935e-01 -1.39823031e+00 1.13564972e-02 -7.99047768e-01
-8.48556399e-01 3.96914363e-01 -1.79190680e-01 -9.69644248e-01
-8.90633881e-01 -2.01226115e-01 -1.45780718e+00 9.26689923e-01
-1.28272581e+00 -6.25138640e-01 -3.51242214e-01 6.36266887e-01
7.96413898e-01 -4.81484115e-01 9.03166354e-01 -2.07313538e-01
-8.12885821e-01 7.41726041e-01 -6.30285561e-01 3.88111919e-01
9.82906282e-01 -1.49652708e+00 1.45625204e-01 3.51592153e-01
-3.39648306e-01 5.60537696e-01 8.08369935e-01 -6.42116189e-01
-1.64316845e+00 -9.49982822e-01 7.80703664e-01 -4.73817110e-01
9.57933486e-01 -3.38776201e-01 -1.07216334e+00 1.17902792e+00
8.58556688e-01 -5.61344564e-01 8.51988614e-01 -4.27769244e-01
-5.43632843e-02 2.82083124e-01 -1.04634714e+00 8.39304149e-01
4.41065103e-01 -5.94986618e-01 -7.67680645e-01 5.17989635e-01
8.37387383e-01 -8.37535322e-01 -9.08696592e-01 -4.47710723e-01
1.70265317e-01 -8.04670572e-01 2.10773751e-01 -7.23397374e-01
7.94030786e-01 2.00609282e-01 5.67546844e-01 -1.67351496e+00
2.75343448e-01 -1.23929703e+00 -1.27486333e-01 1.23342907e+00
2.64399916e-01 -7.60244370e-01 8.53285730e-01 7.54043043e-01
-2.47917175e-01 -6.55043244e-01 -5.70540845e-01 -2.49656916e-01
2.48324960e-01 1.51261255e-01 6.38611376e-01 1.04468954e+00
4.25642818e-01 8.51318240e-01 -1.51187807e-01 -5.07406518e-02
3.52864116e-01 2.26721555e-01 1.28677797e+00 -1.15887177e+00
-2.91661382e-01 -4.28123325e-01 3.18133980e-01 -1.32303369e+00
7.54995346e-01 -7.54196823e-01 4.74869549e-01 -1.00976503e+00
-1.05077036e-01 -3.87439370e-01 3.56618464e-01 8.04339767e-01
-5.06705284e-01 -5.31711102e-01 1.84989437e-01 -1.22767678e-02
-7.00543225e-01 4.79692757e-01 1.83435655e+00 2.15693191e-02
-4.06300634e-01 7.29675055e-01 -5.49552381e-01 6.61751330e-01
1.00114858e+00 -5.94464660e-01 -6.19648159e-01 -6.79484159e-02
-3.72249573e-01 8.55426967e-01 -3.75129618e-02 -5.44746280e-01
6.17178142e-01 -7.25166202e-01 -2.45669693e-01 -1.57567747e-02
6.06668787e-03 -6.88307941e-01 -4.43002403e-01 4.93841648e-01
-8.35097432e-01 1.60627976e-01 -7.83688426e-02 4.20985632e-02
-1.08797289e-01 -5.90335250e-01 7.38951087e-01 -3.09497774e-01
8.84987414e-03 1.77246690e-01 -1.15998816e+00 2.45383754e-01
1.18642974e+00 1.20623320e-01 -6.18145287e-01 -1.06732488e+00
-2.78446257e-01 7.87134886e-01 3.55176419e-01 4.81065780e-01
4.80875224e-01 -9.22116578e-01 -6.99856579e-01 6.73088282e-02
-1.88920721e-01 2.37651587e-01 -2.17150688e-01 4.94863719e-01
-3.15794498e-01 3.60958308e-01 -1.29134595e-01 -5.06309330e-01
-1.70432699e+00 -4.64903899e-02 4.70975608e-01 -4.36836600e-01
-4.98599887e-01 6.25605941e-01 -3.10199298e-02 -9.31418300e-01
4.41918612e-01 -3.12102377e-01 -3.24688673e-01 -5.06839938e-02
5.49168646e-01 2.59806931e-01 -3.10524195e-01 -2.60257900e-01
3.15641642e-01 -3.26957911e-01 -1.84460655e-01 -1.37491629e-01
1.08976281e+00 1.90738052e-01 4.18366641e-02 6.59532666e-01
6.85786724e-01 1.63518593e-01 -1.38036346e+00 -1.96077213e-01
-4.03287783e-02 -5.16212806e-02 -4.48831648e-01 -9.53389645e-01
-7.77839243e-01 6.45081699e-01 -2.36196280e-01 4.75040793e-01
5.42694807e-01 -1.28230542e-01 8.18663538e-01 7.49947309e-01
2.77338922e-01 -7.42808461e-01 4.62801397e-01 1.06148553e+00
6.09387040e-01 -1.12144351e+00 -4.51078206e-01 -1.39317632e-01
-9.57588315e-01 1.13661921e+00 1.41521680e+00 -1.55124953e-02
1.90457448e-01 6.08225048e-01 5.70707262e-01 -1.14885263e-01
-1.40465140e+00 3.39023530e-01 -2.45074838e-01 4.95902419e-01
8.55351090e-01 2.41801411e-01 2.81183422e-01 2.93982387e-01
-5.17864823e-01 -2.71877600e-03 1.18899202e+00 1.22808838e+00
-6.79415166e-01 -1.32956028e+00 -4.00953531e-01 2.97394186e-01
-3.16749930e-01 -6.75520971e-02 -1.00943923e+00 4.60567772e-01
-3.29942405e-01 1.25839007e+00 -1.40847221e-01 -2.56132931e-01
4.10639077e-01 5.46117961e-01 2.60227174e-01 -1.24067092e+00
-1.56052709e+00 -3.59803855e-01 5.47431037e-02 -4.01669174e-01
6.72224909e-02 -5.34186780e-01 -1.39432895e+00 -3.33222717e-01
-9.94270295e-02 6.35458589e-01 2.67641336e-01 8.30716431e-01
7.95315281e-02 9.25911188e-01 9.91592050e-01 -5.87588191e-01
-8.06465030e-01 -1.37370241e+00 -4.21830565e-02 -4.37062904e-02
5.54453254e-01 -4.13182020e-01 -5.25865257e-01 7.00945267e-03] | [12.420877456665039, 8.115405082702637] |
7834e73b-0258-4cea-b2c5-6b23ee1d447c | a-critical-study-on-the-recent-deep-learning | 2111.01604 | null | https://arxiv.org/abs/2111.01604v1 | https://arxiv.org/pdf/2111.01604v1.pdf | A Critical Study on the Recent Deep Learning Based Semi-Supervised Video Anomaly Detection Methods | Video anomaly detection is one of the hot research topics in computer vision nowadays, as abnormal events contain a high amount of information. Anomalies are one of the main detection targets in surveillance systems, usually needing real-time actions. Regarding the availability of labeled data for training (i.e., there is not enough labeled data for abnormalities), semi-supervised anomaly detection approaches have gained interest recently. This paper introduces the researchers of the field to a new perspective and reviews the recent deep-learning based semi-supervised video anomaly detection approaches, based on a common strategy they use for anomaly detection. Our goal is to help researchers develop more effective video anomaly detection methods. As the selection of a right Deep Neural Network plays an important role for several parts of this task, a quick comparative review on DNNs is prepared first. Unlike previous surveys, DNNs are reviewed from a spatiotemporal feature extraction viewpoint, customized for video anomaly detection. This part of the review can help researchers in this field select suitable networks for different parts of their methods. Moreover, some of the state-of-the-art anomaly detection methods, based on their detection strategy, are critically surveyed. The review provides a novel and deep look at existing methods and results in stating the shortcomings of these approaches, which can be a hint for future works. | ['Robert Bergevin', 'Mohammad Baradaran'] | 2021-11-02 | null | null | null | null | ['supervised-anomaly-detection', 'semi-supervised-anomaly-detection'] | ['computer-vision', 'computer-vision'] | [ 7.34365582e-02 -3.09397548e-01 -2.22027346e-01 -2.30416164e-01
9.32344347e-02 -2.91966796e-01 4.09178197e-01 1.73542917e-01
-5.27686715e-01 2.96661437e-01 -2.58844048e-01 -1.07044749e-01
-1.15276035e-02 -6.64184690e-01 -4.59507138e-01 -1.01107740e+00
-2.75810301e-01 2.23904587e-02 4.34295416e-01 -3.42981428e-01
1.67075142e-01 7.82744884e-01 -1.85400283e+00 1.01976514e-01
7.75604069e-01 1.34204054e+00 -2.76450455e-01 5.50939143e-01
-3.17217886e-01 9.25118983e-01 -7.61167943e-01 -3.72710019e-01
3.00074667e-01 -5.42214036e-01 -4.17276919e-01 3.71249050e-01
6.43065929e-01 -6.84975505e-01 -6.99460447e-01 1.38963342e+00
2.30825111e-01 3.07405055e-01 5.91516912e-01 -1.66879690e+00
-3.56582642e-01 1.08725101e-01 -3.35195720e-01 1.06612277e+00
1.19499102e-01 9.31984652e-03 6.88448012e-01 -9.19372857e-01
2.64147967e-01 6.84129119e-01 4.99827653e-01 8.36089671e-01
-4.47808325e-01 -4.30392027e-01 7.40892828e-01 9.63665128e-01
-1.03233123e+00 -3.36373419e-01 1.01976013e+00 -5.29084265e-01
8.42317581e-01 1.54648528e-01 1.04744744e+00 1.46804321e+00
1.93923488e-01 1.13762200e+00 1.94731727e-01 -2.14961126e-01
3.73953730e-01 -3.15949053e-01 2.71262527e-01 6.05920970e-01
5.64341545e-01 2.97583621e-02 -3.24167877e-01 -1.51625231e-01
5.99599898e-01 5.19742787e-01 -1.16068929e-01 -4.80934113e-01
-9.26138341e-01 8.14883471e-01 1.50100395e-01 7.31319427e-01
-4.74791914e-01 -2.89609104e-01 9.51321721e-01 5.30030549e-01
5.18532932e-01 2.76632816e-01 -1.42391443e-01 -3.47680509e-01
-8.57441664e-01 2.32076436e-01 4.52975869e-01 7.55279064e-01
1.39329717e-01 9.10672188e-01 -1.95426852e-01 6.82260334e-01
2.00763047e-01 3.86450440e-01 5.86186409e-01 -6.34410977e-01
1.01495564e-01 6.51115298e-01 -1.00777321e-01 -1.25553298e+00
-4.24289256e-01 -7.25584701e-02 -1.18456686e+00 2.35231593e-01
5.28932631e-01 -1.41617566e-01 -9.44756448e-01 1.26219559e+00
1.21413849e-01 4.11986202e-01 -5.22805192e-02 9.41899717e-01
9.26419854e-01 6.76163435e-01 -3.47743891e-02 -4.26005095e-01
1.20069313e+00 -9.71060455e-01 -1.21950448e+00 -2.40055650e-01
5.22720754e-01 -4.37428206e-01 3.51711273e-01 6.84865236e-01
-6.49917364e-01 -4.36079651e-01 -1.00690770e+00 4.20837462e-01
-6.00464821e-01 -8.47120062e-02 5.10775030e-01 5.30721247e-01
-9.97907937e-01 3.51076931e-01 -1.21547651e+00 -8.17271948e-01
6.59983039e-01 7.71240741e-02 -3.14716548e-01 -9.37213004e-02
-1.18653381e+00 7.97154427e-01 4.06761527e-01 5.37934959e-01
-9.62108552e-01 -6.94799200e-02 -1.05671120e+00 -3.23186010e-01
6.78862810e-01 -2.23638132e-01 1.23556709e+00 -1.32993996e+00
-1.04746449e+00 8.30037296e-01 -1.86930224e-01 -6.94579720e-01
4.93959069e-01 -4.58006293e-01 -9.63068128e-01 4.75991279e-01
-1.00565873e-01 1.03221968e-01 1.20021784e+00 -7.47955859e-01
-9.73761141e-01 -3.68547052e-01 6.00921968e-03 7.28758797e-02
-5.11226714e-01 2.29253292e-01 -2.97711521e-01 -1.09343278e+00
7.76311010e-02 -4.92003769e-01 -3.88999939e-01 2.53943294e-01
-3.17814708e-01 -4.20146704e-01 1.38173187e+00 -5.89016557e-01
1.51877606e+00 -2.15027714e+00 6.90858737e-02 1.12689875e-01
5.43738723e-01 9.18304682e-01 5.42844720e-02 2.63793439e-01
-3.71754378e-01 -2.48037890e-01 -5.37883401e-01 -1.34527400e-01
-4.07932758e-01 3.01610023e-01 -2.82739937e-01 7.30732679e-01
3.37991029e-01 6.27308309e-01 -1.06165957e+00 -2.26659402e-01
6.44639194e-01 1.92816690e-01 -2.08852455e-01 4.30638105e-01
-3.98777537e-02 6.02907062e-01 -5.98273396e-01 1.07418931e+00
4.89771247e-01 2.76112139e-01 -3.49821508e-01 -3.17956805e-02
-1.81535020e-01 -3.66275668e-01 -9.60867345e-01 1.25626171e+00
3.33828539e-01 1.10325181e+00 -1.91602402e-03 -1.74398661e+00
7.64266491e-01 6.04370117e-01 9.61249352e-01 -4.35734302e-01
2.25263745e-01 2.39283800e-01 1.29937232e-01 -8.94753039e-01
2.72968799e-01 3.24908584e-01 2.83134282e-01 5.61678894e-02
6.61421418e-02 5.30780554e-01 5.75514257e-01 5.72630614e-02
1.21028709e+00 -2.10417300e-01 5.36581218e-01 1.08111009e-01
9.61373508e-01 1.07740864e-01 6.82718217e-01 8.84004593e-01
-9.39727902e-01 5.59195220e-01 5.60524702e-01 -1.20275605e+00
-9.11053419e-01 -8.33461702e-01 -1.21470727e-01 8.60680282e-01
1.28717655e-02 -1.84985891e-01 -7.44085491e-01 -1.01941037e+00
-3.48458439e-01 3.37799430e-01 -6.93605483e-01 -2.86630332e-01
-7.62947202e-01 -9.34612632e-01 6.57424986e-01 7.86471069e-01
8.12876761e-01 -1.45885754e+00 -7.23004043e-01 -6.48734868e-02
-2.99030006e-01 -1.28489161e+00 1.00738317e-01 -2.05672577e-01
-1.06846881e+00 -1.50525057e+00 -8.72263134e-01 -7.42247880e-01
9.08636272e-01 4.91264820e-01 9.50304687e-01 3.94315541e-01
-3.03670198e-01 7.63664067e-01 -7.41372526e-01 -6.41380310e-01
-2.11980149e-01 -2.90339202e-01 5.08272946e-01 3.22118342e-01
1.10122287e+00 -3.74990046e-01 -5.55039406e-01 3.61792117e-01
-1.18627357e+00 -7.12658882e-01 3.24908584e-01 7.47102857e-01
4.25892353e-01 8.00819248e-02 3.08671713e-01 -4.42496955e-01
3.65192324e-01 -5.48854351e-01 -8.20983708e-01 -1.23457663e-01
-1.98017001e-01 -5.82879066e-01 7.70676136e-01 -2.67042458e-01
-7.21365869e-01 -1.75485373e-01 -3.47583413e-01 -7.17485726e-01
-7.38881171e-01 1.74368769e-01 -1.18617676e-01 -1.56487495e-01
6.13077343e-01 4.11405206e-01 1.40942454e-01 -3.55535984e-01
-1.53374732e-01 4.37544405e-01 3.18552196e-01 1.29292496e-02
7.60865986e-01 7.43894994e-01 -9.61739570e-02 -1.37543905e+00
-8.39315712e-01 -6.07122838e-01 -8.10396910e-01 -5.05506456e-01
9.56036985e-01 -5.79158008e-01 -2.43897870e-01 1.26506984e+00
-1.12231314e+00 -1.49517898e-02 -4.74031091e-01 5.48986495e-01
-3.80079389e-01 7.70004153e-01 -5.89889407e-01 -9.22328115e-01
-1.45211995e-01 -1.23119593e+00 5.54575264e-01 1.64634377e-01
-2.15579532e-02 -1.14966440e+00 9.04977694e-02 -9.33150649e-02
5.80104589e-01 3.29530954e-01 4.82579857e-01 -1.25285304e+00
-4.15475279e-01 -5.98076105e-01 5.14036827e-02 7.52800584e-01
3.28969270e-01 2.79764354e-01 -9.32047069e-01 -2.29808033e-01
1.57697693e-01 2.26762086e-01 9.42734361e-01 7.92910278e-01
1.59928882e+00 -2.36260653e-01 -3.82691741e-01 6.72546268e-01
9.85909462e-01 6.60472929e-01 6.75691545e-01 5.78653812e-01
9.31645036e-01 4.78962064e-01 6.69530809e-01 4.69749868e-01
2.73007806e-02 5.12327731e-01 8.62073243e-01 6.91007078e-02
3.37737054e-01 3.95909607e-01 5.85497916e-01 6.41044140e-01
-4.74683970e-01 -4.75413501e-01 -8.58039618e-01 5.85718989e-01
-1.97628272e+00 -1.33660603e+00 -1.92591980e-01 2.09533691e+00
-5.99701703e-02 6.39832541e-02 4.56234843e-01 5.14923334e-01
9.27157700e-01 5.26582718e-01 -7.04792857e-01 -2.13521346e-01
-2.31604531e-01 -2.74943382e-01 1.02876894e-01 -7.53398091e-02
-1.85260606e+00 7.57653415e-01 6.50821209e+00 5.27307570e-01
-1.19110298e+00 -3.34214479e-01 4.11752373e-01 -4.00279164e-02
5.74692369e-01 -5.91941237e-01 -5.37409604e-01 6.70671165e-01
7.43456841e-01 2.42462933e-01 3.17725702e-03 1.12094212e+00
2.46754035e-01 -1.45234346e-01 -1.03950238e+00 1.25739813e+00
4.27170277e-01 -1.01717043e+00 3.74595761e-01 -2.18813419e-01
4.26021218e-01 1.97932450e-03 -5.31755574e-02 1.54537290e-01
-4.12381262e-01 -7.14218974e-01 2.29092717e-01 4.91752625e-01
2.59536564e-01 -8.34801197e-01 1.18376839e+00 4.76596840e-02
-1.21983266e+00 -4.01001751e-01 -4.50557739e-01 -1.77216381e-01
3.38221282e-01 6.53104305e-01 1.72020812e-02 4.50274915e-01
1.09676087e+00 1.26157403e+00 -4.40144688e-01 1.44771993e+00
-1.78362504e-01 7.45491982e-01 -1.77972913e-01 1.41810596e-01
4.15879250e-01 -4.21429098e-01 1.19179356e+00 1.09283841e+00
3.42895806e-01 1.03970975e-01 3.05505186e-01 4.26899016e-01
3.91233981e-01 2.38485401e-04 -1.09180200e+00 -2.58604884e-01
1.04458705e-01 9.54261601e-01 -7.64935255e-01 -4.06080753e-01
-8.47079098e-01 9.70341206e-01 -2.10635096e-01 4.74552065e-01
-7.30296969e-01 -5.94600022e-01 9.37533379e-01 -1.96533930e-02
8.10092688e-02 -2.16073781e-01 1.82509441e-02 -1.31913018e+00
2.23582357e-01 -1.00005662e+00 8.29625487e-01 -2.71254808e-01
-1.40792668e+00 6.87651753e-01 3.16873252e-01 -1.76045251e+00
-4.48342413e-01 -9.96466458e-01 -8.60511005e-01 2.45949492e-01
-1.36069071e+00 -7.19467580e-01 -6.46827579e-01 7.08820820e-01
9.49158728e-01 -8.07193875e-01 6.79841101e-01 4.69124287e-01
-1.12551808e+00 4.60268766e-01 -7.20899627e-02 8.26204658e-01
4.44851726e-01 -9.83415186e-01 5.38349509e-01 1.59885776e+00
1.39221132e-01 3.08826953e-01 5.80876529e-01 -5.24639010e-01
-1.01285994e+00 -1.04921818e+00 4.53612596e-01 -4.39310014e-01
6.32709384e-01 -1.56185124e-03 -1.18385768e+00 8.50063443e-01
1.20451406e-01 3.78929794e-01 6.75773144e-01 -3.55265051e-01
9.35531408e-02 -7.11808279e-02 -1.10815585e+00 6.94626749e-01
1.10332763e+00 -1.11318029e-01 -6.73742235e-01 1.52260169e-01
1.85805455e-01 -4.68753517e-01 -1.04257964e-01 5.16626298e-01
3.02319676e-01 -1.35021615e+00 7.16144681e-01 -9.20614660e-01
1.90125838e-01 -3.67711455e-01 9.14894789e-02 -1.32852495e+00
-1.57656357e-01 -3.61949742e-01 -8.17858517e-01 5.93845665e-01
-1.45923272e-01 -7.17732728e-01 7.68779993e-01 3.67284209e-01
-3.97661477e-01 -7.91855335e-01 -1.00180244e+00 -7.97499239e-01
-4.35452610e-01 -7.75828063e-01 2.04447493e-01 9.67186034e-01
-2.18728602e-01 -3.62117052e-01 -4.24446166e-01 2.88754672e-01
6.66423500e-01 -5.35314798e-01 5.01441836e-01 -1.46138251e+00
5.85871041e-01 -7.09888816e-01 -1.25735176e+00 -9.30442929e-01
8.85939300e-02 -2.85095215e-01 -2.07248926e-01 -1.42746115e+00
-2.47423455e-01 2.74919093e-01 -4.60655808e-01 3.80923748e-01
-7.43063316e-02 2.96772748e-01 -2.31492475e-01 2.13821325e-02
-6.58082187e-01 5.79133809e-01 7.86900520e-01 -1.80284008e-01
-1.14156991e-01 5.13180912e-01 -3.14253867e-01 1.32427490e+00
8.63959134e-01 -1.20667510e-01 -3.81205648e-01 -4.42731440e-01
-1.75922457e-02 -4.88681972e-01 4.32753503e-01 -1.32600045e+00
2.30550677e-01 -5.07538468e-02 3.78143996e-01 -7.41358280e-01
2.09122971e-01 -1.22252834e+00 -6.81748390e-01 4.47072119e-01
4.60146107e-02 5.98300457e-01 3.43709260e-01 7.29220510e-01
-6.91453516e-01 -2.83822030e-01 8.05295646e-01 -2.46966571e-01
-1.50273216e+00 7.54672885e-01 -1.12508881e+00 2.58231740e-02
1.41065991e+00 -5.99932015e-01 1.45139366e-01 -5.62777996e-01
-7.77192414e-01 2.87455916e-01 5.64808585e-02 7.35824227e-01
9.38571930e-01 -1.33170128e+00 -4.98883456e-01 4.72094089e-01
3.80959123e-01 1.65319666e-01 5.40437698e-01 9.38465595e-01
-6.67889118e-01 4.97968554e-01 -5.60200989e-01 -8.77829671e-01
-1.17913556e+00 7.21656442e-01 6.07751071e-01 1.47410646e-01
-8.05707097e-01 6.71038628e-01 7.69032687e-02 2.47056440e-01
7.23275423e-01 -3.16127092e-01 -7.92716444e-01 1.81164563e-01
1.01787341e+00 6.71917737e-01 1.96744770e-01 -8.01531732e-01
-4.50473249e-01 4.47942108e-01 -2.33831361e-01 3.21453810e-01
1.11866879e+00 -8.21358431e-03 -1.99642211e-01 5.41503549e-01
6.94534898e-01 -4.22665447e-01 -1.06385744e+00 -1.97902039e-01
-2.45653000e-02 -5.30175388e-01 -1.01094723e-01 -1.08299725e-01
-1.51424956e+00 9.27396715e-01 8.97651076e-01 4.65428442e-01
1.37331116e+00 -3.04555535e-01 6.57138348e-01 6.56683385e-01
2.14919180e-01 -1.12853813e+00 3.11773300e-01 7.09326923e-01
7.98491597e-01 -1.63711846e+00 -1.96259692e-01 -1.39101073e-01
-5.83114862e-01 1.32016373e+00 1.14109874e+00 -2.75893629e-01
8.72053802e-01 1.14777153e-02 3.46503109e-01 -3.97766411e-01
-5.68188094e-02 -1.92920893e-01 4.61477607e-01 9.24515009e-01
2.76661277e-01 -4.43481803e-01 -3.33585329e-02 2.91344047e-01
4.14890110e-01 -3.10518116e-01 4.13247019e-01 1.11872399e+00
-4.73232269e-01 -7.92403579e-01 -4.38966542e-01 7.28217661e-01
-6.29837573e-01 2.17040166e-01 -2.27325201e-01 7.16580153e-01
9.24740806e-02 9.33623552e-01 4.70452785e-01 -2.36029223e-01
5.25084913e-01 8.29503462e-02 8.80331099e-02 -3.31031382e-01
-5.01536280e-02 -4.09255743e-01 -3.52903783e-01 -9.04149115e-01
-8.15610111e-01 -4.87946242e-01 -7.23644435e-01 -2.27601260e-01
1.53891242e-03 9.59889404e-03 2.27843463e-01 1.20107377e+00
7.10186884e-02 5.41902840e-01 3.57724279e-01 -9.56377625e-01
-5.04596718e-02 -8.70188653e-01 -5.79084754e-01 3.40687573e-01
8.28684092e-01 -9.39581215e-01 -4.50111449e-01 -5.38394824e-02] | [7.874264717102051, 1.5453393459320068] |
c4890c9f-e86f-4680-b48f-580bfa91610e | text-guided-eyeglasses-manipulation-with | 2304.12539 | null | https://arxiv.org/abs/2304.12539v1 | https://arxiv.org/pdf/2304.12539v1.pdf | Text-guided Eyeglasses Manipulation with Spatial Constraints | Virtual try-on of eyeglasses involves placing eyeglasses of different shapes and styles onto a face image without physically trying them on. While existing methods have shown impressive results, the variety of eyeglasses styles is limited and the interactions are not always intuitive or efficient. To address these limitations, we propose a Text-guided Eyeglasses Manipulation method that allows for control of the eyeglasses shape and style based on a binary mask and text, respectively. Specifically, we introduce a mask encoder to extract mask conditions and a modulation module that enables simultaneous injection of text and mask conditions. This design allows for fine-grained control of the eyeglasses' appearance based on both textual descriptions and spatial constraints. Our approach includes a disentangled mapper and a decoupling strategy that preserves irrelevant areas, resulting in better local editing. We employ a two-stage training scheme to handle the different convergence speeds of the various modality conditions, successfully controlling both the shape and style of eyeglasses. Extensive comparison experiments and ablation analyses demonstrate the effectiveness of our approach in achieving diverse eyeglasses styles while preserving irrelevant areas. | ['Wei Xu', 'Jingen Liu', 'Ping Liu', 'Jiacheng Wang'] | 2023-04-25 | null | null | null | null | ['virtual-try-on'] | ['computer-vision'] | [ 3.46497655e-01 3.24305706e-02 -9.55492333e-02 -1.40747398e-01
-1.74928859e-01 -7.95862675e-01 6.33638561e-01 -5.23240209e-01
-1.44308269e-01 5.81475437e-01 8.86308253e-02 -1.20320618e-01
1.28474668e-01 -3.83466721e-01 -7.20329225e-01 -7.82509565e-01
4.60575819e-01 -1.89576134e-01 1.75807461e-01 -2.81637043e-01
4.95355904e-01 7.89994895e-01 -1.92633569e+00 3.99106927e-02
8.72163057e-01 7.20397890e-01 4.71332908e-01 6.65543079e-01
8.98868665e-02 1.91534922e-01 -7.66677797e-01 -1.19502142e-01
5.49759567e-01 -4.78326201e-01 -6.95349053e-02 5.48029780e-01
9.60025251e-01 -3.70199054e-01 -3.17513824e-01 9.23420429e-01
5.06988347e-01 -9.91836004e-03 5.88673949e-01 -1.04495800e+00
-9.46182430e-01 -4.81303409e-02 -9.18688178e-01 -1.49510607e-01
5.25337279e-01 4.40570265e-01 8.28144133e-01 -7.36965716e-01
5.53690076e-01 1.11512291e+00 1.82512939e-01 8.04160833e-01
-1.55156755e+00 -9.19237614e-01 2.92418748e-01 -2.33610585e-01
-1.59320104e+00 -8.18443894e-01 6.28220916e-01 -5.17526567e-01
5.98025739e-01 6.20146513e-01 7.59875298e-01 1.06431735e+00
2.78271973e-01 4.89322573e-01 1.23749733e+00 -9.07774508e-01
-1.10714119e-02 6.53090000e-01 -2.30340034e-01 1.03728962e+00
1.57493517e-01 3.48802954e-01 -8.05508733e-01 7.84123540e-02
1.33884633e+00 -1.57931969e-02 -6.76979601e-01 -6.28788650e-01
-1.21296442e+00 3.07707101e-01 2.37112790e-02 -1.42092386e-03
4.56069596e-02 -7.08440915e-02 -1.74767256e-01 2.72528768e-01
3.99997354e-01 7.98871815e-01 -1.95835471e-01 3.17279041e-01
-1.01282859e+00 2.69117922e-01 5.12256563e-01 1.27299964e+00
6.63116813e-01 -2.51982752e-02 -5.31099200e-01 8.84387314e-01
3.06145400e-01 5.46849608e-01 1.67375222e-01 -5.93295097e-01
3.45260024e-01 7.12529778e-01 2.90555865e-01 -8.91339362e-01
-1.84079617e-01 5.17778955e-02 -4.08040673e-01 8.69618654e-01
4.81828481e-01 -1.69114068e-01 -1.36086142e+00 1.98307371e+00
4.43630815e-01 9.43693742e-02 -3.94712061e-01 1.05787778e+00
5.16131461e-01 1.65059000e-01 -1.23560198e-01 -5.48892567e-05
1.60605562e+00 -7.69044280e-01 -1.15977919e+00 -3.73223901e-01
2.48481616e-01 -8.57377231e-01 1.57642996e+00 2.22411573e-01
-1.26469457e+00 -4.05656666e-01 -1.52222598e+00 -3.27422142e-01
-3.06293666e-01 7.40285218e-01 3.08595657e-01 8.79100740e-01
-1.06335211e+00 4.13875908e-01 -7.13155270e-01 -2.54701018e-01
3.29452276e-01 6.48105860e-01 -4.69624102e-01 4.59647566e-01
-7.23174632e-01 1.00422823e+00 -1.45343486e-02 1.91696465e-01
-4.10244077e-01 -6.49111032e-01 -9.87962008e-01 1.62247211e-01
5.00531971e-01 -7.77388871e-01 7.09016681e-01 -1.02546406e+00
-2.19442987e+00 1.09964943e+00 -1.51440457e-01 1.96311355e-01
5.19107044e-01 -9.87614468e-02 -2.99593061e-01 4.59273010e-02
-2.43045151e-01 8.08323443e-01 1.58550668e+00 -1.20383084e+00
-3.16936642e-01 -5.20412505e-01 2.23264992e-02 5.59480309e-01
-5.99549532e-01 9.40048248e-02 -7.62515485e-01 -8.72608900e-01
1.06672607e-01 -1.03169262e+00 1.72189608e-01 4.29697275e-01
-6.97181523e-01 6.41104758e-01 8.09960127e-01 -5.51121354e-01
1.44345701e+00 -2.41474605e+00 5.75604260e-01 1.98569104e-01
4.40750688e-01 1.37483254e-01 -6.43577352e-02 2.87472695e-01
-6.69395179e-02 1.06603019e-01 -4.44170982e-02 -7.20055461e-01
-6.11652210e-02 -1.15764610e-01 -3.85193586e-01 4.65514064e-01
2.63984889e-01 6.66901290e-01 -4.37909633e-01 -2.90087342e-01
2.93935716e-01 4.40398753e-01 -6.53485239e-01 1.78363383e-01
-1.85946792e-01 5.63910365e-01 -3.14128458e-01 6.18348360e-01
7.95774341e-01 -3.38591710e-02 3.16284835e-01 -4.12678391e-01
-2.08555698e-01 -1.41163089e-03 -1.15274823e+00 1.75784695e+00
-4.47355896e-01 6.78381085e-01 4.17505473e-01 -3.58127877e-02
7.81404376e-01 1.68780193e-01 -1.00098506e-01 -6.18770421e-01
2.60622263e-01 4.04429762e-03 -5.76038659e-02 -4.13377404e-01
6.51588976e-01 -1.02929495e-01 1.10913903e-01 5.47141314e-01
-1.48351401e-01 -5.55983305e-01 5.74247241e-02 2.94994563e-02
5.68476617e-01 5.59910893e-01 1.99472636e-01 -3.52360696e-01
2.44767755e-01 -6.08383298e-01 1.58133924e-01 4.55702960e-01
3.96106243e-02 9.66257513e-01 5.41036785e-01 -1.71917379e-01
-9.25594270e-01 -7.52508461e-01 -1.10801170e-02 9.34649527e-01
4.36549425e-01 -3.44984233e-01 -1.00338852e+00 -4.40654427e-01
-4.70553413e-02 6.22854292e-01 -8.72922003e-01 -1.78292379e-01
-4.81688082e-01 -3.54308009e-01 2.72205979e-01 2.58601606e-01
4.45165753e-01 -8.68771136e-01 -8.04888964e-01 -5.36298513e-01
3.57469559e-01 -1.02413261e+00 -1.13178504e+00 -2.77000397e-01
-5.13375819e-01 -9.53074515e-01 -5.74704885e-01 -6.07804239e-01
1.11109900e+00 3.42142254e-01 6.97787404e-01 8.78168046e-02
-5.12282193e-01 1.94356754e-01 -1.11593254e-01 -3.66457731e-01
-2.26015896e-01 1.10625766e-01 5.65399416e-02 3.98512244e-01
-1.45784304e-01 -7.76148677e-01 -7.20851481e-01 5.01498997e-01
-1.00486338e+00 5.52625000e-01 4.97445226e-01 7.77114570e-01
2.21813172e-01 -4.66862500e-01 -2.45509250e-03 -8.15529168e-01
7.13289261e-01 8.67350623e-02 -8.72815192e-01 3.51051569e-01
-4.02720809e-01 1.19233884e-01 3.81154746e-01 -6.42522395e-01
-1.02860916e+00 3.66903208e-02 3.72557431e-01 -6.60133600e-01
-8.12209398e-02 -2.81332642e-01 -4.91251230e-01 -3.45598489e-01
7.04040527e-01 -4.32228334e-02 2.67479151e-01 -3.01645339e-01
5.42521060e-01 6.33177638e-01 1.81857213e-01 -4.55918580e-01
8.61382246e-01 6.29729152e-01 -2.26508677e-01 -8.52128267e-01
-4.38115239e-01 2.05459669e-01 -6.76550150e-01 -1.20115422e-01
9.32029784e-01 -6.01515830e-01 -7.07525015e-01 5.09691298e-01
-9.48421419e-01 -3.38673919e-01 -1.33104861e-01 2.31048167e-01
-4.12073463e-01 1.05204552e-01 -2.72470564e-01 -6.57240152e-01
-3.72936502e-02 -1.54581785e+00 1.42534828e+00 2.82218814e-01
-3.43149811e-01 -6.63366079e-01 -2.19055861e-01 1.88498385e-02
3.37057561e-01 5.84859215e-02 8.26950908e-01 -4.46788371e-02
-1.02115619e+00 -2.03080207e-01 -1.94870576e-01 9.31923240e-02
5.00907779e-01 1.55976132e-01 -1.04548633e+00 -3.55381310e-01
-1.65731966e-01 9.99958254e-03 5.26638031e-01 3.77543092e-01
1.01797771e+00 -2.50806272e-01 -5.07570922e-01 9.45320189e-01
1.01505613e+00 2.21228033e-01 7.37892747e-01 2.19826937e-01
8.61799479e-01 4.89256322e-01 2.35878557e-01 2.48424783e-01
-1.42933667e-01 1.20154285e+00 9.46300924e-02 -3.73560101e-01
-3.26028466e-01 -1.79753527e-01 1.71198800e-01 1.60324395e-01
8.07450619e-03 -2.86763281e-01 -5.05320430e-01 1.56403214e-01
-1.45863986e+00 -6.76207423e-01 4.82816428e-01 2.55545306e+00
8.98696423e-01 -4.13002633e-03 -1.36551410e-01 -2.44586647e-01
7.40965903e-01 2.92246968e-01 -5.57792425e-01 -2.65104681e-01
-7.33700320e-02 9.47141200e-02 4.56322372e-01 6.18495524e-01
-8.92617166e-01 1.03119564e+00 6.84647226e+00 7.11974025e-01
-1.35886741e+00 -2.90939420e-01 3.00059140e-01 -6.25134587e-01
-5.14414549e-01 -9.26600844e-02 -8.62784445e-01 4.45055366e-01
8.21558982e-02 2.67569214e-01 6.97720528e-01 1.76827610e-01
2.54693985e-01 -2.10035115e-01 -1.16656888e+00 9.12243187e-01
2.89138317e-01 -1.27322459e+00 2.23524496e-02 1.71267450e-01
3.92678976e-01 -6.01625502e-01 4.16643143e-01 -2.15115860e-01
-1.02521993e-01 -9.19364989e-01 1.09813499e+00 6.71070874e-01
1.41087258e+00 -3.64960045e-01 -2.31496379e-01 3.34229991e-02
-7.05577254e-01 1.36430487e-02 2.17071518e-01 -3.43518443e-02
-7.91691765e-02 -7.17113391e-02 -5.68174124e-01 2.31506690e-01
4.19994324e-01 2.99776882e-01 -4.33486402e-01 6.31313086e-01
-4.55313653e-01 8.54833499e-02 -2.71881908e-01 -5.80503196e-02
-4.08566833e-01 -4.60674584e-01 8.36176276e-01 8.36138487e-01
1.22563705e-01 7.44844303e-02 -5.49600869e-02 1.35460913e+00
5.17714135e-02 -5.68458736e-02 -7.32737005e-01 -1.12479560e-01
6.99965000e-01 1.21498239e+00 -5.99860966e-01 4.98826755e-03
-3.25694829e-01 1.23043609e+00 3.29509616e-01 8.32732201e-01
-7.64230907e-01 -6.59143746e-01 9.84502912e-01 4.26143736e-01
3.12828362e-01 -4.21645373e-01 -6.16912961e-01 -1.27446663e+00
2.80664682e-01 -9.73393202e-01 -2.09285110e-01 -9.03178513e-01
-6.83348119e-01 7.81529665e-01 1.07604176e-01 -1.14293933e+00
1.21991768e-01 -6.23736143e-01 -5.38851678e-01 1.18046987e+00
-1.22235191e+00 -1.33721650e+00 -2.64744341e-01 4.11231756e-01
3.64443094e-01 -2.04962909e-01 8.00353467e-01 8.27840194e-02
-9.20423746e-01 9.12564635e-01 -2.21962452e-01 -2.54447192e-01
8.81542146e-01 -9.89226997e-01 3.47258478e-01 8.94742012e-01
6.54108450e-02 1.06643891e+00 7.93597400e-01 -5.15000582e-01
-1.58156598e+00 -4.67922568e-01 3.91073912e-01 -6.61851585e-01
3.38056296e-01 -9.94509935e-01 -8.16662908e-01 8.48802805e-01
4.80662018e-01 -2.60515869e-01 5.71042001e-01 -1.26975030e-02
-3.05072010e-01 -1.54420912e-01 -1.04917192e+00 1.33414173e+00
9.97990072e-01 -5.66568494e-01 -2.70991892e-01 3.36241163e-02
4.99679089e-01 -7.62789965e-01 -1.93757877e-01 2.88201153e-01
8.26597214e-01 -9.59199786e-01 7.83527672e-01 -5.14909208e-01
1.90322623e-01 -5.26059866e-01 1.44998774e-01 -1.37608707e+00
-2.26842791e-01 -1.14332891e+00 4.86196950e-03 1.16003704e+00
4.25552934e-01 -7.35194504e-01 6.10633254e-01 8.95222425e-01
-1.66294888e-01 -6.64601505e-01 -6.28110468e-01 -3.20630312e-01
-4.80587691e-01 5.90521097e-02 6.10581100e-01 7.95239389e-01
2.62566149e-01 2.47915983e-01 -5.40465474e-01 3.19484740e-01
1.82593420e-01 1.85177833e-01 9.91459131e-01 -7.16428339e-01
-3.59909326e-01 -6.55718744e-01 -2.16511741e-01 -1.00805116e+00
1.10474668e-01 -4.73812878e-01 -7.91491345e-02 -9.59768653e-01
6.62611052e-03 -3.37567210e-01 1.17635213e-01 6.69882417e-01
-4.18463349e-01 2.60161579e-01 2.42987975e-01 9.48241279e-02
1.12115420e-01 6.37415528e-01 1.59189534e+00 8.44299793e-02
-6.72647715e-01 -1.87701687e-01 -7.62482345e-01 4.99033034e-01
5.53439081e-01 -6.79452792e-02 -7.31398761e-01 -7.07888305e-01
1.26748130e-01 -1.89465340e-02 3.61190528e-01 -6.45178914e-01
1.04365796e-01 -2.77922630e-01 4.02512819e-01 2.67669857e-02
6.61092401e-01 -6.49311244e-01 8.15881938e-02 -5.20260595e-02
-3.54484737e-01 -6.60460591e-02 7.38772452e-01 5.48948646e-01
1.27776757e-01 8.90348107e-02 7.86736727e-01 1.08540341e-01
-1.50285169e-01 2.88893163e-01 -3.20458293e-01 -4.58254248e-01
1.07298052e+00 -6.08294010e-01 -9.31638107e-02 -1.87574491e-01
-7.11126029e-01 -1.34581989e-02 9.69337821e-01 4.65920627e-01
6.21413827e-01 -1.16462040e+00 -4.35451120e-01 1.10474920e+00
2.43042111e-02 -3.74648571e-01 2.16780603e-01 8.50628316e-01
-4.44275111e-01 3.43392789e-02 -2.66178578e-01 -4.23472285e-01
-1.48811901e+00 5.34838200e-01 6.58871233e-01 2.54264742e-01
-6.88839495e-01 8.24867189e-01 6.56157911e-01 -1.65997490e-01
2.60119796e-01 -2.52655298e-01 -1.90263599e-01 -7.03319013e-02
5.96347988e-01 7.44611444e-03 1.71397459e-02 -3.06065053e-01
5.28160436e-03 7.77449250e-01 -2.58169025e-01 -3.90479863e-01
9.23445106e-01 -4.54055250e-01 2.28458270e-02 2.05366179e-01
7.13125706e-01 6.53228223e-01 -1.51813018e+00 2.28486769e-02
-4.83278066e-01 -1.09646773e+00 -1.85027011e-02 -8.38251531e-01
-8.98855686e-01 7.69751072e-01 5.34173012e-01 1.08084343e-01
1.34979129e+00 -3.00546795e-01 4.50961977e-01 -1.76592335e-01
1.25154585e-01 -9.18473303e-01 -8.40282664e-02 -4.05822881e-03
1.14106691e+00 -9.22019064e-01 -6.02240022e-03 -7.22540677e-01
-6.88103676e-01 9.01416719e-01 7.58665502e-01 8.69168043e-02
4.24399257e-01 3.85766059e-01 1.56228364e-01 -4.45682645e-01
-4.62542057e-01 -5.76355644e-02 7.38429248e-01 3.49688888e-01
2.91705012e-01 -1.41799405e-01 -7.99996331e-02 1.98510274e-01
-1.06246695e-01 -2.18611240e-01 2.99993366e-01 8.05460691e-01
-1.59426540e-01 -1.19143689e+00 -4.45906609e-01 2.36247391e-01
-2.53326386e-01 -2.16495737e-01 -4.63981181e-01 9.11547422e-01
9.90744308e-02 6.11953557e-01 4.56000008e-02 -3.64620805e-01
5.01612961e-01 9.22160447e-02 9.70745146e-01 -8.29226911e-01
-3.32928032e-01 2.74864912e-01 -1.10335253e-01 -6.30055487e-01
-9.55264121e-02 -4.44361180e-01 -7.75257587e-01 -1.89771298e-02
-6.30680740e-01 -3.96382987e-01 5.55704713e-01 8.40857089e-01
6.83455825e-01 5.12587488e-01 5.32218099e-01 -1.14888644e+00
-2.37782866e-01 -8.40712368e-01 -8.19913924e-01 2.84187585e-01
5.17914772e-01 -9.55758333e-01 -2.64445662e-01 1.04534604e-01] | [12.687529563903809, -0.22697879374027252] |
c21e48ea-e818-41c5-b235-45d5c80d1ccb | textworld-a-learning-environment-for-text | 1806.11532 | null | https://arxiv.org/abs/1806.11532v2 | https://arxiv.org/pdf/1806.11532v2.pdf | TextWorld: A Learning Environment for Text-based Games | We introduce TextWorld, a sandbox learning environment for the training and evaluation of RL agents on text-based games. TextWorld is a Python library that handles interactive play-through of text games, as well as backend functions like state tracking and reward assignment. It comes with a curated list of games whose features and challenges we have analyzed. More significantly, it enables users to handcraft or automatically generate new games. Its generative mechanisms give precise control over the difficulty, scope, and language of constructed games, and can be used to relax challenges inherent to commercial text games like partial observability and sparse rewards. By generating sets of varied but similar games, TextWorld can also be used to study generalization and transfer learning. We cast text-based games in the Reinforcement Learning formalism, use our framework to develop a set of benchmark games, and evaluate several baseline agents on this set and the curated list. | ['James Moore', 'Ákos Kádár', 'Marc-Alexandre Côté', 'Tavian Barnes', 'Ben Kybartas', 'Xingdi Yuan', 'Mahmoud Adada', 'Layla El Asri', 'Emery Fine', 'Adam Trischler', 'Wendy Tay', 'Ruo Yu Tao', 'Matthew Hausknecht'] | 2018-06-29 | null | null | null | null | ['text-based-games'] | ['playing-games'] | [-3.38276297e-01 -4.79313545e-02 -7.07380660e-03 -4.84667085e-02
-6.06659353e-01 -9.06320333e-01 8.32733512e-01 -1.76305279e-01
-7.82394648e-01 1.02230442e+00 -4.12614122e-02 -5.11078775e-01
-7.63611048e-02 -8.95713210e-01 -4.66168642e-01 -4.81177866e-01
-4.58436519e-01 8.20848644e-01 6.42666161e-01 -1.03685832e+00
-1.07281998e-01 1.44430727e-01 -1.58438063e+00 9.81562063e-02
5.28408766e-01 6.23997390e-01 4.27550316e-01 8.73293161e-01
3.37501198e-01 1.33975792e+00 -8.11804950e-01 -3.69304866e-01
4.02966082e-01 -4.37566012e-01 -7.37460911e-01 -2.52372503e-01
-2.94697434e-01 -5.40714324e-01 -4.57542419e-01 7.34443307e-01
7.30065227e-01 4.31249619e-01 1.85931742e-01 -1.50326109e+00
-8.87212753e-02 7.86906302e-01 4.68015410e-02 3.19973171e-01
4.16182548e-01 6.80509150e-01 1.07047093e+00 -2.25568235e-01
7.63465703e-01 1.22370028e+00 4.31201428e-01 8.88454854e-01
-1.23726237e+00 -7.89777815e-01 1.01273723e-01 -1.43037334e-01
-7.52075255e-01 -2.98479021e-01 1.39005199e-01 -3.31275702e-01
1.24552071e+00 1.61827877e-01 8.89155030e-01 1.40565503e+00
1.16931023e-02 8.69975865e-01 1.00863659e+00 -1.20354025e-02
5.77802837e-01 -1.92060217e-01 -2.39193797e-01 8.63081753e-01
-3.29392761e-01 7.36067414e-01 -3.78957421e-01 -2.98741639e-01
8.87548864e-01 -4.90390956e-01 1.25534624e-01 -4.56427187e-01
-1.07280684e+00 9.89116013e-01 2.15381354e-01 -2.03286603e-01
-4.35106494e-02 5.22482693e-01 6.53064668e-01 5.92806697e-01
7.16449171e-02 9.23143327e-01 -5.37727535e-01 -7.26437867e-01
-2.52653182e-01 1.03796482e+00 9.85911965e-01 9.64604735e-01
7.40489483e-01 4.02078062e-01 -2.88738012e-01 6.63783789e-01
-2.11619623e-02 3.55730712e-01 6.51868582e-01 -1.26143265e+00
4.90544438e-01 4.88010973e-01 2.32001305e-01 -1.79274365e-01
-8.48835766e-01 -2.98486620e-01 -4.29403894e-02 7.30910420e-01
5.54131389e-01 -8.78618836e-01 -8.25642705e-01 1.99819696e+00
8.61574635e-02 1.58153921e-01 2.39108935e-01 8.29433680e-01
9.06440794e-01 4.27114934e-01 2.22700946e-02 1.57727554e-01
1.12313175e+00 -9.49468851e-01 -1.59330666e-01 -5.14643192e-01
8.15499008e-01 -1.35514617e-01 1.34619832e+00 2.52923995e-01
-1.28965962e+00 -9.41755250e-02 -8.68765712e-01 2.57904470e-01
-4.75278527e-01 -4.10164446e-01 1.05761302e+00 6.72488987e-01
-1.33953869e+00 8.40495884e-01 -1.16061652e+00 -9.27094370e-02
2.70461738e-01 6.59928381e-01 -8.85721296e-02 1.85147986e-01
-1.33172596e+00 1.01621318e+00 7.28217065e-01 -5.08980572e-01
-1.34154963e+00 -5.00850499e-01 -1.01009846e+00 -7.01933205e-02
7.93141723e-01 -7.12925315e-01 1.86030447e+00 -5.16126752e-01
-1.99725425e+00 6.93919301e-01 6.96169317e-01 -4.68013883e-01
5.65357268e-01 2.02190489e-01 -9.22151655e-02 -2.09766477e-01
2.01195225e-01 6.52998090e-01 5.37562668e-01 -6.83761775e-01
-6.25682592e-01 6.86045811e-02 7.58885026e-01 5.30310273e-01
3.00981581e-01 1.91411555e-01 -4.42435056e-01 -4.50093657e-01
-6.05633438e-01 -1.06047618e+00 -4.13274020e-01 -6.28290415e-01
-2.81172365e-01 -6.28570914e-02 5.38500965e-01 -2.05468461e-02
6.34260654e-01 -1.97784352e+00 2.18426585e-01 8.32499638e-02
2.26425424e-01 -3.50410379e-02 -4.30882215e-01 6.14966571e-01
6.35306761e-02 -3.83804482e-03 9.76677164e-02 -1.83088765e-01
5.13012350e-01 5.67899108e-01 -1.29735112e-01 -8.52345750e-02
1.67132631e-01 1.15082002e+00 -1.22006571e+00 -2.69007444e-01
3.09083164e-01 -1.88590422e-01 -1.07467175e+00 2.61813700e-01
-8.00477922e-01 4.98501390e-01 -6.66686356e-01 1.63430378e-01
-4.68318630e-03 -1.16149798e-01 3.08012992e-01 7.01365590e-01
-1.11788295e-01 7.75252163e-01 -1.20823097e+00 1.70010710e+00
-3.88401747e-01 2.68962771e-01 -6.96064085e-02 -5.44182479e-01
5.21836638e-01 -1.36083728e-02 4.77136493e-01 -8.00727904e-01
3.89932543e-01 -2.38517478e-01 1.98597923e-01 -2.03699753e-01
8.25570464e-01 -4.55546863e-02 -4.85392392e-01 8.00443590e-01
4.98958021e-01 -6.13413155e-01 8.71917903e-01 3.25077444e-01
1.44682395e+00 5.31903207e-01 1.12554915e-01 -2.61049241e-01
-1.55074999e-01 1.95652634e-01 2.37870574e-01 1.16934359e+00
1.05722822e-01 1.13016360e-01 9.13262069e-01 -3.36080521e-01
-8.69947195e-01 -1.24034667e+00 2.67870784e-01 1.99493504e+00
4.09253277e-02 -8.58935237e-01 -5.90102077e-01 -4.99030322e-01
-1.36387110e-01 6.04282856e-01 -6.80212259e-01 -2.64054686e-01
-3.73081714e-01 -8.14356625e-01 8.74020696e-01 4.49590236e-01
4.75296110e-01 -1.69022107e+00 -7.83362031e-01 2.94268101e-01
9.13235452e-03 -9.76471603e-01 -3.65378350e-01 5.64451158e-01
-4.03125674e-01 -1.29395938e+00 -1.72816500e-01 -7.39094436e-01
1.88408010e-02 -2.79157937e-01 1.64073396e+00 8.98015723e-02
-1.79643586e-01 4.33948845e-01 -4.52042699e-01 -5.09527683e-01
-7.64828563e-01 2.75537431e-01 -2.53066868e-02 -8.14422607e-01
-2.98611611e-01 -7.13942170e-01 -2.71815479e-01 3.72981191e-01
-8.21613133e-01 1.50543541e-01 -3.33103538e-03 9.76831496e-01
1.98131308e-01 -1.12967759e-01 4.34475124e-01 -7.35705853e-01
1.20749104e+00 -2.90435165e-01 -1.33279955e+00 -1.04996867e-01
-1.66391492e-01 4.57917839e-01 6.78071201e-01 -3.97593886e-01
-6.66054428e-01 -9.11709294e-02 -3.78867447e-01 -9.75613575e-03
1.37052342e-01 5.23595452e-01 1.45133168e-01 -1.21706091e-01
9.96203363e-01 2.87052542e-01 6.67929575e-02 -5.87694421e-02
5.24631321e-01 1.29555061e-01 5.03623664e-01 -1.02236426e+00
8.13424706e-01 -1.32275015e-01 -2.09160641e-01 -3.57803434e-01
-4.73832250e-01 2.66987626e-02 7.52790719e-02 -5.89869246e-02
5.44880986e-01 -9.31319058e-01 -1.38591802e+00 4.91930574e-01
-4.69347566e-01 -1.51948369e+00 -6.19343817e-01 2.61046916e-01
-1.08153653e+00 -1.39505014e-01 -8.76098633e-01 -4.65483218e-01
-1.21273838e-01 -1.54682422e+00 8.85097146e-01 1.89537153e-01
-1.64899334e-01 -1.08079183e+00 2.87439764e-01 -1.61218598e-01
4.05084848e-01 2.15637296e-01 6.73574686e-01 -6.98918521e-01
-3.60567123e-01 1.42476842e-01 2.30032668e-01 -1.46843791e-01
-1.01815596e-01 -2.92727083e-01 -8.21420014e-01 -4.30576771e-01
-5.20373464e-01 -1.07083261e+00 5.38935900e-01 2.02372834e-01
8.72313321e-01 -2.83079416e-01 7.48022720e-02 7.42482722e-01
9.89525974e-01 1.77561805e-01 6.07681334e-01 7.17819989e-01
1.65888429e-01 1.08733512e-01 4.44448054e-01 7.84018874e-01
6.72368288e-01 7.79489696e-01 8.29076171e-01 2.04103261e-01
5.04793644e-01 -3.16871673e-01 4.51435536e-01 1.67091519e-01
-2.06863046e-01 -1.74780056e-01 -7.55481362e-01 4.26703282e-02
-1.89646924e+00 -1.23821795e+00 5.70906937e-01 1.97593439e+00
1.12778056e+00 4.83140856e-01 7.86309123e-01 -3.65908533e-01
2.69931585e-01 1.27874106e-01 -6.44010186e-01 -4.15845305e-01
1.39050335e-01 7.27718413e-01 4.26943332e-01 6.61723733e-01
-9.09527183e-01 1.52475572e+00 7.30566454e+00 8.42211604e-01
-1.07764971e+00 -1.53787240e-01 1.90656930e-01 -2.47247532e-01
6.61652908e-02 -2.05933481e-01 -6.10687196e-01 2.42293194e-01
7.84349501e-01 -5.22941053e-01 1.17237818e+00 8.69226515e-01
2.23052725e-01 -3.13196182e-02 -1.20794499e+00 6.55571580e-01
-5.25840938e-01 -1.40743923e+00 -3.51033837e-01 8.47676769e-02
6.74900174e-01 7.39245951e-01 1.55682266e-01 1.25276768e+00
1.59194410e+00 -1.18531704e+00 7.08835959e-01 -1.81418136e-01
8.50742757e-01 -8.48620057e-01 4.60310280e-01 5.51602840e-01
-9.87726450e-01 -1.43682048e-01 -2.81194389e-01 -6.64722860e-01
-3.22219402e-01 -5.22357166e-01 -9.35006857e-01 3.40839416e-01
5.03974140e-01 5.57708502e-01 -5.62953770e-01 9.20856118e-01
-4.73823309e-01 3.77939939e-01 -5.06446540e-01 -3.84996742e-01
5.30339301e-01 -1.08644187e-01 3.52243662e-01 8.53205502e-01
-6.56237155e-02 3.59773673e-02 6.16132557e-01 8.94314468e-01
-4.55913320e-02 -1.21693276e-01 -5.97546458e-01 -9.93005410e-02
3.22221965e-01 1.22369540e+00 -5.77284038e-01 -2.02719301e-01
-1.47575334e-01 7.09722996e-01 5.86438835e-01 3.33970696e-01
-1.07164812e+00 -2.32148379e-01 1.03365612e+00 -1.15683734e-01
2.20068559e-01 -3.63847166e-01 3.40786338e-01 -1.25160396e+00
-4.06144977e-01 -1.22048104e+00 3.98794919e-01 -1.01596749e+00
-9.42782342e-01 7.12941885e-01 2.12838084e-01 -8.92776430e-01
-9.28500473e-01 -6.94806516e-01 -7.47976303e-01 6.29004359e-01
-1.02741528e+00 -6.55740917e-01 -1.75462365e-01 1.08093679e+00
2.71442771e-01 -4.64887291e-01 1.05390561e+00 -1.08671270e-01
-6.32320762e-01 4.84099150e-01 1.04459599e-01 3.16854388e-01
2.65036792e-01 -1.58791912e+00 9.97155070e-01 3.85054827e-01
9.33981165e-02 2.35739067e-01 8.00123870e-01 -4.56761867e-01
-1.25243640e+00 -1.10114408e+00 -1.93533480e-01 -6.71123743e-01
1.11202943e+00 -7.27898657e-01 -2.05838040e-01 1.01849329e+00
-2.11252272e-02 -6.09118026e-04 1.78141743e-01 3.06747407e-01
-9.31074098e-02 3.11382711e-01 -7.79061437e-01 1.06809545e+00
1.13973045e+00 -4.08880234e-01 -4.14286703e-01 3.65385622e-01
6.89948022e-01 -1.24019206e+00 -6.25437260e-01 -1.67875975e-01
4.21008199e-01 -1.11897004e+00 8.93769324e-01 -9.18768823e-01
4.46752638e-01 6.67032152e-02 1.70551986e-01 -1.89830792e+00
-3.02998960e-01 -1.16173244e+00 2.84385890e-01 7.08659232e-01
5.13528645e-01 -6.33375227e-01 1.05953336e+00 4.44206655e-01
-2.27365166e-01 -4.52047527e-01 -6.07147694e-01 -8.57028127e-01
3.10896754e-01 -7.94076085e-01 7.79572964e-01 6.66915715e-01
5.00679433e-01 4.23802674e-01 -2.87451178e-01 -1.73193410e-01
2.23559812e-01 -1.02541074e-01 1.01263082e+00 -1.00136364e+00
-8.53864193e-01 -5.84639728e-01 -2.19556227e-01 -8.66738915e-01
2.31312364e-01 -9.68135655e-01 1.02982700e-01 -1.27328694e+00
-8.98224041e-02 -6.23576760e-01 -9.89597291e-03 8.99300098e-01
-1.07357815e-01 2.04974592e-01 4.80495483e-01 -1.11869767e-01
-1.00882673e+00 5.80549479e-01 1.44309580e+00 1.31507859e-01
-5.18190324e-01 1.73606649e-01 -6.31777167e-01 7.07490146e-01
1.10243595e+00 -3.65273774e-01 -6.08020306e-01 -2.38702610e-01
6.32519960e-01 2.94932663e-01 4.03437734e-01 -1.15925932e+00
-4.84697670e-02 -5.12570739e-01 -2.14383844e-02 8.97907615e-02
4.62874323e-01 -3.33426982e-01 -1.46525294e-01 3.08450460e-01
-4.35514450e-01 3.12705010e-01 6.75035834e-01 1.42717317e-01
1.59488827e-01 -6.03539422e-02 5.63834965e-01 -4.11156267e-01
-7.48953760e-01 3.91542435e-01 -8.28939855e-01 8.95848989e-01
9.60392475e-01 -9.20887291e-02 -4.14213747e-01 -8.47978413e-01
-8.83567393e-01 7.05442369e-01 5.43225348e-01 4.31521803e-01
2.55272537e-01 -1.09741008e+00 -5.99563003e-01 3.21996838e-01
2.55562901e-01 -1.42577767e-01 -1.57340974e-01 2.73806781e-01
-6.34714782e-01 -9.85617749e-03 -6.88870072e-01 -1.71185449e-01
-9.74237680e-01 3.78084749e-01 8.22290003e-01 -7.95018733e-01
-6.82865798e-01 6.47633553e-01 2.31320530e-01 -8.76997054e-01
1.58471197e-01 -4.79416341e-01 -2.16392457e-01 -4.30852801e-01
4.78748649e-01 -4.25104126e-02 -1.88834425e-02 -5.80071062e-02
-1.78428859e-01 -2.63927549e-01 9.19987112e-02 -5.05630016e-01
1.37935007e+00 3.61159533e-01 3.10033083e-01 1.79369256e-01
3.19686800e-01 -2.46499494e-01 -1.63012958e+00 3.12009938e-02
-1.42058477e-01 3.41261253e-02 -2.92496234e-01 -1.05102122e+00
-9.20759439e-01 3.72097939e-01 2.31827110e-01 3.89910609e-01
7.16474175e-01 -4.04211283e-02 3.06150109e-01 8.64494801e-01
7.22158790e-01 -9.95277047e-01 2.84851611e-01 1.25838435e+00
6.36049867e-01 -1.03269744e+00 -3.38088125e-01 1.42634287e-01
-7.29391754e-01 7.92102158e-01 8.82114232e-01 -3.87068450e-01
1.85330003e-01 9.08558965e-01 -6.69319034e-02 -1.50861517e-01
-1.26270652e+00 -5.72012484e-01 -3.39251280e-01 1.04886746e+00
2.29815161e-03 -5.01091592e-02 2.25006968e-01 6.50806785e-01
-8.96670640e-01 8.69634673e-02 8.26632202e-01 1.00549221e+00
-2.96556383e-01 -1.32683682e+00 -1.84549794e-01 4.43842351e-01
-4.79037732e-01 -2.73531407e-01 -3.00363779e-01 1.10871172e+00
-1.22916728e-01 8.20127845e-01 -6.34579733e-02 -6.13528430e-01
3.95613134e-01 -2.37658545e-01 7.34269023e-01 -9.17818904e-01
-1.03349483e+00 -2.80721962e-01 5.72942257e-01 -7.73781061e-01
1.98711723e-01 -3.78577769e-01 -1.26123500e+00 -4.88761663e-01
-1.19461693e-01 3.89997572e-01 3.37832063e-01 1.02784312e+00
6.66423216e-02 7.07350850e-01 2.32416242e-01 -1.16863513e+00
-7.04589069e-01 -8.87466788e-01 -7.53062367e-01 3.97927910e-01
1.44041181e-01 -7.37835944e-01 -1.86086483e-02 -3.83290172e-01] | [3.724057197570801, 1.4171382188796997] |
0ef11e4d-e668-4ec6-8a3c-e16c006879d2 | breastscreening-on-the-use-of-multi-modality | 2004.03500 | null | https://arxiv.org/abs/2004.03500v2 | https://arxiv.org/pdf/2004.03500v2.pdf | BreastScreening: On the Use of Multi-Modality in Medical Imaging Diagnosis | This paper describes the field research, design and comparative deployment of a multimodal medical imaging user interface for breast screening. The main contributions described here are threefold: 1) The design of an advanced visual interface for multimodal diagnosis of breast cancer (BreastScreening); 2) Insights from the field comparison of single vs multimodality screening of breast cancer diagnosis with 31 clinicians and 566 images, and 3) The visualization of the two main types of breast lesions in the following image modalities: (i) MammoGraphy (MG) in both Craniocaudal (CC) and Mediolateral oblique (MLO) views; (ii) UltraSound (US); and (iii) Magnetic Resonance Imaging (MRI). We summarize our work with recommendations from the radiologists for guiding the future design of medical imaging interfaces. | ['Nuno Jardim Nunes', 'Jacinto Carlos Nascimento', 'Francisco Maria Calisto'] | 2020-04-07 | null | null | null | null | ['medical-image-retrieval', 'probabilistic-deep-learning', 'breast-cancer-detection', 'medical-image-retrieval', '3d-medical-imaging-segmentation', 'breast-cancer-detection', 'breast-tumour-classification', 'breast-mass-segmentation-in-whole-mammograms', 'automatic-machine-learning-model-selection', 'mathematical-proofs'] | ['computer-vision', 'computer-vision', 'knowledge-base', 'medical', 'medical', 'medical', 'medical', 'medical', 'methodology', 'miscellaneous'] | [ 2.68258691e-01 6.54234886e-01 -6.80349886e-01 -4.68674451e-01
-8.27603340e-01 -6.12058342e-01 2.68583983e-01 6.52444661e-01
-3.54507178e-01 2.09809795e-01 4.13423389e-01 -1.71898878e+00
-1.82824329e-01 -1.56456202e-01 -4.21356291e-01 -3.67404938e-01
-4.82631713e-01 5.48946798e-01 3.76758516e-01 5.34335300e-02
-9.68967602e-02 7.13534832e-01 -1.10470939e+00 7.35470772e-01
3.88919085e-01 9.51673687e-01 3.86419117e-01 1.44517076e+00
-3.88317816e-02 7.37098455e-01 -1.66213900e-01 -5.68450689e-01
-3.50159794e-01 -4.37543690e-01 -8.73347998e-01 1.90313861e-01
4.46356595e-01 -8.45121384e-01 1.53860778e-01 6.22902036e-01
9.01947021e-01 -6.80537939e-01 5.72291851e-01 -5.99046886e-01
-5.88055193e-01 2.71048814e-01 -1.04437375e+00 7.75397658e-01
7.52742290e-01 4.89942022e-02 -2.56291091e-01 -4.54546899e-01
9.66254652e-01 1.26344550e+00 9.35292780e-01 8.13387215e-01
-9.73008692e-01 -1.21083274e-01 -1.74159452e-01 -1.49525091e-01
-7.49521375e-01 -3.81981313e-01 -2.33575851e-01 -7.27661014e-01
5.14742851e-01 1.06473422e+00 8.31833661e-01 8.35550427e-01
4.57084596e-01 6.56138897e-01 1.15202820e+00 -8.60701263e-01
1.08782994e-02 7.48029649e-01 5.23338079e-01 7.74323106e-01
6.98865712e-01 -1.04919992e-01 -1.12357855e-01 -8.27166736e-01
1.03404129e+00 -2.63770372e-01 -2.00940251e-01 -4.20683950e-01
-8.23342562e-01 5.89855611e-01 -3.13046686e-02 6.20498180e-01
-2.89147526e-01 -3.23740482e-01 5.90003788e-01 -1.00837044e-01
6.78993389e-02 -2.45771945e-01 1.49817448e-02 -8.16946290e-03
-2.80628204e-01 -1.40128374e-01 7.06647813e-01 7.58935392e-01
-2.72807717e-01 -5.24668753e-01 -4.35107112e-01 7.79403508e-01
7.71687925e-01 8.06840241e-01 3.75197798e-01 -5.24284482e-01
2.44894698e-01 2.26969391e-01 5.40011108e-01 -7.60876834e-01
-7.56014049e-01 9.38265547e-02 -6.03150547e-01 3.23806144e-02
5.05028784e-01 -4.52377081e-01 -1.13723493e+00 5.53804934e-01
5.12744069e-01 -8.85900080e-01 -2.19244629e-01 8.58591795e-01
1.60777366e+00 -7.81976581e-02 8.05735528e-01 -2.64801681e-01
2.11762261e+00 -1.76844999e-01 -1.20399499e+00 -1.21307336e-01
8.73922586e-01 -5.62201381e-01 7.02257097e-01 5.32311164e-02
-1.50603163e+00 -2.45533317e-01 -9.71340418e-01 1.78072482e-01
-1.85249329e-01 4.98672664e-01 8.54239166e-01 1.52858460e+00
-1.10079634e+00 1.38620529e-02 -1.53830862e+00 -9.53556061e-01
3.84742409e-01 3.27627897e-01 -4.08333540e-01 -8.34723040e-02
-4.33154851e-01 1.07676852e+00 6.49774168e-03 -2.06336498e-01
-3.06102484e-01 -8.48211646e-01 -9.95374024e-01 -6.07049584e-01
-8.35561305e-02 -5.66747367e-01 1.58092546e+00 -7.97269285e-01
-9.82499778e-01 1.55151331e+00 -5.12874365e-01 8.60129297e-02
6.34610236e-01 -1.54422909e-01 -6.38215542e-01 7.56152689e-01
1.46322340e-01 5.97509265e-01 7.23473914e-03 -1.53504157e+00
-1.08787513e+00 -7.21066773e-01 -4.37732488e-01 1.62079260e-01
-1.59183264e-01 6.03315830e-01 -7.10814416e-01 -1.25576451e-01
2.98863232e-01 -6.22154474e-01 -3.64818752e-01 3.31539989e-01
-2.25848511e-01 1.23525180e-01 6.53075159e-01 -1.34443176e+00
1.47191024e+00 -1.92149460e+00 -5.55149734e-01 2.14993820e-01
3.23825985e-01 2.81559765e-01 4.53273267e-01 7.25271404e-02
-4.50644255e-01 5.54825127e-01 6.67625546e-01 2.13314649e-02
-3.64825100e-01 -2.14334264e-01 5.05178392e-01 5.17577767e-01
-1.94583088e-01 1.07516181e+00 -8.86254728e-01 -9.16608512e-01
1.99752808e-01 3.78488243e-01 5.66567555e-02 2.82559134e-02
7.86653578e-01 7.58497953e-01 -5.29809117e-01 1.28006864e+00
9.02433217e-01 -4.94755715e-01 7.99908817e-01 -2.82310992e-01
-2.84878880e-01 -2.47137129e-01 -8.35572243e-01 1.22603333e+00
1.48561284e-01 1.30168244e-01 7.42111504e-01 -2.79488653e-01
2.89801121e-01 7.85960317e-01 6.02952778e-01 -7.46971488e-01
1.07958943e-01 1.33055761e-01 -9.80936959e-02 -1.45729578e+00
-2.15144634e-01 -1.20053522e-01 6.51829898e-01 3.89947623e-01
-4.03410316e-01 3.67924988e-01 1.24084853e-01 6.42317161e-02
1.07872283e+00 -2.63413221e-01 6.76128685e-01 -4.57472950e-01
1.03995919e-01 2.45973051e-01 -6.73459331e-03 1.12092495e+00
-4.16085958e-01 2.39023194e-01 6.78313434e-01 -5.43348610e-01
-6.38855994e-01 -1.02413106e+00 -7.41076410e-01 9.59346294e-01
2.33632952e-01 -1.60969064e-01 -5.34725547e-01 -7.95763016e-01
2.46986628e-01 3.46379191e-01 -9.51019168e-01 6.75828874e-01
-3.60003233e-01 -9.08304513e-01 4.40331459e-01 9.04894829e-01
1.97257847e-01 -3.10722113e-01 -1.44352376e+00 -3.60447466e-01
-2.82818079e-02 -4.48019177e-01 4.66822162e-02 -2.56459843e-02
-1.33043933e+00 -1.12407613e+00 -1.38058269e+00 -6.80608451e-01
7.95113325e-01 1.68338656e-01 9.90982711e-01 1.28282502e-01
-8.76333296e-01 1.14845967e+00 -2.47970626e-01 -6.27571225e-01
-5.92778027e-01 -6.58724129e-01 -6.52335823e-01 -5.05208015e-01
4.26821083e-01 1.73305020e-01 -9.16927338e-01 2.03208759e-01
-8.50300372e-01 2.89399803e-01 9.64809418e-01 5.79850733e-01
3.60250145e-01 -8.50018620e-01 -7.00261118e-03 -1.12385356e+00
5.25338471e-01 -6.91963792e-01 -1.79237887e-01 4.58935380e-01
-9.30678323e-02 -7.89308548e-01 -7.89747298e-01 -3.64512622e-01
-1.55818295e+00 -8.97803605e-02 -6.57853186e-02 2.49354362e-01
-6.08434141e-01 4.26538110e-01 4.64840114e-01 -3.38168442e-01
1.02341247e+00 -3.75755697e-01 2.14147538e-01 -7.04854846e-01
-2.93239117e-01 8.43191206e-01 6.11374497e-01 -5.71102165e-02
-2.42445290e-01 7.48021722e-01 1.09216012e-01 -9.30338442e-01
3.19004320e-02 -4.64808583e-01 -6.07506394e-01 -5.58795691e-01
1.10390329e+00 -4.76875365e-01 -7.71847010e-01 7.58537129e-02
-9.12321270e-01 -3.15596908e-01 1.98088825e-01 7.38357425e-01
1.54395550e-01 4.08873290e-01 -9.81324673e-01 -1.25055754e+00
-2.75622159e-01 -1.18884778e+00 1.21463585e+00 4.46235240e-01
-5.64218342e-01 -1.32291412e+00 -7.10813776e-02 2.44901285e-01
5.18852532e-01 8.45265388e-01 1.27910113e+00 -1.92457616e-01
2.78546304e-01 -4.55907226e-01 -6.23665571e-01 -5.71389496e-01
2.26336554e-01 -6.68710191e-03 -6.45659268e-01 -2.44958252e-01
-2.40857065e-01 -2.61276156e-01 2.59524405e-01 1.03695416e+00
9.05477762e-01 4.33705479e-01 -1.16886234e+00 3.15979719e-01
1.01794124e+00 9.66612279e-01 7.73082137e-01 3.14706326e-01
2.01644987e-01 8.12849820e-01 8.69801223e-01 4.83372092e-01
2.16500327e-01 2.56002098e-01 3.15780789e-01 -1.03765833e+00
-8.58130977e-02 8.08661431e-02 -2.96902686e-01 -1.77382082e-01
-4.49959993e-01 1.62876427e-01 -1.44035971e+00 4.83500659e-01
-1.19828546e+00 -1.99886113e-01 -4.09296930e-01 2.38207626e+00
5.69098294e-01 -2.87530214e-01 3.22395235e-01 -1.81853622e-01
8.89351010e-01 -6.95901215e-01 -1.55911997e-01 -4.60403383e-01
3.47037852e-01 3.25153433e-02 6.69344306e-01 4.16480958e-01
-1.29677808e+00 -1.48696020e-01 8.34657764e+00 1.67505488e-01
-1.17889845e+00 5.18593431e-01 1.14503336e+00 1.46780714e-01
-2.09958658e-01 -3.36750627e-01 -6.82716489e-01 1.59624591e-01
6.38146281e-01 4.73769337e-01 -6.51220679e-01 5.07497549e-01
-2.25659311e-02 -1.02083874e+00 -1.11540759e+00 8.74145329e-01
-1.48207739e-01 -1.28508556e+00 -1.67743787e-01 3.39380950e-02
1.13261990e-01 -4.76557463e-01 1.16817258e-01 -1.93929330e-01
-1.21713936e-01 -1.03669775e+00 4.50479746e-01 5.83978176e-01
1.69916952e+00 -1.46019936e-01 7.01338589e-01 4.21656370e-02
-7.78297424e-01 1.50653183e-01 2.43261248e-01 5.52325010e-01
-1.52013630e-01 1.64732590e-01 -9.91921723e-01 5.53100467e-01
1.19516110e+00 5.36799692e-02 -1.00668430e+00 1.11244690e+00
6.17587149e-01 5.13638496e-01 4.13744450e-02 1.77163735e-01
-2.76257060e-02 1.19270645e-01 4.55811054e-01 1.74265754e+00
1.92490935e-01 3.78134340e-01 -2.98068166e-01 2.02094227e-01
1.10569322e+00 2.26135984e-01 -4.34753746e-01 -3.13541014e-03
1.29149616e-01 1.27694988e+00 -1.21232724e+00 -4.65017557e-01
-6.82936490e-01 4.61822510e-01 -5.20806193e-01 4.33176041e-01
-3.49781394e-01 -1.41505882e-01 -4.23776180e-01 6.51581824e-01
-2.86245018e-01 6.54450059e-01 -7.10294366e-01 -2.69986540e-01
-2.89272845e-01 -7.27614522e-01 9.99574482e-01 -6.74739897e-01
-6.50786519e-01 2.16689974e-01 6.85221136e-01 -8.75772774e-01
-3.36447865e-01 -8.74958277e-01 -3.58272403e-01 1.05756283e+00
-9.36193705e-01 -1.27652621e+00 -5.85623622e-01 -2.38551553e-02
-2.46989399e-01 1.04274906e-01 1.11149716e+00 4.08365250e-01
2.35341536e-03 6.77921772e-01 2.04168204e-02 -1.54634506e-01
8.13028574e-01 -1.61891294e+00 -1.60837159e-01 -7.88973197e-02
-1.30529797e+00 6.27758741e-01 3.69254351e-01 -9.48852658e-01
-1.75833774e+00 -4.95923251e-01 5.95566332e-01 -2.99024135e-01
1.48389107e-02 2.62562454e-01 -2.56907016e-01 8.85439456e-01
5.58769047e-01 -4.37593386e-02 1.20225453e+00 -1.82897836e-01
5.36044598e-01 1.96902916e-01 -1.75207746e+00 8.07412386e-01
4.20374036e-01 2.08449796e-01 5.88711500e-02 3.62985194e-01
-1.41075835e-01 -1.08251286e+00 -1.22418785e+00 7.20699787e-01
1.28343177e+00 -1.01642489e+00 7.13317990e-01 -7.91986406e-01
3.24593455e-01 4.60483283e-01 3.51543605e-01 -5.33381164e-01
-2.14875221e-01 -5.61088502e-01 5.75690009e-02 7.93986559e-01
4.05514449e-01 -8.05761218e-01 6.18665636e-01 8.62084270e-01
1.27488986e-01 -9.38421011e-01 -7.64081895e-01 3.07836473e-01
-2.71580905e-01 -2.45179266e-01 -3.01956162e-02 6.55027568e-01
6.11142635e-01 -1.53999850e-01 3.11160296e-01 1.56297982e-01
2.69241810e-01 -5.88322222e-01 4.93013471e-01 -8.89476120e-01
-4.84134078e-01 -2.69460678e-01 -3.49718213e-01 -8.08875203e-01
-1.28588355e+00 -3.42169195e-01 -4.62115973e-01 -1.63192809e+00
8.24836135e-01 -3.13410968e-01 2.39931256e-01 2.99938053e-01
-1.50580868e-01 1.34586126e-01 -9.01204124e-02 -1.73450023e-01
-3.29412758e-01 -8.15138519e-01 1.78270948e+00 1.67743891e-01
-3.87548213e-03 7.81647488e-02 -9.69901323e-01 7.06792176e-01
1.94368973e-01 -1.76128596e-01 4.08319607e-02 -3.12303007e-01
-7.10869581e-03 1.09961486e+00 4.13249046e-01 -6.74038649e-01
2.92498954e-02 1.52665675e-01 1.20239353e+00 -1.00313878e+00
1.54612541e-01 -7.13550627e-01 2.78267890e-01 1.18211317e+00
-1.68218642e-01 2.17504010e-01 7.40918577e-01 3.09500217e-01
1.54042631e-01 -3.59459043e-01 8.37346256e-01 -2.84728676e-01
-2.43583038e-01 -5.29213846e-01 -9.67141747e-01 -5.30961096e-01
1.16046596e+00 -4.64098334e-01 -4.10834372e-01 -3.89907062e-01
-1.44966424e+00 3.49340886e-01 2.14439988e-01 -1.14712544e-01
4.20150191e-01 -1.23163283e+00 -6.54700816e-01 1.65600136e-01
-3.87250967e-02 -1.89727873e-01 3.58687550e-01 1.73407817e+00
-1.00214815e+00 6.98335230e-01 -8.21054727e-02 -1.08944643e+00
-1.99638796e+00 1.24688923e-01 5.43476403e-01 -4.68731523e-01
-2.97240347e-01 7.32931376e-01 3.91442776e-01 5.89090108e-04
5.73873878e-01 -9.72049311e-02 -4.97823149e-01 -1.19507119e-01
9.57642198e-01 6.50249362e-01 3.91303807e-01 -4.07637507e-01
-6.15148723e-01 2.10282981e-01 -3.90789807e-01 -2.89677978e-01
6.82461739e-01 -3.02620083e-01 -7.74149001e-02 4.17535543e-01
5.39541900e-01 -1.36189789e-01 -3.22219670e-01 1.26033455e-01
1.82409927e-01 -4.80473787e-01 1.61095727e-02 -1.26902521e+00
-5.07562697e-01 7.49619544e-01 1.44482458e+00 4.84551936e-01
1.11240613e+00 3.65355223e-01 -8.78143981e-02 -1.15074039e-01
-8.43744576e-02 -6.76020443e-01 -1.69308469e-01 -4.70331192e-01
9.50429201e-01 -1.13708615e+00 1.66430354e-01 -8.71631682e-01
-1.08975720e+00 1.46048009e+00 2.80364901e-01 6.37232006e-01
8.43436658e-01 8.29334140e-01 6.55650616e-01 -6.73442066e-01
-4.21960443e-01 9.55505744e-02 5.22749424e-01 1.02761745e+00
1.29775715e+00 1.01989463e-01 -8.67422938e-01 6.82144046e-01
7.96745360e-01 4.76033762e-02 -1.09890647e-01 1.84081519e+00
-3.40912491e-01 -5.72854936e-01 -1.16091299e+00 6.67530239e-01
-1.07478356e+00 5.06820023e-01 -5.92371583e-01 1.00782430e+00
1.32733703e-01 9.13694561e-01 -5.74292839e-02 1.57165706e-01
3.71793628e-01 -1.17805958e-01 5.86005807e-01 -5.68314910e-01
-4.82489586e-01 6.73859835e-01 3.76925766e-01 -5.12241542e-01
-2.05041021e-01 -8.84705842e-01 -7.94042289e-01 1.46308973e-01
-1.75504059e-01 -2.11771563e-01 9.46303487e-01 5.04568279e-01
9.71108973e-02 5.60376465e-01 2.03423686e-02 -9.56688344e-01
-3.16586852e-01 -1.24849606e+00 -8.83544564e-01 7.94175863e-02
2.63115585e-01 -4.21906739e-01 2.12272450e-01 4.28200334e-01] | [15.1630277633667, -2.5372512340545654] |
d3e496cd-b7f2-4b79-ac71-b97e2cfc8f8b | lifelong-ensemble-learning-based-on-multiple | 2205.01982 | null | https://arxiv.org/abs/2205.01982v4 | https://arxiv.org/pdf/2205.01982v4.pdf | Lifelong Ensemble Learning based on Multiple Representations for Few-Shot Object Recognition | Service robots are integrating more and more into our daily lives to help us with various tasks. In such environments, robots frequently face new objects while working in the environment and need to learn them in an open-ended fashion. Furthermore, such robots must be able to recognize a wide range of object categories. In this paper, we present a lifelong ensemble learning approach based on multiple representations to address the few-shot object recognition problem. In particular, we form ensemble methods based on deep representations and handcrafted 3D shape descriptors. To facilitate lifelong learning, each approach is equipped with a memory unit for storing and retrieving object information instantly. The proposed model is suitable for open-ended learning scenarios where the number of 3D object categories is not fixed and can grow over time. We have performed extensive sets of experiments to assess the performance of the proposed approach in offline, and open-ended scenarios. For the evaluation purpose, in addition to real object datasets, we generate a large synthetic household objects dataset consisting of 27000 views of 90 objects. Experimental results demonstrate the effectiveness of the proposed method on online few-shot 3D object recognition tasks, as well as its superior performance over the state-of-the-art open-ended learning approaches. Furthermore, our results show that while ensemble learning is modestly beneficial in offline settings, it is significantly beneficial in lifelong few-shot learning situations. Additionally, we demonstrated the effectiveness of our approach in both simulated and real-robot settings, where the robot rapidly learned new categories from limited examples. | ['Songsong Xiong', 'Hamidreza Kasaei'] | 2022-05-04 | null | null | null | null | ['3d-object-recognition'] | ['computer-vision'] | [ 4.07102518e-02 -9.53869298e-02 1.58401296e-01 -3.83702755e-01
-5.20638406e-01 -3.37033838e-01 5.57995319e-01 -7.57310316e-02
-3.40190113e-01 5.66817641e-01 -1.81547716e-01 2.03000143e-01
-2.78926909e-01 -6.85450315e-01 -8.39336872e-01 -6.19297981e-01
-3.12865794e-01 7.17279851e-01 8.99987817e-02 -2.35284910e-01
3.62735651e-02 6.53117836e-01 -2.12162638e+00 -1.08270116e-01
8.03259850e-01 1.12570465e+00 6.46669626e-01 4.67994481e-01
-2.33236458e-02 6.20005071e-01 -4.27796841e-01 -2.97106117e-01
4.64569688e-01 1.02762625e-01 -4.43589151e-01 5.56698024e-01
1.94744706e-01 -5.16313016e-01 -4.89959180e-01 7.64797390e-01
5.77528715e-01 6.80485547e-01 8.13473761e-01 -1.51432061e+00
-5.99937081e-01 7.10695982e-02 -3.11176360e-01 -2.92336605e-02
4.13489819e-01 4.22782212e-01 6.32039070e-01 -1.22527003e+00
4.36699659e-01 1.27679038e+00 5.77466011e-01 5.41616678e-01
-8.32888067e-01 -6.22781992e-01 3.11569273e-01 4.53125387e-01
-1.01657414e+00 -5.30039549e-01 7.67126977e-01 -3.21651191e-01
1.09915960e+00 -3.84338856e-01 5.64230919e-01 1.25665963e+00
5.06864712e-02 8.60016048e-01 7.40756929e-01 -2.21964329e-01
4.98240352e-01 1.09256476e-01 1.89629063e-01 5.95009267e-01
3.92767251e-01 1.03899516e-01 -2.22027749e-01 -1.38847549e-02
5.72349548e-01 7.71065116e-01 1.64985061e-01 -8.66277158e-01
-1.19560373e+00 7.11609602e-01 4.26576763e-01 2.68379599e-01
-5.67888916e-01 1.17888562e-01 4.78191674e-01 3.47618431e-01
2.42843658e-01 2.54472047e-01 -3.18220615e-01 -2.79915094e-01
-2.10549578e-01 1.63561210e-01 9.91544306e-01 1.42445076e+00
8.66850615e-01 -8.21031183e-02 4.43907678e-02 1.21684110e+00
-1.44446352e-02 5.48739970e-01 7.03550756e-01 -1.04069769e+00
4.64087158e-01 6.04662061e-01 4.98651654e-01 -8.68530691e-01
-3.22245806e-01 -8.34404454e-02 -8.63363028e-01 1.73851267e-01
1.98341951e-01 -8.32077395e-03 -1.07618046e+00 1.49273050e+00
4.15040731e-01 2.13637769e-01 3.10825467e-01 7.69873381e-01
7.09031820e-01 6.68835044e-01 -1.29595876e-01 -1.63243487e-01
1.03800666e+00 -1.05389583e+00 -5.41760147e-01 -3.30757886e-01
4.63148504e-01 -3.03019285e-01 1.08900285e+00 2.65892029e-01
-6.38816595e-01 -8.66357744e-01 -1.07272220e+00 2.28357702e-01
-3.97763550e-01 -2.06505042e-02 6.46306872e-01 3.54580075e-01
-6.34435296e-01 5.95010698e-01 -8.59631777e-01 -8.29640388e-01
5.38428009e-01 2.61425883e-01 -5.61864495e-01 -5.91373444e-01
-7.66224921e-01 9.71916318e-01 5.11911511e-01 -7.47791603e-02
-1.18210661e+00 -1.58368260e-01 -1.00777817e+00 1.16241626e-01
5.85762143e-01 -5.42858958e-01 1.46345127e+00 -6.93195820e-01
-1.49367654e+00 5.84017217e-01 1.19268760e-01 -2.97538966e-01
3.26914668e-01 -3.96417797e-01 -2.55610287e-01 7.30698630e-02
1.29025012e-01 4.96457994e-01 8.58070552e-01 -1.36528409e+00
-7.23518074e-01 -6.69734955e-01 3.33720714e-01 4.36973661e-01
-5.28372347e-01 -5.17679513e-01 -2.15592846e-01 -4.08344418e-01
1.13374315e-01 -1.08241391e+00 -3.01831603e-01 1.83645025e-01
2.72614598e-01 -3.74963284e-01 9.56116498e-01 -1.07623734e-01
2.87186831e-01 -2.33195949e+00 5.45282438e-02 -3.31215680e-01
-1.98037788e-01 1.54104456e-01 -2.84801811e-01 4.99156654e-01
3.03201884e-01 -4.02066976e-01 -1.90499529e-01 -4.41307396e-01
1.18598290e-01 5.65980613e-01 -1.11906737e-01 2.66613603e-01
5.83067350e-02 7.68903613e-01 -1.11747670e+00 -2.07428738e-01
3.69237661e-01 2.56749034e-01 -2.66058058e-01 4.64787006e-01
-1.21444010e-03 2.17854992e-01 -3.48257035e-01 8.85614514e-01
4.38345402e-01 -2.31065109e-01 5.33779822e-02 8.25396702e-02
3.09490144e-01 -3.52925926e-01 -1.23461998e+00 1.92249513e+00
-8.07194531e-01 1.65684208e-01 -1.31540954e-01 -1.30825782e+00
1.22855508e+00 1.59394503e-01 4.44215924e-01 -6.00318491e-01
1.71766758e-01 1.92138389e-01 -2.49118611e-01 -7.24847794e-01
5.08975565e-01 -9.49741527e-02 -3.30141366e-01 4.59028333e-01
4.44542170e-01 -2.10594565e-01 2.55864799e-01 -1.17365263e-01
1.30084741e+00 1.83533907e-01 5.39009035e-01 1.56517565e-01
1.59110919e-01 -1.23121798e-01 4.81533527e-01 8.38695824e-01
-4.40738440e-01 3.51271987e-01 -1.78424656e-01 -6.46029949e-01
-1.02185214e+00 -1.09131265e+00 1.91099331e-01 1.19015229e+00
6.29193425e-01 1.30807459e-01 -1.87613547e-01 -7.41301060e-01
1.65514082e-01 8.52487803e-01 -4.17032391e-01 -3.86114568e-01
-4.31446671e-01 -4.16354656e-01 6.20200709e-02 6.73382580e-01
6.26705766e-01 -1.37478042e+00 -1.16499305e+00 4.15548414e-01
-1.14545329e-02 -1.23326159e+00 -2.12906986e-01 2.89339811e-01
-9.65889156e-01 -1.16802430e+00 -9.13669288e-01 -1.08672690e+00
6.77580833e-01 9.26287353e-01 7.30957091e-01 -4.91340935e-01
-3.95455629e-01 9.63429272e-01 -7.76947141e-01 -4.83223885e-01
-2.13489920e-01 -7.47249499e-02 4.43695068e-01 4.17720154e-02
4.18552458e-01 -9.26101565e-01 -4.04152334e-01 4.55217212e-01
-7.58067846e-01 -2.37086803e-01 8.14835846e-01 1.08170509e+00
4.23910588e-01 1.36218861e-01 8.74998927e-01 -4.11437780e-01
4.32464808e-01 -8.05325806e-01 -3.61672908e-01 3.13881367e-01
-3.03612590e-01 -1.84459221e-02 5.69199681e-01 -9.22435880e-01
-1.06606400e+00 2.16567412e-01 2.70085543e-01 -8.63298357e-01
-3.42507303e-01 2.64440328e-01 -2.16607213e-01 1.76400632e-01
4.95492011e-01 2.61986703e-01 2.02936202e-01 -5.28740704e-01
4.75908190e-01 9.60086703e-01 4.64031428e-01 -6.53499305e-01
6.30670309e-01 3.01143765e-01 -3.85136276e-01 -8.11251938e-01
-5.95408320e-01 -5.78446686e-01 -7.84987092e-01 -4.03085113e-01
4.56622928e-01 -1.00073004e+00 -6.23255551e-01 6.89638376e-01
-1.00661170e+00 -3.86045754e-01 -4.25354749e-01 4.76736009e-01
-1.02716732e+00 1.67018995e-01 -2.88271636e-01 -1.06052780e+00
-2.44507626e-01 -1.05132604e+00 1.09834290e+00 3.59347612e-01
1.48794532e-01 -4.97795105e-01 -7.78173655e-02 1.01259500e-01
1.80827633e-01 2.44665191e-01 7.14938641e-01 -8.60260963e-01
-3.83794457e-01 -2.94549465e-01 -1.20618515e-01 2.12780222e-01
4.95137244e-01 -6.69343829e-01 -8.34672987e-01 -5.34671307e-01
6.07165173e-02 -8.23640585e-01 6.72510147e-01 -1.67897508e-01
1.07823431e+00 -2.31045991e-01 -3.10380399e-01 1.88306838e-01
1.23963630e+00 5.78417242e-01 3.28985661e-01 2.35338941e-01
5.13551950e-01 5.94267726e-01 1.14345026e+00 8.26437116e-01
5.19967794e-01 6.88312650e-01 6.31047547e-01 5.59342980e-01
1.35791630e-01 -1.70321241e-01 2.77356625e-01 6.86054051e-01
4.31889631e-02 -1.85921326e-01 -9.04955447e-01 7.73828566e-01
-2.12180901e+00 -1.02646673e+00 7.36137688e-01 2.15876079e+00
3.28637868e-01 1.45372935e-02 9.07774344e-02 3.22805159e-02
8.26731563e-01 -1.17001645e-01 -1.16173315e+00 -1.22467533e-01
3.87693286e-01 -1.42904475e-01 8.28811303e-02 -3.41782600e-01
-1.15240633e+00 7.00108349e-01 5.35701799e+00 5.17261386e-01
-8.34551692e-01 1.88880682e-01 3.69982630e-01 -9.90300775e-02
3.07757318e-01 -4.84527200e-01 -6.78367913e-01 4.22940731e-01
6.08848751e-01 -2.04334393e-01 4.34103549e-01 1.43493533e+00
-2.14445353e-01 -1.48456082e-01 -1.43985951e+00 1.46809506e+00
3.39938492e-01 -8.05720329e-01 -2.17511624e-01 -2.43437421e-02
7.11306214e-01 9.39290449e-02 -9.22838673e-02 7.98450172e-01
4.17604119e-01 -6.90173924e-01 6.53946102e-01 5.82913816e-01
6.78642452e-01 -8.05864573e-01 7.23602414e-01 8.13569307e-01
-1.15345514e+00 -7.01243281e-01 -6.30860448e-01 -2.01755822e-01
5.73753081e-02 3.45217884e-01 -9.18135643e-01 4.06589031e-01
1.00933504e+00 8.26224327e-01 -3.45789552e-01 1.14596534e+00
4.15002555e-02 -8.16397220e-02 -3.93850416e-01 -3.42053056e-01
1.60209864e-01 9.69920587e-03 5.10374606e-01 8.18261087e-01
7.49088109e-01 5.08577883e-01 3.44330907e-01 4.89630669e-01
-1.99297741e-01 -2.01805830e-01 -1.21935523e+00 -1.94707280e-03
7.25403249e-01 1.20604217e+00 -6.18250906e-01 -3.67474824e-01
-4.47253972e-01 1.07154346e+00 5.37239313e-01 2.32997701e-01
-6.57311678e-01 -4.64768827e-01 5.87487757e-01 -2.72456288e-01
6.62937641e-01 -4.60465372e-01 1.35088474e-01 -1.10389090e+00
1.81616813e-01 -8.08482707e-01 2.10974246e-01 -8.47392559e-01
-1.61690521e+00 4.87860501e-01 1.87899485e-01 -1.61330509e+00
-4.48349833e-01 -5.78821599e-01 -4.58406299e-01 7.40939751e-02
-1.12202203e+00 -1.06830728e+00 -6.78478241e-01 6.21753216e-01
1.11460519e+00 -3.49165738e-01 8.31604600e-01 1.45773262e-01
-3.09673220e-01 4.59871888e-01 4.50462580e-01 -2.33248249e-01
5.51916063e-01 -8.64529550e-01 2.38369644e-01 3.89495581e-01
1.64098293e-01 2.96564728e-01 5.10475338e-01 -5.36589324e-01
-1.68965578e+00 -1.29831147e+00 2.00894520e-01 -2.41457224e-01
3.62524927e-01 -4.90743488e-01 -8.45858634e-01 5.66833496e-01
-2.57961512e-01 2.52205580e-01 4.61899966e-01 7.84527138e-02
-1.53855219e-01 -3.02153647e-01 -1.40465295e+00 3.07074964e-01
1.47250986e+00 -1.97132260e-01 -9.90920126e-01 2.88265049e-01
7.56957650e-01 -1.75302774e-01 -7.36344874e-01 6.87083304e-01
7.67047942e-01 -7.86189437e-01 7.93155551e-01 -3.44123393e-01
2.77742445e-01 9.75894928e-03 -6.11889005e-01 -1.52116430e+00
-2.74644226e-01 -1.36844069e-01 -4.93277371e-01 1.08114684e+00
-3.64258885e-02 -6.50876820e-01 6.43918037e-01 4.54594046e-01
-2.90799797e-01 -7.55323052e-01 -8.13367426e-01 -1.25597632e+00
-3.50773513e-01 -3.61149222e-01 6.93457067e-01 5.51900983e-01
-8.46221223e-02 3.13196301e-01 -3.63427103e-01 1.03624053e-01
7.15537608e-01 4.03657049e-01 1.13824725e+00 -1.28341866e+00
-2.70378590e-01 3.36946128e-03 -6.94810331e-01 -1.00837612e+00
2.35145226e-01 -6.08087540e-01 5.01232147e-01 -1.53386378e+00
3.84210378e-01 -5.84862292e-01 -4.21378791e-01 4.72167730e-01
-8.48714709e-02 -1.07713126e-01 5.83471060e-01 2.71126956e-01
-9.87675607e-01 1.07844567e+00 9.58380759e-01 -4.25525665e-01
-3.20542812e-01 1.92361668e-01 -3.14417660e-01 7.77706027e-01
6.31955504e-01 -2.41056636e-01 -7.23133206e-01 -4.23593462e-01
-3.53587747e-01 8.56766924e-02 3.77280027e-01 -1.23845220e+00
1.89311296e-01 -1.45061195e-01 3.46648693e-01 -6.65569484e-01
7.26447701e-01 -1.04848778e+00 -4.30968031e-03 3.49638999e-01
1.11704946e-01 -1.80477172e-01 5.60655184e-02 8.86596859e-01
1.86110418e-02 -2.81679094e-01 7.16812789e-01 -4.06226814e-01
-1.29578149e+00 4.59493041e-01 -1.09739490e-01 -2.74839997e-01
1.48171532e+00 -5.14492989e-01 5.73929958e-02 -3.83920908e-01
-8.54763150e-01 4.48496461e-01 6.20011330e-01 7.95099974e-01
9.56441820e-01 -1.46009696e+00 -3.00745755e-01 2.40475148e-01
5.55890560e-01 1.83322236e-01 2.51449049e-01 3.38004202e-01
-8.66087079e-02 1.10278703e-01 -4.31234807e-01 -6.93513751e-01
-8.45484793e-01 9.35504794e-01 3.52777839e-02 5.51101472e-03
-7.87490964e-01 6.09807134e-01 2.52316952e-01 -7.73574650e-01
6.55399382e-01 -1.55803621e-01 -2.16493100e-01 8.03566501e-02
4.83282834e-01 6.94961667e-01 -4.63515967e-02 -5.38948476e-01
-2.01752260e-01 5.45365334e-01 1.61338486e-02 2.19419330e-01
1.76255131e+00 -2.97306806e-01 4.24051911e-01 9.99026597e-01
1.10791016e+00 -9.04588580e-01 -1.58333230e+00 -3.73957962e-01
-7.82220960e-02 -4.47355032e-01 -4.68776405e-01 -6.64178491e-01
-7.14828551e-01 7.62421608e-01 1.02008450e+00 -1.52513042e-01
1.01131248e+00 3.13783437e-02 8.78333271e-01 1.16009402e+00
1.37316060e+00 -1.04087698e+00 7.26430535e-01 5.95369756e-01
9.75140274e-01 -1.68404329e+00 -2.84038693e-01 -4.94168997e-02
-6.84889615e-01 9.35581982e-01 8.70309949e-01 -2.38282695e-01
5.45011759e-01 -1.55873209e-01 -3.07901770e-01 -8.40284973e-02
-8.19107175e-01 -2.72241265e-01 -1.59798816e-01 8.60134542e-01
-3.74925375e-01 -5.51252533e-03 2.58697420e-01 6.54629409e-01
5.27814701e-02 9.80177075e-02 4.37610537e-01 1.11840463e+00
-7.50232577e-01 -7.37589419e-01 -2.96046764e-01 5.53968072e-01
2.14267761e-01 5.50596595e-01 -2.88441908e-02 7.54321635e-01
1.60268560e-01 1.09777975e+00 8.54873657e-02 -4.09699500e-01
6.19710267e-01 7.42215291e-02 7.11008728e-01 -7.51689136e-01
1.54550791e-01 -4.34052527e-01 -4.87710759e-02 -5.48071027e-01
-3.75625819e-01 -7.75362611e-01 -1.19524515e+00 4.52446043e-02
-2.57997602e-01 -1.41284212e-01 6.75206900e-01 1.00369120e+00
3.94954979e-01 3.16623479e-01 9.60877120e-01 -1.53134453e+00
-9.57953811e-01 -1.17045426e+00 -6.78256869e-01 5.24916351e-01
2.07058370e-01 -1.30933774e+00 -2.96683133e-01 3.97444516e-02] | [7.6367645263671875, -1.2324292659759521] |
fcf4101f-8e7d-434b-9709-e5f5f02f64cc | graph-propagation-transformer-for-graph | 2305.11424 | null | https://arxiv.org/abs/2305.11424v2 | https://arxiv.org/pdf/2305.11424v2.pdf | Graph Propagation Transformer for Graph Representation Learning | This paper presents a novel transformer architecture for graph representation learning. The core insight of our method is to fully consider the information propagation among nodes and edges in a graph when building the attention module in the transformer blocks. Specifically, we propose a new attention mechanism called Graph Propagation Attention (GPA). It explicitly passes the information among nodes and edges in three ways, i.e. node-to-node, node-to-edge, and edge-to-node, which is essential for learning graph-structured data. On this basis, we design an effective transformer architecture named Graph Propagation Transformer (GPTrans) to further help learn graph data. We verify the performance of GPTrans in a wide range of graph learning experiments on several benchmark datasets. These results show that our method outperforms many state-of-the-art transformer-based graph models with better performance. The code will be released at https://github.com/czczup/GPTrans. | ['Yue Qi', 'Cheng Cheng', 'Qiuying Peng', 'Tong Lu', 'Tianrun Shen', 'Tao Wang', 'Hao Tan', 'Zhe Chen'] | 2023-05-19 | null | null | null | null | ['graph-property-prediction', 'graph-regression', 'graph-representation-learning'] | ['graphs', 'graphs', 'methodology'] | [-3.47364634e-01 2.66067535e-01 -3.13931704e-01 -3.17986012e-01
-2.20554158e-01 -2.95621991e-01 5.57592988e-01 2.77070373e-01
9.70112160e-02 5.04162192e-01 2.01201975e-01 -6.38649225e-01
6.99670799e-03 -1.31992269e+00 -7.80135930e-01 -4.75034386e-01
-1.66583374e-01 4.40480322e-01 4.14454520e-01 -3.42012495e-01
-7.14479312e-02 2.37894237e-01 -8.95234108e-01 1.32092923e-01
7.79804289e-01 6.80551648e-01 1.37132660e-01 3.52714777e-01
-3.63870174e-01 1.10853350e+00 -9.12973210e-02 -7.33712196e-01
-1.88862056e-01 -3.16358536e-01 -9.29247916e-01 -1.09948628e-01
-1.18666152e-02 -2.95752704e-01 -8.49034309e-01 1.06001008e+00
1.75218046e-01 -2.88226875e-03 4.31125641e-01 -1.39781451e+00
-1.09775603e+00 9.85015631e-01 -6.07924342e-01 2.23471820e-01
3.58407795e-01 5.11999801e-03 1.51199079e+00 -8.32675636e-01
4.47255611e-01 1.14782608e+00 5.64312160e-01 3.89921404e-02
-7.04025745e-01 -6.71708226e-01 6.26867652e-01 5.28532207e-01
-1.30960786e+00 -3.56761478e-02 1.10559285e+00 -3.23622108e-01
9.84840453e-01 1.37945235e-01 1.01140320e+00 6.58599317e-01
3.77958268e-01 9.96725678e-01 5.58787286e-01 -1.04391679e-01
-2.95018673e-01 -3.60003352e-01 5.28043509e-01 1.14519298e+00
2.40742728e-01 -1.61046460e-01 -2.26694509e-01 1.87667519e-01
8.74817252e-01 2.21409112e-01 -4.36436385e-01 -2.25288764e-01
-8.22875798e-01 8.93850327e-01 1.18797183e+00 3.33424538e-01
-3.87163967e-01 4.16412890e-01 3.60207289e-01 2.39971891e-01
5.23745358e-01 -9.06392112e-02 -1.96222126e-01 1.81721181e-01
-1.87705189e-01 -3.23242955e-02 5.41922688e-01 1.16881490e+00
8.58204007e-01 4.63236161e-02 -3.35954487e-01 7.08446145e-01
7.85289943e-01 1.18612282e-01 2.12016001e-01 -2.11660296e-01
7.60608852e-01 1.04043436e+00 -4.77454901e-01 -1.30157411e+00
-2.74801999e-01 -7.81435907e-01 -1.08273840e+00 -3.40438962e-01
-1.74928576e-01 6.81004822e-02 -1.05641365e+00 1.57075167e+00
3.06029469e-01 5.64554989e-01 -2.54503697e-01 6.73185110e-01
1.46847105e+00 7.52310574e-01 2.20067009e-01 1.10386662e-01
1.30152535e+00 -1.42743087e+00 -7.16253579e-01 -3.41666907e-01
6.47925019e-01 -3.70534033e-01 1.10232008e+00 -1.07339490e-02
-1.05817425e+00 -5.30234337e-01 -7.76632845e-01 -3.36808413e-01
-4.24148858e-01 -1.77260023e-02 8.65345359e-01 1.52052253e-01
-1.11454487e+00 4.63533968e-01 -8.66670847e-01 -2.68726587e-01
4.91369069e-01 2.70436108e-01 -4.24364537e-01 -3.16077292e-01
-1.42517245e+00 5.01091719e-01 3.49010468e-01 2.90326059e-01
-9.27262306e-01 -4.42964405e-01 -1.22422361e+00 5.21333396e-01
2.94498295e-01 -8.18849504e-01 1.26028216e+00 -5.67885816e-01
-1.19308066e+00 7.26196766e-01 -2.54759073e-01 -4.21162605e-01
1.45902485e-01 -9.17589515e-02 -2.89162189e-01 -6.47633001e-02
-1.05188020e-01 3.62850845e-01 4.84553933e-01 -1.13191962e+00
-2.53927410e-01 -2.55558759e-01 4.41573799e-01 2.13340029e-01
-3.07391286e-01 -7.86179826e-02 -1.13367903e+00 -6.75414264e-01
9.80376229e-02 -7.47729957e-01 -8.94432738e-02 -4.02992725e-01
-7.72862852e-01 -6.35943949e-01 8.65285873e-01 -7.57085919e-01
1.63981152e+00 -1.87679911e+00 1.93068758e-01 1.60562247e-01
8.84688377e-01 3.66729707e-01 -1.92146957e-01 8.40120494e-01
-3.41085464e-01 2.60251641e-01 -5.80472276e-02 -4.25233006e-01
4.34328280e-02 2.40260810e-01 -1.53582424e-01 1.99555323e-01
6.74629584e-02 1.50630510e+00 -1.04570282e+00 -3.71120721e-01
2.17618391e-01 6.83077097e-01 -5.77421427e-01 3.19093227e-01
-1.12526074e-01 2.05173373e-01 -7.88734198e-01 5.58777332e-01
6.57529414e-01 -7.75953531e-01 1.91255048e-01 -3.76181394e-01
2.45306805e-01 6.65168226e-01 -5.73238075e-01 1.42702365e+00
-2.53232539e-01 3.74511570e-01 -3.91971990e-02 -1.09578800e+00
8.67177784e-01 1.03065401e-01 3.25132430e-01 -6.68916166e-01
3.12938541e-01 -2.25465357e-01 1.23757586e-01 -2.37902597e-01
2.72127569e-01 1.02122568e-01 1.01277277e-01 4.14895117e-01
2.83442408e-01 2.84697294e-01 2.93407947e-01 6.41916394e-01
1.17282903e+00 -5.35540245e-02 2.81883627e-01 -1.43603638e-01
6.23550117e-01 -4.18097377e-01 5.65186381e-01 3.50700915e-01
1.36354594e-02 3.20954949e-01 8.87533486e-01 -4.79693532e-01
-5.89931548e-01 -9.29431617e-01 5.05677104e-01 1.04687440e+00
1.14230588e-01 -9.85837042e-01 -5.72836041e-01 -8.26584458e-01
-1.16339251e-01 4.59400684e-01 -7.35476255e-01 -4.29317296e-01
-4.23521876e-01 -4.49205637e-01 1.20031990e-01 7.25379109e-01
6.12440765e-01 -1.20403349e+00 1.68514132e-01 2.51831621e-01
-1.52925313e-01 -9.86742735e-01 -7.36141145e-01 -1.82406768e-01
-7.39584923e-01 -1.26453197e+00 -4.69815195e-01 -1.01115775e+00
8.91493499e-01 3.73939157e-01 1.41557539e+00 7.40513682e-01
1.59209594e-01 1.92885697e-01 -5.47336340e-01 -1.95176214e-01
2.24975049e-02 3.02114397e-01 -5.27894557e-01 -1.60248652e-02
1.38114452e-01 -7.54285932e-01 -4.93506610e-01 1.89229026e-01
-5.66366017e-01 4.05918479e-01 5.55537224e-01 7.61167526e-01
5.65039098e-01 3.24221909e-01 2.39578828e-01 -1.20430899e+00
8.88504267e-01 -7.37576246e-01 -5.49183905e-01 3.60004604e-01
-5.40582120e-01 5.02274930e-02 7.14907944e-01 -3.62143517e-02
-5.92909396e-01 -2.91659713e-01 -4.30835068e-01 -6.52396798e-01
3.34929675e-01 1.11968303e+00 -3.63153100e-01 -1.25490561e-01
-5.17907478e-02 3.70147169e-01 -3.30146253e-01 -4.87852603e-01
3.39485258e-01 2.44776472e-01 1.54121608e-01 -5.13054729e-01
8.27279747e-01 9.41351205e-02 -1.15434319e-01 -3.33622098e-01
-6.65497363e-01 -1.97966859e-01 -4.32130605e-01 -1.96911782e-01
5.95779061e-01 -7.37182260e-01 -7.27665424e-01 7.37133205e-01
-1.10955346e+00 -5.87606728e-01 1.37912348e-01 2.12762624e-01
-6.53788373e-02 3.57706547e-01 -9.47197676e-01 -4.41189617e-01
-5.89061141e-01 -1.23915839e+00 8.04638028e-01 1.80954397e-01
3.32353562e-01 -1.49744952e+00 -7.05415979e-02 1.80817023e-01
3.48159552e-01 -2.67964620e-02 1.01164961e+00 -6.13078117e-01
-8.74730766e-01 -1.39153033e-01 -5.67981660e-01 8.85056853e-02
3.24774235e-01 1.09720849e-01 -5.09378314e-01 -4.09628898e-01
-5.70223987e-01 -8.42207968e-02 1.08963943e+00 2.16898829e-01
1.27713084e+00 -3.58306646e-01 -6.25124574e-01 8.42297792e-01
1.38481927e+00 2.83597503e-02 7.78767586e-01 1.82520151e-02
1.25291669e+00 1.80358008e-01 2.63112783e-01 4.42537814e-02
9.97918785e-01 4.42703396e-01 7.02954471e-01 -2.81081617e-01
-2.89265841e-01 -7.53754139e-01 1.59381941e-01 1.18007088e+00
-6.47043884e-02 -7.37175286e-01 -1.12236786e+00 5.53029895e-01
-1.88951099e+00 -7.12378204e-01 -3.58221412e-01 1.87901664e+00
3.82703155e-01 2.73952574e-01 -7.20367506e-02 4.42053750e-02
8.46515715e-01 4.29620296e-01 -3.87203634e-01 -2.47311443e-01
3.53025794e-01 2.07991853e-01 3.18130165e-01 7.15502203e-01
-8.73832464e-01 1.27838433e+00 5.67670965e+00 5.87335169e-01
-9.55388546e-01 -2.61527505e-02 6.14131510e-01 6.05412185e-01
-7.11209238e-01 4.74231616e-02 -5.50815642e-01 4.97777611e-01
5.55236399e-01 -5.13944864e-01 5.58951497e-01 6.28835559e-01
-7.82794282e-02 4.98143047e-01 -8.42189312e-01 8.37278724e-01
-3.04066371e-02 -1.32486010e+00 4.05444473e-01 -5.34034297e-02
2.81363904e-01 2.87735134e-01 -1.14983357e-01 5.64604998e-01
6.71763420e-01 -9.59140956e-01 2.93945998e-01 2.43324324e-01
5.84344506e-01 -7.15248168e-01 7.74237037e-01 8.24560225e-02
-1.95822251e+00 2.17961688e-02 -2.67622203e-01 4.29850630e-03
1.81317687e-01 7.40280390e-01 -5.83378315e-01 1.10723209e+00
5.61126649e-01 1.39035761e+00 -5.65370262e-01 1.05538785e+00
-8.77373755e-01 7.00707495e-01 -1.35281635e-02 1.39908671e-01
4.25537199e-01 -4.11864609e-01 3.28848094e-01 9.29569542e-01
3.35880667e-01 1.28592521e-01 2.55447507e-01 9.20623779e-01
-5.45364320e-01 1.18870728e-01 -5.80198169e-01 -3.50901753e-01
6.36441767e-01 1.28793550e+00 -6.02967262e-01 -2.91168958e-01
-6.42128468e-01 7.58291066e-01 9.55419004e-01 4.60599869e-01
-9.56828356e-01 -5.79870462e-01 4.67863768e-01 2.34973982e-01
4.21446979e-01 -1.96690068e-01 2.59031028e-01 -1.08080494e+00
9.20045897e-02 -6.40501261e-01 7.52934754e-01 -1.11635661e+00
-1.31525087e+00 8.28675389e-01 -6.42341226e-02 -8.88058066e-01
2.21983016e-01 -4.86374855e-01 -1.14709222e+00 8.67511868e-01
-1.70508277e+00 -1.52441835e+00 -6.90197110e-01 8.21159422e-01
1.09168291e-01 8.22392106e-02 4.88740772e-01 4.53781307e-01
-7.01962471e-01 6.27164543e-01 -3.32457066e-01 6.92479789e-01
9.55951288e-02 -1.15240598e+00 1.13241470e+00 9.14237440e-01
2.73780048e-01 7.59611785e-01 5.21553978e-02 -7.03598619e-01
-1.42760980e+00 -1.28702295e+00 9.98126924e-01 -4.78124805e-02
8.42799783e-01 -4.63776648e-01 -1.11104333e+00 1.37177002e+00
3.48174423e-01 1.70881987e-01 3.68171066e-01 3.23047936e-01
-4.73879039e-01 -1.21366724e-01 -6.58603251e-01 5.45781434e-01
1.30109477e+00 -5.26976705e-01 -4.61163968e-01 3.05697501e-01
1.10528851e+00 -5.46616316e-01 -9.22150195e-01 4.31359082e-01
2.36208871e-01 -8.21733415e-01 8.88963103e-01 -5.91640055e-01
4.39555705e-01 -2.36958429e-01 2.93757886e-01 -1.66226566e+00
-9.04663086e-01 -5.83178997e-01 -3.39723676e-01 1.31251729e+00
4.29049760e-01 -1.04387462e+00 7.24084556e-01 1.78321108e-01
-5.21307528e-01 -1.04012382e+00 -5.19258380e-01 -4.35534090e-01
2.10540965e-02 -3.04007262e-01 9.68854368e-01 1.08609807e+00
6.81622699e-02 6.81463361e-01 -2.51172334e-01 1.95764020e-01
5.64729810e-01 1.93569794e-01 6.88280225e-01 -1.24933028e+00
-2.14390993e-01 -6.43067896e-01 -5.11046112e-01 -1.32912374e+00
2.95295238e-01 -1.31012154e+00 -3.88223290e-01 -2.42269897e+00
3.12147290e-01 -5.48772752e-01 -5.66554010e-01 8.93963814e-01
-5.25199533e-01 -1.22933149e-01 2.33578384e-01 -5.96702285e-02
-6.04481995e-01 8.65374625e-01 1.62009072e+00 -3.51052403e-01
6.31852597e-02 -1.05517201e-01 -9.68614817e-01 4.54294562e-01
1.07075775e+00 -3.83450359e-01 -7.84974515e-01 -8.23739648e-01
2.50631124e-01 -9.87224355e-02 2.32779488e-01 -6.96308196e-01
3.74476343e-01 3.29633243e-02 -1.63852423e-01 -6.34834766e-01
1.21942498e-01 -6.74352944e-01 1.03671275e-01 4.94164139e-01
-1.22889854e-01 3.20613742e-01 4.26463224e-02 4.78713453e-01
-3.82158726e-01 3.58862519e-01 4.68856961e-01 -1.06403813e-01
-6.95735574e-01 1.07809520e+00 1.38597086e-01 1.58476412e-01
8.74842405e-01 2.21978441e-01 -6.62754416e-01 -6.01821125e-01
-6.34215653e-01 7.05617428e-01 1.99071437e-01 5.20476997e-01
7.52195716e-01 -1.58149827e+00 -8.13125789e-01 4.70343798e-01
2.54945070e-01 1.36594489e-01 2.01186433e-01 8.65587175e-01
-4.90628064e-01 4.87403184e-01 -4.71533872e-02 -3.12786728e-01
-1.30509174e+00 5.88171661e-01 5.54065466e-01 -5.72320163e-01
-8.34303439e-01 1.15080225e+00 4.86024380e-01 -4.52888966e-01
1.39574483e-01 -4.31626827e-01 -3.25525999e-01 -4.49907541e-01
3.51511419e-01 5.19693419e-02 -5.91523433e-03 -6.57995820e-01
-3.60202521e-01 6.42030418e-01 -3.31533194e-01 5.22725523e-01
1.20422649e+00 1.52570441e-01 -5.27215719e-01 1.30304977e-01
1.15965474e+00 -1.23299435e-01 -9.39332724e-01 -3.56239229e-01
-3.83150727e-01 -4.24890071e-01 3.29779893e-01 -4.82126385e-01
-1.70873904e+00 1.07163322e+00 -1.62994385e-01 3.76273036e-01
1.09292972e+00 2.06019655e-01 9.80209291e-01 2.08266631e-01
3.94992471e-01 -4.40145999e-01 -1.73818972e-02 6.35200560e-01
1.02398133e+00 -1.01239562e+00 -4.11008596e-02 -7.65952468e-01
-5.12211502e-01 7.30391681e-01 7.35211134e-01 -1.83488175e-01
9.68249977e-01 5.22056930e-02 -3.25545639e-01 -5.63534737e-01
-1.00326836e+00 -2.89906234e-01 4.68929470e-01 5.25615394e-01
6.78850889e-01 1.43456504e-01 -1.65820196e-01 5.27692318e-01
-2.02449504e-02 -1.71849877e-02 2.45599687e-01 7.01318145e-01
-1.90555185e-01 -1.33551979e+00 3.48876923e-01 4.82979447e-01
-1.08376309e-01 -4.53083724e-01 -5.71658373e-01 8.84261310e-01
-3.20295364e-01 8.20438266e-01 -2.73718424e-02 -5.99897146e-01
2.26239324e-01 -2.72398949e-01 3.02859485e-01 -7.41983175e-01
-5.15239775e-01 -2.42569014e-01 1.47843316e-01 -5.52837431e-01
-1.08993784e-01 -1.43656820e-01 -1.47725296e+00 -6.47750497e-01
-2.66917616e-01 4.06427443e-01 1.33767486e-01 6.32479310e-01
6.02846503e-01 1.07121813e+00 4.75709170e-01 -5.49264491e-01
1.80753335e-01 -8.96090269e-01 -4.74380314e-01 3.08012784e-01
2.72810131e-01 -7.38242090e-01 -1.75107509e-01 -3.91251773e-01] | [7.128396034240723, 6.305565357208252] |
39c3aad6-9323-4902-a551-409180f6b77b | semantic-interaction-in-augmented-reality | 2112.05846 | null | https://arxiv.org/abs/2112.05846v1 | https://arxiv.org/pdf/2112.05846v1.pdf | Semantic Interaction in Augmented Reality Environments for Microsoft HoloLens | Augmented Reality is a promising technique for human-machine interaction. Especially in robotics, which always considers systems in their environment, it is highly beneficial to display visualizations and receive user input directly in exactly that environment. We explore this idea using the Microsoft HoloLens, with which we capture indoor environments and display interaction cues with known object classes. The 3D mesh recorded by the HoloLens is annotated on-line, as the user moves, with semantic classes using a projective approach, which allows us to use a state-of-the-art 2D semantic segmentation method. The results are fused onto the mesh; prominent object segments are identified and displayed in 3D to the user. Finally, the user can trigger actions by gesturing at the object. We both present qualitative results and analyze the accuracy and performance of our method in detail on an indoor dataset. | ['Sven Behnke', 'Max Schwarz', 'Peer Schüett'] | 2021-11-18 | null | null | null | null | ['2d-semantic-segmentation'] | ['computer-vision'] | [ 3.92794788e-01 3.34687382e-01 4.45488900e-01 -1.94435641e-01
-2.40596473e-01 -7.99497008e-01 4.99281347e-01 2.89245963e-01
-3.55318785e-01 3.31784487e-01 -1.13737248e-01 -2.18655944e-01
-1.36676177e-01 -7.19093740e-01 -5.71867526e-01 -2.24798664e-01
-9.67988893e-02 7.37191796e-01 6.79698527e-01 -3.59689683e-01
3.13857496e-01 9.50009048e-01 -2.05454564e+00 2.42361411e-01
7.60004103e-01 8.18650424e-01 5.60429573e-01 8.14338386e-01
-1.96012467e-01 1.33290544e-01 -5.65931141e-01 -1.45708099e-01
4.04783338e-01 -2.65169628e-02 -5.97593844e-01 4.61726040e-01
3.28065634e-01 -1.71681643e-01 1.11623272e-01 1.06339478e+00
1.92257911e-01 2.30988428e-01 3.39538217e-01 -9.03450608e-01
-3.54859419e-02 1.03027925e-01 -3.09699953e-01 -3.59796882e-01
1.33020985e+00 6.63351733e-03 4.24281061e-01 -9.46759582e-01
1.19312990e+00 1.16372824e+00 4.65553194e-01 3.65447551e-01
-1.16286862e+00 -2.12356851e-01 1.89081118e-01 -3.14738780e-01
-1.21588969e+00 -9.92086157e-02 9.29224670e-01 -6.61421597e-01
5.97888768e-01 8.25222313e-01 1.10061038e+00 6.86324358e-01
-6.29773289e-02 7.99308360e-01 1.21278369e+00 -5.96987426e-01
4.57216591e-01 4.74772722e-01 2.08761767e-01 7.38670647e-01
-8.95135030e-02 1.46199673e-01 -4.36421335e-01 5.08226342e-02
1.31822193e+00 3.66233259e-01 -5.28411329e-01 -1.06270695e+00
-1.37506950e+00 4.30393703e-02 6.24812543e-01 2.75810361e-01
-4.14499253e-01 2.12694029e-03 -1.82409227e-01 -7.16541186e-02
4.68416750e-01 3.94750267e-01 -3.59406531e-01 -3.06846440e-01
-6.80592775e-01 3.51241469e-01 7.66507745e-01 1.17383373e+00
6.99729443e-01 -5.20288885e-01 2.45552018e-01 6.02359653e-01
2.99244314e-01 3.50985765e-01 -8.53389204e-02 -1.22840142e+00
1.35909513e-01 9.35369849e-01 4.49715793e-01 -8.39458346e-01
-5.14429390e-01 4.84928228e-02 -2.43674189e-01 1.01822114e+00
3.56243461e-01 1.26277626e-01 -9.86160636e-01 7.46094823e-01
7.80993998e-01 -2.36197766e-02 -1.31549165e-01 1.13757968e+00
7.16068029e-01 3.55789632e-01 -2.61555940e-01 1.07470691e-01
1.47707939e+00 -6.64236009e-01 -8.84183347e-01 -1.11718476e-01
5.62257051e-01 -6.84521139e-01 1.47241545e+00 7.25413501e-01
-1.02901781e+00 -5.20202458e-01 -9.72786009e-01 -1.80568650e-01
-6.48943245e-01 3.73316482e-02 6.61195695e-01 5.61675668e-01
-7.28866160e-01 9.02907908e-01 -1.15606415e+00 -7.10441351e-01
2.08010554e-01 3.58089119e-01 -5.01728356e-01 9.56661329e-02
-5.58777571e-01 9.39435184e-01 3.56175900e-01 1.60775170e-01
-3.56796980e-01 -5.35284400e-01 -9.17859256e-01 -2.45812684e-01
4.99397069e-01 -5.67614615e-01 1.28032911e+00 -3.71842861e-01
-1.88047647e+00 1.10309672e+00 1.72298819e-01 1.84283238e-02
8.34450364e-01 -6.18054628e-01 -6.23387918e-02 2.34163642e-01
-6.41452000e-02 5.43405950e-01 2.00201049e-01 -2.02497244e+00
-7.64404476e-01 -7.64181674e-01 4.78778511e-01 4.35380936e-01
2.43470788e-01 -2.03288406e-01 -7.55369186e-01 -2.50699997e-01
6.68077052e-01 -8.32006991e-01 -5.46005666e-01 4.96427387e-01
-5.95381796e-01 1.10659204e-01 1.20435309e+00 -5.67997694e-01
9.03625071e-01 -2.25003004e+00 1.74305007e-01 5.35780549e-01
1.49533125e-02 -7.06103817e-02 5.55563509e-01 2.47318164e-01
2.52093107e-01 -1.58310965e-01 -4.07052130e-01 -7.70438313e-01
1.10266611e-01 2.71502346e-01 8.41770247e-02 3.85447204e-01
-5.44857502e-01 5.81080079e-01 -1.06861210e+00 -5.05369544e-01
8.80779088e-01 6.90607846e-01 -3.67845684e-01 4.38672036e-01
-2.25353390e-01 7.86726832e-01 -4.41782147e-01 6.76722527e-01
5.52629590e-01 1.46007761e-01 1.13596864e-01 -1.00577928e-01
-4.92382556e-01 1.66288525e-01 -1.48650479e+00 2.38328266e+00
-5.85466504e-01 4.04737025e-01 2.60293156e-01 -3.25159639e-01
9.77734089e-01 2.15306848e-01 3.52308840e-01 -4.53696817e-01
1.11675702e-01 2.10213289e-01 -7.12601125e-01 -7.33437657e-01
8.15775692e-01 4.27900732e-01 -1.16287746e-01 1.88565105e-01
-3.61562014e-01 -6.83606505e-01 -8.95300321e-03 3.63893956e-02
8.11239719e-01 8.16973507e-01 3.95205438e-01 -1.50816470e-01
4.28554660e-04 2.25923851e-01 -2.56042868e-01 5.33452570e-01
2.42191866e-01 8.99202466e-01 1.25147551e-01 -2.64258295e-01
-7.32413352e-01 -1.21939528e+00 -2.93511748e-01 7.31436431e-01
7.61746049e-01 -1.12997547e-01 -8.51619720e-01 -5.60898304e-01
-1.45732075e-01 7.93171883e-01 -5.95409811e-01 4.46287274e-01
-4.73437190e-01 1.45875096e-01 -3.39366466e-01 5.52948058e-01
3.75371575e-01 -1.10753345e+00 -1.67114711e+00 5.87347075e-02
1.73395902e-01 -1.02249706e+00 9.08403918e-02 3.17205340e-01
-1.05130160e+00 -1.07388794e+00 -7.70722747e-01 -5.91615856e-01
9.72665608e-01 2.13137135e-01 9.48759139e-01 -1.75197329e-03
-2.91975349e-01 7.84098625e-01 -5.80853999e-01 -3.18270624e-01
-1.85553282e-01 -2.47439027e-01 -2.30203852e-01 -2.93003380e-01
-1.56276539e-01 -7.33873129e-01 -5.98909497e-01 3.90231162e-01
-6.39673173e-01 4.85073239e-01 4.49645072e-02 7.70397857e-02
8.68942499e-01 -2.53148228e-01 -4.74745125e-01 -1.12805140e+00
2.36794397e-01 -2.53264345e-02 -6.71481073e-01 -2.60979198e-02
-7.85407331e-03 -4.68333840e-01 1.85764030e-01 -4.02408272e-01
-9.95609999e-01 5.80593586e-01 1.92890149e-02 -5.16815126e-01
-8.08455229e-01 2.84766316e-01 -5.98977804e-01 -2.56657116e-02
6.07997000e-01 -4.15300369e-01 -2.84324437e-01 -7.95232832e-01
6.46790862e-01 7.31022716e-01 9.50351238e-01 -3.59500259e-01
5.69227934e-01 7.11203098e-01 -1.00964069e-01 -8.25633705e-01
-3.61977011e-01 -5.91437519e-01 -1.31021607e+00 -6.01260006e-01
8.51664662e-01 -2.53348142e-01 -9.07094181e-01 9.09113511e-02
-1.35454369e+00 -4.69537079e-01 -6.53003931e-01 4.47612822e-01
-8.19344401e-01 -1.60257202e-02 -3.02342474e-01 -1.20891130e+00
3.79805475e-01 -1.12556791e+00 1.52343512e+00 2.18061015e-01
-4.19435859e-01 -8.11477900e-01 -2.14210451e-01 1.04096308e-01
-2.80063659e-01 8.03281963e-01 3.83186251e-01 -3.81464750e-01
-5.89653075e-01 -3.42216939e-01 8.96161869e-02 -3.00896555e-01
1.73481241e-01 -1.70653593e-02 -1.10166430e+00 3.92991245e-01
-2.10449919e-01 3.11145574e-01 7.82512352e-02 1.95160612e-01
1.27446222e+00 6.61786124e-02 -8.52390289e-01 4.24048364e-01
1.30967855e+00 4.01803583e-01 7.00528443e-01 3.66789937e-01
8.67186725e-01 9.42291200e-01 9.62137163e-01 1.69477403e-01
1.40746087e-01 1.11502492e+00 6.30355597e-01 -3.48257661e-01
6.28203005e-02 -3.06228995e-01 -1.99736029e-01 3.42354566e-01
-5.25480330e-01 -6.68842345e-02 -9.33205724e-01 1.76087499e-01
-1.93883479e+00 -4.07147437e-01 -5.58100939e-01 2.28418398e+00
4.03249592e-01 3.26162636e-01 6.68848976e-02 5.88915765e-01
4.81348783e-01 -3.23240340e-01 -7.40720928e-02 -2.46835977e-01
3.34716439e-01 2.60474890e-01 3.97748500e-01 5.55904448e-01
-9.23601449e-01 9.07833755e-01 5.85838985e+00 2.77245969e-01
-9.86941278e-01 -1.81319527e-02 -1.61898676e-02 8.19584206e-02
-1.07056446e-01 -1.02307513e-01 -2.84632534e-01 1.81098357e-01
2.14703768e-01 3.19630653e-01 2.78047085e-01 1.06073332e+00
3.31583589e-01 -7.05439925e-01 -1.45023346e+00 1.09630632e+00
-2.57706553e-01 -1.09715366e+00 -3.50966334e-01 1.59361720e-01
4.29002434e-01 -6.09326422e-01 -6.96223229e-02 -1.90436885e-01
1.28754556e-01 -7.51508951e-01 1.04070139e+00 6.92348957e-01
8.40330839e-01 -5.49847484e-01 2.34955996e-01 5.37496567e-01
-1.13729036e+00 2.57772088e-01 1.36922792e-01 -1.82231382e-01
6.33452773e-01 4.02948350e-01 -9.06931520e-01 5.76066792e-01
8.08236897e-01 4.15943444e-01 -2.76753068e-01 1.20277798e+00
-2.29172081e-01 7.52759399e-03 -5.73164046e-01 -1.07378354e-02
-2.29642719e-01 -4.94861066e-01 7.47065604e-01 1.23755467e+00
3.52741212e-01 4.22989368e-01 3.97824526e-01 8.72088432e-01
3.65532488e-01 1.78175271e-02 -9.11123574e-01 3.83047223e-01
3.08950335e-01 1.20265889e+00 -1.46368361e+00 -3.91789347e-01
7.10952431e-02 1.52644885e+00 -1.46398306e-01 3.65153074e-01
-3.76895398e-01 -5.37474811e-01 4.10024554e-01 4.53303993e-01
-3.55794318e-02 -6.57015204e-01 -7.07809150e-01 -7.08872676e-01
1.57658011e-01 -3.90365720e-01 -4.41615954e-02 -1.25198972e+00
-5.65109968e-01 6.59215152e-01 2.60185570e-01 -1.50796735e+00
-8.74913484e-02 -6.36382699e-01 -1.79424062e-01 6.02566421e-01
-7.96996832e-01 -1.17543316e+00 -6.77726328e-01 3.68644714e-01
7.23092616e-01 6.45349205e-01 1.00879133e+00 2.22784027e-01
6.28223270e-03 -1.80117264e-01 -2.62189418e-01 -2.77015507e-01
1.32135212e-01 -1.54343021e+00 4.63058025e-01 5.19891679e-01
4.01944071e-01 6.42978668e-01 9.88625407e-01 -8.19935143e-01
-1.28649628e+00 -5.01755297e-01 4.47180182e-01 -8.86137843e-01
1.32723808e-01 -8.00202489e-01 -1.02796006e+00 6.69967413e-01
-1.78632200e-01 5.46573438e-02 2.32154563e-01 -3.71855423e-02
1.83678284e-01 3.60645086e-01 -1.46254158e+00 8.44543993e-01
1.56316304e+00 -4.66752142e-01 -7.38875031e-01 1.77799001e-01
7.55303264e-01 -1.35670710e+00 -5.40431738e-01 3.76174837e-01
7.42461205e-01 -1.33828413e+00 9.71002519e-01 -7.29781911e-02
1.04649365e-01 -6.03096604e-01 -7.21349418e-02 -1.38547981e+00
1.72144085e-01 -5.14130712e-01 9.66662262e-03 7.77903020e-01
2.58843005e-01 -4.14882064e-01 8.53019834e-01 8.86403561e-01
-4.28327888e-01 -6.06167614e-01 -7.36675560e-01 -5.04963279e-01
-8.07657838e-01 -7.54944384e-01 6.87984943e-01 7.70863295e-01
2.67711312e-01 -2.63534516e-01 1.82927966e-01 5.91521859e-01
4.15053785e-01 1.25169769e-01 1.00137067e+00 -1.58975434e+00
9.43297371e-02 -2.02718467e-01 -5.03531516e-01 -1.32870400e+00
-1.57576859e-01 -5.89258134e-01 2.90807486e-01 -1.93338084e+00
-4.57134396e-01 -6.07171953e-01 3.22050154e-01 2.41813570e-01
1.07336298e-01 2.47820735e-01 2.02729106e-01 2.11286053e-01
-4.42525715e-01 1.77190989e-01 1.24941611e+00 3.92702669e-01
-7.70853996e-01 1.74979091e-01 1.40550667e-02 1.16831946e+00
3.41804475e-01 -1.90523967e-01 -1.69615537e-01 -2.20790058e-01
-1.41192421e-01 1.85984284e-01 6.18503034e-01 -1.31315529e+00
1.68953493e-01 -8.27810317e-02 5.27129054e-01 -8.19546402e-01
9.54778850e-01 -1.49589455e+00 5.64405859e-01 3.02616417e-01
-1.21192575e-01 -1.85573474e-01 7.72147700e-02 3.74326438e-01
2.22932428e-01 -2.13388979e-01 3.47699612e-01 -3.83203417e-01
-7.04687953e-01 -4.18712348e-02 -3.32379669e-01 -7.36277044e-01
1.10708404e+00 -8.96586061e-01 3.48724812e-01 -1.48560241e-01
-1.12756419e+00 1.49326712e-01 9.51352537e-01 2.56530643e-01
8.70568037e-01 -1.09425747e+00 7.36864135e-02 3.13068509e-01
7.83375874e-02 6.36953294e-01 2.03918014e-02 4.63005424e-01
-7.22754598e-01 3.82207148e-02 -1.28438354e-01 -9.13092494e-01
-1.34261572e+00 5.76320529e-01 3.51581693e-01 3.97251666e-01
-9.75306153e-01 5.05688667e-01 7.79144540e-02 -6.83963120e-01
3.74706507e-01 -6.99686646e-01 -3.79077405e-01 -1.43328786e-01
3.23146015e-01 6.96804345e-01 1.97210759e-01 -4.12053317e-01
-2.39340499e-01 7.99849153e-01 7.24957883e-01 -6.26330078e-01
1.14442694e+00 -3.29828978e-01 1.32107809e-01 9.28427577e-01
7.30612874e-01 2.93492824e-01 -1.38575792e+00 1.36236707e-02
5.10745123e-02 -8.96360040e-01 -1.12151206e-01 -8.49598408e-01
-5.21743178e-01 8.68587971e-01 8.60265851e-01 4.74383205e-01
8.43627930e-01 1.93286479e-01 4.59437877e-01 1.76975563e-01
1.07608306e+00 -8.80625725e-01 -1.00998253e-01 2.89405793e-01
9.83096898e-01 -9.04772580e-01 -9.61989462e-02 -1.14946651e+00
-3.92055541e-01 1.20305955e+00 5.27974784e-01 -2.96274889e-02
6.86827660e-01 6.06836200e-01 3.39987755e-01 -6.06927931e-01
1.23080730e-01 -2.59773284e-01 2.66773820e-01 7.83015132e-01
2.03535140e-01 2.98650861e-01 -7.39597529e-02 4.00465220e-01
-6.15811050e-01 1.89444646e-02 4.25674796e-01 1.50887418e+00
-5.73867738e-01 -8.91614795e-01 -6.86887264e-01 1.51358014e-02
-7.05012754e-02 4.39833820e-01 -4.21728075e-01 7.80931771e-01
1.57923505e-01 8.64368379e-01 2.96348572e-01 -2.97291666e-01
9.92694855e-01 -1.31517872e-01 6.44293368e-01 -9.95758593e-01
-6.62233174e-01 1.23354740e-01 -7.09673204e-03 -9.66295660e-01
-4.02399778e-01 -5.18337250e-01 -1.72866595e+00 2.96385556e-01
-2.15820923e-01 6.19225912e-02 1.21229553e+00 6.44730210e-01
2.59704560e-01 4.04937148e-01 3.76490921e-01 -1.69679737e+00
2.14512244e-01 -7.13518918e-01 -5.86353779e-01 3.74385327e-01
1.75341338e-01 -8.15472782e-01 -1.72897115e-01 3.31143647e-01] | [8.4495210647583, -2.7203011512756348] |
e460af21-12c5-47ef-bd6a-d9688aaea137 | computational-asymmetries-in-robust | 2306.14326 | null | https://arxiv.org/abs/2306.14326v1 | https://arxiv.org/pdf/2306.14326v1.pdf | Computational Asymmetries in Robust Classification | In the context of adversarial robustness, we make three strongly related contributions. First, we prove that while attacking ReLU classifiers is $\mathit{NP}$-hard, ensuring their robustness at training time is $\Sigma^2_P$-hard (even on a single example). This asymmetry provides a rationale for the fact that robust classifications approaches are frequently fooled in the literature. Second, we show that inference-time robustness certificates are not affected by this asymmetry, by introducing a proof-of-concept approach named Counter-Attack (CA). Indeed, CA displays a reversed asymmetry: running the defense is $\mathit{NP}$-hard, while attacking it is $\Sigma_2^P$-hard. Finally, motivated by our previous result, we argue that adversarial attacks can be used in the context of robustness certification, and provide an empirical evaluation of their effectiveness. As a byproduct of this process, we also release UG100, a benchmark dataset for adversarial attacks. | ['Michele Lombardi', 'Samuele Marro'] | 2023-06-25 | null | null | null | null | ['adversarial-robustness', 'classification-1'] | ['adversarial', 'methodology'] | [ 3.98155779e-01 5.12977660e-01 9.55207497e-02 -9.17287245e-02
-1.06905425e+00 -1.46593904e+00 3.82035911e-01 2.40090519e-01
-1.93092495e-01 9.39213634e-01 -3.37642908e-01 -9.09509480e-01
-4.54865098e-01 -1.27253008e+00 -1.27053177e+00 -6.87836885e-01
-4.23478991e-01 -8.11455622e-02 3.95849496e-01 -5.72573185e-01
3.19671422e-01 5.86715817e-01 -1.45316398e+00 -1.00262970e-01
5.78095436e-01 1.00551796e+00 -1.08503366e+00 8.34252894e-01
5.84124863e-01 6.42358959e-01 -8.08406353e-01 -1.07170999e+00
7.31953561e-01 -1.54106244e-01 -1.16211271e+00 -3.97562832e-01
6.49291992e-01 -8.25679824e-02 -4.12099689e-01 1.40885317e+00
4.42952901e-01 8.94588381e-02 4.51259017e-01 -1.88072991e+00
-2.29166448e-01 1.06532049e+00 6.78645149e-02 4.03120182e-02
4.22804922e-01 3.61568570e-01 1.13269484e+00 -2.18540832e-01
3.60279411e-01 1.06848347e+00 6.44651771e-01 7.50108123e-01
-1.08328009e+00 -7.28517830e-01 1.93922341e-01 -1.53534412e-01
-1.29119813e+00 -2.58683294e-01 4.82524812e-01 -1.75659820e-01
6.58284187e-01 8.23153436e-01 -1.06587805e-01 1.13374412e+00
1.32715449e-01 3.92282367e-01 1.36037016e+00 -3.56840461e-01
2.27873757e-01 -2.72188336e-03 2.81654388e-01 5.75749576e-01
5.68557441e-01 3.64258021e-01 5.38213085e-03 -4.83603448e-01
1.64067566e-01 -3.45455348e-01 -5.26293099e-01 -7.65210688e-02
-7.14230955e-01 8.19408655e-01 3.01363289e-01 2.39170089e-01
4.74556953e-01 3.00173134e-01 6.75712347e-01 8.26423883e-01
9.19618756e-02 7.81415701e-01 -5.64559221e-01 -7.28630871e-02
-4.80000019e-01 5.35130441e-01 1.00182188e+00 7.93353021e-01
3.41569990e-01 1.26630411e-01 -1.67729743e-02 3.44064087e-02
-7.25559071e-02 5.82412004e-01 5.91098852e-02 -9.77698803e-01
4.86807525e-01 4.03397679e-01 1.09385394e-01 -7.65650094e-01
-1.75125152e-01 -3.28325301e-01 -6.96263194e-01 5.33713162e-01
6.50766134e-01 -2.62421787e-01 -4.38310295e-01 2.20371604e+00
1.09354280e-01 -2.74005700e-02 4.00311828e-01 4.76676524e-01
1.63460940e-01 2.70321488e-01 2.06210688e-02 -2.07783133e-01
1.32002246e+00 -5.83130300e-01 -1.19465791e-01 8.31727087e-02
6.50494933e-01 -3.15021664e-01 9.95960057e-01 7.31334090e-01
-1.28831983e+00 -2.66274035e-01 -1.39133835e+00 2.62213677e-01
-6.48245573e-01 -7.69508481e-01 5.67393780e-01 1.27122116e+00
-8.27344656e-01 6.94951117e-01 -3.51021498e-01 7.93673992e-02
2.44730905e-01 4.87638682e-01 -6.51787758e-01 -3.86570543e-02
-1.49619627e+00 7.21934974e-01 2.87515908e-01 -3.56317461e-01
-1.11628842e+00 -6.39764667e-01 -7.78888285e-01 -5.65866232e-02
5.63737512e-01 -3.31608564e-01 1.18224335e+00 -5.13037562e-01
-1.23814905e+00 8.14067900e-01 3.27966601e-01 -8.95119846e-01
1.03114045e+00 1.38516963e-01 -6.94494069e-01 2.64358044e-01
-1.87263384e-01 -5.59275225e-03 8.86729896e-01 -1.22278929e+00
-6.89651787e-01 -4.62325871e-01 1.10968959e+00 -4.08825397e-01
-2.99338579e-01 1.59149423e-01 2.76478857e-01 -6.78941488e-01
-2.07763910e-01 -1.06793475e+00 -3.21590543e-01 -3.19046021e-01
-6.04258239e-01 -1.16745047e-01 7.40290761e-01 -2.19310403e-01
1.24267924e+00 -2.17189932e+00 2.65115616e-03 7.87274778e-01
1.96860611e-01 3.71752083e-01 2.99867522e-02 1.88250795e-01
-4.05603915e-01 8.98877680e-01 -6.50046349e-01 -1.24394055e-02
4.92274076e-01 1.12898186e-01 -9.57368016e-01 5.92287421e-01
2.34768897e-01 7.14278877e-01 -7.47276723e-01 -2.09481716e-01
-1.05218373e-01 6.76293597e-02 -5.10749936e-01 2.31346637e-02
-3.16717416e-01 1.57790035e-01 -3.27493906e-01 7.72621274e-01
8.75176847e-01 4.82765175e-02 2.41615817e-01 4.11486812e-02
2.89517373e-01 2.34150082e-01 -1.39238989e+00 1.03995430e+00
-3.30426902e-01 2.00899541e-01 1.43999428e-01 -9.65154886e-01
4.71744508e-01 4.01069671e-01 3.48985419e-02 -4.42386448e-01
3.34017664e-01 2.34288871e-01 4.85568717e-02 5.98994419e-02
4.80782062e-01 -2.23466784e-01 -7.43058085e-01 7.03255892e-01
-1.20708734e-01 -3.22838187e-01 2.05022871e-01 3.42380702e-01
1.58640957e+00 -1.26015201e-01 -1.32436320e-01 -2.59271652e-01
8.91089439e-01 -2.35191599e-01 5.77823758e-01 1.02684176e+00
-2.25238308e-01 3.29694659e-01 9.16460574e-01 -2.17525378e-01
-7.80615389e-01 -1.25081289e+00 -2.71693528e-01 9.15738106e-01
-3.18358950e-02 -6.03677928e-01 -1.00058222e+00 -1.48066032e+00
8.27153176e-02 5.95802009e-01 -7.18988001e-01 -4.32991236e-01
-5.38196027e-01 -6.25578701e-01 1.45474327e+00 6.15737319e-01
4.08834368e-01 -6.91832423e-01 -2.53056258e-01 -1.69039488e-01
-7.61524364e-02 -1.06750023e+00 -1.28136994e-02 4.08669382e-01
-4.76790786e-01 -1.49944675e+00 -1.27022162e-01 -2.74427503e-01
8.58129144e-01 -2.65730079e-02 1.13015294e+00 5.48989594e-01
-6.67942241e-02 4.34050769e-01 -3.33986849e-01 -5.15541732e-01
-8.04192245e-01 1.04471557e-01 2.83729941e-01 -5.17418206e-01
-1.83672830e-01 -7.71310806e-01 -3.83893132e-01 5.43892562e-01
-1.27881026e+00 -9.17723536e-01 6.32392615e-02 4.45649534e-01
4.07398283e-01 5.14851391e-01 7.45438457e-01 -1.12493587e+00
5.43767154e-01 -4.74881738e-01 -1.05444074e+00 3.10484558e-01
-5.09802580e-01 -8.08738638e-03 1.33628774e+00 -3.36963832e-01
-2.21266046e-01 -3.05866003e-01 -3.92624378e-01 -1.75271869e-01
-1.57784432e-01 8.63942653e-02 -4.53210860e-01 -5.14486134e-01
1.10280740e+00 6.67442679e-02 -2.69918650e-01 4.09452878e-02
4.59983885e-01 2.64368266e-01 7.91309059e-01 -1.18510020e+00
1.58482122e+00 3.63389939e-01 4.79532003e-01 -1.09891482e-01
-9.64937866e-01 3.83096248e-01 -5.59860915e-02 1.00029208e-01
2.71185070e-01 -5.75549066e-01 -1.42194235e+00 3.60341370e-01
-5.50629258e-01 -5.55187881e-01 -4.11720395e-01 -1.23133995e-01
-5.87802470e-01 7.22422659e-01 -5.59338093e-01 -9.54230487e-01
-3.15583110e-01 -1.29216349e+00 5.16373634e-01 -1.06640728e-02
5.28369546e-02 -7.67420828e-01 -3.91096741e-01 3.26516330e-01
1.84418514e-01 6.80669785e-01 9.81991351e-01 -1.13678420e+00
-3.36016446e-01 -7.30205476e-01 -1.32934257e-01 8.26272845e-01
-2.08500698e-01 1.57629818e-01 -1.20172727e+00 -4.43903714e-01
1.35147572e-01 -6.25325561e-01 5.14658511e-01 -3.71279150e-01
1.57266426e+00 -6.94266140e-01 1.29066914e-01 3.57142180e-01
1.45761192e+00 -3.30283761e-01 7.90238976e-01 3.91451359e-01
2.92873234e-01 3.36590052e-01 5.64573467e-01 2.23624125e-01
1.66594565e-01 4.74546134e-01 8.36081624e-01 3.62083048e-01
3.38435531e-01 -1.89129502e-01 4.04681444e-01 -1.72623143e-01
-9.73533839e-02 -1.63226008e-01 -6.72182322e-01 2.19848514e-01
-1.42813909e+00 -9.92708564e-01 -1.48021774e-02 2.48718548e+00
1.02953649e+00 8.19325089e-01 3.47427011e-01 9.29143727e-01
3.81572247e-01 -2.30598450e-02 -8.32390189e-02 -8.18928778e-01
-3.41197431e-01 7.26832032e-01 7.67597497e-01 6.51652396e-01
-1.04380357e+00 7.70464301e-01 6.06594992e+00 9.69144285e-01
-9.52125549e-01 -1.06059322e-02 5.93064547e-01 3.08410227e-01
-3.46955597e-01 8.25866088e-02 -5.44810414e-01 6.10155165e-01
1.05990112e+00 -3.14875126e-01 5.03920496e-01 1.06714058e+00
-6.20468915e-01 1.41443700e-01 -1.37925363e+00 1.86542884e-01
-1.03419818e-01 -1.12604356e+00 -2.19506145e-01 1.14790678e-01
5.41505098e-01 -5.65096140e-01 4.59866315e-01 5.90059221e-01
8.22594464e-01 -1.24539423e+00 8.83723319e-01 -1.41975433e-01
8.14361572e-01 -1.24597740e+00 7.32011735e-01 2.16045275e-01
-1.03386986e+00 -3.76910001e-01 -2.31886983e-01 -8.51749852e-02
-2.37760484e-01 4.86497641e-01 -4.67454165e-01 1.06202626e+00
4.94184196e-01 -1.36151284e-01 -6.11401260e-01 5.82567394e-01
-7.38821149e-01 6.93356514e-01 -2.73364156e-01 5.12791097e-01
6.78761005e-02 4.14251089e-01 5.98413229e-01 1.10996759e+00
1.41199436e-02 1.33670732e-01 9.88786370e-02 6.08742416e-01
-4.77306396e-01 -4.88159209e-01 -6.12660229e-01 2.21031129e-01
5.61093509e-01 1.14471042e+00 -5.26343644e-01 -1.33469567e-01
1.77582890e-01 6.51497841e-01 6.49369359e-02 -5.19569665e-02
-1.23232007e+00 -6.04507744e-01 6.61945045e-01 -1.55646697e-01
1.57051250e-01 -7.40510076e-02 -3.19255084e-01 -1.00623333e+00
1.96244314e-01 -1.39099348e+00 6.64266229e-01 -2.49476105e-01
-1.46652389e+00 4.43846166e-01 -4.03890200e-02 -1.24758244e+00
-1.28874585e-01 -7.88622141e-01 -6.51192367e-01 8.37791920e-01
-1.55529690e+00 -1.05128288e+00 1.20173834e-01 9.19974506e-01
-1.73076987e-01 8.64692554e-02 1.18412685e+00 6.66570365e-02
-5.76786637e-01 1.40725279e+00 -4.74194109e-01 2.27290183e-01
5.76639056e-01 -1.45251203e+00 5.63293517e-01 1.30933642e+00
-1.52143657e-01 7.75331080e-01 8.39146495e-01 -2.90837795e-01
-1.55728519e+00 -1.18838835e+00 4.86138880e-01 -8.29459369e-01
9.37587261e-01 -3.74143451e-01 -6.66228533e-01 5.97994030e-01
-3.26818675e-01 4.32080895e-01 6.56680405e-01 -5.57624698e-02
-1.14480698e+00 -6.98446184e-02 -1.81014645e+00 7.94744313e-01
1.06087756e+00 -8.10343862e-01 -5.43388605e-01 3.86900067e-01
1.01841152e+00 -4.38923359e-01 -1.07100976e+00 6.60441458e-01
4.37370658e-01 -9.17861223e-01 9.90319431e-01 -7.31323242e-01
3.26458126e-01 -3.59585941e-01 -6.22161150e-01 -9.45316911e-01
9.00546610e-02 -9.74760532e-01 -3.31225514e-01 1.26840150e+00
6.10296667e-01 -9.71196353e-01 4.37781334e-01 6.55937552e-01
-1.00270547e-01 -6.43046439e-01 -1.17683756e+00 -9.81768906e-01
7.61450410e-01 -8.17932963e-01 6.67403698e-01 9.43779349e-01
1.71782240e-01 -3.03502947e-01 3.50201391e-02 6.40785933e-01
5.11248350e-01 -3.34782124e-01 9.57798481e-01 -1.10773993e+00
-5.63535333e-01 -4.37935352e-01 -5.60909748e-01 -1.94506228e-01
4.84762400e-01 -7.50451922e-01 -5.18968627e-02 -6.70143783e-01
-3.47295672e-01 -7.34613836e-01 -4.78068620e-01 8.62561166e-01
-1.28050268e-01 4.91132438e-01 2.91668445e-01 -1.86861098e-01
-4.31654602e-01 -2.52245903e-01 7.14592576e-01 -2.01966614e-01
4.67109859e-01 2.97323644e-01 -1.16877651e+00 6.90645993e-01
9.35762644e-01 -3.17680299e-01 -2.82703787e-01 1.50684923e-01
7.98472047e-01 -1.34761751e-01 6.28011763e-01 -1.22187901e+00
-8.79172012e-02 -7.15291277e-02 -2.59169877e-01 5.41603602e-02
4.98503223e-02 -8.45486641e-01 -2.44908303e-01 6.36181295e-01
-2.93538034e-01 2.80233264e-01 2.93712556e-01 3.74281973e-01
1.12234689e-01 -2.62808055e-01 8.52879763e-01 2.26212777e-02
-1.31268352e-01 2.34797761e-01 -1.13186553e-01 4.49188381e-01
1.16594899e+00 -3.63962688e-02 -9.05566335e-01 -2.00761765e-01
-3.98499817e-01 9.32494029e-02 7.58524537e-01 3.07446066e-02
2.58000433e-01 -8.22232664e-01 -5.49812615e-01 7.70554319e-02
1.15117908e-01 7.85930306e-02 2.70610750e-01 6.23586416e-01
-2.27407277e-01 -1.41239213e-02 -1.47543415e-01 6.24828711e-02
-1.20637465e+00 8.85212243e-01 4.82978076e-01 -3.79675955e-01
-2.57859677e-01 9.91360307e-01 -1.58505142e-01 -4.92466807e-01
5.56101382e-01 -2.28314549e-01 3.82837027e-01 -3.10625404e-01
5.73068917e-01 4.80592668e-01 3.72345686e-01 -1.77609816e-01
-4.31892395e-01 2.68201619e-01 1.33701622e-01 -3.68875384e-01
9.99683499e-01 3.17213386e-01 -2.57179290e-01 -9.59359854e-02
9.21518505e-01 5.20784795e-01 -8.57110262e-01 2.28465334e-01
-1.74740762e-01 -2.68833101e-01 -5.90167880e-01 -1.00268793e+00
-1.00680053e+00 8.14311206e-01 4.19766039e-01 8.74490917e-01
1.19764185e+00 -3.07532340e-01 5.40687680e-01 5.45063436e-01
5.72745979e-01 -1.01531637e+00 -1.56296477e-01 5.47885835e-01
7.05224991e-01 -8.90749335e-01 -4.48990948e-02 -6.07720017e-01
-3.31512809e-01 8.52107525e-01 5.65056264e-01 -2.82321811e-01
4.04914498e-01 4.05535311e-01 -3.51792365e-01 3.18448991e-01
-6.18785799e-01 2.20071629e-01 -1.32162467e-01 6.14926875e-01
4.46278341e-02 5.11342399e-02 -1.94429636e-01 8.61667573e-01
-5.44304192e-01 -3.13951939e-01 8.58801246e-01 1.08078456e+00
-8.06301832e-02 -1.59192526e+00 -4.89774197e-01 -6.36231005e-02
-1.15174305e+00 -1.21566188e-02 -3.58888507e-01 8.55089843e-01
1.86752245e-01 1.20501447e+00 -5.77591062e-01 -5.80648065e-01
2.54346222e-01 8.27086195e-02 4.58887130e-01 -3.42563391e-01
-9.75019813e-01 -7.19503224e-01 2.59522974e-01 -4.53190565e-01
-3.13190650e-03 -9.35224071e-02 -1.09173250e+00 -7.19817817e-01
-1.44922838e-01 4.44864064e-01 6.69042170e-01 8.44328403e-01
1.81979891e-02 4.64594573e-01 1.22689593e+00 -3.62038434e-01
-1.15454030e+00 -5.28466046e-01 -4.19804275e-01 2.82102734e-01
2.41441965e-01 -3.17663342e-01 -8.82881641e-01 -2.14239955e-01] | [5.818338871002197, 7.670707702636719] |
025f91d0-190e-4414-8a4b-61ea126732d7 | global-to-local-expression-aware-embeddings | 2210.15160 | null | https://arxiv.org/abs/2210.15160v2 | https://arxiv.org/pdf/2210.15160v2.pdf | Global-to-local Expression-aware Embeddings for Facial Action Unit Detection | Expressions and facial action units (AUs) are two levels of facial behavior descriptors. Expression auxiliary information has been widely used to improve the AU detection performance. However, most existing expression representations can only describe pre-determined discrete categories (e.g., Angry, Disgust, Happy, Sad, etc.) and cannot capture subtle expression transformations like AUs. In this paper, we propose a novel fine-grained \textsl{Global Expression representation Encoder} to capture subtle and continuous facial movements, to promote AU detection. To obtain such a global expression representation, we propose to train an expression embedding model on a large-scale expression dataset according to global expression similarity. Moreover, considering the local definition of AUs, it is essential to extract local AU features. Therefore, we design a \textsl{Local AU Features Module} to generate local facial features for each AU. Specifically, it consists of an AU feature map extractor and a corresponding AU mask extractor. First, the two extractors transform the global expression representation into AU feature maps and masks, respectively. Then, AU feature maps and their corresponding AU masks are multiplied to generate AU masked features focusing on local facial region. Finally, the AU masked features are fed into an AU classifier for judging the AU occurrence. Extensive experiment results demonstrate the superiority of our proposed method. Our method validly outperforms previous works and achieves state-of-the-art performances on widely-used face datasets, including BP4D, DISFA, and BP4D+. | ['Yu Ding', 'Zhigang Deng', 'Wei Chen', 'Hao Zeng', 'Wei zhang', 'Rudong An'] | 2022-10-27 | null | null | null | null | ['action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [ 2.77591109e-01 -2.10296303e-01 -1.59252495e-01 -6.72497034e-01
-4.75382000e-01 -1.84823319e-01 5.33652484e-01 -3.10314745e-01
-2.42024541e-01 3.58012229e-01 1.33461460e-01 4.30388212e-01
2.56320804e-01 -7.30058610e-01 -2.39099786e-01 -1.04501152e+00
1.52914494e-01 -3.05455178e-01 -3.18712056e-01 -4.10825014e-01
-1.66702718e-01 6.36045754e-01 -1.59377170e+00 2.72004694e-01
5.03732562e-01 1.60142517e+00 -3.23852152e-01 1.54342696e-01
-5.89474998e-02 9.20484602e-01 -5.92999399e-01 -4.05955970e-01
-3.11634727e-02 -8.55184138e-01 -4.03065294e-01 2.60719866e-01
7.79380696e-03 -5.56516171e-01 -1.72444344e-01 1.20580661e+00
4.89462554e-01 9.46818944e-03 7.98221111e-01 -1.38931286e+00
-7.74195910e-01 -9.53281298e-03 -9.29441869e-01 1.08796448e-01
6.09150112e-01 1.49149299e-01 8.29962790e-01 -9.68277395e-01
5.14873326e-01 1.37414622e+00 3.80721658e-01 7.00733960e-01
-1.03649628e+00 -1.10277522e+00 1.28230870e-01 2.01665521e-01
-1.56020689e+00 -5.23249090e-01 1.13492239e+00 -3.17923993e-01
5.34749329e-01 2.76134789e-01 6.22699738e-01 1.09720647e+00
-1.30368531e-01 9.11667526e-01 1.34463358e+00 -1.05904616e-01
-6.10587485e-02 1.55657545e-01 -1.40838891e-01 8.93886149e-01
-5.29565096e-01 -1.68582708e-01 -3.47909898e-01 -1.76190495e-01
7.58745492e-01 9.07929614e-02 -3.21071476e-01 2.29216412e-01
-6.78432643e-01 7.20429599e-01 5.56407571e-01 4.35169697e-01
-5.60715735e-01 -9.39097442e-03 6.26046360e-01 4.70440120e-01
5.39169908e-01 -1.25701427e-01 -1.31803319e-01 -3.65727097e-01
-5.24257898e-01 1.60193279e-01 3.19798827e-01 6.61905050e-01
1.26604867e+00 1.21724848e-02 -3.82365316e-01 1.16223085e+00
2.22468942e-01 4.57223058e-01 4.13264990e-01 -5.68761170e-01
1.12701148e-01 9.10718083e-01 -1.24618456e-01 -1.60838914e+00
-2.94199944e-01 3.77269927e-03 -9.21338320e-01 2.06659481e-01
-1.36152729e-01 -2.21897960e-01 -6.04370713e-01 2.02666974e+00
3.42740446e-01 2.61907667e-01 1.57353342e-01 1.17386532e+00
9.27190900e-01 8.74346793e-01 2.86353379e-01 -4.38381940e-01
1.44134545e+00 -7.20220268e-01 -9.51136112e-01 3.67484093e-02
8.02356780e-01 -7.08852410e-01 1.06119025e+00 1.99660718e-01
-7.88950920e-01 -7.00249791e-01 -9.51745570e-01 -1.32920942e-03
-1.53304383e-01 5.74514329e-01 5.45829117e-01 5.07298946e-01
-7.37002492e-01 2.76334137e-01 -5.31274021e-01 -5.60856685e-02
5.92597961e-01 4.64237541e-01 -7.60977566e-01 8.97518620e-02
-1.34005821e+00 6.67358041e-01 1.45573199e-01 4.68597323e-01
-7.98537135e-01 -2.23247856e-01 -1.20090508e+00 1.17389560e-02
7.38087669e-02 1.60812348e-01 1.10432410e+00 -1.68983912e+00
-1.79851174e+00 1.10239100e+00 -1.82723895e-01 2.78180838e-01
8.43185633e-02 1.38399273e-01 -7.94217765e-01 2.96154499e-01
-5.03394566e-02 4.91837829e-01 1.00103438e+00 -9.67596233e-01
-3.12304109e-01 -5.64237714e-01 -3.10920794e-02 1.77325517e-01
-7.24867642e-01 5.35152495e-01 -4.04570013e-01 -5.70769668e-01
-1.64100379e-01 -7.26328015e-01 1.23990551e-01 2.46681437e-01
-1.29911497e-01 -5.62629402e-01 1.12448978e+00 -5.89967251e-01
1.31611347e+00 -2.61872077e+00 5.64779639e-02 3.84048522e-01
1.35927588e-01 2.45232895e-01 -3.70886952e-01 3.80341969e-02
-2.11341590e-01 -9.76266116e-02 -1.75656214e-01 -2.73145318e-01
6.70492798e-02 2.55281419e-01 2.87257601e-02 5.05816638e-01
7.85651028e-01 8.74154091e-01 -7.99182415e-01 -5.95893562e-01
1.72339126e-01 5.48287868e-01 -3.49964470e-01 4.16834205e-01
5.95017597e-02 3.73562545e-01 -8.97575676e-01 8.27686906e-01
9.44127619e-01 2.01613307e-01 -7.94405416e-02 -3.42327923e-01
4.70312089e-02 -3.45686078e-01 -8.99382651e-01 1.32183957e+00
-6.70811057e-01 3.32800627e-01 3.22819859e-01 -1.07695270e+00
1.63228011e+00 3.58693153e-01 5.27972519e-01 -7.32307076e-01
8.31027627e-01 2.54245758e-01 -1.71821415e-01 -5.51946819e-01
-8.98625925e-02 -3.62531483e-01 -2.01923415e-01 4.42498177e-01
1.88297391e-01 1.15606323e-01 -1.47748590e-01 -2.85580188e-01
9.24728513e-01 5.93039803e-02 3.89416158e-01 -1.13862902e-01
1.23057783e+00 -5.54116726e-01 6.76243067e-01 -2.22404450e-01
-5.14293134e-01 3.99985194e-01 9.12564456e-01 -5.41454315e-01
-4.31677163e-01 -7.54004002e-01 -1.96429014e-01 1.24248481e+00
2.20695153e-01 -6.03117168e-01 -8.47872555e-01 -8.86208415e-01
-1.14539683e-01 1.16611250e-01 -1.09948969e+00 -5.61842442e-01
-3.50692511e-01 -7.12129891e-01 6.88697278e-01 5.58825374e-01
9.29035485e-01 -1.37176085e+00 -3.19311798e-01 9.63645503e-02
3.85490283e-02 -1.07772732e+00 -4.43023384e-01 -1.85050189e-01
-1.45028442e-01 -8.19130599e-01 -6.80061758e-01 -9.41978276e-01
7.80501008e-01 -5.07731326e-02 5.48824787e-01 1.20943531e-01
-2.42454946e-01 -5.96185103e-02 -5.02893031e-01 -1.73640937e-01
-2.77108341e-01 -5.14058530e-01 8.65660608e-02 8.84405613e-01
8.94846380e-01 -6.28598094e-01 -6.30279660e-01 4.18869764e-01
-8.93041313e-01 -2.35357240e-01 7.99004436e-01 7.68001854e-01
5.46299517e-01 -3.18911165e-01 5.64045548e-01 -4.86220717e-01
9.07089651e-01 -3.90159965e-01 -7.25116506e-02 5.17316423e-02
7.28167547e-03 -1.80879682e-01 7.51640141e-01 -5.74255705e-01
-1.25980628e+00 7.99025521e-02 -5.06548762e-01 -6.56506062e-01
-2.67097116e-01 5.04659832e-01 -4.17135239e-01 -2.40696907e-01
4.67278242e-01 3.54225844e-01 3.58301371e-01 -1.59832254e-01
3.22862208e-01 1.05185616e+00 4.31084186e-01 -6.15647316e-01
5.61242998e-01 4.08696145e-01 -4.82597873e-02 -8.29871714e-01
-5.49474001e-01 -1.48735702e-01 -4.55108345e-01 -3.12916756e-01
1.03411150e+00 -9.95500505e-01 -7.72461414e-01 5.48468053e-01
-1.01829016e+00 1.04337178e-01 1.24903880e-01 1.65211633e-01
-5.61232507e-01 2.48786554e-01 -7.35457599e-01 -7.59343445e-01
-4.19622123e-01 -1.21422184e+00 1.37266684e+00 5.00219822e-01
-3.20989013e-01 -4.82144445e-01 -1.62825994e-02 2.73331515e-02
2.62143224e-01 6.82231009e-01 6.56490266e-01 -2.04752058e-01
1.28262028e-01 -3.88165087e-01 -4.99916255e-01 8.01782191e-01
5.48023582e-01 2.86980599e-01 -1.04290736e+00 7.59440893e-03
1.04843378e-02 -7.91210651e-01 5.65482140e-01 -2.11260974e-01
1.43727326e+00 -4.37777430e-01 -1.73444659e-01 6.38609827e-01
9.94640768e-01 3.62225950e-01 6.51647151e-01 -8.54577646e-02
4.74719197e-01 5.97156882e-01 8.13095808e-01 7.83336282e-01
-1.70376655e-02 7.75929153e-01 2.80790716e-01 -3.18325043e-01
2.25272864e-01 -1.10843994e-01 7.37088978e-01 6.32993340e-01
-3.66371930e-01 3.22291344e-01 -3.48155737e-01 1.42640054e-01
-1.70328796e+00 -9.33021903e-01 3.62191170e-01 1.65682554e+00
9.74161565e-01 -1.60474256e-01 -1.15754120e-02 1.00623327e-03
5.27583718e-01 4.98185694e-01 -3.52961183e-01 -7.66579926e-01
-1.78718209e-01 5.32623112e-01 -3.48633975e-01 1.00020401e-01
-1.11227953e+00 1.04469526e+00 4.63415480e+00 1.12555456e+00
-1.54652715e+00 1.56877980e-01 8.36867034e-01 6.58768490e-02
-1.39225349e-01 -4.50701654e-01 -4.24077809e-01 4.64789122e-01
4.52973485e-01 -8.21484029e-02 5.78843281e-02 1.15658271e+00
2.35456645e-01 3.25091749e-01 -8.24408114e-01 1.35612655e+00
3.05813491e-01 -7.55606651e-01 1.98070005e-01 -2.14609936e-01
5.22642672e-01 -5.89953780e-01 1.46522373e-01 5.91764271e-01
-1.90972522e-01 -1.18302226e+00 3.06037545e-01 3.36548060e-01
1.13747168e+00 -1.06016648e+00 8.50005627e-01 -5.86737059e-02
-1.40690053e+00 -3.75877544e-02 -4.11677748e-01 -2.72002816e-01
-1.59711950e-02 6.25609979e-02 -1.60943970e-01 4.97656107e-01
6.07237697e-01 9.90952790e-01 -3.98827374e-01 7.09468424e-02
-4.71686453e-01 3.47448587e-01 -3.69527519e-01 -3.03294450e-01
4.82157439e-01 -4.94019449e-01 2.81399842e-02 1.27174032e+00
2.71043450e-01 4.58179474e-01 -2.02370313e-04 8.44337225e-01
-1.59977093e-01 6.19044840e-01 -5.62555969e-01 -1.11834876e-01
1.49467234e-02 1.77022016e+00 -5.63060753e-02 -2.48921916e-01
-6.50010586e-01 1.33995378e+00 4.23079073e-01 2.46324331e-01
-9.78991389e-01 -5.09571731e-01 1.13879883e+00 -2.11971492e-01
-2.92276051e-02 1.89254016e-01 3.97911012e-01 -1.14471531e+00
3.13081816e-02 -1.01069283e+00 2.93241292e-01 -7.02234507e-01
-1.18979311e+00 8.25116634e-01 -2.97091693e-01 -1.28111827e+00
-1.65638566e-01 -7.65625715e-01 -9.59662855e-01 9.08643007e-01
-1.31996441e+00 -1.39799440e+00 -8.44686985e-01 9.89603996e-01
1.37373030e-01 -3.36577632e-02 1.01421165e+00 4.18699592e-01
-7.92276621e-01 8.92335296e-01 -4.41962481e-01 5.28687537e-01
5.38462281e-01 -5.76790214e-01 -3.03115636e-01 4.56569642e-01
-7.89253265e-02 5.83508611e-01 2.96641648e-01 -9.92866233e-02
-1.19489086e+00 -1.04357851e+00 4.97683436e-01 6.15614504e-02
5.54319203e-01 -4.35041010e-01 -8.84413719e-01 5.95332205e-01
-4.18697409e-02 4.89520907e-01 7.79109776e-01 -2.06496656e-01
-3.74754637e-01 -5.03556252e-01 -1.16482401e+00 6.23441756e-01
1.01473439e+00 -7.63964534e-01 -3.80573720e-01 1.12787731e-01
2.50618190e-01 -1.57058358e-01 -1.04235172e+00 6.70226693e-01
7.12283015e-01 -1.00717783e+00 5.37283599e-01 -5.87961853e-01
6.40233576e-01 -2.71068841e-01 -7.74211213e-02 -1.10282433e+00
-2.28787765e-01 -5.35419285e-01 1.87218711e-01 1.40495765e+00
8.13541096e-03 -5.89704871e-01 6.98372185e-01 2.73027837e-01
9.59345102e-02 -1.08229864e+00 -9.93667543e-01 -5.00674188e-01
-2.38748148e-01 -1.73616558e-01 8.23215306e-01 1.06031120e+00
2.20523149e-01 3.09441000e-01 -3.99101466e-01 -1.90050289e-01
1.34131119e-01 2.62451679e-01 9.22557056e-01 -8.70292783e-01
8.56690407e-02 -7.08334386e-01 -8.78528953e-01 -9.70785677e-01
6.17408812e-01 -7.58975744e-01 -7.23704025e-02 -8.88805807e-01
1.72277033e-01 -3.28442782e-01 -6.31737292e-01 7.48155355e-01
-2.90831149e-01 6.33280396e-01 -5.01933433e-02 -1.09744497e-01
-2.56091118e-01 1.14002979e+00 1.31832886e+00 -2.18106642e-01
-1.25314012e-01 -1.93897948e-01 -6.24825954e-01 7.33790398e-01
8.03810835e-01 1.49194766e-02 -2.77591556e-01 -1.57621838e-02
-1.34691551e-01 -1.85940042e-01 2.67881244e-01 -8.63749802e-01
-3.14916313e-01 -2.12159604e-01 5.95492065e-01 -1.07848011e-01
5.96215487e-01 -7.51761496e-01 -2.55692303e-01 3.48362952e-01
-8.53040367e-02 -2.92377882e-02 1.43340722e-01 1.30395249e-01
-8.28855276e-01 6.74545392e-02 8.12424242e-01 1.35664925e-01
-1.07726038e+00 7.30733275e-01 -2.15007946e-01 -3.80807608e-01
1.40429997e+00 -1.12918220e-01 -1.13899626e-01 -4.50972617e-01
-6.48355305e-01 1.37002587e-01 2.42577285e-01 6.30910814e-01
7.78311849e-01 -1.80759108e+00 -7.53638923e-01 5.89875638e-01
5.91426730e-01 -2.38233402e-01 3.79449159e-01 9.83978748e-01
-3.74913424e-01 -1.76243633e-01 -6.27861738e-01 -4.57893491e-01
-1.55349004e+00 4.28827465e-01 4.08514708e-01 1.48363084e-01
-2.47963622e-01 9.69630063e-01 4.66750950e-01 -3.00398409e-01
-2.62501836e-01 -1.07914492e-01 -4.50685859e-01 2.44125620e-01
7.86016285e-01 -1.16235875e-01 -3.12576503e-01 -1.26667094e+00
-4.28464025e-01 9.59285080e-01 7.28892386e-02 2.86006600e-01
1.02487648e+00 1.52329102e-01 -5.33460319e-01 7.26660490e-02
1.79887986e+00 7.21002445e-02 -1.09868431e+00 -2.90489554e-01
-3.77437621e-01 -4.49191391e-01 -1.50486663e-01 -2.99114287e-01
-1.23353553e+00 1.06895173e+00 4.93865669e-01 -1.87889487e-01
1.64696121e+00 8.85943100e-02 7.13639617e-01 2.07708608e-02
1.35576680e-01 -8.67417991e-01 2.86208659e-01 9.30789784e-02
1.05589092e+00 -1.06521547e+00 -2.84978449e-01 -4.35263127e-01
-8.15260410e-01 1.18771017e+00 1.08490920e+00 -1.41022772e-01
6.02874935e-01 2.25599915e-01 4.24364358e-01 -3.13720763e-01
-3.66287678e-01 -1.64931014e-01 2.33419567e-01 3.51104528e-01
3.30110282e-01 2.25398154e-03 -4.24642920e-01 9.36813414e-01
-2.44221881e-01 1.63532227e-01 -9.71324667e-02 7.77240396e-01
-4.00435716e-01 -8.85993659e-01 -2.92531289e-02 2.88332194e-01
-4.30181116e-01 2.63632774e-01 -5.99126339e-01 8.13473821e-01
4.45022076e-01 7.23618090e-01 1.89508006e-01 -7.77752280e-01
3.97401839e-01 8.05065334e-02 4.55891430e-01 -3.63313645e-01
-4.22791868e-01 -4.66281399e-02 -5.83159141e-02 -7.74255216e-01
-4.87698168e-01 -4.87646639e-01 -1.28200054e+00 -2.45029688e-01
-5.23081720e-02 8.19029361e-02 1.22718342e-01 8.00128341e-01
2.76535481e-01 1.62847057e-01 1.08183002e+00 -7.17352092e-01
-4.18800086e-01 -1.18451226e+00 -6.42418683e-01 1.01168621e+00
2.44328275e-01 -8.72676015e-01 -2.73908257e-01 -7.42117614e-02] | [13.634353637695312, 1.6669858694076538] |
5ac34277-8b67-4d90-8671-9c2574195de8 | sign-segmentation-with-changepoint-modulated | 2104.13817 | null | https://arxiv.org/abs/2104.13817v1 | https://arxiv.org/pdf/2104.13817v1.pdf | Sign Segmentation with Changepoint-Modulated Pseudo-Labelling | The objective of this work is to find temporal boundaries between signs in continuous sign language. Motivated by the paucity of annotation available for this task, we propose a simple yet effective algorithm to improve segmentation performance on unlabelled signing footage from a domain of interest. We make the following contributions: (1) We motivate and introduce the task of source-free domain adaptation for sign language segmentation, in which labelled source data is available for an initial training phase, but is not available during adaptation. (2) We propose the Changepoint-Modulated Pseudo-Labelling (CMPL) algorithm to leverage cues from abrupt changes in motion-sensitive feature space to improve pseudo-labelling quality for adaptation. (3) We showcase the effectiveness of our approach for category-agnostic sign segmentation, transferring from the BSLCORPUS to the BSL-1K and RWTH-PHOENIX-Weather 2014 datasets, where we outperform the prior state of the art. | ['Samuel Albanie', 'Gül Varol', 'Neil Fox', 'Nicolaj C. Stache', 'Katrin Renz'] | 2021-04-28 | null | null | null | null | ['source-free-domain-adaptation'] | ['computer-vision'] | [ 6.55192733e-01 -2.06900224e-01 -2.20060840e-01 -5.04979432e-01
-1.04578352e+00 -9.37916458e-01 7.40726948e-01 -7.25459933e-01
-8.84321630e-01 7.90319920e-01 5.10043502e-01 -2.18182504e-01
-7.37021416e-02 2.13053431e-02 -5.15297771e-01 -5.41327477e-01
-1.63657535e-02 4.52406108e-01 7.82891452e-01 -1.39459074e-01
2.43074596e-01 3.87238055e-01 -1.67303312e+00 4.06149566e-01
1.09609199e+00 6.42137825e-01 -6.11230778e-03 1.00674057e+00
-1.50942564e-01 6.56708241e-01 -4.76615936e-01 -3.41631509e-02
5.23130834e-01 -8.72805059e-01 -1.09454250e+00 2.84251392e-01
9.42356706e-01 -3.66072029e-01 -1.82096392e-01 5.46312213e-01
8.62880290e-01 1.86383665e-01 9.29014385e-01 -1.24520040e+00
-1.86218649e-01 2.09183425e-01 -3.01935136e-01 1.62911072e-01
5.15387692e-02 5.91379464e-01 8.05703402e-01 -5.16425669e-01
1.43013108e+00 9.64329660e-01 6.29561245e-01 1.04941750e+00
-7.68366575e-01 -5.05455315e-01 5.62040150e-01 4.29228634e-01
-7.91533768e-01 -4.77386504e-01 4.58799273e-01 -5.36888957e-01
5.74978352e-01 5.92512228e-02 9.84579623e-01 1.29706347e+00
-7.01958001e-01 1.52803314e+00 1.59334683e+00 -6.63056135e-01
3.20360690e-01 -4.76811886e-01 1.68010339e-01 4.65239495e-01
-8.56458247e-02 2.67173976e-01 -7.08496094e-01 1.15510069e-01
6.59584343e-01 -7.21062601e-01 -2.86311537e-01 -5.32213569e-01
-1.29462302e+00 4.66680586e-01 2.52705336e-01 2.47096270e-01
-3.11216354e-01 3.41809541e-01 4.56400514e-01 2.16037184e-01
8.76267850e-02 2.27216855e-01 -8.35505307e-01 -6.51048183e-01
-1.24947262e+00 2.60915518e-01 7.26960301e-01 9.19555843e-01
1.26084015e-01 -1.35199470e-03 -3.60206842e-01 8.08384657e-01
3.43996733e-01 7.44193494e-01 5.80065846e-01 -1.15484333e+00
3.69586736e-01 1.49277464e-01 -8.94881878e-03 1.70399129e-01
-3.45175982e-01 -6.02833591e-02 -4.57863882e-02 5.43153286e-01
1.25353348e+00 -3.10722500e-01 -1.97840071e+00 1.67102814e+00
3.74533117e-01 6.28254175e-01 9.91208106e-02 1.12143385e+00
8.39258432e-01 -1.02915484e-02 4.42804635e-01 2.30474308e-01
1.04526317e+00 -1.03047967e+00 -3.74275625e-01 -1.74423546e-01
5.63342154e-01 -6.53724968e-01 1.09159267e+00 4.14405316e-01
-7.72877097e-01 -2.03888759e-01 -6.24485016e-01 1.17426164e-01
-3.26322526e-01 7.46173188e-02 7.50899017e-01 6.06486857e-01
-8.72240067e-01 2.67108291e-01 -1.10164928e+00 -8.24662805e-01
7.25490928e-01 1.09227665e-01 -2.94340968e-01 -2.63098270e-01
-8.17201138e-01 7.30189681e-01 4.26858604e-01 1.05730742e-01
-7.16054440e-01 -7.90044248e-01 -5.87282419e-01 -9.82785285e-01
2.74694890e-01 -5.26010692e-01 1.50373530e+00 -1.09948945e+00
-1.77298737e+00 1.10304546e+00 -2.79730350e-01 -4.15333688e-01
1.16993606e+00 -2.08173305e-01 -3.29397976e-01 3.91611099e-01
3.75642814e-02 1.08264983e+00 1.07452619e+00 -1.40152013e+00
-1.17916012e+00 4.32017855e-02 -2.79187471e-01 3.32124412e-01
3.39821517e-01 1.61371335e-01 -7.53409028e-01 -6.91760421e-01
2.23213375e-01 -1.22021294e+00 -2.02974603e-01 1.12535946e-01
-2.45197192e-01 4.00399454e-02 8.72539818e-01 -1.23139918e+00
8.35848451e-01 -1.84780991e+00 2.03518361e-01 3.74291360e-01
-2.62982070e-01 6.24532044e-01 -3.70348096e-01 3.29692774e-02
3.32687855e-01 -1.67068243e-01 -9.55899894e-01 -2.47026110e-05
2.51572300e-02 6.23399258e-01 -2.80220449e-01 4.27552491e-01
2.02791348e-01 1.38135648e+00 -1.13190401e+00 -6.65761232e-01
3.36341828e-01 3.71613652e-01 -4.19855744e-01 -2.07919389e-01
-4.58971083e-01 1.26735079e+00 -3.34945917e-01 1.11946428e+00
4.07210827e-01 -3.40894908e-02 -1.43443108e-01 8.24764147e-02
-1.02380902e-01 1.06994689e-01 -1.13613927e+00 2.02165556e+00
-1.36895090e-01 9.25847113e-01 1.30503029e-02 -4.57934111e-01
2.99593806e-01 2.92077810e-01 6.64307952e-01 -5.09625733e-01
1.81739051e-02 6.71760261e-01 1.00193799e-01 -6.68636978e-01
1.63679942e-01 -2.91060610e-03 -1.57744680e-02 3.47808123e-01
1.00132719e-01 -3.56205463e-01 5.23113310e-01 -6.44309744e-02
1.23952365e+00 8.78392041e-01 1.11422837e-01 3.07699114e-01
3.96947235e-01 4.13907170e-01 4.42714483e-01 8.15889180e-01
-8.75242114e-01 8.99680197e-01 8.44464898e-02 1.33572921e-01
-8.59170318e-01 -1.18609786e+00 -2.92328924e-01 1.17417634e+00
-1.73102230e-01 3.23506951e-01 -5.18050909e-01 -1.19875932e+00
1.36024982e-01 4.88227248e-01 -4.90108699e-01 4.44511682e-01
-1.12548077e+00 -5.67967176e-01 9.55988467e-01 8.24020147e-01
7.98958898e-01 -1.32348180e+00 -8.48137975e-01 1.01133905e-01
-3.05715710e-01 -1.11515427e+00 -5.56679606e-01 7.22590312e-02
-8.60893846e-01 -1.22830331e+00 -1.45989633e+00 -8.47655952e-01
5.07665157e-01 -4.01291668e-01 7.65446424e-01 -2.55535334e-01
-3.01657528e-01 8.17184687e-01 -7.00779140e-01 -3.71163905e-01
-5.06464899e-01 1.02771856e-01 -4.15562361e-01 -2.68598367e-02
3.92292827e-01 -2.79542416e-01 -6.75799966e-01 4.66121018e-01
-6.89207196e-01 -2.12027684e-01 7.20444798e-01 7.45907843e-01
5.36736131e-01 -8.59945059e-01 5.99691749e-01 -6.83526456e-01
3.71196032e-01 -2.46647131e-02 -4.37662601e-01 2.88638562e-01
-3.12254071e-01 1.67217553e-01 -7.57680014e-02 -6.24835610e-01
-1.20842016e+00 5.44652641e-01 -1.48294687e-01 -1.17217667e-01
-6.15381479e-01 1.32211164e-01 -1.09936215e-03 -3.96139950e-01
7.95177281e-01 2.67521739e-01 -1.25795543e-01 -6.14800453e-01
9.23995793e-01 8.49647105e-01 1.15151215e+00 -4.62304026e-01
6.88008368e-01 9.24863696e-01 -2.50456389e-03 -7.82360673e-01
-6.61607146e-01 -1.05144405e+00 -1.40006649e+00 -4.01786685e-01
8.30332756e-01 -6.55091822e-01 -1.77970931e-01 9.56996262e-01
-7.60057271e-01 -1.05108905e+00 -7.44699538e-01 5.79586864e-01
-9.98207927e-01 6.24588609e-01 -3.10637355e-01 -6.64581776e-01
6.33829162e-02 -7.36390412e-01 1.16070449e+00 2.39901423e-01
-3.59271646e-01 -9.43928540e-01 3.45776439e-01 4.74709451e-01
1.34537399e-01 6.15979850e-01 3.65978539e-01 -5.47530115e-01
-6.58134341e-01 -7.76354149e-02 -2.64865667e-01 5.64946592e-01
9.90653187e-02 -1.20082952e-01 -1.14761090e+00 -1.23752713e-01
-9.59807694e-01 -4.01792139e-01 1.32473314e+00 7.74653375e-01
5.14456153e-01 1.24759071e-01 -1.52235612e-01 7.94897854e-01
8.96509826e-01 1.30741164e-01 5.72090983e-01 6.81298971e-01
6.13880396e-01 6.62481844e-01 8.98542225e-01 1.52228892e-01
4.01966304e-01 7.05062926e-01 -1.54656395e-01 -9.11898911e-02
-1.07916892e+00 -3.43158185e-01 2.26602599e-01 4.80563432e-01
-5.28258979e-01 -1.37272894e-01 -1.08844292e+00 1.19212329e+00
-2.11502433e+00 -7.84770489e-01 -3.51460457e-01 2.06322432e+00
8.56991947e-01 -3.03656816e-01 5.53512812e-01 2.13476986e-01
4.20987785e-01 5.69569468e-02 -8.59265149e-01 1.35039177e-03
-3.30244482e-01 3.49509478e-01 9.85185802e-01 5.17251611e-01
-1.37565279e+00 1.43342006e+00 6.40015888e+00 5.77589691e-01
-1.08401573e+00 2.43624970e-01 -2.68096179e-01 -1.17936552e-01
2.27883965e-01 5.99554107e-02 -6.44783914e-01 4.68277752e-01
6.76261544e-01 4.38186228e-01 2.39544645e-01 5.51774621e-01
2.04887077e-01 -1.88542217e-01 -8.46126795e-01 6.91922069e-01
7.70655200e-02 -8.24442625e-01 -3.77443761e-01 -9.74737108e-03
1.02240980e+00 5.31895220e-01 7.64673501e-02 1.93171889e-01
5.98139405e-01 -6.46531641e-01 6.31432891e-01 5.64010918e-01
1.01407671e+00 -2.40563735e-01 5.82684040e-01 -4.99838889e-02
-1.22242868e+00 4.48008254e-02 4.06598687e-01 3.52240652e-01
8.26473534e-01 -1.84056908e-01 -1.17582262e+00 3.64342570e-01
4.74989831e-01 8.71010363e-01 -5.78522801e-01 1.72329307e+00
-1.01180804e+00 1.17213356e+00 -5.97199619e-01 2.23856866e-01
4.57206517e-01 2.07573131e-01 8.18607569e-01 1.48906827e+00
-1.10642761e-02 -7.29698688e-02 2.75688302e-02 3.07304382e-01
2.92859733e-01 -1.65177858e-03 -1.73126861e-01 8.04348886e-02
5.51024526e-02 6.13633335e-01 -9.17210996e-01 -2.98775613e-01
-1.61370650e-01 1.46785188e+00 -3.29630345e-01 8.29627335e-01
-5.11472404e-01 -2.84682810e-01 5.39424002e-01 -1.62976515e-02
7.21723974e-01 -3.57767791e-01 -4.09932584e-01 -1.03637278e+00
6.13061376e-02 -7.26231337e-01 8.11706781e-01 -6.27463996e-01
-1.17148185e+00 1.36393383e-01 2.37914488e-01 -1.48167241e+00
-6.96902633e-01 -7.44159579e-01 -2.45326266e-01 7.78054416e-01
-2.15284038e+00 -1.70575261e+00 -4.30606037e-01 6.60993278e-01
5.50299287e-01 -1.44957244e-01 4.30343509e-01 2.30414048e-01
1.22336000e-01 6.26625299e-01 1.68897301e-01 3.63938093e-01
9.72789347e-01 -1.40271401e+00 6.55580342e-01 1.19447708e+00
2.42309123e-01 3.67125124e-03 5.12291729e-01 -9.78737772e-01
-5.49214005e-01 -1.02020609e+00 8.63192499e-01 -8.05114269e-01
5.70524812e-01 3.13256830e-02 -6.67699933e-01 6.13528013e-01
-3.69230390e-01 1.69982716e-01 3.81027430e-01 -2.88780689e-01
-3.63379210e-01 5.03915668e-01 -1.27367198e+00 6.34572923e-01
1.50696683e+00 -3.24464738e-01 -8.86594713e-01 2.19234332e-01
1.02352008e-01 -7.22714961e-01 -6.32221758e-01 3.99021178e-01
1.00751197e+00 -3.70115787e-01 1.04963052e+00 -8.62523615e-01
5.89580946e-02 -5.19217670e-01 -1.27635136e-01 -1.18990624e+00
2.78086036e-01 -8.21951210e-01 -2.01708704e-01 8.91173661e-01
3.62379014e-01 -5.17721415e-01 1.10973763e+00 6.01257861e-01
-2.54154384e-01 -1.94755599e-01 -1.34377193e+00 -1.06307447e+00
1.09756149e-01 -8.28352690e-01 1.97033927e-01 5.91653466e-01
-4.07263577e-01 -3.63152415e-01 -3.58929753e-01 5.68948314e-02
7.68284857e-01 8.26613978e-02 9.91666675e-01 -1.22040331e+00
-3.35290104e-01 -5.06074429e-01 -6.36510670e-01 -1.33198905e+00
1.02307275e-01 -9.75675225e-01 5.12493432e-01 -1.89242995e+00
-3.28914940e-01 -4.27231789e-01 -2.18050644e-01 8.55941713e-01
-1.08851194e-01 4.89868879e-01 4.35535729e-01 4.57269520e-01
-5.43040991e-01 1.99255377e-01 1.59806883e+00 1.01806142e-03
-5.95939577e-01 3.44615370e-01 -1.59051940e-01 7.77207255e-01
4.32991743e-01 -2.84285337e-01 -2.04127192e-01 -2.25547522e-01
-5.34476340e-01 -5.35467088e-01 5.99422932e-01 -9.49715614e-01
1.45999730e-01 -1.90361381e-01 -9.33670718e-03 -6.51988626e-01
2.12033778e-01 -6.33173406e-01 -4.23638493e-01 4.63938892e-01
-3.33353221e-01 -7.11225629e-01 2.76869863e-01 7.14083970e-01
1.02187693e-02 -1.80669427e-01 9.22714293e-01 -7.59842470e-02
-1.64502358e+00 1.72497362e-01 -3.12486023e-01 6.45093918e-01
7.26544797e-01 -6.51551843e-01 -2.05214873e-01 -3.51757973e-01
-7.70216107e-01 3.61372560e-01 3.78034413e-01 4.61252719e-01
4.36013937e-01 -9.99127567e-01 -9.22083318e-01 1.22383781e-01
4.38548386e-01 -9.10108502e-04 7.15110227e-02 9.35810745e-01
-7.03120470e-01 4.64437246e-01 -3.76234621e-01 -8.67062747e-01
-1.45838809e+00 -3.46227884e-01 2.95349538e-01 4.15481962e-02
-1.00574040e+00 1.13758588e+00 -3.48557025e-01 -7.33756959e-01
3.69052052e-01 -6.68111920e-01 1.36389388e-02 4.89472523e-02
2.70285398e-01 4.23078567e-01 -1.24616124e-01 -8.50355148e-01
-4.19953555e-01 8.64711165e-01 3.11532095e-02 -8.47609460e-01
1.15539587e+00 -1.24839870e-02 5.46850145e-01 3.92547220e-01
7.65393376e-01 -2.42440626e-01 -1.92255425e+00 -2.53375441e-01
4.08663243e-01 -6.39247954e-01 -1.66731462e-01 -1.57855093e+00
-6.69674754e-01 6.70099974e-01 9.51973140e-01 -7.05252290e-01
1.07493651e+00 1.62595421e-01 1.01413655e+00 2.22020388e-01
3.16205680e-01 -1.58490527e+00 -1.78750068e-01 6.26900256e-01
6.94935560e-01 -1.40759337e+00 -2.53069490e-01 -1.20155059e-01
-9.80186582e-01 7.93008745e-01 2.05306813e-01 1.07607886e-01
3.51450980e-01 2.89862268e-02 7.00128317e-01 -6.26711547e-02
-1.54605722e-02 -9.93358254e-01 7.53959298e-01 1.23486876e+00
7.26048723e-02 -1.36112422e-01 -4.53093082e-01 1.11285008e-01
-1.53221106e-02 4.72828090e-01 1.84206918e-01 1.40597200e+00
-2.67907351e-01 -1.29963660e+00 -2.54766792e-01 3.65348160e-01
8.12880173e-02 7.49508366e-02 -6.70428097e-01 1.10837674e+00
2.08598599e-01 7.54546642e-01 -3.41916531e-01 1.48511082e-01
5.16134381e-01 5.81126928e-01 6.66358829e-01 -4.61692780e-01
-3.47375542e-01 6.30758032e-02 3.01795989e-01 -5.03118277e-01
-8.65076602e-01 -1.26429558e+00 -1.52593780e+00 5.42957246e-01
-3.18400376e-02 -4.27787364e-01 8.15781593e-01 1.15365791e+00
9.54310000e-02 4.73354012e-01 -1.63564786e-01 -8.59043300e-01
-3.36235315e-01 -8.69100690e-01 -4.79783118e-01 7.25498378e-01
4.91310269e-01 -6.45114422e-01 -3.57212722e-01 5.84456205e-01] | [9.124917984008789, -6.51456356048584] |
1e5f6a22-2e8a-405f-901e-47e00efc50dc | local-model-reconstruction-attacks-in | 2210.16205 | null | https://arxiv.org/abs/2210.16205v2 | https://arxiv.org/pdf/2210.16205v2.pdf | Local Model Reconstruction Attacks in Federated Learning and their Uses | In this paper, we initiate the study of local model reconstruction attacks for federated learning, where a honest-but-curious adversary eavesdrops the messages exchanged between a targeted client and the server, and then reconstructs the local/personalized model of the victim. The local model reconstruction attack allows the adversary to trigger other classical attacks in a more effective way, since the local model only depends on the client's data and can leak more private information than the global model learned by the server. Additionally, we propose a novel model-based attribute inference attack in federated learning leveraging the local model reconstruction attack. We provide an analytical lower-bound for this attribute inference attack. Empirical results using real world datasets confirm that our local reconstruction attack works well for both regression and classification tasks. Moreover, we benchmark our novel attribute inference attack against the state-of-the-art attacks in federated learning. Our attack results in higher reconstruction accuracy especially when the clients' datasets are heterogeneous. Our work provides a new angle for designing powerful and explainable attacks to effectively quantify the privacy risk in FL. | ['Eoin Thomas', 'Frederic Giroire', 'Giovanni Neglia', 'Chuan Xu', 'Ilias Driouich'] | 2022-10-28 | null | null | null | null | ['inference-attack'] | ['adversarial'] | [-7.62479082e-02 2.03246087e-01 -3.00860077e-01 -6.12115681e-01
-1.10670030e+00 -1.07887352e+00 3.07057947e-01 9.73109305e-02
-1.81198210e-01 6.05757892e-01 -1.63875774e-01 -3.81031662e-01
-2.13575780e-01 -1.19340599e+00 -9.66933429e-01 -1.05243456e+00
-2.52152801e-01 6.41444921e-01 -2.95204818e-02 5.88283464e-02
-2.57637888e-01 5.39919913e-01 -1.10256565e+00 4.06273156e-01
3.44655782e-01 1.10220170e+00 -4.59252566e-01 6.55227482e-01
-3.96707654e-02 9.25873637e-01 -5.31730890e-01 -1.02412605e+00
6.55431509e-01 -3.94806206e-01 -1.04971862e+00 -3.35341722e-01
4.08290476e-01 -5.61942816e-01 -4.38682735e-01 1.08248031e+00
4.70618159e-01 -4.98522639e-01 2.68313419e-02 -1.74957991e+00
-2.72397939e-02 1.13353777e+00 1.31693184e-01 -2.83965856e-01
3.46776336e-01 2.09679201e-01 9.05928493e-01 -2.14619428e-01
5.53006351e-01 1.09409714e+00 7.30711401e-01 8.55946422e-01
-1.24650669e+00 -1.24159610e+00 7.37992898e-02 2.84525782e-01
-1.38359237e+00 -4.24473643e-01 6.83385372e-01 8.90657753e-02
1.22943468e-01 8.28362286e-01 2.00920314e-01 1.34027421e+00
-2.73942947e-04 7.09605753e-01 1.35431266e+00 -1.97789788e-01
3.95109445e-01 5.33610940e-01 1.91520840e-01 7.17301667e-01
2.12095693e-01 2.79227644e-01 -6.47055626e-01 -1.20110118e+00
2.76807070e-01 1.96974501e-01 -4.36361372e-01 -6.07134044e-01
-6.69733346e-01 9.27707136e-01 1.09792069e-01 -1.42699301e-01
-1.49965242e-01 2.30787918e-01 4.56761956e-01 8.86682212e-01
2.95477033e-01 1.36845216e-01 -8.00315857e-01 5.96162006e-02
-4.76458400e-01 2.83671558e-01 1.47033811e+00 8.27529848e-01
8.74698102e-01 -3.51813614e-01 3.91445793e-02 -4.90341745e-02
4.03051078e-01 4.42089289e-01 7.92373996e-03 -1.08318532e+00
4.48396832e-01 4.96230036e-01 5.72951809e-02 -7.00285912e-01
1.20148018e-01 -4.90404546e-01 -6.31815016e-01 9.39435363e-02
5.95590830e-01 -2.99873441e-01 8.04243237e-02 2.02091527e+00
8.67237031e-01 3.34677786e-01 2.46175915e-01 7.61236370e-01
1.54431716e-01 1.64884016e-01 1.58359185e-01 -3.01771551e-01
1.20794916e+00 -8.76523495e-01 -7.20836282e-01 3.15765321e-01
9.46630955e-01 -2.54703730e-01 7.82971978e-01 4.19708461e-01
-1.00300789e+00 1.97843924e-01 -8.67702365e-01 2.37772480e-01
-1.94722190e-01 -6.02580488e-01 9.29732382e-01 1.08041465e+00
-6.31699562e-01 4.91342485e-01 -8.62685263e-01 -1.16694681e-01
7.13915944e-01 6.06417179e-01 -7.21495152e-01 -1.23312376e-01
-1.15515423e+00 2.94090688e-01 -7.02932477e-02 -5.01038849e-01
-1.33767045e+00 -8.90272379e-01 -5.84544718e-01 8.87454748e-02
5.72883844e-01 -9.49575424e-01 1.28990185e+00 -6.38872147e-01
-1.30462790e+00 8.36393952e-01 -6.04716130e-02 -8.94316435e-01
1.03551519e+00 1.77446485e-01 -3.22185606e-01 1.77386209e-01
-3.60541940e-01 -4.93677646e-01 7.15583086e-01 -1.56020021e+00
-6.12832427e-01 -9.13968146e-01 3.44518632e-01 -1.89567387e-01
-5.85091233e-01 1.19683959e-01 3.52720427e-03 -1.52094871e-01
-1.87449738e-01 -8.24886680e-01 -2.46794954e-01 4.23259437e-01
-3.48251790e-01 -4.25844789e-02 1.34714794e+00 -2.90505499e-01
1.14370012e+00 -2.09260130e+00 -2.61996746e-01 5.45549273e-01
4.85605836e-01 2.43211597e-01 6.14340566e-02 7.16536760e-01
2.62601018e-01 3.82679015e-01 -1.72451377e-01 -7.88154423e-01
2.76173770e-01 4.36050624e-01 -6.57640457e-01 8.41299713e-01
-7.15365708e-01 6.87638044e-01 -5.21898746e-01 -5.09246111e-01
-3.17681789e-01 5.15826106e-01 -4.33768600e-01 5.92961907e-01
-2.51844525e-01 5.48534036e-01 -8.27240944e-01 5.38368046e-01
1.08227634e+00 -3.22727859e-01 5.98406255e-01 8.47121254e-02
4.17500138e-01 3.18372369e-01 -1.17756641e+00 1.32557082e+00
-5.20522296e-01 -6.68438524e-02 6.78172648e-01 -4.79681194e-01
8.06954145e-01 5.79069555e-01 4.02498811e-01 -2.04493478e-01
1.00134887e-01 7.98774064e-02 -5.27498245e-01 -3.48862946e-01
-3.50724757e-01 -1.61587253e-01 -1.03026181e-01 1.16145682e+00
-8.29385966e-02 6.25120401e-01 -8.33514094e-01 2.48223066e-01
1.22120488e+00 -1.71083152e-01 3.30361687e-02 -1.61587894e-01
7.43141532e-01 -4.56825554e-01 6.64788246e-01 1.20014441e+00
-3.62820566e-01 5.17108664e-02 6.31791115e-01 -8.10040355e-01
-6.42092168e-01 -1.03092504e+00 5.64570203e-02 1.34310448e+00
1.66304827e-01 -8.61731052e-01 -1.02639925e+00 -1.62267458e+00
4.38123822e-01 6.74093544e-01 -6.45953059e-01 -2.46410921e-01
-3.62027407e-01 -3.01352650e-01 1.17003489e+00 2.41621822e-01
7.84598410e-01 -3.48907977e-01 -1.59054607e-01 7.66931400e-02
-3.54297906e-01 -1.04395759e+00 -3.90533656e-01 8.31522718e-02
-7.58765876e-01 -1.45221663e+00 3.98136675e-01 -2.58493930e-01
5.97961128e-01 1.18926570e-01 8.69939625e-01 4.43575412e-01
1.11009322e-01 6.05485678e-01 1.56453326e-02 -4.84870106e-01
-7.89075017e-01 5.94918579e-02 1.00048659e-02 5.78759253e-01
5.12190878e-01 -6.19261026e-01 -4.78113055e-01 5.44643462e-01
-1.05955291e+00 -3.32122058e-01 9.81628597e-02 6.49003088e-01
5.07253945e-01 -4.83935960e-02 4.42607671e-01 -1.57000387e+00
4.33855146e-01 -5.00170588e-01 -7.59905517e-01 5.87186933e-01
-9.15203452e-01 2.42955938e-01 1.19105434e+00 -4.84167606e-01
-9.51369166e-01 2.98645169e-01 7.16167269e-03 -5.17164528e-01
-2.12869436e-01 -2.89919600e-02 -8.35322201e-01 -5.88155806e-01
6.14912987e-01 4.11686182e-01 3.23368669e-01 -8.80193472e-01
2.48083368e-01 6.07321382e-01 3.29014063e-01 -8.65787268e-01
1.16988003e+00 8.89147520e-01 5.23543000e-01 -5.99035732e-02
-6.74397647e-01 -4.62210178e-03 -3.20931673e-01 9.60798040e-02
2.95941055e-01 -6.68606460e-01 -1.54718208e+00 6.60282731e-01
-9.43772733e-01 -2.02683210e-01 -2.49740630e-01 1.92718923e-01
-5.99046171e-01 5.33956766e-01 -7.29138732e-01 -8.20345521e-01
-6.81945980e-01 -1.12283969e+00 5.21527112e-01 -1.77581668e-01
2.17381939e-01 -1.21587217e+00 -5.79768792e-02 7.29870200e-01
6.08988762e-01 3.58411968e-01 8.56040180e-01 -1.41638935e+00
-7.22598076e-01 -5.47027349e-01 2.05344632e-01 2.24229857e-01
-1.20244585e-01 -4.08203274e-01 -1.26807606e+00 -5.84030032e-01
4.56495941e-01 -4.89659458e-01 3.79096717e-01 -6.02577984e-01
1.57370436e+00 -1.23811233e+00 -1.33691460e-01 1.09773016e+00
1.53145003e+00 -4.85933185e-01 4.47086960e-01 2.36294433e-01
4.68529344e-01 4.42198604e-01 3.70497048e-01 9.06188190e-01
4.43833023e-01 4.48484689e-01 6.92754686e-01 7.56230727e-02
3.55355352e-01 -5.66264868e-01 1.33309707e-01 1.48138747e-01
1.20761611e-01 3.37290876e-02 -5.03957927e-01 -8.26424956e-02
-1.87218988e+00 -9.56470370e-01 -1.25115514e-01 2.52496004e+00
1.04072142e+00 -4.13988531e-01 9.30152237e-02 1.15713151e-03
3.80207211e-01 5.50728571e-03 -5.25314867e-01 -4.85041648e-01
-1.51550427e-01 2.78274566e-01 8.80541623e-01 6.09019935e-01
-1.04569495e+00 7.25929737e-01 5.63902712e+00 5.63906312e-01
-8.95327806e-01 5.05748630e-01 6.14518881e-01 -1.51136756e-01
-4.59150106e-01 3.26033801e-01 -7.81189263e-01 3.79354626e-01
1.17586637e+00 -5.53597271e-01 8.29556108e-01 1.26550019e+00
-3.18051994e-01 5.00911057e-01 -1.44069743e+00 6.67325020e-01
-1.76636294e-01 -1.43810689e+00 4.95459745e-03 4.98592168e-01
3.12313229e-01 -2.38469407e-01 1.75297782e-01 2.10601389e-01
5.35282671e-01 -9.31360841e-01 2.87687808e-01 5.06452322e-01
6.09165311e-01 -1.18239796e+00 6.69126511e-01 7.37161338e-01
-9.74744380e-01 -3.54764640e-01 -4.04295564e-01 1.53460190e-01
-5.31076014e-01 -2.86222082e-02 -6.82531834e-01 7.21966565e-01
4.94402975e-01 -1.05586745e-01 -5.16676009e-01 6.21400714e-01
-6.15312345e-02 9.20693338e-01 -4.59124446e-01 2.97344267e-01
2.21214630e-02 -1.38968989e-01 4.37133849e-01 1.00884163e+00
-1.28076911e-01 1.54802799e-01 1.59874335e-01 8.37124527e-01
-6.20538473e-01 1.41697228e-01 -6.91424131e-01 4.71663952e-01
8.23962450e-01 1.17402804e+00 3.34158123e-01 -2.12833986e-01
-1.52915493e-01 9.27272499e-01 4.91781563e-01 2.16372684e-01
-7.46409595e-01 9.62418690e-03 1.08232379e+00 4.53907661e-02
1.97399165e-02 1.60422355e-01 -1.39818445e-01 -1.20191896e+00
-3.72592323e-02 -1.37734795e+00 1.07491624e+00 4.83534671e-02
-1.51779568e+00 3.87532920e-01 -3.21168125e-01 -7.55846441e-01
-2.13954747e-01 8.40177462e-02 -7.01552749e-01 8.97372544e-01
-1.46907830e+00 -1.56472814e+00 -1.41773462e-01 1.40859258e+00
-3.28700036e-01 -1.39244139e-01 1.53146243e+00 8.24692026e-02
-4.26645488e-01 1.49736214e+00 1.32963032e-01 2.43988708e-01
7.21227348e-01 -9.83879387e-01 8.57679844e-02 7.08539128e-01
1.69098645e-01 8.05980861e-01 6.66291535e-01 -3.79016489e-01
-2.20358181e+00 -1.25790739e+00 8.41701865e-01 -5.30931294e-01
5.36316395e-01 -7.01004982e-01 -9.95446920e-01 1.13091993e+00
-1.31666869e-01 5.75636923e-01 1.20559120e+00 -1.19176619e-01
-1.11559343e+00 -5.79322994e-01 -2.03880715e+00 1.96703896e-01
7.77244210e-01 -9.18311298e-01 1.71648592e-01 5.69279313e-01
8.94475460e-01 -1.12854607e-01 -1.11678350e+00 1.13293573e-01
6.44594491e-01 -1.07729757e+00 8.12511146e-01 -1.25309551e+00
-4.50525194e-01 5.53137474e-02 -4.65115994e-01 -7.03933120e-01
-1.34362828e-03 -1.24417555e+00 -7.48659730e-01 1.31737375e+00
2.41873354e-01 -1.13508427e+00 1.17602038e+00 1.04912019e+00
6.03786290e-01 -6.59813523e-01 -1.28486514e+00 -7.70385146e-01
3.42637807e-01 -4.01846409e-01 1.43455577e+00 1.20565081e+00
-2.48724386e-01 -4.92105901e-01 -3.91923368e-01 7.04985976e-01
1.44610047e+00 1.73732802e-01 1.23188055e+00 -1.18875730e+00
-5.89195251e-01 1.23828307e-01 -2.75203854e-01 -5.10098994e-01
5.86337447e-01 -9.56829607e-01 -7.77668417e-01 -7.73183584e-01
3.80793035e-01 -6.67256176e-01 -3.07600558e-01 7.99199700e-01
1.85063332e-01 -9.12850499e-02 4.57285225e-01 2.69458920e-01
-6.44095600e-01 2.71955967e-01 8.91302764e-01 -5.86066879e-02
2.30959788e-01 5.50651014e-01 -8.26311052e-01 3.58951867e-01
8.28794301e-01 -9.46583211e-01 -3.20195735e-01 2.36413106e-02
4.15043086e-02 2.24067122e-01 6.58830702e-01 -7.00432837e-01
5.70524931e-01 -2.35686839e-01 7.74332359e-02 -1.63809285e-01
1.37982011e-01 -1.43515921e+00 5.57556272e-01 6.84343636e-01
-6.19428933e-01 -2.52328992e-01 -3.94937545e-01 7.40841150e-01
2.31733084e-01 8.66259858e-02 8.09942186e-01 -5.84364682e-02
2.04999939e-01 7.94856071e-01 2.12504074e-01 -7.91967586e-02
1.27271926e+00 2.37782225e-01 -7.03936100e-01 -7.78111458e-01
-7.29892790e-01 2.81594813e-01 8.33429694e-01 -8.50061923e-02
4.82583791e-01 -1.06902301e+00 -7.94434011e-01 6.34941995e-01
1.00159742e-01 -3.09756309e-01 7.50036165e-02 6.97542489e-01
-1.56710908e-01 7.15000033e-02 6.21930100e-02 -2.42684066e-01
-1.71869862e+00 9.19082642e-01 5.66346407e-01 -3.20411265e-01
-6.01254821e-01 4.89429712e-01 -9.71590541e-03 -6.88078880e-01
7.24224865e-01 4.47424024e-01 5.83787918e-01 -4.41630334e-01
8.05776775e-01 5.48670530e-01 7.26841241e-02 -4.94794220e-01
-2.68452853e-01 -4.20043841e-02 -1.47917941e-01 1.09134212e-01
1.15718043e+00 -2.57689446e-01 -4.13199484e-01 3.46806310e-02
1.60077167e+00 1.53065488e-01 -1.08123672e+00 -7.04671621e-01
-2.49865919e-01 -8.58841538e-01 1.84528921e-02 -8.91847610e-01
-1.24423873e+00 5.99292696e-01 5.04590333e-01 2.97623217e-01
1.09274900e+00 9.14305449e-02 9.21744823e-01 4.82553005e-01
9.48492110e-01 -5.99168539e-01 -4.60223347e-01 2.08908282e-02
3.85348678e-01 -8.58063459e-01 -1.47184700e-01 -2.95817077e-01
-5.90296507e-01 9.91887212e-01 5.32462239e-01 3.00186634e-01
7.05221474e-01 4.51817900e-01 2.92359650e-01 1.30039737e-01
-1.16681409e+00 5.94104826e-01 -3.80655497e-01 6.54949427e-01
-3.02046984e-01 2.13469297e-01 1.33191407e-01 1.18852091e+00
-6.68237284e-02 -2.54282773e-01 4.25790042e-01 9.46638286e-01
-8.49395897e-03 -1.65008628e+00 -5.12425959e-01 -5.38189337e-02
-9.76392448e-01 3.75792921e-01 -6.86692178e-01 5.44946551e-01
-2.59173244e-01 1.02053261e+00 -4.70097870e-01 -4.72104222e-01
-1.00723706e-01 4.03079271e-01 1.42671898e-01 2.22065039e-02
-1.24731696e+00 -2.21421525e-01 6.60952181e-02 -1.07210612e+00
8.16858113e-02 -3.14869642e-01 -1.22629249e+00 -1.00676882e+00
-2.76981562e-01 5.40246487e-01 9.22878146e-01 5.84885180e-01
6.07265711e-01 -3.84841412e-01 1.68889558e+00 1.28727332e-01
-1.51172757e+00 -3.51937175e-01 -8.04714441e-01 6.48675799e-01
4.28233415e-01 -2.39249058e-02 -8.27592611e-01 -2.61972934e-01] | [5.802190780639648, 6.805213451385498] |
b2841af9-0884-49c1-928f-d7cdcc9632fd | reward-free-policy-imitation-learning-for | 2304.07988 | null | https://arxiv.org/abs/2304.07988v1 | https://arxiv.org/pdf/2304.07988v1.pdf | Reward-free Policy Imitation Learning for Conversational Search | Existing conversational search studies mainly focused on asking better clarifying questions and/or improving search result quality. These works aim at retrieving better responses according to the search context, and their performances are evaluated on either single-turn tasks or multi-turn tasks under naive conversation policy settings. This leaves some questions about their applicability in real-world multi-turn conversations where realistically, each and every action needs to be made by the system itself, and search session efficiency is often an important concern of conversational search systems. While some recent works have identified the need for improving search efficiency in conversational search, they mostly require extensive data annotations and use hand-crafted rewards or heuristics to train systems that can achieve reasonable performance in a restricted number of turns, which has limited generalizability in practice. In this paper, we propose a reward-free conversation policy imitation learning framework, which can train a conversation policy without annotated conversation data or manually designed rewards. The trained conversation policy can be used to guide the conversational retrieval models to balance conversational search quality and efficiency. To evaluate the proposed conversational search system, we propose a new multi-turn-multi-response conversational evaluation metric named Expected Conversational Reciprocal Rank (ECRR). ECRR is designed to evaluate entire multi-turn conversational search sessions towards comprehensively evaluating both search result quality and search efficiency. | ['Qingyao Ai', 'Zhichao Xu', 'Zhenduo Wang'] | 2023-04-17 | null | null | null | null | ['conversational-search'] | ['natural-language-processing'] | [ 6.07930534e-02 1.42136663e-01 -4.59157646e-01 -2.85965353e-01
-1.05347514e+00 -7.42042780e-01 9.35783029e-01 -3.20684552e-01
-5.33912599e-01 7.09464371e-01 4.77178037e-01 -4.41817045e-01
-2.96179295e-01 -3.72188061e-01 -1.59878582e-01 -5.16845465e-01
4.13510680e-01 8.97351027e-01 1.48196056e-01 -6.70412838e-01
5.02841234e-01 6.25183359e-02 -1.50902283e+00 4.32821721e-01
1.16924965e+00 6.01238608e-01 6.59946322e-01 1.09501088e+00
-1.91141754e-01 7.91436791e-01 -8.76491249e-01 -3.72271985e-01
3.37935239e-02 -7.41581440e-01 -1.36084580e+00 -3.29163313e-01
-1.32347479e-01 -4.79715526e-01 -2.70957559e-01 5.29512405e-01
8.59937787e-01 5.49988806e-01 4.36811239e-01 -1.14261603e+00
-5.19431710e-01 8.16264510e-01 2.21629485e-01 4.19252604e-01
8.73591959e-01 3.76080513e-01 1.29906714e+00 -4.42468822e-01
3.86844963e-01 1.47901833e+00 1.79770291e-01 8.18290770e-01
-8.27823043e-01 -4.43953633e-01 2.09262595e-01 4.00351107e-01
-7.91321337e-01 -4.89376456e-01 7.21454144e-01 -9.57885087e-02
1.14211929e+00 7.61928320e-01 6.46101713e-01 1.36319196e+00
-7.62799084e-02 1.05925083e+00 1.09117591e+00 -3.78926694e-01
-8.79322086e-03 3.18486780e-01 1.25380397e-01 1.64040640e-01
-5.78268886e-01 5.31165712e-02 -5.91708183e-01 -3.85025412e-01
4.06137645e-01 -1.24245629e-01 -4.92893934e-01 2.19536610e-02
-1.13813305e+00 9.42218482e-01 4.27630782e-01 4.04948384e-01
-5.22908866e-01 -1.92314729e-01 4.03033853e-01 7.92990267e-01
3.25673193e-01 9.65857804e-01 -4.18031096e-01 -8.47606838e-01
-4.47125077e-01 4.85669464e-01 1.28420603e+00 8.43389750e-01
4.89489228e-01 -6.01571023e-01 -6.96770847e-01 1.55276310e+00
4.18130271e-02 6.23276651e-01 7.31650352e-01 -1.12424409e+00
4.90685284e-01 6.17116153e-01 2.85389870e-01 -7.90939331e-01
-3.32466006e-01 -1.39650956e-01 -2.97456533e-01 -5.97582042e-01
3.79960418e-01 -8.11450258e-02 -2.17923149e-01 1.70239747e+00
2.89131880e-01 -4.28272635e-01 2.99325466e-01 1.16952908e+00
9.66457427e-01 5.33794105e-01 -3.07426125e-01 -3.95858645e-01
1.47781491e+00 -1.48773730e+00 -7.39968002e-01 -1.73766255e-01
6.99526906e-01 -9.92177308e-01 1.65574443e+00 4.75617796e-02
-1.17838824e+00 -1.35075390e-01 -4.18424755e-01 -1.10505477e-01
-9.78838354e-02 8.81248619e-03 3.63865167e-01 4.54215944e-01
-1.01123118e+00 3.02788205e-02 -3.40695649e-01 -5.15118122e-01
-4.82440114e-01 3.07633281e-01 2.56847858e-01 -4.82875621e-03
-1.66898525e+00 1.19496548e+00 -3.24946076e-01 1.51944816e-01
-7.19177783e-01 -1.94849968e-01 -3.11480910e-01 3.49591345e-01
7.33489096e-01 -6.35054708e-01 1.98489892e+00 -6.61707520e-01
-2.10578012e+00 5.55346549e-01 -2.32582062e-01 -1.93420306e-01
3.97638619e-01 -1.48068503e-01 -1.56246811e-01 1.01111710e-01
-3.34505439e-02 6.20289683e-01 3.51752162e-01 -9.56671476e-01
-6.13924026e-01 1.23227119e-01 7.30309963e-01 9.22871113e-01
-3.20998937e-01 3.65901589e-02 -7.47348726e-01 -3.45189065e-01
-3.65538716e-01 -1.31529629e+00 6.74942089e-03 -7.86549747e-01
-1.90579697e-01 -9.56672132e-01 5.99176049e-01 -4.56109345e-01
1.42307401e+00 -1.50212204e+00 1.82337448e-01 -9.35862288e-02
-1.81529909e-01 3.46618444e-01 -4.65224445e-01 8.05483043e-01
5.34911513e-01 4.11593281e-02 2.78061539e-01 -7.97234103e-02
1.07084051e-01 6.15718961e-02 -1.38439164e-01 -9.57715958e-02
-3.23095143e-01 1.12926829e+00 -9.82800245e-01 -4.92674112e-01
-3.73013467e-02 1.34452567e-01 -8.66347611e-01 7.26964533e-01
-3.52093697e-01 6.13765121e-01 -6.52969360e-01 3.78100395e-01
9.51467007e-02 -4.38033849e-01 1.98174134e-01 2.50132978e-01
1.08840995e-01 8.92547727e-01 -5.28724015e-01 1.37615263e+00
-1.06903529e+00 6.53765440e-01 1.59830943e-01 -7.06293821e-01
8.09274673e-01 3.03633600e-01 3.30951244e-01 -1.20290828e+00
2.87216231e-02 4.70883325e-02 2.48106882e-01 -7.53580093e-01
8.28003347e-01 3.36038738e-01 -2.60376990e-01 1.11369824e+00
-3.69100362e-01 -2.70322978e-01 -3.46392998e-03 2.91595280e-01
1.23243678e+00 -3.38177562e-01 8.32484663e-02 -1.88473407e-02
8.97498190e-01 -3.63219157e-02 2.32884333e-01 1.12859142e+00
-2.87664443e-01 1.91136286e-01 4.18264925e-01 -6.95312247e-02
-6.69537723e-01 -4.03568774e-01 2.17565045e-01 1.86636043e+00
5.32403946e-01 -1.69891313e-01 -8.60973060e-01 -7.53740311e-01
-3.52763116e-01 7.33568311e-01 -2.44942978e-01 -3.11389476e-01
-6.90177679e-01 -3.44674915e-01 5.76273799e-01 -7.92949200e-02
5.39052427e-01 -1.70174873e+00 -5.50671339e-01 2.19268754e-01
-9.60844398e-01 -1.00061011e+00 -1.15042603e+00 -2.31286809e-01
-3.11099738e-01 -1.17435348e+00 -9.35242236e-01 -7.67832875e-01
1.50168359e-01 7.98907936e-01 1.15583074e+00 3.95799875e-01
2.20219851e-01 7.73488343e-01 -7.20750391e-01 -1.12465650e-01
-5.31840801e-01 6.92510307e-01 -1.13228559e-01 -3.25565785e-01
3.13635737e-01 -3.70560199e-01 -1.05140758e+00 1.05873454e+00
-5.45621991e-01 -1.08453698e-01 7.14621127e-01 1.12807083e+00
-2.16305077e-01 -7.30203629e-01 9.20892477e-01 -4.61196244e-01
1.94447124e+00 -2.84696549e-01 -2.09933758e-01 8.19718719e-01
-9.46481705e-01 1.58532426e-01 4.45373893e-01 -7.06993997e-01
-1.10904491e+00 -5.85848212e-01 -3.49146612e-02 -9.61575285e-02
2.70515889e-01 4.53084707e-01 2.06382066e-01 9.16724429e-02
7.74955511e-01 4.86918122e-01 2.14054748e-01 -2.57388860e-01
3.53717685e-01 1.24761605e+00 7.37610385e-02 -8.49886477e-01
1.97933502e-02 -2.29346678e-01 -7.98802793e-01 -6.56157017e-01
-5.02090573e-01 -8.28396380e-01 6.12660274e-02 -4.72799122e-01
4.18137819e-01 -5.37956476e-01 -1.53125668e+00 1.95153520e-01
-1.23784781e+00 -6.33633614e-01 3.06296587e-01 4.60852861e-01
-7.32157052e-01 4.26065475e-01 -6.27809048e-01 -1.11591971e+00
-8.24761868e-01 -1.58760047e+00 1.08793271e+00 2.50480145e-01
-5.46117842e-01 -7.13183105e-01 3.21282864e-01 8.30300152e-01
7.86578298e-01 -9.03135419e-01 7.69092202e-01 -9.71860588e-01
-5.38214505e-01 -4.05250564e-02 -3.64343040e-02 1.87174371e-03
8.46146345e-02 -4.45396841e-01 -6.85775042e-01 -3.15699726e-01
-5.12552299e-02 -7.14703798e-01 4.38946992e-01 1.11968793e-01
9.09851789e-01 -7.48332977e-01 -3.09128970e-01 -6.65492192e-02
3.68708253e-01 4.49518502e-01 3.48841667e-01 3.56436759e-01
3.55721638e-02 9.18025792e-01 8.76409531e-01 1.68641478e-01
8.12383890e-01 1.24076211e+00 8.11832920e-02 4.05094624e-01
9.88315567e-02 -2.42482617e-01 3.33888710e-01 9.51444864e-01
1.96808930e-02 -6.63530111e-01 -7.20285475e-01 5.27192354e-01
-2.07804847e+00 -1.04299414e+00 4.08305287e-01 1.88342154e+00
1.04694307e+00 -2.55305797e-01 3.53438497e-01 -4.51980174e-01
7.32144535e-01 1.09142296e-01 -7.53075480e-01 -7.31883645e-01
2.01428205e-01 -2.64221311e-01 -8.03220198e-02 8.58395696e-01
-3.33796889e-01 1.07650054e+00 6.50106955e+00 7.96333671e-01
-1.13016558e+00 5.40388636e-02 3.64091307e-01 -2.54715413e-01
-5.16064763e-01 8.84879977e-02 -5.92043579e-01 4.67589557e-01
9.08447325e-01 -3.38456959e-01 9.84984815e-01 8.59170437e-01
4.12460119e-01 -1.72378898e-01 -1.07437754e+00 1.07200265e+00
-7.48222917e-02 -1.05949950e+00 -1.04087748e-01 3.93314101e-02
3.72791886e-01 -7.12338462e-02 -9.10436809e-02 9.43228364e-01
6.10256076e-01 -8.18409681e-01 8.08434784e-02 6.38043046e-01
2.13728204e-01 -4.67026651e-01 6.83282077e-01 9.01536703e-01
-6.90925896e-01 -1.70989960e-01 -1.25016838e-01 1.42805632e-02
1.57895520e-01 -2.04843115e-02 -1.33048737e+00 1.33208007e-01
5.95278084e-01 2.37239543e-02 -2.05894068e-01 8.47172737e-01
-8.27387348e-02 4.54032451e-01 -2.65574455e-01 -1.00848627e+00
3.83662283e-01 -2.62725085e-01 5.68348110e-01 1.08180439e+00
1.52698964e-01 3.06359559e-01 1.93262249e-01 5.26567459e-01
-2.41492942e-01 2.63803303e-01 -3.98899138e-01 -8.94888192e-02
9.20530856e-01 1.15792310e+00 -2.47166455e-01 -2.19346911e-01
-1.15438342e-01 7.34437466e-01 4.58810985e-01 4.05374378e-01
-6.47856951e-01 -1.63096055e-01 6.48362279e-01 -1.99315220e-01
-7.46771172e-02 1.93971381e-01 2.24651307e-01 -9.85735893e-01
1.69414610e-01 -1.49133432e+00 3.21469903e-01 -7.97391891e-01
-1.16426575e+00 7.76687741e-01 7.28671923e-02 -9.96078849e-01
-1.02639389e+00 1.36426538e-01 -6.77464128e-01 8.65786135e-01
-1.34438157e+00 -6.91982448e-01 -2.76406676e-01 4.59798574e-01
9.71322119e-01 -1.92965105e-01 7.31855333e-01 1.32300138e-01
-4.78474706e-01 9.66113567e-01 -5.48243113e-02 -2.83805639e-01
1.03937674e+00 -7.95100629e-01 1.33011322e-02 2.20714533e-03
-2.27997378e-01 8.53831947e-01 8.03232372e-01 -2.60343283e-01
-1.45502591e+00 -2.05647588e-01 9.94378328e-01 -4.51697350e-01
3.91975492e-01 -3.69969308e-01 -8.79000306e-01 1.87555417e-01
4.85088915e-01 -6.52441919e-01 4.71489906e-01 5.66000104e-01
1.36969268e-01 2.00296665e-04 -9.58381891e-01 9.50629294e-01
1.13379014e+00 -8.15018952e-01 -6.05702698e-01 5.72951019e-01
8.87373745e-01 -5.05959392e-01 -6.95294738e-01 3.11986119e-01
7.70618200e-01 -8.30720782e-01 8.23508799e-01 -4.32880968e-01
1.32880837e-01 2.94169068e-01 2.81935602e-01 -1.50120962e+00
5.65660782e-02 -9.86277401e-01 4.40724120e-02 9.86512542e-01
5.73105454e-01 -7.56450832e-01 6.33356571e-01 8.72271240e-01
-7.32961893e-02 -1.00677645e+00 -8.25869739e-01 -4.78386313e-01
-8.39270279e-02 -2.88707912e-02 7.53230333e-01 7.13200331e-01
5.16019464e-01 8.39551389e-01 -5.59245169e-01 -2.71628439e-01
-1.89949527e-01 4.42711830e-01 1.22982454e+00 -8.40203345e-01
-2.89438158e-01 -8.53931069e-01 4.95220900e-01 -1.90964854e+00
2.82291591e-01 -6.38633609e-01 4.04694885e-01 -1.50826693e+00
1.75197795e-01 -5.58966875e-01 7.60744661e-02 2.04926208e-01
-4.45880234e-01 -5.89979053e-01 5.95388338e-02 6.55540347e-01
-9.45615232e-01 9.18917120e-01 1.68991518e+00 -1.98566630e-01
-6.34202778e-01 5.37033081e-01 -8.38330746e-01 2.06109717e-01
6.61435008e-01 -2.81120300e-01 -7.32698619e-01 -1.77672505e-01
1.38473064e-01 7.57237852e-01 4.16046381e-02 -2.70089537e-01
7.48639107e-01 -4.19792384e-01 -6.37422442e-01 -4.68068004e-01
4.72816795e-01 -5.42512894e-01 -3.01687628e-01 2.08358526e-01
-1.09124994e+00 2.87350684e-01 -2.58474529e-01 6.33053005e-01
-2.94797361e-01 -2.50566512e-01 2.48445094e-01 -1.01077318e-01
-3.41973454e-01 -4.44507459e-03 -5.42794168e-01 5.24208844e-02
8.49726856e-01 -4.85925749e-02 -4.83565390e-01 -1.21084881e+00
-4.77102101e-01 8.33276391e-01 1.17528573e-01 8.94380569e-01
3.89164925e-01 -1.07123005e+00 -6.33722723e-01 -3.80426675e-01
3.25214356e-01 -2.96078712e-01 1.88250661e-01 7.47066319e-01
1.01992311e-02 1.04168928e+00 1.90916121e-01 -5.57672501e-01
-1.47185946e+00 2.93238670e-01 4.81459677e-01 -7.29139388e-01
-2.04628676e-01 6.55732572e-01 3.14560756e-02 -9.18151140e-01
4.98521775e-01 -1.86686113e-01 -4.13358241e-01 1.40895201e-02
3.06184620e-01 3.08203399e-01 2.74671614e-02 -2.20949039e-01
-1.95632815e-01 1.64118841e-01 -2.28575796e-01 -3.49649668e-01
8.19032192e-01 -4.30955440e-01 8.91337395e-02 4.80451137e-02
1.09819686e+00 -1.03834227e-01 -8.67740929e-01 -5.38828731e-01
-3.21679823e-02 -3.43875319e-01 -2.21165985e-01 -1.11399353e+00
-6.79590404e-01 4.98314470e-01 2.83001661e-01 5.09442747e-01
8.92840624e-01 1.86757401e-01 9.04396713e-01 1.19162500e+00
3.68996710e-01 -1.22936928e+00 5.89350283e-01 7.85438597e-01
1.25021744e+00 -1.43283284e+00 -3.06998610e-01 -1.29250064e-01
-1.07267368e+00 8.95874262e-01 8.45597148e-01 3.48626941e-01
3.50474149e-01 -3.43898833e-01 2.68142909e-01 -2.59803593e-01
-1.21333849e+00 -2.78880358e-01 3.81191313e-01 1.53414428e-01
5.07963359e-01 -1.22666404e-01 -8.05277705e-01 3.79344791e-01
-4.85715330e-01 -2.47050852e-01 6.75586239e-02 6.53585434e-01
-4.41256315e-01 -1.25618446e+00 -2.02467516e-01 4.37461048e-01
-7.26878345e-02 -2.05456614e-02 -9.24979031e-01 4.73135471e-01
-7.89542258e-01 1.58415377e+00 -2.18975678e-01 -4.65020776e-01
6.22207999e-01 1.33140817e-01 1.99166313e-02 -3.88693541e-01
-1.12315655e+00 2.02853000e-03 4.51241434e-01 -5.55506766e-01
-5.29933453e-01 -4.86335307e-01 -7.81903327e-01 -4.22227830e-01
-8.15523207e-01 8.38186681e-01 6.32204056e-01 1.03085303e+00
6.83253169e-01 2.72962272e-01 1.00976074e+00 -4.61054087e-01
-8.79876196e-01 -1.47630262e+00 1.08788818e-01 3.68959427e-01
2.05499843e-01 -6.01437449e-01 -4.44284737e-01 -7.08658874e-01] | [12.136240005493164, 7.827614784240723] |
93d36079-224a-4a4f-8e01-c80d96011ed9 | h-net-unsupervised-attention-based-stereo | 2104.11288 | null | https://arxiv.org/abs/2104.11288v1 | https://arxiv.org/pdf/2104.11288v1.pdf | H-Net: Unsupervised Attention-based Stereo Depth Estimation Leveraging Epipolar Geometry | Depth estimation from a stereo image pair has become one of the most explored applications in computer vision, with most of the previous methods relying on fully supervised learning settings. However, due to the difficulty in acquiring accurate and scalable ground truth data, the training of fully supervised methods is challenging. As an alternative, self-supervised methods are becoming more popular to mitigate this challenge. In this paper, we introduce the H-Net, a deep-learning framework for unsupervised stereo depth estimation that leverages epipolar geometry to refine stereo matching. For the first time, a Siamese autoencoder architecture is used for depth estimation which allows mutual information between the rectified stereo images to be extracted. To enforce the epipolar constraint, the mutual epipolar attention mechanism has been designed which gives more emphasis to correspondences of features which lie on the same epipolar line while learning mutual information between the input stereo pair. Stereo correspondences are further enhanced by incorporating semantic information to the proposed attention mechanism. More specifically, the optimal transport algorithm is used to suppress attention and eliminate outliers in areas not visible in both cameras. Extensive experiments on KITTI2015 and Cityscapes show that our method outperforms the state-ofthe-art unsupervised stereo depth estimation methods while closing the gap with the fully supervised approaches. | ['Daniel S. Elson', 'Stamatia Giannarou', 'Jian-Qing Zheng', 'Baoru Huang'] | 2021-04-22 | null | null | null | null | ['stereo-depth-estimation'] | ['computer-vision'] | [ 2.11215928e-01 3.12798291e-01 1.81036294e-02 -3.91420782e-01
-4.37748224e-01 1.14557706e-02 5.60121179e-01 -6.97996542e-02
-5.44689000e-01 5.43692172e-01 3.94780546e-01 2.62864560e-01
2.38210410e-02 -8.41329396e-01 -6.75606132e-01 -7.71231890e-01
5.01146972e-01 3.63454849e-01 3.12051982e-01 -1.05873324e-01
5.39940298e-01 2.86385387e-01 -1.65530217e+00 3.92576301e-04
9.21329498e-01 8.13987195e-01 4.17141020e-01 1.45292252e-01
-1.96997449e-02 6.68942392e-01 -5.37743466e-03 -2.98812807e-01
6.01895571e-01 -1.84254691e-01 -7.52334177e-01 4.39041436e-01
8.19273531e-01 -7.09153831e-01 -8.20149601e-01 1.21703649e+00
4.47832555e-01 3.00930202e-01 5.29735029e-01 -1.03467751e+00
-2.92506456e-01 1.34853259e-01 -8.67096305e-01 1.68717414e-01
3.25211316e-01 -2.11962145e-02 1.08652055e+00 -8.41658831e-01
6.74931228e-01 1.07055068e+00 3.64174098e-01 3.95814419e-01
-1.14632702e+00 -3.71653497e-01 1.03275955e-01 4.42390054e-01
-1.32345569e+00 -2.75850117e-01 1.05819535e+00 -4.61598009e-01
1.00452971e+00 -3.21232021e-01 8.43289256e-01 8.80758941e-01
5.81698157e-02 9.68771040e-01 7.95474350e-01 -4.17116046e-01
2.37479091e-01 1.93834454e-02 -2.06107944e-01 6.11738026e-01
9.25756544e-02 4.22812313e-01 -6.62273467e-01 3.36215019e-01
1.09878314e+00 4.16600078e-01 -4.87146705e-01 -8.05442989e-01
-1.08048904e+00 9.13781166e-01 7.64324725e-01 1.54074728e-01
-3.79585117e-01 -1.82971820e-01 1.90248907e-01 4.60694581e-02
6.00615859e-01 3.82208437e-01 -2.31324777e-01 6.28853813e-02
-9.08778667e-01 -4.12618592e-02 3.75982374e-01 8.36072385e-01
1.22123027e+00 1.97142716e-02 3.19278777e-01 7.92863846e-01
3.73949021e-01 1.44353807e-01 5.28914511e-01 -1.28157055e+00
6.09638035e-01 9.27334130e-01 -1.99202821e-01 -1.05540264e+00
-1.91163480e-01 -2.54738897e-01 -9.55004156e-01 3.64606529e-01
3.18348706e-01 1.03886247e-01 -8.06657016e-01 1.40750730e+00
3.24077904e-01 3.23434383e-01 9.05087814e-02 1.18463159e+00
6.74768031e-01 3.53195846e-01 -3.71104449e-01 9.40033942e-02
7.33276248e-01 -1.08866322e+00 -4.43550885e-01 -5.73627055e-01
4.62030619e-01 -6.38028026e-01 6.10858023e-01 2.53806859e-01
-8.55714679e-01 -4.82841581e-01 -1.09654522e+00 -5.57908595e-01
-2.19760522e-01 -2.69706935e-01 5.84816635e-01 1.42395288e-01
-9.67553556e-01 5.84059477e-01 -9.39911902e-01 -4.99180883e-01
4.03811872e-01 3.52209717e-01 -7.24934757e-01 -3.52582932e-01
-9.04614747e-01 7.55598962e-01 3.50168914e-01 1.71102673e-01
-5.39366245e-01 -4.55152988e-01 -1.37126517e+00 -1.17740080e-01
9.95538458e-02 -8.39175284e-01 9.35929894e-01 -1.09799039e+00
-1.53735363e+00 1.05367649e+00 -1.98995069e-01 -4.86898452e-01
5.88942528e-01 -3.53233784e-01 1.96795955e-01 3.09573591e-01
3.28276575e-01 1.13860393e+00 7.39533067e-01 -1.07851839e+00
-9.29220200e-01 -5.97744763e-01 2.09408984e-01 6.21368170e-01
-2.66921341e-01 -5.35020411e-01 -6.29812062e-01 -4.70433205e-01
7.16805995e-01 -7.11030364e-01 -2.92309850e-01 1.33130616e-02
-3.58258724e-01 -3.09924153e-03 8.18010867e-01 -5.19018412e-01
8.27129722e-01 -2.18473744e+00 5.11125386e-01 5.04424423e-02
2.38149121e-01 -9.36740339e-02 9.30658430e-02 1.53638437e-01
-1.11827757e-02 -5.65169275e-01 -4.68006045e-01 -6.88163579e-01
-2.89820641e-01 3.35122615e-01 -2.05854755e-02 7.42870510e-01
1.21993080e-01 5.48923850e-01 -9.11598861e-01 -3.45171720e-01
7.89068460e-01 5.47508359e-01 -9.14116323e-01 3.31528008e-01
8.83340463e-03 8.76411736e-01 -1.79538831e-01 3.92674297e-01
7.82939672e-01 -7.94625655e-02 -2.15709731e-01 -2.79683083e-01
-2.81863600e-01 1.93667218e-01 -1.20340168e+00 2.18332577e+00
-3.99543852e-01 8.85289252e-01 -1.41814530e-01 -1.01295722e+00
7.54408181e-01 4.30284031e-02 6.55607104e-01 -8.60595822e-01
2.44831026e-01 1.54752389e-01 -2.22212434e-01 -3.94220769e-01
4.75714892e-01 1.24880699e-02 2.25934565e-01 1.97585583e-01
1.52411938e-01 -5.26942194e-01 7.17299059e-02 4.67154570e-02
7.69336104e-01 3.06568593e-01 3.98736447e-01 -6.60806224e-02
7.16522932e-01 -1.17343068e-01 8.31071973e-01 3.59799504e-01
-3.14418286e-01 1.18456626e+00 -1.06395520e-01 -6.44870281e-01
-1.12070453e+00 -9.37692940e-01 -6.17622286e-02 4.22411680e-01
5.78651547e-01 -2.09736764e-01 -7.71442771e-01 -5.38914979e-01
-1.03561573e-01 2.56415665e-01 -5.30193150e-01 -1.17490135e-01
-4.38386470e-01 -1.63557991e-01 5.58080710e-02 5.86827397e-01
9.54115868e-01 -9.04914021e-01 -6.08005762e-01 8.23047385e-02
-3.51594388e-01 -1.27236521e+00 -4.35915977e-01 2.02738181e-01
-1.15304697e+00 -1.20018768e+00 -8.72119963e-01 -9.14660990e-01
9.09967422e-01 5.68834245e-01 7.60596812e-01 -1.20764688e-01
-2.13487759e-01 3.13855290e-01 -8.14925283e-02 -4.83888313e-02
2.05719709e-01 8.64160731e-02 -8.84573311e-02 1.05179548e-01
6.51856542e-01 -7.71160841e-01 -1.03570592e+00 2.48592421e-01
-8.94507170e-01 2.98951983e-01 5.37317753e-01 1.03926671e+00
6.99964345e-01 -5.39714238e-03 -1.53300658e-01 -7.86851108e-01
-1.71975613e-01 -2.98389822e-01 -7.28652894e-01 -3.10758054e-01
-3.75732243e-01 2.03873277e-01 3.51144701e-01 4.31143381e-02
-1.23823428e+00 4.12530959e-01 -2.01300457e-01 -5.69793701e-01
-2.98282832e-01 2.61268109e-01 -2.47878149e-01 -9.24925059e-02
3.89405161e-01 2.54066855e-01 -4.64250445e-02 -2.30373040e-01
1.41876161e-01 4.34271604e-01 5.93097329e-01 -3.16107385e-02
7.62747586e-01 1.07377446e+00 -4.75498661e-02 -9.66061115e-01
-1.03301036e+00 -9.19450939e-01 -1.09284806e+00 -7.71681890e-02
9.96000767e-01 -1.16416419e+00 -2.60550231e-01 6.92335486e-01
-1.10925150e+00 -3.13357055e-01 -1.49544954e-01 8.12817335e-01
-8.09835017e-01 7.63822556e-01 -5.04317820e-01 -4.77480233e-01
-1.06913783e-01 -1.21531916e+00 1.25712001e+00 3.01010132e-01
-1.87120005e-01 -1.15270257e+00 2.76643813e-01 6.61717117e-01
1.79683156e-02 -2.22899634e-02 4.31238174e-01 -1.50080085e-01
-9.72384274e-01 -5.32882288e-02 -4.16433722e-01 4.87217575e-01
2.25350291e-01 -2.59771734e-01 -1.13580716e+00 -1.95292711e-01
1.67984873e-01 -7.59422407e-02 1.01565850e+00 5.62440395e-01
8.53258669e-01 6.43412843e-02 -1.79566190e-01 1.04901874e+00
1.52504575e+00 -9.99302864e-02 8.44093144e-01 7.65585959e-01
9.84322131e-01 8.20297658e-01 5.55029213e-01 4.53605741e-01
6.71854198e-01 4.87093925e-01 6.69016898e-01 -2.26523608e-01
1.76206827e-02 -4.21871662e-01 9.04374495e-02 7.20971942e-01
-4.37658504e-02 1.76535308e-01 -7.99134851e-01 6.62478805e-01
-1.89389741e+00 -9.27643299e-01 -6.68504685e-02 2.39662647e+00
5.34104526e-01 1.72401130e-01 -4.12401080e-01 3.43011409e-01
6.04615331e-01 2.44680569e-01 -4.85667735e-01 1.00471497e-01
-2.97294527e-01 -2.74106562e-02 4.79594558e-01 7.84579277e-01
-1.26797485e+00 1.07296610e+00 5.20632219e+00 3.36568385e-01
-1.10573065e+00 -1.79257512e-01 4.29397941e-01 9.61805806e-02
-1.29749611e-01 1.53727487e-01 -5.64814448e-01 2.98730463e-01
5.94974458e-02 1.57669097e-01 2.44108066e-01 8.54698539e-01
3.80914003e-01 -5.26261568e-01 -1.19710684e+00 1.39640880e+00
2.54136562e-01 -1.06359649e+00 -6.21074066e-02 2.25987270e-01
1.27528358e+00 2.56651014e-01 1.32321231e-02 -8.32325406e-03
1.59693435e-01 -6.51987970e-01 3.80219758e-01 4.23595279e-01
3.67312133e-01 -7.09565699e-01 9.52980936e-01 3.48348469e-01
-9.95384336e-01 -1.11113198e-01 -5.42679429e-01 -3.40147704e-01
2.25832805e-01 7.11859047e-01 -4.80139703e-01 4.28119600e-01
9.36086476e-01 1.29533327e+00 -3.29377860e-01 1.24907565e+00
-5.06278574e-01 -1.27088979e-01 -3.77384692e-01 5.53312242e-01
3.12566608e-01 -4.64570373e-01 4.94165719e-01 7.38006830e-01
2.42943138e-01 2.14060359e-02 -1.13523249e-02 6.70696676e-01
6.91327304e-02 1.37796067e-02 -9.59324598e-01 5.00549495e-01
1.38653353e-01 1.01822698e+00 -4.64298159e-01 -1.49570450e-01
-6.93082929e-01 1.38709068e+00 4.37709838e-01 3.79907489e-01
-3.96379530e-01 -1.19511180e-01 7.08383977e-01 1.20501474e-01
2.49361157e-01 -2.81829357e-01 -4.46587533e-01 -1.46619034e+00
-1.34447999e-02 -5.33344686e-01 3.62481892e-01 -7.67217517e-01
-1.17248571e+00 3.90447706e-01 -2.59498775e-01 -1.41608310e+00
-3.23490649e-01 -4.22027171e-01 -6.22804999e-01 6.97960377e-01
-1.99461436e+00 -8.19333196e-01 -7.52067387e-01 8.16530824e-01
8.77373517e-01 -9.43246707e-02 4.47781563e-01 3.10813218e-01
-4.68612611e-01 2.14730412e-01 7.40534067e-02 1.52827337e-01
8.02091956e-01 -1.23843169e+00 1.41929865e-01 9.60431516e-01
4.36516330e-02 5.36767244e-01 5.46864569e-01 -4.69057381e-01
-9.88889039e-01 -1.01448786e+00 9.09676254e-01 -1.30517468e-01
4.44870740e-01 -1.40470192e-01 -8.86434734e-01 6.16702855e-01
2.99541324e-01 4.92239185e-03 4.46292251e-01 -2.92712599e-01
-2.91515082e-01 -1.65575445e-01 -9.26697373e-01 7.07951546e-01
1.03724062e+00 -7.66889393e-01 -6.19475067e-01 4.08373289e-02
3.08054000e-01 -5.24245262e-01 -5.63623011e-01 4.03896064e-01
3.51798236e-01 -1.56281567e+00 8.92737269e-01 1.01121821e-01
6.74948871e-01 -3.29474479e-01 -1.31384015e-01 -1.25785291e+00
1.36424045e-04 -2.67594010e-01 9.17370692e-02 8.84969056e-01
1.15787409e-01 -5.64192057e-01 1.25643265e+00 5.82051575e-01
-1.81737483e-01 -4.51738834e-01 -9.21801627e-01 -3.71226251e-01
-2.32687488e-01 -2.59722352e-01 8.35812911e-02 8.83056879e-01
-6.46599308e-02 3.52004707e-01 -2.50158817e-01 2.76088387e-01
9.37780023e-01 2.52280775e-02 9.21306789e-01 -1.27477598e+00
-3.45118254e-01 -4.52681094e-01 -8.16892385e-01 -1.59873807e+00
3.22681695e-01 -6.54451370e-01 2.76628375e-01 -1.55059254e+00
2.41199210e-01 1.21633224e-02 2.56954487e-02 1.40956193e-02
-2.80651990e-02 3.77291560e-01 -1.59109995e-01 2.84081846e-01
-4.24937278e-01 9.53456402e-01 1.17854512e+00 -2.43932590e-01
-2.47577831e-01 2.20468221e-03 -2.82967001e-01 1.18063986e+00
6.85152531e-01 -2.49647737e-01 -4.91905659e-01 -9.57829475e-01
2.30668429e-02 -7.26524442e-02 2.76732206e-01 -1.16639102e+00
6.97192729e-01 -1.06595969e-03 2.58981556e-01 -7.43429780e-01
5.11539936e-01 -1.09461796e+00 -2.51079232e-01 2.10627005e-01
-5.09159490e-02 -9.36856121e-02 -1.11664936e-01 7.20600843e-01
-7.60316372e-01 -1.05753712e-01 8.94049048e-01 -2.21786112e-01
-1.01266944e+00 5.77427268e-01 -2.26953983e-01 -2.98404116e-02
9.09499943e-01 -7.64268339e-01 7.72448629e-02 -4.21390802e-01
-5.51774085e-01 2.28192210e-01 7.89237559e-01 3.07699531e-01
9.84345913e-01 -1.11932206e+00 -4.38679665e-01 5.10968804e-01
3.17934990e-01 6.14129424e-01 4.05152380e-01 9.04148221e-01
-7.65016437e-01 4.38453615e-01 -4.66675460e-01 -9.23347890e-01
-9.82731283e-01 2.80628830e-01 4.27817374e-01 -2.86461934e-02
-8.15425694e-01 7.55708933e-01 7.69525647e-01 -3.78044873e-01
4.53492522e-01 -1.36864990e-01 -3.24388385e-01 -1.03495747e-01
2.92615294e-01 3.00671488e-01 1.06862284e-01 -8.29086006e-01
-1.22016922e-01 9.53670204e-01 -1.28296986e-01 -1.38947964e-01
1.44057810e+00 -5.69226444e-01 -1.09842986e-01 2.00001135e-01
1.39240897e+00 -1.26421094e-01 -1.75717604e+00 -4.62409139e-01
-7.75802732e-02 -7.59964645e-01 3.82048666e-01 2.04392971e-04
-1.40187848e+00 1.14685440e+00 4.99589384e-01 -3.04236799e-01
1.13334620e+00 -2.72556633e-01 8.56062710e-01 3.45073014e-01
4.24658984e-01 -1.19486940e+00 2.51928419e-01 6.97809279e-01
5.57570577e-01 -1.85553312e+00 -1.24920048e-01 -4.99863476e-01
-5.57905674e-01 1.15150785e+00 9.14120555e-01 -3.87791395e-01
5.55659533e-01 -8.46677497e-02 1.51338890e-01 -1.03961945e-01
-2.83382088e-01 -4.42180991e-01 1.74760580e-01 6.56428099e-01
2.93258280e-01 -4.92032915e-01 -3.40937600e-02 -7.56354779e-02
-4.92013581e-02 -1.57797307e-01 4.52138454e-01 8.28863144e-01
-3.45735818e-01 -9.47329938e-01 -1.73654616e-01 1.12268157e-01
-2.09654063e-01 -1.63093343e-01 -2.56817102e-01 6.87354565e-01
7.75143951e-02 8.01842570e-01 3.77990067e-01 -1.52908474e-01
2.47372687e-01 -4.29091930e-01 4.80455130e-01 -7.87248850e-01
-1.46952078e-01 1.19835198e-01 -3.86694223e-01 -6.98420763e-01
-6.45543814e-01 -6.24768317e-01 -1.18589950e+00 -1.30931675e-01
-2.02720866e-01 -1.68959603e-01 4.48137194e-01 1.06945539e+00
2.41063982e-01 7.56216869e-02 6.71262622e-01 -1.17982543e+00
-3.75714414e-02 -6.67994857e-01 -5.44084787e-01 6.57821834e-01
4.42054838e-01 -7.71717966e-01 -5.77439606e-01 6.50820062e-02] | [8.735236167907715, -2.406327486038208] |
0956c109-8158-4b9b-9dcb-23faab38b112 | automatic-generation-of-board-game-manuals | 2109.09507 | null | https://arxiv.org/abs/2109.09507v1 | https://arxiv.org/pdf/2109.09507v1.pdf | Automatic Generation of Board Game Manuals | In this paper we present a process for automatically generating manuals for board games within the Ludii general game system. This process requires many different sub-tasks to be addressed, such as English translation of Ludii game descriptions, move visualisation, highlighting winning moves, strategy explanation, among others. These aspects are then combined to create a full manual for any given game. This manual is intended to provide a more intuitive explanation of a game's rules and mechanics, particularly for players who are less familiar with the Ludii game description language and grammar. | ['Cameron Browne', 'Dennis J. N. J. Soemers', 'Eric Piette', 'Matthew Stephenson'] | 2021-09-20 | null | null | null | null | ['board-games'] | ['playing-games'] | [-1.42873049e-01 4.94330943e-01 3.57640594e-01 3.99792120e-02
-2.11181909e-01 -1.01069915e+00 5.55151403e-01 -9.04987287e-03
-3.21827978e-01 6.48453474e-01 4.39371206e-02 -8.66119504e-01
-3.34340483e-01 -6.42781556e-01 2.26031840e-01 -1.42615527e-01
2.56331652e-01 6.55330896e-01 5.01223981e-01 -8.59476566e-01
5.09627521e-01 4.80036616e-01 -1.58398366e+00 7.72964805e-02
4.16719943e-01 9.37211141e-02 6.55826747e-01 1.22862780e+00
-5.37432194e-01 1.03537869e+00 -8.52966189e-01 -6.18918180e-01
3.20902020e-02 -8.51683736e-01 -1.15165114e+00 -1.06748134e-01
-4.94337887e-01 -2.22115934e-01 -1.12060234e-01 1.09583867e+00
1.55515775e-01 4.29495245e-01 7.51884878e-01 -1.06981194e+00
1.16219372e-01 1.02592945e+00 -7.32807070e-02 1.01282187e-01
1.09391308e+00 1.51219293e-01 7.60109305e-01 -4.42977458e-01
1.06029904e+00 9.25731719e-01 4.52551216e-01 5.96701324e-01
-7.63015866e-01 -5.35637975e-01 -1.18284777e-01 2.59187788e-01
-1.43082774e+00 -1.07204884e-01 6.03941500e-01 -5.28237641e-01
1.25561166e+00 9.09362376e-01 1.17956066e+00 4.68054593e-01
4.51692700e-01 4.10965413e-01 6.71452582e-01 -7.67515719e-01
4.80683558e-02 3.45010221e-01 -4.23482768e-02 5.23713887e-01
2.43072867e-01 3.52290156e-03 -3.49290490e-01 1.47387445e-01
1.16217542e+00 -4.90559042e-01 -9.97988135e-03 -3.19642544e-01
-6.50410652e-01 8.55735183e-01 -4.25137162e-01 5.55343270e-01
-1.96952134e-01 3.12095992e-02 7.17914224e-01 3.83439064e-01
-1.20486028e-01 7.93150485e-01 -4.18136805e-01 -1.05631316e+00
-6.38087809e-01 9.62563992e-01 1.13609433e+00 1.10712504e+00
3.85099947e-01 1.14053331e-01 4.69778657e-01 6.34529173e-01
5.48654616e-01 6.20333180e-02 3.69336456e-01 -8.23493958e-01
2.74316937e-01 6.13278508e-01 2.90801108e-01 -9.75500286e-01
-6.76670134e-01 -6.74981996e-02 -4.71314006e-02 8.46889317e-01
2.84336925e-01 -1.09289341e-01 -3.61254066e-01 1.19743609e+00
-4.26430069e-02 -7.41567314e-01 2.22040504e-01 7.28686213e-01
1.80791223e+00 4.38578367e-01 4.05199872e-03 2.50810552e-02
1.43584239e+00 -7.21952856e-01 -8.71502280e-01 -4.19051170e-01
1.02305305e+00 -8.19366336e-01 1.17048204e+00 4.83757764e-01
-1.64581048e+00 -3.84389699e-01 -8.99760544e-01 1.34271890e-01
-5.54440081e-01 2.59681523e-01 9.23416615e-01 7.11546004e-01
-1.25901163e+00 4.76193398e-01 -7.76355028e-01 -6.53257906e-01
-3.94548506e-01 7.18325496e-01 -1.99144378e-01 3.84315878e-01
-1.20481586e+00 1.07429349e+00 9.88408089e-01 -2.69956917e-01
2.28722677e-01 -2.16086239e-01 -1.33583319e+00 -2.45117024e-02
4.60604936e-01 -4.89194721e-01 1.64930117e+00 -3.36810559e-01
-1.91615665e+00 1.28989148e+00 4.80124682e-01 6.87493086e-02
6.99888766e-01 2.80237377e-01 -3.18057626e-01 -1.80750430e-01
4.07204092e-01 4.01347399e-01 -7.73079023e-02 -8.24089527e-01
-1.06644738e+00 8.78916383e-02 8.87309611e-01 6.87430382e-01
5.72963655e-01 6.15180492e-01 -8.67596745e-01 -5.01457810e-01
-1.99591428e-01 -7.87144065e-01 -1.64667875e-01 -7.68275380e-01
-3.47606808e-01 -2.87341774e-01 4.79194783e-02 -7.79738545e-01
1.84959865e+00 -1.90594864e+00 2.69491315e-01 1.41920537e-01
3.98126245e-01 1.78981442e-02 3.93073201e-01 9.87909138e-01
-3.19999009e-01 1.82053953e-01 4.51591283e-01 1.11183673e-01
4.93241936e-01 2.21314982e-01 2.73708373e-01 -1.74283490e-01
-4.04784739e-01 8.49803329e-01 -9.02701139e-01 -3.85100424e-01
5.78402936e-01 -2.55825907e-01 -5.46929896e-01 1.87493954e-02
1.56347796e-01 2.37520128e-01 -3.74445945e-01 1.74566358e-01
2.85791576e-01 2.25322917e-01 3.89138907e-01 4.07174259e-01
-6.07997656e-01 6.02390468e-01 -1.47096670e+00 1.72622740e+00
-3.79780471e-01 8.02201807e-01 -8.96851793e-02 -5.71834147e-01
9.53356624e-01 2.19617575e-01 2.09205285e-01 -2.16388330e-01
3.74750704e-01 3.30217928e-01 2.82128990e-01 -6.41825378e-01
9.70596373e-01 -1.82461262e-01 -5.59378743e-01 6.24308228e-01
-1.87443808e-01 -6.82853103e-01 1.02431643e+00 4.24050629e-01
9.74111080e-01 5.71429908e-01 1.00974071e+00 -1.75285190e-01
4.34518307e-01 6.77148819e-01 -5.09644393e-03 7.84696579e-01
2.05948561e-01 2.03031242e-01 4.98921722e-01 -4.93857920e-01
-8.46126795e-01 -6.49414897e-01 3.69814396e-01 9.97513533e-01
2.31545955e-01 -1.17657650e+00 -8.92143428e-01 -1.26686752e-01
-7.18339384e-01 9.81688619e-01 -4.85799432e-01 4.91255149e-02
-5.52002370e-01 -4.75698709e-02 3.11876267e-01 3.69945377e-01
2.60757089e-01 -1.36387098e+00 -1.34623373e+00 4.51401204e-01
-2.26333305e-01 -7.49652803e-01 -1.80380747e-01 4.37983632e-01
-1.90322399e-01 -1.09694946e+00 -2.03710362e-01 -7.76735306e-01
1.97750166e-01 -8.65713358e-02 1.31988800e+00 3.88095319e-01
-4.88131881e-01 4.67194647e-01 -6.80579841e-01 -6.17623389e-01
-8.19546103e-01 2.17948258e-01 -2.81281710e-01 -1.39577401e+00
4.13644493e-01 -3.99875671e-01 3.18721056e-01 -1.63662601e-02
-8.09302807e-01 8.40887964e-01 2.24761993e-01 4.86877292e-01
1.80499658e-01 4.18233782e-01 -2.08443508e-01 -1.12550926e+00
1.11851692e+00 1.64809883e-01 -5.23123920e-01 -1.64567716e-02
-1.10284097e-01 -9.09840688e-02 5.70105433e-01 -1.42336890e-01
-8.81068170e-01 -1.20727271e-01 -5.01077473e-01 4.51697618e-01
-4.51424628e-01 7.42592454e-01 -4.10495251e-01 -1.97503477e-01
6.52941406e-01 1.28516510e-01 -3.48571599e-01 -3.49891931e-01
2.94213951e-01 3.85845691e-01 7.37073898e-01 -2.67597616e-01
7.82200396e-01 -4.36863393e-01 -1.19509771e-01 -6.68970883e-01
2.80896932e-01 -4.51604545e-01 -7.88723767e-01 -6.28350198e-01
8.33253860e-01 -2.86367893e-01 -8.52082789e-01 1.73421770e-01
-1.15046751e+00 -7.45228231e-01 -4.47139561e-01 3.78814965e-01
-1.06263840e+00 7.54024908e-02 -3.57834846e-01 -7.26523578e-01
2.95005701e-02 -1.24081993e+00 6.74271822e-01 5.67766249e-01
-1.07862127e+00 -1.32840955e+00 2.53693730e-01 -1.06064193e-01
-1.62090883e-02 1.46719381e-01 9.51928377e-01 -7.91632175e-01
1.58489123e-01 -2.45730758e-01 2.68893301e-01 -4.81465250e-01
1.70121044e-01 1.11787952e-01 -1.87080860e-01 2.16285363e-01
-2.34997556e-01 6.59712702e-02 -3.09543103e-01 2.60150731e-01
6.49101317e-01 -4.49018180e-01 -4.36100811e-01 4.81270880e-01
1.38983488e+00 8.20682168e-01 7.37933815e-01 9.81193721e-01
3.03252280e-01 5.66470742e-01 1.08707309e+00 5.80119133e-01
3.25364709e-01 1.06352174e+00 2.52956092e-01 -1.46874249e-01
8.06468055e-02 -9.94458422e-02 2.27593169e-01 6.96312487e-01
-4.84782636e-01 -3.19625109e-01 -1.18716788e+00 1.96417257e-01
-1.83532870e+00 -1.06839728e+00 -6.10019267e-01 1.63176358e+00
4.55043763e-01 3.95573854e-01 7.06788063e-01 4.24172878e-01
4.54740047e-01 -8.02950636e-02 1.42619208e-01 -1.23811364e+00
3.75481755e-01 5.93756318e-01 4.67937708e-01 8.84542108e-01
-8.36028695e-01 1.42254746e+00 7.58018160e+00 1.07579195e+00
-7.27483869e-01 -1.83296666e-01 -2.48742417e-01 5.55637404e-02
-3.38251978e-01 7.89502710e-02 -5.11160970e-01 6.92027733e-02
6.89889967e-01 -8.54731023e-01 5.47671497e-01 7.60731876e-01
4.49161023e-01 -3.36632520e-01 -7.47274220e-01 1.25002503e+00
-2.48254225e-01 -1.54619825e+00 -1.16184101e-01 1.92106366e-01
1.48324832e-01 -6.46908581e-01 -4.01301354e-01 5.83431900e-01
7.32968986e-01 -1.06188810e+00 1.19864726e+00 2.76831865e-01
9.74624097e-01 -9.94653225e-01 9.25193608e-01 3.61786842e-01
-1.35384178e+00 3.15634191e-01 -3.53116065e-01 -8.75724614e-01
6.99317753e-02 -6.42971992e-01 -7.89175212e-01 9.32451606e-01
5.19386768e-01 2.23918051e-01 -5.56419075e-01 1.05855143e+00
-4.70821291e-01 1.66158661e-01 -9.18102488e-02 -5.51823795e-01
4.76169288e-01 -4.40437257e-01 5.95420897e-01 1.35066307e+00
1.98009402e-01 6.67090356e-01 1.49074093e-01 7.89547622e-01
9.26311314e-01 5.35135865e-01 -7.66267657e-01 -2.15547711e-01
3.47040206e-01 1.22130775e+00 -1.31876147e+00 -2.42699727e-01
-1.56136140e-01 1.17625034e+00 -2.64807612e-01 2.62418911e-02
-8.45185995e-01 -9.67881560e-01 6.58250988e-01 4.43246424e-01
4.27592322e-02 -2.83472389e-01 -3.50735009e-01 -6.18226051e-01
-4.56439883e-01 -9.94400144e-01 2.69270152e-01 -1.37914050e+00
-9.55139175e-02 7.63545692e-01 6.49085402e-01 -1.12099051e+00
-1.29039371e+00 -9.98272240e-01 -1.00698817e+00 1.02851069e+00
-2.14887604e-01 -9.86719072e-01 -2.47350872e-01 2.93987095e-01
7.54556358e-01 -4.14303273e-01 1.11692476e+00 -1.00389414e-01
-2.46152312e-01 2.97116637e-01 -2.85852492e-01 -1.33887574e-01
4.73548956e-02 -1.58365881e+00 6.72562063e-01 7.87786722e-01
-7.15149613e-03 5.17102003e-01 1.33692062e+00 -7.01328039e-01
-7.48695016e-01 -2.79150516e-01 9.60530043e-01 -3.12100440e-01
8.47029150e-01 -1.67097762e-01 -2.04384610e-01 5.54861665e-01
2.15874717e-01 -1.12803495e+00 6.94349825e-01 -1.01944059e-01
6.24282420e-01 6.15847230e-01 -6.53069019e-01 1.13861930e+00
1.21160257e+00 -2.92293459e-01 -6.95767760e-01 1.89840466e-01
1.17233500e-01 -1.11672962e+00 -6.26743615e-01 -3.35751563e-01
3.91661763e-01 -9.46057379e-01 6.67376161e-01 -6.75818920e-01
3.57013792e-01 -3.88634980e-01 4.18198228e-01 -1.31235552e+00
-5.02816200e-01 -1.11607814e+00 8.04955244e-01 1.13201809e+00
4.56588238e-01 -3.15223485e-01 6.17923856e-01 6.86130583e-01
-3.81917894e-01 -3.39281023e-01 -4.92716551e-01 -6.76435411e-01
-1.39138669e-01 -9.58232105e-01 5.91070473e-01 6.62694156e-01
1.05742776e+00 3.32191944e-01 -1.93502545e-01 -4.18257713e-01
-2.35718235e-01 -2.59529620e-01 1.14616585e+00 -1.24569345e+00
-4.35326308e-01 -9.96255636e-01 -9.39728081e-01 -8.77582610e-01
-1.10985585e-01 -7.48188555e-01 9.24220607e-02 -2.23691440e+00
1.23891681e-02 1.44962780e-02 7.16879070e-01 4.33786005e-01
3.76262069e-02 1.54329836e-01 3.49891037e-01 1.34064317e-01
-6.27801061e-01 -2.91595668e-01 9.70193326e-01 2.94119239e-01
-5.80927789e-01 8.82610977e-02 -1.05247974e+00 1.00491226e+00
8.48803520e-01 -5.73819458e-01 -6.28651083e-01 3.73932160e-02
5.77377796e-01 2.04297453e-01 -6.68508485e-02 -1.04699504e+00
2.64897197e-01 -5.54781914e-01 -1.00018889e-01 -5.16294718e-01
2.26966720e-02 -3.92118931e-01 8.92830670e-01 7.36807466e-01
-8.93112794e-02 6.21355474e-01 7.44097173e-01 -2.93622255e-01
-9.07098725e-02 -8.06913435e-01 2.51038671e-01 -4.38218892e-01
-1.05845380e+00 -3.98871750e-01 -1.29785705e+00 -1.69537306e-01
1.14937198e+00 -9.99693751e-01 3.69991452e-01 -9.73173678e-01
-1.42365849e+00 -4.81425896e-02 9.13455427e-01 8.40521082e-02
2.38234654e-01 -1.16822612e+00 -4.62405294e-01 -2.80527603e-02
1.11885414e-01 -3.94087434e-01 2.26595715e-01 4.96694416e-01
-1.43219984e+00 6.06997013e-01 -7.11646855e-01 5.12383096e-02
-1.67912173e+00 2.12071598e-01 3.58740002e-01 -7.11823940e-01
-8.77617955e-01 8.37236822e-01 3.81110549e-01 -4.22071427e-01
-1.60443068e-01 -2.83060670e-01 -8.54117870e-01 -3.81355435e-01
6.10747337e-01 2.55463630e-01 3.58809230e-05 -7.76896834e-01
-6.54379606e-01 4.89336580e-01 1.27827182e-01 -6.16817713e-01
1.19873011e+00 -3.50276589e-01 9.63214785e-02 6.18776917e-01
7.93539956e-02 2.76902169e-01 -5.85862279e-01 4.40285236e-01
4.84919660e-02 -2.46629789e-01 -3.05306226e-01 -9.35584426e-01
-5.73875844e-01 4.80060279e-01 1.07221276e-01 6.77035630e-01
9.10545707e-01 4.14612703e-02 1.65724292e-01 1.82644472e-01
5.46637416e-01 -1.14996350e+00 -4.71672803e-01 7.30851054e-01
9.67717528e-01 -3.28872472e-01 -1.73190281e-01 -6.35612547e-01
-9.03684020e-01 1.58596730e+00 5.15065670e-01 1.61578268e-01
2.04774603e-01 6.17574871e-01 3.02039564e-01 -5.79188228e-01
-4.89854485e-01 -4.69141126e-01 6.97530359e-02 1.14593005e+00
5.21769047e-01 1.02092214e-01 -1.20347059e+00 1.33212364e+00
-1.14880490e+00 -3.29663008e-02 1.12616348e+00 1.00440860e+00
-3.55950713e-01 -1.28666317e+00 -5.83878994e-01 1.66399717e-01
-3.79228622e-01 -4.89392579e-02 -9.79411066e-01 1.60930383e+00
1.57255217e-01 1.11903310e+00 -1.12411797e-01 -6.78969324e-01
5.39487839e-01 -1.36458009e-01 7.08756089e-01 -1.16317308e+00
-9.43557918e-01 9.78274494e-02 6.02700174e-01 -3.63051683e-01
-2.18695011e-02 -6.77765965e-01 -1.56333029e+00 -9.05722380e-01
-1.29123017e-01 4.98121440e-01 6.17218137e-01 1.13541174e+00
-4.47672844e-01 8.55474114e-01 -4.26046759e-01 -1.02170730e+00
5.04021309e-02 -8.88234615e-01 -9.97080028e-01 3.78894806e-02
-6.97022200e-01 -6.75652862e-01 -8.32876191e-03 6.26244722e-03] | [3.419140100479126, 1.4863507747650146] |
25c4ba4d-6003-462e-81c5-90ace0601bd9 | sub-8-bit-quantization-of-streaming-keyword | 2207.06920 | null | https://arxiv.org/abs/2207.06920v2 | https://arxiv.org/pdf/2207.06920v2.pdf | Sub 8-Bit Quantization of Streaming Keyword Spotting Models for Embedded Chipsets | We propose a novel 2-stage sub 8-bit quantization aware training algorithm for all components of a 250K parameter feedforward, streaming, state-free keyword spotting model. For the 1st-stage, we adapt a recently proposed quantization technique using a non-linear transformation with tanh(.) on dense layer weights. In the 2nd-stage, we use linear quantization methods on the rest of the network, including other parameters (bias, gain, batchnorm), inputs, and activations. We conduct large scale experiments, training on 26,000 hours of de-identified production, far-field and near-field audio data (evaluating on 4,000 hours of data). We organize our results in two embedded chipset settings: a) with commodity ARM NEON instruction set and 8-bit containers, we present accuracy, CPU, and memory results using sub 8-bit weights (4, 5, 8-bit) and 8-bit quantization of rest of the network; b) with off-the-shelf neural network accelerators, for a range of weight bit widths (1 and 5-bit), while presenting accuracy results, we project reduction in memory utilization. In both configurations, our results show that the proposed algorithm can achieve: a) parity with a full floating point model's operating point on a detection error tradeoff (DET) curve in terms of false detection rate (FDR) at false rejection rate (FRR); b) significant reduction in compute and memory, yielding up to 3 times improvement in CPU consumption and more than 4 times improvement in memory consumption. | ['Santosh Kumar Cheekatmalla', 'Shiv Vitaladevuni', 'Nikko Strom', 'Alex Escott', 'Yuzong Liu', 'Sree Hari Krishnan Parthasarathi', 'Lu Zeng'] | 2022-07-13 | null | null | null | null | ['keyword-spotting'] | ['speech'] | [ 3.14688325e-01 -1.84524000e-01 -2.08182767e-01 -3.04169744e-01
-5.51820755e-01 -4.30018336e-01 1.30635411e-01 4.10558313e-01
-9.93756533e-01 4.02272642e-01 -4.53671589e-02 -8.12383413e-01
-7.47342631e-02 -6.00665987e-01 -6.78883612e-01 -6.30256176e-01
-4.97061312e-01 -5.47081754e-02 4.35448438e-01 -1.09810844e-01
1.77406117e-01 4.33011174e-01 -1.67604983e+00 3.27070206e-01
1.13336153e-01 1.73980558e+00 -3.34057897e-01 1.11863732e+00
3.39948595e-01 3.65098894e-01 -8.02225113e-01 -4.57400441e-01
5.53989470e-01 8.00018758e-02 -5.33668935e-01 -5.70353568e-01
4.08058912e-01 -6.62019551e-01 -3.82252991e-01 1.19826758e+00
8.86739671e-01 -1.13292798e-01 3.99590045e-01 -1.16140938e+00
-9.43725631e-02 9.05251086e-01 -4.00238693e-01 5.30532300e-01
-2.49472678e-01 5.79660892e-01 7.74188876e-01 -6.57233179e-01
1.87113951e-03 1.01414990e+00 8.86522293e-01 4.20806408e-01
-1.09669185e+00 -1.06708884e+00 -4.50269997e-01 7.72926211e-02
-1.65695059e+00 -5.38492441e-01 1.53997451e-01 -2.46638834e-01
1.59565222e+00 1.82290062e-01 6.85246646e-01 9.49385285e-01
6.52730227e-01 2.09148407e-01 6.13971293e-01 -4.79352355e-01
7.72764504e-01 1.16515219e-01 3.51954639e-01 6.52513206e-01
3.68101001e-01 1.78005755e-01 -7.67272890e-01 -2.86819518e-01
7.92273998e-01 -2.86000937e-01 -1.18301533e-01 2.39949405e-01
-1.02147293e+00 8.42122972e-01 2.29205191e-01 1.77546754e-01
-3.28872114e-01 7.12070465e-01 8.87071848e-01 4.23104376e-01
1.36389256e-01 2.30187595e-01 -6.79471493e-01 -5.94095111e-01
-1.10775578e+00 1.32223099e-01 7.37043321e-01 9.63964939e-01
4.94684905e-01 6.23080790e-01 -4.28061217e-01 5.36107481e-01
2.07514346e-01 3.63963634e-01 1.19605327e+00 -7.68322587e-01
5.07022619e-01 1.18610121e-01 -2.00505510e-01 -5.47631323e-01
-6.55913949e-01 -5.76770782e-01 -1.09973931e+00 1.48804635e-01
1.69371605e-01 -4.41083729e-01 -1.06622934e+00 1.41002512e+00
1.95485160e-01 2.12355644e-01 2.53095448e-01 7.93048501e-01
6.29755855e-01 7.40120053e-01 8.84608775e-02 -1.29392058e-01
1.78748608e+00 -8.62374485e-01 -5.74489951e-01 -3.30607966e-02
7.23874152e-01 -5.87919652e-01 1.04975724e+00 8.27930689e-01
-1.07070947e+00 -8.15546572e-01 -1.68254447e+00 -1.17681809e-01
-3.45050991e-01 3.72390956e-01 5.77761531e-01 1.04203951e+00
-1.25370646e+00 8.96283209e-01 -1.04371226e+00 2.23925505e-02
2.10711986e-01 8.79913688e-01 4.15491452e-03 3.36177826e-01
-1.34081805e+00 3.32787454e-01 7.80586123e-01 -9.30166841e-02
-7.82134473e-01 -9.22633410e-01 -6.03786588e-01 4.59391803e-01
-1.13162935e-01 -3.77823591e-01 1.27561986e+00 -6.90938950e-01
-1.67162538e+00 3.28982979e-01 3.12006831e-01 -1.05693305e+00
-6.87331706e-02 -2.58498788e-01 -6.48113191e-01 4.77531105e-02
-6.06684983e-01 8.60601425e-01 8.56825650e-01 -1.90892354e-01
-7.12961018e-01 -2.65835851e-01 -2.67962068e-01 -2.32450560e-01
-9.86591756e-01 -1.64608255e-01 -2.50827134e-01 -7.51190007e-01
-2.13550210e-01 -6.31019711e-01 -1.94374509e-02 9.33086500e-02
-1.01298124e-01 -3.47934105e-02 5.88002443e-01 -6.82135880e-01
1.50461781e+00 -2.40236807e+00 -6.46491647e-01 3.03273857e-01
8.18832219e-02 5.99033594e-01 1.67623103e-01 7.86335692e-02
-4.62183319e-02 -4.40711305e-02 -2.20277384e-02 -4.02962774e-01
1.34162039e-01 5.22249416e-02 -2.80672044e-01 4.37024355e-01
2.37133875e-01 5.94034374e-01 -6.30124152e-01 1.66472029e-02
2.83235274e-02 5.51342130e-01 -7.25683331e-01 -1.08896324e-03
2.14695618e-01 -4.04461384e-01 4.14191484e-02 4.92589086e-01
4.68944967e-01 -2.55210221e-01 4.06404585e-02 -5.69227159e-01
-8.90077651e-02 3.92502666e-01 -1.40709412e+00 1.66516829e+00
-4.54664528e-01 7.95876801e-01 1.89478278e-01 -7.94019938e-01
9.02018785e-01 3.85931820e-01 6.82574362e-02 -7.56423175e-01
4.73022938e-01 1.96794063e-01 1.54244244e-01 -7.95272067e-02
6.80856705e-01 9.22431350e-02 -9.79731698e-03 2.28559136e-01
5.79580307e-01 2.53373891e-01 2.15310138e-02 -3.16301584e-02
1.30457389e+00 -5.13196707e-01 1.33128464e-01 -4.92724001e-01
7.87078142e-02 -1.83528036e-01 1.70920357e-01 6.72280014e-01
-2.39409029e-01 2.71414608e-01 5.33410132e-01 -4.46721107e-01
-1.17426956e+00 -8.26424122e-01 -4.38997328e-01 1.20843887e+00
-3.88415903e-01 -5.96932113e-01 -8.68216813e-01 1.40143707e-01
6.06751367e-02 6.32512331e-01 -4.20756668e-01 -4.87943530e-01
-4.47531760e-01 -1.06114876e+00 1.31943285e+00 6.29952133e-01
5.89053988e-01 -6.65834606e-01 -1.27359796e+00 2.79191375e-01
8.41871679e-01 -1.12142885e+00 -2.06161544e-01 1.16562581e+00
-1.02930462e+00 -3.70288163e-01 -5.39931059e-01 -4.54623550e-01
2.19348088e-01 -4.60986614e-01 8.30658257e-01 -1.83138624e-01
-3.95064265e-01 -2.94953249e-02 -1.44241989e-01 -3.24916661e-01
-2.20843136e-01 1.26108691e-01 4.32822794e-01 -2.72932887e-01
2.32588172e-01 -5.06639719e-01 -7.09467947e-01 -7.35341832e-02
-9.15474415e-01 -2.16370150e-01 6.86385095e-01 9.41724420e-01
4.67442274e-01 1.97609868e-02 3.26305211e-01 -4.21358466e-01
5.79485595e-01 -3.11406463e-01 -1.10162187e+00 -1.69271246e-01
-1.01046205e+00 3.65310669e-01 8.12781274e-01 -7.37431645e-01
-3.01792026e-01 1.60034955e-01 -4.61568952e-01 -6.90566361e-01
2.51840979e-01 5.98476827e-02 2.43604630e-01 -2.11641237e-01
8.67876887e-01 -1.58708114e-02 -1.04685746e-01 -3.21484178e-01
1.66170701e-01 1.00901926e+00 7.63908386e-01 -3.90152872e-01
3.74441326e-01 6.69568107e-02 -4.40018885e-02 -6.19703412e-01
-2.01615840e-01 -2.86355346e-01 -2.88917482e-01 1.79624736e-01
6.27120674e-01 -1.19078910e+00 -1.01838171e+00 3.43337536e-01
-8.30179691e-01 -4.66100007e-01 -3.65223140e-01 5.16757846e-01
-1.27227023e-01 6.68926463e-02 -8.19189370e-01 -7.91040123e-01
-1.14694047e+00 -1.28714418e+00 1.11579633e+00 2.44277433e-01
-1.48756281e-01 -6.17903829e-01 -4.12282497e-01 -2.70632476e-01
7.17104852e-01 -6.75309300e-02 9.60650623e-01 -7.23758817e-01
-2.41511315e-03 -1.14358678e-01 -1.31874219e-01 7.22558975e-01
-2.30738372e-01 -1.31379738e-01 -1.33951914e+00 -5.15542567e-01
8.47318098e-02 -4.22112644e-01 6.92202270e-01 3.96842003e-01
1.41121662e+00 -4.27954406e-01 -5.15254438e-02 8.98430705e-01
1.36782348e+00 1.39885336e-01 4.10039037e-01 -6.36324212e-02
3.44512373e-01 -1.12455390e-01 3.63955110e-01 8.17029953e-01
-2.62762576e-01 5.97321033e-01 4.30239320e-01 7.88329914e-02
-6.10837303e-02 -3.64387617e-03 5.01035810e-01 9.53027248e-01
3.67230088e-01 -3.14133525e-01 -8.57540488e-01 3.60458940e-01
-1.25779593e+00 -4.24627423e-01 1.68046534e-01 2.55947328e+00
1.14099550e+00 8.25143874e-01 1.20236099e-01 7.44343340e-01
3.44507664e-01 -3.25122446e-01 -7.44781792e-01 -9.76688623e-01
1.73110843e-01 9.05055881e-01 1.20883262e+00 1.49597779e-01
-9.61783230e-01 6.77486002e-01 6.45605469e+00 1.30067122e+00
-1.76425922e+00 2.12644696e-01 9.66103077e-01 -4.92967099e-01
2.55092621e-01 -5.46195686e-01 -1.22568691e+00 5.46109796e-01
2.06294155e+00 1.96887106e-01 5.45203984e-01 1.06830740e+00
-1.03995711e-01 1.02273319e-02 -1.19595504e+00 1.23694217e+00
-2.77291715e-01 -1.34791720e+00 -9.67498645e-02 2.00994965e-03
4.01401430e-01 1.84795409e-01 1.57762468e-01 5.45959651e-01
-2.51042694e-02 -1.04022908e+00 9.62563396e-01 -3.23068239e-02
1.48793900e+00 -1.10279071e+00 9.13216114e-01 1.83473706e-01
-1.01415181e+00 -2.45268881e-01 -5.07483423e-01 -1.83228403e-01
-3.03013027e-01 7.86457717e-01 -9.14563596e-01 9.68906581e-02
9.61430311e-01 -2.45129876e-02 -5.18382251e-01 7.42585123e-01
3.14152271e-01 1.04038835e+00 -6.44555986e-01 -3.90142232e-01
3.75176668e-01 4.15367216e-01 -1.29177347e-01 1.49804211e+00
3.76285762e-01 1.50250688e-01 -3.28360230e-01 4.58603650e-01
-1.31813571e-01 -2.18241274e-01 4.46413830e-02 4.55679512e-03
7.41325915e-01 1.36296737e+00 -8.78492951e-01 -5.22685945e-01
-2.64317952e-02 7.67297447e-01 -6.15796521e-02 -5.56542426e-02
-9.84433651e-01 -1.04021013e+00 6.77457571e-01 -2.02256609e-02
6.05634809e-01 -1.44970998e-01 -4.49670494e-01 -4.83070225e-01
-1.43017575e-01 -7.86876559e-01 7.83722103e-02 -4.81341004e-01
-6.86196446e-01 4.73461479e-01 -5.87484846e-03 -8.61866891e-01
-5.08560836e-01 -9.42279041e-01 -1.88874677e-01 8.95295858e-01
-1.04600143e+00 -3.31788182e-01 -1.06588855e-01 2.53876656e-01
2.55205780e-01 -1.25314951e-01 9.28161323e-01 6.95948184e-01
-3.64821076e-01 1.36897409e+00 3.21799725e-01 1.85922548e-01
3.86167407e-01 -9.09747303e-01 7.79196441e-01 4.98399705e-01
2.86967698e-02 7.49858797e-01 5.85557580e-01 -2.19272241e-01
-1.89891076e+00 -1.30524111e+00 2.47559905e-01 2.50683129e-01
6.62798584e-01 -8.06209981e-01 -6.62826896e-01 3.51749569e-01
-2.00284217e-02 3.27094853e-01 6.24537647e-01 -1.80034965e-01
-2.08790392e-01 -3.73544067e-01 -1.38742900e+00 3.68438214e-01
6.29283488e-01 -2.81820744e-01 8.84851068e-02 2.44619697e-01
1.03888988e+00 -7.90953159e-01 -1.16566086e+00 3.81084025e-01
7.18106210e-01 -8.04255545e-01 8.43048096e-01 -1.71970785e-01
1.19730346e-01 -1.59726247e-01 -3.22410405e-01 -8.12468350e-01
-8.99430886e-02 -7.82503843e-01 -3.35239828e-01 9.33931410e-01
3.60903263e-01 -4.73828346e-01 7.05279529e-01 2.37941325e-01
-4.49959934e-01 -1.12398195e+00 -1.46209025e+00 -6.31586432e-01
-6.07553534e-02 -7.43820548e-01 4.96034026e-01 4.43733037e-01
-4.36032340e-02 3.77210677e-01 -1.72931671e-01 1.55295014e-01
1.90543368e-01 -6.72551572e-01 2.33941138e-01 -7.67760754e-01
-7.44160891e-01 -4.26325351e-01 -7.59048283e-01 -9.31401193e-01
-3.61431003e-01 -5.41155517e-01 -7.13997334e-02 -5.13347507e-01
-3.45411181e-01 -3.63399655e-01 -4.32311267e-01 7.77937829e-01
2.07830176e-01 3.69251579e-01 -1.79116383e-01 -1.90530524e-01
-2.84696877e-01 2.78871030e-01 3.48632365e-01 -1.26559481e-01
4.49979529e-02 -3.14772606e-01 -3.33302915e-01 4.61166114e-01
6.77319288e-01 -4.22855318e-01 -3.88293177e-01 -5.08180976e-01
9.86844897e-02 -4.65363264e-02 3.03188056e-01 -1.85750854e+00
3.37409973e-01 4.30098087e-01 7.09316075e-01 -4.21244174e-01
6.16340995e-01 -5.54614902e-01 2.75393948e-02 9.49693739e-01
-3.53439093e-01 4.53292191e-01 8.24108064e-01 2.32112125e-01
1.17756091e-01 -4.65004623e-01 8.18345368e-01 3.31827193e-01
-6.02760494e-01 -6.30755723e-02 -3.49654585e-01 -1.53123602e-01
6.15892947e-01 -1.48287520e-01 -1.78585827e-01 4.98845242e-02
-4.55542892e-01 -4.31014337e-02 -4.63920040e-03 1.04496025e-01
4.63887781e-01 -1.24209201e+00 -3.15414608e-01 6.30373240e-01
-1.00382805e-01 -8.86954591e-02 9.22375917e-02 3.56217414e-01
-6.10305667e-01 6.21782184e-01 -2.37414062e-01 -7.08589375e-01
-1.08787012e+00 2.04174370e-01 2.43793517e-01 -9.56949443e-02
-4.51583594e-01 1.14646637e+00 -5.33025563e-01 1.62593633e-01
7.47846425e-01 -1.15126300e+00 1.10425904e-01 -2.12863401e-01
8.17885220e-01 4.50702161e-01 8.90512168e-01 -1.38678355e-02
-3.58767271e-01 2.48859733e-01 5.98637313e-02 -2.41569102e-01
9.46658492e-01 6.43745840e-01 2.90498108e-01 5.92868686e-01
1.38391948e+00 -4.05898243e-01 -9.64552045e-01 1.28237024e-01
-1.84914902e-01 2.63424307e-01 6.51534855e-01 -6.64660692e-01
-1.08084559e+00 1.12003684e+00 1.57678771e+00 -1.92855876e-02
1.36833465e+00 -5.28548419e-01 9.72316623e-01 4.10933137e-01
3.76181006e-01 -1.01096272e+00 -2.75308460e-01 5.60731053e-01
2.86448359e-01 -6.47235215e-01 1.18598066e-01 9.77596268e-02
-1.18408918e-01 1.34565043e+00 4.74050403e-01 -5.58339097e-02
7.27490246e-01 8.83006334e-01 -2.26394817e-01 8.41850601e-03
-1.17374849e+00 4.83807653e-01 5.78092337e-02 2.49527737e-01
4.07943249e-01 4.23416868e-02 -4.74447198e-02 9.15646017e-01
-5.59308350e-01 6.06332012e-02 2.84882158e-01 8.68554115e-01
-4.97212082e-01 -6.42061830e-01 -2.11216256e-01 7.76173055e-01
-5.92910945e-01 -5.00970066e-01 3.48524928e-01 5.54293513e-01
2.10688904e-01 6.55435264e-01 6.30039036e-01 -9.67659116e-01
2.37287521e-01 1.40434075e-02 1.52968690e-01 -2.77110487e-01
-1.19495559e+00 -8.88173133e-02 -8.34177062e-02 -5.64325929e-01
3.39733660e-01 -2.25262091e-01 -1.45307088e+00 -3.75195295e-01
-4.08479184e-01 2.42847539e-02 1.19208634e+00 5.66510260e-01
6.92505002e-01 8.68085980e-01 1.65932029e-01 -8.72865438e-01
-1.14169419e+00 -1.16322529e+00 -6.01854086e-01 -3.17629308e-01
3.49864542e-01 -4.47692126e-01 -4.05454904e-01 -2.37740148e-02] | [8.520773887634277, 2.9727752208709717] |
3a7096c2-2f17-405b-90e9-f11efc3a469d | history-based-unsupervised-data-oriented | null | null | https://aclanthology.org/R13-1059 | https://aclanthology.org/R13-1059.pdf | History Based Unsupervised Data Oriented Parsing | null | ['Gholamreza Ghasem-Sani', 'Mohsen Mesgar'] | 2013-09-01 | history-based-unsupervised-data-oriented-1 | https://aclanthology.org/R13-1059 | https://aclanthology.org/R13-1059.pdf | ranlp-2013-9 | ['lexical-analysis'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.387378215789795, 3.6384708881378174] |
a26f2933-6b08-4274-840c-1652c24f58f2 | extreme-learning-machine-based-heterogeneous | null | null | https://ieeexplore.ieee.org/document/8700294 | https://ieeexplore.ieee.org/document/8700294 | Extreme Learning Machine-Based Heterogeneous Domain Adaptation for Classification of Hyperspectral Images | An extreme learning machine (ELM)-based heterogeneous
domain adaptation (HDA) algorithm is proposed for the
classification of remote sensing images. In the adaptive ELM
network, one hidden layer is used for the source data to provide
the random features, whereas two hidden layers are set for target
data to produce the random features as well as a transformation
matrix. DA is achieved by constraining both the source data and
the transformed target data to share the same output weights.
Moreover, manifold regularization is adopted to preserve the local
geometry of unlabeled target data. The proposed ELM-based
HDA (EHDA) method is applied to cross-domain classification
of remote sensing images, and the experimental results using
multisensor remote sensing images demonstrate the effectiveness
of the proposed approach. | ['Li Zhou', 'Li Ma'] | 2019-11-01 | null | null | null | ieee-geoscience-and-remote-sensing-letters | ['classification-of-hyperspectral-images'] | ['computer-vision'] | [ 1.64974183e-01 -1.94654986e-01 5.12163304e-02 -4.02976573e-01
-4.44183856e-01 -5.11626601e-02 2.79762566e-01 -2.10779458e-01
-2.04177856e-01 4.40134704e-01 -1.51263848e-01 -4.01318707e-02
-3.78138989e-01 -9.50802088e-01 -2.34373361e-01 -1.34507906e+00
2.84560956e-02 2.15754330e-01 -2.70628780e-01 -1.45799845e-01
-1.30722955e-01 5.55325150e-01 -1.28263903e+00 -7.81184658e-02
1.11511314e+00 1.04937172e+00 3.21407855e-01 2.03598544e-01
-2.78487504e-01 6.16608918e-01 -2.40684554e-01 2.46999040e-01
4.54545766e-01 -4.12273139e-01 -3.02337676e-01 4.58878160e-01
-2.46018842e-02 1.16443358e-01 -2.27180779e-01 1.21144533e+00
7.81795204e-01 5.93962193e-01 1.00422549e+00 -1.22647536e+00
-9.65703011e-01 3.42928261e-01 -7.09696710e-01 -1.09976381e-01
-2.88464248e-01 -1.14391267e-01 6.45709574e-01 -1.20549607e+00
4.82175767e-01 1.19883764e+00 3.33006263e-01 3.81223172e-01
-1.03157866e+00 -6.50156617e-01 7.29638189e-02 2.49106511e-01
-1.75616598e+00 -2.72069275e-01 1.12293684e+00 -3.98624122e-01
4.77861345e-01 3.83534543e-02 5.33222079e-01 4.49691117e-01
2.64018357e-01 5.58529437e-01 9.84830201e-01 -6.10065997e-01
6.80897772e-01 2.79362470e-01 7.43910298e-02 3.92237186e-01
1.67395156e-02 1.53231800e-01 7.20165074e-02 -2.69488335e-01
5.49906671e-01 3.97662044e-01 -4.66370694e-02 -5.35065889e-01
-7.40812600e-01 1.19149208e+00 8.08863878e-01 3.55742306e-01
-6.47023499e-01 -6.89816952e-01 1.12851858e-02 4.31249648e-01
5.83903909e-01 -1.64336748e-02 -1.07871294e-01 8.34450841e-01
-6.42830193e-01 -2.65711606e-01 2.58025616e-01 7.53028691e-01
1.05123544e+00 4.60607737e-01 1.17030308e-01 1.04268742e+00
6.48424268e-01 7.90466607e-01 7.29467154e-01 -9.49192405e-01
4.03107792e-01 7.70951390e-01 5.72053231e-02 -1.12287593e+00
-3.69920343e-01 -4.65965778e-01 -1.23086023e+00 5.84659040e-01
-9.33697671e-02 -3.73103052e-01 -8.57770026e-01 1.37858498e+00
6.31118476e-01 1.64728582e-01 6.33329809e-01 1.07950449e+00
5.73737919e-01 1.10400736e+00 4.31671739e-02 -3.06265295e-01
5.98076344e-01 -5.71048379e-01 -4.84222502e-01 -5.71853854e-02
3.32296550e-01 -4.41151857e-01 8.64450634e-01 6.10570461e-02
-5.59387505e-01 -7.06148803e-01 -1.09692740e+00 3.58204097e-01
-3.63966793e-01 6.33433610e-02 2.14807212e-01 2.13097408e-01
-5.20244539e-01 1.42623112e-01 -6.00127757e-01 -2.72085398e-01
3.38193536e-01 2.85977781e-01 -2.91791350e-01 -2.46007591e-01
-1.10575759e+00 5.71121693e-01 6.41916335e-01 3.54213476e-01
-5.91661632e-01 -3.78535658e-01 -9.41764832e-01 8.25492218e-02
-2.00133920e-01 -2.84804285e-01 3.91802192e-01 -1.17777455e+00
-1.67264116e+00 7.56822288e-01 1.62886679e-01 -5.77996746e-02
1.80129245e-01 2.89369226e-01 -8.15491617e-01 1.56715289e-02
9.20989644e-03 4.15016264e-01 1.02957988e+00 -1.32115412e+00
-4.75967944e-01 -5.88622749e-01 -7.85633385e-01 4.18368012e-01
-6.67562902e-01 -2.66497463e-01 2.90055007e-01 -5.42890251e-01
7.35840023e-01 -7.38995969e-01 -3.20760787e-01 -4.15195040e-02
-2.47745469e-01 1.27776280e-01 1.30538583e+00 -6.94077075e-01
8.27972829e-01 -2.68587208e+00 4.07347023e-01 7.20286071e-01
2.91714584e-03 2.04456553e-01 -3.72774333e-01 1.74420461e-01
-1.86265945e-01 -3.86567116e-01 -6.40211523e-01 2.11590733e-02
-1.29012659e-01 1.95417255e-01 -2.93909431e-01 7.28125513e-01
-3.81898805e-02 4.75800663e-01 -6.87551498e-01 -5.33334136e-01
5.37015021e-01 5.79031825e-01 -2.20878907e-02 3.17062557e-01
-1.29806489e-01 7.59958386e-01 -6.77172601e-01 4.75217193e-01
9.32118893e-01 -1.41237155e-01 1.62496448e-01 2.60562096e-02
-1.28984571e-01 -7.21614838e-01 -1.50765753e+00 1.37315667e+00
-3.34134132e-01 1.69343397e-01 1.70226127e-01 -8.37724149e-01
1.43735433e+00 3.57757926e-01 7.28828371e-01 -7.32286811e-01
4.20500964e-01 1.37862295e-01 -1.72736302e-01 -6.32602274e-01
-1.35224976e-03 -2.56261975e-01 2.04278342e-02 3.85977119e-01
-8.66224021e-02 -1.01481881e-02 -5.07450879e-01 -3.04256380e-01
2.15975866e-01 1.44111523e-02 2.42701277e-01 -2.07024947e-01
7.90431261e-01 1.65922672e-01 9.31864738e-01 3.81626785e-01
-1.60515353e-01 4.32254851e-01 -4.00929481e-01 -6.38512135e-01
-1.27500737e+00 -1.01976633e+00 -2.35003084e-01 9.40288603e-01
2.30265796e-01 5.17994940e-01 -6.00849509e-01 -4.61113662e-01
-1.31368890e-01 7.30048954e-01 -3.74262631e-01 -4.13754910e-01
-1.17748782e-01 -7.99351752e-01 1.26066402e-01 2.29305789e-01
1.03194177e+00 -9.38171685e-01 -1.99301615e-01 3.08588445e-01
-7.74189308e-02 -5.98929048e-01 -1.66399702e-01 -8.94154534e-02
-1.02472436e+00 -6.07399106e-01 -7.34377027e-01 -1.12481177e+00
7.49294519e-01 1.35366485e-01 2.60158300e-01 -2.81226665e-01
-3.18266042e-02 -5.08990027e-02 -2.11682454e-01 -1.90712079e-01
-4.46073055e-01 -4.30953354e-02 6.24405518e-02 7.80725479e-01
4.97399658e-01 -7.65809715e-01 -3.49614829e-01 2.37018839e-01
-1.00892723e+00 -1.91138878e-01 6.94275737e-01 1.09494889e+00
6.87129438e-01 4.85545158e-01 5.49391389e-01 -4.41673487e-01
3.31362695e-01 -7.43897498e-01 -6.80032134e-01 3.73172730e-01
-6.94375813e-01 -1.29737660e-01 7.77549028e-01 -7.15576291e-01
-1.32648897e+00 3.88904601e-01 2.28731900e-01 -6.95384264e-01
-4.04713362e-01 5.66341817e-01 -5.02502143e-01 -2.31241405e-01
9.00661707e-01 4.56038624e-01 1.16749763e-01 -4.94359642e-01
1.60741821e-01 1.17949533e+00 3.95079017e-01 -2.51425803e-01
8.83202672e-01 5.93663931e-01 -8.49072076e-03 -9.94074166e-01
-7.49314964e-01 -3.81853819e-01 -7.46815324e-01 -1.08284615e-01
1.00139141e+00 -1.04589581e+00 -1.65804386e-01 7.44110525e-01
-3.85136843e-01 -1.50531501e-01 -4.65251267e-01 9.56191242e-01
-4.34866965e-01 9.31568667e-02 -4.17731583e-01 -8.76567483e-01
-4.87061352e-01 -7.24903345e-01 6.38434947e-01 4.23249692e-01
3.89780462e-01 -1.19652247e+00 1.43879727e-01 1.23314351e-01
2.62182951e-01 4.09812421e-01 1.14028347e+00 -5.77571392e-01
-8.83932337e-02 -8.75501782e-02 1.57847166e-01 4.46869642e-01
3.49656999e-01 -2.75476761e-02 -7.40957797e-01 -4.67436612e-01
4.11616206e-01 -2.70188749e-01 5.54357827e-01 4.71568853e-01
6.25829041e-01 -1.98142335e-01 -2.25535899e-01 7.97507167e-01
1.72770834e+00 2.99589813e-01 5.98053455e-01 3.90760899e-01
7.29248226e-01 5.01623988e-01 5.67986250e-01 5.35485268e-01
2.16307849e-01 1.33271217e-01 2.25268006e-01 -5.38055539e-01
3.57735574e-01 -1.77084550e-01 1.97058782e-01 8.49767685e-01
2.43220717e-01 -5.71260899e-02 -8.95550728e-01 4.19364423e-01
-1.70010769e+00 -8.92980814e-01 -1.13003284e-01 2.16668344e+00
3.33283663e-01 -4.00647491e-01 -2.67016143e-01 2.18074754e-01
1.15690017e+00 7.07869977e-02 -9.19904232e-01 5.02863750e-02
-4.81831849e-01 -8.01569968e-02 1.62460119e-01 5.63027442e-01
-1.05799747e+00 8.47875297e-01 5.50445557e+00 6.67805254e-01
-1.47722530e+00 1.48094401e-01 4.34694469e-01 1.43358842e-01
-2.03275174e-01 3.04031614e-02 -3.34722340e-01 3.66858184e-01
4.59691793e-01 2.94105411e-02 5.60973108e-01 8.81097496e-01
3.91358674e-01 2.27899924e-01 -4.03453499e-01 9.66980577e-01
-1.18087918e-01 -8.41708720e-01 3.19281191e-01 7.08052292e-02
1.12356496e+00 1.42175779e-01 2.52514243e-01 1.13459572e-01
5.26741147e-01 -6.32112086e-01 4.09397244e-01 7.21401632e-01
7.12880194e-01 -9.47643697e-01 5.74840486e-01 6.14331722e-01
-1.15259159e+00 -5.72846651e-01 -7.73234189e-01 4.19197418e-02
-1.75719246e-01 5.65152466e-01 -6.67560160e-01 6.42870426e-01
5.74971497e-01 7.32115090e-01 -2.79388368e-01 6.98965788e-01
-8.02968722e-03 4.23527718e-01 -3.98256063e-01 2.88616240e-01
2.79535949e-01 -7.79315114e-01 7.10693955e-01 4.73170280e-01
4.03370231e-01 2.72784799e-01 5.45083165e-01 8.28721404e-01
7.11184442e-02 5.08063972e-01 -6.61988556e-01 2.55923569e-01
6.14822805e-01 1.19750714e+00 -4.04698461e-01 -2.53788799e-01
-1.14087008e-01 9.15588677e-01 2.23265141e-01 7.96976209e-01
-5.20346820e-01 -4.27800059e-01 4.82395381e-01 -3.65069211e-01
3.42917621e-01 -2.46606082e-01 -2.35272497e-01 -1.34551799e+00
-2.18865335e-01 -5.39805412e-01 7.43638337e-01 -9.70728815e-01
-1.50118446e+00 5.30609012e-01 -3.14315051e-01 -1.61185157e+00
-9.47615132e-02 -2.28590131e-01 -8.13241720e-01 1.14039505e+00
-1.81543171e+00 -1.18057370e+00 -5.45947671e-01 1.05416965e+00
9.29582790e-02 -6.24096334e-01 8.41207445e-01 1.24394283e-01
-6.96540952e-01 3.37705642e-01 9.80108857e-01 2.85992265e-01
3.43069553e-01 -6.95009530e-01 -3.17540616e-01 7.60971308e-01
-8.91476572e-02 2.46484488e-01 4.29891557e-01 -5.01049101e-01
-1.18952036e+00 -1.41574979e+00 6.59439445e-01 2.49804214e-01
3.09506118e-01 -8.90238360e-02 -9.46843922e-01 7.57150054e-01
-1.57476202e-01 1.78677469e-01 9.19827521e-01 -4.11487192e-01
-1.54309973e-01 -3.20481777e-01 -1.46123946e+00 3.57343346e-01
2.87217438e-01 -2.66099274e-01 -6.06384516e-01 3.56028914e-01
4.23398346e-01 -4.44892235e-02 -1.26291764e+00 2.98859417e-01
3.03905368e-01 -5.92414975e-01 7.81036377e-01 -5.46299219e-01
6.09597042e-02 -6.17685378e-01 -5.18448293e-01 -1.44386685e+00
-7.73367226e-01 5.90900816e-02 2.83007443e-01 1.15657794e+00
2.71345496e-01 -9.03934598e-01 5.42020619e-01 3.44333112e-01
9.89951286e-03 -2.83733010e-01 -9.21994686e-01 -8.69256139e-01
3.53467882e-01 1.22409523e-01 7.41375029e-01 1.33068240e+00
-3.81224692e-01 4.08285350e-01 -2.83387661e-01 7.50041068e-01
8.96420062e-01 4.64810401e-01 3.94761264e-01 -1.47022569e+00
-7.07199946e-02 -8.10964927e-02 -3.77708793e-01 -3.26526701e-01
3.25135529e-01 -1.01161051e+00 -1.60571694e-01 -1.34680605e+00
5.34420125e-02 -6.62962496e-01 -4.79502797e-01 3.71713430e-01
4.10119593e-02 1.71364158e-01 5.66314720e-02 6.57233179e-01
-1.03745848e-01 1.03226268e+00 9.33665276e-01 -3.03341120e-01
-7.16449618e-01 1.23744071e-01 -3.44841957e-01 5.67105472e-01
8.44996393e-01 -5.51028609e-01 -3.15746188e-01 -5.42780042e-01
-2.94840485e-01 9.91160125e-02 2.01447308e-01 -1.08673024e+00
1.70580074e-01 -4.35832977e-01 8.40844393e-01 -3.05517077e-01
1.72449827e-01 -1.22237051e+00 3.59256625e-01 4.62145209e-01
-2.04365730e-01 -4.68595684e-01 -2.74609804e-01 4.61487502e-01
-4.02464509e-01 -1.10672615e-01 1.12891066e+00 -3.78634245e-03
-7.13222802e-01 5.33703387e-01 -1.61283121e-01 -2.42416427e-01
1.19053173e+00 -3.27309847e-01 1.16250150e-01 -1.57785594e-01
-1.06596243e+00 3.01453233e-01 4.95526254e-01 1.35887057e-01
7.89957106e-01 -1.59758341e+00 -1.15226519e+00 6.77047014e-01
3.27611446e-01 -3.13073630e-03 5.24281383e-01 5.25987208e-01
-3.01416039e-01 -1.76851481e-01 -4.10525799e-01 -6.97283983e-01
-9.79534805e-01 6.19368732e-01 7.17658997e-01 1.72392502e-01
-7.52079129e-01 4.53080952e-01 -1.88574605e-02 -9.43000019e-01
1.09276390e-02 3.08099329e-01 -4.48900044e-01 -2.79067829e-02
4.97686446e-01 4.59205031e-01 -1.21615857e-01 -8.81575763e-01
-1.95380762e-01 7.06048310e-01 3.56045038e-01 -2.23246709e-01
1.64732265e+00 -4.08866853e-01 -2.05106288e-01 5.50199449e-01
1.27421510e+00 -2.56831914e-01 -1.44244063e+00 -6.82629704e-01
-2.11318567e-01 -1.09029062e-01 3.49077135e-01 -5.95119119e-01
-1.19289565e+00 8.59648287e-01 1.10875380e+00 -1.24751292e-01
1.38419843e+00 -2.81114906e-01 5.21403015e-01 5.32937407e-01
1.95069864e-01 -1.24351215e+00 -1.85072929e-01 3.81152064e-01
6.17209196e-01 -1.24986386e+00 -2.26301715e-01 9.79531333e-02
-7.42124856e-01 1.11186934e+00 4.73080218e-01 -1.76382512e-01
7.95514226e-01 -6.27799854e-02 3.36207598e-01 -1.52258769e-01
-2.13699907e-01 5.29123209e-02 -7.78161958e-02 7.16733634e-01
-1.89871818e-01 4.63781245e-02 1.98268369e-01 2.94499964e-01
2.06316710e-01 -1.64530128e-01 1.76426157e-01 9.42347467e-01
-5.50593019e-01 -7.08850563e-01 -8.50001574e-01 2.05145061e-01
3.71248797e-02 1.03634626e-01 -2.39980474e-01 4.67023581e-01
9.34108421e-02 9.88393188e-01 1.73855543e-01 -5.46627879e-01
1.04999073e-01 2.86412627e-01 1.53668532e-02 -3.73757243e-01
-2.09556982e-01 1.52825683e-01 -7.50577152e-01 4.99792583e-02
-3.30823332e-01 -6.33947134e-01 -1.66152620e+00 -2.16538414e-01
-4.45355654e-01 4.28590208e-01 7.50463128e-01 9.05294240e-01
5.19845665e-01 1.25739604e-01 1.59297156e+00 -6.82381451e-01
-6.66891038e-01 -1.03690290e+00 -1.08464229e+00 3.74720961e-01
2.44547158e-01 -4.21459854e-01 -5.55007637e-01 1.71045661e-01] | [9.966907501220703, -1.5246762037277222] |
19f101d4-5642-4671-90f5-161ee464979e | weakly-supervised-conditional-embedding-for | 2306.02928 | null | https://arxiv.org/abs/2306.02928v1 | https://arxiv.org/pdf/2306.02928v1.pdf | Weakly-Supervised Conditional Embedding for Referred Visual Search | This paper presents a new approach to image similarity search in the context of fashion, a domain with inherent ambiguity due to the multiple ways in which images can be considered similar. We introduce the concept of Referred Visual Search (RVS), where users provide additional information to define the desired similarity. We present a new dataset, LAION-RVS-Fashion, consisting of 272K fashion products with 842K images extracted from LAION, designed explicitly for this task. We then propose an innovative method for learning conditional embeddings using weakly-supervised training, achieving a 6% increase in Recall at one (R@1) against a gallery with 2M distractors, compared to classical approaches based on explicit attention and filtering. The proposed method demonstrates robustness, maintaining similar R@1 when dealing with 2.5 times as many distractors as the baseline methods. We believe this is a step forward in the emerging field of Referred Visual Search both in terms of accessible data and approach. Code, data and models are available at https://www.github.com/Simon-Lepage/CondViT-LRVSF . | ['David Picard', 'Jérémie Mary', 'Simon Lepage'] | 2023-06-05 | null | null | null | null | ['image-similarity-search'] | ['computer-vision'] | [-1.62587501e-02 -4.78149146e-01 -4.18689758e-01 -2.60852188e-01
-9.93746996e-01 -9.60263669e-01 8.71926606e-01 4.99721587e-01
-4.70934778e-01 1.04431644e-01 3.01538050e-01 -7.65118301e-02
-1.21759780e-01 -2.70748377e-01 -6.82334185e-01 -4.17722434e-01
2.21153423e-01 3.03366870e-01 1.58378929e-01 -2.55605459e-01
5.47109485e-01 4.15901810e-01 -1.73473251e+00 4.63487864e-01
4.74095583e-01 1.37397754e+00 4.77305561e-01 7.12243259e-01
2.17556283e-01 5.87053418e-01 -4.78282779e-01 -7.73760140e-01
3.60248834e-01 -1.84656620e-01 -9.15837169e-01 1.29478559e-01
1.07599139e+00 6.25407090e-03 -4.10180628e-01 1.05874121e+00
6.18695557e-01 2.99358726e-01 6.91496849e-01 -1.28119111e+00
-1.61800921e+00 8.41062814e-02 -3.95042747e-01 5.53907990e-01
5.85029125e-01 3.68133634e-01 1.21230614e+00 -1.25938833e+00
7.00917840e-01 1.18837595e+00 5.00205576e-01 2.78318346e-01
-1.41157615e+00 -4.24703419e-01 -6.19655512e-02 5.49213946e-01
-1.53718972e+00 -5.24130583e-01 6.66429639e-01 -4.54733700e-01
6.67389929e-01 6.76591754e-01 3.97408426e-01 1.31277347e+00
-1.09940238e-01 9.02436376e-01 1.23294258e+00 -5.99309862e-01
1.08625732e-01 7.23653316e-01 3.35876793e-01 4.55585122e-01
1.47797644e-01 2.03203768e-01 -6.46190286e-01 -6.42606169e-02
5.59724510e-01 -1.51592102e-02 -6.34724349e-02 -6.19434059e-01
-1.11723483e+00 9.96168554e-01 8.25895250e-01 3.81186604e-01
-1.98205784e-01 -3.85630690e-02 4.98070568e-01 5.79132855e-01
4.52479243e-01 7.60502219e-01 -7.73271918e-02 2.21568197e-01
-8.11680913e-01 1.71374857e-01 3.60824585e-01 1.20864415e+00
3.99253756e-01 -2.83835113e-01 -5.27265072e-01 9.62250531e-01
2.16145635e-01 5.31150639e-01 4.40746576e-01 -8.49842906e-01
1.03281967e-01 3.86660486e-01 2.19854683e-01 -1.25919569e+00
1.08141743e-01 -4.26993728e-01 -5.24781525e-01 3.01062614e-01
3.03485364e-01 7.05957532e-01 -9.08422232e-01 1.40904438e+00
1.82268083e-01 -4.11655575e-01 -1.81159243e-01 1.17364609e+00
1.27848315e+00 2.82035172e-01 1.86840683e-01 1.89308792e-01
1.82321560e+00 -1.28520751e+00 -6.24286056e-01 -2.81847149e-01
2.33675435e-01 -1.36162114e+00 1.47788310e+00 2.06573293e-01
-1.07757819e+00 -9.16551888e-01 -1.02582490e+00 -2.93544620e-01
-9.11187768e-01 3.03687215e-01 4.07569855e-01 6.27820373e-01
-1.23090363e+00 3.40713739e-01 1.70381763e-03 -9.31309938e-01
5.07003546e-01 8.00657719e-02 -4.80795711e-01 -3.33461285e-01
-1.01872909e+00 1.07323503e+00 1.80294544e-01 -9.39249992e-02
-8.22313726e-01 -5.78981817e-01 -8.46989036e-01 -1.11110352e-01
3.70618701e-01 -4.42746520e-01 1.07931864e+00 -8.59772921e-01
-8.67596924e-01 1.45095098e+00 -2.21406436e-03 -1.80362582e-01
4.22594130e-01 -4.05206054e-01 -6.66406035e-01 2.28534564e-01
3.69676232e-01 7.63049126e-01 8.51944208e-01 -1.50772500e+00
-2.66364753e-01 -2.23844245e-01 2.35548198e-01 9.81825292e-02
-1.95707649e-01 3.91652018e-01 -7.26436198e-01 -8.03513706e-01
-3.61994833e-01 -1.01269937e+00 -1.43111110e-01 3.39275271e-01
-2.78652042e-01 -3.36542010e-01 5.46458840e-01 -6.40798807e-01
1.06265843e+00 -2.36231637e+00 -6.23897463e-02 6.90065250e-02
2.09114105e-01 5.22755384e-01 -5.00932574e-01 5.92394888e-01
-7.76641145e-02 5.52929342e-02 2.23077655e-01 -3.52994382e-01
2.43308917e-01 -1.93493038e-01 -1.70638431e-02 4.25534755e-01
1.64201513e-01 1.31050324e+00 -1.04239666e+00 -4.99413699e-01
3.73904288e-01 2.96262205e-01 -6.14590310e-02 2.14318901e-01
2.50526369e-02 1.92303509e-01 -1.56651020e-01 8.78667653e-01
7.35469699e-01 -5.46330810e-01 6.80578873e-02 -5.65903068e-01
3.47258616e-03 -2.98765123e-01 -9.51360524e-01 1.66422391e+00
-4.57290113e-01 7.52358735e-01 2.31497083e-02 -6.97467506e-01
8.73466730e-01 -1.76098034e-01 6.58513680e-02 -1.05265486e+00
2.53592700e-01 2.92849075e-02 -5.64108729e-01 -4.73925829e-01
8.40144753e-01 3.55019897e-01 -3.17408770e-01 1.68395340e-01
2.97128469e-01 2.91447751e-02 3.40970725e-01 4.31900591e-01
8.69182527e-01 -1.17993303e-01 3.65584791e-01 -3.79023820e-01
4.46635306e-01 1.38688073e-01 -2.69938968e-02 9.17418182e-01
-4.12135035e-01 6.07781589e-01 2.03093644e-02 -4.13596123e-01
-1.09804535e+00 -1.34604788e+00 -1.21689275e-01 1.27641046e+00
5.76888621e-01 -6.00660980e-01 -1.49680361e-01 -7.44902730e-01
2.16178373e-01 6.29998684e-01 -8.35736513e-01 -3.49403918e-01
-1.82837188e-01 -2.54148096e-01 1.93615302e-01 4.06171978e-01
4.88912404e-01 -1.05931568e+00 -4.12394285e-01 -4.24908668e-01
-6.45264313e-02 -1.15392816e+00 -1.04161656e+00 -3.70825939e-02
-3.41762394e-01 -1.22039402e+00 -8.91323864e-01 -1.15447032e+00
6.36007428e-01 5.98619044e-01 1.41588426e+00 -8.59136805e-02
-6.79287493e-01 8.11417282e-01 -4.64271158e-01 -4.56015356e-02
-4.81998980e-01 -2.60038227e-01 -8.54999349e-02 -1.32561810e-02
7.25042105e-01 1.27209788e-02 -9.05850232e-01 4.73748744e-01
-8.54014993e-01 -3.16540211e-01 6.85798824e-01 8.23791265e-01
6.72348499e-01 -3.97386819e-01 2.26678267e-01 -4.33462292e-01
8.38660836e-01 -5.23704052e-01 -4.64584172e-01 5.62207103e-01
-8.35095286e-01 -7.49442354e-02 3.53918701e-01 -6.40359879e-01
-7.29472041e-01 -3.84487696e-02 2.88250834e-01 -8.02041173e-01
-2.92055964e-01 6.87362207e-03 2.02608436e-01 -2.48324215e-01
7.92946398e-01 2.71272033e-01 -1.83258522e-02 -7.13354468e-01
8.29600275e-01 7.93920219e-01 5.43853164e-01 -1.16025612e-01
7.18797505e-01 2.31417894e-01 -2.44495869e-01 -5.25537550e-01
-6.07620180e-01 -8.82082403e-01 -6.14129782e-01 -4.84854609e-01
8.57193470e-01 -8.12652230e-01 -9.01993275e-01 1.17099896e-01
-8.65211546e-01 5.23554981e-02 -4.43397164e-01 4.00129646e-01
-4.65936035e-01 4.97177511e-01 -5.18921196e-01 -5.71511328e-01
-3.89481068e-01 -1.00482476e+00 1.06231976e+00 1.60772800e-01
-4.21576351e-01 -8.15173864e-01 -1.59872659e-02 6.71501696e-01
5.69655240e-01 8.41021910e-03 8.11223269e-01 -8.92983139e-01
-5.18472672e-01 -2.65604734e-01 -4.92852718e-01 4.48307097e-01
2.01466456e-01 -4.29268688e-01 -9.10146058e-01 -5.75794339e-01
-3.45516592e-01 -7.02468634e-01 9.34900582e-01 1.24107786e-01
9.82799768e-01 -1.96832865e-01 -2.16155022e-01 1.29621834e-01
1.72412765e+00 8.24482068e-02 4.84769344e-01 4.63508248e-01
3.79443288e-01 4.74892914e-01 9.47321951e-01 2.63736278e-01
9.67041552e-02 1.09118688e+00 4.91241366e-01 -2.96058059e-01
-6.51403546e-01 -3.57292771e-01 1.26603216e-01 5.74318171e-01
7.22639710e-02 -2.05593929e-01 -4.60691124e-01 8.64295185e-01
-1.74129760e+00 -1.00935769e+00 -5.43871708e-02 2.40860677e+00
5.81256866e-01 -1.59339175e-01 3.45113307e-01 -1.86325431e-01
8.50519001e-01 3.20066392e-01 -4.50376302e-01 -6.34991109e-01
-2.49130577e-01 -1.04778621e-03 4.85474616e-01 3.48336399e-01
-1.24710155e+00 8.00105989e-01 6.84991360e+00 1.13487220e+00
-7.48132408e-01 2.92324513e-01 5.91910601e-01 -2.42201418e-01
-1.58438131e-01 -2.38364220e-01 -6.02396309e-01 6.44381464e-01
7.57441163e-01 -3.64234149e-02 4.38167155e-01 9.21279728e-01
-9.10559148e-02 -1.45869330e-01 -1.07881534e+00 1.31697333e+00
6.21067524e-01 -1.25049043e+00 3.79589722e-02 -1.77176967e-01
5.88692427e-01 -2.31686741e-01 5.60873806e-01 2.86296070e-01
2.58580089e-01 -1.00990820e+00 8.01516771e-01 4.03736204e-01
8.86941135e-01 -4.15398836e-01 5.88981509e-01 -2.43627861e-01
-1.23848104e+00 -1.59821368e-03 -3.16857845e-01 4.44035649e-01
-1.48262521e-02 2.62515694e-01 -5.98182619e-01 6.24432027e-01
1.04578531e+00 6.09899879e-01 -1.15310311e+00 1.15178120e+00
2.22713739e-01 5.70112057e-02 -5.96432202e-02 -2.21090183e-01
1.47522300e-01 -3.24376218e-04 4.26778167e-01 1.25880945e+00
2.17276543e-01 -4.55369681e-01 1.85768068e-01 7.67904222e-01
-1.17380992e-01 3.50583166e-01 -7.22053349e-01 2.41363533e-02
4.47009712e-01 1.23975456e+00 -7.72516847e-01 -2.82863945e-01
-5.16082227e-01 1.50428855e+00 3.02192926e-01 3.11144888e-01
-8.69846106e-01 -3.69428426e-01 4.42629337e-01 1.23105235e-01
6.42673552e-01 1.67541783e-02 2.17447385e-01 -7.57300675e-01
1.75060689e-01 -8.98830652e-01 6.02118015e-01 -9.75922585e-01
-1.80627203e+00 7.12144315e-01 1.58084020e-01 -1.35229862e+00
2.47070774e-01 -7.49239981e-01 -2.92558044e-01 7.02546477e-01
-1.34017456e+00 -1.45772743e+00 -4.87474352e-01 4.57183331e-01
8.16926479e-01 -2.29622081e-01 7.89404988e-01 3.62281799e-01
-8.35826546e-02 9.03253436e-01 3.46648514e-01 -2.12290406e-01
1.05213463e+00 -1.06944406e+00 2.70107657e-01 4.71062750e-01
4.82589930e-01 5.90711474e-01 7.52810538e-01 -3.44877094e-01
-1.29584193e+00 -9.00675178e-01 1.03754473e+00 -8.00036788e-01
6.95388675e-01 -5.29424191e-01 -4.86579478e-01 4.53247786e-01
7.97751307e-01 8.59361514e-02 9.64095056e-01 5.77495135e-02
-7.06982076e-01 -1.73562273e-01 -1.30865502e+00 6.53130949e-01
1.22665501e+00 -6.94982231e-01 -5.73606133e-01 6.18699312e-01
7.01708376e-01 5.18101864e-02 -1.01259696e+00 1.47318318e-01
5.75603247e-01 -9.02021885e-01 1.31467664e+00 -5.81892431e-01
2.55118906e-01 -1.96767315e-01 -4.99325246e-01 -1.19244611e+00
-8.52519095e-01 -4.39087242e-01 2.36126091e-02 1.35980546e+00
1.61493838e-01 -3.08214456e-01 5.07020891e-01 2.97301382e-01
1.93100497e-01 -6.16520882e-01 -8.69322121e-01 -1.16219950e+00
-2.03720808e-01 -1.82656422e-01 3.69601637e-01 8.84320140e-01
-1.75372809e-01 3.86954367e-01 -5.85315645e-01 -7.38738179e-02
6.31491780e-01 3.05449724e-01 5.27633607e-01 -8.60962570e-01
-2.66354620e-01 -3.84918034e-01 -6.70132518e-01 -1.10950947e+00
-1.53139129e-01 -1.08306050e+00 -2.30062112e-01 -1.44887471e+00
4.27902818e-01 -2.04580635e-01 -5.87744534e-01 2.05561370e-01
1.20035084e-02 8.49509299e-01 5.65883994e-01 2.89455980e-01
-1.05967653e+00 3.96879673e-01 1.13451362e+00 -4.65133876e-01
1.33148775e-01 -2.34059229e-01 -8.35633397e-01 2.09536016e-01
8.06572199e-01 -2.55661696e-01 -3.29275817e-01 -2.94697613e-01
-1.22162230e-01 -4.99547064e-01 7.53186882e-01 -8.48753691e-01
3.26265544e-01 4.91794273e-02 5.27299464e-01 -6.37315750e-01
6.30780518e-01 -9.04532015e-01 2.63481200e-01 3.49531204e-01
-6.47382855e-01 5.68100333e-01 3.11706722e-01 7.07596183e-01
-1.40803874e-01 -3.72704506e-01 5.52514791e-01 -2.53720313e-01
-1.21980762e+00 6.07094727e-03 -1.79343581e-01 -8.50828215e-02
1.25512779e+00 -2.19610929e-01 -5.11021376e-01 -3.56563181e-01
-9.39622045e-01 2.28209957e-01 5.74447155e-01 9.60938931e-01
7.51302123e-01 -1.85235381e+00 -6.63879275e-01 1.30813032e-01
8.18532705e-01 -1.01865768e+00 3.62335593e-01 6.94052815e-01
-2.36553296e-01 5.30767739e-01 -3.05117011e-01 -4.31357771e-01
-1.85330319e+00 1.26221919e+00 -2.03196079e-01 -1.21554442e-01
-2.54553437e-01 8.37776244e-01 9.71801430e-02 -1.15336619e-01
1.59731597e-01 2.32648790e-01 -2.09974080e-01 2.15017736e-01
4.79414344e-01 2.90454596e-01 4.05462608e-02 -9.70228970e-01
-5.60811758e-01 7.05411434e-01 -3.00140172e-01 2.28273317e-01
8.76225591e-01 -3.28143835e-01 2.72455126e-01 3.83120239e-01
1.46242547e+00 9.30380896e-02 -7.42031395e-01 -3.74266773e-01
-4.32411432e-02 -1.08494222e+00 -5.36866449e-02 -1.04114795e+00
-9.34027493e-01 4.84556645e-01 1.30851734e+00 2.55672097e-01
1.04564726e+00 6.42779171e-01 3.96491498e-01 7.44912177e-02
2.26642951e-01 -1.07637298e+00 5.07953227e-01 -1.23169266e-01
1.23785889e+00 -1.58293653e+00 -2.31402088e-02 -4.14293826e-01
-9.15672362e-01 6.48562849e-01 5.26855230e-01 -2.03521892e-01
4.89440292e-01 -2.77987808e-01 1.25479370e-01 -4.26032901e-01
-5.22000670e-01 -4.56208676e-01 5.84913492e-01 7.71287799e-01
2.12135300e-01 -2.27548271e-01 -3.09789032e-01 2.96892017e-01
2.16301486e-01 -1.73983678e-01 -8.90141279e-02 8.59738290e-01
-2.00694174e-01 -1.09502876e+00 -3.76798987e-01 3.00803751e-01
-3.06992918e-01 -2.12656200e-01 -5.30470788e-01 8.51118267e-01
-1.54703967e-02 1.15819943e+00 2.28181239e-02 -4.58227843e-01
5.61856508e-01 -1.45745248e-01 6.07243121e-01 -3.66971642e-01
-7.20842540e-01 -1.80740416e-01 2.08685875e-01 -6.02683187e-01
-4.48815525e-01 -4.66996253e-01 -1.31718308e-01 -3.48111033e-01
-2.77256876e-01 1.97489679e-01 3.85400444e-01 1.49138629e-01
6.41985595e-01 1.35463536e-01 7.26020038e-01 -7.40861237e-01
-5.38073123e-01 -7.78339684e-01 -6.18602276e-01 1.01778102e+00
1.24273911e-01 -6.77307546e-01 -4.11944777e-01 9.89247039e-02] | [10.8936767578125, 1.1003391742706299] |
0868e6e4-784d-428a-bafd-a8276e3797e5 | degreembed-incorporating-entity-embedding | 2112.09933 | null | https://arxiv.org/abs/2112.09933v2 | https://arxiv.org/pdf/2112.09933v2.pdf | DegreEmbed: incorporating entity embedding into logic rule learning for knowledge graph reasoning | Knowledge graphs (KGs), as structured representations of real world facts, are intelligent databases incorporating human knowledge that can help machine imitate the way of human problem solving. However, KGs are usually huge and there are inevitably missing facts in KGs, thus undermining applications such as question answering and recommender systems that are based on knowledge graph reasoning. Link prediction for knowledge graphs is the task aiming to complete missing facts by reasoning based on the existing knowledge. Two main streams of research are widely studied: one learns low-dimensional embeddings for entities and relations that can explore latent patterns, and the other gains good interpretability by mining logical rules. Unfortunately, the heterogeneity of modern KGs that involve entities and relations of various types is not well considered in the previous studies. In this paper, we propose DegreEmbed, a model that combines embedding-based learning and logic rule mining for inferring on KGs. Specifically, we study the problem of predicting missing links in heterogeneous KGs from the perspective of the degree of nodes. Experimentally, we demonstrate that our DegreEmbed model outperforms the state-of-the-art methods on real world datasets and the rules mined by our model are of high quality and interpretability. | ['Yuliang Wei', 'Guodong Xin', 'Yao Wang', 'Hongri Liu', 'Haotian Li'] | 2021-12-18 | null | null | null | null | ['link-prediction', 'knowledge-graphs', 'question-answering'] | ['graphs', 'knowledge-base', 'natural-language-processing'] | [-5.03334701e-01 8.25224280e-01 -8.72859240e-01 -1.60287589e-01
3.53042483e-01 -2.86256909e-01 2.78372794e-01 6.64173841e-01
3.09349895e-01 8.19259346e-01 3.57484609e-01 -5.85696995e-01
-7.27583408e-01 -1.66992116e+00 -7.52251625e-01 -2.21037775e-01
-4.62603450e-01 8.01785171e-01 5.21142900e-01 -4.09318447e-01
-1.06745571e-01 1.38110399e-01 -1.50813615e+00 4.84056920e-01
1.16790223e+00 1.06086195e+00 -2.95054495e-01 1.27828971e-01
-6.59978449e-01 1.34856892e+00 -9.47324485e-02 -1.10023010e+00
-3.14881876e-02 1.05343044e-01 -1.00282657e+00 -3.98754179e-01
1.82210624e-01 5.74678518e-02 -7.56316960e-01 1.16991985e+00
-6.83121756e-03 -2.62756646e-01 6.06995225e-01 -1.58520997e+00
-1.16673493e+00 1.00629401e+00 -2.74261415e-01 1.82310432e-01
4.72983062e-01 -4.63972867e-01 1.54552090e+00 -7.16058671e-01
9.08585370e-01 1.30780649e+00 5.87605000e-01 2.21820667e-01
-8.68168116e-01 -4.74326134e-01 3.73153567e-01 9.92944062e-01
-1.39508498e+00 -9.28593352e-02 8.68036568e-01 -5.02776384e-01
9.89737451e-01 1.62052751e-01 8.82550836e-01 7.29002297e-01
6.55684844e-02 6.81744158e-01 6.18384600e-01 -4.47413146e-01
2.49686867e-01 5.94939590e-01 4.37919319e-01 1.20511079e+00
1.13458204e+00 -2.49880061e-01 -8.06610882e-01 -2.97283471e-01
4.58742112e-01 1.51200324e-01 -3.01741451e-01 -6.75408304e-01
-1.04120970e+00 1.00648808e+00 5.82376540e-01 2.14555547e-01
-3.59993905e-01 -3.51553053e-01 2.04688191e-01 5.44022381e-01
5.08905113e-01 4.64915305e-01 -8.68666232e-01 3.90073240e-01
-3.74317855e-01 6.43109083e-02 1.25113451e+00 1.11594927e+00
7.35349417e-01 -1.89251304e-01 2.51352042e-01 4.83122230e-01
5.76231062e-01 1.93730757e-01 2.99099416e-01 -4.21250641e-01
6.59876049e-01 1.70425260e+00 -2.89464086e-01 -1.72490001e+00
-3.73532265e-01 -4.19720560e-01 -8.51313293e-01 -3.40949565e-01
2.10733071e-01 8.87484998e-02 -4.26987559e-01 1.36901355e+00
6.85054421e-01 2.78022438e-01 3.50885302e-01 5.35046756e-01
1.12141788e+00 3.87528211e-01 -1.31390959e-01 -1.27313644e-01
1.28013957e+00 -8.78193796e-01 -8.97808850e-01 1.06205642e-01
8.97574484e-01 -6.20353222e-02 7.37502277e-01 2.60548949e-01
-5.94903231e-01 -1.52617157e-01 -1.14085114e+00 -8.70609358e-02
-1.00386560e+00 -1.82739452e-01 1.20820880e+00 5.43635488e-01
-7.71323502e-01 5.51424325e-01 -4.61117208e-01 -2.38525882e-01
6.13173366e-01 1.94428578e-01 -4.84560788e-01 -4.49136287e-01
-1.81503880e+00 8.52047503e-01 8.63882959e-01 -2.39843205e-02
-1.20128199e-01 -8.32157493e-01 -9.78622973e-01 2.76147485e-01
1.17305839e+00 -1.00129616e+00 2.23245695e-01 -1.37273401e-01
-8.90805244e-01 5.68363070e-01 2.68631428e-03 -5.02710998e-01
4.51389812e-02 -2.47319341e-02 -1.19091105e+00 3.48028503e-02
3.77351269e-02 -1.03784977e-02 4.64519978e-01 -1.17331862e+00
-5.17358959e-01 -6.13601744e-01 4.02959526e-01 -6.19041361e-02
-7.10873485e-01 -7.72542298e-01 -3.20487708e-01 -1.60283566e-01
1.77346542e-01 -4.09763873e-01 1.06105514e-01 -5.18151112e-02
-8.04962337e-01 -6.07878625e-01 7.80513108e-01 -4.94282573e-01
1.66514635e+00 -1.71193063e+00 1.01132073e-01 6.32984877e-01
8.42276990e-01 4.00094867e-01 2.02947810e-01 5.97136557e-01
4.14009131e-02 3.83397579e-01 3.48677576e-01 5.27646959e-01
1.57994226e-01 5.67612231e-01 -4.36497420e-01 -9.48403403e-02
-4.25679758e-02 1.17314398e+00 -1.08966553e+00 -7.41460502e-01
-8.40859264e-02 -1.68654658e-02 -5.62867820e-01 -2.00726148e-02
-6.27882123e-01 -1.95629388e-01 -6.43135369e-01 9.80329275e-01
3.94226521e-01 -7.61472344e-01 8.22000504e-01 -5.65089285e-01
3.64837945e-01 2.54473597e-01 -1.37667465e+00 1.30002880e+00
-1.48637012e-01 2.13717952e-01 -5.22176802e-01 -1.22188425e+00
8.44091415e-01 1.36857077e-01 3.30248982e-01 -4.15719092e-01
-1.81753010e-01 1.45999536e-01 1.56221539e-01 -9.06497836e-01
2.84183145e-01 1.01831846e-01 2.37078667e-01 1.63276792e-01
9.78912711e-02 3.82946610e-01 4.05062586e-01 7.91890860e-01
1.27476418e+00 -2.43947893e-01 5.04342914e-01 -1.09417595e-01
4.33707982e-01 2.85993248e-01 6.55908883e-01 3.81617635e-01
3.82492393e-01 -4.73047823e-01 9.55317020e-01 -7.39492118e-01
-6.27629220e-01 -1.05698740e+00 -2.48388876e-03 5.23435295e-01
2.57081956e-01 -9.21075046e-01 -4.81939316e-02 -1.13000071e+00
7.48335600e-01 5.65559685e-01 -6.69837058e-01 -3.12069982e-01
5.47455214e-02 -5.22417843e-01 3.06908786e-01 4.99053895e-01
3.54623467e-01 -8.02275300e-01 1.71417817e-01 1.28383979e-01
-1.48840278e-01 -1.31690562e+00 2.35876963e-01 -1.83454484e-01
-9.39697683e-01 -1.91376257e+00 3.38996887e-01 -6.53516054e-01
8.16745758e-01 4.51169722e-03 1.45195615e+00 4.45241213e-01
-1.80833951e-01 3.42586458e-01 -6.22174084e-01 -4.39003587e-01
-7.72493705e-02 2.03587770e-01 3.35516483e-01 -8.22084919e-02
8.26790035e-01 -9.32245553e-01 -3.84062529e-01 1.71995208e-01
-8.88369441e-01 9.70276147e-02 8.10496688e-01 7.50102460e-01
4.76932704e-01 8.08150291e-01 8.96564960e-01 -1.62695837e+00
6.15921557e-01 -1.07715714e+00 -3.68469864e-01 8.35713923e-01
-1.29249036e+00 4.96989578e-01 5.65600693e-01 -1.68578774e-01
-8.73043060e-01 -5.93088150e-01 3.04796636e-01 -3.64388347e-01
2.34388679e-01 1.22478437e+00 -4.75995004e-01 3.61403190e-02
4.51544583e-01 -1.04957208e-01 -3.29344481e-01 -3.43327701e-01
7.17670739e-01 3.53732646e-01 2.76401758e-01 -6.39681160e-01
9.55123842e-01 4.53834623e-01 3.26033443e-01 -5.68013906e-01
-1.10384154e+00 -3.98908108e-01 -6.07254982e-01 -3.24179642e-02
4.03192908e-01 -7.84950912e-01 -9.16915238e-01 -2.70116955e-01
-9.73064482e-01 2.92056978e-01 -3.71459037e-01 4.04136747e-01
-5.28745465e-02 3.05749863e-01 -3.94414365e-01 -5.92054248e-01
-7.64149874e-02 -3.02652657e-01 3.96427780e-01 5.38593158e-02
-9.53058079e-02 -1.36939967e+00 2.06168443e-01 5.54507017e-01
-1.23262051e-02 2.60083407e-01 1.74784207e+00 -9.64078963e-01
-9.43600714e-01 -1.29103482e-01 -2.71877944e-01 -1.53536186e-01
3.17395985e-01 -3.24554145e-01 -5.52582145e-01 1.71406612e-01
-5.50956488e-01 -2.15430498e-01 8.24942827e-01 -2.37494886e-01
1.08012700e+00 -7.75087297e-01 -6.13255560e-01 3.49730819e-01
1.47353065e+00 -1.17088005e-01 6.72766328e-01 2.87764639e-01
9.39852893e-01 7.54033685e-01 4.22560155e-01 2.90279090e-01
9.95010257e-01 3.63976389e-01 5.62366426e-01 2.87952006e-01
8.06327239e-02 -1.04289484e+00 -6.67063296e-02 1.13533545e+00
-2.74354219e-01 -1.66660681e-01 -8.60444605e-01 7.65779674e-01
-2.12363720e+00 -1.10518336e+00 -3.66701454e-01 1.77569306e+00
1.04443431e+00 2.04371125e-01 -1.75013602e-01 4.57244515e-01
6.30972445e-01 -2.47894917e-02 -6.75535321e-01 -8.86106566e-02
-2.01088652e-01 4.76842336e-02 3.27423781e-01 3.52814168e-01
-6.65177166e-01 8.50919306e-01 5.00083923e+00 6.64352715e-01
-4.92221981e-01 -3.93184908e-02 5.19798622e-02 2.08962649e-01
-8.31786096e-01 2.78515100e-01 -8.90989065e-01 3.79193157e-01
6.94060028e-01 -4.64177489e-01 2.61214018e-01 9.57599998e-01
-3.70611519e-01 4.18929085e-02 -1.17430186e+00 7.91787505e-01
5.44281257e-03 -1.75292945e+00 4.60545093e-01 2.35133111e-01
9.92392659e-01 -4.71105725e-01 -2.09930480e-01 5.64136088e-01
7.07882404e-01 -9.27939594e-01 1.45631626e-01 9.17478740e-01
4.83977109e-01 -6.09032869e-01 9.08047795e-01 3.99561733e-01
-1.20670056e+00 -1.61809444e-01 -5.50160289e-01 -1.24958888e-01
-7.68362209e-02 1.23947608e+00 -8.65140200e-01 1.33307540e+00
6.03334427e-01 9.71331775e-01 -6.20941103e-01 7.59016752e-01
-6.16777539e-01 5.96354842e-01 -1.45537719e-01 -1.09372579e-01
-2.47913525e-01 -1.73160478e-01 1.81691185e-01 6.99011862e-01
2.09741846e-01 2.48485401e-01 1.32242963e-01 8.59043181e-01
-4.44921434e-01 1.99301273e-01 -9.72050726e-01 -2.61077344e-01
6.55669987e-01 1.10807621e+00 -3.86223286e-01 -4.68054771e-01
-6.31949127e-01 3.15942109e-01 7.25843489e-01 2.54845381e-01
-3.65764469e-01 -3.43883336e-01 5.57857692e-01 4.38748151e-01
3.57921511e-01 -3.96550819e-02 -1.19432531e-01 -1.57332683e+00
2.73000270e-01 -5.78621447e-01 8.53512645e-01 -5.29151976e-01
-1.75174928e+00 2.64988065e-01 4.18008165e-03 -9.03870344e-01
-1.09223008e-01 -7.52998114e-01 -4.05946612e-01 4.64270025e-01
-1.89761341e+00 -1.13042271e+00 -1.74028635e-01 8.04144979e-01
-1.52647927e-01 -3.12947690e-01 8.41698349e-01 4.02330786e-01
-4.69022244e-01 4.37420219e-01 -4.34101932e-02 2.84281433e-01
3.38834435e-01 -1.26141655e+00 -1.26763090e-01 3.92754138e-01
6.36601508e-01 8.19965303e-01 3.53110999e-01 -1.04944718e+00
-1.69796801e+00 -1.07115531e+00 1.19314849e+00 -5.11753798e-01
1.06061208e+00 -2.85283849e-02 -1.07625818e+00 8.41544509e-01
-1.34328946e-01 3.56593877e-01 1.03304327e+00 9.24773157e-01
-6.38877571e-01 -2.69329309e-01 -9.97837722e-01 5.07153273e-01
1.55851042e+00 -4.13741618e-01 -9.28094447e-01 3.42378020e-01
7.46783793e-01 -3.58720589e-03 -1.16967440e+00 5.00093937e-01
4.47452784e-01 -7.93563187e-01 8.92310262e-01 -1.19204307e+00
4.83718693e-01 -4.49160725e-01 -1.49910033e-01 -1.39653575e+00
-2.92722821e-01 -1.02961712e-01 -1.18156838e+00 1.12584805e+00
7.17082381e-01 -7.07388043e-01 1.06221533e+00 4.18862283e-01
2.55819440e-01 -1.08545852e+00 -6.60200477e-01 -8.45219433e-01
-4.33619887e-01 -2.48736367e-01 8.50426733e-01 1.49474406e+00
6.78807974e-01 6.25531197e-01 -2.01475337e-01 5.49252152e-01
6.96800113e-01 5.94375968e-01 6.06451094e-01 -1.96676922e+00
-1.73007980e-01 -1.28425255e-01 -9.35550213e-01 -7.49050498e-01
3.14714193e-01 -1.31877220e+00 -1.02140296e+00 -1.94056559e+00
1.56494513e-01 -5.88452101e-01 -2.55295366e-01 7.01328039e-01
-1.33062750e-01 -4.57179576e-01 -2.64515787e-01 9.64261442e-02
-7.78724909e-01 6.02504492e-01 1.10328186e+00 -5.29426754e-01
-2.11245995e-02 -2.79222369e-01 -8.73463154e-01 9.31247234e-01
4.94537175e-01 -3.83640915e-01 -9.05667305e-01 -2.90382475e-01
1.16683650e+00 7.16841407e-03 2.90350407e-01 -5.70378780e-01
7.72177756e-01 -2.54123986e-01 1.05551168e-01 -4.35488492e-01
7.28503242e-02 -1.02054203e+00 2.51396716e-01 1.52318761e-01
-6.83312640e-02 -1.10359758e-01 -2.06554785e-01 1.11597586e+00
-4.58289325e-01 -4.67273481e-02 -1.26625597e-01 -1.16356902e-01
-9.91104424e-01 4.99942869e-01 3.30601126e-01 3.17491055e-01
9.05270994e-01 -6.95299997e-04 -7.16200829e-01 -2.69662946e-01
-8.64936471e-01 5.34146428e-01 -2.47134771e-02 4.14592117e-01
8.95256937e-01 -1.67857158e+00 -4.24941897e-01 9.18601006e-02
4.34691846e-01 1.56785041e-01 1.40488535e-01 7.23273396e-01
-2.47114107e-01 5.55611014e-01 1.22754179e-01 9.14059505e-02
-9.85553801e-01 8.90795290e-01 -7.97815025e-02 -7.51145720e-01
-5.87383211e-01 4.69130903e-01 -2.01179177e-01 -4.49378073e-01
1.19523950e-01 -2.97010154e-01 -6.13265455e-01 2.04606742e-01
2.40661189e-01 4.81770545e-01 6.61879778e-02 -3.96857634e-02
-3.31659347e-01 2.47674018e-01 -2.25902721e-01 6.13922060e-01
1.51105905e+00 3.70246246e-02 -4.23692763e-01 5.20598352e-01
7.01574504e-01 2.81864047e-01 -3.38463962e-01 -7.09757090e-01
4.37255651e-01 -6.04449391e-01 -1.41753674e-01 -8.16958189e-01
-1.17135060e+00 5.66864192e-01 -2.27479413e-01 6.71079814e-01
6.55073524e-01 3.62107426e-01 6.59496486e-01 8.60380232e-01
6.62881672e-01 -1.01995134e+00 -2.15726029e-02 2.86484838e-01
5.42966843e-01 -1.09345520e+00 1.99309260e-01 -1.00469351e+00
-5.31397104e-01 1.04278243e+00 6.33008003e-01 2.17471033e-01
1.13178635e+00 1.24573708e-02 -4.54222322e-01 -7.94644773e-01
-1.02547359e+00 -3.20474148e-01 3.87928873e-01 6.31479144e-01
-7.64936805e-02 1.73870102e-01 -2.35586524e-01 1.03944540e+00
-3.04929644e-01 1.66495621e-01 1.92243323e-01 6.99727595e-01
-4.71332133e-01 -1.18755984e+00 5.06937131e-02 9.12014961e-01
-4.45540063e-02 -2.39819497e-01 -4.90841061e-01 7.00593591e-01
5.07905781e-01 9.17404830e-01 -4.14537936e-01 -7.03935742e-01
4.55586523e-01 2.98780501e-01 3.67516518e-01 -6.47698045e-01
2.66410530e-01 -1.04833436e+00 4.55302268e-01 -3.04908365e-01
-3.26804459e-01 -1.91034541e-01 -1.11473930e+00 -6.78814054e-01
-5.71040571e-01 5.36408365e-01 2.11300090e-01 1.02557588e+00
5.85129380e-01 5.88129580e-01 4.17937547e-01 4.48746413e-01
-1.79738566e-01 -6.22543871e-01 -1.05833578e+00 4.04953599e-01
-1.81346267e-01 -1.04954767e+00 -2.05002576e-01 -9.99431014e-02] | [8.831267356872559, 7.8418145179748535] |
76de9d4c-d5bc-427b-850a-185cc860c251 | immfusion-robust-mmwave-rgb-fusion-for-3d | 2210.01346 | null | https://arxiv.org/abs/2210.01346v1 | https://arxiv.org/pdf/2210.01346v1.pdf | ImmFusion: Robust mmWave-RGB Fusion for 3D Human Body Reconstruction in All Weather Conditions | 3D human reconstruction from RGB images achieves decent results in good weather conditions but degrades dramatically in rough weather. Complementary, mmWave radars have been employed to reconstruct 3D human joints and meshes in rough weather. However, combining RGB and mmWave signals for robust all-weather 3D human reconstruction is still an open challenge, given the sparse nature of mmWave and the vulnerability of RGB images. In this paper, we present ImmFusion, the first mmWave-RGB fusion solution to reconstruct 3D human bodies in all weather conditions robustly. Specifically, our ImmFusion consists of image and point backbones for token feature extraction and a Transformer module for token fusion. The image and point backbones refine global and local features from original data, and the Fusion Transformer Module aims for effective information fusion of two modalities by dynamically selecting informative tokens. Extensive experiments on a large-scale dataset, mmBody, captured in various environments demonstrate that ImmFusion can efficiently utilize the information of two modalities to achieve a robust 3D human body reconstruction in all weather conditions. In addition, our method's accuracy is significantly superior to that of state-of-the-art Transformer-based LiDAR-camera fusion methods. | ['Qi Ye', 'Yuchi Huo', 'Jiming Chen', 'Bin Fang', 'Yingfeng Chen', 'Shaohao Zhu', 'Kun Shi', 'Xiangyu Wang', 'Anjun Chen'] | 2022-10-04 | null | null | null | null | ['3d-human-reconstruction'] | ['computer-vision'] | [-7.76998326e-02 -3.78764749e-01 3.06604534e-01 -2.78937340e-01
-7.90206552e-01 -2.77005225e-01 2.99641818e-01 -3.35019886e-01
-3.22957158e-01 4.57683712e-01 1.57190308e-01 4.32922989e-01
2.15894178e-01 -1.18381846e+00 -6.04157209e-01 -8.11896861e-01
1.07274905e-01 7.16210961e-01 2.35150293e-01 -4.47627664e-01
-5.45494556e-01 5.72535157e-01 -1.99540567e+00 2.36425921e-02
6.79420412e-01 1.26551771e+00 -5.27642071e-02 5.98314285e-01
3.94673366e-03 2.62068033e-01 -2.88655490e-01 -6.19526267e-01
7.18246698e-01 -1.37768850e-01 -8.23748857e-02 7.57286549e-02
7.91645408e-01 -3.44563693e-01 -7.19089746e-01 7.96342313e-01
9.64550614e-01 1.24494776e-01 4.79625165e-01 -1.17845690e+00
-1.08084582e-01 -7.13307485e-02 -6.11884296e-01 -2.67034769e-01
8.34585249e-01 3.17445934e-01 5.85622489e-01 -1.12998188e+00
4.95475531e-01 1.32496738e+00 1.02416790e+00 1.58890620e-01
-7.67640471e-01 -8.28347147e-01 -3.73040587e-01 2.58780539e-01
-1.57511687e+00 -4.16666120e-01 7.61613309e-01 -1.83151484e-01
6.76654875e-01 3.80784035e-01 1.47781456e+00 8.21033061e-01
3.63951623e-01 5.24397373e-01 1.19828308e+00 -9.22698006e-02
6.65527508e-02 -4.85144705e-01 -4.03614342e-01 9.68882561e-01
4.27831143e-01 3.84294719e-01 -1.34200931e+00 -2.39258155e-01
7.12844849e-01 4.18343872e-01 -2.34599158e-01 -1.74065188e-01
-1.39542580e+00 4.85304385e-01 7.04572618e-01 -2.66947836e-01
-7.83346117e-01 3.73702019e-01 -5.95563874e-02 1.62559539e-01
5.16368270e-01 -2.32250839e-01 -1.23506591e-01 5.71219809e-03
-1.10591698e+00 6.43410921e-01 5.56151152e-01 8.69269609e-01
1.02580273e+00 1.80355653e-01 2.81437010e-01 5.18378139e-01
7.03068316e-01 1.74881077e+00 -2.06467003e-01 -1.12910390e+00
5.83513737e-01 4.66271967e-01 -1.01469152e-01 -1.05296421e+00
-5.43563485e-01 -2.39165127e-01 -1.14299655e+00 2.72921026e-01
6.34743422e-02 -1.69857576e-01 -1.09278703e+00 1.15870047e+00
9.59807098e-01 2.44114682e-01 5.81861362e-02 1.47891331e+00
1.17387962e+00 5.30326426e-01 -2.74890810e-01 5.75746186e-02
1.54826236e+00 -1.97973356e-01 -6.42906785e-01 -3.99669886e-01
-2.17150375e-01 -8.11897874e-01 5.43219328e-01 3.58839065e-01
-1.06296062e+00 -5.53909957e-01 -9.49488640e-01 -1.76467448e-01
-4.53623803e-03 -2.72234082e-01 4.71641362e-01 5.83802938e-01
-6.94665432e-01 1.13647588e-01 -9.40089047e-01 -3.84087265e-01
8.71224999e-02 1.96314692e-01 -6.72291279e-01 -5.77725172e-01
-1.31012714e+00 8.16949904e-01 -1.67457551e-01 4.40684170e-01
-5.97054482e-01 -7.11452544e-01 -1.16531897e+00 -5.48041880e-01
4.39853445e-02 -1.31634223e+00 8.51876020e-01 -8.57794583e-02
-1.21988249e+00 8.89623702e-01 -1.44188762e-01 -2.59425640e-01
2.71866024e-01 -5.05839884e-01 -2.65895694e-01 5.16650379e-01
1.42738074e-01 5.50491810e-01 9.70162868e-01 -1.35281801e+00
-6.31589890e-01 -9.14163709e-01 -3.42760324e-01 5.16991675e-01
1.75486192e-01 -4.34523910e-01 -5.66104233e-01 -6.25460804e-01
8.76111269e-01 -1.01699078e+00 -2.24433392e-01 3.73693973e-01
-1.02065362e-01 3.77969295e-01 5.76996326e-01 -7.40408540e-01
5.58609188e-01 -1.93242025e+00 2.84581453e-01 5.02369583e-01
1.58998460e-01 -2.98622966e-01 2.50626743e-01 4.51307923e-01
5.29783607e-01 -5.12997270e-01 -3.95549566e-01 -6.78691745e-01
4.20784578e-02 6.32802606e-01 -1.42113179e-01 1.03002441e+00
-3.84826273e-01 8.13324392e-01 -5.85745633e-01 -8.69736195e-01
6.55256629e-01 7.82684088e-01 -2.25211203e-01 2.98702657e-01
1.70792043e-01 5.80555558e-01 -3.62464547e-01 1.56660628e+00
9.04150188e-01 3.94974887e-01 -2.84519166e-01 -6.97268426e-01
-1.73372313e-01 -2.15906322e-01 -1.23801696e+00 2.22257614e+00
-1.57475486e-01 1.57836914e-01 5.63959301e-01 -4.27693069e-01
1.00268221e+00 3.84796321e-01 8.40474010e-01 -9.81432855e-01
2.40302328e-02 2.22861961e-01 -7.23362744e-01 -5.51374555e-01
7.27729857e-01 -6.58399761e-01 -5.44756114e-01 1.93217099e-01
-5.85242175e-02 -7.97974467e-01 -3.48750770e-01 7.73180723e-02
1.01776290e+00 2.56120175e-01 -5.40821254e-02 2.70487815e-01
2.49359850e-02 1.61568731e-01 8.32038820e-01 4.89918590e-01
-9.41416994e-02 9.90084171e-01 -6.77585423e-01 -5.55138826e-01
-7.42017984e-01 -1.57795751e+00 5.23964316e-02 5.77734172e-01
4.70172167e-01 -4.17355508e-01 -2.32384011e-01 -1.48431137e-01
4.79703784e-01 5.97089529e-02 -3.63002628e-01 4.58772369e-02
-6.44827306e-01 -7.18918979e-01 7.65532494e-01 3.78529251e-01
9.65622127e-01 -4.89486754e-01 -1.13111842e+00 -4.42579053e-02
-8.00070882e-01 -1.21146095e+00 4.13958952e-02 -1.57523125e-01
-9.33180749e-01 -1.11178970e+00 -5.98152280e-01 -9.61829275e-02
4.41023171e-01 3.89979035e-01 1.20894384e+00 1.13878883e-01
-5.73753715e-01 7.52855301e-01 -4.73910987e-01 -4.33819592e-01
4.18113917e-01 -4.68804091e-01 3.60362798e-01 -3.60589921e-02
-2.79845912e-02 -7.59055734e-01 -7.66683042e-01 2.51785308e-01
-6.63270712e-01 -4.70184609e-02 6.04179323e-01 4.03180212e-01
8.88337493e-01 1.59085356e-02 -3.95851821e-01 -1.75634190e-01
-2.57435948e-01 -3.95899951e-01 -3.34009022e-01 -3.59042734e-02
-1.01571046e-01 -1.80359736e-01 -1.09437078e-01 1.51211634e-01
-1.08863914e+00 6.14694953e-01 -2.97096312e-01 -6.97089672e-01
-1.47118524e-01 4.76302117e-01 -1.81078032e-01 -2.22891659e-01
5.06166041e-01 2.51684129e-01 4.77465093e-02 -5.77377796e-01
4.82185304e-01 4.86668229e-01 1.13302088e+00 -6.95102930e-01
1.32910514e+00 1.12759614e+00 3.68820548e-01 -1.06870663e+00
-6.78627610e-01 -6.80609822e-01 -7.51998246e-01 -7.82662570e-01
9.05800343e-01 -1.55951130e+00 -6.96285784e-01 6.67913795e-01
-8.00113976e-01 2.89416537e-02 -5.11571109e-01 6.76524103e-01
-5.20958304e-01 3.28956038e-01 -6.42723799e-01 -9.75218475e-01
-5.55869222e-01 -7.06039071e-01 1.62436593e+00 1.85145944e-01
4.10505161e-02 -3.83479625e-01 3.76102537e-01 7.97072530e-01
-3.94040383e-02 8.86768460e-01 1.40766148e-02 5.83139837e-01
-7.55627513e-01 -2.34758303e-01 1.60026670e-01 -2.37112626e-01
7.92750530e-03 -2.51421571e-01 -9.59987342e-01 -1.57729924e-01
-1.07736506e-01 -4.20050830e-01 9.42619503e-01 3.12123448e-01
2.41347224e-01 1.02970108e-01 -2.19994158e-01 1.13902509e+00
1.22742593e+00 -5.18305421e-01 6.99948847e-01 2.46118367e-01
9.07928824e-01 4.19491500e-01 1.00505865e+00 8.22662592e-01
9.52339590e-01 6.99514866e-01 7.44238734e-01 -3.05832356e-01
-3.84211332e-01 -2.48471037e-01 4.92761314e-01 9.44860280e-01
-6.14369154e-01 1.29916504e-01 -8.79486740e-01 3.52009863e-01
-1.87534428e+00 -9.58806276e-01 -2.47282878e-01 2.14041924e+00
5.82092702e-01 -2.73751885e-01 6.38662353e-02 2.92089224e-01
2.54053086e-01 2.01920092e-01 -3.72583449e-01 3.26842040e-01
-3.90299410e-01 4.29629952e-01 7.16331065e-01 1.70157880e-01
-9.56589937e-01 6.64497793e-01 6.14626503e+00 3.78088832e-01
-7.49261796e-01 2.28346020e-01 -3.49234045e-01 -3.15082163e-01
-3.36458623e-01 -6.94379285e-02 -6.48593009e-01 2.52643496e-01
5.96987247e-01 3.16269845e-01 4.37832475e-02 4.56739157e-01
-6.22156402e-03 -4.19954449e-01 -3.96096200e-01 1.40340757e+00
1.80035993e-01 -1.04108596e+00 -1.49060532e-01 2.15978324e-01
4.57590371e-01 2.71097988e-01 -2.27555737e-01 -1.16030447e-01
3.42981577e-01 -6.77219808e-01 1.24010623e+00 1.11482894e+00
8.35973620e-01 -9.44219172e-01 7.08823681e-01 3.91228199e-01
-1.74568391e+00 2.50113666e-01 -4.03186202e-01 -1.99880630e-01
6.56813741e-01 1.29758680e+00 -3.80829811e-01 1.10178196e+00
1.38911426e+00 7.35729754e-01 -2.95670837e-01 9.32300746e-01
-4.20793980e-01 3.77808750e-01 -8.87886882e-01 5.37009239e-01
-2.79813379e-01 -2.76545376e-01 6.35260940e-01 9.80887175e-01
7.56844223e-01 7.21861005e-01 5.43092549e-01 3.99332821e-01
1.30935714e-01 -3.44811678e-01 -7.57399619e-01 4.76980239e-01
4.09978300e-01 1.31767154e+00 -3.64629537e-01 -2.76347518e-01
-1.62875414e-01 1.01329625e+00 -2.23198727e-01 1.49005577e-01
-8.05731773e-01 -8.16039462e-03 1.01724577e+00 3.09016407e-01
2.11833641e-01 -8.22431445e-01 -4.56986755e-01 -1.35311413e+00
3.64355801e-04 -4.63080496e-01 4.61653441e-01 -1.02748907e+00
-1.12109315e+00 2.16605410e-01 1.15584664e-01 -1.41970778e+00
-1.46201178e-01 -1.38569579e-01 -1.25311524e-01 6.59249187e-01
-1.50132775e+00 -1.67809069e+00 -8.44337702e-01 1.15166271e+00
1.92552552e-01 2.05852717e-01 6.24998927e-01 3.76609772e-01
-9.59603563e-02 1.25580356e-01 -4.28452492e-01 1.87537726e-02
7.03270555e-01 -9.87991810e-01 3.12647849e-01 9.96023178e-01
2.60039121e-01 2.07052201e-01 8.43892038e-01 -1.22916961e+00
-2.31409645e+00 -9.29172695e-01 6.22575581e-01 -2.23206863e-01
1.73994657e-02 -3.03026795e-01 -4.00606662e-01 4.55957830e-01
-1.05612695e-01 3.10683757e-01 6.78109229e-01 -1.31251693e-01
-3.82462710e-01 -3.58716846e-01 -1.37653553e+00 2.36972257e-01
1.26004934e+00 -4.74733084e-01 -7.97436535e-01 2.59339303e-01
5.08948445e-01 -9.20519888e-01 -1.25063813e+00 8.82849932e-01
8.77997041e-01 -1.16435707e+00 1.53566098e+00 4.04266357e-01
4.16788086e-02 -7.04734385e-01 -7.48432755e-01 -1.08588684e+00
-1.22149527e-01 -2.47285292e-01 -4.38615054e-01 9.17356551e-01
-3.43385041e-01 -3.83374751e-01 7.75479794e-01 4.05760288e-01
-2.79409617e-01 -2.15897337e-01 -1.39865232e+00 -3.94073606e-01
-5.85790217e-01 -8.02345216e-01 6.69622242e-01 5.22111237e-01
-4.78347659e-01 -2.66369343e-01 -6.70484781e-01 5.73051453e-01
1.56511331e+00 5.13253629e-01 1.09207308e+00 -1.35383666e+00
-1.21255741e-02 4.33737695e-01 -6.56509340e-01 -7.89078832e-01
-3.92690338e-02 -5.72576106e-01 4.67172474e-01 -1.79937291e+00
-1.94216952e-01 -5.01056552e-01 2.38982245e-01 5.49019575e-01
9.72975418e-02 9.11000252e-01 3.38423550e-01 4.67318267e-01
-3.83741260e-01 8.94286633e-01 1.04111111e+00 -2.64815450e-01
1.73761159e-01 -9.53171253e-02 -1.10379085e-01 8.18073988e-01
3.34369540e-01 -3.43513817e-01 1.01418346e-01 -5.29568732e-01
3.52980971e-01 2.11972177e-01 7.43979812e-01 -1.51928520e+00
5.14882982e-01 -2.00181529e-01 8.87353718e-01 -1.25076091e+00
1.08645117e+00 -1.01275909e+00 8.12224090e-01 5.58216810e-01
6.03649020e-01 1.85494930e-01 -1.66122317e-02 6.51503503e-01
-1.88900635e-01 5.57758093e-01 5.85905492e-01 -3.94709766e-01
-7.70861208e-01 6.03457034e-01 -1.79071411e-01 -2.17471808e-01
7.24345505e-01 -4.79716778e-01 -5.05114309e-02 -3.32454890e-01
-6.95831358e-01 2.80157357e-01 7.85859227e-01 1.75252810e-01
1.30476391e+00 -1.58048213e+00 -9.06450033e-01 4.16221648e-01
1.00698456e-01 2.67969429e-01 5.40307939e-01 8.72448623e-01
-6.42855823e-01 -1.25996575e-01 -2.50686973e-01 -9.17993188e-01
-1.44030511e+00 -1.03611052e-01 2.35054374e-01 1.83685780e-01
-1.05921280e+00 7.42365539e-01 -2.83836067e-01 -6.75200760e-01
-4.59131040e-02 -2.74319291e-01 3.74999970e-01 1.59718618e-01
3.72657835e-01 5.62622964e-01 1.58991441e-01 -1.13921165e+00
-5.66457689e-01 1.20595348e+00 8.63653779e-01 -6.10050380e-01
1.36357200e+00 -4.13466126e-01 -1.66422725e-01 4.78357792e-01
6.03151560e-01 1.04027249e-01 -1.17564738e+00 -3.02353442e-01
-8.28691483e-01 -7.23804414e-01 3.66348252e-02 -4.49494183e-01
-1.47983873e+00 8.33699763e-01 6.47300541e-01 -3.24782014e-01
1.42128336e+00 -1.45774195e-03 1.39964306e+00 2.65311509e-01
1.04476881e+00 -9.53715563e-01 -2.06281707e-01 2.86398411e-01
8.46926212e-01 -9.62383807e-01 6.10062778e-01 -5.43314278e-01
-3.72252852e-01 8.85699332e-01 3.36554170e-01 -6.86761141e-02
6.90441549e-01 4.85543191e-01 1.46126121e-01 -6.67616367e-01
-2.99002320e-01 -6.37887597e-01 2.26159573e-01 8.20445240e-01
-9.22158062e-02 2.50533283e-01 1.97630823e-01 1.86229944e-01
-6.29708707e-01 -7.62637258e-02 -3.01045422e-02 1.29561794e+00
-6.79657757e-01 -7.95336366e-01 -1.35592878e+00 2.76446790e-01
-6.29278496e-02 1.72799617e-01 -2.99164414e-01 6.01490855e-01
5.98427415e-01 8.59931588e-01 -7.03052506e-02 -8.45943570e-01
6.86624408e-01 -1.21258758e-01 8.24772358e-01 -2.04686224e-01
-7.36662149e-01 2.31905505e-01 8.66689757e-02 -1.05583346e+00
-9.22970712e-01 -9.54958439e-01 -1.46037245e+00 -5.20855844e-01
-5.40676638e-02 -5.34089617e-02 8.93860757e-01 9.51140881e-01
1.47494256e-01 3.34576070e-02 3.82796466e-01 -1.39221716e+00
2.75923908e-02 -4.92609441e-01 -9.28565621e-01 3.61171007e-01
3.30255151e-01 -1.03844881e+00 -1.39646143e-01 1.80875242e-01] | [7.1825480461120605, -1.2461543083190918] |
2f0a9992-c93e-473c-8ee1-afa6cbbd19b4 | invero-xl-making-cross-lingual-semantic-role | null | null | https://aclanthology.org/2021.emnlp-demo.36 | https://aclanthology.org/2021.emnlp-demo.36.pdf | InVeRo-XL: Making Cross-Lingual Semantic Role Labeling Accessible with Intelligible Verbs and Roles | Notwithstanding the growing interest in cross-lingual techniques for Natural Language Processing, there has been a surprisingly small number of efforts aimed at the development of easy-to-use tools for cross-lingual Semantic Role Labeling. In this paper, we fill this gap and present InVeRo-XL, an off-the-shelf state-of-the-art system capable of annotating text with predicate sense and semantic role labels from 7 predicate-argument structure inventories in more than 40 languages. We hope that our system – with its easy-to-use RESTful API and Web interface – will become a valuable tool for the research community, encouraging the integration of sentence-level semantics into cross-lingual downstream tasks. InVeRo-XL is available online at http://nlp.uniroma1.it/invero. | ['Roberto Navigli', 'Francesco Cecconi', 'Fabrizio Brignone', 'Riccardo Orlando', 'Simone Conia'] | null | null | null | null | emnlp-acl-2021-11 | ['semantic-role-labeling'] | ['natural-language-processing'] | [ 8.08760375e-02 1.98725477e-01 -5.25615215e-01 -5.80036640e-01
-8.59667778e-01 -1.10991824e+00 6.40559316e-01 5.75674474e-01
-5.02561986e-01 9.24746335e-01 4.34430748e-01 -6.02283657e-01
-1.47569686e-01 -5.80914795e-01 -3.54279280e-01 -2.73507982e-01
2.43254721e-01 7.32017457e-01 4.50040221e-01 -6.49380982e-01
1.05503574e-01 1.23694234e-01 -1.64613903e+00 7.14919448e-01
5.78043401e-01 5.38839042e-01 4.54466581e-01 1.13469407e-01
-4.16980147e-01 9.95803058e-01 -1.24830239e-01 -6.08742893e-01
-2.08530709e-01 -3.41937870e-01 -1.27105868e+00 -4.24176753e-01
4.20352966e-02 4.92974967e-01 2.29052290e-01 9.27434444e-01
3.58740121e-01 9.12695453e-02 1.01833008e-01 -1.03971565e+00
-6.43749833e-01 1.05672383e+00 1.31963283e-01 3.45756322e-01
7.98254192e-01 -1.95061982e-01 1.79613256e+00 -6.47920907e-01
1.22770810e+00 1.19946301e+00 3.42983514e-01 5.90016961e-01
-1.13204515e+00 -3.49879771e-01 1.11280628e-01 4.32210475e-01
-9.63100493e-01 -6.06974125e-01 9.11850393e-01 -2.98750550e-01
1.23491096e+00 1.93778232e-01 3.21144402e-01 9.67622340e-01
-3.01078051e-01 9.58355546e-01 1.45460033e+00 -9.30543423e-01
-5.44962287e-02 2.56301492e-01 2.24501714e-01 5.37552774e-01
3.11695844e-01 -4.15418744e-01 -8.28325868e-01 -4.66304012e-02
4.14477974e-01 -6.30317628e-01 -8.47191587e-02 -2.66164690e-01
-1.17694414e+00 7.39586473e-01 -6.52788952e-02 1.09920061e+00
-8.32027271e-02 -5.05479388e-02 8.17503750e-01 4.91338819e-01
6.58764303e-01 7.02391803e-01 -9.97216761e-01 -5.34801483e-01
-5.35522878e-01 3.09347332e-01 9.13653910e-01 6.93931460e-01
4.40213025e-01 -4.78310645e-01 4.44211990e-01 1.14944303e+00
3.68189752e-01 -1.09651819e-01 6.36310577e-01 -1.07225490e+00
5.64825714e-01 7.73577511e-01 3.08796942e-01 -5.53243339e-01
-3.07151258e-01 -3.37833948e-02 8.91018659e-02 -2.72806734e-01
7.58144081e-01 -4.31545265e-02 -1.53125986e-01 1.88557053e+00
3.51121068e-01 -5.26705742e-01 3.97256941e-01 7.33851194e-01
5.68155944e-01 3.32959890e-01 4.85207260e-01 -1.36769295e-01
1.87896717e+00 -6.43331945e-01 -8.28431249e-01 -4.21380430e-01
1.12827182e+00 -9.47263479e-01 1.59901989e+00 6.54607862e-02
-1.18943059e+00 -2.50275016e-01 -7.82157242e-01 -4.37069237e-01
-7.96066105e-01 -1.49881830e-02 9.37660813e-01 4.24007028e-01
-8.04525733e-01 4.27765340e-01 -6.85019493e-01 -8.05364788e-01
6.34957030e-02 -6.08736724e-02 -5.49037099e-01 -9.25515667e-02
-1.69256222e+00 1.06200695e+00 7.56273389e-01 -2.75174379e-01
-1.76060185e-01 -7.12025583e-01 -9.64090824e-01 -2.76068270e-01
7.90648341e-01 -4.03432667e-01 1.45354450e+00 -8.68335843e-01
-1.16792381e+00 1.73724377e+00 -4.23095524e-01 -3.04292858e-01
1.87100261e-01 -4.10310835e-01 -4.42356586e-01 5.62880039e-02
6.29244864e-01 3.53128135e-01 1.08173050e-01 -9.92274702e-01
-9.98853028e-01 -5.94400764e-01 3.94644797e-01 4.86403316e-01
-2.02449128e-01 8.10106039e-01 -1.73889935e-01 -7.25593150e-01
-7.81164551e-03 -8.49529803e-01 1.50696844e-01 -1.79947093e-01
8.79498348e-02 -9.40500557e-01 5.41708052e-01 -4.75793839e-01
1.21835613e+00 -1.99451363e+00 -6.56497478e-02 -4.71296936e-01
-2.82767773e-01 3.36027950e-01 2.86458820e-01 8.15173745e-01
-2.46917710e-01 2.15239540e-01 -3.26954097e-01 -2.44695023e-01
2.51857787e-01 6.76374018e-01 -1.92813560e-01 3.63661259e-01
1.07300341e-01 8.05561244e-01 -1.32991970e+00 -5.62316537e-01
4.40944165e-01 2.05189839e-01 -1.42661393e-01 -6.96778521e-02
-4.51514453e-01 3.26300919e-01 -4.68864918e-01 6.10441685e-01
1.06722087e-01 5.90278991e-02 6.67374074e-01 3.45357984e-01
-6.40330791e-01 1.25646496e+00 -9.52224910e-01 1.86037731e+00
-5.13167441e-01 3.99457663e-01 1.86933979e-01 -9.31545854e-01
6.49023473e-01 6.21946335e-01 4.28045899e-01 -5.88225782e-01
9.82754026e-03 6.91429734e-01 1.41474366e-01 -3.46848547e-01
3.75267386e-01 -4.35051084e-01 -4.53379035e-01 3.69181007e-01
1.18168265e-01 -1.61589503e-01 7.19394863e-01 -9.63401347e-02
6.69898212e-01 5.61908126e-01 8.73016000e-01 -6.19210422e-01
1.08837116e+00 4.59493548e-01 3.92429829e-01 2.43289888e-01
-2.63584435e-01 1.75477624e-01 6.65998816e-01 -4.38983887e-01
-8.62470865e-01 -6.96559370e-01 -4.67336923e-01 1.56633222e+00
-2.48149887e-01 -5.95912516e-01 -6.26591980e-01 -7.50822365e-01
-1.42550692e-01 1.17485595e+00 -3.29613835e-01 5.75259209e-01
-8.53667617e-01 -5.44799089e-01 8.27170312e-01 2.65334249e-01
1.98001444e-01 -1.39714110e+00 -3.95501107e-01 5.02453625e-01
-3.89988244e-01 -1.29054666e+00 8.77763480e-02 3.01457196e-01
-4.45811450e-01 -1.20015121e+00 -1.37873530e-01 -9.94975746e-01
8.60327110e-02 -5.09270057e-02 1.22445536e+00 -7.36512169e-02
2.38195270e-01 1.69423193e-01 -6.69598877e-01 -3.61148536e-01
-5.22847354e-01 2.10208714e-01 -1.32062748e-01 -1.33493796e-01
7.77477443e-01 -5.18290579e-01 -8.78176913e-02 8.14941674e-02
-7.11271346e-01 -9.01014656e-02 -1.83738083e-01 3.41220707e-01
7.09067881e-01 -1.28951430e-01 4.83895689e-01 -1.47054303e+00
7.89769232e-01 -4.27420288e-01 -6.37799382e-01 1.32714748e-01
-6.51349247e-01 1.30636394e-01 8.48971367e-01 2.78610140e-01
-1.21739495e+00 -2.81477030e-02 -7.29977369e-01 4.89970595e-01
-4.31061059e-01 7.17470109e-01 -5.04640281e-01 4.38926399e-01
4.59850013e-01 -8.24143738e-03 -2.64397413e-01 -8.16806078e-01
6.98299944e-01 5.77614248e-01 5.54030657e-01 -9.10133600e-01
4.33048368e-01 3.26124161e-01 -1.23891667e-01 -7.81609237e-01
-1.42913020e+00 -8.64487112e-01 -7.20328212e-01 5.73464595e-02
9.18332279e-01 -8.09817493e-01 -4.57397550e-01 4.30242360e-01
-1.10415733e+00 -4.00153995e-01 -2.79199332e-01 2.09014509e-02
-4.71123487e-01 3.43299478e-01 -5.34694910e-01 -4.41945434e-01
-2.78413683e-01 -6.28056407e-01 9.02486980e-01 -7.10470155e-02
-6.69895172e-01 -1.47716463e+00 9.12819654e-02 7.85359859e-01
-2.55609807e-02 5.10062054e-02 9.59890246e-01 -9.46148813e-01
2.39027981e-02 -2.96090860e-02 -2.36271229e-02 3.61584008e-01
1.94308996e-01 -2.56783873e-01 -7.88655937e-01 1.47712409e-01
2.97891330e-02 -4.36306059e-01 2.70141125e-01 1.52943701e-01
8.58944535e-01 -1.82744831e-01 -1.77160531e-01 1.16399467e-01
1.54428923e+00 3.91400196e-02 3.57337207e-01 7.94227540e-01
4.95735615e-01 9.37808037e-01 1.03061986e+00 -3.16355787e-02
7.39055634e-01 8.03285182e-01 -9.62848440e-02 2.35194027e-01
-2.19305634e-01 -3.80826533e-01 2.86492288e-01 6.52211189e-01
-7.29520470e-02 -1.67828649e-01 -1.26698315e+00 8.10690880e-01
-2.03742313e+00 -7.76093662e-01 -5.15636802e-01 1.86217582e+00
1.19909930e+00 2.19300792e-01 1.01404609e-02 1.79502249e-01
5.80717683e-01 3.07934493e-01 -2.85409596e-02 -5.83668292e-01
-2.39801198e-01 3.00830543e-01 2.15409458e-01 8.28420877e-01
-1.07961023e+00 1.55129707e+00 5.60890102e+00 8.78115475e-01
-8.41814637e-01 4.79509503e-01 1.82725281e-01 3.69089186e-01
-4.06132579e-01 2.71461040e-01 -1.04498374e+00 4.84218359e-01
1.08112299e+00 -2.81999320e-01 3.54992747e-01 9.18987870e-01
4.15368408e-01 -1.14646748e-01 -1.04066336e+00 7.11498976e-01
-1.04338348e-01 -1.44872522e+00 -2.96709955e-01 -2.95443535e-01
2.39275441e-01 3.83958340e-01 -4.24343526e-01 1.55886337e-01
3.56523305e-01 -6.69617712e-01 9.85452473e-01 -1.66726768e-01
7.02912688e-01 -4.98613536e-01 7.54565656e-01 2.74865985e-01
-1.21781397e+00 1.45738661e-01 4.78822812e-02 -5.08785248e-01
4.80898142e-01 4.79025245e-01 -6.15104079e-01 6.31659687e-01
5.86601675e-01 7.75720954e-01 -5.30318797e-01 3.70615691e-01
-8.56998026e-01 9.12310064e-01 -2.01037049e-01 -2.77794152e-01
3.50637823e-01 -1.01850301e-01 6.14838600e-01 1.38892353e+00
-1.84965417e-01 2.73507778e-02 4.01281565e-01 3.23745728e-01
-9.23773348e-02 5.87057531e-01 -5.79396546e-01 -3.37685376e-01
6.03097498e-01 1.04239142e+00 -9.61883962e-01 -4.48467910e-01
-5.73680222e-01 7.66767025e-01 5.56657672e-01 -3.01809222e-01
-5.11268675e-01 -4.30097789e-01 7.95195043e-01 4.15087432e-01
-4.33622338e-02 -4.35307145e-01 -5.54226875e-01 -1.13899326e+00
4.37501110e-02 -8.74667704e-01 8.59883189e-01 -8.34492564e-01
-1.36287415e+00 4.26968217e-01 3.33784938e-01 -6.19347632e-01
-4.72448617e-01 -9.48004484e-01 -1.06019698e-01 8.58356476e-01
-1.63936114e+00 -1.34247184e+00 4.31816578e-01 4.49083030e-01
6.75411940e-01 6.06581494e-02 1.07410550e+00 3.21540982e-01
-6.35604560e-02 7.47997761e-02 -2.64202774e-01 1.48756966e-01
6.57566607e-01 -1.36308944e+00 5.01705229e-01 7.95831978e-01
2.89467156e-01 7.54344642e-01 8.57204616e-01 -4.90592778e-01
-9.45827723e-01 -7.86647618e-01 2.00151777e+00 -8.02035689e-01
1.43253803e+00 -5.90887249e-01 -9.11078632e-01 9.05819476e-01
4.12667245e-01 -1.71137944e-01 9.57144022e-01 4.87187982e-01
-2.26886079e-01 1.92803904e-01 -1.02774632e+00 6.23556674e-01
1.37660480e+00 -8.90535414e-01 -1.42680991e+00 4.01592851e-01
7.81119883e-01 -2.38174617e-01 -8.37637663e-01 1.19499214e-01
3.47494811e-01 -6.33151948e-01 6.99919224e-01 -4.72327977e-01
2.61315197e-01 -4.38230991e-01 -2.18975306e-01 -9.32148278e-01
-1.39468014e-01 -8.27414513e-01 5.53264618e-01 1.60659528e+00
7.41674781e-01 -1.04455304e+00 4.21931624e-01 4.86884356e-01
-4.68771785e-01 -4.03687090e-01 -1.09149384e+00 -5.44006944e-01
1.68180332e-01 -9.51932311e-01 2.72583544e-01 1.10576916e+00
4.34108853e-01 9.34584618e-01 1.33784875e-01 -1.49522752e-01
2.68833131e-01 1.78612128e-01 2.57430911e-01 -1.40398049e+00
-2.24492788e-01 -1.72428533e-01 -4.96605560e-02 -6.47378862e-01
6.14444911e-01 -1.24200368e+00 -1.58965617e-01 -1.66685474e+00
-3.19477618e-01 -5.97879350e-01 -2.00780213e-01 1.05401957e+00
9.98411924e-02 4.52210635e-01 8.70702416e-02 2.63713241e-01
-6.67891145e-01 2.79972434e-01 1.07810867e+00 4.28960711e-01
-2.52241403e-01 -1.57803386e-01 -1.12465644e+00 9.50110376e-01
1.04179966e+00 -9.29154515e-01 -2.91936964e-01 -4.43930596e-01
4.96945202e-01 -3.55211258e-01 -6.86482415e-02 -6.59797609e-01
-1.13193378e-01 -2.22063184e-01 -2.78547674e-01 -3.00803661e-01
1.82116911e-01 -5.71505785e-01 -6.64430410e-02 2.08678558e-01
-2.40269348e-01 1.63377717e-01 1.24061383e-01 -1.37990296e-01
-4.00848091e-01 -7.62349188e-01 5.38811982e-01 -6.90834701e-01
-1.09501958e+00 -3.04384679e-01 -6.87352836e-01 5.37562847e-01
8.62202048e-01 3.97158600e-02 -2.72338390e-01 1.61102742e-01
-6.05425954e-01 9.55212936e-02 6.67385340e-01 6.39090538e-01
-9.06655639e-02 -8.71193171e-01 -6.19391203e-01 -1.60520792e-01
3.35503161e-01 -2.61727303e-01 -2.61405051e-01 3.40582728e-01
-5.56465209e-01 9.14800286e-01 4.59097698e-03 -4.68482524e-02
-1.32142437e+00 3.60225081e-01 4.80254218e-02 -5.37757218e-01
-4.78858024e-01 7.00619876e-01 -3.02216172e-01 -7.57786095e-01
-3.62997681e-01 -1.13248728e-01 -3.63323241e-01 3.54744613e-01
3.08615625e-01 6.22377545e-02 6.32577166e-02 -8.61180007e-01
-7.75832534e-01 1.52916133e-01 2.03826904e-01 -5.38317204e-01
1.50625420e+00 -2.78801173e-01 -6.84326291e-01 9.10744488e-01
8.12724948e-01 4.23595458e-01 -5.78986704e-01 8.21399391e-02
8.00489843e-01 -1.98622927e-01 -1.23220943e-01 -1.05118859e+00
-1.20857507e-01 3.75803798e-01 -9.30954143e-02 5.53584099e-01
7.66824543e-01 6.29701555e-01 6.14677012e-01 2.33801097e-01
6.05350316e-01 -1.29127681e+00 -4.67736214e-01 9.58033860e-01
9.08247948e-01 -1.01447964e+00 -1.98394090e-01 -7.47550547e-01
-7.06102133e-01 6.74812257e-01 2.18243912e-01 -2.11458597e-02
4.96408552e-01 1.98795870e-01 3.56519401e-01 -4.79341030e-01
-8.00326943e-01 -6.55008256e-01 6.72593433e-03 6.38136864e-01
1.18276894e+00 2.63676882e-01 -1.18627501e+00 5.72324634e-01
-6.48314118e-01 7.69480988e-02 2.43640080e-01 9.74194169e-01
-2.42909253e-01 -1.96534109e+00 -1.02508791e-01 -1.18990324e-01
-1.02096450e+00 -5.48123717e-01 -5.21484435e-01 1.09264994e+00
3.51138413e-01 1.07229114e+00 -1.56814620e-01 4.15008277e-01
3.26998264e-01 6.84445500e-01 5.20363808e-01 -1.19274414e+00
-6.07054532e-01 -1.24348097e-01 1.02757430e+00 -5.13125241e-01
-7.98990071e-01 -1.23316336e+00 -1.55277252e+00 1.31146396e-02
1.64584428e-01 3.92009437e-01 6.37049615e-01 1.05197406e+00
1.98397994e-01 2.26521984e-01 -9.46863517e-02 -4.40017611e-01
-1.53683517e-02 -9.48152959e-01 -3.92231911e-01 4.26841855e-01
-1.87233776e-01 -6.85554266e-01 -1.90926626e-01 1.85377449e-01] | [10.328627586364746, 9.451967239379883] |
f8665bad-a6b9-4ca4-9300-7ce5fc039273 | mmc-multi-modal-colorization-of-images-using | 2304.11993 | null | https://arxiv.org/abs/2304.11993v2 | https://arxiv.org/pdf/2304.11993v2.pdf | MMC: Multi-Modal Colorization of Images using Textual Descriptions | Handling various objects with different colors is a significant challenge for image colorization techniques. Thus, for complex real-world scenes, the existing image colorization algorithms often fail to maintain color consistency. In this work, we attempt to integrate textual descriptions as an auxiliary condition, along with the grayscale image that is to be colorized, to improve the fidelity of the colorization process. To do so, we have proposed a deep network that takes two inputs (grayscale image and the respective encoded text description) and tries to predict the relevant color components. Also, we have predicted each object in the image and have colorized them with their individual description to incorporate their specific attributes in the colorization process. After that, a fusion model fuses all the image objects (segments) to generate the final colorized image. As the respective textual descriptions contain color information of the objects present in the image, text encoding helps to improve the overall quality of predicted colors. In terms of performance, the proposed method outperforms existing colorization techniques in terms of LPIPS, PSNR and SSIM metrics. | ['Michael Blumenstein', 'Umapada Pal', 'Saumik Bhattacharya', 'Prasun Roy', 'Subhankar Ghosh'] | 2023-04-24 | null | null | null | null | ['colorization'] | ['computer-vision'] | [ 4.02873248e-01 -5.06154776e-01 1.60458520e-01 -3.06960791e-01
-3.73258650e-01 -5.26588678e-01 2.24613473e-01 1.31947786e-01
-3.61201882e-01 6.27981901e-01 -1.49240419e-01 -1.30079746e-01
2.08703905e-01 -7.38668799e-01 -6.21799290e-01 -7.90534675e-01
5.41180551e-01 2.94948090e-02 1.48574129e-01 -4.90381196e-02
4.82982486e-01 5.44460416e-01 -1.40351593e+00 3.17373782e-01
1.15470207e+00 1.19101322e+00 3.15047354e-01 6.86996043e-01
-4.51614171e-01 7.84253716e-01 -4.98529524e-01 -3.36764276e-01
2.91809112e-01 -5.60258210e-01 -5.41242421e-01 3.33847016e-01
4.31195050e-01 -5.19371748e-01 -1.66954979e-01 1.39425719e+00
1.51910901e-01 1.44710019e-01 5.46006024e-01 -1.31830299e+00
-1.01400089e+00 4.62256908e-01 -9.31693137e-01 -2.26788402e-01
2.16451436e-02 3.05286907e-02 7.31221080e-01 -8.16641629e-01
4.37669128e-01 1.18403757e+00 1.47824496e-01 4.28681612e-01
-1.16722858e+00 -7.03125894e-01 2.57226139e-01 4.23444301e-01
-1.27762425e+00 -8.99866596e-02 1.11221325e+00 -2.15696350e-01
1.12231769e-01 2.80727983e-01 8.74202013e-01 4.19103861e-01
4.33731489e-02 8.53604257e-01 1.18714762e+00 -5.11114299e-01
2.98796803e-01 3.40272605e-01 2.38937698e-02 6.69178843e-01
2.79436380e-01 -2.48890847e-01 -1.87905192e-01 2.93018371e-01
6.02307916e-01 2.90870070e-01 -1.82160750e-01 -2.95283526e-01
-1.05011964e+00 3.04013968e-01 7.05465078e-01 3.88112545e-01
-5.57433367e-01 2.49415249e-01 4.11665142e-02 -1.87611967e-01
2.77848721e-01 1.88448370e-01 -9.10035372e-02 5.66511489e-02
-9.69444275e-01 -6.27791956e-02 4.84864891e-01 8.91035616e-01
1.27962518e+00 5.94925135e-02 -2.10098401e-01 9.53472078e-01
2.19383240e-01 5.98228812e-01 1.63766950e-01 -9.23214078e-01
5.35815895e-01 1.01251316e+00 2.69322574e-01 -1.27225685e+00
-8.29186216e-02 -1.49135977e-01 -9.97580409e-01 5.02604306e-01
1.75800830e-01 -1.22687154e-01 -1.22581351e+00 1.47754097e+00
2.16597036e-01 6.36387691e-02 2.33573690e-01 1.16323912e+00
5.02974689e-01 1.03907835e+00 1.45201892e-01 -3.58177200e-02
1.23441303e+00 -1.10923362e+00 -7.61054277e-01 -4.19246525e-01
-1.92547981e-02 -1.17762363e+00 1.00734138e+00 4.85349000e-01
-1.12556398e+00 -8.10780287e-01 -1.31032872e+00 -1.42602518e-01
-4.00244266e-01 5.41432679e-01 4.19883162e-01 6.17468059e-01
-1.04370022e+00 3.02967191e-01 -6.27620816e-01 -2.13757128e-01
1.09038323e-01 1.11390874e-01 -1.51022956e-01 -4.46972877e-01
-1.00166917e+00 5.49429178e-01 6.78304136e-01 3.85263383e-01
-6.02377653e-01 -3.73162657e-01 -4.46428359e-01 2.67658353e-01
2.81545401e-01 -4.56429869e-01 7.42022753e-01 -1.51884723e+00
-1.26897550e+00 3.75778764e-01 -8.35619792e-02 5.53478859e-02
5.59832454e-01 -6.37802482e-02 -4.90367502e-01 3.12043160e-01
-1.59019783e-01 8.28224480e-01 8.94228995e-01 -1.83513057e+00
-1.15470946e+00 -2.39455387e-01 1.64618582e-01 4.33900356e-01
-3.81118894e-01 -2.76094854e-01 -1.32503808e+00 -6.20608985e-01
2.48944834e-01 -8.74502718e-01 -9.65893343e-02 2.53851831e-01
-6.32496357e-01 2.56783962e-01 8.24494243e-01 -7.97033906e-01
1.08879566e+00 -2.50669789e+00 4.92125079e-02 4.31898385e-01
1.44241646e-01 2.90894777e-01 -3.26727569e-01 2.28086397e-01
3.58875468e-02 1.64342653e-02 -2.83778608e-01 -3.10632557e-01
-1.09223902e-01 7.82041252e-02 6.69614971e-02 1.92943588e-01
3.01132530e-01 5.06679893e-01 -5.38846910e-01 -7.44034469e-01
4.46910888e-01 7.26280928e-01 -5.03053248e-01 3.27930331e-01
-2.37419993e-01 3.55304986e-01 -5.36854923e-01 4.12646443e-01
1.24646246e+00 1.36843733e-02 3.23079750e-02 -6.73265934e-01
-7.77120292e-02 -5.57016850e-01 -1.37684393e+00 1.40558970e+00
-4.79509532e-01 6.21447086e-01 7.97318742e-02 -7.91361511e-01
9.47738945e-01 -3.03409100e-02 6.94244266e-01 -8.44667137e-01
1.44996569e-01 7.50535131e-02 -1.09978817e-01 -3.04324716e-01
8.78648996e-01 6.36041164e-02 1.66873053e-01 4.33272243e-01
-4.08746153e-01 -1.12868600e-01 5.12834430e-01 2.77889311e-01
3.56934816e-01 2.80984864e-02 -2.61543006e-01 1.78691596e-01
7.76567459e-01 1.07473165e-01 4.85743612e-01 4.30798709e-01
-4.16821614e-02 7.49734879e-01 5.21519065e-01 -2.99506366e-01
-1.32026732e+00 -8.40117574e-01 3.05471063e-01 9.54104245e-01
7.07225680e-01 -4.23699692e-02 -9.07226145e-01 -3.47454667e-01
-2.35908628e-01 6.76380754e-01 -6.96000040e-01 -2.67856061e-01
-6.11396432e-01 -5.50304949e-01 1.91744223e-01 3.80627543e-01
7.28082478e-01 -9.51696277e-01 -4.29892063e-01 1.51736140e-01
-2.80496240e-01 -9.94311810e-01 -7.23379374e-01 7.59704337e-02
-5.22073388e-01 -1.21074796e+00 -9.56516206e-01 -9.54699755e-01
9.63574588e-01 4.82671887e-01 5.52468657e-01 3.33177328e-01
-3.60331237e-01 6.31778594e-03 -5.65060854e-01 -1.15448892e-01
-4.71566290e-01 -3.53630483e-01 -4.88980293e-01 4.49851066e-01
4.56682406e-02 -5.76098040e-02 -9.54475880e-01 5.79827502e-02
-1.35167384e+00 5.14087915e-01 8.90148818e-01 7.49553859e-01
6.39571190e-01 3.27171952e-01 2.46519279e-02 -8.26903582e-01
5.47869265e-01 -2.75669489e-02 -6.69683337e-01 7.18702972e-01
-5.98275602e-01 2.13234171e-01 9.75139797e-01 -2.84459919e-01
-1.36042488e+00 1.36979178e-01 -2.80472320e-02 -2.84540176e-01
-1.49026528e-01 3.81064296e-01 -1.21749721e-01 -9.88625139e-02
2.02622443e-01 5.32770336e-01 -1.30876424e-02 -3.55330735e-01
6.83362842e-01 5.64842761e-01 5.84207594e-01 -2.83922613e-01
9.13574934e-01 3.78411472e-01 -1.59508139e-01 -4.40341294e-01
-3.80152196e-01 -1.96309865e-01 -5.57199299e-01 -3.35971892e-01
1.10241044e+00 -7.09138513e-01 -6.40895665e-01 5.24407506e-01
-1.14206898e+00 -6.75857961e-02 2.57064193e-01 4.22382176e-01
-2.64571458e-01 5.16545713e-01 -4.97470617e-01 -8.30928147e-01
-3.65514576e-01 -1.31583571e+00 8.96784186e-01 7.05167294e-01
4.94175345e-01 -7.04953313e-01 -2.86225826e-01 5.04147932e-02
5.08317292e-01 1.14278831e-01 1.27919984e+00 2.14313366e-03
-6.67440176e-01 -2.49152362e-01 -8.82590890e-01 5.51732957e-01
3.30569416e-01 5.37808657e-01 -5.16754746e-01 -1.75536796e-01
-3.44278753e-01 4.96013872e-02 8.41684341e-01 3.11758161e-01
1.27222013e+00 -1.70801222e-01 -5.30575402e-02 6.48945928e-01
1.86846840e+00 7.62369812e-01 6.38844788e-01 3.65910113e-01
8.57269287e-01 4.51059014e-01 6.82373345e-01 5.88587165e-01
2.43865147e-01 4.07790959e-01 6.02409720e-01 -8.56537700e-01
-3.89011145e-01 -1.10042766e-01 7.06753358e-02 5.99539518e-01
1.76720053e-01 -3.40173393e-01 -4.96692777e-01 1.83135271e-01
-1.65442872e+00 -6.64179802e-01 -1.33366243e-03 1.87468493e+00
6.86105430e-01 -8.73394832e-02 -2.15807438e-01 5.81682697e-02
1.12895215e+00 2.22039893e-02 -7.52841234e-01 -3.04738820e-01
-2.33385429e-01 -7.09191114e-02 4.81986225e-01 2.30095014e-01
-8.66297245e-01 9.20146823e-01 5.68542528e+00 6.90391481e-01
-1.49621975e+00 -3.76736224e-01 8.94200504e-01 3.13263297e-01
-3.16223174e-01 -2.44491231e-02 -2.36194327e-01 5.45005322e-01
1.54883295e-01 -8.30625668e-02 9.25716937e-01 5.30901968e-01
2.57707030e-01 -4.53065574e-01 -7.60662258e-01 9.81121302e-01
2.66012043e-01 -8.77676010e-01 3.75407249e-01 -3.78410399e-01
9.51332867e-01 -5.63284159e-01 4.71471816e-01 -1.29160821e-01
2.07698092e-01 -6.04007244e-01 1.04108083e+00 7.08711684e-01
7.56908059e-01 -8.87158036e-01 6.82688832e-01 -6.32952973e-02
-1.28351223e+00 -1.88872352e-01 -4.55841541e-01 5.19919157e-01
1.30383670e-01 3.04008126e-01 -5.38048148e-01 7.07772493e-01
5.10006607e-01 6.15554333e-01 -9.15090621e-01 1.21016526e+00
-1.76358089e-01 3.01755935e-01 -8.68146420e-02 -2.81653050e-02
4.54375893e-01 -5.74110627e-01 -1.19780645e-01 1.01427233e+00
4.27514166e-01 9.16909650e-02 1.28576487e-01 9.97583866e-01
-6.35268837e-02 3.14890027e-01 1.78203970e-01 -2.04338133e-01
3.73732179e-01 1.36794090e+00 -9.28934097e-01 -6.43716455e-01
-4.83526647e-01 1.26699662e+00 -1.32918973e-02 7.83218145e-01
-9.79136467e-01 -6.81507170e-01 3.80485266e-01 -3.68698716e-01
2.29398355e-01 -4.88733388e-02 -3.83354545e-01 -9.94327664e-01
-1.49683533e-02 -7.13586628e-01 1.37869209e-01 -1.14373291e+00
-9.30839896e-01 7.31832683e-01 -4.04838920e-01 -1.48949611e+00
1.98149845e-01 -5.55721819e-01 -5.39511621e-01 1.00176930e+00
-1.60463917e+00 -1.31042671e+00 -7.05452204e-01 7.45318532e-01
3.28426182e-01 1.27421916e-01 2.73031622e-01 5.81156135e-01
-7.92086959e-01 4.53726798e-01 4.61264223e-01 9.14055184e-02
9.58235323e-01 -1.15911484e+00 1.06022097e-01 1.13576627e+00
-6.50778040e-02 4.03836668e-01 7.22608507e-01 -5.89109063e-01
-1.51165938e+00 -1.20639372e+00 1.14611983e-01 5.21258950e-01
2.54002661e-01 -1.97503567e-02 -8.18729758e-01 1.55096561e-01
3.61538887e-01 -2.22988695e-01 1.68838501e-01 -4.29340094e-01
-1.63628891e-01 -5.22295296e-01 -8.85400355e-01 6.49436593e-01
2.64513880e-01 -3.40900183e-01 -8.88262093e-02 3.97970929e-04
5.18374741e-01 -1.12726077e-01 -4.61164206e-01 -1.32180788e-02
6.70048773e-01 -8.49569917e-01 7.95931816e-01 -2.59989053e-01
5.42188287e-01 -7.07688332e-01 -1.51247410e-02 -1.27497959e+00
-3.22355390e-01 2.13843435e-01 6.56363606e-01 1.33165383e+00
4.45086211e-01 -2.55700648e-01 7.55378783e-01 8.39497566e-01
-5.75817153e-02 -3.39270711e-01 -3.47990960e-01 -2.00664133e-01
-1.75566196e-01 -5.05039915e-02 7.16196477e-01 5.77907920e-01
-2.76643693e-01 -1.52892858e-01 -7.29993224e-01 2.05191866e-01
6.07920647e-01 4.63266641e-01 7.31987238e-01 -8.26277494e-01
6.68891426e-03 -5.42420685e-01 -2.09526554e-01 -8.19277465e-01
-1.61944374e-01 -5.51773727e-01 2.40690902e-01 -1.91226518e+00
5.65941572e-01 -5.80205917e-01 -5.43906093e-01 4.70872730e-01
-6.31793141e-01 6.06767654e-01 6.27385437e-01 9.71365049e-02
-6.84578836e-01 4.92758483e-01 1.42649090e+00 -5.21105886e-01
-2.78397769e-01 -3.07421207e-01 -7.89562285e-01 3.97803634e-01
7.18321383e-01 -7.89562687e-02 -1.88920364e-01 -6.55054510e-01
-2.72054709e-02 2.09291086e-01 1.29256845e-01 -1.11694789e+00
3.93538207e-01 -4.45879817e-01 8.55946660e-01 -7.55478382e-01
3.68666619e-01 -1.23712015e+00 3.44122887e-01 5.20817161e-01
-4.41518158e-01 2.60470267e-02 9.62745622e-02 4.14789647e-01
-4.38578546e-01 -2.13502616e-01 1.08063555e+00 6.95220530e-02
-9.56046939e-01 3.25144976e-01 -2.37469807e-01 -4.07155812e-01
1.12822461e+00 -4.13960040e-01 -3.70651931e-01 -4.17497486e-01
-5.78475893e-01 1.40384361e-01 7.81448066e-01 2.74497151e-01
7.55725145e-01 -1.32687724e+00 -5.27979553e-01 2.69510865e-01
1.87251940e-01 -3.50301832e-01 7.26372123e-01 6.15416586e-01
-9.54041481e-01 1.13603763e-01 -6.63238525e-01 -2.70622790e-01
-1.16988075e+00 9.08089042e-01 1.23015270e-01 1.82658315e-01
-3.59540761e-01 4.80348289e-01 4.07472044e-01 3.08742881e-01
2.80658901e-01 -4.59934264e-01 -3.34656239e-01 -5.82780987e-02
4.29072887e-01 2.33428821e-01 -9.63130072e-02 -6.94506168e-01
-7.69850686e-02 9.06617880e-01 -1.98683113e-01 -2.25421384e-01
1.19484365e+00 -5.25546134e-01 -3.42485487e-01 7.23061413e-02
1.44299424e+00 5.29509559e-02 -1.39353395e+00 -2.64025182e-01
-2.92784870e-01 -6.49704099e-01 -5.50649613e-02 -1.02959275e+00
-1.68624377e+00 9.08106685e-01 9.72146749e-01 5.66452965e-02
1.65512669e+00 -3.59899849e-01 7.63454676e-01 7.72737190e-02
-3.22235078e-02 -1.18239212e+00 1.96495131e-01 5.74255139e-02
4.92759287e-01 -1.13491893e+00 -1.18885338e-01 -3.23474705e-01
-8.75207961e-01 1.32941878e+00 6.92565382e-01 -5.22340648e-02
1.79125741e-01 1.79438107e-02 2.74622500e-01 1.90599233e-01
-3.00109386e-01 -2.42142990e-01 4.49393541e-01 3.11514407e-01
2.72855401e-01 -2.35601738e-02 -3.98286939e-01 3.18318278e-01
2.04069436e-01 -2.23258689e-01 5.66648424e-01 5.83042860e-01
-5.34364522e-01 -1.04145944e+00 -6.22750700e-01 2.78468341e-01
-3.68716836e-01 -1.50904000e-01 -4.05894846e-01 4.57302690e-01
2.34570041e-01 1.01576912e+00 -3.63710499e-03 -4.33473080e-01
2.77275711e-01 -2.54046559e-01 1.17802978e-01 -9.03419107e-02
-2.84715563e-01 3.78358364e-01 -4.09921020e-01 -3.83870572e-01
-3.79289061e-01 -1.64906979e-01 -1.28743219e+00 -3.52545410e-01
-2.54216075e-01 2.29495704e-01 1.03141642e+00 5.49399972e-01
3.88686508e-02 7.41543829e-01 8.73044431e-01 -7.22956657e-01
5.68870567e-02 -6.65586114e-01 -8.45918000e-01 8.27454746e-01
3.42346013e-01 -3.18778783e-01 -9.37121063e-02 3.43078732e-01] | [11.191059112548828, -1.237737774848938] |
11d37e9d-27b4-48d4-8d1f-9a7acdfb5481 | schema-guided-semantic-accuracy-faithfulness | 2301.12568 | null | https://arxiv.org/abs/2301.12568v1 | https://arxiv.org/pdf/2301.12568v1.pdf | Schema-Guided Semantic Accuracy: Faithfulness in Task-Oriented Dialogue Response Generation | Ensuring that generated utterances are faithful to dialogue actions is crucial for Task-Oriented Dialogue Response Generation. Slot Error Rate (SER) only partially measures generation quality in that it solely assesses utterances generated from non-categorical slots whose values are expected to be reproduced exactly. Utterances generated from categorical slots, which are more variable, are not assessed by SER. We propose Schema-Guided Semantic Accuracy (SGSAcc) to evaluate utterances generated from both categorical and non-categorical slots by recognizing textual entailment. We show that SGSAcc can be applied to evaluate utterances generated from a wide range of dialogue actions in the Schema Guided Dialogue (SGD) dataset with good agreement with human judgment. We also identify a previously overlooked weakness in generating faithful utterances from categorical slots in unseen domains. We show that prefix tuning applied to T5 generation can address this problem. We further build an ensemble of prefix-tuning and fine-tuning models that achieves the lowest SER reported and high SGSAcc on the SGD dataset. | ['Bill Byrne', 'Weizhe Lin', 'Jinghong Chen'] | 2023-01-29 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [ 5.29494941e-01 9.63757098e-01 2.06695706e-01 -6.79300725e-01
-1.34959590e+00 -8.50364745e-01 1.01461387e+00 -9.04821977e-02
-1.42597437e-01 1.05639744e+00 8.33931327e-01 -3.41518521e-01
9.26824212e-02 -8.06574345e-01 -2.76857466e-01 -8.37138221e-02
6.36125684e-01 1.16468871e+00 3.13056618e-01 -9.56869662e-01
1.77774787e-01 -1.83486715e-01 -1.22901368e+00 9.77462292e-01
9.88287926e-01 7.12825835e-01 -5.94589002e-02 9.10535872e-01
-5.65197229e-01 9.43380415e-01 -1.24513400e+00 -5.01795530e-01
-2.68872023e-01 -1.03066123e+00 -1.61308861e+00 6.63670674e-02
1.08726606e-01 -3.46840501e-01 1.74513176e-01 5.61517119e-01
6.64192021e-01 1.98031276e-01 9.74701107e-01 -1.07468247e+00
-5.07077515e-01 1.11069012e+00 4.98433739e-01 -9.07554403e-02
1.08100927e+00 1.52596906e-01 1.26521754e+00 -7.47156799e-01
8.85236382e-01 1.73249757e+00 6.69208586e-01 1.16968966e+00
-1.54379332e+00 -5.41404188e-02 -3.63454580e-01 -3.07110250e-01
-8.48592699e-01 -7.69042075e-01 4.47264344e-01 -2.49280512e-01
1.63123679e+00 5.76339364e-01 -3.92680317e-02 1.42687619e+00
-3.30892861e-01 7.72150755e-01 1.07804143e+00 -6.96796775e-01
3.94214183e-01 3.91394734e-01 -1.47102863e-01 2.75444001e-01
-4.82959300e-01 -5.43037243e-03 -6.44670010e-01 -2.89844096e-01
4.43515956e-01 -9.43271995e-01 -1.36523368e-02 1.16336748e-01
-1.30536222e+00 1.26293230e+00 -1.13124870e-01 2.03955054e-01
-1.92234188e-01 -1.44654140e-01 7.39231050e-01 6.23061240e-01
4.54654098e-01 1.15485990e+00 -7.88086593e-01 -7.40799248e-01
-5.43217361e-01 7.54500747e-01 1.28456688e+00 1.35262465e+00
5.85985065e-01 2.01746851e-01 -7.83332825e-01 1.29054439e+00
1.30549422e-03 4.69831616e-01 6.50996089e-01 -1.34896076e+00
6.80128515e-01 6.90534711e-01 4.88264203e-01 -5.76617420e-01
-5.57639599e-01 1.34717807e-01 -4.16134775e-01 -4.39806968e-01
7.15460837e-01 -3.33144754e-01 -4.70447928e-01 2.00856996e+00
2.23320499e-01 -9.19874251e-01 6.96313381e-01 7.41051912e-01
1.25163019e+00 6.66557491e-01 3.71273428e-01 -1.46996275e-01
1.43574488e+00 -6.35071397e-01 -7.35641003e-01 -3.06897968e-01
1.18063509e+00 -8.91196012e-01 1.58523560e+00 -1.32371232e-01
-1.19724464e+00 -5.82123756e-01 -5.92365205e-01 -1.88963741e-01
-2.12347567e-01 7.49697387e-02 4.94127572e-01 7.78730810e-01
-1.03841281e+00 3.67190421e-01 -1.62588164e-01 -4.29243177e-01
-2.36740232e-01 -7.73659721e-02 -9.20290500e-02 4.21911657e-01
-1.84510636e+00 1.15100467e+00 5.60151398e-01 -5.03713071e-01
-7.76736438e-01 -5.36137521e-01 -1.02008533e+00 -2.46664003e-01
3.87770772e-01 -6.77456081e-01 1.98584878e+00 -6.06425703e-01
-1.74802208e+00 9.25318539e-01 -2.56501675e-01 -7.19665825e-01
5.33127069e-01 9.51472297e-02 -2.49400347e-01 5.12955859e-02
4.07261938e-01 1.01377273e+00 5.07757246e-01 -9.59839046e-01
-5.26327610e-01 -1.84669327e-02 1.11996748e-01 5.89516282e-01
7.05666915e-02 4.68872115e-02 3.44834566e-01 -3.82412791e-01
2.61168450e-01 -8.67989063e-01 1.03477821e-01 -7.11586058e-01
-6.29568458e-01 -8.83947134e-01 2.41029888e-01 -4.57403868e-01
1.20368719e+00 -1.55847633e+00 -1.82724129e-02 -2.37859055e-01
-2.20916823e-01 1.47426695e-01 -2.33627513e-01 8.53016734e-01
2.50943452e-01 1.63074717e-01 -2.13174865e-01 -1.97620258e-01
6.86294913e-01 3.82472903e-01 -6.77960873e-01 -5.65652013e-01
6.32888019e-01 1.09326613e+00 -9.56253529e-01 -4.46992576e-01
1.53875381e-01 -9.79052186e-02 -6.82619154e-01 6.53393626e-01
-8.14515054e-01 1.71939179e-01 -3.48949045e-01 2.57517397e-01
1.74268052e-01 -1.80898100e-01 2.83677846e-01 1.21724032e-01
1.13511942e-01 1.18931711e+00 -7.14654446e-01 1.65530789e+00
-5.75399339e-01 2.51800418e-01 -3.89418036e-01 -5.79476118e-01
1.20833719e+00 6.43694997e-01 -1.17041409e-01 -9.58568811e-01
-2.00322062e-01 4.79806155e-01 -5.30510060e-02 -5.12606800e-01
1.16941988e+00 -5.26631534e-01 -8.05725873e-01 8.47926736e-01
3.17069322e-01 -8.55862737e-01 7.48211294e-02 4.55925524e-01
1.08652925e+00 3.46927010e-02 3.22321981e-01 -2.26613179e-01
4.64573145e-01 4.69455153e-01 1.78673312e-01 1.02809620e+00
-1.29922464e-01 5.64428151e-01 7.40718067e-01 -1.38597235e-01
-1.31254399e+00 -1.06576669e+00 -1.56125069e-01 1.61382937e+00
-2.65059263e-01 -3.41962337e-01 -1.06520975e+00 -8.26789081e-01
-2.03175187e-01 1.40276575e+00 -3.73749018e-01 -1.85363054e-01
-4.22848582e-01 -4.27149683e-01 1.15741193e+00 3.38970631e-01
3.56783032e-01 -1.44612277e+00 -4.41509068e-01 6.56160414e-01
-9.41590309e-01 -1.18865561e+00 -2.40245774e-01 4.41009328e-02
-4.51706380e-01 -8.08753967e-01 -5.67601621e-01 -5.00077605e-01
2.70911694e-01 -2.03015178e-01 1.67140734e+00 -1.39598832e-01
2.24983290e-01 4.09804255e-01 -6.97491527e-01 -3.27880442e-01
-1.52104914e+00 3.72304291e-01 -1.77974924e-01 -5.12220442e-01
6.30619526e-01 7.70480037e-02 -1.32333323e-01 6.72404289e-01
-7.52009511e-01 2.52253950e-01 2.44481955e-02 1.09614229e+00
2.75374167e-02 -3.87923032e-01 1.19312155e+00 -1.23981750e+00
1.15621400e+00 -4.58921701e-01 1.42995743e-02 2.47385636e-01
-5.88011861e-01 2.70843416e-01 6.04847610e-01 -9.84822437e-02
-1.51476824e+00 -3.95046979e-01 -6.75999343e-01 4.97297376e-01
-5.57034016e-01 1.92571267e-01 -3.56329419e-02 6.80917203e-01
1.27220762e+00 2.74601698e-01 1.62817746e-01 -2.15846181e-01
4.75145072e-01 9.18720305e-01 3.52413476e-01 -1.04147208e+00
2.37486452e-01 -1.68230012e-01 -4.51024085e-01 -6.78454995e-01
-9.59950924e-01 -3.47650081e-01 -2.90289104e-01 -1.33000433e-01
6.60765827e-01 -7.59150028e-01 -4.97289747e-01 1.88041568e-01
-1.37151206e+00 -6.94038868e-01 -4.06604171e-01 2.53346562e-03
-1.12582564e+00 3.80823761e-01 -7.60310948e-01 -1.12432599e+00
-4.75896269e-01 -8.24253440e-01 1.48242486e+00 -3.32506932e-02
-1.21928394e+00 -1.14841557e+00 5.50540090e-02 5.22084177e-01
5.79880118e-01 3.10393684e-02 1.17753136e+00 -1.23463476e+00
7.78343603e-02 2.15942532e-01 -1.35306969e-01 1.39486000e-01
8.52717757e-02 -2.78320611e-01 -9.28415477e-01 3.09941620e-02
-4.40079384e-02 -1.11779904e+00 4.38404828e-01 -3.33379917e-02
4.00935143e-01 -7.71892846e-01 2.58531094e-01 -1.32631317e-01
8.08062732e-01 2.52524167e-01 7.02425361e-01 2.33864397e-01
1.16443351e-01 1.22395551e+00 1.11291122e+00 4.97797579e-01
5.96316218e-01 1.01031303e+00 -1.61901340e-02 2.13495627e-01
-1.05910018e-01 -6.69050455e-01 3.92059177e-01 4.63705063e-01
4.85591739e-01 -4.12847430e-01 -8.67044210e-01 6.91301763e-01
-1.64597571e+00 -1.00972593e+00 -3.60420823e-01 1.72856510e+00
1.63927948e+00 1.75253928e-01 3.74139100e-01 9.72362235e-02
5.15775025e-01 1.17017351e-01 -1.55680716e-01 -9.76851225e-01
-2.80524373e-01 3.18146884e-01 -2.83192575e-01 7.92663097e-01
-7.56027758e-01 1.39646220e+00 6.93565464e+00 9.21939790e-01
-3.71343940e-01 5.13727143e-02 5.09514630e-01 3.89576793e-01
-8.48971665e-01 2.91989720e-03 -9.90469813e-01 4.35109496e-01
1.36286771e+00 -3.09515417e-01 2.29930937e-01 9.83073056e-01
9.58937928e-02 9.08199549e-02 -1.21155834e+00 4.44224536e-01
-1.46535397e-01 -1.24142945e+00 1.60711497e-01 -4.54740524e-01
7.47592151e-01 -4.96032238e-01 -5.80829792e-02 7.74697304e-01
6.94495976e-01 -1.05832982e+00 6.94093227e-01 6.40993565e-02
9.60789263e-01 -4.86596048e-01 7.99160242e-01 4.46333617e-01
-4.95466948e-01 2.96355277e-01 -6.00298047e-01 -4.16548029e-02
2.30129287e-01 1.21560939e-01 -1.88731849e+00 2.73209751e-01
4.12423201e-02 3.14301997e-02 -3.66595626e-01 -1.19157797e-02
-2.73116589e-01 6.31072819e-01 6.13461807e-03 -4.19062495e-01
4.51379210e-01 1.90361470e-01 5.85457087e-01 1.46374357e+00
1.24072634e-01 1.68178901e-01 -1.49351940e-01 1.21953273e+00
2.82761335e-01 7.74971172e-02 -5.96078753e-01 -2.07982853e-01
8.44404042e-01 8.55845213e-01 -9.86948833e-02 -4.67445672e-01
1.48673460e-01 9.76282001e-01 1.57117963e-01 6.29201233e-02
-4.90145743e-01 -1.57971874e-01 6.34993255e-01 -1.91930622e-01
-1.54239506e-01 1.62166223e-01 -2.58435279e-01 -7.42638469e-01
-2.09711462e-01 -1.27097213e+00 5.34677505e-01 -8.08639467e-01
-1.40191996e+00 9.53574181e-01 4.77143750e-02 -1.09427106e+00
-1.38032484e+00 -4.01627868e-01 -2.88549155e-01 1.00661755e+00
-1.10692835e+00 -8.93990338e-01 -7.60810897e-02 4.68690693e-01
1.05540025e+00 -3.47833335e-01 1.50295091e+00 -3.47855985e-01
-1.06419593e-01 7.43298650e-01 -3.98634970e-01 1.27022937e-01
7.35551894e-01 -1.65176511e+00 8.75683486e-01 3.73008728e-01
-1.66968942e-01 5.92077971e-01 1.01377416e+00 -5.39523304e-01
-7.49577224e-01 -1.13252211e+00 1.57557034e+00 -8.54534805e-01
3.34799588e-01 -2.33883619e-01 -9.14407790e-01 4.79246616e-01
1.10983141e-01 -9.08999741e-01 7.76647627e-01 2.25747421e-01
-3.05023670e-01 5.68264782e-01 -1.26072454e+00 4.91150200e-01
1.04781342e+00 -8.67398977e-01 -1.19486463e+00 4.63732839e-01
1.00730741e+00 -7.90910780e-01 -1.04340458e+00 2.25448459e-01
1.10033959e-01 -1.10650277e+00 6.62523150e-01 -9.06861842e-01
6.84930742e-01 3.89998257e-02 -3.54123592e-01 -1.61476040e+00
6.74692690e-02 -8.40821385e-01 1.54480994e-01 1.39250278e+00
8.53758931e-01 -6.09716475e-01 6.65819585e-01 1.04220140e+00
-4.90092784e-01 -1.86429679e-01 -9.17756677e-01 -7.62048364e-01
2.70116568e-01 -3.86494666e-01 9.59550202e-01 9.08757508e-01
4.95790064e-01 8.10242891e-01 -2.38333479e-01 -5.15862763e-01
2.94990987e-01 5.27861854e-03 9.15947437e-01 -8.47809911e-01
-2.58389145e-01 -2.65598059e-01 5.95087670e-02 -1.18315768e+00
2.11428851e-01 -8.95327866e-01 4.25191849e-01 -1.37089050e+00
-2.18985796e-01 -6.27363980e-01 5.34617722e-01 4.18554783e-01
-3.01963955e-01 1.05437540e-01 -8.53709131e-02 4.57047932e-02
-6.08081579e-01 6.86106265e-01 1.13655961e+00 1.42270610e-01
-1.84890613e-01 9.10295639e-03 -9.05407608e-01 3.61776829e-01
8.42301786e-01 -3.31355304e-01 -6.31754458e-01 -3.32537323e-01
1.80909619e-01 5.38581371e-01 1.50358543e-01 -6.41918838e-01
-2.25217834e-01 -3.40770483e-01 2.93341354e-02 -3.95425528e-01
3.83536994e-01 -1.48939654e-01 -5.13870327e-04 1.81975126e-01
-1.15741754e+00 -1.93648577e-01 1.99152492e-02 8.80365595e-02
-2.12329119e-01 -4.96831030e-01 5.48259735e-01 -3.75165373e-01
-6.80576980e-01 -4.51366812e-01 -5.10583878e-01 6.43858135e-01
5.23240387e-01 -2.76327282e-01 -6.94393635e-01 -6.51972771e-01
-7.43924439e-01 6.09016232e-02 2.57300049e-01 5.01085222e-01
6.69751763e-01 -1.46923411e+00 -1.00456393e+00 2.26965517e-01
5.77596068e-01 -1.43235326e-01 2.27042630e-01 3.30961496e-01
-2.65694678e-01 9.63723421e-01 -2.01462698e-03 -7.16003478e-01
-1.05752289e+00 -1.33436710e-01 3.82211655e-01 -2.75278538e-01
-1.70820698e-01 9.65219140e-01 -9.38969031e-02 -1.14465296e+00
3.68045014e-03 -2.11578637e-01 -2.86098737e-02 4.68107946e-02
5.03119648e-01 2.49092907e-01 1.44928366e-01 -7.06110656e-01
-4.81636934e-02 -2.99033135e-01 -6.69357479e-02 -6.37619615e-01
8.38182807e-01 -3.23129475e-01 2.69496869e-02 4.30410206e-01
1.03372717e+00 -2.88717985e-01 -9.81456399e-01 -3.63781989e-01
1.97938472e-01 -3.71748298e-01 -5.67407966e-01 -1.22610199e+00
9.85803604e-02 6.57836914e-01 -7.92798549e-02 8.52167785e-01
6.24886930e-01 1.20881729e-01 9.08680201e-01 5.12308061e-01
6.46615505e-01 -1.48951757e+00 4.58697140e-01 1.06305563e+00
1.17808545e+00 -1.12362194e+00 -7.03922927e-01 -2.21982837e-01
-1.49558914e+00 1.00256097e+00 7.53951252e-01 4.69328195e-01
-3.09708923e-01 -1.11001261e-01 3.25945616e-01 -1.27845556e-01
-1.21465528e+00 -2.44879052e-01 1.32533520e-01 9.56705034e-01
7.96360791e-01 3.01450133e-01 -3.80652905e-01 6.36933267e-01
-9.60090101e-01 -4.80625480e-01 5.39871871e-01 5.01093984e-01
-6.54018700e-01 -1.21922255e+00 -1.81641191e-01 4.31662917e-01
-2.59339839e-01 -2.27353469e-01 -1.06362748e+00 5.65524817e-01
-4.73452985e-01 1.51292372e+00 3.51988226e-02 -3.77323627e-01
5.27353227e-01 6.41488910e-01 2.61127770e-01 -9.58383441e-01
-7.34713554e-01 -1.57780141e-01 1.21870720e+00 -4.51804817e-01
-1.42313287e-01 -5.69229424e-01 -1.30851710e+00 -4.58729535e-01
-1.98210031e-01 6.34814322e-01 3.81313711e-01 8.25828493e-01
2.71105856e-01 3.04308563e-01 8.05721939e-01 -3.08313161e-01
-1.20813859e+00 -1.44027984e+00 -3.49385738e-01 9.74175096e-01
2.28972081e-02 -3.48445058e-01 -4.31665689e-01 -1.88720167e-01] | [12.729377746582031, 8.082128524780273] |
b33d0921-fd85-4cab-a5aa-3685eded5a45 | silent-killer-optimizing-backdoor-trigger | 2301.02615 | null | https://arxiv.org/abs/2301.02615v1 | https://arxiv.org/pdf/2301.02615v1.pdf | Silent Killer: Optimizing Backdoor Trigger Yields a Stealthy and Powerful Data Poisoning Attack | We propose a stealthy and powerful backdoor attack on neural networks based on data poisoning (DP). In contrast to previous attacks, both the poison and the trigger in our method are stealthy. We are able to change the model's classification of samples from a source class to a target class chosen by the attacker. We do so by using a small number of poisoned training samples with nearly imperceptible perturbations, without changing their labels. At inference time, we use a stealthy perturbation added to the attacked samples as a trigger. This perturbation is crafted as a universal adversarial perturbation (UAP), and the poison is crafted using gradient alignment coupled to this trigger. Our method is highly efficient in crafting time compared to previous methods and requires only a trained surrogate model without additional retraining. Our attack achieves state-of-the-art results in terms of attack success rate while maintaining high accuracy on clean samples. | ['Lior Rokach', 'Gallil Maimon', 'Tzvi Lederer'] | 2023-01-05 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [ 4.43043560e-01 6.49394765e-02 -5.98633215e-02 2.01315340e-02
-6.72360659e-01 -1.29214454e+00 8.48715007e-01 -2.78732032e-01
-7.51812875e-01 8.03123653e-01 -2.26206332e-01 -6.95361674e-01
4.39747691e-01 -7.94711828e-01 -1.14706457e+00 -9.20723796e-01
9.50632710e-03 3.34824741e-01 3.27257037e-01 -3.92232597e-01
-2.55208910e-02 6.00582242e-01 -7.52738833e-01 5.17121516e-02
4.88386512e-01 4.62711245e-01 -6.12028182e-01 7.73077428e-01
4.31326807e-01 7.30668902e-01 -1.16424954e+00 -8.02896559e-01
6.15315914e-01 -3.13320994e-01 -6.81065619e-01 -6.10396385e-01
4.95126337e-01 -5.89746416e-01 -5.32038093e-01 1.37372744e+00
6.50074422e-01 -1.64284378e-01 2.98484623e-01 -1.46330225e+00
-4.81644034e-01 8.16278696e-01 -3.26738685e-01 5.80459982e-02
8.93106386e-02 5.77365637e-01 5.75522602e-01 -4.32069421e-01
6.02772713e-01 1.24659288e+00 6.91286325e-01 1.34069681e+00
-1.47817791e+00 -1.40465832e+00 9.76902694e-02 -1.81794748e-01
-1.04097545e+00 -6.52998090e-01 8.37784588e-01 -2.84703642e-01
4.72389966e-01 5.98671138e-01 5.12517750e-01 2.05811715e+00
1.38795286e-01 5.30355632e-01 9.44952667e-01 -2.71434128e-01
5.91110229e-01 3.85600813e-02 -3.28513682e-02 6.42192960e-01
3.39193523e-01 5.56969881e-01 -1.67113632e-01 -8.99126470e-01
4.88629788e-01 7.30209798e-02 -3.40154201e-01 -2.24687994e-01
-7.68756509e-01 1.07042408e+00 4.86884147e-01 -7.54818395e-02
-6.83473721e-02 5.92941523e-01 3.74668330e-01 4.65511292e-01
3.36039484e-01 7.83211231e-01 -3.92877698e-01 1.89541534e-01
-6.15698040e-01 4.79834616e-01 1.17204952e+00 3.74569803e-01
2.31734633e-01 4.60868537e-01 -1.98991299e-01 2.69069374e-01
4.64455262e-02 8.62132192e-01 4.48963881e-01 -6.56095088e-01
3.72095317e-01 6.57822490e-02 6.46187291e-02 -9.21791792e-01
-2.21429933e-02 -5.50430834e-01 -6.07949436e-01 6.76692426e-01
6.57567978e-01 -5.83419561e-01 -1.21790004e+00 2.19263196e+00
6.20143950e-01 4.33984131e-01 1.66791320e-01 5.46015024e-01
1.88697577e-01 3.03186446e-01 -8.54407400e-02 1.40524074e-01
1.13573527e+00 -7.58467615e-01 -4.27440017e-01 -2.17502445e-01
5.41777015e-01 -3.15639466e-01 9.39226866e-01 5.06424010e-01
-8.53248656e-01 8.64499882e-02 -1.26536083e+00 3.12830985e-01
-6.83823466e-01 -6.02951646e-01 4.04955953e-01 1.27799118e+00
-7.13483691e-01 7.07245588e-01 -9.91442621e-01 2.22098842e-01
8.52519214e-01 6.62889183e-01 -5.61286211e-01 5.05107641e-02
-1.48835433e+00 8.02170753e-01 1.97425991e-01 -3.65378767e-01
-1.74356425e+00 -8.93476784e-01 -6.80498481e-01 -2.56758958e-01
2.31820583e-01 -6.37694120e-01 1.16230977e+00 -5.94720662e-01
-1.63689721e+00 6.16852582e-01 3.04219961e-01 -1.02258158e+00
9.26482618e-01 -2.06516549e-01 -4.25643444e-01 1.76869661e-01
-1.31704941e-01 4.13919389e-01 1.41073561e+00 -1.27756190e+00
-1.93353146e-01 -1.91940427e-01 9.66055393e-02 -2.58683980e-01
-5.97464681e-01 1.91038564e-01 9.23812762e-02 -8.36355150e-01
-5.23550570e-01 -9.77268875e-01 -2.45477915e-01 8.29544440e-02
-8.63234341e-01 3.07922900e-01 1.48297489e+00 -3.75801742e-01
9.51855958e-01 -2.17334318e+00 -2.74266656e-02 4.35121864e-01
6.88872218e-01 9.32031989e-01 -2.33693808e-01 2.48693183e-01
-2.14545116e-01 3.31840158e-01 -4.08042103e-01 -4.20143157e-01
1.66145757e-01 1.82186574e-01 -1.06259143e+00 9.12634909e-01
-7.36009702e-02 8.94052088e-01 -1.25429928e+00 1.83708459e-01
-3.15640680e-02 5.23780465e-01 -7.28807330e-01 2.34326661e-01
-2.87800848e-01 3.88747096e-01 -3.42877924e-01 4.31971431e-01
6.38985932e-01 1.31345227e-01 -1.36464700e-01 6.95584938e-02
4.40052629e-01 2.80487031e-01 -7.13369071e-01 1.05136847e+00
-1.45854905e-01 2.90838242e-01 3.42530347e-02 -6.93376958e-01
8.40693533e-01 4.70173866e-01 -4.15266827e-02 1.19999588e-01
4.73277509e-01 2.01597229e-01 1.44200072e-01 -7.54550472e-02
-8.65511224e-02 -1.63789853e-01 -1.83778599e-01 8.24192703e-01
-3.98280993e-02 4.26539183e-02 -3.48181486e-01 1.43756434e-01
1.77837420e+00 -1.59075230e-01 1.43463343e-01 -6.79766573e-03
2.21076429e-01 -1.73804715e-01 4.54855531e-01 1.11263382e+00
-2.04490915e-01 2.40313649e-01 7.11226106e-01 -6.24057114e-01
-1.08494997e+00 -1.18054342e+00 2.04464942e-01 8.78205299e-01
-2.28830829e-01 -2.40364462e-01 -1.21454155e+00 -1.27140903e+00
2.68431276e-01 9.45648968e-01 -8.88441920e-01 -9.09715831e-01
-4.42470580e-01 -8.63159478e-01 1.61901534e+00 1.65652409e-01
7.29771733e-01 -7.88903892e-01 -6.48024857e-01 9.70915705e-02
2.63580561e-01 -8.75230372e-01 -5.70791900e-01 4.27597046e-01
-5.75489819e-01 -1.01875103e+00 -3.44339490e-01 -2.85986930e-01
7.72277594e-01 -9.73044932e-02 6.31023407e-01 -7.99877122e-02
-2.35925063e-01 -1.23395599e-01 -7.45465234e-02 -7.83869863e-01
-8.51984739e-01 -2.14717351e-02 3.47662568e-01 5.00920042e-02
3.11387420e-01 -7.65573740e-01 -3.44131291e-01 1.61385983e-01
-1.14590454e+00 -5.58754504e-01 2.00940147e-01 9.92457986e-01
2.12705731e-01 4.21680547e-02 2.36974105e-01 -1.28012347e+00
7.19007790e-01 -5.71553588e-01 -7.13208973e-01 -2.35653818e-02
-3.95686358e-01 1.27298117e-01 1.25287604e+00 -1.17945123e+00
-2.99097568e-01 -5.90108987e-03 -2.84751475e-01 -9.61783409e-01
-1.40554592e-01 -2.45058432e-01 -2.12167919e-01 -7.00846732e-01
1.31825864e+00 9.53016132e-02 1.18464805e-01 -5.36632299e-01
4.05550063e-01 4.27652001e-01 7.09177256e-01 -5.78624904e-01
1.47768319e+00 5.96584916e-01 2.85510346e-02 -4.82417226e-01
-6.77115679e-01 3.38175625e-01 -1.45907387e-01 2.67972559e-01
3.93888414e-01 -4.96445656e-01 -7.51343608e-01 8.37243319e-01
-1.14565897e+00 -5.15754223e-01 -4.55971658e-01 7.04491213e-02
-6.18862025e-02 1.33782655e-01 -6.13192499e-01 -6.58242345e-01
-6.05612993e-01 -1.10488689e+00 6.19479358e-01 -2.81933337e-01
-2.49280646e-01 -9.95428264e-01 1.07166559e-01 5.09692207e-02
5.22108316e-01 9.46124911e-01 7.31738269e-01 -1.40063691e+00
-2.00806946e-01 -8.57421160e-01 3.27892840e-01 4.41572368e-01
1.68272108e-01 -1.61289610e-02 -1.21051598e+00 -5.97447991e-01
3.85772616e-01 -5.33257902e-01 8.19142759e-01 -2.15529174e-01
1.13616204e+00 -1.06933236e+00 -4.64767218e-01 1.03729117e+00
1.11733174e+00 2.05899850e-01 4.70737934e-01 4.68534529e-01
8.67225051e-01 3.51849720e-02 -2.42529973e-01 2.69222617e-01
-2.23300755e-01 4.54611450e-01 8.42504561e-01 -1.58960640e-01
1.79679990e-01 -4.79224652e-01 5.95781267e-01 -1.61969543e-01
3.95466536e-01 -1.57045081e-01 -7.43410230e-01 1.44592568e-01
-1.56369340e+00 -1.12261117e+00 3.22952479e-01 2.25172901e+00
1.29100895e+00 3.48942548e-01 2.61634827e-01 4.13838685e-01
4.96229529e-01 2.84179658e-01 -8.19806159e-01 -3.65157813e-01
1.07075833e-01 4.36480641e-01 9.73967433e-01 7.74346590e-01
-1.27269435e+00 1.12757158e+00 7.01086378e+00 9.55026388e-01
-1.27365351e+00 3.31614614e-01 4.70745116e-01 -5.37148893e-01
-3.64102364e-01 -2.12307483e-01 -8.82657766e-01 6.08853102e-01
1.04269564e+00 -4.27210405e-02 8.39092255e-01 7.76657820e-01
-3.02095413e-01 6.63575470e-01 -1.11585045e+00 4.31787312e-01
1.21397972e-02 -1.37662816e+00 1.80806845e-01 1.34497941e-01
5.16309559e-01 1.49996504e-01 4.35005069e-01 4.76882458e-02
1.24150503e+00 -1.15125346e+00 6.75466001e-01 1.18521869e-01
9.16039884e-01 -1.06352222e+00 3.82943124e-01 3.44396561e-01
-3.88627261e-01 -1.72467723e-01 -2.78166324e-01 1.30907491e-01
-3.91609162e-01 3.81955981e-01 -8.31254542e-01 2.11858861e-02
4.53306019e-01 1.72255293e-01 -2.65308887e-01 4.55692261e-01
-7.19242394e-01 9.78472114e-01 -6.27421975e-01 -2.89236065e-02
3.34798306e-01 3.27771425e-01 9.24184263e-01 9.84977782e-01
-1.72050327e-01 -2.54482627e-01 -7.01733492e-03 9.90861237e-01
-4.49470997e-01 -5.73714316e-01 -1.03749728e+00 -8.47732052e-02
9.13266778e-01 1.15034163e+00 -2.10456192e-01 -2.94254094e-01
3.74424815e-01 9.53716636e-01 3.60595703e-01 4.14240301e-01
-9.20971096e-01 -7.25134969e-01 9.35146093e-01 -2.70464629e-01
1.88150868e-01 2.39780232e-01 -1.04811825e-01 -1.06300950e+00
-6.88593388e-02 -1.47315812e+00 3.08883667e-01 -7.07230717e-02
-1.44795251e+00 5.83838999e-01 -2.06337065e-01 -9.49793637e-01
-2.24635988e-01 -4.68739778e-01 -7.43712902e-01 8.79593730e-01
-1.03146040e+00 -1.04968870e+00 3.77056152e-01 9.61388469e-01
-3.22632492e-01 -4.74698752e-01 1.19386661e+00 -8.62831101e-02
-7.84821510e-01 1.29310167e+00 1.04013667e-01 4.91934299e-01
5.89469135e-01 -1.14480948e+00 8.89208019e-01 1.22427893e+00
2.78987959e-02 9.87495363e-01 1.16948724e+00 -6.43718839e-01
-1.15350032e+00 -1.34970880e+00 4.39577639e-01 -7.77969658e-01
9.94292140e-01 -8.37306321e-01 -8.15173626e-01 8.14220011e-01
-3.02375834e-02 4.52516973e-01 8.27032745e-01 -3.60745013e-01
-1.23190808e+00 -1.23751521e-01 -1.82150280e+00 9.91526425e-01
8.47647786e-01 -5.87987483e-01 -4.42347527e-01 5.50774872e-01
1.08850741e+00 -6.62441373e-01 -4.60367799e-01 9.78218392e-02
5.45480549e-01 -4.23994720e-01 1.15745604e+00 -1.09976339e+00
9.56987664e-02 -3.03682595e-01 -6.91090971e-02 -1.43225837e+00
-1.06660582e-01 -1.25296640e+00 -5.97515225e-01 7.68078327e-01
2.49120027e-01 -1.18985581e+00 9.23686206e-01 4.94434327e-01
3.06611001e-01 -7.16582000e-01 -1.15252340e+00 -1.08191049e+00
4.08351541e-01 -1.26919359e-01 8.44509602e-01 1.11850810e+00
-4.71507721e-02 -1.84420440e-02 -6.96833432e-01 4.46274340e-01
1.05028021e+00 -4.87980425e-01 9.36466217e-01 -9.57864046e-01
-5.42575061e-01 -3.20536107e-01 -3.38802010e-01 -3.65286142e-01
1.68393239e-01 -8.76671493e-01 -1.42697200e-01 -3.91522467e-01
-2.95830190e-01 -1.15088135e-01 -3.40934902e-01 1.15028095e+00
-2.27875531e-01 5.86344242e-01 1.05845921e-01 1.37292981e-01
-8.55751410e-02 2.91080832e-01 7.89213300e-01 -2.56247669e-01
-1.08863674e-01 4.66530584e-02 -9.27663028e-01 7.18084753e-01
1.00789595e+00 -1.10274220e+00 -3.95962179e-01 -2.40204051e-01
-7.54800960e-02 -3.73336643e-01 7.21568942e-01 -1.00037289e+00
1.51684120e-01 5.06181195e-02 4.15183604e-01 8.14332440e-02
3.58685732e-01 -7.65425920e-01 1.33172557e-01 1.10991871e+00
-5.52762389e-01 -7.39639327e-02 2.84067661e-01 6.09140515e-01
5.15722692e-01 -1.98950991e-01 1.09384775e+00 -4.04022634e-02
1.69194907e-01 4.27759767e-01 -1.93309471e-01 1.58683404e-01
1.06717372e+00 1.99503288e-01 -8.55128169e-01 -1.35290235e-01
-4.58274543e-01 5.55735920e-03 6.73393130e-01 1.07473381e-01
5.27681053e-01 -1.21030843e+00 -4.61528629e-01 5.87795496e-01
-1.81337669e-01 -2.23712146e-01 -2.88078427e-01 3.16753805e-01
-2.84306496e-01 -6.40018471e-03 -2.13235542e-01 -4.09064852e-02
-1.30575967e+00 8.17274451e-01 6.45624101e-01 -3.92326534e-01
-6.34587705e-01 1.02810144e+00 -1.51370332e-01 -4.67753828e-01
4.89801675e-01 1.85070634e-01 1.35378867e-01 -3.84989470e-01
8.37469876e-01 2.38541305e-01 -2.16155112e-01 -8.03008303e-02
-2.97859043e-01 2.08588801e-02 -3.94196451e-01 -2.77240694e-01
1.19293571e+00 7.38139510e-01 -2.67918059e-03 8.79106075e-02
1.29872930e+00 2.34052315e-01 -1.39092457e+00 -2.32966468e-01
-4.11223829e-01 -4.27885830e-01 -1.57606080e-01 -9.93677258e-01
-1.06180215e+00 7.39366531e-01 4.75544333e-01 2.42928863e-01
8.10558975e-01 -4.04615760e-01 8.68097305e-01 7.40279436e-01
3.22821409e-01 -5.31623602e-01 6.65111095e-02 3.76974344e-01
6.30258083e-01 -7.09369183e-01 -1.57995194e-01 -4.01871838e-03
-3.37049931e-01 7.19099224e-01 3.78718972e-01 -6.20037198e-01
4.46784794e-01 7.28408754e-01 1.18133619e-01 2.99541261e-02
-7.44496286e-01 5.26460290e-01 -5.35260364e-02 9.70346451e-01
-5.52590072e-01 -1.29930317e-01 2.06303895e-01 3.99152637e-01
-5.27120829e-01 -2.13388517e-01 5.30955195e-01 9.35218930e-01
-2.27086276e-01 -1.27048004e+00 -5.82488477e-01 2.17113823e-01
-9.62770343e-01 -1.69191480e-01 -6.21229053e-01 6.89421237e-01
1.64238457e-02 8.11802506e-01 -3.84044290e-01 -9.09590662e-01
1.45828366e-01 2.12601677e-01 1.67770967e-01 -5.40816307e-01
-9.23793912e-01 -3.71607661e-01 -1.08005449e-01 -7.01939642e-01
3.00875783e-01 -3.21935147e-01 -8.00913930e-01 -6.50747716e-01
-2.11114064e-01 2.22486317e-01 4.90857422e-01 9.17333245e-01
3.05229515e-01 4.09054726e-01 9.39814985e-01 -7.35201299e-01
-1.29959774e+00 -6.60478473e-01 -2.88550705e-01 2.68701494e-01
8.68991315e-01 -4.24253047e-01 -8.69361222e-01 -1.94718450e-01] | [5.837579250335693, 7.658257961273193] |
8410cc49-0384-4394-94fd-232c6958e5a0 | 190909946 | 1909.09946 | null | https://arxiv.org/abs/1909.09946v1 | https://arxiv.org/pdf/1909.09946v1.pdf | Semi-supervised estimation of event temporal length for cell event detection | Cell event detection in cell videos is essential for monitoring of cellular behavior over extended time periods. Deep learning methods have shown great success in the detection of cell events for their ability to capture more discriminative features of cellular processes compared to traditional methods. In particular, convolutional long short-term memory (LSTM) models, which exploits the changes in cell events observable in video sequences, is the state-of-the-art for mitosis detection in cell videos. However, their limitations are the determination of the input sequence length, which is often performed empirically, and the need for a large annotated training dataset which is expensive to prepare. We propose a novel semi-supervised method of optimal length detection for mitosis detection with two key contributions: (i) an unsupervised step for learning the spatial and temporal locations of cells in their normal stage and approximating the distribution of temporal lengths of cell events and, (ii) a step of inferring, from that distribution, an optimal input sequence length and a minimal number of annotated frames for training a LSTM model for each particular video. We evaluated our method in detecting mitosis in densely packed stem cells in a phase-contrast microscopy videos. Our experimental data prove that increasing the input sequence length of LSTM can lead to a decrease in performance. Our results also show that by approximating the optimal input sequence length of the tested video, a model trained with only 18 annotated frames achieved F1-scores of 0.880-0.907, which are 10% higher than those of other published methods with a full set of 110 training annotated frames. | ['Jinman Kim', 'Michael Fulham', 'Ashnil Kumar', 'Ha Tran Hong Phan', 'David Feng'] | 2019-09-22 | null | null | null | null | ['mitosis-detection'] | ['medical'] | [ 3.56413811e-01 -2.27303594e-01 -6.65257201e-02 1.53726218e-02
-8.17361414e-01 -4.61202174e-01 5.63666523e-01 5.01378596e-01
-9.43412781e-01 8.69512439e-01 -1.55540049e-01 -3.44884545e-02
2.73377061e-01 -6.75443113e-01 -9.21846926e-01 -1.16041839e+00
-2.29182601e-01 4.91203964e-01 4.54414397e-01 4.58645493e-01
3.51968139e-01 6.32764876e-01 -1.35287178e+00 3.38823557e-01
3.48259091e-01 9.27571237e-01 4.51116264e-01 1.26252580e+00
3.04570757e-02 9.98525977e-01 -5.82439423e-01 6.36160821e-02
-2.58907735e-01 -5.09448409e-01 -7.97951639e-01 2.35557511e-01
4.16123234e-02 -4.24999237e-01 -5.15802264e-01 7.04289138e-01
5.00753820e-01 1.08413525e-01 7.34570682e-01 -6.87034011e-01
8.57160166e-02 3.12149048e-01 -3.60373557e-01 8.11693668e-01
2.40786791e-01 2.93486446e-01 5.88806510e-01 -5.59049726e-01
1.12220418e+00 6.69741869e-01 5.58225989e-01 4.50610995e-01
-1.21438932e+00 -2.35040620e-01 -2.03840777e-01 2.68110484e-01
-1.29199016e+00 -3.33226174e-01 3.68863374e-01 -6.28283918e-01
1.08712924e+00 -1.01801746e-01 9.06386793e-01 1.00220430e+00
4.57163304e-01 6.18358433e-01 8.15706372e-01 -4.63659853e-01
2.75582969e-01 -4.41997141e-01 -7.93884173e-02 8.08066726e-01
-8.97295848e-02 4.24306504e-02 -4.84073400e-01 1.53787330e-01
1.07545805e+00 1.69880271e-01 -3.24647039e-01 1.52985081e-01
-1.73289812e+00 6.02480054e-01 4.69874516e-02 7.81887591e-01
-2.27213040e-01 2.49803230e-01 6.43799484e-01 3.91922481e-02
3.69059414e-01 3.72232527e-01 -5.69688380e-01 -4.66934919e-01
-1.16359317e+00 -3.44313507e-04 6.94564700e-01 4.50713038e-01
5.77068090e-01 -1.81769803e-01 -2.68991381e-01 4.15646642e-01
-1.75972685e-01 2.93890744e-01 7.43315279e-01 -1.00135911e+00
-7.54947960e-02 6.65921271e-01 5.17328363e-03 -8.07832658e-01
-4.63041365e-01 -1.26846075e-01 -8.45085800e-01 -1.25191435e-01
1.20961773e+00 -1.99524999e-01 -9.49177265e-01 1.54770672e+00
1.44908279e-01 4.42485303e-01 4.10476699e-02 6.63882792e-01
4.67178404e-01 9.02688265e-01 6.71303784e-03 -6.11592829e-01
1.38877606e+00 -5.72946012e-01 -6.51213229e-01 -1.35367289e-02
8.18265378e-01 -7.33138204e-01 7.10770369e-01 2.44827375e-01
-9.26042736e-01 -3.55822176e-01 -8.94752026e-01 -1.75492659e-01
-5.41070640e-01 2.81794727e-01 4.89442527e-01 1.11137472e-01
-8.95226836e-01 7.93044984e-01 -1.15387356e+00 -6.60540938e-01
5.52267730e-01 5.34979045e-01 -4.27079171e-01 2.27178425e-01
-6.43133461e-01 4.30906266e-01 5.73961139e-01 3.32624093e-02
-1.18210316e+00 -6.55247867e-01 -7.24767447e-01 2.01703385e-01
1.85437843e-01 -4.71557289e-01 1.13760412e+00 -8.25709641e-01
-1.36905193e+00 1.26708853e+00 -4.03107524e-01 -6.37446642e-01
5.40398300e-01 2.41220862e-01 1.73534751e-01 4.98579860e-01
-2.13955149e-01 9.54665303e-01 4.41336304e-01 -8.05021524e-01
-9.26362514e-01 -1.75459817e-01 -2.12608799e-01 -1.52377397e-01
-2.69860148e-01 -5.26340231e-02 -6.24684036e-01 -5.42754471e-01
-1.52464971e-01 -8.51901412e-01 -1.91118941e-01 2.03664619e-02
-1.82776690e-01 -5.08825146e-02 7.72740304e-01 -7.40378320e-01
9.67072427e-01 -2.02899241e+00 2.56870031e-01 -1.89319298e-01
1.84719726e-01 2.94234842e-01 1.03987336e-01 2.27394909e-01
2.70116329e-01 4.87568974e-02 5.27376570e-02 -4.07778114e-01
-4.53961968e-01 1.82688758e-01 -1.05116710e-01 6.70303404e-01
-2.17545033e-02 8.43095303e-01 -1.08925080e+00 -7.73160815e-01
3.50595355e-01 6.48393691e-01 -3.52014564e-02 3.06636870e-01
-2.12302163e-01 6.60626590e-01 9.18912366e-02 7.14019179e-01
7.56831393e-02 -5.01000524e-01 1.77016601e-01 -5.81572205e-02
-1.57916725e-01 -1.28855869e-01 -6.50935590e-01 1.44339871e+00
-4.13649566e-02 1.13321126e+00 -2.80042559e-01 -9.77046251e-01
6.83627069e-01 6.18622363e-01 8.02781343e-01 -2.61011422e-01
4.52420443e-01 3.29775155e-01 -2.63362620e-02 -4.89940524e-01
1.56321719e-01 -2.46272534e-02 1.94240347e-01 2.33381584e-01
2.95142114e-01 1.90931648e-01 8.67866457e-01 -1.47756428e-01
1.39664972e+00 -1.26022827e-02 4.02322054e-01 -1.96943492e-01
6.45211577e-01 -2.71253496e-01 6.41771615e-01 5.51247180e-01
-4.66791451e-01 6.19626462e-01 7.89497316e-01 -9.24520254e-01
-1.32962739e+00 -6.55484080e-01 -5.41470014e-02 8.55216086e-01
-1.64578289e-01 -7.86249191e-02 -9.60179865e-01 -4.82413799e-01
-3.85012686e-01 -5.56718733e-04 -8.14411163e-01 1.47667602e-01
-8.10507536e-01 -9.00825799e-01 5.91287494e-01 5.21619916e-01
2.30944395e-01 -1.28236246e+00 -1.07833672e+00 3.20868164e-01
-2.14964554e-01 -1.45418596e+00 -4.72971410e-01 5.28182149e-01
-1.04157841e+00 -1.25643289e+00 -8.68032515e-01 -1.01980495e+00
9.85560656e-01 -9.04255062e-02 8.14557254e-01 1.52424142e-01
-4.37951475e-01 -1.20661959e-01 -1.29529178e-01 -1.72268033e-01
-3.90284836e-01 1.32782727e-01 -2.72518396e-01 -1.71416655e-01
4.52627271e-01 -3.99464816e-01 -6.15893364e-01 2.61852682e-01
-9.62998569e-01 1.16468422e-01 5.27149081e-01 9.71644044e-01
1.03728497e+00 -3.29097249e-02 2.27996394e-01 -8.04680943e-01
-2.15274245e-01 -1.32438913e-01 -7.84682095e-01 -8.66452008e-02
-2.25189984e-01 -9.44083780e-02 9.26763237e-01 -7.14325070e-01
-6.38115823e-01 2.81039447e-01 -6.98909760e-02 -4.09040660e-01
-4.21425134e-01 4.85451937e-01 2.15808704e-01 -5.89615591e-02
2.48048663e-01 5.19146264e-01 1.42954513e-01 1.22175090e-01
-4.77653444e-01 1.57475457e-01 7.06720829e-01 -2.02043071e-01
1.20434307e-01 8.15643430e-01 3.86406749e-01 -1.11562192e+00
-6.05821550e-01 -6.83235228e-01 -8.27908456e-01 -4.68020767e-01
9.41429019e-01 -6.14165366e-01 -9.11278546e-01 9.61016357e-01
-1.11070538e+00 -7.31846035e-01 -2.58430958e-01 4.18322027e-01
-8.06188226e-01 4.52986956e-01 -1.29135680e+00 -5.39296269e-01
-2.92640239e-01 -1.09845245e+00 1.16867399e+00 4.13935959e-01
-2.34446809e-01 -1.28935325e+00 2.22563192e-01 2.74312913e-01
6.34206235e-02 6.07922971e-01 8.65063190e-01 -6.01963699e-01
-6.62063718e-01 -3.95465195e-01 -4.40828130e-02 -1.67156041e-01
1.21899948e-01 4.78651851e-01 -9.77326393e-01 -4.20612931e-01
-2.64117807e-01 -1.93377003e-01 9.65021014e-01 1.08069658e+00
1.10642886e+00 -5.55757806e-02 -6.72084510e-01 7.51192033e-01
1.42176902e+00 4.13877249e-01 8.81972551e-01 4.28278446e-01
4.34070408e-01 4.16299522e-01 5.06932259e-01 3.96929413e-01
-1.78914100e-01 4.00780916e-01 3.07225734e-01 -2.63269901e-01
-1.74639840e-02 -3.38022714e-03 3.43075663e-01 6.22416556e-01
-1.66868031e-01 -4.23263401e-01 -9.24809158e-01 7.86633670e-01
-1.78593540e+00 -1.16982651e+00 2.53114216e-02 2.40902305e+00
7.99210131e-01 3.37222338e-01 2.59819478e-01 4.81904089e-01
7.78508186e-01 -4.44255508e-02 -5.29330194e-01 -8.48155171e-02
-1.84459761e-01 -8.44810233e-02 5.36305070e-01 4.40969884e-01
-1.25615394e+00 8.34035337e-01 6.34287882e+00 7.57425785e-01
-1.42050874e+00 -2.30376586e-01 1.15447628e+00 -2.31020167e-01
3.21842551e-01 -9.94410515e-02 -8.63155901e-01 7.76776969e-01
1.15946031e+00 4.13418561e-02 9.12872404e-02 2.89478660e-01
5.69469929e-01 -3.90465438e-01 -1.22484434e+00 9.62507188e-01
-1.50311187e-01 -1.66184783e+00 -7.61596709e-02 4.96095061e-01
6.80506945e-01 -8.30970258e-02 -2.96486080e-01 4.04329225e-02
-1.48034662e-01 -9.79478061e-01 4.07528043e-01 6.89753056e-01
6.94553316e-01 -7.50983536e-01 1.15652573e+00 3.87725383e-01
-1.09270811e+00 -6.00971542e-02 -3.03297967e-01 -1.86573669e-01
2.27326944e-01 7.98888266e-01 -9.14780021e-01 -8.24967027e-02
4.75916922e-01 7.95233786e-01 -3.91186237e-01 1.12418818e+00
3.02277446e-01 6.76254392e-01 -4.19372648e-01 -1.66600153e-01
2.66095936e-01 -2.18499526e-02 4.66635555e-01 1.49136257e+00
4.76181746e-01 -1.44568399e-01 1.44504264e-01 4.63544428e-01
-2.07222298e-01 -1.74420282e-01 -6.01166427e-01 -3.99498433e-01
2.92709082e-01 1.29168284e+00 -1.53942049e+00 -3.37236404e-01
-2.64390111e-01 7.54064679e-01 3.58363569e-01 2.48933926e-01
-7.56059945e-01 -3.55651259e-01 2.04004392e-01 2.62550771e-01
5.47957957e-01 -1.82926551e-01 -2.36058682e-02 -7.41839826e-01
-3.16690207e-01 -4.53397483e-01 5.40066600e-01 -4.81083572e-01
-7.02190101e-01 3.00973356e-01 -3.59059542e-01 -1.01150799e+00
-3.08349222e-01 -7.10086763e-01 -5.88600874e-01 2.87689477e-01
-1.25777352e+00 -1.11193609e+00 -3.22571814e-01 1.83365747e-01
8.44117105e-01 6.39926270e-02 7.26735353e-01 2.73306578e-01
-6.74956083e-01 1.96353734e-01 1.64797634e-01 3.00147980e-01
5.27145565e-01 -1.42034268e+00 -7.48059805e-03 9.37563539e-01
1.56057492e-01 1.66334540e-01 9.36760604e-01 -3.74344289e-01
-1.16612864e+00 -8.31602275e-01 1.08291519e+00 -2.10652933e-01
5.30937612e-01 -1.06346302e-01 -8.29463601e-01 6.91380858e-01
4.94917557e-02 1.88370645e-01 8.74110699e-01 -4.96319711e-01
3.49766046e-01 -1.46308383e-02 -9.67994988e-01 4.67477798e-01
5.55861413e-01 -3.90650481e-01 -7.79475123e-02 2.65078306e-01
2.86276847e-01 -5.50888062e-01 -1.09649158e+00 2.70142972e-01
7.27773547e-01 -1.02357042e+00 7.21290588e-01 -4.87977453e-02
5.87034464e-01 -3.05132240e-01 1.39736474e-01 -8.56827676e-01
-2.10345969e-01 -4.83659595e-01 -3.50855112e-01 1.02464736e+00
2.78483570e-01 -1.67517185e-01 1.16954458e+00 9.13066976e-03
-8.20871741e-02 -1.22241485e+00 -1.00094283e+00 -5.09080470e-01
-1.45518810e-01 1.44802064e-01 -8.07606280e-02 6.01152301e-01
5.68238758e-02 1.70144588e-01 -1.29852682e-01 -7.50475824e-02
3.85594338e-01 -1.24265160e-02 6.08508885e-01 -1.01298785e+00
-1.69401810e-01 -4.57439035e-01 -7.28223979e-01 -9.61626887e-01
1.83635563e-01 -3.87159258e-01 2.06943408e-01 -1.23172486e+00
5.63234329e-01 2.37341821e-01 -3.53893250e-01 1.60620645e-01
-1.22306809e-01 2.22690642e-01 -2.59918362e-01 2.30119497e-01
-6.90989017e-01 1.21255778e-02 1.19323635e+00 -1.78621426e-01
-1.52923867e-01 -2.31243685e-01 1.52094260e-01 8.73734176e-01
5.68525374e-01 -3.28348130e-01 -5.89296371e-02 -1.01542570e-01
1.27304420e-01 1.81636035e-01 3.13868046e-01 -1.38272583e+00
4.90526170e-01 -6.38651326e-02 7.18900144e-01 -7.30323493e-01
3.89521033e-01 -5.46383083e-01 1.14706218e-01 9.03384089e-01
-4.11244482e-01 -6.31600916e-02 1.48665115e-01 6.10402882e-01
-2.89673686e-01 -2.70446986e-01 9.48429763e-01 -3.34324270e-01
-7.14404345e-01 3.09646100e-01 -9.59735394e-01 -1.14726424e-01
1.15450501e+00 -5.81457853e-01 -3.02310973e-01 -1.95585266e-01
-5.50040901e-01 -2.23687273e-02 7.27081120e-01 -2.06709743e-01
4.53534037e-01 -9.08277333e-01 -3.55503201e-01 8.74081329e-02
-3.27859633e-02 8.37872848e-02 1.51194066e-01 9.98242855e-01
-1.15965962e+00 5.28737903e-01 -3.57307404e-01 -9.21151519e-01
-1.46040928e+00 5.85075140e-01 3.45793039e-01 -7.53631234e-01
-5.27908266e-01 8.07586551e-01 -4.59824763e-02 1.00895733e-01
2.40231395e-01 -6.81963027e-01 -3.40298265e-01 9.45529267e-02
5.52411377e-01 4.71635014e-01 -6.19241260e-02 -6.08137071e-01
-1.08541697e-01 6.06838048e-01 -2.20823213e-01 2.67120749e-01
1.12811518e+00 -1.00989893e-01 -7.92958811e-02 7.75438488e-01
1.11892343e+00 -3.25612098e-01 -1.63030136e+00 1.90999180e-01
1.55967936e-01 -3.11225235e-01 1.14466228e-01 -3.14531565e-01
-1.05111432e+00 7.37002969e-01 5.89295864e-01 8.94223452e-02
1.16769874e+00 -1.37274638e-01 1.23008609e+00 4.25298870e-01
2.89987385e-01 -9.40109015e-01 1.98504761e-01 5.75660050e-01
1.76180061e-02 -1.19741344e+00 -2.12015539e-01 -1.99363396e-01
-4.96017858e-02 1.35250700e+00 4.30095285e-01 3.39938700e-02
3.73075634e-01 5.96613467e-01 1.51443789e-02 -1.71033531e-01
-9.71258938e-01 -5.39491810e-02 -2.86545724e-01 4.92157131e-01
6.81836247e-01 -2.07731843e-01 -2.38100126e-01 -8.22658651e-03
2.63753831e-01 4.45826799e-01 7.06765294e-01 9.62694645e-01
-5.77198207e-01 -7.72728205e-01 -1.68675706e-01 6.05742812e-01
-9.38770473e-01 2.90484160e-01 -1.06833749e-01 6.04681611e-01
8.02219883e-02 6.00003123e-01 3.98779690e-01 6.66984171e-03
-2.46509612e-01 6.59917891e-02 6.03388727e-01 -3.12069595e-01
-2.33767316e-01 2.89153725e-01 -1.92915559e-01 -3.24322224e-01
-5.46077907e-01 -6.68093026e-01 -1.51922321e+00 -4.12240833e-01
-2.31347516e-01 -5.26726060e-02 1.85996905e-01 1.16475761e+00
2.06503317e-01 6.18046761e-01 1.97729781e-01 -1.22593737e+00
1.13453671e-01 -9.41120863e-01 -6.85720384e-01 5.31229973e-01
5.09104967e-01 -4.48346525e-01 -5.19431531e-01 8.49559307e-01] | [14.54472541809082, -3.209603786468506] |
f76b8532-57af-4946-97cc-aac7f6b5a9d6 | look-before-you-leap-improving-text-based | 2209.06209 | null | https://arxiv.org/abs/2209.06209v1 | https://arxiv.org/pdf/2209.06209v1.pdf | Look Before You Leap: Improving Text-based Person Retrieval by Learning A Consistent Cross-modal Common Manifold | The core problem of text-based person retrieval is how to bridge the heterogeneous gap between multi-modal data. Many previous approaches contrive to learning a latent common manifold mapping paradigm following a \textbf{cross-modal distribution consensus prediction (CDCP)} manner. When mapping features from distribution of one certain modality into the common manifold, feature distribution of the opposite modality is completely invisible. That is to say, how to achieve a cross-modal distribution consensus so as to embed and align the multi-modal features in a constructed cross-modal common manifold all depends on the experience of the model itself, instead of the actual situation. With such methods, it is inevitable that the multi-modal data can not be well aligned in the common manifold, which finally leads to a sub-optimal retrieval performance. To overcome this \textbf{CDCP dilemma}, we propose a novel algorithm termed LBUL to learn a Consistent Cross-modal Common Manifold (C$^{3}$M) for text-based person retrieval. The core idea of our method, just as a Chinese saying goes, is to `\textit{san si er hou xing}', namely, to \textbf{Look Before yoU Leap (LBUL)}. The common manifold mapping mechanism of LBUL contains a looking step and a leaping step. Compared to CDCP-based methods, LBUL considers distribution characteristics of both the visual and textual modalities before embedding data from one certain modality into C$^{3}$M to achieve a more solid cross-modal distribution consensus, and hence achieve a superior retrieval accuracy. We evaluate our proposed method on two text-based person retrieval datasets CUHK-PEDES and RSTPReid. Experimental results demonstrate that the proposed LBUL outperforms previous methods and achieves the state-of-the-art performance. | ['Yifeng Li', 'Tian Wang', 'Chao Liu', 'Xili Wan', 'Jingyi Xue', 'Aichun Zhu', 'Zijie Wang'] | 2022-09-13 | null | null | null | null | ['person-retrieval', 'nlp-based-person-retrival'] | ['computer-vision', 'computer-vision'] | [-3.21878076e-01 -5.12938142e-01 -1.25431612e-01 -1.87919408e-01
-9.07353342e-01 -4.58161384e-01 7.72646308e-01 -8.04911479e-02
-3.06263775e-01 4.35823143e-01 3.35582435e-01 2.56870002e-01
-5.21256447e-01 -6.61157906e-01 -3.86937797e-01 -9.01656032e-01
2.91233420e-01 6.04233980e-01 -1.09946594e-01 -3.68533105e-01
-2.94184033e-02 -4.61493060e-02 -1.52602088e+00 3.96386415e-01
9.05899286e-01 6.84635758e-01 4.08591218e-02 2.07173631e-01
-2.11999163e-01 3.88756394e-01 -3.53874296e-01 -6.34296000e-01
2.62248248e-01 -4.91332531e-01 -7.39805281e-01 1.36785060e-01
5.55682838e-01 -2.50626683e-01 -5.39699733e-01 1.17983019e+00
6.47060990e-01 5.10560200e-02 1.15326393e+00 -1.44722688e+00
-1.04044569e+00 2.72928447e-01 -8.39080334e-01 -5.62038971e-03
5.61880171e-01 -1.38124749e-01 1.03830218e+00 -1.01290178e+00
5.43476522e-01 1.42781305e+00 4.96256649e-01 5.13522267e-01
-1.12124729e+00 -6.50985599e-01 1.54322475e-01 2.38696039e-01
-1.87441063e+00 -1.44568741e-01 8.90852571e-01 -6.08825386e-01
1.40096903e-01 4.42660391e-01 5.35749257e-01 8.43405485e-01
1.53397337e-01 1.03513348e+00 1.19852507e+00 -4.48411405e-01
-1.84082523e-01 4.75006193e-01 2.52370626e-01 6.95982099e-01
-4.55888249e-02 -1.28070787e-01 -7.76859581e-01 -2.64329463e-01
5.79139650e-01 3.61399561e-01 -4.74160224e-01 -4.35229957e-01
-1.39159405e+00 6.12944007e-01 4.92594063e-01 4.44656968e-01
-1.98819548e-01 -1.52550042e-01 2.16889724e-01 5.03842950e-01
2.73518980e-01 -2.29585692e-01 1.54763563e-02 2.09925324e-01
-8.66278470e-01 3.52265388e-01 4.39059645e-01 1.07435489e+00
9.41402435e-01 -4.85054165e-01 -1.88170657e-01 8.46971214e-01
7.35663772e-01 8.62361014e-01 4.09091145e-01 -4.69151765e-01
6.50431752e-01 8.57056260e-01 2.87005454e-01 -1.34196615e+00
-1.09130494e-01 -2.68100291e-01 -1.06597424e+00 -1.96677983e-01
4.94074464e-01 3.95854749e-02 -5.90185463e-01 1.79681218e+00
3.37203324e-01 -1.75925016e-01 1.62619445e-02 1.23144400e+00
8.10596466e-01 7.20799208e-01 8.31327736e-02 -1.90445468e-01
1.45707488e+00 -6.12516999e-01 -7.14674592e-01 7.93092474e-02
5.39392829e-01 -9.38889027e-01 1.07493043e+00 1.59726620e-01
-7.27039039e-01 -7.30218589e-01 -9.70036447e-01 1.59669183e-02
-3.85357112e-01 3.55043292e-01 3.08381617e-01 5.67724943e-01
-9.94925797e-01 2.44722307e-01 -3.28574777e-01 -6.10572338e-01
3.13714780e-02 2.05155388e-01 -6.30749643e-01 -3.83165956e-01
-1.39223754e+00 6.98016882e-01 3.82624537e-01 2.15625331e-01
-3.14289957e-01 -4.02446598e-01 -6.45421386e-01 -2.56263644e-01
1.85731485e-01 -9.06605244e-01 4.82258171e-01 -8.66698325e-01
-9.75288689e-01 9.86414313e-01 -2.35060349e-01 2.61423409e-01
6.47406697e-01 -3.40580672e-01 -7.62844265e-01 2.62079597e-01
4.06225979e-01 5.00018120e-01 1.06070137e+00 -1.58404803e+00
-6.03611827e-01 -6.74638927e-01 -2.30615050e-01 5.27132034e-01
-8.78588021e-01 -4.75794300e-02 -8.74643385e-01 -6.46642983e-01
2.96497375e-01 -1.03248882e+00 3.70287478e-01 7.60865510e-02
-4.94057685e-01 -4.82250750e-01 7.30925202e-01 -6.69733584e-01
1.45333111e+00 -2.11886406e+00 5.74135780e-01 4.36378688e-01
2.94678122e-01 -4.57699969e-02 -4.66285758e-02 7.69557714e-01
1.63286626e-02 -1.66807503e-01 9.95291322e-02 -5.15596032e-01
3.25525641e-01 -2.42715739e-02 -4.16329563e-01 6.91782176e-01
-2.84820557e-01 7.90579438e-01 -8.08100581e-01 -8.58882427e-01
2.34513134e-01 5.51397681e-01 -2.92230308e-01 1.71754591e-03
1.87525168e-01 4.76408899e-01 -5.33593357e-01 7.50708938e-01
7.25531399e-01 -2.91780770e-01 8.17084536e-02 -4.90923941e-01
7.50914067e-02 -4.52619880e-01 -1.40746546e+00 1.80644488e+00
7.93763250e-02 3.23717952e-01 -1.52340665e-01 -9.09975648e-01
9.04489934e-01 3.19291025e-01 6.89488530e-01 -6.76981390e-01
9.77697298e-02 2.02243015e-01 -4.92240429e-01 -4.61438686e-01
5.10816932e-01 -2.78358877e-01 -3.95579159e-01 4.81778324e-01
-7.83244371e-02 3.90161276e-01 -8.29892755e-02 3.82782072e-01
4.50101405e-01 1.52875975e-01 -1.68314829e-01 -5.52809954e-01
8.41481626e-01 -1.23063669e-01 4.83060062e-01 5.90089738e-01
-1.97937757e-01 5.55106521e-01 2.01525420e-01 -3.01815331e-01
-8.41252089e-01 -1.36755693e+00 -1.78950027e-01 1.02386868e+00
6.74926996e-01 -6.26326203e-01 -6.19900584e-01 -5.85384369e-01
1.40449852e-02 1.33177966e-01 -5.69975197e-01 -1.45528674e-01
-2.94658571e-01 -6.31179214e-01 4.33844388e-01 1.00433238e-01
8.90060425e-01 -7.17558682e-01 3.11582908e-02 -1.12347275e-01
-6.27666354e-01 -8.00434172e-01 -9.41440642e-01 -6.36748970e-01
-4.17557091e-01 -8.95323575e-01 -1.21534085e+00 -9.33573306e-01
8.21187794e-01 4.08898532e-01 9.70229208e-01 2.71238416e-01
-2.32771128e-01 7.62554228e-01 -3.72713953e-01 -1.26758620e-01
2.18599200e-01 -9.57786515e-02 4.29620534e-01 4.02875245e-01
7.93571174e-01 -3.75496030e-01 -8.74750435e-01 5.35165608e-01
-1.00381887e+00 -7.21630305e-02 5.11923194e-01 9.97876704e-01
5.98761916e-01 3.13995212e-01 4.91365552e-01 -4.08199787e-01
6.33190811e-01 -4.35052454e-01 -1.14881225e-01 6.48383796e-01
-5.15031934e-01 -1.34477496e-01 4.17388320e-01 -4.75456774e-01
-8.58548045e-01 -1.81502566e-01 3.89366686e-01 -5.79629064e-01
-2.03441381e-02 5.60349345e-01 -4.32016015e-01 2.73530662e-01
4.06644523e-01 6.70726359e-01 1.15179025e-01 -4.31244254e-01
4.21419173e-01 7.52284050e-01 4.27269459e-01 -7.09331691e-01
1.10318577e+00 6.42906368e-01 -1.85673103e-01 -8.63905549e-01
-6.78377867e-01 -6.51541710e-01 -8.93615782e-01 -4.33817595e-01
9.57786381e-01 -1.04327512e+00 -7.54658461e-01 6.93317652e-01
-7.93221653e-01 2.45081723e-01 9.15118605e-02 4.51087981e-01
-3.28522474e-01 6.97869897e-01 -4.92587477e-01 -6.61003292e-01
-3.55339497e-01 -8.34358215e-01 1.18164003e+00 3.58642727e-01
-1.62570357e-01 -1.09812284e+00 1.43068999e-01 5.73269606e-01
1.26806587e-01 1.58364233e-02 9.58937407e-01 -3.52917135e-01
-4.74447578e-01 -3.92939061e-01 -3.72478724e-01 1.01167038e-01
4.14209336e-01 -3.40420127e-01 -6.12691283e-01 -7.40121543e-01
-1.81312844e-01 -1.62874475e-01 6.26482844e-01 3.40143405e-02
6.33654714e-01 -2.22834736e-01 -5.83329141e-01 3.18082601e-01
1.43357980e+00 -2.28506610e-01 6.49360776e-01 3.87107164e-01
7.52861321e-01 8.20378721e-01 7.63359904e-01 3.86963457e-01
8.21177602e-01 7.63802409e-01 -1.68484256e-01 -5.56040928e-02
6.95839450e-02 -4.54917431e-01 4.57925767e-01 9.95445549e-01
-9.12133008e-02 -1.71700090e-01 -7.37175643e-01 6.44474328e-01
-1.93734014e+00 -1.13440621e+00 -1.18116038e-02 2.20962620e+00
7.06301093e-01 -3.15554827e-01 1.74691424e-01 4.97652739e-02
8.60848129e-01 -2.16747820e-02 -1.65055960e-01 3.11274350e-01
-2.95571953e-01 -3.62841964e-01 -1.22506119e-01 2.68969685e-01
-1.20041978e+00 6.98227584e-01 4.84722424e+00 1.06403732e+00
-1.00117052e+00 7.70168826e-02 1.45530149e-01 -2.13239659e-02
-5.09172499e-01 -1.23722348e-02 -9.05631840e-01 6.50683403e-01
1.66908115e-01 -1.44404843e-01 2.77640820e-01 3.49770367e-01
-1.14242323e-01 -2.47328579e-02 -1.17873478e+00 1.55762601e+00
4.19875741e-01 -7.97713935e-01 5.55353224e-01 1.99517354e-01
5.29597819e-01 -5.82706869e-01 4.43095505e-01 4.05323058e-01
-9.67577659e-03 -8.88955653e-01 6.14016533e-01 9.19663131e-01
7.62939274e-01 -7.33419538e-01 6.49375260e-01 5.87341845e-01
-1.55844736e+00 1.36468768e-01 -3.63174528e-01 3.89674842e-01
1.33990988e-01 2.44099915e-01 -2.53437847e-01 1.05697274e+00
1.01128268e+00 8.17004919e-01 -5.89815438e-01 7.51708984e-01
1.94439009e-01 -6.45145848e-02 -1.39185593e-01 1.27494603e-01
-2.86205783e-02 -5.02599120e-01 8.09403062e-01 1.01727641e+00
3.68990242e-01 1.63825318e-01 4.09343064e-01 8.01291704e-01
1.76054806e-01 3.55029494e-01 -4.96821046e-01 4.72616963e-02
4.89840716e-01 1.07596779e+00 -3.69599432e-01 -2.50391871e-01
-4.17885303e-01 1.31407571e+00 2.10438743e-01 5.01437545e-01
-6.26894295e-01 -2.45380580e-01 4.20318931e-01 1.57755181e-01
8.19245204e-02 -2.85963551e-03 5.91336973e-02 -1.44138038e+00
3.24265927e-01 -8.26550126e-01 7.74437249e-01 -7.58957565e-01
-1.94669688e+00 5.19506872e-01 2.09511489e-01 -1.79326642e+00
4.07330245e-02 -2.74996668e-01 -2.87093103e-01 1.12876630e+00
-1.46780360e+00 -1.74399519e+00 -3.14321727e-01 1.27432203e+00
1.64004117e-01 -5.09795725e-01 8.58576834e-01 6.88905418e-01
-4.17749196e-01 8.70191514e-01 3.05816531e-01 2.57220089e-01
1.07642627e+00 -1.12582052e+00 -5.55763423e-01 6.81805551e-01
8.96228328e-02 1.11958802e+00 5.47062635e-01 -7.60923803e-01
-1.62044227e+00 -9.09651101e-01 7.44063616e-01 -5.86265087e-01
5.02997935e-01 -5.47069907e-02 -8.75514090e-01 6.47033155e-01
2.22315043e-01 -3.30875963e-01 8.86855423e-01 1.40981913e-01
-3.96897376e-01 -3.46378803e-01 -9.33093548e-01 7.12845802e-01
8.33260238e-01 -8.38069320e-01 -8.04353595e-01 2.77660757e-01
1.05984502e-01 2.20954921e-02 -1.10235763e+00 3.04226041e-01
6.49038732e-01 -8.83046985e-01 1.08559942e+00 -4.26512510e-01
2.19120547e-01 -5.94110668e-01 -4.61604625e-01 -1.11311352e+00
-2.76020139e-01 -4.12991256e-01 1.39645636e-01 1.54761195e+00
-3.69951800e-02 -5.80602467e-01 4.53880519e-01 6.47452593e-01
2.48083189e-01 -4.85727102e-01 -9.49171841e-01 -6.64423525e-01
3.39394480e-01 -3.74635123e-02 5.14153123e-01 1.09485793e+00
2.36841053e-01 1.61912754e-01 -7.44511306e-01 3.68354678e-01
1.06077528e+00 3.56814772e-01 1.00001228e+00 -1.21723211e+00
-1.48356020e-01 -2.15350866e-01 -4.01636720e-01 -1.32153082e+00
1.89978015e-02 -9.73906875e-01 -1.77151248e-01 -1.45871949e+00
7.32447445e-01 -7.24814475e-01 -4.93020266e-01 2.61260033e-01
-3.39169383e-01 3.91303539e-01 3.47379386e-01 8.26183736e-01
-8.36281300e-01 7.42445409e-01 1.28240597e+00 -3.17886591e-01
-1.48550197e-01 -3.35296214e-01 -6.49596810e-01 5.04639566e-01
2.95072317e-01 -6.93758875e-02 -4.10607398e-01 -4.71171200e-01
1.39753670e-01 8.52374509e-02 5.70332766e-01 -7.86070049e-01
5.44183135e-01 3.24407332e-02 6.34981096e-01 -9.19761002e-01
5.81739962e-01 -8.60748768e-01 3.89301062e-01 2.01846510e-01
4.10633422e-02 3.74434739e-02 -1.51838660e-01 7.49079585e-01
-4.26672339e-01 2.32786909e-01 4.87335891e-01 -9.59108844e-02
-7.72732854e-01 4.13102806e-01 -2.91976742e-02 -2.23299429e-01
9.95327652e-01 -2.39223555e-01 -5.46351790e-01 -5.11659741e-01
-8.00204337e-01 5.71681678e-01 5.50915360e-01 6.94920242e-01
7.83213198e-01 -1.83059371e+00 -7.47474194e-01 2.21855968e-01
3.93837720e-01 -3.23612720e-01 6.90328717e-01 8.96436155e-01
-1.53129131e-01 4.67318684e-01 -2.10668415e-01 -9.20186579e-01
-1.28433645e+00 5.38915515e-01 3.81955504e-01 -1.07049048e-01
-5.54875970e-01 6.52474284e-01 2.62472481e-01 -5.53379357e-01
8.36506188e-02 5.67955494e-01 -3.44295263e-01 2.91264385e-01
5.48872828e-01 1.66296050e-01 -3.29345047e-01 -1.20016015e+00
-3.51208627e-01 1.02870607e+00 -2.10002720e-01 -2.24568248e-01
1.01124799e+00 -5.67613780e-01 -5.03758132e-01 6.48578227e-01
1.29242945e+00 1.26883702e-03 -7.51389265e-01 -6.89540923e-01
-2.39031345e-01 -7.17697740e-01 -2.46638030e-01 -4.83812302e-01
-8.39315295e-01 6.05895400e-01 8.81406009e-01 9.25897732e-02
1.13336527e+00 1.79376990e-01 8.33766222e-01 1.72280312e-01
6.36631787e-01 -1.18152523e+00 3.31646323e-01 5.20410538e-02
9.01234090e-01 -1.28349733e+00 1.79775059e-01 -3.23664248e-01
-7.65056014e-01 8.65133345e-01 6.20104492e-01 -2.60282811e-02
9.04675126e-01 -5.68304360e-01 1.45091966e-01 -4.25304383e-01
-2.57705122e-01 -9.96792838e-02 7.88864434e-01 4.36212838e-01
1.09881788e-01 7.69750327e-02 -3.08237195e-01 5.84013581e-01
-2.16209531e-01 -2.28013426e-01 -1.65785640e-01 8.98664653e-01
-2.24850923e-01 -1.27507770e+00 -6.83434010e-01 2.65767723e-01
-2.23095551e-01 1.55409068e-01 -2.95571357e-01 8.20003092e-01
2.38242537e-01 9.82931912e-01 -6.96211606e-02 -5.70661247e-01
1.81689560e-01 1.31338790e-01 4.33042496e-01 -2.64381766e-01
-1.38235837e-01 3.17663908e-01 -3.96921366e-01 -1.81324244e-01
-6.79601610e-01 -8.10374320e-01 -1.04861581e+00 -4.85972345e-01
-3.28341991e-01 2.32950062e-01 1.13146856e-01 1.10443568e+00
1.44348219e-01 3.62465121e-02 6.36238098e-01 -4.86721188e-01
-4.07935888e-01 -8.40388358e-01 -8.71635675e-01 9.18222785e-01
2.18236148e-01 -7.99283147e-01 -2.62975842e-01 1.01971827e-01] | [14.533700942993164, 0.8887861371040344] |
9ab96587-08a8-418c-ac3c-03640d0c8af0 | surrogate-assisted-active-subspace-and-active | 2105.04979 | null | https://arxiv.org/abs/2105.04979v2 | https://arxiv.org/pdf/2105.04979v2.pdf | Surrogate assisted active subspace and active subspace assisted surrogate -- A new paradigm for high dimensional structural reliability analysis | Performing reliability analysis on complex systems is often computationally expensive. In particular, when dealing with systems having high input dimensionality, reliability estimation becomes a daunting task. A popular approach to overcome the problem associated with time-consuming and expensive evaluations is building a surrogate model. However, these computationally efficient models often suffer from the curse of dimensionality. Hence, training a surrogate model for high-dimensional problems is not straightforward. Henceforth, this paper presents a framework for solving high-dimensional reliability analysis problems. The basic premise is to train the surrogate model on a low-dimensional manifold, discovered using the active subspace algorithm. However, learning the low-dimensional manifold using active subspace is non-trivial as it requires information on the gradient of the response variable. To address this issue, we propose using sparse learning algorithms in conjunction with the active subspace algorithm; the resulting algorithm is referred to as the sparse active subspace (SAS) algorithm. We project the high-dimensional inputs onto the identified low-dimensional manifold identified using SAS. A high-fidelity surrogate model is used to map the inputs on the low-dimensional manifolds to the output response. We illustrate the efficacy of the proposed framework by using three benchmark reliability analysis problems from the literature. The results obtained indicate the accuracy and efficiency of the proposed approach compared to already established reliability analysis methods in the literature. | ['Navaneeth N.', 'Souvik Chakraborty'] | 2021-05-11 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [ 1.50296912e-01 -5.33464551e-02 1.56732023e-01 2.69611347e-02
-8.86834860e-01 -3.13419163e-01 2.49955416e-01 -8.36370364e-02
1.54606085e-02 8.19450736e-01 1.81826651e-01 8.65482092e-02
-6.31187737e-01 -3.34782779e-01 -5.70741534e-01 -1.03216493e+00
-9.43961218e-02 2.70972997e-01 -3.84017617e-01 6.24161065e-02
5.15148163e-01 5.48357069e-01 -1.59440804e+00 -3.08966696e-01
1.19049072e+00 9.93240178e-01 2.03794893e-03 3.09563935e-01
2.46276349e-01 2.57461935e-01 -4.44441557e-01 2.15789601e-01
3.17589223e-01 -4.21435446e-01 -5.47295272e-01 2.11160988e-01
-2.06455603e-01 1.52944140e-02 6.16972260e-02 8.00336719e-01
2.90508181e-01 4.15466070e-01 8.99186134e-01 -1.25665748e+00
-2.61210948e-01 -1.57756247e-02 -4.47181165e-01 -2.58154739e-02
4.52128679e-01 -9.35455784e-02 6.53300524e-01 -1.53022909e+00
3.03272575e-01 1.01389921e+00 5.54103255e-01 2.10299149e-01
-1.65594864e+00 -3.55675668e-01 -2.70199358e-01 -3.99711840e-02
-1.76161563e+00 -8.12752068e-01 1.19712019e+00 -7.88512826e-01
5.75236499e-01 2.94852018e-01 4.17797536e-01 9.18142796e-01
2.71825284e-01 5.20509303e-01 1.22851121e+00 -2.60642081e-01
6.95333183e-01 2.01695502e-01 4.00717288e-01 4.32302713e-01
3.55597734e-01 3.01795542e-01 -5.38160026e-01 -5.99743247e-01
4.78110969e-01 -4.23988290e-02 -2.59958953e-01 -1.02541792e+00
-8.47923636e-01 7.51492321e-01 1.41881838e-01 2.49753535e-01
-6.95923448e-01 -2.71038681e-01 1.36853442e-01 2.95728564e-01
5.61733305e-01 6.14494205e-01 4.05196808e-02 -1.38247520e-01
-1.17317736e+00 -7.98643008e-02 8.12467813e-01 5.04994035e-01
6.61991477e-01 3.47655594e-01 2.73936868e-01 7.67160654e-01
6.81812108e-01 3.19854379e-01 7.83758312e-02 -7.18592286e-01
3.44835609e-01 8.34273398e-01 2.90210545e-01 -1.14778996e+00
-4.12915587e-01 -4.91538107e-01 -1.11476541e+00 4.11967307e-01
2.22572371e-01 -1.16718717e-01 -5.50033092e-01 1.29093814e+00
4.61176306e-01 2.38609716e-01 2.96738386e-01 1.04339790e+00
2.15818673e-01 7.21919835e-01 7.67170638e-02 -8.37642372e-01
5.68443060e-01 -4.73700762e-01 -7.58051634e-01 2.55856931e-01
4.37097341e-01 -7.72844195e-01 9.71298516e-01 3.61589670e-01
-9.41287100e-01 -6.04259610e-01 -1.30240083e+00 4.95906115e-01
-3.09170187e-02 4.40976620e-01 2.28924125e-01 5.60885847e-01
-7.85773754e-01 5.84173977e-01 -8.03756833e-01 -2.14696020e-01
1.27453776e-02 6.14341736e-01 -4.46759135e-01 1.13618448e-02
-1.00557816e+00 9.31713223e-01 2.83070266e-01 3.09800535e-01
-1.04013181e+00 -5.16183078e-01 -8.21087599e-01 -6.49952441e-02
4.00265098e-01 -4.86571312e-01 5.75227141e-01 -7.86586404e-01
-1.63288212e+00 1.25295728e-01 4.39816229e-02 -5.25141470e-02
9.28482264e-02 -4.12512422e-01 -3.92938256e-01 -4.75635454e-02
-1.10960901e-01 -2.26187915e-01 1.38344610e+00 -1.42261672e+00
1.80142686e-01 -5.57957292e-01 -2.14528918e-01 3.21782112e-01
-5.83685696e-01 -1.43669382e-01 -5.80035411e-02 -5.58654308e-01
5.36841452e-01 -1.04201555e+00 -1.69854462e-01 -2.91988373e-01
-3.18305492e-01 -9.18705314e-02 1.06251085e+00 -8.01450193e-01
1.44294822e+00 -2.39434695e+00 8.20769131e-01 5.63143790e-01
1.65675208e-01 9.07787308e-02 1.38517916e-01 7.49594510e-01
-3.18949491e-01 -3.86890084e-01 -6.79879248e-01 -2.68334001e-01
-2.59304672e-01 -1.56107605e-01 -2.73197412e-01 7.62609541e-01
3.78445089e-01 3.19001496e-01 -7.19096899e-01 -4.30386811e-01
2.98649728e-01 6.02864981e-01 -2.17548266e-01 5.13108253e-01
4.26442951e-01 7.73571789e-01 -4.78193551e-01 6.12742186e-01
5.56466341e-01 -2.49739870e-01 1.19996199e-03 -4.51699972e-01
-2.44589165e-01 -2.10414767e-01 -1.68342257e+00 1.67137194e+00
-4.89481419e-01 6.07299320e-02 2.53802329e-01 -1.51335609e+00
1.14107692e+00 6.39821649e-01 9.02111053e-01 -2.92071819e-01
-2.50796545e-02 4.66495335e-01 -1.13425829e-01 -5.13620615e-01
3.34030479e-01 -7.40884617e-02 -5.67739829e-02 4.18733507e-01
-3.59935835e-02 -2.29958091e-02 -2.90026353e-03 1.55379072e-01
8.99527967e-01 2.62768447e-01 4.42716926e-01 -5.19923270e-01
9.33956921e-01 -1.07652359e-01 3.02343249e-01 4.56052758e-02
-3.25347148e-02 2.83519030e-01 3.26120526e-01 -4.48783599e-02
-1.09031975e+00 -1.11939883e+00 -3.91739368e-01 3.21705014e-01
-1.44271299e-01 -3.65964204e-01 -6.88108385e-01 -5.07182240e-01
6.44077659e-02 4.09202963e-01 -4.42821741e-01 -4.34886992e-01
-3.83138329e-01 -7.04875171e-01 -1.50359003e-02 2.01627165e-01
1.11745737e-01 -3.96948159e-01 -3.92912149e-01 1.93299532e-01
2.33466938e-01 -5.74624300e-01 -6.17995933e-02 1.24380007e-01
-1.28051150e+00 -9.49757934e-01 -9.04360771e-01 -5.27001798e-01
1.01487052e+00 1.62416264e-01 7.83308446e-01 -1.59456581e-01
-2.63155043e-01 5.38934171e-01 -1.61881998e-01 6.28659129e-02
-2.24649444e-01 -1.90573856e-01 6.52825058e-01 3.38073552e-01
2.87550632e-02 -5.18500328e-01 -4.09785509e-01 5.21657586e-01
-7.67439961e-01 -4.15964052e-02 7.50338733e-01 1.04991710e+00
5.73204100e-01 3.69791925e-01 1.11149240e+00 -5.86289883e-01
8.06054413e-01 -7.89907575e-01 -7.94223547e-01 1.79191440e-01
-8.22774470e-01 3.38239282e-01 7.97419548e-01 -2.28891134e-01
-9.69070852e-01 6.60723269e-01 3.26932430e-01 -6.11707747e-01
1.23926312e-01 7.41240621e-01 -1.73056483e-01 -1.30901232e-01
7.41891026e-01 1.90914363e-01 2.65515864e-01 -4.45774853e-01
2.53015161e-01 7.52612174e-01 1.54806659e-01 -6.83618903e-01
1.08547127e+00 1.17058538e-01 3.79411936e-01 -9.57038522e-01
-5.59811950e-01 -7.01345086e-01 -6.89320624e-01 -4.26462591e-01
3.66237938e-01 -7.21228898e-01 -4.51763391e-01 1.74664557e-01
-6.58704102e-01 3.33295196e-01 -3.16487491e-01 6.14458144e-01
-6.93407893e-01 5.87038875e-01 -1.61365047e-01 -1.34051919e+00
-4.57714736e-01 -1.26795995e+00 9.70347583e-01 -1.17058143e-01
-4.09126639e-01 -8.49996328e-01 3.62346888e-01 2.70104408e-01
4.28565055e-01 2.77727872e-01 8.41848314e-01 -2.51297504e-01
-3.63007843e-01 -5.43817222e-01 2.55946189e-01 3.83271784e-01
2.04865888e-01 -5.42802364e-02 -7.90760815e-01 -5.41410446e-01
5.27346373e-01 -3.39952379e-01 2.73195207e-01 2.50086188e-01
8.32167685e-01 -5.32950908e-02 -1.32960081e-01 1.13940798e-01
1.37551796e+00 8.77842456e-02 3.45113009e-01 -1.20964520e-01
5.87929070e-01 9.02118027e-01 9.62016940e-01 5.05116999e-01
-2.21565127e-01 8.27077866e-01 1.10176921e-01 1.45844733e-02
3.71480614e-01 3.92320268e-02 4.04158741e-01 1.32102191e+00
-4.67864685e-02 2.65818149e-01 -1.09680367e+00 2.58297741e-01
-1.83297396e+00 -6.34673774e-01 -4.43343520e-02 2.59483671e+00
5.60891271e-01 3.72712016e-02 1.57659486e-01 8.37459326e-01
6.46902144e-01 -3.10544550e-01 -6.15507543e-01 -1.04802161e-01
4.86312136e-02 -5.79174748e-03 -3.10184266e-02 4.25533473e-01
-1.01907337e+00 2.17472747e-01 6.31278896e+00 5.26645303e-01
-9.79084909e-01 -5.10943122e-02 4.03108150e-01 5.59825599e-02
-4.34173942e-02 2.47537836e-01 -4.43545491e-01 4.34229583e-01
1.21704578e+00 -2.85346121e-01 4.36727792e-01 8.75899196e-01
5.18741071e-01 -2.88807213e-01 -1.17768764e+00 1.16771102e+00
1.59352496e-01 -7.56402731e-01 -2.91507363e-01 1.20641984e-01
8.81440461e-01 -7.13552952e-01 7.60707185e-02 1.42403513e-01
-2.83222407e-01 -9.66092110e-01 2.22085848e-01 8.95929575e-01
7.80671656e-01 -7.01520741e-01 5.73407352e-01 4.97798145e-01
-1.08331144e+00 -3.14799666e-01 -2.74125069e-01 -2.75020096e-02
1.82779282e-01 8.19288850e-01 -6.02263570e-01 6.50627911e-01
3.43688160e-01 8.50176990e-01 -3.22141498e-01 1.15240479e+00
3.25943559e-01 6.08751655e-01 -2.26387113e-01 3.09972912e-01
-1.88617915e-01 -7.03608394e-01 8.04794014e-01 4.05870289e-01
6.66648507e-01 1.11097895e-01 2.65941713e-02 9.21651125e-01
2.35270157e-01 2.01370373e-01 -9.53979731e-01 -3.82060111e-01
3.68204325e-01 1.16006756e+00 -5.17317176e-01 -1.05300069e-01
-3.79002571e-01 7.16450036e-01 1.89265490e-01 4.61764306e-01
-4.97385859e-01 -7.25431070e-02 1.88214079e-01 -2.72372123e-02
-2.57225066e-01 -5.58938503e-01 -4.05436337e-01 -1.00525188e+00
1.32882774e-01 -8.95226359e-01 2.54137725e-01 -5.38031936e-01
-1.15936887e+00 3.23259652e-01 1.84610069e-01 -1.76868105e+00
-4.10304785e-01 -2.53272951e-01 -2.08731636e-01 1.11088431e+00
-8.33272517e-01 -6.53146744e-01 -3.04785758e-01 7.06162691e-01
4.99139279e-01 -3.80205631e-01 9.50323939e-01 4.09227520e-01
-9.24801648e-01 2.82168031e-01 6.54719710e-01 -4.43221211e-01
4.20994729e-01 -9.88681495e-01 -2.19216615e-01 8.52927029e-01
-1.65933028e-01 8.80498946e-01 7.87086546e-01 -7.22443402e-01
-2.08570290e+00 -5.57188630e-01 4.93240178e-01 -3.57295334e-01
7.59652555e-01 -2.64959723e-01 -9.00298178e-01 1.09037586e-01
-1.79615170e-01 -1.26385227e-01 8.95815611e-01 1.85137600e-01
9.12028998e-02 -2.04176411e-01 -1.07134080e+00 2.80053020e-01
6.71005428e-01 -8.32983732e-01 -5.92921853e-01 2.53271699e-01
1.22121602e-01 3.07938214e-02 -1.26528621e+00 4.78511542e-01
3.04056168e-01 -3.90476495e-01 9.81395602e-01 -5.02802730e-01
2.03556359e-01 -5.81394494e-01 -1.97459966e-01 -1.30944896e+00
-3.43793452e-01 -5.02724826e-01 -7.69540310e-01 1.11494100e+00
1.85616851e-01 -4.18259054e-01 7.33611345e-01 9.49781597e-01
-2.61004359e-01 -7.89000332e-01 -1.08458340e+00 -7.40501761e-01
-2.96661496e-01 -7.24726096e-02 8.96851644e-02 7.98243999e-01
2.58745551e-01 6.06335998e-01 -5.49313843e-01 2.38795429e-01
1.00507665e+00 1.02369711e-01 6.31073415e-01 -1.49641156e+00
-1.34404063e-01 5.91789708e-02 -5.28324485e-01 -5.16105890e-01
2.16928303e-01 -5.50932467e-01 1.80171505e-02 -1.18486905e+00
1.22522816e-01 -5.79971850e-01 -4.54825073e-01 -1.11436777e-01
-1.32850558e-01 -1.24560604e-02 -6.68819249e-02 6.88995838e-01
-3.60399514e-01 8.99693072e-01 9.61301982e-01 -4.94653471e-02
-3.60830039e-01 4.90371250e-02 -3.22536528e-01 3.22887778e-01
6.51864648e-01 -4.06286597e-01 -8.92618418e-01 1.36691347e-01
1.68733180e-01 5.66269040e-01 1.71843186e-01 -1.32554030e+00
2.56204545e-01 -5.03944606e-02 2.89400727e-01 -5.78197181e-01
6.64474845e-01 -1.28307450e+00 7.82581031e-01 3.65269303e-01
-2.09858477e-01 -6.73642978e-02 -6.36747107e-02 7.75099099e-01
-4.40256447e-01 -3.03291529e-01 7.30347753e-01 3.26945931e-01
-4.16816413e-01 1.04955316e-03 -2.23656148e-01 -3.46801251e-01
1.23931170e+00 -4.12822396e-01 3.56261656e-02 -1.17929697e-01
-9.40854013e-01 -2.55185347e-02 5.05981565e-01 3.29290420e-01
8.98120999e-01 -1.47689545e+00 -3.96011591e-01 3.96001250e-01
1.17864095e-01 -2.23375604e-01 4.40183848e-01 1.10936451e+00
-4.98122238e-02 4.93899673e-01 -2.62134284e-01 -8.11246574e-01
-1.00984192e+00 8.78723621e-01 1.19030215e-01 -6.89649209e-03
-4.48758632e-01 2.65686780e-01 -4.54304181e-02 -1.07921913e-01
4.89181690e-02 3.90243620e-01 -4.85865086e-01 4.73188795e-02
9.93006453e-02 7.62250006e-01 1.75523147e-01 -1.01363337e+00
-2.53631473e-01 7.04772592e-01 3.13624620e-01 -9.05151516e-02
1.36726236e+00 -1.77808553e-01 -2.62258470e-01 7.72179604e-01
1.34253180e+00 -2.25920603e-01 -1.14306068e+00 -9.53575596e-02
3.46518546e-01 -4.80306864e-01 2.69461185e-01 -1.74670920e-01
-8.58209670e-01 6.06329620e-01 9.09961820e-01 2.02800721e-01
1.20489275e+00 -3.96476924e-01 2.12085322e-01 4.13748860e-01
4.74657118e-01 -1.37167025e+00 2.17810258e-01 1.58318788e-01
1.19379675e+00 -1.17153108e+00 2.35943556e-01 -5.02663255e-01
-4.50173974e-01 1.03294730e+00 4.76474911e-01 -1.73751682e-01
8.80031586e-01 -7.35473335e-02 -1.80826515e-01 -2.87250906e-01
-4.33018893e-01 3.13665450e-01 5.20884037e-01 2.57023513e-01
4.53049600e-01 2.27357559e-02 -6.42963409e-01 3.25825036e-01
2.62549430e-01 -5.82132451e-02 3.31458628e-01 9.79631305e-01
-1.17475614e-01 -1.07263839e+00 -7.79471159e-01 5.18346012e-01
-1.00618012e-01 3.39681238e-01 -2.19350055e-01 3.05687249e-01
-4.81032759e-01 1.16669619e+00 -4.03404444e-01 -4.06016320e-01
4.12692845e-01 2.30778947e-01 1.00556314e-01 -5.81223965e-01
-1.55871600e-01 4.68509346e-02 2.24228855e-02 -5.96388400e-01
-5.48016191e-01 -7.88958490e-01 -9.72159147e-01 -5.03685372e-03
-4.76729840e-01 5.59157431e-01 9.15043771e-01 6.63132250e-01
4.67343688e-01 1.70497715e-01 1.10595226e+00 -8.12477350e-01
-9.87852633e-01 -8.89592469e-01 -6.95195794e-01 6.13347530e-01
2.69798368e-01 -1.26361215e+00 -6.49362981e-01 -1.18045874e-01] | [7.782650470733643, 4.226172924041748] |
3988f1fa-5f44-466a-9658-f6aff07c58ac | group-aware-label-transfer-for-domain | 2103.12366 | null | https://arxiv.org/abs/2103.12366v1 | https://arxiv.org/pdf/2103.12366v1.pdf | Group-aware Label Transfer for Domain Adaptive Person Re-identification | Unsupervised Domain Adaptive (UDA) person re-identification (ReID) aims at adapting the model trained on a labeled source-domain dataset to a target-domain dataset without any further annotations. Most successful UDA-ReID approaches combine clustering-based pseudo-label prediction with representation learning and perform the two steps in an alternating fashion. However, offline interaction between these two steps may allow noisy pseudo labels to substantially hinder the capability of the model. In this paper, we propose a Group-aware Label Transfer (GLT) algorithm, which enables the online interaction and mutual promotion of pseudo-label prediction and representation learning. Specifically, a label transfer algorithm simultaneously uses pseudo labels to train the data while refining the pseudo labels as an online clustering algorithm. It treats the online label refinery problem as an optimal transport problem, which explores the minimum cost for assigning M samples to N pseudo labels. More importantly, we introduce a group-aware strategy to assign implicit attribute group IDs to samples. The combination of the online label refining algorithm and the group-aware strategy can better correct the noisy pseudo label in an online fashion and narrow down the search space of the target identity. The effectiveness of the proposed GLT is demonstrated by the experimental results (Rank-1 accuracy) for Market1501$\to$DukeMTMC (82.0\%) and DukeMTMC$\to$Market1501 (92.2\%), remarkably closing the gap between unsupervised and supervised performance on person re-identification. | ['Zheng-Jun Zha', 'Jiebo Luo', 'Tao Mei', 'Lingxiao He', 'Wu Liu', 'Kecheng Zheng'] | 2021-03-23 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Zheng_Group-aware_Label_Transfer_for_Domain_Adaptive_Person_Re-identification_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Zheng_Group-aware_Label_Transfer_for_Domain_Adaptive_Person_Re-identification_CVPR_2021_paper.pdf | cvpr-2021-1 | ['online-clustering'] | ['computer-vision'] | [ 3.66959006e-01 -8.65881704e-03 -1.97115511e-01 -7.13106632e-01
-1.02653408e+00 -6.57094419e-01 5.71189702e-01 9.28241685e-02
-5.93942821e-01 8.09676409e-01 -1.83808655e-01 -4.79728309e-03
-3.14846754e-01 -6.55359983e-01 -4.72963631e-01 -7.77331293e-01
1.57242924e-01 1.02346599e+00 -1.83336750e-01 1.33221924e-01
4.96640205e-02 2.93857008e-01 -1.53986120e+00 6.08084202e-02
1.08396935e+00 8.02808881e-01 1.32003635e-01 2.46278241e-01
-2.11486220e-01 3.54393065e-01 -2.64756054e-01 -5.62688708e-01
5.02392411e-01 -4.98447359e-01 -1.05778992e+00 1.56697899e-01
2.33208209e-01 -1.12813897e-01 7.29211047e-02 1.14127314e+00
5.55908501e-01 3.73816073e-01 9.41101193e-01 -1.20142388e+00
-6.14477634e-01 5.79593062e-01 -6.85123622e-01 -2.35671893e-01
2.13213623e-01 -7.51756206e-02 6.73887730e-01 -7.53983438e-01
5.70722401e-01 1.12520456e+00 7.94386625e-01 8.55386436e-01
-1.58888602e+00 -1.05206382e+00 3.30592066e-01 2.61600733e-01
-1.93641698e+00 -4.62379515e-01 7.34977543e-01 -5.88580191e-01
4.39446658e-01 1.26040503e-01 2.14617580e-01 9.08289433e-01
-7.57602632e-01 4.56090927e-01 1.33412552e+00 -4.63563561e-01
3.06294799e-01 3.17068011e-01 4.54894155e-01 4.56536442e-01
4.30876426e-02 9.43161696e-02 -5.72385609e-01 -2.41715610e-01
3.83043349e-01 -1.76911596e-02 6.39301315e-02 -2.37335160e-01
-9.22900558e-01 6.62526906e-01 2.45568678e-01 -4.62035239e-02
-2.70759881e-01 -1.23326063e-01 3.45547348e-01 2.85675913e-01
5.90692639e-01 3.22819620e-01 -4.56371665e-01 5.21733873e-02
-8.60852897e-01 9.11685638e-03 4.72896874e-01 1.13590157e+00
1.18090987e+00 -3.56396973e-01 -1.56175405e-01 1.24947011e+00
3.07434082e-01 4.63830233e-01 3.85778993e-01 -9.17203784e-01
2.46648028e-01 7.49444962e-01 3.69803041e-01 -6.55290484e-01
-4.35916811e-01 -4.99002248e-01 -8.91595244e-01 -8.60435888e-02
7.27191687e-01 -1.59485161e-01 -8.64951789e-01 1.99303472e+00
6.26034260e-01 4.32261109e-01 6.68247342e-02 8.25770080e-01
3.54703665e-01 4.41585302e-01 5.07073224e-01 -5.17062426e-01
1.38798201e+00 -7.57850170e-01 -4.71634120e-01 -7.87304863e-02
9.62334692e-01 -6.32366896e-01 7.47374713e-01 7.62935579e-02
-7.10599065e-01 -9.40187931e-01 -7.96498835e-01 6.84976429e-02
-3.89825433e-01 2.79158294e-01 2.86755174e-01 8.27794552e-01
-9.96549726e-01 5.79053760e-01 -3.32041681e-01 -5.17115414e-01
4.07344431e-01 8.64657640e-01 -4.01076913e-01 -3.78138095e-01
-1.18190920e+00 3.39607298e-01 4.20445830e-01 2.39576958e-03
-6.55582011e-01 -7.01012135e-01 -6.72294199e-01 -2.24199370e-01
4.38872039e-01 -5.18451691e-01 9.73176777e-01 -9.38091159e-01
-1.49558318e+00 1.29544544e+00 -4.65186775e-01 -3.59667063e-01
5.52996814e-01 4.04401012e-02 -5.24846792e-01 -9.53655094e-02
5.07131636e-01 8.38349402e-01 7.38309324e-01 -1.50016510e+00
-1.06828809e+00 -5.42240143e-01 -2.73347795e-01 4.06204373e-01
-3.02407831e-01 -2.17395633e-01 -3.76923919e-01 -5.90084553e-01
1.97695076e-01 -1.08519566e+00 -1.67521406e-02 -4.28130209e-01
-3.00302207e-01 -6.57301426e-01 4.10051435e-01 -5.82900941e-01
1.00028372e+00 -2.13981557e+00 6.31128810e-03 6.49405420e-01
4.65718210e-02 1.21699937e-01 -1.98173329e-01 1.51553392e-01
-2.01621413e-01 4.50238921e-02 -2.32841372e-01 -7.77624130e-01
4.30221483e-02 6.61920533e-02 -3.52214687e-02 5.31660676e-01
-2.32424840e-01 6.95061505e-01 -1.01757181e+00 -5.88437855e-01
-5.36907976e-03 1.44315049e-01 -3.11489463e-01 2.92866588e-01
9.61068645e-02 8.36730957e-01 -3.35465997e-01 6.46577656e-01
9.32045281e-01 -1.89193293e-01 5.38749516e-01 -1.02813371e-01
-1.26467198e-01 -6.16191700e-02 -1.31920183e+00 1.58747458e+00
-1.42985791e-01 -4.38944995e-02 9.82961059e-02 -9.43283200e-01
1.16905451e+00 7.09459558e-02 7.38742173e-01 -7.12415755e-01
6.24355264e-02 2.53129750e-01 -3.67858440e-01 -1.82070732e-01
5.14821827e-01 -1.82020009e-01 -2.80686319e-01 6.97375834e-01
-1.33086830e-01 6.58729374e-01 -2.97340993e-02 1.01661831e-01
6.29168868e-01 2.39330113e-01 1.05437316e-01 -3.63823175e-01
7.75408208e-01 1.21187031e-01 7.81029224e-01 8.85014474e-01
-4.13308412e-01 2.88193166e-01 -1.18762642e-01 -2.46993944e-01
-9.40522969e-01 -9.38911855e-01 -1.37324840e-01 1.50238287e+00
4.32264477e-01 -1.78255722e-01 -8.67708504e-01 -9.78497684e-01
3.06345187e-02 5.64550698e-01 -5.96505105e-01 -2.70297974e-01
-5.52945375e-01 -7.79259622e-01 4.64115232e-01 3.30836236e-01
7.34190881e-01 -6.82984173e-01 2.79023886e-01 1.27404511e-01
-4.09346908e-01 -9.32533264e-01 -8.22870314e-01 1.88404441e-01
-5.21662414e-01 -9.50278819e-01 -6.64668024e-01 -1.11604369e+00
1.02852702e+00 3.04520220e-01 5.14867485e-01 -3.28667387e-02
1.23959206e-01 4.13857609e-01 -5.15166044e-01 1.35294408e-01
-4.06247050e-01 3.26981485e-01 5.75346470e-01 5.73703289e-01
6.98846519e-01 -4.62866545e-01 -5.87149799e-01 7.25743115e-01
-4.07173723e-01 1.23088650e-01 3.60332787e-01 9.37513649e-01
7.92139709e-01 2.01544717e-01 9.64468479e-01 -1.26173627e+00
3.03400278e-01 -5.40410519e-01 -4.39610720e-01 4.05098826e-01
-9.65670407e-01 -3.17608938e-02 7.41154373e-01 -7.10012734e-01
-1.17490458e+00 5.41756868e-01 3.86858894e-03 -2.12778941e-01
-4.18955952e-01 1.18795119e-01 -4.76360410e-01 -4.29529771e-02
6.62606895e-01 4.32020396e-01 -1.38063326e-01 -7.22786129e-01
4.42263037e-01 9.38419640e-01 7.78567791e-01 -7.46327996e-01
1.06495571e+00 4.59120810e-01 -1.66852355e-01 -3.77328277e-01
-6.52258217e-01 -8.57751369e-01 -1.06412375e+00 -2.27519810e-01
7.70280063e-01 -9.85731602e-01 -9.62434590e-01 6.69002056e-01
-6.73649490e-01 -4.82936561e-01 -3.24522734e-01 3.66894960e-01
-3.68655652e-01 6.28569782e-01 -3.64913642e-01 -8.92789245e-01
-2.55845964e-01 -8.81348252e-01 8.75426471e-01 4.62694466e-01
-3.03990513e-01 -7.04724848e-01 -2.65615582e-01 7.89065182e-01
9.00562853e-02 -2.04041302e-01 9.01411951e-01 -1.02739644e+00
-3.76607597e-01 -2.27500111e-01 -5.20323217e-01 1.93882883e-01
3.21008503e-01 -8.29238296e-01 -1.19881010e+00 -6.21479154e-01
-3.41821671e-01 -2.43979096e-01 5.97360075e-01 1.51057199e-01
9.89662290e-01 -9.75838304e-02 -5.24553001e-01 6.18420243e-01
1.13609087e+00 3.93549055e-01 3.15270990e-01 3.44747186e-01
8.98510754e-01 7.78480113e-01 9.24364090e-01 4.89833623e-01
6.44914448e-01 9.00835454e-01 -4.64451239e-02 1.03582971e-01
-2.34399691e-01 -5.45575738e-01 1.99496821e-01 7.48563290e-01
-8.03004429e-02 9.31103826e-02 -8.51567328e-01 4.45199072e-01
-1.84643114e+00 -6.85814798e-01 -6.53482080e-02 2.57749820e+00
1.05960369e+00 -1.78220794e-01 3.04228872e-01 6.92293644e-02
1.18799746e+00 -4.14794356e-01 -6.91076577e-01 9.88290831e-02
1.20715655e-01 -7.05386931e-03 6.42898679e-01 5.18030584e-01
-1.32722175e+00 1.04227149e+00 5.39238071e+00 1.19780874e+00
-5.76077640e-01 2.90713966e-01 7.38256574e-01 2.95920372e-01
8.27041939e-02 9.65646561e-03 -1.03439593e+00 7.06976712e-01
9.77807879e-01 -1.16769791e-01 7.07231164e-01 7.67876446e-01
1.15990914e-01 -3.48267630e-02 -1.30030251e+00 1.32063091e+00
-3.59320901e-02 -8.94656003e-01 -8.59308094e-02 2.10603043e-01
7.75620401e-01 -5.23967147e-01 3.68039869e-02 4.73909169e-01
4.36758786e-01 -6.66161776e-01 6.01998866e-01 5.57164431e-01
1.24055803e+00 -9.04715717e-01 6.87860489e-01 4.52213287e-01
-1.43164361e+00 -2.86409944e-01 -2.77910024e-01 1.22367129e-01
1.08852826e-01 2.22368047e-01 -9.61529076e-01 7.92510569e-01
6.71038449e-01 5.89754283e-01 -5.33256054e-01 8.55644226e-01
-3.12683452e-03 4.83016670e-01 -1.61531121e-01 3.81684691e-01
-2.24357963e-01 -4.21583980e-01 2.92347103e-01 1.01931465e+00
1.29211396e-01 2.67807573e-01 5.34813404e-01 6.97188020e-01
-3.69382292e-01 3.08269262e-01 -2.41212219e-01 2.96768665e-01
8.29197288e-01 1.06695414e+00 -6.82430089e-01 -4.14971322e-01
-1.15979046e-01 1.28544593e+00 3.94861013e-01 4.65220511e-01
-7.53288329e-01 -1.58571482e-01 5.87848186e-01 2.35707104e-01
-1.00675434e-01 9.70726926e-03 -3.05205435e-01 -7.86794960e-01
-2.15116903e-01 -5.70370436e-01 7.50444889e-01 -3.86421591e-01
-1.63286936e+00 2.92541742e-01 9.87861156e-02 -1.27298439e+00
-1.12501897e-01 -6.00428693e-02 -1.13223098e-01 1.00105488e+00
-1.51445234e+00 -1.32814050e+00 -3.36970836e-01 6.75340056e-01
4.47276860e-01 -2.49838427e-01 9.10482287e-01 6.95933521e-01
-8.09318066e-01 1.18102729e+00 3.02348703e-01 2.53962815e-01
1.13048315e+00 -1.09854472e+00 6.88957572e-02 7.57331729e-01
-3.00410867e-01 5.86493194e-01 2.07511976e-01 -8.27997327e-01
-8.27237487e-01 -1.50567389e+00 9.81566787e-01 -2.22291395e-01
1.79034591e-01 -4.51606691e-01 -8.67797434e-01 5.19357860e-01
-3.53862375e-01 -2.33640924e-01 9.05336678e-01 2.37603068e-01
-4.62602407e-01 -4.27467436e-01 -1.40420437e+00 3.41694653e-01
1.39275920e+00 -6.82902098e-01 -1.30475223e-01 3.42467248e-01
5.31139195e-01 -4.35511433e-02 -1.04481995e+00 2.57270902e-01
3.14851850e-01 -4.44641560e-01 1.05385959e+00 -2.78248191e-01
-3.87450129e-01 -6.17485464e-01 -4.35201414e-02 -9.48512852e-01
-6.62666798e-01 -5.74867070e-01 2.92265147e-01 1.90365684e+00
2.13836521e-01 -8.87495041e-01 8.52477431e-01 1.02306187e+00
1.40965655e-02 -2.10819796e-01 -1.02555764e+00 -8.27597857e-01
-1.46619603e-01 -1.88478217e-01 7.43035376e-01 1.28114438e+00
-1.27234966e-01 1.56934261e-01 -5.56256473e-01 3.44214201e-01
9.82725263e-01 6.33315071e-02 7.33433902e-01 -1.58218074e+00
-1.10897683e-01 -2.03569159e-01 -7.31607601e-02 -1.16386724e+00
3.81478071e-01 -1.27788961e+00 4.92640063e-02 -1.05842090e+00
3.97522837e-01 -1.23071897e+00 -4.57032263e-01 6.91368282e-01
-2.01454192e-01 3.04228276e-01 8.01230073e-02 7.42520571e-01
-7.12450862e-01 3.27577502e-01 7.49997318e-01 -2.62077719e-01
-3.83818030e-01 1.00350276e-01 -8.44404340e-01 3.79058212e-01
7.71040142e-01 -7.44365335e-01 -4.76268828e-01 -7.01917484e-02
-2.11769253e-01 -3.21687460e-01 2.46815175e-01 -9.61953819e-01
4.67526585e-01 -7.00433329e-02 4.05338854e-01 -3.52476835e-01
3.65919024e-01 -7.14446068e-01 3.49154383e-01 2.56630629e-01
-4.86382127e-01 -3.87131900e-01 -1.91015765e-01 8.63066316e-01
1.49408236e-01 -1.97550476e-01 1.04954255e+00 1.36039350e-02
-9.96389091e-01 3.39253008e-01 -1.43951744e-01 -1.65939629e-01
1.13664448e+00 -4.72428769e-01 -2.07532749e-01 -1.40488982e-01
-1.03742337e+00 4.65986103e-01 5.70777476e-01 2.37838492e-01
2.51694769e-01 -1.32011259e+00 -6.59307718e-01 3.88832897e-01
2.21843973e-01 -1.02503844e-01 6.42979205e-01 6.82416081e-01
-2.93159159e-04 9.69227254e-02 -1.01336092e-01 -5.42512894e-01
-1.26700187e+00 6.19524658e-01 3.35516781e-01 -2.86175102e-01
-3.74077827e-01 8.83693695e-01 1.73063725e-01 -7.15556204e-01
3.26557845e-01 4.63875175e-01 -3.61016840e-01 1.39424369e-01
4.35503006e-01 6.34825528e-01 -2.66299754e-01 -9.57118392e-01
-3.41782749e-01 7.83854008e-01 -2.99609691e-01 4.94121239e-02
9.71216381e-01 -6.56125546e-01 -1.05013788e-01 1.92989320e-01
1.07762182e+00 -2.28410825e-01 -1.14528239e+00 -4.72621709e-01
2.23511398e-01 -4.11669046e-01 -4.69020039e-01 -8.75304461e-01
-7.33544588e-01 3.92483175e-01 1.01453424e+00 -6.61906898e-02
1.13785267e+00 8.40719491e-02 8.56082141e-01 1.74631000e-01
5.68598270e-01 -1.46568215e+00 -1.29260644e-01 2.00065419e-01
1.92201495e-01 -1.25939059e+00 -2.23145336e-01 -6.65621817e-01
-6.09820127e-01 6.67498946e-01 7.03381360e-01 2.77448207e-01
4.39124197e-01 -3.74203026e-01 8.11481699e-02 2.33032525e-01
-1.59298271e-01 -4.21692491e-01 1.15471050e-01 9.29328918e-01
-1.35246709e-01 3.96429569e-01 -3.10424596e-01 1.00912952e+00
3.46689224e-02 -8.82249624e-02 3.75558101e-02 6.14395142e-01
-2.15843454e-01 -1.67305088e+00 -6.24720693e-01 4.37045038e-01
-7.33912811e-02 9.40497145e-02 -2.65321136e-01 3.81427407e-01
6.40856087e-01 1.19126403e+00 -9.80926231e-02 -6.62919521e-01
2.67533571e-01 5.83538413e-01 6.25914931e-02 -5.30136526e-01
-5.37769079e-01 1.88101485e-01 1.59833625e-01 -1.89422756e-01
-4.71606493e-01 -9.37085211e-01 -1.37076855e+00 -4.53406394e-01
-2.34430566e-01 3.48239452e-01 4.19839293e-01 9.26963627e-01
5.50241947e-01 -1.74736623e-02 9.47288513e-01 -6.01504028e-01
-3.32695872e-01 -9.49986994e-01 -8.93805683e-01 7.70988584e-01
-7.22276196e-02 -7.76314914e-01 -9.85054076e-02 2.99698919e-01] | [14.836263656616211, 1.0986216068267822] |
4ede4c6b-691e-467a-b264-81a24608b1ac | a-simple-probabilistic-model-for-uncertainty | 1807.09312 | null | http://arxiv.org/abs/1807.09312v1 | http://arxiv.org/pdf/1807.09312v1.pdf | A Simple Probabilistic Model for Uncertainty Estimation | The article focuses on determining the predictive uncertainty of a model on
the example of atrial fibrillation detection problem by a single-lead ECG
signal. To this end, the model predicts parameters of the beta distribution
over class probabilities instead of these probabilities themselves. It was
shown that the described approach allows to detect atypical recordings and
significantly improve the quality of the algorithm on confident predictions. | ['Alexander Kuvaev', 'Roman Khudorozhkov'] | 2018-07-24 | null | null | null | null | ['atrial-fibrillation-detection'] | ['medical'] | [ 4.50590968e-01 5.66505313e-01 -5.78696914e-02 -6.73970938e-01
-8.52293491e-01 -5.97227275e-01 8.95040482e-02 2.23166317e-01
-4.31337021e-02 1.36031830e+00 -2.96808243e-01 -7.10439444e-01
-5.25093079e-01 -3.29996049e-01 -1.96170092e-01 -7.39633381e-01
-5.68216503e-01 7.55829513e-01 -5.20339832e-02 5.52102864e-01
2.37472355e-01 3.99428964e-01 -9.80908692e-01 3.33024800e-01
7.82661021e-01 1.10816693e+00 -3.11257273e-01 9.89565492e-01
5.13020039e-01 2.67456353e-01 -9.47939098e-01 -2.76450574e-01
2.33035550e-01 -4.76254374e-01 -2.27202550e-01 -3.01480204e-01
-2.21024722e-01 -4.53580618e-02 2.25608200e-01 8.59404445e-01
5.12729466e-01 -4.98120904e-01 1.20055509e+00 -1.03009665e+00
3.68961602e-01 9.54274178e-01 -8.29771236e-02 5.50922215e-01
2.75693953e-01 -2.16182858e-01 6.78115904e-01 -6.92074776e-01
7.16455877e-02 5.63600302e-01 1.10180175e+00 -8.42164829e-03
-1.38612711e+00 -4.94611174e-01 -4.27428275e-01 -7.68780485e-02
-1.63774800e+00 -3.77624072e-02 5.01916587e-01 -5.71499407e-01
7.69079089e-01 4.89842951e-01 8.76873553e-01 7.89397359e-01
8.82190704e-01 2.62917072e-01 1.23696601e+00 -4.88765091e-01
3.00780505e-01 3.92918289e-01 5.38451731e-01 3.06616038e-01
8.02819610e-01 6.24164760e-01 -3.42802584e-01 -1.01003778e+00
4.80748475e-01 -2.24414259e-01 -4.26061571e-01 -2.67348409e-01
-1.00910223e+00 6.34199083e-01 -5.70815861e-01 8.90914649e-02
-5.99901259e-01 -1.32519051e-01 1.69553980e-01 2.13166475e-01
2.76003450e-01 5.43997586e-01 -6.45768046e-01 -3.32507670e-01
-1.03137100e+00 3.06932807e-01 1.07833064e+00 8.18199873e-01
7.63164312e-02 8.19799826e-02 -3.28888983e-01 1.71217352e-01
2.61774391e-01 8.13312232e-01 3.65076587e-02 -7.97444224e-01
1.67546943e-01 2.26378087e-02 4.31173593e-01 -4.64778692e-01
-5.31996429e-01 -9.66673076e-01 -9.14014041e-01 4.53076735e-02
4.82157528e-01 -4.56396669e-01 -8.27927411e-01 1.27835321e+00
-2.51116931e-01 1.90588802e-01 1.84929878e-01 3.44564706e-01
1.56380460e-01 4.41885352e-01 1.23576477e-01 -1.06030786e+00
8.30257237e-01 1.14958353e-01 -8.84037852e-01 1.56127766e-01
3.33666712e-01 -4.58797753e-01 1.07114084e-01 1.07452202e+00
-1.00497556e+00 -3.87944609e-01 -1.17230725e+00 1.04150927e+00
4.45846707e-01 1.69336960e-01 4.00209069e-01 1.08736992e+00
-6.75684571e-01 9.32260036e-01 -6.94312096e-01 2.05024898e-01
1.81039140e-01 4.96228248e-01 7.32521489e-02 3.64780277e-01
-1.18353820e+00 1.03050351e+00 6.49546027e-01 1.24895431e-01
-4.48791713e-01 -7.64004350e-01 -4.27016288e-01 5.46778738e-02
-4.69508618e-01 -4.72804099e-01 1.08160734e+00 -8.59853089e-01
-1.03053701e+00 4.83083010e-01 -1.26855850e-01 -8.73494327e-01
6.49690390e-01 -6.59074038e-02 -6.17193818e-01 2.76650228e-02
-1.44395247e-01 -2.13102087e-01 1.10068691e+00 -1.09506941e+00
-6.47910118e-01 -2.52855510e-01 -6.67911530e-01 -1.87864766e-01
5.60245216e-01 -3.26923668e-01 3.49021286e-01 -6.51059687e-01
2.87990868e-01 -8.74654412e-01 -3.93525660e-01 -7.48800099e-01
-3.63674641e-01 -3.42510752e-02 1.68325469e-01 -8.22334170e-01
1.70840704e+00 -2.21469808e+00 -1.53016254e-01 1.16997790e+00
2.92165279e-02 6.71414239e-03 7.10422337e-01 2.34809250e-01
-2.91065633e-01 1.22892871e-01 -6.33253753e-01 4.34596598e-01
-3.07559520e-01 2.31832370e-01 -6.85794592e-01 5.24787784e-01
4.15113494e-02 3.73155743e-01 -5.43328285e-01 -5.18137038e-01
9.72355623e-03 4.22283083e-01 -6.65471107e-02 1.91358939e-01
2.00801298e-01 3.18120390e-01 -3.48883331e-01 3.97279769e-01
5.97365320e-01 -8.91328305e-02 6.98479652e-01 2.44253073e-02
3.39456081e-01 -1.67878326e-02 -1.09378624e+00 7.29920268e-01
-1.47930318e-02 3.34158301e-01 -7.07845926e-01 -1.02519262e+00
1.32633495e+00 7.74790406e-01 4.55973238e-01 2.72624552e-01
1.15197606e-01 3.14414173e-01 4.03652489e-01 -8.27830136e-02
-4.40342784e-01 -5.41416347e-01 2.31474228e-02 3.59295607e-01
-1.02110198e-02 -2.40818232e-01 -3.14552873e-01 -1.14557952e-01
8.58429432e-01 -4.92587909e-02 9.34363067e-01 -5.72282970e-01
3.16368580e-01 -1.56325892e-01 7.75181949e-01 1.16505229e+00
-6.08210675e-02 6.20035350e-01 6.06938124e-01 -5.55821896e-01
-7.94587672e-01 -1.44986606e+00 -1.09208536e+00 -5.16463406e-02
-2.35267177e-01 -4.42453444e-01 -5.42100132e-01 -7.34230042e-01
1.34375721e-01 1.02522588e+00 -5.94586849e-01 -1.40249223e-01
-3.93597126e-01 -1.36850321e+00 6.79655254e-01 6.19180918e-01
-1.92301318e-01 -6.75263762e-01 -8.74885559e-01 4.34652179e-01
-1.86943546e-01 -5.68565369e-01 2.75239110e-01 7.82131195e-01
-1.46301806e+00 -1.14752066e+00 -5.18386364e-01 -3.21847141e-01
5.17953813e-01 -8.90000701e-01 1.33555579e+00 -1.51031733e-01
-2.69360840e-01 2.93736666e-01 -7.81486109e-02 -9.18836117e-01
-4.89732683e-01 -5.02255797e-01 2.10653350e-01 -1.37482569e-01
5.61361909e-01 -7.12780595e-01 -4.70982581e-01 3.65371346e-01
-1.53044313e-01 -4.61514592e-01 5.24177372e-01 1.02605975e+00
7.79656887e-01 8.14282522e-02 1.09445584e+00 -9.95536089e-01
7.65969753e-01 -4.86861557e-01 -4.82851237e-01 3.70779663e-01
-1.34515715e+00 -7.62622729e-02 3.83387655e-01 -4.54430699e-01
-9.21559095e-01 4.65009540e-01 1.67786494e-01 -7.00615272e-02
-1.03138976e-01 3.22834969e-01 2.48947039e-01 1.10240579e-01
8.02942753e-01 2.63683826e-01 -1.48242071e-01 -1.87341154e-01
-2.54272431e-01 8.74981880e-01 6.15706265e-01 -4.66846675e-01
2.39521801e-01 -4.86610644e-02 3.51366311e-01 -5.69136143e-01
-4.93954003e-01 -3.88905913e-01 -7.36856639e-01 -1.92784846e-01
3.44470561e-01 -7.92444170e-01 -3.60732615e-01 -3.02788187e-02
-9.75029171e-01 1.52164429e-01 -4.64366496e-01 9.70457137e-01
-9.27541137e-01 5.17763317e-01 -2.44711906e-01 -1.46145236e+00
-4.17606026e-01 -5.69512069e-01 6.13609254e-01 -2.33304769e-01
-7.25591600e-01 -9.01484847e-01 3.26270789e-01 -2.54885375e-01
-9.31222588e-02 3.03994834e-01 9.51327741e-01 -1.04499018e+00
-1.05027303e-01 -7.95978725e-01 2.53140032e-01 5.32116652e-01
-1.71254963e-01 1.35507151e-01 -1.06517303e+00 -2.71301180e-01
3.86534303e-01 4.50525016e-01 4.42326635e-01 8.29681993e-01
1.09665024e+00 1.03813522e-01 -6.83655202e-01 4.49929714e-01
1.20440865e+00 5.82029521e-01 5.39262295e-01 -8.97172466e-02
-1.80022821e-01 1.13654204e-01 8.71953189e-01 8.22687209e-01
-4.14654851e-01 4.01442885e-01 -2.55050207e-03 4.87776786e-01
5.30354559e-01 5.39517589e-02 -8.11818540e-02 3.31128627e-01
-5.15832543e-01 6.27083555e-02 -1.14140344e+00 3.87859911e-01
-1.83459389e+00 -9.57308292e-01 -2.28555322e-01 2.30445075e+00
6.36448979e-01 4.74868149e-01 1.39628232e-01 6.73110843e-01
8.13534617e-01 -5.18060744e-01 -4.39360648e-01 -5.32232344e-01
-2.10300520e-01 2.22390980e-01 3.63793820e-01 7.42247403e-01
-8.97582173e-01 -2.58287247e-02 8.96931839e+00 3.94819945e-01
-5.32193720e-01 -2.41265267e-01 7.81278074e-01 3.08603793e-01
-1.36802837e-01 3.52645107e-02 -7.62609303e-01 5.01206219e-01
1.21359599e+00 -5.34509420e-01 -1.35172307e-01 8.18256199e-01
3.27263474e-02 -1.86023727e-01 -1.11398399e+00 9.91565824e-01
-2.78920680e-02 -9.10256863e-01 -2.87683047e-02 2.18191445e-02
3.04826528e-01 -4.61790860e-01 -2.40824670e-01 6.70684353e-02
-3.65133941e-01 -1.09127998e+00 3.51202488e-01 1.06433570e+00
8.67728591e-01 -9.68260050e-01 1.35887122e+00 4.78746057e-01
-5.39834261e-01 -2.69766629e-01 -3.38648468e-01 -1.71831384e-01
3.50288421e-01 1.10170591e+00 -1.37456620e+00 3.79661977e-01
4.48094487e-01 5.91375411e-01 -3.45649540e-01 1.43935716e+00
1.35490568e-02 1.21690953e+00 -6.51461959e-01 1.87315241e-01
-5.08992970e-01 -1.42626017e-01 7.78129220e-01 1.30772853e+00
4.89886314e-01 5.08380711e-01 -5.04885055e-02 5.02403200e-01
8.03268313e-01 -8.67811218e-02 -7.97243953e-01 4.00240183e-01
3.85825962e-01 5.50850928e-01 -4.64537710e-01 -4.38180983e-01
5.55938631e-02 2.51798302e-01 -2.69978762e-01 2.80669510e-01
-6.54880881e-01 -4.32263583e-01 -7.55437836e-02 2.76355684e-01
6.34657517e-02 2.84332335e-01 -9.21297431e-01 -8.69932652e-01
1.31006941e-01 -5.98899126e-01 6.82132721e-01 -5.36399245e-01
-1.09492576e+00 9.35743392e-01 3.46519232e-01 -1.44455469e+00
-8.06314707e-01 -5.93688846e-01 -5.43876350e-01 1.17596114e+00
-7.39900887e-01 -2.92896926e-01 3.66645247e-01 4.78528917e-01
2.37473994e-02 -5.60872853e-01 1.31303632e+00 -2.25814924e-01
9.42063555e-02 4.22999471e-01 1.56647936e-01 -2.88371891e-01
5.21310985e-01 -1.42994869e+00 -4.45016026e-02 7.43844092e-01
4.30026688e-02 5.33523023e-01 1.20076180e+00 -9.04394627e-01
-3.45219314e-01 -6.96456730e-01 9.70468283e-01 -8.44296515e-01
1.51737139e-01 3.17430139e-01 -6.79580927e-01 6.07519150e-01
-2.50390112e-01 -2.88349688e-02 9.85017419e-01 3.20792079e-01
1.39071062e-01 -6.84517175e-02 -1.33399963e+00 -1.20275781e-01
4.91123706e-01 -1.25823781e-01 -1.04380381e+00 2.97030479e-01
-2.53661215e-01 -2.81682909e-01 -1.21303821e+00 1.02370703e+00
9.67137218e-01 -1.01593101e+00 9.30116832e-01 -6.97549462e-01
-4.27708536e-01 -1.99651778e-01 -2.21183822e-01 -1.09931147e+00
-3.38167012e-01 -8.18245053e-01 -4.89362568e-01 5.04973948e-01
8.71486366e-01 -5.57778716e-01 5.86731553e-01 4.52692091e-01
1.42489225e-01 -8.40252757e-01 -1.17957532e+00 -4.81965095e-01
-1.60607591e-01 -5.31231642e-01 1.79625198e-01 3.67462993e-01
4.02681470e-01 2.42686048e-02 -5.09199321e-01 4.78333652e-01
1.02720904e+00 2.19020531e-01 -6.25410676e-02 -1.74690449e+00
-6.86784923e-01 2.11128652e-01 -7.65370965e-01 -3.30887049e-01
-1.46887183e-01 -5.16668081e-01 1.33131564e-01 -9.04883027e-01
5.38164154e-02 -6.62687242e-01 -7.36925244e-01 -1.39671033e-02
-2.13555425e-01 2.67556638e-01 -1.82993621e-01 1.65685177e-01
-8.27826113e-02 -2.63163686e-01 5.21149397e-01 7.45233893e-02
-2.48820797e-01 1.06346941e+00 -4.87102091e-01 9.21353519e-01
1.07781982e+00 -8.49213958e-01 -6.05280519e-01 4.96640891e-01
4.41215813e-01 7.91404128e-01 3.23779285e-01 -9.64883566e-01
-2.75168568e-01 1.12395354e-01 8.68825734e-01 -7.20486104e-01
9.48802233e-02 -7.73419380e-01 5.79717696e-01 6.61492229e-01
-2.11247370e-01 1.71437204e-01 7.53433406e-02 1.02111018e+00
-2.44564965e-01 -4.12205517e-01 6.82824671e-01 1.41845107e-01
1.00187458e-01 -1.62414521e-01 -6.08007729e-01 1.82300750e-02
1.03941071e+00 -2.07875967e-01 2.50571132e-01 -3.76596361e-01
-1.66394460e+00 -2.15288833e-01 -9.03934315e-02 -1.52021483e-01
7.41619110e-01 -8.55604410e-01 -7.17721581e-01 4.92957354e-01
8.93764347e-02 -2.80983120e-01 -1.05113842e-01 8.99975240e-01
-4.35173541e-01 4.52301383e-01 -2.70250738e-01 -1.00621772e+00
-1.32013357e+00 3.13003808e-01 7.14615822e-01 -1.66328117e-01
-7.49844909e-01 4.50986207e-01 -1.59743905e-01 2.14065686e-01
3.66607577e-01 -2.21042067e-01 -3.04393798e-01 -3.62410247e-01
5.31538904e-01 2.59011298e-01 2.08145734e-02 -1.33980319e-01
-5.14602900e-01 6.37034297e-01 1.56260461e-01 -2.56689638e-01
8.85045350e-01 -5.69912940e-02 2.00002510e-02 9.09608781e-01
2.43252829e-01 -4.40988541e-02 -9.26053047e-01 2.99260288e-01
9.42200944e-02 -3.18623781e-01 -1.98549896e-01 -1.31096566e+00
-3.07685345e-01 8.28371525e-01 1.05785108e+00 2.62581021e-01
1.16784167e+00 -2.65227050e-01 -8.59527066e-02 5.36567688e-01
5.53300500e-01 -7.15992987e-01 -8.76751125e-01 -1.59885019e-01
7.90924907e-01 -9.37289655e-01 4.37073171e-01 -4.32009608e-01
-8.91496599e-01 1.41112971e+00 -1.64767310e-01 -2.60134935e-01
1.40128207e+00 4.09661412e-01 1.68953478e-01 2.84554929e-01
-6.94861293e-01 5.49297810e-01 3.97810906e-01 1.03703558e+00
3.39812458e-01 5.37856638e-01 -7.71010876e-01 1.31765199e+00
-7.24760965e-02 4.01011556e-01 5.34487307e-01 6.37931526e-01
-3.41244608e-01 -8.71357501e-01 -3.93655866e-01 9.61850643e-01
-7.30353296e-01 -1.62855580e-01 -1.18605681e-01 4.94122177e-01
9.92732048e-02 9.40880418e-01 -2.09665835e-01 -2.97609866e-01
2.29732975e-01 4.24404740e-01 4.48419422e-01 -3.01424950e-01
-3.56888026e-01 5.40352762e-02 4.15741444e-01 -1.72237977e-01
-2.49902695e-01 -1.04359496e+00 -8.37647974e-01 3.56444389e-01
-4.34241891e-01 8.17977369e-01 3.92372280e-01 7.00875998e-01
1.73854828e-01 2.86407113e-01 6.95024848e-01 -2.84188956e-01
-9.98715758e-01 -1.06564879e+00 -1.11820638e+00 -1.07858116e-02
3.24909210e-01 -5.52874207e-01 -7.19610155e-01 1.19879529e-01] | [14.163003921508789, 3.2433977127075195] |
f8d2ddc0-89d3-4fd7-8589-6745b6bb5057 | differentially-private-episodic-reinforcement | 2306.01121 | null | https://arxiv.org/abs/2306.01121v2 | https://arxiv.org/pdf/2306.01121v2.pdf | Differentially Private Episodic Reinforcement Learning with Heavy-tailed Rewards | In this paper, we study the problem of (finite horizon tabular) Markov decision processes (MDPs) with heavy-tailed rewards under the constraint of differential privacy (DP). Compared with the previous studies for private reinforcement learning that typically assume rewards are sampled from some bounded or sub-Gaussian distributions to ensure DP, we consider the setting where reward distributions have only finite $(1+v)$-th moments with some $v \in (0,1]$. By resorting to robust mean estimators for rewards, we first propose two frameworks for heavy-tailed MDPs, i.e., one is for value iteration and another is for policy optimization. Under each framework, we consider both joint differential privacy (JDP) and local differential privacy (LDP) models. Based on our frameworks, we provide regret upper bounds for both JDP and LDP cases and show that the moment of distribution and privacy budget both have significant impacts on regrets. Finally, we establish a lower bound of regret minimization for heavy-tailed MDPs in JDP model by reducing it to the instance-independent lower bound of heavy-tailed multi-armed bandits in DP model. We also show the lower bound for the problem in LDP by adopting some private minimax methods. Our results reveal that there are fundamental differences between the problem of private RL with sub-Gaussian and that with heavy-tailed rewards. | ['Di Wang', 'Sayak Ray Chowdhury', 'Xingyu Zhou', 'Yulian Wu'] | 2023-06-01 | null | null | null | null | ['multi-armed-bandits'] | ['miscellaneous'] | [ 4.79390845e-02 3.14583570e-01 -6.21554613e-01 -4.25144255e-01
-1.22243607e+00 -7.79847443e-01 3.71325538e-02 -6.84587751e-03
-5.56160152e-01 1.39754796e+00 1.63972795e-01 -5.51862776e-01
-5.71919680e-01 -8.82866204e-01 -8.28888655e-01 -1.16074216e+00
-1.23580188e-01 4.09480900e-01 -3.57557118e-01 2.82583963e-02
2.01842800e-01 3.14587891e-01 -1.01105309e+00 -8.60024840e-02
7.75725067e-01 1.36062407e+00 -9.94226038e-02 5.82670629e-01
-5.17959893e-02 9.49088514e-01 -5.03484666e-01 -7.56096423e-01
8.75515163e-01 -2.32936233e-01 -6.63588524e-01 -2.04842761e-01
-1.75227508e-01 -9.15925741e-01 -3.93076479e-01 1.25067520e+00
6.84121668e-01 2.82625645e-01 7.21808672e-01 -1.55931425e+00
-6.02862239e-01 8.51910472e-01 -1.10270441e+00 -8.20260346e-02
1.13665253e-01 4.85441685e-02 1.11309707e+00 2.97947317e-01
4.08529520e-01 1.52114654e+00 2.89418876e-01 8.36442590e-01
-1.23915911e+00 -8.03574204e-01 2.47784108e-01 -2.45308444e-01
-9.44133103e-01 -3.20651382e-01 3.12443256e-01 -2.41188526e-01
5.49145222e-01 2.99743533e-01 2.91428149e-01 1.00471890e+00
3.28684062e-01 1.10548270e+00 1.41222811e+00 -2.35895753e-01
8.12243998e-01 1.49455041e-01 -3.98885831e-02 8.62189382e-02
4.85721856e-01 3.92337918e-01 -1.47608504e-01 -5.90337276e-01
7.81941116e-01 3.21917981e-01 -4.02357988e-02 -6.41418040e-01
-6.69168055e-01 1.16917396e+00 -2.86793590e-01 -3.51191312e-01
-7.17807889e-01 5.73389471e-01 3.83073062e-01 7.37236023e-01
2.57934272e-01 -1.10385828e-01 -4.69796836e-01 -3.17481548e-01
-7.81260431e-01 7.04839408e-01 1.13875437e+00 1.43245482e+00
4.94935870e-01 -1.40382081e-01 -1.07369888e+00 3.94574940e-01
1.36563107e-01 6.95662200e-01 1.82808656e-02 -1.37113845e+00
9.70772743e-01 -3.96602690e-01 1.05437934e+00 -5.83964348e-01
1.94327831e-01 -2.92448372e-01 -5.74612498e-01 2.88951725e-01
6.33286595e-01 -7.15954900e-01 -4.78548765e-01 2.02045441e+00
2.70706058e-01 -5.43993235e-01 3.41049582e-01 7.37098753e-01
-1.19603328e-01 5.31229079e-01 -1.49146855e-01 -9.24537957e-01
9.27170098e-01 -6.23291016e-01 -1.09534287e+00 1.99030891e-01
4.62069690e-01 -3.04229915e-01 8.23485792e-01 5.51357806e-01
-1.24515486e+00 3.87988418e-01 -6.42635286e-01 2.19980821e-01
1.66869074e-01 -4.83144522e-01 6.11474574e-01 1.17990744e+00
-8.79497886e-01 6.82002544e-01 -4.61148083e-01 -1.06425863e-02
7.17294335e-01 4.25955981e-01 1.25099540e-01 -1.41208708e-01
-9.82321978e-01 3.61496240e-01 1.94201693e-01 -2.79857516e-01
-1.40775824e+00 -4.60787535e-01 -2.61615485e-01 2.18497932e-01
1.00162292e+00 -6.30653918e-01 1.58105278e+00 -8.56104672e-01
-1.79101515e+00 5.36150098e-01 2.63492204e-02 -8.85197699e-01
1.18853831e+00 -2.47994170e-01 2.78542250e-01 7.60963112e-02
1.11110501e-01 -1.49672208e-02 8.29299510e-01 -1.09293330e+00
-9.41024363e-01 -6.50828898e-01 4.38763291e-01 3.12224954e-01
-3.48400250e-02 -2.16219321e-01 1.73575357e-01 -3.19155484e-01
-3.71565700e-01 -7.71583200e-01 -4.73041266e-01 -1.30640149e-01
-4.70803410e-01 4.14554551e-02 5.70358455e-01 -4.90069658e-01
1.13689744e+00 -2.09477067e+00 -1.37627676e-01 3.03217500e-01
-5.68086281e-02 -2.72480190e-01 1.25446856e-01 6.11231267e-01
4.84509408e-01 8.48032236e-02 -8.54735002e-02 -5.21043241e-01
6.72932267e-01 6.43070400e-01 -7.15641797e-01 6.01736069e-01
-5.24326086e-01 5.46073258e-01 -7.74448156e-01 -2.78958708e-01
-2.40724698e-01 -2.55620629e-01 -5.01063287e-01 2.88553536e-01
-3.53951126e-01 4.71373826e-01 -9.36656296e-01 6.33086860e-01
1.04774773e+00 6.69088513e-02 1.29317284e-01 5.28380871e-01
-7.20873028e-02 -3.26017886e-01 -1.38923085e+00 1.22462296e+00
-2.69802511e-01 -2.45658562e-01 6.42614365e-01 -9.84104395e-01
6.36878669e-01 3.20349872e-01 4.99038130e-01 -4.89904433e-01
2.93263108e-01 1.98987201e-01 -3.22687685e-01 -2.63717115e-01
5.41362703e-01 -8.01731706e-01 -1.51432842e-01 7.67911315e-01
-1.96260452e-01 5.76747619e-02 -1.92462817e-01 1.94636248e-02
1.07721186e+00 -1.09351173e-01 2.76990533e-01 -2.68224299e-01
1.44845294e-02 -5.74859381e-01 9.08455193e-01 1.33266830e+00
-7.35981107e-01 2.49300972e-01 1.17495751e+00 1.41484171e-01
-9.10103798e-01 -9.55718040e-01 1.17036868e-02 1.17229164e+00
1.10249430e-01 2.81941265e-01 -5.81314266e-01 -7.55534470e-01
6.04782522e-01 1.16331685e+00 -7.09219933e-01 2.31921628e-01
3.81728917e-01 -7.43118286e-01 4.22399223e-01 1.60676152e-01
6.58445001e-01 -5.40465653e-01 -5.53884327e-01 1.16730571e-01
2.30687022e-01 -6.67429507e-01 -4.91690844e-01 4.15915430e-01
-8.14418137e-01 -7.01439440e-01 -1.14545560e+00 -1.88649911e-02
3.19523185e-01 8.89839157e-02 7.13767469e-01 -9.13533866e-01
2.88719267e-01 1.03444374e+00 -2.88506329e-01 -7.50892878e-01
1.35054156e-01 -2.13386297e-01 -9.43680853e-03 1.61968172e-01
2.12455779e-01 -5.18053353e-01 -8.64598930e-01 8.35113302e-02
-1.01412606e+00 -5.32474995e-01 4.88030106e-01 8.28276217e-01
9.45317626e-01 1.84910700e-01 7.59648025e-01 -9.89092171e-01
1.11820519e+00 -7.36146986e-01 -9.37994540e-01 4.87052977e-01
-7.35873401e-01 3.45597535e-01 4.84539360e-01 -3.14609885e-01
-1.24681127e+00 -3.88248563e-02 2.72580206e-01 -5.59398711e-01
1.22987136e-01 3.01433988e-02 -2.68890530e-01 8.34345371e-02
7.38714784e-02 2.11705148e-01 1.85293138e-01 -2.69767702e-01
3.37587237e-01 7.68613815e-01 -3.27143744e-02 -1.22312653e+00
3.43701243e-01 5.74653208e-01 3.75059932e-01 -2.59937525e-01
-9.06225145e-01 -6.39171749e-02 1.04014747e-01 2.92426288e-01
6.56260431e-01 -8.22430253e-01 -1.41223145e+00 1.23467587e-01
-6.60120428e-01 -3.82995039e-01 -8.34218681e-01 5.86802363e-01
-1.57422090e+00 4.60955888e-01 -5.64889848e-01 -1.92106903e+00
-3.22955370e-01 -9.16349649e-01 7.52936661e-01 8.52183476e-02
5.28080523e-01 -8.02720666e-01 1.62816331e-01 2.91386724e-01
2.88366437e-01 5.23472905e-01 7.28601038e-01 -6.43933296e-01
-6.18855298e-01 -9.14139580e-03 -5.73281245e-03 4.29815829e-01
-1.54779255e-01 -7.18694687e-01 -7.55148888e-01 -6.75201654e-01
5.00270367e-01 -6.40515387e-01 5.29775202e-01 8.45098972e-01
1.32231283e+00 -9.88713622e-01 -1.14038534e-01 5.87059736e-01
1.61280179e+00 6.04809999e-01 5.41642368e-01 4.59857941e-01
-7.78474205e-04 6.42743587e-01 1.18661714e+00 1.46912539e+00
2.58534431e-01 9.61489305e-02 8.74845743e-01 4.22835946e-01
1.10342860e+00 -4.29047048e-01 4.75872368e-01 -2.37707317e-01
-5.84290922e-02 -2.28132352e-01 -2.77390748e-01 4.49208766e-01
-2.29972744e+00 -1.10031772e+00 1.72198966e-01 2.96232677e+00
9.35383201e-01 -1.89440191e-01 5.83476841e-01 -2.25254402e-01
7.63484776e-01 5.52288666e-02 -1.01650691e+00 -9.12491500e-01
-2.11547196e-01 2.33067553e-02 1.36006296e+00 4.03817356e-01
-7.30639756e-01 5.82896948e-01 6.15692234e+00 1.14321434e+00
-4.33896810e-01 2.48089120e-01 1.12305045e+00 -6.45032227e-01
-5.45002759e-01 6.12641610e-02 -9.89959180e-01 5.95307112e-01
9.18406963e-01 -6.06220305e-01 8.31367731e-01 1.04293716e+00
3.22950602e-01 -2.34732270e-01 -1.21648622e+00 1.06429267e+00
-6.94562674e-01 -7.36720443e-01 -3.84349614e-01 6.86369658e-01
9.92027819e-01 -3.38408679e-01 4.48026150e-01 4.69590217e-01
1.15041423e+00 -1.09398925e+00 6.76319182e-01 5.42642772e-01
8.86178613e-01 -1.45743012e+00 6.90492630e-01 5.96585751e-01
-4.97878402e-01 -6.48114145e-01 -6.84839427e-01 -1.48978800e-01
5.67026548e-02 5.23726761e-01 -2.09768727e-01 7.23419428e-01
6.68152988e-01 1.53325662e-01 5.42347491e-01 8.45664144e-01
1.30726611e-02 3.68818223e-01 -2.32045472e-01 -1.23193152e-01
5.55079937e-01 -5.21668971e-01 4.89505827e-01 6.76647782e-01
6.24161005e-01 2.30441496e-01 1.00831889e-01 8.44262660e-01
1.43921256e-01 1.23082839e-01 -6.66843891e-01 -1.67629838e-01
4.55589294e-01 7.95914352e-01 -1.18757285e-01 -8.67399350e-02
-2.54466444e-01 9.60834265e-01 1.68809578e-01 5.18147051e-01
-7.84111142e-01 -4.45107758e-01 9.39136684e-01 -1.11063749e-01
6.73979163e-01 2.00817361e-02 -2.48901427e-01 -8.98864508e-01
1.54782301e-02 -6.54246390e-01 7.35326767e-01 -1.26517579e-01
-1.66568553e+00 -1.91112995e-01 9.21827853e-02 -9.31630373e-01
-3.37514371e-01 -2.58379430e-01 -2.99081624e-01 1.03671598e+00
-1.77866733e+00 -7.11172938e-01 7.30019689e-01 1.11526549e+00
-6.22995459e-02 -3.93463969e-02 5.37814200e-01 3.17005701e-02
-3.52556944e-01 7.64343023e-01 1.25215101e+00 -4.78368431e-01
5.40384293e-01 -1.35459900e+00 -5.60744584e-01 3.10042769e-01
-6.99446559e-01 3.75372261e-01 8.38116169e-01 -4.90442574e-01
-1.67706966e+00 -9.44444001e-01 1.57141939e-01 6.52572736e-02
5.47972083e-01 -1.26895070e-01 -3.01155090e-01 9.05445457e-01
3.97198498e-02 -1.24841640e-02 7.26806343e-01 5.46096712e-02
-4.79113199e-02 -2.70065099e-01 -2.01645494e+00 4.57450032e-01
1.00235796e+00 -1.98829755e-01 -1.19375095e-01 4.69622999e-01
7.62593806e-01 -2.94089168e-01 -9.00364995e-01 -1.66702211e-01
6.26877904e-01 -9.52951312e-01 5.29732704e-01 -9.61539984e-01
1.14935443e-01 1.89741299e-01 -7.73957431e-01 -1.05209756e+00
-6.64446875e-02 -1.33452606e+00 -2.16808051e-01 1.28173971e+00
3.16429920e-02 -8.94602716e-01 9.77637589e-01 1.13230979e+00
3.50275367e-01 -6.78193390e-01 -1.40143740e+00 -1.20424020e+00
5.86094558e-01 -2.42226020e-01 7.24642277e-01 6.08909428e-01
-9.71622579e-03 -3.52859288e-01 -9.44569230e-01 5.22560999e-02
9.68810797e-01 3.10867727e-01 7.92831898e-01 -6.19758308e-01
-8.64030898e-01 4.59355935e-02 2.13671774e-01 -1.16890848e+00
1.97411239e-01 -4.75273013e-01 1.15941137e-01 -1.11268473e+00
4.35597628e-01 -4.62147087e-01 -5.31276464e-01 2.92335719e-01
2.88926631e-01 -7.81843722e-01 3.95412266e-01 4.38429639e-02
-7.92753637e-01 9.99447644e-01 1.39844537e+00 1.30479932e-01
-1.91530332e-01 7.32825160e-01 -9.42714453e-01 2.07732707e-01
8.10353339e-01 -6.86073959e-01 -6.49601340e-01 -2.40136664e-02
1.01411909e-01 9.46693718e-01 2.28559449e-02 -3.57930064e-01
-1.87243372e-01 -9.89036679e-01 -1.43213481e-01 -6.05289221e-01
2.01977909e-01 -9.70523596e-01 1.77574251e-02 5.85423589e-01
-6.78430974e-01 -3.18480432e-01 -2.26346552e-01 1.26061010e+00
1.79841399e-01 -6.03168428e-01 7.01080859e-01 -4.69597489e-01
5.69542907e-02 5.95927596e-01 -3.85933667e-01 2.25137919e-01
1.29339516e+00 8.27124808e-03 -3.34900200e-01 -1.11893451e+00
-8.50437403e-01 5.96324682e-01 1.70886129e-01 -4.38665926e-01
3.85423303e-01 -1.16623306e+00 -3.38860840e-01 -4.36860949e-01
-3.73213947e-01 -9.32496488e-02 4.75876689e-01 7.44224548e-01
1.52902737e-01 4.12167758e-01 -3.34835410e-01 -8.99870396e-02
-8.75216663e-01 8.39218795e-01 1.60852537e-01 -7.17773974e-01
-4.03566882e-02 6.11238003e-01 2.91888446e-01 -1.75510719e-01
5.87661862e-01 -2.27110460e-01 2.05661759e-01 -2.53579337e-02
4.26183820e-01 7.75694311e-01 -5.87719440e-01 -2.75309961e-02
2.16002449e-01 -9.80578288e-02 -2.96962589e-01 -6.55015588e-01
1.45422566e+00 -5.99931479e-01 4.12376411e-02 4.51301903e-01
8.05129588e-01 -3.06487381e-02 -1.85581303e+00 -3.60777915e-01
-1.05731592e-01 -8.65313470e-01 -3.73687074e-02 -7.11717904e-01
-9.90747750e-01 7.82926679e-01 7.51186550e-01 2.97877580e-01
9.99264002e-01 -4.14831161e-01 7.07500875e-01 3.50498348e-01
8.79017055e-01 -1.49472857e+00 -3.24555129e-01 3.78550798e-01
7.21257746e-01 -1.05017734e+00 -1.18449330e-01 -2.00019055e-03
-1.12449336e+00 8.01139772e-01 1.07086584e-01 -5.09175137e-02
5.21696091e-01 2.06017613e-01 -5.16666293e-01 2.78504193e-01
-7.50373006e-01 -1.84216887e-01 -4.33354944e-01 8.81075799e-01
-5.01974300e-02 4.38253731e-01 -6.86991930e-01 1.00030816e+00
-7.38797337e-02 3.16522151e-01 6.09442651e-01 1.36318409e+00
-5.43299139e-01 -1.22413075e+00 -5.34530818e-01 5.49264789e-01
-1.10639596e+00 1.27220318e-01 -3.50107998e-02 5.15874982e-01
-3.87226969e-01 1.06673527e+00 -1.24968596e-01 1.07278891e-01
1.47140294e-01 -9.42971110e-02 5.11650324e-01 -2.44022131e-01
-4.24710065e-01 1.27626345e-01 -8.71038903e-03 -5.28373420e-01
-1.65053725e-01 -8.59320223e-01 -7.96691239e-01 -7.58721769e-01
-9.75514937e-04 2.73276597e-01 5.04037797e-01 6.82203531e-01
2.96019405e-01 8.20442569e-03 9.64541733e-01 -2.05424428e-01
-1.86919701e+00 -6.60813093e-01 -1.59616280e+00 1.45563528e-01
4.16171670e-01 -4.84495252e-01 -4.32320148e-01 -7.38072038e-01] | [4.529442310333252, 3.317413568496704] |
c6d4d732-0c2e-40f2-8533-796ba16add29 | dyngfn-bayesian-dynamic-causal-discovery | 2302.04178 | null | https://arxiv.org/abs/2302.04178v2 | https://arxiv.org/pdf/2302.04178v2.pdf | DynGFN: Towards Bayesian Inference of Gene Regulatory Networks with GFlowNets | One of the grand challenges of cell biology is inferring the gene regulatory network (GRN) which describes interactions between genes and their products that control gene expression and cellular function. We can treat this as a causal discovery problem but with two non-standard challenges: (1) regulatory networks are inherently cyclic so we should not model a GRN as a directed acyclic graph (DAG), and (2) observations have significant measurement noise, so for typical sample sizes there will always be a large equivalence class of graphs that are likely given the data, and we want methods that capture this uncertainty. Existing methods either focus on challenge (1), identifying cyclic structure from dynamics, or on challenge (2) learning complex Bayesian posteriors over DAGs, but not both. In this paper we leverage the fact that it is possible to estimate the "velocity" of gene expression with RNA velocity techniques to develop an approach that addresses both challenges. Because we have access to velocity information, we can treat the Bayesian structure learning problem as a problem of sparse identification of a dynamical system, capturing cyclic feedback loops through time. Since our objective is to model uncertainty over discrete structures, we leverage Generative Flow Networks (GFlowNets) to estimate the posterior distribution over the combinatorial space of possible sparse dependencies. Our results indicate that our method learns posteriors that better encapsulate the distributions of cyclic structures compared to counterpart state-of-the-art Bayesian structure learning approaches. | ['Yoshua Bengio', 'Bo wang', 'Leo J. Lee', 'Jason Hartford', 'Alexander Tong', 'Lazar Atanackovic'] | 2023-02-08 | null | null | null | null | ['causal-discovery'] | ['knowledge-base'] | [ 3.43992054e-01 4.44315374e-02 -3.46605659e-01 -1.35885682e-02
-5.58388233e-01 -8.19812715e-01 7.41776288e-01 -5.81064261e-02
9.10939202e-02 8.55371416e-01 5.17516553e-01 -4.89886850e-01
-6.54213667e-01 -6.88679397e-01 -9.34635758e-01 -9.31904197e-01
-5.69663405e-01 1.05197096e+00 2.56659184e-02 6.91069216e-02
7.36355036e-02 4.70144302e-01 -1.04682839e+00 -2.46341340e-02
4.65985090e-01 5.78266501e-01 -8.12151134e-02 1.13769507e+00
-9.43130329e-02 6.68306649e-01 -3.48596901e-01 1.40671715e-01
-1.33801460e-01 -5.18064797e-01 -7.84539402e-01 -1.15661122e-01
3.01304366e-02 1.73385143e-01 -4.23289418e-01 1.03758299e+00
1.54986680e-01 -5.64096831e-02 1.10063541e+00 -1.31946397e+00
-5.58213368e-02 7.74576187e-01 -7.36903131e-01 3.32863122e-01
9.48246866e-02 1.37930840e-01 1.24425125e+00 -3.02279145e-01
8.73094201e-01 1.61696005e+00 5.21022141e-01 2.70238906e-01
-2.08771348e+00 -4.28896844e-01 2.85353333e-01 -2.50006497e-01
-1.41529977e+00 -5.06443262e-01 5.58026791e-01 -8.93450797e-01
6.99916124e-01 1.20554149e-01 6.88128531e-01 1.46850181e+00
5.83372891e-01 4.39854741e-01 7.96115398e-01 -8.23380612e-03
6.21279180e-01 -7.21582949e-01 2.83557653e-01 6.24146283e-01
2.43257701e-01 1.15535803e-01 -5.72861195e-01 -6.31962240e-01
7.75870025e-01 -3.07161808e-02 -3.63915414e-01 -3.63324165e-01
-1.07192791e+00 8.01869452e-01 -3.07845324e-01 1.50087357e-01
-2.77988203e-02 7.48876154e-01 2.44639620e-01 6.08439855e-02
2.43314862e-01 2.92932987e-01 -7.25434482e-01 -2.91803896e-01
-7.29087710e-01 3.12015057e-01 1.35358822e+00 8.29194784e-01
7.13886380e-01 -1.48470208e-01 3.65324393e-02 5.08271873e-01
8.02774847e-01 2.23948628e-01 -6.53629228e-02 -9.84656334e-01
-1.01969279e-02 2.77921826e-01 -6.91252723e-02 -9.78304863e-01
-4.53735679e-01 -3.86654586e-01 -1.10754180e+00 -2.19334453e-01
9.31089640e-01 -4.46932346e-01 -1.18800247e+00 2.18795395e+00
1.76477537e-01 6.48768485e-01 -2.95363903e-01 4.45771962e-01
2.64535964e-01 8.64397347e-01 7.18751252e-02 -4.37640429e-01
1.19134378e+00 1.96674168e-02 -6.52590513e-01 -2.20475212e-01
4.99743938e-01 -3.54717910e-01 4.24854100e-01 4.77497101e-01
-7.66469181e-01 -1.45714143e-02 -8.20146322e-01 1.73811153e-01
-9.96096656e-02 -3.50059479e-01 1.05889070e+00 4.27170187e-01
-1.00985348e+00 7.62760580e-01 -1.40221846e+00 -3.77891481e-01
3.12174767e-01 4.76889968e-01 -1.67221501e-01 -2.86739022e-01
-1.10666120e+00 3.78111511e-01 1.56592339e-01 2.94705391e-01
-1.37921083e+00 -9.64761198e-01 -7.18234956e-01 2.68386960e-01
5.36754131e-01 -8.60973418e-01 8.39488864e-01 -3.93088549e-01
-1.18757820e+00 2.57357389e-01 -3.17864180e-01 -1.83904648e-01
8.92699510e-02 3.69311482e-01 -1.90831032e-02 -1.33075148e-01
-1.63403466e-01 3.86458784e-01 5.95498502e-01 -1.10783541e+00
-2.32247740e-01 -4.95756269e-01 -2.33897492e-01 -2.80518889e-01
2.26932406e-01 -3.70331854e-01 -5.11574745e-01 -2.91621655e-01
2.59787470e-01 -1.14156401e+00 -3.69189739e-01 -3.39604884e-01
-8.33820164e-01 -1.00255132e-01 6.63358390e-01 -3.44225198e-01
1.24891019e+00 -1.82405019e+00 5.70781231e-01 6.31981969e-01
5.35048723e-01 -3.23049515e-01 -9.33011994e-02 5.25908709e-01
-5.66156507e-02 5.27956665e-01 -3.46116215e-01 -2.25676879e-01
-7.64676332e-02 6.91317558e-01 -3.83650452e-01 7.31103837e-01
3.27801734e-01 7.59830654e-01 -1.01105940e+00 -2.30660930e-01
-3.43277723e-01 5.35093844e-01 -4.19369549e-01 7.02189887e-03
-7.54803598e-01 6.58485472e-01 -5.98661423e-01 5.98239541e-01
5.35968184e-01 -6.71683848e-01 8.56146455e-01 -1.06096365e-01
7.26739168e-02 1.88297648e-02 -1.48240292e+00 1.56924307e+00
-1.00675095e-02 4.50659513e-01 2.78839469e-01 -1.32317030e+00
5.64358890e-01 3.15183640e-01 6.74409747e-01 2.84755379e-01
1.37124345e-01 -1.77321717e-01 3.92217845e-01 -2.24947497e-01
-2.86574304e-01 -1.94435075e-01 -1.07013822e-01 4.57521051e-01
4.26254004e-01 -6.38882443e-02 4.11674261e-01 5.72647333e-01
1.87391198e+00 6.99403509e-02 1.38593256e-01 -5.98352194e-01
-9.65049118e-02 -1.56456336e-01 1.17210031e+00 8.95088375e-01
5.71557209e-02 3.38767022e-01 1.71963656e+00 -1.76877156e-01
-9.96139765e-01 -1.15579879e+00 -1.63334176e-01 6.83967769e-01
-3.76690656e-01 -6.19251251e-01 -3.99589717e-01 -3.11861992e-01
2.28911459e-01 1.88058347e-01 -9.08995092e-01 6.01674393e-02
-2.58696884e-01 -1.36405230e+00 5.58209658e-01 3.83685529e-01
-3.87618184e-01 -2.54440844e-01 2.04059273e-01 3.88155848e-01
9.66624394e-02 -9.48746741e-01 -5.28319478e-01 6.57928884e-01
-8.38617444e-01 -1.37036228e+00 -1.91720240e-02 -2.92039007e-01
6.72558963e-01 -3.09673339e-01 1.23687911e+00 -1.46949127e-01
-5.75984299e-01 4.71635133e-01 2.41611093e-01 -3.30753624e-01
-6.66868508e-01 -9.24886689e-02 6.01372495e-02 1.17748007e-02
1.19729370e-01 -1.09321058e+00 -3.67587268e-01 6.75871745e-02
-1.11520803e+00 -1.47400409e-01 4.84043419e-01 1.01244700e+00
8.18913281e-01 3.74089926e-01 4.85454679e-01 -1.16507471e+00
4.66411978e-01 -7.97775507e-01 -1.08191299e+00 2.74425089e-01
-5.74710608e-01 6.01327360e-01 5.19776583e-01 -4.32768136e-01
-9.95050251e-01 3.43371540e-01 1.33685201e-01 -3.25714320e-01
-1.08792171e-01 1.01136184e+00 -2.59857297e-01 2.95506150e-01
4.86836582e-01 6.44057319e-02 2.85699427e-01 -2.32821167e-01
4.14052248e-01 1.26648664e-01 2.96319872e-01 -9.11163449e-01
4.60957766e-01 6.57264292e-01 1.02811170e+00 -9.48174059e-01
-7.50601232e-01 -4.26395088e-01 -5.78744113e-01 -7.09235743e-02
7.62698293e-01 -8.32426906e-01 -1.15284240e+00 2.67518580e-01
-1.05872738e+00 -4.39152241e-01 1.09285176e-01 2.95398116e-01
-6.52886808e-01 2.70089209e-01 -8.52076769e-01 -9.62927997e-01
2.60385215e-01 -1.01982868e+00 1.07179379e+00 -1.30814731e-01
-4.47625697e-01 -1.35296178e+00 6.41631842e-01 -1.21745192e-01
1.77187473e-01 5.02552807e-01 1.31595135e+00 -4.31278020e-01
-9.94348109e-01 -1.38646523e-02 1.12771980e-01 -1.81445345e-01
1.50451623e-02 6.65533245e-01 -6.86733663e-01 -2.47580841e-01
-2.28366256e-01 -1.24755830e-01 1.09195316e+00 1.01900911e+00
9.32682216e-01 -3.07687998e-01 -7.67928362e-01 6.14976823e-01
1.32305181e+00 -1.00196056e-01 5.60673952e-01 -6.04754150e-01
7.05323339e-01 6.88046813e-01 -4.52988259e-02 4.72635388e-01
1.18177921e-01 4.31254119e-01 2.94449210e-01 4.05308247e-01
1.48928180e-01 -4.59578037e-01 2.45166212e-01 6.09064579e-01
1.88760936e-01 -4.47080731e-01 -1.09064054e+00 3.82311344e-01
-2.18496895e+00 -9.51930881e-01 -3.35514247e-01 2.01030326e+00
1.05755258e+00 1.50972486e-01 8.75487104e-02 -3.06674242e-01
6.31296635e-01 -1.02539845e-01 -6.86867714e-01 1.07183948e-01
-8.73707831e-02 9.19771641e-02 4.23022538e-01 7.61626184e-01
-8.27171922e-01 4.34609324e-01 6.91916180e+00 7.09093273e-01
-8.21294844e-01 -2.87318081e-01 9.99195337e-01 3.54491398e-02
-4.07429725e-01 4.99996364e-01 -1.14640403e+00 4.36193168e-01
1.38007057e+00 -1.57159656e-01 4.16930944e-01 3.01970720e-01
4.43797380e-01 -2.44370297e-01 -1.40322649e+00 6.08056664e-01
-4.91056830e-01 -1.49452174e+00 -1.59398407e-01 4.92841214e-01
8.41405869e-01 1.14684977e-01 -1.54062182e-01 1.74972378e-02
1.26247084e+00 -1.34034848e+00 2.62375101e-02 9.28543031e-01
5.51790535e-01 -5.53325057e-01 4.61159110e-01 5.05993664e-01
-9.32093978e-01 -4.83637154e-02 -3.44492674e-01 -4.06124406e-02
3.47176164e-01 1.27428877e+00 -8.38601410e-01 1.98989347e-01
2.57441640e-01 9.41193879e-01 2.06719395e-02 8.27947497e-01
-2.27709070e-01 1.19866920e+00 -6.70690417e-01 8.65757689e-02
-2.30513126e-01 -4.10667658e-01 5.99024653e-01 1.11625314e+00
4.19001848e-01 1.03512153e-01 1.79469541e-01 1.21123099e+00
-1.76959336e-01 -5.60953140e-01 -7.46994734e-01 -6.05916262e-01
4.65522051e-01 1.20100224e+00 -9.86898661e-01 -4.41617332e-02
-2.35939533e-01 3.15757513e-01 2.80942827e-01 5.06224453e-01
-3.48805368e-01 1.64175361e-01 1.02184927e+00 3.03532854e-02
3.48759085e-01 -3.77869010e-01 -1.26782849e-01 -1.09010613e+00
-6.17898166e-01 -7.07309365e-01 5.86083293e-01 -4.04533774e-01
-1.54457474e+00 -2.46215612e-01 -7.63498768e-02 -4.86887366e-01
-5.33361852e-01 -7.14051008e-01 -4.32882607e-01 7.73467779e-01
-9.51762140e-01 -8.91158879e-01 3.59732836e-01 2.29123101e-01
1.58673711e-02 3.95791978e-01 8.00897419e-01 -9.55800787e-02
-7.58552074e-01 -1.63318574e-01 3.61255109e-01 2.85074045e-03
4.52430367e-01 -1.53559947e+00 1.50294438e-01 7.50208378e-01
2.67573237e-01 9.07928348e-01 8.52065265e-01 -1.03491759e+00
-2.06631041e+00 -1.06047428e+00 6.10663295e-01 -7.82907426e-01
1.06989074e+00 -8.43535841e-01 -8.86018634e-01 8.30188692e-01
-4.23671424e-01 2.68399686e-01 7.54361391e-01 6.17234886e-01
-3.47497851e-01 1.42046630e-01 -5.03839433e-01 5.75870395e-01
1.16265213e+00 -3.52298558e-01 -2.02472787e-02 3.57754916e-01
5.57967961e-01 -1.14549629e-01 -1.08649838e+00 3.02862525e-01
5.61675191e-01 -4.53328073e-01 7.81687915e-01 -1.00853610e+00
3.46477896e-01 -6.11236036e-01 -2.57706586e-02 -1.50924456e+00
-4.54203367e-01 -9.49713588e-01 -4.37753737e-01 1.08910763e+00
5.67544937e-01 -1.91830799e-01 8.31535757e-01 6.15040839e-01
1.47210106e-01 -5.58590055e-01 -1.05057359e+00 -5.60812473e-01
5.76607846e-02 -4.53099906e-01 2.88797438e-01 9.41324532e-01
1.19047776e-01 8.22435856e-01 -3.71660978e-01 2.66525388e-01
7.44114339e-01 -7.94188008e-02 5.70768178e-01 -1.54172897e+00
-8.79281759e-01 -2.81884938e-01 -5.22603035e-01 -1.08714986e+00
3.39119554e-01 -8.04932117e-01 3.43741804e-01 -1.25125134e+00
6.11939847e-01 -1.82604536e-01 -1.07649110e-01 2.41502017e-01
-2.95375250e-02 -3.89619231e-01 -5.35594702e-01 5.65678999e-02
-5.06207108e-01 3.05655688e-01 9.93008733e-01 -1.59243673e-01
-8.20014328e-02 -7.86084607e-02 -8.64601612e-01 5.39786518e-01
3.28728378e-01 -5.16688585e-01 -6.08311832e-01 -8.00900012e-02
6.91646099e-01 5.45822978e-01 3.17272991e-01 -4.48215008e-01
5.61769783e-01 -3.75840336e-01 4.01599616e-01 -3.77883732e-01
1.29175648e-01 -4.61311370e-01 7.17032373e-01 3.56764466e-01
-3.65370393e-01 -2.08227918e-01 7.26202503e-02 1.32570159e+00
3.62884700e-02 3.73948328e-02 4.13648218e-01 -1.78668126e-01
-1.21430211e-01 5.25110960e-01 -9.39225435e-01 2.29639292e-01
4.70999539e-01 4.25943285e-01 -3.66181016e-01 -6.44135714e-01
-1.00847960e+00 3.33088756e-01 2.24551409e-01 -1.76519632e-01
1.90963924e-01 -8.52111280e-01 -6.32973135e-01 -5.20167351e-02
-7.79750794e-02 1.53107807e-01 2.50198126e-01 8.19473028e-01
-4.82852682e-02 4.69772488e-01 3.74484301e-01 -7.84176350e-01
-9.18065250e-01 4.02116120e-01 2.11219788e-01 -5.54048717e-01
-3.66578512e-02 8.58339489e-01 3.39792579e-01 -3.15308422e-01
-9.97859910e-02 -1.66772366e-01 -6.64293021e-02 5.45537397e-02
2.84404933e-01 2.29716480e-01 -2.71954119e-01 -1.53471574e-01
-9.47849005e-02 1.98109657e-01 -2.71185413e-02 -4.38894592e-02
1.40629685e+00 5.83959520e-02 -5.71511090e-01 7.62898743e-01
1.19093764e+00 -1.83649257e-01 -1.45656598e+00 -7.21341595e-02
1.87928170e-01 -1.45099401e-01 1.02312535e-01 -6.23374224e-01
-7.99790025e-01 6.70579433e-01 1.63743511e-01 2.38036841e-01
6.78858876e-01 4.07810718e-01 1.26185060e-01 3.07335883e-01
1.08917870e-01 -7.38035440e-01 -3.41019146e-02 6.76977336e-01
3.63841504e-01 -8.47524583e-01 2.49629524e-02 -6.21011674e-01
9.53155905e-02 1.14779520e+00 1.74718767e-01 -1.79874375e-02
1.08782494e+00 8.74871194e-01 -4.62295949e-01 -4.76688743e-01
-1.49560273e+00 5.48628420e-02 1.30864844e-01 5.29999077e-01
5.95548868e-01 1.15721663e-02 -8.92341211e-02 5.08137941e-01
1.45333052e-01 1.43620402e-01 6.19125605e-01 7.68269777e-01
-2.65210688e-01 -1.25395942e+00 -2.02292025e-01 8.53474259e-01
-5.60976684e-01 3.38490191e-03 -2.68022954e-01 5.46735227e-01
-3.06568414e-01 8.51276815e-01 1.23046689e-01 -1.15546286e-01
-1.17708407e-01 1.73791885e-01 5.16741991e-01 -8.09722900e-01
3.96070689e-01 4.47591394e-01 1.87212735e-01 -6.31207645e-01
-1.43253505e-01 -9.54432487e-01 -1.12475514e+00 -3.75603616e-01
-2.76845098e-01 1.67483613e-01 5.49663901e-01 1.05144465e+00
6.27782822e-01 6.77180350e-01 1.40713632e-01 -4.25690919e-01
-7.01101840e-01 -7.61564612e-01 -1.01174557e+00 6.82673603e-02
2.73356527e-01 -6.67383969e-01 -6.24570727e-01 2.97134399e-01] | [7.126335144042969, 4.880430221557617] |
ce805814-9f8c-4512-9521-80447a75f4c0 | data-efficient-visual-place-recognition-using | 2209.08343 | null | https://arxiv.org/abs/2209.08343v2 | https://arxiv.org/pdf/2209.08343v2.pdf | Data Efficient Visual Place Recognition Using Extremely JPEG-Compressed Images | Visual Place Recognition (VPR) is the ability of a robotic platform to correctly interpret visual stimuli from its on-board cameras in order to determine whether it is currently located in a previously visited place, despite different viewpoint, illumination and appearance changes. JPEG is a widely used image compression standard that is capable of significantly reducing the size of an image at the cost of image clarity. For applications where several robotic platforms are simultaneously deployed, the visual data gathered must be transmitted remotely between each robot. Hence, JPEG compression can be employed to drastically reduce the amount of data transmitted over a communication channel, as working with limited bandwidth for VPR can be proven to be a challenging task. However, the effects of JPEG compression on the performance of current VPR techniques have not been previously studied. For this reason, this paper presents an in-depth study of JPEG compression in VPR related scenarios. We use a selection of well-established VPR techniques on well-established benchmark datasets with various amounts of compression applied. We show that by introducing compression, the VPR performance is drastically reduced, especially in the higher spectrum of compression. Moreover, this paper demonstrates how fine-tuning a CNN can be utilised as an optimisation method for JPEG compressed data to perform more consistently with the image transformations detected in extremely JPEG compressed images. | ['Shoaib Ehsan', 'Klaus McDonald-Maier', 'Michael Milford', 'Bruno Ferrarini', 'Mihnea-Alexandru Tomita'] | 2022-09-17 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [ 6.65965438e-01 -8.93960744e-02 1.98034480e-01 9.12564062e-03
-2.45758548e-01 -5.76217413e-01 7.11268008e-01 3.35934132e-01
-8.55736434e-01 4.62720037e-01 -3.56659085e-01 -1.62617847e-01
-1.29738897e-01 -7.64277577e-01 -1.02182937e+00 -5.48713148e-01
-2.11734831e-01 1.39869317e-01 3.70852351e-01 -3.96998048e-01
4.96399850e-01 8.66624296e-01 -2.02573919e+00 -4.96691652e-02
1.68817878e-01 9.82684195e-01 7.99684942e-01 7.78122008e-01
3.87285680e-01 5.57425618e-01 -7.48195946e-01 -1.06021233e-01
6.37423217e-01 -5.26186544e-03 -5.18402934e-01 1.78257167e-01
2.66950279e-01 -3.30190659e-01 -4.90972400e-01 1.04615772e+00
1.41577318e-01 1.27929449e-01 4.44869697e-01 -1.22677314e+00
-1.26507819e-01 9.72648188e-02 -4.09553736e-01 1.15275271e-01
5.85798919e-01 2.17374295e-01 4.66489494e-01 -2.94031143e-01
8.99986923e-01 9.97603178e-01 2.20373079e-01 4.19546068e-02
-1.23100042e+00 -3.44511658e-01 -2.78991401e-01 4.62871164e-01
-1.25553763e+00 -3.02248418e-01 7.19966292e-01 -1.62629545e-01
1.10981667e+00 2.92791665e-01 7.55961537e-01 9.52938616e-01
3.86185259e-01 9.60558802e-02 8.92728567e-01 -4.70506072e-01
3.82490098e-01 -6.38547018e-02 -6.34499729e-01 3.71490806e-01
3.99378687e-01 1.36682272e-01 -6.53116465e-01 2.66075760e-01
7.19399631e-01 -1.34464294e-01 -4.67246830e-01 -4.85926837e-01
-1.33167183e+00 6.37042761e-01 7.55372465e-01 3.69309396e-01
-3.95159453e-01 3.66264880e-01 4.00534242e-01 5.39954126e-01
-2.61870045e-02 3.97780180e-01 -1.29720211e-01 -3.01814944e-01
-8.13325226e-01 1.68170303e-01 5.43353140e-01 9.27820444e-01
7.93159008e-01 -1.26202494e-01 3.30959797e-01 6.70711577e-01
3.98992300e-01 4.47591126e-01 6.28386497e-01 -1.30352902e+00
6.26010239e-01 6.11894071e-01 1.00281447e-01 -1.38139164e+00
-2.50969976e-01 3.11298706e-02 -7.89963484e-01 6.96312249e-01
2.85790890e-01 3.58236194e-01 -6.57178938e-01 1.27005661e+00
6.08004220e-02 -2.10649803e-01 3.76935363e-01 9.62849259e-01
3.07597399e-01 6.83081985e-01 -2.32732117e-01 6.52071312e-02
1.35535967e+00 -5.81770599e-01 -5.95177770e-01 -6.17900610e-01
2.69911498e-01 -8.75225782e-01 8.26769233e-01 6.29535139e-01
-8.15897703e-01 -4.36405897e-01 -1.54968190e+00 -2.14905307e-01
-6.57695651e-01 9.00426060e-02 2.84419805e-01 4.70400453e-01
-1.20572579e+00 7.05449641e-01 -7.84716606e-01 -5.93476176e-01
1.42340392e-01 5.17944276e-01 -9.10233498e-01 -3.76912266e-01
-8.96835864e-01 1.05009687e+00 6.10526383e-01 1.42970726e-01
-8.26803565e-01 -1.87334552e-01 -9.85666275e-01 1.19760849e-01
2.17315763e-01 -2.65528381e-01 9.75877523e-01 -7.13549316e-01
-1.40829086e+00 8.39971781e-01 2.11198613e-01 -9.50808465e-01
8.94892454e-01 7.79407993e-02 5.36820143e-02 4.70077038e-01
-2.21700951e-01 1.01886261e+00 1.03573000e+00 -1.08998120e+00
-5.09844363e-01 -4.32664961e-01 2.12217107e-01 4.98580456e-01
-4.35416661e-02 -3.04997444e-01 -4.65125144e-01 -2.62786776e-01
1.60879791e-01 -9.30596530e-01 -4.75966558e-02 3.42224002e-01
2.01363936e-02 3.66748244e-01 1.09418404e+00 -4.60903496e-01
3.15027207e-01 -2.34167957e+00 8.44356790e-02 1.18947797e-01
-4.33941968e-02 2.18867242e-01 2.91676745e-02 5.62955320e-01
7.26393685e-02 -1.62571535e-01 -2.46937171e-01 -2.60964066e-01
-1.50929689e-01 5.84757447e-01 -1.69752657e-01 8.55105460e-01
-6.22785352e-02 4.20240045e-01 -5.89893341e-01 -2.41401806e-01
8.71234119e-01 7.72297263e-01 -3.96080732e-01 1.54197961e-01
2.51261815e-02 3.93843383e-01 2.15007082e-01 1.32097289e-01
7.02900290e-01 1.59621954e-01 1.39823601e-01 -1.54572085e-01
-2.33850121e-01 -1.12253428e-01 -9.88580465e-01 1.75696123e+00
-6.69522524e-01 1.28307343e+00 1.22937061e-01 -1.17110372e+00
1.05086052e+00 1.16748542e-01 2.39938810e-01 -1.10153031e+00
3.04838717e-01 3.12849849e-01 -7.37474710e-02 -4.87360537e-01
8.93756688e-01 2.19922930e-01 4.83093262e-02 -1.89653352e-01
-9.00268853e-02 -2.30474323e-01 4.56416011e-01 -7.45643824e-02
1.18677747e+00 -5.71337342e-02 3.10679942e-01 5.80197461e-02
4.77093190e-01 9.10617486e-02 -1.07333310e-01 5.22287607e-01
-1.70340493e-01 6.72472417e-01 3.01091373e-01 -3.16375166e-01
-1.22368467e+00 -7.15974808e-01 2.51164176e-02 6.56309426e-01
6.20479643e-01 -1.40369460e-01 -6.76407814e-01 -8.94935504e-02
-2.28183016e-01 6.88067377e-01 -5.19145250e-01 -1.98206142e-01
-4.76218998e-01 -2.98080325e-01 4.68000591e-01 -4.92002396e-03
9.50756073e-01 -1.14445817e+00 -1.70882177e+00 4.25528549e-02
-2.67955512e-01 -1.47621226e+00 2.00124979e-01 3.91666532e-01
-8.42525482e-01 -1.18023610e+00 -7.51204610e-01 -5.49565017e-01
6.74748778e-01 7.12120116e-01 7.46796846e-01 6.76083192e-02
-3.12246680e-01 5.85316658e-01 -6.75560296e-01 -1.03558399e-01
-5.74069381e-01 -2.28505909e-01 -2.76374996e-01 -1.59366965e-01
-8.06296691e-02 -4.86426771e-01 -6.07488036e-01 3.04839849e-01
-1.43416893e+00 -2.59221885e-02 5.78161061e-01 4.26894516e-01
3.90668958e-01 3.29878747e-01 -3.16707075e-01 -2.72146761e-01
4.38696235e-01 -5.07498607e-02 -8.24043751e-01 5.73002100e-02
-2.58528680e-01 1.37490153e-01 6.16556525e-01 -3.10043782e-01
-5.23094952e-01 2.35756978e-01 5.25964908e-02 -4.91352856e-01
-4.22306508e-01 3.38629872e-01 1.31906554e-01 -4.78304356e-01
5.42142570e-01 4.87337172e-01 2.07593724e-01 -7.98339397e-02
2.32916966e-01 6.48207068e-01 6.73756123e-01 4.97806109e-02
6.61114037e-01 7.06055939e-01 3.08623701e-01 -1.18717504e+00
2.17554688e-01 -4.74311203e-01 -3.98038894e-01 -4.87630755e-01
7.68636465e-01 -6.10010028e-01 -7.41568387e-01 3.17968637e-01
-1.20284092e+00 -2.63716429e-01 -2.57276148e-02 4.16068047e-01
-7.77574122e-01 5.40429413e-01 -3.90954651e-02 -6.98715329e-01
7.48210177e-02 -1.51028180e+00 1.02995670e+00 5.11811525e-02
-3.67729180e-02 -7.37550259e-01 -6.41132966e-02 3.08478326e-01
3.83208871e-01 2.30141595e-01 5.27432740e-01 -1.82421342e-01
-6.23521090e-01 -3.06717336e-01 -2.24069819e-01 1.77086353e-01
-8.08206797e-02 -4.61510941e-02 -8.34502041e-01 -4.60252613e-01
3.13586444e-02 -3.17251086e-02 7.01394081e-01 8.23895037e-02
9.64627028e-01 -2.51900047e-01 -1.95375294e-03 5.11936605e-01
1.80160940e+00 2.52514988e-01 1.02603388e+00 9.67427373e-01
2.17430413e-01 5.55992961e-01 8.25217307e-01 3.26784909e-01
3.81716043e-02 9.50553238e-01 1.29794872e+00 1.70410469e-01
4.13577110e-02 -2.16893163e-02 4.21052933e-01 1.05861835e-01
-2.64215827e-01 -4.89299119e-01 -7.77127802e-01 2.51470268e-01
-1.49353182e+00 -7.00793743e-01 -3.88587150e-03 2.45459032e+00
1.48235500e-01 9.80975330e-02 -2.86675304e-01 5.89166522e-01
6.61008537e-01 2.13509932e-01 -2.27585196e-01 -4.91806149e-01
8.03389959e-03 -1.70979738e-01 1.17698193e+00 2.45797947e-01
-9.10244644e-01 4.44937319e-01 5.49730921e+00 4.99928027e-01
-1.41659081e+00 -9.92175415e-02 1.14080921e-01 -1.44317970e-01
2.96661615e-01 -2.00684473e-01 -2.04668820e-01 5.37169158e-01
9.51067805e-01 1.05964750e-01 8.04828346e-01 9.18416142e-01
2.38857239e-01 -6.76740885e-01 -1.09024239e+00 1.57209182e+00
3.35359722e-01 -1.05028546e+00 -9.02211443e-02 3.71243775e-01
2.66232163e-01 2.79853076e-01 2.77697831e-01 -1.67266548e-01
-1.69678077e-01 -9.22658622e-01 9.30184126e-01 1.09696813e-01
5.81442237e-01 -7.22476065e-01 7.87261546e-01 2.88545728e-01
-9.75848019e-01 -2.66801119e-01 -6.36648834e-01 2.82540568e-04
1.49495542e-01 1.79380700e-01 -8.87700319e-01 6.31706595e-01
8.65873396e-01 5.93905807e-01 -7.45697260e-01 1.15634537e+00
-1.40725434e-01 -1.45924479e-01 -5.40365160e-01 1.60950497e-01
3.48601550e-01 -4.94959345e-03 6.10176027e-01 8.40386331e-01
7.59362698e-01 -3.41617584e-01 -4.23396438e-01 4.22740251e-01
-1.83674577e-03 -6.83337301e-02 -1.14432108e+00 2.28530243e-01
2.40938172e-01 9.97904301e-01 -1.02269101e+00 -2.01391429e-02
-1.29623592e-01 1.30232990e+00 1.37534156e-01 1.09672166e-01
-7.12331414e-01 -2.07682893e-01 3.58799636e-01 9.07311216e-02
5.68711579e-01 -5.78446031e-01 1.46472752e-01 -7.84377873e-01
1.01382963e-01 -6.91242635e-01 1.12265863e-01 -1.20994604e+00
-4.01431203e-01 4.69365299e-01 4.87367027e-02 -1.57607305e+00
-2.34159485e-01 -8.33039165e-01 -2.05000177e-01 3.19632292e-01
-1.75384820e+00 -6.46368682e-01 -5.98350585e-01 5.08971035e-01
6.79752171e-01 -4.24041785e-02 7.33738720e-01 1.70799658e-01
-5.21598756e-02 9.79291350e-02 2.31458768e-01 -1.31184638e-01
5.57483137e-01 -9.42554593e-01 -4.01971191e-02 8.71764779e-01
3.17478567e-01 3.95900816e-01 1.03612065e+00 -3.95228654e-01
-1.64393961e+00 -8.96313488e-01 3.85105342e-01 1.52681425e-01
2.49920562e-01 -2.10569978e-01 -6.55388355e-01 4.35881287e-01
4.99824107e-01 -7.35359713e-02 2.38638312e-01 -6.75031781e-01
-2.22004116e-01 -2.30668053e-01 -1.36830366e+00 6.33708715e-01
4.70428765e-01 -3.34144622e-01 -4.63709891e-01 1.31983519e-01
3.26219976e-01 -5.28178811e-01 -6.48550570e-01 1.72362670e-01
3.88846785e-01 -1.26588726e+00 1.08681476e+00 3.34519833e-01
6.48358881e-01 -3.39238316e-01 -4.13457721e-01 -1.24118817e+00
2.09352270e-01 -2.64319927e-01 2.35519066e-01 7.38596439e-01
2.92461738e-02 -5.94511807e-01 4.87034380e-01 4.26768363e-01
1.89103365e-01 -1.43667668e-01 -1.37240529e+00 -8.20567429e-01
-5.94810188e-01 -5.51898956e-01 2.28550911e-01 5.11693537e-01
-8.09422731e-02 -2.03919649e-01 -3.30608279e-01 3.46050471e-01
5.44592559e-01 -2.89188087e-01 8.40716898e-01 -8.56527209e-01
-1.91419765e-01 -2.92191029e-01 -1.24736428e+00 -7.39088476e-01
-2.74824142e-01 -6.58273876e-01 2.30527014e-01 -1.40870750e+00
-2.79376209e-01 -8.52313861e-02 -2.38289982e-02 3.13584059e-01
5.44119716e-01 6.27969503e-01 5.81589699e-01 4.27570701e-01
-4.89397168e-01 4.09435302e-01 9.51456726e-01 -2.88696527e-01
4.05735932e-02 -3.21087092e-01 -6.47077849e-03 5.67159534e-01
8.95535231e-01 -5.01380980e-01 -5.16525745e-01 -4.77228910e-01
4.71786529e-01 9.24124941e-02 6.59050882e-01 -1.47737312e+00
3.86290401e-01 2.79234111e-01 2.73843855e-01 -4.71193939e-01
6.87407434e-01 -1.49863315e+00 4.50486153e-01 8.17237198e-01
-3.31247509e-01 2.89729416e-01 2.34327242e-01 7.19373405e-01
-2.72854358e-01 -5.17360747e-01 7.18029022e-01 -2.22076446e-01
-1.02992499e+00 -3.55248362e-01 -6.77302361e-01 -4.20478642e-01
1.30697501e+00 -5.12051165e-01 -1.66192502e-01 -3.63315254e-01
-3.18731576e-01 -1.93904132e-01 8.53702068e-01 5.19724131e-01
7.62680173e-01 -9.19564247e-01 -1.64862946e-01 3.30231249e-01
1.72969967e-01 1.20141260e-01 3.60783219e-01 6.61117554e-01
-1.31620097e+00 5.23158133e-01 -5.92898786e-01 -8.05896699e-01
-1.54187739e+00 5.23723662e-01 9.89824533e-02 2.18322538e-02
-7.53669679e-01 3.02116871e-01 -2.12036118e-01 -7.39145800e-02
3.19508702e-01 -4.62141156e-01 -3.90687943e-01 -7.73692802e-02
4.23952878e-01 2.87981957e-01 3.03416252e-01 -6.89969838e-01
-2.80296952e-01 5.47616303e-01 1.48944274e-01 -2.28004426e-01
1.36828339e+00 -3.98844391e-01 -7.49687925e-02 2.48085812e-01
1.33216059e+00 -3.70755553e-01 -1.45336533e+00 4.05348241e-01
-3.26718688e-01 -7.29845285e-01 1.64863631e-01 -3.83510470e-01
-1.21358132e+00 8.04111719e-01 8.69308054e-01 3.48911822e-01
1.19867885e+00 -2.03680053e-01 2.46150792e-01 6.11477554e-01
8.61419797e-01 -1.14105630e+00 -1.00320570e-01 2.58562684e-01
1.05886400e+00 -1.15756667e+00 2.79845055e-02 -3.39063406e-01
-4.61237013e-01 1.29791343e+00 4.17885669e-02 -3.01188231e-01
1.34810925e-01 1.70515656e-01 -4.61334363e-02 -5.80436625e-02
-3.36875767e-01 6.32936805e-02 -2.51102060e-01 8.82289290e-01
-2.96722710e-01 -1.95229873e-01 -8.79371688e-02 -4.14269775e-01
-4.96050894e-01 -7.36530125e-02 8.37985218e-01 1.10276830e+00
-3.21185797e-01 -8.69578719e-01 -5.97776711e-01 2.50901226e-02
-2.97736973e-01 2.32740909e-01 -2.75862396e-01 8.79489362e-01
3.63358222e-02 8.99771631e-01 5.57624958e-02 -2.59597570e-01
2.98584580e-01 -3.53770912e-01 5.60485542e-01 -1.58559665e-01
-3.54612142e-01 -3.45524788e-01 -1.03169739e-01 -8.53764415e-01
-7.66067386e-01 -5.72140038e-01 -8.85372579e-01 -2.43019700e-01
1.20644011e-01 -2.19346344e-01 1.32582259e+00 7.78233409e-01
2.16301039e-01 6.33447886e-01 3.54497492e-01 -1.36269796e+00
-2.21007511e-01 -6.09996915e-01 -4.43507373e-01 4.42619324e-01
2.31719628e-01 -7.74303854e-01 -4.99868482e-01 2.46798337e-01] | [7.559793949127197, -1.8618576526641846] |
558c739e-c014-4954-9adb-f8f7af365941 | using-language-models-to-improve-rule-based | null | null | https://aclanthology.org/2022.latechclfl-1.5 | https://aclanthology.org/2022.latechclfl-1.5.pdf | Using Language Models to Improve Rule-based Linguistic Annotation of Modern Historical Japanese Corpora | Annotation of unlabeled textual corpora with linguistic metadata is a fundamental technology in many scholarly workflows in the digital humanities (DH). Pretrained natural language processing pipelines offer tokenization, tagging, and dependency parsing of raw text simultaneously using an annotation scheme like Universal Dependencies (UD). However, the accuracy of these UD tools remains unknown for historical texts and current methods lack mechanisms that enable helpful evaluations by domain experts. To address both points for the case of Modern Historical Japanese text, this paper proposes the use of unsupervised domain adaptation methods to develop a domain-adapted language model (LM) that can flag instances of inaccurate UD output from a pretrained LM and the use of these instances to form rules that, when applied, improves pretrained annotation accuracy. To test the efficacy of the proposed approach, the paper evaluates the domain-adapted LM against three baselines that are not adapted to the historical domain. The experiments conducted demonstrate that the domain-adapted LM improves UD annotation in the Modern Historical Japanese domain and that rules produced using this LM are best indicative of characteristics of the domain in terms of out-of-vocabulary rate and candidate normalized form discovery for “difficult” bigram terms. | ['Mitsunori Ogihara', 'Jerry Bonnell'] | null | null | null | null | latechclfl-coling-2022-10 | ['dependency-parsing'] | ['natural-language-processing'] | [-5.31664938e-02 1.93073913e-01 -5.32942593e-01 -4.56578940e-01
-1.18802440e+00 -9.01364863e-01 8.51572096e-01 3.75934154e-01
-7.82673776e-01 8.75195742e-01 6.31716192e-01 -5.32267332e-01
-2.43534632e-02 -2.81095684e-01 -4.14747417e-01 -1.93292141e-01
3.28680664e-01 8.19090366e-01 3.21357906e-01 -2.20128596e-01
3.68232459e-01 4.43004876e-01 -8.55221510e-01 3.52084368e-01
9.56768751e-01 3.07583421e-01 3.95893335e-01 4.84783322e-01
-4.86780494e-01 7.79981971e-01 -9.16583538e-01 -5.59571743e-01
1.49402291e-01 -1.89063162e-01 -1.36819589e+00 -2.85383731e-01
5.21694243e-01 -2.22099945e-01 -1.18034504e-01 9.23102736e-01
4.81842995e-01 1.93694711e-01 5.49297810e-01 -4.95290428e-01
-6.53986156e-01 1.25717509e+00 3.51226120e-03 7.05016136e-01
4.41822439e-01 -1.50613636e-01 1.06524551e+00 -7.91008055e-01
1.22248316e+00 1.05113912e+00 6.28817618e-01 4.20730859e-01
-1.16219532e+00 -6.39438272e-01 1.91312917e-02 -4.32304898e-03
-1.33410692e+00 -7.29100525e-01 3.97893935e-01 -5.37203848e-01
1.35106349e+00 -3.93899716e-02 3.27232517e-02 9.89351809e-01
1.25380120e-05 6.36887729e-01 8.39837074e-01 -9.44290519e-01
1.12810411e-01 4.50244755e-01 4.42719489e-01 3.42368901e-01
2.05427215e-01 -5.30566156e-01 -6.40843511e-01 -1.67981267e-01
5.87587357e-01 -7.74568081e-01 1.13316067e-02 4.78700101e-01
-9.39864755e-01 6.34634316e-01 -2.49231264e-01 8.75777006e-01
-2.00326711e-01 -3.50015491e-01 7.78038919e-01 1.85309052e-01
7.59925008e-01 8.29583228e-01 -7.18607128e-01 -5.22803485e-01
-1.27897918e+00 1.28723189e-01 9.93605912e-01 1.14445460e+00
5.47788978e-01 -6.72527030e-02 -8.94747674e-02 1.04010999e+00
9.73726138e-02 2.33962357e-01 7.80348837e-01 -8.81907523e-01
5.64602792e-01 7.15289474e-01 1.03725739e-01 -4.25138146e-01
-3.20042253e-01 -1.12847276e-02 5.65017760e-02 -5.51539063e-01
6.61406159e-01 -1.66479528e-01 -9.70696509e-01 1.55861795e+00
3.39609355e-01 -1.63277686e-01 1.40734270e-01 4.61763948e-01
1.06260729e+00 5.68090856e-01 6.91247046e-01 -3.48856539e-01
1.25551975e+00 -4.50964779e-01 -9.86286223e-01 -3.85907888e-01
1.22988427e+00 -1.16719282e+00 1.21322918e+00 1.25890508e-01
-9.66672361e-01 -5.31526566e-01 -8.26620400e-01 -5.09867251e-01
-4.39771563e-01 2.15438113e-01 3.58518839e-01 6.65861011e-01
-6.96737826e-01 4.78412598e-01 -9.68575239e-01 -8.01377296e-01
-4.27705795e-02 2.10840181e-01 -2.84433454e-01 -1.22153133e-01
-1.33143520e+00 1.22847688e+00 6.72886193e-01 -3.70344400e-01
-5.36281347e-01 -7.41695881e-01 -9.47554886e-01 -6.07852563e-02
1.81410685e-01 1.31828442e-01 1.52657342e+00 -9.58383501e-01
-1.14906013e+00 1.20412946e+00 -1.95223704e-01 -3.59120756e-01
2.94328392e-01 -6.19459629e-01 -5.81762433e-01 1.09402798e-01
5.53353548e-01 1.69048533e-01 3.02751094e-01 -1.04446554e+00
-7.25639582e-01 -2.91351825e-01 -1.48858875e-01 1.49118707e-01
-6.70500159e-01 5.77815533e-01 -5.33961475e-01 -6.92315757e-01
-2.38655001e-01 -8.55660141e-01 -2.73348019e-02 -6.54871941e-01
-1.59123659e-01 -6.54302418e-01 7.95473814e-01 -1.16148114e+00
1.80086255e+00 -2.02512527e+00 -4.80888307e-01 2.15124249e-01
-3.12650859e-01 5.06164789e-01 1.75139576e-01 4.35309052e-01
2.45045409e-01 4.89709556e-01 -1.00363865e-01 -3.36585045e-01
-1.15551502e-01 7.44089425e-01 -1.87987208e-01 3.11370730e-01
2.49786705e-01 4.72130477e-01 -9.78132427e-01 -9.86496627e-01
-4.44467962e-02 1.99740455e-01 -3.33497643e-01 2.74027348e-01
-3.26292723e-01 4.82526094e-01 -5.15688241e-01 6.03817880e-01
2.26006404e-01 -6.76522031e-03 7.57558465e-01 2.01690886e-02
-5.85079789e-01 9.03468490e-01 -1.04451489e+00 1.76230371e+00
-4.05298442e-01 7.62029886e-01 -2.31827587e-01 -8.21939170e-01
9.24659967e-01 5.81468761e-01 1.90898046e-01 -6.11476183e-01
-2.02244855e-02 6.20867550e-01 7.90394843e-02 -8.63078892e-01
9.17987108e-01 -6.20140843e-02 -3.29861343e-01 3.75414431e-01
5.66792846e-01 -4.55544405e-02 5.01098871e-01 4.81392354e-01
1.17987323e+00 6.11139655e-01 4.78106648e-01 -3.81182939e-01
5.14468551e-01 5.40444791e-01 7.66757846e-01 7.43491232e-01
6.71700537e-02 3.18794668e-01 2.09138915e-01 -2.93933839e-01
-1.34607816e+00 -4.38653350e-01 -5.60988367e-01 1.59290934e+00
-2.78047651e-01 -5.60647488e-01 -5.24631739e-01 -7.59653866e-01
-1.52878910e-01 1.13533711e+00 -3.13339412e-01 1.94226161e-01
-1.08587146e+00 -6.42231941e-01 1.01861548e+00 4.59209383e-01
2.60664493e-01 -1.13097680e+00 -3.84875357e-01 5.58163464e-01
-3.52267176e-01 -1.67630661e+00 -2.09665030e-01 3.40952903e-01
-5.57590783e-01 -7.82323003e-01 -3.41313571e-01 -9.67439651e-01
3.49086076e-01 -4.55888659e-01 1.19160497e+00 1.80660449e-02
6.65890798e-02 2.81065673e-01 -7.89650321e-01 -3.89657408e-01
-9.03338969e-01 4.52457905e-01 7.85869285e-02 -5.01785457e-01
9.37516689e-01 -2.61796772e-01 -7.86725879e-02 -4.15385738e-02
-8.63007486e-01 -5.15559435e-01 5.47370970e-01 6.66605353e-01
4.53749061e-01 -3.61498147e-01 5.54558456e-01 -1.39596009e+00
6.26469135e-01 -6.78735614e-01 -3.61053288e-01 5.05992770e-01
-7.02756464e-01 1.66850373e-01 6.62542701e-01 -6.56603277e-01
-1.68884039e+00 7.37459362e-02 -2.98086196e-01 3.63393724e-02
-1.86358571e-01 8.64068031e-01 -2.28484586e-01 1.71709269e-01
9.04762506e-01 -7.46794045e-02 -5.67741096e-01 -8.94491613e-01
2.24369660e-01 1.02118444e+00 9.05901730e-01 -1.21770227e+00
4.55785215e-01 -5.63260019e-02 -4.58193094e-01 -8.95032465e-01
-1.09265029e+00 -8.60306084e-01 -1.03047609e+00 -1.10865675e-01
7.73492575e-01 -8.59564245e-01 2.95135509e-02 -9.57490206e-02
-1.20839167e+00 -4.41275150e-01 -2.83830434e-01 5.25728047e-01
1.76615678e-02 5.30747950e-01 -8.38538110e-01 -7.10437834e-01
-3.31463188e-01 -8.68195593e-01 7.51802027e-01 3.52474511e-01
-6.84718549e-01 -1.28358674e+00 2.03919426e-01 2.92135596e-01
3.15247253e-02 7.75162131e-02 1.20050359e+00 -1.46936226e+00
1.97492197e-01 -3.74694675e-01 1.60805628e-01 3.67829114e-01
-9.32416543e-02 3.30285788e-01 -9.59147334e-01 1.41412795e-01
-2.46579140e-01 -3.76690239e-01 1.22385569e-01 1.19997278e-01
6.05306447e-01 -3.21340859e-01 -2.42337167e-01 2.11732939e-01
1.33531308e+00 1.58902749e-01 4.36376482e-01 8.69260132e-01
5.03600717e-01 4.47012901e-01 6.90203726e-01 4.94118899e-01
4.49871033e-01 4.69211489e-01 -3.23961616e-01 1.73691988e-01
-1.32253274e-01 -2.49079943e-01 4.11916018e-01 9.74930584e-01
-1.04398742e-01 -2.11512461e-01 -1.43326759e+00 1.06630230e+00
-1.57125366e+00 -7.60034561e-01 -3.35710078e-01 2.05803204e+00
1.47832274e+00 5.85173428e-01 -1.05967335e-01 -1.49264798e-01
6.82603359e-01 -3.24026287e-01 -1.31899923e-01 -7.90611267e-01
-2.79456317e-01 6.52575374e-01 6.99601233e-01 3.97375673e-01
-8.36537063e-01 1.30019748e+00 6.28626060e+00 7.22676218e-01
-9.31063116e-01 3.04987937e-01 3.36689651e-01 2.62286454e-01
-2.32994944e-01 4.28837627e-01 -1.26722956e+00 3.10866445e-01
1.51323903e+00 -2.71440983e-01 -9.91428867e-02 9.77644265e-01
2.49832541e-01 -9.78364274e-02 -1.20634425e+00 3.23327452e-01
1.38874119e-03 -1.38289618e+00 1.57447532e-01 -1.44541040e-02
6.25473380e-01 1.04899570e-01 -3.58499587e-01 6.79557323e-01
7.25547194e-01 -7.30428874e-01 7.66079485e-01 2.18157843e-01
8.49150956e-01 -4.80623007e-01 1.02915621e+00 3.26293707e-01
-7.84675777e-01 1.90559357e-01 -3.69150519e-01 4.48556095e-02
2.69065380e-01 3.55176866e-01 -1.37912357e+00 5.34121931e-01
5.84263444e-01 3.74238580e-01 -6.38761103e-01 7.31207073e-01
-3.49907398e-01 1.21062744e+00 -5.04336715e-01 1.95267305e-01
3.08650672e-01 2.16458231e-01 5.20564139e-01 1.94714570e+00
7.25092515e-02 2.96688199e-01 4.14656013e-01 4.42355692e-01
-3.36842924e-01 5.21417916e-01 -3.00207436e-01 -3.44668090e-01
9.61789072e-01 1.18985939e+00 -7.10201979e-01 -6.14931226e-01
-6.58066571e-01 7.19892085e-01 4.10237938e-01 1.01080358e-01
-3.62465292e-01 -1.89136222e-01 3.31969649e-01 2.56083280e-01
1.16057649e-01 -3.72277260e-01 -5.60328543e-01 -9.79572117e-01
-1.93759695e-01 -9.28763151e-01 8.57505977e-01 -4.23209131e-01
-1.31914878e+00 5.18068492e-01 1.39040977e-01 -9.14354444e-01
-3.21322352e-01 -5.74267864e-01 -3.78630221e-01 8.69692028e-01
-1.33721936e+00 -1.42602873e+00 6.11681193e-02 4.85840648e-01
7.89186001e-01 -4.50519696e-02 1.05402064e+00 5.42353034e-01
-5.56639731e-01 7.91053295e-01 2.35617399e-01 6.13995194e-01
1.26911378e+00 -1.34457791e+00 1.76518813e-01 1.18005991e+00
3.93867679e-02 9.64153647e-01 8.32679152e-01 -1.05219793e+00
-9.47273612e-01 -8.44017327e-01 1.43939829e+00 -6.82489455e-01
8.73681545e-01 -6.54558390e-02 -1.35220897e+00 9.87965107e-01
1.70170128e-01 -1.42969906e-01 1.14484835e+00 3.56808841e-01
-2.16911241e-01 3.26666206e-01 -1.15646100e+00 1.98830932e-01
6.26515806e-01 -7.08908379e-01 -1.17909431e+00 4.84321415e-01
3.89777482e-01 -4.93519753e-01 -1.51374257e+00 8.11921135e-02
4.82850701e-01 -7.54046366e-02 4.60980654e-01 -7.62422502e-01
3.66543770e-01 -6.19810186e-02 -1.31514236e-01 -8.06896806e-01
-1.16103902e-01 -6.24036372e-01 1.73103184e-01 1.89956403e+00
6.82699800e-01 -1.12530939e-01 3.20606083e-01 1.21586967e+00
-5.37811697e-01 1.51745036e-01 -9.88491952e-01 -8.10965717e-01
4.13985759e-01 -4.18227643e-01 2.50937283e-01 1.50315404e+00
3.95537823e-01 5.53313017e-01 1.43102789e-02 7.47884959e-02
2.71176726e-01 -3.95138085e-01 5.69393635e-01 -1.31009531e+00
5.45718260e-02 -3.93146947e-02 -1.48328900e-01 -8.04549932e-01
4.07772660e-01 -9.33322430e-01 1.62552878e-01 -1.42595804e+00
-4.70667295e-02 -7.56301045e-01 -3.73978727e-02 6.48127019e-01
-3.41822833e-01 2.15233341e-02 -9.74284410e-02 6.88369989e-01
-4.43064898e-01 -5.65643311e-02 6.68496192e-01 3.04631531e-01
-4.90466475e-01 -4.60065693e-01 -4.62243080e-01 8.34216714e-01
4.92816091e-01 -8.31455946e-01 1.44488007e-01 -6.17893696e-01
7.01359585e-02 -1.20630041e-01 -3.88512671e-01 -7.28039384e-01
3.69803190e-01 -2.32149050e-01 3.72272313e-01 -2.04132661e-01
-2.99628645e-01 -5.17864347e-01 -1.61354348e-01 6.51799738e-02
-4.31180149e-01 8.16578418e-02 3.47113013e-01 -7.74287954e-02
-9.74504501e-02 -7.03122675e-01 7.00195432e-01 -3.54537457e-01
-1.17063558e+00 -2.95882165e-01 -5.93527138e-01 5.22308230e-01
7.70187378e-01 -1.80454761e-01 -2.95823097e-01 2.76344791e-02
-6.32503986e-01 4.59974483e-02 3.37403238e-01 1.44045845e-01
-7.76609406e-02 -8.25506628e-01 -8.56413662e-01 -2.49360397e-01
1.64128214e-01 1.15937158e-01 -3.40853393e-01 4.44529235e-01
-6.16108358e-01 4.37988728e-01 -1.24309860e-01 -2.23194599e-01
-1.17382705e+00 1.33321539e-01 7.25041330e-03 -6.90184414e-01
-4.30647701e-01 8.86159122e-01 -4.35501754e-01 -4.13260490e-01
7.27927908e-02 -1.71022549e-01 -4.33842391e-01 2.08904073e-01
4.41205442e-01 2.13878304e-01 4.31657016e-01 -8.85143816e-01
-3.57711524e-01 -1.21573731e-01 -4.22152013e-01 -2.93323040e-01
1.49393559e+00 -2.33950630e-01 -2.62206383e-02 5.25569975e-01
7.87272513e-01 3.24337929e-01 -5.89635134e-01 -6.13780081e-01
8.81763995e-01 -2.96551257e-01 5.88714257e-02 -9.03309047e-01
-2.88803428e-01 4.61342961e-01 3.45610976e-01 3.50598581e-02
7.03971028e-01 1.49093375e-01 7.03270793e-01 3.79635096e-01
-1.53297246e-01 -1.73836410e+00 -2.48100325e-01 8.74504328e-01
3.66505593e-01 -1.18091261e+00 2.04214156e-01 -1.57903641e-01
-7.88880348e-01 1.33930850e+00 8.87110353e-01 2.95341998e-01
4.87972111e-01 4.74408448e-01 2.71107376e-01 -1.97697520e-01
-3.61788750e-01 -3.04858744e-01 2.11343750e-01 5.22351205e-01
9.60896492e-01 -1.15064539e-01 -9.62254226e-01 7.16963410e-01
-2.96718061e-01 -3.81931402e-02 5.78875184e-01 1.27438295e+00
-3.98906469e-01 -1.31256318e+00 -2.69243509e-01 3.35913509e-01
-1.23292387e+00 -4.35582876e-01 -5.09791255e-01 9.66181159e-01
2.08210409e-01 7.91780114e-01 8.07899237e-02 -1.73092857e-02
2.15613216e-01 6.56341076e-01 1.43561915e-01 -1.11574101e+00
-1.03129077e+00 2.78480828e-01 7.41367459e-01 4.20858115e-02
-7.11153328e-01 -8.30464900e-01 -1.52382600e+00 -1.35821715e-01
-5.40350020e-01 4.16606247e-01 6.13166630e-01 1.41142845e+00
-1.91987548e-02 4.12624516e-03 -8.40335488e-02 -3.68340224e-01
-3.60718787e-01 -1.38274801e+00 -2.49807656e-01 7.34823644e-01
-1.40524194e-01 -3.13557565e-01 -9.40698385e-02 6.36155069e-01] | [10.085917472839355, 9.748867988586426] |
2f67a72a-45e2-4c78-9889-2072b026ef9d | inference-in-predictive-quantile-regressions | 2306.00296 | null | https://arxiv.org/abs/2306.00296v1 | https://arxiv.org/pdf/2306.00296v1.pdf | Inference in Predictive Quantile Regressions | This paper studies inference in predictive quantile regressions when the predictive regressor has a near-unit root. We derive asymptotic distributions for the quantile regression estimator and its heteroskedasticity and autocorrelation consistent (HAC) t-statistic in terms of functionals of Ornstein-Uhlenbeck processes. We then propose a switching-fully modified (FM) predictive test for quantile predictability with persistent regressors. The proposed test employs an FM style correction with a Bonferroni bound for the local-to-unity parameter when the predictor has a near unit root. It switches to a standard predictive quantile regression test with a slightly conservative critical value when the largest root of the predictor lies in the stationary range. Simulations indicate that the test has reliable size in small samples and particularly good power when the predictor is persistent and endogenous, i.e., when the predictive regression problem is most acute. We employ this new methodology to test the ability of three commonly employed, highly persistent and endogenous lagged valuation regressors - the dividend price ratio, earnings price ratio, and book to market ratio - to predict the median, shoulders, and tails of the stock return distribution. | ['Nina Kuriyama', 'Katsumi Shimotsu', 'Alex Maynard'] | 2023-06-01 | null | null | null | null | ['unity'] | ['computer-vision'] | [-2.32170895e-01 -3.24908912e-01 -2.30490267e-01 -2.68142462e-01
-7.77275801e-01 -6.05551898e-01 6.38553560e-01 -1.77729815e-01
-2.41101071e-01 1.32427394e+00 1.33974299e-01 -7.94858038e-01
-5.47590971e-01 -8.79905462e-01 -5.08500278e-01 -8.41681421e-01
-2.08953917e-01 4.08266544e-01 2.85554416e-02 1.16338173e-03
5.74467540e-01 3.94818097e-01 -1.42064202e+00 -5.06047368e-01
9.05406117e-01 1.32454956e+00 -2.74441838e-01 7.74728656e-01
2.67260075e-01 9.67134178e-01 -4.48638380e-01 -2.89915919e-01
4.35396075e-01 -3.91675413e-01 -2.09151357e-01 -4.75215256e-01
-6.45364523e-02 -2.44149134e-01 2.48306438e-01 8.74732912e-01
2.78016537e-01 2.68179178e-01 1.23244989e+00 -1.16232681e+00
-9.51236784e-01 4.75663692e-01 -6.89801931e-01 7.03293860e-01
-7.63140544e-02 -1.69483155e-01 1.05431318e+00 -9.30094600e-01
2.57150661e-02 9.38417196e-01 8.32208216e-01 -1.54910594e-01
-1.34010255e+00 -7.37932205e-01 -2.07073197e-01 -2.06826851e-01
-1.15137327e+00 -2.27967471e-01 8.06201249e-02 -7.34797060e-01
6.15453780e-01 1.11713558e-01 3.43965560e-01 9.34734523e-01
1.07200742e+00 -2.07889438e-01 1.39578891e+00 -3.13593358e-01
4.69812453e-01 -5.61765470e-02 2.48509973e-01 -1.26989022e-01
6.34070158e-01 5.67891359e-01 -2.66824633e-01 -4.62680727e-01
9.99453843e-01 1.01423755e-01 1.35884404e-01 2.62646735e-01
-6.42219067e-01 1.03762662e+00 -2.99048007e-01 1.38362423e-01
-1.00685668e+00 2.26441830e-01 1.29784808e-01 7.65938818e-01
9.58691239e-01 1.98468398e-02 -5.64727128e-01 -2.73597270e-01
-1.10898364e+00 4.19252962e-01 1.14684021e+00 7.03327954e-01
1.49992466e-01 5.46830773e-01 -5.78395307e-01 2.48289287e-01
4.52279262e-02 1.01309109e+00 5.14244080e-01 -9.39025760e-01
1.75050899e-01 -1.84849128e-01 6.14999413e-01 -3.66856605e-01
-1.53850183e-01 -5.98037004e-01 -5.45764387e-01 3.32538992e-01
5.68413794e-01 -4.84639466e-01 -3.75647277e-01 1.46729672e+00
-3.55297774e-02 2.41177723e-01 5.77290952e-02 4.39199328e-01
-2.29442582e-01 7.35082150e-01 1.16206825e-01 -8.38873386e-01
9.35615242e-01 -3.26930314e-01 -5.99517226e-01 2.74427354e-01
2.14379027e-01 -6.87919140e-01 6.96216285e-01 3.40784162e-01
-1.05074954e+00 -4.89410579e-01 -5.30131459e-01 4.89676267e-01
9.96562913e-02 -3.13546538e-01 3.10859103e-02 7.37302661e-01
-8.01942766e-01 9.62537467e-01 -5.67440927e-01 3.99943113e-01
-1.81073681e-01 -5.06192297e-02 4.99338582e-02 2.78467983e-01
-9.47518706e-01 8.36389124e-01 -7.89217874e-02 -1.84432250e-02
-3.89347374e-01 -8.31247330e-01 -1.32708743e-01 3.88707012e-01
1.50883226e-02 -4.22512263e-01 1.32404971e+00 -1.23540139e+00
-1.46989262e+00 2.49026790e-01 -6.06884174e-02 -5.48867047e-01
8.10605526e-01 -3.74577284e-01 -5.75189173e-01 -2.67498106e-01
5.15033901e-01 -5.95275998e-01 9.92276967e-01 -4.82433617e-01
-7.14527369e-01 -6.09925628e-01 -9.84929800e-01 -3.02715391e-01
3.38499904e-01 4.13094759e-02 8.43586862e-01 -7.39088118e-01
-1.31132886e-01 -6.89257205e-01 1.95281478e-04 -8.36622059e-01
1.32402703e-02 -2.75606126e-01 1.85701773e-01 -6.77054644e-01
1.23764563e+00 -2.13132095e+00 -7.43890226e-01 7.81234801e-01
-4.58757818e-01 -5.09213507e-01 1.90682083e-01 6.91088498e-01
-4.95453626e-01 3.60970311e-02 2.25292593e-02 3.07981074e-01
1.27972081e-01 2.44825855e-02 -8.37389708e-01 8.44723225e-01
2.84634810e-02 5.65063059e-01 -4.91298378e-01 7.64323100e-02
-2.65565336e-01 1.36279479e-01 -2.30872422e-01 7.95589909e-02
2.40882009e-01 1.89877838e-01 -3.94871414e-01 4.98299867e-01
6.94371223e-01 -1.50612555e-03 -2.27186844e-01 6.25768363e-01
-9.77827191e-01 1.02437153e-01 -1.19203889e+00 8.53317976e-02
-1.12711385e-01 6.34183526e-01 -4.55884933e-01 -1.00695217e+00
1.56767917e+00 3.01978230e-01 1.17635466e-01 -7.52195716e-01
3.06193773e-02 6.07296348e-01 -2.54535917e-02 -1.45737126e-01
4.40379232e-01 -9.42282200e-01 1.01317577e-01 2.31706470e-01
-4.04569656e-01 3.72955441e-01 -2.84902249e-02 -6.24615908e-01
1.01646864e+00 2.41285309e-01 6.62272811e-01 -8.26747596e-01
3.64851564e-01 -3.89741093e-01 6.06450260e-01 9.20967042e-01
2.91983644e-03 3.20348352e-01 1.31819224e+00 -1.23380668e-01
-1.09611392e+00 -1.39158189e+00 -4.97643977e-01 1.34558129e+00
-4.43851352e-01 6.22673333e-01 -6.20568022e-02 6.30320609e-02
6.98442459e-01 1.26488781e+00 -7.87873626e-01 2.29925904e-02
-1.37754112e-01 -7.58295476e-01 2.10581854e-01 8.07597876e-01
2.04871431e-01 -8.48253012e-01 -5.79347014e-01 3.98900926e-01
5.12063324e-01 -4.44881916e-01 -6.58833534e-02 5.14965415e-01
-9.76228476e-01 -9.41483915e-01 -8.82825434e-01 -3.37253541e-01
1.17882423e-01 -2.94840455e-01 9.92767096e-01 -6.92946196e-01
4.12248164e-01 4.65787143e-01 -1.23672619e-01 -5.48103988e-01
-8.35442543e-02 -4.70952243e-01 -1.32848948e-01 1.63554996e-01
2.28515327e-01 -5.31394184e-01 -5.64024031e-01 4.30402458e-01
-5.08353531e-01 -9.87452209e-01 2.19441906e-01 9.94773149e-01
7.30313122e-01 -1.70927256e-01 1.26051378e+00 -6.67859077e-01
9.37632740e-01 -7.75159657e-01 -1.14378870e+00 2.63753504e-01
-1.10524654e+00 9.29171816e-02 5.52061558e-01 -6.53399229e-01
-1.16745591e+00 -6.31201744e-01 3.39780509e-01 -3.08896661e-01
2.11160049e-01 8.71002614e-01 5.90923190e-01 4.64534789e-01
4.48712468e-01 9.22421589e-02 -7.65480697e-02 -3.86894226e-01
-8.86315256e-02 4.08907682e-01 7.39549458e-01 -6.70230210e-01
5.92633963e-01 1.51998460e-01 6.72113121e-01 -6.97985232e-01
-7.14242220e-01 -4.48308408e-01 -3.54785562e-01 -2.91286763e-02
5.08622825e-01 -8.63379121e-01 -9.52781081e-01 3.14067245e-01
-7.00212121e-01 -3.31120729e-01 -7.57446647e-01 9.97914135e-01
-8.60436976e-01 -1.78249687e-01 -7.02453732e-01 -1.50104868e+00
-3.54895324e-01 -4.30158049e-01 4.34193730e-01 2.49265745e-01
-1.75164238e-01 -1.22457635e+00 4.97131526e-01 -4.33199614e-01
4.60028827e-01 7.21641839e-01 9.50103521e-01 -1.25744605e+00
-4.65308242e-02 -5.87682426e-01 -2.70771116e-01 5.88046134e-01
-2.67127246e-01 4.03554469e-01 -4.97217506e-01 -7.99562410e-02
2.32389525e-01 2.26909593e-01 6.08312309e-01 1.13006699e+00
5.73563933e-01 -3.14374357e-01 2.60183752e-01 3.89852136e-01
1.38574898e+00 4.94584203e-01 5.44966459e-01 5.67132831e-01
-2.82387614e-01 5.17323613e-01 8.14322650e-01 1.05052018e+00
-7.35369092e-03 -1.05227791e-01 6.82670996e-02 4.01756167e-01
9.93651807e-01 -1.09095491e-01 6.85273588e-01 5.92685759e-01
-5.10175228e-01 -2.50921603e-02 -8.47391963e-01 2.75650263e-01
-1.63215685e+00 -1.40727961e+00 -5.39398789e-01 2.68547773e+00
6.02185130e-01 2.78122067e-01 3.51546973e-01 -1.53406829e-01
9.25860345e-01 -2.40383670e-01 -4.06704366e-01 -7.67666996e-01
-1.42758980e-01 5.65431535e-01 1.05110204e+00 5.36328673e-01
-7.65752733e-01 2.69917876e-01 6.46265602e+00 4.89774734e-01
-9.38033164e-01 -1.09289147e-01 8.04052234e-01 2.25833759e-01
-2.11333498e-01 2.71155328e-01 -9.83685851e-01 6.48828924e-01
1.50675094e+00 -7.85247326e-01 -1.16722256e-01 1.16178501e+00
5.34134686e-01 -4.57631886e-01 -8.70059311e-01 4.20100957e-01
-4.71335679e-01 -7.51296639e-01 -4.83027548e-01 5.39860606e-01
8.85330141e-01 -2.29132414e-01 5.87671578e-01 4.78397161e-01
1.97828799e-01 -1.10351539e+00 7.65431285e-01 1.33506835e+00
6.78681076e-01 -1.31752288e+00 1.23608601e+00 3.48016173e-01
-9.05152321e-01 -3.55052501e-01 -7.41127610e-01 -5.25594890e-01
-7.39069059e-02 7.81839669e-01 -3.73485744e-01 7.47988671e-02
4.30197626e-01 4.31430787e-01 -1.79979518e-01 1.16940081e+00
1.15546353e-01 9.30345178e-01 -1.06891066e-01 6.17272733e-03
2.42297843e-01 -7.28023767e-01 2.35600367e-01 9.43868577e-01
1.12828231e+00 1.72540814e-01 -2.82169163e-01 7.75644004e-01
3.76206249e-01 2.61829555e-01 -5.94858646e-01 2.19298914e-01
5.44435441e-01 5.25765121e-01 -5.01579642e-01 -2.35844523e-01
-4.77546096e-01 3.46115172e-01 -3.50214630e-01 6.45028591e-01
-6.58429563e-01 -5.16358495e-01 4.57680106e-01 2.64496356e-01
6.34678483e-01 -1.37690619e-01 -5.45269847e-01 -6.71346843e-01
-1.02572836e-01 -4.05955940e-01 4.79550213e-01 -5.87182581e-01
-1.55790663e+00 1.33195028e-01 1.60878509e-01 -1.00540745e+00
-8.33865225e-01 -6.15569413e-01 -1.08182991e+00 1.53101599e+00
-1.47631145e+00 -3.71925056e-01 3.05396378e-01 6.32169485e-01
-1.09622858e-01 -4.00660038e-01 4.22924340e-01 -2.97230691e-01
-3.46345335e-01 1.84206903e-01 1.11908412e+00 -4.88593392e-02
7.31650949e-01 -1.25029480e+00 6.63260743e-02 5.67983568e-01
-6.23549640e-01 5.46104550e-01 9.60247040e-01 -1.00537324e+00
-7.80112863e-01 -7.73039997e-01 1.11753881e+00 -1.39783069e-01
1.19683969e+00 3.93122256e-01 -9.81010616e-01 8.10515046e-01
-9.52578932e-02 1.02285631e-02 7.69240916e-01 1.04456715e-01
-5.47127910e-02 -3.86537433e-01 -7.94420719e-01 1.46681830e-01
7.36253243e-03 -2.60772645e-01 -6.49956584e-01 -3.72392721e-02
2.67721593e-01 2.68934071e-01 -1.33122969e+00 4.30307090e-01
9.16417837e-01 -1.20150638e+00 5.04491210e-01 -5.30968189e-01
4.24622238e-01 3.12950760e-01 -2.31568322e-01 -1.02611029e+00
-5.31721532e-01 -7.52696455e-01 2.69619972e-01 1.25662887e+00
3.25227916e-01 -1.24072361e+00 2.74949908e-01 5.23461640e-01
2.69391298e-01 -5.53317487e-01 -1.49719083e+00 -1.21167314e+00
4.87853259e-01 -2.39069238e-01 5.04356742e-01 6.13307118e-01
-2.97531933e-01 2.78163645e-02 -4.03107285e-01 -3.72794345e-02
6.36165321e-01 4.48932618e-01 4.75882649e-01 -1.76166689e+00
-1.88225731e-01 -4.68776941e-01 -1.29217669e-01 -6.08829618e-01
2.40026578e-01 -3.76622498e-01 -5.56291640e-02 -6.87345922e-01
-1.05645977e-01 1.14996089e-02 -3.97670627e-01 -1.28884658e-01
5.09058498e-02 -3.95317167e-01 -2.12386012e-01 1.84257910e-01
9.18683857e-02 5.64962208e-01 9.59063828e-01 5.99906147e-01
-4.45763201e-01 7.93670297e-01 -4.95904446e-01 5.41428685e-01
8.45696568e-01 -6.73735380e-01 -1.65027604e-01 8.04185450e-01
3.31803381e-01 9.41736698e-01 5.51282585e-01 -5.52970827e-01
4.88947071e-02 -5.73600292e-01 6.74922049e-01 -8.12978148e-01
-3.14520627e-01 -4.61242199e-01 4.07386959e-01 5.11497915e-01
-4.47477579e-01 9.14199233e-01 -6.12084977e-02 7.26399720e-01
-2.76470125e-01 -4.22438204e-01 7.00620055e-01 2.32815668e-01
-6.81342185e-02 -3.96209434e-02 -6.71296835e-01 3.78383338e-01
1.01553178e+00 -1.66930079e-01 -3.78537834e-01 -8.18525314e-01
-5.58912098e-01 8.01416859e-02 4.80754152e-02 -8.20809677e-02
4.72555906e-01 -1.15308619e+00 -1.00293124e+00 2.31500313e-01
-4.52238202e-01 -8.33992720e-01 -9.33319330e-02 1.17443669e+00
-3.23843002e-01 3.94929290e-01 -2.88564205e-01 -1.22347854e-01
-7.57494926e-01 3.46902311e-01 3.47121328e-01 -2.10106492e-01
-3.43665838e-01 2.38700598e-01 2.16843888e-01 3.24354082e-01
-1.94641814e-01 -3.76677841e-01 -5.37138060e-02 5.50569534e-01
4.38911051e-01 1.07338762e+00 -2.49804854e-01 -4.87051308e-01
-2.82043833e-02 4.64425743e-01 5.21003366e-01 -2.24280626e-01
1.40819716e+00 -1.75408170e-01 -2.65840858e-01 1.17166626e+00
9.23472166e-01 1.45720199e-01 -1.32928658e+00 -3.82735655e-02
5.50861359e-01 -2.76675969e-01 -2.31361851e-01 -4.01118040e-01
-4.39397633e-01 6.30817771e-01 3.57310474e-01 5.76369762e-01
7.07154095e-01 -3.77096385e-01 1.26610190e-01 2.05855489e-01
-1.08135670e-01 -1.24426591e+00 -2.85334885e-01 7.83288479e-01
7.79893100e-01 -8.09741974e-01 2.23626330e-01 3.31107348e-01
-8.58044505e-01 1.49245727e+00 -6.87848106e-02 -7.05163300e-01
1.10844123e+00 1.75808087e-01 -2.30914250e-01 3.00888389e-01
-9.72233891e-01 6.06730208e-03 4.07387108e-01 2.48840034e-01
6.12084031e-01 3.77679646e-01 -5.36861897e-01 6.70923412e-01
-4.84919429e-01 -2.02501923e-01 5.90810895e-01 5.13902009e-01
-6.79567933e-01 -5.85458040e-01 -5.78910410e-01 6.18147135e-01
-9.51325715e-01 -1.93066329e-01 2.59222806e-01 1.19686103e+00
-4.13723320e-01 7.14802444e-01 4.79735583e-01 2.98975557e-01
3.95782232e-01 5.34108341e-01 -1.13811359e-01 -1.19275220e-01
-4.26000923e-01 5.67498744e-01 -2.28981301e-01 -1.99205175e-01
-1.97035387e-01 -1.12648726e+00 -7.07442224e-01 -7.55444467e-01
-3.32595080e-01 1.83602050e-01 1.89686254e-01 9.85183001e-01
-2.58538514e-01 8.18297043e-02 9.35215831e-01 -4.49347198e-01
-1.35877526e+00 -9.59657550e-01 -1.50319707e+00 -1.90481588e-01
5.78633845e-01 -4.67155218e-01 -6.67014003e-01 -2.21401691e-01] | [6.232591152191162, 4.1286821365356445] |
82487c95-ce5d-4ef7-ad63-81932c4cd774 | recurrent-residual-convolutional-neural | 1802.06955 | null | http://arxiv.org/abs/1802.06955v5 | http://arxiv.org/pdf/1802.06955v5.pdf | Recurrent Residual Convolutional Neural Network based on U-Net (R2U-Net) for Medical Image Segmentation | Deep learning (DL) based semantic segmentation methods have been providing
state-of-the-art performance in the last few years. More specifically, these
techniques have been successfully applied to medical image classification,
segmentation, and detection tasks. One deep learning technique, U-Net, has
become one of the most popular for these applications. In this paper, we
propose a Recurrent Convolutional Neural Network (RCNN) based on U-Net as well
as a Recurrent Residual Convolutional Neural Network (RRCNN) based on U-Net
models, which are named RU-Net and R2U-Net respectively. The proposed models
utilize the power of U-Net, Residual Network, as well as RCNN. There are
several advantages of these proposed architectures for segmentation tasks.
First, a residual unit helps when training deep architecture. Second, feature
accumulation with recurrent residual convolutional layers ensures better
feature representation for segmentation tasks. Third, it allows us to design
better U-Net architecture with same number of network parameters with better
performance for medical image segmentation. The proposed models are tested on
three benchmark datasets such as blood vessel segmentation in retina images,
skin cancer segmentation, and lung lesion segmentation. The experimental
results show superior performance on segmentation tasks compared to equivalent
models including U-Net and residual U-Net (ResU-Net). | ['Mahmudul Hasan', 'Tarek M. Taha', 'Md Zahangir Alom', 'Vijayan K. Asari', 'Chris Yakopcic'] | 2018-02-20 | null | null | null | null | ['skin-cancer-segmentation', 'lung-nodule-segmentation'] | ['medical', 'medical'] | [ 2.21088767e-01 7.67758787e-02 -2.45395854e-01 -1.32063970e-01
-1.45571575e-01 3.55346985e-02 8.94787908e-02 -2.29115516e-01
-4.30015951e-01 5.64417481e-01 1.37203112e-01 -3.03532451e-01
2.39276618e-01 -1.00840533e+00 -2.69508898e-01 -6.62380099e-01
2.38856480e-01 -1.85377926e-01 7.09784865e-01 -7.63174072e-02
3.03452820e-01 5.44824541e-01 -1.18831050e+00 1.84311196e-01
9.32168961e-01 1.17136514e+00 3.30762178e-01 6.34889603e-01
-6.27236068e-01 1.03010046e+00 -2.95550257e-01 8.04367438e-02
1.39847293e-01 -6.12726390e-01 -8.37722778e-01 3.06669269e-02
-2.33801246e-01 -2.83347189e-01 -3.88052642e-01 9.78183031e-01
6.23105645e-01 1.95682764e-01 6.74752951e-01 -5.66778541e-01
-7.62031555e-01 5.01208484e-01 -7.95226395e-01 4.69280154e-01
-1.03647672e-01 -1.13263130e-01 6.31878257e-01 -2.88813591e-01
4.84359622e-01 1.26566291e+00 8.12180102e-01 8.09346855e-01
-5.95505536e-01 -6.06185377e-01 -9.59159508e-02 7.31075928e-02
-1.13380253e+00 7.85351098e-02 5.79301238e-01 -4.15303648e-01
7.99869299e-01 4.28638309e-02 5.36301494e-01 5.09062707e-01
4.43068236e-01 1.23515701e+00 1.02981710e+00 -2.79924452e-01
-4.48020585e-02 4.85255197e-02 7.09756315e-01 9.18379903e-01
1.66874409e-01 -1.25559941e-01 3.81509513e-01 3.33939195e-01
1.27815294e+00 2.21787274e-01 -3.40454519e-01 2.53369838e-01
-7.20720649e-01 7.84836590e-01 1.04837477e+00 7.39140213e-01
-3.19886833e-01 1.33502021e-01 6.63956046e-01 -3.36898118e-02
3.84254336e-01 6.90776631e-02 -3.37426841e-01 2.95545787e-01
-5.56315124e-01 -1.28555045e-01 3.47362846e-01 7.00884938e-01
5.89752972e-01 3.24336052e-01 -4.25932735e-01 1.15393043e+00
2.07886666e-01 -1.93774238e-01 1.26335335e+00 -3.87492955e-01
1.29728585e-01 9.84487772e-01 -4.75732684e-01 -6.94190025e-01
-8.04128051e-01 -6.00263596e-01 -1.16746163e+00 8.67645219e-02
1.33478105e-01 -4.00110781e-01 -1.43314338e+00 1.23169339e+00
9.53347087e-02 5.62965751e-01 5.52322567e-01 8.52638662e-01
1.46781445e+00 8.51256132e-01 2.15682581e-01 8.20852146e-02
1.28235734e+00 -1.20836067e+00 -6.65851772e-01 2.82266829e-02
8.42149436e-01 -7.80827880e-01 5.25494993e-01 3.80806886e-02
-7.93805540e-01 -9.90471244e-01 -1.02091646e+00 -2.21956760e-01
-5.59234858e-01 3.75349045e-01 8.26307714e-01 6.98176861e-01
-1.05392802e+00 4.91444945e-01 -7.76451349e-01 -5.64122081e-01
7.85660148e-01 4.95816261e-01 -1.62826419e-01 1.92287520e-01
-1.13427460e+00 6.02967441e-01 5.54828107e-01 3.45650107e-01
-5.77947676e-01 -2.26932794e-01 -6.76545024e-01 -3.38843986e-02
1.40949652e-01 -4.70750868e-01 1.11731422e+00 -9.49989796e-01
-1.38761067e+00 7.65674949e-01 5.30872494e-02 -6.53869152e-01
3.57048571e-01 -2.94795297e-02 -2.60894537e-01 3.13052595e-01
-8.82389992e-02 8.45666945e-01 2.12372854e-01 -8.79052103e-01
-6.74724162e-01 -3.33711833e-01 -1.18860774e-01 1.38359576e-01
-1.54185578e-01 -1.36652410e-01 -3.07979912e-01 -6.78394079e-01
3.86273324e-01 -6.62352800e-01 -5.86590707e-01 -2.54183888e-01
-4.83302921e-01 -5.06394506e-01 1.24820673e+00 -6.23071730e-01
1.20788157e+00 -1.95659804e+00 -1.05316229e-01 1.13425523e-01
1.46629810e-01 9.34704125e-01 -1.34344161e-01 -1.80636421e-01
-2.21392155e-01 3.40516359e-01 -2.64231324e-01 1.60760060e-01
-7.08048344e-01 3.12535405e-01 2.61691868e-01 2.50501692e-01
1.72995657e-01 1.12864172e+00 -4.48556304e-01 -5.32232881e-01
4.50566262e-01 4.06826109e-01 -3.55519712e-01 1.49351612e-01
-2.41307542e-01 7.02887893e-01 -8.34028006e-01 6.54171288e-01
6.53744042e-01 -1.20813534e-01 -2.91274130e-01 -2.95720518e-01
-2.02705771e-01 -3.61511886e-01 -9.19197142e-01 1.36993361e+00
-4.42298949e-01 5.80148697e-01 -1.15008734e-01 -1.30491734e+00
1.30588472e+00 3.73174012e-01 4.82940197e-01 -7.76145279e-01
7.96469271e-01 2.90986896e-01 1.64232448e-01 -7.73778737e-01
1.47665575e-01 -1.90691128e-01 3.23268205e-01 1.48421958e-01
1.35406377e-02 4.56801385e-01 1.76915526e-01 -1.91498488e-01
7.59151340e-01 8.01380500e-02 2.80051231e-01 -1.67362571e-01
1.13143861e+00 -2.52981424e-01 7.99386203e-01 3.47204566e-01
-3.27046692e-01 5.82223415e-01 6.12323642e-01 -4.71222788e-01
-6.15751147e-01 -5.16568422e-01 -3.72966141e-01 5.78642428e-01
3.11299503e-01 1.97342157e-01 -8.86864305e-01 -5.03937364e-01
-2.87316829e-01 2.57665396e-01 -6.77927017e-01 -6.22185320e-02
-6.69474661e-01 -9.46685791e-01 7.27258861e-01 8.31050992e-01
1.40781832e+00 -1.38010418e+00 -6.07707143e-01 3.91298473e-01
2.33717754e-01 -9.11919355e-01 -1.83332652e-01 -1.21294130e-02
-1.27367270e+00 -1.28271949e+00 -1.20850241e+00 -1.21606338e+00
8.10401917e-01 2.51982540e-01 5.21717489e-01 2.09144786e-01
-7.26489365e-01 1.52685577e-02 -5.94681978e-01 -2.61882067e-01
-2.41791785e-01 1.49262547e-01 -5.96444845e-01 1.18065141e-01
4.12711143e-01 -3.19120586e-01 -6.49730742e-01 2.34077707e-01
-1.01648700e+00 2.37391926e-02 8.93642187e-01 9.10545468e-01
7.34458923e-01 9.86270085e-02 8.64236057e-01 -1.40271318e+00
6.25316322e-01 -4.55485672e-01 -4.36911374e-01 1.14619605e-01
-3.82706493e-01 -2.32996181e-01 6.88387930e-01 -9.58599150e-02
-1.15613246e+00 -9.94615778e-02 -4.91232127e-01 -2.87768155e-01
-1.61576480e-01 5.77370405e-01 4.79140282e-02 -1.64969936e-01
4.15788919e-01 2.09981203e-01 1.43077865e-01 -5.96392035e-01
2.01743498e-01 1.04876804e+00 5.02382815e-01 -2.11505935e-01
-1.82168428e-02 1.31243229e-01 6.68745637e-02 -9.19420362e-01
-7.42382646e-01 -7.03415692e-01 -6.66442990e-01 -1.06646538e-01
1.55736125e+00 -7.46001065e-01 -5.64038157e-01 9.20027912e-01
-1.13761401e+00 -2.32474595e-01 6.69773743e-02 5.63799262e-01
-2.95538723e-01 4.10268962e-01 -8.32311690e-01 -7.62787700e-01
-6.77518189e-01 -1.43985713e+00 4.71297830e-01 1.14184475e+00
3.71434569e-01 -1.28875184e+00 -3.32505643e-01 3.31643552e-01
5.03817499e-01 6.34513199e-01 9.95096326e-01 -6.46359324e-01
-4.23917890e-01 -3.05014133e-01 -7.78428316e-01 6.96320295e-01
3.29683244e-01 -1.59328416e-01 -8.02659631e-01 5.68744130e-02
-1.54046595e-01 -2.28634641e-01 1.26314580e+00 9.16679263e-01
1.52892399e+00 -1.91193745e-02 -5.32284498e-01 7.19868541e-01
1.72187221e+00 6.85018957e-01 8.86797845e-01 1.41484380e-01
8.97901654e-01 3.05338383e-01 3.08612406e-01 9.19221621e-03
6.00676984e-02 7.06363320e-02 3.99730206e-01 -7.24896431e-01
-4.38757122e-01 1.90103948e-01 -1.29599229e-01 8.48524749e-01
-1.83362573e-01 -5.11056595e-02 -7.88678110e-01 6.04650438e-01
-1.92050099e+00 -6.28826201e-01 -3.34579617e-01 1.84137082e+00
5.37769914e-01 1.51954740e-01 5.11390269e-02 1.05075322e-01
9.53681648e-01 4.60456684e-02 -7.40466535e-01 -6.03176177e-01
2.07912214e-02 6.01977587e-01 7.12498128e-01 1.91221654e-01
-1.18296361e+00 1.12777364e+00 5.37127447e+00 7.97550797e-01
-1.35232198e+00 1.47432610e-01 1.08093286e+00 5.51061392e-01
1.15239993e-01 -3.51545513e-01 -7.38717139e-01 2.76387334e-01
6.08709812e-01 2.41181359e-01 -8.42090622e-02 7.51047432e-01
1.11036018e-01 -1.61030844e-01 -5.56238532e-01 9.11683023e-01
-1.30809307e-01 -1.49148369e+00 2.00459346e-01 -2.36650869e-01
7.75928319e-01 1.09796980e-02 -5.66501617e-02 3.79846931e-01
1.28638208e-01 -1.20147383e+00 -1.15518838e-01 4.23798412e-01
7.93643594e-01 -9.63797867e-01 1.31557631e+00 1.57363877e-01
-1.31986094e+00 -5.97943142e-02 -7.73252845e-01 1.24673545e-01
-4.50544171e-02 5.00792921e-01 -4.45019394e-01 8.04165006e-01
4.93206799e-01 1.09907806e+00 -4.04626936e-01 1.25816047e+00
-1.84099525e-02 5.36923051e-01 6.71184361e-02 -2.20279247e-01
7.75753856e-01 -4.23896641e-01 1.82545930e-01 1.26791012e+00
6.04916401e-02 3.26666892e-01 2.59358257e-01 7.05616176e-01
-1.44182280e-01 3.66938800e-01 -4.60960984e-01 -4.65299226e-02
-1.08467393e-01 1.40628850e+00 -1.10573065e+00 -3.06111187e-01
-4.36824411e-01 7.23663092e-01 2.91139595e-02 2.64589012e-01
-6.04016006e-01 -8.64106894e-01 3.76249939e-01 -1.48215443e-01
1.85916170e-01 7.59155676e-02 -2.91463375e-01 -7.36176014e-01
-5.35549283e-01 -2.60562837e-01 4.29402143e-01 -5.11662722e-01
-1.01337314e+00 8.02786827e-01 -3.15344065e-01 -1.05181789e+00
2.59729564e-01 -8.23290586e-01 -9.89238679e-01 1.03899109e+00
-1.90304780e+00 -1.16065753e+00 -5.26122510e-01 5.13551712e-01
6.77414477e-01 -3.46457183e-01 5.57823479e-01 3.23403686e-01
-1.09573817e+00 3.13012511e-01 -4.79104817e-02 6.46411359e-01
2.74633169e-01 -1.11695910e+00 4.78274561e-02 7.44980037e-01
-4.18426394e-01 6.03725553e-01 -9.52438787e-02 -5.86099565e-01
-8.38208914e-01 -1.53498435e+00 2.15525508e-01 4.47949201e-01
1.90079063e-01 2.89130211e-01 -7.97491550e-01 6.76780581e-01
1.47212565e-01 3.56156617e-01 7.56327271e-01 -3.14404249e-01
1.18615650e-01 -7.96489418e-02 -1.28523564e+00 3.99197340e-01
5.47594547e-01 -1.03902034e-01 -3.43513727e-01 3.11561197e-01
7.12112784e-01 -6.10958457e-01 -8.27891350e-01 5.90487063e-01
4.63085681e-01 -1.02931905e+00 7.62981713e-01 -3.27421606e-01
4.74806458e-01 -2.84959048e-01 2.16964200e-01 -9.23886836e-01
-2.87935019e-01 -1.30886853e-01 4.01247412e-01 1.14846587e+00
2.05371365e-01 -8.16469610e-01 9.36800063e-01 1.74282402e-01
-3.87531161e-01 -1.30012989e+00 -5.71614802e-01 -3.96516681e-01
1.74571142e-01 9.16490406e-02 2.89032608e-01 7.09774256e-01
-6.16701424e-01 2.88725078e-01 -6.52503297e-02 -2.74028271e-01
2.67075956e-01 -1.19213611e-01 4.09575373e-01 -1.24514639e+00
5.78022525e-02 -6.59460902e-01 -6.41405761e-01 -1.24464393e+00
9.60184932e-02 -1.02948546e+00 -2.42419437e-01 -2.12471247e+00
6.73603117e-02 -7.33516812e-01 -5.38379133e-01 4.67566609e-01
-2.44711190e-01 3.01035881e-01 -5.12031093e-02 1.97352499e-01
-2.00234815e-01 3.32162291e-01 1.73498845e+00 -1.23355620e-01
-6.12815917e-01 3.01594526e-01 -6.62007213e-01 7.59458244e-01
1.09164095e+00 1.39284000e-01 -5.02937853e-01 -3.15998763e-01
-6.28130436e-01 1.28845230e-01 2.90325209e-02 -1.22697055e+00
2.46811837e-01 3.16896647e-01 6.43241286e-01 -7.86944091e-01
-1.27689183e-01 -5.73857129e-01 -2.55200833e-01 8.56623769e-01
-3.06163758e-01 -3.57774734e-01 2.82247394e-01 3.71957332e-01
-5.02817512e-01 -5.19216299e-01 1.19323695e+00 -6.65112913e-01
-1.11440563e+00 4.79619145e-01 -3.63242686e-01 -3.56186092e-01
1.30601335e+00 -5.92934072e-01 -3.44090581e-01 1.47899896e-01
-8.26634943e-01 4.04093951e-01 -2.92461306e-01 2.93441772e-01
8.07567179e-01 -1.07934868e+00 -4.56770480e-01 9.76873711e-02
-1.46087065e-01 4.31025952e-01 4.39999580e-01 8.42094243e-01
-1.19966352e+00 7.18882382e-01 -5.82771719e-01 -5.14409900e-01
-1.18819284e+00 2.37920940e-01 6.13608062e-01 -3.57103109e-01
-7.66483545e-01 1.09124041e+00 3.43849629e-01 -4.60399479e-01
2.10427120e-01 -5.56606650e-01 -1.06715679e+00 -9.45007876e-02
4.12429124e-01 5.13323188e-01 -2.26953372e-01 -6.49790645e-01
-1.10004820e-01 9.98778820e-01 -3.82566392e-01 5.18585145e-01
1.01928413e+00 9.89849940e-02 -3.43026996e-01 9.80139449e-02
1.34461796e+00 -5.80115318e-01 -9.15683031e-01 -2.34894753e-01
-8.58677179e-02 9.01638567e-02 5.31689584e-01 -7.18844891e-01
-1.64923203e+00 1.06279540e+00 8.26432943e-01 3.15094702e-02
1.18013060e+00 -3.81561518e-01 1.35795462e+00 1.11540128e-02
4.23422977e-02 -9.61384535e-01 3.14054675e-02 2.81924695e-01
3.71636391e-01 -1.09152305e+00 -1.18880510e-01 -5.35985291e-01
-4.88220632e-01 1.42879438e+00 9.64251280e-01 -4.85020250e-01
8.34071577e-01 -9.83236954e-02 2.89899021e-01 -9.58632678e-03
-2.63271928e-01 -6.56338573e-01 7.22064301e-02 4.40346897e-01
7.76739001e-01 1.86868962e-02 -5.88317394e-01 5.18471301e-01
2.33841583e-01 2.49909282e-01 6.54533982e-01 8.32836926e-01
-7.90105641e-01 -1.12208915e+00 -1.56501800e-01 8.65614295e-01
-7.67547309e-01 1.80944860e-01 -1.19379856e-01 6.97544336e-01
3.44403774e-01 8.39693487e-01 6.85016289e-02 -3.90066594e-01
9.45018232e-02 -2.42673025e-01 3.68341476e-01 -6.96629465e-01
-7.39606678e-01 1.69886038e-01 -2.26348162e-01 -4.03490990e-01
-6.86798811e-01 -1.33403048e-01 -1.81562436e+00 1.55888766e-01
-4.49679345e-01 1.54093981e-01 5.95718861e-01 9.43880141e-01
-3.33879180e-02 1.12729609e+00 5.21189928e-01 -4.38752502e-01
6.83757737e-02 -1.09340894e+00 -7.30217814e-01 -5.41650108e-04
1.69269845e-01 -5.30269980e-01 1.41355991e-01 -9.96619910e-02] | [14.812722206115723, -2.7027487754821777] |
a7382a37-8931-47a0-8599-c68f58d9545e | gantee-generative-adversatial-network-for | 2303.14480 | null | https://arxiv.org/abs/2303.14480v1 | https://arxiv.org/pdf/2303.14480v1.pdf | GANTEE: Generative Adversatial Network for Taxonomy Entering Evaluation | Taxonomy is formulated as directed acyclic concepts graphs or trees that support many downstream tasks. Many new coming concepts need to be added to an existing taxonomy. The traditional taxonomy expansion task aims only at finding the best position for new coming concepts in the existing taxonomy. However, they have two drawbacks when being applied to the real-scenarios. The previous methods suffer from low-efficiency since they waste much time when most of the new coming concepts are indeed noisy concepts. They also suffer from low-effectiveness since they collect training samples only from the existing taxonomy, which limits the ability of the model to mine more hypernym-hyponym relationships among real concepts. This paper proposes a pluggable framework called Generative Adversarial Network for Taxonomy Entering Evaluation (GANTEE) to alleviate these drawbacks. A generative adversarial network is designed in this framework by discriminative models to alleviate the first drawback and the generative model to alleviate the second drawback. Two discriminators are used in GANTEE to provide long-term and short-term rewards, respectively. Moreover, to further improve the efficiency, pre-trained language models are used to retrieve the representation of the concepts quickly. The experiments on three real-world large-scale datasets with two different languages show that GANTEE improves the performance of the existing taxonomy expansion methods in both effectiveness and efficiency. | ['Jian Zhong', 'Jiaqing Liang', 'Zhixu Li', 'Hongwei Feng', 'Yanghua Xiao', 'Jingping Liu', 'Sihang Jiang', 'Zhouhong Gu'] | 2023-03-25 | null | null | null | null | ['taxonomy-expansion'] | ['natural-language-processing'] | [-3.46850716e-02 -1.20966382e-01 -4.84682657e-02 -3.51286054e-01
-2.28902295e-01 -5.84636271e-01 4.52677935e-01 1.70590475e-01
-5.62036157e-01 6.77644193e-01 1.43423468e-01 -2.58382499e-01
-1.86960012e-01 -1.32822907e+00 -1.25525713e-01 -6.12487078e-01
1.78884611e-01 7.15730608e-01 1.94038570e-01 -5.14133573e-01
2.60684043e-02 2.64219999e-01 -1.73735118e+00 4.81441021e-02
1.26939261e+00 6.06924593e-01 4.93457824e-01 -3.33733559e-02
-7.79658437e-01 4.35264885e-01 -7.92887747e-01 -6.79018259e-01
2.82854229e-01 -2.41992444e-01 -5.03472924e-01 -1.67179734e-01
-8.63994509e-02 -1.37604475e-01 -2.92483360e-01 1.26241446e+00
4.32659358e-01 3.22657228e-01 5.54713845e-01 -1.49497139e+00
-6.82538152e-01 7.17043221e-01 -4.62619066e-01 3.59265916e-02
4.34984654e-01 -3.37485939e-01 1.05761528e+00 -9.25821066e-01
5.92712939e-01 1.40523863e+00 5.51717162e-01 6.59023941e-01
-5.43832064e-01 -1.20644915e+00 4.18022275e-01 4.30114865e-01
-1.39385962e+00 1.45888075e-01 7.31111586e-01 -2.75137037e-01
8.89691830e-01 2.14204207e-01 6.35776460e-01 9.91122127e-01
-1.33862600e-01 3.77402663e-01 7.16055036e-01 -2.70697653e-01
2.15890512e-01 2.57702738e-01 3.98783274e-02 4.73859310e-01
2.37034604e-01 8.83155242e-02 -2.00677067e-01 -2.51023710e-01
5.20750940e-01 2.34154120e-01 -8.85394961e-02 -3.53742421e-01
-7.26980329e-01 1.07249784e+00 7.35450625e-01 6.63906395e-01
-1.19902469e-01 -3.56249720e-01 5.50041199e-01 2.47868583e-01
2.01481625e-01 5.69737136e-01 -3.08636755e-01 1.51628628e-01
-6.53725624e-01 1.89009622e-01 4.89068329e-01 1.23019099e+00
7.46773839e-01 -2.64092442e-02 1.86669782e-01 9.71082687e-01
3.32439989e-01 1.22970179e-01 1.14929140e+00 -3.29235017e-01
5.82530499e-01 1.18954206e+00 -2.95977026e-01 -1.03602374e+00
-4.59816098e-01 -7.29468226e-01 -8.81914139e-01 -1.98179245e-01
-1.32477269e-01 5.70362508e-02 -1.00185311e+00 1.75802672e+00
4.25834864e-01 2.87797332e-01 1.75439134e-01 6.57522976e-01
1.21835768e+00 8.04677308e-01 3.04556221e-01 -2.49071524e-01
1.33018017e+00 -9.01241541e-01 -5.74031889e-01 -3.93567383e-01
5.12823641e-01 -7.21828759e-01 1.20559740e+00 2.78111517e-01
-5.45204759e-01 -6.28081262e-01 -9.62419033e-01 7.30697811e-02
-9.35345590e-01 -5.58219627e-02 9.01243746e-01 8.11283350e-01
-6.96134031e-01 2.70804882e-01 -2.84665376e-01 -3.80830526e-01
2.59009421e-01 4.02195662e-01 -1.62940264e-01 -1.38886169e-01
-1.60775232e+00 7.39058614e-01 9.55617070e-01 2.61811214e-03
-5.77208817e-01 -4.49928015e-01 -8.20154965e-01 2.45249107e-01
4.79486614e-01 -6.74002469e-01 1.03173304e+00 -6.81705356e-01
-1.13416231e+00 5.74829578e-01 3.53541300e-02 -3.07158202e-01
2.16743752e-01 1.03641376e-01 -8.20321739e-01 -2.25292996e-01
3.33983272e-01 5.48125982e-01 5.36621928e-01 -1.25046957e+00
-1.07125115e+00 -4.30773854e-01 2.30384290e-01 4.89493251e-01
-8.85371625e-01 -2.18179643e-01 -4.80620921e-01 -9.17270303e-01
2.98148096e-01 -6.30093634e-01 -3.32640111e-01 -4.37032014e-01
-2.14015305e-01 -4.58337069e-01 1.07429659e+00 -3.99867982e-01
1.36617231e+00 -1.83474720e+00 4.09214422e-02 1.09138377e-01
2.12574050e-01 5.89431584e-01 -1.49838552e-01 4.89072412e-01
-1.51555240e-01 1.22350961e-01 -2.26472229e-01 -4.17794883e-02
-2.48010546e-01 5.76804578e-01 -4.84786123e-01 -3.29933763e-01
-1.13117523e-01 8.11915517e-01 -1.32758009e+00 -3.87909502e-01
2.56453574e-01 2.02895001e-01 -4.18036252e-01 2.64662743e-01
-3.04087311e-01 2.36559182e-01 -4.52712208e-01 4.73457128e-01
7.02357352e-01 2.08292186e-01 2.04444960e-01 4.76539247e-02
2.72017896e-01 2.04589620e-01 -1.26360714e+00 1.73996675e+00
-6.80774391e-01 -9.08829048e-02 -4.05521482e-01 -1.19746900e+00
1.30832219e+00 3.25591028e-01 4.38034415e-01 -5.79164147e-01
1.07929088e-01 3.90617520e-01 4.70775254e-02 -4.92588252e-01
6.36841416e-01 -5.61735749e-01 -2.07091242e-01 1.85962483e-01
1.25154465e-01 -2.43643105e-01 3.84386301e-01 2.91397840e-01
8.25197697e-01 7.42604360e-02 4.46176380e-01 1.87392548e-01
7.87402868e-01 -1.38803706e-01 8.63179564e-01 3.72447133e-01
1.17760785e-01 1.89626947e-01 1.22285172e-01 -3.36288869e-01
-7.74747431e-01 -9.25069571e-01 1.73956156e-02 1.01638508e+00
3.97093982e-01 -6.28174722e-01 -2.93989480e-01 -1.13858461e+00
-7.80066177e-02 9.12108660e-01 -3.16602916e-01 -4.48632836e-01
-2.38479689e-01 -7.37193465e-01 5.39554954e-01 6.66056871e-01
6.07201278e-01 -1.27371466e+00 -4.19403702e-01 4.41538334e-01
-3.39360893e-01 -1.04492962e+00 -2.65515208e-01 -3.63983363e-02
-9.18012500e-01 -9.57929075e-01 -5.00488698e-01 -9.36472058e-01
5.49721777e-01 3.67101520e-01 1.00772870e+00 6.98712319e-02
-5.46308607e-02 -1.12161316e-01 -6.39420629e-01 -4.45542186e-01
-3.62309307e-01 2.13183537e-01 1.56209648e-01 -2.66140282e-01
8.74897778e-01 -8.29903126e-01 -1.60481289e-01 3.72594357e-01
-1.08535028e+00 -2.24998057e-01 4.14404184e-01 1.12645745e+00
3.71279508e-01 7.31173456e-01 9.78014112e-01 -1.02337658e+00
9.78904843e-01 -6.58368289e-01 -5.91813385e-01 3.30614418e-01
-9.18048203e-01 3.58988196e-02 1.02056682e+00 -5.44052660e-01
-1.23493934e+00 -9.06861946e-02 -3.01449805e-01 -2.83920348e-01
1.79493781e-02 8.85222197e-01 -4.59820896e-01 2.32589945e-01
5.36286294e-01 2.96461880e-01 -5.43809295e-01 -5.39776504e-01
5.62557578e-01 8.13178480e-01 6.67461812e-01 -4.69705492e-01
1.05663967e+00 1.50594264e-01 -7.75572434e-02 -4.44477022e-01
-7.39725411e-01 -7.82498240e-01 -6.26798093e-01 8.70423838e-02
3.77511084e-01 -8.35600555e-01 -4.48534876e-01 1.48668170e-01
-1.07904375e+00 4.84954149e-01 -3.48153442e-01 4.04925108e-01
-1.67096317e-01 5.45684934e-01 -2.07791910e-01 -7.95651197e-01
-7.58863568e-01 -9.39822555e-01 6.61861539e-01 5.24766922e-01
6.91672787e-02 -8.93698096e-01 -5.94760664e-02 2.14720622e-01
2.02945918e-01 -5.43845911e-03 1.29134548e+00 -8.57105315e-01
-2.81210840e-01 -3.60318691e-01 -2.38338649e-01 3.11310232e-01
3.67962748e-01 -2.23350555e-01 -8.02549541e-01 -3.41898710e-01
8.22071135e-02 -3.11430067e-01 7.82472670e-01 -1.25667989e-01
1.16868305e+00 -2.98621416e-01 -5.08223712e-01 5.63420057e-01
1.46864843e+00 6.99687779e-01 5.75853586e-01 4.05966878e-01
7.99542725e-01 5.65340400e-01 1.00923383e+00 2.98358142e-01
3.42300415e-01 5.36958218e-01 4.16703701e-01 1.51237562e-01
1.12178601e-01 -5.93696773e-01 8.07186067e-02 1.03777635e+00
4.88924421e-02 -3.35604578e-01 -8.34841430e-01 6.51794136e-01
-1.82067454e+00 -8.52584183e-01 -1.13955606e-02 2.37755299e+00
8.58309150e-01 2.71494120e-01 1.08107626e-01 2.28728041e-01
6.83944881e-01 -1.97504088e-01 -6.52271271e-01 -2.87261158e-01
-5.52380010e-02 3.07677716e-01 3.95914912e-02 2.45000124e-01
-8.59501302e-01 1.12819099e+00 5.41235161e+00 1.04374909e+00
-8.54158342e-01 8.07439536e-02 2.41490081e-02 1.64748162e-01
-5.40987432e-01 1.64708585e-01 -8.36866379e-01 4.22425717e-01
5.41391134e-01 -5.33213019e-01 1.55938655e-01 1.09562278e+00
-3.56671423e-01 1.91727176e-01 -7.23121583e-01 1.02585316e+00
-4.24971096e-02 -7.58041024e-01 4.97199744e-01 -1.33227065e-01
5.33581316e-01 -2.53664285e-01 -8.51606950e-02 7.20555127e-01
4.81714785e-01 -9.46620405e-01 2.04353198e-01 2.69366682e-01
9.14259136e-01 -1.13287973e+00 1.00142086e+00 4.81967926e-01
-1.60671663e+00 -3.71115446e-01 -8.01550150e-01 1.61963012e-02
4.01279144e-02 5.94623625e-01 -1.12241626e+00 1.04124475e+00
5.09816289e-01 6.12763822e-01 -6.97571754e-01 1.17985630e+00
-6.13055706e-01 2.02182531e-01 -3.18897694e-01 -1.68809652e-01
4.18249905e-01 -3.58326435e-01 5.76601386e-01 9.84809160e-01
4.38595057e-01 1.99358445e-02 2.93093175e-01 8.38689268e-01
-1.76653877e-01 3.46715689e-01 -5.82884252e-01 -1.65829346e-01
7.74371207e-01 1.39101398e+00 -5.71748555e-01 -3.61104816e-01
-3.53040248e-01 6.88358188e-01 2.73805082e-01 2.24738181e-01
-6.45263433e-01 -7.25273073e-01 3.10850888e-01 -8.39022100e-02
-1.35887833e-02 1.11619700e-02 -2.13973463e-01 -1.25711370e+00
4.55818623e-02 -9.63167012e-01 8.17695141e-01 -7.44607508e-01
-1.54960382e+00 8.23402286e-01 -9.60369129e-03 -1.34759176e+00
-4.09120381e-01 -3.56856465e-01 -4.81235147e-01 1.07439625e+00
-1.30003774e+00 -1.29138529e+00 -5.28052330e-01 6.68409228e-01
5.80201924e-01 -4.93699342e-01 1.09828484e+00 5.08965313e-01
-4.54579890e-01 7.35188603e-01 1.28979340e-01 4.47162762e-02
5.79479039e-01 -1.28144443e+00 5.12731016e-01 9.26779985e-01
2.81217426e-01 8.43021512e-01 4.53425288e-01 -9.10365105e-01
-6.50478244e-01 -1.05440807e+00 1.09072399e+00 -6.23883232e-02
5.43724418e-01 -4.70478266e-01 -1.17532158e+00 5.08790791e-01
-1.33852020e-01 -4.66822028e-01 7.07928240e-01 3.46115053e-01
-5.72921693e-01 -1.04503192e-01 -1.19365191e+00 6.83530271e-01
1.16365290e+00 -2.16333121e-01 -9.52212691e-01 2.39298061e-01
9.15906906e-01 -1.20814770e-01 -7.25532055e-01 5.73159635e-01
3.86619031e-01 -7.58839011e-01 9.83514428e-01 -6.24114275e-01
4.05813217e-01 -3.72800559e-01 3.01435236e-02 -1.50422919e+00
-2.80990899e-01 -3.44206125e-01 -8.73602033e-02 1.49404693e+00
2.20445514e-01 -7.86895990e-01 8.58961225e-01 1.57319158e-01
4.86460663e-02 -8.14223289e-01 -8.41365278e-01 -1.05836689e+00
1.51634961e-01 -2.59403139e-01 1.22080266e+00 1.16529179e+00
9.82562825e-02 8.84762883e-01 -7.14568049e-02 -1.08391814e-01
3.62569571e-01 3.04953098e-01 5.52518189e-01 -1.51640952e+00
-1.02709398e-01 -3.87352884e-01 -5.88451087e-01 -9.17189062e-01
5.64958677e-02 -1.03041029e+00 -1.90533638e-01 -1.63707829e+00
1.88977569e-01 -8.02685380e-01 -3.60823661e-01 6.67421937e-01
-4.74307001e-01 -2.10685313e-01 1.38267428e-01 1.72765911e-01
-1.21134602e-01 8.70642543e-01 1.18872905e+00 -2.41440758e-01
-3.37424904e-01 2.85773188e-01 -7.96798885e-01 6.58776224e-01
7.53145516e-01 -6.75665617e-01 -9.22554672e-01 -2.56671786e-01
1.63889363e-01 -1.20174259e-01 4.21932526e-02 -1.00003660e+00
3.12516540e-01 -8.47519189e-02 1.55877128e-01 -7.30271101e-01
3.31845075e-01 -9.28248584e-01 -2.95825265e-02 3.99101287e-01
-1.22146621e-01 1.68015182e-01 2.01901302e-01 5.95065475e-01
-3.10629517e-01 -7.68751800e-01 5.66738844e-01 -2.36533880e-01
-9.64283645e-01 4.31314766e-01 -1.51213882e-02 2.40304880e-02
9.45358098e-01 -1.89556688e-01 -1.32678479e-01 -3.70339215e-01
-5.06519616e-01 4.23279405e-01 2.68306673e-01 8.57888639e-01
6.87403262e-01 -1.42957604e+00 -4.34113234e-01 5.68971224e-02
4.24808323e-01 1.02171481e-01 4.07871120e-02 -1.02399746e-02
-1.30528271e-01 4.69562888e-01 -3.47595632e-01 -7.43413493e-02
-1.12907600e+00 1.07637036e+00 3.03856313e-01 -5.95810115e-01
-4.05464232e-01 1.09230173e+00 1.87359586e-01 -9.05330300e-01
3.70776564e-01 -1.17579401e-01 -6.61904037e-01 1.22956961e-01
3.36697489e-01 1.67667389e-01 1.64145097e-01 -5.55185020e-01
-3.27559054e-01 4.19251472e-01 -2.71941036e-01 -4.73899990e-02
1.11966503e+00 1.53442416e-02 -2.37132385e-01 2.97960758e-01
9.96627510e-01 -9.52552781e-02 -4.78996694e-01 -5.53373873e-01
1.00763097e-01 -4.94700879e-01 -1.62373930e-01 -9.24562871e-01
-1.12538469e+00 8.68383765e-01 6.07144475e-01 2.45013595e-01
1.50999093e+00 -4.13921535e-01 1.00132644e+00 3.55022430e-01
5.69038391e-01 -8.51624370e-01 -6.24705628e-02 4.47811663e-01
7.46181965e-01 -8.61900508e-01 2.83913165e-02 -7.01383770e-01
-5.49409091e-01 1.13619518e+00 9.00160432e-01 2.21330300e-01
4.48228270e-01 -6.38416782e-02 -3.19866575e-02 -1.73879415e-01
-3.59922230e-01 -4.29225534e-01 2.61809379e-01 9.37211990e-01
2.55722731e-01 -2.77714641e-03 -7.26487875e-01 8.21558297e-01
-5.87172031e-01 -1.00990020e-01 3.27823073e-01 6.09363496e-01
-3.16552788e-01 -1.62161171e+00 -1.17348798e-01 4.93162751e-01
-5.53698577e-02 -2.15924382e-01 -4.26830709e-01 6.05351508e-01
7.09954083e-01 1.13320315e+00 -3.21994781e-01 -6.62515640e-01
4.66252416e-01 1.73463449e-01 8.65987837e-02 -9.54892397e-01
-5.59270084e-01 -1.63001746e-01 -1.24432489e-01 -1.38219029e-01
-2.49960959e-01 -2.92208523e-01 -1.37926579e+00 -1.16202772e-01
-6.74190640e-01 4.52565759e-01 4.53036994e-01 8.68412554e-01
2.41640463e-01 6.19226873e-01 5.64960301e-01 -8.59113224e-03
-6.23395860e-01 -1.07264566e+00 -6.08683169e-01 5.42160869e-01
-4.50756043e-01 -9.36535895e-01 -7.26127476e-02 -2.49757558e-01] | [9.25893497467041, 8.093376159667969] |
49ff4b31-7d84-4fb7-bdaa-a7538dae3776 | on-the-use-of-emojis-to-train-emotion | 1902.08906 | null | http://arxiv.org/abs/1902.08906v2 | http://arxiv.org/pdf/1902.08906v2.pdf | On the Use of Emojis to Train Emotion Classifiers | Nowadays, the automatic detection of emotions is employed by many
applications in different fields like security informatics, e-learning, humor
detection, targeted advertising, etc. Many of these applications focus on
social media and treat this problem as a classification problem, which requires
preparing training data. The typical method for annotating the training data by
human experts is considered time consuming, labor intensive and sometimes prone
to error. Moreover, such an approach is not easily extensible to new
domains/languages since such extensions require annotating new training data.
In this study, we propose a distant supervised learning approach where the
training sentences are automatically annotated based on the emojis they have.
Such training data would be very cheap to produce compared with the manually
created training data, thus, much larger training data can be easily obtained.
On the other hand, this training data would naturally have lower quality as it
may contain some errors in the annotation. Nonetheless, we experimentally show
that training classifiers on cheap, large and possibly erroneous data annotated
using this approach leads to more accurate results compared with training the
same classifiers on the more expensive, much smaller and error-free manually
annotated training data. Our experiments are conducted on an in-house dataset
of emotional Arabic tweets and the classifiers we consider are: Support Vector
Machine (SVM), Multinomial Naive Bayes (MNB) and Random Forest (RF). In
addition to experimenting with single classifiers, we also consider using an
ensemble of classifiers. The results show that using an automatically annotated
training data (that is only one order of magnitude larger than the manually
annotated one) gives better results in almost all settings considered. | ['Mohammed Al-Kabi', 'Mahmoud Al-Ayyoub', 'Wegdan Hussien', 'Yahya Tashtoush'] | 2019-02-24 | null | null | null | null | ['humor-detection'] | ['natural-language-processing'] | [ 2.05173880e-01 3.15091282e-01 -1.48884077e-02 -4.53984261e-01
-3.92337352e-01 -4.59968209e-01 5.04691899e-01 7.58403897e-01
-8.05705488e-01 9.73212481e-01 -3.70190620e-01 -3.83898407e-01
1.44232109e-01 -9.62664306e-01 -4.01227057e-01 -4.73564774e-01
3.91399264e-02 6.83237135e-01 3.10636461e-01 -3.74113321e-01
1.88413501e-01 2.45981291e-01 -1.83045721e+00 2.01099873e-01
9.71138954e-01 1.06315720e+00 3.50391865e-02 3.54906261e-01
-1.78855732e-01 9.63180125e-01 -6.89195335e-01 -8.76766503e-01
-3.35280225e-02 -2.64057785e-01 -1.09554648e+00 3.00939888e-01
-1.51789352e-01 8.51506144e-02 6.58052146e-01 1.05330813e+00
2.03112170e-01 8.07153806e-02 6.17592812e-01 -1.16764474e+00
-2.04244018e-01 4.90090013e-01 -2.60172546e-01 -2.53100187e-01
5.23404956e-01 -3.21992189e-01 8.76764297e-01 -7.49222457e-01
5.68448901e-01 8.85165632e-01 7.16838658e-01 4.03107911e-01
-1.05002224e+00 -4.29652780e-01 -1.64922178e-01 2.11069465e-01
-1.29586446e+00 -2.61848401e-02 8.10573220e-01 -5.28665781e-01
5.15796900e-01 5.26057363e-01 4.11583215e-01 1.20951414e+00
-2.28340238e-01 6.16443336e-01 1.54605615e+00 -8.65783334e-01
3.67438704e-01 1.09081161e+00 3.41653526e-01 6.29745901e-01
1.60166264e-01 -2.02376112e-01 -9.99223813e-02 -3.83938015e-01
4.53653596e-02 -1.94409579e-01 -3.17409933e-01 1.57177113e-02
-7.51897275e-01 1.22278023e+00 1.18187368e-01 7.93570518e-01
-3.65096390e-01 -5.82674384e-01 5.67294538e-01 3.19009393e-01
6.17308795e-01 5.10567367e-01 -6.44872487e-01 -1.51500721e-02
-8.07071209e-01 -5.79185151e-02 1.20830929e+00 6.90885544e-01
8.79709423e-01 -2.68399835e-01 3.48508209e-01 1.02060699e+00
1.20668046e-01 1.78015754e-01 7.45685279e-01 -3.00243318e-01
2.57374704e-01 6.69099510e-01 3.97381157e-01 -1.26520634e+00
-6.24075055e-01 -1.97927848e-01 -7.75989056e-01 8.50573778e-02
7.96396255e-01 -3.45310062e-01 -4.10112381e-01 1.42849016e+00
4.81202245e-01 -2.86530375e-01 2.15883732e-01 8.06994259e-01
8.05725276e-01 5.98028779e-01 2.57978022e-01 -6.06295168e-01
1.43758583e+00 -8.08912754e-01 -8.53623688e-01 -6.35047555e-02
9.14074838e-01 -1.01706147e+00 1.19919801e+00 8.66429508e-01
-5.34771204e-01 -4.81058806e-01 -8.31679761e-01 2.68622130e-01
-7.83603311e-01 3.77372414e-01 5.92998445e-01 1.00310922e+00
-5.13469100e-01 6.11692131e-01 -4.05058473e-01 -5.78035891e-01
1.24902576e-02 3.34547281e-01 -4.23427910e-01 2.71587163e-01
-1.40139794e+00 1.18035007e+00 5.83232343e-01 -3.47729996e-02
-6.69589415e-02 1.08343467e-01 -8.56440067e-01 -9.14176181e-02
4.32145804e-01 5.61767705e-02 1.13994837e+00 -1.68567514e+00
-1.51150382e+00 9.82487023e-01 8.22624639e-02 -4.55713391e-01
4.81306255e-01 1.98462941e-02 -6.17907584e-01 -1.44606262e-01
-6.84463233e-02 2.77740479e-01 6.93502367e-01 -1.32110620e+00
-4.35082585e-01 -2.66098619e-01 1.11294270e-01 -2.39689395e-01
-6.94262683e-01 3.36440146e-01 1.70711830e-01 -4.89712864e-01
-2.09793970e-01 -1.06075215e+00 -1.94936231e-01 -4.95017052e-01
-3.84463787e-01 -3.27647090e-01 6.61888540e-01 -7.29087412e-01
1.27233005e+00 -1.99341571e+00 -1.24563418e-01 3.91067892e-01
-8.57623890e-02 3.42345804e-01 3.15261841e-01 2.96917289e-01
-6.64089397e-02 2.31313512e-01 -2.62950599e-01 -1.72776327e-01
-2.38509309e-02 3.66168946e-01 -8.99268538e-02 1.09993309e-01
2.16317460e-01 2.69512534e-01 -8.61764312e-01 -8.22179019e-01
-6.36290014e-02 3.59706312e-01 -3.76154482e-01 1.68402463e-01
-1.12873182e-01 4.44301307e-01 -2.68868059e-01 3.49506676e-01
4.01957273e-01 -1.14461318e-01 4.39319670e-01 -3.14909145e-02
-4.34606187e-02 -6.36877790e-02 -1.27269411e+00 8.73181999e-01
-9.11375940e-01 4.87676203e-01 -2.12496787e-01 -1.24045205e+00
1.23809731e+00 5.05873621e-01 3.10388863e-01 -2.82450259e-01
5.96669257e-01 4.17905331e-01 1.45858293e-02 -6.94308341e-01
4.91183221e-01 -4.26423520e-01 -2.36065879e-01 4.44929212e-01
5.90707548e-02 4.55823839e-02 2.82923520e-01 -1.24801062e-01
7.88594365e-01 -1.53609440e-01 6.46568656e-01 -4.80224378e-02
8.68915498e-01 1.16334945e-01 4.58472520e-01 3.94253790e-01
-1.37233660e-01 1.03076085e-01 4.56955940e-01 -5.47168374e-01
-8.94093156e-01 -2.02006534e-01 -3.97994012e-01 9.61671054e-01
1.67822707e-02 -4.72521305e-01 -8.85379076e-01 -1.00053024e+00
-4.95377332e-01 8.03460836e-01 -4.22213316e-01 2.34084249e-01
-1.70172364e-01 -1.13504136e+00 3.86357307e-01 1.22252323e-01
4.35366482e-01 -1.19046605e+00 -7.46998072e-01 2.27056846e-01
-3.75570327e-01 -1.19389451e+00 2.90822446e-01 3.22826326e-01
-7.26821125e-01 -1.01075184e+00 -2.77146786e-01 -6.39562666e-01
7.20612347e-01 -2.35219792e-01 1.09071219e+00 4.54593301e-01
3.52511555e-02 -4.73483391e-02 -9.05710459e-01 -6.95708036e-01
-6.64297938e-01 6.49858117e-02 8.83109719e-02 4.52653050e-01
7.71593094e-01 -2.93118984e-01 6.40856922e-02 4.20267195e-01
-1.08949041e+00 -1.32947385e-01 6.06060624e-01 1.10738826e+00
1.12010889e-01 4.27449346e-01 6.96499348e-01 -1.39850271e+00
6.16547883e-01 -5.00460029e-01 -4.03337657e-01 2.04003364e-01
-5.02290845e-01 -4.29023942e-03 1.01464391e+00 -6.18395209e-01
-1.13545382e+00 2.57709712e-01 -3.70083213e-01 1.95743695e-01
-5.85852027e-01 6.62982762e-01 -9.71505195e-02 -2.10272059e-01
1.07260108e+00 -1.70353413e-01 -1.13976292e-01 -4.00399715e-01
-1.16514236e-01 1.13175583e+00 -1.29365832e-01 -4.11577314e-01
6.95623815e-01 8.38863179e-02 -1.10823818e-01 -8.92953038e-01
-1.00065744e+00 -3.00630689e-01 -7.91032434e-01 -3.84245932e-01
7.42331028e-01 -2.61300206e-01 -5.71677923e-01 2.01727718e-01
-1.16617036e+00 -1.10950164e-01 5.65744787e-02 5.73925316e-01
-3.10975701e-01 5.13069749e-01 -4.33763653e-01 -1.12211692e+00
-2.57386923e-01 -9.47570920e-01 7.37797439e-01 1.64162353e-01
-5.95955551e-01 -1.04577899e+00 -1.03503518e-01 6.16483927e-01
1.45390898e-01 3.10053706e-01 9.49068010e-01 -1.17491889e+00
1.05448909e-01 -6.15369678e-01 6.13484271e-02 5.37526786e-01
-3.34137306e-02 1.51947439e-01 -1.02054763e+00 3.40977833e-02
2.22596094e-01 -6.60663784e-01 3.77891243e-01 -3.35459590e-01
9.97862816e-01 -4.56926256e-01 -8.72547552e-03 -3.38321507e-01
1.36781502e+00 2.04906121e-01 5.36943436e-01 4.12619770e-01
1.54657468e-01 1.14562654e+00 1.13113689e+00 4.79232103e-01
2.45274425e-01 8.17478001e-01 2.14417979e-01 -3.77473682e-02
5.82988322e-01 1.78982705e-01 3.27639371e-01 7.62018979e-01
-3.73960167e-01 -1.11350149e-01 -9.34650123e-01 2.93288022e-01
-1.79189920e+00 -9.95728195e-01 -4.51868743e-01 2.38777828e+00
1.15799439e+00 1.62292808e-01 3.14629674e-01 8.66242111e-01
7.82862186e-01 -4.13825631e-01 1.34855643e-01 -8.75926733e-01
6.66913614e-02 4.42502350e-01 1.82380751e-01 5.28962493e-01
-1.23664939e+00 6.65757179e-01 4.80123520e+00 8.58688593e-01
-1.21419382e+00 3.75589460e-01 8.13233912e-01 3.08389783e-01
6.81885034e-02 -8.14037398e-02 -6.66567743e-01 5.57117820e-01
1.07888150e+00 1.96108997e-01 1.83851406e-01 1.12524581e+00
9.56830904e-02 -6.14537537e-01 -8.16552222e-01 9.67537165e-01
9.63025242e-02 -7.36725628e-01 -3.93379182e-01 -1.58518150e-01
5.26596606e-01 -4.88096178e-01 -3.95761102e-01 4.55115825e-01
-5.99901043e-02 -8.23311329e-01 5.67617595e-01 3.04680556e-01
2.96273977e-01 -8.18988323e-01 1.38685250e+00 8.26287687e-01
-5.50640881e-01 -1.24364428e-01 -2.91190267e-01 -2.91431308e-01
2.91383322e-02 8.72150958e-01 -9.04529095e-01 5.26542485e-01
7.30591774e-01 1.42983437e-01 -7.31572449e-01 9.00109708e-01
-2.68640399e-01 5.90996861e-01 -2.76443481e-01 -6.57610297e-01
7.52598122e-02 -1.58724323e-01 1.85734481e-01 1.19592941e+00
2.50713438e-01 5.22685284e-03 2.33206749e-01 3.62906128e-01
1.91735476e-01 9.57797527e-01 -6.18125439e-01 1.89874366e-01
1.02688037e-02 1.59172690e+00 -9.90615904e-01 -4.92329538e-01
-5.30749381e-01 8.26319158e-01 3.45082104e-01 -7.48016089e-02
-9.71369863e-01 -5.72266638e-01 -2.09850162e-01 7.63884932e-02
2.43623778e-02 1.92886069e-01 -2.18302965e-01 -1.14747834e+00
2.64045000e-02 -9.01865840e-01 3.97398829e-01 -5.71325779e-01
-1.51510513e+00 1.12824321e+00 -1.46412745e-01 -1.30076611e+00
-4.05489951e-01 -8.95516753e-01 -2.47616827e-01 6.16509497e-01
-1.32475352e+00 -7.41463423e-01 -5.01811624e-01 3.26769292e-01
3.46363306e-01 6.75624534e-02 1.13796926e+00 6.12963259e-01
-4.59187090e-01 4.27612305e-01 -2.90095270e-01 2.76121974e-01
8.22475612e-01 -1.19929934e+00 -6.56860352e-01 5.46488345e-01
2.30850190e-01 3.65383327e-01 8.45960617e-01 -5.01094460e-01
-6.01273417e-01 -9.09096837e-01 1.41532969e+00 -3.83974463e-01
9.06680584e-01 -2.87390143e-01 -9.47632968e-01 3.62562031e-01
1.99969023e-01 -4.16487418e-02 1.13620234e+00 2.30475709e-01
-1.24783300e-01 4.73157763e-02 -1.32046962e+00 2.04093099e-01
3.40176105e-01 -2.13791832e-01 -5.34664631e-01 6.92086220e-01
1.07047915e-01 -1.40011564e-01 -9.67077971e-01 3.40526432e-01
4.20900524e-01 -1.05510855e+00 3.01314622e-01 -6.93958879e-01
4.79590178e-01 -3.30618799e-01 -4.38604280e-02 -1.16209495e+00
2.97805965e-01 -2.69279242e-01 3.67305994e-01 1.45811975e+00
7.41846681e-01 -5.31119227e-01 6.84735000e-01 8.25221419e-01
3.47807109e-01 -8.67768347e-01 -6.45498514e-01 -7.45964229e-01
-1.28692925e-01 -5.04578769e-01 2.63111591e-01 1.40812886e+00
4.30634111e-01 6.85311794e-01 -5.64760983e-01 -4.56343032e-02
2.49567851e-01 2.28135273e-01 8.16749156e-01 -1.64145148e+00
-3.74955148e-01 -8.55321363e-02 -6.02916777e-01 -3.90343130e-01
3.61086220e-01 -6.82548702e-01 5.24124987e-02 -8.92713249e-01
4.62672766e-03 -8.48960578e-01 2.71053344e-01 6.39563143e-01
-1.38287261e-01 6.04960799e-01 6.26079813e-02 2.20188815e-02
-3.92284065e-01 2.22449854e-01 7.67356038e-01 6.41506398e-03
-1.89303786e-01 3.63251120e-01 -3.42785656e-01 1.05781281e+00
8.83262157e-01 -6.54260814e-01 -1.35194510e-01 1.92113668e-01
4.11240548e-01 -1.35492489e-01 2.53403902e-01 -7.72078574e-01
-5.01645030e-03 -8.90007392e-02 1.50605187e-01 -8.42938796e-02
2.56704926e-01 -1.13696229e+00 1.12875134e-01 3.32093835e-01
-1.30308077e-01 -1.42226875e-01 -9.50637981e-02 3.24947387e-01
-4.28487748e-01 -1.05260825e+00 8.85839462e-01 -1.99687898e-01
-4.92759764e-01 -2.72970051e-01 -6.29360795e-01 -2.50032336e-01
1.31962013e+00 -1.29125848e-01 -8.86479542e-02 -5.03644109e-01
-1.01213336e+00 -2.56894588e-01 3.66389632e-01 1.84840411e-01
1.56999648e-01 -1.13531530e+00 -3.79352033e-01 -1.83812886e-01
1.91175357e-01 -3.10376972e-01 -7.78721422e-02 1.07941568e+00
-4.79615837e-01 2.26516590e-01 -1.89515531e-01 -3.94211888e-01
-1.56324279e+00 7.78440535e-01 -3.04417834e-02 -4.94082004e-01
8.22069794e-02 4.34065282e-01 -4.93561357e-01 -6.90882146e-01
1.10543624e-01 -8.19431245e-02 -8.62924755e-01 4.69291151e-01
2.83876687e-01 2.46304110e-01 4.01304752e-01 -9.37606514e-01
-1.59580052e-01 2.11190179e-01 1.87397078e-01 4.85856719e-02
1.29492009e+00 1.42485037e-01 -4.24782753e-01 5.92586935e-01
8.64785671e-01 2.98387676e-01 -2.62363881e-01 -1.22089908e-01
4.84665990e-01 -4.22404706e-01 -1.09555162e-01 -5.55435061e-01
-7.21227527e-01 8.24455023e-01 3.07292193e-01 8.88865411e-01
1.20934570e+00 -1.04752079e-01 4.92551297e-01 5.37737250e-01
6.62856638e-01 -1.47819364e+00 -3.66701446e-02 3.27285677e-01
5.99197030e-01 -1.57717454e+00 -4.44140173e-02 -8.43311489e-01
-8.78541350e-01 1.39938962e+00 5.63298106e-01 2.18725875e-01
6.94307446e-01 7.47404397e-02 1.49048090e-01 7.24868663e-03
-4.84887451e-01 -3.62475246e-01 1.80528075e-01 3.81833196e-01
6.46478951e-01 1.46619026e-02 -1.02452683e+00 9.51391637e-01
-3.26294512e-01 -4.95687015e-02 8.07747543e-01 8.26393127e-01
-3.44802797e-01 -1.43891263e+00 -5.83092511e-01 5.52650809e-01
-7.16627121e-01 6.02194807e-03 -5.64030349e-01 8.05569589e-01
3.76839995e-01 1.28627014e+00 -2.71133363e-01 -4.42903847e-01
1.90157577e-01 3.79543811e-01 2.17845395e-01 -6.45263016e-01
-7.85375118e-01 -2.24714980e-01 6.26470983e-01 -1.45577535e-01
-9.36755419e-01 -7.08858967e-01 -1.03675675e+00 -2.02287331e-01
-8.46203744e-01 5.99878788e-01 6.91538870e-01 1.18908072e+00
-9.83317867e-02 5.85763669e-03 8.75493526e-01 -7.48617113e-01
-3.73563260e-01 -1.10938227e+00 -3.88805181e-01 7.79569089e-01
-2.51652211e-01 -8.89205515e-01 -5.21758974e-01 1.90898076e-01] | [10.244097709655762, 8.927212715148926] |
b4e9c293-9123-4590-915b-5a4e33ddc3ba | classification-of-complex-systems-based-on | 2008.13503 | null | https://arxiv.org/abs/2008.13503v1 | https://arxiv.org/pdf/2008.13503v1.pdf | Classification of Complex Systems Based on Transients | In order to develop systems capable of modeling artificial life, we need to identify, which systems can produce complex behavior. We present a novel classification method applicable to any class of deterministic discrete space and time dynamical systems. The method distinguishes between different asymptotic behaviors of a system's average computation time before entering a loop. When applied to elementary cellular automata, we obtain classification results, which correlate very well with Wolfram's manual classification. Further, we use it to classify 2D cellular automata to show that our technique can easily be applied to more complex models of computation. We believe this classification method can help to develop systems, in which complex structures emerge. | ['Barbora Hudcova', 'Tomas Mikolov'] | 2020-08-31 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 1.29501522e-01 -2.61950254e-01 3.48458081e-01 3.66684198e-01
2.67598629e-01 -1.00419378e+00 9.75003004e-01 2.21486241e-01
-2.63651520e-01 9.20512319e-01 -2.89506763e-01 -7.41471529e-01
3.59781906e-02 -1.10757864e+00 -1.93624347e-01 -8.91629398e-01
-3.17082435e-01 4.82550591e-01 6.22274458e-01 -7.26546288e-01
2.39564493e-01 9.93892908e-01 -1.60313857e+00 -1.48157761e-01
6.15746915e-01 4.39585149e-01 3.06397583e-03 1.31370258e+00
8.24578404e-02 8.72295976e-01 -7.01032162e-01 3.75348687e-01
3.39640975e-02 -7.84645319e-01 -8.20401073e-01 2.39761304e-02
-5.28699934e-01 2.29503557e-01 -3.32965583e-01 7.34910131e-01
8.12052488e-02 1.66105499e-04 1.22723532e+00 -1.15371573e+00
-4.91389066e-01 1.69868708e-01 1.77522868e-01 3.88435274e-01
4.72417355e-01 1.15839243e-01 2.92537957e-01 -3.70623201e-01
6.51891530e-01 9.67168093e-01 7.34359384e-01 7.55589247e-01
-1.43208563e+00 -7.30755925e-03 -2.29380861e-01 -2.70289093e-01
-1.44286346e+00 -4.37306970e-01 4.13237363e-01 -5.76861501e-01
1.00889885e+00 5.19343734e-01 1.15713227e+00 4.54091251e-01
9.03355658e-01 4.00108516e-01 1.49662745e+00 -7.42820859e-01
5.60500920e-01 -1.27125219e-01 4.52263176e-01 9.13587987e-01
4.35448140e-01 2.67683506e-01 3.76259476e-01 -3.23634773e-01
9.06102717e-01 -1.17455207e-01 -4.41393517e-02 5.22814551e-03
-1.19154036e+00 3.97770077e-01 -2.40211576e-01 9.99109685e-01
-9.56357345e-02 4.26509112e-01 2.77102273e-02 5.19249797e-01
2.76199818e-01 6.08107924e-01 -1.84565455e-01 -3.52955848e-01
-3.81013334e-01 4.89074022e-01 1.34811401e+00 8.76328468e-01
4.72457290e-01 -8.48430544e-02 2.69952834e-01 3.64558995e-01
3.14188376e-02 5.99938750e-01 3.13253105e-01 -1.01614547e+00
-5.55951476e-01 7.19676554e-01 2.38443196e-01 -8.00445914e-01
-7.34250009e-01 -4.64518145e-02 -1.04919016e+00 2.95313716e-01
6.77677333e-01 -5.64719141e-02 -5.72642684e-01 1.43050706e+00
4.93999980e-02 -9.99341011e-02 4.52384561e-01 1.86249152e-01
2.30952919e-01 9.57993031e-01 -3.80304962e-01 -6.99675441e-01
1.05970228e+00 -4.21617508e-01 -7.76704848e-01 7.18267024e-01
8.68662834e-01 -5.42337358e-01 7.43774831e-01 3.76696438e-01
-1.14528930e+00 -1.38353556e-01 -1.05094147e+00 3.81358951e-01
-8.06412220e-01 -1.75764367e-01 7.22015500e-01 8.09933484e-01
-1.36981523e+00 6.60602093e-01 -1.28635693e+00 -7.28062630e-01
-3.16091299e-01 4.46045488e-01 9.59408581e-02 6.31673872e-01
-1.02972305e+00 1.03345084e+00 1.01966105e-01 9.49467649e-04
-6.80883706e-01 8.75490457e-02 -3.52171630e-01 -1.66684180e-01
-5.06572388e-02 -7.16099322e-01 1.19840407e+00 -6.79065287e-01
-1.61036468e+00 7.25398898e-01 -3.62566262e-01 -4.15233195e-01
2.75936186e-01 7.68465579e-01 -5.27012289e-01 1.02129810e-01
-2.79312700e-01 -1.01340197e-01 2.00430557e-01 -1.24675035e+00
-4.17629868e-01 -9.48530436e-02 1.03629470e-01 -1.77157909e-01
-1.18202962e-01 -1.33164108e-01 5.53382263e-02 -4.32322949e-01
-1.21193901e-01 -1.35377038e+00 -6.31545544e-01 -3.19056153e-01
-1.48060992e-01 -4.73707676e-01 6.90709174e-01 1.24557711e-01
1.46478236e+00 -1.77141452e+00 3.51166576e-01 4.48911875e-01
4.81839836e-01 1.03627192e-02 1.99749365e-01 8.49926591e-01
3.73382092e-01 2.51664162e-01 -2.87276864e-01 1.98947936e-01
-4.80025522e-02 4.60591227e-01 -1.83923990e-01 5.23416996e-01
1.19871095e-01 8.20715070e-01 -7.89245963e-01 -4.77518171e-01
4.42169718e-02 1.29599154e-01 -3.88721645e-01 -1.36361450e-01
-1.28939867e-01 2.50424474e-01 -5.40310502e-01 4.24598694e-01
2.34089509e-01 -3.57571632e-01 4.09774333e-02 6.37688875e-01
-5.18450320e-01 -2.16831997e-01 -8.06769609e-01 7.15118289e-01
-3.55503023e-01 7.11532712e-01 -2.02610970e-01 -1.01309502e+00
9.05875921e-01 3.08629245e-01 5.31309605e-01 -1.63009495e-01
4.31162119e-01 3.73245806e-01 3.72197658e-01 -2.09420025e-01
3.63636076e-01 -1.59955338e-01 -4.73897994e-01 9.57513869e-01
-2.35163182e-01 -7.16541708e-01 5.90494454e-01 2.45160848e-01
1.47156382e+00 -5.80567122e-01 4.95592743e-01 -1.02272379e+00
8.15360904e-01 1.91587642e-01 1.12991467e-01 9.08665657e-01
-2.67769843e-01 -9.75861177e-02 6.93426073e-01 -7.41317272e-01
-1.25230265e+00 -1.04452562e+00 -1.93757311e-01 6.42085016e-01
4.67275560e-01 -3.16411436e-01 -1.10062313e+00 3.42803523e-02
-2.22609282e-01 3.81796628e-01 -9.57711399e-01 -1.54190660e-01
-6.76058948e-01 -9.30331945e-01 6.49945259e-01 1.23164088e-01
2.61006683e-01 -9.91310477e-01 -7.47107804e-01 4.91756558e-01
4.24449384e-01 -7.01656103e-01 -1.62146568e-01 2.37181634e-01
-8.21503401e-01 -9.65991974e-01 -6.11063004e-01 -9.49532926e-01
8.08608234e-01 1.86887130e-01 8.60370040e-01 7.08203495e-01
-3.14925879e-01 6.02958679e-01 -2.76702404e-01 -2.19545603e-01
-1.37735581e+00 4.02096272e-01 4.27751660e-01 -5.61254442e-01
2.17434213e-01 -5.70639312e-01 -1.88545987e-01 6.58048391e-01
-1.20362163e+00 8.26978981e-02 -2.80717909e-02 8.51402402e-01
2.76059330e-01 3.28820825e-01 5.00474691e-01 -5.70072889e-01
9.86345232e-01 -3.14899921e-01 -7.22065151e-01 3.97947311e-01
-6.95424497e-01 2.99469203e-01 1.20606482e+00 -5.54388702e-01
-4.39807475e-01 1.89932540e-01 -9.80261434e-03 3.65058452e-01
-1.79222971e-01 1.65540487e-01 2.14506269e-01 -3.62380803e-01
5.66195309e-01 8.24074745e-01 3.84835720e-01 2.39651706e-02
9.14178491e-02 6.92589045e-01 2.03601792e-01 -6.32519245e-01
6.54404402e-01 4.46267486e-01 5.85202396e-01 -1.22877932e+00
8.31083506e-02 -6.84569357e-03 -6.96346164e-01 -4.02308792e-01
6.18573606e-01 -2.06457511e-01 -1.04294193e+00 8.46935630e-01
-1.19813657e+00 -6.68045640e-01 -3.10411125e-01 -6.86925203e-02
-9.85711515e-01 2.17649471e-02 -8.90260279e-01 -1.13929462e+00
1.29200578e-01 -8.64763618e-01 7.84342825e-01 3.58627200e-01
-2.94488549e-01 -1.46172655e+00 6.55892730e-01 -5.99413216e-01
4.60505217e-01 2.41431728e-01 9.71552432e-01 -5.60592771e-01
-3.47564340e-01 -2.05767915e-01 3.78027976e-01 -2.07289770e-01
1.84248462e-01 7.38805294e-01 -3.02332848e-01 -3.25794697e-01
-3.03285439e-02 3.34956646e-01 4.60993677e-01 2.03167200e-01
8.34311843e-01 -3.65875155e-01 -7.43540227e-01 1.20343789e-01
1.22970521e+00 8.40513587e-01 6.90657914e-01 1.10725217e-01
6.09793514e-02 3.89444917e-01 2.14122742e-01 2.56337553e-01
2.97562569e-01 4.45323437e-01 -2.09130749e-01 1.00659241e-03
3.21973175e-01 1.08300209e-01 2.48020872e-01 1.48682463e+00
-5.32457888e-01 -4.22176361e-01 -1.14615381e+00 3.82488966e-01
-1.81630898e+00 -1.24168098e+00 -3.53043407e-01 1.80348110e+00
8.07633400e-01 1.68507412e-01 5.71077466e-01 4.38576758e-01
8.35634530e-01 -3.65901589e-01 -3.30134481e-01 -7.86073029e-01
-3.91399384e-01 1.69373602e-01 4.50787276e-01 7.86318004e-01
-7.05005407e-01 9.25823033e-01 8.97426987e+00 7.71407962e-01
-1.00591803e+00 1.51209310e-02 5.26281774e-01 3.93333822e-01
-3.36715221e-01 5.90958707e-02 -5.89384973e-01 5.35681725e-01
1.17546523e+00 -7.35782623e-01 6.52775884e-01 1.83332354e-01
3.25061560e-01 -4.03706431e-01 -9.32307184e-01 6.38181865e-01
-4.44841117e-01 -1.34559357e+00 -1.12429112e-01 3.07637364e-01
8.92238855e-01 -6.14772558e-01 -2.16758996e-01 1.17092013e-01
5.92666388e-01 -1.05076361e+00 3.23181897e-01 7.87449002e-01
3.72231454e-01 -7.03090787e-01 4.21023071e-01 6.26282454e-01
-1.28058255e+00 5.94158545e-02 -2.54191488e-01 -6.93042576e-01
1.86128050e-01 3.81800860e-01 -6.57617390e-01 -4.44367081e-02
-2.34655142e-02 4.19580817e-01 -4.88520771e-01 8.83419752e-01
4.65589613e-01 4.77462977e-01 -5.89101315e-01 -1.13650048e+00
8.68463516e-02 -4.25908685e-01 4.37290937e-01 1.17646027e+00
5.81913710e-01 7.77790546e-01 -2.42088377e-01 7.52451718e-01
5.34672976e-01 1.39840227e-02 -1.12315309e+00 -3.31071913e-01
4.57986772e-01 9.09664631e-01 -1.55246198e+00 -6.76683724e-01
4.80100401e-02 8.70404899e-01 -1.28144026e-01 3.99578996e-02
-7.46451497e-01 -6.82439089e-01 4.25685257e-01 2.03386948e-01
-8.79313350e-02 -1.03035831e+00 -3.48208964e-01 -1.23522508e+00
-5.73962331e-01 -5.23574233e-01 -2.08133712e-01 -5.24311483e-01
-8.95973861e-01 7.20458210e-01 6.70522675e-02 -1.05077684e+00
-3.62699479e-01 -6.89619482e-01 -7.10396290e-01 3.77680868e-01
-3.60176086e-01 -7.19476402e-01 3.81423719e-02 5.81232131e-01
1.31718397e-01 -3.36513102e-01 9.81025934e-01 -1.34528503e-01
-4.47186619e-01 2.34705999e-01 6.03970408e-01 -1.49426147e-01
-1.43392742e-01 -1.39962578e+00 5.42874992e-01 7.12661684e-01
-5.56180701e-02 6.60544097e-01 1.12385583e+00 -4.13637370e-01
-1.65229845e+00 -7.34830976e-01 8.99779558e-01 -5.86987495e-01
8.30481231e-01 -5.11477172e-01 -6.17353618e-01 2.90848583e-01
1.06179409e-01 -2.28426129e-01 4.77288902e-01 -1.72081545e-01
3.02751482e-01 1.97620660e-01 -9.61395621e-01 9.85860407e-01
1.18733513e+00 -4.01125520e-01 -5.37093341e-01 2.83261448e-01
4.81560349e-01 -4.34897803e-02 -8.66250098e-01 9.99536961e-02
9.25284326e-01 -7.54733086e-01 4.06037778e-01 -4.18923229e-01
1.70034006e-01 -4.47987825e-01 1.15957074e-01 -1.21434772e+00
-4.03049558e-01 -1.01241922e+00 -6.40099272e-02 7.72907376e-01
4.48390484e-01 -1.23116982e+00 2.65916586e-01 5.12300193e-01
2.53828436e-01 -7.09378719e-01 -8.52206171e-01 -1.09158516e+00
6.36441886e-01 7.74401054e-02 6.14320517e-01 8.94227028e-01
6.28639698e-01 8.25787038e-02 9.13375914e-02 -2.50099480e-01
3.38497370e-01 1.89455882e-01 6.76331162e-01 -1.36948442e+00
-2.09078848e-01 -8.13253284e-01 -7.16642618e-01 -1.10428560e+00
8.86318311e-02 -4.54181969e-01 8.16450864e-02 -9.56463516e-01
-7.23613019e-04 -7.09425390e-01 -1.14664353e-01 1.39382929e-01
3.04538339e-01 3.00528854e-01 1.27876527e-03 4.48550493e-01
-5.66454887e-01 2.03189194e-01 1.21021914e+00 2.18320906e-01
-3.99354428e-01 1.54634848e-01 -3.26842368e-01 6.52304232e-01
8.05674255e-01 -4.28621471e-01 -1.18574992e-01 3.14074665e-01
2.92081594e-01 2.11171702e-01 1.48878574e-01 -1.17259479e+00
3.19151789e-01 -4.45256948e-01 -2.17911066e-03 -3.78572434e-01
-5.14666475e-02 -7.51956344e-01 5.85580170e-01 1.11431313e+00
-2.56642699e-01 4.03070718e-01 1.05048075e-01 3.78675163e-01
1.47110328e-01 -3.00721407e-01 8.27448428e-01 -1.06670208e-01
-1.47597685e-01 -5.43351881e-02 -1.72262931e+00 -2.19958022e-01
1.54474223e+00 -3.28515500e-01 -4.28769559e-01 -3.58495593e-01
-1.15097165e+00 3.23786438e-02 1.22654438e+00 -8.66019279e-02
1.71501145e-01 -1.34852934e+00 -1.57475442e-01 2.15411246e-01
-1.99023739e-01 -5.20143628e-01 -2.37059891e-01 8.44110548e-01
-1.33092546e+00 5.14730275e-01 -4.45094764e-01 -7.46002913e-01
-1.15284610e+00 5.41877925e-01 5.43894827e-01 -1.03072152e-01
-1.10754333e-01 4.22891453e-02 -1.12774454e-01 -6.89054281e-02
-5.35088003e-01 -6.72408640e-01 -2.31113043e-02 -2.80898362e-01
4.43295985e-01 4.50083315e-01 -3.45154375e-01 -5.72058201e-01
-3.87855411e-01 7.70387888e-01 3.70650619e-01 -4.01044369e-01
1.05265534e+00 -3.13376397e-01 -6.13612533e-01 1.06976366e+00
9.82989669e-01 2.70838626e-02 -6.64060175e-01 2.39859819e-01
-1.00209296e-01 4.35929820e-02 -4.58443731e-01 -2.36194104e-01
-3.31882834e-01 6.28565013e-01 3.68140250e-01 1.49314070e+00
1.23082232e+00 2.90780496e-02 5.61513007e-01 6.59661055e-01
7.81058311e-01 -9.25305367e-01 -3.84626567e-01 1.00663042e+00
6.22181296e-01 -6.03410661e-01 -1.82148114e-01 -3.18841428e-01
3.20873596e-02 1.55985892e+00 1.15193628e-01 -6.25769556e-01
1.00061989e+00 8.71980309e-01 -3.43375444e-01 -4.59280945e-02
-1.40097785e+00 -2.60873437e-01 -2.86301404e-01 4.54683989e-01
4.18652087e-01 1.81674153e-01 -9.28991258e-01 2.67432809e-01
-2.25606278e-01 4.31325510e-02 1.16993117e+00 1.17558980e+00
-8.06551695e-01 -1.06464756e+00 -3.24469745e-01 3.49796772e-01
-1.51532084e-01 3.19386691e-01 -7.86053300e-01 7.87842393e-01
9.09060054e-03 8.00891399e-01 2.48098522e-01 -4.31454182e-01
7.28180408e-02 -2.76343040e-02 8.42258692e-01 -3.09291065e-01
-4.16337490e-01 -1.65228695e-01 3.20657827e-02 -1.19157650e-01
-5.58256626e-01 -7.22913802e-01 -1.19544494e+00 -9.75336015e-01
-3.25737298e-01 5.87501764e-01 2.57623553e-01 9.82002735e-01
1.26823083e-01 2.31692389e-01 7.52750218e-01 -7.18702137e-01
-5.61470315e-02 -5.26331246e-01 -8.92820776e-01 -5.21928556e-02
4.16718721e-01 -5.89704752e-01 -5.29367745e-01 3.77612948e-01] | [5.593572616577148, 4.19366979598999] |
16b6de23-0d44-4b89-aca7-81fa9e2904b7 | pvt-a-simple-end-to-end-latency-aware-visual | 2211.11629 | null | https://arxiv.org/abs/2211.11629v2 | https://arxiv.org/pdf/2211.11629v2.pdf | PVT++: A Simple End-to-End Latency-Aware Visual Tracking Framework | Visual object tracking is essential to intelligent robots. Most existing approaches have ignored the online latency that can cause severe performance degradation during real-world processing. Especially for unmanned aerial vehicles (UAVs), where robust tracking is more challenging and onboard computation is limited, the latency issue can be fatal. In this work, we present a simple framework for end-to-end latency-aware tracking, i.e., end-to-end predictive visual tracking (PVT++). Unlike existing solutions that naively append Kalman Filters after trackers, PVT++ can be jointly optimized, so that it takes not only motion information but can also leverage the rich visual knowledge in most pre-trained tracker models for robust prediction. Besides, to bridge the training-evaluation domain gap, we propose a relative motion factor, empowering PVT++ to generalize to the challenging and complex UAV tracking scenes. These careful designs have made the small-capacity lightweight PVT++ a widely effective solution. Additionally, this work presents an extended latency-aware evaluation benchmark for assessing an any-speed tracker in the online setting. Empirical results on a robotic platform from the aerial perspective show that PVT++ can achieve significant performance gain on various trackers and exhibit higher accuracy than prior solutions, largely mitigating the degradation brought by latency. Our code will be made public. | ['Changhong Fu', 'Hang Zhao', 'Sebastian Scherer', 'Yiming Li', 'Junjie Ye', 'Ziyuan Huang', 'Bowen Li'] | 2022-11-21 | null | null | null | null | ['visual-tracking', 'visual-object-tracking'] | ['computer-vision', 'computer-vision'] | [-1.01039596e-01 -5.02500474e-01 -2.82444894e-01 1.52993947e-01
-3.83936286e-01 -9.35175538e-01 4.30987120e-01 -7.72586092e-02
-4.33355033e-01 2.89765924e-01 -3.26398820e-01 -4.14385825e-01
-1.54013738e-01 -1.56207964e-01 -7.49648809e-01 -4.91850227e-01
-3.71935487e-01 4.36685570e-02 7.07919300e-01 3.71719562e-02
-2.06733868e-01 5.42388678e-01 -1.61270738e+00 -3.13154042e-01
6.50070906e-01 1.20401096e+00 5.25726318e-01 8.70447099e-01
3.99641007e-01 6.39733732e-01 -5.19270241e-01 -1.02062143e-01
7.44867980e-01 3.48738581e-01 -2.99668133e-01 7.72066712e-02
7.55071580e-01 -6.75750852e-01 -4.24905211e-01 8.83591533e-01
5.50057352e-01 6.36509759e-03 1.86744526e-01 -1.79276371e+00
-1.32221922e-01 7.56581873e-02 -6.56191468e-01 1.13970399e-01
1.59121424e-01 7.26157367e-01 5.81833661e-01 -7.41209090e-01
4.84842986e-01 1.19717324e+00 9.95011210e-01 3.27816904e-01
-9.58600223e-01 -7.79410839e-01 6.24386668e-01 7.22064525e-02
-1.30145717e+00 -4.47144061e-01 1.85346231e-01 -6.36441767e-01
8.11159432e-01 8.40273499e-02 7.83102274e-01 8.29026282e-01
4.16414052e-01 9.00045335e-01 6.99097872e-01 -4.27487232e-02
9.75090191e-02 -2.33493164e-01 1.92183051e-02 6.32080317e-01
6.44967020e-01 5.95891595e-01 -4.27404910e-01 -1.19552404e-01
4.31269139e-01 2.54929870e-01 -3.48516464e-01 -8.32766056e-01
-1.59503198e+00 2.97679156e-01 5.79032600e-01 -3.72883350e-01
-4.65949267e-01 4.95643914e-01 6.27140343e-01 3.42796504e-01
2.39852160e-01 1.71807542e-01 -7.08208919e-01 -2.86764324e-01
-1.04494917e+00 2.79718399e-01 6.88469589e-01 1.56532991e+00
3.86184752e-01 3.22532415e-01 -4.59947169e-01 5.54709276e-03
5.24268568e-01 9.87249970e-01 6.78431019e-02 -1.07081115e+00
3.06853145e-01 3.81626368e-01 6.29156291e-01 -8.35208833e-01
-5.29657602e-01 -6.00486398e-01 -4.24148500e-01 5.69751441e-01
3.72641742e-01 -4.02978033e-01 -7.70682037e-01 1.67611778e+00
8.42104912e-01 3.28124881e-01 7.75357038e-02 1.21703804e+00
5.60305059e-01 3.79377335e-01 1.25713423e-01 -2.55387783e-01
1.42613196e+00 -1.29332471e+00 -6.25997365e-01 -4.31345522e-01
6.17954016e-01 -9.59615827e-01 6.46731913e-01 3.26367676e-01
-6.31373227e-01 -5.40529013e-01 -1.16682100e+00 3.55422437e-01
1.94197930e-02 4.88855124e-01 6.65729880e-01 6.21596456e-01
-1.19063067e+00 4.47905898e-01 -1.30524409e+00 -6.49441838e-01
2.57505834e-01 5.45876443e-01 -1.20501146e-01 3.37887779e-02
-4.88444239e-01 9.70558763e-01 3.12827051e-01 5.71968295e-02
-1.25879490e+00 -9.00494039e-01 -4.94704366e-01 -1.48686752e-01
9.15965319e-01 -7.57043064e-01 1.71239662e+00 -5.86345077e-01
-1.41344774e+00 9.58033502e-02 -8.50628614e-02 -7.05040395e-01
6.98360622e-01 -6.19037449e-01 -1.12814002e-01 -1.02302335e-01
-8.40014406e-03 8.59184980e-01 1.13570499e+00 -1.07233918e+00
-8.59490156e-01 -3.92085947e-02 1.74248248e-01 1.97214022e-01
-5.21107793e-01 7.67551139e-02 -7.76061714e-01 -6.18533432e-01
-1.68748781e-01 -1.42514503e+00 -2.56873339e-01 6.46753907e-01
6.37551248e-02 -9.10133868e-02 1.42501187e+00 -3.60673845e-01
1.01546013e+00 -2.14219093e+00 -1.43393487e-01 -2.64229894e-01
2.74033129e-01 5.86005151e-01 -1.94555223e-01 3.28391612e-01
4.19393301e-01 -5.59483290e-01 2.97126502e-01 -2.84504533e-01
2.13854894e-01 3.87574807e-02 -5.24010181e-01 6.96152747e-01
-6.28436282e-02 7.99619377e-01 -9.97076094e-01 -4.55203980e-01
3.08823496e-01 3.34749401e-01 -4.92384613e-01 5.23932949e-02
-4.66603816e-01 3.54026049e-01 -4.72076535e-01 9.26990986e-01
6.87586546e-01 -4.43788379e-01 1.65407434e-01 -2.55095363e-01
-4.76017445e-01 -1.38431668e-01 -1.20731723e+00 1.49821734e+00
-3.32629442e-01 7.28619874e-01 4.50595707e-01 -2.12828934e-01
5.78288078e-01 1.63211316e-01 5.77254236e-01 -3.05699497e-01
3.02285492e-01 -1.82669275e-02 -3.78201716e-02 -5.03487661e-02
8.70461345e-01 5.97297430e-01 8.45883712e-02 2.17533112e-02
-2.29409114e-01 1.63460493e-01 2.59355098e-01 2.69528002e-01
1.42537415e+00 6.69050455e-01 2.44626880e-01 -1.66649818e-01
1.52429696e-02 5.37951410e-01 8.56036544e-01 7.69670904e-01
-5.61299622e-01 -7.18501806e-02 -1.04187801e-01 -2.83157647e-01
-7.30026484e-01 -9.19945478e-01 5.60628362e-02 1.07957125e+00
6.17777050e-01 -6.57661021e-01 -3.77733558e-01 -6.13083899e-01
2.46832222e-01 1.69404924e-01 -6.71399310e-02 -1.24558546e-01
-3.00206810e-01 -4.63062137e-01 3.94808263e-01 7.17427373e-01
3.67437720e-01 -3.78404558e-01 -1.37863064e+00 3.02803338e-01
1.41845509e-01 -1.56745386e+00 -7.28940427e-01 -4.81120944e-02
-9.83365417e-01 -1.01497447e+00 -5.22749543e-01 -4.97647196e-01
3.88657123e-01 1.10790205e+00 7.80567646e-01 1.01974323e-01
-2.89558053e-01 8.63571703e-01 -5.17336965e-01 -8.14422488e-01
2.15411969e-02 -1.53439760e-01 5.05649686e-01 -3.24330211e-01
-3.00466381e-02 -1.31649628e-01 -7.20164835e-01 6.03356361e-01
-4.46854621e-01 -3.00582349e-02 6.47299230e-01 6.73408449e-01
6.33388877e-01 -4.07654755e-02 1.78015847e-02 -2.59058867e-02
4.01681662e-02 -2.98026532e-01 -1.39621377e+00 1.80789903e-01
-7.54965425e-01 -2.76498079e-01 6.44433200e-01 -9.09486055e-01
-6.39665961e-01 5.09555519e-01 5.07783353e-01 -1.06078064e+00
3.25501770e-01 2.15368062e-01 7.04794973e-02 -6.65014029e-01
3.38119447e-01 4.11021672e-02 1.73238605e-01 -1.35981768e-01
2.52178192e-01 2.39965126e-01 5.98679721e-01 -2.48095825e-01
1.20886445e+00 5.13981104e-01 2.00362042e-01 -7.52570987e-01
-5.70294261e-01 -5.73166788e-01 -2.64679402e-01 -5.52416086e-01
5.86401939e-01 -1.49742532e+00 -1.16174543e+00 2.43121400e-01
-1.03862834e+00 -4.34547693e-01 -1.30627841e-01 8.84884953e-01
-3.66028190e-01 4.94368792e-01 -3.01774472e-01 -9.52013433e-01
-4.66233730e-01 -1.34184396e+00 1.18205893e+00 3.21363062e-01
6.91841021e-02 -6.52929723e-01 -3.25674750e-02 -8.88607875e-02
5.39487720e-01 2.00100183e-01 -2.07927331e-01 -2.96453029e-01
-1.29295707e+00 -3.12115431e-01 -3.93634051e-01 -1.10588714e-01
-2.04594567e-01 1.13959253e-01 -8.72244477e-01 -1.12436366e+00
-4.17555273e-01 -8.85426104e-02 6.53049171e-01 3.17991287e-01
5.99832952e-01 -2.71262944e-01 -8.06052506e-01 8.49870145e-01
1.27939808e+00 1.11981526e-01 -8.50605369e-02 4.96633440e-01
7.38414884e-01 1.67837590e-02 1.50205421e+00 6.60107017e-01
5.17943323e-01 1.15166116e+00 8.85273039e-01 6.03769720e-02
-1.79626659e-01 -2.09445916e-02 9.46725368e-01 6.03530288e-01
1.28544748e-01 -3.65087956e-01 -7.06803381e-01 4.59938079e-01
-2.09786677e+00 -6.68327212e-01 -1.15108557e-01 2.28761601e+00
2.81220764e-01 4.81335297e-02 2.58169085e-01 -2.55538762e-01
5.45739174e-01 1.31804928e-01 -7.67952323e-01 2.93255568e-01
2.18840286e-01 -3.93235028e-01 1.09997559e+00 1.26474932e-01
-1.32449782e+00 1.05137980e+00 5.85988712e+00 6.73236489e-01
-1.16666067e+00 2.95615405e-01 -5.55943668e-01 -4.42817628e-01
5.08192480e-01 3.31504673e-01 -1.14252710e+00 4.77387846e-01
1.12048209e+00 -1.68137804e-01 2.15691850e-01 1.25338185e+00
1.57206699e-01 -8.80161151e-02 -9.89680052e-01 9.99939382e-01
-2.58101463e-01 -1.02334213e+00 -4.05677557e-01 2.06092775e-01
3.70879859e-01 5.05077839e-01 1.34191856e-01 3.02241623e-01
6.60771847e-01 -3.87550831e-01 9.96264517e-01 3.20456624e-01
6.31765723e-01 -4.21628833e-01 5.83538055e-01 3.69578421e-01
-1.80409932e+00 -2.35334009e-01 -5.67132652e-01 -2.47522984e-02
3.25936496e-01 3.64346147e-01 -1.06610060e+00 7.86871910e-01
9.70397711e-01 8.76151264e-01 -6.95667446e-01 1.61669290e+00
5.25882952e-02 4.42756146e-01 -5.69498360e-01 -1.15536518e-01
3.17986012e-01 2.48223111e-01 9.01031256e-01 9.54651535e-01
6.60776198e-01 -1.30944610e-01 8.40945363e-01 1.73610404e-01
2.05433100e-01 -2.30397001e-01 -6.87075138e-01 1.68742202e-02
7.39228845e-01 1.50143528e+00 -7.64363229e-01 -2.16010332e-01
-5.45520663e-01 7.13861585e-01 6.63838312e-02 1.31983776e-02
-1.26948154e+00 -6.66027442e-02 1.03877795e+00 -1.32589221e-01
7.33659089e-01 -6.20934486e-01 1.50843784e-01 -1.03965533e+00
1.20929703e-01 -7.42642283e-01 1.16820499e-01 -6.97539747e-01
-7.70076871e-01 5.67196965e-01 -3.30408327e-02 -1.96744549e+00
1.26785338e-01 -8.24065089e-01 -2.71683127e-01 3.08706492e-01
-1.49281931e+00 -1.39476097e+00 -5.43292403e-01 4.75237936e-01
4.62269545e-01 -1.19079046e-01 3.11726987e-01 3.59529465e-01
-6.27756536e-01 6.75402999e-01 1.37920275e-01 -2.20230386e-01
9.67117786e-01 -8.57595801e-01 4.79612410e-01 1.37431133e+00
-1.22351751e-01 5.56724846e-01 1.01797962e+00 -8.16027462e-01
-2.27235961e+00 -1.56611836e+00 2.48471141e-01 -7.99431860e-01
7.42895365e-01 -1.55453995e-01 -5.02772033e-01 8.19586396e-01
1.85735181e-01 3.14155817e-01 1.91357300e-01 -2.08504975e-01
-2.38247365e-01 -2.52493978e-01 -7.39057839e-01 6.85147643e-01
1.26975346e+00 -4.51291613e-02 -1.12435512e-01 5.11259079e-01
1.05686426e+00 -8.18465233e-01 -1.00330830e+00 5.46963274e-01
6.87043905e-01 -5.23895502e-01 1.01596022e+00 -1.31487206e-01
-5.53935826e-01 -1.05015230e+00 -1.43059164e-01 -1.11574483e+00
-3.64896178e-01 -9.79370117e-01 -5.60428441e-01 1.01510715e+00
1.27204061e-02 -5.72395980e-01 5.87020397e-01 3.47299695e-01
-4.13905174e-01 -3.12218904e-01 -8.31795931e-01 -1.43425727e+00
-6.61491811e-01 -3.62439126e-01 4.94522363e-01 5.34437120e-01
-4.38468993e-01 -2.02935934e-02 -7.92505801e-01 8.08447957e-01
8.44516635e-01 2.03190222e-01 1.26220369e+00 -1.20245695e+00
-2.86447346e-01 -5.86934164e-02 -4.87558901e-01 -1.46906364e+00
-1.32019818e-01 -6.41717732e-01 3.54323208e-01 -1.13483965e+00
-4.55563590e-02 -4.20112997e-01 5.80986515e-02 5.57473600e-01
-8.03561881e-02 1.52202323e-01 7.15436161e-01 4.42600310e-01
-1.20978749e+00 6.71812177e-01 1.15494132e+00 -1.25024751e-01
-7.48187751e-02 -2.24441234e-02 -3.33877116e-01 6.90063119e-01
4.81271833e-01 -5.51835656e-01 -4.79770124e-01 -6.56576276e-01
-7.28703737e-02 7.13264570e-02 6.65407658e-01 -1.47975492e+00
7.25075305e-01 -1.74434409e-01 3.21538538e-01 -7.56238580e-01
4.56094682e-01 -1.30612195e+00 2.68620580e-01 8.39230239e-01
3.45887035e-01 5.70105433e-01 5.99820673e-01 9.22107458e-01
6.71077967e-02 2.63055712e-01 6.04077756e-01 4.40968990e-01
-1.07897985e+00 7.28053987e-01 -3.35314989e-01 -1.61781743e-01
1.49270117e+00 -3.05977434e-01 -6.97999120e-01 -1.63940594e-01
-1.85390070e-01 6.98127866e-01 8.71081531e-01 5.65403223e-01
4.24926013e-01 -1.17522597e+00 -2.23053321e-01 -8.01129118e-02
2.45369643e-01 -2.15188041e-01 2.08113447e-01 1.36822414e+00
-4.70951527e-01 4.58865136e-01 -1.06959775e-01 -9.73955512e-01
-1.52866697e+00 9.87094462e-01 -5.79909869e-02 -1.66811142e-03
-8.00540745e-01 5.59681654e-01 1.60061434e-01 -1.77379236e-01
4.68201011e-01 -4.63970333e-01 1.97051704e-01 -1.68791771e-01
6.36717558e-01 4.30064350e-01 -7.07092732e-02 -5.81812918e-01
-7.03475296e-01 5.85824728e-01 -1.19354032e-01 2.32628495e-01
8.51260841e-01 -2.53732383e-01 3.30391824e-01 -8.33214372e-02
5.06165802e-01 -4.89528701e-02 -1.99539196e+00 -1.73628926e-01
1.90145373e-02 -4.93384808e-01 2.55867749e-01 -5.77368259e-01
-1.13128006e+00 4.56768483e-01 9.65349615e-01 -1.11826032e-01
1.07156253e+00 -3.56297165e-01 8.39808345e-01 6.17570460e-01
8.90246987e-01 -7.91113496e-01 -1.18204258e-01 6.72349274e-01
5.04594088e-01 -1.21303248e+00 2.54724413e-01 -3.87684703e-01
-6.43504798e-01 8.37325335e-01 8.96428943e-01 -3.13345194e-02
4.06651378e-01 6.25383854e-01 5.23011349e-02 1.28350466e-01
-1.15516651e+00 -3.64608109e-01 9.49139670e-02 5.91128051e-01
-2.60546710e-02 -1.21408880e-01 6.09135702e-02 3.30444247e-01
4.22481894e-02 1.48549184e-01 3.21360767e-01 1.34935129e+00
-4.13119465e-01 -8.05443883e-01 -6.77602708e-01 1.08954221e-01
-1.56363189e-01 1.38486221e-01 -6.78549930e-02 9.45974469e-01
-5.65726273e-02 1.06753254e+00 -1.22960806e-01 -6.42593741e-01
4.61738020e-01 -4.79350656e-01 4.23233569e-01 -1.94785908e-01
-6.13189757e-01 9.02787670e-02 9.69015807e-02 -1.04065645e+00
-4.30308908e-01 -6.35162115e-01 -1.01909459e+00 -4.69558775e-01
-5.00118375e-01 -6.32756799e-02 7.18362153e-01 7.30505764e-01
8.61486554e-01 7.37666547e-01 2.54268527e-01 -1.28952956e+00
-8.54029834e-01 -5.96919537e-01 -1.67459637e-01 -3.28313410e-01
6.31138504e-01 -1.23609579e+00 -7.19821006e-02 -5.33656515e-02] | [6.396407127380371, -2.1103551387786865] |
3425c567-27f1-46fd-9019-7b0cb03a71b5 | deep-neural-networks-for-the-sequential | 2006.05587 | null | https://arxiv.org/abs/2006.05587v3 | https://arxiv.org/pdf/2006.05587v3.pdf | Sequential Density Ratio Estimation for Simultaneous Optimization of Speed and Accuracy | Classifying sequential data as early and as accurately as possible is a challenging yet critical problem, especially when a sampling cost is high. One algorithm that achieves this goal is the sequential probability ratio test (SPRT), which is known as Bayes-optimal: it can keep the expected number of data samples as small as possible, given the desired error upper-bound. However, the original SPRT makes two critical assumptions that limit its application in real-world scenarios: (i) samples are independently and identically distributed, and (ii) the likelihood of the data being derived from each class can be calculated precisely. Here, we propose the SPRT-TANDEM, a deep neural network-based SPRT algorithm that overcomes the above two obstacles. The SPRT-TANDEM sequentially estimates the log-likelihood ratio of two alternative hypotheses by leveraging a novel Loss function for Log-Likelihood Ratio estimation (LLLR) while allowing correlations up to $N (\in \mathbb{N})$ preceding samples. In tests on one original and two public video databases, Nosaic MNIST, UCF101, and SiW, the SPRT-TANDEM achieves statistically significantly better classification accuracy than other baseline classifiers, with a smaller number of data samples. The code and Nosaic MNIST are publicly available at https://github.com/TaikiMiyagawa/SPRT-TANDEM. | ['Taiki Miyagawa', 'Akinori F. Ebihara', 'Kazuyuki Sakurai', 'Hitoshi Imaoka'] | 2020-06-10 | sequential-density-ratio-estimation-for | https://openreview.net/forum?id=Rhsu5qD36cL | https://openreview.net/pdf?id=Rhsu5qD36cL | iclr-2021-1 | ['density-ratio-estimation'] | ['methodology'] | [ 1.40145257e-01 -3.74720216e-01 -4.49905127e-01 -4.12145823e-01
-1.12761486e+00 -3.10021430e-01 2.57082164e-01 1.70075625e-01
-7.10481942e-01 8.88110757e-01 -4.91094649e-01 -5.10140717e-01
-3.97293478e-01 -7.51435161e-01 -9.01126444e-01 -6.60646379e-01
-2.65790880e-01 5.88971674e-01 3.97265255e-01 4.14547354e-01
2.01078027e-01 3.62706482e-01 -1.76011336e+00 1.62318721e-01
9.36205387e-01 1.49174261e+00 6.95737749e-02 7.92655885e-01
3.63398224e-01 5.96523106e-01 -5.71071863e-01 -5.14343679e-01
4.96670276e-01 -4.94202673e-01 -6.06261015e-01 -2.94033259e-01
3.96808892e-01 -5.28036773e-01 -2.14082301e-01 1.18901098e+00
4.52326775e-01 1.28935739e-01 7.90549099e-01 -1.52686334e+00
-3.78954500e-01 4.24724996e-01 -6.99262381e-01 3.17609578e-01
1.68814346e-01 8.62506554e-02 1.18937278e+00 -9.24936175e-01
4.62589078e-02 1.07868659e+00 7.09914804e-01 1.63692459e-01
-1.19607198e+00 -1.01498711e+00 5.64131700e-02 5.22572458e-01
-1.82687259e+00 -1.78522527e-01 3.34531575e-01 -3.00906360e-01
6.82927489e-01 2.18126893e-01 4.84319419e-01 1.10455692e+00
1.70574978e-01 1.08461320e+00 1.10560822e+00 -4.40254867e-01
4.04175073e-01 6.36939481e-02 2.16766819e-01 5.60361207e-01
2.84928769e-01 1.38413653e-01 -5.91284037e-01 -3.00817162e-01
5.76354563e-01 9.98834446e-02 -1.94085762e-01 -4.60274033e-02
-9.13513184e-01 7.96580851e-01 -5.56093641e-02 1.30554125e-01
-2.54667729e-01 -4.72408645e-02 2.91213483e-01 4.57017928e-01
4.14075524e-01 1.81520581e-02 -5.37051976e-01 -1.28588438e-01
-1.02967083e+00 4.70816195e-01 7.36506045e-01 8.45729768e-01
4.48186874e-01 -2.17132509e-01 -2.94592619e-01 8.80816519e-01
3.22386444e-01 6.37068093e-01 4.84179676e-01 -8.63131762e-01
4.91296768e-01 1.72799334e-01 1.22581132e-01 -9.28825736e-01
-5.21166362e-02 -5.60687125e-01 -7.64527321e-01 1.20880358e-01
6.80927038e-01 -1.33280590e-01 -6.00935876e-01 1.82657337e+00
1.67796224e-01 1.95621297e-01 -2.05663204e-01 6.84991002e-01
3.16415787e-01 5.62376320e-01 -2.48688340e-01 -2.62684971e-01
1.18179572e+00 -5.59055030e-01 -5.11185467e-01 -3.03934157e-01
2.96069652e-01 -5.54700255e-01 1.09946203e+00 7.37596631e-01
-8.23427439e-01 -4.32393402e-01 -1.21220517e+00 2.52990842e-01
-7.54039958e-02 2.02273250e-01 4.76347327e-01 6.31867647e-01
-7.64012456e-01 5.59631228e-01 -8.45254242e-01 1.22545213e-02
7.02516675e-01 3.40657085e-01 -1.93216905e-01 -2.42988378e-01
-1.14723468e+00 4.05259967e-01 4.66408700e-01 1.11478806e-01
-8.02150249e-01 -4.80497867e-01 -5.76340854e-01 1.99884012e-01
5.18115222e-01 -1.20314747e-01 1.42263567e+00 -7.75266290e-01
-1.31262422e+00 5.66359699e-01 -3.21586251e-01 -7.35617816e-01
8.14182937e-01 -4.02019292e-01 -2.63664126e-01 7.24782720e-02
2.02520236e-01 5.74209213e-01 7.67033219e-01 -7.96094418e-01
-9.35382426e-01 -1.79640338e-01 -3.28105241e-01 -9.63644162e-02
-4.06252474e-01 1.04176022e-01 -3.96015465e-01 -6.43352091e-01
-5.50213009e-02 -7.73149967e-01 -1.88526660e-02 2.71100968e-01
-4.09408778e-01 -5.05245209e-01 5.83795369e-01 -5.90515494e-01
1.12652779e+00 -2.32199097e+00 -3.77403378e-01 1.88876256e-01
1.56926692e-01 4.92091566e-01 -7.60505944e-02 7.36770257e-02
-8.08091536e-02 1.67065367e-01 -2.21021026e-01 -1.94518387e-01
-7.22850710e-02 -2.27960087e-02 -1.81717306e-01 5.13535321e-01
1.39807180e-01 3.50029737e-01 -7.12504089e-01 -4.28446054e-01
-4.76127341e-02 1.84705690e-01 -5.81380010e-01 2.30433196e-01
4.81278151e-02 -3.14092077e-02 -1.59376949e-01 4.77480382e-01
7.73555756e-01 -4.46717948e-01 7.50962794e-02 2.26130895e-03
7.55030289e-02 1.30308077e-01 -1.51483309e+00 7.56277680e-01
-5.71331289e-03 7.85342753e-01 -2.94711351e-01 -1.16410005e+00
9.82680678e-01 2.23482475e-01 3.55410218e-01 -6.64591730e-01
3.19159001e-01 4.56051379e-01 7.18434900e-02 -1.86500445e-01
6.12688549e-02 -8.96564797e-02 1.92109367e-03 2.95596182e-01
6.99910894e-02 3.47998321e-01 4.81446117e-01 -1.82540014e-01
1.07815003e+00 -2.89999008e-01 6.32117331e-01 -1.79981411e-01
2.00574607e-01 -2.56755680e-01 1.16273618e+00 1.25907075e+00
-4.43733394e-01 7.26396978e-01 6.35450304e-01 -2.84433514e-01
-7.80337393e-01 -1.39035392e+00 -3.52611601e-01 6.63254857e-01
6.81226477e-02 -4.35496010e-02 -5.32640100e-01 -7.21836567e-01
1.53413638e-01 8.56436372e-01 -5.13850391e-01 -2.02419594e-01
-2.24688485e-01 -7.85276175e-01 4.42228228e-01 5.00071287e-01
6.96066260e-01 -6.42484725e-01 -4.84938562e-01 -1.05326921e-01
-2.00423107e-01 -1.17748415e+00 -4.15255457e-01 3.47919017e-01
-6.67024314e-01 -1.28354669e+00 -6.11477613e-01 -4.90521997e-01
4.91874725e-01 2.50163317e-01 8.00828338e-01 -1.87858477e-01
-1.69510081e-01 -9.88161191e-02 -4.49902654e-01 -5.24408042e-01
-2.57594764e-01 -1.34780839e-01 2.40859985e-01 1.90300047e-01
7.01639175e-01 -4.03957248e-01 -4.46305782e-01 5.80217540e-01
-7.37145782e-01 -7.72664743e-03 6.60467207e-01 8.34513545e-01
6.91287458e-01 4.03508008e-01 6.33594275e-01 -5.85608900e-01
4.20873553e-01 -6.00656152e-01 -8.62755060e-01 2.73328185e-01
-6.03942633e-01 -2.36900628e-01 8.28487575e-01 -6.86760962e-01
-5.85453689e-01 -3.57641160e-01 -2.98544019e-01 -6.41317725e-01
-2.10286066e-01 4.67005640e-01 -3.91180739e-02 3.08800995e-01
4.69378233e-01 3.67140263e-01 9.89551991e-02 -4.93978083e-01
-4.16845113e-01 7.40302086e-01 4.06968594e-01 -5.04657805e-01
5.80433607e-01 1.44027546e-01 -1.63082480e-01 -7.10421622e-01
-1.10972881e+00 -1.70913190e-01 -3.93706411e-01 -2.63714314e-01
5.29924273e-01 -6.70307577e-01 -9.08543527e-01 7.37611473e-01
-8.64004910e-01 -4.21211213e-01 -1.91192225e-01 8.56368423e-01
-3.28573704e-01 3.28712553e-01 -4.87415731e-01 -1.03636253e+00
-1.90822795e-01 -1.19891584e+00 5.17510653e-01 3.31900001e-01
-1.78756192e-01 -5.33524573e-01 -4.07176077e-01 2.35880405e-01
2.34115124e-01 1.43412560e-01 8.97640646e-01 -1.18459249e+00
-4.05407727e-01 -7.14970946e-01 -3.46272469e-01 8.63569140e-01
4.57469858e-02 1.20614395e-01 -7.13207364e-01 -5.43314934e-01
-2.36645411e-03 -4.19856250e-01 6.90763772e-01 4.73430008e-01
1.66146815e+00 -4.51984286e-01 -1.25492644e-02 5.14830232e-01
1.24686515e+00 6.37774587e-01 5.20791709e-01 2.12131381e-01
1.62836581e-01 3.66063744e-01 8.22032094e-01 6.41499758e-01
2.80583829e-01 5.44599652e-01 3.05024594e-01 3.01475555e-01
2.12377086e-01 -2.19114646e-01 4.19758469e-01 5.51805198e-01
2.54473120e-01 -6.54392183e-01 -8.91459703e-01 3.87243122e-01
-1.74303877e+00 -8.38711798e-01 -4.35009375e-02 2.60361457e+00
8.82806063e-01 4.32323962e-01 3.10548488e-02 4.89698648e-01
8.67855191e-01 1.14845242e-02 -8.23230624e-01 2.96701156e-02
-2.03394979e-01 1.10641435e-01 5.54509699e-01 3.49780738e-01
-1.13865650e+00 3.78369033e-01 6.13834238e+00 1.34125423e+00
-1.01248813e+00 5.79848699e-02 1.09998310e+00 -3.33596975e-01
1.83818623e-01 -1.54446363e-01 -1.07235086e+00 7.81847417e-01
9.98727024e-01 -2.49147162e-01 3.14319372e-01 8.78637731e-01
7.53897205e-02 -4.38324511e-01 -1.18213463e+00 1.14534914e+00
-3.01627442e-02 -1.11156857e+00 -1.33426040e-01 -4.47141100e-03
2.98752785e-01 8.42009708e-02 2.64169961e-01 4.82214987e-01
3.76051188e-01 -8.12092543e-01 1.01494634e+00 3.91674370e-01
7.53432631e-01 -8.17385375e-01 1.02844584e+00 5.38774192e-01
-8.70588243e-01 -3.35351914e-01 -1.90892473e-01 2.63134837e-02
-1.56497464e-01 1.18591571e+00 -7.69554079e-01 3.14570814e-01
1.06881368e+00 4.75393206e-01 -4.82576281e-01 1.27043724e+00
-1.78062901e-01 1.02053618e+00 -6.21144891e-01 -2.22040012e-01
-9.14136544e-02 1.70897190e-02 6.84115529e-01 9.91737247e-01
4.88176584e-01 -2.09907934e-01 2.63901144e-01 7.72191525e-01
-1.96759492e-01 -5.40628806e-02 -1.53628513e-01 -5.60562313e-03
9.35550749e-01 7.31065869e-01 -8.08241665e-01 -2.22112507e-01
-3.25736642e-01 7.03313589e-01 3.24583203e-01 2.80669957e-01
-8.89405608e-01 -5.28576136e-01 4.09008861e-01 2.92406175e-02
3.93163323e-01 -2.03452066e-01 -1.62904680e-01 -9.10622776e-01
3.29884440e-01 -9.41245854e-01 5.36898971e-01 -3.83770376e-01
-1.47841871e+00 4.18853104e-01 2.77830809e-01 -1.35135317e+00
-1.47729874e-01 -9.56401169e-01 -3.80901605e-01 6.18052363e-01
-1.27764452e+00 -4.87596273e-01 -1.50565132e-01 3.15338224e-01
4.13948625e-01 -3.59178096e-01 4.89383310e-01 5.51267922e-01
-9.75195289e-01 1.14709902e+00 3.45129520e-01 3.24536413e-01
6.24104202e-01 -1.07373571e+00 5.46506383e-02 9.19404447e-01
-1.28607735e-01 4.47962999e-01 7.08532095e-01 -4.52713281e-01
-9.96817470e-01 -9.85540867e-01 7.42492974e-01 -3.11147626e-02
5.18541455e-01 -4.89683867e-01 -9.48750377e-01 4.48285550e-01
-2.97187239e-01 2.34025627e-01 8.07689250e-01 -3.45093235e-02
-3.70501727e-01 -4.53772306e-01 -1.19982123e+00 5.74429929e-01
7.14504063e-01 -2.41223305e-01 -2.44087964e-01 3.76865864e-01
5.33408582e-01 -3.05730939e-01 -7.57352889e-01 5.73129773e-01
8.06384265e-01 -1.25015056e+00 7.34737813e-01 -2.03372806e-01
3.34948301e-01 -1.10074945e-01 -4.63314176e-01 -8.97646070e-01
4.72411849e-02 -4.45893079e-01 -1.23536430e-01 9.57787037e-01
4.96932596e-01 -9.18209672e-01 7.36614525e-01 4.59003538e-01
2.81607032e-01 -1.05728948e+00 -1.24006450e+00 -1.30243218e+00
-4.99194413e-02 -9.10978377e-01 4.90696490e-01 6.94366336e-01
-5.95998406e-01 -5.18942922e-02 -4.12724346e-01 5.90008311e-02
7.35485435e-01 -2.31531143e-01 6.61628246e-01 -1.24595261e+00
-4.28442657e-01 -4.45590198e-01 -3.49061400e-01 -1.21598744e+00
-2.94541977e-02 -6.65821910e-01 2.06955254e-01 -1.11655080e+00
3.61217529e-01 -4.84916925e-01 -5.36811054e-01 6.11784995e-01
-2.17221290e-01 1.19677174e-03 5.15193194e-02 2.05738291e-01
-6.61776602e-01 6.53222680e-01 7.25089669e-01 1.50468364e-01
9.28783324e-03 3.39435041e-01 -6.29169106e-01 7.02339590e-01
7.92895794e-01 -8.23056936e-01 -2.01340765e-01 -2.73406982e-01
1.56233504e-01 1.51304588e-01 4.84855890e-01 -1.10340667e+00
2.89047003e-01 -2.78196692e-01 3.64754915e-01 -9.09309030e-01
9.95112211e-02 -6.02754116e-01 -7.26379901e-02 5.21117687e-01
-4.82522845e-01 2.50874907e-02 9.14149210e-02 6.82323694e-01
-1.92705333e-01 -4.08982038e-01 9.96589422e-01 3.07652831e-01
-3.96938890e-01 4.10158217e-01 -5.21495104e-01 1.48394182e-01
1.07108939e+00 -2.27750972e-01 -1.93113491e-01 -3.29556406e-01
-2.05472156e-01 3.10783744e-01 5.13025783e-02 3.47777694e-01
6.64241791e-01 -1.33087909e+00 -8.15021753e-01 3.25047702e-01
1.28190547e-01 9.16040689e-02 2.97051191e-01 9.82881010e-01
-3.27410012e-01 2.29424194e-01 1.23966634e-01 -7.66455054e-01
-1.21555090e+00 2.69668788e-01 3.67247492e-01 -2.51607299e-01
-4.15468037e-01 9.84378755e-01 -3.97215858e-02 -3.78346950e-01
4.02042687e-01 -1.30385846e-01 7.59488940e-02 -1.04416467e-01
7.61575282e-01 5.58482707e-01 4.75079790e-02 -1.47991776e-01
-3.39610428e-01 2.68162843e-02 -3.61580729e-01 -1.36685401e-01
1.26922607e+00 7.78577775e-02 -2.21523009e-02 5.25162458e-01
1.49381208e+00 -2.84077197e-01 -1.39175856e+00 -5.61946750e-01
1.47720575e-02 -7.65440047e-01 4.88373404e-03 -6.91677928e-01
-8.13819528e-01 6.84020340e-01 6.98330343e-01 1.39332369e-01
1.15043664e+00 -2.17706248e-01 6.74223781e-01 4.65463698e-01
3.13354760e-01 -1.03019965e+00 1.70260265e-01 6.30094767e-01
7.05958188e-01 -1.24419081e+00 -8.79379362e-03 -1.69434667e-01
-3.91388267e-01 9.85738218e-01 6.38443053e-01 -8.45961869e-02
7.61552572e-01 3.53629827e-01 -3.68407547e-01 2.11020499e-01
-7.95274138e-01 2.57167816e-01 2.97505975e-01 3.65196049e-01
2.37351671e-01 9.15412605e-02 -2.31323838e-01 8.39429498e-01
-1.36451125e-01 5.06955199e-02 2.72538900e-01 9.85312343e-01
-2.82788038e-01 -7.76403606e-01 -3.14277887e-01 9.56594169e-01
-6.48548782e-01 3.80910328e-03 -6.20446205e-02 7.38071024e-01
1.98800173e-02 9.16476071e-01 2.81985998e-01 -5.86348176e-01
1.47384137e-01 -1.08481020e-01 2.08045319e-01 -3.16933185e-01
3.52528319e-02 -4.28449214e-02 -8.59581158e-02 -4.29561734e-01
-1.75066486e-01 -9.62764621e-01 -9.50215042e-01 -3.83160323e-01
-2.72922873e-01 2.30229214e-01 3.33920598e-01 1.20736110e+00
1.92924470e-01 2.57891893e-01 8.09733629e-01 -6.15818143e-01
-8.49704802e-01 -8.98320973e-01 -6.82612598e-01 8.21247324e-02
3.41035575e-01 -7.75832415e-01 -6.31594718e-01 -3.04435283e-01] | [8.499262809753418, 3.956108570098877] |
58bc52e4-86f7-47a1-a04c-8125e2c6ba10 | classical-versus-quantum-comparing-tensor | 2202.10471 | null | https://arxiv.org/abs/2202.10471v2 | https://arxiv.org/pdf/2202.10471v2.pdf | Classical versus Quantum: comparing Tensor Network-based Quantum Circuits on LHC data | Tensor Networks (TN) are approximations of high-dimensional tensors designed to represent locally entangled quantum many-body systems efficiently. This study provides a comprehensive comparison between classical TNs and TN-inspired quantum circuits in the context of Machine Learning on highly complex, simulated LHC data. We show that classical TNs require exponentially large bond dimensions and higher Hilbert-space mapping to perform comparably to their quantum counterparts. While such an expansion in the dimensionality allows better performance, we observe that, with increased dimensionality, classical TNs lead to a highly flat loss landscape, rendering the usage of gradient-based optimization methods highly challenging. Furthermore, by employing quantitative metrics, such as the Fisher information and effective dimensions, we show that classical TNs require a more extensive training sample to represent the data as efficiently as TN-inspired quantum circuits. We also engage with the idea of hybrid classical-quantum TNs and show possible architectures to employ a larger phase-space from the data. We offer our results using three main TN ansatz: Tree Tensor Networks, Matrix Product States, and Multi-scale Entanglement Renormalisation Ansatz. | ['Michael Spannowsky', 'Jack Y. Araz'] | 2022-02-21 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [ 6.76002959e-03 2.35748991e-01 2.53274471e-01 -1.50947049e-01
-8.21398973e-01 -5.32577157e-01 6.43371940e-01 -5.26136607e-02
-5.38080812e-01 7.04447210e-01 2.86620911e-02 -4.41419303e-01
-4.13858742e-01 -1.00907302e+00 -5.55900753e-01 -9.89902139e-01
-3.43827516e-01 7.82806516e-01 7.26118684e-02 -6.70825660e-01
3.39351267e-01 4.29936737e-01 -1.18511164e+00 1.22622937e-01
5.13558388e-01 1.04854488e+00 -6.62132025e-01 7.73149252e-01
3.49231690e-01 6.16954565e-01 -1.14491753e-01 -6.40443206e-01
6.07945085e-01 -5.97350121e-01 -1.12678957e+00 -7.14979470e-01
3.10672253e-01 -9.75428000e-02 -9.61520255e-01 9.69176948e-01
4.22233760e-01 3.57990623e-01 6.69125199e-01 -8.67050529e-01
-3.86838555e-01 7.41610587e-01 4.06926155e-01 2.00221688e-01
-6.53482378e-02 5.93535841e-01 1.47052884e+00 -3.43085319e-01
7.65206993e-01 1.06135583e+00 6.89594090e-01 6.37593269e-01
-1.90472305e+00 -3.02045941e-01 -1.07915318e+00 1.78461716e-01
-1.26074028e+00 -3.99006724e-01 6.21151984e-01 -9.08603221e-02
1.22563994e+00 3.05594087e-01 8.51780176e-01 8.35276484e-01
6.05376005e-01 1.82587206e-01 1.37911963e+00 -5.85274518e-01
4.82514471e-01 -1.34808451e-01 3.08715671e-01 8.95108163e-01
1.05380327e-01 6.73246980e-01 -6.75872803e-01 -4.04038221e-01
5.47775686e-01 -2.58249998e-01 1.82908669e-01 -7.06477404e-01
-1.53135216e+00 9.24836934e-01 9.29911911e-01 2.98257619e-01
-2.69879550e-01 7.58425176e-01 7.26312697e-01 5.15763223e-01
7.78923705e-02 1.05850017e+00 -2.87205905e-01 -2.35968664e-01
-6.56654716e-01 6.11463428e-01 8.57193470e-01 5.81134915e-01
1.13933456e+00 -2.81296015e-01 -1.32730231e-01 -3.46393278e-03
1.10566892e-01 6.63627982e-01 -6.56073019e-02 -1.24733651e+00
3.06680471e-01 1.83781177e-01 1.52050525e-01 -6.17731035e-01
-7.71221399e-01 -3.53097409e-01 -1.06552577e+00 -3.61043848e-02
5.79497397e-01 1.07193030e-01 -4.63440686e-01 1.51134753e+00
2.55299717e-01 -4.09702927e-01 2.31940597e-01 9.41501796e-01
9.70028266e-02 5.79361081e-01 -3.10644120e-01 -2.41799857e-02
1.20476139e+00 -6.51929617e-01 -4.02648300e-01 2.46184543e-01
1.27989566e+00 -2.61371791e-01 8.12778473e-01 3.67610395e-01
-1.13595200e+00 -5.08091860e-02 -1.14318967e+00 -3.27767491e-01
-3.49873185e-01 -4.60598826e-01 1.39987671e+00 9.38198745e-01
-1.04786444e+00 1.55360293e+00 -1.04273295e+00 -5.00354707e-01
1.43649310e-01 5.88118315e-01 -3.98535192e-01 8.80082622e-02
-1.34163916e+00 1.16176736e+00 5.43720067e-01 3.40824217e-01
-7.41577685e-01 -4.36927289e-01 -3.94663543e-01 1.67403817e-02
-1.51129574e-01 -8.72818768e-01 1.09007287e+00 -1.02579638e-01
-1.77479744e+00 6.45901859e-01 3.24218005e-01 -6.74047470e-01
-9.34826434e-02 4.30331171e-01 -1.46833673e-01 2.79161304e-01
-3.78901392e-01 3.84522498e-01 3.12722743e-01 -3.68485123e-01
3.22584838e-01 -3.32693100e-01 4.88918871e-01 -2.27102190e-01
-1.27089947e-01 -2.07170412e-01 4.33223993e-01 2.00624436e-01
4.92136270e-01 -1.44889092e+00 -4.80343699e-01 -3.29020858e-01
-5.48524976e-01 -1.16284713e-01 6.69124499e-02 -6.80750757e-02
9.54431176e-01 -2.00640559e+00 6.69941068e-01 5.07029533e-01
5.63195407e-01 -6.10690154e-02 -1.70685157e-01 1.03396857e+00
-8.05436261e-03 1.34235784e-01 -1.48014978e-01 -1.55718058e-01
5.60974300e-01 5.18304825e-01 1.19269885e-01 6.33716285e-01
3.34024876e-01 1.17331004e+00 -7.69497156e-01 -3.82673085e-01
7.09226681e-03 1.92580640e-01 -1.08913696e+00 -2.42821112e-01
-2.94806033e-01 6.79236233e-01 -4.92778450e-01 1.99977353e-01
5.77259302e-01 -3.91584098e-01 3.82595897e-01 -6.67663038e-01
-1.54561847e-01 7.54451156e-01 -6.91864431e-01 1.90233576e+00
-3.55105072e-01 5.08129537e-01 -8.59109834e-02 -1.14168417e+00
5.25036395e-01 9.40984637e-02 4.37259048e-01 -1.09935987e+00
2.84542024e-01 6.00121081e-01 5.80002725e-01 -1.61653727e-01
6.54104233e-01 -7.62880504e-01 -4.16357130e-01 6.65434599e-01
4.59735304e-01 -5.92360556e-01 3.33974093e-01 4.92326140e-01
1.55246603e+00 -5.67358173e-02 -3.11844230e-01 -5.37904561e-01
2.16547757e-01 1.09149404e-01 1.49664208e-01 8.80568504e-01
-1.91420287e-01 -4.75797849e-03 8.90153825e-01 -7.98935950e-01
-1.75013483e+00 -1.09870744e+00 -5.41694880e-01 6.35799229e-01
2.78364215e-02 -9.43578303e-01 -7.35925734e-01 -1.96496159e-01
-1.20371170e-01 4.35530454e-01 -5.60970366e-01 -3.72706413e-01
-5.64469874e-01 -1.16839266e+00 8.40211689e-01 -1.03013821e-01
1.58810616e-01 -8.02126646e-01 -3.12371314e-01 1.24481432e-01
1.16011709e-01 -8.19322407e-01 1.91625327e-01 7.84846842e-01
-1.09716547e+00 -8.44729900e-01 -4.33790206e-04 1.17510473e-02
3.21221799e-01 -4.61429536e-01 1.12457299e+00 -1.17915779e-01
-4.46558267e-01 -2.90166587e-03 -2.58537501e-01 3.51462394e-01
-8.11122894e-01 1.91766366e-01 4.79226619e-01 -3.77626449e-01
3.35499138e-01 -7.46760011e-01 -6.99887514e-01 -6.41320944e-02
-8.83791387e-01 -1.89646035e-02 7.31273890e-01 1.20643497e+00
1.32379770e-01 3.98463309e-02 -2.54620850e-01 -6.24102712e-01
4.52189922e-01 -8.28295946e-02 -6.25426948e-01 -2.50558201e-02
-5.37217855e-01 9.68078256e-01 7.76448309e-01 4.08903919e-02
-4.49667186e-01 -3.40108037e-01 -5.64067550e-02 6.45527020e-02
3.64424258e-01 3.45047027e-01 5.71110547e-01 -7.39071369e-01
1.15475845e+00 -5.44500910e-02 1.06787337e-02 -2.24378511e-01
7.06671298e-01 3.69197458e-01 4.59119081e-02 -9.87600446e-01
1.03948879e+00 3.49756569e-01 9.91388202e-01 -6.24356687e-01
-7.53459275e-01 -1.69120178e-01 -9.66649413e-01 1.76106453e-01
8.77071321e-01 -4.00043726e-01 -1.28802323e+00 2.12106466e-01
-1.12844574e+00 -1.14042468e-01 -5.47838211e-01 7.38850176e-01
-7.42323518e-01 2.29089320e-01 -1.26500559e+00 -9.27201092e-01
-2.87825048e-01 -1.19708920e+00 9.91919279e-01 -1.57665431e-01
2.34085441e-01 -9.23947155e-01 4.34744179e-01 3.65004033e-01
7.23523200e-01 2.13883936e-01 1.26189351e+00 -2.97639608e-01
-1.06582665e+00 -3.33688647e-01 -2.12399274e-01 4.05777991e-01
-6.52594507e-01 -1.95005700e-01 -8.41490328e-01 -4.88575459e-01
-2.12727368e-01 -8.53385687e-01 7.43979037e-01 -1.65957406e-01
6.56936586e-01 -1.48131967e-01 -2.68402621e-02 5.93126059e-01
1.48111939e+00 -3.31995845e-01 6.71159685e-01 7.07121715e-02
7.24585056e-01 4.85169202e-01 -2.05042008e-02 1.46334618e-01
2.02506468e-01 8.07939053e-01 2.89271981e-01 2.95256943e-01
3.61228347e-01 -1.34027228e-01 3.33728492e-01 1.62667310e+00
-3.83670032e-01 3.63893986e-01 -9.80610073e-01 -8.88525397e-02
-1.63094842e+00 -1.03877616e+00 -4.74965125e-01 2.23369455e+00
6.97570860e-01 3.20566475e-01 1.29804954e-01 5.95144443e-02
7.11610680e-03 3.17815728e-02 -4.12616432e-01 -7.49840379e-01
-1.34570941e-01 8.75201583e-01 7.70599425e-01 3.43122452e-01
-8.87750685e-01 7.24486589e-01 6.92060995e+00 7.64583111e-01
-8.44680071e-01 5.74012876e-01 2.18456477e-01 -2.55500078e-01
-3.99957091e-01 6.67416096e-01 -2.32421428e-01 1.67738125e-01
1.65537870e+00 1.91342235e-01 1.01330364e+00 1.65578395e-01
-1.73612759e-01 -1.67840213e-01 -1.42229354e+00 1.06533527e+00
-6.29878998e-01 -1.63537300e+00 -2.95649111e-01 4.03421313e-01
7.08319962e-01 7.53413558e-01 -2.39432946e-01 5.04010797e-01
-5.93103794e-03 -9.55728650e-01 4.91436630e-01 8.12612295e-01
5.92944205e-01 -5.91575742e-01 6.53468013e-01 2.13136747e-01
-7.28824258e-01 1.07178027e-02 -6.92172706e-01 -4.07522291e-01
-4.12587356e-03 5.44786394e-01 -2.48044044e-01 6.18796825e-01
3.83521855e-01 1.82085782e-01 -6.06146932e-01 7.15682387e-01
3.31957847e-01 4.52658683e-01 -7.92933702e-01 -5.30198216e-01
6.03776872e-01 -8.24642897e-01 3.48311961e-01 6.92119181e-01
3.13855596e-02 2.63407733e-02 -2.36222178e-01 1.12562263e+00
-4.62468937e-02 -1.55720413e-01 -5.99773586e-01 -6.76254213e-01
-4.54783812e-02 1.37670755e+00 -4.58043963e-01 -6.61037117e-02
2.96955518e-02 8.32160175e-01 4.98695374e-01 8.99829790e-02
-5.20671785e-01 -4.91668999e-01 4.01977479e-01 -9.37420353e-02
-8.62463377e-03 -6.67108655e-01 8.59219208e-02 -1.49146056e+00
-1.06249467e-01 -5.15024841e-01 -2.47656181e-01 -4.74389255e-01
-1.30259371e+00 3.94869089e-01 -3.01795244e-01 -8.99094641e-01
-2.49395177e-01 -1.09173441e+00 -3.22993726e-01 8.78174543e-01
-9.54821527e-01 -8.48239601e-01 4.39214021e-01 1.77331090e-01
-8.02312016e-01 1.59658641e-01 1.28399730e+00 3.09743345e-01
-6.17088795e-01 2.84974605e-01 7.72437215e-01 -1.52129866e-02
1.65375605e-01 -1.39683568e+00 6.39595866e-01 4.22016054e-01
2.59262055e-01 8.97922635e-01 8.86295021e-01 -1.02676541e-01
-2.20239162e+00 -3.83062184e-01 5.95964551e-01 -7.10615218e-01
1.23721325e+00 -9.60701287e-01 -4.32277620e-01 3.42199117e-01
9.80207175e-02 4.22071576e-01 6.25311255e-01 5.83677173e-01
-4.65545505e-01 -1.82869092e-01 -1.17755568e+00 2.52653956e-01
1.09992647e+00 -1.33084917e+00 -7.67642036e-02 9.94684160e-01
5.80985963e-01 -5.09254374e-02 -1.36526477e+00 1.69743642e-01
6.80700302e-01 -1.30264449e+00 7.56148934e-01 -8.36576104e-01
2.89228946e-01 -1.04696207e-01 -2.06007078e-01 -1.15759015e+00
-4.35425252e-01 -1.03040195e+00 -7.56352209e-03 3.83501351e-01
4.27123755e-01 -6.95790589e-01 8.53530109e-01 6.78495467e-01
-7.29123950e-02 -7.90131330e-01 -1.44819844e+00 -8.68577957e-01
6.72479153e-01 -3.91893476e-01 3.07167917e-01 6.55686438e-01
5.13907015e-01 5.90354145e-01 -2.36967877e-01 -6.33593649e-02
9.51811612e-01 -1.04224265e-01 5.67616343e-01 -1.12061656e+00
-4.75049883e-01 -4.39296246e-01 -9.04388368e-01 -7.75031447e-01
-1.12600271e-02 -1.42152274e+00 -1.78789407e-01 -8.47953498e-01
2.88510978e-01 -6.66860461e-01 -3.74155492e-01 -5.34937484e-03
6.97472095e-01 4.75499690e-01 1.27271935e-01 4.13932651e-01
-9.10618067e-01 7.43709028e-01 1.29949248e+00 -7.35277608e-02
3.62272650e-01 -4.82309699e-01 4.92170006e-02 1.93600759e-01
4.97053117e-01 -5.73268771e-01 5.13496734e-02 -2.01018348e-01
1.04976094e+00 3.77291232e-01 6.52395010e-01 -1.19076812e+00
3.56666654e-01 1.54008716e-01 1.00444376e-01 -2.84021497e-01
6.57991171e-01 -1.46913886e-01 1.47500247e-01 7.42883146e-01
-2.88013816e-01 -1.05724677e-01 -1.19729020e-01 3.34346682e-01
3.51115949e-02 -3.74721467e-01 8.36334348e-01 -2.93909192e-01
4.81136627e-02 2.09208667e-01 -1.54105693e-01 -1.51659206e-01
5.65172434e-01 2.79731005e-01 -4.18449789e-01 5.86561933e-02
-7.85016894e-01 -1.03949122e-01 5.97913742e-01 -3.40605021e-01
2.58416776e-02 -1.44582868e+00 -3.09868932e-01 1.01603776e-01
1.70194179e-01 -4.09096897e-01 3.85293156e-01 1.19816494e+00
-1.05475605e+00 8.82266283e-01 -2.96744078e-01 -5.28142035e-01
-2.63137579e-01 4.19275790e-01 7.25131214e-01 -5.05966485e-01
-3.66396666e-01 7.92711496e-01 -2.28598267e-01 -1.07476783e+00
-5.63010991e-01 -5.35346270e-01 7.97391891e-01 -3.64862263e-01
-7.19362199e-02 4.47491884e-01 2.46037081e-01 -6.69882119e-01
-2.65933812e-01 3.01786125e-01 1.99381664e-01 -2.13394165e-01
1.15928948e+00 2.89260447e-01 -7.14287460e-01 5.45853674e-01
1.49231815e+00 -3.54597270e-01 -7.85909653e-01 -3.36982757e-01
1.65472344e-01 -6.07992895e-02 1.62179112e-01 -4.67799604e-01
-4.42367166e-01 1.20847225e+00 6.70646429e-01 7.84306705e-01
4.25943434e-01 1.52575001e-01 9.26107466e-01 1.23822486e+00
9.75306094e-01 -1.04968417e+00 -6.46072552e-02 7.16320395e-01
3.38772774e-01 -1.04871380e+00 -3.48132625e-02 1.27410382e-01
9.57171842e-02 1.25010216e+00 5.17248884e-02 -3.65692824e-01
4.29620653e-01 -2.08926246e-01 -5.10883152e-01 -6.84560895e-01
-7.47094810e-01 -1.55740321e-01 2.39828020e-01 8.95461515e-02
3.02508354e-01 4.84477639e-01 -1.80910304e-01 -1.86902866e-01
-6.92437649e-01 -4.09257948e-01 6.85277164e-01 6.88224196e-01
-1.54979691e-01 -1.63312435e+00 -1.84070811e-01 7.11076498e-01
-2.14371562e-01 -3.19099158e-01 -3.03092692e-03 5.71524203e-01
-3.08110416e-02 6.27435327e-01 -1.30677372e-02 -6.04106188e-01
1.81170642e-01 3.72586727e-01 1.26552534e+00 -5.30707836e-01
-5.56743681e-01 -5.21134377e-01 2.08883971e-01 -9.06856298e-01
-5.12041092e-01 -5.28683841e-01 -1.02063811e+00 -1.04036987e+00
-7.04731405e-01 4.60380524e-01 1.02259791e+00 1.21820951e+00
2.20008925e-01 2.64853030e-01 5.86017728e-01 -1.11389220e+00
-1.11620080e+00 -1.10318720e+00 -8.56813729e-01 3.09939176e-01
3.03949356e-01 -5.82139790e-01 -4.70450759e-01 -8.37823629e-01] | [5.638866901397705, 4.9553542137146] |
4c0e612e-7757-4e51-bd46-10f11989d8ba | ranking-job-offers-for-candidates-learning | null | null | https://aclanthology.org/L14-1615 | https://aclanthology.org/L14-1615.pdf | Ranking Job Offers for Candidates: learning hidden knowledge from Big Data | This paper presents a system for suggesting a ranked list of appropriate vacancy descriptions to job seekers in a job board web site. In particular our work has explored the use of supervised classifiers with the objective of learning implicit relations which cannot be found with similarity or pattern based search methods that rely only on explicit information. Skills, names of professions and degrees, among other examples, are expressed in different languages, showing high variation and the use of ad-hoc resources to trace the relations is very costly. This implicit information is unveiled when a candidate applies for a job and therefore it is information that can be used for learning a model to predict new cases. The results of our experiments, which combine different clustering, classification and ranking methods, show the validity of the approach. | ["Felipe Nav{\\'\\i}o", "N{\\'u}ria Bel", 'Marc Poch', 'Sergio Espeja'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['implicit-relations'] | ['natural-language-processing'] | [ 2.99384389e-02 2.54296690e-01 -5.67269862e-01 -6.51558280e-01
-2.08847731e-01 -3.58324826e-01 8.93459439e-01 7.32501745e-01
-6.33708954e-01 1.16745341e+00 1.91583171e-01 -6.30788028e-01
-1.00191283e+00 -7.90119708e-01 1.97631702e-01 -3.94559562e-01
1.31679296e-01 1.23349309e+00 3.26986462e-01 -4.30243582e-01
5.87889969e-01 7.47877002e-01 -1.82263124e+00 4.07044679e-01
7.72298455e-01 6.53742433e-01 2.99942434e-01 2.86750346e-01
-4.69215930e-01 9.04426694e-01 -5.93137980e-01 -5.82159162e-01
2.60573346e-02 -4.54174161e-01 -1.42474151e+00 -2.26619206e-02
8.00830200e-02 3.74254018e-01 -1.16095133e-01 9.10983324e-01
2.28144556e-01 2.50811726e-01 9.87653017e-01 -1.02943623e+00
-1.25463188e-01 4.07061249e-01 1.11402594e-01 3.86902034e-01
7.56613672e-01 -4.43001032e-01 9.96563852e-01 -5.88704288e-01
9.21789885e-01 9.47527468e-01 6.74405217e-01 4.49557662e-01
-1.23266363e+00 -3.34627271e-01 -7.45343864e-02 4.55058604e-01
-1.56969082e+00 -2.84512341e-01 7.03328609e-01 -6.82457328e-01
9.90362644e-01 6.86508775e-01 4.05234426e-01 4.57056910e-01
-1.33405581e-01 -4.50649066e-03 1.70118475e+00 -1.05400622e+00
5.41859157e-02 1.04881787e+00 3.68515223e-01 8.31048846e-01
3.51342499e-01 -2.62096431e-02 -4.98684138e-01 -3.11953396e-01
3.81822199e-01 -3.09404265e-02 1.65664911e-01 -1.86210632e-01
-5.05360425e-01 5.43775797e-01 2.94500142e-01 1.15117478e+00
-1.94415867e-01 -6.15502656e-01 2.57907152e-01 4.36835021e-01
3.20427656e-01 9.99772310e-01 -5.95754802e-01 -1.05106406e-01
-9.79977190e-01 3.13827187e-01 1.23120618e+00 9.49950755e-01
1.19440162e+00 -7.88133442e-01 9.83349457e-02 9.00187492e-01
1.75385326e-01 -1.94817677e-01 7.78859496e-01 -5.95662773e-01
8.49572569e-02 1.02512336e+00 1.00473322e-01 -1.28889596e+00
-2.27635667e-01 -3.92537951e-01 -3.88049603e-01 5.91013581e-02
4.54737663e-01 1.17190212e-01 -7.78604031e-01 9.73875403e-01
-5.52127846e-02 -1.09533489e-01 -8.43637735e-02 6.97360039e-01
8.20835471e-01 1.50117632e-02 1.73868611e-01 -5.15623033e-01
1.18855882e+00 -5.66286564e-01 -7.69835711e-01 3.94708943e-03
7.26745605e-01 -8.51132631e-01 5.34238279e-01 3.44199121e-01
-7.66572893e-01 -6.28306389e-01 -6.64702952e-01 1.40700936e-01
-9.31921899e-01 -5.00611663e-02 9.49671626e-01 5.02486110e-01
-9.00854826e-01 1.15588510e+00 -4.88243341e-01 -7.66447902e-01
-1.84729636e-01 7.42527783e-01 -3.47073197e-01 -2.69752946e-02
-1.21805453e+00 1.21630883e+00 4.08470064e-01 -1.46494973e-02
4.89869155e-02 -5.80214113e-02 -5.90112686e-01 1.43541228e-02
3.16291034e-01 -4.54057425e-01 7.39781976e-01 -1.00493670e+00
-9.22176123e-01 1.36062956e+00 -3.36439669e-01 -2.84322500e-01
2.39305168e-01 2.54136443e-01 -8.31763566e-01 -8.43471587e-02
4.19163406e-01 3.61592285e-02 3.17714840e-01 -1.16062951e+00
-9.31766570e-01 -3.50924253e-01 1.59086227e-01 1.22340329e-01
-3.79549533e-01 3.54112387e-01 -4.36978877e-01 -7.05287233e-02
4.09877956e-01 -8.29634666e-01 -4.23027664e-01 -9.19232190e-01
-1.59749642e-01 -7.66099036e-01 3.98027450e-01 -6.10550940e-01
1.47579157e+00 -1.56961465e+00 5.65615557e-02 8.61095190e-01
-7.82087669e-02 1.72436878e-01 7.31234193e-01 7.66340554e-01
-1.68077767e-01 4.53597814e-01 1.90292329e-01 5.67276590e-02
-4.55952249e-02 4.87085074e-01 -2.95469221e-02 -3.87350023e-02
-1.66173533e-01 3.17173153e-01 -1.02409160e+00 -1.07026279e+00
-6.75045550e-02 -2.71754265e-01 -4.47139703e-02 -9.54741240e-03
1.89898685e-01 1.87969953e-01 -6.42219305e-01 6.86525702e-01
6.65259659e-02 -1.49756521e-01 7.62163281e-01 2.15541348e-01
-8.62411931e-02 8.27519238e-01 -9.58722174e-01 1.18245912e+00
-3.81359130e-01 4.86576408e-01 -2.14600697e-01 -1.24967444e+00
1.26438856e+00 3.21743011e-01 4.67867166e-01 -5.56946218e-01
-1.07264206e-01 7.84749746e-01 6.75583035e-02 -6.20325923e-01
4.52834994e-01 -2.53316849e-01 -1.95466354e-02 1.68898568e-01
-9.23991278e-02 1.99461535e-01 5.32973886e-01 7.86841363e-02
1.15439022e+00 2.23315150e-01 4.65582281e-01 -4.63100642e-01
9.72299755e-01 5.02162278e-01 6.22830033e-01 7.83422887e-01
1.92986757e-01 -1.34522235e-02 3.15800399e-01 -6.10494018e-01
-8.15971315e-01 -4.54445809e-01 -3.52648318e-01 8.39648485e-01
-6.69322489e-03 -6.08516037e-01 -2.24556461e-01 -9.35489535e-01
1.12998888e-01 3.81480753e-01 -4.16307390e-01 3.37776184e-01
-3.68285805e-01 -1.14313707e-01 1.41210780e-02 3.06329671e-02
1.25181541e-01 -1.13928032e+00 -2.47657046e-01 1.15084268e-01
3.47211696e-02 -6.46314800e-01 2.50085205e-01 5.01360238e-01
-1.01225388e+00 -1.17203498e+00 -3.81909221e-01 -1.10039496e+00
1.11293626e+00 -1.49793983e-01 1.31610262e+00 7.74944603e-01
-1.78980008e-01 2.13119045e-01 -3.87799352e-01 -3.07057858e-01
-3.79469901e-01 5.74642241e-01 1.78086400e-01 -2.80382425e-01
1.05324471e+00 -6.28671110e-01 -7.57517964e-02 4.74918962e-01
-4.81921703e-01 -3.29334229e-01 6.02088988e-01 9.21872258e-01
3.47536206e-01 4.68273044e-01 5.83519042e-01 -1.68683326e+00
8.86055112e-01 -5.34649491e-01 -3.88813078e-01 2.43381724e-01
-1.42185402e+00 1.24835946e-01 2.10978225e-01 -1.84735581e-01
-1.04839563e+00 1.97966635e-01 -4.40516770e-02 3.59207503e-02
-6.24475002e-01 9.63403285e-01 -8.31763223e-02 -4.71711904e-02
8.49578202e-01 -6.26810640e-02 -2.09381714e-01 -8.66931677e-01
-1.99327037e-01 1.00073731e+00 4.00467396e-01 -5.27498007e-01
7.17658758e-01 -1.47916883e-01 2.98660994e-01 -6.65003479e-01
-8.56737077e-01 -1.36593497e+00 -1.12593400e+00 -2.41186127e-01
4.18946922e-01 -3.80535066e-01 -5.02200246e-01 -2.53334135e-01
-9.22509432e-01 1.29049376e-01 -1.29520774e-01 6.86093748e-01
-4.11877096e-01 1.64156452e-01 -2.29515508e-01 -1.05758166e+00
2.08445281e-01 -4.93269891e-01 3.65329683e-01 2.11870566e-01
-8.29530954e-01 -1.23323989e+00 6.24599643e-02 6.41275644e-01
2.04626679e-01 -1.70630649e-01 1.13497353e+00 -1.25361812e+00
-3.49338979e-01 -4.55595315e-01 2.04790145e-01 -7.21957088e-02
3.67491663e-01 -5.26129544e-01 -6.43813193e-01 1.52129261e-02
-2.13143662e-01 -1.42912731e-01 7.16320276e-01 -2.55202651e-01
9.96386409e-01 -4.94127899e-01 -9.34708118e-01 8.68032053e-02
1.32609117e+00 2.55405247e-01 6.89239383e-01 7.64726460e-01
2.42097363e-01 1.18284929e+00 1.08426678e+00 1.22521311e-01
-1.00127727e-01 1.10151052e+00 -1.72315300e-01 -1.93975285e-01
1.05036624e-01 -1.16629153e-01 -2.87407130e-01 6.79270923e-01
-6.35257661e-01 2.43049204e-01 -1.06823647e+00 7.02925622e-01
-1.95891869e+00 -1.25944853e+00 -2.50396729e-01 2.16555500e+00
1.17845500e+00 4.75438118e-01 1.98419556e-01 4.82286483e-01
5.07338107e-01 -4.63228166e-01 7.78988749e-02 -6.69336498e-01
3.62359107e-01 3.92861813e-01 4.99196976e-01 6.73717856e-01
-8.49145770e-01 8.16317797e-01 6.56930971e+00 7.33326733e-01
-6.85617745e-01 -1.76407799e-01 5.63861847e-01 4.13626909e-01
-2.97023743e-01 3.88811499e-01 -9.01855171e-01 3.17513674e-01
8.85683477e-01 -3.56304109e-01 2.55612642e-01 8.67508054e-01
-1.00977486e-02 -3.07526916e-01 -1.19298339e+00 7.65365303e-01
7.34174699e-02 -1.30136979e+00 -3.11407089e-01 3.52397949e-01
6.55503750e-01 -5.30439675e-01 -4.39356089e-01 3.17603618e-01
2.18529746e-01 -1.12732112e+00 1.71701595e-01 9.20944810e-01
4.01011735e-01 -8.27509403e-01 1.13014448e+00 6.29616976e-01
-7.09097564e-01 -1.30376175e-01 -4.63862628e-01 -4.43702906e-01
-3.08238864e-01 3.63616884e-01 -1.43774223e+00 8.50645244e-01
6.34351194e-01 4.69995618e-01 -7.70468652e-01 1.16376293e+00
-2.38609344e-01 6.37374818e-01 -3.92613634e-02 -4.23961580e-01
1.34022981e-01 -2.73553491e-01 3.58514041e-01 1.36646748e+00
2.59608328e-01 1.37822628e-01 2.55077004e-01 6.54182792e-01
3.56806904e-01 4.90662098e-01 -9.89462614e-01 1.69029206e-01
2.55609751e-01 1.26589441e+00 -9.45601523e-01 -3.26258272e-01
-2.41659030e-01 7.40086436e-01 2.02064484e-01 1.95950195e-01
1.43566623e-01 -6.78604901e-01 2.22430766e-01 7.53869176e-01
-4.03463934e-03 -1.34664312e-01 -3.93973440e-02 -5.72744131e-01
-1.80837691e-01 -6.66131675e-01 7.03055203e-01 -5.64365149e-01
-1.36684203e+00 6.17713928e-01 1.87320948e-01 -1.14407718e+00
-8.25603724e-01 -7.82519996e-01 -2.81143248e-01 1.36825871e+00
-1.24763036e+00 -8.16920102e-01 -4.93426658e-02 3.96617770e-01
6.77289218e-02 -6.92446589e-01 1.11123049e+00 3.73492211e-01
9.41042677e-02 2.34659672e-01 -1.45085722e-01 2.49329045e-01
6.38874173e-01 -1.40191960e+00 -3.74158442e-01 3.43821138e-01
4.28262979e-01 8.79607558e-01 9.90179479e-01 -8.16065490e-01
-6.95398688e-01 -5.26950002e-01 2.01490426e+00 -5.73474646e-01
6.04338706e-01 9.31602046e-02 -8.98039579e-01 2.74425000e-01
2.83427089e-01 -4.10541415e-01 8.95436287e-01 8.87729526e-01
1.33622408e-01 -2.68015683e-01 -9.16396320e-01 1.63133755e-01
1.01846004e+00 -6.16122544e-01 -8.73602390e-01 7.06399202e-01
4.23183031e-02 2.01912737e-03 -7.84880996e-01 1.95251971e-01
2.84526885e-01 -8.59780848e-01 7.18642592e-01 -9.31762695e-01
1.89538270e-01 -8.32303390e-02 3.77036840e-01 -1.01666105e+00
-4.48487580e-01 -4.97411340e-01 4.02502179e-01 1.59740341e+00
9.71392512e-01 -4.93573040e-01 1.13300955e+00 9.22504783e-01
8.12316909e-02 -6.25552714e-01 -8.04724991e-01 -8.72993529e-01
-4.35232699e-01 6.18860871e-02 4.86789256e-01 1.44443917e+00
6.88631415e-01 6.47319555e-01 -1.39772713e-01 -2.06233934e-01
1.06998622e-01 4.62108642e-01 5.74593306e-01 -1.87600732e+00
-3.46668214e-01 -5.20958781e-01 -7.48065472e-01 -5.36927164e-01
2.88766712e-01 -1.18488383e+00 -4.82821949e-02 -1.67789495e+00
1.18129171e-01 -1.07521605e+00 -5.87778568e-01 6.07352853e-01
1.85058296e-01 -5.45488335e-02 -9.63456333e-02 5.24833918e-01
-4.95101333e-01 -2.54182935e-01 5.96735060e-01 -1.12786010e-01
-4.15175080e-01 5.21898270e-01 -5.70755839e-01 9.44244683e-01
6.90344632e-01 -9.00416195e-01 -3.19195777e-01 3.05750549e-01
4.25180227e-01 1.89915195e-01 -5.72825931e-02 -1.03637171e+00
5.64907610e-01 -5.29331744e-01 2.52297938e-01 -4.37386781e-01
2.80833334e-01 -1.28442633e+00 3.33065033e-01 5.83674252e-01
-5.97856164e-01 7.75263831e-02 -1.81613326e-01 5.61140358e-01
-4.33200955e-01 -1.14504600e+00 2.12099105e-02 -5.79196572e-01
-8.77824724e-01 2.68684551e-02 -3.75357121e-01 -3.16293180e-01
6.81205750e-01 -3.53165388e-01 2.12057471e-01 -4.28176969e-01
-1.09106851e+00 1.55073851e-02 4.49059159e-01 2.28393793e-01
4.82468247e-01 -1.33478868e+00 -4.76443380e-01 -2.59861082e-01
2.62100428e-01 -6.03179812e-01 -6.10169828e-01 6.37830496e-01
-5.02593815e-01 7.32573569e-01 -2.09571421e-01 -1.48079798e-01
-1.87669075e+00 6.39623821e-01 2.14916572e-01 -6.52370691e-01
-1.55413508e-01 5.14355958e-01 -5.88931203e-01 -4.24053401e-01
1.56080812e-01 3.40220749e-01 -9.83803034e-01 1.41541019e-01
1.34815842e-01 1.27666324e-01 2.68446803e-01 -7.56922364e-01
-5.27327538e-01 1.39761448e-01 7.75401155e-03 -2.40447223e-01
1.32807541e+00 -2.04825364e-02 -4.53078836e-01 5.45886695e-01
6.07148886e-01 4.65821564e-01 -5.99390939e-02 -4.94246304e-01
9.25741196e-01 -8.08728516e-01 -2.42603377e-01 -9.56307888e-01
-5.60740769e-01 4.46990579e-01 3.12218845e-01 7.04009831e-01
9.40823793e-01 2.36510307e-01 -1.68468118e-01 6.83350265e-01
5.85161209e-01 -1.53348649e+00 -2.31741115e-01 5.21792948e-01
5.95042408e-01 -1.19373798e+00 4.41750199e-01 -7.61901200e-01
-4.80092376e-01 1.24751520e+00 5.38419604e-01 3.71220738e-01
7.62355149e-01 -1.07410647e-01 1.09419376e-01 -4.00544256e-01
-7.69818842e-01 -6.55151904e-01 6.73076093e-01 7.07789838e-01
8.20815921e-01 3.33643444e-02 -1.34175980e+00 2.92796433e-01
-2.59655863e-01 1.85939353e-02 2.34855875e-01 9.89439666e-01
-5.34604967e-01 -1.85667467e+00 -2.35055745e-01 6.98524117e-01
-6.54881239e-01 -2.69199684e-02 -9.32318032e-01 8.23668063e-01
6.94057226e-01 1.11090839e+00 -1.29800618e-01 -3.56241941e-01
2.76105970e-01 4.11883712e-01 2.93957770e-01 -1.09531546e+00
-7.74079680e-01 -1.74036205e-01 7.59284377e-01 -7.00354278e-02
-8.60039175e-01 -7.05572844e-01 -9.98870969e-01 -2.89553776e-02
-4.80067849e-01 9.65474904e-01 6.61428750e-01 1.23968005e+00
1.09392516e-01 2.75261998e-01 5.13120294e-01 -2.54105091e-01
-2.26109520e-01 -9.22462046e-01 -8.30663502e-01 8.24594498e-01
-2.83513010e-01 -7.61954606e-01 -1.54357344e-01 1.20006643e-01] | [9.940918922424316, 8.715893745422363] |
4a91257c-4b01-44a6-93fe-9523f59045ef | rubik-s-optical-neural-networks-multi-task | 2304.12985 | null | https://arxiv.org/abs/2304.12985v2 | https://arxiv.org/pdf/2304.12985v2.pdf | Rubik's Optical Neural Networks: Multi-task Learning with Physics-aware Rotation Architecture | Recently, there are increasing efforts on advancing optical neural networks (ONNs), which bring significant advantages for machine learning (ML) in terms of power efficiency, parallelism, and computational speed. With the considerable benefits in computation speed and energy efficiency, there are significant interests in leveraging ONNs into medical sensing, security screening, drug detection, and autonomous driving. However, due to the challenge of implementing reconfigurability, deploying multi-task learning (MTL) algorithms on ONNs requires re-building and duplicating the physical diffractive systems, which significantly degrades the energy and cost efficiency in practical application scenarios. This work presents a novel ONNs architecture, namely, \textit{RubikONNs}, which utilizes the physical properties of optical systems to encode multiple feed-forward functions by physically rotating the hardware similarly to rotating a \textit{Rubik's Cube}. To optimize MTL performance on RubikONNs, two domain-specific physics-aware training algorithms \textit{RotAgg} and \textit{RotSeq} are proposed. Our experimental results demonstrate more than 4$\times$ improvements in energy and cost efficiency with marginal accuracy degradation compared to the state-of-the-art approaches. | ['Cunxi Yu', 'Weilu Gao', 'Yingjie Li'] | 2023-04-25 | null | null | null | null | ['rubik-s-cube'] | ['graphs'] | [ 3.85726541e-01 -1.27644867e-01 -1.31928800e-02 -1.94618642e-01
-2.54281163e-01 -1.93880871e-01 1.93763539e-01 -3.24416101e-01
-6.39684319e-01 7.67770290e-01 -4.43671167e-01 -3.68822455e-01
-3.37528139e-01 -5.25692344e-01 -9.37803090e-01 -1.08547974e+00
5.72715439e-02 1.64135695e-01 1.90473139e-01 1.96217999e-01
4.39691469e-02 6.78772807e-01 -1.36662710e+00 1.20877951e-01
6.68281972e-01 1.28515255e+00 5.35693347e-01 6.38736486e-01
1.24941438e-01 7.82146275e-01 -5.64215958e-01 -2.94814706e-01
1.91359177e-01 -2.34804839e-01 -6.28746212e-01 -5.98012984e-01
4.04154450e-01 -4.06938016e-01 -7.73363352e-01 9.52553809e-01
8.80279362e-01 1.08091719e-01 3.51415783e-01 -9.60579693e-01
-1.00291789e+00 2.76092172e-01 -4.13741589e-01 4.09184724e-01
-5.03713965e-01 3.58739823e-01 6.48125112e-01 -4.55081344e-01
3.47306997e-01 1.05790067e+00 4.61349487e-01 8.76778185e-01
-8.23615372e-01 -9.36275423e-01 -2.14738980e-01 6.87036753e-01
-1.22233331e+00 -6.57415390e-01 6.54362679e-01 -5.24510182e-02
1.58356941e+00 6.98924288e-02 5.86196601e-01 8.61084342e-01
7.61393011e-01 5.59058905e-01 7.98940122e-01 -4.31267619e-01
5.26013851e-01 -2.80303299e-01 -5.35335131e-02 1.06482863e+00
5.23492038e-01 1.51214078e-01 -1.06212342e+00 1.64433494e-01
7.26870477e-01 1.03453532e-01 1.75619517e-02 3.22617516e-02
-8.72953176e-01 3.56879175e-01 6.85870647e-01 5.16223460e-02
-3.09657127e-01 6.92660332e-01 1.48868293e-01 -1.00103781e-01
2.98618227e-01 5.90968668e-01 -3.73166561e-01 -1.24225289e-01
-7.46598959e-01 -2.09544763e-01 2.66964644e-01 7.90678144e-01
5.92744708e-01 3.83941531e-01 -5.21563999e-02 7.42946088e-01
1.78048179e-01 8.90126765e-01 4.39753979e-01 -9.18872595e-01
1.27832457e-01 4.96510506e-01 -9.48432907e-02 -3.50341678e-01
-9.47811365e-01 -5.20799875e-01 -1.01735985e+00 -6.08798638e-02
-1.03625767e-01 -3.06265473e-01 -1.20973575e+00 1.48458838e+00
1.70608878e-01 2.15286821e-01 2.62369782e-01 9.85258996e-01
1.00069964e+00 8.09685051e-01 -3.82153355e-02 -1.43476978e-01
1.35352540e+00 -1.01434529e+00 -5.41865170e-01 -3.37957799e-01
6.86111391e-01 -4.76109594e-01 7.91185915e-01 5.38872838e-01
-9.99003470e-01 -5.29775083e-01 -1.39562511e+00 -3.88233721e-01
-2.08905146e-01 3.63919288e-01 7.86429167e-01 8.65440667e-01
-8.70372355e-01 5.36890090e-01 -1.27778018e+00 -2.06323396e-02
6.48938060e-01 8.22157621e-01 3.74870360e-01 -2.30921805e-01
-8.34985614e-01 5.69588125e-01 2.75933385e-01 1.11525379e-01
-8.59550595e-01 -8.45423281e-01 -3.18696678e-01 1.28977317e-02
1.67424232e-01 -6.84800148e-01 1.15845680e+00 -2.06220821e-01
-2.05865836e+00 3.36750239e-01 -6.50172904e-02 -4.61677790e-01
-2.10639209e-01 -1.37426332e-01 -6.78960621e-01 1.23586655e-01
-4.24950868e-01 8.29339266e-01 6.27810478e-01 -4.18827504e-01
-3.60234618e-01 -4.44747925e-01 -1.74228460e-01 -1.09280296e-01
-8.48346472e-01 1.36737570e-01 -2.48212859e-01 -2.12602690e-01
1.18689761e-01 -1.17340732e+00 3.53246666e-02 1.65896490e-01
-5.01825392e-01 -3.74706060e-01 1.21889353e+00 -1.40433967e-01
7.55808532e-01 -2.08147430e+00 1.44749312e-02 -3.14753443e-01
3.94333541e-01 6.03303373e-01 -2.68703014e-01 1.52920783e-02
4.14657235e-01 -2.52121359e-01 1.34587169e-01 3.94805111e-02
-3.89968425e-01 2.05346271e-01 -5.39949127e-02 4.44399446e-01
1.57974035e-01 1.10961008e+00 -5.89081526e-01 -1.33170709e-01
9.58115608e-02 5.88321269e-01 -5.94358087e-01 -1.56965718e-01
-5.80922544e-01 4.67591941e-01 -6.46295369e-01 7.90941596e-01
5.64781368e-01 -8.03026080e-01 2.38800049e-01 -4.38508868e-01
-2.05006510e-01 4.60199296e-01 -6.62400842e-01 1.80643666e+00
-2.64620036e-01 7.97595143e-01 1.74932387e-02 -1.22316289e+00
7.77988613e-01 -9.91670638e-02 7.64463067e-01 -1.37428916e+00
1.73938602e-01 2.09065899e-01 2.84156084e-01 -6.69738889e-01
3.73851746e-01 9.64114256e-03 1.27470538e-01 4.36191261e-01
1.71986178e-01 -5.13595482e-03 -1.40009031e-01 -2.20853224e-01
1.27153277e+00 2.77237535e-01 -2.03004137e-01 -3.44018042e-01
-4.84644882e-02 3.69410180e-02 4.21727180e-01 4.73668128e-01
-1.31195769e-01 -1.30508319e-01 -1.44481277e-02 -7.32743859e-01
-1.04215467e+00 -9.07561123e-01 -2.57874608e-01 1.18271303e+00
2.83232123e-01 -1.04129083e-01 -6.80212438e-01 -5.69038140e-03
-1.93167016e-01 3.92628998e-01 8.14500600e-02 -3.19514900e-01
-7.95776963e-01 -1.39114964e+00 8.59853327e-01 3.84505451e-01
8.61682534e-01 -6.37040675e-01 -1.26834798e+00 2.55309880e-01
1.81170583e-01 -1.42971706e+00 -6.62449896e-02 4.26132083e-01
-7.89521515e-01 -6.54327273e-01 -3.55682433e-01 -5.67177415e-01
3.73776138e-01 3.95180702e-01 6.34495378e-01 -2.16481939e-01
-8.72741103e-01 -5.53654656e-02 8.37301537e-02 -9.29257989e-01
4.00282331e-02 2.04192787e-01 4.91139799e-01 -3.01445782e-01
3.95695359e-01 -6.14355266e-01 -8.88701081e-01 3.01809572e-02
-7.32049048e-01 3.52058053e-01 1.03241026e+00 5.98998904e-01
7.14405417e-01 2.12694004e-01 4.90303516e-01 -5.12123287e-01
1.63893983e-01 -1.08734697e-01 -8.36555958e-01 4.31736231e-01
-7.62983322e-01 7.12107480e-01 9.49835300e-01 -4.41911042e-01
-7.90672719e-01 -6.18058816e-02 1.75881237e-01 -4.84655499e-01
3.51686746e-01 1.77529886e-01 1.97216511e-01 -5.59912205e-01
7.77116239e-01 2.42697403e-01 -1.91133901e-01 -2.29769215e-01
-3.23445685e-02 6.15431786e-01 4.47710365e-01 -4.59786355e-01
3.56597841e-01 3.56006116e-01 6.64919019e-01 -1.17661178e+00
-7.35352993e-01 7.40934014e-02 -9.84150171e-02 -3.34557205e-01
8.23904216e-01 -8.59903872e-01 -1.29486632e+00 7.68536568e-01
-1.09894395e+00 -4.28520292e-01 5.91286048e-02 7.16473579e-01
-1.71079598e-02 -1.97731629e-02 -6.22518182e-01 -7.74302185e-01
-6.67129278e-01 -1.20009243e+00 9.93239224e-01 6.05398118e-01
4.38713253e-01 -5.79638839e-01 -3.41459990e-01 5.42697191e-01
6.93193913e-01 -2.25207075e-01 1.37831891e+00 -1.13702036e-01
-1.18936372e+00 2.58742511e-01 -5.56439042e-01 9.29235741e-02
-2.91681606e-02 -3.20227265e-01 -1.29301798e+00 -4.56825733e-01
-1.71395376e-01 -6.97753131e-01 9.40246522e-01 6.25088513e-01
1.65786386e+00 -2.76859671e-01 -6.85971737e-01 8.17283630e-01
1.38059056e+00 4.98133868e-01 4.58782643e-01 -1.57296956e-01
6.60178244e-01 -1.20093182e-01 -1.37366563e-01 3.93434912e-01
9.67122018e-02 6.86310172e-01 5.77064931e-01 1.22427173e-01
-2.48202354e-01 1.39966965e-01 4.34061974e-01 1.03839564e+00
-1.73383102e-01 -5.85218787e-01 -8.10009241e-01 3.27913207e-03
-1.36726213e+00 -6.72158778e-01 6.56509325e-02 1.91690576e+00
7.98488975e-01 -1.16273891e-02 -4.26480085e-01 -3.14896971e-01
4.18339700e-01 1.66419134e-01 -1.47171712e+00 -3.59672844e-01
6.52872249e-02 6.65847182e-01 9.36955631e-01 -9.27655175e-02
-8.44936013e-01 6.02599740e-01 5.90662956e+00 9.24115360e-01
-1.76265323e+00 2.51495600e-01 6.19865119e-01 -8.78083527e-01
-3.52455266e-02 -5.03628969e-01 -1.14377284e+00 5.08708894e-01
1.46816170e+00 2.61774868e-01 8.25228691e-01 6.37935817e-01
1.54055908e-01 -1.67684287e-01 -1.12585926e+00 1.17419505e+00
-1.24043308e-01 -1.73865032e+00 -9.67586264e-02 1.80577442e-01
6.87318325e-01 5.02025425e-01 3.09321731e-01 1.92447484e-01
1.86111897e-01 -1.10733998e+00 5.03066540e-01 4.44489866e-01
1.22610259e+00 -6.22561574e-01 2.96935141e-01 4.61718172e-01
-9.33818758e-01 -2.46158764e-01 -6.99637771e-01 -1.13180548e-01
-2.04933062e-01 1.05262125e+00 -7.27309227e-01 4.68611449e-01
9.73432899e-01 3.68612766e-01 -7.66337067e-02 6.45064056e-01
2.84918547e-01 5.45896471e-01 -5.29518545e-01 -7.55226195e-01
1.63506538e-01 -1.78543359e-01 1.47816092e-01 8.94917548e-01
6.16959333e-01 4.24724787e-01 -1.38934050e-02 8.99918914e-01
-6.53056085e-01 -4.94523853e-01 -8.74210596e-02 -2.67277926e-01
5.95324993e-01 1.29238653e+00 -4.23563957e-01 -6.94740862e-02
-3.84876341e-01 8.13195050e-01 2.01070711e-01 1.92387745e-01
-8.80595028e-01 -5.12726426e-01 7.33336627e-01 -6.71238005e-02
7.52560794e-02 -6.22706473e-01 -2.66968429e-01 -5.56748509e-01
2.27640077e-01 -1.39926761e-01 -1.69898063e-01 -8.76127064e-01
-7.61017561e-01 3.59342039e-01 -3.23474020e-01 -7.21992910e-01
2.72997022e-01 -1.09319019e+00 -8.68806615e-02 7.05846012e-01
-1.75754619e+00 -9.18435514e-01 -4.35191244e-01 1.37267828e-01
3.12759340e-01 -2.55178958e-01 7.40570426e-01 4.30632889e-01
-1.05604875e+00 6.32088542e-01 5.81357241e-01 -2.52806693e-01
4.40211833e-01 -5.87380111e-01 3.85869622e-01 7.75834799e-01
-4.81741391e-02 2.51867115e-01 2.62717009e-01 -4.14466918e-01
-2.32747674e+00 -1.46791673e+00 3.59824926e-01 4.88243736e-02
4.53866065e-01 -3.29990566e-01 -5.92714727e-01 2.56237149e-01
1.13411218e-01 2.77967811e-01 6.87545657e-01 -2.31774569e-01
-1.98155597e-01 -4.43911284e-01 -1.05129933e+00 4.23234195e-01
1.40468276e+00 -5.32330394e-01 2.78824955e-01 7.97157943e-01
6.80183470e-01 -5.90878010e-01 -6.40608907e-01 4.56562608e-01
6.27090096e-01 -5.42516351e-01 9.74831283e-01 -3.64137560e-01
3.87215137e-01 -9.02885273e-02 -5.96007006e-03 -8.94661665e-01
-2.66661793e-01 -7.21950889e-01 -3.30443323e-01 4.07808930e-01
2.31196582e-01 -6.11750782e-01 1.14549065e+00 4.29103285e-01
-5.79787612e-01 -1.02161658e+00 -1.43544471e+00 -9.18783188e-01
-1.44724965e-01 -3.31895590e-01 3.16971064e-01 4.23719019e-01
-3.27221066e-01 3.50034028e-01 -3.77058536e-01 4.06328231e-01
4.74313706e-01 -7.26702809e-02 6.00441471e-02 -1.26634681e+00
-5.09944260e-01 -2.44757548e-01 -3.88288528e-01 -1.46718347e+00
-1.90670848e-01 -8.84448409e-01 5.61114065e-02 -1.29088724e+00
3.06928962e-01 -7.27284074e-01 -4.83769417e-01 6.47560060e-01
2.90543646e-01 2.38009989e-01 -7.22823367e-02 3.25669289e-01
-8.68365467e-01 5.48255980e-01 1.28820837e+00 -2.11089283e-01
-1.11559056e-01 -4.17966723e-01 -3.59296709e-01 3.39273274e-01
5.98457277e-01 -5.03750026e-01 -4.56210643e-01 -1.11540449e+00
3.33600432e-01 1.10165700e-02 4.86403227e-01 -1.71795082e+00
7.08565474e-01 -2.17249304e-01 4.76976216e-01 -3.95463228e-01
8.21078002e-01 -3.89248043e-01 3.53862755e-02 8.46075118e-01
-9.59985107e-02 -5.12501858e-02 3.64232808e-01 4.53171313e-01
4.36556786e-01 -4.75809127e-02 9.60691571e-01 5.47067299e-02
-6.33965909e-01 3.75555336e-01 -4.68999505e-01 -1.82558805e-01
9.27606106e-01 -5.71318530e-02 -9.53777134e-01 3.42986524e-01
-5.92533909e-02 -1.43005271e-02 -6.69347867e-02 8.59114006e-02
4.73446339e-01 -9.18744445e-01 -3.10407043e-01 3.67590725e-01
-1.92644387e-01 1.74947992e-01 5.94695926e-01 6.02750540e-01
-5.33147514e-01 1.09898961e+00 -3.01825464e-01 -7.68784642e-01
-9.89610791e-01 2.33813539e-01 5.28805077e-01 -2.47490592e-02
-4.31009710e-01 1.22244906e+00 -1.39398694e-01 6.39738096e-03
1.39928222e-01 -5.15562356e-01 4.31510270e-01 -6.40418112e-01
2.64214218e-01 5.94527960e-01 4.39972788e-01 3.03578049e-01
-3.59420776e-01 5.20538509e-01 -2.86266804e-01 3.86766553e-01
1.38077044e+00 3.51049036e-01 -6.76027441e-04 -1.25878468e-01
9.65540111e-01 -4.05535847e-01 -1.41721141e+00 -3.91857296e-01
-1.76595047e-01 2.30887547e-01 6.19476557e-01 -1.03618610e+00
-1.17452645e+00 9.82710063e-01 1.02321911e+00 -3.28674465e-02
1.34076214e+00 7.52655119e-02 1.18234086e+00 1.13490891e+00
5.59810340e-01 -1.30339098e+00 3.10572714e-01 4.47857171e-01
2.46127710e-01 -8.22923779e-01 2.05215722e-01 -2.02107534e-01
-3.66694760e-03 1.22017801e+00 7.17623949e-01 2.88104624e-01
5.08193791e-01 4.67624187e-01 -4.71727073e-01 -3.00190300e-01
-7.93249667e-01 3.52875382e-01 8.41886401e-02 4.36208487e-01
2.85668790e-01 1.67167261e-01 2.81252623e-01 3.87747943e-01
-1.96913481e-01 4.11083788e-01 3.35861981e-01 1.06187594e+00
-7.22523212e-01 -1.01264822e+00 7.99439773e-02 8.56442094e-01
-2.94101834e-01 -1.46526575e-01 5.96738197e-02 2.58090049e-01
3.08676183e-01 7.18606412e-01 2.11454019e-01 -5.91279566e-01
1.82433799e-02 6.48223609e-02 7.43327379e-01 -4.60220516e-01
-1.78289898e-02 -2.61354208e-01 -2.75863451e-04 -4.37213451e-01
-4.24197286e-01 -3.56894135e-01 -1.73593402e+00 -4.40116882e-01
-2.40663946e-01 -2.30804250e-01 1.18645442e+00 1.25194645e+00
1.09646237e+00 9.51575279e-01 4.42765296e-01 -5.86066484e-01
-5.04680336e-01 -7.10098982e-01 -1.97208568e-01 -5.98698795e-01
3.07584792e-01 -7.10216224e-01 3.27692628e-01 -1.51180431e-01] | [8.289546966552734, 2.511937141418457] |
0d0b7365-2f71-4e84-a1e7-51346b2fe4c7 | progressive-with-purpose-guiding-progressive | 2209.10071 | null | https://arxiv.org/abs/2209.10071v2 | https://arxiv.org/pdf/2209.10071v2.pdf | Progressive with Purpose: Guiding Progressive Inpainting DNNs through Context and Structure | The advent of deep learning in the past decade has significantly helped advance image inpainting. Although achieving promising performance, deep learning-based inpainting algorithms still struggle from the distortion caused by the fusion of structural and contextual features, which are commonly obtained from, respectively, deep and shallow layers of a convolutional encoder. Motivated by this observation, we propose a novel progressive inpainting network that maintains the structural and contextual integrity of a processed image. More specifically, inspired by the Gaussian and Laplacian pyramids, the core of the proposed network is a feature extraction module named GLE. Stacking GLE modules enables the network to extract image features from different image frequency components. This ability is important to maintain structural and contextual integrity, for high frequency components correspond to structural information while low frequency components correspond to contextual information. The proposed network utilizes the GLE features to progressively fill in missing regions in a corrupted image in an iterative manner. Our benchmarking experiments demonstrate that the proposed method achieves clear improvement in performance over many state-of-the-art inpainting algorithms. | ['Jun Chen', 'Muhammad Alrabeiah', 'Kangdi Shi'] | 2022-09-21 | null | null | null | null | ['image-inpainting'] | ['computer-vision'] | [ 4.34363514e-01 -1.86699420e-01 -1.49147168e-01 -1.10032909e-01
-8.02782893e-01 -6.37564585e-02 2.55807281e-01 1.05071813e-01
-2.96183228e-01 7.14048564e-01 3.88474733e-01 2.67162651e-01
-8.76486078e-02 -7.81845748e-01 -8.74238968e-01 -8.46890688e-01
-7.65684701e-04 -2.77685463e-01 -1.72949225e-01 -1.31922722e-01
2.77362078e-01 6.30681932e-01 -1.58688819e+00 5.41521966e-01
9.23320532e-01 1.25563371e+00 4.31197077e-01 5.24487495e-01
-1.44220322e-01 1.19829750e+00 -4.87747014e-01 3.79370041e-02
3.50941360e-01 -5.36054075e-01 -5.87356925e-01 4.20312196e-01
5.42887151e-01 -6.11516893e-01 -6.40325129e-01 1.05683017e+00
3.40242416e-01 4.72568683e-02 3.58885497e-01 -8.07041109e-01
-6.71445012e-01 2.99326062e-01 -8.12214315e-01 1.92637995e-01
2.29022250e-01 9.72605050e-02 1.00514805e+00 -1.12195218e+00
5.94860733e-01 1.00587904e+00 7.18380749e-01 1.18845493e-01
-1.23060799e+00 -3.58193785e-01 -7.28445873e-02 4.07441497e-01
-1.21499646e+00 -3.53439927e-01 1.39269841e+00 -1.60969585e-01
7.29250908e-01 1.09951437e-01 5.69817662e-01 7.50552654e-01
4.90581542e-01 8.26033652e-01 1.03410816e+00 -4.63844478e-01
8.67737755e-02 -1.58688173e-01 -3.12545329e-01 8.24431777e-01
-5.28763346e-02 4.49739732e-02 -8.43797565e-01 9.19880792e-02
9.01131392e-01 4.07268047e-01 -5.00573635e-01 -2.70961702e-01
-8.48271370e-01 7.79738724e-01 7.95555770e-01 4.47319210e-01
-7.16778874e-01 2.69730419e-01 2.43952438e-01 4.18489933e-01
4.11190003e-01 2.91154087e-01 -2.59486169e-01 6.47760788e-03
-1.43336678e+00 2.05074787e-01 3.16897303e-01 6.33606315e-01
1.23437905e+00 1.82479531e-01 -2.27898777e-01 8.66497219e-01
1.12147160e-01 1.32580817e-01 4.22530919e-01 -1.13748682e+00
4.63869214e-01 6.76472247e-01 6.57650381e-02 -1.36758173e+00
-1.86297700e-01 -7.35055387e-01 -1.33595669e+00 2.80637950e-01
-8.63904059e-02 6.89690337e-02 -8.75639796e-01 1.54976666e+00
5.79143316e-02 2.21388504e-01 -9.16437134e-02 8.62202525e-01
5.26549697e-01 7.55321741e-01 -2.68766493e-01 -2.63999671e-01
1.16047335e+00 -1.01540649e+00 -7.39158690e-01 -2.15631247e-01
1.57561287e-01 -9.73901808e-01 9.17866528e-01 3.93557996e-01
-1.33147323e+00 -9.89443660e-01 -1.27884841e+00 -3.90424579e-01
1.34307504e-01 2.51622558e-01 3.30541819e-01 1.20078988e-01
-1.01876426e+00 8.51000965e-01 -8.50387633e-01 7.17059299e-02
6.84706748e-01 6.26338199e-02 -4.08863187e-01 -3.40914190e-01
-9.73012745e-01 6.40931904e-01 2.06589431e-01 1.53595284e-01
-9.03133094e-01 -9.14355457e-01 -9.33037698e-01 3.54754180e-01
2.47219965e-01 -5.81144333e-01 9.48648512e-01 -1.03359818e+00
-1.35077691e+00 4.98573482e-01 -2.64297128e-01 -6.62568033e-01
4.61355597e-01 -5.30497909e-01 -1.66647613e-01 5.91967702e-01
6.92006052e-02 5.07606745e-01 1.40852928e+00 -1.38682973e+00
-6.46342397e-01 -2.13775709e-01 -1.53251573e-01 1.03360876e-01
-4.27899271e-01 -2.20590353e-01 -4.08247530e-01 -1.04673409e+00
3.05962771e-01 -4.80125934e-01 -2.08962575e-01 2.14629471e-01
-1.96900636e-01 2.32770428e-01 1.16374588e+00 -9.71662402e-01
1.15446281e+00 -2.20236850e+00 3.40374380e-01 -3.42347845e-02
4.90355790e-01 1.48581445e-01 -1.56190380e-01 6.37784660e-01
-1.64068073e-01 -2.93469667e-01 -5.93813479e-01 -6.10937059e-01
-2.22617164e-01 2.15274140e-01 -5.53324819e-01 4.21308279e-01
6.16902530e-01 8.16985071e-01 -8.20759952e-01 -4.56780583e-01
4.88148093e-01 7.98301160e-01 -6.73583567e-01 4.66923147e-01
-9.43755358e-02 6.18771195e-01 -2.45922670e-01 5.27543962e-01
1.01240909e+00 -1.66887954e-01 -2.08267808e-01 -5.03853142e-01
-1.52369797e-01 1.10742904e-01 -7.69650400e-01 2.16483521e+00
-5.54577172e-01 7.76027977e-01 4.35694665e-01 -1.23774385e+00
9.47740972e-01 1.89182222e-01 7.00219750e-01 -8.71373951e-01
7.36507848e-02 1.92401990e-01 -2.17281148e-01 -4.29501861e-01
5.73179781e-01 4.01335442e-03 1.65458292e-01 2.52070367e-01
3.05324852e-01 -1.36811227e-01 1.37057588e-01 1.51740402e-01
1.19838285e+00 1.89095929e-01 9.70106795e-02 4.40222248e-02
7.38003194e-01 -2.94706106e-01 5.71653903e-01 6.30089343e-01
-7.36927241e-02 9.14233267e-01 3.63595217e-01 -6.65190220e-01
-1.12818265e+00 -9.61295366e-01 7.02589750e-02 7.28356779e-01
3.84986363e-02 -3.86950284e-01 -7.86401093e-01 -3.77879918e-01
-1.37925774e-01 2.50165135e-01 -6.22195005e-01 -3.83593529e-01
-9.31729555e-01 -4.66214240e-01 1.25887677e-01 4.09610212e-01
8.92935634e-01 -1.08432746e+00 -8.29718769e-01 5.06701291e-01
-3.14965606e-01 -1.20586431e+00 -4.00375515e-01 2.46210814e-01
-9.23281550e-01 -8.96444678e-01 -7.83717275e-01 -9.17751074e-01
6.23973429e-01 3.75131667e-01 1.06187737e+00 1.82672799e-01
-4.62526470e-01 -6.85329512e-02 -4.77966815e-01 1.25728458e-01
-3.85133445e-01 1.88704468e-02 -3.78890276e-01 3.32418054e-01
-1.61748201e-01 -1.07458627e+00 -9.08972025e-01 -1.50716275e-01
-1.31705070e+00 2.19731465e-01 1.02756095e+00 1.08802021e+00
7.45883167e-01 3.60256970e-01 2.44482458e-01 -5.79457223e-01
5.07610440e-01 -2.89987952e-01 -4.10046220e-01 -7.16284961e-02
-2.93154150e-01 2.72444665e-01 9.80318069e-01 -4.39796112e-02
-9.83665526e-01 1.11951217e-01 -3.12450737e-01 -6.42822623e-01
-1.92328785e-02 6.34136379e-01 -1.72607347e-01 -2.47566655e-01
4.16381806e-01 6.23004436e-01 -1.08889982e-01 -6.97735786e-01
3.26966047e-01 3.10876429e-01 8.32592964e-01 -5.49857676e-01
8.58978152e-01 6.30476654e-01 1.91463724e-01 -7.64345169e-01
-9.26818311e-01 -2.69426912e-01 -5.66551208e-01 -1.60382658e-01
6.30866289e-01 -9.73405004e-01 -3.86264533e-01 6.51634157e-01
-1.18725157e+00 3.25112123e-05 -3.96384090e-01 1.31895989e-01
-6.27127290e-01 5.66551685e-01 -8.77089560e-01 -4.60676879e-01
-5.19167483e-01 -1.19077718e+00 1.15960181e+00 2.17921734e-01
1.46159142e-01 -6.97780371e-01 -1.03673466e-01 1.78846851e-01
5.82628846e-01 4.75511014e-01 9.97368932e-01 1.72656134e-01
-7.38969803e-01 -1.80825531e-01 -3.42506588e-01 7.78148770e-01
4.71976221e-01 -3.42473269e-01 -8.64566326e-01 -4.75617975e-01
5.11231840e-01 -3.34850132e-01 1.17769814e+00 3.55495661e-01
1.39371431e+00 -3.82211626e-01 1.64912045e-01 9.08256412e-01
1.64413047e+00 -6.51737824e-02 8.76947582e-01 3.07332784e-01
8.10305357e-01 3.15770745e-01 3.60816598e-01 7.04797506e-01
9.03068297e-03 5.71983218e-01 5.93045712e-01 -3.91327620e-01
-3.76537293e-01 -2.45852336e-01 3.13980967e-01 8.44389141e-01
7.86315575e-02 1.79057136e-01 -5.74171782e-01 5.67508936e-01
-1.77837288e+00 -8.60308409e-01 2.50069976e-01 1.95363379e+00
1.05537558e+00 1.12663768e-01 -3.51512432e-01 3.85881633e-01
4.38625008e-01 5.70387065e-01 -4.99796689e-01 -2.16309160e-01
-2.39013448e-01 6.70696437e-01 3.41230631e-01 4.09044713e-01
-1.11380959e+00 7.05984831e-01 5.32701778e+00 9.86582518e-01
-1.24688280e+00 -5.65447994e-02 7.52052844e-01 3.20870094e-02
-9.31531116e-02 -1.14785805e-02 -2.87494719e-01 4.63936001e-01
6.64816558e-01 -3.30138803e-02 4.32227254e-01 6.94246769e-01
2.38924056e-01 -2.01881438e-01 -9.84157026e-01 1.07038605e+00
1.79770008e-01 -1.60346794e+00 1.28251165e-01 -8.98427591e-02
9.74416852e-01 -2.18495235e-01 4.06854987e-01 -4.79146838e-02
-1.63841352e-01 -1.12390351e+00 7.33944714e-01 6.33648396e-01
6.36385262e-01 -1.05952621e+00 7.84676015e-01 3.02541852e-01
-1.25622797e+00 -1.86630800e-01 -4.23602521e-01 -1.84721258e-02
9.88143161e-02 1.01178229e+00 -4.58785415e-01 8.37597489e-01
7.19551146e-01 1.03218913e+00 -2.86754549e-01 9.12432671e-01
-3.76210093e-01 5.38052857e-01 -2.86955270e-03 8.18681538e-01
4.03995544e-01 -1.65519133e-01 4.05622929e-01 1.01678896e+00
4.12149042e-01 -1.30381316e-01 8.98025855e-02 1.08514476e+00
-3.77716601e-01 -1.76624477e-01 -3.72271061e-01 9.90490243e-02
1.98617443e-01 1.28805006e+00 -5.12663126e-01 -2.37100035e-01
-4.84077454e-01 1.35731578e+00 3.05781692e-01 3.00898284e-01
-7.74885118e-01 -5.18036962e-01 6.26191080e-01 6.16040416e-02
6.58472478e-01 -3.15412760e-01 -2.84953058e-01 -1.08907616e+00
3.62116069e-01 -1.04322529e+00 -4.33694236e-02 -5.93907356e-01
-1.01348734e+00 6.93356693e-01 -4.72327411e-01 -1.42140055e+00
-1.24064669e-01 -2.59012371e-01 -5.01718521e-01 9.81648028e-01
-1.92112887e+00 -1.19344699e+00 -6.10273659e-01 6.25283360e-01
7.61756122e-01 -6.64350763e-03 6.72815740e-01 4.18783814e-01
-3.61181319e-01 4.18000400e-01 2.15726539e-01 1.59337386e-01
6.43035293e-01 -9.77983296e-01 2.58135915e-01 1.05776334e+00
1.41966403e-01 4.24904317e-01 6.36268556e-01 -4.72489864e-01
-1.52446926e+00 -1.29152524e+00 6.60026193e-01 2.03957021e-01
3.12923968e-01 -1.42989531e-01 -9.78548467e-01 3.34113181e-01
2.99053371e-01 3.78926009e-01 1.52936786e-01 -5.76271117e-01
-4.85188514e-01 -3.23883355e-01 -1.11090112e+00 3.26286972e-01
7.39399672e-01 -7.24081516e-01 -4.76635456e-01 4.88190502e-02
6.44257486e-01 -4.39012915e-01 -7.31386900e-01 4.07937318e-01
4.10200566e-01 -1.27938557e+00 1.00093496e+00 -1.36447266e-01
9.53345537e-01 -3.75429958e-01 -1.48615807e-01 -1.30514550e+00
-4.09680307e-01 -7.41596103e-01 -3.70157450e-01 8.81140113e-01
-9.74111483e-02 -1.88676193e-01 9.30151105e-01 7.63027743e-02
-4.04523432e-01 -8.36363554e-01 -9.11936343e-01 -4.50639427e-01
-1.59187898e-01 -2.25745663e-01 3.91807854e-01 6.57566905e-01
-4.56223190e-01 3.38265970e-02 -8.08172166e-01 1.94642887e-01
8.21719050e-01 3.83241773e-01 5.46409667e-01 -9.92229640e-01
-3.77053678e-01 -2.44781017e-01 -4.94238347e-01 -1.26309454e+00
6.45319149e-02 -5.13736665e-01 1.32735491e-01 -1.51610649e+00
2.50407219e-01 -1.82125941e-01 -4.39927608e-01 4.24335152e-01
-1.64827436e-01 5.11353135e-01 2.99162954e-01 2.34465495e-01
-3.29760075e-01 8.46327007e-01 1.28651500e+00 -2.59492993e-01
-2.08644971e-01 -3.32543314e-01 -5.56571841e-01 6.72625303e-01
6.09860897e-01 -4.10827249e-01 -2.71886230e-01 -6.58923209e-01
5.50906770e-02 1.69166818e-01 4.79501992e-01 -1.40427446e+00
2.46749848e-01 8.70112702e-02 7.57705748e-01 -6.35594130e-01
4.20146108e-01 -1.06944191e+00 8.29894319e-02 5.20427108e-01
-2.18150124e-01 -3.56872156e-02 2.40155369e-01 4.65365559e-01
-8.48504901e-01 1.14094496e-01 1.01510203e+00 -1.64288789e-01
-6.24559939e-01 3.68831813e-01 -1.05162159e-01 -2.52355158e-01
6.56024516e-01 -1.42542705e-01 -1.62484925e-02 -4.31455195e-01
-5.89151204e-01 -2.15353698e-01 4.32114631e-01 2.12896690e-01
9.24170375e-01 -1.24686253e+00 -7.30551600e-01 5.35021126e-01
-1.92365259e-01 1.11484602e-01 5.35243154e-01 8.90728056e-01
-7.13588774e-01 1.66340053e-01 -5.57725012e-01 -5.07114708e-01
-1.01581764e+00 4.53438103e-01 2.00427651e-01 -5.16608596e-01
-7.67160773e-01 8.87555122e-01 2.23849654e-01 3.40626031e-01
2.33707756e-01 -5.41801214e-01 4.52142358e-02 -1.33624852e-01
6.38799489e-01 1.96128502e-01 2.80537397e-01 -7.17191875e-01
-1.42721519e-01 6.22087479e-01 -3.65382016e-01 4.04230952e-02
1.67667007e+00 -9.25983265e-02 -3.57138932e-01 8.75194371e-02
1.44047654e+00 1.13302521e-01 -1.69851959e+00 -4.27094698e-01
-1.52202666e-01 -6.63747966e-01 3.06111693e-01 -4.56490725e-01
-1.46718097e+00 9.98689473e-01 5.43932676e-01 4.37574647e-02
1.70655000e+00 -3.71026844e-01 1.21725333e+00 3.99184078e-02
2.08645806e-01 -1.01410103e+00 3.15486759e-01 3.59550595e-01
1.04649830e+00 -9.83340740e-01 2.06729531e-01 -3.11495394e-01
-1.10317960e-01 1.21844804e+00 3.51657897e-01 -5.53623140e-01
5.04032552e-01 2.74345279e-01 -2.48232529e-01 -6.30900124e-03
-5.07873952e-01 3.43809579e-03 2.92890906e-01 3.60703379e-01
3.48360628e-01 -2.85018533e-01 -1.90407932e-01 3.18497032e-01
3.49416547e-02 7.58460462e-02 3.16304445e-01 1.09065497e+00
-4.86636519e-01 -1.18668973e+00 -3.07779521e-01 3.00452322e-01
-5.82202435e-01 -1.50217757e-01 -4.75698002e-02 5.31370521e-01
2.96727061e-01 7.94067502e-01 -5.55267148e-02 -3.02733600e-01
2.08230466e-01 -2.51970351e-01 4.82044607e-01 -3.46649349e-01
-5.31421244e-01 1.93089694e-01 -3.81293207e-01 -9.17812586e-01
-4.37189877e-01 -3.59533727e-01 -1.15404260e+00 -1.61414877e-01
4.94422093e-02 -1.09421520e-03 5.40905654e-01 8.42995942e-01
4.95545805e-01 6.51362956e-01 9.32939231e-01 -1.10127509e+00
-3.52257758e-01 -9.66719389e-01 -5.11715829e-01 5.33698380e-01
8.75594497e-01 -4.62347001e-01 -3.89420778e-01 2.69795179e-01] | [11.266134262084961, -1.7344385385513306] |
2a7a5cca-5a00-47c8-964d-09da343fa816 | evolution-in-groups-a-deeper-look-at-synaptic | 1704.02081 | null | http://arxiv.org/abs/1704.02081v1 | http://arxiv.org/pdf/1704.02081v1.pdf | Evolution in Groups: A deeper look at synaptic cluster driven evolution of deep neural networks | A promising paradigm for achieving highly efficient deep neural networks is
the idea of evolutionary deep intelligence, which mimics biological evolution
processes to progressively synthesize more efficient networks. A crucial design
factor in evolutionary deep intelligence is the genetic encoding scheme used to
simulate heredity and determine the architectures of offspring networks. In
this study, we take a deeper look at the notion of synaptic cluster-driven
evolution of deep neural networks which guides the evolution process towards
the formation of a highly sparse set of synaptic clusters in offspring
networks. Utilizing a synaptic cluster-driven genetic encoding, the
probabilistic encoding of synaptic traits considers not only individual
synaptic properties but also inter-synaptic relationships within a deep neural
network. This process results in highly sparse offspring networks which are
particularly tailored for parallel computational devices such as GPUs and deep
neural network accelerator chips. Comprehensive experimental results using four
well-known deep neural network architectures (LeNet-5, AlexNet, ResNet-56, and
DetectNet) on two different tasks (object categorization and object detection)
demonstrate the efficiency of the proposed method. Cluster-driven genetic
encoding scheme synthesizes networks that can achieve state-of-the-art
performance with significantly smaller number of synapses than that of the
original ancestor network. ($\sim$125-fold decrease in synapses for MNIST).
Furthermore, the improved cluster efficiency in the generated offspring
networks ($\sim$9.71-fold decrease in clusters for MNIST and a $\sim$8.16-fold
decrease in clusters for KITTI) is particularly useful for accelerated
performance on parallel computing hardware architectures such as those in GPUs
and deep neural network accelerator chips. | ['Elnaz Barshan', 'Alexander Wong', 'Mohammad Javad Shafiee'] | 2017-04-07 | null | null | null | null | ['object-categorization'] | ['computer-vision'] | [ 3.02667730e-03 -2.10673466e-01 6.29132867e-01 -3.31194758e-01
6.99427187e-01 -1.63362905e-01 2.03341410e-01 -6.26941770e-02
-7.09611475e-01 6.66800320e-01 -5.61223388e-01 -1.73832789e-01
-3.11433554e-01 -1.34077132e+00 -7.90330291e-01 -9.22781587e-01
-1.65289596e-01 4.74917710e-01 4.51516032e-01 -3.54656219e-01
3.67649019e-01 7.05747306e-01 -2.34793258e+00 3.21262956e-01
8.67879570e-01 9.88337517e-01 5.42170227e-01 8.38324964e-01
-3.05426210e-01 3.67830932e-01 -8.60307157e-01 -6.65206552e-01
1.88286409e-01 -4.96737421e-01 -1.12155005e-01 -2.90483326e-01
5.04300706e-02 3.49564195e-01 -2.19849497e-01 1.43151259e+00
6.98511183e-01 5.10013811e-02 8.72450292e-01 -1.08236206e+00
-7.73716569e-01 9.70987558e-01 -4.11360770e-01 4.73834664e-01
-7.13368535e-01 1.15669049e-01 6.19554281e-01 -5.87091327e-01
2.65384883e-01 1.14318800e+00 7.95979381e-01 7.12925076e-01
-8.42728972e-01 -9.49812531e-01 -2.25148872e-01 3.00993890e-01
-1.47456765e+00 -3.86029556e-02 3.61220092e-01 -2.27926925e-01
1.20513058e+00 1.79775015e-01 1.17770100e+00 5.51543653e-01
6.51269674e-01 4.73940700e-01 6.62383437e-01 -3.24232966e-01
5.52375257e-01 1.07616678e-01 2.20114678e-01 9.91021395e-01
7.05904841e-01 2.11209536e-01 -3.35218728e-01 1.38713211e-01
8.74498665e-01 -1.92970913e-02 2.50130057e-01 1.88768521e-01
-6.12332106e-01 8.74970019e-01 7.57150590e-01 3.64453286e-01
-4.95237529e-01 3.49996388e-01 4.80488747e-01 1.66726768e-01
-1.43518016e-01 5.48307955e-01 -3.09407622e-01 1.04404159e-01
-7.35426009e-01 8.28133970e-02 4.74520802e-01 7.67636180e-01
6.82993829e-01 6.80454910e-01 1.50210038e-01 1.13526344e+00
1.30392313e-01 2.95689911e-01 1.01378036e+00 -6.55211270e-01
-1.71640918e-01 8.85827959e-01 -7.23820448e-01 -9.43759382e-01
-4.55075234e-01 -9.17178929e-01 -1.28222024e+00 5.60061455e-01
-2.31462032e-01 -3.99229288e-01 -1.06467068e+00 1.68538570e+00
2.94206858e-01 2.75211513e-01 2.80271053e-01 5.36364138e-01
1.02602363e+00 7.63292074e-01 1.60501018e-01 2.07881317e-01
1.44373548e+00 -1.08136308e+00 -6.90168589e-02 -6.19960427e-02
3.56921881e-01 -4.34736729e-01 3.74794573e-01 1.89262941e-01
-1.06257820e+00 -1.08736169e+00 -1.45322287e+00 3.59895706e-01
-5.47271848e-01 3.83710414e-01 8.77941012e-01 9.73734498e-01
-1.28641903e+00 6.94679916e-01 -7.00871766e-01 -2.78857172e-01
5.64540744e-01 8.97059441e-01 1.51342526e-01 2.53989905e-01
-1.01941848e+00 5.21194518e-01 8.87942970e-01 -1.38968974e-01
-6.96279228e-01 -7.61761248e-01 -2.79293209e-01 4.23623353e-01
-5.33741653e-01 -9.60552514e-01 7.02761889e-01 -1.22955561e+00
-1.52585256e+00 7.41670549e-01 1.19088635e-01 -6.44559562e-01
2.05901898e-02 4.32382733e-01 -4.41724926e-01 -4.21405844e-02
-3.68118435e-01 1.26796949e+00 7.68507123e-01 -8.63797605e-01
-9.99999464e-01 -3.84390146e-01 -3.28209072e-01 9.61464569e-02
-7.31443286e-01 -4.11694162e-02 -2.41724789e-01 -6.98607266e-01
6.28965050e-02 -1.18303871e+00 -3.81144315e-01 1.02989592e-01
-8.33906606e-02 -3.50906014e-01 8.24402988e-01 4.10615839e-02
7.56901145e-01 -2.03020716e+00 1.17033668e-01 1.26853436e-01
3.16498876e-01 8.12530696e-01 -8.91484842e-02 -1.68136999e-01
-1.04146630e-01 -1.48428202e-01 -1.91634849e-01 3.45152318e-02
-3.46044749e-01 2.76406437e-01 1.73671752e-01 -4.05016951e-02
3.21562648e-01 9.24632668e-01 -3.88732314e-01 -1.89150646e-01
3.72513430e-03 6.07766449e-01 -7.83016682e-01 6.63035661e-02
-4.74728867e-02 5.40151596e-02 -2.94167042e-01 4.30181772e-01
7.47726917e-01 -6.35932460e-02 -6.53659776e-02 -1.45775020e-01
-3.13327640e-01 -4.31452125e-01 -7.73317218e-01 1.43447638e+00
-2.43946820e-01 6.75882578e-01 -1.97522983e-01 -1.11840320e+00
1.54338562e+00 -1.14626162e-01 2.01995343e-01 -7.16395795e-01
5.14155507e-01 1.72339186e-01 8.90260816e-01 -2.25871950e-01
4.49689955e-01 3.07441950e-02 2.20225453e-01 4.18388933e-01
4.51387763e-01 1.02983057e-01 2.76421905e-01 -2.41252765e-01
1.05217600e+00 -4.49875444e-01 -3.64812791e-01 -7.68620014e-01
5.65943301e-01 7.98309818e-02 6.28462791e-01 8.28357875e-01
-3.06910891e-02 3.22786719e-01 2.10184887e-01 -7.03949869e-01
-1.32547224e+00 -1.05137980e+00 -3.56901169e-01 1.12807894e+00
-1.35559617e-02 1.34488925e-01 -1.03555882e+00 1.35644302e-01
-1.41426697e-01 5.34347713e-01 -6.43389285e-01 -6.92242265e-01
-5.24292707e-01 -1.51309526e+00 1.00545633e+00 4.67736304e-01
9.35111046e-01 -1.35763049e+00 -1.03542662e+00 3.39131713e-01
8.17232668e-01 -7.02336192e-01 3.18862826e-01 4.82975274e-01
-1.15814638e+00 -6.69193804e-01 -5.97581089e-01 -1.16729212e+00
9.44147527e-01 -1.59367114e-01 9.14572299e-01 4.50608730e-01
-9.74297404e-01 -4.03073281e-01 -1.01927035e-01 -6.72262430e-01
-4.46415097e-01 1.53455406e-01 1.80063277e-01 -9.15637687e-02
3.47545564e-01 -9.46506023e-01 -7.57966578e-01 -1.64150055e-02
-7.39235282e-01 1.31693661e-01 7.89423525e-01 1.03595018e+00
4.52216178e-01 7.03500032e-01 5.36571085e-01 -8.28781426e-01
6.87537849e-01 -5.36305189e-01 -6.51395380e-01 3.46972406e-01
-5.38634300e-01 2.54131377e-01 9.37990785e-01 -6.93126500e-01
-1.11891353e+00 -6.77145794e-02 -4.62572604e-01 -2.58649498e-01
7.27274567e-02 3.38716149e-01 2.15679958e-01 -2.55871505e-01
6.88719690e-01 4.27404553e-01 -6.98053911e-02 -2.08388448e-01
-1.38079310e-02 4.29873586e-01 6.37076139e-01 -7.32358336e-01
2.76040643e-01 2.28800371e-01 2.26642311e-01 -7.87763059e-01
3.92949861e-03 3.03846866e-01 -4.92295980e-01 -2.58716524e-01
7.88454592e-01 -7.59037793e-01 -8.88303041e-01 8.32487345e-01
-1.05670023e+00 -2.25346442e-02 -8.90503526e-02 3.90210688e-01
-1.25357300e-01 -3.95434856e-01 -8.27032506e-01 -6.98591828e-01
-8.93716216e-01 -1.24774909e+00 3.46326917e-01 9.90417838e-01
6.05271049e-02 -8.22059095e-01 -1.97837457e-01 -1.96428269e-01
6.14693046e-01 -1.23125967e-03 1.47206509e+00 -5.20314991e-01
-3.95978570e-01 8.34044889e-02 -4.40246016e-01 3.36348265e-01
-1.50002927e-01 4.73991483e-01 -7.06125379e-01 -1.41261220e-01
-1.41317872e-02 -6.07370622e-02 9.88024831e-01 4.38631117e-01
1.19744718e+00 1.23544812e-01 -4.21971917e-01 8.70703697e-01
1.69095635e+00 9.43984866e-01 6.63082480e-01 2.30627924e-01
7.07715273e-01 5.65352261e-01 -1.02903694e-01 5.08696795e-01
-7.07328916e-02 3.00125271e-01 3.66109252e-01 3.04274559e-02
-4.66267556e-01 1.67281389e-01 -1.27769962e-01 1.30381143e+00
-1.57262772e-01 -3.31432760e-01 -9.54175532e-01 4.71665531e-01
-1.67935145e+00 -7.69381940e-01 -1.69938996e-01 1.85173857e+00
7.10020781e-01 1.03891954e-01 -2.79125899e-01 9.95551571e-02
1.16405737e+00 -4.49953079e-01 -9.77579534e-01 -8.93362343e-01
-3.47760826e-01 7.99169421e-01 3.11259329e-01 -1.70207232e-01
-5.60474455e-01 9.99800324e-01 5.85248995e+00 9.70418990e-01
-1.34244227e+00 8.21921900e-02 9.55835640e-01 -3.05859715e-01
1.42089620e-01 -4.36937928e-01 -1.12808931e+00 7.67409742e-01
1.10213506e+00 -1.71156585e-01 2.64593393e-01 8.86371434e-01
-3.98206472e-01 3.64543945e-02 -6.94219053e-01 1.01269972e+00
-1.89257517e-01 -1.81790566e+00 1.85526893e-01 -4.84971553e-02
1.08031213e+00 3.34625393e-01 2.85587370e-01 1.27707511e-01
5.14900267e-01 -1.01077330e+00 8.03279757e-01 3.09748441e-01
6.18768156e-01 -1.47053921e+00 6.96342587e-01 1.80889413e-01
-1.03345311e+00 -4.68346566e-01 -1.20761991e+00 -4.16515814e-03
-2.67577112e-01 5.84637880e-01 -5.29616356e-01 -4.41774800e-02
1.19876742e+00 4.55908090e-01 -5.28301239e-01 1.13785374e+00
1.50301799e-01 2.71401018e-01 -3.30813527e-01 -6.78988457e-01
3.22827071e-01 -2.45316759e-01 1.99832380e-01 1.28342366e+00
6.33339465e-01 2.90983886e-01 -6.49718761e-01 1.10170090e+00
-3.82368892e-01 -2.26118967e-01 -3.42748404e-01 1.68493260e-02
7.25678623e-01 1.33746088e+00 -1.28290200e+00 -4.43740934e-01
1.05156176e-01 8.08957517e-01 5.54031551e-01 -1.29305989e-01
-9.58970726e-01 -7.21840978e-01 1.18215847e+00 -4.09888476e-01
6.17382169e-01 -1.24954924e-01 -7.03872204e-01 -5.25703967e-01
-5.69576442e-01 -6.41873538e-01 -8.47861916e-02 -5.69258690e-01
-9.42358255e-01 1.21161234e+00 -5.08494496e-01 -6.78904712e-01
1.75363384e-02 -8.55792701e-01 -9.79895711e-01 7.10299671e-01
-8.75196517e-01 -5.22027671e-01 -3.98573816e-01 4.81689036e-01
6.15168571e-01 -8.22045326e-01 8.09511065e-01 3.74521673e-01
-1.01715422e+00 7.53414333e-01 4.58776116e-01 1.66335721e-02
-2.00226739e-01 -6.60088003e-01 7.18622744e-01 7.14454949e-01
1.23506047e-01 7.11789370e-01 4.11096632e-01 -4.74154860e-01
-1.32728219e+00 -1.29950213e+00 2.57331997e-01 2.56018162e-01
4.15774941e-01 -3.00671518e-01 -8.49991858e-01 -3.54574295e-03
1.90351635e-01 -1.98452905e-01 6.41508043e-01 -3.48148614e-01
-1.22861816e-02 -3.00720423e-01 -1.32641816e+00 7.23752379e-01
1.26657808e+00 6.93480372e-02 -2.30155483e-01 8.17650333e-02
5.21865368e-01 -1.71644405e-01 -7.13977933e-01 5.54566145e-01
6.73518181e-01 -1.30478299e+00 9.06527996e-01 -3.96761626e-01
5.24056196e-01 -2.04740524e-01 8.95501077e-02 -1.13572121e+00
-6.75019145e-01 3.36344913e-02 3.13141137e-01 1.09606695e+00
4.20023650e-01 -7.74221838e-01 1.19692457e+00 4.16683018e-01
-3.95473838e-01 -9.25918758e-01 -7.93795049e-01 -6.31531239e-01
2.50184853e-02 -2.35049473e-03 8.47363889e-01 6.31706953e-01
-7.63408363e-01 6.65385798e-02 3.04557115e-01 9.55161825e-02
6.65700972e-01 -1.05615944e-01 1.64889991e-01 -1.33971214e+00
-4.36081856e-01 -9.16819692e-01 -9.86419499e-01 -3.96916837e-01
1.48067474e-01 -7.58465767e-01 -3.33310403e-02 -1.00608408e+00
1.44219697e-01 -1.09119689e+00 -5.24120331e-01 2.53997028e-01
8.76464546e-02 5.18875241e-01 8.11740011e-02 1.52694046e-01
-3.90474014e-02 5.46478927e-01 8.85799289e-01 -9.14462656e-02
-2.58224644e-02 -1.04189359e-01 -6.11719012e-01 6.22046053e-01
9.61672127e-01 -8.03337455e-01 -7.09343910e-01 -9.00608540e-01
1.27260879e-01 -5.10647058e-01 4.18933295e-03 -1.73947406e+00
7.33125567e-01 3.55983764e-01 7.57825673e-01 -3.45720381e-01
3.48123729e-01 -4.38315600e-01 7.54687250e-01 1.10676253e+00
-2.09960192e-01 6.16207659e-01 5.26356101e-01 1.80522248e-01
-1.85912460e-01 -7.49577880e-01 1.17046213e+00 -2.62286842e-01
-8.47219348e-01 2.77179360e-01 -6.20031893e-01 -3.46965939e-01
1.17851019e+00 -6.74006343e-01 -4.13425535e-01 4.54322547e-01
-4.88763332e-01 -2.65799254e-01 1.26723513e-01 5.01744151e-01
8.35260510e-01 -1.12561095e+00 -6.64207160e-01 4.91989613e-01
-4.69554335e-01 -1.15268812e-01 2.82949626e-01 1.96410224e-01
-1.04723954e+00 1.91709220e-01 -1.08563912e+00 -6.85963631e-01
-1.34344518e+00 1.65866643e-01 4.51887697e-01 3.11437368e-01
-3.49639386e-01 1.76196218e+00 3.04498315e-01 -2.44546264e-01
1.53306974e-02 -2.78955493e-02 -3.27178717e-01 -2.21036568e-01
4.82951462e-01 3.24085206e-01 2.53317475e-01 -4.97209698e-01
-2.83173412e-01 6.33296311e-01 -2.81685442e-01 2.23520681e-01
1.65691698e+00 3.53266031e-01 -6.88067913e-01 -6.19812720e-02
1.24884510e+00 -7.31561184e-01 -1.07656455e+00 1.55803934e-01
-1.22275516e-01 1.35885537e-01 1.18003637e-01 -6.07176125e-01
-1.79251766e+00 8.30954790e-01 9.79363978e-01 -1.16258480e-01
1.27111447e+00 -2.80017853e-01 8.24477434e-01 5.12004137e-01
4.92140502e-01 -1.02838218e+00 1.11467794e-01 6.59628689e-01
4.80253279e-01 -6.93716824e-01 -5.03672920e-02 -1.09649807e-01
-2.69827724e-01 1.36837268e+00 1.11534441e+00 -5.44322014e-01
7.83268631e-01 5.69263935e-01 -2.16068849e-01 -2.51692653e-01
-9.06277001e-01 8.20423663e-02 -2.49546040e-02 6.49248660e-01
2.75235623e-01 1.76727131e-01 -2.68997461e-01 4.69167888e-01
-6.22089744e-01 -2.62405843e-01 2.39031464e-01 8.52684557e-01
-6.47298574e-01 -8.92700374e-01 -1.36210006e-02 6.23989820e-01
-3.08431894e-01 -3.10197473e-01 -1.05201095e-01 3.99292052e-01
7.76991010e-01 5.29265404e-01 6.89632654e-01 -8.01343203e-01
-4.73182537e-02 -3.06728274e-01 6.19168282e-01 -4.98720258e-01
-1.18182302e+00 -5.73068500e-01 -2.93448329e-01 4.45103832e-02
-5.31163923e-02 -3.07293206e-01 -1.68347132e+00 -6.12396955e-01
-2.18900606e-01 2.24536285e-01 1.30632436e+00 4.62787360e-01
7.36809492e-01 9.80085790e-01 4.30900723e-01 -8.28690708e-01
5.83233014e-02 -5.96919239e-01 -5.18673122e-01 -5.93611561e-02
-5.54931343e-01 -5.99668026e-01 7.39144310e-02 -1.36408687e-01] | [8.370542526245117, 3.1074445247650146] |
fe689b32-9c37-4e17-aaef-704512185737 | hih-towards-more-accurate-face-alignment-via | 2104.03100 | null | https://arxiv.org/abs/2104.03100v2 | https://arxiv.org/pdf/2104.03100v2.pdf | HIH: Towards More Accurate Face Alignment via Heatmap in Heatmap | Heatmap-based regression overcomes the lack of spatial and contextual information of direct coordinate regression, and has revolutionized the task of face alignment. Yet it suffers from quantization errors caused by neglecting subpixel coordinates in image resizing and network downsampling. In this paper, we first quantitatively analyze the quantization error on benchmarks, which accounts for more than 1/3 of the whole prediction errors for state-of-the-art methods. To tackle this problem, we propose a novel Heatmap In Heatmap(HIH) representation and a coordinate soft-classification (CSC) method, which are seamlessly integrated into the classic hourglass network. The HIH representation utilizes nested heatmaps to jointly represent the coordinate label: one heatmap called integer heatmap stands for the integer coordinate, and the other heatmap named decimal heatmap represents the subpixel coordinate. The range of a decimal heatmap makes up one pixel in the corresponding integer heatmap. Besides, we transfer the offset regression problem to an interval classification task, and CSC regards the confidence of the pixel as the probability of the interval. Meanwhile, CSC applying the distribution loss leverage the soft labels generated from the Gaussian distribution function to guide the offset heatmap training, which makes it easier to learn the distribution of coordinate offsets. Extensive experiments on challenging benchmark datasets demonstrate that our HIH can achieve state-of-the-art results. In particular, our HIH reaches 4.08 NME (Normalized Mean Error) on WFLW, and 3.21 on COFW, which exceeds previous methods by a significant margin. | ['Jian Cheng', 'Jian Xue', 'Qiang Chen', 'Qinghao Hu', 'Xing Lan'] | 2021-04-07 | null | null | null | null | ['face-alignment'] | ['computer-vision'] | [ 1.40415117e-01 2.40250677e-01 -4.27985340e-01 -6.87063754e-01
-6.41537130e-01 -2.39565447e-01 3.87261182e-01 -9.03812423e-02
-1.20356366e-01 7.75251806e-01 1.22857420e-02 -8.86511579e-02
-3.98829617e-02 -6.84565008e-01 -9.34095085e-01 -9.03282285e-01
2.90991217e-01 2.83259124e-01 -1.45340890e-01 -9.26591828e-02
-1.70912873e-03 3.18058610e-01 -1.23175073e+00 1.52261732e-02
9.72136855e-01 1.42718279e+00 -1.19877398e-01 4.00194637e-02
-7.81396255e-02 4.32632446e-01 -7.96014249e-01 -6.35452271e-01
2.69611657e-01 -3.26231778e-01 -4.84272450e-01 -3.72004688e-01
8.24667156e-01 -3.33174944e-01 -1.74767271e-01 1.12593830e+00
6.03789568e-01 -1.04833283e-01 5.99689364e-01 -1.61736870e+00
-9.08057928e-01 5.25810778e-01 -1.14346623e+00 -2.38774881e-01
3.87533456e-02 -1.92890093e-01 7.59301424e-01 -9.36863124e-01
5.37784874e-01 1.17603230e+00 9.20970440e-01 4.47408557e-01
-1.22219300e+00 -1.14768827e+00 1.05695620e-01 1.82043463e-01
-1.84145725e+00 -8.80663544e-02 7.28656888e-01 -3.30786854e-01
3.95982027e-01 3.63390714e-01 7.00425327e-01 6.46480799e-01
1.34441128e-03 6.77759945e-01 1.05191910e+00 -3.35055053e-01
1.06555156e-01 8.94855522e-03 -8.42183605e-02 7.29857385e-01
2.79473606e-02 -1.50163621e-01 -5.24104118e-01 1.71491534e-01
9.46685970e-01 6.55760169e-02 -2.22599342e-01 -3.89406294e-01
-9.42364514e-01 6.97151601e-01 8.88452947e-01 -7.38738552e-02
2.56784391e-02 2.34877259e-01 1.45058274e-01 -1.26873091e-01
6.80430233e-01 1.31281108e-01 -3.41470778e-01 -1.61136702e-01
-1.19155502e+00 1.45246252e-01 2.61064649e-01 1.09426534e+00
9.33779120e-01 -2.70176142e-01 -4.67512131e-01 1.18355346e+00
-6.00286387e-02 4.63352889e-01 2.10654601e-01 -9.61803377e-01
7.86235869e-01 6.78288817e-01 -2.20531777e-01 -1.46313500e+00
-4.62143511e-01 -3.87894005e-01 -1.23728406e+00 1.19106211e-01
6.62473619e-01 -1.82307512e-01 -8.07016313e-01 1.83645308e+00
4.60644513e-01 2.41575629e-01 -6.22582972e-01 9.13735151e-01
6.24707401e-01 6.07057869e-01 -1.08774886e-01 -1.50254950e-01
1.26601636e+00 -1.01305771e+00 -6.37196779e-01 5.75305037e-02
4.57642645e-01 -6.43678963e-01 1.19075406e+00 2.16500297e-01
-9.19094861e-01 -6.72902226e-01 -1.29414320e+00 -2.42638767e-01
-3.95211786e-01 4.35413897e-01 4.95361388e-01 5.21987259e-01
-1.02472591e+00 5.90905190e-01 -6.97637737e-01 3.72497961e-02
6.55169070e-01 4.46876913e-01 -3.97629827e-01 -3.84611376e-02
-1.05514336e+00 5.77324331e-01 9.26154479e-02 3.19995612e-01
-2.12599132e-02 -1.11235678e+00 -9.18715954e-01 7.58900493e-02
3.58874857e-01 -2.17073336e-01 9.04258370e-01 -6.72204256e-01
-1.50955033e+00 7.56767154e-01 -1.77338168e-01 -1.18777186e-01
6.81291163e-01 -1.42969877e-01 -3.16895247e-01 2.28871182e-02
1.85175925e-01 1.10863245e+00 7.87535906e-01 -9.04997528e-01
-5.51179528e-01 -4.59570169e-01 -2.42235288e-01 6.81559592e-02
-5.96555591e-01 -2.03601271e-01 -8.84674430e-01 -9.72482622e-01
2.52318293e-01 -8.76466334e-01 1.53778329e-01 3.72017711e-01
-7.25873768e-01 -3.22466165e-01 6.83477223e-01 -6.82813764e-01
1.60759223e+00 -2.19113493e+00 8.33837464e-02 4.19849187e-01
2.38265142e-01 -1.01608001e-02 1.14995398e-01 -1.46601856e-01
-3.11231107e-01 2.06064001e-01 -1.75638288e-01 -5.78811169e-01
1.05146326e-01 -5.85521199e-02 -2.31125444e-01 5.28987408e-01
3.32726389e-01 7.87583709e-01 -5.82090974e-01 -6.84671283e-01
1.13037325e-01 7.83431888e-01 -5.22228718e-01 -6.48137704e-02
1.70388892e-01 4.37020898e-01 -1.22176968e-01 6.10893846e-01
1.12334847e+00 -3.17935646e-01 -1.40682325e-01 -5.90719581e-01
-2.04022184e-01 5.27849980e-03 -1.04611695e+00 1.78490603e+00
-2.47051969e-01 6.11806929e-01 -3.59148383e-01 -5.16310155e-01
1.19514942e+00 -2.53273070e-01 6.36705518e-01 -8.30279410e-01
7.50481486e-02 -1.41560093e-01 -4.09523904e-01 6.64415285e-02
6.48442268e-01 2.87245214e-01 -1.00004159e-01 1.32358536e-01
-1.47844672e-01 1.66369509e-02 -3.89140807e-02 -2.22187877e-01
5.26769221e-01 3.54924679e-01 1.92789664e-03 -8.19058418e-02
4.79732931e-01 -1.91956207e-01 9.70529497e-01 3.15342188e-01
-1.55394435e-01 1.16433263e+00 8.46097767e-01 -5.17667294e-01
-1.03827059e+00 -9.39943194e-01 -5.95319450e-01 9.44672883e-01
3.76338661e-01 -6.62714303e-01 -1.25497210e+00 -5.08745253e-01
1.62994966e-01 3.40659916e-01 -8.67032170e-01 -1.17865294e-01
-7.46683419e-01 -8.72489572e-01 5.03152251e-01 7.06874967e-01
7.96279430e-01 -5.15204489e-01 -1.32722884e-01 -9.46552753e-02
-1.71634644e-01 -9.23255563e-01 -7.50400901e-01 -1.88409954e-01
-2.60984540e-01 -9.35866058e-01 -9.63035464e-01 -6.59639895e-01
8.18967760e-01 7.62104169e-02 9.53970373e-01 -6.24974221e-02
-3.82815808e-01 -3.80768418e-01 -1.29897103e-01 -1.20980836e-01
3.49594891e-01 3.55701953e-01 -6.54195920e-02 4.26076315e-02
3.09361845e-01 -3.51139277e-01 -8.80116165e-01 6.31785691e-01
-4.45137352e-01 4.35530096e-01 2.31069252e-01 9.61966574e-01
9.98838067e-01 -6.61390051e-02 3.95719916e-01 -7.90417135e-01
2.07952797e-01 -3.45807552e-01 -7.41729915e-01 3.15360069e-01
-7.66791105e-01 -9.79692303e-03 5.74374020e-01 -2.83601820e-01
-6.60589397e-01 6.57747611e-02 -2.22021211e-02 -4.55752462e-01
1.57172546e-01 1.60259455e-01 -3.06140155e-01 -1.81967765e-01
2.70349771e-01 -1.88397557e-01 9.86012369e-02 -2.96130031e-01
5.79507172e-01 6.42514706e-01 7.84626961e-01 -5.63793004e-01
8.27042878e-01 3.84114176e-01 1.71021193e-01 -5.08912563e-01
-7.35891521e-01 -1.02519222e-01 -7.18644202e-01 -1.94875509e-01
1.08577776e+00 -8.48390460e-01 -9.32793260e-01 2.77806997e-01
-1.03786135e+00 -2.79312611e-01 -7.53049552e-02 2.65155286e-01
-4.18661565e-01 2.76384540e-02 -5.26812434e-01 -3.84375751e-01
-1.82900220e-01 -1.17150676e+00 1.30603051e+00 5.59264004e-01
-9.70474184e-02 -6.23258710e-01 -1.54694721e-01 3.02207768e-01
2.95426518e-01 4.40292835e-01 9.04396951e-01 -1.65067405e-01
-3.26597899e-01 -3.00109476e-01 -6.80474341e-01 2.90209591e-01
-3.63909304e-02 8.28554630e-02 -9.67661440e-01 -1.97133318e-01
-4.50232267e-01 -2.02080816e-01 5.97507000e-01 3.81356031e-01
1.88960695e+00 -3.07914108e-01 -2.77454078e-01 1.13092470e+00
1.16424620e+00 -2.78912503e-02 7.13465035e-01 3.30456525e-01
1.07508993e+00 4.63950604e-01 7.72437394e-01 4.06904578e-01
6.58318162e-01 1.17723978e+00 4.42772627e-01 -3.96402657e-01
-1.73223868e-01 -4.23000395e-01 -8.49574581e-02 5.36089003e-01
-5.25615700e-02 1.47799551e-01 -7.93185592e-01 7.64967920e-03
-1.82574046e+00 -7.73471653e-01 6.75367713e-02 2.41093421e+00
1.14307904e+00 5.41764684e-02 1.13808744e-01 5.24275042e-02
1.06340659e+00 3.25425148e-01 -4.43875462e-01 -2.72880215e-02
-2.47389019e-01 5.89591749e-02 7.31975138e-01 3.35075855e-01
-1.33523309e+00 8.79134178e-01 5.07650375e+00 1.30621421e+00
-1.18232679e+00 -2.28972230e-02 1.19339013e+00 -2.18246385e-01
8.59503523e-02 -3.34330082e-01 -1.21563542e+00 8.77011776e-01
4.28996861e-01 1.10574998e-01 5.11427104e-01 9.22070801e-01
7.53578544e-02 3.56355049e-02 -1.04671788e+00 1.31577396e+00
2.84898281e-01 -1.21291375e+00 -1.82647228e-01 8.86799917e-02
9.77544844e-01 -3.04439455e-01 4.98810679e-01 2.01767892e-01
-5.63925169e-02 -1.36408591e+00 7.67475605e-01 5.35467744e-01
1.39924657e+00 -9.06278253e-01 5.93442321e-01 -4.72367601e-03
-1.32741404e+00 1.29698008e-01 -5.61652839e-01 1.52490526e-01
-2.03870162e-01 5.96823633e-01 -7.12591410e-01 4.25423056e-01
8.41593206e-01 7.61339068e-01 -8.02040637e-01 9.90197301e-01
-3.23489696e-01 2.54115850e-01 -4.93312091e-01 1.15082406e-01
1.10634558e-01 -3.23343158e-01 -6.88598603e-02 9.80557859e-01
4.41315681e-01 -2.76332736e-01 7.59150684e-02 9.10104990e-01
-4.68544155e-01 2.12471008e-01 -2.02748716e-01 5.80064178e-01
8.92979860e-01 1.26476347e+00 -5.64053178e-01 -1.72507718e-01
-3.23816538e-01 8.13523591e-01 4.64900523e-01 2.48734564e-01
-1.18131304e+00 -6.75429225e-01 6.51855826e-01 3.15692313e-02
1.78330302e-01 1.64858326e-01 -8.51413965e-01 -7.71279454e-01
3.74870032e-01 -4.23055798e-01 1.54807001e-01 -8.50938141e-01
-1.06263173e+00 6.49881363e-01 6.43414333e-02 -1.43197572e+00
-3.64730537e-01 -4.58073735e-01 -5.07448137e-01 9.68213737e-01
-1.46796858e+00 -1.20409441e+00 -7.60881841e-01 2.30684191e-01
2.14221194e-01 -2.27047391e-02 5.99090576e-01 5.61447144e-01
-1.08915770e+00 1.33498561e+00 1.70017198e-01 3.75786781e-01
1.06668842e+00 -1.21249449e+00 2.92091042e-01 4.99660224e-01
-1.56143978e-01 5.66091001e-01 3.20134521e-01 -4.95259315e-01
-8.97123933e-01 -1.40880704e+00 7.39563346e-01 -2.59479821e-01
3.61904413e-01 -5.12899995e-01 -8.27258110e-01 4.68436331e-01
-2.34291360e-01 4.47175503e-01 5.74527740e-01 -3.07848491e-02
-5.90547264e-01 -5.82328260e-01 -1.16284263e+00 6.31687939e-01
9.59443569e-01 -3.77248913e-01 -2.55246423e-02 4.21294183e-01
7.54789293e-01 -6.77264035e-01 -9.93767381e-01 5.77945709e-01
8.33079457e-01 -7.94103563e-01 1.12248528e+00 -1.70458347e-01
7.01181650e-01 -5.52811921e-01 -2.97798794e-02 -1.22382843e+00
-3.27991366e-01 -5.12362659e-01 8.41093808e-03 1.48332524e+00
4.56453323e-01 -4.00788784e-01 1.10569489e+00 6.62202179e-01
-9.65459943e-02 -1.26708961e+00 -1.14481127e+00 -4.67870235e-01
1.87892482e-01 -2.09701046e-01 1.24184608e+00 1.07714641e+00
1.61321498e-02 -1.75638109e-01 -3.80533010e-01 -3.38332774e-03
5.48658788e-01 5.81788039e-03 8.38438213e-01 -9.83895063e-01
-3.89317460e-02 -7.72331953e-01 -6.18774116e-01 -1.13909400e+00
1.65901765e-01 -7.45585740e-01 2.25071646e-02 -1.02671731e+00
5.77959642e-02 -7.22128272e-01 -1.66224316e-01 7.42077708e-01
-2.53231168e-01 8.47590089e-01 2.84736067e-01 2.01297358e-01
-4.60028499e-01 6.06573105e-01 1.18865764e+00 -3.17182988e-01
-2.07132041e-01 -1.78243488e-01 -7.19461083e-01 8.44316959e-01
7.84407616e-01 -2.74753660e-01 -2.10722402e-01 -5.64590454e-01
3.00485462e-01 -1.67555228e-01 3.05987626e-01 -1.02816737e+00
5.13601303e-01 -2.33593490e-02 8.53746414e-01 -8.03356647e-01
3.88503700e-01 -8.06739092e-01 6.90804794e-03 -5.99153750e-02
-3.41680646e-01 8.77947584e-02 2.98504978e-02 2.96602935e-01
-2.00200811e-01 2.61099160e-01 7.70683229e-01 5.82324028e-01
-3.59174848e-01 3.97212386e-01 6.35887265e-01 -1.03324465e-01
1.01355338e+00 -3.25862944e-01 -6.74339771e-01 -1.55449912e-01
-4.70124722e-01 3.42661589e-01 4.50256586e-01 4.35775936e-01
4.29685503e-01 -1.83115113e+00 -4.89434183e-01 5.86994886e-01
2.02459633e-01 3.61991733e-01 1.42778054e-01 9.17870581e-01
-7.00571835e-01 3.00581574e-01 -1.62099257e-01 -7.81760156e-01
-1.09533274e+00 1.52221233e-01 2.28383109e-01 4.89045419e-02
-4.49701160e-01 1.00551963e+00 1.72441170e-01 -2.92149216e-01
5.68967342e-01 -3.90631258e-01 -2.45077625e-01 4.90735890e-03
7.08347678e-01 5.76549768e-01 8.40821862e-02 -7.81961203e-01
-2.38200545e-01 1.20458221e+00 -1.78504273e-01 1.81033775e-01
1.09400845e+00 -4.00128998e-02 -3.83713692e-01 3.11556786e-01
1.49207628e+00 -1.35965094e-01 -1.60418820e+00 -3.87533680e-02
-3.36241841e-01 -5.17789781e-01 -8.62800255e-02 -8.18531752e-01
-1.33209002e+00 9.22722578e-01 6.18311763e-01 -9.97132212e-02
1.14838135e+00 -1.70639470e-01 6.85669601e-01 -4.10762355e-02
2.21045375e-01 -1.19891858e+00 -2.56056130e-01 2.54644066e-01
7.63520539e-01 -1.15412450e+00 2.36757889e-01 -8.70480657e-01
-5.10517299e-01 1.07139063e+00 8.18717241e-01 1.46020539e-02
5.41277766e-01 2.04146445e-01 -9.36724097e-02 1.32510781e-01
-1.74089760e-01 2.85808206e-01 4.17348176e-01 4.24204618e-01
3.71728301e-01 1.52691156e-01 -1.36330321e-01 7.30056405e-01
-7.97740698e-01 -6.92308843e-02 -6.43012449e-02 4.61268932e-01
-3.92392837e-02 -7.92056799e-01 -4.49327737e-01 4.13597852e-01
-3.75430882e-01 -4.99573164e-02 -1.27964213e-01 7.26945400e-01
4.36913341e-01 6.34759724e-01 5.80445588e-01 -7.61217296e-01
2.18507186e-01 -1.15612812e-01 2.51583517e-01 -1.57453880e-01
-1.89589053e-01 7.13517740e-02 -3.76579493e-01 -7.42106855e-01
8.03371221e-02 -4.29381430e-01 -1.37964404e+00 -4.52485830e-01
-2.51178086e-01 5.58075774e-03 8.46784770e-01 6.83844268e-01
4.03443992e-01 4.80590254e-01 8.80523801e-01 -8.65935683e-01
-3.37025195e-01 -9.79267359e-01 -4.51533645e-01 3.20509166e-01
2.30601385e-01 -8.62911642e-01 -2.97666192e-01 -1.55814752e-01] | [13.452207565307617, 0.49345308542251587] |
094377cd-ee72-4469-82d2-725e18d0a396 | learning-to-compose-soft-prompts-for | 2204.03574 | null | https://arxiv.org/abs/2204.03574v3 | https://arxiv.org/pdf/2204.03574v3.pdf | Learning to Compose Soft Prompts for Compositional Zero-Shot Learning | We introduce compositional soft prompting (CSP), a parameter-efficient learning technique to improve the zero-shot compositionality of large-scale pretrained vision-language models (VLMs) like CLIP. We develop CSP for compositional zero-shot learning, the task of predicting unseen attribute-object compositions (e.g., old cat and young tiger). VLMs have a flexible text encoder that can represent arbitrary classes as natural language prompts but they often underperform task-specific architectures on the compositional zero-shot benchmark datasets. CSP treats the attributes and objects that define classes as learnable tokens of vocabulary. During training, the vocabulary is tuned to recognize classes that compose tokens in multiple ways (e.g., old cat and white cat). At test time, we recompose the learned attribute-object vocabulary in new combinations to recognize novel classes. We show that CSP outperforms the CLIP on benchmark datasets by an average of 10.9 percentage points on AUC. CSP also outperforms CoOp, a soft prompting method that fine-tunes the prefix context tokens, by an average of 5.8 percentage points on AUC. We perform additional experiments to show that CSP improves generalization to higher-order attribute-attribute-object compositions (e.g., old white cat) and combinations of pretrained attributes and fine-tuned objects. The code is available at https://github.com/BatsResearch/csp. | ['Stephen H. Bach', 'Peilin Yu', 'Nihal V. Nayak'] | 2022-04-07 | null | null | null | null | ['compositional-zero-shot-learning'] | ['computer-vision'] | [ 4.37062830e-01 3.02417427e-02 -2.53883898e-01 -5.24156034e-01
-9.40909386e-01 -8.54879200e-01 7.98944175e-01 1.53210852e-02
-6.20146275e-01 4.76040900e-01 1.90101400e-01 4.49047796e-02
4.02029991e-01 -5.88735819e-01 -1.07265759e+00 -6.57768965e-01
3.25069539e-02 7.01757014e-01 5.00744760e-01 -7.06602708e-02
-1.12359956e-01 1.63313910e-01 -1.95157599e+00 7.17031121e-01
4.65179235e-01 1.04164279e+00 4.29126740e-01 8.27990890e-01
-2.70740300e-01 8.10179293e-01 -5.18286645e-01 -6.26049936e-01
3.41179192e-01 -1.37867466e-01 -5.28525710e-01 -8.18137378e-02
1.09676862e+00 -4.96092498e-01 -2.31944770e-01 9.14414346e-01
4.33903903e-01 2.43835762e-01 7.60826051e-01 -1.49409783e+00
-1.10530674e+00 1.07790637e+00 -1.66018710e-01 3.02102417e-01
-2.35835500e-02 5.83482623e-01 1.44637966e+00 -1.33356977e+00
6.82941437e-01 1.15339899e+00 4.63791937e-01 9.43124652e-01
-1.50675547e+00 -9.91720021e-01 3.93934697e-01 4.55911398e-01
-1.34402430e+00 -5.83946407e-01 5.03717363e-01 -6.56810045e-01
1.15575099e+00 2.49461800e-01 3.43119562e-01 1.48158503e+00
-1.71778291e-01 1.09810066e+00 7.43213713e-01 -1.76839933e-01
3.62707734e-01 1.37707079e-02 3.14682275e-01 5.01407921e-01
4.11431119e-02 2.45090187e-01 -7.83459187e-01 -2.21365750e-01
2.04018682e-01 -3.90367466e-03 3.69060077e-02 -4.10123825e-01
-1.38363922e+00 8.55652571e-01 3.43219817e-01 -2.58553177e-02
-2.45268553e-01 4.19828773e-01 4.59592760e-01 3.54705513e-01
2.05255464e-01 5.87920368e-01 -6.85523689e-01 -4.34868112e-02
-7.51681387e-01 2.04415634e-01 5.25562882e-01 1.46372151e+00
8.34060550e-01 3.97982746e-01 -5.55886984e-01 1.08979964e+00
-1.33279800e-01 7.17190385e-01 4.63306695e-01 -8.20507109e-01
9.59327444e-02 2.93160945e-01 -3.10372293e-01 2.50167791e-02
-4.63688411e-02 -4.42466080e-01 -4.74441171e-01 1.15726767e-02
6.43957183e-02 -3.08427471e-03 -1.44004655e+00 2.04945183e+00
1.70214191e-01 3.96299750e-01 5.44871315e-02 6.64494216e-01
1.19971120e+00 7.32890904e-01 5.27702093e-01 6.07467331e-02
1.53817141e+00 -9.17537153e-01 -3.37095112e-01 -4.69992012e-01
3.62020940e-01 -5.31394958e-01 1.60069430e+00 1.00733653e-01
-9.03581917e-01 -6.12961411e-01 -8.75423551e-01 -2.59864479e-01
-4.11704779e-01 -9.08193812e-02 5.05875885e-01 4.95552383e-02
-8.85038316e-01 2.79746532e-01 -6.67988658e-01 -3.21930975e-01
5.95756412e-01 3.40790093e-01 -2.10364670e-01 -3.83800901e-02
-1.14833426e+00 5.67551076e-01 4.55802172e-01 -7.09201932e-01
-1.58660531e+00 -1.30052185e+00 -1.00297070e+00 1.84885263e-01
5.77198148e-01 -5.62886894e-01 1.78886116e+00 -1.17030251e+00
-1.09748828e+00 1.13269782e+00 -1.51330456e-01 -6.44069433e-01
1.05307251e-01 -5.88873774e-02 -2.24116772e-01 1.70697302e-01
2.98832059e-01 1.26237333e+00 1.15236652e+00 -1.09898198e+00
-7.56107390e-01 -7.75991827e-02 2.03094184e-02 -8.10883939e-03
-2.37833112e-01 2.12531894e-01 -1.84489653e-01 -8.56397092e-01
-2.56106108e-01 -8.91013563e-01 2.73361057e-01 2.64107376e-01
-1.67182907e-01 -4.53663260e-01 5.18092453e-01 -2.22858101e-01
7.14062929e-01 -2.36601377e+00 1.01061195e-01 -3.06915462e-01
3.38101000e-01 3.55003059e-01 -5.47908902e-01 1.75599426e-01
-4.11582701e-02 -1.46353655e-02 -4.30447996e-01 -3.13223273e-01
1.38058752e-01 4.22936112e-01 -6.81107938e-01 -4.92528139e-04
5.23915827e-01 1.17823553e+00 -9.88867581e-01 -4.60083723e-01
4.62320894e-02 2.90843219e-01 -5.80296695e-01 4.47736412e-01
-8.67880225e-01 -2.07879972e-02 4.39301180e-03 8.49915922e-01
2.68543243e-01 -2.33083948e-01 -1.70487911e-01 -6.50759265e-02
6.97107017e-02 1.83953285e-01 -7.90979207e-01 1.51247537e+00
-3.08251739e-01 6.72413886e-01 -3.29180598e-01 -6.60843015e-01
9.03865099e-01 1.63031578e-01 2.56604910e-01 -5.24383903e-01
2.66355667e-02 1.10899948e-01 7.56052881e-02 -3.90547156e-01
4.81149852e-01 -2.71952331e-01 -2.44519278e-01 3.39300036e-01
7.04161644e-01 -3.13658058e-03 1.76615596e-01 4.59311485e-01
1.16922390e+00 7.70955533e-02 3.07893872e-01 -8.43961686e-02
9.14991647e-02 -4.21906486e-02 6.14037275e-01 9.71302688e-01
-1.41010091e-01 6.17236972e-01 4.10053581e-01 -4.80693847e-01
-1.34287930e+00 -1.58125997e+00 -1.26021951e-01 1.99045658e+00
-1.32783979e-01 -4.88637954e-01 -3.64549905e-01 -6.29109502e-01
2.93247014e-01 1.14977717e+00 -7.67563820e-01 -3.97451222e-01
-3.52597266e-01 -2.84250498e-01 5.76961219e-01 8.27059448e-01
9.05435458e-02 -1.29031384e+00 -6.33581877e-01 2.45886385e-01
-2.17819214e-01 -1.25361061e+00 -7.22315252e-01 7.05700457e-01
-4.67770338e-01 -8.73068392e-01 -6.46613717e-01 -9.91107464e-01
5.76601267e-01 1.83068767e-01 1.34039927e+00 -4.75686826e-02
-3.81174684e-01 5.16057909e-01 -5.64383805e-01 -6.32310152e-01
-4.28085268e-01 -2.43611019e-02 1.76880121e-01 3.16657647e-02
7.43395686e-01 -4.50450212e-01 -2.11526513e-01 3.03975910e-01
-1.00095117e+00 1.57696977e-01 3.79990458e-01 9.98385966e-01
8.73503685e-01 -7.54886687e-01 5.25242507e-01 -8.74493301e-01
1.59555823e-01 -5.40649056e-01 -6.94640279e-01 4.89164799e-01
-2.89455563e-01 3.71833980e-01 6.80221200e-01 -1.20305979e+00
-6.05816662e-01 2.65839815e-01 8.32537860e-02 -9.20972943e-01
-2.24532455e-01 3.98042426e-02 -9.56810340e-02 1.04730278e-01
9.39786017e-01 4.92537320e-01 -3.06910127e-01 -4.08289045e-01
6.46830618e-01 6.13723993e-01 9.24689233e-01 -8.08880389e-01
8.67973089e-01 1.69632748e-01 -4.55837637e-01 -6.96855843e-01
-9.39420938e-01 -4.98436153e-01 -3.93811822e-01 4.95718233e-02
9.33813035e-01 -1.15611494e+00 -5.05060554e-01 4.08562809e-01
-9.59934890e-01 -8.36337805e-01 -6.02145553e-01 2.30556577e-01
-7.78398395e-01 -1.38460577e-01 -5.12304842e-01 -4.93422061e-01
-5.71694851e-01 -8.79541457e-01 1.20025241e+00 4.26616333e-03
-3.43689859e-01 -4.47427720e-01 -1.74763307e-01 4.31173742e-02
3.02679539e-01 -1.10971861e-01 1.08976555e+00 -1.10136497e+00
-6.43975317e-01 2.16032743e-01 -6.68902025e-02 3.26492667e-01
-1.56862885e-01 6.42758384e-02 -1.08192861e+00 -2.62293100e-01
-3.60880852e-01 -7.70865619e-01 1.11084127e+00 1.76142573e-01
1.32132685e+00 -4.82364506e-01 -1.75903365e-01 8.27445865e-01
1.23080266e+00 1.27723828e-01 3.02696258e-01 2.89497734e-03
5.52988887e-01 3.17138672e-01 6.01908505e-01 4.98275757e-01
4.05983448e-01 5.71017146e-01 3.89210135e-01 3.94392490e-01
-4.72295731e-01 -5.58780611e-01 4.00333017e-01 4.59625453e-01
2.53688753e-01 -8.25363323e-02 -9.74734962e-01 7.93133974e-01
-1.58286667e+00 -1.12972975e+00 3.86898041e-01 2.25421500e+00
1.27057242e+00 1.19360827e-01 1.39537379e-01 -3.62669349e-01
5.99197686e-01 1.30393803e-02 -8.52593660e-01 -2.62852669e-01
-2.60928035e-01 3.98826718e-01 4.96636450e-01 4.57519323e-01
-1.10391462e+00 1.45186281e+00 5.21776104e+00 7.34453082e-01
-1.09152031e+00 2.45348185e-01 3.03171396e-01 -6.96794868e-01
-5.84341168e-01 -4.24422771e-02 -1.15870476e+00 5.70860684e-01
7.79063523e-01 -2.43017033e-01 7.33509481e-01 1.06034064e+00
-5.84366918e-01 1.87400430e-01 -1.52854502e+00 1.04869807e+00
4.61936206e-01 -1.33552730e+00 2.54972845e-01 -2.76324242e-01
8.02516222e-01 4.91214871e-01 2.48157308e-01 8.60191643e-01
5.66657186e-01 -7.09536374e-01 9.30621028e-01 3.71988803e-01
1.21200180e+00 -3.39482516e-01 3.34476084e-01 3.03396076e-01
-1.11486840e+00 -2.40790829e-01 -6.98329151e-01 2.77949840e-01
-1.13656253e-01 -2.27480996e-02 -1.00265741e+00 -2.81410277e-01
7.88840890e-01 4.76716846e-01 -8.03119302e-01 9.47324097e-01
-4.70276356e-01 7.95249045e-01 -3.18088084e-01 -1.96847737e-01
2.64148861e-01 5.03361821e-01 6.09931946e-01 1.13666534e+00
9.29318815e-02 3.91517460e-01 1.28801152e-01 9.25917923e-01
-3.36755782e-01 -8.65631998e-02 -5.28452873e-01 -2.62344211e-01
7.40600348e-01 1.11005914e+00 -2.43174091e-01 -7.73728311e-01
-4.52431470e-01 8.09169531e-01 4.85245317e-01 4.11020249e-01
-8.94025564e-01 -2.64505237e-01 9.24510777e-01 9.94075909e-02
7.29642987e-01 -3.53668258e-03 -5.03077097e-02 -1.24968159e+00
-1.41104460e-01 -1.06266963e+00 5.34775019e-01 -9.64263380e-01
-1.54900146e+00 6.22570336e-01 2.04930142e-01 -1.13059056e+00
-3.89107049e-01 -5.83362997e-01 -4.96049047e-01 5.08078873e-01
-1.23858440e+00 -1.49455857e+00 -3.72440666e-01 6.73673749e-01
1.05442727e+00 -4.74096775e-01 8.53800535e-01 6.23462610e-02
-1.02563478e-01 1.02327132e+00 -2.07963407e-01 2.44880915e-01
9.57415164e-01 -1.22715962e+00 7.09805012e-01 7.39474356e-01
5.31542838e-01 4.93528038e-01 7.08014369e-01 -6.61952257e-01
-1.17530310e+00 -1.31199741e+00 9.11251485e-01 -5.75865328e-01
8.18484247e-01 -8.50314260e-01 -1.09744132e+00 1.01147890e+00
1.29932448e-01 3.41513485e-01 8.01792443e-01 8.45377743e-02
-1.07365811e+00 -1.94781810e-01 -8.52942526e-01 6.53096139e-01
1.24105847e+00 -7.46522427e-01 -8.86612892e-01 3.56783003e-01
1.29603338e+00 -1.93225846e-01 -6.00681603e-01 5.25039613e-01
6.24000430e-01 -3.39955956e-01 8.96289706e-01 -9.51614559e-01
3.65727097e-01 -1.86101437e-01 -5.74947059e-01 -1.29266489e+00
-5.64158320e-01 -2.88706034e-01 -4.25124556e-01 1.04504430e+00
4.84303832e-01 -4.13970470e-01 4.58228171e-01 5.25843024e-01
-3.63983154e-01 -6.29331708e-01 -7.38930523e-01 -1.10953009e+00
5.21256179e-02 -4.23170805e-01 6.60653830e-01 8.18714619e-01
-3.24945211e-01 5.53687334e-01 -2.81257957e-01 -6.91226050e-02
7.93704689e-01 1.47309899e-01 7.17988133e-01 -1.23319817e+00
-4.45032984e-01 -4.79912877e-01 -4.92741823e-01 -7.47549355e-01
4.23068225e-01 -1.28582919e+00 1.82691187e-01 -1.22241914e+00
3.93446833e-01 -4.31524187e-01 -4.67244923e-01 1.12259269e+00
-3.62348676e-01 2.17181757e-01 4.92966741e-01 1.40253231e-01
-6.90956950e-01 5.73905528e-01 8.43796194e-01 -6.07793450e-01
-5.69112562e-02 -2.21153006e-01 -5.93153358e-01 4.75795418e-01
1.01853466e+00 -6.95742905e-01 -4.48208451e-01 -6.14689291e-01
1.20063417e-01 -4.16938186e-01 6.02247715e-01 -9.28823292e-01
1.80782646e-01 -2.33121559e-01 3.79876941e-01 -6.19073153e-01
4.57649559e-01 -5.43740153e-01 -3.97850201e-02 2.79661596e-01
-8.23309720e-01 -2.08355293e-01 2.71095812e-01 3.63630414e-01
1.26679510e-01 -1.61834925e-01 9.59320068e-01 -8.41867477e-02
-1.12941933e+00 6.87636018e-01 -1.70914009e-01 5.08987665e-01
1.03720486e+00 -1.43166929e-01 -6.24372661e-01 -2.01860651e-01
-7.67420948e-01 3.75691414e-01 4.53058541e-01 7.90370703e-01
7.27538526e-01 -1.37460482e+00 -8.71424913e-01 3.84477973e-01
8.76136899e-01 -1.67033285e-01 8.41769725e-02 5.67453682e-01
-9.30505339e-03 4.75180373e-02 -3.99756461e-01 -6.71701729e-01
-1.43345165e+00 9.74979281e-01 2.03060627e-01 5.10333478e-01
-5.66399097e-01 1.40997136e+00 6.20352745e-01 -3.89032811e-01
5.56034446e-01 -4.37397689e-01 1.66262329e-01 6.90385476e-02
7.29241312e-01 -4.40853313e-02 -2.92805344e-01 -4.94547427e-01
-4.01336640e-01 3.79865229e-01 -2.53337473e-01 -8.24500248e-02
1.30235577e+00 2.81975597e-01 8.63592848e-02 7.84414589e-01
1.16101372e+00 -2.81296611e-01 -1.50027823e+00 -8.24481487e-01
-2.97331475e-02 -2.27455735e-01 -2.78358728e-01 -1.01837182e+00
-7.33865559e-01 1.00424528e+00 5.18594086e-01 -3.61966342e-01
8.54731977e-01 6.26535714e-01 6.73930466e-01 5.82474589e-01
4.36530381e-01 -9.60076511e-01 2.99771786e-01 5.76400280e-01
1.00365698e+00 -1.27969456e+00 -3.50771755e-01 -6.91793635e-02
-1.03873396e+00 8.32427442e-01 9.73267853e-01 1.92068086e-03
2.85470426e-01 4.71310824e-01 -4.21906859e-02 8.02285522e-02
-1.39444220e+00 -5.68833768e-01 3.89669269e-01 6.80574656e-01
1.15848631e-01 4.17722970e-01 2.15173632e-01 7.68689573e-01
-3.73006195e-01 -1.58809543e-01 2.47141957e-01 5.53495944e-01
-7.50554383e-01 -7.64848948e-01 -9.05474275e-02 7.31029987e-01
-1.56101570e-01 -5.76004267e-01 -5.04345179e-01 3.46860468e-01
2.82757103e-01 6.26526475e-01 3.73628974e-01 -4.48197484e-01
8.18568617e-02 5.52013516e-01 4.77708668e-01 -1.04282975e+00
-5.57334602e-01 -2.21495956e-01 -1.26969740e-02 -4.31515872e-01
3.57419103e-02 -6.31777942e-01 -1.34911835e+00 4.90483791e-02
7.73782954e-02 -2.09137321e-01 4.12893921e-01 5.98536253e-01
2.62624681e-01 3.80803883e-01 3.79179537e-01 -3.99562150e-01
-8.79124939e-01 -9.94846344e-01 -2.15490967e-01 5.56562960e-01
5.93707144e-01 -7.85129070e-01 -3.61026883e-01 2.52491951e-01] | [10.093873977661133, 2.0935192108154297] |
0d501fe3-4704-42d6-8276-cc5dbfa72a9c | graph-sampling-for-node-embedding | 2210.10520 | null | https://arxiv.org/abs/2210.10520v1 | https://arxiv.org/pdf/2210.10520v1.pdf | Graph sampling for node embedding | Node embedding is a central topic in graph representation learning. Computational efficiency and scalability can be challenging to any method that requires full-graph operations. We propose sampling approaches to node embedding, with or without explicit modelling of the feature vector, which aim to extract useful information from both the eigenvectors related to the graph Laplacien and the given values associated with the graph. | ['Li-Chun Zhang'] | 2022-10-19 | null | null | null | null | ['graph-sampling'] | ['graphs'] | [ 1.92256600e-01 4.80084121e-01 -2.87922114e-01 1.44082820e-02
-2.26102263e-01 -3.55679274e-01 6.55862391e-01 5.39069176e-01
-3.36853117e-01 5.50864160e-01 4.76820804e-02 -1.90312132e-01
-4.19091046e-01 -8.96378398e-01 -3.15573186e-01 -6.66663527e-01
-7.06498444e-01 5.03278375e-01 -3.53518873e-02 -2.19097465e-01
-3.87145169e-02 7.08911777e-01 -1.08229363e+00 -1.76989809e-01
3.60324383e-01 7.58720279e-01 -1.68559790e-01 1.06143606e+00
-2.93275833e-01 4.30773139e-01 -6.08067274e-01 -3.56287986e-01
8.92786607e-02 -2.75299758e-01 -5.55458903e-01 2.86971293e-02
3.88860226e-01 8.33937973e-02 -9.32880104e-01 1.27111340e+00
3.19067776e-01 1.00817032e-01 1.06410170e+00 -1.58780026e+00
-2.53768981e-01 4.36441123e-01 -4.44981754e-01 2.61745960e-01
5.70349216e-01 -3.95866305e-01 1.39447272e+00 -8.39882433e-01
7.89838135e-01 1.19848943e+00 8.07975292e-01 3.74529928e-01
-1.70990682e+00 -1.49514864e-03 -2.17560798e-01 2.53590405e-01
-1.47626686e+00 -2.22653359e-01 1.25704479e+00 -4.45497453e-01
7.96232283e-01 4.79644477e-01 9.50187385e-01 7.66372800e-01
2.16998070e-01 6.43969178e-01 6.79542720e-01 -5.45260131e-01
1.50797650e-01 1.80035353e-01 3.25216502e-01 1.01244414e+00
6.27372146e-01 -9.67628509e-02 -1.95686311e-01 -6.02122664e-01
5.56867659e-01 1.68172374e-01 -2.81189442e-01 -9.87493575e-01
-9.17991459e-01 1.31721663e+00 4.88998562e-01 2.97886521e-01
-4.60706204e-01 8.78201783e-01 4.92873102e-01 6.06781006e-01
3.73879313e-01 3.01311374e-01 -2.83575267e-01 1.65043056e-01
-6.69408619e-01 -2.17546672e-02 1.22476435e+00 9.84974027e-01
1.13229954e+00 1.88080475e-01 1.68368191e-01 2.97732234e-01
4.92532253e-01 5.11844337e-01 1.06866606e-01 -3.61223310e-01
3.27156931e-01 9.84293163e-01 -4.24345165e-01 -1.55168545e+00
-5.15119374e-01 -4.86930422e-02 -8.74444067e-01 1.47378361e-02
2.99419254e-01 -8.41375515e-02 -6.18616283e-01 1.26825070e+00
3.67625326e-01 2.25337833e-01 -5.86904064e-02 3.96331400e-01
1.05721545e+00 6.10644877e-01 -1.32411122e-01 -2.03447551e-01
1.10135317e+00 -5.21967828e-01 -7.11904943e-01 -1.88137129e-01
8.25444996e-01 -3.32565218e-01 6.33438706e-01 -4.47804332e-02
-6.63630188e-01 -1.19204029e-01 -1.27522624e+00 4.86845011e-03
-7.60212600e-01 2.00909637e-02 8.43133271e-01 7.23050475e-01
-1.38096583e+00 7.11230814e-01 -7.87170708e-01 -3.56770158e-01
5.00960164e-02 7.23303199e-01 -7.76020229e-01 2.64060982e-02
-1.25527442e+00 8.28976572e-01 3.96292686e-01 1.29357696e-01
-4.19746965e-01 -4.27696317e-01 -1.36960661e+00 2.93005705e-01
1.86797008e-01 -3.72600138e-01 4.46081936e-01 -6.84689939e-01
-1.05459785e+00 4.78913397e-01 5.95644675e-03 -3.86259764e-01
1.24855556e-01 1.24815449e-01 -2.01466128e-01 3.56288701e-01
-2.43133858e-01 2.39885792e-01 1.26160121e+00 -9.45315361e-01
2.20912978e-01 -2.99479008e-01 -8.17675516e-02 7.12084249e-02
-6.42705858e-01 -2.95730442e-01 -1.91124544e-01 -5.46966136e-01
3.21960926e-01 -7.58220255e-01 -5.46747506e-01 2.03472003e-01
-4.50291157e-01 -3.51547629e-01 1.02768731e+00 -7.30322063e-01
1.31083679e+00 -2.11608887e+00 5.82890034e-01 7.81157613e-01
6.75748110e-01 2.76971385e-02 -3.81080359e-01 1.08859110e+00
-2.35858873e-01 4.81091365e-02 -1.32788986e-01 7.53997713e-02
2.13174671e-01 3.39656711e-01 -1.62411168e-01 9.48757708e-01
2.08676264e-01 9.92925465e-01 -1.12990379e+00 -5.57546437e-01
3.55994344e-01 9.35847282e-01 -4.12383825e-01 2.09806189e-01
1.58813417e-01 -1.99738950e-01 -5.58470964e-01 2.59955108e-01
3.65237534e-01 -3.74080867e-01 5.59668183e-01 -3.51941228e-01
3.77838671e-01 1.86528444e-01 -1.50439167e+00 1.34182799e+00
-2.30598882e-01 7.76091516e-01 2.12711677e-01 -1.38067472e+00
9.11279202e-01 3.14674437e-01 6.96594179e-01 -1.35690987e-01
-2.51940452e-03 -4.56706434e-02 -1.26233906e-01 -2.56385893e-01
2.94044405e-01 1.98519051e-01 -1.27106637e-01 4.94684696e-01
2.45205984e-01 -3.15320790e-01 2.56400347e-01 6.65285766e-01
1.40913582e+00 -3.64837348e-01 7.61525750e-01 -5.47166049e-01
5.58200419e-01 -4.22861993e-01 2.58152280e-02 4.63380814e-01
-1.35621959e-02 1.14053316e-01 9.17835951e-01 -7.14820266e-01
-8.17045629e-01 -8.23343515e-01 1.28589809e-01 8.07324708e-01
-1.07064083e-01 -9.51574385e-01 -4.28987503e-01 -8.60880077e-01
3.02185148e-01 1.49611607e-01 -8.62409830e-01 -4.23202455e-01
-4.89036262e-01 -3.57298195e-01 1.74206078e-01 3.98973346e-01
-1.69542074e-01 -6.58490658e-01 -3.72387201e-01 2.18820006e-01
4.30519760e-01 -7.87993848e-01 -5.81728935e-01 4.10925090e-01
-9.96656597e-01 -1.06019664e+00 -5.25956631e-01 -8.52682352e-01
1.03745770e+00 4.74588759e-02 1.03669763e+00 4.70488876e-01
-6.36485875e-01 8.70028198e-01 -1.91959292e-01 1.82732433e-01
-2.86532909e-01 2.77881265e-01 -2.35846236e-01 6.62397593e-02
1.92106351e-01 -6.78451300e-01 -2.77781010e-01 -1.56309664e-01
-6.81687415e-01 -2.87363589e-01 5.70501804e-01 1.00606346e+00
3.79729807e-01 -1.12457819e-01 7.20417574e-02 -1.02152276e+00
8.90076697e-01 -4.86368746e-01 -4.79431987e-01 2.85795003e-01
-6.90034449e-01 4.59910929e-01 5.29317617e-01 -3.43885213e-01
1.41990125e-01 3.54832530e-01 2.74847478e-01 -6.29833817e-01
3.42887223e-01 7.80215085e-01 -5.16222864e-02 -5.04813015e-01
6.55699074e-01 2.39019349e-01 1.72406211e-01 -2.78293580e-01
6.62978709e-01 2.33179763e-01 -4.65021022e-02 -1.86108351e-01
1.15442908e+00 2.76386172e-01 6.16144121e-01 -1.54606795e+00
-3.20428818e-01 -7.42571056e-01 -8.83408308e-01 -1.80183202e-01
4.47116017e-01 -3.85162890e-01 -6.09296322e-01 -3.27211678e-01
-1.18714869e+00 1.23559162e-01 -5.76859117e-01 4.66057956e-01
-4.31519032e-01 6.96993291e-01 -5.38080454e-01 -8.75549912e-01
-4.52871382e-01 -6.39734685e-01 1.09016049e+00 -1.23813242e-01
-2.44656891e-01 -1.59775960e+00 3.94816250e-01 -5.88753700e-01
4.10189033e-01 4.65995729e-01 1.03633034e+00 -7.99945831e-01
-3.68431181e-01 -1.00417590e+00 -3.05747896e-01 2.20081806e-02
1.02000810e-01 8.24258849e-02 -7.03070641e-01 -8.43235373e-01
-2.49783337e-01 -1.13578038e-02 7.67910779e-01 2.09865317e-01
6.95648789e-01 -5.61493635e-01 -6.24411404e-01 6.87143087e-01
1.47799337e+00 -4.76322889e-01 1.83698446e-01 -1.67632118e-01
9.38676775e-01 5.28305650e-01 3.27262394e-02 3.63488853e-01
5.07481433e-02 5.75079203e-01 2.43186027e-01 1.69705693e-02
-1.38632700e-01 -4.07253861e-01 2.56653339e-01 1.03018534e+00
-1.34926122e-02 1.44947497e-02 -9.03994262e-01 4.88289505e-01
-1.77688146e+00 -7.91596413e-01 -2.05611840e-01 1.99190295e+00
2.84693897e-01 2.71515977e-02 1.19096719e-01 4.93989676e-01
6.04921460e-01 7.64392734e-01 -3.45422000e-01 -5.92980564e-01
2.90883303e-01 2.43471071e-01 6.31389081e-01 9.17080224e-01
-1.07231283e+00 6.93886757e-01 8.26875782e+00 4.30316240e-01
-7.92654455e-01 -2.16054246e-01 -9.18009207e-02 3.97455871e-01
-6.31113589e-01 2.62317002e-01 -4.18341517e-01 -3.97540425e-04
1.00209153e+00 -5.57891250e-01 7.18261838e-01 1.00641060e+00
-3.40958118e-01 3.77868623e-01 -1.25520492e+00 9.76866007e-01
3.90988737e-01 -1.02775311e+00 7.11556822e-02 3.23820263e-01
3.40852946e-01 -2.11520165e-01 -1.29679054e-01 1.55061752e-01
1.94037482e-01 -1.19921684e+00 -8.63078311e-02 5.04272461e-01
8.40870321e-01 -7.64777064e-01 5.13420939e-01 -4.46328931e-02
-1.86291838e+00 2.57829458e-01 -8.10447037e-01 -4.02596891e-02
-1.72975540e-01 5.50149262e-01 -1.19961393e+00 4.04032052e-01
1.71186447e-01 1.17743003e+00 -7.34643698e-01 7.86556125e-01
-4.57958490e-01 5.99189937e-01 -4.76389408e-01 -3.93929929e-01
1.01946652e-01 -6.03710592e-01 8.51913691e-01 1.19179654e+00
1.83273464e-01 -2.08719879e-01 5.15405595e-01 5.34764230e-01
-2.41489466e-02 3.94087374e-01 -1.24601030e+00 -6.18465781e-01
2.80855030e-01 1.52648997e+00 -9.28261399e-01 -1.98074922e-01
-4.37976927e-01 9.59625661e-01 5.57505429e-01 4.94216621e-01
-2.01186746e-01 -8.40592325e-01 6.76523149e-01 4.78399768e-02
4.40576911e-01 -4.73417133e-01 2.56833524e-01 -1.09616303e+00
-5.10731004e-02 -4.29132968e-01 6.20867431e-01 -7.28066266e-02
-1.20418000e+00 4.93345827e-01 7.84355402e-02 -9.29218888e-01
-5.52762985e-01 -8.87369752e-01 -7.36661315e-01 8.32530200e-01
-1.29772556e+00 -1.05216432e+00 -4.71480004e-02 8.28588665e-01
-1.35996252e-01 -1.51849315e-01 9.91664290e-01 5.40605821e-02
-3.31656724e-01 5.87656200e-01 2.20664486e-01 2.54910856e-01
2.99573913e-02 -1.49815595e+00 6.13370299e-01 5.30268788e-01
6.50761604e-01 5.23517847e-01 8.85105968e-01 -6.34414494e-01
-2.15285730e+00 -7.98367143e-01 1.03508770e+00 -4.46429491e-01
9.42861378e-01 -7.08302498e-01 -7.26433098e-01 7.48907328e-01
-2.18539417e-01 5.47019780e-01 5.04093885e-01 2.99417198e-01
-2.99700499e-01 4.36405477e-04 -9.49697316e-01 3.93768519e-01
7.38385260e-01 -9.05270219e-01 -2.92251140e-01 3.99260789e-01
5.63429177e-01 -3.04655377e-02 -9.63255405e-01 -7.87252709e-02
3.91949326e-01 -1.93985060e-01 1.31317234e+00 -9.38695550e-01
-3.51124257e-01 -4.18801643e-02 -1.47695616e-01 -1.29089999e+00
-3.47685277e-01 -8.48599672e-01 -5.60757935e-01 7.62310147e-01
3.13439548e-01 -7.49921083e-01 1.01536059e+00 3.68234158e-01
5.49184084e-01 -5.38488567e-01 -1.04457378e+00 -7.61383295e-01
-2.53306985e-01 -2.27372944e-01 3.86097252e-01 9.23289955e-01
4.34153885e-01 5.76700389e-01 -4.94723767e-01 1.06004380e-01
7.68222809e-01 6.56651929e-02 6.98783040e-01 -1.65616131e+00
-1.31954104e-01 -3.75223458e-01 -1.33324766e+00 -6.89884365e-01
3.77172887e-01 -1.37145030e+00 -2.91718304e-01 -1.67537808e+00
-1.71989202e-03 -4.65195775e-02 -1.80636838e-01 3.26549679e-01
-5.05638234e-02 5.81002347e-02 -1.23385312e-02 -7.19207600e-02
-3.40282172e-01 6.81307793e-01 9.83291447e-01 -2.77701408e-01
-1.43363848e-02 -1.64227277e-01 -2.02703297e-01 6.57246292e-01
4.36616808e-01 -6.58616781e-01 -6.27755463e-01 1.13583885e-01
4.93528932e-01 8.70793089e-02 2.08891943e-01 -7.38620937e-01
2.02743113e-01 1.09677510e-02 1.67324215e-01 -2.33118415e-01
4.10449028e-01 -9.73609924e-01 1.81470439e-01 8.45544696e-01
-4.07930523e-01 3.78847063e-01 -2.22366825e-01 9.93665099e-01
-1.45210832e-01 -5.11013627e-01 3.70228350e-01 1.31911457e-01
-4.92374003e-01 5.36902189e-01 -1.16234459e-01 5.37087098e-02
1.00511146e+00 -1.15730785e-01 1.62259638e-01 -6.97390795e-01
-7.60318935e-01 -1.67805124e-02 4.52434480e-01 6.15147129e-02
9.03538287e-01 -1.66367805e+00 -6.92808032e-01 3.92030090e-01
2.91013229e-03 -4.67385203e-01 -3.79958153e-01 6.42116487e-01
-5.58451116e-01 2.47990891e-01 1.21140353e-01 -4.66318637e-01
-1.29744160e+00 9.14086401e-01 2.44377285e-01 -3.41844499e-01
-9.33395982e-01 6.06423736e-01 -2.82782316e-01 -4.86258984e-01
1.84821844e-01 -2.69629639e-02 -5.04319787e-01 4.17346269e-01
3.57134014e-01 5.57734787e-01 -1.30727440e-01 -8.02923381e-01
-4.68459010e-01 5.54824531e-01 1.02639265e-01 5.35305813e-02
1.14272666e+00 3.58241111e-01 -3.56142849e-01 5.70516229e-01
1.59032106e+00 -9.24765840e-02 -7.35277772e-01 -2.80525416e-01
2.74290383e-01 -6.07477665e-01 1.97629273e-01 5.93290217e-02
-1.16687047e+00 9.74836171e-01 4.39894140e-01 7.10349023e-01
5.38537383e-01 3.90560292e-02 3.43679994e-01 7.30713904e-01
2.19361901e-01 -7.27792263e-01 -5.66330664e-02 2.24744663e-01
9.03677881e-01 -1.02754676e+00 4.22981024e-01 -6.17014527e-01
-1.52119353e-01 1.40829122e+00 9.50626284e-02 -6.54380023e-01
1.26593447e+00 7.24427924e-02 -1.61271900e-01 -6.30117595e-01
-7.20118582e-01 -2.02765942e-01 7.86963463e-01 7.44930863e-01
5.41606367e-01 1.96540594e-01 -4.10149455e-01 -8.15822408e-02
-2.25088466e-02 -7.52245545e-01 3.73771816e-01 8.96891236e-01
-3.01986337e-01 -1.13013113e+00 -6.80832341e-02 7.75098801e-01
1.03235319e-01 -3.90453860e-02 -6.53905213e-01 9.30430233e-01
-7.76599288e-01 5.71542680e-01 -1.57523647e-01 -3.71602625e-01
2.48245940e-01 1.85422286e-01 3.09963167e-01 -6.91136122e-01
-1.04425356e-01 -1.65089339e-01 7.24208215e-03 -6.92545295e-01
-1.20361663e-01 -7.90490389e-01 -1.05664229e+00 -2.60879457e-01
-4.48022783e-01 5.02436876e-01 7.87018120e-01 6.09278858e-01
1.28206462e-01 2.08524555e-01 7.66642809e-01 -9.93687987e-01
-6.05817020e-01 -8.05182099e-01 -9.26568329e-01 3.45179945e-01
4.73368108e-01 -3.97106022e-01 -8.03429484e-01 -3.33118528e-01] | [7.202200889587402, 6.027388095855713] |
a75660b0-0ae4-439b-9c80-b0bbb4663de2 | can-forward-gradient-match-backpropagation | 2306.06968 | null | https://arxiv.org/abs/2306.06968v1 | https://arxiv.org/pdf/2306.06968v1.pdf | Can Forward Gradient Match Backpropagation? | Forward Gradients - the idea of using directional derivatives in forward differentiation mode - have recently been shown to be utilizable for neural network training while avoiding problems generally associated with backpropagation gradient computation, such as locking and memorization requirements. The cost is the requirement to guess the step direction, which is hard in high dimensions. While current solutions rely on weighted averages over isotropic guess vector distributions, we propose to strongly bias our gradient guesses in directions that are much more promising, such as feedback obtained from small, local auxiliary networks. For a standard computer vision neural network, we conduct a rigorous study systematically covering a variety of combinations of gradient targets and gradient guesses, including those previously presented in the literature. We find that using gradients obtained from a local loss as a candidate direction drastically improves on random noise in Forward Gradient methods. | ['Edouard Oyallon', 'Michael Eickenberg', 'Eugene Belilovsky', 'Stéphane Rivaud', 'Louis Fournier'] | 2023-06-12 | null | null | null | null | ['memorization'] | ['natural-language-processing'] | [ 2.31747925e-02 -3.18874627e-01 -1.88231871e-01 -4.01968002e-01
-4.31480587e-01 -3.32769394e-01 6.50692165e-01 -2.28142574e-01
-1.07741606e+00 1.06666756e+00 2.01636314e-01 -4.07703429e-01
-1.73336118e-01 -6.00106418e-01 -6.69325650e-01 -1.11815512e+00
-1.57166868e-01 2.66119003e-01 1.72521442e-01 -3.95442605e-01
9.06298280e-01 7.04284132e-01 -1.13470078e+00 -1.58063188e-01
4.50533599e-01 1.17392528e+00 1.56190276e-01 5.66540539e-01
-6.59628883e-02 6.23377740e-01 -3.52571070e-01 -2.02678189e-01
5.82306266e-01 -6.19980216e-01 -7.49296784e-01 -4.22450393e-01
7.63748884e-01 -3.25723320e-01 -1.95495218e-01 1.14692223e+00
8.19005728e-01 3.57272714e-01 6.83840454e-01 -5.38626671e-01
-6.93963826e-01 5.10538757e-01 -3.51535171e-01 6.76512122e-01
-2.49391407e-01 3.90691847e-01 8.59787464e-01 -1.41201627e+00
8.34091187e-01 6.93200231e-01 1.21227050e+00 9.80981171e-01
-1.34103990e+00 -3.86448622e-01 2.35701978e-01 9.43094119e-02
-1.09175527e+00 -7.21740603e-01 1.09304416e+00 -2.11572260e-01
9.65381324e-01 1.65544912e-01 6.73831522e-01 7.55180299e-01
1.21163405e-01 6.68806016e-01 1.27692509e+00 -5.68197727e-01
3.85704160e-01 4.52796787e-01 2.51687765e-01 8.97310555e-01
2.54863173e-01 2.18515381e-01 -6.23515666e-01 -3.40433456e-02
1.01175869e+00 -3.42147827e-01 -7.23529100e-01 -5.45257866e-01
-1.08917344e+00 9.39526916e-01 8.15845907e-01 4.57459509e-01
-4.53168571e-01 2.41934210e-01 2.55011711e-02 4.91107821e-01
5.92299044e-01 8.31739008e-01 -4.54899222e-01 -5.48213050e-02
-1.40011668e+00 3.08894962e-01 7.83277750e-01 3.43671679e-01
1.00713170e+00 4.51787770e-01 -5.66357486e-02 8.70179534e-01
3.10561031e-01 3.52423012e-01 7.43168592e-01 -9.12134171e-01
4.71491277e-01 8.97772461e-02 1.32176995e-01 -1.08082068e+00
-5.95393538e-01 -6.91305101e-01 -7.93409884e-01 7.79904425e-01
8.53291988e-01 -3.97748232e-01 -1.07949781e+00 1.78990662e+00
2.51194328e-01 -2.56005794e-01 -2.34309733e-01 1.30340528e+00
3.74872833e-01 3.67651969e-01 -3.10930759e-01 -2.58299708e-01
2.43236333e-01 -1.03821838e+00 -3.17779392e-01 -4.76691425e-01
8.37854028e-01 -5.87229550e-01 1.11787891e+00 2.85737693e-01
-1.35935950e+00 -1.89460933e-01 -1.17106581e+00 -1.38690248e-01
-5.11030972e-01 -9.27542523e-03 8.27021718e-01 4.69662935e-01
-1.62433064e+00 1.29704762e+00 -6.84551239e-01 4.27304134e-02
3.79523873e-01 4.54466105e-01 -3.77180994e-01 2.61312187e-01
-1.02572775e+00 1.31372261e+00 1.24660261e-01 6.52260363e-01
-3.99759054e-01 -7.89008975e-01 -5.39419949e-01 -2.49830261e-01
-1.35030046e-01 -6.52422845e-01 1.00274146e+00 -1.09149766e+00
-1.84031522e+00 7.25124478e-01 -4.01098907e-01 -5.57894349e-01
7.62188256e-01 -9.76034477e-02 2.24665746e-01 1.17912024e-01
-3.32232893e-01 8.99052501e-01 1.04999626e+00 -9.67593312e-01
-3.41760337e-01 -3.87017816e-01 -1.40893504e-01 4.16387558e-01
-4.66366857e-01 -3.74949098e-01 2.96286270e-02 -3.62305880e-01
5.82595348e-01 -9.30964231e-01 -6.46566331e-01 3.81511748e-01
-2.08190873e-01 9.66578424e-02 5.58345616e-01 -5.34558296e-01
1.25520515e+00 -1.64817011e+00 2.64580011e-01 4.89113003e-01
1.19069636e-01 3.43559772e-01 7.32952729e-02 3.13475057e-02
-6.41463324e-02 1.05526805e-01 -2.25014612e-01 -3.12678814e-01
-1.67924494e-01 -5.36446311e-02 -3.31076473e-01 7.11395919e-01
7.08779991e-02 9.15990114e-01 -7.54920542e-01 -2.00387463e-01
3.62613313e-02 5.95707297e-01 -7.26920843e-01 -2.40379080e-01
1.67878971e-01 1.60413772e-01 -1.55084759e-01 2.03094572e-01
6.09817922e-01 -1.33415654e-01 -2.96047926e-01 -1.08991377e-01
-2.75370508e-01 6.68269753e-01 -1.19572699e+00 1.48540521e+00
-5.07243276e-01 8.41085732e-01 3.78097713e-01 -1.10531652e+00
8.28249395e-01 2.29996182e-02 7.85714984e-02 -4.74564880e-01
5.78515418e-02 6.21207058e-01 1.04690030e-01 -1.44957229e-01
3.72137904e-01 -5.19570649e-01 6.71044946e-01 4.24443990e-01
2.54063964e-01 -2.30649129e-01 3.20290238e-01 -8.95596370e-02
7.76257217e-01 1.42659694e-01 -1.43890098e-01 -5.07945955e-01
2.94640243e-01 -3.76074873e-02 2.24143520e-01 9.27653253e-01
-1.78443745e-01 8.95921111e-01 1.96086302e-01 -5.10820985e-01
-8.53861868e-01 -8.02256286e-01 -2.85279304e-01 8.60495746e-01
-2.38948479e-01 1.83147013e-01 -6.04420304e-01 -6.20346606e-01
-1.54857904e-01 3.16986352e-01 -6.94815516e-01 -1.17652543e-01
-8.25265408e-01 -1.03208447e+00 4.04937267e-01 5.78178823e-01
4.96969134e-01 -1.00893402e+00 -5.99827886e-01 3.43589932e-01
4.22788262e-01 -4.46061164e-01 -4.35301900e-01 7.79525101e-01
-1.40897310e+00 -5.68846166e-01 -1.41711235e+00 -7.64257789e-01
9.37058151e-01 -6.37250617e-02 1.02524078e+00 1.25159696e-01
-9.16067809e-02 3.01129609e-01 3.28878552e-01 5.63717708e-02
1.19448863e-01 1.41227305e-01 1.78432226e-01 -1.93157747e-01
1.78482547e-01 -7.37385333e-01 -9.34534609e-01 6.83704019e-02
-4.41068709e-01 -2.58143485e-01 5.39610326e-01 1.32970929e+00
3.77583683e-01 -3.26465815e-01 4.09159303e-01 -9.96874630e-01
9.93040323e-01 -2.53035218e-01 -4.59937990e-01 -1.34348914e-01
-1.10170853e+00 4.78513211e-01 6.96907580e-01 -5.59205830e-01
-9.68557954e-01 -1.14274457e-01 -4.21966046e-01 -1.80777550e-01
3.73241007e-02 5.78418374e-01 6.21229112e-01 -8.41948986e-01
1.11033475e+00 3.49007547e-01 1.21513069e-01 -4.38882440e-01
3.33184928e-01 1.14064984e-01 1.65548354e-01 -4.84554321e-01
6.23303413e-01 3.16127896e-01 2.25171775e-01 -7.72226214e-01
-5.71006477e-01 -3.09730679e-01 -4.80525613e-01 -1.89752355e-01
2.64145643e-01 -4.54697996e-01 -2.96808779e-01 7.08198786e-01
-9.46403921e-01 -6.11105323e-01 -4.57495153e-01 6.27540052e-01
-4.30763185e-01 -1.71260294e-02 -1.04124105e+00 -6.49632931e-01
-3.39366704e-01 -1.06683969e+00 3.13520730e-01 3.68993849e-01
-1.73243761e-01 -1.60425305e+00 1.01070240e-01 -4.35889184e-01
1.23047006e+00 -2.90759653e-01 6.16536617e-01 -3.74906540e-01
-4.28410023e-01 -1.41119242e-01 -1.81048680e-02 4.61292893e-01
-1.60757732e-02 -1.22465469e-01 -9.18057144e-01 -3.02265197e-01
2.37217039e-01 -4.24250662e-01 1.35180032e+00 5.83015382e-01
7.96937406e-01 -4.59557056e-01 -2.63818324e-01 8.78523827e-01
1.42766166e+00 -2.05758512e-01 3.94823521e-01 3.80406082e-01
6.12440169e-01 5.66956401e-01 -7.86134452e-02 4.77310549e-03
-1.56492054e-01 2.75664151e-01 2.27598157e-02 -1.92579970e-01
-3.33195686e-01 -2.10540771e-01 7.79431909e-02 8.59015822e-01
-3.18448335e-01 2.34791934e-01 -7.05261588e-01 5.66857636e-01
-1.44374514e+00 -1.02793062e+00 2.73925304e-01 2.06443644e+00
1.21007872e+00 5.08925319e-01 -1.53798699e-01 3.93275321e-01
3.01936120e-01 3.33153933e-01 -6.12835526e-01 -2.92573690e-01
-2.17872486e-01 5.32407045e-01 7.36082137e-01 9.85967457e-01
-8.68262827e-01 7.22954571e-01 7.67944622e+00 5.41300654e-01
-1.84954536e+00 -1.37002677e-01 7.16523349e-01 -3.98732007e-01
-3.43697250e-01 -1.78708464e-01 -1.10506666e+00 4.13013518e-01
7.84752131e-01 3.44349951e-01 5.14247298e-01 8.81741345e-01
5.29091507e-02 -2.54839301e-01 -1.01857340e+00 9.24563527e-01
-3.12981755e-02 -1.69915736e+00 5.25561348e-02 -2.34084621e-01
1.01639926e+00 3.76377106e-01 5.09315133e-01 3.34383175e-02
-1.18047856e-02 -9.89238560e-01 7.01767147e-01 5.19011259e-01
3.86754990e-01 -4.15515333e-01 3.81639957e-01 2.81155884e-01
-4.65958983e-01 -9.68506485e-02 -4.85103637e-01 -2.94662505e-01
7.18983309e-03 9.39205110e-01 -6.89913332e-01 -3.86997283e-01
6.31387115e-01 5.79664171e-01 -4.70297694e-01 1.29520261e+00
-2.54141837e-01 8.40762317e-01 -8.65325928e-01 -5.13248026e-01
5.60580790e-01 -4.38285649e-01 4.81903642e-01 1.34246552e+00
2.90498048e-01 -3.12605828e-01 -4.71929342e-01 8.97283733e-01
9.16639119e-02 5.59427552e-02 -7.02946544e-01 7.62969255e-02
2.61875719e-01 1.13600361e+00 -7.48321533e-01 -4.70774733e-02
-4.09502834e-02 9.56449211e-01 6.94268346e-01 8.73349488e-01
-7.16806948e-02 -7.50744522e-01 4.32137877e-01 1.44218847e-01
4.63433146e-01 -5.50359666e-01 -7.03706563e-01 -1.04390061e+00
4.14252311e-01 -4.22367632e-01 -1.39791712e-01 -6.00749969e-01
-9.75892782e-01 5.54364681e-01 -1.72688335e-01 -7.55352676e-01
-6.05928481e-01 -1.05183017e+00 -7.18199492e-01 1.23102844e+00
-1.62095857e+00 -5.67269206e-01 2.43115023e-01 5.84223866e-01
2.77863890e-01 6.28591627e-02 5.29620945e-01 3.08122396e-01
-1.48326546e-01 6.61111712e-01 2.04558030e-01 -4.80399001e-03
3.62454027e-01 -1.26270294e+00 1.78790018e-01 8.52705896e-01
3.09463799e-01 1.15062940e+00 9.52020824e-01 -1.55793652e-01
-1.32493854e+00 -4.04665530e-01 9.15534198e-01 -1.39750734e-01
6.24638855e-01 -1.16070881e-01 -1.05100799e+00 4.97483402e-01
-1.61346719e-01 4.84392792e-01 -2.43156757e-02 1.74120277e-01
-2.35540524e-01 -6.85702786e-02 -1.04575634e+00 8.80612195e-01
9.99552786e-01 -5.74929357e-01 -2.63043970e-01 1.21932708e-01
-3.63098904e-02 -4.47232068e-01 -3.12010467e-01 1.64024800e-01
4.50510651e-01 -1.12376535e+00 1.01236689e+00 -5.06940842e-01
2.22154424e-01 -1.29601657e-01 7.68542439e-02 -1.78550065e+00
-2.21692935e-01 -6.84665561e-01 3.20227668e-02 6.93676710e-01
8.67414653e-01 -9.03155327e-01 1.13301265e+00 7.41193354e-01
-5.63082360e-02 -1.16088307e+00 -1.17977250e+00 -6.04083419e-01
5.07898152e-01 -1.80634454e-01 -1.57866269e-01 6.87149584e-01
9.58132520e-02 4.67339128e-01 -2.56392866e-01 -3.12364012e-01
5.73015034e-01 -2.53597319e-01 1.87014192e-01 -9.22010601e-01
-2.73686826e-01 -1.10535228e+00 -2.72450417e-01 -1.70431316e+00
-8.18890333e-02 -9.10993934e-01 3.72101307e-01 -1.45069826e+00
-5.10409176e-01 -8.03705692e-01 -4.35085028e-01 1.88554794e-01
-1.67783201e-01 4.76629972e-01 -2.75425017e-02 1.64752707e-01
-8.60359371e-02 4.90923792e-01 1.29445481e+00 1.82452369e-02
-2.96560645e-01 1.58806831e-01 -5.13048589e-01 9.30568635e-01
8.07768285e-01 -4.97433931e-01 -3.63117367e-01 -5.71917355e-01
4.31331486e-01 -2.56247133e-01 1.94559082e-01 -1.15592825e+00
5.36965549e-01 2.71394532e-02 7.01940954e-01 -1.16848245e-01
6.24247134e-01 -5.42699575e-01 -4.88793284e-01 4.49857414e-01
-7.16442704e-01 1.04358733e-01 3.35669592e-02 2.37990692e-01
-1.64114088e-01 -8.63601208e-01 7.67645597e-01 -3.36367458e-01
-6.21894240e-01 2.29544535e-01 -3.60242009e-01 1.19189456e-01
2.62328118e-01 -3.06065410e-01 -3.59788202e-02 -5.09082556e-01
-6.42994881e-01 -1.91909745e-01 2.72506416e-01 -1.68012604e-01
7.21386194e-01 -1.24829364e+00 -4.35589373e-01 3.63683164e-01
-5.37980437e-01 -2.59174883e-01 -3.33492547e-01 1.10203028e+00
-6.19119644e-01 3.36340755e-01 -1.09102674e-01 -3.16813469e-01
-8.02513838e-01 3.04106176e-01 8.83315384e-01 -1.40119791e-01
-6.18612409e-01 1.43622804e+00 -3.64120573e-01 -4.84857857e-01
4.28077936e-01 -5.01621425e-01 2.84875985e-02 -9.94037092e-03
5.74786365e-01 3.43030959e-01 3.24537903e-01 -2.35495463e-01
-1.81985795e-01 8.07640195e-01 -4.05180492e-02 -5.36988437e-01
1.23369873e+00 -6.53172582e-02 1.42386823e-03 6.00809991e-01
1.38152456e+00 -4.72147837e-02 -1.59429002e+00 -3.06248009e-01
2.15083018e-01 -2.45640889e-01 5.33434927e-01 -7.14275360e-01
-1.22836888e+00 1.23471665e+00 7.19180226e-01 2.86749993e-02
7.14266479e-01 -4.64741826e-01 5.21202564e-01 8.76062095e-01
3.65184754e-01 -1.19976330e+00 -1.64858580e-01 8.00496876e-01
8.70963156e-01 -1.36401927e+00 8.32615942e-02 3.34090889e-01
-1.65919408e-01 1.33546937e+00 4.37246144e-01 -5.02933562e-01
1.01966345e+00 2.30826125e-01 4.29185808e-01 -4.04761881e-02
-5.94952762e-01 -8.73107538e-02 2.47894660e-01 5.03005445e-01
5.24217188e-01 -4.32655275e-01 -5.41732669e-01 -1.94813773e-01
-8.01065937e-02 1.42615065e-01 2.90404499e-01 9.58716393e-01
-5.44785023e-01 -8.51653516e-01 -8.88772458e-02 4.73060250e-01
-5.46046197e-01 -5.12373984e-01 -8.88235122e-02 6.57293797e-01
-2.63234645e-01 5.26060283e-01 -1.15013933e-02 -8.04942548e-02
-3.42791192e-02 2.39896446e-01 8.50749373e-01 -1.78781375e-01
-6.31368816e-01 -2.79027104e-01 -8.22232664e-03 -4.59226608e-01
-4.50214267e-01 -4.34845746e-01 -1.21800923e+00 -1.39249891e-01
-4.97316152e-01 2.05674127e-01 9.12572682e-01 9.84347284e-01
1.05374493e-01 7.31672673e-03 3.20600718e-01 -1.40816712e+00
-9.91536617e-01 -1.09690893e+00 -5.39342999e-01 2.41299763e-01
6.96890414e-01 -6.21114254e-01 -8.92843187e-01 -2.40279645e-01] | [7.996252059936523, 3.505533218383789] |
706385c7-b524-4eb9-ad28-a831eb216e4b | computing-forward-reachable-sets-for | 2209.07780 | null | https://arxiv.org/abs/2209.07780v3 | https://arxiv.org/pdf/2209.07780v3.pdf | Computing Forward Reachable Sets for Nonlinear Adaptive Multirotor Controllers | In multirotor systems, guaranteeing safety while considering unknown disturbances is essential for robust trajectory planning. The Forward reachable set (FRS), the set of feasible states subject to bounded disturbances, can be utilized to identify robust and collision-free trajectories by checking the intersections with obstacles. However, in many cases, the FRS is not calculated in real time and is too conservative to be used in actual applications. In this paper, we address these issues by introducing a nonlinear disturbance observer (NDOB) and an adaptive controller to the multirotor system. We express the FRS of the closed-loop multirotor system with an adaptive controller in augmented state space using Hamilton-Jacobi reachability analysis. Then, we derive a closed-form expression that over-approximates the FRS as an ellipsoid, allowing for real-time computation. By compensating for disturbances with the adaptive controller, our over-approximated FRS can be smaller than other ellipsoidal over-approximations. Numerical examples validate the computational efficiency and the smaller scale of our proposed FRS. | ['Han-Lim Choi', 'Juyeop Han'] | 2022-09-16 | null | null | null | null | ['trajectory-planning'] | ['robots'] | [-2.59940296e-01 2.89331287e-01 4.10891138e-03 4.15590495e-01
-2.67279416e-01 -9.56002355e-01 3.57270002e-01 6.54173940e-02
-3.15943837e-01 9.75104690e-01 -3.79419297e-01 -7.20750988e-01
-6.41687989e-01 -5.21284580e-01 -7.49545634e-01 -7.44155645e-01
-2.40067378e-01 3.09522808e-01 3.17118078e-01 -6.62497997e-01
5.37615307e-02 9.42481339e-01 -1.12864137e+00 -7.67780125e-01
1.02838898e+00 7.97165513e-01 5.05301617e-02 6.05498791e-01
8.77035737e-01 2.79070407e-01 -3.65430564e-01 2.56335348e-01
6.26905739e-01 -2.62359828e-01 -5.37107587e-01 1.46621943e-01
-2.34595895e-01 -2.52632409e-01 -5.99149525e-01 1.38721430e+00
2.71122843e-01 7.59740055e-01 5.35223603e-01 -1.64612329e+00
1.79045826e-01 6.19938485e-02 -1.64976195e-01 -5.27437255e-02
2.84060925e-01 -5.36694787e-02 4.37292844e-01 -4.43517983e-01
5.99483907e-01 9.72763121e-01 4.94259149e-01 2.33034253e-01
-1.04502845e+00 -3.18238288e-01 4.26199697e-02 -1.30269438e-01
-1.90365994e+00 -2.34796833e-02 2.08704069e-01 -5.08609653e-01
6.47744656e-01 7.27829039e-01 7.47861028e-01 2.37754285e-01
8.56896996e-01 7.80045465e-02 6.68420851e-01 5.12046600e-03
1.71319932e-01 1.37842059e-01 -2.42206201e-01 7.50487864e-01
7.13424265e-01 4.03777629e-01 5.50796986e-01 -2.34725922e-01
9.51545537e-01 1.84750617e-01 -6.99739277e-01 -6.90283835e-01
-1.18523359e+00 6.94028795e-01 4.31217700e-01 4.10121977e-02
-2.52177566e-01 -2.01837152e-01 1.70299560e-01 4.16771322e-01
5.20261936e-02 6.56851411e-01 2.40829252e-02 -7.41791576e-02
-3.70662063e-01 5.26586831e-01 8.67115080e-01 1.25545597e+00
2.64949381e-01 4.99867260e-01 3.63174975e-01 9.02460963e-02
5.83724827e-02 8.33491266e-01 -1.34233460e-01 -7.22625017e-01
2.96501458e-01 7.18251646e-01 8.00265133e-01 -9.78820741e-01
-7.07565010e-01 -4.70962465e-01 -7.44367719e-01 4.26859111e-01
5.78488529e-01 -5.71869791e-01 -3.44327182e-01 1.32278371e+00
7.47130871e-01 -1.01252474e-01 3.66652131e-01 1.37481606e+00
-2.07611233e-01 8.66616189e-01 -8.17778647e-01 -5.14314830e-01
1.03685021e+00 -4.72354859e-01 -7.67869115e-01 9.85751674e-02
6.68669403e-01 -5.20083308e-01 5.86763203e-01 5.01478136e-01
-8.69318485e-01 -1.28791809e-01 -1.24181080e+00 5.10021806e-01
2.53504850e-02 1.87715605e-01 -3.58957171e-01 1.38442606e-01
-7.02117503e-01 5.73238969e-01 -7.74365067e-01 -2.31370509e-01
-7.46874452e-01 5.14913857e-01 -3.74733388e-01 4.64362770e-01
-1.24696827e+00 1.36515355e+00 5.60188353e-01 3.50728780e-01
-1.02104473e+00 -2.88474917e-01 -9.12833631e-01 -3.06291729e-01
9.60191190e-01 -3.02629054e-01 1.02830184e+00 -1.85771823e-01
-1.62905228e+00 -8.84589776e-02 4.96097393e-02 -3.21372539e-01
5.89052081e-01 -1.65171206e-01 -3.31187725e-01 1.64479703e-01
-1.61949411e-01 -4.19727206e-01 7.29261696e-01 -1.28154349e+00
-5.87274849e-01 -1.50942668e-01 4.75899786e-01 4.58417177e-01
1.48270220e-01 -2.63068587e-01 -6.50539622e-02 -2.29930401e-01
3.55906993e-01 -1.59728336e+00 -7.20793545e-01 -1.40324295e-01
-5.45815468e-01 2.57032394e-01 8.02629054e-01 -3.36055607e-01
1.25050545e+00 -2.05692959e+00 5.41169763e-01 4.73722100e-01
-9.72529724e-02 3.41454744e-01 3.57453942e-01 8.19381237e-01
2.09236041e-01 -3.04297239e-01 -2.21221372e-02 3.68863791e-01
-1.78914621e-01 1.56109691e-01 -7.04616964e-01 1.21631885e+00
-2.44339034e-02 1.80432588e-01 -1.02890563e+00 -1.02523848e-01
3.43612969e-01 3.53758901e-01 -3.29508156e-01 8.93856883e-02
1.36697933e-01 6.86301291e-01 -7.48225868e-01 4.26991910e-01
5.77652216e-01 5.60244083e-01 1.31869644e-01 7.23684579e-02
-9.86606836e-01 -3.19889039e-01 -1.71971393e+00 9.33134615e-01
-5.09202480e-01 5.43051183e-01 7.25977838e-01 -8.02184224e-01
1.11373365e+00 1.63165361e-01 4.95799989e-01 -2.60733038e-01
5.66230536e-01 4.07178670e-01 -3.98291759e-02 -2.76267171e-01
7.54529238e-01 -1.71003237e-01 -2.76895851e-01 -5.04254885e-02
-5.11179566e-01 -4.75857615e-01 -1.36566609e-01 5.18994182e-02
7.97852695e-01 -1.65324077e-01 6.00969255e-01 -7.15800583e-01
9.34729993e-01 3.43292594e-01 7.23757982e-01 1.09189041e-01
-1.90991536e-01 1.97415039e-01 4.88429159e-01 -1.91109136e-01
-9.13289428e-01 -5.67491412e-01 -3.40439618e-01 5.22699282e-02
1.00101101e+00 -2.28940368e-01 -5.39069891e-01 -2.35317916e-01
2.94203103e-01 7.29840457e-01 -2.10138947e-01 -4.22388852e-01
-6.19226217e-01 -1.88603342e-01 3.36620957e-01 1.71539873e-01
-1.06760286e-01 1.93666872e-02 -9.92145777e-01 3.17032784e-01
7.62503222e-03 -9.42263424e-01 -6.11721933e-01 -7.27562159e-02
-4.90333229e-01 -1.48574185e+00 -4.56954986e-01 -5.26795983e-01
9.63248432e-01 4.32955176e-01 2.37536095e-02 8.34046304e-02
3.21450308e-02 3.15207243e-01 -5.94890267e-02 -3.42507213e-01
-3.97281587e-01 -4.90876317e-01 7.56996393e-01 5.57603613e-02
-6.81263328e-01 1.31607354e-01 -3.35058600e-01 1.03083956e+00
-7.81619787e-01 -6.18887581e-02 -2.65119970e-02 5.23105621e-01
8.08635652e-01 5.82978845e-01 4.50137764e-01 -1.27582893e-01
6.06170356e-01 -3.50227773e-01 -1.61823797e+00 -5.85427694e-02
-4.54777926e-01 2.13001296e-02 1.09495974e+00 -5.11170805e-01
-6.83242559e-01 2.15696782e-01 1.55468360e-01 -5.59180379e-01
3.44959974e-01 4.66551393e-01 -9.91025493e-02 -6.20511055e-01
3.83303314e-01 1.43264204e-01 2.91680038e-01 -2.13024825e-01
2.52550691e-01 8.16133857e-01 5.23319185e-01 -3.83390397e-01
1.15113842e+00 5.56253493e-01 7.24507332e-01 -9.53955770e-01
-3.44159424e-01 -6.29679978e-01 -5.80752969e-01 -6.10789716e-01
3.19364458e-01 -6.90472126e-01 -1.21746242e+00 9.40360501e-02
-1.07457590e+00 -1.36635959e-01 -2.05291927e-01 8.72243047e-01
-7.78196990e-01 2.20877826e-01 -3.04320365e-01 -1.55428410e+00
-3.07552274e-02 -1.33869255e+00 7.15357065e-01 8.35293755e-02
2.08923761e-02 -9.47401166e-01 2.56382376e-01 -2.94114202e-01
1.65299922e-01 1.07440662e+00 2.74059176e-01 -2.90950269e-01
-3.65498543e-01 -6.24861777e-01 3.41267139e-01 -1.45139962e-01
-2.05464497e-01 7.00889975e-02 3.07232849e-02 -9.71352100e-01
4.01759297e-01 1.15020372e-01 1.14615604e-01 1.41550735e-01
2.16105565e-01 -6.99853420e-01 -7.58552015e-01 5.40865183e-01
1.82300997e+00 6.50322855e-01 2.17222050e-01 5.23542583e-01
4.75111306e-01 6.02140784e-01 1.42240524e+00 6.31339014e-01
3.53028476e-01 8.52334499e-01 8.46721888e-01 2.59638071e-01
6.57008469e-01 -2.00172719e-02 5.40186822e-01 5.07756948e-01
-1.47267863e-01 -9.84171480e-02 -1.05825531e+00 7.94573426e-01
-2.05807614e+00 -5.17236710e-01 -4.99946445e-01 2.68117785e+00
4.21816230e-01 1.93319544e-02 1.04855299e-01 2.10750207e-01
9.45281088e-01 -3.97549570e-01 -5.48125565e-01 -6.35206938e-01
2.93737054e-01 -7.06013441e-01 1.05275249e+00 9.16341186e-01
-8.75521004e-01 2.62506038e-01 5.30207872e+00 3.09418261e-01
-1.28521276e+00 -3.39822650e-01 -5.02548635e-01 -1.85096525e-02
9.48476419e-03 4.67407167e-01 -9.28749919e-01 1.44526735e-01
1.07482266e+00 -7.90824115e-01 5.63038766e-01 9.65278745e-01
4.73843247e-01 -2.19400674e-01 -6.01088464e-01 4.01923597e-01
-1.05269335e-01 -7.89646566e-01 -4.61894393e-01 2.26663366e-01
8.64658773e-01 -3.61395180e-01 -2.60754049e-01 7.38018379e-02
-2.59792525e-03 -4.54067737e-01 9.85393286e-01 5.81171215e-01
6.24460399e-01 -1.17746735e+00 7.63409138e-01 7.73489237e-01
-1.39048588e+00 -3.00587147e-01 -4.53540832e-01 -2.60455012e-01
7.56345391e-01 3.21056604e-01 -9.18763936e-01 9.88086522e-01
1.20554417e-02 3.50046247e-01 9.73707885e-02 1.31057048e+00
-2.39220839e-02 -1.22663543e-01 -4.82603341e-01 -4.42987591e-01
5.30417204e-01 -7.02792227e-01 1.48591340e+00 6.02602780e-01
4.89954829e-01 3.26108456e-01 4.94046599e-01 5.56466937e-01
7.60730803e-01 5.86302020e-02 -9.26493943e-01 1.49876669e-01
2.89338499e-01 1.06938982e+00 -5.30827045e-01 -1.27684504e-01
-4.92178984e-02 4.06356871e-01 -3.05313990e-02 4.71832722e-01
-1.10430467e+00 -1.15465653e+00 8.75526130e-01 3.54811251e-01
-1.68387651e-01 -7.92400956e-01 6.85692728e-02 -9.45838988e-01
-1.41382664e-01 -6.07183218e-01 -1.60629265e-02 -5.12489438e-01
-1.71590924e-01 8.40701520e-01 2.78040975e-01 -2.04502058e+00
-3.97491813e-01 -3.84667575e-01 -3.08191061e-01 7.20814288e-01
-1.18271136e+00 -4.63728935e-01 7.52600059e-02 4.90063250e-01
7.13445991e-02 2.24820763e-01 4.92377520e-01 5.11540510e-02
-8.53049755e-01 4.24762722e-03 5.02098322e-01 -3.22207928e-01
3.70973587e-01 -1.11152864e+00 -6.64578527e-02 1.15836167e+00
-7.43957400e-01 8.68501365e-01 1.26949656e+00 -8.48661900e-01
-2.00359273e+00 -1.38403761e+00 4.62997288e-01 -9.32675004e-02
9.69325781e-01 -1.10735372e-01 -7.90918708e-01 9.32438910e-01
-9.65141654e-02 1.06147356e-01 -3.30929458e-01 -7.10336506e-01
4.52574581e-01 5.67777120e-02 -1.19130802e+00 9.05095994e-01
5.41674614e-01 -1.20594904e-01 -2.92505294e-01 2.17998862e-01
7.93822408e-01 -1.05912268e+00 -1.08870518e+00 5.67418218e-01
3.91445845e-01 -3.15266132e-01 6.69845939e-01 -4.47227299e-01
-5.00615537e-01 -9.83384788e-01 8.41234103e-02 -1.67967105e+00
-8.14852491e-02 -1.28105283e+00 -6.31527305e-02 5.57077527e-01
6.62919879e-02 -1.00727773e+00 1.81359440e-01 4.07111913e-01
-5.11553764e-01 -9.35454130e-01 -1.26666653e+00 -1.41441667e+00
-1.11980423e-01 -1.81597665e-01 3.44234258e-01 7.23800302e-01
6.25973701e-01 -1.97049841e-01 -4.99336213e-01 1.01766241e+00
5.71069300e-01 6.03236668e-02 8.51696849e-01 -9.21585560e-01
1.20848760e-01 7.68207610e-02 -3.96201849e-01 -8.95540535e-01
1.31335437e-01 -5.68582177e-01 4.64522064e-01 -1.51409090e+00
-6.69043422e-01 -4.50306386e-01 1.73502818e-01 1.00490429e-01
1.90534934e-01 -1.48979157e-01 1.43365096e-02 1.52407378e-01
-3.08995157e-01 6.90052986e-01 1.43688011e+00 1.26230553e-01
-5.51323712e-01 3.40131998e-01 4.49116863e-02 8.68653536e-01
8.56334448e-01 -3.80251229e-01 -6.04545832e-01 -6.05685860e-02
1.08676672e-01 7.16900051e-01 1.53146684e-01 -1.09389281e+00
2.98042089e-01 -6.79658771e-01 -5.20506978e-01 -7.40336239e-01
3.87908518e-01 -1.29187226e+00 6.79331720e-01 1.07940638e+00
1.06827192e-01 2.20362023e-01 3.72139663e-01 7.11895227e-01
-2.12626681e-01 -2.97770381e-01 8.19342911e-01 4.17667478e-01
-3.04288208e-01 6.82907999e-02 -6.86886966e-01 -2.22583175e-01
1.64861178e+00 -1.10694036e-01 -2.40691990e-01 -3.78878862e-01
-4.80525017e-01 5.89076102e-01 6.09422982e-01 2.85336733e-01
7.47679830e-01 -1.20452905e+00 -2.28300944e-01 4.11664516e-01
-5.05877705e-03 4.11181711e-02 1.51167452e-01 1.13096929e+00
-7.63426483e-01 7.93238282e-01 -6.66075945e-02 -4.03851539e-01
-1.07423019e+00 6.69127584e-01 5.55338144e-01 2.23444477e-01
-5.93739450e-01 9.44274515e-02 -2.76670873e-01 -3.26432943e-01
-1.63652692e-02 -6.73784554e-01 -9.14132148e-02 -2.67446131e-01
5.09256423e-01 7.81757891e-01 -7.43736029e-02 -9.72512066e-01
-4.42835391e-01 7.56361842e-01 4.77145553e-01 -2.22503498e-01
9.93481934e-01 -4.49507892e-01 -1.41881898e-01 1.89344257e-01
8.38914871e-01 2.20487982e-01 -1.31413376e+00 2.40554288e-01
-3.06396633e-01 -4.72041577e-01 4.44096774e-02 -7.97139928e-02
-7.08315134e-01 1.71841219e-01 4.26158719e-02 4.38478857e-01
8.17665875e-01 -5.95526755e-01 6.50525749e-01 6.02538824e-01
7.77554810e-01 -1.00917172e+00 -4.52640444e-01 8.70695591e-01
1.31979501e+00 -5.72617590e-01 -4.82598841e-02 -7.00234234e-01
-7.60921061e-01 1.25238764e+00 6.37082398e-01 -6.00182533e-01
5.46196640e-01 4.30003732e-01 -1.20019317e-01 3.24065745e-01
-5.36143184e-01 -5.82963154e-02 4.10309404e-01 2.70655692e-01
-2.63674349e-01 7.79997483e-02 -7.53869057e-01 4.30123866e-01
-6.31354600e-02 -3.92300785e-01 1.19522643e+00 1.02802110e+00
-7.71498799e-01 -5.40518463e-01 -8.68459940e-01 -1.29408106e-01
-4.46733311e-02 6.27748787e-01 5.02861552e-02 1.13684392e+00
-2.21973479e-01 1.07284462e+00 -5.68307117e-02 -4.33253080e-01
1.04056168e+00 -4.70918030e-01 2.93881655e-01 -3.36098224e-01
-2.09161192e-01 -3.70487794e-02 2.01303780e-01 -4.89599407e-01
2.12328151e-01 -3.88342142e-01 -1.78165293e+00 -2.92467862e-01
-9.61040556e-01 5.89968562e-01 5.69971204e-01 7.06571639e-01
1.63318515e-01 3.66991550e-01 8.43339980e-01 -5.96498668e-01
-1.06324971e+00 -5.31651318e-01 -7.79007554e-01 -2.26169482e-01
8.50472689e-01 -9.88408267e-01 -5.71937859e-01 -6.03443384e-01] | [5.157236576080322, 2.2069144248962402] |
aa05744c-24fb-4d3c-9dca-394d58acdbc8 | towards-personalized-healthcare-in-cardiac | 2207.05138 | null | https://arxiv.org/abs/2207.05138v1 | https://arxiv.org/pdf/2207.05138v1.pdf | Towards Personalized Healthcare in Cardiac Population: The Development of a Wearable ECG Monitoring System, an ECG Lossy Compression Schema, and a ResNet-Based AF Detector | Cardiovascular diseases (CVDs) are the number one cause of death worldwide. While there is growing evidence that the atrial fibrillation (AF) has strong associations with various CVDs, this heart arrhythmia is usually diagnosed using electrocardiography (ECG) which is a risk-free, non-intrusive, and cost-efficient tool. Continuously and remotely monitoring the subjects' ECG information unlocks the potentials of prompt pre-diagnosis and timely pre-treatment of AF before the development of any life-threatening conditions/diseases. Ultimately, the CVDs associated mortality could be reduced. In this manuscript, the design and implementation of a personalized healthcare system embodying a wearable ECG device, a mobile application, and a back-end server are presented. This system continuously monitors the users' ECG information to provide personalized health warnings/feedbacks. The users are able to communicate with their paired health advisors through this system for remote diagnoses, interventions, etc. The implemented wearable ECG devices have been evaluated and showed excellent intra-consistency (CVRMS=5.5%), acceptable inter-consistency (CVRMS=12.1%), and negligible RR-interval errors (ARE<1.4%). To boost the battery life of the wearable devices, a lossy compression schema utilizing the quasi-periodic feature of ECG signals to achieve compression was proposed. Compared to the recognized schemata, it outperformed the others in terms of compression efficiency and distortion, and achieved at least 2x of CR at a certain PRD or RMSE for ECG signals from the MIT-BIH database. To enable automated AF diagnosis/screening in the proposed system, a ResNet-based AF detector was developed. For the ECG records from the 2017 PhysioNet CinC challenge, this AF detector obtained an average testing F1=85.10% and a best testing F1=87.31%, outperforming the state-of-the-art. | ['Kwong-Sak Leung', 'Jia-Min Chen', 'Alex Pui-Wai Lee', 'Kam-Sang Woo', 'Yee Leung', 'Yu Zhou', 'Ya-Fen Chan', 'Sheung-Lai Lo', 'Peng-Fei Liu', 'Wei-Ying Yi'] | 2022-07-11 | null | null | null | null | ['electrocardiography-ecg'] | ['methodology'] | [ 3.23113799e-01 -5.60858063e-02 -5.20325825e-02 -3.01780194e-01
-5.06180167e-01 -3.21953624e-01 -2.49480113e-01 4.96524811e-01
-2.58268118e-01 7.22481787e-01 -4.67339791e-02 -3.87257397e-01
-4.58261758e-01 -7.91717529e-01 -1.81250319e-01 -5.96233785e-01
-4.94174570e-01 1.32454515e-01 -1.19577639e-01 3.51863861e-01
-2.03917958e-02 4.09264207e-01 -1.41693676e+00 8.53570253e-02
1.13886869e+00 1.40063953e+00 -8.79518241e-02 7.04628468e-01
5.43611646e-01 1.12163216e-01 -8.10875237e-01 -5.58674149e-02
9.08705816e-02 -6.40276551e-01 -1.55773029e-01 -5.20544648e-01
-6.24391995e-02 -4.81963128e-01 2.54891105e-02 5.61358333e-01
1.16690540e+00 -4.36439127e-01 3.56136560e-01 -7.73853719e-01
-3.27699282e-03 4.47189718e-01 -3.04319620e-01 3.24764222e-01
4.48679209e-01 8.45387504e-02 2.73059487e-01 -6.20036006e-01
3.11524808e-01 5.07511854e-01 1.26476657e+00 2.42984280e-01
-1.03162384e+00 -6.04803026e-01 -7.47784436e-01 1.36467665e-01
-1.74304867e+00 -3.23112458e-01 6.50487006e-01 -2.34769121e-01
6.47065520e-01 7.69200206e-01 1.16771829e+00 7.88051605e-01
6.34363890e-01 -2.38990322e-01 8.52290511e-01 -4.36826497e-01
2.83260107e-01 2.26213843e-01 1.54658467e-01 4.30292487e-01
7.24789500e-01 1.42946318e-01 -3.45926166e-01 -5.45754969e-01
7.16248751e-01 3.44301522e-01 -3.57137263e-01 2.55652577e-01
-1.16565764e+00 2.79730052e-01 2.15179637e-01 4.20376748e-01
-6.88666046e-01 -1.30848572e-01 5.66769660e-01 1.98348254e-01
2.37054437e-01 2.18847677e-01 -3.49763066e-01 -3.50243926e-01
-9.52322662e-01 -7.84963593e-02 7.33463347e-01 5.29529393e-01
4.92359847e-02 1.47895858e-01 -5.45549870e-01 4.86565232e-01
3.47534090e-01 9.37334359e-01 6.14212215e-01 -6.59392357e-01
2.30900005e-01 6.30308509e-01 4.90070656e-02 -1.26073956e+00
-7.10051954e-01 -1.03228617e+00 -1.56444454e+00 -2.22579002e-01
9.15450081e-02 -4.40602213e-01 -3.71738046e-01 1.45112050e+00
2.96409011e-01 2.81218439e-01 -8.32921639e-02 8.31883967e-01
7.89724231e-01 3.71549517e-01 9.63262916e-02 -9.22255218e-01
1.52509141e+00 2.70479936e-02 -9.85391438e-01 2.23226368e-01
3.60759497e-01 -5.03953457e-01 8.80973637e-01 5.67133546e-01
-1.07917702e+00 -7.84079969e-01 -1.23125756e+00 5.22234619e-01
3.44499052e-01 3.57619643e-01 2.77939975e-01 1.15882123e+00
-8.49579036e-01 7.96282887e-01 -8.37692261e-01 -4.59388256e-01
4.92169231e-01 2.09288463e-01 1.11355931e-02 1.71280533e-01
-1.29076600e+00 6.52100444e-01 -5.19104395e-03 2.24246252e-02
-7.12048590e-01 -9.14950311e-01 -3.56002033e-01 8.91654044e-02
-1.93313554e-01 -9.59696531e-01 5.54705143e-01 -5.44072151e-01
-1.37340939e+00 4.97819394e-01 -1.03362009e-01 -6.13790751e-01
5.23651659e-01 -4.92792904e-01 -8.88171852e-01 3.97462428e-01
-1.76596791e-01 -2.04277307e-01 6.41842127e-01 -5.53249538e-01
-3.02912056e-01 -6.40523970e-01 -6.07712090e-01 9.02122930e-02
-4.52377796e-01 -2.32536048e-01 1.30221531e-01 -7.51302600e-01
2.72982359e-01 -7.57880092e-01 -7.78412595e-02 -1.48245320e-02
-2.31424004e-01 5.04665256e-01 6.21520817e-01 -9.53380346e-01
1.70011127e+00 -2.28458691e+00 -2.23717287e-01 5.51892579e-01
2.94890583e-01 5.63956678e-01 5.02491295e-01 2.34590203e-01
6.69602156e-02 1.47807837e-01 -2.77066916e-01 2.51556069e-01
-5.03690720e-01 1.14369459e-01 5.87819293e-02 7.33606994e-01
-1.84712738e-01 7.15707600e-01 -5.19614398e-01 -2.05372736e-01
1.92700446e-01 8.68361652e-01 -3.16109061e-01 2.85750926e-01
5.42024195e-01 6.94956481e-01 -2.79759943e-01 6.99735522e-01
4.87519860e-01 -1.45972148e-01 5.22236168e-01 -3.67207199e-01
1.38122156e-01 2.31928721e-01 -1.39351559e+00 1.52727473e+00
-1.20066859e-01 1.50064617e-01 -4.41015512e-02 -9.46828127e-01
1.24173832e+00 9.36504126e-01 6.98973835e-01 -6.05804861e-01
1.39685526e-01 2.42884129e-01 1.60577178e-01 -6.44234002e-01
-3.83766383e-01 8.57730433e-02 1.96773142e-01 3.92014027e-01
-2.78806090e-01 3.57770801e-01 -2.49246836e-01 -6.83673024e-02
1.16392756e+00 -2.65694499e-01 5.89165986e-01 -5.51356971e-01
5.74617028e-01 -5.00557423e-01 6.87847376e-01 6.37961507e-01
-7.60803446e-02 4.95578676e-01 -6.77729845e-02 -6.39121413e-01
-5.30044496e-01 -1.09106565e+00 -4.80391681e-01 1.14182562e-01
1.30223110e-02 -6.90601289e-01 -6.29353344e-01 -2.56761134e-01
-1.28763303e-01 3.05701196e-01 4.83864406e-03 -3.86285901e-01
-3.44072372e-01 -8.47545385e-01 9.62252319e-01 3.49072218e-01
6.71343148e-01 -8.36835444e-01 -1.31379640e+00 4.92740870e-01
-4.30175185e-01 -5.77965736e-01 -1.15098976e-01 -1.19513571e-01
-1.36537361e+00 -9.31234777e-01 -7.79291809e-01 -3.36313903e-01
3.65176857e-01 -2.04020619e-01 1.00168216e+00 2.00397372e-01
-6.09529138e-01 1.52484626e-01 -1.82191923e-01 -5.88576436e-01
-3.09020847e-01 -5.71166538e-02 4.91267055e-01 1.23363085e-01
-5.57442047e-02 -9.81249273e-01 -1.26683569e+00 3.15152019e-01
-3.13293308e-01 -3.30791086e-01 5.80040514e-01 5.95833123e-01
7.84850478e-01 -1.49940223e-01 1.04127669e+00 -7.20320165e-01
7.58953810e-01 -3.91859442e-01 -3.18499982e-01 3.77701893e-02
-1.30821598e+00 -6.10431015e-01 5.83435774e-01 -3.91579807e-01
-5.59135735e-01 6.55668005e-02 -2.09122255e-01 7.26600701e-04
-2.44389459e-01 4.17988926e-01 -2.38198519e-01 1.24024414e-01
8.33547473e-01 1.94624692e-01 1.02811716e-01 -3.61284524e-01
-1.64176300e-01 8.59657466e-01 6.36518955e-01 -3.10404986e-01
3.35071653e-01 2.30804197e-02 2.41590351e-01 -8.75852764e-01
-1.69779629e-01 -3.06930929e-01 -1.91420034e-01 -3.34166437e-01
7.17678070e-01 -1.09296048e+00 -9.32419062e-01 3.54309559e-01
-9.02019083e-01 3.26999784e-01 -3.72107655e-01 6.43381119e-01
-1.58311024e-01 4.80682999e-01 -5.34063280e-01 -1.05613863e+00
-1.33188868e+00 -6.22661650e-01 3.70086551e-01 1.39723212e-01
-6.24340296e-01 -3.77952993e-01 8.96587316e-03 3.45595405e-02
7.46343017e-01 7.82764375e-01 6.51268542e-01 -4.28336233e-01
1.95302144e-02 -3.99183244e-01 2.33045533e-01 3.73972088e-01
4.29808408e-01 -2.53194481e-01 -8.70004594e-01 -4.07940894e-01
3.11103404e-01 3.81175607e-01 1.59189880e-01 5.49948037e-01
1.01659560e+00 -5.80288768e-01 -3.91940325e-01 7.32995152e-01
1.16606712e+00 5.51079631e-01 9.45276976e-01 -1.83616877e-01
2.52445847e-01 4.91382815e-02 4.03664947e-01 8.08869481e-01
9.24572349e-02 5.13461471e-01 2.91906804e-01 -2.38727063e-01
1.02487572e-01 2.33070925e-01 3.34957242e-01 8.79958510e-01
-5.33303857e-01 -2.13700403e-02 -8.13599467e-01 2.20707402e-01
-1.55575705e+00 -9.64363575e-01 -6.37568951e-01 2.57151246e+00
8.87031972e-01 9.69395936e-02 2.16644868e-01 7.91602015e-01
6.57337487e-01 -4.79134411e-01 -6.48512006e-01 -4.72760111e-01
-4.49027382e-02 4.82664019e-01 2.81089038e-01 7.98011869e-02
-9.12774384e-01 -2.10580394e-01 5.83489990e+00 5.67397438e-02
-1.24642825e+00 2.14245752e-01 7.59368420e-01 2.03548768e-03
2.03548700e-01 -3.68811846e-01 -3.80717695e-01 6.51395738e-01
1.49078071e+00 -2.20120475e-01 1.34940013e-01 6.22472465e-01
4.60920751e-01 7.39668822e-03 -8.13760281e-01 1.23364651e+00
-1.67648718e-01 -1.17360342e+00 -3.87359619e-01 -5.29019088e-02
2.07206771e-01 -3.32073212e-01 -3.34869266e-01 -1.83435425e-01
-8.77935767e-01 -8.24101210e-01 3.10441166e-01 9.49927807e-01
1.53552651e+00 -8.18441868e-01 1.09198856e+00 3.48376244e-01
-1.14808691e+00 -1.19464234e-01 -4.66912873e-02 -3.28375667e-01
2.53038377e-01 1.33377421e+00 -7.46470690e-01 8.54643822e-01
8.57594848e-01 5.76115608e-01 -4.28268045e-01 1.08160031e+00
4.17579934e-02 1.04954875e+00 -6.23886526e-01 2.66574860e-01
-9.07018006e-01 -1.63164288e-01 5.60577452e-01 1.11636305e+00
8.39671075e-01 5.33986688e-01 -2.43528504e-02 4.80267972e-01
2.73140430e-01 2.30096817e-01 -3.92815024e-01 3.94618750e-01
5.94637990e-01 1.11143970e+00 -3.68296266e-01 -4.17342722e-01
2.59783398e-02 5.49651384e-01 -5.61073780e-01 -1.95233747e-02
-9.60731208e-01 -7.05606818e-01 2.90637523e-01 6.55577540e-01
-2.90529549e-01 1.73473254e-01 -7.84985662e-01 -8.35654199e-01
3.33951980e-01 -9.64783192e-01 4.84289438e-01 -3.27328920e-01
-8.82249117e-01 7.11986005e-01 -3.59508157e-01 -1.52849996e+00
-3.37790340e-01 2.57568359e-01 -5.94835043e-01 1.10785890e+00
-8.51989686e-01 -5.59370697e-01 -6.24930024e-01 6.08743966e-01
-2.04625338e-01 -4.82258014e-02 1.47442412e+00 7.47260511e-01
-3.55239481e-01 7.39419758e-01 -2.62588382e-01 -2.24814162e-01
6.64778709e-01 -7.38654494e-01 -1.07780591e-01 8.52214575e-01
-2.60174364e-01 7.38261759e-01 6.16702318e-01 -8.02256227e-01
-1.40721416e+00 -1.06839836e+00 1.09977388e+00 -1.80273637e-05
-1.86651915e-01 -8.80513061e-03 -8.54207397e-01 -3.14117745e-02
-8.78616422e-02 1.38126642e-01 7.69417584e-01 -2.71618456e-01
2.43366838e-01 -7.81484306e-01 -1.47451591e+00 1.65951133e-01
8.37979078e-01 -2.50748992e-01 -3.95246357e-01 1.07016712e-02
2.66206861e-01 -2.44404271e-01 -1.35353959e+00 8.71215582e-01
1.04829550e+00 -1.01302683e+00 1.08579040e+00 -6.48750588e-02
-5.39702289e-02 -3.34794492e-01 1.23023070e-01 -8.42564404e-01
-4.26266193e-01 -7.54935741e-01 -3.85856867e-01 1.45958972e+00
2.93956339e-01 -7.84648299e-01 3.91214967e-01 4.90147054e-01
-3.97372283e-02 -7.33051062e-01 -1.01044834e+00 -5.06200194e-01
-7.89583147e-01 -3.04542542e-01 2.87385792e-01 7.82656431e-01
1.68103337e-01 2.67480969e-01 -4.68295962e-01 1.90749153e-01
5.90123951e-01 -1.89313889e-01 3.70655149e-01 -1.57327867e+00
-3.80672097e-01 1.29651144e-01 -4.30945724e-01 -1.80018365e-01
-9.16462481e-01 -8.57240319e-01 -4.46889371e-01 -1.48257160e+00
-1.19445361e-01 -5.04546940e-01 -6.49198472e-01 4.08343405e-01
-1.07241496e-01 3.18085283e-01 -1.04899287e-01 1.70294404e-01
-4.61826362e-02 1.45541325e-01 7.39454210e-01 6.56514913e-02
-6.56057715e-01 4.20829803e-01 -5.30067384e-01 5.68657935e-01
9.98717248e-01 -6.31361902e-01 -5.34290135e-01 1.04779147e-01
7.45026991e-02 6.00004971e-01 2.97409505e-01 -1.56378520e+00
8.15481469e-02 4.38749194e-01 7.41227627e-01 -4.61548597e-01
9.95213762e-02 -9.74398613e-01 1.00292397e+00 1.11291897e+00
6.31508753e-02 2.59655923e-01 2.26946086e-01 3.78664315e-01
5.15342131e-02 4.22596514e-01 6.29374504e-01 2.32594088e-01
3.15527260e-01 1.37900367e-01 -5.07648051e-01 -1.23503745e-01
1.09958529e+00 -2.37384275e-01 -9.52944010e-02 -2.37648666e-01
-7.68672526e-01 -1.85394496e-01 -1.25321344e-01 -1.02964777e-03
8.08223784e-01 -1.11696541e+00 -8.09480488e-01 5.07257819e-01
-1.00261331e-01 -1.27966776e-01 4.53978121e-01 1.32439840e+00
-6.61363125e-01 6.47101328e-02 -3.75887632e-01 -8.15868318e-01
-1.41026700e+00 2.22167701e-01 2.31115893e-01 -1.28960252e-01
-9.90047216e-01 3.20453346e-01 -5.50182045e-01 3.77343655e-01
3.97830486e-01 -3.46854240e-01 -2.17396364e-01 1.80444531e-02
7.80309975e-01 7.94243634e-01 5.39488435e-01 -6.56198263e-02
-7.28658378e-01 5.83384514e-01 5.13875902e-01 1.72077894e-01
1.20730495e+00 -1.92062944e-01 -1.46719918e-01 3.50122601e-01
6.24762714e-01 7.49361590e-02 -5.08291841e-01 2.42247820e-01
-1.86785251e-01 -1.54979050e-01 -6.11777902e-02 -1.23141348e+00
-1.13073039e+00 8.04410994e-01 1.41426909e+00 3.05890471e-01
1.66959143e+00 -5.43206871e-01 8.76997352e-01 -3.25420909e-02
4.68228102e-01 -5.21267354e-01 -5.08683681e-01 -1.29104912e-01
7.30055034e-01 -4.67463672e-01 2.42953733e-01 -7.93107748e-02
-5.25199234e-01 1.08991718e+00 -3.68085988e-02 -4.59182262e-02
9.91755724e-01 2.81610727e-01 1.11322969e-01 6.87043509e-03
-4.23699200e-01 5.07545233e-01 2.86344290e-01 7.62671888e-01
6.03693366e-01 2.18997434e-01 -9.71958935e-01 1.23803699e+00
-4.77244891e-02 5.23429692e-01 1.99165344e-01 7.45425582e-01
-2.88121253e-01 -9.24677193e-01 -3.63571078e-01 7.70246565e-01
-8.93018007e-01 -5.13071753e-03 6.44740611e-02 2.26670459e-01
3.39983642e-01 1.21824324e+00 -2.86588699e-01 -3.95873129e-01
4.41080749e-01 1.83288068e-01 1.92289472e-01 -2.39005268e-01
-9.72737789e-01 2.17708156e-01 1.87420338e-01 -6.87748909e-01
-3.92955303e-01 -5.53520620e-01 -1.04999220e+00 -1.98739454e-01
-1.55143529e-01 1.61650300e-01 7.03385353e-01 5.68345904e-01
9.26243246e-01 8.87968957e-01 5.87245464e-01 -2.40388170e-01
-4.52986121e-01 -1.05322123e+00 -7.45017707e-01 1.92084506e-01
2.20501527e-01 -1.72470495e-01 -1.89080670e-01 2.78737307e-01] | [14.200325965881348, 3.239351987838745] |
d6238a94-d27a-4039-ae18-004613c28a40 | elvis-empowering-locality-of-vision-language | 2304.05303 | null | https://arxiv.org/abs/2304.05303v1 | https://arxiv.org/pdf/2304.05303v1.pdf | ELVIS: Empowering Locality of Vision Language Pre-training with Intra-modal Similarity | Deep learning has shown great potential in assisting radiologists in reading chest X-ray (CXR) images, but its need for expensive annotations for improving performance prevents widespread clinical application. Visual language pre-training (VLP) can alleviate the burden and cost of annotation by leveraging routinely generated reports for radiographs, which exist in large quantities as well as in paired form (imagetext pairs). Additionally, extensions to localization-aware VLPs are being proposed to address the needs of accurate localization of abnormalities for CAD in CXR. However, we find that the formulation proposed by locality-aware VLP literatures actually leads to loss in spatial relationships required for downstream localization tasks. Therefore, we propose Empowering Locality of VLP with Intra-modal Similarity, ELVIS, a VLP aware of intra-modal locality, to better preserve the locality within radiographs or reports, which enhances the ability to comprehend location references in text reports. Our locality-aware VLP method significantly outperforms state-of-the art baselines in multiple segmentation tasks and the MS-CXR phrase grounding task. Qualitatively, ELVIS is able to focus well on regions of interest described in the report text compared to prior approaches, allowing for enhanced interpretability. | ['Thijs Kooi', 'Tae Soo Kim', 'Jaewoo Kang', 'Jaewoong Shin', 'Sumin Seo'] | 2023-04-11 | null | null | null | null | ['phrase-grounding'] | ['natural-language-processing'] | [ 1.63953677e-01 2.49157354e-01 -4.55484241e-01 -4.75741029e-01
-1.56967747e+00 -6.42711699e-01 3.80942553e-01 8.05835307e-01
-4.46755946e-01 4.37841117e-01 6.45072281e-01 -9.00904953e-01
-1.38126433e-01 -4.46220636e-01 -7.89918900e-01 -2.40524352e-01
5.15394919e-02 4.30952013e-01 3.30699652e-01 2.81069934e-01
5.99848181e-02 5.85641146e-01 -6.69689894e-01 9.12618399e-01
5.18121958e-01 6.80297732e-01 6.37701929e-01 8.08400452e-01
-1.59742147e-01 1.31805372e+00 -5.02948105e-01 -1.35658115e-01
2.27576923e-02 -5.48572600e-01 -1.05801535e+00 9.84233692e-02
7.44491279e-01 -4.56108958e-01 -4.25161242e-01 6.81857526e-01
6.07805073e-01 1.61142088e-02 6.84254289e-01 -5.08642077e-01
-9.23574924e-01 5.40479243e-01 -1.00913703e+00 9.81145263e-01
2.66620785e-01 3.70049924e-01 1.06740785e+00 -8.13931942e-01
7.48615503e-01 9.27400649e-01 7.94200540e-01 1.89939514e-01
-1.08718407e+00 -3.12925726e-01 1.41038015e-01 -6.31035864e-02
-1.41287208e+00 5.75786121e-02 3.99082899e-01 -4.38707083e-01
1.25610089e+00 4.82674539e-01 4.47387159e-01 8.59167337e-01
3.64663988e-01 9.46315944e-01 9.44221675e-01 -3.75106841e-01
-3.46784703e-02 -4.29983735e-02 5.51620647e-02 1.01414931e+00
8.75567123e-02 -2.60093600e-01 -4.77789968e-01 -2.11086139e-01
1.05215859e+00 2.29672477e-01 -5.42077541e-01 -1.62748501e-01
-1.36094344e+00 6.96628153e-01 9.19295371e-01 4.61525708e-01
-5.45844316e-01 2.55940944e-01 5.72244346e-01 -2.53800601e-01
4.60730195e-01 6.79871261e-01 -9.59320366e-02 4.32346947e-02
-1.27259231e+00 2.93716863e-02 3.45694512e-01 6.80427134e-01
2.51576692e-01 -3.54016423e-01 -7.02824593e-01 7.04528093e-01
1.57987401e-01 1.88351303e-01 6.30265713e-01 -7.39586890e-01
7.97836661e-01 7.17676640e-01 -3.33118200e-01 -8.96121860e-01
-6.52961314e-01 -5.95045269e-01 -6.60740614e-01 -2.09709197e-01
2.73814321e-01 1.43423140e-01 -1.34385037e+00 1.25023711e+00
2.39283461e-02 6.84105828e-02 -1.39068708e-01 9.43659008e-01
9.88795877e-01 5.46902537e-01 3.66532236e-01 1.40228402e-02
1.69560003e+00 -1.09753489e+00 -5.54834127e-01 -2.65628427e-01
9.37826872e-01 -8.56762588e-01 1.41928971e+00 -1.77465215e-01
-1.12766039e+00 -4.13617194e-01 -7.79854953e-01 -4.08131361e-01
2.63430402e-02 1.99197486e-01 4.67455268e-01 2.59218752e-01
-1.18213856e+00 2.25140408e-01 -1.27438080e+00 -2.76002735e-01
7.67546833e-01 3.16251099e-01 -4.05360490e-01 -2.71422803e-01
-7.39588618e-01 1.04268456e+00 4.13335979e-01 -1.53595239e-01
-6.72002733e-01 -1.24117386e+00 -1.00961387e+00 1.23734653e-01
5.44897556e-01 -7.62738347e-01 1.47152054e+00 -3.66196781e-01
-6.00188494e-01 1.00116634e+00 -2.72542894e-01 -6.00235105e-01
6.09939098e-01 -1.98620856e-02 -6.40584975e-02 8.07118237e-01
4.22244072e-01 1.01720953e+00 5.11111975e-01 -1.13093507e+00
-6.11503303e-01 -1.59987152e-01 1.19804442e-01 3.85908633e-01
8.45747534e-03 -1.31636545e-01 -9.47955370e-01 -9.30731177e-01
1.33887321e-01 -8.07288408e-01 -4.17301834e-01 3.00803453e-01
-4.45585161e-01 -2.04950079e-01 9.15515423e-01 -1.08476663e+00
1.19157183e+00 -2.22680020e+00 -3.98897588e-01 2.46605381e-01
6.78723693e-01 1.49309516e-01 6.53189719e-02 1.52261797e-02
-1.11838855e-01 3.82319033e-01 -2.81050235e-01 -5.23429573e-01
-4.27860171e-01 4.96596754e-01 -2.54540920e-01 4.83297557e-01
2.72612184e-01 1.32542133e+00 -8.79426658e-01 -1.03811181e+00
3.80948544e-01 5.09048879e-01 -5.91686487e-01 1.21538341e-01
-1.04472227e-01 6.74143732e-01 -4.28146064e-01 7.41993546e-01
1.08225502e-01 -1.12450361e+00 6.92802444e-02 -3.65030736e-01
7.82844871e-02 5.47531128e-01 -5.60898542e-01 1.83636594e+00
-7.95113385e-01 7.01825738e-01 -2.03811571e-01 -5.55657029e-01
2.87145764e-01 4.42771614e-01 5.57325959e-01 -7.53216088e-01
-8.96505713e-02 1.28268927e-01 1.14226624e-01 -5.90397120e-01
4.06691223e-01 -1.04954392e-01 1.87430307e-01 6.50836170e-01
-2.27421179e-01 -1.35170281e-01 2.02806920e-01 6.50835454e-01
1.21904004e+00 -1.73214391e-01 5.55916607e-01 -9.73444134e-02
2.69388855e-01 2.78112680e-01 -3.67012471e-02 1.19858670e+00
-1.33833125e-01 1.09839559e+00 3.15472633e-01 -1.77816927e-01
-1.00440454e+00 -1.10876453e+00 -2.41616115e-01 8.57571065e-01
2.40587611e-02 -3.89758974e-01 -3.32047313e-01 -1.00124407e+00
-1.04441635e-01 7.07368672e-01 -6.56383693e-01 1.95081085e-01
-9.88341987e-01 -6.01499200e-01 6.54025972e-01 1.18619633e+00
2.40119204e-01 -8.92321348e-01 -9.09940302e-01 9.85813439e-02
-2.80169785e-01 -1.24583983e+00 -7.21689463e-01 3.62341076e-01
-7.27144957e-01 -9.64033544e-01 -1.20591593e+00 -8.24475229e-01
8.39043677e-01 1.07358478e-01 1.28928936e+00 2.28900090e-01
-6.82815790e-01 7.80967832e-01 -2.16598943e-01 -2.22204104e-01
-4.31331307e-01 7.39316344e-02 -6.93342149e-01 -6.37924314e-01
7.85108432e-02 1.85273413e-03 -8.87813985e-01 -1.07043333e-01
-9.88708913e-01 2.48863429e-01 8.99572074e-01 8.38092208e-01
1.06986547e+00 -3.71018440e-01 1.40167207e-01 -1.14901328e+00
6.44205213e-01 -4.64017481e-01 -1.18185811e-01 4.88610536e-01
-4.18814659e-01 2.61226892e-01 3.93077850e-01 -1.90828890e-01
-1.01463687e+00 -2.28161454e-01 -2.31608585e-01 -4.18077886e-01
-2.24726170e-01 5.99066317e-01 7.02436805e-01 -1.55215517e-01
6.53047502e-01 1.97435886e-01 -2.77493656e-01 -3.25978249e-01
4.75985378e-01 4.35063660e-01 8.17603767e-01 -4.67618287e-01
2.86221296e-01 5.07762730e-01 6.86793476e-02 -5.21077275e-01
-1.20252240e+00 -9.23835099e-01 -6.69886410e-01 1.18107729e-01
1.19096601e+00 -8.50984812e-01 -2.94940770e-01 -5.29968619e-01
-1.18310308e+00 -5.89792915e-02 -4.94908810e-01 5.48612356e-01
-3.70093375e-01 3.96776915e-01 -9.68353152e-01 -2.11928919e-01
-3.11628580e-01 -1.62023270e+00 1.35461509e+00 3.74805816e-02
-5.83002090e-01 -1.25298250e+00 -1.95986047e-01 4.91399884e-01
2.52637297e-01 2.27007821e-01 1.23458421e+00 -8.26975524e-01
-6.67664766e-01 -8.56261626e-02 -6.39232159e-01 1.47223264e-01
3.00350040e-01 -4.49872434e-01 -7.47125030e-01 -1.90827638e-01
-8.15380588e-02 -2.87415504e-01 9.49789762e-01 5.92549205e-01
1.58363378e+00 -1.73415929e-01 -4.99545604e-01 7.20250368e-01
1.21163917e+00 2.02025667e-01 4.11888137e-02 1.68398783e-01
1.05081737e+00 2.78998911e-01 2.83627748e-01 7.94213414e-02
3.99057925e-01 3.54371250e-01 2.18621030e-01 -8.78402174e-01
-5.48230767e-01 -2.95337230e-01 -3.97106200e-01 5.51920652e-01
1.77389428e-01 -3.64280283e-01 -1.45535827e+00 7.01500475e-01
-1.58151710e+00 -2.98349231e-01 -2.42213476e-02 1.58160782e+00
1.00833535e+00 5.23766465e-02 -1.82781100e-01 -3.65280032e-01
3.78416151e-01 3.34551454e-01 -4.29958135e-01 -1.01164155e-01
6.67191520e-02 4.52959329e-01 8.32310915e-01 5.16462028e-01
-1.10948503e+00 5.62577426e-01 6.31257200e+00 6.84683621e-01
-1.34393549e+00 4.00218755e-01 1.01262069e+00 -1.01188615e-01
-2.45795429e-01 -2.18059599e-01 -6.06909931e-01 1.89187840e-01
4.99469668e-01 8.53432417e-02 -2.47333184e-01 6.92418337e-01
2.49047488e-01 -2.47905731e-01 -1.48667836e+00 1.06578457e+00
2.34553561e-01 -1.93295157e+00 3.55556160e-01 3.63854282e-02
7.83481896e-01 1.29771262e-01 4.77560520e-01 -2.17461400e-02
-6.41907891e-03 -1.20439005e+00 5.07402599e-01 1.62540644e-01
9.29546237e-01 -4.54155266e-01 8.89146626e-01 1.12725191e-01
-1.16055346e+00 2.87010670e-01 -1.14800252e-01 5.31147420e-01
3.45759332e-01 1.98070437e-01 -1.61942840e+00 6.67533815e-01
5.42259395e-01 5.41762590e-01 -8.98290515e-01 9.50749040e-01
-2.70251930e-01 8.40901077e-01 -2.43276790e-01 3.79878163e-01
8.47457349e-01 4.16079104e-01 3.86335075e-01 1.51437092e+00
6.78220242e-02 3.09521616e-01 5.57390392e-01 9.69748914e-01
-3.05008113e-01 2.51259834e-01 -3.88559639e-01 -8.79791677e-02
2.21180961e-01 8.90286982e-01 -1.20090926e+00 -5.91044307e-01
-5.96233010e-01 8.46702635e-01 1.33237809e-01 2.80123889e-01
-7.56034315e-01 5.61898537e-02 -1.09943591e-01 4.69631642e-01
2.42225945e-01 -1.12295322e-01 -6.37765288e-01 -7.70608604e-01
5.17199636e-02 -7.83846200e-01 7.85697520e-01 -8.70785415e-01
-1.14480126e+00 6.26396656e-01 4.30229940e-02 -9.18602645e-01
-3.39498192e-01 -5.07790387e-01 -3.66158217e-01 9.46684301e-01
-1.56483662e+00 -1.43016803e+00 -1.32241234e-01 5.45623660e-01
8.05882812e-01 2.00992122e-01 4.73947316e-01 3.01835895e-01
-1.40414298e-01 6.43976092e-01 -2.94607431e-01 2.25646824e-01
7.53277540e-01 -1.49026251e+00 2.12399960e-01 8.15134585e-01
4.31325674e-01 9.64338541e-01 2.79225260e-01 -7.32487679e-01
-8.57679009e-01 -1.20503151e+00 7.06363618e-01 -7.27894306e-01
5.81490457e-01 1.02587633e-01 -1.17249930e+00 8.61848235e-01
9.14304703e-02 2.83036619e-01 7.18744755e-01 -1.65835485e-01
-1.99594870e-01 4.94979382e-01 -1.03539407e+00 6.11621976e-01
6.44821942e-01 -9.55426157e-01 -8.44484329e-01 8.10072362e-01
9.56654668e-01 -9.33646441e-01 -9.07612264e-01 3.62084091e-01
4.71478999e-02 -5.33808291e-01 1.25165451e+00 -4.09590662e-01
5.24498641e-01 -2.77708411e-01 2.63400003e-02 -7.53061116e-01
-1.25776142e-01 -1.24569923e-01 3.72356363e-02 7.63049722e-01
5.93237519e-01 -2.76933342e-01 6.68132663e-01 5.12774885e-01
-4.38668042e-01 -1.03111804e+00 -6.89752638e-01 -2.97422737e-01
-2.49523204e-03 -5.51872969e-01 1.59453880e-02 9.59665060e-01
-2.65641063e-01 6.52719811e-02 1.87928796e-01 3.69670153e-01
2.19785348e-01 6.87927008e-02 1.48725688e-01 -3.94339323e-01
-6.61617637e-01 -4.95582044e-01 -2.17758477e-01 -1.19113815e+00
-4.27820273e-02 -1.22831810e+00 -1.20177962e-01 -2.13519931e+00
5.33514023e-01 -4.14082706e-01 -3.91249746e-01 7.87604272e-01
-4.94299680e-01 6.43507838e-01 2.48150736e-01 4.41012681e-01
-6.97621465e-01 -1.29079282e-01 1.51916420e+00 -2.03852519e-01
-2.15260863e-01 -2.23725647e-01 -6.15770221e-01 9.21034396e-01
3.10244620e-01 -6.01238549e-01 -4.44681942e-01 -6.05954826e-01
-5.25712445e-02 3.12579900e-01 5.64656377e-01 -7.49040306e-01
3.29725474e-01 1.45130515e-01 6.18730128e-01 -1.06313872e+00
8.34980160e-02 -6.19947255e-01 -4.34574515e-01 3.95384818e-01
-7.89781272e-01 4.50780213e-01 3.64515185e-01 6.54694319e-01
-4.77976352e-01 -8.37308243e-02 7.51002789e-01 -5.82311392e-01
-6.37348533e-01 2.04854771e-01 -2.64705002e-01 3.09382737e-01
8.63306403e-01 -5.29113561e-02 -3.08448434e-01 -2.14864701e-01
-8.55298698e-01 2.60326654e-01 3.38238269e-01 2.08091542e-01
6.31750584e-01 -6.71054959e-01 -4.65478301e-01 -9.84795168e-02
1.86747596e-01 4.92862314e-01 3.42067212e-01 1.20717061e+00
-1.05117500e+00 8.03401589e-01 2.88259387e-01 -1.07457471e+00
-1.28131342e+00 4.59668308e-01 3.27227771e-01 -7.50288248e-01
-1.20420706e+00 1.12754679e+00 7.99281418e-01 -6.77451491e-02
3.12002718e-01 -8.72645438e-01 5.26452437e-02 -3.95301342e-01
4.87525880e-01 -1.58453539e-01 3.78691256e-01 -4.74783212e-01
-5.52304327e-01 3.79443228e-01 -6.38488352e-01 -5.18914722e-02
1.08007514e+00 -1.10173568e-01 7.71852136e-02 2.83152372e-01
1.19625294e+00 1.46722838e-01 -1.03941429e+00 -2.31350705e-01
1.83038265e-01 -2.79288977e-01 3.01034629e-01 -1.11931026e+00
-8.62958252e-01 9.77626443e-01 6.25266552e-01 -3.58132094e-01
8.94128323e-01 5.47143638e-01 8.69743109e-01 2.94378966e-01
-1.19116612e-01 -4.89769399e-01 3.18993598e-01 8.36079195e-02
7.02905536e-01 -1.37455297e+00 2.04283476e-01 -3.17176580e-01
-9.34970498e-01 9.23770607e-01 5.28165936e-01 1.23397991e-01
3.65361035e-01 5.12417912e-01 2.21658975e-01 -5.04921496e-01
-3.93628806e-01 -5.82853556e-02 8.54880393e-01 3.81526500e-01
8.13746750e-01 7.19747767e-02 -1.31546840e-01 6.86519891e-02
7.66901523e-02 -4.40129608e-01 3.15962642e-01 1.24849141e+00
-1.94395691e-01 -7.95745850e-01 -4.54279542e-01 6.28966331e-01
-9.40408945e-01 -4.81603891e-01 -1.70163810e-01 1.07314110e+00
1.21760830e-01 5.23057580e-01 2.21374422e-01 3.08984548e-01
2.10723042e-01 -1.69468984e-01 3.57400268e-01 -1.05171525e+00
-7.44952261e-01 3.22217405e-01 -5.83531186e-02 -4.97711778e-01
-2.52319276e-01 -3.91239554e-01 -1.62929618e+00 3.03564757e-01
-1.21506318e-01 -9.88496244e-02 4.33361888e-01 1.12555349e+00
1.83224231e-01 1.10745847e+00 -1.45033658e-01 -3.38836730e-01
-3.52607787e-01 -6.61967039e-01 -1.84224755e-01 5.42882383e-01
6.15121961e-01 -3.93161476e-01 4.65787426e-02 2.92580068e-01] | [15.012418746948242, -1.7081348896026611] |
2fa86d8f-b8e4-4d74-a8ab-8a0e9f979163 | learning-in-context-learning-for-named-entity | 2305.11038 | null | https://arxiv.org/abs/2305.11038v3 | https://arxiv.org/pdf/2305.11038v3.pdf | Learning In-context Learning for Named Entity Recognition | Named entity recognition in real-world applications suffers from the diversity of entity types, the emergence of new entity types, and the lack of high-quality annotations. To address the above problems, this paper proposes an in-context learning-based NER approach, which can effectively inject in-context NER ability into PLMs and recognize entities of novel types on-the-fly using only a few demonstrative instances. Specifically, we model PLMs as a meta-function $\mathcal{ \lambda_ {\text{instruction, demonstrations, text}}. M}$, and a new entity extractor can be implicitly constructed by applying new instruction and demonstrations to PLMs, i.e., $\mathcal{ (\lambda . M) }$(instruction, demonstrations) $\to$ $\mathcal{F}$ where $\mathcal{F}$ will be a new entity extractor, i.e., $\mathcal{F}$: text $\to$ entities. To inject the above in-context NER ability into PLMs, we propose a meta-function pre-training algorithm, which pre-trains PLMs by comparing the (instruction, demonstration)-initialized extractor with a surrogate golden extractor. Experimental results on 4 few-shot NER datasets show that our method can effectively inject in-context NER ability into PLMs and significantly outperforms the PLMs+fine-tuning counterparts. | ['Le Sun', 'Xianpei Han', 'Boxi Cao', 'Hua Wu', 'Dai Dai', 'Wei Jia', 'Jie Lou', 'Hongyu Lin', 'Yaojie Lu', 'Jiawei Chen'] | 2023-05-18 | null | null | null | null | ['few-shot-ner'] | ['natural-language-processing'] | [ 5.89476787e-02 3.48815918e-02 7.14247897e-02 -5.50242841e-01
-8.04707229e-01 -9.02536809e-01 2.39368081e-01 4.23073232e-01
-8.34551096e-01 8.24301124e-01 -2.38003686e-01 -4.86220896e-01
-3.83881964e-02 -1.16512346e+00 -1.01764166e+00 -2.49161705e-01
-1.28880329e-02 2.66948879e-01 4.34074312e-01 -3.50361407e-01
4.54347879e-02 3.06643784e-01 -1.79989386e+00 2.19483033e-01
1.26194191e+00 7.40194380e-01 4.68569487e-01 5.37274957e-01
-7.18412280e-01 6.62107229e-01 -1.20809793e+00 -7.96481848e-01
-5.49182892e-02 -2.36980304e-01 -7.92713404e-01 -7.23743796e-01
3.94059032e-01 -1.31990537e-01 -1.91629738e-01 1.11160135e+00
4.52025771e-01 6.15569830e-01 5.25609136e-01 -1.08044064e+00
-6.38362229e-01 9.39098835e-01 -1.18854776e-01 3.18105280e-01
4.73525196e-01 2.46870726e-01 1.02796245e+00 -1.10008323e+00
7.41896033e-01 8.43267441e-01 5.59436560e-01 5.74522018e-01
-1.03061199e+00 -1.08621848e+00 3.93284224e-02 1.92361549e-02
-1.50931919e+00 -1.74190864e-01 2.71977037e-01 -3.09890896e-01
1.31350875e+00 4.03868347e-01 2.42742926e-01 8.41485977e-01
1.88452173e-02 8.06546926e-01 8.61998618e-01 -4.98857737e-01
3.07017148e-01 8.42995495e-02 4.70854282e-01 9.23286140e-01
2.73096323e-01 9.87105668e-02 -2.64917403e-01 -1.47866189e-01
6.39840126e-01 2.28656214e-02 -8.63765329e-02 1.47717327e-01
-9.53909516e-01 5.31078696e-01 2.05578864e-01 6.54029191e-01
-1.22768939e-01 1.08872809e-01 3.18783402e-01 8.22974667e-02
-2.36269072e-01 7.07958341e-01 -9.58364248e-01 -3.68932456e-01
-8.28190506e-01 -2.52915919e-02 8.17285061e-01 1.40438581e+00
1.23292887e+00 2.43981421e-01 -1.83127835e-01 7.83253372e-01
-2.06281155e-01 6.23663604e-01 6.38509929e-01 -6.80615008e-01
6.47308886e-01 1.05386293e+00 1.23905972e-01 -3.35420072e-01
-4.09103394e-01 -1.48366570e-01 -4.44576412e-01 4.02193218e-02
9.04154480e-02 -5.51055729e-01 -8.55348408e-01 1.95183873e+00
4.15722758e-01 5.05268514e-01 2.25617826e-01 2.79945731e-01
1.17379606e+00 7.15599239e-01 5.66090584e-01 -1.30598098e-01
1.37149513e+00 -7.97006726e-01 -4.87249196e-01 -2.28552803e-01
1.06221056e+00 -4.99355882e-01 1.43964541e+00 4.09126841e-02
-8.60126674e-01 -7.00560987e-01 -8.93507898e-01 6.98417276e-02
-9.71408367e-01 2.30892897e-01 4.37234402e-01 7.76350439e-01
-4.20937836e-01 7.96931028e-01 -6.34745538e-01 -7.90431574e-02
7.95224458e-02 4.56740826e-01 -4.74893302e-01 3.65262330e-02
-1.41294718e+00 7.51316130e-01 9.56666946e-01 -1.97785884e-01
-8.28576863e-01 -8.64650905e-01 -1.40043581e+00 2.26486057e-01
6.24154329e-01 -3.55799139e-01 1.16457903e+00 -5.12950122e-01
-1.24014163e+00 5.83209455e-01 -1.27333060e-01 -6.99658617e-02
-1.86448514e-01 -1.77234113e-01 -1.05365765e+00 -1.83382690e-01
4.01198894e-01 4.88341779e-01 5.40485382e-01 -1.18075585e+00
-1.04645228e+00 -1.54232934e-01 5.45485020e-01 -1.05188690e-01
-4.96908009e-01 1.06139436e-01 -2.33852789e-01 -6.84585214e-01
-3.27578366e-01 -6.65240705e-01 -1.21504534e-02 -1.02973735e+00
-5.00298858e-01 -3.96733463e-01 4.71854866e-01 -3.10221285e-01
1.89533412e+00 -2.05670929e+00 -1.93633392e-01 1.67298093e-01
2.43705899e-01 5.97872674e-01 -1.13446862e-01 2.83798873e-01
-2.57751197e-01 4.02854979e-01 -4.64590341e-01 5.64683042e-02
2.67468721e-01 3.96414280e-01 2.31352504e-02 -3.07583511e-01
2.07821250e-01 8.66054714e-01 -1.26946616e+00 -5.20463467e-01
1.01361118e-01 2.49182537e-01 -3.45756263e-01 3.19648832e-01
-2.59457499e-01 1.09281436e-01 -5.75249612e-01 6.94881976e-01
5.52173793e-01 -3.41208540e-02 2.13795617e-01 -2.14879453e-01
-3.83980006e-01 4.31425869e-01 -1.54181802e+00 1.63358843e+00
-5.57094097e-01 9.70795974e-02 -1.79924399e-01 -7.78687418e-01
8.79858375e-01 2.79844910e-01 9.90919992e-02 -5.29593885e-01
1.40946433e-01 3.44822884e-01 -1.41646072e-01 -4.84257877e-01
7.18954861e-01 -3.01981330e-01 -6.82334363e-01 2.97722965e-01
4.87133473e-01 1.14833295e-01 6.61458850e-01 9.52133462e-02
1.54943109e+00 3.33651096e-01 3.19766968e-01 2.04686560e-02
7.30196297e-01 -1.53823253e-02 9.40234244e-01 9.46352243e-01
-1.83731049e-01 -6.42724261e-02 1.73921376e-01 -1.71593517e-01
-7.57362485e-01 -1.27297890e+00 1.43386014e-02 1.69247508e+00
2.58543074e-01 -8.20292771e-01 -9.87202466e-01 -1.28789020e+00
-2.72146761e-01 1.18174529e+00 -3.17541599e-01 -2.55179852e-01
-9.85813200e-01 -3.65348279e-01 1.14211798e+00 8.12621891e-01
4.95167136e-01 -1.62096214e+00 -7.24085212e-01 3.85007769e-01
-2.08073586e-01 -1.00098050e+00 -6.71643913e-01 5.92324495e-01
-6.81045711e-01 -9.35911357e-01 -1.99664950e-01 -1.06533778e+00
7.20994771e-01 -3.24327350e-01 1.22101450e+00 2.44582921e-01
-4.31127727e-01 3.01908731e-01 -2.97410309e-01 -3.34460080e-01
-4.06538188e-01 2.53487408e-01 2.51705293e-02 -3.35152239e-01
7.32963324e-01 -5.33983052e-01 -2.48078793e-01 3.35292965e-01
-1.22784960e+00 -3.37988287e-01 7.27106869e-01 9.57953513e-01
6.85944796e-01 3.48392338e-01 7.63484299e-01 -1.16132939e+00
4.59390759e-01 -3.16422075e-01 -6.27476454e-01 5.37198484e-01
-4.52134907e-01 3.37004185e-01 1.04036248e+00 -6.96659625e-01
-1.18690026e+00 1.28870845e-01 -3.27648103e-01 -3.26211005e-01
-6.04112029e-01 8.16841662e-01 -6.19568884e-01 3.37426186e-01
8.93260419e-01 3.98146510e-01 -8.32066238e-01 -5.76424241e-01
5.03428757e-01 3.58389646e-01 8.66339087e-01 -1.18870771e+00
1.00629079e+00 -2.34632060e-01 -3.75904590e-01 -4.63135213e-01
-6.41232610e-01 -3.81221622e-01 -5.64152539e-01 1.25393853e-01
7.50064135e-01 -6.61260307e-01 -6.95892930e-01 2.29660392e-01
-9.86554861e-01 -4.56698775e-01 -7.45050967e-01 3.74135286e-01
-2.31829017e-01 4.77625370e-01 -6.52794123e-01 -6.64547265e-01
-4.19570923e-01 -1.03207207e+00 7.65772104e-01 8.17672491e-01
-1.11353554e-01 -8.21231961e-01 -2.05726728e-01 1.50697865e-03
9.48504359e-02 1.04463898e-01 1.17033768e+00 -1.06897426e+00
-3.96023661e-01 -2.23758131e-01 -2.10181568e-02 3.77198696e-01
1.23734385e-01 -2.41871640e-01 -6.38647377e-01 1.06704943e-01
-3.13036412e-01 -1.98443264e-01 2.57066280e-01 -4.00673330e-01
8.60567749e-01 -3.71024817e-01 -6.03616714e-01 4.15187925e-01
1.33455563e+00 4.22436357e-01 7.84265876e-01 4.22350466e-01
5.80313981e-01 1.91450343e-01 6.69451475e-01 4.42507684e-01
5.32240629e-01 2.63897270e-01 7.84498872e-04 2.99079299e-01
3.26502100e-02 -3.92455250e-01 5.27014017e-01 7.65515924e-01
9.29964986e-03 -5.23589589e-02 -9.46323037e-01 6.83801889e-01
-1.47615433e+00 -9.47772205e-01 1.74251888e-02 2.27081919e+00
1.24786353e+00 3.36616009e-01 -3.46350849e-01 -2.45480925e-01
8.74512136e-01 -2.79450834e-01 -5.21735907e-01 -1.97363839e-01
-9.15974453e-02 1.10827947e+00 1.79389507e-01 2.61196941e-01
-1.04336143e+00 1.30966854e+00 4.39396715e+00 1.30321395e+00
-8.28905821e-01 6.88547790e-02 -5.54070026e-02 3.13729286e-01
-1.70182884e-01 2.51629382e-01 -1.48050869e+00 7.12904036e-01
1.01718628e+00 -2.90228724e-01 3.88941526e-01 9.61976826e-01
-3.87733161e-01 2.98960432e-02 -1.29254842e+00 7.59521782e-01
-2.03768294e-02 -1.14532268e+00 6.91086724e-02 -2.83363253e-01
5.74160397e-01 -3.41812074e-01 -2.80220956e-01 1.34691608e+00
7.16251671e-01 -7.87117839e-01 6.30165994e-01 3.40395659e-01
1.02591562e+00 -7.34191895e-01 5.71084261e-01 6.96599543e-01
-1.84559894e+00 -1.65050328e-01 -1.58266112e-01 3.46178949e-01
7.99981132e-02 3.00887287e-01 -7.65215635e-01 8.62889707e-01
7.90344954e-01 6.43790215e-02 -4.56394970e-01 8.35362971e-01
-6.41531408e-01 5.03796816e-01 -5.14504075e-01 -1.99336648e-01
1.70903839e-03 1.20105743e-01 4.32791144e-01 1.56054509e+00
4.14340019e-01 5.83426833e-01 3.57998699e-01 1.00048649e+00
-4.63154256e-01 2.57522374e-01 -2.65223652e-01 -2.03877732e-01
7.82626688e-01 1.51138389e+00 -5.75035274e-01 -7.62846887e-01
-2.62420744e-01 9.14174795e-01 6.08081639e-01 3.34698290e-01
-1.06843972e+00 -1.25390744e+00 5.04350662e-01 -4.87877987e-02
5.17413914e-01 -1.15272932e-01 1.13596983e-01 -1.39043736e+00
4.93747443e-02 -8.59144807e-01 7.53911555e-01 -6.80444181e-01
-1.29098403e+00 6.30633056e-01 4.22271639e-02 -1.02891564e+00
-1.43523395e-01 -4.15014237e-01 -7.93661237e-01 8.47415447e-01
-1.25435388e+00 -9.66771662e-01 -3.21043283e-01 7.21389651e-01
4.51386511e-01 -2.09134161e-01 1.14296532e+00 5.77754080e-01
-6.48547947e-01 1.10383952e+00 -1.16023839e-01 6.38126850e-01
7.47964621e-01 -1.40831220e+00 3.39876473e-01 8.30379128e-01
9.62569341e-02 1.12906301e+00 3.44774514e-01 -8.82831812e-01
-1.20490015e+00 -1.22291791e+00 1.06433392e+00 -6.61608517e-01
6.99644506e-01 -2.94959098e-01 -1.08047032e+00 9.87343729e-01
9.94052924e-03 7.27152377e-02 1.06145775e+00 2.51755193e-02
-5.00277400e-01 7.26785436e-02 -1.28434229e+00 6.60700560e-01
1.04374325e+00 -5.10686755e-01 -1.23397446e+00 -1.36218026e-01
7.89672673e-01 -5.38298130e-01 -1.26135015e+00 5.62330961e-01
2.57516950e-01 -5.73802948e-01 8.92514944e-01 -7.31654465e-01
1.15587994e-01 -4.10395235e-01 -1.53820038e-01 -1.22929001e+00
-3.41658480e-02 -4.59866732e-01 -3.49414378e-01 1.69018054e+00
6.81083858e-01 -6.39651477e-01 5.32578170e-01 8.50120544e-01
-5.34929276e-01 -4.70055193e-01 -8.73376489e-01 -1.04399061e+00
1.86083034e-01 -3.76458585e-01 9.82030451e-01 1.13909328e+00
1.80660725e-01 4.28043216e-01 1.01679370e-01 2.98554987e-01
2.77206779e-01 -1.30981831e-02 7.39785016e-01 -1.10488307e+00
-4.79164422e-01 -3.66006315e-01 -1.30158037e-01 -1.10525393e+00
2.56142586e-01 -1.02022886e+00 2.26161927e-01 -1.34576213e+00
2.76097804e-02 -9.57542539e-01 -5.72352290e-01 9.08225298e-01
-5.52608967e-01 -3.26487064e-01 1.35312960e-01 -1.33672237e-01
-7.86328733e-01 2.29365587e-01 7.52817154e-01 -1.28120929e-01
-3.53470087e-01 -3.57714981e-01 -5.08186936e-01 8.20643902e-01
3.74952197e-01 -8.62535179e-01 -7.01973960e-02 -3.03454936e-01
2.33654484e-01 1.28798291e-01 2.12116912e-02 -1.07453346e+00
5.27916253e-01 -2.40418389e-01 3.34638953e-01 -5.31668007e-01
3.41843545e-01 -6.24804258e-01 7.11137876e-02 1.19020410e-01
-7.16937855e-02 7.94798061e-02 3.45948726e-01 3.70407701e-01
-1.39798030e-01 -8.86839807e-01 4.18684214e-01 -3.67717803e-01
-1.10584271e+00 1.43200085e-01 -1.27856553e-01 4.17147666e-01
9.81002271e-01 -1.67493969e-01 -4.36050326e-01 3.68142039e-01
-9.26819265e-01 2.46530488e-01 1.88751519e-01 3.38734537e-01
4.43109035e-01 -1.20817161e+00 -2.73412943e-01 8.99790898e-02
3.37865859e-01 2.93230832e-01 3.65044355e-01 4.13030565e-01
-1.26340151e-01 -8.28581974e-02 2.38531940e-02 -2.35855967e-01
-9.45735455e-01 5.88078678e-01 2.33122557e-01 -5.52209258e-01
-2.89003789e-01 9.03511941e-01 1.23640060e-01 -1.09413803e+00
2.26531938e-01 -4.12590116e-01 -1.78240016e-01 -1.44022480e-01
5.76482236e-01 3.53937119e-01 8.75483304e-02 -4.60922807e-01
-3.98547500e-01 2.58705020e-01 -2.20707078e-02 2.13198345e-02
1.14702642e+00 2.69599736e-01 4.89439219e-02 3.86261791e-01
1.03880012e+00 4.15680349e-01 -6.90503478e-01 -1.41488090e-01
4.21966076e-01 -1.40801370e-01 -6.58861876e-01 -1.11525071e+00
-7.66458154e-01 6.21412635e-01 5.34362614e-01 -9.22014751e-03
1.03252411e+00 -4.84245308e-02 1.18475080e+00 8.24761629e-01
6.89891875e-01 -1.26419842e+00 5.46963178e-02 8.79339516e-01
1.26324534e-01 -9.43038106e-01 -3.50441694e-01 -2.56335288e-01
-3.82449418e-01 1.07885337e+00 1.06791973e+00 6.90591410e-02
4.79734451e-01 4.00941819e-01 -3.50904077e-01 -5.99998087e-02
-2.23610178e-01 -3.59972775e-01 1.14119418e-01 3.73557776e-01
4.07878965e-01 1.13899522e-01 -3.46887976e-01 1.46247375e+00
-2.32870206e-01 -4.40511554e-02 3.72238368e-01 1.49808681e+00
-6.65499687e-01 -1.43327260e+00 -2.70021856e-01 4.67877805e-01
-5.04445612e-01 -5.65639019e-01 -9.14597213e-02 8.22316110e-01
8.36542666e-01 7.48131692e-01 -1.86343402e-01 -4.33752149e-01
8.50974262e-01 6.92940235e-01 2.45057955e-01 -1.04563868e+00
-1.24992776e+00 -1.85750395e-01 1.57042176e-01 -9.98125598e-02
2.19798330e-02 -4.10966039e-01 -2.22551346e+00 1.05421552e-02
-5.56715786e-01 4.97094512e-01 3.95369589e-01 1.09446573e+00
4.73776340e-01 5.93930840e-01 3.30830336e-01 -2.22116545e-01
-4.51213479e-01 -8.55570197e-01 -3.26440930e-01 5.30815363e-01
-2.25016534e-01 -8.33343506e-01 -2.78857380e-01 1.98417962e-01] | [9.610909461975098, 9.418034553527832] |
07c66a4a-88a8-4267-ae4f-69ac2d9f8204 | sap-ri-twitter-sentiment-analysis-in-two-days | null | null | https://aclanthology.org/S14-2091 | https://aclanthology.org/S14-2091.pdf | SAP-RI: Twitter Sentiment Analysis in Two Days | null | ['N', 'Nishta Malhotra', 'Daniel Dahlmeier', 'Akriti Vij', 'Naveen an'] | 2014-08-01 | null | null | null | semeval-2014-8 | ['twitter-sentiment-analysis'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.426507472991943, 3.6914055347442627] |
f4e44e8f-1b9f-48cb-9dbb-0484b28bb8e0 | l1bsr-exploiting-detector-overlap-for-self | 2304.06871 | null | https://arxiv.org/abs/2304.06871v2 | https://arxiv.org/pdf/2304.06871v2.pdf | L1BSR: Exploiting Detector Overlap for Self-Supervised Single-Image Super-Resolution of Sentinel-2 L1B Imagery | High-resolution satellite imagery is a key element for many Earth monitoring applications. Satellites such as Sentinel-2 feature characteristics that are favorable for super-resolution algorithms such as aliasing and band-misalignment. Unfortunately the lack of reliable high-resolution (HR) ground truth limits the application of deep learning methods to this task. In this work we propose L1BSR, a deep learning-based method for single-image super-resolution and band alignment of Sentinel-2 L1B 10m bands. The method is trained with self-supervision directly on real L1B data by leveraging overlapping areas in L1B images produced by adjacent CMOS detectors, thus not requiring HR ground truth. Our self-supervised loss is designed to enforce the super-resolved output image to have all the bands correctly aligned. This is achieved via a novel cross-spectral registration network (CSR) which computes an optical flow between images of different spectral bands. The CSR network is also trained with self-supervision using an Anchor-Consistency loss, which we also introduce in this work. We demonstrate the performance of the proposed approach on synthetic and real L1B data, where we show that it obtains comparable results to supervised methods. | ['Gabriele Facciolo', 'Pablo Arias', 'Axel Davy', 'Jérémy Anger', 'Ngoc Long Nguyen'] | 2023-04-14 | null | null | null | null | ['image-super-resolution'] | ['computer-vision'] | [ 0.564355 -0.0841311 0.01462904 -0.6330556 -1.1841174 -0.52913815
0.5085201 -0.20496042 -0.45690164 0.9171426 -0.07367761 0.03661309
-0.3576116 -0.96834254 -0.86987007 -0.88887316 -0.4343573 -0.05391629
-0.10997073 -0.48050356 -0.24897654 0.833018 -1.4054841 0.3222974
0.8906466 1.1235878 0.13070744 0.47338003 0.4765972 0.44018143
-0.10464366 0.21502222 0.9293893 -0.442249 -0.7832988 0.2294009
1.054323 -0.43928516 -0.08616997 1.4133124 0.4703269 0.13531822
0.39383376 -0.71973926 -0.19139527 0.18721862 -0.95806384 0.4059995
0.05440786 0.00862129 1.1113724 -0.5217622 0.49225733 0.86643475
1.2131239 0.13852088 -1.923754 -0.6250797 -0.32963014 -0.15391263
-1.6672381 -0.58085275 0.593762 -0.5824465 0.59243387 0.27360472
0.46868303 0.5897632 0.0874013 0.01862137 1.5727519 -0.3591596
-0.06332013 0.01867204 -0.15624776 0.28176114 0.2871054 0.53184116
-0.53747445 0.12646282 0.94204736 -0.21272792 -0.77065367 -0.3667762
-1.07017 0.797041 1.0737269 0.28369612 -0.5240987 -0.18637973
-0.08767252 0.3294197 0.80339617 0.46157566 -0.1719474 0.8072728
-1.4430457 0.0344221 0.05776383 0.5910481 1.1171389 0.14882179
0.18291284 0.67713684 0.34985986 0.8964886 0.27371144 -0.8973422
0.23478715 0.26791692 0.50081277 -1.1329169 -0.88557327 -0.9190354
-1.2983214 0.42715013 0.01837542 0.00903254 -0.738562 1.7338547
0.34299535 0.17301187 0.41943 1.3321109 0.5556456 0.779377
-0.38744673 -0.36657614 1.011787 -0.54482573 -0.57233906 -0.46714368
0.37700793 -0.5821733 0.72393584 0.0788265 -0.66213053 -0.6543837
-1.2824211 0.07270584 -0.19217382 0.1359475 0.2756581 0.47003916
-1.1969684 0.72664297 -0.67249966 -0.45407528 0.3692937 0.02683651
-0.5174806 0.14885752 -1.3574518 0.7202528 0.51618457 0.4797306
-0.6388454 -0.758594 -0.8985457 -0.15783273 -0.18268767 -0.36440703
0.67580587 -1.2419941 -1.0732661 1.3864459 0.22968914 -1.0045878
0.40535352 -0.12447486 -0.65488416 0.2956323 0.3736703 0.5222093
0.98616296 -1.4074929 -0.7281252 -0.5926286 -0.23638564 0.2011862
-0.00698402 -0.23380359 0.15821724 -0.5134924 0.5216869 -0.8430362
-0.21958226 0.06773372 -0.27159038 0.7435025 0.4236601 -0.794108
0.5529374 -2.1915662 -0.14008388 0.2421965 -0.09126974 0.1085149
-0.35313946 0.12635036 -0.32232064 -0.21744291 -0.61798674 -0.0976733
-0.43107668 -0.04241995 -0.4653334 1.1855059 0.18877974 0.47068605
-0.6968964 -0.19824578 0.2413587 0.6779944 -0.16275741 0.20956515
0.02786724 0.8655741 0.04321065 0.35914603 1.3289155 -0.23300183
0.14851403 -0.839128 -0.5871183 0.13183 -1.2666665 1.6786711
-0.5483102 0.83400416 0.45749477 -1.0319633 1.1841091 0.07544488
0.4719445 -1.1243484 -0.37108344 0.29401425 -0.23896907 -0.24388042
0.42300838 -0.49167648 0.18724744 0.11510842 -0.25078014 -0.44168383
-0.3522081 -0.19376062 0.42394215 0.01568002 0.43408322 -0.55722296
0.6687444 0.17616947 0.7211468 0.8396164 -0.01807385 0.67154855
-0.11664401 -0.55019546 -1.3796655 -0.7894578 -0.66340744 0.5464063
0.3224985 0.32008734 -0.3737889 -0.12897746 -0.02349047 0.29330263
-0.508966 0.10021956 -0.24430496 -1.2074087 0.4005966 0.10016378
1.0650953 -0.40535858 -0.61839736 0.19773425 -0.3155954 -1.4433981
0.12429681 0.01336818 -1.065406 -0.9694323 -0.37850472 -0.38352016
0.43836656 0.61928797 1.0326794 -0.2766556 -0.3507611 0.05395894
-0.27933785 0.18446293 -0.30692893 -0.05759446 0.03936381 0.53309
-0.09115387 -0.6417114 -0.70863044 0.3508398 -0.9801177 0.25297326
0.52973187 1.0949318 0.9712487 0.28317726 0.20541932 -0.73165095
-0.09884099 -0.09982735 -1.2857906 -0.02859572 -0.5212204 -0.11508563
0.3854083 0.23659435 -1.1257796 0.31402543 -0.04902442 -0.03721768
-0.14766827 0.6429018 -0.06239267 -0.63692415 1.1016492 0.3788459
-0.03867311 -0.58778137 0.08763007 0.9312379 0.9240226 -0.02965772
1.1964078 1.2624321 0.3027296 -1.2471421 -1.5010817 -0.71159744
-0.63675153 -0.16441073 0.84455025 -1.4763361 -0.20176646 0.4969387
-0.779373 -0.44202432 -0.12931159 0.59149456 -0.53533196 0.2886288
-0.34366456 -0.84356177 -0.49072793 -0.79442036 1.2501042 0.15798947
0.410832 -0.7737644 0.33542642 0.50479186 0.60619843 0.64872074
0.24558581 -0.08143524 -0.6931716 -0.10961618 -0.6527274 0.5853967
0.2989578 -0.52827024 -1.2444216 -0.89967966 0.20141108 -0.3356581
1.0691842 0.47088662 0.74298793 -0.19731057 -0.03312442 1.4518402
2.1118762 -0.2971113 0.77500075 0.9036752 0.65949464 0.6673561
0.7612834 0.27686468 0.07844086 0.99942887 0.76749927 -0.84252363
0.02630476 0.29038167 0.09918416 -0.06100927 -0.15964901 0.3059038
-0.91771305 0.574166 -1.6881881 -1.1051396 -0.4147314 2.4572937
0.74564296 -0.19786766 -0.35445744 -0.03306334 0.71890414 0.5943341
-0.59729964 0.35525665 -0.7895174 0.14518476 1.2340736 0.89822304
-1.6319817 0.7077777 5.5321813 0.33071232 -1.4627008 0.21126226
0.52796406 0.11952504 -0.08442224 -0.19109762 -0.76325446 0.09973017
0.6803182 0.14715105 0.35780206 0.41687438 0.6037938 -0.20305467
-0.65613455 1.1764213 0.08932757 -1.4524932 -0.16595139 -0.04293679
1.0747038 0.45876613 -0.07076868 -0.44263598 -0.01581545 -0.9604861
0.5106854 0.5073176 1.0626739 -0.7256211 0.7411715 0.10796674
-1.1997432 -0.04407521 -0.4302136 0.12636144 0.04163369 0.8905169
-0.47435513 1.0172894 1.0985256 0.9787924 -0.3367114 0.92740244
-0.1886503 0.17460264 -0.31042805 0.9729802 0.2147073 -0.53725064
0.67944765 1.1012475 0.38225663 0.10179114 0.214286 0.9193824
0.08913562 -0.07039987 -0.68567866 0.27962893 0.26112425 1.3721986
-0.40533397 -0.02422627 -0.38636944 0.94546884 -0.10591743 0.42465773
-0.61651415 -0.17054166 0.9316563 0.3236439 0.18889587 -0.21977748
-0.05185132 -1.0810335 -0.02634051 -0.84906137 0.32919288 -0.9972706
-1.2543749 0.79745644 -0.22400633 -1.6541533 0.10506348 -0.42511085
-0.03123799 1.2344667 -2.3144808 -1.3605452 -0.8514571 0.4505185
-0.0961566 -0.01322828 0.6288123 0.379083 -0.28724155 0.1170882
0.4871636 0.10136259 0.82732207 -0.98305976 0.2803522 1.2672346
-0.12676443 0.09365053 0.9716904 -0.397193 -0.997736 -1.5115334
0.48313266 0.3024141 0.5466515 -0.12062728 -1.0376555 0.7563578
-0.01458921 0.43440384 0.54987496 -0.37247074 -0.31607175 -0.81498724
-1.2167989 -0.13838245 0.72100717 -0.7597827 -0.30638608 0.6693465
0.5237269 -0.5011717 -0.9310912 0.71383274 0.29547992 -1.420754
1.254956 -0.21116629 0.20385562 -0.7702329 -0.38994026 -1.3324097
-0.37533146 -0.34315938 0.6160059 0.89083725 0.27079245 -0.8117747
0.47921267 -0.02244626 0.09287577 0.2107729 -1.0131326 -1.0468798
-0.20960255 -0.10914776 0.5507801 1.4230525 -0.6884018 0.13123119
-0.5588319 1.1536529 1.3574072 0.7729348 0.61358696 -1.4085977
-0.10792848 -0.07371524 -0.42400223 -0.877267 0.04413131 -0.82531273
0.2386607 -1.3237671 0.06271766 -0.5314088 -0.15477145 0.28752536
0.25886926 0.80471855 -0.19300537 0.51472396 -0.20833622 0.5666073
0.9213244 -0.23128757 -0.19681384 -0.1329458 -0.11269282 0.5434513
0.7814696 -0.38365373 0.08590015 -0.5721297 0.4022571 0.17936082
0.7669873 -1.2699482 0.06743211 -0.0203133 0.5201418 -0.50294644
0.18699601 -0.96945375 0.46937 0.3410334 -0.28340515 -0.51809824
0.2004456 0.4304099 -0.47244707 0.04547464 1.4257277 -0.05094147
-0.7967916 0.32236084 0.0845885 -0.14208323 0.60675013 -0.09392493
-0.26345322 -0.24939194 -0.5005377 0.15990944 0.79648745 -0.10279468
0.3730699 -1.2795337 -0.97848845 0.42025095 0.2968536 0.22092818
0.41983765 0.9452007 -0.79871774 0.22285044 -0.47087058 -0.9321508
-1.1531328 0.16784878 1.114085 -0.14789787 -0.8458143 0.68809575
0.07100702 -0.5696603 -0.17640306 -0.09235144 -0.17232248 0.1266696
0.712578 -0.02437275 0.30457062 -0.96947074 -0.37293813 0.813457
0.33415958 -0.12650079 1.6631218 -0.4547653 -0.44205827 0.26830456
1.1774733 -0.11832878 -1.4472278 -0.5172173 -0.08801445 -0.7344172
0.8027532 -0.46284482 -1.3035456 0.6877575 1.1650004 0.17752461
1.5408128 -0.34450296 0.5340743 0.3801475 0.27499157 -0.97814107
-0.2831797 0.35145426 0.8176733 -1.6396124 0.20017612 -0.2063679
-0.12104016 0.998378 0.17625415 -0.04821714 0.364308 0.02218319
0.29625365 -0.22776514 -0.00899139 -0.48313862 0.20732148 0.5460463
0.30963925 0.1368543 0.05053092 0.02309999 -0.04368481 0.01103431
0.60364044 0.64574474 -0.49203256 -0.5805329 -0.8286859 0.15625258
-0.08298644 -0.23243922 0.14796036 0.60317373 -0.03458069 0.54999524
0.36679637 -0.07017357 0.26074594 -0.38563147 0.36738816 -0.29680416
-0.08456315 0.05193222 -0.03997527 -0.7380056 -1.0013359 -0.5597986
-0.83944523 -0.05421484 -0.15547028 0.01121628 0.6942869 0.6547007
0.2672062 0.17618309 1.0648276 -1.2209939 -0.4610587 -0.8009185
-1.0712031 0.41001353 1.0693513 -0.46174473 -0.5821892 0.1848586 ] | [9.938081741333008, -1.800282597541809] |
60f543a2-f880-4298-8e30-0850f922ab6a | enhancing-cross-lingual-natural-language-1 | 2305.12761 | null | https://arxiv.org/abs/2305.12761v1 | https://arxiv.org/pdf/2305.12761v1.pdf | Enhancing Cross-lingual Natural Language Inference by Soft Prompting with Multilingual Verbalizer | Cross-lingual natural language inference is a fundamental problem in cross-lingual language understanding. Many recent works have used prompt learning to address the lack of annotated parallel corpora in XNLI. However, these methods adopt discrete prompting by simply translating the templates to the target language and need external expert knowledge to design the templates. Besides, discrete prompts of human-designed template words are not trainable vectors and can not be migrated to target languages in the inference stage flexibly. In this paper, we propose a novel Soft prompt learning framework with the Multilingual Verbalizer (SoftMV) for XNLI. SoftMV first constructs cloze-style question with soft prompts for the input sample. Then we leverage bilingual dictionaries to generate an augmented multilingual question for the original question. SoftMV adopts a multilingual verbalizer to align the representations of original and augmented multilingual questions into the same semantic space with consistency regularization. Experimental results on XNLI demonstrate that SoftMV can achieve state-of-the-art performance and significantly outperform the previous methods under the few-shot and full-shot cross-lingual transfer settings. | ['Lijie Wen', 'Philip S. Yu', 'Fukun Ma', 'Yawen Yang', 'Aiwei Liu', 'Xuming Hu', 'Shuang Li'] | 2023-05-22 | null | null | null | null | ['cross-lingual-natural-language-inference', 'cross-lingual-transfer'] | ['natural-language-processing', 'natural-language-processing'] | [-3.79654318e-02 1.03415614e-02 -4.18423921e-01 -7.59132326e-01
-1.29397011e+00 -8.57974231e-01 5.87631345e-01 -4.70071763e-01
-7.53447890e-01 6.58062994e-01 4.89896357e-01 -4.03262794e-01
1.03512608e-01 -5.27870536e-01 -8.45843732e-01 -3.74003023e-01
9.42329764e-01 8.18562806e-01 -7.70532861e-02 -2.41464600e-01
-2.76905239e-01 -3.73875976e-01 -1.01172423e+00 4.61212069e-01
1.35601568e+00 5.98230898e-01 8.55574965e-01 1.76734030e-01
-7.69116282e-01 6.10236883e-01 -3.16658646e-01 -6.65476322e-01
1.55673668e-01 -5.41425586e-01 -8.01370203e-01 -1.71828896e-01
7.17721462e-01 -1.77957281e-01 -2.89338648e-01 1.16699350e+00
4.23888683e-01 2.86384672e-01 5.23597658e-01 -1.02758646e+00
-1.29617453e+00 1.11307704e+00 -1.51642233e-01 5.12258857e-02
5.35514712e-01 2.28163600e-01 1.26150191e+00 -1.38333058e+00
8.29606354e-01 1.62029493e+00 3.48079175e-01 7.37250865e-01
-1.15760279e+00 -9.87151980e-01 5.01639247e-01 7.65921891e-01
-1.23315251e+00 -5.68113148e-01 7.47003794e-01 -1.56064123e-01
9.85214829e-01 3.95493023e-02 4.27922495e-02 1.44090807e+00
-1.56529352e-01 9.48574722e-01 1.23693871e+00 -5.96186697e-01
-3.17924395e-02 3.10914874e-01 3.86750132e-01 6.99816227e-01
-2.07907364e-01 6.26549795e-02 -4.88813072e-01 1.22566029e-01
4.95926917e-01 -2.09352285e-01 -3.18322003e-01 -4.03237194e-02
-1.35252309e+00 1.04523551e+00 1.93279207e-01 1.91751182e-01
-1.85418501e-01 -2.86668181e-01 5.71781635e-01 5.92343807e-01
1.22500218e-01 3.62814486e-01 -7.72627234e-01 2.30212376e-01
-5.50457001e-01 8.65141861e-03 7.02203751e-01 1.51424944e+00
9.37097251e-01 1.35553047e-01 -3.92031014e-01 1.08980751e+00
3.33553255e-01 8.28879654e-01 6.91713333e-01 -7.01642692e-01
9.18647647e-01 5.21102428e-01 -1.82349548e-01 -4.14547563e-01
5.91219869e-03 -2.54846692e-01 -5.91729939e-01 -6.19298160e-01
3.05687785e-01 -2.05337808e-01 -7.41110861e-01 1.96334803e+00
4.15300548e-01 1.44863844e-01 4.90875721e-01 8.13908219e-01
1.12905502e+00 1.15141916e+00 3.33184779e-01 -2.67363310e-01
1.56060469e+00 -1.28831112e+00 -1.04281652e+00 -7.12970197e-01
6.03721023e-01 -9.63458836e-01 1.91257048e+00 1.63136050e-01
-7.94161141e-01 -8.47828984e-01 -8.38254929e-01 -6.22497141e-01
-3.83049965e-01 2.43861362e-01 2.88257569e-01 2.17275307e-01
-4.23637092e-01 -1.27719000e-01 -4.27476138e-01 -3.85596961e-01
-5.78610189e-02 -1.04793645e-01 -2.68745393e-01 -7.20951319e-01
-1.65918636e+00 1.06747103e+00 4.18416917e-01 -6.34985864e-02
-8.61432612e-01 -8.37033749e-01 -1.34287786e+00 -5.99894449e-02
7.35747933e-01 -4.37073350e-01 1.45772672e+00 -8.30952883e-01
-1.51699579e+00 1.00966430e+00 -2.96770841e-01 -1.58774704e-02
1.71269000e-01 -3.08587134e-01 -4.46402997e-01 -3.33238989e-01
5.38075030e-01 5.73176146e-01 5.45216382e-01 -1.03372920e+00
-6.27785206e-01 -2.23885015e-01 6.26256689e-02 3.72888088e-01
-1.63594171e-01 1.89597398e-01 -7.49062777e-01 -7.77990818e-01
-1.32414803e-01 -7.99762964e-01 -8.42042174e-03 -4.15396154e-01
-8.61067623e-02 -9.29567814e-01 5.62218785e-01 -7.27596462e-01
1.03033996e+00 -1.75749695e+00 3.66623938e-01 -2.84437865e-01
-3.48575652e-01 4.61995035e-01 -5.90111494e-01 3.18395555e-01
3.21961612e-01 -2.72498280e-01 -9.34784710e-02 -1.96496725e-01
3.23361933e-01 8.22973907e-01 -5.39373755e-01 2.17089564e-01
-1.11774085e-02 1.14245820e+00 -1.22499156e+00 -6.87284112e-01
1.77910835e-01 2.89411694e-01 -6.20677590e-01 8.64226937e-01
-4.79343563e-01 3.79361153e-01 -2.66298443e-01 3.89906049e-01
4.30366278e-01 -1.95590090e-02 4.48361188e-01 -5.89058578e-01
7.24259438e-03 6.94828391e-01 -8.81024957e-01 2.21831751e+00
-9.85109508e-01 1.93541586e-01 3.65264178e-03 -9.02467370e-01
9.66962397e-01 4.64561552e-01 -1.17414325e-01 -9.38887835e-01
4.31282260e-02 3.29534501e-01 -6.81737810e-02 -8.48521948e-01
2.86163390e-01 -4.30081487e-01 -5.14970243e-01 6.20751798e-01
6.20623708e-01 -8.79484266e-02 8.89916793e-02 2.15619057e-02
4.56703156e-01 4.58832920e-01 3.72022361e-01 -2.03245759e-01
7.30450571e-01 -9.98091623e-02 9.56705093e-01 5.41772246e-01
-5.36475629e-02 1.15408927e-01 -6.53854907e-02 -1.42558515e-01
-7.11709917e-01 -1.40811086e+00 9.47197676e-02 1.76250899e+00
1.21335037e-01 -3.47357333e-01 -7.06367850e-01 -1.01512516e+00
-1.12576820e-01 1.20729029e+00 -3.49372655e-01 -1.41033446e-02
-8.93305302e-01 -1.59816682e-01 4.60719079e-01 5.30724227e-01
2.34952167e-01 -1.30114365e+00 -4.50333692e-02 5.55173576e-01
-6.33021832e-01 -1.57110906e+00 -1.08434474e+00 1.91867813e-01
-2.45550200e-01 -9.00375843e-01 -4.18550074e-01 -1.41175473e+00
7.01011717e-01 6.47742525e-02 1.16888690e+00 -3.73879910e-01
-2.10243482e-02 2.26545736e-01 -3.75374943e-01 -6.48936480e-02
-4.65329826e-01 2.07975477e-01 3.04091007e-01 4.46119718e-02
8.04851413e-01 -1.19449832e-01 -8.96973610e-02 3.32542390e-01
-7.68560410e-01 1.27141908e-01 4.63009387e-01 1.05224812e+00
8.42658758e-01 -5.95293999e-01 7.56281912e-01 -1.17724466e+00
6.61826313e-01 -5.57470322e-01 -6.85300469e-01 7.27454782e-01
-3.61181140e-01 4.60416138e-01 1.13602042e+00 -7.07372725e-01
-1.45274329e+00 -1.78833708e-01 -2.24216983e-01 -4.82056588e-01
-1.50290206e-01 6.96067035e-01 -6.15664065e-01 3.54599804e-01
3.83706182e-01 2.78495550e-01 -4.19650286e-01 -8.35257471e-01
9.27202106e-01 8.15031230e-01 8.36484909e-01 -1.09834957e+00
6.85787201e-01 -4.74320389e-02 -8.77206624e-01 -6.37261927e-01
-1.45345521e+00 -4.24006552e-01 -6.21917069e-01 -4.63109650e-02
9.19368684e-01 -9.03162837e-01 -3.93931448e-01 2.19479829e-01
-1.45331013e+00 -4.84452307e-01 -5.80906272e-02 4.55062926e-01
-4.24448818e-01 1.42428696e-01 -6.43580437e-01 -1.73807040e-01
-5.13581395e-01 -1.26432753e+00 7.79346049e-01 1.57757029e-01
-2.41539747e-01 -1.12649226e+00 1.60396665e-01 5.52955747e-01
3.77778709e-01 -6.61247134e-01 1.27082205e+00 -8.17940414e-01
-4.61354017e-01 2.16725677e-01 -2.95845568e-01 4.86139268e-01
3.49738568e-01 -6.04244590e-01 -7.77564049e-01 -2.56861538e-01
1.98655471e-01 -7.84174740e-01 3.04524064e-01 -1.11909926e-01
7.85892069e-01 -6.04752719e-01 -1.26229286e-01 8.08332086e-01
1.43184388e+00 -6.25145212e-02 1.42404661e-01 -1.33767771e-02
1.04762995e+00 7.54482269e-01 8.70512187e-01 -1.80241987e-01
1.07879913e+00 7.27064788e-01 -1.88371450e-01 1.57033399e-01
-3.87494206e-01 -6.17213547e-01 6.48461938e-01 1.90047944e+00
5.73941588e-01 1.52469566e-02 -9.30278420e-01 6.18231237e-01
-1.76362205e+00 -6.60882950e-01 1.48904368e-01 1.89949000e+00
1.38286185e+00 -1.61967039e-01 -6.12152100e-01 -6.92829132e-01
5.62654614e-01 -3.25533599e-02 -7.05804765e-01 -2.59912401e-01
-1.02553427e-01 2.91129887e-01 2.72606134e-01 8.96799564e-01
-6.57270551e-01 1.67665899e+00 5.11120176e+00 8.82333040e-01
-1.06831956e+00 6.97425604e-01 -1.38670765e-02 4.72147465e-02
-7.73175418e-01 1.14684910e-01 -1.26670825e+00 3.63230914e-01
8.09477746e-01 -5.22349417e-01 5.51083028e-01 7.66318083e-01
-3.28766508e-03 3.91691417e-01 -1.33998775e+00 1.14988399e+00
4.66370493e-01 -1.01707160e+00 2.60046333e-01 -5.21702111e-01
8.17541718e-01 2.57195741e-01 -2.64192402e-01 1.00440001e+00
5.88995814e-01 -8.21473658e-01 6.16080999e-01 3.94131988e-01
1.03139579e+00 -6.60390735e-01 4.49146181e-01 5.28157294e-01
-1.25245202e+00 2.29341760e-01 -6.70358598e-01 1.74151659e-01
3.83154422e-01 -1.23410322e-01 -6.88209236e-01 4.71065909e-01
3.39270443e-01 5.65603852e-01 -3.58846992e-01 4.30733830e-01
-7.23648846e-01 8.16146433e-01 -8.09203759e-02 -8.45503509e-02
4.64486867e-01 -3.44406933e-01 3.01080972e-01 1.26018119e+00
2.45527148e-01 3.36347036e-02 8.47309470e-01 8.30216169e-01
-3.09709042e-01 5.81389427e-01 -4.78289753e-01 4.62630577e-02
6.87390745e-01 1.08782482e+00 7.61304125e-02 -5.45710862e-01
-8.93726587e-01 1.07225609e+00 7.69630611e-01 5.02002239e-01
-7.15607464e-01 -3.17360610e-01 5.70290327e-01 -2.33549997e-01
2.11374536e-01 -1.67872638e-01 -5.86562902e-02 -1.52327442e+00
-3.15924436e-02 -1.36845541e+00 5.25411785e-01 -7.91904628e-01
-1.78718245e+00 7.86195338e-01 -3.52530181e-02 -9.39266503e-01
-4.34634089e-01 -5.85787952e-01 -5.14155447e-01 1.16750336e+00
-1.62383330e+00 -1.45103693e+00 -3.25497836e-02 8.24583888e-01
1.21437037e+00 -3.92636001e-01 9.88622010e-01 3.72893333e-01
-5.16378939e-01 9.02748227e-01 -1.45153239e-01 2.78791755e-01
1.16557562e+00 -1.08908165e+00 5.02543449e-01 9.11121547e-01
3.06704849e-01 7.54376888e-01 5.61695814e-01 -6.73182726e-01
-1.64142168e+00 -1.35475254e+00 1.28743148e+00 -4.73060042e-01
9.06242967e-01 -6.14800572e-01 -1.34055328e+00 1.03794193e+00
8.25062454e-01 -1.08768485e-01 8.50553691e-01 2.29989946e-01
-6.90456748e-01 -1.84952617e-01 -6.75464034e-01 8.93054307e-01
9.42477703e-01 -1.03646779e+00 -1.34498572e+00 3.63376498e-01
1.11415040e+00 -4.16209817e-01 -7.45293558e-01 1.72092915e-01
2.66363740e-01 -1.36615098e-01 6.76156104e-01 -7.78680503e-01
1.32623643e-01 -2.42628142e-01 -3.01922649e-01 -1.49260354e+00
-2.52600431e-01 -4.75432187e-01 1.79037839e-01 1.51514471e+00
4.89792347e-01 -4.72700775e-01 2.34074786e-01 3.14707100e-01
-2.67128229e-01 -5.95602393e-01 -9.04825151e-01 -9.28421676e-01
3.57226521e-01 -5.62860489e-01 5.96871853e-01 1.21708822e+00
6.35697097e-02 1.08414793e+00 -4.35834855e-01 4.77630906e-02
8.59635949e-01 2.43953198e-01 7.19819844e-01 -8.73703063e-01
-3.48289818e-01 -5.89904152e-02 3.90966058e-01 -1.39558852e+00
8.98517728e-01 -1.54406035e+00 4.73066896e-01 -1.44099891e+00
5.51797859e-02 -5.55817962e-01 -2.56261140e-01 6.93983316e-01
-6.79349005e-01 -1.91541851e-01 1.20592844e-02 -1.53622121e-01
-5.77765584e-01 7.90827096e-01 1.21772718e+00 -1.85983628e-01
7.70195425e-02 -5.38479447e-01 -6.80840790e-01 6.63454652e-01
6.12434626e-01 -5.71140766e-01 -6.12124741e-01 -1.01091588e+00
1.30797729e-01 -3.76926665e-03 -1.20632306e-01 -4.52588737e-01
4.18627441e-01 -4.43924576e-01 -1.18420094e-01 -5.75680673e-01
1.56023815e-01 -6.80326462e-01 -3.46216440e-01 5.07227592e-02
-6.12157881e-01 1.15373418e-01 8.06543306e-02 2.94888496e-01
-1.50187343e-01 -3.76353353e-01 6.55146599e-01 -2.70074636e-01
-1.00544393e+00 4.66127247e-01 -1.38605073e-01 7.78998911e-01
6.10560596e-01 3.69235456e-01 -3.60129654e-01 -1.47885188e-01
-4.42133278e-01 6.89479887e-01 2.08505988e-01 1.12795448e+00
5.23122489e-01 -1.68338668e+00 -9.39404726e-01 3.07801902e-01
5.33758044e-01 -9.25842449e-02 1.46512792e-01 4.65796411e-01
1.26603752e-01 5.88220179e-01 -6.16685227e-02 -4.64835286e-01
-1.03803301e+00 7.29730129e-01 1.32552192e-01 -2.02814877e-01
-3.84882033e-01 9.08717334e-01 5.14581382e-01 -1.32353830e+00
3.89356047e-01 -1.19928569e-01 -6.67503998e-02 1.06791824e-01
4.75292325e-01 -1.57038540e-01 -1.42765209e-01 -7.34395802e-01
-3.18574071e-01 5.72892129e-01 -5.28314769e-01 -1.72376275e-01
1.14743638e+00 -3.98385912e-01 -6.16217777e-02 7.11896181e-01
1.35446012e+00 1.04197323e-01 -1.00256419e+00 -9.90643740e-01
1.97282642e-01 -1.63041621e-01 -1.54721126e-01 -7.33566999e-01
-7.88221002e-01 9.32792246e-01 1.72666579e-01 -5.20621240e-01
5.70290685e-01 3.34130079e-01 1.17527330e+00 5.99061191e-01
3.58104587e-01 -1.32228816e+00 6.71244180e-03 9.68017578e-01
1.06336617e+00 -1.33026838e+00 -6.26661718e-01 -3.08136374e-01
-7.61601329e-01 7.69266725e-01 1.05169404e+00 1.02912903e-01
3.38053882e-01 1.43920928e-01 7.46972501e-01 6.06250092e-02
-1.03010845e+00 -1.96123883e-01 4.13757563e-01 5.80895305e-01
6.64064169e-01 2.35092670e-01 -3.45293015e-01 1.15271950e+00
-3.06801081e-01 -2.13474333e-01 1.08144678e-01 5.63685775e-01
-3.42419028e-01 -1.45907378e+00 -2.33368158e-01 1.91647142e-01
-5.56386821e-02 -5.92773557e-01 -4.03403565e-02 4.15725619e-01
2.45229602e-01 8.32735479e-01 -4.94410843e-02 -4.37086932e-02
4.86664414e-01 4.96808767e-01 6.73581362e-01 -1.01036334e+00
-5.24964392e-01 -5.22868298e-02 8.70888084e-02 -4.63703603e-01
-3.50908339e-02 -5.71976840e-01 -1.26543713e+00 2.10606903e-01
-1.81591257e-01 2.47746646e-01 4.10200685e-01 1.47479403e+00
5.26865982e-02 4.30256546e-01 4.60008055e-01 -2.74565905e-01
-7.50805736e-01 -1.06726301e+00 4.64422517e-02 5.41754782e-01
7.20179230e-02 -5.15510499e-01 -2.36463398e-01 2.32247785e-01] | [10.991217613220215, 9.492179870605469] |
7b362d7f-b4d7-451e-a322-ae520da8dbf5 | molecular-de-novo-design-through-deep | 1704.07555 | null | http://arxiv.org/abs/1704.07555v2 | http://arxiv.org/pdf/1704.07555v2.pdf | Molecular De Novo Design through Deep Reinforcement Learning | This work introduces a method to tune a sequence-based generative model for
molecular de novo design that through augmented episodic likelihood can learn
to generate structures with certain specified desirable properties. We
demonstrate how this model can execute a range of tasks such as generating
analogues to a query structure and generating compounds predicted to be active
against a biological target. As a proof of principle, the model is first
trained to generate molecules that do not contain sulphur. As a second example,
the model is trained to generate analogues to the drug Celecoxib, a technique
that could be used for scaffold hopping or library expansion starting from a
single molecule. Finally, when tuning the model towards generating compounds
predicted to be active against the dopamine receptor type 2, the model
generates structures of which more than 95% are predicted to be active,
including experimentally confirmed actives that have not been included in
either the generative model nor the activity prediction model. | ['Thomas Blaschke', 'Hongming Chen', 'Ola Engkvist', 'Marcus Olivecrona'] | 2017-04-25 | null | null | null | null | ['activity-prediction', 'activity-prediction'] | ['computer-vision', 'time-series'] | [ 9.18814361e-01 6.29260302e-01 -2.84145176e-01 -1.69624135e-01
-7.73419559e-01 -8.30976307e-01 7.86164582e-01 3.47629875e-01
-2.09821299e-01 1.70938432e+00 6.48171008e-02 -5.04285693e-01
2.11153805e-01 -8.97937059e-01 -9.18522418e-01 -8.57953608e-01
-1.13034256e-01 6.69169486e-01 6.79217950e-02 -1.91613793e-01
3.10161501e-01 7.88834095e-01 -1.41847050e+00 2.12540746e-01
9.46075618e-01 2.68507063e-01 4.04430926e-01 4.23533857e-01
2.63559997e-01 2.32025608e-01 -8.94097984e-01 -4.33436744e-02
-8.81989021e-03 -7.88191676e-01 -8.45300674e-01 -3.46106946e-01
2.11123690e-01 -1.38127252e-01 3.87311041e-01 5.85381031e-01
6.68543279e-01 2.11293146e-01 1.11713874e+00 -5.64408064e-01
-4.24952567e-01 3.30321282e-01 1.13757037e-01 -2.09573269e-01
7.21382201e-01 1.53363392e-01 8.81508648e-01 -9.47029531e-01
9.18735027e-01 1.15213847e+00 1.55492052e-01 8.85019302e-01
-1.80007064e+00 -7.88953900e-01 -2.01008812e-01 -1.90811113e-01
-1.16862428e+00 -5.33913255e-01 2.88068235e-01 -3.34495902e-01
1.23583496e+00 4.11240727e-01 8.48536253e-01 1.28276813e+00
6.59782231e-01 2.64270991e-01 9.69884157e-01 -2.64620364e-01
8.46307099e-01 2.19555330e-02 -5.16207516e-01 5.32324553e-01
3.54698896e-01 4.40577418e-01 -7.14412570e-01 -6.96947396e-01
4.94406313e-01 -4.00696188e-01 -1.87002942e-01 -3.93615633e-01
-9.01710093e-01 1.26342559e+00 3.41198653e-01 -1.80692207e-02
-5.31414092e-01 -3.37336287e-02 4.28337529e-02 -2.20947847e-01
6.10331111e-02 1.39811993e+00 -5.04658937e-01 2.57623523e-01
-7.74358749e-01 5.30844688e-01 1.00052094e+00 7.43119955e-01
8.55691254e-01 2.20609352e-01 -5.93645535e-02 3.14819127e-01
3.96674395e-01 2.07082301e-01 2.20008388e-01 -5.01682520e-01
-2.43814617e-01 3.59367132e-01 1.96534663e-01 3.57136577e-02
-3.87854069e-01 -4.10640121e-01 -2.35641599e-01 5.29036701e-01
1.33235916e-01 -2.04639375e-01 -1.09171569e+00 1.81733346e+00
4.92646277e-01 6.94656894e-02 2.81983435e-01 4.59346741e-01
8.03502202e-01 1.00777948e+00 4.88801211e-01 -7.93845356e-01
1.02632451e+00 -5.29376507e-01 -1.30301803e-01 1.64733902e-01
7.27325439e-01 -8.06645572e-01 6.22318447e-01 5.51265657e-01
-1.10266781e+00 -3.57993066e-01 -1.37103903e+00 2.82916814e-01
-5.24639189e-01 -1.27771825e-01 9.05270517e-01 8.81276250e-01
-7.10619271e-01 9.54686403e-01 -6.92821026e-01 -9.59753320e-02
3.90954614e-01 7.70838797e-01 -3.45109373e-01 1.78129196e-01
-1.25919783e+00 1.02641046e+00 9.52983797e-01 -4.00314838e-01
-1.70303059e+00 -7.49447107e-01 -6.02223396e-01 -4.28397395e-03
1.17197275e-01 -1.13047409e+00 8.87959480e-01 -4.94326085e-01
-1.70897210e+00 3.52545530e-01 -2.33788833e-01 -5.81469297e-01
-1.52556924e-02 2.57120162e-01 -3.81415308e-01 -1.38125286e-01
9.46492180e-02 1.05379188e+00 6.62935793e-01 -9.83858049e-01
-4.06959295e-01 -1.27069846e-01 6.29686266e-02 2.65415788e-01
3.29163939e-01 -2.36710727e-01 3.71538848e-01 -4.50090349e-01
-3.53038609e-01 -1.06482208e+00 -6.37980461e-01 -5.06355584e-01
-7.92329669e-01 -1.05015390e-01 4.56384301e-01 -1.61549598e-01
9.29359376e-01 -1.26113129e+00 4.70242262e-01 6.95250154e-01
-1.07238203e-01 4.57529008e-01 -3.00954431e-01 9.62144971e-01
-4.38640356e-01 4.75568235e-01 -1.85747504e-01 5.70574224e-01
-3.48059267e-01 -1.24034941e-01 -3.51684541e-01 1.96452975e-01
5.00797570e-01 8.46076906e-01 -9.68788087e-01 1.38652503e-01
-1.68363020e-01 6.01490200e-01 -7.56875992e-01 8.46392736e-02
-9.48420167e-01 9.36174393e-01 -8.30669880e-01 5.64353764e-01
3.28256488e-01 -1.31312802e-01 3.78597558e-01 3.99219953e-02
-9.03237909e-02 6.25182092e-01 -7.86426902e-01 1.40393591e+00
-1.93237722e-01 -3.98406945e-02 -7.46139705e-01 -5.18922687e-01
9.48675156e-01 2.55878568e-01 2.50394672e-01 -5.25095224e-01
-1.75831243e-01 3.50980997e-01 2.98474669e-01 6.54280856e-02
2.07344770e-01 -4.82243031e-01 1.38693824e-01 4.04076993e-01
1.77821189e-01 -3.09741199e-01 4.99663979e-01 7.46942014e-02
1.16976523e+00 4.85054225e-01 4.09179688e-01 -2.17042580e-01
6.75147355e-01 2.89691240e-01 3.07056010e-01 6.47797942e-01
6.67054653e-01 1.37005925e-01 3.37566346e-01 -2.03693956e-01
-1.43758285e+00 -1.01890349e+00 -2.74755776e-01 9.39664423e-01
-3.49307597e-01 -6.19622588e-01 -7.28032053e-01 -3.34433973e-01
-3.34116429e-01 9.60674644e-01 -4.23204362e-01 -4.97095764e-01
-1.93569079e-01 -9.07431781e-01 5.03965437e-01 1.41642503e-02
-6.49067834e-02 -1.02839589e+00 -3.85189682e-01 6.63424671e-01
3.22677433e-01 -2.50892103e-01 -2.00147107e-01 6.76447034e-01
-7.87956417e-01 -1.18650949e+00 -7.80430734e-01 -8.33551943e-01
6.33701980e-01 -4.12815481e-01 8.63820195e-01 -1.96757376e-01
-4.33209926e-01 -2.88962811e-01 -1.14798345e-01 -7.75669813e-01
-9.59435105e-01 2.00936079e-01 1.45526335e-01 -5.90860903e-01
2.40599923e-02 -6.78030789e-01 -6.71039224e-01 2.22464502e-01
-8.11204255e-01 -4.22173403e-02 6.26738429e-01 9.93734419e-01
9.86289561e-01 -1.33153751e-01 1.05533361e+00 -1.06527472e+00
9.75970089e-01 -4.53775465e-01 -6.24362886e-01 2.90099025e-01
-7.18616784e-01 5.05476832e-01 5.35926163e-01 -6.83975756e-01
-9.22535062e-01 7.04401195e-01 -4.70467776e-01 3.35139692e-01
-8.75981376e-02 7.89314091e-01 -4.21605945e-01 -9.77391675e-02
1.10636568e+00 4.27540660e-01 -3.80537771e-02 -3.32818240e-01
4.12707984e-01 9.52252895e-02 1.36231363e-01 -6.60338998e-01
6.17209733e-01 1.07801519e-02 4.39887762e-01 -8.92095268e-01
-4.32816565e-01 -6.73154071e-02 -2.56690919e-01 3.77362967e-01
8.13862085e-01 -8.77593160e-01 -9.25576150e-01 -2.30882242e-01
-1.06467474e+00 -2.32243761e-01 -1.98714063e-01 6.38558269e-01
-8.65940213e-01 -2.03997828e-02 -8.10199529e-02 -5.70042014e-01
-4.03346598e-01 -1.27044463e+00 1.04883587e+00 1.93257421e-01
-6.95054471e-01 -7.81114697e-01 3.79292995e-01 -7.22839534e-02
2.13606935e-02 5.85917532e-01 1.48815382e+00 -1.07213748e+00
-8.15349698e-01 1.97911281e-02 5.57673573e-01 -1.46710157e-01
3.21608812e-01 2.54909974e-02 -5.96027017e-01 -4.77029949e-01
-3.01082045e-01 -5.08932829e-01 7.06589639e-01 2.44113922e-01
7.23150373e-01 -4.48494345e-01 -6.08369708e-01 4.24898684e-01
1.09389174e+00 9.96621966e-01 9.63973641e-01 1.93717733e-01
3.07766289e-01 2.18203470e-01 5.17890453e-01 3.86354089e-01
-4.67128217e-01 8.78296077e-01 3.45849216e-01 -1.97604552e-01
1.56699061e-01 -6.37128890e-01 3.67952615e-01 -2.38711447e-01
-2.08355740e-01 -3.69040489e-01 -4.45983469e-01 1.26791760e-01
-1.31416965e+00 -1.27196860e+00 -2.37468868e-01 2.48135853e+00
1.28993392e+00 1.22989900e-02 4.97198284e-01 -2.58871704e-01
5.60672164e-01 -3.11061412e-01 -9.00865138e-01 -6.44649565e-01
-8.53086486e-02 1.06270337e+00 4.22588646e-01 5.57982743e-01
-8.10676992e-01 1.10478258e+00 7.87220049e+00 9.23536003e-01
-9.86769438e-01 -5.25474370e-01 6.81979001e-01 6.05874732e-02
-6.48700237e-01 5.93997478e-01 -1.11407828e+00 2.75753379e-01
1.27298558e+00 -6.19142830e-01 3.00861478e-01 6.70491934e-01
4.99914318e-01 -2.55799621e-01 -1.43623841e+00 3.33047211e-01
-3.49341601e-01 -2.04798627e+00 4.73992825e-01 5.43644905e-01
8.45135033e-01 -6.20120466e-01 1.45960562e-02 -3.32537602e-04
3.50921780e-01 -1.49621058e+00 3.70717436e-01 4.73663002e-01
8.17676246e-01 -1.23209846e+00 1.44289523e-01 4.04253393e-01
-5.69789112e-01 3.08128774e-01 -4.40168709e-01 4.14539933e-01
2.06728294e-01 2.57956654e-01 -1.66680741e+00 1.82257310e-01
-8.56976062e-02 2.16197342e-01 -3.90707970e-01 1.12702644e+00
-4.48030353e-01 4.42090213e-01 -2.53184259e-01 -4.30351585e-01
2.20491529e-01 -3.40264171e-01 4.78497386e-01 7.48354137e-01
3.99120867e-01 4.81357537e-02 1.76310569e-01 1.13450873e+00
8.98503214e-02 2.65527427e-01 -8.81392837e-01 -3.65716785e-01
3.31199646e-01 6.82664037e-01 -3.74875069e-01 -3.37199837e-01
1.93920493e-01 9.35929835e-01 -7.21918941e-02 5.12973130e-01
-7.13737249e-01 -4.14963692e-01 3.97812933e-01 2.16843277e-01
2.78629094e-01 2.27526575e-02 2.93701887e-01 -3.71040434e-01
-5.79315662e-01 -1.16175973e+00 1.40606437e-03 -8.30615699e-01
-8.69771421e-01 6.33803725e-01 1.32687405e-01 -9.50082123e-01
-5.02265036e-01 -3.90880883e-01 -6.24068081e-01 1.35494006e+00
-8.10725927e-01 -9.50901568e-01 4.14277673e-01 2.96348006e-01
3.27234954e-01 -4.11650300e-01 1.24690413e+00 -6.84271678e-02
-1.71727508e-01 2.01008976e-01 1.12040155e-01 -8.29020679e-01
5.97994149e-01 -1.02905607e+00 3.80992442e-01 3.10200006e-01
-2.85856742e-02 1.10236967e+00 1.07848954e+00 -1.04155922e+00
-1.04503691e+00 -1.11073303e+00 7.95741558e-01 -1.65902376e-01
3.84868234e-01 -4.14814919e-01 -5.51517010e-01 2.26339743e-01
-2.26526707e-01 -8.16523790e-01 1.12533653e+00 -1.28588036e-01
-5.22541478e-02 5.90771735e-01 -1.21873403e+00 7.81506360e-01
1.00419652e+00 -3.84356737e-01 -2.22327262e-01 7.52912164e-01
6.10074222e-01 -4.53823388e-01 -9.68089521e-01 2.10688010e-01
6.34836733e-01 -5.55915654e-01 1.09821332e+00 -1.17491782e+00
3.81167233e-01 -6.40371203e-01 1.49133563e-01 -1.48444676e+00
-3.20012718e-01 -9.13784325e-01 1.79035310e-02 7.37116039e-01
1.18384004e+00 -5.01434565e-01 9.44566846e-01 3.27545524e-01
-9.69225913e-02 -7.25033939e-01 -9.36614752e-01 -9.27226782e-01
2.08719105e-01 3.63318980e-01 6.11902535e-01 3.86505812e-01
1.14496902e-01 1.10835469e+00 -5.58536768e-01 -2.24803522e-01
4.97961603e-02 -2.50058640e-02 6.20215535e-01 -1.26996064e+00
-5.16696036e-01 -1.45831972e-01 -2.53689915e-01 -1.04801404e+00
-2.09557097e-02 -1.08577549e+00 -3.14126581e-01 -1.48065460e+00
1.61799878e-01 -3.87682527e-01 2.29044974e-01 3.81862879e-01
1.40833765e-01 -8.27333927e-02 -2.24759489e-01 3.09257070e-04
5.36312163e-02 5.81863523e-01 1.09887969e+00 -2.45936379e-01
-6.57743037e-01 3.95109504e-01 -1.05664754e+00 1.81866482e-01
7.26139784e-01 -6.35460258e-01 -7.50998437e-01 6.88165009e-01
5.78674972e-01 2.10107252e-01 -1.21696154e-02 -5.94374359e-01
-2.67438352e-01 -4.86767709e-01 7.10517645e-01 -4.76054519e-01
4.05950934e-01 -4.48732555e-01 9.61466551e-01 7.32564211e-01
-4.13649410e-01 -2.65367448e-01 2.31712088e-01 7.20229685e-01
6.08860962e-02 -4.11074430e-01 7.16316164e-01 -2.13698253e-01
-3.21381569e-01 3.37170362e-01 -8.86086524e-01 -5.43873489e-01
1.01234937e+00 -4.47184622e-01 -2.57298589e-01 -3.04512918e-01
-1.15606654e+00 -9.20784697e-02 4.65427786e-01 2.34664261e-01
7.31854796e-01 -1.23850346e+00 -5.19432664e-01 -4.72899042e-02
3.34826201e-01 -2.48937711e-01 -3.17426920e-01 3.48322511e-01
-6.24006331e-01 7.97453225e-01 -1.59904689e-01 -3.59951824e-01
-1.27887583e+00 9.25550580e-01 3.76957357e-01 -2.46486664e-01
3.69897895e-02 7.79785156e-01 3.65767628e-01 -1.66603297e-01
-9.38390344e-02 9.45350975e-02 -1.28019676e-01 -1.34036943e-01
4.96001065e-01 -1.00931421e-01 2.61887521e-01 -3.46933693e-01
-4.15571541e-01 1.51648253e-01 -2.42560461e-01 4.43523517e-03
1.45780587e+00 6.59029722e-01 2.08303362e-01 -9.93450582e-02
1.08254373e+00 6.34232089e-02 -1.05805945e+00 3.66440028e-01
1.08822193e-02 1.11398986e-02 -2.78650612e-01 -1.10449088e+00
-1.84641778e-01 3.25485885e-01 7.47688353e-01 -3.27627987e-01
6.59988642e-01 1.03310794e-02 3.76601011e-01 7.92257845e-01
4.98543024e-01 -9.30627286e-01 1.56144142e-01 2.91617751e-01
9.78293717e-01 -7.54965067e-01 3.22728008e-01 -4.27917868e-01
-4.78221446e-01 1.05264080e+00 1.59697056e-01 6.32873029e-02
2.52855152e-01 -8.75420868e-02 -5.77123344e-01 -3.10150594e-01
-9.90099549e-01 -1.88795954e-01 3.78757030e-01 9.87954021e-01
6.36765003e-01 -5.11300378e-02 -8.48330498e-01 -1.98047444e-01
-2.03127533e-01 -9.64984074e-02 5.04056334e-01 9.18581128e-01
-5.60426176e-01 -1.92730045e+00 -1.16321981e-01 2.59452611e-01
-2.75138170e-01 -5.20156085e-01 -1.04014313e+00 4.52019244e-01
1.38058946e-01 8.65045905e-01 -4.56576914e-01 -1.39495924e-01
2.34101892e-01 2.09388644e-01 7.61590540e-01 -1.04332709e+00
-6.34596765e-01 3.99580121e-01 6.15278721e-01 -4.65361625e-02
-1.62116662e-01 -3.88968498e-01 -1.31311560e+00 4.01235260e-02
-5.96330881e-01 4.91275907e-01 6.30005538e-01 7.53746450e-01
4.08888340e-01 3.48955661e-01 6.46418452e-01 -8.07630956e-01
-2.18420923e-01 -5.30937672e-01 -4.73542482e-01 -1.29816324e-01
1.39891617e-02 -5.49652100e-01 1.89028606e-01 3.83055508e-01] | [4.928395748138428, 5.737380027770996] |
cb060785-3b30-435d-bcf8-ebd4ddf0eb3c | hopfield-model-with-planted-patterns-a | 2304.13710 | null | https://arxiv.org/abs/2304.13710v1 | https://arxiv.org/pdf/2304.13710v1.pdf | Hopfield model with planted patterns: a teacher-student self-supervised learning model | While Hopfield networks are known as paradigmatic models for memory storage and retrieval, modern artificial intelligence systems mainly stand on the machine learning paradigm. We show that it is possible to formulate a teacher-student self-supervised learning problem with Boltzmann machines in terms of a suitable generalization of the Hopfield model with structured patterns, where the spin variables are the machine weights and patterns correspond to the training set's examples. We analyze the learning performance by studying the phase diagram in terms of the training set size, the dataset noise and the inference temperature (i.e. the weight regularization). With a small but informative dataset the machine can learn by memorization. With a noisy dataset, an extensive number of examples above a critical threshold is needed. In this regime the memory storage limits of the system becomes an opportunity for the occurrence of a learning regime in which the system can generalize. | ['Daniele Tantari', 'Gianluca Manzan', 'Luca Camanzi', 'Francesco Alemanno'] | 2023-04-26 | null | null | null | null | ['memorization'] | ['natural-language-processing'] | [ 3.73236448e-01 3.19564402e-01 -1.00352317e-01 -2.85153866e-01
2.98168778e-01 -1.85619682e-01 8.81397307e-01 2.46305913e-01
-9.56650734e-01 8.26695144e-01 -3.47940654e-01 3.24538834e-02
-5.86067915e-01 -1.07643187e+00 -5.87913454e-01 -1.42615223e+00
-3.37065533e-02 7.96787262e-01 5.96750855e-01 -3.11870456e-01
4.46742594e-01 5.56726813e-01 -1.94387865e+00 -3.13970745e-02
9.04966891e-01 1.11987782e+00 2.09261581e-01 5.47732472e-01
-2.64318943e-01 5.84313154e-01 -4.90199924e-01 -5.44005120e-03
2.20589802e-01 -4.99555767e-01 -9.18842912e-01 -2.31261089e-01
2.25495979e-01 4.08026904e-01 -5.48417985e-01 1.33081377e+00
3.88270527e-01 7.08673656e-01 8.08777392e-01 -8.84517610e-01
-4.97706711e-01 5.67841887e-01 5.62208146e-02 5.80739498e-01
-1.83422923e-01 1.89542711e-01 4.06597942e-01 -6.71076775e-01
6.76377177e-01 7.60751247e-01 3.10415059e-01 8.55482399e-01
-1.39344382e+00 -3.59273612e-01 -3.01970541e-01 6.00736141e-01
-1.18893445e+00 -3.86923850e-01 7.60674953e-01 -3.23384851e-01
1.16346586e+00 -3.79685350e-02 1.11215937e+00 7.84662306e-01
4.74343240e-01 1.28039271e-01 1.34011769e+00 -8.86019409e-01
1.06285179e+00 3.63095611e-01 7.15972483e-01 5.87833703e-01
4.61432636e-01 3.47140342e-01 -7.07240820e-01 9.77428854e-02
3.70232701e-01 -1.78789526e-01 -1.89170033e-01 -4.24046487e-01
-6.69721961e-01 6.82346523e-01 4.52486277e-01 6.96664631e-01
-2.31537730e-01 -7.75564462e-02 2.09146336e-01 5.32795906e-01
2.51685202e-01 6.10047102e-01 -6.31113052e-02 2.58369833e-01
-9.54265714e-01 1.83081195e-01 9.20515001e-01 5.10714233e-01
8.22592258e-01 2.30108395e-01 2.57949233e-01 6.25136852e-01
4.26564738e-02 7.36460567e-01 8.85504186e-01 -5.32690287e-01
1.74193885e-02 5.07082999e-01 -5.97471409e-02 -4.39665467e-01
-3.75199139e-01 -3.79403770e-01 -8.39991152e-01 4.05381292e-01
6.64059460e-01 1.09514527e-01 -1.04468095e+00 1.72855699e+00
1.70630068e-01 6.88351691e-02 4.75480169e-01 8.50414455e-01
5.19411445e-01 6.68872058e-01 1.49682954e-01 -7.35966742e-01
1.04088664e+00 -5.24549544e-01 -7.80367374e-01 -3.63940746e-01
4.47075307e-01 -1.25460044e-01 8.34865570e-01 5.75861454e-01
-1.22984207e+00 -2.66784400e-01 -1.27527118e+00 2.07816869e-01
-7.79413760e-01 -7.51534700e-01 4.06787664e-01 4.86320287e-01
-1.06645858e+00 1.09098446e+00 -7.50073612e-01 -4.28734034e-01
2.14384824e-01 5.44249296e-01 -4.31139082e-01 1.11316338e-01
-1.27721691e+00 1.44524932e+00 9.48961675e-01 4.84304540e-02
-6.41298175e-01 -2.91756362e-01 -3.85439515e-01 2.27986693e-01
-1.52247652e-01 -4.49354887e-01 8.64461601e-01 -1.17352736e+00
-1.58593237e+00 1.01128638e+00 -1.02364682e-02 -6.28377557e-01
8.20181593e-02 2.82928944e-01 -3.66746753e-01 4.32896167e-01
-5.35421371e-01 4.64704573e-01 1.04735792e+00 -8.25388908e-01
-4.72115818e-03 -5.88710785e-01 -4.51307654e-01 -9.49581787e-02
-4.75319475e-01 -3.42628777e-01 3.95909399e-01 -2.04865903e-01
3.47065955e-01 -1.11639345e+00 -1.33455843e-01 -4.38119411e-01
2.76948631e-01 -6.14361942e-01 6.76846921e-01 -1.13364115e-01
9.64356482e-01 -2.26491547e+00 4.22635943e-01 5.44296145e-01
-1.92947709e-03 4.03804630e-01 1.49872452e-01 2.53937036e-01
-2.25149557e-01 -8.88007358e-02 -1.83113784e-01 2.55281299e-01
-1.82750300e-01 5.75115919e-01 -2.78122693e-01 5.44734180e-01
3.51048335e-02 6.25630617e-01 -7.95639992e-01 -4.97282505e-01
-1.82697698e-01 3.29243362e-01 -5.62193632e-01 2.77284831e-01
-1.81404054e-01 1.93348959e-01 -3.83070171e-01 -2.44641572e-01
5.11205792e-01 -3.11733484e-01 1.04507729e-01 3.91599566e-01
-2.22739145e-01 2.72919893e-01 -1.10592437e+00 1.31643379e+00
-1.89755887e-01 5.92419922e-01 -1.77579641e-01 -1.37691653e+00
9.70967948e-01 3.78098905e-01 -2.76127434e-03 -6.82913125e-01
1.74151435e-01 3.66621673e-01 4.81289715e-01 -6.98373318e-01
1.06078178e-01 -5.17237008e-01 3.71236593e-01 5.90465248e-01
5.89187801e-01 -1.47739097e-01 3.67964923e-01 -3.64023671e-02
1.10537684e+00 -3.70434314e-01 1.08661950e-01 -8.85012507e-01
3.97805899e-01 -3.15263182e-01 2.17290998e-01 8.64194393e-01
3.39759439e-02 6.13726936e-02 4.43674386e-01 -7.02077210e-01
-1.12785709e+00 -1.10517216e+00 -6.83601320e-01 8.21717024e-01
1.51891252e-02 -6.38905093e-02 -8.27497005e-01 -5.15522435e-02
-1.39527276e-01 6.47262812e-01 -8.33607674e-01 -7.40296602e-01
-5.26847303e-01 -9.85084236e-01 1.94820732e-01 6.45748302e-02
6.95313811e-01 -1.62916672e+00 -1.02442229e+00 2.82909479e-02
4.46408212e-01 -4.67223763e-01 3.51529598e-01 9.92903650e-01
-1.28550553e+00 -8.98832142e-01 -4.02199686e-01 -7.06881046e-01
6.58294439e-01 -3.01501423e-01 9.50198591e-01 1.56888708e-01
-3.27267706e-01 2.67457843e-01 2.12739292e-03 -3.61790329e-01
-5.01497746e-01 1.92511320e-01 5.29205084e-01 -2.06438869e-01
4.95319724e-01 -1.05959356e+00 -7.17724621e-01 -1.54556572e-01
-1.25869083e+00 -1.82307139e-01 4.34886187e-01 9.81447875e-01
3.01173210e-01 1.53576806e-01 7.78990686e-01 -5.19566357e-01
3.88916016e-01 -3.97750169e-01 -6.98314428e-01 2.23661259e-01
-1.13162076e+00 8.42324138e-01 6.57108903e-01 -7.53413379e-01
-1.04785073e+00 -2.41680518e-01 2.30214134e-01 1.21039186e-04
-2.44555652e-01 3.34416449e-01 -2.40913182e-02 -4.64189380e-01
7.57676899e-01 5.00322640e-01 -2.13658251e-02 -2.67214298e-01
3.88395716e-03 5.71202397e-01 1.69195309e-01 -5.45143783e-01
5.63886344e-01 7.00683594e-02 2.76816636e-01 -9.25390899e-01
-7.58607745e-01 -3.51368748e-02 -6.78573549e-01 -4.94753271e-01
7.87207067e-01 -1.73219025e-01 -7.39165425e-01 4.94095802e-01
-8.42746019e-01 -3.78109694e-01 -7.32637405e-01 6.24350429e-01
-7.15003729e-01 -1.03260688e-01 -6.91591620e-01 -9.28655446e-01
-1.60278022e-01 -8.24525893e-01 9.30255279e-02 5.54953992e-01
3.19521502e-02 -9.84691381e-01 2.97682434e-01 -7.00152218e-02
5.95837712e-01 -4.41669196e-01 1.33062637e+00 -1.06241286e+00
-3.77761811e-01 -7.65218437e-02 3.41619909e-01 2.91965842e-01
-5.39487481e-01 -2.62214780e-01 -1.19074059e+00 -1.49554133e-01
4.55130458e-01 -4.32772458e-01 1.20085108e+00 -2.68837698e-02
1.05029178e+00 -4.97789711e-01 2.15829518e-02 1.98504180e-01
1.43136430e+00 3.40117306e-01 6.05271161e-01 2.76259512e-01
8.27687830e-02 6.10915542e-01 -1.69547856e-01 -1.56641472e-02
-3.81811291e-01 3.67197543e-01 1.39978617e-01 7.03945935e-01
3.38228256e-01 -6.95333034e-02 8.82459134e-02 1.17655277e+00
-3.80452693e-01 1.36651531e-01 -1.09554720e+00 2.47568980e-01
-1.69930863e+00 -1.25034952e+00 2.36537024e-01 2.49475193e+00
1.23056602e+00 4.03459638e-01 -2.35127181e-01 5.64641893e-01
4.96443182e-01 1.35782897e-01 -5.13877809e-01 -5.79865217e-01
-2.26512626e-01 6.30790234e-01 3.08948189e-01 5.76065898e-01
-7.00353861e-01 6.78829670e-01 6.59212589e+00 7.27960467e-01
-1.28936696e+00 1.23005755e-01 2.11830616e-01 -2.78215617e-01
-7.51816705e-02 -8.83484930e-02 -6.63880110e-01 5.75807571e-01
1.53675520e+00 -1.32715911e-01 7.41004229e-01 6.54293537e-01
-2.83155233e-01 -6.34854853e-01 -1.04213130e+00 8.31998050e-01
1.41612841e-02 -1.24170160e+00 2.24199612e-02 1.73339978e-01
6.56870723e-01 -1.27426669e-01 1.98154390e-01 2.67412126e-01
-1.69764057e-01 -1.01405311e+00 3.64718348e-01 8.29960704e-01
2.87027448e-01 -6.50616169e-01 5.51184893e-01 6.16851628e-01
-1.91165090e-01 -3.56745541e-01 -6.02998614e-01 -2.14619085e-01
-3.61901850e-01 6.04672790e-01 -1.60420775e-01 -2.40876928e-01
5.36557555e-01 3.01887393e-01 -6.07455373e-01 1.07147634e+00
-1.48726394e-02 7.32435763e-01 -6.11109197e-01 -4.09834683e-01
1.50463492e-01 -6.19745076e-01 4.86890644e-01 8.45124662e-01
2.14831978e-01 4.62679148e-01 -2.44215220e-01 7.00574696e-01
-1.01199426e-01 -8.89498964e-02 -7.29207516e-01 8.23679566e-02
3.58061999e-01 1.00472486e+00 -1.01932621e+00 -4.33923483e-01
1.89414918e-01 6.18157148e-01 5.35938919e-01 5.62796831e-01
-7.43146166e-02 -3.12670410e-01 -8.54652226e-02 2.33607411e-01
1.57921523e-01 -7.80455023e-02 -3.32647681e-01 -1.16761231e+00
-6.73711449e-02 -2.24242911e-01 6.75371364e-02 -6.79582596e-01
-8.80737364e-01 4.71798062e-01 1.37546241e-01 -5.55936456e-01
-2.22009197e-01 -7.09004998e-01 -7.26736128e-01 5.47435522e-01
-1.12945974e+00 -4.04097557e-01 1.94962740e-01 6.54106796e-01
-2.28394940e-01 -3.28891486e-01 1.15291131e+00 -2.69511063e-03
-4.45121557e-01 2.75291264e-01 5.80624282e-01 -1.04393221e-01
2.79117733e-01 -1.29494143e+00 -4.92691815e-01 4.04762864e-01
2.78644711e-01 6.95712090e-01 1.21933222e+00 -1.92338690e-01
-1.33871853e+00 -4.77127433e-01 9.73487616e-01 -9.57429931e-02
7.99985468e-01 -5.45387924e-01 -1.37791610e+00 3.56437415e-01
3.70144725e-01 3.33677709e-01 6.83136404e-01 2.53866427e-02
-2.51867265e-01 -3.60670149e-01 -1.22554922e+00 3.06128472e-01
7.33634472e-01 -6.61777318e-01 -1.13127840e+00 4.13349867e-01
1.00273199e-01 2.64939964e-01 -8.36991847e-01 1.88667644e-02
4.02127892e-01 -9.65922534e-01 4.29779053e-01 -9.43231642e-01
1.50682390e-01 1.32450670e-01 3.17172408e-01 -1.41875327e+00
-1.53667152e-01 -3.80441368e-01 -2.47630835e-01 6.46347463e-01
2.72851795e-01 -6.77176952e-01 5.69399059e-01 6.76982105e-01
3.10215443e-01 -7.67515719e-01 -1.56519353e+00 -7.29236662e-01
3.41470957e-01 9.97289270e-03 1.08906269e-01 7.35502243e-01
6.46903813e-01 4.13493961e-01 2.77596533e-01 -2.21620381e-01
7.31629133e-01 5.68795055e-02 -2.16730714e-01 -1.58284438e+00
-2.52085209e-01 -4.21439320e-01 -5.81939757e-01 -4.02195811e-01
5.12812316e-01 -9.46602523e-01 5.77329844e-02 -8.72891366e-01
3.15844744e-01 -4.51299280e-01 -6.43503904e-01 8.65122601e-02
2.00209111e-01 1.18244486e-02 4.96733375e-02 2.88186640e-01
-7.01977372e-01 3.08822721e-01 8.44692767e-01 3.78893949e-02
-1.74999565e-01 2.11373884e-02 1.25673607e-01 6.36656165e-01
9.39052761e-01 -7.27754831e-01 -3.97132903e-01 2.72095073e-02
6.14625335e-01 8.53645056e-02 2.82889277e-01 -1.18332708e+00
5.47806144e-01 -8.48978013e-02 5.93754351e-01 1.10992203e-02
2.24355534e-01 -7.98654854e-01 9.91522446e-02 6.91732705e-01
-8.75809491e-01 -2.34468639e-01 -6.77053481e-02 6.05281711e-01
2.54539344e-02 -1.05686235e+00 1.08249855e+00 -1.94768265e-01
-2.66559511e-01 1.03595391e-01 -8.49713683e-01 1.35548726e-01
9.28185999e-01 -9.07963216e-02 -3.32890272e-01 -3.00541427e-02
-1.14194143e+00 -5.63789569e-02 3.53528321e-01 1.85631271e-02
5.19261837e-01 -1.12850440e+00 -1.17614940e-01 4.90407735e-01
-1.10366434e-01 -3.35200019e-02 1.36752591e-01 7.33229816e-01
-3.21324766e-01 3.27600986e-01 -6.19393229e-01 -4.14960861e-01
-8.38878453e-01 7.75506556e-01 6.68926716e-01 -1.24991789e-01
-5.07040203e-01 4.92291480e-01 -1.76899090e-01 9.31615308e-02
2.37661198e-01 -9.06778425e-02 -2.71182388e-01 1.03222996e-01
7.45172024e-01 2.12884888e-01 2.09238172e-01 -2.67946601e-01
-7.31206015e-02 2.15493545e-01 2.09016185e-02 -3.15238476e-01
1.47278595e+00 3.66969526e-01 -3.64884585e-01 9.04403627e-01
9.01307642e-01 -4.15116221e-01 -9.63039160e-01 -2.96959072e-01
4.81015921e-01 8.17127898e-02 2.09518433e-01 -4.61478025e-01
-7.96662748e-01 1.23859417e+00 8.49319875e-01 5.68827629e-01
9.94711518e-01 1.43397853e-01 1.68085799e-01 9.31747794e-01
3.87906224e-01 -1.56555295e+00 -2.74598924e-03 8.28903794e-01
6.55090511e-01 -1.00233698e+00 -8.82304907e-02 3.42483640e-01
-8.46452862e-02 1.45553398e+00 5.28776050e-01 -6.48150444e-01
9.45437491e-01 2.91871548e-01 -4.38979149e-01 -1.61044583e-01
-9.48387682e-01 1.46161556e-01 1.97837621e-01 3.37013543e-01
2.13346511e-01 -2.27347557e-02 -4.40135628e-01 5.67352533e-01
-1.94408402e-01 1.01186363e-02 5.11571407e-01 8.86732638e-01
-1.06531739e+00 -9.12610233e-01 -2.83782184e-01 5.58404267e-01
-2.92243809e-01 -5.54357208e-02 -1.84984922e-01 4.99637932e-01
1.81497876e-02 5.36098421e-01 3.05270314e-01 -5.82293347e-02
-2.56357267e-02 9.68097389e-01 9.47444081e-01 -4.80675638e-01
-6.70120120e-01 -4.95558649e-01 -3.39344293e-01 -1.49967253e-01
-5.70450068e-01 -2.21596494e-01 -1.30180538e+00 -3.51514339e-01
-4.57183957e-01 7.95261383e-01 5.43906152e-01 1.26392865e+00
-1.20424293e-01 1.90108091e-01 1.63239598e-01 -7.45959282e-01
-8.28605890e-01 -1.19599092e+00 -1.04037929e+00 1.77542210e-01
3.48336130e-01 -7.17151046e-01 -7.63759792e-01 -7.31187910e-02] | [7.932426929473877, 3.5184669494628906] |
7db4119b-a21a-4f72-a67c-4ebca86218e4 | licd-a-language-independent-approach-for | null | null | https://link.springer.com/chapter/10.1007/978-3-030-15712-8_37 | https://link.springer.com/chapter/10.1007/978-3-030-15712-8_37 | LICD: A Language-Independent Approach for Aspect Category Detection | Aspect-based sentiment analysis (ABSA) deals with processing and summarizing customer reviews and has been a topic of interest in recent years. Given a set of predefined categories, Aspect Category Detection (ACD), as a subtask of ABSA, aims to assign a subset of these categories to a given review sentence. Thanks to the existence of websites such as Yelp and TripAdvisor, there exist a huge amount of reviews in several languages, and therefore the need for language-independent methods in this task seems necessary. In this paper, we propose Language-Independent Category Detector (LICD), a supervised method based on text matching without the need for any language-specific tools and hand-crafted features for identifying aspect categories. For a given sentence, our proposed method performs ACD based on two hypotheses: First, a category should be assigned to a sentence if there is a high semantic similarity between the sentence and a set of representative words of that category. Second, a category should be assigned to a sentence if sentences with high semantic and structural similarity to that sentence belong to that category. To apply the former hypothesis, we used soft cosine measure, and for the latter, word mover’s distance measure is utilized. Using these two measures, for a given sentence we calculate a set of similarity scores as features for a one-vs-all logistic regression classifier per category. Experimental results on the multilingual SemEval-2016 datasets in the restaurant domain demonstrate that our approach outperforms baseline methods in English, Russian, and Dutch languages, and obtains competitive results with the strong deep neural network-based baselines in French, Turkish, and Spanish languages. | ['Azadeh Shakery', 'Heshaam Faili', 'Masoud Jalili Sabet', 'Sajad Movahedi', 'Erfan Ghadery'] | 2019-04-07 | null | null | null | ecir-2019-4 | ['aspect-term-extraction-and-sentiment', 'text-matching', 'aspect-based-sentiment-analysis', 'semantic-textual-similarity', 'aspect-category-detection'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 4.51531112e-02 -3.64846349e-01 -1.44841567e-01 -7.42484510e-01
-7.17920303e-01 -5.44368684e-01 7.78325737e-01 7.45095730e-01
-5.47758460e-01 2.80626804e-01 2.36714169e-01 -4.57705230e-01
3.49075645e-02 -9.19742763e-01 -3.40670794e-01 -5.11960328e-01
3.66879910e-01 3.35269064e-01 4.03517485e-02 -4.77832049e-01
4.72289681e-01 1.36533454e-01 -1.39263487e+00 4.37978595e-01
8.02408636e-01 1.08849216e+00 2.25437939e-01 2.01605171e-01
-5.08782625e-01 4.60718274e-01 -5.02779245e-01 -7.03704476e-01
4.75707604e-03 -6.25854790e-01 -7.27151871e-01 1.40818819e-01
1.47893578e-01 3.82775098e-01 3.63730192e-01 1.02308130e+00
2.31034368e-01 2.74830222e-01 8.58992279e-01 -7.69228697e-01
-4.85111326e-01 7.30473101e-01 -5.75381696e-01 3.22086141e-02
2.65782326e-01 -3.63385439e-01 1.51686811e+00 -1.16154051e+00
5.17670333e-01 1.08945477e+00 5.98089278e-01 3.59097838e-01
-7.99867451e-01 -3.66040379e-01 5.29266238e-01 8.28560516e-02
-9.94773090e-01 -5.90651780e-02 8.59376252e-01 -2.74720609e-01
1.08001220e+00 1.63456798e-01 6.07931793e-01 6.61316931e-01
2.67385423e-01 6.89411521e-01 1.08773935e+00 -6.71853006e-01
5.73848903e-01 6.75791144e-01 7.10712194e-01 2.83626586e-01
2.17147052e-01 -6.24803185e-01 -4.83874679e-01 -5.13591692e-02
-2.98216611e-01 -3.85729112e-02 -1.18584558e-02 -3.82607847e-01
-9.03994262e-01 1.16345978e+00 3.02719921e-01 6.74167275e-01
-3.96878123e-01 -3.97909105e-01 8.11613381e-01 2.64323056e-01
7.97074974e-01 4.53744233e-01 -6.77941024e-01 2.13053674e-02
-7.90747941e-01 1.53163791e-01 1.03472054e+00 7.64448643e-01
7.56007075e-01 -2.88033605e-01 -2.38307504e-04 1.06132436e+00
3.70892316e-01 4.78832990e-01 7.83463299e-01 -2.17157289e-01
5.70546210e-01 1.18134511e+00 -1.78038403e-01 -1.28266609e+00
-3.96688342e-01 -3.78419042e-01 -8.07050645e-01 -2.40498409e-01
1.53870746e-01 1.98589638e-01 -7.19215930e-01 1.44382107e+00
4.33783203e-01 -3.88509423e-01 3.91451508e-01 8.92521799e-01
9.82173324e-01 6.15464032e-01 -2.61506941e-02 -3.68116975e-01
1.63650382e+00 -8.72842491e-01 -5.16495585e-01 -4.14691955e-01
6.80889726e-01 -9.73744929e-01 1.24927270e+00 2.89530307e-01
-5.65596402e-01 -5.54464757e-01 -1.11380482e+00 8.96389633e-02
-8.24862540e-01 2.30393857e-01 5.35677969e-01 7.52921999e-01
-7.89689124e-01 3.92664552e-01 -4.90805775e-01 -7.80119419e-01
3.89734097e-02 7.12820664e-02 -2.96906590e-01 -1.23347096e-01
-1.23555839e+00 7.26923525e-01 2.63612568e-01 8.61660466e-02
-3.18653166e-01 -1.89676866e-01 -1.10199404e+00 1.45728951e-02
3.86274047e-02 -4.50462580e-01 1.08905661e+00 -1.40573835e+00
-1.15596211e+00 1.10587084e+00 -4.27424401e-01 -5.76689243e-01
7.90882483e-02 -1.18325479e-01 -6.85788631e-01 1.01079561e-01
3.77289057e-01 2.71784216e-01 8.08288038e-01 -1.09290206e+00
-7.69330144e-01 -4.79115844e-01 1.86240375e-01 4.30382073e-01
-6.86345160e-01 4.03648585e-01 -4.49644148e-01 -4.70698148e-01
1.36098370e-01 -8.18679571e-01 -1.43967882e-01 -5.36624730e-01
-4.64827299e-01 -5.91567516e-01 6.26842737e-01 -6.80381417e-01
1.33191478e+00 -2.15706182e+00 -1.53036743e-01 2.79736012e-01
-4.78696860e-02 3.06469053e-01 -1.72416389e-01 3.07814091e-01
4.38220017e-02 5.45981191e-02 -4.99043345e-01 -4.80087966e-01
4.66233902e-02 -1.06696337e-01 -2.57574141e-01 2.83762902e-01
1.17271833e-01 4.12755311e-01 -9.52434957e-01 -4.54275519e-01
6.09168112e-02 3.40945274e-01 -5.12904704e-01 1.93876233e-02
-9.73617062e-02 1.89445391e-02 -3.81897926e-01 4.33095366e-01
6.13595188e-01 6.82502985e-03 2.83303499e-01 -2.29965851e-01
-2.02718258e-01 7.95377553e-01 -1.08690071e+00 1.39370263e+00
-8.63649845e-01 4.30953830e-01 -3.22063357e-01 -1.33962369e+00
1.38386798e+00 1.38179466e-01 1.86748549e-01 -6.45244241e-01
3.27031136e-01 5.29421926e-01 8.40432346e-02 -2.11066559e-01
8.01084459e-01 -1.94435477e-01 -5.20538807e-01 4.97948200e-01
-2.82942299e-02 -1.74958840e-01 6.85077429e-01 1.76889688e-01
9.49638009e-01 -6.01941645e-02 4.53444451e-01 -3.24856877e-01
9.07307327e-01 2.16472503e-02 5.33167362e-01 5.37135303e-01
-2.30764419e-01 6.96049273e-01 4.37240183e-01 -5.36242366e-01
-8.52984786e-01 -6.40688598e-01 -1.91990569e-01 9.91490960e-01
1.53845295e-01 -5.35628974e-01 -5.84602356e-01 -9.04042125e-01
-1.16994344e-01 9.45719182e-01 -6.86495066e-01 -2.07846940e-01
-4.29210812e-01 -7.86129653e-01 -1.18850254e-01 2.32863605e-01
5.72527587e-01 -1.23083174e+00 -4.05106962e-01 2.04559103e-01
-2.25895166e-01 -9.39260483e-01 -5.76036155e-01 3.58394712e-01
-5.86919606e-01 -1.07426965e+00 -6.00044131e-01 -1.10869837e+00
6.41162932e-01 4.73261535e-01 1.24729085e+00 -2.85943419e-01
1.23031721e-01 1.19565450e-01 -8.33083153e-01 -3.89153957e-01
-5.13060689e-01 2.61669636e-01 1.39112934e-01 3.81688386e-01
1.12241077e+00 -4.54099588e-02 -5.34390628e-01 2.28734359e-01
-6.41243219e-01 -2.04611942e-01 5.01975834e-01 5.58711648e-01
6.05104983e-01 1.45170372e-02 6.84515059e-01 -9.04278040e-01
8.24878573e-01 -4.25892055e-01 -2.73693651e-01 2.26604611e-01
-6.12907410e-01 -2.63320446e-01 9.42543983e-01 -2.61027604e-01
-7.65445173e-01 -3.82231250e-02 -1.21560872e-01 1.34509355e-01
-1.99493542e-01 9.69789743e-01 -3.99346918e-01 5.59184611e-01
4.79039133e-01 3.73838574e-01 -1.35973871e-01 -3.78192812e-01
2.49229506e-01 1.09038734e+00 -2.07931232e-02 -3.75583768e-02
6.21786535e-01 2.99625903e-01 -4.06725824e-01 -8.34435523e-01
-1.13974845e+00 -1.08483243e+00 -7.02257931e-01 -1.63155600e-01
8.90191257e-01 -8.56666803e-01 -3.18686366e-01 3.95820588e-01
-1.18799520e+00 3.91904026e-01 -1.94987267e-01 6.21828437e-01
-1.61650568e-01 3.35547119e-01 -1.94190457e-01 -6.83960378e-01
-8.49775255e-01 -1.02672076e+00 9.01975155e-01 3.40585172e-01
-5.29541135e-01 -1.00587952e+00 9.35530812e-02 4.01760489e-01
3.93646151e-01 -1.04741603e-01 9.83956993e-01 -1.37212360e+00
9.25287902e-02 -6.36200786e-01 -5.62769733e-02 8.49532783e-01
4.52585369e-01 -1.75818473e-01 -7.86995113e-01 -2.89490163e-01
3.49560887e-01 -9.71015245e-02 8.87084484e-01 3.26472282e-01
6.73963845e-01 -3.13512057e-01 -1.92848280e-01 9.86903161e-02
1.26165330e+00 3.15096200e-01 1.84110373e-01 6.22765005e-01
5.45500100e-01 6.66681051e-01 8.66942823e-01 2.37070978e-01
5.15464783e-01 6.24675691e-01 1.50867641e-01 -5.22703081e-02
1.05786040e-01 -6.98776394e-02 5.82418382e-01 1.54225278e+00
2.80985206e-01 -2.43160039e-01 -6.19496047e-01 8.73723686e-01
-1.57179904e+00 -6.89868927e-01 -2.46612862e-01 2.38615489e+00
8.53416264e-01 3.95058304e-01 1.64318055e-01 3.09906900e-01
8.31353903e-01 2.39494756e-01 -4.81213659e-01 -9.47559416e-01
-1.82114050e-01 -5.14790006e-02 -3.06559037e-02 9.92024243e-02
-1.45801306e+00 8.12585771e-01 4.14337635e+00 9.50631857e-01
-1.07339680e+00 1.16275631e-01 7.24975526e-01 3.36479068e-01
-3.72944176e-01 4.37849127e-02 -9.93765712e-01 4.26976860e-01
9.10474658e-01 -2.56568015e-01 -8.34192615e-03 1.03072333e+00
2.43969068e-01 -1.06878683e-01 -9.03033316e-01 5.04867077e-01
5.29099286e-01 -9.26913321e-01 1.32563353e-01 -2.48859659e-01
7.15662897e-01 -3.59002165e-02 -6.95213377e-02 5.68891644e-01
5.92739433e-02 -5.54717660e-01 6.38432801e-01 4.82825227e-02
3.29674661e-01 -9.92728949e-01 1.14268064e+00 2.14962825e-01
-1.33721983e+00 2.42753297e-01 -6.89764440e-01 1.04508825e-01
1.04608303e-02 8.86007309e-01 -5.65549493e-01 5.61272979e-01
7.16350377e-01 1.03321433e+00 -4.94958907e-01 9.84025061e-01
-2.99610883e-01 7.83011317e-01 -5.54026999e-02 -5.58812380e-01
4.09837753e-01 -5.13772368e-01 4.94943321e-01 1.31359327e+00
3.38651448e-01 -4.42811102e-01 4.81268540e-02 5.30996084e-01
-1.19032249e-01 8.43403637e-01 -5.93408227e-01 7.33796060e-02
2.19714969e-01 1.53073323e+00 -9.96686459e-01 -3.18088204e-01
-6.35085940e-01 8.96312058e-01 1.72202185e-01 7.98372701e-02
-6.38144732e-01 -8.55520189e-01 4.85849261e-01 -1.57185212e-01
5.16935289e-01 9.15927067e-02 -5.44258475e-01 -1.13634849e+00
1.63997695e-01 -9.44769859e-01 3.72656226e-01 -4.15121377e-01
-1.34087515e+00 9.08561349e-01 -5.03497660e-01 -1.59290385e+00
-3.73006798e-02 -5.63461363e-01 -8.06853592e-01 7.66047895e-01
-1.73743796e+00 -1.11574960e+00 -1.37909979e-01 3.47518712e-01
8.30224633e-01 -3.98188829e-01 8.19335938e-01 2.73311764e-01
-4.70085472e-01 5.95951498e-01 2.06263900e-01 2.52617210e-01
6.83846533e-01 -1.27411926e+00 4.56043720e-01 8.80521417e-01
3.46104681e-01 7.46925950e-01 7.36558735e-01 -5.53243577e-01
-1.00100231e+00 -1.26780713e+00 1.72926748e+00 -2.71058321e-01
7.28300750e-01 -4.55387205e-01 -8.92925382e-01 3.01948845e-01
2.69704610e-01 -4.04772758e-01 7.93461740e-01 3.25463593e-01
-5.07816911e-01 -4.03087944e-01 -8.31031859e-01 3.62858415e-01
4.35176760e-01 -5.13802528e-01 -7.46074617e-01 5.30970454e-01
7.82682598e-01 2.54000854e-02 -4.82340604e-01 2.97077656e-01
2.48661935e-01 -8.12270164e-01 5.97474873e-01 -3.98609608e-01
6.50290668e-01 -4.47575331e-01 -1.87697083e-01 -1.57132781e+00
-2.45075002e-02 -1.93006303e-02 4.24231648e-01 1.60073423e+00
6.76989198e-01 -5.37755966e-01 3.33119392e-01 2.92353600e-01
-2.95693457e-01 -7.44704902e-01 -8.22086573e-01 -8.23639452e-01
2.22606868e-01 -5.33934176e-01 6.24568403e-01 7.89798558e-01
1.50424555e-01 6.41409934e-01 9.46979895e-02 -1.03731863e-01
1.82556376e-01 5.50532520e-01 6.29826367e-01 -1.02991986e+00
-8.09918568e-02 -5.36137402e-01 -3.26745361e-01 -7.19702005e-01
2.86915183e-01 -1.10997415e+00 2.14620978e-01 -1.80065668e+00
3.35521579e-01 -3.85542095e-01 -4.92828280e-01 4.04793948e-01
-2.85879493e-01 2.47914732e-01 9.43650529e-02 1.72771305e-01
-8.42070282e-01 6.44809783e-01 6.97033107e-01 -4.11640018e-01
-2.68789351e-01 5.06065309e-01 -9.04778481e-01 8.23209524e-01
8.22144270e-01 -4.08633441e-01 -1.19653493e-01 -1.07715800e-01
4.89418596e-01 -5.50511599e-01 -2.21016183e-01 -7.63765216e-01
5.10092713e-02 1.36566132e-01 -8.10556188e-02 -6.06764793e-01
5.75310215e-02 -7.69854367e-01 -5.16205370e-01 5.33476412e-01
-3.77456218e-01 1.19838670e-01 1.12112425e-02 3.76510203e-01
-6.59753323e-01 -4.45105135e-01 6.99231327e-01 8.33377838e-02
-7.76379526e-01 -1.08424731e-01 -4.93409395e-01 1.16172858e-01
9.02498603e-01 -8.53573233e-02 -1.29542276e-01 -2.18339831e-01
-2.19039425e-01 -6.35806797e-03 3.36664617e-01 7.67457664e-01
6.07845247e-01 -1.18050432e+00 -6.98941648e-01 1.01396285e-01
5.84069073e-01 -1.75344631e-01 3.38444710e-02 7.79134393e-01
-2.63624012e-01 5.38649321e-01 3.80089283e-01 -4.39700693e-01
-1.39464140e+00 4.99082297e-01 9.61412266e-02 -5.03742635e-01
-2.49489173e-01 6.90964103e-01 1.84895799e-01 -9.04170930e-01
-7.57349702e-03 -2.90105313e-01 -8.93157244e-01 5.18230259e-01
5.73780179e-01 -2.47035380e-02 5.44188082e-01 -1.02708352e+00
-6.21618390e-01 7.31886268e-01 -3.09544712e-01 2.45796323e-01
1.37138784e+00 -1.67499572e-01 -3.18340272e-01 7.27431238e-01
1.39947653e+00 2.03212723e-03 -4.20164645e-01 -3.36300224e-01
2.37462804e-01 -1.49754910e-02 -1.01075647e-02 -7.66459763e-01
-1.03362119e+00 8.58147442e-01 3.60050499e-01 3.87147278e-01
1.16994393e+00 -2.82761455e-03 7.60961950e-01 2.61511713e-01
2.03881711e-01 -1.30858040e+00 -8.54799971e-02 8.81567717e-01
6.85585737e-01 -1.25374281e+00 -7.96787590e-02 -2.09467858e-01
-9.40346301e-01 9.60900187e-01 4.35903430e-01 -1.14277430e-01
8.07688713e-01 -2.74036139e-01 3.05633485e-01 -8.88905078e-02
-5.67633092e-01 -2.55996376e-01 5.62294066e-01 2.91609287e-01
6.33053005e-01 1.33250654e-01 -8.62397134e-01 8.59982073e-01
-3.16916555e-01 -4.26466137e-01 4.60508347e-01 7.59629965e-01
-5.61683178e-01 -1.04944301e+00 -7.51184523e-02 6.51385248e-01
-3.71954471e-01 -3.05314243e-01 -3.38156551e-01 5.82395554e-01
-8.43775049e-02 1.29397786e+00 1.76171750e-01 -3.03529024e-01
4.48614836e-01 5.22687100e-02 -2.38049671e-01 -8.47995818e-01
-9.62458670e-01 2.42555775e-02 1.41441047e-01 -2.54059643e-01
-5.88228762e-01 -8.83388102e-01 -1.22338307e+00 3.54736187e-02
-4.68059957e-01 3.75442058e-01 1.00465202e+00 1.21763420e+00
2.82663852e-01 1.68800578e-01 1.06057346e+00 -3.80418003e-01
-3.52528453e-01 -1.04951370e+00 -7.28108108e-01 7.29827046e-01
-4.93577346e-02 -3.05905968e-01 -7.52192736e-01 -1.12415485e-01] | [11.314806938171387, 6.67863655090332] |
9009dc44-2f28-4d5d-969d-5eacae515b7a | araguaia-medical-vision-lab-at-isic-2017-skin | 1703.00856 | null | http://arxiv.org/abs/1703.00856v1 | http://arxiv.org/pdf/1703.00856v1.pdf | Araguaia Medical Vision Lab at ISIC 2017 Skin Lesion Classification Challenge | This paper describes the participation of Araguaia Medical Vision Lab at the
International Skin Imaging Collaboration 2017 Skin Lesion Challenge. We
describe the use of deep convolutional neural networks in attempt to classify
images of Melanoma and Seborrheic Keratosis lesions. With use of finetuned
GoogleNet and AlexNet we attained results of 0.950 and 0.846 AUC on Seborrheic
Keratosis and Melanoma respectively. | ['Larissa Vasconcellos de Moraes', 'Rafael Teixeira Sousa'] | 2017-03-02 | null | null | null | null | ['skin-lesion-classification'] | ['medical'] | [ 6.68234453e-02 -1.14906073e-01 -3.18728030e-01 6.59072772e-02
-3.15195560e-01 -4.76579219e-01 8.93302858e-01 2.14681197e-02
-6.69580460e-01 6.68447793e-01 1.77488074e-01 -4.27460194e-01
-9.15172175e-02 -6.40193701e-01 -2.80168746e-02 -5.70170045e-01
1.60784557e-01 -3.36707413e-01 4.30462323e-02 -1.19014367e-01
9.74462852e-02 6.43993497e-01 -1.02698064e+00 6.84431434e-01
6.31209075e-01 1.10462105e+00 -5.00809312e-01 1.53686178e+00
1.84163868e-01 4.09609139e-01 -5.51283240e-01 -4.48330313e-01
2.84867793e-01 6.25948831e-02 -8.72121811e-01 2.78810948e-01
1.17211461e+00 -2.01567724e-01 -4.34511811e-01 9.03618813e-01
5.57648957e-01 -4.95212913e-01 6.62777722e-01 -7.23096013e-01
-4.04258072e-01 -3.40156317e-01 -8.13464642e-01 9.16063786e-01
2.30253309e-01 4.80075061e-01 3.75815064e-01 -5.82781553e-01
1.00187814e+00 5.86434364e-01 1.20209110e+00 6.62744164e-01
-5.64534009e-01 -2.94747353e-01 -3.65728766e-01 2.81037152e-01
-1.35720873e+00 -2.93936580e-01 -3.15183342e-01 -4.25024480e-01
1.41128504e+00 5.70642650e-01 1.11646473e+00 9.07586575e-01
4.30358529e-01 6.07247651e-01 1.60441577e+00 -5.87451875e-01
-8.40734318e-02 1.37615576e-01 9.14478227e-02 1.07568741e+00
2.32281700e-01 2.29095936e-01 -2.53130823e-01 -1.77480131e-01
1.05789828e+00 -6.06242120e-02 -7.98925459e-02 5.59179127e-01
-5.89786649e-01 6.64152563e-01 8.88289511e-01 9.32798386e-02
-5.44280231e-01 3.21584940e-01 5.33294499e-01 4.50463474e-01
6.02747500e-01 1.02869995e-01 -1.14935059e-02 2.45910048e-01
-8.22570086e-01 -3.41522396e-01 3.97695959e-01 1.36809379e-01
2.87090298e-02 -2.00722843e-01 1.10745005e-01 1.00820673e+00
2.47406840e-01 -2.96813855e-03 5.03719687e-01 -1.76202878e-01
-1.87431976e-01 7.02565372e-01 -2.46897310e-01 -3.72471243e-01
-3.46390694e-01 -1.58075631e-01 -8.74070942e-01 6.06066942e-01
4.60009038e-01 -4.05576587e-01 -2.17743993e+00 6.30882263e-01
1.43967643e-01 4.68837559e-01 2.26157829e-01 8.91331613e-01
1.10398948e+00 -5.90761937e-02 5.91064394e-01 3.24016660e-01
1.26628900e+00 -1.04255855e+00 -4.54203278e-01 -4.17006277e-02
6.92591250e-01 -9.59665060e-01 5.52892745e-01 5.25310576e-01
-1.04190803e+00 -1.78780202e-02 -1.03695655e+00 -2.59382978e-05
-6.40394986e-01 4.17568207e-01 1.09110034e+00 1.08794475e+00
-1.58817840e+00 3.02159280e-01 -8.09294224e-01 -1.22822714e+00
9.04589236e-01 7.02773809e-01 -9.06536877e-01 -8.45757406e-03
-6.75439656e-01 1.21458733e+00 9.52216312e-02 2.14715227e-01
-6.34759545e-01 -8.07784498e-01 -4.83147115e-01 -7.23479807e-01
-2.70796746e-01 -9.91929352e-01 1.08821261e+00 -1.34833443e+00
-1.15658987e+00 1.61331034e+00 -3.23687568e-02 -7.08087981e-01
5.42410731e-01 2.78705079e-02 -8.37519825e-01 4.33201790e-01
-3.99218053e-01 6.65549278e-01 3.87518376e-01 -4.29695994e-01
-1.05227578e+00 -5.32521963e-01 -9.80182290e-02 2.08530203e-01
-2.22925320e-01 1.80824548e-01 -4.38979656e-01 -1.79179341e-01
-3.69266480e-01 -6.82870030e-01 -4.74281311e-01 4.67739016e-01
-8.05488765e-01 -1.11876234e-01 4.80257452e-01 -9.91213560e-01
7.54469931e-01 -1.72067142e+00 -4.41115946e-01 4.21930701e-01
5.02909839e-01 9.05929148e-01 -3.51153493e-01 2.00913459e-01
-3.05966139e-01 5.32609701e-01 4.49362218e-01 3.14504057e-02
-7.62802303e-01 -3.04949790e-01 4.71744090e-01 6.22516930e-01
2.88446277e-01 1.26569736e+00 -5.93011498e-01 -2.31050998e-01
5.25532961e-01 9.04299319e-01 3.44870031e-01 -4.30745482e-01
2.46741503e-01 -3.22939187e-01 -1.54361039e-01 1.27290380e+00
5.66309512e-01 -2.18079969e-01 5.56467194e-03 -2.80097425e-01
7.24127665e-02 -1.65622339e-01 -5.62449992e-01 1.38305664e+00
-4.94368792e-01 1.12551534e+00 9.96048674e-02 -1.69553813e-02
4.11526740e-01 4.06829745e-01 8.81292000e-02 -6.37398362e-01
3.04017484e-01 1.41060576e-01 1.38639033e-01 -5.19089997e-01
8.64911154e-02 -4.46331114e-01 1.07609415e+00 -2.76138969e-02
1.44197419e-01 1.93101138e-01 5.49125075e-02 8.69077668e-02
1.30372334e+00 -3.37043971e-01 6.68357909e-01 -2.01493800e-01
4.35506433e-01 2.56275266e-01 -2.52121061e-01 4.15171057e-01
-4.97547269e-01 6.56991363e-01 4.58273172e-01 -1.02326345e+00
-1.01635849e+00 -1.34204888e+00 -3.86780798e-01 3.79346907e-01
-3.25270146e-01 -3.90969992e-01 -7.26320744e-01 -1.02964032e+00
1.45970499e-02 -6.87095672e-02 -1.21536696e+00 1.23493657e-01
1.02536947e-01 -1.02575159e+00 9.33824360e-01 4.38020825e-01
5.78488469e-01 -7.14355886e-01 -1.65805727e-01 -2.18066111e-01
8.20916891e-01 -5.29582798e-01 -1.03131764e-01 -3.20659637e-01
-4.75667924e-01 -1.76898682e+00 -1.18782759e+00 -6.44035518e-01
8.25145125e-01 -4.40719649e-02 6.67742908e-01 2.27164492e-01
-1.51432943e+00 3.87249738e-01 4.05775942e-02 -6.35777295e-01
-1.40247777e-01 3.50007080e-02 -1.97606802e-01 8.65886919e-03
8.46689582e-01 -1.38355568e-01 -9.96302783e-01 -2.74959028e-01
-7.96842217e-01 -4.04059887e-03 7.45220780e-01 5.66600382e-01
5.49277008e-01 -2.54597723e-01 1.07516266e-01 -1.01821339e+00
9.21324611e-01 -4.90056455e-01 -2.99848169e-01 3.89882505e-01
-6.54421151e-01 -8.97445858e-01 -2.60812361e-02 -3.38107973e-01
-6.75877035e-01 1.57663122e-01 -2.13711202e-01 -2.93845266e-01
-3.54402661e-01 3.38308454e-01 9.46503282e-01 -8.21270764e-01
1.15305269e+00 -8.44632089e-02 3.44853222e-01 -6.43594339e-02
-3.91783044e-02 9.45056617e-01 4.40532386e-01 5.05643964e-01
2.21784964e-01 7.74143636e-01 2.66668946e-01 -1.32816041e+00
-5.87832808e-01 -7.48095453e-01 -3.17459524e-01 -2.63966799e-01
9.22621131e-01 -9.44586575e-01 -3.05363804e-01 1.07999861e+00
-8.95183623e-01 -2.28741348e-01 -1.62300378e-01 -5.26007079e-02
-2.45121457e-02 2.83397317e-01 -7.41011977e-01 -5.14056981e-01
-8.50610375e-01 -7.55706251e-01 7.79726326e-01 9.04852509e-01
-4.41359609e-01 -1.60422087e+00 4.06974703e-01 3.56559724e-01
6.53063357e-01 5.86617947e-01 2.49287814e-01 -3.92334700e-01
-6.59818351e-02 -1.02517879e+00 -4.99441564e-01 6.29631400e-01
4.18699861e-01 5.64206719e-01 -1.15226150e+00 -2.61685044e-01
-8.89543056e-01 -2.89366722e-01 1.19089520e+00 7.49275982e-01
1.02755821e+00 -5.00255078e-03 -4.79485095e-01 1.09978008e+00
1.83928585e+00 4.65402938e-02 1.00102615e+00 4.48284268e-01
4.31202948e-01 3.41243088e-01 4.76424657e-02 -1.12578183e-01
1.53126493e-02 -4.47573289e-02 3.48134160e-01 -6.79510295e-01
-7.22415447e-01 -7.26523697e-02 -4.79636788e-01 -4.96151894e-01
-5.68723559e-01 -2.36293614e-01 -1.34904468e+00 9.53479290e-01
-1.13952029e+00 -2.35363394e-01 -4.15486306e-01 1.74224305e+00
4.35419410e-01 -2.14198932e-01 7.86494985e-02 -2.11276963e-01
7.46122003e-01 -5.01453578e-02 -3.51542324e-01 -6.84253871e-01
-2.31288150e-01 1.00211918e+00 9.37320828e-01 6.22801661e-01
-1.20899069e+00 1.20673382e+00 7.41280270e+00 7.85232544e-01
-1.61602700e+00 1.78884715e-02 7.82906115e-01 -3.21089000e-01
1.59164414e-01 -3.95562530e-01 -5.38316965e-01 1.65996104e-01
9.95870650e-01 -2.21348956e-01 -3.29084806e-02 4.15913463e-01
-2.57368296e-01 -5.52807093e-01 -4.72895294e-01 8.68437946e-01
4.83102798e-02 -2.14464355e+00 8.97685289e-02 3.87168258e-01
1.07778084e+00 5.71857572e-01 4.30088222e-01 -5.45191348e-01
9.37403291e-02 -1.93326437e+00 -7.90815175e-01 9.18899000e-01
1.51199794e+00 -6.41031325e-01 1.00120199e+00 -4.82506067e-01
-4.42114681e-01 2.79343754e-01 -2.11835578e-01 1.12450600e-01
-4.23864394e-01 3.47769916e-01 -1.48796308e+00 -1.08503066e-01
5.27403116e-01 7.92262137e-01 -8.44786346e-01 1.76161110e+00
-2.82298118e-01 7.00524688e-01 -5.03550470e-01 -3.66386145e-01
6.87383354e-01 3.56125057e-01 3.37858737e-01 1.41427755e+00
-4.59301993e-02 -4.58278880e-03 -5.25873661e-01 1.55700669e-01
7.56583437e-02 3.37095886e-01 -5.73545814e-01 -1.94821671e-01
-3.20535377e-02 1.62915707e+00 -5.17839909e-01 -9.32674631e-02
-1.70334026e-01 1.07821167e+00 -3.18685114e-01 5.41927636e-01
-7.10565597e-02 -4.51885581e-01 8.63501012e-01 4.45745349e-01
-7.54151121e-02 3.18840921e-01 -6.03579640e-01 -7.57303357e-01
-4.21003193e-01 -5.45669794e-01 6.25082970e-01 -8.03681374e-01
-1.34914374e+00 7.22993195e-01 -9.43139970e-01 -5.95502496e-01
3.87766846e-02 -1.15870512e+00 -1.36377132e+00 1.47330832e+00
-1.78348720e+00 -1.80305290e+00 -6.60196841e-01 5.51227689e-01
3.32991838e-01 -6.54901981e-01 1.12233353e+00 -4.21845973e-01
-5.12612164e-01 7.75671661e-01 -6.16962500e-02 -2.59663835e-02
7.31481731e-01 -1.36765277e+00 4.27405357e-01 4.95266616e-01
-2.73174673e-01 5.39605916e-01 1.73931867e-01 -5.91551006e-01
-9.53160524e-01 -1.20642233e+00 5.54564595e-01 -4.53888655e-01
9.54726279e-01 4.28686470e-01 -1.48907259e-01 5.51231623e-01
7.94585049e-01 2.29397044e-01 1.37039518e+00 8.36616009e-02
-4.41561729e-01 2.30486140e-01 -1.51397288e+00 5.14302254e-01
3.31732810e-01 -4.35585290e-01 -4.27641608e-02 8.54322255e-01
-1.29259467e-01 -4.82876003e-01 -1.02830458e+00 2.31684625e-01
7.61211514e-01 -8.27376783e-01 8.82790506e-01 -1.12267590e+00
6.32795870e-01 2.26578280e-01 9.56408754e-02 -1.15767753e+00
-1.77662242e-02 -5.69990158e-01 9.93280634e-02 3.03719580e-01
4.45782453e-01 -6.63315535e-01 1.34674251e+00 2.16020361e-01
1.32150009e-01 -1.23899293e+00 -9.33211565e-01 -1.24405123e-01
3.50452542e-01 2.65589207e-02 -1.55747116e-01 7.07355738e-01
-2.99629986e-01 1.36687923e-02 1.23900272e-01 -7.86867365e-03
6.25849903e-01 -7.33035803e-01 4.40016031e-01 -6.97149336e-01
2.60935485e-01 -5.09449244e-01 -1.16849053e+00 1.18899457e-01
-3.39155614e-01 -9.03110981e-01 -8.62100601e-01 -1.81575096e+00
2.15919778e-01 -6.31461665e-02 -7.75414109e-01 6.55271053e-01
1.43813686e-02 1.13226676e+00 -9.23864618e-02 -2.19069883e-01
-1.47360608e-01 -6.50799394e-01 1.01199138e+00 -2.09723726e-01
-1.18472511e-02 2.68003523e-01 -8.38281393e-01 8.94015074e-01
1.16585159e+00 3.86530340e-01 -8.75135288e-02 -3.49806398e-01
1.09687626e-01 -2.92528272e-01 8.42374504e-01 -1.02868998e+00
4.86755371e-01 -1.56885430e-01 9.30488169e-01 -3.04130852e-01
4.57401484e-01 -1.50665283e-01 -1.19700991e-01 8.72133255e-01
-1.81084588e-01 -5.63755453e-01 5.89865565e-01 3.93059731e-01
-1.20900385e-01 2.58406196e-02 1.11152935e+00 -2.60112017e-01
-1.15756774e+00 6.85328245e-01 -5.97600222e-01 -6.85025513e-01
1.50459898e+00 -7.07488239e-01 -6.84483230e-01 -1.05965175e-02
-1.14881086e+00 1.80584386e-01 5.49081802e-01 2.47818038e-01
8.85060012e-01 -9.77042377e-01 -8.98036659e-01 1.48697898e-01
2.81474411e-01 -5.37215650e-01 4.87510413e-01 9.01116729e-01
-1.30095172e+00 4.08595681e-01 -6.60210609e-01 -4.29675400e-01
-1.92259407e+00 -4.10472572e-01 1.19809747e+00 -1.59522351e-02
-4.51312006e-01 1.59662354e+00 -4.10735518e-01 -4.47327569e-02
1.08101420e-01 1.74201399e-01 -9.61845368e-02 -7.33255744e-02
8.20626616e-01 3.43543977e-01 1.04121678e-01 -3.24048400e-02
-3.94859433e-01 5.60787141e-01 -6.94229305e-01 1.77138150e-02
1.02440321e+00 4.17160302e-01 -3.32687616e-01 -1.35728553e-01
1.13049388e+00 -3.57812434e-01 -7.08803713e-01 1.85306549e-01
-3.17948520e-01 -4.96833533e-01 6.62893295e-01 -1.93533611e+00
-1.01687777e+00 9.74811256e-01 1.36057746e+00 1.66163146e-02
1.19257700e+00 -2.40205377e-01 5.48952758e-01 1.48569271e-01
-1.96722046e-01 -9.37643290e-01 -2.36631900e-01 3.45020965e-02
5.34318566e-01 -1.29197669e+00 1.00402020e-01 -6.04808331e-01
-6.26567185e-01 1.51719058e+00 7.35543072e-01 -6.82183564e-01
7.17934370e-01 4.17212993e-01 6.87963724e-01 -3.86989146e-01
-8.82880092e-01 -4.56750095e-01 5.74218690e-01 1.04729378e+00
4.20537502e-01 3.55763793e-01 -4.91534442e-01 -8.04637894e-02
2.25270212e-01 4.55639273e-01 6.73825324e-01 6.47553921e-01
-3.89408141e-01 -8.50379169e-01 1.18187144e-01 1.19819891e+00
-8.37885737e-01 -2.34337091e-01 -1.24475932e+00 1.16869020e+00
2.34556571e-01 3.63228142e-01 2.61206537e-01 -2.60800838e-01
-1.16176317e-02 -1.49385139e-01 6.07987761e-01 -6.74405873e-01
-9.71957862e-01 5.34334667e-02 5.17529607e-01 -4.65556711e-01
-2.12004215e-01 -2.70127714e-01 -7.38871276e-01 -3.44746947e-01
-1.26613304e-01 -3.60585392e-01 1.22474289e+00 5.75693965e-01
5.44761680e-02 3.14402640e-01 1.41021535e-01 1.24066278e-01
-3.23820226e-02 -1.06195140e+00 -9.07756567e-01 -2.38585658e-02
5.59318125e-01 -1.45525439e-02 -1.72600448e-01 -1.59605637e-01] | [15.714887619018555, -3.0264394283294678] |
03064a15-0ecb-4a0d-aa71-e04a481e0c6e | meta-style-adversarial-training-for-cross | 2302.09309 | null | https://arxiv.org/abs/2302.09309v2 | https://arxiv.org/pdf/2302.09309v2.pdf | StyleAdv: Meta Style Adversarial Training for Cross-Domain Few-Shot Learning | Cross-Domain Few-Shot Learning (CD-FSL) is a recently emerging task that tackles few-shot learning across different domains. It aims at transferring prior knowledge learned on the source dataset to novel target datasets. The CD-FSL task is especially challenged by the huge domain gap between different datasets. Critically, such a domain gap actually comes from the changes of visual styles, and wave-SAN empirically shows that spanning the style distribution of the source data helps alleviate this issue. However, wave-SAN simply swaps styles of two images. Such a vanilla operation makes the generated styles ``real'' and ``easy'', which still fall into the original set of the source styles. Thus, inspired by vanilla adversarial learning, a novel model-agnostic meta Style Adversarial training (StyleAdv) method together with a novel style adversarial attack method is proposed for CD-FSL. Particularly, our style attack method synthesizes both ``virtual'' and ``hard'' adversarial styles for model training. This is achieved by perturbing the original style with the signed style gradients. By continually attacking styles and forcing the model to recognize these challenging adversarial styles, our model is gradually robust to the visual styles, thus boosting the generalization ability for novel target datasets. Besides the typical CNN-based backbone, we also employ our StyleAdv method on large-scale pretrained vision transformer. Extensive experiments conducted on eight various target datasets show the effectiveness of our method. Whether built upon ResNet or ViT, we achieve the new state of the art for CD-FSL. Code is available at https://github.com/lovelyqian/StyleAdv-CDFSL. | ['Yu-Gang Jiang', 'Yanwei Fu', 'Yu Xie', 'Yuqian Fu'] | 2023-02-18 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Fu_StyleAdv_Meta_Style_Adversarial_Training_for_Cross-Domain_Few-Shot_Learning_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Fu_StyleAdv_Meta_Style_Adversarial_Training_for_Cross-Domain_Few-Shot_Learning_CVPR_2023_paper.pdf | cvpr-2023-1 | ['cross-domain-few-shot', 'cross-domain-few-shot-learning'] | ['computer-vision', 'computer-vision'] | [ 3.55849236e-01 -1.87990487e-01 1.11944102e-01 -2.81679988e-01
-6.08846128e-01 -8.99944127e-01 6.62852705e-01 -5.45219004e-01
-2.59148002e-01 6.36859417e-01 -3.43808122e-02 -1.30523210e-02
3.01804483e-01 -8.58970106e-01 -9.52961385e-01 -6.03176892e-01
4.50852722e-01 3.44716251e-01 4.81213272e-01 -6.90076590e-01
-3.49904038e-02 3.91135395e-01 -1.18431199e+00 2.42106423e-01
7.82714486e-01 9.57163751e-01 1.53198570e-01 6.14928901e-01
-3.07720333e-01 7.36687899e-01 -8.74481380e-01 -5.68063021e-01
6.51995122e-01 -5.91897249e-01 -5.62696338e-01 -1.74208183e-03
6.50872588e-01 -4.10264194e-01 -5.33088744e-01 1.29784858e+00
5.06310582e-01 3.83595049e-01 5.22733331e-01 -1.42800426e+00
-1.20241153e+00 1.43204764e-01 -5.75122774e-01 7.51401633e-02
2.47224495e-01 7.45795131e-01 6.60312057e-01 -9.56873000e-01
8.63009512e-01 1.27136540e+00 6.21180534e-01 1.30156898e+00
-1.42140365e+00 -9.19417739e-01 2.42592037e-01 1.97504833e-01
-1.07354665e+00 -2.70693839e-01 1.26181579e+00 -3.93426746e-01
4.04929638e-01 1.59185797e-01 5.48199236e-01 1.80904496e+00
-1.02821514e-01 8.18358421e-01 1.33305633e+00 -2.55070537e-01
5.11772215e-01 2.03990012e-01 -2.94210225e-01 3.09374481e-01
-9.94641557e-02 4.53015655e-01 -2.93675601e-01 -3.27238790e-03
7.34280527e-01 1.67879179e-01 -3.89094651e-01 -5.59448183e-01
-1.00197899e+00 6.61546886e-01 6.28921449e-01 1.24585532e-01
7.79376701e-02 2.55751852e-02 5.92045009e-01 5.85828006e-01
4.82233971e-01 5.94191611e-01 -4.79786813e-01 1.08838648e-01
-7.01701701e-01 2.26471439e-01 5.54638028e-01 1.12619984e+00
8.61138463e-01 4.95190471e-01 -3.25917661e-01 1.08167303e+00
-1.47164851e-01 6.12107635e-01 4.44966465e-01 -8.25126171e-01
1.90832838e-01 5.32889903e-01 -7.81554505e-02 -8.10694277e-01
9.16475207e-02 -2.24156916e-01 -1.01173067e+00 5.79801142e-01
4.36810941e-01 -3.75165313e-01 -1.25720584e+00 2.03709626e+00
2.79729843e-01 4.71001029e-01 1.15207145e-02 6.57079875e-01
9.57213640e-01 6.26636922e-01 7.57936090e-02 1.09293774e-01
1.12101138e+00 -9.95003223e-01 -3.94851416e-01 -5.72825611e-01
1.16905712e-01 -6.63940489e-01 1.48364878e+00 1.29301488e-01
-8.96670699e-01 -8.84154499e-01 -1.07531309e+00 1.20102078e-01
-6.08911097e-01 -4.66431528e-01 3.22444022e-01 4.50862288e-01
-8.58377755e-01 7.21475601e-01 -3.92837405e-01 -3.19047391e-01
8.33304703e-01 -1.37229562e-01 -3.41741353e-01 -3.22399944e-01
-1.31924403e+00 6.53582394e-01 3.12262267e-01 -2.30859444e-01
-9.92892683e-01 -1.11797714e+00 -9.82048452e-01 -1.59622058e-01
4.70070839e-01 -5.91714382e-01 1.21737254e+00 -1.49601805e+00
-1.55022609e+00 9.27774787e-01 1.31597251e-01 -2.17223793e-01
8.51665080e-01 -8.39128494e-02 -6.78155184e-01 8.96145999e-02
8.29064772e-02 7.06473589e-01 1.42245936e+00 -1.57221270e+00
-3.41980577e-01 -1.30457178e-01 1.43090129e-01 -1.03009142e-01
-3.39161128e-01 -1.27350852e-01 -3.37712407e-01 -1.17963517e+00
-5.47684968e-01 -9.39377964e-01 -1.68013468e-01 4.76360083e-01
-2.37187505e-01 6.08812599e-03 1.04128480e+00 -4.89052624e-01
8.57571363e-01 -2.41373730e+00 2.57635921e-01 -8.47834796e-02
2.89907992e-01 7.64848590e-01 -5.00369549e-01 3.46375883e-01
-2.47998342e-01 2.50850730e-02 -4.73708600e-01 -1.31498963e-01
-1.60078444e-02 3.75017077e-01 -6.59718812e-01 1.85434088e-01
4.29944605e-01 1.16747749e+00 -1.14046514e+00 -1.49372697e-01
2.30070472e-01 2.58821964e-01 -5.07863462e-01 5.04125655e-01
-4.31932777e-01 6.53599083e-01 -1.21080503e-01 6.35694146e-01
9.65957105e-01 -4.14857566e-02 -1.29880354e-01 -1.93933219e-01
2.51103938e-01 -4.32184458e-01 -1.03761184e+00 1.90833366e+00
-3.82448524e-01 4.21402454e-01 -1.38263330e-01 -8.32365990e-01
1.11197042e+00 1.12808354e-01 1.54596031e-01 -7.65490592e-01
1.97201014e-01 5.94214313e-02 -1.75400719e-01 -1.92648962e-01
8.40729997e-02 -5.39163351e-01 -2.71682560e-01 2.32317179e-01
3.75294536e-01 -4.50971186e-01 3.27344760e-02 2.29712144e-01
1.06484210e+00 2.63154536e-01 1.81385413e-01 -9.42151770e-02
4.51413184e-01 -8.86464268e-02 9.44329023e-01 7.71166623e-01
-7.41994739e-01 7.93415248e-01 4.53042924e-01 -6.11602247e-01
-1.21673453e+00 -1.49401820e+00 1.65701926e-01 1.19156408e+00
2.71387756e-01 -1.06777415e-01 -7.72955179e-01 -1.08820903e+00
2.47517657e-02 7.29198217e-01 -1.04782104e+00 -6.36505783e-01
-4.67129469e-01 -1.88441128e-01 6.12281680e-01 6.07286811e-01
8.30418408e-01 -1.30131888e+00 -2.42050692e-01 2.05035776e-01
1.72354639e-01 -1.05791259e+00 -6.71698570e-01 -4.45946045e-02
-4.32057321e-01 -1.09274936e+00 -1.10480273e+00 -9.05524492e-01
5.49269080e-01 2.17627272e-01 1.12444949e+00 -1.41838923e-01
-1.70989186e-01 2.74650842e-01 -5.29824495e-01 -4.14024234e-01
-6.32602930e-01 -2.83936828e-01 6.83177710e-02 2.46628836e-01
1.74752355e-01 -7.31489301e-01 -7.35953867e-01 3.19333404e-01
-1.21915841e+00 -1.59316137e-02 3.95840496e-01 1.11375797e+00
4.54450369e-01 -1.60862610e-01 6.52031600e-01 -1.19160163e+00
5.22502303e-01 -4.01430130e-01 -3.30670714e-01 2.55210102e-01
-2.98747122e-01 -1.04024857e-01 1.22612596e+00 -9.30654168e-01
-1.20971417e+00 -2.06449345e-01 -1.75823390e-01 -1.10712361e+00
-4.96221125e-01 -2.74262249e-01 -4.37885433e-01 -1.77789986e-01
9.56707954e-01 3.82065475e-01 -3.69790122e-02 -4.66654032e-01
5.97459018e-01 2.43782908e-01 7.54920721e-01 -5.30443549e-01
1.46517789e+00 6.64724648e-01 -3.03664744e-01 -5.68228960e-01
-8.25611532e-01 5.22369100e-03 -5.22411704e-01 -9.13270190e-02
8.37605476e-01 -8.16846192e-01 -1.46768749e-01 9.69511747e-01
-9.81712103e-01 -6.81438208e-01 -7.11923182e-01 -3.37841243e-01
-6.13069415e-01 3.71384859e-01 -5.02548158e-01 -2.02561006e-01
-3.50928098e-01 -1.05291486e+00 8.20381284e-01 3.33573222e-01
-2.47633636e-01 -1.18682611e+00 3.38755280e-01 4.84532192e-02
5.56914389e-01 6.09086275e-01 9.17982399e-01 -6.14925385e-01
-5.53515591e-02 -9.60074291e-02 -2.97602117e-01 7.28056729e-01
3.97453904e-01 -1.55944571e-01 -1.07675827e+00 -5.81142247e-01
-1.60762705e-02 -5.76252937e-01 7.55869269e-01 -1.37638077e-01
1.17365658e+00 -2.63503343e-01 1.18502483e-01 1.03267479e+00
1.55163193e+00 3.30347061e-01 7.43060708e-01 2.61307955e-01
1.05551195e+00 2.90434062e-01 3.77186567e-01 2.48963431e-01
3.32018286e-02 5.28680086e-01 2.45603263e-01 -3.18701595e-01
-4.06769753e-01 -4.74647820e-01 4.65971857e-01 4.34564352e-01
1.02631435e-01 7.47655332e-02 -7.19158292e-01 5.27270496e-01
-1.52766275e+00 -1.16160560e+00 5.04667997e-01 1.86150968e+00
1.00276041e+00 3.60429257e-01 9.36287940e-02 -1.22345448e-01
7.53633499e-01 6.61122680e-01 -1.08246922e+00 -7.32753098e-01
-1.79261506e-01 5.98881721e-01 2.43441135e-01 9.67845395e-02
-1.14329624e+00 1.23515558e+00 5.00185013e+00 1.08263302e+00
-1.31823123e+00 1.56465083e-01 5.06988108e-01 -2.00938791e-01
-3.65466863e-01 -4.78198044e-02 -4.79873836e-01 7.53995478e-01
3.35802168e-01 -2.92492568e-01 6.89968526e-01 1.13004589e+00
-1.78125292e-01 4.39840972e-01 -9.26000714e-01 9.53329325e-01
1.72571450e-01 -1.36795664e+00 3.43643695e-01 -3.63425612e-01
1.05588377e+00 -3.38472612e-02 3.40274632e-01 7.66862571e-01
5.31676710e-01 -6.72378480e-01 6.80056036e-01 3.30176651e-01
1.38517749e+00 -7.10214734e-01 2.44386017e-01 9.43884552e-02
-1.19885516e+00 -1.16744198e-01 -4.62698907e-01 7.14663565e-02
4.17311816e-03 2.68279761e-01 -3.34672362e-01 3.46855551e-01
7.26081014e-01 9.07624125e-01 -8.16670716e-01 5.94787538e-01
-3.61103386e-01 3.11673939e-01 2.19961151e-01 3.59203994e-01
2.24493206e-01 3.55281420e-02 8.16838145e-01 1.01401472e+00
1.37357349e-02 3.45501862e-02 1.06826976e-01 1.13055027e+00
-3.35774213e-01 -2.44580463e-01 -9.81054962e-01 7.54642263e-02
4.06192511e-01 1.11246347e+00 -4.48548883e-01 -5.65092504e-01
-5.28894424e-01 1.45097446e+00 4.00962710e-01 6.33346796e-01
-9.20855522e-01 -5.59474707e-01 1.21054184e+00 -1.57070354e-01
4.64747548e-01 1.17903136e-01 -2.92301089e-01 -1.49026072e+00
-5.47727160e-02 -1.23821902e+00 2.83123612e-01 -6.80600107e-01
-1.79141355e+00 7.87758887e-01 -1.93303898e-01 -1.62160122e+00
7.19050765e-02 -6.17459655e-01 -1.17780113e+00 7.84540355e-01
-1.47805893e+00 -1.31171870e+00 -4.80956674e-01 9.95330036e-01
1.02045608e+00 -4.15210605e-01 7.07487404e-01 5.78272827e-02
-4.80188608e-01 9.46942151e-01 1.70749605e-01 2.93900520e-01
1.08587098e+00 -1.31782901e+00 9.70876038e-01 1.02314341e+00
-1.73132882e-01 3.72334361e-01 7.38064528e-01 -7.14114189e-01
-1.25113690e+00 -1.39979601e+00 1.80684090e-01 -4.80025887e-01
7.85802126e-01 -5.27559459e-01 -1.31911421e+00 6.62320435e-01
2.63364166e-01 3.49056691e-01 6.44707918e-01 -4.57178652e-01
-8.53246331e-01 -2.08171159e-01 -1.31850743e+00 8.72957587e-01
1.29489899e+00 -5.31792760e-01 -7.91196465e-01 8.53001475e-02
9.03867662e-01 -3.20160002e-01 -6.27378285e-01 4.30437565e-01
4.13975954e-01 -9.68802392e-01 1.07995307e+00 -9.21339989e-01
7.19102263e-01 -1.43726557e-01 -1.73409492e-01 -1.75273418e+00
-5.63873649e-01 -5.85887790e-01 -2.12652072e-01 1.36280715e+00
3.16593833e-02 -6.34252965e-01 6.44501925e-01 5.15179634e-01
1.40849026e-02 -6.53037012e-01 -6.86380088e-01 -1.10117471e+00
5.54696441e-01 -2.77466804e-01 7.78705895e-01 1.14699745e+00
-5.69316328e-01 6.91813752e-02 -5.04978240e-01 -9.22661722e-02
8.58348966e-01 5.33999540e-02 1.02761436e+00 -9.51253593e-01
-6.18003309e-01 -4.81222510e-01 -3.44381392e-01 -5.94413817e-01
3.20360899e-01 -7.50785708e-01 -3.70813124e-02 -1.17350817e+00
-5.77671081e-02 -3.31760317e-01 -4.59783226e-01 6.03016078e-01
-4.78156239e-01 4.86700624e-01 6.45674348e-01 6.53181747e-02
-2.97515839e-01 7.58217692e-01 1.69379735e+00 -4.76287663e-01
-9.21709761e-02 -9.88720730e-02 -8.17015052e-01 8.38982105e-01
8.99320424e-01 -3.29153389e-01 -5.29645979e-01 -4.01337028e-01
-2.76227798e-02 -3.86081070e-01 5.10893285e-01 -1.07413769e+00
-1.42751962e-01 -3.76400530e-01 6.24894321e-01 9.81602538e-03
3.27886820e-01 -6.74450755e-01 4.96631442e-03 4.61572081e-01
-1.51656896e-01 -1.55147508e-01 4.06179219e-01 6.24851525e-01
-1.71999596e-02 6.83035105e-02 1.38159788e+00 -3.23321164e-01
-1.36261261e+00 5.55522501e-01 1.25406444e-01 5.50426602e-01
1.21853161e+00 -1.87795386e-01 -4.66337442e-01 -1.51213139e-01
-8.41035485e-01 1.34049028e-01 8.80748689e-01 7.08184540e-01
7.36167669e-01 -1.64994061e+00 -5.51476896e-01 5.32023311e-01
5.01094222e-01 5.39767817e-02 6.11736953e-01 3.03078264e-01
-3.27568740e-01 -4.12215233e-01 -7.05574095e-01 -2.69349366e-01
-9.02876854e-01 1.00081229e+00 3.26775879e-01 -5.17496541e-02
-7.41383374e-01 1.08009827e+00 6.63794696e-01 -5.28108299e-01
-5.28050326e-02 1.06603421e-01 7.20483512e-02 8.52478668e-03
7.18781173e-01 2.73862362e-01 -3.58996242e-01 -4.73786771e-01
-2.45955259e-01 6.00722134e-01 -3.01020294e-01 6.41164184e-02
1.25019264e+00 8.00265670e-02 2.00184390e-01 4.95064080e-01
1.34148896e+00 -1.47121996e-01 -1.91305900e+00 -5.35501301e-01
-4.28911448e-01 -6.09421015e-01 -3.97295594e-01 -8.85916829e-01
-1.12935972e+00 9.41007257e-01 6.73621178e-01 -4.52645794e-02
1.32206583e+00 -1.67984426e-01 1.10934603e+00 2.36071162e-02
3.38272184e-01 -1.05601478e+00 4.20174032e-01 5.95062792e-01
1.01000416e+00 -1.44307470e+00 -4.65898156e-01 -5.55650219e-02
-9.21754241e-01 8.41534734e-01 1.02159667e+00 -5.49016535e-01
5.29186726e-01 -5.18631600e-02 3.54512155e-01 8.81166160e-02
-4.82770294e-01 -9.81523320e-02 2.30439261e-01 9.55552101e-01
-2.94573128e-01 -1.36322491e-02 2.16950670e-01 5.51070988e-01
-1.47759616e-01 -2.34650169e-02 4.24002141e-01 9.30967927e-01
-1.84853658e-01 -1.19553828e+00 -3.49929482e-01 2.68386126e-01
-4.81888093e-02 -8.25012382e-03 -5.44238448e-01 6.12233937e-01
2.94090927e-01 6.57050610e-01 -4.43172790e-02 -5.89134932e-01
6.06624305e-01 9.30589363e-02 4.80006337e-01 -7.36457348e-01
-4.97779459e-01 -2.46879399e-01 -4.31102425e-01 -6.33122921e-01
7.35623986e-02 -3.47161114e-01 -8.56505632e-01 -4.72922534e-01
3.25356394e-01 -2.97176093e-01 1.06134117e-01 8.46598983e-01
3.31838280e-01 6.65346384e-01 9.72320139e-01 -1.00406885e+00
-6.27172947e-01 -6.95640028e-01 -4.84310955e-01 1.03729665e+00
4.74676520e-01 -7.91017890e-01 -4.50136006e-01 7.40870759e-02] | [10.07422924041748, 2.6877658367156982] |
fba7d059-db31-4507-9144-5e11c786cbad | audiolm-a-language-modeling-approach-to-audio | 2209.03143 | null | https://arxiv.org/abs/2209.03143v1 | https://arxiv.org/pdf/2209.03143v1.pdf | AudioLM: a Language Modeling Approach to Audio Generation | We introduce AudioLM, a framework for high-quality audio generation with long-term consistency. AudioLM maps the input audio to a sequence of discrete tokens and casts audio generation as a language modeling task in this representation space. We show how existing audio tokenizers provide different trade-offs between reconstruction quality and long-term structure, and we propose a hybrid tokenization scheme to achieve both objectives. Namely, we leverage the discretized activations of a masked language model pre-trained on audio to capture long-term structure and the discrete codes produced by a neural audio codec to achieve high-quality synthesis. By training on large corpora of raw audio waveforms, AudioLM learns to generate natural and coherent continuations given short prompts. When trained on speech, and without any transcript or annotation, AudioLM generates syntactically and semantically plausible speech continuations while also maintaining speaker identity and prosody for unseen speakers. Furthermore, we demonstrate how our approach extends beyond speech by generating coherent piano music continuations, despite being trained without any symbolic representation of music. | ['Neil Zeghidour', 'Marco Tagliasacchi', 'David Grangier', 'Olivier Teboul', 'Matt Sharifi', 'Olivier Pietquin', 'Eugene Kharitonov', 'Damien Vincent', 'Raphaël Marinier', 'Zalán Borsos'] | 2022-09-07 | null | null | null | null | ['audio-generation'] | ['audio'] | [ 2.38450065e-01 4.60263193e-01 2.53751129e-02 -3.08918685e-01
-1.65010130e+00 -8.65055382e-01 4.99178946e-01 -7.48133436e-02
1.54723912e-01 4.24331963e-01 7.70474315e-01 4.98030260e-02
3.18967015e-01 -4.29638863e-01 -8.36751878e-01 -2.88982183e-01
-2.90162712e-01 2.31758133e-01 -2.15916157e-01 3.69878625e-03
-1.18746348e-01 -1.54972291e-02 -1.73741162e+00 7.41142750e-01
4.46746886e-01 1.03477311e+00 9.41662937e-02 1.12378967e+00
-1.67872503e-01 9.62992609e-01 -9.38107073e-01 -2.54695535e-01
1.84489310e-01 -6.45651698e-01 -8.99709404e-01 2.29930915e-02
3.99888933e-01 -2.98427254e-01 -1.59913078e-01 7.45831907e-01
6.61289811e-01 1.99229091e-01 5.06960809e-01 -1.02172399e+00
-5.68016887e-01 1.57024777e+00 1.25506252e-01 -4.24742162e-01
8.39244783e-01 1.90145284e-01 1.46124220e+00 -1.08800685e+00
5.58799267e-01 1.39112949e+00 7.68384576e-01 6.93705142e-01
-1.60650754e+00 -8.99377108e-01 3.51380371e-02 -1.89414695e-01
-1.56636178e+00 -1.10285854e+00 8.43871713e-01 -3.08239371e-01
8.95056486e-01 2.37544656e-01 6.05908394e-01 1.43490112e+00
-2.22563580e-01 7.59706795e-01 3.88037413e-01 -5.72066188e-01
3.22413355e-01 -1.96010947e-01 -3.76745164e-01 3.22424352e-01
-7.70657361e-01 1.83819637e-01 -1.14810133e+00 -4.04726982e-01
7.07115829e-01 -6.82095945e-01 -3.10033947e-01 3.02968532e-01
-1.42569995e+00 5.21535337e-01 -3.90665904e-02 2.90962249e-01
-4.07485366e-01 6.92528546e-01 5.74867785e-01 5.11379480e-01
1.32485554e-01 6.58793211e-01 -2.03888580e-01 -3.94600064e-01
-1.51677656e+00 4.30124670e-01 7.28854299e-01 1.03953063e+00
4.30590123e-01 9.15032089e-01 -2.85220325e-01 9.59078908e-01
2.02834353e-01 5.63749015e-01 7.68955290e-01 -1.18342471e+00
3.28262538e-01 -3.47496420e-01 9.63319093e-02 -5.44465899e-01
9.38916504e-02 -4.92656887e-01 -4.58977818e-01 -3.48272286e-02
9.34945196e-02 -2.28046626e-01 -6.99250698e-01 2.17141056e+00
1.20171774e-02 6.03192329e-01 2.78677642e-01 6.43155813e-01
6.85631812e-01 1.13330734e+00 -2.11392473e-02 -5.83803058e-02
1.19902468e+00 -8.69184434e-01 -6.65593863e-01 -3.26807201e-02
3.95328343e-01 -9.85381663e-01 1.37046063e+00 6.36761248e-01
-1.70935738e+00 -9.17235732e-01 -9.81930435e-01 -1.56346038e-01
2.25491107e-01 1.79421127e-01 1.99262835e-02 5.05001664e-01
-1.30588138e+00 4.93438005e-01 -7.17716992e-01 2.84676254e-01
-1.32660151e-01 1.12205610e-01 -5.66794574e-02 4.64623481e-01
-1.27432525e+00 1.80223376e-01 3.33498955e-01 -1.50282279e-01
-1.50765932e+00 -9.44665790e-01 -9.91456568e-01 2.79933691e-01
1.17785949e-02 -6.12930715e-01 2.00166512e+00 -1.29889762e+00
-1.85208714e+00 4.46354955e-01 -2.46069491e-01 -8.49575281e-01
1.69231772e-01 -1.06271014e-01 -6.49604023e-01 4.39450055e-01
1.50310159e-01 9.97535229e-01 1.17280948e+00 -1.06611192e+00
-5.73396087e-01 4.27754611e-01 -1.73517391e-01 6.53079674e-02
-2.53319919e-01 2.83163190e-01 -1.80109844e-01 -1.21245885e+00
-1.32043645e-01 -8.50385249e-01 -2.90758461e-02 -1.81896389e-01
-6.24430358e-01 -7.21543953e-02 4.94203717e-01 -7.28760481e-01
1.17752957e+00 -2.33719802e+00 1.79358736e-01 3.04357737e-01
-1.55528590e-01 -6.02488592e-02 -6.14563584e-01 5.53369284e-01
-2.12646231e-01 1.61349416e-01 -1.90978080e-01 -7.10557461e-01
3.95022750e-01 -8.41311812e-02 -1.23655033e+00 -1.24625072e-01
5.07957280e-01 8.05163383e-01 -1.00014901e+00 -4.06899810e-01
-1.33542120e-01 8.06804478e-01 -9.06961501e-01 5.35702765e-01
-6.45465374e-01 6.02139175e-01 -4.34183851e-02 5.11416137e-01
1.56655647e-02 2.14974657e-01 1.35608628e-01 8.25571790e-02
-1.10823438e-01 1.03506422e+00 -1.24734950e+00 2.23301244e+00
-8.57082546e-01 4.89567846e-01 2.95225263e-01 -6.19094610e-01
9.53750014e-01 1.16392243e+00 3.51044416e-01 -4.21511918e-01
-8.88082907e-02 4.31231499e-01 -3.33465874e-01 -2.41882578e-01
6.67756557e-01 -5.10744214e-01 -3.94059628e-01 7.78378546e-01
3.98955792e-01 -5.20583332e-01 -2.18107238e-01 1.58387810e-01
1.03835332e+00 1.15373470e-01 -5.38072996e-02 2.48325348e-01
2.15940014e-01 -4.28500354e-01 2.61974186e-01 7.64430821e-01
4.40593392e-01 9.84393895e-01 2.26426274e-01 9.24032480e-02
-1.03397119e+00 -1.42600441e+00 1.94441929e-01 1.43836164e+00
-5.22279203e-01 -8.19600999e-01 -8.65652621e-01 -1.26477629e-02
-2.66340733e-01 8.19760919e-01 -1.19577229e-01 -2.15531856e-01
-7.74464548e-01 1.92757562e-01 1.40318513e+00 4.81236041e-01
-2.00327888e-01 -1.27507961e+00 -2.06510127e-01 6.52009189e-01
-5.71903169e-01 -1.02723527e+00 -8.37258160e-01 1.63947433e-01
-5.33155501e-01 -4.65560555e-01 -7.48031378e-01 -1.05662346e+00
3.73910815e-02 -2.41387084e-01 1.28333294e+00 -1.62982702e-01
-8.16590860e-02 3.37934047e-01 -3.84977043e-01 -2.45836824e-01
-9.30230558e-01 8.43770653e-02 3.29564631e-01 1.91095769e-01
-4.89165425e-01 -1.02343357e+00 -1.72289416e-01 -8.80826637e-02
-9.56233680e-01 6.97752014e-02 3.33583713e-01 7.80317664e-01
6.07128680e-01 -1.80578575e-01 1.16868925e+00 -4.11403835e-01
7.07373798e-01 -2.98029006e-01 -3.74201685e-01 -4.98025380e-02
-4.94725220e-02 1.81206718e-01 9.08182323e-01 -7.33906746e-01
-9.10495996e-01 1.34144470e-01 -5.58054090e-01 -6.47932112e-01
-6.55061007e-03 5.97276926e-01 -1.52986512e-01 4.55942035e-01
8.62212300e-01 4.01627570e-01 -3.25054705e-01 -5.62469423e-01
7.70900965e-01 8.02311957e-01 1.23680675e+00 -1.04041076e+00
7.59805620e-01 1.42595276e-01 -5.62837958e-01 -8.85476232e-01
-7.17189014e-01 -1.14129215e-01 -2.69969404e-01 -5.20230383e-02
5.36029160e-01 -1.28600585e+00 -5.87975144e-01 1.99647486e-01
-1.28776646e+00 -4.58639652e-01 -8.42406034e-01 4.53508705e-01
-9.92439926e-01 8.55946764e-02 -8.34111989e-01 -1.01057661e+00
-5.43536663e-01 -8.86581004e-01 1.34657621e+00 -4.84398305e-01
-7.20493078e-01 -6.09218001e-01 1.01825051e-01 -1.75583005e-01
4.38361764e-01 1.19922809e-01 8.67419541e-01 -4.15520012e-01
-5.75811327e-01 -5.59289753e-02 3.29903007e-01 4.09800202e-01
6.61580563e-02 -2.82886866e-02 -1.43399072e+00 -1.50107458e-01
-7.17827752e-02 -7.73243546e-01 6.44739687e-01 2.45217577e-01
1.07802010e+00 -8.52023542e-01 1.98822573e-01 6.40784979e-01
7.76571453e-01 -1.37783969e-02 4.05495167e-01 -3.85341614e-01
4.23739254e-01 5.72521389e-01 2.82475263e-01 5.56329608e-01
1.92076579e-01 8.01197290e-01 6.94712996e-02 6.59943968e-02
-6.20750666e-01 -9.12486017e-01 9.76824224e-01 1.33446252e+00
3.58082652e-01 -9.34287440e-03 -6.98840797e-01 8.22695553e-01
-1.46950555e+00 -1.31139278e+00 4.28314686e-01 2.06658316e+00
1.55353844e+00 1.01622164e-01 3.71034771e-01 4.99198437e-01
5.92125237e-01 1.23456270e-01 -2.62482196e-01 -4.86423135e-01
-3.26434970e-02 8.65650773e-01 -2.11388126e-01 6.49097323e-01
-7.12144434e-01 9.96615410e-01 7.17108250e+00 9.93557513e-01
-1.46343482e+00 1.96578354e-01 1.79041281e-01 -5.05881965e-01
-7.82793701e-01 1.05021439e-01 -5.16946197e-01 2.27139369e-01
1.62551904e+00 -3.85215938e-01 7.10491419e-01 6.53642893e-01
4.35342014e-01 6.82352066e-01 -1.50972867e+00 8.64939749e-01
-4.93810773e-02 -1.56297398e+00 3.59687746e-01 -4.23078507e-01
6.65744960e-01 -1.77729741e-01 3.53505611e-01 4.29248780e-01
4.99708742e-01 -1.35567749e+00 1.79014122e+00 2.58327782e-01
1.33226597e+00 -6.76477849e-01 -6.38240203e-02 2.78500259e-01
-1.65619040e+00 -1.19705804e-01 7.66842887e-02 -4.40150462e-02
4.88482744e-01 3.25184196e-01 -1.19861186e+00 3.80244821e-01
2.70753175e-01 5.43114841e-01 -9.82918218e-02 7.33998954e-01
-2.93121397e-01 1.13137281e+00 -2.71280110e-01 5.07142961e-01
1.89147234e-01 4.28935319e-01 6.11498058e-01 1.45468712e+00
6.35511577e-01 -1.84516191e-01 4.30441171e-01 1.13569045e+00
-2.64853448e-01 -2.20844429e-02 -4.68672544e-01 -4.07891154e-01
9.05287862e-01 8.40925574e-01 -1.29543170e-01 -3.59242082e-01
-1.36974035e-02 8.17976892e-01 -3.83077785e-02 4.83340830e-01
-7.50228405e-01 -5.28149962e-01 6.24589443e-01 1.58699647e-01
1.97307676e-01 -3.13366979e-01 -9.18518528e-02 -1.10092652e+00
-2.96926815e-02 -1.11137009e+00 1.80291250e-01 -1.00695741e+00
-1.00283074e+00 9.52952027e-01 -2.65730590e-01 -1.34581184e+00
-1.16492546e+00 7.16774762e-02 -6.34779871e-01 9.71975923e-01
-1.27009118e+00 -1.49715841e+00 3.10753226e-01 5.39488375e-01
7.99580455e-01 -1.28951237e-01 1.30770576e+00 2.31847391e-01
6.98617250e-02 8.34538996e-01 -3.49123389e-01 2.35170022e-01
7.64119685e-01 -1.10704792e+00 7.32433081e-01 6.20630145e-01
7.15213358e-01 6.22238457e-01 8.38093758e-01 -3.40077102e-01
-1.10792613e+00 -1.37727141e+00 1.09328377e+00 -2.61103511e-01
7.87356794e-01 -6.77414060e-01 -8.73262763e-01 8.77134502e-01
3.13528478e-01 -1.46486342e-01 8.11990380e-01 1.12530170e-02
-8.76787126e-01 3.82584892e-02 -6.74516499e-01 6.23646975e-01
8.69026661e-01 -1.29584050e+00 -7.76291490e-01 -1.32225826e-03
1.34428334e+00 -4.69189018e-01 -7.19540656e-01 -6.51134849e-02
6.59477830e-01 -5.50695717e-01 1.07348359e+00 -5.33831537e-01
4.91154671e-01 -3.68437380e-01 -3.28541398e-01 -1.20468020e+00
1.55635983e-01 -1.37457943e+00 2.36209799e-02 1.63672984e+00
6.42994702e-01 -1.20487772e-01 3.60611945e-01 -2.58562472e-02
-6.12225831e-01 -2.14556322e-01 -1.01289678e+00 -8.90621781e-01
6.50172606e-02 -1.10788810e+00 9.06288743e-01 7.34594584e-01
3.17250341e-01 2.61961251e-01 -6.77464545e-01 2.64586389e-01
3.01969379e-01 2.52601296e-01 7.63167620e-01 -8.02900672e-01
-7.38126099e-01 -2.24624902e-01 -2.04357971e-02 -1.13175821e+00
5.42703211e-01 -1.26414692e+00 4.96975094e-01 -9.13040459e-01
-4.09765899e-01 -5.18980086e-01 -1.31743550e-01 6.92571580e-01
4.05799657e-01 3.83084744e-01 2.86990196e-01 1.80920437e-01
-4.39937174e-01 7.23845780e-01 6.87954843e-01 -2.04174504e-01
-4.77394491e-01 1.70685519e-02 -6.20480955e-01 6.19699240e-01
4.82419312e-01 -6.83309674e-01 -4.90398288e-01 -4.90032315e-01
1.82784408e-01 6.53105259e-01 4.57800299e-01 -1.18757379e+00
2.80391932e-01 -2.91679669e-02 4.61294949e-02 -3.64445627e-01
7.89983332e-01 -2.81465739e-01 4.70576912e-01 1.01783782e-01
-1.11798322e+00 -1.82331726e-01 1.96585953e-01 3.76799941e-01
-6.81511164e-01 -1.70558959e-01 5.84099233e-01 -5.47067970e-02
-5.04953116e-02 4.01536673e-02 -6.82892978e-01 3.15080613e-01
2.42371529e-01 4.77356277e-02 2.03495204e-01 -7.55542517e-01
-9.75788534e-01 -1.66887224e-01 2.56297559e-01 4.60775882e-01
7.00620711e-01 -1.71568799e+00 -7.94992983e-01 3.93748999e-01
1.12894543e-01 -4.23631184e-02 -1.04263248e-02 1.01655930e-01
-5.08910678e-02 3.45453143e-01 1.55583724e-01 -4.65460241e-01
-1.01245654e+00 1.20278232e-01 1.84270799e-01 1.04493044e-01
-6.90972567e-01 9.96144176e-01 1.17364116e-01 -2.53903031e-01
6.47840619e-01 -6.93797469e-01 2.57885665e-01 4.30246443e-02
7.32058942e-01 -2.41205841e-01 -2.37030033e-02 -6.28501177e-01
-1.11143157e-01 2.96663761e-01 4.40350622e-01 -1.05939913e+00
1.19443548e+00 -2.86328513e-03 1.54677242e-01 7.52430856e-01
1.08365798e+00 6.96959734e-01 -1.19803441e+00 -2.20913142e-01
7.79128894e-02 -1.05316572e-01 -1.68442786e-01 -6.52051926e-01
-6.98601484e-01 9.50081825e-01 9.98614170e-03 1.67592943e-01
1.01556122e+00 4.71200496e-02 1.16472781e+00 3.59793872e-01
3.23849112e-01 -9.60404992e-01 5.95750332e-01 6.46317244e-01
1.14805377e+00 -3.97841185e-01 -7.57674158e-01 -8.60565826e-02
-7.06312954e-01 1.09332538e+00 -6.74127564e-02 -1.10480368e-01
5.00795424e-01 6.06204450e-01 2.47517928e-01 2.02949360e-01
-1.16656482e+00 -1.46763548e-02 3.12023491e-01 6.97377741e-01
6.56328738e-01 3.61026078e-02 4.85294700e-01 1.12004542e+00
-1.08630025e+00 -6.18151799e-02 5.10458767e-01 5.78488827e-01
-4.70113158e-01 -1.26889849e+00 -5.38685620e-01 -1.57312691e-01
-5.99177897e-01 -5.06269157e-01 -4.29023176e-01 7.73720518e-02
4.05016802e-02 1.10680938e+00 9.30134803e-02 -4.75285709e-01
1.19284675e-01 5.55141687e-01 2.14874521e-01 -9.92734373e-01
-8.84791851e-01 4.36939120e-01 2.65651971e-01 -5.30316353e-01
-6.93263039e-02 -5.63123882e-01 -1.67900836e+00 8.51823166e-02
-1.41415805e-01 4.68174160e-01 5.95718324e-01 6.06476128e-01
3.78164262e-01 6.95293605e-01 8.56459558e-01 -1.05002785e+00
-8.38595927e-01 -7.74767458e-01 -5.57014942e-01 1.58784568e-01
7.74844289e-01 8.21433365e-02 -3.45530868e-01 7.45509088e-01] | [15.584821701049805, 5.803701877593994] |
97307b7e-30fa-4817-b215-d2ad2c2b1939 | does-localization-inform-editing-surprising | 2301.04213 | null | https://arxiv.org/abs/2301.04213v1 | https://arxiv.org/pdf/2301.04213v1.pdf | Does Localization Inform Editing? Surprising Differences in Causality-Based Localization vs. Knowledge Editing in Language Models | Language models are known to learn a great quantity of factual information during pretraining, and recent work localizes this information to specific model weights like mid-layer MLP weights (Meng et al., 2022). In this paper, we find that we can change how a fact is stored in a model by editing weights that are in a different location than where existing methods suggest that the fact is stored. This is surprising because we would expect that localizing facts to specific parameters in models would tell us where to manipulate knowledge in models, and this assumption has motivated past work on model editing methods. Specifically, we show that localization conclusions from representation denoising (also known as Causal Tracing) do not provide any insight into which model MLP layer would be best to edit in order to override an existing stored fact with a new one. This finding raises questions about how past work relies on Causal Tracing to select which model layers to edit (Meng et al., 2022). Next, to better understand the discrepancy between representation denoising and weight editing, we develop several variants of the editing problem that appear more and more like representation denoising in their design and objective. Experiments show that, for one of our editing problems, editing performance does relate to localization results from representation denoising, but we find that which layer we edit is a far better predictor of performance. Our results suggest, counterintuitively, that better mechanistic understanding of how pretrained language models work may not always translate to insights about how to best change their behavior. Code is available at: https://github.com/google/belief-localization | ['Asma Ghandeharioun', 'Been Kim', 'Mohit Bansal', 'Peter Hase'] | 2023-01-10 | null | null | null | null | ['model-editing'] | ['natural-language-processing'] | [ 2.51729459e-01 3.65152538e-01 -2.67555326e-01 -5.66293120e-01
-4.24552649e-01 -7.00647295e-01 7.73200154e-01 2.62804896e-01
-5.69978237e-01 7.79835522e-01 5.54939806e-01 -5.47255039e-01
-1.51433989e-01 -1.03323925e+00 -1.15918386e+00 -4.95766103e-01
2.53246278e-01 2.92554557e-01 3.31142731e-02 -1.20417312e-01
5.72386384e-01 3.93328220e-01 -1.43971968e+00 5.89118659e-01
5.80813587e-01 4.36742961e-01 3.34902018e-01 4.25115407e-01
-1.49134278e-01 9.86266553e-01 -6.90634966e-01 -5.96890390e-01
-6.60708025e-02 -5.75856984e-01 -9.01519597e-01 -3.28696400e-01
5.06690443e-01 -2.05176368e-01 -1.04570456e-01 1.11957920e+00
8.55423808e-02 1.84734151e-01 7.35241473e-01 -9.28550065e-01
-1.02016449e+00 1.27483010e+00 -3.26072425e-01 5.10058641e-01
-4.06952500e-02 1.54023513e-01 1.04149544e+00 -5.66882312e-01
7.48549402e-01 1.41109264e+00 7.89080799e-01 4.65368003e-01
-1.50540459e+00 -6.75573051e-01 6.78613067e-01 1.64099690e-02
-1.28834534e+00 -5.99714577e-01 8.22632611e-01 -5.49565613e-01
9.23572183e-01 2.35033080e-01 7.15692997e-01 1.28801250e+00
4.14292336e-01 5.52472174e-01 1.18766403e+00 -5.24821043e-01
4.88022491e-02 2.99172342e-01 1.79870054e-01 1.00492036e+00
7.14209497e-01 1.78980812e-01 -7.52866983e-01 -3.15398484e-01
8.86736453e-01 -2.98641056e-01 -3.54601979e-01 -8.59844908e-02
-1.05275869e+00 9.48876858e-01 2.84417003e-01 5.90544403e-01
-2.40996897e-01 8.81425202e-01 1.87603757e-01 5.76563776e-01
2.42096469e-01 8.90526474e-01 -7.65447855e-01 -1.07562467e-01
-8.51419806e-01 2.10313305e-01 5.98982871e-01 4.82794195e-01
8.71029496e-01 2.20411912e-01 6.34380206e-02 8.44620705e-01
5.60265958e-01 9.43790227e-02 5.43083549e-01 -1.32814062e+00
2.29375243e-01 2.51106322e-01 -1.73296761e-02 -1.25220156e+00
-2.79744834e-01 -5.48688173e-01 -3.00561398e-01 2.09242463e-01
5.88643849e-01 -2.29813427e-01 -7.17211187e-01 2.33232880e+00
-3.56393635e-01 3.77011113e-02 -1.26201734e-01 6.41065359e-01
3.58632803e-01 4.66824561e-01 2.37295225e-01 4.75843102e-02
1.22163570e+00 -3.34457666e-01 -4.51877981e-01 -6.26091778e-01
9.66315091e-01 -7.51069188e-01 1.15055656e+00 4.85750169e-01
-1.12488496e+00 -4.42592055e-01 -1.11012423e+00 -1.41877398e-01
-5.81033111e-01 -1.36842489e-01 1.10872984e+00 7.63574898e-01
-1.09315991e+00 9.90175784e-01 -9.45638001e-01 -4.89469826e-01
4.06291336e-01 2.06948370e-01 -1.77481905e-01 1.48749620e-01
-1.40964901e+00 1.37786436e+00 3.99388760e-01 6.03863969e-02
-7.56331503e-01 -8.59810352e-01 -7.66478658e-01 9.68343690e-02
3.53773296e-01 -9.99018610e-01 1.25865948e+00 -1.38954771e+00
-1.11874342e+00 8.21009517e-01 -4.22432125e-01 -4.51717317e-01
3.28910425e-02 -5.40680205e-03 -3.30321312e-01 -2.58039713e-01
-8.61067176e-02 5.79592705e-01 7.92390227e-01 -1.76387644e+00
-5.99425733e-01 -3.73517215e-01 4.38742071e-01 -1.13051534e-02
-1.39597863e-01 -9.69536719e-04 -1.70122832e-01 -8.86503041e-01
2.80635923e-01 -6.18419766e-01 5.67887202e-02 1.25527710e-01
-1.11320958e-01 -1.17260970e-01 2.18570217e-01 -4.81262922e-01
1.19892406e+00 -1.87271321e+00 -7.78208002e-02 3.02008003e-01
2.89190590e-01 -1.05623588e-01 -8.23947638e-02 4.26432937e-01
-2.73507357e-01 7.95501590e-01 -3.15026015e-01 -2.65579909e-01
1.94425866e-01 4.03439373e-01 -7.32238114e-01 3.61189425e-01
1.08008057e-01 7.81778216e-01 -8.49599361e-01 -3.80943447e-01
-5.90268224e-02 5.45180619e-01 -8.01747382e-01 -4.40649331e-01
-3.29820186e-01 7.20428377e-02 -2.81289726e-01 2.73339629e-01
4.30321306e-01 -5.66518344e-02 2.44893163e-01 -2.52459019e-01
-7.91724324e-02 7.19900906e-01 -1.13225329e+00 1.49412334e+00
-5.22168458e-01 7.13518381e-01 -2.54670139e-02 -1.16318107e+00
6.10683560e-01 1.66282281e-01 -1.01305861e-02 -4.83856678e-01
1.90530300e-01 1.03011064e-01 4.36587304e-01 -3.73741418e-01
4.89433378e-01 -6.37086928e-01 1.19383380e-01 7.89721906e-01
5.53543307e-02 -2.47164413e-01 1.33977860e-01 2.74231136e-01
1.04576600e+00 2.23134339e-01 -2.25297455e-02 -3.49434078e-01
-5.28190285e-02 5.19359335e-02 7.94371426e-01 1.11123180e+00
1.03685729e-01 4.33920115e-01 8.70745718e-01 -1.49986923e-01
-7.26476610e-01 -9.97981012e-01 -2.13163868e-01 1.07156456e+00
-2.83596516e-01 -3.06155413e-01 -4.78776306e-01 -4.89652842e-01
9.09531116e-02 1.57058692e+00 -9.46940839e-01 -5.53974211e-01
-6.05892301e-01 -6.69538915e-01 7.86000490e-01 6.22013450e-01
1.97041750e-01 -1.03548193e+00 -6.92937970e-01 1.53733388e-01
8.43802541e-02 -3.60923260e-01 -2.90711164e-01 4.94766891e-01
-1.22066689e+00 -8.99942100e-01 -2.21413881e-01 -5.55653691e-01
8.80854309e-01 -3.89856957e-02 1.22444057e+00 4.63902235e-01
3.47925991e-01 5.32092810e-01 -1.64562464e-03 -5.91730952e-01
-6.60069525e-01 3.93606126e-02 -1.44790784e-01 -3.01651835e-01
3.13663870e-01 -7.59798944e-01 -3.12260717e-01 9.12524238e-02
-1.14998972e+00 7.29805604e-03 5.03915071e-01 7.10154891e-01
2.42898539e-01 2.21149921e-01 5.22609949e-01 -1.52215159e+00
9.33661759e-01 -5.28991163e-01 -4.62600768e-01 4.01166379e-01
-7.86819518e-01 5.84787488e-01 5.06681561e-01 -3.54502261e-01
-1.23858166e+00 -4.01814312e-01 -1.71923116e-02 -1.39808550e-01
-2.35157311e-02 9.97307479e-01 5.50663546e-02 2.20698550e-01
9.47053909e-01 1.28024161e-01 -4.45126109e-02 -3.61319095e-01
4.03781474e-01 -3.41198519e-02 1.06149942e-01 -1.19553256e+00
7.05307066e-01 4.41020697e-01 -4.89909232e-01 -3.10023695e-01
-9.64157879e-01 2.32136056e-01 -3.96909833e-01 1.35661468e-01
4.64138657e-01 -6.71485245e-01 -3.84421319e-01 1.39178470e-01
-1.28479683e+00 -7.93591917e-01 -3.94473433e-01 4.08580363e-01
-4.64160919e-01 -1.96015392e-03 -4.72873628e-01 -5.26437640e-01
3.73679936e-01 -9.99326229e-01 3.79023850e-01 -9.53723565e-02
-7.09689081e-01 -1.43255186e+00 -6.40695095e-02 1.98034659e-01
4.94821429e-01 -6.99475929e-02 1.48103809e+00 -4.69447851e-01
-3.45540285e-01 6.91766143e-02 2.06270684e-02 3.34861934e-01
1.01866961e-01 2.74316728e-01 -1.01236296e+00 1.18594885e-01
2.89371103e-01 -1.33088484e-01 1.27005088e+00 3.54999334e-01
1.35016346e+00 -4.01053250e-01 -3.01585823e-01 4.06851739e-01
1.37394226e+00 2.64291078e-01 7.41541803e-01 2.06434056e-01
4.80586499e-01 6.17031217e-01 1.38559937e-01 -3.46992575e-02
3.90334129e-01 4.69510198e-01 1.39787728e-02 1.93485737e-01
-1.87825665e-01 -5.27990639e-01 4.82367337e-01 5.81888199e-01
-5.99209368e-02 -1.29746735e-01 -8.50680232e-01 6.49205327e-01
-1.79116976e+00 -1.22877181e+00 2.25524202e-01 2.25210238e+00
1.47964835e+00 4.99816477e-01 -3.43431860e-01 -1.07544862e-01
4.53661472e-01 3.44678611e-01 -3.31118107e-01 -7.15688050e-01
-9.57565904e-02 1.40629143e-01 4.96567041e-01 9.68185425e-01
-5.22325933e-01 1.08301520e+00 6.16723347e+00 7.05784082e-01
-1.07588565e+00 2.72537708e-01 6.61168993e-01 -4.25113589e-02
-9.75799263e-01 2.82243878e-01 -6.67106986e-01 3.62228006e-01
9.60275531e-01 -2.70323604e-01 6.26749396e-01 4.41988826e-01
2.96435565e-01 -3.93968582e-01 -1.43625546e+00 3.46017569e-01
1.53750435e-01 -1.51862514e+00 3.77159864e-01 -5.76554760e-02
7.04744995e-01 -1.81348652e-01 1.56846449e-01 3.56122613e-01
7.06649005e-01 -1.16753769e+00 1.01632750e+00 8.64869416e-01
8.91769305e-02 -6.28839672e-01 3.77816737e-01 2.77532607e-01
-6.75423503e-01 6.79093003e-02 -5.25646508e-01 -2.22930774e-01
3.35845277e-02 6.85474157e-01 -6.52280092e-01 -5.77253290e-02
3.41213584e-01 3.53833795e-01 -6.37791395e-01 7.42565691e-01
-6.14487708e-01 8.84191632e-01 -3.00504178e-01 1.87391385e-01
-2.72816196e-02 8.43016282e-02 3.62354934e-01 1.21074009e+00
1.87491745e-01 6.12096377e-02 -4.48563576e-01 1.34313774e+00
-2.25430921e-01 -3.21841449e-01 -8.18603158e-01 -2.19615802e-01
5.11055827e-01 6.41885996e-01 -6.74402952e-01 -3.26834738e-01
-3.51422101e-01 6.17537320e-01 4.39321548e-01 6.15487456e-01
-8.51109445e-01 -2.13192955e-01 6.94444001e-01 2.60127068e-01
2.39647835e-01 -2.88733214e-01 -7.60399461e-01 -1.06404090e+00
-1.39249876e-01 -8.35494220e-01 1.99857056e-01 -9.41752672e-01
-1.08899534e+00 3.72923985e-02 2.51607209e-01 -3.29352677e-01
-1.36314094e-01 -6.07861340e-01 -6.14683390e-01 9.07096207e-01
-1.28424335e+00 -7.56899595e-01 4.46458697e-01 2.81620830e-01
3.71156245e-01 2.26474389e-01 8.32898438e-01 1.62865631e-02
-3.90554041e-01 6.12920105e-01 -1.53555572e-01 1.33225918e-01
7.32395768e-01 -1.11043704e+00 2.04007864e-01 7.76184559e-01
6.07932270e-01 1.57757998e+00 8.84797812e-01 -7.50099123e-01
-1.20727265e+00 -7.65028775e-01 9.63939726e-01 -7.74942756e-01
7.37904012e-01 -1.66296393e-01 -1.18899357e+00 1.14124441e+00
2.97207147e-01 -2.68394947e-01 5.09518445e-01 6.06373429e-01
-6.55841470e-01 -4.19011600e-02 -1.03762531e+00 8.72281671e-01
1.12590647e+00 -5.82768500e-01 -9.96184051e-01 1.33333340e-01
6.43869579e-01 -1.16480485e-01 -6.05790973e-01 7.12299123e-02
4.59513992e-01 -8.60329449e-01 7.74125457e-01 -7.63563573e-01
4.94822383e-01 -3.96242738e-01 -2.96977192e-01 -1.66956210e+00
-3.40245008e-01 -1.51669025e-01 1.07580021e-01 1.24755168e+00
1.06723309e+00 -1.09729719e+00 5.79461515e-01 1.01014006e+00
-1.90937072e-01 -6.71812177e-01 -7.19236374e-01 -5.37139356e-01
5.88819444e-01 -7.37961590e-01 3.78429294e-01 1.33842862e+00
-1.92517921e-01 2.44731337e-01 1.09926037e-01 2.14613006e-01
3.17071974e-01 -1.51679352e-01 2.93614388e-01 -1.14594781e+00
-5.21058381e-01 -7.70691991e-01 -3.60621586e-02 -9.04029369e-01
3.90897572e-01 -1.12931621e+00 -1.09498203e-01 -1.73060191e+00
6.17823489e-02 -5.99965334e-01 -4.35486555e-01 9.19428527e-01
-3.76128927e-02 -3.56748067e-02 2.32703790e-01 5.72265051e-02
-6.00617565e-02 1.95075274e-01 1.01469743e+00 -1.36035085e-01
-4.70007630e-03 -1.64352953e-01 -1.37667036e+00 1.08299732e+00
1.06652367e+00 -7.62218177e-01 -6.96846962e-01 -9.13686872e-01
7.40602732e-01 -3.50674331e-01 6.06608152e-01 -7.02043235e-01
3.22563797e-01 -4.49242681e-01 4.82471704e-01 3.55934794e-03
3.42704952e-01 -7.14948416e-01 1.88506037e-01 4.36670929e-01
-6.81418538e-01 1.00646913e-01 6.02874875e-01 4.19686168e-01
9.25748646e-02 -8.09297919e-01 7.07300544e-01 -5.09530485e-01
-7.03529298e-01 -3.71135712e-01 -6.25583768e-01 1.85284957e-01
3.95974606e-01 -2.58147240e-01 -6.20077550e-01 -4.67575163e-01
-7.73456454e-01 -5.60909547e-02 4.24686670e-01 2.63024777e-01
4.43760335e-01 -9.72127020e-01 -5.16910911e-01 -1.35429919e-01
-1.79549083e-01 -2.99308270e-01 -1.07291445e-01 8.12110543e-01
-1.90993428e-01 2.66671509e-01 1.60552114e-01 -4.46595103e-02
-7.36132801e-01 3.41747999e-01 5.31855404e-01 3.86097701e-03
-3.01448256e-01 1.13097548e+00 1.65092796e-01 -4.13020104e-01
1.25873253e-01 -6.98058307e-01 1.25387788e-01 2.96570331e-01
2.70173252e-01 -2.70927120e-02 8.04521665e-02 -1.06796734e-01
-1.76729053e-01 4.22643304e-01 -4.06706244e-01 -3.99529159e-01
1.33692598e+00 7.95910060e-02 -1.98498577e-01 9.34845448e-01
8.51899147e-01 2.44667828e-01 -1.24991429e+00 -1.37411296e-01
-2.70664133e-02 -4.20193762e-01 2.50000298e-01 -1.22415185e+00
-1.16327059e+00 9.34085906e-01 2.01871797e-01 2.40308359e-01
7.11163223e-01 5.46623468e-02 1.07463792e-01 6.95961177e-01
3.31616193e-01 -1.16180456e+00 -2.45640740e-01 4.77873921e-01
1.32116413e+00 -9.68502641e-01 2.14884579e-01 -6.77583227e-03
-6.08912289e-01 8.66172671e-01 5.97624719e-01 -1.78651854e-01
7.41196930e-01 1.79085091e-01 -2.59807911e-02 -3.96704108e-01
-1.00267684e+00 -2.72961170e-03 -1.42429993e-01 5.43763459e-01
7.40218639e-01 -7.95956850e-02 -2.55812049e-01 6.56071723e-01
-4.47919220e-01 -3.81817222e-02 5.87525368e-01 8.79463196e-01
-5.52701652e-01 -1.20112419e+00 -4.23187673e-01 7.50229239e-01
-2.47884184e-01 -4.58515316e-01 -4.72450078e-01 8.64108801e-01
2.56475031e-01 8.22347581e-01 3.31664421e-02 -1.07838504e-01
2.79093385e-01 5.54662645e-01 6.88777685e-01 -1.07402921e+00
-7.37476826e-01 -3.02294284e-01 3.10678929e-01 -3.67042214e-01
-2.98539221e-01 -7.24232078e-01 -1.50563765e+00 -5.25188386e-01
-1.19149312e-01 -4.40283632e-03 5.90181351e-01 9.84381616e-01
3.41974229e-01 6.13410413e-01 -6.11217879e-02 -4.15483534e-01
-4.16466653e-01 -8.41988266e-01 -4.66041565e-01 1.91351473e-02
1.99366093e-01 -8.24691236e-01 -6.09367669e-01 1.98838413e-01] | [9.993653297424316, 7.66853666305542] |
6aae7729-b384-40a6-bf43-e8dd257aa85b | two-stage-voice-anonymization-for-enhanced | 2306.16069 | null | https://arxiv.org/abs/2306.16069v1 | https://arxiv.org/pdf/2306.16069v1.pdf | Two-Stage Voice Anonymization for Enhanced Privacy | In recent years, the need for privacy preservation when manipulating or storing personal data, including speech , has become a major issue. In this paper, we present a system addressing the speaker-level anonymization problem. We propose and evaluate a two-stage anonymization pipeline exploiting a state-of-the-art anonymization model described in the Voice Privacy Challenge 2022 in combination with a zero-shot voice conversion architecture able to capture speaker characteristics from a few seconds of speech. We show this architecture can lead to strong privacy preservation while preserving pitch information. Finally, we propose a new compressed metric to evaluate anonymization systems in privacy scenarios with different constraints on privacy and utility. | ['Patrick A. Naylor', 'Joerg Bitzer', 'Daniel Barreda', 'Francesco Nespoli'] | 2023-06-28 | null | null | null | null | ['voice-conversion', 'voice-conversion'] | ['audio', 'speech'] | [ 1.49333164e-01 2.70624489e-01 1.72789708e-01 -6.84480429e-01
-1.16888535e+00 -9.79566038e-01 4.09920931e-01 2.94375211e-01
-5.58132648e-01 4.49323922e-01 8.36359680e-01 -9.41244960e-02
4.03543524e-02 -3.19690198e-01 -4.37713861e-01 -1.15496002e-01
-6.90621361e-02 1.29089862e-01 -1.63119957e-01 -3.07201631e-02
-2.10299462e-01 7.93499231e-01 -1.40981233e+00 3.81088674e-01
6.49095058e-01 8.98289442e-01 -7.77961612e-01 6.46872997e-01
2.20936581e-01 7.23635554e-02 -7.80152023e-01 -1.03224468e+00
7.99540281e-01 -1.81801111e-01 -8.16123307e-01 -5.48246920e-01
6.35655344e-01 -3.92444938e-01 -5.65689504e-01 9.89390850e-01
1.03520405e+00 9.93510112e-02 -1.59439564e-01 -1.32476699e+00
-2.58760899e-01 7.81268418e-01 1.07619427e-01 1.75810475e-02
5.97515106e-01 2.37922251e-01 9.21985090e-01 -2.33627990e-01
6.19001925e-01 1.18439925e+00 1.06271434e+00 7.71135449e-01
-1.62717772e+00 -6.36882663e-01 -1.40345097e-01 1.05527133e-01
-1.63497078e+00 -1.12923431e+00 6.48400903e-01 -3.44973393e-02
9.75911558e-01 1.09412479e+00 5.65057576e-01 1.44530797e+00
-2.73727417e-01 5.86697936e-01 7.75516272e-01 -3.16087604e-02
2.41128922e-01 5.65781176e-01 3.05306315e-02 -2.71512419e-02
2.88382918e-01 1.79246053e-01 -9.82446313e-01 -7.28742778e-01
-6.91979229e-02 -3.47965121e-01 -4.33683634e-01 -2.96200246e-01
-7.95592785e-01 4.76127058e-01 -7.79976547e-02 1.31566465e-01
-1.89547598e-01 9.00286064e-02 7.48975277e-01 5.61728239e-01
4.30058599e-01 7.17485368e-01 -2.29869202e-01 -3.67752522e-01
-1.24640441e+00 4.58526313e-01 1.39602721e+00 1.26714540e+00
5.37662864e-01 -1.49696469e-01 -4.92338866e-01 4.91941214e-01
5.23252189e-02 3.25793326e-01 4.55517918e-01 -1.06270087e+00
6.08723700e-01 -5.83266728e-02 1.25102270e-02 -1.00131297e+00
-2.00310707e-01 -2.81202704e-01 -6.38928235e-01 -2.40401745e-01
1.79397702e-01 -2.88950086e-01 -3.77928585e-01 1.94866407e+00
2.48069152e-01 2.71790564e-01 2.96893358e-01 7.27913022e-01
4.99215394e-01 2.51704425e-01 9.04288888e-02 -1.72816977e-01
1.51693654e+00 -5.26532888e-01 -1.08228910e+00 3.57282311e-02
2.51394838e-01 -5.55492103e-01 1.07352173e+00 2.62342840e-01
-1.04686344e+00 1.90457627e-02 -9.89219129e-01 -4.26697463e-01
-3.75503778e-01 -3.12397033e-01 4.83279288e-01 1.84636974e+00
-1.18898642e+00 7.51266360e-01 -7.95195639e-01 -6.37485385e-01
3.42682093e-01 6.63034797e-01 -7.48358071e-01 4.75731969e-01
-1.53668976e+00 4.65471834e-01 1.07172042e-01 -2.95701742e-01
-4.68608618e-01 -1.13222110e+00 -7.42446542e-01 4.12750423e-01
4.17553186e-02 -7.88423538e-01 1.23776376e+00 -4.12675023e-01
-1.70052004e+00 9.66900110e-01 -1.56348065e-01 -1.10500813e+00
1.00541532e+00 -3.90742064e-01 -1.05941737e+00 9.94312242e-02
-1.06518842e-01 3.97884637e-01 8.55638444e-01 -9.20628846e-01
-5.17290890e-01 -5.76897204e-01 -2.43940413e-01 1.33999586e-01
-8.57498109e-01 2.02711701e-01 -5.03638864e-01 -8.79950881e-01
-2.62340307e-01 -8.93919170e-01 -6.61382079e-02 -1.91718023e-02
-6.98786676e-01 6.87089801e-01 9.50899243e-01 -1.20678556e+00
1.62133276e+00 -2.50869393e+00 -1.25816777e-01 4.06853437e-01
9.11974236e-02 2.91866243e-01 1.23685554e-01 4.53269511e-01
-7.35237151e-02 5.56142926e-01 -4.98190612e-01 -1.07210541e+00
3.78914446e-01 -5.26103303e-02 -6.29514933e-01 5.95985711e-01
-1.93643823e-01 7.00245500e-01 -4.89552319e-01 -9.07551646e-02
1.58680707e-01 5.41598976e-01 -8.74430358e-01 2.41747245e-01
1.66630194e-01 5.43362021e-01 -3.40716797e-04 4.54453975e-01
1.10780847e+00 6.88347757e-01 3.95244718e-01 -7.62539059e-02
-5.62259257e-02 6.03957176e-01 -1.28379488e+00 2.14051080e+00
-4.36005682e-01 5.29892445e-01 6.29435718e-01 -1.52409568e-01
6.54084265e-01 5.82301855e-01 6.09199345e-01 -4.62749898e-01
1.06659591e-01 1.65988415e-04 -4.78088379e-01 -3.02854449e-01
7.24686205e-01 -3.13006751e-02 -6.02126300e-01 4.89645839e-01
3.91594507e-02 -1.74126089e-01 -3.87546748e-01 2.34744158e-02
1.04557514e+00 -5.45890808e-01 3.52961391e-01 -2.97499955e-01
5.28472841e-01 -7.32567787e-01 8.29795182e-01 6.15285277e-01
-5.44032156e-01 8.57580125e-01 4.34705496e-01 -8.57564807e-02
-8.38432670e-01 -1.03478169e+00 -1.01173796e-01 6.99578762e-01
-2.41056696e-01 -8.97324145e-01 -1.25112891e+00 -5.91598868e-01
4.61515546e-01 1.15778494e+00 -5.22500455e-01 -3.76250505e-01
-5.04545212e-01 -3.36814940e-01 1.55698097e+00 5.57858609e-02
4.70607132e-01 -4.91813034e-01 -3.82075280e-01 -6.20966516e-02
-2.30381846e-01 -1.30193484e+00 -9.80238497e-01 -2.31091641e-02
-4.26091909e-01 -5.03425121e-01 -2.09433883e-01 -1.01788998e-01
-6.94804937e-02 -3.22210491e-02 8.28395069e-01 -5.70576966e-01
-2.13071048e-01 6.38828218e-01 -1.52587160e-01 -5.76118112e-01
-5.10933161e-01 6.80330396e-01 4.64504570e-01 5.50312281e-01
5.15507400e-01 -1.00199783e+00 -4.55053329e-01 1.23264045e-01
-8.28285158e-01 -6.81987405e-01 4.45399731e-02 2.12555170e-01
4.43699241e-01 -1.98317528e-01 4.15723681e-01 -9.46212649e-01
1.08361149e+00 -3.37760478e-01 -3.92097831e-01 2.89710850e-01
-6.54153585e-01 1.43259913e-01 8.06209147e-01 -8.87433887e-02
-8.23556840e-01 3.84474486e-01 -5.51144719e-01 -4.80922580e-01
-2.52654761e-01 -1.20901696e-01 -1.10273516e+00 -3.11402261e-01
4.97216284e-01 1.15351424e-01 3.46131176e-02 -7.80585587e-01
8.64796042e-01 1.07906497e+00 9.08829510e-01 -4.08573151e-01
7.39355445e-01 5.59539795e-01 -1.98408917e-01 -8.51083636e-01
-8.02160725e-02 -4.76426393e-01 -6.78590655e-01 3.70228142e-01
6.21553957e-01 -9.66284156e-01 -9.95811105e-01 4.26050305e-01
-1.01358891e+00 2.35858530e-01 -8.47503424e-01 2.65469551e-01
-6.52223825e-01 8.15871835e-01 -3.02610636e-01 -8.09003711e-01
-7.98686862e-01 -8.45840096e-01 9.59515870e-01 -1.62574351e-01
-6.77630007e-01 -6.66748941e-01 1.59217045e-01 3.26701671e-01
6.47367537e-01 2.36753434e-01 2.45055035e-01 -1.07450771e+00
-2.93820262e-01 -4.60017323e-01 3.24578524e-01 3.47569525e-01
2.45051771e-01 -4.76626009e-01 -1.62009335e+00 -5.23164988e-01
3.61619294e-01 2.23697066e-01 6.40464127e-01 2.85273157e-02
1.23241675e+00 -9.58674729e-01 1.17391972e-02 1.52067745e+00
8.30004275e-01 -3.37346822e-01 6.96882367e-01 1.48438066e-01
6.01036012e-01 7.17204630e-01 1.96288556e-01 8.41003180e-01
2.82533228e-01 9.07279611e-01 6.19645678e-02 1.55678004e-01
-3.51267904e-02 -6.09254718e-01 2.90585756e-01 6.42206430e-01
5.02094150e-01 -2.91306973e-01 -5.56420386e-01 5.45223892e-01
-1.55533111e+00 -1.21804988e+00 1.30471602e-01 2.52368999e+00
8.25245917e-01 -3.37977290e-01 3.64107221e-01 1.65144444e-01
4.95831221e-01 3.37382197e-01 -5.00858545e-01 -8.00083697e-01
-2.70924687e-01 2.82154262e-01 1.07256985e+00 7.17536271e-01
-1.25948036e+00 1.04488468e+00 7.22367477e+00 4.45959806e-01
-1.17366397e+00 1.75357342e-01 2.80713111e-01 -5.17853916e-01
-7.16773510e-01 -5.80806024e-02 -6.45077407e-01 5.56980610e-01
1.51676965e+00 -8.42956841e-01 7.39645064e-01 6.90432906e-01
2.78651506e-01 6.08653724e-01 -1.31203711e+00 1.09451878e+00
1.55905694e-01 -1.01707006e+00 1.28953859e-01 1.49480283e-01
9.38888267e-02 -2.33415768e-01 3.30754340e-01 -1.32966697e-01
1.33109018e-01 -9.40043747e-01 7.60088503e-01 4.63430256e-01
1.14981556e+00 -9.98055935e-01 3.85494798e-01 3.16187479e-02
-1.12705934e+00 -2.72478499e-02 -1.23524621e-01 4.12748665e-01
3.96865278e-01 5.16879559e-01 -8.98683012e-01 8.44066441e-01
8.84341300e-01 4.33512509e-01 -5.70548534e-01 7.96162903e-01
-1.80907831e-01 5.86964488e-01 -7.12175846e-01 4.59281683e-01
-3.48540962e-01 2.07303483e-02 9.87755716e-01 1.36788535e+00
5.27026355e-01 8.98779035e-02 -3.97965223e-01 7.13993192e-01
-4.65058476e-01 2.06431434e-01 -5.85235119e-01 7.97259510e-02
9.72257733e-01 1.02269232e+00 1.47651866e-01 1.16809182e-01
-1.20288245e-01 1.57944369e+00 7.48692900e-02 1.49121433e-01
-6.25728607e-01 -5.95230997e-01 1.42136788e+00 2.37552524e-01
2.63196696e-02 1.10074105e-02 -5.26337802e-01 -1.32357442e+00
2.52262563e-01 -1.05383408e+00 6.19777143e-01 8.76694694e-02
-1.07789814e+00 5.22885442e-01 -2.36396089e-01 -1.10782897e+00
-3.34897041e-01 3.60310413e-02 -3.76876026e-01 1.17155170e+00
-1.10513079e+00 -1.19761753e+00 -1.59190997e-01 8.04679871e-01
-2.75629520e-01 -3.57516706e-01 1.33736730e+00 6.42468989e-01
-5.28745294e-01 1.53634501e+00 2.41684750e-01 -4.25343215e-02
9.67888117e-01 -1.18484437e+00 1.42627943e+00 1.09814000e+00
1.53696477e-01 1.03533494e+00 8.77520740e-01 -4.25294369e-01
-1.54351127e+00 -1.29627252e+00 1.05431259e+00 -6.01411164e-01
2.52594620e-01 -1.00794435e+00 -9.37478602e-01 8.98633540e-01
1.44813761e-01 -3.02702356e-02 1.32589900e+00 3.91574837e-02
-5.69700003e-01 -3.96116644e-01 -1.95738220e+00 3.56686622e-01
1.04826880e+00 -1.20279682e+00 -3.98473978e-01 -1.18647240e-01
1.13958001e+00 -3.37566376e-01 -1.02533054e+00 2.74600871e-02
7.33699381e-01 -1.00664365e+00 9.72553730e-01 -6.48486197e-01
-7.03084350e-01 -2.06356138e-01 -2.70706356e-01 -1.09808099e+00
-1.57532096e-01 -1.63647771e+00 -2.06411764e-01 1.77312231e+00
3.76175374e-01 -8.33565891e-01 1.00577784e+00 1.24865496e+00
1.28050342e-01 1.76919654e-01 -1.45429957e+00 -8.52058291e-01
-1.07227840e-01 -4.37713951e-01 1.22593045e+00 1.14787769e+00
-6.75966451e-03 -1.19855709e-01 -7.81349003e-01 5.15586376e-01
6.14014685e-01 -4.33179468e-01 1.00718331e+00 -1.07592392e+00
-2.70568728e-01 -8.69485736e-02 -6.51402056e-01 -6.27705157e-01
4.45242226e-01 -1.02543974e+00 -6.04481995e-01 -8.73365462e-01
-2.10978627e-01 1.34689948e-02 -2.58905917e-01 3.41042221e-01
1.23938143e-01 -5.86883202e-02 3.56423706e-01 8.84530507e-03
-1.88089535e-01 6.66969538e-01 -2.03667674e-02 -2.38738861e-02
-4.18259501e-01 3.05988103e-01 -1.04629195e+00 2.43607298e-01
9.29357529e-01 -6.50053024e-01 -4.23563629e-01 -4.80176240e-01
-3.11416030e-01 -1.86301589e-01 1.67659253e-01 -1.33538222e+00
4.37154889e-01 2.31242090e-01 -1.03301987e-01 -2.94350803e-01
5.87722003e-01 -1.26431608e+00 5.14581442e-01 2.82311141e-01
-5.08471906e-01 6.76085427e-03 5.55764735e-01 6.79621935e-01
-1.92183435e-01 1.85835600e-01 8.00331235e-01 3.10863316e-01
-3.31230134e-01 4.34154451e-01 -1.23022191e-01 1.15104705e-01
1.02823448e+00 -4.01430018e-02 -1.20785078e-02 -5.97046435e-01
-8.93542588e-01 2.48125911e-01 8.44810367e-01 5.45302808e-01
3.07295501e-01 -1.25972795e+00 -7.42058933e-01 4.79476929e-01
2.10418135e-01 -8.25732350e-01 3.82519543e-01 3.18102270e-01
-4.41290855e-01 3.35787803e-01 -1.93664581e-01 -2.36515682e-02
-1.63410676e+00 6.73904955e-01 4.33274269e-01 -6.11880124e-02
-4.92204159e-01 7.23346233e-01 -4.25038099e-01 -7.26013780e-01
4.34600562e-01 -2.46828333e-01 3.66677046e-01 1.79547146e-01
7.06849098e-01 5.77615976e-01 4.89641815e-01 -6.82778597e-01
-7.22783029e-01 6.07272051e-02 9.62764323e-02 -6.18987083e-01
1.07047963e+00 -4.44696188e-01 7.18686124e-03 7.87077025e-02
1.60270870e+00 6.75176382e-01 -8.37958872e-01 -1.15749292e-01
2.30764598e-01 -7.81254768e-01 -1.73254997e-01 -6.93255603e-01
-8.94694149e-01 6.69294417e-01 8.31683278e-01 1.28872395e-01
1.17816317e+00 -4.74660099e-01 1.08652031e+00 3.16497982e-01
2.83498526e-01 -1.20447993e+00 -1.03479850e+00 1.33685902e-01
7.51621544e-01 -8.07145655e-01 7.06509426e-02 -7.20325470e-01
-7.40433037e-01 6.05666876e-01 -8.36387500e-02 4.19270188e-01
7.24328876e-01 4.76913333e-01 1.57592922e-01 4.53310728e-01
-4.63933706e-01 2.78752327e-01 1.17464758e-01 9.43708301e-01
2.46306553e-01 3.69134724e-01 -3.02684426e-01 1.22665989e+00
-9.71875131e-01 -3.68230194e-01 3.12142611e-01 7.65785694e-01
2.54241258e-01 -1.55882823e+00 -3.37960869e-01 1.24876805e-01
-7.76901722e-01 -1.90966576e-01 -9.64127123e-01 2.34450057e-01
-1.62939265e-01 1.00321412e+00 -8.53895172e-02 -6.84891641e-01
7.56901324e-01 1.25034362e-01 -4.94657904e-02 -4.09142941e-01
-1.35841596e+00 -4.61513400e-01 4.30469364e-01 -1.15232575e+00
8.51114765e-02 -1.15667939e+00 -9.19667304e-01 -8.75316143e-01
2.83314139e-01 3.11319113e-01 7.29265451e-01 8.07101950e-02
1.10022461e+00 1.51095614e-01 8.71645987e-01 -2.82568961e-01
-8.57714236e-01 -5.61417103e-01 -8.43755424e-01 5.88658631e-01
5.53267181e-01 2.86297768e-01 -4.32412714e-01 -8.78839046e-02] | [13.99368953704834, 5.883128643035889] |
a7906667-665a-4901-b226-44c5be81d363 | video-referring-expression-comprehension-via | 2210.02953 | null | https://arxiv.org/abs/2210.02953v1 | https://arxiv.org/pdf/2210.02953v1.pdf | Video Referring Expression Comprehension via Transformer with Content-aware Query | Video Referring Expression Comprehension (REC) aims to localize a target object in video frames referred by the natural language expression. Recently, the Transformerbased methods have greatly boosted the performance limit. However, we argue that the current query design is suboptima and suffers from two drawbacks: 1) the slow training convergence process; 2) the lack of fine-grained alignment. To alleviate this, we aim to couple the pure learnable queries with the content information. Specifically, we set up a fixed number of learnable bounding boxes across the frame and the aligned region features are employed to provide fruitful clues. Besides, we explicitly link certain phrases in the sentence to the semantically relevant visual areas. To this end, we introduce two new datasets (i.e., VID-Entity and VidSTG-Entity) by augmenting the VIDSentence and VidSTG datasets with the explicitly referred words in the whole sentence, respectively. Benefiting from this, we conduct the fine-grained cross-modal alignment at the region-phrase level, which ensures more detailed feature representations. Incorporating these two designs, our proposed model (dubbed as ContFormer) achieves the state-of-the-art performance on widely benchmarked datasets. For example on VID-Entity dataset, compared to the previous SOTA, ContFormer achieves 8.75% absolute improvement on Accu.@0.6. | ['Yuexian Zou', 'Tengtao Song', 'Meng Cao', 'Ji Jiang'] | 2022-10-06 | null | null | null | null | ['referring-expression'] | ['computer-vision'] | [-6.47385791e-02 -1.10991649e-01 -3.06063920e-01 -4.35715586e-01
-1.21508729e+00 -4.62141126e-01 4.56734449e-01 -1.32262021e-01
-4.47096020e-01 5.66273987e-01 5.61600804e-01 1.39467254e-01
1.45964593e-01 -4.39658195e-01 -7.29516149e-01 -7.11213768e-01
3.13864976e-01 1.20624006e-02 2.98129350e-01 -1.85556531e-01
2.31097043e-01 1.79385036e-01 -1.31052518e+00 2.91961521e-01
7.64961004e-01 1.17427027e+00 3.38756144e-01 3.05640455e-02
-2.00844750e-01 8.76086771e-01 -4.80388165e-01 -6.18299007e-01
-9.57965329e-02 -4.41336185e-01 -7.12607145e-01 1.68176860e-01
6.13031924e-01 -4.81782079e-01 -4.06071723e-01 1.00605690e+00
4.36991721e-01 2.45388508e-01 4.66251969e-01 -1.23887074e+00
-7.95896947e-01 4.26100940e-01 -1.06385398e+00 4.46217388e-01
5.57826638e-01 -4.78459802e-03 1.32809389e+00 -1.23000646e+00
6.41931474e-01 1.37089980e+00 1.67935953e-01 5.34055531e-01
-7.81996846e-01 -6.96291327e-01 5.81907034e-01 5.40886939e-01
-1.74197280e+00 -5.68463206e-01 8.60283732e-01 -1.78754687e-01
8.74885559e-01 2.08183959e-01 4.45324898e-01 1.26269233e+00
2.74612270e-02 1.08404267e+00 9.61310506e-01 -2.33819321e-01
4.58750203e-02 1.56842191e-02 -2.22490728e-02 8.71123850e-01
-1.82814106e-01 -1.83322042e-01 -8.36170554e-01 7.29967654e-02
7.52306879e-01 -4.71555553e-02 -6.89245641e-01 -3.70288163e-01
-1.28783524e+00 8.17094207e-01 5.85649431e-01 2.92339325e-01
-3.68027121e-01 2.82979339e-01 5.40458441e-01 -1.59545556e-01
4.20349538e-01 1.11691475e-01 -3.43818516e-01 -2.84201115e-01
-6.89697206e-01 1.21579804e-01 3.51282805e-01 1.30441582e+00
7.20561922e-01 -1.27419084e-01 -4.87818658e-01 7.93321609e-01
3.22046101e-01 3.38758647e-01 4.08732295e-01 -9.10820961e-01
8.54367673e-01 5.39934278e-01 3.21763605e-02 -1.40320444e+00
-1.82663471e-01 -5.14156044e-01 -6.33677483e-01 -5.09333670e-01
2.31109217e-01 6.11595772e-02 -5.97110569e-01 2.06346464e+00
2.36970335e-01 2.47707710e-01 4.43849042e-02 1.24667597e+00
1.05681336e+00 6.92774296e-01 3.32995951e-01 -1.83385149e-01
1.70860052e+00 -1.34461451e+00 -9.18919086e-01 -2.76588470e-01
5.31893849e-01 -5.78515828e-01 1.42236733e+00 7.73766711e-02
-9.15010273e-01 -5.79427242e-01 -9.74204719e-01 -3.03897470e-01
-2.74962813e-01 1.95634469e-01 5.04479408e-01 3.02119732e-01
-9.90109384e-01 -5.74094616e-02 -5.44941723e-01 -3.88689786e-01
5.31830907e-01 1.21209197e-01 -4.67479855e-01 -1.59447044e-01
-1.16124868e+00 7.75392830e-01 2.31507316e-01 3.57281268e-02
-7.51151741e-01 -5.91788709e-01 -9.58098710e-01 1.53480425e-01
7.08521128e-01 -8.66954207e-01 1.06018662e+00 -9.19822991e-01
-1.17921913e+00 1.17301512e+00 -5.55610120e-01 -1.77704856e-01
4.20467645e-01 -6.33415043e-01 -2.58885205e-01 4.53983903e-01
3.50006908e-01 9.39396381e-01 8.81965339e-01 -1.14295518e+00
-6.92281544e-01 -3.66990447e-01 4.70492452e-01 4.20820206e-01
-4.18113112e-01 2.93863207e-01 -1.04275966e+00 -1.02425158e+00
1.90754279e-01 -5.80238879e-01 1.90444216e-01 1.77148074e-01
-3.44549030e-01 -5.11866510e-01 8.16798568e-01 -5.07242858e-01
1.40764296e+00 -2.40211153e+00 3.20253730e-01 -2.60370374e-01
3.52452904e-01 -1.48006365e-01 -1.71852931e-01 2.19610929e-01
-9.61915702e-02 8.90595317e-02 4.82969396e-02 -4.19267088e-01
-3.46035771e-02 2.86928892e-01 -3.96590710e-01 4.18105066e-01
2.70590901e-01 1.10669887e+00 -1.01305401e+00 -8.18026423e-01
-3.62502933e-02 2.98575222e-01 -6.15348876e-01 3.69995296e-01
-1.03618294e-01 4.09667104e-01 -9.39711452e-01 7.35646844e-01
4.68149543e-01 -3.37971687e-01 -2.48630553e-01 -6.25980139e-01
1.35354832e-01 9.62726548e-02 -8.30921769e-01 2.01113772e+00
-3.03890556e-01 4.53306049e-01 -1.47855684e-01 -9.46885884e-01
9.02099848e-01 3.43826711e-01 5.38897336e-01 -8.64560068e-01
1.47252440e-01 5.40209673e-02 -4.01448786e-01 -7.21887290e-01
4.62162018e-01 5.80762466e-03 -2.49220669e-01 1.69647858e-01
1.32283002e-01 2.54584849e-01 -1.14263453e-01 2.82533228e-01
8.38467240e-01 5.24071515e-01 5.55390239e-01 -2.41019070e-01
7.43055105e-01 -1.34979412e-01 6.83951437e-01 3.99861068e-01
-4.66690063e-01 6.71153843e-01 5.91359913e-01 -2.31552988e-01
-6.57763958e-01 -9.44340050e-01 5.53057082e-02 1.35378027e+00
5.91818392e-01 -5.90360105e-01 -7.72162020e-01 -7.56244600e-01
-3.16842943e-01 6.71056747e-01 -6.10745668e-01 -1.00087836e-01
-7.39029169e-01 -4.34757710e-01 5.07963240e-01 8.14540386e-01
8.92861784e-01 -8.85802150e-01 -6.64951921e-01 1.28916940e-02
-6.00428402e-01 -1.53309548e+00 -8.74264359e-01 -1.64897576e-01
-5.76535404e-01 -9.33887303e-01 -7.91761577e-01 -7.76796162e-01
6.20839596e-01 4.95949537e-01 1.33503985e+00 -3.61209847e-02
1.55901358e-01 5.49540579e-01 -6.67164862e-01 -2.57302225e-02
2.87301242e-01 -3.74080939e-03 -8.15567747e-02 1.43700480e-01
5.15686870e-01 -3.39780629e-01 -6.53570175e-01 4.89272743e-01
-6.74648285e-01 3.53894413e-01 5.33863246e-01 1.00484741e+00
8.99083853e-01 -3.71427178e-01 6.27858579e-01 -5.91381371e-01
4.65718776e-01 -4.99275863e-01 -3.16093057e-01 5.10371089e-01
-2.61544079e-01 -8.52514151e-03 4.39831048e-01 -3.81012619e-01
-1.00828063e+00 -1.14671677e-01 -4.60608006e-02 -6.86094582e-01
-8.36876780e-02 4.35841769e-01 -5.09395659e-01 1.19632691e-01
2.44665653e-01 3.33086580e-01 -3.90467256e-01 -2.61060685e-01
5.31042457e-01 5.44668198e-01 6.68180406e-01 -9.77323949e-01
6.24712229e-01 3.36812586e-01 -2.04529375e-01 -3.80800754e-01
-1.12881875e+00 -5.31530797e-01 -5.31054080e-01 -2.27725506e-01
1.03826690e+00 -1.10268664e+00 -5.58567405e-01 1.33916497e-01
-1.36996245e+00 9.14117619e-02 -1.08756416e-01 3.43883723e-01
-8.33295584e-01 3.97733778e-01 -4.44514602e-01 -6.23605132e-01
-4.00226980e-01 -1.20849454e+00 1.44024372e+00 3.16720694e-01
-6.71876743e-02 -6.81354344e-01 -3.95403951e-01 4.27311629e-01
1.87218681e-01 3.36248547e-01 8.19171011e-01 -6.06688321e-01
-7.67157376e-01 6.61698058e-02 -6.98659539e-01 3.21232565e-02
1.08702898e-01 -2.95304298e-01 -1.01828241e+00 -2.04758987e-01
1.63074061e-01 -4.02267665e-01 5.66176236e-01 1.31275788e-01
1.39317715e+00 -2.51422346e-01 -3.42516631e-01 6.45557284e-01
1.21155083e+00 1.05709128e-01 5.03190994e-01 4.72904146e-01
7.03361988e-01 6.62704706e-01 1.07158518e+00 4.34587181e-01
5.61365008e-01 9.90541995e-01 4.40615952e-01 -3.20395567e-02
-1.59177870e-01 -4.10924345e-01 3.12764168e-01 8.14206004e-01
-4.32763807e-02 -4.99695569e-01 -7.09907949e-01 5.27181923e-01
-2.03901863e+00 -7.60635912e-01 1.33099288e-01 1.78971124e+00
7.38870442e-01 -5.85351475e-02 -7.38612935e-02 -3.12906682e-01
5.90528905e-01 4.44959074e-01 -5.98838925e-01 1.20789781e-02
-3.22583854e-01 -1.24780573e-01 1.11554764e-01 1.99830249e-01
-1.11645639e+00 1.15464413e+00 5.31309557e+00 1.04953229e+00
-9.37284470e-01 2.33552232e-01 7.36479521e-01 -2.34126970e-02
-1.63048089e-01 -2.01418564e-01 -8.90520692e-01 3.31944436e-01
4.94406581e-01 -1.76641405e-01 2.41186440e-01 6.85978472e-01
3.30917507e-01 -6.36220798e-02 -1.21007013e+00 1.36837399e+00
4.79437292e-01 -1.07451963e+00 2.79238164e-01 -3.06892067e-01
3.64858955e-01 -2.39347920e-01 8.57570916e-02 4.09750342e-01
-3.70108932e-01 -8.93411338e-01 9.37026739e-01 4.99383301e-01
8.94938588e-01 -6.68912411e-01 8.22242200e-01 1.93693519e-01
-1.33448446e+00 1.26731336e-01 -2.32731327e-01 2.96311080e-01
2.68461019e-01 1.15434527e-01 -2.97403008e-01 7.62230992e-01
8.27023387e-01 9.23251927e-01 -5.91786444e-01 8.17668080e-01
-3.32676321e-01 5.19739389e-01 -1.68266103e-01 7.60937249e-03
5.05766869e-01 -1.06264703e-01 5.02361655e-01 1.13029385e+00
2.69055814e-01 3.89577687e-01 1.50945112e-01 1.02701724e+00
-1.97256759e-01 3.37990075e-01 -3.08729053e-01 1.92677185e-01
5.26589036e-01 1.15253174e+00 -4.18396533e-01 -2.19704106e-01
-7.78358579e-01 1.13542438e+00 4.94219601e-01 5.94366133e-01
-1.14793932e+00 -1.84335366e-01 5.78229010e-01 -8.21645483e-02
3.98063749e-01 -6.56707063e-02 3.28941196e-02 -1.31870127e+00
2.86351115e-01 -8.99735868e-01 4.83048826e-01 -1.13901591e+00
-1.42880177e+00 6.96458519e-01 2.87017584e-01 -1.30650783e+00
-1.96265012e-01 -5.36798835e-01 -3.50953311e-01 8.94557595e-01
-1.63707197e+00 -1.28385627e+00 -6.22280478e-01 7.62986183e-01
8.64385366e-01 -4.37439345e-02 8.20957661e-01 4.06624079e-01
-7.09310412e-01 8.89221847e-01 -3.85439068e-01 2.66125321e-01
8.41556847e-01 -9.26540732e-01 4.24948670e-02 6.90537870e-01
2.68458486e-01 6.94477916e-01 5.51827669e-01 -2.44851664e-01
-1.28324163e+00 -9.80200529e-01 7.67758548e-01 -4.82433259e-01
7.44985998e-01 -2.89295524e-01 -9.22518015e-01 7.07369208e-01
3.12896043e-01 2.31849641e-01 5.41791141e-01 -4.75678109e-02
-5.22844195e-01 -1.61109895e-01 -9.58630085e-01 8.22253585e-01
1.25042534e+00 -6.20349526e-01 -8.21542025e-01 2.11287275e-01
9.91114020e-01 -6.06121123e-01 -7.83500254e-01 3.56146485e-01
4.10805732e-01 -9.17566657e-01 8.89858961e-01 -6.61733985e-01
6.14702523e-01 -3.50475460e-01 -6.51405036e-01 -8.10158432e-01
-1.31155849e-01 -5.52265465e-01 -2.58746237e-01 1.59937632e+00
1.49474844e-01 -4.34227020e-01 6.25722051e-01 5.90760291e-01
-2.01066896e-01 -1.28762043e+00 -1.06983471e+00 -5.32268882e-01
-1.22111149e-01 -2.32923433e-01 5.39674759e-01 8.83395731e-01
2.69069839e-02 6.24155164e-01 -3.85609657e-01 3.60647663e-02
3.32687974e-01 1.99367091e-01 7.16845989e-01 -5.51242530e-01
-1.56803370e-01 -4.23525900e-01 -4.13630813e-01 -1.85929394e+00
2.91942060e-01 -4.43659008e-01 2.58772960e-03 -1.18626535e+00
5.31741261e-01 -3.35448772e-01 -3.93706411e-01 5.18867910e-01
-5.33794105e-01 5.25434241e-02 3.10669988e-01 3.18752408e-01
-9.09227312e-01 8.28278899e-01 1.50494075e+00 -8.14207420e-02
1.86559394e-01 -3.49600852e-01 -8.44857872e-01 7.46310711e-01
5.70885718e-01 -3.00014824e-01 -4.46668923e-01 -7.29875684e-01
-3.84726562e-02 1.93869472e-01 5.10066509e-01 -6.44440770e-01
2.13167354e-01 -1.59180909e-01 1.43165186e-01 -7.94131696e-01
5.96356332e-01 -8.13651919e-01 -3.81703734e-01 -1.75507724e-01
-3.66327107e-01 3.48807096e-01 8.96861553e-02 7.41630256e-01
-5.54365158e-01 -1.35269403e-01 4.53148276e-01 2.69485321e-02
-1.10205567e+00 3.10415179e-01 7.59592652e-02 2.86796123e-01
1.08297753e+00 -3.26512784e-01 -4.92404252e-01 -4.86788481e-01
-4.09759074e-01 4.31607097e-01 3.12027723e-01 5.98059773e-01
8.09261680e-01 -1.43226647e+00 -6.15861714e-01 -1.80125043e-01
4.93065685e-01 8.08666870e-02 2.76912093e-01 1.08463025e+00
-1.72609553e-01 5.19989848e-01 -6.98371530e-02 -7.35948443e-01
-1.17667162e+00 7.32682288e-01 3.22040886e-01 -9.86917391e-02
-7.29400873e-01 1.04483604e+00 6.64597094e-01 6.02061115e-02
4.03172672e-01 -2.77262956e-01 -3.97555083e-01 4.43768315e-02
3.85281920e-01 -3.65025774e-02 -3.18097442e-01 -1.03797483e+00
-4.86778677e-01 8.87428701e-01 -1.89024583e-01 1.15742184e-01
1.05398238e+00 -5.04461050e-01 -1.98309906e-02 3.47266346e-01
1.34701645e+00 5.42183593e-02 -1.20077276e+00 -5.67912877e-01
-5.37214838e-02 -6.00249290e-01 4.19203341e-02 -5.11282504e-01
-1.23566604e+00 8.53386164e-01 4.71154213e-01 -2.74005622e-01
1.36290908e+00 3.91963750e-01 8.08197260e-01 3.12169939e-01
5.48556924e-01 -7.16667116e-01 2.73071259e-01 2.57902861e-01
1.12455332e+00 -1.26400888e+00 1.04694843e-01 -7.43769884e-01
-8.93748820e-01 1.00913846e+00 9.13563907e-01 -2.26924382e-02
2.16664761e-01 -1.18772075e-01 1.09079465e-01 -2.98122734e-01
-6.92876816e-01 -2.48102322e-01 4.30653214e-01 4.03454214e-01
4.62644994e-01 -1.99942634e-01 -3.00810754e-01 9.46125209e-01
-7.02675581e-02 -1.07934043e-01 4.16029952e-02 6.73675835e-01
-3.77705395e-01 -7.64742374e-01 -3.01943004e-01 1.83904618e-01
-4.62381274e-01 -1.65816098e-01 -1.97997868e-01 9.65252936e-01
5.34251705e-02 8.83659303e-01 -7.06239557e-03 -2.26639599e-01
5.66886127e-01 -1.69252023e-01 3.90936345e-01 -4.19314414e-01
-2.80812234e-01 1.46526217e-01 6.53678179e-02 -8.24018657e-01
-8.34130764e-01 -5.65788925e-01 -1.32114482e+00 -3.79991792e-02
-4.03874934e-01 1.03999056e-01 1.91200711e-02 1.02633667e+00
3.90771776e-01 6.03324831e-01 4.68130201e-01 -5.57865262e-01
-4.90441293e-01 -8.62495363e-01 -3.55819225e-01 4.88020301e-01
1.69904068e-01 -1.00945592e+00 -5.00878431e-02 1.06887668e-01] | [10.278319358825684, 1.133104920387268] |
e77c0d60-b57a-4856-ae8b-930a317b6a71 | polyglot-semantic-role-labeling | 1805.11598 | null | http://arxiv.org/abs/1805.11598v1 | http://arxiv.org/pdf/1805.11598v1.pdf | Polyglot Semantic Role Labeling | Previous approaches to multilingual semantic dependency parsing treat
languages independently, without exploiting the similarities between semantic
structures across languages. We experiment with a new approach where we combine
resources from a pair of languages in the CoNLL 2009 shared task to build a
polyglot semantic role labeler. Notwithstanding the absence of parallel data,
and the dissimilarity in annotations between languages, our approach results in
an improvement in SRL performance on multiple languages over a monolingual
baseline. Analysis of the polyglot model shows it to be advantageous in
lower-resource settings. | ['Noah Smith', 'Swabha Swayamdipta', 'Phoebe Mulcaire'] | 2018-05-29 | polyglot-semantic-role-labeling-1 | https://aclanthology.org/P18-2106 | https://aclanthology.org/P18-2106.pdf | acl-2018-7 | ['semantic-dependency-parsing'] | ['natural-language-processing'] | [-3.39213423e-02 1.09878935e-01 -4.01392400e-01 -5.10179818e-01
-9.76854861e-01 -1.16828227e+00 8.86886477e-01 2.82124907e-01
-7.13700175e-01 9.75440979e-01 7.43934989e-01 -4.65478599e-01
7.84500316e-02 -1.33717448e-01 -4.11576390e-01 -2.38171518e-01
3.80982727e-01 5.84928334e-01 4.38024253e-01 -4.53443557e-01
-5.64346053e-02 -1.39783471e-04 -8.98139000e-01 5.45873880e-01
7.36289501e-01 1.80620089e-01 3.61473978e-01 2.71517873e-01
-2.70475894e-01 9.79283690e-01 -4.07960325e-01 -5.45343101e-01
1.94772556e-01 -4.96656835e-01 -1.31941748e+00 -2.84202218e-01
3.60810250e-01 3.89881164e-01 -1.51779249e-01 8.88146639e-01
2.01620758e-01 -2.96644289e-02 3.05907547e-01 -9.32953656e-01
-5.27011216e-01 1.05025280e+00 -2.74901599e-01 3.39939654e-01
5.52404881e-01 -3.87138903e-01 1.43680704e+00 -5.46341896e-01
1.28794205e+00 1.67318332e+00 6.18886054e-01 5.20741165e-01
-1.42491579e+00 -4.18063939e-01 4.39984322e-01 -7.80882612e-02
-1.02194011e+00 -7.70729721e-01 4.85101461e-01 -1.97423518e-01
1.49451566e+00 -4.74206060e-02 2.01481953e-01 8.58099520e-01
-3.40562135e-01 5.46967030e-01 1.52687371e+00 -9.22456264e-01
-3.33228648e-01 2.01463223e-01 3.17833811e-01 5.40342510e-01
2.17802346e-01 -1.20104827e-01 -6.04328930e-01 -7.80166015e-02
4.25774902e-01 -6.50572777e-01 -5.62900305e-02 -3.07107806e-01
-1.36444390e+00 7.68483818e-01 1.58189356e-01 8.13978374e-01
2.71988571e-01 1.23577811e-01 8.10402274e-01 6.06532991e-01
6.58218682e-01 5.99272847e-01 -9.56067026e-01 6.94341063e-02
-4.55948442e-01 9.37628970e-02 9.90845442e-01 8.71020436e-01
5.50258219e-01 -1.16512306e-01 2.79963464e-01 1.15916669e+00
2.84216344e-01 2.13359356e-01 4.81559008e-01 -1.14010298e+00
6.85656250e-01 3.85063142e-01 -2.24630814e-02 -3.04222167e-01
-3.43646526e-01 5.10712154e-02 1.96422920e-01 -1.25827536e-01
6.58436835e-01 -6.31742775e-02 -7.18672991e-01 2.09396482e+00
2.57817894e-01 -4.34788436e-01 5.79077363e-01 6.44573152e-01
4.37920541e-01 3.03959638e-01 7.31725097e-01 -4.14394140e-02
1.60952950e+00 -1.01384830e+00 -4.74136770e-01 -5.34343481e-01
1.22518420e+00 -9.51508999e-01 1.16331935e+00 4.66909632e-02
-9.49957848e-01 -2.72470385e-01 -9.86004412e-01 -5.45322835e-01
-5.53475976e-01 -2.46769600e-02 1.01554537e+00 7.03203201e-01
-1.19307590e+00 3.33275914e-01 -7.50869751e-01 -7.43366897e-01
-1.05625011e-01 -1.06812594e-02 -5.57920694e-01 -3.27098668e-01
-1.39491570e+00 1.22853160e+00 8.05741131e-01 -4.47600007e-01
-4.13422883e-01 -6.15515172e-01 -1.16991270e+00 -4.47289228e-01
3.20462584e-01 -4.62906778e-01 1.40648246e+00 -1.23841190e+00
-1.06157553e+00 1.44474518e+00 -2.44699717e-01 -2.20739126e-01
3.40946525e-01 -3.44726026e-01 -3.31187814e-01 -1.72426015e-01
5.85255265e-01 6.05041444e-01 1.18710749e-01 -1.09954345e+00
-9.15813148e-01 -4.69975799e-01 4.25999165e-01 7.68953085e-01
9.99675617e-02 8.79214704e-01 -4.74780083e-01 -5.98279238e-01
3.74081247e-02 -1.11006951e+00 -1.94346964e-01 -6.88955247e-01
2.59080175e-02 -4.53366250e-01 5.34437239e-01 -9.40880179e-01
9.33093071e-01 -2.22407579e+00 2.10286796e-01 -1.42599583e-01
-3.54839712e-01 -7.43531361e-02 -8.41472298e-02 5.60649574e-01
-3.55304897e-01 2.68788010e-01 -2.48141050e-01 -4.64962333e-01
-7.60347694e-02 7.32906044e-01 -4.18905988e-02 4.42720115e-01
9.44558084e-02 9.28076327e-01 -1.08601117e+00 -4.90692496e-01
5.99566959e-02 3.61317158e-01 -4.20425475e-01 -2.03334346e-01
-2.61709809e-01 7.20680594e-01 -3.41398507e-01 5.67507565e-01
6.10407256e-02 1.47534177e-01 1.04511273e+00 1.56480819e-01
-2.69013107e-01 1.13294351e+00 -8.64537179e-01 2.31262684e+00
-7.59168804e-01 3.97508383e-01 2.48816729e-01 -8.78872752e-01
3.93751889e-01 6.74692750e-01 1.16148815e-01 -7.82163262e-01
-2.71849155e-01 7.25537002e-01 1.98661536e-01 -9.19234306e-02
2.78742701e-01 -4.46905226e-01 -6.58655763e-01 4.96552289e-01
4.14209455e-01 -2.81130224e-01 4.57436860e-01 2.61738598e-01
8.74882042e-01 4.88529563e-01 5.44182777e-01 -7.50015318e-01
7.04027414e-01 3.99941236e-01 5.49837172e-01 4.53585237e-01
-4.15049940e-02 2.99620926e-01 5.56594431e-01 -3.15937012e-01
-1.08257091e+00 -1.08014345e+00 -2.34545678e-01 1.57238257e+00
3.99128068e-03 -5.27908027e-01 -4.25709248e-01 -1.09477818e+00
-1.34962231e-01 8.24769914e-01 -2.15467751e-01 4.66813356e-01
-1.19487453e+00 -7.95796990e-01 9.30942118e-01 4.92604405e-01
-7.03582587e-03 -1.04394913e+00 -1.17753215e-01 3.06105435e-01
-2.09996015e-01 -1.68545985e+00 -1.61424354e-01 4.59297895e-01
-7.06667900e-01 -1.08854437e+00 -2.82453924e-01 -1.05935740e+00
2.14049220e-01 -5.68895526e-02 1.62385511e+00 5.12643270e-02
1.44141123e-01 1.52347296e-01 -3.34068149e-01 -1.69374481e-01
-7.07895577e-01 3.45121086e-01 -1.56724915e-01 -7.45411217e-01
5.21907091e-01 -3.51332545e-01 5.34859151e-02 -1.13866709e-01
-6.57155991e-01 -1.59403324e-01 -2.61967704e-02 7.63869166e-01
3.27418119e-01 -4.50048655e-01 5.73855698e-01 -1.64138925e+00
4.53699797e-01 -6.03069127e-01 -5.14711142e-01 4.21475291e-01
-4.60023075e-01 3.94940555e-01 6.48732781e-01 1.84473127e-01
-1.31419790e+00 -5.28132683e-03 -1.54071838e-01 4.23929453e-01
-2.68053621e-01 4.98844743e-01 -5.30601263e-01 1.83919296e-01
4.80759174e-01 -3.67002547e-01 -3.66978198e-01 -7.99374580e-01
7.12997735e-01 3.19452226e-01 5.16817629e-01 -1.05877125e+00
3.65008920e-01 2.68853515e-01 -2.56447941e-01 -6.57192349e-01
-1.07440650e+00 -5.43216646e-01 -9.46565449e-01 5.69887161e-01
9.87670541e-01 -1.46919990e+00 6.24171235e-02 1.85509190e-01
-1.27201676e+00 -3.89106095e-01 -3.00684452e-01 4.38475490e-01
-3.55992407e-01 4.57174093e-01 -9.67130363e-01 -1.52044415e-01
9.03283656e-02 -1.03210616e+00 9.05593753e-01 -2.88103968e-01
-4.71468627e-01 -1.67106605e+00 2.63252228e-01 4.65894550e-01
7.85010532e-02 -5.62196448e-02 1.26400959e+00 -1.14901900e+00
-2.16956183e-01 1.86431557e-01 -3.16105872e-01 2.67097503e-01
1.76625222e-01 -6.12737477e-01 -8.04162562e-01 -1.87344924e-01
-1.02890551e-01 -6.62933767e-01 5.94589949e-01 -1.19232662e-01
2.54602820e-01 -6.33717552e-02 -2.00728178e-01 4.96712983e-01
1.56011701e+00 7.28263482e-02 1.38396367e-01 6.36620641e-01
9.40952778e-01 1.01142764e+00 5.35506964e-01 -3.41678649e-01
7.59542406e-01 6.19319081e-01 -1.95713833e-01 -5.86419366e-02
-4.23281848e-01 -3.39491665e-01 5.60401440e-01 1.10311317e+00
1.15362413e-01 -1.74547076e-01 -1.20369089e+00 7.84231365e-01
-1.81961060e+00 -3.57604325e-01 -2.92854011e-01 1.85700345e+00
1.03593302e+00 -1.49436845e-02 7.77887926e-02 -3.67085993e-01
4.92399126e-01 3.95571500e-01 7.40948692e-02 -6.23477519e-01
-4.68216300e-01 4.67506200e-01 6.79753244e-01 9.37513113e-01
-9.98779416e-01 1.76029384e+00 7.26246595e+00 5.74571967e-01
-8.10100079e-01 7.86606669e-01 1.93550393e-01 2.13192627e-01
-5.50611675e-01 5.88004470e-01 -9.17607665e-01 3.14340174e-01
1.18660831e+00 -8.95738602e-02 4.55834478e-01 5.62731147e-01
-2.40544856e-01 -2.68646896e-01 -1.21208692e+00 5.59703529e-01
1.56426504e-01 -8.94642174e-01 -8.37739334e-02 -3.68001252e-01
6.24606252e-01 6.41133249e-01 -5.47253430e-01 2.06842706e-01
1.08736300e+00 -9.45130408e-01 7.43320584e-01 -2.83798367e-01
8.00582528e-01 -5.85381508e-01 5.73492825e-01 2.27184549e-01
-1.17002225e+00 2.43623540e-01 -1.13362312e-01 -2.42235065e-01
4.09149408e-01 2.98479777e-02 -6.31292582e-01 8.42955887e-01
4.63440269e-01 9.18048620e-01 -6.47410035e-01 2.29358777e-01
-7.32360542e-01 6.81215644e-01 -2.96075851e-01 4.45136994e-01
4.35136527e-01 -2.59619504e-01 5.81290364e-01 1.65238345e+00
-2.34742239e-01 -2.68834293e-01 7.19064951e-01 1.17293566e-01
-2.62484998e-02 4.18317288e-01 -7.11151481e-01 -1.44807756e-01
3.23456585e-01 9.54793215e-01 -9.24146116e-01 -4.65283096e-01
-8.75950158e-01 1.08359575e+00 6.89811349e-01 1.72801822e-01
-3.64254117e-01 2.33118962e-02 5.91049850e-01 -1.84264574e-02
-6.05134331e-02 -6.61960900e-01 -1.90859467e-01 -1.27414000e+00
-3.82016264e-02 -1.04691350e+00 8.22921336e-01 -5.13521552e-01
-1.28071511e+00 6.18158102e-01 3.57254773e-01 -3.91323835e-01
-4.58782703e-01 -7.74792671e-01 -9.63370595e-03 1.19867456e+00
-1.74085629e+00 -1.64390481e+00 4.17452484e-01 6.89872265e-01
6.88167393e-01 -2.69213654e-02 9.36362743e-01 3.37110996e-01
-3.08931828e-01 3.68478149e-01 -9.43476856e-02 1.77151293e-01
9.43172216e-01 -1.60229945e+00 7.32303917e-01 1.05608964e+00
4.89662170e-01 6.80175185e-01 5.14400363e-01 -7.16041207e-01
-8.24864328e-01 -8.24335158e-01 1.61967313e+00 -8.55276704e-01
1.15597475e+00 -5.79005480e-01 -9.59507287e-01 1.15653455e+00
5.93575776e-01 -9.60539281e-02 6.97308600e-01 6.53128803e-01
-7.13497818e-01 4.02310252e-01 -7.05388248e-01 4.19720232e-01
1.35530078e+00 -1.04756582e+00 -1.06621778e+00 3.56102973e-01
9.26806033e-01 -2.84790039e-01 -8.60878944e-01 3.51147443e-01
4.32735205e-01 -5.11445224e-01 7.17497945e-01 -8.23234320e-01
2.22690701e-01 -2.37880602e-01 -4.17838067e-01 -1.42912447e+00
-1.50442749e-01 -4.37224686e-01 7.29027927e-01 1.45664597e+00
7.09266365e-01 -8.75092566e-01 3.02673370e-01 4.17654663e-01
-2.82466918e-01 1.10761128e-01 -1.00913537e+00 -8.83288324e-01
4.50300187e-01 -3.77852559e-01 6.78272545e-02 1.37281156e+00
1.10357016e-01 9.59836602e-01 1.51270190e-02 5.37052751e-02
4.97190326e-01 1.98873989e-02 3.16520035e-01 -1.26719332e+00
-5.77307105e-01 -8.95122737e-02 -1.75526202e-01 -6.46471381e-01
8.17781329e-01 -1.41882741e+00 6.43253559e-03 -1.51531243e+00
1.63586408e-01 -8.78706753e-01 -4.47663635e-01 8.26084614e-01
-1.69750139e-01 4.03737009e-01 3.17392319e-01 4.77027744e-01
-5.99830925e-01 4.12853360e-02 5.77153027e-01 2.77635366e-01
-2.14531273e-01 -7.31575489e-01 -9.49577928e-01 9.96432424e-01
6.66441202e-01 -8.12323391e-01 -4.83902633e-01 -9.80477750e-01
3.61753792e-01 -2.61032879e-02 -1.16150737e-01 -6.31429672e-01
-1.19762793e-01 -1.07084185e-01 8.43435302e-02 1.09167948e-01
3.61459181e-02 -6.00299299e-01 1.23764537e-01 2.06527457e-01
-1.95318103e-01 4.07425761e-01 2.68917680e-01 1.77758723e-01
-3.95561546e-01 -4.10157174e-01 6.46252930e-01 -5.55185378e-01
-1.11043179e+00 -3.34914804e-01 -2.52427429e-01 6.09008372e-01
6.95985615e-01 6.64895996e-02 -4.00141507e-01 2.81502277e-01
-8.38637829e-01 1.50428206e-01 9.06327128e-01 7.85292208e-01
-4.43048000e-01 -8.86964321e-01 -6.47288680e-01 2.69056503e-02
2.55597770e-01 -3.31701607e-01 -3.54404688e-01 5.57229638e-01
-4.56618398e-01 6.38772845e-01 -1.29320890e-01 -3.50614578e-01
-1.30951846e+00 3.01699013e-01 3.07568938e-01 -7.46276736e-01
-3.85250688e-01 6.50662422e-01 3.72239500e-01 -1.00416720e+00
-3.18295181e-01 1.71302229e-01 -2.46433333e-01 8.01321566e-02
-3.23785990e-02 -1.03389300e-01 1.93846017e-01 -1.03735888e+00
-6.34495318e-01 3.61877024e-01 -2.25428820e-01 -6.59492075e-01
1.33572698e+00 -3.90125066e-01 -3.83599997e-01 8.40299785e-01
1.00198412e+00 7.29711294e-01 -7.97294974e-01 -3.02410156e-01
8.41912389e-01 -2.12067544e-01 -3.38325858e-01 -1.05872321e+00
-5.72008431e-01 3.67317259e-01 -2.02029124e-02 -8.26055259e-02
7.20946431e-01 5.35925448e-01 5.75978041e-01 7.93260932e-02
5.55018485e-01 -1.21756470e+00 -4.45688695e-01 1.08909273e+00
5.20277858e-01 -1.26959169e+00 -5.55430874e-02 -5.81722856e-01
-7.85402536e-01 6.91241801e-01 3.89503688e-01 -1.28764614e-01
4.62048531e-01 3.23264748e-01 4.76889789e-01 -3.40271413e-01
-6.17822587e-01 -4.45935190e-01 -6.11705193e-03 4.77862149e-01
1.18392491e+00 2.64120132e-01 -9.31814969e-01 4.68217015e-01
-2.88727850e-01 -4.96636361e-01 1.53581411e-01 1.15577400e+00
-3.43580102e-03 -1.94369364e+00 6.64209351e-02 -2.69507200e-01
-1.03181815e+00 -5.90930223e-01 -5.05366802e-01 1.27452016e+00
2.65549868e-01 8.68421733e-01 9.20347199e-02 3.59569311e-01
3.13006461e-01 6.75696254e-01 6.28302753e-01 -1.20367599e+00
-7.79816508e-01 1.90772578e-01 8.18108320e-01 -6.44155443e-01
-8.43010247e-01 -8.89651000e-01 -1.24772990e+00 1.91084012e-01
8.81346911e-02 3.14777911e-01 6.83782399e-01 1.19788194e+00
6.00258298e-02 1.73796281e-01 2.60354191e-01 -3.51549834e-01
-1.69094190e-01 -7.59599090e-01 -3.82079840e-01 5.90382576e-01
-4.74932119e-02 -3.99716586e-01 -2.71565914e-01 1.91847369e-01] | [10.50412654876709, 9.651293754577637] |
a9f2ec2c-8266-4e57-aecb-81112c7dd647 | gaudi-a-neural-architect-for-immersive-3d | 2207.13751 | null | https://arxiv.org/abs/2207.13751v1 | https://arxiv.org/pdf/2207.13751v1.pdf | GAUDI: A Neural Architect for Immersive 3D Scene Generation | We introduce GAUDI, a generative model capable of capturing the distribution of complex and realistic 3D scenes that can be rendered immersively from a moving camera. We tackle this challenging problem with a scalable yet powerful approach, where we first optimize a latent representation that disentangles radiance fields and camera poses. This latent representation is then used to learn a generative model that enables both unconditional and conditional generation of 3D scenes. Our model generalizes previous works that focus on single objects by removing the assumption that the camera pose distribution can be shared across samples. We show that GAUDI obtains state-of-the-art performance in the unconditional generative setting across multiple datasets and allows for conditional generation of 3D scenes given conditioning variables like sparse image observations or text that describes the scene. | ['Josh Susskind', 'Afshin Dehghan', 'Daniel Ulbricht', 'Hanlin Goh', 'Shuangfei Zhai', 'Laurent Dinh', 'Zhuoyuan Chen', 'Alexander Toshev', 'Walter Talbott', 'Samira Abnar', 'Pengsheng Guo', 'Miguel Angel Bautista'] | 2022-07-27 | null | null | null | null | ['scene-generation'] | ['computer-vision'] | [ 1.92258447e-01 -2.67720282e-01 3.04864228e-01 -4.52658027e-01
-1.00230968e+00 -8.76833081e-01 9.47950721e-01 -7.31874108e-01
9.07794908e-02 4.98859555e-01 4.38312769e-01 -1.13009028e-01
5.57847209e-02 -8.29356015e-01 -1.09365189e+00 -1.05802190e+00
5.22821769e-02 5.56688309e-01 -1.89811558e-01 2.75232136e-01
-1.78821266e-01 5.32260537e-01 -1.58532441e+00 4.79129218e-02
3.58285248e-01 6.28877103e-01 3.99708539e-01 1.26559341e+00
3.18626106e-01 1.00265563e+00 -5.04756570e-01 -1.38584569e-01
3.83488327e-01 -4.06701773e-01 -3.23790669e-01 5.16441047e-01
9.34803426e-01 -5.89358032e-01 -6.11703038e-01 7.79973447e-01
4.23941255e-01 1.79363370e-01 1.03821647e+00 -1.09887683e+00
-6.91103041e-01 7.37299398e-02 -4.81342077e-01 -1.33123845e-01
5.61056316e-01 2.97668397e-01 1.00530314e+00 -6.45123720e-01
7.03748167e-01 1.32740009e+00 2.57734060e-01 4.14129674e-01
-1.87819278e+00 -4.78593946e-01 2.73581892e-01 -2.17962220e-01
-1.55572820e+00 -2.74105251e-01 8.89781535e-01 -6.98420227e-01
9.07905877e-01 3.07064623e-01 6.18991137e-01 1.73153257e+00
2.98093766e-01 8.22618306e-01 1.08734298e+00 -2.78750688e-01
3.44864815e-01 -1.25705093e-01 -2.88083553e-01 5.64372063e-01
1.43260717e-01 1.22316457e-01 -6.62814975e-01 -2.90411443e-01
1.17136919e+00 -5.24783619e-02 -2.42093965e-01 -8.34285080e-01
-1.22170651e+00 9.07797873e-01 3.39137703e-01 -3.69784027e-01
-4.87506777e-01 4.78373945e-01 -2.99005628e-01 -2.09847420e-01
3.89563829e-01 3.09888035e-01 -2.19427720e-01 -7.05423802e-02
-8.01862001e-01 6.19997025e-01 8.92639816e-01 1.37846065e+00
8.05626214e-01 1.44418508e-01 -9.22315270e-02 2.90648580e-01
4.30531561e-01 1.16841424e+00 -1.46442547e-01 -1.15138137e+00
3.39185357e-01 5.48905768e-02 1.91468224e-01 -7.57430673e-01
7.26995096e-02 -2.08591029e-01 -7.31666446e-01 2.53438294e-01
5.34034856e-02 -3.40009421e-01 -1.29327881e+00 1.96009040e+00
3.14205915e-01 5.06202102e-01 -3.72345094e-03 7.60426939e-01
8.18422318e-01 9.38976824e-01 -9.16075334e-02 1.39187247e-01
9.05181706e-01 -6.06078088e-01 -3.78739446e-01 -3.39036822e-01
-2.91340873e-02 -8.87307107e-01 8.70922208e-01 3.17361474e-01
-1.06215632e+00 -3.98694307e-01 -7.68648148e-01 -2.53156334e-01
1.00029193e-01 -1.22943586e-02 9.14744198e-01 6.65115654e-01
-1.16112649e+00 1.48920724e-02 -1.13742340e+00 -1.26252636e-01
2.87579596e-01 1.07226506e-01 -4.15126950e-01 -4.44035888e-01
-5.31521261e-01 5.33868313e-01 -9.84837022e-03 -1.14697397e-01
-1.59527624e+00 -8.09162378e-01 -1.13715935e+00 -5.61959036e-02
2.69653380e-01 -1.44962776e+00 1.16129529e+00 -4.94716972e-01
-1.66959429e+00 8.83069217e-01 -3.64945561e-01 -3.86234708e-02
3.85922551e-01 -4.11367923e-01 -2.70322543e-02 1.80760548e-01
5.46739511e-02 6.07749641e-01 1.00034440e+00 -1.52588165e+00
-1.80047706e-01 -1.60008535e-01 2.54119784e-01 3.41490567e-01
4.54879373e-01 -4.16340888e-01 -7.84077764e-01 -4.27572846e-01
8.64255130e-02 -1.22133934e+00 -4.38527733e-01 -1.53697774e-01
-8.27115417e-01 4.14922893e-01 7.20597327e-01 -2.72957176e-01
4.80473101e-01 -2.14354587e+00 7.72220433e-01 1.00581810e-01
2.76057959e-01 -5.18202066e-01 -1.63954318e-01 2.94652581e-01
-1.15851209e-01 -5.40828817e-02 -7.89052919e-02 -8.57778013e-01
2.55375534e-01 3.69351923e-01 -6.42558098e-01 5.56645632e-01
2.20029488e-01 9.00940835e-01 -7.17987716e-01 -1.08344890e-01
7.41875589e-01 9.87241447e-01 -9.07405436e-01 5.82247078e-01
-4.95834529e-01 8.57969642e-01 -3.92696738e-01 3.84807259e-01
8.81181538e-01 -4.97309774e-01 1.94595456e-01 -6.76576421e-02
1.52858391e-01 2.65562117e-01 -1.27994776e+00 2.03884387e+00
-5.64917147e-01 5.28690398e-01 -4.78112139e-02 -3.94746959e-01
6.83366954e-01 2.53684789e-01 5.03441513e-01 -1.76770180e-01
2.26041481e-01 -3.16024482e-01 -6.60888791e-01 -2.79071122e-01
5.04092634e-01 -3.41038108e-01 -4.24707353e-01 3.56920987e-01
3.98213863e-01 -8.28606248e-01 -2.12896422e-01 6.25683725e-01
9.81469572e-01 5.44670761e-01 2.66592540e-02 -8.19770098e-02
-1.60043404e-01 -3.05541277e-01 1.36624426e-01 9.34119701e-01
5.36448538e-01 1.02920258e+00 4.43593949e-01 -2.16613889e-01
-1.20767844e+00 -1.89765143e+00 -6.03346042e-02 5.56357443e-01
8.99047852e-02 -3.99798155e-01 -4.00525093e-01 -3.47806334e-01
-3.23727541e-02 9.73632634e-01 -8.03657413e-01 -5.58910482e-02
-1.77523702e-01 -9.20023084e-01 2.90642064e-02 4.84336078e-01
1.08213522e-01 -4.64749694e-01 -4.39589977e-01 -1.69057518e-01
-2.35726073e-01 -1.46295571e+00 -2.24796638e-01 2.70606458e-01
-6.15831017e-01 -8.62191617e-01 -4.36039418e-01 -2.13373810e-01
8.04638326e-01 3.72124344e-01 1.54398417e+00 -5.53370893e-01
-4.99492973e-01 9.33706820e-01 -1.32376850e-01 -2.11656302e-01
-2.07146987e-01 -1.50151506e-01 -1.42110446e-02 7.03227967e-02
1.31702006e-01 -9.22351003e-01 -4.87486213e-01 5.95397269e-03
-8.98545444e-01 5.49168468e-01 3.49319369e-01 4.98756975e-01
8.43677163e-01 -1.23528130e-01 -3.49638641e-01 -9.32437479e-01
-1.13909133e-01 -6.98643565e-01 -8.38284254e-01 -4.34882902e-02
1.79465741e-01 2.72539079e-01 3.36179227e-01 -4.67666686e-01
-1.24253678e+00 3.98735613e-01 -1.91663939e-03 -6.63281024e-01
-5.33194542e-01 7.16992542e-02 -5.71816802e-01 2.06593230e-01
4.81259316e-01 1.55301437e-01 -6.13905191e-01 -5.01351535e-01
7.48127759e-01 2.20146999e-02 7.68526375e-01 -9.14662302e-01
1.15179408e+00 8.13704014e-01 5.20505369e-01 -8.23552310e-01
-9.51933861e-01 -3.06721538e-01 -7.12496161e-01 -2.27622744e-02
1.24702990e+00 -1.41182137e+00 -6.10434055e-01 4.49324459e-01
-1.07464182e+00 -5.87661862e-01 -3.34454149e-01 6.39135242e-01
-9.01189029e-01 -3.97439301e-02 -4.34701174e-01 -8.43541026e-01
4.47800934e-01 -1.10175109e+00 1.79278564e+00 1.92495529e-02
-1.46287322e-01 -1.04160857e+00 3.98428440e-01 1.32891402e-01
7.39020854e-02 7.17006862e-01 7.23445654e-01 3.43519419e-01
-1.42888665e+00 -2.43785948e-01 2.65613571e-02 2.18625680e-01
2.17815250e-01 1.71951339e-01 -1.29966092e+00 -2.26877183e-01
8.32853541e-02 -3.20740312e-01 7.90521741e-01 7.84352005e-01
1.04480445e+00 -1.80076867e-01 -1.49841666e-01 1.29365134e+00
1.71777070e+00 -3.27241629e-01 7.96546698e-01 -2.60537744e-01
1.08163071e+00 1.96389034e-01 -9.02527347e-02 5.63967526e-01
4.20000196e-01 6.79993749e-01 5.68392694e-01 -7.69094899e-02
-1.15995146e-02 -6.94797099e-01 3.47825319e-01 5.55742025e-01
-2.35437736e-01 -6.27183676e-01 -5.57636142e-01 4.73106384e-01
-1.63014710e+00 -1.10632980e+00 -8.10020417e-02 2.05511832e+00
4.04182881e-01 -1.44045547e-01 -1.35100439e-01 -2.96956062e-01
3.10049117e-01 4.60852236e-01 -3.66122425e-01 2.81361118e-02
-3.36674124e-01 3.65206778e-01 5.14001071e-01 7.69565761e-01
-1.04513276e+00 7.29685366e-01 7.70806503e+00 3.69967610e-01
-9.23229754e-01 -9.21783596e-02 5.55797935e-01 -5.73382676e-01
-8.17035735e-01 1.29713848e-01 -9.30823982e-01 1.81980252e-01
8.06600451e-01 4.42425758e-02 4.60644782e-01 8.77466500e-01
-2.02287268e-02 -2.00421005e-01 -1.39835739e+00 1.08802009e+00
4.33854163e-01 -1.32890415e+00 3.41229975e-01 3.86439115e-01
1.21270680e+00 1.70054302e-01 4.40370500e-01 2.52868440e-02
1.19632614e+00 -1.09279871e+00 9.75284636e-01 7.51918852e-01
8.33049417e-01 -5.24882615e-01 9.95405987e-02 3.60831589e-01
-8.47311080e-01 3.22433561e-01 -2.03671977e-01 -6.82462379e-02
5.64677179e-01 6.16372287e-01 -6.88567996e-01 5.25460064e-01
7.21736252e-01 6.85092986e-01 -4.35324073e-01 7.42481291e-01
-6.50311291e-01 6.13139212e-01 -6.02070212e-01 4.18041229e-01
1.16976604e-01 -3.11737686e-01 8.02468479e-01 9.78677273e-01
6.38495922e-01 1.81349590e-02 3.41169864e-01 1.22916698e+00
7.46642490e-06 -4.45328146e-01 -9.58836973e-01 2.94519216e-01
1.04760565e-01 1.13064384e+00 -4.19971704e-01 -1.68732122e-01
-3.16674471e-01 1.22203159e+00 1.62765235e-01 7.99851656e-01
-9.33314145e-01 3.54186922e-01 7.87242949e-01 4.91952896e-02
7.96521604e-01 -7.61505008e-01 -1.20749801e-01 -1.62378204e+00
-1.49524689e-01 -4.35657173e-01 6.99202716e-02 -1.33151567e+00
-1.32356441e+00 4.77000624e-01 4.23829466e-01 -1.08141971e+00
-7.12403238e-01 -6.82370603e-01 -4.86034244e-01 1.06454849e+00
-1.27316260e+00 -1.55484116e+00 -5.51078498e-01 8.75501812e-01
2.88827449e-01 1.68097317e-01 9.98023629e-01 2.21617948e-02
-3.16648036e-01 5.28982542e-02 9.35868844e-02 -1.62625313e-01
5.06368041e-01 -1.56962752e+00 5.07333994e-01 1.10812962e+00
6.14163518e-01 6.99157357e-01 8.74518394e-01 -4.67289507e-01
-1.66046262e+00 -1.03569639e+00 3.55549425e-01 -1.19727314e+00
3.99356067e-01 -9.38763559e-01 -2.30596304e-01 1.20750713e+00
9.29317027e-02 1.84810519e-01 8.04447711e-01 2.62520462e-01
-6.79316103e-01 2.12020487e-01 -8.12827766e-01 6.56142056e-01
1.01476085e+00 -7.72377014e-01 -9.31067616e-02 2.23803803e-01
7.96663642e-01 -8.05734813e-01 -5.29275000e-01 1.30588025e-01
3.09073567e-01 -9.96079087e-01 1.26166344e+00 -4.18428630e-01
6.50844634e-01 -2.44761840e-01 -8.00244093e-01 -1.20847499e+00
-4.32134986e-01 -7.38615930e-01 -2.68964946e-01 9.38353717e-01
2.06409499e-01 -1.24902748e-01 6.27591848e-01 6.88748240e-01
8.79290402e-02 -1.87827021e-01 -7.35927522e-01 -6.52975857e-01
1.93673633e-02 -7.35875428e-01 6.62417948e-01 6.32984936e-01
-8.79807472e-01 5.33111572e-01 -7.70137846e-01 6.10896945e-01
9.81629074e-01 3.37387323e-01 1.29662907e+00 -9.05134559e-01
-7.69112408e-01 3.55580181e-04 -6.01681888e-01 -1.56224644e+00
2.47589856e-01 -6.38974428e-01 2.00779304e-01 -1.42066419e+00
5.62357962e-01 -2.05306917e-01 2.61692047e-01 -5.69302291e-02
-1.16392560e-01 3.93909723e-01 1.70501411e-01 -2.51758769e-02
-6.67569160e-01 6.81740165e-01 1.22776365e+00 6.20991960e-02
-2.84756552e-02 -1.07306495e-01 -7.51072764e-01 6.64830804e-01
1.21487282e-01 -3.20043743e-01 -7.24491060e-01 -8.31633568e-01
4.40169185e-01 3.24767944e-03 1.00824249e+00 -8.40584397e-01
-8.78727138e-02 -4.07503814e-01 7.09039032e-01 -7.30781019e-01
7.69106209e-01 -8.45878303e-01 6.81748092e-01 -3.27143550e-01
-7.29320720e-02 -1.52364895e-01 1.92417666e-01 7.49515295e-01
6.82144761e-02 1.97422341e-01 5.29516459e-01 -1.83216989e-01
-5.00655234e-01 6.65442050e-01 -3.46192658e-01 3.40761282e-02
8.27136755e-01 4.90975156e-02 -1.09028079e-01 -7.86080480e-01
-7.94982255e-01 -1.62097260e-01 8.33848596e-01 3.04379553e-01
3.89713824e-01 -1.47868550e+00 -7.54965067e-01 3.82955045e-01
7.12973326e-02 5.16528249e-01 5.57620168e-01 2.27839887e-01
-5.65068603e-01 1.50607169e-01 4.97008078e-02 -1.11729193e+00
-1.00253844e+00 5.28292716e-01 1.55146495e-01 -1.38646692e-01
-7.82796085e-01 9.34368491e-01 8.09651375e-01 -4.70871270e-01
-1.50549799e-01 -2.96214730e-01 5.19667029e-01 -5.38552880e-01
3.46867710e-01 -7.67126232e-02 -3.66086066e-01 -8.66695285e-01
-1.41861752e-01 6.41362846e-01 1.65687233e-01 -4.52058047e-01
1.57686114e+00 -2.46521384e-01 1.53639764e-01 7.47829318e-01
1.38319767e+00 4.05683011e-01 -1.93544412e+00 -1.16834015e-01
-1.07900596e+00 -8.77138078e-01 1.88287452e-01 -5.82506716e-01
-8.88729811e-01 8.39207113e-01 2.71123052e-01 -1.89137414e-01
9.30461824e-01 4.36600685e-01 3.79817486e-01 1.51277781e-01
5.13398647e-01 -3.72327119e-01 1.91411108e-01 4.90753561e-01
7.47793734e-01 -1.07147706e+00 1.60074115e-01 -4.79031771e-01
-5.54169357e-01 6.14431441e-01 3.33649784e-01 -4.11425084e-01
7.58626580e-01 6.31246090e-01 -8.66251960e-02 -3.40951979e-01
-8.13932598e-01 -9.81672760e-03 4.91417170e-01 7.97982574e-01
2.63711870e-01 3.39137316e-01 8.68264496e-01 1.85488001e-01
-5.60258389e-01 -3.20515931e-01 4.99027103e-01 7.53194034e-01
1.17627278e-01 -9.49560165e-01 -4.05671179e-01 2.80318018e-02
-1.41953051e-01 -1.06535152e-01 -2.22007081e-01 5.71993887e-01
8.73581916e-02 5.05304098e-01 2.76430190e-01 -2.58958131e-01
9.64009911e-02 -1.30063549e-01 1.03770959e+00 -8.02963257e-01
2.67825246e-01 3.54726553e-01 -2.39480183e-01 -6.80786729e-01
-5.21671236e-01 -8.60546052e-01 -5.10605514e-01 -3.61061752e-01
-1.28408521e-01 -1.68602154e-01 6.04061782e-01 6.29438162e-01
2.65705645e-01 5.90455294e-01 7.44974554e-01 -1.36880302e+00
-1.33092850e-01 -6.41354501e-01 -7.71763742e-01 6.12333298e-01
6.27244115e-01 -7.54710913e-01 -6.27412319e-01 4.75850493e-01] | [9.257680892944336, -3.109995126724243] |
c9bb1bd9-b9c9-461e-93e9-c8afb07a9cf7 | clip2video-mastering-video-text-retrieval-via | 2106.11097 | null | https://arxiv.org/abs/2106.11097v1 | https://arxiv.org/pdf/2106.11097v1.pdf | CLIP2Video: Mastering Video-Text Retrieval via Image CLIP | We present CLIP2Video network to transfer the image-language pre-training model to video-text retrieval in an end-to-end manner. Leading approaches in the domain of video-and-language learning try to distill the spatio-temporal video features and multi-modal interaction between videos and languages from a large-scale video-text dataset. Different from them, we leverage pretrained image-language model, simplify it as a two-stage framework with co-learning of image-text and enhancing temporal relations between video frames and video-text respectively, make it able to train on comparatively small datasets. Specifically, based on the spatial semantics captured by Contrastive Language-Image Pretraining (CLIP) model, our model involves a Temporal Difference Block to capture motions at fine temporal video frames, and a Temporal Alignment Block to re-align the tokens of video clips and phrases and enhance the multi-modal correlation. We conduct thorough ablation studies, and achieve state-of-the-art performance on major text-to-video and video-to-text retrieval benchmarks, including new records of retrieval accuracy on MSR-VTT, MSVD and VATEX. | ['Yu Chen', 'Luhui Xu', 'Pengfei Xiong', 'Han Fang'] | 2021-06-21 | null | null | null | null | ['video-text-retrieval'] | ['computer-vision'] | [-9.82778426e-03 -9.59305406e-01 -6.80389047e-01 -3.51476908e-01
-1.06940305e+00 -6.87914729e-01 9.22805965e-01 -1.86901063e-01
-7.35371530e-01 8.64827856e-02 4.73606199e-01 -7.67564178e-02
1.32699773e-01 -2.11879864e-01 -9.21305597e-01 -4.79616314e-01
-1.20634422e-01 3.09332550e-01 2.59018421e-01 -1.76599212e-02
1.43860951e-01 2.07792013e-03 -1.40681148e+00 9.96873915e-01
1.31751150e-01 1.17218351e+00 3.40515643e-01 9.45344090e-01
-3.37494612e-01 1.31139266e+00 -8.52959324e-03 -3.09274316e-01
2.96785116e-01 -3.13539654e-01 -7.47155070e-01 1.37443310e-02
1.02098799e+00 -7.53605962e-01 -1.31041706e+00 5.66707969e-01
5.18975019e-01 3.08629274e-01 1.00137448e+00 -1.16883242e+00
-1.04884827e+00 5.11520743e-01 -8.29400241e-01 5.64182699e-01
6.51806176e-01 1.63249776e-01 1.19683099e+00 -1.26611197e+00
9.65961695e-01 1.62971842e+00 2.98157990e-01 3.91219348e-01
-8.40159237e-01 -7.39808321e-01 3.17894995e-01 7.17530191e-01
-1.64045835e+00 -4.97553051e-01 5.94575107e-01 -6.04486048e-01
1.19666171e+00 2.75658872e-02 7.19195843e-01 1.44385076e+00
2.92276353e-01 1.40444946e+00 4.61627960e-01 -1.45098567e-01
-4.30709094e-01 -2.12278500e-01 -3.19431275e-01 5.94424605e-01
-5.09854436e-01 1.27235860e-01 -1.01740730e+00 2.91863263e-01
6.72531784e-01 3.52709681e-01 -1.14837520e-01 -3.66928816e-01
-1.46012390e+00 5.86496353e-01 1.78157359e-01 4.75230038e-01
-1.62979111e-01 5.51048934e-01 8.66623938e-01 5.43014050e-01
3.53853166e-01 -1.91267699e-01 -2.39979118e-01 -2.40877524e-01
-1.41064858e+00 1.45740479e-01 4.27673519e-01 1.21117723e+00
6.79184496e-01 1.26046643e-01 -6.92359686e-01 7.22318947e-01
3.82000953e-01 9.75256264e-01 5.84214389e-01 -6.70488358e-01
7.88544595e-01 1.53320596e-01 -4.69714582e-01 -1.12535977e+00
1.18651696e-01 2.42913261e-01 -7.34390616e-01 -7.97081172e-01
-9.43803191e-02 3.08361113e-01 -1.06312525e+00 1.41362107e+00
-1.74052224e-01 6.12536609e-01 3.15313153e-02 9.54257846e-01
1.03950489e+00 1.06601477e+00 1.79115981e-01 -2.96250045e-01
1.20143020e+00 -1.34706497e+00 -8.14340651e-01 -1.10999532e-01
8.80998373e-01 -9.04150963e-01 1.12492132e+00 -7.14476556e-02
-1.27518225e+00 -7.95917809e-01 -5.77881753e-01 -6.08518958e-01
-4.53678161e-01 1.10414252e-01 1.58704266e-01 -1.71654075e-01
-1.11237478e+00 1.93993643e-01 -7.39136517e-01 -3.84764910e-01
3.31933349e-01 9.91962031e-02 -4.40978199e-01 -4.96699125e-01
-1.32807708e+00 4.82223958e-01 5.38501918e-01 1.76635049e-02
-1.32068741e+00 -7.47683585e-01 -1.05441725e+00 -2.79161464e-02
4.46598917e-01 -7.33568251e-01 8.95155847e-01 -1.26093709e+00
-1.17808175e+00 1.16186523e+00 -4.53195781e-01 -4.35123414e-01
5.61689079e-01 -4.18660343e-01 -5.27125180e-01 8.86051357e-01
2.11056128e-01 1.07415664e+00 1.31225002e+00 -8.55925500e-01
-7.36123741e-01 -8.46288651e-02 2.51760595e-02 5.24627090e-01
-6.53489172e-01 4.17079180e-01 -1.50680709e+00 -1.04335845e+00
-2.98987538e-01 -9.94051456e-01 4.23920363e-01 4.34613675e-02
-1.01390935e-01 -1.93725929e-01 1.21248782e+00 -7.10906088e-01
1.39957404e+00 -2.46261168e+00 4.81439114e-01 -1.36305332e-01
1.84088692e-01 8.62931535e-02 -7.88303018e-01 5.47549188e-01
-1.32642984e-01 7.76162371e-02 3.43174934e-01 -4.97473568e-01
-8.78905505e-02 1.08523823e-01 -5.75064659e-01 4.63534713e-01
1.57105267e-01 1.42787409e+00 -8.63128126e-01 -1.01640749e+00
5.27298510e-01 5.89591086e-01 -6.54589295e-01 3.20206016e-01
-2.48086572e-01 2.14841440e-01 -5.61094165e-01 8.03409755e-01
3.05858463e-01 -3.63981605e-01 -1.00725085e-01 -4.87568021e-01
1.34431988e-01 8.73855781e-03 -7.59391665e-01 2.20828176e+00
-3.93479049e-01 1.09876490e+00 -2.06786469e-01 -1.02620399e+00
1.96464583e-01 4.69317466e-01 9.96790111e-01 -1.12331283e+00
1.03144804e-02 9.21433792e-02 -5.34598529e-01 -9.49704587e-01
6.59813821e-01 3.42100501e-01 -1.17892988e-01 2.68767327e-01
4.61217701e-01 -2.32244451e-02 3.69327307e-01 6.30201638e-01
8.37997079e-01 2.44551688e-01 -1.54111028e-01 6.88703870e-03
6.68454349e-01 -2.43984550e-01 -7.29329884e-02 7.05243409e-01
-3.71966735e-02 7.69740939e-01 6.51686341e-02 -3.44991535e-01
-9.83511150e-01 -9.75416064e-01 2.50857234e-01 1.64833748e+00
3.89605969e-01 -7.77183235e-01 -1.78456664e-01 -5.25287449e-01
-2.62105204e-02 1.97320461e-01 -5.19913018e-01 -2.11209685e-01
-7.39440143e-01 -9.15880427e-02 7.52591431e-01 4.51937526e-01
5.05089760e-01 -8.85423660e-01 -9.29188132e-02 -5.48736341e-02
-5.84906042e-01 -1.89567804e+00 -1.36730158e+00 -2.57925302e-01
-4.77514893e-01 -8.93844962e-01 -1.09502196e+00 -1.18798947e+00
1.75664946e-01 7.41978645e-01 1.20015073e+00 1.97747275e-01
-3.45979065e-01 1.07737935e+00 -7.23965585e-01 1.96140260e-01
-1.47306267e-02 -1.07621476e-01 6.90536126e-02 9.24572200e-02
3.38473558e-01 -3.46354932e-01 -6.97466314e-01 3.08966309e-01
-1.33112037e+00 -9.70286429e-02 5.82427323e-01 8.23997200e-01
6.47661984e-01 -2.47101516e-01 -1.29856363e-01 -2.64142513e-01
9.28183496e-02 -5.70276737e-01 -2.76234210e-01 6.14533663e-01
-2.90361773e-02 -1.59871534e-01 2.98950911e-01 -9.16044116e-01
-5.48528612e-01 1.13510368e-02 2.24081859e-01 -1.54684746e+00
2.22970724e-01 6.88761771e-01 1.73098683e-01 -3.10624298e-02
5.02737127e-02 8.14311862e-01 -3.49640578e-01 -5.85673936e-02
7.08963871e-01 4.87129658e-01 6.99725866e-01 -6.96167111e-01
8.72547090e-01 6.51750922e-01 -1.60598770e-01 -9.87407267e-01
-6.70533895e-01 -9.84725714e-01 -8.32144737e-01 -2.81954437e-01
1.31597626e+00 -1.64241683e+00 -5.23937166e-01 3.95073533e-01
-1.14309394e+00 -4.29502428e-01 7.48229101e-02 6.14003420e-01
-6.47298336e-01 6.77350700e-01 -8.01295459e-01 -2.19255269e-01
-2.28802651e-01 -1.19596291e+00 1.62406230e+00 -2.31582254e-01
3.35380524e-01 -1.13045704e+00 -1.16162911e-01 3.37956727e-01
1.61563680e-01 -3.41733396e-01 8.16848993e-01 -5.80279708e-01
-8.11870635e-01 3.15854698e-02 -3.38629723e-01 2.19283417e-01
-1.48216531e-01 1.49183840e-01 -6.36123419e-01 -6.99309170e-01
-3.60436201e-01 -7.41250873e-01 1.26067269e+00 4.63994503e-01
1.37408257e+00 -1.88886628e-01 -3.27116936e-01 9.19036269e-01
1.26822758e+00 1.91563312e-02 5.82171321e-01 2.41873860e-01
1.06606865e+00 3.47348690e-01 6.49469256e-01 1.79073706e-01
7.53711224e-01 9.66760993e-01 6.26173168e-02 7.13066198e-03
-3.13236743e-01 -3.92238766e-01 9.03623939e-01 1.20399487e+00
1.57068238e-01 -4.63596821e-01 -8.26946437e-01 6.11076236e-01
-2.00800753e+00 -1.30949855e+00 3.65570575e-01 1.65738297e+00
6.40836358e-01 -1.97057128e-01 4.64363545e-02 -5.54518521e-01
4.85465288e-01 6.38842523e-01 -4.62671489e-01 2.66923547e-01
-4.57981467e-01 -1.70933574e-01 3.94286156e-01 3.68274748e-01
-1.24775064e+00 1.42025018e+00 6.03472996e+00 1.23590946e+00
-1.41288567e+00 1.39920846e-01 4.53874618e-01 -5.98816633e-01
-2.18092874e-01 -1.74517810e-01 -7.60982394e-01 3.16841245e-01
8.77049387e-01 -2.44222969e-01 4.17872339e-01 4.43123579e-01
2.23022789e-01 2.52903014e-01 -1.54928100e+00 1.55136240e+00
5.48188567e-01 -1.57310295e+00 7.85748124e-01 -1.71232983e-01
7.12067962e-01 3.28903556e-01 2.11005196e-01 6.31798029e-01
-1.93333760e-01 -9.76107061e-01 1.02568543e+00 3.42896909e-01
1.13214397e+00 -2.68747568e-01 3.48163098e-01 -5.34978993e-02
-1.83102727e+00 -1.55788772e-02 -8.32354575e-02 5.36562145e-01
2.63423890e-01 -3.98571603e-02 -2.42779776e-01 6.89099491e-01
1.11748707e+00 1.62810934e+00 -6.24684989e-01 5.56681633e-01
3.10160428e-01 2.17201740e-01 -1.81280196e-01 2.89417744e-01
6.67243481e-01 -1.07547669e-02 4.72688824e-01 1.68863297e+00
2.35544726e-01 2.32516583e-02 4.86977279e-01 4.18942481e-01
-2.88718820e-01 2.11729169e-01 -6.81706429e-01 -5.55518746e-01
3.18070769e-01 9.24645662e-01 -4.21769589e-01 -5.39526284e-01
-9.43268061e-01 1.25853932e+00 5.51246991e-03 8.19614947e-01
-9.51368034e-01 1.27774075e-01 4.90951896e-01 -1.95794273e-02
6.54033840e-01 -4.60683674e-01 6.76512003e-01 -1.57694268e+00
1.22702800e-01 -1.18336809e+00 6.82308555e-01 -9.86708581e-01
-1.26589978e+00 4.97957408e-01 1.91507548e-01 -1.36487126e+00
-4.60906833e-01 -5.22862494e-01 -1.43502340e-01 4.69433069e-01
-1.79517388e+00 -1.58028781e+00 -1.20764926e-01 1.38769126e+00
1.09583926e+00 -4.05149072e-01 3.22546035e-01 7.84607589e-01
-2.44234920e-01 7.24644005e-01 2.84960151e-01 5.03684521e-01
1.18511903e+00 -6.15284145e-01 1.19922422e-01 7.67649889e-01
6.38788521e-01 4.11641479e-01 5.78261837e-02 -5.15556872e-01
-2.05840755e+00 -1.22693646e+00 4.54418480e-01 -5.10495424e-01
1.04453313e+00 -5.43943882e-01 -7.37478793e-01 9.17344689e-01
5.33748865e-01 2.97708422e-01 4.06432390e-01 -3.14958423e-01
-6.48333490e-01 5.01437485e-02 -1.60866633e-01 7.48650551e-01
1.08171296e+00 -1.47176969e+00 -4.73988116e-01 4.67436850e-01
9.43266988e-01 -5.81490636e-01 -8.03048849e-01 3.72486413e-01
7.68960774e-01 -4.65757400e-01 1.39138639e+00 -7.72812009e-01
8.19307745e-01 -9.25744176e-02 -5.96693516e-01 -6.86421335e-01
-5.53128868e-02 -7.20621288e-01 -2.57970363e-01 1.13723648e+00
9.98463184e-02 2.39106506e-01 3.33047509e-01 1.81227829e-02
3.14249806e-02 -4.71919864e-01 -9.16225374e-01 -6.80902958e-01
7.58121535e-02 -6.64934993e-01 -1.96154997e-01 1.08631110e+00
-2.00899556e-01 3.92398745e-01 -8.48939657e-01 -5.01786172e-02
2.39743754e-01 9.89436284e-02 7.92905211e-01 -5.46293259e-01
-3.03784400e-01 -4.87637997e-01 -3.42461795e-01 -1.73752844e+00
7.41725504e-01 -1.05181348e+00 5.58302030e-02 -1.06644964e+00
5.96295714e-01 1.53031752e-01 -3.91632587e-01 2.54126251e-01
-1.11539900e-01 4.10016030e-01 5.57875276e-01 6.11430705e-01
-1.32427084e+00 7.03029215e-01 1.36927807e+00 -6.06468379e-01
-2.76089124e-02 -6.36061013e-01 1.51260907e-03 4.30351973e-01
-1.26375034e-01 -3.85960430e-01 -5.72050929e-01 -1.04040265e+00
2.14088768e-01 4.20666605e-01 3.61721307e-01 -7.13282645e-01
4.92328167e-01 -1.67519018e-01 2.61525661e-01 -9.83187914e-01
6.50368214e-01 -9.15020347e-01 -2.18347728e-01 9.37894434e-02
-7.57755816e-01 4.09978330e-01 3.25989068e-01 8.18025589e-01
-7.15419710e-01 2.65996575e-01 3.78410906e-01 -3.70388106e-02
-1.22868466e+00 8.78298819e-01 -3.52726251e-01 3.45491171e-01
9.16800082e-01 -6.41467199e-02 -2.78780133e-01 -7.85331368e-01
-5.34359694e-01 7.48066723e-01 2.97361284e-01 1.13378930e+00
8.47576022e-01 -1.42858303e+00 -7.15969145e-01 1.50583507e-02
4.43972617e-01 -2.85426617e-01 5.44494271e-01 8.70527506e-01
-3.61794174e-01 8.71175170e-01 1.09265074e-02 -1.15380299e+00
-1.45662403e+00 7.93078125e-01 2.28090256e-01 -2.47074902e-01
-7.59296894e-01 7.89685369e-01 6.34318352e-01 1.66216165e-01
6.22523606e-01 -2.95390815e-01 -4.56378683e-02 1.61827415e-01
6.34402752e-01 -2.21380755e-01 -3.47754091e-01 -1.09789324e+00
-3.67514640e-01 1.09835744e+00 -3.84062827e-01 -1.75589129e-01
1.00815725e+00 -5.36915779e-01 -2.42074523e-02 5.19797385e-01
1.93174946e+00 -3.27101886e-01 -1.06585240e+00 -6.28471434e-01
-1.74519196e-01 -5.52901566e-01 1.86516583e-01 -2.52456516e-01
-1.25664926e+00 9.93994713e-01 5.97808063e-01 -3.90309215e-01
1.18802190e+00 1.17154084e-01 9.69823122e-01 7.04319179e-01
-8.78580362e-02 -1.19027841e+00 8.12975347e-01 7.98728645e-01
8.08921099e-01 -1.51539898e+00 1.76393893e-02 -8.00569206e-02
-7.21735477e-01 1.14559841e+00 4.19765800e-01 -3.00867623e-03
8.24080288e-01 2.22951267e-02 2.34637093e-02 -1.10175982e-01
-1.13186443e+00 -2.52496094e-01 8.32430005e-01 2.13141754e-01
3.69760394e-01 -4.53301728e-01 1.24244273e-01 -1.85271110e-02
3.92597228e-01 7.79313818e-02 -3.74883451e-02 8.43649089e-01
1.26056194e-01 -8.81019294e-01 9.21455920e-02 1.75396308e-01
-5.48332512e-01 -5.77521801e-01 -2.31156364e-01 9.61624384e-01
-3.61275882e-01 6.33845985e-01 3.79842162e-01 -5.17064512e-01
4.06400412e-02 -5.16142696e-02 4.82295960e-01 -3.89120162e-01
-3.70367408e-01 6.12747967e-01 -2.31417358e-01 -9.39473689e-01
-9.48834062e-01 -4.84253377e-01 -1.03296554e+00 -2.82701820e-01
2.31244862e-02 -5.30560315e-02 3.94582272e-01 1.15438128e+00
4.76921856e-01 4.29293007e-01 5.99295080e-01 -1.09774446e+00
-3.17119732e-02 -7.41521776e-01 -4.43140954e-01 7.39680469e-01
5.89608312e-01 -3.99152011e-01 -2.40908563e-01 5.58864892e-01] | [10.339366912841797, 0.9456770420074463] |
7c216eb9-4dde-4e4f-ab00-ce80aa588dd3 | on-pursuit-of-designing-multi-modal | 2109.06085 | null | https://arxiv.org/abs/2109.06085v2 | https://arxiv.org/pdf/2109.06085v2.pdf | On Pursuit of Designing Multi-modal Transformer for Video Grounding | Video grounding aims to localize the temporal segment corresponding to a sentence query from an untrimmed video. Almost all existing video grounding methods fall into two frameworks: 1) Top-down model: It predefines a set of segment candidates and then conducts segment classification and regression. 2) Bottom-up model: It directly predicts frame-wise probabilities of the referential segment boundaries. However, all these methods are not end-to-end, i.e., they always rely on some time-consuming post-processing steps to refine predictions. To this end, we reformulate video grounding as a set prediction task and propose a novel end-to-end multi-modal Transformer model, dubbed as GTR. Specifically, GTR has two encoders for video and language encoding, and a cross-modal decoder for grounding prediction. To facilitate the end-to-end training, we use a Cubic Embedding layer to transform the raw videos into a set of visual tokens. To better fuse these two modalities in the decoder, we design a new Multi-head Cross-Modal Attention. The whole GTR is optimized via a Many-to-One matching loss. Furthermore, we conduct comprehensive studies to investigate different model design choices. Extensive results on three benchmarks have validated the superiority of GTR. All three typical GTR variants achieve record-breaking performance on all datasets and metrics, with several times faster inference speed. | ['Yuexian Zou', 'Can Zhang', 'Mike Zheng Shou', 'Long Chen', 'Meng Cao'] | 2021-09-13 | null | https://aclanthology.org/2021.emnlp-main.773 | https://aclanthology.org/2021.emnlp-main.773.pdf | emnlp-2021-11 | ['video-grounding'] | ['computer-vision'] | [ 3.16027910e-01 -8.07389617e-02 -5.66277504e-01 -5.15270710e-01
-1.26784825e+00 -2.65681863e-01 4.06293958e-01 -1.26161858e-01
-3.08584124e-01 3.62737745e-01 3.93749505e-01 -2.10691974e-01
3.84305894e-01 -6.34531140e-01 -1.22449231e+00 -5.52769899e-01
2.89360523e-01 2.00777620e-01 3.96326244e-01 -4.66086715e-02
1.44319296e-01 -2.69586205e-01 -1.28083467e+00 7.12472022e-01
7.45478809e-01 1.32026064e+00 4.48342115e-01 4.98404086e-01
1.98493991e-03 9.44675624e-01 -1.25906482e-01 -6.81165874e-01
-1.56362709e-02 -5.38006365e-01 -7.41377115e-01 2.15398729e-01
2.76150197e-01 -4.14583236e-01 -8.46515834e-01 1.02301049e+00
4.28797752e-01 1.02352351e-01 3.42074513e-01 -1.24417496e+00
-8.17138553e-01 8.71126592e-01 -7.12919772e-01 1.05786704e-01
7.33160973e-01 1.74318790e-01 1.19679868e+00 -1.06588852e+00
4.23716128e-01 1.22656119e+00 5.61023474e-01 4.38340247e-01
-9.28689420e-01 -7.02773750e-01 5.63550532e-01 5.21480083e-01
-1.63872194e+00 -4.25761014e-01 7.15173662e-01 -4.82493788e-01
7.64554560e-01 1.69280931e-01 5.39043188e-01 1.07322979e+00
8.31573159e-02 1.14849186e+00 7.38638699e-01 -1.29806876e-01
-2.17271253e-01 -1.87923566e-01 -1.16323940e-01 9.50656295e-01
-2.60565013e-01 -9.33893397e-02 -9.41567004e-01 1.91803291e-01
5.69969833e-01 7.72130564e-02 -5.73977232e-01 -4.73516397e-02
-1.60723066e+00 7.17757523e-01 4.34602827e-01 1.47787407e-01
-2.84147769e-01 2.29317755e-01 5.64521790e-01 1.75819457e-01
4.97026414e-01 -2.05369398e-01 -1.95767418e-01 -8.49752948e-02
-1.19391584e+00 1.32750362e-01 3.03970426e-01 1.15887380e+00
8.91738892e-01 -3.75678807e-01 -7.70263076e-01 7.52843857e-01
6.22879326e-01 2.61659950e-01 6.04027510e-01 -5.67925692e-01
1.13849354e+00 4.45918471e-01 -3.01587060e-02 -1.12761140e+00
-7.80846626e-02 -2.17690080e-01 -8.20017576e-01 -4.91191715e-01
1.76679883e-02 -5.84792830e-02 -1.01084948e+00 1.67646980e+00
2.32586220e-01 8.32563579e-01 -6.77038059e-02 1.23443007e+00
9.45888042e-01 1.10024810e+00 1.55605018e-01 -2.22866535e-01
1.44659281e+00 -1.40560174e+00 -7.29200065e-01 -2.43938759e-01
6.08179986e-01 -6.49848640e-01 1.16508758e+00 1.88066512e-01
-1.17832005e+00 -7.25892961e-01 -1.00829673e+00 -4.33355004e-01
1.41998222e-02 1.87418878e-01 1.27054662e-01 9.35833603e-02
-8.90078187e-01 3.93040419e-01 -8.26701462e-01 -3.34378667e-02
3.15077633e-01 2.23683968e-01 -3.32765907e-01 -2.59366632e-01
-1.32912362e+00 5.80972373e-01 5.61965227e-01 3.32360357e-01
-9.56703186e-01 -5.35905898e-01 -1.10345519e+00 1.10273115e-01
4.37736571e-01 -8.41449380e-01 1.15862346e+00 -8.56246531e-01
-1.43541479e+00 1.04379272e+00 -6.55901551e-01 -5.56001782e-01
4.96504903e-01 -3.87502760e-01 -5.03896892e-01 2.06758916e-01
2.32012555e-01 7.54862666e-01 8.00421953e-01 -1.09245372e+00
-8.86772871e-01 -4.07759063e-02 1.22463815e-01 4.38926011e-01
-2.91628838e-01 7.36815929e-02 -1.19929910e+00 -9.13136959e-01
1.96539938e-01 -7.54485965e-01 3.71284559e-02 -2.37410188e-01
-6.97221696e-01 -1.69716015e-01 5.90634167e-01 -1.02456582e+00
1.69008911e+00 -2.21460390e+00 3.83135378e-01 -6.14979118e-02
2.12697968e-01 1.02695441e-02 -1.20137617e-01 2.86128163e-01
-5.96331879e-02 1.74156334e-02 -1.29533216e-01 -5.28314948e-01
1.48989722e-01 -1.96702946e-02 -4.51533645e-01 5.03936112e-01
1.81733355e-01 1.08553922e+00 -8.82526159e-01 -8.93409193e-01
2.40739986e-01 4.09567982e-01 -7.02086210e-01 4.62377369e-01
-3.29564810e-01 3.39047521e-01 -4.22517061e-01 6.12047076e-01
3.04113269e-01 -3.81897926e-01 -8.64140019e-02 -5.09963095e-01
4.33406904e-02 2.77833194e-01 -8.98897588e-01 2.13156462e+00
-3.39410663e-01 5.77890456e-01 -3.36764425e-01 -1.17005539e+00
7.88944483e-01 4.68108982e-01 5.24099767e-01 -6.93169475e-01
6.11226000e-02 7.11587369e-02 -4.15647268e-01 -7.90346742e-01
5.85557997e-01 -1.29354507e-01 -2.05294624e-01 1.09801382e-01
1.41306534e-01 4.35887545e-01 2.03763377e-02 1.09600306e-01
8.12471449e-01 5.44248581e-01 -4.87330835e-03 1.98708341e-01
6.75286889e-01 -1.99162737e-01 8.16377044e-01 3.38317096e-01
-6.49370924e-02 9.42698658e-01 4.98292059e-01 -2.77399629e-01
-8.68444800e-01 -8.51971209e-01 2.67916679e-01 1.23874569e+00
6.67369604e-01 -6.70764983e-01 -8.20287824e-01 -6.36103153e-01
-3.43618006e-01 6.02140665e-01 -6.17769718e-01 -3.25337559e-01
-6.71174347e-01 -4.71464276e-01 3.82609755e-01 5.14378250e-01
6.39244437e-01 -9.05606627e-01 -1.67126596e-01 2.18569696e-01
-8.81479979e-01 -1.37117302e+00 -1.10498071e+00 -3.83102357e-01
-5.83782792e-01 -9.72369134e-01 -8.02070618e-01 -1.14385939e+00
4.95283782e-01 3.44395220e-01 1.04803693e+00 2.14253470e-01
4.20019478e-01 5.78495450e-02 -6.47955537e-01 9.73811224e-02
-1.27051637e-01 2.05374405e-01 -2.51270086e-01 5.21929622e-01
3.93028438e-01 -2.21498013e-01 -7.87523687e-01 5.73207855e-01
-7.21841514e-01 5.72171867e-01 5.90148270e-01 7.31256306e-01
1.13083422e+00 -1.70231417e-01 3.46618831e-01 -5.53033710e-01
2.72458166e-01 -5.97627699e-01 -3.87609690e-01 4.90920961e-01
-1.90578446e-01 -5.31509109e-02 6.67189062e-01 -2.92098731e-01
-7.12755084e-01 1.09455645e-01 -2.75080204e-01 -9.34040427e-01
1.27324432e-01 7.25086868e-01 -4.24455732e-01 2.82814950e-01
-1.56712625e-02 5.42449951e-01 -2.80771583e-01 -3.50934297e-01
3.47230226e-01 7.28108108e-01 7.42868066e-01 -4.75959033e-01
8.12761605e-01 1.82450935e-01 -4.30086166e-01 -5.09540379e-01
-1.07115865e+00 -4.08583194e-01 -3.99641961e-01 -3.82914096e-01
1.31291032e+00 -1.17463350e+00 -6.52251065e-01 4.05293226e-01
-1.11550295e+00 -4.93049055e-01 1.08980075e-01 5.48477411e-01
-7.43937910e-01 4.27240491e-01 -6.36879742e-01 -3.82475734e-01
-3.73060524e-01 -1.35526848e+00 1.33215809e+00 1.61685288e-01
1.92298830e-01 -6.30136549e-01 -1.58745885e-01 6.19077861e-01
-1.31866336e-01 1.64827898e-01 6.16832972e-01 -4.11737442e-01
-7.49144852e-01 -1.15387902e-01 -3.81617337e-01 1.13695666e-01
-1.99760571e-01 -5.72195984e-02 -8.03433716e-01 -3.84408295e-01
-3.03974509e-01 -2.89833754e-01 1.03827739e+00 3.33720028e-01
1.58181977e+00 -4.10558462e-01 -4.39036846e-01 9.10943449e-01
1.34028924e+00 2.25085318e-01 8.14893782e-01 3.35880101e-01
1.08960235e+00 3.45894277e-01 1.04861999e+00 2.52236485e-01
7.67045677e-01 9.31335151e-01 4.61580098e-01 -9.88589823e-02
-5.30285649e-02 -7.77638137e-01 6.03162766e-01 9.89467442e-01
4.00250144e-02 -5.48603654e-01 -6.37710154e-01 4.58334506e-01
-2.11878443e+00 -1.16922152e+00 1.55894428e-01 2.13619328e+00
8.40109468e-01 2.45677039e-01 2.80657470e-01 -3.75271440e-02
1.12137282e+00 4.08472121e-01 -5.52605212e-01 1.46858901e-01
3.89226750e-02 -3.09194893e-01 2.40621746e-01 4.85029548e-01
-1.27351153e+00 1.05965638e+00 5.23457909e+00 9.82753158e-01
-1.19591141e+00 2.22204596e-01 9.10154343e-01 -4.35058139e-02
-4.48457181e-01 2.64966651e-03 -8.82485926e-01 8.78234208e-01
7.85727680e-01 -8.21365342e-02 3.79189849e-01 5.21797478e-01
4.48445320e-01 3.42039049e-01 -1.08846354e+00 1.29469025e+00
2.02890873e-01 -1.50538826e+00 1.74435183e-01 -3.02837104e-01
4.21717018e-01 7.60125071e-02 -2.26950184e-01 4.62300807e-01
-2.24744096e-01 -9.82131541e-01 1.32125151e+00 5.32094121e-01
1.01345825e+00 -7.37960935e-01 6.44938886e-01 1.64878115e-01
-1.65670109e+00 -1.73789337e-02 -2.30442941e-01 2.72010148e-01
5.73316157e-01 2.28774324e-01 -2.46277645e-01 7.00769842e-01
6.53159618e-01 9.95401680e-01 -4.13584232e-01 1.08085752e+00
-1.71568811e-01 6.54839814e-01 -7.68049955e-02 2.27956831e-01
3.81367087e-01 -9.78021696e-02 3.73208910e-01 1.32877815e+00
4.87582475e-01 8.81921202e-02 3.53356481e-01 5.44128060e-01
-2.84968883e-01 4.75801900e-02 -1.89473271e-01 1.64823011e-01
5.77912033e-01 9.97376561e-01 -5.77839136e-01 -4.16054100e-01
-8.05524945e-01 1.18918204e+00 2.51583666e-01 4.64937389e-01
-1.48368466e+00 -5.06161392e-01 3.72834116e-01 2.35968322e-01
5.35630822e-01 -4.82776463e-02 -8.75398424e-03 -1.52131295e+00
2.12021813e-01 -9.01957870e-01 4.94016886e-01 -9.29837704e-01
-9.81877387e-01 6.40672743e-01 6.52483804e-03 -1.48252904e+00
2.61138473e-03 -3.65358680e-01 -5.51634729e-01 6.16089106e-01
-1.59589040e+00 -1.36900687e+00 -5.26653230e-01 7.27948129e-01
7.64069319e-01 5.70094995e-02 3.30169946e-01 6.78488672e-01
-9.04851615e-01 9.13962364e-01 -2.40484625e-01 4.53904659e-01
6.38252139e-01 -9.42068458e-01 3.88308406e-01 1.15063524e+00
2.30195716e-01 3.74739826e-01 4.85615969e-01 -4.25641149e-01
-1.37927449e+00 -1.50165331e+00 9.25656796e-01 -3.11996769e-02
5.74667335e-01 -3.71768326e-01 -9.58892822e-01 9.64564681e-01
2.24490076e-01 7.53228888e-02 4.66657609e-01 -1.59153223e-01
-2.86024123e-01 -2.06146613e-01 -5.82614124e-01 6.40495718e-01
1.11265433e+00 -7.48532116e-01 -5.16518176e-01 3.05725396e-01
1.17274165e+00 -7.88694382e-01 -9.61380839e-01 3.68172199e-01
5.21559179e-01 -8.50574791e-01 9.59518015e-01 -4.84652370e-01
8.11857462e-01 -5.12953103e-01 -3.76659930e-01 -8.76197159e-01
-3.94877017e-01 -8.49984586e-01 -3.45563471e-01 1.47734797e+00
4.17600095e-01 -2.57830858e-01 7.21785367e-01 2.93617308e-01
-5.95172524e-01 -1.21271193e+00 -8.25615585e-01 -5.44023633e-01
-2.41505250e-01 -5.59243619e-01 7.28912830e-01 8.07296395e-01
-4.09537926e-02 5.46380579e-01 -6.43562973e-01 2.58096576e-01
3.96830022e-01 3.72699767e-01 6.20454371e-01 -4.59013104e-01
-3.98452997e-01 -3.77633214e-01 -3.23998958e-01 -1.66496944e+00
2.13237971e-01 -9.21717286e-01 3.19936454e-01 -1.59656537e+00
3.57784152e-01 -2.68995404e-01 -4.54205960e-01 5.03220916e-01
-4.68673974e-01 3.48871410e-01 2.13373855e-01 3.83142531e-01
-9.54082549e-01 7.24948645e-01 1.22722018e+00 -1.83764413e-01
-5.99155948e-02 -1.80280834e-01 -6.04166090e-01 6.14406407e-01
5.69537520e-01 -3.46936226e-01 -4.54522759e-01 -8.04739892e-01
2.44967520e-01 4.38260168e-01 4.57971126e-01 -9.88857508e-01
1.45472661e-01 -2.80979007e-01 4.85019088e-02 -9.00326669e-01
3.16382140e-01 -6.26665711e-01 6.09241240e-02 1.50678530e-01
-5.22468448e-01 1.89829618e-01 -1.26410052e-01 6.90790892e-01
-5.87560415e-01 4.02040891e-02 7.58179247e-01 2.17898354e-01
-1.01555169e+00 8.59756827e-01 -1.49186980e-02 2.94998914e-01
1.17896152e+00 -2.51459837e-01 -2.57101983e-01 -3.44923258e-01
-6.41853213e-01 4.04185414e-01 4.21252877e-01 5.87881267e-01
6.57634795e-01 -1.63054419e+00 -7.36430943e-01 3.34509765e-03
1.49558216e-01 1.58711970e-01 3.75217259e-01 1.05625522e+00
-4.98583168e-01 3.67081255e-01 1.46573082e-01 -7.01972187e-01
-1.12458754e+00 7.88072646e-01 3.69640917e-01 -1.86221182e-01
-7.53168702e-01 1.02524495e+00 4.63901579e-01 1.88473195e-01
3.38090092e-01 -4.22961950e-01 -1.39385298e-01 5.47650941e-02
6.06732607e-01 1.95793808e-02 -2.35721380e-01 -9.59104657e-01
-3.55571032e-01 7.32591391e-01 4.27298900e-03 -3.07357572e-02
1.09270942e+00 -4.54626739e-01 6.26170710e-02 4.32052702e-01
1.49336874e+00 -3.31274152e-01 -1.46534228e+00 -3.19655597e-01
-2.61707157e-01 -4.69803065e-01 3.98029722e-02 -3.32608044e-01
-1.26556098e+00 9.09174681e-01 2.18692750e-01 2.16492712e-02
1.29782104e+00 2.81434190e-02 1.33816671e+00 -2.96785738e-02
1.97630480e-01 -8.84169459e-01 1.51164696e-01 3.24715137e-01
8.52702558e-01 -1.13936543e+00 -2.54114747e-01 -4.02733088e-01
-7.26422012e-01 8.47085893e-01 6.95811689e-01 7.05003440e-02
4.04536217e-01 -1.36967540e-01 -2.32025847e-01 5.81990741e-02
-7.90948033e-01 -1.40361458e-01 4.99423236e-01 2.69579470e-01
4.13916081e-01 -1.24889843e-01 -2.59167194e-01 8.56058598e-01
-9.29064751e-02 2.60862112e-01 1.54703349e-01 4.63037789e-01
-4.01995927e-01 -7.53829181e-01 -1.69878200e-01 4.02590752e-01
-4.70574826e-01 -8.57602507e-02 1.86343238e-01 4.81639296e-01
2.18067139e-01 9.77790833e-01 1.66201033e-02 -9.23301578e-01
3.46443415e-01 -9.04553756e-02 2.57739693e-01 -4.57572520e-01
-3.68927121e-01 1.80925250e-01 -1.02348346e-02 -8.04309666e-01
-4.25017864e-01 -6.00062966e-01 -1.25195885e+00 -3.48794401e-01
-4.28945541e-01 2.03242436e-01 1.24944165e-01 1.02359998e+00
4.57286566e-01 6.06202006e-01 6.50476456e-01 -9.05608714e-01
-1.96082681e-01 -7.71485448e-01 -1.86726272e-01 5.27156174e-01
3.07037890e-01 -5.70271075e-01 4.03208360e-02 4.68375623e-01] | [10.148066520690918, 0.7188199758529663] |
c69fe40e-f5c8-4ea6-8c5e-350a0b5d47fd | maps-a-noise-robust-progressive-learning | 2302.04589 | null | https://arxiv.org/abs/2302.04589v1 | https://arxiv.org/pdf/2302.04589v1.pdf | MAPS: A Noise-Robust Progressive Learning Approach for Source-Free Domain Adaptive Keypoint Detection | Existing cross-domain keypoint detection methods always require accessing the source data during adaptation, which may violate the data privacy law and pose serious security concerns. Instead, this paper considers a realistic problem setting called source-free domain adaptive keypoint detection, where only the well-trained source model is provided to the target domain. For the challenging problem, we first construct a teacher-student learning baseline by stabilizing the predictions under data augmentation and network ensembles. Built on this, we further propose a unified approach, Mixup Augmentation and Progressive Selection (MAPS), to fully exploit the noisy pseudo labels of unlabeled target data during training. On the one hand, MAPS regularizes the model to favor simple linear behavior in-between the target samples via self-mixup augmentation, preventing the model from over-fitting to noisy predictions. On the other hand, MAPS employs the self-paced learning paradigm and progressively selects pseudo-labeled samples from `easy' to `hard' into the training process to reduce noise accumulation. Results on four keypoint detection datasets show that MAPS outperforms the baseline and achieves comparable or even better results in comparison to previous non-source-free counterparts. | ['Ran He', 'Aihua Zheng', 'Bo Jiang', 'Jian Liang', 'Yuhe Ding'] | 2023-02-09 | null | null | null | null | ['keypoint-detection'] | ['computer-vision'] | [ 3.01784933e-01 -6.86220219e-03 -7.46865749e-01 -3.96199733e-01
-1.16747093e+00 -8.36317897e-01 6.45494163e-01 2.96631306e-01
-5.19387841e-01 7.72350490e-01 1.08050175e-01 -2.22860113e-01
1.85102969e-01 -4.52938706e-01 -7.08619237e-01 -7.55451322e-01
1.87349349e-01 2.98476219e-01 3.49375606e-01 -7.64760822e-02
-1.18142352e-01 2.18994409e-01 -1.25129306e+00 1.00469068e-01
9.98183608e-01 1.13713872e+00 -1.58961758e-01 5.33149183e-01
4.66379151e-02 5.42114198e-01 -5.70362329e-01 -3.56353223e-01
6.50150955e-01 -2.78359443e-01 -3.98947626e-01 -2.22908095e-01
5.68142593e-01 -3.93556148e-01 -4.07230824e-01 1.26156163e+00
6.00613832e-01 1.82428695e-02 4.31548238e-01 -1.59788108e+00
-7.92798400e-01 3.64221752e-01 -6.53093994e-01 4.47013043e-02
1.95511550e-01 -7.82887638e-02 8.42706263e-01 -8.73756588e-01
5.25498509e-01 7.70954669e-01 6.41689837e-01 8.26974213e-01
-1.44616711e+00 -1.22084427e+00 4.04329419e-01 4.79088724e-02
-1.33295965e+00 -5.17863929e-01 1.01681840e+00 -2.17093259e-01
2.26798326e-01 3.07980567e-01 2.79333383e-01 1.45739770e+00
-3.00639510e-01 1.05342019e+00 1.11661136e+00 -4.42973614e-01
3.93112570e-01 6.10439003e-01 2.76127756e-01 2.88847625e-01
1.89200148e-01 2.61703700e-01 -7.08730996e-01 -7.27768600e-01
5.11102200e-01 -6.24660752e-04 -4.23091292e-01 -9.63050902e-01
-1.20854235e+00 6.96149528e-01 2.25973606e-01 -1.07409991e-01
-2.05707327e-01 -3.11268151e-01 5.13527513e-01 5.36852598e-01
5.20951748e-01 3.70851904e-01 -9.37229335e-01 2.09674701e-01
-8.90335739e-01 3.67236197e-01 5.39938867e-01 1.14378130e+00
7.75363505e-01 -1.73034564e-01 -1.97673246e-01 8.27026904e-01
1.48313701e-01 3.92190039e-01 7.55557120e-01 -5.72327554e-01
6.49420142e-01 6.05390489e-01 2.66765207e-01 -5.85860312e-01
-2.56197359e-02 -5.46952963e-01 -8.66587460e-01 2.03611761e-01
5.82098782e-01 -2.63243347e-01 -1.26565337e+00 2.16830516e+00
6.95244431e-01 3.07132304e-01 2.57861495e-01 6.99016929e-01
7.06078827e-01 4.79020089e-01 4.48101342e-01 -3.26762736e-01
1.33454466e+00 -7.20663846e-01 -5.49580634e-01 -2.78045505e-01
5.70332050e-01 -3.89764190e-01 1.30878115e+00 3.41819644e-01
-6.92295313e-01 -4.45910007e-01 -1.17695475e+00 1.68477207e-01
-5.39247572e-01 1.14110909e-01 3.13015610e-01 7.25296438e-01
-7.89878011e-01 2.28416085e-01 -6.11045122e-01 -1.43785104e-01
5.20619631e-01 4.31686431e-01 -5.63635826e-01 -6.62455484e-02
-1.50858378e+00 4.47048336e-01 4.08057421e-01 -4.62598354e-01
-7.75140762e-01 -1.06901395e+00 -7.47305870e-01 -1.17605455e-01
7.13553548e-01 -5.36656976e-01 1.30331087e+00 -1.10521829e+00
-1.55200291e+00 8.24287236e-01 -4.29525450e-02 -4.56941277e-01
7.36197352e-01 -4.36702639e-01 -4.58574057e-01 -1.23525970e-01
1.44215152e-01 7.29012907e-01 1.22791886e+00 -1.40750778e+00
-8.01035047e-01 -3.12956512e-01 -2.60715336e-01 2.96793193e-01
-6.32306337e-01 -1.73202127e-01 -7.15277016e-01 -8.90209734e-01
-6.52761310e-02 -9.11707878e-01 -2.43534058e-01 6.53631017e-02
-6.87341094e-01 -2.53623933e-01 9.67767417e-01 -2.92098731e-01
1.17419255e+00 -2.35576463e+00 -2.76515812e-01 5.07976234e-01
3.28370064e-01 4.64523375e-01 -2.38484740e-01 1.41866207e-01
-4.23239678e-01 -1.38429135e-01 -1.29029691e-01 -4.59791422e-01
-9.03763250e-02 5.38472719e-02 -5.90472519e-01 5.08151233e-01
-1.79486393e-04 5.99026024e-01 -8.52545321e-01 -3.79538000e-01
-2.99051981e-02 2.75465161e-01 -4.88254845e-01 3.31439823e-01
-2.45210886e-01 5.35755932e-01 -5.43212056e-01 6.47837043e-01
9.36126649e-01 -3.14888924e-01 -4.80703041e-02 4.60859351e-02
1.83614582e-01 2.40576506e-01 -1.39899564e+00 1.52718329e+00
-9.94916409e-02 2.14556858e-01 2.80654699e-01 -9.33558404e-01
8.64210963e-01 4.07821327e-01 5.15937984e-01 -5.60370505e-01
-8.29730406e-02 3.81981917e-02 -3.71152461e-01 -3.60544696e-02
5.20963311e-01 1.36872008e-01 -2.25286141e-01 4.21549588e-01
-6.33861646e-02 4.60066438e-01 -2.89195448e-01 3.40000451e-01
9.56621587e-01 9.08015296e-02 4.23981160e-01 -2.61373688e-02
5.64369857e-01 -3.02351192e-02 9.63972092e-01 8.57457519e-01
-5.73493421e-01 6.27934575e-01 4.51127470e-01 -1.13662653e-01
-8.10204387e-01 -1.10834157e+00 -4.96449992e-02 1.54809570e+00
3.16359162e-01 -4.07939017e-01 -7.49101818e-01 -1.30346656e+00
2.06099197e-01 5.62557101e-01 -6.98107839e-01 -4.46435004e-01
-4.34419811e-01 -6.97032988e-01 5.55947244e-01 5.01642048e-01
5.33688128e-01 -6.03186846e-01 -1.77239209e-01 8.57762918e-02
-7.89768919e-02 -9.38777149e-01 -7.32758999e-01 5.63048661e-01
-5.03597260e-01 -9.65267837e-01 -9.06173289e-01 -6.30587339e-01
7.13439703e-01 3.00156325e-01 7.88651168e-01 -3.27212811e-01
-2.25719605e-02 4.43538904e-01 -2.12932169e-01 -6.32418573e-01
-4.15532887e-01 2.74657995e-01 3.14837426e-01 7.98838511e-02
6.50909483e-01 -4.90329981e-01 -6.55778944e-01 3.72202158e-01
-6.98634744e-01 -2.88200259e-01 5.72838485e-01 9.63902652e-01
7.22036362e-01 -1.45140365e-01 8.63134682e-01 -9.78047669e-01
6.58273458e-01 -6.48505569e-01 -6.45366728e-01 3.26150477e-01
-9.53598022e-01 -2.58545969e-02 7.70746589e-01 -1.11485016e+00
-9.39047456e-01 5.08865476e-01 1.01278305e-01 -7.68793702e-01
-2.47751772e-01 -7.77477995e-02 -5.88278770e-01 -2.36717075e-01
9.11070347e-01 5.60879827e-01 3.45683470e-02 -4.64426547e-01
3.90582889e-01 6.99208975e-01 7.98082054e-01 -5.90320230e-01
1.27659416e+00 3.07025433e-01 -3.91756594e-01 -5.42451203e-01
-7.09232390e-01 -6.12916708e-01 -5.08562088e-01 3.00958544e-01
3.99754971e-01 -1.17953527e+00 -3.99535865e-01 5.05438209e-01
-6.50961876e-01 -2.50081867e-01 -5.77451706e-01 3.52948695e-01
-3.50268364e-01 3.43894184e-01 -2.81204224e-01 -7.45864868e-01
-4.60721791e-01 -8.59835625e-01 9.09188449e-01 2.19522297e-01
-1.90234154e-01 -8.30534041e-01 1.22862525e-01 3.78169841e-03
2.49784470e-01 1.63216904e-01 6.89492762e-01 -1.35828483e+00
-5.14255881e-01 -5.88247061e-01 -1.66479513e-01 4.43954617e-01
2.32694507e-01 -4.76553887e-01 -1.17588329e+00 -3.85346949e-01
-5.68228737e-02 -5.64016640e-01 7.29870737e-01 1.10272646e-01
1.16815174e+00 -2.84433275e-01 -6.38598323e-01 6.66156709e-01
1.10229206e+00 -5.02110575e-04 1.95264056e-01 5.04157722e-01
5.37640572e-01 4.07481194e-01 7.75173247e-01 4.23390031e-01
4.76602405e-01 4.49723542e-01 1.12640478e-01 -3.04858208e-01
1.54474497e-01 -7.58631527e-01 2.23024815e-01 -1.16737425e-01
4.77594912e-01 -1.39968649e-01 -7.90015996e-01 5.17961264e-01
-1.58656085e+00 -6.39538646e-01 2.07018137e-01 2.53857970e+00
1.22869122e+00 1.50480285e-01 4.57860529e-01 1.55362755e-01
7.18418062e-01 4.81677651e-02 -8.32266152e-01 2.37437293e-01
-1.89839881e-02 2.90360823e-02 8.58205914e-01 2.73851782e-01
-1.44791484e+00 9.47160244e-01 6.34903240e+00 9.02092457e-01
-1.37584615e+00 1.92486599e-01 6.94653749e-01 -5.78026958e-02
-1.03191979e-01 -2.26249248e-01 -1.08537579e+00 5.43438196e-01
9.04179871e-01 -1.61125824e-01 1.44658938e-01 1.25141621e+00
-2.22777918e-01 1.13143370e-01 -1.24682724e+00 9.00348604e-01
-1.32198602e-01 -9.91804957e-01 -1.62133530e-01 1.60311922e-01
4.97529805e-01 1.00190200e-01 4.14339513e-01 3.74640077e-01
7.26687551e-01 -4.38240081e-01 6.68500304e-01 1.07725434e-01
1.07326698e+00 -6.99894309e-01 2.62419283e-01 7.03333497e-01
-1.02491319e+00 -1.93862811e-01 -2.22755760e-01 5.45468926e-01
-2.98492551e-01 3.77691150e-01 -9.32885706e-01 3.03398758e-01
7.04872549e-01 4.07922715e-01 -5.61727047e-01 9.21722412e-01
-1.95654497e-01 7.51234829e-01 -5.58640301e-01 2.66992509e-01
-9.07628089e-02 2.65389234e-01 6.16711080e-01 1.05943012e+00
1.12960733e-01 -2.46118288e-02 3.06725919e-01 5.07550359e-01
-3.60778749e-01 2.12909818e-01 -5.04044771e-01 1.80767968e-01
8.49495769e-01 1.04926360e+00 -3.70561808e-01 -4.29171562e-01
-4.70431864e-01 1.02857220e+00 3.23883086e-01 5.66616595e-01
-6.13727510e-01 -3.77522647e-01 9.47735488e-01 1.13463946e-01
1.51972398e-01 7.37506673e-02 -2.89561152e-01 -1.22784007e+00
-5.77039726e-04 -1.01398444e+00 7.63790011e-01 -2.03716382e-01
-1.42317975e+00 4.30785865e-01 3.15352306e-02 -1.53814673e+00
-2.11116508e-01 -3.02764535e-01 -5.98880708e-01 9.82587636e-01
-1.78976345e+00 -1.21655357e+00 -1.17823660e-01 9.99269903e-01
2.26624519e-01 -1.79149210e-01 9.22836602e-01 2.99657345e-01
-5.12838960e-01 1.44343269e+00 2.78336227e-01 1.97547317e-01
1.34956217e+00 -1.36914635e+00 4.59143937e-01 8.84675860e-01
9.14231408e-03 7.62191594e-01 5.80216229e-01 -7.46516347e-01
-1.08484769e+00 -1.23640954e+00 5.91077209e-01 -6.32614791e-01
5.18225431e-01 -8.16940188e-01 -1.29028320e+00 6.34503484e-01
-1.00151084e-01 4.84212160e-01 1.01054537e+00 8.89435858e-02
-8.83568585e-01 -2.96624511e-01 -1.44153821e+00 3.94906133e-01
7.49976516e-01 -5.20094991e-01 -4.95076597e-01 2.12330848e-01
6.89420760e-01 -4.63403553e-01 -6.66852236e-01 3.07310671e-01
3.64187121e-01 -5.32539487e-01 1.23145139e+00 -6.16672456e-01
-2.41399705e-01 -2.52402782e-01 -6.77272454e-02 -1.27603281e+00
-2.84267664e-01 -8.51516008e-01 -3.81083488e-01 1.45269191e+00
3.35027844e-01 -7.16352701e-01 1.39212692e+00 8.53333414e-01
2.57320255e-01 -5.49545050e-01 -1.02062988e+00 -7.33454525e-01
2.17824414e-01 -4.16563243e-01 6.97054327e-01 1.20516920e+00
-5.43587022e-02 2.19903588e-01 -6.44533932e-01 5.18548608e-01
8.42818379e-01 -1.77175447e-01 1.09330380e+00 -1.38436985e+00
-1.63221195e-01 -4.88083139e-02 -3.67599800e-02 -1.45226884e+00
1.44825146e-01 -6.39776111e-01 1.42068649e-02 -7.38118351e-01
-8.58276561e-02 -8.02171052e-01 -6.92214906e-01 6.14948213e-01
-6.07594192e-01 1.96266890e-01 -7.50447344e-03 3.77180040e-01
-6.99346483e-01 4.01246965e-01 9.46096063e-01 7.18459347e-03
-6.18544400e-01 3.93032372e-01 -1.15432286e+00 5.73208690e-01
6.37583792e-01 -5.11978328e-01 -7.43642926e-01 1.30723536e-01
-1.56603128e-01 -2.07899332e-01 2.20186949e-01 -7.82045066e-01
4.11964059e-01 -2.89316952e-01 7.24411011e-01 -4.96854216e-01
3.01641583e-01 -1.02285218e+00 -3.32549661e-01 2.52492577e-01
-6.52540565e-01 -3.88100624e-01 2.40765840e-01 1.00786948e+00
4.77206856e-02 1.27104223e-01 1.01160932e+00 1.76075116e-01
-6.41591012e-01 6.43800974e-01 8.52896646e-03 1.76620543e-01
1.21847928e+00 -2.66706616e-01 -2.03814998e-01 -3.35972965e-01
-6.48842692e-01 4.22867924e-01 4.57767278e-01 4.91239429e-01
5.12543380e-01 -1.31666934e+00 -4.51493233e-01 7.67545998e-01
5.03296018e-01 1.58237040e-01 5.32239266e-02 4.48524177e-01
3.19524914e-01 1.73288137e-01 2.59381741e-01 -5.42391956e-01
-1.46256328e+00 9.04123902e-01 2.84193754e-01 -2.28867278e-01
-5.67737341e-01 1.04491448e+00 3.28932941e-01 -5.74439764e-01
6.57534719e-01 -1.81683347e-01 -9.13616866e-02 6.24421425e-02
6.74589455e-01 1.28641650e-01 -7.43455961e-02 -4.57091242e-01
-3.37244332e-01 1.73947304e-01 -6.79116964e-01 1.45569772e-01
1.09935033e+00 -1.59915358e-01 4.85988528e-01 2.44911253e-01
1.10285842e+00 2.18265995e-01 -1.43835270e+00 -7.47082710e-01
8.49775597e-02 -5.19497156e-01 -4.43967804e-02 -9.28821266e-01
-7.12320447e-01 5.54360509e-01 8.90474319e-01 1.73860520e-01
1.27372515e+00 2.17962340e-02 7.59665251e-01 3.37719411e-01
1.03651188e-01 -1.18949330e+00 4.71574254e-02 1.76561475e-01
3.87635589e-01 -1.49982631e+00 -1.89679354e-01 -3.73486370e-01
-7.26349711e-01 6.73417509e-01 8.25449109e-01 -5.34465946e-02
7.34161794e-01 2.65579969e-01 2.72168875e-01 3.50132942e-01
-7.96169460e-01 1.50600120e-01 3.67445856e-01 8.70716751e-01
-5.72256744e-02 -1.64481461e-01 2.41568014e-01 8.19131553e-01
-1.89483780e-02 -1.06263487e-02 7.58475438e-02 9.20356154e-01
-2.26187721e-01 -1.32727802e+00 -6.40729010e-01 4.20076609e-01
-6.27240896e-01 -2.44055428e-02 -5.90837479e-01 8.33018482e-01
1.78537935e-01 5.46280324e-01 -1.19463623e-01 -4.94266480e-01
4.58176196e-01 3.60684156e-01 -1.57838941e-01 -5.06432056e-01
-5.37126780e-01 4.01848108e-02 -3.19271803e-01 -4.12281066e-01
-2.11681868e-03 -7.06505775e-01 -9.60846603e-01 -5.17134778e-02
-4.70467389e-01 4.05034959e-01 4.83890891e-01 5.58405399e-01
5.91539621e-01 2.61607379e-01 6.90519631e-01 -3.87782395e-01
-1.01997638e+00 -7.68758893e-01 -4.53664362e-01 4.47878122e-01
9.45416689e-01 -5.81694305e-01 -4.54099029e-01 -5.12344800e-02] | [10.393665313720703, 3.1871140003204346] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.